Trehalose phosphate synthases OtsA1 and OtsA2 of Rhodococcus opacus 1CP

Authors

  • Dirk Tischler,

    Corresponding author
    • Interdisciplinary Ecological Center, Environmental Microbiology Group, TU Bergakademie Freiberg, Freiberg, Germany
    Search for more papers by this author
  • Susanne Niescher,

    1. Interdisciplinary Ecological Center, Environmental Microbiology Group, TU Bergakademie Freiberg, Freiberg, Germany
    Current affiliation:
    1. Hegewald Medizinprodukte GmbH, Lichtenberg, Germany
    Search for more papers by this author
  • Stefan R. Kaschabek,

    1. Interdisciplinary Ecological Center, Environmental Microbiology Group, TU Bergakademie Freiberg, Freiberg, Germany
    Search for more papers by this author
  • Michael Schlömann

    1. Interdisciplinary Ecological Center, Environmental Microbiology Group, TU Bergakademie Freiberg, Freiberg, Germany
    Search for more papers by this author

Correspondence: Dirk Tischler, Interdisciplinary Ecological Center, Environmental Microbiology Group, TU Bergakademie Freiberg, Leipziger Street 29, 09599 Freiberg, Germany. Tel.: +49 3731 393739; fax: +49 3731 394015; e-mail: dirk-tischler@email.de

Abstract

Rhodococcus opacus 1CP produces trehalose dinocardiomycolates during growth on long-chained n-alkanes. Trehalose and trehalose-6-phosphate, which are synthesized via the OtsAB pathway, are probable intermediates in the biosynthesis of these biosurfactants. By molecular genetic screening for trehalose-6-phosphate synthases (TPSs and OtsAs), two chromosomal fragments of strain 1CP were obtained. Each contained an ORF whose amino acid sequence showed high similarity to TPSs. To prove the activity of the otsA1 and otsA2 gene product and to detect catalytic differences, both were expressed as His-tagged fusion proteins. Enzyme kinetics of the enriched proteins using several potential glucosyl acceptors showed an exclusive preference for glucose-6-phosphate. In contrast, both enzymes were shown to differ significantly from each other in their activity with different glucosyl nucleotides as glucosyl donors. OtsA1-His10 showed highest activity with ADP-glucose and UDP-glucose, whereas OtsA2-His10 preferred UDP-glucose. In addition, the wild-type OtsA activity of R. opacus 1CP was investigated and compared with recombinant enzymes. Results indicate that OstA2 mainly contributes to the trehalose pool of strain 1CP. OtsA1 seems to be involved in the overproduction of trehalose lipids. For the first time, a physiological role of two different OtsAs obtained of a single Rhodococcus strain was presumed.

Introduction

The nonreducing disaccharide trehalose occurs widespread throughout the microbial world, and it has a multifunctional physiological role for bacteria and yeasts (Argüelles, 2000; Elbein et al., 2003). Four possible biosynthetic routes for trehalose have been described so far: (1) the TreYZ pathway (Maruta et al., 1995); (2) the TreS route (Tsusaki et al., 1996, 1997); (3) the formation of trehalose by a glycosyltransferase (TreT) (Qu et al., 2004); and (4) the OtsAB pathway, which counts as the best characterized and most widely distributed one (Elbein et al., 2003). Active enzymes of OtsAB-, TreYZ-, and TreS-pathway were detected in myco- and corynebacteria (de Smet et al., 2000; Wolf et al., 2003) suggesting a high importance of trehalose and its metabolites for these organisms. Although, Woodruff et al. (2004) have reported on a functional redundancy of these three pathways for Mycobacterium smegmatis, studies revealed that in Mycobacterium tuberculosis, the OtsAB pathway is most important for trehalose biosynthesis (Murphy et al., 2005).

In the OtsAB pathway, a trehalose-6-phosphate synthase (TPS, OtsA) catalyzes the formation of trehalose-6-phosphate from a glucosyl nucleotide and glucose-6-phosphate. In a second step, a trehalose-6-phosphate phosphatase (OtsB) catalyzes the dephosphorylation to yield trehalose. In addition to the delivery of trehalose-6-phosphate for trehalose biosynthesis, TPS may be involved in the biosynthesis of trehalose-containing glycolipids, which are important cell wall constituents in members of the mycolata group (Kretschmer & Wagner, 1983; Asselineau & Lanéelle, 1998; Sutcliffe, 1998; Shimakata & Minatogawa, 2000; Minnikin et al., 2002).

Trehalose lipids are part of the cell wall of Rhodococcus species. Several members of that genus overproduce surface-active trehalose lipids during growth on medium- to long-chained n-alkanes (Rapp et al., 1979; Kim et al., 1990; Espuny et al., 1996; Ueda et al., 2001; Rapp & Gabriel-Jürgens, 2003). These glycolipids are likely produced via the OtsAB route and characterized by high surface and interfacial activity. They show promising stability under changing pH and temperature conditions and are biodegradable (Hommel, 1990; Christofi & Ivshina, 2002). Rhodococcus opacus 1CP was previously shown to produce a novel variant of trehalose dimycolate during growth on long-chained n-alkanes (Niescher et al., 2006) and therefore has been further investigated.

Materials and methods

Strains, plasmids, and culture conditions

Rhodococcus opacus 1CP (Gorlatov et al., 1989) was grown aerobically at 30 °C under shaking on mineral medium (Dorn et al., 1974) containing 20 mM glucose. Those cultures were used to inoculate a 4-L fermentation broth of similar composition in a bioreactor with either glucose (final concentration 70 mM) or n-tetradecane (final concentration 150 mM) as sole carbon source. The aerobic cultivation was similarly performed; pH was regulated to 7.0 during fermentation by the addition of NaOH (1 M stock). Fermentation was stopped after a certain biomass yield was reached (glucose: 100 g wet biomass; n-tetradecane: 375 g wet biomass). Cells were harvested by centrifugation (5000 g, 4 °C, 30 min), washed once with 27 mM phosphate buffer (pH 7.2), and stored in aliquots at −80 °C.

The Escherichia coli strains DH5α (GIBCO-BRL) and BL21 CodonPlus(DE3)-RP (Stratagene) were used for cloning and expression studies, respectively. In general, the strains were grown aerobically under constant shaking at 37 °C on lysogeny broth (LB) medium (Sambrook et al., 2001). For selection purposes, ampicillin was added to a final concentration of 100 μg mL−1. For expression experiments, chloramphenicol was added additionally to a final concentration of 50 μg mL−1. Plasmids used and obtained in this study are listed in the Supporting Information section.

DNA preparation and general in vitro manipulation

Genomic DNA from Rhodococcus was prepared as described earlier (Eulberg et al., 1997). Standard protocols were used for restriction digestion of DNA (genomic and plasmid), agarose gel electrophoresis, plasmid DNA extraction and purification, cloning procedures, preparation of competent cells, and heat-shock transformation, as well as blue/white screenings (Inoue et al., 1990; Marchuk et al., 1991; Sambrook et al., 2001).

PCR and hybridization procedures

For the identification of otsA genes, degenerate oligonucleotides (otsA_fw/otsA_rev) were designed based on an alignment of different TPSs. Oligonucleotides used for sequence completion were designed according to existing sequence information (otsA_fwII/otsA_revII) or by comparison with genomic data available for Rhodococcus jostii RHA1 (ABG91926) (otsA2_fwIRHA1/otsA2_revIIRHA1). Oligonucleotides used in this study are listed in the Supporting Information section. Regular PCR protocols were performed with 0.5 μg of genomic template DNA or appropriate amounts of plasmid DNA and respective primers applying Taq DNA polymerase (Fermentas).

For hybridization experiments, DNA was initially digested with the respective restriction enzymes and separated by gel electrophoresis. DNA labeling, dot blot, colony blot, and Southern blot hybridizations, and detection procedures were performed according to the DIG Application Manual for Filter Hybridisation (Roche) using a DIG DNA Labelling and Detection kit Nonradioactive (Eulberg et al., 1997). For dot blot and Southern blot hybridization, positively charged nylon membranes were used. Colony blot hybridization was performed on nitrocellulose membranes. In general, prehybridization was carried out for 2 h at 68 °C followed by an overnight hybridization at the same temperature.

Construction of expression plasmids

The genes otsA1 and otsA2 were amplified from pSN87 and pSN29 LIV, respectively, using the primers 1CPfw_otsA1pet/1CPrev_otsA1pet or 1CPotsA2pet_fw2/1CPotsA2pet_rev3. Products obtained were purified by agarose gel electrophoresis, ligated into a T-tailed EcoRV site of pBluescript II SK(+), and propagated in E. coli DH5α. Plasmids were isolated from these clones and subjected to DNA sequencing with T7- or T3-primer in order to confirm the correctness of insert sequences.

The genes otsA1 and otsA2 were obtained from restriction digests with BamHI/NdeI or NdeI/BclI and ligated into a BamHI/NdeI-digested and purified plasmid pET16bP yielding pSN87E1 and pSN29E2a, respectively. Genes then could be expressed with an N-terminal His10-tag.

Gene expression

Expression clones were grown at 30 °C in LB media with antibiotics. At an optical density of 0.6 (546 nm), cultures were induced with 0.1 mM isopropylthio-β-d-galactoside (IPTG) and further incubated (between 19 and 24 h) at 16 °C. Controls were cultivated without IPTG. Cells were harvested at 4 °C by centrifugation (5000 g, 30 min). Pellets were washed twice with ice-cold Tris buffer (25 mM Tris–HCl, pH 7.5) and, after final centrifugation, suspended in the same buffer. Cell suspension aliquots were stored at −80 °C.

Determination of enzyme activities

TPS activity was determined in duplicate using a colorimetric detection method (anthrone method) for trehalose phosphate (Lapp et al., 1971; Pan et al., 1996, 2002) and using the MBTH method (3-methyl-2-benzothiazolinone hydrazone method) suitable for the detection of reducing sugars (Anthon & Barrett, 2002). Calibration curves were performed with respective sugars. Typically, assay mixtures of 25 μL contained 250 nmol UDP-glucose, 250 nmol glucose-6-phosphate, 250 nmol MgCl2, 12.5 pmol heparin (Roth), 1.25 μmol Tris–HCl (pH 8.0), and an appropriate amount of enzyme. Incubation was carried out for 15 min at 37 °C prior to sugar determination.

The pH dependence of TPSs was determined without heparin. The enzyme was incubated in an appropriate buffer (50 mM Bis–Tris propane/HCl, pH 9.0–6.3; or 50 mM succinate buffer, pH 6.0–4.0) for 20 min prior adding substrates and thereby initiating the reaction.

Some assay mixtures were incubated with 1 μg μL−1 of clarified cell extract (6.9 mg protein mL−1), which was freshly prepared from glucose-grown cells of strain 1CP.

Substrate specificity was verified by applying other glucosyl donors (ADP-, GDP-, or TDP-glucose) or alternative glucosyl acceptors (fructose-6-phosphate and glucosamine-6-phosphate; Sigma).

Protein determination

Protein concentrations were determined with the Bio-Rad protein quantification kit using bovine serum albumin as a standard (Bradford, 1976).

Discontinuous sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) was performed according to Laemmli (1970), and gels were stained with Coomassie Brilliant Blue R-250 and/or silver-stained (Merril et al., 1981).

Enrichment of recombinant TPSs

DNaseI was added to thawed cell suspensions. Cells were disrupted by two passages through a French pressure cell (SIM-AMINCO Spectronic Instruments). From the cell lysates, cell debris and nondisrupted cells were removed by centrifugation at 4 °C (100 000 g, 60 min).

His10-tagged TPSs were purified on Ni-sepharose columns (column volume = CV: 1 mL, HiTrap Chelating HP or HiTrap-FF; GE Healthcare) using an ÄKTA FPLC system (GE Healthcare), equipped with an UV monitor (254 nm).

Crude extracts were spiked with imidazole (10 mM). The column was pre-equilibrated with a mixture of 98% buffer A (50 mM NaH2PO4, 300 mM NaCl, pH 8.0) and 2% buffer B (buffer A plus 0.5 M imidazole). Proteins were eluted by increasing the amount of buffer B resulting in a linear gradient from 0.01 to 0.5 M imidazole over 8 CV. One millilitre of fractions were collected and assayed for TPS activity. 1 mM dithiothreitol (DTT) was immediately added to active fractions. Most active fractions were applied to a Sephadex G-25 column (GE Healthcare) equilibrated with Tris buffer (50 mM Tris–HCl, pH 7.5, containing 1 mM DTT). Fractions were collected at a flow rate of 2 mL min−1 and assayed for TPS activity. Most active fractions were pooled (c. 3 mL in total).

TPS purification with HiTrap-FF was performed similarly, but with different buffers: buffer A1 (10 mM Tris–HCl, pH 7.5, 500 mM NaCl, 2 mM DTT) and buffer B1 (buffer A1 plus 0.5 M imidazole). Most active fractions were pooled and subjected to ultrafiltration (Vivaspin 500, Vivascience Sartorius; centrifugation at 4 °C, 10 000 g, 30 min). Concentrated protein was eluted from the membrane with 750 μL buffer (10 mM Tris–HCl, pH 7.5, 2 mM DTT).

To achieve the cleavage of the His10-tag from recombinant enzymes, the proteins were digested with Factor Xa protease (2 U μL−1; Novagen). Digestion was performed at 20 °C for 3 h in cleavage buffer (Novagen) with appropriate amounts of TPS and Xa protease.

Enrichment of wild-type OtsA

1CP biomass was suspended in an equal volume of 10 mM Tris–HCl (pH 7.5, 2 mM DTT, 0.8 U mL−1 DNaseI), and cells were disrupted by three passages trough a French press. Cell debris and whole cells were removed by centrifugation (100 000 g, 4 °C, 1 h).

A MonoQ column (CV: 1 mL, flow rate of 2 mL min−1) was used to enrich OtsA from cell-free crude extract. After column equilibration (25 mM Tris–HCl, pH 7.5, 2 mM DTT), protein samples of up to 150 mg total protein were loaded, and unbound protein was removed by washing with 6 CV of the same buffer. Bound proteins were eluted by a linear increasing two-step gradient of 0–500 mM NaCl and 500–1000 mM NaCl over 25 and 2 CV, respectively. Fractions were collected and assayed for OtsA activity.

Attempts to further enrich OtsA from strain 1CP by additional chromatographic steps resulted in an almost complete loss of active enzyme (not shown).

DNA sequence analysis

Sequencing and data analyses were performed as described elsewhere (Altschul et al., 1997; Tischler et al., 2009). DNA sequences obtained were submitted to GenBank database (DQ469313 and DQ469314).

Results

Identification of otsA1 and otsA2

A PCR with genomic DNA of R. opacus 1CP using degenerated primers yielded a 762-bp DNA fragment. The product obtained showed high similarity to TPSs of R. jostii RHA1 (ABG96494) and of Nocardia farcinica IFM 10152 (BAD55392) and thus clearly verified the amplicon as a putative OtsA-encoding gene fragment. The labeled fragment (otsA_a1) was applied as a probe during hybridization experiments against 1CP DNA. Colorimetric detection revealed two strong hybridization signals for the BamHI-digested 1CP DNA (Supporting Information). Cloning of both fragments led to the clones, pSN87 and pSN29, carrying a 2.6-kb and a 3.7-kb insert, respectively. Sequence analysis of both clones indicated the presence of two distinct otsA genes, of which the one of pSN29 corresponded to the initially obtained PCR product. By comparing their sequences on protein level, the genes showed 69% identity over 468 amino acids.

For repeated hybridization experiments, two new, more specific, probes were designed. Probe otsA_a2 was generated from the 762-bp PCR fragment, whereas probe otsA_b was created using the insert of pSN87. Two complementary signal patterns were obtained by applying the probes against BamHI-digested genomic 1CP DNA confirming above results (Supporting Information). In the following, the otsA gene on pSN87 will be designated as otsA1, the one harbored by pSN29 as otsA2.

After generation of various subclones, DNA sequence analysis of the main clones, pSN87, pSN29 and pSN29 LIV, revealed the allocation of several open reading frames (ORFs) (Fig. 1). Five ORFs were identified in plasmid pSN87 carrying the otsA1 gene. The assembly of these ORFs clearly corresponds to the arrangement of annotated genes found in the close vicinity of the otsA gene of R. jostii RHA1 (ABG91926), which was highly similar to otsA1 (ORF4). Sequencing of pSN29 revealed that the start region of otsA2 was not located on the plasmid. PCR with otsA2-specific primers (otsA2_fwIRHA1/otsA2_revIIRHA1; designed using the sequence available for strain RHA1, McLeod et al., 2006) yielded a fragment of expected size. Cloning of the 1.83-kb amplicon led to clone pSN29 LIV, which carried the complete otsA2 gene. Sequence analysis of pSN29 and pSN29 LIV revealed the presence of six ORFs. The arrangement of annotated ORFs around otsA2 (ORF2) was similar to the corresponding otsA gene of R. jostii RHA1 (ABG96494). Interestingly, some encoded proteins are supposed to be involved in cell envelope biogenesis according to their in silico annotation.

Figure 1.

Schematic view of plasmids pSN87, pSN29 1 and pSN29 LIV. Plasmid pSN87 harbors otsA1 of Rhodococcus opacus 1CP in a 2590-bp BamHI-insert (a). Plasmids pSN29 and pSN29 LIV together carry otsA2 in 4247-bp of R. opacus 1CP DNA (b). Restriction maps of plasmids show respective DNA-fragments. Thin black bars indicate the DNA fragments used as hybridization probes. Inserts of subclones prepared for sequencing purposes (grey bars) are shown below the restriction maps. Organization of open reading frames (ORFs) localized on the inserts are presented as arrows indicating the direction of transcription. The loci of otsA1 and otsA2 genes appear in black.

Phylogenetic analysis

Because most bacteria encode solely a single TPS, the identification of two different TPS genes in a Rhodococcus strain was unexpected and revealed questions on the phylogeny of respective proteins. By means of blastp (Altschul et al., 1997), genomes of mycolata strains (Sutcliffe, 1998) were screened for TPSs. Interestingly, the screening showed that all genomes of Rhodococcus strains sequenced so far encode two TPSs. Always one of those is homologous to OtsA1, while the other equals OtsA2 of strain 1CP. By expanding the comparisons to further bacteria, plants, and fungi, TPS proteins were found to be widespread among organisms, but usually only a single TPS per genome was found.

A phylogenetic tree was calculated on the basis of an alignment of TPSs of bacteria, plants, and fungi (Fig. 2). The distance tree clearly allows differentiating between TPSs of Actinobacteria, Proteobacteria, and eukarya.

Figure 2.

Dendrogram of (putative) trehalose-6-phosphate synthases. The distance tree was created by means of the program mega version 5 (Tamura et al., 2011) using ClustalW to generate the alignment and the Maximum–Likelihood method for phylogenetic tree construction. Default settings were applied throughout the procedure, while the calculation of 1000 bootstraps (shown in percentage) was performed. Representatives reported with activity and/or structural data were marked with a black triangle. In addition to bacteria several functionally characterized TPSs of eukarya [fungi: Aspergillus niger ATTC 11414 (AAB05869, AAA21861), Beauveria bassiana ARSEF2860 (ADV16369), Candida albicans SC5314 (CAA69223), Ogataea angusta (CAB38058), Saccharomyces cerevisiae C836 (CAA43580), Saccharomyces cerevisiae YNN 295 (CAA48296), Saccharomycopsis fibuligera sdu (AEV53962), and Zygosaccharomyces rouxii KACC30010 (AAK69413) and plantae: Arabidopsis thaliana (Q9SYM4), Camellia sinensis (AFJ06909), Nicotiana tabacum (BAI99252), Oryza sativa (AEB53177), and Selaginella pulvinata (ADD71944)] were included to obtain a more robust distance tree. The corresponding branches were compressed.

Characterization of recombinant OtsA1 and OtsA2

pSN87E1 and pSN29E2a harboring respective otsA genes were used for expression and allowed the production of recombinant proteins after induction with IPTG. Soluble fractions of cell disruption were used for isolation of TPS on Ni-sepharose columns, which was controlled by loading samples from the procedure on an SDS-PAGE (Fig. 3).

Figure 3.

TPS purification analysis by SDS-PAGE. SDS-PAGE of cell-free crude extracts, eluate from Ni-Sepharose chromatography, and eluate from gel filtration containing recombinant (a) OtsA1-His10 and (c) OtsA2-His10, as well as (b) the substrate and product of a Factor Xa-mediated proteolytic cleavage of an active OtsA1-His10 fraction. Black arrows indicate the recombinant enzyme. Molecular weights of the marker (Fermentas) are indicated. The lane of Factor Xa-treated sample shows two bands, corresponding to non-tagged and His-tagged OtsA1. Furthermore, a 30-kDa band fits to the molecular weight of the protease Factor Xa.

Isolation of recombinant OtsA1-His10 after gel filtration led to considerable, but incomplete purification of active enzyme (Fig. 3a). Theoretical molecular sizes had been calculated for OtsA1 and OtsA2 to 53.4 and 55.3 kDa, respectively. Whereas purified OtsA2-His10 yielded a single band of c. 56 kDa on SDS-PAGE gels (Fig. 3c), OtsA1-His10 showed two significant bands, of which the lower one fits to the expected size (Fig. 3a). SDS-PAGE of the Xa protease-treated fusion protein yielded two bands, which proved a successful cleavage. The new band corresponds in size to that of an OtsA1 without the His10-tag. The cleaved protein preparation showed very low activity. On the contrary, activities of OtsA2 preparations after treatment with Factor Xa protease were similar to the His10-tagged enzyme, but seemed to be strongly dependent on heparin. In the presence of the polyanion, activity of the His10-free enzyme was increased more than threefold (Table 1). However, for so far unexplained reasons, it could not be traced by SDS-PAGE whether the treatment of OtsA2-His10 with Factor Xa protease succeeded in a complete removal of the His10 tag from the protein (data not shown).

Table 1. Comparison of recombinant and wild-type TPSs
Glucosyl donorRelative TPS activity −/+ the presence of heparin [specific activity (U mg−1) obtained with UDP-glucose and glucose-6-phosphate was set to 100%]
Recombinant TPSsWild-type TPSs from biomass grown on the substrate
OtsA1-His10OtsA2-His10OtsA2aGlucosen-tetradecane
+−/c.e.++++
  1. a

    By means of Factor Xa protease, the His-tag of OtsA2-His10 was removed prior characterization.

  2. −/c.e., the assays contained no heparin, but freshly prepared clarified crude extract of glucose-grown strain 1CP; n.d., not determined.

ADP-glucosen.d.100535660n.d.n.d.n.d.58n.d.47
GDP-glucosen.d.25727380n.d.n.d.n.d.74n.d.60
TDP-glucosen.d.50252631n.d.n.d.n.d.25n.d.37
UDP-glucose100100 (0.17)7067100 (4.13)27100 (3.88)5100 (0.09)22100 (0.06)

Substrate preferences of both His10-tagged glycosyltransferases were determined by the ability to utilize ADP-, GDP-, UDP- or TDP-glucose as glucosyl donors and glucose-6-phosphate as glucosyl acceptor. Although both recombinant proteins used all of the tested glucosyl donors, OtsA1-His10 and OtsA2-His10 showed clear differences in their specific activities and substrate preferences (Table 1). Using fructose-6-phosphate or glucosamine-6-phosphate as alternative glucosyl acceptors, both enzymes showed no activity, no matter which of the glucosyl donors were used.

The polyanion heparin in assays seemed to have no effect on the activity of recombinant OtsA1-His10, whereas the activity of OtsA2-His10 was significantly increased, most of all when UDP-glucose was used. Tests with OtsA2-His10 using UDP-glucose as substrate resulted in maximum enzyme activity at pH 7 and temperatures between 35 and 40 °C. A drastic decrease in enzyme activity was observed at pH values below 6 and above 8 and temperatures below 20 and above 50 °C. The low stability of OtsA1-His10 circumvented the determination of respective activity values.

Time-dependent stability differed drastically for both proteins. While purified OtsA2-His10 kept a stable activity over several days when stored at 4 °C, OtsA1-His10 almost lost its entire activity overnight under identical storage conditions. Storage of the OtsA2-His10 gel filtration pool at −20 °C for 7 days yielded 76% of its initial activity, whereas OtsA1-His10 had lost its activity after thawing.

Wild-type TPS activity of strain 1CP

Crude extracts of glucose- and n-tetradecane-grown strain 1CP biomass were used for a chromatographic expression analysis. A different expression pattern of wild-type OtsAs according to possible roles in trehalose lipid formation was expected. But, the low stability of OtsAs allowed solely a single chromatographic step to evaluate the made assumption. To obtain comparable data, crude extracts of both types were loaded consecutively onto a MonoQ column, and identical conditions were applied for respective chromatography runs.

The specific activities 2.4 mU mg−1 (glucose) and 0.7 mU mg−1 (n-tetradecane) of crude extracts indicated a divergent expression of TPS genes during growth on different carbon sources. Surprisingly, in both cases, the TPS-active fractions eluted as a single peak around 150 mM NaCl from MonoQ runs. Most active fractions obtained were pooled and applied for characterization (Table 1). Interestingly, the specific activity of the enriched TPS from glucose (4 mU mg−1)- and n-tetradecane-grown biomass (12 mU mg−1) differed, and also, the addition of 2.5 μM heparin to the assay increased differentially the specific activity. The value for glucose-grown biomass changed by factor 22 to about 87 mU mg−1, whereas the value for n-tetradecane-grown biomass only changed by factor 5 to about 55 mU mg−1.

Discussion

Two TPS genes otsA1 and otsA2 have been identified in strain 1CP. OtsA1-His10 and OtsA2-His10 showed broad specificity toward the glucosyl donor, and all tested glucosyl nucleotides were utilized when glucose-6-phosphate served as an acceptor. Both enzymes showed no detectable activity when alternative glucosyl acceptors were applied. OtsA1-His10 showed highest activity with ADP-glucose and UDP-glucose, whereas OtsA2-His10 was most active with UDP-glucose and a bit less with GDP-glucose. In comparison, activities of a recombinant TPS of M. tuberculosis and of a TPS purified from M. smegmatis were highest with ADP- and GDP-glucose using glucose-6-phosphate as acceptor (Lapp et al., 1971; Pan et al., 1996, 2002). Mycobacterial and both rhodococcal recombinant enzymes hardly converted TDP-glucose. The purified wild-type TPS of M. smegmatis was also less stimulated by that pyrimidine sugar nucleotide. In contrast, a recombinant TPS obtained from Propionibacterium freudenreichii showed highest activity when TDP-glucose was applied (Cardoso et al., 2007).

Activity and substrate specificity of TPSs were reported to be affected by polyanions, such as heparin (Lapp et al., 1971; Pan et al., 1996), and by components of cell extracts (Cardoso et al., 2007). Polyanions were found to stimulate enzyme activity when pyrimidine sugar nucleotides were used as substrates. Also, OtsA2-His10 activity was significantly increased in the presence of heparin (about 43%), especially, with UDP-glucose as substrate. The activity of OtsA2-His10 was only slightly stimulated by the polyanion when other glucosyl nucleotides were supplied as substrates. Heparin seemed to have no effect on OtsA1-His10, no matter if glucosyl donors with pyrimidine or purine bases were utilized. Interestingly, in case of P. freudenreichii, the addition of crude extract to purified TPS changed the substrate specificity drastically. Thus, without crude extract, various glucosyl nucleotides were converted, and with crude extract, TPS became specific for ADP-glucose. However, OtsA2-His10 showed no such change in substrate preferences after incubation with cell extract.

Rhodococci, mycobacteria, propionibacteria, and streptomycetes belong to the same taxonomic class. Still, their TPSs show clear differences in substrate preferences, possibly attributed to an early separated evolution. For the recombinant TPSs of strain 1CP, the observed differences in substrate specificities may also reflect an evolutionary differentiation between the two enzymes coinciding with a functional separation.

The overproduction of trehalose lipids by R. opacus 1CP depends on the carbon source supplied (Niescher et al., 2006). Hence, a differential expression of otsA genes dependent on the trehalose demand was expected. Indeed, TPS activity was determined from glucose- and n-tetradecane-grown 1CP biomass. Substrate specificity of wild-type OtsA was determined in TPS pools obtained from analytical anion-exchange chromatography runs. Obtained activity values were compared to the data available from the recombinant OtsAs (Table 1). The patterns of substrate preference for the wild-type OtsA obtained from glucose- and n-tetradecane-grown biomass were highly similar and most comparable to those of recombinant OtsA2-His10. That clearly indicates a major role of OtsA2 of strain 1CP in trehalose metabolism and trehalose lipid formation for the cell-wall-linked trehalose dicorynomycolates. An involvement of OtsA2 in trehalose lipid formation for the cell wall is reinforced by the presence of ORFs adjacent to the otsA2 gene with putative function in cell envelope biogenesis.

However, on first sight, it seemed that only a single OtsA was enriched from crude extract, because only one peak with TPS activity was eluted at the same salt concentration from the MonoQ column. But, the pools had different specific activities, and even more striking was the difference in heparin activation. Mainly, the pool obtained from glucose-grown biomass was affected by the presence of the polyanion, whereas only little effects were determined from the pool obtained from n-tetradecane-grown biomass. Because it was demonstrated that OtsA1-His10 had no dependence on heparin, this possibly demonstrates the existence of OtsA1. Taking also the much lower specific activity and stability of recombinant OtsA1 into account, it can be reasoned that from n-alkane-degrading biomass, both OtsA1 and OtsA2 were enriched, leading to the assumption that OtsA1 can be involved in the overproduction of trehalose lipids.

So far, Actinobacteria, which were investigated for trehalose formation as well as for involved TPS activity, were found to harbor solely one otsA gene. Thus, this was the first time that two genes from the same strain were investigated and found to be active. In case of R. opacus 1CP, OtsA2 was found to be the key player in trehalose-6-phosphate synthesis independent of the carbon source supplied. Besides the biochemical data, that conclusion is supported by the phylogenetic analysis made herein. OtsA2 forms a distinct branch with homologous enzymes of other rhodococci and mycobacteria in a phylogenetic tree (Fig. 2). Mycobacteria harbor only one TPS, which was shown to be part of the major supply line for trehalose in mycobacteria (Murphy et al., 2005). OtsA2 homologous enzymes of rhodococci seem to be involved in general trehalose metabolism and normal trehalose lipid biosynthesis.

OtsA1 of strain 1CP and only homologous enzymes of rhodococci form another branch in the phylogenetic tree. Still the recombinant OtsA1 of strain 1CP was found to be active and showed different properties compared with OtsA2. But, the physiological function of OtsA1 and related enzymes of rhodococci can only be presumed to be part of the biosynthetic route, which leads to the overproduction of trehalose lipids during growth on n-alkanes.

Acknowledgements

The work was supported by the Deutsche Bundesstiftung Umwelt and by the Saxon State Ministry of Environment and Agriculture (SMUL no. 13.8811.61/182). The authors declare having no conflict of interest while publishing the manuscript.

Ancillary