Bioresource Technology 175 (2015) 578–585
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
Enhancing microalgal biomass productivity by engineering a microalgal–bacterial community Dae-Hyun Cho a, Rishiram Ramanan a, Jina Heo a,b, Jimin Lee a, Byung-Hyuk Kim a, Hee-Mock Oh a,b, Hee-Sik Kim a,b,⇑ a b
Sustainable Bioresource Research Center, Korea Research Institute of Bioscience and Biotechnology (KRIBB), Daejeon 305 806, Republic of Korea Major of Green Chemistry and Environmental Biotechnology, University of Science & Technology (UST), Daejeon 305 350, Republic of Korea
h i g h l i g h t s Phycosphere bacterial diversity analyzed in C. vulgaris by DGGE and pyrosequencing. Growth promoting and inhibiting microorganisms from C. vulgaris were co-cultivated. Four isolated bacterial strains improved algal growth, flocculation and lipid content. Algae supplied DOC, bacteria in return, supplied DIC and low molecular weight DOC. Engineered consortium significantly enhanced algal biomass and lipid productivity.
a r t i c l e
i n f o
Article history: Received 3 October 2014 Received in revised form 28 October 2014 Accepted 29 October 2014 Available online 6 November 2014 Keywords: Chlorella vulgaris Phycosphere bacteria Growth enhancement Artificial microalgal bacterial community Biodiesel
a b s t r a c t This study demonstrates that ecologically engineered bacterial consortium could enhance microalgal biomass and lipid productivities through carbon exchange. Phycosphere bacterial diversity analysis in xenic Chlorella vulgaris (XCV) confirmed the presence of growth enhancing and inhibiting microorganisms. Co-cultivation of axenic C. vulgaris (ACV) with four different growth enhancing bacteria revealed a symbiotic relationship with each bacterium. An artificial microalgal–bacterial consortium (AMBC) constituting these four bacteria and ACV showed that the bacterial consortium exerted a statistically significant (P < 0.05) growth enhancement on ACV. Moreover, AMBC had superior flocculation efficiency, lipid content and quality. Studies on carbon exchange revealed that bacteria in AMBC might utilize fixed organic carbon released by microalgae, and in return, supply inorganic and low molecular weight (LMW) organic carbon influencing algal growth and metabolism. Such exchanges, although species specific, have enormous significance in carbon cycle and can be exploitated by microalgal biotechnology industry. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction In large scale microalgal cultivation systems, the role of bacteria and other microorganisms cannot be ignored but are understudied (Unnithan et al., 2014). In natural ecosystems, many studies have shown the influence of these organisms over each other (Ashen and Goff, 2000; Geng and Belas, 2010). These interactions have been either mutualistic or commensalistic or parasitic and are often considered species specific (Ashen and Goff, 2000; Sapp et al., 2007). Recently, certain class of bacteria widely known as ⇑ Corresponding author at: Sustainable Bioresource Research Center, Korea Research Institute of Bioscience and Biotechnology (KRIBB), Daejeon 305 806, Republic of Korea. Tel.: +82 42 860 4326; fax: +82 42 860 4594. E-mail address: [email protected] (H.-S. Kim). http://dx.doi.org/10.1016/j.biortech.2014.10.159 0960-8524/Ó 2014 Elsevier Ltd. All rights reserved.
Plant Growth Promoting Bacteria (PGPB) has been acknowledged to be enhancing algal growth (Gonzalez and Bashan, 2000; Hernandez et al., 2009). Most studies on algal–bacterial interactions only address algal growth promotion and often speculate on the mode of interaction, inadequately addressing the role of algae in those interactions (Gonzalez and Bashan, 2000; Henderson et al., 2008). Moreover, algal–bacterial researchers have only dealt with effect of one species of bacteria on the growth and physiology of algae (Gonzalez and Bashan, 2000; Kim et al., 2014a). There have been no systematic studies so far that have addressed the role of several microorganisms in the mini-ecosystem surrounding algal cell walls called phycosphere (Kim et al., 2014a). Besides, use of an artificially engineered consortium to alter the dynamics of this mini-ecosystem has not been endeavored (Brenner et al., 2008).
D.-H. Cho et al. / Bioresource Technology 175 (2015) 578–585
Chlorella is not only a commercially exploited species but also widely studied green algae with respect to interactions with other organisms viz., bacteria and virus. Azospirillum sp. and Bacillus sp. have been implicated in growth promotion of unicellular microalgae Chlorella vulgaris, and has been reported to influence cell morphology, lipid, and pigment production (Gonzalez and Bashan, 2000). Azospirillum is a rhizosphere-dwelling, N2-fixing bacterium that is very versatile in nitrogen fixation assimilating NH+4, NO 3, or NO 2 under microaerobic conditions while also denitrifying under anaerobic conditions and hence can act as a general PGPB for numerous plant species and algae, including Chlorella (Gonzalez and Bashan, 2000; Steenhoudt and Vanderleyden, 2000). While the role of Azospirillum has been studied well, the roles of other PGPB and bacteria, in general, are under studied. Hence, in this study, the influence of phycosphere bacteria on microalgal growth was ascertained. Chlorella, a model algae used thus far for studying interactions, isolated from environmental samples was selected (Cho et al., 2013), and the associated microbial diversity as well as the effect of most isolated strains on the host were studied. Based on the results of the study, an artificial bacterial consortium was developed and their growth patterns with algae were characterized and results on mechanism of the interaction were also presented. 2. Methods 2.1. Samples and culture condition C. vulgaris OW-01 (NCBI accession number JQ664295) and Scenedesmus sp. YC001 (NCBI accession number KC439160) used in this study, were isolated from swine wastewater in Gonju, Korea and from an open pond in Daejeon, Korea respectively. Both cultures were grown in BG11 medium (Cho et al., 2013) and xenic unialgal cultures of C. vulgaris (CV) and Scenedesmus sp. (SC) were maintained by routine serial subculture. Axenic cultures of both strains were obtained in consequent treatment of ultrasonication, fluorescence activated cell sorter (FACS), and micropicking (Cho et al., 2013) and were continuously monitored for confirmation of axenicity using the said protocols. Microalgal strains were grown in 1 L Erlenmeyer flask constituting 300 ml BG11 medium for 14 days (constant stirring at 100 rpm, 25 °C, light intensity of 100 lmol m2 s1).
579
2.4. DNA extraction and PCR The microalgal biomass was washed twice with TE buffer (Tris 10 mM, EDTA 1 mM, pH 8.0) followed by centrifugation at 4800g for 5 min and mild centrifugation at 1000g for 5 min to eliminate free living bacteria. The biomass was resuspended in 1.5 ml distilled water and was centrifuged at 10,000g for 3 min at room temperature. DNA extractions were carried out in accordance with eukaryotic microalgal nucleic acids extraction (EMNE) method (Kim et al., 2012). The purity and quantity of DNA were examined by electrophoresis on 1% agarose gel and measured using absorbances at 260 and 280 nm with a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies Inc., Wilmington, DE, USA). One microliter of extracted DNA was used to amplify 16S rRNA genes by PCR using an MJ mini Thermal cycler (Bio-Rad, Hercules, CA, USA) with primers, 27f (50 -AGA GTT TGA TCC TGG CTC GA-30 ) and 518r (50 -ATTACCGCGGCTGCTGG-30 ) as described elsewhere (Cho et al., 2013). 2.5. Diversity analysis 2.5.1. Denaturing gradient gel electrophoresis Three different kinds of samples were used for extraction of genomic DNA: (1) XCV, (2) supernatant of XCV culture medium after centrifugation at 3000g and (3) filtered XCV (> 1 lm) culture medium (Minisart HY syringe filter, Sartorius, Germany). DGGE was performed as mentioned in an earlier study (Lee et al., 2013). Each DGGE band of interest was excised from the gel and cut bands were amplified as template for PCR. Forward and reverse strands sequences were assembled with SeqMan software (DNA STAR, Madison, WI) and homology searches of these assembled sequences were performed with the GenBank database using the Basic Local Alignment Search Tool (BLAST) in the NCBI (http:// www.ncbi.nlm.nih.gov/). 2.6. Pyrosequencing The PCR products were analyzed using pyrosequencing with a 454 Genome Sequencer FLX Instrument (Roche 454 Life Sciences, Branford, CT, USA). The raw reads were deposited into the NCBI short-reads archive database. The sequences obtained in this study were compared using Silva rRNA database.
2.2. Biomass determination
2.7. Isolation and identification of microalgal associated bacteria
Growth of green algae was determined by dry cell weight (DCW) and by monitoring the cell count using hemocytometer (Peters et al., 2011). In co-culture experiments, microalgae and bacteria were separated by ultrasonication and centrifuged in the presence of 40% Histodenz (Sigma, USA). For microalgal DCW determination, the cells were separated by Histodenz. The bacterial cell numbers were monitored by cell counting using epifluorescence microscopy preceded by DAPI or SYBR green staining in both bacterial fraction and microalgal fraction, as a small fraction of bacteria were still attached to microalgal cell wall even after ultrasonication and Histodenz treatment.
In order to isolate microalgae associated bacteria, algae was cultured in three different liquid medium which suit bacterial growth (R2A, TSA, BG11 + glucose 100 ppm). Subsequently, culture broths were spread on the corresponding agar plate medium for picking up single colonies. After 3 days of cultivation, each single, discriminated colony was plated further and incubated at 25 ± 1 °C and cultivated for 3 days. Each isolated bacterial strain was identified by sequencing 16S rRNA gene using colony PCR (Cho et al., 2013).
2.3. Lipid content and fatty acid composition The total lipids were extracted as described previously (Lee et al., 2010). The fatty acid composition was determined using the protocol supplied by MIDI Inc. and gas chromatography (GC2010, Shimadzu, Koyto, Japan). Each fatty acid was identified and quantified based on comparing the retention times and peak areas with FAME Mix, C8-C24 (18918-1AMP, Supelco, Sigma–Aldrich Co. LLC., St. Louis, MO, USA).
2.8. Co-culture of isolated bacteria with microalgae In the co-cultivation studies, the inoculum ratio was one of the first determining factors. The ratio of cell numbers of algae and bacteria in the exponentially growing xenic culture (6–8 days) was determined by FACS and the same population ratio was used in co-cultivation experiments throughout this study (Powell and Hill, 2013). For e.g. the total cell numbers of bacteria was kept constant at 1 105 cells/ml in all co-cultivation studies involving either one, two, three or four strains of bacteria. Control cultures without algae were also established in BG11 medium and BG11 medium supplemented with glucose. The cultures were stirred at
580
D.-H. Cho et al. / Bioresource Technology 175 (2015) 578–585
100 rpm, at 25 °C with a light intensity of 100 lmol m2 s1 into 50 ml of the BG11 medium in 250 ml Erlenmeyer flask for 24 days. 2.9. Chemical analyses
profiles were analyzed using Genescan Software version 3.7 (Applied Biosystems, CA, USA).
2.13. Statistical analysis
DOC and DIC were measured in supernatant of algal samples using total organic carbon analyzer (Multi N/C 3100, Analyticjena, Germany) (Kim et al., 2014b). Total carbohydrates were determined using a protocol reported earlier after filtering the algal supernatant with Amicon Ultra-4 Centrifugal Filter Units (>50, 10–50, 3–10 and 63 kDa; EMD Millipore, USA) and obtaining different fractions (Taylor, 1995).
Growth pattern of C. vulgaris and bacteria in response to various co-cultivation experiments were analyzed through correlation analysis. The growth measurements were transformed into a single common scale using Z-transformation so that the mean and variance (average = 0 and standard deviation = 1) were normalized (Cho et al., 2014). The Z-scores obtained were used to compute Pearson linear correlation analysis (SPSS 18.0, USA).
2.10. Fluorescence activated cell sorting FACS analysis was performed using a BD FACSAria cell sorter (Becton Dickinson, USA) to ascertain the algal and bacterial cell numbers as well to determine the algal morphology. Parameters such as cell aggregation as well as nile red fluorescence were determined using FACS analysis (Montero et al., 2011) and were performed at least twice for each strain. 2.11. Microscopic analysis The samples were stained with nile red and/or DAPI (40 -6diamidino-2-phenylindole), SYBR Green to determine the effect of bacteria on algal lipid droplet synthesis as well to ascertain the bacterial cell numbers (Ramanan et al., 2013). For scanning electron microscopy analysis, the samples were fixed in 2.5% paraformaldehyde–glutaraldehyde mixture buffered with 0.1 M phosphate (pH 7.2) for 2 h, followed by fixing with 1% osmium tetroxide in the same buffer for 1 h, dehydrated in graded ethanol, and substituted by isoamyl acetate and finally dried at the critical point in CO2. The samples were sputtered with gold in a sputter coater (SC502, Polaron) and observed using HITACHI S4300N scanning electron microscope (Hitachi, Japan). 2.12. Terminal restriction fragment length polymorphism Amplicons were purified using a PCR purification kit (Qiagen, Germany) as directed by the supplier, and eluted in 20 ll sterile water. Purified PCR products (approximately 100 ng) were digested with 5 U of HaeIII and HhaI (Fermentas, USA), in a 20 ll reaction volume. Restriction reactions were performed at 37 °C for 12 h. Aliquots (8 ll) of restriction digests were examined by 2.5% agarose gel electrophoresis using SYBR Green I staining. Total restriction fragment analysis used 1 ll of digested samples mixed with 1 ll of formamide (with loading buffer and DNA fragment length standard [Rox 2500, ABI]). The mixture was denatured at 94 °C for 5 min and snap-cooled on ice before electrophoresis on 7% polyacrylamide gel for 10 h at 250 V using an ABI377 automated DNA sequencer (Applied Biosystems, CA, USA). T-RFLP
3. Results and discussion 3.1. Growth characteristics of axenic and xenic cultures Cell growth of ACV and XCV was compared in BG11 medium (without any instantaneous/external carbon source) and maximum cell number reached to 1.86 107 and 3.01 107 cell/ml, respectively after 10 days of cultivation (Supplementary Fig. 1). Specific growth rate, cell density, lipid content and flocculation efficiency of XCV was 0.29 day1, 2.49 g/L, 28%, and 62%, respectively (Table 1). All analyzed parameters were higher in XCV indicating that bacteria influenced algal growth and metabolism. Although Chlorophyll-a concentration (wt/vol. culture) was higher in XCV (600 lg/L) but its content (wt/cell) was higher in ACV (0.034 pg/cell) compared to XCV (0.018 pg/cell) subtly indicating the presence of heterotrophic bacteria in the latter.
3.2. Bacterial diversity analysis 3.2.1. DGGE analysis Since a huge difference was observed in the growth rates of ACV and XCV, subsequent diversity analyses were carried out by DGGE and pyrosequencing. DGGE followed by sequence analysis of XCV revealed five bands of bacteria constituting Rhizobium sp., Hyphomonas sp., Terrimonas sp., Flavobacterium sp. and Mesorhizobium sp. which were later pure cultured as well as two bands of bacteria which were unculturable. Mild centrifugation of the XCV followed by washing and DGGE resulted in a different band pattern. Two bands belonging to Rhizobium sp., and Hyphomonas sp., were present in all the samples but were prominent in XCV and not in filtered sample, indicating their close association with algae (Supplementary Fig. 2). On the contrary, some bands diagnosed later as Flavobacterium, Mesorhizobium and uncultured bacteria were more prominent in supernatant or filtered samples. Although DGGE analysis revealed the bacterial diversity associated with the algal strain, abundance of each bacteria associated with the algal strain could not be calculated using DGGE.
Table 1 Comparison of microalgal growth and physiology parameters in ACV, XCV and AMBC.
Growth rate (max, day1)a Cell mass (g/L)b Lipid content (%)c FAME composition Flocculating activity 1 2 3
ACV
XCV
AMBC
0.22(±0.02) 1.3(±0.013) 22.4(±3.6) C18:2, C16:0, C16:2, C18:3 3(±60.4)%
0.29(±0.014) 2.49(±0.24) 28(±4.2) C18:2, C16:0, C16:3, C18:1 62(±6.3)%
0.47(±0.011) 3.31(±0.21) 28(±2.6) C16:0, C16:3, C18:0, C18:3 80(±7.6)%
All C. vulgaris strains were grown under photoautotrophic condition in BG11 medium under same conditions mentioned in Section 2. Cell mass was measured on 24 days cultivation after separation using Histodenz (Section 2). Lipid content was represented as the weight ratio of the extracted lipid from 1 g of cell.
D.-H. Cho et al. / Bioresource Technology 175 (2015) 578–585
3.2.2. Pyrosequencing analysis In order to determine the abundance of bacteria in XCV as well as to benefit from deep sequencing analysis to improve coverage of associated bacteria, 454 pyrosequencing analysis was performed. Results pointed to similar community as revealed by the DGGE (Supplementary Fig. 3). Nevertheless, Sphingomonas which was not detected in DGGE analysis, was prevalent in pyrosequencing (20%) while Flavobacterium (37%) and Hyphomonas (21%) remained the most abundant bacteria (Supplementary Fig. 3). Rhizobium (12%) and Bradyrhizobium (4%), the former was detected in DGGE, were lesser dominant strains. Interestingly, Exophiala sp., which is a widespread fungus known to be pathogenic to a variety of eukaryotic hosts, was the only associated eukaryote identified in pyrosequencing. From the analysis of bacterial diversity, it was understood that the phycosphere of XCV composed of both growth promoting and inhibiting microbial strains. 3.3. Isolation of microalgal associated bacteria Out of a total of 28 microorganisms identified by pyrosequencing, 14 microorganisms were isolated from XCV using a variety of culture medium. Out of these, 13 were bacteria from 4 different phylums and 10 orders (Supplementary Table 1), and one was fungus. Phylogenetically, isolated bacteria belonged to actinomycetes, firmicutes, a,b-proteobacteria and bacteroidetes (Fig. 1) but Cytophaga–Flavobacterium–Bacteroides (CFB) and a-proteobacteria phylum predominated in actual abundance (Supplementary Fig. 3). Lesser prominent strains such as Microbacterium sp.,
581
Ochrobactrum sp., Achromobacter sp., and Exophiala sp. have been reported to be algicidal, of which only Microbacterium sp., and Exophiala sp., could be isolated. The strains identified by 454 pyrosequencing and those isolated from XCV have been tabulated along with their characteristics (Supplementary Table 1). 3.4. Growth enhancement Four dominant bacterial strains (Flavobacterium, Hyphomonas, Rhizobium and Sphingomonas) as indicated by DGGE and pyrosequencing, were selected for further co-cultivation with ACV along with two least dominant strains (Microbacterium and Exophiala) reported to be algicidal (Fig. 2). All these strains were isolated from XCV and pure cultured. The co-cultivation studies with 4 dominant bacterial strains showed a generic increase in the algal cell numbers (>100%) when compared to axenic cultures (Fig. 2A and Supplementary Fig. 4). While co-cultivation with both Microbacterium sp. and Exophiala sp. resulted in growth inhibition with different growth pattern. Microbacterium sp. seems to have slowed down the growth rate of C. vulgaris OW-01 whereas Exophiala inhibits the growth of the algae (Fig. 2B and Supplementary Fig. 4). These results demonstrated that the dominant bacterial strains had a growth enhancing effect on algae while the isolated Microbacterium strain competed with algae for some essential nutrients resulting in a reduction in growth rate of algae, and Exophiala strain was algicidal. Conversely, the growth pattern of four dominant bacterial strains was determined in the presence of algae with a control in
Fig. 1. Phylogenetic tree of isolated bacterial strains (JM1-18) from the phycosphere of C. vulgaris.
582
D.-H. Cho et al. / Bioresource Technology 175 (2015) 578–585
algal growth (Supplementary Table 2 and Fig. 6). In addition, there is a possibility of these relationships being driven by growth phase of algae as indicated by correlation analysis (Supplementary Table 2) and bacterial growth pattern in co-culture experiments (Fig. 3). Nonetheless, co-cultivation studies confirmed that all these bacteria strains facilitated algal growth. 3.5. C. vulgaris – artificial algal–bacterial consortium (CV-AMBC) Further co-cultivation studies were conducted to determine the synergistic effect of various combinations of these four bacterial strains on microalgal growth. Growth curve experiments with different combinations followed by correlation analysis with axenic culture revealed that four dominant strains together yielded a statistically significant synergistic effect on C. vulgaris OW-01 (P 6 0.01) when compared to other combinations or each bacterial species alone (Supplementary Table 2 and Fig. 6). Biomass of ACV, XCV and CV-AMBC reached to 1.3, 2.49 and 3.31 g/L, respectively after 24 days (Fig. 3B) and the specific growth rate of CV-AMBC was highest at 0.47 day1 (Table 1). Taken together, the growth promotion might be a result of elimination of harmful microorganisms as well as the synergistic effect of all growth promoting bacteria on algae. 3.6. Physiological characteristics of CV-AMBC
Fig. 2. Co-cultivation of axenic C. vulgaris (C) with selected microorganisms. Green algae were cultured in 125-ml Erlenmeyer flasks in BG11 medium at 25 °C at a light intensity of 100 lE m2 s1. (A) Co-cultivation with four dominant phycosphere bacteria, closed circle; control (ACV), open circle; Flavobacterium (F), reversed closed triangle; Hyphomonas (H), open triangle; Rhizobium (R), closed square; Sphingomonas (S), (B) co-cultivation of growth inhibiting bacteria, circle; Exophiala (E), triangle; Microbacterium (M).
absence of algae. It should be noted that BG11 medium was a minimal medium with no instantaneous carbon source and hence did not support bacterial growth. No significant change in bacterial cell count was found through microscopic observation after DAPI staining in control even after 10 days of cultivation (Supplementary Fig. 5). Interestingly, all four cultures exhibited similar growth patterns with differing cell numbers in co-cultured flasks. Cell numbers of Rhizobium and Flavobacterium increased drastically in the late stationary phase of algae (24th day) reaching twice that of algal cell numbers although the algae–bacteria ratio of the initial inoculums was 10:1 (Fig. 3A and Supplementary Fig. 6). The cell numbers of Sphingomonas and Hyphomonas showed a mild increase in late stationary phase of algae, equalizing with algal cell numbers. It should be noted that algal cell numbers doubled in the exponential phase (8–16th day), while remaining almost identical in the stationary phase (16–25th day) while the bacterial growth propelled during the corresponding period. Correlation analysis of growth pattern of algae and bacteria in co-cultivation studies indicated that none of the bacterial strains encourage a typical mutualistic interaction with algae (P > 0.05), where a gradual increase in the growth rate of algae and bacteria over time is expected. Co-cultivation studies with each bacterium did not show statistically significant relationship possibly because of interrelationship between each bacterial species to enhance
Although co-cultivation with bacterial consortium suggested growth enhancement, very little information was gathered from the growth studies with respect to algal cell morphology. To study the morphology of C. vulgaris in the presence and absence of bacterial community, microscopic and flow cytometric analyses were employed (Supplementary Fig. 7). Flow cytometric analysis confirmed a definite increase in cell aggregation (Supplementary Fig. 8). Further examination of the cultures with SEM also corroborated with flow cytometric analysis (Supplementary Fig. 9). Confocal and FACS analysis of nile red stained cells indicated a mild increase in TAG content (20%) when co-cultivated with bacterial consortium (Supplementary Figs. 8 and 10), while FAME analysis showed a significant shift towards Oleic and palmitic acids (Fig. 4). Growth rates, lipid content and flocculation efficiency of CV-AMBC was superior to that of xenic and axenic cultures (Table 1). Although there is a possibility that increase in lipid content might be attributed to the increase in growth rate of algae in AMBC cultures resulting in depletion of nutrients, such nutrient depletion is usually accompanied with increase in polyunsaturated fatty acids in Chlorella unlike increase in saturated and monounsaturated fatty acids observed in this study (Shekh et al., 2013). Moreover, batch scale microalgal cultures using BG11 or TAP medium are usually not limited by availability of major nutrients but carbon and light (Ramanan et al., 2013). Hence increase in lipid content and shift in the FAME profile might be due to the presence of bacteria. Although the changes to algal metabolism can be attributed to the presence of these bacteria, it is necessary to confirm the absence of other bacteria in this defined consortium. Hence, t-RFLP analysis was performed and the population of bacteria in CV-AMBC was confirmed to be Flavobacterium sp., Hyphomonas sp., Rhizobium sp. and Sphingomonas sp. in 24 day cultures. The pseudo peaks were detected using pure cultures of respective bacteria as well by the use of two different restriction enzymes (Supplementary Fig. 11). 3.7. Mechanism behind change in microalgal physiology upon co-cultivation After confirming that the artificial consortium of bacteria and algae mutually effect each other’s growth and physiology, further
D.-H. Cho et al. / Bioresource Technology 175 (2015) 578–585
583
Fig. 3. (A) Time course study of cell numbers of each bacteria and C. vulgaris under co-cultivation. (B) Difference in the microalgal cell numbers in axenic, xenic and CV-AMBC confirming growth enhancement. Abbreviation as denoted in Fig. 2 and text.
Fig. 4. FAME profile of axenic, xenic and CV-AMBC.
studies were performed to ascertain the mechanism behind this relationship. As mentioned before, since the BG11 medium was devoid of external carbon and bacterial growth was not observed
in control cultures (Supplementary Fig. 5), it was hypothesized that fixed carbon, referred to as algogenic organic matter (AOM) in earlier studies (Henderson et al., 2008), was supplied to the bacterial consortium by the microalgae upon co-cultivation, especially in the late stationary phase. To confirm this hypothesis, time course experiments to monitor the levels of both dissolved organic carbon (DOC) and dissolved inorganic carbon (DIC) in the supernatants of ACV and AMBC cultures were performed. Results of these experiments suggest that fixed carbon released as DOC by algae was directly consumed by bacteria as DOC levels in AMBC culture was consistently much lesser than ACV culture especially after the stationary phase. In fact, on the 24th day, the DOC levels in AMBC culture was >10 times lesser than ACV culture (Supplementary Fig. 12A). Meanwhile, as algae supplied fixed carbon to bacteria, bacterial growth propelled during this corresponding period resulting in the release of DIC which might have been utilized by the algae to sustain growth (Supplementary Fig. 12B). Besides, when the supernatant of the AMBC and ACV cultures were segregated based on the size and scrutinized for total carbohydrate content, a more revealing finding was observed. The fraction containing LMW carbohydrates (
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
Enhancing microalgal biomass productivity by engineering a microalgal–bacterial community Dae-Hyun Cho a, Rishiram Ramanan a, Jina Heo a,b, Jimin Lee a, Byung-Hyuk Kim a, Hee-Mock Oh a,b, Hee-Sik Kim a,b,⇑ a b
Sustainable Bioresource Research Center, Korea Research Institute of Bioscience and Biotechnology (KRIBB), Daejeon 305 806, Republic of Korea Major of Green Chemistry and Environmental Biotechnology, University of Science & Technology (UST), Daejeon 305 350, Republic of Korea
h i g h l i g h t s Phycosphere bacterial diversity analyzed in C. vulgaris by DGGE and pyrosequencing. Growth promoting and inhibiting microorganisms from C. vulgaris were co-cultivated. Four isolated bacterial strains improved algal growth, flocculation and lipid content. Algae supplied DOC, bacteria in return, supplied DIC and low molecular weight DOC. Engineered consortium significantly enhanced algal biomass and lipid productivity.
a r t i c l e
i n f o
Article history: Received 3 October 2014 Received in revised form 28 October 2014 Accepted 29 October 2014 Available online 6 November 2014 Keywords: Chlorella vulgaris Phycosphere bacteria Growth enhancement Artificial microalgal bacterial community Biodiesel
a b s t r a c t This study demonstrates that ecologically engineered bacterial consortium could enhance microalgal biomass and lipid productivities through carbon exchange. Phycosphere bacterial diversity analysis in xenic Chlorella vulgaris (XCV) confirmed the presence of growth enhancing and inhibiting microorganisms. Co-cultivation of axenic C. vulgaris (ACV) with four different growth enhancing bacteria revealed a symbiotic relationship with each bacterium. An artificial microalgal–bacterial consortium (AMBC) constituting these four bacteria and ACV showed that the bacterial consortium exerted a statistically significant (P < 0.05) growth enhancement on ACV. Moreover, AMBC had superior flocculation efficiency, lipid content and quality. Studies on carbon exchange revealed that bacteria in AMBC might utilize fixed organic carbon released by microalgae, and in return, supply inorganic and low molecular weight (LMW) organic carbon influencing algal growth and metabolism. Such exchanges, although species specific, have enormous significance in carbon cycle and can be exploitated by microalgal biotechnology industry. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction In large scale microalgal cultivation systems, the role of bacteria and other microorganisms cannot be ignored but are understudied (Unnithan et al., 2014). In natural ecosystems, many studies have shown the influence of these organisms over each other (Ashen and Goff, 2000; Geng and Belas, 2010). These interactions have been either mutualistic or commensalistic or parasitic and are often considered species specific (Ashen and Goff, 2000; Sapp et al., 2007). Recently, certain class of bacteria widely known as ⇑ Corresponding author at: Sustainable Bioresource Research Center, Korea Research Institute of Bioscience and Biotechnology (KRIBB), Daejeon 305 806, Republic of Korea. Tel.: +82 42 860 4326; fax: +82 42 860 4594. E-mail address: [email protected] (H.-S. Kim). http://dx.doi.org/10.1016/j.biortech.2014.10.159 0960-8524/Ó 2014 Elsevier Ltd. All rights reserved.
Plant Growth Promoting Bacteria (PGPB) has been acknowledged to be enhancing algal growth (Gonzalez and Bashan, 2000; Hernandez et al., 2009). Most studies on algal–bacterial interactions only address algal growth promotion and often speculate on the mode of interaction, inadequately addressing the role of algae in those interactions (Gonzalez and Bashan, 2000; Henderson et al., 2008). Moreover, algal–bacterial researchers have only dealt with effect of one species of bacteria on the growth and physiology of algae (Gonzalez and Bashan, 2000; Kim et al., 2014a). There have been no systematic studies so far that have addressed the role of several microorganisms in the mini-ecosystem surrounding algal cell walls called phycosphere (Kim et al., 2014a). Besides, use of an artificially engineered consortium to alter the dynamics of this mini-ecosystem has not been endeavored (Brenner et al., 2008).
D.-H. Cho et al. / Bioresource Technology 175 (2015) 578–585
Chlorella is not only a commercially exploited species but also widely studied green algae with respect to interactions with other organisms viz., bacteria and virus. Azospirillum sp. and Bacillus sp. have been implicated in growth promotion of unicellular microalgae Chlorella vulgaris, and has been reported to influence cell morphology, lipid, and pigment production (Gonzalez and Bashan, 2000). Azospirillum is a rhizosphere-dwelling, N2-fixing bacterium that is very versatile in nitrogen fixation assimilating NH+4, NO 3, or NO 2 under microaerobic conditions while also denitrifying under anaerobic conditions and hence can act as a general PGPB for numerous plant species and algae, including Chlorella (Gonzalez and Bashan, 2000; Steenhoudt and Vanderleyden, 2000). While the role of Azospirillum has been studied well, the roles of other PGPB and bacteria, in general, are under studied. Hence, in this study, the influence of phycosphere bacteria on microalgal growth was ascertained. Chlorella, a model algae used thus far for studying interactions, isolated from environmental samples was selected (Cho et al., 2013), and the associated microbial diversity as well as the effect of most isolated strains on the host were studied. Based on the results of the study, an artificial bacterial consortium was developed and their growth patterns with algae were characterized and results on mechanism of the interaction were also presented. 2. Methods 2.1. Samples and culture condition C. vulgaris OW-01 (NCBI accession number JQ664295) and Scenedesmus sp. YC001 (NCBI accession number KC439160) used in this study, were isolated from swine wastewater in Gonju, Korea and from an open pond in Daejeon, Korea respectively. Both cultures were grown in BG11 medium (Cho et al., 2013) and xenic unialgal cultures of C. vulgaris (CV) and Scenedesmus sp. (SC) were maintained by routine serial subculture. Axenic cultures of both strains were obtained in consequent treatment of ultrasonication, fluorescence activated cell sorter (FACS), and micropicking (Cho et al., 2013) and were continuously monitored for confirmation of axenicity using the said protocols. Microalgal strains were grown in 1 L Erlenmeyer flask constituting 300 ml BG11 medium for 14 days (constant stirring at 100 rpm, 25 °C, light intensity of 100 lmol m2 s1).
579
2.4. DNA extraction and PCR The microalgal biomass was washed twice with TE buffer (Tris 10 mM, EDTA 1 mM, pH 8.0) followed by centrifugation at 4800g for 5 min and mild centrifugation at 1000g for 5 min to eliminate free living bacteria. The biomass was resuspended in 1.5 ml distilled water and was centrifuged at 10,000g for 3 min at room temperature. DNA extractions were carried out in accordance with eukaryotic microalgal nucleic acids extraction (EMNE) method (Kim et al., 2012). The purity and quantity of DNA were examined by electrophoresis on 1% agarose gel and measured using absorbances at 260 and 280 nm with a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies Inc., Wilmington, DE, USA). One microliter of extracted DNA was used to amplify 16S rRNA genes by PCR using an MJ mini Thermal cycler (Bio-Rad, Hercules, CA, USA) with primers, 27f (50 -AGA GTT TGA TCC TGG CTC GA-30 ) and 518r (50 -ATTACCGCGGCTGCTGG-30 ) as described elsewhere (Cho et al., 2013). 2.5. Diversity analysis 2.5.1. Denaturing gradient gel electrophoresis Three different kinds of samples were used for extraction of genomic DNA: (1) XCV, (2) supernatant of XCV culture medium after centrifugation at 3000g and (3) filtered XCV (> 1 lm) culture medium (Minisart HY syringe filter, Sartorius, Germany). DGGE was performed as mentioned in an earlier study (Lee et al., 2013). Each DGGE band of interest was excised from the gel and cut bands were amplified as template for PCR. Forward and reverse strands sequences were assembled with SeqMan software (DNA STAR, Madison, WI) and homology searches of these assembled sequences were performed with the GenBank database using the Basic Local Alignment Search Tool (BLAST) in the NCBI (http:// www.ncbi.nlm.nih.gov/). 2.6. Pyrosequencing The PCR products were analyzed using pyrosequencing with a 454 Genome Sequencer FLX Instrument (Roche 454 Life Sciences, Branford, CT, USA). The raw reads were deposited into the NCBI short-reads archive database. The sequences obtained in this study were compared using Silva rRNA database.
2.2. Biomass determination
2.7. Isolation and identification of microalgal associated bacteria
Growth of green algae was determined by dry cell weight (DCW) and by monitoring the cell count using hemocytometer (Peters et al., 2011). In co-culture experiments, microalgae and bacteria were separated by ultrasonication and centrifuged in the presence of 40% Histodenz (Sigma, USA). For microalgal DCW determination, the cells were separated by Histodenz. The bacterial cell numbers were monitored by cell counting using epifluorescence microscopy preceded by DAPI or SYBR green staining in both bacterial fraction and microalgal fraction, as a small fraction of bacteria were still attached to microalgal cell wall even after ultrasonication and Histodenz treatment.
In order to isolate microalgae associated bacteria, algae was cultured in three different liquid medium which suit bacterial growth (R2A, TSA, BG11 + glucose 100 ppm). Subsequently, culture broths were spread on the corresponding agar plate medium for picking up single colonies. After 3 days of cultivation, each single, discriminated colony was plated further and incubated at 25 ± 1 °C and cultivated for 3 days. Each isolated bacterial strain was identified by sequencing 16S rRNA gene using colony PCR (Cho et al., 2013).
2.3. Lipid content and fatty acid composition The total lipids were extracted as described previously (Lee et al., 2010). The fatty acid composition was determined using the protocol supplied by MIDI Inc. and gas chromatography (GC2010, Shimadzu, Koyto, Japan). Each fatty acid was identified and quantified based on comparing the retention times and peak areas with FAME Mix, C8-C24 (18918-1AMP, Supelco, Sigma–Aldrich Co. LLC., St. Louis, MO, USA).
2.8. Co-culture of isolated bacteria with microalgae In the co-cultivation studies, the inoculum ratio was one of the first determining factors. The ratio of cell numbers of algae and bacteria in the exponentially growing xenic culture (6–8 days) was determined by FACS and the same population ratio was used in co-cultivation experiments throughout this study (Powell and Hill, 2013). For e.g. the total cell numbers of bacteria was kept constant at 1 105 cells/ml in all co-cultivation studies involving either one, two, three or four strains of bacteria. Control cultures without algae were also established in BG11 medium and BG11 medium supplemented with glucose. The cultures were stirred at
580
D.-H. Cho et al. / Bioresource Technology 175 (2015) 578–585
100 rpm, at 25 °C with a light intensity of 100 lmol m2 s1 into 50 ml of the BG11 medium in 250 ml Erlenmeyer flask for 24 days. 2.9. Chemical analyses
profiles were analyzed using Genescan Software version 3.7 (Applied Biosystems, CA, USA).
2.13. Statistical analysis
DOC and DIC were measured in supernatant of algal samples using total organic carbon analyzer (Multi N/C 3100, Analyticjena, Germany) (Kim et al., 2014b). Total carbohydrates were determined using a protocol reported earlier after filtering the algal supernatant with Amicon Ultra-4 Centrifugal Filter Units (>50, 10–50, 3–10 and 63 kDa; EMD Millipore, USA) and obtaining different fractions (Taylor, 1995).
Growth pattern of C. vulgaris and bacteria in response to various co-cultivation experiments were analyzed through correlation analysis. The growth measurements were transformed into a single common scale using Z-transformation so that the mean and variance (average = 0 and standard deviation = 1) were normalized (Cho et al., 2014). The Z-scores obtained were used to compute Pearson linear correlation analysis (SPSS 18.0, USA).
2.10. Fluorescence activated cell sorting FACS analysis was performed using a BD FACSAria cell sorter (Becton Dickinson, USA) to ascertain the algal and bacterial cell numbers as well to determine the algal morphology. Parameters such as cell aggregation as well as nile red fluorescence were determined using FACS analysis (Montero et al., 2011) and were performed at least twice for each strain. 2.11. Microscopic analysis The samples were stained with nile red and/or DAPI (40 -6diamidino-2-phenylindole), SYBR Green to determine the effect of bacteria on algal lipid droplet synthesis as well to ascertain the bacterial cell numbers (Ramanan et al., 2013). For scanning electron microscopy analysis, the samples were fixed in 2.5% paraformaldehyde–glutaraldehyde mixture buffered with 0.1 M phosphate (pH 7.2) for 2 h, followed by fixing with 1% osmium tetroxide in the same buffer for 1 h, dehydrated in graded ethanol, and substituted by isoamyl acetate and finally dried at the critical point in CO2. The samples were sputtered with gold in a sputter coater (SC502, Polaron) and observed using HITACHI S4300N scanning electron microscope (Hitachi, Japan). 2.12. Terminal restriction fragment length polymorphism Amplicons were purified using a PCR purification kit (Qiagen, Germany) as directed by the supplier, and eluted in 20 ll sterile water. Purified PCR products (approximately 100 ng) were digested with 5 U of HaeIII and HhaI (Fermentas, USA), in a 20 ll reaction volume. Restriction reactions were performed at 37 °C for 12 h. Aliquots (8 ll) of restriction digests were examined by 2.5% agarose gel electrophoresis using SYBR Green I staining. Total restriction fragment analysis used 1 ll of digested samples mixed with 1 ll of formamide (with loading buffer and DNA fragment length standard [Rox 2500, ABI]). The mixture was denatured at 94 °C for 5 min and snap-cooled on ice before electrophoresis on 7% polyacrylamide gel for 10 h at 250 V using an ABI377 automated DNA sequencer (Applied Biosystems, CA, USA). T-RFLP
3. Results and discussion 3.1. Growth characteristics of axenic and xenic cultures Cell growth of ACV and XCV was compared in BG11 medium (without any instantaneous/external carbon source) and maximum cell number reached to 1.86 107 and 3.01 107 cell/ml, respectively after 10 days of cultivation (Supplementary Fig. 1). Specific growth rate, cell density, lipid content and flocculation efficiency of XCV was 0.29 day1, 2.49 g/L, 28%, and 62%, respectively (Table 1). All analyzed parameters were higher in XCV indicating that bacteria influenced algal growth and metabolism. Although Chlorophyll-a concentration (wt/vol. culture) was higher in XCV (600 lg/L) but its content (wt/cell) was higher in ACV (0.034 pg/cell) compared to XCV (0.018 pg/cell) subtly indicating the presence of heterotrophic bacteria in the latter.
3.2. Bacterial diversity analysis 3.2.1. DGGE analysis Since a huge difference was observed in the growth rates of ACV and XCV, subsequent diversity analyses were carried out by DGGE and pyrosequencing. DGGE followed by sequence analysis of XCV revealed five bands of bacteria constituting Rhizobium sp., Hyphomonas sp., Terrimonas sp., Flavobacterium sp. and Mesorhizobium sp. which were later pure cultured as well as two bands of bacteria which were unculturable. Mild centrifugation of the XCV followed by washing and DGGE resulted in a different band pattern. Two bands belonging to Rhizobium sp., and Hyphomonas sp., were present in all the samples but were prominent in XCV and not in filtered sample, indicating their close association with algae (Supplementary Fig. 2). On the contrary, some bands diagnosed later as Flavobacterium, Mesorhizobium and uncultured bacteria were more prominent in supernatant or filtered samples. Although DGGE analysis revealed the bacterial diversity associated with the algal strain, abundance of each bacteria associated with the algal strain could not be calculated using DGGE.
Table 1 Comparison of microalgal growth and physiology parameters in ACV, XCV and AMBC.
Growth rate (max, day1)a Cell mass (g/L)b Lipid content (%)c FAME composition Flocculating activity 1 2 3
ACV
XCV
AMBC
0.22(±0.02) 1.3(±0.013) 22.4(±3.6) C18:2, C16:0, C16:2, C18:3 3(±60.4)%
0.29(±0.014) 2.49(±0.24) 28(±4.2) C18:2, C16:0, C16:3, C18:1 62(±6.3)%
0.47(±0.011) 3.31(±0.21) 28(±2.6) C16:0, C16:3, C18:0, C18:3 80(±7.6)%
All C. vulgaris strains were grown under photoautotrophic condition in BG11 medium under same conditions mentioned in Section 2. Cell mass was measured on 24 days cultivation after separation using Histodenz (Section 2). Lipid content was represented as the weight ratio of the extracted lipid from 1 g of cell.
D.-H. Cho et al. / Bioresource Technology 175 (2015) 578–585
3.2.2. Pyrosequencing analysis In order to determine the abundance of bacteria in XCV as well as to benefit from deep sequencing analysis to improve coverage of associated bacteria, 454 pyrosequencing analysis was performed. Results pointed to similar community as revealed by the DGGE (Supplementary Fig. 3). Nevertheless, Sphingomonas which was not detected in DGGE analysis, was prevalent in pyrosequencing (20%) while Flavobacterium (37%) and Hyphomonas (21%) remained the most abundant bacteria (Supplementary Fig. 3). Rhizobium (12%) and Bradyrhizobium (4%), the former was detected in DGGE, were lesser dominant strains. Interestingly, Exophiala sp., which is a widespread fungus known to be pathogenic to a variety of eukaryotic hosts, was the only associated eukaryote identified in pyrosequencing. From the analysis of bacterial diversity, it was understood that the phycosphere of XCV composed of both growth promoting and inhibiting microbial strains. 3.3. Isolation of microalgal associated bacteria Out of a total of 28 microorganisms identified by pyrosequencing, 14 microorganisms were isolated from XCV using a variety of culture medium. Out of these, 13 were bacteria from 4 different phylums and 10 orders (Supplementary Table 1), and one was fungus. Phylogenetically, isolated bacteria belonged to actinomycetes, firmicutes, a,b-proteobacteria and bacteroidetes (Fig. 1) but Cytophaga–Flavobacterium–Bacteroides (CFB) and a-proteobacteria phylum predominated in actual abundance (Supplementary Fig. 3). Lesser prominent strains such as Microbacterium sp.,
581
Ochrobactrum sp., Achromobacter sp., and Exophiala sp. have been reported to be algicidal, of which only Microbacterium sp., and Exophiala sp., could be isolated. The strains identified by 454 pyrosequencing and those isolated from XCV have been tabulated along with their characteristics (Supplementary Table 1). 3.4. Growth enhancement Four dominant bacterial strains (Flavobacterium, Hyphomonas, Rhizobium and Sphingomonas) as indicated by DGGE and pyrosequencing, were selected for further co-cultivation with ACV along with two least dominant strains (Microbacterium and Exophiala) reported to be algicidal (Fig. 2). All these strains were isolated from XCV and pure cultured. The co-cultivation studies with 4 dominant bacterial strains showed a generic increase in the algal cell numbers (>100%) when compared to axenic cultures (Fig. 2A and Supplementary Fig. 4). While co-cultivation with both Microbacterium sp. and Exophiala sp. resulted in growth inhibition with different growth pattern. Microbacterium sp. seems to have slowed down the growth rate of C. vulgaris OW-01 whereas Exophiala inhibits the growth of the algae (Fig. 2B and Supplementary Fig. 4). These results demonstrated that the dominant bacterial strains had a growth enhancing effect on algae while the isolated Microbacterium strain competed with algae for some essential nutrients resulting in a reduction in growth rate of algae, and Exophiala strain was algicidal. Conversely, the growth pattern of four dominant bacterial strains was determined in the presence of algae with a control in
Fig. 1. Phylogenetic tree of isolated bacterial strains (JM1-18) from the phycosphere of C. vulgaris.
582
D.-H. Cho et al. / Bioresource Technology 175 (2015) 578–585
algal growth (Supplementary Table 2 and Fig. 6). In addition, there is a possibility of these relationships being driven by growth phase of algae as indicated by correlation analysis (Supplementary Table 2) and bacterial growth pattern in co-culture experiments (Fig. 3). Nonetheless, co-cultivation studies confirmed that all these bacteria strains facilitated algal growth. 3.5. C. vulgaris – artificial algal–bacterial consortium (CV-AMBC) Further co-cultivation studies were conducted to determine the synergistic effect of various combinations of these four bacterial strains on microalgal growth. Growth curve experiments with different combinations followed by correlation analysis with axenic culture revealed that four dominant strains together yielded a statistically significant synergistic effect on C. vulgaris OW-01 (P 6 0.01) when compared to other combinations or each bacterial species alone (Supplementary Table 2 and Fig. 6). Biomass of ACV, XCV and CV-AMBC reached to 1.3, 2.49 and 3.31 g/L, respectively after 24 days (Fig. 3B) and the specific growth rate of CV-AMBC was highest at 0.47 day1 (Table 1). Taken together, the growth promotion might be a result of elimination of harmful microorganisms as well as the synergistic effect of all growth promoting bacteria on algae. 3.6. Physiological characteristics of CV-AMBC
Fig. 2. Co-cultivation of axenic C. vulgaris (C) with selected microorganisms. Green algae were cultured in 125-ml Erlenmeyer flasks in BG11 medium at 25 °C at a light intensity of 100 lE m2 s1. (A) Co-cultivation with four dominant phycosphere bacteria, closed circle; control (ACV), open circle; Flavobacterium (F), reversed closed triangle; Hyphomonas (H), open triangle; Rhizobium (R), closed square; Sphingomonas (S), (B) co-cultivation of growth inhibiting bacteria, circle; Exophiala (E), triangle; Microbacterium (M).
absence of algae. It should be noted that BG11 medium was a minimal medium with no instantaneous carbon source and hence did not support bacterial growth. No significant change in bacterial cell count was found through microscopic observation after DAPI staining in control even after 10 days of cultivation (Supplementary Fig. 5). Interestingly, all four cultures exhibited similar growth patterns with differing cell numbers in co-cultured flasks. Cell numbers of Rhizobium and Flavobacterium increased drastically in the late stationary phase of algae (24th day) reaching twice that of algal cell numbers although the algae–bacteria ratio of the initial inoculums was 10:1 (Fig. 3A and Supplementary Fig. 6). The cell numbers of Sphingomonas and Hyphomonas showed a mild increase in late stationary phase of algae, equalizing with algal cell numbers. It should be noted that algal cell numbers doubled in the exponential phase (8–16th day), while remaining almost identical in the stationary phase (16–25th day) while the bacterial growth propelled during the corresponding period. Correlation analysis of growth pattern of algae and bacteria in co-cultivation studies indicated that none of the bacterial strains encourage a typical mutualistic interaction with algae (P > 0.05), where a gradual increase in the growth rate of algae and bacteria over time is expected. Co-cultivation studies with each bacterium did not show statistically significant relationship possibly because of interrelationship between each bacterial species to enhance
Although co-cultivation with bacterial consortium suggested growth enhancement, very little information was gathered from the growth studies with respect to algal cell morphology. To study the morphology of C. vulgaris in the presence and absence of bacterial community, microscopic and flow cytometric analyses were employed (Supplementary Fig. 7). Flow cytometric analysis confirmed a definite increase in cell aggregation (Supplementary Fig. 8). Further examination of the cultures with SEM also corroborated with flow cytometric analysis (Supplementary Fig. 9). Confocal and FACS analysis of nile red stained cells indicated a mild increase in TAG content (20%) when co-cultivated with bacterial consortium (Supplementary Figs. 8 and 10), while FAME analysis showed a significant shift towards Oleic and palmitic acids (Fig. 4). Growth rates, lipid content and flocculation efficiency of CV-AMBC was superior to that of xenic and axenic cultures (Table 1). Although there is a possibility that increase in lipid content might be attributed to the increase in growth rate of algae in AMBC cultures resulting in depletion of nutrients, such nutrient depletion is usually accompanied with increase in polyunsaturated fatty acids in Chlorella unlike increase in saturated and monounsaturated fatty acids observed in this study (Shekh et al., 2013). Moreover, batch scale microalgal cultures using BG11 or TAP medium are usually not limited by availability of major nutrients but carbon and light (Ramanan et al., 2013). Hence increase in lipid content and shift in the FAME profile might be due to the presence of bacteria. Although the changes to algal metabolism can be attributed to the presence of these bacteria, it is necessary to confirm the absence of other bacteria in this defined consortium. Hence, t-RFLP analysis was performed and the population of bacteria in CV-AMBC was confirmed to be Flavobacterium sp., Hyphomonas sp., Rhizobium sp. and Sphingomonas sp. in 24 day cultures. The pseudo peaks were detected using pure cultures of respective bacteria as well by the use of two different restriction enzymes (Supplementary Fig. 11). 3.7. Mechanism behind change in microalgal physiology upon co-cultivation After confirming that the artificial consortium of bacteria and algae mutually effect each other’s growth and physiology, further
D.-H. Cho et al. / Bioresource Technology 175 (2015) 578–585
583
Fig. 3. (A) Time course study of cell numbers of each bacteria and C. vulgaris under co-cultivation. (B) Difference in the microalgal cell numbers in axenic, xenic and CV-AMBC confirming growth enhancement. Abbreviation as denoted in Fig. 2 and text.
Fig. 4. FAME profile of axenic, xenic and CV-AMBC.
studies were performed to ascertain the mechanism behind this relationship. As mentioned before, since the BG11 medium was devoid of external carbon and bacterial growth was not observed
in control cultures (Supplementary Fig. 5), it was hypothesized that fixed carbon, referred to as algogenic organic matter (AOM) in earlier studies (Henderson et al., 2008), was supplied to the bacterial consortium by the microalgae upon co-cultivation, especially in the late stationary phase. To confirm this hypothesis, time course experiments to monitor the levels of both dissolved organic carbon (DOC) and dissolved inorganic carbon (DIC) in the supernatants of ACV and AMBC cultures were performed. Results of these experiments suggest that fixed carbon released as DOC by algae was directly consumed by bacteria as DOC levels in AMBC culture was consistently much lesser than ACV culture especially after the stationary phase. In fact, on the 24th day, the DOC levels in AMBC culture was >10 times lesser than ACV culture (Supplementary Fig. 12A). Meanwhile, as algae supplied fixed carbon to bacteria, bacterial growth propelled during this corresponding period resulting in the release of DIC which might have been utilized by the algae to sustain growth (Supplementary Fig. 12B). Besides, when the supernatant of the AMBC and ACV cultures were segregated based on the size and scrutinized for total carbohydrate content, a more revealing finding was observed. The fraction containing LMW carbohydrates (
