High level expression of genes cloned in phage λgt11

Gene, 78 (1989) 93-99

93

Elsevier GEN 02978

High level expression of genes cloned in phage lgtll (Recombinant DNA; plasmids; vectors; lac promoter; homo-oligopeptides; leprae; E. coli host; T cells; Not1 restriction endonuclease)

fusion protein; Mycobacterium

Neil G. Stoker, Kathie A. Grant, Hazel M. Dockrell, Cohn R. Howard, Nathalie F. Jouy and Keith P.W.J. McAdam Departmentof ClinicalSciences, London School of Hygiene and Tropical Medicine, London WClE 7HT (U.K.) Received

by K.F. Chater:

Accepted

after revision:

9 December 16 January

1988 1989

SUMMARY

cloning vectors have been constructed which allow genes originally cloned in ,J gtl 1 to be expressed at a high level in Escherichia coli. They are based on the pEMBL and pUC vectors, with the genes transcribed from the lac promoter. The EcoRI site in the vector has been altered to be in the same reading frame as the site used for cloning in 1 gtl 1. Cloned proteins are expressed fused to a 2-kDa leader sequence containing a run of six Aparagine residues which considerably improves the stability of the recombinant proteins, but does not interfere with immunological assays. Using these vectors, the Mycobacterium leprae I8-kDa protein was expressed at 20 mg per litre of culture and constituted 15% of total cell protein. Plasmid

Phage I gtll (Young and Davis, 1983) is a vector widely used for cloning and expressing cDNA from eukaryotes and genomic DNA from prokaryotes. DNA is cloned into a unique EcoRI site at the 3’ end of the 1acZ gene, and, where the reading frame of the cloned gene coincides with that of IacZ, a BGal fusion protein is produced (containing the first

1007 aa of BGal). Clones of interest can be detected by screening the library with a monoclonal antibody or hyperimmune serum. Once a clone has been isolated it is generally desirable to subclone the insert into a plasmid vector for further analysis and improved expression. For example, subcloning the insert into pUR290 (Rtither and Milller-Hill, 1983) produces the same gene fusion construct as was present in the phage, though

Correspondence to: Dr. N.G.

phy; kb, kilobase

INTRODUCTION

Sciences, Keppel

London Street,

School London

Stoker,

of Hygiene WClE

7HT

Department

of Clinical

and Tropical (U.K.)

Medicine,

Tel. 01-6368636;

MLsom, nucleotide; peripheral

Fax 01-436 5389.

or 1000 bp; MCS,

PAGE, blood

polyacrylamide-gel mononuclear

saline (140 mM NaCl/l.5 Abbreviations: tosidase;

aa, amino acid(s); Ap, ampicillin;

bp, base pair(s);

min; HPLC,

BCG, MibovB BCG; cpm, counts per

high-performance

isopropyl-/I-D-thiogalactopyranoside; thoxycarbonyl;

HPLC,

0378-I 119/89/$03.50

jGa1, p-galac-

liquid chromatography; FMOC,

high performance

0 1989 Elsevier

IPTG,

9-fluorenylme-

liquid chromatogra-

Science Publishers

B.V. (Biomedical

phenylmethylsulfonyl single stranded; EDTAjlOO 0.5%

electrophoresis;

mM KH,P0,/2.7

fluoride,

mM NaCl;

(v/v) Triton

TENT

X-100;

cloning

site;

oligo, oligodeoxyriboPBMC,

cells; PBS, phosphate-buffered

TEN buffer,

/?-D-galactopyranoside;

Division)

multiple

Myobaccerium leprue sonicate;

mM KCl); PMSF,

SDS, sodium dodecyl 50 mM buffer,

XGal,

sulfate; ss,

Tris . HCl pH 8.0/l mM TEN

buffer

containing

5-bromo-4-chloro-3-indolyl-

[ 1, designates

plasmid-carrier

state.

94

this is often expressed much more efficiently. One disadvantage with such constructs is that most (often 90% or more) of the recombinant protein is fiGal, so that even ifit makes up 25 y0 of the total cell protein, only 2-3 y0 is actually foreign peptide. More importantly, we have encountered interference in immunological assays from the /IGal portion of the fusion protein. We describe the development of plasmid expression vectors which overcome these problems, using the observation of Sung et al. (1986) that homo-oligopeptides stabilise the expression of preproinsulin in E. coli. A nucleotide sequence coding for six consecutive Asn residues has been inserted close to the start of the 1acZ gene in pUC and pEMBL vectors, and the EcoRI site shifted into the appropriate frame. Foreign peptides are stably expressed from these vectors fused to a 2-kDa leader peptide.

MATERIALS

AND METHODS

(a) Chemicals and reagents

General chemicals were obtained from BDH Ltd (Poole, U.K.) and Sigma (Poole, U.K.). Restriction enzymes and DNA-modifying enzymes were obtained from Gibco-BRL (Paisley, U.K.) and Boehringer-Mannheim (Lewes, U.K.). The monoclonal antibody L5 (B&ton et al., 1985) was kindly provided by Drs. Warwick Britton and Tony Basten. (b) Bacterial and bacteriophage strains E. coliK-12 strain BMH71-18 has been described (Messing et al., 1977). Bacteriophage 2 clone Y3 179 (Young et al., 1985) was kindly provided by Dr. R.A. Young. Bacteriophage fl R408 (Russel et al., 1986) was purchased from NBL Enzymes Ltd. (Cramlington, U.K.). (c) Plasmids

Plasmids used were pUR290 (RUther and Muller-Hill, 1983), pEMBL8( +) and pEMBL8( -) (Dente et al, 1983), pUC8 (Vieira and Messing, 1982), pUC18 (Norrander et al., 1983) and pUC8-2 (Hanna et al., 1984).

(d) General methods

Standard methods were used for DNA manipulation (Maniatis et al., 1982). Phage I DNA was prepared as described by Davis et al. (1980). DNA fragments were gel-purified using the method of Dretzen et al. (198 1). Plasmid mini-preparations were performed using the method of Holmes and Quigley (1981). Double-stranded DNA was sequenced using the method of Murphy and Kavanagh (1988). Site-directed mutagenesis was performed using the method of Nakamaye and Eckstein (1986) (Amersham International, Amersham, U.K.). (e) Oligo and peptide synthesis

Oligos were synthesized using phosphoramidite chemistry in a Pharmacia Gene Assembler, and were used without further purification. The peptide MTMITNYNNNNNNMY was synthesized using F-MOC chemistry on a Beckman System 990 synthesizer and purity confirmed by aa analysis and reverse-phase HPLC. (f) Construction of pNGS8(+)

Plasmid pNGS8( +) (see Fig. 1) was constructed by synthesizing the partially complementary oligos, 5’-AATTAC(AAC),ATGG-3’ and 5’-AATTCCAT(GTT),GT-3’. The non-phosphorylated oligos annealed, were ligated to EcoRI-digested pEMBL8(+), introduced into BMH71-18 and selected on plates containing 50 pg Ap/ml, 5 mM IPTG and 40 pg XGal/ml. White colonies were picked and checked by DNA sequencing. (g) Construction of pNGS18( + )

Plasmid pNGS 18( +) was constructed by replacing the MCS of pNGS8(+) with that from pUC18. The large EcoRI-Hind111 fragment from pNGS8( +) and the small EcoRI-Hind111 fragment from pUC18 were gel-purified, ligated together and used to transform BMH7 1-18. Plasmids from transformants were checked by DNA sequencing. (h) Construction of pNGSZO(+)

Plasmid pNGS20( +) (see Fig. 6) was constructed by site-directed mutagenesis of pNGSlS( +). An

95

oligo (5’-GACGGCCAGTGCGGCCGCAAGCTTGCATGC-3’) was synthesized, phosphorylated, and used to mutagenize ss pNGS 18( +) (prepared by infectionofBMH71-18[pNGS18(+)]withphagefl R408). Plasmids from transformants were screened for the presence of a Not1 site and subsequently checked by DNA sequencing. (i) Construction

producing pML18 (Fig. 1). BGa.l fusion protein @Gal-18K) was purified by preparative SDS-PAGE, and was used in human T cell proliferation assays. Control antigen included jIGal expressed from pUR290 and purified in the same way, as well as enzyme obtained from commercial sources. In proliferation assays with human peripheral blood mononuclear cells, there was a substantial antigen-specific response to the /IGal portion of the molecule, masking any response to the foreign peptide. Profiles from two donors shown in Fig. 2 demonstrate significant proliferation to purified jIGal. The antigen specificity was demonstrated by the cloning of /3Gal responsive T cells (H.M. D., unpublished results).

of pNGSZO(-) and pNGS21

The small PvuII fragment from pNGS20(+), which contains the MC& was purified and ligated to the large PvuII fragments from pEMBL8(-) and pUC8. Plasmids from transformants were screened by restriction digestion, and the orientations of the inserts were checked by DNA sequencing.

(b) Expression without /?Gal

RESULTS

An immunological response to the leader peptide is a potential hazard with any fusion protein, but the shorter the leader peptide is, the less likely it is to contain epitopes which will interfere in the assays. We recloned the insert from Y3179 into pUC8-2 (a derivative of pUC8 with the EcoRI site in the same reading frame as in ;1 gtll), producing pML22 (Fig. 1). Since the cloning site is near the beginning

AND DISCUSSION

(a) Immunological assays using BGal fusion proteins

The EcoRI fragment which encodes the C-terminal two-thirds of the M. leprue 18-kDa protein was recloned from the Izgtl 1 clone Y3 179 into pUR290,

Bg Fig. 1. Cloning of the M. leprae 1%kDa gene into plasmid and subcloned Open arrows protein

into plasmid

vectors

pUR290,

show the ZacZ coding region transcribed

coding sequence.

intergenic

expression

B, BamHI;

region containing

expression pUC8-2

from the lac promoter;

Bg, BgZII; C, ClaI; E, EcoRI;

the origin of replication.

vectors.

and pNGS8( H, HindIII;

The 3.2 kb EcoRI insert in Y3179 (open bar) was excised + ) generating

plasmids

the blackened

portion

P, &I;

pML18,22 of the arrows

and 27, respectively. shows the 18-kDa

Pv, PvuII; S, SalI; Sm, SmaI. fl(IG),

phage fl

96

formyl methionine,

including

the EcoRI site). It was

hoped that this would allow inducible

expression

of

the foreign peptide without most of /IGal, thus providing most of the benefits

of the previous

system,

with fewer drawbacks. However, the resulting fusion protein was produced in variable and extremely small quantities

(Fig. 4~).

(c) Construction

of pNGSS( + )

One explanation from pML22

the short (1-kDa)

Fig. 2. T-cell proliferation

assays. Bar diagram

ation of human PBMC in the presence and B) were both

U.K. BCG vaccinees.

(i) M. leprae sonicate

protein

Institute

the WHO

(Bgal-18K);

Asn-18K

protein;

leader sequence. mined optimal

(v) synthetic The antigens

were performed scales

obtained

from heparinised assays

10% autologous round-bottom with

incubated pulsed

U.K.)

added

Amersham,

Inc., Watertown, Pangbourne,

fusion protein deter-

were

three

times.

1640 medium

25 mM HEPES/Z

mM

of streptomycin

and

g/ml

Roskilde,

concentrations

and

air with 5% CO,. Cultures

were

1 PCi [3H]-thymidine/well U.K.), harvested

MA) and counted ofoptifluor

Denmark)

indicated,

fibre paper using a PHD Ceil Harvester counter after addition

(Fig. 3), and foreign peptides are expressed fused to a 2-kDa leader peptide. pEMBL8( +) was used so

over Histo-

washed

in RPM1

plates (Nunc,

at the

on day 6 with

was not protecting the heterologous peptide to the same extent as the /IGal portion of fiGal-18K. It has previously been shown that certain homo-oligopeptides stabilise the expression of preproinsulin in E. coli (Sung et al., 1986). Accordingly, a synthetic DNA linker containing the coding sequence for six consecutive Asn residues was inserted close to the start of the 1ucZ gene in pEMBL8( + ), producing pNGS8(+) (Figs. 1 and 3). The EcoRI site in this vector is in the same reading frame as in A gtll

Cells were plated at 2 x 10s per well in

at 37°C in humidified

International,

and

U.K.) containing

96 microtitre

antigens

of the Asn-18K

blood after centrifugation

penicillin/100

plasma.

because

acid sequence

with no extra antigen added. Note that

Poole,

IU/ml

amino

of 10 pg/ml (i) and 1 ng/ml (ii-v).

were performed

(Flow, Rickmansworth, glutamine/lOO

degradation,

(iv) purified

have been used on the Y axes. PBMC

1077 (Sigma,

Proliferation

peptide

(Bgal);

by

N-terminal

Mill Hill

/IGal-18K

were used at the previously

concentrations

different

Research,

(ii)purified BGal

used were

kindly provided

for Medical

Bank;

(iii) purified

Controls

paque

Antigens

(batch CD99) (MLson),

Dr. R. Rees, National (U.K.) through

showing prolifer-

of antigen. The donors (A

for the low level of expression

was proteolytic

(Amersham

@w

TNYNNNNNNMEF T M I ATOACCITOA~TICOI*TT*C**CUCUCMCMC~*TG-O*GCTC

16 h later onto glass

(Cambridge

Technology,

on a Beckmann

scintillation

scintillant

PW=J+

(Canberra

Packard,

U.K.). Results shown are the mean cpm from tripli-

cate wells.

EC&l

T&i-

KpnlsMIimHI

Fig. 3. Sequences

Sal,

T

spill

TGzi-

Not,

of the start of the IacZ genes in plasmids

used

in this study. The genes are shown from the start codon to a point beyond

the MCS. The first 19 nt and the last 9 nt in each case

are identical, the EcoRI

of the 1ucZ gene it would result in a leader peptide of 33 aa, compared with 1007 aa in PGal (excluding the

sscl

~COOEPbU;I;TCT*GAEULEb(;CTOCAOPCb~~e

and are shown italicised.

part in a-complementation nation).

The aa sequences

sites are shown. The ability of each plasmid is indicated

up to to take

(see blue or white desig-

97

that ss DNA could be produced for subsequent sitedirected mutagenesis.

A

C

6

E

D

(d) Expression from pNGSS(+)

The EcoRI fragment from Y3179 was recloned into pNGS8( + ), producing pML27 (Fig. l), and expression of the 18-kDa protein was compared to that from pML22 (Fig. 4). After a 2-h induction with IPTG, significant synthesis could be seen from pML27, where none was detected from pML22. After overnight induction, yields of over 20 mg of recombinant protein per litre of culture were obtained, estimated to be 15% total cell protein by densitometry of a Coomassie blue-stained gel (Fig. 5). For good yields, rich broth and good aeration (no more than 50 ml in a 250~ml flask) were required. The expressed protein was found to be insoluble, and was purified as shown in Fig. 5. Throughout, the identity of the 18-kDa protein was confirmed by probing Western blots with the mouse monoclonal antibody L5 as described by Vega-

Fig. 5. PurificationofAsn-18K. containing

with IPTG; (C-E) following (E) afler

gel filtration.

SDS-PAGE staining.

arrow. The molecular

_B

C

D

E

F

resuspended

200 PM PMSF,

at room

room temperature.

Fig. 4. Expression

of recombinant

Whole-cell

lysates

of BMH71-18

(B) pMLl8,

(C) pML22, (D) pNGSB( +). (E and F) pML27, after

2 h induction nant

with (A-D,

Asn-18K

standards

are shown

were diluted containing

protein

18-kDa

fusion

proteins.

containing

(A)no

plasmid,

F) or without is indicated

(E) IPTG. The recombi-

by an arrow.

on the left (in kDa).

1: 50 into 2 x TY (tryptone

yeast

0.2% glucose and grown at 37°C to&,,,

was added to 1 mM, and the incubation 20 h. For analysis, 7.2) at A,,

cells were harvested,

= 20, and aliquots

The Mrs of

Overnight

broth

= 0.5. IPTG

was continued resuspended

cultures

extract)

for up to

in PBS (pH

were mixed with SDS sample

buffer. Ofeach

sample 2 ~1 were analyzed

44% monomer:

0.8% bis) and stained

by SDS-PAGE with Coomassie

(16%; blue.

was as follows: 2 g cells (wet

in 20ml

temperature.

were added

TEN

buffer

containing

and 0.7 ng pepstatimml. for 30 min on

MgCl,

followed

(to 3 mM)

and

by 30 min mixing at

EDTA was added to 10 mM and the lysate buffer,

mM Tris dialysed

dissolved

in 6 M guanidine

HCI pH 7.5 and fractionated

using a TSK G3000SW collected,

by an

at 5000 x g for 10 min. The pellet was washed

with TENT

chloride/lo

is indicated

are shown on the left

acid was added to 1.3 mg/ml, followed by

(to 7 pg/ml)

was centrifuged

buffer; by 0.1%

bis) and Coomassie-blue

was added to 250 pg/ml and incubated

30 min mixing

twice

protocol

analysed

protein

0.5 ng leupeptin/ml

ice. Solid deoxycholic DNase

Asn-18K

with TENT

were

masses of standards

(in kDa). The purification were

of(B); (C) 5000 x g pellet

38 : 2 monomer:

(14%,

ofBMH71-18

after 20-h induction

after one wash

The samples

The recombinant

weight)

Whole-celllysates and (B) pML27,

show fractionation

lysis; (D) pellet

Lysozyme

A

(A) no plasmid

against

gel filtration

(Anachem).

hydro-

by HPLC Peaks

were

PBS and concentrated.

Lopez et al. (1988). Using pNGS8( + ), four other fusion proteins, derived from Schistosoma mansoni and Onchocerca volvzdus, have been successfully expressed at a high level. In contrast to the M. leprae 18-kDa protein, these recombinant proteins are found in the soluble fraction (P. Francis and J. Bradley, personal communication). The recombinant protein produced by pML27 (Asn- 18K) was purified and used in T-cell proliferation assays. Significant responses to the protein were observed (Fig. 2; H.M.D., N.G.S., S.P. Lee, M. Jackson, K.A.G., N.F.J., S.B. Lucas, C.R.H., R. Hussain and K.P.W.J. McA, submitted for publication). As a control, the leader peptide was synthesized and tested in the T cell assays. No significant responses were observed with cells from BCG-vaccinated donors. It is of interest that no T cell epitopes

98

are predicted in the peptide using the algorithm of Rothbard and Taylor (1988). We conclude that the fusion protein is suitable for use in assays with uncloned T cells. (e) Refinements

As a further refinement, the pNGS8( +) MCS was replaced with that from pUC18, producing pNGS18(+). The reading frame following the EcoRI site was then restored to allow a-complementation and a blue-white selection in cloning experiments. A Not1 site was introduced in the process, producing pNGS20( +) (Fig. 6). The (-) derivative, and a pUC derivative were constructed by purifying the small PvuII fragment from pNGS20(+) and replacing the equivalent fragments in pEMBL8(-) and pUC8. These constructs were named pNGS20(-) and pNGS21, respectively. The pUC plasmids have recently been shown to lack a niclbom site (Minton et al., 1988), reducing any ‘imaginary hazard’ in genetic manipulation experiments. (f) Conclusions

We have constructed plasmids which allow the efficient expression of foreign proteins fused to a

short leader peptide in E. coli. The insertion of a homo-oligopeptide in the leader sequence had a dramatic effect on the stability of the recombinant protein, as originally reported, although the reason for this is not known. Because of the widespread use of il gtl 1, the MCS is arranged with the EcoRI site closest to the start of the gene, and is in the appropriate reading frame. Recombinants may be identitied using /3Galc+complementation. The ss DNA may be prepared from pNGS20( +) and pNGS20(-) for the purposes of nucleotide sequencing, site-directed mutagenesis and ss probe synthesis. The vectors have been used in this and other laboratories to express proteins from A4. leprae, S. mansoni and 0. volvulus. These plasmids are particularly useful where T-cell proliferation assays are to be performed, due to the lack of background response to the leader peptide. Another approach to this problem is to synthesize a larger fusion protein, and to remove the leader by enzymatic means (e.g., Nagai and Thogersen, 1987), but this introduces extra steps into the purification process. We are currently developing a method for the aflinity purification of recombinant proteins generated in this system using monoclonal antibodies raised to the leader peptide. It is our experience that most BGal fusion proteins are insoluble if expressed in large quantities, whereas four out of live proteins expressed from these vectors have been found to be soluble. This would be a great advantage if affinity purification is to be used. Furthermore, if antibody studies are to be performed, it is preferable to avoid the strong denaturing agents such as guanidine hydrochloride, which are necessary with insoluble proteins.

ACKNOWLEDGEMENTS

Fig. 6. Restriction

map of pNGS20(

+). Most abbreviations

listed in the legend to Fig. 1. Additional N, NorI; SC, SucI; Sp, SphI; X, XbaI.

are

ones used are: K, KpnI;

We thank Drs. Warwick Britton and Tony Basten for their gift of L5 monoclonal antibody, Dr. Richard Young for Y3179, Drs. Rabia Hussain and John Raynes for advice on the protein purification, Bala Ramesh and Brent Ryan for technical assistance, John Garbera for artwork and Dr. Tim Harrison for comments on the manuscript. This work was supported by the Wellcome Trust and the Medical Research Council (U.K.).

99 REFERENCES B&ton, W.J., Hellqvist, L., Basten, A. and Raison, R.L.: Mycobucreriwn leprue antigens involved in human immune responses, I. Identification offour antigens by monoclonal antibodies. J. Immunol. 135 (1985) 4171-4177. Davis, R.W., Botstein, D. and Roth, J.R.: Advanced Bacterial Genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1980. Dente, L., Cesareni, G. and Cortese, R.: pEMBL: a new family of single-stranded plasmids. Nucleic Acids Res. 11 (1983) 1645-1655. Dretzen, G., Bellard, M., Sassone-Corsi, P. and Chambon, P.: A reliable method for the recovery of DNA fragments from agarose and acrylamide gels. Anal. Biochem. 112 (1981) 295-298. Hanna, Z., Fregeau, C., Prefontaine, G. and Brousseau, R.: Construction of a family of universal expression vectors. Gene 30 (1984) 247-250. Holmes, D.S. and Quigley, M.: A rapid boiling method for the preparation ofbacterial plasmids. Anal. Biochem. 114 (1981) 193-197. Maniatis, T., Fritsch, E.F. and Sambrook, J.: Molecular Cloning. A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1982. Messing, J., Gronenbom, B., Mttller-Hill, B. and Hofschneider, P.-H.: Filamentous coliphage Ml3 as a cloning vehicle: insertion of a Hind11 fragment of the luc regulatory region in Ml3 replicative form in vitro. Proc. Natl. Acad. Sci. USA 74 (1977) 3642-3646. Minton, N.P., Chambers, S.P., Prior, SE., Cole, ST. and Gamier, T.: Copy number and mobilization properties of pUC plasmids. Focus 10 (1988) 56.

Murphy, G. and Kavanagh, A.: Speeding-up the sequencing of double-stranded DNA. Nucleic Acids Res. 16 (1988) 5198. Nagai, K. and Thogersen, H.C.: Synthesis and sequence-specific proteolysis of hybrid proteins produced in Escherichia coli. Methods Enzymol. 153 (1987) 461-481. Nakamaye, K.L. and Eckstein, F.: Inhibition of restriction endonuclease NciI cleavage by phosphorothioate groups and its application to oligonucleotide-directed mutagenesis. Nucleic Acids Res. 14 (1986) 9679-9698. Norrander, J., Kempe, T. and Messing, J.: Construction of improved Ml3 vectors using oligodeoxynucleotide-directed mutagenesis. Gene 26 (1983) 101-106. Rothbard, J.B. and Taylor, W.R.: A sequence pattern common to T cell epitopes. EMBO J. 7 (1988) 93-100. Russel, M., Kidd, S. and Kelley, M.R.: An improved tilamentous helper phage for generating single-stranded plasmid DNA. Gene 45 (1986) 333-338. Rtither, U. and Mllller-Hill, B.: Easy identification of cDNA clones. EMBO J. 2 (1983) 1791-1794. Sung, W.L., Yao, F.-L., Zahab, D.M. and Narang, S.A.: Short synthetic oligodeoxyribonucleotide leader sequences enhance accumulation of human proinsulin synthesized in Escherichia coli. Proc. Natl. Acad. Sci. USA 83 (1986) 561-565. Vega-Lopez, F., Stoker, N.G., Locniskar, M.F., Dockrell, H.M., Grant, K.A. and McAdam, K.P.W.J.: Recognition of mycobacterial antigens by sera from patients with leprosy. J. Clin. Microbial. 26 (1988) 2474-2479. Vieira, J. and Messing, J.: The pUC plasmids, an Ml3mp7derived system for insertion mutagenesis and sequencing with synthetic universal primers. Gene 19 (1982) 259-268. Young, R.A. and Davis, R.W.: Efficient isolation of genes by using antibody probes. Proc. Natl. Acad. Sci. USA 80 (1983) 1194-l 198.