AmpG is required for BlaXc beta-lactamase expression in Xanthomonas campestris pv. campestris str. 17

Authors

  • Tsuey-Ching Yang,

    1. Department of Biotechnology and Laboratory Science in Medicine, National Yang-Ming University, Taipei, Taiwan
    Search for more papers by this author
  • Tzu-Fan Chen,

    1. Department of Biotechnology, Asia University, Wufeng, Taichung, Taiwan
    Search for more papers by this author
  • Jeffrey J.P. Tsai,

    1. Department of Biomedical Informatics, Asia University, Wufeng, Taichung, Taiwan
    Search for more papers by this author
  • Rouh-Mei Hu

    Corresponding author
    1. Department of Biotechnology, Asia University, Wufeng, Taichung, Taiwan
    2. Department of Biomedical Informatics, Asia University, Wufeng, Taichung, Taiwan
    • Department of Biotechnology and Laboratory Science in Medicine, National Yang-Ming University, Taipei, Taiwan
    Search for more papers by this author

Correspondence: Rouh-Mei Hu, Department of Biotechnology, Asia University, Wufeng, Taichung 413, Taiwan. Tel.: +886 4 2332 2456 ext. 1834; fax: +886 4 2331 6699; e-mail: rmhu@asia.edu.tw

Abstract

The chromosomal ampRXc-blaXc module is essential for the β-lactam resistance of Xanthomonas campestris pv. campestris. BlaXc β-lactamase is expressed at a high basal level in the absence of an inducer and its expression can be further induced by β-lactam. In enterobacteria, ampG encodes an inner membrane facilitator involved in the recycling of murein degradation compounds. An isogenic ampG mutant (XcampG) of X. campestris pv. campestris str. 17 (Xc17) was constructed to investigate the link between murein recycling and blaXc expression. Our data demonstrate that (1) XcampG is susceptible to β-lactam antibiotics; (2) AmpGXc is essential for expression of blaXc; (3) AmpGs of Xc17, Stenotrophomonas maltophilia KJ (SmKJ) and Escherichia coli DH5α can complement the defect of XcampG; (4) overexpression of AmpGXc significantly increased blaXc expression; and (5) AmpGXc from Xc17 is able to restore β-lactamase induction of the ampNXc-ampGXc double mutant of SmKJ. In Xc17, ampGXc can be expressed from the promoter residing in the intergenic region of ampNXc-ampGXc and the expression is independent of β-lactam induction. AmpN, which is required for β-lactamases induction in SmKJ, is not required for the β-lactam antibiotic resistance of Xc17.

Introduction

Many Gram-negative bacteria carry β-lactamases, which are the leading cause of resistance to β-lactam antibiotics. Studies in the chromosomally encoded AmpC β-lactamases (BlaampC) of Enterobacteriaceae have shown that expression of blaampC is regulated by the transcription factor AmpR and linked to peptidoglycan (murein) recycling (Normark, 1995; Park & Uehara, 2008). Gram-negative bacteria, such as Escherichia coli, recycle about 45% of their murein each generation (Goodell, 1985). Three genes, ampG, ampD and nagZ, required for murein recycling are involved in the regulation of β-lactamase induction. AmpG is a permease responsible for the import of murein degradation compounds, GlcNAc-anhMurNAc peptides (Cheng & Park, 2002). In the cytoplasm, GlcNAc is removed from GlcNAc-anhMurNAc peptides by the β-N-acetylglucosaminidase activity of NagZ to produce anhMurNAc peptides (Cheng et al., 2000; Bacik et al., 2012). AmpD, a cytoplasmic N-acetylmuramyl-l-alanine amidase, cleaves the amide bond between the anhMurNAc and l-alanine of GlcNAc-anhMurNAc peptides and anhMurNAc peptides to release the peptides (Lee et al., 2009; Carrasco-López et al., 2011). The derived amino acids (or peptides) and amino sugars are further processed to produce the murein precursor UDP-MurNAc pentapeptide (Park & Uehara, 2008). AmpR is a LysR family protein. Regulatory activity of AmpR is dependent on the cytoplasmic level of cell-wall degradation products and the murein precursor. The current model of β-lactamase induction based on studies of the chromosmal blaampC in Pseudomonas aeruginosa and Enterobacter cloacae, as well as the blaampC-ampR operon of Citrobacter freundii carried on a plasmid in E. coli, suggests that anhMurNAc tripeptide and UDP-MurNAc peptapeptide are the activating ligand (AL) and the repressing ligand (RL), respectively, of AmpR (Jacobs et al., 1994; Hanson & Sanders, 1999). In noninduced cells, AmpR binds to UDP-MurNAc peptapeptide (RL) to repress blaampC expression. Induction by β-lactam antibiotics leads to an accumulation of anhMurNAc tripeptide (AL) which displaces UDP-MurNAc peptapeptide from AmpR to activate blaampC expression (Lindberg & Normark, 1987; Jacobs et al., 1994; Hanson & Sanders, 1999; Uehara & Park, 2002; Lister et al., 2009). AmpG and NagZ, regulating the import of cell-wall recycling compounds and the synthesis of AL, are required for the activation of blaampC by AmpR (Korfmann & Sanders, 1989; Lindquist et al., 1993; Vötsch & Templin, 2000; Asgarali et al., 2009; Kong et al., 2010; Zhang et al., 2010). Dysfunction of ampD results in an accumulation of AL and, as a consequence, an overexpression of BlaampC in non-induced cells (Jacobs et al., 1995; Langaee et al., 1998; Moya et al., 2008; Schmidtke & Hanson, 2008).

Stenotrophomonas maltophilia, an important opportunistic pathogen of humans, and Xanthomonas campestris pv. campestris, the causal agent of black rot disease in cruciferous plants, are closely related members of the Xanthomonadaceae. Stenotrophomonas maltophilia expresses two chromosomally mediated β-lactamases, BlaL1 and BlaL2. BlaXc β-lactamase in X. campestris is a homologue of BlaL2. AmpR is required for the expression and induction of blaL1, blaL2 and blaXc (Hu et al., 2008; Yang et al., 2011). Our previous studies on Xc17 and SmKJ showed that the expression pattern of β-lactamases in these two species differs in the following aspects: (1) in the absence of β-lactam inducer, the blaL1 and blaL2 promoters are inactive (Hu et al., 2008), while blaXc has a high basal expression (Weng et al., 1999; Yang et al., 2011); (2) induction of blaL1 and blaL2 is fast and significant (Hu et al., 2008), whereas induction of blaXc is slow and moderate (Yang et al., 2011); and (3) like most Gram-negative bacteria, β-lactamase activity of SmKJ is detected in the cellular pellet, while in Xc17, the majority of β-lactamase activity is present in the supernatant (Yang et al., 2011).

In SmKJ, AmpGSm is required for the induction of blaL1 and blaL2 (Huang et al., 2010). Unlike in E. coli in which AmpGEc can work on its own, the activity of AmpGSm in SmKJ requires AmpNSm encoded by the same operon (Huang et al., 2010). An ampN-linked ampG homologue was identified in X. campestris. As there are many differences in β-lactamase expression between Xc17 and SmKJ, it is interesting to compare the transcriptional regulation of the ampN-ampG genes and their regulatory roles in β-lactamase expression between these two species.

Materials and methods

Bacterial strains, plasmids and growth conditions

Bacterial strains and plasmids used in this study are listed in Table 1. Unless otherwise specified, bacteria were grown under conditions as previously described (Huang et al., 2010; Yang et al., 2011).

Table 1. Bacteria and plasmids used in this work
Bacteria/plasmidsDescriptionReferences
  1. a

    The Sm strain SmKJNG was named as KJΔN2GxylEΩ and the plasmids pRKEc-G, pRKSm-G and pRKSm-NG were named as pRK-EcG, pRK-G and pRK-NG, respectively, in a previous paper (Huang et al., 2010).

Bacteria
Xanthomonas campestris pv. campestris
Xc17A local purified virulent Xc17 strainYang et al. (1988)
XcblablaXc mutant derived from Xc17, blaXc::Gm, GmrYang et al. (2011)
XcampGampGXc mutant derived from Xc17, ampGXc::Gm, GmrThis work
XcampNampNXc mutant derived from Xc17, ampNXc::Gm, GmrThis work
 Stenotrophomonas maltophilia
SmKJA clinical isolate from TaiwanHu et al. (2008)
SmKJNGaampNSm-ampGSm double mutant derived from SmKJ, ΔampNSm, ampGSm::xylE-GmΩHuang et al. (2010)
 Escherichia coli
DH5αsupE44 ΔlacU169 (ϕ80 lacZΔM15) hsdR17 recA1 endA1 gyrA96 thi-1 relA1Hanahan (1983)
Plasmids
pBBE. coli cloning vector derived from pOK12, KmrYang et al. (2009a)
pUCGMGentamicin resistance gene cassette donor, Apr, GmrSchweizer (1993)
pFY13-9Promoter-probing vector derived from pRK415, using lacZ as the reporter, TcrLee et al. (2003)
pRK415Broad-host-range plasmid, TcrYang et al. (1988)
pBBampGXc::GmA truncated ampGXc gene in pBB, Kmr, GmrThis work (Fig. 1)
pBBampNXc::GmA truncated ampNXc gene in pBB, Kmr, GmrThis work (Fig. 1)
pFYampGXcPampGXc-lacZ transcriptional construct in pFY13-9, TcrThis work (Fig. 1)
pFYblaXcPblaXc-lacZ transcriptional construct in pFY13-9, TcrYang et al. (2011)
pRKXc-GComplete ampGXc gene from Xc17 in pRK415, TcrThis work
pRKEc-GaComplete ampGEc gene from E. coli in pRK415, TcrHuang et al. (2010)
pRKSm-GaComplete ampGSm gene from S. maltophilia KJ in pRK415, TcrHuang et al. (2010)
pRKSm-NGaComplete ampNSm-ampGSm genes from S. maltophilia KJ in pRK415, TcrHuang et al. (2010)

Construction of ampGXc and ampNXc mutants and complementary plasmid

The restriction maps with the locations of oligonucleotides of ampG genes used in this research are illustrated in Fig. 1. A DNA fragment containing the entire coding sequence of ampGXc (1308 bp) was amplified by PCR using Pfu DNA polymerase with primers ampG-F (AAGCTTATGACCCCGCCCGCATCCAC) and ampG-R (GGATCCTCAATCCGGCCGCCGGTCAT; Fig. 1), and then cloned into pBB to get pBBampGXc. A gentamicin resistance cassette, excised from pUCGM, was inserted into the SacI site of cloned ampGXc to get pBBampGXc::Gm (Fig. 1/Table 1). pBBampGXc::Gm was introduced into Xc17 by electroporation at 25 μF/2.5 kV/200 Ω. pBB, carrying the origin of replication of p15A (Selzer et al., 1983), cannot replicate in Xc17. The double-crossover recombinant, XcampG, with an ampGXc::Gm allele was selected by genamicin resistance and kanamycin sensitivity. The recombinant chromosome was checked by PCR and restriction fragment length analysis. The ampNXc mutant, XcampN, was constructed by using a similar strategy to that for XcampG construction, except that the primers ampN-F (CCGACTCCCCATTCCAGC) and ampN-R (CGCTGCGATAACCGCCGC) were utilized to amplify the ampNXc-containing DNA fragment (874 bp), and that a gentamicin resistance cassette was inserted into the StuI site of ampNXc (Fig. 1).

Figure 1.

Genomic organization of the ampG regions in Xc17, SmKJ and Escherichia coli DH5α and the DNA regions cloned in the plasmids used in this paper. A Gm cassette (black inverted triangle) was inserted into the StuI and SacI sites of pBBampNXc and pBBampGXc, respectively, to form pBBampNXc::Gm and pBBampGXc::Gm for allelic exchange. The insert DNA of pBBampGXc was sucloned into pRK415 to form pRKXc-G. The cloning information of pRKSm-G, pRKSm-NG and pRKEc-G has been described previously (Huang et al., 2010). The yajG gene in Escherichia coli encoding a putative lipoprotein is nonessential for expression of β-lactamase (Lindquist et al., 1993). In Xc17, the ampRXc-blaXc module is not located in the vicinity of ampNXc-ampGXc.

The ampGXc gene cloned in pBBampGXc was subcloned downstream of the lac promoter in the broad-host-range vector pRK415 to form pRKXc-G (Fig. 1/Table 1). Complementary activity of pRKXc-G, pRKSm-G and pRKEc-G (Fig. 1/Table 1) harboring ampG from Xc17, SmKJ and E. coli DH5α, respectively (Huang et al., 2010), was analysed. To distinguish different ampG orthologues, ampG was subscript-labelled with the abbreviation of the species name thereafter. As pRK415 does not have the lac repressor gene, cloned ampGs were expressed from a derepressed lacZYA promoter of E. coli present in pRK415.

Antibiotic susceptibility

Antibiotic susceptibility was determined in Mueller–Hinton medium by a microdilution method according to the guidelines of the Clinical Laboratory Standards Institute (CLSI, 2006). Each inoculum contained 1 × 104 CFU per spot. Minimum inhibitory concentrations (MICs) of each antibiotic were read after 24 h of incubation.

Determination of promoter activity

The PampGXc (523 bp) was amplified with primers ampGp-F (CTCGAGGGTGGAGCAACCGCGGCG) and ampGp-R (TCTAGACAGCCCGCTGCCGAAGCC), and cloned into the broad-host-range plasmid pFY13-9 upstream of the promoterless lacZ gene to yield the PampGXc-lacZ transcriptional fusion (pFYampGXc in Fig. 1). For promoter activity assay, overnight bacterial cultures were adjusted in fresh medium to an initial OD550 of 0.35. In the case of induction, ampicillin was added to a final concentration of 25 μg mL−1. Promoter activities were monitored by determining the β-galactosidase activity in the cells after 1 and 7 h of cultivation as previously described (Hu et al., 2005).

Western blotting

Western blotting with polyclonal anti-Bla antibody was carried out as previously described (Hu et al., 2005).

β-Lactamase activity assay

Xc17 and derived strains were adjusted to an OD550 of 0.35 in fresh media with or without ampicillin (25 μg mL−1) and incubated at 28 °C. The β-lactamase activity in the culture supernatant was analysed 4 h after subculture according to the methods described previously (Yang et al., 2011). For S. maltophilia KJ and its derivatives, overnight cultures were adjusted to an OD420 of 0.15 in fresh media with or without cefuroxime (60 μg mL−1). Cells were harvested from 2-h cultures and the β-lactamase activity in the pellet was assayed. Activity was calculated using a molar absorption coefficient for nitrocefin of 20 500 M−1 cm−1. One unit (U) of β-lactamase activity was defined as the amount of enzyme required to hydrolyse 1 nmol of nitrocefin per minute at 28 °C. The protein concentration was determined using the Bio-Rad protein assay reagent. Bovine serum albumin was used as the standard for protein quantification.

Results and discussion

Sequence analysis of AmpNXc and AmpGXc proteins

Sequence analysis suggested that AmpGXc (XCC3845) has 11 transmembrane helices (Sonnhammer et al., 1998). AmpGXc shares high global identity with its orthologues in other Xanthomonas spp. (94–100%) and Stenotrophomonas sp. SKA14 (78%), and moderate identity with the clinical strain S. maltophilia KJ (35%) and E. coli (34%). A homologue of ampN (XCC3846) was identified upstream of ampGXc. The deduced AmpNXc sequence consists of a conserved exonuclease–endonuclease–phosphatase domain (cd08372) and has 78% global identity to the AmpNSm protein of SmKJ (Huang et al., 2010).

AmpGXc, but not AmpNXc, is required for the β-lactam antibiotic resistance in Xc17

Isogenic mutants of ampNXc and ampGXc (named XcampN and XcampG, respectively) were constructed by insertional mutagenesis in Xc17 (Fig. 1). In comparison with their parental strain, XcampN and XcampG showed a normal growth rate and can elicit symptoms during infection (data not shown). This is consistent with the observation of E. coli that AmpGEc is not required for normal growth (Jacobs et al., 1994). MIC data (Table 2) demonstrated that XcampG was susceptible to penicillin, ticarcillin, carbenicillin and cefuroxime (all with MICs < 4 μg mL−1), whereas the MIC values were not changed in XcampN, suggesting that ampGXc, but not ampNXc, is required for the β-lactam antibiotic resistance. As AmpG is responsible for uptake of murein peptides, the increase of β-lactam susceptibility in XcAmpG indicates that the β-lactam resistance in Xc17 is linked to murein recycling. In S. maltophilia KJ, AmpNSm was proposed to act as an accessory protein of AmpGSm (Huang et al., 2010). As AmpNXc is not required for β-lactam resistance, it is possible that AmpGXc could work without AmpNXc. This hypothesis was confirmed in the complementation assay in S. maltophilia KJ described below. As XcAmpN does not exert any different phenotype in comparison with Xc17, the biological function of ampNXc remains unknown.

Table 2. The MIC (μg mL−1) of different antibiotics for Xc17, SmKJ and their derived strains; the cefoxitin resistance is independent of the Bla β-lactamase; measurements were performed in triplicate
StrainAntibiotic
AmpicillinTicarcillinCarbenicillinCefuroximeCefoxitin
X. campestris pv. campestris
Xc1725612812812864
XcampG< 4< 4< 4< 464
XcampN25612812812864
XcampG(pRKXc-G)25612812812864
XcampG(pRKSm-G)2561281286464
XcampG(pRKSm-NG)2561281286464
XcampG(pRKEc-G)1286412812864
S. maltophilia
SmKJ> 1024102410241024512
SmKJNG< 8< 8< 83232
SmKJNG(pRKSm-G)< 8< 8< 83232
SmKJNG(pRKXc-G)> 1024102410241024512

The ampNXc-ampGXc intergenic region contains promoter activity

The expression patterns of ampG homologues have been studied in several species. In most cases, ampG is the second gene of an operon. The gene upstream of ampG, for example yajG in E. coli (Fig. 1), ampF in P. aeruginosa, and ampN in X. campestris and S. maltophilia, varies between species. The genes yajG and ampF are nonessential for β-lactamase expression (Lindquist et al., 1993; Huang et al., 2010; Kong et al., 2010). In S. maltophilia KJ, ampNSm and ampGSm overlap by four nucleotides (Fig. 1). Expression of ampGSm is dependent on the promoter activity upstream of ampNSm. An insertional mutation in ampNSm provokes a polar effect on the expression of ampGSm (Huang et al., 2010). In Xc17, an intergenic region of 342 bp (designated PampGXc) lies between ampNXc and ampGXc. As the ampNXc mutant has normal phenotypes, it is possible that ampGXc can be expressed from its own promoter. Experiments were carried out using a PampGXc-lacZ transcriptional fusion in pFY13-9 (Fig. 1/Table 1). The β-galactosidase expressed from PampGXc was assayed in Xc17, XcampN and XcampG in the presence or absence of ampicillin. In comparison with the basal level of pFY13-9 (< 10 Miller units), PampGXc indeed contains a promoter activity. The data (Table 3) demonstrated that ampGXc can be expressed from its own promoter. There was no significant change in the expression level in the ampGXc or ampNXc mutants (Table 3). The expression from PampGXc was noninducible by ampicillin (Table 3). This is similar to P. aeruginosa and S. maltophilia in which expression of ampGs is noninducible by β-lactam (Huang et al., 2010; Kong et al., 2010).

Table 3. Promoter activities of ampGXc, and blaXc genes measured with promoter-lacZ transcriptional fusions in Xc17, XcampN and XcampG after 1 and 7 h of cultivation
PromoterStrainβ-Galactosidase activity (Miller units)a
1 h7 h
  1. a

    β-Galactosidase activities are presented as noninduced/induced values. The basal level of expression from the parental plasmid pFY13-9 is below 10 Miller units. Results are mean ± SD of nine readings from three independent experiments.

  2. b

    na, not applicable. As XcampG is β-lactam antibiotic sensitive, the induced expression was not tested.

P ampGXc Xc1783 ± 9/82 ± 359 ± 18/60 ± 15
XcampN95 ± 4/102 ± 1051 ± 7/44 ± 10
XcampG74 ± 9/nab44 ± 8/nab
P blaXc Xc17458 ± 31/410 ± 56371 ± 33/1269 ± 161
XcampN506 ± 4/500 ± 7363 ± 14/1211 ± 37
XcampG58 ± 10/nab43 ± 4/nab

AmpGXc is essential for blaXc expression in Xc17

Expression of blaXc was analysed by promoter assay using the transcriptional fusions PblaXc-lacZ (Yang et al., 2011), by Western blot with polyclonal antibodies raised against BlaXc, and by β-lactamase activity assay. In agreement with the MIC assay, the blaXc promoter (PblaXc) activity of XcampG was significantly reduced to only 13–14% of that of Xc17 (Table 3). Compared with Xc17, the activity showed no significant change in XcampN (Table 3). Western blotting data demonstrated that Xc17 expressed the 31-kDa BlaXc protein (Fig. 2, lane 1) while XcampG did not (Fig. 2, lane 3). The β-lactamase activity assay reconfirmed the loss of the β-lactamase activity (< 0.1 U mL−1) in XcampG. In summary, AmpGXc is required for expression of blaXc.

Figure 2.

Western blot detected the BlaXc protein in Xc17, Xcbla, XcampG and XcampG complemented with different ampG orthologues. The arrow indicates BlaXc (31 kDa) expressed in Xc17 (lane 1). The blaXc-deficient mutant Xcbla was used as a negative control (lane 2). M represents the protein size marker (in kDa).

ampGs from Xc17, SmKJ and E. coli can complement the defect of XcampG

Complete coding sequences of ampG genes were cloned from the chromosome of Xc17, SmKJ and E. coli DH5α into the broad-host-range plasmid pRK415 to form pRKXc-G, pRKSm-G and pRKEc-G, respectively (Fig. 1). Their complementation activity was verified by MIC, Western blot and β-lactamase activity assays. Plasmid pRKSm-NG carrying the entire ampNSm-ampGSm operon from SmKJ was also tested. As shown in Table 2, the MICs for XcampG(pRKXc-G), XcampG(pRKSm-G), XcampG(pRKSm-NG) and XcampG(pRKEc-G) were (almost) restored to the wild-type level. Western blotting showed that XcampG(pRKXc-G), XcampG(pRKSm-G), XcampG(pRKSm-NG) and XcampG(pRKEc-G) were able to synthesize BlaXc (Fig. 2, lanes 4–7 respectively). The β-lactamase activities measured from XcampG complemented with ampGXc, ampGSm and ampGEc were 0.62 ± 0.15, 1.61 ± 0.51, and 0.51 ± 0.06 U mL−1, respectively. In comparison with the wild-type (1.37 ± 0.19 U mL−1) and XcampG (< 0.01 U mL−1), all AmpGs successfully restored β-lactamase production. It is noteworthy that ampGSm, which failed to induce β-lactamase expression in the ampNSm-ampGSm double mutant of SmKJ (Huang et al., 2010), can complement the ampGXc mutant. As ampGSm is also able to restore the MIC of an Xc17-derived ampNampG double mutant to the wild-type level (data not shown), AmpNXc is not needed by AmpGSm in Xc17. In Gram-negative bacteria, the carbohydrate backbone of peptidoglycan is generally conserved, consisting of β-(1-4)-linked N-acetylglucosamine (GlcNAc) and N-acetylmuramic acid (MurNAc) (Schleifer & Kandler, 1972). In E. coli, about 3–6% of the murein subunits have a 1,6-anhydroMurNAc residue (Höltje et al., 1975). Experiments have demonstrated that the presence of the disaccharide GlcNAc-anhMurNAc is the main requirement for the substrates of AmpGEc (Cheng & Park, 2002). However, the disaccharide of the muropeptides of X. campestris pv. campestris is characterized by the lack of an acetyl group on GlcN residues (Erbs et al., 2008). As AmpGEc is active in Xc17, the murein peptides of X. campestris pv. campestris (GlcN-MurNAc peptides) might also be recognized by AmpGEc.

ampGXc from Xc17 is able to complement the ampNG double mutant SmKJNG

It has been reported that AmpGSm fails to complement the ampNSm-ampGSm polar mutant (SmKJNG), while AmpGEc can partially restore the AmpG function of SmKJNG (Huang et al., 2010). Whether ampGXc can complement the defect of SmKJNG is of debate. To address this question, MICs and β-lactamase activity of SmKJNG(pRKXc-G) were tested. As shown in Table 2, the autologous ampGSm, pRKSm-G, could not complement the antibiotic resistance of SmKJNG, whereas ampGXc of Xc17 (pRKXc-G) successfully restored it to the wild-type level. In the presence of an inducer, the specific β-lactamase activity of the wild-type strain (SmKJ) was 397 ± 73 U mg−1 of total protein. No β-lactamase activity [< 0.1 U (mg protein)−1] was detected in SmKJNG. The β-lactamase activity in SmKJNG(pRKXc-G) almost attained the level of the wild-type, with 330 ± 66 U (mg protein)−1. In summary, AmpGXc successfully complemented SmKJNG, indicating that AmpGXc, like AmpGEc, can work independently of AmpNSm.

AmpGXc, AmpGSm and AmpGEc share only 35% sequence similarity. However, cross-complementation suggested that they have a similar biological activity. This might suggest that only a few amino acids are enough to determine their substrate specificity and transportation activity.

AmpG has a dosage effect on expression of blaXc

Induction of β-lactamase depends upon an adequate cytoplasmic concentration of AL. The high basal expression of blaXc in Xc17 suggested that the quantity of AL is high enough to induce blaXc even in the absence of inducer. As AmpG is essential for the import of the precursor of AL, it is possible that the quantity of AmpG could affect the expression level of blaXc. To test this hypothesis, β-lactamase activity was assayed in Xc17 overexpressing AmpGXc. In comparison with Xc17, the basal activity of β-lactamase in Xc17(pRKXc-G) increased about 3.1-fold (from 0.91 ± 0.18 to 2.85 ± 0.37 U mL−1) in the absence of inducer, and about 2.3-fold (from 1.37 ± 0.19 to 3.18 ± 0.27 U mL−1) in the presence of inducer. Overexpression of AmpGXc significantly increased the expression of β-lactamase. This dosage effect of AmpGXc has not been observed in other systems. In E. cloacae and S. maltophilia, AmpG has no dosage effect on the production of β-lactamases (Korfmann & Sanders, 1989; Huang et al., 2010).

Based on our observation, we conclude that the transcriptional regulation of blaXc fits the AL concentration-dependent activation model. The high basal expression of BlaXc suggests a high concentration of AL in the noninduced cell. The dosage effect of AmpGXc on BlaXc expression suggests that AL cannot be efficiently turned over. In E. coli, P. aeruginosa and S. maltophilia, AmpD is responsible for catalysing AL to keep the β-lactamase genes silent in the absence of inducer (Jacobs et al., 1995; Langaee et al., 1998; Yang et al., 2009b). The constitutive presence and rapid accumulation of AL in X. campestris could be explained by a low expression level or a low enzymatic activity of AmpD. Further experimentation will be necessary to clarify these points.

Acknowledgements

We thank Dr Timothy Williams for careful proofreading. This study was supported by Asia University grant 98-NSC-06 to R.M.H. and National Science Council of Taiwan grants NSC 99-2632-E-468-001-MY3 and 101-2221-E-468-022- to J.P.T. and R.M.H., respectively.

Ancillary