Streptococcus gallolyticus subsp. gallolyticus from Human and Animal Origins: Genetic Diversity, Antimicrobial Susceptibility, and Characterization of a Vancomycin-Resistant Calf Isolate Carrying a vanA-Tn1546-Like Element

Abstract

The aim of this work was to characterize the antibiotic susceptibility and genetic diversity of 41 Streptococcus gallolyticus subsp. gallolyticus isolates: 18 isolates obtained from animals and 23 human clinical isolates. Antibiotic susceptibility was determined by the semiautomatic Wider system and genetic diversity by pulsed-field gel electrophoresis (PFGE) with SmaI. Animal isolates grouped separately in the PFGE analysis, but no statistical differences in antimicrobial resistance were found between the two groups. The LMG 17956 sequence type 28 (ST28) strain recovered from the feces of a calf exhibited high levels of resistance to vancomycin and teicoplanin (MIC, ≥256 mg/liter). Its glycopeptide resistance mechanism was characterized by Southern blot hybridization and a primer-walking strategy, and finally its genome, determined by whole-genome sequencing, was compared with four closely related S. gallolyticus subsp. gallolyticus genomes. Hybridization experiments demonstrated that a Tn1546-like element was integrated into the bacterial chromosome. In agreement with this finding, whole-genome sequencing confirmed a partial deletion of the vanY-vanZ region and partial duplication of the vanH gene. The comparative genomic analyses revealed that the LMG 17956 ST28 strain had acquired an unusually high number of transposable elements and had experienced extensive chromosomal rearrangements, as well as gene gain and loss events. In conclusion, S. gallolyticus subsp. gallolyticus isolates from animals seem to belong to lineages separate from those infecting humans. In addition, we report a glycopeptide-resistant isolate from a calf carrying a Tn1546-like element integrated into its chromosome.

INTRODUCTION

New taxonomic criteria have recently been applied to the Streptococcus bovis/equinus complex, mainly on the basis of the genetic diversity of the sodA gene, which is considered the best target for adequate identification (1). These taxonomic advances have permitted the study of the epidemiological correlations between particular subspecies and specific human pathologies (2), including gastrointestinal colonization by Streptococcus gallolyticus subsp. gallolyticus (formerly Streptococcus bovis biotype I) and its coincidence with colorectal cancer (3).

The rates of colonization by S. gallolyticus subsp. gallolyticus are around 5 to 10% in humans but could be much higher in animals (4). Bacteremia and endocarditis are the main relevant clinical manifestations in humans and in avian species (5, 6). The mortality rate in a broiler flock outbreak was 4.3% (7), and the organism has also been implicated in other veterinary pathologies (8, 9).

Similarities and differences between S. gallolyticus subsp. gallolyticus isolates from humans and animals have been described previously (8, 10), including the existence of particular virulent clones with increased invasion and adherence abilities, which favor bloodstream infections (11–13). A recent multilocus sequence typing (MLST)-based study demonstrated a lack of specificity for any particular host or geographical location (10).

Antibiotic resistance is a not infrequent feature of S. gallolyticus subsp. gallolyticus isolates. Resistance to macrolides and tetracyclines and high-level resistance to aminoglycosides are most commonly reported (14–16). Resistance to glycopeptides has been reported occasionally as a consequence of the acquisition of enterococcal vanA or vanB mechanisms (17–19).

Transmission of S. gallolyticus subsp. gallolyticus strains between animals and humans has not been documented yet, although in view of the example of the related genus Enterococcus, zoonotic potential cannot be ruled out. A major prevalence of endocarditis caused by S. bovis in rural areas of France and Spain has been demonstrated in this sense (20, 21).

We have characterized an S. gallolyticus subsp. gallolyticus collection including isolates recovered from animals and humans and have compared their genetic diversity and antimicrobial susceptibility profiles. Furthermore, we have characterized a glycopeptide-resistant isolate carrying a vanA-Tn1546-like element that was recovered from the feces of a calf.

MATERIALS AND METHODS

Bacterial isolates.

A total of 41 S. gallolyticus subsp. gallolyticus isolates (18 from animals and 23 from humans) were included in the study (Tables 1 and 2; see Fig. 1). Some of the human and animal isolates were kindly provided from the public BCCM/LMG strain collection (http://bccm.belspo.be/about/lmg.php) (13), whereas the bacteremic isolates were obtained at the Ramón y Cajal University Hospital (1) in Madrid, Spain. All strains were identified to the subspecies level by PCR amplification of an internal fragment of the sodA gene and further nucleotide sequencing (1). Matrix-assisted laser desorption ionization–time of flight (MALDI-TOF) mass spectrometry (MS) using the Bruker Biotyper system (Bruker Daltonics, Bremen, Germany) was also performed as part of the bacterial identification scheme (22).

TABLE 1.

Main characteristics of the 18 animal isolates

Isolate	Origin	Yr	Country	STa	Antibiotic resistance
					Phenotypeb	Genotype
LMG 14621	Dove abscess	1994	Belgium	19	Fo, Lv
LMG 14622	Dove	1994	Belgium	19	Fo
LMG 14623	Dove liver	1994	Belgium	20	Fo, Er, Cd, Mn, St	erm(B), tet(M)
LMG 14634	Cow	1994	Belgium	7	Fo, Er, Cd, Mn, St	erm(B), tet(M)
LMG 14821	Dove gut	1994	Belgium	21	Fo, Cd, Mn, St	tet(M)
LMG 14823	Dove gut	1994	Belgium	22	Fo, Er, Cd, Mn, St	erm(B), tet(M)
LMG 14855	Horse gut	1994	Belgium	23
LMG 14856	Horse gut	1994	Belgium	24
LMG 14870	Cow gut	1994	Belgium	25	Er, Cd, St	erm(B)
LMG 14876	Cow tonsil	1994	Belgium	26	Fo, Er, Cd, Mn, St	erm(B), tet(M)
LMG 14878	Pig lung	1994	Belgium	19	Fo, Er
LMG 15572	Goat rumen	1994	Australia	27
LMG 15573	Goat rumen	1994	Australia	27
LMG 16005	Cow gut	1995	Belgium	3	Er, Cd, Mn, SxT, St	erm(B), tet(M)
LMG 17956	Cow gut	1997	Netherlands	28	Cd, Q/D, SxT, Va	Tn1546-like
LMG 22782	Dog	2000	Belgium	17	Er, Cd, Mn, Q/D	erm(B), tet(M)
05WDK43740 002	Cow gut	Unknown	Unknown	13	Fo
DSM16831	Koala feces	1990	Australia	1	Fo, Gm	aac(6′)-aph(2″)

^aST, sequence type.

^bFo, fosfomycin; Lv, levofloxacin; Er, erythromycin; Cd, clindamycin; Mn, minocycline; St, streptomycin at 1,000 mg/liter; SxT, trimethoprim-sulfamethoxazole; Q/D, quinupristin-dalfopristin; Va, vancomycin; Gm, gentamicin at 500 mg/liter.

TABLE 2.

Main characteristics of the 23 human-invasive isolates

Isolate	Origin	Yr	Country	STa	Antibiotic resistance
					Phenotypeb	Genotype
003080/00	Human gut	2000	Germany	5	Fo, Gm	aac(6′)-aph(2″)
005950/03	Human heart valve	2003	Germany	11	Fo, Er, Cd, Mn, St, Q/D	erm(B), tet(M)
006718/00		2000	Germany	7	Mn	tet(M)
007849/02		2002	Germany	8	Fo, Er, Cd, Lv, St, Q/D	erm(B)
010288/01		2001	Germany	3	Er, Cd, Mn, St	erm(B), tet(M)
010672/01		2001	Germany	6	Mn, Gm	tet(M), aac(6′)-aph(2″)
021702/06		2006	Germany	9	Fo, Mn, Gm	tet(M), aac(6′)-aph(2″)
12932/01		2001	Germany	3	Fo, Er, Cd, Mn	erm(B), tet(M)
B1	Bacteremia	2004	Spain	34	SxT
B6		2009	Spain	35	Fo, Cd, St
B11		2005	Spain	7	Fo, Er, Cd, St, Lv, Q/D	erm(B)
B13		2004	Spain	36	Fo, Er, Cd, St, Lv	erm(B)
B14		2004	Spain	5
B19		2006	Spain	7	Fo, Er, Cd, Lv, St, Q/D	erm(B)
B22		2006	Spain	37
B28		2009	Spain	38	Fo, Mn, Gm	erm(B), aac(6′)-aph(2″)
B29		2009	Spain	39	Er, Cd, Mn	erm(B), tet(M)
B35		2003	Spain	26	Er, Cd, St
B41		2010	Spain	5	Er	erm(B)
B49		2003	Spain	6
B51		2003	Spain	40	Fo
B52		2003	Spain	5	Er, Cd	erm(B)
K6236/35_MS		Unknown	Unknown	12

^aST, sequence type.

^bFo, fosfomycin; Gm, gentamicin at 500 mg/liter; Er, erythromycin; Cd, clindamycin; Mn, minocycline; St, streptomycin at 1,000 mg/liter; Q/D, quinupristin-dalfopristin; Lv, levofloxacin; SxT, trimethoprim sulfamethoxazole.

FIG 1.

Dendrogram based on the Dice coefficient of the PFGE patterns. The source and sequence type (ST) of each strain are also given.

Antimicrobial susceptibility testing.

Susceptibility was determined using the Wider semiautomatic microdilution system (Fco. Soria Melguizo, Madrid, Spain), and results were interpreted according to the guidelines of the Clinical and Laboratory Standards Institute by using the criteria described for enterococci (23). The presence or absence of the erm(B), tet(M), and aac(6′)-aph(2″) genes was determined by PCR using specific primers [erm(B)-F (GAAAAGTACTCAACCAAATA) and erm(B)-R (AGTAACGGTACTTAAATTGTTTA), tet(M)-F (GTTAAATAGTGTTCTTGGAG) and tet(M)-R (CTAAGATATGGCTCTAACAA), and aac-aph-F (CCAAGAGCAATAAGGCATA) and aac-aph-R (CACTATCATAACCACTACCG)].

Genetic diversity.

Clonal relatedness was determined by pulsed-field gel electrophoresis (PFGE) with SmaI by using a protocol initially described for Streptococcus suis serotype 2 (24). We constructed a dendrogram using Phoretix software, version 5.0 (Nonlinear Dynamics Ltd., Newcastle, United Kingdom), based on the Dice coefficient.

Tn1546 characterization.

A primer-walking scheme described previously was used to characterize Tn1546 (25). The location of the glycopeptide resistance mechanism was assessed by hybridization of the I-CeuI-digested genomic DNA of the LMG 17956 ST28 strain, and of positive- and negative-control strains (the vanA-containing strains Enterococcus faecalis RC715 and Enterococcus faecium RC714 and the vancomycin-susceptible strain S. gallolyticus subsp. gallolyticus ATCC BA-2069, respectively), with probes for the 16S rRNA genes and vanA.

Initially, the nitrocellulose membrane was hybridized with the 16S rRNA gene probes obtained from the E. faecalis and E. faecium control strains and the glycopeptide-resistant S. gallolyticus isolate after universal PCR with V3–V4 primers.

In a second stage, the same membrane, after adequate washes, was newly hybridized with a vanA probe from a PCR vanA fragment obtained from an Enterococcus faecium control strain (25). All probes were labeled by random primer labeling with Redivue [³²P]dCTP (Amersham, Little Chalfont, Buckinghamshire, United Kingdom). Prehybridization and hybridization were carried out in Rapid buffer (Amersham) at 60°C for 30 min and at 56°C for 18 h, respectively. The membrane was washed twice at 56°C in 2× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate)–0.1% SDS and then twice at room temperature with 1× SSC–0.1% SDS and 0.7× SSC–0.1% SDS, successively. Autoradiography was carried out by filter exposure for 72 h at −80°C.

Conjugative transfer of vancomycin resistance and plasmid detection.

We tested the ability of vancomycin resistance to be transferred by conjugation by using the E. faecalis JH2-2 and E. faecium OG-RF1 strains (both of which are resistant to rifampin and fusidic acid) as recipients. Conjugation was developed by the filter mating method, and transconjugants were selected onto m-Enterococcus agar supplemented with 25 mg/liter of fusidic acid, 30 mg/liter of rifampin, and 16 mg/liter of vancomycin. The enterococcal model was chosen for conjugation because enterococci are the natural hosts for the vancomycin resistance determinants and also because of the lack of a validated S. gallolyticus subsp. gallolyticus recipient strain.

The presence of plasmids in the donor and recipient strains was explored after a plasmid extraction protocol and subsequent electrophoresis.

Whole-genome sequencing.

Total DNA from the vanA-containing S. gallolyticus subsp. gallolyticus LMG 17956 ST28 isolate was obtained by using a QIAamp kit (Qiagen, Hilden, Germany). A library was generated with 100 ng of DNA by using the Xpress Plus Fragment Library kit (Life Technologies, Eggenstein-Leopoldshafen, Germany). Quality was measured by the High Sensitivity DNA kit in the 2200 TapeStation system (Agilent Technologies, Palo Alto, CA, USA). Pyrosequencing was performed by using Ion Torrent technology with an Ion 316 v2 chip. The preliminary assembly obtained with MIRA software (26) was further completed by Era7 Bioinformatics. Protein-coding genes were annotated by performing a BLAST search against the COGs (clusters of orthologous groups) database (http://www.ncbi.nlm.nih.gov/COG/) (27, 28).

Comparison of genomes.

The genome sequence of the S. gallolyticus subsp. gallolyticus LMG 17956 ST28 isolate was compared with the previously sequenced genomes of S. gallolyticus subsp. gallolyticus UCN34 (BioProject accession no. PRJNA46061), S. gallolyticus subsp. gallolyticus ATCC 43143 (PRJDA162103), and S. gallolyticus subsp. gallolyticus ATCC BAA-2069 (PRJNA63617). Chromosomal rearrangements were inferred by aligning the genome sequences of these strains with MAUVE software (29). Groups of orthologous genes were defined by the OrthoMCL algorithm (30). Species-specific (exclusive) orthologous gene groups were represented using the VennDiagram package (version 1.6.5) of the R programming language (31, 32).

Gene family expansions.

Orthologous groups that had experienced significant gene expansion in the LMG 17956 lineage were identified using the probabilistic and phylogenetic framework provided by BadiRate software (33). In addition to the four S. gallolyticus strains, two closely related outgroup species were included in this analysis: Streptococcus equinus and Streptococcus thermophilus. In particular, the multiple-sequence alignments (MSAs) of their 1:1 orthologous protein-coding genes (a single gene copy per strain/species) were built with MAFFT, version 7 (34). These MSAs were concatenated using in-house-developed Perl scripts, leading to a concatenation that comprised 996,954 positions. The phylogenetic relationships among the six strains/species were estimated from this concatenation by RAxML, version 7.2.8, using a general time-reversible (GTR) model of DNA substitution with gamma distribution, with S. thermophilus as the outgroup species (to root the phylogenomic tree) (35, 36).

Statistical analysis.

The chi-square test was used for the antibiotic resistance analysis.

Nucleotide sequence accession numbers.

The genome of the vancomycin-resistant S. gallolyticus subsp. gallolyticus LMG 17956 ST28 isolate was deposited in the European Nucleotide Archive (ENA) under accession numbers CCBC010000001 to CCBC010000260 (http://www.ebi.ac.uk/ena/data/view/CCBC010000101).

RESULTS

Genetic diversity analysis.

The genetic diversity analysis based on the dendrogram calculated from the PFGE patterns clustered all isolates into five distinct groups (Fig. 1). Human bacteremic isolates clustered in groups II, III, and IV, whereas almost all animal isolates clustered in group I. It should be noted that the glycopeptide-resistant LMG 17956 ST28 strain, obtained from the feces of a calf, was grouped with the bacteremic human isolates of group IV (Fig. 1).

Antimicrobial susceptibility.

In general, all isolates presented low resistance rates, with no significant differences between animal and human isolates (Table 3). The most relevant result was the glycopeptide resistance (MIC values of both vancomycin and teicoplanin, 256 mg/liter) detected in the S. gallolyticus subsp. gallolyticus LMG 17956 ST28 strain obtained from the feces of a calf.

TABLE 3.

Comparison of antimicrobial resistance in animal and human isolates

Antibiotica	No. (%) of isolates with resistance		Pb
	Animal isolates (n = 18)	Human isolates (n = 23)
Fosfomycin	10 (55.5)	11 (47.8)	NS
Erythromycin	8 (44.4)	11 (47.8)	NS
Clindamycin	9 (50.0)	11 (47.8)	NS
Minocycline	7 (38.8)	8 (34.7)	NS
Gentamicin (500 mg/liter)	1 (5.5)	4 (17.3)	NS
Streptomycin (1,000 mg/liter)	7 (38.8)	8 (34.7)	NS
SxT	2 (11.1)	1 (4.3)	NS
Q/D	2 (11.1)	4 (17.3)	NS
Levofloxacin	1 (5.5)	4 (17.3)	NS
Vancomycin	1 (5.5)	0 (0)	NS

^aSxT, trimethoprim-sulfamethoxazole; Q/D, quinupristin-dalfopristin.

^bStatistical significance of differences in antimicrobial resistance between animal and human isolates. NS, not significant.

Molecular characterization of glycopeptide resistance.

We were unable to transfer vancomycin resistance by conjugation into the E. faecalis JH2-2 and E. faecium OG-RF1 enterococcal recipients after independent experiments. Moreover, the presence of plasmids was ruled out for both the donor and the recipient strains after independent negative plasmid extractions.

Initially, we followed the primer-walking strategy approximation to characterize the glycopeptide resistance mechanism. Using this technique, we identified the first 10,850 bp of Tn1546, but amplification of the vanY-vanZ region was not possible. Positive hybridization in the same I-CeuI fragment was observed with both the vanA and 16S rRNA genes, indicating integration of the Tn1546-like element into the bacterial chromosome (Fig. 2).

FIG 2.

Hybridization experiments performed to determine if the Tn1546-like element is located in plasmids or on the bacterial chromosome. (A) CeuI digestion of the total DNAs of a negative-control vancomycin-susceptible S. gallolyticus subsp. gallolyticus strain (lane 2), positive-control vanA-containing Enterococcus faecalis (lane 3) and Enterococcus faecium (lane 4) strains, and the vanA-containing S. gallolyticus subsp. gallolyticus LMG 17956 ST28 strain (lane 5). Lanes 1 and 6 correspond to lambda and low-range markers. (B) Hybridization with the 16S rRNA genes of the CeuI fragments. The positive bands correspond to chromosomal fragments. (C) Hybridization with the vanA gene. Arrows point to the Tn1546-like element, which was integrated into the bacterial chromosome.

Whole-genome sequencing of the vancomycin-resistant S. gallolyticus subsp. gallolyticus LMG 17956 ST28 strain.

For further characterization of the Tn1546 element, we sequenced the whole genome of the vancomycin-resistant S. gallolyticus subsp. gallolyticus LMG 17956 ST28 strain. This genome was deposited in the European Nucleotide Archive.

Bioinformatic analysis of contigs in comparison with the canonical Tn1546 sequence from the work of Arthur et al. (37) demonstrated an inversion of the proximal part of the transposon and a deletion of 1,835 bp in the distal Tn1546-like fragment (from bp 9016 to 10851), including the vanY and vanZ genes (Fig. 3).

FIG 3.

Genetic map of the Tn1546-like element detected in the LMG 17956 ST28 strain in comparison with the map of canonical Tn1546. The primers used in the primer-walking strategy are indicated above the map of the canonical Tn1546 element (25). At the right of the vanA and vanX genes, the vanY-vanZ region present on the consensus Tn1546 is absent. Note the partial duplication of the vanH gene.

The genome assembly comprises a total of 2,698,137 bp, with 2,410 genes, 58 tRNAs, 4 copies of rRNA genes, and 1 other noncoding RNA. In agreement with the analysis of clusters of orthologous groups (COGs), 35% of the genetic content was related to metabolic processes and 28% to information storage and processing, while another 17% of the genetic content contributed to cellular processes and signaling. However, 32% of the genome was poorly characterized, including those genes for which assignation to COGs was not possible (Fig. 4).

FIG 4.

Distribution of COGs in the whole-genome sequence of the LMG 17956 ST28 isolate.

Comparison of the four S. gallolyticus strains revealed that the LMG 17956 ST28 genome presented extensive rearrangements, contained a high number of transposase-encoding genes (n = 111) (mainly related to IS1167, IS1272, IS1548-like, IS4-like, IS630-SpnII, and IS66 elements), and was larger than the others (Fig. 5). Indeed, 246 genes are exclusive to the LMG 17956 ST28 strain (114 of these 246 genes had been described only in Streptococcus pneumoniae and Streptococcus pyogenes) (Fig. 6).

FIG 5.

MAUVE comparison of the genome sequences of the S. gallolyticus subsp. gallolyticus ATCC BAA-2069, ATCC 43143, and UCN34 strains with that of the vancomycin-resistant LMG 17956 ST28 isolate. Note the extensive rearrangements in our isolate. The arrow points to the Tn1546-like element, and the red lines in the distal section correspond to transposases and IS.

FIG 6.

Venn diagram showing the exclusive and shared ortholog groups defined by OrthoMCL for the 4 S. gallolyticus genomes.

The likelihood framework provided by BadiRate allows testing of whether these LMG 17956 ST28 gene acquisitions can be explained merely by the stochasticity underlying the mutational process or whether they have some selective meaning. Remarkably, our analyses identified some orthologous groups that were significantly expanded in the LMG 17956 ST28 lineage, including two related to transposable elements (an IS110 element and a recombinase), as well as a few associated with antibiotic resistance {streptogramin A acetyltransferase, virginiamycin lyase, the 6′-aminoglycoside nucleotidyltransferase [ANT(6′)], and a complete system for bacteriocin production} (Table 4).

TABLE 4.

Gene families significantly expanded in the S. gallolyticus subsp. gallolyticus LMG 17956 ST28 isolate

Molecular function	No. of genes in the following S. gallolyticus subsp. gallolyticus strain:				Statistical supporta
	ATCC 43143	ATCC BAA-2069	UCN34	LMG 17956 ST28
Transposase	0	1	0	7	28.789
Site-specific recombinase	0	0	0	4	28.621
Streptogramin A acetyltransferase	0	0	0	2	1.968
Virginiamycin B lyase (streptogramin B lyase)	0	0	0	2	1.968
Sequence-specific DNA binding	0	0	0	2	1.968
Dysgalacticin DysA2	0	0	0	2	1.968
UPF0236 protein in vanSb 3′ region	0	0	0	2	1.968
Uncharacterized protein	0	0	0	2	1.968
Uncharacterized protein	0	0	0	2	1.968

^aThe statistical support of these expansions was measured as the Akaike information criterion ratio between the first and the second best-fit models, as described in reference 32.

DISCUSSION

For humans, aside from its involvement in endocarditis, the major clinical interest of S. gallolyticus subsp. gallolyticus is the association with colorectal cancer. A new “bacterial driver-passenger model” has been proposed to explain the role of the gut microbiota in the colorectal cancer process (3). This model differentiates between the “bacterial drivers” in the microbiota (essentially the genus Bacteroides and the family Enterobacteriaceae), which are directly related to the process of carcinogenesis, and the tumor-foraging opportunistic pathogens called “bacterial passengers.”

The S. bovis complex belongs to the bacterial passenger category, mainly as a consequence of its efficient metabolic process, which permits adaptation to adverse environments. The prevalence of colonization of healthy human populations and animals has scarcely been investigated (8), and the presence of these organisms might be underestimated by a low fecal load.

The recent changes in the taxonomy of the S. bovis complex preclude a reliable retrospective analysis of most of the published epidemiological studies, except those that have investigated genetic diversity by the PFGE technique (38, 39). The multilocus sequence typing (MLST) tool has recently been adapted for the S. gallolyticus subsp. gallolyticus population, enabling investigators to start deciphering the genetic population structure of this organism (10). In our work, the PFGE patterns of the animal strains (group I) were clearly separated from those of strains causing bacteremic episodes in humans (groups II, III, and IV); moreover, the high genetic diversity among our isolates was corroborated by MLST. Nevertheless, it should be noted that the vancomycin-resistant strain LMG 17956 ST28 recovered from a calf was grouped with the human-invasive isolates. This fact implies a risk for introduction of a vancomycin-resistant isolate into human compartments, particularly in view of the genetic plasticity of this strain, which could facilitate the acquisition of new resistant/virulent determinants.

The only published study focusing on genetic differences between animal and human isolates failed to distinguish between strains of the two collections by use of the randomly amplified polymorphic DNA (RAPD) and amplified rRNA gene restriction analysis (ARDRA) techniques (40). Currently, PFGE remains the gold standard technique for exploring genetic diversity in epidemiological studies, and differences between the work of Sasaki et al. (40) and our work are probably the consequence of the different methodologies used. Hence, the whole-genome sequencing strategy is the best option for objective comparison of results between laboratories. In spite of the particular characteristics of each origin, the existence of host-adapted genetic lineages, as in the E. faecium population, cannot be ruled out (41).

In general, S. gallolyticus subsp. gallolyticus is susceptible to antimicrobial compounds frequently used for humans. However, previous reports have described resistance to macrolides (45 to 59%) and tetracyclines (56 to 78%) and high-level resistance to aminoglycosides (35 to 43%) (1, 14–16, 42). Therefore, the therapy of S. gallolyticus subsp. gallolyticus infections with the standard penicillin antibiotic regime is uncomplicated in the clinical routine, since resistant strains have not yet been described.

In our work, we detected a glycopeptide-resistant isolate recovered from the feces of a calf. Similar reports in the literature are limited and involve strains of both animal and human origins, clinical and colonization sources, and the enterococcal genetic mechanisms vanA and vanB (19–21). The gut microbiota is a complex ecosystem, with coexistence, and possibly colocalization, of bacterial organisms, which are frequently genetically related and able to exchange genetic material (genetic exchange communities) (43). These communities might include both Enterococcus species and the S. bovis complex. Our hypothesis is that the Tn1546-like element was originally located on an Enterococcus strain, was transmitted to S. gallolyticus subsp. gallolyticus by conjugation, and finally was integrated into the chromosome, provoking a deletion in the distal section.

The VanZ protein has been associated with teicoplanin resistance (44). Our isolate exhibited high-level resistance to both glycopeptides, even though the vanY-vanZ region was deleted. This result indicates that other pathways might contribute to the final inhibitory concentration of teicoplanin.

Although the primer-walking method failed in the characterization of the Tn1546-like element, whole-genome bacterial sequencing is an increasingly widely used strategy that allowed us to decode the structure of this transposon. In enterococci, differences between Tn1546-related elements from animals and humans (25, 45), as well as heterogeneity in Tn1546-like elements with IS1216 insertions (46, 47), have been described previously. These features has not been reported previously for S. gallolyticus subsp. gallolyticus. In any case, the presence of a Tn1546-like element in an S. gallolyticus subsp. gallolyticus strain genetically close to those causing bacteremia and endocarditis in humans is certainly a matter of concern due to the zoonotic transmission possibility.

The comparative genomic analyses demonstrated that the LMG 17956 ST28 strain is quite different from the three described previously (48–50), due mainly to the presence of numerous insertion elements (IS) and other transposable elements. This unusual large number of transposable elements could provide considerable genome plasticity and might explain the large size of this strain's genome. Indeed, this strain contains 246 exclusive genes, some of which show statistical support for a lineage-specific gene family expansion. Our results are in agreement with those of Richards et al. (51), whose findings show a higher number of gene gains for bovine than for human isolates of Streptococcus agalactiae. These accelerations of the bovine gene gain rates probably reflect the adaptive pressure in the calf gut environment, where the microbiome diversity might facilitate horizontal gene transfer events, especially the transfer of antibiotic resistances and transposable elements.