P. winnipegensis strain JUPW001 was isolated from a blood sample of a 17-year-old woman patient at Juntendo University Hospital, Tokyo, Japan in September 2022. The patient was admitted for suspected central line-associated bloodstream infection due to recurrent fever and elevated C-reactive protein (CRP). Treatment with piperacillin/tazobactam (4.5 g/day) was started and her symptoms improved 2 days after onset. Two sets of blood cultures taken at the time of fever were positive after 36 hours. Three days after treatment, blood cultures were negative. The samples of the positive blood culture were stained and were cultured on AccuRate™ separated Sheep Blood Agar/Chocolate Agar EXII (Shimadzu Diagnostics Co., Kyoto, Japan) and BTB lactose agar (Eiken Chemical Co. Ltd., Tokyo, Japan) at 35°C. The type strain of P. winnipegensis NCTC 14396^T was obtained from National Collection of Type Cultures (Salisbury, UK). Type strains of P. kaohsiungensis CCUG 55854^T, P. mexicana CCUG 49454^T, P. japonensis CCUG 48231^T, and P. spadix CCUG 53828^T were obtained from the Culture Collection, University of Göteborg (CCUG), Sweden. All isolates were cultured aerobically at 35°C on 5% sheep blood agar plates (Becton, Dickinson Diagnostic Systems, MD, USA) or Luria-Bertani broth under aerobic conditions at 35 °C.

The genomic DNA of the clinical isolate was extracted using DNeasy blood and tissue kits (Qiagen, Tokyo, Japan) and Genomic-tips 20/G (Qiagen). For short-read sequencing, DNA library was prepared using Nextera XT DNA Library Prep Kit (Illumina, San Diego, CA, USA). The genomes were sequenced on the Illumina MiniSeq platform (300 cycles). The raw reads were trimmed and assembled using CLC Genomic Workbench version 10.0.1 (CLC bio, Aarhus, Denmark). For long-read sequencing, DNA library was prepared using Native Barcoding Kit 24 V14 SQK-NBD 114.24 (Oxford Nanopore Technologies, Oxford, UK). Sequencing was performed on R10.4.1 flowcell using MinION Mk1B (Oxford Nanopore Technologies). MinKNOW (version 24.02.8) and Guppy (version 7.3.11) (Oxford Nanopore Technologies) were used for base calling and adapter trimming of raw data. The sequences determined by MiniSeq and MinION were assembled using Unicycler v0.5.0 (Wick et al., 2017). The genome relatedness of the relevant strains was estimated using an average nucleotide identity (ANI) calculator (Yoon et al., 2017) and a Type (Strain) Genome Sever³. ANI values were calculated using reference genomes from P. winnipegensis (NCTC 14396^T; genome accession number GCF_004283755). Virulence factors were detected by the virulence factor database (VFDB⁴) and MacSyfinder v2.0 (Abby et al., 2014) for secretion systems (SS).

Genome completeness and contamination were assessed using CheckM2 v1.0.1 with lineage_wf and default settings (Parks et al., 2015). A phylogenetic tree was constructed using the kSNP4 software based on pangenome SNPs⁵ (Gardner et al., 2015), and visualized using iTol ver.6⁶. The strains of P. broegbernensis (DSM 12573^T; NZ_JACHGU010000001), P. daejeonensis (DSM 17801^T; NZ_PDWN01000010), P. dokdonensis (DSM 21858^T; NZ_LDJL01000001), P. gei (KCTC 32298^T; NZ_QOVG01000010), P. helianthin (NBRC 110414^T; NZ_JAGKTC010000001), P. indica (CCM 7430^T; NZ_BMCL01000001), P. japonensis (CCUG 48231^T; NZ_PDWW01000010), P. kalamensis (DSM 18571^T; NZ_PDWQ01000001), P. kaohsiungensis (CCUG 55854^T; NZ_PDWO01000010), P. koreensis (KCTC 12208^T; NZ_PDWM01000010), P. mexicana (CCUG 49454^T; NZ_PDWV01000100), P. putridarboris (LMG 25968^T; NZ_JBBWWT010000001), P. sacheonensis (DSM 19373^T; NZ_PDWS01000010), P. sangjuensis (DSM 28345^T; NZ_PDWR01000010), P. spadix (CCUG 53828^T; NZ_RDQN01000001), P. suwonensis (DSM 17175^T; NZ_PDWP01000010), P. taiwanensis (DSM 22914^T; NZ_PDWK01000100), P. winnipegensis (NCTC 14396^T; NZ_SHMH01000001), P. wuyuanensis (DSM 100640^T; NZ_PDWU01000010), P. yeongjuensis (DSM 18204^T; NZ_PDWT01000010), Stenotrophomonas maltophilia (NCTC 10257^T; GCF_900186865.1) were used as reference strains.

Theoretical masses of proteins encoded in the genomes of Pseudoxanthomonas were calculated for the following genomes as part of the development of a genomically predicted protein mass database toolkit (GPMsDB-tk): P. japonensis (CCUG 48231^T; NZ_PDWW01000010), P. kaohsiungensis (CCUG 55854^T; NZ_PDWO01000010), P. mexicana (CCUG 49454^T; NZ_PDWV01000100), P. spadix (CCUG 53828^T; NZ_RDQN01000001), and P. winnipegensis (NCTC 14396^T; NZ_SHMH01000001) (Sekiguchi et al., 2023). The genome sequences were obtained from the NCBI database⁷. The prediction of genes from the genomes obtained in this study was performed using GPMsDB-dbtk v1.0.1⁸ (Sekiguchi et al., 2023).

Alpha-cyano-4-hydroxycinnamic acid (CHCA) was used as a matrix. To prepare this matrix solution, 10 mg of 4-CHCA was dissolved in 1 mL of solvent consisting of 1% (v/v) trifluoroacetic acid, 35% (v/v) ethanol, 15% (v/v) acetonitrile, and milliQ water. A full loop of bacterial cells was dispersed in 200 μL of distilled water in a microtube and mixed with 800 μL of ethanol. The suspensions were briefly vortexed and centrifuged at 15, 000 g for 2 min. The pellets were then dried for 5 min. The pellets were suspended in 50 μL of 70% formic acid, vortexed, suspended in 50 μL of acetonitrile, and centrifuged at 15, 000 g for 2 min. Supernatants were analyzed by MALDI-MS according to the manufacturer’s instruction.

MALDI-MS measurements were performed in positive linear mode using MALDI-8020 RUO (Shimadzu Corporation, Kyoto, Japan) and Microflex LT/SH (Bruker Daltonics, Germany) equipped with a 200 Hz Nd: YAG laser (355 nm) and 60 Hz nitrogen laser (337 nm), respectively. Before sample analysis, the MALDI-MS instrument was mass-calibrated externally using six peaks with m/z 4365.4, 5381.4, 6411.6, 7274.0, 8369.8, and 10300.1 from Escherichia coli DH5α. More than five individual mass spectra were acquired for each bacterial extract in the range of m/z 2, 000-20, 000. The assignment of the peak was performed using eMSTAT Solution™ software (Shimadzu Corporation). Species identifications were performed by the MicrobialTrack software v1.1.0 (Shimadzu Corporation) (Sekiguchi et al., 2023; Nakagawa et al., 2025) and MBT Compass 4.1 with Microflex LT/SH (Bruker Daltonics).

For biomarker validation, the theoretical mass of L29, L33 and CspA was calculated for 15 Pseudoxanthomonas type strains: P. broegbernensis (DSM 12573^T; NZ_JACHGU010000001), P. daejeonensis (DSM 17801^T; NZ_PDWN01000010), P. dokdonensis (DSM 21858^T; NZ_LDJL01000001), P. gei (KCTC 32298^T; NZ_QOVG01000010), P. helianthin (NBRC 110414^T; NZ_JAGKTC010000001), P. indica (CCM 7430^T; NZ_BMCL01000001), P. kalamensis (DSM 18571^T; NZ_PDWQ01000001), P. koreensis (KCTC 12208^T; NZ_PDWM01000010), P. putridarboris (LMG 25968^T; NZ_JBBWWT010000001), P. sacheonensis (DSM 19373^T; NZ_PDWS01000010), P. sangjuensis (DSM 28345^T; NZ_PDWR01000010), P. suwonensis (DSM 17175^T; NZ_PDWP01000010), P. taiwanensis (DSM 22914^T; NZ_PDWK01000100), P. wuyuanensis (DSM 100640^T; NZ_PDWU01000010), and P. yeongjuensis (DSM 18204^T; NZ_PDWT01000010). A phylogenetic tree was constructed using SRplot with the unweighted pair group method with arithmetic mean (UPGMA) (Tang et al., 2023).

The whole-genome sequence of P. winnipegensis JUPW001 obtained in this study has been deposited in GenBank under the accession number AP044739⁹.

Samples of the positive blood culture were stained as Gram-negative rod. After 24 h of incubation, a large number of oxidase-positive and yellow-colored colonies were observed on the agars. The isolate was identified as “Pseudoxanthomonas species” using MALDI-MS with MBT Compass 4.1 by Microflex LT/SH (Bruker Daltonics).

The ANI value using whole-genome sequence of the isolate was 97.4% identical to the sequence of P. winnipegensis (NCTC 14396^T; NZ_SHMH01000001). Therefore, the isolate was confirmed to be P. winnipegensis and designated as P. winnipegensis JUPW001.

P. winnipegensis JUPW001 had flagellum, type 1 secretion system (T1SS), T2SS, T4SS, T5SS and type IVa pilus (T4aP). In particular, T4SS included VirB2 to VirB6 and VirB8 to VirB11.

As shown in Figure 1, the phylogeny was separated by the genus level between Pseudoxanthomonas and Stenotrophomonas. The phylogenetic tree revealed three clades: A, B, and C. P. winnipegensis JUPW001 belonged to clade B and the species was close to P. spadix (Figure 1). In clade A, P. kaohsiungensis (GenBank accession no. NZ_PDWO01000010) was isolated from humans (Kuo and Lee, 2018). Other previously reported clinical isolates, P. japonensis (accession no. NZ_PDWW01000010) and P. mexicana (accession no. NZ_PDWV01000100), belonged to clade C (Thierry et al., 2004). The other species were mainly isolated from soil or water environments.

The bacterial identification for Pseudoxanthomas species was performed using MicrobialTrack, which is a software for peak matching by predicted mass values using the annotated protein sequences. The MicrobialTrack analysis revealed that P. winnipegensis JUPW001, P. winnipegensis NCTC 14396^T, P. spadix CCUG 53828^T, P. japonensis CCUG 48231^T, P. mexicana CCUG 49454^T, and P. kaohsiungensis CCUG 55854^T were correctly identified with high reliability (Table 1). In this analysis, the number of ribosomal proteins peaks in each strain were from 19 to 28 and annotated proteins except for ribosomal proteins were from 8 to 19. In contrast, Microflex LT/SH analysis using the Biotyper database correctly identified P. kaohsiungensis, P. mexicana, and P. spadix at the species level, whereas P. japonensis was misidentified and P. winnipegensis was classified only at the genus level (Table 1).

The candidate mass peaks of MALDI-MS analysis for the five type strains, including P. kaohsiungensis, P. japonensis, P. mexicana, P. spadix and P. winnipegensis, are shown in Figure 2. In the MALDI-MS profiles, 21 major mass peaks were commonly detected and successfully annotated with predicted protein names in P. winnipegensis JUPW001 and NCTC 14396^T (Table 2). Of these 21 annotated peaks, 17 were predicted as ribosomal subunit proteins, and the remaining four were co-chaperonin GroES, cold shock-like protein (CspA), DNA-binding protein HU and translation initiation factor IF-1. As shown in Table 2, compared to the theoretical mass peaks of P. winnipegensis NCTC 14396^T, most of the peaks were different from the theoretical mass peaks of P. spadix CCUG 53828^T, P. japonensis CCUG 48231^T, P. mexicana CCUG 49454^T, and P. kaohsiungensis CCUG 55854^T. To distinguish five Pseudoxanthomonas at the species level using MALDI-MS, the combination of appropriate peaks of ribosomal L29, L33 and CspA were selected as biomarkers.

As shown in Figure 2 and Table 2, the corresponding theoretical peaks of ribosomal protein L29 were m/z 6983.1 for P. winnipegensis NCTC 14396^T; m/z 7565.7 for P. spadix CCUG 48231^T; m/z 7257.5 for P. japonensis CCUG 48231^T and m/z 7032.1 for P. kaohsiungensis CCUG 55854^T (Figure 2; Table 2). The peak at m/z 7296.6 for L29 of P. mexicana CCUG 49454^T were not detected, although the theoretical peaks were calculated by GPMsDB-dbtk. The corresponding theoretical peaks of the ribosomal protein L33 were m/z 6039.0 for P. winnipegensis NCTC 14396^T and P. spadix CCUG 48231^T; m/z 6163.2 for P. japonensis CCUG 48231^T and P. mexicana CCUG 49454^T; m/z 6122.2 for P. kaohsiungensis CCUG 55854^T (Figure 2; Table 2). The corresponding theoretical peaks of CspA were m/z 7155.9 for P. winnipegensis NCTC 14396^T; m/z 7183.9 for P. spadix CCUG 48231^T; m/z 7141.9 for P. japonensis CCUG 48231^T; m/z 7125.9 for P. mexicana CCUG 49454^T and m/z 7178.9 for P. kaohsiungensis CCUG 55854^T (Figure 2; Table 2).

Compared to the amino acid sequence of L29 in P. winnipegensis, there were 6 amino acid substitutions in P. spadix, 8 in P. japonensis, 8 in P. mexicana and 11 in P. kaohsiungensis (Figure 3A). Compared to the amino acid sequence of L33 in P. winnipegensis, there were 5 substitutions in P. japonensis, 5 in P. mexicana, and 6 in P. kaohsiungensis (Figure 3B). Compared to the amino acid sequence of CspA of P. winnipegensis, there were 1 substitution in P. spadix, 4 in P. japonensis, 3 in P. mexicana, and 10 in P. kaohsiungensis (Figure 3C).

MALDI-MS proteotyping was able to distinguish P. winnipegensis, P. spadix, P. japonensis, P. mexicana and P. kaohsiungensis using the three biomarkers, including L29, L33 and CspA.

Cluster analysis using L29, L33 and CspA revealed that 20 Pseudoxanthomonas type strains were distinguished by the three theoretical mass peaks (Supplementary Table S1 and Supplementary Figure S1). The close cluster such as P. japonensis, P. putridarboris and P. wuyuanensis were theoretically separated by L29 or CspA, whereas P. sacheonensis and P. yeongjuensis were theoretically separated by CspA (Supplementary Figure S1).