Complete genome sequence of Photobacterium ganghwense C2.2: A new polyhydroxyalkanoate production candidate

Abstract

Polyhydroxyalkanoates (PHAs) are biodegradable bioplastics that can be manufactured sustainably and represent a promising green alternative to petrochemical‐based plastics. Here, we describe the complete genome of a new marine PHA‐producing bacterium—Photobacterium ganghwense (strain C2.2), which we have isolated from the Black Sea seashore. This new isolate is psychrotolerant and accumulates PHA when glycerol is provided as the main carbon source. Transmission electron microscopy, specific staining with Nile Red visualized via epifluorescence microscopy and gas chromatography analysis confirmed the accumulation of PHA. This is the only PHA‐producing Photobacterium for which we now have a complete genome sequence, allowing us to investigate the pathways for PHA production and other secondary metabolite synthesis pathways. The de novo assembly genome, obtained using open‐source tools, comprises two chromosomes (3.5, 2 Mbp) and a megaplasmid (202 kbp). We identify the entire PHA synthesis gene cluster that encodes a class I PHA synthase, a phasin, a 3‐ketothiolase, and an acetoacetyl‐CoA reductase. No conventional PHA depolymerase was identified in strain C2.2, but a putative lipase with extracellular amorphous PHA depolymerase activity was annotated, suggesting that C2.2 is unable to degrade intracellular PHA. A complete pathway for the conversion of glycerol to acetyl‐CoA was annotated, in accordance with its ability to convert glycerol to PHA. Several secondary metabolite biosynthetic gene clusters and a low number of genes involved in antibiotic resistance and virulence were also identified, indicating the strain's suitability for biotechnological applications.

INTRODUCTION

The adverse effects plastic waste has on our biosphere (Chae & An, 2018; Eriksen et al., 2014; Sebille et al., 2015) demand a global need to implement plastic clean‐up strategies and replace petrochemical‐based plastics with biodegradable, bio‐based polymers (Haward, 2018). Polyhydroxyalkanoates (PHAs) are a group of thermoplastic biopolyesters (Harding et al., 2007; Raza et al., 2018; Zhang et al., 2018) which are biodegradable and immunologically inert (Wang et al., 2014). The most common PHA is polyhydroxybutyrate (PHB), which can be produced by diverse bacteria (Inoue et al., 2016; Koller et al., 2011; Muhammadi et al., 2015; Sathiyanarayanan et al., 2017), which synthesize and store it as intracellular reserves of carbon and energy (Cavaillé et al., 2016; Keshavarz & Roy, 2010; Sedlacek et al., 2019; Slaninova et al., 2018). The most prevalent bacteria used in industrial bioplastic production are (1) Cupriavidus necator H16 (Yield10 Bioscience; CJ CheilJedang; Tianjin GreenBio Materials Co.; TianAn Biologic Materials Co.; Bio‐On Srl.), (2) Alcaligenes sp. (Biomer; HB Industrial S.A.); and (3) genetically engineered Escherichia coli that received PHA synthesis genes from naturally PHA producing bacteria such as C. necator H16 (Patent no. US5480794A, former Metabolix), Rhodospirillum rubrum (Patent no. US5849894A, CJ CheilJedang Corp), or Ralstonia eutropha modified to express the synthase gene from Pseudomonas fluorescens GK‐13 (Danimer Scientific; Noda et al., 2005). However, commercialization and production of bacterial PHA are constrained by its expensive substrates such as refined sugars, starch, or valuable plant oil (Koller & Marsalek, 2015), making its price twofold that of conventional, petroleum‐based plastics (average cost of PHB was reported to be approx. 4.88 USD/kg; Raza et al., 2018).

To decrease the production costs, a PHA producing strain should be able to grow to high cell densities and accumulate large amounts of PHA at the account of inexpensive carbon resources such as glycerol (Gahlawat & Soni, 2017; Poblete‐Castro et al., 2014), waste cooking oil (Sangkharak et al., 2020; Vastano et al., 2019), or other low‐cost biomass (whey, starch, spent coffee grounds, wastewaters, wheat, and rice straw, lignin, etc.; Alcântara et al., 2020). As biodiesel production is increasing, the glycerol market has expanded rapidly, and using this by‐product as a cheap substrate could be integrated into a circular economy approach (El‐malek et al., 2020). In this context, we isolated a new strain of Photobacterium ganghwense that can convert glycerol to biodegradable polymers (PHA) in the form of poly‐3‐hydroxybutyrate (PHB).

The Photobacterium genus encompasses Gram‐negative, facultative‐anaerobic, and motile bacteria, which are widespread throughout marine environments where some species live symbiotically with marine animals (Urbanczyk et al., 2011). This genus is relatively new, with 22 of the 28 existing species described within the last 15 years (Labella et al., 2017; Machado & Gram, 2017). Although several draft genomes are available, complete genomes exist for only three species (P. damselae, P. profundum, and P. gaetbulicola). Neither one of them is documented as a PHA producer.

The biotechnological potential of this genus is yet to be explored and most studies have focused on individual members’ pathogenicity toward animals and humans (Abushattal et al., 2020; Fumanal et al., 2020; Rivas et al., 2013; Romalde, 2002). Information regarding PHA production within the Photobacterium genus is scarce and, to our knowledge, only two species (P. leiognathi and P. phosphoreum) have been described to accumulate intracellular PHAs when provided with glycerol and peptone as carbon and nitrogen substrates (Boyandin et al., 2008). None of these PHA‐producing Photobacterium has a complete genomic sequence publicly available.

In this study, we report the isolation of a new PHA‐producing Photobacterium ganghwense (Park et al., 2006) strain (C2.2) and its complete genomic sequence, the first complete sequence available for this species. Furthermore, we present strain C2.2’s PHA production phenotype, its genetic basis, and provide valuable insights into other predicted metabolic capacities, gene transfer, structural modifications, virulence, and antibiotic resistance.

MATERIALS AND METHODS

Bacterial isolation and PHA screening

A wet sand sample was collected in February 2018 from a Black Sea beach on the Romanian shoreline (location coordinates: 44.215993; 29.656308; marine water temperature 2°C, pH 7, salinity 17‰; Appendix Table A1). To detach the bacterial cells from sand particles, the sample was placed on a rotative incubator (150 rpm) in filter‐sterilized marine water for 5 days at 20°C. After the detachment step, the aqueous phase was collected by centrifugation (6000 g, 5 min, 24°C), serially diluted up to 10⁻⁴ and spread (100 µl) on solid artificial seawater media (ASW) supplemented with 1 g/L yeast extract, 1 ml trace element solution, and 10 ml vitamin solution (ASW‐Y media; Xiao & Jiao, 2011). Morphologically distinct colonies grown on ASW‐Y media were selected and isolated. To screen for PHA‐production the solid ASW media was supplemented with sterile glycerol (2% w/v) and Nile Red staining solution (0.25 mg/ml) immediately after autoclavation. Plates with glycerol and Nile Red for PHA screening were inoculated with the previously isolated bacterial strains and incubated in the dark for 14 days (20°C). The formation of fluorescent granules (PHA accumulation) was evaluated daily using wet mount slides and viewed through epifluorescence microscopy (Zeiss Axioplan, Carl Zeiss, Germany; FS00 filter (λ_ex: 545/±25 nm, λ_em: 560–710 nm).

Bacterial strain identification

The bacterial strain was identified by 16S rRNA gene sequence similarity. The genomic DNA was extracted with the PureLink Genomic DNA Extraction kit (Invitrogen, #K1820), according to the manufacturer's instructions. DNA quantity and quality were evaluated by 0.8% agarose gel electrophoresis and Qubit dsDNA BR assay kit (Life Technologies, #Q32850). 16S rRNA gene sequence was amplified with the universal primers 27F/1429R. The PCR product was purified and sequenced at Genetic Lab (http://www.geneticlab.ro/). Strain identification was based on the 16S rRNA gene sequence similarity using the EzTaxon database (http://www.eztaxon.org/). Phylogenetic analyses were conducted using the Molecular Evolutionary Genetics Analysis (MEGA) version 7.0 (Kumar et al., 2016), with bootstrap values generated from 1000 replicates using the Neighbor‐Joining method (Saitou & Nei, 1987). The type strain Cupriavidus necator N‐1 (DSM 13513) was chosen as a root for the phylogenetic tree. The closest neighbors and type strain from the EzBioCloud database (Yoon et al., 2017) were included in the tree.

Confirmation of PHA accumulation

Transmission electron microscopy (TEM) was used for PHA granule visualization. For this, cells were pre‐cultured in ASW‐Y rich media (supplemented with 5 g/L tryptone), collected by centrifugation, and washed in sterile ASW. Freshly‐collected cells were resuspended in ASW without NH₄Cl and supplemented with 2% (w/v) glycerol and 2 g/L urea as carbon and nitrogen sources. The culture (starting at OD₆₀₀ of 0.2) was incubated at 20°C, 150 rpm. For TEM investigation, 1 ml of bacterial culture was harvested at 24 and 72 h. TEM was performed as described elsewhere (Pokrovskaya et al., 2012). Briefly, cells were fixed with 2.5% glutaraldehyde in PHEM buffer (60 mM PIPES, 25 mM HEPES, 10 mM EGTA, 4 mM MgSO₄·7H₂O), processed and imaged using Hitachi HT7800 TEM with a Xarosa camera (EMSIS GmbH).

Growth characteristics and polymer accumulation

The growth dynamics and polymer accumulation were monitored in 250 ml flasks with 100 ml of ASW media supplemented with 1 ml trace element solution, 10 ml vitamin solution, 2% pure glycerol (PG) as sole carbon, and 0.2% urea as the nitrogen source. The flasks were equipped with breather screw caps with ePTFE membranes, for better aeration. Biological replicates were started at OD600 equal to 0.2 and grown for 21 days, at 20°C and 150 rpm. The growth of C2.2 culture was monitored by measuring the optical density at 600 nm (OD₆₀₀). At each time point, three flasks (biological replicates) were sacrificed, and the cell dry weight (CDW) was determined following the biomass harvesting at 11500 g for 5 min. The pellet was freeze‐dried at −55°C (CHRIST ALPHA1‐2 LDPlus, Fisher Scientific) and weighed. The total amount and the composition of PHAs were determined following methanolysis of freeze‐dried, ground samples. Approximately, 30 mg of dry pellet were subjected to methanolysis with 1.5% sulfuric acid/methanol (3 ml) and chloroform (3 ml) at 100°C for 72 h, in screw‐capped test tubes. Benzoic acid (2 mg/test tube) was used as an internal standard. Poly[(R)‐3‐hydroxybutyric acid] (PHB; #363502, Sigma) was used as positive controls. After methanolysis, 2 ml of ammonia solution (12.5%) was added to separate the organic and aqueous phases. The organic phase containing methyl ester derivatives was analyzed by a GC‐MS system (Varian Saturn 2000) equipped with an HP 5 MS (5%)—diphenyl (95%)—dimethylsiloxane capillary column (30 m, 0.25 mm diameter, 1 μm film thickness). 1 μl sample was injected with 5.0 purity helium as the carrier gas, 13.81 psi, total flow 9 ml/min, column flow 1 ml/min, purge flow 3.0 ml/min, temperature increment of 10°C/min from 50°C to 260°C, injector temperature 220°C. Data analysis was performed using Saturn 2000 MS Workstation.

Genomic DNA extraction

Photobacterium ganghwense C2.2 DNA was extracted from cells cultivated on ASW‐Y rich media, at 20°C with 150 rpm for 24 h. The extraction was carried out using the PureLink Genomic DNA Extraction kit (Invitrogen, #K1820), according to the manufacturer's instructions. The DNA quantity and quality were evaluated by 0.8% agarose gel electrophoresis and Qubit dsDNA BR assay kit (Life Technologies, #Q32850).

Genome sequencing

Sequencing of P. ganghwense C2.2’s genome was done at the Functional Genomics Center Zurich (FGCZ, ETH Zurich and the University of Zurich, Switzerland; http://www.fgcz.ch/), using the PacBio RSII platform (Pacific Biosciences). The SMRTbell library was generated using the DNA Template Prep Kit 1.0 (Pacific Biosciences, USA, #100‐259‐100). The start concentration of the genomic DNA was measured using a Qubit dsDNA BR assay kit (Life Technologies, #Q32850). Then the genomic DNA (5 μg) was mechanically sheared to obtain an average size distribution of 15–20 kb, using a Covaris gTube (Kbiosciences, GB, #520079). Fragment size distribution was assessed using Bioanalyzer 2100 12 K DNA Chip assay (Agilent, #5067‐1508). Sheared genomic DNA (5 μg) was DNA‐damage repaired and end‐repaired using polishing enzymes suggested by the manufacturer. To create the SMRTbell template, a blunt‐end ligation reaction was performed, which was then followed by exonuclease treatment. To select the size of the SMRTbell template and enrich for large fragments (>12 kbp) we used a BluePippin device (Sage Science). The size‐selected library was quality inspected and quantified on the Agilent Bioanalyzer 12 kb DNA Chip and a Qubit Fluorimeter (Life technologies), respectively. A ready‐to sequence SMRT bell‐Polymerase Complex was created using the P6 DNA/Polymerase binding kit 2.0 (Pacific Biosciences, #100‐236‐500) according to the manufacturer's instructions. The Pacific Biosciences RS2 instrument was programmed to load and sequence the sample on one SMRT cell v3.0 (Pacific Biosciences, #100‐171‐800). A MagBead loading (Pacific Biosciences, #100‐133‐600) method was chosen to improve the enrichment of the longer fragments. To assess the adapter dimer contamination, the sample loading efficiency, the obtained average read length, and the number of filtered sub‐reads, a sequencing report was generated for every cell, via the SMRT portal. NanoPlot (De Coster et al., 2018) was used to assess the quality of the subreads.

Genome assembly

The raw data was initially assembled by the Hierarchical Genome‐Assembly Process 3 (HGAP3) pipeline (Pacific Biosciences). The final assembly was done in‐house using open‐source tools. For this, a subassembly was initially created using the wtdbg2 assembler (Ruan & Li, 2019), with an estimated genome size parameter set at the HGAP3 assembly size value of 5.8 Mbp. The final assembly was generated with the Flye assembler (Kolmogorov et al., 2019) using the wtdbg2‐generated subassembly, the HGAP assembly size as a reference, and two rounds of polishing. To assess the completeness of the assembly we used: Bandage (Wick et al., 2015) for graph visualization, Quast (Mikheenko et al., 2018) for obtaining the technical metrics, and BUSCO (Seppey et al., 2019) for evaluating the gene content and presence of single‐copy orthologs. Circlator (fixstart; Hunt et al., 2015) was used for changing the start position of the contigs to the origin of replication.

The whole‐genome taxonomic affiliation was assessed by in silico genome hybridization with OrthoANI (Lee et al., 2016) to compare the Photobacterium C2.2 genome to the existing draft genome of the type strain P. ganghwense DSM22954^T (ASM102945v1).

Genome annotation

ORF prediction and genome annotation were performed using NCBI PGAP (Prokaryotic Genome Annotation Pipeline). CGView (Stothard & Wishart, 2005) was employed for genome visualization. Besides PGAP annotation, CDS were also functionally annotated via the eggNOG‐mapper v2 using the DIAMOND mapping mode (Huerta‐Cepas et al., 2017, 2019); KEGG annotation through KoalaBLAST (Kanehisa et al., 2007, 2016); as well as Pathway Tools (Karp et al., 2015) and the integrated PathoLogic annotation tool, with the MetaCyc database (Caspi et al., 2012) for metabolic pathway mapping. For the PHA depolymerase annotation, we used the PHA Depolymerase Engineering Database (Knoll et al., 2009) and BLASTp (Altschul et al., 1990). Annotation quality and genome completeness were verified through a DIAMOND alignment against the UniProt TrEMBL database (Buchfink et al., 2014; Ravintheran et al., 2019; The Uniprot Consortium, 2019; Watson & Warr, 2019). Secondary Metabolite Biosynthetic Gene Clusters (smBGCs) were predicted using antiSMASH v5.0 (Blin et al., 2019). Contigs were screened for antimicrobial and virulence marker genes using Abricate (https://github.com/tseemann/abricate; Seemann, 2017) with the Resfinder database (http://genomicepidemiology.org/; Zankari et al., 2012), and the Virulence Factor Database (VFDB; Liu et al., 2019), respectively. Prophage regions were identified using PHASTER (Arndt et al., 2016). A workflow of the sequencing, assembly, and annotation processes is presented in Appendix Figure A1.

RESULTS AND DISCUSSION

Isolation and characterization of P. ganghwense strain C2.2

A total number of 82 bacterial isolates were obtained from the sediment samples. From all the isolates, PHA‐production screening done by Nile Red staining highlighted strain C2.2, as its cell fluorescence was observed after 24 h (Figure 1) and maintained throughout the entire incubation period (14‐days). Intracellular PHA granule accumulation was demonstrated by TEM imaging after growing the C2.2 in liquid media supplemented with glycerol as a sole carbon source (Figure 1). Cells with numerous small inclusions were already present after 24 h of incubation. After 72 h of incubation, the PHA granules and the cells had shown an apparent growth in size. Enlargement of the cell size may be a mechanism to maximize the capacity for granule storage.

1 FIGURE. Epifluorescence microscopy of Nile Red stained cells (a, b) and TEM (c, d). P. ganghwense C2.2 cells after 24 (a, c) and 72 h (b, d) cultivation on 2% pure glycerol, as sole carbon source, at 20°C. The bars represent 1 μm

16S rRNA gene sequence and phylogenetic analysis

16S rRNA gene comparison against the EzTaxon database showed the close taxonomical relation of strain C2.2 to members of the Photobacterium genera, especially with P. ganghwense DSM22954^T. Strain C2.2 shared a 16S rRNA gene sequence similarity of 99.86% and a 99.68% genome identity (see below) with the P. ganghwense DSM22954^T type strain. Based on the 16S rRNA gene sequence analysis, the isolated strain was grouped into a distinct cluster, together with two P. ganghwense strains, distant from the other three Photobacterium species with complete genomic sequences (Figure 2).

2 FIGURE. 16S rRNA gene‐based phylogenetic relationships of the “C.2.2” strain. The numbers shown at the tree nodes indicate bootstrap values (in %) based on 1000 replications. The scale bar indicates 0.02 substitutions per nucleotide position. Photobacterium strains with a complete genomic sequence available in public databases are underlined

Polymer accumulation on pure glycerol

Pure glycerol (PG) was used to assess the PHB‐production phenotype of strain C2.2. Intracellular polymer accumulation (% CDW) and final PHB production (g/L) were obtained from 2% PG. Culture optical density, CDW, and PHB content increased steadily throughout the cultivation. They stabilized after 7 days (OD₆₀₀ of 33.3 (±0.3)) and remained in close range until the 14th day. Peak values were recorded on the 14th day and reached 65.4% PHB content, with 4 (±0.3) g/L PHB (Figure 3). Strain C2.2 showed the highest overall PHB production (g/L) among those reported for PHA‐producing Gammaproteobacteria in similar conditions (use of PG as a sole carbon source and shake flask cultivations; see comparative Table 1). This and the moderate halotolerance of the C2.2 strain indicate its suitability for larger‐scale PHA‐production testing.

3 FIGURE. Time‐dependent cell dry weight (CDW) and PHA accumulation. Changes in P. ganghwense C2.2 cell dry weight (dark gray) and PHA accumulation (light gray; in g/L) when cultivated in shake flasks, on 2% pure glycerol, at 20°C. Bar graphs represent mean values (± SD) of three independent experiments

1 TABLEProduction of PHA from pure glycerol by Gammaproteobacteria strains and Cupriavidus necator

Strain	Time (h)	Culture volume (ml)a	Pure glycerol (%; v:v)	CDW (g/L)	PHA content (% of CDW)	References
Photobacterium ganghwense C2.2	96 14 days	100	2	3.1 6.3	53 65.4	Current study
Vibrio harveyi MCCB 284	72	200	2	3	68	Mohandas et al. (2017)
Vibrio spp. M11/M14/M20/M31	24 h after stationary phase	200	1	0.31/0.31/0.44/0.45	30.2/31.5/42.8/24	Chien et al. (2007)
Vibrio proteolyticus	48	NM	1	~ 1.6	<10	Hong et al. (2019)
Salinivibrio sp. M318	48	50	3	7.2	39	Van Thuoc et al. (2019)
Zobellella denitrificans MW1	100	300	2	3.7	73.5	Ibrahim and Steinbüchel (2010)
Aeromonas spp. AC_01/AC_02/AC_03	48	NM	1	1.69/1.48/1.2	7.8/5/3.6	Możejko‐Ciesielska and Pokoj (2018)
Halomonas sp. KM‐1	60	20	2	NM	40.5	Kawata and Aiba (2010)
Cupriavidus necator DSM 545b	88	100	2 3	~6.35 ~7.18	~77 ~79	Sun et al. (2020)
Cupriavidus necator DSM 545b	33.5	Fed‐batch 1500	24.9	82.5	62	Cavalheiro et al. (2009)

A PHB content (wt%) was expressed as a percentage of PHA mass in dry cell mass.

Abbreviation: NM, not mentioned.

^aThe culture volume in shake flask, if not mentioned otherwise.

^bCupriavidus necator DSM 545 is used in the industrial production of PHA.

Genome of P. ganghwense strain C2.2

The genome assembly of P. ganghwense stain C2.2 was covered 113x, with a total size of 5,744,420 bp, GC content of 50.34%, and is comprised of two chromosomes and one plasmid (Table 2). Genome comparison of strain C2.2 to the type strain of P. ganghwense DSM22954^T (ASM102945v1) showed an average nucleotide identity of 99.68% further supporting the classification of the C2.2 strain as a member of the P. ganghwense species. In terms of genome size, typically, Photobacterium species have genomes ranging from 4.2 to 6.4 Mbp, and a GC content between 38.7% and 50.9% (Machado & Gram, 2017). Thus, the 5.74 Mbp size genome and 50.34% GC content of strain C2.2 is similar to that of other metabolically versatile Photobacterium species (e.g., P. profundum and P. halotolerans) and to that of the P. ganghwense type strain (ASM102945v1). The C2.2 genome arrangement (Figure 4) into two circular chromosomes is observed for the other two Photobacterium species with a complete genomic sequence available and appears specific for the Vibrionaceae family (Machado & Gram, 2017). As shown for other species of the genus (Machado & Gram, 2017; Vesth et al., 2010), the second chromosome and the plasmid (202 kbp) can be a source of genomic plasticity and strain‐specific differences. In most cases, the studied Photobacterium strains have plasmids that range in size from 35 to 80 Kbp (Machado & Gram, 2017). The C2.2 megaplasmid is the second‐largest plasmid (202.454 bp) of the genus (the largest—319.190 bp—belonging to P. damselae strain Phdp Wu‐1; GenBank assembly accession no. GCA_003130755.1).

2 TABLEThe metrics for the in‐house Flye assembly generated with QUAST

Assembly metrics	Assembly metrics
# contigs	3	GC (%)	50.34
Contig 1 length (bp)	3,515,384	N50	3,515,384
Contig 2 length (bp)	2,026,582	N75	2,026,582
Contig 3 length (bp)	202,454	Avg. coverage depth	113
Total length	5,744,420	# N's per 100 kbp	0

4 FIGURE. The two chromosomes and plasmid of P. ganghwense strain C2.2. The genome maps consist of genome information displayed circularly (from the outside in): CDS, tRNA, rRNA, G + C content, and GC skew

Genome annotation, by PGAP, yielded 4,983 coding sequences (CDS), 188 tRNAs, 55 rRNAs (Table 3). Assembly completeness evaluation (BUSCO, Appendix Table A2) indicated that single‐copy orthologs were 100% and, respectively, 99.6% complete for Bacteria and Vibrionales lineages. According to the PathoLogic results, the CDS encode for enzymes involved in 304 metabolic pathways. After annotation, a DIAMOND blast against the UniProt TrEMBL database showed that 92.35% of CDS have over 90% similarity to protein sequences from TrEMBL, indicating the sequencing method did not have a negative impact on assembly quality and protein prediction. No CRISPR arrays were annotated.

3 TABLEGenome features for strain C2.2

Features	Features
Genome size (bp)	5,744,420	CDS (coding)	4.875
Chromosome	2	tRNAs	188
Plasmid	1	rRNAs	55
Genes	5.198	ncRNAs	4
CDS (total)	4.951	Pseudogenes	76

Genomic basis for PHA accumulation in P. ganghwense strain C2.2. The PHA‐positive phenotype of strain C2.2. was confirmed by the presence of a complete phaCAB operon (Figure 5) via functional annotations of clusters of orthologous groups (COGs). The phaCAB is located on chromosome 2, and also includes a phaP gene encoding a phasin family protein (FH974_19300), a surface protein with a role in PHA granule stabilization and production (Figure 3). The only annotated phaC (FH974_19305) encodes for a class I poly‐hydroxyalkanoic acid synthase, which polymerizes CoA thioesters of short carbon chain length hydroxyalkanoic acids (C₃‐C₅). phaA (FH974_19295) and phaB (FH974_19290) encode for: acetyl‐CoA acetyltransferase—the first enzyme in the PHA synthesis pathway, and an acetoacetyl‐CoA reductase. phaCAB cluster organization and its similarity of protein sequences with the functional ones of Cupriavidus necator H16 (DSM 428; Kutralam‐Muniasamy et al., 2018; Figure 5), suggest an operational PHA‐synthesis pathway in strain C2.2.

5 FIGURE. Simplified comparative analysis of PHA gene clusters from the PHA model organism C. necator H16 and P. ganghwense C2.2. The coding regions are indicated as follows: PHA synthase gene phaC (black); precursor‐generating enzymes genes phaA (dark gray) and phaB (light gray); surface protein gene phaP (oblique lines); PHA dependent transcriptional regulator phaR (white). The numbers above the coding regions indicate the sequence similarities with the protein counterparts of C. necator H16

Additionally, three putative phaP, four phaA, nine phaB, and one phaD, encoding hypothetical PHA synthesis transcriptional regulators (De Eugenio et al., 2010), were annotated for strain C2.2.

Unlike other PHA‐producing bacteria, genes encoding the PHA‐dependent transcriptional regulatory proteins (phaR/Q/F) and PHA depolymerase (phaZ) were not identified in the genome of P. ganghwense strain C2.2. The absence of PHA depolymerase was validated via a BLASTp search against the PHA Depolymerase Engineering Database (http://www.ded.uni‐stuttgart.de/; DED). This finding may explain the prolonged stability of accumulated PHB in the shake flask experiments (Figure 3). In a recent paper, de Vogel et al. (2021) highlighted similar features for two Vibrio strains (V. proteolyticus ATCC 15338 and V. alginolyticus ATCC 33787) that lack sequences similar to PhaZ depolymerases. For those strains, the authors re‐assigned two initially annotated lipases as putative extracellular PHA depolymerases. The putative depolymerases were found to be most similar to the PhaZ7 depolymerase present in Paucimonas lemoignei, with demonstrated depolymerase activity for extracellular amorphous PHA (native PHA granules; Handrick et al., 2001). A BLASTp hit similar to this supposed extracellular depolymerases (92% coverage, 82%–84% identity) was found in the genome of Photobacterium sp. C2.2 (FH974_07335). Its presence may indicate the ability of strain C2.2 to degrade PHB only after PHB’s extracellular release succeeding the cell death. The confirmation of depolymerase activity for this putative lipase needs to be addressed in future experiments.

In contrast, C. necator H16, the model strain for PHA production studies, has five well‐characterized intracellular and two extracellular depolymerases, as well as two oligomer hydrolases (Brigham et al., 2012). The absence of a depolymerase may prove advantageous for biotechnological applications, as PHA‐producing strains are usually genetically engineered to inactivate the depolymerization step for a higher yield in PHA production and increase the PHA molecular mass (Adaya et al., 2018; Kadouri et al., 2003; Kobayashi & Kondo, 2019).

Likewise, except for the extracellular PHA oligomer hydrolase PhaY, all the genes for enzymes involved in the catabolism of hydroxyacyl monomers were annotated in strain C2.2 genome: FabG (FH974_05790, FH974_24585), acetoacetate CoA synthetase AacS (FH974_21375), 3‐oxoacid‐CoA‐transferase subunits A and B (ScoA ‐ FH974_04750; ScoB ‐ FH974_04745, FH974_24545), and 3‐hydroxyisobutyrate dehydrogenase MmsB (FH974_21025).

As expected, strain C2.2. harbored all four genes (glpK, glpD, glpC, glpB) of the glycerol conversion pathway: glpK (FH974_00925) encodes for a glycerol kinase that converts glycerol to glycerol 3‐phosphate; glpD (FH974_00520) encodes for a glycerol‐3‐phosphate dehydrogenase that converts glycerol 3‐phosphate to dihydroxyacetone phosphate, which feeds into the glycolysis pathway, where it ultimately becomes acetyl‐CoA, a substrate for PHB synthesis (Kok et al., 1998); glpC (FH974_06490), and glpB (FH974_06495) encode subunits for an anaerobic glycerol‐3‐phosphate dehydrogenase (Cole et al., 1988).

Previous studies have shown the efficient use of glycerol for PHA production (Phithakrotchanakoon et al., 2015; Rodríguez‐Contreras et al., 2015; Tanadchangsaeng & Yu, 2012). In the case of C. necator H16, when glycerol was used instead of expensive sugars (e.g., glucose), the molecular weight of the resulting polymer was reduced, but its thermal and mechanical properties remained unchanged (Tanadchangsaeng & Yu, 2012).

Other pathways. Several metabolic pathways of biotechnological interest were annotated: a complete pathway for the degradation of phenylacetate (Teufel et al., 2010) to acetyl CoA, with a total of 51 genes involved in the degradation and metabolism of xenobiotics; complete pathways for the synthesis of various isoprenoids; genes involved in the metabolism of various terpenoids and polyketides; pathways for the degradation of various carbohydrates, such as starch, glycogen, chitin, etc. Such metabolic versatility could prove useful in expanding the range of raw substrates for the production of PHAs.

Prediction of smBGCs. Secondary metabolites have great potential for biotechnological applications (e.g., antibiotics, pigments, growth hormones, antitumor agents, and others). Since the production of bioactive molecules is poorly studied in Photobacterium (Čihák et al., 2017), we screened the genome of strain C2.2 for putative smBGCs. Based on their homology to known smBGCs, we predicted seven such gene clusters: four smBGCs on the large chromosome and three on the small chromosome. Strain C2.2 megaplasmid CDSs included no smBGCs nor genes involved in horizontal gene transfer. The smBGCs identified on the larger chromosome encode for antimicrobial functions: bacteriocin (FH974_15530), thiopeptide (FH974_04020 ‐ FH974_04040), and betalactone (FH974_13860 ‐ FH974_13885). These compounds provide a competitive advantage and protection from other microbial community members but could also mediate interspecies interactions (Čihák et al., 2017). The smBGCs identified on the smaller chromosome encode for: ectoine (ectABC operon), aryl polyene (T2PKS; FH974_24595), and polyketide (T1PKS; FH974_21825) synthesis. Polyketides like aryl polyene pigments are broadly distributed within Bacteria (Grammbitter et al., 2019). Polyketides, together with ectoine, serve as protection against reactive oxygen species and may have a protective role for strain C2.2 in marine environments (Das et al., 2015; Schöner et al., 2016). Further research is needed to investigate whether there is a link between the production of these compounds and certain environmental conditions.

Antimicrobial susceptibilities and resistance genes. Since many strains of the Photobacterium genus are well known for their virulence and antimicrobial resistance (Chiu et al., 2013; Fuertes‐Perez et al., 2019; Labella et al., 2017; Nonaka et al., 2012), we screened the P. ganghwense C2.2 genome for the acquired resistance mechanisms using Abricate. Surprisingly, strain C2.2 had only one antibiotic resistance gene (qnrS5) associated with resistance to fluoroquinolone (Han et al., 2012).

Abricate revealed several putative virulence‐associated genes: cheW‐2 chemotaxis protein, fliG—flagellar motor protein, fliM, and fliN, both polar flagellar switch proteins, involved in cell signaling and motility in liquid environments. These genes represent only a small fraction of the 15 core virulence genes of Gammaproteobacteria (Vázquez‐Rosas‐Landa et al., 2017).

Acquisition of additional virulence genes by horizontal gene transfer could be possibly mediated by one of the five phage regions predicted by PHASTER (Srividhya et al., 2007; Vázquez‐Rosas‐Landa et al., 2017). However, all five prophage regions are incomplete, with the largest being 103 kb in length and located on the small chromosome.

Auspiciously, a low number of antibiotic resistance and virulence genes along with the incomplete prophage regions are an indicator of genome stability, which may be beneficial in the potential biotechnological applications of P. ganghwense strain C2.2.

CONCLUSIONS

We obtained the first complete genomic sequence for a Photobacterium ganghwense strain. Moreso, strain C2.2 is capable of using glycerol, as a sole carbon source, to produce PHA granules. Genome analysis revealed the presence of all the genes required to synthesize PHA from glycerol, supporting the PHA‐producing phenotype. Observation of PHA accumulation dynamics showed a sustained increase in PHA content, with peak CDW and PHB content on day 14th. Gene annotation indicated the lack of a PHA depolymerase in the strain C2.2 genome. Although a putative lipase with the presumed ability to degrade extracellular amorphous PHA was annotated, our findings suggest that strain C2.2 is naturally prone to accumulate PHA for extended periods, a feature of great biotechnological importance. Also, the multitude of secondary metabolic pathways combined with the low number of genes involved in antibiotic resistance and virulence can be a plus from an applied science perspective. Our findings highlight the biotechnological potential of P. ganghwense strain C2.2, increasing the existing knowledge regarding PHA‐producing bacteria. The complete genome of a Photobacterium ganghwense contributes to the understanding of the Photobacterium genus.

ETHICS STATEMENT

None required.

ACKNOWLEDGMENTS

The authors gratefully acknowledge the work done by Andrea Patrignani and Giancarlo Russo from the ETH Zurich, Functional Genomics Center Zurich (FGCZ) for PacBio sequencing and HGAP initial assembly of Photobacterium sp. strain C2.2 complete genome. We also thank Randi Olsen at the Advanced Microscopy Core Facility (AMCF), UiT the Arctic University of Norway, Tromsø, Norway. We thank Assoc. prof. Amelia Rotaru for her constructive comments and suggestions on the paper and for editing this manuscript. This work was supported by the Romanian National Authority for Scientific Research and Innovation, CCCDI – UEFISCDI, project number 13/2017 ERANET‐Marine Biotechnology MARPLAST and project number 1/2018 ERANET‐Marine Biotechnology METAMINE, within PNCDI III.

CONFLICT OF INTEREST

None declared.

AUTHOR CONTRIBUTIONS

Irina Andreea Lascu: Data curation (equal); Formal analysis (equal); Investigation (equal); Methodology (equal); Validation (equal); Writing‐original draft (equal). Ioana Mereuta: Data curation (equal); Formal analysis (equal); Investigation (equal); Methodology (equal); Validation (equal); Writing‐original draft (equal). Iulia Chiciudean: Conceptualization (equal); Investigation (equal); Writing‐original draft (equal); Writing‐review & editing (lead). Hilde Hansen: Investigation (equal); Writing‐original draft (equal); Writing‐review & editing (equal). Sorin Marius Avramescu: Investigation (equal); Writing‐original draft (supporting). Ana‐Maria Tanase: Conceptualization (equal); Funding acquisition (lead); Project administration (lead); Writing‐original draft (equal). Ileana Stoica: Supervision (supporting).

DATA AVAILABILITY STATEMENT

Sequencing data are available through the NCBI Sequence Read Archive under SRR10983439: https://www.ncbi.nlm.nih.gov/sra/SRR10983439. The genome sequence of P. ganghwense strain C2.2 was deposited in the GenBank database under the accession numbers CP071325 ‐ CP071327; BioProject no. PRJNA548038: https://www.ncbi.nlm.nih.gov/bioproject/PRJNA548038. The strain is available from the DSMZ‐German Collection of Microorganisms and Cell Cultures under the accession number DSM 109767: https://www.dsmz.de/collection/catalogue/details/culture/DSM‐109767. All bioinformatics tools and databases used in this study are listed in Appendix Table A3. The workflow of this study is presented in Appendix Figure A1.

APPENDIX

A1 FIGURE. Flow diagram showing the overview of the study design

A1 TABLEGeneral features of P. ganghwense strain C2.2 based on MIGS mandatory information

Items	Description
Submitted to INSDC	SRA, BioProject, Genome
Investigation type	Bacteria
Project name	Complete genome sequence of Photobacterium ganghwense C2.2
Geographic location	44.215993, 29.656308
Geographic location name	Romania, Constanta, Black Sea
Collection date	2/15/2018
Environment (biome)	Shore
Environment (feature)	Seashore
Environment (material)	Sea sand
Environmental package	Sediment
Subspecific genetic lineage	Strain C2.2
Number of replicons	2
Extrachromosomal elements	1
Estimated size	5,744,420 bp
Observed biotic relationship	Free‐living
Trophic level	Heterotroph
Relationship to oxygen	Aerobic
Sequencing method	PacBio RSII
Finishing strategy (status; coverage; contigs)	Complete; 113; 3
Sediment depth	0 m
Sediment elevation	0 m
Sediment particle classification	Sand
Sediment pH	7
Sediment salinity	17‰
Sediment temperature	2°C

A2 TABLEBUSCO evaluation of the completeness of P. ganghwense C2.2 genome assemblies

BUSCO	Bacteria BUSCOs	Vibrionales BUSCOs
HGAP3	Flye	HGAP3	Flye
Complete BUSCOs	99.2% (123)	100% (124)	98.2% (1420)	99.6% (1439)
Complete and single‐copy BUSCOs	98.4% (122)	99.2% (123)	97.9% (1425)	99.3% (1435)
Complete and duplicated BUSCOs	0.8% (1)	0.8% (1)	0.3% (5)	0.3% (4)
Fragmented BUSCOs	0% (0)	0% (0)	0.7% (10)	0.1% (2)
Missing BUSCOs	0.8% (1)	0% (0)	1.1% (15)	0.3% (5)
Total BUSCO groups searched	124	1445

A3 TABLEThe bioinformatics tools, software versions, specific settings, and databases used in this study

Program	Version	Specific parameters used
NanoPlot	1.28.2	‐‐plots hex dot
wtdbg2	2.5	‐A ‐c rs ‐g 5.8m ‐S 2
Flye	2.6	‐‐pacbio‐raw [subreads.fastq] ‐‐subassemblies [wtdbg2.fasta] ‐‐genome‐size 5.8m ‐i 2
Bandage	0.8.1
seqtk	1.3‐r106	seq ‐r
Quast	5.0.2	‐f ‐‐pacbio
BUSCO	4.0.1	‐‐auto‐lineage‐prok ‐m geno ‐‐long
DIAMOND	0.9.29.130	‐‐db [TrEMBL db] ‐f 6 ‐‐max‐target‐seqs 1
EggNOG Mapper	2.0.1	‐m diamond
Abricate	0.5	‐‐db [vfdb/resfinder]
Pathway Tools	23.5
antiSMASH	5.1.2
Database	Version
VFDB	2020 ‐01‐17
Resfinder	2019‐10‐01
TrEMBL	2019‐12‐11
EggNOG	5.0.0