Genomic insight for algicidal activity in Rhizobium strain AQ_MP

Occurrence of Harmful Algal Blooms (HABs) creates a threat to aquatic ecosystem affecting the existing flora and fauna. Hence, the mitigation of HABs through an eco-friendly approach remains a challenge for environmentalists. The present study provides the genomic insights of Rhizobium strain AQ_MP, an environmental isolate that showed the capability of degrading Microcystis aeruginosa (Cyanobacteria) through lytic mechanisms. Genome sequence analysis of Rhizobium strain AQ_MP unraveled the algal lytic features and toxin degradative pathways in it. Functional genes of CAZymes such as glycosyltransferases (GT), glycoside hydrolases (GH), polysaccharide lyases (PL) which supports algal polysaccharide degradation (lysis) were present in Rhizobium strain AQ_MP. Genome analysis also clarified the presence of the glutathione metabolic pathway, which is the biological detoxification pathway responsible for toxin degradation. The conserved region mlrC, a microcystin toxin-degrading gene was also annotated in the genome. The study illustrated that Rhizobium strain AQ_MP harbored a wide range of mechanisms for the lysis of Microcystis aeruginosa cells and its toxin degradation. In future, this study finds promiscuity for employing Rhizobium strain AQ_MP species for bioremediation, based on its physiological and genomic analysis.


Introduction
Cyanobacteria are the photosynthetic organisms found in both freshwater and marine environment (Pal et al. 2018), which can be pelagic and benthic (Wehr et al. 2015). Although cyanobacteria have few positive trends, such as fixing atmospheric nitrogen but on the other hand, few species release toxic secondary metabolites such as dermatoxins, hepatotoxins, cytotoxins, and neurotoxins (Carmichael 2001;Pearson et al. 2010;Schmidt et al. 2014). Cyanobacteria have another ability to avoid predation by grazers (Lampert 1987). They can form elongated shapes, colonies and release some toxic secondary metabolites. The increased anthropogenic activities these days have increased the concentration of nitrogen and phosphate in water bodies which in a major contributor to the proliferation of algal bloom (Beaver et al. 2014). Algal bloom affects the entire water body by hindering light penetration and toxin release (Paerl et al. 2011;Pal et al. 2020). The recreational activities and drinking of these water affected by algal blooms often disturb animal health (Pal et al. 2020). Some of these algal bloom species release lethal toxins. The Microcystis species, such as M. aeruginosa, M. fosaquae, M. wesenbergii, M. ichthyoblabe, and M. phertaare are the leading cause of almost 90% of the harmful algal blooms in freshwater. M. aeruginosa is the most commonly observed cyanobacterial species causing harmful algal blooms and releases toxin, mainly microcystins (MCs) (Kim et al. 2019), which are neurotoxin and hepatotoxins. To date, more than 90 types of microcystins are released. Still, microcystin-LR is the most abundant and highly toxic variant (Pal et al. 2020). Toxic mechanism of MCs is due to the disruption of cytoskeleton formation and inhibition of protein phosphatase. In humans, MCs can enter through toxin-contaminated water or diet; they can cause oxidative stress, leading to cell damage (Beaver et al. 2014). Microcystins are readily water-soluble because their log of Communicated by Erko Stackebrandt. octanol and water distribution ratio is approximately -1 (pH 7). The essential binding sites of microcystins to the protein phosphatase are methyl-dehydroalanine (Mdha), and 3-amino-9-methoxy-2,6,8-trimethyl-10-phenyl-4,6-decadienoic acid (ADDA) groups. Microcystin toxicity depends on the amino acid combination present at the two different positions of the peptide ring. Microcystin-LR has amino acid leucine and arginine at the variable peptide ring, and Microcystin-RR has arginine and arginine (Rinehart et al. 1994;Corbel et al. 2014). Zhang et al. (2017) reported an innate effective bacterium of Sphingopyxis sp. having microcystin-degrading capability. It contained enzymes of microcystin degradation, i.e., mlrA, mlrC and mlrD. The literature reported nearly 60 strains of Microcystin-degrading bacteria which were mainly categorized among 22 genera of three bacterial phylum (Actinobacteria, Proteobacteria, and Firmicutes), and widely belong to α-and β-Proteobacteria (Park et al. 2001;Zhang et al. 2011a, b;Bourne et al. 1996;Jiang et al. 2011;Rapala et al. 2005). In addition, Pal et al. 2018 reported three bacteria showing algicidal activity, viz., Rhizobium sp., Methylobacterium zatmanii, and Sandaracinobactor sibiricus, where Rhizobium sp. revealed both algicidal and microcystin degradation property. In Lake Taihu, Stenotrophomonas F6 showed intense algicidal activity contrary to prevailing cyanobacterial bloom species, mainly via; secretion of extracellular algicidal compounds (Lin et al. 2016). Pedobactor sp. showed algalytic activity against Microcystis aeruginosa by the secretion of mucous-like substance in a lake in Japan (Yang et al. 2012). Zhu et al. (2016) isolated Rhizobium sp. TH capable of degrading microcystin-LR under environmental condition. The associated gene cluster mlr, a microcystin-degrading gene, was cloned and verified by studying the heterologous expression. Mann et al. (2013) reported bacterium Formosa agariphila (KMM 3901T) has a broad potential of algal polysaccharide degradation. Similarly, in this study, bacteria was isolated which has algicidal and microcystin degradation property. Rhizobium showed algal lytic property as revealed via: interaction studies of Rhizobium-Microcystis; analyzed by SEM and the genome sequence analysis for genes encoding for microcystin degradative pathway and cyanobacterial polysaccharide lytic enzymes. Rhizobium genome sequence analysis confirmed the presence of pathways and genes responsible for algal lysis and also conserved region for microcystin degradation.

Isolation and cultivation of M. aeruginosa and Rhizobium strain AQ_MP used for interaction studies
Microcystis aeruginosa was isolated from water samples of Ambazari Lake, Nagpur, Maharashtra. Culture was maintained by sub-culturing for 30-35 days as defined by Sangolkar et al. (2009). M. aeruginosa was poured first with the OD 678 0.03, cultured in 250 mL of conical flasks using BG-11 medium under continuous cool fluorescent light (12:12 light and dark cycle, 3000 lx) at 25 ± 1 °C and mechanically shaken a day thrice. Different bacterial cultures were isolated, purified and characterized from the lake water sample published in previous studies (Pal et al. 2018). Fresh bacterial culture was prepared by inoculating Rhizobium strain AQ_MP into 250 ml conical flask containing 100 ml of Luria broth media and kept overnight in a shaking incubator (120 rpm, 30 °C) for 24 h. Growth curve of Rhizobium strain AQ_MP is shown in Supplementary Fig. 1.

M. aeruginosa cells lysis after exposure to Rhizobium strain AQ_MP
Microcystis aeruginosa culture was purified as axenic culture, by ultra-sonication followed by sequential antibiotic treatment, i.e., kanamycin (100 µg/ml), ampicillin (50 µg/ ml), and imipenem (50 µg/ml), as per method reported by Pal et al 2018. Purified M. aeruginosa suspension (100 ml) was interacted with 10 ml of 1 OD 600 bacterial (Rhizobium) suspension. Two controls were set: one was Microcystis culture without Luria broth and another was Microcystis culture with Luria broth. The experiment was done for 10 days, where control and treated samples were collected every day. Samples (1 ml each) were centrifuged at 10,000 rpm for 10 min and fixed in 5% glutaraldehyde, and kept at 4 °C for Scanning Electron Microscopy (SEM) analysis. Cells (control and experimental) were centrifuged at 10,000 rpm and suspended and fixed with 5% v/v glutaraldehyde in 0.1 M phosphate buffer (30 min). Fixed control and experimental cells were centrifuged again, and the supernatant was discarded. Cell pellets were washed three times with 0.1 M phosphate buffer (15 min); dehydration was done with 35% ethanol (15 min), 50% ethanol (15 min), 75% ethanol (15 min), 95% ethanol (15 min), and two times with 100% ethanol (15 min). The gold coating of samples was done using Tescan SEM equipment. The cells were then examined in Vegag3 software, operating at 5.0 kV by a German scanning electron microscope (Kim et al. 2019). The total chlorophyll concentration was estimated for the algae-bacteria mixture for 10 days by a method described by (Gupta and Pawar 2018), wherein filtered dried biomass was extracted with 80% of ice-chilled acetone for estimation of chlorophyll using wavelength 663.2 nm and 646.8 nm. Cellulolytic activity of Rhizobium sp. using cellulose as substrate was determined by qualitative plate-zymography technique with Congo red staining (Bohra et al. 2019a, b), depicting cellulose as model substrate for Cyanobacterial polysaccharides (Supple Fig. 2).

Whole-genome sequencing and annotation of functional genes: Rhizobium strain AQ_MP
The qualities and quantities of the extracted bacterial DNA were checked by resolving on 1% Agarose gel followed by NanoDrop. The gDNA was used as template for molecular identification by targeting the bacterial 16S rRNA gene using Sanger sequencing technique. PCR amplified fragment of the bacterial 16S rDNA with single distinct band of DNA was observed on the agarose gel and specific primers were used for sequencing reaction of 16S amplicons. For whole-genome sequencing, QC passed genomic DNA sample was used for the paired-end sequencing library preparation (after confirmation), using Illumina TruSeqNano DNA Library Prep Kit (Srinivasan and Rajamohan 2020;Tikariha et al. 2016). Gene prediction and functional annotation were done using Rapid Annotation using Subsystem Technology (RAST) server (Overbeek et al. 2014). CG Viewer server was also used to create a circular genome that allowed visualizing sequence feature information in sequence analysis outcomes. In this server only, PROKKA annotation was also done to merge CDS, tRNA, tmRNA and rRNA subunits. InterProScan database using Pfam system was used to check the presence of the genetic domain of microcystin protein family. dbCAN database was used to check the annotate CAZymes.

Interaction study of Microcystis aeruginosa and Rhizobium strain AQ_MP
In this study, control and experimental Microcystis aeruginosa culture with Rhizobium strain AQ_MP culture were kept for 10 days to check the interaction and degradation of M. aeruginosa cells. After collecting samples, SEM analysis showed the contact between M. aeruginosa and Rhizobium strain AQ_MP cells. In 10 days of interaction, M. aeruginosa cells were lysed in flask experiments, as demonstrated in previous studies by Pal et al. (2018). It was observed that control Microcystis culture with and without Luria broth was not showing any degradation, also Microcystis cell numbers were increasing. But in the experimental setup, Microcystis cells were found to be lysed (Fig. 1). Many studies have already been suggested that M. aeruginosa have large mucilaginous aggregates comprised of a mucus substance called phycosphere. This mucous region typically comprises associated epiphytic bacteria (Kim et al. 2019). Zhang et al. (2011a), Gumbo and Cloete (2011) have already published some data on the interaction mechanism between bacteria and M. aeruginosa, depicting that the M. aeruginosa cell membrane gets damage, followed by the release of some extracellular substances. These extracellular substances serve as useful nutrients for bacteria growth. Such bacteriacyanobacteria sometimes shows epiphytic relationships, where the dominant species flourish the most. Total Chlorophyll content was also estimated, which showed that the chlorophyll levels were decreased at the end of the 10th day ( Supplementary Fig. 2).

Lysis of M. aeruginosa cells exposed to Rhizobium strain AQ_MP and its mechanism of action
Congo red assay showed a clear yellow zone indicating utilization of cellulose ( Supplementary Fig. 3) by Rhizobium culture. It was further proved by the presence of CAZyme GH1 and GH5 (Supplementary Table 1), which remains solely responsible for cellulose degradation through Cellulase enzymes. Nobles, et al. 2001 reported the presence of cellulose in different Cyanobacteria as their cell wall building block component. In addition, Kim and Han (2003) studied cellulase as an algalytic enzyme. Thus, Congo red assay representatively verified the degradative potential of the isolated Rhizobium against Cyanobacterial polysaccharides. Further, SEM analyses showed the cellular interactions between the Rhizobium strain AQ_MP and M. aeruginosa cells. On the first day of experiment, M. aeruginosa was observed as a dominant organism. As the incubation continued, the Rhizobium strain AQ_MP cells emerged as a dominant organism indicating the decline of M. aeruginosa.

Whole-genome sequence statistics: Rhizobium strain AQ_MP
The filtered high-quality PE reads of the bacterial samples were assembled into scaffolds using SPAdes assembler (v-3.13.0). Nanodrop reading was observed as 152 ng/ µl. Total data were 616 Mb; the total number of bases was 615,855,661, total number of reads was observed 2,069,397, total number of scaffolds was 122. The average scaffold size was (bp) 43,485. Max and min scaffold size was 620,217 bp and 200 bp. Whole-genome was submitted to NCBI/Genbank under the accession number JACJVI010000000 as Rhizobium strain AQ_MP. The degree of genomic similarity of the strain with closely related species was calculated using OrthoANI software, where average nucleotide identity values between closely related species were found ranged from 73.86 to 84.51%, which is lower than that of standard ANI cut-off for a new species is 95-96% (Lee et al. 2016;Yoon et al. 2017). It was found that the genome of Rhizobium rosettiformans have 84.51% relatedness to Rhizobium strain AQ_MP genome (Fig. 2). Two most closely related type strains with Rhizobium strain AQ_MP (Accession No. MF185100) include Rhizobium rosettiformans (Accession No. CP032405.1) and Rhizobium sp. (Accession No. CP058350.1). Using NCBI genome data, proteins/enzymes present in the genome were downloaded from NCBI. RNA from genomic FASTA (fna) file was downloaded from NCBI website and checked for the 16S rRNA gene relatedness with the old 16S rRNA partial sequence submitted to NCBI with the accession number MF185100 (Pal et al. 2018). It was found that Rhizobium strain AQ_MP genome has three 16S rRNA sequences. One sequence was used to prepare a phylogenetic tree with Rhizobium sp. (Acc. No. MF185100) partial sequence, and other related sequences showed the best match in BLAST search ( Supplementary Fig. 5). The 16S rRNA sequence present in the genome was closely associated with the 16S rRNA partial sequence submitted by Pal et al. 2018. FASTA file downloaded from NCBI was then uploaded into the RAST server to check the M. aeruginosa lytic and microcystin degradative pathways. CG viewer server database results showed a circular representation of the Rhizobium strain AQ_MP genome. From outward to inward: ORF (circle 1), CDS (circles 2 and 3), GC skew (circle 4), GC content (circle 5), and ORF (circle 6) are shown in Fig. 3. RAST server data results depicted the presence of Size of the sequence was found 1482 bp and 494aa (Fig. 4). Some mlr (microcystin-degrading gene) genes from NCBI were compared with conserved regions of Rhizobium strain AQ_MP, and match was observed and checked in Pfam and InterProScan (Mitchell et al. 2019;Finn et al. 2017). The result of InterProScan for the Rhizobium strain AQ_MP conserved genome sequence is represented in Fig. 4. At domain level, target conserved genome sequence were classified under the protein family IPR009197 and domain IPR015995, IPR010799. This protein family was nearly related to mlrC domain. This family signifies the C-terminus of a bacterial gene cluster product associated with the degradation of the toxin microcystin and is encoded in the mlr gene cluster. Phylogenetic relationship of selected mlrC sequence (from NCBI) was compared with mlrC sequences of Rhizobium strain AQ_MP (Fig. 4 C). Glycosyltransferase gene was also compared as outer protein family. It was seen that Rhizobium strain AQ_MP mlrC conserved region was related to Sphingopyxis sp. and Sphingomonas mlrC genes.

Microcystin degradative pathway
Analysis of the whole-genome sequences depicted that Rhizobium strain AQ_MP followed a Glutathione metabolic pathway (Sies et al. 1980) for the degradation of microcystins, in which glutathione-S-transferase (gst) (MBC2773493.1) and gamma-glutamyltransferase (tgm) (MBC2775265.1) enzymes were present. Microcystin has ADDA and Mdha site in which in the Mdha site, glutathione was attached and formed Microcystin-RR-glutathione (Beaver et al. 2014). Due to the presence of Gamma-glutamyltransferase enzyme Microcystin-RR-glutathione, get cleaved, and gamma glutamic acid was released, which leads to the formation of Microcystin-RR-cysteine-Glycine (Wang et al. 2018;Lance et al. 2014). Cyc-gly Dipeptidase (dug) was used to cleave the gamma-glutamylcysteine intermediate's glycine to get the cysteine-conjugated product, as a result, oxidized by acetyl transferase-acetyl co-A (acat) to form the mercapturic acid metabolite (Manahan 2003), which is Microcystin-RR-Mercapturic acid (Fig. 5). Another pathway was found to degrade microcystin, wherein Rhizobium strain AQ_MP, three enzymes jointly denoted as microcystinase operate in a sequential pathway to degrade MC. The first enzyme Microcystinase C precursor (mlrC)  (Goldberg et al. 1995). The final enzyme degrades the products formed by the first two enzymes and releases ADDA from the tetrapeptide intermediate (Fig. 5). Genes denoting linearized microcystinase (mlrB) was the conserved region of beta-lactamase (scaffold 22, sequence = 33,495-35,177), was annotated via RAST, InterProScan, and checked via Uniprot identity (supplementary Fig. 6).
RAST server data exhibited the presence of siderophores in the Rhizobium strain AQ_MP genome. Iron is a crucial element essential for key biotic processes. The common bacterial groups necessitate iron for existence and progression. Bioavailability of Iron is limited, which is a persistent source of pressure in many biological structures. There is accruing indication that Fe restricts phytoplankton biomass in the equatorial Pacific (Wells et al. 1994), the North Pacific gyre (Martin et al. 1989;Liu et al. 2014), and the Southern Ocean (Helbling et al. 1991). For microorganisms, obtaining iron is a significant challenge; capturing and integrating iron governs their existence. Cyanobacteria and algae, particularly responsible for biomass's primary production, require ten times higher iron content than nonphotosynthetic prokaryotes (Brand 1991). Bacteria, fungi, microalgae, and many higher plants have established specific approaches for low iron bioavailability. Siderophores secretion is one of the approaches among them. Siderophores are the molecules that chelate iron with high affinity (Guerinot Fig. 4 Gene structure of annotated mlrC gene in Rhizobium strain AQ_MP (A protein structure of annotated mlrC conserved sequence through InterProScan tool to identify the multimodular domains of mlrC. B Presence of mlrC gene in genome identified by RAST server. C Phylogenetic tree of different bacterial mlrC gene with Rhizobium strain AQ_MP conserved region) 1 3 1994). Siderophores extracellularly solubilize the iron from minerals of organic substances and transport them into cells when there is iron deprivation. Photosynthesis and capturing light energy are closely related to photosynthetic pigments. In photosynthesis, iron plays a crucial role in chlorophylla production (Imai et al. 1999). Studies say cyanobacteria requires higher iron uptake than other algae (Brand 1991). Liu et al. (2014) have suggested that due to the presence of siderophores, photosynthetic pigment synthesis was inhibited in M. aeruginosa. In this study, Rhizobium strain AQ_ MP is connected to siderophore release, which indicated the inhibition of photosynthetic pigment synthesis. It could be responsible for inhibiting M. aeruginosa growth due to the low bioavailability of iron (Martin et al. 1989). In previous studies, it was observed that bacteria and cyanobacteria tend to compete in freshwater for the low bioavailability of iron (Liu et al. 2014). Protein FASTA sequences of these loci were downloaded from NCBI and were submitted to Phyre software, and Pymol generates the structures of different siderophore and CAZymes ( Supplementary Fig. 7).
To control algal blooms, significant removal of nitrogen sources through denitrification pathways have been highlighted and reported by Jiang et al. (2020). Rhizobium strain AQ_MP harbored denitrification pathway genes (Supplementary Fig. 8) illustrating their role in removal of nitrogenous sources and thereby inhibit cyanobacterial growth, as Microcystin degradative pathways in the Rhizobium strain AQ_MP genome: Microcystin-RR has two groups, ADDA and Mdha groups as important part of microcystins, where ADDA allows binding of protein phosphatases to 1 and 2A which are the target enzymes. Mdha group covalently binds to the enzyme protein phosphatase's cysteine part. 1. The first step is the cleavage of the peptide ring at the ADDA-arginine bond, followed by subsequent degradation of the linear microcystin-LR product to yield a tetrapeptide intermediate and the ADDA moiety. In Rhizobium strain, three enzymes jointly denoted as microcystinase operate in a sequential pathway to degrade MC. The first enzyme Microcystinase C precursor (mlrC) linearizes Microcystin through the cleavage of the peptide ring at the ADDA-arginine bond. The second enzyme linearizes microcystinase (mlrB) cleaves this linear intermediate at the alanine-leucine bond, yielding a peptide intermediate of ADDA-Glu-Mdha-Ala. The final enzyme degrades the products formed by the first two enzymes and releases ADDA from the tetrapeptide intermediate. 2. Glutathione (GSH) is a peptide commonly found in the biotransformation of phase II enzymes. Phase II enzyme form a glutathione conjugate, which is a prevalent type of xenobiotic modification, i.e., glutathione-S-transferase (gst). This reaction occurs between the sulfhydryl group of the reduced glutathione and a nucleophilic center on the toxin. Mdha group is responsible for the nucleophilic center of microcystins. GSH-conjugation is generally the first sequence of the reaction, eventually producing an N-acetyl-cysteine which is mercapturic acid conjugate; can be efficiently removed from the cell. Gammaglutamyltransferase (tgm) is responsible for the enzymatic cleaving of the γ-glutamic acid group of the GSH, forming the intermediate γ-glutamylcysteine. The glycine of this γ-glutamylcysteine intermediate is cleaved by a dipeptidase (dug) to form cysteine-conjugate, which tends to get oxidized by Acetyl Transferase-Acetyl Co A (acat) to produce the mercapturic acid metabolite nitrogen acts a good source for the growth of cyanobacteria. Thus functional denitrifying genes existed in the genome, which includes nar, nir, nos and nor, where nir gene is responsible for converting nitrite to ammonia (Tikariha and Purohit 2019).

Conclusion
Rhizobium strain AQ_MP was isolated from the lake water, which showed the lysis of harmful Cyanobacterial species Microcystis aeruginosa. Scanning electron microscopy (SEM) and chlorophyll estimation revealed the algicidal property of Rhizobium strain AQ_MP. Genome analysis predicted that Rhizobium strain AQ_MP have secretion ability for extracellular substances like CAZymes responsible for algal polysaccharide degradation. The SEM analysis, presence of glutathione metabolic genes, toxin (microcystin) degradative pathways and functional gene clusters for polysaccharide degradation (CAZymes) ensured the algal lytic characteristics of Rhizobium strain AQ_MP.