- Email: [email protected]
New insights into the regulation of the pac gene from Escherichia coli W ATCC 11105
FEMS Microbiology Letters 177 (1999) 7^14
New insights into the regulation of the pac gene from Escherichia coli W ATCC 11105 Ana Roa, Jose¨ Luis Garc|¨a * Department of Molecular Microbiology, Centro de Investigaciones Biolo¨gicas, Consejo Superior de Investigaciones Cient|¨¢cas, Vela¨zquez 144, 28006 Madrid, Spain Received 25 March 1999; received in revised form 28 May 1999 ; accepted 2 June 1999
Abstract The regulation of the pac gene encoding the penicillin G acylase of Escherichia coli W ATCC 11015 has been investigated by a molecular approach using lacZ as a reporter gene. This analysis revealed that a region of 170 bp located upstream of the pac structural gene contains the regulatory sequences that control its expression. The cAMP receptor protein is involved not only in the catabolite repression of penicillin G acylase production caused by glucose but also in the induction of pac gene expression by phenylacetic acid. Primer extension analyses have demonstrated that the transcription of the pac gene can be initiated from at least three different promoters. Although all these promoters are functional, their relative activity depends on the transcribed gene, the P1 and P3 promoters being more active in the presence of the pac gene, whereas the P2 promoter was stronger when the upstream region of the pac gene was fused to the lacZ reporter. A deletion of the region surrounding the 310 box of the P3 promoter produced a constitutive expression of the fused gene indicating that this sequence is required for phenylacetic acid induction and suggesting that the expression of the pac gene is regulated by a repression mechanism. This work reveals that the regulation of the pac gene is more complex than previously envisioned and provides new clues to investigate further this interesting regulatory system. ß 1999 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. Keywords : Phenylacetic acid; Penicillin G acylase; Pac regulation
1. Introduction Penicillin G acylase (PGA, penicillin G amidohydrolase, EC 3.5.1.11) catalyzes the hydrolysis of benzyl penicillin and it is one of the most important industrial enzymes since it is used for the commercial
* Corresponding author. Tel.: +34 (91) 5611800. Fax: +34 (91) 5627518; E-mail: [email protected]
production of semisynthetic penicillins [1]. Although di¡erent PGAs have been described in many microorganisms, the enzyme from Escherichia coli W ATCC 11105 has been the most extensively studied so far [1]. The expression of the pac gene is subjected to several regulatory controls, including temperature, oxygen, catabolite repression, and induction by phenylacetic acid (PAA) [1,2]. In addition, a complex maturation process is required to render the active form of PGA [2]. The enzyme is produced as an inactive cytoplasmic precursor of 93 kDa which be-
0378-1097 / 99 / $20.00 ß 1999 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 1 0 9 7 ( 9 9 ) 0 0 2 8 1 - 5
FEMSLE 8874 15-7-99
8
A. Roa, J.L. Garc|¨a / FEMS Microbiology Letters 177 (1999) 7^14
comes catalytically active after its translocation to the periplasmic space followed by an autoproteolytic disruption into two subunits (K and L) that removes an internal spacer polypeptide [2]. Since the pac gene has been cloned and sequenced [3,4], very few studies have been done on the characterization and regulation of the pac promoter [5^7]. Some results reported in the available literature that are a matter of discrepancy (see below) and the current pending questions about PAA induction and catabolite repression prompted us to reevaluate the regulatory elements involved in the expression of this important gene. In the present work, we have analyzed the sequences involved in the expression of the pac gene of E. coli W ATCC 11105 using a lacZ fusion approach that has provided a new scenario to understand the complex regulation of this gene.
2. Materials and methods 2.1. Bacteria and plasmids The strains and plasmids used in this work are shown in Table 1. E. coli cells were grown with shaking in Luria broth (LB) [13] or M63 minimal medium [14]. Media were supplemented with L-leucine (100 Wg ml31 ), L-proline (100 Wg ml31 ), ampicillin (100 Wg ml31 ), tetracycline (12.5 Wg ml31 ), glucose or glycerol when required.
2.2. DNA and RNA manipulations Plasmid DNA was prepared by the rapid alkaline method [13]. Transformation of E. coli cells was carried out using the RbCl method [13]. Polymerase chain reaction (PCR) was performed using the Gene ATAQ Controller (Pharmacia LKB). DNA fragments were puri¢ed using L-agarase (New England Biolabs, Beverly, MA, USA). DNA sequencing was carried using the Pharmacia T7 sequencing kit. RNA extraction and primer extension analysis were performed as previously described [12], using the oligonucleotides PRS551 (5P-GCCAGGGTTTTCCCAGTC-3P) and PAC (5P-GCTCCAATAATACATCAGGGAAG-3P) that hybridized within the 5P coding sequences of the lacZ and pac genes, respectively. PCR ampli¢cation was performed using the plasmid pPGA1 as template and the oligonucleotides PEC1 (5P-CCCAAGCTTTTCATTGTATCCTTCTGG-3P; the BamHI site is underlined) and PEC2 (5P-CGCGGATCCAGCGGTGAATAAAGCG-3P ; the HindIII site is underlined) as primers. 2.3. Enzyme assays L-Galactosidase (L-Gal) activity was determined according to the method of Miller [14] using 2-nitrophenyl-L-galactopyranoside as substrate. PGA activity was assayed using 6-nitro-3-phenylacetamido benzoic acid (NIPAB) as substrate [12]. Usually
Table 1 Bacterial strains and plasmids used in this study Strain or plasmid E. coli MC4100 SBS688 HB101 W ATCC 11105 Plasmids pSKS107 pRS550 pPGA1 pAJ19 pSKSK pSKSP pSKSE pRS55E
Relevant genotype or properties
Source
v(lacIPOZYA)U169, thi MC4100 vcrp39 proA2, leuB, thi Vitamin B12 auxotroph
M. Casadaban J. Pe¨rez CIB collection ATCC
amp, promoter-less lacZYA amp, kan, promoter-less lacZYA tet, Ppac -pac from E. coli W amp, vPpac -pac from E. coli W amp, Ppac (K. citrophila)-lacZ amp, Ppac -pac from E. coli W amp, Ppac (E. coli)-lacZ amp, vPpac (E. coli)-lacZ
[8] [9] [10] [11] [12]. This study This study This study
FEMSLE 8874 15-7-99
A. Roa, J.L. Garc|¨a / FEMS Microbiology Letters 177 (1999) 7^14 Table 2 PGA activity of E. coli HB101 cells harboring di¡erent pac-containing plasmids Plasmid
pPGA1 pAJ19 pSKSP
PGA activitya (nmol min31 ml31 of culture) 3PAA
+PAA
10 þ 5 220 þ 15 20 þ 5
240 þ 10 270 þ 15 260 þ 20
9
studying the pac promoter by using a simpli¢ed experimental approach based on the construction of lacZ fusions. 3.2. Construction of L-Gal fusions
a Cells were grown for 24 h at 28³C in LB medium supplemented with tetracycline (12.5 Wg ml31 ) (pPGA1) or ampicillin (100 Wg ml31 ) (pSKSP and pAJ19) in the absence or presence of 0.1% PAA. PGA activities were determined using NIPAB as substrate.
three independent determinations were made for each activity measured.
3. Results and discussion 3.1. Expression of the pac gene in E. coli K12 strains To con¢rm that the expression of the cloned pac gene was still inducible by PAA as occurs in the parental E. coli W ATCC 11105 [10], we determined the PGA activity produced in an E. coli K12 recombinant strain harboring the pac gene. Table 2 shows that the production of PGA in E. coli HB101 (pPGA1) (Fig. 1B) was inducible by PAA, whereas E. coli HB101 (pAJ19) produced PGA constitutively. Sequencing of plasmid pAJ19 [11] revealed that the cloned fragment only contains the ¢rst 53 bp of the 5P non-coding region of the pac gene that exclusively harbors the 310 box of the previously proposed pac promoter [6] (Fig. 1B, hereafter named P3 promoter). Therefore, the expression of the pac gene in the pAJ19 recombinants is more likely directed by a promoter present in the plasmid. However, we should also consider that Oh et al. [4] have ascribed the pac promoter to the sequences TTGCTA (335 box) and TATACA (310 box) located downstream of the HindIII site (Fig. 1A). Interestingly, the PAA induction of pac expression can be restored (Table 2) by reconstructing in plasmid pSKSP the complete native upstream regulatory region using a 170-bp PCR ampli¢ed fragment (Fig. 1B). Therefore, this result suggested that the 170-bp fragment contained the regulatory signals required to control pac expression, opening the possibility of
Although a preliminary analysis of pac expression by fusing its non-coding upstream region to the lacZ gene had been performed [5], this study presented some technical drawbacks hampering the interpretation of the results. In this sense, the promoter analysis was conducted in E. coli HB101, a lacZ strain that displays a L-Gal background activity. In addition, the lacZ fusion was constructed using the available restriction sites of the pac sequence which caused that the cloned region was too large and rendered a chimeric L-Gal fused to the PGA signal peptide that might cause toxic e¡ects [15]. To avoid all these drawbacks, we have constructed plasmid pSKSE (Fig. 1B), which contains the ¢rst 170 bp of the 5P non-coding region of the pac gene fused to the ATG codon of the lacZ gene. When this plasmid was transformed into E. coli MC4100 (vlacZ), we observed that the production of L-Gal was inducible by PAA (Table 3) supporting our suggestion that the PCR-ampli¢ed sequence contains the signals required for PAA induction. In addition, this result suggests that the presence of the complete pac gene is not essential for PAA induction, refuting the hypothesis that the regulatory gene responsible for such induction is located within the pac structural gene [7]. Nevertheless, we cannot discard that such regulatory element could play another role (see below). The e¡ect of PAA seems to be mediated by a regulatory mechanism that is also present in a heterologous host like E. coli HB101, lacking the pac gene. Since we have recently shown that the genomes of E. coli W ATCC 11105 and E. coli K12 encode the pathway for PAA degradation [16], we cannot rule out the possibility that the gene(s) involved in the regulation of this pathway could also be involved in pac gene expression. On the other hand, we have also observed that the expression of the lacZ fusion in E. coli MC4100 (pSKSE) decreased when glucose was added to the medium (Table 3). Interestingly, the inhibitory e¡ect of glucose was lower on the uninduced cells (basal level) than on the cells induced by PAA. Neverthe-
FEMSLE 8874 15-7-99
10
A. Roa, J.L. Garc|¨a / FEMS Microbiology Letters 177 (1999) 7^14
Fig. 1. Plasmid constructions and nucleotide sequence of the 5P non-coding region of the E. coli W ATCC 11105 pac gene. A: Sequence of the 5P non-coding region of pac gene from E. coli W ATCC 11105. CRP1 and CRP2 indicate the putative CRP binding sites. The 335 and 310 boxes of the P1, P2 and P3 promoters are underlined. ATG in bold face represents the translation initiation codon of the pac gene. RBS shows the position of the ribosome binding site. +1 indicates the transcription start sites of the promoters. B : Schematic representation of the plasmid constructions. White boxes represent the putative CRP binding sites. The 335 and 310 boxes of P1, P2 and P3 promoters are represented by dashed boxes. The lacZ gene is shown by a dotted box. The pac gene is represented by a striped box. The oligonucleotides used for the PCR ampli¢cation are indicated as primers PEC1 and PEC2. Abbreviations: amp, ampicillin resistance gene; tet, tetracycline resistance gene; kan, kanamycin resistance gene ; B, BamHI; H, HindIII; S, SacI; Sa, SalI, Sm, SmaI. H* indicates a partial digestion with HindIII. H(Kl.) indicates a HindIII digestion followed by a treatment with the Klenow fragment of DNA polymerase.
FEMSLE 8874 15-7-99
A. Roa, J.L. Garc|¨a / FEMS Microbiology Letters 177 (1999) 7^14
less, PAA was still able to induce the production of L-Gal in the presence of glucose, in both LB and M63 minimal medium, but the induction e¡ect was signi¢cantly reduced (Table 3). The L-Gal production can be repressed up to 95% in LB medium containing 10 mM glucose, whereas the presence of 20 mM glycerol only produced a limited repression (25%) (data not shown). These results are in agreement with previous ¢ndings [5] and strongly suggest that the pac promoter is regulated by catabolite repression that can be most probably ascribed to the presence of two putative CRP (cAMP receptor protein) binding sites, CRP1 and CRP2, which are located within the 170-bp 5P non-coding region of the pac gene cloned in pSKSE (Fig. 1A). To determine the in£uence of the CRP protein on the regulation of the pac promoter the expression of the lacZ fusion in pSKSE was analyzed in the E. coli CRP3 mutant SBS668. The production of L-Gal in E. coli SBS688 (pSKSE) cells induced with PAA was reduced to the basal levels independently of the absence or the presence of 5 mM glucose in the culture medium (Table 3). As expected, the addition of 5 mM glucose to the culture medium did not generate an inhibitory response on the uninduced cells (Table 3). The levels of L-Gal in the uninduced cells of both the wild-type and the CRP3 strains were quite similar and they were not drastically a¡ected by the addition of glucose suggesting that the basal expression of lacZ fusion is mainly generated from a CRPindependent transcription activity. Surprisingly, these results not only con¢rmed that the catabolite repression of pac promoter was a CRP-dependent process but, more important, they revealed that the CRP protein was directly involved in the PAA induction.
11
3.3. Analysis of pac promoter In spite of the high similarity of the pac genes from E. coli W ATCC 11105 and Kluyvera citrophila, the pac promoter of K. citrophila was located overlapping the sequence of the CRP2 binding site [12], this is, far upstream from the promoter (P3 promoter) proposed for the pac gene of E. coli W ATCC 11105 [6] (Fig. 1A). This ¢nding pointed to the possibility that other promoters might be involved in the expression of the E. coli pac gene. To investigate this possibility, we constructed plasmid pRS55E that contains a lacZ gene fused to a fragment lacking the 310 box of the previously proposed pac promoter [6] or P3 promoter (Fig. 1B). Interestingly, E. coli MC4100 (pRS55E) cells cultured in LB medium showed a high L-Gal production (3000 U after 20 h of culture) that was not induced by PAA (data not shown), but that was repressed by 5 mM glucose (1000 U after 20 h of culture) (data not shown). These results strongly supported the existence of a pac promoter upstream of the HindIII site, and also indicated that the deleted region was involved in the PAA induction. In this sense, the high constitutive activity found in these cells suggests that the pac expression might be regulated by a repressor that recognizes a binding site within the deleted region. A primer extension analysis carried out in E. coli MC4100 (vlacZ) transformed with plasmid pSKSE, showed the presence of three bands of di¡erent intensities (Fig. 2). The same pattern of bands was observed when the RNA was extracted after 20 h (Fig. 2, lane E1) and 3 h (Fig. 2, lane E3) of incubation, suggesting that the relative amounts and the length of the transcripts were not a¡ected by the
Table 3 L-Gal activity of E. coli MC4100 and E. coli SBS688 pSKSE transformants Strain
L-Gal activity (units)a LB
MC4100 (pSKSE) SBS688 (pSKSE)
LB+5 mM glucose
M63+5 mM glucose
3PAA
+PAA
3PAA
+PAA
3PAA
+PAA
330 þ 60 410 þ 50
1500 þ 130 400 þ 50
220 þ 60 330 þ 60
430 þ 50 400 þ 60
130 þ 40 230 þ 50
720 þ 60 230 þ 50
a Activities are expressed in Miller units. Cells were grown for 24 hours at 28³C in LB or M63 minimal medium supplemented with ampicillin (100 Wg/ml) in the absence or presence of 0.1% PAA and in the absence or presence of 5 mM glucose.
FEMSLE 8874 15-7-99
12
A. Roa, J.L. Garc|¨a / FEMS Microbiology Letters 177 (1999) 7^14
phila (Fig. 3, lane E2) [12]. Upstream of the main start site we found the sequences TTAGTT and CATAAT that might represent the 335 and 310 boxes of the P2 promoter (Fig. 1A). These sequences overlap the putative CRP2 binding site and do not perfectly match with the consensus sequence of the E. coli c70 promoters. The 335 box of P2 promoter is preceded by a putative A+T rich enhancer se-
Fig. 2. Primer extension analysis of the transcription start sites of lacZ fusions. Total mRNA was isolated from cells cultured at 28³C in LB medium containing ampicillin (100 Wg ml31 ) and 0.1% PAA. Lane E1, mRNA isolated from E. coli MC4100 (pSKSE) after 20 h of culture; lane E2, mRNA isolated from E. coli MC4100 (pSKSK) after 3 h of culture; lane E3, mRNA isolated from E. coli MC4100 (pSKSE) after 3 h of culture. The size of the extended products was determined by comparison with a DNA sequencing ladder of the pac promoter region using the plasmid pSKSE as template. The primer extensions and the sequencing reactions were performed using the primer PRS551.
phase of growth. The major band corresponded to a transcriptional start site located at 377 bp (P2 promoter) from the ATG translation initiation codon whereas the two minor bands corresponded to putative start sites located at a distance of 381 bp (P1 promoter) and 330 bp (P3 promoter), respectively (Figs. 1 and 3). Interestingly, the major band corresponded to that previously found in the transformants carrying the plasmid pSKSK containing the lacZ gene fused to the pac promoter of K. citro-
Fig. 3. Primer extension analysis of the pac transcription start sites. Total mRNA was isolated from cells cultured at 28³C in LB medium containing tetracycline (12.5 Wg ml31 ) and 0.1% PAA. Lane E1, mRNA isolated from E. coli W ATCC 11105 (pPGA1) cells after 6 h of incubation; lane E2, mRNA isolated from E. coli HB101 (pPGA1) cells after 6 h of incubation; lane E3, mRNA isolated from E. coli HB101 (pPGA1) cells after 15 h of incubation. The size of the extended products was determined by comparison with a DNA sequencing ladder of the pac promoter region using the plasmid pPGA1 as template. The primer extensions and the sequencing reactions were performed using the primer PAC.
FEMSLE 8874 15-7-99
A. Roa, J.L. Garc|¨a / FEMS Microbiology Letters 177 (1999) 7^14
quence [17]. The start site located at 381 bp could be ascribed to the 335 (TATCAA) and 310 (CACGAT) boxes of the putative P1 promoter. The existence of a P1 and P2 promoters explains the expression of lacZ observed with plasmid pRS55E. Finally, the putative start site located at 330 bp from the ATG codon corresponds to that determined by Valle et al. [6] and, thus, the P3 promoter is built by the 335 (TAGATA) and 310 (TAGTAT) boxes that surround the HindIII site (Fig. 1A). It was surprising that the major transcription start site did not correspond to the start site previously located at 3-30 bp by Valle et al. [6]. To rule out the possibility that the band of the P3 promoter might be generated by an unspeci¢c arrest of the reverse transcriptase or by a degradation/processing of pac mRNA, we reproduced the experiment carried out by Valle et al. [6]. Therefore, a primer extension analysis was performed using as hosts both the E. coli strains W ATCC 11105 and HB101 transformed with plasmid pPGA1. Surprisingly, in this case, the main transcription start site was located at 330 bp from the ATG codon (Fig. 3). In contrast, the bands corresponding to the P1 and P2 promoters were signi¢cantly reduced compared to that of P3 promoter. The relative intensities and the positions of the bands were similar when the RNA was extracted after 6 or 15 h of culture (Fig. 3, lanes E2 and E3) indicating that length of the transcripts was independent of the incubation time. The same pattern was observed in both W and HB101 strains suggesting that the dramatic change in the intensities of the bands should not be ascribed to the presence of speci¢c regulatory elements in the parental strain. The above ¢ndings strongly suggest that the pac gene a¡ects its own expression, modifying the transcription rate of the P1, P2 and P3 promoters. In this sense, it has been suggested that the pac gene might control its own expression [7], and although this argument cannot be used to explain the PAA induction observed in plasmid pSKSE, we should consider that the PacR protein could modulate the a¤nity of RNA polymerase for the alternative binding sites. Moreover, the upstream region of pac gene presents several regions that could be involved in DNA bending facilitating the interactions of di¡erent factors (CRP, PAA-dependent repressor, or PcaR) that might function as transcriptional switches between
13
tandem promoters [17]. Finally, it has been demonstrated that some genes such as the luxAB genes can activate or repress transcription from a subset of promoters due to an intrinsically curved DNA segment in the 5P coding sequence of the luxA gene [18]. Summarizing, the results presented in this work provide novel experimental evidence that the regulation of the pac gene is more complex than previously envisioned. The existence of several alternative promoters raises new questions about the nature of the regulatory elements and mechanisms that control the expression of this gene that should be further investigated.
Acknowledgements We thank M.A. Prieto, E. D|¨az, and R. Lo¨pez for critical reading of the manuscript. The artwork of A. Hurtado and the technical assistance of E. Cano and M. Carrasco are gratefully acknowledged. This work was supported by the Comisio¨n Interministerial de Ciencia y Tecnolog|¨a (Grant AMB94-1038-CO2-02 and Grant AMB97-603-CO2-02).
References [1] Sudhakaran, V.K., Deshpande, B.S., Ambedkar, S.S. and Shewale, J.G. (1992) Molecular aspects of penicillin and cephalosporin acylases. Process. Biochem. 27, 131^143. [2] Virden, R. (1990) Structure, procesing and catalytic action of penicillin acylase. Biotech. Gen. Eng. Rev. 8, 189^218. [3] Schumacher, G., Sizmann, D., Haug, H.Buckel,P. and Bo«ck, A. (1986) Penicillin acylase from E. coli: unique gene-protein relation. Nucleic Acids Res. 14, 5713^5727. [4] Oh, S.-J., Kim, Y.-C., Park, Y.-W., Min, S.-Y., Kim, I.-S. and Kang, H.-S. (1987) Complete nucleotide sequence of the penicillin G acylase gene and the £anking regions, and the expression in Escherichia coli. Gene 56, 87^97. [5] Merino, E., Balba¨s, P., Recillas, F., Becerril, B., Valle, F. and Bol|¨var, F. (1992) Carbon regulation and role in nature of the Escherichia coli penicillin acylase (pac) gene. Mol. Microbiol. 6, 2175^2182. [6] Valle, F., Gosset, G., Tenorio, B., Oliver, G. and Bol|¨var, F. (1986) Characterization of the regulatory region of the Escherichia coli penicillin acylase structural gene. Gene 50, 119^122. [7] Jiang, L., Wang, Y. and Yang, S. (1997) Location of regulatory gene in penicillin G acylase operon (pacR) of E. coli D816. Gen. Anal. 14, 51^54. [8] Casadaban, M.J., Mart|¨nez-Arias, A., Shaphira, S.K. and Chou, J. (1983) L-Galactosidase gene fusions for analyzing
FEMSLE 8874 15-7-99
14
[9]
[10]
[11]
[12]
[13]
A. Roa, J.L. Garc|¨a / FEMS Microbiology Letters 177 (1999) 7^14 gene expression in Escherichia coli and yeast. Methods Enzymol. 100, 293^308. Simons, R.W., Houman, F. and Kleckner, N. (1987) Improved single and multicopy lac-based cloning vectors for protein and operon fusions. Gene 53, 85^96. Garc|¨a, J.L. and Buesa, J.M. (1986) An improved method to clone penicillin acylase genes: Cloning and expression in Escherichia coli of penicillin G acylase from Kluyvera citrophila. J. Biotechnol. 3, 187^195. Prieto, M.A., D|¨az, E. and Garc|¨a, J.L. (1996) Molecular characterization of the 4-hydroxyphenylacetate catabolic pathway of Escherichia coli W: Engineering a mobile aromatic degradative cluster. J. Bacteriol. 178, 111^120. Roa, A. and Garc|¨a, J.L. (1997) Identi¢cation of the pac promoter from Kluyvera citrophila. FEMS Microbiol. Lett. 151, 9^16. Sambrook, J,, Fritsch, E.F. and Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
[14] Miller, J.H. (1972) Experiments in Molecular Genetics. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. [15] Snyder, W.B. and Silhavy, T.J. (1995) Beta-galactosidase is inactivated by intermolecular disul¢de bonds and is toxic when secreted to the periplasm of Escherichia coli. J. Bacteriol. 177, 953^963. [16] Ferra¨ndez, A., Min¬ambres, B., Garc|¨a, B., Olivera, E.R., Luengo, J.M., Garc|¨a, J.L. and D|¨az, E. (1998) Catabolism of phenylacetic acid in Escherichia coli. Characterization of a new aerobic hybrid pathway. J. Biol. Chem. 273, 25974^ 25986. [17] Pe¨rez-Mart|¨n, J., Rojo, F. and de Lorenzo, V. (1994) Promoters responsive to DNA bending: a common theme in prokaryotic gene expression. Microbiol. Rev. 58, 268^290. [18] Forsberg, A.J., Pavitt, G.D. and Higgins, C.F. (1994) Use of transcriptional fusions to monitor gene expression: a cautionary tale. J. Bacteriol. 176, 2128^2132.
FEMSLE 8874 15-7-99
New insights into the regulation of the pac gene from Escherichia coli W ATCC 11105 Ana Roa, Jose¨ Luis Garc|¨a * Department of Molecular Microbiology, Centro de Investigaciones Biolo¨gicas, Consejo Superior de Investigaciones Cient|¨¢cas, Vela¨zquez 144, 28006 Madrid, Spain Received 25 March 1999; received in revised form 28 May 1999 ; accepted 2 June 1999
Abstract The regulation of the pac gene encoding the penicillin G acylase of Escherichia coli W ATCC 11015 has been investigated by a molecular approach using lacZ as a reporter gene. This analysis revealed that a region of 170 bp located upstream of the pac structural gene contains the regulatory sequences that control its expression. The cAMP receptor protein is involved not only in the catabolite repression of penicillin G acylase production caused by glucose but also in the induction of pac gene expression by phenylacetic acid. Primer extension analyses have demonstrated that the transcription of the pac gene can be initiated from at least three different promoters. Although all these promoters are functional, their relative activity depends on the transcribed gene, the P1 and P3 promoters being more active in the presence of the pac gene, whereas the P2 promoter was stronger when the upstream region of the pac gene was fused to the lacZ reporter. A deletion of the region surrounding the 310 box of the P3 promoter produced a constitutive expression of the fused gene indicating that this sequence is required for phenylacetic acid induction and suggesting that the expression of the pac gene is regulated by a repression mechanism. This work reveals that the regulation of the pac gene is more complex than previously envisioned and provides new clues to investigate further this interesting regulatory system. ß 1999 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. Keywords : Phenylacetic acid; Penicillin G acylase; Pac regulation
1. Introduction Penicillin G acylase (PGA, penicillin G amidohydrolase, EC 3.5.1.11) catalyzes the hydrolysis of benzyl penicillin and it is one of the most important industrial enzymes since it is used for the commercial
* Corresponding author. Tel.: +34 (91) 5611800. Fax: +34 (91) 5627518; E-mail: [email protected]
production of semisynthetic penicillins [1]. Although di¡erent PGAs have been described in many microorganisms, the enzyme from Escherichia coli W ATCC 11105 has been the most extensively studied so far [1]. The expression of the pac gene is subjected to several regulatory controls, including temperature, oxygen, catabolite repression, and induction by phenylacetic acid (PAA) [1,2]. In addition, a complex maturation process is required to render the active form of PGA [2]. The enzyme is produced as an inactive cytoplasmic precursor of 93 kDa which be-
0378-1097 / 99 / $20.00 ß 1999 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 1 0 9 7 ( 9 9 ) 0 0 2 8 1 - 5
FEMSLE 8874 15-7-99
8
A. Roa, J.L. Garc|¨a / FEMS Microbiology Letters 177 (1999) 7^14
comes catalytically active after its translocation to the periplasmic space followed by an autoproteolytic disruption into two subunits (K and L) that removes an internal spacer polypeptide [2]. Since the pac gene has been cloned and sequenced [3,4], very few studies have been done on the characterization and regulation of the pac promoter [5^7]. Some results reported in the available literature that are a matter of discrepancy (see below) and the current pending questions about PAA induction and catabolite repression prompted us to reevaluate the regulatory elements involved in the expression of this important gene. In the present work, we have analyzed the sequences involved in the expression of the pac gene of E. coli W ATCC 11105 using a lacZ fusion approach that has provided a new scenario to understand the complex regulation of this gene.
2. Materials and methods 2.1. Bacteria and plasmids The strains and plasmids used in this work are shown in Table 1. E. coli cells were grown with shaking in Luria broth (LB) [13] or M63 minimal medium [14]. Media were supplemented with L-leucine (100 Wg ml31 ), L-proline (100 Wg ml31 ), ampicillin (100 Wg ml31 ), tetracycline (12.5 Wg ml31 ), glucose or glycerol when required.
2.2. DNA and RNA manipulations Plasmid DNA was prepared by the rapid alkaline method [13]. Transformation of E. coli cells was carried out using the RbCl method [13]. Polymerase chain reaction (PCR) was performed using the Gene ATAQ Controller (Pharmacia LKB). DNA fragments were puri¢ed using L-agarase (New England Biolabs, Beverly, MA, USA). DNA sequencing was carried using the Pharmacia T7 sequencing kit. RNA extraction and primer extension analysis were performed as previously described [12], using the oligonucleotides PRS551 (5P-GCCAGGGTTTTCCCAGTC-3P) and PAC (5P-GCTCCAATAATACATCAGGGAAG-3P) that hybridized within the 5P coding sequences of the lacZ and pac genes, respectively. PCR ampli¢cation was performed using the plasmid pPGA1 as template and the oligonucleotides PEC1 (5P-CCCAAGCTTTTCATTGTATCCTTCTGG-3P; the BamHI site is underlined) and PEC2 (5P-CGCGGATCCAGCGGTGAATAAAGCG-3P ; the HindIII site is underlined) as primers. 2.3. Enzyme assays L-Galactosidase (L-Gal) activity was determined according to the method of Miller [14] using 2-nitrophenyl-L-galactopyranoside as substrate. PGA activity was assayed using 6-nitro-3-phenylacetamido benzoic acid (NIPAB) as substrate [12]. Usually
Table 1 Bacterial strains and plasmids used in this study Strain or plasmid E. coli MC4100 SBS688 HB101 W ATCC 11105 Plasmids pSKS107 pRS550 pPGA1 pAJ19 pSKSK pSKSP pSKSE pRS55E
Relevant genotype or properties
Source
v(lacIPOZYA)U169, thi MC4100 vcrp39 proA2, leuB, thi Vitamin B12 auxotroph
M. Casadaban J. Pe¨rez CIB collection ATCC
amp, promoter-less lacZYA amp, kan, promoter-less lacZYA tet, Ppac -pac from E. coli W amp, vPpac -pac from E. coli W amp, Ppac (K. citrophila)-lacZ amp, Ppac -pac from E. coli W amp, Ppac (E. coli)-lacZ amp, vPpac (E. coli)-lacZ
[8] [9] [10] [11] [12]. This study This study This study
FEMSLE 8874 15-7-99
A. Roa, J.L. Garc|¨a / FEMS Microbiology Letters 177 (1999) 7^14 Table 2 PGA activity of E. coli HB101 cells harboring di¡erent pac-containing plasmids Plasmid
pPGA1 pAJ19 pSKSP
PGA activitya (nmol min31 ml31 of culture) 3PAA
+PAA
10 þ 5 220 þ 15 20 þ 5
240 þ 10 270 þ 15 260 þ 20
9
studying the pac promoter by using a simpli¢ed experimental approach based on the construction of lacZ fusions. 3.2. Construction of L-Gal fusions
a Cells were grown for 24 h at 28³C in LB medium supplemented with tetracycline (12.5 Wg ml31 ) (pPGA1) or ampicillin (100 Wg ml31 ) (pSKSP and pAJ19) in the absence or presence of 0.1% PAA. PGA activities were determined using NIPAB as substrate.
three independent determinations were made for each activity measured.
3. Results and discussion 3.1. Expression of the pac gene in E. coli K12 strains To con¢rm that the expression of the cloned pac gene was still inducible by PAA as occurs in the parental E. coli W ATCC 11105 [10], we determined the PGA activity produced in an E. coli K12 recombinant strain harboring the pac gene. Table 2 shows that the production of PGA in E. coli HB101 (pPGA1) (Fig. 1B) was inducible by PAA, whereas E. coli HB101 (pAJ19) produced PGA constitutively. Sequencing of plasmid pAJ19 [11] revealed that the cloned fragment only contains the ¢rst 53 bp of the 5P non-coding region of the pac gene that exclusively harbors the 310 box of the previously proposed pac promoter [6] (Fig. 1B, hereafter named P3 promoter). Therefore, the expression of the pac gene in the pAJ19 recombinants is more likely directed by a promoter present in the plasmid. However, we should also consider that Oh et al. [4] have ascribed the pac promoter to the sequences TTGCTA (335 box) and TATACA (310 box) located downstream of the HindIII site (Fig. 1A). Interestingly, the PAA induction of pac expression can be restored (Table 2) by reconstructing in plasmid pSKSP the complete native upstream regulatory region using a 170-bp PCR ampli¢ed fragment (Fig. 1B). Therefore, this result suggested that the 170-bp fragment contained the regulatory signals required to control pac expression, opening the possibility of
Although a preliminary analysis of pac expression by fusing its non-coding upstream region to the lacZ gene had been performed [5], this study presented some technical drawbacks hampering the interpretation of the results. In this sense, the promoter analysis was conducted in E. coli HB101, a lacZ strain that displays a L-Gal background activity. In addition, the lacZ fusion was constructed using the available restriction sites of the pac sequence which caused that the cloned region was too large and rendered a chimeric L-Gal fused to the PGA signal peptide that might cause toxic e¡ects [15]. To avoid all these drawbacks, we have constructed plasmid pSKSE (Fig. 1B), which contains the ¢rst 170 bp of the 5P non-coding region of the pac gene fused to the ATG codon of the lacZ gene. When this plasmid was transformed into E. coli MC4100 (vlacZ), we observed that the production of L-Gal was inducible by PAA (Table 3) supporting our suggestion that the PCR-ampli¢ed sequence contains the signals required for PAA induction. In addition, this result suggests that the presence of the complete pac gene is not essential for PAA induction, refuting the hypothesis that the regulatory gene responsible for such induction is located within the pac structural gene [7]. Nevertheless, we cannot discard that such regulatory element could play another role (see below). The e¡ect of PAA seems to be mediated by a regulatory mechanism that is also present in a heterologous host like E. coli HB101, lacking the pac gene. Since we have recently shown that the genomes of E. coli W ATCC 11105 and E. coli K12 encode the pathway for PAA degradation [16], we cannot rule out the possibility that the gene(s) involved in the regulation of this pathway could also be involved in pac gene expression. On the other hand, we have also observed that the expression of the lacZ fusion in E. coli MC4100 (pSKSE) decreased when glucose was added to the medium (Table 3). Interestingly, the inhibitory e¡ect of glucose was lower on the uninduced cells (basal level) than on the cells induced by PAA. Neverthe-
FEMSLE 8874 15-7-99
10
A. Roa, J.L. Garc|¨a / FEMS Microbiology Letters 177 (1999) 7^14
Fig. 1. Plasmid constructions and nucleotide sequence of the 5P non-coding region of the E. coli W ATCC 11105 pac gene. A: Sequence of the 5P non-coding region of pac gene from E. coli W ATCC 11105. CRP1 and CRP2 indicate the putative CRP binding sites. The 335 and 310 boxes of the P1, P2 and P3 promoters are underlined. ATG in bold face represents the translation initiation codon of the pac gene. RBS shows the position of the ribosome binding site. +1 indicates the transcription start sites of the promoters. B : Schematic representation of the plasmid constructions. White boxes represent the putative CRP binding sites. The 335 and 310 boxes of P1, P2 and P3 promoters are represented by dashed boxes. The lacZ gene is shown by a dotted box. The pac gene is represented by a striped box. The oligonucleotides used for the PCR ampli¢cation are indicated as primers PEC1 and PEC2. Abbreviations: amp, ampicillin resistance gene; tet, tetracycline resistance gene; kan, kanamycin resistance gene ; B, BamHI; H, HindIII; S, SacI; Sa, SalI, Sm, SmaI. H* indicates a partial digestion with HindIII. H(Kl.) indicates a HindIII digestion followed by a treatment with the Klenow fragment of DNA polymerase.
FEMSLE 8874 15-7-99
A. Roa, J.L. Garc|¨a / FEMS Microbiology Letters 177 (1999) 7^14
less, PAA was still able to induce the production of L-Gal in the presence of glucose, in both LB and M63 minimal medium, but the induction e¡ect was signi¢cantly reduced (Table 3). The L-Gal production can be repressed up to 95% in LB medium containing 10 mM glucose, whereas the presence of 20 mM glycerol only produced a limited repression (25%) (data not shown). These results are in agreement with previous ¢ndings [5] and strongly suggest that the pac promoter is regulated by catabolite repression that can be most probably ascribed to the presence of two putative CRP (cAMP receptor protein) binding sites, CRP1 and CRP2, which are located within the 170-bp 5P non-coding region of the pac gene cloned in pSKSE (Fig. 1A). To determine the in£uence of the CRP protein on the regulation of the pac promoter the expression of the lacZ fusion in pSKSE was analyzed in the E. coli CRP3 mutant SBS668. The production of L-Gal in E. coli SBS688 (pSKSE) cells induced with PAA was reduced to the basal levels independently of the absence or the presence of 5 mM glucose in the culture medium (Table 3). As expected, the addition of 5 mM glucose to the culture medium did not generate an inhibitory response on the uninduced cells (Table 3). The levels of L-Gal in the uninduced cells of both the wild-type and the CRP3 strains were quite similar and they were not drastically a¡ected by the addition of glucose suggesting that the basal expression of lacZ fusion is mainly generated from a CRPindependent transcription activity. Surprisingly, these results not only con¢rmed that the catabolite repression of pac promoter was a CRP-dependent process but, more important, they revealed that the CRP protein was directly involved in the PAA induction.
11
3.3. Analysis of pac promoter In spite of the high similarity of the pac genes from E. coli W ATCC 11105 and Kluyvera citrophila, the pac promoter of K. citrophila was located overlapping the sequence of the CRP2 binding site [12], this is, far upstream from the promoter (P3 promoter) proposed for the pac gene of E. coli W ATCC 11105 [6] (Fig. 1A). This ¢nding pointed to the possibility that other promoters might be involved in the expression of the E. coli pac gene. To investigate this possibility, we constructed plasmid pRS55E that contains a lacZ gene fused to a fragment lacking the 310 box of the previously proposed pac promoter [6] or P3 promoter (Fig. 1B). Interestingly, E. coli MC4100 (pRS55E) cells cultured in LB medium showed a high L-Gal production (3000 U after 20 h of culture) that was not induced by PAA (data not shown), but that was repressed by 5 mM glucose (1000 U after 20 h of culture) (data not shown). These results strongly supported the existence of a pac promoter upstream of the HindIII site, and also indicated that the deleted region was involved in the PAA induction. In this sense, the high constitutive activity found in these cells suggests that the pac expression might be regulated by a repressor that recognizes a binding site within the deleted region. A primer extension analysis carried out in E. coli MC4100 (vlacZ) transformed with plasmid pSKSE, showed the presence of three bands of di¡erent intensities (Fig. 2). The same pattern of bands was observed when the RNA was extracted after 20 h (Fig. 2, lane E1) and 3 h (Fig. 2, lane E3) of incubation, suggesting that the relative amounts and the length of the transcripts were not a¡ected by the
Table 3 L-Gal activity of E. coli MC4100 and E. coli SBS688 pSKSE transformants Strain
L-Gal activity (units)a LB
MC4100 (pSKSE) SBS688 (pSKSE)
LB+5 mM glucose
M63+5 mM glucose
3PAA
+PAA
3PAA
+PAA
3PAA
+PAA
330 þ 60 410 þ 50
1500 þ 130 400 þ 50
220 þ 60 330 þ 60
430 þ 50 400 þ 60
130 þ 40 230 þ 50
720 þ 60 230 þ 50
a Activities are expressed in Miller units. Cells were grown for 24 hours at 28³C in LB or M63 minimal medium supplemented with ampicillin (100 Wg/ml) in the absence or presence of 0.1% PAA and in the absence or presence of 5 mM glucose.
FEMSLE 8874 15-7-99
12
A. Roa, J.L. Garc|¨a / FEMS Microbiology Letters 177 (1999) 7^14
phila (Fig. 3, lane E2) [12]. Upstream of the main start site we found the sequences TTAGTT and CATAAT that might represent the 335 and 310 boxes of the P2 promoter (Fig. 1A). These sequences overlap the putative CRP2 binding site and do not perfectly match with the consensus sequence of the E. coli c70 promoters. The 335 box of P2 promoter is preceded by a putative A+T rich enhancer se-
Fig. 2. Primer extension analysis of the transcription start sites of lacZ fusions. Total mRNA was isolated from cells cultured at 28³C in LB medium containing ampicillin (100 Wg ml31 ) and 0.1% PAA. Lane E1, mRNA isolated from E. coli MC4100 (pSKSE) after 20 h of culture; lane E2, mRNA isolated from E. coli MC4100 (pSKSK) after 3 h of culture; lane E3, mRNA isolated from E. coli MC4100 (pSKSE) after 3 h of culture. The size of the extended products was determined by comparison with a DNA sequencing ladder of the pac promoter region using the plasmid pSKSE as template. The primer extensions and the sequencing reactions were performed using the primer PRS551.
phase of growth. The major band corresponded to a transcriptional start site located at 377 bp (P2 promoter) from the ATG translation initiation codon whereas the two minor bands corresponded to putative start sites located at a distance of 381 bp (P1 promoter) and 330 bp (P3 promoter), respectively (Figs. 1 and 3). Interestingly, the major band corresponded to that previously found in the transformants carrying the plasmid pSKSK containing the lacZ gene fused to the pac promoter of K. citro-
Fig. 3. Primer extension analysis of the pac transcription start sites. Total mRNA was isolated from cells cultured at 28³C in LB medium containing tetracycline (12.5 Wg ml31 ) and 0.1% PAA. Lane E1, mRNA isolated from E. coli W ATCC 11105 (pPGA1) cells after 6 h of incubation; lane E2, mRNA isolated from E. coli HB101 (pPGA1) cells after 6 h of incubation; lane E3, mRNA isolated from E. coli HB101 (pPGA1) cells after 15 h of incubation. The size of the extended products was determined by comparison with a DNA sequencing ladder of the pac promoter region using the plasmid pPGA1 as template. The primer extensions and the sequencing reactions were performed using the primer PAC.
FEMSLE 8874 15-7-99
A. Roa, J.L. Garc|¨a / FEMS Microbiology Letters 177 (1999) 7^14
quence [17]. The start site located at 381 bp could be ascribed to the 335 (TATCAA) and 310 (CACGAT) boxes of the putative P1 promoter. The existence of a P1 and P2 promoters explains the expression of lacZ observed with plasmid pRS55E. Finally, the putative start site located at 330 bp from the ATG codon corresponds to that determined by Valle et al. [6] and, thus, the P3 promoter is built by the 335 (TAGATA) and 310 (TAGTAT) boxes that surround the HindIII site (Fig. 1A). It was surprising that the major transcription start site did not correspond to the start site previously located at 3-30 bp by Valle et al. [6]. To rule out the possibility that the band of the P3 promoter might be generated by an unspeci¢c arrest of the reverse transcriptase or by a degradation/processing of pac mRNA, we reproduced the experiment carried out by Valle et al. [6]. Therefore, a primer extension analysis was performed using as hosts both the E. coli strains W ATCC 11105 and HB101 transformed with plasmid pPGA1. Surprisingly, in this case, the main transcription start site was located at 330 bp from the ATG codon (Fig. 3). In contrast, the bands corresponding to the P1 and P2 promoters were signi¢cantly reduced compared to that of P3 promoter. The relative intensities and the positions of the bands were similar when the RNA was extracted after 6 or 15 h of culture (Fig. 3, lanes E2 and E3) indicating that length of the transcripts was independent of the incubation time. The same pattern was observed in both W and HB101 strains suggesting that the dramatic change in the intensities of the bands should not be ascribed to the presence of speci¢c regulatory elements in the parental strain. The above ¢ndings strongly suggest that the pac gene a¡ects its own expression, modifying the transcription rate of the P1, P2 and P3 promoters. In this sense, it has been suggested that the pac gene might control its own expression [7], and although this argument cannot be used to explain the PAA induction observed in plasmid pSKSE, we should consider that the PacR protein could modulate the a¤nity of RNA polymerase for the alternative binding sites. Moreover, the upstream region of pac gene presents several regions that could be involved in DNA bending facilitating the interactions of di¡erent factors (CRP, PAA-dependent repressor, or PcaR) that might function as transcriptional switches between
13
tandem promoters [17]. Finally, it has been demonstrated that some genes such as the luxAB genes can activate or repress transcription from a subset of promoters due to an intrinsically curved DNA segment in the 5P coding sequence of the luxA gene [18]. Summarizing, the results presented in this work provide novel experimental evidence that the regulation of the pac gene is more complex than previously envisioned. The existence of several alternative promoters raises new questions about the nature of the regulatory elements and mechanisms that control the expression of this gene that should be further investigated.
Acknowledgements We thank M.A. Prieto, E. D|¨az, and R. Lo¨pez for critical reading of the manuscript. The artwork of A. Hurtado and the technical assistance of E. Cano and M. Carrasco are gratefully acknowledged. This work was supported by the Comisio¨n Interministerial de Ciencia y Tecnolog|¨a (Grant AMB94-1038-CO2-02 and Grant AMB97-603-CO2-02).
References [1] Sudhakaran, V.K., Deshpande, B.S., Ambedkar, S.S. and Shewale, J.G. (1992) Molecular aspects of penicillin and cephalosporin acylases. Process. Biochem. 27, 131^143. [2] Virden, R. (1990) Structure, procesing and catalytic action of penicillin acylase. Biotech. Gen. Eng. Rev. 8, 189^218. [3] Schumacher, G., Sizmann, D., Haug, H.Buckel,P. and Bo«ck, A. (1986) Penicillin acylase from E. coli: unique gene-protein relation. Nucleic Acids Res. 14, 5713^5727. [4] Oh, S.-J., Kim, Y.-C., Park, Y.-W., Min, S.-Y., Kim, I.-S. and Kang, H.-S. (1987) Complete nucleotide sequence of the penicillin G acylase gene and the £anking regions, and the expression in Escherichia coli. Gene 56, 87^97. [5] Merino, E., Balba¨s, P., Recillas, F., Becerril, B., Valle, F. and Bol|¨var, F. (1992) Carbon regulation and role in nature of the Escherichia coli penicillin acylase (pac) gene. Mol. Microbiol. 6, 2175^2182. [6] Valle, F., Gosset, G., Tenorio, B., Oliver, G. and Bol|¨var, F. (1986) Characterization of the regulatory region of the Escherichia coli penicillin acylase structural gene. Gene 50, 119^122. [7] Jiang, L., Wang, Y. and Yang, S. (1997) Location of regulatory gene in penicillin G acylase operon (pacR) of E. coli D816. Gen. Anal. 14, 51^54. [8] Casadaban, M.J., Mart|¨nez-Arias, A., Shaphira, S.K. and Chou, J. (1983) L-Galactosidase gene fusions for analyzing
FEMSLE 8874 15-7-99
14
[9]
[10]
[11]
[12]
[13]
A. Roa, J.L. Garc|¨a / FEMS Microbiology Letters 177 (1999) 7^14 gene expression in Escherichia coli and yeast. Methods Enzymol. 100, 293^308. Simons, R.W., Houman, F. and Kleckner, N. (1987) Improved single and multicopy lac-based cloning vectors for protein and operon fusions. Gene 53, 85^96. Garc|¨a, J.L. and Buesa, J.M. (1986) An improved method to clone penicillin acylase genes: Cloning and expression in Escherichia coli of penicillin G acylase from Kluyvera citrophila. J. Biotechnol. 3, 187^195. Prieto, M.A., D|¨az, E. and Garc|¨a, J.L. (1996) Molecular characterization of the 4-hydroxyphenylacetate catabolic pathway of Escherichia coli W: Engineering a mobile aromatic degradative cluster. J. Bacteriol. 178, 111^120. Roa, A. and Garc|¨a, J.L. (1997) Identi¢cation of the pac promoter from Kluyvera citrophila. FEMS Microbiol. Lett. 151, 9^16. Sambrook, J,, Fritsch, E.F. and Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
[14] Miller, J.H. (1972) Experiments in Molecular Genetics. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. [15] Snyder, W.B. and Silhavy, T.J. (1995) Beta-galactosidase is inactivated by intermolecular disul¢de bonds and is toxic when secreted to the periplasm of Escherichia coli. J. Bacteriol. 177, 953^963. [16] Ferra¨ndez, A., Min¬ambres, B., Garc|¨a, B., Olivera, E.R., Luengo, J.M., Garc|¨a, J.L. and D|¨az, E. (1998) Catabolism of phenylacetic acid in Escherichia coli. Characterization of a new aerobic hybrid pathway. J. Biol. Chem. 273, 25974^ 25986. [17] Pe¨rez-Mart|¨n, J., Rojo, F. and de Lorenzo, V. (1994) Promoters responsive to DNA bending: a common theme in prokaryotic gene expression. Microbiol. Rev. 58, 268^290. [18] Forsberg, A.J., Pavitt, G.D. and Higgins, C.F. (1994) Use of transcriptional fusions to monitor gene expression: a cautionary tale. J. Bacteriol. 176, 2128^2132.
FEMSLE 8874 15-7-99