Combination of degradation pathways for naphthalene utilization in Rhodococcus sp. strain TFB

Rhodococcus sp. strain TFB is a metabolic versatile bacterium able to grow on naphthalene as the only carbon and energy source. Applying proteomic, genetic and biochemical approaches, we propose in this paper that, at least, three coordinated but independently regulated set of genes are combined to degrade naphthalene in TFB. First, proteins involved in tetralin degradation are also induced by naphthalene and may carry out its conversion to salicylaldehyde. This is the only part of the naphthalene degradation pathway showing glucose catabolite repression. Second, a salicylaldehyde dehydrogenase activity that converts salicylaldehyde to salicylate is detected in naphthalene-grown cells but not in tetralin-or salicylate-grown cells. Finally, we describe the chromosomally located nag genes, encoding the gentisate pathway for salicylate conversion into fumarate and pyruvate, which are only induced by salicylate and not by naphthalene. This work shows how biodegradation pathways in Rhodococcus sp. strain TFB could be assembled using elements from different pathways mainly because of the laxity of the regulatory systems and the broad specificity of the catabolic enzymes.


Introduction
Rhodococcus sp. strain TFB is a Gram-positive bacterium able to metabolize a wide variety of aromatic compounds, such as phthalate (Tomás-Gallardo et al., 2006), tetralin (Tomás-Gallardo et al., 2009), naphthalene, benzene, biphenyl, salicylate or toluene. Members of the genus Rhodococcus are known for their ubiquity in contaminated environments (Bell et al., 1998), which is explained by the genetic rearrangement and plasticity of the Rhodococcus genomes (Larkin et al., 2006). Other characteristics of this genus are its resistance to genetic manipulation (Larkin et al., 1998) and the redundancy of metabolic pathways and genes in their genomes (van der Geize and Dijkhuizen, 2004;McLeod et al., 2006), which complicates the elucidation of complete biodegradation pathways.
Naphthalene is a polycyclic aromatic hydrocarbon formed by two aromatic rings sharing two carbon atoms. It is one of the most common aromatic compounds present in the environment, and it is released as coal tar and coal tar products, such as creosote (Mueller et al., 1989). The genetics of naphthalene degradation pathways has been studied in detail in Gram-negative bacteria, such as Pseudomonas (Yen and Serdar, 1988) or Ralstonia species (Fuenmayor et al., 1998). In all known pathways, naphthalene catabolism starts with a dihydroxylation producing 1,2-dihydroxynaphthalene, followed by the cleavage of the hydroxylated aromatic ring, resulting in the formation of salicylate (upper pathway). Salicylate is further metabolized via catechol or gentisate pathways depending on the bacteria (lower pathway). In Pseudomonads, naphthalene degradation takes place via catechol formation (Eaton and Chapman, 1992), whereas in Ralstonia sp. strain U2 and Polaromonas naphthalenivorans CJ2, naphthalene catabolism follows the gentisate pathway (Fuenmayor et al., 1998;Jeon et al., 2006). Genetic organization is also different. In Pseudomonas putida G7, the best characterized naphthalene-degrading bacterium, nah genes are organized in two operons in plasmid pNAH7 (Schell et al., 1990). One operon codes for enzymes involved in the conversion of naphthalene to salicylate and the other encodes enzymes for the conversion of salicylate to pyruvate and acetyl coenzyme A via catechol production. In Ralstonia sp. strain U2, the nag genes are organized in a single operon coding for all enzymes involved in the conversion of naphthalene to pyruvate and fumarate through the gentisate pathway (Zhou et al., 2001). In both cases, transcription of nag and nah operons is controlled by LysR-type transcriptional regulators (Park et al., 2002a,b;Jones et al., 2003). These proteins act as positive regulators recognizing salicylate as the inducer molecule.
In previous reports, combining proteomic, genetic and biochemical data, we have described the phthalate and tetralin degradation pathways in Rhodococcus sp. strain TFB (Tomás-Gallardo et al., 2006;2009). Using the same approach, we show in this paper that naphthalene degradation in strain TFB seems to take place by combining, at least, three coordinated but independently regulated sets of genes.

Proteome analysis of naphthalene TFB grown cells and identification of naphthalene-induced proteins
Equal amounts of soluble protein extract from Rhodococcus sp. strain TFB grown exponentially on either glucose or naphthalene were labelled with Cy5 or Cy3 dyes respectively. Proteins specifically expressed in naphthalene-(green spots) or glucose-grown cells (red spots) were detected after scanning (Fig. 1A). A total of 103 (11.7%) out of 883 distinct spots showed naphthalene induction. Of the 103 spots, 49 could be individually selected from the two-dimensional differential in gel electrophoresis (2D-DIGE) gel ( Fig. 1B) for analysis by mass spectrometry, which identified 16 different proteins (Table 1). Of these proteins, eight are also involved in the metabolism of tetralin in TFB (Tomás-Gallardo et al., 2009): the α and β subunits of the tetralin dioxygenase (ThnA1 and ThnA2; spots 7 and 12 respectively), the ferredoxin reductase (ThnA4, spot 5), the cis-biphenyl-2,3-dihydrodiol-2,3-dehydrogenase (ThnB, spots 2 and 11), a possible sterol transfer protein (ThnU, spot 3), the extradiol dioxygenase (ThnC, spot 8 and 9), the aldehyde dehydrogenase (ThnV, spot 4) and the β-ketoadipyl-CoA thiolase (ThnI, spot 13). We also found three proteins (spots 6, 10 and 17) previously identified as induced by tetralin in TFB proteome but not directly involved in tetralin biodegradation. Spot 6 is similar to a group of proteins with chaperone-like functions (Neuwald et al., 1999). Spot 10 was identified as an undecaprenyl pyrophosphate A. Naphthalene-versus glucose-grown cells were compared using DIGE technology. Green spots are those specifically induced by naphthalene. B. Silver stained gel of naphthalene-versus glucose-grown cells used for DIGE analysis with the identified spots marked with a circle. C. Tetralin-versus naphthalene-grown cells using DIGE technology where yellow spots are those induced by both substrates, the green spots are induced only by tetralin, and the red spots are induced only by naphthalene. synthetase, an enzyme involved in cell wall synthesis. Induction of these spots might respond to the stress situation created when growing on naphthalene (Pumphrey and Madsen, 2007). Spot 17 was identified as a protein similar to a transposase from Bacillus halodurans.
Spots 14, 15 and 21, present only on naphthalene cell extracts, were similar to proteins from R. opacus B4. Spot 14 was identified as an hypothetical protein with homology to the Cupin superfamily (Dunwell et al., 2004), and spot 15 was identified as a NAD(P)H-quinone oxidoreductase, an enzyme involved in cell detoxification in response a oxidative stress in different bacteria (Hong et al., 2008;Zhang et al., 2009). Spot 21 was identified as a possible heparin-binding hemaglutinin, an enzyme involved in mycobacterial aggregation (Menozzi et al., 1996).
To confirm that some of the identified proteins were induced in both tetralin-or naphthalene-grown cells, a 2D-DIGE analysis was performed using the soluble protein fraction from tetralin-grown cells labelled with Cy5 and the same fraction from naphthalene-grown cells labelled with Cy3 (Fig. 1C). As shown, most of the spots identified as naphthalene-induced proteins were also induced by tetralin (yellow spots), while few proteins are specifically induced by only one of the molecules.

Analysis of key enzymatic activities on naphthalene degradation
Proteins similar to those known as involved in the production of salicylate from naphthalene were not identified by the proteomic analysis of the cells grown on naphthalene. To test whether salicylate was an intermediate in naphthalene metabolism in TFB, salicylaldehyde dehydrogenase activity was measured in cell-free extracts from TFB cells grown on glucose, tetralin, salicylate, naphthalene or glucose plus naphthalene. Salicylaldehyde-dependent NADH production was detected only in naphthalene-or naphthalene plus glucose-grown cells with similar activities (0.11 ± 0.04 U mg −1 protein and 0.14 ± 0.05 U mg −1 protein respectively). Induction the enzymatic activity that produces salicylate in the presence of naphthalene suggests that salicylate is an intermediate in naphthalene biodeg-radation. Interestingly, this activity is specifically induced by naphthalene but not by salicylate and is not subjected to catabolite repression by glucose.
To elucidate if salicylate is metabolized via catechol or gentisate, catechol 1,2-dioxygenase, catechol 2,3dioxygenase and gentisate dioxygenase activities were assayed with cell-free extracts of TFB cells grown on glucose, naphthalene, salicylate and glucose plus salicylate. Catechol dioxygenase activity was undetectable in any condition (data not shown). However, gentisate dioxygenase activity was measured in naphthalene (0.2 ± 0.01 U mg −1 protein), salicylate (0.7 ± 0.21 U mg −1 protein) and glucose plus salicylate (1.2 ± 0.53 U mg −1 protein) TFB grown cells. This activity was 5.6-fold higher in salicylate-than in naphthalene-grown cells, thus indicating that salicylate is the real inducer. Besides, expression of its coding gene is not subjected to glucose-mediated catabolite repression as the activity found on salicylate plus glucose-grown cell extracts is even higher than that of salicylate-grown cells extracts.

Proteomic analysis of TFB cells grown on salicylate
The different pattern of key enzymatic activities in salicylate-and naphthalene-grown cells suggested that several genes involved in naphthalene degradation could be induced by salicylate. A proteomic analysis was performed comparing the soluble proteome of TFB cells growing on each substrate. A total of 156 and 163 spots were detected in the naphthalene and salicylate proteomes respectively. After matching, 10 spots induced by salicylate were selected for identification by mass spectrometry. Three of the salicylate-induced spots (spots 1, 2 and 3) were also present in the naphthalene proteome although in lower quantity (Fig. 2), and seven were specific of salicylate-grown cells. Spots 1, 2 and 3 were identified as a gentisate 1,2-dioxygenase, a 3-maleylpyruvate isomerase and a mycothiol-dependent maleylpyruvate isomerase, respectively, very similar to those found in different Rhodococcus strains as involved in gentisate catabolism (Table 2). Further analysis of the 3-maleylpyruvate isomerase sequence revealed that this protein belongs to the fumaryl-acetoacetate hydrolases family, just like NagK, the fumaryl-pyruvate hydrolase involved in gentisate metabolism in Ralstonia sp. strain U2 (Zhou et al., 2001) and Corynebacterium glutamicum (Shen et al., 2005). Those three enzymes could be responsible for gentisate degradation in TFB following a micothiol-dependent pathway as it has been described in R. jostii RHA1 (Dosanjh et al., 2008), Rhodococcus sp strain NCIMB 12038 (Liu et al., 2011) and Corynebacterium glutamicum (Feng et al., 2006). No direct function on salicylate degradation could be assigned to the identified spots specifically induced by salicylate. Spot 4 is similar to a hypothetical protein of the Cupin superfamily different to the naphthalene-induced protein (spot 14) in the DIGE analysis but identified as a tetralin-induced protein in a previous analysis (Tomás-Gallardo et al., 2009). Spot 5 is similar to a 3-methyl-2-oxobutanoate hydroxymethyltransferase, a protein involved in biosynthesis of pantotenate and CoA. Its induction could be explained as a compensation for the inhibition of the activity by salicylate as it has been described in Salmonella typhimurium (Primerano and Burns, 1982). Spots 6 and 9 are proteins similar to an alanine dehydrogenase and a 3-hydroxyacyl-CoA dehydrogenase, respectively, both involved in central metabolism. Spot 7 was identified as an acyl-[acyl-carrierprotein] desaturase, an enzyme involved in fatty acid biosynthesis. Spot 8 was identified as a flavodoxin, an enzyme that catalyses electron transfer reactions. Spot 10 was not identified. Naphthalene-induced spots different from those selected in the 2D-DIGE analysis were sent for identification by mass spectrometry. Only two of seven spots were identified in this analysis ( Fig. 2 and Table 1). Spots 11 and 12 are similar to a flavin-binding mono-oxygenase and an esterase/lipase, respectively, both located together in R. jostii RHA1 genome but with unknown functions. Spots 13, 14, 15, 16 and 17 could not be identified because their peptide mass fingerprinting did not match any protein in the databases.

Cloning and identification of nag genes of Rhodococcus sp. strain TFB
A protein similar to the putative gentisate 1, 2-dioxygenase (NagI) encoded by the gene ro01866 from R. jostii RHA1 was identified in the proteomic analysis. Based on its DNA sequence, primers FwGDORH and RvGDORH (Table 3) were designed to amplify an intragenic region. A 489 bp polymerase chain reaction (PCR) product, obtained using total DNA from Rhodococcus sp. strain TFB as template, was cloned and sequenced. Its nucleotide sequence shows 95% identity to the corresponding region of the Rhodococcus sp. RHA1 ro01866 gene. The putative nagI PCR product was used to screen a TFB library, identifying a positive cosmid, named pMPO1109. The sequence of a 7192 bases fragment of pMPO1109 showed the presence of seven open reading frames (ORFs) (Fig. 3A), whose deduced amino acid sequences were screened against the peptide sequences obtained from the proteome analysis. Remarkably, peptides from spots 1, 2 and 3 perfectly matched sequences of the products of the ORFs named nagI, nagK and nagL respectively ( Table 2). Upstream of the nag genes, there is an ORF encoding a protein similar to IclR-like transcriptional regulators (orf4). The most similar characterized protein is Ncg12921, an IclR type regulator involved in gentisate degradation regulation in Corynebacterium (Shen et al., 2005). Divergently transcribed to the IclR regulator gene, there are three more ORFs that seem to encode a benzoate transporter (orf1), a salicylate monooxygenase (orf2) and a CorA-like putative magnesium transport protein (orf3). Because the naphthalene and tetralin-induced thn genes of TFB are located in the linear mega-plasmid pTFB1 of approximately 1.1 Mb (Tomás-Gallardo et al., 2009), we decided to analyse the location of nag genes using the 489 bp internal nagI fragment as the probe in Southern blotting experiments with TFB total DNA separated by pulse field electrophoresis gel. The two large plasmids that TFB bears were easily detected as previously reported, but none of them hybridized to the nagI probe (Fig. 3B). As described for R. jostii RHA1, we found that the nag genes are located on the TFB chromosome (McLeod et al., 2006).

Transcriptional characterization of thn, nag, orf1, orf2, orf3 and orf4 genes
In order to characterize the expression pattern of some of the identified genes and their transcriptional organization, retrotranscription (RT)-PCR and quantitative real-time PCR (qPCR) analyses were carried out.
For the determination of the transcriptional units of thn and nag genes, RT-PCR assays were carried out using RNA obtained from glucose-, naphthalene-or salicylategrown cells using the oligonucleotides described in Table 3. Intergenic regions corresponding to nag genes were detected in both naphthalene and salicylate growing cells (Fig. 3C). These results suggest that nagIKL genes are transcribed as an operon, while orf4 is constitutively transcribed as an independent and low expressed cistron. PCR products of the expected sizes for each intergenic region of the thn genes were detected in naphthalene-but not salicylate-growing cells (not shown), and indicated that the transcriptional units for the thn genes are the same as those described when For qPCR, total RNA isolated from TFB cells grown to mid-log phase with glucose, naphthalene, salicylate, and naphthalene or salicylate plus 2 h in the presence of glucose was converted to cDNA as described in Materials and Methods. As shown in Table 4, catabolic thnA1 and thnD genes, and the regulatory thnS gene are induced by naphthalene but not by salicylate. Besides, the presence of glucose provokes a partial catabolite repression of these thn genes (2.4-fold for thnS and 3.6-fold for thnA1 and thnD). On the contrary, the highest induction of the nagIKL genes is achieved in the presence of salicylate. A lower level of induction is also detected when naphthalene is the carbon source, but this induction may be due to the salicylate produced during naphthalene degradation. Intriguingly, nagIKL expression is not subjected to glucose-mediated catabolite repression at all when induced by salicylate, while it is fully repressed when induced by naphthalene. However, this difference may be explained by an indirect effect of the carbon catabolite repression of thn genes, which prevents conversion of naphthalene into the real inducer salicylate. Expression of orf1, orf2, orf3 and orf4 genes was not detected in any condition (not shown).

Inducer profile of TFB thn genes
Taking in consideration that the expression of thn genes is a prerequisite for naphthalene degradation in TFB, we performed a more complete analysis of the induction profile of TFB thn genes that was carried out by examining expression of a thnA1::gfp translational fusion in the presence of molecules that may share some structural or functional features with tetralin and/or naphthalene. As shown in Fig. 4, cyclohexane and benzene (monocyclic  molecules) induce thn gene expression at similar levels as tetralin. In addition, naphthalene and cis-decalin, molecules formed by two aromatic or aliphatic rings, respectively, sharing two carbon atoms, also induce thn genes, but in a less efficient mode (33% and 40% of the levels obtained with tetralin respectively). On the other hand, monocyclic (phthalate and salicylate) or bicyclic (5,6,7,8tetrahydro-2-naphtol and biphenyl) molecules containing substitutions do not induce thn genes at all.

Discussion
Using proteomic and genetic approaches, we show in this work that TFB naphthalene degradation ability may be the result of the combination of different catabolic pathways. Genes encoding the enzymes of these pathways are located in different replicons, have their own specific induction profile and respond differently to the presence of an alternative carbon source.
In TFB, naphthalene induces the tetralin degradation enzymes, as happens in Sphingomonas macrogolitabida strain TFA, an α-proteobacterium whose tetralin degradation pathway and thn genes have been thoroughly characterized. In TFA, naphthalene can be initially transformed (although it is not able to grow on naphthalene), and the specific activity of TFA-ThnC (the extradiol dioxygenase), which is 65% identical to TFB-ThnC, is 1.6-fold higher when using 1,2 dihydroxynaphthalene as the substrate than that using the corresponding tetralin intermediate (Andújar et al., 2000). Because Thn proteins are the most prominently induced in naphthalene-gown cells, we propose that the ThnA1A2A3A4 (the initial dioxygenase complex), ThnB (dehydrogenase) and ThnC enzymes, which catalyse the first three steps of the tetralin degradation pathway, may also convert naphthalene into 2-hydroxychromene-2-carboxylate (HCCA) in TFB (Fig. 5). In TFB, both, the thn structural and the thn regulatory genes (a two-component system formed by ThnST) are located at the pTFB1 mega-plasmid and are subjected to catabolite repression (Table 4). Recently, a cyclic AMP receptor protein (CRP)-like protein has been identified as involved in the repression of the thn genes in the presence of glucose (Tomás-Gallardo et al., 2012). In TFB, thn genes are also induced by compounds that can hardly be considered as substrates of the encoded enzymes (i.e. cis-decalin or cyclohexane, Fig. 4). On the other hand, TFB-thn genes are not induced by compounds like biphenyl, which is a good substrate of TFB-ThnD (Tomás-Gallardo et al., 2009). This proves the laxity of the regulatory system in recognizing putative inducers. A broad spectrum of inducers provokes the gratuitous induction of the pathway but can confer an advantage in the degradation of new molecules. Besides, in contrast with the regulation of naphthalene degradation in Pseudomonas, neither TFB-nor TFA-thn genes are induced by salicylate (López-Sánchez, 2009; Table 4). In pseudomonads, HCCA is converted to salicylate in three steps that involve an isomerase, a hydratase-aldolase (named NarC in Rhodococcus or NahE in Pseudomonas) that yields pyruvate and salicylaldehyde, and a salicylaldehyde dehydrogenase that converts the latter into salicylate. We were not able to identify these enzymes or the encoding genes either by proteomics or by PCR using degenerate primers (not shown). The isomerase activity has been shown not to be essential for growth on naphthalene (Royle et al., 1981;Eaton and Chapman, 1992) and has not been identified in other naphthalene degrading rhodococci. Therefore, it is possible that TFB lacked this enzyme, which could explain its high doubling time (35 h;Tomás-Gallardo et al., 2006) when growing with naphthalene as carbon and energy source. In both R. opacus R7 (Di Gennaro et al., 2010) and Rhodococcus sp. NCIMB12038 (Kulakov et al., 2005), the narC encoding the hydratase-aldolase is not located in the vicinity of the nar genes, and the same could be happening in TFB. However, the similarity of the TFB narC to other rhodococcal narC genes seems not to be high enough to allow its detection by PCR. A less likely possibility is that ThnE and ThnF, the hydratase and aldolase of the tetralin degradation pathway, could use naphthalene intermediates as substrates. Another possibility is that the naphthalene-induced spot identified as an esterase/lipase (spot 12, Fig. 2 and Table 1), which according to Pfam carries an alpha/beta hydrolase domain, was involved in this conversion by an unknown mechanism. Functional characterization of this enzyme would be required to confirm this hypothesis.
We have detected a salicylaldehyde dehydrogenase activity in naphthalene-grown cells that is undetectable in salicylate-grown cells. This indicates that the regulation of the upper pathway (conversion of naphthalene into salicylate) is different in rhodococci as compared with pseudomonads, in which salicylate induces both the upper and the lower pathway (conversion of salicylate into central intermediates) (Yen and Serdar, 1988). Besides, this activity in TFB is neither induced by tetralin nor shows glucose catabolite repression when induced by naphthalene, thus indicating that its encoding gene, which remains unidentified, does not belong to the thn genes regulon, which showed a partial but obvious catabolite repression by glucose (Table 4).
Two different pathways to convert salicylate into gentisate have been described in the literature, one via a salicylate-5-hydroxylase activity that involves two structural subunits that form a salicylate-5-hydroxylase linked to an electron transport chain (Suemori et al., 1993;Fuenmayor et al., 1998) and the other via a CoAdependent pathway as it has been proposed in Streptomyces sp strain WA46 (Ishiyama et al., 2004) and R. opacus R7 (Di Gennaro et al., 2010). In TFB, only a protein similar to a flavin-binding monooxygenase in R. jostii RHA1 has been identified as induced by naphthalene (spot 11, Fig. 2 and Table 1), but its role in the catabolism of this compound remains unclear as no salicylate-5-hydroxylase activity nor the CoA-dependent pathway have been detected in TFB cell extracts (not shown).
Gentisate degradation seems to take place in TFB as it has been described in other rhodococci (Liu et al., 2011) as the TFB nag genes are very similar (more than 90% identical) to other rhodococcal genes found in RHA1, R. opacus B4, or Rhodococcus sp. NCIMB 12038. Proteomic, gene expression and gentisate dioxygenase activity measurement indicate that TFB nag genes are specifically induced by salicylate and not by naphthalene, and their products are involved in gentisate metabolism. Upstream of nagI and orf4, a palindromic region (TTCTGCACAGCAGAA and TTCTGCTGTGCAGAA respectively), also found upstream of the R. opacus B4 and RHA1 gentisate dioxygenase genes (ROP_15440 and RHA1_ro01866 respectively), could be recognized by the IclR-like putative regulator of these genes (the orf4 product). This type of transcriptional regulators can act as activators or repressors and there is not a clear consensus sequence for their DNA-binding sites (Molina-Henares et al., 2005).
Although further experiments should be carried out to completely identify each enzyme and reaction involved in conversion of HCCA into salicylaldehyde and salicylate into gentisate, the data presented in this paper provide evidences supporting a biodegradation pathway for naphthalene in TFB (Fig. 5) in which at least three sets of enzymes, differently regulated, participate: the Thn enzymes induced by naphthalene or tetralin but not by salicylate, the salicylaldehyde dehydrogenase induced by naphthalene but not by salicylate or tetralin, and the Nag enzymes induced by salicylate but not by naphthalene or tetralin. Therefore, it represents a paradigmatic example of how a biodegradation pathway can be assembled using genes belonging to different regulons.

Experimental procedures
Bacterial strains, culture conditions, plasmids and primers Rhodococcus sp. strain TFB (Tomás-Gallardo et al., 2006) was grown on minimal medium (MM) (Dorn et al., 1974) at 30°C with tetralin supplied in gas phase, naphthalene supplied as solid phase, 12 mM of sodium salicylate or 12 mM of glucose as carbon and energy sources. Lysogeny broth (LB; Bertani, 2004) was used as rich media. Escherichia coli strains were grown in LB medium containing the appropriate antibiotics. Escherichia coli DH5α (Hanahan, 1983) was used for DNA manipulation.
For carbon catabolite repression assays, cells were grown on MM with glucose until optical density (OD)600 = 1, diluted 10-fold on MM with naphthalene plus glucose or salicylate plus glucose. After incubation for 12 h, cultures were diluted fivefold in the same medium and incubated until OD600 = 1 was reached.
Plasmids and primers used in this work are listed in Table 3.

Enzyme assays
Cell-free extracts were prepared by cell disruption using a French press. Exponentially growing cells (OD600 = 1) on MM supplemented with glucose, tetralin, naphthalene, salicylate, glucose plus naphthalene or glucose plus salicylate were harvested, washed once with phosphate buffer (17 mM Na2HPO4, 7.35 mM KH2PO4, pH 7.2), resuspended in 2 ml of phosphate buffer and disrupted. Cell extracts were centrifuged at 10,000 g for 20 min at 4°C, and the supernatants were collected for further assays.
Salicylaldehyde dehydrogenase activity was performed as described by Shamsuzzaman and Barnsley (1974) with several modifications. The reaction was determined from the rate of increase of absorbance at 340 nm because of the formation of NADH. The reaction started by adding 20 μl of fresh cell extract (50-250 μg soluble protein) to a 980 μl of reaction mixture containing 20 mM tetrasodium pyrophosphate buffer HCl pH 8.5, 0.1 mM salicylaldehyde and 10 mM NAD + . Production of NADH was calculated using an extinction coefficient of 3.84 mM −1 cm −1 . One enzyme unit (U) was defined as the amount of enzyme that catalyses the production of 1 μmol of NADH per minute.
Gentisate dioxygenase activity was performed as described by Adams and colleagues (2006) with several modifications. The reaction started with the addition of 20 μl of cell extract (50-250 μg soluble protein) to the reaction mixture containing 979 μl phosphate buffer (described earlier) and 1 μl sodium gentisate 0.25 M. The activity was determined by measuring the increase of absorbance at 330 nm. The activity was calculated using an extinction coefficient for maleylpyruvate of 10.8 mM −1 cm −1 . One enzyme unit (U) was defined as the amount of enzyme that catalyses the production of 1 μmol of maleylpyruvate per minute.
Protein sample preparation, fluorophore labelling and 2D electrophoresis analysis TFB cell free extracts, fluorophores labelling, 2D electrophoresis and protein identification were carried out as described in Tomás-Gallardo and colleagues (2006). For DIGE analysis, fluorescent gel images were obtained by scanning using a Typhoon TM 9400 Scanner and analysed with DeCyder Differential Analysis software version 5.01 (GE Healthcare, GE Healthcare Europe GmbH, Freiburg, Germany). Proteins selected for identification where those with an expression ratio level higher than 1.5 and a t-test lower than 0.05. For standard 2D analysis, silver-stained gel images were obtained by scanning using ImageScanner III software (GE Healthcare) and analysed with ImageMaster Platinum software version 7.0 (GE Healthcare). Theoretical molecular masses and isoelectric points of the proteins of interest were calculated using EXPASY tools. Identification of the selected spots was carried out by mass spectrometry at the Proteomic Unit of Centro Nacional de Investigaciones Cardiovasculares in Madrid, Spain.

Total RNA sample preparation, RT-PCR and qPCR assays
Total RNA of Rhodococcus sp. strain TFB was isolated following the method described in Tomás- Gallardo and colleagues (2006). PCR amplification of 16S rDNA and some thn genes with RNA samples as template was conducted to test for DNA contamination. DNA-free RNA samples (2 μg) were converted to cDNA using the High Capacity cDNA Reserve Transcription kit (Applied Biosystems, Alcobendas, Madrid, Spain) following the manufacturer's instructions. cDNA was cleaned using the QIAquick ® PCR Purification kit (Qiagen, Limburg, Netherlands). Equal amounts of cDNA were used as templates in PCR reactions with primers listed in Table 3. Semiquantitative PCR reactions were performed with 30 ng of cDNA using the puReTaq Ready-To-Go PCR beads (GE Healthcare) and consisted of 20 cycles of 30 s at 94°C, 30 s at 59°C and 45 s at 72°C. As loading control, primers f27 and r519 amplifiying 16S gene (Hugenholtz et al., 1998) were used with 20 pmol of cDNA. qPCR was performed using SYBR Green technology in an ABI Prism 7000 Sequence Detection System (Applied Biosystems) following the method described previously by Yuste and colleagues (2006). Samples were initially denatured by heating at 95°C for 10 min. A 40-cycle amplification and quantification programme was followed (95°C for 15 s and 60°C for 1 min, with a single fluorescence measurement per cycle according to manufacturers' recommendations). A final extension cycle (72°C, 1 min) was performed. cDNAs (10 ng) from three biological samples of each condition were amplified in quadruplicate in separate PCR reactions using 0.3 mM of each primer. All PCR products were between 50 and 100 bp in length. Standard curves were made using serial dilutions from 50 to 0.01 ng of TFB genomic DNA to quantify the relative abundance of transcripts. Primers used are described on Table 3. Value of quantification (Qt) is expressed as quantity of transcript relative to the standard curve.

Screening for genes encoding gentisate dioxygenases
A total TFB DNA library was screened by colony hybridization using a probe targeting gentisate dioxygenase gene. The probe was obtained by PCR using primers FwGDORH and RvGDORH (Table 3), whose design was based on the peptide sequence obtained from the proteomic analysis and the nucleotide sequence of ro01866 from R. jostii RHA1 that encodes a putative gentisate dioxygenase. PCR was carried out using 50 ng of total DNA as template and the puRe-Taq Ready-To-Go PCR Beads (GE Healthcare) under the following amplification conditions: 95°C for 10 min, followed by 30 cycles of 95°C for 30 s, 65°C for 30 s and 72°C for 30 s. The 489 bp fragment was cloned into EcoRV-cleaved pBluescript II SK+ (Stratagene, Santa Clara, California, USA) to obtain pMPO1102. This fragment was used to probe a TFB library by Southern analysis (Sambrook et al., 1989) to get cosmid pMPO1109. A 7192 bp fragment from pMPO1109 was commercially sequenced.
The obtained DNA sequence was compared with those in the databases using BLASTN 2.2.6 (Altschul et al., 1997) and submitted to GenBank under Accession Number JF754465.

Plasmid isolation, pulsed field gel electrophoresis, linearity determination and Southern blot hybridization
For detection of extra chromosomal DNA, plasmid isolation and pulse field gel electrophoresis assays were developed as described by König and colleagues (2004). Yeast Chromosome PFG marker (New England Biolabs, IZASA, Barcelona, Spain) was used as reference. After electrophoresis, DNA was transferred to Hybond-N+ nylon membranes (GE Healthcare) and subjected to Southern hybridization with a nagI probe.

Fluorescence measurements of gfp fusions
Fluorescence of exponentially grown TFB cells harbouring the plasmid pMPO633 was measured using a POLARstar Omega Fluorimeter (BMG Labtech, Ortenberg, Germany) with the EM520 emission filter, the 485BP1 excitation filter and an adjusted gain of 1000. The fluorescence measurement was normalized using the OD600 of each culture.