Conversion of volatile fatty acids into polyhydroxyalkanoate by Ralstonia eutropha

Transcription

1 Journal of Applied Microbiology ISSN ORIGINAL ARTICLE Conversion of volatile fatty acids into polyhydroxyalkanoate by Ralstonia eutropha P. Chakraborty 1, W. Gibbons 2 and K. Muthukumarappan 3 1 Givaudan Flavors Corporation, Cincinnati, OH, USA 2 Department of Biology and Microbiology, South Dakota State University, Brookings, SD, USA 3 Department of Agriculture and Biosystems Engineering, South Dakota State University, Brookings, SD, USA Keywords bioreactor, condensed corn solubles, fermentation, polyhydroxybutyrate, Ralstonia eutropha, volatile fatty acids. Correspondence Panchali Chakraborty, Givaudan Flavors Corporation, 1199 Edison Drive, Cincinnati, OH 45216, USA. panchali.chakraborty@gmail.com : received 3 July 2008, revised 26 September 2008 and accepted 8 November 2008 doi: /j x Abstract Aims: The aims of this study were to optimize condensed corn solubles (CCS) as a medium for growth of Ralstonia eutropha and to determine the effects of individual volatile fatty acids (VFAs) on polyhydroxyalkanoate (PHA) production. Methods and Results: A CCS medium of concentration 240 g l )1 with a carbon : nitrogen ratio of 50 : 1 was developed as the optimal medium. Cultures were grown in 1-l aerated flasks at 250 rev min )1 at 30 C for 120 h. Comparable growth rates were observed in CCS vs a defined medium. At 48 h, VFAs were fed individually at different levels. Optimal levels of all the acids were determined to maximize PHA production. An overall comparison of the VFAs indicated that butyric and propionic acids provided the best results. Conclusion: An optimized CCS medium supported growth of R. eutropha. Butyric and propionic acids were the most efficient carbon sources to maximize PHA production when added at the 5 g l )1 level. Significance and Impact of the Study: The study shows that a byproduct of ethanol industry can be effectively used as a low cost medium for PHA production, thus partly reducing the cost of commercialization of biopolymers. Introduction Polyhydroxyalkanoate (PHA) is a naturally occurring biopolymer with synthetic polymer-like properties (Lee 1996a). PHAs are polyesters of 3-hydroxy fatty acid monomers. The carboxyl group of one monomer forms an ester bond with hydroxyl group of the neighbouring monomer. Due to environmental concerns regarding the use of synthetic polymers, there has been a growing public and scientific interest regarding the development and use of biopolymers. Polyhydroxyalkanoates are used for making bottles, cosmetics, containers, pens, golf tees, films, adhesives, nonwoven fabrics, toner and developer compositions, ion-conducting polymers and as latex for paper coating applications (Rutherford et al. 1997). It can be used to make laminates with other polymers such as polyvinyl alcohol. Poly (3-hydroxybutyrate-3-hydroxyvalerate) [P(HB-HV)] can be used in reconstructive surgery. High production costs have limited the use of biopolymers, however if these costs can be reduced, there would be widespread economic interest (Nishida and Tokiwa 1992). The synthesis of PHA is a two stage process (Lee 1996b). Cells are first grown to a high cell density in a nutrient-balanced medium, and then at least one essential nutrient is limited (e.g. nitrogen, phosphorous, etc.), while excess carbon is provided. This unbalanced nutritional state triggers intracellular production and accumulation of PHA. Several bacteria can be used to produce PHA, including Ralstonia eutropha. Ralstonia eutropha is a Gram-negative organism that grows on a variety of carbon sources, including several sugars and volatile fatty acids (VFAs) such as acetate, propionate, lactate, butyrate, valerate (Kunioka et al. 1989; Madison and Huisman 1999). At high concentrations, these acids can be inhibitory or toxic 1996 Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 106 (2009)

2 P. Chakraborty et al. Conversion of VFAs into PHA depending on ph and thus may result in low growth rate, acid utilization rate and yield of PHA (Axe and Bailey 1995). However, as VFAs are gradually consumed, cell activity may resume. At low initial VFA levels, substrate limitation can also result in slow acid utilization rates. However, when VFAs are present at appropriate levels, the undissociated fatty acids molecules enter the cytoplasm, are activated, and metabolized into CO 2, cell biomass, or PHA (Yu et al. 2002). Production of PHAs also depends on the C : N ratio. Low C : N ratios divert more carbon to energy formation and anabolic reactions for cell growth, leaving less for PHA production. However, at high C : N ratio, PHA productivity is increased due to insufficient supply of nitrogen for growth (Shimizu et al. 1999). The aims of this particular study were to: (i) develop a low cost medium to grow R. eutropha to high cell density, and (ii) determine the effects of different concentrations of VFAs on acid utilization rates, fermentation efficiency (FE), PHA concentration and PHA productivity. A carbon rich byproduct of corn ethanol production, condensed corn solubles (CCS), was used as an alternative medium. In addition to optimizing the CCS concentration, nitrogen was supplemented to maximize cell growth. When nitrogen became limiting at 48 h, VFAs (acetic, butyric, lactic and propionic acids) were added to aid PHA production. Our long-term objective is to produce these VFAs from another byproduct of corn ethanol production. This would reduce the cost of producing PHA. Materials and methods Culture, maintenance and inoculum propagation The ATCC type strain of R. eutropha was used. The culture was routinely transferred to nutrient broth, and incubated on a reciprocating shaker (250 rev min )1 ) at 30 C for 24 h. For short-term maintenance, the culture was stored on tryptic soy agar (TSA) slants covered with mineral oil and stored at 5 C. Long-term storage was via lyophilization followed by freezing at )20 C. Inoculum for all trials was prepared in a stepwise manner, by transferring the culture from TSA plates into 100 ml of the specific media (described below), then incubating the flasks for 24 h on a rotary shaker (250 rev min )1 ) at 30 C. These cultures were subsequently used, at a 1% (v v) rate, to inoculate aerated shake flasks experimental trials. Media Nutrient broth and a defined medium (Kunioka et al. 1989) were used as control media. Kunioka s medium was prepared by adding 2Æ5 g beef extract, 5 g yeast extract, 5 g ammonium sulfate and 5 g peptone to 1 l deionized water. Different amounts of CCS and deionized water were mixed to obtain different concentrations of CCS (80, 240, 400 and 700 g l )1 ) in the basal media. The phs of all media were adjusted to 7Æ0 using 10 mol l )1 NaOH. The media were then centrifuged at g for 7 min at C. The supernatant was filtered through Whatman filter paper #113 and dispensed into appropriate vessels prior to autoclaving. Experimental design Experimental trials were conducted in 1l conical flasks that contained 800 ml of different media. Filter sterilized air (1 l l )1 min )1 ) was provided through a glass sparger inserted through a rubber stopper, with a companion filtered vent tube. About four to five drops of antifoam (Cognis Clerol FBA 5059; Cognis, Cincinnati, OH) were added to the medium before inoculation. Flasks were incubated for a minimum of 48 h at 30 C and 250 rev min )1. Three replications of each trial were performed to assess growth rate and maximum cell population in each media. As no growth was obtained in 400 g l )1, and 700 g l )1, the next set of experiments compared two concentrations of CCS media (80 and 240 g l )1 ), with ph control. Subsequent trials evaluated the effect of controlling ph in the optimal range (7Æ5 8Æ5) by addition of 10 N H 2 SO 4 to these media. After the optimal CCS concentration was found to be 240 g l )1, this medium was supplemented with different levels of 178 g l )1 stock solution of filter-sterilized NH 4 HCO 3 solution to bring the C : N ratio to 30 : 1, 50 : 1, 70 : 1 and 90 : 1. The carbon level was based upon carbon sources that R. eutropha can metabolize (i.e. glycerol and organic acids).the ph was further adjusted to 7Æ0 by adding 10 N H 2 SO 4 before inoculation. A CCS medium containing 240 g l )1 with a C : N ratio of 50 : 1 was found to be optimum for growth of R. eutropha. Three replications were made for each C : N ratio to determine growth rate and maximum cell population. Once maximum growth was reached, individual VFAs (acetic, butyric, lactic and propionic acids) were added at three different levels. Stock solutions of these acids were prepared and filter sterilized before use. For acetic, butyric, and propionic acids, levels of 1, 3 and 5 g l )1 were used, while 2, 4 and 8 g l )1 of lactic acid were added. Incubation was continued for an additional 72 h. Three replications were performed for each level of each fatty acid to determine the effect of individual VFA on cell viability, acid utilization and PHA production. Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 106 (2009)

3 Conversion of VFAs into PHA P. Chakraborty et al. Analytical methods Samples were collected at regular intervals during incubation and ph was measured using an Acumet 950 ph meter (Thermo Fisher Scientific, Waltham, MA). Viable cell counts were made via pour plates using TSA. Samples were also analysed via a Waters HPLC system for sugars, organic acids and glycerol. These samples were first filtered through a nonsterile 0Æ2-lm filter to remove solids, then kept at )20 C until analysis. An Aminex HPX87H column (Bio-Rad), operated at 65 C with a heliumdegassed, 4 mmol l )1 H 2 SO 4 mobile phase at a flow rate of 0Æ6 mlmin )1 was used. Peaks were detected using a refractive index detector. Standard solutions of maltose, glucose, lactic acid, acetic acid, propionic acid, succinic acid and glycerol (at 3 and 30 g l )1 ) were used to calibrate the integrator. Samples collected at 0, 72 and 120 h were also tested for ammonia and phosphate using Hach Ammonium (no. HCT 102) and Phosphate (no. HCT 122) Unicell tests. At 24 and 120 h, 50 ml samples were collected to determine cell dry weights. To determine cell dry weight, samples were centrifuged and the precipitate was dried in a hot air oven at 80 C for 2 days. An additional 50 ml sample was collected at 120 h to determine PHA levels. To measure polyhydroxybutyrate (PHB), 50 ml samples of broth were centrifuged, and pellets were then lyophilized and ground using a mortar and pestle. The method developed by Braunegg et al. (1978) was used to extract simultaneously and derivatize PHB to the 3-hydroxyalkanoate methyl esters of the monomers. In this method, mg of ground cells were digested by adding 5 ml of digest solution and incubating at C for 4 h. The digest solution contained 50% chloroform, 42Æ5% methanol, 7Æ5% sulfuric acid (v v%). After cooling, sample was washed with 2 ml of water, and the bottom layer (containing the chloroform and methyl esters of PHB) was collected and placed in a GC vial with anhydrous sodium sulfate (to remove residual water). Vials were kept at )20 C until analysis. Polyhydroxybutyrate was quantified using a Hewlett- Packard 5890 Series II gas chromatograph with a flame ionization detector (GC-FID) (Braunegg et al. 1978; Comeau et al. 1988). Split injection was used onto a Supelco SSP-2380 capillary column (30 m 0Æ25 mm I.D. with 0Æ20-lm film). The inlet head pressure was 28 psi and the temperature programme started at 50 C for 4 min, then increased by 3 C min )1 to a final temperature of 146 C for 4 min. The injector and detector temperatures were 230 and 240 C respectively. Purified poly-(3-hydroxybutyric acid co-3-hydroxyvaleric acid) obtained from Sigma-Aldrich was used for a standard calibration. The co-polymer consisted of 88% 3-hydroxybutyric acid and 12% 3-hydroxyvaleric acid. Co-polymer concentrations of 2 10 mg ml )1 chloroform were digested as above, then analysed by GC-FID. Retention times were 14Æ9 min for methylated 3-hydroxybutyric acid, and 17Æ8 min for methylated 3-hydroxyvaleric acid. Statistical analysis All trials were performed in triplicate various fermentation parameters were analysed to determine least significant differences between treatments using randomized complete block design. FE was calculated as the percentage of substrate supplied that was consumed. Data were analysed using the PROC GLM procedure of sas software (SAS, Cary, NC) to determine F-values and LS means. Exponential regression equations were used to determine growth rates and acid utilization rates for each replication, from which averages were calculated. They were statistically analysed by ancova to test homogeneity of slopes. Statistical data were analysed at the significant level of P <0Æ05. Results The first objective of this study was to formulate a medium to maximize growth of R. eutropha. The second objective was to determine cell viability, additional growth, PHA production, and the utilization rates and FE for individual VFAs. Optimization of CCS medium Ralstonia eutropha did not grow in the CCS media at concentrations of 400 and 700 g l )1, therefore, these data are not shown. For the remaining media (Table 1), the highest cell populations occurred at 24 h in the defined medium (Kunioka s), followed by nutrient broth and CCS at 240 g l )1. Growth rates in Kunioka s medium and nutrient broth were not significantly different from 240 g l )1 CCS media. The lowest growth rate and cell population were observed in 80 g l )1 CCS medium, probably due to suboptimal nutrient concentrations. Table 1 Growth of Ralstonia eutropha in different media without ph control Media growth rate (h )1 ) biomass Defined medium 0Æ29 a 6Æ3 10 9a Nutrient broth 0Æ24 a 3Æ3 10 9b CCS medium 80 g l )1 0Æ21 b 9Æ3 10 8c CCS medium 240 g l )1 0Æ29 a 2Æ6 10 9b Mean values within column not sharing common superscript differ significantly (P < 0Æ05) Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 106 (2009)

4 P. Chakraborty et al. Conversion of VFAs into PHA Table 2 Growth of Ralstonia eutropha in different concentration of CCS media with or without ph control CCS media growth rate (h )1 ) biomass 240 g l )1 with ph control 0Æ24 a 2Æ5 10 9a 80 g l )1 with ph control 0Æ18 b 1Æ1 10 9b 240 g l )1 without ph control 0Æ29 a 2Æ6 10 9a 80 g l )1 without ph control 0Æ20 b 9Æ3 10 8b Mean values within column not sharing common superscript differ significantly (P < 0Æ05). Table 3 Growth of Ralstonia eutropha in 240 g l )1 CCS medium supplemented with different levels of nitrogen Carbon to nitrogen ratio Ammonia concentration (mg l )1 ) growth rate (h )1 ) biomass 30 : Æ19 a 6Æ a 50 : Æ25 b 4Æ b 70 : Æ20 ab 4Æ b 90 : Æ19 a 2Æ c Mean values within column not sharing common superscript differ significantly (P < 0Æ05). The optimal ph of R. eutropha is 7Æ0, although its ph tolerance extends to 8Æ5. At 24 h, ph levels in Kunioka s and the 240 g l )1 CCS media were c. 8Æ2, while ph was 8Æ5 for the nutrient broth and 8Æ7 for the 80 g l )1 CCS media. The increase in ph of the CCS medium was likely due to the consumption of organic acids. To determine if high ph levels had any effect on growth, trials with the 80 and 240 g l )1 CCS media were repeated with ph control (Table 2). The maximum cell populations and growth rates were again significantly higher for the 240 g l )1 CCS medium trials compared to the 80 g l )1 medium. However for each type of medium, there was no significant difference for either parameter, with or without ph control. Thus, we concluded that ph in the range tested was not a critical factor for cell population or growth rate. Moreover, the 240 g l )1 CCS medium was considered as the optimal medium for growth of R. eutropha, with ph control not necessary (due to intrinsic buffering). In the prior trials, we observed that cell populations in the CCS media without nitrogen supplementation peaked at 24 h. We next investigated whether supplementation with NH 4 HCO 3 would boost growth (Table 3). When nitrogen was added to the medium, cell numbers continued to increase until 48 h, peaking at higher levels. Thus, nitrogen was a limiting factor in the basal CCS medium. The highest growth rates and cell populations were achieved at the carbon to nitrogen ratio of 50 : 1. Nitrogen limitation slightly reduced cell populations at the 90 : 1 ratio, while ammonia toxicity could have reduced growth in the 30 : 1 trials. Organic acids are present in the basal CCS medium as a result of prior metabolism of bacterial contaminants and yeast during ethanol fermentation. Lactic, butyric and acetic acid present in CCS were completely utilized by 42 h for media with C : N ratios of 50 : 1 and greater, and by 78 h in the 30 : 1 medium (data not shown). Propionic acid utilization rates were much slower than that observed for the other acids, with little change in concentrations until after 36 h. In the media containing 50 : 1 and higher C : N levels, propionic acid was depleted by 118 h. Propionic acid (pka = 4Æ87) is better utilized at a ph of 7Æ5 (when added individually), due to an increase in the concentration of undissociated propionic acid as an uncoupler (Kim et al. 2005). Slightly higher growth rates and cell populations as well as higher VFA utilization efficiency at the 50 : 1 level, helped us conclude that C : N ratio of 50 : 1 was best for the growth of R. eutropha. Effects of volatile fatty acids Different VFAs were added at 48 h since prior trials showed the cell population peaked by this time. As the initial 48 h of all trials were replicates, that data were pooled to show the average growth response of R. eutropha on the CCS medium. Figure 1 is an average fermentation curve, showing cell populations and VFA utilization curves for R. eutropha in CCS medium during the initial 48 h. Table 4 provides rates of cell growth and VFA, ammonia, and phosphate utilization during the first 48 h. Concentration (g l 1 ) E E E Time (h) 1 00E+07 Figure 1 Growth and organic acid utilization by Ralstonia eutropha in the CCS medium though 48 h. The values for average cell count ( ), acetic acid utilization ( n), butyric acid utilization (e), lactic acid utilization ( h), propionic acid utilization ( ) and succinic acid utilization ( ) are indicated. Values are the mean of three replications with standard deviation showed by error bars. Cell population (CFU ml 1 ) Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 106 (2009)

5 Conversion of VFAs into PHA P. Chakraborty et al. Table 4 Growth and nutrient utilization rates of Ralstonia eutropha in a nitrogen supplemented*, 240 g l )1 CCS medium through 48 h cell population growth rate (h )1 ) Ammonia utilization rate (mg l )1 h )1 ) Phosphate utilization rate (g l )1 h )1 ) 2Æ Æ13 1Æ7 0Æ022 Acetic Butyric Lactic Propionic Succinic VFA utilization rate (g l )1 h )1 ) 0Æ033 0Æ026 0Æ054 0Æ010 0Æ021 VFA fermentation efficiency (%) 77Æ6 76Æ2 94Æ2 35Æ6 62Æ7 *Nitrogen supplemented to achieve a carbon to nitrogen ratio of 50 : 1, ammonia concentration of 200 mg l )1. The exponential phase lasted c. 16 h; however, cell numbers continued to increase at a lower rate reaching 2Æ CFU ml )1 by 48 h. Ralstonia eutropha achieved a maximum growth rate of 0Æ13 h )1 in the CCS medium in the aerated shake flasks. As the wild type strain of R. eutropha (H16) was used, the 1Æ5 2Æ0 gl )1 of glucose present in the CCS medium was not metabolized due the lack of glucose transport and phosphorylation enzymes (Kim et al. 1995; Lee et al. 1999; Madison and Huisman 1999). The generation time calculated from the growth of the organism in the CCS medium was 5Æ3 h. The CCS medium also contained g l )1 of glycerol, and this was reduced by almost 50% by 48 h (data not shown). Cell growth was also supported by metabolism of the VFAs present in CCS. Lactic acid was consumed at the fastest rate (0Æ054 g l )1 h )1 ), followed by acetic, succinic and propionic. Butyric acid was not metabolized until stationary phase. Over 94% of lactic acid was used, likely because less energy is needed to transport lactate to the central metabolic pathway (Shi et al. 1997). At least two-thirds of the acetic, butyric and succinic acids were also consumed within 48 h, while only 35% of the propionic acid was utilized. At 48 h, only minimal amounts of lactic, acetic and butyric acids were present, while c. 0Æ5 gl )1 of propionic was present. Ammonia levels fell from 200 to 60 mg l )1 by 72 h, while phosphate levels fell from 2Æ67 to 1Æ06 g l )1. Effects of acetic acid addition Figure 2 shows the average of three fermentation trials where 5 g l )1 acetic acid was added at 48 h. Trends were similar when either 1 or 3 g l )1 acetic acid were added, with cell numbers continuing to rise through 72 h. Acetic acid levels were falling immediately after addition. Succinic acid utilization was not affected by addition of acetic acid, but propionate utilization was repressed. Propionic acid utilization is ph dependent, and as other acids were used, the ph increased, allowing better utilization of propionic acid (Kim et al. 2005). Table 5 summarizes the acetic acid addition trials. The maximum cell population was directly related to the level of acetate fed to the culture. However, there was no Concentration (g l 1 ) E E E E Time (h) 1 00E+06 Figure 2 Growth and VFA utilization by Ralstonia eutropha with acetic acid fed at 5 g l )1. The values for cell count ( ), acetic acid utilization ( n), butyric acid utilization (e), lactic acid utilization ( h), propionic acid utilization ( ) and succinic acid utilization ( ) are indicated. Values are the mean of three replications with standard deviation showed by error bars. Table 5 Acetic acid feeding of Ralstonia eutropha Parameters Acetic acid added (g l )1 ) cell population 3Æ a 4Æ b 5Æ b Fermentation efficiency (%) 100 a 100 a 70Æ6 b Acid utilization rate 0Æ018 a 0Æ039 b 0Æ048 b (g l )1 h )1 ) Ammonia utilization 1Æ3 a 1Æ0 a 1Æ3 a rate (mg l )1 h )1 ) Phosphate utilization 0Æ017 b 0Æ013 b 0Æ020 a rate (g l )1 h )1 ) PHA concentration (g l )1 ) 1Æ7 a 1Æ9 a 2Æ9 a Cell dry weight (g l )1 ) 5Æ4 a 6Æ5 a 9Æ9 a PHA productivity (g l )1 h )1 ) 0Æ013 a 0Æ016 a 0Æ024 a PHA content (%) 30Æ79 a 29Æ2 a 29Æ3 a Mean values within row not sharing common superscript differ significantly (P < 0Æ05). Cell population (CFU ml 1 ) 2000 Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 106 (2009)

6 P. Chakraborty et al. Conversion of VFAs into PHA significant difference in maximum cell population between the 3 and 5 g l )1 trials, perhaps due to the eventual limitation of nitrogen at 96 h. Acid utilization rates followed a similar trend, with higher rates observed at higher acid levels. The reduced FE at the 5 g l )1 level was due to fermentation being halted at 120 h, as the trend line in Fig. 2 suggests that acetic acid would have been depleted by 140 h. The ammonia was present at 39 mg l )1 at 72 h and was almost depleted by 96 h. The concentration of phosphate 0Æ48 g l )1 at 72 h, and was utilized by 120 h. The presence of small amounts of ammonia and phosphate corresponded with the lack of additional growth after 72 h. As expected we found that the utilization rates of both ammonia and phosphate was higher in the first 72 h, as the nutrients were directed towards anabolism (data not shown). The highest PHA concentration, dry cell weight and PHA productivity were obtained when acetic acid was fed at 5 g l )1, but there were no significant difference between treatments. Based on trials with the CCS basal medium, feeding 5 g l )1 acetic acid provided the best results. Although FE was highest at 3 g l )1, acetic acid would also have been completely utilized had the 5 g l )1 trial been extended. Effects of butyric acid addition Figure 3 shows the average of three fermentation trials where 5 g l )1 butyric acid was added at 48 h. The trends were similar in trials when either 1 or 3 g l )1 butyric acid were added, with cell numbers continuing to rise through 48 h, and butyric acid levels falling immediately after addition. Succinic and propionic acid utilization followed similar trends as observed during acetic acid addition trials. Concentration (g l 1 ) E E E E E+06 Time (h) Figure 3 Growth and VFA utilization by Ralstonia eutropha with butyric acid fed at 5 g l )1. The values for cell count ( ), acetic acid utilization ( D), butyric acid utilization (e), lactic acid utilization ( h), propionic acid utilization ( ) and succinic acid utilization ( ) are indicated. Values are the mean of three replications with standard deviation showed by error bars. Cell population (CFU ml 1 ) Table 6 summarizes the results when butyric acid was fed at different levels. The highest cell population was reached when the organism was fed butyric acid at 5gl )1, as expected. Butyric acid consumption rates were significantly higher at the 5 g l )1 feeding rate, but FE was similar at the 1 and 5 g l )1 feeding rate. Ammonium and phosphate utilization rates decreased as the cells reached stationary phase, and followed the same trends as observed in the acetic acid feeding trials. Polyhydroxyalkanoate concentration, productivity, PHA content and dry cell weight were directly proportional to the amount of butyric acid fed, although dry cell weight was not significantly different. The acid was almost completely utilized at the 5 g l )1 feeding level, resulting in the highest rate and extent of PHA production. The PHA concentration and productivity were significantly higher at 5 g l )1. Effects of lactic acid addition Figure 4 shows the average of three fermentation trials where 8 g l )1 lactic acid was added at 48 h. The trends were similar when either 2 or 4 g l )1 lactic acid were added, with cell numbers rising through 72 h. Levels of lactic acid were falling immediately after addition. One difference with the prior acetic and butyric trials was that the batch of CCS used in these trials contained higher initial levels of lactic acid, such that the total initial organic acid concentration was 4Æ9 gl )1. This reduced the growth rate, so that the 48-h cell population was lower. Succinic and propionic acid utilization were both repressed until the majority of lactic acid was consumed. Table 7 summarizes the results when lactic acid was fed at different levels. The highest cell population was reached Table 6 Butyric acid feeding of Ralstonia eutropha Parameters Butyric acid added (g l )1 ) cell population 5Æ a 4Æ b 6Æ a Fermentation efficiency (%) 100 a 95Æ6 b 98Æ9 ab Acid utilization rate 0Æ022 a 0Æ041 a 0Æ068 b (g l )1 h )1 ) Ammonia utilization 1Æ4 a 1Æ4 a 1Æ3 a rate (mg l )1 h )1 ) Phosphate utilization 0Æ011 a 0Æ016 b 0Æ017 b rate (g l )1 h )1 ) PHA concentration (g l )1 ) 0Æ4 a 1Æ5 a 4Æ6 b Cell dry weight (g l )1 ) 9Æ8 a 8Æ8 a 14Æ5 a PHA productivity (g l )1 h )1 ) 0Æ003 a 0Æ013 b 0Æ037 c PHA content (%) 3Æ9 a 17Æ4 b 31Æ9 b Mean values within row not sharing common superscript differ significantly (P < 0Æ05). Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 106 (2009)

7 Conversion of VFAs into PHA P. Chakraborty et al. Concentration (g l 1 ) E E E E E+06 Time (h) Figure 4 Growth and VFA utilization by Ralstonia eutropha with lactic acid fed at 8 g l )1. The values for cell count ( ), acetic acid utilization ( n), butyric acid utilization (e), lactic acid utilization ( h), propionic acid utilization ( ) and succinic acid utilization ( ) are indicated. Values are the mean of three replications with standard deviation showed by error bars. Table 7 Lactic acid feeding of Ralstonia eutropha Parameters Lactic acid added (g l )1 ) cell population 4Æ a 6Æ b 5Æ c Fermentation efficiency (%) 100 a 100 a 70Æ7 b Acid utilization rate 0Æ037 ab 0Æ055 bc 0Æ080 c (g l )1 h )1 ) Ammonia utilization 1Æ0 a 1Æ1 a 1Æ0 a rate (mg l )1 h )1 ) Phosphate utilization 0Æ013 a 0Æ010 a 0Æ010 a rate (g l )1 h )1 ) PHA concentration (g l )1 ) 0Æ1 a 0Æ8 a 2Æ4 b Cell dry weight (g l )1 ) 3Æ5 a 4Æ0 a 6Æ0 a PHA productivity (g l )1 h )1 ) 0Æ001 a 0Æ006 a 0Æ020 b PHA content (%) 4Æ0 a 17Æ0 b 40Æ7 c Mean values within row not sharing common superscript differ significantly (P < 0Æ05). when the organism was fed 4 g l )1 lactic acid. The 8 g l )1 level may have resulted in metabolic inhibition that reduced growth (Tsuge et al. 2001). When lactic acid was fed at 2 and 4 g l )1, it was completely used by the end of the trial, compared to 70% FE at 8 g l )1. Again, the high acid level in the 8 g l )1 trial appears to have inhibited metabolism. Lactic acid consumption rates were similar at the 4 and 8 g l )1 levels. The rates of ammonia and phosphate utilization were somewhat slower than those observed when acetic and butyric acids were fed. Polyhydroxyalkanoate concentrations and productivities, along with cell dry weights, were substantially lower for the 2 and 4 g l )1 lactic acid levels, compared to acetic and butyric acid. The 8 g l )1 addition level showed improvement, especially in PHA content, even though the low cell numbers limited PHA concentration. The PHA Cell population (CFU ml 1 ) concentration, productivity and PHA content were significantly higher at 8 g l )1 addition. These data suggest that the optimal level of lactic acid falls between 4 and 8 g l )1. Effects of propionic acid addition Figure 5 shows the average of three fermentation trials where 5 g l )1 propionic acid was added at 48 h. Similar trends were observed in the trials when either 1 or 3 g l )1 propionic acid was added. The cell numbers increased through 72 h, and propionic acid was used immediately after it was fed. Utilization of acetic and succinic acids was repressed by the addition of propionic acid, but the other organic acids present in CCS were utilized. The results obtained by feeding different levels of propionic acid, shown in Table 8, indicate that the highest Concentration (g l 1 ) Time (h) 1 00E E E E E+06 Figure 5 Growth and Acid utilization by Ralstonia eutropha with propionic acid fed at 5 g l )1. The values for cell count ( ), acetic acid utilization ( n), butyric acid utilization (e), lactic acid utilization ( h), propionic acid utilization ( ) and succinic acid utilization ( ) are indicated. Values are the mean of three replications with standard deviation showed by error bars. Table 8 Propionic acid feeding of Ralstonia eutropha Parameters Propionic acid added (g l )1 ) cell population 2Æ a 4Æ b 6Æ c Fermentation efficiency (%) 93Æ3 a 96Æ3 a 68Æ6 b Acid utilization rate (g l )1 h )1 ) 0Æ024 a 0Æ042 b 0Æ046 b Ammonia utilization 0Æ8 a 0Æ9 a 0Æ9 a rate (mg l )1 h )1 ) Phosphate utilization 0Æ013 a 0Æ010 a 0Æ013 a rate (g l )1 h )1 )) PHA concentration (g l )1 ) 0Æ7 a 2Æ1 a 4Æ3 b Cell dry weight (g l )1 ) 6Æ0 a 10Æ0 ab 14Æ0 bc PHA productivity (g l )1 h )1 ) 0Æ005 a 0Æ020 b 0Æ036 c PHA content (%) 10Æ2 a 20Æ7 b 29Æ3 c Mean values within column not sharing common superscript differ significantly (P < 0Æ05). Cell population (CFU ml 1 ) 2002 Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 106 (2009)

8 P. Chakraborty et al. Conversion of VFAs into PHA Table 9 Comparison of volatile fatty acid feeding of Ralstonia eutropha at optimal levels Parameters Volatile fatty acid Acetic Butyric Lactic Propionic cell population 5Æ a 6Æ ab 5Æ a 6Æ b Fermentation efficiency (%) 70Æ6 a 95Æ6 b 70Æ7 a 68Æ6 a Acid utilization rate (g l )1 h )1 ) 0Æ048 a 0Æ041 a 0Æ080 b 0Æ046 a PHA concentration (g l )1 ) 2Æ9 ab 4Æ6 a 2Æ4 b 4Æ3 a Cell dry weight (g l )1 ) 9Æ9 ab 14Æ5 b 6Æ0 a 14Æ7 b PHA productivity (g l )1 h )1 ) 0Æ024 ab 0Æ037 a 0Æ020 b 0Æ036 a PHA content (%) 29Æ2 a 31Æ9 a 40Æ7 a 29Æ3 a Mean values within column not sharing common superscript differ significantly (P < 0Æ05). cell population and acid utilization rate were reached at the 5 g l )1 propionic acid level. However, the 5 g l )1 propionic acid level resulted in significantly lower FE than that at 1 and 3 g l )1. Ammonia and phosphate utilization were again somewhat reduced, as in the lactic acid trials. Polyhydroxyalkanoate concentration, productivity, PHA content and cell dry weights were highest at 5 g l )1 of propionic acid addition. Comparison of acetic, butyric, lactic and propionic acids The best levels of acetic, butyric, lactic and propionic acids were compared, based on the parameters shown in Table 9. Butyrate was best in terms of all the parameters, with propionic acid yielding similar results, except that FE was lower. Lactic acid resulted in the lowest cell population, dry cell weight and PHA concentration. Acetic acid resulted in intermediate cell population, dry cell weight, acid consumption rate and PHA content. But PHA concentration and productivity were low. There were no significant differences between the ammonia utilization rates for the different treatments, but acetic and butyric acid additions showed somewhat faster utilization rates for phosphates (data not shown). In all trials, there were only small amounts of ammonia or phosphate left in the media at the end of each trial. Discussion These findings show that CCS at a concentration of 240 g l )1 can be used as an inexpensive medium for growth of R. eutropha. The growth rates in Kunioka s medium and nutrient broth were not significantly different from 240 g l )1 CCS media. This might be due to the presence of similar utilizable compounds present in the CCS medium as a result of enzymatic degradation and byproducts of yeast fermentation. CCS levels of 400 and 700 g l )1 failed to support growth, perhaps due to a combination of high osmotic pressure, resulting from levels of dissolved nutrients (Madison and Huisman 1999). For example, the 400 g l )1 CCS medium contained c. 3Æ4 gl )1 glucose, 40 g l )1 glycerol and 6 g l )1 total organic acids, while the levels in the 700 g l )1 CCS medium were 5Æ9 gl )1 glucose, 70 g l )1 glycerol and 10Æ4 gl )1 total organic acids. We found that the intrinsic buffering capacity of CCS was sufficient to maintain ph in an appropriate range without active ph control. Nitrogen is a critical nutrient in the growth phase for anabolic reactions, but nitrogen deficiency is typically used to trigger PHA production when excess carbon is present (Kim et al. 1994). When nitrogen supplementation to the CCS medium was evaluated, a C : N ratio of 50 : 1 was found to result in optimal growth rate, maximum cell numbers and acid utilization rates. Therefore, we concluded the optimum medium formulation to be CCS at 240 g l )1 with C : N ratio of 50 : 1 with an initial ph of 7. Ammonia and phosphate levels were monitored during the initial growth phase in the trials evaluating fed batch addition of VFAs. Ideally, one or both of these nutrients becomes limiting at the end of exponential phase, to trigger the shift from reproductive metabolism to PHA synthesis (Madison and Huisman 1999). Because nitrogen and phosphate were not depleted until 72 h, this could have contributed to the continued increase in cell numbers observed after 48 h. It is likely that at least some of the VFAs fed at 48 h were utilized for growth, until the point at which nitrogen became limiting. Researchers have found that the complete lack of nitrogen may suppress enzyme activity in PHA synthesis (Shimizu et al. 1999). Thus a small amount of ammonia in the media might be necessary to trigger PHA synthesis. It is also important to maintain phosphorous and magnesium levels at 0Æ35 g l )1 and 10 mg l )1 respectively (Asenjo et al. 1995). So the presence of residual phosphate in the medium might help improve PHA production. Ralstonia eutropha can metabolize several VFAs as additional carbon sources under nitrogen deficiency, but these can be toxic, depending on concentration and ph. At ph levels below the pkas for VFAs (lactic 3Æ86, acetic 4Æ76, Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 106 (2009)

9 Conversion of VFAs into PHA P. Chakraborty et al. butyric 4Æ83, propionic 4Æ87), the undissociated form predominates, and these readily cross the cell membrane. Once inside they rapidly dissociate and acidify the cytoplasm (Salmond et al. 1984). At high VFA levels, this can reduce the proton gradient across the membrane, increase osmotic pressure, and reduce acid utilization rate, growth rate and yield (Lawford and Rousseau 1993; Axe and Bailey 1995). At ph levels closer to the optimum for R eutropha (c. 7Æ0), VFAs would be in the dissociated form in the medium. While the anions would not be transported as readily, once inside the cell they would not cause the adverse effects of the undissociated form. For example, Kim et al. (2005) noted that propionic acid is better utilized at a ph of c. 7Æ5. VFA concentration is also important at more neutral ph levels, as Chung et al. (1997) showed that 5 g l )1 propionate produced less PHA than 2 3 g l )1 propionate at ph 7Æ5. Therefore, VFAs can only be effective carbon sources when ph and VFA concentration are carefully regulated. In this study, Ralstonia eutropha was able to utilize acetic, butyric, lactic and propionic acids when fed at concentrations up to 5 g l )1. However, when the fatty acids were present at high concentrations, a reduction in initial growth and FE were observed, which ultimately resulted in lower PHA production. For example, the lactic acid supplementation trials resulted in the lowest cell population, dry cell weight and PHA concentration. We theorize that this was occurred because these trials contained the highest initial concentration of organic acids. When grown on a mixture of acetate, lactate and butyrate, Ralstonia eutropha used lactate first, because a large amount of ATP is required for transport of acetate and butyrate (Shi et al. 1997). In this study, the apparent highest acid utilization rates of lactic acid as compared to all other VFAs further supports the fact. With respect to PHA production, butyric acid yielded the best results compared to the other VFAs. This correlates with the findings that butyrate is more energetically efficient than acetate and lactate for PHB production. The glyoxalate pathway is required for growth of R. eutropha on butyrate, and it competes for butyrate with the PHB biosynthetic pathway. When the nitrogen source was depleted, R. eutropha shifts from growth to PHB production. The lack of nitrogen blocks amino acid synthesis pathways and NADH consumption rate decreases. However, there is no significant difference in flux of butyrate into the TCA cycle, where isocitrate dehydrogenase generates NADH. Ralstonia eutropha can direct 67% of butyrate into the TCA cycle, whereas only 33% of either acetate or lactate can enter. The accumulating NADH is redirected to PHB biosynthesis where it is used to convert acetoacetyl coa to (R)-3-hydroxybutyryl-coA (Shi et al. 1997). Researchers have reported that inexpensive carbon sources generally result in lower growth rates of PHA producing organisms, due to inefficient use of certain nutrients (Lee et al. 1999; Tsuge et al. 2001; Tsuge 2002). In this study, the wild-type strain of R. eutropha was not able to utilize the glucose and did not completely use the high amount of glycerol present in the CCS medium. Also the presence of high concentrations of organic acids can reduce growth rates, and levels below 1 g l )1 are recommended (Gorenflo et al. 2001; Tsuge et al. 2001). Therefore, most studies that use VFAs as additional carbon sources only add the organic acids after the organism had reached optimum growth, to trigger PHA granule formation (Guocheng et al. 2001; Wang and Yu 2001; Yu et al. 2002). The initial VFA concentration of CCS was about 3 4 g l )1, and this might have resulted in overall lower growth rates. Moreover, the trials were carried out in 1-l shake flasks, in which did not permit precisely controlled aeration. The cumulative effect of these factors might have lowered the PHA concentration and productivity. Recombinant microbes can be used to increase growth rate, as they can utilize the nutrients more efficiently. Substrate utilization genes can be inserted in PHA producers, or PHA biosynthesis genes can be inserted in organisms that have wide range of utilizable substrates. For example, the recombinant strains of R. eutropha (H16) containing glucose-utilizing genes of Escherichia coli, and E. coli harbouring the R. eutropha genes had higher PHA productivity and concentration as compared to the wild-type strain (Lee 1996a; Lee et al. 1999). Molecular engineering can also be used to improve metabolic pathways so as to increase PHA synthesis and or production of new types of PHAs. As the carbon source is a major contributor to PHA cost, inexpensive sources of carbon are important (Lee 1996b). From an economical point of view, the use of purified media to increase PHA yield will significantly increase the production cost. This study showed that R. eutropha is capable of growing in the 240 g l )1 CCS medium and the growth rate is comparable to Kunioka s medium. We also concluded that the VFAs were not inhibitory, provided they were added at the stationary phase of growth. VFAs can be added at high levels once the organism attained its maximum growth, and these acids can be efficiently diverted towards PHA production by R. eutropha. As the use of pure VFAs would also be cost prohibitive, we are separately developing a mixed culture system to produce a mixture of VFAs from another ethanol production byproduct. The data obtained from this VFA study can be translated to determine the feeding rate of mixed VFAs to maximize PHA production Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 106 (2009)

10 P. Chakraborty et al. Conversion of VFAs into PHA Acknowledgements This project was funded by South Dakota Corn Utilization Council and the South Dakota Agricultural Experiment Station. References Asenjo, J.A., Schmidt, A.S., Anderson, P.R. and Andrews, B.A. (1995) Effect of single nutrient limitation on polyb-hydroxybutyrate molecular weight distribution in Alcaligenes eutrophus. Biotechnol Bioeng 46, Axe, D.D. and Bailey, J.E. (1995) Transport of lactate and acetate through the energized cytoplasmic membrane of Escherichia coli. Biotechnol Bioeng 47, Braunegg, G., Sonnleitner, B. and Lafferty, R.M. (1978) A rapid gas chromatographic method for the determination of poly-b-hydroxybutyric acid in microbial biomass. Eur J Appl Microbiol 6, Chung, Y.J., Cha, H.J., Yeo, J.S. and Yoo, Y.J. (1997) Production of poly(3-hydroxybutyric-co-3hydroxyvaleric) acid using propionic acid by ph regulation. J Ferm Bioeng 83, Comeau, Y., Hall, K.J. and Oldha, W.K. (1988) Determination of poly-b-hydroxybutyrate and poly-b-hydroxyvalerate in activated sludge by gas-liquid chromatography. Appl Environ Microbiol 54, Gorenflo, V., Schmack, G., Vogel, R. and Steinbuchel, A. (2001) Development of a process for the biotechnological large scale production of 4-hydroxyvalerate containing polyesters and characterization of their physical and mechanical properties. Biomacromology 2, Guocheng, D., Chen, J., Yu, J. and Lun, S. (2001) Continuous production of poly-3-hydroxybutyrate by Ralstonia eutropha in a two stage culture system. J Biotechnol 88, Kim, B.S., Lee, S.C., Lee, S.Y., Chang, H.N., Chang, Y.K. and Woo, S.I. (1994) Production of poly (3-hydroxybutyric acid) by fed batch culture of Alcaligenes eutrophus with glucose concentration control. Biotechnol Bioeng 43, Kim, H.Y., Park, J.S., Shin, H.D. and Lee, Y.H. (1995) Isolation of glucose utilizing mutant of Alcaligenes. eutrophus, its substrate selectivity, and accumulation of polyb-hydroxybutyrate. J Microbiol 33, Kim, J.H., Kim, B.G. and Choi, C.Y. (2005) Effect of propionic acid on poly (b-hydroxybutyric-co-hydroxyvaleric acid production by Alcaligenes eutrophus. Biotechnol Lett 14, Kunioka, M., Nakamura, Y. and Doi, Y. (1989) New bacterial co-polyesters produced in Alcaligenes eutrophus from organic acids. Polym Commun 29, Lawford, H.G. and Rousseau, J.D. (1993) Effects of ph and acetic acid on glucose and xylose metabolism by genetically engineered ethanologenic Escherichia coli by weak acids. Appl Biochem Biotechol 39, Lee, S.Y. (1996a) Bacterial polyhydroxyalkanoates. Biotechnol Bioeng 49, Lee, S.Y. (1996b) Plastic bacteria? Progress and prospect for polyhydydroxyalkanoate production in bacteria. Tibtech 14, Lee, S.Y., Choi, J.I. and Wong, H.H. (1999) Recent Advances in polyhydroxyalkanoate production by bacterial fermentation: mini review. Int J Biol Macromol 25, Madison, L.L. and Huisman, G.W. (1999) Metabolic engineering of poly (3-hydroxyalkanoates): from DNA to plastic. Microbiol Mol Biol Rev 63, Nishida, H. and Tokiwa, Y. (1992) Effects of higher order structure of poly (3-hydroxybutyrate) on its biodegradation I. Effects of heat treatment on microbial degradation. J Appl Polym Sci 46, Rutherford, D.R., Hammer, W.J. and Babu, G.N. (1997) Poly(b-hydroxybutyrate) Pressure Sensitive Adhesive Compositions. U.S. Patent (March 1997). Salmond, C.V., Kroll, R.G. and Booth, I.R. (1984) The effect of food preservatives on ph homeostasis in Escherichia coli. J Gen Microbiol 130, Shi, H., Shiraishi, M. and Shimizu, K. (1997) Metabolic flux analysis of biosynthesis of poly(b-hydroxybutyric acid) in Alcaligenes eutrophus from various carbon sources. J Ferm Bioeng 84, Shimizu, H., Kozaki, Y., Kodama, H. and Shioya, S. (1999) production strategy for biodegradable copolymer P(HB-co-HV) in fed-batch culture of Alcaligenes eutrophus. Biotechnol Bioeng 62, Tsuge, T. (2002) Metabolic improvements and use of inexpensive carbon sources in microbial production of polyhydroxyalkanoates. J Biosci Bioeng 94, Tsuge, T., Tanaka, K. and Ishizaki, A. (2001) Development of a novel method for feeding a mixture of L-lactic acid and acetic acid in fed batch culture of Ralstonia eutropha for poly-d-3-hydroxybutyrate formation. J Biosci Bioeng 91, Wang, J. and Yu, J. (2001) Kinetic analysis on formation of poly (3-hydroxybutyrate) from acetic acid by Ralstonia eutropha under chemically defined conditions. J Ind Microbiol Biotechnol 26, Yu, J., Si, Y., Keung, W. and Wong, R. (2002) Kinetics of modeling of inhibition and utilization of mixed volatile fatty acids in the formation of polyhydroxyalkanoates by Ralstonia eutropha. Proc Biochem 37, Journal compilation ª 2009 The Society for Applied Microbiology, Journal of Applied Microbiology 106 (2009)