The Journal of Immunology

Conservation of Structural and Functional Features in a Primordial CD80/86 Molecule from Rainbow Trout (Oncorhynchus mykiss), a Primitive Teleost Fish1,2

Yong-An Zhang,3* Jun-ichi Hikima,3* Jun Li,* Scott E. LaPatra,† Yan-Ping Luo,* and J. Oriol Sunyer4*

In mammals, interaction of CD28 with CD80 or CD86 molecules provides costimulatory signals for T cell activation that leads to increased IL-2 gene and protein expression by activated T cells. Thus far, CD80 and CD86 have been cloned and functionally characterized only in mammals and birds. To shed light into the evolution of CD80 and CD86, we have cloned and functionally characterized a rainbow trout (rt) molecule (rtCD80/86) that shows the highest degree of sequence conservation and phylogenetic relationship with CD80 and CD86 molecules. Moreover, its genomic organization was almost identical to that of human CD86. Rainbow trout possess one membrane-bound and two soluble CD80/86 transcripts, all of which are derived from the same rtCD80/86 gene. The membrane-bound form exhibited its highest degree of expression in lymphoid tissues, particularly on B cells. Incubation of trout leukocytes with LPS and bacteria leads to up-regulation of rtCD80/86 gene expression. Importantly, we show that trout and other teleost fish contain a single CD80/86 gene, thus suggesting that this gene may represent the ancestor from which CD80 and CD86 arose by gene duplication in more evolved species. To gain further insights into the function of rtCD80/86, we have identified and cloned trout IL-2 and have shown that recombinantly produced trout CD80/86 up-regulates the expression of IL-2 in trout blood leukocytes. Significantly, this finding indicates that the capacity to modulate IL-2 expression is a primordial function that has been conserved both in fish and mammalian CD80/CD86 molecules throughout 350 million years of evolution. The Journal of Immunology, 2009, 183: 83–96.

cell activation requires a costimulatory signal that is de- the extracellular domains in both CD80 and CD86 have homology livered through the interaction between surface molecules with the Ig variable region, whereas the membrane proximal do- T of APCs and T cells (1). The two most important cell mains are more homologous to the Ig constant region (5). Al- costimulatory molecules on APCs are CD80 (B7-1) and CD86 though both CD80 and CD86 are mainly expressed as surface mol- (B7-2), both of which belong to the B7 family. To exert their ecules, alternatively spliced variants of mammalian CD80 and by guest on October 1, 2021. Copyright 2009 Pageant Media Ltd. costimulatory activity, CD80 and CD86 bind to CD28, a receptor CD86 have been reported to encode for soluble products (6–12). on T cells. Engagement of CD28 by CD80 or CD86 is critical for However most of the biological activities attributed to CD80/ triggering the production by activated T cells of IL-2, a cytokine CD86 are based on studies of their membrane-bound forms, that induces T cell proliferation (2–4). which are believed to be the main players in providing costimu- CD80 and CD86 are type I membrane glycoproteins, and both latory signals to T cells. Nevertheless, one of the soluble CD86 are members of the Ig superfamily. They both consist of two ex- splice variants has been shown to be present at very small tracellular Ig-like domains that are linked to a transmembrane do- amounts in plasma and has been demonstrated to induce the main and a cytoplasmic tail. The membrane distal Ig domains of production of IL-2 in T cells (7).

http://classic.jimmunol.org In mammals, the CD80 and CD86 genes are closely linked in chromosome 3 and have very similar genomic organizations and *Department of Pathobiology, School of Veterinary Medicine, University of Penn- sylvania, Philadelphia, PA 19104; and †Clear Springs Foods Inc., Research Division, domain structures (13). Thus, it is likely that these two genes arose Buhl, ID 83316 by gene duplication from a common ancestor. Very little is known, Received for publication February 25, 2009. Accepted for publication April 24, 2009. however, about the structure, function, and evolution of CD80/ The costs of publication of this article were defrayed in part by the payment of page CD86 molecules in nonmammalian species. Besides mammals, charges. This article must therefore be hereby marked advertisement in accordance CD80 and CD86 molecules have only been cloned in birds (14). Downloaded from with 18 U.S.C. Section 1734 solely to indicate this fact. From a functional perspective, our knowledge of CD80/CD86 co- 1 This work was supported by U.S. Department of Agriculture Grants USDA-NRI stimulatory roles comes from studies performed largely in mam- 2006-01619 and USDA-NRI 2007-01719 (to J.O.S.) and by U.S. National Science Foundation Grant NSF-MCB-0719599 (to J.O.S.). malian species. In this regard, a major outcome of CD28 engage- 2 The nucleotide sequences of rainbow trout CD80/86 and IL-2 presented in this ment by CD86 or CD80 is the subsequent increase in IL-2 gene article have been submitted to GenBank under accession numbers EU927451– expression and protein synthesis by activated T cells (2–4). In EU927453, FJ467621–FJ467623, and FJ607781 (for CD80/86) and FJ571512 and FJ571513 (for IL-2). addition, it is also well known that some mitogens (such as PHA) 3 Y.-A.Z. and J.-i.H. contributed equally to this work. can induce the production of IL-2 by T cells (15). Although the main role of IL-2 on naive T cells is to promote their proliferation 4 Address correspondence and reprint requests to Dr. J. Oriol Sunyer, Department of Pathobiology, University of Pennsylvania School of Veterinary Medicine, 413 upon Ag activation (16), it can also induce apoptosis of activated Rosenthal Building, 3800 Spruce Street, Philadelphia, PA 19104. E-mail address: T cells (17). Moreover, IL-2 is also known to exert a variety of [email protected] biological roles on other immune cells, including NK cells (18), B Copyright © 2009 by The American Association of Immunologists, Inc. 0022-1767/09/$2.00 cells (19), and monocytes (20).

www.jimmunol.org/cgi/doi/10.4049/jimmunol.0900605 84 CHARACTERIZATION OF A PRIMORDIAL CD80/86 MOLECULE IN TROUT

Beyond mammals and birds, IL-2 has only been cloned in were retrieved from Atlantic salmon. 5Ј-RACE was performed initially Takifugu rubripes (fugu), a teleost fish species (21). Fugu IL-2 with the reverse primers sIL2-R1, -R2, and -R3, whose designs were based Ј appears to have a secondary structure comprised of three helices on the salmon EST sequences using the 5 -tailed cDNA from PHA-stim- ulated head kidney leukocytes as a template. Having isolated a 5Ј-end instead of the four typically found in IL-2 from mammals and partial sequence, the 3Ј-RACE was performed using the gene-specific birds. An important finding that strengthened the identification of primers IL2-F4 and IL2-F5 (see Table I) based on the sequenced 5Ј-end fugu IL-2 as a true ortholog of its mammalian counterpart was the cDNA of trout IL-2. conservation of synteny between the chromosomal areas contain- Identification of other B7 family members from teleost fish ing the fugu and mammalian IL-2 genes (21). Although the bio- logical roles of mammalian and avian IL-2 have been well char- Amino acid sequences for human B7-DC (NM_025239), B7-H1 (NM_014143), B7-H2 (NM_015259), B7-H3 (NM_001024736), and acterized, virtually nothing is known about the function of fish B7-H4 (NM_024626) were used in TBLASTN-based searches to identify IL-2. In this regard, the study on the fugu IL-2 molecule showed orthologous sequences within teleost fish using the NCBI, TIGR, and En- that its gene expression could be induced after the treatment of fish sembl databases as described above. leukocytes with PHA. However, its inducibility through the action Genomic analysis of the rtCD80/86 gene of a costimulatory molecule (like CD86 or CD80) remains to be investigated in fish. Genomic DNA was isolated from the head kidney of rainbow trout with a Thus, the objective of this study was 2-fold. First, to establish blood and cell culture DNA mini kit (Qiagen) according to the manufac- turer’s instructions. The trout genomic DNA was used as a PCR template. the presence of a CD80 or CD86 ortholog in teleost fish; second, The rainbow trout CD80/86 gene was PCR-amplified using the Platinum to evaluate whether the ability to induce IL-2 expression is con- Pfx DNA polymerase (Invitrogen) with three specific primer sets as fol- served in the fish CD80 or CD86 molecule. The identification of lows: 1) CD86-F2 and CD86ss-qR1; 2) CD86g-F2 and CD86g-R4; and 3) CD28 in teleost fish provided the rationale for the existence of CD86m-qF3 and CD86ml-R3 (where F is forward primer, R is reverse primer, g refers to genomic amplification, q refers to quantitative real-type CD80- or CD86-like molecules in these species (22, 23). The PCR, and ss is short secretory isoform; see Table I). The PCR profile was aforementioned hypothesis is now supported in this study by the as follows: 94°C for 3 min followed by 33 cycles of 94°C for 30 s, 56°C cloning of a rainbow trout (rt)5 gene (rtCD80/86) whose de- for 30 s, and 68°C for 3 min and finally 68°C for 5 min. The PCR products duced primary sequence and overall structure appear to resem- were cloned and sequenced as described below. ble those of both CD80 and CD86. We subsequently identified Cloning and sequence analysis and cloned trout IL-2 with the goal of evaluating a putative involvement of trout CD80/86 in IL-2 expression. We found The obtained PCR products were cloned into pCR-Blunt II-TOPO vector (Invitrogen). Plasmid DNA was isolated from positive colonies using the that recombinant rtCD80/86 induced the up-regulation of IL-2 QIAprep spin miniprep kit (Qiagen) and sequenced with an ABI 377 se- expression on trout leukocytes. quencer (Applied Biosystems) at the DNA Core Facility of the School of Veterinary Medicine, University of Pennsylvania (Philadelphia, PA). Materials and Methods Comparing the obtained nucleotide and amino acid sequences for sim- Fish ilarity with those in GenBank was performed using BLAST. Prediction of the open reading frame (ORF), the signal peptide, and the transmembrane Rainbow trout (Oncorhynchus mykiss) (50–100 g) obtained from Lime- helices was performed with the programs Translate, SignalP, and stone Springs Fish Farm were maintained in aquarium tanks using a water TMHMM, respectively, at the ExPASy proteomics server (ca.expasy.org/).

by guest on October 1, 2021. Copyright 2009 Pageant Media Ltd. recirculation system involving extensive biofiltration, UV sterilization Multiple sequence alignments were generated using ClustalW (version units, and thermostatic temperature control. Water temperature was main- 1.81) and BioEdit (version 7.0.9.0), and phylogenetic trees were con- tained continuously at 16–17°C. The fish were fed daily with commercial structed from the ClustalW-generated alignments using the maximum par- trout pellets, and feeding was terminated 48 h before sacrifice. Fish were simony methods within the PHILIP program (version 3.68) and were boot- acclimated to the laboratory conditions for at least 2 wk before being used strapped 1000 times. in experiments. Expression and purification of recombinant trout CD80/86 in Isolation of cDNAs for rtCD80/86 and IL-2 Escherichia coli The amino acid sequences for human CD80 and CD86 (GenBank acces- The pET-23b expression vector (Novagen) was modified by replacing the sion nos. NM_005191 and NM_175862, respectively) were used in N-terminal T7 tag with a 6ϫHis tag and used for constructing the expres- TBLASTN-based searches to identify orthologous sequences within the

http://classic.jimmunol.org sion plasmids. The DNA fragments encoding for the extracellular domain teleost expressed sequence tag (EST) databases at the National Center for (aa 30–231) of rtCD80/86m (where the suffix “m” designates the mem- Biotechnology Information (NCBI; www.ncbi.nlm.nih.gov/) and The In- brane isoform) were amplified from a plasmid containing rtCD80/86m us- stitute for Genomic Research (TIGR; www.tigr.org/db.shtml) and the ge- ing the primers CD86(E)-F and CD86(X)-R (Table I) and the proofreading nome database at Ensembl (www.ensembl.org/index.html). ESTs display- Platinum Pfx DNA polymerase (Invitrogen). The resulting amplification ing similarity to human CD80 and CD86 were retrieved from Atlantic was digested with EcoRI and XhoI and ligated to the modified pET23b salmon (Salmo salar) and other teleost species. Forward primers sCD86- vector, which was digested with the same restricted enzyme. The plasmid F1, -F2, and -F3 (see Table I) were designed based on the salmon EST DNA was prepared and sequenced to confirm that the construct (p23bHis-

Downloaded from sequences for 3Ј-RACE amplification of trout CD80/86 cDNA fragments. rtCD80/86) encodes the extracellular domain of rtCD80/86 fused with an For 5Ј-RACE of trout CD80/86, cDNA produced from head kidney leu- N-terminal 6ϫHis tag for purification. As a negative control, an expression kocytes was tailed with dATP and terminal deoxynucleotidyl transferase plasmid (p23bHis-rtIgD) containing an unrelated gene (encoding the con- (Promega); PCR was then performed with the gene-specific reverse prim- stant domain (aa 347–845) of trout IgD (GenBank accession no. ers CD86-R1, -R2, and -R3, whose designs were based on the sequenced AY870261)) was constructed as described above for rtCD80/86. 3Ј-RACE amplification products of trout CD80/86. The obtained products The expression plasmids p23bHis-rtCD80/86 and p23bHis-rtIgD were were cloned and sequenced as described below. subsequently transformed in BL21(DE3)-CodonPlus-RIL expression host The amino acid sequence for fugu IL-2 (GenBank accession no. competent cells (Stratagene). Bacterial cells were cultured in Luria-Bertani NM_001037994) was used to search homologous sequences within salmo- broth. Recombinant proteins were expressed by induction with isopropyl- nid EST databases at NCBI and TIGR. Candidate ESTs (TIGR: TC77909; ␤-D-thiogalactopyranoside (IPTG) and purified from inclusion bodies. NCBI: EG762969 and EG764237) encoding for putative IL-2 sequences Briefly, 10 ml of an overnight-saturated culture of bacteria transformed with the aforementioned constructs were inoculated into 1 liter of Luria- ␮ ␮ 5 Bertani broth containing 100 g/ml ampicillin and 34 g/ml chloramphen- Abbreviations used in this paper: rt, rainbow trout (prefix); EST, expressed sequence icol and grown at 37°C with shaking. The OD of this culture was al- tag; HKL, head kidney leukocyte; IgC, Ig constant region-like; IgV, Ig variable re- 600 lowed to reach 0.6 before induction with 0.4 mM IPTG and allowed to gion-like; IPTG, isopropyl-␤-D-thiogalactopyranoside; m, membrane isoform (suffix); ORF, open reading frame; PKC, protein kinase C; qPCR, quantitative real-time PCR; grow for an additional 4–6 h. Cells were harvested by centrifugation and sl, long secretory isoform (suffix); ss, short secretory isoform (suffix); TM, transmem- resuspended in 25 ml of native lysis buffer (50 mM NaH2PO4, 300 mM brane; UTR, untranslated region. NaCl, 10 mM imidazole, and 10 mM 2-ME (pH 8.0)). Cell suspensions The Journal of Immunology 85

Table I. Primers used for gene cloning and expression

Gene Primer Sequences (5Ј–3Ј) Application

a Ј Ј Adaptors APT CCAGACTCGTGGCTGATGCA(T)16V 3 -or5-RACE AP CCAGACTCGTGGCTGATGCA 3Ј-or5Ј-RACE CD80/86 sCD86-F1b CACCCAGAGTATGCAAACCGTA 3Ј-RACE sCD86-F2b AGGGGCTGTACGAATGCCACA 3Ј-RACE sCD86-F3b CCGACAACGGTTACCCTCCTA 3Ј-RACE CD86-R1 GGTAACCGTTGTCAGAGGAACA 5Ј-RACE CD86-R2 CTGTAGTTGGCTGTCACACTGA 5Ј-RACE CD86-R3 CATTCGTACAGCCCCTCATCT 5Ј-RACE CD86-F1 GGCTGTTAAGAAAGGCAAAGCA cDNA amplification CD86m-R1 GGATTCCTTCAGGTCAGCTGTA cDNA amplification CD86s-R1 ACACAGTATCTCAGTCCTCACA cDNA amplification CD86-F2 GCTGCTGTATGCTACAATGGCA Genomic amplification CD86g-F2 AACCTCAGCTGTGCTGTAGG Genomic amplification CD86m-R2 CACCTCAGCAACCCCTTTATCA Genomic amplification CD86ml-R3 GTTTTCATTGGTTCTTCTGTGGA Genomic amplification CD86g-R4 CATCCTGTCTCCTCTATCAC Genomic amplification CD86(E)-F AAGGAATTCAAAGATGTCACCATGGTGATAc CD80/86 production CD86(X)-R AATCTCGAGTCAGCAGACAGTGTGTTCCTGTd CD80/86 production IL-2 sIL2-R1e GCTAAATGTAATATGTTCTGAAGGA 5Ј-RACE sIL2-R2e TTCACAGGTGCATTCTGAAGTGT 5Ј-RACE sIL2-R3e TTATCGAGGCATTCTGCTTTCAC 5Ј-RACE IL2-F4 TACACCTTACAGGACACTATCTA 3Ј-RACE IL2-F5 TCTACAAACCCTTCAGATTATCAT 3Ј-RACE ␤-Actin Bactin-F ATGGAAGATGAAATCGCC RT-PCR Bactin-R TGCCAGATCTTCTCCATG RT-PCR IL-2 IL2-F4 TACACCTTACAGGACACTATCTA RT-PCR IL2-R4 CGAGGCATTCTACTTTCACAGT RT-PCR ␤-Actin Bactin-qF GGACTTTGAGCAGGAGATGG Real-time PCR Bactin-qR ATGATGGAGTTGTAGGTGGTCT Real-time PCR IL-1␤ IL1b-qF ACCGAGTTCAAGGACAAGGA Real-time PCR IL1b-qR CATTCATCAGGACCCAGCAC Real-time PCR TNF-␣ TNFa2-qF CAAGAGTTTGAACCTCATTCAG Real-time PCR TNFa2-qR GCTGCTGCCGCACATAAAG Real-time PCR CD80/86m CD86m-qF3 GTGTTTCCTGGTTCTGGTATCTA Genomic amplification and real-time PCR CD86m-qR3 AACTTGCTGCTCCCTTTCCTC Real-time PCR by guest on October 1, 2021. Copyright 2009 Pageant Media Ltd. CD80/86sl CD86-qF1 TAGTGAACAGCAGTGGGTGGA Real-time PCR CD86sl-qR1 GAGCCTTCTTTTTCTTCCTGAG Real-time PCR CD80/86ss CD86-qF1 TAGTGAACAGCAGTGGGTGGA Real-time PCR CD86ss-qR1 AATGACAGTGATAGAAACACAGTA Real-time PCR and genomic amplification IL-2 IL2-F6 TTTCCTTTTTGACGCTTTTTCTCA Real-time PCR IL2-R4 CGAGGCATTCTACTTTCACAGT Real-time PCR

a V ϭ A, C, or G. b Salmon CD86. c The restriction enzyme EcoRI recognition sequence is underlined. d http://classic.jimmunol.org The restriction enzyme XhoI recognition sequence is underlined. e Salmon IL2.

were mechanically disrupted using an SLM Aminco French pressure cell desalting columns (GE Healthcare). As control fractions, we collected the press (SLM Instruments) and then centrifuged at 12,000 ϫ g for 20 min to fractions from the unrelated recombinant protein sample eluting at the pellet the inclusion bodies. same position (elution time) as recombinant trout CD80/86. Control frac-

Downloaded from Inclusion bodies were solubilized with 25 ml of denaturing lysis buffer tions were treated identically as the fractions containing CD80/86 and,

(50 mM NaH2PO4, 10 mM Tris-HCl, 8 M urea, and 10 mM 2-ME (pH thus, were also concentrated and desalted into PBS (pH 7.5) with PD-10 8.0)). Recombinant proteins with a His tag were batch bound to a Ni-NTA desalting columns (GE Healthcare). Concentration of the purified recom- Superflow resin (Qiagen) by rocking at room temperature for 1 h. The binant protein was determined by SDS-PAGE and Bradford assay (Bio- Ni-NTA resin was washed three times with 30 ml of denaturing wash Rad). Endotoxin (LPS) concentration within the purified recombinant pro- buffer (containing the same component as denaturing lysis buffer (pH 6.5)) tein samples and control samples was tested at the Cell Center Services before being packed into an empty column. The bound protein was eluted facility at the University of Pennsylvania. using a denaturing elution buffer (containing the same component as the denaturing lysis buffer (pH 5.8)). Fractions containing the recombinant Isolation of PBLs and head kidney leukocytes (HKLs) rtCD80/CD86 or the unrelated recombinant protein were pooled and re- folded by gel filtration using a Pharmacia Superdex 200 column (GE Collection of trout leukocytes from blood and head kidney was done as Healthcare). The column was previously equilibrated with the refolding previously reported by us (24, 25). Briefly, fish blood was obtained from buffer (20 mM Tris-HCl, 150 mM NaCl, 400 mM arginine, 2 mM reduced the caudal vein and immediately diluted (1/5) with DMEM (Invitrogen) glutathione, and 0.2 mM oxidized glutathione (pH 7.5)), and protein frac- supplemented with 100 U/ml penicillin, 100 ␮g/ml streptomycin, and 25 tions were eluted at 0.5 ml/min with the same refolding buffer using a fast U/ml heparin and then placed on ice. Trout head kidneys were removed protein liquid chromatography instrument (GE Healthcare). Protein elution aseptically and pressed through a 100-␮m nylon mesh and suspended in was monitored by absorbance at 280 nm. The fractions containing trout DMEM supplemented with 100 U/ml penicillin, 100 ␮g/ml streptomycin, CD80/86 were concentrated and desalted into PBS (pH 7.5) with PD-10 and 25 U/ml heparin. The cell suspensions from head kidney or blood were 86 CHARACTERIZATION OF A PRIMORDIAL CD80/86 MOLECULE IN TROUT

layered onto a 51/34% discontinuous Percoll (GE Healthcare) density gra- reverse primer, and q refers to quantitative real-time PCR; see Table I). The dients and centrifuged at 400 ϫ g for 30 min. The band of leukocytes lying specificity of the PCR amplification for all primer sets was verified from at the interface was collected and the cells were washed three times with the dissociation curves. The relative expression levels for trout CD80/86 or DMEM supplemented with 100 U/ml penicillin, 100 ␮g/ml streptomycin, IL-2 were determined using the trout ␤-actin gene as an internal reference and 10 U/ml heparin. Leukocytes were used for the experiments described (normalizer). The p values were calculated by Student’s t test. in the sections below. ϩ RT-PCR was performed using Taq DNA polymerase (Invitrogen) with Rainbow trout IgM B cells were isolated from PBLs using a MACS the primer sets for trout IL-2 (IL2-F4 and IL2-R4) and ␤-actin (Bactin-F separator and the cell separation column LS (Miltenyi Biotec) with a bi- and Bactin-R) (Table I). The amplification of the trout ␤-actin gene served otinylated anti-trout IgM Ab (mAb 1.14). Ten million PBLs were used for Ϫ as an internal reference for each sample. The PCR profile for IL-2 expres- the MACS isolation according to the manufacturer’s instruction. The IgM sion was as follows: 94°C for 3 min, 36 cycles of 94°C for 30 s, 60°C for cells in the flow-through were incubated at 17°C in DMEM for2hina 30 s, and 72°C for 25 s followed by 72 °C for 1 min. The PCR profile for 24-well cell culture plate, and then the nonadherent IgMϪ cells (in the ␤-actin was 94°C for 3 min and 25 cycles of 94°C for 30 s, 55°C for 30 s, medium) and the remaining adherent cells were collected separately for and 72°C for 30 s followed by 72 °C for 1 min. The PCR products were total RNA extraction. separated on 1.6% agarose gels and visualized by staining with 100 ␮g/ml Stimulation of PBLs and HKLs with LPS and Vibrio bacterin ethidium bromide solution (Invitrogen). The isolated PBLs or HKLs (5 ϫ 105 cells per 100 ␮l of medium per well in a 96-well microplate (Nalge Nunc)) were incubated at 17°C with 50 ␮g/ml LPS (E. coli 0111:B4; Invitrogen) or with formalin-killed Vibrio Results anguillarum/Vibrio ordalii bacterin (Vibrogen-2; Novartis Animal Health) Isolation and sequence analysis of trout CD80/86 cDNA at doses of 1/200 and 1/20 of the original bacterin in DMEM supplemented Using the amino acid sequences of human CD80 and CD86 to with 10% FBS in an incubator with 5% CO2. After 24 and 48 h of incu- bation, the stimulated cells were collected, washed with PBS and then search salmonid EST databases, we found several Atlantic salmon subjected to total RNA extraction. EST sequences (NCBI: EG833543, DW580717, DR695874, Induction of IL-2 expression by PHA, PMA, and rtCD80/86 DW580718; TIGR: TC111220) displaying 23–29% identity to ei- ther CD80 or CD86. The full-length cDNA for a rtCD80/86 or- PBLs (5 ϫ 105 cells per 100 ␮l of medium per well in a 96-well micro- plate) were incubated at 17°C with PHA (0, 1, 10, and 100 ng/ml; tholog was obtained by the RACE technique using primers based Sigma-Aldrich) or PMA (0, 0.1, and 10 ␮g/ml; Sigma-Aldrich) to an- on the aforementioned salmon EST sequences and the cDNA from alyze the ability of the aforementioned mitogens to induce IL-2 gene rainbow trout HKLs as a template. Using forward primers (Table ϫ 5 expression. In a different set of experiments, PBLs (5 10 cells per I) for 3Ј-RACE, we obtained several partial cDNA products en- 100 ␮l medium per well in a 96-well microplate) were incubated at 17°C with recombinant rtCD80/86 at doses of 2 and 5 ␮g/ml in an coding for a molecule that showed roughly equal sequence identity

incubator with 5% CO2. In parallel, PBLs were also incubated with to mammalian CD80 and CD86; thus, we decided to name this equivalent volumes of control fractions or PBS as negative controls (see molecule rtCD80/86. Based on the consensus cDNA sequence of above paragraphs in this section under the heading Expression and pu- these rtCD80/86 fragments, 5Ј-RACE was performed using gene- rification of recombinant trout CD80/86 in Escherichia coli). To test the effect on the expression of IL-2 by the trace amounts of contami- specific reverse primers (Table I). Two full-length cDNA se- nating LPS present in the rtCD80/86 and control fractions, PBLs were quences were subsequently compiled, one of which was comprised also incubated with 0.1 and 10 ␮g/ml LPS (E. coli 0111:B4; Invitro- of 1361 bp (the membrane isoform rtCD80/86m, GenBank acces- gen). After 12 h of incubation, cells were collected and washed with by guest on October 1, 2021. Copyright 2009 Pageant Media Ltd. sion no. FJ467621; Fig. 1A) while the other one was 940 bp (short PBS and then subjected to total RNA extraction. secretory isoform (ss) rtCD80/86ss, GenBank accession no. Total RNA extraction and cDNA synthesis FJ467623; Fig. 1B). The secretory isoform is probably generated Total RNA was extracted from rainbow trout lymphoid and nonlymphoid by a differential use of the transcription termination signal that tissues (whole blood, head kidney, trunk kidney, spleen, thymus, heart, gill, results in the exclusion of both transmembrane and cytoplasmic liver, brain, intestine, skin, and muscle), stimulated PBLs and HKLs, as well as sorted IgMϩ and IgMϪ cells. Total RNA was extracted with the domains. The ORFs of the membrane (EU927451; 283 aa) and RNeasy mini kit (Qiagen). Isolated RNA was treated with an RNase-free secretory (EU927452; 227 aa) isoforms of trout CD80/86 were DNase set (Qiagen) according to the manufacturer’s instruction. To nor- confirmed by PCR with primer pairs CD86-F1/CD86m-R1 and malize gene expression levels for each sample, equivalent amounts of the total CD86-F1/CD86s-R1, respectively (Table I). Interestingly, among http://classic.jimmunol.org RNA (100 ng) were used for cDNA synthesis with the SuperScript first-strand synthesis system for RT-PCR (Invitrogen) in a 20-␮l reaction volume. The the PCR products amplified with primer pair CD86-F1/CD86m- synthesized cDNA (20 ␮l) was then diluted with 80 ␮l of RNase-free water R1, we obtained an additional secretory isoform (rtCD80/86sl, and used as a template for regular RT-PCR or real-time PCR. EU927453; 251 aa) slightly larger than the aforementioned Quantitative real-time PCR (qPCR) and regular RT-PCR rtCD80/86ss secretory product (Fig. 1B). This alternatively spliced analyses of trout CD80/86 and IL-2 expression variant is characterized by the loss of a large portion of exon 6, which encodes for the transmembrane (TM) domain. The absence Downloaded from qPCR was used to determine the abundance of trout CD80/86 and IL-2 transcripts in tissues and stimulated cells. The qPCRs were performed in of the highly hydrophobic segment represented by the TM domain duplicate and each contained 4 ␮l of a diluted cDNA template (4 ng of total in both secretory isoforms was further demonstrated by hydropho- RNA equivalents), 5 ␮l of Power SYBR Green PCR master mix (2ϫ; bicity plot analysis (data not shown), strongly suggesting that both Applied Biosystem), and 150 nM forward and reverse primers (except for isoforms are not membrane bound. IL-2 primers, which were 250 nM; see Table I) in a 10-␮l reaction volume. The amplification profile consisted of an initial denaturation step at 95°C The deduced amino acid sequence of rtCD80/86m exhibits a for 10 min, and then 40 cycles of 95°C for 15 s and 60°C for 1 min typical structure for members of the Ig superfamily, including followed by melting (dissociation stage) from 72°C to 95°C in an ABI mammalian CD80 and CD86 (Fig. 1B). Thus, it consists of an Prism 7500 sequence detection system (Applied Biosystems). extracellular Ig variable region-like (IgV) domain and an Ig con- Amplification efficiencies for all qPCR primer sets were analyzed fol- lowing the protocol reported elsewhere (26). For all of the primer sets used stant region-like (IgC) domain followed by a TM domain and a in this study, the efficiencies were Ͼ90% as shown: 97% for CD86m-qF3 cytoplasmic tail in which two potential sites for protein kinase C and CD86m-qR3 (R2 Ͼ 0.998); 94% for CD86-qF1 and CD86sl-qR1 (R2 Ͼ (PKC) phosphorylation at locations 255 (S) and 263 (T) were pre- 2 Ͼ 0.998); 92% for CD86-qF1 and CD86ss-qR1 (R 0.999); 99% for dicted. The extracellular regions of mammalian CD80 and CD86 IL1b-qF and IL1b-qR (R2 Ͼ 0.986); 102% for TNFa2-qF and TNFa2-qR (R2 Ͼ 0.992); 91% for IL2-F6 and IL2-R4 (R2 Ͼ 0.998); and 93% for are highly glycosylated (27, 28). Similarly, one and seven potential Bactin-qF and Bactin-qR (R2 Ͼ 0.999) (where F is forward primer, R is N-linked glycosylation sites (N-X-T/S) exist in the IgV and the The Journal of Immunology 87

FIGURE 1. Sequences and do- main structure of rainbow trout CD80/86 transcripts. A, Nucleotide and deduced amino acid sequences of rtCD80/86m (GenBank accession no. FJ467621). The start and stop codons are in boldface. The putative polyad- enylation signals are in boldface and italics. The predicted signal peptide and TM hydrophobic region are in boldface and underlined. The poten- tial N-glycosylation sites are in bold- face and circled. B, Domain structure of trout CD80/86 transcripts. The sequences have been submitted to GenBank (rtCD80/86m, EU927451; rtCD80/86ss, EU927452; rtCD80/ 86sl, EU927453). All three trout CD80/86 transcripts consist of an IgV domain (Ig-V) followed by an IgC domain (Ig-C). The membrane-bound form, rtCD80/86m, has a TM domain and a cytoplasmic tail that contains two putative PKC phosphorylation sites. The small gray circles on the Ig domains indicate N-glycosylation sites. S---Sdenotes interchain disul- by guest on October 1, 2021. Copyright 2009 Pageant Media Ltd. fide bonds. The circled P denotes pu- tative PKC phosphorylation sites pre- dicted with the NetPhosK 1.0 server. http://classic.jimmunol.org

Downloaded from IgC domains of rtCD80/86m, respectively (Fig. 1; see supplemen- TM domains, the highest percentage of identity is to chicken tal Table I).6 Among the eight glycosylation sites, the only one (30%) and mouse (28%) CD86, whereas the trout cytoplasmic located in the IgV domain (N100; Fig. 2A) is conserved between domain showed a higher degree of identity to both mouse CD80 rtCD80/86m and mammalian/avian CD80. However, within the (27%) and human CD86 (27%). When making comparisons us- IgC domain of rtCD80/86m there are three glycosylation sites con- ing the full-length sequences, we found that rtCD80/86m was served in the mammalian CD80 (locations 161, 208, and 229) or slightly more similar to the known CD86 sequences (18–19%) CD86 (locations 156, 168, and 200) sequences. than to those of CD80 (16–19%). Importantly, rtCD80/86m A comparison of each domain of rtCD80/86m with the cor- showed a higher degree of sequence identity to CD80 or CD86 responding domains of human, mouse, and chicken CD80 and members (16–19%) than to any of the other five members (B7- CD86 (Table II) revealed that the signal peptide, IgV, and IgC DC, B7-H1, B7-H2, B7-H3, and B7-H4) of the B7 family (11– domains of rtCD80/86m showed the highest degree of homol- 16%, see supplemental Table I). ogy with the corresponding domains of mouse CD86 with As shown in the multiple alignment (Fig. 2A), rtCD80/86m ex- amino acid identities of 21, 25, and 21%, respectively. For the hibits many structural features common to mammalian CD80 and CD86 molecules. The cysteine residues critical to form the inter- 6 The online version of this article contains supplemental material. chain disulfide bonds within the IgV (C56 and C127) and IgC 88 CHARACTERIZATION OF A PRIMORDIAL CD80/86 MOLECULE IN TROUT by guest on October 1, 2021. Copyright 2009 Pageant Media Ltd. http://classic.jimmunol.org Downloaded from

FIGURE 2. Comparison of trout CD80/86 with known CD80 and CD86 sequences. A, Multiple sequence alignments of trout CD80/86m with mammalian and avian CD80 and CD86 molecules. The identical and similar residues from the aligned sequences are shaded black and gray, respectively. Dashes indicate gaps introduced to maintain maximal sequence alignment. The secondary structure of the human CD80 is shown above the alignment. Conserved cysteine residues within IgV and IgC domains are indicated by a dollar sign ($) below the alignment. The amino acids from human CD80 participating in the binding with CD28/CTLA-4 are indicated by an asterisk (*) above the alignment (30, 53), whereas those from CD86 are indicated by an asterisk below the alignment (32). The amino acids contributing to the CD80 (IgV domain) homodimer interface are indicated by a filled circle (F) above the alignment (29, 31). Within the IgC domains, the amino acids contributing to CD80/CTLA-4 binding are indicated by a pound sign (#) (33, 54). The conserved N-glycosylation sites (N-X-T/S) between trout CD80/86 and other CD80 or CD86 are indicated by an ampersand (&) above or below the alignment, whereas those conserved between trout CD80/86 and chicken CD80/CD86 are indicated by an “at” sign (@) above and below the alignment. GenBank accession numbers of the sequences used in the multiple alignments are shown as follows: human (Hum) CD80, NM_005191; mouse (Mou) CD80, NM_009855; chicken (Chi) CD80, AM050146); Hum CD86, NM_175862; Mou CD86, NM_019388; Chi CD86, AM050135; trout (Tro) CD80/86, EU927451. B, Hydrophobicity profiling of human CD80, CD86, and trout CD80/86 molecules. The profile shows a hydrophobic signal peptide and a TM region as well as a hydrophilic C terminus for all three molecules. The plot was generated using the Kyte-Doolittle method. The Journal of Immunology 89

Table II. Protein homologies between trout CD80/86 and other (C180 and C241) domains of mammalian B7 molecules are con- a vertebrate CD80 and CD86 served in rtCD80/86m. Interestingly, the IgC domain of rtCD80/ 86m possesses four more cysteines (locations 166, 171, 230, and Identity (%) 262) than their mammalian counterparts. Similarly, the corre- Protein Species Signal IgV IgC TM CYTb Overall sponding domain of chicken CD80 also contains four additional cysteines, two in the IgV domain (C57 and C145) and two in the CD80 H. sapiens 61615238 16 M. musculus 13 15 18 18 27 19 IgC domain (C171 and C262), whereas the IgV domain of chicken G. gallus 10 23 10 8 13 16 CD86 contains only one additional cysteine (C149) (Fig. 2A). CD86 H. sapiens 13 18 9 19 27 18 In mammals, the IgV domains of CD80 and CD86 are respon- M. musculus 21 25 21 28 6 19 sible for binding to their receptors (CD28 and CTLA-4) (29). The G. gallus 15 18 18 30 18 18 binding occurs through the interaction of the CDR3-analogous c c CD80/86 S. salar 80 86 86 ND ND 85 loop (MYPPPY loop) from CTLA-4 and CD28, with the CD80/ D. rerio 17 29 31 13 18 26 ␤ Ј G. aculeatus 11 29 40 21 20 30 CD86 surface comprised primarily of the -strands G, F, C, C , O. latipes 29 23 25 13 20 19 and CЉ (29–32). The MYPPPY loops have previously been re- T. rubripes NDc 36 41 13 14 33 ported to be conserved in both the CD28 and the CTLA-4 of te- a Percentages indicate amino acid identity. Protein sequences were aligned with leost fish (22, 23); thus, it would be conceivable that the residues ClustalW (Version 1.81). Signal cleavage sites were predicted with SignalP (version involved in receptor binding in rtCD80/86m are similar to those of 3.0). The transmembrane domains were predicted with TMHMM (version 2.0). Bold- face indicates the highest identity percentage between each domain of rtCD80/86m the mammalian CD80 and CD86 molecules. Confirming the pre- and the corresponding domains of tetrapod. vious hypothesis, we found that the residues Y72, Q74, D77, E126, b Cytoplasmic domain. c Percentage could not be determined because of the lack of sequence data. by guest on October 1, 2021. Copyright 2009 Pageant Media Ltd. http://classic.jimmunol.org Downloaded from

FIGURE 3. Phylogenetic tree of vertebrate B7 family members. The tree was constructed in accordance with the maximum parsimony method of the PHYLIP program, using the amino acid sequences of the IgV and IgC domains. Percentage values shown for each node represent 1000 bootstrap replications. In the tree construction, GenBank accession numbers for the teleost fish B7 molecules reported in this paper are shown in supplemental Table II; GenBank accession numbers for known B7 family members are as follows: human (Hum; Homo sapiens) CD80, NM_005191; mouse (Mous; Mus musculus) CD80, NM_009855; rat (Rat; Rattus norvegicus) CD80, NM_012926; hamster (Ham; Mesocricetus auratus) CD80, AB085742; dog (Dog; Canis lupus familiaris) CD80, NM_001003147; sheep (Shee; Ovis aries) CD80, AY390555); pig (Pig; Sus scrofa) CD80, NM_214087; rabbit (Rabb; Oryctolagus cuniculus) CD80, D49843; chicken (Chic; Gallus gallus) CD80, AM050146; human CD86, NM_175862; mouse CD86, NM_019388; dog CD86, NM_001003146; sheep CD86, AY395982; pig CD86, AY826403; rabbit CD86, D49842; chicken CD86, AM050135; human B7-H1, NM_014143; mouse B7-H1, NM_021893; pig B7-H1, NM_001025221; chicken B7-H1, XM_424811; human B7-DC, NM_025239; mouse B7-DC, BC104138; chicken B7-DC (EST), CF251191; human B7-H2, NM_015259; mouse B7-H2, NM_015790; chicken B7-H2, Y08823; human B7-H3, NM_025240; mouse B7-H3, NM_133983; rat B7-H3, AY190319; chicken B7-H3, XM_413702; human B7-H4, NM_024626; mouse B7-H4, AY280973; chicken B7-H4, XM_416546. The fish species involved in this tree are rainbow trout (Trou; O. mykiss), Atlantic salmon (Sal; S. salar), medaka (Meda; O. latipes), zebrafish (Zebf; D. rerio), stickleback (Stib; G. aculeatus), fugu (Fugu; T. rubripes), and tetraodon (Tetr; Tetraodon nigroviridis). 90 CHARACTERIZATION OF A PRIMORDIAL CD80/86 MOLECULE IN TROUT

FIGURE 4. Schematic diagrams showing the gene organization of trout CD80/86 and human B7 family members. A, PCR cloning strategy and exon/intron organization of the rtCD80/86 gene and the three differ- ent rtCD80/86 transcripts. The letters a, b, and c indicate the genomic DNA fragments produced by PCR using the following primer sets: a, CD86-F2 and CD86ss-qR1; b, CD86g-F2 and CD86g-R4; and c, CD86m-qF3 and CD86 ml-R3 (see Table I). The cod- ing regions are indicated by solid black and gray boxes, and the un- translated regions (5Ј-or3Ј-UTR) by open boxes. The gray boxes indicate the regions in which a frame shift oc- curred in the rtCD80/86sl transcript. The size of each exon (above) and in- tron (below) of the rtCD80/86 gene is shown as bp numbers. B, Comparison of the exon/intron and domain orga- nization between trout CD80/86 and human B7 family members. Draw- ings are not proportional to the actual exon and intron sizes. GenBank ac- cession number and Ensemble gene identifiers are shown on the right.

and K138, which are critical for the interaction between mamma- thermore, the only two hydrophilic residues (D105 and T107) at

by guest on October 1, 2021. Copyright 2009 Pageant Media Ltd. lian CD80 and CTLA-4 (30), are conserved in rtCD80/86m. In the dimer interface of human CD80 are also conserved in trout contrast, only two similar residues (Y85 and S93) in rtCD80/86m CD80/86 and mouse CD80. show conservation with the corresponding area of mammalian Based on hydropathic analysis (Fig. 2B), the hydropathy plot of CD86 (32), indicating that the contact interface between rtCD80/ rtCD80/86m is very similar to that of human CD80 and CD86, 86m and its receptors is more similar to that of CD80 than to that showing a hydrophobic signal peptide and TM region, and a hy- of CD86. drophilic cytoplasmic region for all three molecules. Particularly in Within the IgC domain, the motif SQD(P/N)(E/V)(S/T)(E/ the extracellular IgV and IgC domains rtCD80/86m shows a nearly K)L(Y/F) between the ␤-strands D and E (residues 212–220) have identical hydrophobicity profile with that of human CD86, whereas been suggested to contribute also to the binding of CD80 and the hydrophobicity profile of the trout intracytoplasmic domain is http://classic.jimmunol.org CD86 molecules to their receptors CD28 and CTLA-4 (27). In more similar to that of human CD80. rtCD80/86m, a conserved motif exists at the corresponding loca- tion except for the residues at locations 212, 213, and 218 (Fig. 2A). By site-mutagenesis studies, four residues (Q213, D214, Identification of CD80/86 homologs from other teleost and E218, and L219) located in the aforementioned motif and another phylogenetic analysis three (F157, P160, and I162) in ␤-strand A were reported to par- Taking advantage of available genome databases (i.e., Ensembl ge-

Downloaded from ticipate in the contact interface between human CTLA-4 and CD80 nome database) from several teleost fish, we used the trout CD80/86m (33). Among these residues, P160, D214, and L219 are conserved sequence to search for homologous sequences from zebrafish (Danio in rtCD80/86m. rerio), stickleback (Gasterosteus aculeatus), medaka (Oryzias lati- In mammals, CD80 and CD86 adopt different oligomeric pes), and fugu (T. rubripes). Importantly, only one gene homologous states on the cell surface; CD80 exists predominantly as a dimer to rtCD80/86m could be identified from each fish species (see sup- whereas CD86 exists as a monomer (34). Roughly 13 aa within plemental Fig. 1 and supplemental Table II). None of the analyzed the IgV domain were revealed to contribute at the human CD80 genomes appeared to contain distinct orthologs for CD80 and CD86 homodimer binding interface (29, 31). Among these residues, genes. It should be stressed that in rainbow trout and other teleost fish eleven (V51, V62, L66, A67, G87, M91, I103, F104, I106, we found orthologs for all of the other members of the B7 family V113, and L115) are hydrophobic, whereas at the same loca- (B7-H1/DC, B7-H3, and B7-H4) except for B7-H2 (see Supplemental tions most (9/13) of the residues of human and mouse CD86 are Table II), thus strongly suggesting that rtCD80/86m and the other hydrophilic. Similar to mammalian CD80, many (7/13) of the teleost CD80/86m genes are not B7-H/DC members. The latter is residues in the corresponding areas of rtCD80/86 (V52, I61, further supported by the phylogenetic tree of Fig. 3, which shows that L66, P67, L91, V104, and W115) and mouse CD80 (V52, P61, teleost fish B7-H1, B7-H3, and B7-H4 molecules group with their G87, L103, Y104, I113, and L115) are also hydrophobic. Fur- mammalian counterparts whereas teleost CD80/86 members form a The Journal of Immunology 91

FIGURE 5. Expression patterns of rtCD80/86 transcripts in tissues and by guest on October 1, 2021. Copyright 2009 Pageant Media Ltd. blood leukocytes of rainbow trout. The rtCD80/86 expression in lymphoid and nonlymphoid tissues (A) and MACS-separated peripheral blood IgMϩ and IgMϪ cells (B) were analyzed by qPCR and normalized against the expression of ␤-actin. Abbreviations for tissues and cells are as follows: BL, blood; HK, head kidney; TK, trunk kidney; SP, spleen; TH, thymus; HE, heart, GI, gill, LI, liver; BR, brain; IN, intestine; SK, skin; MU, mus- cle, AD, adherent IgMϪ cells; and NA, nonadherent IgMϪ cells. The letters a, b, and c in B indicate that the differences of rtCD80/86m, rtCD80/86sl, and rtCD80/86ss expression between the IgMϩ cells and the adherent or Ϫ http://classic.jimmunol.org nonadherent IgM cells, respectively, are significant ( p Ͻ 0.05). The val- ues represent the mean Ϯ SEM of three individual fish. The p values were calculated by Student’s t test. FIGURE 6. Trout CD80/86 expression in PBLs and HKLs stimulated with E. coli LPS and Vibrio bacterin. A, Trout PBLs were stimulated with sister clade clustering together with the clade containing mammalian 50 ␮g/ml LPS or PBS for 12, 24, and 48 h. Expression of the three trout CD80/86 transcripts (rtCD80/86m, rtCD80/86sl, and rtCD80/86ss) as well Downloaded from and avian CD80 and CD86. as that of TNF-␣ and IL-1␤ was detected by qPCR and normalized against Genomic organization of trout CD80/86 the expression of PBS control (set as 1). The values represent the mean Ϯ SEM of six individual fish. Asterisks indicate significant differences be- ,ءء ;p Ͻ 0.05 ,ء) The genomic organization of the rtCD80/86 gene was determined by tween the PBS control (C) and LPS-treated leukocytes a PCR-based approach using rainbow trout genomic DNA as tem- p Ͻ 0.01). The p values were calculated by Student’s t test. Actual p values plate. The full-length rtCD80/86 gene was comprised of 5,328 bp were also shown on the 24-h graphs of rtCD80/86sl and rtCD80/86ss. B, (GenBank accession no. FJ607781), which was obtained by assem- Trout PBLs and HKLs were incubated with two doses of commercial for- bling three genomic DNA fragments amplified by PCR (Fig. 4A). The malin-killed Vibrio bacterin or with PBS for 24 and 48 h. V1 and V10 exon/intron boundaries of the three transcripts derived from the indicate leukocytes stimulated with 1/200 and 1/20 diluted bacterin, re- rtCD80/86 gene were also analyzed (Fig. 4A). The full-length spectively. The expression levels of rtCD80/86m, rtCD80/86ss, and TNF-␣ rtCD80/86 gene consists of eight exons and seven introns, including were detected by qPCR and were normalized against the expression of PBS Ϯ two large exons encoding for the IgV (303 bp) and IgC (279 bp) control (set as 1). The values represent the mean SEM of three individual fish. The asterisks indicate significant differences between PBS control (C) Ͻ ءء Ͻ ء -domains. Although the first four exons are common in all three tran Ј and bacterin-treated leukocytes ( , p 0.05; , p 0.01). The p values scripts, the 3 regions of the two soluble transcripts have a different were calculated by Student’s t test. alternative splicing pattern (Fig. 4A). Thus, the rtCD80/86sl transcript 92 CHARACTERIZATION OF A PRIMORDIAL CD80/86 MOLECULE IN TROUT

FIGURE 7. Multiple alignment of trout IL-2 with known IL-2 sequences. Identical and similar amino acid residues from the aligned sequences are shaded back and gray, respectively. Dashes indicate gaps introduced to maintain maximal sequence alignment. The signal peptide and ␣-helixes A–D of human IL-2 are denoted above the alignment. The IL-2 family signature (TELXXLXCLXXEL) found in mammalian sequences is shown below the alignment. The conserved cysteine residues critical for disulfide bond formation are indicated by a dollar sign ($) above (human IL-2) and below (fish IL-2) The putative N-glycosylation sites (N-X-T/S) are .(ء) the alignment. The biologically important Leu53, Asp56, and Gln173 are shown with an asterisk indicated by an ampersand (&) below the alignment. The identity percentages among trout IL-2 and other IL-2 molecules are shown at the end of each sequence. The accession numbers of the various IL-2 sequences are as follows: human (Hum), NM_000586; mouse (Mou), NP_032392; dog (Dog), NM_001003305; cat (Cat), NM_001043337; sheep (She), NM_001009806; pig (Pig), X56750; chicken (Chi), NM_204153; duck (Duc), AY193713; fugu (Fug), NM_001037994; and trout (Tro), FJ571512.

is translated differently due to a frame shift at the codon encoding the Regulation of rtCD80/86 gene expression by E. coli LPS and aa 225 located just after the TM domain (encoded by exon 6), whereas Vibrio bacterin

by guest on October 1, 2021. Copyright 2009 Pageant Media Ltd. the exon 5 is spliced normally, like that of the membrane isoform Regulation of rtCD80/86m, rtCD80/86ss, and rtCD80/86sl gene (rtCD80/86m). In contrast, the rtCD80/86ss transcript is spliced at the expression by LPS was studied in vitro using trout PBLs. After intron 5 and is ends abruptly at a stop codon. 12 h of incubation only the two secretory isoforms showed an Fig. 4B shows the comparison of the exon/intron organization increase, albeit minor, in their expression levels when compared between the rtCD80/86 gene and human genes of representative with control values. Gene expression levels of all three isoforms in- members of the B7 family. With regard to the exon number and creased moderately but in a significant manner, after 24 h of incuba- domain organization, the trout CD80/86 gene is almost identical to tion (Fig. 6A). After 48 h, only the membrane isoform showed a that of the human CD86 gene, except for the signal peptide region. significant increase in its expression levels. As a control of immune As can be observed, the gene organization of the IgV and IgC activation, we showed that IL-1␤ and TNF-␣ expression levels were http://classic.jimmunol.org domains is conserved in all members of the B7 family. significantly increased after 24 and 48 h of incubation. A similar gene Tissue distribution of trout CD80/86 transcripts expression trend was obtained with rtCD80/86m and one of the se- cretory isoforms (rtCD80/86ss) when incubating trout PBLs and The gene expression of rtCD80/86m, rtCD80/86sl, and rtCD80/ HKLs with a Vibrio bacterin (Fig. 6B). 86ss transcripts was analyzed by qPCR in several rainbow trout tissues (Fig. 5A) and PBLs (Fig. 5B). The expression levels of the

Downloaded from membrane isoform rtCD80/86m were the highest of the three iso- Cloning and expression analysis of trout IL-2 forms in all of the lymphoid tissues analyzed (Fig. 5A). Within the lymphoid tissues, the strongest expression levels of the membrane In mammals, a major effect of CD28 crosslinking by CD80 or isoform were detected in blood leukocytes. Interestingly, we found CD86 is the induction of IL-2 expression mainly by activated T that within blood leukocytes, IgMϩ B lymphocytes showed their cells (2–4). To be able to study whether trout CD80/86 had a highest levels of expression when compared with the levels ex- homologous role, we first had to identify and clone a trout IL-2 pressed by the adherent and nonadherent IgMϪ cell populations ortholog. (Fig. 5B). Within the nonlymphoid tissues, the expression of We initially identified two salmon EST sequences (GenBank rtCD80/86m was the highest in the gills and to a lesser degree in accession nos. EG762969 and EG764237) encoding for a molecule the liver, probably as a result of the large amount of blood leuko- with a high degree of sequence similarity to available IL-2 se- cytes contained in these two tissues. With regard to the two se- quences (data not shown). Based on these EST sequences, reverse cretory isoforms, they were both expressed at low levels in all primers were designed to perform 5Ј-RACE using rainbow trout tissues except for rtCD80/86ss, which showed the highest expres- cDNA isolated from the head kidney as a template. This yielded a sion levels of all three isoforms in tissues with a high content of cDNA fragment that was similar to that of Fugu IL-2, the only muscle cells, including the heart and muscle. teleost fish IL-2 molecule reported thus far (21). Subsequently, this The Journal of Immunology 93

FIGURE 9. Trout IL-2 expression in PBLs stimulated with recombinant trout CD80/86 protein (rtCD80/86). Trout PBLs were incubated with two doses of the rtCD80/86 (2 or 5 ␮g/ml) or an equal volume of concentrated control fractions. An additional control included trout PBLs incubated with an equal volume of PBS. After 12 h of incubation, the expression of trout IL-2 was detected by qPCR and normalized against the expression of the PBS control sample (set as 1). The values represent the mean Ϯ SEM of p Ͻ ,ء) seven individual fish. Asterisks indicate significant differences .p Ͻ 0.01). The p values were calculated by Student’s t test ,ءء ;0.05

We checked for the inducibility of the trout IL-2 gene by PMA and PHA, both of which are known to induce the expression of IL-2 in mammalian T cells (15, 35). Similarly, the expression of FIGURE 8. Purification and refolding of recombinant trout CD80/86 trout IL-2 was induced by both PMA and PHA in a dose-depen- (rtCD80/86). A, Refolding of rtCD80/86 by gel filtration chromatography. dent fashion in trout blood leukocytes (see supplemental Fig. 2).

by guest on October 1, 2021. Copyright 2009 Pageant Media Ltd. Affinity-purified rtCD80/86 (0.5 mg) was injected into a Superdex 200 gel filtration column (GE Healthcare) for refolding. The refolded rtCD80/86 Production of rtCD80/86 and its role in stimulating IL-2 (hatched peak) was eluted at a flow rate of 0.5 ml/min. Protein was mon- expression in trout PBLs ␮ itored by absorbance at 280 nm. B, Refolded rtCD80/86 (2 g) was ana- To study whether rtCD80/86 had a role homologous to that of its lyzed by electrophoresis on a 4–15% SDS-polyacrylamide gel under re- mammalian counterparts in the stimulation of IL-2 expression, we ducing conditions. Proteins were stained with Coomassie blue R250. Lane produced recombinant rtCD80/86 in which the TM and cytoplas- 1, purified and refolded rtCD80/86 (ϳ30 kDa); lane 2, lysate from E. coli 4 h after induction with 0.4 mM IPTG; M, molecular marker. mic domains were deleted. As shown in Fig. 8, we were able to produce a highly purified preparation of soluble rtCD80/86. After applying a refolding protocol that used gel filtration chromatogra-

http://classic.jimmunol.org phy (see Material and Methods), the recombinant rtCD80/86 ap- fragment was used to design gene-specific forward primers to per- peared to elute largely in the form of tetramers (Fig. 8A). Similar form 3Ј-RACE, which yielded a fragment corresponding to the to the aforementioned observation, the formation of dimers and 3Ј-end of trout IL-2. The compiled full-length cDNA sequence of other multimers has been shown also in recombinantly produced trout IL-2 contained a 67-bp 5Ј-untranslated region (UTR), a mammalian soluble CD80 molecules (36). With SDS-PAGE the 568-bp (IL-2A) or 76-bp (IL-2B) 3Ј-UTR, and a single ORF en- purified protein showed a molecular mass of 30 kDa, which is in

Downloaded from coding a 142-aa IL-2 molecule. Interestingly, five copies of an agreement with the expected size deduced from its amino acid mRNA instability motif (ATTTA) and a typical poly(A) signal sequence (Fig. 8B). It should be stressed that without the refolding (AATAAA) exists upstream of the poly(A) tail of IL-2A but not of step, all of the recombinant proteins formed insoluble high m.w. IL-2B (sequences not shown; GenBank accession nos. FJ571512 aggregates when run both by gel filtration and SDS-PAGE, thus and FJ571513 for IL-2A and IL-2B, respectively). indicating that the refolding step was critical for the production of When compared with other known IL-2 amino acid sequences, soluble tetramers of trout CD80/86 (data not shown). trout IL-2 is predicted to have a 20-aa signal peptide (Fig. 7). The Stimulation of trout blood leukocytes with two different con- ␣-helix B portion of trout IL-2 exhibits the IL-2 family signature centrations of recombinant rtCD80/86 lead to a significant up-reg- (T-E-[L/F]-X (2)-L-X-C-L-X (2)-E-L). Two cysteine residues lo- ulation of IL-2 gene expression in a dose-dependent manner (Fig. cated at positions 98 and 151 are conserved in all aligned mam- 9). Significantly, seven of the seven tested fish showed an increase malian, avian, and teleost IL-2 molecules, which are thought to be in IL-2 gene expression when the higher dose of recombinant involved in disulfide bond formation. Furthermore, an additional rtCD80/86 was used. In contrast, the control fraction sample did pair of cysteines (positions 91 and 148) were found only in the not induce any IL-2 gene expression when compared with the PBS avian and teleost IL-2 molecules. Two potential glycosylation sites control. It should be stressed that the levels of contaminating LPS (positions 65 and 140) were found in the trout IL-2 molecule. in the control fraction and the recombinant rtCD80/86 preparations 94 CHARACTERIZATION OF A PRIMORDIAL CD80/86 MOLECULE IN TROUT

were comparable, and very low (Ͻ0.08 ng/ml). We performed rainbow trout, only one rtCD80/86 homolog could be identified for experiments using much higher dose of E. coli LPS (1,000 ng/ml) each of the teleost species analyzed, and thus none of the analyzed that failed to induce increases in IL-2 expression (data not shown), fish genomes contained distinct orthologs for the mammalian thus indicating that the minute amounts of LPS contained in the CD80 and CD86 genes. This suggests that teleost CD80/86 rep- recombinant rtCD80/86 preparation did not have an influence on resents the ancestral gene from which CD80 and CD86 arose by the obtained IL-2 expression levels. Similarly LPS also failed to gene duplication before the emergence of birds, because birds have induce IL-2 expression in fugu leukocytes (21). already been reported to contain true orthologs of mammalian CD80 and CD86 genes (13, 14). During the writing of this article, Discussion a study was reported online that included a systematic in silico Thus far, CD80 and CD86 have only been cloned and functionally evolutionary analysis of B7 family members using available ge- characterized in mammals and birds. To shed light into the evo- nome and EST databases from lower vertebrates (37). This study lutionary origins and primordial functions of CD80 and CD86, our identified a CD80/86 gene in the genome of several teleost fish studies focused on the search for CD80 and CD86 orthologs in species that corresponds to the same CD80/86 gene we have iden- teleost fish. Thus, in the present study we have cloned and func- tified in the analyzed fish genomes (see supplemental Table II). tionally characterized a rainbow trout molecule (rtCD80/86) that Similar to our findings, this study showed the presence of one shows a high degree of homology to CD80 and CD86 molecules. single CD80/86 gene (termed B7R) in each of the analyzed fish In addition, we have identified and cloned trout IL-2 and have genomes. In agreement with our conclusion and supported by el- shown that recombinantly produced trout CD80/86 induces the egant syntenic analyses, the aforementioned study concludes that expression of IL-2 in trout blood leukocytes. fish B7R probably represents the ancestor from which CD80 and We initially identified salmon EST sequences with a high degree CD86 were derived in more evolved species (37). of homology to mammalian CD80 and CD86 molecules. Based on Tissue expression of rtCD80/86 was detected in all tissues an- these sequences, primers were designed to perform 5Ј- and 3Ј- alyzed. The membrane form (rtCD80/86m) was significantly more RACE using rainbow trout cDNA as a template. This led to the expressed in the lymphoid tissues, thus suggesting a larger in- cloning of a trout cDNA whose deduced amino acid sequence volvement of this isoform in fish immunity. However, the two showed comparable degrees of similarity with that of mammalian secretory forms were expressed at low levels both in lymphoid and CD80 or CD86, and thus we named this trout molecule rtCD80/86. nonlymphoid tissues, although the short secretory form rtCD80/ Three different transcripts for the rtCD80/86 gene were identified 86ss was predominantly expressed in the heart and muscle. Sim- by PCR, a membrane form (rtCD80/86m), a short secretory iso- ilarly as in rainbow trout, expression of membrane CD80 and form (rtCD80/86ss), and a long secretory isoform (rtCD80/86sl). CD86 in mammals is restricted to lymphoid cells including B cells, The short secretory isoform is lacking the exons coding for both macrophages/monocytes, dendritic cells, and activated T cells and the TM and cytoplasmic domains, whereas the longer secretory mainly play a role in immune modulation (13, 38). However, the isoform only lacks the TM domain. However, the latter isoform expression of membrane forms of mouse and rat CD80 and/or contains a frame shift at the C terminus, corresponding to the cy- CD86 molecules has also been shown in nonlymphoid tissues (i.e., toplasmic tail of the membrane form. Comparable soluble forms of heart, tongue, and liver), although to a much lesser degree than in

by guest on October 1, 2021. Copyright 2009 Pageant Media Ltd. CD80 and CD86 molecules have also been described in several lymphoid tissues (39, 40). Although CD80 and CD86 genes are not mammalian species (6–12). normally expressed in the human muscle cells, after activation Similar to mammalian CD80 and CD86 molecules, the trout with IL-4 or CD40 those genes can be induced in myoblasts (41). membrane form is composed of two extracellular Ig-like domains, Interestingly, we found that in blood leukocytes and HKLs, a hydrophobic TM region, and an intracellular cytoplasmic tail IgMϩ B cells showed a significantly higher degree of rtCD80/86m (Fig. 1B). Several lines of evidence strongly support the hypothesis expression when compared with the adherent and nonadherent that the rtCD80/86 gene represents an ortholog of mammalian and populations of IgMϪ leukocytes, suggesting an important role for avian CD80 and CD86 genes: 1) the deduced amino acid sequence the IgMϩ B cell as an APC. The putative role of trout IgMϩ B cells of rtCD80/86 showed higher degrees of similarity (16–19% iden- as APCs is perhaps related to the high phagocytic capacity that we http://classic.jimmunol.org tity) to the CD80 and CD86 sequences than to sequences of any of have recently reported for teleost fish B cells (42), a hypothesis the other five members of the B7 family (see supplemental Table that will be examined in future studies. I); 2) trout orthologs of mammalian B7-DC/H1, B7-H3, and In mammals, CD86 has been shown to be expressed constitu- B7-H4 were identified (see supplemental Table II), thus excluding tively at low levels and is rapidly up-regulated by a variety of the possibility that rtCD80/86 could be an ortholog of any the stimuli including LPS, whereas CD80 is expressed only inducibly aforementioned B7 molecules; 3) the rtCD80/86 molecule showed and later, after the activation of CD86 (1, 2, 43). In this regard

Downloaded from a closer phylogenetic relationship to mammalian and avian CD80 rtCD80/86 appeared to behave more like a CD86 molecule, as we and CD86 than to any of the other members of the B7 family (Fig. found that its gene expression was significantly up-regulated by E. 3); 4) the genomic organization of the rtCD80/86 showed the high- coli LPS (Fig. 6A) and Vibrio bacterin (see Fig. 6B). The mem- est degree of homology to that of human CD86 (Fig. 4B); 5) most brane form transcript (rtCD80/86m) showed the highest levels of of the key residues of mammalian CD80 and CD86 involved in the up-regulation and, although the increases in rtCD80/86m expres- binding to their receptors (CD28 and CTLA4) are conserved in sion were moderated, they were comparable to the up-regulation rtCD80/86 (Fig. 2A); and 6) the number of potential N-linked gly- levels induced by LPS for mammalian CD86 (44, 45). cosylation sites (eight sites) in rtCD80/86 resembled that of mam- Data on the functional roles of CD80 and CD86 are restricted to malian CD80 and CD86 molecules (eight sites) and is more than mammalian species (1, 13) and to a much lesser degree to birds that of any of the other members of the B7 family (4–7 sites) (see (14). Perhaps the most important functional outcome resulting supplemental Table I). from the engagement of CD80 and/or CD86 with CD28 is the We could only identify one CD80/86 gene in rainbow trout. To up-regulation of gene expression and protein synthesis of IL-2, verify whether other teleost fish also contained a single CD80/86 mainly by Th cells (2–4). Thus, we resolved to determine whether gene, we searched for rtCD80/86 homologs in available teleost fish this role in rtCD80/86 was conserved. To carry out such studies, genome drafts and EST databases. Similar to what we found in we needed first to characterize trout IL-2, because its identification The Journal of Immunology 95

had not thus far been reported. Therefore, we isolated cDNA en- IL-2 gene expression in T cells in a TCR-independent fashion coding for trout IL-2 that represents the second IL-2 molecule to (46, 47). be cloned in a teleost fish. The primary amino acid sequence de- In conclusion, this study represents the first cloning and func- duced from the trout IL-2 gene contains the key residues charac- tional characterization of an ancestral CD80/86 ortholog in a poiki- terizing the IL-2 family signature (T-E-[L/F]-X (2)-L-X-C-L-X lotherm species, as well as the cloning of the second IL-2 gene in (2)-E-L) and the two cysteine residues involved in disulfide bond a teleost fish. Significantly, we show that recombinantly produced formation that are conserved in all known IL-2 molecules. The rtCD80/86 up-regulates IL-2 expression levels in blood leuko- rainbow trout IL-2 showed the highest degree of sequence simi- cytes, thus indicating that this capacity to modulate IL-2 expres- larity to that of the recently reported fugu IL-2 molecule (21). The sion is a primordial function that has been conserved in both fish trout IL-2 transcript could be significantly up-regulated by PMA and mammalian CD80/CD86 molecules throughout 350 million and PHA, both of which are also known to up-regulate the expres- years of evolution. sion levels of IL-2 in mammals (15, 35). More importantly, IL-2 Although teleost fish appear to have most of the cellular and transcripts were able to be significantly up-regulated by recombi- molecular machinery necessary to elicit primary and secondary nantly produced rtCD80/86 (Fig. 9). Our data show that the addi- adaptive immune responses, it is clear that degree of such re- tion of recombinant rtCD80/86 alone to trout blood leukocytes sponses (e.g., specificity, affinity maturation, and duration) differ to provided a signal that was sufficient to induce an increase in the a great extent between fish and mammals. It is likely that differ- expression levels of IL-2. In this regard, IL-2 production in mam- ences in the mechanisms of Ag uptake and presentation between mals normally requires two signals induced by the interaction of fish and mammals account, at least in part, for some of the differ- the TCR with the MHC/peptide complex (signal 1) and by the ences observed in their primary and secondary responses. For ex- engagement of CD80 or CD86 with CD28, which provides a co- ample, we do not know yet which cells are the main APCs in stimulatory signal (signal 2) (2). However, and in support of our teleosts, as the presence of bona fide dendritic cells in these species results, it has been shown by several groups that engagement of remain to be demonstrated. Thus, the characterization of trout CD28 alone (signal 2) by crosslinked anti-CD28 Abs is sufficient CD80/86 and IL-2 molecules presented in this article will allow us to induce stimulation of IL-2 gene expression in T cells, although in the near future to determine further the mechanisms and cells to significantly lower levels than the simultaneous coengagement involved in Ag presentation in this species. In turn, this will enable us to identify some of the cellular and molecular elements respon- of CD28 and TCR (46–48). Moreover, other studies have also sible for the differences seen between immune responses of fish shown that up-regulation of IL-2 on T cells, or their activation, can and mammals. From a practical perspective, future studies on Ag happen independently of TCR engagement (49–51). presentation processes in fish will be instrumental for improving In this study the functional extracellular domain of rtCD80/86m delivery strategies and adjuvant formulations with the goal of de- was recombinantly produced in a prokaryotic expression system signing more efficacious fish vaccines. (E. coli). Similarly, a functional receptor binding domain of human CD86 (IgV-like domain) was also produced in E. coli (32, 52). This indicated that, similar to what we observed with rtCD80/86, Acknowledgments the bacterially produced human molecule retains the conforma- We thank Drs. Carsten Sanders and Serdar Turkarslan and Prof. Fevzi by guest on October 1, 2021. Copyright 2009 Pageant Media Ltd. Daldal for providing the French pressure cell press and related technical tional structure required to interact with its receptor and that gly- support. cosylation of the binding domain of CD86 does not play a role in binding to its receptor. In support of the latter, it has also been Disclosures shown that soluble deglycosylated human CD80 produced recom- The authors have no financial conflict of interest. binantly retains its ability to bind to both CD28 and CTLA-4 (36). Moreover, it has been reported that the potential N-linked glyco- References sylation sites in human CD80 are localized to regions essentially 1. Greenwald, R. J., G. J. Freeman, and A. H. Sharpe. 2005. The B7 family revis- opposite of the site for receptor interaction and are therefore un- ited. Annu. Rev. Immunol. 23: 515–548. 2. Sharpe, A. H., and G. J. Freeman. 2002. The B7-CD28 superfamily. Nat. Rev. http://classic.jimmunol.org likely to play a role in the interactions between CD80 with CD28/ Immunol. 2: 116–126. CTLA-4 molecules (33), a fact that further explains the observed 3. Carreno, B. M., and M. Collins. 2002. The B7 family of ligands and its receptors: binding of deglycosylated CD80 and CD86 products to their new pathways for costimulation and inhibition of immune responses. Annu. Rev. receptors. Immunol. 20: 29–53. 4. Wang, S., and L. Chen. 2004. T lymphocyte co-signaling pathways of the B7- It has been shown that CD80 but not CD86 forms homodimers CD28 family. Cell. Mol. Immunol. 1: 37–42. on cell surfaces (34). Recombinantly produced soluble human 5. Bhatia, S., M. Edidin, S. C. Almo, and S. G. Nathenson. 2006. B7-1 and B7-2: similar costimulatory ligands with different biochemical, oligomeric and signal- Downloaded from CD80 maintains also its capacity to form dimers as well as higher ing properties. Immunol. Lett. 104: 70–75. m.w. oligomers, depending on its concentration in solution (5, 30). 6. Yang, S., and G. K. Sim. 1999. New forms of dog CD80 and CD86 transcripts Similarly, our gel filtration studies showed that recombinant that encode secreted B7 molecules. Immunogenetics 50: 349–353. 7. Jeannin, P., G. Magistrelli, J. P. Aubry, G. Caron, J. F. Gauchat, T. Renno, rtCD80/86 did not form monomers in solution; instead, it appeared N. Herbault, L. Goetsch, A. Blaecke, P. Y. Dietrich, et al. 2000. Soluble CD86 to form tetramers. It has been shown that the residues at the dimer is a costimulatory molecule for human T lymphocytes. Immunity 13: 303–312. interface of human CD80 are primarily hydrophobic and, thus, 8. Faas, S. J., M. A. Giannoni, A. P. Mickle, C. L. Kiesecker, D. J. Reed, D. Wu, W. L. Fodor, J. P. Mueller, L. A. Matis, and R. P. Rother. 2000. Primary structure critical for inducing the dimerization of CD80 (30), whereas in the and functional characterization of a soluble, alternatively spliced form of B7-1. same area most of the residues of CD86 are hydrophilic. Similar to J. Immunol. 164: 6340–6348. 9. Yang, S., K. S. Sellins, T. Powell, E. Stoneman, and G. K. Sim. 2001. Novel mammalian CD80, many (seven of the 11 hydrophobic) residues in transcripts encoding secreted forms of feline CD80 and CD86 costimulatory mol- the corresponding area of rtCD80/86 are also hydrophobic, a fact ecules. Vet. Immunol. Immunopathol. 81: 15–21. that might explain the degree of oligomerization observed in re- 10. Terzo, E. A., M. P. de Villarreal, V. Mick, F. Mun˜oz, B. Amorena, D. de Andre´s, and J. M. Pe´rez de la Lastra. 2006. Molecular cloning of multiple forms of the combinant rtCD80/86 (Fig. 8). In addition, the oligomeric state of ovine B7-2 (CD86) costimulatory molecule. Vet. Immunol. Immunopathol. 114: recombinant rtCD80/86 may have contributed to is ability to effi- 149–158. 11. Choi, I., B. Cho, S. D. Kim, D. Park, J. Y. Kim, C. G. Park, D. H. Chung, ciently crosslink its ligand on trout T cells, similar to the reported W. S. Hwang, et al. 2006. Molecular cloning, expression and functional charac- and above-mentioned ability of crosslinked anti-CD28 to induce terization of miniature swine CD86. Mol. Immunol. 43: 480–486. 96 CHARACTERIZATION OF A PRIMORDIAL CD80/86 MOLECULE IN TROUT

12. Kakoulidou, M., R. Giscombe, X. Zhao, A. K. Lefvert, and X. Wang. 2007. 34. Bhatia, S., M. Edidin, S. C. Almo, and S. G. Nathenson. 2005. Different cell Human soluble CD80 is generated by alternative splicing, and recombinant sol- surface oligomeric states of B7-1 and B7-2: implications for signaling. Proc. uble CD80 binds to CD28 and CD152 influencing T-cell activation. Scand. J. Im- Natl. Acad. Sci. USA 102: 15569–15574. munol. 66: 529–537. 35. Grabstein, K., S. Dower, S. Gillis, D. Urdal, and A. Larsen. 1986. Expression of 13. Collins, M., V. Ling, and B. M. Carreno. 2005. The B7 family of immune- interleukin 2, interferon-␥, and the IL 2 receptor by human peripheral blood regulatory ligands. Genome Biol. 6: 223. lymphocytes. J. Immunol. 136: 4503–4508. 14. O’Regan, M. N., K. R. Parsons, C. A. Tregaskes, and J. R. Young. 1999. A 36. Davis, S. J., S. Ikemizu, A. V. Collins, J. A. Fennelly, K. Harlos, E. Y. Jones, and chicken homologue of the co-stimulating molecule CD80 which binds to mam- D. I. Stuart. 2001. Crystallization and functional analysis of a soluble deglyco- malian CTLA-4. Immunogenetics 49: 68–71. sylated form of the human costimulatory molecule B7-1. Acta Crystallogr. D 15. Kim, H. P., J. Imbert, and W. J. Leonard. 2006. Both integrated and differential Biol. Crystallogr. 57: 605–608. regulation of components of the IL-2/IL-2 receptor system. Cytokine Growth 37. Hansen, J. D., L. D. Pasquier, M. P. Lefranc, V. Lopez, A. Benmansour, and Factor Rev. 17: 349–366. P. Boudinot. 2009. The B7 family of immunoregulatory receptors: A comparative 16. Lan, R. Y., C. Selmi, and M. E. Gershwin. 2008. The regulatory, inflammatory, and evolutionary perspective. Mol. Immunol. 46: 457–472. and T cell programming roles of interleukin-2 (IL-2). J. Autoimmun. 31: 7–12. 38. Lenschow, D. J., T. L. Walunas, and J. A. Bluestone. 1996. CD28/B7 system of 17. Zambricki, E., A. Shigeoka, H. Kishimoto, J. Sprent, S. Burakoff, C. Carpenter, T cell costimulation. Annu. Rev. Immunol. 14: 233–258. E. Milford, and D. McKay. 2005. Signaling T-cell survival and death by IL-2 and 39. Inaba, K., M. Witmer-Pack, M. Inaba, K. S. Hathcock, H. Sakuta, M. Azuma, IL-15. Am. J. Transplant. 5: 2623–2631. H. Yagita, K. Okumura, P. S. Linsley, S. Ikehara, et al. 1994. The tissue distri- 18. Becknell, B., and M. A. Caligiuri. 2005. Interleukin-2, interleukin-15, and their bution of the B7-2 costimulator in mice: abundant expression on dendritic cells roles in human natural killer cells. Adv. Immunol. 86: 209–239. in situ and during maturation in vitro. J. Exp. Med. 180: 1849–1860. 19. Armitage, R. J., B. M. Macduff, J. Eisenman, R. Paxton, and K. H. Grabstein. 40. Damoiseaux, J. G., H. Yagita, K. Okumura, and P. J. van Breda Vriesman. 1998. 1995. IL-15 has stimulatory activity for the induction of B cell proliferation and Costimulatory molecules CD80 and CD86 in the rat; tissue distribution and ex- differentiation. J. Immunol. 154: 483–490. pression by antigen-presenting cells. J. Leukocyte Biol. 64: 803–809. 20. Civallero, M., S. Barni, R. Nano, and E. Capelli. 2000. Dendritic cells and in- 41. Behrens, L., M. Kerschensteiner, T. Misgeld, N. Goebels, H. Wekerle, and terleukin-2: cytochemical and ultrastructural study. Histol. Histopathol. 15: R. Hohlfeld. 1998. Human muscle cells express a functional costimulatory mol- 1077–1085. ecule distinct from B7.1 (CD80) and B7.2 (CD86) in vitro and in inflammatory 21. Bird, S., J. Zou, T. Kono, M. Sakai, J. M. Dijkstra, and C. Secombes. 2005. lesions. J. Immunol. 161: 5943–5951. Characterisation and expression analysis of interleukin 2 (IL-2) and IL-21 ho- 42. Li, J., D. R. Barreda, Y. A. Zhang, H. Boshra, A. E. Gelman, S. Lapatra, L. Tort, mologues in the Japanese pufferfish, Fugu rubripes, following their discovery by and J. O. Sunyer. 2006. B lymphocytes from early vertebrates have potent phago- synteny. Immunogenetics 56: 909–923. cytic and microbicidal abilities. Nat. Immunol. 7: 1116–1124. 22. Bernard, D., B. Riteau, J. D. Hansen, R. B. Phillips, F. Michel, P. Boudinot, and 43. Hathcock, K. S., G. Laszlo, C. Pucillo, P. Linsley, and R. J. Hodes. 1994. Com- A. Benmansour. 2006. Costimulatory receptors in a teleost fish: typical CD28, parative analysis of B7-1 and B7-2 costimulatory ligands: expression and func- elusive CTLA4. J. Immunol. 176: 4191–4200. tion. J. Exp. Med. 180: 631–640. 23. Bernard, D., J. D. Hansen, L. Du Pasquier, M. P. Lefranc, A. Benmansour, and 44. Borriello, F., J. Oliveros, G. J. Freeman, L. M. Nadler, and A. H. Sharpe. 1995. P. Boudinot. 2007. Costimulatory receptors in jawed vertebrates: conserved Differential expression of alternate mB7-2 transcripts. J. Immunol. 155: CD28, odd CTLA4 and multiple BTLAs. Dev. Comp. Immunol. 31: 255–271. 5490–5497. 24. Boshra, H., J. Li, R. Peters, J. Hansen, A. Matlapudi, and J. O. Sunyer. 2004. 45. Delgado, M., W. Sun, J. Leceta, and D. Ganea. 1999. VIP and PACAP differ- Cloning, expression, cellular distribution, and role in chemotaxis of a C5a re- entially regulate the costimulatory activity of resting and activated macrophages ceptor in rainbow trout: the first identification of a C5a receptor in a nonmam- through the modulation of B7.1 and B7.2 expression. J. Immunol. 163: malian species. J. Immunol. 172: 4381–4390. 4213–4223. 25. Boshra, H., T. Wang, L. Hove-Madsen, J. Hansen, J. Li, A. Matlapudi, C. J. Secombes, L. Tort, and J. O. Sunyer. 2005. Characterization of a C3a 46. Williams, T. M., D. M. Moolten, H. Makni, H. W. Kim, J. A. Kant, and receptor in rainbow trout and Xenopus: the first identification of C3a receptors in M. Kamoun. 1992. CD28-stimulated IL-2 gene expression in Jurkat T cells oc- nonmammalian species. J. Immunol. 175: 2427–2437. curs in part transcriptionally and is cyclosporine-A sensitive. J. Immunol. 148: 26. Pfaffl, M. W., G. W. Horgan, and L. Dempfle. 2002. Relative expression software 2609–2616. tool (REST) for group-wise comparison and statistical analysis of relative ex- 47. Siefken, R., S. Klein-Hessling, E. Serfling, R. Kurrle, and R. Schwinzer. 1998. A pression results in real-time PCR. Nucleic Acids Res. 30: e36. CD28-associated signaling pathway leading to cytokine gene transcription and T 27. Freeman, G. J., G. S. Gray, C. D. Gimmi, D. B. Lombard, L. J. Zhou, M. White, cell proliferation without TCR engagement. J. Immunol. 161: 1645–1651.

by guest on October 1, 2021. Copyright 2009 Pageant Media Ltd. J. D. Fingeroth, J. G. Gribben, and L. M. Nadler. 1991. Structure, expression, and 48. Rudd, C. E., and M. Raab. 2003. Independent CD28 signaling via VAV and T cell costimulatory activity of the murine homologue of the human B lympho- SLP-76: a model for in trans costimulation. Immunol. Rev. 192: 32–41. cyte activation antigen B7. J. Exp. Med. 174: 625–631. 49. Baroja, M. L., J. L. Ceuppens, J. Van Damme, and A. Billiau. 1988. Cooperation 28. Freeman, G. J., J. G. Gribben, V. A. Boussiotis, J. W. Ng, V. A. Restivo, Jr., between an anti-T cell (anti-CD28) monoclonal antibody and monocyte-produced L. A. Lombard, G. S. Gray, and L. M. Nadler. 1993. Cloning of B7-2: a CTLA-4 IL-6 in the induction of T cell responsiveness to IL-2. J. Immunol. 141: counter-receptor that costimulates human T cell proliferation. Science 262: 1502–1507. 909–911. 50. Verwilghen, J., P. Vandenberghe, G. Wallays, M. de Boer, N. Anthony, 29. Schwartz, J. C., X. Zhang, S. G. Nathenson, and S. C. Almo. 2002. Structural G. S. Panayi, and J. L. Ceuppens. 1993. Simultaneous ligation of CD5 and CD28 mechanisms of costimulation. Nat. Immunol. 3: 427–434. on resting T lymphocytes induces T cell activation in the absence of T cell 30. Ikemizu, S., R. J. Gilbert, J. A. Fennelly, A. V. Collins, K. Harlos, E. Y. Jones, receptor/CD3 occupancy. J. Immunol. 150: 835–846. D. I. Stuart, and S. J. Davis. 2000. Structure and dimerization of a soluble form 51. Unutmaz, D., P. Pileri, and S. Abrignani. 1994. Antigen-independent activation of B7-1. Immunity 12: 51–60. of naive and memory resting T cells by a cytokine combination. J. Exp. Med. 180: http://classic.jimmunol.org 31. Stamper, C. C., Y. Zhang, J. F. Tobin, D. V. Erbe, S. Ikemizu, S. J. Davis, 1159–1164. M. L. Stahl, J. Seehra, W. S. Somers, and L. Mosyak. 2001. Crystal structure of 52. Zhang, X., J. C. Schwartz, S. C. Almo, and S. G. Nathenson. 2002. Expression, the B7-1/CTLA-4 complex that inhibits human immune responses. Nature 410: refolding, purification, molecular characterization, crystallization, and prelimi- 608–611. nary X-ray analysis of the receptor binding domain of human B7-2. Protein Expr. 32. Zhang, X., J. C. Schwartz, S. C. Almo, and S. G. Nathenson. 2003. Crystal Purif. 25: 105–113. structure of the receptor-binding domain of human B7-2: insights into organiza- 53. Fargeas, C. A., A. Truneh, M. Reddy, M. Hurle, R. Sweet, and R. P. Sekaly. tion and signaling. Proc. Natl. Acad. Sci. USA 100: 2586–2591. 1995. Identification of residues in the V domain of CD80 (B7-1) implicated in 33. Peach, R. J., J. Bajorath, J. Naemura, G. Leytze, J. Greene, A. Aruffo, and functional interactions with CD28 and CTLA4. J. Exp. Med. 182: 667–675. P. S. Linsley. 1995. Both extracellular immunoglobin-like domains of CD80 54. Guo, Y., Y. Wu, M. Zhao, X. P. Kong, and Y. Liu. 1995. Mutational analysis and Downloaded from contain residues critical for binding T cell surface receptors CTLA-4 and CD28. an alternatively spliced product of B7 defines its CD28/CTLA4-binding site on J. Biol. Chem. 270: 21181–21187. immunoglobulin C-like domain. J. Exp. Med. 181: 1345–1355. Supplemental Figure legends

Supplemental FIGURE 1. Multiple alignment of teleost fish CD80/86 sequences. The identical (black) and similar (grey) residues from the aligned sequences are shaded.

Dashes indicate gaps introduced to maintain maximal sequence alignment. Conserved cysteine residues critical for disulfide bond formation within IgV and IgC domains are indicated by “$”, while additional cysteine residues are indicated by “©”. The sequences used in the multiple alignment are represented in Supplemental table II.

Supplementary FIGURE 2. Trout IL-2 expression in PBLs stimulated with PMA and

PHA. The trout PBLs were incubated with two doses of PMA (0.1 and 10 µg/ml) or three doses of PHA (1, 10, and 100 ng/ml), or with PBS, for 12 hours. IL-2 gene expression was detected by RT-PCR. The expression of β-actin was detected as an internal control.

The sizes of the PCR products are shown on the right.

Supplemental Table I. Comparison between trout CD80/86 and human B7 family molecules

Alternative Accession Length Glycosylation Mw1 pI Identity Protein names number (aa) sites (kDa) (pH) (%)

IgV: 1 2 B7R EU927451 283 27.6 8.34 100 rtCD80/86 IgC: 7

B7-1 IgV: 3 3 B7 NM_005191 288 29.3 6.43 16 huCD80 IgC: 5 CD28LG1 B7-2 IgV: 2 B70 NM_175862 329 35.2 6.78 18 huCD86 IgC: 6 CD28LG2 CD273 IgV: 2 PD-L2 NM_025239 273 28.7 8.49 15 huB7-DC IgC: 3 PDCD1LG2 CD274 IgV: 1 PD-L1 NM_014143 290 31.0 6.49 13 huB7-H1 IgC: 3 PDCD1LG1 CD275 IgV: 2 ICOS-L NM_015259 302 31.4 4.95 16 huB7-H2 IgC: 4 B7RP-1

IgV: 2 4 CD276 NM_025240 316 30.9 4.49 16 huB7-H3 IgC: 2

B7X IgV: 1 B7S1 NM_024626 282 28.4 5.20 11 huB7-H4 IgC: 6 VTCN1

1 Theoretical molecular weight before modification. 2 rt = rainbow trout. 3 hu = human. 4 The huB7-H3 used in this table is the isoform containing only two Ig domains.

Supplemental Table II. Genes for B7 family molecules identified in teleost fish

Protein Species Chromosomal location Ensembl Gene ID mRNA1

CD80/86 Salmo salar ND2 ND TIGR: TC111220

Chr 15 Danio rerio ND GB: CN017839 (39,718,341-39,735,630)

Group I Gasterosteus aculeatus ENSGACG00000008066 GB: DN656970 (6,449,986-6,452,113)

Chr 13 Oryzias latipes ENSORLG00000010731 TIGR: TC69013 (20,863,824-20,873,683)

Scaffold 172 Takifugu rubripes ND ND (289,229-293,023)

B7-H1/DC Oncorhynchus mykiss ND ND TIGR: TC155993

-H1/DCa Salmo salar ND ND TIGR: TC80363

-H1/DCb ND ND TIGR: TC76785

-H1/DCc3 ND ND TIGR: TC72076

Chr 21 Danio rerio ENSDARG00000059310 GB: DN833503 (548,496-549,822) Group XIV GB: DT956548 Gasterosteus aculeatus ENSGACG00000017360 (7,425,305-7,428,114) GB: DT956717 Scaffold 243 Takifugu rubripes ENSTRUG00000005708 GB: BU808255 (101,609-104,788) Chr 4 Tetraodon nigroviridis ENSTNIG00000004945 ND (1,300,607-1,302,512)

TIGR: TC138276 Oncorhynchus mykiss ND ND B7-H3 GB: CA352727

GB: DY700751 Salmo salar ND ND TIGR: TC71395

Chr 18 GB: CU638774 Danio rerio ENSDARG00000003061 (44,906,707-44,943,735) GB: NM_001080622

Group II GB: DN679146 Gasterosteus aculeatus ENSGACG00000016547 (16,172,372-16,188,797) GB: DW637643

GB: DK193899 Chr 3 Oryzias latipes ENSORLG00000013032 GB: DK244157 -H3a (27,032,398-27,037,739) GB: DK216942

-H3b4 GB: BJ889017

Scaffold 418 Takifugu rubripes ENSTRUG00000002758 GB: CA589617 (8,352-12,021) Chr 5 Tetraodon nigroviridis ENSTNIG00000011968 ND (1,812,347-1,836,725)

B7-H4 Oncorhynchus mykiss ND ND GB: CA369063

Salmo salar ND ND TIGR: TC76415

Chr 24 Danio rerio ENSDARG00000062729 GB: XM_001337711 (23,816,932-23,825,189) Group XVI GB: DN659573 Gasterosteus aculeatus ENSGACG00000008295 (16,313,860-16,316,272) GB: DN658251 Scaffold 1479 ENSORLESTG00000000078 Oryzias latipes TIGR: TC64208 (19,349-29,399) ENSORLG00000019982 Scaffold 433 Takifugu rubripes ENSTRUG00000001237 GB: CA846291 (7,348-9,093) Chr 2 Tetraodon nigroviridis ENSTNIG00000015655 ND (19,805,631-19,807,784) Scaffold 4244 ENSTNIG00000002118 (47,937,244-47,938,319)

1 If assembled sequence was available from TIGR, the EST sequences from NCBI (GB) were not shown. 2 ND, not detectable. 3 Soluble isoform (without TM domain) of salmon B7-H1/DC. 4 Truncated isoform (without IgC domain) of medaka B7-H3.