Signs of biofilm formation in the genome of Labrenzia sp. PO1.

Labrenzia sp. are important components of marine ecology which play a key role in biochemical cycling. In this study, we isolated the Labrenzia sp. PO1 strain capable of forming biofilm, from the A. sanguinea culture. Growth analysis revealed that strain reached a logarithmic growth period at 24 hours. The whole genome of 6.21813 Mb of Labrezia sp. PO1 was sequenced and assembled into 15 scaffolds and 16 contigs, each with minimum and maximum lengths of 644 and 1,744,114 Mb. A total of 3,566 genes were classified into five pathways and 31 pathway groups. Of them, 521 genes encoded biofilm formation proteins, quorum sensing (QS) proteins, and ABC transporters. Gene Ontology annotation identified 49,272 genes that were involved in biological processes (33,425 genes), cellular components (7,031genes), and molecular function (7,816 genes). We recognised genes involved in bacterial quorum sensing, attachment, motility, and chemotaxis to investigate bacteria's ability to interact with the diatom phycosphere. As revealed by KEGG pathway analysis, several genes encoding ABC transporters exhibited a significant role during the growth and development of Labrenzia sp. PO1, indicating that ABC transporters may be involved in signalling pathways that enhance growth and biofilm formation.

Introduction

Bacteria and phytoplankton are two important marine ecosystem components (Sarmento and Gasol, 2012). During the past few decades, several investigations found interactions between phytoplankton-bacteria and their potential activities during harmful algal bloom (HAB) occurrence and termination (Buchan et al., 2014, Teeling et al., 2012). Research findings have already shown that bacteria live in free and particulate attached forms and interact with phytoplankton in the aquatic world. Specific interactions often occur in the phycosphere, a micro-environment that defines close spaces around phytoplankton cells, in which bacteria of phycosphere grow according to the extracellular products secreted by phytoplankton (Bell and Mitchell, 1972). In return, bacteria generate important components (Bertrand et al., 2015, Kazamia et al., 2012) to promote phytoplankton growth (Amin et al., 2012). Phytoplankton‐bacteria symbiosis is considered a phenomenon of significant importance in the availability of nutrients and essential biogeochemical cycles (Buchan et al., 2014, Durham et al., 2019).

In terms of our knowledge, bacterial behaviour processes for the control of phycosphere bacterial attachment and motility, have not been thoroughly studied. The widespread HAB community has long been established as strong associations between dinoflagellate and bacterial phycosphere while bacteria's pivotal function in dinoflagellate growth has been already confirmed (Alavi et al., 2001). It has long suspected that the related bacteria are responsible for the host dinoflagellates' toxic production (Lafay et al., 1995). Some researchers have questions about the association between dinoflagellate toxicity and bacteria (Green et al., 2010, Jauzein et al., 2015). Most of the studies support the most prevalent bacterial community within the marine diatom and dinoflagellate phycosphere being Rhodobacteraceae, particularly the roseobacter clade  (Geng and Belas, 2010). Successful phycosphere colonisation can be improved by unique genetic characteristics, such as motility, chemotaxis and surface attachment (Raina et al., 2019, Slightom and Buchan, 2009). Chemotaxis allows bacteria to sense differences in local food concentrations and thus regulates motility. Thus genes related to motility and flagellar seem essential for attachment and biofilm formation in many roseobacter (Bruhn et al., 2007, Miller and Belas, 2006). Many bacteria have a biphasic 'swim- or-stick' lifestyle in the phycosphere to easily find the food sources, thus minimising energy costs until finding the food. Bacteria use chemotaxis to find phytoplankton cells in the motile phase (Seymour et al., 2017). In the phycosphere, ‘switch’ is turned on, which allows the bacteria to migrate into the sessile lifestyle, the phenotype entails the loss of flagella and the further development of biofilm (Geng and Belas, 2010).

Microalgae-bacteria interactions studied thus far; nutrient seems to play a significant role (Fuentes et al., 2016). Overall, the association of microbial cells with the substratum surface at the seawater/surface interface under specific nutritional conditions would possibly lead significantly to the initiation and progress of microbial colonisation in marine environments (Dang and Lovell, 2016). This colonisation depends on the the species' ability to acquire essential nutrients from the surrounding environment (Naik et al., 2018). Vitamin B12 and iron are critical for phytoplankton development (Croft et al., 2005), and Iron is another essential bacteria micronutrient.

Acquisition of iron and transportation have significant phycosphere consequences. Some bacteria associated with dinoflagellate allow dinoflagellates to assimilate iron (Amin et al., 2009, Roe et al., 2013). For transporting these nutrients, transporter proteins are required and are a significant factor in host-bacterial interactions (Bleves et al., 2020, Eisenreich et al., 2019).

Furthermore, various transporters have identified as responsible for the transportation of signalling molecules. Quorum sensing (QS) is one of the most well-studied bacterial cell-to-cell communication signalling mechanism. Many bacteria have demonstrated QS by secreting small signalling molecules called autoinducers, which can evaluate changes in the populations of bacteria and co-ordinate the gene expression in the population as a whole (Waters and Bassler, 2005). Autoinducers of bacteria are acyl‐homoserine lactones (AHLs) used to regulate functions that are collectively beneficial to perform, including the motility, virulence, and biofilms (Antunes et al., 2010, Daniels et al., 2004, Hammer and Bassler, 2003).

Akashiwo sanguinea is a common alga that causes environmental issues worldwide. It has been the most prevalent producer of harmful algal blooms (HABs) in China's coastal regions (Menden-Deuer and Montalbano, 2015, White et al., 2014) (Lu and Hodgkiss, 2004). Observations of the epibiotic microbial communities of laboratory cultures of A. sanguinea revealed that the species abundance and composition of the microbial community are directly related to the physiological state of the dinoflagellates. These insights are the foundation of our ongoing investigations to understand further the functions of different phycosphere bacteria (Bolch et al., 2011). Genus Labrenzia is a member of the class Alphaproteobacteria (Amiri Moghaddam et al., 2018). Labrenzia consists of more than 100 genera and over 300 species (Pujalte et al., 2014a, Pujalte et al., 2014b, Xu et al., 2019). Representatives of Rhodobacteraceae exhibit very diverse physiology in marine environments and are the leading players in biogeochemical cycling (Pujalte et al., 2014a, Pujalte et al., 2014b, Xu et al., 2019). Genomes of Labrenzia sp. PHM005 and Labrenzia sp. CP4 strains and transcriptome of Labrenzia aggregata LZB033 have been well studied previously (Galan et al., 2019). However, information about the genome of Labrenzia species is limited. In the present study, we quantitatively examined the De novo genome of Labrenzia sp. PO1 isolated from the logarithmic growth phase of A. sanguinea algae. De novo genome assembly enables functional genomics research in species with limited genetic information, especially in non-model organisms. In the present study, the complete genome sequence of Labrenzia sp. PO1 was obtained using second-and third-generation sequencing technologies. Further, pathway enrichment, as well as enrichment of broader gene ontology categories, were performed. The current results provide new insight into Labrenzia sp. PO1 at the genomics level, suggesting several gene targets that might be involved in the nutrient transport and biofilm formation.

Materials and methods

Bacterial strain isolation

Plate streaking and serial dilution methods were used to isolate the culturable bacteria from the cultures of A. sanguinea under Log growth phase. The isolated bacteria have been maintained at 30 oC on Zobell marine agar 2216E(agar:15 g, ferric phosphate: 0.5 g,yeast extract: 0.5 g) in per litre seawater and R2A (agar: 15 g, MgSO47H2O K2HPO4: 0.3 g,sodium pyruvate: 0.3 g,soluble acid: 0.5 g,casamino acids: 0.5 g,protease peptone:0.5 g and yeast extract :0.5 g) in per litre distilled water.

Slow-growth of less culturable strains, was promoted by streaking the cultures of A. sanguinea are streaked on agar plates of R2A and 2216E medium and incubated for over one month until single colonies were obtained. Bacterial DNA (Omega, USA) Kit was used to extract the genomic DNA following the  manufacturer's instructions.

The ribosomal sequences 16S subunits were amplified with primers 27F/1492R (27F: 5 pour -AGAGTTGATCCTGCTCAG-3 um −1492R: 5 pour -TACGGCTACCTGTTACGACT −3 um:) with initial denaturation at 94 oC for 5 minutes, accompanied by 30 cycles of 94 oC for one minute each, annealing at 55 oC for one minute, extension at 72 oC for 1 minute, and final elongation at 72 oC for 10 min.The taxon of bacteria was defined in GenBank using BLAST (https:/blast.ncbi.nlm.nih.gov/Blast.cgi) for Labrenzia PO1 sequence.

Bacterial strain growth

Labrenzia sp. PO1 isolates were preserved on Zobell marine agar 2216E (per L of seawater, 15 g of agar, 0.1 g of ferric phosphate, 1 g of yeast extract, 5 g of peptone) at 30 °C before collecting single colonies. PO1 subcultured on 2216E agar. A single colony was reinoculated on 100 mL 2216E and cultured at 30 °C and 200 rpm for at least 18 h. One hundred μL of this crop was inoculated on 100 mL of fresh 2216E agar and cultivated at 30 °C. Samples of bacteria were obtained from 10-fold serial dilution in autoclave-purified water at 0 and 2 h, respectively, from 12 to 42 h after inoculation. By patting the ideal dilutions on 2216E agar, colony-forming units were quantified. The growing sample was examined three times, and the experiment was repeated three times.

Scanning electron microscopy method

The Labrenzia sp. PO1 streaked, and a single colony was used to inoculate 100 mL of 2216E medium at 30 °C and 200 rpm for 24 h. One mL of bacterial solution was taken and centrifuged at 8000 rpm for 3 min; after discarding the supernatant, the bacterial pellet was washed three times with phosphate buffer. Then 1 mL 2.5% pentanediol fixing solution was added and fixed at four-degree Celcius for 4 h, again centrifuged at 8000 rpm for 3 min, and discarded the supernatant. Washing three times with phosphate buffer. Dehydrated with 15%, 50%, 80%, and 100% ethanol gradients respectively. Each gradient was allowed to stand for 7 min and then centrifuged as mentioned above, and the upper phase was discarded. After dehydration, resuspended in 1 mL  sterilised pure water, taking 1 μL bacterial suspension onto a sterilised silicon wafer, and freeze-dried for 12 h. After drying and dehydration, the sample was sprayed with a conductive coating, and its morphology was observed by scanning electron microscope.

Bacterial biofilm formation

A single colony was selected from streaked bacteria and inoculated in sterilised 2216E medium and incubated at 200 rpm, 30 °C for 24 h. 200 µL of the bacterial solution was added to the wells of 96-well plates, and three biological replicates were performed for each strain. After incubation at 30 °C for 12 h, wells were washed three times with 0.9% physiological saline to remove free bacteria. After drying at 60 °C, bacteria were stained with 0.1% crystal violet for 5 min and washed with 0.9% physiological saline three times. After drying, 250 μL of 33% acetic acid was added to each well. After the crystal violet was fully dissolved, absorbance at OD590 was determined with a microplate reader (OD590 > 0.2 is positive). Isolated and purified bacteria were incubated in 100 mL of 2216E liquid medium at 200 rpm and 30 °C with shaking to the bacterial suspension logarithmic phase (concentration of the bacteria reached 108 cell mL−1 after 24 h of incubation). The bacterial solution to be tested was mixed with A. sanguinea liquid (algae concentration up to 104 cell mL−1, logarithmic growth period) in a ratio of 1:1, i.e., add 1 mL of bacterial liquid and 1 mL of algae liquid to each well of a 12-well plate. A blank control group was set up, and 1 mL of algal solution and 1 mL of 2216E medium were added to the control. Kept in a light incubator and shake manually at regular intervals 2 to 3 times a day (within 2 to 3 days). They were observed under the microscope to determine whether the isolated bacteria have algae dissolving ability.

Preparation of DNA and library construction

The PacBio (BGI, Shenzhen, China) and MGI2000 (MGI, Shenzhen, China) sequencing platforms were used to sequence Labrenzia sp. PO1 genome. A 50-mL culture of Labrenzia sp. PO1 was centrifuged for 10 min at 8000 rpm for genomic DNA extraction. The cell pellets were ground in liquid nitrogen and treated with 5 mL  DNA extraction buffer. After that, the cells were resuspended in 5 mL phenol/chloroform/isoamyl alcohol (25:24:1) and centrifuged twice at 9000 rpm for 16 min. The cells were subsequently resuspended in an equivalent volume of chloroform and centrifuged for 16 min at 12,000 rpm. Finally, the pellet was resuspended in 2.5 × volume of 100% ethanol to precipitate genomic DNA. The third-generation PacBio sequencing technology was used to sequence Labrenzia sp.PO1. The purity, integrity, and concentration of genomic DNA were determined using gel electrophoresis (1.2% agarose gel, 120 V, 50 min) and Qubit fluorometer version 2.0 (Invitrogen, Carlsbad, CA, USA) before library construction. A total of 10 μg genomic DNA was fragmented using g-TUBE (Covaris, Woburn, MA, USA) for library construction, and 10-kb DNA fragments were obtained using the Agencourt AMPure XP-Medium kit (Beckman Coulter, Brea, CA, USA). Exonuclease ExoVII (New England Biolabs, Ipswich, MA, USA) digestion was performed to generate the DNA fragments and subject to end repair were linked to the hairpin structure on both sides to form the SMRTbell dumbbell structure. The library was constructed successfully for PacBio sequencing after purification and sorting of DNA fragments. After the annealing and binding polymerase steps, qualified library fragments were sequenced. The genome sequencing data of Labrenzia sp.PO1 were uploaded to the National Center for Biotechnology Information Search database (NCBI) (BioProject ID: PRJNA579065; BioSample accession: SAMN13135800; SRA accession: PRJNA579065). The genome sequence data of Labrenzia sp. PO1 was uploaded to the National Center for Biotechnology Information Search database (NCBI)(SUB7234302PRJNA614627SAMN14435810JAAXCR000000000 Labrenzia sp. PO1). The library construction methods used for MGI2000 (MGI, Shenzhen, China) and PacBio (BGI, Shenzhen, China) platforms were identical.

Fragment assembly and gene annotation

The HGAP (Chin et al., 2013) software used to correct the sequencing data, high-quality error-corrected data used for assembly, and Finally, it optimised the assembly results to obtain high-quality genome assembly sequences. Canu (Koren et al., 2017) is an assembly software specially adapted for PacBio sequences. Its corrects algorithm reads and trims suspicious regions, and finally assembles high-quality genome sequences. At the present study present, we use SMRT Analysis software's P_ModificationDetection and P_MotifFinder modules for genome base modification and Motif analysis. Glimmer (Gene Locator and Interpolated Markov ModelER) (Delcher et al., 1999) is a tool for finding microbial DNA genes. The software uses interpolated Markov models (IMMs) to identify coding regions and distinguish them from non-coding DNA, with low false-positive predictions. Glimmer is the preferred software for microbial gene structure prediction in TIGR (The Institute for Genomic Research) database. So far, Glimmer has successfully annotated the gene structure of more than 100 bacteria. We used Glimmer version 3.02 software to predict the gene structure of the bacterial genome.tRNAscan-SE (Lowe and Eddy, 1997) can scan the whole genome of tRNA. The software integrates various analysis tools, which can identifying 99% of tRNA genes with extremely low false positives. While providing tRNA prediction results, the software can also predict the secondary structure of tRNA. We use tRNAscan-SE version 1.3.1 for tRNA prediction of the bacterial genome. RNAmmer (Lagesen et al., 2007) integrates a hidden Markov model, predicting 5 s/8s, 16 s/18 s, 23 s/28 s ribosomal RNA of the whole gene. We use RNAmmer version 1.2 to predict the rRNA of the bacterial genome. The Rfam database can be used to identify non-coding RNA (ncRNA) of prokaryotes. The database contains a large number of RNA family data information, and each RNA family contains a large number of multiple sequence alignments and covariance models (CMs). The Rfam 12.0 version (Griffiths-Jones et al., 2005), which contains 2450 families. This version is the first version based on Infernal 1.1. Use cmscan in Infernal to find ncRNA. We use Rfam 12.0 and Infernal to find the ncRNA structure of the genome. To explore the repetitive sequences in the bacterial genome, we use RepeatMasker to find various types of repetitive sequences in the whole genome. We use trf407b Linux (Benson, 1999) and MISA to search for the tandem repeat TR and microsatellite sequence SSR of the strain.CRISPR (clustered regularly interspaced short palindromic repeat sequences) is a cluster of regularly spaced short palindromic repeat sequences. Mining CRISPRs in Environmental Datasets, (MinCED) is a tool for analysing bacterial genome and metagenomic CRISPR elements. We used MinCED0.2.0(github.com/ctSkennerton/minced) to predict the CRISPR structure of the bacterial genome. For the generation corresponding functional annotation data, these sequences were compared with protein databases, such as SwissProt (Version: release-2017-07), Kyoto Encyclopedia Genes and Genomes (KEGG)(Version: 81), and Gene Ontology (GO) (Version: releases_2017-09-08) using BLAST. Because each alignment result was more than one, therefore we selected the optimal alignment result to ensure that the biological significance as gene annotation. A separate BLAST search was conducted to compare sequence with the Eukaryotic Cluster of Orthologous Groups (KOG) and KEGG databases against both the KEGG () and the KOG () databases. The SwissProt () and GO databases () were also used for GO and SwissProt annotations. The kinase database () and HMMER (v.3.0) with default parameters were used for the annotation and prediction of Protein Phosphatases Database (EKPD) and kinase domains in the Eukaryotic Protein Kinases.

Phylogenetic Analysis

For sequence alignment studies, 18S rRNA sequences were retrieved from the NCBI database. The NCBI blast feature was used to determine the similarity of Labrenzia sp. PO1 sequence with that of the reference strains. The accession numbers of the reference strains are listed in (Table 5). A Phylogenetic tree was constructed for the reference strains and Labrenzia sp. PO1 using the MEGA software (Version: 7.0.26; algorithm: neighbour-joining method, the maximum composite likelihood method for the evolutionary distances; related website: ).

Table 5

Gene annotation results of Labrenzia sp. PO1 according to the database.

Database	NR	SwissProt	KEGG	GO	COG
Number	5622	3819	2831	3887	4878
%	93.62	63.60	47.14	64.73	81.23

NR: Non-redundant protein database; KEGG: Kyoto Encyclopedia of Genes and Genomes; GO: Gene Ontology Database; COG: Clusters of Orthologous Groups.

Statistical Analysis

Analysis of variance (ANOVA) was performed, and statistically significant differences among group mean at P < 0.05 were determined by the least significant difference (LSD).

Results

Labrenzia sp. PO1 Isolation

Bacterial strain Labrenzia PO1 was isolated and identified in the present study. Their 16S rRNA sequence accession numbers and other related 16S rRNA sequences in the NCBI GenBank database shown in (Table1). This strain was isolated in the log Phase, this isolated strain being Rhodobacteraceae members, particularly the Roseobacter clade (Table 1).

Table 1

Characteristics of Labrenzia sp. PO1.

Number	GenBank accession No.	Reference sequence	% similarity	Taxonomic family	Species	Isolation stage	Biofilm forming ability
PO1	MH443361	KP301106.1	9%	Rhodobacteraceae	Labrenzia sp.	Log	+

Growth and characteristics of Labrenzia sp. PO1

The Labrenzia sp. PO1 strain belonged to the Labrenzia and was isolated from the A. sanguinea culture. Growth analysis (Fig. 1 A) revealed that strain reached a period of logarithmic growth. For Labrenzia sp. PO1, the maximum number of cells, 2.0 × 104, was reached at 24 h. The Labrenzia sp. PO1 morphological characteristics were evaluated by SEM (Fig. 1). The colony of Labrenzia sp. PO1 is round, with smooth edges. The PO1 strain exhibited behavior and shape on the plate shown in (Fig. 2). However, their color PO1 was milky white. If the plate was left for more than a week, a blackening phenomenon occurred from the edges to the single colony's centre.Labrenzia sp. PO1 cells can form a biofilm (Table 1).

Fig. 1

Growth curve of Labrenzia sp. PO1.

Fig. 2

SEM image of Labrenzia sp. PO1.

Genome sequencing and de novo assembly

A total of 6.21813 Mb raw data and 6005 annotated genes were generated by PacBio (BGI, Shenzhen, China) and MGI2000 (MGI, Shenzhen, China) platforms. After eliminating short adapter sequences and poor-quality data, high-quality subreads were generated by the PacBio platform. The statistical data for Labrenzia sp. PO1 subreads are shown in (Table 2), and the subreads' length distribution  in (Fig. 3). As shown in (Table 2), 15 scaffolds and 16 contigs were generated following the sequencing assemblyA total of 73,939 subreads were generated from 545,739,862 bases (Table 3). The scaffolds and contigs N50 values were 1,576,508 bp and 1,576,508, respectively. The GC content of the scaffolds and contigs was 59.12%. The GC content and GC depth Poisson distributions analysis showed that sequence bias or DNA contamination absence indicated  high-quality genome assembly (Fig. 4). The (G − C)/(G + C) method of calculation was used to analyse GC skew, distribution and and ncRNA annotation. Fig. 5 shows a genomic diagram with variables such as genome length, gene density, non-coding RNA (ncRNA) density, and GC rate. The Labrenzia sp. PO1 contains 5,527,806 CDSs. The annotation and assembly results were used to predict 119 ncRNA copies, which included including nine rRNA and 53 tRNA copies (Table 4).

Table 2

Statistical data of sample subreads.

Sample Name	Seq Type (#)	Total Number (#)	Total Length (Mb)	N50 Length (Mb)	N90 Length (Mb)	Max Length (Mb)	Min Length (Mb)	GC Content (%)
PO1	Scaffold	15	6,218,140	1,576,508	352,075	1,744,114	644	59.12
PO1	Contig	16	6,218,130	1,576,508	352,075	1,744,114	644	59.12

Total number (#): Total number of contigs or scaffolds. Total length (Mb): Total length of assembly. N50 Length (Mb): Used to determine the assembly continuity; the higher the better. N50 is a weighted median statistic indicating that 50% of the total genome length is contained in transcripts that are equal to or larger than this value. N90 length (Mb): similar to N50 length. Max length (Mb): the maximum length of a scaffold or a contig. Min length (Mb): the minimum length of a scaffold or a contig. Gap number (Mb): number of gaps in the sequence. GC content (%): the percentage of G and C bases in the assembly.

Fig. 3

Labrenzia sp. PO1 subread length distribution.

Table 3

Summary of clean data assembly for Labrenzia sp. PO1.

Sample	Total bases (bp)	Number of subreads	Average length (bp)	N50 (bp)
Labrenzia PO1	545,739,862	73,939	7380	10,292

Fig. 4

Correlation between the GC content and read depth distribution for Labrenzia sp. PO1. Abscissa, GC content; ordinate, average depth. The scatter plot shows a Poisson distribution of sequencing data with low GC bias.

Fig. 5

Genomic diagram of Labrenzia sp. PO1. Outer to inner: Genome (sorted by length), gene density (in 50,000 bp non-overlapping windows), ncRNA density (in 100,000 bp non-overlapping windows), GC rate (in 20,000 bp non-overlapping windows), GC skew (in 20,000 bp non-overlapping windows).

Table 4

RNA distribution of Labrenzia sp. PO1.

Type	Total bases (bp)	Copy number	Average length (bp)	% of genome
CDS	5,527,806	6005	920.53	88.64
rRNA	12,050	9	1338.89	0.19
tRNA	4143	53	78.17	0.07
ncRNA	52,328	119	439.73	0.84

CDS: coding sequence; rRNA: ribosomal RNA; tRNA: transfer RNA; ncRNA: non-coding RNA

Annotation of genome sequence

All Labrenzia sp. PO1 gene sequences were blasted in five databases, including COG, Swiss-Prot, NR, KEGG, and GO. Table 5 shows the number and percentage of annotated genes in each database. The Gene Ontology Association (GOA) had developed the GO database in 1988 and included the following three categories: biological process, molecular function, and cellular components. The cellular component category can provide information about macromolecular complexes, subcellular structures, and their locations. Gene and product of gene described by molecular function. The biological process described ordered combination of molecular functions to achieve broader biological functions. A total of 49,272 genes of Labrenzia sp. PO1 were annotated based on the GO database (Fig. 6). Of all genes within the biological process category, 1459 and 2017 were involved in metabolic and cellular processes, respectively. Moreover, one gene was involved in the extracellular activity. The top three categories with most genes were membrane (1197 genes), membrane part (1115 genes), and cellular (833 genes), respectively. In the molecular function category, 1458 and 2290 genes were involved in binding and catalytic activities, respectively. Furthermore, only nine genes were associated with the regulation of molecular function. To understand the biological system's important functions, molecular information about genes was obtained using high-throughput techniques and genome sequencing. This molecular information was also used to generate the KEGG database. Biological pathways are divided into eight categories, which are further subdivided. Each subdivision is annotated with a related gene that is displayed graphically. The accession numbers of the reference strains are listed in (Table 6).

Fig. 6

Distribution of functional annotations in the GO database. The ordinate is the annotation item and the abscissa is the number of corresponding genes.

Table 6

The National Center for Biotechnology Information (NCBI) accession numbers of the reference strains for phylogenetic analysis.

Species ID	Genome (Mb)	Scaffolds	Contigs	GC%	Accession ID	No. of genes	No. of specific genes
Labrenzia sp. PO1	6.21813	15	16	59.12	MH443361	6005	538
Labrenzia aggregata	6.60619	3	3	58.9	CP019630.1	5657	391
Labrenzia alba	6.88095	32	41	56.4	CXWE00000000.1	6234	1619
Labrenzia alexandrii	5.4074	3	6	56.4	CM011002.1	4887	1033
Labrenzia suaedae	5.14283	14	14	60.2	FRBW00000000.1	4626	982
Roseovarius pacificus	4.54703	52	52	62	FRBR00000000.1	4482	1539
Stappia indica	4.62317	17	17	67	MBQE00000000.1	4572	816
Stappia stellulata	4.62317	12	15	64.7	AUIM00000000.1	4254	689
Epibacterium multivorans	4.17562	45	45	59.7	CYSD00000000.1	3926	1255

In the present study, 3566 genes of Labrenzia sp. PO1 were categorised into 5 KEGG classes: cellular processes (12.43%), metabolic processes (65.49%), environmental information processing (18.01%), organismal systems (1.98%), and genetic information processing (6.04%) (Fig. 7). Metabolic processes were further divided into the following 12 subclasses: amino acid metabolism, global maps, energy metabolism, carbohydrate and lipid metabolism, vitamin and cofactor metabolism, nucleotide metabolism, biodegradation of xenobiotic compounds, metabolism of other amino acids, biosynthesis of secondary metabolites, terpenoid biosynthesis, and polyketide and glycan biosynthesis. Biofilm formation and QS-related genes were identified by further analysis of cellular processes. A total of 47, 134, and 340 genes were responsible for biofilm formation, QS, and ABC transport, respectively.

Fig. 7

Distribution of functional annotations in the KEGG database. The ordinate is the annotation item and the abscissa is the number of corresponding genes.

Predict genome CRISPR structure

CRISPR structure is a short palindrome repeat sequence in regular clusters; the length of the repeat sequence is generally 21–47 bp. There are three CRISPR elements on the chromosome of PO1 genome, indicating that PO1 also has a bacterial-specific CRISPR immune system, which can effectively defend against foreign DNA infestation (Table.7).

Table 7

Predicting CRISPR element information on the genome.

Position on chromosome (bp)	Number of repeats	repeat average length	Spacer average length	Repeat sequence
1672458–1672732	5	23	40	CCGCCTTCACGGCCACCGCCGCC
2097288–2097446	3	27	39	GGTCGTGTCGATGGCGGTGTAACGCTG
3452386–3452635	4	34	38	TTGCCATTGCCACCGGAATTGCCGTTGCCGTTGC

Identification of key genes

Genes involved in biofilm formation were identified from the genomic sequence and protein annotation data. waaA encodes 3-deoxy-D-manno-octulosonic acid transferase, which is responsible for inner core assembly and phosphorylation. Labrenzia sp. PO1 ompF and lpxC (encodes a protein involved in lipid biosynthesis) gene sequences were obtained in this study. Curli is another extracellular protein fibre involved in cell-surface, cell-cell interaction and  colonisation of host cells. csgD is a master regulator of biofilm formation and can activate curli fimbriae synthesis. Hence, the csgD gene was sequenced and annotated in this study. Lipoproteins encoding genes are that participate in adhesion and biofilm formation were annotated and sequenced in Labrezia sp. PO1. Labrenzia PO1 genomes contained five chemotaxis genes in a single operon‐like structure (cheA, cheR, cheW,cheB and cheY) (Table 8). Flagellar genes were present in the genome of Labrenzia PO1, flgf, MotA and MotB. In addition, Labrenzia PO1 contained genes related to pilus formation. Gene such as tadD that are involved in PGA biosynthesis in Labrezia sp. PO1 were annotated and sequenced (Table 8). Several gram-negative bacteria use ATP-binding cassette transporters (ABCs) to facilitate intermembrane transport of substrates. Substrates such as ions, amino acids, antibiotics, and materials required for biofilm formation are often recognised and transported by ABC transporter proteins.

Table 8

Genes involved biofilm formation and metabolism of Labrenzia sp. PO1.

Gene_ID	CDS_Length	CDS_GC(%)	Pep_Length	Gene	Protein
orf03123-1943	615	59.02439024	204	lprI	Lipoprotein
orf00150-5717	915	57.26775956	304	yfhB	Phenazine biosynthesis protein
orf03073-1913	912	63.26754386	303	mll1393	Phenazine biosynthesis protein
orf05066-3183	804	54.47761194	267	aapP	ABC transporter permease
orf06171-3887	1011	59.74282888	336	dppB	Dipeptide transport system permease protein DppB
orf04493-2810	1605	56.38629283	534	ddpA	ABC transporter substrate-binding protein
orf02722-1684	1989	58.57214681	662	rpoD	RNA polymerase sigma factor RpoD
orf05475-3441	1605	58.87850467	534	mdlC	Acetolactate synthase
orf07489-4715	1407	60.69651741	468	thrC	Threonine synthase
orf05456-3428	1845	59.94579946	614	ilvD	Dihydroxy-acid dehydratase
orf00218-127	780	62.17948718	259	hisF	Imidazole glycerol phosphate synthase
orf01654-1020	1179	62.68023749	392	serC	Phosphoserine aminotransferase
orf05907-3716	1311	59.72540046		glyA	Glycine/serine hydroxymethyltransferase
orf04860-3049	729	53.22359396	242	ydhC	HTH-type transcriptional regulator
orf03087-1922	1020	58.92156863	339	ilvC	Ketol-acid reductoisomerase
orf03820-2385	918	59.80392157	305	ilvE	Branched-chain-amino-acid aminotransferase
orf00217-126	744	62.90322581	247	hisA	Phosphoribosylformimino-5-aminoimidazole
orf04013-2510	522	60.53639847	173	ydjA	Putative NAD(P)H nitroreductase
orf00485-283	606	57.42574257	201	leuD	3-isopropylmalate dehydratase
orf05748-3611	1338	60.61285501	445	DHS1	Phospho-2-dehydro-3-deoxyheptonate aldolase
orf07227-4550	1920	61.71875	639	dxs	1-deoxy-D-xylulose-5-phosphate synthase
orf07242-4559	1299	61.0469592	432	ribBA	Riboflavin biosynthesis protein
orf07316-4609	1083	57.34072022	360	oprF	Outer membrane protein
orf05847-3680	1851	47.70394381	616	NGR_a03550	glycosyltransferase
orf07526-4736	1317	62.2627183	438	waaA	3-deoxy-D-manno-octulosonic acid transferase
orf00520-5950	1338	62.0328849	445	cpxA	Sensor protein
orf02281-1400	1143	53.10586177	380	CsgD	DNA-binding transcriptional regulator, CsgD
orf03001-1865	945	59.57671958	314	lpxC	UDP-3-O-acyl-N-acetylglucosamine deacetylase
orf02494-1535	606	62.04620462	201	oprF	Outer membrane porin F Precursor
orf07976-5019	777	62.03346203	258	TadD	Flp pilus assembly protein TadD
orf04743-2967	807	58.11648079	268	yabD	LuxR family transcriptional regulator
orf00446-262	534	61.2359550	177	LuxR	LuxR family transcriptional regulator
orf04516-2826	1143	56.86789151	380	Atu1380	Membrane-associated protease RseP, regulator of RpoE
orf05683-3570	843	61.32858837	280	znuC	Zinc import ATP-binding protein
orf04044-2527	588	57.65306122	195	CheY	chemotaxis protein
orf01688-1044	891	58.81032548	296	CheR	chemotaxis protein
orf01683-1041	492	60.36585366	163	CheW	chemotaxis protein
orf01681-1039	2595	60.61657033	864	CheA	Chemotaxis protein
orf01687-1043	1161	61.32644272	386	cheB	Chemotaxis response regulator protein-glutamate methylesterase
orf05140-3228	1620	61.17283951	539	MotB	Motility protein B Chemotaxis protein
orf04616-2888	897	58.63991081	298	MotA	Motility protein A Chemotaxis protein
orf00027-5630	1911	56.41025641	636	epsL1	protein involved in exopolysaccharide biosynthesis

The activity of dppB is regulated by five substrate-binding proteins (SBP) that function in the periplasm. These SBPs affect biofilm formation because four other SBPs bind to ABC transporter dipeptides. Genes encoding SBPs and zinc ABC transporter SBP znuC of Labrenzia sp. PO1 was sequenced and annotated. Furthermore, genes encoding putative ABC transporter permease, which can increase biofilm formation, were also sequenced and annotated. There are three stages of biofilm formation: initial attachment, microcolony formation, and community development. The bacterial cells in biofilms, therefore, are metabolically distinct from their free-living forms. It assumed that extracellular signalling molecules (autoinducers of QS) could regulate specific transcription factors to modulate bacterial expression in biofilms and induce the transition from a planktonic to a sessile lifestyle. The transcriptional regulator luxR a key player of QS regulates several genes, including those that encode proteins involved in biofilm formation.

Furthermore, mll1393 was sequenced and annotated in the present study. The QS-related mll1393 gene encodes proteins for the biosynthesis of phenazine compounds, promoting biofilm formation. The bacterial response to challenging environmental conditions during biofilm formation. This response requires sensor and regulator proteins during each phase of biofilm formation and genes cpxA encoding sensory histidine kinase in Labrenzia sp. PO1 was also sequenced and annotated.

Phylogenetic analysis of Labrenzia sp. PO1

A phylogenetic tree was constructed using the MEGA program with 18S rRNA sequences of Labrenzia sp. PO1 sample and reference strains. A neighbour-joining analysis showed that Labrenzia sp. PO1 hsve a common ancestor with all reference strains (Fig. 8). The evolutionary position of Labrenzia sp. PO1 is lower than that of Tropicibacter multivorans. Roseovarius pacificus, Labrenzia alba, Labrenzia alexandrii, Stappia indica, Labrenzia suaedae, and Stappia stellulata. In terms of kinship, Labrenzia sp. PO1 showed a higher kinship than Labrenzia aggregata.

Fig. 8

A phylogenetic tree diagram. The evolutionary history was inferred using the neighbor-joining method. The optimal tree with the sum of branch length of 0.01309240 is shown. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1,000 replicates) is shown next to the branches. The evolutionary distances were computed using the maximum composite likelihood method and are in the units of the number of base substitutions per site. The analysis involved six nucleotide sequences. Codon positions included were 1st + 2nd + 3rd + noncoding. All ambiguous positions were removed for each sequence pair. The final dataset comprised 1,739 positions. Evolutionary analyses were conducted in MEGA7.

Discussion

Rhodobacteraceae is one of the most diverse aquatic bacterial groups (Garrity, 2005, Giovannoni and Rappé, 2000) and is often found in the maritime environment. Members of this family are phenotypically and ecologically diverse (Pujalte et al., 2014a) and play a key role in biochemical cycling. We isolated Labrenzia sp. PO1 a member of Rhodobacteraceae from A. sanguinea. Microbial interaction studies showed that Labrenzia sp. PO1 coexists with A. sanguinea, a prominent species found in the South China Sea that forms large blooms (Lu and Hodgkiss, 2004). A phylogenetic tree revealed that Labrenzia sp. PO1 is phylogenetically linked to Tropicibacter multivorans, Roseovarius pacificus, Labrenzia alba, Labrenzia alexandrii, Stappia indica, Labrenzia suaedae, and Stappia stellulata strains. We sequenced and analysed Labrenzia sp. PO1 and submited to the GenBank database under the accession number JAAXCR000000000.

The molecular and genetic pathways for biofilm formation are not well understood in Labrenzia sp. PO1. Hence, whole-genome sequencing of Labrenzia sp. PO1 was performed to investigate the molecular mechanism of biofilm formation. GO and KEGG pathway enrichment was performed to identify essential gene-encoding proteins  that are involved in biofilm formation. Motility, chemotaxis, and attachment are likely to lead to efficient phytoplankton surface colonisation. We recognise genes involved in bacterial quorum sensing, attachment, motility, and chemotaxis to investigate bacteria's ability to interact with the diatom phycosphere. Flagellar motors were regulated by signal transduction cascade that mediated by CheAYW. CheB and CheR regulate methyl‐accepting chemotaxis proteins methylation state togeather (Wuichet et al., 2007). These data indicate that Labrenzia PO1 can chemotaxis, pili, and flagella structures construction because we find genes of chemotaxis, pili, and flagella in this strain. Extracellular polymeric matrix biosynthesis activates the function of curli subunit gene D (csgD) a central transcriptional regulator in the formation of biofilm; curli allows bacterial cells to migrate from the planktonic stage into multicellular condition. CsgD is important for forming biofilms; individual bacteria do not express CsgD at levels similar to the multicellular (Liu et al., 2014). Bistable characteristics of CsgD expression was high in a biofilm subpopulation, while CsgD expression was low in another bacterial subpopulation. The upregulated CsgD gene bacteria create the extracellular matrix and shape the biofilm structure that supports the low CsgD expressing bacteria because bistable CsgD high expression saves energy in biofilm formation, it can provide the bacterial community as a whole with additional benefits while maintaining population growth capacity (Grantcharova et al., 2010). Ironically, csgD encoding a DNA-binding transcriptional regulator were annotated and sequenced in the Labrenzia PO1 strain (Table 8). These factors are responsible for the transport and assembly of the curli fibres and maybe play an important role in biofilm formation (Xie et al., 2016). Surface polysaccharides often play a significant role in biofilm formation (Beloin et al., 2008, O'Toole et al., 2000). The evolutionary conserved waa gene is responsible for inner core assembly and phosphorylation in Labrenzia sp. PO1 (Whitfield et al., 1999); this gene was sequenced in this study (Table 8). For attachment, genes involved in exopolysaccharide production and biofilm formation were distributed in the bacteria's genome (Table 8). The initial steps in biofilm formation are reversible fixation and motility (Wood, 2009). Furthermore, the genes involved in adhesion, such as tadC, can promote biofilm formation. Several studies have shown that tight adherence (tad) genes are required for adhesion to the surfaces of the bacterium as well as for rough colony autoaggregation, morphology, and Flp pilli development (Bhattacharjee et al., 2001, Clock et al., 2008, Kachlany et al., 2000, Perez et al., 2006, Planet et al., 2003). Our data show that genome of Labrenzia sp. PO1 contains TadD gene which supports adhesion of this strain. QS autoinducers AHLs have widely been shown to modulate important biological functions motility and biofilm formation (Bassler, 2002, Waters and Bassler, 2005). In addition to AHL biosynthesis, a transcriptional regulator, encoded by a luxR family gene, is required to perceive AHLs and coordinate gene expression among bacterial populations. Labrenzia sp. PO1 genome contained luxR family genes. QS signals are required for biofilm formation. QS is a cell–cell signaling process that bacterial strains use to coordinate behavior (Dobretsov et al., 2009). Several QS gene regulation mechanisms occur in gram-negative bacteria via direct interaction of a self-inducer with a cytosolic transcription factor. LuxR protein is a transcriptional regulator that binds as a ligand to an autoinducer to induce the  dimerisation of QS-regulated genes and promote its binding to DNA (Choi and Greenberg, 1991). Histidine kinase receptors activate a cascade signalling to upregulate LuxR transcription factor and induce QS gene expression (Miyamoto et al., 1996, Miyamoto et al., 2003). Several gram-negative bacteria contain LuxR proteins that participate in QS signal generation (Fuqua et al., 1994). Bacteria with lux R genes have been hypothesised to identify and respond to exogenous signals from other bacterial populations (Hudaiberdiev et al., 2015). We identified Histidine kinase receptors and LuxR‐type regulators in the genome of Labrenzia sp. PO1 (Ball et al., 2017, Chen and Xie, 2011). We have found that phenazine production dramatically affects the bacterial colony's morphology (Dietrich et al., 2008). Several studies showed that phenazine involves in flow cell biofilm formations; a lack of phenazine production prevents cell attachment to surfaces and removes the formation of biofilms (Maddula et al., 2008, Maddula et al., 2006). A previous study has shown that QS can modulate phenazine production to induce adhesion and formation of biofilms (Cao et al., 2017). Another recent study showed that pyo expression could greatly influence bacterial biofilm formation (Cao et al., 2017). We identified gene of Phenazine biosynthesis protein in the genome of Labrenzia sp. PO1 (Ball et al., 2017, Chen and Xie, 2011) which may be important for the attachment of the cell to surface. Sigma factors are dissociable subunits of prokaryotic RNA polymerase that regulate recognition and transcription of genes. Direct gene transcription from the promoter of genes needed to support basic cellular functions is primary sigma factors (Wilson and Lamont, 2006). In addition to primary sigma factors, bacteria also contain secondary sigma factors to respond to environmental stresses. The sigma factor that regulates biofilm formation in Pseudomonas aeruginosa has been identified previously. Sigma factor-related genes were identified in the present study.

The bacterial capacity of biofilm formation is greatly enhanced upon the activation of ABC transporters (Zhu et al., 2008). Bacteria utilise transport proteins to transfer substrates via the internal and external membrane (Chung et al., 2009). Increased nutrient uptake via ABC transporters can help the bacteria adapt quickly to the evolving host microenvironment. Transporters also participate in signal transduction (Mai-Prochnow et al., 2015) and are required for bacterial survival (Garmory and Titball, 2004). Moreover, ABC transporters use ATP for intermembrane transport of substrates (Chung et al., 2009, Moussatova et al., 2008). Several ABC importers play a crucial role in the distribution of nutrients. ABC importers in Yersinia pestis, Proteus mirabilis, and Brucella abortus (Gutiérrez-Preciado et al., 2015, Vitreschak et al., 2002) import a wide variety of substrates such as carbohydrates, sugars, metals, peptide, amino acids, and other metabolites. ABC exporters are responsible for transporting substrates such as antibiotics, lipids, and proteins. It has been reported that limited nutrient availability might not promote the interaction between A. sanguinea and Labrenzia sp. PO1. Genes that may be involved in transport, namely transporters, were identified for the PO1 strain. In this study, we identified ABC transporter-related genes that are involved in regulating biofilm formation in Labrenzia sp. PO1 (Table 8). Earlier research in Pseudomonas aeruginosa, Listeria monocytogenes, Burkholderia pseudomallei, and Rhizobium leguminosarum bv. viciae 3841 revealed the importance of transporter proteins in bacteria (Lee et al., 2018, Pinweha et al., 2018, Vanderlinde et al., 2010, Wang et al., 2009, Zhu et al., 2008). Gene sequencing of a zinc-specific transporter showed that it could enhance biofilm formation in Labrenzia PO1. Biofilm formation is a bacterial response against stressful environmental conditions and requires sensor and regulator proteins at each step of the biofilm formation process (Król et al., 2019). We have previously shown that cpxA encodes sensor proteins that might play a role in biofilm formation under environmental stress (Beloin et al., 2008).

Conclusions

In summary, this study provides useful insights into the genomic influence of Labrenzia sp. PO1 on biofilm formation. The Labrenzia sp. PO1 strain was isolated from the A. sanguinea culture. Growth analysis revealed that strain reached a period of logarithmic growth period at 24 hours. We identified genes related to transporter proteins, and quorum sensing that participates in biofilm formation in Labrenzia sp. PO1. To our knowledge, this is the first study that provides the complete genome sequence of Labrenzia sp. PO1 using a third-generation sequencing platform. We also analysed and predicted KEGG and GO terms for potential genes and proteins that promote biofilm formation by assembling contigs and scaffolds. These data may provide new insights into the molecular mechanisms of Labrenzia sp. PO1 and help develop innovative strategies for limiting the growth of the harmful red tide.

Author Contributions

M.Z and Y.C conceived the study. M.Z and H.C contributed to the selection of the data. Y.C and H.C performed the pre-sequencing steps. M.Z and S.L write the manuscript.X.Y,Z.H and L.O revise the manuscript.

Funding information

This work was supported by Shenzhen science and technology application demonstration project (KJYY20180201180253571), and Shenzhen Grant Plan for Science and Technology(JCYJ20170818092708470).

Declaration of Competing Interest

The authors declare that there are no conflicts of interest.