Host specicity in the Richeliadiatom symbiosis

revealed by hetR gene sequence analysis

S. Janson,
1
*
J. Wouters,
2
B. Bergman
2
and
E. J. Carpenter
3
1
Department of Marine Sciences, Institute of Natural
Sciences, University of Kalmar, Box 905, S-391 29
Kalmar, Sweden.
2
Department of Botany, Stockholm University, S-106 91
Stockholm, Sweden.
3
Marine Sciences Research Center, State University of
New York at Stony Brook, NY 11794-5000, USA.
Summary
The lamentous heterocyst-forming cyanobacterium
Richelia intracellularis forms associations with dia-
toms and is very abundant in tropical and subtropical
seas. The genus Richelia contains only one species,
R. intracellularis Schmidt, although it forms associa-
tions with several diatom genera and has considerable
variation in size and morphology. The genetic diver-
sity, and possible host specicity, within the genus
Richelia is unknown. Using primers against hetR, a
gene unique for lamentous cyanobacteria, specic
polymerase chain reaction (PCR) products were
obtained from natural populations of R. intracellularis
laments associated with three diatom genera. Phylo-
genetic analyses of these sequences showed that
they were all in the same clade. This clade contained
only the R. intracellularis sequences. The genetic
afliation of hetR sequences of R. intracellularis to
those of other heterocystous cyanobacteria strongly
suggests that it was not closely related to endosym-
biotic Nostoc spp. hetR sequences. Sequences from
R. intracellularisHemiaulus membranaceus sampled
in the Atlantic and Pacic Oceans were almost identi-
cal, demonstrating that the genetic relatedness was
not dependent on geographical location. All sequences
displayed a deep divergence between symbionts from
different genera and a high degree of host specicity.
Introduction
Cyanobacteria occur in many different environments and
in association with a wide variety of organisms. Richelia
intracellularis Schmidt is the most widespread heterocyst-
forming cyanobacterium in marine pelagic environments.
The nitrogen xation associated with R. intracellularis
aggregates was rst reported by Mague et al. (1974) and,
together with members of the non-heterocystous cyano-
bacterial genus Trichodesmium, is believed to be of major
importance in the nitrogen budgets of tropical oceans
(Venrick, 1974; Carpenter and Romans, 1991; Capone
et al., 1997). The genus contains one species, R. intracel-
lularis (Geitler, 1932), and is mainly found as an endophyte
in diatoms in the genera Hemiaulus and Rhizosolenia.
Since its discovery inside the diatom Rh. styliformis col-
lected in the Red Sea (Ostenfeld and Schmidt, 1901),
the common occurrence and wide distribution of R. intra-
cellularis have been conrmed repeatedly. Lemmermann
(1905) reported R. intracellularis to be present in the gen-
era Rhizosolenia and Hemiaulus collected in the Pacic
Ocean, and also described an epiphyte attached to Chae-
toceros, named Calothrix rhizosoleniae. Karsten (1907)
observed R. intracellularis inside Rhizosolenia spp. and
epiphytically attached to Chaetoceros spp. in the Indian
Ocean. The underlying genetic basis for the diversity
of R. intracellularis, occurring extra- or intracellularly and
associated with different diatom genera, has not been
examined.
The hetR gene has only been detected in lamentous
cyanobacteria (Buikema and Haselkorn, 1991; Janson et
al., 1998), and its function has been assigned to hetero-
cyst and akinete differentiation (Buikema and Haselkorn,
1991; Legane s et al., 1994). The protein encoded by
hetR appears to be a serine type of protease (Zhou et al.,
1998). The unique distribution of hetR among lamentous
cyanobacteria and its relatively high sequence variation
between closely related strains makes hetR a powerful
genetic marker for lamentous cyanobacteria. A recent
study showed that the phylogeny of hetR within Tricho-
desmium spp. is congruent with that of 16S rRNA gene
sequences (S. Janson, unpublished).
Richelia intracellularis laments inhabiting diatoms of
the genera Rhizosolenia and Hemiaulus, as well as
attached to the outside of the diatom Chaetoceros sp.,
were collected during two research cruises, one in the
Caribbean Sea and one in the South Pacic Ocean. In
order to study the genetic variation and host specicity
of R. intracellularis, the nucleotide sequence of partial
hetR genes was determined from the different diatom
associations observed.
Environmental Microbiology (1999) 1(5), 431438
Q 1999 Blackwell Science Ltd
Received 14 April, 1999; revised 25 May, 1999; accepted 31 May,
1999. *For correspondence. E-mail Sven.Janson@ng.hik.se; Tel.
(46) 0480 447310; Fax (46) 0480 447305.
Results and discussion
Observations with microscopy
Epiuorescence microscopy revealed that nearly all Hemi-
aulus spp. contained two short endophytic Richelia intra-
cellularis (the microsymbiont) laments, each with one
terminal heterocyst, as described previously (Kimor et al.,
1978; 1992; Heinbokel, 1986; Villareal, 1994). The Rhizo-
solenia clevei var. communis (sensu Sundstro m, 1984)
diatoms each contained 24 endophytic R. intracellularis
laments, each consisting of one terminal heterocyst and
410 vegetative cells (Fig. 1). Filaments that were mor-
phologically similar to R. intracellularis, in the sense that
they were short, with 410 vegetative cells and with a single
terminal heterocyst, were observed attached to the out-
side of the chain-forming diatom Chaetoceros sp. (Fig. 1)
but never inside the cells. The heterocysts of the laments
attached to Chaetoceros sp. were of smaller diameter than
the adjacent vegetative cells, in contrast to the endophytic
R. intracellularis, and the autouorescence of the hetero-
cysts was less intense (Fig. 1). The laments were also
slightly tapered. The R. intracellularis laments seen on
Chaetoceros sp. matched that described by Lemmermann
(1905), who also noted differences in cell diameter and
pigmentation of the heterocyst in extracellular microsym-
bionts compared with those inside diatoms.
Life history and symbiont specicity of the symbiosis
The hetR gene sequences obtained from two species of
Hemiaulus were consistently divided over two lineages
with very high statistical support (Figs 25). The sequence
obtained from samples of H. membranaceus collected in
the Caribbean Sea (Atlantic Ocean) was most closely
related to the sequence obtained from H. membranaceus
from the South Pacic Ocean. The sequences from the
Q 1999 Blackwell Science Ltd, Environmental Microbiology, 1, 431438
Fig. 1. Epiuorescent micrographs of Richelia intracellularis collected in the South Pacic Ocean.
A and B. R. intracellularis laments attached to the outside of the chain-forming diatom Chaetoceros sp. Note the smaller heterocysts (arrows)
at one end of the yellow autouorescent R. intracellularis laments, relative to the vegetative cells. The laments are also slightly tapered. Red
uorescent chloroplasts are seen within the Chaetoceros sp. cells. In (B), incident illumination allows the contours of the Chaetoceros to be
visible. Dark areas in the background are the pores of the Poretics lter used to concentrate the sample.
C. One R. intracellularis lament is seen inside Rhizosolenia clevei var. communis. Note the larger heterocyst, relative to the vegetative cells,
at the end of the R. intracellularis lament. The scale bar represents 50mm.
432 S. Janson, J. Wouters, B. Bergman and E. J. Carpenter
two samples of H. hauckii, both collected at the same
station, were most closely related to each other. The two
samples of Rh. clevei var. communis were from the same
station as one of the H. membranaceus samples. Still, the
two sequences from Rh. clevei var. communis were phylo-
genetically closest to each other, implying that the genetic
diversity of microsymbionts does not depend on the geo-
graphical location. The sequence identity, as illustrated
in Table 1, between hetR sequences from all R. intracellu-
laris samples was 85% or higher. Within the same diatom
species, the variation was very low, 1% or less. The sequ-
ence similarities within the same diatomgenus, Hemiaulus,
were <96% for the microsymbionts from the two species.
In hetR sequences from the non-heterocystous planktonic
cyanobacteria Trichodesmium spp., two closely related
species scored 96.7%, while the similarity within one
species was 99.6% (S. Janson, unpublished). Assuming
an equal rate of nucleotide substitution for R. intracellularis
and Trichodesmium spp., which both inhabit oligotrophic
waters, then the two Hemiaulus species each contain a dif-
ferent microsymbiont species. The two sequences from
epiphytes of Chaetoceros sp. were less informative in this
respect, because the host identity could not be determined
beyond the genus level. However, they were most closely
related to sequences from Rh. clevei var. communis and,
in most cases, they appeared as two sister groups to
these sequences.
This suggests that R. intracellularis laments bearing
hetR genes of different lineages are host specic in their
infection, or that the microsymbiont is transferred vertically
during host cell division with reinfection being redundant.
In the RhizosoleniaRichelia symbiosis, the laments of
Q 1999 Blackwell Science Ltd, Environmental Microbiology, 1, 431438
Fig. 2. Phylogenetic tree of hetR nucleotide sequences from
natural populations of Richelia intracellularis. The sequence
denoted (a) was from the Atlantic Ocean; all other R. intracellularis
sequences were from the southern Pacic Ocean. Maximum
likelihood was used as a tree-building method with the transition/
transversion ratio set to the observed value of 2.0. The likelihood
value of the tree was ln(L) 1545.804. The hetR sequence of
Trichodesmium sp. IMS 101 was used as an outgroup. The
numbers at the nodes are the percentage of 500 bootstrap
replicates, showing values over 50 only. The scale indicates the
branch length at which 10% of the sequence has changed.
Abbreviations as in Table 2.
Fig. 3. Phylogenetic tree inferred from hetR nucleotide sequences from a variety of cyanobacteria. The sequence denoted (a) was from the
Atlantic Ocean; all other R. intracellularis sequences were from the southern Pacic Ocean. The resulting neighbour-joining tree from the
unweighed distance matrix obtained using the one-parameter model by Jukes and Cantor (1969) is shown. The numbers at the nodes are the
percentage of 500 bootstrap replicates; numbers within brackets are the corresponding value for maximum parsimony analysis of the data.
Values below 50 are not shown. The scale bar represents the change in nucleotide sequence of 10%. Abbreviations as in Table 2.
Host specicity in the Richeliadiatom symbiosis 433
R. intracellularis are divided in the middle, and laments
are carried over to the other end of the host cell by force
of cytoplasmic streaming, before the host cell completes
its division cycle (Taylor, 1982). In a laboratory culture of
the RhizosoleniaRichelia symbiosis, fast-growing host
cells that nished the formation of the cell wall septa before
the microsymbiont had been transported to the daughter
cell and the diatom cells that apparently lost their micro-
symbiont were unable to grow in nitrogen-decient media
(Villareal, 1989; 1990). The reinfection process has never
been observed; vertical transfer is therefore likely to be
the preferred route of microsymbiont acquisition.
In most cases, symbiotic host plants are continuously
reinfected, e.g. in the GunneraNostoc symbiosis (Berg-
man et al., 1992). The microsymbiont in such plant
microbe symbioses are `broad specic' in their associations,
and co-evolution of the partners does not seem to occur to
any great extent (Doyle, 1998). In laboratory experiments,
Q 1999 Blackwell Science Ltd, Environmental Microbiology, 1, 431438
Fig. 4. Phylogenetic tree inferred from hetR sequences using quartet puzzling. The sequence denoted (a) was from the Atlantic Ocean; all
other R. intracellularis sequences were from the southern Pacic Ocean. The substitution model was that of Scho niger and Haeseler (1994).
The quartet puzzling was resampled 10 000 times, and the percentage of the number of times that each conguration occurred in all trees is
shown at the nodes. The likelihood of the tree was ln(L) 4199.93. Note the high resolution within the R. intracellularis sequences, while the
relationship to the sequences from other heterocystous cyanobacteria is unresolved. Abbreviations as in Table 2.
Fig. 5. Phylogenetic tree inferred from hetR
amino acid sequences. The sequence denoted
(a) was from the Atlantic Ocean; all other R.
intracellularis sequences were from the southern
Pacic Ocean. The distances were calculated for
the translated amino acid sequences using the
PAM correction (Dayhoff et al., 1978). The
numbers at the nodes are the percentage of 500
bootstrap replicates, not showing values below
50. The values in brackets are derived from
maximum parsimony analysis. The scale bar
represents the change in nucleotide sequence of
10%. Abbreviations as in Table 2.
434 S. Janson, J. Wouters, B. Bergman and E. J. Carpenter
Nostoc sp. strains isolated from a variety of plant groups
and genetically diverse can infect the same Gunnera sp.
strain (Bonnet and Silvester, 1981; Johansson and Berg-
man, 1994; Rasmussen and Svenning, 1998). The two
hetR sequences from endosymbiotic Nostoc sp. strains
used here can infect the same host species. These sequ-
ences were 90% similar, much lower than two R. intra-
cellularis sequences from the same species (Table 1).
The parallel divergence of host and symbionts (Figs
25) may reect a high degree of mutual benets of the
association. For instance, the retention of the nitrogen-
xing microsymbiont could be advantageous in tropical
oceans where sources of combined nitrogen are scarce.
The infrequent observations of epiphytic and free-living
R. intracellularis suggest that the microsymbiont is not
so competitive in its free-living state, where it might not
be able to regulate its position in the water column. Indeed,
the ultrastructure of the microsymbiont of Rh. clevei
var. clevei showed that gas vesicles were absent (Janson
et al., 1995).
Relatedness to other cyanobacterial species
In order to examine the phylogenetic afliations of hetR
sequences from R. intracellularis with those from other
cyanobacteria, hetR sequences from heterocystous cya-
nobacteria of various morphology were determined.
Sequences from heterocystous cyanobacteria constituted
a separate lineage in all analyses (see Figs 35). Among
these were two strains that, like R. intracellularis, also
live endosymbiotically. One was isolated from the African
terrestrial angiosperm Gunnera perpensa, Nostoc sp. GP
9401. The hetR sequence from the other endosymbiotic
isolate, Nostoc sp. PCC 9229 from Gunnera monoica,
was already known (A. Matveyev, F. Lotti and B. Bergman,
unpublished). The hetR sequences from these strains
were related more to each other than to any other sequence
used in the analysis, and they appear to be distantly
related to the R. intracellularis hetR sequences. Nodularia
sp. BC 9408, an isolate fromthe Baltic Sea, is a heterocys-
tous cyanobacterial phytoplankton that forms laments that
are several hundred cells long. The hetR sequence from
Nodularia sp. BC 9408 clustered with the sequences
from the two endosymbiotic Nostoc sp. strains (see Figs
35). One morphological similarity between Nodularia
spp. and Nostoc spp. is the production of multiple hetero-
cysts and akinetes; otherwise, they exhibit different morpho-
logical and ecological properties. Strains of Nodularia and
symbiotic Nostoc other than those used in this study are in
the same clade in 16S rRNA-based trees (Turner, 1997),
indicating that the hetR-based tree is not merely reecting
the similarities in heterocyst and akinete patterns. The hetR
sequence from Calothrix sp. PCC 7103 clustered with the
R. intracellularis sequences when using the nucleotide
sequence with both distance and parsimony analyses.
The habitat of this strain is not known, but both Calothrix
spp. and R. intracellularis have a polar lament with the
heterocyst attached at one end. The maximum likelihood
analysis, as implemented by quartet puzzling, strongly
indicates that the data represent a star phylogeny at the
node of all heterocystous sequences, and the Calothrix
sp. PCC 7103 sequence was not grouped together with
R. intracellularis sequences (see Fig. 4). Fischerella sp.
PCC 7521 displays non-polarity, true branching of the la-
ments, and heterocysts are developed at multiple locations
on the lament. When analysing the translated amino acid
sequences of hetR, the sequence from Fischerella sp.
PCC 7521 was a sister group to the R. intracellularis
sequences, with a high bootstrap support, and the Calothrix
sp. PCC7103 sequence was ancestral to these sequences,
in both distance and parsimony analysis (see Fig. 5). The
Fischerella sp. PCC7521 and R. intracellularis, but not the
Calothrix PCC 7103, hetR amino acid sequences were in
the same clade in a quartet puzzling analysis (data not
shown).
The conicting results obtained using the nucleotide ver-
sus the translated amino acid sequence of hetR and using
different analytical methods can be explained by the lack
of a close relationship between R. intracellularis sequences
and extant hetRsequences. However, when the third codon
position was removed from the nucleotide sequences
and analysed with maximum parsimony, all the resulting
trees were identical to the amino acid sequence tree with
respect to the Calothrix sp. PCC 7103/Fischerella sp.
Q 1999 Blackwell Science Ltd, Environmental Microbiology, 1, 431438
Table 1. Comparison between nucleotides of
hetR gene fragments obtained from Richelia
intracellularis laments associated with four
diatom genera.
Sequence
1 2 3 4 5 6 7 8
1 RHH9804-4
2 RHH9804-5 99.3
3 RHM9565 95.8 96.2
4 RHM9847 95.5 96.0 98.9
5 RCH9811 85.0 85.5 86.2 85.9
6 RCH9814 85.7 86.2 86.5 86.2 92.4
7 RRHC9847-2 85.9 86.4 87.1 86.6 92.0 96.9
8 RRHC9847-3 86.4 86.8 87.5 87.1 92.0 96.4 99.6
The names of the sequences are according to Table 1.
Host specicity in the Richeliadiatom symbiosis 435
PCC 7521/R. intracellularis clade. Although the bootstrap
support for this constellation was low, it indicates that the
mutations in the third codon position, to some extent, were
saturated when comparing sequences outside the R. intra-
cellularis clade. If hetR sequences from a closer relative of
R. intracellularis are obtained, the resolution in this region
of the tree could be improved. While the genetic afliation
of R. intracellularis to other cyanobacteria is still not clear,
our data suggest that it is not closely related to endosym-
biotic Nostoc sp. strains.
Concluding remarks
Richelia intracellularis laments have been observed both
endo- and epiphytically on certain diatoms by taxonomists,
but diagnostic genetics have not been applied until now
to investigate the diversity of symbiotic cyanobacteria in
open oceans. The phylogenetic analyses used on our
data suggest that all R. intracellularis laments belong to
a monophyletic group, distant from other cyanobacterial
endosymbionts. Moreover, the high degree of host speci-
city between diatom hosts and R. intracellularis is unusual
in plantmicrobe symbiosis in which the symbiont is trans-
ferred horizontally, suggesting that the microsymbiont is
transferred vertically. The data also imply that R. intracellu-
laris comprises several species, but this should be con-
rmed by further genetic analysis.
Experimental procedures
Collection
Richeliadiatom samples were collected during a cruise in
the Bahamas and northern Caribbean Sea in January 1995
on the R/V Seward Johnson (Fort Pierce, FL, USA). Those
from the South Pacic Ocean were obtained from waters off
Fiji during a research cruise in MarchApril 1998 on the R/V
Roger Revelle (San Diego, CA, USA). The Hemiaulus mem-
branaceus sample (RHM9665) was collected with the aid
of a microcapillary system (Guillard and Keller, 1984) and
sequentially transferred to ve droplets of ltered sea water.
Several samples were pooled together to yield about 50 diatom
cells, each containing no more than two laments of R. intra-
cellularis, and the DNA was extracted as described previously
(Janson et al., 1998). The samples from the South Pacic
Ocean were either picked with a micropipette from 1% (w/v)
agarose/sample mixtures or cut out from transparent lters
under a dissecting microscope. The agarose or lter pieces
were placed on a clean microscope slide and examined with
an epiuorescence microscope to ensure that R. intracellu-
laris laments were present and that no other cyanobacteria
were contaminating the sample. The two Chaetoceros sp.
samples (RCH9811 and RCH9814) contained about 10 R.
intracellularis laments each; the two H. hauckii samples
(RHH9804-4 and RHH9804-5) contained six and ve diatom
cells with two R. intracellularis laments in each cell; the H.
membranaceus sample (RHM9847) contained 10 diatom cells
with two R. intracellularis laments in each cell; the two Rhizo-
solenia clevei var. communis samples (RRHC9847-2 and
RRHC9847-3) contained four R. intracellularis laments
each. The samples were then placed in a 0.5 ml PCR tube
and frozen at 208C for at least 2h before PCR amplication.
PCR and cloning
The 448 bp fragment of hetR, corresponding to number 159
606 in Nostoc PCC 7120 (Buikema and Haselkorn, 1991),
was amplied from the H. membranaceus sample from the
Atlantic Ocean with primers hetr1 and hetr2 (Janson et al.,
1998). PCR buffer was added to all other samples without
DNA extraction. A mixture of Taq and Pwo DNA polymerase
(High-delity PCR; Boehringer) was added after the initial
denaturation step of 5min at 948C followed by 45 cycles with
Q 1999 Blackwell Science Ltd, Environmental Microbiology, 1, 431438
Table 2. The hetR gene sequences determined in this study
Sequence Accession
Taxa Habitat and origin name number
Calothrix sp. PCC
a
7103 Uncertain C7103 AF135804
Calothrix sp. PCC
a
7507 Free-living, Sphagnum bog, Vierwaldsta ttersee, Switzerland C7507 AF135805
Fischerella sp. PCC
a
7521 Free-living, hot spring, Yellow Stone, WY, USA F7521 AF135806
Nodularia sp. BC
b
9408 Free-living plankton, Baltic Sea N9408 AF135807
Nostoc sp. GP
c
9401 Endophytic in Gunnera perpensa, Mbeya, Tanzania, Africa NC9401 AF135808
R. intracellularisChaetoceros sp. Epiphytic plankton, Pacific Ocean (228318S, 1668118E) RCH9811 AF135809
R. intracellularis Chaetoceros sp. Epiphytic plankton, Pacific Ocean (158108S, 1638308E) RCH9814 AF135810
R. intracellularis Hemiaulus hauckii Endophytic plankton, Pacific Ocean (328318S, 1708288E) RHH9804-4 AF135811
R. intracellularis H. hauckii Endophytic plankton, Pacific Ocean (328318S, 1708288E) RHH9804-5 AF135812
R. intracellularis H. membranaceus Endophytic plankton, Atlantic Ocean (238558N, 758388W) RHM9565 AF135813
R. intracellularis H. membranaceus Endophytic plankton, Pacific Ocean (18808S, 178808W) RHM9847 AF135814
R. intracellularis Rhizosolenia clevei var. communis Endophytic plankton, Pacific Ocean (18808S, 178808W) RRHC98472 AF135815
R. intracellularis Rh. clevei var. communis Endophytic plankton, Pacific Ocean (18808S, 178808W) RRHC9847-3 AF135816
The name of the strain or host that the sequences were retrieved from is given in the first column. The second column indicates the habitat and
where it was sampled. The third and fourth columns give the relevant name for the sequence and the GenBank accession number respectively.
For further details, see text
a. PCC, Pasteur Culture Collection.
b. BC, Baltic Sea cruise.
c. GP, Gunnera perpensa.
436 S. Janson, J. Wouters, B. Bergman and E. J. Carpenter
1 min at 458C, 1.5 min at 728C and 1 min at 948C. The PCR
buffer provided by the supplier of the enzymes was supple-
mented with 2.5% (v/v) nal concentration of DMSO in order
to lower the annealing temperature and improve the permea-
bilization of the cells. All clones were derived from single PCR
reactions and were cloned according to the instructions pro-
vided with a TA cloning kit (Invitrogen).
Sequencing and sequence analyses
Plasmids used in sequencing were prepared with Qiaprep
spin miniprep (Qiagen). The inserts were sequenced bidirec-
tionally using an ABI model 373 An automated sequencer
(Applied Biosystems) and dye terminator chemistry, using pri-
mers M13F and M13Rfromthe cloning vector. Two sequences
from the same plasmid library never showed more variation
than could be accounted for by Taq or Pfu polymerase errors
(Wintzingerode et al., 1997). Consequently, the sequences
reported here are the consensus of each analysed library.
The sequences were all of equal length, and the alignment
was therefore performed manually using the SEAVIEW software
(Galtier et al., 1996). Trees with bootstrap analysis of the data
were inferred using PHYLOWIN (Galtier et al., 1996) operating on
a Linux i386 platform. For quartet puzzling, the program PUZZLE
was used (Strimmer and Haeseler, 1996). The program TREE-
VIEW (Page, 1996) was used to examine and edit tree les. The
GenBank accession numbers for the sequences determined
in this study and those used for phylogenetic analyses are
listed in Tables 2 and 3.
Acknowledgements
We thank the captain and crew on R/V Roger Revelle.
Research was partially supported by a grant to E.J.C. from
the US NSF, by a grant to B.B. from the Swedish Foundation
for International Co-operation in Research and Higher Educa-
tion (STINT) and to S.J. from the Swedish Natural Science
Research Council (NFR).
References
Bergman, B., Johansson, C., and Soderback, E. (1992) Tans-
ley review no. 42. The NostocGunnera symbiosis. New
Phytol 122: 379400.
Bonnet, H.T., and Silvester, W.B. (1981) Specicity in
the GunneraNostoc endosymbiosis. New Phytol 89:
121128.
Buikema, W.J., and Haselkorn, R. (1991) Characterization of
a gene controlling heterocyst differentiation in the cyano-
bacterium Anabaena 7120. Genes Dev 5: 321330.
Capone, D.G., Zehr, J.P., Paerl, H.W., Bergman, B., and
Carpenter, E.J. (1997) Trichodesmium, a globally signi-
cant marine cyanobacterium. Science 276: 12211229.
Carpenter, E.J., and Romans, K. (1991) Major role of the
cyanobacterium Trichodesmium in nutrient cycling in the
North Atlantic Ocean. Science 254: 13561358.
Dayhoff, M.O., Schwartz, R.M., and Orcutt, B.C. (1978) A
model of evolutionary change in proteins. In Atlas of Protein
Sequences and Structure, Vol. 5. Dayhoff, M.O. (ed.). Silver
Springs: Natural Biomedicine Research Foundation, pp.
345352.
Doyle, J.J. (1998) Phylogenetic perspectives on nodulation:
evolving views of plants and symbiotic bacteria. Trends
Plant Sci 3: 473478.
Galtier, N., Gouy, M., and Gautier, C. (1996) SEAVIEW and
PHYLOWIN, two graphic tools for sequence alignment and
molecular phylogeny. Comput Appl Biosci 12: 543548.
Geitler, L. (1932) Cyanophyceae. In Rabenhorst's Kryptoga-
menora von Deutschland, O

sterreich und der Schweiz,

Vol. 14. Kolkwitz, R. (ed.). Leipzig: Akademische-Verlags-
gesellschaft, pp. 804805.
Guillard, R.R.L., and Keller, M.D. (1984) Culturing dinoagel-
lates. In Dinoagellates. Spector, D. (ed.). Orlando: Aca-
demic Press, pp. 391442.
Heinbokel, J.F. (1986) Occurrence of Richelia intracellularis
(Cyanophyta) within the diatoms Hemiaulus hauckii and
H. membranaceus off Hawaii. J Phycol 22: 399403.
Janson, S., Rai, A.N., and Bergman, B. (1995) The intracel-
lular cyanobiont Richelia intracellularis : ultrastructure and
immuno-localisation of phycoerythrin, nitrogenase, Rubisco
and glutamine synthetase. Mar Biol 124: 18.
Janson, S., Matveyev, A., and Bergman, B. (1998) The pre-
sence and expression of hetR in the non-heterocystous
cyanobacterium Symploca PCC 8002. FEMS Microbiol
Lett 168: 173179.
Johansson, C., and Bergman, B. (1994) Reconstitution of the
Gunnera manicata Linden symbiosis: cyanobacterial spe-
cicity. New Phytol 126: 643652.
Q 1999 Blackwell Science Ltd, Environmental Microbiology, 1, 431438
Table 3. The hetR gene sequences used in phylogenetic analyses that were not determined in this study.
Taxa Habitat and origin Reference Accession number
Leptolyngbya sp. PCC
a
73110 Uncertain Janson et al . (1998) AF013036
Nostoc sp. PCC
a
7120 Uncertain Buikema and Haselkorn (1991) M37779
Nostoc sp. PCC
a
9229 Endophytic in Gunnera spp. A. Matveyev, unpublished X92989
Symploca sp. PCC
a
8002 Free-living, mud flat Janson et al . (1998) AF013035
Trichodesmium sp. IMS
b
101 Free-living plankton, Atlantic Ocean S. Janson, unpublished AF091323
T. contortum Free-living plankton, Atlantic Ocean S. Janson, unpublished AF013031
T. erythraeum Free-living plankton, Atlantic Ocean S. Janson, unpublished AF013034
T. tenue Free-living plankton, Atlantic Ocean S. Janson, unpublished AF013033
T. thiebautii Free-living plankton, Pacific Ocean S. Janson, unpublished AF013032
The name of the strain that the sequences were retrieved from is given in the first column. The second column indicates the habitat and where it
was sampled. The third and fourth columns give the relevant reference for the sequence and the GenBank accession number respectively. For
further details, see text
a. PCC, Pasteur Culture Collection.
b. IMS, culture collection held at the Institute of Marine Science, University of North Carolina, Morehead City, NC, USA.
Host specicity in the Richeliadiatom symbiosis 437
Jukes, T.H., and Cantor, C.R. (1969) Evolution of protein
molecules. In Mammalian Protein Metabolism, Vol. III.
Muntu, H.N. (ed.). New York: Academic Press, pp. 21
132.
Karsten, G. (1907) Das Indische Phytoplankton nach dem
Material der Deutschen Tiefsee-Expedition 18981899.
Dtsch Tiefsee-Exped 18981899 2: 423548.
Kimor, B., Gordon, N., and Neori, A. (1992) Symbiotic asso-
ciations among the microplankton in oligotrophic marine
environments, with special reference to the Gulf of Aqaba,
Red Sea. J Plankton Res 14: 12171231.
Kimor, B., Reid, F.M., and Jordan, J.B. (1978) An unusual
occurrence of Hemiaulus membranaceus Cleve (Bacillario-
phyceae) with Richelia intracellularis Schmidt (Cyanophy-
ceae) off the coast of Southern California in October
1976. Phycologia 17: 162166.
Legane s, F., Ferna ndez-Pin as, F., and Wolk, C.P. (1994)
Two mutations that block heterocyst differentiation have
different effects on akinete differentiation in Nostoc ellipso-
sporum. Mol Microbiol 12: 679684.
Lemmermann, E. (1905) Die Algenora der Sandwich-Inseln.
Ergebnisse einer Reise nach dem Pacic, H. Schauinsland
1896/97. Engler's Bot Jb 34: 607663.
Mague, T.H., Weare, M.M., and Holm-Hansen, O. (1974)
Nitrogen xation in the north Pacic Ocean. Mar Biol 24:
109119.
Ostenfeld, C.H., and Schmidt, J. (1901) Plankton fra det
Rde hav og Adenbugten. Copenhagen: Vidensk Meddel
Naturh Forening i Kbhvn, pp. 141182.
Page, R.D.M. (1996). TREEVIEW: an application to display phy-
logenetic trees on personal computers. Comp Appl Biosci
12: 357358.
Rasmussen, U., and Svenning, M.M. (1998) Fingerprinting of
cyanobacteria based on PCR with primers derived from
short and long tandemly repeated repetitive sequences.
Appl Environ Microbiol 64: 265272.
Scho niger, M., and von Haeseler, A. (1994) A stochastic
model for the evolution of autocorrelated DNA sequences.
Mol Phyl Evol 3: 240247.
Strimmer, K., and von Haeseler, A. (1996) Quartet puzzling:
a quartet maximum likelihood method for reconstructing
tree topologies. Mol Biol Evol 13: 964969.
Sundstro m, B.G. (1984) Observations on Rhizosolenia clevei
Ostenfeld (Bacillariophyceae) and Richelia intracellularis
Schmidt (Cyanophyceae). Bot Mar 27: 345355.
Taylor, F.J.R. (1982) Symbioses in marine microplankton.
Ann Inst Oce anogr Paris Suppl 58: 6190.
Turner, S. (1997) Molecular systematics of oxygenic photo-
synthetic bacteria. Pl Syst Evol Suppl 11: 1352.
Venrick, E.L. (1974) The distribution and signicance of
Richelia intracellularis Schmidt in the North Pacic Central
Gyre. Limnol Oceanogr 19: 437445.
Villareal, T.A. (1989) Division cycles in the nitrogen-xing
Rhizosolenia (Bacillariophyceae)Richelia (Nostocaceae)
symbiosis. Br Phycol J 24: 357365.
Villareal, T.A. (1990) Laboratory culture and preliminary
characterization of the nitrogen-xing RhizosoleniaRichelia
symbiosis. Mar Ecol 11: 117132.
Villareal, T.A. (1994) Widespread occurrence of the Hemiau-
luscyanobacterial symbiosis in the Southwest North
Atlantic Ocean. Bull Mar Sci 54: 17.
Wintzingerode, F., Go bel, U.B., and Stackebrandt, E. (1997)
Determination of microbial diversity in environmental sam-
ples: pitfalls of PCR-based rRNA analysis. FEMS Microbiol
Rev 21: 213229.
Zhou, R., Wei, X., Jiang, N., Li, H., Dong, Y., Hsi, K.-L., et al.
(1998) Evidence that HetR protein is an unusual serine-
type protease. Proc Natl Acad Sci USA 95: 49594963.
Q 1999 Blackwell Science Ltd, Environmental Microbiology, 1, 431438
438 S. Janson, J. Wouters, B. Bergman and E. J. Carpenter