Expression of the Pseudomonas putida OCT Plasmid Alkane Degradation Pathway Is Modulated by Two Different Global Control Signals: Evidence from Continuous Cultures

Abstract

Expression of the genes of the alkane degradation pathway encoded in the Pseudomonas putida OCT plasmid are subject to negative and dominant global control depending on the carbon source used and on the physiological status of the cell. We investigated the signals responsible for this control in chemostat cultures under conditions of nutrient or oxygen limitation. Our results show that this global control is not related to the growth rate and responds to two different signals. One signal is the concentration of the carbon source that generates the repressing effect (true catabolite repression control). The second signal is influenced by the level of expression of the cytochome o ubiquinol oxidase, which in turn depends on factors such as oxygen availability or the carbon source used. Since under carbon limitation conditions the first signal is relieved but the second signal is not, we propose that modulation mediated by the cytochrome o ubiquinol oxidase is not classical catabolite repression control but rather a more general physiological control mechanism. The two signals have an additive, but independent, effect, inhibiting induction of the alkane degradation pathway.

Many organisms, particularly plants and algae, produce different types of hydrocarbons (20, 21). It therefore comes as no surprise that many bacterial strains have acquired the ability to use hydrocarbons as carbon sources. However, in most cases hydrocarbons are not preferred growth substrates, and induction of the pathways responsible for their degradation is inhibited to different extents when other carbon sources are simultaneously present (reviewed in references 3 and 4). This regulatory process, called catabolic repression (15), affects the assimilation of several carbon sources in many bacterial species. It has been studied mainly in Escherichia coli and Bacillus subtilis, in which it was found to be the consequence of a complex global regulatory response; from the mechanistic point of view, the global regulatory responses are significantly different in these two bacterial species (22, 23, 26). In the case of pseudomonads, the mechanisms responsible for the global control are not well understood yet. However, the available evidence shows that this regulatory process is, again, mechanistically different from that of E. coli or B. subtilis (4). Furthermore, recent evidence suggests that catabolic repression is just one of several kinds of regulatory responses to environmental and physiological signals that modulate the expression of catabolic pathways for hydrocarbons in Pseudomonas spp. (7, 28, 30). Factors other than the carbon source itself, such as the availability of alternative sigma factors (28), global regulatory proteins like IHF (30), or components of the electron transport chain that are used to monitor the physiological status (7, 19), can have a substantial impact on induction of these pathways.

The alkane degradation pathway encoded in the OCT plasmid of Pseudomonas putida GPo1 is a useful model to study the global regulation responses that affect the induction of catabolic pathways. The genes of this pathway are grouped in two clusters, alkBFGHJKL and alkST (Fig. 1) (31, 32). The first cluster encodes all but one of the different proteins required for the terminal oxidation of the alkane to the corresponding fatty acid, which is further metabolized through β-oxidation. The alkST cluster codes for the transcriptional regulator of the pathway, designated AlkS, and for AlkT. The latter protein forms part of the alkane hydroxylase system and is the first enzyme of the pathway, which is composed of a membrane-bound alkane hydroxylase (AlkB) and two soluble proteins, AlkG (a rubredoxin) and AlkT (a rubredoxin reductase). In the presence of alkanes, the AlkS activator induces expression of the alkBFGHJKL cluster from the promoter PalkB and expression of the alkST genes from the promoter PalkS2 (2, 13, 18). However, the levels of induction achieved are down-modulated depending on the carbon source used and on the physiological status of the cell (2, 7, 9, 25, 34). Induction is efficient when cells are grown in a minimal salts medium at the expense of citrate but is reduced three- to fourfold when organic acids, such as succinate, lactate, or pyruvate, are used as carbon sources. Repression is much stronger when cells are grown exponentially in a rich medium, such as Luria-Bertani (LB) medium (34). This superimposed control occurs by ill-defined mechanisms that ultimately regulate the levels of the AlkS activator, which is an unstable protein present in limiting amounts even under induced conditions (35). Inactivation of the crc gene, which encodes a protein involved in the catabolic repression of several pathways for sugars and other carbon sources (10, 14, 33), partially relieves the repression observed when cells are grown in a rich medium (35). However, under the experimental conditions used, Crc showed little effect on the repression mediated by succinate or pyruvate. Inactivation of the genes encoding the cytochrome o ubiquinol oxidase (Cyo) partially alleviates the repression caused both by organic acids and by the rich medium (7). The effects of these two factors have been studied in cells growing in batch cultures. However, batch cultures do not allow clear separation of the effects of factors such as the concentrations of different nutrients, oxygen availability, and growth rate, and although they offer useful information, the information is limited in several respects. In this study we analyzed modulation of the alkane degradation pathway in chemostat cultures under different nutrient or oxygen limitation conditions. The results provide new experimental evidence indicating that the expression of this pathway is modulated by integration of several different signals.

FIG. 1.

Regulation of the genes encoding the P. putida GPo1 alkane degradation pathway. The genes are grouped into two clusters, alkBFGHJKL and alkST, both of which are regulated by the AlkS protein. When no alkanes are present, alkS is expressed at low levels from promoter PalkS1; AlkS negatively modulates the expression of this promoter, resulting in low constant expression of the gene. When alkanes become available, AlkS activates transcription from the PalkB and PalkS2 promoters, generating a positive amplification loop for alkS expression. Induction of these two promoters by alkanes is negatively modulated by a dominant global control when cells are grown in the presence of alternative carbon sources, such as some organic acids (succinate or pyruvate) or amino acids.

MATERIALS AND METHODS

Bacterial strains and culture media.

Strain PBS4 was derived from P. putida KT2442 by insertion into its chromosome of a PalkB::lacZ transcriptional fusion and of the alkS gene (34). P. putida PBS4C1 was derived from strain PBS4 by inactivation of the crc gene (containing a crc::tet allele) (35). Similarly, strain PBS4B1 was derived from PBS4 by inactivation of the cyoB gene (7). Strains were grown at 30°C in LB medium or in M9 minimal salts medium (24); the latter medium was supplemented with trace elements (1) and succinate as the carbon source. Succinate was added at a concentration of either 10 mM (carbon limitation conditions) or 20 mM (nonlimited conditions and batch cultures). To obtain sulfur limitation conditions, the MgSO₄ in the M9 salts medium was replaced by MgCl₂, and Na₂SO₄ was added at a concentration of 15 μM. For nitrogen limitation conditions, the concentration of NH₄Cl was decreased from 18.8 to 1.88 mM. Cell growth was monitored by measuring turbidity at 540 nm.

Continuous cultures.

Custom-made reactors with working volumes of 90 to 110 ml were used. The growth temperature was 28°C. Flasks were inoculated with 20 ml of an overnight culture, and the preparations were diluted with 80 ml of the appropriate minimal salts medium. In nutrient limitation assays, cells were grown at a dilution rate (D) of 0.05 h⁻¹ for at least three residence times (3 to 4 days), until a steady state was reached. Dicyclopropylketone (DCPK), which is a nonmetabolizable inducer of the alkane degradation pathway, was added to the culture medium at a concentration of 0.05% (vol/vol). The cultures were stirred at 500 rpm with a Teflon-coated magnetic bar. The rate of airflow was 300 ml min⁻¹. For oxygen limitation assays, a BIOSTAT MD fermentor was fitted with a 2-liter M2 culture vessel and was operated at a 1-liter scale. To obtain oxygen-limited conditions, the vessel was sparged with a sterile air-nitrogen mixture. The percentage of O₂ in the culture was monitored with an Ingold oxygen sensor, which was calibrated with 99.8% nitrogen (0% O₂) and 100% air (21% O₂). The medium flow rate ranged from 0.8 to 3.4 ml/min (D = 0.05 to 0.2 h⁻¹).

To grow cells at the maximal growth rates, the D of the chemostat culture was adjusted to just below the washout level (0.85 h⁻¹ for strain PBS4 and 0.7 h⁻¹ for strains PBS4B1 and PBS4C1). The turbidity was monitored by measuring the A₅₄₀, and the maximal growth rate (μ_max) was calculated by using equation μ_max = D + 1/t[ln (X/X₀)], where X is the A₅₄₀ at time t and X₀ is the A₅₄₀ at time zero.

Succinate concentration determination.

To determine the succinate concentration in a culture, 5 ml of culture supernatant was filtered through a 0.22-μm-pore-size filter. Then, 10 μl of the filtrate was injected into a high-performance liquid chromatograph equipped with a 20-cm Waters Spheroids C₈ column and a variable-wavelength detector. Chromatography was performed at a flow rate of 0.8 ml min⁻¹ by using 0.2 M phosphoric acid as the solvent. Succinate was detected at 210 nm, and its concentrations in the sample supernatants were calculated with the help of a standard curve made with samples containing known succinate concentrations.

Assay for β-galactosidase activity.

Culture samples were obtained from the reactors and analyzed for β-galactosidase activity as described by Miller (16). At least two independent assays with duplicate samples were performed in each case.

Real-time reverse transcription-PCR assays.

Twenty-milliliter samples were removed from a reactor and centrifuged at 4°C. The pellets were frozen in liquid nitrogen and kept at −80°C. Total RNA was extracted by using the phenol-guanidine thiocyanate mixture Tri Reagent LS (Molecular Research Center, Inc.). Residual DNA was removed by treatment with DNase I. Total RNA was quantified after it was stained with SYBRgreen II (Molecular Probes) by using a fluorimeter, and its integrity was checked by electrophoresis in 1.5% (wt/vol) agarose gels. Reverse transcription reaction mixtures contained 200 ng of total RNA, 5 U of avian myeloblastosis virus reverse transcriptase (Retroscript; Ambion) per μl, random oligoprimer decamers at a concentration of 5 μM, and 0.5 U of RNase inhibitor per μl. cDNA quantitation was performed by real-time PCR by using a sequence detection system 7700 and Taqman probes obtained from Applied Biosystems. In the case of promoter PalkB, the primers used were 5′-GAAGCAGCGTTGTTGCAGTG-3′ (lacZ forward) and 5′-CAACGGTAATCGCCATTTGA-3′ (lacZ reverse), and the Taqman probe was 6-carboxyfluorescein-5′-CGGCAGATACACTTGCTGATGCGG-3′. For promoter PalkS, the primers were 5′-GCTTTCTGGCATTGGCGTAG-3′ (alkS forward) and 5′-GCCACTGTCTCAGCGCATC (alkS reverse), and the Taqman probe was6-carboxyfluorescein-5′-ACGAATATCCCTACGGGCGAATGACTGT-3′.The melting temperatures of the primers were 58 to 60°C, whereas the melting temperature of the Taqman probes was 70°C. The reaction mixtures (25 μl) contained 12.5 μl of Taqman master mixture (which included the heat-activated TaqGold enzyme [Applied Biosystems]), each oligonucleotide primer at a concentration of 300 nM, 200 nM Taqman probe, and 2-μl portions of different dilutions of the reverse transcriptase product, corresponding to 0.5 to 16 ng of the starting total RNA. The reaction conditions were as follows: 10 min at 95°C for enzyme activation and 40 cycles of 1 min at 95°C and 30 s at 60°C. Fluorescence due to hydrolysis of the probe was measured twice per cycle. The increase in fluorescence versus reaction cycle was plotted, and the threshold cycle value (C_T) (11) was obtained by positioning the threshold baseline at the mid-exponential phase of the curve. The threshold baselines from different PCR runs were standardized by loading a reference sample in duplicate in each plate. The PCR efficiencies for the primer pairs used were checked by using the slopes obtained from real-time PCR performed with serial dilutions of genomic DNA. The slopes obtained were similar (−1/logE = −3.4 ± 0.2), which allowed reliable comparison of the results obtained. Conversion of C_T values to standardized mRNA equivalents (R) was performed by defining a reference curve with the equation C_T = −3.44 logR + 20. By using this equation, an mRNA equivalent was defined as the amount of mRNA that yielded a C_T of 40 under our experimental conditions. Mean values were obtained from at least three experiments.

S1 nuclease protection assays.

Cells were collected by centrifugation and chilled, and total RNA was obtained by using the phenol-guanidine thiocyanate mixture Tri Reagent LS (Molecular Research Center, Inc.). S1 nuclease protection assays to determine the origins and amounts of the transcripts of interest were performed as described previously (2) by using the same amount of RNA in each sample. The single-stranded DNA probe used, which was added in a large excess to titrate the mRNA, was generated by linear PCR as described previously (2). Two oligonucleotides (which had the same transcriptional start site) were used as primers in separate assays, and their 5′ ends hybridized at position 29 or −237 relative to the cyoA translational start site. The plasmid used as the substrate (pPCYOA), which was linearized with ClaI, contained a DNA fragment spanning positions −997 to 150 relative to the cyoA translation start site. This DNA fragment was obtained by PCR amplification from P. putida PBS4 chromosomal DNA and was cloned into pGEM-T Easy (Promega).

RESULTS

Effect of nutrient limitation on expression of promoter PalkB.

P. putida strain PBS4 was used to analyze expression of the PalkB and PalkS2 promoters under different nutrient limitation conditions. This strain contains PalkB::lacZ and the alkS gene inserted into the chromosome and has been extensively characterized in previous studies (2, 7, 34, 35). It allowed us to monitor expression of the PalkB promoter by measuring the accumulation of β-galactosidase. When excess succinate was supplied (conditions of no limitation, allowing maximal growth rates) (Table 1) and in the absence of a pathway inducer, the level of expression of the PalkB promoter under steady-state growth conditions was very low (20 to 30 Miller units). If the nonmetabolizable inducer DCPK was included in the growth medium, the β-galactosidase levels remained very low (about 42 Miller units) (Table 1 and Fig. 2A). This agrees with the previously reported catabolic repression effect of succinate on PalkB induction in batch cultures (34). However, decreasing the growth rate by limiting the concentration of succinate (carbon limitation conditions; D = 0.05 h⁻¹) led to a 52-fold increase in the activity of the PalkB promoter (Fig. 2A and Table 1). The background expression levels in the absence of the inducer DCPK were, however, very low (30 ± 10 Miller units). Similarly, decreasing the growth rate (D = 0.05 h⁻¹) by limiting the availability of sulfur and by limiting the availability of nitrogen increased PalkB activity much less (17- and 3-fold, respectively) (Fig. 2A and Table 1), although the background levels in the absence of DCPK remained low (about 30 Miller units). This indicated that the growth rate per se was not responsible for the levels of expression of PalkB and that the repression observed was, to a large extent, due to true carbon catabolite repression by succinate.

TABLE 1.

Expression of the PalkB promoter in chemostat cultures under different nutrient limitation conditions and in the presence of the inducer DCPK

Limitation	Strain	D (h−1)	Succinate concn (mM)a	A540	PalkB::lacZ activity (Miller units)b	Repression releasec
Noned,e	PBS4	0.85	7	0.20 ± 0.01	42 ± 3	1
	PBS4C1	0.7	6	0.20 ± 0.02	530 ± 25	13
	PBS4B1	0.7	9.2	0.25 ± 0.01	5,016 ± 648	119
Carbonf	PBS4	0.05	<0.1	0.60 ± 0.04	2,200 ± 141	52
	PBS4C1	0.05	<0.1	0.53 ± 0.02	2,700 ± 216	64
	PBS4B1	0.05	<0.1	0.51 ± 0.01	9,900 ± 1,512	236
Sulfurd	PBS4	0.05	11.9	0.45 ± 0.02	697 ± 45	17
	PBS4C1	0.05	10.9	0.40 ± 0.03	579 ± 18	14
	PBS4B1	0.05	19.1	0.40 ± 0.02	2,122 ± 88	50
Nitrogend	PBS4	0.05	13.2	0.39 ± 0.02	130 ± 3	3
	PBS4C1	0.05	12	0.40 ± 0.05	43 ± 3	1
	PBS4B1	0.05	13.2	0.40 ± 0.03	5,275 ± 553	126
O2 and carbonf	PBS4	0.05	<0.1	0.30 ± 0.01	14,993 ± 486	357
	PBS4C1	0.05	<0.1	0.31 ± 0.01	15,325 ± 193	365
O2d	PBS4	0.7	6.0	0.25 ± 0.01	4,380 ± 120	104
Carbong	PBS4	0.05	<0.1	0.40 ± 0.07	3,041 ± 202	72
	PBS4C1	0.05	<0.1	0.40 ± 0.01	3,105 ± 388	74

^aActual succinate concentrations measured in the culture media under steady-state growth conditions.

^bAverage ± standard deviation for three separate experiments.

^cPalkB::lacZ activity relative to that observed in strain PBS4 at the maximal growth rate.

^dSuccinate (20 mM) was supplied as the carbon source (conditions of no limitation; part of the succinate remained unused in the culture medium).

^eMaximal growth rate.

^fSuccinate (10 mM) was supplied as the carbon source (conditions of carbon limitation; all of the succinate was used by cells and could not be detected in the culture medium).

^gSuccinate (10 mM) and 100% O₂ saturation, performed under the conditions used for the O₂ limitation assay.

FIG. 2.

Activities of promoters PalkB and PalkS2 in cells growing under different nutrient limitation conditions. (A) Cells were grown in chemostats with succinate as the carbon source and either at the maximal growth rate (conditions of no limitation) or at a D of 0.05 h⁻¹ by limiting the concentration of either succinate (carbon limitation), SO₄²⁻ (sulfur limitation), or NH₄⁺ (nitrogen limitation). The nonmetabolizable compound DCPK (0.05%, vol/vol) was included in the culture medium to induce transcription from the PalkB and PalkS2 promoters. After cultures reached a steady state, samples were taken, and β-galactosidase levels were measured. The values are averages from two independent assays with duplicate samples (standard deviations, less than 15%). (B) Cells were grown as described above, but samples were processed to obtain total RNA. The activities of promoters PalkB and PalkS2 were measured by real-time reverse transcription-PCR. RNA levels were normalized relative to the amount of total RNA. The values are averages from two independent assays with triplicate samples (standard deviations, less than 20%). NL, nonlimited; ScL, succinate limited; O₂L, oxygen limited (samples from the assay whose results are shown in Fig. 3); wt, wild type.

Expression of the PalkB promoter upon inactivation of the crc or cyoB gene.

When cells are grown in batch cultures, inactivation of the crc gene relieves the repression imposed on PalkB activity by the presence of amino acids in the growth medium but not the repression exerted by organic acids, such as succinate or pyruvate (35). Under similar conditions, inactivation of the cyoB gene relieves the repression caused both by the LB medium and by succinate (7). Neither of these genes is essential for cell viability (7, 35). To get a more detailed view of the effects of these two genes, chemostat cultures were set up with bacterial strains PBS4C1 and PBS4B1, which were derived from strain PBS4 by inactivation of the crc and cyoB alleles, respectively. When cells were grown at the maximal rates with succinate as the carbon source, the activity of the PalkB promoter increased 13-fold in crc-deficient strain PBS4C1 relative to the activity of wild-type strain PBS4, while in the case of cyoB-deficient strain PBS4B1 the increase was 119-fold (Fig. 2A and Table 1). This indicates that crc has an effect on the catabolic repression of the PalkB promoter imposed by succinate, although its influence is much less than that exerted by the cytochrome o ubiquinol oxidase. When the succinate concentration was decreased to obtain carbon limitation conditions, the activity of the PalkB promoter in the crc-deficient strain increased, but the levels observed were essentially the same as those in the wild-type strain grown under carbon limitation conditions (Fig. 2A and Table 1). In the case of the strain lacking the cytochrome o ubiquinol oxidase, however, PalkB activity increased considerably, reaching values that were 4.5-fold higher than those of the wild-type strain grown under the same conditions (Fig. 2A and Table 1). This suggests that induction of PalkB is repressed not only by carbon availability (catabolite repression) but also by some other signal sensed by the cytochrome o ubiquinol oxidase. Under sulfur limitation conditions, PalkB activities were similar in the wild-type and crc-deficient strains, but again the activity increased about threefold in the strain lacking the cytochrome o ubiquinol oxidase (Fig. 2A and Table 1). A similar result was obtained when cells were grown under nitrogen limitation conditions; inactivation of the crc gene did not increase PalkB activity, while inactivation of cyoB led to a 40-fold increase in promoter activity.

Effect of oxygen limitation on expression of the PalkB promoter.

In E. coli, the cytochrome o ubiquinol oxidase is the main terminal oxidase and accommodates most of the electron flow when cells are grown with an ample supply of oxygen (6, 27). When the oxygen supply becomes limiting, the levels of the cytochrome o ubiquinol oxidase decrease, and the function of this enzyme is replaced by other terminal oxidases with higher affinities for oxygen. The cytochrome o ubiquinol oxidase present in Pseudomonas species is believed to play a similar role (5, 12, 36). Given the strong influence of the cytochrome o ubiquinol oxidase on PalkB activity, we analyzed the effect of oxygen availability on PalkB expression. To do this, strain PBS4 was grown in a batch culture to the stationary phase in a fermentor by using 20 mM succinate as the carbon source and an ample supply of oxygen (100% saturation) to obtain a high turbidity value (A₅₄₀, about 1.6). Initially, the PalkB promoter remained silent in spite of the presence of the inducer DCPK (Fig. 3A). However, the activity increased steadily when cells approached the stationary phase of growth, a point at which the oxygen concentration decreased to less than 60%. Culture medium containing 10 mM succinate (C limitation conditions) was then added to the fermentor, and the D was adjusted to 0.05 h⁻¹. The culture was allowed to stabilize until steady-state conditions were reached. At this point, the turbidity was about 0.4, the oxygen concentration was close to 100% saturation, and the β-galactosidase level was up to 2,400 Miller units (Fig. 3A). After several hours in this steady state, the oxygen supply was decreased to 40% saturation. Although the turbidity remained constant, PalkB activity immediately increased, reaching values close to 15,000 Miller units (a sixfold increase) (Fig. 3A). This correlation between oxygen availability and PalkB activity suggests that a decrease in oxygen concentration leads to a decrease in the amount or activity of the cytochrome o ubiquinol oxidase, which is no longer able to exert its repressive effect on PalkB activity. When the same assay was performed with strain PBS4C1, the results obtained were similar (Fig. 3B). This suggests that the crc function is not involved in the response to oxygen availability. However, the response to oxygen availability was different in the case of strain PBS4B1. In this case, PalkB activity was high both when oxygen was present at saturating concentrations and when the oxygen concentration was adjusted to 40% (Fig. 3C).

FIG. 3.

Effect of oxygen availability on the activity of the PalkB promoter. P. putida strains PBS4 (wild type [wt]), PBS4C1 (crc), and PBS4B1 (cyoB) were grown to the stationary phase in a fermentor by using 20 mM succinate as the carbon source in the presence of the inducer DCPK at a concentration of 0.05% (vol/vol). Culture medium containing 10 mM succinate and 0.05% (vol/vol) DCPK was then added to the fermentor, and the D was adjusted to 0.05 h⁻¹. The culture was allowed to stabilize until a steady state was reached, as characterized by turbidity values (A₅₄₀) of about 0.4. The oxygen supply was then decreased from 100 to 40% saturation. PalkB activity was monitored by measuring β-galactosidase activity.

Expression of the cytochrome o ubiquinol oxidase under different growth conditions.

As mentioned above, the levels of the cytochrome o ubiquinol oxidase in E. coli are high when oxygen is abundant and decrease when the oxygen tension decreases (29). In P. putida, the situation is less well understood. The different components of the oxidase are encoded in the cyoABCDE cluster (12). S1 nuclease protection assays allowed us to localize the promoter for the cyoA gene 319 nucleotides upstream from the translation start site. RNA was obtained from cells growing in a fermentor under succinate-limited conditions and at either 100 or 40% oxygen saturation. Expression of cyoA was efficient when cells were grown in the presence of high oxygen concentrations but decreased four- to fivefold when the oxygen tension was decreased to 40% (Fig. 4). Interestingly, the activity of this promoter was low in cells growing in batch cultures with citrate (a nonrepressing substrate) as the carbon source, while use of succinate as the carbon source resulted in fourfold-higher expression (Fig. 4). In summary, the levels of expression of cyoA were high under conditions that led to repression of the PalkB promoter but decreased under growth conditions characterized by a low repression effect.

FIG. 4.

Expression of the cyoA gene under different growth conditions. Cells grown under carbon limitation conditions (10 mM succinate) and in the presence of either high or low oxygen concentrations (100 and 40% saturation, as described in the legend to Fig. 3) were collected, and total RNA was purified. Expression of the cyoA promoter was determined by S1 nuclease protection assays. A similar analysis was performed with cells growing logarithmically in batch cultures by using either succinate (Scc) or citrate (Cit) at a nonlimiting concentration as the carbon source.

The expression pattern of the alkS gene is similar to that of PalkB.

Expression of the alkS gene was monitored by real-time reverse transcription-PCR. Expression of the lacZ gene from promoter PalkB was analyzed in parallel. Under maximal-growth-rate conditions, the activities of the PalkB and PalkS2 promoters were very low. However, under succinate-limited conditions transcription increased about 3-fold for PalkB and 1.8-fold for PalkS (Fig. 2B). Inactivation of the cyoB gene or a decrease in the supply of oxygen led to considerably higher levels of expression of both the PalkB and PalkS2 promoters (25- to 28-fold higher for PalkB and 8- to 12-fold higher for PalkS2 compared to the levels observed under maximal-growth-rate conditions). Inactivation of crc led to PalkB and PalkS2 activities that were somewhat higher than those of the wild-type strain, but the effect was small. The overall expression pattern was similar to that observed by measuring β-galactosidase activity, although the absolute values were higher for the β-galactosidase measurements. It has been shown that the levels of β-galactosidase resulting from promoter PalkB faithfully reproduce the mRNA expression patterns in batch cultures (34). The quantitative differences most likely arise because mRNA turns over faster than β-galactosidase, an enzyme that is relatively stable in P. putida.

DISCUSSION

The use of cells growing under steady-state conditions in chemostats has allowed us to study aspects of the physiological control of the P. putida OCT plasmid alkane degradation pathway that could not be analyzed with batch cultures. When cells were grown at the maximal growth rates with succinate as the carbon source, transcription from the two AlkS-dependent promoters of the pathway, PalkB and PalkS2, was very low, in spite of the presence of the inducer DCPK in the medium. This is consistent with the repressing effect of succinate that was described when cells were grown in batch cultures (2, 34). The expression of these promoters increased considerably when the concentration of succinate was decreased to limiting values. Limiting the concentration of sulfur relieved repression to a lesser extent, while nitrogen limitation had a very small effect. It should be noted that in all three cases, the growth rate was the same (D was adjusted to 0.05 h⁻¹). The fact that limiting the succinate concentration resulted in an increase in PalkB expression, while limiting the nitrogen source did not, indicates that the repression effect is not due to the growth rate itself but rather to true catabolite repression control. Inactivation of the crc gene did not have a large effect on the repression imposed by succinate. This is consistent with previous observations made with batch cultures, suggesting that the role of crc is evident only when cells are using amino acids as carbon sources. However, under maximal-growth-rate conditions, inactivation of the crc gene led to increased expression of the PalkB promoter. Therefore, crc may have some effect on the repression imposed by succinate, although the effect is not large.

Inactivation of the cyoB gene, which codes for an essential subunit of the cytochrome o ubiquinol oxidase, had a large effect on succinate-imposed repression. When cells were growing at the maximal growth rates, a situation in which the repressive effect of succinate was most evident, inactivation of cyoB led to a great increase in the levels of β-galactosidase expressed from promoter PalkB. Interestingly, under carbon limitation conditions PalkB expression was about 4.5-fold higher in the cyoB-deficient strain than in the wild-type strain. Since under these conditions the repressive effect of succinate was relieved, the effect of cyoB mutation indicates that the carbon catabolite repression induced by succinate is not the only signal that has an impact on the expression of promoter PalkB. The cytochrome o ubiquinol oxidase must therefore transmit a different signal. The role of the cytochrome o ubiquinol oxidase has been studied mainly in E. coli, in which it has been found to be the main terminal oxidase when cells are grown with an ample supply of oxygen (29). Under oxygen limitation conditions, the level of expression of the cyo genes decreases, and expression of the cyo genes is replaced by the cytochrome d ubiquinol oxidase, which has a higher affinity for oxygen. This suggests that the signal transmitted by the cytochrome o ubiquinol oxidase for controlling the PalkB promoter could be affected by the oxygen supply. To test this hypothesis, cells were grown in a fermentor under carbon limitation conditions and at saturating oxygen concentrations. When the culture reached a steady state, the oxygen concentration was decreased to 40% saturation. This was immediately followed by an increase in PalkB expression. Since under the conditions used succinate-dependent repression had already been eliminated, the increased PalkB expression should be attributed to the release of a different repressing signal. If the repressing effect of high oxygen concentrations is indeed transmitted through the cytochrome o ubiquinol oxidase, a cyoB-deficient mutant should be insensitive to changes in oxygen concentration. This was observed to be the case. Our finding that expression of the cyoA gene (and presumably of the entire cyo operon) decreased fourfold when the oxygen concentration was decreased from 100 to 40% saturation further supports the hypothesis that the cytochrome o ubiquinol oxidase is a link between oxygen availability and modulation of the PalkB promoter. Furthermore, if succinate was replaced by a carbon source such as citrate, which does not induce carbon catabolite repression (34), the expression of cyoA decreased significantly even in the presence of high oxygen concentrations. The activity of the E. coli cyoA gene has also been shown to vary depending on the carbon source used (6). In our case, this striking observation further linked the levels of the cytochrome o ubiquinol oxidase with control of the PalkB promoter but indicated that the signal transmitted by this oxidase is not only the oxygen concentration but also the activity of the electron transport chain. Oxygen availability, therefore, is just one of the different physiological or environmental signals that can be transmitted through the cytochrome o ubiquinol oxidase. Succinate can feed electrons directly to the electron transport chain through succinate dehydrogenase, which may explain why it stimulates expression of the cytochrome o ubiquinol oxidase. We propose, therefore, that promoter PalkB is subject to two different repressing signals. One signal is a classical carbon catabolite repression control. The other signal is related to the activity of the electron transport chain and is transmitted through the cytochrome o ubiquinol oxidase. A decrease in the concentration of the repressing carbon source does not relieve the repression transmitted through the electron transport chain. Therefore, rigorously speaking, the second signal cannot be considered a classical catabolite repression control signal but rather is a more general physiological control mechanism. Oxygen depletion or the carbon source used affects the activity of the electron transport chain. It is clear that the activity of the electron transport chain is a good mirror of the physiological status, and it is therefore a good way to integrate different input signals and transmit a global regulation response. This strategy is used in other cases, including in the switch from aerobic to anaerobic metabolism in E. coli (8) and in the regulation of expression of photosynthesis genes in Rhodobacter sphaeroides (17). In the case of P. putida, the precise way in which the information concerning the activity of the cytochrome o ubiquinol oxidase is transformed into a transcription regulation signal is still not known.

The levels of expression of alkS followed those of PalkB. The transcription from PalkS2 was always two- to sevenfold lower than that from PalkB. The difference between mRNA levels arising from the PalkS2 and PalkB promoters was small when expression of both promoters was low, while the difference in expression was clearly more important when PalkB activity was high. In other words, moderate increases in PalkS2 expression resulted in progressively higher activity of PalkB. This is in good agreement with the cascade amplification scheme of the pathway (2) (Fig. 1) and with the proposal that catabolite repression is directed to control the levels of the AlkS regulator, an unstable protein that is always present in limiting amounts (35). Therefore, small changes in the AlkS concentration have large effects on PalkB expression.