The gfc operon is involved in the formation of the O antigen capsule in Aeromonas hydrophila and contributes to virulence in channel catfish

Aquaculture 512 (2019) 734334

Contents lists available at ScienceDirect

Aquaculture journal homepage: www.elsevier.com/locate/aquaculture

The gfc operon is involved in the formation of the O antigen capsule in Aeromonas hydrophila and contributes to virulence in channel catfish

T

Charles M. Thurlowa,1, Mohammad J. Hossaina,1,2, Dawei Suna,b,1, Priscilla Bargera,c, Luke Fosheec, Benjamin H. Beckd, Joseph C. Newtonc, Jeffery S. Terhuneb, Mark A. Sapere, ⁎ Mark R. Lilesa, a

Department of Biological Sciences, Auburn University, Auburn, AL 36849, USA School of Fisheries, Aquaculture and Aquatic Sciences, Auburn University, AL 36849, USA c Department of Pathobiology, Auburn University, Auburn, AL 36849, USA d United States Department of Agriculture, Agricultural Research Service, Aquatic Animal Health Research Unit, Auburn, AL 36832, USA e Department of Biological Chemistry, University of Michigan, Ann Arbor, MI 48109-5606, USA b

ARTICLE INFO

ABSTRACT

Keywords: Aeromonas hydrophila Pathogenesis O-antigen Capsule Fish LPS

A hypervirulent A. hydrophila (vAh) pathotype has been identified as the etiologic agent involved in disease outbreaks in farmed carp and catfish species in China and the Southeastern United States, respectively. To assess the role of the LPS O-antigen in vAh virulence, the O-antigen ligase (waaL) and O-antigen polymerase (wzy) genes were deleted. While neither waaL nor wzy were found to be required for vAh virulence, a waaL mutant was found to have a polar effect on an adjacent group 4 capsule (gfc) genetic operon that was predicted to play a role in capsule assembly. Mutations in the gfc operon attenuated vAh virulence in channel catfish, particularly the mutant lacking gfcD that is predicted to encode an outer membrane protein required for capsular polysaccharide export. Furthermore, the vAh gfcD mutant was found to lack significant biofilm-forming capacity or buoyancy compared to wild-type vAh and induced an adaptive immune response that protected catfish from vAh challenge. This study indicates the importance of the capsular polysaccharide assembly process in the pathogenesis of this highly virulent A. hydrophila pathotype.

1. Introduction

and the US, respectively, showed they share a recent common ancestor. VAh-related MAS outbreaks were first recorded in China in the late 1980s, and the first documented occurrence in the US was a case in 2004 in Mississippi (Hossain et al. 2014; Pang et al. 2015). Since that time, vAh-induced MAS has spread across the Southeastern United States and has led to the mortality of over 25 million pounds of channel catfish in the State of Alabama alone (Rasmussen-Ivey et al. 2016b). Currently, vAh isolates are widespread in the US and China and pose a continuous threat for commercial fish farming in these countries. There have been a large number of predicted virulence factors seemingly acquired by horizontal gene transfer in vAh strains (Hossain et al. 2013; Pang et al. 2015); however, their roles in virulence have not been studied experimentally. Two putative virulence factors are the LPS O-antigen and O-antigen capsule. The LPS O-antigen has been shown to play important roles in the protection and adhesion against the host

The bacterial pathogen A. hydrophila is the causative agent of motile Aeromonas septicemia (MAS) in channel catfish (Ictalurus punctatus) and their hybrid (I. punctatus × I. furcatus). The highly virulent A. hydrophila (vAh) isolates affiliated with sequence type 251 (ST251) are responsible for epidemic MAS outbreaks in catfish and are highly virulent as compared to typical opportunistic A. hydrophila isolates from diseased fish (Pridgeon and Klesius 2011). Historically, A. hydrophila has been considered an opportunistic pathogen in stressed or immunocompromised channel catfish. However, vAh strains act as primary pathogens in healthy channel catfish causing high mortalities in mature, market-sized fish (Bebak et al. 2015). Comparative genomic (Hossain et al. 2013) and phylogenetic (Hossain et al. 2014) analyses of vAh strains isolated from Asian carp and channel catfish from China

⁎ Corresponding author at: Department of Biological Sciences, Auburn University, Room 101, Rouse Life Sciences Building, 120 West Samford Avenue, Auburn, AL 36849, USA. E-mail address: [email protected] (M.R. Liles). 1 C.M.T., M.J.H. and D.S. contributed equally to this manuscript. 2 Present address: Department of Biological Chemistry, Johns Hopkins University School of Medicine, Baltimore, MD 21205, USA.

https://doi.org/10.1016/j.aquaculture.2019.734334 Received 28 January 2019; Received in revised form 28 June 2019; Accepted 22 July 2019 Available online 24 July 2019 0044-8486/ © 2019 Elsevier B.V. All rights reserved.

Aquaculture 512 (2019) 734334

C.M. Thurlow, et al.

immune system in gram negative organisms, including Aeromonas hydrophila (Goebel et al., 2008 and Merino et al. 1996). Capsular polysaccharides of other bacterial pathogens have been shown to play significant roles in bacterial pathogenesis by aiding host colonization (Magee and Yother 2001), invasion (Merino et al. 1997), and/or preventing phagocytosis by host macrophages (Locke et al. 2007). In E. coli, capsules are classified in four different groups based on their assembly and secretion to the outer membrane (Whitfield 2006). The group 4 capsule (G4C) consists of oligosaccharide repeats comparable to that of the LPS O-antigen, hence the name “O-antigen capsule” (Peleg et al. 2005; Whitfield 2006). The G4C polysaccharides have been previously characterized in E. coli (Peleg et al. 2005), Shigella sonnei (Caboni et al. 2015), and Salmonella enterica serovar Typhimurium (Gibson et al. 2006), and in S. sonnei the G4C was found to be responsible for virulence modulation in a rabbit infection model (Caboni et al. 2015). In this study, we determined the role of LPS O-antigen and G4C assembly in the virulence of A. hydrophila to channel catfish.

pDMS197 (Edwards et al. 1998). To generate a waaL deletion mutant, two primer sets (LiupF/LiupR and LidnF/LidnR; Table 2) were used to amplify approximately 400 bp upstream and downstream sequences of the waaL gene, respectively, using ML09-119 genomic DNA as template. The chloramphenicol acetyltranferase gene (cat) was amplified from plasmid pMJH46 (Hossain et al. 2015) using primers catF and catR (Table 2). Primers LiupR and LidnF contained reverse complemented sequences of catF and catR primers, respectively, at their 5′-ends. The two PCR fragments were fused to the flanking cat cassette by splicing by overlap extension PCR (SOE) (Horton et al. 1989) using primers LiupintF and Lidn-intR (Table 2). The suicide plasmid pDMS197 was digested with XbaI and blunt ended using the DNA terminator end repair kit (Lucigen, Middleton, WI). The PCR product was ligated to pDMS197 using quick ligase ligation kit (NEB, Ipswich, MA) and electroporated into E. coli SM10λpir to generate plasmid pDMS197waaL that contained sequences upstream and downstream of the waaL gene and therefore replaced waaL with a cat cassette. The suicide plasmid pDMS197waaL was introduced into A. hydrophila ML09-119 by conjugation with E. coli SM10λpir bearing plasmid pDMS197waaL. Single cross-over mutants of A. hydrophila ML09-119 were selected on TSA plate supplemented with chloramphenicol, tetracycline and colistin. Double cross-over mutants were obtained by plating isolates onto LB agar plates supplemented with 15% sucrose and 12.5 μg ml−1 chloramphenicol. Mutants grown on this selective medium were subjected to phenotypic and genotypic characterizations. The wzy gene from vAh strain ML09-119 was deleted according to the methods described above using primers WzyupF, Wzyup-intF, WzyupR, WzydnF, Wzydn-intR and WzydnR as listed in Table 2. The complete deletion of waaL from ML09119 waaL::catae and wzy from ML09-119 wzy::catae was confirmed by PCR followed by Sanger sequencing at the Auburn University Genomics and Sequencing Laboratory. The presence of the O-antigen was determined by LPS extraction using a LPS Extraction Kit (Portsmouth, NH).

2. Materials and methods 2.1. Animal welfare statement All channel catfish challenges were conducted under the approval of the Institutional Animal Care and Use Committee (IACUC) of Auburn University in compliance with U.S. regulatory standards for the humane care and use of laboratory animals. 2.2. Bacterial strains, plasmids and culture conditions Bacterial strains and plasmids used in this study are listed in Table 1. Oligonucleotide primers used for mutant construction, complementation, and sequencing are listed in Table 2. Primers were purchased from Eurofins Operon (Huntsville, AL). A. hydrophila ML09-119, a well characterized vAh strain, isolated from a diseased catfish derived from an epidemic outbreak of MAS in channel catfish in Alabama in 2009 (Hossain et al. 2013; Tekedar et al. 2013), was used in this study. Wild type strain AL06-06, which was isolated from a diseased fish without history of an epidemic MAS outbreak (Hossain et al. 2013), was considered a reference strain. A. hydrophila ML09-119 and its respective mutant strains were routinely grown on trypticase soy broth (TSB) and trypticase soy agar (TSA) at 30 °C with and without shaking at 200 rpm. E. coli strains SM10λpir and CC118λpir were grown on 2 × YT medium at 37 °C. When required, bacterial growth media were supplemented with sucrose (15%), chloramphenicol (12.5 μg ml−1or 25.0 μg ml−1), tetracycline (10 μg ml−1), and/or colistin (10 μg ml−1).

2.5. Generation of ML09-119 waaL::cat and ML09-119 gfc operon mutants by recombineering A. hydrophila isogenic mutants gfcB, gfcC, gfcD and gfcBCD were generated by recombineering according to methods described previously (Hossain et al. 2015), using primer pairs specific to the respective gene sequences (Table 2). Briefly, a genome-integrated cat cassette with flanking Flippase Recombinase Target (FRT) sequences was PCR amplified from the genomic DNA of mutant A. hydrophila ML09-119vgr3::cat using primer pairs ymcC-FRT-Fn and ymcC-FRT-R, ymcB-FRT-F and ymcB-FRT-R, Lipo-FRT-F and Lipo-FRT-R, and LipoFRT-F and ymcC-FRT-R for deletions of gfcB, gfcC, gfcD and gfcBCD, respectively, from A. hydrophila ML09-119 (Table 2). These primers contained a 60 bp sequence homologous to the targeted genes at their 5′ and 3′-ends. Purified PCR amplicons containing the cat cassette bracketed by FRT regions flanking homologous regions specific to each of gfc operon genes were introduced by electroporation into electrocompetent A. hydrophila ML09-119 containing the recombinogenic plasmid pMJH65. PMJH65 was deemed necessary for the integration of these PCR amplicons due to the integration of the lambda Red recombinase machinery into this vector (Hossain et al. 2015). Electroshocked A. hydrophila ML09-119 cells were recovered in SOC medium overnight at 30 °C, with shaking at 200 rpm. Mutants were selected on 2 × YT plates supplemented with 25 μg ml−1 chloramphenicol after overnight growth at 37 °C. Mutants were streaked onto TSA plates and incubated overnight at 37 °C for the removal of the temperature sensitive plasmid pMJH65. Mutants were transferred to TSB broth with 25 μg ml−1 chloramphenicol and grown overnight at 30 °C. The removal of plasmid pMJH65 from mutant strains was confirmed by their lack of growth in TSB supplemented with tetracycline (10 μg ml−1). Markerless mutants of A. hydrophila were generated using flp recombinase plasmid pCMT-flp according to methods described

2.3. Phylogenetic analysis of gfc operon The evolutionary history of each of the genes in the gfc operon was inferred using the Maximum Likelihood method based on the JTT matrix-based model using MEGA7.0 (Kumar et al. 2016). The tree with the highest log likelihood was used for inference. The percentage of trees in which the associated taxa clustered together is shown next to the branches of each tree. Initial tree(s) for the heuristic search were obtained automatically by applying Neighbor-Join and BioNJ algorithms to a matrix of pairwise distances estimated using a JTT model, and then selecting the topology with superior log likelihood value. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. All positions containing gaps and missing data were eliminated. 2.4. Generation of O-antigen ligase (ML09-119 waaL::catae) and O-antigen polymerase (ML09-119 wzy::catae) mutants by allelic exchange Non-polar targeted gene deletions in the A. hydrophila ML09-119 O– antigen ligase gene waaL were constructed using suicide plasmid 2

Aquaculture 512 (2019) 734334

C.M. Thurlow, et al.

Table 1 List of bacterial strains and plasmids used in this study. Bacterial strains and plasmids

Relevant features

References

Wild-type virulent Aeromonas hydrophila (vAh) isolate Reference A. hydrophila (rAh) isolate Deletion of gfcD through the integration of cat cassette Deletion of gfcC through the integration of cat cassette Deletion of gfcB through the integration of cat cassette Deletion of gfcD containing cat cassette flanked by FRT regions Deletion of gfcC containing cat cassette flanked by FRT regions Deletion of gfcB containing cat cassette flanked by FRT regions Deletion of gfcBCD containing cat cassette flanked by FRT regions Unmarked deletion of gfcD gene Unmarked deletion of gfcC gene Unmarked deletion of gfcB gene Unmarked deletion of gfcBCD operon A. hydrophila gfcBCD operon mutant with pBBC2 A. hydrophila with pMJH97 A. hydrophila gfcD gene mutant with pMJH97 A. hydrophila with pBBC2 Replacement of waaL gene with cat gene by allelic exchange Replacement of waaL gene with cat gene by recombineering Replacement of wzy gene with cat gene Replacement of vgr3 gene with cat gene

Hossain et Hossain et This study This study Hossain et This study This study This study This study This study This study Hossain et This study This study This study This study Hossain et This study Hossain et This study Hossain et

thi-1thr leu tonA lacY supE recA::RP4-2-TcT::Mu Kmrλpir Δ(ara-leu) araD ΔlacX74 galE galKphoA20 thi-1 rpsE rpoB argE(Am) recA1 λpir F−mcrAΔ(mrr-hsdRMS-mcrBC) endA1 recA φ80dlacZΔM15ΔlacX74 araD139 Δ (ara,leu)7697 galU galK rpsL (StrR) nupG λ− tonA

Simon et al. (1983) Herrero et al. (1990) Lucigen (Middleton, WI)

Suicide vector, sacB, TetR Suicide vector contains cat cassette flanked by upstream and downstream regions of waaL Suicide vector contains cat cassette flanked by upstream and downstream regions of wzy CmR, waaL cloned into plasmid pGNSBAC CmR and conjugally transferrable plasmid for cloning gene(s) of interest for complementation of mutant strains cat-oriR-oriT-gfcABCD complement vector CmR and conjugally transferrable recombinogenic plasmid TetR and conjugally transferrable recombinogenic plasmid TetR and temperature-sensitive Flp recombinase plasmid

Edwards et al. (1998) This study This study This study Hossain et al. (2015) Hossain et al. (2015) Hossain et al. (2015) Hossain et al. (2015) Hossain et al. (2015)

Bacterial strains A. hydrophila A. hydrophila ML09-119 A. hydrophila AL06-06 ML09-119 gfcD::cat ML09-119 gfcC::cat ML09-119 gfcB::cat ML09-119 gfcD::cat::FRT ML09-119 gfcC::cat::FRT ML09-119 gfcB::cat::FRT ML09-119 gfcBCD::cat::FRT ML09-119 gfcD ML09-119 gfcC ML09-119 gfcB ML09-119 gfcBCD ML09-119 gfcBCD (pBBC2) ML09-119 (pMJH97) ML09-119 gfcD (pMJH97) ML09-119 (pBBC2) ML09-119 waaL::catae ML09-119 waaL::cat ML09-119 wzy::catae ML09-119 vgr3::cat E. E. E. E.

coli coli SM10λpir coli CC118λpir coli 10G

Plasmids pDMS197 pDMS197waaL pDMS197wzy pGNS-BAC-waaL pMJH97 pBBC2 pMJH46 pMJH65 pCMT-flp

al. (2013) al. (2013) al. (2015)

al. (2015)

al. (2015) al. (2015) al. (2015)

into the A. hydrophila gfcBCD, gfcB, gfcC and gfcD mutants was confirmed by PCR using primer pairs CCatR and ymcA-CM-1F, and pABC-R and p15AF (Table 2).

previously (Hossain et al. 2015). Both pMJH65 and pCMT-flp are necessary for carrying out the targeted deletion event and removal of the antibiotic cassette, due to the presence of the lambda Red recombinase and flp recombinase found on each vector, respectively (Hossain et al. 2015). The genotypes of mutants gfcB, gfcC, gfcD and gfcBCD were verified by PCR followed by Sanger sequencing using primer pairs ymcC-upF and ymcC-dnR, ymcA-tetA-dnR and ymcC_tetA-uF, lipo-upF and lipo-dnR, and ligaF and ymcC-dnR, respectively that were specific for DNA sequences immediately upstream and downstream of the chloramphenicol-resistant cassette insertion site of targeted genes.

2.7. Purification of LPS and SDS-PAGE analysis The LPS from AL06-06, ML09-119, ML09-119 waaL::cat (recombineering generated), ML09-119 wzy::catae and ML09-119 gfcD were extracted by phenol/chloroform using the iNtRON LPS Extraction Kit following a protocol modified for previously described methods and the extraction kit manufacturer's recommendations (Marolda et al. 2006). Briefly, A. hydrophila cultures were grown for 16 h at 30 °C and 5 ml of the culture (OD600 = 1.0) was subjected to centrifugation at 13,000 ×g for 30 s. Bacterial pellets were suspended in 1 ml of lysis buffer (containing phenol) and vortexed vigorously until the pellet was dissolved. The samples were then treated with 200 μl of chloroform, vortexed for 30 s, and incubated at room temperature for 5 min. This solution was subjected to centrifugation at 13,000 ×g for 10 min at 4 °C. The top aqueous layer was carefully aspirated to prevent disturbance of the phenol layer and protein precipitate. For secondary purification of the LPS, 800 μl of purification buffer was added to every 400 μl of supernatant recovered and incubated for 10 min at −20 °C. Each treated aliquot was then subjected to centrifugation at 13,000 ×g for 15 min at 4 °C. The resulting pellet was washed with 1 ml of 70% EtOH and dried. Then 70 μl of 10 mM Tris-HCL (pH 8.0) was added to the dried pellet, boiled for 1 min and stored at 4 °C for later use. The LPS was separated with Tris-Glycine NB 12% (NuSep, Inc. Germantown,

2.6. Complementation of waaL and gfc operon mutants To complement the avirulent phenotype of A. hydrophila waaL::cat mutant, the wild type waaL gene was cloned into the HindIII restriction digestion site of pGNS-BAC after PCR amplification of waaL using primers Ligase F-HindIII and Li234R-HindIII (Table 2). A. hydrophila gfcB, gfcC, gfcD and gfcABCD mutants were complemented with the complete gfcABCD operon cloned into the plasmid pMJH97 (Hossain et al. 2015). The resulting plasmid pBBC2 was introduced into each of the gfcABCD, gfcB, gfcC and gfcD mutants by conjugation with E. coli SM10λpir (pBBC2) according to the methods described previously (Hossain et al. 2015). Transconjugants of A. hydrophila mutants complemented with respective plasmids were selected on 2 × YT plates supplemented with chloramphenicol and colistin. The mutant colonies were identified as A. hydrophila by a positive oxidase test to distinguish them from oxidasenegative E. coli SM10λpir. The introduction of complementing plasmid 3

Aquaculture 512 (2019) 734334

C.M. Thurlow, et al.

Table 2 List of primers used in this study. For some primers (denoted by asterisks) the first four phosphodiester internucleotide linkages at the 5′ end of the oligonucleotide were replaced with phosphorothioate internucleotide linkages. Primer name

Sequence in 5′−3′ direction

LidnF LidnR Lidn-intR Liup-intF LiupF LiupR Ligase F-HindIII Li234R-HindIII WzydnF Wzydn-intR WzydnR WzyupF Wzyup-intF WzyupR CatF CatR CCatR ymcA-CM-1F pABC-R p15AF Lipo-upF Lipo-dnR Lipo-FRT-F Lipo-FRT-R ymcC-FRT-R ymcC-upF ymcC-dnR ymcB-FRT-F ymcB-FRT-R ymcA-tetA-dnR ymcC_TetA-uF ymcC-FRT-Fn Liga-F

TTAGCTCACTCATTAGGCAAGATCGGCTCTATGCAACT TGATTATGATGTAATGACTGG GGCAGTTACCATTCATGAGT AGAAGCGGTGCTGATAACG ACTTAAGCTCGCCGAACTC AGTTCAACGGAAGTCTACGCTGTCGAGGCCATGTG AGTCTAAGCTT GACCAGCGCATTGAGAGAGAGG AGTCTAAGCTT GCTCAAGCCAACAACCGCGAA TTAGCTCACTCATTAGGCCTAGCTGTGGTGCCAGAATA CTGATGTTATTATTGACCAAG CATTCAATATAGTGTCTGGA CCGCGACAACAACTCCTT GTGACGCCACCGATGATA AGTTCAACGGAAGTCTACGCACTTCCTGTATCAAGATT GTAGACTTCCGTTGAACT GCCTAATGAGTGAGCTAA GGCCGTCGACCAATTCTCATGTT GCGACAAAAATAAGGCTGCCA TGAGTCGCAAGAATGGCCT TCACATATTCTGCTGACGCACC CCGAATGGTAATCCACAGTT TAGAACAGCTGGTCACGAGA C*A*A*C*TGCTCGCCCTTTTTGATGAAAAAAGATCGGCTCTATGCAACTTTTGA GTGTAGGCTGGAGCTGCTTC T*A*G*A*GATATCAATATTCGTATTGCCAATCTCCTTGCTAATCGAGTACCAGA CATATGAATATCCTCCTTAGT T*T*G*G*CAGCCTTATTTTTGTCGCTTTGCAATCATTTTCCTATTCCACTGAGTAACCCATCT CATATGAATATCCTCCTTAGT GTCCAGCTCTTGTAGTAACTGC TAATGCGAATGACGGCTCCACC ATGAAAAGATACCTGATGGTTAGTCTAATACCATTTTTTGCTTTGGCTCAAGCTGATGTA GTGTAGGCTGGAGCTGCTTC CATTCTGGTACTCGATTAGCAAGGAGATTGGCAATACGAATATTGATATCTCTATACTG CATATGAATATCCTCCTTAGT CTTGCCAGAGACGGAACTTGAA GGCCTGATCCATTGCAGAA C*C*A*T*CAGGTATCTTTTCATAATGCGTACGGCTTTAATATTATCATTTCAATAACCGGTAA GTGTAGGCTGGAGCTGCTTC ATGACCAGGCTGGCGAAGATCC

MD). Gel electrophoresis was carried out at 250 V for approximately 1 h. The gel was then observed through silver staining on a Gel Doc™ EZ Imager reading system (BIO-RAD, Hercules, CA).

A. hydrophila strains ML09-119 and ML09-119 gfcD were recovered from glycerol stocks at −80 °C and incubated on TSA for 24 h at 30 °C. A single colony of each strain was inoculated into TSB medium and incubated overnight at 30 °C with shaking at 200 rpm. Overnight cultures were standardized to OD600 = 1.5 and diluted into TSB medium (1:100). Triplicates of each strain were incubated for 16 h in 75 ml of TSB at 30 °C with shaking at 200 rpm. Bacterial cells were pelleted by centrifugation at 20,000 ×g for 15 min at 4 °C and the supernatant was decanted and sterilized by passage through disposable 0.22 μm vacuum filters. Sterile supernatants were used as the starting point for secreted protein purification.

2.8.2. Ion-exchange chromatography Prior to ion-exchange chromatography, crude protein solutions were filtered through low-binding 0.45 μm syringe filters. Following filtration, the protein solution was loaded at a flow rate of 1 ml/min onto a 5-ml HiTrap QHP prepacked column (GE Healthcare) previously equilibrated with 10 column volumes of loading buffer (20 Mm TrisHCl, pH 7.6) connected to a Biologic LP Chromatography system (BioRad). Samples were eluted at a flow rate of 1 ml/min using a linear gradient from 0 to 100% elution buffer (1 M NaCl in 20 mM Tris-HCl, pH 7.6). Eluate was monitored at 280 nm and visualized on LP Data View software (v1.03, Bio-Rad). Fractions of 2 ml were automatically collected with Model 2110 fraction collector (Bio-Rad), and fractions containing proteins were pooled and de-salted by passage through a 5ml HiTrap Desalting prepacked column (GE Healthcare). Protein concentrations were determined using a Bradford assay kit (Coomassie Plus, Thermo Scientific). Fractions were then submitted for protein identification and quantitation via liquid chromatography- tandem mass spectrometry (HPLC MS/MS).

2.8.1. Ammonium sulfate precipitation Secreted proteins were precipitated from cell-free supernatants by the addition of ammonium sulfate crystals (Fisher Scientific) to achieve 65% saturation, followed by an incubation at 4 °C on a rotary platform shaker with gentle mixing for 24 h. Precipitated proteins were collected by centrifugation at 30,000 ×g for 45 min at 4 °C, then dissolved in 10 ml cold Tris buffer (20 mM Tris-HCl, pH 7.6) + protease inhibitor (Complete tablets, mini, EDTA-free (Roche)). Resuspended proteins were dialyzed twice, for 18 h and 12 h, respectively, against the same buffer in 10 kDa dialysis cassettes (Slide-A-Lyzer, Thermo Fisher). After dialysis, the total volume was adjusted to 20 ml by the addition of cold Tris buffer.

2.8.3. Proteomics analysis Proteomics analyses were performed as previously described with minor modifications (Ludwig et al. 2016). Samples were prepared for analysis as follows: 20 μg of protein per samples was diluted to 35 μl in NuPAGE LDS sample buffer (Invitrogen), reduced with dithiothreitol, and denatured at 70 °C for 10 min prior to loading onto Novex NuPage 10% Bis-Tris protein gel (Invitrogen). The gel was separated as a short stack (10 min, 200 V constant) and stained overnight with Novex Colloidal Blue Staining kit (Invitrogen). Gels were destained and each lane was cut into single molecular weight fractions and equilibrated in 100 mM ammonium bicarbonate. Each plug was then digested overnight with Trypsin Gold (Mass Spectrometry grade (Promega)) following manufacturer's instructions and peptide extracts were

2.8. Secreted proteins

4

Aquaculture 512 (2019) 734334

C.M. Thurlow, et al.

reconstituted to 0.1 μg μl−1 in 0.1% formic acid.

the supernatants were collected and filtered through 0.22 μm filters. Following filtration, 500 μl of the bacterial supernatant was added to an equal volume of 1% suspension of washed sheep erythrocytes in 1× PBS (pH = 7.2). All replicates were incubated at 30 °C for 40 min, and hemolytic activity was determined every 10 min for the duration of the 40 min incubation. To determine hemolytic activity, each replicate was subjected to centrifugation at 2000 ×g for 5 min at 4 °C to remove unlysed erythrocytes, the supernatants were aspirated, and the absorbance of each replicate supernatant was determined at 540 nm in triplicates using a Tecan Infinite M1000Pro Plate Reader (Tecan Group Ltd., Männedorf, Switzerland). A negative control was prepared by mixing the erythrocyte suspension with sterile TSB and treated similarly to the samples. The absorbance corresponding to 100% hemolysis was determined by adding sodium dodecyl sulfate (SDS, 0.1% final concentration) to blank tubes causing the lysis of all erythrocytes present.

2.8.4. Mass spectrometry Prepared peptide digests (8 μl) were injected onto a 1260 Infinity nHPLC stack (Agilent Technologies) and separated using a 71 μm I.D. X 15 cm pulled-tip C-18 column (Jupiter C-18300 Å, 5 μm (Phenomenex)) running in-line with a Thermo Orbitrap Velos Pro hybrid mass spectrometer, equipped with a nano-electrospray source (Thermo Fisher). All data were collected in CID mode. nHPLC was configured with binary mobile phases comprised of 0.1% formic acid (solvent A) and 0.1% formic acid in 85% acetonitrile (solvent B). Following each parent scan (300–1200 m/z at 60 K resolution), fragmentation data (MS2) were collected on the top most intense 15 ions. For data-dependent scans, charge-state screening and dynamic exclusion were enabled with a repeat count of 2, repeat duration of 30 s., and exclusion duration of 90 s. 2.8.5. Mass spectrometry data conversion and searches Xcalibur RAW files were collected in profile mode, centroided, and converted to mzXML using ReAdW v3.5.1 (IonSource). The mgf files were then created using MzXML2Search (included in Trans-Proteomics Pipeline v3.5) for all scans. The data were searched using SEQUEST (Thermo Fisher), which was set for two maximum missed cleavages, a precursor mass window of 20 ppm, trypsin digestion, variable modification C at 57.0293, and M at 15.9949. Searches were performed with a species-specific subset of the UniRef 100 database.

2.10. Bactericidal serum assay To determine the survival rate of organisms when exposed to complement-mediated killing, a basic serum bactericidal assay was used according to the methods described previously (Hsieh et al. 2012). Briefly, all organisms tested were grown on TSA plates and incubated at 30 °C for 24 h. After 24 h, individual colonies were selected and inoculated in 25 ml of TSB and allowed to grow overnight in a 30 °C water-bath with shaking. The overnight culture (1 ml) was added to 9 ml (1:10) of fresh TSB and returned to the water bath to grow to an OD600 = 0.4. The bacterial suspension was then diluted in sterile 1xPBS to a final concentration of 1.0 × 106 CFU ml−1. Aliquots of 5 μl of suspension were inoculated 1:1 in a mixture of normal catfish serum and sterile 1xPBS. Heat-inactivated normal catfish serum was also used to compare survival rates. All samples were incubated for 1 h at 25 °C, after which samples were serially diluted and plated on TSA plates. Plate counts were performed after a 24 h incubation and the percent survival was calculated by dividing the number of CFUs observed on normal catfish serum plates by the number seen on plates spread with heat-inactivated serum. This number was multiplied by 100 and then compared using the Statistical Analysis System (SAS) GLIMMX Procedure and Tukey-Kramer at the P ≤ .05 confidence level. This experiment was repeated 4 times and triplicates were used each time for each organism.

2.8.6. Peptide filtering, grouping, quantification and statistical analyses The list of peptide IDs generated based on SEQUEST search results were filtered using Scaffold (Protein Sciences, Portland, Oregon), which filters and groups all peptides to generate and retain only high confidence IDs while also generating normalized spectral counts across all samples to allow for relative quantification. Filter cut-off values were set with minimum peptide length of > 5 amino acids, with no MH+ charge states, peptide probabilities of > 80% C.I., and with the number of peptides per protein ≥2. Proteins probabilities were then set to a > 99% C.I. with false discovery rate < 1.0. Scaffold incorporates the two most common methods for statistical validation of large proteome data sets, the false discovery rate (FDR) and protein probability (Weatherly et al. 2005). Relative quantification across samples were then performed via spectral counting (Old et al. 2005) and, when relevant, spectral count abundances were normalized between samples (Beissbarth et al. 2004). Proteins present in at least two experimental replicates were included in significance analyses. To determine statistical significance, two non-parametric statistical analyses were performed between each pair-wise comparison, including reproducibilityoptimized test statistic (ROTS, bootstrapping value = 1000) combined with single-tail t-test (P < .05) (Pursiheimo et al. 2015; Suomi et al. 2017). These were then sorted according to the highest statistical relevance in each comparison. For protein abundance ratios determined by normalized spectral counts, a fold change threshold ≥1.5 was set for significance. Protein abundance of proteins present in only one experimental group was set as the average of the normalized quantitative value.

2.11. Determination of buoyancy of A. hydrophila strains Buoyancy differences between vAh strain ML09-119, ML09-119 gfcD and the gfcD complemented mutant (ML09-119 gfcD (pBBC2) were observed according to the methods described previously (Peleg et al. 2005). Each strain was grown in culture tubes at 30 °C with shaking in 2 ml TSB to an OD600 = 1.0. The medium was underlaid with 2 ml Percoll (55% Percoll, 25 mM phosphate buffer, pH 6.5) and subjected to centrifugation at 2000 ×g at RT for 10 min. Cells containing capsules were more buoyant and formed a band at the Percoll-medium interface, while cells lacking capsules were less buoyant and formed a pellet. 2.12. Capsule visualization by bright field microscopy

2.9. Hemolytic assay

A. hydrophila strains ML09-119 (pMJH97), ML09-119 gfcD (pMJH97), and ML09-119 gfcD (pBBC2 containing gfcABCD) were grown overnight in TSB with 25 μg ml−1 of chloramphenicol at 30 °C with shaking at 200 rpm. The empty vector pMJH97 was introduced into the wild-type ML09-119 and the gfcD mutant so that capsule visualization could be compared in these strains in the presence of chloramphenicol. Following overnight incubation, the three strains were adjusted to an OD600 = 1.0 (~1 × 109 CFU ml−1) and diluted 1:100 into fresh TSA agar plates containing 0.25% agar and chloramphenicol (25 μg ml−1). Each plate was incubated statically for 48 h at 30 °C. After 48 h, 10 μl of bacteria were aspirated from the soft agar

The hemolytic activity from ML09-119 (pMJH97), ML09-119 gfcD (pMJH97), and ML09-119 (pBBC2) were analyzed according to a previously described protocol with modifications (Santos et al. 1999). Each bacterial strain was recovered from a glycerol stock at −80 °C and incubated on TSA for 24 h at 30 °C. A single colony of each strain was inoculated into fresh TSB medium and incubated overnight at 30 °C with shaking at 200 rpm. Overnight cultures of different strains were normalized to OD600 = ~1.5 and diluted in triplicates into fresh TSB medium (1:100) and grown for 16 h post-inoculation at 30 °C with shaking at 200 rpm. After centrifugation at 10,000 ×g for 10 min at RT, 5

Aquaculture 512 (2019) 734334

C.M. Thurlow, et al.

plates and stained following a previously described capsule staining technique (Croxatto et al. 2007). Each of the three replicate bacterial strains were spread onto a microscope slide. After drying, each smear was negatively stained using 1% Congo Red solution (unfiltered). Following staining with Congo Red, slides were flooded with Maneval's modified stain and allowed to stain for 1 min. Stained slides were then gently washed with sterile distilled water, air dried, and viewed at 100× using an Olympus BX61 QI Imaging Microscope using CellSens Dimensions Imaging software (Olympus America Inc., PA).

approximately 15 g. Each treatment contained three replicates of 50-l tanks in a randomized block design with 10 fish/replicate tank. The fish were acclimatized for at least 12 days prior to challenge with wild type A. hydrophila strains or its respective mutant. The water temperature was maintained at 28 °C during the acclimatization and throughout the course of the experiment. All tanks were on flow-through at a rate of 0.6 l min−1 throughout the experiments to maintain optimal water quality. The fish were fed daily at 2% body weight (BW) daily up until the challenge. Cryopreserved (−80 °C) stocks of A. hydrophila strains were streaked on TSA plates and incubated overnight at 30 °C to obtain well isolated colonies. A single isolated colony for each strain was grown in TSB separately overnight and 50 μl of each culture was transferred to fresh TSB (1:100 dilution) to grow until OD600 = 0.8–1.0. Spent growth media was removed by subjecting the culture to centrifugation at 8000 ×g for 10 min and resuspended with fresh TSB to adjust OD600 to ~1.0 for each strain. For each infection trial, three groups of 10 fish were injected IP with 200 μl of bacterial cells adjusted to the approximate concentration of 1.0 × 106 CFU/fish. Fish were inoculated within 1 h of resuspension of bacterial cells in fresh TSB. The fish were monitored for mortality for 7 days and percent mortalities were calculated. Dead fish were necropsied, and portions of kidney and liver collected and assessed for the presence of A. hydrophila by bacterial culture and identification.

2.13. Biofilm formation assay To quantify differences in biofilm formation between ML09-119 (pMJH97), ML09-119 gfcD (pMJH97) and the gfcD complemented mutant (ML09-119 gfcD (pBBC2), a biofilm assay was conducted following methods previously described with modifications (O'Toole 2011). The empty vector was introduced to ML09-119 and ML09-119 gfcD so that growth rates and attachment could be compared under similar growth conditions in the presence of chloramphenicol (Fig. 4). Three isolates were incubated in 2 ml TSB containing 25 μg ml−1of chloramphenicol at 30 °C overnight with shaking at 200 rpm. Each strain was adjusted to an OD600 = 1.0 and diluted (1:100) into 6 ml of fresh TSB containing chloramphenicol (25 μg ml−1). The 6 ml aliquot was then separated into three replicate tubes containing 2 ml of TSB with chloramphenicol and incubated statically for 24 h at 30 °C. Each replicate tube was then standardized to an OD600 = 1.5 and diluted (1:100) into a 1 ml master mix containing fresh TSB with chloramphenicol. Four replicates of 100 μl of the master mix was placed into 4 wells of a Greiner 96-well flat bottom polystyrene plate. The plate was incubated statically at 30 °C for 48 h. Unbound cells were removed from the plates through gentle washing (2×) with 300 μl of sterile ultrapure water. One hundred twenty-five microliters of 0.1% crystal violet was added to each well of the 96 well plate and incubated at RT for 15 min. To remove unbound crystal violet, each well was washed three times with 300 μl of sterile ultrapure H2O and dried. One hundred twenty-five microliters of 30% acetic acid in sterile dH2O was placed into each well to solubilize the crystal violet bound to the attached cells. Absorbance on each plate was quantified at OD550 using a Tecan Infinite M1000Pro plate reader.

2.16. Prevention of MAS disease in catfish by vaccination with A. hydrophila ML09-119 gfcD Survivors of the initial challenge with ML09-119 gfcD were housed in their respective tanks (under the same aquaria conditions as stated above) for 21 days. Following the 21-day acclimation period, one fish from each replicate tank was sacrificed for bleeding. All blood samples were coagulated for 12 h at 4 °C, centrifuged and the serum was aspirated, retained, and stored at −20 °C until analysis. The remaining survivors were challenged via IP injection with wild-type ML09-119 at a concentration of 5 × 106 CFU/fish. Fish in the negative control treatment group were injected with sterile TSB, whereas the positive control fish were naïve fish that were housed in adjacent tanks during the acclimation period, initial challenge, and 21-day incubation period. 2.17. Enzyme linked immunosorbent assay

2.14. Fluorescent microscopy of biofilm formation

IgM anti-Aeromonas antibodies produced by the catfish post-vaccination were quantified through enzyme linked immunosorbent assays (ELISAs) using the HRP anti-mouse IgG (H + L) Protein Detector ELISA Kit following the manufacturer's recommendations with some modifications (KPL SeraCare, Milford, MA, USA). ML09-119 was sub-cultured overnight at 30°C with shaking in 5 ml TSB and cells were collected by centrifugation at 10,000 ×g for 10 min at room temp. The TSB was aspirated, and cells were diluted in 1× coating solution at a final cell concentration of ~1 × 107 CFU ml−1. One hundred microliters of cells in coating solution were placed into each well of the 96 well plates and incubated for 1 h at room temp. The coating solution with unbound bacteria was removed and 300 μl of 1% blocking solution was placed into each well and incubated at room temp for 15 min. The plate was then washed 3 times with 1× wash solution and dried. For the ELISA, 100 μl of 1% blocking solution was placed into each well of the 96 well plate. Serum from each fish was diluted 1/10 into the 1% blocking solution in each well and serially diluted to a 1/160 dilution. The plates were covered and incubated for 1 h at room temperature. Following the 1 h incubation, each well of the plate was washed 3× with 300 μl of 1× wash solution. The plate was dried and 100 μl of the 9E1 mouse immunoglobin (IgG) anti-catfish IgM antibody (University of Mississippi, Jackson, MS, USA), diluted 1:35, was placed into each well and incubated for 1 h atRT. Unbound antibodies were removed from wells and each well was then washed 3× with 1× wash solution and dried. One hundred microliters of 1:1000 diluted HRP conjugated

Biofilm visualization was carried out on wild-type strain ML09-119 (pMJH97), ML09-119 gfcD (pMJH97) and ML09-119 gfcD (pBBC2) following methods previously described with modifications (Shibata et al. 2006). Each strain was grown overnight in TSB containing 25 μg ml−1 chloramphenicol at 30 °C. The cultures were standardized to an OD600 = 1.5 and each strain was diluted (1:100) into 25 ml fresh TSB with chloramphenicol (25 μg ml−1). The 25 ml cultures were placed into sterile petri dishes containing UV sterilized coverslips (22 mm × 22 mm × 1 mm) and were incubated statically at 30 °C for 48 h. The coverslips were then extracted from the broth and gently washed of all non-bound cells (3×) with 25 ml of sterile dH2O. Coverslips were then stained in the dark with 400 μl of SYBR Gold nucleic acid stain (Invitrogen, CA) (1:1000 dilution) for 15 min, at which time they were rinsed gently with sterile H2O and dried. The coverslips were fixed using a drop of SlowFade Gold Antifade (supplied concentration) (Thermo Fisher Scientific, MA). Fixed cells were viewed using a FITC filter (excitation from 465 to 505 nm; emission from 515 to 565 nm) at 40× and 100× using an Olympus BX61 QI Imaging Microscope using CellSens Dimensions Imaging software (Olympus America Inc., PA). 2.15. Virulence studies in channel catfish Channel catfish fingerlings were obtained from Auburn University North Fisheries Unit. Each fish was roughly 3.5 months old and weighed 6

Aquaculture 512 (2019) 734334

C.M. Thurlow, et al.

anti-mouse goat IgG (supplied in kit) was placed into each well and incubated for 1 h at RT. The plate was emptied, washed 3× with 1× wash solution and dried. The reaction was initiated using 100 μl of the ABTS peroxidase substrate system (at recommended manufacturer's concentration). The reaction was stopped after 15 min using 100 μl of 1× stop solution and the plate scanned at OD405 using a Tecan Infinite M1000Pro plate reader. Titers were reported as positive if the serum anti-Aeromonas antibody levels measured at least 2× the levels of the plate blank and negative control (naïve fish) spectrophotometer readings. The blank represented wells that were not subjected to a primary Ab, all remaining steps were performed concurrent to the negative control and the fish sera of each treatment. Titers were also compared to positive control serum, which was extracted from fish that survived the initial encounter with ML09-119 (21 days post challenge).

operon to vAh virulence. 3.2. Annotation of the A. hydrophila ML09-119 gfc operon The annotated genome for A. hydrophila ML09-119 deposited in GenBank (accession # CP005966.1, Jan. 2014) predicts three open reading frames (ORFs) upstream of waaL with homology to gfcD, gfcC, and gfcB (synonymous with ymcA, ymcB and ymcC, respectively) (Tekedar et al. 2013). In Escherichia coli K-12 substrain MG1655 there are two operons, yjbEFGH and gfcABCD, that have homology with the A. hydrophila ML09-119 gfc operon and have been implicated in exopolysaccharide production (Ferrieres et al. 2007; Peleg et al. 2005). A maximum likelihood phylogenetic analysis indicated that the predicted amino acid sequences from A. hydrophila ML09-119 were more closely related to the GfcB, GfcC and GfcD sequences from E. coli K12 compared to their paralogs YjbF, YjbG and YjbH, respectively (Fig. S2 and data not shown). Multiple alignments of the GfcD amino acid sequences demonstrated that the A. hydrophila ML09-119 GfcD sequence annotated in GenBank (AHML_15465) had 266 amino acids truncated from its amino-terminus (Fig. S3). A full-length GfcD sequence was obtained using GeneMark.hmm to predict ORFs from the genomic region between waaL and gfc (Lukashin and Borodovsky 1998). The subsequent generation of A. hydrophila ML09-119 mutants and assessment of their respective virulence was therefore conducted for each of the predicted gfcBCD loci that were hypothesized to play a role in G4C assembly and potentially contribute to A. hydrophila virulence. Later, in silico searches were conducted to identify A. hydrophila ML09-119 homologs of the other genetic loci implicated in G4C assembly in E. coli, specifically the genetic loci gfcA (synonymous with ymcD), yccZ, etp and etk (Peleg et al. 2005). A search of the region upstream of gfcB with SeqBuilder (DNAStar) revealed an 84 amino acid reading frame with characteristics typical of GfcA homologs from diverse bacterial species: a predicted signal peptidase 1 cleavage site between residues 20 and 21 (SignalP 4.1); high percentage of Ala, Gly, Ser, and Thr residues; and a basic residue at the C-terminus. Although not predicted in the annotation in the original genome entry (CP005966.1, 2014) it is now annotated as a hypothetical protein (NC_021290.1, Apr. 2017). Based on these observations, we predict that A. hydrophila ML09-119 encodes a gfcABCD operon. Additional in silico searches were conducted to identify A. hydrophila ML09-119 homologs of the other genetic loci implicated in G4C assembly in E. coli, specifically the genetic loci yccZ (gfcE), etk, and etp but did not return any significant BLAST hits. A search of GenBank using the gfcABCD operon sequence revealed that this operon is present in all sequenced A. hydrophila isolates of epidemic disease outbreaks in channel catfish in the United States and carp isolates in China (data not shown). A search of the Conserved Domain Database predicted that the GfcD protein is homologous with the bacterial putative lipoprotein pfam06082 in the pfam database for E. coli K12 and EPEC 0127:H6 strains (Fig. S4).

2.18. Statistical analyses Statistical analysis was performed using Statistical Analysis System (SAS) software (SAS Institute Inc., Cary, NC, USA) and R (R Foundation for Statistical Computing, Vienna, Austria). Statistical significance was determined by one-way analysis of variance (ANOVA) and Tukey post hoc multiple comparisons tests. Quantitative data are presented as means ± standard error. Differences were considered statistically significant if a P value is < 0.05. 3. Results 3.1. Virulence of A. hydrophila ML09-119 O-antigen ligase and O-antigen polymerase mutants in channel catfish To test whether the LPS O-antigen plays a significant role in vAh virulence to a catfish host, ML09-119 waaL::catae, an O-antigen mutant missing most of the LPS ladder, and ML09-119 wzy::catae, an O-antigen semirough mutant, were generated by deleting the predicted O-antigen ligase gene waaL and O-antigen polymerase gene wzy, respectively. Allelic exchange methods were used to generate a waaLup:cat:waaLdn and a wzyup:cat:wzydn cassette to replace waaL and wzy with the cat gene in each of their respective mutants. We observed a reduction of LPS with regard to O-antigen repeats, except for a single O-antigen unit present in the ML09-119 wzy::catae mutant (Fig. S1 and data not shown), whereas the ML09-119waaL::catae mutant was observed to have reduced but not a complete absence of O-antigen units (Fig. S1). An experimental intraperitoneal (IP) challenge of channel catfish with wild-type ML09-119 and these respective O-antigen mutants demonstrated that the O-antigen ligase mutant was significantly attenuated in its virulence in channel catfish, whereas the O-antigen polymerase mutant retained its virulence when compared to wild-type ML09-119 (Fig. 1, P = .0001 and P = .75 respectively). However, complementation of ML09-119 waaL::catae with the wild-type waaL gene did not restore virulence of ML09-119 waaL::catae mutant in channel catfish (data not shown). This suggested that there was a polar effect of the ML09-119 waaL::catae mutant on adjacent genetic loci. To evaluate this hypothesis, we generated an O-antigen mutant using recombineering to precisely delete the waaL coding region of ML09-119 without affecting its flanking sequences. The experimental challenge of channel catfish with the waaL mutant generated by recombineering revealed that this mutant was equally virulent to channel catfish as compared to wildtype ML09-119 (Fig. 1, Panel B). These findings suggested that the waaL mutant constructed by allelic exchange interfered with the adjacent gfc operon transcription, particularly gfcD that is located 122 bp upstream of waaL (Fig. 2). Sequence analysis of the waaL flanking region in the mutant generated by allelic exchange revealed that a predicted gfcD transcriptional termination site (AGATAATAAAAGGGCGAGCAGTTTG CAACTGCTCGCCCTTTTTGATGAAAA) was deleted. Due to the lack of attenuation of virulence observed in the O-antigen mutants, we therefore decided to focus primarily on evaluating the contribution of the gfc

3.3. Assessing the virulence of A. hydrophila ML09-119 with mutations in the gfc operon To determine the contribution of gene products in the gfc operon to capsule biosynthesis and A. hydrophila virulence, markerless in-frame deletion mutants were constructed for gfcB, gfcC and gfcD (Table 1). The cumulative percent survival of channel catfish fingerlings IP injected with different A. hydrophila mutants revealed that each of the genes inactivated in the gfc operon contributed to vAh virulence (Fig. 3). Catfish challenged with wild-type A. hydrophila ML09-119 indicated an average survival rate of 27% ± 22% post-IP injection with 1.0 × 106 CFU/fish and the onset of mortality was very rapid with most mortalities observed by 6 h post-injection. Based on Tukey post hoc multiple comparisons, a significant difference was observed between the survival rates of the fish challenged with the wild-type compared to 7

Aquaculture 512 (2019) 734334

C.M. Thurlow, et al.

Fig. 1. Percent survival of channel catfish after challenge with vAh strain ML09-119, O-antigen ligase mutant (waaL) or O-antigen polymerase mutant (wzy). (Panel A) Percent survival of fish challenged with the waaL mutant generated by allelic exchange was significantly different when compared to the percent survival of fish challenged with the wild-type (ML09-119) (P = .0001). (Panel B) Percent survival of fish challenged with the waaL mutant generated by recombineering or the wzy mutant were not significantly different when compared to the percent survival of fish challenged with the wild-type (P = .75). Error bars represent the standard error between the replicates of each treatment. Error bars represent the standard error between the three replicate tanks of each treatment.

each of the gfc operon mutants (P < .05), with each mutant having a survival rate between 89 and 100%. All the fish that were challenged with ML09-119 gfcD survived when monitored for at least 15 days. Due to this, ML09-119 gfcD was chosen for further evaluation in subsequent experiments. Virulence was partially restored upon introduction of the gfc operon into the gfcD or gfcBCD mutants, with a decreased survival

rate (~56%) observed in the complemented gfc operon mutant (Fig. 3). Tukey pairwise comparisons indicated no significant differences between the average survival rate of the fish challenged with the wildtype and the fish challenged with the gfcBCD (pBBC2) complemented mutant (P = .44). The survival rate for the gfcD or gfcBCD mutants complemented with pBBC2 were significantly higher than the wild-type

Fig. 2. Genetic organization of the O-antigen capsule assembly cluster in wild-type ML09-119 and its individual deletion mutants. The predicted gfcABCD operon is flanked by waaL and wzzA of the LPS core and O-antigen biosynthesis gene clusters, respectively. The mutant strains used to study the role of the gfc operon in virulence and other vAh phenotypes are also depicted. 8

Aquaculture 512 (2019) 734334

C.M. Thurlow, et al.

Fig. 3. Percent survival of channel catfish challenged with vAh strain ML09-119, its mutants, and complemented mutant. Percent survival of fish challenged with each gfc operon mutant was significantly increased compared to the wild-type (P < .05). Error bars represent the standard error between the three replicates of each treatment.

(P < .05). In all cases, data are expressed as mean ± SE of three replicate tanks each with 10 channel catfish fingerlings.

3.6. A. hydrophila gfcD exhibits a reduction in its ability to form a biofilm To test whether the absence of gfcD directly affected bacterial biofilm formation, the wild-type ML09-119 (pMJH97), ML09-119 gfcD (pMJH97) and the gfcD complemented mutant ML09-119 gfcD (pBBC2) were quantified for their ability to form a biofilm (Fig. 5). Both the wild-type strain and the complemented mutant exhibited significantly higher biofilm formation compared to the gfcD mutant after 48 h when incubated statically at 30 °C (P < .001) (Fig. 5, Panel A). The wild-type and the complemented mutant were observed to have similar levels of biofilm formation (P = .72). To visualize differences in biofilm formation observed in the biofilm assay, the wild-type ML09-119 (pMJH97), ML09-119 gfcD (pMJH97) and the complemented mutant ML09-119 gfcD (pBBC2) were grown on glass coverslips for 48 h statically at 30 °C and stained with SYBR Gold (Fig. 5, Panel B). Fluorescent microscopy indicated that there was a significant reduction in the number of A. hydrophila gfcD cells attached to the coverslip after 48 h. In contrast, there were no observable differences in bacterial cell attachment between the wild-type and gfcD complemented mutant. These fluorescent micrographs correlate with the microtiter biofilm assay indicating a reduction in biofilm formation by the GfcD-deficient mutant.

3.4. The A. hydrophila gfcD mutant lacks a group 4 capsule Bright field microscopy was conducted to visualize G4C in the wildtype A. hydrophila ML09-119 (pMJH97), ML09-119 gfcD (pMJH97) and the gfcD complemented mutant ML09-119 gfcD (pBBC2). Growth kinetics data indicated that there was no significant difference in growth rates between the wild type and gfcD mutant (Fig. S5). The halo around each cell indicating the presence of the G4C was observed in the wildtype and the gfcD complement but was absent in the gfcD mutant (Fig. 4). No other phenotypic differences were observed. 3.5. The A. hydrophila gfcD mutant lacks buoyancy Because E. coli homologs of gfcB, gfcC and gfcD have been found responsible for G4C assembly and the lack of those homologs have been shown to result in decreased buoyancy in Percoll (Peleg et al. 2005), the buoyancy of A. hydrophila ML09-119, ML09-119 gfcD and the complemented mutant ML09-119 gfcD (pBBC2) was determined. Buoyancy was reduced in the absence of GfcD when compared to the wild-type strain (Fig. S6). The buoyancy was restored in ML09-119 gfcD upon complementation (Fig. S6).

Fig. 4. Modified Maneval's capsule staining of the A. hydrophila ML09-119 (pMJH97) (left), ML09-119 gfcD (pMJH97) (middle), and complemented ML09-119 gfcD (pBBC2) mutant (right). Cells producing capsules result in a halo around each cell after positive Maneval's stains (black arrow). No halo was observed around ML09119 gfcD cells after 48 h of incubation. All cells were observed at 40× magnification. ⁎Asterisks represent treatments that were significantly different when compared to the wild-type (P < .05). 9

Aquaculture 512 (2019) 734334

C.M. Thurlow, et al.

Fig. 5. The biofilm-forming capacity of different A. hydrophila strains. (Panel A) Biofilm formation was reduced in the gfcD mutant relative to its wild-type strain ML09-119. The wild-type strain ML09-119 and its gfcD mutant complemented with gfc operon encoding plasmid pBBC2 were observed to have significantly higher levels of biofilm formation compared to the gfcD deficient mutant (P < .001). There were no significant differences between the wild-type and the gfcD mutant complemented with pBBC2 (P = .72). Error bars represent the standard error between the three replicate cultures for each treatment. (Panel B) Fluorescence microscopy of biofilms formed by the wild-type ML09-119 (pMJH97) (A), ML09-119 gfcD mutant (pMJH97) (B), and ML09-119 gfcD mutant complemented with pBBC2 (C). Observable cell density in each micrograph represents relative attachment of post-washed bacteria.

3.7. LPS-PAGE analysis of vAh wild-type and mutants

3.9. Hemolysis of RBCs

To identify the effect gfcD played on the LPS O-antigen, LPS was extracted from the reference strain AL06-06, wild type strain ML09119, ML09-119 gfcD mutant, ML09-119 wzy::catae mutant, and ML09 waaL::catae mutant and visualized via silver staining, which showed a reduced presence of LPS from the wzy mutant relative to wild-type ML09-119 but no significant differences with the gfcD mutant. In contrast, the waaL mutant was observed to have a reduced number of Oantigen repeats relative to wild-type ML09-119 (Fig. S1).

Among the proteins that were significantly altered in their abundance, the hemolysin Ahh1 was observed to have a strong reduction in abundance in the supernatant of the gfcD mutant relative to that of the WT culture (Fig. 6). Due to this oberservation, we tested the hemolytic activity of the WT in comparison to the gfcD mutant and its complement. Hemolysis was confirmed in every strain after 20 min (P < .05 when compared to the blank) and expressed as % hemolysis relative to a positive control of complete red blood cell lysis in the presence of 0.1% SDS. The hemolytic activity observed in the gfcD mutant was comparable to the WT at the 10, 20 and 40 min timepoints (Fig. 7; P > .8). Under these culture conditions, there was an observed decrease in hemolytic activity of roughly 30% observed in the WT at the 30 min time point when compared to the gfcD mutant; however, there was no observed significant differences in hemolytic activity between the two strains at the final time point (P = .009 and P = .9, respectively). Interestingly, there was a significant increase in hemolytic activity observed in the complemented mutant when compared to the gfcD mutant or WT, with the complemented mutant reaching 100% hemolysis within 10 min, whereas the gfcD mutant and WT only reached 100% hemolysis by 30 to 40 min (P < .05). There was no significant difference in hemolytic activity in the three strains after 40 min (P > .05).

3.8. Secreted proteins The altered phenotypes observed for the gfcD mutant (lack of buoyancy, reduced biofilm formation and attachment to a substrate) were consistent with previous literature on the functions associated with the G4C for other Gram-negative bacteria, but these observed phenotypes did not explain the complete attenuation of virulence observed for the gfcD mutant. Given the predicted role for GfcD in G4C export, we hypothesized that the lack of GfcD might impact the secretion of A. hydrophila virulence factor(s). This hypothesis was tested by conducting an analysis of secreted proteins in the wild-type (WT) and gfcD mutant. Two hundred sixty-eight proteins were identified that were differentially secreted by ML09-119 WT or the gfcD mutant. A reproducibility-optimized test statistic (ROTS) analysis returned 27 proteins that significantly (FDR = 0.1, P = .01) varied in their abundance (Fig. 6). Protein abundance ratios of all 27 ROTS-identified proteins were above the significant fold change threshold (Table 3). Of the 27 proteins that varied significantly in their abundance, five proteins were absent or decreased in the gfcD mutant, while 22 were absent or decreased in the WT secretome. The five proteins absent from the gfcD mutant were all putative A. hydrophila virulence factors, including an adhesin, proteases, and a hemolysin that were secreted by WT at high observed abundance relative to other secreted proteins. Proteins found in greater abundance in the gfcD mutant secretome were predicted to be involved in carbohydrate metabolism, pyrimidine, lysine, and cell wall synthesis, and transport. Of interest were the outer membrane proteins and surface antigen proteins present at higher relative abundance in the gfcD secretome (Table 3).

3.10. Vaccination of channel catfish with ML09-119 gfcD To determine the vaccination efficacy of A. hydrophila ML09-119 gfcD in channel catfish in preventing fish from developing MAS, the survivors of the initial challenge with ML09-119 gfcD were maintained for 21 days post-vaccination (Fig. 8). Following the 21-day acclimation period, survivors were sub-challenged by IP injection with wild-type ML09-119 at a dose of 1.0 × 106 CFU/fish. A. hydrophila ML09-119 gfcD protected channel catfish from MAS with an 89% ± 11% survival rate, which was significantly higher than the 25% ± 14% survival rate of naïve fish challenged with the wild-type (P = .006). There was no significant difference observed between the survival rates of fish challenged with a negative control (TSB alone) and ML09-119 gfcD protected channel catfish (P = .62). In all cases, data is expressed as mean 10

Aquaculture 512 (2019) 734334

C.M. Thurlow, et al.

Fig. 6. Average quantitative protein values of differentially secreted proteins of ML09-119 wild-type and gfcD mutant +/− the standard error of three replicate cultures of each strain. ML09-119 gene-specific locus tags are presented parenthetically.

percent survival ± SE of three replicate tanks. These data indicated that the gfcD mutant can elicit a strong protection against subsequent challenge with vAh.

3.12. Serum resistance in vAh wild-type and mutants Resistance to serum is another important phenotype that has been reported to be modulated by the presence of an LPS O-antigen. A comparison between at ML09-119 and ML09-119 gfcD indicated that they were completely resistant to complement-mediated killing (Fig. 9), with ML09-119 observed to have a 170.9% survival rate and ML09-119 gfcD mutant observed to have a 98.3% survival rate. In contrast, the ML09-119 wzy::cat and ML09-119 waaL::cat mutants were completely sensitive to complement- mediated killing showing zero survival (Fig. 9). A complement-sensitive E. coli strain HB101 was used as a positive control and showed 100% killing due to serum complement. As has been observed in E. coli, the lack or disruption of an LPS O-antigen (as with wzy or waaL mutants) resulted in serum sensitivity, whereas the gfcD mutant maintained serum resistance (Peleg et al. 2005).

3.11. Adaptive immune response in vaccinated fish To determine the relative IgM antibody titers found in the blood serum for gfcD inoculated fish, fish serum obtained from naïve fish, fish challenged with ML09-119 WT or gfcD mutant was subjected to five serial dilutions (1:10, 1:20, 1:40, 1:80 and 1:160), and serum absorbance levels in all three treatments were standardized against a blank (Fig. S7). Averages and standard error of each standardized absorbance levels in the replicate fish serums of each treatment indicated that ML09-119-specific IgM immunoglobins in the fish that were vaccinated with the gfcD mutant were substantially higher than both the positive control (fish that survived the initial encounter with ML09-119) and naïve fish. In addition, the naïve fish showed an average decrease of 93 ± 0.75% in their relative absorbance levels when compared to the absorbance levels of the gfcD vaccinated fish. Fish that survived the initial encounter with ML09-119 were observed to have a 54 ± 4% decrease in absorbance when compared to the fish that were vaccinated with the gfcD mutant. These results indicated significant IgM antibody response to IP inoculation with the gfcD mutant.

4. Discussion The significant roles of O-antigens in the virulence of other bacterial pathogens (Beaz-Hidalgo and Figueras 2013; DeShazer et al. 1998; Goebel et al. 2008; Lerouge and Vanderleyden 2002; Merino et al. 1996; Nagy et al. 2006; West et al. 2005) prompted us to investigate the involvement of the O-antigen biosynthesis and LPS core gene cluster in vAh pathogenesis. The annotation of the ML09-119 O-antigen gene products required for the synthesis and transport of activated sugars to 11

Aquaculture 512 (2019) 734334

C.M. Thurlow, et al.

Table 3 Differentially secreted proteins of ML09-119 WT and gfcD mutant. Reproducibility Z-score = 6.73, FDR = false discovery rate. A protein abundance fold change ≥1.5 was considered significant. Secreted protein

ROTSstatistic

P value

FDR

Protein abundance fold change

Significant experimental group

Quantitative protein value WT

Quantitative protein value gfcD

Hemolysin ahh1 Major adhesin Aha1 Maltoporin A0KHF6 Superoxide dismutase Outer membrane protein A Outer membrane protein F2X2X2 Glyceraldehyde-3-phosphate dehydrogenase Ribulose-phosphate 3-epimerase Oligopeptide ABC transporter Riboflavin biosynthesis protein RibA Major outer membrane protein OmpAII Carbonic anhydrase, family 3 Dihydroorotase Surface antigen family protein Alanine racemase Periplasmic nitrate reductase Acetyl-CoA carboxylase Outer membrane protein P5 Arginine ABC transporter Vitamin B12 transporter BtuB Dihydrodipicolinate synthase Extracellular protease Protease LasA Phospholipid-cholesterol acyltransferase Putative deoxyribonuclease YcfH Basic endochitinase A0KGX5 Probable phosphatase

−8.82 −4.25 2.59 3.73 4.01 3.64 4.63

0.0003 0.0019 0.0070 0.0033 0.0023 0.0036 0.0013

0 0 0 0 0 0 0

6.7 6 2.5 4.2 7.37 5.4 8.4

WT WT gfcD gfcD gfcD gfcD gfcD

122 6 4 3 3 2 2

18 0 11 13 20 11 17

4.68 6.05 3.2 3.79 2.84 3.86 4.41 3.24 2.8 2.75 2.03 2.08 2.29 1.62 −1.72 −1.75 1.77

0.0009 0.0006 0.0047 0.0029 0.0054 0.0026 0.0016 0.0043 0.0057 0.0060 0.0106 0.0099 0.0084 0.0187 0.0176 0.0173 0.0160

0 0 0 0 0 0 0 0 0 0 0.05 0.052 0.055 0.071 0.074 0.076 0.08

12 14 6 10 6 14 11 7 5 4 7 6 4 2.7 1.5 2 1.8

gfcD gfcD gfcD gfcD gfcD gfcD gfcD gfcD gfcD gfcD gfcD gfcD gfcD gfcD WT WT gfcD

1 1 1 1 1 0 0 0 0 0 1 1 0 4 24 2 4

12 14 6 10 6 14 11 7 5 4 7 6 4 10 16 0 7

1.97 −1.99 2

0.0138 0.0130 0.0123

0.083 0.086 0.09

5 2.0 4

gfcD WT gfcD

0 51 0

5 26 4

the outer membrane and ligation of these sugars to the growing O-antigen chain (Hossain et al. 2013). Further, the O-antigen biosynthesis gene cluster contained genes required for the synthesis of nucleotide sugars D-rhamnose, D-mannose, GDP-L-fucose, Qui3NAc and DFucP3NAc. A total of five different glycosyltransferase genes and one acetyltransferase gene were also discovered, which were predicted to assemble the nucleotide sugar repeat onto the membrane lipid undecaprenol pyrophosphate (Und-PP). The O-antigen flippase, O-antigen polymerase (Wzy), and O-antigen ligase (WaaL) gene products were predicted to be present in the ML09-119 O-antigen and the LPS core biosynthetic gene clusters, which suggests that ML09-119 relies on the Wzy-dependent pathway to synthesize the O-antigen, attach it to lipid A, and export the LPS to the outer leaflet of the outer membrane (Raetz and Whitfield 2002; Whitfield 2006). Upon deleting waaL or wzy there was an observed lack of LPS Oantigen for the wzy mutant, whereas the waaL mutant had an

apparently reduced number of O-antigen repeat units (Fig. S1). This was expected given that reduced number of O-antigen repeats would be predicted for a mutant lacking O-antigen ligase activity (Pei et al. 2015). When tested by IP injection in catfish, the ML09-119 waaL::catae mutant showed an attenuation of virulence, whereas the wzy mutant was fully virulent. However, complementation of ML09-119 waaL::catae with the wild-type waaL did not restore virulence to ML09-119 waaL::catae despite multiple attempts. Further analysis of this mutant suggested that there was a polar effect of the waaL mutation on an adjacent gfc operon (see below). We therefore generated a new version of the waaL mutant via recombineering that lacked the presumed polar effect on transcription (waaL::cat mutant) and this was found to be fully virulent in catfish. There remain some unexplained aspects concerning the original waaL::catae mutant, including why it had only a partial reduction of O-antigen repeat units and decreased virulence in fish. However, since the re-make of this mutant did not have attenuated

gfcD gfcD

12

Fig. 7. Hemolytic activity of ML09-119, ML09-119 gfcD and ML09-119 gfcD complemented mutant after 40 min. ML09119, ML09-119 gfcD carry an empty vector that contains the chloramphenicol acetyltransferase (cat) gene. ML09-119 gfcD (pBBC2) contains the same empty vector with the gfcBCD operon.ML09-119 gfcD was insignificantly different in hemolytic activity at time 10, 20, and 40 when compared to the wild-type (P > .8) ML09-119 pBBC2 was significantly different compared to the wild-type and ML09-119 gfcD at ever time point other than time 40 (P < .05). All data has been expressed as the percent of hemolysis of the positive control. All error bars represent the standard error between three replicate cultures of each strain. ⁎ Asterisks represent treatments that were significantly different when compared to naïve fish inoculated with the wildtype (P < .05).

Aquaculture 512 (2019) 734334

C.M. Thurlow, et al.

Fig. 8. Protective effects of ML09-119 gfcD mutant. Naïve fish and survivors of the initial challenge with ML09-119 gfcD (represented as ML09-119 and ML09-119 gfcD, respectively) were challenged with the wild-type ML09-119 after 21 days. Survival rates of fish previously challenged with ML09-119 gfcD were significantly higher than that of naïve fish (P = .006). The negative control represents a population of naïve fish that were challenged with sterile TSB (sham). No significant difference was observed between the survival rates of the negative control (Sham treated) and ML09-119 gfcD (P = .62). Error bars represent the standard error between the three replicates tanks of each treatment. ⁎ Asterisks represent treatments that were significantly different when compared to the complement-sensitive E. coli strain, HB101 (P < .05).

virulence and the principal goal of this study was to determine the gene products that contribute to A. hydrophila virulence, we focused our subsequent experiments on the downstream gfc operon that was hypothesized to have been involved in the attenuated virulence phenotype observed for the first waaL::catae mutant. The annotated genome for A. hydrophila ML09-119 deposited in GenBank predicted three open reading frames, gfcB, gfcC, and gfcD that were upstream of waaL. CDD analysis of this operon indicated homology to gene products of the gfc operon (Fig. S3), which has been identified in at least 29 genera of Gram negative bacteria and been shown in E. coli O127 and O157 pathogenic strains to play a role in capsule assembly (Sathiyamoorthy et al. 2011). The crystal structure of GfcB, an outer membrane-anchored lipoprotein, showed that it has a flattened β-barrel structure. GfcC is a soluble periplasmic protein containing two β-grasp domains (Sathiyamoorthy et al. 2011), and because it has some characteristics comparable to Wza, it may function as an export protein. GfcD is predicted to be an outer membrane lipoprotein with 22-strand beta-barrel that may function as a transport protein for the formed polysaccharide chain (Sathiyamoorthy et al. 2011). Due to the expected requirement of the gfc operon on G4C assembly, we characterized ML09-119 wild-type and the gfcD mutant for capsule formation, biofilm formation, buoyancy, serum resistance and virulence. As expected, the gfcD mutant lacked an observable G4C under bright field microscopy compared to wild-type ML09-119. The absence of a formed capsule was also confirmed by the reduced buoyancy of the gfcD mutant (Peleg et al. 2005), which could be reversed upon complementation (Fig. S6). The waaL and ML09-119 wzy::catae mutants

were observed to have increased susceptibility to fish serum, which was not observed in ML09-119 or ML09-119 gfcD (Fig. 9). This is expected to be attributable to increased complement-mediated killing in the waaL and wzy mutants that is due to increased deposition of complement in cells that have reduced presence or length of O-antigen chains (Goebel et al. 2008; Schiller et al. 1989). We observed reductions in both the presence of the O-antigen and shortened O-antigen chains in both the vAh waaL and wzy mutants (Fig. S1). This reduction was not observed in ML09-119 or ML09-119 gfcD, supporting the idea that GfcD does not play a role in the formation of the LPS O-antigen. Further, there was no observed virulence attenuation in the waaL or wzy mutants, which indicates that expression of the vAh LPS O-antigen is not critical to vAh virulence when IP injected into fish but could play a role in complement resistance and avoidance of the innate immune response (Goebel et al. 2008). Further, the gfcD ortholog in E. coli, yjbH, has also been shown to play a role in secretion of exopolysaccharides (Ferrieres et al. 2007). The pathogenic role of the group 4 capsular (G4C) polysaccharide has not been previously determined in A. hydrophila but has been shown to play a role in virulence in other Gram-negative bacteria, including enabling bacteria to resist innate immune killing and prevent phagocytosis and other immune modulating machinery (Attridge and Holmgren 2009; Caboni et al. 2015; Croxatto et al. 2007; Nakhamchik et al. 2007). To determine the role for the gfc operon in vAh virulence, aquaria challenges were carried out with mutants in gfcB, gfcC, gfcD and gfcBCD. All mutants were observed to have an attenuation of virulence in channel catfish, with ML09-119 gfcD showing a complete attenuation Fig. 9. Survival rate in percent of wild type ML09-119 compared to LPS mutants of ML09-119 (waaL::cat, wzy, and gfcD) after exposure to normal catfish serum. HB101, a complement-sensitive strain of E. coli, was used as a positive control. Error bars represent the standard error between the three replicate cultures for each treatment. ⁎Asterisks represent treatments that were significantly different when compared to the complement-sensitive E. coli HB101 (P < .05).

0b

HB101

ML09-119 waaL::cat

0b

ML09-119 wzy

0b

ML09-119 gfcD

98.26 a

* *

170.88 a

ML09-119 0

50

100 150 Percent Survival

200

250

13

Aquaculture 512 (2019) 734334

C.M. Thurlow, et al.

of virulence whereas mutants in gfcB and gfcC were observed to have significant reductions in virulence compared to wild-type vAh (Fig. 3). Initial in silico examination of the gfc operon predicted the presence of gfcBCD in vAh. Genetic complementation was carried out to confirm that the attenuated virulence was due to disruption of the targeted gene (s), by introducing the intact gfcABCD operon. Partial restoration of virulence was observed after introduction of the plasmid-borne gfcABCD operon into the gfcD and gfcBCD mutants, which was anticipated due to the lack of antibiotic selection for the complementing plasmid in vivo (Fig. 3). Given the phenotypes associated with the gfcD mutant (lack of buoyancy, reduced biofilm formation and attachment to a substrate), we were surprised to observe such a dramatic attenuation of virulence in fish that experienced rapid onset of mortality when IP injected with the WT vAh strain. While G4C formation in other bacterial pathogens has been shown to play a role in biofilm formation and attachment to host tissues (Croxatto et al. 2007), the reduced virulence observed in this study was under challenge conditions that bypassed the fish tegument defenses via IP injection. We therefore considered the hypothesis that the gfc operon, and particularly GfcD that is predicted to be essential for G4C export, affects virulence factor secretion. The observations of altered protein secretion in the gfcD mutant relative to the WT strain lend support to this hypothesis while not providing a mechanistic understanding. Future studies should explore the role of GfcD and the other proteins predicted to be involved in G4C export and assembly in the processes required for protein secretion. Due to the full attenuation of virulence observed for ML09-119 gfcD, it was of interest to challenge surviving fish with wild-type ML09-119 after 21 days of initial mutant challenge. Fish that had received an IP dose of ML09-119 gfcD showed much higher survival rate compared to naïve fish (Fig. 8), indicating that the fish had mounted an adaptive immune response in response to vaccination. This was further supported by the observation of a significant increase in blood serum IgM levels specific to the wild-type ML09-119 antigens (Fig. S7). The serum IgM levels observed from ML09-119 gfcD vaccinated fish were significantly higher than the titers found in naïve fish or from fish that survived the initial encounter with WT vAh. The enhanced adaptive immune response observed for the gfcD mutant is hypothesized to result from the absence of the G4C resulting in greater surface antigen exposure to complement and other immune modulating cells (Attridge and Holmgren 2009; Goebel et al. 2008). This hypothesis was supported by observations from the secretome analysis that revealed significant increases in multiple outer membrane and surface antigen proteins in the gfcD mutant relative to the wild-type, which have been shown to stimulate protective immunity (Abdelhamed et al. 2017; Rahman and Kawai 2000). The results of the secretome analysis also indicated changes in levels of secreted proteins such as hemolysins that are considered virulence factors in A. hydrophila (Beaz-Hidalgo and Figueras 2013; Pang et al. 2015; Rasmussen-Ivey et al. 2016a). Likewise, multiple degradative enzymes were decreased or absent in the gfcD mutant secretome. The changes in secreted virulence factors may contribute to the reduced virulence observed for the gfcD mutant, while the increased host exposure to multiple antigenic bacterial surface proteins may be responsible for the strong adaptive immune response observed in fish vaccinated with the gfcD mutant. The lack of gfcE, etp, and etk homologs in the ML09-119 G4C operon is puzzling. These three genes, encoding an outer membrane helical pore, a tyrosine phosphatase, and tyrosine kinase, follow gfcABCD in the G4C operon of E. coli and Shigella species but also are paralogous to wza, wzb, and wzc elsewhere on the E. coli chromosome and implicated in group 1 capsule (colanic acid) secretion, E. coli K30 capsule secretion, and other capsules (Nadler et al. 2012; Sathiyamoorthy et al. 2011). In E. coli and other species, a Wzc octamer (CapB in Staphylococcus aureus) undergoes cyclic autophosphorylation and dephosphorylation (with Wzb) which may aid in synthesis or secretion of group 1 EPS or capsule through the Wza pore (Nadler et al. 2012; Olivares-Illana et al. 2008; Temel et al. 2013). Although the ML09-119

gfcBCD genes are more homologous to gfcBCD in E. coli, the lack of wza, wzb, and wzc homologs makes the operon more similar to yjbEFGH present in many organisms and apparently responsible for secretion of an unknown exopolysaccharide (Ferrieres et al. 2007). GfcD, a large outer membrane β-barrel protein, may provide the exit portal for ML09119 capsule in lieu of Wza. Understanding how the ML09-119 capsule is secreted will help clarify why the wzc gene is apparently missing in vAh genomes. Supplementary data to this article can be found online at https:// doi.org/10.1016/j.aquaculture.2019.734334. Acknowledgements The authors are thankful for the funding that made this research possible, from the U.S. Department of Agriculture (Grant numbers: #2013-67015-21313, #MIS-371530), the Alabama Agricultural Experiment Station (Grant number ALA021-1-09005) and the State of Alabama Innovation Fund (Grant number 225804). References Abdelhamed, H., Ibrahim, I., Nho, S.W., Banes, M.M., Wills, R.W., Karsi, A., Lawrence, M.L., 2017. Evaluation of three recombinant outer membrane proteins, OmpA1, Tdr, and TbpA, as potential vaccine antigens against virulent Aeromonas hydrophila infection in channel catfish (Ictalurus punctatus). Fish Shellfish Immunol. 66, 480–486. Attridge, S.R., Holmgren, I., 2009. Vibrio cholerae O139 capsular polysaccharide confers complement resistance in the absence or presence of antibody yet presents a productive target for cell lysis: implications for detection of bactericidal antibodies. Microb. Pathog. 47, 314–320. Beaz-Hidalgo, R., Figueras, M.J., 2013. Aeromonas spp. whole genomes and virulence factors implicated in fish disease. J. Fish Dis. 36, 371–388. Bebak, J., Wagner, B., Burnes, B., Hanson, T., 2015. Farm size, seining practices, and salt use: risk factors for Aeromonas hydrophila outbreaks in farm-raised catfish, Alabama, USA. Prev. Vet. Med. 118, 161–168. Beissbarth, T., Hyde, L., Smyth, G.K., Job, C., Boon, W.M., Tan, S.S., Scott, H.S., Speed, T.P., 2004. Statistical modeling of sequencing errors in SAGE libraries. Bioinformatics 20, 31–39. Caboni, M., Pedron, T., Rossi, O., Goulding, D., Pickard, D., Citiulo, F., MacLennan, C.A., Dougan, G., Thomson, N.R., Saul, A., Sansonetti, P.J., Gerke, C., 2015. An O antigen capsule modulates bacterial pathogenesis in Shigella sonnei. PLoS Pathog. 11. Croxatto, A., Lauritz, J., Chen, C., Milton, D.L., 2007. Vibrio anguillarum colonization of rainbow trout integument requires a DNA locus involved in exopolysaccharide transport and biosynthesis. Environ. Microbiol. 9, 370–382. DeShazer, D., Brett, P.J., Woods, D.E., 1998. The type II O-antigenic polysaccharide moiety of Burkholderia pseudomallei lipopolysaccharide is required for serum resistance and virulence. Mol. Microbiol. 30, 1081–1100. Edwards, R.A., Keller, L.H., Schifferli, D.M., 1998. Improved allelic exchange vectors and their use to analyze 987P fimbria gene expression. Gene 207, 149–157. Ferrieres, L., Aslam, S.N., Cooper, R.M., Clarke, D.J., 2007. The yjbEFGH locus in Escherichia coli K-12 is an operon encoding proteins involved in hexopolysaccharide production. Microbiology 153, 1070–1080. Gibson, D.L., White, A.P., Snyder, S.D., Martin, S., Heiss, C., Azadi, P., Surette, M., Kay, W.W., 2006. Salmonella produces an O-antigen capsule regulated by AgfD and important for environmental persistence. J. Bacteriol. 188, 7722–7730. Goebel, E.M., Wolfe, D.N., Elder, K., Stibitz, S., Harvill, E.T., 2008. O antigen protects Bordetella parapertussis from complement. Infect. Immun. 76, 1774–1780. Herrero, M., de Lorenzo, V., Timmis, K.N., 1990. Transposon vectors containing nonantibiotic resistance selection markers for cloning and stable chromosomal insertion of foreign genes in gram-negative bacteria. J. Bacteriol. 172, 6557–6567. Horton, R.M., Hunt, H.D., Ho, S.N., Pullen, J.K., Pease, L.R., 1989. Engineering hybrid genes without the use of restriction enzymes: gene splicing by overlap extension. Gene 77, 61–68. Hossain, M.J., Waldbieser, G.C., Sun, D.W., Capps, N.K., Hemstreet, W.B., Carlisle, K., Griffin, M.J., Khoo, L., Goodwin, A.E., Sonstegard, T.S., Schroeder, S., Hayden, K., Newton, J.C., Terhune, J.S., Liles, M.R., 2013. Implication of lateral genetic transfer in the emergence of Aeromonas hydrophila isolates of epidemic outbreaks in channel catfish. PLoS One 8. Hossain, M.J., Sun, D., McGarey, D.J., Wrenn, S., Alexander, L.M., Martino, M.E., Xing, Y., Terhune, J.S., Liles, M.R., 2014. An Asian origin of virulent Aeromonas hydrophila responsible for disease epidemics in United States-farmed catfish. mBio 5. Hossain, M.J., Thurlow, C.M., Sun, D., Nasrin, S., Liles, M.R., 2015. Genome modifications and cloning using a conjugally transferable recombineering system. Biotechnol. Rep. 8, 24–35. Hsieh, P.F., Lin, T.L., Yang, F.L., Wu, M.C., Pan, Y.J., Wu, S.H., Wang, J.T., 2012. Lipopolysaccharide O1 antigen contributes to the virulence in Klebsiella pneumoniae causing pyogenic liver abscess. PLoS One 7, e33155. Kumar, S., Stecher, G., Tamura, K., 2016. MEGA7: molecular evolutionary genetics analysis version 7.0 for bigger datasets. Mol. Biol. Evol. 33, 1870–1874. Lerouge, I., Vanderleyden, J., 2002. O-antigen structural variation: mechanisms and

14

Aquaculture 512 (2019) 734334

C.M. Thurlow, et al. possible roles in animal/plant-microbe interactions. FEMS Microbiol. Rev. 26, 17–47. Locke, J.B., Colvin, K.M., Datta, A.K., Patel, S.K., Naidu, N.N., Neely, M.N., Nizet, V., Buchanan, J.T., 2007. Streptococcus iniae capsule impairs phagocytic clearance and contributes to virulence in fish. J. Bacteriol. 189, 1279–1287. Ludwig, M.R., Kojima, K., Bowersock, G.J., Chen, D.Q., Jhala, N.C., Buchsbaum, D.J., Grizzle, W.E., Klug, C.A., Mobley, J.A., 2016. Surveying the serologic proteome in a tissue-specific kras(G12D) knockin mouse model of pancreatic cancer. Proteomics 16, 516–531. Lukashin, A.V., Borodovsky, M., 1998. GeneMark.Hmm: new solutions for gene finding. Nucleic Acids Res. 26, 1107–1115. Magee, A.D., Yother, J., 2001. Requirement for capsule in colonization by Streptococcus pneumoniae. Infect. Immun. 69, 3755–3761. Marolda, C.L., Lahiry, P., Vines, E., Saldias, S., Valvano, M.A., 2006. Micromethods for the characterization of lipid A-core and O-antigen lipopolysaccharide. Methods Mol. Biol. 347, 237–252. Merino, S., Rubires, X., Aguilar, A., Tomas, J.M., 1996. The O:34-antigen lipopolysaccharide as an adhesin in Aeromonas hydrophila. FEMS Microbiol. Lett. 139, 97–101. Merino, S., Aguilar, A., Rubires, X., Abitiu, N., Regue, M., Tomas, J.M., 1997. The role of the capsular polysaccharide of Aeromonas hydrophila serogroup O:34 in the adherence to and invasion of fish cell lines. Res. Microbiol. 148, 625–631. Nadler, C., Koby, S., Peleg, A., Johnson, A.C., Suddala, K.C., Sathiyamoorthy, K., Smith, B.E., Saper, M.A., Rosenshine, I., 2012. Cycling of Etk and Etp phosphorylation states is involved in formation of group 4 capsule by Escherichia coli. PLoS One 7. Nagy, G., Danino, V., Dobrindt, U., Pallen, M., Chaudhuri, R., Emody, L., Hinton, J.C., Hacker, J., 2006. Down-regulation of key virulence factors makes the Salmonella enterica serovar Typhimurium rfaH mutant a promising live-attenuated vaccine candidate. Infect. Immun. 74, 5914–5925. Nakhamchik, A., Wilde, C., Rowe-Magnus, D.A., 2007. Identification of a wzy polymerase required for group IV capsular polysaccharide and lipopolysaccharide biosynthesis in Vibrio vulnificus. Infect. Immun. 75, 5550–5558. Old, W.M., Meyer-Arendt, K., Aveline-Wolf, L., Pierce, K.G., Mendoza, A., Sevinsky, J.R., Resing, K.A., Ahn, N.G., 2005. Comparison of label-free methods for quantifying human proteins by shotgun proteomics. Mol. Cell. Proteomics 4, 1487–1502. Olivares-Illana, V., Meyer, P., Bechet, E., Gueguen-Chaignon, V., Soulat, D., LazeregRiquier, S., Mijakovic, I., Deutscher, J., Cozzone, A.J., Laprevote, O., Morera, S., Grangeasse, C., Nessler, S., 2008. Structural basis for the regulation mechanism of the tyrosine kinase CapB from Staphylococcus aureus. PLoS Biol. 6, 1321–1331. O'Toole, G.A., 2011. Microtiter dish biofilm formation assay. J. Vis. Exp. 30 (47), 2437. Pang, M.B., Jiang, J.W., Xie, X., Wu, Y.F., Dong, Y.H., Kwok, A.H.Y., Zhang, W., Yao, H.C., Lu, C.P., Leung, F.C., Liu, Y.J., 2015. Novel insights into the pathogenicity of epidemic Aeromonas hydrophila ST251 clones from comparative genomics. Sci. Rep. 5. Pei, D., Jiang, J., Yu, W., Kukutla, P., Uentillie, A., Xu, J., 2015. The waaL gene mutation compromised the inhabitation of Enterobacter sp. Ag1 in the mosquito gut environment. Parasit. Vectors 8, 437. https://doi.org/10.1186/s13071-015-1049-1. PubMed PMID: 26306887; PubMed Central PMCID: PMC4549878. Peleg, A., Shifrin, Y., Ilan, O., Nadler-Yona, C., Nov, S., Koby, S., Baruch, K., Altuvia, S., Elgrably-Weiss, M., Abe, C.M., Knutton, S., Saper, M.A., Rosenshine, I., 2005. Identification of an Escherichia coli operon required for formation of the O-antigen capsule. J. Bacteriol. 187, 5259–5266.

Pridgeon, J.W., Klesius, P.H., 2011. Virulence of Aeromonas hydrophila to channel catfish Ictaluras punctatus fingerlings in the presence and absence of bacterial extracellular products. Dis. Aquat. Org. 95, 209–215. Pursiheimo, A., Vehmas, A.P., Afzal, S., Suomi, T., Chand, T., Strauss, L., Poutanen, M., Rokka, A., Corthals, G.L., Elo, L.L., 2015. Optimization of statistical methods impact on quantitative proteomics data. J. Proteome Res. 14, 4118–4126. Raetz, C.R.H., Whitfield, C., 2002. Lipopolysaccharide endotoxins. Annu. Rev. Biochem. 71, 635–700. Rahman, M.H., Kawai, K., 2000. Outer membrane proteins of Aeromonas hydrophila induce protective immunity in goldfish. Fish Shellfish Immunol. 10, 379–382. Rasmussen-Ivey, C.R., Figueras, M.J., McGarey, D., Liles, M.R., 2016a. Virulence factors of Aeromonas hydrophila: in the wake of reclassification. Front. Microbiol. 7. Rasmussen-Ivey, C.R., Hossain, M.J., Odom, S.E., Terhune, J.S., Hemstreet, W.G., Shoemaker, C.A., Zhang, D.H., Xu, D.H., Griffin, M.J., Liu, Y.J., Figueras, M.J., Santos, S.R., Newton, J.C., Liles, M.R., 2016b. Classification of a hypervirulent Aeromonas hydrophila pathotype responsible for epidemic outbreaks in warm-water fishes. Front. Microbiol. 7. Santos, J.A., Gonzalez, C.J., Otero, A., Garcia-Lopez, M.L., 1999. Hemolytic activity and siderophore production in different Aeromonas species isolated from fish. Appl. Environ. Microbiol. 65, 5612–5614. Sathiyamoorthy, K., Mills, E., Franzmann, T.M., Rosenshine, I., Saper, M.A., 2011. The crystal structure of Escherichia coli group 4 capsule protein GfcC reveals a domain organization resembling that of Wza. Biochemistry 50, 5465–5476. Schiller, N.L., Hatch, R.A., Joiner, K.A., 1989. Complement activation and C3 binding by serum-sensitive and serum-resistant strains of Pseudomonas aeruginosa. Infect. Immun. 57, 1707–1713. Shibata, A., Goto, Y., Saito, H., Kikuchi, T., Toda, T., Taguchi, S., 2006. Comparison of SYBR green I and SYBR gold stains for enumerating bacteria and viruses by epifluorescence microscopy. Aquat. Microb. Ecol. 43, 223–231. Simon, R., Priefer, U., Puhler, A., 1983. A broad host range mobilization system for in vivo genetic engineering: transposon mutagenesis in gram negative bacteria. Nat. Biotechnol. 1, 784–791. Suomi, T., Seyednasrollah, F., Jaakkola, M.K., Faux, T., Elo, L.L., 2017. ROTS: an R package for reproducibility-optimized statistical testing. PLoS Comput. Biol. 13. Tekedar, H.C., Waldbieser, G.C., Karsi, A., Liles, M.R., Griffin, M.J., Vamenta, S., Sonstegard, T., Hossain, M., Schroeder, S.G., Khoo, L., Lawrence, M.L., 2013. Complete genome sequence of a channel catfish epidemic isolate, Aeromonas hydrophila strain ML09-119. Genome Announc. 1. Temel, D.B., Dutta, K., Alphonse, S., Nourikyan, J., Grangeasse, C., Ghose, R., 2013. Regulatory interactions between a bacterial tyrosine kinase and its cognate phosphatase. J. Biol. Chem. 288, 15212–15228. Weatherly, D.B., Atwood, J.A., Minning, T.A., Cavola, C., Tarleton, R.L., Orlando, R., 2005. A heuristic method for assigning a false-discovery rate for protein identifications from mascot database search results. Mol. Cell. Proteomics 4, 762–772. West, N.P., Sansonetti, P., Mounier, J., Exley, R.M., Parsot, C., Guadagnini, S., Prevost, M.C., Prochnicka-Chalufour, A., Delepierre, M., Tanguy, M., Tang, C.M., 2005. Optimization of virulence functions through glucosylation of Shigella LPS. Science 307, 1313–1317. Whitfield, C., 2006. Biosynthesis and assembly of capsular polysaccharides in Escherichia coli. Annu. Rev. Biochem. 75, 39–68.

15