lable at ScienceDirect

Fish & Shellfish Immunology 26 (2009) 293–304

Contents lists avai

Fish & Shellfish Immunology

journal homepage: www.elsevier .com/locate / fs i

Characterization of the interferon genes in homozygous rainbow troutreveals two novel genes, alternate splicing and differential regulationof duplicated genes

Maureen K. Purcell a,*, Kerry J. Laing b, James C. Woodson a, Gary H. Thorgaard c, John D. Hansen a

a U.S. Geological Survey, Western Fisheries Research Center, 6505 NE 65th St., Seattle, WA 98034, USAb Fred Hutchinson Cancer Research Institute, 1100 Fairview Ave N., Seattle, WA 98109, USAc School of Biological Sciences, Washington State University, Pullman, WA 99164, USA

a r t i c l e i n f o

Article history:Received 25 September 2008Received in revised form24 November 2008Accepted 25 November 2008Available online 3 December 2008

Keywords:Hot Creek troutInfectious hematopoietic necrosis virusDNA vaccinePoly I:CQuantitative reverse transcriptase PCRType I interferonInterferon gamma

* Corresponding author. Tel.: þ1 206 526 6282x252E-mail address: mpurcell@usgs.gov (M.K. Purcell).

1050-4648/$ – see front matter � 2008 Elsevier Ltd.doi:10.1016/j.fsi.2008.11.012

a b s t r a c t

The genes encoding the type I and type II interferons (IFNs) have previously been identified in rainbowtrout and their proteins partially characterized. These previous studies reported a single type II IFN(rtIFN-g) and three rainbow trout type I IFN genes that are classified into either group I (rtIFN1, rtIFN2) orgroup II (rtIFN3). In this present study, we report the identification of a novel IFN-g gene (rtIFN-g2) anda novel type I group II IFN (rtIFN4) in homozygous rainbow trout and predict that additional IFN genes orpseudogenes exist in the rainbow trout genome. Additionally, we provide evidence that short and longforms of rtIFN1 are actively and differentially transcribed in homozygous trout, and likely arose due toalternate splicing of the first exon. Quantitative reverse transcriptase PCR (qRT-PCR) assays weredeveloped to systematically profile all of the rainbow trout IFN transcripts, with high specificity at anindividual gene level, in naı̈ve fish and after stimulation with virus or viral-related molecules. Cloned PCRproducts were used to ensure the specificity of the qRT-PCR assays and as absolute standards to assesstranscript abundance of each gene. All IFN genes were modulated in response to Infectious hematopoieticnecrosis virus (IHNV), a DNA vaccine based on the IHNV glycoprotein, and poly I:C. The most inducible ofthe type I IFN genes, by all stimuli tested, were rtIFN3 and the short transcript form of rtIFN1. Geneexpression of rtIFN-g1 and rtIFN-g2 was highly up-regulated by IHNV infection and DNA vaccination butrtIFN-g2 was induced to a greater magnitude. The specificity of the qRT-PCR assays reported here will beuseful for future studies aimed at identifying which cells produce IFNs at early time points after infection.

� 2008 Elsevier Ltd. All rights reserved.

1. Introduction

The interferons are powerful pleiotropic cytokines that areinduced in response to viruses by many different cell types andhave diverse roles in regulating the immune system. Mammalianspecies possess multiple type I IFN genes, which are organized intogene families, such as IFN-a, b, 3, k, s and u [1]. The selectiveadvantage provided by the extensive duplication of type I IFN genesin mammals is not known. Type III IFN is a newly characterized IFNfamily (IFN-l) that appears functionally similar to type I IFN and isproduced by a variety of cell types [2]. Type I and III IFNs are distinctfrom type II IFN (IFN-g), which is produced by a restricted set ofimmune cells [1]. IFN regulates the transcription of >300 cellulargenes, known collectively as the IFN stimulated genes (ISGs) [3,4].

Type I and type II IFN genes have been reported for a range offinfish species, including fugu (Takifugu rubripes), common carp

; fax: þ1 206 526 6654.

All rights reserved.

(Cyprinus carpio), zebrafish (Danio rerio), channel catfish (Ictaluruspunctatus), Atlantic salmon (Salmo salar) and rainbow trout(Oncorhynchus mykiss) [5–13]. It has been hypothesized that theteleost type I IFNs are homologous to mammalian type III IFN [8];however, since this hypothesis is currently disputed, we will followZou et al. [13] in calling this gene family the type I IFNs.

Previous studies have identified three type I IFN genes in rainbowtrout (rtIFN1, rtIFN2 and rtIFN3) [13]. The rtIFN1 and rtIFN2 genesare highly similar in sequence, encode a protein with two cysteineresidues and are classified as group I type I IFNs [13]. The rtIFN3 geneencodes a protein that possesses 4 cysteine residues and as such, isclassified as a type I group II IFN [13]. Recombinant rtIFN1 and rtIFN2display potent anti-viral activity in vitro while recombinant rtIFN3has minimal anti-viral activity [13]. In contrast to type I IFN, onlya single rainbow trout IFN-g gene has been reported to date [12].However, other finfish species (zebrafish, channel catfish andcommon carp) possess at least two IFN-g genes [6,10,14].

Infectious hematopoietic necrosis virus (IHNV) is a negative-sensesingle stranded RNA virus within the genus Novirhabdovirus [15].

mailto:mpurcell@usgs.govwww.sciencedirect.com/science/journal/10504648http://www.elsevier.com/locate/fsi

Table 1Sequencing primer sets, real-time PCR primer and probe sequences and GenBankaccession numbers.

Transcript Name Sequence(50-30) Accession

Sequencing primersIFN1-Long 911-F28 gaagactacggaacaacatttcg AJ580911

911-R1132 tgtacaaaatacgtttttattcacaIFN1-Short 890-F1 gaaactcatctggataactaacagcga AY788890

890-R737 acatgtacaaaagggaaatacgaaataIFN1-Promoter SasaIFNA1/A2 F2 accaaggcctgtatttattaagcat DQ354152

OmyIFN1 R119 gcttaagtgaccgaagtggtgtt FJ477854IFN2 754-F1 tggaaaagctaaaagcaaaataaac AJ582754

754-R805 gcacaatacatttttattcacaatttcIFN3 738-F1 agctgttccattcaagttctt AM235738

738-R805 gttgtacgtattttttatttaattIFN-g IFN-g F7 caccgattgaggactattgag AJ616215

IFN-g R640 ctacacctagccttcacagtagaIFN-g Intron 1 IFN-g F5 ggctggatgactttaggatg AM489415

IFN-g R4 cgttgaacagctggtccttgqRT-PCR AssaysIFN1-Long IFN1L 138F cacgcgaagttattagcagttgaa FJ184370

IFN1L 253R aaattatagttgaaccacaatgaaatattattcIFN1L 164T 60FAM-caaagctcgcgaatagcctattc

tcgc-TAMRAIFN1-Short IFN1S 24F gcgaaacaaactgctatttacaatgtata FJ184371

IFN1S 119R tcacagcaatgacacacgctcIFN1S 58T 60FAM-cagagctggagttgtatttttctta

ttatttgcagtatgc-TAMRAIFN1-Total IFNIT 131R ttcttgaagtaccgtttcagtctcctct FJ184370

IFN1T 28F ctggacgatttcctcaacattctagaa FJ184371IFN1T 69T 60FAM-ccttaattcctgtgtatcacctgccat

gaaacc-TAMRAIFN2 IFN2 5F aaagctaaaagcaaaataaacagctctt FJ184369

IFN2 107R cagcaatgacacacactctgcaIFN2 47T 60FAM-tgcagagttggacgtgtctttttcttat

tctttg-TAMRAIFN3/4a IFN3 270F tggaggctatgcgatatgtgg FJ184372

IFN3 370R acgatatatgacgttttggaacatgttIFN3 317T-MGB MGB-60FAM-tctgtcacgtggaacaa-NFQ

IFN-g1 IFN-g1 299F tcggccagatgctgaacc FJ184374IFN-g1 402R cctccccaggaaatagtgtttcIFN-g1 333T-MGB MGB-60FAM-aatgattgagagtct

gaaata-NFQIFN-g2 IFN-g2 207F gttgatgagtgtggttctggacg FJ184375

IFN-g2 329R ttctgggtctcctgaaccttccIFN-g2 279T 60FAM-agtgagggagaggctggaccag

gtcaag-TAMRA

a This assay was designed against IFN3 but does not discriminate between IFN3and IFN4 (FJ184373).

M.K. Purcell et al. / Fish & Shellfish Immunology 26 (2009) 293–304294

We have previously shown in trout that IHNV infection or vacci-nation with a DNA vaccine encoding the IHNV glycoprotein rapidlyinduces the type I and II IFN pathways [16,17]. However, in theseprevious studies the rtIFN1 gene (alias IFN-a1) was the only type IIFN investigated. The goal of the present study was to develophighly specific quantitative reverse transcriptase PCR (qRT-PCR)assays for all known rainbow trout IFN genes and systematicallyprofile these genes following vaccination, IHNV infection and polyI:C stimulation. During the course of this study, we found evidenceof additional transcript forms. To eliminate possible allelic variants,we amplified and sequenced the IFN genes in homozygous rainbowtrout. Here, we confirm the identification of two new rainbow troutIFN genes (a second type I group II IFN (rtIFN4) and a second IFN-ggene (rtIFN-g2)), and report the development and validation ofspecific qRT-PCR assays for rainbow trout IFN gene families. Tran-scriptional profiles of these genes following viral infection orstimulation with viral-related molecules are presented.

2. Materials and methods

2.1. Homozygous rainbow trout

Specific pathogen free homozygous rainbow trout from the HotCreek and OSU strains were obtained from Washington StateUniversity trout hatchery. The trout lines were created by andro-genesis as previously described [18]. The trout were transferred toWestern Fisheries Research Center, reared in sand-filtered, UV-treated freshwater at 15 �C and fed a commercial semi-moist pelletdiet (Bioproducts). Three unhandled homozygous trout wereeuthanized with buffered tricaine methanesulfonate (MS222;Argent Chemical Laboratories) and anterior kidney, heart, posteriorkidney, spleen, gill, skin, anterior intestine, posterior intestine, liver,muscle, ovary and brain were removed, sampled into RNAlater(Qiagen Inc.) and stored at �80 �C until extraction.

2.2. DNA vaccination and infection of homozygous trout

The construction and efficacy of the pIHNw-G DNA vaccine havebeen previously reported [19–22]. Hot Creek trout, each weighingapproximately 83 g, were injected with the IHNV DNA vaccine(pIHNw-G) vaccine plasmid or controls as previously described[17]. Briefly, trout were intra-muscularly (I.M.) injected with 50 mlvolumes of phosphate buffered saline (PBS), 10 mg of pcDNA3.1vector DNA diluted in PBS or 10 mg of pIHNw-G vaccine in PBS. Eachfish was sampled at 7 days post-vaccination (p.v.). Tissues from theI.M. site and anterior kidney were snap-frozen in liquid nitrogenand stored at �80 �C until RNA extraction.

IHNV infection of homozygous trout, weighing approximately300 g, has been previously described [16]. Briefly, 1 � 106 plaqueforming units (pfu) of IHNV (WRAC strain; ATCC VR-1392) in 100 mlPBS was injected intraperitoneally (I.P.); control fish received onlythe PBS. Five fish were sacrificed from each group at days 3, 7 and28 post-injection and the anterior kidney was sampled, frozen inliquid nitrogen and stored at �80 �C until RNA extraction.

2.3. Poly I:C stimulation of anterior kidney leukocytes in vitrofrom non-homozygous fish

Rainbow trout were obtained from Clear Spring Foods Inc. (Buhl,ID) and held at a constant temperature of 15 �C until use, as describedabove. Poly I:C stimulation of anterior kidney leukocyte obtainedfrom non-homozygous trout has been previously described [23].Briefly, anterior kidney tissue from five individual animals wasaseptically removed and leukocytes were enriched by centrifugationin a discontinuous (34%/51%) Percoll gradient (Sigma–Aldrich) [24].The leukocytes were adjusted to 2 � 107 cells/mL and 100 mL of cells

were plated into a 96 well tissue-culture plate and 5 mg/mL of poly I:C(Sigma–Aldrich) were added to the appropriate wells and incubatedfor 6, 12 or 24 h; a 50 mg/mL dose was incubated for 24 h only.

2.4. RNA extraction, cDNA synthesis and sequencing

RNA was extracted using the RNeasy Mini-Kit with in-columnDNAse I treatment following manufacturer’s instructions (QiagenInc.). RNA was used to synthesize cDNA as previously described [25]using 1 mg RNA to initiate cDNA synthesis. Primers designed againstpreviously characterized rainbow trout IFN sequences [12] (Table 1)were used to amplify the genes or gene products from Hot Creekhomozygous trout RNA. All conventional PCRs were performedusing the following parameters: 94 �C for 30 s, 55 �C for 30 s and72 �C for 1 min for all primers except those for amplifying IFN-gintron 1 [13], which used an annealing temperature of 63 �C.Amplified products were subjected to electrophoresis on a 1%agarose gel (1.5% agarose gel for intron primers) and PCR productswere cloned using the Topo TA cloning kit (Invitrogen). Clones werescreened by PCR with the M13 forward and reverse primers andrecombinant clones were sequenced as previously described [25].Sequences were aligned visually with Sequencher V.4.5 (GeneCodes Inc.). Multiple alignments of cDNA and amino acid sequences

M.K. Purcell et al. / Fish & Shellfish Immunology 26 (2009) 293–304 295

were performed using ClustalW2 [26] and formatted using thealignment editor of MEGA 4 [27]. Phylogenetic comparisons wereperformed from Clustal generated multiple alignments of aminoacid sequences using MEGA 4 with Poisson correction andcomplete deletion of gaps [27]. Tree topologies were validated bybootstrapping 1000 times.

2.5. Absolute quantitative reverse transcriptase PCR

Primer and probe sequences used for real-time qRT-PCR areprovided in Table 1. Our standard qRT-PCR methodology using theABI 7900HT (Applied Biosystems) has been described in detailpreviously [25]. Plasmids encoding the gene of interest were usedto create standard curves to estimate absolute mRNA copy number.Copy number was determined by dividing Avogadro’s number bythe molecular weight (mw) of the plasmid (mw ¼ (number ofnucleotides � 607.4) þ 157.9). CT values were converted to copynumber by finding the slope (m) and y-intercept (b) of the line ofbest fit and using the equation: copy number ¼ antilog(CT � b)/m.The amplification efficiency of the plasmids was calculated usingthe formula E ¼ (10�1/m � 1) � 100 (Applied Biosystems Inc. Guideto Performing Relative Quantification of Gene Expression; availablefrom www.appliedbiosystems.com; accessed September 9, 2008).The expression levels of specific IFN gene products were normal-ized against expression levels of a housekeeping gene, acidicribosomal phosphoprotein P0 (ARP) [25,28]. The calculation of foldchange relative to control groups has been previously described[25]. Briefly, fold increase was calculated as (Ts/Tn)/(Cs/Cn) where Tsequals the treated sample assayed for the specific gene and Tnequals the treated sample assayed for the normalizer gene (ARP)and Cs and Cn equals the control group with the specific andnormalizing gene, respectively. To assess statistical significance, weused either a Mann–Whitney U test (two experimental groups) ora Kruskal–Wallis test followed by a Dunn’s post-hoc test (threeexperimental groups) to compare groups. Statistical analyses ofqRT-PCR data was performed using GraphPad InStat V.3.01 statis-tics package (GraphPad Software, Inc.). P values less than 0.05 wereconsidered significant.

3. Results

3.1. Gene sequencing and qRT-PCR assay development

Confirmation of IFN gene expression was performed ina homozygous strain of rainbow trout by verifying the productionof transcripts for each speculative IFN gene family member. Duringsequence confirmation experiments, alternate transcripts andduplicated genes were identified (described below for each gene).Gene specific qRT-PCRs were developed and used in an initialexperiment to assess baseline expression levels for each IFN innaı̈ve trout. Carefully quantified plasmids were used as absolutestandards and verified that all assays had similar PCR efficienciesranging from 81 to 100%. The qRT-PCR data generated for unhan-dled fish (baseline expression) is presented as unnormalized values(mRNA copy number per reaction). The stimulation experiments(vaccine, virus or poly I:C) are presented as normalized valuesrelative to the ARP housekeeping gene. The normalized ratiosreflect absolute transcript abundance and comparisons can bemade across assays (transcripts) and a ratio of 1 would indicateequal expression of both the specific IFN gene and the house-keeping gene. To facilitate these comparisons graphically, the y-axisscale was kept consistent for each transcriptional regulationexperiment unless otherwise noted. A fold change estimate isprovided for the treatment group(s) relative to the mock control foreach experiment. DNAse I treatment of cDNA samples provedeffective as no detectable amplification was observed for any

qRT-PCR in controls containing RNA but lacking reverse transcrip-tase enzyme.

3.1.1. IFN1-evidence of alternate splicingShort and long transcripts of rtIFN1 were evident in the

GenBank databases (long transcript – AJ580911 and short transcript– AY788890). Primers based on these sequences confirmed that theshort and long transcripts were expressed in an individual homo-zygous trout (FJ184370, FJ184371). Based on the available genomicrtIFN1 sequence (AM489415) and the upstream promoter region(FJ477854) (Fig. 1A), we predicted the short and long transcriptslikely arise from alternate 50 exon usage (Fig. 1B). The short tran-script starts with exon 1a and joins all subsequent exons to encodea deduced 176 amino acid protein with a putative signal peptide.The long transcript splices exon 1a0 to exon 1b leading to a shorterdeduced 160 amino acid protein lacking the putative signal peptide.A likely cryptic intron exists between exons 1a0 and 1a and thisregion was highly similar to the proximal promoter region from theAtlantic salmon type I IFN (SasaIFNA1) (Fig. 1A). The proximalpromoter contains predicted NF-kB binding site, two interferonregulatory factor (IRF) binding sites and a TATA box (Fig. 1A).SasaIFNA1 has a distal promoter that leads the production of a longtranscript. Overall, there was high nucleotide similarity betweenSasaIFNA1 and rtIFN1 in the distal promoter region except fora 40 bp deletion in rtIFN1 where the SasaIFNA1 distal promoterstarts (Fig. 1A). Upstream of the distal promoter is a degenerateNF-kB binding site (Fig. 1A).

Three qRT-PCR assays were developed to specifically assess (1)rtIFN1-total (both short and long transcript forms), (2) the longtranscript form and (3) the short transcript form (Table 1). Plasmidsencoding the two rtIFN1 transcript forms were used to verify thespecificity of the assays and the plasmid DNA served as absolutestandards. Constitutive expression of the housekeeping gene ARP isprovided as a reference (Fig. 2A). The cDNA reactions were initiatedwith 1 mg of total RNA but ARP expression varied across tissues(Fig. 2A). The rtIFN1-total assay detected constitutive expression inall tissues tested with the highest expression in the gill (Fig. 2B).The relative constitutive expression of the long and short transcriptforms across tissues was relatively equal (Fig. 2C and D). The sum ofthe mRNA copies of the rtIFN1-long and rtIFN1-short transcriptswas less than the rtIFN1-total mRNA copies suggesting that the IFNtotal assay was detecting additional transcripts. This additionalamplification was not due to genomic DNA contamination becauseno detectable amplification was observed in cDNA control samplescontaining RNA but lacking reverse transcriptase. Additionally,validation using cloned IFNs confirmed that, although the rtIFN1-total assay detects the long and short rtIFN1 transcript forms withequal efficiency and does not detect rtIFN2 and rtIFN3 (data notshown).

3.1.2. rtIFN2IFN2 from homozygous trout was sequenced and a single tran-

script form was observed (FJ184372). There was no evidence ofshort or long forms in the databases and we did not conduct 50

RACE to determine if they were present in homozygous trout. Thededuced amino acid sequence for rtIFN2 contains a predicted signalpeptide similar to the rtIFN1-short transcript. We developed a qRT-PCR assay for this gene that was specific for rtIFN2 (Table 1). TheqRT-PCR assay was designed to the 50 untranslated region (UTR) toreduce the likelihood that alternative transcript forms would bedetected if they existed. Constitutive expression of rtIFN2 was lowoverall, with the highest expression in the gill and spleen (Fig. 2E).

3.1.3. rtIFN3: evidence of gene duplicationSequencing, in homozygous trout, revealed two different

rtIFN3 transcripts which we have designated as rtIFN3 (FJ184372)

http://www.appliedbiosystems.com

A

Short transcript with signal peptide

Long transcript lacking signal peptideATG

Exon 2 Exon 3 Exon 4Exon 1a’ Exon 1b

Exon 1a’ Exon 2 Exon 3 Exon 4Exon 1a Exon 1b

ATG

Exon 2 Exon 3 Exon 4Exon 1a Exon 1b

Gene

Exon 5*

Exon 5

Exon 5

B

Short transcriptstart codon

Long transcriptstart codon

Fig. 1. Promoter region of rtIFN1 and hypothetical alternate splicing of exon 1. (A) Alignment of the rtIFN1 (FJ477854) upstream gene sequence with corresponding Atlantic salmonregion (SasaIFNA1; DQ354152). Dots represent conserved nucleotides in the salmon gene relative to the trout gene, and dashes represent gaps in the alignment. The rtIFN1 sequencepartially overlaps the SasaIFNA1 distal promoter (shaded) and contains predicted APF/c-Jun and IRF sites (in uppercase bold). Immediately upstream of this area is a degenerate NF-kB site. The beginning positions of rainbow trout exons 1a0 , 1a and 1b are indicated by white triangles, and the end of exon 1a0 is marked with a black triangle. The rainbow troutputative cryptic intron (italicized) aligns with the region identified as the SasaIFNA1 proximal promoter region (shaded) which contains predicted NF-kB, IRF and TATA binding site(in uppercase bold). Donor and acceptor splice motifs on either end of this intron are boxed. The SasaIFNA1 transcription start site is indicated with a sideways arrow. The startcodons are boxed in black. The rtIFN1 start site in exon 1a is predicted to encode the rtIFN1-short transcript while the start site in exon 1b is predicted to encode the rtIFN1-longtranscript. (B) Hypothetical splicing model for rtIFN1 in homozygous trout based on the genomic sequences (panel A and AM489415) The transcript sequences for two alternatesplice variants of the trout IFN1 gene are encoded over 5 exons labeled, from the 50 end, as exon 1a0 , exon 1a, exon 1b and exons 2 through 5. Exons 1a–5 combine to make the shortIFN1 transcript, which encodes a signal peptide to allow putative protein secretion. In the longer transcript, exon 1a is replaced by exon 1a0 , removing a sequence encoding theputative signal peptide. The asterisk designates the location of the cryptic intron described in panel A.

M.K. Purcell et al. / Fish & Shellfish Immunology 26 (2009) 293–304296

ARP

IFN1 long IFN1 short

IFN3/4

IFN-γ1 IFN-γ2

IFN2

IFN1 Total

mR

NA

Cop

ies

Tissues Tissues

0.0E+00

5.0E+04

1.0E+05

1.5E+05

2.0E+05

2.5E+05

0

20

40

60

0

20

40

60

0

20

40

60

0

20

40

60

0

20

40

60

mR

NA

Cop

ies

0

20

40

60

mR

NA

Cop

ies

mR

NA

Cop

ies

0

20

40

60

Ova

ry

Ant

. Kid

.

Ant

. Int

.

Post

. Int

.

Post

. Kid

.

Sple

en

Gill

Liv

er

Mus

cle

Skin

Bra

in

Ova

ry

Ant

. Kid

.

Ant

. Int

.

Post

. Int

.

Post

. Kid

.

Sple

en

Gill

Liv

er

Mus

cle

Skin

Bra

in

Ova

ry

Ant

. Kid

.

Ant

. Int

.

Post

. Int

.

Post

. Kid

.

Sple

en

Gill

Liv

er

Mus

cle

Skin

Bra

in

Ova

ry

Ant

. Kid

.

Ant

. Int

.

Post

. Int

.

Post

. Kid

.

Sple

en

Gill

Liv

er

Mus

cle

Skin

Bra

in

Ova

ry

Ant

. Kid

.

Ant

. Int

.

Post

. Int

.

Post

. Kid

.

Sple

en

Gill

Liv

er

Mus

cle

Skin

Bra

in

Ova

ry

Ant

. Kid

.

Ant

. Int

.

Post

. Int

.

Post

. Kid

.

Sple

en

Gill

Liv

er

Mus

cle

Skin

Bra

in

Ova

ry

Ant

. Kid

.

Ant

. Int

.

Post

. Int

.

Post

. Kid

.

Sple

en

Gill

Liv

er

Mus

cle

Skin

Bra

in

Ova

ry

Ant

. Kid

.

Ant

. Int

.

Post

. Int

.

Post

. Kid

.

Sple

en

Gill

Liv

er

Mus

cle

Skin

Bra

in

A B

D

F

HG

E

C

Fig. 2. Constitutive gene expression in three unhandled homozygous trout measured using quantitative RT-PCR. One microgram of total RNA was used to initiate cDNA synthesis foreach sample. (A) housekeeping gene (ARP), (B) IFN1 total, which measures both the short and long forms of the IFN1 transcript, (C) IFN1 long form, (D) IFN1 short form, (D) IFN2, (F)IFN3/4, (G) IFN-g1, (H) IFN-g2. All data are presented as the mean of the mRNA copies per reaction �SEM. Absolute copy number was determined based on plasmid DNA standardcurves. Abbreviations used include: ant ¼ anterior, post ¼ posterior, kid ¼ kidney and int ¼ intestine.

M.K. Purcell et al. / Fish & Shellfish Immunology 26 (2009) 293–304 297

and rtIFN4 (FJ184373). The rtIFN3 and rtIFN4 transcripts differ by28 single base differences but encode highly similar deducedproteins (Fig. 3). We attempted to develop a specific rtIFN3 assaythat did not cross-react with rtIFN4 by using a qRT-PCR probecontaining a minor groove binding (MGB) linker (Applied Bio-systems Inc.) (Table 1). The rtIFN3 assay was not specific andrecognized rtIFN4 with similar efficiency despite the probe beingmismatched with rtIFN4 by 3 nucleotides. The combined rtIFN3/4gene had low constitutive expression overall with no detectable

expression in the major immune organs (kidney, spleen, gill andintestine) and the highest expression was observed in ovary(Fig. 2F).

3.1.4. rtIFN-g: evidence of gene duplicationTwo unique IFN-g sequences were observed in homozygous

rainbow trout (Fig. 4A). We have designated these sequences HotCreek (HC) IFN-g1 and IFN-g2 in Fig. 4A and Table 2. IFN-g1 hashigh amino acid similarity (98.9%) to the previously reported

Fig. 3. Two distinct type I IFN group II genes were present in homozygous trout. Alignment of rtIFN3 (AM235738) with two IFN3-related sequences (IFN3 and IFN4) fromhomozygous (Hot Creek strain; HC) rainbow trout. Nucleotide (nt) sequences are shown in lower case, and the predicted amino acid (aa) sequence of AM235738 is shown inuppercase. Start (atg) and stop (tga) codons are in bold and underlined and the flanking 50 and 30 UTRs are italicized. Dots (.) represent identical nt in HC IFN3 and HC IFN4 relative tortIFN3, and dashes (-) represent gaps in the alignment. Amino acids changes produced by differences in the nt sequence are shown in shaded text above the alignment. Boxed areasrepresent predicted N-linked glycosylation sites. The two HC IFN sequences share 96.7% nt and 97.8% aa identity. Their nt and aa identities to AM235738 are shown.

M.K. Purcell et al. / Fish & Shellfish Immunology 26 (2009) 293–304298

rainbow trout gene (rtIFN-g [12]) (Fig. 4; Table 2). IFN-g2 is moresimilar to the Atlantic salmon gene (SasaIFN-g) (92.8%) (Fig. 4A;Table 2). Of 6 putative alpha helices in rtIFN-g, the amino aciddifferences between HC IFN-g1 and HC IFN-g2 were predominantlyin the alpha helix called region D (Fig. 4A). Phylogenetic analysessupport the clustering of the HC IFN-g1 with the previouslydescribed rtIFN-g and the HC IFN-g2 with the SasaIFN-g; thesalmonid sequences cluster separately from other fish IFN-g genes(Fig. 4B). Thus we propose that the nomenclature rtIFN-g1 for thepreviously reported gene [12] and rtIFN-g2 for the novel IFN-greported here.

To assess whether additional genomic copies of IFN-g arepresent in homozygous trout, we amplified the region of the rtIFN-g genes from exon 1 to exon 2 (these exons flank the highly poly-morphic intron 1) [12]. Genomic DNAs from homozygous Hot Creektrout as well as two other inbred rainbow trout lines (OSU and Arlee[18]) were analyzed. Three to five bands were seen in all threestrains of homozygous trout (Fig. 4C). Two bands were sequencedfrom Hot Creek trout. An 864 bp band matched with HC IFN-g1 inthe coding region while an 813 bp band matched with HC IFN-g2;these intron sequences have been deposited to GenBank (FJ222819and FJ222820) but the remaining bands have not beencharacterized.

Two qRT-PCR assays were developed to specifically assess rtIFN-g1 and rtIFN-g2 (Table 1) and the assay specificity was confirmed

using plasmids encoding the genes. The rtIFN-g1 and -g2 genes hadsimilar constitutive expression with the highest expression inimmune-related organs (gill, spleen and anterior intestine) (Fig. 2Fand G).

3.2. IFN transcriptional regulation following IHNV DNA vaccinationof homozygous trout

Total IFN1 expression was up-regulated w27-fold at the I.M.site of pIHNw-g vaccination 7 days p.v. (Fig. 5A). The long IFN1transcript was highly induced at the I.M. site (w21-fold) (Fig. 5B)but higher induction was observed for the short IFN1 transcript(w42-fold) (Fig. 5C). IFN1 up-regulation was more modest in theanterior kidney; only induction of the short IFN1 transcript wassignificantly greater in vaccinated fish than in fish receiving themock (PBS) control (Fig. 5A–C). Expression of rtIFN2 was signifi-cantly induced by vaccination at the I.M. site (4.3-fold) relative tothe mock control but not relative to the vector treatment (Fig. 5D).In the anterior kidney, rtIFN2 was not significantly up-regulatedby the vaccine relative to either the mock of vector control. IFN3/4had the highest magnitude of induction (w340-fold) in responseto the DNA vaccine at the I.M. site (Fig. 5E). This induction wassignificantly greater than both the mock and vector controlgroups. IFN3/4 was up-regulated in the anterior kidney but notsignificantly.

D. rerio IFN- 1-2

I. punctatus IFN- 2b

I. punctatus IFN- 2a

I. punctatuss IFN- 1

D. rerio IFN- 1-1

G. gallus IFN-

O. mykiss IFN-

S. salar IFN-

100

99

77

7059

100

91 0.1

A

B

C

OSU

Arl

ee

Hot

Cre

ek

750 bp

1500 bp

1000 bp**

Fig. 4. Two distinct genes for IFN-g are present in homozygous trout. (A) Alignment of IFN-g cDNA sequences from rainbow trout (rtIFN-g; AJ616215) and Atlantic salmon (SasaIFN-g;AJ841811) with two IFN-g related sequences (IFN-g1 and IFN-g2) from homozygous (Hot Creek strain; HC) rainbow trout. Nucleotide sequences are shown in lower case and the predictedamino acid sequence of rtIFN-g is shown in uppercase. Dots (.) represent identical nucleotides relative to rtIFN-g and dashes (-) represent gaps in the alignment. Predicted amino acidssubstitutions are shown in shaded text above the alignment. Region D and the nuclear localization site (NLS) (RRKR) domain of IFN-g are indicated with underlining. (B) Neighbor-joiningtree depicting the relationships between HC IFN-g1 and IFN-g2 with other fish IFN-g genes. Phylogenies were generated from amino acid alignments and bootstrapped 1000 times.Bootstrap values are shown. Accession numbers for all genes are given in Table 2. (C) PCR amplification of the rtIFN-g intron 1 from genomic DNA representing three different strains ofhomozygous trout (Hot Creek, Arlee and OSU). The first intron of rtIFN-g contains 30 bp and microsatellite repeats as previously described [12]. Lanes were loaded as follows: (1) PCR DNAmarker (Amresco, Inc.), (2) Hot Creek, (3) Arlee and (4) OSU. The asterisks designate bands corresponding to HC IFN-g1 and HC IFN-g.2 (GenBank FJ222819 and FJ222820).

M.K. Purcell et al. / Fish & Shellfish Immunology 26 (2009) 293–304 299

IFN-g1 and IFN-g2 were both up-regulated at the I.M. site (w25-and 83-fold, respectively) after DNA vaccination; transcript levelswere significantly greater in vaccinated fish than either the mock orvector treated control groups (Fig. 5F and G). In the anterior kidney,

IFN-g1 was down-regulated in the vector and vaccine grouprelative to the mock group, although not significantly. IFN-g2 wasup-regulated, although not significantly, in the anterior kidneys ofthe DNA vaccinated group.

Table 2Pairwise comparison of Hot creek IFN-g1 and IFN-g2 with other vertebrate IFN-gamino acid sequences.

Species IFN-g % Amino acid identity toa Accession #

HC IFN-g1 HC IFN-g2

Rainbow trout (HC strain) IFN-g1 100 90.6 FJ184374Rainbow trout (HC strain) IFN-g2 90.6 100 FJ184375Rainbow trout IFN-g 98.9 91.7 AJ616215Atlantic salmon IFN-g 89.4 92.8 AJ841811Channel catfish IFN-g1 22.2 23.2 DQ124249Channel catfish IFN-g2a 31.9 32.4 DQ124250Channel catfish IFN-g2b 32.1 32.6 DQ124251Zebrafish IFN-g1-1 25.4 25.6 AB194272Zebrafish IFN-g1-2 34.4 35.4 Nm_212864Chicken IFN-g 20.7 22.8 X99774

a Percent identity score calculated by Global pairwise alignment of full-lengthamino acid sequences using the Needle program.

Ant. Kid. I.M. Site

0.0000

0.0003

0.0005

0.0008

0.0010

Ant. Kid. I.M. Site

Ant. Kid. I.M. Site

Ant. Kid. I.M. Site

IFN

1 T

otal

/AR

P R

atio

0.0000

0.0003

0.0005

0.0008

0.0010

IFN

1 L

ong/

AR

P R

atio

0.0000

0.0003

0.0005

0.0008

0.0010

IFN

2 /A

RP

Rat

io

0.0000

0.0003

0.0005

0.0008

0.0010

A

B

D

F

ab a b

a ab b

ab a b

0.7x1.9x 3.5x

26.7x

0.7x1.6x

3.3x

20.6x

a ab b

a ab b

2.3x 2.5x1.6x 4.3x

0.4x

0.8x

1.6x

24.8x

a a b

IFN

- 1/

AR

P R

atio

Fig. 5. Interferon gene expression 7 days after DNA vaccination. Expression was monitoredPCR. (A) IFN1 total, which measures both the short and long forms of the IFN1 transcript, (B)y-axis, data are presented as mean normalized ratios (specific gene copy number/ARP copycompared across genes. Fold changes estimates of the vector and vaccine groups relative todesignated with a distinct lower-case letter while groups that are not significantly differen

M.K. Purcell et al. / Fish & Shellfish Immunology 26 (2009) 293–304300

3.3. IFN transcriptional regulation following IHNV infection inhomozygous trout

Homozygous trout infected with IHNV and control (mockinfected) trout were assessed for changes in IFN transcription intheir anterior kidneys over time. Only 2/15 fish injected with virusdied in this study (at 11 d and 14 d p.i.); no mortalities wereobserved in the mock control groups. Total rtIFN1 expression wassignificantly induced at day 3 (3 d) (w21-fold) and day 7 (7 d)(w11-fold) post-infection (p.i.) but not at day 28 (28 d) (w2-fold)relative to the mock injected fish (Fig. 6A). The rtIFN1-short formhad much higher induction in response to IHNV infection relative tothe long form, although both forms were significantly induced at3 d and 7 d p.i. (Fig. 6B and C). For instance, at 3 d p.i. the short formwas induced w68-fold while the long form was only induced w5-fold. Expression of rtIFN2 was significantly up-regulated in infectedfish at 3 d and 7 d p.i. (w14 and 22-fold) relative to the mock

Ant. Kid. I.M. Site

Ant. Kid. I.M. Site

Ant. Kid. I.M. Site

0.0000

0.0003

0.0005

0.0008

0.0010

IFN

1 sh

ort/

AR

P R

atio

0.0000

0.0003

0.0005

0.0008

0.0010

0.0000

0.0003

0.0005

0.0008

0.0010

IFN

3/4/

AR

P R

atio

G

E

C

PBS

Vector

pIHNw-G

a ab b

1.1x3.4x 3.3x

41.5x

344x

2.8x

4.3x1.0x

b

a

a a

ab b

0.1x

1.6x

1.3x

82.7x

a a b

IFN

- 2/

AR

P R

atio

at both the I.M. site of vaccine delivery and the anterior kidney using quantitative RT-IFN1 long form, (C) IFN1 short form, (D) IFN 2, (E) IFN3/4, (F) IFN-g1, (G) IFN-g2. On thenumber) � SEM. These ratios reflect absolute transcript abundance and can be directlyPBS injected fish are shown above the respective bars. Statistically different groups aret share the letter.

0.0000

0.0150

0.0300

0.0000

0.0010

0.0020

0.0030

0.0000

0.0010

0.0020

0.0030

0.0000

0.0010

0.0020

0.0030

0.0000

0.0150

0.0300

0.0000

0.0010

0.0020

0.0030

0.0000

0.0010

0.0020

0.0030

3 7 28

3 7 28

3

3

7

7

28

28 3 7 28

3 7 28

3 7 28

IFN

1 T

otal

/AR

P R

atio

IFN

1 L

ong/

AR

P R

atio

IFN

1 Sh

ort/

AR

P R

atio

IFN

2 /A

RP

Rat

io

DaysDays

PBS

IHNV

20.6x

10.6x

1.9x

4.5x 2.3x 1.4x

67.5x

20.3x

13.6x22.1x

0.6x

0.9x

*

*

* *

*

*

**

IFN

3/4/

AR

P R

atio 198x

1.9x

*

IFN

-γ1/

AR

P R

atio

IFN

-γ2/

AR

P R

atio

257x

49.3x1.5x

*

*

1050x

93.4x1.3x

*

*

0.2x

A

B

D E

F G

C

Fig. 6. Interferon gene expression in the anterior kidney following IHNV infection. (A) IFN1 total, which measures both the short and long forms of the IFN1 transcript, (B) IFN1 longform, (C) IFN1 short form, (D) IFN 2, (E) IFN3/4, (F) IFN-g1, (G) IFN-g2 (note the y-axis scale change for panels F and G). On the y-axis, data are presented as mean normalized ratios(specific gene copy number/ARP copy number) � SEM. These ratios reflect absolute transcript abundance and can be directly compared across genes. Fold change estimates of theinfected group relative to the PBS injected fish are shown above the respective bars. Statistical significance between mock and IHNV infected groups is designated with an asterisk.

M.K. Purcell et al. / Fish & Shellfish Immunology 26 (2009) 293–304 301

control. Expression of IFN2 was highest at 7 d p.i. whereas rtIFN1induction was greatest at 3 d p.i. (Fig. 6D). Expression levels ofrtIFN2 in the infected fish were similar to those of mock control fishat 28 d p.i. (Fig. 6D). Of all the type I IFN genes, IFN3/4 was thehighest induced (198-fold) at 3 d p.i. (Fig. 6E). This induction wastransient and expression in the infected group was not significantlygreater than the mock controls at either 7 or 28 d p.i. (Fig. 6E).

IFN-g1 and IFN-g2 were both significantly induced at 3 and7 d p.i. in the virally-infected fish relative to the mock control fish(Fig. 6F and G). Overall, IFN-g2 had higher induction than IFN-g1.For instance, IFN-g2 was induced 1050-fold at 3 d p.i. while IFN-g1was induced only 257-fold. The IFN-g1 and IFN-g2 overall tran-script abundances were approximately one log greater than thetype I IFNs (note the change in the y-axis for Fig. 6F and G).

3.4. IFN transcription following poly I:C stimulation innon-homozygous trout

All rtIFN1 transcript forms were induced in a dose and timedependent manner in poly I:C stimulated anterior kidney leukocytepreparations (Fig. 7A–C). The IFN1 short form had the highestoverall induction (w67–550-fold) (Fig. 7C). IFN2 was also highlyinduced by poly I:C ranging from w30- to 445-fold (Fig. 7D). IFN3/4

had the highest induction of all the type I IFN genes ranging fromw105- to 1819-fold (Fig. 7E). For all type I IFNs, the single 10-foldhigher poly I:C dose of 50 mg/ml stimulated at most a 1.5–2.0-foldincrease in gene expression. The type II IFNs were also significantlyinduced in response to poly I:C stimulation but to much lowerlevels and in a different kinetic manner relative to the type I IFNs.IFN-g2 had a higher response to poly I:C (w3- to 42-fold) relative toIFN-g1 (w2–10-fold) (Fig. 7F and G). The 10-fold higher dose of polyI:C (50 mg/ml) did not stimulate higher IFN-g expression.

4. Discussion

Short and long transcript forms for type I IFNs have beenreported in both Atlantic salmon and zebrafish [8,29]. The long andshort zebrafish IFN transcripts are derived by differential promoteruse and alternative splicing [8]. The constitutively expressed longtranscript in zebrafish contains an alternate 50 exon (exon 10) that isspliced (in frame) downstream of the first start codon. Uponinfection, a different promoter splices exon 1 to the remainingexons; this short transcript encodes a signal peptide and thereforegenerates a secreted form of IFN [8]. In this study, the short andlong transcripts of rtIFN1 appear to derive from a similar alternativesplicing mechanism as zebrafish. Based on sequence analogy with

0.0000

0.0100

0.0200

0.0300

0.0400

0.0000

0.0100

0.0200

0.0300

0.0400

0.0000

0.0100

0.0200

0.0300

0.0400

0.0000

0.0100

0.0200

0.0300

0.0400

0.0000

0.0100

0.0200

0.0300

0.0400

0.0000

0.0100

0.0200

0.0300

0.0400

0.0000

0.0100

0.0200

0.0300

0.0400

IFN

1 T

otal

/AR

P R

atio

0 µg/ml 5 µg/ml 50 µg/ml

0 µg/ml 5 µg/ml 50 µg/ml

nd nd

IFN

1 L

ong/

AR

P R

atio

IFN

1 sh

ort/

AR

P R

atio

0 µg/ml 5 µg/ml 50 µg/ml

IFN

2/A

RP

Rat

io

IFN

3/4/

AR

P R

atio

0 µg/ml 5 µg/ml 50 µg/ml

nd nd

nd nd

nd nd

6 hrs12 hrs24 hrs

*20x

*69x

*124x

*280x

*5.2x

*8.8x8.9x

**

13.3x*

67x

*258x

*425x

*550x

*31.1x

*139x

*223x

*445x

0 µg/ml 5 µg/ml 50 µg/ml

*105x

*1084x

*818x

*1819x

nd nd

0 µg/ml 5 µg/ml 50 µg/ml 0 µg/ml 5 µg/ml 50 µg/ml

nd nd2.1x

*10.1x *

7.4x*

7.3x

nd nd3.4x

*41.1x *

34.5x *26.3x

Poly I:C dose Poly I:C dose

A

B C

E

F G

D

IFN

- 2

/AR

P R

atio

IFN

- 1/

AR

P R

atio

Fig. 7. Interferon gene expression following in vitro poly I:C stimulation of anterior kidney leukocytes measured using quantitative RT-PCR. (A) IFN1 total, which measures both theshort and long forms of the IFN1 transcript, (B) IFN1 long form, (C) IFN1 short form, (D) IFN 2, (E) IFN3/4, (F) IFN-g1, (G) IFN-g2. On the y-axis, data are presented as meannormalized ratios (specific gene copy number/ARP copy number) � SEM. These ratios reflect absolute transcript abundance and can be directly compared across genes. Fold changeestimates relative to the mock control group (0 mg/mL poly I:C) are shown above the respective bars. The discrepancy between fold change estimates and the normalized ratios wasbecause specific gene expression varied in the mock controls at different time points. Analysis using the 50 mg/mL dose was only conducted at 24 h post-stimulation; other timepoints were not determined (nd). Statistical significance between mock controls and poly I:C treated is designated with an asterisk.

M.K. Purcell et al. / Fish & Shellfish Immunology 26 (2009) 293–304302

zebrafish IFN and signal peptide prediction algorithims, the shortrtIFN1 transcript encodes a signal peptide that would likely lead tocytokine secretion. This short form of rtIFN1 was the most inducibleafter viral infection, vaccination or poly I:C stimulation. The longrtIFN1 transcript lacks a traditional secretion signal, and ispresumably not secreted, which raises questions regarding thefunctional role of the long transcript form. Alternatively, the longtranscript could possess a non-traditional signal peptide that wouldlead to its secretion. Atlantic salmon have two group I type I IFNgenes (SasaIFNA1 and SasaIFNA2) and both genes can produce longand short transcripts that differ in the length of the 50UTR, but haveidentical coding regions [11,30]. These short and long transcriptsare produced from the same gene via differential promoter usage (aproximal or distal promoter region, respectively). Reporterconstructs containing the proximal promoter of salmon IFN arehighly induced by poly I:C relative to constructs containing thedistal promoter [29]. The SasaIFNA1 proximal promoter was highlysimilar to the rtIFN1 genomic region immediately upstream of thetranscription start of the short transcript (Fig. 1A). This high

conservation supports the inference that this region is the proximalpromoter for rtIFN1 leading to the inducible rtIFN1-short form. TheSasaIFNA1 distal promoter leads to the production of the longtranscript. The rtIFN1 gene possessed a highly similar region(Fig. 1A) except for a deletion in the region identified as the start ofthe SasaIFNA1 distal promoter. Immediately upstream of distalpromoter region is a degenerate NF-kB binding site. It is alsopossible that rtIFN2 also encodes a short and long transcript formbut this was not investigated in our study.

Expression of rtIFN1 and rtIFN2 generally agreed witha previous study that used semi-quantitative PCR to monitorexpression [13]. However, Zou et al. [13] reported higher consti-tutive expression of rtIFN2 across a range of tissues, and greatertranscript abundance, relative to rtIFN1 after poly I:C stimulation.These results are in contrast to our study as we observed similarconstitutive expression of rtIFN2 and rtIFN1. This discrepancy maybe because our qRT-PCR assay for IFN2 was designed towards the 50

end of the gene. If other transcript forms of rtIFN2 existed, they maynot be detected by our assay. Additionally, the rtIFN2 was less

M.K. Purcell et al. / Fish & Shellfish Immunology 26 (2009) 293–304 303

inducible in response to vaccine and virus relative to rtIFN1-short.However, these transcripts differed kinetically in vivo as rtIFN1-short was most abundant at 3 d p.i. while rtIFN2 was most abun-dant at 7 d p.i. Thus, rtIFN2 may have peaked at a time point notexamined in our study. However, this kinetic difference was notevident in vitro, where rtIFN1-short and rtIFN2 had relativelysimilar kinetic responses to poly I:C.

Rainbow trout and zebrafish possess a second major group oftype I IFNs (type I group II IFNs) [13]. During the course of our study,we identified a second type I group II IFN gene that was closelyrelated to rtIFN3; we designated this gene rtIFN4, consistent withthe current nomenclature. These two genes have high sequenceidentity (99% amino acid and 97% nucleotide) and may be confusedfor allelic forms in non-homozygous animals; our use of homozy-gous fish removes this variable. Attempts to make a qRT-PCR assaythat distinguishes between rtIFN3 and rtIFN4 were not successfuland so the assay used in this study detects both genes. Zou et al.previously demonstrated that constitutive expression of rtIFN3, intissues, was restricted with high expression observed in the testesor ovaries, and detectable rtIFN3 expression was only observed inanterior kidney leukocytes preparations when the cells werestimulated with high concentrations of poly I:C (100 mg/ml). Weobserved a similar restricted constitutive expression of rtIFN3/4with detectable expression limited to the ovaries and skin.However, in our hands, rtIFN3/4 showed the highest response toinfection, vaccination or poly I:C stimulation relative to bothrtIFN1-short and rtIFN2. Expression of rtIFN3/4 appears transientafter viral infection as no significant elevated expression wasobserved at 7 d p.i., in contrast to rtIFN1 and rtIFN2. Our resultsclearly indicate that rtIFN3/4 is highly regulated in response to viralstimuli suggesting a functional role in the anti-viral response.Conflicting with this hypothesis, recombinant rtIFN3 did notpossess the same anti-viral activity of rtIFN1 or rtIFN2 nor did itinduce the transcription of Mx during a previous investigation,although incorrect folding of the protein could not be ruled out [13].

In this study we have identified a second IFN-g gene in homo-zygous trout, designated IFN-g2. This gene is highly similar to thepreviously described rainbow trout IFN-g1 and possesses typicalIFN-g motifs, including a nuclear localization site shown to berequired for IFN-g1 function [12]. Studies in mammals havedemonstrated that IFN-g, STAT1 and one of its receptor subunits(IFNGR1) are translocated to the nucleus and directly interact withthe promoters of IFN-g inducible genes [31]. Zebrafish, commoncarp and channel catfish have duplicated IFN-g genes. However, theIFN-g1 and -g2 sequences reported in these other species are moredivergent than the rainbow trout sequences reported here. Forinstance, channel catfish IFN-g1 and -g2 share only approximately20% amino acid identity [10] whereas the rtIFN-g1 and -g2 have90% identity. The IFN-g1 and -g2 genes reported in the other fishspecies most likely diverged early in teleost evolution. The dupli-cated rainbow trout IFN-g genes may represent a more recentsalmonid-specific genome duplication event [32]. Thus, it ispossible that rainbow trout possess more than two IFN-g genes.The length polymorphisms observed in intron 1 also suggesthomozygous fish may have additional IFN-g genes or pseudogenes.The rtIFN-g1 had higher constitutive expression than rtIFN-g2 butboth were primarily detected in tissues of immunological signifi-cance (gill, spleen, kidney and intestine). Expression of rtIFN-g2was induced to higher magnitude than IFN-g1 in response to viralinfection, vaccination or poly I:C stimulation; however, both geneswere significantly induced. Although IFN-g1 and -g2 responded topoly I:C, it differed kinetically from the type I IFNs and did notappear to be affected by increasing poly I:C dose.

The primary goal of this study was to develop specific assays tosystematically profile all of the rainbow trout IFN genes after stim-ulation with virus or viral-related molecules. The transcriptional

profiling of the IFN genes in this study were largely in agreementwith our previous studies [16,23,25]. Early after IHNV infection, allIFN genes examined were significantly up-regulated within 3 d p.i.but differed in the magnitude and timing of response. The type II IFNgenes (rtIFN-g1 and -g2) had high induction following viral stimu-lation but this response was transient and did not persist into theadaptive immune phase (28 d p.i.). It is not known which cells areproducing IFN-g in the early phase of infection. The two rainbowtrout IFN-g genes are similar to the common carp IFN-g2 gene whichhas recently been shown to be associated with T lymphocytes [6].Previous studies have demonstrated a cytotoxic T lymphocyte (CTL)response in rainbow trout after infection with a highly relatedaquatic rhabdovirus (Viral hemmorhagic septicemia virus) [33] but nostudies have examined the ability of virus specific CTLs to producertIFN-g. Early after IHNV DNA vaccination (7 d p.v.), a generalizedanti-viral response is associated with systemic induction of ISGs[17]. In the present study, we confirmed that the IFN genes them-selves were only induced consistently at the I.M. site of vaccinationand not at secondary sites such as the anterior kidney. The specificityof the qRT-PCR assays developed in this study will be useful forfuture studies aimed at identifying which cells are producing IFN atearly time points after infection.

We predict that additional IFN genes exist in rainbow troutbased on several lines of evidence. PCR analysis of the IFN-g intronsuggests that homozygous trout have >3 gene copies. Additionally,other teleost fish possess a second more distantly related IFN-ggene that may have evolved by tandem duplication [10]. The type IIFNs in trout are also likely to be further duplicated beyond the 4genes reported to date. The ‘total’ rtIFN1 assay described heredetects a greater number of transcripts than would be expectedfrom the short and long transcript copies combined. During therevision of this manuscript, it was reported that Atlantic salmonhave multi-gene clusters encoding 11 genes representing threedifferent IFN subgroups [34]. The Atlantic salmon gene clusterswere identified by large-scale sequencing of IFN containinggenomic regions and this approach, in combination with geneticand physical mapping, will likely identify additional rainbow troutIFN genes. The homozygous trout lines provide a valuable tool forcharacterizing the complex IFN gene families. The functional rolesof the multiple IFN genes have yet to be determined.

Acknowledgements

The authors would like to acknowledge the assistance of EricLandis (National Marine Fisheries Service) and Gael Kurath(Western Fisheries Research Center). Paul Wheeler (WashingtonState University) provided the homozygous trout used in this study.Dr. Scott LaPatra (Clear Springs Food Inc.) provided the outbredtrout used in this study. The authors thank Dr. Jun Zou (Universityof Aberdeen) for helpful discussion. The project was supported bythe National Research Initiative of the USDA Cooperative StateResearch, Education and Extension Service, grant number 2006-35204-17393 and grant number 02066006. The use of trade, firm,or corporation names in this publication is for the information andconvenience of the reader. Such use does not constitute an officialendorsement or approval by the U.S. Department of Interior or theU.S. Geological Survey of any product or service to the exclusion ofothers that may be suitable.

References

[1] Samuel C. Antiviral actions of interferons. Clinical Microbiology Reviews2001;14:778–809.

[2] Ank N, West H, Paludan SR. IFN-lambda: novel antiviral cytokines. Journal ofInterferon and Cytokine Research 2006;26:373–9.

[3] Sen GC. Viruses and interferons. Annual Review of Microbiology 2001;55:255–81.

M.K. Purcell et al. / Fish & Shellfish Immunology 26 (2009) 293–304304

[4] Kalvakolanu DV. Alternate interferon signaling pathways. Pharmacology andTherapeutics 2003;100:1–29.

[5] Zou J, Yasutoshi YB, Dijkstra JM, Sakai M, Ototake M, Secombes C. Identifica-tion of an interferon gamma homologue in fugu, Takifugu rubripes. Fish andShellfish Immunology 2004;17:403–9.

[6] Stolte EH, Savelkoul HFJ, Wiegertjes G, Flik G, Verburg-van Kemenade LBM.Differential expression of two interferon gamma genes in common carp (Cyp-rinus carpio L.). Developmental and Comparative Immunology 2008;32:1467–81.

[7] Altman SM, Mellon MT, Distel DL, Kim CH. Molecular and functional analysisof an interferon gene from the zebrafish, Danio rerio. Journal of Virology2003;77:1992–2002.

[8] Levraud JP, Boudinot P, Colin I, Benmansour A, Peyrieras N, Herbomel P, et al.Identification of the zebrafish IFN receptor: implications for the origin of thevertebrate IFN system. Journal of Immunology 2007;178:4385–94.

[9] Long S, Wilson M, Bengten E, Bryan L, Clem LW, Miller NW, et al. Identificationof a cDNA encoding channel catfish interferon. Developmental and Compar-ative Immunology 2004;28:97–111.

[10] Milev-Milovanovic I, Long S, Wilson M, Bengten E, Miller NW, Chinchar VG.Identification and expression analysis of interferon gamma genes in channelcatfish. Immunogenetics 2006;58:70–80.

[11] Robertsen B, Bergan V, Rokenes T, Larsen R, Albuquerque A. Atlantic salmoninterferon genes: cloning, sequence analysis, expression, and biologicalactivity. Journal of Interferon Cytokine Research 2003;10:601–12.

[12] Zou J, Carrington AC, Collet B, Dijkstra JM, Yoshiura Y, Bols N, et al. Identifi-cation and bioactivities of IFN-gamma in rainbow trout Oncorhynchus mykiss:the first Th1-type cytokine characterized functionally in fish. Journal ofImmunology 2005;175:2484–94.

[13] Zou J, Tafalla C, Truckle J, Secombes C. Identification of a second group of type IIFNs in fish sheds light on IFN evolution in vertebrates. Journal of Immunology2007;179:3859–71.

[14] Igawa D, Sakai M, Savan R. An unexpected discovery of two interferon gamma-like genes along with interleukin (IL)-22 and -26 from teleost: IL-22 and -26genes have been described for the first time outside mammals. MolecularImmunology 2006;43:999–1009.

[15] Tordo N, Benmansour A, Calisher C, Dietzgen RG, Fang RX, Jackson AO, et al.Family Rhabdoviridae. In: The eighth report of the international committee fortaxonomy of viruses. San Diego: Academic Press; 2004.

[16] Landis ED, Purcell MK, Thorgaard GH, Wheeler PA, Hansen JD. Transcriptionalprofiling of MHC class I genes in rainbow trout infected with infectioushematopoietic necrosis virus. Molecular Immunology 2008;45:1646–57.

[17] Purcell MK, Nichols KM, Winton JR, Kurath G, Thorgaard GH, Wheeler P, et al.Comprehensive gene expression profiling following DNA vaccination ofrainbow trout against Infectious hematopoietic necrosis virus. MolecularImmunology 2006;43:2089–106.

[18] Young WP, Wheeler PA, Fields RD, Thorgaard GH. DNA fingerprinting confirmsisogenicity of androgenetically derived rainbow trout lines. Journal of Heredity1996;87:77–80.

[19] Corbeil S, LaPatra SE, Anderson ED, Jones J, Vincent B, Hsu YL, et al. Evaluationof the protective immunogenicity of the N, P, M, NV and G proteins of

infectious hematopoietic necrosis virus in rainbow trout Oncorhynchus mykissusing DNA vaccines. Diseases of Aquatic Organisms 1999;39:29–36.

[20] Corbeil S, LaPatra SE, Anderson ED, Kurath G. Nanogram quantities of a DNAvaccine protect rainbow trout fry against heterologous virus strains of infec-tious hematopoietic necrosis virus. Vaccine 2000;18:2817–24.

[21] Corbeil S, Kurath G, LaPatra SE. Fish DNA vaccine against infectious hemato-poietic necrosis virus: efficacy of various routes of immunization. Fish andShellfish Immunology 2000;10:711–23.

[22] LaPatra SE, Corbeil S, Jones GR, Shewmaker WD, Kurath G. The dose-dependenteffect on protection and humoral response to a DNA vaccine against IHN virus insubyearling rainbow trout. Journal of Aquatic Animal Health 2000;12:181–8.

[23] Rodriguez MF, Wiens G, Purcell MK, Palti Y. Characterization of Toll-likereceptor 3 gene in rainbow trout (Oncorhynchus mykiss). Immunogenetics2005;57:510–9.

[24] Braun-Nesje R, Kaplan G, Seljelid R. Rainbow trout macrophages in vitro:morphology and phagocytic activity. Developmental and ComparativeImmunology 1982;6:281–91.

[25] Purcell MK, Kurath G, Garver KA, Herwig RP, Winton JR. Quantitativeexpression profiling of immune response genes in rainbow trout followinginfectious haematopoietic necrosis virus (IHNV) infection of DNA vaccination.Fish and Shellfish Immunology 2004;17:447–62.

[26] Larkin MA, Blackshields G, Brown NP, Chenna R, McGettigan PA, McWilliam H,et al. Clustal W and Clustal X version 2.0. Bioinformatics 2007;23:2947–8.

[27] Tamura K, Dudley J, Nei M, Kumar S. MEGA4: molecular Evolutionary Geneticsanalysis (MEGA) software version 4.0. Molecular Biology and Evolution2007;24:1596–9.

[28] Pierce AL, Dickey JT, Larsen DL, Fukada H, Swanson P, Dickhoff WW. A quanti-tative real-time RT-PCR assay for salmon IGF-I mRNA and its application in thestudy of GH regulation of IGF-I gene expression in primary culture of salmonhepatocytes. General and Comparative Endocrinology 2004;135:401–11.

[29] Bergan V, Steinsvik S, Xu H, Kileng O, Robertsen B. Promoters of type Iinterferon genes from Atlantic salmon contain two main regulatory regions.FEBS Journal 2006;273:3893–906.

[30] Robertsen B. The interferon system of teleost fish. Fish and Shellfish Immu-nology 2006;20:172–91.

[31] Ahmed CM, Johnson HM. IFN-gamma and its receptor subunit IFNGR1 arerecruited to the IFN-gamma-activated sequence element at the promoter siteof IFN-gamma-activated genes: evidence of transactivational activity inIFNGR1. Journal of Immunology 2006;177:315–21.

[32] Allendorf FW, Thorgaard G. Tetraploidy and the evolution of salmonid fishes.In: Turner BJ, editor. The evolutionary genetics of fishes. New York: PlenumPress; 1994. p. 1–53.

[33] Utke K, Bergmann S, Lorenzen N, Kollner B, Ototake M, Fischer U. Cell-mediatedcytotoxicity in rainbow trout, Oncorhynchus mykiss, infected with viral hae-morrhagic septicaemia virus. Fish and Shellfish Immunology 2007;22:182–96.

[34] Sun B, Robertsen B, Wang Z, Bin L. Identification of an Atlantic salmon IFNmultigene cluster encoding three IFN subtypes with very different expressionproperties. Developmental and Comparative Immunology, in press, doi:10.1016/j.dci.2008.10.00 [ Epub Nov. 11.].

http://doi:10.1016/j.dci.2008.10.00http://doi:10.1016/j.dci.2008.10.00

Characterization of the interferon genes in homozygous rainbow trout reveals two novel genes, alternate splicing and differential regulation of duplicated genesIntroductionMaterials and methodsHomozygous rainbow troutDNA vaccination and infection of homozygous troutPoly I:C stimulation of anterior kidney leukocytes in vitro from non-homozygous fishRNA extraction, cDNA synthesis and sequencingAbsolute quantitative reverse transcriptase PCR

ResultsGene sequencing and qRT-PCR assay developmentIFN1-evidence of alternate splicingrtIFN2rtIFN3: evidence of gene duplicationrtIFN-gamma: evidence of gene duplication

IFN transcriptional regulation following IHNV DNA vaccination of homozygous troutIFN transcriptional regulation following IHNV infection in homozygous troutIFN transcription following poly I:C stimulation in non-homozygous trout

DiscussionAcknowledgementsReferences