ABSTRACT
Streptococcus tigurinus is a newly described member of the Streptococcus mitis group. Due to the difficulty in distinguishing viridans group streptococci (VGS) by phenotype, analysis of 16S rRNA sequences is necessary for the accurate identification of most species. Through a laboratory policy of analyzing all clinically significant isolates from the VGS group by16S rRNA gene sequencing, we identified 14 S. tigurinus isolates from 11 patients. The Vitek 2 system most commonly gave an excellent rating to an incorrect identification (e.g., Streptococcus mitis), as did matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS) (e.g., Streptococcus oralis). S. tigurinus strains were recovered from numerous body sites, including the blood, peritoneal fluid, bone, synovial fluid, a perianal abscess, and an arm wound. Retrospective chart review indicated that most isolates were clinically significant, with bacteremia (n = 5), soft tissue infections (n = 3) osteomyelitis (n = 2), infected joint prosthesis (n = 2), and peritonitis (n = 2) being the most common, thus expanding the spectrum of disease associated with S. tigurinus.
INTRODUCTION
Streptococcus tigurinus was first described after isolation from a patient with infective endocarditis. Isolates of alpha-hemolytic, catalase-negative Gram-positive cocci were recovered from multiple blood cultures of a 74-year-old patient. The isolates were initially identified as viridans group streptococci (VGS). However, analysis of 16S rRNA sequences revealed that the isolates were distinct members of the Streptococcus mitis group (SMG) and were closely related to Streptococcus pneumoniae, Streptococcus pseudopneumoniae, Streptococcus mitis, Streptococcus infantis, and Streptococcus oralis (1). This novel Streptococcus strain was named Streptococcus tigurinus. Since the first report of S. tigurinus infection in 2012, S. tigurinus has been isolated from several additional sterile sites, including cerebrospinal fluid, heart valves, and joint fluid (2, 3). It has been associated with serious invasive infections, including infective endocarditis, culture-negative endocarditis, spondylodiscitis, bacteremia, meningitis, prosthetic joint infections, and thoracic empyema (2, 4). Small colony variants of S. tigurinus were associated with a prosthetic joint infection (3). S. tigurinus has also been isolated from the subgingival plaque of a patient with periodontitis (5). In vitro studies have shown S. tigurinus to be highly virulent in a rat model of experimental endocarditis (6).
Because of its relatedness to SMG, S. tigurinus has been hypothesized to be a commensal of the human oral cavity. Data to support this hypothesis, however, are conflicting (7). In a screen of saliva specimens from 31 volunteers, S. tigurinus was not identified among the more than 600 strains of alpha-hemolytic bacterial colonies isolated (2). In a second study that used a specific 16S rRNA gene real-time TaqMan PCR assay, however, S. tigurinus was detected in the saliva and subgingival plaque of 53% of patients sampled. S. tigurinus was not associated with periodontal disease in this study (8). Thus, the prevalence of S. tigurinus in the human oral cavity and the role of oral S. tigurinus in causing invasive infection warrant further investigation (7).
Due to the difficulty in differentiating members of the SMG using conventional and commercial phenotypic test methods, our laboratory analyzes all clinically significant isolates of SMG by16S rRNA gene sequencing. Through 16S sequencing, our laboratory identified 14 S. tigurinus isolates from 11 patients. The phenotypic characteristics and clinical significance of these isolates were investigated in this study.
MATERIALS AND METHODS
This study was completed with the Puget Sound Veterans Administration Medical Center institutional review board approval (MIRB 01012).
Bacterial strains.Fourteen isolates of S. tigurinus were isolated during routine culture procedures at the VA Puget Sound Healthcare System in Seattle, Washington. Isolates stored between 2000 and 2012 were from a randomly saved collection. However, after better recognition of S. tigurinus, we identified the 6 most recent patients in a 2.5-year period. Specimens were routinely cultured on sheep blood agar (SBA), chocolate agar, and MacConkey agar (Remel, Lenexa, KS). VGS colonies were identified using the biochemical- and sequencing-based approaches discussed below. Isolates were stocked at −70°C using an alphanumeric labeling system. All stocked clinical isolates were grown on sheep blood agar at 35°C in air enriched with 6% to 7% CO2.
Biochemical differentiation.Biochemical tests were performed with the Vitek 2 GP ID card and the Vitek 2 microbial identification system following the manufacturer's recommendations (bioMérieux, Durham, NC). Briefly, each S. tigurinus strain was subbed from a frozen stock onto SBA and passaged on SBA twice prior to analysis. Bacterial suspensions were made in a 0.9% sodium chloride solution and adjusted to a McFarland standard of 0.50 to 0.63 before testing with the GP ID card.
16S rRNA gene sequencing.The identity of the isolates was determined by 16S rRNA gene sequencing. Strains were subcultured from freezer stocks onto SBA and were incubated aerobically for 24 h. The DNA was extracted by using a Prepman apparatus (PE Applied Biosystems, Foster City, CA) following the manufacturer's instructions. Primers (synthesized by TIB Molbiol, Adelphia, NJ) VAB1 (5′-TGGAGAGTTTGATCCTGGCTCAG-3′) and VAB2 (5′-GTATTACCGCGGCTGCTGG-3′) were used to generate a 543-bp amplicon of the 16S rRNA gene. PCR was performed as previously described (9), and sequencing reactions were performed using the ABI Prism BigDye Terminator cycle sequencing kit (Applied Biosystems, Foster City, CA, USA), with sequences generated using the Biosystems 3500 genetic analyzer, according to the manufacturer's instructions. The sequences were analyzed and compared to information in the BIBI and GenBank databases (10). The phylogenetic tree was constructed with MUSCLE, a program for generating multiple alignments (11), with MEGA 5.1 software (http://www.megasoftware.net/).
Matrix-assisted laser desorption ionization–time of flight mass spectrometry analysis.Twelve S. tigurinus isolates were analyzed by matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS). For analysis, a portion of a single colony of each isolate was applied to a polished steel MSP 96-target plate as a thin film using a sterile wooden stick and analyzed using the MicroFlex LT mass spectrometer and database v3.3 (Bruker Daltonics, Billerica, MA), as previously described (12).
Antibiotic susceptibility testing.The antibiotic susceptibility testing of all isolates was determined using disk diffusion methods on Mueller-Hinton agar with 5% sheep blood, as recommended by the Clinical and Laboratory Standards Institute (CLSI), or using a Gram-positive panel on the Trek Sensititre (Thermo Scientific, Oakwood Village, OH). CLSI interpretive criteria for viridans group Streptococcus species were used.
RESULTS
By 16S rRNA gene sequencing, 14 alpha-hemolytic Streptococcus isolates from 11 patients were found to be identical or to have greater than 99.4% similarity to the type strain of S. tigurinus (i.e., DSM 24864) and approximately 98.5% similarity to Streptococcus cristatus and Streptococcus sanguis (Fig. 1). The published sequence of the S. tigurinus type strain differs by 13 bp of about 500 from S. mitis, and our strains differ by 0 to 3 bp from the type strain. The most variable 2-bp difference occurred between approximately bp 200 to 202.
Unrooted tree showing the phylogenetic relationships of the S. tigurinus strains isolated at the VA Medical Center to closely related species. The tree was inferred from 16S rRNA gene sequences by using the neighbor-joining method. The scale bar indicates the substitution rate per nucleotide. Sequences for all strains other than S. tigurinus were acquired from the GenBank database. Names and ATCC accession numbers are those cited in the GenBank database. Sequences for all S. tigurinus strains included in the dendrogram were generated by sequencing the first approximately 500-bp portion of the 16S rRNA gene at the Puget Sound VA Medical Center, Seattle, WA.
Strains of S. tigurinus identified in this study formed small, smooth, white to gray, alpha-hemolytic and nonhemolytic colonies on sheep blood agar, similar to many VGS. None of the strains were correctly identified by phenotypic methods. Biochemical and MALDI-TOF MS identification of the isolates was not successful. The Vitek 2 microbial identification system gave scores of excellent or very good identification as S. mitis or S. oralis for 8 of the isolates tested (66.7%) and a low discrimination identification for the remaining 4 isolates tested (33.3%), even though all identifications were incorrect (Table 1). S. tigurinus is not in the Vitek 2 database. MALDI-TOF MS was also used to identify each isolate. As spectra for S. tigurinus are not in the Bruker or the Vitek MS databases (13), the Bruker MALDI-TOF MS misidentified all of the isolates tested, usually as Streptococcus oralis (Table 1).
Identifications given by the bioMérieux Vitek 2 microbial identification system and the Bruker MALDI-TOF MSa
The clinical significance of recovery of S. tigurinus from each patient was determined through retrospective chart review. Patient demographic information, sites of isolation, and clinical findings are summarized in Table 2. All isolates of S. tigurinus were isolated from male patients, with a mean age of 60.8 years and an age distribution of 26 to 81 years of age. S. tigurinus strains were recovered from numerous body sites, including the blood (n = 5) and peritoneal fluid (n = 2), and 2 isolates each were associated with osteomyelitis and prosthetic joint infection. S. tigurinus was recovered in 3 soft tissue infections, including a perianal abscess, purulent cellulitis on an arm following trauma, and an arm wound at the site of intravenous drug administration. S. tigurinus represented the only organism isolated in 6 cases (55%). In other cases, it was recovered as a predominant species along with other organisms, including Enterococcus spp., other Streptococcus spp., and Leptotrichia wadei, a fastidious Gram-negative rod that is part of the normal human oral and intestinal flora (14). Leukocytosis was evident in 5 patients at the time of specimen collection, with inflammatory cells seen at the site of infection for 4 additional patients (Table 2).
A Summary of Patient Data and the Clinical Significance of Isolates of S. tigurinus
Four of the five patients with S. tigurinus bacteremia had significant comorbidities, including neutropenia, presence of a prosthetic joint, alcoholic cirrhosis, diabetes, and end-stage renal disease. A fifth patient with S. tigurinus bacteremia, patient 5, had no significant comorbidities but was emergently intubated 1 day prior to development of symptoms (fever, elevated WBCs) following anaphylactic shock from the administration of intravenous contrast. One of the patients with S. tigurinus bacteremia, patient 3, developed both paraspinal and epidural abscesses, and all four specimens were positive for S. tigurinus as the only organism. As Gram-positive cocci were seen in pleural fluid, the epidural abscesses were attributed to possible hematogenous spread of S. tigurinus from an oral site. S. tigurinus was isolated from the peritoneal fluid of 2 patients (patients 4 and 6). S. tigurinus was also isolated from the knee joint of a patient who underwent a knee revision arthroplasty 5 months prior (patient 7).
S. tigurinis was isolated from nonsterile sites in four patients (Table 2). In one patient, S. tigurinus was isolated from a wrist abscess at the site of illegal drug injection (patient 8); in a second patient, S. tigurinus was isolated from exudate collected from the anterior arm following trauma (patient 9). S. tigurinus was isolated along with Peptostreptococcus anaerobius in a patient with a recurrent perianal abscess (patient 10) and with Streptococcus anginosus from toe bone tissue in a patient with peripheral neuropathy (patient 11).
Antibiotic susceptibility results were available for 8 patients in this study (Table 3). Of those that were tested, all strains were susceptible to vancomycin and penicillin and all except strain F7522 were susceptible to levofloxacin. As shown, there was variable resistance to erythromycin, clindamycin, and tetracycline.
Antibiotic susceptibility of the S. tigurinus isolates identified in this studya
DISCUSSION
Although SMG isolates are commonly oral flora, our data suggest that the gastrointestinal tract and/or skin might also be sources from which S. tigurinus can access sterile sites and cause infection. In this study, seeding from oral sources is probable for 6 cases and possible for 2 more. Recent dental procedures or significant tooth decay were noted in the medical records for 4 of the 11 patients (36%). In addition, one patient was emergently intubated 24 h prior to isolation of S. tigurinus from blood specimens. It is possible that oral flora was introduced into the patient's bloodstream during the intubation procedure. Likewise, an aspiration event in one patient may have introduced S. tigurinus into the bloodstream. S. tigurinus was isolated from a patient who admitted to injecting drugs into his wrist several days before the development of the abscess. While not documented for this patient, it is possible the patient licked a needle prior to its use for injection of illicit drugs. One study estimated that one-third of injection drug users lick their needles prior to injecting (15). The skin could also have been a source of infection in this patient.
For the 2 patients who were diagnosed with peritonitis, the source of S. tigurinus was probably intestinal rather than oral. Spontaneous bacterial peritonitis (SBP) has been proposed to occur following a disturbance in the gut flora. This disturbance allows for overgrowth and translocation of enteric bacteria, typically Gram-negative enteric bacteria (16–18). Cirrhosis also predisposes an individual to overgrowth of bacterial flora and may lead to increased intestinal permeability (19, 20). Deficiencies of local immune response in patients with liver cirrhosis may also contribute to the development of SBP. The most frequent infection in patients with cirrhosis and ascites is SBP (19). The two patients with peritonitis in this study were documented to have cirrhosis and/or ascites. Thus, we cannot exclude the possibility that the intestine was the source of infection for these patients. The gastrointestinal tract was also the probable source of infection for the patient with a perianal abscess. Skin may have been the source of S. tigurinus for a patient with cellulitis and for a patient with chronic ulceration and osteomyelitis of a toe.
Accurate identification of S. tigurinus remains a challenge. S. tigurinus is not contained in the database for the Vitek 2 microbial identification system. As S. tigurinus spectra are not present in either of the currently available Biotyper (Bruker) or Vitek MS (bioMérieux) databases, S. tigurinus is not reliably identified by the most commonly used MALDI-TOF MS systems at this time (13). Currently, sequencing methods must be used to successfully identify S. tigurinus. Although our isolates were identified as S. tigurinus using standard procedures (21), there was sequence microheterogeneity of 2 bp in the 200-bp region of the 16S rRNA gene sequence. We detected no association between a particular sequence and a clinical condition.
In summary, using 16S rRNA gene sequencing, our laboratory identified 14 isolates of S. tigurinus from 11 patients and demonstrated that the usual methods of identification, MALDI-TOF MS and phenotypic biochemical testing, would not have correctly identified the organism. This study expands the spectrum of infection that can be attributed to S. tigurinus and suggests there are sources of S. tigurinus in addition to the oral mucosa. Our study adds to the growing body of literature demonstrating that S. tigurinus can cause severe, invasive infection and enlarges the spectrum of disease associated with this organism. To our knowledge, this is the first report of isolation of S. tigurinus from peritoneal fluid and a perianal abscess. Furthermore, including this report, 4 prosthetic joint infections caused by S. tigurinus have been described, highlighting the need for future studies on S. tigurinus biofilm formation (2, 3). As we are able to identify and describe the clinical significance of more of the VGS, we may find there are sufficient differences to warrant routine species-level identification of clinically significant isolates and to warrant improvements in the databases to do so.
ACKNOWLEDGMENTS
We thank the personnel of the microbiology laboratory at Veterans Administration Puget Sound Health Care System, especially Janelle Lee and Uyen Bui, for their support.
FOOTNOTES
- Received 8 June 2015.
- Returned for modification 7 July 2015.
- Accepted 30 August 2015.
- Accepted manuscript posted online 9 September 2015.
- Copyright © 2015, American Society for Microbiology. All Rights Reserved.
