Metagenomics-Based Discovery of Malachite Green-Degradation Gene Families and Enzymes From Mangrove Sediment

Malachite green (MG) is an organic contaminant and the effluents with MG negatively influence the health and balance of the coastal and marine ecosystem. The diverse and abundant microbial communities inhabiting in mangroves participate actively in various ecological processes. Metagenomic sequencing from mangrove sediments was applied to excavate the resources MG-degradation genes (MDGs) and to assess the potential of their corresponding enzymes. A data set of 10 GB was assembled into 33,756 contigs and 44,743 ORFs were predicted. In the data set, 666 bacterial genera and 13 pollutant degradation pathways were found. Proteobacteria and Actinobacteria were the most dominate phyla in taxonomic assignment. A total of 44 putative MDGs were revealed and possibly derived from 30 bacterial genera, most of which belonged to the phyla of Proteobacteria and Bacteroidetes. The MDGs belonged to three gene families, including peroxidase genes (up to 93.54% of total MDGs), laccase (3.40%), and p450 (3.06%). Of the three gene families, three representatives (Mgv-rLACC, Mgv-rPOD, and Mgv-rCYP) which had lower similarities to the closest sequences in GenBank were prokaryotic expressed and their enzymes were characterized. Three recombinant proteins showed different MG-degrading activities. Mgv-rPOD had the strongest activity which decolorized 97.3% of MG (300 mg/L) within 40 min. In addition, Mgv-rPOD showed a more complete process of MG degradation compared with other two recombinant proteins according to the intermediates detected by LC-MS. Furthermore, the high MG-degrading activity was maintained at low temperature (20°C), wider pH range, and the existence of metal ions and chelating agent. Mgv-rLACC and Mgv-rCYP also removed 63.7% and 54.1% of MG (20 mg/L) within 24 h, respectively. The results could provide a broad insight into discovering abundant genetic resources and an effective strategy to access the eco-friendly way for preventing coastal pollution.


INTRODUCTION
The application of synthetic dyes is extensive in the fields of textile, pharmaceutical, cosmetics, paper making, solar cells, and acrylic industries (Chang and Lin, 2001). Due to various risks on environment and human health, strategies for the removal of dyes have attracted increasing attention from health professionals and environmentalists (Srivastava et al., 2004;Venil and Lakshmanaperumalsamy, 2010;Gopinathan et al., 2015). Malachite green (MG), in particular, is one typical of synthetic dyes and is widely used in dyeing of silk, leather, and paper, which has been shown to harm humans and animals because of toxicity, mutagenicity, and carcinogenicity (Srivastava et al., 2004;Gopinathan et al., 2015). MG is a triphenylmethane dye and highly soluble in water (Yong et al., 2015). MG can be metabolized into leucomalachite green (LMG) (Yong et al., 2015) and remains in fish muscles, fat, and organs with a half-life of about 10 days and even longer (Chen and Miao, 2010). MG is also persistent in the environment and its half-life in sediment can range from 12.9 to 50.34 days (Jiang, 2011). Despite the global ban on MG in aquaculture, 1-15% of dyes are estimated being discharged within effluents, which seriously inhibit survival, development, and reproductive of aquatic organisms (Chang et al., 2016). The removal or reduction of MG can be processed by chemicals methods or biological treatments. Chemicals methods, such as photodegradation (Algubury, 2016;Pathania et al., 2016;Liang et al., 2017), have some disadvantages of high costs and secondary pollutant generation (Tayabali et al., 2017). Bioremediation strategy may constitute an alternative approach to conventional physicochemical methods, benefiting from the potential of indigenous microorganisms to metabolize anthropogenic compounds (Grosser et al., 1991;Tayabali et al., 2017).
Mangrove ecosystems constitute 60-70% of the coastline in the tropical and subtropical regions on Earth, which receive nutrient-rich aquaculture effluent from nearby farming activities and accumulated organic contaminnat from industrial wastewater discharge (Gomes Gomes et al., 2008;Giri et al., 2011). The highly productive and diverse microbial community living in tropical and subtropical mangrove ecosystems continuously transforms varied nutrients into sources of nitrogen, phosphorus, and other nutrients that can be used by the plants (Ouyang et al., 2017;Yang et al., 2017). Due the ability to absorb waste and pollutants, mangroves are considered as significant sinks for pollutants from freshwater discharges as well as from contaminated tidal water (Marchand et al., 2016). Therefore, mangrove sediments are suitable for exploring MG degrading microorganisms because of input of carbon in the form of litter which then acts as a substrate for decomposition by microbe. The bioremediation potential of microorganisms isolated from hydrocarbon-contaminated environments was as active as or even higher than those originating from non-contaminated sediments (Jones et al., 2011). For now, a few culturable microorganisms with the degradation activity of dyestuff pollutant are isolated from mangrove sediment microbiome. Two strains, P1 and D1, isolated from mangrove sediments previously, were found with the decolorization ability against a variety of dyes, including nitomill brill crimson, methyl red, and nitro green B (Srinivasan et al., 2014). A mangrove-derived strain, Aplanochytrium sp., was reported with MG degrading activity which could remove 86.32% of MG within 5.5 days (Gomathi et al., 2013). Despite global advancement in understanding the microbial diversity in mangrove sediments, more than 90% of environmental microorganisms remain unculturable (Amann et al., 1990;Schloss and Handelsman, 2005;Kimura, 2006). MG degradation potential could be difficult to be assessed and utilized by using traditional culture-dependent method.
Nowadays, culture-independent metagenomic library has successfully used for discovering of novel biosynthetic gene from diverse environments. Lac15, a laccase isolated from marine bacterial metagenome, could degrade several industrial dyes belonged to reactive azo class under alkalescent conditions (Fang et al., 2011). Besides, a novel bacterial laccase Lac21 was isolated from metagenomic library of the South China Sea. Lac21 could remove 80% of Reactive Deep Blue M-2GE (50 mg/L) within 24 h (Fang et al., 2012). However, the reliance on relatively low-throughput of clone libraries combined with activity-based screening could limit screening output (Pope and Moran, 2010;Hess et al., 2011). In addition, metagenomic library focuses on the individual functional gene, not conveniently provides the capacity and diversity of the functional gene in a certain environmental sample. Sequence-based screening of metagenomics combined with the databases, such as Kyoto Encyclopedia of Genes and Genomes (KEGG), Cluster of Orthologous Groups of Proteins (COG), etc., provides a high-performance method on screening the sequences and abundance of potential functional genes (Hugenholtz and Tyson, 2008;Simon and Daniel, 2009). For now, several MDGs have been isolated from MG-degrading strains and metagenomics libraries, including laccase (LACC) (Murugesan et al., 2009), peroxidase (POD) (Ulson de Souza et al., 2007), cytochrome p450 (CYP) (Jefferson and Jones, 2003), triphenylmethane reductase (TMD) (Kim et al., 2008) and triphenylmethane dye oxidase (TpmD) (Ren et al., 2006). However, MDGs resources and abilities from different environment, such as mangrove sediment, were still cryptical. Moreover, the enzymatic properties of MDGs from varied gene families, such as degrading activity, stability against complex conditions, and more radical degradation products, are still strongly desired.
In this study, the culture independent metagenomic method was applied in understanding bacterial abundance and diversity, assessing the gene families, and revealing the novel MDGs genes of the mangrove sediment along South China Sea. Based on the metabolism pathway analysis, some potential MDGs from different gene families were selected and prokaryotic expressed. The biochemical characterization of the corresponding enzymes of MDGs, including temperature, pH, metal ions, metal-chelator, salinity, ionic detergent, and degradation pathway, were tested to further analyze their adaptability and potential to serve as tool enzymes in MG bioremediation.

Sampling and Environment DNA (eDNA) Extraction
The sediment samples were collected from mangrove forest located in Longhai City, Fujian, China (Zini Mangrove Nature Reserve, 24 • 20 N, 117 • 45 E). About 50 g of sediments were collected in a 50 mL of sterile centrifugal tube and stored on ice.
Stored the sediments samples in ultra-low temperature freezer (−80 • C) after they were brought back to lab. eDNA was extracted using the chemical lysis and enzyme digestion method. It was purified with low melting-point agarose [Sangon Biotech (Shanghai) Co., Ltd., China] and agarase (Takara Biotechnology Co., Ltd., Japan). Three replications of extraction and purification were performed and pooled together to avoid the extraction biases.
Sequencing and Bioinformatics Analysis of Metagenome eDNA sample was used for shotgun paired-end library construction. DNA sample was break into the fragments with the length of about 400 bp using ultrasonic breakers (Covaris, United States). End-repair, adaptor jointing, and purification of the DNA fragments were performed to construct the paired-end library. The concentration of the library was measured by Qubit 2.0 Fluorometer (Invitrogen, United States). High-throughput sequencing using Illumina HiSeq 2500 (Illumina Inc., United States) was performed by Shanghai Majorbio Bio-Pharm Technology Co. (China). Approximately 10 GB of data were generated and the raw reads containing "N" or adaptors were removed from the dataset and the clean reads were retaining for further analysis. The raw data of Illumina sequencing has been deposited in Sequence Read Archive (SRA) database 1 under the accession numbers SRR5824292.
All clean reads were assembled using SOAPdenovo with Kmer of 43-47. The best Kmer was identified on the basis of contig numbers, contig N50, contig length, etc. MetaGeneMark was used to predict the open reading frames (ORFs) based on the contigs and singletons obtained. ORFs were aligned using BLAST+ in Nr database of NCBI (cut-off E-value of 1e-5) and were annotated with the functional information. The putative MDGs mentioned before were filtered out according to the gene annotation. The bacteria origin of putative MDGs was predicted based on sequence homology with the protein sequences in NCBI Nr database using BLAST+ (cut-off E-value of 1e-5). COGs functional classification was conducted in STRING database 2 and KEGG 3 was used to obtain the biological pathways with a BLAST algorithm in KEGG database. All reads were aligned in SSU rRNA database of SILVA (cut-off E-value of 1e-5) to investigate taxonomic assignment of sampling site.

Full-Length Amplification of mgv-laccase and mgv-p450 Genes
Due to relatively lower similarities (about 50% similarities) with known sequences compared with other putative MDGs, three representative MDGs belonging to POD, LACC, and CYP, respectively, were chosen to be expressed in E. coli BL21 (DE3) cells and their MG-degrading activities were further detected. Three genes were labeled as mgv-peroxidase, mgv-laccase, and mgv-p450, respectively. Among them, mgv-laccase and mgv-p450, which had not complete ORFs, were amplified by TAIL-PCR with Genome Walking Kit (Takara Biotechnology Co., Ltd., Japan) according to the instructions. In brief, eDNA solution was used as the template for TAIL-PCR. Thermal asymmetric cycle was performed for three times, and the programs and reagent doses were mixed strictly according to the instructions. The degenerate primers for thermal asymmetric amplification were provided in the kit, and the specific primers (Table 1) for mgv-laccase and mgv-p450 were synthesized by Invitrogen (Shanghai) Co., Ltd. (United States). The possible DNA fragments were retrieved with TIANgel Midi Purification Kit (Tiangen Biotech Co., Ltd., China) after 1% agarose gel electrophoresis, and were cloned into pMD19-T Vector (Takara Biotechnology Co., Ltd., Japan). Then, DNA sequencing was conducted by using 3730xl DNA Analyzer (Thermo Fisher Scientific Co., Ltd., United States). Sequences were assembled with DNAman (Version 6.0.3.99) and ORFs of mgv-laccase and mgv-p450 were predicted in ORF Finder 4 .
Frontiers in Microbiology | www.frontiersin.org and mgv-peroxidase have been deposited in GenBank database under the accession numbers MF461728, MF461729, and MF461730, respectively.

Biochemical Analysis of MG-Degrading Activity of Recombinant Proteins
The MG-degrading activity of recombinant proteins was measured by the decolorization efficiency of MG with the formulae as follows: Decolorization efficiency (%) = (A − B)/A × 100, where A is the initial absorbance of MG and B is the final absorbance of MG at a wavelength of 622 nm. To value the cost of MG degradation, same quality (100 µg) of the three proteins were used in decolorization experiment.

Determination of MG Biodegradation Intermediates
The method for the determination of MG biodegradation intermediates referred to the previous studies (Du et al., 2011(Du et al., , 2013Wang et al., 2012;Yang et al., 2015). To determine the MG biodegradation intermediates, 20 mg/L MG was mixed with the three recombination proteins and incubated at 30 • C for 12 h. The products were desalted with C18 solid phase extraction column (Waters Sep-pak, United States). All the samples were analyzed using LC-MS (UPLC-Tof MS System, Waters, United States). The mobile phase contained H 2 O and acetonitrile. The initial proportion of acetonitrile was 5% and reached 100% within 15 min. The injection volume was 10 µl, and other parameters were set at the default settings.

Metagenomic Analysis
Approximately 10 GB of data was generated and 9.4 GB was remained as clean data. Subsequently, 33,756 contigs were assembled with best Kmer of 45 (Supplementary Table S1) and 44,743 ORFs were predicted. Based on SSU rRNA genes from metagenomic sequencing, microbial community showed great diversity on phylum level, mainly including Proteobacteria (50.19%) and Actinobacteria (17.03%), followed by other minor groups ( Figure 1A). On genus level, a total of 666 bacterial genera were found in our datasets. Due to the lack of adequate contig length, 70.57% genus of total microbes remained unknown ( Figure 1B). Twenty-three COG categories (Supplementary Figure S3) and 258 pathways in KEGG (Supplementary Table S2) were found. Thirteen metabolic pathways of common pollutant, such as styrene, ethylbenzene, aromatic compounds, chlorinated cyclohexane, and benzene degradation, were also detected in KEGG analysis.
Phylogenetic tree based on the amino acid sequences showed that mgv-laccase clustered within LACC of Pelobacter seleniigenes (Supplementary Figure S1A), mgv-p450 gathered with CYP of Salinisphaera shabanensis (Supplementary Figure S1B) and mgv-peroxidase was close to POD from Salinibacter ruber (Supplementary Figure S1C).

MG-Degrading Characteristics Analysis
Mgv-rPOD showed strong MG-degrading activity. After incubated with Mgv-rPOD for only 40 min, the MG decolorization efficiency could reach 99.0%, 99.2%, 99.4%, 96.3%, 97.8%, and 97.3% in 10 mg/L, 20 mg/L, 30 mg/L, 100 mg/L, 200 mg/L, and 300 mg/L of MG, respectively ( Figure 3B). The Mgv-rPOD had the optimum temperature 30 • C for color removal with 96.9% of MG-decolorizing efficiency and still maintained the high MG-degrading activity at 20 • C ( Figure 3C). Also, experimental results showed that high MG decolorizing activity was often between 7.0 and 10.0 and the observed optimum pH for the reaction was 9.0 ( Figure 3D). Cu 2+ had obviously negative effect, which caused the loss of 65.9% MG-degrading activity of Mgv-rPOD. Meanwhile, lots of bubbles were generated in the reaction with Cu 2+ . Besides, the use of other metal ions did not show much variation in the activity of decolorizing MG (Figure 6A). Different concentrations of EDTA (1, 10, and 100 mM) and NaCl (1, 10, and 100 mM) could not affect the decolorizing efficiency. However, there was a sharp decrease with 0.1% SDS and an absence of MG-degrading activity with 1% SDS (Figure 6B). Mgv-rLACC could decolorize 66.9% and 63.7% of MG at a concentration of 10 mg/L and 20 mg/L ( Figure 4B) in 24 h. However, in higher concentration of MG, Mgv-rLACC did not show obvious MG-decolorizing characteristic. The optimum pH of Mgv-rLACC was 4.0 ( Figure 4D) and the optimum temperature was 50 • C ( Figure 4C). 10 mM and 100 mM sodium chloride additive dramatically inhibited the decolorizing activity of MG, up to 48.1% and 0%, respectively ( Figure 6B). Metal-chelators, EDTA, had a negative impact on Mgv-rLACC ( Figure 6B) due to the lack of Cu 2+ which may serve as an electron transfer during MG degradation. There were no significant differences observed in the variation of metal ions ( Figure 6A).
With increasing of the MG concentration, the Mgv-rCYP showed the trend of gradual reduction. The rates of MG degradation incubated with MG concentration of 10 mg/L, 20 mg/L, and 30 mg/L for 24 h were 66.5%, 54.1%, and 36.2%, respectively ( Figure 5B). As Mgv-rLACC, no decolorizing phenomenon was detected with higher concentration of MG ( Figure 5B). The optimum pH of Mgv-rCYP for MG degradation was 7.0 ( Figure 5D) and the optimum temperature was 40 • C ( Figure 5C). Cu 2+ , Ni 2+ , and SDS could inhibit MG-degrading activities significantly (Figures 6A,B). MG-degrading characteristics of three recombination proteins were comparatively analyzed (Supplementary Table S3).

DISCUSSION
Advanced metagenomic sequencing has been widely used in studying the structures and functions of microbial communities (Andreote et al., 2012;Soares et al., 2017). There were evidences from metagenomics analysis showed that abundant microorganisms and genes existing in mangrove environment could be implicated in the process of pollutant degrading (Andreote et al., 2012;Gonçalves et al., 2015). In this study, a broad diversity of bacterial and MDGs profiles were confirmed in mangrove sediments. However, the actual bioactivities of the functional genes were still unclear due to the lack of characteristics of complete gene products. Metagenomics based TAIL-PCR was an effective and relatively low-cost method for obtaining complete genes. The results that MDGs from different gene families were expressed and the corresponding MG-degrading characteristics were tested to further prove the validity of the strategy.
Mangroves are one among the world's most productive ecosystems and are of ecological, economic, and societal importance (Mumby et al., 2004;Donato et al., 2011). Microbes play a key role in maintaining this productivity and conserve this ecosystem (Andreote et al., 2012). From our dataset of mangrove sediments, 666 bacterial genera (Figure 1) and 13 pollutant degradation pathways (Supplementary Table S2) were found. Genes that were found more frequently in one community than others are assumed to endow beneficial function on that community (Hugenholtz and Tyson, 2008). It was implied that microbes in the mangrove could constitute a large gene pool related to pollutant transformation with potential biotechnological and environmental implications. Besides, the apparent associations between various bacterial taxa and functional genes were analyzed ( Figure 2D). Thirteen genera were identified to be the original genera of the 44 MDGs. Of them, 12 genera have been reported to be involved in organic pollutant degradation. Nitrosomonas could be the most active genera with the most diverse MDGs, including four glutathione peroxidase genes and one peroxidase gene. It was reported that the genus had significant activities to degrade halogenated aliphatic (Wahman et al., 2005), trichloroethylene (Suttinun et al., 2010) and PAHs (Chang et al., 2002). The Shewanella genus could also remove MG, methyl violet B , naphthylamine sulfonic azo dye (Hong et al., 2007) and naphthol green B dye (Xiao et al., 2012). This information provided important clues to discover microbial and genetic resources from the mangroves. According to the previous studies, lots of members of Proteobacteria and Bacteroidetes harbor many functional genes, including biodegradation genes, phenol degradation genes and PAHs degradation genes (Klankeo et al., 2009;Fang et al., 2013;Zhao et al., 2016). However, the functions of Proteobacteria and Bacteroidetes for MG degradation have been poorly studied. In our work, 18/30 and 8/30 of those bacterial origin, respectively, belonged to the phyla of Proteobacteria and Bacteroidetes, which indicated that Proteobacteria and Bacteroidetes also played a crucial role of MG degradation in mangrove sediments.
Forty-four genes were predicted to be related to MG-degrading and belonged to three gene classes (LACC, POD, and CYP, Figure 2C). Relative lower amino acid identities from 50% to 70% (Figure 2B) suggested most genes were new sequences which were not reported in other environments. The diversity and richness of functional gene represented by the metagenomes implied the potential of mangrove microorganisms in environment restoration. Three MDGs from three gene families were expressed and the MG-degrading characteristics from the corresponding recombinant proteins at varied testing levels were verified. Mgv-rPOD showed the highest MG-degrading activity among the three MG-degrading enzymes (Figure 3). Due to the usual low temperature of natural seawater (about 10-35 • C), the low-temperature adaptability of Mgv-rPOD which kept a high degradation activity (95.8%) at 20 • C, has great advantage applying in marine environment. A strong anti-interference ability to environmental factors, including the adaptabilities to metal ions (except Cu 2+ ), metal-chelator (represented by EDTA), and the stability in different salinity further showed that Mgv-rPOD could be a promising bio-resource for removing MG from complex wastewaters. In addition, Mgv-rPOD had advantages in MG-degrading activity over not only the enzymes from this study but also other peroxidases reported in previous works. A manganese peroxidase, which showed the MG-degrading activity, was discovered from the white rot fungus Irpex lacteus F17. The manganese peroxidase could efficiently degrade 96% of MG at the concentration of 200 mg/L within 1 h; however, only about 75% of MG was degraded by using this enzyme at the MG concentration of 300 mg/L (Yang et al., 2016). By comparison, Mgv-rPOD could degrade more than 97% of MG within 30 min at the concentration of 300 mg/L, and the concentration was higher than most MG-degrading enzymes (Li et al., 2009;Saravanakumar et al., 2013;Zhang et al., 2013). Besides, many MG-degrading enzymes could only work at a narrow pH range. For instance, Saravanakumar et al. (2013) isolated manganese peroxidase isozyme H4 from Phanerochaete chrysosporium. However, about 50% MG-degrading activity of the enzymes was lost when the working pH was deviated from the optimum pH 4.5 . Unfortunately, it is very difficult to ensure the pH of nature environment when this enzyme was used in the remediation of MG pollution. Conversely, Mgv-rPOD possessed a wider pH range (7.0-11.0) for MG degradation, which suggested that Mgv-rPOD was more suitable for practice abatement of MG pollution. The possible mechanisms of MG degradation by Mgv-rPOD, Mgv-rLACC, and Mgv-rCYP were speculated according to the intermediates detected by LC-MS (Figure 7). All the processes of MGdegrading by the three recombination proteins were begin with a series of N-demethylation reactions, which was consistent with the former reports (Kedderis and Hollenberg, 1983;Cha et al., 2001;Murugesan et al., 2009). The products degrading by Mgv-rLACC and Mgv-rCYP were retained the triphenyl structure, and no further intermediates were detected, which was consistent with the previous study (Cha et al., 2001;Murugesan et al., 2009). The products of Mgv-rPOD suggested a further process of MG degradation. Mgv-rPOD not only mediated the N-demethylation but could also disconnect the triphenyl structure. Therefore, Mgv-rPOD was an attractive option for the remediation of MG pollution due to the high MG-degrading activity, high antiinterference, wide pH working condition, and the more complete degradation process.

CONCLUSION
Three recombinant proteins, especially Mgv-rPOD, were discovered and characterized as promising enzymes for the remediation of MG pollution in nature environment. Meanwhile, metagenomics provided an efficient path to view the microbial community structure and the functional (metabolic) potential of microbial community. The study further confirmed not-yetcultivated bacteria from mangrove environments were a potential source for novel biocatalysts.