Focusing on Gordonia Infections: Distribution, Antimicrobial Susceptibilities and Phylogeny

The immunosuppression conditions and the presence of medical devices in patients favor the Gordonia infections. However, the features of this aerobic actinomycete have been little explored. Strains (n = 164) were characterized with 16S rDNA and secA1 genes to define their phylogenetic relationships, and subjected to broth microdilution to profile the antimicrobial susceptibilities of Gordonia species that caused infections in Spain during the 2005–2021 period. Four out of the eleven identified species were responsible for 86.0% of the infections: Gordonia sputi (53.0%), Gordonia bronchialis (18.3%), Gordonia terrae (8.5%) and Gordonia otitidis (6.1%). Respiratory tract infections (61.6%) and bacteremia (21.9%) were the most common infections. The secA1 gene resolved the inconclusive identification, and two major clonal lineages were observed for G. sputi and G. bronchialis. Species showed a wide antimicrobial susceptibility profile. Cefoxitin resistance varies depending on the species, reaching 94.2% for G. sputi and 36.0% for G. terrae. What is noteworthy is the minocycline resistance in G. sputi (11.5%), the clarithromycin resistance in G. bronchialis secA1 lineage II (30.0%) and the amoxicillin–clavulanate and cefepime resistance in G. terrae (21.4% and 42.8%, respectively). G. sputi and G. bronchialis stand out as the prevalent species causing infections in Spain. Resistance against cefoxitin and other antimicrobials should be considered.

Unlike Nocardia spp., the most common aerobic actinomycete that received an increasing interest in the last decade [19], there is very limited information about the Gordonia species involved in human infections.They are rare, predominating the descriptions of catheter-related bloodstream infections, sternal wound infections and soft tissue and bone Antibiotics 2023, 12, 1568 2 of 18 infections [9,11,15,20,21].Patients with immunocompromised status due to congenital or acquired immunodeficiencies (AIDS infection, hypogammaglobulinemia, leukemia, solid organs cancer, transplant recipients, Hodgkin's disease, etc.), or other underlying diseases (chronic hepatitis B virus infection, diabetes mellitus, chronic obstructive pulmonary disease, vasopressor medication, obesity), are susceptible to suffer Gordonia infections [5,6,[11][12][13].Also, immunocompetent patients are susceptible due to the presence of foreign bodies including central lines and shunts, being the main risk factor of this acquisition [6,17,20,22].
As for other organisms, the accurate species identification is crucial to understand the epidemiology, clinical features, treatment and outcome of Gordonia infections.So far, the most reliable is the sequencing of 16S rDNA [6], supported with secA1 and gyrB genes [23,24].However, secA1 was preferred because of difficulties on amplification and sequencing of the gyrB gene.Gordonia phylogeny, based in both genes, has been mainly explored for type or reference strains [23].
The features of this aerobic actinomycete have been little explored, and because of its rarity, the infections produced with Gordonia species are not well understood.The aims of this work were to study the species distribution of the Gordonia infection in Spanish patients, and their respective antimicrobial susceptibilities.To gain insight in the knowledge of the diversity of Gordonia species involved in clinical infections, phylogenies based on 16S rDNA and secA1 genes were also studied.
In our study, G. sputi (n = 2) and G. bronchialis (n = 2) were responsible for urinary tract infections (UTIs), not described previously.G. terrae has been described as an emerging pathogen [6], and the cause of bacteremia [6,15], cholecystitis [33], granulomatous mastitis [34], palpebral abscess [35] and peritonitis [30].Here, G. terrae caused respiratory tract and soft tissue infections (it was isolated from the respiratory samples of ten patients with tuberculosis suspicion and from one wound sample).G. otitidis has been reported as responsible for external otitis [36] and bacteriemia [15], and here it was found as an agent of respiratory infections and UTI (seven sputum and one urine sample).For the remaining Gordonia species, not-yet-described infections were observed in our study, as occurring for (i) G. aichiensis previously collected in blood as the cause of bacteremia [37] and here it was isolated as being responsible for respiratory tract infections in four patients; (ii) G. araii has been responsible for post-surgical injury and skin infection [22,38] and here it also appeared in respiratory infections; (iii) G. hongkongensis has been collected from blood and peritoneal dialysis effluent [16], and it was detected in two sputum and one abscess samples; (iv) G. iterans has been obtained from sputum as the cause of pneumonia [39], and here in bacteremia and soft tissue infection; and last, (v) G. polysoprenivorans has been described as the cause of bacteremia [5,12,15] and endocarditis [10], and here it was collected from two sputum and one abscess samples as responsible for respiratory and soft tissue infections.To our knowledge, there have not been reported descriptions of infections caused by G. alkanivorans and G. jinhuaensis.In our study, these species were isolated from sample of patients with soft tissue and respiratory infections, respectively.
The ability of some Gordonia species to attach to and colonize the surface of medical devices (central lines and shunts) with biofilm formation produces a significant number of catheter-related bloodstream or intravascular device infections [22].This occurs for G. polyisoprenivorans that produces biosurfactants that help to form biofilm onto the catheter, decreasing antibiotic penetration [10].This ability, together with the possible presence of a urinary catheter, could explain the UTIs caused by G. sputi, G. bronchialis and G. otitidis.However, this acquisition route was not confirmed or ruled out in our study.In this work, other species (G.terrae, G. otitidis, G. aichiensis, G. hongkongensis, G. iterans, G. polyisoprenivorans and G. jinhuaensis) can be the probable etiological agent of respiratory infections, due to the clinical information provided at the time of their submission.However, some of these isolations could be colonization more than true infections such as that occurring for Nocardia in patients with chronic pulmonary disease [19].

Antimicrobial Susceptibilities of Gordonia Species
In previous works, the data about the antimicrobial susceptibility profiles of Gordonia species were reduced to the results of the strain that produce the clinical case, or a low number of strains, or just a review of the literature.Fortunately, Gordonia spp.are usually described as antibiotic-sensitive [6,20,23,30,41] as it is shown in Table 2 for the 164 strains of the 11 species identified as Gordonia spp.Their wide susceptibility spectrum includes β-lactams and carbapenems; amikacin and tobramycin; clarithromycin; doxycycline and minocycline; ciprofloxacin and moxifloxacin; trimethoprim and sulphametoxazole; and linezolid.
However, the results showed some details: (i) the lack of activity of cefoxitin against Gordonia strains, with overall resistance rates of 76.6%, and this cefoxitin resistance affects all species studied, and it ranged from 35.7% for G. terrae to 94.2% for G. sputi, and (ii) some species harbored a differential resistance profile, as occurring for G. sputi against minocycline (11.5%),G. bronchialis respective to clarithromycin (30.0%) and G. terrae against amoxicillin-clavulanate and cefepime (21.4% and 42.8%, respectively).Regarding tigecycline, when the corresponding EUCAST susceptibility breakpoint criteria of ≤0.5 mg/L for Staphylococcus spp.and Enterobacterales organisms was used, a high resistance rate was found for G. bronchialis (40%), and for G. sputi and G. terrae (≈30% for each one) [42].
Gordonia spp.degrade xenobiotics, environmental pollutants and biodegradable natural polymers, and transform or synthesize possibly useful compounds for environmental and industrial biotechnology [43,44].These abilities could orientate toward thinking that Gordonia should be intrinsically an antimicrobial multiresistant bacteria.A further note is that susceptible profiles of the different Gordonia species were mainly observed in our study.This feature differs greatly from those described for the Nocardia species [45,46].Both organisms proceed from environmental sites (mainly soil) and they are also subjected to the effect of bioactive compounds as antimicrobial products synthesized using actinobacteria in this niche including Streptomyces spp., among others [47].In general, Nocardia strains show an expanded resistance profile, which affects different classes of antibiotics, and with specific drug pattern types according to the Nocardia species [48].At this moment, these events do not occur for Gordonia spp.

Moxifloxacin
2.4.Phylogenetic Analysis of Gordonia spp.2.4.1.16S rDNA and secA1 phylogeny Two major clusters were observed in the phylogenetic tree based on 16S rDNA sequences (adjusted to 943-pb, Figure S1): G. sputi, G. aichiensis, G. otitidis, G. jinhuaensis and G. polyisoprenivorans grouped together in one cluster, and G. alkanivorans and G. bronchialis in a second cluster.Meanwhile, G. araii and G. iterans, and G. terrae and G. hongkongensis, were placed in minor third and fourth clusters.The bootstrap values were low (≤50), except for the G. terrae and G. hongkongensis cluster (96).In contrast, the values for branches that grouped strains belonging to the same species were high (average of 90), except for the branches of G. aichiensis and three G. otitidis strains (63 and 47, respectively).In another study [23], similar clusters were observed for 23 Gordonia species, except for G. alkanivorans (it was grouped together with G. terrae).
Through the secA1 gene, some 16S rDNA clustering remained similar for G. sputi, G. aichiensis and G. otitidis, but not for G. polyisoprenivorans that grouped together with G. bronchialis, and for G. alkanivorans that grouped together with G. hongkongensis (Figure 1).This different position of G. polyisoprenivorans was also previously seen [23].Three independent branches were seen for G. jinhuaensis, G. araii and G. iterans.G. terrae and G. hongkongensis gathered in the same branch with the 16S rDNA gene (with no differences in a studied fragment of 1.166 bp), but they were separated with the secA1 gene, whereas G. terrae appeared in a unique cluster, and G. hongkongensis did with G. alkanivorans.The strain CNM20140419 Gordonia spp. was placed between the two branches for the 16S rDNA gene of G. polyisoprenivorans and G. jinhuaensis.Meanwhile, in the tree of the secA1 gene, this strain placed in a cluster was together with G. alkanivorans and G. hongkongensis.
The partial secA1 gene provided significant base diversity among all strains (HGDI = 0.954), and also for G. sputi and G. bronchialis groups (0.89 and 0.75, respectively), but it was very low for G. terrae and G. otitidis groups (Table 3).For the G. sputi group, the identity or sequence similarity respective to the type strain ranged from 100.0% to 95.6%; for G. bronchialis, from 100.0% to 88.7%; for G. terrae, from 100.0% to 99.6%; and for G. otitidis, from 100.0% to 96.0%.Low values of identity for secA1 showed the great diversity of this gene for a single species.This could be due to the presence of different clonal lineages of intra-species that are onwards seen.In a previous work, Kang et al. reported a range of secA1 identity values from 82 to 98% for different species [23].strains with those of reference strains (n = 175 strains), and Williamsia muralis reference strain MA140/96 as outgroup using MEGA7 [49].The evolutionary history was inferred by using the Maximum Likelihood method based on the Tamura 3-parameter model with 1000 replications for bootstrap values.The optimal tree is shown.The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree.A discrete Gamma distribution was used to model evolutionary rate differences among sites (5 categories (+G, parameter = 0.5004)).secA1 sequence-based (337 bp alignment positions) phylogenetic tree of clinical Gordonia strains with those of reference strains (n = 175 strains), and Williamsia muralis reference strain MA140/96 as outgroup using MEGA7 [49].The evolutionary history was inferred by using the Maximum Likelihood method based on the Tamura 3-parameter model with 1000 replications for bootstrap values.The optimal tree is shown.The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree.A discrete Gamma distribution was used to model evolutionary rate differences among sites (5 categories (+G, parameter = 0.5004)).a Analyzed size in number of base pars; b HGDI, S2 and SD correspond to the Hunter and Gaston discrimination index, the variance and standard deviation, respectively; c the nucleotide diversity (defined as the average number of nucleotide differences per site between two sequences) is expressed as a percentage among strains of each group (Nei and Kumar, 2000); d respective to reference strain of each species; e this analysis includes 164 clinical strains and 10 reference strains; f 6 SNPs were identified respective to G. sputi DMS 43,896 strain (secA1-type) and respective to the prevalent secA-type (A8-type) for clonal lineages I and II, respectively; g data of clonal lineage II respective to G. sputi DMS 43,896 strain were 16-20 (18.4, 19); h SNPs were identified respective to G. bronchialis DMS 43247 strain (A1-type) and respective to prevalent secA-type (A4-type) for clonal lineage I and II, respectively.
for clonal lineage I respective to G. sputi DSM 43896T (≤98.4% for ≤7 SNPs), while lower identity remained for lineage II respective to secA1-type A8 (≤95.6% for ≤20 SNPs, respectively).Therefore, lineage II was more diverse than lineage I. Until 2014, the involvement of clonal lineage II in the human infections of the Spanish patients was double than lineage I, but since 2015, the clinical implications of both lineages were similar.These lineages were not seen so clear in the secA1 gene phylogenies for G. sputi strains using G. bronchialis DSM 43247 as the outgroup (Figure S2).[49].The evolutionary history was inferred by using the Maximum Likelihood method based on the Tamura 3-parameter model with 1000 replications for bootstrap values.The optimal tree is shown.The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree.A discrete Gamma distribution was used to model evolutionary rate differences among sites (5 categories (+G, parameter = 0.0500)).
The nucleotide identity for all G. sputi strains (≤95.6% for ≤20 SNPs/454 bp) increased for clonal lineage I respective to G. sputi DSM 43896T (≤98.4% for ≤7 SNPs), while lower identity remained for lineage II respective to secA1-type A8 (≤95.6% for ≤20 SNPs, respectively).Therefore, lineage II was more diverse than lineage I. Until 2014, the involvement of clonal lineage II in the human infections of the Spanish patients was double than lineage I, but since 2015, the clinical implications of both lineages were similar.These lineages were not seen so clear in the secA1 gene phylogenies for G. sputi strains using G. bronchialis DSM 43247 as the outgroup (Figure S2).

Gordonia Species with Medium Prevalence: G. bronchialis
The 16S-haplotype of the reference strain was found in all G. bronchialis strains (n = 30).This is unlike the secA1 partial sequences, which showed 15 segregating sites respective to the G. bronchialis DSM 43247 strain, resulting in nine secA1-haplotypes (Table 3).Nine of these SNPs were very prevalent (≥46%) (positions 480, 507, 540, 543, 651, 672, 690, 738 and 801 respective to the complete secA1 gene of G. sputi ATCC 29,627 GenBank accession no.JAAXP0010000001).Three amino acid replacements were produced, Asp-180 → Glu, Glu-231 → Gln and Arg-258 → Gln, in 14, 21 and 2 strains, respectively.Three clonal lineages could be observed in the phylogenetic tree (Figure 3), created with the same suggested model mentioned above: clonal lineage I, represented by the strain G. bronchialis DSM 43247T with secA1-type A1 (eight strains), and types A2-A3 (one strain for each one); clonal lineage II, with the predominant secA1-type A4 (thirteen strains), and types A5-A8 (one strain for each one); and clonal lineage III, represented by type A9 (two strains).In the studied period, clonal lineage II produced double the clinical cases than other lineages.It should be highlighted that seven out of nine clarithromycin-resistant strains belonged to clonal lineage II (n = 18).30).This is unlike the secA1 partial sequences, which showed 15 segregating sites respective to the G. bronchialis DSM 43247 strain, resulting in nine secA1-haplotypes (Table 3).Nine of these SNPs were very prevalent (≥46%) (positions 480, 507, 540, 543, 651, 672, 690, 738 and 801 respective to the complete secA1 gene of G. sputi ATCC 29,627 GenBank accession no.JAAXP0010000001).Three amino acid replacements were produced, Asp-180 → Glu, Glu-231 → Gln and Arg-258 → Gln, in 14, 21 and 2 strains, respectively.Three clonal lineages could be observed in the phylogenetic tree (Figure 3), created with the same suggested model mentioned above: clonal lineage I, represented by the strain G. bronchialis DSM 43247T with secA1-type A1 (eight strains), and types A2-A3 (one strain for each one); clonal lineage II, with the predominant secA1-type A4 (thirteen strains), and types A5-A8 (one strain for each one); and clonal lineage III, represented by type A9 (two strains).In the studied period, clonal lineage II produced double the clinical cases than other lineages.It should be highlighted that seven out of nine clarithromycin-resistant strains belonged to clonal lineage II (n = 18).
Two outbreaks of sternal wound infections with G. bronchialis, where nurse scrub was identified as the vector for contamination of the surgical wounds, were described [50,51].To investigate the clonal relation among strains collected in four patients hospitalized in the same ward during 6 months with previous cardiac surgery, secA1 typing was applied.Three secA1-types were detected (A1, A7 and A9).To distinguish two strains with secA1type A1, the gyrB gene sequencing was carried out, resulting in two different alleles, discarding the outbreak scenario.[49].The evolutionary history was inferred by using the Maximum Likelihood method based on the Tamura 3-parameter model with 1000 replications for bootstrap values.The optimal tree is shown.The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree.A discrete Gamma distribution was used to model evolutionary rate differences among sites (5 categories (+G, parameter = 0.0500)).
Two outbreaks of sternal wound infections with G. bronchialis, where nurse scrub was identified as the vector for contamination of the surgical wounds, were described [50,51].To investigate the clonal relation among strains collected in four patients hospitalized in the same ward during 6 months with previous cardiac surgery, secA1 typing was applied.Three secA1-types were detected (A1, A7 and A9).To distinguish two strains with secA1-type A1, the gyrB gene sequencing was carried out, resulting in two different alleles, discarding the outbreak scenario.

Gordonia Species with Low Prevalence: G. terrae and G. otitidis
In the case of G. terrae, a common haplotype of 16S rDNA was shared by the reference DSM 43249T strain and the studied strains.Three secA1-haplotypes were seen: one silent SNP in 12 strains (identity of 99.8%); one silent SNP and the mutation Ile-177 → Val in one strain (identity of 99.6%); and the secA1-type of the type strain, in one strain.Regarding G. otitidis, seven strains showed the same 16S haplotype compared to the reference strain, and three strains showed one change (position 433).A unique secA1-haplotype was observed for all strains, except one strain that showed 18 silent SNPs respective to the reference strain (identity of 96%).These strains with changes in 16S or secA1 genes did not show any difference in the clinical or susceptibility features.

Other Gordonia species
Regarding the other seven detected species, differences respective to the 16S rDNA haplotype of the corresponding reference strain were found for G. alkanivorans (position 704, one strain), G. iterans (position 403, three strains) and G. jinhuaensis (position 1357, one strain).Meanwhile, identical 16S haplotypes of the reference strain were found for strains belonging to G. aichiensis (eight strains), G. araii (two strains) and G. polyisoprenivorans (four strains) species.It should be noted that 16S rDNA sequences of three G. hongkongensis strains showed a full identity respective to the reference G. hongkongensis strain HKU50, but also respective to G. terrae NRRL B-16283 (GenBank accession nos.NR_152022.1 and CP029604.1,respectively).Differences in a partial sequence of the secA1 gene were detected for two G. araii strains (identities of 99.1-98.1%,4-8 silent SNPs/432 bp), for three strains of G. iterans (99.8%, 1 SNP/432 bp, Met-94 → Ileu), for one strain of G. jinhuaensis (99.8%, 1 silent SNP/506 bp) and for three out of four G. polyisoprenivorans strains (99.7-98.6%,1-6 silent SNPs/443 bp), respective to the corresponding reference strains.The same secA1-haplotype compared to the corresponding reference strain was found for all strains of G. aichiensis, G. alkanivorans and G. hongkongensis.In addition, when the secA1 of G. hongkongensis strain HKU50 was compared to G. terrae DSM 43249T, the identity was 94.31% (24 SNPs/422 bp), allowing the correct species assignment.
In the adaptation and survival of bacteria, including manipulating host cells and competing for resources with other microorganisms, the protein secretion plays many roles in bacteria [52].This happens in the cell envelope biogenesis and maintenance, and in the delivery of adherence and pathogenic effector proteins to the cell surface [53].The Sec translocase is responsible for the translocation of unfolded proteins across membranes.The two components of the Sec translocase are the cytoplasmic motor protein, SecA1, and the membrane-embedded channel, SecYEG.During protein secretion, SecYEG engages with the cytosolic motor ATPase SecA1, and together they pass pre-proteins with a short N-terminal cleavable signal sequence across the membrane, whilst still having an unfolded conformation [53,54].The vital role and variability of SecA1 in bacteria becomes connected to the secA1 gene as a strong candidate to perform identification and phylogeny.So, it had been previously described for Gordonia species [23], as for other related genera such as a Mycobacterium and Nocardia [55,56].To obtain an accurate identification of Gordonia strains, different approaches have been used, in special genome-based approaches [57].The analysis of the secA1 gene can provide a quick and simple panoramic view of the diversity of the populations of Gordonia species (as it has been seen for G. sputi and G. bronchialis) with clinical and resistance implications.

Strains and Target Genes for Identification
Strains were submitted from different Spanish hospitals to the Reference and Research Laboratory for Taxonomy of National Center for Microbiology (CNM, Majadahonda, Madrid, Spain) for species identification.The strains were isolated from patients with signs and symptoms of bacterial infection in 31 provinces of 13 Spanish Autonomous Communities between 2005 and 2021 (17 years).Only one strain per patient was considered in this study.Strains were grown on Columbia agar supplemented with 5% (v/v) sheep blood and buffered charcoal-yeast extract agar (BCYE) for 48-72 h at 37 • C under aerobic conditions.DNA was extracted with the boiling method and amplifications were performed using Ready-To-Go PCR Beads (Amersham Biosciences, Buchinghamshire, UK) with the following targets to identification: (i) a full 16S rDNA gene sequence analysis, with primers fD1 and rP2 for amplification [58], and E781 and U1115 for sequencing [59]; (ii) the secA1 gene with primers SecA1-F and SecA1-R [23]; and (iii) the gyrB gene with primers UP-1 and UP-2 for no identified strains [23] (Table S1).Specific primers for the gyrB sequence of G. bronchialis 123F and 1248Rev were designed to study the possible relation of G. bronchialis strains.The amplification products were electrophoresed and purified using an ExoSAP-IT reagent (GE Healthcare, Livingston, NJ, USA) and sequenced with capillary electrophoresis in an ABI Prism 3100 apparatus (Applied Biosystems, Foster City, CA, USA).

16S rDNA and secA1 Phylogeny
Sequences were assembled using SEQ-Man software (DNASTAR, Inc., Madison, WI, USA) and adjusted for a phylogenetic analysis to coincide with the length of the shortest sequence using BioEdit software [63].The Hunter-Gaston discrimination index (HGDI) [64] and summary statistics for the analysis of multi-locus population genetics were assessed using DNA Sequence Polymorphism (DnaSP) software [65].
A phylogenetic assessment of all collected populations of Gordonia strains was undertaken with the 16S rDNA gene and secA1 genes.For the species with high and medium prevalence, dendograms based on the secA1 gene were created.Phylogenetic trees were constructed using maximum-likelihood methods [66], with bootstrap analyses based on 1000 resamplings.Branches corresponding to partitions that were reproduced in <50% of bootstrap replicates were collapsed.The evolutionary distance between the nucleotide and amino acid sequences of the secA1 gene was determined using the Tamura 3-parameter mode with gamma distribution (five categories and by assuming that a certain fraction of sites is evolutionary invariable), as it is suggested as the best DNA model with MEGA7 [49].Williamsia muralis-type strain MA140/96 (GenBank accession no.NR_037083.1)was included as an outgroup for the phylogeny of the 16S rDNA and secA1 genes.
Sixty-five strains that belonged to different species were analyzed with a matrixassisted laser desorption ionization time of flight MS (MALDI-TOF MS)-based system (Vitek MS, SARAMIS premium software, bioMerieux, Madrid, Spain), following the procedure recommended by the manufacturer.Briefly, target slides were inoculated into the spots by picking a freshly grown overnight colony and overlaid with a matrix solution (1 µL of a-cyano-4-hydroxycinnamic acid) using a complete protocol of protein extraction with formic acid and acetonitrile.

Conclusions
The two main species responsible for the Gordonia infections in Spain are G. sputi and G. bronchialis (71.3%).Species assignation was correctly performed in 87.2% of the clinical Gordonia strains with the 16S rDNA gene.The secA1 gene resolves the inconclusive identification, giving a finer species distinction among closely related species (G.terrae and G. hongkongensis), discriminates among clinical strains and defines phylogenetic relationships at the inter-and intra-species level.In this way, the secA1 gene provides a panoramic view of their diversity with clinical and resistance implications.To our knowledge, this is the first study that explores susceptibilities and the phylogenies of a wide population of Gordonia clinical strains to better understand the increasing clinical importance of the infections caused by this genus.

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/antibiotics12111568/s1, Figure S1.16S rDNA sequence-based (943 bp alignment positions) phylogenetic tree of clinical Gordonia strains with those of reference strains and Williamsia muralis reference strain MA140/96 as outgroup (n = 175 strains) using MEGA7 [49].The evolutionary history was inferred by using the Maximum Likelihood method based on the Tamura 3-parameter model with 1000 replications for bootstrap values.The optimal tree is shown.The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree.A discrete Gamma distribution was used to model evolutionary rate differences among sites (5 categories (+G, parameter = 0.1000)); Figure S2.secA1 sequence-based phylogenetic tree of clinical G. sputi strains with G. sputi DSM 43896, and G. bronchialis DSM 43247 as outgroup in Figure S2a (n = 89 strains, 453 bp alignment positions).secA1 sequence-based phylogenetic tree of clinical G. bronchialis strains with G. bronchialis DSM 43247, and G. sputi DSM 43896 as outgroup in Figure S2b (n = 32 strains, 335 bp alignment positions), using MEGA7 [49].The evolutionary history was inferred by using the Maximum Likelihood method based on the Tamura 3-parameter mode with 1000 replications for bootstrap values.The optimal tree is shown.The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree.A discrete Gamma distribution was used to model evolutionary rate differences among sites (5 categories (+G, parameter = 2.8492 and 0.1977)); Table S1.Primers and conditions of amplification used for Gordonia species.

Figure 1 .
Figure1.secA1 sequence-based (337 bp alignment positions) phylogenetic tree of clinical Gordonia strains with those of reference strains (n = 175 strains), and Williamsia muralis reference strain MA140/96 as outgroup using MEGA7[49].The evolutionary history was inferred by using the Maximum Likelihood method based on the Tamura 3-parameter model with 1000 replications for bootstrap values.The optimal tree is shown.The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree.A discrete Gamma distribution was used to model evolutionary rate differences among sites (5 categories (+G, parameter = 0.5004)).

Figure 1 .
Figure1.secA1 sequence-based (337 bp alignment positions) phylogenetic tree of clinical Gordonia strains with those of reference strains (n = 175 strains), and Williamsia muralis reference strain MA140/96 as outgroup using MEGA7[49].The evolutionary history was inferred by using the Maximum Likelihood method based on the Tamura 3-parameter model with 1000 replications for bootstrap values.The optimal tree is shown.The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree.A discrete Gamma distribution was used to model evolutionary rate differences among sites (5 categories (+G, parameter = 0.5004)).

Figure 2 .
Figure 2. secA1 sequence-based phylogenetic tree of clinical G. sputi strains with reference G. sputi DSM 43,896 (n = 88 strains, 453 bp alignment positions) using MEGA7[49].The evolutionary history was inferred by using the Maximum Likelihood method based on the Tamura 3-parameter model with 1000 replications for bootstrap values.The optimal tree is shown.The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree.A discrete Gamma distribution was used to model evolutionary rate differences among sites (5 categories (+G, parameter = 0.0500)).

Figure 3 .
Figure 3. secA1 sequence-based phylogenetic tree of clinical G. bronchialis strains with reference G. bronchialis DSM 43247 (n = 31 strains, 335 bp alignment positions) using MEGA7[49].The evolutionary history was inferred by using the Maximum Likelihood method based on the Tamura 3-parameter model with 1000 replications for bootstrap values.The optimal tree is shown.The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree.A discrete Gamma distribution was used to model evolutionary rate differences among sites (5 categories (+G, parameter = 0.0500)).

Funding:
S.P.-R. is a student for a BSc of the Master's Degree in Microbiology Applied to Public Health and Research in Infectious Diseases (plan of studies M138) of the Universidad de Alcalá in collaboration with the Instituto de Salud Carlos III.N.G. is contracted via the MCIN/AEI grant PTA-2019-016623-I.M.V. is contracted via the grant PEJ CAM 2021-/TL/BMD-21100 from the Programa Operativo Empleo Juvenil e Iniciativa Empleo Juvenil (YEI).Institutional Review Board Statement: Not applicable.Informed Consent Statement: Not applicable.

Table 1 .
Distribution of the Gordonia species in clinical infections in Spain according to sample source in the 2005-2021 period.

Table 2 .
Antimicrobial susceptibilities of the Gordonia species responsible for clinical infections in Spain over the 17-year study period.

Table 3 .
Summary statistics for the analysis of 16S rRNA and secA1 population genetics of the high, medium and low prevalent species of Gordonia isolated from clinical samples in Spain between 2005 and 2021.