Vol. 141: 53–69, 2020 DISEASES OF AQUATIC ORGANISMS Published online September 17 https://doi.org/10.3354/dao03521 Dis Aquat Org

Multilocus sequence analysis of diverse Streptococcus iniae isolates indicates an underlying genetic basis for phenotypic heterogeneity

Taylor I. Heckman1, Matt J. Griffin2, Alvin C. Camus3, Benjamin R. LaFrentz4, Danny Morick5, Rita Smirnov6, Tamir Ofek6, Esteban Soto1,*

1Aquatic Animal Health Laboratory, Department of Medicine & Epidemiology, School of Veterinary Medicine, University of California, Davis, CA 95616, USA 2Department of Pathobiology and Population Medicine, College of Veterinary Medicine, Mississippi State University, Stoneville, MS 39762, USA 3College of Veterinary Medicine, University of Georgia, Athens, GA 30602, USA 4USDA-ARS, Aquatic Animal Health Research Unit, Auburn, AL 36832, USA 5Morris Kahn Marine Research Station, Department of Marine Biology, Leon H. Charney School of Marine Sciences, University of Haifa, Haifa 3498838, Israel 6Central Fish Health Laboratory, Department of Fisheries and Aquaculture, Ministry of Agriculture & Rural Development, Nir David 1080300, Israel

ABSTRACT: Streptococcus iniae is a Gram-positive, opportunistically zoonotic bacterium infective to a wide variety of farmed and wild fish species worldwide. Outbreaks in wild fish can have detri- mental environmental and cultural impacts, and mortality events in aquaculture can result in sig- nificant economic losses. As an emerging or re-emerging pathogen of global significance, under- standing the coalescing factors contributing to piscine streptococcosis is crucial for developing strategies to control infections. Intraspecific antigenic and genetic variability of S. iniae has made development of autogenous vaccines a challenge, particularly where the diversity of locally endemic S. iniae strains is unknown. This study genetically and phenotypically characterized 11 S. iniae isolates from diseased wild and farmed fish from North America, Central America, and the Caribbean. A multilocus sequence analysis (MLSA) scheme was developed to phylogenetically compare these isolates to 84 other strains of Streptococcus spp. relevant to aquaculture. MLSA gen- erated phylogenies comparable to established genotyping methods, and isolates formed distinct clades related to phenotype and host species. The endothelial Oreochromis mossambicus bulbus arteriosus cell line and whole blood from rainbow trout Oncorhynchus mykiss, Nile tilapia Oreo - chromis niloticus, and white sturgeon Acipenser transmontanus were used to investigate the per- sistence and virulence of the 11 isolates using in vitro assays. In vivo challenges using an O. niloticus model were used to evaluate virulence by the intragastric route of infection. Isolates showed signif- icant differences (p < 0.05) in virulence and persistence, with some correlation to genogroup, estab- lishing a basis for further work uncovering genetic factors leading to increased pathogenicity.

KEY WORDS: Streptococcus iniae · Multilocus sequence analysis · MLSA · Streptococcosis · Aquaculture · Genotype

Resale or republication not permitted without written consent of the publisher

1. INTRODUCTION pected to approach 10 billion by 2050 (United Na tions 2019). Infectious diseases such as strepto coc cosis Aquaculture is the world’s fastest growing food pro- have substantial economic impacts on the industry duction sector and plays a critical role in providing through diseases outbreaks, treatment ex pen ditures, livelihoods and protein sources for a population ex - and other production losses (Lucas et al. 2012, FAO

*Corresponding author: [email protected] © Inter-Research 2020 · www.int-res.com 54 Dis Aquat Org 141: 53–69, 2020

2018). A major etiologic agent of pis cine streptococco- sequence analysis (MLSA) identifies variation in sis, Streptococcus iniae, is estimated to cost the global housekeeping genes in a reproducible and dissem- aquaculture industry over 100 million US$ annually inable manner. MLSA has been used successfully to (Shoemaker et al. 2001), and the number of hosts, genetically characterize a number of bacterial species habitats, and countries impacted continues to expand. (Martens et al. 2008, Macheras et al. 2011, Glaeser & Infection by S. iniae is multisystemic, but is commonly Kämpfer 2015, Whatmore et al. 2016, Liu et al. 2017), associated with menin gitis, pano phthalmitis, and sep- and publicly accessible databases facilitate inclusion ticemia. Pathogenesis varies depending on the bacte- of MLSA and whole-genome data from isolates across rial strain, host species, and environmental conditions the globe (Peacock et al. 2002, Patchanee et al. 2012, (Agnew & Barnes 2007). The host range of S. iniae in- Glaeser & Kämpfer 2015). However, no MLSA scheme cludes over 30 species of fresh, euryhaline, and salt- has been established for S. iniae. Herein, isolates pre- water fish (Agnew & Barnes 2007, Silayeva et al. viously typed by Rep-PCR (Soto et al. 2017a, T. I. 2020). Outbreaks with high mortalities occur in com- Heckman et al. unpubl. data) were used to evaluate mercially valuable fish, in cluding tilapia Oreochromis an MLSA method for S. iniae. The MLSA was de- spp. (Perera et al. 1994, Ortega et al. 2018), rainbow signed to be inclusive of other piscine streptococcal trout Oncorhynchus mykiss (Eldar et al. 1994, Erfan- pathogens, giving it broader relevance to fish health manesh et al. 2012), white sturgeon Acipenser trans- research and diagnostics. To further characterize montanus (Soto et al. 2017a), and others. Ornamental these isolates and improve laboratory techniques species (Russo et al. 2006), wild marine fish (Eldar et used to assess S. iniae pathogenicity, in vitro and in al. 1999, Fer guson et al. 2000, Keirstead et al. 2014, vivo assays were employed to assess isolate virulence Berzak et al. 2019), and mammalian species including in relevant fish hosts. Through the in vivo assay, this humans (Weinstein et al. 1997, Koh et al. 2004, 2009, study also validates a biologically relevant intra - Facklam et al. 2005, Lau et al. 2006, Sun et al. 2007) gastric gavage challenge model for S. iniae infections. can also be affected. As such, S. iniae is a pathogen of concern in the fields of aquaculture, conservation, and animal health. Disease transmission be tween 2. MATERIALS AND METHODS farmed and wild fish has been implicated in several outbreaks (Zlotkin et al. 1998, Colorni et al. 2002, Bro- 2.1. Bacteria mage & Owens 2002, Berzak et al. 2019), illustrating the need for a better appreciation of the disease and A total of 11 clinical isolates of Streptococcus iniae its transmission directly between hosts and indirectly from 5 wild and farmed fish species from North in the environment. Despite this, the pathogenic America, Central America, and the Caribbean were mechanisms used by S. iniae and its epidemiology re- used in all aspects of this study; 35 additional S. iniae main incompletely understood, and vaccination efforts isolates from piscine and mammalian sources, as well have been met with variable success (Bachrach et al. as representative S. agalactiae, S. dysgalactiae and 2001, Creeper & Buller 2006, Agnew & Barnes 2007, S. icta luri isolates, were included for the MLSA Eyngor et al. 2008). A contributing factor to these dif- (Table 1). Isolates were stored in 1 ml aliquots in ficulties is our incomplete knowledge of the antigenic brain heart infusion broth (BHI; MP Biomedicals) and genetic diversity of S. iniae, especially related to with 20% glycerol at −80°C. Before each assay, iso- relevant differences in pathogenesis between strains. lates revived from frozen stocks were grown at 30°C To this end, S. iniae collected from wild and for 24 h in BHI, with shaking, or for 48 h on trypticase cultured fish species in marine and freshwater envi- soy agar supplemented with 5% sheep’s blood (SBA; ronments across North America, Central America, University of California, Biological Media Services), and the Caribbean were genetically and phenotypi- unless otherwise noted. The 0.5 McFarland standard cally characterized. Genotyping schemes, such as (~1.5 × 108 CFU ml−1) used for the in vivo and in vitro repetitive sequence mediated fingerprinting (Rep- virulence assays corresponded to a bacterial suspen- PCR) and pulsed field gel electrophoresis (PFGE), sion in phosphate-buffered saline (PBS) with an opti- have been used for S. iniae typing (Weinstein et al. cal density measurement of 0.14−0.155 at 600 nm, 1997, Chou et al. 2014), but the portability of data read on a UV/Vis photometer (BioPhotometer Plus, from these techniques is cumbersome, results are of- Eppendorf AG). Isolates were phenotypically charac- ten poorly reproducible, and reliable comparisons re - terized using the API 20 STREP system following the quire isolates to be processed simultaneously. In con- manufacturer’s instructions and read at 48 h (bio- trast to these image-based approaches, multilocus Mérieux). Heckman et al.: Multilocus sequence analysis of Streptococcus iniae 55

Table 1. Streptococcus isolates used in this study. Isolates in bold were used for both multilocus sequence analysis (MLSA) typing and further phenotypic characterization

Isolate Origin Geography Reference

S. iniae ECO86-17 Spotted rose snapper Lutjanus guttatus Central America T. I. Heckman et al. unpubl. data S. iniae B8 Wild reef fish S. Caribbean Basin Soto et al. (2017a) S. iniae K08-409H Wild reef fish N. Caribbean Basin Soto et al. (2017a) S. iniae F15-4-3 Tilapia Oreochromis spp. California Soto et al. (2017a) S. iniae WS-6B White sturgeon Acipenser transmontanus California Soto et al. (2017a) S. iniae ARK PB03-62B Albino rainbow shark Epalzeorhynchos frenatum Florida Soto et al. (2017a) S. iniae LSU 01-105 Tilapia Oreochromis spp. Minnesota Soto et al. (2017a) S. iniae LSU 10-070 Tilapia Oreochromis spp. Florida Soto et al. (2017a) S. iniae LSU 94-034 Tilapia Oreochromis spp. Massachusetts Soto et al. (2017a) S. iniae LSU 96-525 Tilapia Oreochromis spp. Iowa Soto et al. (2017a) S. iniae LSU 94-036 Tilapia Oreochromis spp. Illinois Soto et al. (2017a) S. iniae WS-6H White sturgeon A. transmontanus California Soto et al. (2017a) S. iniae 831 Pygmy hippo Choeropsis liberiensis Texas This study S. iniae 832 Pygmy hippo C. liberiensis Texas This study S. iniae 837 Pygmy hippo C. liberiensis Texas This study S. iniae 105-04 Human Homo sapiens California Facklam et al. (2005) S. iniae 4780-01 Human H. sapiens California Facklam et al. (2005) S. iniae 4787-01 Human H. sapiens California Facklam et al. (2005) S. iniae 4989-04 Human H. sapiens California Facklam et al. (2005) S. iniae 143-01 Human H. sapiens California Facklam et al. (2005) S. iniae 2388-02 Human H. sapiens California Facklam et al. (2005) S. iniae 1056-03 Human H. sapiens Pennsylvania Facklam et al. (2005) S. iniae SS1440 Human H. sapiens Canada Facklam et al. (2005) S. iniae SS1543 Human H. sapiens Canada Facklam et al. (2005) S. iniae 8278 Tilapia Oreochromis spp. Israel This study S. iniae 8679 Hybrid striped bass Morone saxatilis × M. chrysops Israel This study S. iniae 11042 Tilapia Oreochromis spp. Israel This study S. iniae 11979 Tilapia Oreochromis spp. Israel This study S. iniae 12302 Hybrid striped bass M. saxatilis × M. chrysops Israel This study S. iniae 12424 Hybrid tilapia O. aureus × O. niloticus Israel This study S. iniae 14957 Hybrid striped bass M. saxatilis × M. chrysops Israel This study S. iniae 15091 Hybrid striped bass M. saxatilis × M. chrysops Israel This study S. iniae 15414 Hybrid striped bass M. saxatilis × M. chrysops Israel This study S. iniae 15843 Barramundi Lates calcarifer Israel This study S. iniae 15957 Barramundi L. calcarifer Israel This study S. iniae 16029 Hybrid striped bass M. saxatilis × M. chrysops Israel This study S. iniae 16616 Hybrid striped bass M. saxatilis × M. chrysops Israel This study S. iniae 17105 Hybrid striped bass M. saxatilis × M. chrysops Israel This study S. iniae 17737 Barramundi L. calcarifer Israel This study S. iniae 19735 Barramundi L. calcarifer Israel This study S. iniae 20130 Barramundi L. calcarifer Israel This study S. iniae 20960 Hybrid striped bass M. saxatilis × M. chrysops Israel This study S. iniae WS-10A White sturgeon A. transmontanus California This study S. iniae WS-10B White sturgeon A. transmontanus California This study S. iniae WS-10C White sturgeon A. transmontanus California This study S. iniae WS-10D White sturgeon A. transmontanus California This study S. ictaluri CNA2848 Catfish Ictalurus spp. Unknown This study S. agalactiae RUSVM-CR Nile tilapia O. niloticus North America This study S. dysgalactiae subsp. White sturgeon A. transmontanus Uruguay This study equisimilis STC3

2.2. Buoyant density assays CPS (Håkansson et al. 1987). Assays were performed as described previously (Millard et al. 2012), with Buoyant density assays can be used to estimate the some modifications. Briefly, a standard isotonic Per- quantity of capsular polysaccharide (CPS), as the coll solution was prepared by mixing 9 parts Percoll buoyant density is inversely related to the amount of with 1 part 1.5 M NaCl. Each isolate was grown to 56 Dis Aquat Org 141: 53–69, 2020

mid-exponential phase in BHI and 1 ml pelleted by strain NGBS061, GenBank Acc. No. CP007631; S. centrifugation, washed in PBS and resuspended in agalactiae strain SGEHI2015-113, GenBank Acc. No. 500 μl PBS. The suspension was layered onto the Per- CP025026, S. agalactiae strain SG-M29, GenBank coll solution and centrifuged at 4000 × g for 90 min. Acc. No. CP021866) and verified in silico by BLASTn The experiment was repeated with new cultures 3 searches of Streptococcus spp. in GenBank (Table S1 times. in the Supplement at www. int-res. com/articles/ suppl/ d141p053_ supp. pdf). Primer sequences are listed in Table 2. PCR using Phusion high-fidelity DNA poly- 2.3. Genomic DNA extraction merase (Thermo Scientific™) was carried out following manufacturer recommendations with 3% added Swabs of each isolate from freezer stocks were DMSO, and annealing temperatures 6°C below the used to inoculate 5 ml aliquots of BHI and were ex- suggested melting temperature for each primer set panded at 30°C overnight with shaking at 150 rpm. A (Table 2). Aliquots of amplification reactions, along 1 ml aliquot of the expanded bacterial suspension with concurrently run molecular weight standards was centrifuged for 10 min at 5000 × g (7500 rpm). (Quick-Load® Purple 100 bp DNA Ladder; New Eng- Genomic DNA (gDNA) was isolated from the con- land BioLabs), were electrophoresed through 1% centrated pellet using the DNeasy® Blood and Tis- agar ose gels supplemented with SYBR® Safe DNA gel sue kit (Qiagen) following manufacturer recommen- stain (Invitrogen) (1 μl ml−1) and visualized under UV dations for Gram-positive bacteria. The quality and light to confirm the presence of appropriate sized quantity of recovered DNA was assessed using a bands. NanoDrop™ One Microvolume UV-Vis Spectropho- PCR products were purified using the AccuPrep® tometer (Thermo Scientific™) and samples with 260/ PCR Purification Kit (Bioneer) or QIAquick PCR Pu- 280 ratios of 1.8−2.0 were cryogenically stored rification Kit (Qiagen), and their concentration and (−20°C) until further analysis. purity were assessed by Nanodrop. Purified products and corresponding forward primers were diluted and submitted for Sanger sequencing at the University of 2.4. MLSA California, Davis Sequencing facility (UC Davis, CA, USA) or through GENEWIZ (South San Francisco, The purified gDNA from each isolate was used for CA, USA). Sequences for each housekeeping gene PCR amplification and sequencing of 9 housekeeping were aligned by MUSCLE with default settings in genes: arcC, glnA, groEL, gyrB, mutS, pheT, prkC, Geneious Prime (v.2019.0.4). Sequence ends were an - rpoB, and tkt. Degenerate primers were initially de- notated with a 0.01 error probability limit and trimmed signed based on alignments of publicly available to the region of quality bases shared by all isolate se- Streptococcus spp. genomes available in the National quences. A representative trimmed se quence of each Center for Biotechnology Information’s Microbial housekeeping gene was used in BLAST searches of a Genome Database (S. iniae strain YSFST 01-82, Gen- localized database populated by S. iniae, S. agalac- Bank Acc. No. CP010783; S. iniae strain FP5228, Gen- tiae, S. dysgalactiae, S. ictaluri, and S. pyogenes Bank Acc. No. CP024843; S. iniae strain QMA0248, genomes downloaded from GenBank. The database GenBank Acc. No. CP022392; S. agalactiae strain represented 46 fish and mammalian isolates derived WC1535, GenBank Acc. No. CP 016 501, S. agalactiae worldwide (Table 3). Trimmed sequences for each

Table 2. Primers used in the MLSA scheme for genotyping Streptococcus iniae

Name Forward primer Reverse primer

Carbamate kinase (arcC) GCW AAA GCA CAA CAA GAA GC CGC CAD CCA CGR CCW GCA TC Glutamine synthetase (glnA) MAA ATG GGY TTT GAA GTD GAA GC RTC AAT TTC CCA TTG WGA MAY Chaperonin (groEL) TAA ATT TTC AGC AGA TGC SCG Y ACT TCA AGY TCT GTY TCC ATA CC Gyrase B (gyrB) GGW GAR GAT GTT CGT GAA GG TCC ATT GTT GTT TCC CAA AG DNA mismatch repair protein (mutS) WAA AAA TTC TGA RCG YTA TGG AAG GTT GAT TGC CCA GAA AT Phenylalanine-tRNA ligase subunit beta (pheT) GGT CAA CCW ATG CAT GCT TT WCA TYG GCC ACA TVA GTT C Serine/threonine-protein kinase (prkC) TAT TTG CTG GTC GTT ATC GS YCC CAT SGC ATA AAT ATC AC RNA polymerase B (rpoB) TGT TGG TAC TGG TAT GGA AAA CGT TGT CCA CCA AAT Transketolase (tkt) CAG AAG ATG TKA AAG GAC GTT GCC ATK GCA AAT TCA CGW AC Heckman et al.: Multilocus sequence analysis of Streptococcus iniae 57 , Israel CP007587 Complete China NZ_CP032400 Complete , USA GCA_003674285.1 Scaffold Accession number Assembly level , USA GCA_003674315.1 Contig , USA GCA_003674305.1 Scaffold O. niloticus O. niloticus, , USA GCA_003674265.1 Contig M. chrysops , USA GCA_003697605.1 Contig , South Korea NZ_CP024843 Complete × , Reunion GCA_003674875.1 Scaffold , Israel GCA_003674925.1 Scaffold , Israel GCA_003674945.1 Scaffold , USA ASM18801v3 Contig , USA GCA_003675085.1 Contig , Thailand GCA_003674935.1 Contig M. chrysops , Brazil CP018623 Complete × , Australia CP029632 Complete , Australia GCA_003675145.1 Scaffold , Australia GCA_003674715.1 Contig , Australia GCA_003674565.1 Scaffold , USA GCA_003674215.1 Scaffold , Thailand GCA_003674865.1 Scaffold , Australia GCA_003675245.1 Scaffold , Australia GCA_003674385.1 Scaffold , China CP005941 Complete , South Korea NZ_CP010783 Complete , Australia NZ_CP022392 Complete , Australia GCA_003674425.1 Contig , Australia GCA_003675285.1 Scaffold , Australia GCA_003675195.1 Scaffold , Australia GCA_003697565.1 Contig , Australia GCA_003697585.1 Scaffold , Australia GCA_003674995.1 Contig , Australia GCA_003674985.1 Contig Epalzeorhynchos bicolor , Taiwan NZ_CP017952 Complete , Israel CP007586 Complete I. geoffrensis Inia geoffrensis , Italy NC_004116 Complete sp., USA GCA_003674205.1 Contig sp., USA GCA_003674235.1 Scaffold sp., Honduras GCA_003674165.1 Scaffold sp., USA GCA_003674345.1 Contig sp., USA GCA_003674395.1 Contig M. saxatilis Morone saxatilis C. striata Channa striata L. calcarifer L. calcarifer , USA NC_002737.2 Complete , Australia NZ_CP015238.2 Complete , Japan AP012976 Complete , Germany NC_019042 Complete Ictalurus punctatus Pteropus alecto , UK NZ_CM001076 Complete P. olivaceus P. P. olivaceus P. Paralichthys olivaceus Oncorhynchus mykiss Oncorhynchus mykiss Oncorhynchus mykiss Netuma thalassina Chromobotia macracanthus C. macracanthus L. calcarifer L. calcarifer Lates calcarifer L. calcarifer L. calcarifer L. calcarifer L. calcarifer L. calcarifer L. calcarifer O. niloticus O. niloticus Oreochromis niloticus Scortum barcoo H. sapiens H. sapiens H. sapiens H. sapiens Homo sapiens Oreochromis Oreochromis Oreochromis Oreochromis Oreochromis Bos taurus strains with full genome assemblies available on GenBank used in MLSA Streptococcus ATCC 27957ATCC Cow 167 Human AC-2713 Human Table 3. Table dysgalactiae equisimilis equisimilis subsp. subsp. subsp. QMA0271 Giant catfish S13 Nile tilapia 2603V/R Human M1 GAS SF370 Human NS53 Human 707-05 Channel catfish QMA0190 Striped snakehead QMA0233 Bone, barramundi QMA0248 Barramundi QMA0373 Barramundi QMA0448 Hybrid striped bass QMA0458 Red-tailed black shark QMA0462 Clown loach QMA0463 Clown loach QMA0466 Tilapia QMA0468 Tilapia QMA0490 Tilapia SF1 Olive flounder YM011 Attenuated strain from GX005, Nile tilapia YSFST01-82 Olive flounder QMA0189 Rainbow trout QMA0249 Bone, barramundi QMA0371 Jade perch QMA0445 Tilapia QMA0446 Tilapia QMA0447 Hybrid striped bass FP5228 Olive flounder ISNO Attenuated strain from ISET0901, Nile tilapia 89353 Nile tilapia ISET0901 Nile tilapia QMA0071 Barramundi QMA0188 Rainbow trout QMA0074 Barramundi QMA0080 Barramundi QMA0142 Barramundi QMA0141 Amazon river dolphin QMA0084 Black flying fox QMA0140 Amazon river dolphin QMA0155 Barramundi QMA0165 Barramundi QMA0177 Barramundi QMA0187 Striped snakehead QMA0186 Rainbow trout S. iniae S. iniae S. iniae S. iniae S. iniae S. iniae S. iniae S. iniae S. iniae S. iniae S. iniae S. iniae S. iniae S. iniae S. pyogenes S. pyogenes S. iniae S. iniae S. iniae S. iniae S. iniae S. iniae S. dysgalactiae S. dysgalactiae S. iniae S. iniae S. dysgalactiae S. ictaluri S. iniae S. iniae S. iniae S. agalactiae S. agalactiae Isolate S. agalactiae Origin S. iniae S. iniae S. iniae S. iniae S. iniae S. iniae S. iniae S. iniae S. iniae S. iniae S. iniae S. iniae 58 Dis Aquat Org 141: 53–69, 2020

housekeeping gene were concatenated alphabeti- mal essential media-2 + HEPES + 10% fetal bo vine cally for each isolate. Concatenates were exported to serum (MEM) and grown to confluence at 25°C. A 0.5 MEGA-X (Kumar et al. 2018) and aligned by MUS- McFarland standard was generated for each isolate CLE using the default settings. A maximum likeli - and 100 μl added to their respective cell wells (MOI hood tree was generated using the Tamura 3 parame- 1:100). Sterile PBS (100 μl) was added to uninfected ter model with a Gamma distribution (Tamura 1992), cells as a negative control. Cells were incubated for selected based on Bayesian and Akaike information 3 h at 20 or 30°C, washed 3 times with 1 ml of MEM to criteria in MEGA-X. The percentage bootstrap confi- remove non-adherent bacteria, and placed back in dence levels calculated from 1000 re-samplings of the the same incubator for 24 h. original data. The Cytotox96© Non-Radioactive Assay (Pro mega) was used according to the manufacturer’s instructions to quantitate release of the stable cytosolic enzyme 2.5. Whole-blood survival assays lactate dehydrogenase (LDH). An enzymatic assay is used to measure LDH release into culture supernatant, Isolate survival in whole blood was assessed using converting a tetrazolium salt (INT) into a red formazan methods adapted from Locke et al. (2007a). Whole, product. The amount of color produced is pro por - lithium-heparinized blood was collected from healthy, tional to the number of lysed cells. Adsorption at anesthetized rainbow trout, Nile tilapia, and white 490 nm was measured using a Cytation™ 5 Imaging sturgeon by caudal venipuncture. Aliquots of 300 μl Reader (BioTek). Adsorption values were standard- were transferred to sterile microcentrifuge tubes. A ized against negative controls and percent toxicity 0.5 McFarland suspension was generated for each calculated by dividing experimental treatment values bacterial isolate from 48 h cultures on SBA and sus- by the positive control value. Statistical significance pensions diluted to ~1.5 × 105 CFU ml−1 in sterile PBS. was determined using Kruskal-Wallis with Dunn’s A 2 μl (~300 CFU) aliquot of bacterial suspension was multiple comparisons test in GraphPad Prism v.8.3.0. added to each tube of blood, in triplicate, and incu- bated for 1 h with shaking at 20°C for the trout and sturgeon blood and 30°C for the tilapia blood; 2 μl of 2.7. Intragastric challenge bacterial suspension were also added to 300 μl sterile PBS to establish baseline CFU. Six 25 μl aliquots of Challenges were conducted under the University blood or PBS were spot-plated onto SBA and colonies of California, Davis School of Veterinary Medicine counted after 72 h. Percent survival was calculated by Institutional Animal Care and Use Committee ap - dividing colony counts from blood by respective colo - proved protocol no. 19645. To investigate in vivo ny counts from PBS. Spots with >100 colonies were pathogenesis, Nile tilapia fingerlings were chal- ex cluded from analysis due to in creased probability of lenged with each of the 11 North American S. iniae counting error. Experiments were repeated 3 times isolates. Fresh 0.5 McFarland solutions (~1.5 × 108 using blood from the same group of fish. Statistical CFU ml−1) were generated for each isolate. A total of significance was determined in GraphPad Prism 16 fish were gavage challenged with 0.1 ml of the v.8.3.0 (GraphPad Software) using Brown-Forsythe McFarland solutions (~1.5 × 107 CFU) or with a sterile and Welch ANOVA tests with Dunnett’s T3 multiple PBS control using a 20G × 1.5’’, 1.9 mm flexible plas- comparisons test, with individual variances computed tic feeding needle (Cadence Science®). Fish were for each comparison. kept in aerated flow-through tanks at 28−30°C, and morbidity and mortality was recorded daily for 21 d. Moribund fish or those exhibiting abnormal swim- 2.6. Cytotoxicity assays ming, lethargy, or exophthalmia were euthanized with buffered MS-222 (500 mg l−1) in sodium bicar- To test the ability of S. iniae to kill endothelial cells, bonate and necropsied. Posterior kidney and brain the Oreochromis mossambicus bulbus arteriosus cell swabs were plated on SBA and incubated at 30°C. line (TmBs) (Lewis & Marks 1985, Soto et al. 2017b) Whole fingerlings were fixed in 10% neutral was challenged with the different bacterial isolates at buffered formalin, decalcified, processed routinely, 20 and 30°C. This cell line was chosen as the heart is and stained with hematoxylin and eosin (H&E) for commonly affected in S. iniae infection (Keirstead et light microscopic examination. Select sections were al. 2014, Woo & Cipriano 2017). TmBs were plated in stained with a Brown and Hopps Gram stain. At the 24 well dishes (2.5 × 105 cells well−1) in 500 μl of mini- end of the 21 d challenge, all surviving fish were Heckman et al.: Multilocus sequence analysis of Streptococcus iniae 59

euthanized. Three fish per treatment were fixed and strain isolate (ATCC29178), which fell into Clade E. processed for histologic evaluation. Three fish per QMA0084 from a black flying fox Pteropus alecto treatment were necropsied and posterior kidney and formed its own lineage but shared a high degree of brain swabs plated on SBA then incubated at 30°C homology with the fish isolates. The second dolphin for 48 h. Survival curve analysis and statistics were isolate (QMA0141) and isolates from pygmy hippos performed using GraphPad Prism v.8.3.0. Statistical (831, 832, 837) were the most divergent lineages. significance was determined by Log-rank (Mantel- Including isolates of different streptococcal species Cox) and Gehan-Breslow-Wilcoxon tests. slightly reduced intraspecies resolution but grouped isolates according to their respective species and sub species, consistent with phylogenies derived 3. RESULTS from whole genome comparisons (Gao et al. 2014) (Fig. 2). The concatenated sequences of S. iniae, S. 3.1. Genetic characterization agalactiae, S. dysgalactiae, S. ictaluri, and S. pyoge- nes shared 87.3% pairwise identity. The MLSA primers successfully amplified target regions of the selected housekeeping genes in Strep- tococcus iniae, S. agalactiae, S. dysgalactiae, and 3.2. Morphology and hemolysis S. ictaluri. MLSA of S. iniae isolates generated a max- imum likelihood tree consisting of 5 major clades, The North American isolates formed small translu- A−E (Fig. 1), with several lineages consisting of 1 or 2 cent colonies within 24 h on SBA that became white isolates. There was an overall high level of similarity within 48 h. Colony size differed between phylo - between isolates among the chosen housekeeping groups, with Clade D isolates forming larger (≤3 mm) genes (99.9% pairwise identity). Isolates as signed to colonies than isolates in Clades A, B, and E (≤1.5 mm) the same vertical line had no bp differences. Clusters (Fig. S1A). Isolates from Clade D were also clearly β- were largely determined by only a few single nucleo- hemolytic, while isolates from Clades A and E had a tide polymorphisms (SNPs) in the gyrB, pheT, prkC, tight ring of β-hemolysis with a surrounding zone of and rpoB genes, with exception of lineages from α-hemolysis. Isolates from Clade B showed the most pygmy hippopotamuses (Choeropsis liberiensis; 831, limited hemolysis on SBA, with a diffuse ring of α-he- 832, and 837) and the second Amazon river dolphin molysis (Figs. 1 & S1A). Hemolytic activity increased isolate (Inia geoffrensis; QMA0141), which had multi- in agar stabs under more anaerobic conditions, con- ple SNPs in every gene fragment. sistent with reports by Facklam et al. (2005) (Fig. S1B). Isolates from North America fell into 4 of the 5 Isolate WS-6H, a Clade B isolate by rep-PCR, formed clades (A, B, D, and E), 3 of which consisted exclu- its own lineage with MLSA and was morphologically sively of isolates from this region (A, B, and D). Clade distinct from all other isolates (Fig. S1C). All tested E, the largest, contained isolates from diverse species isolates are suspected to be serotype I based on their and countries. Clade A contained isolates from tilapia ability to react with arginine dihydrolase (ADH) and in the USA and humans from Canada; Clade B, iso- ribose in the API 20 STREP system (Barnes et al. lates from fish in California; and Clade D, isolates 2003b), although caution should be ex ercised in con- from outbreaks in wild and farmed marine fish in cluding serotypes by this method (Barnes & Ellis Central America and the Caribbean. Clade C, the sec- 2003). Buoyant density as says indicated that the Cali- ond largest clade, consisted of isolates from marine fornia isolates displayed reduced capsulation, as they and freshwater fish from Israel and Australia. Several traveled furthest through the Percoll gradient (Figs. 1 smaller lineages of 1 or 2 isolates did not fit into the & S2). Isolate Eco86-17 from Costa Rica had a higher other major clades, likely a result of undersampling. density than the Caribbean isolates, indicating slightly Isolates WS-6H, QMA0188, and QMA0187 from a reduced capsulation, while LSU 96-525 from Clade A white sturgeon, rainbow trout, and striped snakehead had the lowest overall density, suggesting overex- Channa striata, respectively, branched separately pression of CPS. from the main cluster. Tilapia isolate ISET0901 and its attenuated derivative ISNO grouped together, as did the isolates from clown loaches Chromobotia macra- 3.3. Resistance to whole-blood killing canthus. Non-human mammalian isolates were gen- erally divergent from the other isolates, except for Establishing bacteremia is important in strepto- QMA0140, the original Amazon River dolphin type coccal pathogenesis, and resistance to phagocytic 60 Dis Aquat Org 141: 53–69, 2020

Fig. 1. Maximum likelihood tree of Streptococcus iniae isolates generated from MLSA. Isolates used in further phylogenetic analysis are in bold. Evolutionary history was inferred using the maximum likelihood method based on the Tamura 3-parameter model (Tamura 1992). The tree with the highest log likelihood (−7587.37) is shown. Percentage of trees in which the associated taxa clustered together is shown next to the branches. Initial trees for the heuristic search were obtained automatically by apply- ing neighbor-join and BioNJ algorithms to a matrix of pairwise distances estimated using the maximum composite likelihood ap- proach, and then selecting the topology with superior log-likelihood value. A discrete Gamma distribution was used to model evolutionary rate differences among sites (5 categories [+G, parameter = 0.0500]). The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. Evolutionary analyses were conducted in MEGA X (Kumar et al. 2018) Heckman et al.: Multilocus sequence analysis of Streptococcus iniae 61

Fig. 2. Maximum likelihood tree of Streptococcus sp. isolates generated from MLSA. Isolates used in further phylogenetic analysis are in bold. Evolutionary history was inferred using the maximum likelihood method based on the Tamura 3-parame- ter model (Tamura 1992). The tree with the highest log likelihood (−10418.63) is shown. Percentage of trees in which the associ- ated taxa clustered together is shown next to the branches. Initial trees for the heuristic search were obtained automatically by applying neighbor-join and BioNJ algorithms to a matrix of pairwise distances estimated using the maximum composite likeli- hood approach, and then selecting the topology with superior log likelihood value. A discrete Gamma distribution was used to model evolutionary rate differences among sites (5 categories (+G, parameter = 0.0598)). The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. Evolutionary analyses were conducted in MEGA X (Kumar et al. 2018) killing is known to correlate with differences in vir- blood from 3 relevant fish hosts. Survival differed ulence (Zlotkin et al. 2003, Locke et al. 2007b). As between isolates and by host, but all isolates de- such, the ability of isolates to survive exposure to monstrated some resistance to killing (Fig. 3). Over- whole-blood killing factors was investigated in fresh all, survival was highest in rainbow trout blood 62 Dis Aquat Org 141: 53–69, 2020

generally showed lower survival than isolates in Clades A, D, and E, but variability within groups and across blood types precluded statistical compar- isons. Differences in survival were not due to differ- ences in isolate growth rate or incubation tempera- ture (Figs. S3 & S4).

3.4. Cytotoxic effects of S. iniae on tilapia endothelial cells

The ability of isolates to damage or kill host cells was assessed by colorimetric assay measuring re lease of LDH from monolayers of TmBs cells. Three main patterns of virulence were observed, and temperature was positively correlated with isolate cytotoxicity (Fig. 4). Most isolates induced a higher percentage of LDH release at 30°C compared to 20°C. For the re- maining isolates there was no statistically significant difference between temperatures (p > 0.05). At 30°C, the marine Clade D showed significantly higher cyto- toxic effects than isolates in Clades A, B, or E (p < 0.05). Clade B isolates consistently did not cause re- lease of LDH above background levels. Members of Clades A and E induced similar trends in cytotoxicity, and there was no statistically significant difference between the 2 groups at either temperature.

15 X 20°C

30°C

10

YZ

5 % LDH released

Fig. 3. Percent isolate survival after incubation for 1 h in whole heparinized blood from 3 fish species incubated at biologically relevant temperatures. Species from top to bot- 0 7 3 tom: Nile tilapia O. niloticus, white sturgeon Acipenser trans- B8 4- montanus, rainbow trout Oncorhynchus mykiss. The experi- 94-03496-525 F15- WS-6B Eco86-1 ment was performed twice in biological duplicate using the K08-409H PB 03-62B same groups of fish. Error bars: SE. Color and patterning LSU 01-105LSU 10-070LSU LSU LSU 94-036 reflect isolate phylogroup: Clade A: solid red; Clade B: Fig. 4. Percent cytotoxicity as measured in terms of lactate orange with vertical stripes; Clade D: teal with dots; Clade E: dehydrogenase (LDH) release from Oreochromis mossam- blue with horizontal stripes bicus bulbus arteriosus cell lines compared to lysed cells serving as a positive control. Cells were incubated with bacteria for 3 h, washed, and incubated for 24 h at 20 or where, in addition to persisting, most isolates were 30°C. Experiments were carried out in technical triplicate also able to replicate (Facklam et al. 2005). Survival and repeated 4 times. Colors denote clade: Clade A: red; Clade B: orange; Clade D: teal; Clade E: blue. Letters ranged from 37−292%, compared to 21−117% in denote statistical significance at 30°C (p < 0.001). Signifi- tilapia blood and 17−133% in sturgeon blood. There cance was determined using 1-way ANOVA with post hoc were some trends related to phylogroup, as Clade B Tukey HSD; error bars: SE Heckman et al.: Multilocus sequence analysis of Streptococcus iniae 63

3.5. Intragastric challenge cal signs and swift autolysis due to high water temper- atures and cannibalism limited collection of moribund In vitro assays lack the immunological complexities or freshly dead fish. In the collected fish, the most of a living organism, so an in vivo challenge using the common gross lesion was bilateral or unilateral ex- intragastric method was used to test the virulence of ophthalmia. Other gross lesions included re duced representative isolates. Tilapia challenged by Clade E mucus and congestion of dorsal and caudal fin bases isolate LSU 10-070 had significantly lower survival (Fig. 6). Isolates LSU 10-070 and LSU 94-034 were re- (56.25%) than the control fish challenged with sterile covered from the brain tissue of freshly dead fish. The PBS (p < 0.005). There were some mortalities in tanks tissues from tilapia challenged with LSU 94-036 and challenged with each of the tested isolates from Clade K08-409H were heavily cannibalized before collection A and one isolate from Clade D, but the survival and were not sampled for re-isolation of bacteria. No curves were not significantly different from the con- histological changes indicating infection were noted trols (Fig. 5). In most cases, fish were observed eating in the controls. Fig. 7 depicts histopathologic changes and swimming normally before dying rapidly, al- in moribund and surviving fish following intragastric though a few fish did display disoriented swimming challenge with isolate LSU 10-070. Examination of a with or without additional clinical signs. Lack of clini- moribund fish revealed extensive severe granuloma- tous meningo encephalitis and ventri culitis dominated Control 100 by macrophages laden with Gram-positive cocci, LSU 94-034 LSU 96-525 neutro phils, and scattered lymphocytes and plasma LSU 94-036 cells. Similar severe inflammatory changes affected F15-4-3 the intraocular humors, uveal tract, choroid rete, and WS6B periocular connective tissues. Edema, foci of necrosis, 50 ARK PB 03-62B LSU 01-105 and less intense inflammatory infiltrates were wide-

% Survival% LSU 10-070 spread in cranial adipose, skeletal muscle, and associ- Eco86-17 ated interstitial areas. Some surviving fish were free K08-409H of microscopic changes, while the meninges and cra- 0 B8 0 5 10 15 20 nial adipose of others contained small organizing Days post inoculation granulomas characterized by central regions of de - generate macrophages, surrounded by mantles of Fig. 5. Survival curve for Nile tilapia fingerlings intragas- lymphocytes, with scattered plasma cells and coarse trically challenged with North American isolates for 21 d. eosinophilic granulocytes. Bacteria were not ob - In total, 16 fish were challenged per treatment. Star: statis- served in routine H&E sections or with tissue Gram tical significance as determined by log-rank (Mantel-Cox) stains. Visceral organs were not affected in any of the and Gehan-Breslow-Wilcoxon tests. Color reflects isolate phylogroup: Clade A: red; Clade B: orange; Clade D: teal; fish examined. No histological changes indicating in- Clade E: blue fection were noted in the controls

4. DISCUSSION

During the early emergence of Strep- tococcus iniae infections in fish and humans, efforts to define an effective A identification and typing workflow sys- tem were complicated by similarities of S. iniae to other strep to coccal species and by the lack of commercial bacterial identification systems that included S. iniae in their databases (Lau et al. BC2003, Facklam et al. 2005, Roach et al. 2006). Genetic techniques such as 16S rRNA sequencing (Berridge et al. 1998, Fig. 6. Tilapia collected and euthanized during the intragastric challenge with Streptococcus iniae. (A) Congestion along the dorsal fin; (B) severe exoph- Zhou et al. 2011), PFGE (Weinstein et thalmia with hyperemia; (C) gradient of exophthalmia in challenge survivors al. 1997), and Rep-PCR (Chou et al. 64 Dis Aquat Org 141: 53–69, 2020

AB

CD

EF

Fig. 7. Histologic sections of Streptococcus iniae isolate LSU 10-070 (Clade E) induced lesions in Nile tilapia following intragas- tric challenge. (A) Brain from moribund fish with granulomatous meningitis and inflammatory infiltrate within the third ventri- cle (arrows) below the optic tectum (H&E; scale bar = 500 μm). (B) Eye with heavy infiltration of the iris (I) and posterior chamber (PC) by inflammatory cell infiltrate (H&E; scale bar = 200 μm). (C) High magnification image of inflammatory infiltrate within the third ventricle containing macrophages with cytoplasm distended by bacterial cocci (H&E; scale bar = 10 μm). (D) Optic tec- tum of the brain infiltrated by macrophages laden with Gram-positive cocci (Brown & Hopps; scale bar = 10 μm). (E) Developing meningeal granuloma in a 21 d post-challenge survivor. A central focus of degenerate macrophages is surrounded by a broad mantle of lymphocytes (H&E; scale bar = 100 μm). (F) Higher magnification image of developing granuloma bordered by lym- phocytes and eosinophilic granular cells. The absence of bacteria in these chronic lesions suggests resolution on the infection (H&E; scale bar = 20 μm Heckman et al.: Multilocus sequence analysis of Streptococcus iniae 65

2014) were developed to resolve this issue, but each from tilapia in the USA and their similarities may be approach has distinct advantages and limitations in related to a closer genetic relationship and host type. their replicability and ability to differentiate closely Clade C contained isolates from fish cultured in related species and strains (Li et al. 2009). An alterna- Israel and Australia. Unfortunately, isolates of Clade tive to these methods is MLSA. MLSA characterizes C were unavailable at the time of study for inclusion bacterial isolates by using the concatenated sequence into the phenotypic and challenge ex periments and fragments of multiple ‘house keeping’ genes to deter- were only utilized in the phylogenetic analyses. mine phylogenetic relationships. It generates precise, Many of the S. iniae reference genomes used were reproducible data that can be compared to sequence generated by Silayeva et al. (2020) in a study investi- data from other organisms through database query gating the role of mutator strains in S. iniae epidemio - and is useful for analysis of strains recovered around logy and evolution. The authors found 6 major the globe and identification of unknown isolates. The clades, one lineage with 2 strains, and 3 lineages molecular markers chosen for MLSA should have low with a single strain. Our analysis incorporated iso- heterogeneity within a species or genotype but pro- lates from each of these lineages and produced a vide the highest amount of separation between spe- maximum likelihood tree with some differences from cies or genotypes (Martens et al. 2008). The gene the one generated by Silayeva et al. (2020) using fragments chosen were largely homogenous be tween non-recombinant core genome SNPs. Comparably, S. iniae isolates but were still able to differentiate some clades were geographically diverse (C, E), groups similarly to established methods (Fig. 1). Fur- while others displayed some degree of endemism (A, thermore, this MLSA scheme was able to incorporate B). Both trees maintained the grouping of human iso- and differentiate divergent Streptococcus spp., in- lates from Canada with tilapia cases from the USA cluding S. pyogenes, which is an almost exclusively (Clades A vs. E2 in Silayeva et al. 2020), the singular human pathogen (Bessen 2009), giving this scheme lineage of the striped snakehead isolate QMA0187 broader applicability than typical species-specific from Thailand, and the second Amazon river dolphin MLSA formats. isolate QMA0141 from the USA (Fig. 1). The remain- The 11 North American isolates used in the study ing lineages, however, were either condensed into were previously genotyped by an established Rep- Clade A or split into smaller lineages with only one or PCR method that yielded 4 major clusters (T. I. Heck- 2 members. Inclusion of an additional mutation man et al. unpubl. data). These clusters were largely repair gene into the analysis could potentially re - maintained in the MLSA maximum likelihood phylo - solve these differences, but the reduced heterogene- geny, where there were 5 major clades consisting of ity in the housekeeping genes compared to the 3 or more isolates in the intraspecific comparison tree whole genome could still be beneficial. ‘Mutator’ (Fig. 1). Clade D, containing marine isolates from strains like QMA0141 can have enormous branch Costa Rica and the Caribbean, maintained the same lengths in whole genome analysis due to deleterious isolates from the Rep-PCR, while Clades A, B, and E changes in their mutation repair genes, potentially were expanded by additional isolates. White stur- distorting evolutionary evaluations. QMA0141 had geon isolate WS-6H, co-infected with Acipenserid the highest mutation rate among the isolates investi- herpesvirus 2, had previously grouped with the other gated by Silayeva et al. (2020), although the pygmy sturgeon isolates from California by Rep-PCR, but hippo isolates used in this study had an even larger formed its own lineage in the MLSA generated max- number of SNPs in the mutS fragment as well as the imum likelihood tree (Fig. 1). This distinction sup- other fragments investigated. This lends support to ports the use of MLSA over Rep-PCR, as WS-6H has the hypothesized role of mutators in the jump to relevant differences from the other sturgeon isolates. atypical hosts. It would be interesting to further char- The colony morphology of WS-6H differed from acterize these isolates using whole genome and Clade B isolates (Fig. S1), had a lower density indica- mutation rate analyses. tive of more capsulation, and better resisted killing in The importance of host type in streptococcal infec- whole blood (data not shown). Previous work has tion was demonstrated in the whole-blood survival as- also demonstrated that WS-6H has a different bio- say. The 11 representative North American isolates chemical profile than other sturgeon isolates (Piere - showed varying degrees of survival in blood from 3 zan et al. 2020). The remainder of the isolates fell into commercially valuable fish species at biologically rel- 2 different clades but showed similar patterns of vir- evant temperatures. All isolates survived to some de- ulence and no significant differences in the in vitro gree in all blood types, but survival and replication assays. These clades both contained isolates cultured was conspicuously higher in rainbow trout blood 66 Dis Aquat Org 141: 53–69, 2020

(Fig. 3). Additionally, although isolates did not display Clade B showed no cytotoxic effects. One of the iso- the exact same relative patterns across host types, lates representing this clade was from sturgeon, a there were still trends related to phylogroup. Isolates cold- or temperate-water host. However, decreasing from the marine Clade D largely showed higher sur- the assay temperature did not result in increased vival, while isolates in Clade B were generally more LDH release. In cell cultures challenged with Clade susceptible to killing. Isolates in Clades A and E B isolates, bacterial numbers approximated those of shared similar levels of survival. Because blood from the other clades, but produced no microscopic dam- same fish was not used across trials, intragroup vari- age to the cell monolayer. ability was significant in this assay, and the isolate Results of the in vitro assays suggested that mar- demonstrating the highest level of survival also varied ine Clade D isolates would be the most virulent. between hosts. This speaks to the complicated nature Despite this, isolates from the intermediate in vitro of host−pathogen interactions, where virulence de- phenotype (Clades A and E) caused the highest pends on multiple factors related to the microbe and mortality in the tilapia challenge (Fig. 5). This dis- its target species. It would be interesting to repeat the crepancy may be due to differences in the func- in vivo challenge in trout and sturgeon to determine tional capabilities of an isolated cell type versus a whether mortality patterns differed from that in fully immunocompetent animal. However, it is im - tilapia. While the complete array of pathogenic mech- portant to note that these isolates were originally anisms used by S. iniae to initiate infection and pro- collected from tilapia, suggesting certain isolates duce disease remain un clear, establishment of bac- may be better adapted to cause disease in specific teremia is crucial for the dissemination of bacteria hosts. Additionally, most of the isolates that caused from local sites of infection to target organs (Fuller et mortality had been recently passaged through al. 2001, Zlotkin et al. 2003, Locke et al. 2007b). Dif- tilapia, while isolates from Clades B and D are of ferences in survival between isolates may be related unknown passage number. Efforts to standardize to proposed virulence factors affecting resistance to passage number were not possible due to the innate immune clearance such as expression of a cell inability of isolates in Clade B to fulfill Koch’s pos- surface Fc binding factor (Barnes et al. 2003b), poly- tulates when originally cultured (Soto et al. 2017a). saccharide de acetylase (Milani et al. 2010), M-like Regardless, the challenge did demonstrate differ- protein (Locke et al. 2008), or extent of capsulation ences in virulence, and validated effectiveness of (Barnes et al. 2003a, Buchanan et al. 2005, Locke et al. the intragastric challenge model for S. iniae. 2007b, Lowe et al. 2007). In addition to investigations of S. iniae associated In addition to virulence factors involved in resist- disease pathogenesis, isolates are used to model ance to the host immune response, S. iniae and re- other streptococcal species pathogenic to humans lated streptococcal species express proteins that (Miller & Neely 2004), test dietary supplements damage tissues directly or via the host inflammatory such as immunostimulants, probiotics and prebi- reaction (Fuller et al. 2001, 2002, Bromage & Owens otics for aquaculture (Li & Gatlin 2004, Yildirim- 2002, Miller & Neely 2005, Locke et al. 2007a, Bolotin Aksoy et al. 2007), and to study fish immune et al. 2007). Cytotoxicity of the different S. iniae iso- responses (Costa et al. 2012, Harvie et al. 2013). lates was investigated in a tilapia endothelial cell line Intraperitoneal injection is the most commonly by a colorimetric assay for LDH release. Cellular used infection route for in vivo S. iniae studies, damage was significantly higher at 30°C compared although it bypasses host protective barriers and to 20°C (Fig. 4), supporting previous observations may cause mortalities (up to 100%) inconsistent linking increased temperature to outbreaks of strep- with natural infections (Perera et al. 1997, Locke et tococcosis (Bromage & Owens 2009). The trend may al. 2007b). Although transmission dy na mics of S. be related to higher bacterial replication rates at iniae infection remain unresolved, evidence sup- 30°C (Fig. S3) or to changes in virulence factor ex- ports transmission through the consumption of pression, which could enhance tissue invasion and contaminated tissue, including cannibalism of subsequent cellular injury. However, further experi- moribund or dead fish by tank mates (Shoemaker mentation would be necessary to determine specific et al. 2000, Bromage & Owens 2002). As a result, causes. At 30°C, the isolates displayed cytotoxicity gavage challenge may more closely approximate patterns that mirrored their survival in whole blood, natural outbreaks of disease. Direct delivery of S. but with less intragroup variability. Isolates in marine iniae by gavage was used successfully in a limited Clade D exhibited the greatest cytotoxicity, isolates study by Perera et al. (1997) and its effectiveness in Clades A and E were intermediate, and those in further validated in this investigation. Intragastric Heckman et al.: Multilocus sequence analysis of Streptococcus iniae 67

challenge proved to be a useful alternative and Acknowledgements. This study was funded by the Floyd straightforward method for the delivery of a stan- and Mary Schwall Fellowship in Medical Research and the USDA National Institute of Food and Agriculture, Animal dardized bacterial dose by a biologically relevant Health project CALV-AH-377. We thank the members of the exposure route. Overall, disease progression was Aquatic Animal Health Laboratory for their support and the consistent with previous reports (Perera et al. Microbiology Graduate Group at UC Davis. We especially 1997, 1998, Bromage & Owens 2002). Mortalities thank Dr. Thomas deMaar at the Gladys Porter Zoo in Brownsville TX for collecting the pygmy hippo isolates, and began within 72 h of gavage challenge and contin- Dr. Patricia Lynn Shewmaker from the CDC for providing ued over the 21 d trial, reaching a ~45% the human S. iniae isolates. Additional thanks to Susan Yun, maximum (Fig. 5). Exophthalmia was the most Fernanda de A. Sebastião, and Eric Pomaranski for guiding commonly ob served gross lesion (Fig. 6), although and training T.I.H.. M.J.G. acknowledges the assistance of most fish died without clinical signs, similar to the Cynthia Ware in initial validation of MLSA primers. ‘acute’ form of disease reported by Bromage & Owens (2002). Histologic evaluation revealed typi- LITERATURE CITED cal lesions of meningoencephalitis, panophthalmi- tis, and cellulitis with abundant intracellular cocci Agnew W, Barnes AC (2007) Streptococcus iniae: an aquatic pathogen of global veterinary significance and a chal- in moribund specimens. Some survivors were free lenging candidate for reliable vaccination. Vet Microbiol of microscopic changes, suggesting infection had 122: 1−15 never occurred. In others, early granuloma forma- Bachrach G, Zlotkin A, Hurvitz A, Evans DL, Eldar A (2001) tion in the absence of bacteria was associated with Recovery of Streptococcus iniae from diseased fish previ- a shift in the inflammatory reaction from predomi- ously vaccinated with a streptococcus vaccine. Appl Environ Microbiol 67:3756−3758 nantly macrophages to lymphocytes and indicates Barnes AC, Ellis AE (2003) Variation in arginine dihydrolase resolution of the infection. The presence of granu- activity in Streptococcus iniae may be an artefact of the lomas is consistent with descriptions of chronic S. assay. Bull Eur Assoc Fish Pathol 23: 163−166 iniae induced lesions and suggests elimination of Barnes AC, Young FM, Horne MT, Ellis AE (2003a) Strepto- coccus iniae: serological differences, presence of capsule the infection (Perera et al. 1998, Suanyuk et al. and resistance to immune serum killing. Dis Aquat Org 2010). 53: 241−247 MLSA characterization of a diverse panel of S. Barnes AC, Horne MT, Ellis AE (2003b) Streptococcus iniae iniae isolates from wild and farmed fish species de - expresses a cell surface non-immune trout immunoglob- ulin-binding factor when grown in normal trout serum. monstrated a genetic basis for strain phenotype, and Fish Shellfish Immunol 15: 425−431 further validation of the protocol as a predictor of vir- Berridge BR, Fuller JD, de Azavedo J, Low DE, Bercovier H, ulence is warranted. The MLSA scheme proved Frelier PF (1998) Development of specific nested oligo- robust, and readily discriminated between Strepto - nucleotide PCR primers for the Streptococcus iniae 16S- coccus spp. relevant to aquaculture and human 23S ribosomal DNA intergenic spacer. J Clin Microbiol 36: 2778−2781 health. Increasing the number of disparate isolates Berzak R, Scheinin A, Davidovich N, Regev Y, Diga R, Tch- included in analyses, as well as the inclusion of addi- ernov D, Morick D (2019) Prevalence of nervous necrosis tional nucleotide and genome sequences in public virus (NNV) and Streptococcus species in wild marine databases as they become available will increase the fish and crustaceans from the Levantine Basin, Mediter- ranean Sea. Dis Aquat Org 133: 7−17 applicability of this genotyping method and expand Bessen DE (2009) Population biology of the human re - understanding of S. iniae genetic diversity. In addi- stricted pathogen, Streptococcus pyogenes. Infect Genet tion, intragastric challenge proved to be an effective Evol 9: 581−593 and biologically relevant alternative to other chal- Bolotin S, Fuller JD, Bast DJ, de Azavedo JCS (2007) The two-component system sivS/R regulates virulence in lenge methods. Results from this study will be used Streptococcus iniae. FEMS Immunol Med Microbiol 51: to inform the design for future investigations by pro- 547−554 viding a diverse group of genetically profiled isolates Bromage ES, Owens L (2002) Infection of barramundi Lates of differing virulence. calcarifer with Streptococcus iniae: effects of different routes of exposure. Dis Aquat Org 52:199−205 Bromage E, Owens L (2009) Environmental factors affecting Data availability. The Genbank accession numbers for the the susceptibility of barramundi to Streptococcus iniae. MLSA nucleotide sequences are MN 955874− MN955922, Aquaculture 290: 224−228 MN955923−MN955971, MN 95 59 72− MN956020, MN956021− Buchanan JT, Stannard JA, Lauth X, Ostland VE, Powell MN956069, MN 95 60 70 − MN956118, MN9561 19− MN956167, HC, Westerman ME, Nizet V (2005) Streptococcus iniae MN 95 61 68− MN 956216, MN956217− MN95 62 65, and MN 95 phosphoglucomutase is a virulence factor and a target 62 66 −MN956314 . The other raw data supporting the conclu- for vaccine development. Infect Immun 73: 6935−6944 sions of this paper will be made available by the authors, Chou L, Griffin MJ, Fraites T, Ware C and others (2014) Phe- without undue reservation, to any qualified researcher. notypic and genotypic heterogeneity among Streptococ- 68 Dis Aquat Org 141: 53–69, 2020

cus iniae isolates recovered from cultured and wild fish Koh TH, Sng LH, Yuen SM, Thomas CK, Tan PL, Tan SH, in North America, Central America and the Caribbean Wong NS (2009) Streptococcal cellulitis following prepa- Islands. J Aquat Anim Health 26: 263−271 ration of fresh raw seafood. Zoonoses Public Health 56: Colorni A, Diamant A, Eldar A, Kvitt H, Zlotkin A (2002) 206−208 Streptococcus iniae infections in Red Sea cage-cultured Kumar S, Stecher G, Li M, Knyaz C, Tamura K (2018) MEGA and wild fishes. Dis Aquat Org 49: 165−170 X: molecular evolutionary genetics analysis across com- Costa G, Danz H, Kataria P, Bromage E (2012) A holistic puting platforms. Mol Biol Evol 35: 1547−1549 view of the dynamisms of teleost IgM: a case study of Lau SKP, Woo PCY, Tse H, Leung KW, Wong SSY, Yuen KY Streptococcus iniae vaccinated rainbow trout (Onco- (2003) Invasive Streptococcus iniae infections outside rhynchus mykiss). Dev Comp Immunol 36:298−305 North America. J Clin Microbiol 41: 1004−1009 Creeper JH, Buller NB (2006) An outbreak of Streptococcus Lau SKP, Woo PCY, Luk W, Fung AMY and others (2006) iniae in barramundi (Lates calcarifera) in freshwater Clinical isolates of Streptococcus iniae from Asia are cage culture. Aust Vet J 84: 408−411 more mucoid and β-hemolytic than those from North Eldar A, Bejerano Y, Bercovier H (1994) Streptococcus shiloi America. Diagn Microbiol Infect Dis 54: 177−181 and Streptococcus difficile: two new streptococcal spe- Lewis DH, Marks JE (1985) Microcultures of Sarotherodon cies causing a meningoencephalitis in fish. Curr Micro- mossambicus (Peters) cells: their use in detecting fish biol 28: 139−143 viruses. J Fish Dis 8: 477−478 Eldar A, Perl S, Frelier PF, Bercovier H (1999) Red drum Sci- Li P, Gatlin DM (2004) Dietary brewers yeast and the prebi- aenops ocellatus mortalities associated with Streptococ- otic GrobioticTM AE influence growth performance, cus iniae infection. Dis Aquat Org 36:121−127 immune responses and resistance of hybrid striped bass Erfanmanesh A, Soltani M, Pirali E, Mohammadian S, Taher- (Morone chrysops × M. saxatilis) to Streptococcus iniae imirghaed A (2012) Genetic characterization of Strepto- infection. Aquaculture 231:445−456 coccus iniae in diseased farmed rainbow trout (Oncho- Li W, Raoult D, Fournier PE (2009) Bacterial strain typing in rhynchus mykiss) in Iran. Sci World J 2012: 594073 the genomic era. FEMS Microbiol Rev 33:892−916 Eyngor M, Tekoah Y, Shapira R, Hurvitz A, Zlotkin A, Lublin Liu Y, Lai Q, Shao Z (2017) A multilocus sequence analysis A, Eldar A (2008) Emergence of novel Streptococcus iniae scheme for phylogeny of Thioclava bacteria and pro- exopolysaccharide-producing strains following vaccina- posal of two novel species. Front Microbiol 8: 1321 tion with nonproducing strains. Appl Environ Microbiol Locke JB, Colvin KM, Varki N, Vicknair MR, Nizet V, 74: 6892−6897 Buchanan JT (2007a) Streptococcus iniae β-hemolysin Facklam R, Elliott J, Shewmaker L, Reingold A (2005) Iden- streptolysin S is a virulence factor in fish infection. Dis tification and characterization of sporadic isolates of Aquat Org 76: 17−26 Streptococcus iniae isolated from humans. J Clin Micro- Locke JB, Colvin KM, Datta AK, Patel SK and others (2007b) biol 43: 933−937 Streptococcus iniae capsule impairs phagocytic clear- FAO (2018) The state of the world fisheries and aquaculture ance and contributes to virulence in fish. J Bacteriol 189: 2018 — meeting the sustainable development goals. 1279−1287 FAO, Rome Locke JB, Aziz RK, Vicknair MR, Nizet V, Buchanan JT Ferguson HW, St John VS, Roach CJ, Willoughby S, Parker (2008) Streptococcus iniae M-like protein contributes to C, Ryan R (2000) Caribbean reef fish mortality associated virulence in fish and is a target for live attenuated vac- with Streptococcus iniae. Vet Rec 147: 662−664 cine development. PLOS ONE 3: e2824 Fuller JD, Bast DJ, Nizet V, Low DE, De Azavedo JCS (2001) Lowe BA, Miller JD, Neely MN (2007) Analysis of the poly- Streptococcus iniae virulence is associated with a distinct saccharide capsule of the systemic pathogen Streptococ- genetic profile. Infect Immun 69: 1994−2000 cus iniae and its implications in virulence. Infect Immun Fuller JD, Camus AC, Duncan CL, Nizet V and others (2002) 75: 1255−1264 Identification of a streptolysin S-associated gene cluster Lucas JS, Southgate PC, Tucker CS (eds) (2012) Aquacul- and its role in the pathogenesis of Streptococcus iniae ture: farming aquatic animals and plants, 3rd edn. Black- disease. Infect Immun 70: 5730−5739 well Publishing, Southgate Gao XY, Zhi XY, Li HW, Klenk HP, Li WJ (2014) Compara- Macheras E, Roux AL, Bastian S, Leão SC and others (2011) tive genomics of the bacterial genus Streptococcus illu- Multilocus sequence analysis and rpoB sequencing of minates evolutionary implications of species groups. Mycobacterium abscessus (sensu lato) strains. J Clin PLOS ONE 9: e101229 Microbiol 49: 491−499 Glaeser SP, Kämpfer P (2015) Multilocus sequence analysis Martens M, Dawyndt P, Coopman R, Gillis M, De Vos P, (MLSA) in prokaryotic taxonomy. Syst Appl Microbiol Willems A (2008) Advantages of multilocus sequence 38: 237−245 analysis for taxonomic studies: a case study using 10 Håkansson S, Holm SE, Wagner M (1987) Density profile of housekeeping genes in the genus Ensifer (including for- group B streptococci, type III, and its possible relation to mer Sinorhizobium). Int J Syst Evol Microbiol 58:200−214 enhanced virulence. J Clin Microbiol 25: 714−718 Milani CJE, Aziz RK, Locke JB, Dahesh S, Nizet V, Buchanan Harvie EA, Green JM, Neely MN, Huttenlocher A (2013) JT (2010) The novel polysaccharide deacetylase homo- Innate immune response to Streptococcus iniae infection logue Pdi contributes to virulence of the aquatic pathogen in zebrafish larvae. Infect Immun 81: 110−121 Streptococcus iniae. Microbiology 156: 543−554 Keirstead ND, Brake JW, Griffin MJ, Halliday-Simmonds I, Millard CM, Baiano JCF, Chan C, Yuen B and others (2012) Thrall MA, Soto E (2014) Fatal septicemia caused by the Evolution of the capsular operon of Streptococcus iniae zoonotic bacterium Streptococcus iniae during an out- in response to vaccination. Appl Environ Microbiol 78: break in Caribbean reef fish. Vet Pathol 51:1035−1041 8219−8226 Koh TH, Kurup A, Chen J (2004) Streptococcus iniae discitis Miller JD, Neely MN (2004) Zebrafish as a model host for in Singapore. Emerg Infect Dis 10:1694−1696 streptococcal pathogenesis. Acta Trop 91: 53−68 Heckman et al.: Multilocus sequence analysis of Streptococcus iniae 69

Miller JD, Neely MN (2005) Large-scale screen highlights rate dynamics with adaptation to novel immunological the importance of capsule for virulence in the zoonotic landscapes. Infect Genet Evol 85:104435 pathogen Streptococcus iniae. Infect Immun 73: Soto E, Richey C, Stevens B, Yun S and others (2017a) Co- 921−934 infection of Acipenserid herpesvirus 2 (AciHV-2) and Ortega C, García I, Irgang R, Fajardo R, Tapia-Cammas D, Streptococcus iniae in cultured white sturgeon Acipen - Acosta J, Avendaño-Herrera R (2018) First identification ser transmontanus. Dis Aquat Org 124: 11−20 and characterization of Streptococcus iniae obtained Soto E, Yun S, Lewis J, Kearney MT, Hansen J (2017b) Inter- from tilapia (Oreochromis aureus) farmed in Mexico. action of Francisella noatunensis subsp. orientalis with J Fish Dis 41: 773−782 Oreochromis mossambicus bulbus arteriosus cell line. Patchanee P, Chokboonmongkol C, Zessin KH, Alter T, Por- Microb Pathog 105: 326−333 naem S, Chokesajjawatee N (2012) Comparison of multi- Suanyuk N, Tanmark N, Tantikitti C, Supamattaya K and locus sequence typing (MLST) and repetitive sequence- others (2010) Streptococcus iniae infection in cultured based PCR (rep-PCR) fingerprinting for differentiation of Asian sea bass (Lates calcarifer) and red tilapia (Ore- Campylobacter jejuni isolated from broiler in Chiang ochromis sp.) in southern Thailand. Songklanakarin J Sci Mai, Thailand. J Microbiol Biotechnol 22: 1467−1470 Technol 32: 341−348 Peacock SJ, de Silva GDI, Justice A, Cowland A, Moore CE, Sun JR, Yan JC, Yeh CY, Lee SY, Lu JJ (2007) Invasive infec- Winearls CG, Day NPJ (2002) Comparison of multilocus tion with Streptococcus iniae in Taiwan. J Med Microbiol sequence typing and pulsed-field gel electrophoresis as 56: 1246−1249 tools for typing Staphylococcus aureus isolates in a micro- Tamura K (1992) Estimation of the number of nucleotide epidemiological setting. J Clin Microbiol 40: 3764−3770 substitutions when there are strong transition-transver- Perera RP, Johnson SK, Collins MD, Lewis DH (1994) Strep- sion and G+C-content biases. Mol Biol Evol 9: 678−687 tococcus iniae associated with mortality of Tilapia nilot- United Nations (2019) World population prospects 2019: ica × T. aurea hybrids. J Aquat Anim Health 6: 335−340 data booklet. United Nations Department of Economic Perera RP, Johnson SK, Lewis DH (1997) Epizootiological and Social Affairs, Population Division, New York, NY aspects of Streptococcus iniae affecting tilapia in Texas. Weinstein MR, Litt M, Kertesz DA, Wyper P and others Aquaculture 152: 25−33 (1997) Invasive infections due to a fish pathogen, Strep - Perera RP, Fiske RA, Johnson SK (1998) Histopathology of to coccus iniae. N Engl J Med 337: 589−594 hybrid tilapias infected with a biotype of Streptococcus Whatmore AM, Koylass MS, Muchowski J, Edwards-Small- iniae. J Aquat Anim Health 10: 294−299 bone J, Gopaul KK, Perrett LL (2016) Extended multi- Pierezan F, Khalid S, Heckman TI, Ang J, Byrne BA, Soto E locus sequence analysis to describe the global popula- (2020) Outbreaks of severe myositis in cultured white tion structure of the genus Brucella: phylogeography and sturgeon (Acipenser transmontanus L.) associated with relationship to biovars. Front Microbiol 7:2049 Streptococcus iniae. J Fish Dis 43: 485−490 Woo PTK, Cipriano RC (eds) (2017) Fish viruses and bacte- Roach JCM, Levett PN, Lavoie MC (2006) Identification of ria: pathobiology and protection. CAB International, Streptococcus iniae by commercial bacterial identifica- Boston, MA tion systems. J Microbiol Methods 67:20−26 Yildirim-Aksoy M, Lim C, Davis DA, Shelby R, Klesius PH Russo R, Mitchell H, Yanong RPE (2006) Characterization of (2007) Influence of dietary lipid sources on the growth Streptococcus iniae isolated from ornamental cyprinid performance, immune response and resistance of Nile fishes and development of challenge models. Aquacul- tilapia, Oreochromis niloticus, to Streptococcus iniae ture 256: 105−110 challenge. J Appl Aquacult 19: 29−49 Shoemaker CA, Evans JJ, Klesius PH (2000) Density and Zhou SM, Fan Y, Zhu XQ, Xie MQ, Li AX (2011) Rapid iden- dose: factors affecting mortality of Streptococcus iniae tification of Streptococcus iniae by specific PCR assay uti- infected tilapia (Oreochromis niloticus). Aquaculture lizing genetic markers in ITS rDNA. J Fish Dis 34:265−271 188: 229−235 Zlotkin A, Hershko H, Eldar A (1998) Possible transmission Shoemaker CA, Klesius PH, Evans JJ (2001) Prevalence of of Streptococcus iniae from wild fish to cultured marine Streptococcus iniae in tilapia, hybrid striped bass, and fish. Appl Environ Microbiol 64: 4065−4067 channel catfish on commercial fish farms in the United Zlotkin A, Chilmonczyk S, Eyngor M, Hurvitz A, Ghittino C, States. Am J Vet Res 62: 174−177 Eldar A (2003) Trojan horse effect: phagocyte-mediated Silayeva O, Engelstädter J, Barnes AC (2020) Evolutionary Streptococcus iniae infection of fish. Infect Immun 71: epidemiology of Streptococcus iniae: linking mutation 2318−2325

Editorial responsibility: David Verner-Jeffreys, Submitted: May 28, 2020; Accepted: July 30, 2020 Weymouth, UK Proofs received from author(s): September 7, 2020