Comprehensive genome analysis of a pangolin-associated Paraburkholderia fungorum provides new insights into its secretion systems and virulence

Genome datasets

The genome sequence of pangolin Pf (accession numbers: CP028829–CP028832) can be downloaded from the GenBank database at the National Center for Biotechnology Information (NCBI) website. It was assembled from the previous pangolin genome data (accession numbers: SRR3949798, SRR3949787, SRR3949728, SRR3949765, and SRR11935558) generated from the genome sequencing of an infected female Malayan pangolin (Tan et al., 2016b; Choo et al., 2020). Briefly, the raw reads generated from three P. fungorum-infected tissues (cerebrum, cerebellum and fetus) of the Malayan pangolin (manuscript in preparation) were pooled and assembled by mapping the reads to the reference genome, Paraburkholderia fungorum ATCC BAA-463 (Accession numbers: CP010024–CP010027) downloaded from NCBI using CLC Assembly Cell software with default parameters (D’Cruze et al., 2011). The mapped reads were assembled into a draft genome sequence for downstream analyses (Table 1).

Genome annotation

The pangolin Pf genome sequence was submitted to the Rapid Annotation using Subsystem Technology (RAST) server for gene prediction and functional annotation (Aziz et al., 2008). To ensure the uniformity in the annotations for comparative analysis, the genome sequences of 17 closely related species (Table 1), were retrieved from NCBI (http://www.ncbi.nlm.nih.gov) (Pruitt, Tatusova & Maglott, 2007) and annotated by the RAST server.

The core-genome SNP-based phylogenetic inference

The single nucleotide polymorphism (SNP) refers to a single-base variation in a DNA sequence (Vignal et al., 2002). The SNPs are the most abundant genetic variation, in a region that can be used to assess bacterial diversity and for taxonomic classification. The core-genome SNP-based phylogenetic analyses have successfully been used to infer the taxonomic relationships of many bacterial species (Ang et al., 2016; Choo et al., 2016a; Choo et al., 2014; Zheng et al., 2017). PanSeq was used to align all genome sequences and SNPs were identified in the conserved genomic regions among all species (Laing et al., 2010). All SNPs were extracted and concatenated into a long sequence for each genome. Phylogenetic tree based on the core-genome SNP was constructed using the genome sequences of the pangolin Pf and 17 selected closely related species (Table 1). The tree was constructed using MEGA-X (Kumar et al., 2018). Neighbour-joining tree was inferred using the Kimura’s two parameter model and nodal support was estimated using 1,000 replicates.

Whole-genome Average Nucleotide Identity (ANI) analysis

To evaluate the genetic relatedness between the genome of pangolin Pf and its closely related species, the ANI values were calculated based on the method as previously described (Ang et al., 2016; Rodriguez & Konstantinidis, 2014; Goris et al., 2007; Tan et al., 2016a). Two-way BLAST was chosen and only the forward and reversed-matched orthologs were used in the calculations. For robustness, the BLAST match was set for at least 50% identity at the nucleotide and amino acid level and a sequence coverage of at least 70%.

Genome Island and prophage prediction

All putative genomic islands (GIs) in the pangolin Pf genome were predicted using IslandViewer4, which is based on unique features in codon usage, dinucleotide sequence composition and the presence of mobile element genes (Bertelli et al., 2017; Langille & Brinkman, 2009). The RAST provided GenBank file was submitted into IslandViewer4 web server for GI prediction.

The identification of putative prophages in the pangolin Pf genome was performed using the PHAST (PhAge Search Tool) with default thresholds (Zhou et al., 2011). PHAST uses a web server to perform a series of database comparisons and phage feature identification analyses, locating and annotating prophage sequences.

Comparative virulence gene analysis

To identify the potential virulence factors, the RAST-predicted protein-coding genes in the genome of pangolin Pf were BLAST searched against the Virulence Factors Database (VFDB) (Chen et al., 2016; Chen et al., 2005) using the BLASTP of the BLAST software package (Altschul et al., 1997). In-house developed Perl scripts were used to screen out the genes that were identified as orthologous to virulence genes with at least 40% sequence identity and at least 40% sequence coverage in query and subject from the BLAST search. The filtered results were then further processed using in-house developed R scripts for clustering and constructing a graphical heat map with dendrograms, sorted according to similarities across the strains and genes.

Genome summary

The genome sequence of pangolin Pf was assembled from the raw sequencing reads generated from the infected tissues (cerebrum and cerebellum) of a pregnant female pangolin and its infected fetal muscle (accession ID: SRR3949798, SRR3949787, SRR3949728, SRR3949765, and SRR11935558). The assembled genome of pangolin Pf has a genomic size of 7.7 Mbp (7,746,865 bp), covering 86% of the reference genome P. fungorum ATCC BAA 463 and with a GC content of 62.2%. The pangolin Pf genome had a total number of 222 contigs with an N50 metric of 69,666 bp (Table 1). The genome was plotted using a circular representation by including genomic features such as CDS, tRNA, GC content, and GIs (Fig. 1). When mapping the read sequences back to the assembled genome sequence, they covered 99% of the genome with an average read depth of 29X, suggesting the high quality of the assembled genome sequence and its suitability for downstream analyses.

Confirmation of the taxonomic position of the pangolin Pf

We previously discovered the identity of the pangolin Pf based on the preliminary analyses using two bacterial classification markers such as rec A and 16S rRNA genes (Fig. S1). The single marker gene approach may not be able to clearly differentiate all closely related members of Burkholderiales (Jin et al., 2020). Therefore, we used a more robust and reliable whole-genome method or the core-genome SNP approach to further confirm the identity of the pangolin Pf (Fig. 2). The core-genome SNP-based phylogenetic tree classified the pangolin Pf as P. fungorum species.

To further validate the taxonomic position of pangolin Pf, the average nucleotide identity (ANI) was calculated based on the method proposed by Goris and his colleagues to measure the relatedness between the members of the Burkholderiales and Paraburkholderiales (Goris et al., 2007). Using the nucleotide sequence of pangolin Pf as the reference genome, the ANI was calculated by performing a pairwise comparison between the reference genome with the genome of other members of Burkholderiales and Paraburkholderiales (Table 1). ANI results indicated that our pangolin Pf strain is closely related to P. fungorum ATCC BAA-463 with an ANI value of 98.49% (passing the threshold of 97% to define a species), whereas other species showed ANI values below the threshold (Fig. 3). Taken together, the results of our ANI and the core-genome SNP-based phylogenetic analyses are consistent and confirm that pangolin Pf is indeed P. fungorum.

Functional analysis

RAST predicted 6,746 protein-coding genes and, 82 RNAs (Table S1). As anticipated, a large number of genes were enriched in basic functions such as carbohydrates (746 genes), amino acid and derivative metabolism (718 genes), cofactors, vitamins, prosthetic groups and pigments (401 genes), fatty acids, lipids and isoprenoids (318 genes), protein metabolism (300 genes), membrane transport (213 genes), and RNA metabolism (193 genes), which are important for bacterial survival (Fig. 4). We also observed 196 genes related to stress responses that may play an important role in bacterial adaptation to various environments (Table 2) (Koh & Roe, 1995; Koh & Roe, 1996; May et al., 1989).

Besides that, we identified 158 putative genes related to virulence, disease and defense including resistance to antibiotics and toxic compounds (132 genes), the host cell invasion and intracellular resistance (14 genes) and the ribosomally synthesized antibacterial peptides (12 genes), probably helping bacteria to inhibit the growth of similar or closely related bacterial strain (bacteriocins).

Toxin-antitoxin (TA) systems

The genome of pangolin Pf has two putative toxin-antitoxin (TA) systems; higAB and ygiUT. The ygiT is the transcriptional repressor while ygiU has been predicted to be a cyanide hydratase induced upon biofilm formation, acting as a global regulator controlling biofilm formation by inducing motility (Shah et al., 2006). The TA gene has been suggested to play a role in mediating growth arrest during stress situations (Butt et al., 2013; Fivian-Hughes & Davis, 2010). Both higAB and ygiUT were located closely with each other, forming the toxin-antitoxin system.

Urease

The pangolin Pf genome has several putative genes important for the urea decomposition and urease subunits such as Urea ABC transporter urtBCDE, Urease accessory proteins ureDEFG, Urease alpha, beta, and gamma subunits in the genome of pangolin Pf. The Urea ABC transporter urtBCDE involves in the uptake of urea and, in response to nitrogen limitation (Beckers et al., 2004). Urease, a nickel containing metalloenzyme, is a virulence factor enabling bacteria to survive by hydrolysing urea as the sole nitrogen source in nutrient limiting conditions to ammonia (Lin et al., 2012). For example, in other Gram-negative bacteria it has been shown that urease enables survival of the bacteria in strong acidic conditions of the stomach by neutralizing gastric acid with the released ammonia and thus playing a major role in the pathogenesis of gastroduodenal diseases (Cussac, Ferrero & Labigne, 1992).

Stress response

There were 196 putative stress response genes identified in the pangolin Pf genome (Table 2). They were mostly related to oxidative stress, osmotic stress, detoxification, heat shock, cold shock, various polyols ABC transporter, ATP-binding component, periplasmic substrate-binding protein, and permease component. A large number of these stress response genes are associated with osmotic stress such as the ones related to Choline and Betaine Uptake and Betaine Biosynthesis. The proVWX operon has been reported to be remarkably stimulated in response to hyperosmotic stress encoding for a high affinity transport system for glycine-betaine (Kudva et al., 2016; Price et al., 2008). The glycine-betaine have a role as a cryoprotectant that help bacteria to stabilize their cell membranes and adapt to cold environments (Subramanian et al., 2011).

Several oxidative stress-related genes were also identified in the pangolin Pf genome. For instance, the presence of oxidative stress-related genes for Paraquat-inducible (Pqi) protein A and B (O’Grady & Sokol, 2011). These genes are the members of the soxRS regulon (Koh & Roe, 1995) and their expression can increase during the carbon or phosphate starvation in the bacterial stationary phase (Koh & Roe, 1996).

The pangolin Pf also had genes related to periplasmic stress such as rpoE, rseA, rseB, and htrA (Table 2). RpoE is an important transcription factor, which functions as effector molecules responding to extracytoplasmic stimuli and is essential for Burkholderia to cope with thermal stress (Vanaporn et al., 2008). Besides that, htrA is a virulence factor in B. cenocepacia important for the growth under the exposure to stress and survival in vivo (Flannagan et al., 2007).

Membrane transport

The pangolin Pf genome has putative genes encoding ABC transporters related to membrane transport such as Phosphonate ABC transporter phosphate-binding periplasmic component; Oligopeptide ABC transporter, periplasmic oligopeptide-binding protein OppA, Oligopeptide transport system permease protein OppC & OppB, ATP-binding protein OppD & OppF; Branched-chain amino acid transport ATP-binding protein LivF & LivG, permease protein LivH & LivM); Dipeptide transport system permease protein DppB & DppC, ATP-binding protein DppD & DppF; Hopanoid-associated RND transporter, HpnN; Phosphonate ABC transporter permease protein phnE. The OppABC functions in the recycling of cell wall peptides (Goodell & Higgins, 1987). While the DppBCDF operon encodes an ABC transporter responsible for the utilization of di/tripeptides in Pseudomonas aeruginosa (Pletzer et al., 2014). The LivH and LivM permeases and the LivG and LivF ATPases are essential, to mediate the transport of these branched-chain amino acids into the cytoplasm (Adams et al., 1990). We also identified genes related to the resistance to antibiotics and toxic compounds such as the RND efflux system, the inner membrane transporter CmeB, outer membrane lipoprotein CmeC & NodT family; Cobalt/zinc/cadmium efflux RND transporter, membrane fusion protein, CzcB & CzcC family; Membrane fusion protein of RND family multidrug efflux pump, as well as the genes related to Ton and Tol transport systems in the pangolin Pf genome. The CmeABC gene is crucial to maintain a high-level resistance to fluoroquinolones and contributes significantly to the emergence of fluoroquinolone-resistant mutants (Yan et al., 2006). The RND multidrug efflux system is mainly responsible for the intrinsic multidrug resistance in Gram-negative bacteria (Guglierame et al., 2006).

Secretion systems

In Gram-negative bacteria, the components of secretion systems are highly specialized macromolecule nanomachines secrete substrates across bacterial inner and outer membranes (Abby et al., 2016). The secreted substrates are responsible for a bacterium’s response to its environment and in physiological processes such as survival, adhesion, adaptation, and pathogenicity (Costa et al., 2015; Duangurai, Indrawattana & Pumirat, 2018). In the pangolin Pf genome, we identified several putative secretion systems which will be discussed below (Table 3).

Comparative virulence gene analysis

To have a better understanding and more complete insights into the virulence of pangolin Pf, we performed a comparative virulence gene analysis of the genomes with known pathogenic Burkholderia sp. and Paraburkholderia sp... Our analysis showed that pangolin Pf shared some common virulence genes with these pathogenic Burkholderia strains such as epsI, pilQ, bsaX, escS, and hrcS (Fig. 5 and Table S2). Most of these predicted genes were associated with secretion systems especially Type III, and VI. The pangolin Pf also showed the presence of Type IV pilli (pilABCDERT) which are involved in adhesion (Piepenbrink & Sundberg, 2016). We found another set of common genes (eg. bsaX, escS, and hrcS) among the strains related to flagella that are involved in motility and invasion. Type III Secretion System is the main secretion system that mediates the secretion of effector molecules directly into host cells in B. pseudomallei and B. mallei (Beeckman & Vanrompay, 2010). Besides that, we found a T6SS cluster (TssA-TssM) in the pangolin Pf genome, probably playing an important role in bacterial competition (Spiewak et al., 2019).

Genomic Island and prophage analysis

We wondered whether the pangolin Pf had acquired the horizontally transferred Genomic Islands (GIs) over the evolutionary time. GIs are clusters of genes that are inserted into a bacterial genome during a single horizontal gene transfer event, playing important roles in microbial evolution, virulence, drug resistance and/or adaptation to different environments (Zhang et al., 2014). To examine this, we predicted the GIs in the pangolin Pf genome using the IslandViewer4 (Bertelli et al., 2017). We identified putative GIs predicted in the pangolin Pf genome such as harboured genes related to Molybdenum cofactor biosynthesis (MoaD and MoaE), Molybdenum transport proteins (modABCE) and Molybdopterin biosynthesis protein moeA (Table S3). The molybdopterin biosynthesis pathway is important for the molybdopterin cofactor syntheses, which are required for a variety of molybdoenzymes such as nitrate reductase that which plays an important role in the denitrification process in bacteria under oxygen limiting conditions (Leimkühler & Iobbi-Nivol, 2016). The nitrate reductase and the Molybdopterin biosynthetic pathway have also been associated with bacterial virulence (Andreae, Titball & Butler, 2014; Filiatrault et al., 2013; Fritz et al., 2002). In addition, we identified T6SS genes in the GIs of pangolin Pf, suggesting that these genes might have originated from other sources.

There are three incomplete prophages predicted in the pangolin Pf genome, in region 1 with about 7.8 kb in length with 10 coding regions; region 2, 8.2 kb in length with 10 coding regions; and region 3 with 8.8 kb in length and with 6 coding regions (Fig. S2; Table S4), suggesting that pangolin Pf might have acquired these prophages a long time ago. No intact prophages or recently evolved complete prophages were detected in the genome. We cannot rule out the possibility that the secretion systems of pangolin Pf might provide a strong defense to the bacterium in preventing the integration of phages into its genome.