- Email: [email protected]
Heat-inactivation of plasmid-encoded CI857 repressor induces gene expression from Ind− lambda prophage in recombinant Escherichia coli
FEMS Microbiology Letters 177 (1999) 327^334
Heat-inactivation of plasmid-encoded CI857 repressor induces gene expression from Ind3 lambda prophage in recombinant Escherichia coli Frank Ho¡mann a , Anna Ar|¨s b , Xavier Carbonell b , Manfred Rohde c , Jose¨ L. Corchero b;1 , Ursula Rinas a , Antonio Villaverde b; * a
Biochemical Engineering, GBF National Research Center for Biotechnology, Mascheroder Weg 1, 38124 Braunschweig, Germany Institut de Biologia Fonamental and Departament de Gene©tica i Microbiologia, Universitat Auto©noma de Barcelona, Bellaterra, 08193 Barcelona, Spain c Microbiology Division, GBF National Research Center for Biotechnology, Mascheroder Weg 1, 38124 Braunschweig, Germany
b
Received 27 January 1999; received in revised form 30 April 1999; accepted 22 June 1999
Abstract We have observed significant cell lysis upon temperature up-shift of recombinant Escherichia coli cultures harboring CI857repressed lambda-based expression vectors. This event, that becomes evident about 30^40 min after the heat shock, takes place when using the lambda promoter system in Ind3 lysogenic strains, but not in others commonly employed for recombinant gene expression. These results strongly suggest that the thermosensitive CI857 repressor, encoded by the expression vector, competes with CI Ind3 molecules for binding to the prophage operator region, allowing for expression of lytic genes from the integrated Ind3 viral genome upon temperature up-shift. Transcription of viral lytic genes does not include unspecific expression of a reporter sulA: :lacZ gene fusion carried in the prophage genome. These results prompt, however, to carefully evaluate the limitations of expression systems based on pL /pR -CI857 in bacterial strains modified through lambda Ind3 gene transfer vehicles. ß 1999 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. Keywords : Lambda bacteriophage; CI857 repressor ; Recombinant protein; Structural gene ; Cell lysis
1. Introduction The overproduction of heterologous proteins in Escherichia coli causes a severe metabolic burden
* Corresponding author. Tel.: +34 (93) 5812148; Fax: +34 (93) 5812011; E-mail: [email protected] 1 Present address: VEHB, MailStop G 18, Center for Disease Control, 1600 Clifton Road, Atlanta, GA 30333, USA.
[1^5], which is often accompanied by direct toxicity of the products on the producing cells. Although still unidenti¢ed, there are several indications suggesting a multifactorial nature of the deleterious e¡ects associated to recombinant protein synthesis [6^14]. In this line, we have recently reported induction of SOS genes in E. coli indicator strains in response to thermal induction of plasmid-encoded recombinant genes contained in CI857-controlled, lambda-based expression vectors [15]. Since the nature of this in-
0378-1097 / 99 / $20.00 ß 1999 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 1 0 9 7 ( 9 9 ) 0 0 3 3 4 - 1
FEMSLE 8924 29-7-99
328
F. Ho¡mann et al. / FEMS Microbiology Letters 177 (1999) 327^334
triguing event remained still unsolved, we were prompted to investigate in more detail the possible mechanisms linking thermal induction of recombinant gene expression and SOS transcription. By analyzing the growth pattern of an indicator strain carrying a sulA: :lacZ gene fusion, we evidenced a partial but signi¢cant lysis of the culture about 30^40 min after the temperature up-shift, accompanied by the release of incomplete lambda virus particles to the medium. This observation revealed expression of lytic genes from the Ind3 prophage genome, used as a vehicle for the reporter sulA: :lacZ gene fusion, that is probably caused by an interference of the thermosensitive CI857 repressor with the natural activities of the CI Ind3 repressor encoded by the lambda prophage DNA, its inactivation when the bacterial culture is submitted to a thermal up-shift and the consequent derepression of lytic genes. Therefore, these results show the necessity to carefully evaluate the limitations of expression systems based on pL /pR -CI857 in bacterial strains modi¢ed through lambda Ind3 gene transfer vehicles.
2. Materials and methods 2.1. Bacterial strains, plasmids and proteins The E. coli strain GC4581, F3 vlac-pro (V cIind1 (sulA: :lacZ) RecA ) [16], was used to monitor expression of the SOS sulA (formerly s¢A) gene through L-galactosidase activity during plasmid-encoded recombinant gene expression. A vrecA derivative of GC4581 (named LMA1000) has been constructed by P1 transduction from the donor strain FL8641/pFL352, (F3 ) araD139 v(argF-lac)169 £hD5301 fruA25 relA1 rpsL150(Strr ) rbsR22 deoC1 v(gal-VG) Vb2: : x((cea pup-lacZ)2(Hyb) lacY ) v(recA-srlR)306 srlR301: :Tn10-84/pFL352(bla recA ) [17], followed by clone selection by tetracycline resistance and further testing of the RecA3 phenotype by determining ultraviolet radiation sensitivity. Strains MC1061, hsdR2 mcrB araD139 v(araABCleu)7679 vlacX74 galU galK rpsL thi strA [18], BL21, hsdS gal3 Lon3 OmpT3 [19], GE864 MC4100, v(gal-GV ) b2V : :((cea Pup-lacZ)Z (Hyb)
lacY ) [17], and TG1, supE hsdv5 thi v(lac-proAB) F'(traD36 proAB lacIq lacZvM15) [20], were also employed for recombinant gene expression. The construction of the expression vector pVFGFB, a derivative of pCYTEXP1, has been described before [21]. It harbors the gene for human basic ¢broblast growth factor (bFGF). Plasmid pJVP1 derives from pJLA602 [22] and it encodes a complete VP1 protein from foot and mouth disease virus (FMDV) [5]. All these multi-copy plasmids harbor an ampicillin resistance gene and the CI857 repressor gene under the control of its own promoter. Expression of recombinant genes in the described expression vectors is controlled by the lambda pR and pL promoters placed in tandem. 2.2. Culture conditions and recombinant gene expression Luria-Bertani (LB) medium [18] with 40 Wg ml31 streptomycin and 100 Wg ml31 ampicillin was used for culture growth. Aliquots from overnight cultures were used to inoculate fresh medium at 1:50, in 100ml shaker £asks, with a working volume of 20 ml. These cultures were incubated at 250 rpm and 28³C. At an optical density (OD550 ) of 0.2^0.5 U, recombinant gene expression was induced by transfer to a pre-warmed bath at 42³C and further incubation at the same agitation rate. When indicated, the potent SOS inducers nalidixic acid or mitomycin C were added simultaneously to the temperature up-shift at 40 or 1.6 Wg ml31 , respectively. 2.3. Monitoring of culture growth, cell integrity and lacZ gene expression Cell growth and lysis were monitored by regular measurements of OD550 . L-Galactosidase activity was determined as described previously [15]. The enzymatic activity was also measured in the cell-free culture supernatant and the cell pellet obtained after centrifugation at 14000 rpm for 3 min. Cell pellets were re-suspended in the original volume of LB medium. Cell permeabilization was achieved by vortexing in chloroform-containing Z bu¡er (0.06 M Na2 HPO4 , 0.04 M NaH2 PO4 , pH 7.2, 0.01 M KCl, 1 mM MgSO4 , 0.04 M L-mercaptoethanol).
FEMSLE 8924 29-7-99
F. Ho¡mann et al. / FEMS Microbiology Letters 177 (1999) 327^334
2.4. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) For SDS-PAGE analysis, cell pellets were re-suspended in 50 mM sodium phosphate bu¡er (pH 7) to an OD of 4.5. After sonication for 2 min, soluble and insoluble fractions were collected by centrifugation for 45 min at 38 000Ug and 4³C. The insoluble fraction was washed with sodium phosphate bu¡er, centrifuged again and re-suspended in the same bu¡er in 1/4 of the original volume. SDS-PAGE analysis was performed according to the method of Laemmli [23]. Samples were mixed with loading bu¡er (1:1) consisting of 7% (w/v) SDS, 35% (w/v) glycerol, 150 mM dithiothreitol, 250 mM Tris-HCl (pH 6.8), boiled for 10 min and ran immediately on 9^16% polyacrylamide gradient gels. 2.5. Procedures for bacteriophage detection Samples from lysed cultures were plated on con£uent cultures of MC1061 and MC1061 carrying pJLA602 and incubated at di¡erent temperatures from 28 to 42³C to detect phage plaques. Plasmid pJLA602 was added to provide CI857 repressor to reproduce as much as possible the situation observed in cell culture. Furthermore, aliquots of these samples were also negatively stained and prepared for visualization by transmission electron microscopy following the procedure described by Valentine and co-authors [24]. Brie£y, thin carbon ¢lms were prepared by sublimation of carbon onto freshly cleaved mica. Samples were then adsorbed onto small pieces of the carbon ¢lm for 30 s and negatively stained with 4% uranyl acetate dissolved in water (pH 5.0). After air-drying, samples were examined in a Zeiss transmission electron microscope EM910 at 80 kV and at calibrated magni¢cations using a cross-lined grating replica.
329
pL tandem promoter and its thermosensitive CI857 repressor. The use of the reporter strain GC4581 allows for monitoring of the SOS sulA gene transcription. The time-dependent evolution of the reporter L-galactosidase activity in the induced recombinant culture (Fig. 1A) was comparable to previously published results on the induction of the SOS gene upon temperature up-shift in temperatureinducible expression systems [15,25]. However, sulA gene expression is also observed in the strain harboring the parental vector pCYTEXP1 without any cloned gene, suggesting that the biosynthesis of a complete protein is not required for the activation of the sulA transcription. On the other hand, frequent measurements of the OD in these cultures indicated a lytic episode in both strains (Fig. 1B) that was not noticed in previous experiments with less frequent OD measurements [15]. The drop in the OD started around 30^40 min after the temperature shift and seemed to involve the lysis of a major part of the cell population. This OD decrease was reproducible and time-coincident when comparing di¡erent independent experiments.
3. Results and discussion 3.1. Cell lysis during induction of recombinant gene expression Protein bFGF was produced in E. coli as directed from an expression vector based on the lambda pR /
Fig. 1. Kinetics of sulA : :lacZ expression (A) and cell growth (B) after thermal up-shift of the indicator strain GC4581 carrying VpR /pL -CI857-based expression vectors. Plasmids are pVFGFB (squares) and the parental pCYTEXP1 (triangles). Time 0 indicates the temperature shift to 42³C.
FEMSLE 8924 29-7-99
330
F. Ho¡mann et al. / FEMS Microbiology Letters 177 (1999) 327^334
Fig. 2. Cell lysis after thermal up-shift of recombinant E. coli GC4581 carrying pCYTEXP1. L-Galactosidase activity was determined in the culture samples (solid squares), in the cell-free culture supernatant (solid circles) and in re-suspended cell pellets (solid triangles). The OD of the culture is also indicated (open circles). Time 0 indicates the thermal up-shift.
To con¢rm that the decrease in the OD was caused by cell lysis, L-galactosidase activity was measured in the cell-free medium and in the cells collected by centrifugation(Fig. 2). The OD drop coincided with a decrease of the intracellular L-galactosidase activity and an accompanying increase of the activity in the cell-free culture supernatant. Release of intracellular proteins was also con¢rmed by pulse-labelling and by analyzing the cell proteins and the medium proteins by SDS-PAGE prior to and after the drop of the OD (not shown). This analysis revealed that proteins present in the soluble cell fraction prior to the OD decrease were later found in the cell-free culture supernatant. Furthermore, the amount of proteins belonging to the soluble cell fraction was signi¢cantly reduced in the cell pellet recovered by centrifugation after the drop of the culture OD. To investigate the strain speci¢city of the lytic episode, the growth pattern in response to temperatureinduced heterologous gene expression was also examined using E. coli TG1, BL21, GE864 and MC1061, transformed with the speci¢ed expression vectors (Fig. 3). Although the growth rate of these strains was reduced after induction, revealing the detrimental e¡ects of recombinant gene expression on host cell metabolism, the OD values in these cultures did not decrease upon temperature up-shift with neither bFGF nor FMDV VP1 production. Therefore, it is concluded that cell lysis after temperature-induced gene expression is speci¢c to the re-
Fig. 3. Strain speci¢city of cell lysis. Time-dependent evolution of the OD of either recombinant E. coli GC4581 (solid down triangles) or TG1 (solid up triangles) cultures carrying pVFGFB and recombinant E. coli GC4581 (open down triangles), BL21 (open circle), GE864 (open up triangles) and MC1061 (open squares) cultures carrying pJVP1. Time 0 indicates the thermal up-shift.
porter strain GC4581. In addition, the presence of the vector is required, since the addition of the genotoxic drug nalidixic acid to the plasmid-free GC4581 strain during thermal shock does not promote cell lysis (Fig. 4), proving that SOS induction alone is not a su¤cient requisite for loss of cell integrity.
Fig. 4. Kinetics of sulA : :lacZ expression and cell growth of E. coli GC4581 (squares) and GC4581:pCYTEXP1 (triangles) after simultaneous thermal up-shift and nalidixic acid addition (time 0). Time-dependent evolution of the speci¢c L-galactosidase activity (closed symbols) and the OD (open symbols). As a control, the evolution of the OD of GC4581 not treated with nalidixic acid but exposed to the heat shock is also shown (diamonds).
FEMSLE 8924 29-7-99
F. Ho¡mann et al. / FEMS Microbiology Letters 177 (1999) 327^334
3.2. Expression of prophage lytic genes A main feature of strain GC4581 is the presence of an integrated prophage [16]. Ind3 lambda phages are widely used as vehicles to introduce chromosomal modi¢cations in enterobacteria. They encode a non-hydrolyzable CI repressor to prevent induction of viral gene expression once the prophage genome has been inserted. The results presented here are compatible with the hypothesis of a competition between CI857 and CI Ind3 repressors for the natural pL and pR operator regions in the GC4581 prophage
331
genome. A higher amount of CI857 molecules is expected because of the high gene dosage resulting from the plasmid location of this gene. Therefore, in the plasmid-carrying strains, expression of prophage lytic promoters could be repressed by either CI857 homodimers or CI857/CI Ind3 heterodimers and consequently, the thermal up-shift could lead to transcription of viral genes. This process could become irreversible under our working conditions and might allow for the expression of the lambda lytic genes in a high percentage of the cell population. This is also in agreement with the observation that
Fig. 5. Empty head-like lambda particles observed in GC4581:pVFGFB cultures. Samples were taken 1 h after heat-induced lysis (about 2 h after temperature up-shift). Such particles were not observed 5 min after the temperature up-shift. A depicts a survey, whereas in B and C, the head-like structures (circles) are shown at higher magni¢cations. The measured diameter of the head-like structures is approximately 50 nm in width and 55 nm in height. Bars represent 400 nm in A, 200 nm in B and 50 nm in C.
FEMSLE 8924 29-7-99
332
F. Ho¡mann et al. / FEMS Microbiology Letters 177 (1999) 327^334
Fig. 6. Kinetics of cell growth (A and B) and sulA: :lacZ expression (C and D) in GC4581:pJVP1 (A and C) and LMA1000:pJVP1 (B and D). Cells were growing at 28³C and either kept at this temperature (open circles), kept at this temperature and treated with mitomycin C (open diamonds) or submitted to temperature up-shift (closed circles). Time 0 indicates the thermal up-shift or the addition of mitomycin C.
cell lysis does not occur in GC4581 upon temperature up-shift when this strain is transformed with IPTG-inducible expression vectors not carrying the CI857 repressor (data not shown). To con¢rm that lysis was caused by prophage induction, we looked for the presence of infectious particles after lysis by standard plaque assays. However, no phage plaques were detected by application of supernatants of lysed cultures to indicator plates, neither after incubation at 28, 37 and 42³C, nor when pre-incubating the plates for 5 h at 28³C before transferring them to either 37 or 42³C. In addition, supernatants from the plasmid-free strain GC4581 also failed to produce lambda plaques, despite that the lambda prophage used to construct GC4581 was described as plaque-forming [16,26]. Wild-type lambda lysates, however, produced the expected plaques on MC1061 plates at 37³C, as well as supernatants from a wild-type lambda lysogen (data not shown). The defectiveness of the prophage genome prompted us to investigate the viral dependence of the cell lysis by electron microscopy. The analyses of supernatants from lysed cultures revealed head-like struc-
tures of a size compatible with that of lambda tailless particles (Fig. 5). These particles, which are devoid of nucleic acid, were not detected in the cell-free culture supernatant directly after the thermal shift, but became abundant in samples from extensively lysed cultures. The presence of lambda empty capsids in the cell-free culture con¢rms that the observed lysis is due to heat-inactivation of the CI857 repressor and the consequent transcription of at least some viral lytic genes. Considering these results, it could not be excluded that the previously reported transcription of the sulA: :lacZ fusion in response to temperature up-shift of GC4581 carrying CI857-controlled, lambda-based expression vectors [15] could be enhanced by an extended transcription of lambda lytic genes. The possibility of an unspeci¢c transcription of the reporter gene does not apply to E. coli GE864, in which SOS transcription has also been observed during thermal induction of a CI857-repressed recombinant gene [25]. Having a di¡erent origin and lacking any lambda prophage [17], this strain does not show detectable cell lysis upon thermal up-shift (Fig. 3). To in-
FEMSLE 8924 29-7-99
F. Ho¡mann et al. / FEMS Microbiology Letters 177 (1999) 327^334
vestigate this possibility, we constructed a RecA3 derivative of GC4581 that was further transformed with pJVP1. After thermal induction, this host showed cell lysis like the parental strain but not lacZ transcription (Fig. 6), con¢rming the RecA-dependence of the observed SOS expression. The function of the lacZ gene fusion as a reporter for SOS induction is thus not hampered by a possible unspeci¢c transcription as a part of the viral genome. In conclusion, the cell lysis reported here in the CI857-based cell host system must be taken into account for the interpretation of cellular responses in recombinant bacteria when using indicator strains modi¢ed with lambda gene transfer vehicles.
[7]
[8]
[9]
[10]
[11]
Acknowledgements We are indebted to M. Defais for generously providing strain GE864 and FL8641/pFL352 and to M. Blanco for strain GC4581 and helpful discussions. An `Acciones Integradas' Spanish-German co-operation programme was supporting part of the work: HA96-0028, HA97-0136 (MEC) and 314-A1-e-dr (DAAD). Work in Barcelona was also supported by CICYT (Grant BIO98-0527) and by the Fundacio¨ Maria Francesca de Roviralta, Spain. A.A. and J.L.C. were recipients of predoctoral fellowships from the Ministerio de Educacio¨n y Cultura, Spain.
[12]
[13]
[14]
[15]
References [16] [1] Marston, F.A.O. (1986) The puri¢cation of eukaryotic polypeptides synthesized in Escherichia coli. Biochem. J. 240, 1^12. [2] Bentley, W.E., Mirjalili, N., Andersen, D.C., Davis, R.H. and Kompala, D.S. (1990) Plasmid-encoded protein: the principal factor in the `metabolic burden' associated with recombinant bacteria. Biotechnol. Bioeng. 35, 668^681. [3] Dong, H., Nilsson, L. and Kurland, C.G. (1995) Gratuitous overexpression of genes in Escherichia coli leads to growth inhibition and ribosome destruction. J. Bacteriol. 177, 1497^ 1504. [4] Glick, B.R. (1995) Metabolic load and heterologous gene expression. Biotechnol. Adv. 13, 247^261. [5] Corchero, J.L. and Villaverde, A. (1998) Plasmid maintenance in Escherichia coli recombinant cultures is dramatically, steadily and speci¢cally in£uenced by features of the encoded proteins. Biotechnol. Bioeng. 58, 625^632. [6] Pham, T.T. and Coleman, J.E. (1985) Cloning, expression,
[17]
[18]
[19]
[20]
[21]
333
and puri¢cation of gene 3 endonuclease from bacteriophage T7. Biochemistry 24, 5672^5677. Bedouelle, H., Guez, V., Vidal-Cros, A. and Hermann, M. (1990) Overproduction of tyrosyl-tRNA synthetase is toxic to Escherichia coli : a genetic analysis. J. Bacteriol. 172, 3940^3945. Vidal, M., Cairo¨, J., Mateu, M.G. and Villaverde, A. (1991) Molecular cloning and expression of the VP1 gene of footand-mouth disease virus serotype C1 : E¡ect on bacterial cell viability. Appl. Microbiol. Biotechnol. 35, 788^792. Sisk, W.P., Bradley, J.D., Kingsley, D. and Patterson, T.A. (1992) Deletion of hydrophobic domains of viral glycoproteins increases the levels of their production in Escherichia coli. Gene 112, 157^162. Laity, J.H., Shimotakahara, S. and Scheraga, H. (1993) Expression of wild-type and mutant bovine pancreatic ribonuclease A in Escherichia coli. Proc. Natl. Acad. Sci. USA 90, 615^619. Sheu, S.Y. and Lo, S.J. (1995) Deletion or alteration of hydrophobic amino acids at the ¢rst and the third transmembrane domains of hepatitis B surface antigen enhances its production in Escherichia coli. Gene 160, 179^184. Viaplana, E. and Villaverde, A. (1996) Polylinker-encoded peptides can confer toxicity to recombinant proteins produced in Escherichia coli. Biotechnol. Prog. 12, 723^727. Yike, I., Zhang, Y., Ye, J. and Dearbon, D.G. (1996) Expression in Escherichia coli of cytoplasmatic portions of the cystic ¢brosis transmembrane conductance regulator; apparent bacterial toxicity of peptides containing R-domain sequences. Protein Expr. Purif. 7, 45^50. Viaplana, E., Rebordosa, X., Pin¬ol, J. and Villaverde, A. (1997) Secretion-dependent proteolysis of recombinant proteins is associated with inhibition of cells growth in Escherichia coli. Biotechnol. Lett. 19, 373^377. Ar|¨s, A., Corchero, J.L., Benito, A., Carbonell, X., Viaplana, E. and Villaverde, A. (1998) The expression of recombinant genes from bacteriophage lambda strong promoters triggers the SOS response in Escherichia coli. Biotechnol. Bioeng. 60, 651^659. Sommer, S., Bailone, A. and Devoret, R. (1985) SOS induction by thermosensitive replication mutants of miniF plasmid. Mol. Gen. Genet. 198, 456^464. Larminat, F. and Defais, M. (1989) Modulation of the SOS response by truncated RecA proteins. Mol. Gen. Genet. 216, 106^112. Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989) Molecular cloning. A laboratory manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Studier, F.W. and Mo¡at, B.A. (1986) Use of the bacteriophage T7 RNA polymerase to direct selective high-level expression of cloned genes. J. Mol. Biol. 189, 113^130. Carter, P., Bedouelle, H. and Winter, G. (1985) Improved oligonucleotide site-directed mutagenesis using M13 vectors. Nucleic Acids Res. 13, 4431^4442. Seeger, A., Schneppe, B., McCarthy, J.E., Deckwer, W.-D. and Rinas, U. (1995) Comparison of temperature- and isopropyl-L-D-thiogalacto-pyranoside-induced synthesis of basic
FEMSLE 8924 29-7-99
334
F. Ho¡mann et al. / FEMS Microbiology Letters 177 (1999) 327^334
¢broblast growth factor in high-cell-density cultures of recombinant Escherichia coli. Enzyme Microb. Technol. 17, 947^953. [22] Schauder, B., Blo«cker, H., Frank, R. and McCarthy, J.E.G. (1987) Inducible vectors incorporating the Escherichia coli atpE translational initiation region. Gene 52, 279^283. [23] Laemmli, U.K. (1979) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 27, 680^685. [24] Valentine, R.C., Shapiro, B.M. and Stadtman, E.R. (1968)
Regulation of glutamate synthetase XII. Electron microscopy of the enzyme from Escherichia coli. Biochemistry 7, 2143^ 2152. [25] Benito, A., Viaplana, E., Corchero, J.L., Carbonell, X. and Villaverde, A. (1995) A recombinant foot-and-mouth disease virus antigen inhibits DNA replication and triggers the SOS response in E. coli. FEMS Microbiol. Lett. 129, 157^162. [26] Huisman, O. and D'Ari, R. (1983) E¡ect of suppressors of SOS-mediated ¢lamentation on s¢A operon expression in Escherichia coli. J. Bacteriol. 153, 169^175.
FEMSLE 8924 29-7-99
Heat-inactivation of plasmid-encoded CI857 repressor induces gene expression from Ind3 lambda prophage in recombinant Escherichia coli Frank Ho¡mann a , Anna Ar|¨s b , Xavier Carbonell b , Manfred Rohde c , Jose¨ L. Corchero b;1 , Ursula Rinas a , Antonio Villaverde b; * a
Biochemical Engineering, GBF National Research Center for Biotechnology, Mascheroder Weg 1, 38124 Braunschweig, Germany Institut de Biologia Fonamental and Departament de Gene©tica i Microbiologia, Universitat Auto©noma de Barcelona, Bellaterra, 08193 Barcelona, Spain c Microbiology Division, GBF National Research Center for Biotechnology, Mascheroder Weg 1, 38124 Braunschweig, Germany
b
Received 27 January 1999; received in revised form 30 April 1999; accepted 22 June 1999
Abstract We have observed significant cell lysis upon temperature up-shift of recombinant Escherichia coli cultures harboring CI857repressed lambda-based expression vectors. This event, that becomes evident about 30^40 min after the heat shock, takes place when using the lambda promoter system in Ind3 lysogenic strains, but not in others commonly employed for recombinant gene expression. These results strongly suggest that the thermosensitive CI857 repressor, encoded by the expression vector, competes with CI Ind3 molecules for binding to the prophage operator region, allowing for expression of lytic genes from the integrated Ind3 viral genome upon temperature up-shift. Transcription of viral lytic genes does not include unspecific expression of a reporter sulA: :lacZ gene fusion carried in the prophage genome. These results prompt, however, to carefully evaluate the limitations of expression systems based on pL /pR -CI857 in bacterial strains modified through lambda Ind3 gene transfer vehicles. ß 1999 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. Keywords : Lambda bacteriophage; CI857 repressor ; Recombinant protein; Structural gene ; Cell lysis
1. Introduction The overproduction of heterologous proteins in Escherichia coli causes a severe metabolic burden
* Corresponding author. Tel.: +34 (93) 5812148; Fax: +34 (93) 5812011; E-mail: [email protected] 1 Present address: VEHB, MailStop G 18, Center for Disease Control, 1600 Clifton Road, Atlanta, GA 30333, USA.
[1^5], which is often accompanied by direct toxicity of the products on the producing cells. Although still unidenti¢ed, there are several indications suggesting a multifactorial nature of the deleterious e¡ects associated to recombinant protein synthesis [6^14]. In this line, we have recently reported induction of SOS genes in E. coli indicator strains in response to thermal induction of plasmid-encoded recombinant genes contained in CI857-controlled, lambda-based expression vectors [15]. Since the nature of this in-
0378-1097 / 99 / $20.00 ß 1999 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 1 0 9 7 ( 9 9 ) 0 0 3 3 4 - 1
FEMSLE 8924 29-7-99
328
F. Ho¡mann et al. / FEMS Microbiology Letters 177 (1999) 327^334
triguing event remained still unsolved, we were prompted to investigate in more detail the possible mechanisms linking thermal induction of recombinant gene expression and SOS transcription. By analyzing the growth pattern of an indicator strain carrying a sulA: :lacZ gene fusion, we evidenced a partial but signi¢cant lysis of the culture about 30^40 min after the temperature up-shift, accompanied by the release of incomplete lambda virus particles to the medium. This observation revealed expression of lytic genes from the Ind3 prophage genome, used as a vehicle for the reporter sulA: :lacZ gene fusion, that is probably caused by an interference of the thermosensitive CI857 repressor with the natural activities of the CI Ind3 repressor encoded by the lambda prophage DNA, its inactivation when the bacterial culture is submitted to a thermal up-shift and the consequent derepression of lytic genes. Therefore, these results show the necessity to carefully evaluate the limitations of expression systems based on pL /pR -CI857 in bacterial strains modi¢ed through lambda Ind3 gene transfer vehicles.
2. Materials and methods 2.1. Bacterial strains, plasmids and proteins The E. coli strain GC4581, F3 vlac-pro (V cIind1 (sulA: :lacZ) RecA ) [16], was used to monitor expression of the SOS sulA (formerly s¢A) gene through L-galactosidase activity during plasmid-encoded recombinant gene expression. A vrecA derivative of GC4581 (named LMA1000) has been constructed by P1 transduction from the donor strain FL8641/pFL352, (F3 ) araD139 v(argF-lac)169 £hD5301 fruA25 relA1 rpsL150(Strr ) rbsR22 deoC1 v(gal-VG) Vb2: : x((cea pup-lacZ)2(Hyb) lacY ) v(recA-srlR)306 srlR301: :Tn10-84/pFL352(bla recA ) [17], followed by clone selection by tetracycline resistance and further testing of the RecA3 phenotype by determining ultraviolet radiation sensitivity. Strains MC1061, hsdR2 mcrB araD139 v(araABCleu)7679 vlacX74 galU galK rpsL thi strA [18], BL21, hsdS gal3 Lon3 OmpT3 [19], GE864 MC4100, v(gal-GV ) b2V : :((cea Pup-lacZ)Z (Hyb)
lacY ) [17], and TG1, supE hsdv5 thi v(lac-proAB) F'(traD36 proAB lacIq lacZvM15) [20], were also employed for recombinant gene expression. The construction of the expression vector pVFGFB, a derivative of pCYTEXP1, has been described before [21]. It harbors the gene for human basic ¢broblast growth factor (bFGF). Plasmid pJVP1 derives from pJLA602 [22] and it encodes a complete VP1 protein from foot and mouth disease virus (FMDV) [5]. All these multi-copy plasmids harbor an ampicillin resistance gene and the CI857 repressor gene under the control of its own promoter. Expression of recombinant genes in the described expression vectors is controlled by the lambda pR and pL promoters placed in tandem. 2.2. Culture conditions and recombinant gene expression Luria-Bertani (LB) medium [18] with 40 Wg ml31 streptomycin and 100 Wg ml31 ampicillin was used for culture growth. Aliquots from overnight cultures were used to inoculate fresh medium at 1:50, in 100ml shaker £asks, with a working volume of 20 ml. These cultures were incubated at 250 rpm and 28³C. At an optical density (OD550 ) of 0.2^0.5 U, recombinant gene expression was induced by transfer to a pre-warmed bath at 42³C and further incubation at the same agitation rate. When indicated, the potent SOS inducers nalidixic acid or mitomycin C were added simultaneously to the temperature up-shift at 40 or 1.6 Wg ml31 , respectively. 2.3. Monitoring of culture growth, cell integrity and lacZ gene expression Cell growth and lysis were monitored by regular measurements of OD550 . L-Galactosidase activity was determined as described previously [15]. The enzymatic activity was also measured in the cell-free culture supernatant and the cell pellet obtained after centrifugation at 14000 rpm for 3 min. Cell pellets were re-suspended in the original volume of LB medium. Cell permeabilization was achieved by vortexing in chloroform-containing Z bu¡er (0.06 M Na2 HPO4 , 0.04 M NaH2 PO4 , pH 7.2, 0.01 M KCl, 1 mM MgSO4 , 0.04 M L-mercaptoethanol).
FEMSLE 8924 29-7-99
F. Ho¡mann et al. / FEMS Microbiology Letters 177 (1999) 327^334
2.4. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) For SDS-PAGE analysis, cell pellets were re-suspended in 50 mM sodium phosphate bu¡er (pH 7) to an OD of 4.5. After sonication for 2 min, soluble and insoluble fractions were collected by centrifugation for 45 min at 38 000Ug and 4³C. The insoluble fraction was washed with sodium phosphate bu¡er, centrifuged again and re-suspended in the same bu¡er in 1/4 of the original volume. SDS-PAGE analysis was performed according to the method of Laemmli [23]. Samples were mixed with loading bu¡er (1:1) consisting of 7% (w/v) SDS, 35% (w/v) glycerol, 150 mM dithiothreitol, 250 mM Tris-HCl (pH 6.8), boiled for 10 min and ran immediately on 9^16% polyacrylamide gradient gels. 2.5. Procedures for bacteriophage detection Samples from lysed cultures were plated on con£uent cultures of MC1061 and MC1061 carrying pJLA602 and incubated at di¡erent temperatures from 28 to 42³C to detect phage plaques. Plasmid pJLA602 was added to provide CI857 repressor to reproduce as much as possible the situation observed in cell culture. Furthermore, aliquots of these samples were also negatively stained and prepared for visualization by transmission electron microscopy following the procedure described by Valentine and co-authors [24]. Brie£y, thin carbon ¢lms were prepared by sublimation of carbon onto freshly cleaved mica. Samples were then adsorbed onto small pieces of the carbon ¢lm for 30 s and negatively stained with 4% uranyl acetate dissolved in water (pH 5.0). After air-drying, samples were examined in a Zeiss transmission electron microscope EM910 at 80 kV and at calibrated magni¢cations using a cross-lined grating replica.
329
pL tandem promoter and its thermosensitive CI857 repressor. The use of the reporter strain GC4581 allows for monitoring of the SOS sulA gene transcription. The time-dependent evolution of the reporter L-galactosidase activity in the induced recombinant culture (Fig. 1A) was comparable to previously published results on the induction of the SOS gene upon temperature up-shift in temperatureinducible expression systems [15,25]. However, sulA gene expression is also observed in the strain harboring the parental vector pCYTEXP1 without any cloned gene, suggesting that the biosynthesis of a complete protein is not required for the activation of the sulA transcription. On the other hand, frequent measurements of the OD in these cultures indicated a lytic episode in both strains (Fig. 1B) that was not noticed in previous experiments with less frequent OD measurements [15]. The drop in the OD started around 30^40 min after the temperature shift and seemed to involve the lysis of a major part of the cell population. This OD decrease was reproducible and time-coincident when comparing di¡erent independent experiments.
3. Results and discussion 3.1. Cell lysis during induction of recombinant gene expression Protein bFGF was produced in E. coli as directed from an expression vector based on the lambda pR /
Fig. 1. Kinetics of sulA : :lacZ expression (A) and cell growth (B) after thermal up-shift of the indicator strain GC4581 carrying VpR /pL -CI857-based expression vectors. Plasmids are pVFGFB (squares) and the parental pCYTEXP1 (triangles). Time 0 indicates the temperature shift to 42³C.
FEMSLE 8924 29-7-99
330
F. Ho¡mann et al. / FEMS Microbiology Letters 177 (1999) 327^334
Fig. 2. Cell lysis after thermal up-shift of recombinant E. coli GC4581 carrying pCYTEXP1. L-Galactosidase activity was determined in the culture samples (solid squares), in the cell-free culture supernatant (solid circles) and in re-suspended cell pellets (solid triangles). The OD of the culture is also indicated (open circles). Time 0 indicates the thermal up-shift.
To con¢rm that the decrease in the OD was caused by cell lysis, L-galactosidase activity was measured in the cell-free medium and in the cells collected by centrifugation(Fig. 2). The OD drop coincided with a decrease of the intracellular L-galactosidase activity and an accompanying increase of the activity in the cell-free culture supernatant. Release of intracellular proteins was also con¢rmed by pulse-labelling and by analyzing the cell proteins and the medium proteins by SDS-PAGE prior to and after the drop of the OD (not shown). This analysis revealed that proteins present in the soluble cell fraction prior to the OD decrease were later found in the cell-free culture supernatant. Furthermore, the amount of proteins belonging to the soluble cell fraction was signi¢cantly reduced in the cell pellet recovered by centrifugation after the drop of the culture OD. To investigate the strain speci¢city of the lytic episode, the growth pattern in response to temperatureinduced heterologous gene expression was also examined using E. coli TG1, BL21, GE864 and MC1061, transformed with the speci¢ed expression vectors (Fig. 3). Although the growth rate of these strains was reduced after induction, revealing the detrimental e¡ects of recombinant gene expression on host cell metabolism, the OD values in these cultures did not decrease upon temperature up-shift with neither bFGF nor FMDV VP1 production. Therefore, it is concluded that cell lysis after temperature-induced gene expression is speci¢c to the re-
Fig. 3. Strain speci¢city of cell lysis. Time-dependent evolution of the OD of either recombinant E. coli GC4581 (solid down triangles) or TG1 (solid up triangles) cultures carrying pVFGFB and recombinant E. coli GC4581 (open down triangles), BL21 (open circle), GE864 (open up triangles) and MC1061 (open squares) cultures carrying pJVP1. Time 0 indicates the thermal up-shift.
porter strain GC4581. In addition, the presence of the vector is required, since the addition of the genotoxic drug nalidixic acid to the plasmid-free GC4581 strain during thermal shock does not promote cell lysis (Fig. 4), proving that SOS induction alone is not a su¤cient requisite for loss of cell integrity.
Fig. 4. Kinetics of sulA : :lacZ expression and cell growth of E. coli GC4581 (squares) and GC4581:pCYTEXP1 (triangles) after simultaneous thermal up-shift and nalidixic acid addition (time 0). Time-dependent evolution of the speci¢c L-galactosidase activity (closed symbols) and the OD (open symbols). As a control, the evolution of the OD of GC4581 not treated with nalidixic acid but exposed to the heat shock is also shown (diamonds).
FEMSLE 8924 29-7-99
F. Ho¡mann et al. / FEMS Microbiology Letters 177 (1999) 327^334
3.2. Expression of prophage lytic genes A main feature of strain GC4581 is the presence of an integrated prophage [16]. Ind3 lambda phages are widely used as vehicles to introduce chromosomal modi¢cations in enterobacteria. They encode a non-hydrolyzable CI repressor to prevent induction of viral gene expression once the prophage genome has been inserted. The results presented here are compatible with the hypothesis of a competition between CI857 and CI Ind3 repressors for the natural pL and pR operator regions in the GC4581 prophage
331
genome. A higher amount of CI857 molecules is expected because of the high gene dosage resulting from the plasmid location of this gene. Therefore, in the plasmid-carrying strains, expression of prophage lytic promoters could be repressed by either CI857 homodimers or CI857/CI Ind3 heterodimers and consequently, the thermal up-shift could lead to transcription of viral genes. This process could become irreversible under our working conditions and might allow for the expression of the lambda lytic genes in a high percentage of the cell population. This is also in agreement with the observation that
Fig. 5. Empty head-like lambda particles observed in GC4581:pVFGFB cultures. Samples were taken 1 h after heat-induced lysis (about 2 h after temperature up-shift). Such particles were not observed 5 min after the temperature up-shift. A depicts a survey, whereas in B and C, the head-like structures (circles) are shown at higher magni¢cations. The measured diameter of the head-like structures is approximately 50 nm in width and 55 nm in height. Bars represent 400 nm in A, 200 nm in B and 50 nm in C.
FEMSLE 8924 29-7-99
332
F. Ho¡mann et al. / FEMS Microbiology Letters 177 (1999) 327^334
Fig. 6. Kinetics of cell growth (A and B) and sulA: :lacZ expression (C and D) in GC4581:pJVP1 (A and C) and LMA1000:pJVP1 (B and D). Cells were growing at 28³C and either kept at this temperature (open circles), kept at this temperature and treated with mitomycin C (open diamonds) or submitted to temperature up-shift (closed circles). Time 0 indicates the thermal up-shift or the addition of mitomycin C.
cell lysis does not occur in GC4581 upon temperature up-shift when this strain is transformed with IPTG-inducible expression vectors not carrying the CI857 repressor (data not shown). To con¢rm that lysis was caused by prophage induction, we looked for the presence of infectious particles after lysis by standard plaque assays. However, no phage plaques were detected by application of supernatants of lysed cultures to indicator plates, neither after incubation at 28, 37 and 42³C, nor when pre-incubating the plates for 5 h at 28³C before transferring them to either 37 or 42³C. In addition, supernatants from the plasmid-free strain GC4581 also failed to produce lambda plaques, despite that the lambda prophage used to construct GC4581 was described as plaque-forming [16,26]. Wild-type lambda lysates, however, produced the expected plaques on MC1061 plates at 37³C, as well as supernatants from a wild-type lambda lysogen (data not shown). The defectiveness of the prophage genome prompted us to investigate the viral dependence of the cell lysis by electron microscopy. The analyses of supernatants from lysed cultures revealed head-like struc-
tures of a size compatible with that of lambda tailless particles (Fig. 5). These particles, which are devoid of nucleic acid, were not detected in the cell-free culture supernatant directly after the thermal shift, but became abundant in samples from extensively lysed cultures. The presence of lambda empty capsids in the cell-free culture con¢rms that the observed lysis is due to heat-inactivation of the CI857 repressor and the consequent transcription of at least some viral lytic genes. Considering these results, it could not be excluded that the previously reported transcription of the sulA: :lacZ fusion in response to temperature up-shift of GC4581 carrying CI857-controlled, lambda-based expression vectors [15] could be enhanced by an extended transcription of lambda lytic genes. The possibility of an unspeci¢c transcription of the reporter gene does not apply to E. coli GE864, in which SOS transcription has also been observed during thermal induction of a CI857-repressed recombinant gene [25]. Having a di¡erent origin and lacking any lambda prophage [17], this strain does not show detectable cell lysis upon thermal up-shift (Fig. 3). To in-
FEMSLE 8924 29-7-99
F. Ho¡mann et al. / FEMS Microbiology Letters 177 (1999) 327^334
vestigate this possibility, we constructed a RecA3 derivative of GC4581 that was further transformed with pJVP1. After thermal induction, this host showed cell lysis like the parental strain but not lacZ transcription (Fig. 6), con¢rming the RecA-dependence of the observed SOS expression. The function of the lacZ gene fusion as a reporter for SOS induction is thus not hampered by a possible unspeci¢c transcription as a part of the viral genome. In conclusion, the cell lysis reported here in the CI857-based cell host system must be taken into account for the interpretation of cellular responses in recombinant bacteria when using indicator strains modi¢ed with lambda gene transfer vehicles.
[7]
[8]
[9]
[10]
[11]
Acknowledgements We are indebted to M. Defais for generously providing strain GE864 and FL8641/pFL352 and to M. Blanco for strain GC4581 and helpful discussions. An `Acciones Integradas' Spanish-German co-operation programme was supporting part of the work: HA96-0028, HA97-0136 (MEC) and 314-A1-e-dr (DAAD). Work in Barcelona was also supported by CICYT (Grant BIO98-0527) and by the Fundacio¨ Maria Francesca de Roviralta, Spain. A.A. and J.L.C. were recipients of predoctoral fellowships from the Ministerio de Educacio¨n y Cultura, Spain.
[12]
[13]
[14]
[15]
References [16] [1] Marston, F.A.O. (1986) The puri¢cation of eukaryotic polypeptides synthesized in Escherichia coli. Biochem. J. 240, 1^12. [2] Bentley, W.E., Mirjalili, N., Andersen, D.C., Davis, R.H. and Kompala, D.S. (1990) Plasmid-encoded protein: the principal factor in the `metabolic burden' associated with recombinant bacteria. Biotechnol. Bioeng. 35, 668^681. [3] Dong, H., Nilsson, L. and Kurland, C.G. (1995) Gratuitous overexpression of genes in Escherichia coli leads to growth inhibition and ribosome destruction. J. Bacteriol. 177, 1497^ 1504. [4] Glick, B.R. (1995) Metabolic load and heterologous gene expression. Biotechnol. Adv. 13, 247^261. [5] Corchero, J.L. and Villaverde, A. (1998) Plasmid maintenance in Escherichia coli recombinant cultures is dramatically, steadily and speci¢cally in£uenced by features of the encoded proteins. Biotechnol. Bioeng. 58, 625^632. [6] Pham, T.T. and Coleman, J.E. (1985) Cloning, expression,
[17]
[18]
[19]
[20]
[21]
333
and puri¢cation of gene 3 endonuclease from bacteriophage T7. Biochemistry 24, 5672^5677. Bedouelle, H., Guez, V., Vidal-Cros, A. and Hermann, M. (1990) Overproduction of tyrosyl-tRNA synthetase is toxic to Escherichia coli : a genetic analysis. J. Bacteriol. 172, 3940^3945. Vidal, M., Cairo¨, J., Mateu, M.G. and Villaverde, A. (1991) Molecular cloning and expression of the VP1 gene of footand-mouth disease virus serotype C1 : E¡ect on bacterial cell viability. Appl. Microbiol. Biotechnol. 35, 788^792. Sisk, W.P., Bradley, J.D., Kingsley, D. and Patterson, T.A. (1992) Deletion of hydrophobic domains of viral glycoproteins increases the levels of their production in Escherichia coli. Gene 112, 157^162. Laity, J.H., Shimotakahara, S. and Scheraga, H. (1993) Expression of wild-type and mutant bovine pancreatic ribonuclease A in Escherichia coli. Proc. Natl. Acad. Sci. USA 90, 615^619. Sheu, S.Y. and Lo, S.J. (1995) Deletion or alteration of hydrophobic amino acids at the ¢rst and the third transmembrane domains of hepatitis B surface antigen enhances its production in Escherichia coli. Gene 160, 179^184. Viaplana, E. and Villaverde, A. (1996) Polylinker-encoded peptides can confer toxicity to recombinant proteins produced in Escherichia coli. Biotechnol. Prog. 12, 723^727. Yike, I., Zhang, Y., Ye, J. and Dearbon, D.G. (1996) Expression in Escherichia coli of cytoplasmatic portions of the cystic ¢brosis transmembrane conductance regulator; apparent bacterial toxicity of peptides containing R-domain sequences. Protein Expr. Purif. 7, 45^50. Viaplana, E., Rebordosa, X., Pin¬ol, J. and Villaverde, A. (1997) Secretion-dependent proteolysis of recombinant proteins is associated with inhibition of cells growth in Escherichia coli. Biotechnol. Lett. 19, 373^377. Ar|¨s, A., Corchero, J.L., Benito, A., Carbonell, X., Viaplana, E. and Villaverde, A. (1998) The expression of recombinant genes from bacteriophage lambda strong promoters triggers the SOS response in Escherichia coli. Biotechnol. Bioeng. 60, 651^659. Sommer, S., Bailone, A. and Devoret, R. (1985) SOS induction by thermosensitive replication mutants of miniF plasmid. Mol. Gen. Genet. 198, 456^464. Larminat, F. and Defais, M. (1989) Modulation of the SOS response by truncated RecA proteins. Mol. Gen. Genet. 216, 106^112. Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989) Molecular cloning. A laboratory manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Studier, F.W. and Mo¡at, B.A. (1986) Use of the bacteriophage T7 RNA polymerase to direct selective high-level expression of cloned genes. J. Mol. Biol. 189, 113^130. Carter, P., Bedouelle, H. and Winter, G. (1985) Improved oligonucleotide site-directed mutagenesis using M13 vectors. Nucleic Acids Res. 13, 4431^4442. Seeger, A., Schneppe, B., McCarthy, J.E., Deckwer, W.-D. and Rinas, U. (1995) Comparison of temperature- and isopropyl-L-D-thiogalacto-pyranoside-induced synthesis of basic
FEMSLE 8924 29-7-99
334
F. Ho¡mann et al. / FEMS Microbiology Letters 177 (1999) 327^334
¢broblast growth factor in high-cell-density cultures of recombinant Escherichia coli. Enzyme Microb. Technol. 17, 947^953. [22] Schauder, B., Blo«cker, H., Frank, R. and McCarthy, J.E.G. (1987) Inducible vectors incorporating the Escherichia coli atpE translational initiation region. Gene 52, 279^283. [23] Laemmli, U.K. (1979) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 27, 680^685. [24] Valentine, R.C., Shapiro, B.M. and Stadtman, E.R. (1968)
Regulation of glutamate synthetase XII. Electron microscopy of the enzyme from Escherichia coli. Biochemistry 7, 2143^ 2152. [25] Benito, A., Viaplana, E., Corchero, J.L., Carbonell, X. and Villaverde, A. (1995) A recombinant foot-and-mouth disease virus antigen inhibits DNA replication and triggers the SOS response in E. coli. FEMS Microbiol. Lett. 129, 157^162. [26] Huisman, O. and D'Ari, R. (1983) E¡ect of suppressors of SOS-mediated ¢lamentation on s¢A operon expression in Escherichia coli. J. Bacteriol. 153, 169^175.
FEMSLE 8924 29-7-99