The PaaX Repressor, a Link between Penicillin G Acylase and the Phenylacetyl-Coenzyme A Catabolon of Escherichia coli W

The pac gene, encoding the penicillin G acylase from Escherichiacoli W, is regulated by the PaaX repressor of the phenylacetatecatabolic pathway. pac expression depends on the synthesis ofphenylacetyl-coenzyme A. PaaX and the cyclic AMP receptor protein(CRP) bind in vitro to the Ppac promoter region. A palindromicsequence proposed as the PaaX operator is located upstream ofthe –35 box overlapping a CRP binding site, an unusualposition that suggests a novel regulatory mechanism.

INTRODUCTION

Top

Introduction

Penicillin G acylase (PAC) (penicillin G aminohydrolase, EC3.5.1.11) is a member of a large enzyme family conventionallyknown as ß-lactam acylases because they are used forthe semisynthesis of ß-lactam antibiotics (30). PAChas a broad substrate range and is able to hydrolyze differentesters and amides of phenylacetic acid (PA) and other aromaticand aliphatic acids, which has made it one of the most importantenzymes used at the industrial scale worldwide (1, 5, 6, 11,33). Although many PAC enzymes have been identified and characterizedin different microorganisms, the PAC from Escherichia coli Wis by far the best-studied enzyme of this family (30, 33).

Despite the ability of PAC to hydrolyze penicillin G, it doesnot have a function in bacterial antibiotic resistance, andits physiological role still remains unclear. Nevertheless,its broad substrate range together with the fact that its synthesisis activated by PA has favored the proposal of PAC as a scavengerenzyme for natural compounds containing a phenylacetate or hydroxyphenylacetateresidue in ester or amide linkage (10, 21). Supporting thishypothesis, we have demonstrated that the pac gene encodingthe PAC is located in the vicinity of the hpa cluster responsiblefor the degradation of 3- and 4-hydroxiphenylacetate in thechromosome of E. coli W (Fig. 1) (10, 23, 24). This observationsuggested the implication of evolutionary selective forces whichfavored the clustering of physiologically interdependent genes,like pac and hpa. However, the recently identified paa clusterresponsible for the catabolism of PA was located very far fromthe pac gene (Fig. 1) (13). Therefore, new data are needed tosolve this apparent paradox.

View larger version (23K):
[in this window]
[in a new window]
FIG. 1. General scheme of the PAC regulatory system. The top of the figure shows the relative localizations of the pac gene (black block), the paa cluster (gray block) for the degradation of PA, and the hpa cluster (cross-hatched block) for the degradation of 4-hydroxyphenylacetate in the E. coli chromosome, referred to as the E. coli K-12 linkage map (10). The regulatory protein PaaX is represented by a gray diamond. The – indicates transcriptional repression. The black cross represents an undetermined factor that controls the dependent PA expression of the pac gene. The synthesis and maturation processes of the PAC protein are shown. The pac gene encodes a polypeptide precursor (preproPAC) which has a molecular mass of 92 kDa and is composed of, in the direction of N to C, a signal peptide (S), an ${alpha}$ subunit ( ${alpha}$ ), a connecting peptide (C), and a ß subunit (ß). The signal peptide directs the export of a polypeptide precursor into the periplasm and is removed to form another precursor (proPAC) after the translocation. Periplasmic processing steps consist of autoproteolysis ( ${alpha}$ +ß) of the precursor and folding of the mature PAC.

It is well known that the synthesis of PAC is probably one ofthe most complex processes for bacterial proteins describedso far since it is subjected to sophisticated regulatory controlsat both the transcriptional and translational levels (Fig. 1)(16, 18, 21, 22, 27, 29, 31). Despite the fact that this systemwas deeply studied, the regulatory proteins involved in PA inductionstill remained unclear. Although a pacR regulatory gene locatedinside the pac gene has recently been identified (7), severaldata obtained by using Ppac::lacZ fusions suggested that otherfactors might be involved in the induction process (21).

Using different genetic and biochemical approaches, both invivo and in vitro, we have demonstrated that the PaaX regulatorthat controls the expression of the PA-coenzyme A (CoA) catabolonis also involved in the transcriptional regulation of the pacgene. This work not only settles the basis for clarifying thepuzzling data about this complex regulatory system obtainedso far but definitively supports the implication of PAC in thePA-CoA catabolon and provides interesting evolutionary evidencethat helps to explain how the cell is able to integrate andtune the regulation of genes that are involved in the same catabolicprocesses.

Localization of a PaaX binding site in the Ppac promoter region.

Top

Localization of a paax...

It has been reported that PaaX negatively controls the expressionof the paa catabolic cluster and that the PA-CoA generated bythe PA-CoA ligase (PaaK) acts as the true inducer of the system(12, 13). Since PAC synthesis is induced by PA, we analyzedthe possibility that PaaX might directly or indirectly controlthe transcription of the pac gene. A detailed analysis of thePpac promoter region showed a potential PaaX binding site locatedbetween positions –125 and –111 with respect tothe transcriptional start site described for the pac gene (27)(Fig. 2). This putative operator has a palindromic sequencewith a pseudodyad axis through a central T base, in perfectagreement with the consensus sequence for the PaaX operatorsdescribed so far (12) (Fig. 2). The inverted sequences of theconsensus PaaX binding region are well conserved in the Ppacpromoter since only one and two nucleotide changes are detectedon the right and left half of the binding site, respectively(Fig. 2).

View larger version (26K):
[in this window]
[in a new window]
FIG. 2. In vitro demonstration of the PaaX function on the Ppac promoter. (A) Nucleotide sequence of the Ppac promoter of E. coli W. The putative –35 and –10 boxes, the ribosome binding site (RBS), and the transcription start site (+1) are underlined. The ATG start codon of pac is shown in italics. The PaaX putative operator located at –118 relative to the transcription start site is indicated by gray letters. Putative CRP binding motifs (CRP 1 [–110.5] and CRP 2 [–68.5]) are boxed. The locations of the primers PAC5' and PAC3' used to amplify the PPAC DNA fragment are shown. (B) Alignment of the nucleotide sequences of the described operators for PaaX. The Pz promoter drives the expression of the paaZ gene, while Pa controls the expression of the operon paaABCDEFGHIJK; both promoters control expression from the paa cluster of the PA degradation pathway in E. coli. The promoters are numbered with respect to the transcription start site (+1) of each respective promoter. Nucleotides that are not conserved relative to the consensus site (12) are indicated by gray letters. (C) In vitro binding of PaaX to the Ppac promoter by gel retardation assay. Lanes 2 and 3 contain 31 and 630 ng, respectively, of total protein of PaaX-free extracts from strain W14(pUC18). Lanes 5 to 8 contain 6.3, 12.6, 31, and 630 ng, respectively, of total protein of extracts obtained from E. coli W14(pAFX). A – indicates that no extract was added to the reaction mixture. (D) PaaX released by PA-CoA. Lanes 2 and 3 contain 31 ng of total protein extracts from the strain E. coli W14(pAFX). PA-CoA (125 µM) was added to the reaction mixture in lane 3. The – indicates that no extract was added to the reaction mixture. The DNA probe used for the experiments whose results are shown in panels C and D was PPAC at 1 nM.

To test the ability of PaaX to bind a Ppac promoter, we performedgel retardation assays (15) using a DNA fragment of 269 bp (PPAC)that covers the entire Ppac promoter as a probe (Fig. 2). ThePPAC fragment was generated by PCR with 10 ng of the plasmidpPGA1 (17) and the primers PAC5' (5'-CGGAATTCTTTACATACAGATAATGACCTGAGC-3')and PAC3' (5'-CGGGATCCTCTATTTTTCATTGTATCCTCTGGC-3'). Cell extractsfrom E. coli W14(pAFX), which expresses the paaX gene, and fromE. coli W14(pUC18), which does not produce PaaX (12), were usedfor the assays. The strain E. coli W14 (14) is a derivativeof E. coli W (9), which lacks the whole paa cluster, includingthe paaX repressor and the paaK ligase coding genes. As expected,whereas the cell extract containing PaaX was able to retardthe migration of the PPAC probe in a protein concentration-dependentmanner, no protein-DNA complexes were detected when the controlextract was tested (Fig. 2). Furthermore, we proved that theinteraction of PaaX with Ppac was specific, since a 100-foldexcess of unlabeled PPAC fragment prevented the formation ofthe PaaX-DNA complex, whereas an excess of salmon DNA did not(data not shown). More important, PA-CoA inhibited the bindingof PaaX to the Ppac promoter, suggesting that it should be thetrue inducer of the system (Fig. 2).

Demonstration that PaaX regulates in vivo the expression of the pac gene.

Top

Demonstration that paax...

To establish in vivo the role of PaaX in pac expression, wefirst compared the PAC activity (2) produced by the mutant E.coli W14 ( ${Delta}$ paa) (14) with that of the wild-type E. coli W (Fig.3). Remarkably, the paa mutant W14 produced a large amount ofPAC in the absence of PA, strongly suggesting that the paa clusteris involved in the regulation of the Ppac promoter and that,most likely, PaaX behaves as a repressor. More interesting,when E. coli W14 was transformed with the low-copy-number plasmidpAAD, which harbors the paa cluster (13), the expression ofthe Ppac promoter recovered its normal PA dependence (Fig. 3),confirming that the Ppac promoter was regulated by a gene ofthe paa cluster. To unequivocally ascribe a function to PaaXin Ppac regulation, we compared the PAC activities of threestrains: E. coli W14(pAAD46), carrying a pAAD derivative plasmidlacking the PaaK ligase-coding gene (13); E. coli W14(pAFX2),harboring a plasmid that expresses in trans the paaX gene underthe control of the Plac promoter (13); and the wild type, E.coli W (Fig. 3). The PAC activity was strongly reduced in theabsence of PaaK ligase, suggesting that PA-CoA is the true inducerof PAC synthesis. In addition, the overexpression of paaX produceda hyperrepression of the Ppac promoter.

View larger version (26K):
[in this window]
[in a new window]
FIG. 3. In vivo demonstration of the role of PaaX in the expression driven by the Ppac promoter. (A) Genetic organization of the paa pathway for the catabolism of PA in E. coli. Relevant genes are indicated by blocks. The regulatory gene paaX and the PA-CoA ligase-coding gene paaK are represented by hatched and black blocks, respectively. The arrows show the direction of gene transcription. Circles indicate Pa, Pz, and Px promoters. Plasmids containing different subcloned DNA fragments are indicated with continuous lines. The location of the insertion of the Tn1000 transposon in plasmid pAAD46 is indicated by a white triangle. (B) Diagram showing the PAC activity produced in E. coli W and W14 cells containing different plasmids. Cells were grown overnight in Luria broth at 28°C in the absence (black columns) or in the presence (hatched columns) of 1 mM PA. The crude protein extracts were obtained as described previously (12). PAC activity was measured as described previously (2), with 1 U of PAC activity being defined as the amount of enzyme required to form 1 nmol of 6-aminopenicillanic acid per min at 37°C. Protein concentrations were determined by the method of Bradford (4).

In vitro analysis of CRP and IHF binding to the Ppac promoter.

Top

In vitro analysis of...

The implication of the presence of the global regulators cyclicAMP receptor protein (CRP) and integration host factor (IHF)in the pac regulatory apparatus has been demonstrated by invivo experiments (27, 29, 32), but direct evidence of the bindingof these proteins to the Ppac promoter has not been obtained.Using the PPAC fragment as a probe, we tested the binding ofpurified CRP (kindly provided by A. Kolb) and IHF (kindly providedby F. Boccard) to the Ppac region by use of a gel retardationassay (Fig. 4). These analyses have demonstrated that CRP interactsdirectly with the Ppac promoter. Remarkably, inasmuch as theamount of the CRP increased, two distinct CRP-PPAC complexeswere detected, a result that is in perfect agreement with thosefor the two putative CRP binding sites predicted in the Ppacregion (Fig. 2). These results also suggest that CRP binds withdifferent affinities to both sites.

View larger version (19K):
[in this window]
[in a new window]
FIG. 4. In vitro binding of purified global regulators CRP and IHF to the Ppac promoter. (A) CRP binding to the Ppac promoter. A gel retardation assay was carried out in the presence of 200 µM cyclic AMP in an electrophoresis buffer. The DNA probe used was PPAC at 1 nM. The – indicates the absence of purified CRP. Lanes 2 to 5 contain 50, 100, 150, and 300 nM of purified CRP, respectively. (B) IHF binding to the Ppac promoter. The DNA probes used were PR-PG at 1 nM and PPAC at 1 nM. Lanes 2 and 6 contain 100 nM protein, lanes 3 and 7 contain 200 nM protein, and lanes 4 and 8 contain 300 nM protein.

The role of IHF as a positive transcriptional regulator of thepac gene has been demonstrated by using an IHF^– strain(29). Furthermore, two putative sites for IHF binding in thePpac promoter between the positions –76 and –88and positions –137 and –149 have been proposed (29).This finding appears to be in agreement with that of the paaregulatory system, where CRP, IHF, and PaaX coregulate the catabolicoperons (12). However, our assay showed that IHF was not ableto retard the migration of the PPAC probe, whereas it was ableto retard our control IHF probe, PR-PG (15). This result suggeststhat IHF does not bind to the Ppac promoter region, and therefore,the response observed in vivo is most likely due to an indirecteffect on other global regulatory systems.

Occurrence of paaX-homologous genes in several strains used for the industrial production of PAC.

Top

Occurrence of paax-homologous...

The PA induction of PAC synthesis is not restricted to E. coliW, since other PAC producer bacteria, like Bacillus megaterium,have shown similar responses to PA (34, 35). Taking into accountthat many other bacteria mineralize PA by homologous paa clusters(10), we have investigated by Southern blot analysis the presenceof paaX-homologous genes in several PAC producer microorganisms(19). The paaX DNA fragment (950 bp) used as a probe containingthe complete paaX gene was generated by PCR with primers X5-Bam(5'-TCGGATCCGTAAACTTGTTACTTTTATCC-3') and X3-Sac (5'-CCGGAGCTCGACCATCTATCTG-3')with plasmid pAAD as a template. The PAAX probe was labeledby the random primer method using the digoxigenin system (BoehringerMannheim). The results shown in Fig. 5 suggest that the threePAC producer strains tested, Kluyvera citrophila ATCC 21285,a close relative of E. coli (3); B. megaterium ATCC 14945 (20);and Providencia rettgeri ATCC 31052 (8), did not contain DNAsequences homologous to paaX.

View larger version (20K):
[in this window]
[in a new window]
FIG. 5. Presence of paaX-homologous genes in heterologous microorganisms. Results of Southern blot analysis of chromosomal DNAs from E. coli W, E. coli W14, E. coli HB101 (28), E. coli MC4100 (25), K. citrophila, B. megaterium, and P. rettgeri are shown. Total DNAs were digested with EcoRI and were probed with fragment PAAX. The sizes of the bands corresponding to the EcoRI fragments that contain sequences homologous to the paaX gene are indicated.

The finding of the lack of a paaX-homologous gene in K. citrophilais in agreement with the fact that it is not able to mineralizePA (unpublished data) and that PAC production is constitutivein this strain (17). Nevertheless, a detailed analysis of thePpac promoter from K. citrophila revealed a nucleotide sequence(5'-GAAATGATTCGCTTT-3') located between positions –64and –78 of its corresponding transcriptional start site(26), which shows only three nucleotide mismatches comparedwith the consensus PaaX operator binding site (Fig. 2). Thisfinding might explain the unexpected PA-dependent expressionof the pac gene from K. citrophila when it was cloned and expressedin E. coli K-12 strains (26). Moreover, as shown in Fig. 5,some E. coli K-12 strains harbor the paaX gene, and therefore,we assumed that we could ascribe to this gene a role in theexpression of the cloned heterologous pac gene. To determineif PaaX might also control the expression of the pac gene fromK. citrophila, we transformed this strain with the plasmid pAFX2expressing in trans the paaX gene from E. coli W. Remarkably,the PAC activity of K. citrophila (pAFX2) was 4.5-fold lowerthan that of the wild-type strain (data not shown), stronglysuggesting that PaaX is also able to recognize the homologousoperator sequence found in K. citrophila and repress pac expression.This result poses the intriguing question of why a pac genethat can be regulated by PaaX still survives in a strain whichdoes not have PaaX and which does not use PA as a carbon source.Perhaps this is because PAC may not be exclusively dedicatedto the generation of PA, and it may fulfill other missions notyet defined. In summary, the results presented above allow usto unequivocally establish that the pac gene of E. coli W isregulated by the PaaX repressor of the paa catabolic cluster.The induction by PA of PAC activity is mediated by the synthesisof the PA-CoA derivative, the first intermediate of the PA catabolicpathway. Moreover, it is worth mentioning the finding that thepalindromic sequence proposed as the PaaX operator for the Ppacpromoter is located upstream of the –35 box, overlappingone of the CRP binding sites. This location is certainly veryunusual and suggests a novel mechanism of repression that willrequire more-sophisticated analyses. Finally, our results reveala quite complex evolutionary scheme for the PA-CoA catabolon.Taking into account both the role of PAC as a PA scavenger enzymeand the fact that both systems, the pac and paa genes, are controlledby the same regulator (PaaX), one might expect that these genesevolved in a coordinated way. However, other data suggest thatthese genes have followed independent evolutionary tracks, sincepac is located far from the paa cluster in E. coli W, it isabsent in E. coli K-12 but contains the paa cluster, and itis alone in K. citrophila. These observations suggest that thepac gene has been acquired by E. coli W as a peripheral pathwayto funnel the PA esters and amides to a resident PA centralpathway. Thereafter, it was further subjugated to the disciplineof PaaX repression by introducing the PaaX palindromic sequenceswithin the Ppac promoter.

ACKNOWLEDGMENTS

We thank E. Díaz for helpful discussions. We are indebtedto F. Boccard and A. Kolb for the kind gift of purified IHFand CRP proteins, respectively. We gratefully acknowledge thetechnical assistance of E. Cano and I. Alonso.

This work was supported by the Comisión Interministerialde Ciencia y Tecnología (grants ABM97-603-C02-02, BIO2003-05309-C04-02,and CICYT-P4 [Optimización de la producción decefalosporina]).

FOOTNOTES

* Corresponding author. Mailing address: Department of Molecular Microbiology, Centro de Investigaciones Biológicas, CSIC, Ramiro de Maeztu 9, 28040 Madrid, Spain. Phone: 34-918373112. Fax: 34-915360432. E-mail: auxi{at}cib.csic.es.

REFERENCES

Top