Corynebacterium conjunctivae: A New Corynebacterium Species Isolated from the Ocular Surface of Healthy Horses.

Twenty-two unidentified Gram-positive, rod-shaped organisms were recovered from the conjunctival surface of apparently healthy horses and subjected to a polyphasic taxonomic analysis. Based on cellular morphology and biochemical criteria, the isolates were tentatively assigned to the genus Corynebacterium, although they did not match any recognized species. Comparative 16S rRNA gene sequencing studies demonstrated that all of the isolates were phylogenetically members of the genus Corynebacterium. The isolates shared 99.4 to 100% 16S rRNA gene sequence similarity among the strains and 96.5% similarity with Corynebacterium tapiri 2385/12T, which was the closest phylogenetically related species. The DNA G+C content was 58.4 mol%. The major fatty acids were C15:0, C16:0, C17:1 ω8c and C18:1 ω9c, while the predominant mycolic acids consisted of C30:0, C32:0 and C34:0. The isolates were distinguished from related Corynebacterium species by a number of phenotypic properties. On the basis of phenotypic and phylogenetic evidence, it is proposed that the unknown isolates from horses be classified in the genus Corynebacterium as Corynebacterium conjunctivae sp. nov. The type strain of C. conjunctivae is ICM19-01138T (DSM 109759T = CCUG 73728T).

1. Introduction

Eye surfaces are continually exposed to the environment, providing a habitat for different bacterial and fungal microorganisms. This microbiota contributes to the ocular defense mechanism by producing antibacterial substances that prevent colonization by pathogenic microorganisms [1,2]. Despite this, eye infections are very common in horses. On many occasions, these infections are caused by members of the conjunctiva commensal microbiota that act as opportunistic pathogens after damage to the integrity of the ocular surface or surrounding areas [3]. Therefore, better knowledge of the normal equine conjunctiva microbiota may reveal its role in the pathogenesis of eye infections and help clinicians to prevent complications arising from the normal conjunctiva microbiota and possibly identify better treatment measures. Various studies have focused on identifying the normal ocular microbiota in horses, and most of them confirm that Gram-positive bacteria are predominant, with the genus Corynebacterium among the most frequently identified [4,5,6,7,8,9,10]. With the exception of those species associated with ocular pathology, mainly in humans, such as Corynebacterium bovis, Corynebacterium lowii, Corynebacterium oculi, Corynebacterium accolens, Corynebacterium macginleyi and Corynebacterium minutissimum [11,12,13,14], corynebacteria isolated from normal ocular flora are commonly identified only to the genus level. Only the isolation of Corynebacterium pseudotuberculosis and Corynebacterium mastitidis from normal cattle and mouse eyes has been reported [15,16]. This drawback limits the availability of information on the diversity of Corynebacterium species present in the normal flora of the eye surface. The presented work was carried out in healthy animals from the John Thomas Vaughan Large Animal Teaching Hospital at Auburn University. The purpose of this study was to characterize some commensal Corynebacterium-like organisms isolated from the conjunctival sac of healthy adult horses using a polyphasic approach. Based on the results of this study, we propose a new Corynebacterium species, Corynebacterium conjunctivae, as a commensal organism of the equine eye.

2. Materials and Methods

2.1. Animal Care

All animal procedures were conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and approved by the Animal Care and Use Committee of Auburn University (protocol#2013–2252). Eight healthy horses were included in this study [17]. The sampling period was between June and September 2014. All animals were free of ocular diseases after routine ophthalmic examination and had no reported history of recent antibiotic treatment.

2.2. Sample Collection

Bilateral conjunctival swabs were collected using routine procedures in 8 horses. One eye of one horse was excluded due to the presence of corneal scars. Samples were obtained along the surface of the ventral conjunctival fornix, taking special care not to contaminate the swabs with contact to the vibrissae, eyelids or eyelashes. The samples were placed in individual sterile vials on ice and processed immediately in the laboratory.

2.3. Bacterial Sampling

Swabs were plated onto tryptic soy agar (Becton Dickinson, El Paso, TX, USA) with 5% sheep blood (Hardy Diagnostics, Santa Maria, CA, USA) and then incubated at 35 °C for 18–24 h. After incubation, colonies that were morphologically different (a maximum of 5 colonies per plate) were picked and sub-cultured on TSB (Becton Dickinson, El Paso, TX, USA) under the same conditions. Individual purified isolates were cryopreserved in 20% glycerol at −80 °C until further use.

2.4. 16S rRNA Gene Sequencing

The phylogenetic similarity of twenty-two isolates was established by the sequencing of their 16S rRNA genes. Approximately 1400 bases of the 16S rRNA gene of the isolates were determined using PCR-amplified products and further sequenced as described previously [18]. The EzTaxon server (http:/eztaxon-e.ezbiocloud.net, accessed on 12 December 2021) was used to identify the phylogenetic neighbors and calculate pair-wise 16S rRNA gene sequence similarities [19]. The newly determined sequences were aligned with the sequences of the most closely related species and other representative species within the genus Corynebacterium retrieved from GenBank, using the program SeqTools Version 8.4.053 [20]. Three different algorithms were used to reconstruct the phylogenetic trees: neighbor-joining [21] using the programs SeqTools and treeview [20,22], as well as maximum-parsimony and maximum-likelihood using the software package MEGA (Molecular Evolutionary Genetics Analysis) version 5 [23]. Kimura’s two-parameter method was used to calculate the genetic distances for the neighbor-joining and maximum-likelihood algorithms [24], and close-neighbor-interchange (search level = 2, random additions = 100) was applied in the maximum-parsimony analysis. The stabilities of the groupings were estimated by bootstrap analysis (1000 replications).

2.5. DNA G+C Content

The HPLC method of Mesbah et al. [25] was used to determine the G+C content of the DNA of a representative isolate (strain ICM19-01138T) at the DSMZ (Braunschweig, Germany).

2.6. Fatty Acid Composition and Cell Wall Analyses

The cell fatty acid–fatty acid methyl ester (CFA–FAME) profile and the presence of wall murein acid and mycolic acids of the type strain (ICM19-01138T) were determined by the identification service of the DSMZ (DSMZ, Braunschweig, Germany; https://www.dsmz.de/services/microorganisms/biochemical-analysis, accessed on 9 October 2021).

2.7. Morphological, Physiological and Biochemical Characteristics

The API Coryne (version 2.0), API 50CH and API ZYM systems (bioMérieux España S.A, Madrid, Spain) were used to biochemically characterize the equine isolates following the manufacturer’s instructions. The CAMP test, with Staphylococcus aureus ATCC 25923 and lipophilic requirements, was conducted according to standard procedures [26].

2.8. PFGE Typing

Isolates were characterized by pulsed-field gel electrophoresis (PFGE) according to the specifications of Galán-Relaño et al. [27], but the restriction enzyme used was XbaI (MBI Fermentans), and the pulse times ranged from 0.1 to 20 s for a period of 22 h. Visual comparisons of the band patterns of isolates run in the same gel were used to determine the similarities between the restriction endonuclease digestion profiles. Strains were considered different when they differed in at least one band.

2.9. Susceptibility Testing

The equine (EQUIN1F) antimicrobial susceptibility panel for veterinary organisms (Trek Sensititre™, Thermo Fisher Scientific, Oakwood Village, OH, USA) was used to determine the antimicrobial susceptibility of the isolates. The following antimicrobials were included: amikacin, ampicillin, azithromycin, ceftiofur, chloramphenicol, doxycycline, enrofloxacin, erythromycin, gentamicin, imipenem, oxacillin, penicillin, rifampicin, tetracycline, ticarcillin, ticarcillin-clavulanic acid and trimethoprim-sulfamethoxazole. The concentrations for each antimicrobial included in the commercial panel are available at https://assets.thermofisher.com/TFS-Assets/MBD/Specification-Sheets/Sensititre-Plate-Layout-EQUIN1F.pdf, accessed on 17 July 2022). The breakpoints recommended by the Clinical and Laboratory Standards Institute were used to determine the antimicrobial susceptibility [28,29]. Staphylococcus aureus ATCC 29,213 was included as a quality control.

3. Results and Discussion

3.1. Phenotypic Characteristics

The 22 isolates consisted of Gram-positive, short rod-shaped cells, approximately 0.5 µm in width and 1 µm in length. The isolates were non-motile and did not produce spores. When cultured aerobically on Columbia blood agar plates, the isolates formed small colonies (approximately 1–2 mm diameter after 24 h incubation at 37 °C) that were circular, convex, dull, cream, and non-hemolytic. All isolates were catalase-positive and oxidase-negative and grew under anaerobic conditions. They were non-lipophilic and did not display a positive CAMP reaction with S. aureus after 48 h. Using the commercial API Coryne system, the equine isolates produced the numerical biochemical profile 0101304, corresponding to Corynebacterium pseudotuberculosis; however, the equine isolates could be easily differentiated because they did not produce acid from the maltose (C. pseudotuberculosis is positive) and additionally did not produce a reverse CAMP test [12]. Moreover, the colonies of equine isolates were not dry and opaque like those of C. pseudotuberculosis [30]. All equine isolates were also biochemically very homogeneous when they were characterized by the API 50CH and API ZYM systems. The biochemical characteristics are detailed in the species descriptions below. Three strains did not give a positive reaction for lipase C14 (ICM19-01138T, ICM19-01139 and ICM19-01141).

3.2. Phylogenetic Analysis

Sequence searches of GenBank based on their 16S rRNA gene sequences showed that the equine isolates were members of the genus Corynebacterium, exhibiting 99.4 to 100% similarity among them, thereby demonstrating high genealogical homogeneity. The 16S rRNA gene sequence of the strain ICM19-01138T exhibited 96.5% sequence similarity with the type strain of Corynebacterium tapiri 2385/12T and less than 96.0% with all other Corynebacterium species. The 16S rRNA sequence divergence values of > 3% indicate that the isolates from the horses represent a hitherto unknown species [31]. The phylogenetic relationship of the equine isolates within the genus Corynebacterium, based on the neighbor-joining algorithm, revealed that the 22 equine isolates clustered together and formed a distinct lineage to C. tapiri 2385/12T (Figure 1). The equine isolates exhibited weak associations (<70% bootstrap values) with C. tapiri and other validated Corynebacterium species (Figure 1). The GenBank accession numbers for the 16S rRNA gene sequences of the equine isolates are shown in Figure 1.

Figure 1

Neighbor-joining tree based on 16S rRNA showing the phylogenetic relationships within the genus Corynebacterium of Corynebacterium conjunctivae sp. nov. Mycobacterium smegmatis MC2 155T was used as an out-group. Bootstrap values (expressed as a percentage of 1000 replications) higher than 70% are given at the branching points. The nodes (groupings) also obtained in the maximum-likelihood and parsimony trees are indicated by open circles. The corresponding nodes also obtained in the parsimony tree are indicated by filled circles. Bar, 1% sequence divergence.

3.3. Chemotaxonomic and G+C Analysis

The wall murein acid hydrolysates of ICM19-01138T showed the presence of meso-diaminopimelic acid. The main mycolic acids of this strain consisted of short-chain mycolic acids (C28:0, 5%; C30:0, 28 %; C32:0, 53 % and C34:0, 14 %). The predominant cellular fatty acids (>1%) of the novel bacterium consisted of straight-chain saturated (C15:0, 23.4%; C16:0, 17.6%; C17:0, 11.6% and C18:0, 1.7%) and monounsaturated (C16:1 ω9c, 1.2%; C17:1 ω8c, 21.7% and C18:1 ω9c, 21.5%) fatty acid types. These chemotaxonomic characteristics resembled those of corynebacteria [4], supporting the phylogenetic results. The G+C content of the type strain ICM19-01138T was 58.4 mol%, which is also consistent with the values reported for members of the genus Corynebacterium (46–74 mol%) [12].

3.4. Antimicrobial Susceptibility and PFGE Analysis

All isolates were susceptible to amikacin, azithromycin, chloramphenicol, enrofloxacin, gentamicin, cefazolin, doxycycline, imipenem and rifampicin, whereas a few isolates were resistant to trimethoprim/sulfamethoxazole (9%), oxacillin (4.5%), penicillin (4.5%), erythromycin (4.5%) and tetracycline (9%). These results are in line with the high susceptibility detected to most of the commonly used veterinary antimicrobials in Corynebacterium isolates previously isolated from the conjunctiva of healthy horses [32].

A visual comparison of the restriction endonuclease digestion profiles generated by the 22 equine isolates revealed 10 different profiles represented by isolates ICM19/01138, ICM19/01139, ICM19/01141, ICM19/01144, ICM20/00190, ICM20/00191, ICM20/00193, ICM20/00195, ICM20/00199 and ICM20/00202 (Figure 2; lines 2–23). Corynebacterium isolates were recovered from all animals included in the study, supporting its high frequency of isolation from normal equine ocular flora [23]. These data, together with the relatively high genetic diversity detected by PFGE analysis and commonly observed in widespread microorganisms [33], suggest that the new Corynebacterium species identified in this study may be part of the commensal flora present in the conjunctival sac of equids. The formal description of Corynebacterium conjunctivae will allow its recognition and identification in the future and will improve knowledge of its distribution on ocular surfaces.

Figure 2

PFGE patterns of Corynebacterium conjunctivae sp. nov generated after XbaI macrorestriction. Lanes 1 and 12, Salmonella serotype Branderup strain H9812; Lanes 2 to 11, strains ICM19/01138T (four isolates), ICM19/01139 (four isolates), ICM19/01141 (five isolates), ICM19/01144 (two isolates), ICM20/00190, ICM20/00191, ICM20/00193, ICM20/00195 (two isolates), ICM20/00199 and ICM20/00202, respectively.

4. Conclusions

The polyphasic analyses based on phenotypic, chemotaxonomic and phylogenetic data indicated that the unidentified isolates from the ocular surface of apparently healthy horses belong to a distinct genomic species of the genus Corynebacterium, for which the name Corynebacterium conjunctivae sp. nov. is proposed, with strain ICM19-01138T (DSM 109759T = CCUG 60112T) as the type strain. C. conjunctivae can be differentiated from its nearest phylogenetic relative by the characteristics shown in Table 1.

Table 1

Characteristics that differentiate Corynebacterium conjunctivae sp. nov. from its closest phylogenetic relative Corynebacterium tapiri.

Characteristic	C. conjunctivae(22 Isolates)	C. tapiri 2385/12Tb
Nitrate reduction	−	+
Hydrolysis of:
Esculin	−	−
Urea	+	+
Production of:
Pyrazinamidase	−	+
Esterase C4	+	−
Ester lipase C8	+	−
β-glucuronidase	−	+
α-glucosidase	−	+
Alkaline phosphatase	+	+
Acid phosphatase	+	−
Acid production from:
Glucose	+	+
Ribose	+	+
Maltose	−	+
Saccharose	−	+

Data obtained from Baumgardt et al. [34].

Description of Corynebacterium conjunctivae sp. nov.

Corynebacterium conjunctivae (con.junc.ti’vae. N.L. n. conjunctiva, the membrane joining the eyeball to the eyelids; N.L. gen. n. conjunctivae, of conjunctiva).

Cells are Gram-positive, non-spore-forming, non-motile rods, approximately 0.5 µm in width and 1 µm in length. They are facultatively anaerobic, non-lipophilic and CAMP-negative with S. aureus. On Columbia blood agar, after 24 h incubation at 37 °C, colonies are cream, circular, convex, dull, approximately 1–2 mm in diameter and non-hemolytic. Urea is hydrolyzed but not aesculin and gelatin. Nitrate is not reduced. Acid is produced from glucose, D-fructose, D-mannose, ribose and trehalose but not from maltose, lactose, sucrose, glycogen, D-xylose, mannitol, N-acetyl-glucosamine, erythritol, glycerol, L-arabinose, L-xylose, adonitol, β-methyl-xyloside, galactose, dulcitol, inositol, sorbitol, L-rhamnose, sorbose, α-methyl-D-mannoside, α-methyl-D-glucoside, melibiose, inulin, melezitose, D-raffinose, xylitol, amygdalin, arbutin, salicin, cellobiose, β-gentibiose, D-turanose, D-lyxose, D-tagatose, D-arabinose, L-fucose, gentiobiose, D-fucose, D-arabitol, L-arabitol, 2-ketogluconate or 5-ketogluconate. Activity for alkaline and acid phosphatases, esterase C4, lipase C14, ester lipase C8, naphthol-AS-BI-phosphohydrolase, leucine arylamidase and β-glucosidase is detected. Pyrrolidonyl arylamidase, α-mannosidase, α-galactosidase, α-fucosidase, β-glucuronidase, pyrazinamidase, β-galactosidase, α-glucosidase, N-acetyl-β-glucosaminidase, chymotrypsin, trypsin, valine arylamidase and cystine arylamidase are not produced. Activity for lipase C14 is variable. Long-chain cellular acids are of the straight-chain saturated and monounsaturated types, with C15:0, C16:0, C17:1 ω8c and C18:1 ω9c predominating. Cell-wall murein is based on meso-diaminopimelic acid. Corynomycolic acids are present (C28–C34).

The type strain, ICM19-01138T (DSM 109759T = CCUG 60112T), was isolated from the ocular surface of an apparently healthy horse.