Biotechnol Lett DOI 10.1007/s10529-015-1918-x

ORIGINAL RESEARCH PAPER

Isolation and characterization of an interactive culture of two Paenibacillus species with moderately thermophilic desulfurization ability Jia Wang . Batzaya Davaadelger . Joelle K. Salazar . Robert R. Butler III . Jean-Franc¸ois Pombert . John J. Kilbane . Benjamin C. Stark

Received: 19 May 2015 / Accepted: 15 July 2015 Ó Springer Science+Business Media Dordrecht 2015

Abstract Objective To isolate and characterize novel thermophilic bacteria capable of biodesulfurization of petroleum. Results A culture containing two Paenibacillus spp. (denoted ‘‘32O-W’’ and ‘‘32O-Y’’) was isolated by repeated passage of a soil sample at up to 55 °C in medium containing dibenzothiophene (DBT) as sulfur source. Only 32O-Y metabolized DBT, apparently via

Electronic supplementary material The online version of this article (doi:10.1007/s10529-015-1918-x) contains supplementary material, which is available to authorized users.

the 4S pathway; maximum activity occurred from 40 to 45 °C, with some activity up to at least 50 °C. 32OW enhanced DBT metabolism by 32O-Y (by 22–74 % at 40–50 °C). With sulfate as sulfur source, 32O-Y and 32O-W grew well up to 58 and 63 °C, respectively. Selection of a mixed culture of 32O-Y and 32O-W at 54 °C increased DBT metabolism 36–42 % from 40 to 45 °C. Genome sequencing identified desulfurization gene homologs in the strains consistent with their desulfurization properties. Conclusion The 32O-Y/32O-W culture may be a useful starting point for development of an improved thermophilic petroleum biodesulfurization process.

J. Wang  B. Davaadelger  J. K. Salazar  R. R. Butler III  J.-F. Pombert  J. J. Kilbane  B. C. Stark (&) Department of Biology, Illinois Institute of Technology, Chicago, IL 60616, USA e-mail: [email protected]

Keywords Biodesulfurization  Dibenzothiophene  Directed evolution  Interaction  Paenibacillus  Thermophiles

J. Wang e-mail: [email protected]

Introduction

B. Davaadelger e-mail: [email protected]

Removal of organic sulfur from petroleum by microorganisms (‘‘biodesulfurization’’) was discovered more than 25 years ago, and remains a possible cost-effective and environmentally friendly alternative to chemical desulfurization methods (Kilbane 2006). Despite considerable work, however, efforts to advance this technology to a stage at which it can be used at an industrial level have been hampered by both low biodesulfurization enzyme activities and the need for the process to be performed at thermophilic

J. K. Salazar e-mail: [email protected] R. R. Butler III e-mail: [email protected] J.-F. Pombert e-mail: [email protected] J. J. Kilbane e-mail: [email protected]

123

Biotechnol Lett

temperatures (Kilbane 2006). One approach to solve the second problem has been the isolation and characterization of thermophilic bacteria with desulfurization abilities (Konishi et al. 1997; Ishii et al. 2000; Kirimura et al. 2001; Kayser et al. 2002). The (mesophilic) enzyme pathway for desulfurization, originally discovered in Rhodococcus erythropolis IGTS8, involves the metabolism of dibenzothiophene (DBT) to 2-hydroxybiphenyl (2-HBP), releasing the sulfur as sulfite and preserving the caloric value of DBT; it is known as the ‘‘4S’’ pathway (Denome et al. 1994; Kilbane 2006). DBT is an excellent substrate in such studies because it is a significant source of organic sulfur in petroleum and cannot be efficiently and selectively desulfurized through the currently applied hydrodesulfurization method (McFarland et al. 1998; Song 2002). The mesophilic 4S pathway has been found in a number of bacteria (Omori et al. 1992; Izumi et al. 1994; Ohshiro et al. 1996; Yu et al. 2006; Mohebali et al. 2007; Bhatia and Sharma 2010), and the operon (‘‘dszABC’’) and its homologs encoding the three enzymes involved have been isolated and characterized (Denome et al. 1994; Kilbane and Robbins 2007). Thermophilic versions of the 4S pathway have also been discovered and characterized (Konishi et al. 1997; Ishii et al. 2000; Furuya et al. 2001; Kirimura et al. 2001; Kayser et al. 2002; Li et al. 2003), and operons, such as tdsABC (Ishii et al. 2000) and bdsABC (Kirimura et al. 2001), encoding the enzymes involved in these 4S pathways have also been characterized. The tdsABC encoded enzymes are active up to about 60 °C (Konishi et al. 1997). In an effort to identify and characterize additional thermophilic biodesulfurization competent cultures, we used selection at up to 55 °C in medium with DBT as the sole sulfur source using a locally obtained soil sample as the starting inoculum, to select a culture named 32O. From this culture two species of Paenibacillus were isolated. As described below, these two species may be a useful starting point for development of an improved thermophilic biodesulfurization process.

Materials and methods Bacterial strains Culture 32O was originally isolated by repeated passage of a soil sample obtained from near the Field

123

Museum in Chicago, IL, in minimal medium with DBT as sole sulfur source; the early passages were done at 30 and 37 °C, and ultimately the culture was selected in the same medium at 55 °C. Two strains, named 32O-Y and 32O-W, were isolated from this culture by repeated passages on LB plates. The experiments described in this report were conducted on both pure cultures of 32O-W and 32OY, as well as a reconstituted mixed culture of 32O-W and 32O-Y (denoted 32O-Y ? W #1). 32O-Y ? W #1 was also subjected to further selection at 54 °C, in an attempt to improve desulfurization at higher temperatures; this produced a culture denoted 32OY ? W #2, on which experiments were also conducted. Rhodococcus erythropolis strain IGTS8 (Kayser et al. 1993) was used as a desulfurization positive control. The 32O strains were identified by 16S rDNA sequence analysis using neighbor-joining phylogenetic inferences performed by SeqWright (Houston, TX) followed by maximum likelihood (ML) reconstructions (performed by us). Growth media Media used in this study included LB medium (pH 7.0) and a modified version of the minimal medium CDM. Stock solutions for modified CDM included 1 M phosphate buffer (pH 7.2) [Na2HPO4 (86.6 g/l) and KH2PO4 (53 g/l)]; 1 M NH4Cl; 1 M MgCl2; 0.3 M CaCl2; trace elements (pH 6.7) (FeCl36H2O (2.04 g/l), ZnCl2 (70 mg/l), MnCl24H2O (100 mg/l), CoCl26H2O (200 mg/l), CuCl22H2O (20 mg/l), NiCl26H2O (20 mg/l), Na2MoO42H2O (40 mg/l) and H3BO4 (20 mg/l)); vitamin mix [cyanocobalamine B12 (100 mg/l), pyridoxamine-2HCl B6 (300 mg/l), Ca–D(?) pantothenate (100 mg/l), thiamine dichloride B1 (200 mg/l), nicotinic acid (200 mg/l), 4-aminobenzoic acid (160 mg/l), and D(?)biotin (20 mg/l)]; yeast extract (50 g/l); and 1 M glucose. Stock solutions of phosphate buffer, NH4Cl, MgCl2, CaCl2, glucose and yeast extract were prepared separately and sterilized by autoclaving. Trace elements and vitamin mix were sterilized by filtration. To prepare 1 liter of working modified CDM, the following solutions were mixed with autoclaved H2O up to 1 liter: 50 ml phosphate buffer, 10 ml NH4Cl, 1 ml MgCl2, 1 ml CaCl2, 1 ml trace elements, 1 ml vitamin mix, 1 ml yeast extract and

Biotechnol Lett

60 ml glucose solution. As sulfur sources, 125 ll of either 40 mM DBT (dissolved in ethanol) or 40 mM Na2SO4 was added per 50 ml of modified CDM to give final concentrations of 0.1 mM. Milli-Q purified water was used to prepare all solutions and media and to rinse all glassware. Desulfurization activity of resting cells Pre-cultures were prepared by inoculating cells into 50 ml modified CDM containing 0.1 mM DBT in 250 ml flasks and incubating until the mid-growth phase at 30 °C, 200 rpm for R. erythropolis IGTS8 and 50 °C, 250 rpm for all 32O cultures. Pre-cultures were transferred into 50 ml of the same CDM medium in 250 ml flasks with an initial OD600 of *0.03, and incubated under the same growth conditions described above. Cells were harvested at an OD600 of 0.4–0.7 by centrifugation at room temperature, washed twice with ice-cold phosphate buffer (0.05 M, pH 7.2), and again pelleted by centrifugation at 4 °C. Cells were resuspended in the same ice-cold phosphate buffer, 5 ml of which was used immediately in resting cell reactions, while the rest was transferred to a preweighed aluminum dish and dried in a 90–100 °C oven to determine dry cell weight. In the resting cell reaction, 5 ml cell suspension was warmed to the desired test temperature (25–60 °C) for 2 min, followed immediately by addition of 75 ll DBT (54 mM in ethanol); the final concentration of DBT was 0.8 mM. The resting cell reaction was allowed to proceed at the corresponding temperature with a shaking speed of 200 rpm for 1 h. One ml of each reaction was then taken for the Gibbs assay (Pan et al. 2013) to determine the concentration of 2-HBP (measuring the A610 of each reaction). A control sample (cell suspension prepared as above but without DBT) was also subjected to the Gibbs assay, and that A610 value was subtracted from the A610 value for the 1 h time point to give the net activity. The Gibbs reagent (2, 6-dichloroquinone-4-chloroimide) reacts with the hydroxyl group of 2-HBP to produce a blue product, the amount of which thus measures the amount of DBT metabolized through the 4S pathway. The A610-concentration equivalence was determined from Gibbs assays on samples of 2-HBP of known concentrations. One unit of desulfurization activity (nmol/mg DCW.min) is defined as 1 nmol of 2-HBP produced by 1 mg dry cell weight of cells in 1 min.

Desulfurization activity of growing cells Cells were inoculated into 50 ml modified CDM medium containing 0.1 mM DBT in 250 ml flasks and incubated at the desired temperatures (40, 45, 50 and 55 °C), at 250 rpm until the mid-point of the log phase. These pre-cultures were inoculated into 50 ml of the same CDM medium in 250 ml flasks with initial OD600 of 0.03, and each incubated at the same temperature as its pre-culture at 250 rpm for 48 h. One ml of each culture was used to perform the Gibbs assay to determine the accumulated amount of 2-HBP (lmol/ml), and the rest of each culture was harvested for dry cell weight determination as described above. The Gibbs reaction controls in the growing cell assays were performed on CDM medium immediately after inoculation from pre-cultures, but before addition of DBT. The same inoculation and growth protocol was used for preliminary experiments (Fig. 1), but with growth monitored at OD600 for 7 days. Growth rate in CDM-SO2 4 (0.1 mM) Pre-cultures were prepared by inoculation of cells in modified CDM medium with sulfate (SO2 4 ) (0.1 mM) replacing DBT as the sulfur source; the growth conditions were 250 rpm at 45–66 °C. Each preculture was inoculated into 50 ml of the same CDMsulfate medium with initial OD600 of around 0.05 and incubated at 250 rpm at the same temperature. Bacterial growth was monitored by measuring the turbidity at 600 nm until late log phase. The OD600 values were converted to log10 (OD600), and plotted versus hours. The greatest slope (determined with at least 3 points) was defined as the maximum growth rate (log10 (OD600)/h). These values were then converted to generations/h. Investigation of novel desulfurization traits Attempts to amplify dsz/tds homologs from strains 32O-Y and 32O-W using primers based on conserved sequences of known dsz and tds operons were unsuccessful with two sets of primers, one based on sequences that are highly conserved in desulfurization operons in different species (forward and reverse primers 50 -TAN GAC CGN GCN GAN GAN TT-30 and 50 -TTG TTN TCG CTG GAN GCN TT-30 ,

123

Biotechnol Lett Fig. 1 Growth (50 °C, 250 rpm) of pure cultures of strains 32O-Y, 32O-W, and a mixed 32O-Y plus 32O-W culture (32O-Y ? W #1) in CDM medium containing 0.06 M glucose as carbon source and 0.1 mM DBT as sulfur source. Growth in this experiment was measured by OD600 rather than DCW. Plating of the mixed culture confirmed, from colony morphology, that both 32OW and 32O-Y grew in the mixed culture

respectively, where ‘‘N’’ is equimolar amounts of A, T, C and G) and one based on the tdsABC desulfurization operon from Paenibacillus sp. A11-2 (forward and reverse primers 50 -CAT GAT CAT ATG AGG AGG CAA TCG ATG CGT CAA ATG CAT CTT 30 and 50 CAT GAT ACT AGT TCA AGA GTA AAA GCT GGG GGT-30 , respectively). Following this, whole genomic DNA was isolated from both strains (PowerSoil DNA Extraction Kit, MoBio, Carlsbad, CA, USA) and subjected to genomic analysis using both hybrid Illumina and PacBio sequencing methods. Draft genomes (to be reported elsewhere, including exact locations of all dsz/tds homologs) were assembled using SMRT portal 2.2.0 (https://github.com/ PacificBiosciences/SMRT-Analysis). BLASTN and TBLASTN/BLASTP searches for dsz/tds related sequences in the genomes of both 32O-Y and 32O-W were run for the tdsA, tdsB, tdsC and tdsD genes from Paenibacillus sp. A11-2 and for the dszA, dszB, dszC and dszD genes from R. erythropolis IGTS8 (GenBank Protein Accessions: BAA94831.1, BAA94832.1, BAA94833.1, BAB13707.1, P54995.1, P54997.1, P54998.1, AAC38226.1). The threshold E value was set at 1e-20. Pfam searches were performed to identify

123

Dsz/Tds functional protein domains in all the potential orthologs found in BLAST searches for both strains. Conserved domain families for the DszA (DszA/ BdsA/TdsA), DszB (DszB/BdsB/TdsB) and DszC (DszC/BdsC/TdsC) proteins were retrieved from the NCBI conserved domain database (CDD) (MarchlerBauer et al. 2015) under accession numbers TIGR03860, cd13554 and cd01163, respectively, and the underlying multiple sequence alignments downloaded as multifasta files. Hidden Markov models (HMM) for the DszA, DszB and DszC proteins were built from each multifasta file with hmmbuild and then proteins tentatively displaying these conserved domains were searched for in the 32O-Y and 32O-W proteomes with hmmsearch, both from the HMMER 3.1b1 package (Eddy 2011). Functions were further validated by bidirectional BLAST hits against the NCBI non-redundant database and independent InterProScan 5 web queries (Jones et al. 2014; Mitchell et al. 2015). Putative dsz gene sequences for strains 32O-W and 32O-Y were deposited in GenBank under accession numbers KR057780 to KR057808 and KR057810 to KR057888, respectively.

Biotechnol Lett

Results General characteristics of strains 32O-Y and 32OW 32O-Y was identified by 16S rDNA sequencing as Paenibacillus naphthalenovorans; 32O-W was also identified as Paenibacillus, but the species to which it belongs is uncertain (Supplementary Fig. 1 and Supporting Information). In pure culture, strain 32O-Y grew in CDM medium with DBT as sulfur source; strain 32O-W, however, was not, but, as confirmed by the appearance of both colony types on plated samples from mixed cultures, grew in mixed cultures with 32O-Y (Fig. 1). The maximum growth rates of the 32O-Y and mixed cultures were similar in this experiment, although the mixed culture reached a maximum OD600 of about three times that of 32O-Y. This result, and the ability of 32O-Y cultures to accumulate 2-HBP (as detected by the Gibbs assay; see below) indicated that only 32O-Y contained a fully functioning 4S pathway. Desulfurization activity of resting cells Desulfurization activities of resting cells of IGTS8, 32O-Y, 32O-Y ? W #1 and 32O-Y ? W #2 cultures were also compared at various temperatures (Fig. 2). The activity of IGTS8 reached its maximum of 0.55 units at 30–35 °C, and dramatically decreased when the temperature exceeded 35 °C. The optimal temperature for 32O-Y is 40 °C, where activity is 0.13 units; the activity of 32O-Y is 0.1 units at 45 °C. The activity of 32O-Y ? W #1 is higher than that of the pure 32OY culture by 22, 36 and 74 % at 40, 45 and 50 °C, respectively. This increase indicated that 32O-W can enhance the activity of 32O-Y; possible reason(s) for this observation are discussed later. After the high temperature selection, the activity of 32O-Y ? W #2 was increased by 42 and 36 % at 40 and 45 °C, respectively, but was nearly identical at 50 °C compared to that of 32O-Y ? W #1. Desulfurization activity of growing cells Data in Fig. 3 demonstrate the accumulated 2-HBP (mM) produced by 32O-Y, 32O-Y ? W #1 and 32OY ? W #2 cultures after 48 h incubation in CDMDBT (0.1 mM) at various temperatures. When

incubation was between 40 and 50 °C, 32O-Y ? W #1 and 32O-Y ? W #2 converted more DBT to 2-HBP than 32O-Y alone. This result confirmed that 32O-W, as a helper, can improve the desulfurization activity of 32O-Y. Dry cell weights of these three cultures grown at various temperatures were also compared after incubation in CDM-0.1 mM DBT for 48 h (Fig. 4). At 45 °C, the two mixed cultures have slightly less biomass despite accumulating more 2-HBP than 32O-Y; at 50 °C both the biomass and 2-HBP produced by the mixed cultures are greater than those produced by 32O-Y. This suggests that, depending on the growth temperature, 32O-W may improve the desulfurization activity of 32O-Y in more than one way. Growth rate in CDM-SO2 4 Temperature-dependent growth rates of pure 32O-Y and 32O-W cultures were also tested using 0.1 mM sulfate as the sulfur source, and their maximum thermostabilities were thus investigated (Fig. 5). 50 °C is the optimal temperature for both 32O-Y and 32O-W. 32O-Y can grow faster than 32O-W when the temperature is lower than 58 °C, but cannot grow at 60 °C; the two strains have similar growth rates at 58 °C. At 63 °C, the growth rate of 32O-W is decreased by 45 % compared to its highest growth rate (at 50 °C), and no growth occurs at 66 °C. Analysis of 32O-Y and 32O-W dsz/tds related genes BLASTN searches returned zero hits when either 32OY or 32O-W genomes were queried with any of the eight dsz/tds genes mentioned in ‘‘Materials and methods’’ section. Both TBLASTN and BLASTP searches, however, returned the same hits in all 16 searches, and suggested that dsz/tds homologous genes do exist in 32O-Y and 32O-W. The dszA search against the 32O-Y genome identified the same genes as the tdsA search. This was also true for the 32O-Y dszC/tdsC searches; a 32O-Y tdsB but no dszB homolog was also identified (Fig. 6). 32O-Y has 15 putative dszA/tdsA homologs, one tdsB homolog, and four dszC/tdsC homologs. 32O-Y also has one homolog (49 % identity of inferred amino acid sequence) of the tdsD gene, which encodes the

123

Biotechnol Lett Fig. 2 Temperature dependent desulfurization activity of resting cells of R. erythropolis IGTS8, 32O-Y, 32O-Y ? W #1 and 32OY ? W #2. Values are averages of at least three independent experiments. Error bars are omitted for clarity, as duplicates agreed to less than 5 % of the means

Fig. 3 DBT degradation by growing cultures of 32O-Y, 32O-Y ? W #1 and 32OY ? W #2 at various temperatures. Assays measured 2-HBP accumulation in the growth medium after 48 h of growth. Values are averages of at least three independent experiments. Error bars are omitted for clarity, as duplicates agreed to less than 5 % of the means

NADH-FADH2 oxidoreductase utilized in the pathway, and which is not linked to the dszABC/tdsABC genes in characterized species. The inferred amino

123

acid sequence similarities were uniformly higher (by a few percentage points) for the tdsA genes (38–48 %) compared to the dszA genes (37–44 %). The four

Biotechnol Lett Fig. 4 Dry cell weight (DCW) accumulation by growing cultures of 32O-Y, 32O-Y ? W #1 and 32OY ? W #2 at various temperatures. DCW values represent the biomass (g/l) produced after incubation for 48 h. Values are averages of at least three independent experiments. Error bars are omitted for clarity, as duplicates agreed to less than 5 % of the means

Fig. 5 Growth rates of 32O-Y and 32O-W at various temperatures in CDM-0.1 mM SO2 4 . Values are averages of at least three independent experiments (standard deviations indicated)

123

Biotechnol Lett Fig. 6 Putative 4S pathway desulfurization genes in strains 32O-Y and 32O-W. Y-axis, number of individual loci in the genomes of strains 32O-Y and 32O-W for which TBLASTN/ BLASTP results showed a significant (E value lower than 1e-20) homology to inferred amino acid sequences of thermophilic (tds) and mesophilic (dsz) desulfurization genes isolated from Paenibacillus sp. A11-2 (a) and Rhodococcus erythropolis IGTS8 (b), respectively

dszC/tdsC homologs have, overall, slightly greater identity (inferred amino acid sequences) to those of dszC than to tdsC (28–44 versus 26–45 %) and the closest match in each case of comparable similarity as the closest dszA/tdsA matches. The tdsB homolog has lower amino acid sequence identity (only 28 %). In 32O-Y the D gene homolog is located close to two A gene homologs, there are two cases where a C gene homolog is immediately adjacent to an A gene homolog, and there are three cases in which a number (2–7) of A gene homologs are closely linked to each other. There is, however, no case in which homologs to all three of A, B, and C are in the same operon. In general the dszABC/tdsABC homologs are found in several fairly widely separated loci. The BLAST search identified three homologs for the A gene and no homologs for either the B or C genes in 32O-W. The amino acid sequence similarities to known A genes are similar to those for the A genes from 32O-Y, and, as in 32O-Y, the similarities are slightly greater to tds versions (41–45 %) than to dsz versions (39–43 %). There are two tdsD homologs in 32O-W, one of which is 99 % identical in amino acid sequence to TdsD from Paenibacillus sp. A11-2. The

123

32O-W A and D homologous genes are found generally in widely separated loci, although one of the A homologs and one of the D homologs are closely linked. The Conserved Domain Database search did identify a DszB/TdsB homolog in 32O-W, but even this additional search could find no DszC/TdsC homolog in this strain. In addition, family protein domains found in the dszA/tdsA, dszB/tdsB, dszC/tdsC and dszD/tdsD homologs from 32O-Y are identical or similar to the domains found in the corresponding tds genes from Paenibacillus sp. A11-2. The same is true for the homologous sequences of dszA/tdsA and dszD/tdsD from 32O-W.

Discussion Comparison of the 32O-Y/32O-W culture relative to other desulfurization competent thermophiles The protocols used to measure resting cell desulfurization specific activities previously reported for various thermophiles vary somewhat, so that no

Biotechnol Lett

precise comparison with our cultures can be made. Generally, however, the peak specific activity of each of our cultures (32O-Y, 32O-Y ? W #1, and 32OY ? W #2) was comparable to or greater than those reported for M. phlei WU-F1 (Furuya et al. 2001), Paenibacillus A11-2 (Konishi et al. 1997), and B. subtilis WU-S2B (Kirimura et al. 2001), but only about one-fourth that of M. phlei GTIS10 (Kayser et al. 2002). The resting cell desulfurization temperature optima for our three cultures were similar to those of B. subtilis WU-S2B (Kirimura et al. 2001) and M. phlei WU-F1, but about 5° lower than for M. phlei GTIS10 (Kayser et al. 2002) and Paenibacillus A11-2 (Konishi et al. 1997). In this regard, it should also be noted that the overall thermostability of the 4S pathway measured in whole cells of any species may be a function of both the inherent stabilities of the 4S enzymes as well as other factors. This has been shown for the 4S pathway in M. phlei strain GTIS10, which has 4S enzymes that are identical to those of R. erythropolis strain IGTS8, but metabolizes DBT by the 4S pathway at temperatures far above those at which IGTS8 is active (Kayser et al. 2002).

metabolize DBT but could grow in the mixed culture by being fed sulfur from IGTS8’s metabolism of DBT. Also similar to our results, growth of E. cloacae with IGTS8 appeared to stimulate IGTS8’s rate of DBT metabolism. Analysis of both accumulated 2-HBP and dry cell weight in the growing culture experiments suggests that there might be more than one way that this increase in DBT metabolism of 32O-Y occurs. For example, at 45 °C, 32O-W might enhance the rate of sulfur metabolism of 32O-Y without increasing the overall biomass; but at 50 °C, 32O-W might increase the growth of itself or 32O-Y to maintain the higher activity of mixed cultures. The 32O mixed culture may be a useful candidate for further attempts to increase thermostable DBT metabolism. The interaction between 32O-Y and 32OW suggests a possibility of establishing a more active 4S pathway by mixing microbes, each of which contributes to DBT metabolism in different ways. The overall performance of the mixed culture might be improved by cooperative modifications in both strains.

Desulfurization activity of the 32O-Y/32O-W culture

Based on the current data that 32O-W may possess a partial 4S pathway, can enhance the desulfurization ability of 32O-Y, tolerate DBT well and grow up to at least 63 °C, it is possible that 32O-W might be a good host for genetic engineering and expression of thermophilic desulfurization genes. In addition, 32O-Y can grow reasonably well at temperatures of at least 58 °C and thus is already metabolically adapted to both high temperature growth and use of DBT as sole sulfur source. This means that 32O-Y is also a promising starting candidate for selection of a strain with even greater desulfurization activity and thermostability.

Results of assays with both resting and growing cultures are consistent in showing that DBT metabolism by 32O-Y was aided by co-culturing with 32OW. This may occur in part because 32O-W may contain a partially functioning DszABC/TdsABC pathway composed of homologs to DszA/TdsA, DszB/TdsB, and DszD/TdsD, but not DszC/TdsC. Expression of the A, B, and D proteins would presumably allow 32O-W to catalyze the third and fourth steps in the 4S pathway, utilizing DBT sulfone and hydroxyphenylbenzenesulfinate produced during DBT metabolism by 32O-Y (Kilbane 2006). Alternatively, 32O-W could utilize sulfate produced by 32OY metabolism of DBT. In either case this would create an additional ‘‘sulfur sink’’ that forces 32O-Y to increase its DBT metabolism. A similar result has been reported previously for mixed cultures of R. erythropolis IGTS8 and Enterobacter cloacae growing in minimal-DBT medium (Kayser et al. 1993). E. cloacae is itself unable to

Thermostabilities of 32O-Y and 32O-W

DszABC/TdsABC homologs In this study, we found that the amino acid sequences of the DszABC/TdsABC homologs in 32O-Y and 32O-W have 26–48 % identity with the corresponding Tds and Dsz proteins from Paenibacillus sp. A11-2 and R. erythropolis IGTS8, respectively. This is somewhat lower than the similarities between the same three proteins when Paenibacillus sp. A11-2 is

123

Biotechnol Lett

compared with R. erythropolis IGTS8 (Ishii et al. 2000), B. subtilis WU-S2B is compared with either Paenibacillus sp. A11-2 or R. erythropolis IGTS8 (Kirimura et al. 2004), or Gordonia alkalinivorans 1B is compared to R. erythropolis IGTS8 (Alves et al. 2007). That the 32O-Y and 32O-W homologs are putative 4S pathway enzymes is also supported by the identification in all of them of protein domains corresponding to those of the Dsz/Tds proteins from R. erythropolis IGTS8 and Paenibacillus sp. A11-2. Other of the characteristics of the dszABC/tdsABC homologs from 32O-Y and 32O-W are also somewhat unusual. None of the dszABC/tdsABC homologs in 32O-Y are arranged in an operon containing one copy of each gene, as is the typical arrangement (dszABC from the mesophiles R. erythropolis IGTS8 (Denome et al. 1994) and Gordonia alkanivorans 1B (Alves et al. 2007), and tdsABC or the homologous bdsABC from the thermophiles Paenibacillus sp. A11-2 (Ishii et al. 2000) or B. subtilis (Kirimura et al. 2004)). It is also interesting that 32O-Y, and to a lesser extent 32OW, may have multiple copies of some of the dsz/tds genes, especially dszA/tdsA. These findings imply that the dsz/tds homologs or the desulfurization trait of 32O-Y might be novel. If this is borne out in further studies, these two strains may provide us an additional resource for developing an improved thermostable biocatalyst for desulfurization. It is important to note, however, that it is not yet known which of the dsz/tds homologs in the two strains actually encode DBT metabolic enzymes and which may encode homologous proteins with functions unrelated to desulfurization. The 32O-Y and W system may also be useful as a model to study synergistic interactions between two bacterial species, including aspects of microbial ecology/physiology and genomic evolution. In addition, since neither strain appears to contain a dsz/tds operon, they may provide an opportunity to identify the primordial sources of the dszABCD/tdsABCD genes and the molecular processes by which those individual genes assembled themselves into operons. Acknowledgments This work was supported by a Grant (No. 6600019855) from the Saudi Arabian Oil Company (Saudi Aramco). Supporting information Supplementary Fig. 1—Data for the identification of strain 32-Y as Paenibacillus naphthalenovorans and strain 320-W as Paenibacillus sp.

123

References Alves L, Melo M, Mendonca D, Simoes F, Matos J, Tenreiro R, Girio FM (2007) Sequencing, cloning and expression of the dsz genes required for dibenzothiophene sulfone desulfurization from Gordonia alkalinivorans strain 1B. Enzyme Microbial Technol 40:1598–1603 Bhatia S, Sharma D (2010) Biodesulfurization of dibenzothiophene, its alkylated derivatives and crude oil by a newly isolated strain Pantoea agglomerans D23W3. Biochem Eng J 50:104–109 Denome SA, Oldfield C, Nash LJ, Young KD (1994) Characterization of the desulfurization genes from Rhodococcus sp. strain IGTS8. J Bacteriol 176:6707–6716 Eddy SR (2011) Accelerated profile HMM searches. PLoS Comput Biol 7:e1002195 Furuya T, Kirimura K, Kino K, Usami S (2001) Thermophilic biodesulfurization of dibenzothiophene and its derivatives by Mycobacterium phlei WU-F1. FEMS Microbiol Lett 204:129–133 Ishii Y, Konishi J, Okada H, Hirasawa K, Onaka T, Suzuki M (2000) Operon structure and functional analysis of the genes encoding thermophilic desulfurizing enzymes of Paenibacillus sp. A11-2. Biochem Biophys Res Commun 270:81–88 Izumi Y, Ohshiro T, Ogino H, Hine Y, Shimao M (1994) Selective desulfurization of dibenzothiophene by Rhodococcus erythropolis D-1. Appl Environ Microbiol 60:223–226 Jones P, Binns D, Chang HY, Fraser M, Li W, McAnulla C, McWilliam H, Maslen J, Mitchell A, Nuka G, Pesseat S, Quinn AF, Sangrador-Vegas A, Scheremetjew M, Yong SY, Lopez R, Hunter S (2014) InterProScan 5: genomescale protein function classification. Bioinformatics 30:1236–1240 Kayser KJ, Bielaga-Jones BA, Jackowski K, Odusan O, Kilbane JJ (1993) Utilization of organosulfur compounds by axenic and mixed cell cultures of Rhodococcus rhodochrous IGTS8. J Gen Microbiol 139:3123–3129 Kayser KJ, Cleveland L, Park HS, Kwak JH, Kolhatkar A, Kilbane JJ (2002) Isolation and characterization of a moderate thermophile, Mycobacterium phlei GTIS10, capable of dibenzothiophene desulfurization. Appl Microbiol Biotechnol 59:737–746 Kilbane JJ (2006) Microbial biocatalyst developments to upgrade fossil fuels. Curr Opin Biotechnol 17:305–314 Kilbane JJ, Robbins J (2007) Characterization of the dszABC genes of Gordonia amicalis F.5.25.8 and identification of conserved protein and DNA sequences. Appl Microbiol Biotechnol 75:843–851 Kirimura K, Furuya T, Nishii Y, Ishii Y, Kino K, Usami S (2001) Biodesulfurization of dibenzothiophene and its derivatives through the selective cleavage of carbon-sulfur bonds by a moderately thermophilic bacterium Bacillus subtilis WU-S2B. J Biosci Bioeng 91:262–266 Kirimura K, Harada K, Iwasawa H, Tanaka T, Iwasaki Y, Furuya T, Ishii Y, Kino K (2004) Identification and functional analysis of the genes encoding dibenzothiophene-desulfurizing enzymes from thermophilic bacteria. Appl Microbiol Biotechnol 65:703–713

Biotechnol Lett Konishi J, Ishii Y, Onaka T, Okumura K, Suzuki M (1997) Thermophilic carbon-sulfur-bond-targeted biodesulfurization. Appl Environ Microbiol 63:3164–3169 Li FL, Xu P, Ma CQ, Luo LL, Wang XS (2003) Deep desulfurization of hydrodesulfurization-treated diesel oil by a facultative thermophilic bacterium Mycobacterium sp. X7B. FEMS Microbiol Lett 223:301–307 Marchler-Bauer A, Derbyshire MK, Gonzales NR, Lu S, Chitsaz F, Geer LY, Geer RC, He J, Gwadz M, Hurwitz DI, Lanczycki CJ, Lu F, Marchler GH, Song JS, Thanki N, Wang Z, Yamashita RA, Zhang D, Zheng C, Bryant SH (2015) CDD: NCBI’s conserved domain database. Nucl Acids Res 43:D222–D226 McFarland BL, Boron DJ, Deever W, Meyer JA, Johnson AR, Atlas RM (1998) Biocatalytic sulfur removal from fuels: applicability for producing low sulfur gasoline. Crit Rev Microbiol 24:99–147 Mitchell A, Chang HY, Daugherty L, Fraser M, Hunter S, Lopez R et al (2015) The InterPro protein families database: the classification resource after 15 years. Nucl Acids Res 43:D213–D221 Mohebali G, Ball AS, Rasekh B, Kaytash A (2007) Biodesulfurization potential of a newly isolated bacterium,

Gordonia alkanivorans RIPI90A. Enzyme Microb Technol 40:578–584 Ohshiro T, Hirata T, Izumi Y (1996) Desulfurization of dibenzothiophene derivatives by whole cells of Rhodococcus erythropolis H-2. FEMS Microbiol Lett 142:65–70 Omori T, Monna L, Saiki Y, Kodama T (1992) Desulfurization of dibenzothiophene by Corynebacterium sp. strain SY1. Appl Environ Microbiol 58:911–915 Pan J, Wu F, Wang J, Xu L, Khayyat NH, Stark BC, Kilbane JJ (2013) Enhancement of desulfurization activity by enzymes of the Rhodococcus dsz operon through coexpression of a high sulfur peptide and directed evolution. Fuel 112:385–390 Song C (2002) New approaches to deep desulfurization for ultra-clean gasoline and diesel fuels: an overview. Prepr Paper Am Chem Soc Div Fuel Chem 47:438–444 Yu B, Ma C, Zhou W, Wang Y, Cai X, Tao F, Zhang Q, Tong M, Qu J, Ping X (2006) Microbial desulfurization of gasoline by free whole-cells of Rhodococcus erythropolis XP. FEMS Microbiol Lett 258:284–289

123