CURRENTMICROBIOLOGYVol. 20 (1990), pp. 83-90

Current Microbiology © Springer-Verlag New York Inc. 1990

Expression of Heat-labile Enterotoxin Genes is under Cyclic AMP Control in Escherichia coli Isidre Gibert, Virtudes Villegas, and Jordi Barb6 Department of Genetics and Microbiology,Facultyof Sciences, AutonomousUniversityof Barcelona, Barcelona, Spain

Abstract. The expression of the linked toxA and toxB genes coding for the A and B subunits of the heat-labile enterotoxin of Escherichia coli, respectively, has been studied. For this reason, t o x A - l a c Z and t o x B - l a c Z fusions were constructed through a promoter-probe vector. Results obtained show that the toxB gene has a promoter from which it may be transcribed independently from the toxA gene. Nevertheless, the expression of toxB gene is about 25-fold higher from the toxA promoter than from its own promoter. Furthermore, the presence of glucose in the culture medium decreased the transcription from both toxA and toxB promoters. Likewise, cya and crp mutants exhibited a lower level of tox gene expression than did the wild-type strain. The addition of cyclic AMP increased the expression from toxA and toxB promoters in both cya and wild-type strains, but not in a crp mutant. All these data suggest that the cyclic AMP receptor protein-cyclic AMP complex positively modulates toxA and toxB gene transcription in E. coll.

Enterotoxigenic Escherichia coli carries the genes for enterotoxin production, which may originate a diarrheal disease in humans and some farm animals. At least five types of E. coli enterotoxins have been characterized [2]: the human (LT-H) and porcine (LT-P) heat-labile toxins, the human (STI-H) and porcine (STI-P) heat-stable enterotoxins, and the heat-stable enterotoxin II (STII). The genes encoding these enterotoxins are present on transmissible plasmids. LT consists of two subunits, LTA and LTB [7, 9]. LTB binds to the receptor (GM1 ganglioside) of the target cells, and after binding takes place, LTA stimulates the adenylate cyclasecAMP system of the cells [10], resulting in diarrhea in the intestine. Besides their similar modes of action, LTA and LTB also share similar immunological properties with Vibrio cholerae enterotoxin subunits A and B, respectively [7, 30]. The genes for LT-H have been designated elth or tox [10, 37], whereas the genes for LT-P have been designated elte [8, 30]. Both -tox and elt genes have been cloned and sequenced [19, 29, 33, 38]. LT-H and LT-P are closely related immunologically but are not identical [12, 33, 38]. The LT genes may be flanked by inverted and directed repeats [34], but

transpositional activity has not been demonstrated. The toxA and toxB genes form an operon, with toxA proximal to the promoter [29]. The terminal amino acid codon for toxA overlaps with the initiation codon of toxB by one base pair [30, 39]. Recently, the toxB genes have been subcloned, and very slight B subunit synthesis was detected [26, 35], although it has not been determined whether this transcription of the toxB gene was carried out from either a vector promoter or from its own promoter. In this latter case, it would be interesting to know the regulation of the expression from both toxA and toxB promoters as well as the relative strength of each of them. In this respect, it is also known that medium composition can affect LT production in the laboratory [15], although this aspect of the LT toxin production has not been exhaustively studied. For all these reasons, in the present work we have tried to determine the hypothetical presence of an independent promoter for the toxB gene as well as the effect of the cyclic AMP (cAMP) on the expression of both toxA and toxB genes. To do that, several gene fusions between the l a c Z gene and either the toxA or the toxB gene have been constructed by using a promoter probe-vector.

Address reprint requests to: Dr. Jordi Barbr, Departmentof Geneticsand Microbiology,Facultyof Sciences, AutonomousUniversity of Barcelona, Bellaterra, 08193-Barcelona, Spain.

84

CURRENT MICROBIOLOGY Vol. 20 (1990)

Table I. Escherichia coli K12 strains used in this work Strain

Genotype

Source

C600

thi-1 thr-1 leuB6 lacY1 tonA21 supE44 araD139 A(ara-leu)7697 AlacX74 galU galK hsdR rpsL endAl hsdR17 supE44 thi-I recA1 gyrA96 relA1 A(lacZYAargF) U169 F'C8OdlacZAM15 lacZ thi galK phoA8 ilv rpsL recA56 hsdR xyl ilvA argH1 AlacX74 As TP2100 but Acya As TP2100 but Acrp As C600 but harboring 1032H-19 As MC1061 but harboring pUA76 As DH5aF' but harboring pUA77 As MC1061 but harboring pUA78 As MC1061 but harboring pUA82 As TP2006 but harboring pUA76 As TP2139 but harboring pUA76 As TP2006 but harboring pUA78 As TP2139 but harboring pUA78 As TP2006 but harboring pUA82 As TP2139 but harboring pUA82 As TP2006 but harboring pUA85 As TP2100 but harboring pUA54 As TP2006 but harboring pUA54 As TP2139 but harboring pUA54

This laboratory

MC1061 DH5e~F'

CB977 TP2100 TP2006 TP2139 UA4551 UA4611 UA4624 UA4626 UA4642 UA4653 UA4655 UA4640 UA4657 UA4654 UA4656 UA4659 UA4481 UA4480 UA4482

D N A techniques. DNA techniques were carried out as previously described [13]. Computer analyses of both toxA and toxB sequences were performed with the Staden-Plus personal computer DNA software from Amersham.

M. Casadaban

Results

This laboratory

Fusion o f tox genes to the lacZ gene. 1032H-19 is a

C,F. Beck J.R, Guest J.R, Guest J,R, Guest H, Danbara This work This work This work This work This work This work This work This work This work This work This work This laboratory This laboratory This laboratory

Materials and M e t h o d s Bacterial strains and plasmids. Bacterial strains and plasmids used in this work are listed in Tables 1 and 2. Plasmid pUA85, harboring a fusion between the recA and the lacZ genes, was constructed from the pCB267 promoter-probe vector and a 1.25 kb BamHI DNA fragment from the pUA80 plasmid carrying the promoter and the NH2-terminal regions of the recA gene. Plasmid pUA80 had been previously constructed from the pUC7 cloning vector, and 6.9 kb PstI DNA segment from pDR1453 [25] plasmid containing the recA gene. Growth conditions. Cultures were grown at 37°C with shaking in either LB rich medium [2211or AB minimal medium [6], supplemented with 10/xg/ml thiamine, 0.2% (wt/vol) glucose, and 0.4% (wt/vol) casamino acids, When necessary to determine the possible effect of methionine and lysine on tox gene expression, the cells were grown in AB minimal medium with glucose and all amino acids with the exception of methionine and lysine, which were added at 200/zg/ml when the cultures were at the middle of the exponential phase of growth. Antibiotics were added to the culture medium at the concentrations indicated by Maniatis et al, [20]. LB plates containing 40/zg/ml of 5-bromo-4-chloro-3-indolyl-fl-o-galactopyranoside (X-Gal) were used to screen colonies for/3-galactosidase production. fl-Galactosidase assay. The fl-galactosidase assay was performed as described [22]. Enzyme concentrations were calculated from the formula given by Casaregola et al, [5],

pBR322 derivative plasmid harboring two HindIII fragments of 1680 and 670 bp from the enterotoxic plasmid pMYI900 inserted into the unique HindIII site [18]. These two HindIII fragments contain the toxAB genes and are able to produce heat-labile toxin type h [35]. toxA and toxB genes have been sequenced, and their direction of transcription is also known [35]. To perform the gene fusions between toxAB genes and the lacZ gene, the plasmid pCB267 was used as the vector. This plasmid pCB267 is a promoter-probe vector with a high copy number derived from the pKO plasmid; it contains the lacZ and phoA indicator genes divergently oriented and separated by a polylinker [27]. In addition, in this polylinker there are also translational stop codons in all three reading frames upstream of the initiation codons as well as ribosome-binding sites before the ATG codon. t o x A - l a c Z fusion was constructed following the strategy shown in Fig. 1. Plasmid 1032H-19 was digested by HindIII, and the fragment of 1.68 kb containing the promoter and the NHz-terminal region of the toxA gene was ligated with plasmid pCB267, which previously had been digested with HindIII. The ligation products were transformed into Escherichia coli CB977, and blue colonies on X-Gal-supplemented plates were selected. Afterwards, the plasmids present in these clones were tested by restriction endonuclease analysis to corroborate that the toxA promoter was located upstream of the tacZ gene. One of the appropriate plasmids was kept and designated as pUA76. To obtain a t o x A B - I a c Z fusion, we followed the schedule shown in Fig. 2. 1032H-19 plasmid was digested by both BamHI and PstI enzymes, and the fragment of 3.42 kb was ligated with plasmid pUC19, which had been previously digested with BamHI and PstI. After transformation in E. coti DH5aF', white colonies in X-Gal-supplemented plates were purified, and the plasmid obtained was called pUA77. Afterwards, plasmid pUA77 was digested by SmaI. The fragment of 2.29 kb, originated by the SmaI site located within the toxB gene and by the SmaI site of the pUC19 plasmid flanking the BamHI site insertion, was ligated in the SmaI site of

I. Gibert et al.: Enterotoxin Genes in Escherichia coli

85

Table 2. Plasmids used in this work Plasmid

Genetic characteristics

Source

1032H-19 pCB267 pUC19 pUA76 pUA77 pUA78 pUA82 pUA85 pUA54

Amp R ToxAB ÷ Amp R LacZ PhoAAmp R LacZ + As pCB267 but harboring toxA-lacZ fusion As p U C t 9 but harboring toxAB genes As pCB267 but harboring toxAB-lacZ fusion As pCB267 but harboring toxB-lacZ fusion As pCB267 but harboring recA-lacZ fusion As pFR97 but harboring ubiG-lacZ fusion

H. Danbara C.F. Beck This laboratory This work This work This work This work This work This laboratory

plasmid pCB267. This 2.29 kb fragment contains the toxA gene and the NH2-terminal region of the toxB gene. The ligation products were transformed in E. coli CB977 strain, and after being plated on X-Galsupplemented plates, blue colonies were purified and the presence of t o x A B - l a c Z fusion was tested by restriction analysis. One of these plasmids, pUA78, represented the desired fusion. Finally, and to determine whether the toxB may be expressed independently from the toxA promoter site, the scheme followed to obtain a toxBlacZ fusion is outlined in Fig. 3, pUA78 plasmid was digested by HindlII and the fragment of 0.425 kb containing the COOH-terminal region of the toxA gene, and the possible promoter region of the toxB gene as well as the NH2-terminal region of the toxB gene was ligated to the HindlII-digested pCB267 plasmid. After transformation in E. coli CB977, blue colonies were purified. By using restriction analysis, we were able to show that plasmids present in these blue colonies were all oriented with the toxB gene upstream of the lacZ gene, indicative that the toxB gene may be expressed in the absence of the toxA promoter. Behavior of the different tox-lacZ fusions. As cited above, cells containing the t o x B - l a c Z fusion gave blue color in X-Gal-supplemented plates, showing that the toxB gene may be transcribed from a promoter that is exclusively its own. For this reason, we compared the relative strengths of the toxA and toxB promoters by analyzing the basal level of expression of toxA-lacZ, t o x A B - l a c Z and toxB-lacZ fusions. Table 3 shows that basal expression of the toxA-lacZ fusion is about 25-fold higher than the basal level of the t o x B - l a c Z fusion, indicative that the toxA promoter is considerably stronger than the toxB promoter. In agreement with this, expression of the toxB gene from the toxA promoter was considerably higher than from its own promoter. On the

EcoRI

Hlncl ~I

Hind/n

/'-°

amH~ Hnic~lE~~ I

d|gestlon

Hn i cn'lT

~

Xbal

p U A 7@

~

H|nd n' T

Fig. 1. Schematic diagram of the procedure of construction of the toxA-lacZ fusion. In all plasmids, the sequence ( ~ ) means the polylinker region. Only the positions of the restiction sites relevant as landmarks to plasmid construction are given. The directions of the transcription of several genes are indicated (arrows).

other hand, plasmid pCB267 did not give any /3galactosidase activity when introduced in either MC1061 or TP21000 strains (data not shown). Furthermore, Table 3 also shows that expression of t o x A - l a c Z fusion is lower in media in which glucose is present, this inhibitory effect being stronger in rich medium than in minimal medium. Likewise, the basal level of expression of t o x A - l a c Z fusion is considerably lower in cya and crp mutants than in

CURRENT MICROBIOLOGY Vol. 20 (1990)

86

BamHI EcoRI PSt ]~..~ ~ H

fff"

"~ ~

Ind ]lI

l P s ~ m a T

~I:OXB"~"-- EH]O;~I\

7

[/¢

.....

\

/B.mH,.n0.st,"<.

""-~.~HInOIH

/~

double

/

digestion

ndllI Amp

pUA 7 7 ,oxE 612KD

Sm3T digestion

toxA

SmaI

/ /

phoA

A,

Hind ]El !gKb

\ toxA

pUA76 mp

1019Kb -EcORI

Sma'r

the wild-type strain (Table 3). On the other hand, the basal level of a pCB267 derivative harboring a r e c A - l a c Z fusion was not affected by either the presence of glucose or the the cya and crp mutations (Table 3), in agreement with previous data showing that this gene is not under glucose repression [1]. This fact indicates that the behavior of the t o x A - l a c Z fusion against the glucose must not be attributed to the pCB267 vector but to the toxA gene regulation, t o x A B - l a c Z and t o x B - l a c Z fusions showed the same behavior as the t o x A - l a c Z fusions; however, and to confirm data presented in Table 3, the effect of cAMP over the expression of toxA and toxB genes in the wild-type, cya, and crp strains was studied. Results obtained showed that cAMP increases the expression of the three fusions in both wild-type and cya strains, but did not produce any increase in these fusions when present in a crp mutant (Table 4). Likewise, transcription of r e c A - l a c Z fusion was not affected by cAMP addition (Table 4). Furthermore, the effect of cAMP

Fig. 2. Schematic diagram of the procedure of construction of the t o x A B - h u Z fusion. In all plasmids, the sequences ( ~ ) and (~/~) mean the polylinker regions. Only the positions of the restriction sites relevant as landmarks to plasmid construction are given. The directions of the transcription of several genes are indicated (arrows).

was higher in cells growing in rich medium than in minimal medium (Table 4). Finally, it is worth noting that the behavior of all t o x - l a c Z fusions against the glucose and the cAMP was the same as that shown by an u b i G - l a c Z fusion (Tables 3 and 4), which is under the cAMP-CRP complex control [131. Analysis of t o x A and toxB sequences. Our data show that toxB presents its own promoter from which it may be transcribed, besides coming from the toxA promoter. The sequence of both toxA and toxB genes is known [19, 29, 33, 38], and it has been established that there is an overlapping between the terminal acid codon for the toxA gene and the initiation codon of the toxB gene by one base pair [30, 39]. To confirm the data obtained in vivo about the promoter of the toxB gene, we have performed a computer study of the upstream region of the first codon of the toxB gene. This analysis showed the presence of the sequences TTCAGA and TATATA

I. Gibert et al.: Enterotoxin Genes in Escherichia coil

87

Table 3. Basal level of transcription of the toxA-lacZ, toxAB-lacZ, and toxB-lacZ fusions in several strains of Escherichia coil growing in different culture media Specific units of fl-galactosidase" strain and media Wild Type Fusion toxA-lacZ toxAB-lacZ toxB-lacZ ubiG-lacZ b recA-lacZ c

LB 770 438 30 6.2 230

Cya

Crp

LB + glucose AB + glucose AB + glycerol LB + glucose AB + glucose LB + glucose AB + glucose 240 t50 8 3.3 240

550 216 12 3.5 220

830 435 40 9.3 230

150 120 4 1.8 250

140 110 5 1,6 220

130 120 5 1.6 230

140 105 3 1.8 240

The ~-galactosidase assays were performed in exponentially growing cultures at concentrations of about 5 × l0 s cfu/ml. All values are reproducible to within an error of -+ 10%. b The behavior of an ubiG-lacZ fusion whose transcription is repressed by the glucose [13] is also shown as a positive control. As a control of a possible effect of the medium in the copy number of the plasmids, the behavior of plasmid pUA85, a pCB267 derivative harboring a fusion between the recA and lacZ genes (see Materials and Methods), was also studied.

(Fig. 4), spaced 20 nucleotides apart, which are consistent with the - 3 5 and - 10 region of a promoter, and the following the method of Staden [31], present a homology score of 72 and 80 respectively with the - 3 5 and - 1 0 regions of known E. coil promoters. Table 3 indicates that cAMP can modulate the expression of both toxA and toxB genes. For this reason, we also performed a computer search for homology with consensus sequences derived from known binding sites of the cAMP-CRP complex in the upstream regions of toxA and toxB start codons. Several slightly different cAMP-CRP consensus sequences have been proposed, such as A A A G T G T G A C A , A / G - - T G T G / C A C A - - C / T and TGTGNsCACA [11, 32]. Using these consensus sequences, we were able to localize a DNA region upstream of the toxA promoter and another upstream of the toxB promoter that showed a homology score of 66 and 77 respectively with the cAMPCRP consensus sequences. The toxB-related cAMP-CRP consensus sequence was located upstream of the - 3 5 region of the promoter of this gene found by us. In the toxA gene, the cAMP-CRP consensus sequence was overlapping the more distal of the two putative - 3 5 regions that the promoter of this gene presents [38].

Discussion In this study, we have shown that the expression of the toxB gene may be accomplished from the toxA promoter and from its own promoter, which is overlapping the COOH-terminal region of the toxA

Sinai

Hlnd m

jo,, ,o Hind RT

Fig. 3. Schematic diagram of the procedure of construction of the toxB-lacZ fusion. In all plasmids, the sequences ( ~ ) , (5"~) and ( ~ ) mean the polylinker regions. Only the positions of the restriction sites relevant as landmarks to plasmid construction are given, The directions of the transcription of several genes are indicated (arrows).

gene. These data are in agreement with previous results showing that plasmids harboring the 670 bp HindlII fragment of pMY1900 containing the toxB gene gave rise to a low production of the LTB sub-

CURRENT MICROBIOLOGYVol. 20 (1990)

88

Table 4. Effect of cAMP addition in the expression of the t o x A - l a c Z , t o x A B - l a c Z , and t o x B - l a c Z fusions in several strains of Escherichia coli growing in different culture media

Relative specific units of/3-galactosidase" strain and media Wild Type Fusion

Cya-

Crp-

LB + glucose

AB + glucose

LB + glucose

AB + glucose

LB + glucose

AB + glucose

2.15 2.1 2.5

1.9 1.8 2

5.8 4,8 6

4.8 3.6 4.4

1 1 1

1 1 1

2,2 1

2.1 1

6.3 1

6.5 1

1 1

1 1

toxA-lacZ toxAB-lacZ toxB-lacZ ubiG-lacZ b recA-lacZ C

The/3- galactosidase assays were performed in exponentially growing cultures at concentrations of about 5 × 108 cfu/ml and 2 h after cAMP addition at 10 mM, Each value is referred to the basal level of its respective fusion growing in the absence of cAMP and is reproducible to within an error of -+ 10%. b The behavior of an u b i G - l a c Z fusion whose transcription is regulated by the cAMP-CRP complex [13] is also shown as a positive control of the cAMP effect. c As a control of a possible effect of the cAMP in the copy number of the plasmids, the behavior of plasmid pUA85, a pCB267 derivative harboring a fusion between the recA and a c Z genes (see Materials and Methods), was also studied.

® -t60

-150

-140

-~JO

-t20

-ilO

COTSCACTCT TTCTTTATC;G CTTCAACTAC ACATTTTATC C T ~ T G G A -80

-IO0

- 70

"60

TGTTTTATAA -50

A~AACAT~T TGACATCATG TTOCAT_~G 6TTAAACAAA AC/~GTGOCC TTATCTTTTT I (-~sl ~1 ~ tOxA

_a,c

-20

-20

TTCTTGTATG A ~ G T

®

-I0

-~

I

Ira,,

TTTTCCCTCG ATG ~ ~u~T ATA ACT TTC ATT Met Lys Asn lle Thr Phe Ile

® toxA -tTO

-I~0

-15~

-t40

-120

-120

(~ACC~C~G OTTGTO~.GA TTCA~CAAGA ACAATTACAG GTC~TACTTG TAATGAOGAG toxA -90 -aO -70 - t 10 -~00 ACCC_~GAATC TC~OC~CAAT ATATCTCAGO AAATATCAAT ~ T T A A

-~0 GACK)CAGATA

toxA tox B -30

-20

-I0

-~ ;

~--

TTTTCAGACT ATCAGTCAGA GOTTC~ACATA TAT~CAC~a¢x, TTC:OCVVkTCxA ATT AT(3 P@,T

® GT~ ~ Val W s

Q TOT TAT Cys Tyr

u n i t [26, 36]. T h e a u t h o r s o f t h i s w o r k p r o p o s e d t h a t this synthesis could be due to an inefficient prom o t e r in t h e v e c t o r o r t o a v e r y w e a k p r o m o t e r in t h e c l o n e d f r a g m e n t . O u r r e s u l t s s h o w n in T a b l e 2 give support to the second hypothesis. The possible

Met ASh Lys

Fig. 4. Nucleotide sequences of upstream regions of the first codon of toxA (A) and toxB (B) genes. Each nucleotide sequence is numbered from its respective transcription initiation site (+ 1). The putative -35 and - 1 0 regions in both sequences are underlined. The putative CRP binding sites found by us in both sequences after computer analysis are enclosed in solid boxes. The sequences shown are from [38] and [35] respectively.

role of the toxB promoter within the tox operon may c o n s i s t in o v e r c o m i n g t h e p o l a r i t y o f t h e p o l y c i s t r o n i c o p e r o n s t o m a i n t a i n t h e p r o d u c t i o n o f a sufficiently large quantity of the B subunit of the LT p r o t e i n to e n s u r e a p r o p e r m o l a r r a t i o b e t w e e n b o t h

I. Gibert et al.: Enterotoxin Genes in Escherichia coli subunits, since it is known that holotoxin consists of one molecule of subunit A and five molecules of subunit B [14]. So, the normal transcription of the toxB gene would be from the toxA promoter, whereas the toxB p r o m o t e r would collaborate to maintain a high synthesis of toxB mRNA. In this way, a sufficient level o f subunit B would be achieved by the combination of two processes: the presence o f toxB p r o m o t e r reported in this work, and the previously proposed discriminated role of the local secondary structure of m R N A in the efficiency of both toxA and toxB m R N A translations [39]. A similar case of overlapping operons has been reported in the f r d - a m p C system of Escherichia coli [24]. The a m p C operon of E. coli K-12 codes for the fumarate reductase complex and is composed of four genes [3, 28]. The a m p C p r o m o t e r has been located i n the COOH-terminal coding region of the FrdD protein [17]. Then, it is known that different a m p C up-promoter mutations affect the f r d D gene product, changing several amino acid residues in the FrdD protein [16]. It has been speculated that this overlapping between f r d D and a m p C genes in E. coli could have been generated by a deletion of the region between these two operons in a species coding for an inducible ampC/3-1actamase [24]. Thus, ampC/3-1actamase production would be regulated by the growth rate under anaerobic conditions, because the f r d D operon is repressed under aerobic growth. We must not forget that the absence of oxygen is a condition prevailing in the large intestine, the natural ecological niche of E. coli. In this respect, a relationship between the evolution of the f r d - a m p C operon overlap and the antibiotic therapy has been pointed out. Nevertheless, there is still not a full explanation for the complexity of the structure of the tox genes: the toxB promoter and toxB encoding region overlapping the toxA encoding sequence. Further work in this field is necessary to explain this interesting question. It is also noting that results presented in this study show that the c A M P stimulates the expression of tox genes in a way similar to what happens with the synthesis of the heat-stabile enterotoxin I in E. coli [21]. Nevertheless, it had been previously reported that glucose supports the maximal synthesis of L T h [15]. A possible explanation for this apparent contradiction between the stimulating effect of cAMP in the tox genes and the positive effect of the glucose may be found in the fact that in this last case L T h activity was measured directly from the supernatant of enterotoxigenic-strain cultures, whereas in our work we have directly analyzed the

89 expression of tox genes. So, all these data suggest that the glucose must probably be required in the process of release of the L T from the cells to the media; however, our data about the role of c A M P upon the expression of the tox genes are clear (Table 4). F u r t h e r m o r e , it is also interesting that cAMP increases transcription from both toxA and toxB promoters, indicative that both promoters are under the same host regulation. It had already been described that the addition of methionine and lysine stimulates the heat-labile enterotoxin production by enterotoxigenic strains of E. coli growing in a basal salts medium [15]. Nevertheless, the addition of these amino acids did not produce any effect on the transcription o f either o f the t o x - l a c Z fusions (data not shown), showing that this previously described stimulation of both amino acids is not at the transcriptional level. H o w e v e r , the regulation of tox gene expression in E. coti is not so well known. So, several c h r o m o s o m a l mutations resulting in L T overproduction in E. coli have been isolated, although it has not been extensively characterized [4, 23]. In this respect, plasmids harboring the t o x lacZ fusions constructed in this study may be very useful in isolating mutants with either increased or decreased L T synthesis, enabling us to determine in an easier way whether these mutations are implied in L T leakage or in regulation of the tox gene transcription. ACKNOWLEDGMENTS

We thank C.F. Beck, M. Casadaban, H. Danbara, and J.R. Guest for their generous gifts of several plasmids and strains and J.M. Cuartero for drawing the figures. This work was supported by grant No. 1129/86of the Fondo de Investigaciones Sanitarias, Spain (to J,B.). Literature Cited

1. Barb6 J, Gibert I, Guerrero R (1987) Modulating action of cyclic AMP on the expression of two SOS genes in Escherichia coli. Can J Microbiol 33:704-709 2. Betley MJ, Miller VL, Mekalanos JJ (1986) Genetics of bacterial enterotoxins. Annu Rev Microbiot 40:577-605 3. Boman HG, Nordstr6m K, Normark S (1974) Penicillin resistance in Escherichia coli K-12: synergism between penicillinases and a barrier in the outer part of the envelope. Ann NY Acad Sci 235:569-586 4. Bramucci MG, Twiddy EM, Baine WB, Holmes RK (1981) Isolation and characterization of hypertoxinogenic (htx) mutants of Escherichia coli KL320 (pCG86). Infect Immunol 32:1034-1044 5. Casaregola S, D'Ari R, Huisman O (1982)Quantitative evaluation of recA gene expression in Escherichia coli. Mol Gen Genet 185:430-439 6. Clark D, MaalCe O (1967) DNA replication and the division cycle of Escherichia coli. J Mol Biot 23:99-112

90 7. Clements JD, Yancey RJ, Finkelstein RA (1980) Properties of homogeneus heat-labile enterotoxin from Escherichia coli. Infect Immun 29:91-97 8. Dallas WS (1983) Conformity between heat-labile toxin genes from human and porcine enterotoxigenic Escherichia coli. Infect Immunol 40:647-652 9. Dallas WS, Falkow S (1979) The molecular nature of heatlabile enterotoxin (LT) of Escherichia coli. Nature 277:4064O7 I0. Dallas WS, Gill DM, Falkow S (1979) Cistrons encoding Escherichia coli heat-labile toxin. J Bacteriol 139:850-858 11. de Crombrugghe B, Busby S, Buc H (1984) Cyclic AMP receptor protein: rote in transcription activation. Science 224:831-838 12. Geary S J, Marchlewicz BA, Finkelstein RA (1982) Comparison of heat-labile enterotoxins from porcine and human strains of Escherichia coli. Infect Immun 36:215-220 13. Gibert I, Llagostera M, Barb6 J (1988) Regulation of ubiG gene expression in Escherichia coli. J Bacteriol 170:13461349 14. Gill DM (1976) The arrangement of subunits in cholera toxin. Biochemistry 15:1242-t248 15. Gilligan PH, Robertson DC (1979) Nutritional requirements for synthesis of heat-labile enterotoxin by enterotoxigenic strains of Escherichia coli. Infect Immun 23:99-107 16. Grundstr6m T, Jaurin B (1982) Overlap between ampC and frd operons on the Escherichia coli chromosome. Proc Natl Acad Sci USA 79:1111-1115 17. Jaurin B, Grundstr6m T, Edlund T, Normark S (1981) The E. coli fl-lactamase attenuator mediates growth rate dependent regulation. Nature 290:221-225 18. Kirii Y, Danbara H, Komase K, Arita H, Yoshikawa M (1987) Detection of enterotoxigenic Escherichia coli by colony hybridization with biotinylated enterotoxin probes. J Clin Microbiol 25: t 962-1965 19. Leong J, Vinal AC, Dallas WS (1985) Nucleotide sequence comparison between heat-labile toxin B-subunit cistrons from Escherichia coli of human and porcine origin. Infect Immun 48:73-77 20. Maniatis T, Fritsch EF, Sambrook J (1982) Molecular cloning: a laboratory manual. Cold Spring Harbor, New York: Cold Spring Harbor Laboratory 21. Martfnez-Cadena MG, Guzman-Verduzco LM, Stieglitz H, Kupersztoch-Portnoy YM (1981) Catabotite repression of Escherichia coli heat-stable enterotoxin activity. J Bacteriol 145:722-728 22. Miller JM (1972) Experiments in molecular genetics. Cold Spring Harbor, New York: Cold Spring Harbor laboratory 23. Neill RJ, Ivins BE, Holmes RK (1983) Synthesis and secretion of the plasmid-coded heat-labile enterotoxin of Escher# chia coli in Vibrio cholerae. Science 221:289-91 24. Normark S, Bergstr6m S, Edlund T, GrundstrOm T, Jaurin B, Lindberg FP, Olsson O (1983) Overlapping genes. Annu Rev Genet 17:499-525

CURRENT MICROBIOLOGY Vol. 20 (t990)

25. Sancar A, Rupp WP (1979) Physical map of recA gene. Proc Natl Acad Sci USA 76:3144-3148 26. Sanchez J, Bennett PM, Richmond MH (t982) Expression of elt-B, the gene encoding the B subunit of the heat-labile enterotoxin of Escherichia coli, when cloned in pACYC184. FEMS Microbiol Lett 14:1-5 27. Schneider K, Beck CF (1987) New expression vectors for identifying and testing signal structures for initiation and termination of transcription. Methods Enzymol 153:452461 28. Spencer ME, Guest JR (1973) Isolation and properties of fumarate reductase mutants of Escherichia coli. J Bacteriol 114:563-570 29. Spicer EK, Noble JA (1982) Escherichia coli heat-labile enterotoxin. Nucleotide sequence of the A subunit gene. J Biol Chem 257:5716-5721 30. Spicer EK, Kavanaugh WM, Dallas WS, Falkow S, Konigsberg WH, Schaffer DE (1981) Sequence homologies between A subunits of Escherichia coli and Vibrio cholerae enterotoxins. Proc Natl Acad Sci USA 78:50-54 31. Staden R (1982) An interactive graphics program for comparing and aligning nucleic acid and amino acid sequence. Nucleic Acids Res 10:2951-2961 32. van den Elzen PJM, Maat J, Waiters HHB, Veltkamp E, Nijkamp JJ (1982) The nucteotide sequence of the bacteriocin promoters of plasmids Clo DF13 and Col El: role of lexA repressor and cAMP in the regulation of promoter activity. Nucleic Acids Res t0:1913-1928 33. Yamamoto T, Yokota T (1980) Cloning of deoxyribonucleic acid regions encoding a heat-labile and heat-stable enterotoxin originating from an enterotoxigenic Escherichia coti strain of human origin. J Bacteriol 143:652-660 34. Yamamoto T, Yokota T (1981) Escherichia coli heat-labile enterotoxin genes are flanked by repeated deoxyribonucleic acid sequences. J Bacteriol 145:850-860 35. Yamamoto T, Yokota T (1983) Sequence of heat-labile enterotoxin of Escherichia coli pathogenic for humans. J Bacteriol 155:728-733 36. Yamamoto T, Yokota T, Kaji A (1981) Molecular organization of heat-labile enterotoxin genes originating in Escherichia coli of human origin and construction of heat-labile toxoid-producing strains. J Bacteriol 148:983987 37. Yamamoto T, Tamura T, Yokota T, Takano T (1982) Overlapping genes in the heat-labile enterotoxin operon originating from Escherichia coli human strain. Mol Gen Genet 188:356-359 38. Yamamoto T, Tamura T, Yokota T (1984) Primary structure of heat-labile enterotoxin produced by Escherichia coli pathogenic for humans. J Biol Chem 259:5037-5044 39. Yamamoto T, Sujama A, Mori N, Yokota T, Wada A (1985) Gene expression in the polycistronic operons of Escherichia coli heat-labile toxin and cholera toxin: a new model of translational control. FEBS Lett 183:377-380