- Email: [email protected]
A new host-vector system allowing selection for foreign DNA inserts in bacteriophage λgtwes
Gene, 8 (1979) 69--80 © Elsevier/North-Holland Biomedical Press
69
A NEW HOST-VECTOR SYSTEM ALLOWING SELECTION FOR FOREIGN
DNA INSERTS IN BACTERIOPHAGE )~gtICES (Minimum insert, duplication, phages T5 and ),, plasmid ColIb)
J O H N DAVISON, F R A N ~ O I S E B R U N E L and MIREILLE M E R C H E Z Unit of Molecular Biology, International Institute of Cellular Pathology, Avenue Hippocrate, 75, B 1200 Brussels (Belgium)
(Received May 30th, 1979) (Revision received and accepted September 5th, 1979)
SUMMARY
An improved vector ()~gtWES.T5-622) for EcoRI fragments has been derived from EK2 vector kgtWES.kB' by replacing the )~B fragment with two identical 1.1 Md fragments from the pre-early region of bacteriophage T5. The new vector has two advantages which facilitate elimination of parentaltype recombinants in an in vitro recombination experiment. Firstly, the 1.1 Md insert is too small to be re-inserted into kgtWES in a sin~e copy. Secondly the 1.1 Md T5 fragment carries T5 gene A3 which prevents growth of phage retaining this fragment when the Escherichia coil host carries plasmid ColIb. Thus, essentially all plaques are due to phage with donor DNA inserts and are free of T5 DNA fragments. The size usually given as the theoretical minimum size for insertion into the kgt series of vectors is 0.66 Md. We have shown that this size is an underestimate and that the lower limit is about 1.6 Md. A precise estimate is difficult since there is strong selection, among phage having small inserts, for those which have acquired additional genetic material by duplication of the k DNA.
INTRODUCTION
Thomas et al. (1974) have constructed a derivative of bacteriophage )~ which is useful for the cloning of EcoRI fragments. This vector has been modified (Tiemeier et al., 1976, Leder et al., 1977) to produce a NIH certified EK2 cloning vehicle ),gtWES.),B'. This vector carries two EcoRI sites on either side of the kB fragment and during cloning, this fragment is removed and replaced by donor DNA (Fig. 3). The system has the advantage that, while all of the genes essential for plaque formation are located on the right and left arms of Abbreviations: Md, megadaltons; pfu, plaque-forming units.
70 the phage DNA, a molecule having only these two fragments cannot be packaged. Therefore, phages carrying inserts of up to 9 Md (Thomas et al., 1974) can be isolated as plaque forming phages. However, a viable phage is also often regenerated by the re-insertion of the kB fragment. The problem can be avoided by physical removal of the kB fragment prior to ligation using RPC-5 chromatography (Tiemeier et al., 1976) or preparative gel electrophoresis (Thomas et al., 1974). Alternatively, the vector can be treated with a second restriction enzyme SstI, which cuts the kB fragment twice but does not cut the left and right arms (Tiemeier et al., 1976). Neither of these solutions is completely satisfactory. The physical methods are tedious and require relatively large quantities of DNA, while the SstI procedure is inefficient since a viable phage can be generated by re-incorporation of most of the kB fragment, but omitting the central SstI fragment. This publication describes a new type of XgtWES vector in which the kB fragment has been removed and replaced by DNA of bacteriophage T5. Recombinant phages resulting from a re-incorporation of the T5 DNA fragment into the vector can be eliminated by genetical means, so that there is a very strong selection for recombinants carrying donor DNA EcoRI fragments. MATERIALS AND METHODS The triple mutant T5rislris3ris4 has been described previously (Davison and Brunel, 1979, Brunel and Davison, 1979). E. coil strains W3110 and its colicinogenic derivative Collb/W3110 were obtained from Dr. D.J. McCorquodale. Mutant T5A3h-12 is a mutant of T5 able to grow on ColIb/W3110 (Mizobuchi et al., 1971). ~gtWES.kB' and its host E. coli LE392 thy- were obtained from Leder et al. (1977). E. coli MM21 is a spontaneous streptomycin-resistant derivative of LE392 thy- which was made colicinogenic for ColIb by conjugation with ColIb/W3110. The origins of ~,-T5 hybrids ~gtWES.T5-6 and kgtWEg.T5-73 (which contain respectively the 1.1 Md and the 1.6 Md EcoRI fragments of T5rislris3ris4) are specified in the accompanying paper (Brunel et al., 1979). The phage referred to as k ÷ in this manuscript is in fact kcII68. Methods of construction of in vitro recombinants are given in the accompanying paper (Brunel et al., 1980). The experiments are exempt from the NIH Guidelines (1978). RESULTS
Construction of kgtWES. T5-622 During the cloning of EcoRI fragments of T5rislris3ris4 in XgtWES.XB' it was observed that the 1.1 Md fragment located between the r/sl and r/s3 mutations was never cloned alone, but always in combination with the XB fragment of the vector, or the 1.6 Md fragment of T5 (Brunel et al., 1979). In an effort to isolate a XgtWES clone having a single 1.1 Md fragment, the
71
DNA of AgtWES.T5-6 (which has both the XB and the T5 1.1 Md fragments, Fig. 1, Fig. 3) was restricted with EcoRI endonuclease and then re-ligated. The DNA was transfected into LE392thy- and the resulting plaques screened for their ability to complement the A2am231 mutant of T5. Since the 1.1 Md fragment carries a complete A2 gene, a positive complementation test indi-
Fig. 1. Construction of vector hgtWES.TS-622. DNA from phage hgtWES.T5-6 was cleaved with EcoRI endonuclease (slot c) and then re-ligated using T4 DNA ligase. This DNA was then used to transfect LE392thy- and the resulting plaques were tested for their ability to complement mutant TSA2am231. Of 64 plaques tested six gave a positive complementation response and the DNA of these was tested for its EcoRI restriction pattern by agarose gel electrophoresis (slots d, e, f, g, h, i). One isolate hgtWES.TS-622 (slot f) lacks the AB fragment present in AgtWES.T5-6 (slot c). Slots a and b contain EcoRI digests of hgtWES.hB’ and TStilris3ris4, the parents of agtWES.TS-6.
72
cates the presence of the 1.1 Md fragment (Brunel and Davison, 1980). Of 64 lysates tested only six were found which were able to complement TSam231. The DNA from these six lysates was then submitted to EcoRI restriction anal. yses and the results are given in Fig. 1. Five of the reconstituted phage are identical to the parent, having retained both the ~,B fragment and the T5 1.1 Md fragment. The sixth, kgtWES.T5.622 (Fig. 1 slot f), contains only the 1.1 Md fragment but close inspection shows that the band has too much fluorescence for a fragment present only once per molecule and suggests that this recombinant carries two or more copies of the 1.1 Md fragment. To verify this suggestion, the DNA of recombinant kgtWES.T5-622 was cleaved with BarnHI endonuclease and the products separated by agarose gel electrophoresis. BamHI cuts the vector on either side of the inserted DNA but does not cleave the 1.1 Md insert. Thus, the number of inserts can be estimated from the increase in molecular weight of this Barn fragment. The results of such an experiment are given in Fig. 2 and are expressed diagrammatically in Fig. 3. The molecular weight increase of the largest BamHI fragment is 2.13 Md. This corresponds to 1.94 copies of the 1.1 Md EcoRI frag-
Fig. 2. E c o R I and BamHI analysis of kgtWES.TS-622 and its relatives. DNA preparations were digested to completion with EcoRI, BamHI or both and subjected to agarose gel electrophoresis. (a) TSr/slr/s3r/s4 + EcoRI; (b) kgtWES.kB' + EcoRI; (c) kgtWES.kB' + BamHI; (d) kgtWES.kB' + EcoRI + BamHI; (e) kgtWES.T5-6 + EcoRI; (f) kgtWES.T5-6 + BamHI; (g) kgtWES.TS-6 + E c o R I + BamHI; (h)kgtWES.T5-622 + EcoRI; (i) kgtWES.T5622 + BamHI; (j) kgtWES.TS-622 + E c o R I + BamHI; (k) k÷ + E c o R L
73
hgt WES .~ B'
14.0 3.6
t
~gt WES ~-6
~',gt Wi=S .T5-622
12.'7
1/.,.0
36
t
11o 14.0
3.6
t
3.1
I
~ 3.1
~lt
t
~2
i1.111.11 14.5
9.2
I |2.5 ~
t
`7.5
9.2
t
`7~ 92 7.5
Fig. 3. Physical map of kgtWES.T5-6 and kgtWES.T5-622. The molecular weight estimates given for kgtWES.TS-6 and kgtWES.T5-622 are derived from several gels using EcoRI fragments o f T5 and k + as standards. The figures given for kgtWES.kB' are those of Tiemeier et al. (1976) and are in agreement with our estimations. It should be noted that the kB fragment o f kgtWES.T5-6 is inverted relative to the parent xgtWES.XB'. The orientations of the 1.1 Md T5 inserts are not yet known since these lack sites for the c o m m o n l y used restriction endonucleases. Downward arrows indicate EcoRI cleavage sites and upward arrows indicate BamHI cleavage sites.
ment per molecule. A similar conclusion was reached by electrophoretic separation and counting o f EcoRI fragments of T5 DNA uniformly labeled in vivo with 32p (Brunel and Davison, 1979). This m e t h o d gave an estimate of 1.88 copies of 1.1 Md fragment per molecule. These observations suggest that the theoretical figure o f 0.66 Md usually given as the lower size limit for DNA inserts into kgtWES (Tiemeier et al., 1976) is an underestimate and that the true value is in excess of 1.1 Md.
Cloning of a 1.6 Md T5 fragment in kgtWES In the accompanying publication the cloning of the 1.6 Md T5 EcoRI-F fragment was reported (Brunel and Davison, 1979). This recombinant was odd in that the molecular weight o f the left arm of ~ was increased. To investigate whether this is a general p h e n o m e n o n , a new series of 1.6 Md recombinants was isolated. This was facilitated by the observation that the 1.6 Md EcoRI fragment of T5 is the only fragment that can be cloned when T5 ÷ DNA is the donor (Brunel et al., 1979). Parental t y p e ~,WES.XB' recombinants were reduced by the use o f endonuclease SstI which cuts the kB fragment twice b u t does not cleave the left or right arms (Tiemeier et al., 1976). A series of plaques was isolated and the DNA screened by EcoRI restriction analysis (Fig. 4). A b o u t 50% of the plaques carry a fragment slightly smaller than kB. This represents a reconstituted kB fragment from which the central Sst portion is missing and shows that the use of SstI is n o t a very efficient w a y to select recombinant phage. The remainder of the plaques all have 1.6 Md inserts o f T5. Comparison of the reconstituted ~gtWES.XB* phage (Fig. 4) with the kgtWES.T5 phage shows that in every case the latter have an increased size for the ~, left arm, whereas the former are normal. More rarely recombinants, having 1.6 Md inserts, are seen that have an increased molecular weight for
74
Fig. 4. Cloning of the 1.6 Md T5 EcoRI DNA fragment in kgtWES.kB'. DNA from the kgtWES.kB' was treated with SstI and EcoRI endonucleases and ligated to T5 + DNA which had been treated only with EcoRI. The photograph shows EcoRI analysis of DNA from independently isolated recombinant plaques. About half of the plaques show a fragm e n t (2.4 Md) which belongs neither to kgtWES, kB' nor to T5 ÷. This is a kB* fragment which has been re-ligated in the absence of the central 3st fragment (0.7 Md). The remaining plaques all contain a 1.6 Md T5 DNA insert since this is the only clonable EcoRI fragment of T5 ÷. The left two slots contain EcoRI treated DNA of kgtWES.kB' and T5r/slr/s4, respectively. (For kgtWES, kB* see review of Enquist and Szybalski, 1978.)
the right arm of kWES instead of the left (Davison, unpublished results). These results suggest that the 1.6 Md insert is very close to the lower limit for insertion into kgtWE8 and that such recombinants are packaged inefficiently, leading to strong selection for additional DNA. A similar phenomenon has been studied by BeUett et al. (1971) who found that a k~b80 hybrid carrying a natural deletion of 23% became more stable by the acquisition of duplications in the kDNA. A XgtWES recombinant having only a 1.6 Md insert would have an effective deletion of 22% relative to k+.
Genetic properties of kgtWES. T5.622 Three closely linked pre~arly genes, A1, A2, A3, have been identified in bacteriophage T5 (Hendrickson and McCorquodale, 1971). The A2 gene is concerned with T5 second step DNA transfer into the cell and the A3 gene product (of both T5 and related phage BF23) causes abortive infection on E. coli hosts colicinogenic for plasmid ColIb (Lanni, 1969; Mizobuchi et al., 1971; McCorquodale et al., 1979). The 1.1 Md fragment located between the
75
risl and ris3 EcoRI restriction sites carries t he genes A2 and A3 (Brunel and Davison, 1979) and t he presence o f the A3 gene in kgtWES.T5-6 prevents g ro w t h o f this phage in E. coli MM21 which carries ColIb (Brunel et al., 1980). T h e new h y b r i d kgtWES.T5.622 described above carries t w o copies o f the 1.1 Md fragment and its plating ability (Table I) on MM21 is 6.10 -7, some t w e n t y times lower than its par e nt kgtWES.T5-6 which has a single copy. In contrast, phages ~*, kgtWES.kB' and kgtWES.T5-73 (which all lack t he 1.1 Md fragment) grow normally on MM21. This suggests t hat kgtWES.T5-622 has advantages over kgtWES.~B' as a vector. T he presence o f DNA b e t w e e n t h e left and right arms o f k is essential f or plaque f o r m a t i o n but the 1.1 Md parental DNA can be c o m p l e t e l y eliminated by plating on t he selective host MM21. This situation provides an e x t r e m e l y strong selection f o r r e c o m b i n a n t phage having foreign DNA inserts. T o f u r t h e r characterize t he p r o p o s e d host-vector system, it was necessary to measure th e ability o f MM21 t o act as a recipient in t he Mandel and Higa
TABLE I EFFECT OF ColIb ON THE PLAQUE-FORMING ABILITY OF PHAGE ~,gtWES.T5-622 AND ITS RELATIVES Phage
Yield ratio pfu on LE392thy- [ColIb] (=MM21) pfu on LE392thy-
T5 T5A3h-12 k+
1"10-' 5"10 -2 9-10-1 8"10 -~ 8"10 -~ 1"10 -5 6"10 -~
kgtWES.kB' kgtWES.T5-73 kgtWES.T5-6 kgtWES.T5-622
TABLE II EFFECT OF ColIb ON THE TRANSFECTION EFFICIENCY BY INTACT DNA OF kgtWES.T5-622 AND ITS RELATIVES DNA
Number of transformants per ~ g DNA LE392 LE392thy-[ ColIb ] (=MM21)
k ÷
3.1.10 s 4.0"10 ~ 2.5"10 s 2.5-105
kgtWES.kB' ~,gtWES.T5-6 ~gtWES.T5-622
2.0-10 s 2.2-10 s 0 0
76 ( 1 9 7 0 ) t r a n s f e c t i o n assay using i n t a c t D N A f r o m various derivatives o f )~. T a b l e I I s h o w s t h a t t h e p r e s e n c e o f C o l I b d o e s n o t greatly a f f e c t t h e transf e c t i o n ability w h e n i n t a c t )~+ o r )~gtWES.)~B' D N A are used. H o w e v e r , w i t h )~gtWES.T5-6 o r )~gtWES.TS.622 n o p l a q u e s were seen o n M M 2 1 , t h o u g h n o r m a l t r a n s f e c t i o n was o b t a i n e d using LE392thy-. T h e results given a b o v e s h o w t h a t p l a q u e f o r m a t i o n b y p h a g e )~gtWES.T5-
Fig. 5. Elimination of the T5 DNA fragments of kgtWES.TS-622 by selection on MM21. 'I~ne DNA of phage kgtWES.TS-622 was cleaved with EcoRI endonuclease, re-ligated in the absence of donor DNA, and transfected into MM21. The rare plaques obtained were isolated and the DNA subjected to EcoRI restriction analysis. The leftmost 12 slots contain EcoRI digests of DNA from phage able to grow on MM21, The rightmost (control) slot contains EcoRI digest of the parental kgtWES.TS-622 DNA. DNA/DNA hybridisation following transfer of DNA from this gel to nitrocellulose sheets (Southern, 1975) showed that, with the exception of the 1.1 Md fragment of kgtWES.TS-622, all fragments contain kgtWES DNA and none contain T5 DNA (J. Davison, unpublished data).
77 622 is extremely rare on E. coli MM21 and that transfection by intact DNA is below the limit of detection of the experiment. It was, therefore, expected that in a simulated in vitro genetic recombination experiment, plaque formation would be entirely dependent on the addition of foreign Ec0RI fragments. To test this, the DNA of kgtWES.T5-622 was cleaved with EcoRI, re-ligated and transfected into MM21. Surprisingly, plaques were obtained at low frequency (10-2). EcoRI analysis of the DNA from these plaques (Fig. 5) reveals that the 1.1 Md T5 fragments are absent. However, the lack of DNA caused by the loss of these fimgments is compensated in each isolate either by increased molecular weight of the left or right EcoRI fragments of ~gtWES, or by the presence of another EcoRI fragment. Southern blot hybridization experiments (Southern, 1975) show that all of the DNA fragments contain material homologous to kgtWES and none of them contain T5 specific sequences (J. Davison, unpublished data). However, in the absence of heteroduplex analysis, it is impossible to decide whether the additional DNA is derived entirely from kgtWES duplication, or from other sources such as E. coil or transposons. DISCUSSION
The 1.1 Md DNA fragment located between the r/sl and r/s3 mutations of
T5rislris3ris4 contains two intact pre~early genes, A2 and A3. When cloned in bacteriophage kWES.kB', both of these genes are expressed and can be detected; A2 by complementation of a T5A2 mutant and A3 by its ability to prevent growth on a strain carrying the plasmid ColIb. The 1.1 Md fragment has never been cloned alone but always in the presence of other DNA fragments and a deliberate attempt to isolate a ~ WES recombinant having only the 1.1 Md fragment resulted in a phage carrying two such fragments. This strongly suggests that the theoretical figure of 0.66 Md usually given as the lower limit for the size of the clonable fragment in kgtWE8 (Tiemeier et al., 1976) is too low. Although not specifically mentioned by the authors, a similar conclusion may be reached from the data of Velten et al. (1976) on the cloning of bacteriophage T4 DNA fragments. These workers found that fragments having a size of less than 1.5 Md were only isolated in recombinants carrying multiple inserts. We have cloned the 1.6 Md EcoRI fragment of bacteriophage T5 on many occasions and have found that, although it can be inserted separately, all clones contain added genetic material (probably duplications) in the left or right arms of kgtWES. This suggests that a 1.6 Md insert is the lower practical limit for insertion into ~gtWE8 and that although such phages are viable, there is strong selection for increased DNA content. Feiss et al. (1977) have shown that k DNA molecules are packaged normally and retain normal infectivity down to a lower limit of 75% of the normal DNA content, after which both packaging and infectivity are reduced. kgtWES phages with 1.1 Md or 1.6 Md inserts would have a DNA content of 76% and 78% respectively and so would be expected to have normal viability.
78 A possible explanation is provided by the observations that while kb211imm21 (73%) is viable (Sternberg and Weisberg, 1977), kgt.O (also 73%), from which .kgtWES was derived, is inviable in the absence of additional DNA (Thomas et al., 1974). This suggests that DNA sequences present in kb221imm21 but absent in kgt.O may contribute to viability when DNA content is limiting. A possible candidate for this would be the red region of k which is present in kb221imm21 but absent in kgt.O, and is necessary for a normal burst size, though not for viability. In the case of kgtWES, inefficient suppression of the amber mutations may also cause the reduced viability. An investigation into the effectiveness of the kgtWES.T5-622/MM21 host vector system as a way of selecting in vitro recombinants revealed that plaque formation in the absence of added foreign DNA is reduced about 100-fold. These plaques result from phage which have acquired DNA in addition to the left and right EcoRI fragments of kgtWES. The additional DNA may be in the left or right EcoRI fragments and when it includes an EcoRI site, a third EcoRI fragment is generated. The additional DNA probably results from duplication of kgtWES genetic material, though alternate sources such as E. coil DNA and transposons have not been ruled out. It should be noted that such rearrangements are not peculiar to the host-vector system described here, though the powerful selection system facilitates their detection when donor DNA is omitted. Molecules containing additions or duplications have also been rarely observed during routine cloning with XgtWES.kB' (Davison, unpublished observations) and during growth of a k ~ 8 0 hybrid carrying a large deletion (BeUet et ai., 1971). A number of systems have been described which enable phenotypic recognition of reconstituted parental-type plaques from those carrying new DNA inserts. These include the use of phage carrying a removable segment of DNA containing the lac or sue genes which enable colour recognition of parentaltype plaques (Rambach and Tiollais, 1974; Murray et al., 1974). The new vector XgtWES.T5-622 described in this report has the advantage that virtually all plaques obtained in an in vitro recombination experiment are due to phagecarrying donor DNA inserts. This is particularly useful when cloning a relatively simple genome since it avoids the inconvenience of using a hybridisation probe to detect the clones. The mechanism by which AgtWES.T5-622 is unable to grow on a ColIb host is not fully understood. It is known, both in T5 and in closely related phage BF23, that a wild type A3 gene product is the only phage gene necessary for abortive infection, though host and plasmid coded gene products are also needed (McCorquodale et al., 1979). Abortive infection results in massive leakage of potassium and nucleotides from the cell, suggesting that the phenomenon may be analogous to killing of sensitive cells by colicin Ib, and indicating that the A3 gene product may inactivate the immunity protein which protects ColIb cells from the colicin they produce (McCorquodale et al., 1979). It should be noted that the left and right EcoRI fragments of AgtWES.T5. 622 are identical to those of kgtWES.kB' (Leder et al., 1977) and that the re-
79
combinants formed using the two vectors are subsequently indistinguishable. Furthermore, the 1.1 Md T5 DNA fragment cannot compromise the safety features of kgtWES since the fragment is eliminated by selection on the ColIb host MM21. The only disadvantage with the host-vector system described here is that the NIH Guidelines (1978, Section II-D-l-a-1) specifically exclude the use of E. coli hosts carrying self-transmissible plasmids as EK1 hosts. The rationale behind this has been to prevent mobilisation of small non-conjugative plasmids, when these are used as vectors, from K-12 to more healthy strains of E. coli. However, there is no scientific reason to extend this prohibition to the use of phage as vectors since transfer of phage DNA by conjugation must certainly be trivial as compared to direct transfer by phage infection. In contrast, the British Genetic Manipulation Advisory Group (GMAG) has recently approved kgtWES.T5-622 as being equivalent to kgtWES.kB'. They require, however, that both phages are grown on a recA host, which unfounded requirement greatly reduces the yield of the phage. REFERENCES Bellett, A.J., Busse, H.G. and Baldwin, R.L., Tandem genetic duplications in a derivative of phage k, in Hershey, A.D. (Ed.), The Bacteriophage k, Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y., 1971, pp. 501--505. Brunel, F. and Davison, J., Restriction insensitivity in bacteriophage T5, III. Characterization of EcoRI sensitive mutants by restriction analysis, J. Mol. Biol., 128 (1979) 527-544. Brunel, F., Davison, J. and Merchez, M., Cloning of bacteriophage T5 DNA fragments in plasmid pBR322 and bacteriophage kgtWES, Gene, 8 (1979) 53--68. Davison, J. and Brunel, F., Restriction insensitivity in bacteriophage T5, I. Genetic characterization of mutants sensitive to EcoRI restriction, J. Virol., 29 (1979) 11--16. Enquist, L. and Szybalski, W., Coliphage lambda as a safe vector for recombinant DNA experiments in Kurstak, E. and Maramorosch, K. (Eds.), Viruses and Environment, Academic Press, New York, 1978, pp. 625--652. Feiss, M., Fisher, R.A., Crayton, M.A. and Enger, C., Packaging of the bacteriophage k chromosome: Effect of chromosome length, Virology, 77, (1977) 281--293. Hendrickson, H.E. and McCorquodale, D.J., Genetic and physiological studies of bacteriophage T5. An expanded genetic map of T5, J. Virol., 7 (1971) 612---618. Lanni, Y.T., Functions of two genes in the first-step-transfer DNA of bacteriophage T5, J. Mol. Biol., 44 (1969) 173--184. Leder, P., Tiemeier, D. and Enquist, L., EK2 derivatives of bacteriophage k useful in the cloning of DNA from higher organisms: the kgtWES system, Science, 196 (1977) 175-176. Mandel, M. and Higa, H., Calcium dependent bacteriophage infection, J. Mol. Biol., 53 (1970) 159--162. McCorquodale, D.J., Shaw, A.R., Moody, E.E., Hall, R.A. and Morgan, A.F., Is abortive infection by phage BF23 of E. coli harboring ColIb plasmids due to cell killing by internally liberated colicin Ib?, J. Virol., 31 (1979) (in press). Mizobuchi, K., Anderson, G.C. and McCorquodale, D.J., Abortive infection by bacteriophage BF23 due to colicin Ib factor, I. Genetic studies of nonrestricted and amber mutants of bacteriophage BF23, Genetics, 68 (1971) 323--340. Murray, N.E. and Murray, K., Manipulation of restriction targets in phage ~, to form receptor chromosomes for DNA fragments, Nature, 251 (1974) 476--481.
80 Rambach, A. and TioUais, P., Bacteriophage k having EcoRI endonuclease sites only in the nonessential region of the genome, Proc. Natl. Acad. Sci. USA, 71 (1974) 3927-3930. Southern, E_M., Detection of specific sequences among DNA fragments separated by gel electrophoresis, J. Mol. Biol., 78 (1975) 503--517. Sternberg, N. and Weisberg, R., Packaging of coliphage k DNA, II. The role of gene D protein, J. Mol. Biol., 117 (1977) 733--759. Thomas, M., Cameron, J.R. and Davis, R.W., Viable molecular hybrids of bacteriophage lambda and eukaryotic DNA, Proc. Natl. Acad. Sci. USA, 71 (1974) 4579--4583. Tiemeier, D., Enquist, L. and Leder, P., An improved derivative of a bacteriophage kEK2 vector useful in the cloning of recombinant DNA molecules: kgtWES.kB, Nature, 263 (1976) 526--527. Velten, J., Fukada, K. and Abelson, J., In vitro construction of bacteriophage k and plasmid molecules containing DNA fragments from bacteriophage T4. Gene, 1 (1976) 9 3 106. Communicated by A. Skalka.
69
A NEW HOST-VECTOR SYSTEM ALLOWING SELECTION FOR FOREIGN
DNA INSERTS IN BACTERIOPHAGE )~gtICES (Minimum insert, duplication, phages T5 and ),, plasmid ColIb)
J O H N DAVISON, F R A N ~ O I S E B R U N E L and MIREILLE M E R C H E Z Unit of Molecular Biology, International Institute of Cellular Pathology, Avenue Hippocrate, 75, B 1200 Brussels (Belgium)
(Received May 30th, 1979) (Revision received and accepted September 5th, 1979)
SUMMARY
An improved vector ()~gtWES.T5-622) for EcoRI fragments has been derived from EK2 vector kgtWES.kB' by replacing the )~B fragment with two identical 1.1 Md fragments from the pre-early region of bacteriophage T5. The new vector has two advantages which facilitate elimination of parentaltype recombinants in an in vitro recombination experiment. Firstly, the 1.1 Md insert is too small to be re-inserted into kgtWES in a sin~e copy. Secondly the 1.1 Md T5 fragment carries T5 gene A3 which prevents growth of phage retaining this fragment when the Escherichia coil host carries plasmid ColIb. Thus, essentially all plaques are due to phage with donor DNA inserts and are free of T5 DNA fragments. The size usually given as the theoretical minimum size for insertion into the kgt series of vectors is 0.66 Md. We have shown that this size is an underestimate and that the lower limit is about 1.6 Md. A precise estimate is difficult since there is strong selection, among phage having small inserts, for those which have acquired additional genetic material by duplication of the k DNA.
INTRODUCTION
Thomas et al. (1974) have constructed a derivative of bacteriophage )~ which is useful for the cloning of EcoRI fragments. This vector has been modified (Tiemeier et al., 1976, Leder et al., 1977) to produce a NIH certified EK2 cloning vehicle ),gtWES.),B'. This vector carries two EcoRI sites on either side of the kB fragment and during cloning, this fragment is removed and replaced by donor DNA (Fig. 3). The system has the advantage that, while all of the genes essential for plaque formation are located on the right and left arms of Abbreviations: Md, megadaltons; pfu, plaque-forming units.
70 the phage DNA, a molecule having only these two fragments cannot be packaged. Therefore, phages carrying inserts of up to 9 Md (Thomas et al., 1974) can be isolated as plaque forming phages. However, a viable phage is also often regenerated by the re-insertion of the kB fragment. The problem can be avoided by physical removal of the kB fragment prior to ligation using RPC-5 chromatography (Tiemeier et al., 1976) or preparative gel electrophoresis (Thomas et al., 1974). Alternatively, the vector can be treated with a second restriction enzyme SstI, which cuts the kB fragment twice but does not cut the left and right arms (Tiemeier et al., 1976). Neither of these solutions is completely satisfactory. The physical methods are tedious and require relatively large quantities of DNA, while the SstI procedure is inefficient since a viable phage can be generated by re-incorporation of most of the kB fragment, but omitting the central SstI fragment. This publication describes a new type of XgtWES vector in which the kB fragment has been removed and replaced by DNA of bacteriophage T5. Recombinant phages resulting from a re-incorporation of the T5 DNA fragment into the vector can be eliminated by genetical means, so that there is a very strong selection for recombinants carrying donor DNA EcoRI fragments. MATERIALS AND METHODS The triple mutant T5rislris3ris4 has been described previously (Davison and Brunel, 1979, Brunel and Davison, 1979). E. coil strains W3110 and its colicinogenic derivative Collb/W3110 were obtained from Dr. D.J. McCorquodale. Mutant T5A3h-12 is a mutant of T5 able to grow on ColIb/W3110 (Mizobuchi et al., 1971). ~gtWES.kB' and its host E. coli LE392 thy- were obtained from Leder et al. (1977). E. coli MM21 is a spontaneous streptomycin-resistant derivative of LE392 thy- which was made colicinogenic for ColIb by conjugation with ColIb/W3110. The origins of ~,-T5 hybrids ~gtWES.T5-6 and kgtWEg.T5-73 (which contain respectively the 1.1 Md and the 1.6 Md EcoRI fragments of T5rislris3ris4) are specified in the accompanying paper (Brunel et al., 1979). The phage referred to as k ÷ in this manuscript is in fact kcII68. Methods of construction of in vitro recombinants are given in the accompanying paper (Brunel et al., 1980). The experiments are exempt from the NIH Guidelines (1978). RESULTS
Construction of kgtWES. T5-622 During the cloning of EcoRI fragments of T5rislris3ris4 in XgtWES.XB' it was observed that the 1.1 Md fragment located between the r/sl and r/s3 mutations was never cloned alone, but always in combination with the XB fragment of the vector, or the 1.6 Md fragment of T5 (Brunel et al., 1979). In an effort to isolate a XgtWES clone having a single 1.1 Md fragment, the
71
DNA of AgtWES.T5-6 (which has both the XB and the T5 1.1 Md fragments, Fig. 1, Fig. 3) was restricted with EcoRI endonuclease and then re-ligated. The DNA was transfected into LE392thy- and the resulting plaques screened for their ability to complement the A2am231 mutant of T5. Since the 1.1 Md fragment carries a complete A2 gene, a positive complementation test indi-
Fig. 1. Construction of vector hgtWES.TS-622. DNA from phage hgtWES.T5-6 was cleaved with EcoRI endonuclease (slot c) and then re-ligated using T4 DNA ligase. This DNA was then used to transfect LE392thy- and the resulting plaques were tested for their ability to complement mutant TSA2am231. Of 64 plaques tested six gave a positive complementation response and the DNA of these was tested for its EcoRI restriction pattern by agarose gel electrophoresis (slots d, e, f, g, h, i). One isolate hgtWES.TS-622 (slot f) lacks the AB fragment present in AgtWES.T5-6 (slot c). Slots a and b contain EcoRI digests of hgtWES.hB’ and TStilris3ris4, the parents of agtWES.TS-6.
72
cates the presence of the 1.1 Md fragment (Brunel and Davison, 1980). Of 64 lysates tested only six were found which were able to complement TSam231. The DNA from these six lysates was then submitted to EcoRI restriction anal. yses and the results are given in Fig. 1. Five of the reconstituted phage are identical to the parent, having retained both the ~,B fragment and the T5 1.1 Md fragment. The sixth, kgtWES.T5.622 (Fig. 1 slot f), contains only the 1.1 Md fragment but close inspection shows that the band has too much fluorescence for a fragment present only once per molecule and suggests that this recombinant carries two or more copies of the 1.1 Md fragment. To verify this suggestion, the DNA of recombinant kgtWES.T5-622 was cleaved with BarnHI endonuclease and the products separated by agarose gel electrophoresis. BamHI cuts the vector on either side of the inserted DNA but does not cleave the 1.1 Md insert. Thus, the number of inserts can be estimated from the increase in molecular weight of this Barn fragment. The results of such an experiment are given in Fig. 2 and are expressed diagrammatically in Fig. 3. The molecular weight increase of the largest BamHI fragment is 2.13 Md. This corresponds to 1.94 copies of the 1.1 Md EcoRI frag-
Fig. 2. E c o R I and BamHI analysis of kgtWES.TS-622 and its relatives. DNA preparations were digested to completion with EcoRI, BamHI or both and subjected to agarose gel electrophoresis. (a) TSr/slr/s3r/s4 + EcoRI; (b) kgtWES.kB' + EcoRI; (c) kgtWES.kB' + BamHI; (d) kgtWES.kB' + EcoRI + BamHI; (e) kgtWES.T5-6 + EcoRI; (f) kgtWES.T5-6 + BamHI; (g) kgtWES.TS-6 + E c o R I + BamHI; (h)kgtWES.T5-622 + EcoRI; (i) kgtWES.T5622 + BamHI; (j) kgtWES.TS-622 + E c o R I + BamHI; (k) k÷ + E c o R L
73
hgt WES .~ B'
14.0 3.6
t
~gt WES ~-6
~',gt Wi=S .T5-622
12.'7
1/.,.0
36
t
11o 14.0
3.6
t
3.1
I
~ 3.1
~lt
t
~2
i1.111.11 14.5
9.2
I |2.5 ~
t
`7.5
9.2
t
`7~ 92 7.5
Fig. 3. Physical map of kgtWES.T5-6 and kgtWES.T5-622. The molecular weight estimates given for kgtWES.TS-6 and kgtWES.T5-622 are derived from several gels using EcoRI fragments o f T5 and k + as standards. The figures given for kgtWES.kB' are those of Tiemeier et al. (1976) and are in agreement with our estimations. It should be noted that the kB fragment o f kgtWES.T5-6 is inverted relative to the parent xgtWES.XB'. The orientations of the 1.1 Md T5 inserts are not yet known since these lack sites for the c o m m o n l y used restriction endonucleases. Downward arrows indicate EcoRI cleavage sites and upward arrows indicate BamHI cleavage sites.
ment per molecule. A similar conclusion was reached by electrophoretic separation and counting o f EcoRI fragments of T5 DNA uniformly labeled in vivo with 32p (Brunel and Davison, 1979). This m e t h o d gave an estimate of 1.88 copies of 1.1 Md fragment per molecule. These observations suggest that the theoretical figure o f 0.66 Md usually given as the lower size limit for DNA inserts into kgtWES (Tiemeier et al., 1976) is an underestimate and that the true value is in excess of 1.1 Md.
Cloning of a 1.6 Md T5 fragment in kgtWES In the accompanying publication the cloning of the 1.6 Md T5 EcoRI-F fragment was reported (Brunel and Davison, 1979). This recombinant was odd in that the molecular weight o f the left arm of ~ was increased. To investigate whether this is a general p h e n o m e n o n , a new series of 1.6 Md recombinants was isolated. This was facilitated by the observation that the 1.6 Md EcoRI fragment of T5 is the only fragment that can be cloned when T5 ÷ DNA is the donor (Brunel et al., 1979). Parental t y p e ~,WES.XB' recombinants were reduced by the use o f endonuclease SstI which cuts the kB fragment twice b u t does not cleave the left or right arms (Tiemeier et al., 1976). A series of plaques was isolated and the DNA screened by EcoRI restriction analysis (Fig. 4). A b o u t 50% of the plaques carry a fragment slightly smaller than kB. This represents a reconstituted kB fragment from which the central Sst portion is missing and shows that the use of SstI is n o t a very efficient w a y to select recombinant phage. The remainder of the plaques all have 1.6 Md inserts o f T5. Comparison of the reconstituted ~gtWES.XB* phage (Fig. 4) with the kgtWES.T5 phage shows that in every case the latter have an increased size for the ~, left arm, whereas the former are normal. More rarely recombinants, having 1.6 Md inserts, are seen that have an increased molecular weight for
74
Fig. 4. Cloning of the 1.6 Md T5 EcoRI DNA fragment in kgtWES.kB'. DNA from the kgtWES.kB' was treated with SstI and EcoRI endonucleases and ligated to T5 + DNA which had been treated only with EcoRI. The photograph shows EcoRI analysis of DNA from independently isolated recombinant plaques. About half of the plaques show a fragm e n t (2.4 Md) which belongs neither to kgtWES, kB' nor to T5 ÷. This is a kB* fragment which has been re-ligated in the absence of the central 3st fragment (0.7 Md). The remaining plaques all contain a 1.6 Md T5 DNA insert since this is the only clonable EcoRI fragment of T5 ÷. The left two slots contain EcoRI treated DNA of kgtWES.kB' and T5r/slr/s4, respectively. (For kgtWES, kB* see review of Enquist and Szybalski, 1978.)
the right arm of kWES instead of the left (Davison, unpublished results). These results suggest that the 1.6 Md insert is very close to the lower limit for insertion into kgtWE8 and that such recombinants are packaged inefficiently, leading to strong selection for additional DNA. A similar phenomenon has been studied by BeUett et al. (1971) who found that a k~b80 hybrid carrying a natural deletion of 23% became more stable by the acquisition of duplications in the kDNA. A XgtWES recombinant having only a 1.6 Md insert would have an effective deletion of 22% relative to k+.
Genetic properties of kgtWES. T5.622 Three closely linked pre~arly genes, A1, A2, A3, have been identified in bacteriophage T5 (Hendrickson and McCorquodale, 1971). The A2 gene is concerned with T5 second step DNA transfer into the cell and the A3 gene product (of both T5 and related phage BF23) causes abortive infection on E. coli hosts colicinogenic for plasmid ColIb (Lanni, 1969; Mizobuchi et al., 1971; McCorquodale et al., 1979). The 1.1 Md fragment located between the
75
risl and ris3 EcoRI restriction sites carries t he genes A2 and A3 (Brunel and Davison, 1979) and t he presence o f the A3 gene in kgtWES.T5-6 prevents g ro w t h o f this phage in E. coli MM21 which carries ColIb (Brunel et al., 1980). T h e new h y b r i d kgtWES.T5.622 described above carries t w o copies o f the 1.1 Md fragment and its plating ability (Table I) on MM21 is 6.10 -7, some t w e n t y times lower than its par e nt kgtWES.T5-6 which has a single copy. In contrast, phages ~*, kgtWES.kB' and kgtWES.T5-73 (which all lack t he 1.1 Md fragment) grow normally on MM21. This suggests t hat kgtWES.T5-622 has advantages over kgtWES.~B' as a vector. T he presence o f DNA b e t w e e n t h e left and right arms o f k is essential f or plaque f o r m a t i o n but the 1.1 Md parental DNA can be c o m p l e t e l y eliminated by plating on t he selective host MM21. This situation provides an e x t r e m e l y strong selection f o r r e c o m b i n a n t phage having foreign DNA inserts. T o f u r t h e r characterize t he p r o p o s e d host-vector system, it was necessary to measure th e ability o f MM21 t o act as a recipient in t he Mandel and Higa
TABLE I EFFECT OF ColIb ON THE PLAQUE-FORMING ABILITY OF PHAGE ~,gtWES.T5-622 AND ITS RELATIVES Phage
Yield ratio pfu on LE392thy- [ColIb] (=MM21) pfu on LE392thy-
T5 T5A3h-12 k+
1"10-' 5"10 -2 9-10-1 8"10 -~ 8"10 -~ 1"10 -5 6"10 -~
kgtWES.kB' kgtWES.T5-73 kgtWES.T5-6 kgtWES.T5-622
TABLE II EFFECT OF ColIb ON THE TRANSFECTION EFFICIENCY BY INTACT DNA OF kgtWES.T5-622 AND ITS RELATIVES DNA
Number of transformants per ~ g DNA LE392 LE392thy-[ ColIb ] (=MM21)
k ÷
3.1.10 s 4.0"10 ~ 2.5"10 s 2.5-105
kgtWES.kB' ~,gtWES.T5-6 ~gtWES.T5-622
2.0-10 s 2.2-10 s 0 0
76 ( 1 9 7 0 ) t r a n s f e c t i o n assay using i n t a c t D N A f r o m various derivatives o f )~. T a b l e I I s h o w s t h a t t h e p r e s e n c e o f C o l I b d o e s n o t greatly a f f e c t t h e transf e c t i o n ability w h e n i n t a c t )~+ o r )~gtWES.)~B' D N A are used. H o w e v e r , w i t h )~gtWES.T5-6 o r )~gtWES.TS.622 n o p l a q u e s were seen o n M M 2 1 , t h o u g h n o r m a l t r a n s f e c t i o n was o b t a i n e d using LE392thy-. T h e results given a b o v e s h o w t h a t p l a q u e f o r m a t i o n b y p h a g e )~gtWES.T5-
Fig. 5. Elimination of the T5 DNA fragments of kgtWES.TS-622 by selection on MM21. 'I~ne DNA of phage kgtWES.TS-622 was cleaved with EcoRI endonuclease, re-ligated in the absence of donor DNA, and transfected into MM21. The rare plaques obtained were isolated and the DNA subjected to EcoRI restriction analysis. The leftmost 12 slots contain EcoRI digests of DNA from phage able to grow on MM21, The rightmost (control) slot contains EcoRI digest of the parental kgtWES.TS-622 DNA. DNA/DNA hybridisation following transfer of DNA from this gel to nitrocellulose sheets (Southern, 1975) showed that, with the exception of the 1.1 Md fragment of kgtWES.TS-622, all fragments contain kgtWES DNA and none contain T5 DNA (J. Davison, unpublished data).
77 622 is extremely rare on E. coli MM21 and that transfection by intact DNA is below the limit of detection of the experiment. It was, therefore, expected that in a simulated in vitro genetic recombination experiment, plaque formation would be entirely dependent on the addition of foreign Ec0RI fragments. To test this, the DNA of kgtWES.T5-622 was cleaved with EcoRI, re-ligated and transfected into MM21. Surprisingly, plaques were obtained at low frequency (10-2). EcoRI analysis of the DNA from these plaques (Fig. 5) reveals that the 1.1 Md T5 fragments are absent. However, the lack of DNA caused by the loss of these fimgments is compensated in each isolate either by increased molecular weight of the left or right EcoRI fragments of ~gtWES, or by the presence of another EcoRI fragment. Southern blot hybridization experiments (Southern, 1975) show that all of the DNA fragments contain material homologous to kgtWES and none of them contain T5 specific sequences (J. Davison, unpublished data). However, in the absence of heteroduplex analysis, it is impossible to decide whether the additional DNA is derived entirely from kgtWES duplication, or from other sources such as E. coil or transposons. DISCUSSION
The 1.1 Md DNA fragment located between the r/sl and r/s3 mutations of
T5rislris3ris4 contains two intact pre~early genes, A2 and A3. When cloned in bacteriophage kWES.kB', both of these genes are expressed and can be detected; A2 by complementation of a T5A2 mutant and A3 by its ability to prevent growth on a strain carrying the plasmid ColIb. The 1.1 Md fragment has never been cloned alone but always in the presence of other DNA fragments and a deliberate attempt to isolate a ~ WES recombinant having only the 1.1 Md fragment resulted in a phage carrying two such fragments. This strongly suggests that the theoretical figure of 0.66 Md usually given as the lower limit for the size of the clonable fragment in kgtWE8 (Tiemeier et al., 1976) is too low. Although not specifically mentioned by the authors, a similar conclusion may be reached from the data of Velten et al. (1976) on the cloning of bacteriophage T4 DNA fragments. These workers found that fragments having a size of less than 1.5 Md were only isolated in recombinants carrying multiple inserts. We have cloned the 1.6 Md EcoRI fragment of bacteriophage T5 on many occasions and have found that, although it can be inserted separately, all clones contain added genetic material (probably duplications) in the left or right arms of kgtWES. This suggests that a 1.6 Md insert is the lower practical limit for insertion into ~gtWE8 and that although such phages are viable, there is strong selection for increased DNA content. Feiss et al. (1977) have shown that k DNA molecules are packaged normally and retain normal infectivity down to a lower limit of 75% of the normal DNA content, after which both packaging and infectivity are reduced. kgtWES phages with 1.1 Md or 1.6 Md inserts would have a DNA content of 76% and 78% respectively and so would be expected to have normal viability.
78 A possible explanation is provided by the observations that while kb211imm21 (73%) is viable (Sternberg and Weisberg, 1977), kgt.O (also 73%), from which .kgtWES was derived, is inviable in the absence of additional DNA (Thomas et al., 1974). This suggests that DNA sequences present in kb221imm21 but absent in kgt.O may contribute to viability when DNA content is limiting. A possible candidate for this would be the red region of k which is present in kb221imm21 but absent in kgt.O, and is necessary for a normal burst size, though not for viability. In the case of kgtWES, inefficient suppression of the amber mutations may also cause the reduced viability. An investigation into the effectiveness of the kgtWES.T5-622/MM21 host vector system as a way of selecting in vitro recombinants revealed that plaque formation in the absence of added foreign DNA is reduced about 100-fold. These plaques result from phage which have acquired DNA in addition to the left and right EcoRI fragments of kgtWES. The additional DNA may be in the left or right EcoRI fragments and when it includes an EcoRI site, a third EcoRI fragment is generated. The additional DNA probably results from duplication of kgtWES genetic material, though alternate sources such as E. coil DNA and transposons have not been ruled out. It should be noted that such rearrangements are not peculiar to the host-vector system described here, though the powerful selection system facilitates their detection when donor DNA is omitted. Molecules containing additions or duplications have also been rarely observed during routine cloning with XgtWES.kB' (Davison, unpublished observations) and during growth of a k ~ 8 0 hybrid carrying a large deletion (BeUet et ai., 1971). A number of systems have been described which enable phenotypic recognition of reconstituted parental-type plaques from those carrying new DNA inserts. These include the use of phage carrying a removable segment of DNA containing the lac or sue genes which enable colour recognition of parentaltype plaques (Rambach and Tiollais, 1974; Murray et al., 1974). The new vector XgtWES.T5-622 described in this report has the advantage that virtually all plaques obtained in an in vitro recombination experiment are due to phagecarrying donor DNA inserts. This is particularly useful when cloning a relatively simple genome since it avoids the inconvenience of using a hybridisation probe to detect the clones. The mechanism by which AgtWES.T5-622 is unable to grow on a ColIb host is not fully understood. It is known, both in T5 and in closely related phage BF23, that a wild type A3 gene product is the only phage gene necessary for abortive infection, though host and plasmid coded gene products are also needed (McCorquodale et al., 1979). Abortive infection results in massive leakage of potassium and nucleotides from the cell, suggesting that the phenomenon may be analogous to killing of sensitive cells by colicin Ib, and indicating that the A3 gene product may inactivate the immunity protein which protects ColIb cells from the colicin they produce (McCorquodale et al., 1979). It should be noted that the left and right EcoRI fragments of AgtWES.T5. 622 are identical to those of kgtWES.kB' (Leder et al., 1977) and that the re-
79
combinants formed using the two vectors are subsequently indistinguishable. Furthermore, the 1.1 Md T5 DNA fragment cannot compromise the safety features of kgtWES since the fragment is eliminated by selection on the ColIb host MM21. The only disadvantage with the host-vector system described here is that the NIH Guidelines (1978, Section II-D-l-a-1) specifically exclude the use of E. coli hosts carrying self-transmissible plasmids as EK1 hosts. The rationale behind this has been to prevent mobilisation of small non-conjugative plasmids, when these are used as vectors, from K-12 to more healthy strains of E. coli. However, there is no scientific reason to extend this prohibition to the use of phage as vectors since transfer of phage DNA by conjugation must certainly be trivial as compared to direct transfer by phage infection. In contrast, the British Genetic Manipulation Advisory Group (GMAG) has recently approved kgtWES.T5-622 as being equivalent to kgtWES.kB'. They require, however, that both phages are grown on a recA host, which unfounded requirement greatly reduces the yield of the phage. REFERENCES Bellett, A.J., Busse, H.G. and Baldwin, R.L., Tandem genetic duplications in a derivative of phage k, in Hershey, A.D. (Ed.), The Bacteriophage k, Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y., 1971, pp. 501--505. Brunel, F. and Davison, J., Restriction insensitivity in bacteriophage T5, III. Characterization of EcoRI sensitive mutants by restriction analysis, J. Mol. Biol., 128 (1979) 527-544. Brunel, F., Davison, J. and Merchez, M., Cloning of bacteriophage T5 DNA fragments in plasmid pBR322 and bacteriophage kgtWES, Gene, 8 (1979) 53--68. Davison, J. and Brunel, F., Restriction insensitivity in bacteriophage T5, I. Genetic characterization of mutants sensitive to EcoRI restriction, J. Virol., 29 (1979) 11--16. Enquist, L. and Szybalski, W., Coliphage lambda as a safe vector for recombinant DNA experiments in Kurstak, E. and Maramorosch, K. (Eds.), Viruses and Environment, Academic Press, New York, 1978, pp. 625--652. Feiss, M., Fisher, R.A., Crayton, M.A. and Enger, C., Packaging of the bacteriophage k chromosome: Effect of chromosome length, Virology, 77, (1977) 281--293. Hendrickson, H.E. and McCorquodale, D.J., Genetic and physiological studies of bacteriophage T5. An expanded genetic map of T5, J. Virol., 7 (1971) 612---618. Lanni, Y.T., Functions of two genes in the first-step-transfer DNA of bacteriophage T5, J. Mol. Biol., 44 (1969) 173--184. Leder, P., Tiemeier, D. and Enquist, L., EK2 derivatives of bacteriophage k useful in the cloning of DNA from higher organisms: the kgtWES system, Science, 196 (1977) 175-176. Mandel, M. and Higa, H., Calcium dependent bacteriophage infection, J. Mol. Biol., 53 (1970) 159--162. McCorquodale, D.J., Shaw, A.R., Moody, E.E., Hall, R.A. and Morgan, A.F., Is abortive infection by phage BF23 of E. coli harboring ColIb plasmids due to cell killing by internally liberated colicin Ib?, J. Virol., 31 (1979) (in press). Mizobuchi, K., Anderson, G.C. and McCorquodale, D.J., Abortive infection by bacteriophage BF23 due to colicin Ib factor, I. Genetic studies of nonrestricted and amber mutants of bacteriophage BF23, Genetics, 68 (1971) 323--340. Murray, N.E. and Murray, K., Manipulation of restriction targets in phage ~, to form receptor chromosomes for DNA fragments, Nature, 251 (1974) 476--481.
80 Rambach, A. and TioUais, P., Bacteriophage k having EcoRI endonuclease sites only in the nonessential region of the genome, Proc. Natl. Acad. Sci. USA, 71 (1974) 3927-3930. Southern, E_M., Detection of specific sequences among DNA fragments separated by gel electrophoresis, J. Mol. Biol., 78 (1975) 503--517. Sternberg, N. and Weisberg, R., Packaging of coliphage k DNA, II. The role of gene D protein, J. Mol. Biol., 117 (1977) 733--759. Thomas, M., Cameron, J.R. and Davis, R.W., Viable molecular hybrids of bacteriophage lambda and eukaryotic DNA, Proc. Natl. Acad. Sci. USA, 71 (1974) 4579--4583. Tiemeier, D., Enquist, L. and Leder, P., An improved derivative of a bacteriophage kEK2 vector useful in the cloning of recombinant DNA molecules: kgtWES.kB, Nature, 263 (1976) 526--527. Velten, J., Fukada, K. and Abelson, J., In vitro construction of bacteriophage k and plasmid molecules containing DNA fragments from bacteriophage T4. Gene, 1 (1976) 9 3 106. Communicated by A. Skalka.