MGG

Mol Gen Genet (1981) 184:200 207

© Springer-Verlag 1981

Interplasmidic and Intraplasmidic Recombination in Escherichia coil K-12 Avraham Laban and Amikam Cohen Department of Microbiological Chemistry, The Hebrew University, Hadassah Medical School, 91010 Jerusalem, Israel

Summary. The construction of plasmids which facilitate the study of interplasmidic and intraplasmidic recombination is described. In this system, a single recombination event between two mutated Tc ~ genes on separate plasmids or on one plasmid leads to a change in the host phenotype from sensitivity to resistance to tetracycline. Recombination proficiencies have been determined for different E. coli K-12 strains: both interplasmidic and intraplasmidic recombination are independent of the recBC gene product. RecA mutations decrease the proficiency of plasmidic recombination 40-100 fold. Intraplasmidic and interplasmidic recombination via the recE pathway are more efficient than via the recBC pathway. Intraplasmidic recombination, but not interplasmidic recombination via the recE pathway is independent of a functional recA product.

Introduction Plasmids which contain directly repeated DNA sequences are subjected to deletion during propagation in bacteria (Brutlag et al. 1977; Cramer et al. 1977; Cohen et al. 1978; Honigman et al. 1981). The involvement of repetitive sequences in this process, the infrequent occurrence of deletion products in plasmid preparations of recA mutants, and the structure of the deletion products, have led to the conclusion that deletion takes place in these plasmids by intraplasmidic recombination (Cramer et al. 1977; Cohen et al. 1978). While the presence of deletion products in preparations of plasmid DNA carrying sequence duplications suggests the occurrence of intraplasmidic recombination, the ratio of deletion products to the original plasmids in these preparations does not necessarily reflect the degree of recombination proficiency of the bacterial strain. Since both the original plasmid and its deletion product belong to the same incompatibility group, any selective advantage for replication or maintenance of the deletion product would lead to a rapid change in the plasmid population following intraplasmidic recombination. A selective advantage of cells which harbor the deletion product would lead to an enrichment of these cells in the bacterial culture. In an attempt to determine the relative proficiency of intraplasmidic recombination in different mutants, we constructed plasmids, in which a single intraplasmidic recombination event would lead to a detectable change in its host phenotype. These plasmids contain a duplication of the tetracycline resistance gene (Tc~), with each one of the two copies of the gene carrying Offprint requests to." Dr. Amikam Cohen

0026-8925/81/0184/0200/$01.60

a mutation at a different site. Intraplasmidic recombination between the repetitive sequences would lead to deletion of a fragment between the sites of recombination and the formation of a functional tetracycline resistance (Td) gene (Fig. 3B). Thus, intraplasmidic recombination in these plasmids may be followed by determining the proportion of tetracycline resistant cells in the transformed culture. Using a similar approach, a rate of interplasmidic recombination is estimated by determining the number of tetracycline resistant cells in clones which carry two plasmids, each of which contains a tetracycline resistance gene with a mutation at a different site (Fig. 3A). In this system, a single recombination event between the sites of mutations in the two plasmids may lead to the formation of a functional Tc r gene (Fig. 3A). Using these systems, we have determined the proficiency of interplasmidic and intraplasmidic recombinations in E. coli K-12 strains, carrying various mutations which affect the rate of recombination. Recombination in E. coli has been demonstrated to occur through three pathways recBC, recE, and reeF. The recBC pathway is the major pathway of E. coli wild type cells. This pathway depends on the activity of the recA and recB recC gene products (for review see Clark 1973, 1974). Mutations in the recB recC gene are indirectly suppressed by two mutations: The sbcA mutation, which activates the recE pathway by derepressing the synthesis of the ATP-independent exonuclease VIII (Barbour et al. 1970), and by the sbcB mutation which activates the reef pathway by inactivating exonuclease I (Kushner et al. 1972). All three pathways depend on a functional recA gene product for formation of recombinants following conjugation (Clark 1974). On the other hand, recombination of 2 phage DNA through the recE pathway takes place at a relatively high proficiency in sbcA recA recB recC mutants (Gillen and Clark 1974). The relative proficiency of interplasmidic and intraplasmidic recombination through the recBC and recE recombination pathways and the dependency of this process on the recA gene product is presented. Materials and Methods Bacteria and Growth Conditions

Bacterial strains used in this work were derived from E. coli K-12. Their origin and relevant genotypes are presented in Table 1. Most of the strains used belong to two sets of isogenic strains. One, constructed in Dr. A.J. Clark's laboratory and isogenic to ABl157 (ABl157, JC2926, JC5459, JC5519 and JC8679), was obtained from Dr. Jerry Cohen. The other set, prepared by Dr. R.E. Malone (REM199, REM200, REM201,

201 PvulS EcoRI

Table 1. Bacterial strains

Designation

Pertinent genotype

Source and/or reference

C600 ABl157 REM199 JC2926 REM201 DM455 DR100 JC5495 JC5519 REM202 JC8679 JC5183 DR107

-

Bachmann 1972 Bachmann 1972 Stahl and Stahl 1977 A.J. Clark Stahl and Stahl 1977 Mount 1971 this laboratory A.J. Clark Stahl and Stahl 1977 Stahl and Stahl 1977 Gillen et al. 1977 Barbour et al. 1970 this laboratory

recA13 recA1 recA99 recA recA13 recB21 recB21 recC22 recB21 recB21 recC22 sbcA23 recB21 recC22 sbcA5 recA recB21 recC22 sbcA5

Psi

I

",5;......x/--s°''_ _"z'F'--"'~ AvoI

pvul~ j

~

.

.

Hindm

Psll

3

Pvull "EcoRI ~

3 Kb

/

/

// //' / / 1/ / I/

EcoRI Hindll ~r~-BomHI

A¥Ol

k Pvull

I Hind 'IT[ 1 ligase

]Sal I

IS, nucl. . . . l[,qase

Pvull • EcoRI

t

~..I~--

,,o:/

EcoRI

.

S I nuclease ~T4 ligase

Pvull ~ T ~ \

REM202), was obtained from Dr. F. Stahl. The recA mutation in DR100 and DR107 was transferred by PI tranduction to strains C600 and JC5185, repsectively, from CS101 (recA srl::TnlO), as described by Metzer et al. (1979). Tetracycline-sensitive derivatives of UV-sensitive tetracycline-resistant transductants were derived by penicillin selection (Kleckner et al. 1978). The UV survival curves of DR100 and DR107 were identical to that of JC2926. Bacteria were grown in LB medium (Lennox 1955). Those carrying drug resistant plasmids were grown in the presence of antibiotics (10 gg/ml tetracycline, 100 pg/ml ampicillin or 20 gg/ml chloramphenicol), recA and recA + genotypes were checked routinely by testing for the ability of the cells to propagate )~ red- gain- phage (Zissler et al. 1971) and for sensitivity to UV irradiation.

.

,,o-

HindII2

EcoRI

Hind ~I

\ / PvullN N @ PVulI

P~,I

~ F - ~

Enzymes

The T4 ligase and all restriction enzymes except EcoRI and BamHI were obtained from New England Biolabs. EcoRI and BamHI were prepared in our laboratory by Z. Silberstein. D N A cleavage with restriction enzyme was carried out in 60 m M Tris HCI, pH 7.5, 60 m M MgClz, 50 m M NaC1 and 1 m M dithiothreitol. Ligation was carried out as described by Sugino and Goodman (1977). $1 nuclease was obtained from Miles Laboratories. Digestion with $1 nuclease was according to Weintraub and Groudine (1976), at 37 ° C in 0.03 M sodium acetate buffer, pH 4.8, 0.2 M NaC1, 0.2 m M ZnSO4 and 20 gg/ml Salmon sperm single-stranded D N A and 1 pg of plasmid D N A substrate. Plasmid Purification, Transformation and Gel Electrophoresis

Purification of plasmids, agarose gel electrophoresis and transformation were performed as described previously (Cohen et al. 1978). Separation of HinfI restriction endonuclease fragments was by electrophoresis on 5% polyacrylamide gel. HindIII, SalI and B a m H I D N A fragments were separated on 1% agarose gel. Results Construction o f Plasmids

Intraplasmidic recombination was studied by using plasmids containing a duplication of the Tc ~ gene, with each one of the two gene copies having a mutation at a different site. Interplas-

z / . . . .

AvQI

111%

Hin(Jlit

~

L2 Fig. 1. Diagrammatic representation of the construction of pAL210. The Tc r gene of pBR322 or pBR325 was mutagenized as described in the text. pAL199 and pAL201 were separately digested with restriction endonuclease PvuII. Following electrophoresis, the generated fragments of pAL201 were isolated from the agarose gel and the fragment which contains the mutated Tc r gene was inserted into the PvuII site of pAL199 by T4 ligase, to generate pAL210. Heavy line indicates DNA insertion, • - site of deletions

midic recombination was studied using cells which harbor two species of plasmids, each of which has a Tc r gene mutated at a different site. The construction of plasmids which serve as substrates for interplasmidic recombination (pAL199 and pAL201) and for intraplasmidic recombination (pAL210) is illustrated in Fig. 1. Mutations in the tetracycline resistance gene and the ampicillin resistance gene (Ap r) of pBR322 (Bolivar et al. 1977) and pBR325 (Bolivar 1978) were induced in vitro: plasmid D N A preparations were digested with restriction endonucleases which produce cohesive, single-strand ends. The restricted plasmids were treated

202 A

B C

DE

F

G

H

I

J -5.4

-4,4 -3.7

-0.4

A substrate for intraplasmidic recombination, pAL210, was constructed as follows: pAL199 and pAL201 were digested with restriction endonuclease PvuII. The PvuII fragment of pAL201 which contains the mutated Tc r gene was purified by agarose gel electrophoresis and electroelution. This fragment was inserted into the PvuII site of pAL199. The structure of pAL210 was confirmed by restriction endonuclease analysis with BamHI, HindIII and EcoRI (not shown). Reversion frequencies of the mutations in the Tc r gene of pALl99, pAL201, and pAL206 were less than 10 -9.

The Recombination Products

Fig. 2. Restriction endonuclease fragments of plasmids mutated in the Tcr gene. Following in vitro mutagenesis, transformation and clone selection on the appropriate medium, plasmids were purified from ampicillin-resistant, tetracycline-sensitiveclones (pAL199 and pAL206) or chloramphenicol-resistant tetracycline-sensitive clones (pAL201). Restriction endonuclease digestion and electrophoresis were performed as described in the text. Digestion products of pBR322 are presented in lanes A, F; of pAL199 in lanes B, D, G; of pAL206 in lanes C, E, H; of pBR325 in lane I; and of pAL201 in lane J. DNA preparations were digested with SalI - lanes A, B, C. EeoRI - lanes D, E. HinfI - lanes F, G, H. HindIII - lanes I, J. Closed circular and open circular forms of pALl99 and pAL206 are visible in lanes B, C (compare to the linear form of the same plasmids in lanes D, E). HinfI fragments 7a and 7b of pBR322 have the same electrophoretic mobility (Sutcliffe 1978). The deletion of HinfI fragment 7a in pAL199 is deduced from densitometer tracing of lanes G, F (not shown) and from the fact that HinfI 7a fragment is located between fragments HinfI1 and HinfI8 (Sutcliffe 1978)

with $1 nuclease as described above, and then the blunt ends were ligated using T4 DNA ligase. Restriction endonucleases HindIII, BamHI and SalI were used for mutagenesis of the Tc r gene, and PstI was used for mutagenesis of the Apr gene. Alternatively, mutations were induced by inserting a )~ DNA fragment into restriction sites using the appropriate restriction enzyme and T4 DNA ligase. Restriction analysis of plasmids following $1 digestion and ligation indicates that in some of the plasmids this procedure has led to deletion of the corresponding restriction site, while in others it led to a deletion of a larger fragment in the vicinity of the restriction site. This may be due to a trace of doublestrand-specific nuclease in the preparation. Examples of pBR322 and pBR325 plasmids mutagenized in the Tc ~gene are presented in Fig. 2. pAL206 has a short deletion at the SalI site, which makes it resistant to this restriction endonuclease (lane C). Its HinfI restriction pattern (lane H) is indistinguishable from that of pBR322 (lane F), suggesting that there are no extensive deletions in this plasmid. On the other hand, pAL199 lost a fragment of about 700 nucleotides, corresponding to fragments HinfI7a, 8, and part of 1 (lane G) which are located between 300 and 1,000 nucleotides away from the E c o R I site of pBR322 (Sutcliffe 1978). The structure of this plasmid was confirmed by restriction endonuclease analysis with TaqI, HaeII, AluI, BamHI and SaII. The plasmid pAL201 was constructed by deleting the PstI site of pBR325, as described above, and insertion of a 0.4 Kb 2 HindlII fragment into the plasmid's HindIII site (Fig. 2, lane J). Since pAL199 and pAL206 carry a functional Ap ~ gene and pAL201 carries a functional chloramphenicol resistance (Cmr), gene, but a mutated Apr gene, pAL199 or pAL206 could be stably maintained with pAL201 in cells growing in the presence of ampicillin and chloramphenicol.

Cells which have been transformed with both pAL199 and pAL201 were grown in the presence of ampicillin and chloramphenicol, as described in legend to Fig. 5. Tetracycline resistant clones were obtained by plating the transformed cells on tetracycline plates. Plasmid preparations of sixteen clones were used for transformation of recA (DR100) cells, followed by selection of tetracycline resistant transformants. The tetracycline resistant clones of DR100 cells were then tested for resistance to chloramphenicol and ampicillin. All tetracycline resistant DRI00 clones which were tested (160 clones) were also found resistant to chloramphenicol and ampicillin. This result suggests that resistance to tetracycline in the clones which were tested, results from recombination of the two plasmids between the mutation sites of the Tc r gene (Fig. 3). Such a recombination process would lead to the formation of a dimer of pAL199 and pAL201 with two Tc r genes, one functional, and the other carrying both mutations. To verify this prediction, the recombination product which confers tetracycline resistance on its host was purified, digested with AvaI restriction endonuclease, and the restriction products were ligated and transformed back into DR100 cells. Ampieillin resistant and chloramphenicol resistant clones were selected on the appropriate plates and then tested for resistance to tetracycline. While all ampicillin resistant clones were resistant to tetracycline, none of the chloramphenicol resistant clones were tetracycline resistant. Further support for the structure of the recombinant product which is presented in Fig. 3A is obtained by restriction analysis of A vaI fragments of the recombination product of pAL199 and pAL201 (Fig. 4). While the 4.4 Kb Aval fragment has one site of each restriction endonucleases, HindIII (lane E), BamHI (lane G) and SalI (lane I), as in the functional Tc r gene in pBR322 (Bolivar et al. 1977), the 5.2 Kb fragment has a deletion which includes both the BamHI (lane F) and SalI (lane H) sites and an insertion of 0.4 Kb in the HindIII site (lane D). Following transformation of ABl157 cells with pAL210, tetracycline resistant clones were isolated and their plasmids were used to transform DR100 cells. Restriction endonuclease analysis of plasmids isolated from the tetracycline resistant DR 100 clones indicate that these plasmids are indistinguishable from pBR322, as proposed in Fig. 3B (data not shown). Since pAL210 has only one origin of replication, intraplasmidic recombination would yield only one viable plasmid - pBR322.

Proficiency o f Interplasmidic and Intraplasmidic Recombination

For analysis of interplasmidic recombination, cells were transformed with both pAL199 and pAL201, and the transformed clones were maintained in the presence of ampiciline and chloramphemicol. For intraplasmidic recombination, cells were trans-

203

A EcoRI 1/~Dff-~

PvulI

/ ~L~5?r"b '¢

Pvu./_^, o? X\

~.'~ "" J "

k'

PvuT[ E c o R I

/~ / / ~

..... '

/ S~

f v'3.)'K; ^./ 5~rPst!-

~Pvull

APr." 7 /

,ind,,

~

/

•

p~__~.~\

,~m.

/

~ ~mL /

.~

RecombinationProduct

l'

~_"~_~'

'.N .ind..

/ 8 tOrt

Sali~-~-~

-,

Eco RI

_\ ~,~

^^r -r~r r~r

.. 3,6 \

.

II

7

/

BornHI .,'b'--- S - -

lava I

B Eco RI 1 Hind ]]I

PSt~Ava (~r,

~:~fKb

I

HindI T [ ~

Cm/ ~"-EcORI H'nd~'~ r -'- \ B~m__~ -

EcoRI

~

Avoi,~_....~

\Sol I.

Fig. 3A, B. Regeneration of a functional Tc r gene by interplasmidic and intraplasmidic recombination of two mutated Tc' genes. A functional Tc' gene is generated by interplasmidic recombination between pAL199 and pAL201 (A) or by intraplasmidic recombination in pAL210 (B), if the crossing-over site is between the mutations on the Tc' genes. The proposed structure of the recombination product is presented, Heavy line indicates DNA insertion and the sequence between mutations. • site of deletion

formed with pAL210 and maintained in the presence of ampiciline. Escherichia colt K-12 cells (ABl157) which have been transformed with both pAL199 and pAL201, or cells of the same strain which have been transformed with pAL210 were inoculated into liquid medium and grown at 3 7 ° C in the presence of ampiciline (pAL210) or ampiciline and chloramphemicol (pAL199 and pAL201). Samples were taken at time intervals for determination of the proportion of tetracycline resistant cells in the cultures (Fig. 5). Both in cultures which have been transformed with pAL199 and pAL201 and in those transformed with pAL210 the ratio of tetracycline resistant cells to ampicillin resistant cells remains almost constant throughout all the growth phases of the cultures.

The average ratio of tetracycline resistant cells to total cells was 2.3 x 10 _4 for cultures transformed with both pAL199 and pAL201 and 3.0 x 10 . 4 for cultures transformed with pAL210. While different ratios of tetracycline-resistant to tetracyclinesensitive cells have been observed in different strains (See Table 1, 2), in all strains tested, the proportion of tetracycline resistant cells in the culture remained constant throughout all growth phases. Recombination may have taken place in the growing culture, before plating, or, as has been shown to be the case in the intrachromosomal recombination system (Zieg and Kushner 1977), on the selective plates. The latter possibility could have led to an apparent constant ratio of tetracycline resistant to tetracycline sensitive cells in the culture, independently of the growth phase. To test this possibility, "respreading

204

A

B C

D

E

F G

H

I J

Table 2. Interplasmidic recombination of pAL199 and pAL201 Strain

Relevant genotype

Recombination Relative proficiency recombina(Tc r cells/ tion Ap' cells) proficiency a

-5.2 -4.4 -3.0 -1.4 --1.0 -0.8 -0.4

Fig. 4. Restriction endonuclease fragments of the product of interplasmidic recombination between pAL199 and pAL201. A recombination product which carries a functional Tc r gene was isolated and subjected to Aval digestion. The AvaI restriction products were separated by agarose gel electrophoresis (lane C), purified, and each product was digested with HindlII (lanes D, E), BamHI (lanes F, G), or SalI (lanes H, I). Digestion products of the 4.4 Kb AvaI fragment are presented in lanes E, G, I, and of the 5.2 Kb AvaI fragment in lanes D, F, H. Linear DNAs of pAL199, pAL201 and pBR322 are presented as size markers in lanes A, B and J, respectively

ABl157

-

2.3x10 -4(10)

JC2926

recA13 recB21 recC22 recB21 recC22 sbcA23 recB21 recC22 sbcA5 recA recB21 recC22 sbcA5

4 . 2 x 10 - 6 (8)

1.8 x 10 - 2

1.4 x 5.7 x 1.I x 2.9 x

0.6 2.4 4.8 (1.0) 0.13 (2.6 x 10-2) b

JC5519 JC8679 JC5183 DR107

10 .4 10 -4 10 .3 10 .5

(7) (7) (4) (7)

(1.0)

Results express the average of the indicated number of experiments, each one with an independent transformant. The number of experiments is presented in the parenthesis a Recombination proficiency/recombination proficiency in ABl157 b Recombination proficiency/recombination proficiency in the isogenic RecA ÷ strain in JC5183

I

I

I

I

I00

o

A

B ""

I0

c) 109

105

ioe

l04

I

v

I

0

107 ,

i

i 2

i 5

i 4

i 5

I

103

i

i

1234;-

i

Time (hrs)

Fig. 5A, B. Tetracycline resistant cells in growing cultures of AB1157 harboring pAL199 and pAL201 or pAL210. 10 ml LB cultures were inoculated with fewer than 100 cells harboring pAL199 and pAL201 (A) for analysis of interplasmidic recombination or with cells harboring pAL210 (B) for analysis of intraplasmidic recombination. After cell density had reached 107 cells/ml, samples were taken at time intervals, and the number of tetracycline resistant cells (e) and ampicillin resistant cells (o) was determined

I

I

I

2 5 Time(hrs)

I

4

Fig. 6. Respreading experiment. A sample was taken at to from a culture of AB 1157 harboring pAL210 at a logarithmic phase of growth. After dilution, 0.1 ml aliquots were spread on LB afar plates, supplemented with ampicillin or tetracycline to yield about 10 to 100 drug resistant colonies per plate, and the plates were incubated at 37 ° C. At the indicated time intervals, 0.1 ml phosphate buffer was added to the plates. The plates were respread, incubated overnight, and the number of tetracycline resistant colonies (e) and ampicillin resistant colonies (o) was determined

plates. These results suggest that recombination takes place in the liquid culture or immediately following plating. Recombination in recB recC and recA Mutants

e x p e r i m e n t s " (Zieg a n d K u s h n e r 1977) were conducted. Cultures h a r b o r i n g pAL210, at the logarithmic phase of growth, were diluted a n d plated on several ampicillin and tetracycline plates, which were incubated at 37 ° C. A t time intervals, plates were removed from the incubator, 0.1 ml buffer was added to the plates, the plates were respread a n d further incubated. The results presented in Fig. 6 indicate that growth of tetracycline resistant clones is initiated within one h o u r following plating, a n d proceeds at a rate similar to that of the clones on the ampicillin

Escherichia coli K-12 strains, carrying mutations in the recB and recC genes were transformed with pAL199 and pAL201,

or with pAL210. Cultures of the transformed clones were grown in L-broth in the presence of antibiotics as described above, a n d the ratio of tetracycline resistant to tetracycline sensitive cells was determined (Tables 2, 3). While in conjugation, mutations in the recB or recC genes lower the proficiency of recombinant formation by a b o u t two orders of magnitude (Clark 1973), these mutations h a d little effect on the proficiency of interplas-

205 Table 3. Intraplasmidic recombination in pAL210

Strain

Relevant genotype

Recombination Relative a proficiency recombina(Tc r cells/ tion Ap ~cells) proficiency

ABl157 REM199 JC2926 REM201 DM45 DR100 JC5459 JC5519 REM202 JC8679 JC5183 DR107

-

3.0 × 10 -4 (13) 3.3 × 10 -4 (4) 4.6 x 10 -6 (6) 7.5 x 10-6 (4) 2.3 x 10 6(4) 6.0 x 10 -6 (5) 6.0 x 10 .6 (2) 2.3 x 10 .4 (5) 1.1 x 10 .4 (4) 7.3 x 10-3 (2) 9.8 x 10 3 (15) 9.9 x 10 .3 (8)

recA13 recA1

recA99 recA recA13 recB21 recB21 recC22 recB21 recB21 recC22 sbcA23 recB21 recC22 sbcA5 recA recB21 recC22 sbcA5

1.0 1.1 1.5 x 10 2 2.5x 10 -2 7.6 x 10 -3 2.0 x 10 .2 2.0 x 10 2 0.76 0.36 24 32 (1.0) 33 (1.0) b

Results express the average of the indicated number of experiments, each one with an independent transformant. The number of experiments is presented in parenthesis a Recombination proficiency/recombination proficiency in AB1157 b Recombination proficiency/recombination proficiency in JC5183

midic or intraplasmidic recombination in the plasmids described in this paper. The proficiency of interplasmidic recombination in recB recC mutants is more than half of that in recB ÷ recC + strains (Table 2) and the proficiency of intraplasmidic recombination in the same mutants is one third (in REM202) to three quarters (in JC5519) of the recombination proficiency in isogenic strains carrying the wild type alleles of the recB recC mutations (Table 3). The proficiency of both interplasmidic and intraplasmidic recombination is lowered by recA mutations by a factor of 40150. Little difference was observed between the proficiencies of intraplasmidic recombination (Table 3) in strains carrying different recA mutations, with the possible exception of DM455 which carries an amber recA mutation (recA99) (Mount et al. 1976). Recombination proficiency in this strain is one third to one fourth of that in other recA mutants. Interplasmidic and intraplasmidic recombination products were isolated from recA cells and subjected to restriction endonuclease analysis as described above. These products were indistinguishable from recombination products in recA ÷ cells. Recombination via the recE Pathway

The major recombination pathway in recB recC sbcA E. coli strains is the recE pathway (Barbour et al. 1970). Conjugal recombination via the recE pathway depends on a functional reeA gene product. On the other hand, recombination of 2 phage D N A via this pathway is reeA independent (Gillen and Clark 1974). The relative proficiency of interplasmidic and intraplasmidic recombination in recB recC sbcA (JC8679 and JC5183) mutants is presented in Tables 2 and 3. Activation of the recE pathway by an sbcA mutation increases the proficiency of interplasmidic recombination two to five fold. On the other hand, the proficiency of intraplasmidic recombination is more than twenty fold higher in sbcA mutants than in sbcA + strains. Differences between the proficiencies of interplasmidic and intraplasmidic recombination are observed also in sbcA recA mutants: recA mutation reduces the proficiency of interplasmidic recombination via

the recE pathway forty fold. On the other hand, recA mutation has no detectable effect on the proficiency of intraplasmidic recombination in sbcA mutants. The proficiency of intraplasmidic recombination in a recA recB recC sbcA mutants (DR107) is the same as in the isogenic recB recC sbcA (JC5185) strain, thirty fold higher than in strains which do not carry the sbcA mutation, and more than a thousand fold higher than in sbcA ÷ strains carrying the same recA mutation (DR100). Discussion

We have described the construction of plasmids which facilitate direct determination of the proficiency of intraplasmidic and interplasmidic recombination. In this system, a recombination event leads to a detectable change in the bacterial phenotype from tetracycline sensitivity to resistance. The independence of the recombination process of external D N A transfer, the ease with which the substrates of recombination can be manipulated in vitro, and the convenience of the recombination assay may make this system a useful tool for both in vivo and in vitro studies of bacterial recombination. The constant frequency of tetracycline resistant cells in cultures harboring plasmids with a duplication of mutated Tc r gene, indicates that the recombination products have no selective advantage over the parent plasmid, and that cells harboring the recombination product do not have a selective advantage over the rest of the culture. This constant ratio depends on the genotype of the host (Tables 2, 3) and on the distance between the mutations (Laban and Cohen, in preparation). We propose that the ratio of tetracycline resistant to tetracycline sensitive cells in a culture is an index of the relative proficiency of recombination of the bacterial host strain. The system which is presented in this paper differs from conjugal recombination in the degree of dependency on host functions. A functional recA gene product is essential for recombinant formation in conjugation. The frequency of transfer of genetic markers from Hfr to F - cells is lowered by five orders of magnitude by a recA mutation in the recipient, and this residual transfer does not appear to be due to recombination (Clark 1974). On the other hand, a recA mutation lowers the proficiency of intraplasmidic and interplasmidic recombination only 50 to 100 fold, and the recombination products are indistinguishable from those obtained in recA + cells. Recombination in recA mutants may be due to a recA independent process, or to leakiness of the recA mutation. The observation that similar recombination proficiencies have been determined in all recA mutants tested makes the latter possibility unlikely and suggests that recombination takes place in these strains via a recA-independent pathway. The recA protein catalyzes ATP-dependent annealing of homologous D N A sequences (McEntee et al. 1979; Shibata et al. 1979). This protein may also be involved in the protection of nicked, gapped or single stranded D N A from degradation by nucleases (Marsden et al. 1974). While annealing is required for both processes, it is conceivable that in the absence of the recA protein, intermediate products in the D N A transfer stage of conjugation are more susceptible to degradation than intermediate products in the intraplasmidic and interplasmidic recombination process. This idea is consistent with the observation that incoming D N A in minicells of F - recA strains is susceptible to extensive degradation following conjugation (Khachatourians et al. 1975) and that a recA mutation in the recipient cell decreases the proficiency of F' transfer in conjugation about 10 fold (Gillen et al. 1981).

206 Mutations in the recB recC gene reduce the frequency of recombinant formation in conjugation about 100 fold (Clark 1973). On the other hand, little reduction in the proficiency of interplasmidic or intraplasmidic recombination is observed as a result of mutations in this gene. Recombination via the recBC pathway depends on a functional recB reeC gene product - exonuclease V (Goldman and Linns 1970) and is stimulated by the presence of Chi sequences on the recombining D N A (McMilin et al. 1974; Lam et al. 1974). Thus, the difference between the response of conjugal recombination and intraplasmidic or interplasmidic recombination to recB recC mutations may be explained, as it is explained in 2 recombination (Stahl and Stahl 1977) by the absence of Chi sequences from the plasmids. It would be of interest to determine the effect of Chi sequences, inserted into the plasmids on the proficiency of recombination via the recBC pathway. The difference in the degree of dependence on exonuclease V between conjugal recombination and intraplasmidic or interplasmidic recombination may also reside in the difference in the secondary structure of the recombination substrates. Taylor and Smith (1980) have recently proposed that the function of exonuclease V in recombination involves the formation of single strand loops which may serve as a substrate for the r e c A protein. The presence of single strand regions on supercoiled plasmids (Kato et al. 1973) may facilitate recB recC independent recombination. The s b c A mutation activates the recE recombination pathway by derepressing the synthesis of exonuclease VIII (Barbour et al. 1970). While the proficiency of interplasmidic recombination in recB reeC s b e A mutants is only 2-5 fold higher compared to the wild type strains, that of intraplasmidic recombination is 20-50 fold higher. Even more striking is the observation that intraplasmidic recombination via the recE pathway, but not interplasmidic recombination via the same pathway, is r e c A independent. Recombination of 2 phage via the recE or the red pathways is r e c A independent when D N A replication is not blocked (Gillen and Clark 1974; Stahl et al. 1979). The reason for the independence of intraplasmidic recombination of the recA function is not clear. It is possible that in this system, as well as in recombination via the red or the recE pathways (Stahl 1979) D N A replication intermediates serve as substrates for a r e c A independent recombination process. In the intraplasmidic recombination system, a transient pairing of homologous sequences cis to each other on the plasmid may facilitate r e c A independent recombination. The occurrence of single stranded D N A regions in supercoiled plasmids (Kato et al. 1973) may also contribute to this reaction. Acknowledgments. We thank J. Cohen, A.J. Clark, E. Metzer and

F.W. Stahl for bacterial and phage strains, Y.S. Halpern and E. Yagil for their critical review of this manuscript, and Z. Silberstein and E. Sheehter for collaboration on parts of this work. This work was supported by a grant from the United States-Israel Binational Science Foundation (BSF), Jerusalem, Israel.

References

Bachmann BJ (1972) Pedigree of some mutant strains of Escherichia coli K-12. Bacteriol Rev 44:1-56 Barbour SD, Nagaishi H, Templin A, Clark AJ (1970) Biochemical and genetic studies of recombination proficiency in Eseheriehia coll. II Rec + revertants caused by indirect suppression of R e c mutations. Proc Natl Acad Sci USA 67:128-135 Bolivar F (1978) Construction and characterization of new cloning

vehicles. III Derivatives of plasmid pBR322 carrying unique EcoRI site for selection of EcoRI generated recombinant DNA molecules. Gene 4:121-136 Bolivar F, Rodriguez RL, Green PJ, Betlach MC, Heyneker HL, B oyer HW, Crosa JH, Falkow S (1977) Construction and characterization of new cloning vehicles. II A multi-purpose cloning system. Gene 2:95-113 Brutlag D, Fry K, Nelson T, Hung P (1977) Synthesis of hybrid bacterial plasmids containing highly repeated satellite DNA. Cell 10:509-519 Clark AJ (1973) Recombination deficient mutants of E. coli and other bacteria. Annu Rev Genet 7:67-86 Clark AJ (1974) Progress toward a metabolic interpretation of genetic recombination of Eseherichia coli and bacteriophage Lambda. Genetics 78 : 259-271 Cohen A, Ram D, Halvorson HO, Wensink PC (1978) Deletions within E. coli plasmids carrying yeast rDNA. Gene 3:135-147 Cramer JH, Farrelly FW, Barnitz JT, Rownd RH (1977) Construction and restriction endonuclease mapping of hybrid plasmids containing S. eerevisia ribosomal DNA. Mol Gen Genet 151:229-244 Gillen J, Clark AJ (1974) The recE pathway of bacterial recombination. In: Grell RF (ed) Mechanism in recombination. Plenum Press, New York, pp 123-136 Gillen JR, Karn AE, Nagaishi H, Clark AJ (1977) Characterization of deoxyribonuclease determined by lambda reverse as exonuclease VIII of Eseherichia eoli. J Mol Biol 113:27-41 Gillen JR, Willis DK, Clark AJ (1981) Genetic analysis of the recE pathway of genetic recombination in Eseherichia eoli K-12. J Bacteriol 145 : 521-532 Goldmark P J, Linn S (1970) An endonuclease activity from Escherichia coli absent from certain ree strains. Proc Natl Acad Sci USA 67 : 434-444 Honigman A, Oppenheim AB, Hohn B, Hohn T (1981) Plasmid vector for positive selection of DNA inserts controlled by the 2P promotor, repressor and antitermination function. Gene 13:301-310 Kato AC, Bartok MJ, Frazer MJ, Denhart DT (1973) Sensitivity of superhelical DNA to a single-strand specific endonuclease. Biochina Biophys Acta 308:68-78 Khachatourians GG, Paterson MC, Sheehy R J, Dorp BV, Worthy TE (1975) DNA degradation in minicells of Eseheriehia eoli K-12. II Effect of recA and recB21 mutations on DNA degradation in minicells and detections of exonuclease-V activity. Mol Gen Genet 138:179-192 Kushner SR, Nagaishi H, Clark AJ (1972) Indirect suppression of recB and recC mutations by exonuclease I deficiency. Proc Natl Acad Sci USA 69:1366-1370 Lain ST, Stahl MM, McMilin KD, Stahl, FW (1974) Rec mediated recombinational hot spot activity in bacteriophage lambda. II A mutation which causes hot spot activity. Genetics 77:425 433 Lennox ES (1955) Transduction of linked genetic characters of the host by bacteriophage P1. Virology 1 : 190-206 McEntee K, Weinstock GM, Lehman IR (1979) Initiation of general recombination catalyzed in vitro by the recA protein of Escherichia coli. Proc Natl Acad Sci USA 76:2615-2619 McMilin KD, Stahl MM, StahI FW (1974) Rec-mediated recombinational hot spot activity in bacteriophage lambda. I. Hot spot activity associated with spi- deletion and bio substitution. Genetics 77 : 409425 Marsden HS, Pollard EC, Ginoza W, Randall EP (1974) Involvement of recA and exr genes in the in vivo inhibition of recBC nuclease, J Bacteriol 118:465-470 Metzer E, Levitz R, Halpern YS (1979) Isolation and properties of Escheriehia coli K-12 mutants impaired in the utilization of Vaminobutyrate. J. Bacteriol 137 : 1111-1118 Mount DW (1971) Isolation and genetic analysis of a strain of Escherichia eoli K-12 with an amber recA mutation. J Bacteriol 107 : 388-389 Shibata T, Dasgupta RP, Cunningham RP, Radding CM (1979) Purified Escherichia coli recA protein catalyzes homologous pairing of superhelical DNA and single-stranded fragments. Proc Natl Acad Sci USA 76:1638-1642

207 Stahl FW (1979) Genetic recombination, thinking about it in phage and fungi. W.H. Freeman and Co, San Francisco, p 204 Stahl FW, MeMilin KD, Stahl MM, Grosemann JM, Lain S (1974) The distribution of crossovers along unreplicated lambda bacteriophage chromosomes. Genetics 77 : 395-408 Stahl FW, Stahl MM (1977) Recombination pathway specificity of Chi. Genetics 86 : 715-725 Sugino A, Goodman HM, Heyneker HL, Shiner J, Boyer HW, Cosserelli N (1977) Interaction of bacteriophage T4 RNA and DNA ligases in joining of duplex DNA at base-paired ends. J Biol Chem 252: 3987-3994 Sutcliffe JG (1978) pBR322 restriction map derived from the DNA sequence: accurate DNA size markers up to 4361 nucleotide pairs long. Nucleic Acid Res 5:2721 2728 Taylor A, Smith GR (1980) Unwinding of DNA by the recBC enzyme. Cell 22 : 447 457

Weintraub H, Groudine M (1976) Chromosomal subunits in active genes have an altered conformation. Science 193:848-856 Zieg J, Kushner SR (1977) Analysis of genetic recombination between two partially deleted lactose operons of Escheriehia coli K-12 J Bacteriol 131 : 123 132 Zissler JE, Singer E, Shaeffer F (1971) The role of recombination in growth of bacteriophage lambda. I. The gamma gene. In: Hershey AD (ed) The bacteriophage lambda. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, pp 455-468

C o m m u n i c a t e d by F.W. Stahl

Received May 20/September 11, 1981