3.1 Microbial community revealed by metagenomic sequencing
Metagenomic shotgun sequencing of the endophthalmitis sample revealed a diverse microbial community. After trimming low-quality reads and filtering human sequences, a total of 1.78 million microbial reads were obtained for subsequent analysis. At the kingdom level, only bacteria were identified. The bacterial sequences were classified as 2 phyla, 2 genera, and 4 species. The microbial community was dominated by Streptococcus_mitis_oralis_pneumoniae (97.98%), followed by Pseudomonas_unclassified (0.96%), Helicobacter pylori (0.72%) and Neisseria_unclassified (0.33%) (Fig. 1A). Streptococcus is a common pathogen causing severe infectious endophthalmitis (Yospaiboon et al. 2018). The virulence of pathogens is one of the factors affecting the prognosis of endophthalmitis (Sadaka et al. 2012). Compared with coagulase-negative Staphylococcus endophthalmitis, Streptococcus endophthalmitis has a worse visual prognosis (Lemley and Han 2007)(Durand 2017).
3.2 Abundance and categories of ARG types and subtypes
In summary, 6 ARG types comprising 14 ARG subtypes were identified in the endophthalmitis sample. Beta-lactam resistance genes were the dominant ARG type (4.41 × 10− 1 copies/16S rRNA gene), followed by ARGs for macrolide-lincosamide-streptogramin (2.19 × 10− 1 copies/16S rRNA gene) and multidrug resistance (1.79 × 10− 1 copies/16S rRNA gene) (Fig. 1B). The high abundance of multidrug resistance may be one of the reasons why the patient was ineffectively treated with systemic and topical antibiotics. The top 5 ARG subtypes with the highest abundance were mefA, PBP-2X, multidrug transporter, PBP-1A, and PBP-1B (Additional Fig. 1). These results were valuable for guiding rational clinical medication prescription.
3.3 Streptococcus genome recovered from metagenomes
Metagenome assembly and genome binning based on the MetaWRAP pipeline resulted in one near-complete genomic bin with 23 scaffolds. The genome size of the obtained bin was 2 Mbp, encoding 1974 genes with a mean gene length of 907 bp. The genome was estimated at 99.78% completeness and 0.846% contamination, with a low GC content of 39.7%. Only 0.76% of this genome was occupied by repeat elements. Interspersed repeats were the predominant type of repeat region, which accounted for 0.55% of the whole genome.
The genomic bin has a single circular chromosome, and no plasmids were found. Blast searching genomic sequences against the NT database showed that the bin was classified as Streptococcus but could not be identified at the species level. Metagenomic taxonomy results showed that Streptococcus_mitis_oralis_pneumoniae were the most dominant species, accounting for 97.98%. Because S. mitis, S. pneumoniae, S. pseudopneuomoniae, and S. oralis are closely related viridans group streptococcus species, they are difficult to discriminate at the species level (Scholz et al. 2012)(Ikryannikova et al. 2013). Therefore, there was a need to carry out genome comparisons of the Streptococcus MAG against related Streptococcus spp.
3.4 Comparative genome analysis
We used ANI and digital DDH for genome comparison analyses. The ANI is regarded as the most relevant comparative parameter to determine bacterial species. A whole-genome pairwise ANI > 95% indicates the same species (Jain et al. 2018). DDH is generally used to determine the genomic similarity among strains (Goris et al. 2007), and 70% similarity is considered the gold-standard threshold of DDH values for species boundaries (Meier-Kolthoff et al. 2014). Table 1 contains the digital ANI and DDH values between this Streptococcus MAG and 5 type strains of related species. The digital DDH and ANI values of the Streptococcus MAG against the 5 type strains of related species were below 95% and 70%, respectively, which showed that the Streptococcus MAG represented a new species. This new Streptococcus strain was named Streptococcus sp. v1. nov. Consistent with the metagenomic taxonomic classification results, both the DDH and ANI values showed that Streptococcus sp. v1. nov. was highly similar to Streptococcus gwangjuense ChDC B345, Streptococcus mitis NCTC 12261, Streptococcus pseudopneumoniae CCUG 49455, Streptococcus pneumoniae NCTC 7465, and Streptococcus oralis subsp. dentisani CECT 7747. Streptococcus gwangjuense isolated from human pericoronitis is closely related to the mitis group of the genus Streptococcus (Park et al. 2019).
Table 1
The digital DDH and ANI values of comparisons between Streptococcus sp. v1. nov. and 5 type strains of related species.
Strains
|
DDH (%)
|
ANI (%)
|
Streptococcus gwangjuense ChDC B345 T
|
53.1
|
93.48
|
Streptococcus mitis NCTC 12261 T
|
54.7
|
92.84
|
Streptococcus pseudopneumoniae CCUG 49455 T
|
48.8
|
92.3
|
Streptococcus pneumoniae NCTC 7465 T
|
46.6
|
91.86
|
Streptococcus oralis subsp. dentisani CECT 7747 T
|
31.7
|
86.98
|
T: Type strain. |
The TCS function of JSpeciesWS can rapidly compare selected genomes against a continuously updated reference database (ftp://ftp.ncbi.nlm.nih.gov/genomes/genbank) (Richter et al. 2016). TCS analysis showed that the draft genome of Streptococcus sp. v1. nov. was closest to the genome of Streptococcus mitis SK321, with a Z-score of 0.99859. To identify genetic differences between the two strains, we performed variant calling.
Comparing Streptococcus sp. v1. nov. with the reference Streptococcus mitis SK321, a total of 39979 high-quality SNPs were identified, including 28914 transitions (Ti) and 11065 transversions (Tv), with an average Ti/Tv ratio of 2.61. A > G|T > C type (14497) and G > A|C > T (14417) type accounted for the majority of all SNPs. The gain and loss of mutated genes are perceived as one of the most important contributors to functional changes (Nei and Rooney 2005). A total of 1588 InDels were identified, consisting of 809 insertions and 779 deletions. Moreover, 29 SVs containing 17 deletions, 1 inversion, and 11 translocations were uncovered in Streptococcus sp. v1. nov. genome. For CNV, 2 duplications and 43 deletions were identified (Fig. 2). Genomic variation is an important evolutionary driving force (Zhou et al. 2019).
3.5 Pan-core genome analysis
To find different characteristics between Streptococcus sp. v1. nov. and Streptococcus mitis, a comparative genome analysis for Streptococcus sp. v1. nov. and TCS top 25 hit Streptococcus strains (Additional Table 1) was performed by the BPGA pipeline. The core–pan plot showed that as the number of given genomes increased, the number of core genes decreased gradually (Fig. 3a). The core genome curve set plateaued, while the pan-genome trend curve grew continuously, indicating an open pan-genome and a conserved core genome. With the addition of new genomes, the number of genes in the core genome decreased from 1915 to 824, and the number of pan genome genes increased from 1915 to 5120. The pan-genome of all the strains had 718 core genes, 2591 accessory genes and 1810 unique genes. When compared with these related Streptococcus strains, Streptococcus sp. v1. nov. had the most accessory genes. A dendrogram of core genes showed that Streptococcus sp. v1. nov. had the closest phylogenetic relationship with Streptococcus mitis SK321 (Fig. 3b).
Functional analyses of COGs in the 26 Streptococcus genomes revealed that the highest proportion of core genome genes were related to “metabolism”, and the majority of unique gene families were mostly associated with “information storage and processing”. Previous research suggests that information storage and processing categories are linked to intracellular survival (Yang et al. 2018). The core genes were mainly enriched in [J] translation, ribosomal structure & biogenesis, [R] general function prediction only, and [E] amino acid transport & metabolism. In the categories of [K] transcription, [L] replication, recombination & repair, [M] cell wall/membrane/envelope biogenesis, and [V] defense mechanisms, the accessory and unique genes accounted for a greater proportion than the core genes (Fig. 4a). The KEGG analysis revealed that genes associated with “metabolism” accounted for the largest proportion of the core, accessory, and unique genomes. Among these genes, most were related to “carbohydrate metabolism”, “membrane transport”, “overview”, “amino acid metabolism”, “replication and repair”, and “nucleotide metabolism” (Fig. 4b).
In our new strain Streptococcus sp. v1. nov., 86 genes were unique. KEGG pathway annotations of these unique genes indicated that most KEGG Orthologies (KOs) were associated with the ABC transporter system, including ATP-binding cassette, subfamily B, bacteria (K06147); putative ABC transport system permease protein (K02004); putative ABC transport system ATP-binding protein (K02003); raffinose/stachyose/melibiose transport system permease protein (K101190); ABC-2 type transport system ATP-binding protein (K01990); and raffinose/stachyose/melibiose transport system permease protein (K10118) (Table 2). Pathogens acquire essential nutrients from the host by select ABC transporters, which enables rapid adaptation to changing host microenvironments and mediates toxicity (Tanaka et al. 2018)(Vitreschak et al. 2002)(Gutiérrez-Preciado et al. 2015). A total of 9 virulence factors were identified in the Streptococcus sp. v1. nov. genome using the virulence factor database (VFDB), including capsule, PavA, hyaluronic acid capsule, CBPs, PfbA, autolysin, and PsaA. Following the functional classification scheme in the VFDB, 71% of the virulence factors are related to “offensive function”, which contributes to successful infection of the host cells and tissues by colonization and toxicity (Heermann and Fuchs 2008). This result also explains to a certain extent why the patient’s condition progressed rapidly.
Table 2
KO annotation of the unique genes of Streptococcus sp. v1. nov.
KO
|
Definition
|
Score
|
K07407
|
alpha-galactosidase [EC:3.2.1.22]
|
369
|
K06147
|
ATP-binding cassette, subfamily B, bacterial
|
295
|
K09458
|
3-oxoacyl-[acyl-carrier-protein] synthase II [EC:2.3.1.179]
|
294
|
K00652
|
8-amino-7-oxononanoate synthase [EC:2.3.1.47]
|
263
|
K02004
|
putative ABC transport system permease protein
|
198
|
K02003
|
putative ABC transport system ATP-binding protein
|
184
|
K10119
|
raffinose/stachyose/melibiose transport system permease protein
|
168
|
K01990
|
ABC-2 type transport system ATP-binding protein
|
139
|
K10118
|
raffinose/stachyose/melibiose transport system permease protein
|
114
|