APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Aug. 2000, p. 3624–3627 Vol. 66, No.

8
0099-2240/00/$04.00⫹0
Copyright © 2000, American Society for Microbiology. All Rights Reserved.

Production of Poly(3-Hydroxybutyrate) by Fed-Batch

Culture of Recombinant Escherichia coli with
a Highly Concentrated Whey Solution
WOO SUK AHN, SI JAE PARK, AND SANG YUP LEE*
Department of Chemical Engineering and BioProcess Engineering Research Center,
Korea Advanced Institute of Science and Technology, Yusong-gu,
Taejon 305-701, Korea
Received 27 March 2000/Accepted 31 May 2000

Fermentation strategies for the production of poly(3-hydroxybutyrate) (PHB) from whey by recombinant
Escherichia coli strain CGSC 4401 harboring the Alcaligenes latus polyhydroxyalkanoate (PHA) biosynthesis
genes were developed. The pH-stat fed-batch cultures of E. coli CGSC 4401 harboring pJC4, a stable plasmid
containing the A. latus PHA biosynthesis genes, were carried out with a concentrated whey solution containing
280 g of lactose equivalent per liter. Final cell and PHB concentrations of 119.5 and 96.2 g/liter, respectively,
were obtained in 37.5 h, which resulted in PHB productivity of 2.57 g/liter/h.

Whey is a major by-product in the manufacturer of cheese or the production of PHB from whey (18). Even though a rela-
casein from bovine milk, representing 80 to 90% of the volume tively high concentration of PHB (69 g/liter) could be pro-
of milk transformed. It contains approximately 4.5% (wt/vol) duced, cell broth had to be intermittently removed due to the
lactose, 0.8% (wt/vol) protein, 1.0% (wt/vol) salts, and 0.1 to volumetric limitation of the fermentor caused by the low sol-
0.8% (wt/vol) lactic acid (18). Only half of the whey produced ubility of lactose in the feeding solution (ca. 210 g of lactose
annually in the United States is recycled into useful products per liter) (18).
such as food ingredients and animal feed, and the rest is re- In this study, we report the fermentation strategies for the
garded as a pollutant due to its high biological oxygen demand. production of PHB from whey in recombinant E. coli without
Disposal of whey is being managed at considerable cost. Poly- removing culture broth during fermentation. A highly concen-
hydroxyalkanoates (PHAs) are polyesters, which are accumu- trated whey solution was successfully employed for the efficient
lated as energy and/or carbon storage materials by numerous production of PHB from whey to a high concentration by
microorganisms, usually when a nutritional component such as recombinant E. coli.
nitrogen, phosphorus, sulfur, oxygen, or magnesium is limited Experimental procedures. E. coli CGSC 4401 (E. coli Ge-
in the presence of an excess carbon source (1, 10, 11, 14, 16). netic Stock Center, New Haven, Conn.), CGSC 3121, CGSC
PHAs have been considered to be good substitutes for petro- 2507, DSM 499 (German Collection of Microorganisms and
leum-derived synthetic plastics because of their similar mate- Cell Cultures, Braunschweig, Germany), and KCTC 2223 (Ko-
rial properties to synthetic polymers and complete biodegrad- rean Collection for Type Cultures, Taejon, Korea) were used
ability after disposal (3). A major problem in commercializing in this study. The plasmid pJC4 containing the A. latus PHA
PHAs is their high production cost. Much effort has been biosynthesis genes has been described previously (6). E. coli
devoted to lower the production cost of PHA by developing strains were transformed with pJC4 by electroporation (7).
better bacterial strains and efficient strategies for fermentation Cells were maintained as a 15% (vol/vol) glycerol stock at
and recovery of PHAs (5, 10, 12). Among these bacterial ⫺75°C after growing in Luria-Bertani (LB) medium (pH 6.7)
strains, recombinant Escherichia coli strains harboring the Ral- or chemically defined MR medium (described below) contain-
stonia eutropha and Alcaligenes latus PHA biosynthesis genes ing 20 g of lactose per liter.
have been successfully employed for the production of poly Bovine whey powder was obtained from SamIk Co., Seoul,
(3-hydroxybutyrate) (PHB) to a high concentration with high Korea. Crude whey solution was prepared by dissolving 700 g
productivity (6, 15, 17). Economic evaluation of the process for of whey powder in 1 liter of distilled water. To remove exces-
the production of PHB suggested that the major contributor to sive proteins in whey solution, the pH of the whey solution was
the overall PHB production cost was carbon substrate cost (up adjusted to 4.5 by the addition of 37% (wt/vol) HCl (19). The
to 50%) (4). Therefore, it is desirable to produce PHB from solution was autoclaved at 121°C for 15 min and centrifuged at
cheap carbon source or even from a waste product such as 11,000 ⫻ g in a sterilized bottle for 15 min to remove aggre-
whey by using recombinant E. coli. Several researchers re-
gates. By adding diatomaceous earth (Sigma Co., St. Louis,
ported that PHB could be produced from whey-based medium
Mo.) to 2% (wt/vol), small protein particles could be removed
by recombinant E. coli in flask culture (8, 9, 13). Recently, we
by filtration with Whatman no. 3 filter paper (Whatman Co.,
carried out high-cell-density culture of a recombinant E. coli
Maidstone, England). The pH of the filtered solution was ad-
strain harboring the R. eutropha PHA biosynthesis genes for
justed to 6.5 with 12 N NaOH.
Flask cultures were carried out in a 250-ml flask containing
100 ml of MR medium in a shaking incubator at 30°C and 200
* Corresponding author. Mailing address: Department of Chemical
Engineering and BioProcess Engineering Research Center, Korea Ad- rpm. Whey powder (40 g/liter) was added as a carbon source in
vanced Institute of Science and Technology, 373-1 Kusong-dong, Yu- flask culture. The MR medium (pH 6.9) contains (per liter)
song-gu, Taejon 305-701, Korea. Phone: 82-42-869-3930. Fax: 82-42- 6.67 g of KH2PO4, 4 g of (NH4)2HPO4, 0.8 g of MgSO4 䡠 7H2O,
869-3910. E-mail: [email protected]. 0.8 g of citric acid, and 5 ml of trace metal solution. The trace

3624
VOL. 66, 2000 POLY (3-HYDROXYBUTYRATE) PRODUCED BY E. COLI WITH WHEY 3625

metal solution contains (per liter of 5 M HCl) 10 g of FeSO4 䡠 TABLE 1. PHB production from whey in flask cultures of various
7H2O, 2 g of CaCl2, 2.2 g of ZnSO4 䡠 7H2O, 0.5 g of MnSO4 䡠 recombinant E. coli strains (pJC4)
4H2O, 1 g of CuSO4 䡠 5H2O, 0.1 g of (NH4)6Mo7O24 䡠 4H2O, Cell concn PHB concn PHB content
and 0.02 g of Na2B4O7 䡠 10H2O. Two different feeding solu- E. coli strain
(g/liter) (g/liter) (wt%)
tions were used for the fed-batch cultures. In fermentation A,
a concentrated whey solution (containing 210 g of lactose CGSC 4401 6.6 5.0 76
equivalent per liter) plus 4.5 g of MgSO4 䡠 7H2O per liter was CGSC 2507 6.8 3.4 50
CGSC 3121 3.0 1.3 43
used. In fermentations B and C, highly concentrated whey DSM 499 6.1 2.7 45
solution (containing 280 g of lactose equivalent per liter) plus KCTC 2223 6.2 0.57 9.2
6 g of MgSO4 䡠 7H2O per liter was used. Feeding solutions
were prepared as described above.
With the recombinant E. coli strain CGSC 4401(pJC4), fed-
batch cultures were carried out at 30°C in a 6.6-liter jar fer- using the whey solution containing 210 g of lactose per liter
mentor (Bioflo 3000; New Brunswick Scientific Co., Edison, (fermentation A). During the cultivation, the DOC was ini-
N.J.) containing 1.3 liter of MR medium plus pretreated whey tially maintained at 30%. When the OD600 reached 180 (cell
solution equivalent to 20 g of lactose per liter. Seed cultures concentration of ca. 60 g/liter), the DOC was maintained at
(130 ml) were prepared in the same medium. The culture pH 20% until the end of cultivation. In 49 h, cell and PHB con-
was controlled at 6.95 by the automatic addition of 28% (vol/ centrations had reached 83.1 and 46.8 g/liter, respectively, re-
vol) NH4OH. The level of dissolved oxygen concentration sulting in a PHB content of 56.3 wt% and a productivity of 1.15
(DOC) was controlled by automatically increasing the agita- g/liter/h. Because the lactose concentration in the feeding so-
tion speed up to 1,000 rpm and varying the pure oxygen per- lution was somewhat low (210 g/liter), a relatively large volume
centage. Nutrient feeding solution was added by the pH-stat of nutrient feeding solution (7.4 liter) was added to the fer-
feeding strategy. When the pH rose to a value greater than its mentor to achieve high cell density. Due to the volumetric
set point (6.95) by 0.1 due to the depletion of lactose, an limitation of the fermentor caused by the large volume of
appropriate volume of feeding solution was automatically feeding solution, culture broth had to be removed to prevent
added to increase the lactose concentration in the culture flooding during the cultivation. When lactose was depleted, 2,
broth to 20 g/liter. The feeding volume was calculated online 2.5, and 2 liters of culture broth were sequentially removed
with the fermentation software AFS3.42 (New Brunswick Sci- from the fermentor. During the entire cultivation, 7.4 liters of
entific Co.). Foam formation was suppressed by adding Anti- nutrient feeding solution was added and 6.5 liters of culture
foam 289 (Sigma Chemical Co., St. Louis, Mo.) during the broth was removed.
initial stage of fed-batch cultures. Fed-batch culture with highly concentrated whey solution
Cell growth was monitored by measuring the optical density containing 280 g of lactose per liter. In fermentation A, a large
at 600 nm (OD600; DU Series 600 Spectrophotometer; Beck- volume of culture broth had to be removed due to the volu-
man, Fullerton, Calif.). The PHB concentration was deter- metric limitation of the fermentor caused by the low concen-
mined by measuring the concentration of 3-hydroxybutyric acid tration of the carbon source (210 g of lactose per liter), which
methyl ester, which was prepared by methanolysis of PHB, by resulted in a relatively low cell concentration and PHB pro-
gas chromatography (Donam Co., Seoul, Korea) with a fused ductivity. To solve this problem, we examined the possibility of
silica capillary column (SPB-5, 30 m by 0.32 mm [inside diam- further concentrating the whey solution. In our previous study
eter], 0.25-␮m-thick film; Supelco, Bellefonte, Pa.) with ben- (18), the whey solution was concentrated based on the solubil-
zoic acid as an internal standard (2). Cell concentration, de- ity data for lactose in water (200 to 210 g/liter). It was reasoned
fined as dry weight of cells per liter of culture broth, was that the actual solubility of lactose in whey solution might be
determined as previously described (14). The residual cell con- different from this value. It was found that whey solution could
centration was defined as the cell concentration minus the be highly concentrated to contain 280 g of lactose equivalent
PHB concentration. The PHB content (percentage of weight) per liter by pretreatment of whey solution. The feeding solu-
was defined as the percentage of the ratio of PHB concentra- tion containing 280 g of lactose per liter was sufficient to allow
tion to cell concentration. The concentrations of lactose, ga- us to achieve a high density of cells and high concentration of
lactose, and glucose were measured by high-performance liq- PHB with a higher cell yield than that obtained with whey
uid chromatography (L-4200 UV-Vis detector, L-600 pump, solution containing 210 g of lactose per liter (cell yield in-
D-2500 chromato-integrator; Hitachi, Tokyo, Japan) equipped creased from 0.42 to 0.52 g/g of lactose) without removing
with an ion-exchange column (Aminex HPX-87H, 300 mm by culture broth during fermentation. Even though lactose was
7.8 mm; Hercules, Calif.) using 0.01 N H2SO4 as a mobile concentrated to a level greater than its water solubility (210
phase. g/liter) at room temperature, whey solution containing 280 g of
Flask cultures. E. coli CGSC 4401, CGSC 3121, CGSC 2507, lactose per liter was stable at room temperature, which seemed
DSM 499, and KCTC 2223 harboring pJC4 were cultivated for to be mainly due to the presence of impurities, such as the
96 h in chemically defined MR medium supplemented with remaining proteins and salts in whey solution. Lactose precip-
40 g of whey powder per liter at 30°C to examine the efficiency itates appeared in oversaturated whey solution containing
of PHB production from whey. The results of flask cultures are 280 g of lactose per liter, only when it was deposited in a ⫺4°C
summarized in Table 1. The highest cell and PHB concentra- refrigerator for longer than 1 week. Concentrated whey solu-
tions of 6.6 and 5.0 g/liter, respectively, were obtained when E. tion can be easily prepared on a large scale by evaporation. For
coli CGSC 4401(pJC4) was cultivated in MR plus 40 g of whey example, a falling film evaporator, which is widely employed in
per liter. From these flask culture results, recombinant E. coli the milk industry, can be used. The pH-stat fed-batch culture
CGSC 4401(pJC4) was chosen as the strain for PHB produc- of E. coli CGSC 4401 harboring pJC4 was carried out by using
tion from whey. highly concentrated whey solution containing 280 g of lactose
Fed-batch culture with whey solution containing 210 g of per liter (fermentation B). Culture broth removal during fed-
lactose per liter. Fed-batch culture of E. coli CGSC 4401 har- batch cultivation could be avoided and a higher cell concen-
boring pJC4 was carried out by the pH-stat feeding strategy tration could be achieved by using this highly concentrated
3626 AHN ET AL. APPL. ENVIRON. MICROBIOL.

FIG. 1. Time profiles of cell concentration (F), PHB concentration (■), residual cell concentration (Œ), and PHB content (E) (A) and cell growth rate (F) and
PHB synthesis rate (E) (B) during fed-batch culture of E. coli CGSC 4401(pJC4) with feeding solution containing 280 g of lactose equivalent per liter (fermentation C).

whey feeding solution. The initial DOC was maintained at 30% tent and cell concentration at the same time. In fermentations
of air saturation. When the OD600 reached 240 (cell concen- A and B, the DOC was lowered from 30% to 20% when the
tration of ca. 80 g/liter), the DOC was lowered to 20%. In cell concentration reached 70% of the final cell concentration.
fermentation B, the final cell concentration, PHB concentra- When the DOC was lowered to 20%, it was observed that the
tion, and PHB content of 102.9 g/liter, 59.6 g/liter, and 57.9 cell growth rate rapidly increased initially and then continu-
wt%, respectively, were obtained in 42 h, which resulted in a ously decreased; however, the PHB synthesis rate reincreased
productivity of 1.42 g of PHB/liter/h. (results not shown). From these results, we reasoned that ap-
A new strategy to achieve high PHB content. Even though a propriate control of the DOC could allow us to obtain higher
higher cell concentration of 102.9 g/liter was achieved without PHB content and PHB concentration. This strategy was ap-
removing culture broth by employing highly concentrated whey plied in the fed-batch culture by using highly concentrated
solution, the final PHB content was still lower than 60%. The whey solution containing 280 g of lactose per liter (fermenta-
PHB content is a very important factor, which contributes tion C). Wang and Lee (17) suggested that fed-batch cultiva-
significantly to the cost of production of PHB in large-scale tion with recombinant E. coli could be divided into two phases:
fermentation (4). A new strategy had to be developed to obtain an active growth phase during which PHB content is kept
higher PHB content. Therefore, we examined the effect of the constant at a low level and an active PHB synthesis phase
DOC during the fed-batch culture to achieve high PHB con- during which PHB is actively accumulated with the concomi-
VOL. 66, 2000 POLY (3-HYDROXYBUTYRATE) PRODUCED BY E. COLI WITH WHEY 3627

tant increase in PHB content (17). It was reported that a REFERENCES

sufficient DOC during the active growth phase was important 1. Anderson, A. J., and E. A. Dawes. 1990. Occurrence, metabolism, metabolic
to achieve high final cell and PHB concentrations. In fermen- role, and industrial uses of bacterial polyhydroxyalkanoates. Microbiol. Rev.
54:450–472.
tation C, the DOC was maintained at 40% during the active 2. Braunegg, G., B. Sonnleitner, and R. M. Lafferty. 1978. A rapid gas chro-
growth phase. By doing this, the cell growth rate and PHB matographic method for the determination of poly-␤-hydroxybutyric acid in
synthesis rate increased much higher than those obtained in microbial biomass. Eur. J. Appl. Microbiol. Biotechnol. 6:29–37.
fermentations A and B after lowering the DOC in the later 3. Byrom, D. 1987. Polymer synthesis by micro-organisms: technology and
economics. Trends Biotechnol. 5:246–250.
phase of fermentation. In another fed-batch cultivation similar 4. Choi, J., and S. Y. Lee. 1997. Process analysis and economic evaluation for
to fermentation B, the DOC was rapidly lowered from 40% to poly(3-hydroxybutyrate) production by fermentation. Bioprocess Eng. 17:
5% at the OD600 of 180 and was maintained at 5% until the 335–342.
end of fermentation. However, the final cell and PHB concen- 5. Choi, J., and S. Y. Lee. 1999. Efficient and economical recovery of poly(3-
hydroxybutyrate) from recombinant Escherichia coli by simple digestion with
trations obtained were only 79 and 63.2 g/liter, respectively chemicals. Biotechnol. Bioeng. 62:546–553.
(results not shown). Even though the PHB content (80%) was 6. Choi, J., S. Y. Lee, and K. Han. 1998. Cloning of the Alcaligenes latus
high, the cell concentration was lower than that obtained in polyhydroxyalkanoate biosynthesis genes and use of these genes for en-
fermentation B. These results suggested that the timing of hanced production of poly(3-hydroxybutyrate) in Escherichia coli. Appl. En-
viron. Microbiol. 64:4897–4903.
decreasing the DOC and how to decrease the DOC have con- 7. Dower, W. J., J. F. Miller, and C. W. Ragsdale. 1988. High efficiency trans-
siderable effects on the final cell concentration and PHB formation of Escherichia coli by high voltage electroporation. Nucleic Acids
content. To achieve a higher PHB content, the DOC was Res. 16:6127–6145.
decreased stepwise from 40% to 30% and then to 15% in 8. Fidler, S., and D. Dennis. 1992. Polyhydroxyalkanoate production in recom-
binant Escherichia coli. FEMS Microbiol. Rev. 103:231–236.
fermentation C (Fig. 1B). Whenever the DOC was decreased 9. Janes, B., J. Hollar, and D. Dennis. 1990. Molecular characterization of the
during the active PHB synthesis phase, the PHB synthesis rate poly-beta-hydroxybutyrate biosynthetic pathway of Alcaligenes eutrophus
sharply increased. The final PHB synthesis rate was much H16, p. 175–190. In E. A. Dawes (ed.), Novel biodegradable microbial
higher than that obtained in fermentations A and B. The final polymers. Kluwer Academic, Dordrecht, The Netherlands.
10. Lee, S. Y. 1996. Plastic bacteria? Progress and prospects for polyhydroxyal-
cell concentration, PHB concentration, and PHB productivity kanoate production in bacteria. Trends Biotechnol. 14:431–438.
obtained in fermentation C were 119.5 g/liter, 96.2 g/liter, and 11. Lee, S. Y. 1996. Bacterial polyhydroxyalkanoates. Biotechnol. Bioeng. 49:1–
2.57 g/liter/h, respectively, which are much higher than those 14.
reported previously (18). By optimally controlling the level of 12. Lee, S. Y. 1996. Poly(3-hydroxybutyrate) extrusion by cells of recombinant
Escherichia coli. J. Microbiol. Biotechnol. 6:147–149.
DOC, the higher PHB concentration and PHB content could 13. Lee, S. Y., A. P. J. Middelberg, and Y. K. Lee. 1997. Poly(3-hydroxybutyrate)
be achieved at the same time. production from whey using recombinant Escherichia coli. Biotechnol. Lett.
In this study, we have demonstrated that environmentally 19:1033–1035.
friendly PHB polymer can be efficiently produced from envi- 14. Lee, S. Y., and H. N. Chang. 1994. Effect of complex nitrogen source on the
synthesis and accumulation of poly(3-hydroxybutyric acid) by recombinant
ronmentally polluting whey with high PHB content and pro- Escherichia coli in flask and fed-batch cultures. J. Environ. Polym. Degrad.
ductivity by the pH-stat fed-batch culture with a highly con- 2:169–176.
centrated whey solution. It was also found that higher cell and 15. Schubert, P., A. Steinbüchel, and H. G. Schlegel. 1988. Cloning of the
PHB concentrations could be obtained by optimal timing of Alcaligenes eutrophus genes for synthesis of poly-␤-hydroxybutyric acid
(PHB) and synthesis of PHB in Escherichia coli. J. Bacteriol. 170:5837–5847.
the DOC reduction. The strategies developed in this study 16. Steinbüchel, A. 1992. Biodegradable plastics. Curr. Opin. Biotechnol. 3:291–297.
should provide an industrially attractive solution to whey dis- 17. Wang, F., and S. Y. Lee. 1997. Production of poly(3-hydroxybutyrate) by
posal and utilization problems by allowing economical produc- fed-batch culture of filamentation-suppressed recombinant Escherichia coli.
tion of higher-value biodegradable polymer. Appl. Environ. Microbiol. 12:4765–4769.
18. Wong, H. H., and S. Y. Lee. 1998. Poly(3-hydroxybutyrate) production from
whey by high-density cultivation of recombinant Escherichia coli. Appl. Mi-
We thank Mary Berlyn for kindly providing the CGSC strains. crobiol. Biotechnol. 50:30–33.
This work was supported by the Korea-Australia International Co- 19. Yellore, V., and A. Desai. 1998. Production of poly-3-hydroxybutyrate from
operative Project from the Korean Ministry of Science and Technology lactose and whey by Methylobacterium sp. ZP24. Lett. Appl. Microbiol.
(MOST) and by the Brain Korea 21 program. 26:391–394.