Strain 610A2^T was isolated from a 26-year-old male patient with pulmonary TB under treatment at the tuberculosis clinic of the Chaoyang District Center for Disease Control and Prevention (39.89 N, 116.40 E) in our previous surveillance under ethics approval No. ICDC-2022010 (Li et al., 2023). It was first identified as Stenotrophomonas by amplification and sequencing with 16S rDNA universal primers for bacteria (16S-U: 5′AGA GTT TGA TCM TGG CTC AG 3′ and/L: 5′ CCG TCA ATT CMT TTR AGT TT 3′).

Genome of 610A2^T was sequenced using the PacBio sequel II and DNBSEQ platform at the Beijing Genomics Institute (BGI, Shenzhen, China). Four SMRT cell zero-mode waveguide arrays for sequencing were used to generate the sub-read set. The PacBio subreads (length < 1 kb) were removed. Canu and GATK (https://www.broadinstitute.org/gatk/) were used for self-correction and single-base corrections, to improve the accuracy of the genome sequences.

Gene prediction was performed using glimmer3 with Hidden Markov models (http://www.cbcb.umd.edu/software/glimmer/). Other genomic elements, such as RNA, tandem repeats, genomic regions, small satellite DNA, and microsatellite DNA, were identified using tRNAscan SE, RNAmmer, Tandem Repeat Finder (http://tandem.bu.edu/trf/trf.html), Genomic Island Suite of Tools (http://www5.esu.edu/cpsc/bioinfo/software/GIST/), PHAge Search Tool (PHAST, http://phast.wishartlab.com), and CRISPRFinder. Eleven databases, namely, KEGG (Kyoto Encyclopedia of Genes and Genomes, http://www.kegg.jp), COG (Clusters of Orthologous Groups, http://www.ncbi.nlm.nih.gov), NR (Non-Redundant Protein Database databases, ftp://ftp.ncbi.nih.gov/blast/db/FASTA/nr.gz), Swiss-Prot (http://ngdc.cncb.ac.cn), GO (Gene Ontology, http://www.geneontology.org), TrEMBL, EggNOG (http://eggnog.embl.de), VFDB (Virulence Factors of Pathogenic Bacteria, http://www.mgc.ac.cn), ARDB (Antibiotic Resistance Genes Database, http://www.argodb.org/), CAZy (Carbohydrate-Active enZYmes Database, http://www.cazy.org), and T3SS (Type III secretion system effector protein, EggNOG database 1.0.1, http://effectivedb.org), were used for general function annotation and pathogenicity/drug resistance analysis.

Complete 16S rDNA gene sequences were extracted from the sequenced genome and reference genomes of 26 species of the genus Stenotrophomonas, except for S. detusculanense and S. indologenes, which do not have publicly available genome sequences. ClustalX was used for multiple sequence alignments, and a bootstrap neighbour-joining (NJ) phylogenetic tree was constructed using Treeview (1.6.6) with 1000 replicas. ANI was estimated using the ANI calculator (http://enve-omics.ce.gatech.edu/ani/). Syntenies were performed using MUMmer and BLAST. Core/Pan genes were clustered by CD-HIT rapid clustering of similar proteins software with a threshold of 50% pairwise identity and 0.7 length difference cutoff in amino acid. Hcluster_sg software was used to perform gene family clustering, and a phylogenetic tree was constructed using TreeBeST with the NJ method.

The strain’s growth ability was tested on 4 media, tryptone soy (TSA; Difco), Luria-Bertani (LB; Difco), Columbia blood agar (OXOID), and Mueller-Hinton (MH; Difco), at 37 °C for 24–72 h. Growth at different temperatures (4, 15, 20, 25, 28, 35, 37, 40, and 45 °C), pH (pH 4–11, at intervals of 1 pH unit, 37 °C) and NaCl concentrations (0–10%, w/v, in intervals of 1%) was also evaluated using LB broth and agar at 37 °C. Cellular motility was monitored using 0.75% semi solid agar medium, and morphological features were observed using a transmission electron microscope (HT7700; Hitachi, Japan). Utilisation of carbon sources and enzyme production were tested with API 20NE (48 h, 28 °C), API ZYM (4 h, 28 °C), and API 50 CH (inoculated with AUX medium, 48 h, 28 °C) (20050, 50300, 25200; BioMerier France, France), according to the manufacturer’s instructions. Cellular fatty acids and matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF) were analysed using the Sherlock Microbial Identification System (MIDI) with the MIDI Sherlock software program (version 6.3), RTSBA6 (6.21) library, and Autoflex speed TOF/TOF (Bruker Daltonics GmbH, Germany), according to the manufacturers’ instructions.

Two kinds of AST plates, the customised AST plate for Chinese Pathogen Identification Net (CHNENF, Trek Diagnostic Systems Ltd, West Sussex, United Kingdom) and the Sensititre^™ MYCOTB minimum inhibitory concentration (MIC) plate (Trek Diagnostic Systems, Cleveland, OH, USA), were used to determine the strain’s sensitivity to 17 drugs commonly used for gram-negative strains and 12 anti-TB drugs. MICs were determined according to the standards of the Clinical and Laboratory Standard Institute (CLSI, 2018) and manufacturer’s instructions.

Biofilm formation was determined using crystal violet staining. Fresh bacterial suspension was prepared and adjusted to 0.5 McFarland. After 1:30-fold dilution with tryptone soy broth (TSB; Difco), 150 μL/well suspension was added to a 96-well flat-bottom microtitre plate, and the plate was incubated at 37 °C. At each time point, bacterial turbidity (OD_{630 nm}) was measured using an enzyme-linked immunosorbent assay (ELISA) reader (Bio-Rad 680; Bio-Rad Laboratories, Japan). The cultures were discarded, and the wells were washed 3 times with PBS (pH 7.3) to remove planktonic cells. The biofilms were stained with 200 μL/well of 0.1% crystal violet for 15 min. The wells were washed again, and 200 μL of 99% ethanol was added and OD_{403 nm} was measured after 15 min. The biofilm index was calculated as OD_{403 nm}/OD_{630 nm}. Six clinically isolated S. maltophilia strains, including strain 11066 (GDMCC 1.4335), were used as controls.

Animal experiments were conducted with the approval of the Laboratory Animal Welfare & Ethics Committee of the National Institute for Communicable Disease Control and Prevention (Issue number 2023-021).

Using a random number method, female Kunming mice aged 6–7 weeks and weighing 30–35 g were randomly divided into infection and control groups. Five mice in each infection group were intranasally inoculated with 50 μL of 1.5 × 10⁸ CFU/mL of tested strains, and mice in the control groups were inoculated with PBS. At 4 h, 12 h, 1 day, 2 days, 3 days, 5 days, and 7 days post-infection, the mice were sacrificed and the lung and spleen tissues were collected aseptically.

The left lobes of the lungs were fixed in 10% neutral formalin for 24 h, followed by tissue processing and paraffin embedding. The paraffin blocks were sectioned at 2 µm, stained with haematoxylin and eosin, and blindly examined under a microscope by an expert in the field of laboratory animal pathology. Pulmonary clearance and occurrence of disseminated infection were monitored via quantitative bacteriology of the lung and spleen homogenates, respectively. Briefly, tissues were homogenised (1000 rpm) on ice in 1 mL of sterile normal saline by using a homogeniser (RZ-GR96A; Beijing Guoke Rongzhi Biotechnology Co., Ltd., China). Then, 100 μL of homogenates and 10-fold serial dilutions were inoculated on LB agar. The number of colonies was counted 24–48 h after incubation at 37°C. Bacterial colony counts were normalised according to the wet tissue weight and calculated as CFU/g. The levels of murine tumour necrosis factor alpha (TNF-α), granulocyte-macrophage colony-stimulating factor (GM-CSF), gamma interferon (IFN-γ), interleukin (IL)-2, IL-4, IL-6, IL-10, IL-12p70, and IL-17A were quantified using Luminex technology (R&D Systems, Minneapolis, MN, USA) and reagents (12002798; BIO-RAD, USA).

For comparison of multiple groups, one-way ANOVA was performed using GraphPad Prism version 9.3.1 for Windows (GraphPad Software, San Diego, California, USA).

The total length of 610A2^T complete genome was 4,681,496 bp, and GC content was 63.29% with an N50 length of 11,086 bp. A total of 4,070 CDSs, 70 tRNAs, 6 5S rRNAs, 4 16S rRNAs, 4 23S rRNAs, 37 sRNAs, 353 tandem repeats, 175 minisatellite DNAs, and 90 microsatellite DNAs were annotated.

In the phylogenetic tree based on the complete 16S rDNA gene (1545 bp), 610A2^T was clustered in a clade including S. terrae, S. nitritireducens, S. humi, and S. pictorum and formed a subclade between S. terrae and S. humi ( Figure 1 ). Dispensable gene heat maps and phylogenetic trees based on corePan and genefamily1 results showed completely consistent positioning ( Figures 2A, B ). When compared with the 26 Stenotrophomonas type species, 610A2^T displayed lower ANI values ranging from 87.53 to 80.22, less than 95%, the generally accepted threshold for species (Maturana and Cardenas, 2021; Deng et al., 2022). Synteny analyses at the nucleotide and amino acid levels were compared with 8 type strains: S. humi DSM 18929, S. sepilia SM16975, S. geniculata FLMAT1, S. maltophilia NCTC10257, S. terrae DSM 18941, S. nitritireducens DSM 12575, S. rhizophila DSM 14405, and S. acidaminiphila T0-18. S. terrae had the highest homology, with a gene median identity of 91.34% and map length rate of 63.327% ( Table 1 ). These data suggest that 610A2^T represents a novel species within the genus Stenotrophomonas, and it is closely related to, but distinct from, S. terrae. The name Stenotrophomonas pigmentata sp. nov. is proposed for the distinct brown water-soluble pigment it produces.

At the infection dose of 50 μL of 1.5 × 10⁸ CFU/mL, no mortality was observed in either the infection or control group. The mice infected with 610A2^T showed a weight loss of 3.9 ± 0.9%, 8.0 ± 0.7%, 6.4 ± 1.3%, 2.3 ± 1.2%, 0.8 ± 1.3% at 12 h, 1 day, 2 days, 3 days, and 5 days post-infection, respectively. On the seventh day, weight gain resumed. In addition, no significant differences were observed in the physical signs between the infection and control groups.

The bacterial load per tissue weight (g) in the lung measured via viable count was 1.1 ± 1.5 × 10⁷ CFU/mL·g at 4 h, 1.8 ± 3.6 × 10⁵ CFU/mL·g at 12 h, and 2.0 ± 4.0 × 10³ CFU/mL·g at 1 day. Bacteria could not be isolated from lung homogenates 2 days post-infection. 610A2^T was isolated only from the spleen homogenate of all infected mice within 4 h, with a viable bacterial count of 1.1 ± 2.2 × 10³ CFU/mL g, indicating that 610A2^T has strong invasive ability but the host can also quickly clear it.

Histological examination showed that pulmonary damage in the infected group constantly included alveolar septa thickening, congestion, bleeding, inflammatory cell infiltration, compensatory alveolar dilation, mucosal epithelial oedema, and vacuolar degeneration ( Figure 3 ). Although the viable count showed that the bacteria in the lungs had been cleared after 2 days, obvious bloody exudate and a large number of inflammatory cells were detected in the alveoli in 1–3 days, which is consistent with the symptoms for pulmonary congestion. After 7 days, the lung structure of the infected group was restored to normal, whereas that of the control group was normal.

Eight of the 9 quantified cytokines showed significant differences between infected and control groups ( Figure 4 ). The means of TNF-α, GM-CSF, IL-4, IL-6, IL-10, and IL-17A at each timepoint were higher than those of the control group, but only the values at 4 h had statistical significance because of the variability caused by individual differences. IFN-γ and IL-2 showed a consistent trend of change, lagging behind other cytokines and showing significant differences from the control group on the 5th day; however, at this time point, TNF-α and IL-6 had recovered to levels no different from the control group.

Functional annotation revealed that the 610A2^T genome contained 304 virulence factors of pathogenic bacteria in VFDB and 100 human disease pathway genes in KEGG database, of which 16 genes were related to bacterial infectious diseases. In details, 53 genes were related to flagellar biosynthesis, flagellar basal body protein, flagellar motor switching, and chemotaxis; 32 type IV pili genes; 3 outer membrane usher protein genes; 24 iron/haeme uptake and utilisation relative genes; and 1 biofilm-controlling response regulator. 610A2^T harbours 2 haemolysin-encoding genes, hlyB and hlyIII, as well as the haemolysin activation/secretion protein gene fhaC, a transcriptional regulator of haemolysin slyA, and 3 HlyD family secretion proteins, which are consistent with the haemolytic activities on Columbia blood agar containing sheep blood.

The annotation results showed that 610A2^T may possess 4 types of secretion systems, consisting of 3 T3SS genes, 5 T4SS effectors, 1 T5SS gene, 19 T6SS genes (tssA-H, tssK, tssL, and tssM), concentrated in 2 regions of the chromosome. In addition, 610A2^T harbours 9 lipopolysaccharide biosynthesis genes; 11 haeme biosynthesis and utilisation genes; 4 superoxide dismutase genes (Fe-Mn and Cu-Zn families); 1 superoxide oxidase gene; 3 glutathione peroxidase genes; 2 metalloendopeptidase OMA1; 1 melanin-producing gene cluster containing MarR transcriptional regulator, 4-hydroxyphenylpyruvate dioxygenase, and homogentisate 1,2-dioxygenase; 3 catalase genes; 12 alginate biosynthesis and regulation genes; 2 nitrate reductase genes; 1 isocitrate lyase gene; and 1 succinate dehydrogenase gene.

Owing to the slow growth rate of 610A2^T, its AST could not be tested using the BD Phonix-100 equipment (BD, Sparks, MD, USA) with BD Phonix™ NMIC 413 panels. Among the drugs detected with CHNENF and MYCOTB plates, 610A2^T was sensitive to 10 drugs: azithromycin (8 μg/mL), chloramphenicol (16 μg/mL), tetracycline (8 μg/mL), streptomycin (16 μg/mL), tigecycline (0.5 μg/mL), ofloxacin (1 μg/mL), moxifloxacin (0.06 μg/mL), and rifampin (0.12 μg/mL), with amikacin (<4 μg/mL) and colistin (<0.25 μg/mL) being the most sensitive. 610A2^T was resistant to ertapenem (8 μg/mL) and rifabutin (0.5 μg/mL). Fifteen drugs exceeded the maximum detection concentrations of the assay: cycloserine (>256 μg/mL), kanamycin (>40 μg/mL), para-aminosalicylic acid (>64 μg/mL), ethionamide (>40 μg/mL), isoniazid (>4 μg/mL), ethambutol (>32 μg/mL), cefotaxime (>16 μg/mL), ceftazidime (>32 μg/mL), ceftazidime/avibactam (>16/4 μg/mL), meropenem (>2 μg/mL), ampicillin/sulbactam (>32/16 μg/mL), ciprofloxacin (> 0.12 μg/mL), nalidixic acid (> 32 μg/mL), ampicillin (> 32 μg/mL), and trimethoprim-sulfamethoxazole (TMP-SMX, > 8/152 μg/mL).

Comparison against the ARDB, CARD, and KEGG databases revealed that 610A2^T carried 41 antibiotic-resistant relative genes on its genome, including 12 genes, aac6-I, acrb, adeb, ceob, emre, macb, mexc, mexe, oprn, smed, smee, and ykkd, coding for multi-drug resistance efflux pump for chloramphenicol, aminoglycoside, fluoroquinolone, macrolide, and tetracycline resistance; extended-spectrum beta-lactamase (ESBL) gene blaCTX-M, subclass B3 metallo-β-lactamase gene blaB, and pbp1a and pbp2b for cephalosporin, penicillin, and β-lactam resistance; rosA and rosB for fosmidomycin resistance; as well as some special resistant genes, bacA for bacitracin, ksgA for kasugamycin, dfrA-26 for trimethoprim, and vanA for vancomycin and teicoplanin.

610A2^T grew in all 4 tested media. The colonies on nutrient agar were translucent, smooth, and moist, with brown water-soluble pigment ( Figure 5A ); on blood agar, they were greyish white with a stimulating ammonia odour and a haemolytic ring ( Figure 5B ). The strain grew under conditions of 15–37 °C, pH 6–8, and NaCl concentration of 0–2%, but it did not grow below 4 °C and above 40 °C and did not tolerate 5% salt concentration like S. maltophilia and S. terrae. During semi-solid puncture culture, bacteria grew diffusely along the puncture line; however, no pigment was produced, indicating motility; pigment production was aerobic. 610A2^T was rod-shaped gram-negative, and without spores or capsules. Under electron microscopy, the bacterial body was about 1.5–2 × 0.5 μM with a single extreme flagella ( Figure 5C ).

610A2^T grew slower than S. maltophilia strains in TSB. After 48 h, the cultures of 610A2^T showed turbidity, but the entire growth curve was flat without a logarithmic phase ( Figure 6A ). Although the growth rate was slow, the biofilm index showed that 610A2^T had a strong biofilm-forming ability, which was significantly higher than that of all S. maltophilia strains after 3 h of cultivation ( Figure 6B ).

The biochemical results showed that, in the API 20NE test, 610A2^T was positive for nitrate reduction, hydrolysis of quercetin and gelatin, assimilation of N-acetylglucosamine, maltose, and citric acid. In API ZYM, it was positive for alkaline phosphatase, esterase lipase (C8), leucine arylamidase, trypsin, acid phosphatase, and naphthol-AS-BI phosphate hydrolase. In API 50CH, only starch fermentation was detected. The major differentiating feature of 610A2^T was that it could not assimilate glucose and mannose, but it could ferment starch.

In the MALDI-TOF spectrum, the 4 main peaks of 610A2^T were located at 5254.065 Da, 4857.675 Da, 2770.713 Da, and 6118.187 Da, of which the 4857 Da peak was genus-specific and the 2770.713 Da peak was 610A2^T-specific. In the local Bruker_MSP library, only 5 Stenotrophomonas species were found: S. acidaminiphila, S. maltophilia, S. pictorum, S. rhizophila, and S. nitritireducens. The identification scores of 610A2^T for these species ranged from 1.625 to 1.370, with the highest similarity to S. pictorum. These data indicate that 610A2^T is a distinct species closely associated with these species of Stenotrophomonas.

The predominant fatty acids of 610A2^T were iso-C_15:0, iso-C_14:0, summed Feature 3 (C_16:1ω7c/_16:1ω6c), iso-C_16:0, and C_15:0 anteiso and sum in Feature 9 (iso-C_17:1ω9c). It contained all 3 fatty acids, iso-C_11:0, iso-C_11:03 3OH, and iso-C_13:0 3OH, characteristic of the genus Stenotrophomonas, but no matches were found in RTSBA6. 610A2^T also contained all characteristic fatty acids of S. terrae and S. humi: iso-C_14:0 (14.2% and 15.7%), iso-C_15:1 (4.6% and 2%), iso-C_16:0 (8% and 12.7%), and iso-C_17:1ω9c (7.2% and 4.6%, respectively) (Heylen et al., 2007). The contents of these 4 fatty acids in 610A2^T were 15.45%, 1.5%, 11.9%, and 5.22%, respectively, indicating that the similarity of 610A2^T to S. humi is higher than that to S. terrae.

Stenotrophomonas pigmentata sp. nov. (pig.men.ta’ta. L. fem. adj. pigmentata, pigmented, coloured).

The type strain 610A2^T was isolated from patients with pulmonary TB. It is a Gram-negative bacillus with a single extreme flagella and can produce brown water-soluble pigment. Growth is observed at 15–37 °C, pH 6–8, and 0–2% NaCl concentration, but it does not grow below 4 °C, above 40 °C, and at 5% salt concentration. 610A2^T yielded positive results for nitrate reduction, quercetin and gelatin hydrolysis, and starch fermentation. The 4 main peaks in the MALDI-TOF spectrum were located at 5254.065 Da, 4857.675 Da, 2770.713, Da and 6118.187 Da. The predominant fatty acids are iso-C_15:0, iso-C_14:0, summed Feature 3 (C_16:1ω7c/_16:1ω6c), iso-C_16:0, and C_15:0 anteiso and sum in Feature 9 (iso-C_17:1ω9c). 610A2^T has multi-drug resistance and intrinsic resistance to β-lactams, carbapenems, and trimethoprim-sulfamethoxazole, harbouring several multi-drug resistance efflux pump and antibiotic resistant genes. This strain is pathogenic to mice.

The complete genome of type strain 610A2^T comprises 4681496 bp and GC content of 63.29%, and the complete genome and 16S rDNA gene sequence are deposited in GenBank under accession numbers CP130832.1 and OR936313, respectively. The type strain is 610A2^T (=GDMCC 1.4134 ^T =JCM 36488 ^T).