Enhancement of dibenzothiophene biodesulfurization by weakening the feedback inhibition effects based on a systematic understanding of the biodesulfurization mechanism by Gordonia sp. through the potential “4S” pathway

Abstract

Gordonia sp. JDZX13 (source: industrial petroleum soil) shows good potential for dibenzothiophene (DBT) biodesulfurization. With the GC/MS analysis of metabolites and PCR-sequencing verification of the key desulfurization operon (dszA/dszB/dszC), the valuable “4S” pathway of DBT biodesulfurization in Gordonia sp. is identified. The key rate-limiting factors (2-hydroxybiphenyl/sulfate ions) suggest significant feedback inhibition effects on cell growth (μ_x) and biodesulfurization efficiency. Moreover, the qRT-PCR analysis of the dszA/dszB/dszC operon transcriptions also indicates the prominent negative effects on the key desulfurization enzyme activity, particularly under sulfate ion stress. Based on the abovementioned analysis, the oil/aqueous ratio in the two-phase system was optimized as 1 : 2 for better weakening of the inhibition effect, and a higher DBT removal efficiency (improved by 100.7%) was achieved. In addition, the DBT biodesulfurization mechanism is proposed. Related methods and mechanisms would be useful to further guide similar biocatalysis processes in the near future.

---

Introduction

Sulfur dioxide emissions from fuel combustion are the primary cause of acid rain and air pollution.^1–4 During further refining of crude oil, the content of polycyclic aromatic sulfur-containing heterocyclic compounds (PASH) gradually increases, particularly with respect to the dibenzothiophene compound (C_X-DBT) content.^5,6 C_X-DBT (>70%), such as 4,6-dimethyl dibenzothiophene (4,6-DMDBT), 4-methyl dibenzothiophene (4-MDBT), and dibenzothiophene (DBT), are recognized as the main organosulfur compounds, and the total sulfur content in diesel catalysis is restricted to 500 mg L⁻¹.^7,8 With increasing concern for environmental safety, efficient sulfur-removal of C_X-DBT in the petroleum refining process has attracted considerable attention in last few decades.^9,10

One of the alternative options for the removal of sulfur from fossil fuel is biodesulfurization.^11–13 Sulfur atoms form 0.5–1% of the dry weight of bacterial cells, such as the structure of some enzyme cofactors (coenzyme A, thiamine and biotin), amino acids and proteins (cysteine, methionine, and disulfur bonds).^7,14,15 Organic sulfur in crude oil can be utilized by specific microbial activity, resulting in a highly targeted removal effect.^16,17 With the advantages of moderate reaction conditions, low cost, and environmental friendliness, biodesulfurization is recognized as an alternative sulfur-removal technology.^13,18,19 It was estimated that the investment cost of a biological desulfurization plant is only 2/3^rd of the traditional hydrodesulfurization, and operating costs and carbon dioxide emissions are reduced by 15% and 70–80%, respectively.²⁰ To date, the three ring-destructive pathways for DBT metabolism via biodesulfurization have been recognized as the destruction of the benzene ring structure or “Kadama” pathway, “C–S” reduction pathway and specific “4S” pathway.^21–24 The “4S” metabolic pathway of the DBT-removal process can catalyze CS bond breaking to form 2-hydroxybiphenyl and water-soluble sulfates without destroying the benzene ring structure.²⁵ This pathway does not affect the combustion value of oil and is recognized as the most valuable potential application pathway. Therefore, it is very significant to promote the commercialization process of biodesulfurization by deeply understanding the biodesulfurization mechanism of strains with the potential “4S” metabolic pathway.

Nowadays, although various desulfurization strains have been gradually isolated, industrial needs have still not been satisfied. It has been suggested that the bacterial desulfurization efficiency needs to be maintained at 1.2–3.0 mmol g⁻¹ h⁻¹ in industrial applications. However, the highest desulfurization efficiency of the most current engineered strain is only 0.28 mmol g⁻¹ h⁻¹.²⁶ There are also several bottlenecks that limit the commercialization process such as the biocatalyst's organic solvent tolerance and feedback inhibition effects of metabolites accumulated during the bioconversion process, which result in a low desulfurization rate.^15,27–30 The major end metabolic products in the “4S” metabolic pathway of DBT biodesulfurization are 2-HBP and sulfate ions, and their effects on cell growth, key gene expression level and biodesulphurization efficiency are not well understood in detail.

In our previous study, the novel bacterial strain Gordonia sp. JDZX5 (isolated from the Karamay industrial oilfield) showed good potential in dibenzothiophene (DBT) biodesulfurization.³¹ The main aim of this study is to investigate the DBT biodesulfurization mechanism and weaken the end-product inhibition effects by (a) identifying the metabolic pathway via the analysis of metabolic products and key desulfurization gene operon; (b) revealing the detailed effects of the major end-products 2-HBP and sulfate ions on cell growth, transcriptional levels of dszA/dszB/dszC and desulfurization efficiency; and (c) optimizing the oil/aqueous ratio to weaken the major feedback inhibition effects.

Methods

Materials and methods

Strain, media, and culture conditions. Gordonia sp. JDZX13 was isolated from crude oil-contaminated samples of the Karamay Oilfield, Xinjiang, China.³¹ It was deposited in the Culture and Information Center of Industrial Microorganisms of China University (CICIM) with the number B7054. The detailed strain characteristics are listed in Table 1. Strain JDZX13 was cultured in a basic salt medium (BSM). The basal salts of the BSM are as follows (g L⁻¹): K₂HPO₄: 5.0, NaH₂PO₄·2H₂O: 2.0, NH₄Cl: 2.0 and MgCl₂·6H₂O: 0.2. The trace elements are as follows (mg L⁻¹): CaCl₂: 20.0, FeCl₃·6H₂O: 4.0, CoCl₂·6H₂O: 4.0, MnCl₂·4H₂O: 0.8, NH₄Mo₄·2H₂O: 0.2, ZnCl₂: 0.2, CuCl₂·2H₂O: 0.1, AlCl₃·6H₂O: 0.1 and H₃BO₃: 0.05. Carbon source: 10.0 g L⁻¹ glycerol. Sulfur source: 0.3 mM dibenzothiophene (DBT) or 4,6-dimethyldibenzothiophene (4,6-DMDBT). The initial pH of the media was adjusted to 7.0. The strain was previously adapted in 0.3 mM DBT at 35 °C and 170 rpm. It was incubated in fresh media once for 48 h.

Table 1 The main characteristics of the strains used in this study

Species	Strain	Microbial type	Optimal T/pH	Description and source
Gordonia sp.	JDZX13	Desulfurizing-bacteria, chemoheterotroph	20–40 °C, pH 5.0–11.0	Crude oil-contaminated samples of the Karamay oilfield, Xinjiang, China

---

Identification of metabolites in biodesulfurization process by GC/MS. The medium sample was previously acidified to pH 2.0 by 1 mM HCl. DBT and 2-HBP were extracted with the equal volumes of ethyl acetate. The GC/MS system was equipped with a DB-5 fused-silica capillary column model SE-54 (30 m × 0.32 nm × 0.33 μm, Agilent, California, USA) with a flame ionization detector. The temperature of the detector and injector was set at 280 °C. The initial column temperature was set at 150 °C and held for 15 min. Then, it was increased to 200 °C at the rate of 5 °C min⁻¹ and held for 3 min. Finally, it was increased to 270 °C at the rate of 10 °C min⁻¹ and held for 5 min. The analysis time was set at 40 min. The gas mass spectrometer conditions were set as follows: OV21: capillary column (30 m × 0.25 mm), column temperature: 210 °C, injector temperature: 300 °C, fid detector temperature: 325 °C, hydrogen flame ionization detector, carrier gas: N₂, column pressure: 40 kPa and injection volume: 10 μL.

Identification of key biodesulfurization gene operon and phylogenetic analysis. Whole DNA was extracted according to the bacterial genomic extraction kit (FastaGen, Shanghai, China). The related gene sequences of dszA/dszB/dszC from closely related strains were blasted by the DNAMAN software, and the PCR primers were designed by Primer 6.0 software in the conserved region (Table 2). The PCR reaction conditions were set as follows: 94 °C, thermal denaturation: 5 min, 94 °C melt: 30 s, 55 °C anneal: 30 s, 72 °C extend 60 s. The 2× Ultra-Pfu Master Mix containing high fidelity enzymes was chosen for the amplification reaction. The PCR product was purified using the B-type small DNA fragment Gel Extraction Kit (BioDev-Tech, Beijing, China). The dszA/dszB/dszC gene sequences of JDZX5 were compared to closely related strains in the National Center for Biotechnology Information using the Blast Alignment tool. The phylogenetic tree was constructed using the MEGA 6.0 software.

Table 2 The designed primers for dszA/dszB/dszC identification and analysis of the transcription level

Gene		Primer sequence (5′ → 3′)
PCR identification of dszA/dszB/dszC	dszA	F: 5′-GGCTGGCAATGTGACTCAT-3′
		R: 5′-TATTCGGTGCGGAAGTAGC-3′
	dszB	F: 5′-TTGAGGCTGGTGTTCAGA-3′
		R: 5′-GGTTATCTGCTCAACGGCA-3′
	dszC	F: 5′-GGTTATCTGCTCAACGGCA-3′
		R: 5′-TCTTCACCATCAGTTCGCC-3′
qRT-PCR analysis of gene transcription levels	dszA	F: 5′-AGGTATTCCAGCCGGTCCT-3′
		R: 5′-ACATCAAAGCCCAGGTGAAG-3′
	dszB	F: 5′-CTGGTGCCGATCAACAGTC-3′
		R: 5′-AACAGGGCGTCAACCTCAC-3′
	dszC	F: 5′-ATTGTCGCCTTCATCCAGTC-3′
		R: 5′ATCCTCGGTTGCCTGTTG-3′
	16S rDNA	F: 5′-TCCTGGTGTAGCGGTGAAAT-3′
		R: 5′-ACGGAACTCGTGAAATGAGC-3′

---

Biodesulfurization experimental procedure for analyzing the effects of 2-HBP and sulfate ions

Procedure for the analysis of the effect of 2-HBP and sulfate ions on cell growth and desulfurization.
Growth cell. Strain JDZX13 was inoculated in 250 mL flasks with 100 mL enrichment medium containing 2-HBP/sulfate ions (0, 0.05, 0.1 or 0.2 mmol L⁻¹) and 0.3 mmol L⁻¹ DBT. The culture conditions were set at 35 °C and 170 rpm for 80 h. Resting cells, which are a type of dormant cell, when suspended in a phosphate buffer solution, does not grow and reproduce, however, still maintain the catalytic ability of enzymes. Resting cells were employed as the indicator of desulfurization enzyme activity.^32,33 60 mL of re-suspended resting cells in potassium phosphate buffer (pH 7.0) containing 2-HBP/sulfate ions (0, 0.05, 0.1 or 0.2 mmol L⁻¹) and 0.3 mM DBT were used. DBT biodesulfurization was carried out in 250 mL flasks at 30 °C and 170 rpm for 12 h.

Assay procedure for analyzing cell growth and desulfurization. During the course of the biodesulfurization process, samples were collected at 2 h intervals to detect the biomass, DBT and 2-HBP concentrations by GC/MS. The pH value was measured by a pH meter (PHB-3TC, Sartorius, Germany) with a gel-filled electrode. Biomass was measured by a spectrophotometer (IV-1100D, Meipuda, China) at the absorbance of 620 nm. Specific growth rate (μ_x) was calculated from the slope of the semilogarithmic plot of biomass versus bioleaching time according to eqn (1). X₀ and X_t indicate the biomass at the beginning and end of the exponential phase, respectively, and t indicates the duration of the exponential phase. Desulfurization activity was determined based on the quantity of DBT degraded per hour per liter (mM h⁻¹ L⁻¹ DBT) and the quantity of 2-HBP produced per hour per liter (mM h⁻¹ L⁻¹ 2-HBP).

μ_x = ln(X_t/X₀)/t (1)

Fluorescent quantitative PCR experiments

5% resting cells in phosphate buffered saline suspension was inoculated in a fresh medium. After 12 h growth, the cells were harvested by centrifugation at 6000 × g for 5 min. Then, 1 mL of Tris–EDTA buffer (30 mM Tris, 10 mM EDTA, pH 8.0) was added. After centrifugation at 6000 × g for 2 min, the supernatant was discarded. 0.1 mL of 100 mg mL⁻¹ lysozyme and 25 mg mL⁻¹ lysostaphin were added and incubated at 37 °C in a water bath for 2 h. 0.6 mL of TRIzol (Sangon, Shanghai, China) was added and shaken at room temperature for 10 min. RNA was extracted according to the RNeasy Mini extraction kit. The extracted RNA was tested by gel-electrophoresis and stored at −70 °C. Reverse transcriptional test was performed by the AMV First Strand cDNA Synthesis Kit (Roche, Basel, Switzerland). The primers of dszA/dszB/dszC for fluorescent quantitative PCR were designed using the Oligo 7 software (Table 2). The reagents of the PCR system are as follows (20 μL): SybrGreen qPCR Master Mix: 10.0 μL, forward primer: 0.4 μL, reverse primer: 0.4 μL, dd H₂O: 8.8 μL and template (cDNA): 0.4 μL. The reaction mixtures were incubated at 95 °C for 10 min, followed by 45 cycles of 10 s at 95 °C and 30 s at 60 °C. PCR was carried out in a 96-microwell plate by a fluorescent quantitative PCR instrument (ABI Stepone plus, AB, USA).

Oil/aqueous ratio regulation experiment

To investigate the desulfurization capacity of the strain JDZX5 in an oil/aqueous two-phase system, resting cells were employed as the model. Hexadecane was used as the imitated model oil, whereas DBT was selected as the organic sulfur compound model. N-Hexadecane with 0.3 mM DBT was mixed as the oil phase and potassium phosphate buffer (pH 7.0) with resting cells were inoculated as the seed liquid in the oil/aqueous two-phase system. Different oil/aqueous ratios, such as pure water, 2 : 1, 1 : 1, 1 : 2 and 1 : 3, were selected for optimizing the biodesulfurization process. The DBT in the upper layer was determined by measuring the amount of residual DBT through GC/MS analysis every 12 h.

Statistical analysis

All experiments were performed in triplicate and the statistical analysis of the experimental data was performed by one-way analysis of variance, which is presented as mean values ± SD. The statistical analysis was performed with the software SPSS 17.0 (SPSS Inc., Chicago, USA).

Results and discussion

Identification of biodesulfurization metabolic pathway in Gordonia sp. JDZX13

Identification of metabolites (DBT/4,6-DMDBT as substrate). The GC/MS analysis of the main metabolites in C_X-DBT biodesulfurization is presented in Fig. 1. Two major peaks were presented at the residence time of 7.93 min and 18.45 min. After a comparative analysis in the GC/MS data library, the compounds of the peaks were identified as 2-HBP and DBT, respectively (Fig. 1A). Similarly, two peaks were presented at the residence time of 15.48 min and 25.11 min. After a comparative analysis in the GC/MS data library, the compounds of the peaks were identified as 4,6-dimethyl-hydroxybiphenyl (4,6-DMHBP) and 4,6-dimethyldibenzothiophene (4,6-DMDBT), respectively (Fig. 1B). Moreover, the metabolites in different phases were also analyzed to better reveal the detailed biodesulfurization pathway (Fig. 1C). In the beginning phase of biosulfurization and 60 h of the blank control system (without strain), only the DBT peak was detected, which indicates that DBT cannot be spontaneously degraded. During the course of DBT biodesulfurization, the peak area of DBT gradually decreased along with an increase in the peak area of 2-HBP. At 40 h, the peak area of 2-HBP was almost equal to that of DBT. At 60 h, the DBT peak disappeared, and 2-HBP gradually became the maximum peak, which indicate that almost all the DBT was completely converted to 2-HBP. This result is consistent with the previous study.¹² Moreover, 0.1 mmol L⁻¹ 2-HBP was used as the sole carbon source for culturing the strain JDZX13 in the BSM medium. The content of the HBP remained stable, which indicates that 2-HBP is the end-product of DBT biodesulfurization by the strain JDZX13.

Fig. 1 Identification of the metabolites of the C_X-DBT biodesulfurization process by GC/MS analysis (DBT and 4,6-DMDBT as substrates). (A) DBT as the substrate; (B) 4,6-DMDBT as the substrate; and (C) metabolite identification at different times.

Identification of key pathyway genes and phylogenetic analysis. To further solve the desulfurization pathway, the typical desulfurization genes (dszA/dszB/dszC) were identified. With the whole DNA as the template, the target genes were amplified by designed primers. The PCR products were assayed using 0.8% agarose gel electrophoresis (Fig. 2A). Three single bands with the size of 1200 bp, 700 bp, and 550 bp were obtained. The size of the bands was consistent with the sequence size of the typical desulfurization genes dszA/dszB/dszC.³⁴ After further sequencing analysis, the dszA/dszB/dszC operon was identified as the key functional element in the DBT desulfurization pathway of strain JDZX13. High homology genes were selected from the GenBank and the phylogenetic tree was constructed using the MEGA 6.0 software with the neighbor-joining method (Fig. 2B–D). The similarity between strain JDZX13 dszA and Gordonia sp. WQ-01 dszA (DQ448811.1) reached 99%. However, the values between Brevibacillus brevis HN1 dszA (KJ002079.1), Agrobacterium tumefaciens dszA (AY960127.1), and Rhodococcus erythropolis dszA (AY714058.1) were 91%, 88% and 88%, respectively. The similarity between dszB of strain JDZX13 and Gordonia sp. WQ-01 dszB (Q448813.1) was up to 99%, whereas it was only 88% compared to the other desulfurization strain Rhodococcus sp. FMF (AJ514947.2). Similarly, dszC of strain JDZX13 showed 99% similarity with that of Gordonia sp., whereas only 89% with Mycobacterium goodii X7B dszC (EU527978.1). The abovementioned results indicate that the species differences of the key desulfurization gene dszA/dszB/dszC are significant.

Fig. 2 Characterization, phylogenetic tree analysis of the desulfurization genes and desulfurization pathway of C_X-DBT of Gordonia sp. JDZX13. (A) Agarose electrophoresis of PCR products; (B) dszA phylogenetic tree analysis; (C) dszB phylogenetic tree analysis; (D) dszC phylogenetic tree analysis; and (E) desulfurization pathway of C_X-DBT by Gordonia sp. JDZX13.

Based on the dual identification of metabolites by GC/MS and desulfurization gene by PCR/sequecing, the valuable “4S” desulfurization pathway can be presented, as shown in Fig. 2E. It is suggested that Gordonia sp. JDZX13 metabolized C_X-DBT by the desulfurization gene (dszA, dszB and dszC) encoding the three enzymes (DSZA, DSZB and DSZC). With the action of the first enzyme DszC, C_X-DBT was oxidized to a C_X-DBT sulfone substance. Subsequently, the C_X-DBT sulfone substance was hydrolyzed to C_X-DBT sulfone formation of acid salts by DszA. Finally, the sulfite acid salts and hydroxybiphenyl substances were formed by the action of DszB enzyme. Some other strains, such as Gordonia sp. F.5.25.8 and strain AK6 (incomplete “4S” pathway), were also identified according to similar principals.^27,34

Effects of 2-HBP on cell growth, biodesulfurization and transcriptional level of dszA/dszB/dszC

Effects of 2-HBP on μ_x of growth cell. It is commonly recognized that feedback inhibition effects might be generated by the accumulation of end products.^7,13 With DBT as the sole sulfur source, DBT will be finally converted to the intermediate 2-HBP and sulfate ions through the “4S” pathway. It has been indicated that excess 2-HBP and sulfate ions are likely to affect cell performance in the desulfurization process.^22,35 Therefore, the effects of 2-HBP and sulfate ions on cell growth and desulfurization were studied to further improve the desulfurization performance.

The effects of different 2-HBP densities on μ_x are depicted in Fig. 3A. The appearance time of μ_max was obviously lengthened with an increase in 2-HBP. μ_max was tested at about 21.44 h in the blank system, whereas the 0.05 mmol L⁻¹, 0.1 mmol L⁻¹ and 0.2 mmol L⁻¹ 2-HBP systems were tested at 22.40 h, 31.40 h and 55.00 h, respectively. Moreover, the value of μ_max in the blank system was 0.147 h⁻¹, while the values were 0.139 h⁻¹, 0.131 h⁻¹ and 0.110 h⁻¹ in the related systems containing additional 2-HBP. These results indicate that the accumulation of 2-HBP is harmful and not only lengthens the lag phase but also threatens cell growth in biodesulfurization. It was reported that 0.52 mM 2-HBP showed high toxicity for the cell growth of another desulphurizing-bacteria, R. erythropolis, compared to ethanol and dimethyl formamide.³⁵

Fig. 3 Effects of different 2-HBP concentrations on cell growth and desulfurization of Gordonia sp. JDZX13. (A) Specific growth rate; (B) biomass and DBT removal efficiency by growth cells at 60 h; (C) DBT removal efficiency by resting cells at 12 h; and (D) gene transcriptional level of dszA/dszB/dszC genes. a–c and x–z represent the statistically significant differences (c > b > a; z > y > x).

Effects of 2-HBP on biodesulfurization. The biomass and DBT removal efficiency of growth cell under different 2-HBP densities are listed in Fig. 3B and Table 3. In the blank and 0.05 mmol L⁻¹ 2-HBP systems, DBT was completely eliminated. The DBT removal rates in the 0.1 mmol L⁻¹ and 0.2 mmol L⁻¹ 2-HBP systems were 98.8% and 20.2%, respectively. These results show that 2-HBP is highly toxic to cell growth, which also inhibits the biodesulfurization performance. Resting cells can be recognized as a model for evaluating desulfurization enzyme activity owing to its good microbial stability (not for cell growth).^32,33 The profile of the DBT removal efficiency in different systems was similar to that of growth cells (Fig. 3C). The DBT removal rate decreased rapidly with an increase in 2-HBP concentration. The values in the blank, 0.05 mmol L⁻¹, 0.10 mmol L⁻¹ and 0.20 mmol L⁻¹ 2-HBP systems were 20.89%, 15.71%, 11.51% and 3.31%, respectively. These results indicate that 2-HBP strongly inhibits the desulfurization enzyme activity.

Table 3 The main process parameters of growth cells in DBT biodesulfurization (at 60 h) under different concentrations of 2-HBP and sulfate ions in the aqueous phase

Condition	Concentration (mmol L^−1)	μmax (h^−1)	μmax time (h)	Biomass (g L^−1)	Removal DBT (mmol L^−1)	Removal productivity (mmol L^−1 h^−1)
Blank	0	0.147	21.44	1.32	0.3	0.227
2-HBP	0.05	0.139	22.40	1.10	0.3	0.273
	0.10	0.131	31.40	0.99	0.296	0.299
	0.20	0.110	55.00	0.34	0.061	0.179
Sulfate ion	0.05	0.157	21.00	3.99	0.294	0.074
	0.10	0.163	20.10	5.64	0.287	0.051
	0.20	0.171	19.20	5.59	0.047	0.008

---

Effects of 2-HBP on transcriptional levels of dszA/dszB/dszC of resting cells. To further prove the effects of different 2-HBP densities on the transcriptional level of key desulfurization related genes, fluorescent quantitative PCR was performed (Fig. 3D). With an increase in 2-HBP, the transcriptional levels of the desulfurization genes dszA/dszB/dszC gradually decreased. Compared to the blank system, the transcriptional levels of the desulfurization genes dszA/dszB/dszC were 0.89, 0.71 and 0.65 times, respectively, in 0.05 mmol L⁻¹ 2-HBP. Compared to the blank system, the transcriptional levels of desulfurization genes dszA/dszB/dszC were 0.81, 0.66 and 0.59 times, respectively, in the 0.1 mmol L⁻¹ 2-HBP system. Compared to the blank system, the transcriptional levels of the desulfurization genes dszA/dszB/dszC were 0.77, 0.59 and 0.56 times, respectively, in the 0.2 mmol L⁻¹ 2-HBP system. These results are also very consistent with the performance of cell growth and biodesulfurization. It was proved that HBP is a potential inhibitor of the second (dszC) and fourth steps (dszB) of the “4S” pathway, with the resting cells of Pseudomonas putida CECT 5279 and R. erythropolis IGTS8 under aqueous conditions.³¹

Effects of SO₄²⁻ on cell growth, biodesulfurization and transcription of dszA/dszB/dszC

Effects of SO₄²⁻ on μ_x of growth cell. The effects of different sulfate ion densities on μ_x are depicted in Fig. 4A and Table 3. The appearance time of μ_max was obviously shortened with an increase in sulfate ions. μ_max was tested at about 21.44 h in the blank system, whereas the 0.05 mmol L⁻¹, 0.1 mmol L⁻¹, and 0.2 mmol L⁻¹ sulfate ion systems were tested at 19.20 h, 20.10 h, and 21.00 h, respectively. Moreover, the values of μ_max in the blank system was 0.147 h⁻¹, whereas the values were 0.171 h⁻¹, 0.163 h⁻¹ and 0.157 h⁻¹, respectively, in the related systems containing additional sulfate. At low sulfate ion concentration, cell growth was accelerated. In contrast to the situation of 2-HBP, which was eliminated from the cell, part of the sulfate ion remains in the cell as a sulfur source for synthetic cysteine, methionine, vitamins and other material necessary for life, the other part is excreted.³⁵ However, there was a slight inhibition with an increase in sulfate ion concentration, compared to the lower sulfate ion concentration. The reason for this may be that the sulfur source was not the limiting factor and feedback inhibition was gradually generated by excess sulfate ions.^21,25

Fig. 4 Effects of different sulfate ion concentrations on cell growth and desulfurization of Gordonia sp. JDZX13. (A) Specific growth rate; (B) biomass and DBT removal efficiency by growth cells at 60 h; (C) DBT removal efficiency by resting cells at 12 h; and (D) gene transcriptional levels of dszA/dszB/dszC genes. a–c and x–z represent the statistically significant differences (c > b > a; z > y > x).

Effects of SO₄²⁻ on biodesulfurization. The biomass and DBT removal efficiency of growth cell under different sulfate ion densities are listed in Fig. 4B. In the blank and 0.05 mmol L⁻¹ sulfate ion systems, DBT was completely eliminated. The DBT removal rates of the 0.1 mmol L⁻¹ and 0.2 mmol L⁻¹ sulfate ions systems were 95.6% and 15.6%, respectively. These results show that sulfate ion promotes cell growth; however, it inhibits DBT metabolism. This might be due to the fact that excessive sulfate ions inhibit the expression of desulfurization genes or even directly reduce enzyme activity, which result in a lower desulfurization efficiency.^36–38 Moreover, the DBT removal capacity of the resting cells gradually decrease with an increase in sulfate ions (see Fig. 4C). Compared to the blank system, the DBT removal rates of the 0.05, 0.1, and 0.2 mmol L⁻¹ sulfate ions systems decreased by 23.70%, 33.12%, and 37.80%, respectively. These results show that desulfurization enzyme activity was slightly inhibited under higher sulfate ion concentration.

Effects of SO₄²⁻ on transcriptional levels of dszA/dszB/dszC of the resting cells. The effects of the sulfate ion on the transcriptional level of dszA/dszB/dszC of the resting cells are presented in Fig. 4D. To confront the increase in sulfate ions, the transcription of desulfurization genes dszA/dszB/dszC firstly up-regulated and then decreased. Under 0.05 mmol L⁻¹ sulfate ions, the transcriptional levels of the desulfurization gene dszA/dszB/dszC were 3.61, 2.00 and 2.62 times, respectively, its original level. The reason for this might be that more sulfur intermediates were utilized for the synthesis of cellular substances, which sequentially stimulate the expression of the desulfurization gene.^12,39,40 Under 1 mmol L⁻¹ sulfate ion, the transcriptional levels of dszA/dszB/dszC were only 0.33, 0.14 and 0.53 times, respectively, the original level. Different from 2-HBP, the expression of dszB and dszC down-regulated compared to dszA, which indicates that the third and fourth steps were inhibited by sulfate ions. Compared to the equal 2-HBP concentration, the negative effect of sulfate ions on the desulfurization gene was more significant. Under 0.2 mmol L⁻¹ sulfate ion, the transcriptional levels of the desulfurization gene dszA/dszB/dszC severely down-regulated, which were only 0.015, 0.004 and 0.011 times, respectively, the original level. These results also prove that the DBT removal performance of the resting cells was greatly inhibited under a higher sulfate concentration.

Oil/aqueous ration regulation for weakening feedback inhibition of rate-limiting factors

As the major end-products of DBT biodesulfurization, the accumulation of 2-HBP and sulfate ions would form prominent feedback inhibition effects on cell growth and biodesulfurization efficiency. Generally, the feedback inhibition effects of sulfate ions can be weakened by eliminating excess sulfate ions into the extracellular aqueous phase. Meanwhile, considering that 2-HBP is a hydrophobic organic compound in water, the feedback inhibition might be partially alleviated by appropriately introducing an oil phase. For DBT desulfurization under an oil/aqueous two-phase, bacterial cells primarily grew in the aqueous phase, and then a large number of cells gathered at the oil–aqueous interface due to the hydrophobic interaction between the cells.¹⁴ The sulfur-containing compounds were utilized and desulfurization was persistently performed in the oil phase. Excess sulfate ions were excreted into the water phase, whereas the other inhibitor factor 2-HBP entered the oil phase. However, considering the differences in the feedback inhibition effects between 2-HBP and sulfate ions, the oil/aqueous ratios should be regulated to better weaken the dual feedback inhibition effects.

The DBT removal efficiencies with different oil/aqueous ratios are presented in Fig. 5. The DBT removal efficiencies of the aqueous, 2 : 1, 1 : 1, 1 : 2 and 1 : 3 systems were 20.89%, 13.89%, 24.40%, 41.92%, and 44.14%, respectively. Compared to the aqueous phase, although an oil phase was introduced, the removal efficiency of the 2 : 1 (oil/aqueous) system presented a decreasing trend. The reason for this might be that a higher oil/aqueous ratio reduced the effective space of cell growth, resulting in lower biomass for further biodesulfurization.^2,41 Along with the up-regulation of the oil/aqueous ratio, the DBT removal efficiency and productivity gradually increased. This result is also consistent with the transcriptional levels of dszA/dszB/dszC under different 2-HBP and sulfate ion concentrations. Compared to 2-HBP, the accumulation of sulfate ions would generate more negative feedback inhibition effects on desulfurization enzyme activity. However, the highest DBT-removal efficiency was achieved in the 1 : 3 oil/aqueous ratio system (Table 4). Considering the amount of water used, the DBT-removal efficiency of the oil/aqueous ratio 1 : 2 system reached 0.188 mmol L⁻¹ water, which was the highest among all the systems. Thus, the oil/aqueous ratio of 1 : 2 (improved by 100.7%) was determined as the best choice for better balancing the feedback inhibition effects of 2-HBP and sulfate ions. A detailed comparison of the desulfurization efficiency between Gordonia sp. JDZX5 and other common desulphurizing-strains is given in Table 5, which indicates its potential industrial application in DBT biodesulfurization. Based on all the abovementioned results and analyses, the overall mechanism of DBT biodesulfurization is summarized in Fig. 6.

Fig. 5 Effects of oil/water ratio in the two-phase DBT removal efficiency.

Table 4 Comparison of the DBT-removal productivity between different systems

Strategy	Condition	DBT-removal efficacy (×10^−3 mmol L^−1)	DBT-removal efficacy (×10^−3 mmol L^−1 water)	Production of DBT-removal (×10^−3 mmol L^−1 h^−1)	Production of DBT-removal (×10^−3 mmol per L water per h)
a Present not applicable.
Blank	Aqueous	62.67	NA^a	5.22	NA^a
Regulation	2 : 1	41.67	125.01	3.47	10.42
	1 : 1	73.20	146.40	6.10	12.20
	1 : 2	125.30	187.95	10.44	15.66
	1 : 3	132.42	176.56	11.04	14.71

---

Table 5 Comparison of the desulfurization efficiency between Gordonia sp. JDZX5 and other related reported strains

Desulfurization organism	Substrate	Reaction time	Total S reduction (%)	Reference
Lysinibacillus sphaericus DMT-7	0.2 mM DBT	15 d	60%	17
Rhodococcus erythropolis SHT87	0.25 mM DBT	75 h	∼100%	19
Gordonia sp. F.5.25.8	1.0 mM DBT	15 d	73%	27
Microbacterium sp. NISOC-06	1.0 mM DBT	15 d	94.8%	42
Gordonia sp. JDZX5	0.3 mM DBT	47 h	100%	This study

---

Fig. 6 Overall mechanism of DBT the biodesulfurization process by Gordonia sp. JDZX13.

Conclusions

Based on the GC/MS analysis of metabolites and PCR/sequencing of key biodesulfurization genes (dszA/dszB/dszC), the C_X-DBT biodesulfurization process by Gordonia sp. was identified as the “4S” pathway. The effects of key rate-limiting factors (2-HBP and sulfate ions) on μ_x, desulphurization efficiency and transcriptional level of dszA/dszB/dszC were investigated. The feedback inhibition effects of 2-HBP and sulfate ions on cell growth, transcription of dszA/dszB/dszC and biodesulfurization efficiency were significant, especially under sulfate ion stress. Moreover, the DBT removal efficiency was improved by 100.7% by better weakening the feedback inhibition effects. Taken together, these results and mechanism can be applicable to further guide similar biocatalysis processes in the near future.

Acknowledgements

This study was supported by grants from the National Natural Science Foundation of China (No. 21606110), the Natural Science Foundation of Jiangsu Province (No. BK20150133), the funding of Key Laboratory of Industrial Biotechnology, Ministry of Education (KLIB-KF201504), the funding of Key Laboratory of Carbohydrate Chemistry and Biotechnology, Ministry of Education (KLCCB-KF201601), the Priority Academic Program Development of Jiangsu Higher Education Institutions, and the 111 Project (No. 111-2-06).

Notes and references

1. Y. Hou, M. Y. Masoomi, M. Bagheri, A. Morsali and S. W. Joo, RSC Adv., 2015, 5, 81356–81361.
2. V. C. Srivastava, RSC Adv., 2012, 2, 759–783.
3. F. Cui, G. C. Li, X. B. Li, M. H. Lu and M. S. Li, Catal. Sci. Technol., 2015, 5, 549–556.
4. A. Stanislaus, A. Marafi and M. S. Rana, Catal. Today, 2010, 153, 1–68.
5. T. A. Zepeda, B. Pawelec, J. L. G. Fierro and T. Halachev, Appl. Catal., B, 2007, 71, 223–236.
6. L. Alves, S. M. Paixão, R. Pacheco, A. F. Ferreirab and C. M. Silva, RSC Adv., 2015, 5, 34047–34057.
7. B. Pawelec, R. M. Navarro, J. M. Campos-Martin and J. L. G. Fierro, Catal. Sci. Technol., 2011, 1, 23–42.
8. K. Vertin and A. Reek, SAE Technical Paper, 2014, no. 2014-01-1624.
9. F. F. Duan, C. Q. Chen, G. Z. Wang, Y. Z. Yang, X. G. Liu and Y. Qin, RSC Adv., 2014, 4, 1469–1475.
10. H. Y. Zhang, B. S. Guo, Z. L. Hui and M. X. Jian, Chem. Eng. Commun., 2007, 194, 938–945.
11. Y. Nie, C. Li and Z. Wang, Ind. Eng. Chem. Res., 2007, 46, 5108–5112.
12. A. K. Chauhan, A. Ahmad, S. P. Singh and A. Kumar, Int. Biodeterior. Biodegrad., 2015, 104, 105–111.
13. N. Akhtar, M. A. Ghauri, M. A. Anwar and K. Akhtar, FEMS Microbiol. Lett., 2009, 301, 95–102.
14. G. Gün, Y. Yürüm and G. D. Doğanay, Turk. J. Chem., 2015, 39, 255–266.
15. I. Martinez, V. E. Santos, A. Alcon and F. Garcia-Ochoa, Process Biochem., 2015, 50, 119–124.
16. C. Song, Catal. Today, 2003, 86, 211–263.
17. A. Bahuguna, M. K. Lily, A. Munjal, R. N. Singh and K. Dangwal, J. Environ. Sci., 2011, 23, 975–982.
18. D. Pokorna and J. Zabranska, Biotechnol. Adv., 2015, 33, 1246–1259.
19. F. Davoodi-Dehaghani, M. Vosoughi and A. A. Ziaee, Bioresour. Technol., 2010, 101, 1102–1105.
20. S. Aggarwal, I. A. Karimi and G. R. Ivan, Mol. BioSyst., 2013, 9, 2530–2540.
21. M. Soleimani, A. Bassi and A. Margaritis, Biotechnol. Adv., 2007, 25, 570–596.
22. D. Boniek, D. Figueiredo, A. F. B. dosSantos and M. A. deResende Stoianoff, Clean Technol. Environ., 2015, 17, 29–37.
23. I. Torktaz, Z. Etemadifar and P. Derikvand, Bioinformation, 2012, 8, 246–250.
24. J. J. Kilbane II and J. Robins, Appl. Microbiol. Biotechnol., 2007, 75, 843–851.
25. R. Abro, A. A. Abdeltawab, S. S. Al-Deyab, G. Yu, A. B. Qazi, S. R. Gao and X. C. Chen, RSC Adv., 2014, 4, 35302–35317.
26. G. Mohebali and A. S. Ball, Microbiology, 2008, 154, 2169–2183.
27. S. C. C. Santos, D. S. Alviano, C. S. Alviano, M. Padula, A. C. Leitao, O. B. Martins, C. M. S. Ribeiro, M. Y. M. Sassaki, C. P. S. Matta, J. Bevilaqua, G. V. Sebastian and L. Seldin, Appl. Microbiol. Biotechnol., 2006, 71, 355–362.
28. A. Nuhu, Rev. Environ. Sci. Bio/Technol., 2013, 12, 9–23.
29. X. Ma, L. Sun and C. Song, Catal. Today, 2002, 77, 107–116.
30. L. Alves and S. M. Paixao, Bioresour. Technol., 2011, 102, 9162–9166.
31. X. Zhan, S. S. Feng, L. Zhang, X. Huang, J. D. Tang, Q. Q. Wu and H. L. Yang, Microbiology, 2016, 43, 1171–1180.
32. I. B. W. Gunam, K. Yamamura, I. N. Sujaya, N. S. Antara, W. R. Aryanta, M. Tanaka, F. Tomita, T. Sone and K. Asano, J. Microbiol. Biotechnol., 2013, 23, 473–482.
33. J. Calzada, A. Alcon, V. E. Santos and F. Garcia-Ochoa, Biochem. Eng. J., 2012, 66, 52–60.
34. W. Ismail, W. S. El-Sayed, A. S. Abdul Raheem, M. E. Mohamed and A. M. El Nayal, Front. Microbiol., 2016, 7, 266–279.
35. A. Caro, K. Boltes, P. Letón and E. García-Calvo, Biodegradation, 2008, 19, 599–611.
36. A. Caro, K. Boltes, P. Letón and E. García-Calvo, Biochem. Eng. J., 2007, 35, 191–197.
37. A. Abin-Fuentes, M. E.-S. Mohamed, D. I. C. Wang and K. L. J. Prather, Appl. Environ. Microbiol., 2013, 79, 7807–7817.
38. B. L. McFarland, D. J. Boron, W. Deever, J. A. Meyer, A. R. Johnson and R. M. Atlas, Crit. Rev. Microbiol., 1998, 24, 99–147.
39. W. Li, M. D. Wang, H. Chen, J. M. Chen and Y. Shi, Biotechnol. Lett., 2006, 28, 1175–1179.
40. J. J. Kilbane II, Curr. Opin. Biotechnol., 2006, 17, 305–314.
41. L. Alves and S. M. Paixão, Bioresour. Technol., 2011, 102, 9162–9166.
42. M. Papizadeh, M. R. Ardakani, G. Ebrahimipour and H. Motamedi, World J. Microbiol. Biotechnol., 2010, 26, 1195–1200.

---