- Email: [email protected]
Impact of plasmid presence and induction on cellular responses in fed batch cultures of Escherichia coli
Biotechno ELSEVIER
Journal of Biotechnology
46 (1996) 255-263
Impact of plasmid presence and induction on cellular responses in fed batch cultures of Escherichia coli Lena Andersson
*, Shaojun
Yang, Peter Neubauer ‘, Sven-Olof Enfors
Department of Biochemistry and Biotechnology, Royal Institute of Technology IKTHJ, S-100 44 Stockholm, Sweden Received 21 October
1995; revised 27 December
1995; accepted 28 December
1995
Abstract Fed batch cultivations of plasmid-free and recombinant Escherichia coli were employed in order to determine cellular responses and effects of plasmid presence and induction on the host cell physiology. While plasmid presence was shown to have minor influence on overall biomass yield, induction with 0.1 mM IFTG led to a marked reduction. The number of dividing cells, measured as colony forming ability, was influenced by plasmid presence and to a larger extent by induction. The latter caused a decline in the number of dividing cells to less than 10% of the population within 10 h. However, this cell segregation did not affect the specific rate of product formation, which was approximately constant throughout the cultivations. Analysis of the in vivo degradation rate of the product indicated that it was proteolytically stable. The cellular content of the stringent response signal substance, ppGpp, peaked immediately after transition from batch to fed batch mode to stabilise at a higher value than in the batch phase. When the specific growth rate declined below 0.06 h- ’ an additional rise in ppGpp concentration was observed. Keywords: Recombinant
protein production;
Cellular responses;
Stringent response; Proteolysis;
1. Introduction Efforts in finding strategies to maximise the productivity of recombinant protein production in E. coli are well documented in literature. Numerous studies have focused on gaining more optimal solutions by molecular biological methods and vector constructions (Imanaka, 1986; Das, 19901. Others have been concerned with enhancing the productivity
* Corresponding author. Tel. (+46-g) 7907503, fax. (+46-g) 7231890. ’ Present address: Institut fur Angewandte Biochemie, Fachbereich Biochemie/Biotechnologie, Martin-Luther-Universitiit Halle-Wittenberg, D-06099 Halle, Germany. Elsevier Science B.V. PII SOl68-1656(96)00004-l
Fed batch; E. coli
by engineering means by employing high cell density cultures and control strategies to avoid inhibiting mechanisms like by-product formation and oxygen limitation (Gleiser and Bauer, 1981; O’Connor et al., 1992). However, growing cells to high densities aerobically includes the use of severely energy limited processes, of which the physiological consequences are not so well described. Recombinant protein production is an interdisciplinary art: both genetic and process engineering are essential parts in the development of a bioprocess. The link between the two is the host cell and its metabolic and regulatory characteristics. The amount of protein produced from a recombinant strain is strongly determined by factors such as host cell-vec-
256
L. Andersson
et cd. /Journal
tor interactions, plasmid segregational instability and the extent of cellular stress responses (Zabriskie and Arcuri, 1986; Kim et al., 1992; Bailey, 1993). A commonly reported host cell-vector interaction is the negative effect of plasmid presence on growth rate and overall growth yield (Sea and Bailey, 1985; Reinikainen and Virkajlrvi. 1989). Possible causes for increased maintenance metabolism are redirection of cellular resources to support plasmid-related activities and inhibitory mechanisms on host cell metabolism. Alterations in levels of E. coli cell proteins and ribosomal components have been shown to occur after introduction of multicopy plasmids (Bimbaum and Bailey, 1991). The levels of stress proteins were higher for recombinant strains, while metabolic enzymes showed lower values. Many studies have reported growth rate related segregational plasmid instability (Ryan and Parulekar, 1991; Flickinger and Rouse, 1993). Another segregation phenomenon observed at slow growth of plasmid-free E. coli, is the occurrence of an increasing fraction of cells that have lost its colony forming ability but still exhibits metabolic activity (Andersson et al., 1995). The physiological status of this type of subpopulation is still not known. A possible loss of biosynthetic capacity may however have detrimental influence on protein production. Cellular stress responses like heat shock and stringent response may also have profound effects on the final amount and quality of product. Besides affecting many other processes, these global regulation systems are known to induce or augment protease activity (Voellmy and Goldberg, 1980; Goff et al., 1984). There is strong evidence that elevated levels of the nucleotide guanosine S-diphosphate 3’-diphosphate (ppGpp), evoked by depletion of amino acids or by severe energy/carbon limitation, trigger the stringent response (Gallant, 1979; Cashel and Rudd, 1987; Stouthamer and Bulthuis, 1990). Intracellular concentrations of ppGpp have also been shown to be inversely correlated to specific growth rate and the amount of stable RNA transcripts (Arbige and Chesbro, 1982; Bremer and Dennis, 1987). In this paper, we have focused on the behaviour of plasmid-free and recombinant E. coli during growth in energy limited cultures. Biomass yield, segregation to non-dividing cells and levels of ppGpp
of Biotechnology 46
(19%)
255-263
were compared. The rate of product formation and proteolysis in vivo were determined during production of the recombinant protein, ZZ.
2. Material and methods 2. I. Strain and plasmid The bacterial strain used, RB791 (lacIqL,), was kindly provided from the E. coli Genetic Stock Centre. The plasmid pRIT28NZZT is a derivate of pRIT28 (Hultman et al., 19881, into which a gene encoding protein ZZ has been inserted. This protein is composed of two IgG-binding staphylococcal protein A domain analogues (Zl (Nilsson et al., 1985). The production of the 15 kDa protein is under control of the lacUV5 promoter. The expression was induced by addition of 0.1 mM isopropyl P-Dthiogalactopyranoside, IPTG, to the cultures (and feed solutions). 2.2. Cultiuation medium The culture medium (main and precultures) had the following composition (g 1-l 1: Na,SO, 2.0; NH,Cl 0.5; Na,HPO,. 2H,O 3.0; KH,PO, 7.0; (NH,)2S0, 1.0; thiamine 0.1; ampicillin 0.07. The medium also contained trace component solution 2 ml 1-l (Holme et al., 1970) and 1 M MgSO, 2 ml l- ’ . The initial glucose concentration was 10 g 1- ’ . The feed solution contained essentially the same components as the initial medium with the exceptions that no MgSO, was added, the trace component solution and glucose concentrations were 8 ml I-’ and 150 g l-l, respectively. Additional 8 ml MgSO, (1 Ml was added intermittently to the cultivations to avoid precipitation. The pH was controlled at 7 with 4 M H,SO, and 25% NH,OH. Adecanol LG- 109 (antifoam) was added when necessary. 2.3. Cultiuation conditions Inoculum was prepared from shake flask cultures grown at 37°C. The experiments were carried out in a 7 1 bioreactor with an initial culture volume of 3.5 1. Temperature was kept at 37°C. Air flow rate was 2.0 1 min-‘. The stirrer speed was kept at 800 RPM
L.. Andersson et al. / Journal of Biotechnology 46 (1996) 255-263
and the dissolved oxygen tension did not fall below 20% in any cultivation. The cultivations were started as batch cultures and the feed, at a rate of 40 ml h- ‘, was started when the glucose concentration was between 1 and 2 g 1-l. 2.4. Analytical
methods
Optical density was measured at 580 nm. Dry weight was determined by centrifuging 3 X 5 ml of cell suspension in preweighed tubes, washing the samples with distilled water and drying overnight at 105°C before weighing. Analyses of number of dividing cells (colony forming units, cfu) were performed on nutrient agar plates, with and without ampicillin (0.07 g l- ’ >, incubated at 37°C for 24 h. Dissolved oxygen tension was continuously measured with a polarographic oxygen electrode. The oxygen and carbon dioxide content in the effluent gas were analysed with a paramagnetic analyser (Servomex Oxygen Analyser 540A, Sybron/Taylor, UK) and an IR analyser (Binos, Leybold-Hereaus, Germany), respectively. Plasmid content in the cells was detected from 0.25 to 1 ml process samples by a mini-preparation method based on the alkaline sodium dodecyl method of Bimboim and Doly (Bimboim and Doly, 1979). Plasmid purity was controlled by agarose gel electrophoresis in 0.8% agarose gels and plasmid concentration was measured with a GeneQuant RNA/DNA Calculator (Pharmacia Biotech, Uppsala, Sweden). Total protein concentration in the medium was measured by the Bradford method (Bradford, 1976) with BSA as standard. Samples were taken from the bioreactor, centrifuged and filtered before freezing at - 20°C. Samples for analysis of the protein ZZ concentration were subjected to high pressure homogenisation prior to protein purification using IgG-affinity chromatography (Yang et al., 1994). The elution of ZZ was monitored by following the absorbance at 280 nm. The extinction coefficient used was 0.152 emg ml-’ cm-’ according to the amino acid composition (Gill and von Hippel, 1989). SDS-PAGE analysis of the IgG-binding fraction was performed with 3.5% stacking gel and 16% separation gel (Laemmli, 1970).
2.51
Proteins were stained with Coomassie Brilliant Blue (CBB) R-250. 1 pg of purified protein was loaded to each electrophoresis lane. To analyse effects of proteolysis on the product, in vivo proteolysis of ZZ protein at different culture times after induction was measured by using chloramphenicol to stop de novo synthesis of proteins (Yang and Enfors, 1995). Samples, withdrawn from cultivations at different culture times, were diluted with minimal salt medium to about OD = 4 and incubated at 37°C. The mineral salt medium consisted of (g 1-l): (NI-I,),SO, 7; KI-I,pO, 1.6; Na,HPO, .2H,O 6.6 and 100 pg ml-’ chloramphenicol with an initial pH of 7.0. Samples were removed after 0, 30 and 60 min of incubation, centrifuged and stored at -20°C. The composition of the IgG-binding protein fraction was determined by Western blotting analysis as described elsewhere (Yang and Enfors, 1995). The same amount of cells was loaded to each electrophoresis lane. Cellular ppGpp concentration was analysed by an HPLC method on a Supercosil LC-18T column as described elsewhere (Neubauer et al., 1995).
3. Results 3.1. Efsects of plasmid growth kinetics
presence
and induction
on
Effects of plasmid presence and chemically induced gene expression (0.1 mM IPTG) were studied in glucose limited fed batch cultivations of E. coli. Fig. 1A shows growth curves from three types of cultures: host strain not bearing plasmid, host strain transformed with plasmid pRIT28NZZT encoding the intracellular protein ZZ, but without induction and finally with induction after 3 h of glucose limited growth. Initially, the accumulated biomass (g) increased almost linearly with the same rate for the different cultivations but at about 10 h, when /.Lwas about 0.05 h- ‘, the rates started to decline gradually (Fig. lA, C). The induced culture then grew at lower rate and the final amount of biomass achieved was considerably lower (23%) compared to the other cultures. The growth rate of the plasmid containing but non-induced culture was intermediate. The results indicated that induction with 0.1 mM IPTG led
258
et al./Joumul of Biotechnology 46 (1996) 255-263
L. Anderssan
to an increased maintenance requirement, as a larger fraction of the substrate was used for non-growth purposes, whereas the contribution by plasmid presence alone was less drastic. These findings were further reflected in data on specific glucose consumption rate, qs (Fig. 1B). The values of qs for all cultures rapidly declined during the first 10 h to approach a lower limit where most of the sugar feed was consumed for maintenance purpose. The qsvalue for the induced culture was about 20% higher than for the other two cultures at 30 h. The number of dividing ceils, measured as colony
\
d
\ 0. ‘CL-_:=*;-+ I
0
10
la I
,-
20
I
30
Time (h) Fig. 2. Comparison of colony forming ability (cfu ml-’ ) of plasmid-free cells (without plasmid), non-induced, plasmid-bearing cells (without induction) and induced, plasmid-bearing cells (induction). Induction was performed at 3 h (indicated by arrow) with 0.1 mM IPTG.
Ol
0.2
/
1
1
j
I
I c
-
0.15-
--CL
0.1 -:%kE
L
:
A .
0.05 -
‘“L-, b
$-.-
81
0
1 0
I IO
,
Time
$=-kP I 20
I
I 30
(h)
Fig. I. Comparison of cell growth during glucose limited fed batch cultivations of three types of cells; plasmid-free cells (without plasmid), non-induced, plasmid-bearing cells (without induction) and induced, piasmid-bearing cells (induction). (A) Cell mass (g); (B) specific substrate uptake rate, qs; (hh ’ ) (C) specific growth rate, k (hh’ ).
forming ability (cfu ml- ’ >, differed considerably among the three cultures, as can be seen in Fig. 2. The cfu-values for the induced culture started to drop after induction (3 h) and after 15 h the colony forming ability was only 10% of the maximum value. In the cultivation with plasmid-bearing cells, without induction, the cfu-values exhibited a similar pattern with the difference that the decline appeared later, at about 10 h. The difference in colony forming ability on non-selective media and selective media (with ampicillin) was insignificant (not shown), indicating that the observed response was not caused by formation of plasmid-free segregants. Colony forming ability in the cultivation with plasmid-free cells remained high until 25 h, when it started to decline. To investigate if the loss of colony forming ability was a consequence of cell lysis the amount of extracellular proteins in the culture media was analysed (Fig. 3). During the first 15 h no substantial difference in accumulation of extracellular proteins could be seen between the induced culture and the plasmid-free culture. Thereafter, the extracellular protein content in the induced culture increased drastically to become 130% higher (9.3 g) than in the plasmid-free culture (4.0 g> at the end of the experiment. These results suggest that cell lysis occurred in later stages
L Andersson et al./Joumal
of Biotechnology 46 (1996) 255-263
259
of all cultures, but that it was more pronounced in the induced culture. The onset of increasing accumulation rate of extracellular proteins correlates with the time when p approached 0.03 hh’ in all cultures. Assuming that the cellular protein content was 50%, cell lysis would account for about 65% of the difference in finally achieved biomass between the cultures. However, this amount of extracellular protein corresponds only to 26% of the final sum of cell- and extracellular proteins, which rules out cell lysis as an explanation for the more than 90% loss of colony forming ability. 1
I
r
,
20
3.2. Cellular responses and product fonnation
I
I
3
30
Time (h)
Data on intracellular product formation showed that a low amount of protein ZZ was present before induction (Fig. 4). Following induction, the specific protein content (mg ZZ per g cells) increased roughly linearly throughout the cultivation. The final specific amount of ZZ corresponded to approximately 5% of the total cell mass. To measure the rate of proteolysis in vivo, samples were taken from the bioreactor intermittently, and incubated O-60 min with 100 pg ml-’ chloramphenicol prior to analysis of the composition of the IgG-binding protein fraction. Western blot analyses of samples taken at different times during the process showed that the amount of the full
lrlf
’
’
I
10
’
’
20
’
’
’
I
30
Time (h) Fig. 3. Extracellular protein content (g) of plasmid-free cells (without plasmid), non-induced, plasmid-bearing cells (without induction) and induced, plasmid-bearing cells (induction).
Fig. 4. Product formation of protein ZZ (mg ZZ per g cells) in a glucose limited fed batch cultivation. The production system was induced with 0.1 mM IPTG after 3 h of limited growth (as indicated).
length product was constant during the 60 min of incubation, indicating that no degradation took place (results not shown). However, several IgG-binding bands with lower molecular mass than the main band were also visible. These bands remained essentially constant for all samples during the incubation with chloramphenicol. In contrast, SDS-PAGE analysis of the IgG-binding protein fraction recovered by affinity chromatography did not reveal any other IgG-binding bands than tbe product band (results not shown). The concentration of the main effector of stringent response, ppGpp, was determined in fed batch cultivations for production of protein ZZ. The experiments were started as batch cultures and a constant feed of glucose was supplied to the bioreactor before the initially added glucose was exhausted (0 h) to exclude any period of glucose starvation. Recombinant protein production was induced after 3 h of limited growth. Fig. 5A shows the cellular ppGpp content during the different phases of an experiment with induced cells. The first measurement of ppGpp (60 nmol per g cells), at -0.5 h, was taken during growth in batch phase. This concentration was lower than the succeeding values from the fed batch phase. Other cultures exhibited a similar pattern of cellular ppGpp content (data not shown). The downshift from
L. Andersson
260
(4 3 = s m
600
of Biotechnology
46 (19961255-263
t
i
B 2 c
+
et al. /Journal
400
; E g 200 c3 E? $ 3 J
0
I
2,s
5
f
0.1
0.05 Specific
35
growth
rate (h’)
Fig. 5. (A) cellular ppGpp concentration (nmol per g cells) during different phases in the cultivation for production of recombinant protein ZZ. The culture was started as a batch culture and a constant feed of glucose was supplied into the bioreactor before the initial added glucose was exhausted (0). The last measurement was done after that the substrate feed had been shut off for 30 min. (B) Cellular ppGpp concentration (nmol per g cells) as function of specific growth rate
batch to fed batch mode (glucose limitation) resulted in a sharp peak in ppGpp content (500 nmol per g cells) and the concentration remained elevated at approximately 100 nmol per g cells during the subsequent 5 h. A second, significant rise was observed in ppGpp concentration at 10 h before it stabilised at a new, higher level at about 175 nmol per g cells, where it remained throughout the cultivation. The last measurement, at 35.5 h, was done after that the substrate feed had been shut off for 30 min, reflecting the response when the cells were exposed to complete energy/carbon starvation. Cellular ppGpp content as a function of specific growth rate during the fed batch phase (3-35 h), is presented in Fig. 5B.
The ppGpp content increased at a time when the specific growth rate had declined to a low value around 0.06 h- ’ . Plasmid-free and non-induced plasmid-bearing cultures also showed a high ppGpp peak in the transition from batch to fed batch and an increased ppGpp level during the late stages of the fed batch phase (data not shown).
4. Discussion The fed batch technique is widely employed in industrial processes. Often, at least during part of the process, a constant feed rate is applied. The charac-
L. Andersson et al./Joumal
of Biotechnology 46 (1996) 255-263
teristic features of such cultures arc progressively increasing substrate limitation and declining specific growth rate. The present investigation revealed a number of physiological responses during such conditions. Furthermore, differences in behaviour between plasmid-free, non-induced plasmid-bearing and induced cultures were found. Induction of plasmid encoded protein ZZ with 0.1 mM IPTG was shown to have strong influence on biomass formation due to increased maintenance metabolism and cell lysis. On the other hand, a high content of plasmids, above 200 ng plasmid per g cells throughout the experiments (data not shown), was shown to have a somewhat smaller effect on the biomass yield. A substantial fraction of the loss in biomass could be explained by analysis of extracellular protein content. Consequently, the observed increase in specific substrate consumption rate could only to a lesser extent be comprised by an additional energy demand. Cell lysis may still, from a macroscopic point of view, be regarded as part of the maintenance demand. The expression system used resulted in relatively low amounts of recombinant protein, with a final value at about 5% of the cell mass. This implies that production only exerts a modest demand on metabolic activity of the host cell and that the lethal effect on cell growth should be attributed to IPTG itself. Previous studies have shown that addition of IPTG to wild-type E. coli leads to elevated expression of several stress response proteins (Kosinski et al., 1992). The observed loss of biomass in the present study corresponded to a 23% decrease in final, conceivable product concentration, which makes utilisation of IPTG as inducer quite questionable, even at as low dosage as 0.1 n&I. Instead, IPTG might be replaced by lactose or glycerol-P-galactoside (Hengge and Boos, 1983; Neubauer et al., 1992). The phenomenon of cells that respirate but have lost their ability to form colonies on nutrient agar plates, viable but non-culturable cells, seems to be associated with stress of nutrient limitation and is well recognised in studies of bacterial starvation (Mason et al., 1986a; Oliver et al., 1991; Kaprelyants et al., 1993). Moreover, it has been shown that it is possible to resuscitate some of these bacterial forms in liquid medium, thus the state appears to be reversible (Nilsson et al., 1991). In the present work, a
261
segregation of the populations where an increasing fraction of the cells had lost their ability to form colonies occurred in all cultures, but at distinctly different times. These results suggest that the cultures were exposed to different magnitudes of stress and that induction with IPTG appears to have triggered cell segregation. The segregation that took place during the first 10 h after induction did not result in reduced specific productivity (on dry weight basis) in spite of the reduction of the cfu-values by more than 90%. Thus, the non-dividing population contributes to the protein production. The Western blotting analyses of cells incubated with chloramphenicol to inhibit protein synthesis indicated that the product was stable with respect to proteolysis during all phases of the process. The IgG-binding bands with lower molecular mass than the main band revealed by this analysis were not found when the IgG-binding protein fraction recovered by affinity chromatography was analysed by SDS-PAGE and CBB staining. This divergence can be explained from the fact that the applied analysis methods differ in their accuracy with regard to qualification and quantification of proteins. Thus, the Western blotting analysis detects extremely low concentrations of proteins whereas SDS-PAGE with CBB staining gives a more reliable estimation of the amount of protein. The cellular ppGpp content increased transiently after the first carbon/energy down-shift from batch to fed batch mode and stabilised at a higher level than in the batch phase. It is interesting to note that even if the culture was not exposed to carbon starvation, but limitation, the measured peak of ppGpp, 500 nmol per g cells is similar to those that have been observed during complete glucose starvation (Chaloner-Larsson and Yamazaki, 1978). Later in the cultivations at a low specific growth rate an additional, significant rise in ppGpp content was observed, indicating conditions of severe energy/carbon limitation. The last ppGpp measurement, 275 nmol per g cells, was taken after the culture had been exposed to complete carbon starvation for 30 min. This low value, when compared to the first peak, might be explained by a limitation in the amount of cellular ATP available for synthesis of ppGpp at that late stage of the cultivation.
262
L. Andersson et al./ Journal of Biotechnology 46 (1996) 255-263
Acknowledgements We want to thank Dr. Mary Polacco at the E. coli Genetic Stock Centre for providing the strain, Prof. Mathias UhlCn, Dr. Maria Murby and co-workers for providing the plasmid. This investigation was supported by the Swedish Research Council for Engineering Sciences and by a grant from the German Academic Exchange Service to P.N.
References Andersson, L., Strandberg, L. and Enfors, S.-O. (19951 Cell segregation and lysis have profound effects on the growth of E. coli in high cell density fed bath cultures. Biotechnol. Prog., in press. Arbige, M. and Chesbro, W. (1982) relA and related loci are growth rate determinants for Escherichia coli in a recycling fermenter. J. Gen. Microbial. 128, 693-703. Bailey, J.E. (19931 Host-vector interactions in Escherichia cob. Adv. B&hem. Eng. Biotechnol. 48, 29-52. Bimbaum, S. and Bailey, J.E. (1991) Plasmid presence changes the relative levels of many host cell proteins and ribosome components in recombinant Escherichia coli. Biotechnol. Bioeng. 37, 736-745. Bimboim, H.C. and Doly, J. (1979) A rapid alkaline extraction procedure for screening of recombinant plasmid DNA. Nucleic Acids Res. 7, 1513-1523. Bradford, M. (1976) A rapid sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72, 248-254. Bremer, H. and Dennis, P.P. (1987) Modulation of chemical composition and other parameters of the cell by growth rate. In: F.C. Neidhardt, Ingraham, J.L., Low, K.B., Magasanik, B., Schaechter, M. and Umbarger, H.E. @is.), Escherichia coli and Salmonella ryphimurium Cellular and Molecular Biology. Am. Sot. Microbial., Washington D.C., 1527-1542. Cashel, M. and Rudd, K.E. (19871 The stringent response. In: F.C. Neidhardt, Ingraham, J.L., Low, K.B., Magasanik, B.. Schaechter, M. and Umbarger, H.E. (Eds.), Escherichia coli and Salmonella typhimurium Cellular and Molecular Biology. American Society of Microbiology, Washington D.C., 14101438. Chaloner-Larsson, G. and Yamazaki, H. (1978) Effects of the spoT and reZA mutation on the synthesis and accumulation of ppGpp and RNA during glucose starvation. Can. J. B&hem. 56, 264-272. Das, A. (1990) Overproduction of proteins in Escherichia coli: Vectors, hosts, and strategies. Meth. Enzymol. 182, 93-l 12. Flickinger, M.C. and Rouse, M.P. (1993) Sustaining protein synthesis in the absence of rapid cell division: An investigation of plasmid-encoded protein expression in Escherichia coli during very slow growth. Biotechnol. Prog. 9, 555-572.
Gallant, J.A. (1979) Stringent control in E. coli. Ann. Rev. Genet. 13,393-415. Gill. S.C. and von Hippel, P.H. (1989) Calculation of protein extinction coefficients from amino acid sequence data. Anal. B&hem. 182, 319-326. Gleiser, I.E. and Bauer, S. (1981) Growth of E. coli W to high cell concentration by oxygen linked control of carbon source concentration. Biotechnol. Bioeng. 23, 1015-1021. Gaff, S.A., Casson, L.P. and Goldberg, A.L. (19841 Heat shock regulatory gene htpR influences rates of protein degradation and expression of the ion gene in Escherichia coli. Proc. Natl. Acad. Sci. USA 81, 6647-6651. Hengge, R. and Boos, W. (1983) Maltose and lactose transport in Escherichia coli. Examples of two different types of concentrative transport systems. B&him. Biophys. Acta 737, 443478. Holme, T., Arvidsson. S., Lindholm, B. and Pavlu, B. (19701 Enzymes: Laboratory-scale production. Process Biochem. 5, 62-66. Hultman, T., Stihl, S., Moks, T. and Uhlen, M. (1988) Approaches to solid phase DNA sequencing. Nucleosides Nucleotides 7, 629-638. Imanaka, T. (1986) Application of recombinant DNA technology to the production of useful biomaterials. Adv. B&hem. Eng. Biotechnol. 33, l-27. Kaprelyants, A.S., Gottshal, J.C. and Kell, D.B. (19931 Dormancy in non-sporulating bacteria. FEMS Microbial. Rev. 104, 271286. Kim, E., Fu, J., Wilson, D.B. and Shuler, M.L. (1992) Expression of human epidermal growth factor by Escherichia coli in continuous culture. Biotechnol. Lett. 14, 339-344. Kosinski, M.J., Rinas, U. and Bailey, J.E. (1992) Isopropyl-P-Dthiogalactopyranoside influences the metabolism of Escherichia coli. Appl. Microbial. Biotechnol. 36, 782-784. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London) 227, 680-685. Mason, C.A., Bryers, J.D. and Hamer, G. (1986a) Activity, death and lysis during microbial growth in a chemostat. Chem. Eng. Commun. 45, 163-176. Neubauer, P., Hofmann, K., Holst, 0.. Mattiasson, B. and Kruschke, P. (1992) Maximizing the expression of a recombinant gene in Escherichia coli by manipulation of induction time using lactose as inducer. Appl. Microbial. Biotechnol. 36, 739-744. Neubauer, P., Ahman, M., Tomkvist, M., Larsson, G. and Enfors, S.-O. (1995) Response of guanosine tetraphosphate to glucose fluctuations in fed batch cultivations of Escherichia coli. J. Biotechnol., in press. Nilsson, B., Abrahmstn, L. and Uhltn, M. (1985) Immobilization and purification of enzymes with staphylococcal protein A gene fusion vectors. EMBO J. 4, 1075- 1080. Nilsson, L., Oliver, J.D. and Kjelleberg, S. (1991) Resuscitation of Vibrio uulnificus from the viable but nonculturable state. J. Bacterial. 173, 5054-5059. O’Connor, G.M., Sanchez-Riera, F. and Cooney, C.L. (1992)
L Andersson et al. /Journal Design and evaluation of control strategies for high ccl! density fermentations. Biotechnol. Bioeng. 39, 293-304. Oliver, J.D., Nilsson, L. and Kjelleberg, S. (19911 Formation of nonculturable Vibrio uulnificus cells and its relation to the starvation state. Appl. Environ. Microbial. 57, 2640-2644. Reinikainen, P. and Virkaj%rvi, I. (1989) Escherichia coli growth and plasmid copy numbers in continuous cultivations. Biotechnol. Lett. 4, 225-230. Ryan, W. and Parulekar, S. (1991) Recombinant protein synthesis and plasmid instability in continuous cultures of Escherichia coli JM103 harboring a high copy number plasmid. Biotechnol. Bioeng. 37, 415-429. Seo, J.-H. and Bailey, J.E. (1985) Effects of recombinant plasmid content on growth properties and cloned gene product formation in Escherichiu coli. Biotechnol. Bioeng. 27, 1668-1674. Stouthamer, A.H. and Bulthuis, B.A. (19901 Energetics of growth at low growth rate and its relevance for the maintenance
of Biotechnology
46 (1996) 255-263
263
concept. In: R.K. Poole, Bazin, M.J. and Keevil, C.W. (Eds.1, Microbial growth dynamics. Society for General Microbiology, Oxford, New York, Tokyo, 85-102. Voellmy, R. and Goldberg, A.L. (19801 Guanosine-T-diphosphate-3’-diphosphate (ppGpp) and the regulation of protein breakdown in Escherichia coli. J. Biol. Chem. 255, 10081014. Yang, S., Bergman, T., Veide, A. and Enfors, S.-O. (19941 Effects of amino acid insertions on the proteolysis of a staphylococcal protein A derivate in Escherichia co/i. Eur. J. B&hem. 226, 847-852. Yang, S. and Enfors, S.-O. (1995) The influence of energy sources on the proteolysis of a recombinant staphylococcal protein A in Escherichia coli. Eur. J. Biochem. 233, 969-975. Zabriskie, D.W. and Arcuri, E.J. (19861 Factors influencing productivity of fermentations employing recombinant microorganisms. Enzyme Microb. Technol. 8, 706-717.
Journal of Biotechnology
46 (1996) 255-263
Impact of plasmid presence and induction on cellular responses in fed batch cultures of Escherichia coli Lena Andersson
*, Shaojun
Yang, Peter Neubauer ‘, Sven-Olof Enfors
Department of Biochemistry and Biotechnology, Royal Institute of Technology IKTHJ, S-100 44 Stockholm, Sweden Received 21 October
1995; revised 27 December
1995; accepted 28 December
1995
Abstract Fed batch cultivations of plasmid-free and recombinant Escherichia coli were employed in order to determine cellular responses and effects of plasmid presence and induction on the host cell physiology. While plasmid presence was shown to have minor influence on overall biomass yield, induction with 0.1 mM IFTG led to a marked reduction. The number of dividing cells, measured as colony forming ability, was influenced by plasmid presence and to a larger extent by induction. The latter caused a decline in the number of dividing cells to less than 10% of the population within 10 h. However, this cell segregation did not affect the specific rate of product formation, which was approximately constant throughout the cultivations. Analysis of the in vivo degradation rate of the product indicated that it was proteolytically stable. The cellular content of the stringent response signal substance, ppGpp, peaked immediately after transition from batch to fed batch mode to stabilise at a higher value than in the batch phase. When the specific growth rate declined below 0.06 h- ’ an additional rise in ppGpp concentration was observed. Keywords: Recombinant
protein production;
Cellular responses;
Stringent response; Proteolysis;
1. Introduction Efforts in finding strategies to maximise the productivity of recombinant protein production in E. coli are well documented in literature. Numerous studies have focused on gaining more optimal solutions by molecular biological methods and vector constructions (Imanaka, 1986; Das, 19901. Others have been concerned with enhancing the productivity
* Corresponding author. Tel. (+46-g) 7907503, fax. (+46-g) 7231890. ’ Present address: Institut fur Angewandte Biochemie, Fachbereich Biochemie/Biotechnologie, Martin-Luther-Universitiit Halle-Wittenberg, D-06099 Halle, Germany. Elsevier Science B.V. PII SOl68-1656(96)00004-l
Fed batch; E. coli
by engineering means by employing high cell density cultures and control strategies to avoid inhibiting mechanisms like by-product formation and oxygen limitation (Gleiser and Bauer, 1981; O’Connor et al., 1992). However, growing cells to high densities aerobically includes the use of severely energy limited processes, of which the physiological consequences are not so well described. Recombinant protein production is an interdisciplinary art: both genetic and process engineering are essential parts in the development of a bioprocess. The link between the two is the host cell and its metabolic and regulatory characteristics. The amount of protein produced from a recombinant strain is strongly determined by factors such as host cell-vec-
256
L. Andersson
et cd. /Journal
tor interactions, plasmid segregational instability and the extent of cellular stress responses (Zabriskie and Arcuri, 1986; Kim et al., 1992; Bailey, 1993). A commonly reported host cell-vector interaction is the negative effect of plasmid presence on growth rate and overall growth yield (Sea and Bailey, 1985; Reinikainen and Virkajlrvi. 1989). Possible causes for increased maintenance metabolism are redirection of cellular resources to support plasmid-related activities and inhibitory mechanisms on host cell metabolism. Alterations in levels of E. coli cell proteins and ribosomal components have been shown to occur after introduction of multicopy plasmids (Bimbaum and Bailey, 1991). The levels of stress proteins were higher for recombinant strains, while metabolic enzymes showed lower values. Many studies have reported growth rate related segregational plasmid instability (Ryan and Parulekar, 1991; Flickinger and Rouse, 1993). Another segregation phenomenon observed at slow growth of plasmid-free E. coli, is the occurrence of an increasing fraction of cells that have lost its colony forming ability but still exhibits metabolic activity (Andersson et al., 1995). The physiological status of this type of subpopulation is still not known. A possible loss of biosynthetic capacity may however have detrimental influence on protein production. Cellular stress responses like heat shock and stringent response may also have profound effects on the final amount and quality of product. Besides affecting many other processes, these global regulation systems are known to induce or augment protease activity (Voellmy and Goldberg, 1980; Goff et al., 1984). There is strong evidence that elevated levels of the nucleotide guanosine S-diphosphate 3’-diphosphate (ppGpp), evoked by depletion of amino acids or by severe energy/carbon limitation, trigger the stringent response (Gallant, 1979; Cashel and Rudd, 1987; Stouthamer and Bulthuis, 1990). Intracellular concentrations of ppGpp have also been shown to be inversely correlated to specific growth rate and the amount of stable RNA transcripts (Arbige and Chesbro, 1982; Bremer and Dennis, 1987). In this paper, we have focused on the behaviour of plasmid-free and recombinant E. coli during growth in energy limited cultures. Biomass yield, segregation to non-dividing cells and levels of ppGpp
of Biotechnology 46
(19%)
255-263
were compared. The rate of product formation and proteolysis in vivo were determined during production of the recombinant protein, ZZ.
2. Material and methods 2. I. Strain and plasmid The bacterial strain used, RB791 (lacIqL,), was kindly provided from the E. coli Genetic Stock Centre. The plasmid pRIT28NZZT is a derivate of pRIT28 (Hultman et al., 19881, into which a gene encoding protein ZZ has been inserted. This protein is composed of two IgG-binding staphylococcal protein A domain analogues (Zl (Nilsson et al., 1985). The production of the 15 kDa protein is under control of the lacUV5 promoter. The expression was induced by addition of 0.1 mM isopropyl P-Dthiogalactopyranoside, IPTG, to the cultures (and feed solutions). 2.2. Cultiuation medium The culture medium (main and precultures) had the following composition (g 1-l 1: Na,SO, 2.0; NH,Cl 0.5; Na,HPO,. 2H,O 3.0; KH,PO, 7.0; (NH,)2S0, 1.0; thiamine 0.1; ampicillin 0.07. The medium also contained trace component solution 2 ml 1-l (Holme et al., 1970) and 1 M MgSO, 2 ml l- ’ . The initial glucose concentration was 10 g 1- ’ . The feed solution contained essentially the same components as the initial medium with the exceptions that no MgSO, was added, the trace component solution and glucose concentrations were 8 ml I-’ and 150 g l-l, respectively. Additional 8 ml MgSO, (1 Ml was added intermittently to the cultivations to avoid precipitation. The pH was controlled at 7 with 4 M H,SO, and 25% NH,OH. Adecanol LG- 109 (antifoam) was added when necessary. 2.3. Cultiuation conditions Inoculum was prepared from shake flask cultures grown at 37°C. The experiments were carried out in a 7 1 bioreactor with an initial culture volume of 3.5 1. Temperature was kept at 37°C. Air flow rate was 2.0 1 min-‘. The stirrer speed was kept at 800 RPM
L.. Andersson et al. / Journal of Biotechnology 46 (1996) 255-263
and the dissolved oxygen tension did not fall below 20% in any cultivation. The cultivations were started as batch cultures and the feed, at a rate of 40 ml h- ‘, was started when the glucose concentration was between 1 and 2 g 1-l. 2.4. Analytical
methods
Optical density was measured at 580 nm. Dry weight was determined by centrifuging 3 X 5 ml of cell suspension in preweighed tubes, washing the samples with distilled water and drying overnight at 105°C before weighing. Analyses of number of dividing cells (colony forming units, cfu) were performed on nutrient agar plates, with and without ampicillin (0.07 g l- ’ >, incubated at 37°C for 24 h. Dissolved oxygen tension was continuously measured with a polarographic oxygen electrode. The oxygen and carbon dioxide content in the effluent gas were analysed with a paramagnetic analyser (Servomex Oxygen Analyser 540A, Sybron/Taylor, UK) and an IR analyser (Binos, Leybold-Hereaus, Germany), respectively. Plasmid content in the cells was detected from 0.25 to 1 ml process samples by a mini-preparation method based on the alkaline sodium dodecyl method of Bimboim and Doly (Bimboim and Doly, 1979). Plasmid purity was controlled by agarose gel electrophoresis in 0.8% agarose gels and plasmid concentration was measured with a GeneQuant RNA/DNA Calculator (Pharmacia Biotech, Uppsala, Sweden). Total protein concentration in the medium was measured by the Bradford method (Bradford, 1976) with BSA as standard. Samples were taken from the bioreactor, centrifuged and filtered before freezing at - 20°C. Samples for analysis of the protein ZZ concentration were subjected to high pressure homogenisation prior to protein purification using IgG-affinity chromatography (Yang et al., 1994). The elution of ZZ was monitored by following the absorbance at 280 nm. The extinction coefficient used was 0.152 emg ml-’ cm-’ according to the amino acid composition (Gill and von Hippel, 1989). SDS-PAGE analysis of the IgG-binding fraction was performed with 3.5% stacking gel and 16% separation gel (Laemmli, 1970).
2.51
Proteins were stained with Coomassie Brilliant Blue (CBB) R-250. 1 pg of purified protein was loaded to each electrophoresis lane. To analyse effects of proteolysis on the product, in vivo proteolysis of ZZ protein at different culture times after induction was measured by using chloramphenicol to stop de novo synthesis of proteins (Yang and Enfors, 1995). Samples, withdrawn from cultivations at different culture times, were diluted with minimal salt medium to about OD = 4 and incubated at 37°C. The mineral salt medium consisted of (g 1-l): (NI-I,),SO, 7; KI-I,pO, 1.6; Na,HPO, .2H,O 6.6 and 100 pg ml-’ chloramphenicol with an initial pH of 7.0. Samples were removed after 0, 30 and 60 min of incubation, centrifuged and stored at -20°C. The composition of the IgG-binding protein fraction was determined by Western blotting analysis as described elsewhere (Yang and Enfors, 1995). The same amount of cells was loaded to each electrophoresis lane. Cellular ppGpp concentration was analysed by an HPLC method on a Supercosil LC-18T column as described elsewhere (Neubauer et al., 1995).
3. Results 3.1. Efsects of plasmid growth kinetics
presence
and induction
on
Effects of plasmid presence and chemically induced gene expression (0.1 mM IPTG) were studied in glucose limited fed batch cultivations of E. coli. Fig. 1A shows growth curves from three types of cultures: host strain not bearing plasmid, host strain transformed with plasmid pRIT28NZZT encoding the intracellular protein ZZ, but without induction and finally with induction after 3 h of glucose limited growth. Initially, the accumulated biomass (g) increased almost linearly with the same rate for the different cultivations but at about 10 h, when /.Lwas about 0.05 h- ‘, the rates started to decline gradually (Fig. lA, C). The induced culture then grew at lower rate and the final amount of biomass achieved was considerably lower (23%) compared to the other cultures. The growth rate of the plasmid containing but non-induced culture was intermediate. The results indicated that induction with 0.1 mM IPTG led
258
et al./Joumul of Biotechnology 46 (1996) 255-263
L. Anderssan
to an increased maintenance requirement, as a larger fraction of the substrate was used for non-growth purposes, whereas the contribution by plasmid presence alone was less drastic. These findings were further reflected in data on specific glucose consumption rate, qs (Fig. 1B). The values of qs for all cultures rapidly declined during the first 10 h to approach a lower limit where most of the sugar feed was consumed for maintenance purpose. The qsvalue for the induced culture was about 20% higher than for the other two cultures at 30 h. The number of dividing ceils, measured as colony
\
d
\ 0. ‘CL-_:=*;-+ I
0
10
la I
,-
20
I
30
Time (h) Fig. 2. Comparison of colony forming ability (cfu ml-’ ) of plasmid-free cells (without plasmid), non-induced, plasmid-bearing cells (without induction) and induced, plasmid-bearing cells (induction). Induction was performed at 3 h (indicated by arrow) with 0.1 mM IPTG.
Ol
0.2
/
1
1
j
I
I c
-
0.15-
--CL
0.1 -:%kE
L
:
A .
0.05 -
‘“L-, b
$-.-
81
0
1 0
I IO
,
Time
$=-kP I 20
I
I 30
(h)
Fig. I. Comparison of cell growth during glucose limited fed batch cultivations of three types of cells; plasmid-free cells (without plasmid), non-induced, plasmid-bearing cells (without induction) and induced, piasmid-bearing cells (induction). (A) Cell mass (g); (B) specific substrate uptake rate, qs; (hh ’ ) (C) specific growth rate, k (hh’ ).
forming ability (cfu ml- ’ >, differed considerably among the three cultures, as can be seen in Fig. 2. The cfu-values for the induced culture started to drop after induction (3 h) and after 15 h the colony forming ability was only 10% of the maximum value. In the cultivation with plasmid-bearing cells, without induction, the cfu-values exhibited a similar pattern with the difference that the decline appeared later, at about 10 h. The difference in colony forming ability on non-selective media and selective media (with ampicillin) was insignificant (not shown), indicating that the observed response was not caused by formation of plasmid-free segregants. Colony forming ability in the cultivation with plasmid-free cells remained high until 25 h, when it started to decline. To investigate if the loss of colony forming ability was a consequence of cell lysis the amount of extracellular proteins in the culture media was analysed (Fig. 3). During the first 15 h no substantial difference in accumulation of extracellular proteins could be seen between the induced culture and the plasmid-free culture. Thereafter, the extracellular protein content in the induced culture increased drastically to become 130% higher (9.3 g) than in the plasmid-free culture (4.0 g> at the end of the experiment. These results suggest that cell lysis occurred in later stages
L Andersson et al./Joumal
of Biotechnology 46 (1996) 255-263
259
of all cultures, but that it was more pronounced in the induced culture. The onset of increasing accumulation rate of extracellular proteins correlates with the time when p approached 0.03 hh’ in all cultures. Assuming that the cellular protein content was 50%, cell lysis would account for about 65% of the difference in finally achieved biomass between the cultures. However, this amount of extracellular protein corresponds only to 26% of the final sum of cell- and extracellular proteins, which rules out cell lysis as an explanation for the more than 90% loss of colony forming ability. 1
I
r
,
20
3.2. Cellular responses and product fonnation
I
I
3
30
Time (h)
Data on intracellular product formation showed that a low amount of protein ZZ was present before induction (Fig. 4). Following induction, the specific protein content (mg ZZ per g cells) increased roughly linearly throughout the cultivation. The final specific amount of ZZ corresponded to approximately 5% of the total cell mass. To measure the rate of proteolysis in vivo, samples were taken from the bioreactor intermittently, and incubated O-60 min with 100 pg ml-’ chloramphenicol prior to analysis of the composition of the IgG-binding protein fraction. Western blot analyses of samples taken at different times during the process showed that the amount of the full
lrlf
’
’
I
10
’
’
20
’
’
’
I
30
Time (h) Fig. 3. Extracellular protein content (g) of plasmid-free cells (without plasmid), non-induced, plasmid-bearing cells (without induction) and induced, plasmid-bearing cells (induction).
Fig. 4. Product formation of protein ZZ (mg ZZ per g cells) in a glucose limited fed batch cultivation. The production system was induced with 0.1 mM IPTG after 3 h of limited growth (as indicated).
length product was constant during the 60 min of incubation, indicating that no degradation took place (results not shown). However, several IgG-binding bands with lower molecular mass than the main band were also visible. These bands remained essentially constant for all samples during the incubation with chloramphenicol. In contrast, SDS-PAGE analysis of the IgG-binding protein fraction recovered by affinity chromatography did not reveal any other IgG-binding bands than tbe product band (results not shown). The concentration of the main effector of stringent response, ppGpp, was determined in fed batch cultivations for production of protein ZZ. The experiments were started as batch cultures and a constant feed of glucose was supplied to the bioreactor before the initially added glucose was exhausted (0 h) to exclude any period of glucose starvation. Recombinant protein production was induced after 3 h of limited growth. Fig. 5A shows the cellular ppGpp content during the different phases of an experiment with induced cells. The first measurement of ppGpp (60 nmol per g cells), at -0.5 h, was taken during growth in batch phase. This concentration was lower than the succeeding values from the fed batch phase. Other cultures exhibited a similar pattern of cellular ppGpp content (data not shown). The downshift from
L. Andersson
260
(4 3 = s m
600
of Biotechnology
46 (19961255-263
t
i
B 2 c
+
et al. /Journal
400
; E g 200 c3 E? $ 3 J
0
I
2,s
5
f
0.1
0.05 Specific
35
growth
rate (h’)
Fig. 5. (A) cellular ppGpp concentration (nmol per g cells) during different phases in the cultivation for production of recombinant protein ZZ. The culture was started as a batch culture and a constant feed of glucose was supplied into the bioreactor before the initial added glucose was exhausted (0). The last measurement was done after that the substrate feed had been shut off for 30 min. (B) Cellular ppGpp concentration (nmol per g cells) as function of specific growth rate
batch to fed batch mode (glucose limitation) resulted in a sharp peak in ppGpp content (500 nmol per g cells) and the concentration remained elevated at approximately 100 nmol per g cells during the subsequent 5 h. A second, significant rise was observed in ppGpp concentration at 10 h before it stabilised at a new, higher level at about 175 nmol per g cells, where it remained throughout the cultivation. The last measurement, at 35.5 h, was done after that the substrate feed had been shut off for 30 min, reflecting the response when the cells were exposed to complete energy/carbon starvation. Cellular ppGpp content as a function of specific growth rate during the fed batch phase (3-35 h), is presented in Fig. 5B.
The ppGpp content increased at a time when the specific growth rate had declined to a low value around 0.06 h- ’ . Plasmid-free and non-induced plasmid-bearing cultures also showed a high ppGpp peak in the transition from batch to fed batch and an increased ppGpp level during the late stages of the fed batch phase (data not shown).
4. Discussion The fed batch technique is widely employed in industrial processes. Often, at least during part of the process, a constant feed rate is applied. The charac-
L. Andersson et al./Joumal
of Biotechnology 46 (1996) 255-263
teristic features of such cultures arc progressively increasing substrate limitation and declining specific growth rate. The present investigation revealed a number of physiological responses during such conditions. Furthermore, differences in behaviour between plasmid-free, non-induced plasmid-bearing and induced cultures were found. Induction of plasmid encoded protein ZZ with 0.1 mM IPTG was shown to have strong influence on biomass formation due to increased maintenance metabolism and cell lysis. On the other hand, a high content of plasmids, above 200 ng plasmid per g cells throughout the experiments (data not shown), was shown to have a somewhat smaller effect on the biomass yield. A substantial fraction of the loss in biomass could be explained by analysis of extracellular protein content. Consequently, the observed increase in specific substrate consumption rate could only to a lesser extent be comprised by an additional energy demand. Cell lysis may still, from a macroscopic point of view, be regarded as part of the maintenance demand. The expression system used resulted in relatively low amounts of recombinant protein, with a final value at about 5% of the cell mass. This implies that production only exerts a modest demand on metabolic activity of the host cell and that the lethal effect on cell growth should be attributed to IPTG itself. Previous studies have shown that addition of IPTG to wild-type E. coli leads to elevated expression of several stress response proteins (Kosinski et al., 1992). The observed loss of biomass in the present study corresponded to a 23% decrease in final, conceivable product concentration, which makes utilisation of IPTG as inducer quite questionable, even at as low dosage as 0.1 n&I. Instead, IPTG might be replaced by lactose or glycerol-P-galactoside (Hengge and Boos, 1983; Neubauer et al., 1992). The phenomenon of cells that respirate but have lost their ability to form colonies on nutrient agar plates, viable but non-culturable cells, seems to be associated with stress of nutrient limitation and is well recognised in studies of bacterial starvation (Mason et al., 1986a; Oliver et al., 1991; Kaprelyants et al., 1993). Moreover, it has been shown that it is possible to resuscitate some of these bacterial forms in liquid medium, thus the state appears to be reversible (Nilsson et al., 1991). In the present work, a
261
segregation of the populations where an increasing fraction of the cells had lost their ability to form colonies occurred in all cultures, but at distinctly different times. These results suggest that the cultures were exposed to different magnitudes of stress and that induction with IPTG appears to have triggered cell segregation. The segregation that took place during the first 10 h after induction did not result in reduced specific productivity (on dry weight basis) in spite of the reduction of the cfu-values by more than 90%. Thus, the non-dividing population contributes to the protein production. The Western blotting analyses of cells incubated with chloramphenicol to inhibit protein synthesis indicated that the product was stable with respect to proteolysis during all phases of the process. The IgG-binding bands with lower molecular mass than the main band revealed by this analysis were not found when the IgG-binding protein fraction recovered by affinity chromatography was analysed by SDS-PAGE and CBB staining. This divergence can be explained from the fact that the applied analysis methods differ in their accuracy with regard to qualification and quantification of proteins. Thus, the Western blotting analysis detects extremely low concentrations of proteins whereas SDS-PAGE with CBB staining gives a more reliable estimation of the amount of protein. The cellular ppGpp content increased transiently after the first carbon/energy down-shift from batch to fed batch mode and stabilised at a higher level than in the batch phase. It is interesting to note that even if the culture was not exposed to carbon starvation, but limitation, the measured peak of ppGpp, 500 nmol per g cells is similar to those that have been observed during complete glucose starvation (Chaloner-Larsson and Yamazaki, 1978). Later in the cultivations at a low specific growth rate an additional, significant rise in ppGpp content was observed, indicating conditions of severe energy/carbon limitation. The last ppGpp measurement, 275 nmol per g cells, was taken after the culture had been exposed to complete carbon starvation for 30 min. This low value, when compared to the first peak, might be explained by a limitation in the amount of cellular ATP available for synthesis of ppGpp at that late stage of the cultivation.
262
L. Andersson et al./ Journal of Biotechnology 46 (1996) 255-263
Acknowledgements We want to thank Dr. Mary Polacco at the E. coli Genetic Stock Centre for providing the strain, Prof. Mathias UhlCn, Dr. Maria Murby and co-workers for providing the plasmid. This investigation was supported by the Swedish Research Council for Engineering Sciences and by a grant from the German Academic Exchange Service to P.N.
References Andersson, L., Strandberg, L. and Enfors, S.-O. (19951 Cell segregation and lysis have profound effects on the growth of E. coli in high cell density fed bath cultures. Biotechnol. Prog., in press. Arbige, M. and Chesbro, W. (1982) relA and related loci are growth rate determinants for Escherichia coli in a recycling fermenter. J. Gen. Microbial. 128, 693-703. Bailey, J.E. (19931 Host-vector interactions in Escherichia cob. Adv. B&hem. Eng. Biotechnol. 48, 29-52. Bimbaum, S. and Bailey, J.E. (1991) Plasmid presence changes the relative levels of many host cell proteins and ribosome components in recombinant Escherichia coli. Biotechnol. Bioeng. 37, 736-745. Bimboim, H.C. and Doly, J. (1979) A rapid alkaline extraction procedure for screening of recombinant plasmid DNA. Nucleic Acids Res. 7, 1513-1523. Bradford, M. (1976) A rapid sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72, 248-254. Bremer, H. and Dennis, P.P. (1987) Modulation of chemical composition and other parameters of the cell by growth rate. In: F.C. Neidhardt, Ingraham, J.L., Low, K.B., Magasanik, B., Schaechter, M. and Umbarger, H.E. @is.), Escherichia coli and Salmonella ryphimurium Cellular and Molecular Biology. Am. Sot. Microbial., Washington D.C., 1527-1542. Cashel, M. and Rudd, K.E. (19871 The stringent response. In: F.C. Neidhardt, Ingraham, J.L., Low, K.B., Magasanik, B.. Schaechter, M. and Umbarger, H.E. (Eds.), Escherichia coli and Salmonella typhimurium Cellular and Molecular Biology. American Society of Microbiology, Washington D.C., 14101438. Chaloner-Larsson, G. and Yamazaki, H. (1978) Effects of the spoT and reZA mutation on the synthesis and accumulation of ppGpp and RNA during glucose starvation. Can. J. B&hem. 56, 264-272. Das, A. (1990) Overproduction of proteins in Escherichia coli: Vectors, hosts, and strategies. Meth. Enzymol. 182, 93-l 12. Flickinger, M.C. and Rouse, M.P. (1993) Sustaining protein synthesis in the absence of rapid cell division: An investigation of plasmid-encoded protein expression in Escherichia coli during very slow growth. Biotechnol. Prog. 9, 555-572.
Gallant, J.A. (1979) Stringent control in E. coli. Ann. Rev. Genet. 13,393-415. Gill. S.C. and von Hippel, P.H. (1989) Calculation of protein extinction coefficients from amino acid sequence data. Anal. B&hem. 182, 319-326. Gleiser, I.E. and Bauer, S. (1981) Growth of E. coli W to high cell concentration by oxygen linked control of carbon source concentration. Biotechnol. Bioeng. 23, 1015-1021. Gaff, S.A., Casson, L.P. and Goldberg, A.L. (19841 Heat shock regulatory gene htpR influences rates of protein degradation and expression of the ion gene in Escherichia coli. Proc. Natl. Acad. Sci. USA 81, 6647-6651. Hengge, R. and Boos, W. (1983) Maltose and lactose transport in Escherichia coli. Examples of two different types of concentrative transport systems. B&him. Biophys. Acta 737, 443478. Holme, T., Arvidsson. S., Lindholm, B. and Pavlu, B. (19701 Enzymes: Laboratory-scale production. Process Biochem. 5, 62-66. Hultman, T., Stihl, S., Moks, T. and Uhlen, M. (1988) Approaches to solid phase DNA sequencing. Nucleosides Nucleotides 7, 629-638. Imanaka, T. (1986) Application of recombinant DNA technology to the production of useful biomaterials. Adv. B&hem. Eng. Biotechnol. 33, l-27. Kaprelyants, A.S., Gottshal, J.C. and Kell, D.B. (19931 Dormancy in non-sporulating bacteria. FEMS Microbial. Rev. 104, 271286. Kim, E., Fu, J., Wilson, D.B. and Shuler, M.L. (1992) Expression of human epidermal growth factor by Escherichia coli in continuous culture. Biotechnol. Lett. 14, 339-344. Kosinski, M.J., Rinas, U. and Bailey, J.E. (1992) Isopropyl-P-Dthiogalactopyranoside influences the metabolism of Escherichia coli. Appl. Microbial. Biotechnol. 36, 782-784. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London) 227, 680-685. Mason, C.A., Bryers, J.D. and Hamer, G. (1986a) Activity, death and lysis during microbial growth in a chemostat. Chem. Eng. Commun. 45, 163-176. Neubauer, P., Hofmann, K., Holst, 0.. Mattiasson, B. and Kruschke, P. (1992) Maximizing the expression of a recombinant gene in Escherichia coli by manipulation of induction time using lactose as inducer. Appl. Microbial. Biotechnol. 36, 739-744. Neubauer, P., Ahman, M., Tomkvist, M., Larsson, G. and Enfors, S.-O. (1995) Response of guanosine tetraphosphate to glucose fluctuations in fed batch cultivations of Escherichia coli. J. Biotechnol., in press. Nilsson, B., Abrahmstn, L. and Uhltn, M. (1985) Immobilization and purification of enzymes with staphylococcal protein A gene fusion vectors. EMBO J. 4, 1075- 1080. Nilsson, L., Oliver, J.D. and Kjelleberg, S. (1991) Resuscitation of Vibrio uulnificus from the viable but nonculturable state. J. Bacterial. 173, 5054-5059. O’Connor, G.M., Sanchez-Riera, F. and Cooney, C.L. (1992)
L Andersson et al. /Journal Design and evaluation of control strategies for high ccl! density fermentations. Biotechnol. Bioeng. 39, 293-304. Oliver, J.D., Nilsson, L. and Kjelleberg, S. (19911 Formation of nonculturable Vibrio uulnificus cells and its relation to the starvation state. Appl. Environ. Microbial. 57, 2640-2644. Reinikainen, P. and Virkaj%rvi, I. (1989) Escherichia coli growth and plasmid copy numbers in continuous cultivations. Biotechnol. Lett. 4, 225-230. Ryan, W. and Parulekar, S. (1991) Recombinant protein synthesis and plasmid instability in continuous cultures of Escherichia coli JM103 harboring a high copy number plasmid. Biotechnol. Bioeng. 37, 415-429. Seo, J.-H. and Bailey, J.E. (1985) Effects of recombinant plasmid content on growth properties and cloned gene product formation in Escherichiu coli. Biotechnol. Bioeng. 27, 1668-1674. Stouthamer, A.H. and Bulthuis, B.A. (19901 Energetics of growth at low growth rate and its relevance for the maintenance
of Biotechnology
46 (1996) 255-263
263
concept. In: R.K. Poole, Bazin, M.J. and Keevil, C.W. (Eds.1, Microbial growth dynamics. Society for General Microbiology, Oxford, New York, Tokyo, 85-102. Voellmy, R. and Goldberg, A.L. (19801 Guanosine-T-diphosphate-3’-diphosphate (ppGpp) and the regulation of protein breakdown in Escherichia coli. J. Biol. Chem. 255, 10081014. Yang, S., Bergman, T., Veide, A. and Enfors, S.-O. (19941 Effects of amino acid insertions on the proteolysis of a staphylococcal protein A derivate in Escherichia co/i. Eur. J. B&hem. 226, 847-852. Yang, S. and Enfors, S.-O. (1995) The influence of energy sources on the proteolysis of a recombinant staphylococcal protein A in Escherichia coli. Eur. J. Biochem. 233, 969-975. Zabriskie, D.W. and Arcuri, E.J. (19861 Factors influencing productivity of fermentations employing recombinant microorganisms. Enzyme Microb. Technol. 8, 706-717.