Biochemical and spectroscopic characterization of purified Latex Clearing Protein (Lcp) from newly isolated rubber degrading Rhodococcus rhodochrous strain RPK1 reveals novel properties of Lcp

Background Biodegradation of rubber (polyisoprene) is initiated by oxidative cleavage of the polyisoprene backbone and is performed either by an extracellular rubber oxygenase (RoxA) from Gram-negative rubber degrading bacteria or by a latex clearing protein (Lcp) secreted by Gram-positive rubber degrading bacteria. Only little is known on the biochemistry of polyisoprene cleavage by Lcp and on the types and functions of the involved cofactors. Results A rubber-degrading bacterium was isolated from the effluent of a rubber-processing factory and was taxonomically identified as a Rhodococcus rhodochrous species. A gene of R. rhodochrous RPK1 that coded for a polyisoprene-cleaving latex clearing protein (lcpRr) was identified, cloned, expressed in Escherichia coli and purified. Purified LcpRr had a specific activity of 3.1 U/mg at 30 °C and degraded poly(1,4-cis-isoprene) to a mixture of oligoisoprene molecules with terminal keto and aldehyde groups. The pH optimum of LcpRr was higher (pH 8) than for other rubber-cleaving enzymes (≈ pH 7). UVvis spectroscopic analysis of LcpRr revealed a cytochrome-specific absorption spectrum with an additional feature at long wavelengths that has not been observed for any other rubber-cleaving enzyme. The presence of one b-type haem in LcpRr as a co-factor was confirmed by (i) metal analysis, (ii) solvent extraction, (iii) bipyridyl assay and (iv) detection of haem-b specific m/z values via mass-spectrometry. Conclusions Our data point to substantial differences in the active sites of Lcp proteins obtained from different rubber degrading bacteria. Electronic supplementary material The online version of this article (doi:10.1186/s12866-016-0703-x) contains supplementary material, which is available to authorized users.


Background
Natural rubber is an important biopolymer that has been produced for more than a century by cultivating the rubber tree (Hevea brasiliensis). Natural rubber obtained by tapping of the rubber trees is used for countless applications, for example for the production of tires, sealings, latex gloves and many, many other items. The main component of rubber latex is the hydrocarbon poly(cis-1,4-isoprene). Chemosynthetic rubber is also produced at a scale that is almost comparable to that of the natural compound.
Despite the economic importance of rubber and the enormous amounts of rubber waste materials that are permanently released into the environment, complete degradation in nature is rarely detected and wastes continue to accumulate. Knowledge of the reasons for this is limited. In fact, application is made of this extremely slow natural degradation for example in the use of rubber tyres to provide attachment sites for creating artificial coral reefs. However, microorganisms that can attack rubber have been detected in many ecosystems in which the physical parameters (temperature, pH, salinity) are moderate [1][2][3][4][5][6][7]. It is also well known that the initial microbial attack on rubber depends on the ability to produce and secrete rubber-cleaving enzymes into the environment. Only two types of rubber-cleaving enzymes are known. One is the rubber oxygenase RoxA that was first isolated from Xanthomonas sp. 35Y [8,9] and so far has been found only in Gram-negative bacteria [10]. RoxA of Xanthomonas sp. 35Y is a c-type dihaem dioxygenase and cleaves poly(cis-1,4-isoprene) into a C 15 compound with a terminal keto and aldehyde group (12oxo-4,8-dimethyl-trideca-4,8-diene-1-al, ODTD) as the main product [11][12][13]. The other rubber cleaving enzyme is a protein designated as latex clearing protein (Lcp) [1]. It shares no significant sequence homology with RoxA, with cytochrome c peroxidases or with dihaeme 7,10-diol synthases [14] and is present in Gram-positive rubber degrading bacteria such as Streptomyces sp. K30 [1] and other Actinobacteria. G. polyisoprenivorans VH2 and Streptomyces sp. K30, two well-studied Gram-positive rubber degraders, oxidatively cleave poly(cis-1,4-isoprene) to products of different sizes but with the same keto and aldehyde end groups as in RoxA-generated ODTD [15][16][17]. There have been different reports published for the cofactor and metal-contents of the Lcps from Streptomyces sp. K30 and of G. polyisoprenivorans VH2 [15,17,18], and at present there are currently only two biochemically characterized Lcp proteins.
In this study, we used a waste pond at a rubberprocessing factory in Thailand as a natural enrichment environment for rubber-degrading microorganisms and a source for the isolation of new rubber degrading strains. Taxonomic analysis revealed that one isolated strain was a member of the genus Rhodococcus, a taxon that had not been previously identified as having the ability to utilise rubber as a sole source of carbon and energy but that is well known for its members to have a high potential for the biodegradation of recalcitrant compounds [19]. Biochemical and biophysical characterization of the purified recombinant Lcp protein of Rhodococcus rhodochrous strain RPK1 revealed some unexpected properties not previously described for any other rubber-degrading enzyme in addition to properties shared with the two other characterized Lcp proteins.

Results and discussion
Taxonomic identification of isolate RPK1 Isolate RPK1 had a high rubber-degrading activity compared to other rubber degraders in liquid culture, as revealed by pronounced disintegration of rubber pieces (Fig. 1a). However, isolate RPK1 did not form clearing zones on an opaque polyisoprene latex mineral salts agar while known clear zone formers such as Xanthomonas sp. 35Y [8] or Streptomyces coelicolor strain 1A [3] formed large clearing zones. Isolate RPK1 developed colonies with an intense red colour upon growth and prolonged incubation on NB agar (Fig. 1b). Microscopic examination revealed non-motile cells. Depending on the growth phase the cells were coccoid (cells from late stationary phase), rod-shaped (cells from early and late log phase) or long rods (up to 1 ×5 μm), partially branched and star-like in exponentially growing cultures ( Fig. 1c-e). Isolate RPK1 was catalase positive and Grampositive. It grew well at 43°C but no colonies developed at 45°C. Strain RPK1 tolerated the presence of 3 % NaCl (in NB). It accumulated storage compounds that were stainable by Nile red (polyhydroxyalkanoates or triacylglycerols) and strain RPK1 synthesised polyphosphate granules as shown by staining with DAPI (4' ,6-diamidine-2-phenylindole) and the use of DAPIpolyphosphate-specific emission filters in fluorescence microscopy ( Fig. 1e, f ). Isolate RPK1 utilised complex media (NB, LB medium) and grew with mineral salts media containing D-mannitol, fructose, acetate, benzoate or octane as a single carbon source. Glucose, sucrose, gluconate, pentane, petroleum or pyridine (excluding Rhodococcus pyridinivorans) were not used for growth. Polymers such as polyhydroxybutyrate (PHB), casein or starch were also not utilised by RPK1. These characteristics, in combination with the red colour of the colonies and the variable morphology of the cells indicated that the isolate RPK1 could be a member of the genus Rhodococcus. To verify this assumption we determined the DNA-sequence of the PCR-amplified 16S rRNA gene (accession No KU140418) and compared the sequences to the database by BLAST search. The 16S rRNA gene was 99.7 and 99.2 % identical to Rhodococcus MK3027 and to R. rhodochrous MTCC11081, respectively. Together with the biochemical and morphological data we concluded that isolate RPK1 is a member of the species R. rhodochrous. It differed from the rubber degrading Xanthomonas sp. 35Y [8,9], Streptomyces sp. K30 [1], and other rubber degrading streptomycetes [3] by its inability to produce clearing zones on opaque polyisoprene latex agar plates. Previously, bacteria with a strong rubber-degrading activity but with no ability to form clearing zones had been isolated and identified as Gordonia polyisoprenivorans or Gordonia westfalica [20].
Identification of the gene coding for the latex clearing protein in R. rhodochrous strain RPK1 BLAST analysis revealed that many Actinobacteria and all known rubber-degrading Actinobacteria for which the genome sequences have been determined have at least one gene that codes for a so-called latex clearing protein (lcp) that is suspected to be responsible for the initial oxidative attack on the polyisoprene carbon backbone [1,[21][22][23]. Remarkably, non clearing zone formers such as G. polyisoprenivorans also have functional lcp genes [20]. This indicates that the Lcp protein apparently is not directly responsible for the formation of clearing zones during growth on opaque latex-agar. Alignment of the amino acid sequences of the Lcp proteins from different species (most of them annotated as Lcp protein but without verified function or biochemical characterization) revealed conserved regions within the Lcp amino acid sequences [15]. We identified a hypothetical lcp gene in the genomes of R. rhodochrous strain MTCC11081, Rhodococcus sp. MK3027 and Rhodococcus sp. ARG-BN062 by screening of the published genome sequences for the presence of lcp-like sequences. The deduced amino acid sequence of these hypothetical Lcp proteins included the DUF2236 domain that constitutes the central part of most if not all Lcp proteins [15]. Two oligonucleotides based on the upstream and downstream regions of the lcp genes of these Rhodococcus strains were generated (Lcp Rr -PstI_for and Lcp Rr -HindIII_rev) and a PCR reaction was performed with the chromosomal DNA of R. rhodochrous strain RPK1. A 1.5 kbp DNA fragment was obtained and its DNA sequence was determined (accession number KU140417). Analysis of the DNA sequence revealed one large open reading frame of 1227 bp that coded for a peptide of 408 amino acids (45.2 kDa, Table 1). The deduced amino acid sequence revealed strong similarities to the postulated Lcp proteins of R. rhodochrous strains and of several other formation of red-coloured colonies of R. rhodochrous RPK1 during growth on NB agar; (c) morphology of stationary R. rhodochrous RPK1 cells in bright field microscopy, note almost coccoid cells; (d) R. rhodochrous RPK1 cells during growth on NB medium supplemented with acetate (bright field and fluorescent image stained with Nile red). Note, star-like branched cells typical for R. Rhodochrous; (e) R. rhodochrous RPK1 cells during growth on NB medium supplemented with acetate (bright field and fluorescent image stained with Nile red, note, presence of Nile-red-stainable granules, possibly representing PHB granules or triacylglycerol bodies; (f) R. rhodochrous RPK1 cells during growth on NB medium supplemented with acetate (bright field and fluorescent image stained with DAPI and examined for presence of polyphosphate granules using DAPI-polyphosphate-specific emission filters). Note, presence of cell-pole localized polyphosphate granules in most cells Rhodococcus sp. strains (81 to 99 % identical amino acids). High degrees of similarities were also detected to many other genome-deduced sequences of putative Lcp proteins from bacteria including those from many streptomycetes and other Actinobacteria. When we compared the overall amino acid sequence of Lcp Rr with that of the only two other biochemically characterized Lcp proteins, a 70 % (76 %) identity (similarity) and a 57 % (66 %) identity (similarity) was determined to the Lcp VH2 from G. polyisoprenivorans [20] and Lcp K30 protein from Streptomyces sp. K30 [1], respectively (Additional file 1). A 30 amino acid long sequence at the N-terminus of Lcp Rr was predicted to code for a signal peptide that enabled the secretion of the protein. The molecular mass of the predicted mature protein amounted to 42.2 kDa (Table 1).

Expression and purification of Lcp Rr
The DNA sequence coding for the Lcp Rr signal peptide was replaced by a Strep-tag coding sequence and the modified gene was cloned under the control of an Lrhamnose-dependent promoter into p4782.1 and subsequently transformed to E. coli JM109. The Lcp Rr protein was purified from the combined cells of a 4.8 L E. coli (p4782.1::lcp Rr ) culture as described in the method section. A yield of approximately 7.7 mg of purified Lcp Rr protein (5.3 mg/mL in BCA assay) was obtained after the final purification step. The Lcp proteins of Streptomyces sp. K30 (Lcp K30 ) and of the rubber oxygenase RoxA of Xanthomonas sp. strain 35Y (RoxA Xsp ) were also purified and were used for comparison purposes. All purified proteins were separated by SDS-PAGE and checked for purity. As shown in Fig. 2, the Lcp Rr , Lcp K30 and RoxA Xsp proteins were almost homogenous. Both Lcp Rr , and Lcp K30 proteins migrated at a slightly higher apparent molecular mass (50 and 47 kDa) as deduced from the gene sequences (44.5 and 43.3 kDa, respectively). Most remarkably, concentrated Lcp Rr had a brownish colour. This was in sharp contrast to the red colour of the concentrated solutions of Lcp K30 or of RoxA Xsp and indicated that there were substantial differences of the Lcp Rr protein in comparison to the other rubber-cleaving enzymes. n.d not determined -not detectable (below detection limit) (a) deduced from [8,9,15] and SignalP4.1 Server (b) activities of Lcp Rr at 37°C were not reproducible in oxygen consumption assay (c) position in native sequence

Biochemical properties of Lcp Rr
The purified Lcp Rr protein was tested for its rubber cleaving activity using both the oxygen consumption and the HPLC-based rubber cleavage product assay. The oxygen consumption assay (Fig. 3a) confirmed that Lcp Rr cleaved poly(cis-1,4-isoprene) latex in an oxygendependent manner; specific activities of 0.9 U/mg and of 3.1 U/mg were determined for Lcp Rr at pH 8 and at 23°C and 30°C, respectively (Table 1). Variable data were determined for the specific activity of Lcp Rr at 37°C possibly because of the decreasing stability of the Lcp Rr protein at higher temperatures (see below). The HPLC ( Fig. 3b) and Fuchsin assay (Additional file 2) revealed that Lcp Rr produced the same mixture of polyisoprene cleavage products (C 20 and higher oligo-isoprenoids with terminal keto and aldehyde groups) that had been determined for Lcp K30 . ODTD was only detectable in trace amounts for Lcp Rr or for Lcp K30 but was the main product of the RoxA Xsp -derived rubber cleavage products. Determination of the activities of purified Lcp Rr at different pH values using the HPLC-based product assay revealed a pH optimum of around pH 8 ( Fig. 4) that was about one pH unit higher than the pH optimum that had been previously determined for RoxA Xsp or for  enzyme (control without enzyme) was added after 3.75 min (arrow) and oxygen saturation was recorded. Activities were calculated from the slopes of the graphs during the first minutes. The mixtures were extracted with ethylacetate after 90 min of incubation at 23°C and the formed cleavage products were analysed by HPLC (b). ODTD was the main cleavage product of RoxA Xsp but was present only in trace amounts in Lcp Rr and Lcp K30 preparations. The graph for Lcp Rr is given in its inverse orientation. Assays with each enzyme were performed at least three times; one typical dataset is shown Lcp K30 and Lcp1 VH2 [9,15,17]. The stability of all the Lcp preparations decreased upon incubation in buffer at 37°C (Fig. 5a). In accordance with this, the concentration of rubber degrading products in an in vitro latex cleavage assay with Lcp Rr or Lcp K30 increased for only 4-8 h (Fig. 5b). RoxA Xsp , on the other hand, was much more stable and continuously produced ODTD molecules for up to 70 h [24].
Lcp Rr is a b-type cytochrome and revealed remarkable differences to Lcp K30 Concentrated solutions of Lcp Rr had a brown colour while the Lcp K30 solutions were red. Figure. 6 shows a comparison of the UVvis spectra of the purified Lcp K30 and Lcp Rr proteins in the as isolated (oxidised) and in the dithionite-reduced state. Lcp Rr (and Lcp K30 ) as isolated both showed similar absorption maxima at 407 (412) nm and at 535 (544) nm that are typical for haemcontaining proteins in the oxidised form. However, purified Lcp Rr had an additional broad absorption maximum around 645 nm. The absorption band at 645 nm was absent in Lcp K30 and in other biochemically characterized RoxA proteins such as RoxA Xsp and RoxA Cco [10] and was responsible for the different (brown) colour of Lcp Rr . When the Lcp preparations were chemically reduced by the addition of sodium dithionite, the absorption bands at 407 (412) nm and 535 (544) nm shifted to 428 (430) nm and 560 (562) nm. A comparison of the reduced spectra of both Lcp proteins showed differences in the Q-bands (500 -600 nm). Apparently, Lcp Rr is far less pronounced in this region than Lcp K30 . Nevertheless, these data corresponded to the Soret and Q-bands that are typical for haem-containing proteins and strongly indicated that Lcp Rr is a haem-containing Fig. 4 pH optimum of Lcp Rr . The pH optimum was determined using the HPLC-based product assay in a pH range of 5 to 11 using acetate buffer (pH 5 -pH 6, diamonds), phosphate buffer (pH 6 -pH 9, open circles), carbonate buffer (pH 9 -pH 11, squares), or HEPES (pH 7, closed circle). Assays were performed with two biological and two technical replicates. Error bars indicate standard deviation To confirm that Lcp Rr is a haem-containing protein and to determine its haem type, a metal analysis and a spectral analysis by the haem-bipyridyl assay were performed. 6.5 μg Fe/mL Lcp Rr protein solution (5.3 mg protein/mL) were determined. This corresponded almost perfectly with a stoichiometry of one atom Fe per one Lcp Rr molecule. It was of interest that low amounts of copper (2.8 μg/mL) were also identified and corresponded to 0.36 atoms Cu per one Lcp Rr molecule. Zinc was detected at the detection limit (0.1 μg/mL) and Nickel was below the detection limit (<0.1 μg/mL); other metals (vanadium to zinc tested) were also not detected in significant amounts. Divalent cations such as magnesium or calcium were not present (below the detection limit of 0.1 μg/mL) ( Table 1). The presence of approximately one third of an atom% Cu per Lcp Rr molecule was unexpected because only traces of copper had been previously detected in Lcp K30 [18]. The determined amount of copper in Lcp Rr , however, was too high to be explained by an error in the determination of the metal or protein concentration. One possibility could be that the amount of copper was due to a contamination of the protein by traces of copper present in either the growth medium or in the buffer ingredients. For example, the used batch of NaCl that was present in some purification buffers was only of 98 % purity and could contain traces of heavy metals. However, sub-stoichiometric amounts of copper (precise concentration not known) had been also detected in Lcp1 VH2 of G. polyisoprenivorans [15]. Addition of an equimolar concentration or of a 10-fold molar excess of copper ions [Cu(II)Cl 2 ] to the assay mixture with purified Lcp Rr had no detectable effect on the UVvis spectrum or on the activity of Lcp Rr . The addition of 50 μM CuSO 4 to the Lcp Rr -expression culture produced no increased activity or yield of Lcp Rr . At present, there is no convincing explanation for the finding of variable sub-stoichiometric amounts of copper in the purified Lcp proteins from R. rhodochrous RPK1.
An absorption maximum of 556 nm was determined using the bipyridyl assay for Lcp Rr and for haemoglobin that was used as a b-type cytochrome control protein (Additional file 3). This result indicated the presence of a b-type haem in Lcp Rr . In contrast to the covalently linked c-type cytochromes, the haem groups of the btype cytochromes are not covalently linked to the peptide chain and can be therefore extracted by an acid solvent extraction [25]. Acid solvent extraction of the purified Lcp Rr yielded a coloured supernatant and a non-coloured precipitate. In contrast, solvent extraction of the c-type cytochromes such as RoxA Xsp or of other commercially available cytochrome c enzymes yielded a non-coloured supernatant and a red precipitate which is in agreement with the covalent attachment of porphyrin to the polypeptide. MALDI-ToF analysis of the purified Lcp Rr resulted in the identification of ions with m/z values of 616 (data not shown) which is typical for haem b [26]. Taken together, all these results indicated that Lcp Rr is a b-type cytochrome similar to Lcp K30 [18] Notably, MALDI-ToF analysis of Lcp Rr also revealed an ion species with m/z values of 619 besides that of 616 which could correspond to a verdo-haem [27]. As the activity of purified Lcp Rr rapidly and substantially decreased during storage, the haem species with m/z value of 619 could represent a haem degradation product of the inactivated Lcp Rr .

Lcp Rr is insensitive to most chelating inhibitors
Metal-dependent proteins are often inhibited by chelating compounds. Therefore, a variety of known chelator compounds was tested for their effects on the activity of Lcp Rr using the HPLC-based activity assay. EDTA, tiron, or phenanthroline had no significant effect on the Fig. 6 UVvis spectrum of Lcp Rr and Lcp K30 as isolated (black lines) and after reduction with dithionite (grey lines). Both Lcp proteins show a prominent band at 407 (412 nm in case of Lcp K30 ) that is characteristic for porphyrines. After reduction with dithionite a characteristic shift of the α-band to 428 (430) nm as well as an increase in the Q-band region (560/562 nm) was observable. Assays were repeated at least three times with two separate protein batches. A typical experiment is shown activity (Fig. 7). Ethyl xanthogenate partially inhibited Lcp Rr by ≈ 40 % similar to that for Lcp K30 but was different from the Lcp purified from G. polyisoprenivorans (Lcp1 VH2 ) that completely inhibited Lcp VH2 at 2 mM xanthogenate [15]. The only compound that had a strong effect on the activity of Lcp Rr was the metal chelator diethyl dithiocarbamate (82 % inhibition, Fig. 7). However, diethyl dithiocarbamate had no effect on the UVvis spectrum of Lcp Rr and this excluded a direct effect of the inhibitor at the haem site. Carbon monoxide, that completely inactivated RoxA and led to a prominent band at 415 nm by UVvis spectroscopy [18], had no effect on the absorption spectrum of Lcp Rr or Lcp K30 as isolated and this was in agreement with the presence of an oxidised (Fe 3+ ) haem centre. Carbon monoxide had no inhibitory effect on the polyisoprene cleavage during the assay for the HPLC-based product when sufficient oxygen was also present in the assay mixture. However, when Lcp Rr was incubated in carbon monoxidesaturated and oxygen-free buffer before it was added to an oxygenated polyisoprene latex assay solution, a lag phase of Lcp Rr in its ability to consume oxygen was observed. The oxygen consumption and polyisoprenecleaving activities were recovered within 30 to 50 min of incubation and exposure of the assay solution to air. The same result was obtained when Lcp K30 was exposed to carbon monoxide. Addition of carbon monoxide to the dithionite-reduced Lcp Rr or Lcp K30 had visible effects on the UVvis spectra as revealed by the increase of the α-band of Lcp Rr and Lcp K30 . The effect of carbon monoxide on the UVvis spectrum of Lcp was reversible by addition of a dioxygen atmosphere and indicated that the binding of carbon monoxide to the chemically reduced haem group in Lcp was reversible. This is different to RoxA Xsp that binds carbon monoxide more strongly and completely inhibits the activity. An apparent consequence of these data is that the haems of the Lcp proteins undergo a reversible Fe 3+ to Fe 2+ reduction during oxidative polyisoprene-cleavage and that the reduced Lcp proteins were the carbon monoxide-sensitive molecular species. This is the first evidence for a switch in the oxidation state of the active haem site of an Lcp protein during catalysis.

Lcp Rr but not Lcp K30 is accessible for external ligands
Previous studies on rubber oxygenase RoxA had revealed that the active haem site in RoxA had only one axial amino acid ligand. The other axial ligand was a dioxygen molecule that was stably bound to haem in a Fe 3+ --O 2 − transition state. The oxygen molecule in RoxA could be partially removed by the addition of imidazole thereby moving the negative charge from the oxygen molecule to the iron atom (Fe 2+ ). This charge transfer resulted in a small visible change of the UVvis spectrum as revealed by an increase of the absorption of the Q-bands at 549 nm [12,[15][16][17]. When imidazole was added to the dithionite-reduced RoxA, substantial increases in the Soret and Q-bands were determined Fig. 7 Inhibition of Lcp Rr by potential inhibitors. Activity was determined using the HPLC-based detection of polyisoprene cleavage products after 2 h of incubation of polyisoprene latex with Lcp at room temperature and subsequent solvent extraction of the products. The final concentration of inhibitors was 1 mM. The 100 % value corresponded to the area of the C 35 (23 min) product peak (see Fig. 3b). Sensitivity to carbon monoxide was tested by replacing 0.5 volumes of (oxygenated) assay buffer by a deoxygenated and CO-saturated assay buffer. Cleavage products were solvent-extracted after 2 h of incubation time. At least two repetitions with two technical replicates were performed for each inhibitor. Error bars indicate the upper and lower values of one experiment compared to the reduced RoxA bands without imidazole [18,24]. This increase in absorption was interpreted as the result of the binding of the imidazole molecule to the (now) free sixth (axial) coordination site of the haem iron. When we performed an analog experiment with purified Lcp Rr and with purified Lcp K30 we found remarkable differences between both Lcp proteins: addition of imidazole to the dithionite-reduced Lcp K30 had no effect on the UVvis spectrum and there was no detectable increase of the Q-bands. This indicated that the 6th coordination site of the haem apparently was not accessible for imidazole and the Lcp K30 protein was present in a "closed state". Binding of the substrate (polyisoprene) would therefore require a conformational change of the Lcp K30 structure. In contrast, addition of imidazole to the dithionite-reduced Lcp Rr protein resulted in a substantial increase of the Soret-and Qbands (Fig. 8) and this can be explained by the binding of imidazole to the reduced haem. Similar results were obtained when both the Lcp proteins were treated with mercaptoethanol: no change of the UVvis spectrum was determined for Lcp K30 while prominent changes were detected for the Lcp Rr protein (Additional file 4). In conclusion, Lcp K30 and Lcp Rr seem to rest in a different conformation in their as isolated states. While Lcp K30 as isolated was in a six-fold coordinated "closed" state, the haem group of Lcp Rr was readily accessible to external ligands and substrates, and this indicated a five-fold coordinated "open" state. Further evidence for this can be found in the UVvis spectra of five-fold coordinated myoglobin in the oxidised (met myoglobin) and reduced (desoxy-myoglobin) state. The UVvis spectra of the latter proteins showed similarities to the corresponding spectra of Lcp Rr , particularly in the region of the less pronounced Q-bands of reduced Lcp Rr compared to Lcp K30 as well as in the 645 nm region in the oxidised (as isolated) state [28]. The presence of the 645 nm absorption band in Lcp Rr might be also explained by a charge transfer phenomenon of a charged residue/ion in close neighbourhood to the haem group in Lcp Rr and in its absence in Lcp K30 [29]. Unfortunately, only the RoxA structure [30] but no Lcp structure was available to obtain direct support for our assumption.
lcp genes are frequently present in the genomes of Actinobacteria [22] and many rubber degrading species have been described for members from this group [3-5, 23, 31, 32]. Most of the rubber degrading actinomycetes such as Streptomyces sp. K30 [1], Streptomyces coelicolor 1A, and many others [3,6,33] produce clearing zones on opaque polyisoprene latex agar. However several of the most potent rubber degraders do not produce clearing zones and apparently need a close contact to the rubber material they degrade. Two well known rubber degrading Gordonia species (G. polyisoprenivorans and G. westfalica) [20,22] and also the strain from this study (R. rhodochrous RPK1) belong to this group of nonclearing zone formers. One might speculate that the Lcp proteins of non-clearing zone formers constitute a group that have an open conformation with free access to the active site (no conformational change is needed) and that the other Lcp proteins that have an ability to form a clearing zone have a closed form. The prototype of the first group would be Lcp Rr and the prototype of the latter would be Lcp K30 . It will be necessary to biochemically investigate more Lcp proteins and to solve the structure of Lcp proteins to find evidence for or against this hypothesis.

Conclusions
This study extends the list of biochemically characterized rubber-degrading non-clearing zone formers by latex clearing proteins (Lcp) to the genus Rhodococcus (besides Gordonia). The detection of rubber-cleaving activity with purified Lcp Rr and the absence of clearing Fig. 8 UVvis spectrum of Lcp Rr and Lcp K30 before and after the addition of 1 mM imidazole (as isolated/after reduction) in the area of the Q-band. No significant change was observable for Lcp K30 whereas for Lcp Rr a substantial increase at 560 nm was determined upon addition of imidazole. Assays were repeated two times with two separate protein batches. A typical experiment is shown zones during growth on polyisoprene latex agar raises the question of whether the designation "latex clearing protein" has been well-chosen. Rubber oxygenase B (RoxB) would be an appropriate alternative. However, the designation Lcp has been used is several previous publications and has also been used for many annotated genes in genome-sequenced Actinobacteria. Re-classification of Lcp as RoxB therefore could lead to the confusion of other research workers.
The isolation and characterization of the Lcp protein of R. rhodochrous RPK1 in this study showed that all the so far studied Lcp proteins can differ in some spectroscopic features and/or in spatial arrangements of their metal ions/cofactors and indicate the presence of two or even more subgroups of Lcp proteins. It will be necessary to study more Lcp proteins to reveal the complete variability of rubber degrading enzymes present in rubber-degrading organisms.

Methods
Bacterial strains, plasmids and culture conditions Table 2 shows the bacterial strains, plasmids and oligonucleotides that were used in this study. R. rhodochrous strain RPK1 was grown with nutrient broth (NB) medium or in mineral salts medium (MSM, 9 g/L Na 2 HPO 4 ×12 H 2 O, 1.5 g/L KH 2 PO 4 , 1 g/L NH 4 NO 3 , 0.2 g/L MgSO 4 ×7 H 2 O, 0.02 g/L CaCl 2 ×2 H 2 O, 1.2 mg/L Fe(III)ammonium citrate with solid rubber pieces or with water-soluble carbon sources as indicated at 30°C. Pieces (1 cm ×1 cm) of heat-sterilised vulcanised rubber (commercial but not-powdered rubber protecting gloves) were added to the sterile mineral salts medium during the enrichment and growth of R. rhodochrous on rubber (0.6 % [wt/vol]). Plasmid-carrying recombinant E. coli strains were grown with LB medium at 22°C or 37°C in the presence of the appropriate antibiotic (ampicillin or kanamycin). Polyisoprene latex was kindly provided by Weber and Schaer, Hamburg (Germany) and was used after 3 washing steps in 0.1 % (wt/vol) Nonidet P40. For purification of Lcp Rr , recombinant E. coli cells were grown in LB medium supplemented with 0.1 % (wt/vol) L-rhamnose at 22°C. Utilization of carbon sources was tested on mineral salts agar with separately filter-sterilised carbon sources at the following end concentrations (sugars, sugar alcohols and sugar acids at 0.5 % [wt/vol], sodium acetate [0.25 %, wt/vol], sodium benzoate [0.1 %, wt/vol]). Volatile compounds (alkanes) were applied by adding a quantity of 100 μL to a sterile filter paper placed in the lid of a petri disk. The plates were sealed with parafilm and incubated separately at 30°C. Utilization of PHB was tested on PHB overlay plates as described previously [34]. Growth at different temperatures was tested on NB agar plates.

Enrichment and isolation of rubber-degrading microorganisms
Liquid from a waste pond at a rubber latex processing factory in Thailand (Namom rubber factory at Namom, Songkhla) was used as an inoculum to enrich for rubber-degrading microorganisms in a mineral salts medium (MSM) that had been supplemented with 1×1 cm pieces of rubber gloves as a sole source of carbon and energy. After two weeks of incubation at 30°C, 0.1 volumes (without pieces of rubber) were transferred to fresh medium and incubated for an additional month. Substantial disintegration of the new rubber pieces became visible and indicated that active rubber-degrading microorganisms were present. Several bacterial strains were isolated from this enrichment culture by repeated purification from streaks onto NB and LB agar plates. Each isolate was subsequently tested for its ability to degrade rubber in liquid MSM with rubber pieces as carbon source. One isolate (designated as isolate RPK1) with strong rubber-degrading activity was selected for this study.
Cloning and heterologous expression of lcp Rr , and determination of the 16S rRNA gene sequence of isolate RPK1 The lcp Rr gene was amplified using the chromosomal DNA from R. rhodochrous strain RPK1 as template and the oligonucleotides Lcp Rr -complete_for and Lcp Rr -com-plete_rev as PCR primers and Takara Primestar DNA polymerase as the proof-reading polymerizing enzyme. The DNA sequence of the product was determined and is available under the accession no KU140417. Alternatively, the coding sequence of mature Lcp Rr was amplified from chromosomal DNA using Lcp Rr -mature-PstI_for and Lcp Rr -mature-HindIII_rev as primers. The DNA products were purified via agarose gel electrophoresis, cleaved with restriction enzymes PstI and HindIII and ligated into plasmid pUC9::lcp K30 that had been cleaved by the same restriction enzymes. The coding sequence for strep-tagged lcp Rr was cut out using HindIII and NdeI and was subsequently ligated into the expression plasmid p4782.1 and transformed to competent E. coli JM109 cells. A part of the 16S rRNA gene of the isolate RPK1 was PCR-amplified using the primers 16S-universal_for and 16S-universal_rev. The DNA sequence of the resulting PCR product was determined (1412 bp) and revealed a strong similarity to the 16S rRNA genes of several Rhodococcus sp. strains. The 16S rRNA gene sequence of isolate RPK1 was determined after PCR amplification using the primers (16S-Rr-complete_for and 16S-Rr-complete_rev), that were specific for the ends of the known 16S rRNA gene sequences of R. rhodochrous strains taken form the NCBI data base, and is now available under the accession no KU140418.

Purification of Lcp Rr , Lcp K30 and of RoxA Xsp
Purification of the rubber oxygenase of Xanthomonas sp. 35Y (RoxA Xsp ) and latex clearing protein Lcp K30 was performed as described previously [12,18] Lcp Rr was purified as follows: eight individual 600 mL LB cultures in 3 litre Erlenmeyer flasks were inoculated each with 0.02 volumes of a seed culture of E. coli JM109 harbouring the plasmid p4782.1::lcp Rr that had been grown with the same medium. It was important that L-rhamnose (0.1 %, wt/vol) was present right from the beginning in the main cultures to maximise the yield of the expressed Lcp Rr protein. Cells of the main culture were harvested by centrifugation after ≈ 20 h of growth at 22°C and were immediately used for protein purification. The cell pellet was resuspended in 100 mM potassium phosphate buffer, pH 7.7, containing 150 mM sodium chloride (KPN, 2 mL KPN/g cell wet weight). A soluble cell extract was prepared by two French press steps and subsequent centrifugation at 40,000 g for 40 min. The supernatant (≈ 60 mL) was directly applied to a 10 mL Strep-Tactin HC gravity flow column that had been equilibrated with KPN buffer. The column was washed with at least five volumes of KPN buffer before the Lcp Rr protein was eluted by ≈ 30 mL of 5 mM desthiobiotin dissolved in KPN. Lcp Rr -containing fractions were combined, desalted by running through a G25 Sephadex (26/ 160) Hiprep desalting column (53 mL bed volume) that had been equilibrated with 1 mM potassium phosphate (KP) buffer, pH 7.0 and subsequently concentrated to 1-2 mL via ultrafiltration (10 kDa cut-off ). Remaining impurities were removed by chromatography on a Superdex 200 column (16/600, equilibrated with 1 mM KP, pH 7) at a flow rate of 1 mL/min. Combined Lcp Rr -containing fractions were ultrafiltrated (10 kDa cut-off ) and concentrated to ≈ 1.5 mL. Aliquots of the purified Lcp Rr protein were stored on ice for about 3 days (Lcp K30 up to 1 week) or shock-frozen with liquid nitrogen and stored at -70°C.

Determination of the cytochrome type of Lcp Rr
The haem type of Lcp Rr was determined by the bi-pyridyl assay as described elsewhere [35]. Purified RoxA Xsp , cytochrome c (horseheart, type III, Sigma, St. Louis, USA) (both c-type cytochromes) and haemoglobin (b-type) (bovine, Sigma, St. Louis, USA) were used as controls for known c-type and b-type cytochromes, respectively. 25 μL