- Email: [email protected]
Selection of a suitable strain from recombinant Escherichia coli strains with the same genetic structure expressing periplasmic hGM-CSF
JOURNAL OF BIOSCIENCE AND BIOENGINEERING Vol. 96, No. 6, 578–580. 2003
Selection of a Suitable Strain from Recombinant Escherichia coli Strains with the Same Genetic Structure Expressing Periplasmic hGM-CSF SEYED SAFA-ALI FATEMI,1,2 BAGHER YAKHCHALI,1* JAFAR TOWFIGHI DARIAN,2 SEYED ABBAS SHOJAOSADATI,2 ALI REZA ZOMORODIPOUR,1 ALI ASGHAR KARKHANE,1 AND FERDOWS RASTGAR JAZII1 Industrial Fermentation Group, National Research Center for Genetic Engineering and Biotechnology, P.O. Box 14155-6343, Tehran, Iran1 and Biotechnology Group, Chemical Engineering Department, Faculty of Engineering, Tarbiat Modarres University, P.O. Box 14155-4838, Tehran, Iran2 Received 27 June 2003/Accepted 24 September 2003
The selection of a suitable strain among five recombinant Escherichia coli strains with the same genetic structure that expresses the human granulocyte macrophage colony stimulating factor (hGM-CSF) was carried out based on four criteria: growth rate, expression level, plasmid stability and feasibility for protein extraction. There was no significant difference in growth between the five strains, while a suitable expression level, a high plasmid stability and a good feasibility for protein extraction from periplasmic space were observed for one of the recombinant strains. This strain expressed 27% hGM-CSF relative to total proteins and had 96% plasmid stability after 7-d subcultures on an antibiotic-free LB medium. [Key words: recombinant hGM-CSF, Escherichia coli, plasmid stability, strain selection, extraction feasibility]
plasmic hGM-CSF expression from the bacteriophage T7 promoter have been used for hGM-CSF production (7). The five strains K1, K2, K3, M and P harboring pBZY obtained from the cloning of the hgm-csf gene have been used for hGM-CSF production. Inocula were prepared by growing recombinant E. coli to exponential phase in Luria–Bertani (LB) broth containing 30 mg/ml kanamycin. Test cultures were inoculated with the same amount of the inoculum and grown in 250-ml shake flasks containing 50 ml of LB broth at 30°C and 250 rpm. Lactose (10 g/l) and 1 mM IPTG were used for the induction of the hGM-CSF expression when the optical density of the cultures reached 1.5 and 2.4, respectively. Cell growth was monitored by measurement of OD600. hGM-CSF expression was analyzed by SDS–PAGE and measured with respect to the total proteins of the cell lysate and periplasmic fraction by densitometry. The total protein of the periplasmic fraction was determined by the Bradford method. Plasmid stability was determined by the replica method (8). For determination of plasmid copy number ratio, the same specific quantity of M and P cells was harvested 4 h after induction with IPTG in LB medium. The plasmids of the cells were extracted and analyzed by OD measurement at 280 and 260 nm (9) using a UV-Vis spectrophotometer (DU 530; Beckman, Fullerton, CA, USA). Periplasmic proteins were extracted by applying osmotic shock to evaluate the translocation of hGM-CSF to periplasmic space. The cells induced during overnight E. coli culture were resuspended in 6.7 ml osmotic shock buffer (500 mM sucrose, 0.5 mM EDTA and 1 mM PMSF in 200 mM Tris–HCl, pH 8) per g cell dry weight and incubated for 20 min on ice. Ten milliliters of cold water per g cell dry
Granulocyte macrophage colony stimulating factor (GMCSF) is a glycoprotein that can stimulate the proliferation and differentiation of granulocytes and macrophage progenitor cells (1). The gene encoding hGM-CSF has been cloned and expressed in mammalian (2), bacterial (3) and yeast (4) cells. Recombinant hGM-CSF has considerable therapeutic uses in the treatment of myelodysplasia and AIDS, and improvement of host defense against potential infection following major traumas such as burns (5). The major objective of fermentation in research and industry is to maximize volumetric productivity (g l–1 h–1) using high-cell-density cultivation (6). For high-cell-density cultivation and production of recombinant proteins, a rapidly growing strain that has a stable plasmid and that over expresses an easily extractable protein is required. In this study, we investigate the probable physiological differences between clones obtained from the cloning of the hgm-csf gene in Escherichia coli for the selection of a strain suitable for high-cell-density cultivation and efficient production of hGM-CSF. Such selection is based on maximum growth rate, expression level, plasmid stability and feasibility for protein extraction from periplasmic space. E. coli strain BL21 (DE3) as a host and a chemically inducible expression vector (pBZY) that directs efficient peri* Corresponding author. e-mail: [email protected] phone: +98-21-6419738 fax: +98-21-6419834 The costs of publication of this article were supported in part by Grants-in-Aid for Publication Scientific Research Results from the Ministry of Education, Culture, Sports, Science and Technology, Japan (no. 143050). 578
VOL. 96, 2003
NOTES
579
FIG. 1. SDS–PAGE analysis of hGM-CSF expression level in five recombinant E. coli strains before and after induction. Lanes 1–5, strains K1, K2, K3, M and P, respectively, before induction; lane 6, hGM-CSF standard; lanes 7–11, strains P, M, K3, K2 and K1, respectively, after induction. (A) Cultures were treated with 1 mM IPTG when OD600 reached 2.4 and the samples were prepared 12 h after induction. (B) Cultures were treated with 10 g/l lactose when OD600 reached 1.5 and the samples were prepared 4 h after induction.
weight was added to the suspension. The resulting mixture was further incubated on ice for 30 min, and centrifuged at 20,000 rpm for 10 min at 4°C. The supernatant was collected as the periplasmic fraction. Five strains obtained from the transformation of E. coli BL21(DE3) using a plasmid containing the hgm-csf gene were used for the selection of a strain suitable for production of hGM-CSF on a large-scale. Growth for 16 h did not significantly differ among the five strains. Therefore, it is clear that all strains are the same regarding this criterion. The expression of recombinant hGM-CSF was evaluated by SDS–PAGE of the cell lysate before and after treatment with IPTG (Fig. 1A) or lactose (Fig. 1B). It can be clearly seen that all the strains expressed hGM-CSF after induction with lactose or IPTG. Two distinct protein bands after induction show cytoplasmic (unprocessed) and periplasmic (processed) hGM-CSFs corresponding to the standards used. Western blot analysis using antiserum specific for hGMCSF confirmed hGM-CSF expression (data not shown). The results indicate that lactose could be use as an inducer, which is also suitable for increasing production compared with IPTG because of its cost effectiveness and toxicity effects. The higher hGM-CSF expression level upon induction with IPTG could be due to the longer induction time and inducer property of the IPTG. Under two induction conditions mentioned above, strain M showed only one distinct processed hGM-CSF band. It seems that all of the hGMCSFs expressed by strain M were processed and secreted into periplasmic space. Therefore, it could be a candidate strain suitable for hGM-CSF production. The other strains did not show a very significant difference in expression quality. SDS–PAGE scanning showed densities of 20%, 23%, 21%, 30% and 27% of the processed hGM-CSF bands relative to the total proteins for strains K1, K2, K3, M and P, respectively, when the cultures were treated with IPTG. Figure 2 shows a comparison of plasmid stability among the five strains during 7-d subcultures in an antibiotic-free LB medium. As predicted, plasmid stability decreased with time, but the rate of decrease differed among strains. The highest plasmid stability of 96% was obtained for strain P; strains K1, K2, K3 and M showed 83%, 74%, 75% and 83%
plasmid stabilities, respectively, after 7-d subcultures. According to the above results, strains K2 and K3 are more inferior than the others in respect to expression level and plasmid stability. Thus, strains K1, M, and P were considered for further evaluations. The next criterion studied was feasibility of strains K1, M, and P, for hGM-CSF extraction from periplasmic space which was investigated using osmotic shock. It was found that there is no substantial difference in the electrophoretic pattern of total protein from periplasmic space between strains K1 and P (Fig. 3A, B). Although strains P and K1 had shown an acceptable yield in protein extraction from periplasmic space (about 20 mg hGM-CSF per liter of culture), such a result was not obtained from strain M. A study on the protein extractability of strain M from periplasmic space using osmotic shock showed that hGM-CSF is not released from periplasmic space. Previous studies have shown the formation of periplasmic inclusion bodies or association of recombinant proteins including hGM-CSF (10–12) with cellular membranes upon induction. Thus, different strategies were used to confirm these possibilities. Treatment with 1% (v/v) Triton X-100 plus 1 mM 2-mercaptoethanol as a reducing agent, sonication, or a combination of both Triton X-100 treatment and sonication was carried out. Following either of these treatments, it was found that hGM-CSF can be solubilized and released along with other cellular proteins (Fig. 4). There-
FIG. 2. Comparison of plasmid stability for five recombinant E. coli strains during 7-d subcultures in LB medium.
580
J. BIOSCI. BIOENG.,
FATEMI ET AL.
FIG. 3. SDS–PAGE for evaluation of hGM-CSF extractability from periplasmic space using osmotic shock. (A) Strain P, (B) strain K1. Lane 1, hGM-CSF standard; lane 2, total proteins of cell lysate; lane 3, total soluble proteins of periplasmic space; lane 4, total insoluble proteins after osmotic shock.
FIG. 4. hGM-CSF released from strain M using detergent and sonication, and examined by SDS–PAGE. Lane 1, hGM-CSF standard; lane 2, total protein from cell lysate; lane 3, total insoluble proteins after osmotic shock; lane 4, periplasmic fraction obtained by osmotic shock; lane 5, total soluble proteins after treatment with Triton X-100; lane 6, total insoluble proteins after treatment with Triton X-100; lane 7, total soluble proteins after sonication; lane 8, total insoluble proteins after sonication; lane 9, total soluble proteins after treatment with Triton X-100 and sonication; lane 10, total insoluble proteins after treatment with Triton X-100 and sonication.
fore, hGM-CSF did not precipitate as inclusion bodies and it remained soluble in the inner membrane after processing. A distinct processed protein band after 4-h induction (Fig. 1B) verified the highest expression level and processing rates of hGM-CSF in strain M. A twofold plasmid copy number of strain M relative to that of strain P can result in a high transcription rate for strain M that results in the high hGM-CSF expression level and probably saturates all translocation sites after signal processing and inhibits the hGM-CSF transport. The signal peptide was present because all expressed hGMCSFs were processed (Fig. 1); its sequence has been confirmed previously. Sec mutations cause defects in the transport of many proteins through membranes (13), which can be another reason for trapping hGM-CSF in the membrane of strain M. Because osmotic shock is more feasible than treatment with Triton X-100 and sonication in the largescale production of recombinant proteins and the aim of this research was the selection of a suitable strain with easily extractable products, no more studies on strain M were con-
ducted. Finally, strain P was selected as a strain suitable for hGM-CSF production because of its higher plasmid stability and expression level than strain K1 and better feasibility for partitioning hGM-CSF in periplasmic space than strain M. In this research, different colonies with apparently the same genetic structure obtained from the cloning of hgm-csf in E. coli differed in expression level, plasmid stability and localization of recombinant proteins. It has been shown that the localization of recombinant proteins is affected by the growth conditions for E. coli such as temperature and medium composition (12), signal sequence and Sec mutation (13). The mutation, activation or deactivation of some genes during transformation can be the causative factor for the physiological difference between recombinant strains with the same genetic structure that should be investigated. REFERENCES 1. Seetharam, R. and Sharma, S. K.: Purification and analysis of recombinant proteins. Marcell Dekker, New York (1991). 2. Wong, G. G., Wirek, J. S., Temple, P. A., Wilkens, K. M., Leary, A. C., Luxenberg, D. P., and Shoemarker, S. S.: Human GM-CSF: molecular cloning of the complementary DNA and purification of the natural and recombinant proteins. Science, 28, 810–815 (1985). 3. Burgess, A. W., Begley, C. G., Johnson, G. R., Lopez, A. F., Williamson, D. J., Mermod, J. J., Simpson, R. J., Schmittz, A., and Delemarter, J. F.: Purification and properties of bacterially synthesized human granulocyte-macrophage colony stimulating factor. Blood, 69, 43–51 (1987). 4. Lee, F., Yokota, T., Otsuka, T., Gemmell, L., Larson, N., Luh, J., Arai, K-I., and Rennick, D.: Isolation of cDNA for a human granulocyte-macrophage colony stimulating factor by functional expression in mammalian cells. Proc. Natl. Acad. Sci., 82, 4360–4364 (1985). 5. Weiss, M. and Belohradskey, B. H.: Granulocyte-macrophage colony stimulating factor (GM-CSF). A variety of possible applications in clinical medicine. Infection, 20, S81–S83 (1992). 6. Riesenberg, D. and Guthke, R.: High-cell-density cultivation of microorganisms. Appl. Microbiol. Biotechnol., 51, 422–430 (1999). 7. Borjaliloo, S., Zomorodipour, A., Yakhchali, B., and Shojai, S.: Comparison of T7-and lac-based systems for the periplasmic expression of human granulocyte macrophage colony stimulating factor in Escherichia coli. Iranian J. Biotechnol., 1, 101–108 (2003). 8. Beal, C., Dangio, C., and Corrieu, G.: pH influences growth and plasmid stability of recombinant Lactococcus lactis subsp. Lactis. Biotechnol. Lett., 20, 679–682 (1998). 9. Sambrook, J. and Russell, D. W.: Molecular cloning, a laboratory manual, p. A8–A21. Cold Spring Harbor Press, New York (2001). 10. Libby, R. T., Baredt, G., Kronheim, S. R., March, C. J., Urdal, D. L., Chiaverotti, T. A., Tushinski, R. J., Mochizuki, D. Y., Hopp, T. P., and Cosman, D.: Expression and purification of native hGM-CSF from an E. coli secretion vector. DNA, 6, 221–229 (1987). 11. Greenberg, R., Lundell, D., Alroy, Y., Bonitz, S., Condon, R., Fossetta, J., Frommer, B., Gewain, K., Katz, M., Leibowitz, P. J., Narula, S. K., Kastelein, R., and Kimmenadae, A. V.: Expression of biologically active, mature hGM-CSF with an E. coli secretory expression system. Curr. Microbiol., 17, 321–332 (1988). 12. Jeong, K. J. and Lee, S. Y.: Secretory production of human leptin in E. coli. Biotechnol. Bioeng., 67, 398–407 (2000). 13. Snyder, L. and Champness, W.: Molecular genetics of bacteria. American Society for Microbiology Press, New York (1997).
Selection of a Suitable Strain from Recombinant Escherichia coli Strains with the Same Genetic Structure Expressing Periplasmic hGM-CSF SEYED SAFA-ALI FATEMI,1,2 BAGHER YAKHCHALI,1* JAFAR TOWFIGHI DARIAN,2 SEYED ABBAS SHOJAOSADATI,2 ALI REZA ZOMORODIPOUR,1 ALI ASGHAR KARKHANE,1 AND FERDOWS RASTGAR JAZII1 Industrial Fermentation Group, National Research Center for Genetic Engineering and Biotechnology, P.O. Box 14155-6343, Tehran, Iran1 and Biotechnology Group, Chemical Engineering Department, Faculty of Engineering, Tarbiat Modarres University, P.O. Box 14155-4838, Tehran, Iran2 Received 27 June 2003/Accepted 24 September 2003
The selection of a suitable strain among five recombinant Escherichia coli strains with the same genetic structure that expresses the human granulocyte macrophage colony stimulating factor (hGM-CSF) was carried out based on four criteria: growth rate, expression level, plasmid stability and feasibility for protein extraction. There was no significant difference in growth between the five strains, while a suitable expression level, a high plasmid stability and a good feasibility for protein extraction from periplasmic space were observed for one of the recombinant strains. This strain expressed 27% hGM-CSF relative to total proteins and had 96% plasmid stability after 7-d subcultures on an antibiotic-free LB medium. [Key words: recombinant hGM-CSF, Escherichia coli, plasmid stability, strain selection, extraction feasibility]
plasmic hGM-CSF expression from the bacteriophage T7 promoter have been used for hGM-CSF production (7). The five strains K1, K2, K3, M and P harboring pBZY obtained from the cloning of the hgm-csf gene have been used for hGM-CSF production. Inocula were prepared by growing recombinant E. coli to exponential phase in Luria–Bertani (LB) broth containing 30 mg/ml kanamycin. Test cultures were inoculated with the same amount of the inoculum and grown in 250-ml shake flasks containing 50 ml of LB broth at 30°C and 250 rpm. Lactose (10 g/l) and 1 mM IPTG were used for the induction of the hGM-CSF expression when the optical density of the cultures reached 1.5 and 2.4, respectively. Cell growth was monitored by measurement of OD600. hGM-CSF expression was analyzed by SDS–PAGE and measured with respect to the total proteins of the cell lysate and periplasmic fraction by densitometry. The total protein of the periplasmic fraction was determined by the Bradford method. Plasmid stability was determined by the replica method (8). For determination of plasmid copy number ratio, the same specific quantity of M and P cells was harvested 4 h after induction with IPTG in LB medium. The plasmids of the cells were extracted and analyzed by OD measurement at 280 and 260 nm (9) using a UV-Vis spectrophotometer (DU 530; Beckman, Fullerton, CA, USA). Periplasmic proteins were extracted by applying osmotic shock to evaluate the translocation of hGM-CSF to periplasmic space. The cells induced during overnight E. coli culture were resuspended in 6.7 ml osmotic shock buffer (500 mM sucrose, 0.5 mM EDTA and 1 mM PMSF in 200 mM Tris–HCl, pH 8) per g cell dry weight and incubated for 20 min on ice. Ten milliliters of cold water per g cell dry
Granulocyte macrophage colony stimulating factor (GMCSF) is a glycoprotein that can stimulate the proliferation and differentiation of granulocytes and macrophage progenitor cells (1). The gene encoding hGM-CSF has been cloned and expressed in mammalian (2), bacterial (3) and yeast (4) cells. Recombinant hGM-CSF has considerable therapeutic uses in the treatment of myelodysplasia and AIDS, and improvement of host defense against potential infection following major traumas such as burns (5). The major objective of fermentation in research and industry is to maximize volumetric productivity (g l–1 h–1) using high-cell-density cultivation (6). For high-cell-density cultivation and production of recombinant proteins, a rapidly growing strain that has a stable plasmid and that over expresses an easily extractable protein is required. In this study, we investigate the probable physiological differences between clones obtained from the cloning of the hgm-csf gene in Escherichia coli for the selection of a strain suitable for high-cell-density cultivation and efficient production of hGM-CSF. Such selection is based on maximum growth rate, expression level, plasmid stability and feasibility for protein extraction from periplasmic space. E. coli strain BL21 (DE3) as a host and a chemically inducible expression vector (pBZY) that directs efficient peri* Corresponding author. e-mail: [email protected] phone: +98-21-6419738 fax: +98-21-6419834 The costs of publication of this article were supported in part by Grants-in-Aid for Publication Scientific Research Results from the Ministry of Education, Culture, Sports, Science and Technology, Japan (no. 143050). 578
VOL. 96, 2003
NOTES
579
FIG. 1. SDS–PAGE analysis of hGM-CSF expression level in five recombinant E. coli strains before and after induction. Lanes 1–5, strains K1, K2, K3, M and P, respectively, before induction; lane 6, hGM-CSF standard; lanes 7–11, strains P, M, K3, K2 and K1, respectively, after induction. (A) Cultures were treated with 1 mM IPTG when OD600 reached 2.4 and the samples were prepared 12 h after induction. (B) Cultures were treated with 10 g/l lactose when OD600 reached 1.5 and the samples were prepared 4 h after induction.
weight was added to the suspension. The resulting mixture was further incubated on ice for 30 min, and centrifuged at 20,000 rpm for 10 min at 4°C. The supernatant was collected as the periplasmic fraction. Five strains obtained from the transformation of E. coli BL21(DE3) using a plasmid containing the hgm-csf gene were used for the selection of a strain suitable for production of hGM-CSF on a large-scale. Growth for 16 h did not significantly differ among the five strains. Therefore, it is clear that all strains are the same regarding this criterion. The expression of recombinant hGM-CSF was evaluated by SDS–PAGE of the cell lysate before and after treatment with IPTG (Fig. 1A) or lactose (Fig. 1B). It can be clearly seen that all the strains expressed hGM-CSF after induction with lactose or IPTG. Two distinct protein bands after induction show cytoplasmic (unprocessed) and periplasmic (processed) hGM-CSFs corresponding to the standards used. Western blot analysis using antiserum specific for hGMCSF confirmed hGM-CSF expression (data not shown). The results indicate that lactose could be use as an inducer, which is also suitable for increasing production compared with IPTG because of its cost effectiveness and toxicity effects. The higher hGM-CSF expression level upon induction with IPTG could be due to the longer induction time and inducer property of the IPTG. Under two induction conditions mentioned above, strain M showed only one distinct processed hGM-CSF band. It seems that all of the hGMCSFs expressed by strain M were processed and secreted into periplasmic space. Therefore, it could be a candidate strain suitable for hGM-CSF production. The other strains did not show a very significant difference in expression quality. SDS–PAGE scanning showed densities of 20%, 23%, 21%, 30% and 27% of the processed hGM-CSF bands relative to the total proteins for strains K1, K2, K3, M and P, respectively, when the cultures were treated with IPTG. Figure 2 shows a comparison of plasmid stability among the five strains during 7-d subcultures in an antibiotic-free LB medium. As predicted, plasmid stability decreased with time, but the rate of decrease differed among strains. The highest plasmid stability of 96% was obtained for strain P; strains K1, K2, K3 and M showed 83%, 74%, 75% and 83%
plasmid stabilities, respectively, after 7-d subcultures. According to the above results, strains K2 and K3 are more inferior than the others in respect to expression level and plasmid stability. Thus, strains K1, M, and P were considered for further evaluations. The next criterion studied was feasibility of strains K1, M, and P, for hGM-CSF extraction from periplasmic space which was investigated using osmotic shock. It was found that there is no substantial difference in the electrophoretic pattern of total protein from periplasmic space between strains K1 and P (Fig. 3A, B). Although strains P and K1 had shown an acceptable yield in protein extraction from periplasmic space (about 20 mg hGM-CSF per liter of culture), such a result was not obtained from strain M. A study on the protein extractability of strain M from periplasmic space using osmotic shock showed that hGM-CSF is not released from periplasmic space. Previous studies have shown the formation of periplasmic inclusion bodies or association of recombinant proteins including hGM-CSF (10–12) with cellular membranes upon induction. Thus, different strategies were used to confirm these possibilities. Treatment with 1% (v/v) Triton X-100 plus 1 mM 2-mercaptoethanol as a reducing agent, sonication, or a combination of both Triton X-100 treatment and sonication was carried out. Following either of these treatments, it was found that hGM-CSF can be solubilized and released along with other cellular proteins (Fig. 4). There-
FIG. 2. Comparison of plasmid stability for five recombinant E. coli strains during 7-d subcultures in LB medium.
580
J. BIOSCI. BIOENG.,
FATEMI ET AL.
FIG. 3. SDS–PAGE for evaluation of hGM-CSF extractability from periplasmic space using osmotic shock. (A) Strain P, (B) strain K1. Lane 1, hGM-CSF standard; lane 2, total proteins of cell lysate; lane 3, total soluble proteins of periplasmic space; lane 4, total insoluble proteins after osmotic shock.
FIG. 4. hGM-CSF released from strain M using detergent and sonication, and examined by SDS–PAGE. Lane 1, hGM-CSF standard; lane 2, total protein from cell lysate; lane 3, total insoluble proteins after osmotic shock; lane 4, periplasmic fraction obtained by osmotic shock; lane 5, total soluble proteins after treatment with Triton X-100; lane 6, total insoluble proteins after treatment with Triton X-100; lane 7, total soluble proteins after sonication; lane 8, total insoluble proteins after sonication; lane 9, total soluble proteins after treatment with Triton X-100 and sonication; lane 10, total insoluble proteins after treatment with Triton X-100 and sonication.
fore, hGM-CSF did not precipitate as inclusion bodies and it remained soluble in the inner membrane after processing. A distinct processed protein band after 4-h induction (Fig. 1B) verified the highest expression level and processing rates of hGM-CSF in strain M. A twofold plasmid copy number of strain M relative to that of strain P can result in a high transcription rate for strain M that results in the high hGM-CSF expression level and probably saturates all translocation sites after signal processing and inhibits the hGM-CSF transport. The signal peptide was present because all expressed hGMCSFs were processed (Fig. 1); its sequence has been confirmed previously. Sec mutations cause defects in the transport of many proteins through membranes (13), which can be another reason for trapping hGM-CSF in the membrane of strain M. Because osmotic shock is more feasible than treatment with Triton X-100 and sonication in the largescale production of recombinant proteins and the aim of this research was the selection of a suitable strain with easily extractable products, no more studies on strain M were con-
ducted. Finally, strain P was selected as a strain suitable for hGM-CSF production because of its higher plasmid stability and expression level than strain K1 and better feasibility for partitioning hGM-CSF in periplasmic space than strain M. In this research, different colonies with apparently the same genetic structure obtained from the cloning of hgm-csf in E. coli differed in expression level, plasmid stability and localization of recombinant proteins. It has been shown that the localization of recombinant proteins is affected by the growth conditions for E. coli such as temperature and medium composition (12), signal sequence and Sec mutation (13). The mutation, activation or deactivation of some genes during transformation can be the causative factor for the physiological difference between recombinant strains with the same genetic structure that should be investigated. REFERENCES 1. Seetharam, R. and Sharma, S. K.: Purification and analysis of recombinant proteins. Marcell Dekker, New York (1991). 2. Wong, G. G., Wirek, J. S., Temple, P. A., Wilkens, K. M., Leary, A. C., Luxenberg, D. P., and Shoemarker, S. S.: Human GM-CSF: molecular cloning of the complementary DNA and purification of the natural and recombinant proteins. Science, 28, 810–815 (1985). 3. Burgess, A. W., Begley, C. G., Johnson, G. R., Lopez, A. F., Williamson, D. J., Mermod, J. J., Simpson, R. J., Schmittz, A., and Delemarter, J. F.: Purification and properties of bacterially synthesized human granulocyte-macrophage colony stimulating factor. Blood, 69, 43–51 (1987). 4. Lee, F., Yokota, T., Otsuka, T., Gemmell, L., Larson, N., Luh, J., Arai, K-I., and Rennick, D.: Isolation of cDNA for a human granulocyte-macrophage colony stimulating factor by functional expression in mammalian cells. Proc. Natl. Acad. Sci., 82, 4360–4364 (1985). 5. Weiss, M. and Belohradskey, B. H.: Granulocyte-macrophage colony stimulating factor (GM-CSF). A variety of possible applications in clinical medicine. Infection, 20, S81–S83 (1992). 6. Riesenberg, D. and Guthke, R.: High-cell-density cultivation of microorganisms. Appl. Microbiol. Biotechnol., 51, 422–430 (1999). 7. Borjaliloo, S., Zomorodipour, A., Yakhchali, B., and Shojai, S.: Comparison of T7-and lac-based systems for the periplasmic expression of human granulocyte macrophage colony stimulating factor in Escherichia coli. Iranian J. Biotechnol., 1, 101–108 (2003). 8. Beal, C., Dangio, C., and Corrieu, G.: pH influences growth and plasmid stability of recombinant Lactococcus lactis subsp. Lactis. Biotechnol. Lett., 20, 679–682 (1998). 9. Sambrook, J. and Russell, D. W.: Molecular cloning, a laboratory manual, p. A8–A21. Cold Spring Harbor Press, New York (2001). 10. Libby, R. T., Baredt, G., Kronheim, S. R., March, C. J., Urdal, D. L., Chiaverotti, T. A., Tushinski, R. J., Mochizuki, D. Y., Hopp, T. P., and Cosman, D.: Expression and purification of native hGM-CSF from an E. coli secretion vector. DNA, 6, 221–229 (1987). 11. Greenberg, R., Lundell, D., Alroy, Y., Bonitz, S., Condon, R., Fossetta, J., Frommer, B., Gewain, K., Katz, M., Leibowitz, P. J., Narula, S. K., Kastelein, R., and Kimmenadae, A. V.: Expression of biologically active, mature hGM-CSF with an E. coli secretory expression system. Curr. Microbiol., 17, 321–332 (1988). 12. Jeong, K. J. and Lee, S. Y.: Secretory production of human leptin in E. coli. Biotechnol. Bioeng., 67, 398–407 (2000). 13. Snyder, L. and Champness, W.: Molecular genetics of bacteria. American Society for Microbiology Press, New York (1997).