Evidence for Evolving Toll-IL-1 Receptor-Containing Adaptor Molecule Function in Vertebrates

This information is current as Con Sullivan, John H. Postlethwait, Christopher R. Lage, of September 30, 2021. Paul J. Millard and Carol H. Kim J Immunol 2007; 178:4517-4527; ; doi: 10.4049/jimmunol.178.7.4517 http://www.jimmunol.org/content/178/7/4517 Downloaded from

References This article cites 60 articles, 30 of which you can access for free at: http://www.jimmunol.org/content/178/7/4517.full#ref-list-1 http://www.jimmunol.org/ Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

by guest on September 30, 2021 *average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2007 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Evidence for Evolving Toll-IL-1 Receptor-Containing Adaptor Molecule Function in Vertebrates1

Con Sullivan,* John H. Postlethwait,† Christopher R. Lage,* Paul J. Millard,‡ and Carol H. Kim2*

In mammals, Toll-IL-1R-containing adaptor molecule 1 (TICAM1)-dependent TLR pathways induce NF-␬B and IFN-␤ re- sponses. TICAM1 activates NF-␬B through two different pathways involving its interactions with TNFR-associated factor 6 and receptor-interacting 1. It also activates IFN regulatory factor 3/7 through its interaction with TANK-binding kinase-1, leading to the robust up-regulation of IFN-␤. In this study, we describe the role of zebrafish (Danio rerio) TICAM1 in activating NF-␬B and zebrafish type I IFN. Zebrafish IFN is unique in that it cannot be categorized as being ␣-or␤-like. Through comprehensive sequence, phylogenetic, and syntenic analyses, we fully describe the identification of a zebrafish TICAM1 ortholog.

Zebrafish TICAM1 exhibits sequence divergence from its mammalian orthologs and our data demonstrate that these sequence Downloaded from differences have functional consequences. Zebrafish TICAM1 activates zebrafish IFN; however, it does so in an apparently IFN regulatory factor 3/7-independent manner. Furthermore, zebrafish TICAM1 does not interact with zebrafish TNFR-associated factor 6, thus NF-␬B activation is dependent upon its interaction with receptor-interacting protein 1. Comparative genome analysis suggests that TICAM1 and TICAM2 evolved from a common vertebrate TICAM ancestor following a duplication event and that TICAM2 was lost in teleosts following the divergence of the rayfin and lobefin fishes 450 million years ago. These studies provide evidence, for the first time, of the evolving function of a vertebrate TLR pathway. The Journal of Immunology, http://www.jimmunol.org/ 2007, 178: 4517–4527.

he recent discovery of a functional type I IFN in the ze- the signal cascade (reviewed in Refs. 10 and 11). One such adaptor brafish, Danio rerio, provided the first insight into how known as TIR domain-containing adaptor molecule 1 (TICAM1, T these essential antiviral cytokines evolved (1). Fish type I also known as TRIF) plays an essential role in TLR3 and TLR4 IFNs are unique in that they possess introns and, based upon phy- signaling, allowing for the up-regulation of IFN-␤ (12, 13). logenetic analyses, cluster away from mammalian type I IFNs (1, Mammalian TICAM1s are comprised of proline-rich N- and 2). The mechanisms underlying activation of fish type I IFNs have C-terminal domains and a central TIR domain essential for in- by guest on September 30, 2021 yet to be explored. In mammals, transcription of type I IFNs can be teractions with other TIR domains (12, 13). Each domain plays initiated by ligand activation of TLR3, TLR4, TLR7, TLR8, and an important role in signaling. The TIR domain is responsible TLR9 (3–9). TLRs transduce signals from extracellular stimuli for interacting with TLR3 or TICAM2. The N-terminal domain into the cell through interactions with adaptor molecules. Both can interact with TNFR-associated factor (TRAF)6 or form a TLRs and adaptor possess Toll/IL-1R (TIR)3 domains that complex consisting of TANK-binding kinase (TBK)-1, IFN reg- facilitate the protein-protein interactions responsible for triggering ulatory factors (IRF)3 and 7 (IRF7), TRAF3, I␬B kinase-related kinase ␧, and NAK-associated protein-1 (14–21). The C-termi- nal domain interacts with receptor-interacting protein (RIP)1 *Department of Biochemistry, Microbiology, and Molecular Biology, University of and RIP3. Maine, Orono, ME 04469; †Institute of Neuroscience, University of Oregon, Eugene, OR 97403; and ‡Department of Chemical and Biological Engineering and The Lab- In this study, we report the identification of a novel zebrafish oratory for Surface Science and Technology, University of Maine, Orono, ME 04469 TICAM1 ortholog that exhibits unique structural and functional Received for publication September 6, 2006. Accepted for publication January features. Our results provide a glimpse into the evolutionary his- 10, 2007. tory of TLR-mediated type I IFN induction in a basally diverging The costs of publication of this article were defrayed in part by the payment of page vertebrate TICAM1-mediated pathway. Our findings suggest that charges. This article must therefore be hereby marked advertisement in accordance ␬ with 18 U.S.C. Section 1734 solely to indicate this fact. zebrafish TICAM1 activates NF- B and type I IFN through mech- 1 This work was supported in part by Grant R15AI049237-02 (to C.S., C.R.L., and anisms not observed in mammalian TICAM1 orthologs and that C.H.K.) from the National Institute for Allergy and Infectious Disease, Grant R01 while zebrafish type I IFN does not group with other avian or RR10715 (to J.H.P.) from the National Center for Research Resources, and Grant mammalian clades (1, 2), its TICAM1-dependent induction indi- R15AI065509-01 (to P.J.M.) from National Institute for Allergy and Infectious Dis- ease. All institutes are components of the National Institutes of Health and its contents cates the presence of this potent antiviral pathway in early are solely the responsibility of the authors and do not necessarily represent the official vertebrates. views of National Center for Research Resources or National Institutes of Health. 2 Address correspondence and reprint requests to Dr. Carol H. Kim, Department of Biochemistry, Microbiology, and Molecular Biology, 5735 Hitchner Hall, University of Maine, Orono, ME 04469. E-mail address: [email protected] Materials and Methods 3 Abbreviations used in this paper: TIR, Toll/IL-1R; TICAM, TIR domain-containing Nomenclature adaptor molecule; TRAF, TNFR-associated factor; TBK, TANK-binding kinase; IRF, IFN regulatory factor; RIP, receptor-interacting protein; ISRE, IFN-stimulated regu- Nomenclature rules for zebrafish, mouse, and human and proteins latory element; ZFL, zebrafish liver cell; RHIM, RIP homotypic interaction motif; follow different conventions. To minimize confusion in presenting these HA, hemagglutinin; poly(I:C), polyinosinic-polycytidylic acid. data, gene names will be presented in italicized capital letters (e.g., TICAM1) and protein names will be presented in standard capital letters Copyright © 2007 by The American Association of Immunologists, Inc. 0022-1767/07/$2.00 (e.g., TICAM1). www.jimmunol.org 4518 TICAM SIGNALING DIFFERS IN BASALLY DIVERGING VERTEBRATES

DNA constructs Cell culture

Zebrafish orthologs of TICAM1, RIP1, and TBK1 were identified through basic 293H cells (Invitrogen Life Technologies) were cultured at 37°C, 6% CO2 local alignment search tool sequence analyses of zebrafish genome and ex- in DMEM (high glucose) supplemented with 10% heat-inactivated FBS. pressed sequence tag databases using human and mouse orthologs and ZFL cells were grown in sealed vessels at 28°C in LDF medium, which available zebrafish sequence data. Each zebrafish ortholog was cloned and consists of 50% Leibovitz’s L-15 medium, 35% DMEM (high glucose), deposited in GenBank (zebrafish TICAM1, accession no. DQ848679; ze- and 15% F-12 nutrient mixture (Ham) supplemented with 5% heat-inacti- brafish RIP1, accession no. DQ848680; zebrafish TBK1, accession no. vated FBS, 0.5% heat-inactivated SeaGrow trout serum (East Coast Bio- DQ860098). Zebrafish IFN promoter was cloned from a DraI-digested, logics), 50 ng mlϪ1 mouse epidermal growth factor, and 1ϫ insulin-trans- zebrafish genomic DNA pool through nested PCR, according to the pro- ferrin-selenium-X. Unless otherwise noted, all culture products were tocol outlined in the GenomeWalker Universal kit (BD Clontech). First- purchased from Invitrogen Life Technologies. round PCR used primers AP1 (5Ј-GTAATACGACTCACTATAGGGC- 3Ј) and IFN PRO (5Ј-GTTATTATCCTGTATCGGCCAAGC-3Ј); nested, Luciferase reporter assays second-round PCR used primers AP2 (5Ј-ACTATAGGGCACGCGTGGT- 3Ј) and IFN PRO NESTED (5Ј-CATTCGCAAGTAGACGCAGAG-3Ј), 293H and ZFL cells were plated in 24-well plates (Corning) so that they and the product from the first round. Subsequently, the IFN promoter were 90–95% confluent on the day of transfection. Using Lipofectamine (GenBank accession no. DQ855952) was subcloned in pGL3-Basic (Pro- 2000 (Invitrogen Life Technologies), 400 ng of TICAM1 construct (mouse TICAM1 TRAF6 TICAM1, zebrafish TICAM1, or indicated deletion construct), 400 ng of mega) for use in luciferase reporter assays. Mouse , , and ␬ RIP1, along with the NF-␬B-luciferase reporter vector pBIIx and the IFN- reporter construct (NF- B: pBIIx-luc; ISRE: ISRE-luc; zebrafish minimal stimulated regulatory element (ISRE)-luciferase reporter vector ISRE-luc, Mx (ISRE) promoter: Mx187-luc; or zebrafish IFN promoter: pGL3-IFN were provided by R. Medzhitov (Yale University, New Haven, CT). The Pro), and 10 ng of pRL-CMV, which served as a Renilla luciferase internal Mx187-luc reporter (22) was used as a secondary means to measure ISRE control to normalize data, were used to transfect 293H or ZFL cells. At 24 h posttransfection, 293H cells were lysed and luciferase activities were activity in zebrafish liver cells (ZFL). The zebrafish expression vector Downloaded from frm2bl (23) was a gift from P. Gibbs (University of Miami, Miami, FL). measured using the Dual-Luciferase Reporter Assay system (Promega). At pcDNA3.1-p35, which encodes the antiapoptotic protein p35, was provided 48 h posttransfection, ZFL cells were lysed and luciferase activities were by W. Kaiser and M. Offermann (Emory University, Atlanta, GA). measured using the Dual-Luciferase Reporter Assay system. All firefly For luciferase assays, mouse TICAM1, zebrafish TICAM1, and deletion luciferase light outputs were normalized to Renilla luciferase activities. TICAM1 DeepVent Data are represented as fold induction over empty vector control. Error bars constructs of zebrafish were amplified by PCR using Ն polymerase (New England Biolabs) and subcloned in the zebrafish expres- indicate the SEM for 3 replicates in a representative experiment. sion vector frm2bl at the KpnI and SpeI sites, resulting in the removal of a GFP stuffer fragment and the insertion of the desired TICAM1 fragment. Coimmunoprecipitation http://www.jimmunol.org/ Mouse TICAM1 possesses a KpnI restriction site in its coding sequence. To 293H cells were plated in 25-cm2 flasks (Corning) so that they were 90– facilitate its subcloning in frm2bl, mouse TICAM1’s native KpnI site was 95% confluent on the day of transfection. Using Lipofectamine 2000, cells destroyed without changing the amino acids it encodes via the QuikChange were transfected with indicated amounts of plasmids totaling 8 ␮g(3␮gof XL Site-Directed Mutagenesis kit (Stratagene). The N-terminal deletion zebrafish or mouse TICAM1 construct, 3 ␮g of interactor construct, 1 ␮g ⌬ ( N) of zebrafish TICAM1 lacks the sequence encoding the first 311 aa. of pAdVAntage, and 1 ␮g of pcDNA3.1-p35). At 48 h posttransfection, ⌬ The C-terminal deletion ( C) lacks the sequence encoding the final 81 aa. cells were washed twice with PBS (Invitrogen Life Technologies) and ⌬ ⌬ A construct consisting of the TIR domain alone ( N C) was created by lysed for 20 min in ice-cold buffer containing 50 mM Tris-HCl (pH 7.4), amplification of a product that lacked the nucleotide sequence encoding the 150 mM NaCl, 1 mM EDTA, and 1% Triton X-100 into which a general N-terminal 311 aa and C-terminal 81 aa. A zebrafish TICAM1 with a TIR use protease-inhibitor mixture containing 4-(2-aminoethyl)benzenesulfo- ⌬ domain deletion ( TIR) was created by PCR sewing. An N-terminal nyl fluoride hydrochloride, aprotinin, bestatin HCl, E-64, EDTA, and leu- by guest on September 30, 2021 TICAM1 with a C-terminal overlap was amplified along with a C-terminal peptin hemisulfate (Sigma-Aldrich) was added to 10% of total volume. TICAM1 with an N-terminal overlap. Amplicons were combined, dena- Cell lysates were centrifuged at 16,000 ϫ g for 15 min and soluble frac- tured, and spliced in the presence of DeepVent polymerase, which filled tions were collected. Soluble lysates were transferred to plugged Handee in the overlapping PCR product. Following overlap extension, the spliced MiniSpin columns (Pierce) containing 10 ␮l of packed gel volume of product was amplified, and the product was subcloned in frm2bl. An RIP washed anti-FLAG M2 or anti-HA agarose affinity gel resin (Sigma- homotypic interaction motif (RHIM) domain mutant was created by site- Aldrich). Samples were incubated on a rotating platform overnight at 4°C. directed mutagenesis. The resulting TICAM1 product contained changes of Following overnight incubation, samples were washed eight times in 25 amino acids 547–550 from Met-Ile-Gly-Asn to Ala-Ala-Ala-Ala (Fig. 1A). mM Tris-HCl (pH 7.2), 150 mM NaCl, 0.05% Tween 20, and eluted in 40 For coimmunoprecipitation assays, full-length mouse and zebrafish ␮lof2ϫ lane marker nonreducing sample buffer (Pierce) during a 5-min ϫ TICAM1 were subcloned in p3 FLAG-CMV-14 (Sigma-Aldrich) following incubation at 100°C. Samples were reduced in 2-ME, separated by PAGE, amplification with Phusion polymerase (New England Biolabs). An N-termi- and transferred to BioTrace nitrocellulose membranes (Pall Corporation) nal zebrafish TICAM1 construct lacking the first 311 aa was subcloned in using the Mini Trans-Blot cell (Bio-Rad). Membranes were blocked in ϫ p3 FLAG-CMV-14. Except for zebrafish TLR3, potential mouse and/or ze- TBS (pH 7.5) with 0.1% Tween 20 containing 5% nonfat dried milk and brafish TICAM1 interactors (zebrafish RIP1, zebrafish TBK1, mouse TRAF6, exposed to rabbit anti-FLAG polyclonal (Affinity Bioreagents), anti-HA and zebrafish TRAF6) were subcloned in pCMV3Tag9 (Stratagene). Zebrafish polyclonal (Invitrogen Life Technologies), or anti-myc polyclonal Ab (Af- TLR3 was subcloned in pcDNA3.1D/V5-His-TOPO with a hemagglutinin finity Bioreagents) overnight at 4°C. Following overnight incubation, Ј (HA) tag and a stop codon introduced at the 3 end by PCR. pcDNA3.1-p35 membranes were washed three times in TBST and incubated in goat anti- was transfected into cells used in coimmunoprecipitation experiments to rabbit secondary Ab conjugated to alkaline phosphatase (Bio-Rad) for 1 h counter the apoptotic effects of TICAM1. pAdVAntage plasmid (Promega) at room temperature. Membranes were washed three times in TBST. CDP- was included in each coimmunoprecipitation to bolster protein translation. Star AP substrate (Novagen) was applied to membranes according to the Phylogenetic reconstruction manufacturer’s recommendations and the membranes were exposed to film. Representatives from a broad range of TIR domain-containing proteins were aligned with DIALIGN (24), a multiple sequence alignment program Results that compares amino acid sequences via a segment-to-segment approach with no gap penalties imposed. This alignment method is useful in com- Identification of a full-length zebrafish TICAM1 ortholog paring proteins with similar domains that are otherwise unrelated. Molec- exhibiting considerable sequence divergence from mammalian ular phylogenetic analyses were performed using PHYLIP, software version orthologs 3.6b (distributed by the author, J. Felsenstein, University of Washington, Se- attle, WA at http://evolution.gs.washington.edu/phylip.html) (25). Sequences The TIR domain of the zebrafish TICAM1 was originally de- were bootstrapped 1000 times with the program SEQBOOT and these scribed by Meijer et al. (28). The TICAM1 TIR domain cDNA bootstrapped amino acid sequences were used to compute distance matri- sequence was used to search zebrafish expressed sequence tag and ces with the program PROTDIST, according to the Jones-Taylor-Thornton genome resources. A predicted full-length TICAM1 sequence was (26) model of amino acid replacement. Phylogenetic trees based upon these data were generated by the neighbor-joining method (27), using the pro- identified using the online resources of the Zebrafish Genome gram NEIGHBOR, and from these trees, an extended majority rule con- Project and zebrafish TICAM1 was cloned. The open reading frame sensus tree was created with the program CONSENSE. of zebrafish TICAM1 is 1701 bp. Comparisons of syntenic regions The Journal of Immunology 4519 Downloaded from http://www.jimmunol.org/ by guest on September 30, 2021

FIGURE 1. TICAM1 alignment and domain architecture. A, Alignment of human, mouse, and zebrafish TICAM1 orthologs using AlignX (Vector NTI; Invitrogen Life Technologies), a ClustalW-based algorithm. Dashes indicate gaps in amino acid sequence. The TRAF6-binding motif, TIR domain, and RHIM domain are identified by underlines and annotations. The arrow above the sequence Met-Ile-Gly-Asn at positions 547–550 identifies the amino acid residues targeted by QuikChange mutagenesis to make the Ala-Ala-Ala-Ala RHIM domain mutant. Zebrafish TICAM1 is 566 aa; human TICAM1 is 712 aa; and mouse TICAM1 is 732 aa. B, Both the N- (1–318 aa) and C-terminal (456–566 aa) domains of zebrafish TICAM1 are truncated, calling into question their capacity to interact with homologs of proteins known to interact in mammals.

among zebrafish, mouse, and human genomes indicate that ze- consensus positions 265–270 was identified in zebrafish TICAM1. brafish TICAM1 is orthologous to mammalian TICAM1s. Despite Some sequence divergence was also noted in the BB loop of TIR this conserved synteny, the zebrafish TICAM1 gene encodes a domain and the RHIM domain (consensus positions 682–720), smaller protein (566 aa) than does the mouse (732 aa) or human raising questions about zebrafish TICAM1’s capacity to interact ortholog (712 aa). A ClustalW-based alignment (AlignX, Vector with TLR3 and RHIM domain-containing proteins (RIP1, RIP3), NTI; Invitrogen Life Technologies) identified significant gaps in respectively. Zebrafish TICAM1 shares 26% sequence identity the N and C termini of zebrafish TICAM1 amino acid sequence and 46% sequence similarity with human and mouse across when compared with human and mouse, raising questions about aligned residues, excluding gaps. Zebrafish TICAM1 is com- zebrafish TICAM1’s capacity to transduce downstream signals prised of 318-aa N terminus, a 137-aa TIR domain, and a 110-aa (Fig. 1A). Furthermore, no identifiable TRAF6-binding motif at C terminus (Fig. 1B). 4520 TICAM SIGNALING DIFFERS IN BASALLY DIVERGING VERTEBRATES Downloaded from

FIGURE 2. TICAM1 and TICAM2 share a close evolutionary relationship.

A, Zebrafish TICAM1 forms a mono- http://www.jimmunol.org/ phyletic group with mouse and human TICAM1 and mouse and human TICAM2, exhibiting 100% bootstrap support. Representatives from a broad range of TIR domain-containing pro- teins were aligned with DIALIGN (55), a multiple sequence alignment program that compares amino acid se- quences via a segment-to-segment ap- proach with no gap penalties im- by guest on September 30, 2021 posed. All bootstrap values Ͻ91% are shown while those Ն91% are not shown. B, The TICAM paralogon from the . Segments of 1, 5, 15, and 19 are paralogous. Data from National Cen- ter for Biotechnology Information Gene Database: (www.ncbi.nlm.nih. gov//query.fcgi?db ϭ gene). The Journal of Immunology 4521

FIGURE 3. Comparative genomics of TICAM regions. Comparisons Downloaded from between human and zebrafish ge- nomes show regions in each con- taining TICAM1 (A)orTICAM2 (B) and the near-neighbors. Zebrafish data are from the Zv6 version of the zebrafish genome (www.ensembl. org/Danio_rerio/index.html). http://www.jimmunol.org/ by guest on September 30, 2021

TICAM1 and TICAM2 evolved in a gene duplication event The zebrafish TICAM1 gene on Chr:22 (ENSDART Mammalian TLRs use four different TIR domain-containing adap- 00000062685) is embedded in one of two zebrafish duplicates of tor proteins to transduce signals: MyD88, Mal, TICAM1, and the region of Hsa19p13 that surrounds TICAM1, with orthologous TICAM2 (TRAM) (28, 29). All but TICAM2 have been identified in near-neighbors in both genomes (Fig. 3A, bottom panel). The other fish, and accumulating evidence indicates that TICAM2 does not exist copy of the human region lies on zebrafish 2 in the in fishes (30). Phylogenetic analysis demonstrates that TICAM1 and Zv6 version of the zebrafish genome (Fig. 3A, top panel). Both TICAM2 form a monophyletic group with 100% bootstrap support zebrafish duplicate regions include genes slightly further away on (Fig. 2A). These findings are supported by Iliev et al. (30) and suggest Hsa19p13.3 (Fig. 3A, middle panel). The genomic neighborhood that TICAM1 and TICAM2 are sisters from a gene duplication event. of TICAM2 is better conserved in zebrafish than is the TICAM1- The TICAM gene pair is part of a paralogon, or a group of paralogous containing region, despite the lack of a TICAM2 ortholog (Fig. regions within the same species (31, 32), in the human genome (Fig. 3B). Genes that immediately neighbor TICAM2 in the human ge- 2B). The paralogon includes Hsa19p13 with FEM1A, TICAM1, nome have orthologs that are neighbors in the zebrafish genome TMED1, and SEMA6B; Hsa5q21-q23 with FEM1C, TICAM2, (FEM1C and TMED7), but without an ortholog of TICAM2 be- TMED7, and SEMA6A; Hsa15q21-q25 with FEM1B, TMED3, and tween them (Fig. 3B). In humans, a sequence appropriate to encode SEMA6D; and Hsa1q21-q22 with SEMA6C and other genes. These a human membrane-bound isoform of TICAM2 has been identified four paralogous chromosome segments probably originated in the two but not described (GenBank accession no. AY304581). It appar- rounds of genome duplication that likely preceded the vertebrate ra- ently occurs through an alternative splicing event that combines diation (33). the open reading frame of TMED7 with TICAM2, its 3Ј neighbor. 4522 TICAM SIGNALING DIFFERS IN BASALLY DIVERGING VERTEBRATES

FIGURE 4. A hypothesis for the history of TICAM genes. R1, First round of gene duplication; R2, second round of gene duplication. A, Predicted portion of a chromosome from the last pregenome duplication com- mon ancestor of vertebrates possessed genes in follow- ing order: FEM1, TICAM, TMED, SEMA6, and KCNN. B, Following the first round of genome duplication (Post R1), one of the TICAM genes was lost. C, After the second round of genome duplication (Post R2), the genes that became TICAM1 and TICAM2 appeared. In the preduplication teleost (D), the TICAM2 gene was lost, but the TICAM1 gene was retained. In humans (E), TICAM1 and TICAM2 were both retained and evolved distinct functions in TLR signaling. Downloaded from

To further test for the presence of a zebrafish ortholog of TICAM2, and TMED7 are adjacent in the frog would suggest that the gene http://www.jimmunol.org/ we performed 3Ј RACE to determine whether such a membrane- has been lost from the frog genome in an event independent from bound form of TICAM2 could be identified in zebrafish. A single the loss from the fish genomes. Given that rayfin fish genomes lack fragment was amplified, but analysis showed it to consist solely of TICAM2 (30), this gene may have been lost early after the diver- sequences orthologous to TMED7 (data not shown). gence of rayfin and lobefin fishes 450 million years ago. Comparative genomic analysis of available sequence data re- veals that in the opossum (Monodelphis domesticus), a marsupial diverging from the placental lineage ϳ175 million years ago (34), Zebrafish TICAM1 fails to interact with TRAF6 but can part of ENSMODT0000001740 on chromosome 6 is orthologous associate with TLR3, RIP1, and TBK1 to TICAM2 and part to TMED7; the opossum FEM1C ortholog Mammalian TICAM1 interacts with TLR3, TRAF6, TBK1, and by guest on September 30, 2021 (ENSMODT00000019962) is unlinked, appearing on chromosome RIP1 (15, 17, 40, 41) and these interactions potentiate downstream 3. In Xenopus, as in zebrafish, FEM1C (ENSXETT00000040604) antiviral signaling, leading to the activation of NF-␬B and IFN-␤. and TMED7 (ENSXETT00000040594) genes are adjacent to each Zebrafish TICAM1’s unique structural features, particularly the other, but TICAM2 does not appear between them. In chicken presence of a hydrophilic glutamine residue in the typically hy- ␪ (Gallus gallus), TMED7 (ENSGALT00000034557) is on chromo- drophobic 1 position, the absence of an obvious TRAF6-binding some Z, but FEM1C (ENSGALT00000013361) is unassigned motif, the truncation of its N- and C-terminal domains, and the making it difficult to draw conclusions. The nonvertebrate chordate sequence differences observed in the RHIM domain (Fig. 1), call Ciona intestinalis has an ortholog of TICAM genes (ENSCINT into question its capacity to interact with TLR3, TRAF6, TBK1, 00000022578), of TMED7/3/1 (ENSCINT00000003864), and of and RIP1 and transduce vital immune signals. To address zebrafish FEM1 genes (ENSCINT00000027118). These three genes are not TICAM1’s capacity to associate with these known interactors, co- adjacent in the Ciona genome, but the FEM1 ortholog has loci that immunoprecipitation experiments were performed (Fig. 5). are close neighbors with loci that are close neighbors of FEM1C Zebrafish TLR3-HA and zebrafish TICAM1–3ϫFLAG were and FEM1B in human. overexpressed in 293H cells for 48 h, and following a coimmu- These genomic data suggest the following history for TICAM1 noprecipitation experiment, were shown to associate with one an- and TICAM2 (Fig. 4). A chromosome in the last pregenome other (Fig. 5A, middle panel, lane 1). Empty vector controls were duplication common ancestor of vertebrates possessed FEM1, included to show that the interactions were not the result of non- TICAM, and TMED genes in that order, in addition to SEMA6, specific interactions (Fig. 5A, middle panel, lanes 2 and 3). KCNN, and other genes (Fig. 4A). After the first round of genome Through coimmunoprecipitation, it was shown that zebrafish duplication (R1), the TICAM gene on one of the duplicate chro- TICAM1 associated with zebrafish RIP1 (Fig. 5B, middle panel, mosomes was lost (Fig. 4B), and after the second round of genome lane 1) when overexpressed in 293H cells. Despite its truncated N duplication (R2), produced what became TICAM1 and TICAM2 terminus, zebrafish TICAM1 also is capable of interacting with (Fig. 4, C and E). In the teleost lineage, the TICAM2 gene was zebrafish TBK1 (Fig. 5C, lower middle panel, lane 1). Like mam- eventually lost, but TICAM1 was retained (Fig. 4D). Some of the mals, this interaction relies on the N terminus of TICAM1, as a other genes in the paralogon are present in duplicate copy in te- TICAM1 protein lacking the N terminus failed to associate with leosts today, due to a genome duplication event in the teleost an- TBK1 (Fig. 5C, lower middle panel, lane 3). Despite the similarity, cestry ϳ300 million years ago (35–39) (data not shown in the the functional significance of the TICAM1-TBK1 interaction in figure). According to this proposed history, the current absence of zebrafish, based upon the observations described later, is in ques- TICAM2 from the genome databases of chicken and Xenopus tion. In contrast, following overexpression in 293H cells, zebrafish could be due either to its absence from the animals’ genomes or to TICAM1 failed to coimmunoprecipitate zebrafish TRAF6 or a hole in the genome libraries, although the finding that FEM1C mouse TRAF6 (Fig. 5D, middle panel, lanes 1 and 2). Analysis of The Journal of Immunology 4523 Downloaded from http://www.jimmunol.org/

FIGURE 5. Zebrafish TICAM1 associates with TLR3, RIP1, and TBK1, but not TRAF6. A, Zebrafish TICAM1 (middle panel, lane 1) coimmunopre- cipitates with zebrafish TLR3 (top panel, lane 1). B, Zebrafish RIP1 (middle panel, lane 1) coimmunoprecipitates with zebrafish TICAM1 (top panel, lane 1). C, Zebrafish TBK1 (lower middle panel, lane 1) coimmunoprecipitates with zebrafish TICAM1 (top panel, lane 1). Zebrafish TICAM1 lacking the N-terminal residues (top middle panel, lane 3) fails to associate with zebrafish TBK1 (lower middle panel, lane 3), despite evidence of strong expression in whole cell lysates. D, Zebrafish TRAF6 (middle panel, lanes 1) fails to coimmunoprecipitate with zebrafish TICAM1 (top panel, lane 1), despite strong expression levels in the whole cell lysate (bottom panel, lane1). Similarly, mouse TRAF6 fails to coimmunoprecipitate with zebrafish TICAM1 (top panel, lane 2), despite strong expression levels in the whole cell lysate (bottom panel, lane 2). Mouse TICAM1 coimmunoprecipitates with both zebrafish TRAF6

(middle panel, lane 4) and mouse TRAF6 (middle panel, lane 5). by guest on September 30, 2021

zebrafish genome databases fails to uncover a sequence for a sec- man 293H and zebrafish ZFL cells to assay their effects on NF-␬B ond TRAF6 gene, and thus the potential for another TRAF6 co- activation, using the reporter construct pBIIx-luc (43) (Fig. 6, A evolving with TICAM1 in a way that would allow for their protein- and B). In 293H cells (Fig. 6A), full-length zebrafish TICAM1 protein interaction to be maintained appears extremely unlikely. retained a capacity to activate NF-␬B (9.3-fold induction), albeit at Zebrafish TRAF6 exhibits 58% identity and 73% conservation levels below that of full-length mouse TICAM1 (21.6-fold induc- with mouse TRAF6. Mouse TICAM1 associated both with ze- tion). Similarly in ZFL cells (Fig. 6B), full-length TICAM1 acti- brafish and with mouse TRAF6 (Fig. 5D, middle panel, lanes 4 and vated NF-␬B (3.3-fold induction), but in this instance, mouse 5). These data, coupled with the alignment data, imply that ze- TICAM1 exhibited diminished NF-␬B activation (1.2-fold induc- brafish TICAM1 does not interact with zebrafish TRAF6 and that tion). Overexpression of a zebrafish TICAM1 N-terminal deletion NF-␬B activation occurs through an alternative mechanism. Based construct enhanced NF-␬B activation over zebrafish full-length upon the coimmunoprecipitation data collected, the evidence in- TICAM1 in both 293H and ZFL cells, with 24.2-fold (Fig. 6A) and dicates that NF-␬B activation may occur through a TICAM1-RIP1 4.2-fold (Fig. 6B) activations noted, respectively. These data indi- association. cate that the TIR and C-terminal domains of zebrafish TICAM1 are responsible for NF-␬B activation and are corroborated by the ␬ Zebrafish TICAM1 activates NF- B in human 293H and diminished activities noted upon overexpression of deletion con- zebrafish ZFL cells structs missing both the N and C termini (ZFL: 1.4-fold induction; Alignment of zebrafish TICAM1 with mammalian orthologs re- 293H: 1.1-fold induction) or the C-terminal domain (ZFL: 2.0-fold veals clear sequence differences that may confound its capacity to induction; 293H: 2.1-fold induction) (Fig. 6, A and B). Deletion of transduce signals (Fig. 1A). Through overexpression of TICAM1 the TIR domain also diminished NF-␬B activation in ZFL cells and various deletion constructs, it has been shown that full-length (1.4-fold induction) (Fig. 6B), but appeared not to disrupt NF-␬B human TICAM1 can activate NF-␬B, IFN-␤ promoter, and ISRE signaling in 293H cells (11.2-fold induction), when compared with luciferase reporters and that individual TICAM1 domains (N ter- the full-length TICAM1 result (9.3-fold induction) (Fig. 6A). Fur- minus, C terminus, and/or TIR domains) contribute differentially thermore, site-directed mutagenesis of residues 547–550 (Met- to this signaling capacity (13, 40, 42). To examine zebrafish Ile-Gly-Asn) in the RHIM domain of zebrafish TICAM1, which TICAM1 and its domains’ signaling capacities, full-length ze- correspond to residues 687–690 in human TICAM1 and 688– brafish TICAM1, along with several zebrafish TICAM1 domain 691 in mouse TICAM1 (Val-Gln-Leu-Gly), to Ala-Ala-Ala- deletion constructs, a zebrafish TICAM1 RHIM mutant construct, Ala, as previously demonstrated by Meylan et al. (40), led to a and a mouse TICAM1-positive control, were overexpressed in hu- diminished capacity for zebrafish TICAM1 to activate NF-␬B 4524 TICAM SIGNALING DIFFERS IN BASALLY DIVERGING VERTEBRATES Downloaded from http://www.jimmunol.org/

FIGURE 6. Overexpression of zebrafish TICAM1 activates NF-␬B and type I IFN promoter, but in an ISRE-independent manner. Luciferase assays demonstrate the effect of mouse TICAM1, zebrafish TICAM1, and zebrafish TICAM1 mutants on NF-␬B activation in 293H (A) and ZFL (B) cells. A RHIM domain mutation causes a decrease in NF-␬B-luciferase activation relative to the control zebrafish TICAM1 in 293H cells (A). C, Unlike mouse TICAM1, zebrafish TICAM1, and deletion mutants demonstrate limited ISRE-luciferase activation in 293H cells. D and E, In ZFL cells, minimal ISRE-luc and by guest on September 30, 2021 Mx187-luc activation is observed upon overexpression of mouse TICAM1, zebrafish TICAM1, and zebrafish TICAM1 deletion constructs. F, Despite the absence of strong ISRE-luciferase activity, type I IFN promoter-luciferase activation is observed upon overexpression of mouse TICAM1, zebrafish TICAM1, and zebrafish TICAM1 deletion constructs. All firefly luciferase activities are normalized to constitutively active Renilla luciferase levels. Data are presented as fold induction over empty vector. Error bars represent one SEM. Empty, empty vector control; Mouse, full-length mouse TICAM1, Danio, full-length zebrafish TICAM1; ⌬N, N-terminal deletion (1–311 aa) of zebrafish TICAM1; ⌬C, C-terminal deletion (485–566 aa) of zebrafish TICAM1; ⌬TIR, deletion of TIR domain (312–482 aa) of zebrafish TICAM1; ⌬N⌬C, deletion of N- and C-termini of zebrafish TICAM1 leaving TIR domain (312–482 aa); RHIM Mutant, site-directed mutagenesis of amino acids 547–550 from Met-Ile-Gly-Asn to Ala-Ala-Ala-Ala.

(2.4-fold activation over empty vector control and 3.9-fold re- cells (136-fold above empty vector) (Fig. 6C); however, when duction compared with wild-type full-length TICAM1) (Fig. compared with the 293H data, it exhibits minimal activation of 6A). Collectively, these data demonstrate an essential role for ISRE in ZFL cells (ISRE-luc: 1.2-fold above empty vector; the C-terminal RHIM domain in zebrafish TICAM1-mediated Mx187-luc: 4.9-fold above empty vector) (Fig. 6, D and E). De- NF-␬B activation. spite the unexpected absence of robust ISRE activation in ZFL cells, as seen in 293H cells, zebrafish type I IFN was activated by Zebrafish TICAM1 activation of type I IFN occurs through zebrafish TICAM1 (9.5-fold induction) (Fig. 6F). Mouse TICAM1 N-terminal-independent and ISRE-independent mechanisms strongly activated the zebrafish type I IFN promoter (22.9-fold Mammalian TICAM1 has been shown to play an essential role in induction) (Fig. 6F). Overexpression of the zebrafish TICAM1 N- TLR3- and TLR4-mediated IFN-␤ activation through an IRF3/ terminal deletion construct led to an even more enhanced type I IRF7-dependent mechanism (12, 16, 44, 45). TBK1 is brought into IFN activation (16.9-fold induction) than with full-length zebrafish a complex with TICAM1 at its N terminus via TRAF3 and phos- TICAM1 (Fig. 6F). These results correlated with the observed phorylates IRF3 and IRF7, allowing for its eventual nuclear NF-␬B activation, in which the N-terminal deletion construct ex- translocation and up-regulation of IFN-␤ (15, 17, 19–21). hibited stronger activation (Fig. 6, A and B). The C-terminal de- Overexpression of zebrafish TICAM1 led to minimal activation letion retained a capacity to activate IFN promoter (Fig. 6F) (10.8- of an ISRE-luciferase reporter in 293H cells (2.6-fold above fold induction), but it appeared not to be strongly ISRE-induced in empty vector) (Fig. 6C) and no activation in ZFL cells (Fig. 293H cells (2.4-fold induction) or ZFL cells (ISRE-luc: 1.7-fold 6D). Similarly, using the zebrafish-derived Mx187-luc construct, induction; Mx187-luc: 3.4-fold induction) (Fig. 6, D and E). In- which is a minimal IFN-inducible Mx promoter construct possess- terestingly, site-directed mutagenesis of the C-terminal RHIM do- ing two ISREs, a 3.4-fold induction over empty vector control was main at residues 547–550 from Met-Ile-Gly-Asn to Ala-Ala-Ala- observed in ZFL cells (Fig. 6E). In contrast, mouse TICAM1 over- Ala disrupted IFN promoter activation (4.0-fold induction over expression robustly activated ISRE-luciferase reporter in 293H empty vector control), leading to a 2.4-fold reduction from the The Journal of Immunology 4525

wild-type TICAM1 and 2.7-fold reduction from the C-terminal likely in describing what has happened to TICAM2 at each verte- deletion TICAM1 construct (Fig. 6F). The apparent difference in brate class, if they indeed have been lost. Thus, it is a strong pos- IFN promoter activation capacity between the C-terminal deletion sibility that fishes, and perhaps amphibians and birds, rely solely TICAM1 construct and the RHIM mutant indicates not only an upon TICAM1, MyD88, and Mal to transduce signals from TLRs. activating role for the RHIM domain in IFN activity, but also a It is possible that additional TIR domain adaptor proteins may possible inhibitory role for other portions of the C terminus. Con- exist, but this appears unlikely based upon the plethora of genome structs lacking the TIR domain retained a capacity to activate the data available for representatives of each vertebrate class. The sig- IFN promoter construct (6.7-fold induction) (Fig. 6F), but again nificance of TICAM2’s absence in basally diverging vertebrates exhibited minimal ISRE activation in 293H (1.9-fold induction) clearly needs to be explored, as it has been shown to be essential (Fig. 6C) or ZFL (ISRE-luc: 1.5-fold induction; Mx197-luc: 2.0- to TLR4-mediated IFN-␤ induction in mammals. It is also clear fold induction) cells (Fig. 6, D and E). Overexpression of the com- that the role of MyD88, Mal, and TICAM1 in lower vertebrate bined N-terminal, C-terminal deletion, which consists largely of TLR signaling also needs to be fully characterized. The data pre- the TIR domain alone, resemble the data presented by Yamamoto sented here attempt to describe TICAM1’s function in zebrafish et al. (13). The IFN promoter construct was activated 2.4-fold in TLR signaling, beginning with an experiment designed to test ZFL cells (Fig. 6F); the ISRE-luc reporter was induced 1.5-fold in whether zebrafish TICAM1 can interact with zebrafish TLR3, as it 293H cells (Fig. 6C) and 1.3-fold in ZFL cells (Fig. 6D); the does in mammals. Mx187-luc reporter was activated 1.1-fold above empty vector TLR3 recognizes dsRNA (3) and endogenous cellular RNA (47) control (Fig. 6E). These data show that the TIR domain of ze- and unleashes a potent MyD88-independent antiviral pathway (12,

brafish TICAM1, when expressed alone, has a dominant-negative 13). Human TLR3 is a 904-aa long protein expressed on the cell Downloaded from effect. surface of human fibroblast epithelial cells (48) and within den- dritic cells (49, 50). It possesses a horseshoe-shaped ectodomain Discussion consisting of 23 leucine-rich repeats (51, 52), a transmembrane Phylogenetic analyses imply that fish type I IFNs form clades dis- domain, and a TIR domain required for downstream signaling (3). tinct from those of other mammalian and avian type I IFNs and TLR3 orthologs have been cloned from a broad range of verte-

thus cannot be classified as being ␣-or␤-like, except by function brates, from zebrafish to humans (53). The actual role of TLR3 in http://www.jimmunol.org/ (1, 2). These findings, coupled with the unique gene structure of mediated antiviral responses is under scrutiny (54); however, spe- fish type I IFNs, raise interesting questions about how they are cific evidence does indicate a real immunological function. Spe- triggered to signal and whether TLR-mediated pathways are re- cifically, TLR3 is believed to interact with dsRNA intermediates of sponsible for their induction. In describing the first fish type I IFN, single-stranded RNA viruses like respiratory syncytial virus (55) which we cloned from the zebrafish, our laboratory noted its an- or West Nile virus (56). Interactions were also noted between tiviral potential, as it was inducible, in ZFL cells, by the TLR3 TLR3 and mouse CMV, a DNA virus (57), and TLR3 and reovi- agonist polyinosinic-polycytidylic acid (poly(I:C)) and was pro- rus, a dsRNA virus (3, 54). Interestingly, TLR3 was shown to be tective of cells infected with snakehead rhabdovirus (1). Subse- activated in response to nonviral dsRNAs as well. dsRNAs from quently, we noted that zebrafish type I IFN exerts its antiviral eggs of the helminth parasite Schistosoma mansoni were shown to by guest on September 30, 2021 effects through the induction of antiviral proteins like Mx (22) and trigger an immune response through the TLR3 signal transduction can itself be induced in vivo by snakehead rhabdovirus (46). cascade (58). TLR3 activation by ligands appears to occur in the The findings described herein illustrate a TICAM1-dependent acidic compartments of early phagolysosomes or endosomes, as an mechanism by which fish type I IFN expression can be induced. In acid pH is required (59). Disruption of acidic pH blocked TLR3 mammals, TLR3 and TLR4 signaling through TICAM1 leads to activation by poly(I:C). Additionally, TLR3 must form multimers the specific up-regulation of IFN-␤. Direct IFN-␣ up-regulation to be active (59). Upon activation, the antiviral signal cascades through TICAM1-dependent pathways has yet to be described and through the adaptor molecule TICAM1 (12, 13). The TLR3- indeed may not occur. Furthermore, based upon the current liter- TICAM1 interaction requires their respective BB loops. The con- ature, direct IFN-␣ induction appears restricted to MyD88-depen- sensus sequence for the BB loop, a conserved sequence contained dent signaling pathways (6). These findings are interesting in light within the TIR domain and shown to be important for TLR and ␾ ␾ of the fact that zebrafish type I IFN, as already discussed, cannot TICAM1 signaling, is RDx 1 2G, where x represents any residue ␣ ␤ ␾ ␾ be defined as -or -like in terms of phylogeny but may be con- and represents any hydrophobic residue. The mutation of the 2 sidered ␤-like in terms of its induction by TICAM1 (Fig. 6F). residue from proline to histidine results in the inability of the hu- Comprehensive analyses of syntenies and phylogenies from rep- man TICAM1 to interact with human TLR3 and thereby up-reg- ␤ ␾ resentative organisms of each vertebrate class provide excellent ulate human IFN- (12). Similarly, when the 2 position of human insight into how TLR signaling has evolved (Figs. 2 and 3). For TLR3 was mutated from alanine to histidine, TLR3 failed to in- example, our data confirm a recent hypothesis proposed by Iliev teract with human TICAM1 and therefore failed to up-regulate et al. (30) about TICAM2’s absence in fish and goes further by IFN-␤ (12). The BB loop of the putative zebrafish TICAM1 differs indicating that TICAM2 most likely arose from an early gene du- from the consensus sequence and from the mammalian TICAM1s ␾ plication (Fig. 4) and then was lost after the divergence of rayfin (Fig. 1A). The proline at 2 is conserved in known mammalian and lobefin fishes 450 million years ago. Our inability to identify TICAM1s. In the putative zebrafish TICAM1, this position con- ␾ TICAM2 in the chicken and Xenopus leaves open the possibility tains an alanine instead. In addition, in zebrafish, 1 is occupied by that TICAM2 was lost from each of these genomes in events in- a glutamine, and the x position contains an alanine. We believed dependent of TICAM2’s loss in the fish genomes. Indeed, in Xe- that the differences noted in these positions may be important to nopus, the maintained arrangement of the flanking genes FEM1C TLR3-TICAM1 signaling in zebrafish; however, our coimmuno- and TMED7 and the absence of TICAM2 makes this a likely sce- precipitation data indicate that TLR3 and TICAM1 can associate nario. Unfortunately, similar evidence could not be derived from and thus this interaction appears conserved despite sequence dif- the chicken genome due to sequence gaps. Although our hypoth- ferences (Fig. 5A). It is noteworthy that human and mouse TLR3 ␾ esis is not the most parsimonious, it is clear, based upon the history contain alanines in the 2 positions, while human and mouse ␾ of the TICAM genes (Fig. 4), that this explanation is the most TICAM1 contain prolines in their 2 positions. These occurrences 4526 TICAM SIGNALING DIFFERS IN BASALLY DIVERGING VERTEBRATES

␾ are reversed in zebrafish, where TLR3 possesses a proline in the 2 deletion of the C terminus of TICAM1 did not alter IFN pro- ␾ position and putative TICAM1 possesses an alanine in its 2 po- moter activation in ZFL cells, mutation of the C-terminal RHIM sition. This “complementary switch” may be noteworthy from an domain at residues 547–550 from Met-Ile-Gly-Asn to Ala-Ala- evolutionary perspective, but may also have practical impacts on Ala-Ala resulted in a diminished capacity for IFN activation TLR3-TICAM1 interactions and signaling. (Fig. 6F). These findings show a clear role for the RHIM do- In mammals, TBK1 interacts with the N terminus of TICAM1 main in IFN activation but also indicate that other portions of and then phosphorylates IRF-3 and IRF-7, leading to up-regulation the C terminus may negatively regulate IFN activation. Further of IFN-␤ (15–18). Although zebrafish TICAM1 retains the capac- investigation of the zebrafish pathway may provide the context ity to coimmunoprecipitate TBK1 through its N terminus (Fig. necessary to resolve some of the discrepancies noted in char- 5C), indicating this interaction can occur, it is unclear what role the acterizing the role and importance of TICAM1’s domains in TBK1 interaction plays in the up-regulation of zebrafish type I tetrapods. IFN. In fact, upon deletion of the N terminus responsible for the TICAM1-TBK1 interaction in mammals, enhanced IFN activation Acknowledgments was noted (Fig. 6F), indicating that the N terminus, and thus po- We thank Brian Niland, Akshata Nayak, and Jeremy Charette for their tentially the TBK1 interaction, has one or both of the following technical assistance. This is Maine Agricultural and Forest Experiment effects: it may facilitate an inhibitory portion of the zebrafish Station publication 2923. TICAM1 pathway, or it may sterically hinder TICAM1’s interac- tion with other protein partners that are yet to be described. Disclosures It was noteworthy that zebrafish TICAM1, lacking an apparent Downloaded from The authors have no financial conflict of interest. TRAF6-binding motif with the consensus PxExx[Ac/Ar] (P repre- sents proline, x represents any amino acid, and [Ac/Ar] represents an acidic or aromatic amino acid), failed to coimmunoprecipitate References 1. Altmann, S. M., M. T. Mellon, D. L. Distel, and C. H. Kim. 2003. Molecular and zebrafish TRAF6 (Fig. 5D). This finding confounds TRAF6’s role functional analysis of an interferon gene from the zebrafish, Danio rerio. J. Virol. in mediating a NF-␬B response, at least in the zebrafish, and sup- 77: 1992–2002. 2. Robertsen, B. 2006. The interferon system of teleost fish. Fish Shellfish Immunol. ports the assertion by Gohda et al. (60) that the TICAM1-TRAF6 http://www.jimmunol.org/ 20: 172–191. interaction may be nonessential in the mammalian TLR3 pathway. 3. Alexopoulou, L., A. C. Holt, R. Medzhitov, and R. A. Flavell. 2001. Recognition Our finding is in contrast to reports by Sato et al. (15) and Jiang of double-stranded RNA and activation of NF-␬B by Toll-like receptor 3. Nature et al. (41), who showed that TICAM1-TRAF6 interaction is im- 413: 732–738. ␬ 4. Diebold, S. S., T. Kaisho, H. Hemmi, S. Akira, and C. Reis e Sousa. 2004. Innate portant for NF- B activation in human 293 cells. Furthermore, antiviral responses by means of TLR7-mediated recognition of single-stranded overexpression of an N-terminal deletion zebrafish TICAM1 in RNA. Science 303: 1529–1531. 293H and ZFL cells, causing enhanced NF-␬B activation relative 5. Heil, F., H. Hemmi, H. Hochrein, F. Ampenberger, C. Kirschning, S. Akira, G. Lipford, H. Wagner, and S. Bauer. 2004. Species-specific recognition of sin- to full-length zebrafish, was noted (Fig. 6, A and B). These data gle-stranded RNA via Toll-like receptor 7 and 8. Science 303: 1526–1529. lend additional credence to an aforementioned potential inhibitory 6. Kawai, T., S. Sato, K. J. Ishii, C. Coban, H. Hemmi, M. Yamamoto, K. Terai, role the N terminus plays in zebrafish TICAM1. Indeed, our data M. Matsuda, J. Inoue, S. Uematsu, et al. 2004. Interferon-␣ induction through by guest on September 30, 2021 Toll-like receptors involves a direct interaction of IRF7 with MyD88 and TRAF6. indicate that TICAM1 signaling is driven by the TIR domain and Nat. Immunol. 5: 1061–1068. C terminus, with the RHIM domain playing an important role. Our 7. Lund, J. M., L. Alexopoulou, A. Sato, M. Karow, N. C. Adams, N. W. Gale, finding that zebrafish TICAM1 and RIP1 can interact with one A. Iwasaki, and R. A. Flavell. 2004. Recognition of single-stranded RNA viruses by Toll-like receptor 7. Proc. Natl. Acad. Sci. USA 101: 5598–5603. another (Fig. 5B), and that mutation of the TICAM1 RHIM do- 8. Hoshino, K., T. Kaisho, T. Iwabe, O. Takeuchi, and S. Akira. 2002. Differential main at residues 547–550 from Met-Ile-Gly-Asn to Ala-Ala-Ala- involvement of IFN-␤ in Toll-like receptor-stimulated dendritic cell activation. ␬ A Int. Immunol. 14: 1225–1231. Ala disrupts NF- B activation in 293H cells (Fig. 5 ), suggests 9. Hemmi, H., O. Takeuchi, T. Kawai, T. Kaisho, S. Sato, H. Sanjo, M. Matsumoto, that this activation occurs because of this association, as is the case K. Hoshino, H. Wagner, K. Takeda, and S. Akira. 2000. A Toll-like receptor in mammals (40, 42). recognizes bacterial DNA. Nature 408: 740–745. 10. Kopp, E., and R. Medzhitov. 2003. Recognition of microbial infection by Toll- The evolutionary history of TLR-mediated induction of type I like receptors. Curr. Opin. Immunol. 15: 396–401. IFN responses provides essential perspective into the origins and 11. Akira, S., and K. Takeda. 2004. Toll-like receptor signalling. Nat. Rev. Immunol. complexities of antiviral immunity. The data presented herein rep- 4: 499–511. 12. Oshiumi, H., M. Matsumoto, K. Funami, T. Akazawa, and T. Seya. 2003. TI- resent the first insights into the underlying mechanisms basally CAM-1, an adaptor molecule that participates in Toll-like receptor 3-mediated diverging vertebrates use to counter pathogens through TICAM1- interferon-␤ induction. Nat. Immunol. 4: 161–167. dependent signaling. Our findings also demonstrate that TICAM1- 13. Yamamoto, M., S. Sato, K. Mori, K. Hoshino, O. Takeuchi, K. Takeda, and S. Akira. 2002. Cutting edge: a novel Toll/IL-1 receptor domain-containing dependent induction of type I IFNs is an ancient mechanism, al- adapter that preferentially activates the IFN-␤ promoter in the Toll-like receptor though the means by which TICAM1 induces this particular signaling. J. Immunol. 169: 6668–6672. zebrafish type I IFN is through an apparently alternative signaling 14. Doyle, S., S. Vaidya, R. O’Connell, H. Dadgostar, P. Dempsey, T. Wu, G. Rao, R. Sun, M. Haberland, R. Modlin, and G. Cheng. 2002. IRF3 mediates a TLR3/ pathway. Although the complete details of how TICAM1 induces TLR4-specific antiviral gene program. Immunity 17: 251–263. this zebrafish type I IFN are yet to be determined, our findings may 15. Sato, S., M. Sugiyama, M. Yamamoto, Y. Watanabe, T. Kawai, K. Takeda, and be predictive in defining undiscovered TICAM1-dependent, IRF3/ S. Akira. 2003. Toll/IL-1 receptor domain-containing adaptor inducing IFN-␤ (TRIF) associates with TNF receptor-associated factor 6 and TANK-binding ki- 7-independent type I IFN induction schemes in mammals, perhaps nase 1, and activates two distinct transcription factors, NF-␬B and IFN-regulatory through a RHIM-dependent mechanism. In mammals, the TICAM1- factor-3, in the Toll-like receptor signaling. J. Immunol. 171: 4304–4310. TBK1 interaction is essential to type I IFN activation (15). Our data 16. Fitzgerald, K. A., D. C. Rowe, B. J. Barnes, D. R. Caffrey, A. Visintin, E. Latz, B. Monks, P. M. Pitha, and D. T. Golenbock. 2003. LPS-TLR4 signaling to show that while this interaction also occurs in zebrafish (Fig. 5C), its IRF-3/7 and NF-␬B involves the toll adapters TRAM and TRIF. J. Exp. Med. importance in mounting the TLR-mediated antiviral response is 198: 1043–1055. 17. Han, K. J., X. Su, L. G. Xu, L. H. Bin, J. Zhang, and H. B. Shu. 2004. Mech- unclear. In particular, our surprising finding that a TICAM1 anisms of the TRIF-induced interferon-stimulated response element and NF-␬B construct lacking the N-terminal domain necessary for interac- activation and apoptosis pathways. J. Biol. Chem. 279: 15652–15661. tion with TBK1 exhibited an enhanced activation of the ze- 18. McWhirter, S. M., K. A. Fitzgerald, J. Rosains, D. C. Rowe, D. T. Golenbock, and T. Maniatis. 2004. IFN-regulatory factor 3-dependent is brafish type I IFN promoter (Fig. 6F) indicated that other parts defective in Tbk1-deficient mouse embryonic fibroblasts. Proc. Natl. Acad. Sci. of the TICAM1 protein play a role in this activation. Although USA 101: 233–238. The Journal of Immunology 4527

19. Sasai, M., H. Oshiumi, M. Matsumoto, N. Inoue, F. Fujita, M. Nakanishi, and 42. Kaiser, W. J., and M. K. Offermann. 2005. Apoptosis induced by the Toll-like T. Seya. 2005. Cutting edge: NF-␬B-activating kinase-associated protein 1 par- receptor adaptor TRIF is dependent on its receptor interacting protein homotypic ticipates in TLR3/Toll-IL-1 homology domain-containing adapter molecule-1- interaction motif. J. Immunol. 174: 4942–4952. mediated IFN regulatory factor 3 activation. J. Immunol. 174: 27–30. 43. Saksela, K., and D. Baltimore. 1993. Negative regulation of immunoglobulin ␬ 20. Oganesyan, G., S. K. Saha, B. Guo, J. Q. He, A. Shahangian, B. Zarnegar, light-chain gene transcription by a short sequence homologous to the murine B1 A. Perry, and G. Cheng. 2006. Critical role of TRAF3 in the Toll-like receptor- repetitive element. Mol. Cell Biol. 13: 3698–3705. dependent and -independent antiviral response. Nature 439: 208–211. 44. Oshiumi, H., M. Sasai, K. Shida, T. Fujita, M. Matsumoto, and T. Seya. 2003. 21. Hacker, H., V. Redecke, B. Blagoev, I. Kratchmarova, L. C. Hsu, G. G. Wang, TIR-containing adapter molecule (TICAM)-2, a bridging adapter recruiting to M. P. Kamps, E. Raz, H. Wagner, G. Hacker, et al. 2006. Specificity in Toll-like Toll-like receptor 4 TICAM-1 that induces interferon-␤. J. Biol. Chem. 278: receptor signalling through distinct effector functions of TRAF3 and TRAF6. 49751–49762. Nature 439: 204–207. 45. Doyle, S. E., R. O’Connell, S. A. Vaidya, E. K. Chow, K. Yee, and G. Cheng. 22. Altmann, S. M., M. T. Mellon, M. C. Johnson, B. H. Paw, N. S. Trede, L. I. Zon, 2003. Toll-like receptor 3 mediates a more potent antiviral response than Toll- and C. H. Kim. 2004. Cloning and characterization of an Mx gene and its cor- like receptor 4. J. Immunol. 170: 3565–3571. responding promoter from the zebrafish, Danio rerio. Dev. Comp. Immunol. 28: 46. Phelan, P. E., M. E. Pressley, P. E. Witten, M. T. Mellon, S. Blake, and 295–306. C. H. Kim. 2005. Characterization of snakehead rhabdovirus infection in ze- 23. Gibbs, P. D., and M. C. Schmale. 2000. GFP as a genetic marker scorable brafish (Danio rerio). J. Virol. 79: 1842–1852. throughout the life cycle of transgenic zebra fish. Mar. Biotechnol. 2: 107–125. 47. Kariko, K., H. Ni, J. Capodici, M. Lamphier, and D. Weissman. 2004. mRNA is 24. Morgenstern, B. 2004. DIALIGN: multiple DNA and protein sequence alignment an endogenous ligand for Toll-like receptor 3. J. Biol. Chem. 279: 12542–12550. at BiBiServ. Nucleic Acids Res. 32: W33–W36. 48. Matsumoto, M., S. Kikkawa, M. Kohase, K. Miyake, and T. Seya. 2002. Estab- 25. Felsenstein, J. 1989. Phylip: phylogeny inference package (version 3.2). Cladis- lishment of a monoclonal antibody against human Toll-like receptor 3 that blocks tics 5: 164–166. double-stranded RNA-mediated signaling. Biochem. Biophys. Res. Commun. 26. Jones, D. T., W. R. Taylor, and J. M. Thornton. 1992. The rapid generation of 293: 1364–1369. mutation data matrices from protein sequences. Comput. Appl. Biosci. 8: 49. Matsumoto, M., K. Funami, M. Tanabe, H. Oshiumi, M. Shingai, Y. Seto, 275–282. A. Yamamoto, and T. Seya. 2003. Subcellular localization of Toll-like receptor 27. Saitou, N., and M. Nei. 1987. The neighbor-joining method: a new method for 3 in human dendritic cells. J. Immunol. 171: 3154–3162. reconstructing phylogenetic trees. Mol. Biol. Evol. 4: 406–425. 50. Muzio, M., D. Bosisio, N. Polentarutti, G. D’Amico, A. Stoppacciaro, Downloaded from 28. Meijer, A. H., S. F. Gabby Krens, I. A. Medina Rodriguez, S. He, W. Bitter, R. Mancinelli, C. van’t Veer, G. Penton-Rol, L. P. Ruco, P. Allavena, and B. Ewa Snaar-Jagalska, and H. P. Spaink. 2004. Expression analysis of the Toll- A. Mantovani. 2000. Differential expression and regulation of Toll-like receptors like receptor and TIR domain adaptor families of zebrafish. Mol. Immunol. 40: (TLR) in human leukocytes: selective expression of TLR3 in dendritic cells. 773–783. J. Immunol. 164: 5998–6004. 29. Jault, C., L. Pichon, and J. Chluba. 2004. Toll-like receptor gene family and 51. Choe, J., M. S. Kelker, and I. A. Wilson. 2005. Crystal structure of human TIR-domain adapters in Danio rerio. Mol. Immunol. 40: 759–771. Toll-like receptor 3 (TLR3) ectodomain. Science 309: 581–585. 30. Iliev, D. B., J. C. Roach, S. Mackenzie, J. V. Planas, and F. W. Goetz. 2005. 52. Bell, J. K., I. Botos, P. R. Hall, J. Askins, J. Shiloach, D. M. Segal, and

Endotoxin recognition: in fish or not in fish? FEBS Lett. 579: 6519–6528. D. R. Davies. 2005. The molecular structure of the Toll-like receptor 3 ligand- http://www.jimmunol.org/ 31. Coulier, F., C. Popovici, R. Villet, and D. Birnbaum. 2000. MetaHox gene clus- binding domain. Proc. Natl. Acad. Sci. USA 102: 10976–10980. ters. J. Exp. Zool. 288: 345–351. 53. Phelan, P. E., M. T. Mellon, and C. H. Kim. 2005. Functional characterization of 32. Leveugle, M., K. Prat, N. Perrier, D. Birnbaum, and F. Coulier. 2003. ParaDB: full-length TLR3, IRAK-4, and TRAF6 in zebrafish (Danio rerio). Mol. Immu- a tool for paralogy mapping in vertebrate genomes. Nucleic Acids Res. 31: 63–67. nol. 42: 1057–1071. 33. Dehal, P., and J. L. Boore. 2005. Two rounds of whole genome duplication in the 54. Edelmann, K. H., S. Richardson-Burns, L. Alexopoulou, K. L. Tyler, ancestral vertebrate. PLoS Biol. 3: e314. R. A. Flavell, and M. B. Oldstone. 2004. Does Toll-like receptor 3 play a bio- 34. Hedges, S. B., and S. Kumar. 2002. Genomics: vertebrate genomes compared. logical role in virus infections? Virology 322: 231–238. Science 297: 1283–1285. 55. Rudd, B. D., E. Burstein, C. S. Duckett, X. Li, and N. W. Lukacs. 2005. Differ- 35. Amores, A., A. Force, Y. L. Yan, L. Joly, C. Amemiya, A. Fritz, R. K. Ho, ential role for TLR3 in respiratory syncytial virus-induced chemokine expression. J. Langeland, V. Prince, Y. L. Wang, et al. 1998. Zebrafish hox clusters and J. Virol. 79: 3350–3357. vertebrate genome evolution. Science 282: 1711–1714. 56. Wang, T., T. Town, L. Alexopoulou, J. F. Anderson, E. Fikrig, and R. A. Flavell.

36. Amores, A., T. Suzuki, Y. L. Yan, J. Pomeroy, A. Singer, C. Amemiya, and 2004. Toll-like receptor 3 mediates West Nile virus entry into the brain causing by guest on September 30, 2021 J. H. Postlethwait. 2004. Developmental roles of pufferfish Hox clusters and lethal encephalitis. Nat. Med. 10: 1366–1373. genome evolution in ray-fin fish. Genome Res. 14: 1–10. 57. Tabeta, K., P. Georgel, E. Janssen, X. Du, K. Hoebe, K. Crozat, S. Mudd, 37. Naruse, K., M. Tanaka, K. Mita, A. Shima, J. Postlethwait, and H. Mitani. 2004. L. Shamel, S. Sovath, J. Goode, et al. 2004. Toll-like receptors 9 and 3 as es- A medaka gene map: the trace of ancestral vertebrate proto-chromosomes re- sential components of innate immune defense against mouse cytomegalovirus vealed by comparative gene mapping. Genome Res. 14: 820–828. infection. Proc. Natl. Acad. Sci. USA 101: 3516–3521. 38. Postlethwait, J. H., Y. L. Yan, M. A. Gates, S. Horne, A. Amores, A. Brownlie, 58. Aksoy, E., C. S. Zouain, F. Vanhoutte, J. Fontaine, N. Pavelka, N. Thieblemont, A. Donovan, E. S. Egan, A. Force, Z. Gong, et al. 1998. Vertebrate genome F. Willems, P. Ricciardi-Castagnoli, M. Goldman, M. Capron, et al. 2005. Dou- evolution and the zebrafish gene map. Nat. Genet. 18: 345–349. ble-stranded RNAs from the helminth parasite Schistosoma activate TLR3 in 39. Taylor, J. S., I. Braasch, T. Frickey, A. Meyer, and Y. Van de Peer. 2003. Ge- dendritic cells. J. Biol. Chem. 280: 277–283. nome duplication, a trait shared by 22000 species of ray-finned fish. Genome Res. 59. de Bouteiller, O., E. Merck, U. A. Hasan, S. Hubac, B. Benguigui, G. Trinchieri, 13: 382–390. E. E. Bates, and C. Caux. 2005. Recognition of double-stranded RNA by human 40. Meylan, E., K. Burns, K. Hofmann, V. Blancheteau, F. Martinon, M. Kelliher, Toll-like receptor 3 and downstream receptor signaling requires multimerization and J. Tschopp. 2004. RIP1 is an essential mediator of Toll-like receptor 3-in- and an acidic pH. J. Biol. Chem. 280: 38133–38145. duced NF-␬B activation. Nat. Immunol. 5: 503–507. 60. Gohda, J., T. Matsumura, and J. Inoue. 2004. Cutting edge: TNFR-associated 41. Jiang, Z., T. W. Mak, G. Sen, and X. Li. 2004. Toll-like receptor 3-mediated factor (TRAF) 6 is essential for MyD88-dependent pathway but not toll/IL-1 activation of NF-␬B and IRF3 diverges at Toll-IL-1 receptor domain-containing receptor domain-containing adaptor-inducing IFN-␤ (TRIF)-dependent pathway adapter inducing IFN-␤. Proc. Natl. Acad. Sci. USA 101: 3533–3538. in TLR signaling. J. Immunol. 173: 2913–2917.