ENHANCING THE METABOLIC CAPACITY OF CYANOBACTERIA FOR BIOLOGICAL HYDROGEN PRODUCTION: BIOFUEL APPLICATIONS OF CYANOTHECE AND ARTHROSPIRA SPP.

Nicholas John Skizim

A DISSERTATION PRESENTED TO THE FACULTY OF PRINCETON UNVERSITY IN CANDIDACY FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

RECOMMENDED FOR ACCEPTANCE BY THE DEPARTMENT OF CHEMISTRY Advisor: G. Charles Dismukes

April 2012

 Copyright Nicholas John Skizim, 2012. All rights reserved.

ENHANCING THE METABOLIC CAPACITY OF CYANOBACTERIA FOR BIOLOGICAL HYDROGEN PRODUCTION: BIOFUEL APPLICATIONS OF CYANOTHECE AND ARTHROSPIRA SPP.

Nicholas John Skizim Advisor: G. Charles Dismukes

Abstract The dissertation presented herein focuses on understanding the biological pathways leading to hydrogen production in cyanobacteria and optimizing their capacity to funnel intracellular reductant to hydrogen. With increasing global population and the continued development of previously unindustrialized nations, it is now more important than ever to meet the worldwide demand for energy to allow economies to continue to develop and grow. Coupling this demand to the decreasing availability of our finite reserves of fossil fuels and the increasing global climate, the development of clean and renewable energy sources is clearly a critical societal need. The biological production of hydrogen from cyanobacteria, photosynthetic bacteria which solely require minimal nutrients and sunlight to grow, is one promising approach to meeting this need. The energy-dense fuel hydrogen is completely carbon-free and can be used in combustion engines or for electricity generation through fuel cells. The production of hydrogen from nitrogen-fixing cyanobacteria (e.g. Cyanothece spp.) and non-nitrogen-fixing cyanobacteria (e.g. Arthrospira spp.) is explored within this dissertation.

iii

In Chapter 1, we characterize hydrogen production from the nitrogen-fixing (diazotrophic) cyanobacterium Cyanothece sp. Miami BG 043511. This organism is capable of producing H2 via the bidirectional hydrogenase as well as the nitrogenase complex. We illustrate that dark, autofermentative H2 production is catalyzed by the hydrogenase, whereas photo-fermentative H2 production (photo-H2) is catalyzed by nitrogenase and is dependent upon excitation of Photosystem I. Internal carbohydrate reserves supply the reductant for both pathways; this reductant is therefore shared between the two H2 producing modes. We illustrate that by allowing increased time for autofermentation, the available reductant (generated from carbohydrate catabolism) is increased and an increased photo-H2 production rate is observed. For photo-H2, this reductant is supplied to Photosystem I via non-photochemical reduction of the plastoquinone pool by either of two NADH dehydrogenases. By inhibiting the class of NADH dehydrogenase that does not pump protons, we observe a 2-fold increase in photo-H2 production rate, due to the production of additional ATP (via the proton gradient). In Chapter 2, we advance our understanding of Cyanothece sp. Miami BG 043511 and demonstrate methods to increase its catabolic rate. Internal carbohydrate catabolism serves as the source of reductant for biological H2 production from this strain, and by increasing the rate of catabolism in the strain we are able to supply increased reductant to hydrogenase and nitrogenase and subsequently observe increased H2 yield. We demonstrate that the addition of ATP sinks to the cell under fermentative conditions (e.g. cyanophycin synthesis and/or amino acid import) causes an increased redox poise of the cell, due to increased glycolytic flux to generate ATP by substrate level phosphorylation. While increased H2 production is observed under conditions showing a moderate increase in redox poise, further increased redox stress causes the cell to increase production of ethanol and formate, fermentative pathways that recycle NADH to NAD+. The growth of Cyanothece in medium with increased osmotic strength is shown to cause hyperaccumulation of osmolytes during photosynthetic growth and the subsequent hypotonic salt stress
iv

of these cells by transfer to low salt media correlates with increased photo-H2 production (maximal increase 9-fold). Thirdly, co-fermentation of Cyanothece with the lactic acid producing cyanobacterium Synechococcus sp. PCC 7002 (WT and lactate dehydrogenase overexpression mutant) causes increased H2 production from the mixed culture. A maximal increase of 2.5-fold is observed from the mixed culture of Cyanothece and the overexpression mutant; this increase is attributed to (a) providing Cyanothece with increased substrate (lactate) for photo-H2 production and (b) removing extracellular lactate, leading to an increase in the catabolic rate of the Synechococcus strain. In Chapter 3, we present a two channel instrument that allows the simultaneous detection of dissolved H2 (measured electrochemically) and intracellular NAD(P)H concentration (measured by NAD(P)H fluorescence). This tool is applied to the cyanobacterium Arthrospira maxima and utilized to study its metabolic response under fermentative conditions to the availability of the macronutrient nitrate. Nitrate is shown to compete with protons for intracellular reductant; a clock-like delay is observed where all extracellular nitrate must be consumed prior to the onset of fermentative H2 production. Moreover, nitrate is shown to induce a metabolic switch from the glycolytic to the oxidative pentose phosphate (OPP) pathway. This pathway generates more reductant per glucose equivalent catabolized, but provides this reductant in the form of NADPH rather than NADH, the latter being the obligate substrate for hydrogenase. A mechanism for the utilization of this NADPH by a membrane associated nitrate reductase is presented. This pathway would not only utilize NADPH for nitrate reduction, but would also pump protons and lead to PMF dependent ATP generation (so called nitrate respiration). This is the first observation of nitrate respiration in cyanobacteria and supports the intense competition between nitrate and proton reduction in Arthrospira maxima. In Chapter 4, we continue our studies with Arthrospira maxima and illustrate methods to increase the yield of autofermentative H2 production by eliminating H2 backpressure. We define the
v

continuous removal of H2 as milking and illustrate that by selective milking of H2 by its electrochemical consumption produces an increase in H2 yield (11-fold) and rate (3.4-fold). Smaller increases are observed when H2 is milked non-selectively by dilution of the biomass in the incubation media (3.7-fold yield increase, 3.1-fold rate increase). The addition of a mixture of excreted carbon fermentative products (lactate, acetate, and ethanol) that form in competition with other NADH sinks is shown to increase H2 yield by 1.4-fold. We attribute this to shifting the reversible reaction equilibria toward substrate (NADH) and thereby increasing the availability of NADH to hydrogenase. Photosynthetic activity of the cultures after an exhaustive 7 day autofermentation is demonstrated to be completely recoverable by low intensity light incubation, allowing regeneration of carbohydrate reserves by subsequent growth/fermentation cycles. Lastly, in Chapter 5, we illustrate degenerate primers to amplify a fragment of the large subunit of the bidirectional hydrogenase ([NiFe] hydrogenase) from cyanobacteria. Two established limitations of [NiFe] hydrogenases are their oxygen sensitivity and relatively slow turnover rates. These primers will allow the screening of large collections of cyanobacterial culture collections for novel hydrogenases that may not suffer from these limitations. The method described herein was applied to screen a collection of 24 axenic strains of cyanobacteria (from the University of Hawaii Culture Collection) and 10 strains were identified as containing hydrogenase. Two such strains (which not only exhibited presence of hydrogenase but also grew well in liquid culture) were shown to possess both in vitro and in vivo hydrogenase activity. Throughout the thesis we illustrate methods to increase the biological production of H2 from the biotechnologically relevant strains of the genera Cyanothece and Arthrospira presented herein. We demonstrate physiological approaches to redirect metabolism for increased H2 yield and identify future targets (physiological and genetic) to be explored to further increase biofuel production from these strains and bring them closer to the realization of their biotechnological potential.
vi

Acknowledgements First and foremost, I want to thank my advisor, Charles Dismukes, for the opportunity to perform research in his lab over the past 5 years and for exposing me to the exciting fields of photosynthesis and bioenergy sciences. Chuck has instilled in me a great excitement for discovery, high expectations for myself and others, and a highly critical mindset when analyzing results. Additionally, the opportunities Ive had to attend international conferences and government funding agency reviews, to participate in grant writing and manuscript reviewing, and to extensively collaborate with and mentor other students have undoubtedly strengthened me as a scientist; for all these opportunities I have Chuck to thank. I am forever indebted to Gennady Ananyev for teaching me the art of instrument design and for being such a close collaborator and friend over my time in graduate school. I would never have guessed that my time at Princeton would involve building an electrical circuit for a hydrogen electrode, but sure enough, I learned that and many other techniques from Gennady that I will carry with me for the rest of my scientific career. I thank all the members of the Dismukes lab (past and present) for making the lab an exciting, welcoming, supportive, and fun environment and for their collaborations on various scientific projects: Derrick Kolling, Damian Carrieri, Kelsey McNeely, Nicholas Bennette, Elizabeth Burrows, Tiago Guerra, David Vinyard, David Robinson, Anagha Krishnan, Xiao Qian, Graeme Gardner, Paul Smith, Clyde Cady, Kumar Kenchappa, Mario Cornejo, Jorge Montalvo, and Rob Brimblecombe. I have learned a multitude of skills from each of these individuals and thank them all for helping to make this thesis a reality. Additionally I have had the pleasure of mentoring a number of undergraduate scientists who have contributed to the works presented here and would like to acknowledge them as well: Jonathan Yao, Jeremy Priestner, Michael Abrams, Kandyce Perry, Tom Giordano, and Jennifer Sun.

vii

This journey would not have been possible without the constant support from my family and friends, especially my wife Siobhan, who has always offered a non-scientists perspective on my projects and provided much needed advice and motivation throughout the PhD program in general. The constant support from Siobhan, my family, and my friends completing their own journeys through graduate school have made this thesis possible. Besides these individuals in the Dismukes lab, Id especially like to thank Kate Keets, Ben Brooks, Ashley Donovan, Pearl Dickerson, Tom Umile, Dave Grizzanti, Chase Zachary, Maria Korolev, Piper Harron, Rob Harron, Amanda Tricarico, and Amber Hibberd. This work would not have been possible without the gracious funding from various agencies and institutions. The Air Force Office of Scientific Research is acknowledged for funding the BioSolarH2 multi-university research initiative (AFOSR-MURI-FA9550-05-1-0365), as well as the Department of Energy Genomes to Life program (DE-PS02-07ER07-13) and Princeton University Department of Chemistry for summer support. I would also like to thank Istvan Pelczer, John Eng, and Sallie Dunner at Princeton University, as well as Melissa Grunwerg at Rutgers University, for support throughout my time in graduate school. In Chapter 1, we thank Todd Michael and Randy Kerstetter for their contributions towards obtaining genomic data on Cyanothece, and we thank Kelsey McNeely and Nicholas Bennette for aid in reviewing the manuscript. In Chapter 2, L. Tiago Guerra is personally acknowledged for helpful discussions about the hypotonic salt stress response in Synechococcus and for the development of the soluble sugar extraction method used therein. In Chapter 3, we thank David Vinyard, Kelsey McNeely, and Josh Rabinowitz for careful reading of the manuscript and suggesting improvements. In Chapter 4 we acknowledge Damian Carrieri for helpful discussions and for providing information regarding carbohydrate composition in Arthrospira maxima and Elizabeth Burrows for editorial assistance. In Chapter 5, we acknowledge funding of E.H.B. by the Busch-Waksman Postdoctoral Fellowship, of D.R.J.K. by the American Chemical Society Alternative Energy Postdoctoral
viii

Fellowship and the Dreyfus Postdoctoral Fellowship in Environment Chemistry, of M.A. by the Williams College Center for Environmental Studies, and of D.C. by Princeton University. Jessica Anicas is thanked for her work to produce axenic strains from the Hawaii Culture Collection. Our colleagues from the BioSolarH2 team are thanked for several years of open and stimulating discussions about the various projects reported herein. Walter Kozumbo, our program director at the Air Force Office of Scientific Research, is personally thanked for his support.

ix

Table of Contents Abstract ...................................................................................................................................................................................... iii Acknowledgements .............................................................................................................................................................. vii Table of Contents ...................................................................................................................................................................... x List of Tables............................................................................................................................................................................ xii List of Figures ......................................................................................................................................................................... xiii List of Schemes ........................................................................................................................................................................xv Chapter 1: Metabolic Pathways for Photobiological Hydrogen Production by Nitrogenaseand Hydrogenase- Containing Unicellular Cyanobacteria Cyanothece .............................................. 1 Abstract ........................................................................................................................................................................................ 2 Introduction ................................................................................................................................................................................ 3 Methods ........................................................................................................................................................................................ 7 Results ........................................................................................................................................................................................... 9 Discussion ................................................................................................................................................................................. 14 References ................................................................................................................................................................................ 26 Supplemental Information................................................................................................................................................. 29 Chapter 2: Methods to Increase Catabolic Rate in Cyanothece for Increased Biological H2 and Ethanol Production .............................................................................................................................................35 Abstract ..................................................................................................................................................................................... 36 Introduction ............................................................................................................................................................................. 37 Methods ..................................................................................................................................................................................... 41 Results ........................................................................................................................................................................................ 45 Discussion ................................................................................................................................................................................. 50 References ................................................................................................................................................................................ 62 Supplemental Information................................................................................................................................................. 65 Chapter 3: Metabolic Regulation of Carbohydrate Catabolism Revealed in Real-time ..............67 Abstract ..................................................................................................................................................................................... 68 Introduction ............................................................................................................................................................................. 69 Methods ..................................................................................................................................................................................... 72 Results ........................................................................................................................................................................................ 75 Discussion ................................................................................................................................................................................. 81 References ................................................................................................................................................................................ 88 Supplemental Information................................................................................................................................................. 91 Chapter 4: Enhancing Biological Hydrogen Production from Cyanobacteria by Removal of Excreted Products ...............................................................................................................................................97 Abstract ..................................................................................................................................................................................... 98 Introduction ............................................................................................................................................................................. 99 Methods .................................................................................................................................................................................. 103 Results ..................................................................................................................................................................................... 106 Discussion .............................................................................................................................................................................. 111 References ............................................................................................................................................................................. 122

Chapter 5: Development and Application of a DOP-PCR Based Genetic Screen for [NiFe] Hydrogenases: Bioprospecting for Efficient Hydrogen Producing Organisms .......................... 125 Abstract .................................................................................................................................................................................. 126 Introduction .......................................................................................................................................................................... 127 Methods .................................................................................................................................................................................. 129 Results ..................................................................................................................................................................................... 132 Discussion .............................................................................................................................................................................. 135 References ............................................................................................................................................................................. 143

xi

List of Tables Table 3.S1 BLAST search results of translated nucleotides from A. maxima against dissimilatory nitrate reductase narGHI .......................................................................................................................................... 95 Table 5.1 Conserved amino acid regions used to create candidate degenerate primers and their corresponding DNA sequence ............................................................................................................................. 141 Table 5.2 Culture numbers, species (if known), and results from the hydrogenase screen ............... 142

xii

List of Figures Figure 1.1 Photo-H2 production in Cyanothece miami in response to selective excitation of PSI and to the photosynthetic electron transfer inhibitors DCMU and DBMIB .................................................. 20 Figure 1.2 Effect of dark preincubation time (autofermentation) on photo-H2 capacity ........................ 22 Figure 1.3 Effect of specific NADH dehydrogenase inhibitors flavone and rotenone on H2 production in Cyanothece ................................................................................................................................................................. 23 Figure 1.4 Effect of inhibiting membrane-coupled ATP generation on H2 production in Cyanothece24 Figure 1.S1 Experimental design for the continuous monitoring of photo- and dark-H2 production in Cyanothece ...................................................................................................................................................................... 29 Figure 1.S2 Demonstration of light saturation by 670 and 735 nm LEDs ..................................................... 30 Figure 1.S3 Plastoquinone pool size in native Cyanothece cells and in the presence flavone and rotenone........................................................................................................................................................................... 31 Figure 1.S4 Simultaneous detection of H2 production and intracellular reduced pyridine nucleotide concentration in Cyanothece during dark autofermentative conditions .............................................. 32 Figure 2.1 Total carbohydrate composition before and after anaerobic incubation with and without light in Cyanothece miami ......................................................................................................................................... 56 Figure 2.2 Introduction of an ATP sink under fermentative conditions ......................................................... 57 Figure 2.3 Carbohydrate content of Cyanothece in response to growth in high osmostic strength medium ............................................................................................................................................................................ 58 Figure 2.4 Increased photo-H2 in response to hypo-osmotic salt stress ........................................................ 59 Figure 2.5 Co-fermentation of diazotrophic Cyanothece miami with Synechococcus PCC 7002 WT and LdHAEx mutant .................................................................................................................................................... 60 Figure 2.S1 Photo-H2 production is limited by total available carbohydrate ............................................... 65 Figure 2.S2 Skeletal outline of carbohydrate catabolism via glycolysis or the oxidative pentose phosphate pathway ..................................................................................................................................................... 66 Figure 3.1 Dissolved H2 production rate and residual NAD(P)H concentration during dark anaerobic incubation in A. maxima ............................................................................................................................................ 85 Figure 3.2 Effect of extracellular nitrate on the rate of dissolved H2 production and residual NAD(P)H concentration during dark anaerobic incubation in A. maxima ........................................... 86 Figure 3.3 Effects of pentochlorophenol (PCP) and oxythiamine diphosphate (Oxy-TPP) on dissolved H2 production rate and residual NAD(P)H concentration during dark anaerobic incubation in the presence of nitrate ................................................................................................................... 87
xiii

Figure 3.S1 Linear response of fluorescence signal to exogenous NADH ...................................................... 93 Figure 3.S2 Scatterplot of phase-2 H2 oxidation current magnitude versus change in residual NAD(P)H concentration from ten different experiments ............................................................................ 94 Figure 4.1 Effect of sample dilution on headspace H2 yield by A. maxima.................................................. 115 Figure 4.2 Headspace H2 concentration above A. maxima indicating absence of H2 uptake and elimination of H2 production in 20% H2 atmosphere ................................................................................ 116 Figure 4.3 Sustained autofermentative dark H2 production by Arthrospira maxima under milking conditions ..................................................................................................................................................................... 117 Figure 4.4 Photosynthetic viability of cultures before and after 10 day autofermentation, measured by FRR fluorometry .................................................................................................................................................. 118 Figure 4.5 Dissolved H2 production rate and intracellular residual NAD(P)H concentration in A. maxima in the presence of a mixture of carbon fermentative products ............................................ 119 Figure 4.6 Eh-pH Pourbaix stability diagram of H2, H2O, NADH, and NAD+ at different H2 partial pressures ...................................................................................................................................................................... 120 Figure 5.1 Sequence alignment of hoxH amino acid sequences from 16 strains of cyanobacteria, created with Clustal X2 ........................................................................................................................................... 137 Figure 5.2 PCR amplification results illustrating accuracy of degenerate primers ................................ 139 Figure 5.3 In vivo electrochemical detection of dissolved hydrogen from two strains exhibiting hydrogenase activity................................................................................................................................................ 140

xiv

List of Schemes Scheme 1.1 Proposed model for coupling of carbohydrate catabolism to photo-H2 production in Cyanothece via non-photochemical reduction of the PQ pool, excitation of PSI, and nitrogenase activity .............................................................................................................................................................................. 25 Scheme 2.1 Abbreviated scheme indicating pathways for carbohydrate degradation and ATP sinks with Cyanothece ............................................................................................................................................................ 61 Scheme 3.1 Redox energy and phosphory energy conversion steps during catabolism that terminate in proton reduction and nitrate/nitrite reduction ......................................................................................... 83 Scheme 4.1 Simplified functional model of dark fermentative metabolic pathways in A. maxima that produce ATP, reductant, excreted carbon fermentative products, and molecular H2 ................. 121

xv

Chapter 1: Metabolic Pathways for Photobiological Hydrogen Production by Nitrogenase- and Hydrogenase- Containing Unicellular Cyanobacteria Cyanothece

Co-authors: Gennady M. Ananyev, Anagha Krishnan, and G. Charles Dismukes

This article is similar to the peer-reviewed article: Nicholas J. Skizim, Gennady M. Ananyev, Anagha Krishnan, and G. Charles Dismukes (2011) Metabolic pathways for photobiological hydrogen production by nitrogenase- and hydrogenase- containing unicellular cyanobacteria Cyanothece. Journal of Biological Chemistry. doi: 10.1074/jbc.M111.302125 Used with permission by ASBMB.

Abstract Current biotechnological interest in nitrogen-fixing cyanobacteria stems from their robust respiration and capacity to produce hydrogen. Here we quantify both dark- and light- induced H2 effluxes by Cyanothece sp. Miami BG 043511 and establish their respective origins. Dark, anoxic H2 production occurs via hydrogenase utilizing reductant from glycolytic catabolism of carbohydrate (auto-fermentation). Photo-H2 is shown to occur via nitrogenase and requires illumination of PSI, while production of O2 by co-illumination of PSII is inhibitory to nitrogenase above a threshold pO2. Carbohydrate also serves as the major source of reductant for the PSI pathway mediated via nonphotochemical reduction of the plastoquinone pool by NADH dehydrogenases type-1 and type-2 (NDH-1, NDH-2). Redirection of this reductant flux exclusively through the proton-coupled NDH-1 by inhibition of NDH-2 with flavone increases the photo-H2 production rate by 2-fold (at the expense of the dark-H2 rate), due to production of additional ATP (via the proton gradient). Comparison of photobiological hydrogen rates, yields, and energy conversion efficiencies reveals opportunities for improvement.

Background: Cyanothece produces H2 catalyzed by hydrogenase or nitrogenase. Results: Photo-H2 is nitrogenase mediated (via PSI) with reductant originating from catabolism and ATP from photophosphorylation. Conclusion: Forcing additional ATP production by inhibiting NDH-2 increases the photo-H2 production rate at the expense of dark-H2. Significance: Pathways for accelerating generation of ATP and reductant are identified as targets for further optimization of H2 yield by Cyanothece.

Introduction The biotechnological potential of photobiological hydrogen production by cyanobacteria and algae to renewable fuel production has not yet been realized. However, experimental and theoretical investigations of the maximum attainable rates (1), advances in culturing methods (23), enzyme manipulation (4-5), and metabolic engineering in model strains (6-8) have provided insight into the fundamental mechanisms and revealed new opportunities and limitations. Biological hydrogen production can be mediated by either of two broad classes of enzymes, the hydrogenases or the nitrogenases (8-10). Both classes are O2 sensitive, requiring anoxic environments to function maximally (11). Here we shall describe hydrogen metabolism by Cyanothece, a genus of unicellular, aerobic, nitrogen-fixing (diazotrophic) cyanobacteria. Unlike the heterocyst-forming diazotrophic cyanobacteria (e.g. Nostoc and Anabaena spp.) that spatially separate oxygen-evolving photosynthesis from oxygen-sensitive nitrogen fixation (11), Cyanothece performs strictly regulated temporal separation of photosynthesis and nitrogen fixation (12). This temporal separation ensures that an intracellular anoxic environment conducive to nitrogenase activity is maintained. Deactivation of photosynthesis and an increased cellular respiration are thought to maintain this anoxia (12-14), regulated by an intrinsic circadian rhythm (15-16), as even non-growing (stationary phase) cells exhibit this cycling (15, 17). Specifically, is has been shown in Cyanothece ATCC 51142 that one of the 4 copies of psbA, the gene encoding the D1 subunit of Photosystem II, is highly up-regulated throughout the dark phase of the circadian cycle. This isoform of D1 (psbA4) varies significantly from the isoforms expressed during the light phase, especially at the C terminal residues involved in binding the Mn-cluster and allowing O2 evolution (18). It is hypothesized that expression of this isoform and incorporation of the translated D1 protein into the reaction center would lead to a PSII core incapable of evolving O2. Expression of this isoform during the dark phase would allow the cell to retain a viable PSII quaternary framework so that an active D1 can quickly be re-incorporated during the subsequent light phase.
3

Illumination of Cyanothece while this inactive D1 were expressed could afford photobiological H2 production by enzymes normally sensitive to O2. A homolog of this isoform (psbA4) is found in the draft genome of the Cyanothece strain studied herein. The enzymatic reduction of dinitrogen to ammonia in Cyanothece is catalyzed by a Monitrogenase, the most common and most efficient of the nitrogenase classes. Mo-nitrogenase (nitrogenase hereafter) requires 16 ATP and 8 electrons per N2 fixed, with 2 of these electrons diverted to the obligate reduction of H+ to H2. The observed ratio of proton reduction to nitrogen reduction by this class of enzyme is at least 25% (from the reaction stoichiometry) and at maximum 100% (in the absence of N2, where the nitrogenase functions like an ATP-powered hydrogenase, requiring 4 ATP per H2 produced) (7-8). Of note, the alternative nitrogenases (V-nitrogenase and Fe-nitrogenase) found in some strains naturally favor a higher ratio of proton reduction to nitrogen reduction. Cyanothece species accumulate carbohydrate, primarily in the form of glycogen granules, during the day and subsequently degrade this carbohydrate at night to provide the energy and reductant for nitrogenase function and O2 respiration (19). This natural diurnal cycling, efficient conversion of intracellular carbohydrate to energy, efficient respiration and intracellular anoxia, and presence of both classes of hydrogen producing enzymes makes Cyanothece species among the best candidates for H2 production (20-21). Cyanothece sp. Miami BG 043511 is a marine strain that has been reported to possess both the bidirectional hydrogenase (Class 3d, bidirectional NAD-linked H2ase) and nitrogenase, but to lack the membrane bound respiratory or uptake hydrogenase (Class 2a, cyanobacterial uptake hydrogenase) (22), typically found expressed alongside nitrogenase (23). We have, however, identified the complete gene coding for the uptake hydrogenase in the strain and have observed its transcription under both photoautotrophic and auto-fermentative conditions (Skizim & Krishnan, in preparation). We can therefore assert that all three hydrogen metabolizing enzymes found in cyanobacteria are present in this strain.
4

NADH and reduced ferredoxin (FDx) are the direct substrates for the bidirectional hydrogenases and nitrogenases, respectively, as established by in vitro assays of enzymes from multiple microorganisms (24-25). The NADH-dependent reduction of H+ to H2 by hydrogenase is often cited as thermodynamically unfavorable, as the standard potential at pH 7 for NAD+/NADH is -320 mV, compared to -420 mV for H+/H2. However, standard conditions refer to 1 bar H2 pressure, conditions that are never found in biological cells. By contrast, the calculated thermodynamic redox potential for H+/H2 is nearly identical to that of NAD+/NADH in a 0.1% H2 atmosphere (1000 ppm, or 0.001 bar) (26). Consequently, pyridine nucleotides can serve as efficient reductant sources under biological conditions, especially when H2 backpressure is minimized. In addition to NADH and FDx, NADPH can be utilized indirectly, either by the reduction of NADH catalyzed by the pyridine nucleotide transhydrogenase (if expressed and active), or by the reduction of the plastoquinone (PQ) pool via NAD(P)H dehydrogenase and subsequent reduction of ferredoxin through excitation of PSI (Scheme 1.1). Cyanobacteria are known to possess two NAD(P)H dehydrogenase (NDH) enzymes to exchange reductant between NAD(P)H and the lipid soluble PQ pool (27-29). NDH type-1 is homologous to Complex I of the respiratory chain of mitochondria and bacteria (the NAD(P)H:quinone oxidoreductase). Interestingly however, of the 14 minimal subunits that form the complex in E. coli, only 11 are found in the genomes of cyanobacteria. The missing subunits (genes nuoE, nuoF, and nuoG) encode the NADH dehydrogenase module, which in E. coli functions as the energy input device, and consequently the mechanism by which cyanobacteria use NAD(P)H via this enzyme is still unknown (30). Experimental evidence has verified that NADPH and NADH can both be oxidized by the complex in purified cyanobacterial membranes (29) and a variety of mechanisms for electron donation to the complex have been supposed; this remains an open question (30). NDH type-1 (NDH-1) is capable of oxidizing both NADH and NADPH and (of importance to the present study) translocates (pumps) protons upon each reducing equivalent
5

(NAD(P)H) exchanged. NDH type-2 is comprised of a single subunit, does not pump protons, and can only oxidize NADH. The contribution of dark (autofermentative) hydrogen production arising from either hydrogenase- or nitrogenase- mediated pathways has been largely ignored in cyanobacteria. We show herein that hydrogen production in Cyanothece sp. Miami BG 043511 can be of the dark, autofermentative type (primarily via hydrogenase), as well as light-induced (via nitrogenase). By utilizing monochromatic excitation sources and detection of dissolved hydrogen, we monitor the kinetics of hydrogen evolution from both light- and dark-pathways within the same experimental incubation, thus allowing visualization of the independent responses of each pathway (light and dark) to applied stresses. We show that both intracellular reductant and ATP availability are limiting factors for maximal photo-H2 production, and that by increasing reductant availability via dark anaerobic incubation, or by channeling the flow of reductant through one of the specific NADH dehydrogenases to increase ATP availability, we can substantially increase the rate of photo-H2 production.

Methods Growth of Culture. Cyanothece sp. Miami BG 043511 was obtained from the University of Hawaii Culture Collection, where it was maintained as Synechococcus sp. Miami BG 043511 (many Cyanothece species were originally misclassified as Synechococcus (31)). Cultures were grown in ASP2 medium (32) without combined inorganic nitrogen at 30C under diurnal conditions (12 hr light, 12 hr dark) with a light intensity of ~30 E s-1 m-2. Cultures were not bubbled with any gases nor shaken. Dissolved H2 Rate Electrode & LED Illumination System. The first generation of our homebuilt H2 microcell is described elsewhere (33). Our second (current) generation electrochemical microcell is also a reverse Clark-type electrode for measuring dissolved H2 concentration (see Supporting Information). This 2nd generation microcell has an increased sensitivity of ~2 x 10-9 Coulomb H2, fast response time (100 ms), and micro-volume sample chamber (6.5 L). During measurement, H2 is constantly consumed by oxidation at a Pt/Ir electrode, allowing the instantaneous rate of H2 production to be measured. The microcell responds linearly to H2 and performs for several months using an oversized Ag/AgCl reference electrode (~0.5 g AgCl) to consume the H+ product of H2 oxidation. In this article, H2 production rate is presented as the current (nA) arising from the oxidation of dissolved H2. H2 yield was determined by integration with respect to time to yield electrical charge, which was converted to moles of H2 by Faradays second law. The microcell is equipped with light emitting diodes (LEDs) at wavelengths of 67010 and 73510 nm (FWHM), for exciting both Photosystems (670 nm) or PSI only (735 nm), respectively. The PAR intensity (400-700 nm) of the 670 nm LED at the sample is 1,010 E and only 2.8 E for the 735 nm LED. As a result, minimal excitation of PSII from the spectral tail of the 735 nm LED occurs, as confirmed by electrochemical O2 measurements in another Clark-type cell. Culture Incubation. Stationary phase cells were taken from photoautotrophic growth conditions between hours 10 and 11 after the onset of the light cycle (circadian dawn),
7

concentrated (40x) by centrifugation, re-suspended in fresh media, and placed in the 6.5 l sample chamber of the electrochemical cell. When biochemical inhibitors were added to the cultures, this was done five minutes prior to concentration and incubation on the hydrogen rate electrode. Inhibitors stocks were made with water or DMSO as solvent, not ethanol (exogenous ethanol may serve as a reductant source). The cell was covered by a quartz disc and removed from light, where cellular respiration created anoxia within a matter of seconds. Dissolved H2 was continuously consumed as described above. When pulsed illumination was supplied to the culture, it was done using repeated pulses of 10s duration, separated by 90s dark time, i.e. 10% duty cycle. Turbidity and Dry Weight Measurements. Cell density was measured as turbidity by absorbance at 730 nm. Optical densities were measured spectrophotometrically on a Thermo Scientific Evolution 60 UV-Vis. Dry weights of cells were taken by filtering cells through Whatman GF/C glass microfiber filters (1.4 um pore size) and drying the filters in an 80C oven for 24 hours.

Results Differential excitation and inhibition of Photosystems I and II. The selective excitation of PSI (denoted 735-photo-H2) and of PSII + PSI together (denoted 670-photo-H2) is shown in Figure 1.1. Cells were first subjected to 12 hours of complete darkness and anoxia prior to the 120 minutes of pulsed illumination shown. The sharp rise in H2 oxidation current corresponds to the 10s of illumination during each 100s light-dark cycle (repeated 72x). When both Photosystems are excited by 670 nm light (Fig 1.1A), the H2 production arising from each successive flash gradually decreases, completely ceasing after approximately 100 minutes. H2 production also gradually decreases but more slowly both when near-infrared 735 nm light (Fig 1.1C) is used (which only excites PSI) and when PSII is chemically inhibited (Fig 1.1B) with DCMU (3-(3,4-dichlorophenyl)1,1-dimethylurea). DCMU acts as a quinone analog, which binds to the QB pocket of PSII and blocks the transfer of electrons from QA (Scheme 1.1). The decrease in the absence of O2 production must therefore be due to the depletion of accessible intracellular reductant, which is common to all three conditions. The faster decay seen with light that excites both PSI + PSII is expected due to the O2 sensitivity of nitrogenase, and increases throughout the duration of the photoperiod, consistent with an intracellular buildup of O2 from PSII. DBMIB (2,5-dibromo-3-methyl-6-isopropyl-pbenzoquinone) blocks the plastoquinone binding site on cytochrome b6f and therefore prevents the flow of electrons into PSI. When cells were treated with DMBIB (Fig 1.1D) and illuminated with light, photo-induced H2 production is completely eliminated, indicating that the photo-H2 signal is completely dependent on electron transfer through PSI. This illustrates that photo-H2 is entirely nitrogenase mediated, as we would still expect to see 670-photo-H2 via PSII (steps 1234b7 in Fig 1.1E) in the presence of DBMIB if hydrogenase were mediating the process, as PQH2 oxidation by reverse electron flow through the NADH dehydrogenases is likely to occur (34-35). Further, the fact that H2 yield does not decrease when PSII is inhibited (in fact, the average yield increases 2.35x 0.75 with DCMU and 2.5x 1.3 with 735 nm illumination, based on triplicate
9

measurements) indicates that reductant from intracellular catabolism (glycogen) rather than water serves as the immediate electron source for photo-H2 in this strain (steps 84356). Interestingly, the fact that we do observe photo-H2 when 670 nm light is used confirms that Cyanothece strains are capable of maintaining lower levels of intracellular O2 during the dark phase of their circadian cycle. This may be due to higher levels of respiration by this strain (though this would steal electrons that could reduce protons to H2), or the incorporation of an inactive D1 protein into PSII during dark periods (as described in the introduction). Alternatively, the possibility also exists that the nitrogenase found in Cyanothece exhibits a higher tolerance for O2 and that it is this quality which allows it to produce H2 in response to visible light. Increased dark pre-incubation leads to increased photo-H2. Figure 1.2 illustrates the effect of the dark, anaerobic pre-incubation period on 735-photo-H2. As above, cells were taken from photoautotrophic growth and placed in the sample chamber of the electrochemical cell. Dark periods (autofermentation) of 0.75, 1.5, 4, 8, 12, and 16 hours were allowed before the cells were exposed to three hours of pulsed illumination with 735 nm light. Figure 1.2A shows the rate of H2 production measured for each of these experiments, and Figure 1.2B illustrates the yield of 735photo-H2 produced (as picomoles H2) during each three-hour illumination window. (The integral of the hydrogen production rate is only taken over the three hours period of illumination.) With increasing dark time supplied to the cells before illumination, the yield of 735-photo-H2 over this fixed 3 hour period continuously increases as a linear function of dark time up to 16 hours. From the linear regression fit shown in Figure 1.2B, the yield of 735-photo-H2 increases by 8.4- fold from time zero to 16 hours. This result shows that during the auto-fermentation period prior to illumination, intracellular reductant accumulates, presumably by the degradation of glycogen to mono- and di-saccharides, the immediate glycolytic precursors, which is subsequently funneled to proton reduction during illumination. Indeed we have directly observed an increase in redox poise of the cells throughout autofermentative conditions (Fig 1.S4).
10

To investigate the effect of biochemical inhibitors on both photo- and dark- H2 production, an experimental design consisting of four periods of pulsed illumination each separated by 4h dark time was instituted (Fig. 1.S1). Inhibition of the NADH dehydrogenases (Types 1 and 2). Figure 1.3 illustrates the effect on hydrogen production rate (A) and yield (B) when catabolically derived reductant is funneled to PQ reduction through each of the specific NADH dehydrogenases found in the strain. Figure 1.3A shows the rate of hydrogen production of (i) control cells, where both enzymes are active, (ii) cells treated with 50 M flavone, where only NDH-1 is active, and (iii) cells treated with 20 M rotenone, where only NDH-2 is active. Each condition was subjected to a 20-hr incubation with four pulse trains of saturating light at 735 nm (Fig. 1.S2). The average 735-photo-H2 production rate when cells were treated with flavone was 16.4 mol H2 gDW-1 h-1, or 2-fold higher than the control rate. This increase occurs at the expense of a 2.5-fold lower dark-H2 production rate. These effects are evident in Figure 1.3B, where the cumulative yield of H2 (both dark- and light-) is plotted as a function of time. These data reveal that the carbohydrate pool, which serves as precursor to cellular reductant for conversion to H2, is shared between the photo-H2 and dark-H2 pathways, as the increased photo-H2 is coupled to a decline in the dark-H2 rate, a redirection of reductant between the pathways. We also point out the decrease in dark-H2 production in Figure 1.3 common to treatment with both flavone and rotenone. Because the rate of glycolysis is regulated by the cellular energy charge (ATP availability) as well as the NADH/NAD+ poise, it is reasonable to expect inhibitors that affect proton pumping and equilibration of the intracellular reductant pools to have upstream effects on glycolysis. The overall decrease in dark-H2 production common to both inhibitors may well be due to a down-regulation of glycolysis in response to increased NAD(P)H which cannot reduce the plastoquinone pool due to the action of these inhibitors. There also may be some net flux of reductant out of the PQ pool (PQH2 to NADH) leading to dark-H2 production in the native
11

system. When NDH-2 is inhibited with flavone, this reaction would have to proceed via NDH-1 (requiring ATP) and become unfavorable. We believe this explains the observation of decreased dark-H2 prior to the first illumination period with flavone treatment. In flavone-treated cells, the light saturated rate of 735-photo-H2 production starts to decline after 60 minutes of pulsed illumination and drops to 50% of the initial rate after approximately 2-3 hours. This decrease in rate is common to control cells without flavone treatment as well, and is presumably due to the depletion of substrate for nitrogenase (reductant, protons, or ATP). Interestingly, in a similar experiment composed of a 20 hour anoxic incubation of a 5 mL sample in sealed glass vials under continuous saturating white light illumination (1.81 mW cm-2 s-1), only 15.7% of the total intracellular carbohydrate was depleted over the 20 hour incubation (same experiment length as rate data presented here). Therefore, there exists significant room to improve the light saturated 735-photo-H2 rate, if the dark catabolic rate and subsequent generation of NADH could be enhanced. Decreased capacity for ATP Phosphorylation. Two well-documented protonophores were used to collapse the proton gradient and uncouple photophosphorylation from photosynthetic electron transfer. Both 2,4-dinitrophenol (DNP) and carbonyl cyanide-ptrifluoromethoxyphenylhydrazone (FCCP) are ionophores that shuttle protons across biological membranes and were used to collapse the proton motive force responsible for generation of ATP via the F1F0 complex (36). Additionally, N,N'-dicyclohexylcarbodiimide (DCCD) was employed to directly inhibit ATP synthesis by blocking the flow of protons through the F0 channel and subsequently prevent phosphorylation of ADP by the F1 complex (37). The disruption of ATP synthesis and its effect on cellular energy charge by each of these inhibitors is shown in Figure 1.4. Figure 1.4A illustrates the cumulative yield of H2 produced by the cells (both dark-H2 and photoH2) as a function of time (the white areas indicate the four 60-minute pulsed illumination periods, as before). Both FCCP (5 M) and DNP (50 M) increase the dark-H2 rate. This increase is attributed
12

to an increased rate of carbohydrate catabolism to compensate for the loss of ATP and collapse of the proton gradient caused by the protonophores. The bidirectional hydrogenase is the expected enzymatic outlet for this increase in dark-H2 via NADH oxidation. By contrast, both protonophores decrease the photo-H2 production rate; the rate approaches zero at the end of each 1 hr pulse train. This decrease is attributed to the loss of ATP production by photophosphorylation which is essential for nitrogenase-dependent photo-H2 production. This decrease is illustrated as well in Figure 1.4B, in which the average photo-H2 rates during the 1 hr illumination periods for each of the four pulse trains are shown. All three inhibitors initially decrease the observable photo-H2 by more than 50% (illumination period #1). While FCCP and DNP retain their inhibition activity, the inhibition effect of DCCD decreases successively with each subsequent pulse train. Because DCCD is a cross-linker that physically binds to the F0 channel, we hypothesize that the cell can effectively sequester the DCCD by this binding activity. Therefore, by synthesizing more F0 channels, the cell can recover from this initial inhibition. Because DNP and FCCP do not bind to proteins like DCCD does, the cell cannot sequester them, and their capacity for inhibition remains unchanged with time.

13

Discussion By designing an experimental method that exposes Cyanothece to segments of both illumination and darkness during a single anoxic incubation, our electrochemical H2 sensor can observe (with high kinetic resolution) the response of both H2 producing enzymes to environmental and biochemical stresses and thereby allows the current mechanistic study. As such we are able to determine the relative contributions from hydrogenase and nitrogenase towards light- and dark- H2 production, identify shared pools for both processes, and indicate factors limiting the yield of each. The complete elimination of photo-H2 production when DBMIB is added to the sample (inhibiting electron flow through Photosystem I) can be attributed to loss of ATP generation from cyclic electron flow (photophosphorylation) and/or elimination of the production of reduced ferredoxin, based on the known outcome under aerobic conditions (36). Both are obligate substrates for nitrogenase, and the complete elimination of photo-H2 under these conditions indicates that the sole enzyme responsible for photo-induced H2 production in Cyanothece is nitrogenase. Our data illustrate an increased capacity for photo-H2 production with increasing dark pre-incubation, independent of PSII-dependent water oxidation. This result indicates the tight coupling of photo-H2 to dark, anaerobic metabolism, where substrate builds up and becomes kinetically more accessible as anaerobic conditions progress. This increased availability of reductant translates into a greater photo-H2 production rate once the pulsed illumination commences. ATP levels fall during dark anoxia (38) and thus only accumulation of reductant can be responsible for this increase due to dark time. Scheme 1.1 presents a minimal diagram of reductant and ATP generation to guide interpretation and assignment of the pathways. Under dark anoxia, cells catabolize their glycogen and sugar reserves to produce glucose-6-phosphate (G6P), which enters either glycolysis or the oxidative pentose phosphate pathway to produce NADH or NADPH, respectively. Subsequent
14

oxidation of pyruvate by pyruvate:ferredoxin oxidoreductase (PFOR) generates reduced ferredoxin in the dark (PFOR is the primary source of reduced ferredoxin during dark, anaerobic, and anabolic nitrogen fixation for unicellular nitrogen fixers (8)). NAD(P)H has a reducing potential of -320 mV, substantially less than that of ferredoxin at -415 mV. As the latter is the obligate electron donor for nitrogenase-dependent H2 production, NAD(P)H can only convert into reduced ferredoxin through either (a) being energized by PSI and light, or (b) interconversion by ferredoxin:NADP+ oxidoreductase (FNR) in the dark, a thermodynamically uphill reaction that requires buildup of NAD(P)H. Our results demonstrate the predominance of the former pathway, where catabolically generated NAD(P)H exchanges hydride (reductant) with the PQ pool via either of two NADPH dehydrogenases. This hydride is energized to the level of reduced ferredoxin and free proton (coupled to ATP generation) by photo-excitation of PSI and made available for nitrogenase. This sudden shift in equilibrium between NAD(P)H and reduced ferredoxin and ATP manifests as the sharp increase in hydrogen evolution observed upon illumination. Hydride exchange from NAD(P)H to the plastoquinone pool can occur in cyanobacteria by either of two NADH dehydrogenases present in the organism: NDH type-1 and NDH type-2. As mentioned earlier, NDH-1 translocates a proton from the stroma to lumen as it reduces plastoquinone (Scheme 1.1) and thus contributes to the intracellular proton gradient. NDH-2 lacks this ability. The channeling of all hydride flux from NAD(P)H to PQ through NDH-1 (rather than both NDH-1 and NDH-2) would provide the most ATP generation per hydride equivalent exchanged (by the proton pumping of NDH-1) and best accommodate the high ATP demand of nitrogenase. In the case of cells treated with flavone, we indeed see this effect. The maximal sustained (2.5 hours) rate of hydrogen production in Cyanothece sp. Miami BG 043511 observed in these experiments was 16.4 mol H2 gDW-1 h-1 (15.8 ml H2 L-1h-1) with flavone. To ensure that the increase in 735photo-H2 was not due to an increase in the plastoquinone concentration (arising from regulatory

15

effects of NDH-1 or NDH-2), we measured the PQ pool size and confirmed that it did not change following treatment with either flavone or rotenone (Fig. 1.S3). Our findings reveal that of the two factors limiting the rate of nitrogenase mediated 735photo-H2, reductant limitation is stronger than ATP limitation under most physiological conditions, as evidenced by the increase in 735-photo-H2 production rates with increasing prior dark time where reductant accumulates (Fig. 1.S4), and by the decrease in H2 production rate as the illumination period exhausts the available pool. Only when cells have ample reductant available is an ATP limitation observed, leading to the observed increase in 735-photo-H2 when reductant flux is forced through NDH-1. Further evidence of reductant limitation as the most severe limitation is that photo-H2 production increases in the presence of exogenous reduced carbon (3.3-fold increase when 685 mM glycerol was added to cells). Our H2 production rate of un-supplemented stationary phase cells is 1.64 mol H2 mg Chl-1 h-1. Based on carbohydrate degradation and hydrogen production yields after 20 hrs, our numbers correspond to a 24% energy conversion efficiency (ECE) of carbohydrate to hydrogen from the strain. Higher rates have been reported by Borodin and colleagues (21) for the same strain grown in a photobioreactor where actively growing cells were held at a chlorophyll concentration of ~4 g/mL (6.3 mL H2 L-1 h-1, or calculated to 70 mol H2 mg Chl-1 h-1). More recently, extremely high rates of 373 mol H2 mg Chl-1 h-1 and 465 mol H2 mg Chl-1 h-1 (supplemented with 50 mM glycerol) have been reported for Cyanothece sp. ATCC 51142 (39). If confirmed, these would set a new benchmark, as they are significantly higher than the maximum rates published to date for both heterocystous diazotrophs (167.6 mol H2 mg Chl-1 h-1 for Anabaena variabilis sp. ATCC 29413 PK84 mutant) (40) and for strains which produce H2 via the bidirectional hydrogenase only (3.1 mol H2 mg Chl-1 h-1 for Arthrospira maxima (41)). However, our own attempt to replicate the glycerol rate data using the ATCC 51142 strain have yet to confirm these results, and the reported rates exceed the maximum theoretical rate of nitrogenase mediated H2 production of 40 mol H2 mg Chl-1 hr-1 as
16

calculated by Bothe, et al. based on maximal carbon fixation rates of roughly 100 mol mg Chl-1 hr-1 and a C/N ratio of 6 in cyanobacteria (8). It is interesting to note that while the addition of the protonophores decreases the yield of photo-H2, they substantially increase the dark-H2 production yield. This confirms that the two mechanisms of hydrogen production (light- and dark-) are catalyzed by different enzymes that respond differently to cellular energy stress. While the decreased photophosphorylation with membrane uncouplers in the light serves to limit nitrogenase of its key energy source (ATP), under dark conditions this disruption stimulates the cell to accelerate the rate of carbohydrate catabolism and thus increase ATP generation by substrate level phosphorylation (glycolysis) (Scheme 1.1, step 8b). Faster carbohydrate catabolism under anoxia increases the ratio of NADH:NAD+ within the cell, which the bidirectional hydrogenase serves to alleviate by reducing protons to H2 and regenerating NAD+ (6). The analogous response of increased fermentative hydrogen production in the presence of FCCP was seen in Cyanothece sp. PCC 7822 (42) and in the non-photosynthetic bacterium E. coli when ATP levels were decreased by the introduction of futile ATPases that continually hydrolyze ATP (43). The model in Scheme 1.1 presents multiple targets identified herein for improving H2 production to better match the solar cycle and improve the H2 yield. For instance, future improvements to cyanobacterial photo-H2 production via nitrogenases can be anticipated by: 1) maximizing the total pool size of carbohydrate (e.g. by growth in high salt media) and hence NADH and NADPH turnover, 2) increasing the expression of NDH-1 thus increasing both ATP generation and reductant flux from carbohydrate catabolism to PSI, 3) engineering carbohydrate catabolism so that less glycolytic flux and more OPP flux occurs thereby producing more reductant for delivery to ferredoxin via PSI, 4) increasing the rate of nitrogenase mediated H2 production by lowering its ATP demand, and 5) engineering nitrogenase so a larger percentage of electrons reduce H+ instead of N2. In fact, genetic engineering of the Mo-nitrogenase from the non-photosynthetic bacterium
17

Azotobacter vinelandii has shown that replacement of a single amino acid in the enzyme results in approximately 80% of the electrons being redirected to H2 from N2 (4). Additionally, by disrupting the two genes encoding homocitrate synthase in Anabaena sp. PCC 7120, citrate can be incorporated into the FeMo cofactor of nitrogenase, thus favoring H2 production in a N2 atmosphere (5). Lastly, the knockout of anaerobic pathways which recycle NADH to NAD+ (i.e. lactate dehydrogenase, as has been done in Synechococcus sp. PCC 7002 (6)) should increase the accessible intracellular reductant pool under anoxia. If coupled to the removal of the two hydrogenases within the cell, this approach could divert a substantially increased flux of reductant to nitrogenase for H2 production. The recent development of a transformation system in Cyanothece sp. PCC 7822 with the use of ssDNA (44) may enable realization of the above mentioned genetic manipulations for further improving H2 production by strains of this genus. These manipulations will undoubtedly help overcome the reductant and ATP limitations we observe from in vivo nitrogenase dependent H2 production rates. Additionally, from a biotechnological standpoint, minimizing O2 production from the system has to be a future research focus as well, in order to realize the potential of photo-H2 pathways. This could involve 1) the search for O2 insensitive nitrogenases, 2) the development of a large scale optical filtering system such that only near-infrared light is transmitted to the bioreactor, and/or 3) the selective expression of an inactive D1 isoform to inactivate PSII dependent O2 evolution under H2 producing conditions. In conclusion, Cyanothece strains are promising because of their ability to grow on minimal media without combined nitrogen, their robust energy metabolism during anoxia, their ability to maintain low-oxic intracellular conditions conducive to H2 generating enzymes, and their strong coupling of carbohydrate catabolism to PQ reduction allowing rapid photo-H2 production in response to illumination. While the net yield of H2 from Cyanothece is not increased over dark fermentative levels by the addition of illumination periods (see control in Fig. 1.3B), the rate of H2
18

production from the photo- pathway is significantly increased. Biotechnologically, this affords the most rapid conversion of intracellular reductant (which is shared between the dark- and photopathways) to H2 (2-fold increase observed here). We imagine a system where reductant could accumulate under anoxia (perhaps in a strain with the hydrogenase knocked out) and then be quickly extracted (by conversion to H2) during an illumination period. The cycle could then be repeated. Lastly, nitrogenase-mediated H2 production is unique in being unidirectional and not inhibited by H2 accumulation, a quality not found with the bidirectional hydrogenase. While significant challenges remain in developing the potential of aquatic microbes for biological hydrogen production, including overall energy conversion efficiencies and scalability concerns, we hope that through the use of the above targets, subsequent studies can further our understanding and bring us closer to realizing the biotechnological potential of nitrogenase-mediated H2 production from Cyanothece.

19

Figures

Figure 1.1. Immediate source of reductant for photo-H2 production in Cyanothece sp. Miami BG 043511 is glycogen. (A) Red light ( = 670 nm) excites both PSII and PSI, while (C) near-infrared light (NIR, = 735 nm) excites PSI only and has no direct effect on PSII. (B) DCMU (10 M) inhibits PSII by blocking reduction of QB. An increase in photo-H2 is observed both when red light is supplied to the culture in the presence of DCMU (2.35x 0.75 (SD)) and when NIR light is used for illumination (2.49x 1.3 (SD)) due to diminished O2 poisoning from elimination of PSII-dependent O2 evolution. (D) DBMIB (40 M) blocks transfer of electrons from reduced PQ to PSI which blocks the nitrogenase-mediated pathway (PSI dependent), but would not prevent H2ase mediated photoH2 production via PSII (1234b7). The lack of photo-H2 with this treatment indicates a photo-H2 pathway entirely mediated by nitrogenase, while the low level dark-H2 current is due to hydrogenase (E) A minimal model for photo-H2 production shows possible routes for electron flow to nitrogenase and hydrogenase with reductant originating from the reduced PQ pool (formed from H2O via visible light and PSII) or glycogen (via NADH dehydrogenase). In both cases NIR light and
20

PSI is required to transfer electrons to the level of reduced ferredoxin (substrate for nitrogenase) in order to observe photo-H2. Hydrogenase mediated H2 production may occur in the dark via NADH formed via carbohydrate catabolism (8b7). H2 electrode calibration: 100 nA = 14.3 mol H2 gDW1 h-1.

21

Figure 1.2. Effect of dark, preincubation time (autofermentation, AF) on photo-H2 capacity. Cells were taken from photoautotrophic growth and placed under dark, anaerobic conditions for lengths of time ranging from 0.75 to 16 hrs before the onset of a three hour period of pulsed illumination (735 nm LED, represented by the horizontal gray bars). Dissolved H2 was measured as current arising from the oxidation of dissolved H2. The rate (A) of hydrogen evolution and cumulative yield (B) of 735-photo-H2 arising from the pulsed illumination window are shown. Error bars indicate 5% variability in H2 yield. The horizontal black bars (AF time & hv) represent the time segments for the top trace only (16h dark).

22

Figure 1.3.The effect on H2 production of selective inhibitors of the two NADH dehydrogenases. (A) The rate of dissolved H2 production was measured continuously from cells where both NADH dehydrogenases were functional (CTRL, C), where only NDH Type-1 was active (50 M flavone, F), and where only NDH Type-2 was active (20 M rotenone, R). (B) The cumulative yield of dark- and photo- H2 evolution is shown, with the four 1-hr photo-incubation periods (735-photo) indicated by vertical white rectangles. The light-saturated maximal rate of photo-H2 production (extracted from the four photo- incubation periods) was observed when only NDH type-1 was active, and was quantified as 16.4 mol H2 gDW-1 h-1. This rate increase of 2-fold (compared to the control) was at the expense of a dark-H2 production rate approximately 2.5-fold slower. This indicates a shared pool of reductant for both pathways (light- and dark-).

23

Figure 1.4 Effect of inhibiting membrane-coupled ATP generation on H2 production. Both FCCP (5 M) and DNP (50 M) are protonophores (uncouplers of the proton gradient); DCCD (250 M) blocks proton flow through the F0 channel of ATPase. All three of these inhibitors prevent the potential energy stored in the proton motive force from being translated into chemical energy as ATP. The cumulative yields of H2 production (A) and individual photo-H2 production rates from the four pulsed illumination windows (B) are shown. While the cells treated with DNP and FCCP show inhibited photo-H2 capacity, their dark-H2 production rate is increased, consistent with increased glycolytic flux to compensate for stress introduced by the uncouplers.

24

Scheme 1.1. Proposed model for coupling of carbohydrate catabolism to photo-H2 production in Cyanothece where non-photochemical reduction of PQ by catabolically derived NAD(P)H is the primary source of reductant for PSI-dependent, nitrogenase-mediated photo-H2. Sites of proton pumping coupled to ATP generation are denoted by blue arrows, and sites of inhibitor action are denoted with a red X. An abbreviated scheme of carbohydrate catabolism showing the reduction of NADH by glycolysis and the reduction of NADPH by the oxidative pentose phosphate pathway (OPP) is shown in the inset at left. Stoichiometries are not included. Abbreviations: CHO carbohydrates, G6P glucose-6-phosphate, GAP glyceraldehyde phosphate, PYR pyruvate, AcCoA acetyl coenzyme-A, QA& QB PSII-bound plastoquinone molecules, PQ the pool of membrane soluble plastoquinones, b6f cytochrome b6f, FNR ferredoxin:NADP+ oxidoreductase, N2ase nitrogenase, TH pyridine nucleotide transhydrogenase, [NiFe]-H2ase bidirectional hydrogenase.

25

References 1. 2. James B, Baum G, Perez J, & Baum K (2009) Technoeconomic boundary analysis of biological pathways to hydrogen production. in Report No. SR-560-46674 (NREL), p 207. Carrieri D, Momot D, Brasg IA, Ananyev G, Lenz O, Bryant DA, & Dismukes GC (2010) Boosting Autofermentation Rates and Product Yields with Sodium Stress Cycling: Application to Production of Renewable Fuels by Cyanobacteria. Appl. Environ. Microbiol. 76(19):6455-6462. Laurinavichene TV, Kosourov SN, Ghirardi ML, Seibert M, & Tsygankov AA (2008) Prolongation of H2 photoproduction by immobilized, sulfur-limited Chlamydomonas reinhardtii cultures. Journal of Biotechnology 134(3-4):275-277. Barney BM, Igarashi RY, Dos Santos PC, Dean DR, & Seefeldt LC (2004) Substrate Interaction at an Iron-Sulfur Face of the FeMo-cofactor during Nitrogenase Catalysis. Journal of Biological Chemistry 279(51):53621-53624. Masukawa H, Inoue K, & Sakurai H (2007) Effects of Disruption of Homocitrate Synthase Genes on Nostoc sp. Strain PCC 7120 Photobiological Hydrogen Production and Nitrogenase. Appl. Environ. Microbiol. 73(23):7562-7570. McNeely K, Xu Y, Bennette N, Bryant DA, & Dismukes GC (2010) Redirecting Reductant Flux into Hydrogen Production via Metabolic Engineering of Fermentative Carbon Metabolism in a Cyanobacterium. Appl. Environ. Microbiol. 76(15):5032-5038. McKinlay JB & Harwood CS (2010) Photobiological production of hydrogen gas as a biofuel. Current Opinion in Biotechnology 21(3):244-251. Bothe H, Schmitz O, Yates MG, & Newton WE (2010) Nitrogen Fixation and Hydrogen Metabolism in Cyanobacteria. Microbiol. Mol. Biol. Rev. 74(4):529-551. Tsygankov A (2007) Nitrogen-fixing cyanobacteria: A review. Applied Biochemistry and Microbiology 43(3):250-259. Ghirardi ML, Posewitz MC, Maness PC, Dubini A, Yu JP, & Seibert M (2007) Hydrogenases and hydrogen photoproduction in oxygenic photosynthetic organisms. Annual Review of Plant Biology 58:71-91. Gallon JR (1992) Reconciling the incompatible: N2 fixation and O2. New Phytologist 122(4):571-609. Mitsui A, Kumazawa S, Takahashi A, Ikemoto H, Cao S, & Arai T (1986) Strategy by which nitrogen-fixing unicellular cyanobacteria grow photoautotrophically. Nature 323(6090):720-722. Stckel J, Welsh EA, Liberton M, Kunnvakkam R, Aurora R, & Pakrasi HB (2008) Global transcriptomic analysis of Cyanothece 51142 reveals robust diurnal oscillation of central metabolic processes. Proceedings of the National Academy of Sciences 105(16):6156 -6161. Schneegurt MA, Tucker DL, Ondr JK, Sherman DM, & Sherman LA (2000) Metabolic rhythms of a diazotrophic cyanobacterium, Cyanothece sp. strain atcc 51142, heterotrophically grown in continuous dark. Journal of Phycology 36(1):107-117. Mitsui A & Suda S (1995) Alternative and cyclic appearance of H2 and O2 photoproduction activities under non-growing conditions in an aerobic nitrogen-fixing unicellular cyanobacterium Synechococcus sp. Current Microbiology 30(1):1-6. Suda S, Kumazawa S, & Mitsui A (1992) Change in the H2 photoproduction capability in a synchronously grown aerobic nitrogen-fixing cyanobacterium, Synechococcus sp. Miami BG 043511. Archives of Microbiology 158(1):1-4. Mitsui A, Cao S, Takahashi A, & Arai T (1987) Growth synchrony and cellular parameters of the unicellular nitrogen-fixing marine cyanobacterium, Synechococcus sp. strain Miami BG 043511 under continuous illumination. Physiologia Plantarum 69(1):1-8.

3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.

26

18. 19. 20. 21. 22. 23.

24. 25.

26. 27. 28. 29. 30. 31. 32. 33.

34.

Toepel J, Welsh E, Summerfield TC, Pakrasi HB, & Sherman LA (2008) Differential Transcriptional Analysis of the Cyanobacterium Cyanothece sp. Strain ATCC 51142 during Light-Dark and Continuous-Light Growth. J. Bacteriol. 190(11):3904-3913. Schneegurt M, Sherman DM, Nayar S, & Sherman LA (1994) Oscillating behavior of carbohydrate granule formation and dinitrogen fixation in the cyanobacterium Cyanothece sp. strain ATCC 51142. J. Bacteriol 176:1586-1597. Min H & Sherman LA (2010) Hydrogen Production by the Unicellular, Diazotrophic Cyanobacterium Cyanothece sp. Strain ATCC 51142 under Conditions of Continuous Light. Appl. Environ. Microbiol. 76(13):4293-4301. Borodin VB, Rao KK, & Hall DO (2002) Manifestation of behavioural and physiological functions of Synechococcus sp. Miami BG 043511 in a photobioreactor. Marine Biology 140(3):455-463. Vignais PM, Billoud B, & Meyer J (2001) Classification and phylogeny of hydrogenases. FEMS Microbiol Rev 25(4):455-501. Ludwig M, Schulz-Friedrich R, & Appel J (2006) Occurrence of hydrogenases in cyanobacteria and anoxygenic photosynthetic bacteria: Implications for the phylogenetic origin of cyanobacterial and algal hydrogenases. Journal of Molecular Evolution 63(6):758768. Maroti J, Farkas A, Nagy IK, Maroti G, Kondorosi E, Rakhely G, & Kovacs KL (2010) A Second Soluble Hox-Type NiFe Enzyme Completes the Hydrogenase Set in Thiocapsa roseopersicina BBS. Appl. Environ. Microbiol. 76(15):5113-5123. Burgdorf T, van der Linden E, Bernhard M, Yin QY, Back JW, Hartog AF, Muijsers AO, de Koster CG, Albracht SPJ, & Friedrich B (2005) The soluble NAD(+)-reducing [NiFe]Hydrogenase from Ralstonia eutropha H16 consists of six subunits and can be specifically activated by NADPH. Journal of Bacteriology 187(9):3122-3132. Vincent KA, Parkin A, & Armstrong FA (2007) Investigating and Exploiting the Electrocatalytic Properties of Hydrogenases. Chemical Reviews 107(10):4366-4413. Howitt CA, Udall PK, & Vermaas WFJ (1999) Type 2-NADH Dehydrogenases in the Cyanobacterium Synechocystis sp. Strain PCC 6803 Are Involved in Regulation Rather Than Respiration. J. Bacteriol. 181(13):3994-4003. Cooley JW & Vermaas WFJ (2001) Succinate Dehydrogenase and Other Respiratory Pathways in Thylakoid Membranes of Synechocystis sp. Strain PCC 6803: Capacity Comparisons and Physiological Function. J. Bacteriol. 183(14):4251-4258. Peschek GA, Obinger C, & Paumann M (2004) The respiratory chain of blue-green algae (cyanobacteria). Physiol Plant 120(3):358-369. Battchikova N, Eisenhut M, & Aro E-M (2011) Cyanobacterial NDH-1 complexes: Novel insights and remaining puzzles. Biochimica et Biophysica Acta (BBA) - Bioenergetics 1807(8):935-944. Zehr JP, Waterbury JB, Turner PJ, Montoya JP, Omoregie E, Steward GF, Hansen A, & Karl DM (2001) Unicellular cyanobacteria fix N2 in the subtropical North Pacific Ocean. Nature 412(6847):635-638. Reddy KJ, Haskell JB, Sherman DM, & Sherman LA (1993) Unicellular, aerobic nitrogenfixing cyanobacteria of the genus Cyanothece. J. Bacteriol. 175(5):1284-1292. Ananyev G, Carrieri D, & Dismukes GC (2008) Optimization of metabolic capacity and flux through environmental cues to maximize hydrogen production by the cyanobacterium "Arthrospira (Spirulina) maxima". Applied and Environmental Microbiology 74(19):61026113. Nixon PJ & Rich PR (2006) Chlororespiratory Pathways and Their Physiological Significance - The Structure and Function of Plastids. Advances in Photosynthesis and Respiration, eds Wise RR & Hoober JK (Springer Netherlands), Vol 23, pp 237-251.
27

35. 36. 37. 38. 39. 40. 41. 42.

43. 44.

Klamt S, Grammel H, Straube R, Ghosh R, & Gilles ED (2008) Modeling the electron transport chain of purple non-sulfur bacteria. Mol Syst Biol 4. Falkowski PG & Raven JA (1997) Aquatic Photosynthesis (Blackwell, Malden). Danon A & Stoeckenius W (1974) Photophosphorylation in Halobacterium halobium. Proceedings of the National Academy of Sciences 71(4):1234-1238. Bennette NB, Eng JF, & Dismukes GC (2011) An LCMS-Based Chemical and Analytical Method for Targeted Metabolite Quantification in the Model Cyanobacterium Synechococcus sp. PCC 7002. Analytical Chemistry 83(10):3808-3816. Bandyopadhyay A, Stckel J, Min H, Sherman LA, & Pakrasi HB (2010) High rates of photobiological H2 production by a cyanobacterium under aerobic conditions. Nat Commun 1(9):139. Dutta D, De D, Chaudhuri S, & Bhattacharya SK (2005) Hydrogen production by Cyanobacteria. (Translated from eng) Microb Cell Fact 4:36 (in eng). Carrieri D, Ananyev G, Costas AMG, Bryant DA, & Dismukes GC (2008) Renewable hydrogen production by cyanobacteria: Nickel requirements for optimal hydrogenase activity. International Journal of Hydrogen Energy 33(8):2014-2022. van der Oost J, van Walraven HS, Bogerd J, Smit AB, Ewart GD, & Smith GD (1989) Nucleotide sequence of the gene proposed to encode the small subunit of the soluble hydrogenase of the thermophilic unicellular cyanobacterium Synechococcus PCC 6716. Nucleic Acids Res 17(23):10098. Koebmann BJ, Westerhoff HV, Snoep JL, Nilsson D, & Jensen PR (2002) The Glycolytic Flux in Escherichia coli Is Controlled by the Demand for ATP. J. Bacteriol. 184(14):3909-3916. Min H & Sherman LA (2010) Genetic Transformation and Mutagenesis via ssDNA in the Unicellular, Diazotrophic Cyanobacteria of the Genus Cyanothece. Appl. Environ. Microbiol.:AEM.01456-01410.

28

Supplementary Information

Figure 1.S1. Experimental design for the continuous monitoring of photo- and dark- H2 production. Four 60-minute photoperiods (pulse trains) are represented by clusters of vertical black lines representing pulsed illumination (via LED) at a duty cycle of 10%, i.e. 10 seconds light followed by 90 seconds dark, with this cycle repeated 36 times. The first pulse sequence starts after 2 hours dark incubation; each of the following is separated by four hours of dark time. A representative dataset is shown for Cyanothece when illuminated with red LED ( = 670 nm, panel A) as well as infrared LED ( = 735 nm, panel B). Hydrogen is measured as current arising from the oxidation of dissolved H2 on a Pt/Ir electrode. The yield of photo-H2 is largest and the extent of dampening of the maximal rate (pulse #36 rate minus pulse #1 rate, denoted VIS or IR) is smallest, when 735 nm light is used for illumination.

29

Figure 1.S2. Demonstration of saturating light intensity by 670 nm and 735 nm LEDs. Optical power of the LED flashers (red data points = 670 nm; black data points = 735 nm) was modulated and the corresponding peak photo-H2 production after a 10 second flash was plotted. Photo-hydrogen production saturates at an optical power of <0.5 mW cm-2 while the optical power of the LEDs throughout the experiments presented in this manuscript is 7.8 mW cm-2 for the 670 nm LED and 7.0 mW cm-2 for the 735 nm LED, an intensity over 10-fold above the level necessary to saturate the system with light.

30

Figure 1.S3. Plastoquinone (PQ) pool size in wild type Cyanothece miami cells, as well as cells in the presence of 50M flavone and 20M rotenone. The area between the fluorescence induction curve in the presence and absence of DCMU represents the size of the PQ pool, which does not change with the addition of either NADH dehydrogenase inhibitor.

31

Figure 1.S4. Simultaneous detection of H2 production and intracellular reduced pyridine nucleotide concentration during dark autofermentative conditions. H2 was measured electrochemically by Pt/Ir electrode, while intracellular NAD(P)H was measured by fluorescence (see SI Methods). The redox poise of the cell (blue) increases during the course of dark autofermentative time.

32

Additional Materials & Methods Taxonomic Verification The following primers were used to amplify the 16S rDNA region for taxonomic identification: CYA106F: 5-CGG ACG GGT GAG TAA CGC GTG A-3, CYA359F: 5-GGG GAA TYT TCC GCA ATG GG-3, CYA781R(a): 5-GAC TAC TGG GGT ATC TAA TCC CAT T-3, and CYA784R(b): 5-GAC TAC AGG GGT ATC TAA TCC CTT T-3 (1). Amplification parameters were as described in the aforementioned reference, and sequencing was done by GeneWiz (South Plainfield, NJ).

Electrochemical Hydrogen Detection Our second (current) generation electrochemical reverse Clarke-type microcell is based on the schematic of a fuel cell for measuring dissolved H2 concentration. It utilizes a thin (5 m) Teflon membrane-covered Pt/Ir electrode biased at +220 mV for H2 oxidation. To increase the instrumental sensitivity and stability, a 20 m thick proton conducting membrane made of DuPont DE 1020 Nafion (Ion Power, Inc., USA) doped with black micro-dispersed Pd particles was inserted between the Teflon membrane and the Pt/Ir electrode. This 2nd generation H2 microcell has an increased sensitivity of ~2 x 10-9 Coulomb H2, fast response time (100 ms), and micro-volume for samples (6.5 L). During measurement both H2 and AgCl are constantly consumed in redox reactions, allowing the instantaneous rate of H2 production to be measured, rather than simply the H2accumulation.

Optical fluorescence unit for NAD(P)H measurement. The device is based on sample illumination at an angle of 45 by pulsed (250 ms) UV-367 nm LED (Nichia, Japan). An additional interference filter (365 10 nm, Intor, USA) is used to greatly reduce artifacts from excitation light on measuring signal. NAD(P)H fluorescence is selected by an interference filter 450 30 nm (Intor, USA), and the output signal from the amplified photodiode
33

S5591 (Hamamatsu, Japan) is filtered by a preamplifier in the DC-100 Hz range, with gain 20X (Model 113, EG&G, USA). All parts of the optical unit are constructed with black Delrin that has a very low UV parasitic fluorescence. In order to further reduce background UV fluorescence by a factor of 3-5 fold, the irradiated parts were covered by carbon containing black enamel NG SF 100 (Dupli-Color, USA). Intermittent UV-365 nm measuring light has a negligible effect on dark H2 production in Arthrospira maxima. The UV LED and fluorescence photodiode are controlled by a National Instruments data acquisition card. Each 200 ms fluorescence measurement is immediately followed by a 200 ms window where background fluorescence (without UV excitation) is measured. Measurements are taken every second and individual data points are bunched into groups of 10 and averaged, in order to increase signal relative to noise.

Supplementary References 1. Nubel U, Garcia-Pichel F, & Muyzer G (1997) PCR primers to amplify 16S rRNA genes from cyanobacteria. Appl. Environ. Microbiol. 63(8):3327-3332.

34

Chapter 2: Methods to Increase Catabolic Rate in Cyanothece for Increased Biological H2 and Ethanol Production

Co-authors: Gennady M. Ananyev, Yu Xu, Donald A. Bryant, and G. Charles Dismukes

35

Abstract Cyanobacteria of the genus Cyanothece are promising biotechnological candidates for the production of molecular hydrogen and liquid fuels (e.g. ethanol) from photosynthetically accumulated glycogen. Herein, three approaches are presented to increase the catabolic rate of Cyanothece and increase H2 production from the system. A 1.6-fold increase in fermentative H2 production is observed when the ATP sinks of cyanophycin synthesis and amino acid import are introduced to the cell by supplementation with NH4Cl and Glu/Gln, respectively; a 2.5-fold increase in H2 production is observed when Cyanothece is subjected to a co-fermentation with a transgenic strain of Synechococcus sp. PCC 7002 overproducing lactic acid. The maximal observed increase in H2 production was 9-fold, when Cyanothece cells were forced to hyper-accumulate carbohydrate by growth in high osmotic strength medium and later subjected to hypo-osmotic salt stress to induce catabolism of hyper-accumulated carbohydrate.

36

Introduction The solar driven splitting of water offers a potential source of renewable H2, but the efficiency and scalability of current solar based methods are inadequate as alternatives to fossil derived H2. This has spurred research into microorganisms that convert water into H2 and O2 using solar energy (1-2). One such class of organisms that exhibit this behavior is the cyanobacteria. Cyanobacteria are a diverse group of prokaryotes that perform oxygenic photosynthesis and are credited with converting the earths atmosphere from reducing to oxidizing approximately 2.8 billion years ago (3). Besides their use as biological hydrogen producers, cyanobacteria and microalgae have also gained significant attention for their potential as biodiesel and biomass feedstocks (4-5) and because they can be grown over non-arable land and thereby dont compete with food crops (6). Biological hydrogen production can be mediated by either of two broad classes of enzymes, the hydrogenases or the nitrogenases (7-9). Both classes are O2 sensitive, requiring anoxic environments to function maximally (10). Herein we focus on hydrogen production by Cyanothece, a genus of unicellular, aerobic, nitrogen-fixing (diazotrophic) cyanobacteria (11). Unlike the heterocyst-forming diazotrophic cyanobacteria (e.g. Nostoc and Anabaena spp.) that spatially separate oxygen-evolving photosynthesis from oxygen-sensitive nitrogen fixation (10), Cyanothece performs strictly regulated temporal separation of photosynthesis and nitrogen fixation (12). This temporal separation ensures that an intracellular anoxic environment conducive to nitrogenase and hydrogenase activity is maintained during the dark period of the diurnal cycle. Being capable of quickly achieving and maintaining anoxia, Cyanothece species are promising candidates as platforms for renewable energy generation (13-15). In addition, because a community of Cyanothece cells will possess nitrogenase in each individual cell, rather than solely within the heterocysts (which differentiate at a frequency of approximately 8% (16)), a Cyanothece culture

37

will possess more nitrogenase protein per volume of culture than heterocytous nitrogen-fixing cyanobacteria. The enzymatic reduction of dinitrogen to ammonia in Cyanothece is achieved by an ATP dependent reaction catalyzed by Mo-nitrogenase (nitrogenase hereafter), the most common and most efficient of the nitrogenase classes. For extensive reviews of the Mo-nitrogenase and the alternative nitrogenases, see reviews by Bothe and Tsygankov (7-8). Cyanothece species accumulate carbohydrate, primarily in the form of glycogen granules, during the day and subsequently degrade this carbohydrate at night to provide the needed energy and reductant for nitrogenase function and O2 respiration (17). Cyanothece species are therefore among the best candidates for biological H2 production because of this natural diurnal cycling, efficient conversion of intracellular carbohydrate to energy, efficient respiration and intracellular anoxia, and presence of both classes of hydrogen producing enzymes (13-14). Cyanothece sp. Miami BG 043511 is a marine strain which possesses three hydrogen metabolizing enzymes: the bidirectional hydrogenase (Class 3d, bidirectional NAD-linked H2ase), the membrane bound respiratory or uptake hydrogenase (Class 2a, cyanobacterial uptake hydrogenase, typically found co-expressed with nitrogenase) (18-19), and the above mentioned nitrogenase. NADH and reduced ferredoxin (FDx) are the direct substrates for the bidirectional hydrogenases and nitrogenases, respectively, as established by in vitro assays of enzymes from multiple microorganisms (20-21). The NADH-dependent reduction of H+ to H2 by hydrogenase is often cited as thermodynamically unfavorable, as the standard midpoint potential (Eh) at pH 7 for NAD+/NADH is -320 mV, compared to -420 mV for H+/H2. However, standard conditions refer to 1 bar H2 pressure, conditions that are never found in biological cells. By contrast, the calculated thermodynamic redox potential for H+/H2 is nearly identical to that of NAD+/NADH in a 0.1% H2 atmosphere (1000 ppm, or 0.001 bar) (22). Consequently, pyridine nucleotides can serve as efficient reductant sources under biological conditions, especially when H2 backpressure is
38

minimized. In addition to NADH and FDx, NADPH can be utilized indirectly, either by the reduction of NADH catalyzed by the pyridine nucleotide transhydrogenase (encoded by genome but in vivo activity unknown), or by the reduction of the plastoquinone (PQ) pool via NAD(P)H dehydrogenase and subsequent reduction of ferredoxin through excitation of PSI (as extensively described in (15)). Previous work (15) has identified the functional pathways providing reductant to nitrogenase in Cyanothece and suggested various approaches to increase the biological H2 yield from the strain. While Cyanothece is capable of producing both photo-H2 and dark, autofermentative H2, the rate of H2 production from the photo- pathway is significantly increased relative to the dark- pathway. Biotechnologically, this affords the most rapid conversion of intracellular reductant (which is shared between the dark- and photo- pathways) to H2 (2-fold increase observed in (15)). A system has been envisioned where reductant would accumulate under anoxia and then be quickly extracted (by conversion to H2) during an illumination period. This cycle could be repeated ad infinitum. In the current article, we describe mechanisms to increase the catabolic rate in Cyanothece such that the cycling of carbohydrates is improved and thereby the available reductant for conversion of protons to molecular H2 is increased. Herein we present three such methods: 1) introduction of an ATP sink, 2) hypotonic salt stress to promote catabolism of osmolytes, and 3) co-fermentation with organic acid producing cyanobacteria to supply exogenous sources of reductant. Previous studies have shown that organic substrates (e.g. glucose, pyruvate, and ethanol) as well as inorganic substrates (e.g. sulfide) can supply the reductant for nitrogenase in addition to intracellular glycogen (23-24). In the co-fermentation method we present here, in addition to providing Cyanothece with a biological source of exogenous organic acids and alcohols, the consumption of said fermentative products (produced by the non-diazotrophic Synechococcus sp. PCC 7002) is projected to increase autofermentative H2 production by Synechococcus by milking its fermentative products and decreasing the backpressure of its fermentative reactions. Indeed, the
39

removal of excreted products has recently been shown to increase biological H2 production in the cyanobacterium Arthrospira maxima (Ananyev, et al., submitted).

40

Methods Growth of cultures Cyanothece sp. Miami BG 043511 (Cyanothece miami) was obtained from the University of Hawaii Culture Collection, where it was maintained as Synechococcus sp. Miami BG 043511. Cultures were grown in ASP2 medium (11) without combined inorganic nitrogen at 30C under diurnal conditions (12 hr light, 12 hr dark) with a light intensity of ~20 E s-1 m-2, as described previously (15). Cultures were not bubbled with any gases nor shaken. Synechococcus sp. PCC 7002 was grown in liquid A+ medium (25) supplemented with 2 M Ni2SO4 and sparged continuously with 2% CO2 in air (v/v). Cells were grown at 35 C for three days to mid-late exponential phase in 250 mL flasks using white fluorescence lights at an intensity of approximately 200 E m-2 s-1.

Anthrone assay and soluble sugar extraction Total carbohydrate was measured using the anthrone reagent (26). 1 mL aliquots of cell suspensions were pelleted by centrifugation, flash frozen in liquid N2, and stored at -80 C until time of assay. Said cell pellets were boiled in 100 L of 33% H2SO4 (in water) for one hour before 900 L of anthrone reagent (0.2 g anthrone (Sigma) per 100 mL of 71% sulfuric acid in water) was added and cells subsequently boiled for an additional 15 minutes. The difference in absorption of 620 and 820 nm was used for quantification. Standards were prepared using glucose at concentrations of 10 150 g per 100 L. For the extraction of soluble sugars, cell pellets were resuspended in 80% ethanol and incubated at 65 C for 3 hours. Hexane was used to extract pigments and the ethanol/water phase was collected, evaporated in a centrivap at 60 C until completely dry (27), and the resulting pellet resuspended in dH2O and analyzed by the aforementioned anthrone method.

Hydrogen measurements by gas chromatography

41

H2 gas was measured from the headspace of air tight glass incubation vials by gas chromatography. Gas samples (200 L) were taken from the headspace of the incubation vials by air-tight syringe (Hamilton) and injected into the gas chromatograph. For the data presented in Figure 2.2, a GowMac (Newark, NJ) Series 580 gas chromatograph equipped with a 6 13X molecular sieve packed column and thermal conductivity detector was used. For the data presented in Figure 2.4, a PerkinElmer Clarus 680 GC equipped with a Restek ShinCarbon ST 80/100 packed column (2m length, 2 mm ID) and thermal conductivity detector was used. Ultra high purity argon (Airgas, Inc.) was used as the carrier gas for both instruments, running at 30 mL min-1.

Quantification of excreted fermentative products by 1H NMR Fermentation products were measured utilizing cryoprobe-assisted proton nuclear magnetic resonance spectroscopy (1H NMR) on a Bruker Avance-II 500 MHz spectrometer at a controlled temperature of 295 K as reported in (28). Samples from the extracellular fermentation medium were spiked with 10% D2O containing 30 ng mL-1 3-(trimethylsilyl)propionic-2,2,3,3-d4 acid, sodium salt (TSP, Sigma). Excitation sculpting was used for water suppression (29). Peaks corresponding to each metabolite were integrated and normalized to the internal standard (TSP). Linearity was observed over the concentration range between 10 M and 1 mM. TopSpin 2.1/ICON software was used to acquire the data and MNova v.5.2.5 (Mestrelab Research S.L., Satiago de Compostela, Spain) was used for analysis and integration of signals.

42

Dissolved H2 rate electrode & LED illumination system The first generation of our home-built H2 microcell is described elsewhere (30). Our current generation electrochemical microcell is also a reverse Clark-type electrode for measuring dissolved H2 concentration. The second generation of home-built electrochemical cell is based on fuel cell technology for measuring the rate of dissolved H2 concentration change. It uses a thin 5 m Teflon membrane covering a 6 mm Pt/Ir electrode. For H2 oxidation the electrode is poised at +220 mV vs. Ag/AgCl in 100 mM KCl. Between the Teflon membrane and the Pt/Ir electrode is a 20 m thick H+conducting membrane Nafion DE 1020 (Ion Power, Inc., USA) doped with micro-dispersed Pd black particles. This microcell has an increased sensitivity of ~2 x 10-9 Coulomb H2 (concentration 3.2 nM H2), fast response time (100 ms), and micro-volume sample chamber (6.5 L). During measurement, H2 is constantly consumed by oxidation at the Pt/Ir electrode, allowing the instantaneous rate of H2 production to be measured. The microcell responds linearly to H2 and performs for several months using an oversized Ag/AgCl reference electrode (~0.5 g AgCl) to consume the H+ product of H2 oxidation. In this article, H2 production rate is presented as the current (nA) arising from the oxidation of dissolved H2. H2 yield was determined by integration with respect to time to yield electrical charge, which was converted to moles of H2 by Faradays second law. The microcell is equipped with a near infrared light emitting diode (LED) at 73510 nm (FWHM) for selectively exciting Photosystem I, as described in (15).

For incubation, stationary phase cells were taken from photoautotrophic growth conditions between hours 10 and 11 after the onset of the light cycle (circadian dawn), concentrated (40x) by centrifugation, re-suspended in fresh media, and placed in the 6.5 l sample chamber of the electrochemical cell. The cell was covered by a quartz disc and removed from light, where cellular respiration created anoxia within 2-3 minutes. Dissolved H2 was continuously consumed as

43

described above. When pulsed illumination was supplied to the culture, it was done using repeated pulses of 10s duration, separated by 90s dark time, i.e. 10% duty cycle.

44

Results I. Catabolic rate and H2 production under various metabolic modes Figure 2.1 illustrates the total carbohydrate content of Cyanothece miami cells taken from the stationary phase of growth (approximately 4 weeks). Carbohydrate (measured by the anthrone assay) was quantified from cells before and after 20 hours of incubation under two physiological conditions: i) dark anoxic and ii) light anoxic. For the light anoxic condition (for photo-H2 production), light was supplied as continuous white light from cool fluorescent lamps (200 E m-2 s1).

Carbohydrate was consumed in both conditions, with maximal consumption (50%) under

fermentative conditions (dark/anoxic), as is expected because the lowest energy payout per glucose equivalent catabolized is obtained under this mode. Hydrogen production is not observed under dark oxic and/or light oxic incubations, consistent with the behavior of hydrogen evolving enzymes that are inactivated by O2. Interestingly, H2 production from light/anoxic conditions (photo-H2) is 15-fold higher than dark, fermentative H2 from the strain. Conditions supporting photo-H2 production (0.95% 0.08% headspace accumulation) correspond with carbohydrate consumption of less than 30%. Accessing the remaining carbohydrate (70% of the initial amount) is therefore a major research thrust in efforts to increase H2 production from Cyanothece. Scheme 2.1 presents an overview of pathways for carbohydrate degradation in Cyanothece along with reactions that consume both ATP and reductant. Three mechanisms to increase the catabolic rate in Cyanothece are proposed and presented herein: 1) introducing ATP sinks which stimulate glucose oxidation, 2) inducing the catabolism of osmolytes, and 3) supplementing the system with exogenous lactate. Because carbohydrate is not exhaustively catabolized under photo-H2 producing conditions, we have focused on increasing the catabolic rate in Cyanothece rather than increasing the absolute carbohydrate content at the start of anaerobic incubation. Nevertheless, insufficient carbohydrate content can also limit photo-H2 production, as observed in Figure 2.S1. Cyanothece cultures were
45

incubated in an electrochemical cell (6.5 L) under dark anaerobic conditions and subjected to four periods of pulsed illumination. Both dark- and photo-H2 production was significantly increased (3fold over 20 hours) when exogenous reduced carbon was provided to the culture (5% glycerol shown here). A similar increase is observed when glucose and other mono- and disaccharides are supplemented to the culture.

II. Introduction of ATP sinks to fermentative metabolism Cyanothece possesses a bidirectional [NiFe] hydrogenase that is responsible for the production of H2 under fermentative conditions (15). We therefore aimed to increase this dark, hydrogenase-mediated pathway to H2 evolution by introducing ATP requiring reactions to Cyanothece. The addition of ATP sinks to Cyanothece is hypothesized to up-regulate the flux of carbon catabolism through glycolysis, as this is the soles means of producing ATP (through substrate level phosphorylation) in the absence of terminal electron acceptors (i.e. O2). Figure 2.2A shows H2 production over 7 days from Cyanothece cultures in response to the addition of NH4Cl. Cyanothece strains store nitrogen in the non-ribosomally synthesized amino acid polymer cyanophycin (multi-L-arginyl-poly (L-aspartic acid)) (11, 31). This synthesis is catalyzed by cyanophycin synthetase in an ATP requiring reaction (32). Cyanothece was incubated in 10 mL air-tight glass vials which were covered from light by coating with black electrical tape. For the dark fermentative experiments presented here, 10 mM glucose was supplied to all of the samples, including the control, to eliminate the effect of carbohydrate limitation and maximize the effect of introduction of ATP sinks. While the addition of 1 M NH4+ slightly increased fermentative H2 production, higher concentrations of NH4+ were inhibitory towards H2 (50% inhibition with 5 M, 85% inhibition with 10 M). However, the profile of other excreted fermentative products (Figure 2.2B) indicates that formate and ethanol production is significantly increased in the presence of 5 and 10 M NH4+. This indicates a cell with
46

an increased redox poise: the production of formate allows the consumption of pyruvate (produced via glycolysis to provide ATP) by a pathway that does not produce additional internal reductant (the alternative pathways via (a) pyruvate:ferredoxin oxidoreductase and (b) pyruvate dehydrogenase complex both do generate additional reductant within the cell). Additionally, the production of ethanol allows for the excretion of acetyl-CoA by a fermentative reaction that recycles the maximal amount of NADH to NAD+ per turnover (2 per cycle). Another process that requires ATP is amino acid import in Cyanothece. Glutamine and glutamate are imported by ABC-type transporters (33) as are most amino acids in Cyanothece. Hydrogen production under dark fermentative conditions from Cyanothece in response to 2 mM glutamate and glutamine is shown in Figure 2.2C. The addition of either amino acid stimulates hydrogen production 1.6 fold.

III. Hyper-accumulation of carbohydrates and hypotonic salt stress Cyanothece miami is a marine strain normally grown in the laboratory in ASP2 medium without combined nitrogen. This growth media contains 308 mM NaCl (total osmolarity approximately 330 mM). To increase the carbohydrate content in Cyanothece, cells were grown in augmented growth media: ASP2 supplemented with additional NaCl (between 250 mM and 1.0 M additional) or with the non-ionic osmolyte polyethylene glycol (PEG-3200). Figure 2.3A illustrates the total carbohydrate composition of Cyanothece (normalized to chlorophyll content) from growth in media of increasing osmolarity. Under medium-high light intensity (110 E m-2 s-1) the presence of additional NaCl inhibited growth rate, while under low light intensity (10 E m-2 s-1), the growth rate was comparable among the NaCl concentrations tested. Interestingly, the presence of PEG3200 stimulated growth of Cyanothece under the higher light intensity. Strains of the genus Cyanothece are known to produce exopolysaccharides (34-36) which have been hypothesized to assist the cell in maintaining suboxic conditions during its dark phase. By providing a physical
47

barrier to atmospheric O2, we hypothesize that PEG addition aids young cells which have not yet produced significant amounts of exopolysaccharide through this same mechanism. Figure 2.3B illustrates total carbohydrate and soluble sugars (osmolytes) accumulated in response to growth in increased osmotic media. It shows that soluble sugars (red bars) increase in response to growth in hyperosmotic conditions; they are the subclass of the total carbohydrates that increase in the strain (insoluble glycogen content, obtained from subtracting the soluble content from the total content, does not appear to significantly increase). Indeed, hyperaccumulation of soluble sugars in response to high salt has been previously reported in Arthrospira platensis (37) and maxima (38). To induce hypercatabolism of osmolytes, Cyanothece cells were taken from the growth phase (where they accumulated increased carbohydrate), pelleted by centrifugation, and resuspended in regular osmotic strength ASP2 media. Increased photo-H2 production during 20 hours of continuous light incubation is presented in Figure 2.4A. This effect of increased photo-H2 is coincident with the level to which carbohydrate was hyper-accumulated (and therefore to the extent of hypo-osmotic shock that the cultures experienced). Similarly grown cultures were also assayed for hydrogen production by incubation with an electrochemical cell that continually monitors dissolved H2 production. This electrode is poised at a bias of +220 mV such that it continually consumes (oxidizes) H2 produced from the culture. Under these milking conditions, hydrogen does not accumulate above the culture and therefore the H2 producing enzymes do not suffer from backpressure of product inhibition or uptake of H2 in the headspace. Because Cyanothece miami contains the unidirectional uptake (respiratory) hydrogenase, monitoring H2 production electrochemically is expected to produce the highest H2 production rates as the activity of the uptake enzyme would be minimized. Figure 2.4B illustrates the photo-H2 yield from Cyanothece cells grown in increased osmolarity over 4 separate illumination periods. It indicates, as above, that the extent to which photo-H2 production is increased depends upon the extent of hyper-accumulation of carbohydrate and accordingly the extent of hypo-tonic salt stress
48

prior to incubation. The maximal increase in photo-H2 (9-fold) was observed here, when cells 1.0 M were grown in 1.3 M NaCl and resuspended in 0.3 M NaCl prior to incubation.

IV. Co-fermentation with organic acid producing cyanobacterium Previous work has shown that the addition of organic acids and alcohols can supplement Cyanothece and lead to increased photo-H2 production (23). The cyanobacterium Synechococcus sp. PCC 7002 is a fast growing, prolific cyanobacterium that produces lactate as its major fermentative product. In fact, the removal of lactate dehydrogenase from the strain was targeted as a means to improve autofermentative H2 production from the strain. Its removal provided a 5-fold increase in autofermentative H2 production, as excess internal reductant was available for hydrogenase (39). A transgenic strain has been constructed in which lactate dehydrogenase is overexpressed on a pAQ1-Ex plasmid (40) behind the strong cyanophycin promoter (PcpcAB). This strain produces lactate (and interestingly, acetate) continuously under photoautotrophic conditions. In an effort to supplement Cyanothece with lactate produced from Synecococcus, the strains were grown separately (Cyanothece in ASP2 medium lacking combined nitrogen and Synecococcus strains in A+ medium with 18 mM NaNO3) and combined (Figure 2.5B) immediately prior to anaerobic incubation (50:50 ratio of culture (Cyanothece:Synechococcus) on a per volume basis, 20:80 ratio on a per weight basis). The combined culture was exposed to a 20 hour anaerobic incubation with 4 periods of pulsed illumination (60 minutes in duration each) separated by dark time. Figure 2.5A illustrates the cumulative yield of H2 from the combined culture. The vertical white rectangles illustrate the four pulsed illumination periods. Co-fermentation of Cyanothece and Synechococcus 7002 WT produced a 1.8-fold increase in cumulative H2 production and cofermentation of Cyanothece with the mutant Synechococcus 7002 overexpressing lactate dehydrogenase (LdhAEx) showed a 2.5-fold increase.

49

Discussion In the present work we have demonstrated three methods to increase the catabolic rate in Cyanothece to supply increased internal reductant to reduce protons to molecular H2. The first approach is centered upon the introduction of an ATP sink. As mentioned previously, Cyanothece is capable of H2 production catalyzed by either hydrogenase or nitrogenase. Because nitrogenase produces H2 through an ATP-requiring reaction, introducing a separate sink to consume ATP is not a promising approach, as it would steal ATP that could otherwise be used by nitrogenase. The hydrogenase-mediated pathway, on the other hand, reduces H+ to H2 coupled to the oxidation of NADH to NAD+ and is not dependent on ATP hydrolysis. We therefore chose to investigate the addition of ATP sinks to the hydrogenase-mediated pathway, i.e. the dark fermentative pathway. We note that the addition of nitrogen compounds (NH4+ and amino acids, as described herein) has been shown to inhibit nitrogenase. Therefore by utilizing these compounds to induce ATP sinks, we ensure that the effects observed are due to H2 production by hydrogenase (the dark fermentative pathway). The addition of ATP sinks to Cyanothece was hypothesized to up-regulate the flux of carbon catabolism through glycolysis, as this is the soles means of producing ATP (through substrate level phosphorylation) in the absence of terminal electron acceptors (i.e. O2). Similar approaches have been taken in Cyanothece sp. PCC 7822 where FCCP was shown to increase fermentative hydrogen production (41) and in the non-photosynthetic bacterium E. coli, where the introduction of futile ATPases (to decrease ATP availability) also increased fermentative hydrogen production (42). This hypothesis appears to be valid, as the addition of 1 mM NH4+, 2 mM glutamate, or 2 mM glutamine each slightly stimulated fermentative H2 production. Interestingly, the strengthening of the ATP sink of cyanophycin synthesis (by the addition of more NH4+) caused an increased intracellular redox poise (evidenced by the increased flux to the carbon fermentative reactions that process glycolytically derived pyruvate and lead to excretion of carbon fermentative products) but did not
50

correlate with increased H2 production. Of note, production of the fermentative product ethanol was increased 40-fold. Ethanol is an energy rich molecule that has been the focus of many early generation biofuels. Considerable public and scientific interest has been focused on advanced method to create ethanol that have improved economics and decreased environmental impact than corn-based ethanol (43). Such methods have included cellulosic ethanol (44), ethanol production from algal biomass, and direct production of ethanol from cyanobacteria (38, 45). This increased production of fermentative ethanol in response to the addition of an ATP sink is therefore an attractive mechanism that could lead to increased ethanol yields from industrial ethanol producing strains. The presence of nitrogenase in unicellular Cyanothece is itself an ATP sink under native conditions. In fact, the interdependence of hydrogenase and nitrogenase within the context of dark metabolism in Cyanothece is complex and may explain why these organisms exhibit such intense photo-H2 signals. As mentioned above, nitrogen fixation is an energy intense process, and in the absence of light, the energy required to maintain nitrogenase must come from the breakdown of intracellular carbohydrates, e.g. glycogen. In the absence of O2 (conditions required for optimal nitrogenase activity and expected to occur nightly in communities of Cyanothece), the glycolytic breakdown of glucose is the sole means for ATP generation to the cell. This natural ATP sink of nitrogenase, therefore, is expected to demand a sufficient catabolic rate to provide nitrogenase with ATP and reductant. This high metabolic rate (demanded by nitrogenase) would support an increased redox poise within the cell and allow the rapid onset of nitrogenase-mediated photo-H2 production when anaerobic cells are illuminated (a feature observed in (15)). The second approach presented herein was the hyper-accumulation of carbohydrate in Cyanothece in response to growth in increased osmotic strength medium and the subsequent hypoosmotic salt stress to induce catabolism of said carbohydrates. This approach led to increased photo-H2 production corresponding to the extent of hypo-osmotic stress felt by the culture. The
51

maximal increase observed from this approach was a 9-fold boost in photo-H2 production and was observed when Cyanothece was incubated under milking conditions to continually consume the produced H2 and thereby eliminate H2 backpressure on the system. While nitrogenase is relatively insensitive to H2 feedback inhibition because of its reaction coupling to ATP hydrolysis (it is effectively irreversible), Cyanothece possesses a dedicated uptake hydrogenase that serves to recycle H2 produced by nitrogenase. This enzyme is unidirectional due to its membrane association mechanism. It immediately passes electrons along a membrane cascade after oxidizing H2, and in this sense is irreversible. By milking Cyanothece of its photo-H2, the substrate (and accordingly, activity) of this uptake enzyme is minimized. The approach of hypo-osmotic salt stress has recently been studied in Arthrospira maxima (38) and was shown to cause an increased catabolic rate which correlated with the increased production of carbon fermentative products but not with increased H2 production. Carrieri, et al. have suggested that the recycling of NADH to NAD+ by the pathways which produce excreted carbon products may be a mechanism for the cell to regenerate NAD+ and to eliminate small molecules which are osmotically active from the cell. An additional explanation could be proposed based upon the catabolic pathways that produce and utilize NADH and NADPH within the cell. There are two catabolic pathways for breakdown of carbohydrates under anoxia distributed among bacteria: Glycolysis (Embden-Meyerhof-Parnas pathway, EMP) and the Oxidative Pentose Phosphate pathway (OPP), summarized briefly in Scheme 2.1 and in detail in Figure 2.S2. In cyanobacteria their presence has been deduced from genome sequences and experimentally confirmed by the detection (46) and quantification (25, 38) of excreted fermentation products and intracellular metabolite pools (47). The yields of ATP and NAD(P)H differ between the two. While glycolysis provides NADH and ATP, the OPP pathway provides NADPH (more reductant per glucose catabolized) and can produce ATP if coupled to EMP in a hybrid pathway described in Figure 2.S2. The [NiFe] hydrogenase requires NADH as its cognate electron donor (20-21, 48) while both NADH
52

and NADPH can serve as indirect electron donors to nitrogenase through non-photochemical reduction of the plastoquinone pool catalyzed by NADH dehydrogenase types 1 and 2 (15). Electrons are then transferred from reduced plastoquinol (PQH2) through PSI and ferredoxin, ultimately to nitrogenase. Osmolytes (synthesized by growth in hyperosmotic conditions) may be consumed through this OPP pathway upon hypo-ionic salt stress in cyanobacteria, providing NADPH to the cell. This NADPH cannot be consumed by the [NiFe] bidirectional hydrogenase directly, but could power photo-H2 production via nitrogenase through the pathway mentioned above. Further, this pathway couples the oxidation of NAD(P)H to a more favorable electron acceptor: the PQ pool. The midpoint potential of this pool (oxidized under these conditions because of illumination) is ~0 mV, compared to -420 mV for the [NiFe] hydrogenase at standard conditions. This pool is therefore poised to favorably accept electrons from NADH or NADPH (-320 mV) that could otherwise provide electrons to competing fermentative pathways. The activity of PSI to energize these electrons to the level of reduced ferredoxin (substrate for nitrogenase) allows the production of photo-H2 by this pathway. Additionally, similar to the osmotic hypothesis discussed in (38), the oxidation of glucose through the OPP pathway would excrete more carbon equivalents as CO2 and accordingly produce less low molecular weight metabolites (e.g. pyruvate, acetate, lactate) which are osmotically active. The third method presented here, co-fermentation of Cyanothece with the organic acid producing strains of Synechococcus, led to a 2.5-fold increase in cumulative H2 production from the combined culture. While this increase was partly due to increased photo-H2 production from Cyanothece in the presence of excreted organic acids (lactate), as was previously reported by Luo & Mitsui (23), the dark H2 production rate from the combined culture was significantly increased as well. We hypothesize that this increased dark rate was mostly attributable to increased autofermentative H2 production from Synechococcus. The milking of the Synechococcus cells by the consumption of its product H2 (electrochemically) along with its carbon fermentative products
53

(through uptake by Cyanothece) is hypothesized to stimulate its catabolic rate and fermentative reactions. A similar effect of stimulated catabolic rate and autofermentative H2 production in response to the removal or consumption of excreted products was recently described in the cyanobacterium Arthrospira maxima (Ananyev, et al., submitted). Interestingly, the LdhAEx strain of Synechococcus sp. PCC 7002 has been shown to excrete acetate in addition to lactate under photoautotrophic conditions (Bryant, unpublished data). The exact mechanism of this increased acetate production is not yet known. The green alga Chlamydomonas reinhardtii has been shown to photo-assimilate acetate under anaerobic conditions (49) through functioning of the citric acid and glyoxylate cycles utilizing the [FeFe] hydrogenase as a terminal electron acceptor. It is possible that a similar mechanism could be occurring in Cyanothece cultures in the presence of acetate, utilizing nitrogenase as the effective terminal electron acceptor. Further studies will have to study this potential occurrence, perhaps through the addition of [13C]-labeled acetate and/or succinate and tracking of the 13C label among intracellular metabolic intermediates using LC-MS/MS. Synechococcus grows considerably faster than does Cyanothece, doubling every 3-4 hours compared to once every 24 hours compared to Cyanothece miami. Additionally, because it cannot fix atmospheric nitrogen, Synechococcus is routinely grown on A+ medium, a salt water-like medium containing 18 mM NaNO3. Cyanothece is grown in ASP2 medium in the absence of combined nitrogen. While its growth rate is slower without combined inorganic nitrogen, Cyanothece is less likely to be contaminated without nitrate in the medium, and said growth conditions ensure that nitrogenase is expressed and functional among Cyanothece cultures. For these reasons, growth of the two strains was carried out separately and the strains combined immediately prior to the time of co-fermentation incubation. If the cells were attempted to be co-cultivated on medium containing NaNO3, we anticipate the Synechococcus sp. outcompeting the Cyanothece sp.

54

Cyanothece is a promising genus of cyanobacteria for biotechnological applications due to its ability to produce H2, ethanol, exopolysaccharides, and cyanophycin. It is capable of a wide range of metabolisms and specifically is capable of maintaining intracellular anoxia conducive to nitrogenase and hydrogenase functioning. Through the experiments presented here, we have advanced our understanding of what governs carbohydrate catabolism in Cyanothece and demonstrated that we can improve its rate of catabolism through various physiological approaches. We look forward to further studies to increase our understanding, overcome remaining obstacles, and improve the functional metabolic pathways in Cyanothece with an end goal of fully realizing the biotechnological potential of the genus.

55

Figures

Figure 2.1. Total carbohydrate composition of Cyanothece miami before and after anaerobic incubation with (light) and without (dark) continuous illumination for 20 hours. The corresponding H2 production yield from each (% H2 in headspace above culture) is listed above each column. Less than 30% of the carbohydrate pool is consumed in 20 hours from a light/anaerobic incubation (3rd column), conditions which produce the maximal yield of H2 (15x larger yield than dark anoxic H2). Accessing the remaining carbohydrate is identified as a primary target for increasing the photo-H2 yield.

56

Figure 2.2. Introduction of an ATP sink under fermentative conditions (e.g. cyanophycin synthesis (A/B) or amino acid import (C)) causes an increased catabolic rate and corresponding increased redox poise within the cell. (A) The addition of low concentration NH4+ stimulates fermentative H2 production while higher concentrations inhibit H2 but (B) stimulate the production of other fermentative products that recycle NADH to NAD+. (C) The addition of glutamine and glutamate (2 mM each) stimulates H2 production under fermentative conditions.

57

Figure 2.3. Carbohydrate content of Cyanothece in response to growth in high osmotic strength medium. Augmented growth medium with supplemented NaCl or PEG-3200 causes increased total carbohydrate (A) and soluble sugar (B) accumulation in Cyanothece.

58

Figure 2.4. Increased photo-H2 in response to hypo-osmotic salt stress. Measured by GC in panel A and measured by dissolved H2 rate electrode in panel B. (A) The extent of increase in photo-H2 corresponds with the extent of previously hyper-accumulated carbohydrates and accordingly, the extent of hypotonic salt stress. (B) Quadrants 1 through 4 represent the photo-H2 yield from each of the four illumination windows of the incubation within the electrochemical micro-cell.

59

Figure 2.5. Co-fermentation of diazotrophic Cyanothece miami with Synechococcus PCC 7002 WT (red) and mutant LdhAEx (overexpressing lactate dehydrogenase, blue). Lactate is the predominate fermentative product from PCC 7002 WT. The LdhAEx strain produces lactate and acetate. Experimental design described in panel B.

60

Scheme 2.1. An abbreviated scheme indicating pathways for carbohydrate degradation and ATP sinks within Cyanothece. Approaches reported herein to increase catabolic rate in Cyanothece miami are shown: (1) introduction of ATP sinks (cyanophycin synthesis, amino acid import), (2) increased accumulation and subsequent degradation of osmolytes, and (3) supplementation with reduced organic acids (e.g. lactate) to provide additional reductant (NADH).

61

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. Kruse O & Hankamer B (2010) Microalgal hydrogen production. Current Opinion in Biotechnology 21(3):238-243. Quintana N, Van der Kooy F, Van de Rhee M, Voshol G, & Verpoorte R (2011) Renewable energy from Cyanobacteria: energy production optimization by metabolic pathway engineering. Applied Microbiology and Biotechnology 91(3):471-490. Peschek GA, Obinger C, & Paumann M (2004) The respiratory chain of blue-green algae (cyanobacteria). Physiol Plant 120(3):358-369. Chisti Y (2007) Biodiesel from microalgae. Biotechnology Advances 25(3):294-306. Williams PJlB & Laurens LML (2010) Microalgae as biodiesel & biomass feedstocks: Review & analysis of the biochemistry, energetics & economics. Energy & Environmental Science 3(5):554-590. Dismukes GC, Carrieri D, Bennette N, Ananyev GM, & Posewitz MC (2008) Aquatic phototrophs: efficient alternatives to land-based crops for biofuels. Curr Opin Biotechnol 19(3):235-240. Tsygankov A (2007) Nitrogen-fixing cyanobacteria: A review. Applied Biochemistry and Microbiology 43(3):250-259. Bothe H, Schmitz O, Yates MG, & Newton WE (2010) Nitrogen Fixation and Hydrogen Metabolism in Cyanobacteria. Microbiol. Mol. Biol. Rev. 74(4):529-551. Ghirardi ML, Posewitz MC, Maness PC, Dubini A, Yu JP, & Seibert M (2007) Hydrogenases and hydrogen photoproduction in oxygenic photosynthetic organisms. Annual Review of Plant Biology 58:71-91. Gallon JR (1992) Reconciling the incompatible: N2 fixation and O2. New Phytologist 122(4):571-609. Reddy KJ, Haskell JB, Sherman DM, & Sherman LA (1993) Unicellular, aerobic nitrogenfixing cyanobacteria of the genus Cyanothece. J. Bacteriol. 175(5):1284-1292. Mitsui A, Kumazawa S, Takahashi A, Ikemoto H, Cao S, & Arai T (1986) Strategy by which nitrogen-fixing unicellular cyanobacteria grow photoautotrophically. Nature 323(6090):720-722. Min H & Sherman LA (2010) Hydrogen Production by the Unicellular, Diazotrophic Cyanobacterium Cyanothece sp. Strain ATCC 51142 under Conditions of Continuous Light. Appl. Environ. Microbiol. 76(13):4293-4301. Borodin VB, Rao KK, & Hall DO (2002) Manifestation of behavioural and physiological functions of Synechococcus sp. Miami BG 043511 in a photobioreactor. Marine Biology 140(3):455-463. Skizim NJ, Ananyev GM, Krishnan A, & Dismukes GC (2011) Metabolic pathways for photobiological hydrogen production by nitrogenase- and hydrogenase- containing unicellular cyanobacteria Cyanothece. Journal of Biological Chemistry. Wong FCY & Meeks JC (2002) Establishment of a functional symbiosis between the cyanobacterium Nostoc punctiforme and the bryophyte Anthoceros punctatus requires genes involved in nitrogen control and initiation of heterocyst differentiation. Microbiology 148(1):315-323. Schneegurt M, Sherman DM, Nayar S, & Sherman LA (1994) Oscillating behavior of carbohydrate granule formation and dinitrogen fixation in the cyanobacterium Cyanothece sp. strain ATCC 51142. J. Bacteriol 176:1586-1597. Vignais PM, Billoud B, & Meyer J (2001) Classification and phylogeny of hydrogenases. FEMS Microbiol Rev 25(4):455-501. Ludwig M, Schulz-Friedrich R, & Appel J (2006) Occurrence of hydrogenases in cyanobacteria and anoxygenic photosynthetic bacteria: Implications for the phylogenetic
62

17. 18. 19.

20. 21.

22. 23. 24. 25. 26. 27. 28.

29. 30.

31. 32.

33. 34. 35.

origin of cyanobacterial and algal hydrogenases. Journal of Molecular Evolution 63(6):758768. Maroti J, Farkas A, Nagy IK, Maroti G, Kondorosi E, Rakhely G, & Kovacs KL (2010) A Second Soluble Hox-Type NiFe Enzyme Completes the Hydrogenase Set in Thiocapsa roseopersicina BBS. Appl. Environ. Microbiol. 76(15):5113-5123. Burgdorf T, van der Linden E, Bernhard M, Yin QY, Back JW, Hartog AF, Muijsers AO, de Koster CG, Albracht SPJ, & Friedrich B (2005) The soluble NAD(+)-reducing [NiFe]Hydrogenase from Ralstonia eutropha H16 consists of six subunits and can be specifically activated by NADPH. Journal of Bacteriology 187(9):3122-3132. Vincent KA, Parkin A, & Armstrong FA (2007) Investigating and Exploiting the Electrocatalytic Properties of Hydrogenases. Chemical Reviews 107(10):4366-4413. Luo Y-H & Mitsui A (1994) Hydrogen production from organic substrates in an aerobic nitrogen-fixing marine unicellular cyanobacterium Synechococcus sp. strain Miami BG 043511. Biotechnology and Bioengineering 44(10):1255-1260. Luo Y-H & Mitsui A (1996) Sulfide as electron source for H2-photoproduction in the cyanobacterium Synechococcus sp., strain Miami BG 043511, under stress conditions. Journal of Photochemistry and Photobiology B: Biology 35(3):203-207. McNeely K, Xu Y, Ananyev G, Bennette N, Bryant DA, & Dismukes GC (2011) Characterization of a nifJ Mutant of Synechococcus sp. strain PCC 7002 Lacking Pyruvate:Ferredoxin Oxidoreductase. Appl. Environ. Microbiol.:AEM.02792-02710. Hassid WZ & Abraham S (1957) [7] Chemical procedures for analysis of polysaccharides. Methods in Enzymology, (Academic Press), Vol Volume 3, pp 34-50. Schoor A, Erdmann N, Effmert U, & Mikkat S (1995) Determination Of The Cyanobacterial Osmolyte Glucosylglycerol By High-Performance Liquid-Chromatography. Journal Of Chromatography A 704(1):89-97. Carrieri D, McNeely K, De Roo AC, Bennette N, Pelczer I, & Dismukes GC (2009) Identification and quantification of water-soluble metabolites by cryoprobe-assisted nuclear magnetic resonance spectroscopy applied to microbial fermentation. Magnetic Resonance in Chemistry 47(S1):S138-S146. Hwang TL & Shaka AJ (1995) Water Suppression That Works. Excitation Sculpting Using Arbitrary Wave-Forms and Pulsed-Field Gradients. Journal of Magnetic Resonance, Series A 112(2):275-279. Ananyev G, Carrieri D, & Dismukes GC (2008) Optimization of metabolic capacity and flux through environmental cues to maximize hydrogen production by the cyanobacterium "Arthrospira (Spirulina) maxima". Applied and Environmental Microbiology 74(19):61026113. Li H, Sherman D, Bao S, & Sherman L (2001) Pattern of cyanophycin accumulation in nitrogen-fixing and non-nitrogen-fixing cyanobacteria. Archives of Microbiology 176(1):918. Ziegler K, Diener A, Herpin C, Richter R, Deutzmann R, & Lockau W (1998) Molecular characterization of cyanophycin synthetase, the enzyme catalyzing the biosynthesis of the cyanobacterial reserve material multi-L-arginyl-poly-L-aspartate (cyanophycin). European Journal of Biochemistry 254(1):154-159. Flores E & Herrero A (2005) Nitrogen assimilation and nitrogen control in cyanobacteria. Biochem Soc Trans 33(Pt 1):164-167. Chi Z, Su CD, & Lu WD (2007) A new exopolysaccharide produced by marine Cyanothece sp. 113. Bioresource Technology 98(6):1329-1332. Madamwar D, Garg N, & Shah V (2000) Cyanobacterial hydrogen production. World Journal Of Microbiology & Biotechnology 16(8-9):757-767.
63

36. 37. 38.

39. 40.

41.

42. 43. 44. 45. 46. 47. 48. 49.

De Philippis R, Margheri MC, Materassi R, & Vincenzini M (1998) Potential of unicellular cyanobacteria from saline environments as exopolysaccharide producers. Appl Environ Microbiol 64(3):1130-1132. Reed RH, Borowitzka LJ, Mackay MA, Chudek JA, Foster R, Warr SRC, Moore DJ, & Stewart WDP (1986) Organic solute accumulation in osmotically stressed cyanobacteria. FEMS Microbiology Letters 39(1-2):51-56. Carrieri D, Momot D, Brasg IA, Ananyev G, Lenz O, Bryant DA, & Dismukes GC (2010) Boosting Autofermentation Rates and Product Yields with Sodium Stress Cycling: Application to Production of Renewable Fuels by Cyanobacteria. Appl. Environ. Microbiol. 76(19):6455-6462. McNeely K, Xu Y, Bennette N, Bryant DA, & Dismukes GC (2010) Redirecting Reductant Flux into Hydrogen Production via Metabolic Engineering of Fermentative Carbon Metabolism in a Cyanobacterium. Appl. Environ. Microbiol. 76(15):5032-5038. Xu Y, Alvey RM, Byrne PO, Graham JE, Shen G, & Bryant DA (2011) Expression of Genes in Cyanobacteria: Adaptation of Endogenous Plasmids as Platforms for High-Level Gene Expression in Synechococcus sp. PCC 7002: Photosynthesis Research Protocols. Methods in Molecular Biology, ed Carpentier R (Humana Press), Vol 684, pp 273-293. van der Oost J, van Walraven HS, Bogerd J, Smit AB, Ewart GD, & Smith GD (1989) Nucleotide sequence of the gene proposed to encode the small subunit of the soluble hydrogenase of the thermophilic unicellular cyanobacterium Synechococcus PCC 6716. Nucleic Acids Res 17(23):10098. Koebmann BJ, Westerhoff HV, Snoep JL, Nilsson D, & Jensen PR (2002) The Glycolytic Flux in Escherichia coli Is Controlled by the Demand for ATP. J. Bacteriol. 184(14):3909-3916. Searchinger T, Heimlich R, Houghton RA, Dong FX, Elobeid A, Fabiosa J, Tokgoz S, Hayes D, & Yu TH (2008) Use of US croplands for biofuels increases greenhouse gases through emissions from land-use change. Science 319(5867):1238-1240. Lynd LR, Zyl WHv, McBride JE, & Laser M (2005) Consolidated bioprocessing of cellulosic biomass: an update. Current Opinion in Biotechnology 16(5):577-583. Deng M-D & Coleman JR (1999) Ethanol Synthesis by Genetic Engineering in Cyanobacteria. Applied and Environmental Microbiology 65(2):523-528. Stal LJ & Moezelaar R (1997) Fermentation in cyanobacteria. FEMS Microbiology Reviews 21(2):179-211. Bennette NB, Eng JF, & Dismukes GC (2011) An LCMS-Based Chemical and Analytical Method for Targeted Metabolite Quantification in the Model Cyanobacterium Synechococcus sp. PCC 7002. Analytical Chemistry 83(10):3808-3816. Aubert-Jousset E, Cano M, Guedeney G, Richaud P, & Cournac L (2011) Role of HoxE subunit in Synechocystis PCC6803 hydrogenase. FEBS Journal. Gibbs M, Gfeller RP, & Chen C (1986) Fermentative Metabolism of Chlamydomonas reinhardii: III. Photoassimilation of Acetate. Plant Physiol 82(1):160-166.

64

Supplemental Information

Figure 2.S1. Photo-H2 production is limited by the total available carbohydrate pool. Cells supplemented with 5% glycerol (680 mM) exhibit an increase in both dark- and photo-H2 production. The inset shows cumulative H2 production is increased approximately 3 fold over the 20 h incubation.

65

Figure 2.S2. Skeletal outline of carbohydrate catabolism via glycolysis (left) or the oxidative pentose phosphate pathway (right) focusing solely on carbon products and reductants. Through glycolysis and the oxidation of pyruvate (see box on lower right), a maximum of 4 equivalents NADH are generated per glucose catabolized. With a hybrid of the OPP pathway and glycolysis, 3 glucose-6phosphate molecules enter OPP, ultimately yielding 2 molecules of fructose-6-phosphate and 1 molecule of glyceraldehyde-3-phosphate (GAP). The 2 molecules of fructose-6-phosphate can be isomerized to glucose-6-phosphate and be recycled while the GAP can proceed to lower glycolysis and pyruvate oxidation. The maximal theoretical yield of this hybrid scheme is 6 NADPH and 2 NADH per glucose equivalent.

66

Chapter 3: Regulation of Carbohydrate Catabolism in Cyanobacteria Revealed in Real-time

Co-authors: Gennady M. Ananyev and G. Charles Dismukes

67

Abstract Pyridine nucleotides, NAD(P)H, are universal hydride (H-) carriers that regulate metabolic pathways and transfer hydrogen and electrons intracellularly to make fuels. We have refined a fluorescence method for quantifying intracellular NAD(P)H and combined it with an electrochemical sensor for dissolved H2 to continuously monitor their concentrations during dark anaerobic carbohydrate catabolism (autofermentation) in a biotechnologically relevant cyanobacterium. Two temporal phases of NAD(P)H production are resolved, originating from prior photosynthesis and carbohydrate catabolism. These sources deliver hydride to competing terminal sinks (e.g. oxygen, nitrate, carbon dioxide and protons). The competition between these sinks under anoxia reveals precise clock-like timing of NAD(P)H and H2 production that persists for days. This method reveals, in real-time, metabolic switching between two carbohydrate catabolic pathways that differ in NAD(P)H yield by several fold, controlled by nitrate availability. Opportunities for increasing biohydrogen production are revealed. We present evidence for energy-coupled nitrate reduction (nitrate respiration) in Arthrospira.

68

Introduction Microalgae and cyanobacteria are metabolically versatile microorganisms that make their own energy and food using sunlight and nutrients via photosynthesis, and catabolize these reserves for metabolic energy during respiration (typically darkness) and fermentation (typically anoxia) (1). They are capable of storing hydrogen equivalents (universal energy carrier) as carbohydrates, lipids, molecular hydrogen (H2), and organic alcohols. Although they are recognized as the most efficient phototrophs, to realize their full potential for bioenergy applications, further improvements in metabolic efficiencies are needed and are being widely sought (2-6). The production and utilization of the main intracellular energy currencies (NADPH, NADH and ATP) is a major determinant of their biological productivity and has been a target for improvements (7-8). However, this goal has been hampered by the lack of methods to continuously monitor intracellular NAD(P)H concentrations on timescales corresponding to metabolic fluxes. Here we introduce fluorometric and electrochemical methods that continuously and simultaneously measure intracellular pyridine nucleotides and extracellular dissolved H2, respectively. We apply them to investigate what limits energy conversion via autofermentative carbohydrate catabolism in a biotechnologically important cyanobacterium. Fermentation, the hydrolytic breakdown of intracellular carbohydrates (glycogen and osmolytes) without light or terminal electron acceptors (oxygen, nitrate, etc.) is widely distributed in cyanobacteria and microalgae (9-11). As a low energy yielding metabolism used during anoxia (9), it relies upon carbohydrate disproportionation into oxidized and reduced metabolites that are excreted, and may yield H2 if hydrogenase is present. There are two catabolic pathways for the breakdown of carbohydrates under anoxia distributed among the bacteria, as summarized in Scheme 3.1: Glycolysis (Embden-MeyerhofParnas pathway, EMP) reduces 2 NAD+ and phosphorylates 2 ADP for each glucose converted to 2 equivalents pyruvate (Pyr); the Oxidative Pentose Phosphate pathway (OPP) oxidizes glucose
69

yielding NADPH in variable yield and regenerates no ATP. Because the yields of NAD(P)H and ATP via these pathways differ greatly, the pathways are tightly regulated in heterotropic bacteria. In cyanobacteria these pathways have been suggested by comparative genome sequences and experimentally confirmed by the detection (9) and quantification (10-11) of excreted fermentation products, but their regulation has not been well studied. To investigate this regulation in cyanobacteria we examined the temporal dynamics of NAD(P)H and H2 production under anoxia, using nutrients (CO2, nitrate and nitrite) to redirect carbohydrate catabolism. We used Arthrospira maxima because it contains a single class of hydrogenase and exhibits the highest reported fermentative H2 yield to date among cyanobacteria (12). It contains a [NiFe]hydrogenase which exhibits high homology to the soluble NAD+-linked hydrogenases of the proteobacteria (13), utilizes NAD+/NADH as its cognate substrates (14-17), and is expressed constitutively at equal levels in both oxic and anoxic conditions (12). Its catalytic activity is reversibly inhibited by O2 and reactivated by NAD(P)H, a general characteristic of the [NiFe]hydrogenases (12-13). The H2 yield in fermenting cyanobacteria is typically a small fraction of the theoretical maximum of 12 H2 per glucose equivalent converted to CO2 (i.e., if 100% of catabolic flux could be directed through the OPP pathway to form 12 NADPH) (18-20). By contrast, catabolism via Glycolysis generates only 2 NADH and 2 pyruvates. While only NADH is known to efficiently reduce protons to H2 via [NiFe]-hydrogenase (14-17), its enzyme-mediated equilibration with NADPH (via TH) and with reduced ferredoxin (FDx, via FNR) are potentially other reductant sources for H2 production. While both enzymes are encoded in the A. maxima genome, their activity under anoxia is unknown. FDx is the obligate electron donor to the assimilatory nitrate (narB) and nitrite (nirA) reductases(21), which are encoded by the A. maxima genome and targeted to the

The yield of reductant and ATP (per glucose equivalent catabolized) varies between EMP and OPP. The yield of EMP (glucose 2 pyruvate) is NADH/ATP = 2/2 and the yield from OPP (glucose pyruvate + 3CO2) is maximally NAD(P)H/ATP = 7/0. The pyruvate generated from each mode can generate an additional NADH through its oxidation to acetyl-CoA. See Figure 2.S2 for complete pathways. 70

cytoplasm as NARc and NIR, respectively. Scheme 3.1 predicts competition for intracellular reductant between two terminal electron acceptors: a) protons (H+) derived from carbohydrate catabolism, and b) nitrate (NO3-) imported from the medium. Prior work has shown that nitrate suppresses fermentative H2 production in cyanobacteria (12, 22).

71

Methods Arthrospira (Spirulina) maxima (CS-328) was obtained from the Tasmanian CSIRO Collection of Living Microalgae and was grown at 30 C in batch culture in 2.8 L Fernbach flasks containing 500
2+

and 200 mM

additional NaHCO3 (23-24). Cultures were illuminated by cool fluorescent lamps in 12 hr light/dark cycles. Optimal light intensity was about 302

s1. For simultaneous H2 and NAD(P)H

measurements samples were used after the 10th hour of illumination (i.e. 2 hours before the 12h dark cycle starts).

H2 sensor based on fuel cell technology. The first generation of H2 detecting microcell and electronic circuit is described elsewhere (25). Our second generation of home-built electrochemical (Clark type) microcell is based on fuel cell technology for measuring dissolved H2 concentration and uses a thin (~500 nm) Teflon membrane covered Pt/Ir electrode (12.6 mm2) biased at +220 mV for H2 oxidation. To increase the instruments sensitivity and stability a 20 m thick proton conducting membrane made of DuPont DE 1020 Nafion (Ion Power, Inc., USA) doped with black microdispersed palladium particles was inserted between the Teflon membrane and the Pt/Ir electrode (26). The second generation H2 microcell has an increased sensitivity of 2 x 10-9 C, fast response time (0.1 s), and micro-volume for samples (8 L). Resulting protons are consumed at an oversized Ag/AgCl reference electrode, allowing the rate of H2 production to be measured, rather than simply the accumulated concentration (25). The microcell responds linearly to H2 and performs for several months with excess of AgCl (~0.5 g) in the Ag/AgCl reference electrode compartment. In this article, H2 production is presented in terms of electrical current (I, in nA) generated by H2 oxidation over a period of about 20-24 h, which can be converted to gas volume using Faradays second law after digital integration of kinetic data.

72

Total electric charge (coulombs) = dQ = I*dt

[SI Eq. 1]

Optical fluorescence unit for NAD(P)H measurement. The device is based on sample illumination at an angle of 45 by pulsed (250 ms) UV-367 nm LED (Nichia, Japan). An additional interference filter (365 +/- 10 nm, Intor, USA) is used to greatly reduce artifacts from excitation light on measuring signal. NAD(P)H fluorescence is selected by an interference filter 450 +/- 30 nm (Intor, USA), and the output signal from the amplified photodiode S5591 (Hamamatsu, Japan) is filtered by a preamplifier in the DC-100 Hz range, with gain 20X (Model 113, EG&G, USA). All parts of the optical unit are constructed with black Delrin that has a very low UV parasitic fluorescence. In order to further reduce background UV fluorescence by a factor of 3-5 fold, the irradiated parts were covered by carbon containing black enamel NG SF 100 (Dupli-Color, USA). Intermittent UV-365 nm measuring light has a negligible effect on dark H2 production in Arthrospira maxima. The UV LED and fluorescence photodiode are controlled by a National Instruments data acquisition card. Each 200 ms fluorescence measurement is immediately followed by a 200 ms window where background fluorescence (without UV excitation) is measured. Measurements are taken every second, and individual data points are bunched into groups of 10 and averaged, to increase signal relative to noise.

In solution, both NADPH and NADH have indistinguishable ultraviolet absorption and blue fluorescence spectra. A large portion (91%) of this emission is quenched in vivo due to the Soret absorption bands of chlorophyll a and those of -carotene. Accordingly, exogenous NADH was used to standardize the fluorescence signal and demonstrate its linearity and reproducibility (Figure 3.S1). Quantification of the change in steady-state NAD(P)H fluorescence with a limit of detection (LOD) below 0.9 M and a response time of 1 s per individual measurement is achieved.

73

Cultures were examined before and following auto-fermentation for contaminating microbial growth using optical microscopy. All figures presented are from representative single measurements of replicated experiments unless otherwise noted.

Analytical assays for nitrate and nitrite quantification. Nitrate was measured by the nitrification of salicylic acid (27). Because nitrite interferes with the assay, it was first removed by briefly boiling the 1 ml sample in the presence of 50 l amidosulfuric acid (10% w/v). 200 l were then taken and combined with 0.5 ml sodium salicylate (5% w/v in concentrated H2SO4) and mixed immediately. After 20 minutes, 5 ml of 4 N NaOH was added to stop the reaction, and absorbance was taken at 430 nm. Nitrite was measured by the nitrite dependent azo-coupling of sulfanilamide and N-(1napthyl)-ethylenediamindihydrochloride, the Griess reaction (28). 100 l sample was added to 1.4 ml ddH2O, and to this mixture 0.5 ml sulfanilamide (58 mM in 3 N HCl) and 0.5 ml N-(1-napthyl)ethylenediamindihydrochloride (0.39 mM) were added. Absorbance at 550 nm was taken after 10 minutes. Concentrations of both nitrate and nitrite were calculated by standard curves of nitrate and nitrite prepared in Zarrouks media.

Replacement of thiamine diphosphate by oxythiamine diphosphate. A sample of Arthrospira maxima was taken from photoautotrophic growth in normal Zarrouks medium and transferred to a Petri dish to which 50 M oxythiamine was added. The culture was kept under photoautotrophic conditions (~30 E m-2 s-1, identical light intensity as used for growth) for 90 minutes to allow uptake of oxythiamine and incorporation of inactive cofactor oxythiamine diphosphate into thiamine-requiring enzymes. Cells that were still actively growing (mid-log phase, 9-10 days post culture inoculation) were utilized for this experiment to ensure that oxythiamine (and/or thiamine) uptake would be actively occurring due to the demands of cellular biosynthesis.

74

Results I. Simultaneous in vivo detection of H2 production and NAD(P)H concentration To examine the dynamics of H2 and NAD(P)H concentrations in cells we developed a twochannel instrument to simultaneously measure H2 oxidation current and intrinsic NAD(P)H fluorescence emission, building on procedures developed previously for photosynthetic organisms (see SI) (29-30) to improve sensitivity, stability, and time resolution over instruments commercially available (www.walz.com). Briefly, NAD(P)H fluorescence was excited using an LED at wavelength 365 10 nm and detected at 455 10 nm (quantum yield is 1,000-fold higher than NAD(P)+). Quantification of the change in NAD(P)H fluorescence yield was achieved with a limit of detection (LOD) below 0.9 M using a measurement time of 1 s. Correction for fluorescence absorption by cells was made via calibration against added NADH and found to be linear at the cell density used here (Figure 3.S1). Electrochemical detection of H2 (LOD = 1 nM) was done in the same cell (sample volume 8 L) using a membrane covered electrode that consumes all detectable H2 within 0.1s (H2 rate electrode) as described previously (12). Figure 3.1 gives the H2 rate and NAD(P)H fluorescence intensity over 24 h of dark anaerobic incubation. Prior to incubation, these cells were grown photoautotrophically and given an additional 3 hours of pre-illumination at a light intensity of 120 E m-2 s-1, four-fold higher than that used for growth, to deliberately accumulate a larger pool of NADPH and reduced ferredoxin (FDx). They were subsequently transferred to dark anoxia in fresh media containing nitrate (30 mM) and bicarbonate (0.2 M). Two main temporal phases of H2 production are observed in cyanobacteria incubated in dark anoxia (12). Phase-1 H2 was previously attributed to photosynthetically accumulated reductant as the yield increases with the light intensity used in the pre-illumination period and decreases with the pre-incubation dark period (glycogen catabolism during respiration on O2), prior to transfer to anoxia for measurement (12). Figure 3.1A shows that when no added bicarbonate is present phase-1 H2 production starts within minutes as soon as
75

anoxia is established, while a longer delay occurs if additional CO2 is present (t1 =60 minutes at 0.2M added bicarbonate, as shown). Phase-1 H2 production arises from two possible reductant sources that avoided consumption by the competing terminal electron acceptors CO2 and NO3-. One source is due to photosynthetically generated reductant (FDx or NADPH) and the other is derived from anaerobic carbohydrate catabolism (NAD(P)H). Figure 3.1B shows that in the same sample the NAD(P)H concentration rises immediately upon transfer to anoxia without delay even in the presence of HCO3- and NO3-, reaches a maximum extending up to 1 hour and then, in synchrony with the onset of phase-1 H2 production, decays linearly for the next 8.5 h. The rate of NAD(P)H disappearance (taken from the linear slope) is a small fraction of the flux into both H2 production and NO3- reduction (described below). This reveals that the influx to and efflux from the NAD(P)H pools are tightly balanced, resulting in a small residual pool size arising from much larger turnover rates, consistent with the understanding of yeast and other bacteria (31) (see also Figure 3.S2). This small pool size of NAD(P)H underscores the very limited redox buffering capacity and is a main reason for the strong competition for cellular reductant. A second rapid rise in H2 production (phase-2 H2) and NAD(P)H fluorescence occurs 2-20 h later following the onset of dark anoxia (Figure 3.1A). This second phase produces significantly more H2 than phase-1 and does so continuously with a sustained rate extending approximately 7 days beyond the 24 h period shown, after which cell lysis occurs (not shown). Phase-2 arises from NADH formed by catabolism of carbohydrates (glycogen and osmolytes) (12). A variable time delay between the rise in phase-2 NAD(P)H fluorescence and H2 production is observed on the order of 530 minutes reflecting intracellular events (Figure 3.1A, insert). This delay may be due to several factors: (a) competition for NADH consumed by other metabolic pathways (such as reduction of pyruvate to lactate and/or ethanol), (b) thermodynamic requirement for the NADH/NAD+ ratio to surpass a threshold sufficient to overcome the 100 mV difference in standard reduction potentials

76

of NADH/NAD+ (-320 mV) and H2/2H+ (-420 mV), and/or (c) possible in vivo allosteric regulation of hydrogenase by pyridine nucleotides or ATP (15).

II. Triggering of H2 production upon nitrate depletion Figure 3.2A/B/C reveals the temporal dynamics of phase-2 NAD(P)H concentration and H2 evolution rate under the same conditions as in Figure 3.1, both in the absence and presence of variable concentrations of nitrate. Without pre-illumination to build up intracellular reductant, nitrate (15 mM) delays the onset of H2 production and causes a significant drop in NAD(P)H for approximately 5 hours. An inflection point occurs at hour 5 where the residual [NAD(P)H] sharply increases and H2 production turns on within minutes. The rate saturates at a near constant value over the next 13 h. Figure 3.2C shows that varying the extracellular nitrate concentration systematically increases the delay time for the onset of phase-2 H2 from 0 to 20 hours (t2); this relationship is a linear function depending upon the initial nitrate concentration. These rapid transitions and tight coupling of NAD(P)H/NAD(P) ratio to H2 production suggest a triggering event with clock-like precision as nitrate is consumed and proton reduction takes over in the cell. The sharp nature of this transition was unexpected and reveals that the nitrate concentration regulates the rate of reductant generation (as described below). As expected, the extracellular NO3- concentration decreases with time during anoxic metabolism (Figure 3.2D); nitrate is taken up in two temporal phases, initial rapid influx into the cells upon addition, followed by slower uptake in parallel with excretion of reduced nitrite. Nitrate reduction is thermodynamically favored over H+ reduction, and its eventual depletion from the culture medium allows the reductant generated from carbohydrate catabolism to be more effectively channeled to H2 production. The measured nitrate uptake rate (between 3.9 and 5.1 mol min-1 gDW-1) agrees with those from other bacteria. It is comparable to the anoxic uptake rate in E. coli of 52 mol min-1 gDW-1 (reported rate adjusted to 10 gDW L-1, 50% protein, in A. maxima)
77

when the culture is growing on nitrate (32), and significantly faster than the nitrate uptake rate in aerobic A. platensis of 0.074 mol min-1 gDW-1 (33). Figure 3.2D also shows that the terminal products of nitrate reduction include significant quantities of excreted nitrite, although the majority is retained in the cell. Together, these data suggest dissimilatory nitrate reduction in A. maxima (11) (anoxic respiration on nitrate coupled to nitrite excretion has not been reported in cyanobacteria previously).

III. Nitrate respiration in A. maxima & metabolic switching between EMP and OPP pathways The FDx-dependent nitrate reductase found in cyanobacteria (narB) is often found associated with the thylakoid membranes (21), suggesting the possibility of respiratory electron transport coupled to NO3- reduction. In support of this proposed function, six open reading frames in the A. maxima draft genome exhibit significant homology (57% sequence similarity) to regions of the three subunits of the quinol-nitrate oxidoreductase (narGHI), the membrane-bound dissimilatory nitrate reductase expressed during anoxic metabolism in proteobacteria (e.g. Anaeromyxobacter spp.) (34). Specifically, ORF#1825 exhibits homology to the transmembrane domain containing region of narI, the subunit, whose canonical function is to anchor the complex to the membrane and couple quinol oxidation to nitrate reduction (35). This could serve as a membrane association mechanism in A. maxima for narB (known to exist) or narG (present, but possibly inactivated by frameshift mutations), both of which contain a molybdenum active site (Table 3.S1). This gene product would allow for PQH2 dependent reduction of NO3 to couple to NAD(P)H dependent reduction of plastoquinone (PQ) by NAD(P)H dehydrogenase (NDH-1). Such a model would support the proton motive force both by H+ translocation via NDH-1 and the PQ/PQH2 cycle, and allow ATP generation by energy-coupled nitrate reduction, so called nitrate respiration (denoted ~PMF in Scheme 3.1).

78

In support of this hypothesis, when pentachlorophenol (PCP), a specific inhibitor of the membrane bound quinol-nitrate oxidoreductase (36), is added to A. maxima cultures along with nitrate, the phase-1 H2 yield is increased 3.3-fold and the phase-2 delay time is unchanged (Figure 3.3A), while the residual NAD(P)H concentration decreases 2-fold (Figure 3.3B). The decreased capacity for ATP generation by nitrate respiration in the presence of PCP is predicted to cause the organism to divert a larger portion of carbohydrate through glycolysis instead of the OPP pathway, as glycolytic ATP generation (~P in Scheme 3.1) would be the only means of phosphorylating ADP in the absence of the nitrate respiratory pathway (and all other potential respiratory pathways). This redirection of carbohydrate catabolism from OPP to EMP pathway will increase the NADH production (which hydrogenase can access) at the expense of a larger loss in NADPH production, thus explaining both the 3.3-fold increase in phase-1 H2 production (Figure 3.3A) and the 2-fold lower NAD(P)H concentration (Figure 3.3B). As the OPP pathway produces much more reductant (as NADPH) than the EMP-glycolysis (as NADH), it should be the preferred pathway in the presence of excess nitrate, as it could support much greater ATP generation via nitrate respiration (~PMF in Scheme 3.1). To confirm this hypothesis, cells were incubated for 1.5 hours with oxythiamine (Oxy-TPP), using a protocol developed to partially replace the native cofactor thiamine diphosphate with oxythiamine diphosphate (see SI), and thereby partially inactivate the thiamine-dependent transketolase of the OPP pathway (37-38). Following this treatment, carbohydrate catabolism should be redirected from OPP to EMP. This experiment produces a lower NAD(P)H concentration during the initial 5 hours (Figure 3.3D) and a 30% longer delay time necessary to reduce the available nitrate to nitrite (phase-2 H2 delay time, Figure 3.3C). This longer delay time is consistent with a slower rate of reductant generation per glucose equivalent catabolized, as predicted. This outcome might also occur if the NARm prefers NADPH rather than NADH, as NARm is coupled to NDH-1, which may exhibit a preference for NADPH over NADH in some cyanobacteria (39). Figure 3.3C also reveals
79

that Oxy-TPP causes an increased NAD(P)H concentration and corresponding H2 evolution rate during hours 8 through 20. As illustrated in Scheme 3.1, this is likely due to stimulation of the lower branch of EMP-glycolysis by accumulation of product nitrite. The resulting production of pyruvate is needed for PFOR-mediated FDx reduction (FDx is the obligate substrate for NIR). We expect a slower rate of reduction of nitrite in the presence of Oxy-TPP as a consequence of the slower conversion of nitrate to nitrite, hence the build-up in NAD(P)H.

80

Discussion To our knowledge this is the first demonstration of nitrate respiration in a cyanobacterium a finding that was enabled by our new tool for simultaneous quantification of H2 and NAD(P)H. Our results show that supplementation of A. maxima cultures with nitrate (30 mM) reduces H2 production by more than 200-fold, illustrating an extremely efficient redirection of reductant within the cell. The model in Scheme 3.1 summarizes the reductants produced by anoxic carbohydrate catabolism, emphasizing the nitrate induced changes in NAD(P)H and H2 production. Together, these data reveal the activation of an upstream metabolic switch upon nitrate addition that redirects carbohydrate catabolism from the EMP to the OPP pathway, resulting in an accelerated rate of reductant generation and production of significantly more NADPH per glucose equivalent catabolized. Cells can thereby produce significantly more ATP by nitrate respiration and this ATP can be used to accelerate the energy-coupled import of nitrate and other essential nutrients (for later use when nutrients are scarce). Following consumption of nitrate below a threshold (or with PCP inhibition and/or Oxy-TPP substitution), carbon catabolism switches back to the EMP pathway and A. maxima must rely upon much less efficient ATP generation by substrate level phosphorylation. These results show that the competition between terminal electron acceptors (proton and nitrate) within cyanobacteria controls the relative fluxes through EMP/OPP pathways, presumably via regulation at the entry into OPP and EMP pathways as common in other bacteria (40). The term metabolic switching has also been used to denote switching between oxygenic photosynthesis and nitrogen fixation that coordinate these incompatible metabolisms in various cyanobacteria Trichodesmium (41), Cyanothece (42), and Synechococcus (43). Similarly, switchable genetic circuits have been engineered into model organisms by synthetic biologists (44). This model of OPP/EMP metabolic switching can account, not only for stimulation of anoxic H2 production, but also the observed significant increase in glycogen accumulation in many
81

cyanobacteria (up to 50% of cell dry weight in A. platensis) upon nitrate depletion from the aerobic culture medium (45-46). Because removal of nitrate lowers the ATP demand and elevates the ATP pool size, it slows carbohydrate catabolism and activates gluconeogenesis, enabling significant build up of glycogen reserves. In closing, we call attention to a biotechnological implication. Reprogramming this metabolic switch from EMP to OPP in the absence of nitrate could provide the organism with significantly more reductant from carbohydrate catabolism and increase the maximal theoretical energy conversion efficiency of carbohydrate to H2 from 33% (4 NADH/glucose) to 67% (8 NAD(P)H/glucose) for acetogenic fermentation, provided that the NADPH formed from OPP could be completely converted to H2. Realizing such an increase would be a major step in improving the feasibility of renewable solar to hydrogen production catalyzed by cyanobacteria.

82

Figures

Scheme 3.1. The redox and phosphoryl energy conversion steps during catabolism that terminate in (top) proton reduction (H2 evolution) and (bottom) nitrate/nitrite reduction. Glucose-6-phosphate (G6P), derived from carbohydrate breakdown, is the switching point between entry into either the upper glycolytic pathway or the oxidative pentose phosphate (OPP) pathway. These pathways converge at GAP and share lower glycolysis, but differ in the yield and type of reductants formed (NADH vs NADPH). The oxidation of pyruvate via pyruvate-ferredoxin oxidoreductase (PFOR) generates reduced ferredoxin (FDx). NARc and NARm are the cytoplasmic and (proposed) membrane-associated nitrate reductases, respectively. Nitrite reductase (NIR) catalyzes the reduction of nitrite to ammonium. NARc and NIR are ferredoxin-dependent, while NARm is an
83

energy-linked quinol:nitrate oxidoreductase that uses NAD(P)H and pumps H+ across the cytoplasmic membrane. ATP coupling sites (via substrate level phosphorylation and via nitrate respiration) are labeled ~P and ~PMF, respectively.

84

Figure 3.1. Dissolved H2 rate (A) and residual NAD(P)H concentration (B) during anaerobic incubation for A. maxima cells that were pre-illuminated for 3 hours under 4x higher light intensity than used for growth. The concentration of residual intracellular NAD(P)H rises and precedes two phases of extracellular H2 release. H2 production has two reductant sources: leftover photosynthetic reductant (NADPH & FDx) generated by prior strong pre-illumination, and reductant generated from anaerobic carbohydrate catabolism (NADH, NADPH, FDx). Competition for reductant utilization affects the delay times of these phases and yield of H2. The delay time of phase-1 H2 production (t1 ~1 hour, 0.2M bicarbonate) increases with the concentration of bicarbonate in media consistent with completion for CO2 reduction. The delay time of phase-2 H2 production (8.8 hrs), denoted t2, depends on competition for reductant between hydrogenase and nitrate reductase.

85

Figure 3.2. Effect of extracellular nitrate concentration (15mM) on the rate of dissolved H2 production (A) and residual intracellular NAD(P)H concentration (B) during dark anaerobiosis of A. maxima cells without pre-illumination. The phase-2 delay time, t2, increases linearly with the concentration of extracellular nitrate over a wide dynamic range (C). The slope of the linear regression (k = 5.1 mol min-1 gDW-1) represents the rate of nitrate consumption (by reduction) within the cell, a process in competition with proton reduction (H2 evolution). During this time and following, nitrate is continuously imported from the medium and some nitrite is excreted (20% of total nitrate after 45 hours) (D). The 10-fold lower cell density in D causes increased time necessary for nitrate consumption but a similar rate (k = 3.9 mol min-1 gDW-1). In the absence of pre-illumination and presence of nitrate, the residual NAD(P)H concentration (B) initially falls and then starts to recover at a constant rate at hour 2. An inflection point at hour 5 signals the depletion of NO3- inside the cell, diverting available reductant to NADH and proton reduction (H2 evolution).

86

Figure 3.3. H2 production (A) and residual intracellular NAD(P)H concentration (B) of A. maxima during dark anaerobic incubation in the absence of nitrate (black), with 30 mM NO3- (red), and with 30 mM NO3- plus 20 M pentachlorophenol (PCP) toinhibit the energy-coupled quinol nitrate oxidoreductase (blue). H2 production (C) and residual intracellular NAD(P)H concentration (D) in the presence of 30 mM NO3- with (red) and without (black) oxythiamine diphosphate (Oxy-TPP) treatment. In the presence of Oxy-TPP, transketolase of OPP is inactivated and catabolism is diverted from OPP to EMP, producing less reductant per glucose during the initial 6 hours of incubation (D). We note that younger, exponentially growing, cells were used for these experiments because of their more robust utilization of nitrate. Accordingly the phase-2 H2 delay is slightly shorter than observed if stationary phase cells are used.

87

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. Benemann J, Weissman J, Koopman B, & Oswald W (1977) Energy production by microbial photosynthesis. Nature 268(5615):19-23. Kruse O, Rupprecht J, Bader K-P, Thomas-Hall S, Schenk PM, Finazzi G, & Hankamer B (2005) Improved Photobiological H2 Production in Engineered Green Algal Cells. Journal of Biological Chemistry 280(40):34170-34177. Hallenbeck PC & Benemann JR (2002) Biological hydrogen production; fundamentals and limiting processes. International Journal Of Hydrogen Energy 27(11-12):1185-1193. McKinlay JB & Harwood CS (2010) Photobiological production of hydrogen gas as a biofuel. Current Opinion in Biotechnology 21(3):244-251. Ghirardi ML, Posewitz MC, Maness PC, Dubini A, Yu JP, & Seibert M (2007) Hydrogenases and hydrogen photoproduction in oxygenic photosynthetic organisms. Annual Review of Plant Biology 58:71-91. Quintana N, Van der Kooy F, Van de Rhee M, Voshol G, & Verpoorte R (2011) Renewable energy from Cyanobacteria: energy production optimization by metabolic pathway engineering. Applied Microbiology and Biotechnology 91(3):471-490. Wall J, Harwood CS, & Demain AL (2008) Bioenergy (ASM Press, Washington). Reaves ML & Rabinowitz JD (2011) Metabolomics in systems microbiology. Current Opinion in Biotechnology 22(1):17-25. Stal LJ & Moezelaar R (1997) Fermentation in cyanobacteria. FEMS Microbiology Reviews 21(2):179-211. Carrieri D, Momot D, Brasg IA, Ananyev G, Lenz O, Bryant DA, & Dismukes GC (2010) Boosting Autofermentation Rates and Product Yields with Sodium Stress Cycling: Application to Production of Renewable Fuels by Cyanobacteria. Appl. Environ. Microbiol. 76(19):6455-6462. McNeely K, Xu Y, Ananyev G, Bennette N, Bryant DA, & Dismukes GC (2011) Characterization of a nifJ Mutant of Synechococcus sp. strain PCC 7002 Lacking Pyruvate:Ferredoxin Oxidoreductase. Appl. Environ. Microbiol.:AEM.02792-02710. Ananyev G, Carrieri D, & Dismukes GC (2008) Optimization of Metabolic Capacity and Flux through Environmental Cues To Maximize Hydrogen Production by the Cyanobacterium "Arthrospira (Spirulina) maxima". (Translated from Eng) Applied and Environmental Microbiology 74(19):61026113 (in Eng). Germer F, Zebger I, Saggu M, Lendzian F, Schulz R, & Appel J (2009) Overexpression, Isolation, and Spectroscopic Characterization of the Bidirectional [NiFe] Hydrogenase from Synechocystis sp. PCC 6803. Journal of Biological Chemistry 284(52):36462-36472. Maroti J, Farkas A, Nagy IK, Maroti G, Kondorosi E, Rakhely G, & Kovacs KL (2010) A Second Soluble Hox-Type NiFe Enzyme Completes the Hydrogenase Set in Thiocapsa roseopersicina BBS. Appl. Environ. Microbiol. 76(15):5113-5123. Burgdorf T, van der Linden E, Bernhard M, Yin QY, Back JW, Hartog AF, Muijsers AO, de Koster CG, Albracht SPJ, & Friedrich B (2005) The soluble NAD(+)-reducing [NiFe]Hydrogenase from Ralstonia eutropha H16 consists of six subunits and can be specifically activated by NADPH. Journal of Bacteriology 187(9):3122-3132. Tamagnini P, Axelsson R, Lindberg P, Oxelfelt F, Wunschiers R, & Lindblad P (2002) Hydrogenases and hydrogen metabolism of cyanobacteria. Microbiology And Molecular Biology Reviews 66(1):1-20. Aubert-Jousset E, Cano M, Guedeney G, Richaud P, & Cournac L (2011) Role of HoxE subunit in Synechocystis PCC6803 hydrogenase. FEBS Journal. Woodward J, Orr M, Cordray K, & Greenbaum E (2000) Biotechnology: Enzymatic production of biohydrogen. Nature 405(6790):1014-1015.
88

11. 12.

13. 14. 15.

16. 17. 18.

19. 20. 21. 22. 23. 24. 25.

26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37.

McNeely K, Xu Y, Bennette N, Bryant DA, & Dismukes GC (2010) Redirecting Reductant Flux into Hydrogen Production via Metabolic Engineering of Fermentative Carbon Metabolism in a Cyanobacterium. Appl. Environ. Microbiol. 76(15):5032-5038. Zhang YHP, Evans BR, Mielenz JR, Hopkins RC, & Adams MWW (2007) High-Yield Hydrogen Production from Starch and Water by a Synthetic Enzymatic Pathway. PLoS ONE 2(5):e456. Flores E, Frias JE, Rubio LM, & Herrero A (2005) Photosynthetic nitrate assimilation in cyanobacteria. Photosynth Res 83(2):117-133. Gutthann F, Egert M, Marques A, & Appel J (2007) Inhibition of respiration and nitrate assimilation enhances photohydrogen evolution under low oxygen concentrations in Synechocystis sp. PCC 6803. Biochim Biophys Acta 1767(2):161-169. Carrieri D, Ananyev G, Brown T, & Dismukes GC (2007) In vivo bicarbonate requirement for water oxidation by Photosystem II in the hypercarbonate-requiring cyanobacterium Arthrospira maxima. J Inorg Biochem 101(11-12):1865-1874. Carrieri D, Ananyev G, Costas AMG, Bryant DA, & Dismukes GC (2008) Renewable hydrogen production by cyanobacteria: Nickel requirements for optimal hydrogenase activity. International Journal of Hydrogen Energy 33(8):2014-2022. Ananyev G, Carrieri D, & Dismukes GC (2008) Optimization of metabolic capacity and flux through environmental cues to maximize hydrogen production by the cyanobacterium "Arthrospira (Spirulina) maxima". Applied and Environmental Microbiology 74(19):61026113. Harrison DK & Kessler M (1989) A multiwire hydrogen electrode for in vivo use. Phys Med Biol 34(10):1397-1412. Cataldo DA, Maroon M, Schrader LE, & Youngs VL (1975) Rapid colorimetric determination of nitrate in plant tissue by nitration of salicylic acid. Commun. Soil Sci. Plant 6:71-80. Granger DL, Taintor RR, Boockvar KS, & John B H, Jr. (1995) Determination of Nitrate and Nitrite in Biological Samples Using Bacterial Nitrate Reductase Coupled with the Griess Reaction. Methods 7(1):78-83. Duysens LNM & Sweep G (1957) Fluorescence Spectrophotometry of Pyridine Nucleotide in Photosynthesizing Cells. Biochimica Et Biophysica Acta 25(1):13-16. Mi HL, Klughammer C, & Schreiber U (2000) Light-induced dynamic changes of NADPH fluorescence in Synechocystis PCC 6803 and its ndhB-defective mutant M55. Plant and Cell Physiology 41(10):1129-1135. Bakker BM, Overkamp KM, van Maris AJA, Ktter P, Luttik MAH, van Dijken JP, & Pronk JT (2001) Stoichiometry and compartmentation of NADH metabolism in Saccharomyces cerevisiae. FEMS Microbiology Reviews 25(1):15-37. Denis KS, Dias FM, & Rowe JJ (1990) Oxygen regulation of nitrate transport by diversion of electron flow in Escherichia coli. Journal of Biological Chemistry 265(30):18095-18097. Lodi A, Binaghi L, Solisio C, Converti A, & Borghi M (2003) Nitrate and phosphate removal by Spirulina platensis. Journal of Industrial Microbiology & Biotechnology 30(11):656-660. Altschul SF, Madden TL, Schffer AA, Zhang J, Zhang Z, Miller W, & Lipman DJ (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Research 25(17):3389-3402. Bertero MG, Rothery RA, Palak M, Hou C, Lim D, Blasco F, Weiner JH, & Strynadka NCJ (2003) Insights into the respiratory electron transfer pathway from the structure of nitrate reductase A. Nature Structural Biology 10(9):681-687. Bertero MG, Rothery RA, Boroumand N, Palak M, Blasco F, Ginet N, Weiner JH, & Strynadka NCJ (2005) Structural and Biochemical Characterization of a Quinol Binding Site of Escherichia coli Nitrate Reductase A. Journal of Biological Chemistry 280(15):14836-14843. Meshalkina LE & Kochetov GA (1979) The functional identity of the active centres of transketolase. Biochimica et Biophysica Acta (BBA) - Enzymology 571(2):218-223.
89

38.

39. 40. 41. 42.

43.

44. 45. 46.

Shreve DS, Holloway MP, Haggerty JC, & Sable HZ (1983) The catalytic mechanism of transketolase. Thiamin pyrophosphate-derived transition states for transketolase and pyruvate dehydrogenase are not identical. Journal of Biological Chemistry 258(20):1240512408. Battchikova N, Eisenhut M, & Aro E-M (2011) Cyanobacterial NDH-1 complexes: Novel insights and remaining puzzles. Biochimica et Biophysica Acta (BBA) - Bioenergetics 1807(8):935-944. Lehninger AL (1982) Principles of biochemistry (Worth Publishers, New York, N.Y.). Berman-Frank I, Lundgren P, Chen Y-B, Kpper H, Kolber Z, Bergman B, & Falkowski P (2001) Segregation of Nitrogen Fixation and Oxygenic Photosynthesis in the Marine Cyanobacterium Trichodesmium. Science 294(5546):1534-1537. Stckel J, Jacobs JM, Elvitigala TR, Liberton M, Welsh EA, Polpitiya AD, Gritsenko MA, Nicora CD, Koppenaal DW, Smith RD, & Pakrasi HB (2011) Diurnal Rhythms Result in Significant Changes in the Cellular Protein Complement in the Cyanobacterium Cyanothece 51142. PLoS ONE 6(2):e16680. Steunou AS, Bhaya D, Bateson MM, Melendrez MC, Ward DM, Brecht E, Peters JW, Kuhl M, & Grossman AR (2006) In situ analysis of nitrogen fixation and metabolic switching in unicellular thermophilic cyanobacteria inhabiting hot spring microbial mats. Proceedings of the National Academy of Sciences of the United States of America 103(7):2398-2403. Purnick PEM & Weiss R (2009) The second wave of synthetic biology: from modules to systems. Nat Rev Mol Cell Biol 10(6):410-422. Aoyama K, Uemura I, Miyake J, & Asada Y (1997) Fermentative metabolism to produce hydrogen gas and organic compounds in a cyanobacterium, Spirulina platensis. Journal of Fermentation and Bioengineering 83(1):17-20. Vonshak A, Guy R, & Guy M (1988) The Response Of The Filamentous Cyanobacterium Spirulina-Platensis To Salt Stress. Archives Of Microbiology 150(5):417-420.

90

Supplemental Information Choice of Arthrospira maxima A. maxima possesses the most efficient (fastest) electron transport chain in vivo (H2O PSII PSI NADPH) in comparison to all eukaryotic algae and plants (1-2), and a robust photosynthetic energy metabolism that allows it to survive in solutions at pH 9.0 - 11.5 and dissolved Na2CO3 concentrations of 0.4 1.0 M. Also, in natural conditions A. maxima lives as a monoculture and in the laboratory it has shown high resistance to contamination. A. maxima displays one of the highest biomass productivities of cyanobacteria: 8.4 gDW L-1 culture (after 12 days), with up to 45% of biomass as glycogen, 10-12% osmolytes, and 19.5 % as chloroform-methanol extractable lipids (our data). Optimizing growth conditions of A. maxima has increased the fermentative H2 yield to 36 mL H2 gDW-1 with a maximum accumulation of 18% hydrogen in the vials headspace (by volume) (2). A significantly higher yield of 230 mL H2 gDW-1 is reached when hydrogen is milked from the culture by a fuel-type H2 microcell which rapidly extracts and oxidizes the evolved gas (as presented here).

Relationship of residual and actual NAD(P)H concentrations The NAD(P)H fluorescence records the intracellular residual concentration that is not instantaneously consumed by metabolic reactions faster than the fluorometer response time. It represents the difference between the NAD(P)H formation and utilization reactions, representing the steady-state redox poise of the cell, with 0 corresponds to the (normalized) initial level of residual NAD(P)H at the start of the measurement. For example, the increase in H2 production rate of 100 mol gDW-1 h-1 (as in phase 2 of Figure 3.1) correlates with an increase in residual NAD(P)H of only 0.4 mol gDW-1 h-1 at the onset of Phase-2. Because the actual NAD(P)H production rate must be large enough to support the observed H2 production rate (1:1 molar equivalence of NADH:H2) as well as other unmeasured fermentative pathways, the actual flux into the NAD(P)H
91

pool is at least 100.4 mol gDW-1 h-1 (i.e. 250-fold larger than the residual change we monitor, see Figure 3.S2). For comparison, formation rates of NADH in E. coli are reported in the range of 12 14 mmol NADH gDW-1 h-1 (3-4) and the occurrence of a much smaller residual concentration is consistent with the understanding that the total pool size (NADH + NAD+) is present only in catalytic amounts and that formation and utilization rates of NADH are continually matched to minimize residual concentration (5).

92

Figure 3.S1. Linear response of fluorescence signal to exogenous NADH. NADH of various concentrations was assayed for its fluorescence signal (450-520 nm) in the presence and absence of concentrated A. maxima filaments (10 gDW L-1). The filaments quenched 91% of the native NADH fluorescence. Remaining fluorescence (in the presence of A. maxima filaments) is plotted and used to determine absolute concentration change of residual intracellular NAD(P)H from observed fluorescence signal.

93

Figure 3.S2. Scatterplot of phase-2 H2 oxidation current magnitude versus change in residual NAD(P)H concentration ([NAD(P)H]) from ten different experiments. The residual NAD(P)H is the fluorescence-detected reduced pyridine nucleotide pool. The regression line indicates that changes in residual [NAD(P)H] less than 0 (i.e. unable to be seen as increase in fluorescence) correlate with positive H2 oxidation current. Therefore a portion of the intracellular reductant that generates H2 is not detected as NAD(P)H fluorescence, either because it is oxidized by hydrogenase faster than the fluorometer response time (1 s), or because there is another reductant source other than NADH. H2 conversion rate: 100 nA = 46.7 mol H2 gDW-1 h-1.

94

Table 3.S1. Translated nucleotides from A. maxima ORFs were searched against all translated nucleotide sequences in Genbank using the BLAST algorithm (tblastx) (6). Six ORFs exhibited unique homology to subunits of the dissimilatory nitrate reductase (narGHI); their degree of homology and coverage is listed. Of note, the single ORF exhibiting the largest degree of uninterrupted coverage (41.1%) is or_1825, encoding a portion of the subunit (narI) which serves as the canonical membrane anchor and interactor with quinone. The region found in or_1825 codes for the transmembrane domain containing region of narI. Subunits and (narG and narH) appear to have been inactivated by frameshift mutations. Homology is defined as conservation of residue or mutation to residue with similar functionality (conserved mutation).

95

Supplementary References 1. 2. Ananyev G & Dismukes GC (2005) How fast can Photosystem II split water? Kinetic performance at high and low frequencies. Photosynthesis Research 84(1-3):355-365. Ananyev G, Carrieri D, & Dismukes GC (2008) Optimization of metabolic capacity and flux through environmental cues to maximize hydrogen production by the cyanobacterium "Arthrospira (Spirulina) maxima". Applied and Environmental Microbiology 74(19):61026113. Berros-Rivera SJ, Bennett GN, & San K-Y (2002) The Effect of Increasing NADH Availability on the Redistribution of Metabolic Fluxes in Escherichia coli Chemostat Cultures. Metabolic Engineering 4(3):230-237. Sauer U, Canonaco F, Heri S, Perrenoud A, & Fischer E (2004) The Soluble and Membranebound Transhydrogenases UdhA and PntAB Have Divergent Functions in NADPH Metabolism of Escherichia coli. Journal of Biological Chemistry 279(8):6613-6619. Bakker BM, Overkamp KM, van Maris AJA, Ktter P, Luttik MAH, van Dijken JP, & Pronk JT (2001) Stoichiometry and compartmentation of NADH metabolism in Saccharomyces cerevisiae. FEMS Microbiology Reviews 25(1):15-37. Altschul SF, Madden TL, Schffer AA, Zhang J, Zhang Z, Miller W, & Lipman DJ (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Research 25(17):3389-3402.

3. 4. 5. 6.

96

Chapter 4: Enhancing Biological Hydrogen Production from Cyanobacteria by Removal of Excreted Products

Co-authors: Gennady M. Ananyev and G. Charles Dismukes

97

Abstract Hydrogen is produced by a [NiFe]-hydrogenase in the cyanobacterium Arthrospira (Spirulina) maxima during autofermentation of photosynthetically accumulated glycogen under dark anaerobic conditions. Herein we show that elimination of H2 backpressure by continuous H2 removal (milking) can significantly increase the yield of H2 in this strain. We show that milking by continuous selective consumption of H2 using an electrochemical cell produces the maximum increase in H2 yield (11-fold) and H2 rate (3.4-fold), which is considerably larger than through milking by non-selective dilution of the biomass in media (increases H2 yield 3.7-fold and rate 3.1fold). Exhaustive autofermentation under electrochemical milking conditions consumes >98% of glycogen and 27.6% of biomass over 7-8 days and extracts 39% of the energy content in glycogen as H2. Non-selective dilution stimulates H2 production by shifting intracellular equilibria competing for NADH from excreted products and terminal electron sinks into H2 production. Adding a mixture of the carbon fermentative products shifts the equilibria towards reactants, resulting in increased intracellular NADH and an increased H2 yield (1.4-fold). H2 production is sustained for a period of time up to 7 days, after which the PSII activity of the cells decreases by 80-90%, but can be restored by regeneration under photoautotrophic growth.

98

Introduction The solar driven splitting of water offers a potential source of renewable H2, but the efficiency and scalability of current solar based methods are inadequate as alternatives to fossil derived H2. This has spurred research into microorganisms that convert water into H2 and O2 using solar energy (1-3). The use of solar energy for the biological production of H2 can be divided into two stages: an aerobic photosynthetic stage leading to accumulation of energy storage macromolecules such as starch (in green algae) or glycogen (in cyanobacteria), followed by anaerobic catabolism of these storage molecules by metabolic reactions that provide intracellular reductant for H2 production (4). One biological route to hydrogen production is by fermentation of the cells own carbohydrate reserves, so-called autofermentation. This form of metabolism occurs under conditions where light and oxygen are not available and accordingly the cell is incapable of oxidative- or photo-phosphorylation for ATP generation. Microalgae have also attracted significant attention for their potential use as biodiesel and biomass feedstocks (5-6). Energy demands on the cell such as macromolecule repair, maintenance of ion gradients, and osmolyte synthesis require ATP which is produced by substrate level phosphorylation in the glycolytic reactions, along with NADH (7). This NADH must be recycled to NAD+ for the reactions of glycolysis to continue, and one such route to NADH oxidation is by the bidirectional [NiFe]hydrogenase (8-9). This enzyme catalyzes the reversible reduction of protons to H2 coupled to the regeneration of NAD+. It is not present in all cyanobacteria genomes (10). The cyanobacterium studied here, Arthrospira maxima sp. CS-328 (A. maxima), exhibits the highest known rate and yield of autofermentative H2 production reported to date among the cyanobacteria (up to 36 ml of H2/g dry weight and a maximum 18% H2 in the headspace) (11). The maximal yield of reductant generated during acetogenic fermentation through glycolysis alone (absence of any respiratory electron acceptors) is four equivalents of NADH and four equivalents of ATP per glucose catabolized to 2 acetates:
99

C6H12O6 + 4NAD+ + 4ADP + 4Pi + 2H2O 4NADH + 4H+ + 4ATP + 2CH3COOH + 2CO2 This maximum yield of 4 NADH per glucose, known as the Thauer limit (12), is one third of the theoretical yield of 12 NAD(P)H if each glucose were to be completely oxidized by the enzymes of the oxidative pentose phosphate (OPP) pathway, which has never been achieved in vivo. However, in vitro reconstitution of the 13 enzymes of the OPP pathway has been demonstrated to produce this maximum yield (13-14). Recent work by our group has demonstrated that nitrate can induce a switch to the OPP pathway during autofermentation (Ananyev, et al., in review). H2 production by autofermentation is termed indirect biophotolysis, as reductant powering the evolution of H2 originates from previously accumulated carbohydrate reserves (glycogen, osmolytes and soluble sugars) rather than directly from water. This process differs from direct photolysis (usually catalyzed by an [FeFe]-hydrogenase found in some eukaryotes, e.g. Chlamydomonas reinhardtii) where reductant originating from water is directly transferred from PSII through PSI to the hydrogenase (15-16). Indirect biophotolysis has an advantage over the direct route when the O2 sensitivity of hydrogenase is considered. Because indirect biophotolysis occurs in the dark, O2 production by PSII is circumvented. To date, algal systems performing direct biophotolysis have had to utilize harsh methods to inactivate PSII (e.g. by sulfur deprivation (17) or copper cations (18)) or selectively excite PSI through the use of infrared illumination (19). The bidirectional hydrogenase (encoded by the hox gene cluster) is the only hydrogen metabolizing enzyme in A. maxima. It is constitutively expressed at constant level, but reversibly inactivated by O2; its activity is restored immediately upon the onset of anaerobiosis. Biotechnologically, hydrogen production by A. maxima is very promising in that: i) routine headspace H2 accumulation of 10 to 12% (occasionally up to 18%) is achieved (11), ii) at the end of exhaustive autofermentation (7-8 days) the culture is still viable and able to be returned to a photobioreactor for another cycle of carbohydrate accumulation (we have demonstrated that at least 3 cycles are possible), and iii) the H2 yield is dependent upon the initial carbohydrate content
100

of the cells, which can be up to 48% of cell dry weight after nitrate depletion, phosphate depletion, or growth in high salt (above 1M NaCl) (20-22). For comparison to non-phototrophic systems, the maximal observed H2 headspace accumulation is 74% from Clostridium spp. (which contains the more efficient [FeFe]-hydrogenase) fermenting on exogenous glucose (23). In Clostridia species, the presence of the gaseous products in the fermentative medium and headspace negatively affect biomass accumulation and H2 production (24). Removing CO2 from the headspace (decreasing its partial pressure from 24% to 5%) was shown to increase H2 accumulation by 87% (25). Various studies of Clostridia have established other key metabolites whose concentration in the culture medium directly affect H2 production: acetic and butyric acids (26), ammonia (27), glucose, and initial biomass (23). We note that the higher H2 headspace accumulation by Clostridia is most likely due to the functioning of the [FeFe]hydrogenase, which operates at a stronger driving force with FDx as its electron donor. Building upon the knowledge gained from studying the Clostridia systems, we have identified secondary metabolites and physiological conditions to increase the autofermentative H2 yield by A. maxima. While other studies have focused on creating transgenic strains where pathways competing with hydrogen production are eliminated (e.g. lactate production (28) and nitrate reduction (29)), herein, we focus on manipulating the equilibrium and thermodynamics of the hydrogenase reaction itself. Because the redox couples of NADH/NAD+ and H2/H+ are nearly identical under biological conditions, the reaction equilibrium is especially sensitive to the concentrations of reactants and products. We define milking as the act of accelerating the reaction rate by removing the product H2 gas and thereby shifting the equilibrium of the hydrogenase reaction. We use two methods for milking: i) selective removal of H2 gas from the reaction by consuming it with a fuel-type electrochemical cell, and ii) non-specifically diluting the biomass concentration used for fermentation (along with the spent medium) and as a result, diluting the excreted products, salinity of medium, and extracellular nitrate concentration. Furthermore, we illustrate that by incubating
101

cultures with a mixture of fermentative products (lactate, formate, ethanol) to shift their reaction equilibria to substrates, we are able to redirect NADH to hydrogenase and increase the H2 yield.

102

Methods Arthrospira maxima sp. CS-328 was obtained from the Tasmanian CSIRO Collection of Living Microalgae and was grown at 30C by batch culture in 2.8 L Fernbach flasks containing 500 mL modified Zarrouks medium (initial pH 9.0) supplemented with 1 M Ni2+, 10 mM NO3-, 1 mM H2PO4, and 200 mM additional NaHCO3 (30-31). Cultures were illuminated by cool fluorescent lamps in 12 hr light/dark cycles. Optimal light intensity was about 30-70 Em-2s-1. For simultaneous H2 and NAD(P)H measurements we used samples taken after the 10th hour of the illumination cycle. A two channel device was fabricated that allows simultaneous, continuous, monitoring of dissolved H2 gas and fluorescence from reduced pyridine nucleotides with a time resolution longer than 0.1 s as described below and in (Ananyev, et al., in review).

H2 sensor based on fuel cell design. The first generation of the H2 Clark-type electrode (covered by gas permeable membrane) is described elsewhere (11). The second generation of home-built electrochemical cell is based on fuel cell technology (see review (32)) for measuring the rate of dissolved H2 concentration change. It uses a thin 5 m Teflon membrane covering a 4 mm Pt-Ir electrode. For H2 oxidation the electrode is poised at +220 mV vs. Ag/AgCl in 100 mM KCl. Between the Teflon membrane and the Pt-Ir electrode is a 20 m thick H+-conducting membrane Nafion DE 1020 (Ion Power, Inc., US) doped with micro-dispersed Pd black particles. The second generation H2 microcell has a sensitivity of ~2 nC (concentration 2.6 nM H2), fast response time (>0.1 s for 95 % signal response), and small sample volume (8 L). Such small sample volume and thickness (0.63 mm) provide fast H2 consumption throughout the whole sample. During measurement, H2 is efficiently consumed, so the H2 production rate (rather than concentration) is measured (11). The microcell has a linear response to H2 and is capable of working for months (~11-12) at stable dark current with excess AgCl (~0.5 g) in the reference electrode compartment. In this study, H2 rate is

103

presented as electrical current (I, nA) and can be converted to gas volume by using Faradays second law (32) after digital integration of the current (total electrical charge dQ = I*dt (coulomb).

Optical fluorescence unit for NAD(P)H measurement. The design is based on intermittent illumination of the sample from the top at a 45 angle by a pulsed illumination (250 ms) UV-365 nm LED (Nichia, Japan). This source (optical power 1,000 W/cm2) was pulsed at a fixed duty cycle of 20% (0.2 s on, 0.8 s off). This illumination has been found to have a negligible effect on photosynthetic electron transfer and autofermentative dark H2 production. Using an additional interference filter 365 10 nm (Intor, Inc., US) greatly reduced artifacts from excitation light. NAD(P)H fluorescence emission was selected by interference filter centered at 450 30 nm (Intor, Inc., US). The signal from an amplified photodiode S5591 (Hamamatsu Photonics, Japan) was filtered by a preamplifier in the range of DC-100 Hz (Model 113, EG&G, US). All parts of the optical unit were made using black Delrin that has very low UV induced fluorescence. The irradiated parts were coated with carbon containing black enamel NG SF 100 (Dupli-Color, US), to further reduce the UV background fluorescence an additional 3-5 fold.

For measurements, A. maxima cells were taken from the growth phase (~1g DW/L), concentrated to ~15 g DW/L, and transferred to the sample cavity of the electrode. After covering the electrode cavity, cellular respiration consumes the dissolved O2 and the culture becomes anaerobic (within 23 minutes, measured electrochemically by bare Pt electrode). All figures presented are from representative measurements of replicated experiments (3 or more) unless otherwise noted.

Chlorophyll fluorescence and quantum efficiency. Chlorophyll-a (Chl-a) variable fluorescence yield, which is proportional to PSII quantum efficiency, was measured with a home-built laser-based fast repetition rate fluorometer (FRRF) described elsewhere (33). Cultures were examined before and
104

after autofermentation for quantum efficiency of PSII and contaminant microbial growth using optical microscopy.

Fermentative Hydrogen Incubation in Glass Air-tight Vials. For vial measurements, cells were taken from the growth phase and transferred to glass vials, leaving 55-75% of the volume as headspace. Each was made air-tight by crimp top cap containing rubber septa, and the headspace was purged for 20 minutes with argon. Cultures incubated in darkness were wrapped carefully in aluminum foil.

H2 measurement by Gas Chromatography. H2 gas was measured from the headspace by gas chromatography. Gas samples (200 L) were taken from the headspace of the incubation vials by air-tight syringe (Hamilton Company, US) and injected into a Gow-Mac (Newark, NJ, US) gas chromatograph equipped with a 6 13X molecular sieve packed column and thermal conductivity detector. Argon (Airgas, Inc.) was the carrier gas, running at 30 mL/minute.

Specific Volumes and Cell Densities for Experiments. In Figure 4.1, the total vial volume is 250 mL, with 40 mL sample at 15 g DW/L. In Figure 4.2, the total vial volume is 58 mL, with 25 mL sample at 14 g DW/L. In Figure 4.3, the sample volume is 8 L at 35 g DW/L.

105

Results I. Autofermentative H2 production yields at different biomass concentrations. Batch cultures of A. maxima cells produce H2 by indirect biophotolysis (11, 16). Photoautotrophic growth of A. maxima produces biomass yields of up to 8.2 g DW/L at a quantum efficiency up to 6.2% when grown with 650 10 nm LED illumination ((20) and Ananyev et al., unpublished data). A. maxima cells can be composed of up to 48% carbohydrates (glycogen and osmolytes) and during (dark) autofermentation, carbohydrates are converted to H2 and a variety of excreted secondary metabolites (succinate, lactate, acetate, formate, ethanol, and CO2 have been reported (11)). The genes for the corresponding enzymes have been identified in the genome (20). To identify the upper limit of H2 producing capacity, we diluted the sample (containing growth medium and cells) with double deionized water prior to anaerobic incubation. Figure 4.1A shows dark autofermentative H2 accumulation in the headspace at different degrees of dilution. The sample diluted 2-fold shows a total H2 increase of 19%, while samples diluted by 4- and 8-fold show decreases of 15.4 and 64% respectively. However, the H2 accumulation normalized to biomass (cell dry weight) is significantly higher among all dilutions tested (Figure 4.1B). The results show that decreasing the concentration of biomass and residual nutrients by dilution prior to autofermentation increases the H2 yield between 2.5- and 3-fold. The cultures that were hypotonically stressed the most (dilutions of 4- and 8-fold) exhibited cell lysis (rupture) after 9 and 5 days autofermentation, respectively. This is a result of an increased rate of catabolizing intracellular storage carbohydrates, soluble sugars, and osmolytes (trehalose and glucosyl-glycerol) in response to the hypotonic stress.

106

II. H2 uptake from headspace and equilibrium constant. The bidirectional hydrogenase present in A. maxima is reversible and accordingly could take up H2 and reduce NAD+ under certain conditions. Figure 4.2 shows the kinetics of H2 production and consumption by A. maxima at 30C starting from an initial headspace concentration of 20% H2 (80% N2 to balance) during 6 days of dark or light incubation. A simple calculation shows that the amount of H2 withdrawn for GC measurement (0.2 mL x 7 measurements = 1.4 mL, that is 4.2% of headspace) is coincident with the decrease observed in the experimental light time course. Illumination suppresses hydrogenase activity and we therefore utilize this light time course as our control. When this control is subtracted from the dark time course, the differential trace indicates a near complete elimination of H2 production activity in an atmosphere of 20% H2 above the culture. Interestingly, while H2 uptake is not observed in the experiment even in the presence of 20% H2 in the atmosphere, biological H2 production is inhibited due to the H2 backpressure. Comparing the data in Figure 4.1 and Figure 4.2 shows that the initial 20% H2 in the headspace inhibits dark auto-fermentative H2 production by 6-8-fold (depending on the time point in incubation). Therefore the equilibrium constant for the hydrogenase reaction, Keq = k1/k-1, is equal to ~6-8, where k1 is the rate constant of forward reaction (at initial [H2] = 0) and k-1 is rate constant of the back reaction (at initial [H2] = 20%). In other words, a headspace of 18-20% hydrogen almost completely eliminates biological H2 evolution.

III. Selective milking of H2 from A. maxima electrochemically. A faster, more exhaustive, and selective method for the removal of the product H2 directly from the medium during autofermentation is through the use of a fuel cell electrode (see Materials and Methods). When the Pd/Nafion coated electrode is separated from the culture by a gas permeable membrane and polarized by +220 mV, such a device is specific to hydrogen gas and does not react to other excreted products. The fuel cell generates electrical current arising from the
107

oxidation of H2 (linear over a wide range of H2 concentrations). This electrical current reflects the rate of H2 production; it can be converted to electrical charge by integration over a period of

hydrophobic membrane prevents diffusion of any products between the sample and the anode electrochemical compartment except H2 gas. Figure 4.3A shows sustained H2 production during 7 days of dark autofermentation. The A. maxima culture was conditioned for maximum H2 production by the approach described in (11). Oxygen was fully consumed from the 8 L sample within 2-3 min, and H2 production started during that same period, indicative of (i) a hydrogenase quickly activated upon anoxia and (ii) the depletion of nitrate, a electron acceptor competing for reductant, initially supplied in the growth medium. The average electrical current density during the initial 4 days of dark anaerobic incubation reaches about 5.6 A/cm2 at the anode surface. After 4-5 days the H2 production rate decreases and eventually reaches zero after the 7th day. During this time, the total carbohydrate content of the cells decreased drastically from 48% (initial) to below 1% (terminal) (20). The carbohydrate availability therefore defines the capacity for fermentative H2 production, and an initial carbohydrate content of 48% allows A. maxima cells to survive for at least 7 and up to 10-12 days, after which point H2 production ceases. In most cases such cells remain viable and can be used for repeated autotrophic growth in fresh Zarrouks medium. Figure 4.3B shows the digital integration of the data in Figure 4.3A. This concentration kinetic shows the maximal cumulative production of H2 (yield) by A. maxima cells in the absence of any inhibition by H2 uptake (280 mL H2/g DW). The plateau of the H2 yield kinetic after 7 days indicates the point at which H2 production ceases due to completed carbohydrate consumption. The data representing headspace H2 accumulation from undiluted control cells was taken from Figure 4.1 and replotted here (diamond trace) to illustrate the 11-fold increase in autofermentative H2

108

yield (over 7 days) when the product H2 is continually consumed electrochemically (milking conditions).

IV. Status of photosynthetic viability following several days of autofermentation. For biotechnological applications, it is important to know if the cultures are still viable after a single cycle of autofermentation. PSII activity can serve as a general gauge of cell integrity. Loss of PSII activity over several days of autofermentation is due to degradation of the PSII reaction center (including the D1 protein) and the water-oxidizing Mn4Ca1O5 active site. Figure 4.4A shows the PSII activity, measured by Chl-a variable fluorescence yield, before and after 10 days of autofermentation. The results indicate that within 24h following the anaerobic incubation, total recovery of PSII activity is achieved for the sample. Initially, about 80-90% of A. maxima cells have lost their PSII activity (initial quantum efficiency Fv/Fm is about 0.05), but it is recoverable. The application of low intensity light (average 2 Em-2s-1, 660 nm) recovers 50% of PSII activity in 50 min (Figure 4.4B), and about 70-80% of PSII activity was restored to the initial level after 24 h of low light illumination. Of course, the overall survival and viability of cells after autofermentation depends on a number of factors, including: i) microbial contamination; ii) initial carbohydrate content; iii) length of anaerobic incubation; iv) incubation temperature; v) stirring of bioreactor; vi) biomass concentration; vii) pH of media; and viii) net yield of H2 produced. Balancing all factors, our current and previous (11) studies have found that the maximal incubation period for milking H2 production from A. maxima is about 7-8 days, which allows greater than 70-80% recovery of cell viability.

V. H2 production in the presence of external organic acids and ethanol.

109

Another potential mechanism to increase the yield of fermentative H2 production is to redirect the partitioning of the NAD(P)H produced by the catabolic reactions (glycolysis and OPP pathways). In principle, this may be possible by altering the concentrations of the secondary metabolites that require NADH to form and thus compete with H2 formation. A. maxima excretes six main secondary metabolites: lactate, formate, acetate, ethanol, CO2 and H2 (11, 20). The fraction of each metabolite is variable and depends on many factors, including cell age and osmotic stress (20). We focused our approach on suppressing the generation of the metabolites whose production consumes the most reducing power, such as lactate, formate, and especially ethanol (see Scheme 4.1 for the number of NADH molecules involved in reactions). Because H2 production must accompany a pathway for the excretion of the main carbon product of glycolysis (pyruvate), and because the production of acetate from pyruvate generates additional NADH which can be utilized by hydrogenase, we consider acetate production the optimal variety of carbon fermentative product. Figure 4.5 demonstrates a 35% increase in the maximal rate of H2 production as a result of adding a mixture of lactate, formate and ethanol (3 mM of each) to fermenting cultures. The addition of these products is expected to shift the equilibrium of their fermentative reactions from products to substrates and thereby increase the available NADH for hydrogenase. Not only does the maximal H2 production rate increase after the addition of the mixture, but also the time before the maximal H2 production rate is reached decreases 45%, and the rate at which the NAD(P)H is consumed by the system (measured as the slope of the NAD(P)H kinetic) is decreased as well (from 0.93 M/h in control to 0.49 M/h with treatment). Together these data confirm a decreased ability to recycle NADH by the lactate-, formate-, and ethanol- pathways, translating directly into increased NADH oxidation by hydrogenase and accordingly increased H2 production.

110

Discussion In this article we present a novel, so called milking, approach to increase H2 production by a factor of up to 11-fold in yield and 3.4-fold in rate. When transferred to dark anoxic conditions, cyanobacteria quickly begin to catabolize their carbohydrate reserves for maintenance of energetic demands (ATP, NAD(P)H, and reduced FDx) and produce fermentative by-products (34), including molecular hydrogen. Scheme 4.1 presents a simplified functional model of dark fermentative metabolic pathways in A. maxima that produce energy, excreted carbon products (organic acids, ethanol and CO2), and molecular H2 (for detailed pathways see (20)). An important consideration for attaining maximal H2 yield is the partitioning of carbon and reductant during autofermentation. Figure 4.6 shows an Eh-pH Pourbaix stability diagram indicating the redox potential of H2/H+ at various partial pressures of H2. These data show that the reduction of protons by [NiFe]hydrogenase is a process operating close to thermodynamic equilibrium under biologically relevant conditions (intracellular pH between 7 and 7.5) at 0.001 bar, and as such the system is especially sensitive to changes in H2 product concentration and competing pathways. The removal of H2 by milking is therefore expected to increase the net rate of H2 formation and thus provides a mechanistic explanation for the large, 11-fold, increase observed under milking conditions. The Pourbaix diagram additionally indicates that acidification of the cytoplasm is another mechanism that would favor proton reduction from NADH. Previous work has found that the hydrogenase is constitutively expressed in A. maxima under oxic and anoxic conditions (11) and immediately activated by NAD(P)H upon transfer to an anoxic environment (35). Moreover, dynamic electrochemical investigation has found that product H2 is a common and potent inhibitor of [NiFe] hydrogenases in general (36). Additionally, the accumulation of excreted carbon products (organic acids and alcohols) in the medium may cause feedback inhibition and slow the reactions of glycolysis.

111

In this article we report three ways to increase the yield of auto-fermentative H2 by the cyanobacterium A. maxima. The first is based on optimizing the concentration of biomass in the reaction vessel by dilution. This approach increases H2 production from 3.6 to 13.2 mL/g DW/day, a 3.7-fold increase. In addition to decreasing the extracellular concentration of fermentative products, this dilution approach decreases the availability of extracellular nitrate (an electron acceptor in competition with protons) thereby shifting metabolism from nitrate reduction to proton reduction, while also decreasing the osmolarity of the culture medium, which may induce catabolism of some osmolytes and soluble sugars in A. maxima. The second approach is based on milking by continuous removal of H2 from the medium throughout the autofermentative incubation by its selective consumption at an electrode. By eliminating product inhibition, milking increased H2 yield from 3.6 to 40 mL H2/g DW/d, or 11fold compared to non-milking conditions. Biotechnologically, this approach could be used to continuously consume H2 produced from cyanobacterial cultures and provide electrical current. A similar biotechnological approach has been developed utilizing hydrogenases adsorbed to electrodes to consume H2 and generate electrical current (37-38). A clear implication of this finding is in the design of bioreactors, where sufficient headspace is required to prevent excessive H2 backpressure from accumulating on the system. Alternatively, the inclusion of an inorganic H2 sponge or electrode to consume the H2 would be highly beneficial to attaining maximal H2 production rates. The latter approach would allow direct electrochemical oxidation of H2 to produce electricity. The third method focuses on changing the profile of fermentative products by adding a mixture of formate, lactate, and ethanol to shift the equilibria of their fermentative reactions towards substrates (NADH) and thereby increase the NADH available to hydrogenase. This is based upon the principle of Le Chatelier, stating that a chemical system at equilibrium will react to a change in concentration of its reactants or products and respond by creating a new equilibrium.
112

The range of concentrations of excreted organic acids by A. maxima is up to 3.0 mM (20) with H2 in the headspace in the range 12-18% (11). In the present work we have chosen to utilize a mixture of formate, lactate, and ethanol to inhibit the fermentative production of each. In a metabolic mode consisting of glycolysis alone, pyruvate will be generated from the oxidation of glucose equivalents and the carbon corresponding to this pyruvate must be excreted from the cell (Scheme 4.1). Accordingly, our aim is to funnel all this pyruvate to acetate via pyruvate ferredoxin oxidoreductase (PFOR) or pyruvate dehydrogenase complex (PDC) as this route will generate more internal reductant to the cell that could be converted to H2. Under this metabolic mode, glycolysis is required to generate ATP within the cell. We note, however, that if carbon could be catabolized by the enzymes of the OPP pathway, glucose equivalents could be more fully oxidized to CO2, with less of a flux required from pyruvate to acetate. The functioning of this pathway would generate more reductant for the cell per glucose catabolized (4 NADH for glycolysis alone, 8 NAD(P)H for a glycolysis/OPP hybrid pathway shown to occur in our experiments, 12 NADPH for OPP alone which occurs in vitro (13-14)). While the functioning of the glycolysis/OPP hybrid pathway directly provides ATP by the carbon flux through lower glycolysis, complete catabolism of glucose through OPP (yielding 12 NADPH) is not directly coupled to substrate level ADP phosphorylation. Most importantly, maximal H2 yield depends on the total amount of carbohydrate stored in the cell prior to fermentation. The best way to cause the cell to accumulate highest levels of carbohydrate is by nitrate depletion, low light (40-70 Em-2s-1), and a prolonged growing period. Under said conditions, 0.48 g of carbohydrate per 1 g DW is achievable and could provide a theoretical maximal H2 yield of (1490 * 0.48) = 716 mL H2/g DW. Our maximal observed experimental H2 yield is 282 mL/g DW corresponding to an energy conversion efficiency (ECE, carbohydrate to hydrogen) of 282/716 = 39%. We note that this ECE is very similar to the ECE expected (33%) from 4 equivalents H2 per glucose as outlined by the Thauer limit. The fact that it is
113

slightly higher than the 33% ECE of the Thauer limit may indicate that there is some contribution from the OPP pathway to reductant generation in the native system. We close in calling attention to Arthrospira maxima as a promising candidate for biological H2 production. Building upon the role of [NiFe]-hydrogenase to recycle NADH to NAD+ and allow fermentative reactions to continue (9), it follows that microbial phototrophs with high metabolic rates and high cellular ATP demands are promising platforms for efficient [NiFe] hydrogenases and efficient H2 production. A. maxima routinely experiences anaerobic conditions during the dark phase of its diurnal cycle and has evolved to thrive in an environmental niche consisting of high osmolarity and alkalinity (39). These environmental conditions demand a high catabolic rate for the strain to satisfy the ATP requirements demanded by this environment, and we observe this high metabolic rate translate directly into high rates of fermentative H2 production from the strain.

114

Figures

Figure 4.1. Effect of sample dilution on (A) total headspace H2 yield by A. maxima measured by gas chromatography, and (B) H2 yield normalized to cell dry weight. Conversion coefficient: 1 mL H2 = 45 mol H2. Initial carbohydrate content was measured as 48 2% and decreases to ~1% over time at which point H2 production stalls. Vertical marks in panel B indicate the point at which cell lysis is observed.

115

Figure 4.2. H2 headspace concentration above A. maxima cultures at 30C during anaerobic incubation with and without illumination (30 E m-2 s-1, supplied to suppress hydrogenase activity). A headspace of 20% H2 (added at start of incubation) near completely eliminates biological H2 evolution. For each GC measurement 200 L of headspace gas was withdrawn for sampling, and is responsible for the gradual decrease in headspace H2 observed in both traces. The bottom trace illustrates the biological H2 accumulation of ~1% (calculated by taking the difference in the top traces) in the 20% H2 atmosphere and illustrates that H2 uptake from the culture under these conditions is negligible.

116

Figure 4.3. Sustained autofermentative dark H2 production by A. maxima measured in fuel-cell type H2 cell. (A) The rate of H2 production, measured electrochemically by oxidation current is shown over the course of 7 days. (B) The cumulative yield of H2 increases linearly until the 7th day, when glycogen reserves are depleted and cell lysis occurs. By milking H2 production (continuously consuming the excreted H2 at the electrode), the yield of H2 production (348.3 mC, 3.61 mol H2) is increased 11-fold compared to a non-milking system where H2 accumulates in the headspace. Headspace accumulation of control (undiluted) cells over 7 days is shown by diamonds in B (data taken from Figure 4.1).

117

Figure 4.4. Photosynthetic viability of cultures before and after 10 days of autofermentation. The kinetics of PSII activity were measured by single turnover laser excitation flashes. A. maxima filaments were subjected to a train of 50 single turnover flashes (STFs) (each STF is 60 s in duration) applied at 20 Hz. Measurements were repeated every 2 minutes and samples were kept in darkness between measurements. (A) 700 STF trains were averaged and the Fv/Fm value for each individual STF plotted, illustrating viable PSII by oscillations in Fv/Fm. (B) The average Fv/Fm value was calculated from each STF train and plotted as a function of time. Within 24 hours, Fv/Fm reaches 70-80% of its starting level, indicating the culture is capable of regaining full photosynthetic activity after autofermentative incubation.

118

Figure 4.5. H2 production (red) and intracellular residual NAD(P)H concentration (blue) in A. maxima cells (control, A) illustrating the effect of adding a mixture of formate, lactate, and ethanol (3 mM each) in order to shift equilibrium of competing fermentative reactions toward substrates, notably NADH (B). The addition of this mixture causes an increased average rate of fermentative H2 production and decreases the time prior to H2 production reaching its maximal rate, 2.2h compared to 4h in the control. Additionally, the rate of decrease in the NAD(P)H kinetic is slower in the presence of the mixture (0.49 M/h with mixture compared to 0.93 M/h in control) indicating a buildup of NADH within the cell in response to the shifted reaction equilibrium with this treatment.

119

Figure 4.6. Eh-pH Pourbaix stability diagram of H2, H2O, NADH, and NAD+ at different H2 partial pressures. A 10-fold change in the H2 pressure results in a change of 29.6 mV in the H2/H+ reduction potential. At 0.1% headspace H2 concentration (1000 ppm, or 0.001 bar) the redox potentials of the H2/H+ and NADH/NAD+ couples are approximately equal.

120

Scheme 4.1. A simplified functional model of dark fermentative metabolic pathways in A. maxima that produce ATP, reductant, excreted carbon products (organic acids, ethanol, CO2), and molecular H2. Increasing H2 production from cyanobacteria is possible by three approaches: i) removing H2 gas by oxidation on Pt-Ir electrode in a fuel cell (so called milking), ii) decreasing the concentration of excreted products by dilution, and iii) adding external organic acids and ethanol (marked as subject of manipulation) to shift the reaction equilibrium of competing fermentative reactions toward substrates and thereby increase NADH availability to hydrogenase. Production (+) or consumption (-) of NADH equivalents is shown for the various fermentative reactions.

121

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. Kruse O & Hankamer B (2010) Microalgal hydrogen production. Current Opinion in Biotechnology 21(3):238-243. Quintana N, Van der Kooy F, Van de Rhee M, Voshol G, & Verpoorte R (2011) Renewable energy from Cyanobacteria: energy production optimization by metabolic pathway engineering. Applied Microbiology and Biotechnology 91(3):471-490. Schtz K, Happe T, Troshina O, Lindblad P, Leito E, Oliveira P, & Tamagnini P (2004) Cyanobacterial H2 production a comparative analysis. Planta 218(3):350-359. Prince RC & Kheshgi HS (2005) The Photobiological Production of Hydrogen: Potential Efficiency and Effectiveness as a Renewable Fuel. Critical Reviews in Microbiology 31(1):1931. Chisti Y (2007) Biodiesel from microalgae. Biotechnology Advances 25(3):294-306. Williams PJlB & Laurens LML (2010) Microalgae as biodiesel & biomass feedstocks: Review & analysis of the biochemistry, energetics & economics. Energy & Environmental Science 3(5):554-590. Lehninger AL (1982) Principles of biochemistry (Worth Publishers, New York, N.Y.). Tamagnini P, Axelsson R, Lindberg P, Oxelfelt F, Wunschiers R, & Lindblad P (2002) Hydrogenases and hydrogen metabolism of cyanobacteria. Microbiology And Molecular Biology Reviews 66(1):1-20. Carrieri D, Wawrousek K, Eckert C, Yu J, & Maness P-C (2011) The role of the bidirectional hydrogenase in cyanobacteria. Bioresource Technology 102(18):8368-8377. Bartacek J, Zabranska J, & Lens PNL (2007) Developments and constraints in fermentative hydrogen production. Biofuels Bioproducts & Biorefining-Biofpr 1(3):201-214. Ananyev G, Carrieri D, & Dismukes GC (2008) Optimization of metabolic capacity and flux through environmental cues to maximize hydrogen production by the cyanobacterium "Arthrospira (Spirulina) maxima". Applied and Environmental Microbiology 74(19):61026113. Thauer RK, Jungermann K, & Decker K (1977) Energy conservation in chemotrophic anaerobic bacteria. Bacteriol Rev 41(1):100-180. Woodward J, Orr M, Cordray K, & Greenbaum E (2000) Biotechnology: Enzymatic production of biohydrogen. Nature 405(6790):1014-1015. Zhang YHP, Evans BR, Mielenz JR, Hopkins RC, & Adams MWW (2007) High-Yield Hydrogen Production from Starch and Water by a Synthetic Enzymatic Pathway. PLoS ONE 2(5):e456. Hallenbeck PC (2005) Fundamentals of the fermentative production of hydrogen. (I W a Publishing), pp 21-29. Hallenbeck PC & Benemann JR (2002) Biological hydrogen production; fundamentals and limiting processes. International Journal of Hydrogen Energy 27(11-12):PII S03603199(0302)00131-00133. Ghirardi ML, Posewitz MC, Maness PC, Dubini A, Yu JP, & Seibert M (2007) Hydrogenases and hydrogen photoproduction in oxygenic photosynthetic organisms. Annual Review of Plant Biology 58:71-91. Surzycki R, Cournac L, Peltiert G, & Rochaix JD (2007) Potential for hydrogen production with inducible chloroplast gene expression in Chlamydomonas. Proc Natl Acad Sci USA 104:17548-17553. Meuser J, Boyd E, Ananyev G, Karns D, Radakovits R, Narayana Murthy U, Ghirardi M, Dismukes G, Peters J, & Posewitz M (2011) Evolutionary significance of an algal gene encoding an [FeFe]-hydrogenase with F-domain homology and hydrogenase activity in Chlorella variabilis NC64A. Planta 234(4):829-843.

12. 13. 14. 15. 16. 17. 18. 19.

122

20.

21. 22. 23. 24. 25. 26. 27. 28. 29.

30. 31. 32. 33. 34. 35. 36. 37. 38.

Carrieri D, Momot D, Brasg IA, Ananyev G, Lenz O, Bryant DA, & Dismukes GC (2010) Boosting Autofermentation Rates and Product Yields with Sodium Stress Cycling: Application to Production of Renewable Fuels by Cyanobacteria. Appl. Environ. Microbiol. 76(19):6455-6462. Aoyama K, Uemura I, Miyake J, & Asada Y (1997) Fermentative metabolism to produce hydrogen gas and organic compounds in a cyanobacterium, Spirulina platensis. Journal of Fermentation and Bioengineering 83(1):17-20. Dephilippis R, Sili C, & Vincenzini M (1992) Glycogen And Poly-Beta-Hydroxybutyrate Synthesis In Spirulina Maxima. Journal Of General Microbiology 138:1623-1628. Zhang HS, Bruns MA, & Logan BE (2006) Biological hydrogen production by Clostridium acetobutylicum in an unsaturated flow reactor. Water Research 40(4):728-734. Alshiyab H, Kalil MS, Hamid AA, & Yusoff WMW (2008) Removal of headspace CO2 increases biological hydrogen production by C. acetobutylicum. Pak J Biol Sci 11(19):2336-2340. Park W, Hyun SH, Oh SE, Logan BE, & Kim IS (2005) Removal of headspace CO2 increases biological hydrogen production. Environmental Science & Technology 39(12):4416-4420. Van Ginkel S & Logan BE (2005) Inhibition of biohydrogen production by undissociated acetic and butyric acids. Environmental Science & Technology 39(23):9351-9356. Salerno MB, Park W, Zuo Y, & Logan BE (2006) Inhibition of biohydrogen production by ammonia. Water Research 40(6):1167-1172. McNeely K, Xu Y, Bennette N, Bryant DA, & Dismukes GC (2010) Redirecting Reductant Flux into Hydrogen Production via Metabolic Engineering of Fermentative Carbon Metabolism in a Cyanobacterium. Appl. Environ. Microbiol. 76(15):5032-5038. Baebprasert W, Jantaro S, Khetkorn W, Lindblad P, & Incharoensakdi A (2011) Increased H2 production in the cyanobacterium Synechocystis sp. strain PCC 6803 by redirecting the electron supply via genetic engineering of the nitrate assimilation pathway. Metabolic Engineering 13(5):610-616. Carrieri D, Ananyev G, Brown T, & Dismukes GC (2007) In vivo bicarbonate requirement for water oxidation by Photosystem II in the hypercarbonate-requiring cyanobacterium Arthrospira maxima. J Inorg Biochem 101(11-12):1865-1874. Carrieri D, Ananyev G, Garcia Costas AM, Bryant DA, & Dismukes GC (2008) Renewable hydrogen production by cyanobacteria: nickel requirements for optimal hydrogenase activity. International Journal of Hydrogen Energy 33:2014-2022. Korotcenkov Ghenadii SDH, and Joseph R. Stetter (2009) Review of Electrochemical Hydrogen Sensors. Chemical Reviews 109(3):1402-1433. Ananyev G & Dismukes GC (2005) How fast can Photosystem II split water? Kinetic performance at high and low frequencies. Photosynthesis Research 84(1-3):355-365. Stal LJ & Moezelaar R (1997) Fermentation in cyanobacteria. FEMS Microbiology Reviews 21(2):179-211. Germer F, Zebger I, Saggu M, Lendzian F, Schulz R, & Appel J (2009) Overexpression, Isolation, and Spectroscopic Characterization of the Bidirectional [NiFe] Hydrogenase from Synechocystis sp. PCC 6803. Journal of Biological Chemistry 284(52):36462-36472. Armstrong FA, Belsey NA, Cracknell JA, Goldet G, Parkin A, Reisner E, Vincent KA, & Wait AF (2009) Dynamic electrochemical investigations of hydrogen oxidation and production by enzymes and implications for future technology. Chem. Soc. Rev. 38:3651. Shastik ES, Vokhmyanina DV, Zorin NA, Voronin OG, Karyakin AA, & Tsygankov AA (2011) Demonstration of hydrogenase electrode operation in a bioreactor. Enzyme and Microbial Technology 49(5):453-458. Vincent KA, Cracknell JA, Lenz O, Zebger I, Friedrich B, & Armstrong FA (2005) Electrocatalytic hydrogen oxidation by an enzyme at high carbon monoxide or oxygen levels. Proc Natl Acad Sci USA 102(47):16951-16954.
123

39.

Vonshak A (1997) Spirulina platensis (Arthrospira) : physiology, cell-biology, and biotechnology (Taylor & Francis, London ; Bristol, PA) pp xvii, 233 p.

124

Chapter 5: Development and Application of a DOP-PCR Based Genetic Screen for [NiFe] Hydrogenases: Bioprospecting for Efficient Hydrogen Producing Organisms

Co-authors: Elizabeth H. Burrows, Michael Abrams, Diana Chien, Derrick R.J. Kolling, Susan L. Brown, and G. Charles Dismukes

125

Abstract Degenerate (universal) primers were designed and implemented to screen cyanobacterial culture collections for genes encoding the large subunit of bidirectional hydrogenase (hoxH, active site) by PCR. This NAD+-dependent soluble hydrogenase can evolve H2 from anaerobic cells under reductive stress and is promising as an enzymatic route toward sustainable biological hydrogen production. However, oxygen sensitivity and relatively slow turnover are limiting factors in obtaining maximal H2 yields, and thus the identification of novel hydrogenases from less-studied cyanobacteria is a research thrust in the field. The method described herein allows for rapid screening of these less-studied cyanobacteria. A collection of 24 axenic cyanobacterial cultures (from the University of Hawaii Culture Collection) was screened and hoxH was identified in 10 of the 24 strains.

126

Introduction Cyanobacteria are a diverse group of prokaryotes that perform oxygenic photosynthesis and are credited with converting the earths atmosphere from reducing to oxidizing approximately 2.8 billion years ago (1-2). The phylum consists of more than 10,000 species, many of which occupy extreme environments such as high salt (3), acidity/alkalinity (4-5), or temperature (6). In addition to photosynthesis, they are capable of aerobic respiration and acidogenic fermentation (7). Many strains are also capable of hydrogen metabolism, which can be mediated by nitrogenase, or by either of two distinct [NiFe]-hydrogenases (8-10). Nitrogenase is a molybdenum-containing protein complex capable of reducing atmospheric N2 to NH4+ in an ATP-requiring reaction that utilizes reduced ferredoxin as an electron source. With every complete cycle (1 equivalent N2 reduced), an obligate equivalent of H2 is released. In almost all nitrogen-fixing cyanobacteria, a membrane-bound respiratory hydrogenase is co-expressed with nitrogenase (11). Its canonical function is to oxidize the H2 byproduct from nitrogenase to H+ and e-, which are immediately transferred along an intra-membrane cascade to a terminal electron acceptor (e.g. O2). In this sense, the enzyme functions unidirectionally and is classified as an uptake hydrogenase. A second distinct soluble hydrogenase is found in many cyanobacteria, and has no observed correlation to nitrogen fixation, energy metabolism (ATP), nor to any specific physiological environment (9). This hydrogenase is a heteropentameric enzyme, containing a [NiFe] active site and NAD+-reducing (diaphorase) moiety (12). It is denoted a bidirectional enzyme as it can either reduce protons or oxidize H2, depending on the redox state of the cell. This hydrogenase is of substantial interest to the scientific community as a possible biological platform for renewable H2 generation (13). [NiFe]-hydrogenases have two well recognized limitations relating to their biotechnological application: (a) they are inhibited by oxygen and (b) they function at a relatively slow rate compared to the [FeFe]-hydrogenases (found in some green algae), which are reported to exhibit
127

catalytic rates 100x faster than [NiFe]-hydrogenases (14-15). There has existed, therefore, a significant research effort to identify novel hydrogenases from less studied species of cyanobacteria that may not suffer from these limitations. Previous screens have utilized gas chromatography to assay headspace H2 concentration from cultures under dark, anaerobic conditions (16), focused on genetic methods to isolate lambda clones containing hupB (17), and pioneered the development of chemochromic sensors for visual detection of evolved H2 from cultures (18). Large collections of cyanobacteria are available from culture collections that could be screened for hydrogenase activity or its genetic presence in an effort to target candidate organisms and enzymes for more intense biochemical characterization and also as a means to establish the frequency of hydrogenase occurrence in nature. To address these questions, we have designed degenerate (universal) primers to amplify a 792-bp fragment of genomic DNA coding for the large subunit of the bidirectional hydrogenase from cyanobacteria (hoxH), utilizing two regions in the protein sequence that exhibit complete conservation among cyanobacterial sequences from 16 diverse species (Figure 5.1). We applied these primers to screen 24 strains (for which no genomic information is known) and identified 10 of them as containing [NiFe]-hydrogenase. Similar approaches of degenerate primer design have been developed to screen for the [FeFe]-hydrogenase in eukaryotic algae (19) and for the membrane-bound uptake [NiFe]-hydrogenase (Group 1 as classified by Vignais, et al.) (20) among bacteria (21). Because the bidirectional variety of [NiFe]-hydrogenases (Group 3d) are capable of evolving H2, they are the more promising enzymes to identify for biotechnological applications, and consequently are the topic of our current work.

128

Methods Culture Growth Strains were obtained from the University of Hawaii at Manoa (University of Hawaii Culture Collection) where they were maintained as axenic cultures. They were transported and maintained on agar plates in solid BG0 medium. The subset of the strains which grew well in liquid medium were inoculated into liquid BG-11 medium (22) and grown at 35C with 24-h illumination at approximately 120 E m-2 s-1 prior to culture assays.

Primer Design, DNA Extraction, and Amplification Parameters Primers (Table 5.1) were ordered from Integrated DNA Technologies (Coralville, IA) and PCR enzymes and reagents from Applied Biosystems (Carlsbad, CA). To extract genomic DNA, the cultures were concentrated, exposed to a freeze/thaw cycle, and incubated at 37C for 30 minutes in the presence of lysozyme (40 mg/ml). DNA was extracted twice by a phenol/chloroform method and washed with ethanol. The PCR reaction mixture (50 L total volume) contained 5 L of 10x reaction buffer (150 mM Tris-HCl, 500 mM KCl, 15 mM MgCl2, pH 8.0), 1 L of 10 mM dNTP mix, 2 L of both forward and reverse primer at 100 M, 50 ng DNA as template, 2.5 units of Amplitaq polymerase, and nuclease-free water. Cycling was carried out on an Applied Biosystems GeneAmp PCR System 9700 with the following parameters: Initial Denaturation (94C for 4 minutes), then Touchdown Amplification (13 cycles of 94C for 30 sec, 50C for 40 sec, 72C for 70 sec) with the annealing temperature (initially 50C) decreased by 1C per cycle, then Standard Amplification (20 cycles of 94C for 30 sec, 55C for 40 sec, 72C for 70 sec), and lastly a final elongation at 72C for 7 minutes. Amplification products were analyzed by electrophoresis on a 1% agarose gel in TAE buffer and ethidium bromide was used for visualization of DNA.

Hydrogenase Activity and Chlorophyll Assays

129

Methyl viologen assays were performed to quantify hydrogenase activity in a method similar to Zorin, et al. (23). Samples were grown to high biomass for each strain (to maximize the possibility of detecting hydrogen), and as much material as possible was spun down and resuspended in 5 mL of optimized BG-11 (24). Samples were placed in sealed 10 mL glass GC vials with a septa crimptop, and purged with argon for at least 10 minutes. Immediately prior to removing the feed and exhaust needles, 1 mL of BugBuster was added through the septa cap via syringe, followed by 0.5 mL of 100 mM sodium dithionite in 30 mM NaOH and 1 mL of 10 mM methyl viologen in 50 mM potassium phosphate buffer at pH = 6.9 . H2 production was measured by gas chromatography, by injecting 200 L from the headspace into a Gow-Mac Series 580 GC equipped with a thermal conductivity detector. A 6 13X molecular sieve column and argon carrier gas (30 mL/min flow) was used to perform the chromatographic separation. Chlorophyll was determined by methanol extraction from the sample and a subsequent absorbance measurement at 665nm (25). The limit of detection for this assay (over a 3 hour incubation) was 0.16 mol H2 (mg Chla)-1 hr-1.

Hydrogen Rate Electrode Measurements The first generation of our home-built H2 microcell is described elsewhere (5). Our second (current) generation electrochemical microcell is also a reverse Clark-type electrode for measuring dissolved H2 concentration based on fuel cell technology. It consists of a Teflon membrane covered Pt/Ir electrode biased at +220 mV for H2 oxidation. This 2nd generation microcell has an increased sensitivity of ~2 x 10-9 Coulomb H2 (LOD = 0.10 mol H2 (mg Chla)-1 hr-1), fast response time (100 ms), and micro-volume sample chamber (6.5 L). During measurement, H2 is constantly consumed by oxidation at the Pt/Ir electrode, allowing the instantaneous rate of H2 production to be measured. The microcell responds linearly to H2 and performs for several months using an oversized Ag/AgCl reference electrode (~0.5 g AgCl) to consume the H+ product of H2 oxidation. The microcell is equipped with a light emitting diodes (LED) at a wavelength of 67010 nm
130

(FWHM) for illumination of the Photosystems and the assay of photo-H2 production. Pulsed illumination was supplied to the cultures at a frequency of 0.01 Hz and a duty cycle of 10%, i.e. repeated cycles of 10 seconds light followed by 90 seconds dark.

131

Results A multiple sequence protein alignment of hoxH is shown in Figure 5.1. Sixteen strains were analyzed and aligned with Clustal X2: Anabaena variabilis ATCC29413, Nostoc PCC7120, Nostoc PCC7422, Spirulina subsalsa, Arthrospira platensis, Arthrospira maxima, Prochlorothrix hollandica, Synechococcus elongatus spp. PCC 7941 and 6301, Synechocystis PCC6803, Cyanothece spp. ATCC51142, PCC8801, PCC7424, and PCC7425, Microcystis aeruginosa, Acaryochloris marina, and Synechococcus PCC7002. The alignment yielded 6 candidate regions of conservation (denoted A through F), which were used to create degenerate primers (Table 5.1). Combinations of the three forward (A,B,C) and three reverse (D,E,F) primers were tested in PCR amplifications utilizing genomic DNA known to contain hoxH. The combination of forward primer A (5-GGN TTY GAR AAR TTY TGY GAR-3) and reverse primer D (5-NGC NAR NGG NCC NAC NCK RTA-3) was chosen for the screen, because it provided the most specific amplification of the desired target (complete absence of non-specific amplified bands) and most efficient amplification (intensity of band). To verify the specificity towards hydrogenase and to demonstrate that the primers would identify hoxH presence in a variety of cyanobacteria, fully sequenced (model) strains were assayed with our degenerate primers (Figure 5.2). The five positive-control sequenced strains known to contain [NiFe]-bidirectional hydrogenase each showed clear bands of the appropriate size fragment (792 bp), while the two negative-control strains known to lack the enzyme, as well as a no-template control (NTC), failed to yield a fragment of the corresponding length, as expected. The results from the screen of the 24 cultures from the University of Hawaii collection is shown in Table 5.2. Ten species were identified as containing hoxH among the strains, representing genera Leptolyngbya, Oscillatoria, Hapalosiphon, Synechococcus, Nostoc, and Pseudocytonema. This distribution of hydrogenase among our diverse, but small, culture collection

132

suggests a rough estimate for frequency of hydrogenase presence in cyanobacteria of ~40%, consistent with the ~50% estimate obtained assaying genomes deposited in GenBank (11). In order to further corroborate the PCR results, all strains that grew in liquid media, regardless of whether they were identified to contain hoxH, were screened for in vitro and in vivo hydrogenase activity (Table 5.2). In future applications, only characterizing the positive hits would be necessary. Strains were first transferred to liquid culture and preliminary characterization was conducted to determine optimal growth conditions. Then, in vitro hydrogenase activity was measured on all strains that accumulated sufficient biomass and that had a phenotype conducive to vial experiments (some strains grew in thick biofilms or globs), a total of 10 strains. Of the ten, two of the strains, which were also identified as containing hoxH, illustrated significant in vitro hydrogenase activity (utilizing methyl viologen as exogenous source of reductant) of 0.24 0.15 mol H2 (mg Chla)-1 hr-1 (#2, Nostoc muscorum) and 0.39 0.31 mol H2 (mg Chla)-1 hr-1 (#22, taxonomy unknown). Additionally, both of these strains exhibited biological H2 production under anoxia when measured in vivo by a H2 rate electrode (Figure 5.3). None of the other strains tested illustrated detectable in vitro or in vivo hydrogenase activity. Limits of detection (LOD) were 0.16 mol (mg Chla)-1 hr-1 for the in vitro screen (measured over a 3 h period) and 0.10 mol H2 (mg Chla)-1 hr-1 for the in vivo screen (which measures an instantaneous rate). We note that of the total 24 strain, only 3 strains both grew well in liquid media (single fixed conditions) and also tested positive for hydrogenase (strains 2, 22, 24). Two of those exhibited in vitro and in vivo hydrogenase activity as identified above and in Table 5.2 (strains 2, 22). It is likely, therefore, that many of the other 7 strains identified as containing hydrogenase would also exhibit activity if conditions allowing for significant cell growth were surveyed (not examined for these widely diverging strains). However, from a biotechnological standpoint, strains must not only possess hydrogenase, but also be amenable to culturing. Because the two strains (#2 and #22) possess both attributes, they are promising targets for future biochemical studies. The observation
133

of both in vitro and in vivo hydrogenase activity from these two strains illustrates the utility of the genetic screen as a proxy for functional hydrogenase.

134

Discussion Herein, we have demonstrated (for the first time, to our knowledge) novel degenerate primers that can be used to screen a wide range of cultures for the genetic presence of bidirectional hydrogenase in a high-throughput fashion. In addition to targeted genetic engineering of known hydrogenases in model organisms (e.g. to increase their catabolic rate, expression level, or oxygen tolerance), bioprospecting should be pursued as an alternative means to identifying enzymes with the same advantages. Our primers allow for the robust screening and bioprospecting of the vast quantity of cyanobacteria culture collections in existence worldwide as the first stage of this pursuit for these enzymes. It is important to note that our primers are specific to bidirectional hydrogenase and will not spuriously amplify the uptake hydrogenase found in nitrogen-fixing cyanobacteria, thereby causing a false positive identification. This is due to the fact that the amino acid sequence used to design primer A is unique to the bidirectional enzyme, and does not have an equivalent motif in the uptake enzyme. Additionally, the expected fragment size (792 bp) is conserved among cyanobacteria, and this parameter can be used to distinguish the desired amplicon from nonspecifically amplified bands of different size. The use of Touchdown PCR (26) to incrementally decrease the annealing temperature during the first amplification phase of the PCR also serves to limit non-specific interactions between the primers and the genomic DNA. With Touchdown PCR, we are able to (a) use a standard concentration of MgCl2 in the reaction vessel (1.5 mM) (27), (b) use a 25/25/25/25 percent mixture of A/T/G/C at nucleotide positions where 100% degeneracy is found (rather than substitute the base with an inosine) (28-30), and (c) omit the addition of DMSO, TEAC, betaine, or any other chemical additive that destabilizes the structure of the primer-DNA complex and serves to decrease non-specific amplification (31-32). By contrast, without the Touchdown method, we

135

were unable to cleanly amplify our desired fragment of hoxH even with the incorporation of the three above-mentioned additives. We originally envisioned a screen without the need to extract DNA from the cultures via a separate phenol/chloroform extraction step, but instead where a small volume of cell suspension would be added directly to the PCR mixture, providing the necessary DNA template, so-called whole-cell PCR. Such methods rely upon the initial denaturation step of the PCR thermal profile to lyse the cells and release template DNA. While this approach worked for some of the model cyanobacterial strains we work with, it failed for the majority of tested strains. We attempted whole-cell PCR following various pre-treatments, including multiple freeze/thaw cycles in the presence and absence of two different non-ionic detergents (Triton X-100 and n-dodecyl -Dmaltoside) but were unable to achieve DNA release and successful whole-cell PCR from 100% of our model strains under any conditions. We did, however, observe that a single freeze/thaw cycle prior to the lysozyme treatment of our DNA extraction method substantially increased the yield, and subsequently incorporated it into our standard protocol (see Methods). We note that strains of the genus Cyanothece were especially hard to lyse, which may well be due to the extensive exopolysaccharides known to be produced from strains of the genus, which may serve to protect the cell from osmotic and/or temperature stress (33-34). We anticipate that our primers and cycling parameters could find wide use in the field and, in conjunction with screening for O2 tolerance, could lead to the identification of naturally occurring strains of cyanobacteria with hydrogenases exhibiting characteristics promising for biotechnological applications.

136

Figures

137

Figure 5.1. Sequence alignment of hoxH sequences from 16 strains of cyanobacteria, created with Clustal X2. Three forward primers (A, B, C) and three reverse primers (D, E, F) were designed based upon regions exhibiting complete (100%) conservation among strains.

138

Figure 5.2. Amplification results illustrating accuracy of degenerate primers. Five sequenced strains known to contain [NiFe]-bidirectional hydrogenase each show clear bands of appropriate size (792 bp, lanes 1-5), while two sequenced strains known to lack the enzyme and a no-template control (NTC) failed to amplify a fragment of appropriate length (lanes 8-10). Lanes: 1-Cyanothece sp. ATCC51142, 2-Cyanothece sp. Miami BG043511, 3-Synechococcus sp. PCC7002, 4-Arthrospira maxima, 5-Synechocystis sp. PCC603, 8-Chlamydomonas reinhardtii, 9-Phaeodactylum tricornutum, 10-NTC.

139

Figure 5.3. In vivo electrochemical detection of dissolved hydrogen. Strains #2 and #22, which exhibited appreciable in vitro hydrogenase activity, also exhibited in vivo hydrogen evolution under anaerobic conditions and in response to pulsed LED illumination by red light (start and stop of pulsed illumination is indicated by upward and downward arrows, respectively). Strain #2 (Nostoc muscorum) exhibited significant dark H2 production (0.3 1.0 mol H2 (mg Chla)-1 hr-1), which decreased to zero during each of the illumination periods. Strain #22 showed a transient level of dark H2 production (maximum 0.5 mol H2 (mg Chla)-1 hr-1) at the immediate onset of anoxia, which declined to zero. However, it showed positive photo-H2 activity in response to illumination.

140

Tables

Table 5.1. Conserved amino acids regions used to create the candidate degenerate primers, and their corresponding DNA sequence. The degeneracy, both total and normalized (degeneracy per base) are shown. Forward primer A and reverse primer D were selected as they produced the most specific amplification.

141

Table 5.2. Culture numbers, species (if known), and results from screen. 10 of the 24 strains tested were shown to possess the [NiFe]-bidirectional hydrogenase and are indicated by +. Strains which grew in liquid medium (10 total) were assayed for in vitro and in vivo hydrogen production. Limits of detection (LOD) were 0.16 and 0.10 mol H2 (mg Chla)-1 hr-1 for the in vitro and in vivo screens, respectively. I/B indicates strains with insufficient biomass for activity assays.

142

References 1. Dismukes GC, Klimov VV, Baranov SV, Kozlov YN, DasGupta J, & Tyryshkin A (2001) The origin of atmospheric oxygen on Earth: The innovation of oxygenic photosynthesis. Proceedings of the National Academy of Sciences of the United States of America 98(5):21702175. Peschek GA, Obinger C, & Paumann M (2004) The respiratory chain of blue-green algae (cyanobacteria). Physiol Plant 120(3):358-369. Garcia-Pichel F, Nbel U, & Muyzer G (1998) The phylogeny of unicellular, extremely halotolerant cyanobacteria. Archives of Microbiology 169(6):469-482. Steinberg CEW, Schfer H, & Beisker W (1998) Do Acid-tolerant Cyanobacteria Exist? Acta hydrochimica et hydrobiologica 26(1):13-19. Ananyev G, Carrieri D, & Dismukes GC (2008) Optimization of metabolic capacity and flux through environmental cues to maximize hydrogen production by the cyanobacterium "Arthrospira (Spirulina) maxima". Applied and Environmental Microbiology 74(19):61026113. Steunou AS, Bhaya D, Bateson MM, Melendrez MC, Ward DM, Brecht E, Peters JW, Kuhl M, & Grossman AR (2006) In situ analysis of nitrogen fixation and metabolic switching in unicellular thermophilic cyanobacteria inhabiting hot spring microbial mats. Proceedings of the National Academy of Sciences of the United States of America 103(7):2398-2403. Stal LJ & Moezelaar R (1997) Fermentation in cyanobacteria. FEMS Microbiology Reviews 21(2):179-211. Tsygankov A (2007) Nitrogen-fixing cyanobacteria: A review. Applied Biochemistry and Microbiology 43(3):250-259. Tamagnini P, Axelsson R, Lindberg P, Oxelfelt F, Wunschiers R, & Lindblad P (2002) Hydrogenases and hydrogen metabolism of cyanobacteria. Microbiology And Molecular Biology Reviews 66(1):1-20. Carrieri D, Wawrousek K, Eckert C, Yu J, & Maness P-C (2011) The role of the bidirectional hydrogenase in cyanobacteria. Bioresource Technology 102(18):8368-8377. Ludwig M, Schulz-Friedrich R, & Appel J (2006) Occurrence of hydrogenases in cyanobacteria and anoxygenic photosynthetic bacteria: Implications for the phylogenetic origin of cyanobacterial and algal hydrogenases. Journal of Molecular Evolution 63(6):758768. Aubert-Jousset E, Cano M, Guedeney G, Richaud P, & Cournac L (2011) Role of HoxE subunit in Synechocystis PCC6803 hydrogenase. FEBS Journal. Dismukes GC, Carrieri D, Bennette N, Ananyev GM, & Posewitz MC (2008) Aquatic phototrophs: efficient alternatives to land-based crops for biofuels. Curr Opin Biotechnol 19(3):235-240. Fontecilla-Camps JC, Volbeda A, Cavazza C, & Nicolet Y (2007) Structure/Function Relationships of [NiFe]- and [FeFe]-Hydrogenases. Chemical Reviews 107(10):4273-4303. Adams MWW (1990) The structure and mechanism of iron-hydrogenases. Biochim Biophys Acta 1020(2):115-145. Asada Y & Kawamura S (1986) Screening for cyanobacteria that evolve molecular hydrogen under dark and anaerobic conditions. Journal of Fermentation Technology 64(6):553-556. Gubili J & Borthakur D (1996) The use of a PCR cloning and screening strategy to identify lambda clones containing the hupB gene of Anabaena sp. strain PCC7120. Journal of Microbiological Methods 27(2-3):175-182. Seibert M, Benson DK, & Flynn T (2001) US 6,277,589 B1.

2. 3. 4. 5.

6.

7. 8. 9. 10. 11.

12. 13. 14. 15. 16. 17. 18.

143

19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34.

Boyd ES, Spear JR, & Peters JW (2009) [FeFe]-hydrogenase genetic diversity provides insight into molecular adaptation in a saline microbial mat community. Appl. Environ. Microbiol.:AEM.00582-00509. Vignais PM, Billoud B, & Meyer J (2001) Classification and phylogeny of hydrogenases. FEMS Microbiol Rev 25(4):455-501. Kim JYH, Jung HJ, & Cha HJ (2007) Universal degenerate oligonucleotide-primedpolymerase chain reaction for detection and amplification of NiFe-hydrogenase genes. Enzyme and Microbial Technology 42(1):1-5. Rippka R, Deruelles J, Waterbury JB, Herdman M, & Stanier RY (1979) Generic assignments, strain histories, and properties of pure cultures of cyanobacteria. J. Gen. Microbiol. 111:1-61. Zorin NA (1986) Redox Properties And Active-Center Of Phototrophic Bacteria Hydrogenases. Biochimie 68(1):97-101. Burrows EH, Chaplen FWR, & Ely RL (2008) Optimization of media nutrient composition for increased photofermentative hydrogen production by Synechocystis sp. PCC 6803. International Journal of Hydrogen Energy 33(21):6092-6099. Porra R (2002) The chequered history of the development and use of simultaneous equations for the accurate determination of chlorophylls a and b. Photosynthesis Research 73(1):149-156. Don RH, Cox PT, Wainwright BJ, Baker K, & Mattick JS (1991) Touchdown PCR to circumvent spurious priming during gene amplification. Nucleic Acids Research 19(14):4008. Anonymous (1990) PCR protocols: a guide to methods and applications (Academic Press, San Diego, California) p xviii + 482 pp. Knoth K, Roberds S, Poteet C, & Tamkun M (1988) Highly degenerate, inosine-containing primers specifically amplify rare cDNA using the polymerase chain reaction. Nucleic Acids Research 16(22):10932. Batzer MA, Carlton JE, & Deininger PL (1991) Enhanced evolutionary PCR using oligonucleotides with inosine at the 3-terminus. Nucleic Acids Research 19(18):5081. Patil RV & Dekker EE (1991) PCR amplification of an Escherichia coli gene using mixed primers containing deoxyinosine at ambiguous positions in degenerate amino acid codons. Nucleic Acids Research 19(11):3184. Pomp D & MEDRANO JF (1991) Organic Solvents as Facilitators of Polymerase Chain Reaction (Eaton, Natick, MA, ETATS-UNIS) (0736-6205) p 2. Roux KH (1995) Optimization and troubleshooting in PCR. Genome Research 4(5):S185S194. Chi Z, Su CD, & Lu WD (2007) A new exopolysaccharide produced by marine Cyanothece sp. 113. Bioresource Technology 98(6):1329-1332. De Philippis R, Margheri MC, Materassi R, & Vincenzini M (1998) Potential of unicellular cyanobacteria from saline environments as exopolysaccharide producers. Appl Environ Microbiol 64(3):1130-1132.

144