Extension of bacteriophage A host range: Selection, cloning, and

Proc. Nati. Acad. Sci. USA Vol. 81, pp. 6080-084, October 1984 Genetics

Extension of bacteriophage A host range: Selection, cloning, and characterization of a constitutive X receptor gene (LamB protein/insertion sequence 3/transposidon/cosmid transduction)

GERT E. DE VRIES, CHRIS K. RAYMOND, AND ROBERT A. LUDWIG* Department of Biology, Thimann Laboratories, University of California, Santa Cruz, CA 95064

Communicated by Dale Kaiser, May 21, 1984

ABSTRACT A set of plasmids has been constructed that carry a constitutive lOB gene (LamBc phenotype) from Escherichia coli and that confer functional phage A receptors to bacteria other than E. coli. This E. coli LamBc strain has been selected to escape both maltose-inducible and glucose-repressible control. Constitutivity results from an IS-3 insertion, carrying a mobile promoter, proximal to lamB. The LamBc DNA has been cloned into both broad and narrow host-range plasmids, and the resulting pTROY plasmids have been transferred to diverse bacteria. Both Salmonela typhimuriumi pTROY and Klebsiefla pneumoniae/pTROY strains efficiently adsorb phage A; Pseudomonas aeruginosa/pTROY strains do not. Introduction of a functional E. coli LamB protein into foreign bacteria will allow these bacteria carrying pTROY plasmids to be infected by phage A recombinant DNA libraries, phage K::Tn insertion mutagenesis vectors, and in vivo Kpackaged cosmids.

(it) select pseudorevertants in which lamB was derepressed and subsequently scored for constitutivity. The efficacy of this protocol required that it be possible to select strains that expressed lamB from a population of cells in which lamB was severely repressed. Because in wild-type E. coli there exists no known physiological condition under which lamB is sufficiently repressed to allow such a selection (19), lamBrepressed strains were first selected. Here, we report the selection, characterization, and cloning of such a constitutive lamB gene as well as initial studies of its function in other bacteria.

Escherichia coli phage X is a powerful tool for molecular genetic analysis (1-3). However, its power has been largely limited to E. coli, being essentially the only bacterium able to adsorb phage X (4). A single receptor protein, the product of the lamB gene, facilitates X adsorption to E. coli (5, 6). All signals necessary to export and position LamB protein in the outer membrane appear to be encoded in the lamB structural gene itself (7). Therefore, it appeared feasible to extend X host range by introducing a functional E. coli lamB gene into diverse bacteria. LamB protein is a major constituent of the E. coli outer membrane (8). It spans the membrane, is associated with peptidoglycan, and is generally classified as a porin (8-10). LamB protein serves as receptor for several coliphages (5, 11, 12) and participates with maltose-binding protein (MalE protein) in the transport of maltose and maltodextrins across the outer membrane (9, 13, 14). The lamB gene is part of the malK-lamB operon in the MalB region (15); the unlinked MalA and MalB regions comprise the maltose regulon (16). The expression of MalB genes, including lamB, is positively regulated by the malT gene product, located in MalA, and by the catabolite receptor protein-cAMP complex (17, 18); lamB is, thus, maltose inducible and glucose repressible. Because these ancillary proteins controlling lamB expression might neither exist nor function in bacteria other than E. coli, the selection of a constitutive lamB gene whose expression was uncoupled from maltose and glucose control was considered expeditious. At a minimum, we hypothesized that new host bacteria expressing a constitutive lamB gene (LamBc phenotype) would adsorb phage X and thus allow injection of X-packaged DNA. This would greatly facilitate recombinant DNA analyses in diverse bacteria. We undertook to select E. coli LamBc strains in two steps: (i) select strains in which lamB was severely repressed, and

MATERIALS AND METHODS Bacterial Strains and Media. Bacteria, phages and plasmids used are listed in Table 1. Either LB (32) or RM [R medium (33) containing 10 mM MgSO4 and 0.1% maltose] were used as bacterial growth media. Minimal medium consisted of M9 salts (32) with carbon sources supplied at 0.4%, amino acids at 0.01%, and vitamins at 1 ,g/ml. E. coli derivatives were screened for maltose phenotypes on EMB-maltose indicator plates (33). Antibiotic concentrations (pg/ml) used were as follows: E. coli kanamycin (Km), 50; tetracycline (Tc), 10; and ampicillin (Ap), 100; Klebsiella pneumoniae Km, 50; Tc, 10; Ap, 1000; Salmonella typhimurium Km, 100; Tc, 10; Ap, 100; Pseudomonas aeruginosa Km, 2000; Tc, 30. All bacterial conjugations and other manipulations were carried out at 370C unless otherwise specified. Phages and Plasmids. Phage stocks were prepared from plate lysates as described (33). Cosmid pN4 is a 46-kilobase (kb) triple cointegrate (pLAFR::pBR322::pLAFR) having both pLAFR (29) copies in the same orientation; pSil is a pJB8 (31) cosmid derived from a Rhizobium sp. ORS571 genomic library (unpublished results). Cosmid transducing lysates (34) were prepared by XcI857 infection of a cosmidcontaining E. coli strain at 420C. Phage Adsorption Assays. In vivo X receptor activity of E. coli strains was estimated two ways: (0) cosmid transduction (see below), and (ii) adsorption of [3H]thymidine-labeled X particles to cells, measured as a function of cell concentration, as described in the text. Cells were grown in minimal medium with either maltose or glucose as carbon source, harvested, washed with SM buffer (32), and diluted in SM buffer; 0.5-ml cell samples were mixed with 5ka of 3H-labeled phage (5.6 x 10" phage mCi'1; 1 Ci = 37 GBq) and incubated 30 min at 350C. Cells and adsorbed X particles were removed by centrifugation, and unadsorbed phage in the supernatant were counted by liquid scintillation. Bacteria other than E. coli to be tested for X receptor activity were grown to late exponential phase in RM broth at 370C. Antibiotics were added where appropriate to select maintenance of plasmids. Culture samples (0.2 ml) were added to phage dilutions (0.2 ml) and incubated at 370C for 20 min to adsorb

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Abbreviations: Km, kanamycin; Tc, tetracycline; Ap, ampicillin; kb, kilobase(s). *To whom reprint requests should be addressed.

6080

Genetics: de Vries

et

Proc. Natl. Acad Sci USA 81 (1984)

aL

Table 1. Bacteria, phages, and plasmids Genotype, phenotype Strain E. coli Pro-, Leu-, recA, hsdS(R-, M-) HB101::TnS malT, AR, KmR ECG1 malK::IS-3, LamBc, KmR ECG1::Tn5-132 malK::IS-3, LamBc, TcR, KmR ECG1OAamB XR Pro- Su6' MM294[::pRP4ATnl, Tcs] recA Tra(inc P-1)', KmR K. pneumoniae nifA, his, hsdR nifD, hsdR hsdR S. typhimurium met, val, hsdR P. aeruginosa met Bacteriophages

HB101 ECG1 ECG10 ECG16 ECG17 JMSu6 SM10

CK263 CK2901 KP5022 SL4213 PA02175 AcI857 A::Tn5 A::Tn5-132 Plvir

b221, Oam, Pam, rex::Tn5, cI857 b221, Oam, Pam, rex::Tn5-132, c1857 Plasmids tet+, oriT pRK290 ACOS, BamHI linker(RI-BamHI-RI)

pRK290 pLAFRB pBR322 pTROY9, -10 pTROY11 pJB8

pLAFRB malK::IS-3, LamBc pBR322 malK::IS-3, LamBc TcO pBR322 ACOS, TcO, ApR pJB8/Rhizobium sp. ORS571 insert pLAFR::pBR322::pLAFR triple cointegrate colEl::pRK2 (pRP4), KmR, Tra+

pS11 pN4 pRK2013

phage. Cells scribed.

were

then plated

or

treated further,

as

de-

RESULTS Selection of an E. coli LamBC Mutant Strain. An E. coli strain whose lamB expression was independent of maltose induction and glucose repression was selected in two steps. In the first step, the phage X-resistant (XR) strain ECG1 (Table 1) was selected as a mutant that expressed lamB at a very low level. By phage P1 transductional analysis and complementation tests, it was established that strain ECG1 was MalT-, which accounted for its XR phenotype and its ability to still adsorb X-packaged DNA at low levels (Table 2; refs. 35 and 36). In the second step of the selection procedure for a constitutive lamB mutant, phage X-sensitive (XS) revertants of ECG1 that expressed lamB at a high level were sought. ECG1 was infected with X::TnS-132 (Table 1), which carries the TcR transposon TnS-132, a derivative of TnS that transposes as does Tn5 (27). Because Tn5-132 transposes at low Table 2. Xd(pN4) cosmid transduction frequencies Phage dilution 100 10-1 10-2 1o-3 10-5 10-6 Recipient strain 2 NT NT 30 >500 440 ECG 1 7 NT NT NT >500 45 ECG10 NT NT NT 0 0 0 ECG17 0 0 108 ECG1/ECG1O (1000:1) >500 480 440 TcR colonies were scored after infection with the designated phage dilution. The Xd(pN4) cosmid transducing lysate contained -5.7 x 109 XcI857 per ml. The cosmid transducing particle titer was not determined in this experiment. NT, not tested.

6081

Ref.

This work This work This work This work 20 21 22, 23 22, 23 24 25 26 2 27

28 29; this work 30 This work This work 31 This work This work 28

frequency (-10-4 per infected cell), derivatives of ECG1 able to adsorb many X particles should stand a statistically greater chance of acquiring TnS-132 (TcR) transpositions. In a culture to which many more phage were added than bacteria, cells that expressed LamB protein at a high level could be multiply infected and, thus, were more likely to become TcR than ECG1 cells, which express lamB at a very low level. Therefore, ECG1 was infected with X::Tn5-132 at a moi of 50 and plated on LB + Tc. TcR colonies containing TnS-132 transpositions were isolated at a frequency of _10-8. TcR strains were scored for their LamB phenotypes by crossstreaking against XcI857 on either R plus maltose (RM) or R plus glucose (RG) at 42°C. Wild-type LamB+ strains are Xs on RM but because of glucose catabolite-mediated repression of lamB are XR on RG (17, 18). LamB- strains are XR on both media. One strain, ECG16 (Table 1; Fig. 1), was Xs on both RG and RM; it thus had the desired LamBc phenotype. ECG16 remained Mal-, KmR, like parental strain ECG1. However, in the absence of both Tc and Km selection, ECG16 lost one or the other resistance determinant (data not presented). This segregation did not affect lamB constitutivity as measured by X adsorption, however, and may have resulted from interactions between TnS and Tn5-132 elements present in ECG16 (37). Since phage P1 transductional analysis of ECG16 again showed that both TnS and TnS-132 elements were unlinked to LamB and Mal phenotypes (data not presented), this segregation was deemed irrelevant to the isolation of a constitutive lamB mutant. A derivative of ECG16 was, therefore, isolated by maintenance under Km selection only, to minimize problems of genome instabilities. This strain, called ECG10, was found to be stably KmR without selection and remained LamBC (Fig. 1). Tests of X adsorption to ECG10 showed that its phage adsorption efficiency was unaffected by cell growth on either maltose or

Genetics: de Vries

6082

probe:LamB

et

aL

Proc. NatL Acad Sci USA 81 (1984)

Pvull

Bgl II

l_

o

Vo

P-

QD

W

0

0

X

LUJ

LU

..Wll.

..Ir'.

"

0

_

o

-

!:":.iii.,

L-

0

w

L)

L)

LU

x

u]

wl

......

1.

0

W-

P-

U] L.U

D

LULLJ

.,

..L

i ,. "Ill.

FIG. 1. Southern genomic blot hybridizations of ECG strains with a lamB probe. Total bacterial DNAs were isolated by the cleared lysate procedure (32), subjected to CsCl gradient equilibrium centrifugation, precipitated twice with ethanol, digested to completion with either Pvu II or Bgl II endonucleases, subjected to electrophoresis through 0.7% agarose, and transferred to nitrocellulose sheets. The hybridization probe was a Sal I fragment isolated from pTROY9 containing malK distal and lamB proximal sequences (see Fig. 2) and was nick-translated with [a-32P]dCTP. Chromosomal DNA preparations were purified by equilibrium centrifugation (32), digested to completion with the indicated restriction endonuclease, subjected to agarose gel electrophoresis, and transferred to nitrocellulose. Hybridizations were conducted in 5 x SSPE (1 x SSPE = 180 mM NaCl/10 mM NaPO4, pH 7.7/1 mM EDTA) (32), 0.3% NaDodSO4/50%o formamide/100 Ag of carrier salmon sperm DNA per ml at 42°C. After hybridization, filters were serially washed with 2 x SSPE/0.1% NaDodSO4 at 420C, 2x SSPE at 420C, and 0.3x SSPE at 650C (melting temperature - 170C). The entire malK-lamB operon is carried by a single Bgl II fragment.

glucose and was -1/10th that of maltose-induced LamB+ strain HB101 (Table 3). Cloning and Characterization of LamBC Plasnids. Because the malK-lamB operon is contained on a single Bgl II DNA fragment in wild-type cells (38-40), Bgi II was chosen to digest ECG10 genomic DNA for subsequent cloning. ECG10 DNA was cloned into the BamHI site of the broad host range, restricted copy-number cosmid pLAFRB (Table 1). DNA from an ECG10 genomic library in pLAFRB was used to transform LamB- strain ECG17, a nonreverting XR derivative of ECG10 that exhibited no residual X adsorption when

tested by cosmid transduction (Tables 1 and 2). TcR transformants were infected with a cosmid-transducing lysate of cosmid pS11 (Table 1), which confers ApR and is compatible with pLAFRB (31). ApR transductants were screened for Xs LamBc phenotypes, and plasmids from candidate strains were isolated and again tested for their capacity to transform ECG17 to TcR. Two distinct plasmids conferring TcR were isolated (pTROY9 and pTROY10) that carried the LamBc gene, as evidenced by X cross-streaking tests. Restriction endonuclease mapping analysis of pTROY9 and pTROY10 revealed that both carried a 7.5-kb Bgl II fragment with restriction sites similar to those reported for malK-lamB (38-41). Plasmid pTROY10 carried an additional 2.2-kb Bgl II fragment, unlinked in the original ECG10 genome, now distal to lamB. Both pTROY9 and pTROY10 were found to have 1.3 kb more DNA in the common Bgl II insert than is present in wild-type HB101; this additional DNA carried two HindIII sites and one Pvu II site (Fig. 2). The HindIII sites, their position, and the size (1.3 kb) of the additional DNA suggested that an IS-3 insertion (42) was present in malK DNA of pTROY9 and pTROY10. To test this suggestion, the HindIII internal fragment (42) of the putative IS-3 element was isolated from pTROY9, radiolabeled, and used to probe HB101 and ECG genomic DNAs in hybridization experiments. Fragments homologous to this probe were observed to be multiply represented in these genomes (Fig. 3), consistent with the statuis of an IS element. That this element corresponded to IS-3 was suggested by a more detailed restriction endonuclease analysis. IS-3 contains a single internal Pvu II site between the HindIII sites (42); genomic DNA digested with Pvu II, therefore, exhibits two DNA fragments per IS-3 element, whereas Bgl II, which does not cleave IS-3, yields a single DNA fragment per IS-3 element. HB101 and ECG1 were found to carry several genomic IS-3 elements, consistent with previous estimates of four to five (43). Moreover, ECG10, ECG16, and ECG17 were found to carry an additional IS-3 element (Fig. 3). It is also evident from the hybridization experiments, using as probe a DNA fragment containing distal malK and proximal lamB sequences, that the LamB-containing fragment of ECG10 and ECG16 has been enlarged by 1.3 kb (Fig. 2). ECG17 has suffered further lamB deletion/rearrangement, which may relate to its LamB- phenotype. Because the ECG10 malK::IS-3 insertion was coincident with the LamBc phenotype, it was hypothesized that IS-3 had created the constitutive promoter for lamB. Indeed IS-3 carries a weak promoter that, when inserted in one orientation (orientation II), promotes expression of distal genes (42). To test the hypothesis that an IS-3 promoter facilitated 3IS5

S

HP H

Bg

X CO Bg

o

7

pTROY9

Table 3. Adsorption of 3H-labeled phage X to E. coli strains Carbon Cells per ml Strain 108 106 source 109 107 HB101

Glucose 90 3 0 0 Maltose 93 72 69 12 0 ECGI Glucose