Systematic investigation of the emerging pathogen of Tsukamurella species in a Chinese tertiary teaching hospital

ABSTRACT Tsukamurella species have been clinically regarded as rare but emerging opportunistic pathogens causing various infections in humans. Tsukamurella pneumonia has often been misdiagnosed as pulmonary tuberculosis due to its clinical presentation resembling tuberculosis-like syndromes. Tsukamurella species have also been confused in the laboratory with other phylogenetic bacteria, such as Gordonia. This study aimed to investigate the clinical, microbiological, and molecular characteristics; species distribution; and antimicrobial susceptibility of Tsukamurella species. Immunodeficiency and chronic pulmonary disease appeared to be risk factors for Tsukamurella pneumonia, and the presence of bronchiectasis and pulmonary nodules on imaging was highly correlated with this infection. The study confirmed that groEL (heat shock protein 60) and secA (the secretion ATPase) genes are reliable for identifying Tsukamurella species. Additionally, the ssrA (stable small RNA) gene showed promise as a tool for discriminating between different Tsukamurella species with the shortest sequence length. In terms of antimicrobial susceptibility, quinolones, trimethoprim/sulfamethoxazole, amikacin, minocycline, linezolid, and tigecycline demonstrated potent in vitro activity against Tsukamurella isolates in our study. The study also proposed a resistance mechanism involving a substitution (S91R) within the quinolone-resistance-determining region of the gyrA gene, which confers resistance to levofloxacin and ciprofloxacin. Furthermore, we found that disk diffusion testing is not suitable for testing the susceptibilities of Tsukamurella isolates to ciprofloxacin, imipenem, and minocycline. In conclusion, our systematic investigation may contribute to a better understanding of this rare pathogen. Tsukamurella species are rare but emerging human pathogens that share remarkable similarities with other mycolic acid–containing genera of the order Actinomycetales, especially Mycobacterium tuberculosis. Consequently, misdiagnosis and therapeutic failures can occur in clinical settings. Despite the significance of accurate identification, antimicrobial susceptibility, and understanding the resistance mechanism of this important genus, our knowledge in these areas remains fragmentary and incomplete. In this study, we aimed to address these gaps by investigating promising identification methods, the antimicrobial susceptibility patterns, and a novel quinolone resistance mechanism in Tsukamurella species, utilizing a collection of clinical isolates. The findings of our study will contribute to improve diagnosis and successful management of infections caused by Tsukamurella species, as well as establishing well-defined performance and interpretive criteria for antimicrobial susceptibility testing.

Accurate and rapid identification of the pathogen is crucial for diagnosing and managing infectious diseases.Tsukamurella species share many common features with other phylogenetic bacterial groups comprising Nocardia, Gordonia, Rhodococcus, Corynebacterium, and Mycobacterium (5,18).Gordonia, a genus that bears phenotyp ical similarity to Tsukamurella, has been increasingly isolated in our laboratory and often misidentified as Tsukamurella species by conventional phenotypic identification method or matrix-assisted laser desorption ionization-time-of-flight mass spectrome try (MALDI-TOF MS) system.Traditionally, accurately identifying Tsukamurella species using conventional phenotypic or biochemical methods in most clinical microbiology laboratories is challenging, leading to the underestimated incidence of Tsukamurella infection (35,36).Several molecular methods have been demonstrated to be compara tively rapid and reliable tools for identifying Tsukamurella species, including sequencing of 16S ribosomal RNA (rRNA), the essential secretory protein-encoding gene secA (the secretion ATPase), rpoB (beta-subunit of DNA-dependent RNA polymerase), groEL (60 kDa chaperonin and heat shock protein), and transfer messenger RNA coding gene ssrA (stable small RNA) (37)(38)(39).Among these five gene targets mentioned, the 16S rRNA gene has been considered impractical for distinguishing species within the Tsukamur ella genus owing to its highly conserved sequence (35).However, it has been found to be sufficiently discriminative in identifying Tsukamurella species from phylogeneti cally related groups (5,35,40).Regarding the other four housekeeping genes, Teng et al. recommended that sequencing the groEL gene displayed good performance in identifying Tsukamurella species at the species level (37).Furthermore, Pérez et al. (41) provided evidence of the secA1 gene being a useful target for differentiating Tsukamur ella pulmonis from other clinically significant Tsukamurella species.Nevertheless, the discriminatory capacity of these genes warrants further evaluation using a broader range of clinical isolates, encompassing species other than T. pulmonis and Tsukamurella tyrosinosolvens (37,41).However, identification by sequencing is still beyond the reach of many routine clinical microbiology laboratories.MALDI-TOF MS has been proven as a potential routine identification tool for Tsukamurella species (29).Teng et al. found that the MALDI-TOF MS showed good performance for genus-level identification of 60 Tsukamurella isolates using a commercial database but insufficient performance for species-level identification.However, 98.3% of Tsukamurella isolates were correctly identified to species level with score ≥2.0 using the expanded in-house database (29).In light of varying antimicrobial susceptibilities among different Tsukamurella spp., the continuous expansion of MALDI-TOF MS databases with available spectral data for each Tsukamurella species, as well as other phylogenetically related bacteria, would be highly beneficial in enabling accurate, rapid, and cost-effective identification of Tsukamurella species in routine clinical laboratories.
To date, the knowledge of well-defined susceptibility or resistance patterns of Tsukamurella species is fragmentary and primarily derived from sporadic reports.The Clinical and Laboratory Standards Institute (CLSI) guidelines have recommended the broth microdilution method for antimicrobial susceptibility testing (AST) of aerobic actinomycetes (42).In addition, CLSI M62 has proposed tentative breakpoints for aerobic actinomycetes, which depend mainly upon organism population distributions, clinical data, breakpoints used for other organisms, and experience of experts in the field (43).CLSI has also emphasized the necessity to gather more data for the optimal manage ment of infections caused by aerobic actinomycetes including Tsukamurella species (43).The disk diffusion method is easy and convenient for AST of Nocardia species and the other aerobic actinomycetes in many clinical laboratories (18,44).However, well-estab lished interpretive criteria for testing Tsukamurella species or other aerobic actinomy cetes using the disk diffusion method are lacking.The breakpoints for Staphylococcus spp. or Gram-positive bacteria are usually used to perform agar disk assays with Actinomycetales (32,44,45).Tsukamurella isolates are generally susceptible to amika cin, ciprofloxacin, imipenem, doxycycline, vancomycin, linezolid, and sulfamethoxazole but resistant to penicillin, oxacillin, piperacillin/tazobactam, and cephalosporin (1,38).Nevertheless, the susceptibility patterns may differ among various Tsukamurella spp.(46).Good clinical outcomes and favorable prognoses are highly dependent on accurate identification and appropriate antibiotic therapy with source control (1).Accordingly, it is necessary to clarify the clinical and microbiology features, adequate diagnostic methods, and antimicrobial susceptibility to effectively treat Tsukamurella infections.
In the present study, we retrospectively collected clinical isolates that were initially identified as Tsukamurella species along with relevant clinical information in the Clinical Microbiology Laboratory of Peking Union Medical College Hospital, Beijing, China, from 2018 to 2022.We aimed to investigate the clinical, microbiological, and molecu lar characteristics; species distribution; and antimicrobial susceptibility of Tsukamurella species.

Patients and bacterial isolates
We analyzed 22 clinical isolates (from 17 patients) of Tsukamurella and Gordonia species, which were initially recognized as Tsukamurella species by the Clinical Microbiology Laboratory of Peking Union Medical College Hospital, Beijing, China, between January 2018 and October 2022.A total of 22 preserved isolates from 17 patients were ini tially identified as Tsukamurella spp.using a MALDI-TOF MS system (Autof MS1000; Autobio Labtec Instruments Co, Ltd., Zhengzhou, China).The samples were prepared as previously described (47), and the identification of isolates was performed using Autof Acquirer V2.0.18 (Autof MS1000, Autobio) following the manufacturer's criteria for species and genus identification (47).The Autof MS1000 database contained 4 species of Tsukamurella and 12 species of Gordonia, encompassing Tsukamurella inchonensis, Tsukamurella paurometabola, T. pulmonis, T. tyrosinosolvens, Gordonia aichiensis, Gordonia alkanivorans, Gordonia amicalis, Gordonia australis, Gordonia bronchialis, Gordonia humi, Gordonia kroppenstedtii, Gordonia otitidis, Gordonia rubripertincta, Gordonia sihwensis, Gordonia sputi, and Gordonia terrae.Misidentification referred to discrepant identification results using the MALDI-TOF MS system compared with molecular methods.Chart review focused on 11 patients who were ultimately confirmed to have Tsukamurella infection, including 10 cases of probable Tsukamurella pneumonia and 1 case of CR-BSI.The patient records were evaluated to determine patient demographics, clinical characteristics of infection (including underlying diseases and pulmonary manifestations), laboratory test results, prior/empirical antibiotic use, therapy, and outcomes.Furthermore, all isolates were thoroughly phenotypically characterized, including observations on colony appearance and cell morphology using Gram stain, Ziehl-Neelsen stain, and modified Ziehl-Neelsen stain.

Phylogenetic analysis
The nucleotide sequences of 16S rRNA, secA, rpoB, and groEL for type strains of 16 Tsukamurella species and ssrA gene for 11 Tsukamurella species available in GenBank were included in this study as described by Teng et al. (2,37).The phylogenetic analysis was performed with Molecular Evolutionary Genetic Analysis software (version 11.0; https://megasoftware.net/) using the neighbor-joining (NJ) method, with all positions containing gaps and missing data eliminated from the data set.The significance of the cluster nodes was determined by bootstrapping with 1,000 randomizations, and the NJ tree was, therefore, generated.The phylogenetic trees were manipulated and adjusted using the online tool The Interactive Tree of Life (https://itol.embl.de)(49).The 16S rRNA, secA, rpoB, and groEL sequences of Mycobacterium tuberculosis (Gen Bank accession no.NR116692; CP074075: 3619823-3620337, CP074075: 760955-761238, and CP074075: 529466-530142, respectively) and ssrA sequences of Tsukamurella soli (GenBank accession no.KX932003) were used as outgroups.

Antimicrobial susceptibility testing
In vitro susceptibility was determined using the broth microdilution methodology according to the CLSI M24-3E protocol (42).The minimum inhibitory concentrations (MICs) of 34 tested antibiotics against the 22 isolates were determined and interpreted after 72-h incubation at 37°C (Tables 3 and 4) (43).
Ceftriaxone, amoxicillin/clavulanic acid (2:1), tobramycin, amikacin, minocycline, doxycycline, clarithromycin, vancomycin, imipenem, moxifloxacin, ciprofloxacin, trimethoprim/sulfamethoxazole, linezolid, and rifampin were interpreted with reference to the breakpoints for testing Nocardia spp.and other aerobic actinomycetes (43).Besides, the interpretation of cefoxitin and meropenem referred to the breakpoints for rapidly growing mycobacteria, whereas the interpretation of ethambutol and isoniazid referred to the M. tuberculosis complex (43).No breakpoints or interpretive criteria were available for some of the antimicrobial agents tested; therefore, only an MIC was reported without interpretation.
We routinely used the disk diffusion method to perform AST on 15 antimicrobial disks (Table 5) for aerobic actinomycetes on Mueller-Hinton agar with 5% sheep blood.The inocula were prepared according to the CLSI M24-3E protocol (42).The results of the inhibition zone diameter were recorded after 72-h incubation at 37°C and interpreted based on the thresholds adapted from the guidelines of the Antibiogram Committee of the French Society of Microbiology devoted to Gram-positive bacteria (50) and CLSI guidelines devoted to Staphylococcus aureus (51) (Table S1).S. aureus (ATCC 29213) and Escherichia coli (ATCC 35218) were used for quality control.

Nucleotide sequence accession numbers
The 16S rRNA, secA, rpoB, groEL, and ssrA gene sequences of 15 isolates assigned as Tsukamurella spp.were deposited in GenBank with accession numbers as shown in Fig. 2. Meanwhile, the 16S rRNA gene sequences of eight isolates assigned as Gordonia spp.were also deposited in GenBank (Table S2).

Clinical characteristics
A total of 15 isolates identified as Tsukamurella spp.were recognized as pathogens causing probable pulmonary infection in 10 patients (12 isolates) and CR-BSI in 1 patient (three isolates).The clinical characteristics of probable pulmonary infection and CR-BSI caused by Tsukamurella spp. in 11 patients are summarized in Tables 1  and 2, respectively.Of 10 patients with probable Tsukamurella pulmonary infection, 7 were aged ≥60 years, and 1 was a teenager.Out of 10 patients, 7 were female.Six patients received steroid administration, yet five patients had an immune disease and the other four patients had chronic lung disease.All the patients with proba ble Tsukamurella pulmonary infection presented with associated airway constitutional symptoms; pulmonary nodules on imaging seemed to suggest this kind of infection.In addition, 4 out of 10 patients had bronchiectasis, whereas other pathogens were isolated in 6 patients, including Pseudomonas aeruginosa, S. aureus, Enterobacter cloacae, and Pneumocystis jirovecii, at the same time.Tuberculosis was clinically suspected in five patients.Unfortunately, three patients had been on anti-tuberculosis therapy for years as an empirical diagnosis of negative tuberculosis.In particular, patient 10 was diagnosed with tuberculosis in a local hospital on the basis of a sputum smear showing acid-fast bacilli.While the misdiagnosis was discovered by sputum culture, it was positive for T. pulmonis in 2020 and for T. inchonensis in 2018 and 2019 but negative for mycobac terium tuberculosis culture.Five patients showed good improvement after treatment with levofloxacin, sulfamethoxazole, or trimethoprim.Furthermore, a 51-year-old woman presented to the general surgical department complaining of intermittent abdominal distension with vomiting for 3 months (Table 2).She underwent surgery for a mesentery tumor on 29 December 2021.Bacteremia caused by T. tyrosinosolvens was diagnosed in this patient with CR-BSI in the presence of a peripherally inserted central catheter (PICC) on 12 January 2022.The patient had a good clinical outcome after ciprofloxacin and amikacin therapy (Table 2).

Phenotypic characterization
The colonies of Tsukamurella isolates were white-grayish to yellow colored, dry, wrinkled, and resembled a rough membrane with irregular spreading fringes on Columbia blood agar plates after 48-h incubation at 37°C under aerobic conditions (Fig. 1A through F).G. bronchialis displayed white-grayish creamy colored, dry, and membrane-like colonies with irregular spreading fringes, similar to Tsukamurella isolates (Fig. 1J).Besides, the other Gordonia isolates exhibited white-grayish to yellow colored but smooth, butyrous, and circumscribed small colonies under the same condition (Fig. 1G through I).The colony appearance of the same Tsukamurella or Gordonia species may show different characteristics, yet different Tsukamurella and Gordonia species have similarities (Fig. 1A through J).Gram staining showed Gram-positive, slender, rod-shaped bacilli for Tsukamurella isolates or slightly short and thick cocci-rod bacilli for Gordonia isolates (Fig. 1K through M).The results of modified Ziehl-Neelsen staining (Fig. 1N) and Ziehl-Neelsen staining (Fig. 1P and Q) were positive and partially positive for Tsukamurella isolates, respectively.Ziehl-Neelsen staining and modified Ziehl-Neelsen staining results were negative for the Gordonia isolates in this study (Fig. 1O and R).

Phylogenetic analysis
A total of 16 Tsukamurella species (16S rRNA, secA, rpoB, and groEL) or 12 Tsukamurella species (ssrA gene) were employed to construct phylogenetic trees (Fig. 2).Phylogenetic trees were inferred from (A) partial 16S rRNA (1,221 nucleotide positions), (B) partial rpoB (284 nucleotide positions), (C) partial secA (607 nucleotide positions), (D) partial ssrA (366 nucleotide positions), and (E) partial groEL (677 nucleotide positions) gene sequence data using the NJ method (Fig. 2).Comparative 16S rRNA gene sequence analysis revealed a low sequence variability among species of the genus Tsukamurella (Fig. 2A).Some recently recognized novel species in the genus Tsukamurella, including T. sputi, T. conjunctivitidis, T. sinensis, T. ocularis, and T. hominis, could not be distinguished by 16S rRNA gene.The other four housekeeping genes showed higher variation than the 16S rRNA region.Each individual housekeeping gene assembled in the 15 Tsukamurella isolates (four species) in our study formed monophyletic clusters with the corresponding Tsukamurella type strain but exhibited intraspecies polymorphisms.Isolate 20TM02422 was the most isolated compared with other isolates in the T. pulmonis species in the above five loci of the phylogenetic tree, especially for the 16S rRNA gene.The T. tyrosinosolvens isolates 19TP0049 and 22TM02858 showed a more distant relationship with the other T. tyrosinosolvens isolates in four loci of the phylogenetic tree (except for 16S rRNA gene, Fig. 2).No reference ssrA sequence of T. sputi is available in the National Center for Biotechnology Information (NCBI) database.The ssrA sequence of T. sputi isolate 22NC06989 in this study was submitted with the accession number OP902932.Compared with the phylogenetic tree constructed from all five loci, the ssrA gene had the highest intraspecies discrimination with the shortest sequence length in the 12 closely related species of the genus Tsukamurella.Unfortunately, no ssrA sequences were available in the NCBI database for investigating the regional divergence of the four Tsukamurella species, including T. asaccharolytica, T. hominis, T. conjunctivitidis, and T. ocularis.
The MIC value of all the Tsukamurella isolates to ceftazidime was >32 µg/mL, whereas the MIC range of Gordonia isolates was 4-8 µg/mL.All the Tsukamurella isolates showed high MICs to cefoxitin and amoxicillin clavulanic acid (2:1), which were >128 and >64/32 µg/mL, respectively.However, all the Gordonia isolates showed higher susceptibility to cefoxitin and amoxicillin/clavulanic acid (2:1) with the MIC of 16 and ≤2/1 µg/mL, respectively.All the Tsukamurella and Gordonia isolates were susceptible to amikacin and linezolid with MIC values of ≤1-2 µg/mL.Except for one T. tyrosinosolvens isolate (21TM00974), which was intermediate to minocycline with an MIC of 2 µg/mL, the other Tsukamurella and all Gordonia isolates were susceptible to minocycline with an MIC of ≤1 µg/mL.The MIC values of all the Tsukamurella and Gordonia isolates against daptomycin were more than 2 µg/mL, while the MIC range against tigecycline was 0.12-0.5 µg/mL.Regarding obligate anti-tuberculosis agents, including isoniazid, ethionamide, and ethambutol, all the Tsukamurella and Gordonia isolates showed high resistance with MIC values of more than 8 µg/mL, more than 20 µg/mL, and 16 µg/mL, respectively.Overall, the Gordonia isolates showed higher susceptibilities than the Tsukamurella isolates to penicillin, chloramphenicol, aminogly cosides, tetracyclines, and macrolides (Tables 3 and 4).A clear differentiation between Tsukamurella and Gordonia spp. was also observed for antimicrobial susceptibilities to vancomycin (Tables 3 and 4).For carbapenems, only two T. pulmonis isolates (22TM01897 and 21TM01062) were not susceptible to imipenem and meropenem in this study (Table 3).One T. tyrosinosolvens isolate (22TM00764) demonstrated extremely high resistance to levofloxacin and ciprofloxacin but intermediate to moxifloxacin (Table 3).The other isolates in this study were susceptible to quinolones.Trimethoprim/sulfamethoxazole appeared to be a potent antimicrobial agent with strong in vitro antimicrobial activity (Tables 3 and 4).Moreover, rifampin was not a good choice against infection caused by Tsukamurella and Gordonia because of high non-susceptibility (Tables 3 and 4).
Based on the results of AST by disk diffusion method, the interpretation categories of all Tsukamurella and Gordonia isolates to amoxicillin/clavulanic acid, linezolid, and trimethoprim/sulfamethoxazole demonstrated 100% agreement with the results of the broth microdilution method in this study, with complete resistance to amoxicillin/clavu lanic acid in Tsukamurella isolates, complete susceptibility in Gordonia isolates, and all susceptibility to linezolid and trimethoprim/sulfamethoxazole irrespective of Tsukamur ella and Gordonia isolates (Tables 3 to 5).Conversely, the disk diffusion method for ceftriaxone susceptibility showed a 60% of susceptibility against Tsukamurella isolates compared with 100% susceptibility by the broth microdilution method (Tables 3 and  5).In addition, completely opposite interpretation categories to cefoxitin for Gordonia isolates and vancomycin for Tsukamurella isolates were found between these two methods (Tables 3 to 5).The disk diffusion method failed to detect the ciprofloxacin resistant isolate 22TM00764 with an inhibitory zone diameter of 38 mm.Although the inhibitory zone diameter of isolates 22TM00764 to moxifloxacin was 24 mm, it was still interpreted as susceptible based on the breakpoint adopted in this study (50,51).For carbapenems, the Tsukamurella isolates were found to be slightly more frequently non-susceptible to imipenem by the broth microdilution method than by the disk diffusion method (Tables 3 and 5).All the Gordonia isolates were found to be susceptible to cephalosporins using the disk diffusion method, whereas Tsukamurella isolates showed varying degrees of non-susceptibility to this class of agents, with the highest susceptibility rate of 100% to cefepime (Table 5).The disk diffusion test con firmed tigecycline as an in vitro active agent against Tsukamurella and Gordonia isolates (Tables 3 to 5).

Mutation within gyrA
The quinolone-resistance-determining region (QRDR) of the gyrA gene in Tsukamurella species was determined to stretch from amino acids 75 to 114 by multiple-sequence alignment with NCBI reference sequence of M. tuberculosis H37Rv (NP_214520.1)(52).In this study, a unique mutation S91R located within the QRDR of the gyrA gene occurred in the quinolone-resistant isolate of 22TM00764.It may be implicated that this DNA gyrase subunit substitution in 22TM00764 conferred high resistance to levofloxacin and ciprofloxacin but intermediate to moxifloxacin.We considered that this mutation in Tsukamurella species demonstrated identical functionality with the corresponding A90V mutation in the gyrA gene of M. tuberculosis H37Rv (NP_214520.1), which was the second most frequent mutation associated with lower level quinolone resistance and moxifloxacin effectiveness (M24-A3) (42).

DISCUSSION
Tsukamurella species are rare but emerging pathogens causing human infections at different body sites, with an underestimated prevalence of pulmonary infection because they share remarkable similar clinical features with other mycolic acid-containing genera of the order Actinomycetales, especially M. tuberculosis (1,3).In this context, the understanding of this rare infectious agent is needed to find methods for accurate and rapid identification, to better understand its epidemiology within hospitals, to increase knowledge of its clinical, microbiological and molecular characteristics, to establish specific criteria for susceptibility testing, and to investigate its mechanisms of resistance, which would otherwise remain fragmentary and incomplete.This study was novel in exploring the infections caused by various Tsukamurella species in China and aimed to elaborate their characteristics to improve the understanding of the clinical, microbiological, and molecular aspects, thus covering species distribution, antimicrobial susceptibilities, and quinolone resistance mechanism of Tsukamurella infections.
Considering that Tsukamurella lung disease manifests as tuberculosis-like syndromes (18), our study may be conceivable that pulmonary infection caused by Tsukamurella has been underestimated in China, the world's second-largest country with a high tuberculo sis burden (53).Among the 10 cases of pulmonary infection, T. inchonensis and T. sputi were described for the first time as potential pathogens of pulmonary infections (15).In this study, reinfection by another Tsukamurella species was observed in a patient with immunoglobulin G (IgG4)-related diseases for more than 3 years.Immunodeficiency and chronic pulmonary disease appeared to be risk factors for Tsukamurella pneumonia; the presence of bronchiectasis and pulmonary nodules on imaging was also highly correlated with this infection.Cases of pulmonary Tsukamurella infection were primar ily diagnosed as mycobacterial infection, with extreme side effects to empirical antituberculosis therapy (13).As anti-tuberculosis therapy is usually of prolonged duration, various adverse drug reactions may cause associated morbidity and even mortality if not recognized early (54).In our study, 5 out of 10 patients were suspected of having tuberculosis without isolation of mycobacterium tuberculosis.Of these, one patient was misdiagnosed as having tuberculosis due to the detection of acid-fast bacilli in a sputum smear, similar to a clinical case reported by Liu et al. (19), and three patients were receiving anti-tuberculosis therapy at the outset.Thus, an effective identification method is needed to develop an accurate diagnosis of Tsukamurella infections.Furthermore, six patients with Tsukamurella infection had good clinical outcomes after treatment with quinolones or sulfonamides.We also encountered a case of CR-BSI caused by T. tyrosinosolvens in a patient with underlying cancer who was immunosuppressed and underwent surgery within 30 days.She recovered after removal of the infected line and received antibiotic treatment of ciprofloxacin and amikacin.The combination of catheter removal and appropriate antibiotic treatment is considered an essential therapeutic method for the treatment of catheter-related infections caused by Tsukamurella species (55,56).
The phenotypic similarities of Tsukamurella, Gordonia, Rhodococcus, Actinomyces, Nocardia, and Mycobacterium have resulted in unreliable identification using pheno typic or commercially available automated systems (18).MALDI-TOF MS has been employed as a user-friendly, rapid, and cost-effective means of identifying bacteria, fungi, and mycobacteria in routine clinical laboratories (57).However, except for 2 isolates of T. inchonensis, the other 20 isolates were all misidentified, including specieslevel misidentification of 13 Tsukamurella isolates and genus-level misidentification of 7 Gordonia isolates under the MALDI-TOF MS platform in our routine study.Teng et al. reported that none of the 60 isolates belonging to five different Tsukamurella species were correctly identified at the species level by MALDI-TOF MS with the original Bruker database V.6.0.0.0 (29).However, after optimizing the database by adding, the mass spectrum profiles of all 11 currently recognized Tsukamurella species as a reference; 59 of the 60 isolates were correctly identified to the species level (29).Therefore, once the database has been optimized by adding more Gram-positive rods, MALDI-TOF MS may serve as a beneficial tool for the routine identification of Tsukamurella species in clinical microbiology laboratories (1,29,58).Although sequence-based identification is considered to be a comparatively rapid and reliable molecular technique for the identification of Tsukamurella species (41), it has not yet been adopted as a routine identification method in clinical microbiology laboratories because it is expensive, time-consuming, and technically demanding for this purpose (29).Consistent with previous studies (37,41,59), this study also suggested that 16S rRNA gene sequencing was insufficient to discriminate among Tsukamurella spp.but was enough to discriminate Tsukamurella and Gordonia.Four additional housekeeping genes, rpoB, secA, ssrA, and groEL, were successfully used in this study for species-level identification in 16 and 12 closely related species of the genus Tsukamurella, respectively.Contrary to previously recommended target genes of groEL and secA (37,41), we confirmed the identification performance of these two genes while also demonstrating the efficacy of the ssrA gene as a promising tool.The ssrA gene exhibited the highest intraspecies discrimination but the shortest sequence length in the 12 closely related species of the genus Tsukamurella.Nevertheless, the inclusion of more sequences from the four novel Tsukamurella species, including T. asaccharolytica, T. hominis, T. conjunctivitidis, and T. ocularis, may further help evaluate the efficacy of this gene for Tsukamurella identification.
Microbiological morphology is unreliable for accurate identification of species from the order Actinomycetales, including Tsukamurella and Gordonia (18).However, the phenotypic characteristics could remind us when the morphology did not match the identification result in the clinical laboratory or a simple distinction was made for different genera by rapid staining.In terms of phenotypic characteristics, for example, Tsukamurella and Gordonia isolates grew well on Columbia blood agar plates at 37°C for 48 h.This rapid growth characteristic was completely different from that of M. tuberculosis and nontuberculous mycobacteria (60).Overall, compared with the colony appearance of Gordonia species, Tsukamurella species were generally bigger, drier, wrinkled, and resembled a rough membrane with irregular spreading fringes.Tsukamur ella and Gordonia species are both Gram-positive bacilli and are difficult to distinguish by Gram staining.However, it appears that Tsukamurella isolates are slender rod-shaped bacilli, whereas Gordonia isolates are short and thick cocci-rod bacilli.Furthermore, positive modified Ziehl-Neelsen staining and partially positive Ziehl-Neelsen staining were clearly observed in Tsukamurella isolates but not in any of the Gordonia isolates.Therefore, tuberculosis cannot be diagnosed by the detection of acid-fast bacilli alone, a lesson to be learned from the misdiagnosis of case 10 in our study and the previously described case (19).
The optimal management of Tsukamurella infections remains to be determined because of the paucity of antimicrobial susceptibility for this genus.This study revealed the in vitro antimicrobial susceptibility profiles of 34 antibiotics for Tsukamurella and Gordonia isolates using the microbroth dilution method.Previous studies have shown that Tsukamurella isolates are resistant to penicillin, cefoxitin, and cephalosporin, which are commonly prescribed for the treatment of nontuberculous mycobacterial infections (5,38,46).However, Tsukamurella isolates in this study were susceptible to the thirdand fourth-generation cephalosporins.The differentiate antimicrobial susceptibility to amoxicillin/clavulanic acid (2:1) provided an additional method to discriminate between Tsukamurella and Gordonia species in terms of drug susceptibility.Previous studies summarized that Tsukamurella isolates were susceptible to aminoglycosides, macrolides, and vancomycin (46,56), but these antibiotics were not included in our recommen ded list without susceptibility testing, except for amikacin.Regarding anti-tuberculosis agents, rifampin had probable efficacy, but isoniazid, ethionamide, and ethambutol were completely ineffective in terms of in vitro susceptibilities against Tsukamurella and Gordonia isolates.Based on our results, quinolones, trimethoprim/sulfamethoxazole, amikacin, minocycline, linezolid, and tigecycline showed higher in vitro susceptibilities against Tsukamurella and Gordonia isolates.Although one T. tyrosinosolvens isolate in this study was not susceptible to quinolones, we still recommended quinolones as first-line or empirical therapy in Tsukamurella infection because it was commonly confused with tuberculosis (13,19).However, a favorable prognosis would be achieved by individualized treatment based on identification of the cause of infection and antimi crobial susceptibilities.Sulfonamides were good choice for the treatment Tsukamurella pulmonary infections with co-isolation of P. jirovecii (61), but not with that of P. aeru ginosa which is intrinsically resistant to sulfonamides (51,62).Our data may help to establish exclusive breakpoints for Tsukamurella species, even for the order Actinomyce tales.Notably, we proposed that the S91R substitution within the QRDR of the gyrA gene could confer high resistance to levofloxacin and ciprofloxacin and intermediate to moxifloxacin.Chauffour et al. concluded that mutations in residues at positions 87, 89, 91, and 92 in the QRDR of the gyrA gene confer resistance by impacting the binding of the quinolones to mycobacterial cells (52).
AST using the disk diffusion method is easy, universal, and convenient for most clinical laboratories.However, this study found disk diffusion to be an inappropriate method for testing ciprofloxacin, imipenem, and minocycline, based on the abnormal results of undetected non-susceptible isolates.For the other antimicrobials in this study, disk diffusion testing could be used to test the susceptibilities of Tsukamurella and Gordonia isolates after the clinical standard breakpoint has been established.
This study was undertaken because of the confusion between Tsukamurella and other mycolic acid-containing genera of the order Actinomycetales and the need for accurate identification methods to avoid misdiagnosis in clinical settings.To the best of our knowledge, we have investigated the susceptibilities of four species of Tsukamurella to the widest range of antimicrobial agents, as well as neatly compared the susceptibility discrimination between Tsukamurella and Gordonia.For the first time, we described T. inchonensis and T. sputi as potential pulmonary pathogens and elucidated a quinolone resistance mechanism in T. tyrosinosolvens.However, there are several limitations in this study.Firstly, the pathogenicity of pulmonary infections caused by Tsukamurella species is not completely certain, as they were all isolated from sputum samples.In addition, although the point mutation of the gyrA gene was found to confer quinolone resistance in T. tyrosinosolvens, this needs to be confirmed by further studies with phenotypic and genotypic experiments.Finally, the identification efficiency of the MALDI-TOF MS system for this rare pathogen should be further evaluated by establishing an in-house database.
In conclusion, this study highlighted the emergence of Tsukamurella species as a potential pathogen mainly for pulmonary infections in patients with various risk factors.The molecular methods accurately identified this pathogen to avoid misdiagnosis as pulmonary tuberculosis.We confirmed that groEL and secA genes are reliable for identifying Tsukamurella species.Additionally, we proposed the ssrA gene as a promising tool with the highest intraspecies discrimination but the shortest sequence length.This study remarkably supported quinolones, trimethoprim/sulfamethoxazole, amikacin, minocycline, linezolid, and tigecycline with potent in vitro activity against Tsukamur ella isolates.This study was novel in elucidating the mechanism conferring quinolone resistance in T. tyrosinosolvens.

12 TABLE 4
of NS MIC range N (%) of NS MIC range N (%) of NS MIC range N (%) of NS MIC range N (%) of NS inhibitory concentration; NA, not available; NS, non-susceptible.Results of antimicrobial susceptibility testing using broth microdilution methodology among Gordonia isolates according to species a inhibitory concentration; NA, not available; NS, non-susceptible.

TABLE 1
Clinical characteristics of 10 patients with probable pulmonary infection caused by Tsukamurella species a November/December 2023 Volume 11 Issue 6 10.1128/spectrum.01644-236

TABLE 2
Clinical characteristics of one patient with CR-BSI caused by T. tyrosinosolvens a a CRP, C-reactive protein; MALDI-TOF MS, matrix-assisted laser desorption ionization-time-of-flight mass spectrometry; PICC, peripherally inserted central catheter.

TABLE 3
Results of antimicrobial susceptibility testing using broth microdilution methodology among Tsukamurella isolates according to species a