Bioresource Technology 98 (2007) 281287

Coconut water as a medium additive for the production of docosahexaenoic acid (C22:6 n3) by Schizochytrium mangrovei Sk-02
Panida Unagul a,b, Caetharin Assantachai c, Saranya Phadungruengluij c, Manop Suphantharika c, Morakot Tanticharoen a,b, Cornelis Verduyn c,
b a BIOTEC, 113 Thailand Science Park, Phaholyothin Rd, Klong 1, Klong Luang, Pathum Thani 12120, Thailand School of Bioresources and Technology, King Mongkuts University of Technology Thonburi (Bangkhuntien), 83 Moo 8 Tientalay 25 Road, Thakham, Bangkhuntien, Bangkok 10150, Thailand c Faculty of Science, Department of Biotechnology, Mahidol University, Rama 6 Rd, Bangkok 10400, Thailand

Received 16 September 2005; received in revised form 10 January 2006; accepted 17 January 2006 Available online 23 March 2006

Abstract The eVect of coconut water (CW) on biomass and docosahexaenoic acid (DHA, C22:6 n3) formation by Schizochytrium mangrovei Sk-02 was studied in a yeast extract-diluted sea water medium. Optimal CW-level was ca. 33% (v/v), resulting in a biomass level of 28 g/l with a DHA-content of 20% (w/w) or 6 g DHA/l, almost 50% higher than in non-supplemented cultures at the same initial sugar level. Study on the growth-promoting eVects of coconut water suggested that it could be (partially) mimicked by addition of trace elements; the fatty acids present in CW did not appear to be incorporated or eVect fatty acid formation by the organism. CW-addition was also eVective in media with other nitrogen sources such as casitone, peptone and tryptone. Its inclusion (at 50% v/v) increased biomass levels two-to-three-fold with concomitant increases in the DHA-level. 2006 Elsevier Ltd. All rights reserved.
Keywords: Coconut water; Docosahexaenoic acid; Growth promoter; Schizochytrium; Yield

1. Introduction Docosahexaenoic acid (DHA) is an essential polyunsaturated fatty acid (PUFA) and insuYcient dietary intake has been implicated in a variety of human diseases (Horrocks and Yeo, 1999). The traditional source of DHA is Wsh oil but declining Wsh stocks, seasonal variability in oil composition, oVensive odor and potential chemical contamination have stimulated research into alternative sources, in particular production by heterotrophic marine protists as Cryptheconidium cohnii and thraustochytrids, including Schizochytrium and Ulkenia sp. (Ratledge, 2004; Sijtsma and de Swaaf, 2004; Ward and Singh, 2005). These organisms can be grown in complex media with carbohydrates as
*

carbon and energy source. Yeast extract is a suitable nitrogen source for these organisms and has the beneWt of signiWcant vitamin levels. The aim of this study was to evaluate the eVect of coconut water (CW) to improve DHA yields in a glucoseyeast extract-(diluted) sea water medium. 2. Methods 2.1. Microorganism and cultivation Schizochytrium mangrovei sp. Sk-02 isolated in a mangrove forest was used in this study (Fan et al., 2001). Pure cultures were maintained on slants in solid GPY-medium containing 15 g/l each agar (Difco, USA) and artiWcial sea salts (ASS) (Sigma, USA) at 20 C and sub-cultured every month. To prepare an inoculum, cells from a slant were streaked on a plate with the medium described above and

Corresponding author. Tel.: +66 2 02201 5318; fax: +66 2 246 3026. E-mail address: frcvd@mahidol.ac.th (C. Verduyn).

0960-8524/$ - see front matter 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2006.01.013

282

P. Unagul et al. / Bioresource Technology 98 (2007) 281287

incubated at 25 C for 24 h. Two circular wells ( 1 cm) were then aseptically cut in the agar with a large cork-borer and the plate was Xushed with 10 ml of a 15 g/l solution of ASS. After two hours, zoospores were collected from the wells into a sterile tube, the OD660 was assayed and diluted to 7.0 by adding 15 g/l sterilized ASS. A 5% (v/v) inoculum was then transferred to a 250 ml Erlenmeyer containing 100 ml of standard medium containing (g/l) yeast extract (Difco, USA) 10, glucose 60 and ASS 15. Where indicated, the medium was dissolved in (diluted) mature coconut water (CW) obtained from a local manufacturer of canned coconut meat. In order to remove any larger particles, CW was Wrst passed through a cheese cloth. CW contained monosaccharides (a total of 11 g/l fructose and glucose, see results), but practically it was diYcult to adjust the total initial monosaccharides level to 60 g/l (as in the standard medium) for media with various levels of diluted CW, hence this was not pursued. Rather, controls were performed in which the glucose level in the standard medium was increased to 71 g/l by supplementation with the level of glucose and fructose found in undiluted CW. In addition, in an experiment where medium components were dissolved in undiluted CW, the amount of glucose added was reduced to 49 g/l to give an initial monosaccharides level of 60 g/l as in the standard medium. To obtain fatty acid-free coconut water, a freeze-dried aliquot (25 ml, yielding 1 g dry material) was extracted twice with 10 ml hexane. The hexane layers were combined and evaporated with a gentle stream of pure nitrogen gas at 40 C. FAs were then assayed as described below. A non-extracted freeze-dried sample was used as a control and gave results identical to non-frozen coconut water. In order to identify possible growth-promoting factors in coconut water, the latter was replaced by addition of (combinations of) complete vitamin or trace element solutions (Verduyn et al., 1992), acetic acid or DLmalic acid, with the latter two at concentrations found in undiluted CW to standard medium. Concentrated vitamin solution contained (mg/l): biotin, 0.05; calcium panthotenate, 1; cyanocobalamin, 1; myo-inositol, 25; nicotinic acid, 1; para-aminobenzoic acid, 0.2; pyridoxineHCl, 1; riboXavin 1, and thiamineHCl, 1, whereas concentrated trace elements solution consisted of (mg/l): CaCl2 2H2O, 4.5; CoCl2 6H2O, 0.3; CuSO4 5H2O, 0.3; FeSO4 7H2O, 3; H3BO4, 1; KI, 0.1; MnCl2 4H2O, 1; NaMoO4 2H2O, 0.4 and ZnSO4 7H2O, 4.5. Where indicated, trace element and vitamin solutions were added to media at 1 ml/l. Glucose was sterilized separately at 110 C as a 50% solution in water or coconut water as required, whereas other medium components were sterilized together at 120 C. Sterilization of media supplemented with CW did not lead to any precipitates or excessive browning. In an additional experiment, medium dissolved in pure CW was Wlter-sterilized (0.45 m). This gave the same results as heatsterilized medium. Flasks were shaken at 200 rpm and 25 C for up to four days, Samples were taken regularly and centrifuged in 2 ml Eppendorf tubes (12,000 rpm for 10 min). The pellet was washed twice with distilled water and subse-

quently freeze-dried, whereas the supernatant was frozen at 20 C for further analysis. 2.2. Fatty acid analysis Freeze-dried biomass (ca. 10 and 20 mg dry weight) was weighed accurately and transferred to amber reaction vials. Two milliliter of 4% sulfuric acid in methanol was added, as well as C17:0 as an internal standard and butylated hydroxytoluene (BHT) as anti-oxidant at a Wnal concentration of 0.1 g/l. After mixing on a vortex, the vials were incubated in a water bath at 90 C for 1 h. After cooling, the mixture was extracted twice with a mixture of 1 ml of hexane and 1 ml of water. Further extractions did not increase recovery of fatty acid methyl esters (FAMEs). The extract was centrifuged (5 min at 3000g), and a small amount of solid anhydrous sodium sulfate was added to absorb water. The liquid phase was then injected on a Shimadzu GC-17A gas chromatograph equipped with a Omegawax 250 (30 m 0.25 mm) column (Supelco, USA), an auto-injector and Xame ionization detector. Injector and detector temperatures were 250 and 260 C, respectively, with helium as carrier gas at a linear velocity of 30 cm/s. Column temperature was held at 200 C for 10 min, then increased to 230 at 10 C/min and kept at this temperature for 14 min. The procedure was checked by methylation of pure palmitic and docosahexaenoic acid (Sigma, USA). Identity of FAMEs was initially conWrmed with GC/MS; for routine analysis peak quantiWcation was performed by comparison with four dilutions of a mixed PUFA-standard (no. 189-19) as well as pure DHAFAME (both Sigma, USA). 2.3. Sugar and acid analysis Free sugars were assayed by HPLC (Waters 2690 injector equipped with a model 410 diVerential refractometer) on a Sugar-Pak I (Waters, Milford, USA) column (300 6.5 mm) at 90 C with a Xow rate of 0.5 ml/min. The mobile phase consisted of double distilled water containing 50 mg/l EDTA. Before injection, samples were de-fatted by hexane extraction and subsequently Wltered through a 0.45 m Wlter. Injection volume was 50 l. Organic acids were assayed by HPLC (Waters 290 injector) with a model 2996 photodiode array detector on an IC-Pak Ion-exclusion 50A column (150 7.8 mm) at 55 C with a Xow rate of 0.7 ml/min. The mobile phase consisted of 0.1% phosphoric acid (pH 2.0). Before injection (50 l) samples were Wltered through a 0.45 m Wlter. Peaks that could be identiWed by this system were citric acid (retention time (RT) 4.42 0.08 min), DL-malic acid (RT 5.18 0.10), succinate (RT 6.02 0.09), whereas fumaric acid and acetic acid overlapped (RT both 7.47 0.10) and could not separated even when two columns were used in series. Hence acetic assayed was assayed by GC on a Shimadzu GC-17A instrument equipped with FID-detector and a 80/120 Carbopack BDA/4% Carbowax 20M column (20 m 0.2 mm). Column

P. Unagul et al. / Bioresource Technology 98 (2007) 281287

283

chamber was controlled at 190, and injection port and detector at 210 C, respectively with nitrogen as carrier gas. Response for fumaric acid (1 g/l) was negligible in the latter system. All the experiments were conducted on at least two independent occasions and in duplicate. Data shown pertain to a typical experiment with the mean standard deviation. Biomass levels and FA-weight contents diVered by no more than 5% and 7%, respectively, between independent experiments. Yields are deWned as gram biomass (or DHA) produced per gram glucose consumed. Actual initial glucose levels, assayed within 5 min after inoculation was used. Average volumetric biomass- and DHA-productivities were calculated over the entire period of cultivation until glucose exhaustion and is expressed as mg biomass or DHA formed per liter per hour. 3. Results 3.1. Proximate composition of mature coconut water Crude proximate bulk composition of coconut water is shown in Table 1. A signiWcant amount of monosaccharides in the form of ca. equal levels of glucose and fructose (total 11 g/l) was observed, with the remainder consisting of sucrose. Together these sugars accounted for almost half of the dry solids. Nitrogen was only a minor component as shown by CHN-analysis. From the latter it can be calculated that undiluted coconut water contained ca. 10 mg/l N, whereas this was approximately 1.0 g/l in 10 g/l YE. Fatty acids made up 16% of the solids, with decanoic acid (C10:0) and myristic acid (C12:0) as the main components (Table 1). Long-chain highly unsaturated fatty acids were absent. Organic acids were also found, with malic acid, acetic acid and succinic acid as major components. Coconut water contained a high level of manganese (0.19 mg/l) as compared to 10 g/l YE (0.007 mg/l, not shown), but sodium content was relatively low at 0.2 g/l as compared to 4.9 g/l in the artiWcial sea salts used as main salt source. A low amount of magnesium (0.045 g/l) was found as compared to the level

in 15 g/l ASW (0.54 g/l, Sigma information sheet). In contrast, a signiWcant amount of potassium was found in CW. A second batch of coconut water collected from the same factory half a year later showed a proximate analysis, which diVered less than 10% for the bulk components listed in Table 1 (data not shown). 3.2. Addition of coconut water or other growth-factors to standard medium Dissolving medium components in pure coconut water (CW) resulted in a higher maximal biomass level (27.5 g/l) as compared to standard medium (16.5 g/l). It should be noted that in addition to the 60 g/l glucose added, undiluted CW contained 11 g/l monosaccharides (Table 1), resulting in an initial total sugar level (excluding sucrose, which was not consumed) of 71 g/l. In order to allow easier comparison with standard medium, controls were performed in which Wrstly, glucose addition to media with undiluted CW was reduced to 49 g/l (giving 60 g/l total initial sugar as in the standard medium) and secondly additional glucose and fructose (11 g/l in total) was added to standard medium to give a total initial sugar level of 71 g/l. Obviously, absolute biomass levels increased when extra sugar was added, i.e. from ca. 17 to 19 g/l in the standard medium, but biomass or DHA-yields on glucose were not eVected (Table 2) and hence could be compared directly. In agreement with this observation, reducing total sugar levels in media with undiluted CW from 71 to 60 g/l reduced absolute biomass level, but not biomass yield (Table 2). Time course of biomass, fatty acid formation and glucose consumption for a cultivation with undiluted CW as well as a cultivation in standard medium supplemented with extra sugar to obtain identical initial sugar levels are shown in Fig. 1. To reduce the number of lines, no distinction was made between fructose and glucose, which were both completely consumed. It has been shown previously that cultivation on glucose or fructose gave similar results for strain Sk-02 (Unagul et al., 2005). Consumption of sucrose present in CW was not observed (data not shown) and has been omitted. The total fatty acid

Table 1 Proximate analysis of a typical batch of undiluted mature coconut water (mean of duplicate analysis SD) Solids content Total carbohydrate Glucose Fructose Sucrose Organic acids Acetic acid Citric acid DL-malic acid Succinic acid Elemental analysis C30.99H7.27N0.026 40.0 2.0 17.8 5.0 0.4 6.1 0.4 6.7 0.6 4.5a 0.7 0.1 0.05 0.03 2.5 0.2 0.9 0.1 Fatty acidsa C8:0 C10:0 C12:0 C14:0 C16:0 C18:0 C18:1 C18:2 (Trace) elements Manganese (mg/l) Magnesium Potassium Sodium 6.6 0.35 1.78 2.23 0.93 0.48 0.14 0.32 0.07 0.19 0.02 0.046 0.009 0.41 0.06 0.23 0.03

Data are expressed in g/l unless otherwise noted with the exception of CHN which is in % w/w. a Excluding non-identiWed peaks.

284

P. Unagul et al. / Bioresource Technology 98 (2007) 281287

Table 2 Biomass, FA-contents and yields as well as average volumetric biomass- and DHA-productivity (R) of S. mangrovei Sk-02 following growth in a standard medium (10 g/l YE, 60 g/l glucose in 15 g/l ASW with initial pH 6.5) supplemented with coconut water (CW) or various deWned growth-factors Addition(s) None (control) None, no ASW 17% CW 33% CW 66% CW 100% CW2 100% CW 100% CW + 0.1 M MES 100% CW, hexane extr3 100% CW, no ASW Glucose + fructose4 Trace elements (1 ml/l) Vitamins (1 ml/l) Glucose + fructose4 + trace elements (1 ml/l) Total sugar1 Max biomass Final pH (g/l) (g/l) 60 60 62 64 66 60 71 71 71 71 71 60 60 71 16.9 0.6b 8.5 0.6a 23.5 0.6d 28.6 0.5f 28.4 0.5f 24.0 0.6d 27.5 0.4e,f 26.8 0.4e 27.0 0.6e,f 7.5 0.8b,c 19.0 0.5c 24.5 0.6d 15.9 0.7b 27.9 0.9f 7.1a 7.0a 7.8b 8.2c 8.4c,d 8.3d 8.6d 6.9a 8.7c 7.3a 7.3a 7.9b 7.0a 8.0b Total FA (% w/w) 51.4a 52.1a,b 50.9a 49.4a 50.7a 50.1a 49.6a 50.6a 49.4a 54.2b 51.9a 50.3a 49.1a 50.6a DHA DHA (g/l) (% w/w) 19.9 21.1 20.0 19.2 20.1 20.3 20.7 21.3 20.4 19.6 21.1 21.3 18.9 20.9 3.4 0.3b 1.8 0.2a 4.7 0.2d 5.5 0.3d,e 5.7 0.2e 4.9 0.2d 5.7 0.2e 5.7 0.3e 5.5 0.3d,e 3.4 0.2b 4.0 0.2c 5.2 0.2d 3.0 0.3b 5.8 0.3e Ybiomass (g/g sugar) 0.28 0.14 0.38 0.45 0.43 0.40 0.39 0.38 0.38 0.25 0.27 0.41 0.27 0.39 Rbiomass YDHA (mg/l h) (g/g sugar) 235 89 326 397 394 333 382 372 375 104 264 340 221 388 0.057 0.030 0.076 0.086 0.086 0.082 0.080 0.080 0.077 0.048 0.056 0.087 0.050 0.082 RDHA (mg/l h) 47 19 65 76 79 68 79 79 79 47 56 72 42 81

A diVerent superscript in a column indicates a statistical diVerence (p < 0.05); TFA D total fatty acid; MES D 4-morpholineethanesulfonic acid monohydrate. 1 Total sugar at time of inoculation, excluding sucrose in CW, which was not consumed. 2 Glucose addition reduced to 49 g/l. 3 Fatty acids removed by extraction with hexane. 4 5 + 6 g/l, equivalent to content in 100% coconut water.

Biomass, glucose (g/l) and TFA (%w/w)

75

A
pH TFA
8

75

B
8

DHA (g/l), medium pH

50

50

DHA
25

4 25

biomass glucose
0 0 1 2 3 4

0 0 1 2 3 4

Time (days)

Time (days)

Fig. 1. Biomass, TFA-formation, glucose consumption and medium pH during growth of S. mangrovei Sk-02 in standard medium (60 g/l glucose, 10 g/l YE and 15 g/l ASW) with addition of 11 g/l glucose plus fructose (A) or standard medium dissolved in pure coconut water (B). Note: Glucose (65 g/l) and fructose (6 g/l) have been pooled and displayed as glucose.

(TFA) and DHA-contents of the biomass were comparable in both media (ca. 50% and 20% (w/w), respectively, Fig. 1). A signiWcant increase in medium pH was observed during cultivation, especially in the case of CW-addition (Fig. 1B) but addition of a buVer (0.1 M MES), which maintained pH in the range of 6.57, did not result in a signiWcant change in maximal biomass level (Table 2). Further experiments with reduced levels of coconut water suggested that optimal results in terms of absolute biomass- and DHA-levels were obtained with ca. 3366% (v/v) CW, whereas below ca. 33% (v/v) CW, biomass yields began to decrease (Table 2). In terms of volumetric DHA-productivity, values of 72 79 mg/l h were obtained in the presence of 33100% (v/v)

CW, whereas this amounted to a maximum of only 56 mg/ l h in media without CW, i.e. in the standard medium with monosaccharides increased to 71 g/l (Table 2). Biomass volumetric productivity followed the same trend as DHA-productivity. The FA-composition varied relatively little for all conditions tested, hence only a selection is shown in Table 3. Palmitic acid and DHA each made up ca. 40% of the total fatty acid (TFA), whereas the other main fatty acid was docosapentanoic acid (DPA, C22:5 n6) at ca. 8% of TFA. Pentadecanoic acid (C15:0) was always high in control medium. Minor and unidentiWed peaks amounted to less than 2% of the total area and were only marginally higher in cells grown in the presence of CW. At the end of

P. Unagul et al. / Bioresource Technology 98 (2007) 281287 Table 3 Fatty composition (as % of TFA) in selected samples Addition FA (% of TFA) C14:0 None (control) 33% CW 100% CW 100% CW, no ASS Trace elements 3.9 4.0 4.4 3.1 3.8 C15:0 5.3a 2.8b 2.3b 3.1b 3.0b C16:0 43.4a 42.4a 42.9a 49.0b 41.0a C18:0 1.0 0.8 0.8 0.8 0.8 C22:5 n6 7.0 7.2 7.4 6.5 8.2 C22:6 n3 38.7a 41.5b 40.5a,b 36.5a 42.3b

285

Others1 0.7a 1.3b 1.7b 1.0a 0.9a

See Table 2 for details. A diVerent superscript in a column indicates a statistical diVerence (p < 0.05). 1 Including C10:0, C12:0, C16:1, C18:0, C18:1 and unidentiWed peaks.

the cultivation, malic acid as well as acetic acid present in CW had been consumed completely, whereas there was a doubling of the succinic acid concentration in the medium (data not shown). Extraction of the fatty acids from coconut water with hexane before inoculation gave results comparable to complete coconut water. Replacement of coconut water in standard medium with either a trace element solution, a vitamin solution or a combination thereof, showed that trace elements could largely mimic the eVect of coconut water, with DHA-levels up to ca. 6 g/l (Table 2), whereas vitamin addition had no signiWcant eVect. Combination of trace elements and vitamins gave results identical to addition of trace elements only. Addition of other speciWc components found in CW, such as malic acid and/or acetic acid, to standard medium did not result in more than a 5% change in biomass or FA-formation (not signiWcant) as compared to the standard medium. Combination of these acids with vitamins and trace elements gave results identical to the latter (data not shown). Finally, omitting sea salts from the standard medium reduced biomass yield by half, but in the presence of undiluted CW a biomass yield comparable to that in standard medium which contained sea salts was obtained (Table 2). 3.3. Addition of CW to media with N-sources other than YE Replacement of YE with other N-sources (casitone, peptone or tryptone) generally resulted in poor growth and high residual glucose levels, whereas addition of 50% (v/v) CW signiWcantly improved biomass levels, resulting in a two-to-three-fold increase in biomass and DHA-levels (Table 4).

4. Discussion 4.1. Availability and proximate composition of coconut water Coconut water is a common waste product in SouthEast Asia. Thailand produced an estimated 200,000 tons of coconut water in 2001 (calculated from total output, Agricultural Statistics of Thailand crop year 20012002). A small amount is used in food gels and sports drinks, but most is discarded as waste in diluted liquid form. It can be obtained free if transportation costs are paid. The main sugars in CW are fructose, glucose and sucrose (this study, Santosa et al., 1996), but the latter was not consumed in agreement with carbon uptake studies for various Schizochytrium and Thraustochytrium species (Bahnweg, 1979). It has been established that the composition of CW depends on age of the coconut, soil conditions, locality, etc. (Santosa et al., 1996). Nevertheless, the analytical proximate composition found in the present study is in reasonable agreement with the data from Santosa et al. (1996). They reported a solids content of 55 g/l for mature coconut water from Indonesia, with ca. 64% consisting of sugars. Malic acid was reported as the main organic acid and C12:0 as the major fatty acid. 4.2. EVect of coconut water and its constituents on biomass and fatty acid levels Threefold diluted CW functioned as an eVective growth promoter for cultivation of Sk-02 (cf. Table 2). Although CW contained a signiWcant amount of short-chain to medium-chain fatty acids (cf. Table 1), these were appar-

Table 4 EVect of coconut water addition (50% v/v) on biomass-, TFA-content, absolute DHA-level following growth of S. mangrovei Sk-02 in media with various nitrogen sources (1 g total nitrogen/l) N-source No CW Time (h) Casitone Peptone Tryptone 72b 96a 72b
1

50% CW Biomass (g/l) 7.2b 3.5a 10.8c TFA (% w/w) 48.7a 58.5b 57.4b DHA (g/l) 1.4b 0.7a 2.2c Time (h)1 72 72 72 Biomass (g/l) 17.2b 11.5a 22.7c TFA (% w/w) 55.9b 48.4a 57.6b DHA (g/l) 3.6b 2.1a 5.1c

All media contained 60 g/l glucose and 15 g/l ASW (initial pH 6.5). A diVerent superscript in a column indicates a statistical diVerence (p < 0.05). 1 Time refers to cessation of growth.

286

P. Unagul et al. / Bioresource Technology 98 (2007) 281287

ently not incorporated into the biomass and did not eVect overall TFA-composition (cf. Table 3). The positive eVect of coconut water on biomass formation could be partly ascribed to its content of monosaccharides, but these only increased biomass levels by 1015% as shown by increasing the monosaccharide level in standard medium to that in media dissolved in undiluted CW. However, biomass (and DHA-) yields on sugar were independent over the range of monosaccharides (6071 g/l) tested and hence could be compared directly. Due to the complex nature of CW, extensive fractionation was not attempted with the exception of removal of fatty acids, which did not aVect results. Further tests were performed with addition of various components found in CW to standard medium. Weak acid addition did not improve biomass formation; apparently the acid concentrations in CW were too low, although acetic acid is a precursor of acetyl-CoA and hence of fatty acids. Malic acid is an important intermediate in the provision of NADPH required for FA-formation, which is thought to occur via malic enzyme in many oleaginous organisms (Ratledge, 2004). This then left (trace) elements and vitamins as the remaining possibilities. In this respect it can be noted that coconut water is commonly sold by manufacturers of biochemicals in a freeze-dried form as an additive to plant growth media due to its high content of (some) vitamins and trace elements (e.g. Sigma-Aldrich catalog 2005). Addition of the former to a standard medium did not aVect the growth, indicating that YE provided suYcient vitamin(s), but a standard trace element solution was able to mimic the eVect of CW, taking into account the eVect of the extra monosaccharides in the latter. By comparison of the reported trace element levels in coconut water (Santosa et al., 1996) with those in YE (Difco manual) and ASW (Sigma information sheet), it can be concluded that only manganese levels would increase signiWcantly upon CWaddition. Its eVect remains to be conWrmed, but it was noted by Santosa et al. (1996) that manganese was the most important trace element in CW with levels up to Wvefold higher than common trace elements such as iron and zinc. Finally, an eVect of further components in CW, for instance major elements such as magnesium or potassium, cannot be excluded. In this respect it was noted that, whereas in standard medium omission of sea salts reduced biomass yield by ca. 50% as shown for other Schizochytrium sp. (Yokochi et al., 1998), CW-addition vastly improved biomass levels, despite a low sodium-content in the latter (cf. Table 1). It was shown previously for Sk-02 that replacement of artiWcial sea water with an equivalent amount of sodium chloride reduced biomass yields signiWcantly (Unagul et al., 2005), suggesting that other elements in sea water were required. The most abundant elements in sea water after chloride and sodium are magnesium and sulfate (Sigma information sheet), but in the absence of seawater or sea salts, their only sources were YE and the inoculum and a limitation might occur. In recent experiments with Sk-02 it was shown that addition of both trace elements as well as

low levels of sodium chloride and magnesium sulfate (ca. 0.8 and 0.6 g/l, respectively) to standard YEG-media without ASW resulted in up to 25 g/l biomass (Unagul et al., 2006). It can be concluded that in the present work, CW provided (some of) the elements normally present in ASW (including magnesium and sulfate), but at levels insuYcient to permit maximal biomass formation. The potassium content of CW may also have been beneWcial as this element has been reported to stimulate biomass formation of Thraustochytrium aureum when sodium was present at low levels (Garrill et al., 1992). In general, the results here pointed to the potential of signiWcant limitations in the simple organic nitrogen sourceglucoseASW media commonly used for growing thraustochytrids. 4.3. Production of DHA in various media Though no attempt was made in this study to optimize the absolute DHA-level, for instance by further increasing glucose levels, the DHA-levels obtained (up to 6 g/l) compared favorably with other shake Xask studies on DHAproduction: among the high values reported, Fan et al. (2002) obtained a maximal DHA-levels of up to 3.1 g/l for various S. mangrovei strains on glucoseYE media, whereas 4.2 g/l was obtained with Schizochytrium limacinum SR21 (Yokochi et al., 1998) using glucosecorn steep liquor ASW. In contrast, poor growth was reported for SR21 on other N-sources such as peptone and tryptone (Yokochi et al., 1998). This was also conWrmed for Sk-02 in the present study (cf. Table 4) but CW-addition relieved part of the apparent limitation(s), although biomass levels were always inferior to those obtained with YE. Hence it is likely that further limitations still exist. In general, conditions in shake Xask are far from optimal as signiWcant oxygen limitation occurs. Use of an aerated bioreactor and fed-batch techniques should lead to vastly improved biomass levels (Ward and Singh, 2005). Obviously, laboratory-grade YE as used in the present study would be prohibitively expensive, but it might be replaced by a cheap food-grade or brewers YE, supplemented with coconut water as a cheap source of growthfactors. Acknowledgements This research was supported by a grant from the Center for Genetic Engineering and Biotechnology (BIOTEC), Bangkok, Thailand to CV. We thank Professor Lilian Vrijmoed and M.Sc. Keith W. Fan (City University of Hong Kong) for providing the strain. References
Bahnweg, G., 1979. Studies on the physiology of the Thraustochytriales. 1 Growth requirements and nitrogen nutrition of Thraustochytrium spp., Schizochytrium spp., Japanochytrium spp., Ulkenia spp. and Labyrinthuloides. VeroV. Inst. Meeresf. Bremerh. 17, 245268.

P. Unagul et al. / Bioresource Technology 98 (2007) 281287 Fan, K.W., Chen, F., Jones, E.B.G., Vrijmoed, L.L.P., 2001. Eicosapentaeinoic acids production by and okara-utilizing potential of thraustochytrids. J. Ind. Microbiol. Biotechnol. 27, 199202. Fan, K.W., Vrijmoed, L.L.P., Jones, E.B.G., 2002. Physiological studies of subtropical mangrove Thraustochytrids. Bot. Mar. 45, 5057. Garrill, A., Clipson, N.J.W., Jennings, D.H., 1992. Preliminary observations on the monovalent cation relations of Thraustochytrium aureum, a fungus requiring sodium for growth. Mycol. Res. 96, 295304. Horrocks, L.A., Yeo, Y.K., 1999. Health beneWts of docosahexaenoic acid (DHA). Pharmacol. Res. 40, 211225. Ratledge, C., 2004. Fatty acid biosynthesis in microorganisms being used for single cell oil production. Biochimie 86, 817824. Santosa, U., Kubo, K., Ota, T., Tadokoro, T., Maekawa, A., 1996. Nutrient composition of kopyor coconuts (Cocos nucifera L.). Food Chem. 57, 299304. Sijtsma, L., de Swaaf, M.E., 2004. Biotechnological production and applications of the -3 polyunsaturated fatty acid docosahexaenoic acid. Appl. Microbiol. Biotechnol. 64, 146153.

287

Unagul, P., Assantachai, C., Phadungruengluij, S., Suphantharika, M., Verduyn, C., 2005. Properties of the docosahexaenoic acid-producer Schizochytrium mangrovei Sk-02: eVects of glucose, temperature and salinity and their interaction. Bot. Mar. 48, 387394. Unagul, P., Assantachai, C., Phadungruengluij, S., Pongsuteeragul, T., Suphantharika, M., Verduyn, C., 2006. Biomass and docosahexaenoic acid formation by Schizochytrium mangrovei Sk-02 at low salt concentrations, Bot. Mar. 49, in press. Verduyn, C., Postma, E., ScheVers, W.A., van Dijken, J.P., 1992. EVect of benzoic acid on metabolic Xuxes in yeasts: a continuous culture study on the regulation of respiration and alcoholic fermentation. Yeast 8, 501517. Ward, O.P., Singh, A., 2005. Omega-3/6 fatty acids: Alternative sources of production. Proc. Biochem. 40, 36273652. Yokochi, T., Honda, H., Higashihara, T., Nakahara, T., 1998. Optimization of docosahexaenoic acid production by Schizochytrium limacinum SR21. Appl. Microbiol. Biotechnol. 49, 7276.