Gene, 127 (1993) 99-103

0 1993 Elsevier Science Publishers B.V. All rights reserved. 0378-l 119/93/$06.00

GENE 06994

99

A new cloning vector and expression strategy for genes encoding proteins toxic to Escherichia coli

(Saccharomyces cereuisiae; POL3 gene; DNA polymerase 6; bacteriophage T7)

William Clay Brown and Judith L. Campbell

Braun Laboratories, Division of Chemistry, California Institute of Technology, Pasadena, CA 91125, USA

Received by G. Wilcox: 20 April 1992; Revised/Accepted: 21 September/4 November 1992; Received at publishers: 14 December 1992

SUMMARY

Here, we describe a modification of a plasmid, pT7-7 [Tabor and Richardson, Proc. Natl. Acad. Sci. USA 262 (1985) 1074-10781, that allows expression of inserted genes from the phage T7 RNA polymerase promoter. The modification is designed to suppress readthrough transcription from cryptic promoters and start points on the plasmid, in order to reduce expression in the absence of T7 RNA polymerase and thus improve the vector for use in the expression of highly toxic gene products. This vector (pT7SC) was used to stably clone the POL3 gene (encoding DNA polymerase 6) of Saccharomyces cereuisiae, which destabilizes all other cloning and expression vectors tested. Previously described expres- sion strategies proved ineffective in overexpressing the POL3 gene. A new strategy was developed which relies on induction by infection with mutant T7 phage. This system efficiently overproduced the POL3 gene product.

INTRODUCTION

Since the cloning of DNA has become a fairly routine procedure, the use of bacteria as a means of expressing the cloned genes has become common. Very few failures were encountered when the genes used were from Escher- ichia coli or other prokaryotic organisms, but as more and more eukaryotic genes have been cloned and charac- terized a problem has arisen in their expression in bacte- ria. Often the gene product is toxic, as evidenced initially by instability of the plasmid containing the gene. As a means of countering this problem, expression vectors

Correspondence to: Dr. J.L. Campbell, Divisions of Biology and Chem-

istry, 147-75 Caltech, Pasadena, CA 91125, USA. Tel. (818)356-6053;

Fax (818) 449-0756.

Abbreviations: A, absorbance (1 cm); Ap, ampicillin; bla, gene encoding

B-lactamase (Bla); bp, base pair(s); A, deletion; HSV, herpes simplex

virus; IPTG, isopropyl-B-o-thiogalactopyranoside; kb, kilobase or

1000 bp; moi, multiplicity of infection; nt, nucleotide(s); ORF, open

reading frame; PA, polyacrylamide; POW, gene encoding yeast DNA

polymerase 6; PolIk, Klenow (large) fragment of E. co/i DNA polymer-

ase I; S., Saccharomyces; SDS, sodium dodecyl sulfate.

were developed that rely on a bacteriophage T7 pro- moter, which should not be recognized by the RNA poly- merases of E. co/i, to ensure that the expression of the gene is tightly regulated (Tabor and Richardson, 1985; Studier and Moffat, 1986). The T7 RNA polymerase may be delivered to the system in a number of ways. A copy of the T7 RNA polymerase gene may be carried on another plasmid under the control of a thermolabile phage h repressor and induced by temporarily raising the temperature (Tabor and Richardson, 1985). Because repression is not complete, however, genes are often unstable even in these systems. Another method uses a h lysogen cell line in which the T7 RNA polymerase gene is under the control of the IPTG-inducible lac promoter. To make the control of RNA polymerase gene expression more stringent a plasmid containing the gene for lyso- zyme is present (Studier and Moffat, 1986). T7 lysozyme specifically cleaves the residual RNA polymerase that is produced prior to induction. When IPTG is added, the cell is induced to produce much higher levels of T7 RNA polymerase which rapidly outstrips the lysozyme activity and goes on to transcribe the gene of interest. The final

100

method routinely used consists of infecting the cells con- taining the gene of interest with an Ml3 phage that car- ries the T7 RNA polymerase gene, also under EGC control, and inducing with IPTG. These latter systems have not been universally effective in allowing for overexpression of eukaryotic genes in bacteria. Often the yield of protein is quite low, if the protein is produced at all.

Three nuclear RNA polymerases - a, 6 and a have been characte~zed and shown to be encoded by essential genes in Saccharomyces cerevisiae. Previous work has shown that DNA polymerase 6 was encoded by gene CffC2 (Sitney et al., 1989; Boulet et al., 1989), which has since been renamed POL3. We and others have observed that POL3-expressing plasmids are unstable in E. coEi (Simon et al., 1991; P. Burgers, per. comm.) Observations made while handling the gene suggested that uncon- trolled expression, even at low levels, is responsible for the instability, thus DNA polymerase 6 must be toxic to E. co&. The goal of the present study is the development of a vector and expression strategy that allows stable cloning and expression of genes such as PUL3 that are toxic to E. coii,

EXPERI~E~AL AND DISCUSSION

(a) Construction of a vector which suppresses read- through tran~ription

To stabilize the POL3 gene of S. cereu&siae, a new vector was developed, based on those of Tabor and Richardson (1985). The bia promoter and the Rho-independent tran- scription terminators 7; and TZ from the rrnB operon of E. co& (Brosius, 1984a) were cloned into pT7-7 (Tabor and Richardson, 1985) upstream from the T7 promoter (Fig. 1). The promoter was directed away from the clon- ing region and was followed downstream by T, and TZ which occur in the inverted orientation from that found in vivo but are still functional, though with decreased efficiency (Brosius, 1984a). Some genes are so unstable in bacterial vectors that there is difficulty in obtaining DNA appropriate for further characterization, especially sequence analysis. In order to allow for the cloning of toxic genes with unknown sequence and orientation, the extra promoter and terminator elements were also inserted into the last restriction site of the pT7-7 poly- linker, again such that the promoter is directed away from the cloning region. The completed plasmid is shown in Fig. 1 and has been named pT7SC (for stringent control).

(b) Cloning the POL 3 gene A map of the POL3 gene of S. cerevisiae, which encodes

the catalytic subunit of DNA polymerase 6 is shown in Fig. 2A. This gene has a 3279-bp ORF, which is predicted

pT7 SC 4121 bp

Fig. 1. Construction of pT7SC. Methods: Plasmid pKK223-3 (Phar- macia) was digested with Hind111 + SC&. The 82%bp fragment contain- ing the bla gene upstream sequences and the rsnE terminators was purified from a low-melting-agarose gel by phenol extraction (Perbal, 1984). The HindtII overhang was filled-in with PolIk to yield a bhmt- ended molecule. Piasmid pT7-7 was cut with EgJIX and the overhangs were filled in. The 828-bp fragment was then blunt-end ligated into pT7-7,Or~entation was determined by cutting recombinants with &WI. Plasmid containing the properly oriented fra~ent was propagated and then digested with CJaI. The overhangs were filled in and the 828-bp fragment was inserted by blunt-end ligation. Orientation was again checked by digestion with LtraI. The DNA replication origin fori) was from ColEI.

to give rise to a 124-kDa protein (Boulet et al., 1989). Constructs of POL3 beginning at either the A4iuI site or the N&I site and running to the 3’ HindIII site were unstable in expression vectors pT7-7, pET3C (Studier and Moffat, 1986) and pKK223-3 (Amman et al., 1983; Brosius, 1984b). Constructs truncated to the &rr~I site were stable in pT7-7 but protein was not observed upon induction. This restriction site occurs down-stream from the putative exonuciease sequences in region IV (Wong et al., 1989; Simon et al., 1991) and this may have a bear- ing on both the stability of this construct and the lack of expressed protein. The entire POL3 gene, however, was successfully and stably cloned into pT7SC. A diagram of this construct is shown in Fig. 2B. Inappropriate expres- sion is apparently suppressed by the presence of the pro- moters and terminator sequences, and the POL3-carrying plasmid is stable in a variety of commonly used cell lines. Another gene which had previously proven difficult to work with due to toxicity, yeast ~SP6~, was also stabi- lized in this vector (Smiley et al., 1992), suggesting that this vector may be generally useful,

(c) Expression of POL3 by infection with mutant T7 bacteriophage

The pT7SC/POL3 construct relies on a T7 promoter for the expression of the POW gene, and therefore a cell

101

_ Start: ATG I” II “I 111 I ” Stop: TAA

Consmed Regions

pT7 SCIPOL3

Fig. 2. Linear map of S. cereuisiae POU gene (A), cloned in the pT7SC plasmid (B). (A) ORF is shown as an open box, non-coding sequences are medium thickness and shaded. The blackened boxes in the ORF correspond to the conserved regions associated with u-like DNA poly- merases (Wong et al., 1988). (B) The POW gene was obtained from a YCp50 construct (between the Sal1 and Hind111 sites) in which the upstream MluI site of POL.? was inserted at the vector Sal1 site such that each was regenerated (pBL304). Plasmid pBL304 was digested with Sal1 + Hind111 and ligated into these same sites in pT7SC. The resultant recombinant was then cut with XbaI+MluI. The ends were filled in and the plasmid was re-circularized. This construct (B) was used for cloned gene expression.

line, BL21 (Studier and Moffat, 1986), that contains a chromosomal copy of the T7 RNA polymerase gene was used for expressing POL3. This cell line carries the POL3 gene-containing plasmid without alteration, but the DNA polymerase 6 is not detectable upon induction with IPTG (Fig. 3). A second approach was to use a protease- deficient cell line and to provide the T7 RNA polymerase by infecting with mGPl-2 which is an Ml3 strain that contains the gene for T7 RNA polymerase under lac con- trol. Again DNA polymerase 6 was not detected upon infection and induction with IPTG (Fig. 3).

The protein encoded by gene 4 of bacteriophage T7 is toxic and very difficult to clone (S. Tabor, per. comm.) but is stable in E. coli during infection by T7 phage. Possible phage T7 activities involved in suppression of cellular functions include that of gene 2 protein, which binds to and inhibits E. coli RNA polymerase, and of

Source of ‘I7 RNA Pal

(- : uniduced, + : induced)

C124kDa

Fig. 3. Expression of POI.3 is dependent on infection by mutant T7 bacteriophage. Fresh overnight cultures are diluted lOO-fold in LB with 20 ug Ap/ml and incubated at 30°C to an A,,, = 1. Induction is through either the addition of IPTG to 0.4 mM, infection with mGPl-2 at an moi of 5 with addition of IPTG or by infection with bacteriophage T7356- at an moi of 50. Incubations are then continued at 30 “C for 1 h (T7 infection) or 3 h (IPTG and mGPl-2 induction). A 1 ml aliquot is removed from each sample, the cells are pelleted and then resus- pended in 80 pl of cracking buffer (60 mM TrisHCl pH 6.8/l% mer- captoethanol/lO% glycerol/3% SDS/O.01 % bromophenol blue). Half of this volume is loaded onto a 0.1% SDS-7.5% PA gel. E. coli RNA polymerase was run as molecular weight standard. The proteins are visualized by staining with Coomassie blue. The source of the T7 RNA polymerase is indicated above the lanes. Cells are E. coli HMS 174 containing either vector or vector with the POL3 gene and are either uninduced (-) or induced (+). Only the lane containing cells harboring POW plasmid that are infected with bacteriophage T7356- displays production of a 124-kDa protein (arrow) corresponding to DNA poly- merase 8, the POW product.

gene 0.7, which encodes a protein kinase that also inacti- vates the host RNA polymerase. There are several genes in the early expressed portion of the genome that encode rather small proteins that have been detected upon infec- tion but whose activities have not been characterized. These proteins may be peptide inhibitors of proteases or other metabolic enzymes. We reasoned that DNA poly- merase 6 might be more compatible with the host during infection by T7. A mutant phage that produces T7 RNA polymerase but that does not progress through its life cycle because of amber mutations in genes 3,5 and 6 was used. Genes 3 and 6 encode very potent endo- and exo- nucleases which normally degrade the host DNA during an infection. Gene 5 encodes the catalytic subunit of T7 DNA polymerase, which is essential for replicating the phage genome. Upon infection with these mutant phage, effective overproduction of DNA polymerase 6 is observed (Fig. 3). Within several minutes of infection

102

large amounts of DNA polymerase 6 are detectable. The cells generally lyse between 150 and 180 min after infec- tion. These experiments have been repeated a number of times with all three systems and the results are invariant, that is polymerase 6 is not produced using the traditional systems but large amounts are obtained using phage infection.

(d) Partially deleted and truncated constructs of DNA polymerase 6

In order to carry out simple structure function studies, several ~01338 mutants were constructed. The stable con- struct of the full-length PUL3 gene was cut at restriction sites tanking sequences predicted to encode exonuclease function and religated (Fig. 4 B and D). The first internal deletion, formed by EcoRV digestion, removes a 252”bp segment which contains the exoI’, exo1 and exoI1 sequences. Site-specific mutations in these sequences in vivo resulted in the apparent reduction of proofreading activity of polymerase 6 (Simon et al., 1991). Ligation

HindIll

product = 124 ku

Stan: ATG Stop: TAA

Produa = 114kDa

Iindfli

c HindIII HindIII

EmRV Ndel EcoRV

Product = 9% ma

I I MI11 I

Fig. 4. Deletion and truncation constructs of PUL3. The thin single lines represent removed sequences, medium shaded boxes are noncod- ing regions and the large open boxes are the ORFs with the six con- served regions indicated in black. The full-length construct is shown (A) as it appears in pT7SCjPOt3. This was digested with EcoRV and re-circularized (B). The resulting plasmid would give rise to a 1 ICkDa protein. A truncated form of POL.3 (C) was constructed by digesting both pBL304 and pT7SC with N&I +HindIII. The digests were then mixed for ligation. Re~rnbin~~ were screened first by size and then restriction mapping to identify the proper construct which should yield a 9%kDa protein. A larger deletion (D) was constructed by digesting pT7SC/POW with NcoI + EcoRV. The ends were filled in and the plas- mid was re-circularized. This should give rise to a 78-kDa protein.

regenerates a single EcoRV site which can be used to ensure the gene is in frame. The other internal deletion retains the first nine triplets at the N terminus but removes the next 1212 bp to the second EcoRV site, com- bining the effects of the first deletion and the following truncation. An N-terminally truncated form of the gene was also cloned into the Ndel site of pT7SC as shown in Fig. 5C. The N-terminal truncation beginning at the NldeI site removes 660 bp from the N terminus, while retaining all of the six conserved regions. This construct differs from the others in that the vector ribosome binding site and ATG spacing are retained (see position of XbaI and NdeI sites in Fig. 1).

In a number of trials both the full-length (124 kDa) and truncated (98 kDa) versions of the protein are overex- pressed. When bacteriophage infection is accompanied by over-production of protein, the protein profile observed is radically altered as evidenced by the lanes of 124-kDa and 98-kDa induction in Fig. 5, a representative SDS-PA gel of these experiments. The internal deletions (114 and 78 kDa) did not give rise to protein upon induc- tion (data not shown).

Site-specific mutants of PUL3 affecting the deleted exo- nuclease region showed DNA polymerase activity in vitro and complemented ~013 mutants (Simon et al., 1991). However, another study of site specific mutants of the homologous regions in HSV Pol (Gibbs et al., 1991) showed some mutants completely lacked polyme~zation activity. These authors suggested that the exonuclease region does not fold as a separate domain but interacts with the polymerase active site. If this is the case then deleting these regions may destabilize the protein and this

N- -98

Fig. 5. Synthesis of various forms of DNA polymerase S by infection with bacte~ophage T7356-. Cell lines contained the constructs indicated by molecular sizes (in kDa) and are either infected (+) or uninfected (-). Expression is detected only for the full-length (124 kDa) and truncated (98 kDa) constructs. The samples and gel were handled as described in the legend to Fig. 3.

103

is why expression from the 11CkDa and 7%kDa con- structs is not detected. In keeping with this interpretation, as stated above in section (b), a construct deleting the entire N terminus through the exonuclease domain (BsmI site) could be stably cloned in any vector, though the protein is not detected upon induction.

(e) Conclusions The vector presented here, pT7SC, successfully sup-

presses the read-through transcription of the toxic POL3 gene of S. cereuisiae thus stabilizing it in E. coli. The strategies most commonly used for expression of genes in T7-based vectors were not effective in overexpressing the POL3 gene. Infection with mutant T7 bacteriophage led to high levels of overexpression. We have obtained up to 15 mg of protein from as little as 3 grams of cells. The protein is not found in the soluble cell extract but is included and found in the insoluble material. Purification of active protein from the inclusion bodies will be pub- lished elsewhere (Brown et al., 1993).

There are many eukaryotic and also prokaryotic pro- teins that have potential clinical or industrial use but are not abundant. The ability to express them in bacteria would provide a large and readily available supply of such proteins. While initial gene isolation and renatu- ration of overexpressed proteins have become common, a problem remains in inducing the cells harboring the gene of interest to produce protein, when the product is for some reason incompatible with the host. The strategy we describe is a way of removing cellular control and directing the cell to produce these proteins in abundance. We are currently investigating the possibility of placing the early gene region of T7 on a plasmid to reduce the problem of cell lysis associated with phage infection.

ACKNOWLEDGEMENTS

The authors are grateful to Dr. Stan Tabor (Harvard University) for providing pT7-7 and mGPl-2 as well as advice, Dr. Peter Burgers (Washington University, St.

Louis) for providing pBL304, and Dr. Louis J. Roman0 (Wayne State University, Detroit) for providing T7356- phage. This work was supported by NIHGMS 25508 and NRSA CA08692 to W.C.B.

REFERENCES

Amann, E., Brosius, J. and Ptashne, M.: Vectors bearing a hybrid trp-

lac promoter useful for regulated expression of cloned genes in Esch-

erichia coli. Gene 25 (1983) 167-178.

Boulet, A., Simon, M., Faye, G., Bauer, G.A. and Burgers, P.M.J.: Struc-

ture and function of the Saccharomyces cereuisiae CDC2 gene encod-

ing the large subunit of DNA polymerase III. EMBO J. 8 (1989)

1849-1854.

Brown, W.C., Duncan, J.A. and Campbell, J.L.: Purification and charac-

terization of the Saccharomyces cereuisiae DNA polymerase 8 over-

produced in Escherichia coli. J. Biol. Chem. 268 (1993) 982-990.

Brosius, J.: Toxicity of an overproduced foreign gene product in Escher-

ichia cofi and its use in plasmid vectors for the selection of transcrip-

tion terminators. Gene 27 (1984a) 161-172.

Brosius, J.: Plasmid vectors for the selection of promoters. Gene 27

(1984b) 151-160.

Gibbs, J.S., Weisshart, K., Digard, P., deBruynkops, A., Knipe, D.M.

and Coen, D.M.: Polymerization activity of an a-like DNA polymer-

ase requires a conserved 3’-5’ exonuclease active site. Mol. Cell. Biol.

11 (1991) 4786-4795.

Perbal, B.: A Practical Guide to Molecular Cloning. Wiley, New York,

1984, p. 221.

Simon, M., Giot, L. and Faye, G.: The 3’ to 5’ exonuclease activity

located in the DNA polymerase 6 subunit of Saccharomyces cereuis- iae is required for accurate replication. EMBO J. 10 (1991)

2165-2170.

Sitney, K.C., Budd, M.E. and Campbell, J.L.: DNA polymerase III, a

second essential DNA polymerase, is encoded by the S. cereuisiae CDC2 gene. Cell 56 (1989) 599-605.

Smiley, J.K., Brown, WC. and Campbell, J.L.: The 66 kDa component

of yeast SFI, stimulatory factor I, is hsp60. Nucleic Acids Res. 20

(1992) 4913-4918.

Studier, F.W. and Moffat, B.A.: Use of bacteriophage T7 RNA polymer-

ase to direct selective high-level expression of cloned genes. J. Mol.

Biol. 189 (1986) 113-130.

Tabor, S. and Richardson, C.C.: A bacteriophage T7 RNA polymer-

ase/promoter system for the controlled exclusive expression of

specific genes. Proc. Natl. Acad. Sci. USA 262 (1985) 1074-1078.

Wong, S.W., Wahl, A.F., Yuan, P.-M., Arai, N., Pearson, B.E., Arai, K.-

I., Kern, D., Hunkapiller, M.W. and Wang, T.S.-F.: Human DNA

polymerase a gene expression is cell proliferation dependent and its

primary structure is similar to both prokaryotic and eukaryotic

replicative DNA polymerases. EMBO J. 7 (1988) 37-47.