Isolation and phenotypic characterization of Pseudomonas aeruginosa pseudorevertants containing suppressors of the catabolite repression control-defective crc-10 allele

Abstract

The amiE gene encodes an aliphatic amidase capable of converting fluoroacetamide to the toxic compound fluoroacetate and is one of many genes whose expression is subject to catabolite repression control in Pseudomonas aeruginosa. The protein product of the crc gene, Crc, is required for repression of amiE and most other genes subject to catabolite repression control in this bacterium. When grown in a carbon source such as succinate, wild-type P. aeruginosa is insensitive to fluoroacetamide (due to repression of amiE expression). In contrast, mutants harboring the crc-10 null allele cannot grow in the presence of fluoroacetamide (due to lack of repression of amiE). Selection for succinate-dependent, fluoroacetamide-resistant derivatives of the crc-10 mutant yielded three independent pseudorevertants containing suppressors that restored a degree of catabolite repression control. Synthesis of Crc protein was not reestablished in these pseudorevertants. All three suppressors of crc-10 were extragenic, and all three also suppressed a Δcrc::tetA allele. In each of the three pseudorevertants, catabolite repression control of amidase expression was restored. Catabolite repression control of mannitol dehydrogenase production was also restored in two of the three isolates. None of the suppressors restored repression of glucose-6-phosphate dehydrogenase or pyocyanin production.

Catabolite repression control, Pseudomonas aeruginosa, Suppressor mutation, Fluoroacetamide

1 Introduction

When presented with both a rapidly metabolizable carbon source and less rapidly metabolizable carbon source(s), most microorganisms utilize the compounds sequentially rather than simultaneously. The rapidly metabolizable compound is consumed first, then the less rapidly metabolizable compound(s). Repression of genes required for metabolism of the less rapidly metabolizable compound(s) is called catabolite repression control (CRC) [1].

Mechanistically, CRC is best understood in the Gram-negative enterics. In Escherichia coli, transport of its preferred carbon source, glucose, results in dephosphorylation of the glucose-specific IIA protein (IIA^glc) of the phosphoenolpyruvate phosphotransfer system. Because adenylate cyclase is stimulated by phosphorylated IIA^glc, dephosphorylation of IIA^glc lowers cAMP levels. Hence transcription of genes requiring cAMP and the cAMP receptor protein (CRP) is reduced (i.e. repressed). Dephosphorylated IIA^glc also mediates CRC by inhibiting permeases and enzymes required for catabolism of certain other carbon sources (reviewed in [2,3]).

In contrast, Pseudomonas aeruginosa exhibits a different pattern of CRC, preferring intermediates of the trichloroacetic acid (TCA) cycle to glucose (reviewed in [4]). Furthermore, neither Vfr (the CRP homologue of P. aeruginosa [5]) nor cAMP plays a role in CRC in P. aeruginosa [6,7]. Mutants defective in CRC of independently inducible regulatory units have been isolated and the responsible gene, crc, mapped [8], cloned [9], and sequenced [7]. The deduced amino acid sequence of Crc displays 25–32% identity with members of a family of DNA repair enzymes possessing apurinic-apyrimidinic endonuclease activity. A virtually identical crc gene has been identified in Pseudomonas putida, where it also is required for CRC of multiple catabolic regulatory units [10]. In addition to its role as a global carbon metabolism regulator, Crc is a component of a signal transduction pathway required for biofilm development and for type IV pilus-dependent twitching motility in P. aeruginosa [11]. To date, no specific biochemical activity has been ascribed to Crc [4,7], although recent evidence was presented for Crc-mediated posttranscriptional regulation of the bkdR message in P. putida [12]. In the present effort to learn more about CRC in P. aeruginosa, a suppressor analysis was initiated. The isolation and phenotypic characterization of pseudorevertants harboring suppressors of the defective crc-10 allele is described here.

2 Materials and methods

2.1 Bacterial strains and growth conditions

P. aeruginosa strain PAO1 (crc⁺, prototroph [13]) and its derivatives PRP705 (crc-5), PRP710 (crc-10), PRP720 (crc-20) [8], and PAO8020 (Δcrc::tetA) [7] have been described. Basal salts medium contained 50 mM KPO₄ (pH 7.0), 15 mM (NH₄)₂SO₄, 1 mM MgCl₂, and 2 μM FeSO₄. All growth was at 37°C. Fluoroacetamide (FAA, Aldrich Chemical Co., Milwaukee, WI, USA) stocks, prepared by dissolving FAA in water (100 mg ml⁻¹) and filter sterilizing, were stored protected from light at 22°C. Tetracycline (Tc) was used at 100 μg ml⁻¹; FAA at 2.5 mg ml⁻¹. Carbon sources were 20 mM (mannitol and glucose) or 40 mM (citrate, fumarate, succinate, and lactate).

2.2 Determination of FAA sensitivity

Cells grown overnight on basal salt agar plates supplemented with the appropriate carbon source were washed from the plate in medium containing the same carbon source. The cells were used to inoculate duplicate side arm flasks containing medium supplemented with the appropriate carbon source with or without FAA. Growth was followed with a Klett–Summerson colorimeter fitted with a no. 66 filter.

2.3 Assays

Cells harvested from mid-exponential phase cultures (125 Klett turbidity units) were disrupted in a French pressure cell at 16 000 psi, and centrifuged (200 000×g, 20 min, 4°C). Amidase, mannitol dehydrogenase, and glucose-6-phosphate dehydrogenase activities in the supernatant fraction of these crude extracts were determined as described [9]. Pyocyanin levels in the cell-free supernatants of cultures grown to saturation (14–16 h) in LB medium were measured as described [14].

3 Results and discussion

TCA cycle intermediates effect strong CRC of many independent, inducible operons/regulons of P. aeruginosa, including the mannitol regulon, the Hex (hexose metabolism) regulon and the amidase operon (reviewed in [4]). amiE, a structural gene of the amidase operon, encodes an amidase [15] capable of cleaving a broad range of aliphatic amides including FAA. Hydrolysis of FAA yields ammonium and fluoroacetate. The resultant fluoroacetate is thought to be converted in the cell to the TCA cycle inhibitor fluorocitrate [16]. Under conditions permissive for amidase expression, sufficient fluoroacetate is generated to render FAA bacteriostatic. Selection for FAA^r has been employed to obtain amidase mutants [16,17]. Because CRC of amidase is lost in crc mutants, we reasoned that even in the presence of a repressing carbon source such as succinate, amidase levels in PRP710 (crc-10) should be high enough to render it sensitive to FAA, while PAO1 (crc⁺) should be resistant. If so, then selection of succinate-dependent FAA^r derivatives of PRP710 (crc-10) should yield a class of mutants in which CRC of amidase expression, and perhaps other CRC responsive operons, has been reestablished.

3.1 Growth of PRP710 (crc-10) is inhibited by FAA

When growing in basal salts–succinate media, the doubling time (T_d) of the mutant strain PRP710 (crc-10) was about 50–55 min (Fig. 1A). The T_d for strain PAO1 was similar (not shown). As expected, growth of PRP710 was severely inhibited by FAA (T_d≥800 min), while PAO1 was partially refractory to this compound (T_d about 100 min) (Fig. 1A).

1

Amidase levels in various strains and the effect of FAA on their growth. A: Growth in basal salts supplemented with succinate was monitored by turbidity as described in Section 2. Cultures containing 2.5 mg ml⁻¹ FAA: PRP710, ○; PAO1, ◻; PAO8030, ×; PAO8031, △; PAO8032, ?. Cultures lacking FAA: PRP710, ● (for clarity not all FAA-free cultures are shown). Data were from a single representative study. B: Amidase activities in cells grown three generations in basal salts media supplemented with succinate and lactamide (diagonal hatching) or mannitol and lactamide (stippled). Average values and standard deviations were derived from three independent succinate–lactamide cultures; for mannitol–lactamide, data were from a single experiment.

3.2 FAA sensitivity and amidase activities are correlated

When grown in succinate medium supplemented with lactamide (an inducer of amiE) amidase levels in PRP710 were 10-fold higher than levels in PAO1 (Fig. 1B). This extent of amidase derepression in PRP710 was consistent with that reported for the crc mutants PRP720 and PAO8020 [7–9]. Like PRP710, growth of these strains in succinate medium was severely inhibited by FAA (not shown). When grown in medium supplemented with mannitol, a non-repressing carbon source that allowed high levels of amidase expression in PRP710 and PAO1 (Fig. 1B), both PRP710 and PAO1 were extremely sensitive to FAA (not shown). Hence, FAA resistance was associated with low amidase levels, while FAA sensitivity was associated with high amidase levels.

3.3 Isolation of FAA^r pseudorevertants

Consistent with the results obtained in liquid media (see Section 3.1), PAO1 formed normal-appearing colonies after 14–18 h of incubation on basal salt agar plates supplemented with succinate and FAA, while PRP710 did not. However, following incubation for 36–48 h a few FAA^r PRP710 colonies appeared. By picking a single FAA^r colony from each of 24 independent platings of PRP710 on basal salts–succinate–FAA agar plates 24 independent, spontaneous, FAA^r derivatives were obtained. A FAA^r phenotype could arise not only from restoration of succinate-mediated CRC of amidase expression, but also from a number of other mechanisms, including the loss of amidase expression, altered FAA transport, or alterations in FAA-sensitive targets. Indeed, 21 of the FAA^r mutants also grew on basal salts–FAA agar plates supplemented with the non-repressing carbon source mannitol. Hence, FAA^r in these 21 mutants was not due to restoration of CRC. They were not studied further. The remaining three mutant strains, designated PAO8030, PAO8031, and PAO8032, grew on basal salt media supplemented with succinate and FAA and on mannitol agar plates, but not mannitol–FAA agar plates. Hence FAA^r in these three mutants required a repressing carbon source.

When cultured in basal medium supplemented with succinate and challenged with FAA, strains PAO8031 and PAO8032 grew at the same rate as PAO1 (T_d about 100 min), while the growth rate of pseudorevertant PAO8030 (T_d about 240 min) was intermediate to that of PAO1 and PRP710 (Fig. 1A). With the non-repressing carbon source mannitol, the pseudorevertants exhibited a FAA^s phenotype (not shown) and expressed high amidase levels (Fig. 1B). Hence FAA^r in the pseudorevertants was associated with the restoration of succinate-mediated CRC of amidase expression.

Lactate is known to exert only weak CRC of amiE [18]. Consistent with weak repression (i.e. higher amidase expression) PAO1 grew slowly, and neither PRP710 nor PAO8030 exhibited sustained exponential phase growth in basal salt–lactate medium supplemented with FAA (Fig. 2A). Interestingly, pseudorevertants PAO8031 and PAO8032 were more refractory to FAA in lactate than was PAO1 (T_d 120 min vs. 220 min, respectively). Strains growing in basal salt–lactate medium supplemented with FAA exhibited an unusual triphasic growth curve (Fig. 2A) characterized by a lag phase, a burst of rapid growth consisting of about one cell division, followed by a slower exponential phase. It has been shown that maximal expression of amidase lags about one generation after addition of inducer [19]. The second phase (rapid growth) may represent growth that occurs before amidase levels have become high enough to produce inhibitory levels of fluoroacetate.

2

Effect of FAA on growth of various strains in lactate, fumarate, and citrate. Growth in basal salts medium supplemented with lactate (A), fumarate (B), or citrate (C) was monitored as described in Section 2. Cultures containing FAA (2.5 mg ml⁻¹): PAO1, ◻; PRP710, ○; PAO8030, ×; PAO8031, △; PAO8032, ?. Cultures lacking FAA: PRP710, ● (for clarity not all FAA-free cultures are shown). Data were from a single representative study.

Like succinate, fumarate imposes strong CRC on many operons. PAO1 grown in fumarate exhibited no sensitivity to FAA, and pseudorevertants PAO8031 and PAO8032 were also very refractory (Fig. 2B). Generation times for PAO1, PAO8031, and PAO8032 in basal salt–fumarate medium with FAA were about half those without FAA (T_d about 55 vs. 100 min, respectively). Pseudorevertant PAO8030 was slightly more FAA^r in fumarate than in succinate (T_d about 200 vs. 240 min, respectively), and even PRP710 exhibited modest growth (T_d about 300 min).

Amidase levels in PAO1 grown in citrate medium with lactamide were about seven-fold higher than in PAO1 grown in succinate (4.0 vs. 0.6 μmol min⁻¹ mg protein⁻¹). This result is consistent with the reported inability of citrate to cause strong CRC of amidase expression in strain PAC1 growing in pyruvate [18]. Specific activities of about 5 μmol min⁻¹ mg protein⁻¹ were also measured in PRP710 and pseudorevertant PAO8030 grown in citrate medium plus lactamide. However, amidase-specific activities in pseudorevertants PAO8031 and PAO8032 were below 0.5 μmol min⁻¹ mg protein⁻¹. Hence, amidase expression in pseudorevertants PAO8031 and PAO8032 was significantly more sensitive to CRC imposed by the weakly repressing carbon source citrate.

In agreement with the relatively high levels of amidase found in citrate-grown PAO1, PAO1 grew more slowly in basal citrate medium plus FAA (Fig. 2C) than in succinate medium with FAA (T_d 240 vs. 100 min). This difference could not be ascribed to differences in the growth rates sustained by citrate and succinate since PAO1 grew at similar rates in succinate and citrate in the absence of FAA. Pseudorevertants PAO8031 and PAO8032 were expected to be FAA^r in citrate medium due to their low levels of amidase; and this indeed was the case. Although amidase levels in PAO8031 and PAO8032 grown in either succinate or citrate were similar, these two pseudorevertants were substantially more resistant to FAA when grown in citrate. This difference may be due to dilution of fluorocitrate by citrate, which could protect aconitase, the presumptive target of fluorocitrate, from inhibition. Both pseudorevertant PAO8030 and PRP710 grew well in basal citrate medium plus FAA, despite relatively high amidase levels (about 25 μmol min⁻¹ mg protein⁻¹). The basis for this phenotype appears to be associated with loss of Crc since several other crc mutants (PRP705, PRP720, and PAO8020), which were inhibited by FAA in succinate, exhibited little or no sensitivity to FAA when grown in citrate. Furthermore, a plasmid carrying crc⁺ causes PRP705 to become as sensitive to FAA in citrate as PAO1 is (data not shown). One might speculate that expression of aconitase is higher in crc mutants than in the wild-type parent PAO1. If this were the case, then the combination of higher target levels and lower inhibitor concentrations (due to dilution by citrate) might protect crc mutants from FAA, despite their high amidase levels.

3.4 Additional characterization of FAA^r pseudorevertants

A Western blot analysis indicated that PRP710 and its CRC⁺ pseudorevertants PAO8030, PAO8031, and PAO8032 produce little or no Crc protein (Fig. 3). This observation suggested that the mutations responsible for suppressing crc-10 are extragenic. Allelic replacement was used to test this hypothesis. A lysate of phage F116 was prepared in PAO8020 (Δcrc::tetA), and used to transduce [20] PAO1, PRP710, and each of the pseudorevertants. Selection for Tc^r allowed recovery of individuals in which the existing crc allele has been replaced by the Δcrc::tetA allele. If the suppressing mutations in the pseudorevertants are intragenic, then they will also be lost by allelic exchange at a high frequency, and the transductants will be sensitive to FAA in succinate (like PAO8020). However, if the suppressors are unlinked to crc-10, and they can suppress the Δcrc::tetA allele, then the transductants should be resistant to FAA on succinate (like the pseudorevertants). Twenty-four Tc^r transductants from each transduction were tested for their ability to grow on succinate plus FAA agar plates. As expected, none of the Tc^r PAO1 transductants grew because replacement of the crc⁺ allele of PAO1 with the Δcrc::tetA allele resulted in loss of CRC and derepression of amidase. Identical results were obtained with the PRP710 (crc-10) transductants. In contrast, 100% of the Tc^r PAO8030, PAO8031, and PAO8032 transductants grew on succinate plus FAA. Thus, the suppressors of crc-10 contained in these pseudorevertants are extragenic, are not informational suppressors such as suppressors of nonsense codons, and are capable of suppressing the Δcrc::tetA null allele.

3

Western blot analysis of Crc production. Crude cell lysates were prepared as described (Section 2) from cells grown in basal salts medium supplemented with succinate. Aliquots containing 10 μg of protein were resolved by SDS-polyacrylamide gel electrophoresis, transferred to an Immobilon-P membrane and probed with polyclonal rabbit anti-Crc antisera and a goat-antirabbit IgG-alkaline phosphatase reporter as previously described [7]. Lane 3 contains 20 ng of purified Crc protein (29.8 kD mol weight). All five crude cell protein preparations show one weak and three strong reacting bands that are not Crc protein. As shown previously [7], these bands also are reactive with preimmune rabbit sera, most likely reflecting natural exposure of the rabbits to P. aeruginosa.

To determine if succinate-mediated CRC was restored to operons other than that containing amiE, mannitol dehydrogenase (MDH) and glucose-6-phosphate dehydrogenase (ZWF) activities were measured. MDH levels in PAO8030 were similar to those in PRP710 (i.e. no repression). However, MDH levels in PAO8031 and PAO8032 were intermediate between those of PAO1 and PRP710 (Fig. 4A). ZWF activities in the three pseudorevertants were indistinguishable from those of PRP710 (Fig. 4B). We conclude that CRC of amidase expression was partially restored in PAO8030, PAO8031, and PAO8032; CRC of MDH expression was partially restored in PAO8031 and PAO8032; and CRC of ZWF expression was not reestablished in any of these pseudorevertants.

4

MDH, ZWF, and pyocyanin levels in various strains. Extracts were prepared from cells grown three generations in basal salts medium supplemented with mannitol or glucose, and MDH and ZWF activities were determined, respectively. A: MDH; mean and standard deviation from three independent cultures are shown. B: ZWF; mean and standard deviation from six independent cultures are shown. C: Pyocyanin; levels in culture supernatants grown to saturation in LB were measured. The mean and standard deviation from two independent cultures are shown. In all cases, values are expressed relative to levels in PAO1 which were 8.4 nmol min⁻¹ mg protein⁻¹ for MDH, 23.5 nmol min⁻¹ mg protein⁻¹ for ZWF, and 2.9 μg ml⁻¹ culture for pyocyanin.

Visual inspection of colonies on agar plates indicated that production of the blue-green pigment pyocyanin is upregulated in PRP710. A quantitative analysis demonstrated that PRP710, and each of the pseudorevertants, produced at least 2.5 times as much pyocyanin as PAO1 (Fig. 4C). Pyocyanin concentrations in PAO8020 (Δcrc::tetA) cultures were only a fraction of those of PAO1 (not shown).

Histidase and urocanase, enzymes of the histidine utilization pathway (Hut), are also subject to succinate-mediated CRC [21]. Succinate was previously found to impose CRC on Hut in mutants harboring the crc-5 allele [22], suggesting that CRC of this pathway is independent of Crc. Consistent with these results, the urocanase levels of PRP710 and PAO1 grown in basal succinate medium with histidine were the same (data not shown). Hence P. aeruginosa contains at least two succinate-responsive CRC pathways: a Crc-independent pathway that controls Hut, and a Crc-dependent pathway that controls expression of amidase, MDH, ZWF, and various other genes. A similar pattern of Crc-dependent and -independent regulation of various catabolic genes was described recently in a P. putida crc mutant [10]. It is tempting to speculate that mutational alterations which broaden the range of genes controlled by a preexisting carbon source-responsive repressor, such as an element that normally controls only Hut, are responsible for suppression of crc-10. Alternatively, if Crc participates in a signaling cascade, then mutational alterations in elements that function either ‘upstream’ or ‘downstream’ of Crc could result in a degree of Crc-independent repression.

Acknowledgements

We thank Worth Calfee, Chantel Sabus, Elizabeth Batten, and Ann Rinaldi for technical assistance; Bill Proctor, Paul Hager, and Tim Brickman for discussions; and P. Hager for reading the manuscript. C.S. and M.J.C. were supported by a training program funded by Glaxo Wellcome, Research Triangle Park, NC (P.V.P.). This work was supported in part by a Starter Research Grant from the East Carolina University School of Medicine (D.N.C.).