reversion of the gal3 mutation of escherichia coli: partial deletion of the insertion sequence

Melee. gen. Genet. 142, 263--275 (1975) © by Springer-Verlag 1975

Reversion of the gal3 Mutation of Escherichia coli: Partial Deletion of the Insertion Sequence

Asad Ahmed and Eric Johansen

Department of Genetics, University of Alberta, Edmonton, Alberta T6G 2E9, Canada

Received October 6, 1975

Summary. The gal3 mutation of E. cell is an insertion of a DNA sequence, 1,100 base pairs in length, into the operator-promoter region of the galactose operon. This mutation reverts spontaneously to gal + by excision of the insertion to produce stable, inducible revertants, or by tandem duplications of the gal operon to produce unstable, constitutive revertants. The nature of a third class of revertants, which are stable and constitutive, is the subject of the present study.

The stable, constitutive class of revertants included approximately 30% of all gal + revertants obtained from a gal3(2) strain. Although the constitutive reversions could be transduced by ~, the efficiency was found to be extremely poor and the rare transductants which did appear seemed to originate from abnormal transducing particles. I t was concluded that these reversions were not normally packaged by 2.

In order to facilitate the packaging of these reversions, the chlD-pgl region was deleted from the parent gal3(Z) strain. Unexpectedly, the gal3 mutation in the majority of these deletions reverted to produce stable, constitutive reversions exclusively. The explanation proposed was that the chlD-pgl deletions had also removed part of the gal operator-promoter up to the gal3 insertion, so that simple excisions of the insertion yielded stable, constitutive revertants by connecting the gal structural genes to a different promoter. These revertants were not considered to be true representatives of the stable, constitutive class. The specificity of deletion end-points at the insertion was found only in the gal3(~) strain, and not in gal+, gal+(X), or gal3 strains. Moreover, the frequency of spontaneous chlD-pgl deletions increased 10- to 15-fold in presence of the gal3 insertion.

A 2gal phage bearing a true stable, constitutive reversion (galc200) was isolated from the revertant strain by subsequent deletion of the chID-pgl segment (LI31). Electron micrographs of ~gal+ and 2gale200 A31(chID pgl) DNA heteroduplexes were interpreted to indicate that the stable, constitutive reversion had arisen by a deletion of 8/4 of the gal3 insertion sequence.

The main conclusions are: (i) the stable, constitutive reversions of gal3 can arise by partial deletions of the insertion sequence, apparently by elimination of the nucleotide sequence which causes polarity; (ii) the chlD-pgl deletions may exhibit preferential termination at the right extremity of the gal3 insertion in presence of prophage ~; and (iii) the gal3 insertion appears to inhibit the production of ~gal particles by providing a nucleotide sequence which is recognized and degraded by a specific endonucleasc. I t is suggested that inhibition of transducing particle formation by gal3 and the preferred termination of deletions at gal3 might represent related phenomena.

Introduction We have recent ly shown tha t the gal3 m u t a t i o n of Escherichia cell, first

described by Morse, Lederberg, and Lederberg (1956), was caused by the in- sert ion of a DNA sequence, 1,100 nucleot ide pairs in length, in the operator- p romoter (OP) region of the galactose (gal) operon (Ahmed and Scraba, 1975). This f inding provided satisfac¢ory exp lana t ion for the various unusua l features

264 A. Ahmed and E. Johansen

exhibited by the gal3 mutation (Hill and Echols, 1966; Morse, 1967; Adhya and Shapiro, 1969) such as its spontaneous revertibility, failure to respond to chemical mutagens, extreme polarity, and the failure of nonsense suppressors to relieve the polarity.

The most puzzling feature of the gal3 mutation has been the spontaneous production of three different kinds of gal + revertants. (Hill and Echols, 1966; Morse, 1967; Morse and Pollock, 1969). The first class consists of revertants which are stable gal + and synthesize the gal enzymes in an inducible manner. The second class of revertants is unstable, i.e., segregates the original gal3 mutation at a high rate, and synthesizes the enzymes in a constitutive manner. The third class includes revertants which are stable gal + but synthesize the enzymes in a constitutive manner. We found that the stable, inducible revertants arise by accurate excision of the insertion sequence resulting in the restoration of the wild type gal + nucleotide sequence (Ahmed, 1975). The unstable, constitutive revertants appeared to arise by tandem duplications of the gal operon in such a way that the second copy of the gal operon was connected to a different promoter. This new configuration resulted in constitutive expression of the operon, and the genetic instability was ascribed to internal recombination between the two copies of the gal operon.

In the present communication, we have examined the nature of the stable and constitutive class of revertants of the gal3 mutation. We find that one of these reversions arises by deletion of a part of the insertion sequence. We also present evidence showing that the presence of the insertion in the gal operon causes severe inhibition of 2gal particle formation, and that the right terminus of the gal3 insertion seems to constitute a preferred end-point for spontaneous deletions.

Materials and Methods

1. Bacterial Strains All strains used in these studies were derived from E. coli K12 and are listed in Table 1.

2. Media LMT broth contained 12 g Baeto-tryptone, 5 g Bacto-yeast extract, 1 g glucose, 10 g

NaC1, 1.2 g MgSO~, and 50 mg thymine per liter. Tryptone-Mg++ broth contained 10 g Baeto- trypLone, 8 g NaC1, 1.2 g MgS04, and 2 g glucose per liter. Other media have been described earlier (Ahmed, 1975; Ahmed and Scraba, 1975).

3. Preparation o/Lysates and Transduction Procedure Cells were grown to early log phase in tryptone-Mg ++ broth at 30 °. The cultures were

transferred to a 46 ° incubator for 18 rain, and then incubated with vigorous aeration at 37 °. Lysis usually occurred after 2-3 hrs at which time a few drops of chloroform were added.

Transduction was performed by mixing equal volumes of the phage and an overnight culture of the recipient in LMT broth, incubating for 20 min at 30 °, and plating on EMB- galaetose plates. Gal + papillae appeared after 2-3 days at 31 °.

4. Selection o] chlD-pgl Deletions Chlorate-resistant mutants were selected by the procedure of Adhya, Cleary, and Campbell

(1968) using a BBL GasPak 150 anaerobic chamber at 31 °. The chlD-pgl deletions among

Reversion of the gal3 Mutation

Table 1. List of strains

265

Strain Relevant genotype Origin

gal3 gal3(~) gal+ A9 E J100

E J101

E J200

A31

A45

31.1

45.14

46.1

F- thr leu lac gal3 2- thi ga13(),¢I857) gal +

gat3 A9(cMD pgl) ()~cI857)

galClO0 zJ9(chlD pgl) (Zci857)

gal3 (ZgaleJO0 A9(~hlD pgl)) ().ci857)

gale200 ().ci857)

galc200 d31(chtD pgl) (),ci857)

gale200 Ag5(chlD pgl) (~cI857)

gal3 (kgalc200 231(chID PgO) (;~cI857)

gal3 (kgatC200 Ag5(chID pgl)) (;~cI857)

gal3 (kgal+) (kci857)

Obtained from M. L. Morse

gal3 lysogenized with ~cI857

Stable, inducible revertant of gal3 zJ(chlD pgl) derived from gal3 (~) Stable, constitutive revertant derived from zJ9

HFT-produeing transduetant obtained from a lysate of EJ100

Stable, constitutive revertant derived from gal3(Z) (vhlD pgl) deletion derived from E J200

Another deletion from EJ200

HFT-prodneing strains obtained from lysates of z131 and 245. The kgal harbors galc200 reversion and the (chlD pgl) deletion

H~T-producing strain carrying kgal + (46.1) described by Ahmed and Scraba (1975). The gal sequence is derived from a wild-type strain

these were identified by testing for the pgl marker by the method of Kupor and Fraenkel (1969). The deletions were tested for the presence of prophage 2cI857 by failure to grow at 42 °, for uvrB by exposure to a preselected dose of UV, and for nadA, aroG and bio markers by omission of the appropriate supplement from the minimal medium. The gal operons carrying gal3 were distinguished from gal deletions by the ability of the former to produce gal+ revertants.

The procedure of Alper and Ames (1975), modified by the substitution of galactose for glycerol and 2-deoxy-D-galaetose, was used in attempts to select gal+-vhlD deletions which were expected to arise by simultaneous excision of the insertion sequence and the chlD gene from the gal3(~) strain.

5. Preparation o / D N A Heteroduplexes and Electron Microscopy Phage was purified by eentrifugation in block CsC1 gradients by the procedure of Miller

(1972) except that the phage buffer was used at p i t 7.4 (instead of pH 7.9). The DNA heteroduplexes were prepared and mounted for electron microscopy by the formamide technique of Davis, Simon, and Davidson (1971). Length estimates are based upon measurements of 18 heteroduplexes.

Experimental Results I t was fe l t t h a t the most d i rec t app roach to s t u d y the na tu r e of s table , con-

s t i t u t ive revers ions of gal3 would be to isola te 2gal phages bear ing these reversions, and to examine these as ~gale/~gal + he te rodup lexes wi th the DS,TA of a known Zgal+ phage b y e lec t ron microscopy. As shown below, i t was no t poss ible to do so d i r ec t ly and cer ta in special me thods had to be used.


1. Stable, Cowstitutive Reversions Derived/rom gal3(Z)

Forty-three stable, constitutive revertants were identified among a total of 150 gal + revertants obtained by plating gal3(2) on EMB-galactose. Their stability was checked by three successive cycles of re-streaking on tetrazolium- galactose plates and their constitutivity was confirmed by galactokinase assays. Heat-induced lysates from these revertants were used to transduce a gal3 recipient. We found that extremely few gal+ transductants (approximately 2 % of a wild type LFT lysate) appeared. The majority of these gal + colonies were constitutive for the gal enzymes indicating that the reversions had been transduced. However, unlike true transductants, these colonies produced neither gal- seg- regants nor H F T lysates. These transductants were therefore considered to arise from abnormal transducing particles. One possible explanation envisaged for the production of abnormal particles is that extra DNA was brought in by the stable, constitutive reversion so that the amount of DNA necessary to produce normal transducing particles exceeded the packaging capacity of the phage particle. Such would be the case if these reversions were caused by tandem or inverted duplications or inversions of the gal operon. Hence, it was considered desirable to delete the chlD-pgl region located between gal and att2 (Feiss, Adhya, and Court, 1972) to allow packaging by 2gal particles of any extra DNA that might accompany the stable, constitutive reversions.

2. Stable, Constitutive Reversions Derived/rom gal3(2 ) Carrying chlD-pgl Deletions

A total of 2,250 chlorate-resistant mutants derived from gal3(~) were tested for simultaneous loss of the pgl marker. In this manner, 26 extended chlD-pgl deletions were isolated. Of these 26 deletions, 11 failed to produce gal + revertants suggesting that these deletions had extended into the gal operon. The remaining 15 continued to produced gal+ revertants indicating that the gal3 mutation had been retained. These deletions showed the following extents: 9 were gal3 A (chlD pgl) (A) ; 3 were gal3 zJ (chlD pgl A) ; and 3 were gal3 A (chID pgl A bio uvrB). I t was surprising to find that 14 out of these 15 deletions produced constitutive revertants exclusively, and no inducible revertants were ever found. Only one deletion (viz., 312, which was gal3 zJ (chlD pgl A bio uvrB)), behaved like the oiiginal gal3 mutation in the production of predominantly inducible reversions.

As expected, the presence of a chlD-pgl deletion in the gal3(~) strain greatly facilitated the isolation of a ~gal phage bearing a constitutive reversion. A lysate from strain E J100 (a stable, constitutive revertant derived from gal3 zJ9(chlD pgl) (~)) was used to transduce gal3 and a transductant which segregated gal- and produced normal HFT lysate was isolated. The genetic constitution of this strain, designated EJ101, was believed to be gal3 (2galclO0 A9(chlD Tgl)) (~). However, in view of our interpretation concerning the origin of these reversions (see Discussion), we do not consider these as typical representatives of the stable, constitutive class. This strain was, therefore, not utilized for electron microscopy of DIqA heteroduplexes.

Reversion of the gal3 Mutation

Table 2. Influence of the gal3 insertion on the frequency of chlD-pgl deletions

267

Parent strain

Chlorate-resistant % A (chlD pgl) A (chlD pgl) Major class of mutants per among per gal + revertants viable cell chlorate-resistant viable cell produced by

mutants A (chlD pgl)

gal + 10.6 × 10 _6 0.2 2.1 × 10 -s - - gaP(X) 3.0 × 10 -6 0.4 1.2 × 10 -s - - gal3 12.1 × 10 -6 2.6 31.5 × 10 -s inducible gal3(X) 4.3 × 10 -6 3.2 13.8 × 10 -s constitutive gale2OO(X) 3.8 × 10 -6 5.4 20.5 × 10 -6 - -

3. Influence o/the Insertion Sequence and ~ on Deletion Frequency and End-points

We unde r took a sys t ema t i c s t u d y of the influence of the gal3 inser t ion and p rophage ,% on the f requency and end-po in t s of chlD-pgl delet ions. The frequencies of spontaneous ch lora te - res i s tan t mu ta t i ons and ehlD-pgl dele t ions in var ious strains, isogenic except for the inser t ion sequence and ~, are p resen ted in Table 2. I t is clear t h a t the f requency of spontaneous ehlD-pgl delet ions was increased 10- to 15-fold in presence of the gal3 inser t ion. This effect was also de t ec t ab le in a s table , cons t i tu t ive reversion, gale200, der ived f rom gal3(,~). The observed frequencies of de le t ions were somewhat r educed in the presence of ()~).

The ex ten t s of ehID-pgl dele t ions wi th respec t to the ad j acen t loci viz., att~ or (,~), bio, and uvrB on the r igh t side, and gal, aroG, and nadA on the left , were also de t e rmined for the d i f ferent s t ra ins l i s ted in Table 2. These d a t a are no t p resen ted because we fa i led to not ice any s ignif icant p a t t e r n or consis tency in the ex ten t s of var ious dele t ions among these loci. The on ly in te res t ing fea ture observed was t h a t de le t ions arising from the gal +, gal+(~), or gal3 s t ra ins which t e r m i n a t e d be tween gal and chlD d id not, in a n y de tec tab le manner , a l t e r the i n t eg r i t y of the gal operon. Thus, de le t ions of th is class de r ived from gal + and gal+ (A) were st i l l inducib le gal +, and those de r ived f rom gal3 were sti l l gal3 and produced m a i n l y inducib le rever tan t s . I n contras t , the m a j o r i t y of such de le t ions f rom gal3()~) produced on ly cons t i tu t ive r eve r t an t s ind ica t ing t h a t the r egu la to ry region (OP) of the gal operon had been p re fe ren t i a l ly impai red . P r e l i m i n a r y a t t e m p t s to select gal+-chlD (or gale-chlD) dele t ions t h a t would dele te the ent i re inser t ion sequence and chID were wi thou t success.

4. Construction o/ ~gal Phages Carrying Stable, Constitutive Reversions o/ gal3 A stable , cons t i tu t ive r e v e r t a n t des igna ted EJ200 (gale200) was i so la ted

f rom gal3(~). As shown in Table 3, th is was a typ ica l r e v e r t a n t because i t synthesized high levels of k inase in a cons t i tu t ive manner , d id no t produce gM- segregants (gM- segregat ion f requency < 2 . 8 × 10-5), and p roduced poor L F T lysa tes con- t a in ing abnorma l t r ansduc ing par t ic les . Two chlD-pgl dele t ions of EJ200, viz., d31 and A45 which had r e t a i n e d gale200 and (~) were found to produce norma l L F T lysa tes capable of t r ansduc ing a gal3 recipient . Two gal + t r ansduc tan t s , des igna ted 31.1 and 45.14 (which had or ig ina ted from LJ31 and A45, respect ive ly)

2 Molec. gen. Genet.


Table 3. Properties of HFT-produeing strains derived from the stable, constitutive revertant E J200

Strain Genotype Galactokinase Kind Ratio specific of kgal/k activity a lysate (uninduced)

gal + ()~) gal+(k) <0.001 LFT 4.8 × 10 -7 gal3 gal3 0.002 - - - - E J200 galc200()~) 0.101 LFT b 9.7 × 10 -9 331 gale200 A31(chID pgl)(k) 0.054 LFT 4.7 x 10 -7 A45 gale200 A45(chID pgl)(k) 0.095 LFT 1.3 X 10 -7 31.1 gal3 (),galc200 A31(chlD pgl))()~) 0.160 HFT 1.9 X 10 -a 45.14 gal3 ()~galC200 A45(chlD pgl))(~) 0.061 HFT 2.5 x 10 -S

46.1 gal3 (~,gal +) (k) 0.001 lIFT 2.5 X 10 o

a Expressed as ~zmoles galactose phosphorylated per mg protein per min. b Lysates contained abnormal transducing particles. Transductants were constitutive but failed to produce gal- segregants and HFT lysates.

behaved as normal heterogenotes. These strains produced t t F T lysates, segregated gal- colonies, and synthesized galactokinase in a consti tutive manner. The strain 31.1 was selected for DNA heteroduplex analysis.

5. Electron Microscopy o/~gal+/~gal c DNA Heteroduplexes I-Ieteroduplexes prepared by annealing the DNA of ~gal + and ~galc200 zJ (ehlD

pgl) phages are expected to display two essential features. These are (i) a subst i tution loop located near the left end (caused by different points of subst i tution of bacterial DS~A in the two independent ly derived ~gal phages), and (ii) a loop corresponding to the chID-pgl deletion expected to he located to the r ight of the first loop. This basic pa t te rn is subject to fur ther variat ion depending upon the alteration tha t m a y accompany the gale200 reversion. This provides a convenient means of identification of the nature of the stable, consti tutive reversion.

Electron micrographs of ~gal + (46.1)/~gale200 ~31(chID pgl) (31.1) heteroduplexes are presented in Fig. 1. The DNA molecules were mainly linear and contained two substi tut ion loops at fixed positions. A few molecules had circu- larized, but retained the two loops. The left end carried a short homologous region of mean length 0.60-4-0.01 ~m followed by a small substi tut ion loop (designated "L-loop") whose non-homologous (single-stranded) arms were 0.09:~ 0.01 fxm and 0.83-4-0.03 fxm long. The L-loop appears to have been formed by the different end-points oi bacterial substitutions in the two ~gal genomes. Between this small L-loop and the large subst i tut ion loop (designated "R-loop") was a region of DNA homology which was 1.09-4-0.01 ~m long. The two single-stranded arms of the R-loop, which seems to be caused by the ehlD-pgl deletion, were 2.11:[:0.06 and 0.084-4-0.004 ~m in length. The remaining par t of the duplex, which extended from the R-loop to the r ight end, was completely homologous

Reversion of the gal3 Mutation 269

Fig. l a--e . Electron mierographs o2 ~tgal+ (46.1)/~galc200 J31(chlD pgl) (31.1) DNA heteroduplexes. A complete heteroduplex is shown in (a) and portions of heteroduplexes are shown in (b-e). Arrows indicate the 270 base pair fragment of the gal3 insertion located adjacent to the chlD-pgl deletion. Horizontal bars represent 0.5 ~m. Photographs were taken by Dr.

D. Scraba


(a) 0 4.9 13.4

A K T £ OP chlD pgl go/ f

go~3

46.5

......... I '

art B.P'

53.5 Agal ÷(46.1 )

100

R

(b) -~ IZ2 ~ -

04.9 5.7 14.6,,31.7 46.5

'o,~.-! [-~'~/FTN pg/ I

• V"I Aga/3 A31(chlD pgI) L- loop R- loop

100 Agal+(46.1)

AgalC(31.1) R

Fig. 2. (a) Revised physical map of 2gal + (46.1) chromosome (Ahmed and Scraba, 1975) showing new co-ordinate assignments in Agal units (1 ~gal unit:400 base pairs.) Straight and wavy lines denote phage and bacterial sequences, respectively. The location of various phage and bacterial genes is indicated below the lines. Upright numerals refer to map co- ordinates and slanted numerals indicate map distances. (b) Schematic representation of 2gal+(46.1)/Agale200A31(chlDpgl) (31.1) heteroduplex. Arrow indicates the location of chlD-pgl deletion (co.ordinates 14.6-31.7) adjacent to the partially deleted gal3 insertion sequence on the Agale strand. The lengths of Agal + and 2gal c strands are estimated to be

approximately 40,000 and 35,800 base pairs, respectively

between the two genomes and measured 8.39:~0.06 [zm in length. Knowing that the long arm of the R-loop (i.e., the chlD + pgl + segment, 2.11 b~m) was contributed by the ~gal + genome, and assuming that the short arm of the L-loop (0.09 ~m) was derived from ~gal +, the overall lengths of ,~gal + and ~galc DNA were calculated to be 12.30~=0.08 ~m and 11.00±0.07 [zm, respectively. These calculations are in close agreement with independent length measurements of ~ (14.3 ~m), ~gal + (12.5 ~m), and ~gal c (11.3 ~m) homoduplexes indicating that the assumption concerning the assignment of the two arms of the L-loop was correct.

The different regions of homology and non-homology in the heteroduplex were identified by the use of the physical map of ~gal+(46.1) (Ahmed and Scraba, 1975). In order to do so, it was considered necessary to convert the co-ordinates (presented in ~ units) into new 2gal units. The physical map of 2gal + (46.1) with the new co-ordinate assignments is presented in Fig. 2(a). According to om current estimates, the length of this chromosome is 86.1 A units or 40,050 base pairs which is taken as 100 ,~gal units (1 2gal uni t=400 base pairs). According to this map, the region between co-ordinates 0 and 4.9 contains the A genes A and W, the region between co-ordinates 4.9 and 13.4 contains the gal operon, co-ordinate 13.4 marks the site of the gal3 insertion, the segment between co-ordinates 13.4 and 46.5 contains chID, pgl, and phr genes from E. coli, the attZ (B.P') site maps at co-ordinate 46.5, and the remaining segment to the right contains A nucleotide sequences from art to gene R. These measurements also permit more accurate estimates of the lengths of the gal operon and the gal3 insertion to


3,410 and 1,110 base pairs, respectively. Previous estimates were reported by Ahmed and Scraba (1975).

Knowing the overall length (12.3 ~m--~100 ~gal units) and the arrangement of genes along the ~gal + strand, the interpretation of the physical structure of ~gal c becomes a relatively simple matter. A schematic drawing of the 2gal+/ ),galc A(chlD pgl) heteroduplex is presented in Fig. 2 (b). Several interesting features which become obvious are described below:

(i) within the L-loop, the short arm (0.78~0.05 2gal units) consists of bacterial DNA from ~gal + whereas the long arm (6.72=[=0.24 ~gal units) contains phage DNA from ~gal°;

(if) the region of homology from co-ordinate 5.7 to 14.6 (i.e., between the two loops) which is 3,560 base pairs in length contains the gal operon (indicating that the galo200 reversion was not caused by inversion or inverted duplication of the gal operon) ; and

(iii) the most intriguing observation is that the large R-loop, apparently caused by the chlD-pgl deletion, is not a true deletion loop. Several micrographs presented in Fig. 1 (b-e) clearly show that it is a substitution loop in which the long arm (17.2~=0.49 2gal units=6,870 base pairs) containing the chlD+ and pgl + genes is derived from 2gal +, and the very short arm (0.68-4-0.03 ~gal uni t s= 270 base pairs) is derived from )~gal e. This raises the problem as to how a single deletion event could give rise to a substitution loop, and what the possible origin of the short non-homologous segment in 2gal c could be. This problem is resolved if one recalls that 2gal c was derived from a bacterial strain which harbored the gal3 insertion in addition to all sequences homologous to ~gal +. This 270 base pair long segment must, therefore, represent a remaining portion of the original gal3 insertion (1,110 base pairs) with the chlD-pgl deletion located immediately adjacent to it.

Discussion

The implications of these findings on our current understanding of the nature of the gal3 insertion and its revertants are discussed below.

1. Pre/erential Termination o/Deletions at the Right Terminus o] the gal3 Insertion Considerable difficulty was encountered in the isolation of ~gal phages bearing

the stable, constitutive reversions of gal3 for electron microscopy. In order to circumvent this problem, we at tempted to delete the chID-pgl region located between att2 and gal and found that the presence of the gal3 insertion exercised profound influence on the frequencies and extents of deletions to the right. The data presented in Table 2 show that there was a 10- to 15-fold increase in the frequency of chlD-pgl deletions in strains carrying the gal3 insertion. This effect, however, appears insignificant when compared to the 30- to 2,000-fold increase in the frequency of deletions reported for the insertion sequence IS1 by Reif and Saedler (1975). In this respect, the gal3 insertion seems to resemble IS2 more closely than IS1.

The simultaneous presence of the gal3 insertion and prophage ~ seemed to have an interesting effect on the extents of the ehlD-pgl deletions produced. Deletions arising from strains carrying the gal3 insertion only produced revertants which were mainly stable and inducible, of the kind known to arise by excision of


OP

K I" E gal3 chlD pgl I I .... I

L~ch/D -pg/ I ...................... 1

Deletion l K F E gal3 A

i i -:= ;-- II II

Excision 1 g T E A

I i ,~ li II

A II II

Constitutive Reversion

Fig. 3. A scheme to illustrate the exclusive origin of stable, constitutive revertants from chID-pgl deletions in a gal3(,~) strain. The deletions preferentially end at the right terminus of the gal3 insertion, so that simple excisions of the insertion yield constitutive revertants by

connecting the gal structural genes to a different promoter

the insertion (Ahmed, 1975). Such deletions seemed to have random end-points at both the left and the right. In contrast, the majority of chlD-pgl deletions arising from gal3(k) produced stable and constitutive revcrtants only. These deletions had variable end-points at the right because some terminated in pgl .while others ended in prophage 2 or uvrB, but their left end-point was located in a shSrt region between gal and chlD. I t is unlikely that the constitutivity of these r6versions was due to slight readthrough transcription initiating from the P~ promoter of prophage 2cI857 (Adhya, Gottesman, and de Crombrugghe, 1974) at 31 °, because constitutive revertants were also produced by deletions which had removed (4) entirely. We interpret the appearance of constitutive revertants exclusively to mean that the presence of (2) confers a unique specificity on chlD- pgl deletions to terminate at or near the right terminus of the gal3 insertion. A simple scheme by which such deletions would produce constitutive revertants is presented in Fig. 3. The chlD-pgl deletion extended up to the right end of the gal3 insertion, so that ordinary excisions of the insertion would yield stable, constitutive revertants only. Direct support for the termination of deletions at the right end of gal3 was provided by electron micrographs (Fig. 1). I t is possible that a similar specificity resides at the left terminus of gal3 for deletions extending leftwards (i.e., into gal, aroG, and nadA). However, we found no evidence for deletions extending from the left terminus of gal3 rightwards to chlD and pgl. A similar specificity of deletion end-points has been observed for the int-promoted deletions of 4 which extend on either side from the cross-over point XOP within the att P.P' site (Davis and Parkinson, 1971). Reif and Saedler (1975) have also shown that one end-point of deletions is frequently at the site of IS1 insertion and the second end-point is variable, located either to the left or right of the insertion. I t may be that the ends of the insertion sequence constitute preferred sites for endonucleolytic attack which is followed by unidirectional degradation outwards 1.

1 It must be borne in mind that extended deletions with one end-point fixed at gal3 would also result from recombination of the gal3 insertion with similar sequences located elsewhere on the genome, with concomitant release of circles.


2. Nature o/the Stable, Constitutive Reversions o/gal3 The nature of a stable, constitutive reversion, galc200, was studied by electron

microscopy of ~gal+/~gal e heterodup]exes. In addition to the galo200 reversion, the ~gal e strand also contained an extended chlD-pgl deletion (A31). The electron micrographs showed a substitution loop (R-loop) located near the OP end of the gal operon. This loop consisted of a long arm (6,870 base pairs) corresponding to the chlD-pgl deletion and a very short arm (270 base pairs) representing the remaining 1/4 portion of the gal3 insertion.

Our interpretation of this configuration is that it represents an aggregate of two independent genetic events, namely the gale200 reversion and the chlD-pgl deletion. First, a stable, constitutive reversion (gale200) occurred by deletion of 3/4 of the gal3 insertion. This deletion may have included the nucleotide sequence(s) responsible for polarity (such as transcription termination sites). Elimination of polarity signals together with the failure of the repressor to bind the disrupted operator is probably sufficient to explain constitutive expression of the gal operon in this class of revertants. In the second step, a spontaneous deletion (LJ31) removed the region extending from the right end of the gal3 to the pgl gene. The segment bearing the gale200 reversion and LJ31(chlDpgl) was subsequently incorporated into the genome of a transducing phage by defective excision leading to the establishment of an H F T line (31.1). DNA heteroduplexes between ,~gal + with ),gale of the aforementioned origin are expected to yield the observed configuration (Fig. 2b).

I t must be remembered, however, that these electron micrographs do not prove that the constitutive reversion occurred by partial deletion of the insertion sequence. In particular, an alternate explanation tha t the constitutive reversion arose by inversion of the insertion sequence (Saedler, Reif, Hu, and Davidson, 1974) followed by a chlD-pgl deletion which extended into, and removed, 3/4 of the insertion sequence cannot be ruled out. However, it does not appear to be com- patible with the genetic evidence presented earlier. The electron micrographs do eliminate other explanations such as inversions or inverted duplications of the gal operon as the basis for stable, constitutive reversions (Ahmcd, 1975).

3. Inhibition o/ ~gal Particle Production by gal3 We have repeatedly observed tha t the presence of the gal3 insertion in the gal

operon interferes, in some manner, with the production of ~gal particles. This drastic inhibition is noticeable both in LFT as well as I / F T lysates. For example, the data in Table 3 show that, in LFT lysates, the ratio of ~gal particles to ~ was 4.8 × 10 -7 for gal + (~) strain, and only 9.7 × 10 -9 for galc200(~). Thus a 50-fold reduction in the proportion of ~gal particles was observed in the gal e strain which retained a portion of the gal3 insertion. Similar poor LFT lysates were invariably obtained with 42 other gal e revertants derived from gal3(~). In each case, the proportion of ~gal particles was drastically reduced, and even these particles were not normal since true transductants were never produced.

Initially, we at tr ibuted this inhibition of transducing particle formation to packaging difficulties caused by extra DNA which was thought to accompany the gal e reversion. This assumption appeared to be reinforced by the finding tha t


chlD-pgl deletions (such as A31 and A45), which would allow for packaging of extra gal DNA, caused almost complete restoration of ~gal c formation in LFT lysates (Table 3). However, the electron microscopic evidence presented earlier suggests that galo200 was not caused by the addition of DNA but rather by deletion of a part of the gal3 insertion. I t appears very unlikely tha t the presence of only 270 extra nucleotides of gal3 DNA in gale200 could pose such a serious problem unless the basis of inhibition is other than simply the packaging capacity of the phage particle.

This inhibition of ~gal production does not seem to be exercised at the level of defective excision but at a later step such as packaging because similar inhibition was expressed in H F T lysates. During our previous studies (Ahmed, 1975) involving CsC1 density-gradient eentrifugation of radioactive H F T lysates, it was noticed that ~gal particles harboring the gal3 insertion (such as ,~gal3 (46.1.8 A/2) and particular]y Revl02) were produced in extremely small quantities compared to ~t. On the other hand, yield of ~gal particles was normal in ,~gal + (46.1) in and the inducible revertants (Revl04 and 109) from which gal3 had been excised. This inhibition of ~gal production could not have been due to limitations on packaging capacity because ~gal + (46.1) contains 86.1% of ~ DNA, and addition of 2.4% extra gal3 DNA to this genome should have been easily accommodated. Finally, we note that this inhibition was partially or totally relieved in both LFT and l I F T lysates by the introduction of chlD-pgl deletions (Table 3).

In order to explain these observations, we propose tha t gal3 and its constitutive reversions inhibit the production of ~gal particles by providing a specific nucleotide sequence which is highly susceptible to cndonucleolytic degradation, rather than by exhausting the packaging capacity of the phage particle. This nueleotide sequence could be recogzlized by a restriction kind of endonuclease leading to selective, pre-mature degradation of the ~gal DNA pool. Alternatively, the insertion could interfere with the packaging process in an unspecified manner by acting as a hot spot for recombination. Whatever be the true mechanism, it is conceivable tha t inhibition of transducing particle formation, its relief by chlD-pgl deletions, and the preferred termination of deletions at gal3 are different aspects of one and the same phenomenon.

Acknowledgements. We are grateful to Dr. D. Scraba for electron microscopy and Bey Clark for loveable technical assistance. E. Johansen was supported by the Government of Alberta STEP Program for undergraduates. This research work was supported by NRC Grant A-4412 to A. Ahmed.

References Adhya, S., Cleary, P., Campbell, A. : A deletion analysis of prophage lambda and adjacent

genetic regions. Proc. nat. Acad. Sci. (Wash.) 61, 956-962 (1968) Adhya, S., Gottesman, M., de Crombrugghe, B. : Release of polarity in E. coli by gene N of

phage )t: Termination and antitermination of transcription. Proc. nat. Acad. Sci. (Wash.) 71, 2534-2538 (1974)

Adhya, S., Shapiro, J. A. : The galactose operon of E. coli K-12. I. Structural and pleiotropic mutations of the operon. Genetics 62, 231-247 (1969)

Ahmed, A. : Mechanism of reversion of the gal3 mutation of Escherichia coll. Molec. gen. Genet. 136, 243-253 (1975)


Ahmed, A., Seraba, D. : The nature of the gal3 mutation of Escherichia coll. Molec. gen. Genet. 136, 233-242 (1975)

Alper, M., Ames, B. N. : Positive selection of mutants with deletions of the gal-chl region of the Salmonella chromosome as a screening procedure for mutagens that cause deletions. J. Baet. 121, 259-266 (1975)

Davis, R .W. , Parkinson, J . : Deletion mutants of bacteriophage lambda. III . Physical structure of att~. J. molec. Biol. ~6, 403423 (1971)

Davis, R .W. , Simon, M., Davidson, N.: Electron microscope heteroduplex methods for mapping regions of base sequence homology in nucleic acids. In: Methods in enzymology (L. Grossman and K. Moldave, ed.), vol. 21(D), p. 413-428. New York: Academic Press 1971

Feiss, M., Adhya, S., Court, D. L. : Isolation of plaque forming, galactose-transducing strains of phage lambda. Genetics 71, 189-206 (1972)

Hill, C. W., Echols, H. : Properties of a mutant blocked in inducibility of messenger RNA for the galactose operon of Escherichia coll. J. molec. Biol. 19, 38-51 (1966)

Kupor, S, R., Fraenkel, D. : 6-Phosphogluconolactonase mutants of Escherichia coli and a maltose blue gene. J. Bact. 100, 1296 1301 (1969)

lVIiller, J. H. : Experiments in molecular genetics. Cold Spring Harbor, New York: Cold Spring Harbor Laboratory, 1972

Morse, M. L. : Reversion instability of an extreme polar mutant of the galactose operon. Genetics ~6, 331-340 (1967)

Morse, M. L., Lederberg, E. M., Lederberg, J. : Transductional heterogenotes in Escherichia coll. Genetics 41, 758-779 (1956)

Morse, M. L., Pollock, B. F. : Reversion instability in the galactose operon of Escherichia coll. J. Bact. 99, 567-569 (1969)

Reif, H. J. , Saedler, H. : IS1 is involved in deletion formation in the gal region of E. coli K12. Molec. gen. Genet. 187, 17-28 (1975)

Saedler, H., Reif, H. J. , Hu, S., Davidson, N. : IS2, a genetic element for turn-off and turn-on of gene activity in E. coll. Molec. gen. Genet. 132, 265-289 (1974)

C o m m u n i c a t e d by G. B e r t a n i

Asad Ahmed and Eric Johansen Department of Genetics University of Alberta Edmonton, Alberta T6G 2E9 Canada

reversion of the gal3 mutation of escherichia coli: partial deletion of the insertion sequence

Documents