MGG
Molec. gen. Genet. 166, 337 346 (1978)
© by Springer-Verlag 1978
Scale of the Genetic Map and Genetic Control of Recombination after Conjugation in Escherichia coli K-12 Hot Spots of Recombination S.E. Bresler, S.V. Krivonogov, and V.A. Lanzov Leningrad Institute of Nuclear Physics
Summary. There exist many regions on the genetic map of E. coli, remarkable for very high frequency of genetic exchanges between the donor and recipient chromosome after conjugation. We call these regions fre (frequent recombination exchange). Two of them were localized: frel near to the gene tsx and fre2 adjacent to metB. The conjugational transfer of fre is characterized by high negative interference in the corresponding region of the map. The effect called Fre is genetically determined. It is slightly present on the Rec BC pathway of recombination and becomes drastic on the Rec F pathway. The effect is sharpened by an increase of temperature till 43 ° C during and after conjugation. The effect is absolutely dependent on the genes recA and recF. It is assumed that region fre contains many hot spots of recombination, i.e. sites of initiation, where a recF-dependent endonuclease starts the process. The scale of the genetic map of E. coli K-12 in the areas not including the fre regions is about 24 min both on the Rec BC and the Rec F pathways. In the regions including fre, the scale drops to 5 rain on the Rec BC pathway and to about 1 min on the Rec F pathway. These strong variations explain the discrepancies in the mapping distances found in different works. If a plasmid F' containing the fre region is transmitted during conjugation it becomes extremely unstable. A fragment of D N A containing the fre region is always lost from the plasmid. It leads to its shortening or sometimes to the killing of the cell. The Fre effect is seen also in P1 transduction. These facts pose many questions. Suggestive answers are discussed.
For offprints contact: Prof. S.E. Bresler, Leningrad Nuclear Physics Institute, Gatehina, Leningrad district 188350, U.S.S.R.
Introduction The scale of the genetic map )~, i.e. the average distance between two adjacent genetic exchanges is a fundamental characteristics of the recombination process. The first measurements of this quantity for the conjugation of E. coli K-12 by Jacob and Wollman (1959) gave a value )~=5 min, the entire map being equal to 100 min (Jacob and Wollman, 1959; Bachmann et al., 1976). Afterwards other workers used different approaches to measure the mapping distances and modified slightly the mapping function of Holdane. They found various values of )~ from 10 min (Verhoef and de Haan, 1966; Wu, 1967) till 20 min (Wood and Walmsley, 1969). The reason of these discrepancies remained obscure. In last years the genetic determination of recombination was successfully studied by Clark and coworkers (for ref. see Clark, 1974). The main determinant was found to be the recA gene. Its damage blocks recombination almost entirely. The recB and recC genes are coding for exonuclease V, a degradation enzyme (Buttin and Wright, 1968). Its specificity allows for its participation in recombination (Barbour and Clark, 1970; Karu et al., 1973; Lieberman and Oishi, 1974). Mutations in these genes reduce the yield of recombinants about 10-100 times. An additional mutation sbcB in the structural gene of exonuclease I (Kushner et al., 1972; Templin et al., 1972) which inhibits the activity of this enzyme leads practically to a total recovery of the yield of recombinants in cells of genotype recB recC- sbcB-. According to Clark and coworkers this is a token of the existence of a second recombination pathway characterized by different enzymes. The main genetic determinant of this alternative pathway is recF. Its damage reduces the recombinants yield about 100 times. It is widely accepted to call this additional pathway of recombinat i o n - the RecF pathway in contrast to the more corn-
0026-8925/78/0166/0337/$02.00
338
S.E. Bresler et al. : Scale of the Genetic M a p and Genetic Control of Recombination
mon RecBC pathway. The general assumption of two alternative recombination pathways is liable to criticism. But we shall use the same terminology throughout this paper because it facilitates the description of experiments and exposure of data. The aim of our work was to find the correct value of 2 and its dependence on the recombinational pathway.
Materials and Methods Bacterial Strains used in the work are listed in Table 1. Bacteriophage tsx.
T6 was used for the identification of the marker
Media (slightly modified minimal TPG, maximal AP, selective agar) and conditions of conjugation were described earlier (Bresler et al., 1968).
Table 1. Characteristics of bacterial strains used Strain
Genotype
Source
H
thi dnaBT43
R4
gal tonA
Dr. F. Bonhoeffer Dr. P.G. de Haan
JC 9247 (P4 x )
recF143 metB
Dr. A.J. Clark
Z 1035 (OR 11)
dnaBT43
Z 436 (OR 7)
his
C
prototroph
Dr. R. Curtiss III Dr. R. Curtiss III Dr. B. Low Dr. B. Bachmann
Hfr:
P K 191
thi sup56 (proA-lac)
K L 16
prototroph
R1
metB
Dr. B. Low Dr. T.S. Iljina
R 25
rif
Dr. T.S. Iljina
P 72
metB thi
Dr. F. Jacob
F':
Z 517
F' ORF-1 purE+-laC + in F - V(purE-lac)
Dr. R. Curtiss III
Mutagenesis by trimethoprim was affected according to the m a n u a l of Miller (1972)
F ' 94
F'lac in F
Dr. W. Hayes
Transduction by phage Plvir. The recipients were grown on the T P G minimal m e d i u m till the late exponential phase. Then CaCI; was added till a final concentration 5- 10 4 M and a phagolysate of Plvir prepared from Hfr C by the method of agar cultivation (the average a m o u n t of phage was 1 per bacterial cell). After 20 rain at 37 ° C the process was interrupted by the addition of 1 M Nacitrate. The initial phage Plvir was obtained from Dr. B. Wolf,
AB 1157
thr ara leu proA tsx lac his str argE thi
Dr. P. HowardFlanders
JC 7623
as AB 1157 but recB21 recC22 sbcB15
Dr. A. J. Clark
AB 2463
as AB 1157 but recA13
Dr. P. HowardFlanders
JC 5743
as JC 7623 but recB21
Dr. A.J. Clark
JC 5547
as AB 1157 but recB21 recC22 recA13
Dr. A.J. Clark
X 5036
lacZproC tsx trp str
Dr. F. Jacob
E C K 026
proC tsx purE str
this laboratory
E C K 033
as JC 7623 but arg + xyl + metB recF143 JC 9247 × JC 7623
E C K 034
as E C K 033 but thyA
prototroph
F-:
Experimental Approach The analysis of crosses was effected on the basis of the mapping function of Holdane: w = 1/2[1-exp(-2l/2)],where w is the probability of r e c o m b i n a t i o n , / - - t h e physical distance between the markers, 2 - - t h e scale of the map. In each experiment we measure /~--the linkage of a nonselected proximal marker with a selected distal one, i.e. the probability to find the nonselected marker a m o n g the recombinants for the selected one. By definition p = 1-w and this enables to compute the scale 2 if we know the distances l between the markers. The precision of our measurements will depend on the knowledge of the distances between the genetic loci in question (Bachmann et al., 1976). But we can obtain an average value of 2 by interpolation if we measure # for a set of nonselected markers. If we plot the function In 1 / ( 2 g - 1 ) versus l we must obtain a straight line with the slope equal to 2/2. On the other hand if we are comparing two different values of the scale 21 and 22 we can compare the linkages measured for the same pair of markers )~1/22 = 1 n ( 2 ~ 2 - 1 ) / l n ( 2 g l - 1). In this case the precision of the ratio does not depend of the accuracy of the distance between the markers studied. To obtain reproducible data for #, it is important to standardize all conditions of cultivation, conjugation and selection of recombinant clones. We noticed that the use of maximal media for conjugation and of minimal media for subsequent selection of recombinants gives significantly reduced values of the linkage #. To avoid this step down we performed all procedures of cultivation, conjugation and selection on minimal media supplied with all necessary additions. The optimal concentrations of amino acids in the media in case of auxotrophic strains were chosen according to recommendations of Verhoef and de H a a n (1966).
from E C K 033 by a trimethoprim mutagenesis
E C K 035
as E C K 033 but recBC ÷ K L 16 x E C K 034
E C K 036
as JC 7623 but thyA
E C K 037
as Eck 036 but his + sbcB +
PK 191 x E C K 036
E C K 038
as E C K 036 but rec +
P K 191 x E C K 037
E C K 039
as JC 7623 but his +sbcB +
Plvir(C) x JC 7623
E C K 040
as JC 7623 but lac + proC
Plvir(X 5036 lac ÷) x JC 7623
JC 817l
as JC 7623 but recL152
Dr. A.J. Clark
M O 61l
as JC 7623 but endoI trpVBC
Dr. M. Oishi
from JC 7623 by a trimethorpim mutagenesis
Strains JC 9247 and JC 8171 were kindly supplied by D r . G . B . Smirnov, strain JC 5547 by Dr. M.G.R. Hart, strain JC 7623 by Dr. M. Oishi
S.E. Bresler et al. : Scale of the Genetic Map and Genetic Control of Recombination
339
We shall denote the linkage g by two indices (the first one showing the marker of primary selection) and show in brackets the number of clones analyzed. For example #p,oA.m~=0.62(300) denotes the linkage of two loci proA and lac, in case when the former was used for primary selection. In this cross 300 recombinant clones were scored. The values of/~ were found averaging the results of 2 3 experiments. The statistical scattering is seen from the Figures.
~
~
O ¢'J
Results and Discussion Scale of the Genetic Map on Both Recombination Pathways. It was found that in some crosses the ideal
conditions are really fulfilled and we obtain a unique scale of the genetic map. In Fig. 2 a are sden the plots of/~ versus l for the selected distal marker ArgE + and a set of nonselected markers seen in Figure 1. The data were obtained from crosses of the donor Hfr R4 with the recipients AB 1157 rec + (Rec BC recombination pathway) or JC 7623 r e c B - r e c C - s b c B - (Rec F pathway). As expected a line is obtained which reflects changes in # beginning with total linkage ( # = 1) till total absence of linkage (#=0.5). If we plot In (i/2~_i) versus l, as deduced from the mapping function of Holdane, we obtain straight lines 1 and 2 (Fig. 2b). Therefrom a scale of the genetic map 2 = 2 4 min, identical for both recombination pathways is computed. Similar results were obtained for the donor P 4 x (data not presented). If we use for the crosses a donor Hfr C a different situation is encountered. Curve 3 in Figure 2a and b shows some departure from a constant mapping scale on the Rec BC pathway. Curve 4 in Figure 2a refers to the recombination on the Rec F pathway. It manifests a drastic difference with the regular function found in case of the donor Hfr R4. What is p e c u l i a r - t h e linkage /~ drops near the marker Tsx to a very low value, an order of magnitude lower than the limiting value 0.5 predicted by the mapping function. As Holdane's deduction was based on the conservation of genetic material, we must assume that in our case, when the linkage between the distal marker ArgE + and the proximal markers near Tsx is measured some genetic material in the region proAtsx is lost with a high probability and does not participate in subsequent recombination with a far away marker like ArgE +. How does this happen and what becomes of this D N A fragment adjacent to the gene tsx? We suggest that in this region of the map a hot spot of recombination is situated, i.e. a cluster of points on the chromosome which are recognized by the initiating endonuclease of recombination and readily cleaved. Practically one cut is always administered to D N A in this hot spot. The second cut is
Fig. 1. Genetic map ofE. coli (Bachmann et al., 1976). The markers used in this work are shown and the directions of transfer of the donors
G 1
0,75 1,2 0,5
0,25 4
tnl 1 /
r
i
i
i
b
2,u- 1 3
2
,'o I 11'5 12~]t(m~ol Arg E
Thr Leu ProC LacTsx
Fig. 2.a Dependence of the linkage (g) of the selected marker ArgE + and different markers (Thr +, Leu +, ProA *, Lac +, Tsx s) in crosses: 1) R 4 x A B 1157; 2) R 4 x J C 7623; 3) Hfr C x A B 1157; 4) Hfr C × JC 7623 on the distance between the markers (/). b Dependence of lnl/2~_1 on 1 for the same crosses
also quite probable in the same region. Then a small fragment of DNA, containing the respective markers is lost for recombination with the distal ArgE ÷. Sometimes the second cut is made outside the anomalous region, where the regular recombination mechanism is valid. These big D N A pieces may contain
340
S.E. Bresler et al. : Scale of the Genetic M a p and Genetic Control of Recombination
S
~s c E 8~ ff "S
~3 c 2 _E
I[ Is
I,1
,;
201
F4"oCLocProa Leu Thr Arg E Distance between markers (min)
Fig. 3. Gradient of transfer of the markers in crosses of the donor Hfr C with recipients of the Rec BC (AB 1157) and the Rec F (ECK 040) recombination pathways. After a 5 rain contact the conjugating mixture was diluted 1 : 500 by fresh preheated medium and incubated for 80 min at 37 ° C. Afterwards the mating pairs were disrupted and the cells incubated with aeration for 60 rain more for phenotypic expression of recombinants
ArgE + and yield the recombinants for ArgE + and the Tsx s marker. The small fragments excized from the hot spot may be cleaved or rather integrated into another chromosome of the same cell. The latter possibility is more likely because we are dealing with the strain ExoI E x o V - in which D N A degradation is slowed down. Another argument is that if we measure the gradient of transfer of markers along the chromosome we do not find any anomalous behaviour when the region proA-lac is transmitted. The yield of markers in both recipient cells is decreasing exponentially (Fig. 3). Hence we conclude that no irreversible loss of markers takes place. The donor Hfr C, in contrast with Hfr R4, transmits proximally a region of D N A with a very high frequency of genetic exchanges per unit length. We shall denote this region as fre (Frequent Recombination Exchanges) and call the whole phenomenon - Fre effect. The recombination probabilities of markers inside the l°re region were studied by normalizing the data according to a proximal marker. For this purpose two strains (JC 7623 and ECK 040) were used. The recipient ECK 040 was obtained by means of transduction of the marker P r o C - instead of Lac(see Table 1). The linkage #proA_lac=0.62(300) in a cross Hfr C x JC7623. A value 2 = 2 min is estimated, according to Holdane's expression. For the same pair of markers on the Rec BC pathway 2 = 5 min was obtained (cross Hfr C × AB 1157). Taking the nearest pair of markers ProC+Tsx s, a value of #;roC_tsx= 0.8(100) was measured in the cross Hfr C × ECK 040. The resulting value of 2 is 1 rain. Hence the evidence
for the Fre effect is of two kinds. First of all the Fre effect considered earlier is demonstrated by a very low probability to find the respective markers in a linkage group with another far away marker. Instead of the probability # = 0 . 5 , # = 0 . 1 is found which is a drastic alteration. Secondly inside the fre region an extremely small scale of genetic recombination is encountered. Adjacent genetic markers are disjoined and the linkage of neighbouring genes is disrupted. This is called generally-negative interference. The hot points are not equally distributed. Near tsx the genetic exchanges are predominant. Therefore we can use the Poisson distribution in this region only as a rough estimate. And in the vicinity of tsx the genetic scale becomes extremely short and the exchange extraordinary frequent. So the Fre effect exhibits pronounced polarity. We considered up to now the Fre effect on the Rec F pathway of recombination. From Figure 2a we see that it is present to some extent on the Rec BC pathway. But it is so weak that we shall use the corresponding crosses as control devoid of anomalities. We shall introduce a correction into the mapping function of Holdane to allow for the loss of genetic linkage in the hot spot or fre region. We shall introduce the correction into the main expression ~t= 1 w = 1/2(1 + e -2l/'~) (1 --(p). (p is the loss or translocation of genetic material in the fre region. We suggested that thefre region contained many points of initiation for the endonuclease of recombination. If they are situated at random inside this small region, then the distribution function of the small fragments of D N A excised by recombination from the hot spot will confirm Poisson's law: dw/dlAu = 1~co.e -IA"I~, where lab is the distance between markers A and B, co - the average value of the excised fragment. Hence the correction ~0 for the loss of genetic material will be: (p = a . e -tA"/°~= a. e L-l/~o. L is the distance from the hot spot tsx till the selected marker ArgE +. Therefore the distance of the recombining marker l=L--lAB, is a constant. Finally we obtain the corrected expression for the mapping function: g = */2(1 +e -2l/;) (1 -a.e-L-t/°°). To compare it with experiment we shall use the data of Figure 2a, where a primary selection for ArgE + marker was performed. We shall take as a matter of fact that the fre region is near the marker Tsx ~ (this will be proved later) and denote L as the distance between ArgE + and Tsx s. In Figure 2a an ideal curve for # is drawn (the curve 1,2 with dotted continuation) which tends exponentially to 0.5. If we divide the real value of #, plotted in Figure 2 a (curve 4) by the ideal value (curve 1,2), then we obtain the correction factor I-q) as function of L-I. If we plot
S.E. Bresler et al. : Scale of the Genetic Map and Genetic Control of Recombination
the value of lnqo calculated from the data versus L-l, we obtain a straight line as expected. We find that a=0.9. That makes the maximal correction near tsx 1-cp=0.1 and the parameter o9=2 min. This value was obtained in case when the frequency of recombination was normalized to a distal marker ArgE +. We have also data for recombination near to the hot spot, normalized to an adjacent marker in the region proA-lac. As already stated the average value o f / t in this region is 2 min in good accord with the value of co. Therefore we think that our description of the phenomenon is correct. The donor Hfr C transmits its c h r o m o s o m e anticlockwise (Fig. 1). But the fre effect is easily seen in case of a donor with an oposite direction of transfer, provided it contains t h e f r e region. We performed the cross of Hfr H with the recipient JC 7623. In this case we selected for two markers on the flanks of t h e f r e region, i.e. for Pro + and Gal + (see Fig. 1). The marker Tsx ~ is in between and it belongs to the hot spot. What happens is that about 70% of the recombinants Pro ÷ Gal ÷ Str" do not contain the intermediate marker Tsx ~. The experimental value l ~ p r o A g a l _ t s x = 0 . 3 0 ( l O 0 ) . In a control cross Hfr H x AB 1157 the corresponding figure /~proAg~l_ts~=0.90(100). Thus all experiments described demonstrate the Fre effect, the loss of genetic linkage and very probable cleavage of the c h r o m o s o m e at the hot spot of recombination, if primary selection is effected for the markers on the flanks. Localization o f the Fre Region on the Map. The pro-
nounced polarity of the Fre effect allows its localization. It must be situated somewhere between the point of origin of transfer of Hfr C c h r o m o s o m e (13 min on the E. coli map) and the locus tsx (9 min). More precisely its localization can be effected by means of the donor Hfr O R 7, for which Tsx s is the initially transferable marker. For the control cross O R 7 x A B 1157 the measurement of linkages gives lhe,-proA = 0.73(200) and # l e u - l a c = 0.65(200). But in case of cross O R 7 × JC 7623 the linkages are #~e,_proA= 0.67(200) and #~e,-~ac=0.35(200) correspondingly. The last figures are the result of a strong polar Fre effect on the Rec F pathway of recombination. Therefore we are aware that the f r e region is localized between the point of origin of transfer of the donors Hfr O R 7 c h r o m o s o m e and the locus tsx. We shall call this hot spot f r e l because we found similar regions in other parts of the map. Genetic Control o f the Fre Effect. As we already mentioned the Fre effect is fully manifested on the Rec F pathway of recombination, i.e. when we go from a rec + strain to r e e B - recC s b c B - . To be sure that
341
Table 2. Linkage (la) of markers near to t h e f r e region as a function of the recipients genotype (cross with the donor Hfr C) Recipients Strain
Linkage Genotype
]gleu-Iac
]glac-tsx
AB 1157
rec +
0.67 (700)
0.80 (100)
JC 7623
r e c B 2 l recC22 sbcB15
0.30 (700)
0.30 (1oo)
ECK 038 rec + s b c B +
0.65 (300)
0.80 (loo)
JC 8171
recB21 recC22 s b c B l 5 recL152
0.25 (200)
MO 611
recB21 recC22 sbcB15 endI
0.32 (200)
no other mutations are responsible for this difference we performed a control experiment. We converted a widely used recipient JC 7623 into a rec ÷ to switch it to the alternative Rec BC pathway. For this purpose a thy mutation was induced in the strain JC 7623 by mutagenesis (resulting strain E C K 036). Then by means of two consecutive crosses with the donor Hfr P K 191 we changed the genotype of the recipient E C K 036, selecting for his+sbcB ÷ (strain E C K 037) and finally for t h y + r e c B + r e c C + ( E C K 038). The Fre effect was registered by the lack of linkage of the selected marker Leu ÷ and the nonselected Lac + which is near to the f r e region. F r o m Table 2 we see that the recipient E C K 038 is characterizied by the same linkage of both genes as a usually employed wild type rec + (AB 1157). The recipient of the Rec F pathway (JC 7623) differs from the strains of the Rec BC pathway by two defects in the exonucleases V and I. We studied also the intermediate case E x o V - (strain E C K 039 obtained by transduction of JC 7623 to the his+sbcB + genotype) and found that the Fre effect was still visible but weak. In this case we deal with the residual recombination in cells of the Rec BC pathway. Therefore we were obliged to study recombination on a maximal medium and the absolute figures were changed. In the cross Hfr C x E C K 039 we obtain #leu.lac----0.46(600), in the control Hfr C x J C 7623 #l~,.io~,=0.27(400). Hence the absence of both ExoV and ExoI is important for the demonstration of the Fre effect. The introduction of a r e c F - mutation into the strain JC 7623 (resulting strain E C K 033) leads to a marked inhibition of the Fre effect. In this case the experiment was also performed on a maximal medium. The cross Hfr C × E C K 033 gives /hac_~sx= 0.65(100), in the control Hfr C x A B 1157 #Zac_tsx= 0.75(100). F r o m Table 2 we see also that a genetic damage
342
S.E. Bresier et al. : Scale of the Genetic Map and Genetic Control of Recombination
of the gene recL and endI (coding for endonuclease I) on the Rec F pathway of recombination are indifferent for the Fre effect. Thus the products of four genes (recF, recB, recC, sbcB) determine the Fre effect. The first one is indispensable, three others must be inactivated.
Table 3. Linkage of markers (/0 Leu+ and Lac + measured in crosses of the donor Hfr C with the recipients on the Rec BC and Rec F pathways of recombination at a normal and increased temperature
Dependence of the Fre Effect on Temperature, Independence of Protein Synthesis. According to data of Lieberman and Oishi (1974), the enzyme ExoV is partly inactivated in vitro at temperatures higher than 40 ° C. It is probable that the same properties are peculiar for this enzyme in vivo. In Table 3 we present data on the Fre effect at elevated temperatures. We see that on the Rec BC pathway the Fre effect is strongly enhanced and on the Rec F pathway exempt from ExoV it increases too, if the temperature is increased till 43 ° C during conjugation and 2 h of postconjugational period. Hence the increased temperature is favourable for the action of enzymes responsible for the Fre effect. To verify if the Fre effect is dependent on a constitutive or inducible enzyme, an experiment with conjugation of Hfr C x JC 7623 was performed on a T P G medium supplied with chloramphenicol (100/~g/ml). After 60 rain conjugation the cells were washed on a millipore filter and plated. The linkage frequencies were practically identical in the control (#leu_lac = 0.27(200) and in the run (#le,-lac = 0.16(200) showing that the Fre effect does not depend of de novo protein synthesis.
Conjugation
Incubation on pIates after conjugation
Growth of recombinants
AB 1157
JC 7623
37 43 43
37 37 43
37 37 37
0.65 0.50 0.37
0.27 0.17 0.11
The Gene recF and the Scale of the Genetic Map. Because the gene recF controls the Fre effect it was important to measure the genetic scale 2 in conditions of eliminated Fre effect, i.e. using a recipient recB+reeC+recF-. To design the desired strain we introduced into the strain E C K 033 a mutation thyA and after a short conjugation with the donor Hfr K L 16 selected for the recombinant E C K 035 (thy + reeB + recC + recF- sbcB-). The latter was used as recipient in crosses with the donor Hfr C. The results are presented in Table 4 and compared with crosses on the Rec BC and Rec F pathways. The analysis was limited to the region leu-lac of the map. What we found using Holdane's function and the data for leu-lac linkage was an increase of the mapping scale till 2 = 2 7 min. On the other hand it is just about the size of the transfered D N A piece. We suggest that in case of the recB+recC +recF strain recombination takes place only near to the edges of the transferred D N A piece without any endonuclease cleavage in the middle of the transferred D N A fragment. If so Holdane's mapping function must not conform with the recombinational events. The no-
Temperature during
#leu-lac
in case of recipients:
The value # was found by means of analysis of 200 clones
In ~P/o 0
0,5
1
1,5
I )
IX
~
IL-I
I;
~sx LQc ProA Leu i'hr {rain) Fig. 4. Dependence of the correcting function (0 on the distance (L--/) between the corresponding marker and the hot spot of recombination
Table 4. Linkage of markers in the crosses of Hfr C with recipients on different recombination pathways Recipient
Number of clones analyzed
AB 1157 800 AB 1157 400 JC 7623 800 JC 7623 400
Linkage of markers Xyl +
ArgE ÷
Leu +
ProA +
Lac +
1 1
1 0.59 1 0.50
0.64 0.31 0.52 0.24
0.58 0.22 0.46 0.19
0.53 0.19 0.21 0.09
The yield for the selected marker is taken as unit
tion of the mapping scale 2 must be abandoned because recombination in a r e c F - recipient depends entirely on the size of the transfered D N A piece. This was proved in three crosses of donors O R 11, R4 and P4 x with recipient E C K 035 recF-. The linkage of markers MetB ÷ and Leu +, #,,et~-~e, was found correspondingly equal to 0.81(300), 0.84(300) and 0.94(300). All three donors have the same point of
S.E. Bresler et al. : Scale of the Genetic Map and Genetic Control of Recombination origin but the two former are recF + and the latter is recF-. They transmit Leu + on the 4 rain, MetB + on the 18 min, recF on the 24 rain. All three figures for g are amazing. They would give according to Holdane a scale 2 much bigger than the c h r o m o s o m e length. An extreme case is the cross with Hfr P4 x because both parental cells are exempt from recF gene. In case of the donors O R 11 and R4 the recF + gene was transmitted into the recipient and influenced the recombination to some extent. Still the main result of all three crosses is practically the s a m e - t o t a l absence of recombination exchanges on a big piece of the chromosome. The recombination in this particular case is taking place at the edges of the transferred D N A fragment and the latter is integrated as a whole into the recipient chromosome. N o w we see the meaning of the recF-product. It is evidently an initiating endonuclease or some controlling subunit of the latter. It makes intermediate splits in the transmitted piece of D N A and it recognizes specifically the cluster of hot points and cleaves readily the fre region. The recF product must be extremely important for recombination.
The Region fre2. It is improbable that the hot spot of recombination frel is unique in the E. coli chromosome. We looked for other hot spots and found a second one, calledfre2 near the gene metB. The effect of fre2 is easily seen if we use Hfr C and Hfr R4 as donors. In Table 4 are presented data on the linkage of the selected marker Xyl + with a set of nonselected markers. For comparison in the lines 1 and 3 of the Table values of g are given if the selection is effected for ArgE +. We see in case of the donor Hfr C that the linkage is very much decreased if we pass from primary selection for ArgE + to selection for Xyl + . The difference is obvious even on the Rec BC pathway of recombination and it is much stronger on the Rec F pathway. In case of the donor R4 the linkage #~yt-arg~ drops from 0.73(400) till 0.56(400) and/~xy~-t~r from 0.62(400) till 0.50(400) if we go from a recipient on the Rec BC pathway to a recipient on the Rec F pathway. On the other hand after primary selection for ArgE + the linkages on both pathways are identical. It is clear that fre2 is localized between the markers ArgE ÷ and Xyl ÷. A more accurate localization is made by means of donor P72, which shows the Fre2 effect and demonstrates its strong polarity (Table 5). If we measure the scale of the genetic m a p using a small region near the hot spot, i.e. between argEmetB, we find 2 = 1 . 2 min. (By the cross Hfr R1 metB- x JC 7623 ; ]AargE_metB 0.69(200). If we measure the same linkage on the Rec BC pathway we obtain =
343
Table 5. Linkage of markers in crosses of the donor P 72 metB with recipients on different recombination pathways Recipient
AB 1157 JC 7623
Linkage of markers MetB-
ArgE-
Thr +
Leu +
0.42 0.12
0.53 0.34
0.92 0.81
1 1
The linkage was found by means of analysis of 300 clones
)~=9.4 rain (cross Hfr R1 metB- x A B 1157, [largE.metB =0.91(200). For the neighbouring part of the c h r o m o s o m e argE-leu, that does not contain fre2, the scale of the map is )~---24 rain. The scale 2 is shortened 20 times on the Rec F pathway and about 2 times on the Rec BC pathway. The ~'e2 region is strongly dependent on recF. This circumstance was found in crosses of the donor R 25 with the recipient E C K 033 (data not presented).
Investigation of the Fre Effect by means of the long F' Plasmid. The long F ' plasmid ORF-1 transmits a c h r o m o s o m e fragment of E. coli with the sequence of genes purE-fi'el-tsx-proC-lac. The short plasmid F'-lac transmits only a short fragment around the lac operon. As a rule the transmisssion of the plasmid is not accompanied by recombination of the genetic material. The transconjugants are secondary F'-containing strains. We studied the genetic characteristics of a plasmid ORF-1 transmitted into a recipient with very outspoken Fre effect (JC 7623) and for the sake of control into a recipient with slight Fre effect (AB 1157). We found a drastic difference in the results, a new manifestation of the Fre effect. The donor and recipient cells were cultivated on the AP maximal medium and mixed in the same medium in the ratio 1:1. The transconjugants were plated on EMB-lac maximal agar, supplied with streptomycin for the counterselection of donors. The coloured clones allowed the scoring of Lac + Str r transconjugants. The efficiency of conjugation was measured as the ratio of these Lac+StF clones to all recipient clones on the plates. To normalize the figures we measured the same ratio in case of the short plasmid F'-lac which is known to be transmitted in a very efficient way. All results are presented in Table 6. We see first of all that after the conjugation ORF-1 x JC 7623 the yield of recipient cells is 2-3 times lower than in case of the control plasmid F'-lac. The yield of transconjugants is also proportionally reduced. Nothing of the kind is seen if we observe conjugation ORF-1 x AB 1157 where the Fre effect is virtually absent. So a considerable inhibition of growth or killing of the cell population takes place.
344
S.E. Bresler et al. : Scale of the Genetic M a p and Genetic Control of Recombination
Table 6. Relative yield of recipient cells and transconjugants Lac + S t r ~. Inheritance of different markers after conjugational transmission of plasmid ORF-1 into the recipients on the Rec BC or the Rec F pathways of recombination
Table 7. Inheritance of the marker Tsx s in transconjugants after crosses of different recipients with the plasmid bearing strain ORF1 Recipient
Recipient
AB 1157 JC 7623
Normalized yield Re cipients
Transconjugants
1.0 0.3-0.5
0.79 0.25
Inheritance a m o n g the Lac + Str ~ transconjugants of the markers PurE +
Tsx s
ProC +
1.0 (60)
1.0 (100) 0.13 (70)
1.0 (60)
The yield was normalized as the ratio to the n u m b e r of F'-lac transconjugants
Strain
Genotype
AB 2463 JC 5743 E C K 039 JC 5547 E C K 033
recA13 reeB21 recB21 recC22 recB21 recC22 recA13 recB21 recC22 sbcB15 recF143
Normalized yield of transconjugants
Inheritance of the marker Tsx~ a m o n g the transconjugants
0.84 0.52 0.80 0.50 0.47
1.0(50) 0.12(50) 0.07(100) 1.0(50) 1.0(50)
But the most striking p h e n o m e n o n was the almost universal loss of the Tsx S marker from the transmitted plasmids. To demonstrate the loss of tsx we verified the sensitivity of the transconjugants to the phage MS2 and used them as donors in a new conjugation with the recipient strains X 5036 L a c - P r o C - T s x r and E C K 026 L a c - P u r E . To effect counterselection of donors F ' ( J C 7623) we used the amino acid auxotrophies of the strain JC 7623. We registered the transmission of the ORF-1 plasmid by means of the Lac +, Pro + or P u r e + characters. In the same time we looked specifically for the Tsx s marker. Most of the secondary transconjugants were Tsx r because the marker Tsx S was irreversibly lost. A small fraction of the transconjugants had preserved the Tsx ~ marker intact. After repeated conjugation of plasmid strains F ' ( X 5036) taken from separate clones belonging to this small fraction with a recipient JC 7623 the Fre effect could be observed once more. The Tsx ~ locus was excized with a high probability. Obviously the plasmid is shortened by this piece of D N A situated in the vicinity of the hot spot. In some cases this act is even lethal for the cell. Probably the corresponding plasmid is degraded and loses its markers totally. On the Rec BC pathway of recombination the loss of the Tsx s marker is observed no more than in 1% of the transconjugants.
is entirely eliminated. In all respects the conjugating strains behave normally and the plasmid ORF-1 is transmitted with high efficiency to the recipient. Hence we see that the Fre effect in plasmids is a case of recA-controlled recombination which is totally
Genetic Control of the Fre Effect in Plasmids. The plasmid experiment enables us to confirm the genetic determination of the Fre effect and specifically to study the role of the recA gene. As we have already mentioned the recB + recC + recipient cells do not manifest the Fre effect (Table 7). The damage in ExoV by means of one (recB-) or two (recB- recC-) mutations makes the Fre effect evident. But if we introduce into this strain exempt from ExoV the mutations recA- or recF- the Fre effect (loss of the Tsx s marker)
The Fre effect, the finding of several hot spots of recombination, where the genetic exchanges become extremely frequent poses m a n y new problems. First of all it would be important to know these regions on the m a p of E. coli. Two of them near tsx and metB were already discussed. We have preliminary data concerning a third fre region in the interval trphis. H o w m a n y more can be found is to be seen. The second question is about their chemical structure. We must mention that L a m et al. (1974), Stahl
recF-dependent. Fre Effect during Transduction. It was interesting to look into the manifestation of the Fre effect in Pl transduction because in case of generalized transduction it was established that double-stranded D N A pieces are transmitted into the recipient (Ebel-Tsipsis et al., 1972). The experiment was performed as follows. A strain E C K 040 was designed as recipient. We could compare in the same run the linkage of two pairs of markers Leu +-Ara + (distance 0.3-0.4 rain; Bachmann et al., 1976) and ProC+-Tsx S (0.3 min), the former situated far from the fre region, the latter near to the fre region. We obtained the values #~...... =0.87(300), iZproC_t~x=0.42(lO0). The first pair behaves as totally linked, the second as practically unlinked. This difference is probably a witness of the Fre effect. Of course the dependence on the recF gene must be studied to make sure that this difference has no other origin.
General Discussion
S.E. Bresler et al. : Scale of the Genetic Map and Genetic Control of Recombination et al. (1975), Stahl and Stahl (1975) described some artificially induced mutations in phage 2 and later in E. coli (Malone et al., 1976) which behaved like initiation sites of r e c A - d e p e n d e n t recombination. They were called chi and m a p p e d on the E. coli chromosome. They are covering the c h r o m o s o m e rather uniformly with average distances about 10-15 genes. These chi sites manifest a negative interference, they are absolutely dependent on the r e c B C function and independent of recF. Hence we see that these artificially introduced mutations have some features in c o m m o n with the natural f r e regions of the cell, but are essentially different. We can hypothesize that structurally the hot spots are composed of IS sequences. In favour of this statement we can quote the fact that F factor is intergrated in sites overlapping with the f r e regions found by us. The third and mostly important question is the genetic determination of the Fre effect. As we told it is absolutely recA and r e c F - d e p e n d e n t and it is inhibited by the r e c B C function. The role of the r e c F gene in recombination is remarkable. It appears that in a r e c F - strain the Fre effect is cancelled and the transmitted donor c h r o m o s o m e is integrated as one piece without intermediate genetic exchanges. Being r e c F - the cell integrates only very big pieces of donor DNA. Very interesting is the inhibitory action of the r e c B C product on the Fre effect. Why ExoV plays such a strange role? This enzyme has both exonucleolytic activity (for one- and double-stranded D N A ) and endonucleolytic as well (only for one-stranded D N A ) (Karu et al., 1973). In the presence of the DNA-binding protein the exonuclease function of this enzyme is inhibited and it effects unwinding of the D N A double helix (Mackay and Linn, 1976) - a situation where ExoV can act as an initiating endonuclease. We suggest that ExoV acts as an efficient endonuclease of recombination mainly near the ends of the transmitted piece of D N A because the ends are eventually one-stranded. In the absence of the r e c F product ( r e c F strain) this is the only recombination process in the cell. But in the presence of the r e c F endonuclease the donor D N A is cleaved into m a n y pieces which integrate independently. And the r e c F product has a high affinity towards the hot spots in the Fre regions of the chromosome. It is feasible that the ExoV enzyme is unable to make breaks inside the donor D N A at a perceptable rate but is competing with the r e c F product for the sites of attachment. This kind of competitive inhibition, quite c o m m o n in enzymology can explain the smothering of the Fre effect in rec + cells. N o w we would like to emphasize the enormous complications in the mapping of markers on the chro-
345
mosome. It is quite natural that very different scales of the m a p were found by different research workers. Another point that deserves explanation is the stability of the genetic material in the hot spot or Fre region of the cell. Obviously the attack by the endonuclease of recombination (suggestively the recFproduct) is limited only to the moments of conjugational transfer. It can be argued that the one-stranded structure of the transferred D N A chain is the general reason for the existence of the Fre effect. But the preliminary results of transduction are against this explanation. The possibility of enzymes induction by the act of conjugation must be rejected because of experiments with chloramphenicol. We suggest that the explanation is the proteinized chromatin-like structure of the bacterial chromosome and plasmid (Griffith, 1976). Only during transfer by conjugation the chromatin-like substance is depleted of proteins and is liable to D N a s e attack. The same must be true for the general stability of genetic material in the cell. The multiple chromosomes of the bacterial cell do not recombine but the recently synthesized D N A is able to sister exchange probably because it is free of proteins. The specific role of the recF product in effecting intermediate genetic exchanges along the chromosome is in perfect accord with the findings of Ganesan and Seawell (1975) concerning postreplicative repair of UVdamaged D N A by sister exchange. These authors found a strong dependence of recombinational repair on the integrity of the r e c F gene.
References
Bachmann, B.J., Low, K.B., Taylor, A.L.: Recalibrated linkage map of Echerichia coli K-12. Bact. Rev. 40, 116 (1976) Barbour, S.D., Clark, A.J.: Biochemical and genetic studies of recombination proficiency in Escherichia coll. I. Enzymatic activity associated with recB + and recC + genes. Proc. nat. Acad. Sci. (Wash.) 65, 955 (1970) Bresler, S.E., Lanzov, V.A., Lukjaniec-Blinkova, A.A.: On the mechanism of conjugation in Escherichia coli K-12. Molec. gen. Genet. 102, 269 (1968) Buttin, G., Wright, M.R.: DNA breakdown in Escherichia eoli. Cold Spr. Harb. Symp. quant. Biol. 33, 259 (I968) Clark, A.J. : Recombination deficient mutants of E. coli and other bacteria. Ann. Rev. Genet. 7, 67 (1974) Clark, A.J. : Progress toward a metabolic interpretation of genetic recombination of Escherichia coli and bacteriophage lambda. Genetics 78, 259 (1974) Ebel-Tsipsis, J., Fox, M.S., Botstein, D. : Generalized tranduction by phage P22 in Salmonella typhimurium. II. Mechanism of integration of transducing DNA. J. molec. Biol. 71, 449 (1972) Ganesan, A.K., Seawell, P.C.: The effect of lexA and recF mutations on post-replication repair and DNA synthesis in Escheriehia coli K-12. Molec. gen. Genet. 141, 189 (1975) Griffith, J.D.: Visualization of prokaryotic DNA in a regularly
346
S.E. Bresler et al. : Scale of the Genetic Map and Genetic Control of Recombination
condensed chromatin-like fiber. Proc. nat. Acad. Sci. (Wash.) 73, 563 (1976) Jacob, F., Wollman, E.L.: Sexuality and the genetics of bacteria. New York and London: Academic Press 1961 Karu, A.E., MacKay, V., Goldmark, P.J., Linn, S.: The recBC deoxyribonuclease of Escherichia coil K- 12. Substrate specificity and reaction intermediates. J. biol. Chem. 248, 4874 (1973) Kushner, S.R., Nagaishi, H., Clark, A.J.: Indirect suppresiion of recB and recC mutations by exonuclease I deficiency. Proc. nat. Acad. Sci. (Wash.) 69, 1366 (1972) Lam, S.T., Stahl, M.M., McMilin, K.D., Stahl, F.W.: Rec-mediated recombinational hot spot activity in bacteriophage lambda. II. A mutation which causes hot spot activity. Genetics 77, 425 (1974) Lieberman, R.P., Oishi, M.: The recBC deoxyribonuclease of Escherichia coli: Isolation and characterization of the subunit proteins and reconstitution of the enzyme. Proc. nat. Acad. Sci. (Wash.) 71, 4816 (1974) Mackay, V., Linn, S. : Selective inhibition of the DNase activity of the recBC enzyme by the DNA binding protein from Escherichia coll. J. biol. Chem. 251, 3716 (1976) Malone, R.E., Stahl, M.M., Chattoraj, D.K., Stahl, F.W.: Chi recombination hot spot sites. In: Bacteriophage meeting, Cold Spring Harbor, N.Y.: Cold Spring Harbor Laboratory 1976
Miller, J.H. : Experiments in molecular genetics. Cold Spring Harbor, N.Y.: Cold Spring Harbor Laboratory 1976 Stahl, F.W., Crasemann, J.M., Stahl, M.M.: Rec-mediated recombinational hot spot activity in bacteriophage lambda. III. Chi mutations are site-mutations stimulating rec-mediated recombination. J. molec. Biol. 94, 203 (1975) Stahl, F.W., Stahl, M.M. : Rec-mediated recombinational hot spot activity in bacteriophage lambda. IV. Effect of heterology on chi-stimulated crossing over. Molec. gen. Genet. 140, 29 (1975) Templin, A., Kushner, S.R., Clark, A.J. Genetic analysis of mutations indirectly suppressing recB and recC mutations. Genetics 72, 205 (1972) Verhoef, C., de Haan, P.G. : Genetic recombination in Escherichia coli. I. Relation between linkage of unselected markers and map distance. Mutation Res. 3, 101 (1966) Wood, T.H., Walmsley, R.H. : Conjugation in Escherichia coil K-12 and its modification by irradiation. Biophys. J. 9, 391 (1969) Wu, T.T.: Recombination frequencies of proximal markers in bacterial conjugation. J. theor. Biol. 17, 40 (1967) Communicated
by D. Goldfarb
Received April 10, 1978