acetobutylicum ATCC 824 *

Transcription

1 FEMS Microbiology Letters 108 (1993) Federation of European Microbiological Societies /93/$06.00 Published by Elsevier 319 FEMSLE Determination in Clostridium of plasmid copy number and acetobutylicum ATCC 824 * stability Sang Yup Lee a, Lee D. Mermelstein b and Eleftherios Terry Papoutsakis a a BioProcess Engineering Research Center, Korea Advanced Institute of Science and Technology, Taejon, Korea and b Department of Chemical Engineering, Northwestern University, Evanston, Illinois, USA (Received 16 November 1992; revision received 19 January 1993; accepted 3 February 1993) Abstract: The copy number and stability of several plasmid vectors in Clostridium acetobutylicum ATCC 824 were determined. The protocols were modified from the traditional ones to overcome the problems associated with unusual behavior of C. acetobutylicum cells on solid medium. The plasmid copy numbers of psyl2, pfnk1, pfnk3, and pfnk5 in strain ATCC 824 were 14, 8, 6, and 6, respectively, psyl2 and pfnk1 were segregationally stable, since the fractions of plasmid-carrying cells after 60 generations of growth without antibiotic (erythromycin) were 73% and 77%, respectively. Vector pfnk1 carrying fermentative genes was found to be rather unstable. The observed instability seemed to be due to the complex host-plasmid interactions by amplified expression of enzymes involved in the tightly regulated primary metabolism of C. acetobutylicum. Key words: Clostridium acetobutylicum; Shuttle vector; Plasmid stability; Plasmid copy number Introduction The fermentation by Clostridium acetobutylicum can be used for the production of butanol and acetone as an alternative to petrochemical production. Metabolic engineering of C. acetobutylicum by genetic manipulations has the potential to improve the organism's inherent fer- Correspondence to: Sang Yup Lee, BioProcess Engineering Research Center, Korea Advanced Institute of Science and Technology, Daeduk Science Town, Taejon , Korea. * Part of this work was presented at the conference 'Biochemical Engineering VII' in Santa Barbara, CA, USA, March, mentative characteristics. Once appropriate shuttle vectors have been developed [1,2], it is important to have quantitative information on the plasmid copy number and stability for recombinant fermentation applications. Growth of recombinant cells in the presence of selective pressure may allow the stable maintenance of the plasmid-carrying cell population, but the use of antibiotics is not practical in industrial fermentations. The plasmid copy number is an important factor which can affect plasmid stability and expression of cloned genes. In addition, it may affect the cell metabolism directly when the cloned genes are involved in the primary metabolic pathways. The plasmid copy number and stability have been studied quite extensively for relatively

2 320 well characterized organisms such as Escherichia coli [3-5]. However, there have been no systematic studies reported on the plasmid copy number in C. acetobutylicum. There has been one recent report on the plasmid stability in C. acetobutylicum NI-4081, an autolysin negative mutant strain [6]. The replicon used in their study was Bacillus subtilis plasmid plm13. They suggested that the instability of plm13 derived shuttle vectors in NI-4081 was due to the formation of plasmid multimers. The most commonly used method for determining a plasmid's copy number employs the comparison of the quantity of plasmid DNA to the quantity of host chromosomal DNA. The plasmid copy number is therefore reported as copy number per unit host chromosome. This technique cannot be used in C. acetobutylicum for the following reasons. First, isolation of intact chromosomal DNA is extremely difficult due to the presence of highly active nucleases. Secondly, the genomic size of C. acetobutylicum is still unknown. In this study protocols to determine the plasmid copy number and plasmid stability in C. acetobutylicum were developed and used to determine the copy number and stability of the E. coli-c, acetobutylicum shuttle vector psyl2 [2] and B. subtilis-c, acetobutylicum shuttle vectors [1]. Materials and Methods Bacterial strains and plasmid DNA E. coli and B. subtilis strains used for the transformation and isolation of plasmid DNA were DBll [7] and BR151 [8], respectively. Recombinant C. acetobutylicum ATCC 824 strains harboring psyl2 [2], pfnk1, pfnk3, and pfnk5 [1] were used in this study, psyl2 (8.7 kb) is a shuttle vector between C. acetobutylicum and E. coli. pfnk1 (2.4 kb) is a shuttle vector between C. acetobutylicum and B. subtilis, pfnk3 (4.5 kb) and pfnk5 (4.3 kb) are derivatives of pfnk1, which carry the C. acetobutylicum acetoacetate decarboxylase and,phosphotransbutyrylase genes, respectively. Growth conditions C. acetobutylicum and E. coli strains were routinely grown and maintained as described previously [9]. B. subtilis was grown in Luria-Bertani (LB) medium. For plasmid stability studies, C. acetobutylicum cells harboring plasmids were grown and subcultured in 2 x YTG (per liter: tryptone 16 g, yeast extract 10 g, NaC1 4 g, glucose 5 g). When required, antibiotics were added at the following concentrations: tetracycline, 12 /zg/ml; ampicillin, 50 mg/ml; erythromycin, 35 /zg/ml for E. coli DBll and 40 mg/ml for C. acetobutylicum. DNA manipulation and transformation All DNA manipulations including plasmid DNA isolation and transformation were carried out as previously reported [1,2]. Determination of the number of cells A Coulter counter (Coulter Diagnostics, Hialeah, FL) was used to determine the number of cells. The electrolyte used was Isotone II (per liter: NaCI 7.93 g, EDTA 0.38 g, KCI 0.40 g, NaF 0.30 g, NaH2PO g, Na2HPO g, Coulter Diagnostics). Cells grown to different optical densities were diluted in Isotone II to give approximately 104 cells/ml, mixed well, and a 500 ~1 sample was passed through a 30 ~m aperture orifice tube. The actual number of cells was determined by subtracting the background count of similarly treated growth medium. For each sample, the cell counting was repeated at least three times and the sample standard deviation was calculated. Estimation of plasmid amount Plasmid was prepared from 1.5 ml of culture at several growth stages by the method described earlier [2], either digested or not digested with restriction enzyme, and analyzed by agarose gel electrophoresis. BamHI digested pbr322 and HindlII digested A DNA were used as intensity standards. The gels stained in ethidium bromide solution for 15 and 25 min were photographed on Polaroid 665 P/N film. The negatives were scanned with an LKB Bromma Ultrascan XL laser densitometer (Pharmacia LKB Biotechnol-

3 321 ogy Inc., Piscataway, N J) using the LKB Gelscan XL (v.1.21) computer program. The amount of the plasmid was determined from the ratio of that plasmid to the standard DNA from density tracings. Determination of plasmid stability Cells were grown to an OD600 of 0.3 in 2 x YTG supplemented with erythromycin. The culture was diluted, grown under non-selective conditions to an OD600 of 0.6 to 0.9, and appropriate dilutions were plated on both selective and nonselective plates before transferring into fresh medium. This process was repeated for generations. The plasmid stability was determined by comparing the number of colonies present on antibiotic and non-antibiotic containing plates. Results and Discussion Determination of plasmid copy number As discussed in the Introduction, the traditional methods of determining a plasmid's copy number per unit host chromosome could not be used for C. acetobutylicum. A method that does not require either intact chromosomal DNA isolation or information on the genome size was developed by modifying the methods of Projan et al. [10] and Lewington and Day [11]. First, the correlation between the culture optical density and the absolute number of cells per unit volume was made. Second, plasmid DNA was isolated from a known culture volume and run on an agarose gel along with a pre-determined amount of standard DNA of known size. By comparing the band intensity of plasmid DNA to that of standard DNA, the amount of plasmid DNA was calculated. Since the size of the plasmid was known, the number of plasmid DNA molecules could also be calculated. By dividing the number of plasmid DNA molecules by the corresponding number of cells, the plasmid copy number per cell was obtained. The copy number obtained is a minimum estimate since we assume a 100% yield in the plasmid preparation o Z O ] " a ' l - l - a ' l. i. l, l. l. 0, Time (hr) Fig. 1. Growth of C. acetobutylicum ATCC 824 and number of cells determined by Coulter counting (solid line) and plating (broken line). Error bars represent the sample standard deviation (%). Due to the variation in plating efficiency for cultures at different growth stages (Fig. 1), the true number of cells could not be determined by plating. This unusual behavior of cell growth on solid media is due to cell differentiation [12]. Long et al. [13] also observed a sharp decrease in colony forming units (CFUs) when the culture reached the stationary growth phase. Alternatively, a Coulter counter was used as in Begg and Donachie's study [14]. We did not employ a cell fixing step because no cell breakage occurred for at least 30 min in Isotone II. There was a linear relationship between the number of cells and the OD. The correlating equation was: (No. of ceils) = x 108 x (OD600) [15]. The OD values should be measured with appropriately diluted samples to use the above correlation correctly. This linear interpolated equation is valid with a +7% error for a culture with an OD600 larger than 0.5. When plasmid was linearized by restriction digestion, direct fluorescence intensity compari-

4 322 son could be made with standard DNA. However, when undigested, a correction factor of 1.41 [11] was used to adjust the fluorescence intensity of CCC DNA to that of L DNA. Using this correlation factor, the copy number estimated with undigested DNA was the same as that obtained with linearized DNA. The intensity of OC DNA was not corrected since it binds ethidium bromide to the same extent as the L DNA [11]. The copy number of psyl2 was estimated for cultures grown in 2 x YTG with erythromycin to four different optical densities. A copy number of 12, 14, 15, and 14 was obtained for cultures grown to OD60 o of 0.663, 1.208, 1.332, and 1.398, respectively. The average copy number of psyl2 in C. acetobutylicum ATCC 824 was 14. The copy number of pfnk1 was similarly determined and was found to be ca. 8 in ATCC 824. The difference between the copy numbers of psyl2 (8.7 kb) and pfnk1 (2.4 kb) does not seem to be due to the difference in the plasmid size, but rather due to the different nature of the replicons employed in two vectors as will be mentioned in the next section. The copy number of either pfnk3 or pfnk5 was ca. 6. It is not yet known whether this reduction in the copy number is due either to the presence of the clostridial fermentative genes or the larger sizes of the plasmids. Estimation of plasmid stability Since C. acetobutylicum has been known to exhibit unusual growth characteristics on solid media, we first investigated the suitable growth stage to determine plasmid stability. Most reproducible results were obtained when cells grown to an OD600 of 0.6 to 0.9 were used. Since the replicons of psyl2 and pfnk1 were derived from C. butyricum plasmid pcbu2 [2] and B. subtilis plasmid plm13 [1], respectively, different stability characteristics are expected. pfnk3 and pfnk5 may also show different stability characteristics from pfnk1 due to the presence of clostridial fermentative genes. The segregational stability of psyl2, pfnk1, pfnk3, and pfnk5 is shown in Fig. 2. About 73% of the initial psyl2 containing ATCC 824 transformants was resistant to erythromycin after 60 generations (Fig. 2). psyl2 is therefore considered ::iii o....o 50 "%,i "~, ~o] psvta '~... "~ 20"]-'--OI-- pfnk1 ""-o '~ ]... ~"... pfnk3 10 t ---t3--" pfnk5 i-r' O l,,,,.,,, NO. of generations Fig. 2. Stability of psyl2, pfnk1, pfnk3 and pfnk5 in C. acetobutylicum ATCC 824 following repeated subculturing in non-selective media over about 60 generations. to be segregationally stable in ATCC 824. The structural stability was also studied by isolating plasmid from C. acetobutylicum during the subcultures, re-transforming E. coli, and analyzing the plasmid restriction map and phenotypes. psyl2 isolated from the cells grown for 30 or 60 generations showed the same restriction map and expressed antibiotic markers (both erythromycin and tetracycline resistance) in E. coli. psyl2 is, therefore, structurally stable in ATCC 824, as well. pfnk1 was also found to be segregationally stable since only 23% of cells lost the plasmid after 61 generations (Fig. 2). pfnk1 isolated from ATCC 824 was used to transform B. subtilis. pfnk1 isolated from re-transformed B. subtilis showed the same restriction map as the original plasmid and expressed all the plasmid encoded phenotypes. This suggests that pfnk1 is structurally stable, as well. We also investigated the stability of pfnk3 and pfnk5. They were found to be less stable than pfnk1. This could be simply due to the fact that the cells containing pfnk3 and pfnk5 grew slower than the cells harboring pfnk1, while the latter exibited a similar growth rate to the cells without the plasmid. Vectors carrying fermentative genes were based on plm13, which was shown to be unstable in C. acetobutylicum NI-4081 [6]. They suggested that the instability was due to the formation of plasmid multimers in this strain. They also showed

5 323 that the vectors derived from plm13 were stably maintained in a mutant strain NI-4082, where almost all plasmid was in the monomeric form. This seems to be untrue at least for ATCC 824 since pfnk1, a derivative of pim13, was stably maintained. It should be mentioned that we also observed the multimer formation of pfnk1, pfnk3, and pfnk5 in ATCC 824 [15]. None of the four vectors tested (Fig. 2) appears to follow an exponentially decaying stability curve, which is frequently observed as a result of a takeover of the cell population by faster-growing non-plasmid carrying cells [16]. It is not known at this point why these plasmids do not show this expected behavior, and some in fact appear to reach a plateau in their copy number per cell. In this study we determined the copy number and stability of several plasmids using protocols we developed for C. acetobutylicum. These methods can be used for characterizing recombinant strains of C. acetobutylicum. Acknowledgements This work was carried out at Northwestern University and was supported by Grant No. BCS from the National Science Foundation, USA. References 1 Mermelstein, L.D., Welker, N.E., Bennett, G.N. and Papoutsakis, E.T. (1992) Bio/Technol. 10, Lee, S.Y., Bennett, G.N. and Papoutsakis, E.T. (1992) Biotechnol. Lett. 14, Aiba, S. and Koizumi, J. (1984) Biotechnol. Bioeng. 26, Seo, J.-H. and Bailey, J.E. (1986) Biotechnol. Bioeng. 28, Chew, L.C.K., Tacon, W.C.A. and Cole, J.A. (1988) FEMS Microbiol. Lett. 56, Azeddoug, H., Hubert, J. and Reysset, G. (1992) J. Gen. Microbiol. 138, Macrina, F., Tobian, J.A., Jones, K.R., Evans, R.P. and Clewell, D.B. (1982) Gene 19, Wu, L. and Welker, N.E. (1989) J. Gen. Microbiol. 135, Cary, J.W., Petersen, D.J., Papoutsakis, E.T. and Bennett, G.N. (1988) J. Bacteriol. 170, Projan, S.J., Carleton, S. and Novick, R.P. (1983) Plasmid 9, Lewington, J. and Day, M.J. (1986) Lett. Appl. Microbiol. 3, Jones, D.T. and Woods, D.R. (1986) Microbiol. Rev. 50, Long, S., Jones, D.T. and Woods, D. (1984) Appl. Environ. Microbiol. 20, Begg, K.J. and Donachie, W.D. (1978) J. Bacteriol. 133, Lee, S.Y. (1991) Ph.D. thesis, Dept. of Chem. Eng., Northwestern Univ., Evanston, IL. 16 Imanaka, T., Tsunekawa, H. and Aiba, S. (1980) J. Gen. Microbiol. 118,