Genome sequencing and annotation of Afipia septicemium strain OHSU_II.

We report the 5.1 Mb noncontiguous draft genome of Afipia septicemium strain OHSU_II, isolated from blood of a female patient. The genome consists of 5,087,893 bp circular chromosome with no identifiable autonomous plasmid with a G + C content of 61.09% and contains 4898 protein-coding genes and 49 RNA genes including 3 rRNA genes and 46 tRNA genes.

Direct link to deposited data

http://www.ncbi.nlm.nih.gov/nuccore/APJI00000000.

Experimental design, materials and methods

The genus Afipia was first proposed by Brenner et al. in 1991 [1], containing Afipia felis sp. nov., Afipia clevelandensis sp. nov., Afipia broomeae sp. nov., and three unnamed Afipia genospecies. This genus comprises of gram-negative, motile, oxidase-positive bacteria belonging to the alpha-2 subgroup of the class Proteobacteria. Subsequently, the isolation of Afipia birgiae, Afipia massiliensis and A. felis genospecies A from hospital water supply using amoeba co-cultivation was described in 2002 [2]. More recently, our lab isolated a novel species A. septicemium OHSU_I and OHSU_II from the blood cultures of 2 patients with poorly defined clinical illnesses [3]. The bacteria were originally living in an inactive state in SP4 broth medium designed to culture fastidious mycoplasmas [4]. The bacteria gradually adapted to grow in modified SP4 medium supplemented with 5% fetal bovine serum, vitamin B12, CKM (CaCl2/KNO3/MgSO4) and NAD/NADP (β-nicotinamide adenine dinucleotide/β-nicotinamide adenine dinucleotide phosphate, 10 mg/L) and could form microscopic colonies on 1.5% Noble agar plates prepared using modified SP4 medium kept in room temperature [3]. The subsequent single-colony derived microbes could grow in YM (yeast malt) and BHI (brain heart infusion) broth without serum supplement.

Whole-genome sequencing was performed on DNA extracted from single-colony clones of A. septicemium OHSU_II grown in SP4 broth cultures. Briefly, 50 ng of purified DNA of a bacterial culture derived from a single colony were subjected to DNA library construction using the Nextera DNA Sample Prep Kit (Illumina). A mixture of 4 to 6 separate DNA libraries was usually analyzed for each run involving 2 × 250 bp pair-end sequencing. The sequencing data, sufficient to cover bacterial genomes of ~ 5 million bps 378-fold, was obtained by Illumina MiSeq platform. The raw reads generated from genomic DNA sequencing were first assembled into 42 contigs using CLC Genomics Workbench (CLC Bio version 6.0) using the de novo assembly after filtering out and trimming low-quality reads.

The 16S rRNA gene sequence of A. septicemium was extracted from genome data and compared using NCBI BLAST under default settings (e.g. considering only the high-scoring segment pairs (HSP) from the best 250 hits) with the most recent database release in GenBank. The analysis showed that among all the matches for the 16S rRNA gene sequences, the species with the highest bit score of 2700 was uncultured Afipia sp. clone C-12 (AY568503), corresponding to an identity of 99.7%, followed by 2 uncultured Afipia clones and 2 uncultured bacterial clones, with bit scores ranging from 2676 to 2641. The highest scoring bacteria of currently recognized known species were Afipia lausannensis (bit score 2638) and A. broomeae (bit score 2630).

Fig. 1 shows the phylogenetic relatedness of A. septicemium with other closely related bacteria species based on the 16S rRNA sequences. The phylogenetic relatedness of difficult to culture A. septicemium with several other uncultured Afipia species using both Neighbor-Joining [5] and Minimal Evolution [6] criteria inferred by MEGA4 program [7] strongly supports the assigning of the newly identified microbes to the genus Afipia. It is important to note that the microbes in this genus, or even in this family, are known to have very high levels of homology in their 16S rRNA gene sequences [3], [8]. Thus, what presently classified as Afipia bacteria based on 16S rRNA sequences are likely to be a heterogeneous group of Bradyrhizobiaceae. Genomic content differences between A. septicemium and other established Afipia species of bacteria including A. broomeae are more than 22% [3], which strongly supports that A. septicemium is a distinct Afipia species.

Fig. 1

Phylogenetic tree highlighting the position of Afipia septicemium relative to other members of Afipia and Bradyrhizobiaceae bacterium SG-6C. This tree is inferred from 1487 aligned positions of the 16S rRNA gene sequence under the Neighbor-Joining criteria (A) and Minimum Evolution criteria (B) using MEGA4. GenBank Accession numbers of sequences used in the analyses are shown in parentheses. The branches are scaled in terms of the expected number of substitution per site. Numerals indicate bootstrap percentages over 50 after 500 replications.

A working draft genome of the A. septicemium OHSU_II-C1 was constructed using the complete genome of Bradyrhizobiaceae bacterium strain SG-6C (GenBank accession number CM001195) and the 3 supercontigs (1.1, 1.2 and 1.3 without including supercontig 1.4 of 128 Kb) of A. broomeae (GenBank accession numbers NZ_KB375282.1, NZ_KB375283.1 and NZ_KB375284) [9] as reference genomes using CONTIGutor [10]. Sequence of supercontig 1.4 of A. broomeae could not be aligned with Bradyrhizobiaceae bacterium SG-6C genome sequence and revealed significant homology with several plasmids of alph-2 Proteobacteria, and therefore was not used as a reference for draft genome construction. The draft genome of A. septicemium OHSU_II-C1 consists of a 5,087,893 bp circular chromosome with no identifiable autonomous plasmid with a G + C content of 61.09% and contains 49 RNA genes including 3 rRNA genes and 46 tRNA genes predicted by RNAmmer 1.2 [11] and tRNAscan-SE-1.23 [12].

The functional annotation was carried out by RAST (Rapid Annotation using Subsystem Technology) hosted by Fellowship for Interpretation of Genomes [13]. The functional comparison of genome sequences available on the RAST server revealed the closest neighbors of A. septicemium as Afipia sp. 1NLS2 (score 537) followed by Nitrobacter winogradskyi Nb-255 (score 517), A. clevelandensis ATCC 49720 (score 496) and Oligotropha carboxidovorans OM4 (score 486). A. broomeae genome was not available on RAST server for functional comparison.

A total of 5061 coding regions (2337 from a positive strand and 2724 from a negative strand) were found in the genome of which 3487 (69%) could be functionally annotated. The genome coding density is 87% with an average gene length of 875 bp. Fig. 2 revealed that the annotated genome has 135 genes potentially involved in virulence, disease and defense including 109 genes for resistance to antibiotics and toxic compounds such as cobalt–zinc–cadmium, fluoroquinolones, arsenic, beta-lactamase and chromium compounds.

Fig. 2

Subsystem distribution of Afipia septicemium (based on RAST annotation server).

Additionally, the annotated genome has 18 genes involved in phages, prophages, transposable elements and plasmids (Fig. 2). Further examination of the 18 genes revealed that there is a cluster of 14 gene transfer agent (GTA) genes from the genomic position 3,629,808 to 3,652,259 (Fig. 3). These GTA genes have a similar organization of the first GTA described in the purple photosynthetic alpha-2 Proteobacteria Rhodobacter capsulatus (RcGTA) [14]. These GTA genes include terminase, the phage portal protein, the prohead protease, the major capsid protein, the head–tail adaptor protein, the major tail protein, the tail tape measure protein, and host specificity proteins (Fig. 3). This particular region identified in A. septicemium genome revealed the highest similarity with that of Brucella abortus, a known pathogenic member of Rhizobiales, when compared using the RAST database (Fig. 3). Several potentially pathogenic Afipia bacteria such as A. felis, Afipia sp. 1NLS2 and A. birgiae are also found to possess similar numbers of RcGTA genes in their genomes. However, only the terminase gene is found in genomes of A. clevelandensis and A. broomeae.

Fig. 3

Comparative analysis of genomic regions containing RcGTA like genes for Afipia septicemium, Brucella abortus, Agrobacterium tumefaciens and Azorhizobium caulinodans. Sets of homologous proteins located in the genomic regions are presented in the same color, with the phage major capsid protein in the middle of the region as number 1 while other numbers are randomly assigned by RAST. The arrow for each gene is pointing in the direction of transcription.

Alphaproteobacteria are known to transfer their genomes into their eukaryotic hosts and to receive DNA from other organisms by bidirectional lateral sequence transfer through mechanisms such as transformation, transduction, conjugation, membrane vesicles, intracellular nanotube or GTAs [15]. Recently A novel putative GTA (BaGTA) was recently discovered in the mouse-infecting bacterium Bartonella grahamii, another species of alphaproteobacteria within the order of Rhizobiales. It was suggested that BaGTA contributed to the genomic plasticity of Bartonella to allow the microbe to evolve adaptively towards infecting a broader range of different hosts [16]. It is presently unclear if BaGTAs or any novel GTAs other than the RcGTA-like genes are present in A. septicemium genome.

Nucleotide sequence accession number

A. septicemium draft genome shotgun (WGS) project has been deposited at DDBJ/EMBL/GenBank under the project accession APJI00000000. It has a bioproject ID of PRJNA191806 and a GOLD ID of Gi51905.

Conflict of interest

The authors declare that there is no conflict of interest on any work published in this paper.

Specifications
Organism/cell line/tissue	Afipia septicemium OHSU_II isolated from blood of a female patient
Sex	N/A
Sample source location	Oregon Health and Science University, Portland, Oregon
Sequencer or array type	Sequencer; Illumina Miseq
Data format	Processed
Experimental factors	Microbial strain
Experimental features	draft genome sequencing of A. septicemium, assembly and annotation
Consent	N/A