Carbon-Source-Dependent Expression of the PalkB Promoter from the Pseudomonas oleovorans Alkane Degradation Pathway

Abstract

Pseudomonas oleovorans GPo1 can metabolize medium-chain-length alkanes by means of an enzymatic system whose induction is regulated by the AlkS protein. In the presence of alkanes, AlkS activates the expression of promoter PalkB, from which most of the genes of the pathway are transcribed. In addition, expression of the first enzyme of the pathway, alkane hydroxylase, is known to be influenced by the carbon source present in the growth medium, indicating the existence of an additional overimposed level of regulation associating expression of the alk genes with the metabolic status of the cell. Reporter strains bearing PalkB-lacZ transcriptional fusions were constructed to analyze the influence of the carbon source on induction of the PalkB promoter by a nonmetabolizable inducer. Expression was most efficient when cells grew at the expense of citrate, decreasing significantly when the carbon source was lactate or succinate. When cells were grown in Luria-Bertani rich medium, PalkB was strongly down-regulated. This effect was partially relieved when multiple copies of the gene coding for the AlkS activator were present and was not observed when the promoter was moved to Escherichia coli, a heterologous genetic background. Possible mechanisms responsible for PalkB regulation are discussed.

The genetics and enzymology of bacterial metabolism of n-alkanes have been well characterized for Pseudomonas oleovorans GPo1, which harbors a large plasmid, named OCT (9), encoding the enzymes required to oxidize medium-chain-length (C₆ to C₁₂) n-alkanes to the corresponding terminal acyl coenzyme A derivatives, which then enter the β-oxidation cycle (see reference 50 for a review) (Fig. 1). The genes coding for these enzymes are clustered in two operons, alkBFGHJKL and alkST (50) (Fig. 1). The alkS gene codes for a transcriptional regulator which, in the presence of alkanes, activates expression of the alkBFGHKJL operon (53). This operon is transcribed from a single promoter, PalkB (28). The first enzyme of the pathway, alkane hydroxylase, has attracted much attention due to its ability to oxidize alkanes, alkenes, and related products, yielding alcohols or epoxides (for reviews, see references 51 and 52). Its use as a biocatalyst requires the development of strains harboring the alkane hydroxylase but not the subsequent enzymes of the pathway. This enzyme is composed by three different subunits: a membrane-bound hydroxylase and two soluble proteins, rubredoxin and rubredoxin reductase, which act as electron carriers between NADH and the hydroxylase (36, 38, 49). Alkane hydroxylase is expressed at high levels upon induction, which has been shown to affect the membrane lipid fatty acids (10, 37).

FIG. 1.

The alkane oxidation pathway and transcriptional fusions constructed. (A) Terminal oxidation of medium-chain-length n-alkanes by the enzymes of the pathway encoded in the OCT plasmid, and genes involved. The genes are clustered in two operons; expression from the PalkB promoter is activated by the AlkS regulator in the presence of inducers (adapted from reference 50). (B) PalkB-lacZ and PalkS-lacZ transcriptional fusions used throughout this work. The presumed binding region of the AlkS activator at PalkB (53) is indicated. The DNA segments are not drawn to scale.

The amount of newly synthesized alkane hydroxylase in P. oleovorans varies depending on the carbon source present in the growth medium, which indicates that its expression may be modulated by catabolite repression (20, 46). The term catabolite repression describes a number of regulatory processes that ensure that when the cell is exposed to a preferred carbon source, the catabolic pathways for other, nonpreferred substrates are not induced, even if the appropriate inducers are present (31). Previous analyses of the expression of alkane hydroxylase in cells growing on different carbon sources had been done mainly by measuring enzyme activity in cells harboring either the complete OCT plasmid or the complete alk pathway cloned into a broad-host-range plasmid (20, 46). To identify the minimum determinants that lead to catabolic repression of the alk operon, we have constructed reporter strains containing exclusively the AlkS regulator and the PalkB promoter fused to a reporter gene. Expression of the PalkB promoter was analyzed when these strains were grown at the expense of different carbon sources. We conclude that PalkB promoter activity is modulated by the carbon source used and that its expression in a rich medium is strongly repressed. Our results suggest that repression in Luria-Bertani (LB) rich medium occurs by an interference with AlkS function.

MATERIALS AND METHODS

Bacterial strains and plasmids.

The strains and plasmids used throughout this work are listed in Table 1. General procedures for DNA manipulations were as indicated elsewhere (42). Promoter PalkB (positions −525 to +66 relative to the transcription start point) was obtained by PCR amplification from plasmid pGEc47 (17) with adequate primers, cut with PstI, and cloned between the PstI and SmaI sites of plasmid pUC19Ω, yielding plasmid pPB7 (promoter sequence was verified to ensure that no mutations had been accidentally introduced during amplification). A transcriptional fusion between PalkB and lacZ was obtained by cloning the KpnI-PstI (blunt) fragment from pPB7 containing this promoter between the EcoRI and BamHI (filled-in) sites of plasmid pUJ8 (13), yielding plasmid pUJPB1. The PalkB-lacZ fusion, obtained from pUJPB1 as a NotI fragment, was cloned at the NotI site of mini-Tn5 Km and mini-Tn5 Tet (13), generating plasmids pPBK2 and pPBT1, respectively. These suicide donor plasmids served to deliver the PalkB-lacZ fusion into the chromosomes of Pseudomonas putida KT2442 and KT2440rpoN, respectively. The alkS gene was obtained from plasmid pGEc228 (16) as a HindIII-HpaI fragment and cloned between the SalI and SmaI sites of plasmid pT7-12 (GIBCO-BRL), generating pTS1. This plasmid was cut with EcoRI and HindIII, and the fragment harboring alkS was cloned between the corresponding sites of pUJ8, yielding pUJS1. The alkS gene was excised from this plasmid as a NotI fragment and cloned at the NotI site of the mini-Tn5 suicide donor pJMT6 (43), yielding pTLS1, which served to deliver the alkS gene into the chromosomes of KT2442 and KT2440rpoN. High-copy-number plasmids containing the alkS gene, the PalkB-lacZ fusion, or both were constructed as indicated below, using as vector the broad-host-range high-copy-number plasmid pKT231 (4). To obtain plasmid pHCS1, containing alkS, the HindIII-BsaAI DNA fragment including alkS of plasmid pTS1 was inserted between the HindIII and SmaI sites of pKT231. Plasmid pHCP1, containing the PalkB-lacZ fusion, was constructed by cloning at the HindIII site of pKT231 a 3.8-kbp DNA fragment containing the PalkB-lacZ fusion, excised from plasmid pPBK2 with HindIII endonuclease. This DNA fragment was also cloned into the HindIII site of plasmid pHCS1, obtaining plasmid pHCPR1, which therefore harbors both the alkS gene and the PalkB-lacZ fusion.

TABLE 1.

Strains and plasmids used

Strain or plasmid	Description	Reference or source
E. coli
CC118(λpir)	CC118 lysogenized with λpir phage; host for mini-Tn5 vectors	21
TG1	Host for DNA manipulations	42
W3110	K-12, F−	3
W3110-B1	W3110 with a PalkB-lacZ fusion and alkS in the chromosome	This study
P. putida
KT2442	hdsR; Rif derivative of KT2440	19
KT2440rpoN	hsdR rpoN-Kmr	27
PBS4	KT2442 with a PalkB-lacZ fusion and alkS in the chromosome	This study
PS16	KT2442 with a PalkS-lacZ fusion in the chromosome	This study
PBS18-1	KT2440 rpoN with a PalkB-lacZ fusion and alkS in the chromosome	This study
Plasmids
pKT231	Kmr Smr; RSF1010-derived vector	4
pRK2013	Kmr Mob+ Tra+; donor of transfer functions	18
pT7-12	Apr; ColE1-derived vector	GIBCO-BRL
pUC19Ω	Apr; pUC19 derivative with the Ω interposon at the NarI site	S. Harayama
pUJ8	Apr; vector for construction of lacZ fusions	13
Mini-Tn5 Km	Apr Kmr; mini-Tn5 suicide donor plasmid	13
Mini-Tn5 Tet	Apr Tetr; mini-Tn5 suicide donor plasmid	13
pJMT6	Apr Tetr; mini-Tn5 suicide donor plasmid	43
pGEc47	Tcr; alkBFGHJKL and alkST operons in pLAFR1	17
pGEc228	Apr; SalI fragment containing the alkST operon	16
pPB7	Apr; PalkB promoter (−525 to +66) cloned into pUC19Ω	This study
pUJPB1	Apr; PalkB-lacZ fusion into pUJ8	This study
pPBK2	Apr Kmr; PalkB-lacZ fusion at the NotI site of mini-Tn5 Km	This study
pPBT1	Apr Tetr; PalkB-lacZ fusion at the NotI site of mini-Tn5 Tet	This study
pTS1	Apr; alkS gene from pGRc228 cloned in vector pT712	This study
pUJS1	Apr; alkS gene cloned into pUJ8	This study
pTLS1	Apr Telr; alkS gene cloned at the NotI site of pJMT6	This study
pHCS1	Smr; alkS gene cloned into pKT231	This study
pHCP1	Smr; PalkB-lacZ fusion cloned into pKT231	This study
pHCPR1	Smr; PalkB-lacZ fusion and alkS gene cloned into pKT231	This study
pUJPS16	Apr; PalkS-lacZ fusion into pUJ8	This study
pTPS16	Apr Telr; PalkS-lacZ fusion cloned at the NotI site of pJMT6	This study

A transcriptional fusion between the PalkS promoter and lacZ gene was obtained by cloning between the EcoRI (filled-in) and BamHI sites of plasmid pUJ8 a 407-bp DNA fragment containing the PalkS promoter, obtained by PCR amplification with primers hybridizing 344 nucleotides (nt) upstream and 53 nt downstream, respectively, from the AlkS transcriptional start site. After verification of the sequence, the PalkS-lacZ fusion was excised from the resulting plasmid (pUJPS16) as a NotI segment and cloned at the NotI site of the mini-Tn5 suicide donor pJMT6, generating pTPS16, which served as the vector for introducing the PalkS-lacZ fusion into the P. putida KT2442 chromosome.

Media and culture conditions.

Cells were grown at 30°C (unless otherwise stated) in LB medium or in minimal M9 salts medium (42) supplemented with trace elements (6). Carbon sources were added to a concentration of 30 mM, except Casamino Acids, whose final concentration was 0.2% (wt/vol). Where indicated, the nonmetabolizable inducer dicyclopropylketone (DCPK; 0.05% [vol/vol], final concentration) was added to induce expression of the PalkB promoter. Antibiotics were used at the following concentrations (in micrograms per milliliter): ampicillin, 100; kanamycin, 50; tetracycline, 12; and streptomycin, 50. Potassium tellurite was used at 80 μg/ml.

Conjugal transfers.

Plasmid DNA was introduced into P. putida by conjugation, using plasmid pRK2013 as the donor of transfer functions in triparental matings, as described previously (11).

Assay for β-galactosidase activity.

The activity of PalkB and PalkS promoters was monitored by assaying β-galactosidase accumulation in cells harboring either a PalkB-lacZ fusion and the alkS gene or the PalkS-lacZ fusion. Cells were grown in LB or in minimal salts medium; where indicated, PalkB expression was induced by addition of DCPK (a nonmetabolizable inducer) up to 0.05% (vol/vol). β-Galactosidase activity was measured as described by Miller (33) and is expressed as Miller units.

S1 nuclease analyses of the mRNA originated at the PalkB promoter.

Total RNA was isolated from bacterial cultures as described previously (34). S1 nuclease reactions were performed as described previously (2), using 20 μg of total RNA and an excess of a single-stranded DNA (ssDNA) hybridizing to the 5′ region of the mRNA. The ssDNA probe was generated by linear PCR, using as the substrate plasmid pPB7, which contains a DNA fragment including the PalkB promoter (positions −525 to +66 relative to the PalkB transcription start site). Prior to use as a template for the amplification reaction, the plasmid was cut with PstI (target at positions −524 relative to the PalkB start site). The primer used for amplification was a ³²P-end-labeled 21-mer oligonucleotide, complementary to the mRNA originated at PalkB, whose 5′ end hybridized 68 nt downstream from the PalkB transcription start site. Therefore, the ssDNA generated extended from positions +68 (5′ end) to −524 (3′ end) relative to the PalkB start site.

RESULTS

Construction of a reporter strain containing a transcriptional PalkB-lacZ fusion.

To have a reliable system for monitoring in vivo PalkB activity under different growth conditions, we constructed a PalkB-lacZ transcriptional fusion. The DNA fragment containing the PalkB promoter that was used spanned positions −525 to +66 relative to the transcription start site (Fig. 1B), which should contain the target for the AlkS activator and for any additional unidentified regulatory protein that may participate in regulation of the promoter but does not include alkB translational start site. The PalkB-lacZ fusion and the alkS gene (expressed from its own promoter) were inserted via mini-Tn5 transposons into the chromosome of P. putida KT2442 (19), a well-characterized strain that does not grow on alkanes and that is closely related to P. oleovorans, which has been classified as a P. putida strain (47). Expression of PalkB promoter was analyzed in four independent isolates containing both the PalkB-lacZ fusion and alkS by growing cells to stationary phase either in LB medium or in minimal salts medium containing citrate as the carbon source, in the absence or presence of the nonmetabolizable inducer DCPK, which mimics the inducing effect of octane (20) and has been routinely used as an inducer in most studies of this pathway. In all cases, synthesis of β-galactosidase in noninduced cultures was low, while addition of DCPK efficiently induced the expression of the reporter gene. Since the four isolates had the same behavior, yielding similar induction levels, one of them, named PBS4, was selected for further analyses.

PalkB expression in cells growing at the expense of different carbon sources.

Expression of the PalkB promoter in cells of strain PBS4 growing in a defined medium with different carbon sources, or in rich LB medium, was monitored by following the rate of increase of β-galactosidase after induction of freshly diluted cultures relative to noninduced cultures. Basal expression levels in the absence of inducer were very similar in all growth media tested, being low (20 to 80 Miller units, depending on cell density) during the exponential and early stationary phases of growth and slowly increasing with cell density. In the presence of DCPK, PalkB expression was efficiently induced when cells grew at the expense of citrate, reaching induction values (relative to noninduced cultures) in the range of 70- to 80-fold at mid-exponential phase and increasing up to 95- to 120-fold when cultures approached stationary phase (Fig. 2). When lactate or succinate was used as a carbon source, induction was significantly lower; at mid-exponential phase, induction was about three- to fourfold lower than when cells grew at the expense of citrate, although differences were smaller in stationary phase. These results indicate that lactate and succinate allow only a partial induction of PalkB, while citrate supports a higher induction.

FIG. 2.

Induction of the PalkB promoter in culture media with different carbon sources. P. putida PBS4, harboring a PalkB-lacZ fusion and the alkS gene integrated into the chromosome, was grown in duplicate in LB medium or in minimal salts medium supplemented with citrate (Cit), succinate (Scc), or lactate (Lct). At a cell density of about 0.08, the nonmetabolizable inducer DCPK was added to one of the flasks, leaving the other one as a noninduced control. Aliquots were taken at different times, and β-galactosidase levels were measured. The plot shows the induction of PalkB observed as a function of cell density, calculated as the level of β-galactosidase detected in the presence of inducer divided by that observed in the absence of inducer. A minimum of three to five independent assays were performed for each medium; representative results are shown. Maximum β-galactosidase levels observed corresponded to about 10,000 Miller units. Basal levels in the absence of inducer were low during the exponential and early stationary phases, slowly increasing with cell density (20 to 80 Miller units), and had similar values in all growth media. Basal levels were higher in overnight cultures, which explains why induction values frequently declined in overnight cultures in minimal salts media. The value corresponding to the highest cell density was taken from overnight cultures.

When strain PBS4 was grown in LB medium, PalkB expression was very low during the exponential phase of growth both in the absence and in the presence of inducer; at mid-exponential phase, induction was about 37-fold lower than in cells grown in a defined medium with citrate as the carbon source. Nevertheless, induction increased significantly when cultures approached stationary phase, eventually reaching very high values in stationary-phase cultures (Fig. 2).

Since stability of β-galactosidase could lead to misinterpretation of the results for stationary-phase cultures, the mRNA produced from PalkB in induced and noninduced cultures was analyzed by S1 nuclease assays in the presence of an excess of probe, to allow titration of the mRNA present. In the absence of inducer no mRNA was detected, but in its presence PalkB expression followed the same pattern as that of β-galactosidase: transcription increased steadily immediately after induction when cells grew at the expense of citrate, until a plateau was reached (Fig. 3). The increase in mRNA was slower in cells growing at the expense of lactate or succinate. When cells were grown in LB medium, transcription from PalkB was not detected until cells entered the late exponential phase, reaching high levels in stationary phase (Fig. 3). These results show that the levels of β-galactosidase analyzed in previous sections faithfully reproduce the PalkB expression pattern.

FIG. 3.

Analysis of the transcripts originated at the PalkB promoter in cells grown with different carbon sources. P. putida PBS4, harboring a PalkB-lacZ fusion and the alkS gene integrated into the chromosome, was grown in duplicate in LB medium or in minimal salts medium supplemented with citrate (Cit), succinate (Scc), or lactate (Lct). At a cell density of about 0.08, the nonmetabolizable inducer DCPK was added to one of the flasks, leaving the other one as a noninduced control. Aliquots were taken at different times (intervals of 15 to 60 min, depending on the growth phase), and total RNA was purified. The amount of mRNA originated at PalkB promoter region was analyzed by S1 nuclease assays in the presence of an excess of a ³²P-labeled ssDNA hybridizing to the 5′ end of the transcript (see Materials and Methods). Probe sequences protected from S1 nuclease by hybridization with transcripts originated at the PalkB region were identified in a denaturing polyacrylamide gel, and their amounts were quantified in a Bio-Rad Molecular Imager. A single band (shown in the inserts) of a size corresponding to an mRNA originated at PalkB was obtained. Plots show the amount of signal detected in induced cultures represented versus the cell density observed at each sampling time. No mRNA originated at PalkB could be detected in noninduced cultures.

Differences in β-galactosidase accumulation could be due either to a direct modulating effect on PalkB or to the presence of an additional overlapping promoter which is active only under certain growth conditions. The above analyses to investigate the amounts of mRNA originated from PalkB under different growth conditions also indicated that transcription of the PalkB-lacZ fusion originated in all cases at the same start site, which corresponded to that described for the PalkB promoter (28), and no additional promoters were detected in the vicinity of PalkB (not shown). We therefore conclude that the repression effects observed occur at the level of transcription from PalkB.

Possible factors affecting repression in LB medium.

Silencing of a promoter when cells are grown in LB medium has been observed in some other Pseudomonas catabolic pathways (12, 22, 23, 32, 48) and has been proposed to be an effect that is related to the nature of the compounds in LB medium and is relieved at a certain point of cell growth, presumably when the responsible component(s) is consumed. To test whether this was also the case for PalkB, cells were grown in spent LB medium (medium that has already supported cell growth, filtered and sterilized after adjustment of the pH to 7.0). Interestingly, no repression of PalkB was observed in spent LB medium, induction being efficient from the start of the exponential phase (Fig. 4). Addition of Casamino Acids to minimal salts medium containing citrate or lactate led also to inhibition of the PalkB promoter in the exponential phase of growth. This finding suggests that either the Casamino Acids or the products resulting from their metabolism could be responsible for the repression effect of LB medium, although an effect related to the growth rate in the different culture conditions cannot be excluded.

FIG. 4.

Induction of the PalkB promoter in spent LB medium or in minimal salts medium in the presence of Casamino Acids. P. putida PBS4, harboring a PalkB-lacZ fusion and the alkS gene integrated into the chromosome, was grown in LB medium, in spent LB medium (see text), or in minimal salts medium supplemented with either lactate plus Casamino Acids (Lct+CS) or citrate plus Casamino Acids (Cit+CS). Induction of PalkB was determined as indicated for Fig. 2; the plot shows the induction observed expressed as a function of cell density.

Comparison of the repression levels versus doubling times observed in each medium did not show a correlation between repression and fast growth, neither in defined media nor in LB medium (Fig. 5). For example, the spent LB medium, supporting growth with a doubling time of 51 min, exerted basically no repression on PalkB, while succinate or lactate, supporting slower growth (doubling times of 56 and 60 min, respectively), decreased PalkB induction three and fourfold, respectively, relative to the induction levels observed when cells grew at the expense of citrate. To compare PalkB expression in a particular medium at different growth rates, cells were grown at different temperatures in LB medium or in minimal salts medium supplemented with citrate. The doubling time of strain PBS4 in LB medium at 30°C was 43 min and increased to 50 and 84 min when the incubation temperature was lowered to 25 and 20°C, respectively (Fig. 5). When cells were grown at 25 and 20°C, repression of PalkB during the exponential phase was also observed when LB medium was used but not when minimal salts medium with citrate as the carbon source was used. At mid-exponential phase, PalkB expression in LB medium at 20°C was 42-fold lower than when cells used citrate as the carbon source at the same temperature (Fig. 5). This repression level was high, despite the fact that cells grew slowly. In conclusion, our data indicate that no correlation exists between growth rate and PalkB repression. Rather, it seems that a component of the medium triggers a signal which associates the metabolic status of the cell with PalkB expression.

FIG. 5.

Comparison of the growth rate of P. putida PBS4 in different media with the level of PalkB repression. The doubling times and PalkB repression values for cells grown in minimal salts medium with different carbon sources or in LB medium are represented. Unless indicated otherwise, the growth temperature was 30°C. Repression values denote the induction level of PalkB promoter observed at a cell density of 0.5 (A₆₀₀) in cells grown with the indicated carbon source, measured as indicated for Fig. 2, relative to the induction level observed when cells used citrate as the carbon source (induction in citrate divided by the induction in any other carbon source). All values correspond to the average of several independent assays (error bars are shown). Cit, citrate; Lct, lactate; Scc, succinate; Sp-LB, spent LB medium; Cit+CS, citrate plus Casamino Acids; LB, LB medium.

Expression of the PalkS promoter in cells grown at the expense of different carbon sources.

Since the carbon source present in the medium did not seem to affect PalkB expression in noninduced cultures, we considered that modulation of PalkB induction might be exerted by interfering with AlkS, perhaps limiting transcription of the alkS gene, as has been observed for some other positively regulated catabolic pathways (22, 23). We investigated this possibility by measuring alkS expression under different growth conditions. For this purpose, we constructed a transcriptional fusion bearing the lacZ reporter gene immediately downstream from the sequences driving expression of the alkS gene (Fig. 1). The PalkS-lacZ fusion was inserted into the chromosome of P. putida KT2442 by a mini-Tn5 suicide donor, and β-galactosidase production was measured in cells growing at the expense of different carbon sources. As shown in Fig. 6, transcription from the PalkS promoter was low in all cases, particularly in the exponential phase of growth, irrespective of the carbon source present in the culture medium. Expression was not affected by addition of the inducer DCPK (not shown). Since PalkS activity was basically identical when cells grew in LB medium or in minimal salts medium supplemented with citrate as the carbon source, we conclude that the AlkS levels in exponential phase are high enough to account for the high induction of PalkB observed when cells grow at the expense of citrate and that the low expression of PalkB in LB medium in exponential phase is not due to a repression effect on PalkS.

FIG. 6.

Expression of the PalkS promoter in cultures grown at the expense of different carbon sources. P. putida PS16, harboring a PalkS-lacZ fusion integrated into the chromosome, was grown in LB or in minimal salts medium supplemented with citrate (Cit) or lactate (Lct). Levels of β-galactosidase were measured at different cell densities in each medium. The plot shows the amount of β-galactosidase observed in each case (in Miller units) as a function of cell density (A₆₀₀).

An increase in alkS gene dosage partially relieves PalkB repression in rich medium but not its modulation by organic acids.

If modulation of the PalkB promoter relies on the existence of a repressing factor which competes with AlkS for promoter binding or inhibits AlkS activity in other ways, an increase in AlkS levels might allow the repressing effect to be overcome. This possibility was evaluated by analyzing the effect of increasing the copy number of the alkS gene on PalkB repression, which was achieved by cloning alkS into a broad-host-range high-copy-number plasmid. To investigate whether an increase in alkS gene copy number is accompanied by an increase in AlkS protein levels, we analyzed PalkB expression in cells having different relative copy numbers of alkS and PalkB, using citrate as the carbon source (condition yielding the highest PalkB expression levels). When a multicopy plasmid harboring the PalkB-lacZ fusion was introduced into strain PBS4 (PalkB-lacZ in multicopy and alkS in monocopy), the levels of β-galactosidase upon induction with DCPK were similar to those seen when both PalkB-lacZ and alkS were in monocopy (strain PBS4) (Fig. 7). Nevertheless, when the multicopy plasmid introduced into strain PBS4 contained both the PalkB-lacZ fusion and the alkS gene, a clear increase in PalkB expression was observed upon induction with DCPK. On the one hand, these results suggest that a single copy of the PalkB promoter can titrate out the AlkS protein produced from a single copy of the alkS gene, which agrees with the low expression levels of PalkS observed in Fig. 6. In addition, they show that an increase in the alkS gene dosage leads to higher levels of the AlkS protein. The increase in alkS gene copy number led also to higher alkS mRNA levels (not shown).

FIG. 7.

Effect of increasing the number of copies of the alkS gene on the level of AlkS protein. Expression of the PalkB promoter was analyzed in strains PBS4 (PalkB-lacZ and alkS in monocopy, integrated into the chromosome; rectangles), PBS4 harboring plasmid pHCP1 (PalkB-lacZ in multicopy and alkS in monocopy; triangles), and PBS4 harboring plasmid pHCPR1 (both PalkB-lacZ and alkS in multicopy; circles). Cells were grown in duplicate in minimal salts medium with citrate as carbon source; at an optical density of about 0.08, the inducer DCPK was added to one of the flasks, leaving the other one as a noninduced control. The plot shows the levels of β-galactosidase observed (in Miller units) at different times after induction versus the cell density observed at the moment of sampling. Open symbols correspond to expression observed in the absence of inducer; filled symbols indicate expression in the presence of inducer. LC and HC indicate low copy and high copy, respectively.

Considering this result, we analyzed PalkB expression in cells harboring alkS in multicopy and PalkB-lacZ in monocopy (strain PBS4 with plasmid pHCS1), grown either in minimal salts medium containing citrate or lactate as the carbon source or in LB medium. As shown in Fig. 8A, the presence of multiple copies of the alkS gene did not relieve the repression effect observed when cells grew with lactate as a carbon source: at mid-exponential phase, PalkB induction was about 70-fold when cells grew at the expense of citrate and about 15-fold when lactate was the carbon source used, values that are essentially the same as those observed when alkS was in monocopy. When cells were grown in LB medium, the levels of β-galactosidase reproducibly started to increase earlier when alkS was in multicopy than when it was in monocopy; when cells approached stationary phase, PalkB induction was about 10-fold higher when alkS was in multicopy than when it was in monocopy, although expression was far lower than when cells used citrate as the carbon source (Fig. 8A). It should be noted that the presence of alkS in multicopy did not affect the basal expression of PalkB in the absence of inducer. Analysis of the levels of mRNA originated at PalkB by S1 nuclease assays confirmed the results obtained by measuring β-galactosidase activity (Fig. 8B). We conclude that the increase in alkS copy number, which leads to higher levels of AlkS protein, partially relieves PalkB repression in rich medium but not its modulation by organic acids, which suggests that differences exist in the way repression occurs in each case.

FIG. 8.

Effect of an increase in alkS gene dosage on PalkB modulation in different growth media. (A) P. putida PBS4 harboring plasmid pHCS1 (alkS gene in multicopy and the PalkB-lacZ fusion inserted into the chromosome) was grown in LB medium (open rectangles) or in minimal salts medium supplemented with either lactate (Lct; open triangles) or citrate (Cit; open circles), and induction of the PalkB promoter was assayed as indicated for Fig. 2. The plot shows the induction observed expressed as a function of cell density. PalkB induction in strain PBS4 (PalkB-lacZ and alkS in monocopy) grown in LB is also shown for comparison (filled rectangles). LC, in monocopy; HC, alkS in multicopy. (B) Levels of mRNA originated at the PalkB promoter in strain PBS4 harboring plasmid pHCS1 (alkS gene in multicopy and PalkB-lacZ in monocopy), grown in LB medium, determined as indicated for Fig. 3. The graph shows the mRNA levels observed in samples taken at different times after induction; no mRNA originated at PalkB was detected in noninduced cells.

Transfer of PalkB to Escherichia coli relieves repression in rich medium.

The results described above suggest that the PalkB promoter is affected by an unknown negative factor whose expression or activity depends on the growth substrate used. If a factor specific to P. putida is involved, transfer of the PalkB/AlkS system to an unrelated genetic background would eliminate the repression effect. To test this hypothesis, and considering that the PalkB promoter is expressed in E. coli in an AlkS-dependent way (17), the PalkB-lacZ fusion and the alkS gene were inserted into the chromosome of E. coli W3110 via mini-Tn5 transposons. PalkB expression was analyzed by measuring β-galactosidase production in several independent isolates growing on LB medium, all yielding the same result: the PalkB promoter was efficiently induced by DCPK, and no repression effect during the exponential phase was observed (Fig. 9A). Analysis of the transcripts originated at PalkB confirmed that transcription from PalkB was initiated at the expected position and that expression started immediately and efficiently just after addition of the inducer (Fig. 9B). These results agree with the hypothesis that PalkB repression in LB medium relies on a factor or a mechanism that is not present, or is not active, in E. coli. The behavior of this strain in minimal medium was not analyzed since it has been shown that modulation of PalkB by organic acids in minimal medium does not take place in E. coli (46).

FIG. 9.

Induction of the PalkB promoter in LB medium when transferred to E. coli. E. coli W3110-B1, harboring the PalkB-lacZ fusion and the alkS gene integrated into the chromosome, was grown in LB medium, and induction of the PalkB promoter was assayed by measuring either β-galactosidase activity as indicated for Fig. 2 (A) or the transcripts originated at the promoter as indicated for Fig. 3 (B). The plots show either induction of the PalkB promoter as a function of cell density (A) or the amounts of transcripts originated at PalkB expressed as a function of cell density (B).

DISCUSSION

The results presented indicate that the inhibition of alkane hydroxylase synthesis observed when cells grow at the expense of some organic acids (20, 46) occurs by a direct modulation of the activity of PalkB promoter. Highest expression of PalkB was observed when cells grew at the expense of citrate, declining about three- to fourfold when the carbon source was lactate or succinate. This finding agrees with the general observation that certain organic acids, most frequently succinate, induce catabolic repression in Pseudomonas (15, 22, 30).

Catabolite repression is well understood in E. coli and in Bacillus subtilis (25, 39–41) but not in Pseudomonas (7). Important mechanistic differences seem to exist among these microorganisms. For example, the levels of the cyclic AMP (cAMP)-cAMP receptor protein complex play an important role in catabolic repression in E. coli but not in B. subtilis or Pseudomonas. In P. aeruginosa and P. putida, the levels of cAMP do not vary appreciably with the carbon source, as they do in enteric bacteria (30), and the only cAMP receptor protein analog known, named Vfr, is not involved in catabolic repression but is involved in quorum sensing (1). In addition, glucose is the preferred carbon source in E. coli and B. subtilis, but in Pseudomonas species, organic acids are usually preferred (30). Therefore, carbon catabolite repression in Pseudomonas probably occurs through mechanisms different, at least in some aspects, from those known to be operative in E. coli and B. subtilis.

Succinate and lactate induce catabolic repression on many Pseudomonas promoters (15, 29, 35). Nevertheless, as it occurs for PalkB, it is at present unclear how these organic acids modulate promoter activity. We consider it unlikely that factors such as the modifications in DNA topology caused by general chromatin-associated proteins in response to a nutritional shift-up (5, 24, 44) could mediate PalkB repression, since the effect did not depend on the nutritional shift-up itself but on which particular nutrient was available. Probably, repression is driven by one or more key metabolites whose levels vary depending on the carbon source being used. In our case, the final target seems to be the PalkB promoter. The simplest way to modulate the activity of a positively regulated promoter such as PalkB is by interfering with the ability of the AlkS regulator to activate transcription. The alkS gene seemed to be transcribed with similar efficiencies in the presence of all carbon sources tested, which allows us to discard the notion that PalkB expression could be modulated by varying the transcription levels of the alkS gene. Moreover, the presence of multiple copies of the alkS gene did not relieve the modulation effect caused by lactate or succinate.

When cells were grown in a rich medium such as LB or minimal salts medium supplemented with Casamino Acids, a marked repression of PalkB was observed during the exponential phase of growth. Repression was much stronger than that caused by lactate or succinate, since PalkB induction in LB medium decreased about 37-fold relative to the levels observed in minimal medium with citrate. Repression sharply disappeared when cells entered into stationary phase. A similar strong repression effect of LB medium or Casamino Acids has been observed also for some other promoters of catabolic pathways of Pseudomonas, for example, for the Pu and Ps promoters of the toluene pathway encoded by the P. putida TOL plasmid pWW0 (8, 12, 22, 23, 32) and for the Po promoter of the phenol degradation pathway of Pseudomonas sp. strain CF600 (48). These promoters are recognized by an RNA polymerase associated with the alternative sigma factor ς⁵⁴ (14, 26, 45). Since overexpression of ς⁵⁴ partially relieved the repression of the Pu promoter, it was proposed that repression might be mediated by changes in the activity of the ς⁵⁴ factor itself (8). The PalkB promoter does not show any of the characteristics typical of ς⁵⁴-dependent promoters, and it has been assumed that it is recognized by ς⁷⁰-RNA polymerase (28). Insertion of the PalkB-lacZ fusion and the alkS gene into the chromosome of a P. putida strain lacking ς⁵⁴ showed that the absence of ς⁵⁴ does not affect significantly PalkB expression in LB medium; repression was also observed (not shown). Therefore, PalkB repression in LB medium probably occurs by mechanisms different from those affecting other Pseudomonas ς⁵⁴-dependent promoters.

The factor triggering PalkB repression in LB medium is not known. It is worth noting that growth rate did not correlate with the repression level in rich medium. The use of a spent LB medium eliminated PalkB repression, suggesting that LB medium includes one or more components, which are consumed during cell growth, that trigger PalkB repression. This component(s) might be one or more amino acids or the carbon-to-nitrogen ratio, since addition of Casamino Acids to a minimal salts medium caused PalkB repression as well. Repression in LB did not seem to be caused by the preferential recognition of PalkB by a form of RNA polymerase associated with a stationary-phase sigma factor, since when citrate was the carbon source used, PalkB was efficiently expressed in exponential phase. In addition, analysis of the transcripts originated at the PalkB region under different growth conditions showed that repression occurred by regulation of PalkB itself and excluded the presence of overlapping promoters that could be activated only under certain growth conditions.

Transcription of the alkS gene in LB medium was as efficient as in minimal medium with citrate as the carbon source, indicating that repression in LB medium is not mediated by repression of the alkS gene. Nevertheless, the presence of multiple copies of the alkS gene partially relieved the repression effect. On the one hand, this suggests that repression in LB medium occurs through a mechanism different from that responsible for PalkB modulation by lactate or succinate. On the other hand, the effect of increasing the alkS gene dosage strongly suggests that PalkB repression in LB medium is mediated by a factor that interferes with the ability of the AlkS regulator to activate transcription. Interference could occur by inhibition of AlkS itself or by the presence of a repressor that competes with AlkS for promoter binding. This negative factor seems to be specific of P. putida, since transfer of the PalkB/AlkS system to E. coli completely eliminated the repression effect. This is in contrast to what occurs at the Pu (21) and Po (48) promoters mentioned above, since their transfer to E. coli did not eliminate the repressing effect in LB medium. Therefore, all data suggest that several mechanisms for catabolic repression exist in Pseudomonas, each affecting different promoters or different classes of promoters.