The Cyclic AMP Receptor Protein, Crp, Is Required for the Decolorization of Acid Yellow 36 in Shewanella putrefaciens CN32

Shewanella shows good application potentials in the decolorization and detoxification of azo dye wastewater. However, the molecular mechanism of decolorization is still lacking. In this study, it was found that Shewanella putrefaciens CN32 exhibited good decolorization ability to various azo dyes, and a global regulatory protein cAMP receptor protein (Crp) was identified to be required for the decolorization of acid yellow 36 (AY) by constructing a transposon mutant library. Then, the molecular mechanism of AY decolorization regulated by Crp was further investigated. RT-qPCR and electrophoretic mobility shift assay (EMSA) results showed that Crp was able to directly bind to the promoter region of the cymA gene and promote its expression. Riboflavin acting as an electron shuttle could accelerate the AY decolorization efficiency of S. putrefaciens CN32 wild-type (WT) but did not show a promoting effect to Δcrp mutant and ΔcymA mutant, further confirming that Crp promotes the decolorization through regulating electron transport chains. Moreover, the mutant with cymA overexpression could slightly enhance the AY decolorization efficiency compared with the WT strain. In addition, it was found that MtrA, MtrB, and MtrC partially contribute to the electron transfer from CymA to dye molecules, and other main electron transport chains need to be identified in future experiments. This study revealed the molecular mechanism of a global regulator Crp regulating the decolorization of azo dye, which is helpful in understanding the relationship between the decolorization and other metabolic processes in S. putrefaciens CN32.


INTRODUCTION
With the rapid development of the textile, printing, and dyeing industries, more than 700,000 tons of commercial synthetic dyes are produced worldwide each year (Guo et al., 2020b), of which azo dyes account for 60-70% (Hameed and Ismail, 2018;Kong et al., 2018), mainly due to their properties of low cost, easy synthesis, high coloring efficiency, and good stability to various oxidizing agents. The dyeing industries consume huge amounts of azo dyes, and about 10-15% of dye molecules that are not effectively bound to clothing is discharged into the surrounding environment in the form of dyeing wastewater (Zahir et al., 2014;Zhang et al., 2019), causing serious environmental pollution (Yesilada et al., 2003). Dyeing wastewater not only affects the transparency of the surrounding water but also poses a serious threat to the ecological environment and even human health because azo dye molecules and their degradation products have strong mutagenic, teratogenic, and carcinogenic effects (Guo et al., 2020a;Samir et al., 2020). The most remarkable characteristic of azo dyes is that their molecules contain one or more azo groups (Yan et al., 2012). Because of their stubborn structural properties, azo dyes are very stable in nature and difficult to degrade (Liu W. et al., 2017). Therefore, decolorization and detoxification are very necessary before the discharge of azo dye-containing wastewater (Liu et al., 2017b).
Physical, chemical, and biological strategies are generally applied to treat azo dye-containing wastewater (Pandey et al., 2007). Compared with physicochemical methods, which are limited by high energy consumption and secondary pollution, biological methods attracted more and more attention due to their advantages of high decolorization efficiency, low operation cost, and environmental friendliness (Saratale et al., 2011;Samir et al., 2020;Wang et al., 2020). The mechanism of biological decolorization of azo dyes mainly includes bioflocculation (Liu et al., 2009), biological adsorption (Saratale et al., 2011), electron reduction , and enzymatic degradation (Dawkar et al., 2010;Baweja et al., 2016;Yang et al., 2018Yang et al., , 2020. Moreover, the decolorization efficiency of azo dyes can also be enhanced by the combined use of different biological methods at the same time. For example, the degradation efficiency of insoluble Sudan red can be accelerated by the synergistic effect of enzyme-catalyzed biodegradation and non-specific reductive decolorization . In addition, much effort has been made to improve the decolorization efficiency of azo dyes (Imran et al., 2016). For example, some electron shuttles, such as riboflavin and methylene blue, have been found to be able to promote electron transfer from the cell surface to the dye molecules, thereby accelerating the biodegradation of azo dyes (Liu et al., 2016).
Many strains have been reported to be able to decolorize azo dyes; among them, Shewanella strains are the most concerned species due to their excellent decolorization performance Liu et al., 2016Liu et al., , 2018. Shewanella, as a species of facultative anaerobic bacteria with remarkable respiratory pathways, is able to utilize various terminal electron acceptors under anaerobic conditions, including various pollutants, such as azo dyes and heavy metal ions (Fredrickson et al., 2008;Ding et al., 2014). Therefore, Shewanella strains show good potential in the field of environmental remediation. In Shewanella, many components in the electron transfer pathway are necessary for its decolorization ability Xiao et al., 2012;Liu et al., 2016). In the cytomembrane, the tetraheme c-type cytochrome CymA receives electrons from the quinone pool and then transfers them to multiple respiratory pathways, such as Mtr, Dms, NarfA/B, and NrfA pathways (Schwalb et al., 2003;Gao et al., 2009). In Shewanella oneidensis MR-1, the Mtr respiratory pathway, including MtrA, MtrB, MtrC, and OmcA, has been found to be involved in decolorization processes of various azo dyes Xiao et al., 2012). However, studies on the molecular mechanism of the decolorization of azo dyes by Shewanella are still lacking, especially on the global regulatory factors related to the decolorization of azo dyes.
In this study, cAMP receptor protein (Crp), a global transcription regulator (Gao et al., 2012), was found to be essential for the decolorization of azo dyes in Shewanella putrefaciens CN32 through the construction of a transposon mutant library and the selection of mutants with different decolorization abilities and transposon locations. And then, the molecular mechanism of the decolorization of azo dyes regulated by Crp was also investigated. We found that, with the assistance of cAMP, Crp was able to directly bind to the promoter region of the cymA gene and promote its expression, thereby promoting decolorization through regulating electron transport chains. Thus, this study has a certain theoretical significance for revealing the molecular mechanism of the decolorization of azo dyes by Shewanella strains.

Bacterial Strains, Plasmids, Primers, and Culture Conditions
The strains and plasmids used or constructed in this study are shown in Table 1, and the primers used in this study are listed in Table 2. Escherichia coli and S. putrefaciens CN32 strains were grown aerobically at 37 • C and 30 • C, respectively, in Luria-Bertani (LB) medium which contains tryptone 10 g/L, yeast extract 5 g/L, and NaCl 10 g/L, for genetic manipulation. If necessary, kanamycin of 50 µg/ml was added into the medium. The decolorization medium for S. putrefaciens CN32 consisted of yeast extract 2 g/L, NH 4 Cl 1 g/L, NaCl 0.5 g/L, Na 2 HPO 4 ·12H 2 O 7.52 g/L, NaH 2 PO 4 ·2H 2 O 7.13 g/L, and filter-sterilized sodium lactate 20 mM. The transformants were screened on the LKT medium, which is based on the LB medium added with 50 µg/ml of kanamycin and 20 µg/ml of potassium tellurite.

Decolorization Assay of Azo Dyes
S. putrefaciens CN32 was cultured in LB medium for 12 h to obtain the seed culture and then inoculated into 60 ml serum vials containing 50 ml of the decolorization medium with azo dyes to an initial OD 600 of 0.1. To ensure anaerobic condition, decolorization systems were purged with nitrogen gas for 2 min. Subsequently, serum vials were sealed with rubber stoppers immediately and then cultured in an incubator at 30 • C without shaking for decolorization. The decolorization samples of different times were centrifuged at 10,000 rpm, 25 • C for 1 min. The absorbance of the supernatant was determined at 497 nm for Congo red, 520 nm for amaranth, 414 nm for acid yellow 36 (AY), 465 nm for methyl orange, 618 nm for amino black, and 714 nm for naphthol green to determine the concentration of residual azo dyes. The decolorization rate (%) was calculated based on the following equation: (ODx − ODy)/ODx × 100%, where ODx and ODy refer to the absorbance of the initial and decolorized

Transposon Mutagenesis and Location of Transposon Insertion Sites
S. putrefaciens CN32 (recipient strain) was mixed with E. coli UQ3022 (donor strain), which carries a plasmid pRL27 containing a mini-Tn5 transposon and a kanamycin resistance gene (Larsen et al., 2002). Both recipient strain and donor strain were cultured overnight in LB medium. The donor strain was washed with LB medium twice and mixed with the recipient strain in a 1:1 ratio. The mixture of these two strains was spotted on a solid LB agar plate and incubated at 30 • C for 8 h. Then the cells were scraped off from the agar plate and plated onto an LB agar plate with 50 µg/ml of kanamycin and 20 µg/ml of potassium tellurite (to inhibit the cell growth of the donor strain) and incubated at 30 • C for 36 h. Single black colonies presented were purified, and their decolorization abilities were determined in the decolorization medium with 200 mg/L AY. The mutants with significantly altered decolorization capacity were selected and preserved. Then, the genomic DNA of these mutants was extracted for localization of the transposon insertion site. The extracted genomic DNA was self-linked after digestion by BamHI or SpeI and then transformed into E. coli UQ3021 (Larsen et al., 2002) and selected on the LB agar plate added with 50 µg/ml of kanamycin. Plasmids extracted from the transformant were sequenced using primer Tn5-seqF and Tn5-seqR and BLAST in NCBI to map the location of the mini-Tn5 transposon. The target genes inserted by the mini-Tn5 transposon were deleted to confirm its function in regulating AY decolorization in S. putrefaciens CN32.

Construction of Deletion Mutants and Complementation Strains
The in-frame deletion of mutants for crp, mtrA, mtrB, mtrC, undA, and cymA was performed based on the principle of homologous recombination. The primer positions for the target gene in-frame deletion are shown in Figure 1A, and the genomic arrangement of mtrA, mtrB, mtrC, undA, and cymA in S. putrefaciens CN32 is shown in Figure 1B. To delete the target gene, approximately 1,000-bp fragments upstream and downstream the targeted gene were amplified with respective primers D5F/D5R and D3F/D3R, using the S. putrefaciens CN32 genome DNA as a PCR template. These two fragments were ligated into the suicide vector pK19mobsacB after being digested by the corresponding restriction endonuclease (Schäfer et al., 1994). Then, the constructed plasmid was transformed into E. coli DH5α and introduced from E. coli DH5α into S. putrefaciens CN32 wild type (WT) using a helper plasmid, pRK2013, by triparental conjugation (Figurski and Helinski, 1979). The conjugation experiment was performed according to a previous study (Liu et al., 2017a). Briefly, the donor and recipient strains were conjugated in a 1:1 ratio and spotted on an LB agar plate. LKT medium plates were used to screen for    were amplified with primers MtrC-D5F1/MtrC-D5R1 and MtrB-D3F1/MtrB-D3R1, respectively, and for the construction of undA mtrCAB, upstream and downstream fragments were amplified with primers UndA1-D5F1/UndA1-D5R1 and MtrB-D3F1/MtrB-D3R1, respectively.
To construct a complementation plasmid, a DNA fragment carrying the ribosome binding site (RBS) and open reading frame (ORF) of the targeted gene was amplified by primers CF/CR using the S. putrefaciens CN32 genome as a PCR template and then ligated with the plasmid pBBR1MCS-2-P aacC1 with the corresponding restriction site (Liu et al., 2017a). And then the complementation plasmid was introduced into corresponding mutants for complementation or into S. putrefaciens CN32 WT for overexpression.

RNA Extraction and Real-Time RT-PCR Analysis
In order to analyze the regulatory effect of Crp on the cymA gene, the seed samples of WT and crp mutant were inoculated into the decolorization medium with 200 mg/L AY and cultured for 4 h. The cells were collected by centrifugation at 10,000 rpm, 4 • C for 5 min, and RNA was extracted by the TRIzol reagent (Tiangen, China) according to the protocol provided by the manufacturer. Real-time RT-PCR was performed using primers CymA-QF/CymA-QR, with SYBR Green Master Mix (Biosharp, China) according to the specification from the manufacturer. Signal intensities of PCR products were standardized to those of the 16S rRNA gene amplified with primers 16S QF/16S QR. The experiments were performed with at least three replicates.

Purification of His 6 -Crp Protein
To prepare the His 6 -Crp protein, the coding region of crp was amplified by PCR using primers Crp-EF/Crp-ER. The PCR product was digested with BamHI/HindIII and cloned into pET-28a (+) to generate pET28-crp. The resulting plasmid was confirmed by sequencing and then transformed The initial concentration of azo dyes is 50 mg/L. The images of azo dye solution obtained before decolorization and after 24 h of decolorization. The cell pellets were collected from decolorization solution using a 2 ml tube with 10,000 rpm.
into E. coli BL21(DE3) for overexpression of His 6 -Crp (six-histidine tag on the N-terminus of the Crp protein). After induction with 0.4 mM isopropyl β-D-thiogalactoside (IPTG) at 16 • C for 20 h, the soluble recombinant His 6 -Crp protein was purified by Ni-agarose resin chromatography (CoWin Biosciences, China).

Electrophoretic Mobility Shift Assay (EMSA)
DNA probe-carrying promoter regions of cymA were PCRamplified by using primers CymA-PF and CymA-PR. The PCR products purified from agarose gel were labeled with digoxigenin (DIG) using terminal transferase. The 3 -terminal DIG-labeled probe of 0.15 nM was incubated with various quantities of His 6 -Crp and 1 µM cAMP in a binding reaction. EMSA experiments were operated as described previously (Sun et al., 2016). To confirm specificity of protein-DNA interaction, a 300-fold excess of unlabeled specific probe or non-specific DNA was added into the binding mixture before incubation.

DNase I Footprinting
DNase I footprinting assay was carried out to determine the binding site of Crp in the promoter region of cymA. PCR was conducted using 5 -terminal FAM-labeled forward primer CymA-FootF and regular reverse primer CymA-FootR, and the PCR products were purified to obtain FAM-labeled footprinting probes. The mixtures of 20-µl volume containing various concentrations of purified His 6 -Crp, 300-ng probes, and 2 µM of cAMP were incubated at 25 • C for 40 min to achieve the binding reaction of the Crp protein with the probe. Then 1 U of DNase I (NEB, United States) was added to the mixture and incubated at 37 • C for 10 s. The DNase I treatment was terminated by adding 10 µl of 0.5 M EDTA solution and heated at 80 • C for 10 min. The DNA fragments were purified and capillary sequenced in a 3730XL DNA Genetic Analyzer (ABI, United States). The data were processed and analyzed with the GeneMarker program, v2.2.0.

Effect of Exogenous Riboflavin on Decolorization
In order to analyze the effect of exogenous riboflavin on AY decolorization of S. putrefaciens CN32 WT, crp, and cymA, the strains were cultured in LB medium to an OD 600 of around 1.0 and used as seed culture. Then, the cells in seed culture were collected by centrifugation at 25 • C, 8,000 rpm for 5 min, and were washed once using the decolorization base medium (decolorization medium without adding yeast extract). The washed cells were inoculated into 60 ml serum vials containing 50 ml of the decolorization base medium with 200 mg/L of AY to an initial OD 600 of 0.1. After purging with nitrogen gas for 2 min, the decolorization ability of WT, crp, and cymA with and without adding 2 µM riboflavin was determined at different time intervals.

Decolorization of Various Azo Dyes by S. putrefaciens CN32
The decolorization ability of S. putrefaciens CN32 to various azo dyes was analyzed under anaerobic condition ( Table 3). The results showed that S. putrefaciens CN32 was able to decolorize all the tested azo dyes within 48 h. Especially to AY (Figure 2) and methyl orange, more than 90% of decolorization rates were achieved within 8 h. As shown in Figure 2, the scanning spectrum of AY and its degradation products in a range from 240 to 720 nm was investigated. It can be seen that the characteristic absorption peak at around 414 nm disappeared after decolorization of 4 h, suggesting that S. putrefaciens CN32 shows good potential in the treatment of azo dye-containing wastewater. In previous studies, several Shewanella strains have been used to decolorize azo dyes, such as S. oneidensis MR-1 (Liu et al., 2016;Li et al., 2018), Shewanella decolorationis S12 (Xu et al., 2007), Shewanella sp. RQs-106 (Zhou et al., 2018), Shewanella aquimarina (Meng et al., 2012), and Shewanella algae (Meng et al., 2014). In S. oneidensis MR-1, a transmembrane electron transport chain Mtr respiratory pathway, which includes cytochromes MtrC and OmcA and related proteins MtrA and MtrB, has been proven to be responsible for the anaerobic decolorization of azo dyes . However, the further regulatory mechanism of the decolorization of azo dyes in Shewanella species is largely unknown. In this study, AY was selected for investigating the molecular mechanism of azo dye decolorization by S. putrefaciens CN32 due to its very high decolorization efficiency.

Crp Promotes the Decolorization Ability of S. putrefaciens CN32 Under Anaerobic Conditions
To identify the underlying decolorization mechanisms of AY by S. putrefaciens CN32 in anaerobic respiration, approximately 1,000 mini-Tn5 transposon-inserted mutants screened by a kanamycin agar plate were selected for decolorization ability analysis using 60 ml serum vials. Compared with the CN32 WT strain, the decolorization efficiencies of 10 transposon-inserted mutants were significantly altered (more than 50% change), in which three mutants that exhibited poor decolorization ability to AY were inserted into different sites of the crp gene (sputcn32_0652), suggesting that Crp may be involved in biodecolorization. Thus, an in-frame deletion mutant of the crp gene ( crp) and a complementation strain (C-crp) were constructed. As shown in Figure 3A, the anaerobic decolorization efficiency of the crp mutant to AY was seriously decreased compared with the WT, and the C-crp mutant was obviously restored to the WT level. At the same time, the growth of WT, crp mutant, and C-crp mutant was determined under the anaerobic decolorization condition. The results (Figure 3B) showed that the WT, crp mutant, and C-crp mutant did not show obvious cell growth during the anaerobic AY decolorization process. Therefore, these results indicated that Crp promotes AY decolorization in S. putrefaciens CN32. Crp is a global regulatory factor which regulates different metabolic processes in many bacteria by forming complexes with cAMP; for example, Crp is involved in carbon catabolite repression in many bacteria (Deutscher et al., 2006); in E. coli, Crp regulates many stress responses to protect cells from harmful environments including starvation and osmotic shock (Battesti et al., 2011;Kalia et al., 2013); in Pseudomonas aeruginosa, Crp (Vfr) regulates biofilm formation through controlling type IV pili (Beatson et al., 2002;Persat et al., 2015). In this study, it was found that Crp is necessary for biological decolorization in S. putrefaciens CN32.
CymA Is Necessary for the Anaerobic Decolorization of AY in S. putrefaciens CN32 Except for the crp gene, two transposon-inserted mutants with a significant decrease in decolorization efficiency were identified with different insertion sites in the cymA gene (sputcn32_0286). CymA, a c-type cytochrome, is a component of the Mtr respiratory pathway, which is a critical transmembrane electron transfer channel in dissimilatory metal-reducing strains . A previous study showed that the cymA mutant of S. oneidensis MR-1 almost lost complete decolorization capability . Genome analysis found that the amino acid sequence of CymA in S. putrefaciens CN32 exhibits 95.7% similarity to that in S. oneidensis MR-1. To confirm the function of CymA in S. putrefaciens CN32 during AY decolorization, an in-frame gene deletion mutant cymA and a complementation strain C-cymA were constructed. The anaerobic decolorization ability of the cymA mutant was seriously decreased compared with that of the WT strain ( Figure 4A); the C-cymA mutant was obviously restored to the WT level. And the WT, cymA mutant, and C-cymA mutant exhibited similar cell growth (Figure 4B), suggesting that CymA is necessary for the anaerobic decolorization of AY in S. putrefaciens CN32. As the cymA mutant showed a similar phenotype to the crp mutant, a double mutant crp cymA was constructed to analyze their relationship. The results showed that the crp cymA mutant exhibited a similar decolorization efficiency to the crp and  (Lanes 1, 2, 3, 4, and 5). A 300-fold-excess unlabeled specific competitor (Probe cymA, Lane S) and non-specific competitor (Lane N) were added to the mixture to perform specific or non-specific competition assays, respectively. Lanes A and B: Reaction mixtures without addition of cAMP used as negative controls. cymA mutants, which is obviously lower than that of the WT strain, indicating that Crp and CymA may be involved in the same decolorization pathway.

Crp Directly Activates the Transcription of cymA
Crp is a critical global transcriptional regulator in prokaryotes through forming a complex with cAMP. In S. oneidensis MR-1, the complex of Crp and cAMP can regulate the transcription of multiple cytochrome c genes including omcA and mtrC (Kasai et al., 2015). However, the relationship between Crp and the cymA gene is still unclear. To determine whether Crp regulates the transcription of the cymA gene in S. putrefaciens CN32, real-time reverse-transcription PCR (RT-qPCR) was carried out (Figure 5A). Compared with that in the WT strain, the cymA expression level in the crp mutant was significantly decreased, and almost no cymA expression was detected in the crp FIGURE 6 | The effect of exogenous riboflavin on AY decolorization. Values are means ± SD (n = 3). The cells were cultured in decolorization medium with (+) or without (-) adding 2 µM riboflavin. The initial AY concentration of is 200 mg/L. Significance analysis of decolorization rates of the same strain with and without riboflavin was performed using Student's t-test; asterisks **p < 0.01, and NS indicates not significant. mutant, indicating that Crp plays a critical role in activating the expression of the cymA gene. To identify the regulation that Crp exerts on the transcription of the cymA gene, EMSA was performed to investigate whether Crp directly binds to the upstream regions of the cymA gene. When a labeled DNA probe containing the region from −183 to −477 bp upstream of cymA was incubated with the Crp-cAMP complex, shifted bands were observed ( Figure 5B). When there is no cAMP in the mixture, no shifted band was observed ( Figure 5B). These findings suggested that the Crp-cAMP complex can bind to the promoter region of the cymA gene. Taken together, Crp directly activates the transcription of the cymA gene through forming a complex with cAMP. To further investigate the mechanism by which the Crp-cAMP complex activates cymA, a DNase I footprinting assay was carried out to identify the precise binding site of the Crp-cAMP complex at the upstream regions of the cymA gene, and a protected region from −375 to −338 nt upstream of the cymA start codon was revealed ( Figure 5C). Subsequently, the transcription start site (TSS) of cymA was predicted using an online BDGM promoter prediction tool 1 . A possible TSS was found to be a "G" located at 300 nt upstream of the cymA start codon ( Figure 5D). Based on the above results, it is possible that the Crp-cAMP complex activates cymA expression by recruiting the RNA polymerase to the promoter region of cymA.

Effect of Exogenous Riboflavin on Decolorization
Previous studies have reported that flavins produced from Shewanella genus (Marsili et al., 2008) and other chemical substances, such as methylene blue (Liu et al., 2016) and humic acids (Liu et al., 2011), were able to act as shuttles to accelerate the electron transfer from the cell surface to pollutant molecules, 1 https://fruitfly.org/seq_tools/promoter.html  such as azo dyes and heavy metal ions. In order to analyze the effect of electron shuttle on AY decolorization by S. putrefaciens CN32, 2 µM riboflavin was added into the decolorization systems of WT, crp, and cymA. The results showed that riboflavin could significantly improve the decolorization efficiency of the WT strain, but no significant promotion effect was observed for the crp mutant and cymA mutant (Figure 6). These results further demonstrated that the regulation of Crp to the AY decolorization efficiency is achieved by regulating the members of electron transport chains including CymA.

Analysis of the Electron Transport Pathway From CymA to AY
CymA is an electron transfer hub supporting multi-branched respiratory chains (Marritt et al., 2012). Recent studies showed that the Mtr respiratory pathway of S. oneidensis MR-1 not only participates in the dissimilatory reduction of multiple metal ions (Toporek et al., 2019) but also plays a critical role in the decolorization process of a variety of textile dyes . Genome analysis found that an mtr-like gene cluster exists in the genome of S. putrefaciens CN32, including MtrC, MtrA, MtrB, and UndA, and their amino acid sequences exhibit 53.5, 88.6, 83.4, and 26.6% similarity to MtrC, MtrA, MtrB, and OmcA, respectively, in S. oneidensis MR-1. To determine whether the Mtr respiratory pathway is involved in AY decolorization in S. putrefaciens CN32, the in-frame deletion mutants mtrC, mtrA, mtrB, and undA and the triple mutants mtrCAB and undA mtrCAB were constructed. As shown in Figure 7, mtrC, mtrA, mtrB, undA, mtrCAB, and undA mtrCAB showed 10. 6, 9.8, 14.5, 3.8, 22.8, and 17.6% decreases in the decolorization efficiency of AY at 4 h, respectively, indicating that the Mtr respiratory pathway only partially contributes to the transfer of electron required for the AY decolorization process in S. putrefaciens CN32. This is consistent with the results previously reported . At the same time, we also noticed that the blocking of the Mtr pathway could not completely inhibit the AY decolorization efficiency and that the influence of the Mtr pathway on the decolorization of MR-1 was more significant than that of CN32. In S. putrefaciens CN32, the decolorization efficiency of Mtr mutants (except cymA mutant) decreased only about 20% compared with the WT strain; however, in S. oneidensis MR-1, a decrease of more than 60% in decolorization efficiency was reported , indicating that MtrA, MtrB, and MtrC only partially contribute to the AY decolorization in S. putrefaciens CN32. This is similar to the latest published result, which showed that MtrA, MtrB, and MtrC did not play a major role in the decolorization of methyl orange by S. putrefaciens CN32 under microaerobic conditions in 96-well plates (Min et al., 2020). Thus, S. putrefaciens CN32 exhibits a more complex electron transfer process than did S. oneidensis MR-1 in the decolorization of azo dyes. Therefore, it is necessary to identify other main electron transfer pathways from the CymA electron transfer hub to azo dye acceptors in future studies.

Mechanism of Crp Regulating the AY Decolorization in S. putrefaciens CN32
A molecular mechanism model of Crp regulating AY biodecolorization was proposed in Figure 8. Crp and cAMP form a complex and then directly activate the transcription of cymA; MtrA, MtrB, and MtrC partially contribute to electron transfer from cells to AY acceptors, and other major electron transfer pathways from CymA to dye molecules need to be identified in future studies; flavin acts as a shuttle to accelerate the electron transfer from cells to dye molecules through the switch between the oxidation state and reduction state. AY dye molecules are decomposed into colorless degradation products under the action of electron reduction. Based on the above molecular regulation mechanism, we tried to improve the decolorization ability of S. putrefaciens CN32 by overexpression of cymA in the WT strain (O-cymA mutant). The result showed that the decolorization ability of the O-cymA mutant was slightly improved compared to that of the WT strain. The above results indicate that Crp is necessary to activate the expression of cymA, thereby promoting AY decolorization through accelerating electron transfer from cells to dye molecules.

CONCLUSION
In this study, AY was selected as an electron acceptor to reveal the molecular mechanism of S. putrefaciens CN32 decolorization of azo dyes. By constructing a transposon mutant library, the cAMP receptor protein Crp was identified as a necessary regulator for AY decolorization in S. putrefaciens CN32. Crp can directly bind to the promoter region of the cymA gene and activate the expression of the cymA gene, thereby supporting AY decolorization. MtrA, MtrB, and MtrC partially contribute to the electron transfer from CymA to AY molecules, and other major electron transfer pathways need to be identified in future studies. Furthermore, the overexpression of cymA could slightly enhance the decolorization efficiency of AY in S. putrefaciens CN32. This study will help us understand the molecular mechanism of azo dye decolorization in other Shewanella strains.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

AUTHOR CONTRIBUTIONS
WL and CL designed the research. WL and YC performed decolorization experiments. YC and XZ operated gene deletion and complementation experiments. JL and JZ carried out protein expression and purification experiments. YC, SW, and DS performed RNA extraction, RT-qPCR and DNase I footprinting assays. WL, CL, and DS wrote and revised the manuscript. All authors contributed to the article and approved the submitted version.