The LuxI/LuxR-Type Quorum Sensing System Regulates Degradation of Polycyclic Aromatic Hydrocarbons via Two Mechanisms

Members of the Sphingomonadales are renowned for their ability to degrade polycyclic aromatic hydrocarbons (PAHs). However, little is known about the regulatory mechanisms of the degradative pathway. Using cross-feeding bioassay, a functional LuxI/LuxR-type acyl-homoserine lactone (AHL)-mediated quorum sensing (QS) system was identified from Croceicoccus naphthovorans PQ-2, a member of the order Sphingomonadales. Inactivation of the QS system resulted in a significant decrease in PAHs degradation. The QS system positively controlled the expression of three PAH-degrading genes (ahdA1e, xylE and xylG) and a regulatory gene ardR, which are located on the large plasmid. Interestingly, the transcription levels of these three PAH-degrading genes were significantly down-regulated in the ardR mutant. In addition, bacterial cell surface hydrophobicity and cell morphology were altered in the QS-deficient mutant. Therefore, the QS system in strain PQ-2 positively regulates PAH degradation via two mechanisms: (i) by induction of PAH-degrading genes directly and/or indirectly; and (ii) by an increase of bacterial cell surface hydrophobicity. The findings of this study improve our understanding of how the QS system influences the degradation of PAHs, therefore facilitating the development of new strategies for the bioremediation of PAHs.


Introduction
Quorum sensing (QS) is a process of bacterial cell-cell communication that controls many important population-level behaviors, such as bioluminescence, biofilm formation, antibiotic resistance, and virulence factor production [1][2][3]. Bacteria produce and release signal molecules whose concentration accumulates as bacterial population density increases. When the signal molecules pass a specific threshold, QS alters global patterns of gene expression. In general, Gram-negative bacteria communicate using acyl-homoserine lactones (AHLs) as signal molecules [4]. AHLs are composed of a homoserine-lactone ring and a 4~18 carbon acyl chain that is occasionally modified by an oxo-or hydroxyl group at the 3-C position. The canonical AHL-mediated QS system was firstly discovered in Vibrio fischeri and then identified in many other Gram-negative bacteria. This system consists of two components, LuxI-type and LuxR-type proteins. The LuxI-type proteins are AHL synthases that catalyze the synthesis of AHLs, while LuxR-type proteins are transcription factors responsible for the perception of AHLs. LuxR-family proteins possess an N-terminal AHL-binding domain and a C-terminal DNA-binding domain. In the absence of AHLs, most LuxR proteins are 2. Results 2.1. Identification of an AHL-Type QS System in C. naphthovorans PQ-2 C. naphthovorans PQ-2 can produce AHL molecules, but genes responsible for the AHL-type QS systemremain unidentified. According to the genome annotation information from NCBI, we found a pseudogene with a length of 515 bp (AB433_RS00085) that was annotated to encode an autoinducer synthase. This pseudogene contains only a partial coding region. In contrast, the full-length sequence of the gene (locus tag: Ga0111307_123190) can be obtained from the JGI database. This gene has a length of 657 bp and was designated luxI. Notably, a transposase gene (AB433_RS00065) and an integrase gene (AB433_RS00100) lie in the flanking sequence of the luxI gene.
To determine whether the gene product of luxI is indeed an AHL synthase, the luxI gene was expressed in the Escherichia coli BL21 (DE3) strain that cannot produce any AHLs [18]. The Agrobacterium tumefaciens A136 was used as a biosensor strain to detect AHLs with medium and long acyl chains [19]. As expected, the culture extract of E. coli harboring the empty plasmid did not affect the biosensor strain. However, the recombinant E. coli sample induced a blue coloration on the indicator plate, which results from the expression of β-galactosidase in A. tumefaciens A136 ( Figure 1A). TLC analysis showed that the strain PQ-2 produces three AHLs, C6-HSL and C8-HSL with an oxo- 3 of 15 or hydroxyl group at the 3-C position, and an unknown AHL that needs to be further investigated ( Figure 1B). Similar results were also observed in the crude AHLs extract from the recombinant E. coli cells ( Figure 1B). Int oxo-or hydroxyl group at the 3-C position, and an unknown AHL that needs to be further investigated ( Figure 1B). Similar results were also observed in the crude AHLs extract from the recombinant E. coli cells ( Figure 1B). C8-HSL and acidified ethyl acetate were used as positive control and negative control (NC), respectively. (B) Identification of AHLs by TLC. M1~M3, standard AHLs; S1, AHLs extracted from PQ-2; S2, AHLs extracted from E. coli BL21 expressing luxI. (C) Cross-feeding bioassay for detection of AHLs in C. naphthovorans strains. ΔluxI C and ΔluxR C represent the complemented version of ΔluxI and ΔluxR, respectively. (D,E) Relative expression levels of luxI (D) and luxR (E). The transcription level of the wild type was defined as 1.0. Experiments were performed in three biological replicates, and similar trends were observed. The representative data from three separate experiments are shown. Statistical significance of differences was analyzed by t-test; **, p< 0.01; ***, p< 0.001.
To further confirm whether the luxI gene is active in C. naphthovorans PQ-2, an rpsL-based markerless gene deletion system for Sphingomonads was employed to construct the ∆luxI strain [20]. Cross-feeding bioassay results showed that the ∆luxI strain failed to induce a blue coloration of A. tumefaciens A136, while expression of the luxI gene under its native promoter regained the ability to generate a blue coloration ( Figure 1C). Therefore, the luxI gene is responsible for the synthesis of AHLs in C. naphthovorans PQ-2.
Given that the expression of luxI depends on a LuxR-type autoinducer-responsive regulator, the cognate gene for LuxR was identified from the PQ-2 genome. The locus tag Ga0111307_123182, which locates 6.3 kb upstream of the luxI gene, was predicted to encode a LuxR homolog. Accordingly, we renamed this gene luxR. Similar to the strain lacking luxI, the ∆luxR strain no longer induced a blue coloration of the biosensor strain, which could be restored by genetic complementation ( Figure 1C). In addition, the transcription of luxI was significantly reduced in the ∆luxR strain ( Figure 1D), while the transcription of luxR was down-regulated in the ∆luxI strain ( Figure 1E). Both transcription levels were restored by complementation ( Figure 1D,E). Collectively, we can conclude that C. naphthovorans PQ-2 possesses a functional AHL-type QS system, which is composed of an AHL synthase LuxI and an autoinducer-responsive regulator LuxR. Identification of the LuxI/LuxR QS system. (A) Analysis of AHLs using the biosensor strain A. tumefaciens A136. AHLs were extracted from culture supernatants of C. naphthovorans PQ-2, E. coli BL21 harboring pET-28b(+) (BL21/pET28b) or pET-28b(+) expressing luxI (BL21/pET28b-luxI). C8-HSL and acidified ethyl acetate were used as positive control and negative control (NC), respectively. (B) Identification of AHLs by TLC. M1~M3, standard AHLs; S1, AHLs extracted from PQ-2; S2, AHLs extracted from E. coli BL21 expressing luxI. (C) Cross-feeding bioassay for detection of AHLs in C. naphthovorans strains. ∆luxI C and ∆luxR C represent the complemented version of ∆luxI and ∆luxR, respectively. (D,E) Relative expression levels of luxI (D) and luxR (E). The transcription level of the wild type was defined as 1.0. Experiments were performed in three biological replicates, and similar trends were observed. The representative data from three separate experiments are shown. Statistical significance of differences was analyzed by t-test; **, p < 0.01; ***, p < 0.001.
To further confirm whether the luxI gene is active in C. naphthovorans PQ-2, an rpsL-based markerless gene deletion system for Sphingomonads was employed to construct the ∆luxI strain [20]. Cross-feeding bioassay results showed that the ∆luxI strain failed to induce a blue coloration of A. tumefaciens A136, while expression of the luxI gene under its native promoter regained the ability to generate a blue coloration ( Figure 1C). Therefore, the luxI gene is responsible for the synthesis of AHLs in C. naphthovorans PQ-2.
Given that the expression of luxI depends on a LuxR-type autoinducer-responsive regulator, the cognate gene for LuxR was identified from the PQ-2 genome. The locus tag Ga0111307_123182, which locates 6.3 kb upstream of the luxI gene, was predicted to encode a LuxR homolog. Accordingly, we renamed this gene luxR. Similar to the strain lacking luxI, the ∆luxR strain no longer induced a blue coloration of the biosensor strain, which could be restored by genetic complementation ( Figure 1C). In addition, the transcription of luxI was significantly reduced in the ∆luxR strain ( Figure 1D), while the transcription of luxR was down-regulated in the ∆luxI strain ( Figure 1E). Both transcription levels were restored by complementation ( Figure 1D,E). Collectively, we can conclude that C. naphthovorans PQ-2 possesses a functional AHL-type QS system, which is composed of an AHL synthase LuxI and an autoinducer-responsive regulator LuxR.

The AHL-Type QS System Is Crucial for the Degradation of PAHs
To investigate the degradation of PAHs, C. naphthovorans PQ-2 was dropped onto minimal medium plates containing different PAHs (fluorene, phenanthrene, anthracene, fluoranthene, and pyrene) as the sole carbon and energy source. As shown by the results in Figure 2A, growth of PQ-2 was observed on all the plates tested, suggesting that this bacterium can use various PAHs as the sole carbon and energy source. Notably, PQ-2 had the highest degradation ability for phenanthrene and anthracene, which forms a halo around bacterial colonies. Phenanthrene was then used in subsequent studies.

The AHL-Type QS System Is Crucial for the Degradation of PAHs
To investigate the degradation of PAHs, C. naphthovorans PQ-2 was dropped onto minimal medium plates containing different PAHs (fluorene, phenanthrene, anthracene, fluoranthene, and pyrene) as the sole carbon and energy source. As shown by the results in Figure 2A, growth of PQ-2 was observed on all the plates tested, suggesting that this bacterium can use various PAHs as the sole carbon and energy source. Notably, PQ-2 had the highest degradation ability for phenanthrene and anthracene, which forms a halo around bacterial colonies. Phenanthrene was then used in subsequent studies. To explore the involvement of the AHL-type QS system in the degradation of PAHs, growth of the ∆luxI and ∆luxR strains was measured in liquid culture media. Deletion of luxI and luxR per se did not affect bacterial growth in nutrient-rich P5Y3 medium ( Figure S1). However, in minimal medium with phenanthrene as the sole carbon and energy source, the colony-forming units (CFUs) of strains lacking either luxI or luxR were significantly reduced when compared to the wild type ( Figure 2B). CFUs of the ∆luxI strain were restored by exogenous addition of AHLs, which were extracted from the wild type of PQ-2. On the contrary, the addition of AHLs was unable to restore the CFUs of the ΔluxR strain. The residual phenanthrene in the supernatants was analyzed by HPLC. Both ∆luxI and ∆luxR strains displayed an obvious reduction in phenanthrene degradation ( Figure 2C,D). The degradation ability of the ∆luxI strain but not the ΔluxR strain was almost The representative data from three separate experiments are shown. Statistical significance of differences was analyzed by t-test: *, p < 0.05; **, p < 0.01; ***, p < 0.001.
To explore the involvement of the AHL-type QS system in the degradation of PAHs, growth of the ∆luxI and ∆luxR strains was measured in liquid culture media. Deletion of luxI and luxR per se did not affect bacterial growth in nutrient-rich P5Y3 medium ( Figure S1). However, in minimal medium with phenanthrene as the sole carbon and energy source, the colony-forming units (CFUs) of strains lacking either luxI or luxR were significantly reduced when compared to the wild type ( Figure 2B). CFUs of the ∆luxI strain were restored by exogenous addition of AHLs, which were extracted from the wild type of PQ-2. On the contrary, the addition of AHLs was unable to restore the CFUs of the ∆luxR strain.
The residual phenanthrene in the supernatants was analyzed by HPLC. Both ∆luxI and ∆luxR strains displayed an obvious reduction in phenanthrene degradation ( Figure 2C,D). The degradation ability of the ∆luxI strain but not the ∆luxR strain was almost reversed to the wild type when supplemented with the AHLs extracted from the PQ-2 ( Figure 2C). Moreover, the expression of either luxI or luxR in the corresponding mutants partially recovered the ability of phenanthrene degradation ( Figure 2D). These results collectively indicate that the AHL-mediated QS system plays an important role in the degradation of phenanthrene in C. naphthovorans PQ-2.

PQ-2 Degrades Phenanthrene through the Salicylic Acid Pathway
The common microbial pathway for PAHs degradation involves the formation of catechol, which will then be transformed into the tricarboxylic acid (TCA) cycle [8]. To determine the degradation pathway of phenanthrene in C. naphthovorans PQ-2, the degradation products were analyzed by GC-MS ( Figure S2). After degradation by strain PQ-2 for 48 h, the samples were separated into neutral and acidic fractions. Phenanthrene was detected in the neutral fraction, which results from the incomplete degradation. By contrast, the acidic metabolites of phenanthrene contain salicylic acid and catechol, suggesting that the strain PQ-2 also transforms phenanthrene to the common intermediate catechol.

The Large Plasmid Is Responsible for the Degradation of Phenanthrene
The complete genome of C. naphthovorans PQ-2 is composed of a chromosome (3.54 Mb) and two plasmids, P1 (0.19 Mb) and P2 (0.13 Mb). Sequence analysis demonstrated that the large plasmid P1 possesses a putative PAH-degrading cluster (from AB433_RS17995 to AB433_RS18220), which shares high identities with several aromatic compound degradation gene clusters in other Sphingomonas strains ( Figure 3A). For example, the nucleotide sequence from AB433_RS18095 to AB433_RS18120 of PQ-2 has 95% identity with the phnIJKN cluster of Sphingomonas sp. 14DN-61, which has been confirmed to be involved in PAHs degradation [21]. Similar to the aforementioned luxI gene, several transposase genes are located in the flanking sequence of the putative PAH-degrading cluster. reversed to the wild type when supplemented with the AHLs extracted from the PQ-2 ( Figure 2C). Moreover, the expression of either luxI or luxR in the corresponding mutants partially recovered the ability of phenanthrene degradation ( Figure 2D). These results collectively indicate that the AHL-mediated QS system plays an important role in the degradation of phenanthrene in C. naphthovorans PQ-2.

PQ-2 Degrades Phenanthrene through the Salicylic Acid Pathway
The common microbial pathway for PAHs degradation involves the formation of catechol, which will then be transformed into the tricarboxylic acid (TCA) cycle [8]. To determine the degradation pathway of phenanthrene in C. naphthovorans PQ-2, the degradation products were analyzed by GC-MS ( Figure S2). After degradation by strain PQ-2 for 48 h, the samples were separated into neutral and acidic fractions. Phenanthrene was detected in the neutral fraction, which results from the incomplete degradation. By contrast, the acidic metabolites of phenanthrene contain salicylic acid and catechol, suggesting that the strain PQ-2 also transforms phenanthrene to the common intermediate catechol.

The Large Plasmid Is Responsible for the Degradation of Phenanthrene
The complete genome of C. naphthovorans PQ-2 is composed of a chromosome (3.54 Mb) and two plasmids, P1 (0.19 Mb) and P2 (0.13 Mb). Sequence analysis demonstrated that the large plasmid P1 possesses a putative PAH-degrading cluster (from AB433_RS17995 to AB433_RS18220), which shares high identities with several aromatic compound degradation gene clusters in other Sphingomonas strains ( Figure 3A). For example, the nucleotide sequence from AB433_RS18095 to AB433_RS18120 of PQ-2 has 95% identity with the phnIJKN cluster of Sphingomonas sp. 14DN-61, which has been confirmed to be involved in PAHs degradation [21]. Similar to the aforementioned luxI gene, several transposase genes are located in the flanking sequence of the putative PAH-degrading cluster.
To explore whether the large plasmid P1 is involved in phenanthrene degradation, PQ-2 was treated with rifampicin to eliminate the plasmid. The strain lacking the P1 (ΔP1) was screened in the presence of 20 μg/mL rifampicin. Compared to the wild type, the ΔP1 strain was no longer able to degrade phenanthrene ( Figure 3B), indicating that the plasmid P1 is required for phenanthrene degradation.  To explore whether the large plasmid P1 is involved in phenanthrene degradation, PQ-2 was treated with rifampicin to eliminate the plasmid. The strain lacking the P1 (∆P1) was screened in the presence of 20 µg/mL rifampicin. Compared to the wild type, the ∆P1 strain was no longer able to degrade phenanthrene ( Figure 3B), indicating that the plasmid P1 is required for phenanthrene degradation.

The AHL-Type QS System Regulates the Expression of PAH-Degrading Genes
To investigate how the AHL-mediated QS system regulates phenanthrene degradation in C. naphthovorans PQ-2, we analyzed the transcription levels of genes within the PAH-degrading gene cluster in the ∆luxI strain. Based on gene annotation and the degradative pathway of phenanthrene in PQ-2, eight PAH-degrading genes within the cluster were chosen for qRT-PCR analysis, including ahdA2e, ahdA1e, bphC, xylG, xylE, xylX, xylY and ahdA2c. Among them, the transcription levels of three PAH-degrading genes (ahdA1e, xylE and xylG) were significantly down-regulated in the ∆luxI strain, which could be fully (ahdA1e and xylE) or partially (xylG) rescued when supplemented with exogenous AHLs extracted from PQ-2 broth ( Figure 4A). The ahdA1e gene encodes the α subunit of the aromatic ring-hydroxylating dioxygenase, while the xylE and xylG genes encode a catechol 2, 3 dioxygenase (C23O) and a 2-hydroxymuconic semialdehyde dehydrogenase, respectively. Notably, the promoter region of all three PAH-degrading genes possess the conserved "lux box" element.
We also determined the expression of AB433_RS18080 gene (now designated as ardR for aromatic degradation regulator), which is predicted to encode a sigma-54-dependent Fis family transcriptional regulator. As shown by the results in Figure 4A, the expression of ardR was dramatically reduced in the absence of luxI, which could be partially restored when exogenous addition of the PQ-2 AHLs extract. These results indicate that ArdR may be involved in the regulation of PAHs degradation. To confirm, the ardR gene was deleted from the plasmid P1 and then the resultant mutant (ΔardR) was subjected to phenanthrene degradation assay ( Figure 4B). Deletion of the ardR gene led to a significant decrease in the degradation of phenanthrene, which could be relieved by the expression of the ardR gene in trans. The upstream sequence of ardR also contains a conserved "lux box" sequence. Interestingly, the transcription levels of these three PAH-degrading genes were sharply reduced in the ΔardR strain when compared to the wild type ( Figure 4C). Relative expression of ahdA1e but not xylE and xylG was partially restored in the complemented version of ardR (ΔardR C ) in trans. We also determined the expression of AB433_RS18080 gene (now designated as ardR for aromatic degradation regulator), which is predicted to encode a sigma-54-dependent Fis family transcriptional regulator. As shown by the results in Figure 4A, the expression of ardR was dramatically reduced in the absence of luxI, which could be partially restored when exogenous addition of the PQ-2 AHLs extract. These results indicate that ArdR may be involved in the regulation of PAHs degradation. To confirm, the ardR gene was deleted from the plasmid P1 and then the resultant mutant (∆ardR) was subjected to phenanthrene degradation assay ( Figure 4B). Deletion of the ardR gene led to a significant decrease in the degradation of phenanthrene, which could be relieved by the expression of the ardR gene in trans. The upstream sequence of ardR also contains a conserved "lux box" sequence. Interestingly, the transcription levels of these three PAH-degrading genes were sharply reduced in the ∆ardR strain when compared to the wild type ( Figure 4C). Relative expression of ahdA1e but not xylE and xylG was partially restored in the complemented version of ardR (∆ardR C ) in trans.

The AHL-Type QS System Affects Bacterial Cell Surface Hydrophobicity
Bacterial cell surface hydrophobicity is crucial for the degradation of organic pollutants in the environment [15,22]. To determine whether the AHL-type QS system affects cell surface properties in C. naphthovorans PQ-2, microbial adherence to hydrocarbons (MATH) assay was performed to determine cell surface hydrophobicity ( Figure 5A). Compared to the wild type, deletion of the luxI gene led to a~2-fold decrease in the cell surface hydrophobicity, which could be recovered by expression of the luxI gene in trans. To further investigate this, we measured the content of glycosphingolipids by Enzyme-linked immunosorbent assay (ELISA), but no significant change was observed in all tested strains ( Figure S3). We next determined the content of hydrophobic proteins on the cell surface using the fluorescent probe bis-ANS ( Figure 5B). The ∆luxI strain displayed a significant decrease in the bis-ANS fluorescence when compared to the wild type, while expression of the luxI gene restored the fluorescence of the complemented strain to the level of the wild type. Therefore, the ∆luxI strain has lower amount of hydrophobic proteins than the wild type on the cell surface. (B) Percentage of residual phenanthrene after three days degradation by the ΔardR and its complemented version (ΔardR C ). (C) The relative transcriptional levels of PAH-degrading genes in the ΔardR. Experiments were performed in three biological replicates, and similar trends were observed. The representative data from three separate experiments are shown. Statistical significance of differences was analyzed by t-test: *, p< 0.05; **, p< 0.01; ***, p< 0.001.

The AHL-Type QS System Affects Bacterial Cell Surface Hydrophobicity
Bacterial cell surface hydrophobicity is crucial for the degradation of organic pollutants in the environment [15,22]. To determine whether the AHL-type QS system affects cell surface properties in C. naphthovorans PQ-2, microbial adherence to hydrocarbons (MATH) assay was performed to determine cell surface hydrophobicity ( Figure 5A). Compared to the wild type, deletion of the luxI gene led to a ~2-fold decrease in the cell surface hydrophobicity, which could be recovered by expression of the luxI gene in trans. To further investigate this, we measured the content of glycosphingolipids by Enzyme-linked immunosorbent assay (ELISA), but no significant change was observed in all tested strains ( Figure S3). We next determined the content of hydrophobic proteins on the cell surface using the fluorescent probe bis-ANS ( Figure 5B). The ΔluxI strain displayed a significant decrease in the bis-ANS fluorescence when compared to the wild type, while expression of the luxI gene restored the fluorescence of the complemented strain to the level of the wild type. Therefore, the ΔluxI strain has lower amount of hydrophobic proteins than the wild type on the cell surface.
Cell morphology was also observed by scanning electron microscopy (SEM) ( Figure 5C). Results showed that the cell surface morphology of the ΔluxI strain was quite different from that of the wild type. The wild type strain exhibited a rough cell surface with many extracellular matrix components, whereas the ΔluxI strain had a smooth cell surface without extracellular matrix components.   For (A,B), experiments were performed in three biological replicates, and similar trends were observed. The representative data from three separate experiments are shown. Statistical significance of differences was analyzed by t-test: **, p < 0.01; ***, p < 0.001. Figure 6. Model for QS regulation on phenanthrene degradation in C. naphthovorans PQ-2. The QS system positively regulates PAH degradation via multiple mechanisms. The LuxR-AHL complex may bind to the "lux box" and then directly regulate the transcriptional levels of three PAH-degrading genes, including ahdA1e, xylE and xylG. The LuxR-AHL complex may also indirectly modulate the expression of these PAH-degrading genes, which is mediated by the transcriptional regulator ArdR. Moreover, the QS system enhances bacterial cell surface hydrophobicity (CSH), thus affecting the attachment of cells to PAHs. Similar to other AHL-type QS systems, the LuxR-AHL complex in PQ-2 has a positive feedback effect on the expression of luxR and luxI. For clarity, the secretion and uptake of AHLs are omitted in the model.
The transcription of the three PAH-degradative genes is also regulated by the ardR gene, which is located within the PAH-degrading gene cluster. The transcriptional regulators for PAHs degradation have been reported in several Pseudomonas and Novosphingobium species [13,28,29]. It is also possible that the AHL-type QS system in C. naphthovorans PQ-2 indirectly modulates the expression of PAH-degrading genes, which is mediated by the transcriptional regulator ArdR ( Figure 6). The expression of PAH-degrading genes may be regulated in a hierarchical manner. The AHL-mediated QS system is at the apex of the regulatory cascade, while the specific transcriptional regulator sets up the second cascade driving the expression of PAH-degrading genes.
In addition to up-regulation of the degradative enzymes, our results show that the AHL-type QS system in C. naphthovoransPQ-2 also positively controls bacterial cell surface hydrophobicity. Many studies have demonstrated that there is a positive correlation between cell surface hydrophobicity and PAHs degradation [15,22,30,31]. PAH-degrading bacteria normally have high cell surface hydrophobicity, which stimulates the direct attachment of bacterial cells to hydrophobic Figure 6. Model for QS regulation on phenanthrene degradation in C. naphthovorans PQ-2. The QS system positively regulates PAH degradation via multiple mechanisms. The LuxR-AHL complex may bind to the "lux box" and then directly regulate the transcriptional levels of three PAH-degrading genes, including ahdA1e, xylE and xylG. The LuxR-AHL complex may also indirectly modulate the expression of these PAH-degrading genes, which is mediated by the transcriptional regulator ArdR. Moreover, the QS system enhances bacterial cell surface hydrophobicity (CSH), thus affecting the attachment of cells to PAHs. Similar to other AHL-type QS systems, the LuxR-AHL complex in PQ-2 has a positive feedback effect on the expression of luxR and luxI. For clarity, the secretion and uptake of AHLs are omitted in the model.
The Sphingomonadaceae, the largest family within the order Sphingomonadales, are renowned for their ability to degrade recalcitrant compounds and xenobiotics including PAHs [10,24]. C. naphthovorans PQ-2 belongs to the family Erythrobacteraceae, which is the second largest family within the order Sphingomonadales [17]. Our results demonstrate that C. naphthovorans PQ-2 can degrade PAHs containing three and four rings, such as phenanthrene, anthracene, fluorene, fluoranthene and pyrene. Among them, PQ-2 has the highest degradation ability for three-ring molecules phenanthrene and anthracene. These two PAHs are known to be more susceptible to biodegradation, since they are more volatile and more soluble in water [25]. It is unclear whether PQ-2 is capable of using PAHs with five or more than five rings.
The genes involved in PAHs degradation in the Sphingomonadaceae are very often located on large plasmids (megaplasmids) [10,23,26]. Consistently, our results show that the large plasmid P1 in C. naphthovorans PQ-2 is responsible for the degradation of phenanthrene. This megaplasmid possesses a PAH-degrading gene cluster which is composed of 32 open reading frames. There are several transposase genes located in the flanking regions of the gene cluster, implying that this cluster may be transferred horizontally from other bacteria. The degradative pathway for phenanthrene in C. naphthovorans PQ-2 is also in agreement with that in the members of the family Sphingomonadaceae [8,10,27]. Under aerobic conditions, the aromatic ring-hydroxylating dioxygenase (including the α subunit AhdA1e and β subunit AhdAle2) catalyzes the initial reaction in the degradative pathway, resulting in the modification of the aromatic ring. After several steps, C. naphthovorans PQ-2 degrades phenanthrene to the common intermediate catechol. Subsequently, the catechol 2, 3 dioxygenase XylE and 2-hydroxymuconic semialdehyde dehydrogenase XylG convert catechol to 2-hydroxymuconate, which will be further transformed to TCA cycle ( Figure 6).
Although the research on regulation of the PAH-degrading pathways is limited, the existing studies demonstrated that the AHL-type QS system is involved in the regulation of PAHs degradation process in P. aeruginosa CGMCC1.860 [13]. The AHL-type QS system in this bacterium directly and indirectly controls the expression of catechol 2,3-dioxygenase gene, resulting in the enhanced aromatics biodegradation [13]. Similar results were observed in our studies on C. naphthovorans PQ-2. The AHL-type QS system in C. naphthovorans PQ-2 positively regulates the expression of three PAH-degrading genes (including ahdA1e, xylE and xylG), as well as the regulatory gene ardR. It is worth mention that the upstream regions of these genes contain a conserved "lux box" sequence (5 -NNCT-N 12 -AGNN-3 ). Therefore, the LuxR-AHL complex may bind to the "lux box" site and then directly regulate the expression of these target genes, thus affecting phenanthrene degradation ( Figure 6).
The transcription of the three PAH-degradative genes is also regulated by the ardR gene, which is located within the PAH-degrading gene cluster. The transcriptional regulators for PAHs degradation have been reported in several Pseudomonas and Novosphingobium species [13,28,29]. It is also possible that the AHL-type QS system in C. naphthovorans PQ-2 indirectly modulates the expression of PAH-degrading genes, which is mediated by the transcriptional regulator ArdR ( Figure 6). The expression of PAH-degrading genes may be regulated in a hierarchical manner. The AHL-mediated QS system is at the apex of the regulatory cascade, while the specific transcriptional regulator sets up the second cascade driving the expression of PAH-degrading genes.
In addition to up-regulation of the degradative enzymes, our results show that the AHL-type QS system in C. naphthovorans PQ-2 also positively controls bacterial cell surface hydrophobicity. Many studies have demonstrated that there is a positive correlation between cell surface hydrophobicity and PAHs degradation [15,22,30,31]. PAH-degrading bacteria normally have high cell surface hydrophobicity, which stimulates the direct attachment of bacterial cells to hydrophobic surface including PAHs and the partition of dissolved PAHs to the cell surface [22]. As a result, PAHs enter into bacterial cell for biodegradation. Thus, the QS system of C. naphthovorans PQ-2 facilitates the attachment of bacterial cells to PAHs and then enhances the uptake of PAHs. Once PAHs enter into the cytoplasm of the bacterial cell, they will induce the expression of PAH-degrading genes and then the degradation of PAHs (Figure 6). Due to the presence of glycosphingolipids, the cell surface of sphingomonads is more hydrophobic than those of other bacteria [10]. However, the enhanced cell surface hydrophobicity in C. naphthovorans PQ-2 is not derived from glycosphingolipids, but related to hydrophobic proteins on the cell surface. There are many factors that can affect cell surface hydrophobicity. For example, it has been reported that biosurfactants secreted by Bacillus subtilis and P. aeruginosa enhance bacterial cell surface hydrophobicity, resulting in higher uptake and use of pyrene [22,32]. The detailed mechanism in C. naphthovorans PQ-2 needs to be further investigated.
In contrast to our results, a recent study demonstrated that the AHL-type QS system negatively regulates phenanthrene removal in N. pentaromativorans US6-1, a member belongs to the family Sphingomonadaceae [15]. Deletion of the AHL-type QS system in US6-1 increases phenanthrene removal efficiency. This phenomenon is also related to PAH-degrading genes, since the relative expression levels of 12 PAH-degrading genes are up-regulated in QS-deficient mutants. Interestingly, QS-deficient mutants have significantly higher cell surface hydrophobicity, which partially results from the increase of hydrophobic glycosphingolipids. Therefore, the two AHL-type QS systems play an opposite role in the regulation of PAHs degradation in N. pentaromativorans US6-1 and C. naphthovorans PQ-2. At present, the underlying mechanism remains unclear. It should be noted that N. pentaromativorans US6-1 was isolated from muddy sediment [33], while C. naphthovorans PQ-2 was isolated from marine biofilm [17]. Biofilm-based bioremediation is one of the most efficient approaches for the decontamination of pollutants [14,34]. Thus, C. naphthovorans PQ-2 has a great potential in bioremediation of PAHs.

Expression of AHL Synthase of Strain PQ-2 in E. coli
The full-length sequence of luxI encoding AHL synthase was obtained from the JGI database (locus tag: Ga0111307_123190). PCR amplification was performed with the primers containing restriction enzyme sites (Table S1). The resulting PCR product was cloned into the expression vector pET-28b(+), and then the recombinant plasmid pET-28b(+)-luxI was transformed into E. coli BL21 (DE3). Finally, the recombinant vector was verified by sequencing.

Extraction and Analysis of AHLs
The AHLs produced by E. coli and C. naphthovorans were extracted from culture supernatants as previously described [36,37]. 100 mL of overnight bacterial culture was centrifuged at 8000 rpm for 10 min. The supernatant was extracted three times with an equal volume of acidified ethyl acetate. The combined extract from the organic phase was evaporated on a rotatory evaporator at 50 • C. The dried residue was dissolved into 1 mL of acidified ethyl acetate and stored at −20 • C.

Cross-Feeding Bioassay
The AHLs produced by bacteria were measured by cross-feeding bioassay as described elsewhere [16,36]. The tested bacteria and the biosensor A. tumefaciens A136 were streaked in parallel on LB-P5Y3 plate (containing equal amounts of LB and P5Y3) supplemented with X-gal (40 µg/mL). The cross-feeding plate was then cultivated at 30 • C for 12 h before color visualization.
For analysis of AHL extracts, 2 µL of each AHL extract was dropped onto a sterile filter paper on the indicator plate. The plate was incubated for 12 h at 30 • C and then photographed. C8-HSL and acidified ethyl acetate were used as positive control and negative control, respectively.

Markerless Gene Deletion Mutagenesis and Complementation
The gene deletion mutant of C. naphthovorans PQ-2 was constructed by an rpsL-based markerless gene deletion system [20] with a little modification. In brief, two fragments flanking the target gene were amplified by PCR with primers containing the restriction enzyme sites and then were joined by fusion PCR (Table S1). The fusion fragment and the suicide plasmid pAK405 were ligated together by T4 DNA ligase and transformed into E. coli WM3064 (DAP auxotroph) [35]. The resulting plasmid was transferred into C. naphthovorans PQ-2 via conjugation. Integration of the mutagenesis constructs into the chromosome was selected by resistance to kanamycin and verified by PCR. The correct transconjugant was grown in P5Y3 broth for three days and then plated onto the P5Y3 plate supplemented with streptomycin. Kanamycin-sensitive and streptomycin-resistant colonies were screened by PCR for deletion of the targeted gene. Finally, the deletion mutations were verified by sequencing.
The broad-host-range plasmid pBBR1MCS-5 was used for the genetic complementation of mutants [36]. A fragment containing the gene of interest and its native promoter was amplified by PCR and then cloned into pBBR1MCS-5. The resulting recombinant plasmid was then transferred into the corresponding mutant via conjugation.

RNA Isolation and Quantitative Real-Time PCR (qRT-PCR)
Total RNA from C. naphthovorans strains was isolated using the RNAiso Plus Kit (TaKaRa, Dalian, China) according to the manufacturer's instructions. The qRT-PCR analysis was performed with a CFX Connect Real-Time PCR Detection System (BioRad, Hercules, CA, USA) as described previously [36]. The cycle threshold (C T ) values for each gene of interest were normalized against the C T values of the 16S rRNA gene. The relative expression level of each gene of interest was determined from three replicates in a single experiment.

Analysis of Phenanthrene and Its Metabolites
For phenanthrene analysis, the sample was extracted with ethyl acetate and dried under vacuum. Then it was dissolved in acetonitrile, filtrated, and subjected to HPLC analysis [38]. The column for the measurement was InertsilODS-3 C18 column (4.6 × 250 mm, 5 µm), and the mobile phase was methanol/water (90:10) at a flow rate of 1.0 mL/min using a Waters 2487 dual-wavelength detector.
The phenanthrene metabolites were analyzed by gas chromatography-mass spectrometry (GC-MS), as previously described [39,40]. For brief, phenanthrene was allowed to degrade for 48 h. The culture was centrifuged, and the supernatant was then acidified to pH 2.3 with HCl and extracted with ethyl acetate. The organic phase was extracted three times with an equal volume of NaOH (10 mmol/L). The remaining organic phase was dried over anhydrous sodium sulfated and concentrated to 5mLof ethyl acetate (neutral fraction). The aqueous phase was acidified to pH 2.3 with HCl and extracted with ethyl acetate (acidic fraction). The GC-MS analysis of neutral and acidic fractions was carried out with/without derivatization.

Microbial Adhesion to Hydrocarbons (MATH) Test
The MATH test was carried out to determine bacterial cell surface hydrophobicity as described previously [15,41] with some modifications. After incubation, bacterial cultures were resuspended in phosphate-buffered saline (PBS) and the OD 600 values were measured (OD 0 ). Then, 4 mL of xylene was added to 8 mL of the suspension. The mixtures were vortexed for 15 s, and allowed phase separation for 20 min at room temperature. Subsequently, the OD 600 of the aqueous phase of the suspension was measured (OD 1 ). The cell surface hydrophobicity values were calculated according to the following equation: Cell surface hydrophobicity (%) = [(OD 0 − OD 1 )/OD 0 ] × 100%. (1)

Determination of Hydrophobic Proteins Using Fluorescent Dye
The fluorescent probe bis (8-anilinonaphthalene-1-sulfonate) (bis-ANS) was used to determine hydrophobic proteins [42]. Bacterial cultures were resuspended in PBS (pH = 7.4) and the bacterial cell density was adjusted to 0.15 at OD 600 . 30 µL bis-ANS was added to 170 µL of these samples to achieve a final concentration of 5 µmol/L. The fluorescence intensity of bis-ANS was measured by the UV-Vis spectrophotometer (Molecular Devices SpectraMax M4, Sunnyvale, CA, USA). The excitation and emission wavelengths were 385 nm and 530 nm, respectively.

Scanning Electron Microscopy (SEM)
C. naphthovorans strains were cultivated in marine minimal medium supplemented with phenanthrene. 1 mL of bacterial culture was centrifuged at 4000 rpm for 5 min. The pellet was fixed with 4% (v/v) glutardialdehyde overnight at 4 • C, washed with PBS and dehydrated by passage through a graded ethanol series (30%, 50%, 70%, 80%, 90%, and 100%, v/v), subsequently dehydrated with liquid carbon dioxide and coated with platinum powder. The bacterial cell morphology was observed with SEM (Hitachi SU8010, Tokyo,Japan) [43].

Conclusions
Sphingomonads are biotechnologically interesting organisms for their potential in the bioremediation of PAHs. This study indicates that the AHL-mediated QS system in C. naphthovorans PQ-2 positively regulates the degradation of PAHs via two mechanisms: (i) by induction of PAH-degrading genes expression directly and/or indirectly and (ii) by an increase of cell surface hydrophobicity. The findings of this study improve our understanding of the involvement of QS regulation in the degradation of PAHs in Sphingomonads, therefore facilitating the development of new strategies for the bioremediation of PAHs.

Conflicts of Interest:
The authors declare no conflict of interest.