Abstract

We describe a rapidly growing mycobacterial (RGM) species closely related to Mycobacterium wolinskyi. The bacterium was isolated from three biopsies obtained from a 78-year-old Venezuelan woman who presented with a chronic superficial erythematous plaque in the right nasal cavity. The three isolates, LTG2003(1–3), exhibited identical phenotypic, biochemical, and molecular characteristics and showed susceptibility to amikacin and linezolid, intermediate susceptibility to cefoxitin and imipenem, and resistance to ciprofloxacin, clarithromycin, sulfamethoxazole-trimethoprim, and tobramycin. Conventional diagnostic methods, including biochemical assays, hsp65 PCR restriction enzyme analysis (PRA-hsp65), the Genotype Mycobacterium CM v2.0 linear probe assay, and Sanger sequencing of the 16S rRNA, hsp65, and rpoB genes, failed to identify the microorganism to the species level. MALDI-TOF mass spectrometry yielded genus-level scores suggestive of M. wolinskyi. Whole-genome sequencing confirmed that the isolates represent a new species. The type strain, designated LTG 2003-1 (WFCC 1312 = P7483 = CDBB-500-3), has been deposited in two internationally recognized microbial culture collections in Mexico and Brazil. The genome of this organism has also been deposited in GenBank as Mycobacterium sp. LTG2003 with accession number JBONDW000000000. Based on our findings, we propose that the LTG2003(1-3) isolates represent a new species, for which we propose the name Mycobacterium venezuelense sp. nov.

Introduction

Nontuberculous mycobacteria (NTM) are environmental microorganisms that are generally nonpathogenic to humans, except in individuals with underlying immunosuppression or other predisposing conditions (Cook, 2010; Ratnatunga et al., 2020). Once considered of limited clinical relevance, NTMs are now recognized as important opportunistic pathogens, driven in part by increasing global detection rates and advances in diagnostic methods (Moore et al., 2010; Johnson and Odell, 2014; Donohue, 2018). Although pulmonary disease is the most common clinical presentation, NTMs are also responsible for infections of the skin, soft tissues, and postsurgical wounds, including cases reported in Venezuela (Weitzul et al., 2000; Piquero et al., 2004; Henkle and Winthrop, 2015; Torres-Coy et al., 2016; Zimmermann et al., 2017; Pérez-Alfonzo et al., 2020; Da Mata-Jardín et al., 2020; García-Ruza et al., 2024).

Rapidly growing mycobacteria (RGM), a subgroup of NTMs, were historically regarded as having low virulence (Falkinham, 1996; De Groote and Huitt, 2006; Mohananey et al., 2018). However, their clinical significance has increased, particularly among immunocompromised patients. RGMs are frequently associated with postsurgical and device-related infections, outbreaks in healthcare settings, and they often exhibit intrinsic or acquired multidrug resistance, complicating therapeutic management (van Ingen et al., 2012). Clinically relevant species include Mycobacterium abscessus, Mycobacterium fortuitum, and Mycobacterium chelonae, among others (Brown-Elliott and Wallace, 2002; Brown-Elliott and Philley, 2017).

Isolation and identification of RGMs remain challenging. These organisms may grow more slowly than other bacteria in polymicrobial samples and can therefore be missed or overgrown in routine cultures (Han et al., 2007). In addition, accurate species-level identification is difficult because traditional biochemical methods lack sufficient discriminatory power. Molecular approaches, including PCR–restriction enzyme analysis of hsp65 (PRA-hsp65) and partial Sanger sequencing of the 16S rRNA, hsp65, and rpoB genes, have improved diagnostic accuracy but remain insufficient to reliably differentiate rare, emerging, or closely related species. Although MALDI-TOF mass spectrometry (MS) has substantially increased the speed of mycobacterial identification, its performance is limited by the incompleteness of available reference spectral libraries, hindering the recognition of novel taxa (Wang et al., 2025).

The taxonomic classification of the genus Mycobacterium has undergone major revisions in recent years, including a proposal to divide the genus into five separate genera—Mycobacterium, Mycolicibacterium, Mycolicibacillus, Mycolicibacter, and Mycobacteroides—based on shared molecular markers (Gupta et al., 2018). Subsequent phylogenomic analyses, however, recommended retaining a single unified genus due to the clinical, diagnostic, and practical implications of renaming medically relevant species (Meehan et al., 2021). As both nomenclatures continue to appear in the literature, the present study adopts the traditional genus name Mycobacterium for clarity and consistency (Tortoli et al., 2019).

Here, we report the case of a 78-year-old woman from Guárico State, Venezuela, who presented with a chronic granulomatous lesion of the nostril persisting for more than two years. A rapidly growing mycobacterium was isolated from three independent biopsies; however, multiple diagnostic approaches—including PRA-hsp65, the GenoType Mycobacterium CM v2.0 - line probe assay (Hain Lifescience GmbH, Nehren, Germany), and partial sequencing of the 16S rRNA, hsp65, and rpoB genes—failed to achieve species-level identification. MALDI-TOF MS (Bruker, Billerica, USA) suggested affiliation with the M. fortuitum group but yielded scores below the species-identification threshold. Whole-genome sequencing (WGS) using Illumina technology ultimately demonstrated that the three isolates represent a novel species, for which we propose the name Mycobacterium venezuelense sp. nov.

Materials and methodsClinical case presentation

A 78-year-old woman with diabetes mellitus, residing in Guárico State, Venezuela, was clinically evaluated in June 2009 for a persistent lesion in her right lower nasal cavity (photographic record not provided). According to the patient, the lesion first appeared in May 2008 following a traumatic laceration of the nasal cavity with soft tissue tearing. Transient clinical improvement was observed after a short course of cefoxitin 1 g every 12 h for 6 days; however, the lesion recurred shortly after each treatment. Subsequent treatment with cefadroxil 500 mg every 24 h for 7 days, prescribed at the Dermatology Center of the Dr. Jacinto Convit Institute of Biomedicine at the Central University of Venezuela, Caracas, proved ineffective, and the infection progressively worsened. A biopsy, submitted for histopathological analysis, revealed the presence of granulomas (report not provided). The patient subsequently received empirical intravenous treatment for 4 months with linezolid 600 mg every 12 h and amikacin 500 mg every 12 h for 4 weeks, resulting in complete clinical resolution of the lesion. She was declared cured in March 2010.

Primary mycobacterial culture in Venezuela

Because infection with a nontuberculous mycobacterium (NTM) was suspected, an initial biopsy specimen was submitted to the Tuberculosis Laboratory of the same institute for decontamination, culture, assessment of colony morphology, and species identification using PCR–restriction enzyme analysis of hsp65 (PRA-hsp65). The biopsy fragment was decontaminated by incubation in a 0.75% hexadecylpyridinium chloride (HPC) solution for 2 min, following the method described by Gao et al. (2005). The HPC was then removed by brief rinsing in sterile water, after which the tissue was homogenized in phosphate-buffered saline (PBS). The resulting suspension was inoculated onto Löwenstein–Jensen (LJ) medium and incubated at 37 °C. Cultures were examined for growth every 2–3 days. After 6 days, acid-fast bacteria were isolated. PCR–restriction enzyme analysis of hsp65 (PRA-hsp65) (Telenti et al., 1993) was used to identify the strain; however, the resulting restriction pattern was inconclusive and did not correspond to any known Mycobacterium species. Two additional biopsy specimens obtained subsequently yielded isolates with PRA-hsp65 patterns identical to that of the initial strain. The three isolates were designated LTG2003-1 to LTG2003-3. The patient received empirical treatment for 4 months with linezolid and for 4 weeks with amikacin, resulting in complete clinical resolution of the lesion. She was declared cured in March 2010.

All three isolates were sent to the Laboratory of Bacteriology and Bioassays in Tuberculosis and other Mycobacteria (LBBTM), National Institute of Infectology (INI), Oswaldo Cruz Foundation (FIOCRUZ), Rio de Janeiro, Brazil, for comprehensive species identification. Analyses included phenotypic characterization, drug susceptibility testing, line probe assays, Sanger sequencing of the 16S rRNA, rpoB, and hsp65 genes, MALDI-TOF mass spectrometry, and whole-genome sequencing. At LBBTM/INI, the isolates were cataloged as LTG2003(1), LTG2003(2), and LTG2003(3) under collection code BK912/719.

Morphological and phenotypic characterization in Brazil

Acid-fast bacilli were confirmed by Ziehl–Neelsen (ZN) staining. Colony morphology on Löwenstein–Jensen (LJ) medium, 7H10, and 7H10 + PANTA. Growth characteristics, including pigment production, temperature tolerance (30 °C, 37 °C, and 42 °C), and growth rate—were evaluated under both light and dark incubation conditions. Pigment production was assessed by exposing cultures incubated in the dark to bright white fluorescent light for 3 h, followed by an additional 18-h incubation to differentiate photochromogenic from scotochromogenic species (Forbes et al., 2018).

Classical biochemical assays were performed as described by Kent et al. (1985). These included growth on LJ medium supplemented with 5% NaCl; growth on MacConkey agar without crystal violet; citrate and iron utilization; Tween 80 hydrolysis; potassium tellurite and nitrate reduction; urea hydrolysis; and the semiquantitative catalase test.

Identification by PCR–restriction enzyme analysis of hsp65 (PRA)

A 441-bp fragment of the hsp65 gene was amplified by PCR using standard primers. The PCR products were digested separately with the restriction enzymes BstEII and HaeIII, and the resulting fragments were resolved by electrophoresis on a 3% agarose gel. This approach generates species-specific banding patterns characteristic of individual Mycobacterium species (Telenti et al., 1993). The observed restriction patterns were compared with established PRA-hsp65 databases and published reference tables for species identification.

Identification by Genotype Mycobacterium molecular line-probe assay

Species identification was further attempted using the Genotype Mycobacterium CM v2.0 line-probe assay (LPA; Hain Lifescience, Nehren, Germany), which detects the Mycobacterium tuberculosis complex (MTBC) and 13 clinically relevant NTMs, including M. avium, M. intracellulare, M. abscessus complex, the M. fortuitum group, M. kansasii, M. gordonae, M. xenopi, and others. DNA was extracted from pure cultures using the GenoLyse kit (Hain Lifescience, Nehren, Germany), followed by multiplex PCR targeting the 23S rRNA gene, according to the manufacturer’s instructions. Biotinylated amplicons were hybridized to species-specific probes immobilized on nitrocellulose strips, and resulting banding patterns were visually interpreted (Lee et al., 2009; Caulfield et al., 2019).

Molecular identification

Identification was performed by partial sequencing the 16S rRNA, rpoB, and hsp65 genes. DNA was extracted from cultured isolates using the BIO GENE K205-2 gDNA extraction kit (described later). Target gene regions were amplified using published primers and protocols specific to mycobacteria (Cooksey et al., 2004). Amplicons were purified using the commercial GFX PCR DNA and Gel Band Purification Kits (Cytiva, Massachusetts, USA) and subjected to Sanger sequencing on an ABI 3730xL sequencer, in both forward and reverse directions. The resulting sequences were compared to reference sequences in the NCBI GenBank database using the BLASTn algorithm. Species identification was based on maximum sequence similarity, applying the established thresholds of ≥99% identity for the 16S rRNA gene and ≥97% for the hsp65 and rpoB genes (Kim and Shin, 2018; Hall et al., 2003). Phylogenetic analyses were performed to support identification when BLAST results were inconclusive.

Identification by MALDI-TOF mass spectrometry

This technique is based on the ionization of sample proteins and the subsequent separation of ions by mass, generating a mass spectral profile that is compared against a reference protein database (Oliva et al., 2021). MALDI-TOF MS was performed using the Bruker Biotyper system (Bruker Daltonics, Karlsruhe, Germany) with the Mycobacteria RUO Library v4.0 (Bruker Daltonics, Karlsruhe, Germany) for spectral matching. Colonies were subjected to ethanol-formic acid extraction following the manufacturer’s protocol (Genc et al., 2018). Spectral acquisition was carried out using a Microflex LT/SH system, and analysis was performed with Bruker Biotyper software. Identification scores ≥2.0 were considered reliable for species-level identification (Costa-Alcalde et al., 2019).

Drug susceptibility testing

DST of the RGM was performed using the commercial Sensititre™ RAPMYCO1 susceptibility test plate assay (Thermo Fisher Scientific, Waltham, USA), which includes 15 antibiotics: amikacin, cefoxitin, moxifloxacin, ciprofloxacin, clarithromycin, doxycycline, imipenem, linezolid, trimethoprim-sulfamethoxazole, tobramycin, cefepime, amoxicillin/clavulanic acid, ceftriaxone, minocycline and tigecycline (Wetzstein et al., 2020). All tests were performed in triplicate and plates were incubated at 30 °C for three to 7 days. Test interpretation was according to the Clinical and Laboratory Standards Institute (CLSI) M24S, which recommends Mycobacterium peregrinum ATCC® 27853 and Staphylococcus aureus ATCC ® 29213 as microorganisms for quality control, using breakpoints described for RGM (CLSI, 2023; Bhalla et al., 2018; Woods et al., 2011) and published data (CLSI, 2018; Brown-Elliott et al., 2016).

DNA extraction, purification, and whole-genome sequencing

For DNA preparation from the three isolates, bacterial cells from three loops were resuspended in 200 μL of 1X Tris-EDTA (TE) buffer. After vigorous vortexing for 5 min, the suspension was incubated at 95 °C for 30 min for bacterial inactivation. Genomic DNA was then extracted using the BIO GENE gDNA Extraction Kit K205-2 (Bioclin, Belo Horizonte, Brazil) according to the manufacturer’s instructions. DNA quality and concentration were assessed using the Qubit® dsDNA High Sensitivity (HS) Assay Kit (Thermo Fisher Scientific, Waltham, USA). Whole-genome sequencing (WGS) was performed on an Illumina MiSeq platform (Illumina, San Diego, CA, USA). Sequencing libraries were prepared using the Nextera XT Kit (Illumina), generating 2 × 150 bp paired-end reads.

Genomic analysis and initial specie identification

WGS data were analyzed using the Bacass v2.3.1 pipeline, implemented in Nextflow DSL2 with Docker/Singularity containers for reproducibility. Genome assembly was carried out with Unicycler v0.4.8 and Kraken2 (standard 8 GB database, April 2025) was used to assess contamination, and results were visualized using Krona plots (Wick et al., 2017; Wood et al., 2019). CheckM v1.2.3 was used to evaluate genome completeness (Gurevich et al., 2013; Parks et al., 2015). To determine taxonomic placement, average nucleotide identity (ANI; using the NCBI Prokaryotic Genome Annotation Pipeline (PGAP) taxonomy verification tool) and digital DNA–DNA hybridization (dDDH; using the automated Type Strain Genome Server (TYGS) platform, Meier-Kolthoff and Göker, 2019; Meier-Kolthoff et al., 2014) values were calculated by comparing the genome with the genomes of Mycobacterium species present in GenBank. Species assignment was based on cutoffs of ANI > 95% and dDDH > 70%; subspecies assignment used ANI ≤ 97% and dDDH ≤ 80% (Iversen et al., 2025). Annotation was performed using both the NCBI Prokaryotic Genome Annotation Pipeline (PGAP) and the RAST server, which provide structural and functional annotation for prokaryotic genomes.

Phylogenetic analysis

Phylogenetic relationships were inferred using Mashtree, which combines the min-hash algorithm from the Mash tool (Ondov et al., 2011) with the neighbor-joining (NJ) algorithm from QuickTree (Howe et al., 2002). A Mash distance matrix was computed and used to construct a dendrogram with default parameters. The resulting tree was visualized and annotated using iTOL (Interactive Tree of Life) (Meier-Kolthoff et al., 2013; Kreft et al., 2017). Species labels corresponding to M. wolinskyi genomes (GCF accessions from GenBank) were included for comparison with the genomes of the three LTG2003 isolates (1, 2, and 3).

Taxonomic identification

Phylogenetic trees based on ANI values were constructed to complement the Mash-based analysis, providing further resolution of genomic relatedness among the isolates and reference strains. WGS data was uploaded to the Type (Strain) Genome Server (TYGS) platform2 for detailed genome-based taxonomic analysis (Meier-Kolthoff and Göker, 2019; Meier-Kolthoff et al., 2022). TYGS calculates ANI and dDDH values and constructs phylogenomic trees based on WGS comparisons. This analysis was supplemented with current nomenclature and taxonomic information from the TYGS’s linked database and List of Prokaryotic names with Standing in Nomenclature (LPSN).3 The results from TYGS, received on 31st of March of 2025, provided further resolution for species delineation and taxonomic placement (Turenne, 2019).

ResultsPhenotypic and biochemical characterization LTG2003(1–3)

All three isolates, LTG2003(1–3) (later renamed BK912/719(1–3) at the LBBTM, INI-FIOCRUZ), formed smooth, creamy, non-pigmented colonies that grew at 30 °C, 37 °C, and 42 °C after 5 days of incubation on either L-J or 7H10 + PANTA medium, with no differences in morphology (Figure 1 and Supplementary Figure 1). Colonies on 7H10 medium were photographed after 5 days, highlighting their non-pigmented appearance. As the three isolates originated from separate biopsies of the same lesion and exhibited identical colonial morphology, each was independently subjected to biochemical testing and molecular analyses.

(A) Cultures of isolates LTG2003(1–3) grown on 7H10 + PANTA medium (upper panel). (B) Colonies from each culture, captured using a stereoscopic microscope at 1.6X magnification (lower panel). Cultures of the isolates LTG2003(1–3) on 7H10 + PANTA medium (upper line), and image of the colonies of each culture, captured with a stereoscopic microscope at 1.6X (lower line).

All isolates tested positive for urea hydrolysis, tellurite reduction (on days 7 and 10), arylsulfatase activity (on day 15), and semi-quantitative catalase. They also showed growth on Löwenstein-Jensen medium supplemented with NaCl, iron, and citrate (after 3 weeks). Weak positivity was observed for Tween 80 hydrolysis, and on MacConkey agar, weak positivity appeared on day 11 for the first two isolates and complete positivity for the last isolate. Negative results were specific for the pigmentation test, indicating an absence of pigment production, classifying the isolates as non-chromogenic whites (Group IV - Runyon classification). When compared with the identification scheme proposed by Fernández de Vega F. et al. (2005), the biochemical profile most closely resembled that of M. mageritense, M. smegmatis, and M. wolinskyi. However, M. wolinskyi typically tests negative for arylsulfatase after 3 days, whereas our isolates showed weak positivity at 3 days and full positivity at 15 days for this test. Table 1 presents a summary of the phenotypic and biochemical results.

Sample codesUreaTween 80 HydrolysisTellurite ReductionArylsulfataseSemi-quantitative catalaseGrowth in MacConkeyGrowth in NaClIron UptakeCitrateGrowth ratePigmentation18 h3d5d10d7d10d3d15d15d5d11d15d3S15d3S15d3S7d15d7d15dAfter exposed to lightLTG2003 (1)POSPOSPOS-weakPOS-weakPOSPOSPOS-weakPOSPOSNEGPOS-weakPOSPOSPOSPOS–POS<7d–NEGNEGWhite-Non-chromogenic (Group IV)LTG2003 (2)POSPOSPOS-weakPOS-weakPOSPOSPOS-weakPOSPOSNEGPOS-weakPOSPOSPOSPOS–POS<7d–NEGNEGWhite-Non-chromogenic (Group IV)LTG2003 (3)POSPOSPOS-weakPOS-weakPOSPOSPOS-weakPOSPOSNEGPOSPOSPOSPOSPOS–POS<7d–NEGNEGWhite-Non-chromogenic (Group IV)

Phenotypic and biochemical characteristics of the three isolates (Kent et al., 1985).

Identification by molecular techniques

PRA-hsp65 method yielded restriction patterns (BstEII: 231/131/79-bp; HaeIII: 139/117/58/51/40/36-bp), that did not correspond to any previously described species (Telenti et al., 1993. Devallois et al., 1997). Supplementary Figure 2 presents the results of the Genotype Mycobacterium CM assay, which classified the three isolates within the M. fortuitum group. MALDI-TOF MS produced low-confidence genus-level scores ranging from 1.63 to 1.82, all suggesting M. wolinskyi but falling below the ≥2.0 threshold required for species assignment (Table 2). Partial Sanger sequencing of the 16S rRNA, rpoB, and hsp65 genes also failed to definitively identify the isolates. Phylogenetic trees for each marker (Figures 2–4), constructed using the neighbor-joining method with 1,000 bootstrap replicates, consistently placed the isolates within a clade closely related to M. wolinskyi, however, bootstrap support was insufficient for definitive species-level resolution. As expected, the 16S rRNA gene lacked discriminatory power. While rpoB and hsp65 provided greater resolution, their sequences remained inconclusive for assignment to a recognized species (GenBank accession numbers SUB15741956, SUB15743129, and SUB15743112, respectively). These findings prompted whole-genome sequencing.

Sample codesIdentification – MALDI TOFScoreLTG2003(1)Mycobacterium wolinskyi1.63LTG2003(2)Mycobacterium wolinskyi1.82LTG2003(3)Mycobacterium wolinskyi1.69

Results of MALDI-TOF mass spectrometry analysis of isolates LTG2003(1), (2), and (3), performed at the Bacteriology and Bioassays Reference Laboratory, National Institute of Infectology – FIOCRUZ.

Score interpretation: 2.00–3.00: High-confidence species-level identification. 1.70–1.99: Low-confidence (genus-level) identification. 0.00–1.69: Identification not reliable or not possible.

Phylogenetic tree inferred from GBDP distances based on 16S rRNA gene sequences of LTG2003 (1–3), using FastME 2.1.6.1 (Lefort et al., 2015). Branch lengths are scaled according to the GBDP distance formula d5. Numbers above branches represent GBDP pseudo-bootstrap support values >60% from 100 replicates; the average branch support was 71.3%. The tree was rooted at the midpoint (Farris, 1972).

Phylogenetic tree of the LTG2003 strains (1, 2, and 3) using neighborhood-joining statistical methods and the Kimura two-parameter nucleotide substitution model for the rpoB gene. M. litorale was used as an outgroup to root the tree. A comparative analysis was performed using partial sequencing data from LTG2003 strains and FASTA sequences of the rpoB gene available in GenBank from the species with the greatest ANI similarity to the study samples.

Phylogenetic tree of LTG2003 strains (1–3) using neighborhood-joining statistical methods and the Kimura two-parameter nucleotide substitution model for the hsp65 gene. Mycobacterium litorale was used as an outgroup to root the tree. A comparative analysis was performed using partial sequencing data from the LTG2003 strains and FASTA sequences of the hsp65 gene available in GenBank from the species with the greatest ANI similarity to the study samples.

Antibiotic susceptibility of the LTG2003 strains

Accurate species identification, DST, are essential for selecting effective therapies against mycobacterial infections. In this case, MIC testing was performed on the three isolates in triplicate. The isolates were resistant in each test (of the triplicate) to ciprofloxacin, clarithromycin, sulfamethoxazole, trimethoprim, and tobramycin; showed intermediate susceptibility to cefoxitin, moxifloxacin, and imipenem; and remained susceptible to amikacin and doxycycline. The MIC value for tigecycline was 0.12 μg/mL, for cefepime was >32 μg/mL, for amoxicillin/clavulanic acid was 16/8 μg/mL, for ceftriaxone was >64 μg/mL, and for minocycline was 1 μg/mL for each sample in each triplicate (Table 3).

AntibioticsLTG2003(1)LTG2003(2)LTG2003(3)MIC quality control ranges (μg/mL)MIC (μg/mL)MIC categoriesMIC (μg/mL)MIC categoriesMIC (μg/mL)MIC categoriesM. peregrinum ATCC® 700,686S. aureus ATCC ® 29,213Amikacin1S1S1S≤1–41–4Cefoxitin32I32I32I4–321–4Moxifloxacin2I2I2I≤0.06–0.250.016–0.12Ciprofloxacin4R4R4R≤0.12–0.50.12–0.5Clarithromycin16R16R16R≤0.06–0.50.12–0.5Doxycycline0.12S0.12S0.12S0.12–0.50.12–0.5Imipenem8I8I8I2–160.016–0.06Linezolid1S1S1S1–81–4Trimethoprim-sulfamethoxazole8/152R8/152R8/152R≤0.25/4.8–2/38≤0.5/9.5Tobramycin16R16R16R2–80.12–1Tigecycline0.12–0.12–0.12–0.03–0.250.03–0.25Cefepime>32–>32–>32––aAmoxicillin/clavulanic acid16/8–16/8–