Gene 94 (1990) 109-113 Elsevier
109
GENE 03704
Construction of a shuttle vector consisting of the EschericMa coil plasmid pACYCI77 inserted into the S t r e p t o m y c e s c a n l e y a phage T G I (Recombinant DNA; thienamycin; DNA cloning; deletion mutants; ff,~ostrepton resistance; kanamycin-resistance marker; bacteriophage; lysogens) Forrest Foor and Nancy Morin Merck Sharp & Dohme Research Laboratories, Rahway, NJ 0706.5 (U.S.A.) Received by K.F. Chater: I August 1989 Revised: 27 March 1990 Accepted: 20 June 1990
SUMMARY
The Escherichia coli plasmid, pACYC 177, was inserted into the single Pstl site of a deletion derivative of the Streptom~ces cattleya phage, TG 1. The hybrid molecule can be propagated as a phage in S. cattleya and as a plasmid in E. coli and is readily transferred between the two species by transfection and transformation. The kanamycin-resistance-encoding gene derived from pACYC 177 is not expressed in lysogens of the hybrid phage. Analysis of deletion mutants of the hybrid phage indicated that at least 7.5 kb ofphage DNA is dispensable. Some ofthe deletion mutants fail to lysogenize S. cattleya (Lyg- phenotype). The locations of these deletions are consistent with the location of the phage art site as previously established by Southern hybridization analysis. The thiostrepton-resistance-encoding gene derived from Streptomyces azureus was inserted into Lyg ÷ and Lyg- deletion derivatives and is expressed in S. cattleya.
INTRODUCTION
S. cattleya produces the important wide spectrum p-lactam antibiotic thienamycin (Kahan et ai., 1979). The availability of recombinant DNA techniques in this organism would be extremely useful in studies directed toward understanding the regulation and production ofthis compound. Unfortunately, none of the available Streptomyces plasmid or phage vectors (Hopwood et al., 1987) have been successfully introduced into this organism. We have described the isolation and characterization of a temperate bacteriophage (TG1) of S. cattleya having properties making it potentially useful as a DNA cloning vector (Foor and Morin, 1984; Foor et al., 1985). TGI
Correspondenceto: Dr. F. Foor, R80Y-235, Merck & Co., Inc., Rahway, NJ 07065 (U.S.A.) Tel. (201)594-6791; Fax (201)594-5468. Abbreviations: bp, base pair(s); cfu, colony-forming units; kb, kilobase(s) or 1000 bp; Kin, kanamycin; Lyg ÷/Lyg-, ability/inability of phages to form stable lysogens; pfu, plaque-forming units; R, resistance/resistant; S., Streptomyces; s, sensitive/sensitivity; Th, thiostrepton; tsr, Th R gene; wt, wild type; YC, YD, YME media, see legend to Fig. 1. 0378-1! 19/90/$03.50© 1990ElsevierScience PublishersB.V.(BiomedicalDivision)
forms stable lysogens by integration at a unique site on the chromosome and contains a double-stranded 41-kb DNA molecule with cohesive ends. A unique Pstl site was shown to be in a dispensable region of the phage genome, and a 2-kb deletion mutant (TO2) was obtained which retained this site.
EXPERIMENTAL AND DISCUSSION
(a) Construction and characterization of phage-plasmid hybrids A shuttle vector which could replicate in E. coil and S. cattleya would allow the use of the considerable genetic and biochemical techniques available in E. coli for the modification of molecules intended for use in S. cattleya. For this reason we decided to insert the E. coli plasmid pACYCI77 (Chang and Cohen, 1978), which also carries a KmR gene reported to express in S. lividans (Schottel et al., 1981), into the PstI site of TG2. The hybrids (TG6 and TGT, Fig. 1) were indistinguishable from wt by plaque morphology or ability to form iysogens. DNA isolated from the hybrid phages was transformed into E. coli by selecting
i10 for Km R. Plasmid D N A isolated from the transformants was identical to the phage D N A (as judged by digestion with Hindlll, except that the 3. l-kb fragment carrying the phage cos site could not be dissociated by heating to 72 ° C) and could be transfected with equal efficiency (typically 104-10s plaques per pg of D N A ) into protoplasts of S. cattleya. The restriction pattern of D N A from phage obtained by transfection with plasmid D N A isolated from E. coil was indistinguishable from that from the original hybrid phage. Our transfection results differ from those of Harris et al. (1983), with a ~C31 phage-plasmid hybrid, who found that D N A isolated from E. coil transfected less efficiently than D N A isolated from the phage. Thus, in spite of their large size, TG6 and T G 7 can be stably maintained an
11
I
pACYC177
II I
!
II
Km R
I I'
I
I
I
att
I
llIlll
I
---
TG6
top
LYO
o TG11
Z~ 2.0kb
,TG13 I TG8 I TG9
2.8kb 3.1 kb 4.7 kb
+ ÷ -
t TG12 TG14 TGI0
6.0kb 4.1kb 3,0kb
" " "
+
(b) Insertion of the tsr gem We decided to insert the 5. azureus tsr gene (Thompson et ai., 1982), since this marker is selectable in a wide variety of streptomycetes. A 1.8-kb B a m H l fragment carrying tsr was inserted into the Barn HI site of T G 15 and TG30. Each derivative carrying the tsr gene was tested for its ability to
pACYC177
|
i
Jl .d
,
,, ,
!
,i.i
I ]
t I _
_ _
TO7
I:==I
top
I
I !
I TG30
!TG15 lTG16
Km R /~ Lyg 2.0kb ÷ 3,1kb 4.Skb
as plasmids in E. coil and transferred back into S. cattleya as phages without restriction, rearrangement, or deletion. S. cattleya is naturally resistant to about 2 pg Km/ml. Unfortunately, there was no detectable increase in Km resistance in lysogens containing the hybrid phages, and the Km R gene is therefore only useful as a selectable marker in E. coll. To make space for a Streptomyces-selectable marker, mutants offiie hybrid phages with deletions to the left of the Km R gene were isolated (Fig. 1). Several of the deletions overlap considerably with that region of the phage D N A previously shown to contain the phage attachment (att) site (Foor et al., 1985). Deletion of the art site should result in an inability to lysogenize the host (Chater et al., 1982). The phages containing these deletions were tested for their ability to form stable lysogens of S. cattleya (Lyg + phenotype). Deletions which extended into the att region were L y g - , while those which did not extend into the region were Lyg + (Fig. 1). Although these results are consistent with deletion of the art site, deletion of a related attachment function, such as int, would also result in a L y g - phenotype. Since the deletion in T G 9 ( L y g - ) can extend at most 1.3 kb to the left of the second KpnI site from the left end of the map as drawn in Fig. 1, at least one function required for the Lyg + phenotype is located in this region. Likewise, there can be no Lyg functions more than 200 bp to the right of this site, since T G 3 0 remains Lyg + .
-
1 kb
Fig. !. Partial restriction maps of hybrids TG6 and TG7 and locations
of deletions. TG6 and TG7 were constructed by linearizing pACYCI77 with Pstl, inserting the plasmid into the Pstl site of TG2, and transfecting S. cattleya(Foor and Morin, 1984).Only the 12-kbportion at the left end of the phage restriction map is shown. The remainder of the map (indicated by the three dashes at the right end) is identical to the TGI map (Foot et al., 1985).The location of the pACYC177 DNA is indicated by the open bar immediatelybelow the map. The locations of the plasmid replication (rap) functions and the KmR gene are also indicated. The hybrids contain 4.4% (1.8 kb) more DNA than the wt and are sensitive to pyrophosphate. Deletion mutants were isolated by three cycles of killing in the presence of 25 mM Na. pyrophosphate pH 6.0, in YME. YME is YD broth (Foot et al., 1985)lacking divalent cations. Deletions located outside ofthe plasmid portion ofth ~. hybrid occur at a frequency of !% or less and were selected by transformation of E. coli RRI
(Bethesda Research Laboratories) to KmR. The approximate location and size of several such deletions are indicated by the bars below the maps of the parental phage. Phages were tested for their ability to form lysogens of S. catdeya (Ly8+ phenotype) as follows. A 10 pl aliquot of a phage lysate, typically containing 2 x 107 pfu, was added to 2 x 107 8ermlings in !.0 ml of YC broth (Foor et al., 1985) and incubated 3.5 h at 28°C Surviving cells (0.1 mi) were spread in a soft agar overlay on a YC agar plate. Samples (10/AI)containing about l0 s pfu ofphages TGI, TG28 (a virulent mutant ofTGi; Foor et al., 1985), and the phage being tested for the Lyg phenotype were spotted on the bacterial lawn. All cultures were lysed by TG28 and unaffected by TGI or the test phage. Lysis by TG28 indicates that the survivors are not merely phenotypically resistant but are resistant due to the presence in the cells of a phage in a repressed state. The survivors were allowed to sporulate, and spores were collected, washed, and tested for lysis by superinfecting phage. Cultures initially infected with Lyg+ phages remained resistant to lysis by TGI and the test phage aiter sporulation, while those infected with Lyg- phages became sensitive. Presumably Lyg- phages are lost after sporulation, because they are unable to integrate into the chromosome. The Lyg phenotype ofeach deletion mutant is indicated by + or - . The region determined by Southern hybridization analysis to contain the phage art site (Foor et al., 1985)is indicated by the open bar above each map.
I!1 transduce S. cattleya to Th R. TG20 and TG21 (Tha Lyg- derivatives of TGI5 with the tsr gene in opposite orientations) failed to transduce a nonlysogen, but did transduce a TGI lysogen, while TG38 (Fig. 2) and TG40 (Th R Lyg + derivatives of TG30 with the tsrgene in opposite orientations) transduced both cell types. The Lyg- phages presumably form double lysogens by homologous recombination between the superinfecting phage and the resident prophage, a process which is independent of the phage attachment functions. Such double lysogens should be unstable, since a reversal of this process would result in the excision and loss ofthe superinfected phage. The TG20 and TG21 double lysogens were tested for stability of the Th R phenotype by growth in the absence of Th. A single colony from a plate lacking Th was transferred to 25 ml YME, and
:'
I,
l'!
IA
a~
t I
KmR I
,
------
TG38
Th g ,!
$.$.kb deletion
EDTA
i
i
------
" all
"rG43
Tn R
BamHI.PsII linker
G~TCCGGG~.~CG:.~GT?;CC~T?GC~'GC~ GCCCG~','GC~C~CGGT,~CG
-----aft
TGS6
Th R
lkb
Fig. 2. Leftend of the restrictionmaps of TG38, TG43 and TG56. The right end of each map is identical to the TGI map (Foor et el., 1985). TG38 was constructed by inserting a 1.8-kb BamHl fragment (from plJ39, obtainedfrom D. Hopwood),whichcarries tsr and its promoter, into the BarnHl site of TG30 and transforming E. coli. TG43 was obtainedbytreatingTG38withEDTA(Fooret ell.,1985)and transducing $. cattleya to Thn. The location and amount of DIqA deleted in the formation of TG43 is indicatedby the bar belowthe map of TG38. The deletion does not removethe Pml ~ite 179bp downstreamfrom the tsr stop codon [sitenot shown;see Bibb et al. (1985)for the sequenceoftsr and surroundingDIqA].TG56was obtainedfromTG43by replacingthe remainingpACYCI77 DNA betweenthe BamHl and PstI sites with a 26-bp fragmentderivedfrom the polylinkerof the plasmid nVX (Seed, 1983). The sequence of the linker is also shown. The locations and direction of transcriptionof KmR and ThR are indicatedby the arrows.
the culture grown to saturation (equivalent to approx. 34 doublings). Portions (0.1 ml) were sporulated on 10~ FBL plates. 10% FBL consisted of 1.0g yeast extract (Difco)/l.0 g glucose/2.0 ml phosphate buffer/0.05 g MgSO4 • 7H20/20 g Bacto-Agar per liter of distilled water. The phosphate buffer consisted of 91 g KH,PO4/95g Na2HPO4 per liter ofdistilled water. After the medium was autoclaved, 0.1ml of a solution of CoCi,-6H,O (100 mg/ml)/0.1 ml of FeSO4- 7H20 (25 mg/ml)/0.1 ml of ZnSO4" 7H20 (10mg/ml) were added per liter. Spores were collected, plated for single colonies, and tested for Thg by replication. With the TG20 double lysogen 3.2~ (,3/4o4) became Th s, and with TG21 5.6~ (st/9oT) became Ths. Single iysogens of the Th R Lyg + phages TG38 and TG40 are extremely stable; more than 103 isolates of each were tested, and no Ths colonies were detected. Comparable results to these were obtained in similar experiments with single and double lysogens of S. coelicolor using q~C31 derivatives (Chater et el., 1982). The presence of the marker allowed us to determine the frequency of transduction of S. cattleya as a function of germination time (Fig. 3). The time at which cells were most susceptible to lysogeny corresponded to the emergence of the germ tube at 3 h. The number of Th ~ transductants decreased rapidly thereafter. At 4 h no mycelial branches were observed, while at 5 h about 90~ ofthe germlings had at least one branch. At this point no Th R transductants were obtained. The susceptibility of the germlings to lysis was also restricted to this same time period (see the curve in Fig. 3 showing total cfu in the infected samples), in separate experiments the yield of phage was similarly dependent on the germination time (data not shown). The shortness of the time period during which germlings were effectively transduced is unusual and is important to keep in mind during cloning experiments requiring transduction of a Streptomyces host. Since both lyric growth and lysogeny were affected in the same way, it is possible that the surface receptor for the phage is developmentally regulated and is only present during this time period. The presence of the drug marker also allowed us to compare the host range of the phage as previously determined by plaque formation to that measurable by transduction. The results (Table I) show that most strains supporting the formation of plaques by TGI, also yield The iysogens with TG38, S. gr/seus being the only exception. It may be that this strain lacks the phage integration site, or that the additional DNA present in TG38 has a negative effect, possibly rendering it more susceptible to restriction by this particular host. Two strains, S. echinatus and S. r#nosus, yielded substantial numbers of ThR colonies without supporting the formation of plaques. The phage may be useful as a cloning vector in strains such as these as well.
112 lo 9
,
¢
loe
TABLE I
Total (uninfected)
Formation of plaques by TGI and of TG38 Th R lysogens with various Streptomyces spp.
ld ~
10a
t
e
d
)
10s ,~, ThR (infected)
104 ld
/' olgeritube, ~!beg,ns /
emergence
~
branching
ld 0
1
2
3
4
5
6
7
Strain
Strain No. a
EOP b
Th R colonies e
S. cattleya S. avermitilis S. coelicolor S. echinatus S. fradiae S. grieseolus S. griseus $. !~vendulae S. lividans S. phaeochromogenes S. rimosus S. venezuelae S. viridochromogenes
MA4297 NRRL8165 A3(2) NRRL2587 MA 194 ATCC11796 ATCCI0137 MA330 1326 MAI93 NRRL2234 ATCC14585 NRRL3414
(1) 1 < 10- s < 10 -9c 10 - 2 10- ~ 10- 3 10- 3 < 10 -9 I0- 4 < I0 -9~ 10- 2 10- $d
516 140 0 545 20 93 0 22 0 15 437 40 438
Germination time (h) Fig. 3. Transduction of 5. cattleya as a function of germination time. Spores were suspended in YD at a concentration of 2 x 107 spores per ml and germinated at 3"/°C. At the indicated times two 1.0-ml samples of the culture were removed. One sample was infected with 10#1 (2 × 10? pfu) ofa TG38 lysate, while the other was not. The morphological state of the germlings (presence of germ tubes or mycelial branches) was determined microscopically. Samples were incubated for 2 h at 28°C to allow expression of tar. The cfu were determined by serial dilution and plating of each sample on YME supplemented with 5 mM Na3' citrate, with and without Th (Squibb, 10/Ag/ml). The results are plotted as the concentration of total (ThR+ Th s) cfu in the infected (circles) and uninfected (squares) samples and of Th R cfu in the infected sample (triangles).
To allow more space for the cloning of foreign DNA the pACYCI77 DNA to the left ofthe tsrgene was deleted from TG38 by treatment with EDTA to produce TG43 (Fig. 2). The remaining pACYC177 DNA in TG43 (to the right of tsr) was replaced with a 26-bp linker. Unexpectedly, this derivative (TG56, Fig. 2) transduced S. cattleya to Th R at a 1000-fold lower frequency. The resulting lysogens grew significantly more slowly in the presence of Th, when compared to lysogens of TG38 or TG43, suggesting that high level expression of the tsr gene in the latter phages is dependent upon sequences (possibly a strong promoter) present in this region of the pACYC177 DNA. The dependence of tsr expression on sequences upstream from the normal tsr promoter could prevent use of the upstream BamHI site as a cloning site, since insertions may significantly reduce the frequency of transduction. (c) Conclusions
(1) The results presented in this paper demonstrate the suitability of TG 1 and its derivatives as cloning vectors in the thienamycin-producer S. cattleya. The insertion of the E. coil plasmid pACYCI77 and of the S. azureus tsr gone does not affect phage growth or the ability to form lysogens.
" Strain number prefixes: MA, Merck & Co., Inc.; NRRL, Agricultural Research Culture Collection, Northern Regional Research Center, USDA, Peoria, IL; ATCC, American Type Culture Collection, Rockville, MD. $. coeficolor A3(2) and $. iividans 1326 were obtained from D. Hopwood, John Innes Institute, Norwich, England. b EOP is the titer of TGl on the test strain divided by that on $. cattleya. Results are from Foot et al. (1985), except those for $, avermitilis, $. coelicolor, $.fradfae, and $. phaeochromo&enes, which have not previously been published. c Spot test of a high titer lysate on a lawn seeded with spores showed a slight inhibition of cell growth. d Very small plaques. o The number of Th R colonies obtained from 0.2 ml of a culture transduced with TG38. Spores of the strain to be transduced were suspended in 1.0 ml YD at a concentration of 2 x 107 spores per ml and germinated for 3.0 h at 37°C. A 10-/d aliquot of a phage lysate, containing 2 x 107 pfu, was added to the germlings, and the mixture incubated for 3.5 h at 28°C. A portion (0.2 ml) was plated on minimal medium (Foor etal., 1982) with and without Th. $. coelicolor and 8. lavendulae were plated on R2YE (Thompson et ai., 1980) with and without Th and 5. avermitilis on NK-1 with and without Th. NK-1 contains per liter 1.75 g K2HPO4/0.75 g KH2PO4/I.0 g (NH4)2SO4/0.5 g Na3" citrate/0.2$ g MgSO4.7HaO/l.2 g Na.glutamate/5 ml trace elements (Foor and Morin, 1984), and is adjusted to pH 7.0 with NaOH. ARer autoclaving 50 ml of 40% glucose is added. To confirm that the Th Rcolonies were true lysogens representative isolates were purified and tested for phage release and immunity to iysis by TG38.
These results are similar to those obtained by Suarez and Chater (1980) with pBR322 inserted into the Streptomyces phage ~C31. In addition, Streicher (1983) inserted the S. cattleya argG gene into the PstI site of TG1 and TGI2 and was able to complement Arg- mutants of S. catdeya. (2) Previous results (Foot etal., 1985) with deletion mutants of TGI showed a minimum of 3.8 kb of dispensable DNA in the vicinity of the single Pstl site of the TGI genome. The present results with deletion mutants of the hybrids TG6 and TG7 extend this region to a minimum of 8.2 kb of dispensable DNA. The wt (TGI) genome is about 41.2 kb in length. The shortest genome (TGI2) is
113 36.0 kb and the longest (TG6, TG7), 43.0 kb. Thus the packaging range for viable phage extends at least from 87-104% of the wt genome length. These results are very similar to those obtained with ~C31 (Chater et al., 1981a). (3) The au site was previously shown by Southern hybridization analysis of lysogen DNA to map in a region extending from 1.6-4.'/kb from that end ofthe phage DNA nearest the single Pstl site (Foor et al., 1985). The analysis in this paper of the ability of various deletion mutants to form stable lysogens is in agreement with those results and further limits the art site to a region from 1.6-4.0 kb from the end of the DNA. The f'mding that the tsr derivatives of TGIS, but not TG30, fail to give drug-resistant transductants also supports these results. Lyg- phages can integrate in the presence of a resident prophage, forming double lysogens which are stable enough to be useful for the expression and maintenance of cloned genes. Cloning of chromosomal DNA into a Th R Lyg- phage should allow stable insertion at homologous sites in the bacterial genome for complementation or gene disruption analysis. (4) The locations ofthe an site and the dispensable DNA relative to the end of the phage DNA, as well as the length of the phage genome, are very similar to those observed for ¢C31 (Chater et al., 198 la,b; 1982). Although their restriction maps are dissimilar, these structural similarities suggest that the two phages are related. Southern hybridization analysis and antigenic cross-reactivity studies show that this is indeed the case (N.M. and F.F., unpublished observations). (5) Several of the phage derivatives described in this paper ate potentially useful as vectors for the cloning of DNA. For example, the Lyg ÷ phage TG8 has space for 3. l kb of DNA and single sites for Pstl, SstIl, and Xhol. The Lyg ÷ phage TG30 and the Lyg- phage TGI6 could be used as replacement vectors for cloning with Pstl. Removal of the pACYC177 with Pstl would add 3.85 kb of additional space for foreign DNA, giving room for a total ofS.9 kb of DNA in the case of TG30 and about 8.5 kb of DNA in the case ofTGl6. TG43 has space for about 4 kb of DNA and both Pstl and BamHl cloning sites. The availability of these vectors makes it possible to clone DNA in this organism with a variety of techniques.
ACKNOWLEDGEMENTS
We thank Keith Chater, David Hopwood, and Brian Seed for helpful discussions and for providing us with plasmids, Stanley Cohen for providing pACYC177, Doug MacNeil, Tanya MacNeil, and Stanley Streicher for many helpful conversations, and Sara Currie for providing strains.
REFERENCES Bibb, MJ., Bibb, MJ., Ward, J.M. and Cohen, S.N.: Nucleotide sequences encoding and promoting expression of three antibiotic resistance genes indigenous to Streptom)~ces. Mol. Gen. Genes. 199 (1985) 26-36. Chang, A.C.Y. and Cohen, S.N.: Construction and characterization of amplifiable multicopy DNA cloning vehicles derived from the PiSA cryptic miniplasmid. J. Bacteriol. 134 (1978) i 141-1156. Chater, K.F., Bruton, C3., King, A.A. and Suarez, J.E.: The expression of Stre,vtomyces and Escherichia coil drug-resistance determinants cloned into the Sweptomyces phage ~ 3 1 . Gene 19 (1982) 2i-3~ Chater, K.F., Bruton, C3., Springer, W. and Suarez, J.F_: Dispensable sequences and packaging constraints of DNA from the S~r~vtomyces temperate phage ~ 3 1 . Gene 15 (1981a) 249-256. Chater, K.F., Bruton, CJ. and Suarez, J.E.: Restriction mapping of the DNA of the Streptomyces temperate phage ~ 3 1 and its derivatives. Gene 14 (1981b) 183-194. Foor, F. and Morin, N.: DNA cloning vector TGI, derivatives, and processes of making. U.S. Patent 4,460,689. July 17, 1984. Foot, F., Tyler, B. and Morin, N.: Effect ofglycerol, phosphate, and pH on cellular differentiation and production of the ~-Iactam antibiotic thienamycin by Streptomyces cauleya. Develop. Ind. Microbiol. 23 (1982) 305-314. Foor, F., Roberts, G.P., Morin, N., Snyder, L., Hwang, M., Gibbons, P.H., Paradise, MJ., Stotish, R.L, Ruby, C.L, Wolanski, B. and Streicher, S.L.: Isolation and characterization of the Streptomyces cattleya temperate phage TGI. Gene 39 0985) ! 1-16. Harris, J.E., Chater, K.F., Bruton, CJ. and Piret, J.M.: The restriction mapping ofc gene deletions in Streptomycesbacteriophage ¢~31 and their use in cloning vector development. Gene 22 (1983) 16"/-174. Hopwood, D.A., Bibb, M.J., Chater, K.F. and Kieser, T.: Plasmid and phage vectors for gene cloning and analysis in Streptomyces.Methods Enzymol. 153 (1987) 116-166. Kahan, J.S., Kahan, F.M., Goegelman, R., Currie, S.A., Jackson, M., Stapley, E.O., Miller, T.W., Miller, A.K., Hendlin, D., Mochales, S., Hernandez, S., Woodruff, H.B. and Birnbaum, J.: Thienamycin, a new p-lactam antibiotic, !. Discovery, taxonomy, isolation and physical properties. J. Antibiot. 32 (1979) 1-12. Seed, B.: Purification ofgenomic sequences from bacteriophage libraries by recombination and selection in rive. Nucleic Acids Res. i I (1983) 242"/-2445. Schottel, J.L, Bibb, M.J. and Cohen, S.N.: Cloning and expression in Streptomyces lividans of antibiotic resistance genes derived from Escherichia coll. J. Bacteriol. 146 (1981) 360-368. Streicber, S.L.: Cloning and expression of the argG gene from $treptomyces cattleya in Escherichia coil and $. cattleya. Abstracts of the Annual Meeting oftbe American Society for Microbiology. American Society for Microbiology, Washington, DC, 1983, p. 134. Suarez, J.E. and Chater, K.F.: DNA cloning in $treptomyces: a bifunctional replicon comprising pBR322 inserted into a $treptomyces phage. Nature 286 (1980) 52"/-529. Thompson, C.J., Ward, J.M. and Hopwood, D.A.: DNA cloning in Streptomyces: resistance genes from antibiotic-producing species. Nature 286 (1980) 525-52"/. Thompson, C.J., Kieser, T., Ward, J.M. and Hopwood, D.A.: Physical analysis ofantibiotic-resistance genes from Streptomyces and their use in vector construction. Gene 20 (1982) 5 !-62.
109
GENE 03704
Construction of a shuttle vector consisting of the EschericMa coil plasmid pACYCI77 inserted into the S t r e p t o m y c e s c a n l e y a phage T G I (Recombinant DNA; thienamycin; DNA cloning; deletion mutants; ff,~ostrepton resistance; kanamycin-resistance marker; bacteriophage; lysogens) Forrest Foor and Nancy Morin Merck Sharp & Dohme Research Laboratories, Rahway, NJ 0706.5 (U.S.A.) Received by K.F. Chater: I August 1989 Revised: 27 March 1990 Accepted: 20 June 1990
SUMMARY
The Escherichia coli plasmid, pACYC 177, was inserted into the single Pstl site of a deletion derivative of the Streptom~ces cattleya phage, TG 1. The hybrid molecule can be propagated as a phage in S. cattleya and as a plasmid in E. coli and is readily transferred between the two species by transfection and transformation. The kanamycin-resistance-encoding gene derived from pACYC 177 is not expressed in lysogens of the hybrid phage. Analysis of deletion mutants of the hybrid phage indicated that at least 7.5 kb ofphage DNA is dispensable. Some ofthe deletion mutants fail to lysogenize S. cattleya (Lyg- phenotype). The locations of these deletions are consistent with the location of the phage art site as previously established by Southern hybridization analysis. The thiostrepton-resistance-encoding gene derived from Streptomyces azureus was inserted into Lyg ÷ and Lyg- deletion derivatives and is expressed in S. cattleya.
INTRODUCTION
S. cattleya produces the important wide spectrum p-lactam antibiotic thienamycin (Kahan et ai., 1979). The availability of recombinant DNA techniques in this organism would be extremely useful in studies directed toward understanding the regulation and production ofthis compound. Unfortunately, none of the available Streptomyces plasmid or phage vectors (Hopwood et al., 1987) have been successfully introduced into this organism. We have described the isolation and characterization of a temperate bacteriophage (TG1) of S. cattleya having properties making it potentially useful as a DNA cloning vector (Foor and Morin, 1984; Foor et al., 1985). TGI
Correspondenceto: Dr. F. Foor, R80Y-235, Merck & Co., Inc., Rahway, NJ 07065 (U.S.A.) Tel. (201)594-6791; Fax (201)594-5468. Abbreviations: bp, base pair(s); cfu, colony-forming units; kb, kilobase(s) or 1000 bp; Kin, kanamycin; Lyg ÷/Lyg-, ability/inability of phages to form stable lysogens; pfu, plaque-forming units; R, resistance/resistant; S., Streptomyces; s, sensitive/sensitivity; Th, thiostrepton; tsr, Th R gene; wt, wild type; YC, YD, YME media, see legend to Fig. 1. 0378-1! 19/90/$03.50© 1990ElsevierScience PublishersB.V.(BiomedicalDivision)
forms stable lysogens by integration at a unique site on the chromosome and contains a double-stranded 41-kb DNA molecule with cohesive ends. A unique Pstl site was shown to be in a dispensable region of the phage genome, and a 2-kb deletion mutant (TO2) was obtained which retained this site.
EXPERIMENTAL AND DISCUSSION
(a) Construction and characterization of phage-plasmid hybrids A shuttle vector which could replicate in E. coil and S. cattleya would allow the use of the considerable genetic and biochemical techniques available in E. coli for the modification of molecules intended for use in S. cattleya. For this reason we decided to insert the E. coli plasmid pACYCI77 (Chang and Cohen, 1978), which also carries a KmR gene reported to express in S. lividans (Schottel et al., 1981), into the PstI site of TG2. The hybrids (TG6 and TGT, Fig. 1) were indistinguishable from wt by plaque morphology or ability to form iysogens. DNA isolated from the hybrid phages was transformed into E. coli by selecting
i10 for Km R. Plasmid D N A isolated from the transformants was identical to the phage D N A (as judged by digestion with Hindlll, except that the 3. l-kb fragment carrying the phage cos site could not be dissociated by heating to 72 ° C) and could be transfected with equal efficiency (typically 104-10s plaques per pg of D N A ) into protoplasts of S. cattleya. The restriction pattern of D N A from phage obtained by transfection with plasmid D N A isolated from E. coil was indistinguishable from that from the original hybrid phage. Our transfection results differ from those of Harris et al. (1983), with a ~C31 phage-plasmid hybrid, who found that D N A isolated from E. coil transfected less efficiently than D N A isolated from the phage. Thus, in spite of their large size, TG6 and T G 7 can be stably maintained an
11
I
pACYC177
II I
!
II
Km R
I I'
I
I
I
att
I
llIlll
I
---
TG6
top
LYO
o TG11
Z~ 2.0kb
,TG13 I TG8 I TG9
2.8kb 3.1 kb 4.7 kb
+ ÷ -
t TG12 TG14 TGI0
6.0kb 4.1kb 3,0kb
" " "
+
(b) Insertion of the tsr gem We decided to insert the 5. azureus tsr gene (Thompson et ai., 1982), since this marker is selectable in a wide variety of streptomycetes. A 1.8-kb B a m H l fragment carrying tsr was inserted into the Barn HI site of T G 15 and TG30. Each derivative carrying the tsr gene was tested for its ability to
pACYC177
|
i
Jl .d
,
,, ,
!
,i.i
I ]
t I _
_ _
TO7
I:==I
top
I
I !
I TG30
!TG15 lTG16
Km R /~ Lyg 2.0kb ÷ 3,1kb 4.Skb
as plasmids in E. coil and transferred back into S. cattleya as phages without restriction, rearrangement, or deletion. S. cattleya is naturally resistant to about 2 pg Km/ml. Unfortunately, there was no detectable increase in Km resistance in lysogens containing the hybrid phages, and the Km R gene is therefore only useful as a selectable marker in E. coll. To make space for a Streptomyces-selectable marker, mutants offiie hybrid phages with deletions to the left of the Km R gene were isolated (Fig. 1). Several of the deletions overlap considerably with that region of the phage D N A previously shown to contain the phage attachment (att) site (Foor et al., 1985). Deletion of the art site should result in an inability to lysogenize the host (Chater et al., 1982). The phages containing these deletions were tested for their ability to form stable lysogens of S. cattleya (Lyg + phenotype). Deletions which extended into the att region were L y g - , while those which did not extend into the region were Lyg + (Fig. 1). Although these results are consistent with deletion of the art site, deletion of a related attachment function, such as int, would also result in a L y g - phenotype. Since the deletion in T G 9 ( L y g - ) can extend at most 1.3 kb to the left of the second KpnI site from the left end of the map as drawn in Fig. 1, at least one function required for the Lyg + phenotype is located in this region. Likewise, there can be no Lyg functions more than 200 bp to the right of this site, since T G 3 0 remains Lyg + .
-
1 kb
Fig. !. Partial restriction maps of hybrids TG6 and TG7 and locations
of deletions. TG6 and TG7 were constructed by linearizing pACYCI77 with Pstl, inserting the plasmid into the Pstl site of TG2, and transfecting S. cattleya(Foor and Morin, 1984).Only the 12-kbportion at the left end of the phage restriction map is shown. The remainder of the map (indicated by the three dashes at the right end) is identical to the TGI map (Foot et al., 1985).The location of the pACYC177 DNA is indicated by the open bar immediatelybelow the map. The locations of the plasmid replication (rap) functions and the KmR gene are also indicated. The hybrids contain 4.4% (1.8 kb) more DNA than the wt and are sensitive to pyrophosphate. Deletion mutants were isolated by three cycles of killing in the presence of 25 mM Na. pyrophosphate pH 6.0, in YME. YME is YD broth (Foot et al., 1985)lacking divalent cations. Deletions located outside ofthe plasmid portion ofth ~. hybrid occur at a frequency of !% or less and were selected by transformation of E. coli RRI
(Bethesda Research Laboratories) to KmR. The approximate location and size of several such deletions are indicated by the bars below the maps of the parental phage. Phages were tested for their ability to form lysogens of S. catdeya (Ly8+ phenotype) as follows. A 10 pl aliquot of a phage lysate, typically containing 2 x 107 pfu, was added to 2 x 107 8ermlings in !.0 ml of YC broth (Foor et al., 1985) and incubated 3.5 h at 28°C Surviving cells (0.1 mi) were spread in a soft agar overlay on a YC agar plate. Samples (10/AI)containing about l0 s pfu ofphages TGI, TG28 (a virulent mutant ofTGi; Foor et al., 1985), and the phage being tested for the Lyg phenotype were spotted on the bacterial lawn. All cultures were lysed by TG28 and unaffected by TGI or the test phage. Lysis by TG28 indicates that the survivors are not merely phenotypically resistant but are resistant due to the presence in the cells of a phage in a repressed state. The survivors were allowed to sporulate, and spores were collected, washed, and tested for lysis by superinfecting phage. Cultures initially infected with Lyg+ phages remained resistant to lysis by TGI and the test phage aiter sporulation, while those infected with Lyg- phages became sensitive. Presumably Lyg- phages are lost after sporulation, because they are unable to integrate into the chromosome. The Lyg phenotype ofeach deletion mutant is indicated by + or - . The region determined by Southern hybridization analysis to contain the phage art site (Foor et al., 1985)is indicated by the open bar above each map.
I!1 transduce S. cattleya to Th R. TG20 and TG21 (Tha Lyg- derivatives of TGI5 with the tsr gene in opposite orientations) failed to transduce a nonlysogen, but did transduce a TGI lysogen, while TG38 (Fig. 2) and TG40 (Th R Lyg + derivatives of TG30 with the tsrgene in opposite orientations) transduced both cell types. The Lyg- phages presumably form double lysogens by homologous recombination between the superinfecting phage and the resident prophage, a process which is independent of the phage attachment functions. Such double lysogens should be unstable, since a reversal of this process would result in the excision and loss ofthe superinfected phage. The TG20 and TG21 double lysogens were tested for stability of the Th R phenotype by growth in the absence of Th. A single colony from a plate lacking Th was transferred to 25 ml YME, and
:'
I,
l'!
IA
a~
t I
KmR I
,
------
TG38
Th g ,!
$.$.kb deletion
EDTA
i
i
------
" all
"rG43
Tn R
BamHI.PsII linker
G~TCCGGG~.~CG:.~GT?;CC~T?GC~'GC~ GCCCG~','GC~C~CGGT,~CG
-----aft
TGS6
Th R
lkb
Fig. 2. Leftend of the restrictionmaps of TG38, TG43 and TG56. The right end of each map is identical to the TGI map (Foor et el., 1985). TG38 was constructed by inserting a 1.8-kb BamHl fragment (from plJ39, obtainedfrom D. Hopwood),whichcarries tsr and its promoter, into the BarnHl site of TG30 and transforming E. coli. TG43 was obtainedbytreatingTG38withEDTA(Fooret ell.,1985)and transducing $. cattleya to Thn. The location and amount of DIqA deleted in the formation of TG43 is indicatedby the bar belowthe map of TG38. The deletion does not removethe Pml ~ite 179bp downstreamfrom the tsr stop codon [sitenot shown;see Bibb et al. (1985)for the sequenceoftsr and surroundingDIqA].TG56was obtainedfromTG43by replacingthe remainingpACYCI77 DNA betweenthe BamHl and PstI sites with a 26-bp fragmentderivedfrom the polylinkerof the plasmid nVX (Seed, 1983). The sequence of the linker is also shown. The locations and direction of transcriptionof KmR and ThR are indicatedby the arrows.
the culture grown to saturation (equivalent to approx. 34 doublings). Portions (0.1 ml) were sporulated on 10~ FBL plates. 10% FBL consisted of 1.0g yeast extract (Difco)/l.0 g glucose/2.0 ml phosphate buffer/0.05 g MgSO4 • 7H20/20 g Bacto-Agar per liter of distilled water. The phosphate buffer consisted of 91 g KH,PO4/95g Na2HPO4 per liter ofdistilled water. After the medium was autoclaved, 0.1ml of a solution of CoCi,-6H,O (100 mg/ml)/0.1 ml of FeSO4- 7H20 (25 mg/ml)/0.1 ml of ZnSO4" 7H20 (10mg/ml) were added per liter. Spores were collected, plated for single colonies, and tested for Thg by replication. With the TG20 double lysogen 3.2~ (,3/4o4) became Th s, and with TG21 5.6~ (st/9oT) became Ths. Single iysogens of the Th R Lyg + phages TG38 and TG40 are extremely stable; more than 103 isolates of each were tested, and no Ths colonies were detected. Comparable results to these were obtained in similar experiments with single and double lysogens of S. coelicolor using q~C31 derivatives (Chater et el., 1982). The presence of the marker allowed us to determine the frequency of transduction of S. cattleya as a function of germination time (Fig. 3). The time at which cells were most susceptible to lysogeny corresponded to the emergence of the germ tube at 3 h. The number of Th ~ transductants decreased rapidly thereafter. At 4 h no mycelial branches were observed, while at 5 h about 90~ ofthe germlings had at least one branch. At this point no Th R transductants were obtained. The susceptibility of the germlings to lysis was also restricted to this same time period (see the curve in Fig. 3 showing total cfu in the infected samples), in separate experiments the yield of phage was similarly dependent on the germination time (data not shown). The shortness of the time period during which germlings were effectively transduced is unusual and is important to keep in mind during cloning experiments requiring transduction of a Streptomyces host. Since both lyric growth and lysogeny were affected in the same way, it is possible that the surface receptor for the phage is developmentally regulated and is only present during this time period. The presence of the drug marker also allowed us to compare the host range of the phage as previously determined by plaque formation to that measurable by transduction. The results (Table I) show that most strains supporting the formation of plaques by TGI, also yield The iysogens with TG38, S. gr/seus being the only exception. It may be that this strain lacks the phage integration site, or that the additional DNA present in TG38 has a negative effect, possibly rendering it more susceptible to restriction by this particular host. Two strains, S. echinatus and S. r#nosus, yielded substantial numbers of ThR colonies without supporting the formation of plaques. The phage may be useful as a cloning vector in strains such as these as well.
112 lo 9
,
¢
loe
TABLE I
Total (uninfected)
Formation of plaques by TGI and of TG38 Th R lysogens with various Streptomyces spp.
ld ~
10a
t
e
d
)
10s ,~, ThR (infected)
104 ld
/' olgeritube, ~!beg,ns /
emergence
~
branching
ld 0
1
2
3
4
5
6
7
Strain
Strain No. a
EOP b
Th R colonies e
S. cattleya S. avermitilis S. coelicolor S. echinatus S. fradiae S. grieseolus S. griseus $. !~vendulae S. lividans S. phaeochromogenes S. rimosus S. venezuelae S. viridochromogenes
MA4297 NRRL8165 A3(2) NRRL2587 MA 194 ATCC11796 ATCCI0137 MA330 1326 MAI93 NRRL2234 ATCC14585 NRRL3414
(1) 1 < 10- s < 10 -9c 10 - 2 10- ~ 10- 3 10- 3 < 10 -9 I0- 4 < I0 -9~ 10- 2 10- $d
516 140 0 545 20 93 0 22 0 15 437 40 438
Germination time (h) Fig. 3. Transduction of 5. cattleya as a function of germination time. Spores were suspended in YD at a concentration of 2 x 107 spores per ml and germinated at 3"/°C. At the indicated times two 1.0-ml samples of the culture were removed. One sample was infected with 10#1 (2 × 10? pfu) ofa TG38 lysate, while the other was not. The morphological state of the germlings (presence of germ tubes or mycelial branches) was determined microscopically. Samples were incubated for 2 h at 28°C to allow expression of tar. The cfu were determined by serial dilution and plating of each sample on YME supplemented with 5 mM Na3' citrate, with and without Th (Squibb, 10/Ag/ml). The results are plotted as the concentration of total (ThR+ Th s) cfu in the infected (circles) and uninfected (squares) samples and of Th R cfu in the infected sample (triangles).
To allow more space for the cloning of foreign DNA the pACYCI77 DNA to the left ofthe tsrgene was deleted from TG38 by treatment with EDTA to produce TG43 (Fig. 2). The remaining pACYC177 DNA in TG43 (to the right of tsr) was replaced with a 26-bp linker. Unexpectedly, this derivative (TG56, Fig. 2) transduced S. cattleya to Th R at a 1000-fold lower frequency. The resulting lysogens grew significantly more slowly in the presence of Th, when compared to lysogens of TG38 or TG43, suggesting that high level expression of the tsr gene in the latter phages is dependent upon sequences (possibly a strong promoter) present in this region of the pACYC177 DNA. The dependence of tsr expression on sequences upstream from the normal tsr promoter could prevent use of the upstream BamHI site as a cloning site, since insertions may significantly reduce the frequency of transduction. (c) Conclusions
(1) The results presented in this paper demonstrate the suitability of TG 1 and its derivatives as cloning vectors in the thienamycin-producer S. cattleya. The insertion of the E. coil plasmid pACYCI77 and of the S. azureus tsr gone does not affect phage growth or the ability to form lysogens.
" Strain number prefixes: MA, Merck & Co., Inc.; NRRL, Agricultural Research Culture Collection, Northern Regional Research Center, USDA, Peoria, IL; ATCC, American Type Culture Collection, Rockville, MD. $. coeficolor A3(2) and $. iividans 1326 were obtained from D. Hopwood, John Innes Institute, Norwich, England. b EOP is the titer of TGl on the test strain divided by that on $. cattleya. Results are from Foot et al. (1985), except those for $, avermitilis, $. coelicolor, $.fradfae, and $. phaeochromo&enes, which have not previously been published. c Spot test of a high titer lysate on a lawn seeded with spores showed a slight inhibition of cell growth. d Very small plaques. o The number of Th R colonies obtained from 0.2 ml of a culture transduced with TG38. Spores of the strain to be transduced were suspended in 1.0 ml YD at a concentration of 2 x 107 spores per ml and germinated for 3.0 h at 37°C. A 10-/d aliquot of a phage lysate, containing 2 x 107 pfu, was added to the germlings, and the mixture incubated for 3.5 h at 28°C. A portion (0.2 ml) was plated on minimal medium (Foor etal., 1982) with and without Th. $. coelicolor and 8. lavendulae were plated on R2YE (Thompson et ai., 1980) with and without Th and 5. avermitilis on NK-1 with and without Th. NK-1 contains per liter 1.75 g K2HPO4/0.75 g KH2PO4/I.0 g (NH4)2SO4/0.5 g Na3" citrate/0.2$ g MgSO4.7HaO/l.2 g Na.glutamate/5 ml trace elements (Foor and Morin, 1984), and is adjusted to pH 7.0 with NaOH. ARer autoclaving 50 ml of 40% glucose is added. To confirm that the Th Rcolonies were true lysogens representative isolates were purified and tested for phage release and immunity to iysis by TG38.
These results are similar to those obtained by Suarez and Chater (1980) with pBR322 inserted into the Streptomyces phage ~C31. In addition, Streicher (1983) inserted the S. cattleya argG gene into the PstI site of TG1 and TGI2 and was able to complement Arg- mutants of S. catdeya. (2) Previous results (Foot etal., 1985) with deletion mutants of TGI showed a minimum of 3.8 kb of dispensable DNA in the vicinity of the single Pstl site of the TGI genome. The present results with deletion mutants of the hybrids TG6 and TG7 extend this region to a minimum of 8.2 kb of dispensable DNA. The wt (TGI) genome is about 41.2 kb in length. The shortest genome (TGI2) is
113 36.0 kb and the longest (TG6, TG7), 43.0 kb. Thus the packaging range for viable phage extends at least from 87-104% of the wt genome length. These results are very similar to those obtained with ~C31 (Chater et al., 1981a). (3) The au site was previously shown by Southern hybridization analysis of lysogen DNA to map in a region extending from 1.6-4.'/kb from that end ofthe phage DNA nearest the single Pstl site (Foor et al., 1985). The analysis in this paper of the ability of various deletion mutants to form stable lysogens is in agreement with those results and further limits the art site to a region from 1.6-4.0 kb from the end of the DNA. The f'mding that the tsr derivatives of TGIS, but not TG30, fail to give drug-resistant transductants also supports these results. Lyg- phages can integrate in the presence of a resident prophage, forming double lysogens which are stable enough to be useful for the expression and maintenance of cloned genes. Cloning of chromosomal DNA into a Th R Lyg- phage should allow stable insertion at homologous sites in the bacterial genome for complementation or gene disruption analysis. (4) The locations ofthe an site and the dispensable DNA relative to the end of the phage DNA, as well as the length of the phage genome, are very similar to those observed for ¢C31 (Chater et al., 198 la,b; 1982). Although their restriction maps are dissimilar, these structural similarities suggest that the two phages are related. Southern hybridization analysis and antigenic cross-reactivity studies show that this is indeed the case (N.M. and F.F., unpublished observations). (5) Several of the phage derivatives described in this paper ate potentially useful as vectors for the cloning of DNA. For example, the Lyg ÷ phage TG8 has space for 3. l kb of DNA and single sites for Pstl, SstIl, and Xhol. The Lyg ÷ phage TG30 and the Lyg- phage TGI6 could be used as replacement vectors for cloning with Pstl. Removal of the pACYC177 with Pstl would add 3.85 kb of additional space for foreign DNA, giving room for a total ofS.9 kb of DNA in the case of TG30 and about 8.5 kb of DNA in the case ofTGl6. TG43 has space for about 4 kb of DNA and both Pstl and BamHl cloning sites. The availability of these vectors makes it possible to clone DNA in this organism with a variety of techniques.
ACKNOWLEDGEMENTS
We thank Keith Chater, David Hopwood, and Brian Seed for helpful discussions and for providing us with plasmids, Stanley Cohen for providing pACYC177, Doug MacNeil, Tanya MacNeil, and Stanley Streicher for many helpful conversations, and Sara Currie for providing strains.
REFERENCES Bibb, MJ., Bibb, MJ., Ward, J.M. and Cohen, S.N.: Nucleotide sequences encoding and promoting expression of three antibiotic resistance genes indigenous to Streptom)~ces. Mol. Gen. Genes. 199 (1985) 26-36. Chang, A.C.Y. and Cohen, S.N.: Construction and characterization of amplifiable multicopy DNA cloning vehicles derived from the PiSA cryptic miniplasmid. J. Bacteriol. 134 (1978) i 141-1156. Chater, K.F., Bruton, C3., King, A.A. and Suarez, J.E.: The expression of Stre,vtomyces and Escherichia coil drug-resistance determinants cloned into the Sweptomyces phage ~ 3 1 . Gene 19 (1982) 2i-3~ Chater, K.F., Bruton, C3., Springer, W. and Suarez, J.F_: Dispensable sequences and packaging constraints of DNA from the S~r~vtomyces temperate phage ~ 3 1 . Gene 15 (1981a) 249-256. Chater, K.F., Bruton, CJ. and Suarez, J.E.: Restriction mapping of the DNA of the Streptomyces temperate phage ~ 3 1 and its derivatives. Gene 14 (1981b) 183-194. Foor, F. and Morin, N.: DNA cloning vector TGI, derivatives, and processes of making. U.S. Patent 4,460,689. July 17, 1984. Foot, F., Tyler, B. and Morin, N.: Effect ofglycerol, phosphate, and pH on cellular differentiation and production of the ~-Iactam antibiotic thienamycin by Streptomyces cauleya. Develop. Ind. Microbiol. 23 (1982) 305-314. Foor, F., Roberts, G.P., Morin, N., Snyder, L., Hwang, M., Gibbons, P.H., Paradise, MJ., Stotish, R.L, Ruby, C.L, Wolanski, B. and Streicher, S.L.: Isolation and characterization of the Streptomyces cattleya temperate phage TGI. Gene 39 0985) ! 1-16. Harris, J.E., Chater, K.F., Bruton, CJ. and Piret, J.M.: The restriction mapping ofc gene deletions in Streptomycesbacteriophage ¢~31 and their use in cloning vector development. Gene 22 (1983) 16"/-174. Hopwood, D.A., Bibb, M.J., Chater, K.F. and Kieser, T.: Plasmid and phage vectors for gene cloning and analysis in Streptomyces.Methods Enzymol. 153 (1987) 116-166. Kahan, J.S., Kahan, F.M., Goegelman, R., Currie, S.A., Jackson, M., Stapley, E.O., Miller, T.W., Miller, A.K., Hendlin, D., Mochales, S., Hernandez, S., Woodruff, H.B. and Birnbaum, J.: Thienamycin, a new p-lactam antibiotic, !. Discovery, taxonomy, isolation and physical properties. J. Antibiot. 32 (1979) 1-12. Seed, B.: Purification ofgenomic sequences from bacteriophage libraries by recombination and selection in rive. Nucleic Acids Res. i I (1983) 242"/-2445. Schottel, J.L, Bibb, M.J. and Cohen, S.N.: Cloning and expression in Streptomyces lividans of antibiotic resistance genes derived from Escherichia coll. J. Bacteriol. 146 (1981) 360-368. Streicber, S.L.: Cloning and expression of the argG gene from $treptomyces cattleya in Escherichia coil and $. cattleya. Abstracts of the Annual Meeting oftbe American Society for Microbiology. American Society for Microbiology, Washington, DC, 1983, p. 134. Suarez, J.E. and Chater, K.F.: DNA cloning in $treptomyces: a bifunctional replicon comprising pBR322 inserted into a $treptomyces phage. Nature 286 (1980) 52"/-529. Thompson, C.J., Ward, J.M. and Hopwood, D.A.: DNA cloning in Streptomyces: resistance genes from antibiotic-producing species. Nature 286 (1980) 525-52"/. Thompson, C.J., Kieser, T., Ward, J.M. and Hopwood, D.A.: Physical analysis ofantibiotic-resistance genes from Streptomyces and their use in vector construction. Gene 20 (1982) 5 !-62.
