Biochimie ( !991 ) 73, 329-334 © Socirt6 franqaise de biochimie et biologie moirculaire / Elsevier. Paris

32'-)

Positive and negative regulatory elements in the dnaA-dnaN-recF operon of Escherichia coil I Prrez-Roger, M Garcfa-Sogo*, JP Navarro-Avifi6, C L6pez-Acedo, F Macifin. ME Armengod** ]nstituto de I,westigaciones Citol6gicas, Amadeo de Saboya, 4, 46010-Valencia, Spain

(Received 19 November ! 990; accepted 31 January 1991 )

Summary - - The reeF gene of E coli lies within a cluster of genes which play essential roles in DNA replication; the gene order is dnaA dnaN reeF gyrB. Each of these genes has its own promoters which, with the exception of dnaA promoters, reside entirely within the translated region of the respective preceding gene. In this report, we analyze the effect of the dnaA and dnaN promoters on reeF expression by translational fusions between recF and the iacZ reporter gene. Our results indicate that recF is a distal gene of the dnaA operon, and support the previous proposal that dnaN and recF constitute a transcriptional unit under control of the dnaN promoters. They also suggest that dnaA, dnaN and recF are predominantly expressed from the same mRNA although transcriptional and/or posttranscriptional mechanisms should be specifically involved in lowering expression of the recF gene. Recently, we have localized 3 tandem transcription termination sites in the second half of the dnaN gene. downstream from the recF promoters. Neither of them shows the typical features of simple terminators and apparently they do not work in a minimal system of in vitro transcription. In this report, we present evidence that only one of them is dependent on the Rho protein. Although the operon structure allows coordinate expression of dnaA, dnaN and recF, the presence of internal promoters (the dnaN and recF promoters), which appear to be inducible by DNA damage, and intracistronic terminators, whose activity is inversely proportional to the efficiency of translation, permits expression of individual genes to be independently regulated in response to altered growth conditions. recF / dnaA operon / promoters / terminators / Rho factor

Introduction T he recF gene of E coli controls one of the recombination pathways as well as UV sensitivity and has recently been involved in stable replication [1]; however, its precise function and expression pattern are still largely unknown, recF lies within a cluster of genes which participate in DNA metabolism (see fig 1). These genes are dnaA, which encodes a D N A binding protein essential for initiation of DNA replication at the chromosomal origin, dnaN, which determines the 13 subunit and processivity factor of D N A polymerase III, recF and gyrB, which encodes the 13 subunit of DNA gyrase, an enzyme required for maintenance of chromosomal superhelical density. The 4 adjacent genes are all transcribed in the same direction (from dnaA toward gyrB) and are organized

*Present address: Department of Medicine, University of Medicine and Dentistry of New Jersey, Robert Wood Johnson Medical School, 675 Hoes Lane, Piscataway, NJ 08854-5635, USA **Correspondence and reprints Abbreviations: bp, base pair; UV, ultraviolet light

in a very compact fashion. Thus, the dnaA-dnaN and recF-gyrB intercistronic regions are 4 and 30 bases respectively, and d n a N and recF overlap by 1 base [2, 3]. This organization suggests that the expression of these genes is interrelated. In fact, the dnaN gene seems to be cotranscribed with dnaA as an operon [4, 5] and we have demonstrated that the dnaN and recF genes constitute a transcriptional unit under control of the dnaN promoters [6] (see also fig 1). Curiously. recF has its own promoter region located in the middle of the d n a N structural gene [7]. Therefore, transcripts initiated at the recF promoters have an unusually long, non-translated leader region. We have recently identified and characterized 3 tandem transcription termination sites (TI, T2 ,~.,,""a T3) ,--~,--~-...~.vv.. sible for transcriptional polarity in the dnaN-recF operon (Armengod et al, submitted). These sites are located in the dnaN gene, downstream from the recF promoter region (fig 1). To define the ability of TI, T2 and T3 to terminate transcription started at the recF promoters, D N A fragments containing these promoters together with T-flanking sequences of variable length, including or not the terminators, were inserted

.~.~0

I Pdrez-Roger et al

Plasmid pMLB1034 191 contains the lacZ gene lacking a romoter, ribosome binding site and the first 8 codons for -galactosidase. It also contains 3 unique restriction sites into which DNA fragments harboring promoter and translation initiation sites can be inserted to form gene fusions. The order of the 3 restriction sites before the lacZ gene is EcoRl, Smal and BamHl. Derivatives of plasmid pMLBI034 encoding ~galactosidase fusion proteins were usually isolated by growing transformants of E coil K-12 strain MCI000 on LATamp plates 161 containing 40 pg/ml of X-Gal (5-bromo-4chloro-3-indolyl-~-D-galactopyranoside) and screening of blue colonies. In all constructions the orientation and size of inserts were checked by restriction mapping. Plasmid plCI03 was constructed by cloning a 3555-bp chromosomal EcoRl fragment obtained from pJC605 (see 171; and fig !) into the EcoRl site of pMLBI034. The fragment contains a long stretch of the dnaA gene without its promoter region, the dnaN and reef genes and the NH2-terminal coding sequence of the gyrB gene 12, 31. These 4 genes are transcribed in the same direction and. in plC 103, they are correctly oriented with respect to the lacZ gene but there is no translational fusion between gyrB, the last gene on the chromosomal fragment, and lacZ. plC356, a recF-iac'Z translational fusion, was formed from plCl03 by deletion of the DNA fragment between the Ncol site, on chromosomal DNA (bp 2586 in fig 1), and the SmaI site on pMLB 1034 DNA, followed by filling in of the NcoI protruding ends mediated by DNA polymerase I (Klenow fragment) and recircularizing in the presence of HindllI linkers (sequence GAAGC'IWC). It should be noted that the NcoI site is finally regenerated, plC361 was constructed from plC356 by deletion of the DNA fragment between the EcoRI and XhoI sites located at bp -3 and 1490, respectively (fig 1), followed by filling in of the protruding ends mediated by DNA polymerase I (Klenow fragment) and recircularizing in the presence of Sail linkers (sequence CGTCGACG). plC176 [6] includes the EcoRl chromosomal fragment extending from -3 to 3553 (fig 1). This plasmid contains a small deletion in the dnaN sequence, from bp ! 474 to ! 506, hereafter designated zldnaN176. Both ends of the deletion are joined throu~zh a Hindlll linker (seauence (GAAGCTI'C). It-should be pointed out that the deletion of a few nucleotides and the insertion of the linker in plC176 neither change the dnaN reading frame nor produce nonsense

imo the ga/K transcription fusion ~ector p K O - ! [8]. GatK ac~i,,'ities for the resulting p l a s m i d s were a.~sa)ed and nonrralized by using the ~ - l a c t a m a s e gene to e x c l u d e .any effect o f p l a s m i d copy n u m b e r on the results. In N I00 host !61, the termination was 4 4 % effective tor T I, 62% for T ! + T2 and 80% for T ! + T2 + T3. N o termination activity was l b u n d associated with d n a N -sequences located d o w n s t r e a m from T3. Apparently, the terminators do not affect d n a N e x p r e s s i o n under normal p h y s i o l o g i c a l conditions. H o w e v e r . they reduce e x p r e s s i o n o f this gene in the p r e ~ n c e o f the translation i n h i b i t o r fusidic acid. W e c o n c l u d e d that the n o r m a l c o u p l i n g o f transcription and translation o f the d n a N gene a l l o w s R N A polym e r a s e c o m i n g from the d n a N promoters to c o n t i n u e through the terminators and contribute to the expression o f the r e c F gene. T h e situation is very different for t r a n ~ r i p t s initiated at the r e c F promoters b e c a u s e they are not being translated w h e n reaching the terminators and are m o s t l y stopped (see fig 1). Since d n a A and d n a N are o ~ a n i z e d in an operon under control o f the dnaA promoters, it s e e m s r e a s o n a b l e to propose that r e e f is a distal gene o f this operon b e c a u s e all transcription o f d n a N o v e r c o m i n g terminators T 1, T2 and T3 s h o u l d be able to reach the r e e f gene i n d e p e n d e n t l y o f the promoters ( d n a A or d n a N promoters) at w h i c h transcription had started. T h e a i m o f this work was to further characterize the T I , T2 and T3 terminators and to test the h y p o t h e s i s that r e e f is a distal gene o f the d n a A operon. Materials

~

and Methods

Bacterial xtrains and plasmids The E coil K-12 strains used were MCI000, CSH42 and JBRR3. which have previously been described 19-111.

m :b

tU

u

o to

Q.

t',l

i

!

.,....

i,,.L,

1,.I I I

I] ; "

I

I

I

PIP2

)k d.,,. O~

,-

:

}

L

I

II

.,jl,.

e)

P1 P 2 P 3 P 4

r,,, P, t T1

t t T2 "1"3

Fig 1. Structure of the dnaA-dnaN-recF-gyrB chromosomal region of E coll. Relevant restriction sites are indicated. Coordinate 0 is the middle of the EcoRl site located at the beginning of the dnaA structural gene [2]. The A's of the ATG starts o[ the dnaN and r e e f structural genes, according to Ohmori et ai [2] and Blanar et al [3], are at bp 1346 and 2446, respectively. The main transcription initiation points are at b p - 2 9 7 a n d - 2 1 4 (for dnaA promoters PI and P2), 934, 1077, 1105 and 1319 (for dnaN promoters PI, P2, P3 and P4) and 1806 and 1826 (for reeF promoters P2 and PI ) [2, 6, 7]. Terminators in the second half of dnaN are designated TI, T2 and T3 (see text).

Regulatory elements of the dnaA-dnaN-recF mutations in the gene, although obviously the dnaN gene product becomes inactive [6]. plC409, a recF-lacZ fusion, was constructed by replacing the wild-type BstEII-Ncol fragment (bp 1056-2586, in fig I) of plC356 with the equivalent fragment of plC!76 carrying the dnaN176 mutation, plC435 was formed by inserting the 945-bp EcoRl fragment (bp -948 to -3, in fig 1) from pBF110 [12] into the unique EcoRl site of p1C409 in such a way that an entire dnaA gene was generated. plC340, a dnaN-lacZ translational fusion, was formed from plCl03 by deletion of the DNA fragment between the Xhol site, on chromosomal DNA (bp 1490 in fig !), and the Smal site, on pMLBI034 DNA, followed by filling in of the Xhol protruding ends mediated by DNA polymerase I (Klenow fragment) and recircularizing in the presence of HindIll linkers (sequence CCAAGCTTGG). pIC371 was obtained by digestion of pIC103 with Bai31 from the unique Ncol site (at bp 2586 in fig 1). Bai31-cut DNA was treated with DNA polymerase I (Klenow fragment), digested with Smal (to eliminate chromosomal DNA sequences between the dnaN and lacZ genes) and recircularized. Deletion end-point in plC371 was located by DNA sequencing, plC420, a dnaA-lacZ fusion, was constructed by inserting the 1012-bp HpaI-Pvull fragment (bp -856 to 156, in fig 1) of pBFll0 into the Smal site of pMLBI034, plC428 was formed by digesting pIC420 with Clal, treating the digested DNA with DNA polymerase I (Klenow fragment) to fill in the ends and then redigesting that DNA with BamHI; the resultant 548-bp ClaI-BamHI fragment carrying the fusion region of pIC420 was inserted between the Sinai and BamHI sites of pMLB1034 to produce piC428. Plasmids piC437, piC438, piC439 and plC440 were construtted by inserting the 393-bp EcoRI fragment of p1C428, containing the dnaA promoters, into the EcoRI site of pIC340, piC356, pIC371 and pIC409, respectively, in such a way that an entire dnaA gene was generated. In these constructions, ligation mixtures were transformed into CSH42 (dnaA46) and ,i o selected on LATamp at ,+2 C. DNA manipulations

Standard methods were used for construction and cloning of plasmids, bacterial transformation, purification of plasmid DNA, and restriction enzyme analysis [13]. Enzyme assays

For determination of I]-galactosidase activity, cells harboring lacZ plasmids were grown in LBT [61 with 40 l.t~ml of ampicillin. Assays were performed, unless otherwise noted, on sodium dodecyl suifate--chloroform-permeabilized cells as described by Miller [ 10]. ~-Lactamase activity was measured as specified by Andrup et al[ 141.

Results and Discussion Effect o f Rho f a c t o r on terminators TI, T2 and T3

Prokaryotic transcription terminators may be operationally classified into simple or complex depending on whether or not they can work in a purified transcription system in the absence of any added termination proteins [15]. While features of complex terminators are poorly understood, simple terminators consist of 2 parts: a), a stable, GC-rich dyad s y m m e t r y

331

(allowing an RNA stem-loop structure with AG typically - 2 0 to - 3 0 kcal/moil: and bl. a stretch of several consecutive uridine residue~ in the transcript sequence. The terminated transcript ends within, or just distal to, the oligo (rU) sequence [15]. DNA sequence analysis performed by using the Microgenie program (Beckman Instruments. Inc) indicates that termination points T I , T2 and T3 are located in DNA regions which do not show the typical structure of simple terminators (data not shown). Moreover. previous results suggest that these signals do not work m a minimal system of in vitro transcription ([7]: and unpublished results). Therefore, it can be postulated that they fall into the category of complex terminators. Most, if not all, of the complex terminators identified in E coli up to now depend on the Rho protein [15]. Therefore, we decided to analyze the in vivo effect of this protein on T l, T2 and T3. To this end we used the host strain JBRR3, which has a temperature-dependent Rho activity [ 11 ]. DNA fragments carrying the reeF promoters together with 3'-flanking sequences of variable length, including or not some or all the terminators, were inserted into the lacZ transcription fusion vector pCB267 116]. The resulting plasmids were transformed into JBRR3 and the [3-galactosidase activity was measured at 30 and 42°C (table I). These plasmids were also transformed into M C I 0 0 0 . We used this strain as a control, since the effect of temperature is a factor which m a y vary the activity of some terminators [17, 18]. In these experiments we made use of the pCB267 vector because the 13-galactosidase assay is cheaper and easier to perform than the galactoi.: . . . . . . . . . . lkllli:lbC d33aff,

I..I . . . . . . . . . llUW'k.v,,..i.,

" It

~h~HId ,:,HvuJtu

~

nt~int~d !

nHl

thai

utilization of lacZ in place of galK transcriptional fusions to quantify the s t r e n ~ h of regulatory signals can lead to an unrealistic picture of the transcriptional acttvity reaching the indicator gene. This is due to the fact that the translation efficiency of lacZ is highly sensitive to changes in upstream RNA structure, ie DNA sequences in the 5' untranslated regions between the promoter and lacZ may strongly influence the expression of this gene by altering the accessibility of the lacZ ribosome binding site [8, 16, 19]. This can explain w h y in the Rho+ control hosts (see M C i 0 0 0 at any temperature or JBRR3 at .300 C in table I) the percentage of termination for plC406, which carries T I and T2, appears greater than that obtained for piC394, which contains T I. T2 and T3. However. since the aim of these experiments was not to quantify the relative strength of the termination signals (already carried out with the pKO-I system) but to analyze whether the percentage of termination related to a specific D N A region varied depending on the Rho activity, the lacZ fusion plasmids could be used. As shown in table I, the in vivo termination activity for pIC404 and pIC406 in MC 1000 is maintained in

•~;,

i Pcrez-Rogcr ct a/

l'abte I. Eft~2cl o! lhc Rho factor on terminalo, s TI. T2 and T3. JBRR3

MC 1000 Ph+.smid~ Fr~lgme~lt h

Terminator fl-Gal ttnitx c .

~0 ~(

pCB267 plC308 piC4(~4 plC4()6 piC394

i 486-- i 923 1647-2060 1429.-2158 ! 6 i 8--2212

TI T I. T2 T I. T2. T3

4YC

pCB267 plC ~98 plC404 piC4(~ piC394

l--h~,6-! 923 i 647-2069 !429--2 ! 58 1618-2212

TI T I. T2 TI.T2. T3

.

.

Bla ttnit.~~ .

0.15 9.25 5.70 1.71) 4.43 0.40 20.71 10. ! I 3.2 ! 7.94

.

.

[J-Gal/Bla d

% Tc

fl-Gal

Bla

ttnit,~,

ltttits c

fl-Gal/Bla d

% Te

.

9.00 9.(~) 9.05 7.90 9. I 0

0.02 ! .03 0.63 0.__ 0.49

40 80 53

0.13 8.90 5.20 i .8.', 4.70

6.60 7.00 7.30 7.20 7.50

0.02 1.27 0.7 ! 0.25 0.63

45 82 5!

9.81 9.48 8.05 8.72 9.02

0.04 2. i 8 1.26 0.37 0.88

43 85 61

! 1.42 54.43 32.83 ! 4.07 38.84

21.99 21.90 20.34 19.03 20.85

0.52 2.49 ! .61 0.74 1.86

45 89 32

+,DNA fragments were inserted into the Sinai site of pCB267. Only recombinant plasmids carrying the thtaN sequences correctl3 oriented with respect to the h+cZ gene are shown. ~'The inserted fragment is defined by the location of the restriction sites u.,~-d to obtain it. The numbering system is that ,3f figure I. ,[3-Galactosidase units (]3-Gal units) and 13-1actamase units (Bla units) were detem'fined as described in M a t e r i a l s ttttc+ Methods but cells were grown to log phase at the temperature indicated in the table. The results represent the averages of at least 5 independent experiments with the standard deviation usually not exceeding 10%. Note that the weak read-through from some undefined promoters on pCB267 is augmented several-fold in the rh+~ strain at 42°C. A similar effect has been observed by other authors using vectors carrying galK as indicator gene [ ! 1. 17]. al~13allBla = [~galactosidase units/l~-Iactamase units, eTh'e percentage termination was calculated according to the lbrmula: 100-- [~--Gai/Bla of terminator plasmid - ~-Gal/Bia of pCB267 x 100. [3-Gal/Bla of plC398 - I]-Gal/Bla of pCB267 J B R R 3 both at 30 and 42°C. Therefore, it appears that terminators T! and T2 are Rho-independent. However, the tenninatien efficiency for plC394 decreases from 61% in the M C I 0 0 rho+ host to 32% in J B R R 3 at 42°C. This suggests that termination at T3 is dependent on the host Rho factor because T3 is the only te~nination signal carried by piC394 that is not present in plC406. Since T ! and T2 are located upstre,',m from T3 in p i C 3 9 4 and seem to be Rhoindependent, we cannot expect a great decrease in the efficiency o f termination tbr this plasmid after thermal inactivation o f the Rho protein. Experiments are at present underway to further analyze the degree of T3 terminator Rho dependence. On the other hand. the apparent Rho independence of TI and T2 does not rule out the possibility that termination at these signals mediated by other protein factor,,, particularl 3 if we consider that. as mentioned betbre, neither of them shows the typical structure of simple terminators. T a n d e m Rho-dependent and Rho-independent termination sites have also been detected in the lacZ and rho genes of E coli [20, 21] as v~:ell as in the hisC gene of S t y p h i m u r i u m [22]. E w r e s s i o n +~" the recF gem" d e p e n d e n t on the dnaA prontolers

To analyze.+ the effect of the d n a A promoters on r e e F expressmn we constructed a series of translational

fusions between r e e F and the l a c Z reporter gene carried by plasmid p M L B 1034. These fusions contain 144 bp o f r e c F - c o d i n g sequence at the NH2-terminal region, the translation initiation signals of recF, and 5'-flanking sequences of increasing length, including or not the d n a A promoters. Piasmids piC361 is a r e c F - l a c Z fusion where expression of the r e e F gene is solely mediated by its own promoters. As can be seen in table II, the ~-Gai/Bla ratio for plC361 is very low. This must be due. at least partially, to the activity of the T l, T2 and T3 terminators which, as mentioned before, accumulatively stop transcription started at the r e e F promoters by = 80% ( A r m e n g o d et al, submitted). Expression of r e e F increases significantly when the d n a N promoters are included in the fusion plasmid (compare p l C 3 5 6 and plC361 in table II). This is consistent with our previous results indicating that d n a N and r e e F constitute an operon under control o f the d n a N promoters [6, 71. The 20-fold difference between the I]-Gal/Bla ratios o f p l C 3 5 6 and plC361 suggests that r e e F expression depends on the d n a N promoters to a greater extent than on the r e c F promoters, at least when the d n a N - r e c F region is present on a multicopy plasmid and cells are growing under normal physiological conditions. It should be noted, however, that the 13-Gal/Bla ratio o f plC356 is -- 37 times lower than that obtained for plC340, a d n a N l a c Z fusion where d n a N expression is under control o f its own promoters (see table II).

Regulatory elements of the dllaA-dnaN-re~'F The unequal expression of the dnaN and recF genes mediated by the thtaN promoters suggests that subsequent transcriptional and/or translational signals modulate the synthesis of the reeF gene product. However, differences in measured 13-galactosidase activity might also be due to different specific activity or stability of the fusion proteins. At present, we can rule out autoregulation of the d n a N gene in order to explain the lower 13-Gal/Bla ratio of plC356 (which carries the entire dnaN gene) since a similar ratio is obtained for plC409, a plC'~Sf-derived plasmid carryin~ the dnaN176 deletion (see table II). Moreover, we have recently shown that the TI, T2 and T3 terminators, located in the second half of dnaN, do not affect transcription coming from the d n a N promoters, at least when the dnaN transcription and translation processes are tightly coupled (Armengod et al, submitted). Thus, table II shows that the 13-Gal/Bla ratio of plC371, a dnaN-lacZ fusion containing T1, T2 and T3, is similar to that obtained for plC340, a dnaNlacZ fusion in which these terminators are not present. It therefore seems reasonable to conclude that T I , T2 and T3 are not responsible for the unequal expression of the d n a N and reeF genes mediated by the d n a N promoters, at least under the conditions used in the experiments in table II. Maximal expression of the reeF gene, as measured by the I]-galactosidase activity, is reached when the dnaA promoters are inchJded in the fusion plasmid. Thus, plC435 and plC440 show a 13-Gal/Bla ratio = 7fold greater than that obtained for plC356 where, as seen before, the expression of reeF is solely mediated by the dn a N and reeF promoters (table II). Note that plC435 and plC440 carry the dnaN176 deletion. We

33

considered this to be advantageous because it has been reported that multicopy plasmids carrsing, both dn~,.A and dnaN genes are somewhat unstable, at least when transformed in a dnaA46 recA1 strain [2]. We have previously shown that the d n a N I 7 6 deletion neither changes the dnaN reading frame nor affects reeF expression [6]. This can also be inferred from results in table II (compare the [3-Gal/Bla ratio of plasmids plC356 and p1C409). We later constructed piC438, a recF-lacZ fusion carrying both wild-type dnaA and d n a N genes. As can be seen in table II. no differences in the [3-Gal/Bla ratio were detected between piC438 and its similar pIC440, which carries the dnaN176 deletion. The fact that reeF expression increases when the dnaA promoters are present in the fusion plasmid strongly supports the idea that reeF is a distal gene of the dnaA operon. Note that the increase in the reeF expression is remarkable despite the fact that piC435, pIC440 and piC438 provide multiple copies of the wild-type dnaA gene. It is known that overproduction of the DnaA protein from multicopy plasmids leads to decreased expression of the dnaA promoters [23.24]. Finally, it should be noted that a 5-fold increase in dnaN expression is obtained when the dnaA promoters and the ,~ntire dnaA gene are included in the dnaNlacZ fusions (see pIC340, pIC371, piC437, and piC439, in table II). This supports the previous proposal that d~iaN is mostly expressed from the dnaA promoters [4, 5]. Our results suggest that both d n a N and reeF genes are predominantly transcribed from the dnaA promoters: however, it is known that regulatory regions may not respond normally when present in a large number

Table II. Expression of the reeF gene dependent on the dnaA promoters. Plasmid~'

Fragment b

p1C361 piC356 plC409 plC435 plC440 plC438

1490--2586 (-3)-2586 (-3)-2586 (-948)-2586 (-389)-2586 (-389)-2586

plC340 plC437 plC37 ! plC439

(-3)-i490 (-389)- 1490 (-3)-2419 (-389)-2419

Promoters~

fl-Gal units d

Bla unit~~

,CCGal/Bla~

reeF reeF reeF reeF reeF reeF

0.21 3.77 3.00 20.63 16.84 13.96

! 0.35 8.39 6.80 7.94 4.58 3.99

0.02 0.45 0.44 2.60 3.68 3.50

dnaN dnaA, dnaN dnaN dnaA, dnaN

92.85 412.70 158.50 453.30

5.36 4. I i 9.52 5.05

! 7.32 i 00.4 i 16.65 89.76

dnaN. dnaN, ~hlaA, dnaN, dnaA. dnaN, dnaA, dnaN,

•~See plasmids construction in Materials and Methods. Note that plC409, plC435 and plC440 contain the dnaN176 deletion. bChromosomal fragment fused to the lacZ gene. The numbering system is that of figure 1. cPromoters included in the chromosomal fragment, dl3-Galactosidase units (13-Gal units) and [~-Iactamase units (Bla units) were determined as described in Materials and Methods. The results represent the averages of at least 3 independent experiments with the standard deviation usually not exceeding 10%. el~-Gal/Bla = [:l-galactosidase units/13-1actamase units. The I~-Gal/Bla ratio for pMLB1034 was 0.004. The recF-lacZ and dnaN-lacZ fusions are shown on the upper and lower parts of the table, respectively.

3.L4

! P6rez-Roger et al

o | copies per ceil. as h a p p e n s in the e x p e r i m e n t s sho,,vn in table II. Theretbre. e x p e r i m e n t s using a single cop3' o f the d m ; A - t h ~ a N - r e c F region should be p e r f o r m e d to d e t c m f i n e precisely to what extent d n a N and r e e f are expressed from the promoters o f the preceding genes and to what extent this is done from their o w n promoters. The m u l t i p l e regulator).' e l e m e n t s found w i t h i n the d n a A o p e r o n (16. 71; Annengod et al, s u b m i t t e d ) could be used to a c c o m p l i s h discoordinate regulation as well as differential e x p r e s s i o n o f the operon genes. In tact, it has been reported that s o m e o f the d n a N promoters are inducible in a d a m b a c k g r o u n d u n d e r conditions o f s i m u l t a n e o u s l y reduced d n a A tran:,u,:ription [~51. Moreover, we h a v e found that both tbzaN and r e c F expression, but not d n a A expression, are induced by D N A d a m a g e (Navarro-Avifi6 et al, in prepa:ation). By contrast, the T l, T2 and T3 terminators m a y contribute to reduce expression o f the distal g,~nes o f the operon u n d e r c i r c u m s t a n c e s w h e r e d n a N translation h a p p e n s to be s l o w e d d o w n ( A r m e n g o d et al, submitted). F i n a l l y , transcriptional and/or post-transcriptional m e c h a n i s m s should be i n v o l v e d in specifically lowering e x p r e s s i o n o f the r e e F gene. T h i s is supported by the e x p e r i m e n t s h o w n in table II, as well as previous results indicating that signals within the r e e f coding f r a m e cause transcription t e r m i n a t i o n or m R N A instability ([7, 26]; A r m e n g o d et al, submitted). Acknowledgments We wish to thank A Komberg for the gift of plasmid pBF110. We are ve~, grateful to E Knecht and F Thompson for help also thanks Glaxo SA for providing us with the nitrocefine. This work was supported by grants PB85-0263 and PB880461. the International Pro=re'am of Molecular Cytology of me IIC-KUMC and by an Institutional Grant from the Areces Foundation. ! PR and F M are predoctoral fellows from MEC and C LA from CCEV-GV.

References 1 Torrey TA. Kogoma T ~1987} Genetic analysis of constitutive stable DNA replication in rnh mutants of E coli K 12. :il~: Gen Gone( 208. 420-427 2 Ohmori H. Minoru K. Nazata T. Sakakibara Y (1984) Structural analysis of the d~'[aA and dnaN genes of E coli. Gt'ne cAmslJ 28. 159-170 Btanar MA. Sandier SJ. Armengod ME. Ream LW, Clark .-MI (1984) Molecular analysis of the recF gene of E coli. Proc Natl Acad Sci USA 8 I. 4622-4626 4 Sakakibara Y. Tsukano H. Sako T (1981) Organization and transcription of the dnaA and dnaN genes of E coli. Gene tAms() 13.47-55 5 Sako T. Sakakibara Y (1980) Coordinate expression of Escherichia coli dnaA and dnaN genes. Mol Gen Genet ! 79. 521-526 6 Armengod ME. Garc/a-Sogo M. Lambfes E (1988) Transcriptional organization of the dnaN ~nd recF genes of E coli K- 12. J Biol Cheat 263, 12109-12 ! 14

7 Armengod ME. Lambfes E (1980) Overlapping arrangement of the recF and dnaN operons of Escherichia coli: positive and negative control sequences. Gene (Amst) 43. 183-196 8 McKenny K, Shimatake H, Cgurt D, Schemissner U, Brady C, Rosenberg M ( ! 981 ) A system ~o study promoter and terminator signals recognized by Escherichia coli RNA polymerase. In: Gene Amplification and Analysis (Chirikjian JG, Papas TS, eds) Elsevier Scientific l~ublishing Co, Amsterdam, vol 2, 383--415 9 Silhavy T J, Berman ML, Enquist LW (1984) Experiments with Gene Fusions. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY 10 Miller JH (1972) Experimenls in Molecular Genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY i l Dalrymple B, Arber W (1986) The characterization of terminators of RNA transcription on IS30 and an analysis of their role in IS element-mediate polarity. Gene (Amst) 44, 1-10 12 Fuller RS, Komberg A (1983) Purified dnaA protein in initiation of replication at the E coli chromosomal origin of replication. Proc Natl Acad Sci USA 80, 5817-5821 13 Mania(is T, Fritsch EF, Dambrook J (1982) Molecular Chining. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY 14 Andrup L, Atlung T, Ogasawara N, Yashikawa H, Hansen FG (1988) Interaction of the B subtillis DnaA-like protein with the E coli DnaA protein. J Bacteriol 170, 1333-1338 15 Yager TD, von Hippel PH (1987) Transcript elongation and termination in Escherichia coli. In: E coli and S thyphimurium, Celhdar and Molecular Biology (Neidhardt FC, ed) American Society for Microbiology, Washington, vol 2, ! 241-1275 i 6 Schneider K, Berck CF (1987) New expression vectors for identifying and testing signal structures for initiation and termination of transcription. Methods En:ymol 153. 452--461 17 Luk KC, Szybalski W (1082) Characterization of the cloned terminators ta,, tL3 and t, and the nutR antitermination site of coliphage lambda. Gel~e tAms() 20, 127-134 i8 Madden KA, Landy A (1989) Rho-dependent transcription termination in the t3.'rT operon of E coll. Gene (Amst) 76, 271-280 19 Xian-Ming Y, Munson LM, Reznikoff WS (1984) Molecular cloning and sequence analysis of trp-lac fusion deletions. J Mol Biol 172, 355-362 20 Richardson JP. Ruteshouser EC (1986) Rho factor dependent transcription termination: interference by a mutant rho. J Mol Biol 189.413-419 21 Matsumoto Y. Shigesada K, Hirano M, Imai M (1986) Autogenous regulation of the gene for transcription termination factor rho in E coli: localization and function of its attenuators. J Bacteriol 166, 945-958 22 Alifano P, Ciampi MS, Nappo AG, Bruni CB, Carlomagno MS (1988) In vivo analysis of the mechanisms responsible for strong transcriptional polarity in a 'sense" mutant within an intercistronic region. Cell 55, 351-360 23 Braun RE, O'Day K, Wright A (1985) Autoregulation of the DNA replication gene dnaA in E coli K-12. Cell 40, 159-169 24 Atlung T, Clausen ES, Hansen FG (1985) Autoregulation of the dnaA gene of E coli K I2. Mol Biol Gen Genet 200, 442-450 25 Quifiones A, Messer W (1988) Discoordinate gene expression in the dnaA-dnaN operon. Mol Gen Genet 213, 118-124

26

Sandier SJ, Clark AJ (1990) Factors affecting expression of the reeF gene of E coli K- 12. Gene 86, 35--43