Discovery and Analysis of Invertebrate Igvj-C2 Structure from Amphioxus Provides Insight Into the Evolution of the Ig Superfamily

Discovery and Analysis of Invertebrate IgVJ -C2 Structure from Amphioxus Provides Insight into the Evolution of the Ig Superfamily This information is current as of October 1, 2021. Rong Chen, Lijie Zhang, Jianxun Qi, Nianzhi Zhang, Ling Zhang, Shugang Yao, Yanan Wu, Bo Jiang, Zhenbao Wang, Hongyu Yuan, Qiujin Zhang and Chun Xia J Immunol published online 7 March 2018

http://www.jimmunol.org/content/early/2018/03/07/jimmun Downloaded from ol.1700906

Supplementary http://www.jimmunol.org/content/suppl/2018/03/07/jimmunol.170090 Material 6.DCSupplemental 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

by guest on October 1, 2021 • Fast Publication! 4 weeks from acceptance to publication

*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 © 2018 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Published March 7, 2018, doi:10.4049/jimmunol.1700906 The Journal of Immunology

Discovery and Analysis of Invertebrate IgVJ-C2 Structure from Amphioxus Provides Insight into the Evolution of the Ig Superfamily

Rong Chen,*,† Lijie Zhang,* Jianxun Qi,‡ Nianzhi Zhang,* Ling Zhang,* Shugang Yao,* Yanan Wu,* Bo Jiang,* Zhenbao Wang,* Hongyu Yuan,* Qiujin Zhang,x and Chun Xia*,{

The emergence of adaptive immunity in jawed vertebrates depended on the appearance of variable immune receptors, BCRs and

TCRs, which exhibit variable-J–constant (VJ-C)–type Ig superfamily folds. Hitherto, however, the structures of IgV-J-IgC–type molecules had never been characterized in invertebrates, leaving the origin of BCR/TCR-type molecules unknown. Using x-ray crystallography, the structure of a V -C2 molecule, named AmpIgV -C2, was determined in amphioxus (Branchiostoma ﬂoridae).

J J Downloaded from The ﬁrst domain shows typical V folding, including the hydrophobic core, CDR analogs, and eight conserved residues. The second domain is a C2-type Ig superfamily domain, as deﬁned by its short length and the absence of b-strand D- and C1-typical motifs.

AmpIgVJ-C2 molecules form homodimers, using “three-layer packing dimerization,” as described for TCRs and BCRs. The AmpIgVJ-C2 V domain harbors a diglycine motif in b-strand G and forms a b-bulge structure participating in V–V intermolecular interaction. By immunohistochemistry, AmpIgVJ-C2 molecules were primarily found in mucosal tissues, whereas PCR and sequence analysis indicated considerable genetic variation at the single-gene level; these ﬁndings would be consistent with an http://www.jimmunol.org/ immune function and a basic ability to adapt to binding different immune targets. Our results show a BCR/TCR-ancestral like molecule in amphioxus and help us to understand the evolution of the adaptive immune system. The Journal of Immunology, 2018, 200: 000–000.

he Ig superfamily (IgSF) is one of the largest, and it function of a protein is closely related to its structure. Therefore, contains diverse groups of proteins that are widely dis- examining the structures of the Ig domains provides clues re- T tributed in human and other vertebrate genomes. Ig do- garding the origin and evolutionary pathway of this family. mains are primarily grouped into four sets: variable (V) set, The adaptive immune system (AIS) similar to that in mammals

constant-1 (C1) set, constant-2 (C2) set, and intermediate (I) set (the definition of AIS used in this article), which is characterized by by guest on October 1, 2021 (1). The typical Ig domain, which is ∼100 aa in length, was first diverse molecules containing one or more Ig domains, including identified in Abs through sequence comparison (2). Structural BCRs/Abs, TCRs, and MHC classes I and II, arose in ancient fish studies of Abs have shown that Ig domains possess a b-sandwich ∼500 million years ago (6). These characteristic components of structure composed of seven to nine strands (3). Examining the the AIS have been found to exist only in jawed vertebrates. The species distribution of Ig domains, it was found that IgV and IgI emergence of the RAG transposon, along with two rounds of domains were present in the earliest metazoa and associated with whole-genome duplication (7), are believed to have contributed to kinase domains, suggesting that the ancestral function of Igs in- the genesis of the AIS, which is based on clonal selection volved cell signaling (4). Proper C1 set domains are only found in of lymphocytes with somatically rearranged/mutated genes for vertebrates; C2 set domains, including CD4 and other cell surface IgSF domain-containing immune receptors (BCRs and TCRs). receptors, could be found in the most primitive metazoa (5). The Although jawless fish also have an AIS based on clonal selection

*Department of Microbiology and Immunology, College of Veterinary Medicine, China The coordinates and structural factors presented in this article have been submitted to Agricultural University, Haidian District, Beijing 100094, China; †Institute of Veterinary the Protein Data Bank (https://deposit-pdbj.wwpdb.org/deposition/) under accession Medicine, Jiangsu Academy of Agricultural Sciences, Key Laboratory of Veterinary numbers 5XPW and 5XPV. The sequences of the primer pairs presented in this article Biological Engineering and Technology, Ministry of Agriculture, National Center for have been submitted to the National Center for Biotechnology Information’s GenBank Engineering Research of Veterinary Bio-products, Xuanwu District, Nanjing 210014, (https://www.ncbi.nlm.nih.gov/genbank/) under accession numbers 1008457267 and China; ‡Chinese Academy of Sciences Key Laboratory of Pathogenic Microbiology 1008457291. Immunology, Institute of Microbiology, Chinese Academy of Sciences, Chaoyang Dis- x Address correspondence and reprint requests to Prof. Chun Xia, Department of trict, Beijing 100101, China; College of Life Sciences, Fujian Normal University, { Microbiology and Immunology, Key Laboratory of Animal Epidemiology and Zoo- Fujian 350117, China; and Key Laboratory of Animal Epidemiology and Zoonosis, nosis, Ministry of Agriculture, College of Veterinary Medicine, China Agricultural Ministry of Agriculture, College of Veterinary Medicine, China Agricultural University, University, Haidian District, Beijing 100094, China. E-mail address: xiachun@cau. Haidian District, Beijing 100094, China edu.cn ORCIDs: 0000-0003-0932-8523 (Lijie Zhang); 0000-0002-2027-8523 (S.Y.); 0000- The online version of this article contains supplemental material. 0001-8326-2492 (Y.W.); 0000-0001-9459-5030 (B.J.); 0000-0002-3444-0624 (C.X.). Abbreviations used in this article: AIS, adaptive immune system; C1, constant-1; C2, Received for publication June 21, 2017. Accepted for publication February 12, 2018. constant-2; I, intermediate; DSCAM, Down syndrome cell adhesion molecule; IEM, This work was supported by the 973 Project of the China Ministry of Science and immunogold electron microscopy; IgSF, Ig superfamily; LRR, leucine-rich repeat; Technology (Grant 2013CB835302) and the State Key Program of the National rmsd, root-mean-square deviation; V, variable; VCBP, V domain–containing chitin– Natural Science Foundation of China (Grant 31230074). binding protein; VJ-C, variable-J–constant.

R.C., Lijie Zhang, J.Q., Ling Zhang, H.Y., S.Y., Q.Z., B.J., Z.W., and Y.W. performed Ó experiments; R.C., C.X., and N.Z. analyzed data; and C.X. supervised all research. Copyright 2018 by The American Association of Immunologists, Inc. 0022-1767/18/$35.00 R.C. wrote the manuscript, and revisions were made by C.X.

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1700906 2 DISCOVERY AND ANALYSIS OF AMPHIOXUS IgVJ-C2 of lymphocytes with somatically rearranged/mutated genes, with a Materials and Methods proposed role for cytosine deaminases similar to that in jawed Identification of the AmpIgVJ-C2 gene in the amphioxus vertebrates, in jawless fish these genes encode immune receptors genome and transcriptome in the leucine-rich repeat (LRR) category and not the IgSF cate- The V domains of the TCR, Ig L chain, and Ig NAR of sharks (IDs: TCRa: gory (8). The BCR, TCR, and MHC molecules of the AIS appear ADW95876, TCRb: ADW95928, TCRg: ADW95913, TCRd: to have emerged rapidly after the separation of jawless and jawed ADW95927, Ig L chain: A49633, and Ig NAR: 1T6V) were used as query vertebrates and before separation of the shark/ray lineage and sequences to search for AmpIgVJ-C2 in the amphioxus genome (http:// bony vertebrates (7). No convincing candidate precursor-type www.genome.jp/tools/blast/) and transcriptome. BLASTP and default pa- molecule has been found for these AIS molecules. Similar IgV rameters were used for the search. The TMHMM Server (v. 2.0) was used to predict the transmembrane helices in different proteins. Searches for and Ig C1 sets, similar rearrangement mechanisms, and a close genomic information related to AmpIgVJ-C2 (JGI: 122935) were per- linkage between BCR and TCR genes in a shark genome have led formed using the National Center for Biotechnology Information database, to agreement that the diverse AgRs of jawed vertebrates share a the JGI Genome Portal (http://genome.jgi-psf.org/Brafl1/Brafl1.home. common ancestor (6, 9). A unique property of the V sets of TCRs html), and Ensembl. and BCRs is that they contain a J segment, which forms a Amphioxi diglycine bulge and is important for dimerization; these types of V domains have been termed V V domains or V sets (10). For this Branchiostoma belcheri and Branchiostoma japonicum adult Chinese J J amphioxi were cultured at Fujian Normal University (22). The amphioxi reason, IgSF molecules in invertebrates with a VJ-type domain are were kept in artificial seawater and fed plankton. the logical focus of attention when looking for candidate structures representative of the ancestral BCR/TCR molecule. How- Southern blot analysis Downloaded from ever, only two diversified Ig-I/Ig-V fold structures, the Down Genomic DNA (∼35 mg) was digested by the restriction enzyme HindIII syndrome cell adhesion molecule (DSCAM) and V domain–con- (TaKaRa Biotechnology, Dalian, China) at 37˚C over 8 h. After purifica- taining chitin–binding proteins (VCBPs), have been resolved in tion, 15–20 mg of digested DNA was loaded on a 7% agarose gel and run at 25 V overnight at 4˚C. DNA was transferred to a positively charged nylon invertebrates, and neither belongs to the VJ set (11, 12). The corresponding three-dimensional structures show multiple IgI membrane (Amersham Biosciences). Prehybridization was carried at 37˚C for 2 h in hybridization buffer containing 50% formamide. DIG probes domains in DSCAM (13) and two tandem V sets in VCBPs (14). (forward primer 59-TGCGGACAGATACATACTCA-39 and reverse primer VCBP structures show a unique V–V region pairing, which is 59-CATGCACATGTACGTTCCTA-39) were synthesized using a PCR DIG http://www.jimmunol.org/ intramolecular and head-to-tail rather than intermolecular and Probe Synthesis Kit (Roche Applied Science), and hybridization was head-to-head as found in BCRs and TCRs (14). However, the most performed at 4˚C overnight using fresh hybridization buffer that contained the DIG probes. The blot was washed twice in low-stringency buffer (23 important Ig variable-J–constant (VJ-C) connected structures, SSC and 0.1% SDS) at room temperature for 5 min each, followed by two which show higher similarity with vertebrate TCRs and BCRs, washes in high-stringency buffer (0.13 SSC and 0.1% SDS) at 65˚C for have never been characterized in cephalochordates or primitive 15 min each. Finally, the blot was detected using Anti-Digoxigenin-AP and vertebrates. a DIG Random Probe Detection Kit II (Roche Applied Science). With the development of genomic sequencing and transcriptomics, Protein preparation effective strategies have become available to recognize immune by guest on October 1, 2021 receptors, as well as to analyze the origin and evolution of adaptive The IgVJ-C2 and IgVJ domains of AmpIgVJ-C2 were synthesized and expressed using pET21a vector and BL21 (DE3). The selenomethionine- immunity (15, 16). Amphioxus (lancelet) is in the basal taxon of the substituted V-C tandem AmpIgVJ-C2 proteins (Se-AmpIgVJ-C2 -VC) chordates, and its genome reveals that amphioxus is representative were expressed in the auxotrophic Escherichia coli strain B834 (DE3) of the common ancestor of all chordates prior to the two rounds (Novagen; Merck, Darmstadt, Germany). All three protein types were of whole-genome duplication (17). Genomic analysis shows expressed as inclusion bodies, which were dissolved by 6 M Gua-HCl. The proteins were then refolded by the gradual dilution method using buffer that amphioxus has an extraordinarily complex and diversified containing 400 mM L-Arg HCl, 2 mM EDTA, 5 mM GSH, 0.5 mM GSSH, gene repertoire corresponding to the innate immune system, as and 100 mM Tris-HCl (pH 8) at 277 K for 12 h. Next, the soluble shown by the vastly expanded TLR, scavenger receptor, Nod- AmpIgVJ-C2 proteins were concentrated and sequentially purified using a like receptor, LRR, and C1q gene models (16). Several immune Superdex 200 size-exclusion column (GE Healthcare) in Tris buffer pathways, including the TNF/IL-1R network, the TLR and (20 mM Tris [pH 8] and 50 mM NaCl). An additional purification step was carried out using a RESOURCE Q ion-exchange column (GE Healthcare) Nod-like receptor networks, and the complement system, seem and Tris buffers (Buffer A: 20 mM Tris [pH 8] and 5 mM NaCl; Buffer B: to have a role in amphioxus mucosal immunity. The composi- 20 mM Tris [pH 8] and 500 mM NaCl). tion and regulation of these pathways are similar to those of Crystallization vertebrate cellular pathways (18). Moreover, the pro-MHC region and homologs of proteasomes and chaperones potentially The crystallization experiment was performed using the sitting-drop vapor- involved in antigenic peptide processing and presentation by diffusion method at 291 K. Crystals of the native and selenomethionine- substituted V-C tandem AmpIgVJ-C2 proteins were grown for 3–4 mo in MHC have been recognized in the amphioxus genome and 0.2 M lithium sulfate monohydrate, 0.1 M Bis-Tris (pH 5.5), and 25% w/v in a cDNA library (17, 19). In addition, amphioxus RAG1-like polyethylene glycol 3350 at a protein concentration of 5 mg/ml. Crystals of protein, which contains the core region of vertebrate RAG1, is the V domain of AmpIgVJ-C2 appeared in 0.5 M ammonium sulfate, functional in V(D)J recombination with the assistance of 0.1 M sodium citrate tribasic dehydrate, and 1 M lithium sulfate mono- mammalian RAG2 (20), and an active RAG transposon illu- hydrate (pH 5.6) after 1 wk at a protein concentration of 15 mg/ml. minating the origins of V(D)J recombination was recently Data collection and processing discovered (21). These findings indicate that amphioxus has Prior to data collection, all crystals were cryoprotected in a reservoir buffer almost acquired the molecular basis for the emergence of AIS; with 33% (v/v) glycerol and were flash cooled at 100 K. Diffraction data for however, an amphioxus IgVJ-C molecule with similarity to the V domain of AmpIgVJ-C2 were collected on an R-AXIS IV++ image BCRs/TCRs had not been identified. plate detector using a Rigaku rotating-anode x-ray generator with a radiation ˚ In the current study, we identified an amphioxus gene encoding wavelength of 1.5418 A. Data for native and selenomethionine-substituted AmpIgVJ-C2 V-C were collected on beamline NE3A at the KEK synchro- an IgVJ-C–type molecule. We found, using x-ray crystallography, tron facility (Tsukuba, Japan) at a wavelength of 0.98051 A˚ using an ADSC that it forms homodimers and that, especially, the VJ domain Q270 image plate detector. The data collection and refinement statistics are shows high similarity with BCR/TCR structures. summarized in Table I. The IgVJ-C2 structure was determined using the The Journal of Immunology 3 single-wavelength anomalous diffraction method, with selenomethionine as White resin (Sigma), and ultrathin sections (80 nm) were collected on 100- an anomalous signal (23). In particular, heavy atoms were briefly searched mesh nickel grids. The samples were incubated with an anti–AmpIgVJ-C2 by SHELXD (24), and the initial phases were determined using Phaser (25). mAb (25 mg/ml) overnight at 4˚C. A goat anti-mouse IgG secondary Ab Density modification was performed by DM (26). Approximately 90% of the labeled with 10-nm gold particles (Sigma) was used for detection. After AmpIgVJ-C2 residues were automatically traced by ARP/wARP (27). The staining with uranyl acetate, the sections were analyzed using a Hitachi H- structure of the IgVJ domain of AmpIgVJ-C2 was then solved by molecular 7500 transmission electron microscope (Hitachi, Tokyo, Japan) at 80 kV. replacement using the AmpIgVJ-C2 structure as the start model. Extensive Controls, using preimmune mouse IgG as the primary Ab, were run at the model building and restrained refinement were performed using COOT (28) same concentration as the mAb in parallel. Preimmune mouse IgGs were and REFMAC5 (29), respectively. Refinement cycles were performed using purified using a Protein A Sepharose column. the PHENIX package (30). The stereochemical quality of the models was validated in PROCHECK (31). Accession numbers Structural analysis The coordinates and structural factors generated in this study have been submitted to the Protein Data Bank (https://deposit-pdbj.wwpdb.org/ Structural analysis was performed using PyMOL software and the CCP4 program. deposition/) under the following accession numbers: 5XPW for IgV -C2 The structural similarity comparison was performed using the Dali server. Res- J and 5XPV for the IgVJ domain of AmpIgVJ-C2. idues 1–100 from the V domain of AmpIgVJ-C2 were used for V domain structural comparisons, and residues 101–180 from the C domain of AmpIgVJ- C2 were used for C domain structural superimpositions. The interaction force Results between the interfaces was analyzed using LIGPLOT. The PDB codes for the AmpIgV -C2 gene discovered in the amphioxus genome and cited proteins and the regions for structural and sequence comparisons are as J follows: sIg: camel single-chain Ab, 1MEL:A (residues 2–133); IgNAR: shark cDNA transcriptome IgNAR, 1T6V:O (residues 2–113); TCRd: 1TVD:A (residues 1–116); VCBP: The complete amphioxus (Branchiostoma floridae) genome and

1XT5:A (residues 1–135); NITR: 2QQQ:A (residues 3–111); TCRa: 2VLM:D Downloaded from (residues 3–111); TCRb: 2VLM:E (residues 5–115); IgH: 2X7L:A (residues RNA transcriptome were subjected to a primordial and novel IgV- 1–116); IgL: 2X7L:B (residues 2–133); DSCAM: 2V5M:A (residues 209– C molecule candidate search; the results revealed that only one 307); CD2: 1HNF:A (residues 105–182); B2m: 3GBL:A (residues 1–97); full gene (JGI: 122935), which we designated AmpIgVJ-C2, had IgH: 2X7L:A (residues 121–221); axonin: 1CS6:A (residues 299–388); CD4: 25 28 an E-value . e (about 5e ). AmpIgVJ-C2 encodes 233 aa, 3CD4:A (residues 97–178); ACX: 1ACX:A (residues 1–107); GGT: 1GGT:A including an N-terminal signal peptide, an Ig-V domain, an Ig-C (residues 43–185); and Bence-Jones: Bence-Jones protein l L chain 2MCG:A (residues 2–110 for the V domain and 111–214 for the C domain). domain, a transmembrane region with positively charged amino acids, and a short C-terminal cytoplasmic segment (Supplemental http://www.jimmunol.org/ Diversity of the AmpIgVJ-C2 gene in amphioxi Fig. 1A, 1D). The charged amino acids in the transmembrane region The adult Chinese amphioxi B. belcheri and B. japonicum were examined. of IgR are a typical feature of the conserved AgR transmembrane mRNA extraction and cDNA library construction were performed. The motif (33), which recruits downstream signaling molecules, such as sequences of the primer pairs used to search for diversity were as follows: CD3. The positively charged residue in AmpIgV -C2 is arginine, forward primer, 59-TTCGAGTGCACCTTCCCGATCAGTGC-39 and re- J verse primer, 59-CCGTTGTCTGCCATGCACATGTACGT-39. PCR was and the position of this residue is more similar to that found performed with ExTaq polymerase (TaKaRa Biotechnology, Dalian, in chicken TCRb than to that found in mammals (Supplemental China) in a total volume of 50 ml, as follows: 98˚C for 5 min, 72˚C for Fig. 1C). The AmpIgVJ-C2 gene is located in the amphioxus 1 min, followed by 32 cycles of 94˚C for 50 s, 60˚C for 50 s, and 72˚C for genome B. floridae v1.0 scaffold_104: 461541–472650. The 50 s, with a final elongation step at 72˚C for 10 min. The PCR products region surrounding AmpIgV -C2 bears no obvious synteny by guest on October 1, 2021 were inserted into the pMD-18T vector (TaKaRa Biotechnology), and the J positive clones were selected and sent to Sangon Biotech (Shanghai, with the genomes of vertebrates, but it is notable that a number China) for sequencing. Once the same sequence was verified from two of the neighboring genes encode IgSF, LRR, or C-type lectin independent PCR reactions in an individual sample, this sequence was domains. The IgSF, LRR, and C-type lectin domain families regarded as a new allele sequence. Finally, these sequences were submitted have been extensively used for the creation of immune varia- to GenBank (National Center for Biotechnology Information) under accession numbers 1008457267 and 1008457291. tion between and within vertebrate species, and we speculate that this amphioxus genomic region may encode several im- Localization of AmpIgVJ-C2 proteins in amphioxus tissues mune functions. determined using immunohistochemistry When comparing the gene with those found in agnathan lamprey, Individuals from both Chinese amphioxus species were fixed in 4% (v/v) AmpIgVJ-C2 was found to be similar to the reported lamprey polyoxymethylene for 48 h, dehydrated, and embedded in paraffin using TCR-like molecule, which is regarded as an evolutionary precursor routine methods. Serial paraffin sections (4 mm) were prepared, incubated of modern TCR genes (34). Furthermore, the neighboring LRR with an anti–AmpIgVJ-C2 mAb (12 mg/ml; made by Institute of Genetics domain genes are homologs of lamprey VLRs, which function as and Developmental Biology, Chinese Academy of Sciences) overnight at 4˚C, and detected using goat anti-mouse IgG conjugated to HRP and a Fast the alternative adaptive immune receptors of jawless vertebrates DAB Kit (Zhongshan Jinqiao Biotechnology). The secondary anti-mAb in- (8) (Supplemental Table I). This finding indicates that AmpIgVJ-C2 cubation was also performed at 4˚C. The conventional method for producing has a genetic relationship with the conventional immune genes of mAbs was used (32). In particular, purified AmpIgVJ-C2 protein was used to vertebrates; however, the genome synteny among amphioxus, immunize a BALB/c mouse three times. Approximately 3 d after the final immunization, the lymphoid cells of the immunized mouse were isolated humans, and other vertebrates revealed no obviously similar re- from the spleen and fused with myeloma cells (SP2/0) using polyethylene gion corresponding to AmpIgVJ-C2. glycol. The hybridoma cells were screened using hypoxanthine/aminopterin/ thymine medium. Every hybridoma cell, each of which had single-Ab Nucleotide variation in AmpIgVJ-C2 transcripts specificity, was subcultured independently. After examination of the immu- nological potencies of the different Abs from the different hybridoma cells Some invertebrate immune receptors, such as VCBPs, DSCAM, using Western blotting and ELISA, the best hybridoma cell line was injected and FREPs, generate diversity via certain primitive mechanisms into the abdominal cavity of a mouse (BALB/c) for propagation. Specific (11, 35, 36). Unlike for VCBPs, there is only one copy of the high-affinity mAbs from ascites fluid were then purified sequentially using a AmpIgV -C2 gene in the amphioxus genome (Supplemental Fig. HiTrap Protein A HP column (5 ml; GE Healthcare Life Sciences) and gel J filtration. The isotype of the mAb is IgG2a. 1B). To verify the variability of AmpIgVJ-C2, we tested the cDNA library from a single amphioxus (B. belcheri or B. japonicum). Ultrastructural detection of AmpIgVJ-C2 protein in cells When the transcripts of 22 validated AmpIgVJ-C2 genes were Amphioxus tissue samples were fixed in 2.5% glutaraldehyde for 24 h. After isolated from B. belcheri using RT-PCR, 9 mutational transcripts dehydration, the samples were embedded in low temperature-resistant LR were found. In addition, 11 mutated AmpIgVJ-C2 sequences were 4 DISCOVERY AND ANALYSIS OF AMPHIOXUS IgVJ-C2

FIGURE 1. Alignment of the amino acid sequences deduced from AmpIgVJ-C2 transcripts from B. belcheri (BB) and B. japonicum (BJ). Mutational amino acids are highlighted in green. The second topology of AmpIgVJ-C2 is indicated above the sequence. CDR-like regions are labeled (CDR1L–3L).

The identities among the AmpIgVJ-C2s from different species are shown at the bottom of the ﬁgure. The accession numbers for GenBank PopSet are as Downloaded from follows: B. belcheri, 1008457267 (https://www.ncbi.nlm.nih.gov/popset/) and B. japonicum, 1008457291 (https://www.ncbi.nlm.nih.gov/popset/). BF,

AmpIgVJ-C2 protein form B. ﬂoridae.

found among 24 validated cDNA clones in one B. belcheri. In- the AmpIgVJ-C2 structure was used for detailed structural anal- stead of gene segment rearrangement, the diversity of the yses throughout this study. A structural study showed that V do- AmpIgVJ-C2 molecule is generated by point mutations, similar to mains are present in species as primitive as amphioxus (14); http://www.jimmunol.org/ the FREPs of snails (35) (Fig. 1). We interpret these findings to however, to our knowledge, this is the first study to reveal a V-C mean that AmpIgVJ-C2 in amphioxus may show a primordial Ig- combination structure and a V domain of the VJ category in a like mutational pattern. nonvertebrate species. According to the solved structures, the first AmpIg-V domain consists of A, B, C, C9,C99, D, E, and F and Structural analysis reveals an IgV-C2–organized molecule nine b strands, with a root-mean-square deviation (rmsd) of 1.8 A˚ The structure of AmpIgVJ-C2 confirmed that it is an Ig-VJ and Ig- with IgH-V (Fig. 2A, 2B). The second Ig domain lacks C99 and D C2 tandem molecule. Crystals of the V (AmpIgV) and AmpIgVJ-C2 strands and is a typical Ig-C2 domain (Fig. 2C, see below). The 15 79 domains of AmpIgVJ-C2 diffracted to 2.0 and 1.9 A,˚ respectively. conserved disulfide bonds are Cys -Cys in the Ig-V domain and 125 165 The data collection and refinement statistics are summarized in Cys -Cys in the Ig-C2 domain. These findings verify that the by guest on October 1, 2021 Table I. Because the V domains from the AmpIgVJ structure and IgV-C2 structure of AmpIgVJ-C2 resembles modern immune re- AmpIgVJ-C2 structure are nearly the same (Supplemental Fig. 2), ceptors, including BCRs and TCRs, which are characterized by

Table I. Data collection and reﬁnement statistics

AmpIgVJ-C2 V AmpIgVJ-C2 V-C Se AmpIgVJ-C2 V-C Data Collection Space group P22121 P3121 P3121 Cell dimensions a, b, c (A)˚ 45.41, 67.79, 74.63 41.76, 41.76, 213.72 41.76, 41.76, 213.71 a, b, g (˚) 90.0, 90.0, 90.0 90.0, 90.0, 120.0 90.0, 90.0, 120.0 Wavelength (A)˚ 1.5418 0.98051 0.98051 Absorption (Se) Peak Resolution (A)˚ 50.00–1.90 (1.97–1.90) 50.00–2.00 (2.07–2.00) 50.00–2.25(2.33–2.25) Rsym or Rmerge 0.049 (0.186) 0.094 (0.439) 0.093 (0.283) Ι /sΙ 35.7 (11.21) 24.8 (3.02) 35.5 (12.6) Completeness (%) 99.9 (100.0) 98.2 (93.1) 99.4 (100.0) Redundancy 7.6 (7.6) 15.9 (6.8) 19.1 (18.5) Reﬁnement Resolution (A)˚ 23.35–1.90 34.25–2.00 No. of reﬂections 18,730 14,607 Rwork/Rfree 0.181/0.207 0.226/0.266 No. of atoms Protein 1,580 1,396 Water 202 72 B factors Protein 25.18 45.96 Water 38.42 57.18 rmsd values Bond lengths (A)˚ 0.005 0.06 Bond angles (˚) 0.819 0.828

Rmerge = + |I2 〈I〉|/+I, where I is the integrated intensity of a given reﬂection. R = + || Fobs | 2 |Fcalc ||/+hkl |Fobs |. Rfree was calculated using 5% of the data omitted from reﬁnement. Ι /sΙ = average (Ι /sΙ). Data in parentheses are for the highest-resolution shell. The Journal of Immunology 5

FIGURE 2. Overall structure of

AmpIgVJ-C2 and comparison of the V domain of AmpIgVJ-C2 with IgH-V and the AmpIgVJ-C2C domain. (A) Overall structure of AmpIgVJ-C2. The V and C domains are marked. The disulﬁde bond is shown as pink sticks. The strands are shown in green, and the loops are shown in gold. (B) Structural comparison of the V domain of

AmpIgVJ-C2 and the most similar V set (i. e., IgH-V). IgH-V is shown in a purple Downloaded from color. The major differences are in the BC loop, FG loop, and A9 strand; these are shown in red. AmpIgVJ-C2 has longer BC and FG loops compared with IgH-V. (C) Structural comparison of the V domain and the C domain of AmpIgVJ-C2. The V and C

domains are shown in gray and green, re- http://www.jimmunol.org/ spectively. The C99 and D strands of the V domain of AmpIgVJ-C2 are shown in red. by guest on October 1, 2021

one V domain and one or several C domains. However, this IgH-V (∼6%), the IgV fold of AmpIgVJ-C2 most closely resem- structure differs from those of the two Ig-V jointed VCBPs. bles IgH-V at the structural level, with an rmsd of 1.8 A.˚ A similar result was observed for IgL, TCRa, and TCRb (Fig. 4). With The AmpIg-V domain of IgVJ-C2 exhibits the typical Ig-V fold that acquired mutation capability in CDR analogs regard to the sequence identity, it shows the highest homology to the l L chain (25%), which, although a modern molecule, can also The typical V-type Ig domain differs from the C1, C2, and I sets by form Bence-Jones structures that have been speculated to have the C9 and C99 strands inserted between the C and D strands (1). similarities with primitive AgRs (37, 38). The major structural The front and back sheet of the AmpV domain of IgV -C2 consist J differences between the V domain of AmpIgV -C2 and IgH-V can of AGFCC9C99 and BED, respectively, indicating a classical Ig-V J be observed in the BC and FG loops, which are homologous to the fold (Fig. 3A). In the absence of an A9 strand, the arrangement of CDR1 and CDR3 loops, respectively. Compared with IgH-V, the the b strands of the V domain of AmpIgVJ-C2 is more similar to that of the V domain of the Ig L chain (3DVG) than to those of BC and FG loops of the V domain of AmpIgVJ-C2 are extraor- TCRs. The conserved disulfide bond formed by Cys15 and Cys79 dinarily long, which may indicate the presence of a different in the B and F strands, respectively, connects the front and back dominant Ag-recognition region (Fig. 1B). Compared with the 28 sheets. The invariant Trp packs against the bond to form the first V domain of VCBP from the same amphioxus, AmpIgVJ-C2 hydrophobic core of the V set (Fig. 3B). The eight invariant has an rmsd of 2.2 A,˚ as well as longer BC and FG loops and a conserved residues that form the backbone of the V set Ig domain shorter C9C99 loop (Fig. 4). These findings indicate that, among 8 15 28 and are detected in AmpIgVJ-C2 are as follows: Gly ,Cys ,Trp , elucidated protein structures, the V domain of AmpIgVJ-C2 has Arg53, Leu64, Asp73,Tyr77, and Cys79 (Figs. 3A, 4). Compared the most sequence and structural similarity to the V domains of with the V domains of adaptive immune receptors, the BC, C9C99, modern AgRs, despite its presence in cephalochordates. and FG loops had the coordinates of the CDR1, CDR2, and CDR3 In amphioxus VCBP3, the two tandem V regions exhibited loops, respectively. To deduce the structural relationship between a head-to-tail relationship, forming opposing CDR-like loops. the V set domains of AmpIgVJ-C2 and modern immune receptors, However, similar to the heterodimeric AgRs in jawed vertebrates we compared the V domain of AmpIgVJ-C2 with the V domains (BCRs and TCRs), the V domain dimer of AmpIgVJ-C2 adopts of diversified Ag recognition molecules at the structural and se- head-to-head packing, positioning six CDR-analogous loops (the quence levels. Although it has the lowest sequence identity with BC, C9C99, and FG loops) from two V domains to form a merged 6 DISCOVERY AND ANALYSIS OF AMPHIOXUS IgVJ-C2 Downloaded from http://www.jimmunol.org/ by guest on October 1, 2021

FIGURE 3. Structure of the V domain of AmpIgVJ-C2 and its diversified CDR-analogous loops. (A) The overall structure of the V domain of AmpIgVJ-C2. The nine b strands and the CDR-analogous loops (BC, C9C99, and FG loops) are labeled. The conserved 8 aa of the Ig-V set in AmpIgVJ-C2 are shown 28 as sticks. (B) The hydrophobic core of the V domain of AmpIgVJ-C2. This core consists of an invariant tryptophan residue (Trp ) packed against the 15 79 disulfide bond (Cys -Cys ). (C) A distribution comparison of the CDR-analogous loops in the AmpIgVJ-C2 dimer and the CDR loops in the IgH-l dimer. The V-domain dimer of AmpIgVJ-C2 is shown as a white surface. The CDR-analogous loops (BC, C9C99, and FG loops) of AmpIgVJ-C2 are shown as orange loops and surface. The CDR loops of IgH-l are violet and superimposed with AmpIgVJ-C2. (D) V sites in the CDR-analogous loops in AmpIgVJ-C2. These sites have point mutations according to the transcript sequence alignment, and they map to the V domain of the AmpIgVJ-C2 structure. These sites are shown as cyan sticks. A close-up schematic of the diverse residues in the FG (E), C9C99 (F), and BC (G) loops. The structures of other AmpIgVJ-C2 transcripts have been modeled according to the resolved AmpIgVJ-C2 structure using SWISS-MODEL, and they overlap with the resolved AmpIgVJ-C2 structure. The mutated residues are shown as sticks and are labeled accordingly. region for potential binding of ligands (Fig. 3C). To verify the CDR-analogous loops. In B. belcheri, these sites in CDR- 22 23 39 variability of AmpIgVJ-C2 CDR-analogous loops, we further analogous loops include Ala and Arg in the BC loop, Phe mapped the mutational sites in the CDR-analogous loops and b and Thr41 in the C9C99 loop, Ser40 in the C9C99 loop, and Gly88 in strands of the AmpIgVJ-C2 structure. Interestingly, we can find the FG loop (Figs. 1, 3D–G). These results indicate that AmpIgVJ-C2 mutational sites throughout the entire AmpIgVJ-C2. Conversely, has the ability to generate sequence diversity and might bind to there are more mutational sites in the b strands than in the varied ligands; however, the mechanism and the manner used by The Journal of Immunology 7 Downloaded from http://www.jimmunol.org/ by guest on October 1, 2021

FIGURE 4. Sequence comparison of the V and C domains of AmpIgVJ-C2 with other immune receptors. (A) Sequence comparison of the V domains between AmpIgVJ-C2 and the other immune receptors. The conserved eight residues of the Ig-V set are denoted by brown stars. The J-chain motifs are denoted by blue triangles, and the J-segment motifs of AmpIgVJ-C2 are indicated. Conserved amino acids are highlighted in green. The sequence identity and rmsd between AmpIgVJ-C2 and other V domains are provided to the right of each sequence. (B) Sequence comparison of the C domains between AmpIgVJ-C2 and other immune receptors. In the C2 and I category sequences, residues belonging to the conserved eight residues shared by the Ig-V set and Ig-I set are shaded green. IntheC1 category sequences, the characteristic residues of Ig-C1 domains (5) are shaded cyan. The sequence identity and rmsd are provided to the right of each sequence.

AmpIgVJ-C2 are different from those of modern adaptive immune acid), and this rather unique fragment plays a major role in AgR receptors. dimerization and combinational diversity. A corresponding J segment–like sequence was observed in AmpIgV -C2, which b J A J segment with a -bulge motif is preserved in AmpIgVJ-C2 contains the sequence GDGTRYQLXV in the G strand of the V The J chain in the V domains of BCRs and TCRs is characterized by domain (Fig. 5D). A b-bulge, which is the typical structural fea- the conserved motif FGXGTXLXV (where X represents any amino ture of the J chain found in AgRs, was also observed in the 8 DISCOVERY AND ANALYSIS OF AMPHIOXUS IgVJ-C2

FIGURE 5. The J-corresponding segment of

AmpIgVJ-C2. (A) Structural comparison of the J-corresponding segments from AmpIgVJ-C2 (green), VCBP (yellow), IgH (pink), TCRa (blue), and TCRd (sand). The J-corresponding segment motifs are shown as sticks. (B) The b-bulge and J-corresponding segment motifs of AmpIgVJ-C2. The b-bulge is formed by a salt bridge with a 8.3-A˚ distance. The J-corresponding segment motifs are labeled and are shown as sticks. The electron density for these residues is shown in green mesh. (C) The b-bulge and J segment motifs of TCRa, which exhibit structural similarities with

AmpIgVJ-C2. The J segment motifs are labeled and are shown as sticks. (D) Comparison of the J-correspond- Downloaded from ing segment motifs from AmpIgVJ-C2, VCBP, IgH, TCRa, and TCRd. The consensus sequence is shown at the bottom of the alignment. The two most conserved G residues are shown in red. The other motifs (F/W, T, and L/V and V) are shown in yellow, orange, and green, respectively. http://www.jimmunol.org/

AmpIgVJ-C2 structure, located in the middle of the G strand amino acids from the strongly twisted edge strands C9 and G, and by guest on October 1, 2021 (Fig. 5A, 5D). The bulge of AmpIgVJ-C2, which is formed by a the two outer layers are formed by the central b strands C and F salt bridge between Asp90 and Arg93, was found to be 2.9 A˚ larger (Fig. 6B, 6D). The residues that constitute the inner layer of the V 32 29 88 than the bulges of BCRs/TCRs (Fig. 5B, 5C). Based on super- domains of the AmpIgVJ-C2 homodimer are Gly ,Tyr , Gly , 90 imposition, the J-corresponding segment of AmpIgVJ-C2 is most and Ser , whereas the C and F strands form the outer layers similar to the J chain of TCRb in terms of structure (Fig. 5A–C) (Fig. 6A, 6C). and is less similar to the J-corresponding segment of VCBP in terms The additional inner face is primarily formed by a hydrophobic of sequence and structure (Fig. 5A, 5D); the J-corresponding segment interaction that is absent from the two-layer packing interface. of VCBP does not have a diglycine motif or a bulge structure Therefore, we analyzed the intermolecular force of the AmpIgVJ-C2 (Fig. 5A, 5D). The findings for AmpIgVJ-C2 indicate that V domains dimer. Like TCRs and BCRs, hydrophobic bonds are the dominant 29 37 with a J-segment already existed at the primitive cephalochordate forces in the AmpIgVJ-C2 dimer; only four residues (Tyr , Asn , level. Gly88, and Ser90) from one V domain form three pairs of hydrogen bonds with corresponding residues in another V domain Discovery of the three-layer packing dimerization pattern of V (Table II, Supplemental Fig. 3). In total, 27% of the intermo- domains in the AmpIgVJ-C2 structure lecular forces of AmpIgVJ-C2 are derived from residues in the Due to selective pressure to recognize more diverse invaders Gstrand(Gly85,Ala86,Gly87,Gly88,andSer90)(TableII, (compared with innate immune receptors), modern Ag-binding Supplemental Figs. 2, 3). The J gene–related segments encode molecules have adopted the V-V domain packing dimer to ex- these residues, which indicates that the AgR-like J segment pand their Ag-recognition repertoire. The crystal structures of induces dimer packing, even though AmpIgVJ-C2 and AmpIgVJ AmpIgVJ-C2 and AmpIgVJ reveal that both are dimers, verifying are homodimers, whereas typical BCRs and TCRs are (based on) their ability to undergo dimerization (Fig. 6, Supplemental Fig. 2). heterodimers. The rmsd between the dimers of AmpIgVJ-C2 and AmpIgVJ is only 0.655 A˚ (Supplemental Fig. 2), and they share the same The second domain of AmpIgVJ-C2 is an Ig-C2 domain dimer-packing pattern through their V domains (Fig. 6, The second domain of AmpIgVJ-C2, which has a classical b-barrel Supplemental Fig. 2E). configuration, contains two antiparallel b sheets that consist of The dimer interfaces of the adaptive immune receptors do not A, B, and E strands and C, C9, F, and G strands. Similar to Ig-Vs, possess ordinary b-sheet packing in terms of face-to-face contact the conserved aromatic amino acid Trp137 of the C domain of of the molecules. According to the dimer structures of AmpIgVJ-C2, AmpIgVJ-C2 (AmpIg-C2) is packed toward the disulfide bond the two V domains interact with each other using the same Cys125-Cys165 to form the conserved hydrophobic core and sup- front-sheet-to-front-sheet mode, which defines the so-called three- ports the total structure with other hydrophobic amino acids layer packing found in AIS AgRs. The inner layer consists of (Fig. 7A). The distance between the two cysteines is 40 residues, The Journal of Immunology 9

FIGURE 6. Similar dimerization pattern between

AmpIgVJ-C2 and Ig. (A) The three-layer packing mode of the V-V dimer of AmpIgVJ-C2. Hydrophobic residues that form the inner layer are shown as yellows and are labeled. The C and F strands that form the outer layers are shown in violet. (B) Three-layer packing of the IgH-L V-V dimer. The IgH-L dimer is shown using colors similar to those used for AmpIgVJ-C2. The inner-layer residues of the three-layer packing are shown as yellow sticks. C and F strands are in violet. (C)

Three-layer packing interface of AmpIgVJ-C2. (D) The three-layer packing interface of IgH. Downloaded from http://www.jimmunol.org/ which indicates that it is a C-type Ig domain. Based on the this structure exhibits considerable differences from other absence of a D strand and of most C1-specific residues, this elucidated C2 structures. configuration should be classified as an Ig-C2 domain The AmpIgVJ-C2 protein is located in immunity-associated (Fig.4B).IntheAmpIgVJ-C2 C2 domain, six of the eight characteristic Ig-V/Ig-I–type motifs are conserved (Fig. 4B), cells and organs of amphioxus which is reminiscent of some other C2 category sequences and Immunohistochemistry was used to characterize the expression of is in agreement with continuity between the several IgSF do- AmpIgVJ-C2 in different organs. Tissues from different individ- main categories (5). During the search for the identity struc- uals from both types of Chinese amphioxus were used to avoid by guest on October 1, 2021 ture, the Ig-I type, such as those found in axonin (rmsd = 1.9 A)˚ individual and species differences. Strongly reactive signals were and DSCAM (2V5M, rmsd = 1.9 A),˚ exhibited the most simi- observed for anti–AmpIgVJ-C2 Ab in tissue sections of the gill, larity with AmpIgVJ-C2 C2. Compared with Ig-C1–type and hepatic diverticulum, and gut, based on immunohistochemical other Ig-C2–type molecules, AmpIgVJ-C2C2hasgreater staining (Fig. 8A–F). These tissues constitute the digestive and structural deviations, with rmsd values of 3.0 A˚ (2X7L) and 2.2 respiratory organs in amphioxus and are considered primordial A˚ (3CD4) (Fig. 7B–D). In summary, the above findings reveal immune tissues, which are continuously exposed to various po- that the AmpIgVJ-C2 constant domain has a C2 structure, but tential pathogens.

Table II. Interactions between the V domains of AmpIgVJ-C2

AmpIgVJ-C2 -V(B) Hydrophobic Interactions and Van der Waals AmpIgVJ-C2 (A) AmpIgVJ-C2 -V(A) Hydrogen Bonds Contacts T23 L39 V25 V25, I82, I84 N27 I82, S90 Y29 Y29(Oh)-S90(OG) I82, S90 S34 G88, S90 N37 N37(ND2)–G85(O), N37(OD1)–G85(N) I82, E83, I84, G85, G88 Y38 I84, L39 T23, L39, I84 A45 G85, A86 T47 G87, G88 I82 V25, N27, Y29, N37, I82 E83 N37 I84 V25, N37, Y38, L39 G85 G85(N)–N37(OD1), G85(O)–N37(ND2) N37, A45, T47 A86 A45, T47 G87 T47 G88 S34, N37, T47 S90 S90(OG)–Y29(Oh) Y29, N27, S34 Hydrophobic interactions are underlined. 10 DISCOVERY AND ANALYSIS OF AMPHIOXUS IgVJ-C2 Downloaded from http://www.jimmunol.org/ by guest on October 1, 2021

FIGURE 7. Overall structure of the Ig-C2 domain of AmpIgVJ-C2 and its comparison with the I and C sets. (A) Total structure of the C domain of 106 108 121 AmpIgVJ-C2. The disulﬁde bridge and invariant tryptophan are shown as hot pink sticks. The hydrophobic residues (Leu and Leu on strand A; Met , Ser123, and Ala127 on strand B; Phe135 and Leu139 on strand C; Thr145 and Gly142 on strand C9; Leu150 and Leu152 on strand E; Ala167 and Tyr163 on strand F; 178 180 and Val and Ile on strand G) are represented by orange lines. The strands are labeled accordingly and show that the C domain of AmpIgVJ-C2 is an Ig- C2 domain. (B) Comparison of the C domain of AmpIgVJ-C2 and the axonin Ig-I domain revealed high total structural similarity with the I set. The major differences are in the C9 and A strands of AmpIgVJ-C2 and are shown in hot pink. (C) Comparison between the C domain of AmpIgVJ-C2 and the IgH C1 domains, showing less similarity with the Ig-H C1 domain. The rmsd between the two structures is 3 A.˚ The different regions consist of the C9 strand and loop regions. (D) Comparison between the C domain of AmpIgVJ-C2 and the CD4 C2 domains, showing an overall structure similar to that of the C2 domain. The different regions consist of the C9 strand, the AB loop, and the BC loop. The rmsd between the structures is 2.2 A.˚

Recently, the anatomical structure of amphioxus intestine was organ is a major immune response site (39). To determine whether shown to differ from those of vertebrates by its abundant ciliated the subcellular localization of the AmpIgVJ-C2 molecule is re- cell protrusions and secretory granules, which indicate that this lated to defense or immunity, immunogold electron microscopy The Journal of Immunology 11 Downloaded from http://www.jimmunol.org/ by guest on October 1, 2021

FIGURE 8. Immunohistochemical and IEM localization of AmpIgVJ-C2. (A, C, and E) Immunohistochemical localization of AmpIgVJ-C2. Positive immunohistochemical signals were detected in hepatic diverticulum (A), gill (C), and gut (E) sections. mAb binding to AmpIgVJ-C2 in the tissues is represented by diffuse brown staining. (B, D, and F) Negative controls for (A), (C), and (E), respectively. For (B), (D), and (F), preimmune mouse IgG at the same concentration as that of the mAb is used as the primary Ab in the controls. All experiments were run in parallel under the same conditions. (G) IEM localization of AmpIgVJ-C2. Immunogold labeling localized to intestinal epithelial cells (asterisk). A high density of labeled particles was observed in cilia, cell protrusions, and secretory granules. (H) Control section using puriﬁed IgG from preimmune mouse serum. C, cilia; CP, cell protrusion; S, secretory granule.

(IEM) was used to label thin sections of the intestine. AmpIgVJ- 8H). These staining patterns are consistent with a secretory form + C2–bound gold particles were primarily localized in cell protru- of AmpIgVJ-C2, which agrees with the idea that AmpIgVJ-C2 sions (especially in secretory granules) and cilia. Gold particles granules are occasionally secreted from the surfaces of cells in the were also observed in secretory granules in cytoplasm (Fig. 8G, intestinal epithelium. 12 DISCOVERY AND ANALYSIS OF AMPHIOXUS IgVJ-C2

Discussion Overall, these findings help to illuminate the evolution of the Substantial effort has been made to understand the emergence of diversified IgV-C proteins of jawed vertebrates, and our study the AIS by searching for diversified immune genes in invertebrates importantly revealed the structural presence of VJ-C2 organized and vertebrates. However, when referring to lower organisms, molecules in cephalochordates. sequence comparison alone is not always sufficient for studying the evolution of molecules, especially when the sequence identity is Acknowledgments low (,30%). For this reason, structural comparison has become a We thank George F. Gao (Chinese Academy of Sciences) and Hans (J.M.) useful tool for the study of molecular evolution (40). Thus far, few Dijkstra (Fujita Health University) for helpful suggestions. We also ac- Ig domain–containing molecular structures have been elucidated knowledge the assistance provided by the staff at the KEK synchrotron fa- in invertebrates; these include FREPs in snails, DSCAMs in flies, cility in Japan. and VCBPs in amphioxus (12, 14, 35). In invertebrates, the elu- Disclosures cidated AmpIgVJ-C2 structure is the first structurally determined The authors have no financial conflicts of interest. V-C–organized molecule and the first structurally determined VJ domain. At the sequence level, some other V-C–organized molecules in invertebrates have been discussed with regard to their References possible relation to BCR/TCR evolution, but their structures have 1. Bork, P., L. Holm, and C. Sander. 1994. The immunoglobulin fold. Structural not been determined (41). classification, sequence patterns and common core. J. Mol. Biol. 242: 309–320. Although not identical, the presence of the J segment–like 2. Hill, R. L., R. Delaney, R. E. Fellows, and H. E. Lebovitz. 1966. The evolutionary origins of the immunoglobulins. Proc. Natl. Acad. Sci. USA 56: 1762– structure in AmpIgVJ-C2, and especially the characteristic 1769. Downloaded from 3. Smith, D. K., and H. Xue. 1997. Sequence profiles of immunoglobulin and b-bulge in the G strand, provides evidence that AmpIgVJ-C2 is in immunoglobulin-like domains. J. Mol. Biol. 274: 530–545. the same family as the BCR/TCR molecules. The similarities 4. Halaby, D. M., and J. P. Mornon. 1998. The immunoglobulin superfamily: an between AmpIgVJ-C2 and BCR/TCR extend to the head-to-head insight on its tissular, species, and functional diversity. J. Mol. Evol. 46: 389– interaction of the V domains and their three-layer packing mode. 400. 5. Williams, A. F., and A. N. Barclay. 1988. The immunoglobulin superfamily– AmpIgVJ-C2 is the only intermolecular V-V dimer structure that domains for cell surface recognition. Annu. Rev. Immunol. 6: 381–405. has been identified in invertebrates. The homodimer configuration 6. Venkatesh, B., A. P. Lee, V. Ravi, A. K. Maurya, M. M. Lian, J. B. Swann, of AmpIgV -C2 is reminiscent of the L chain dimer of Bence- Y. Ohta, M. F. Flajnik, Y. Sutoh, M. Kasahara, et al. 2014. Elephant shark ge- http://www.jimmunol.org/ J nome provides unique insights into gnathostome evolution. [Published erratum Jones protein (38) in humans, which is also a homodimer. The appears in 2014 Nature 513: 574.] Nature 505: 174–179. 7. Flajnik, M. F., and M. Kasahara. 2010. Origin and evolution of the adaptive AmpIgVJ-C2 dimer has a greater resemblance to the dimer formed immune system: genetic events and selective pressures. Nat. Rev. Genet. 11: 47– by junctional adhesion molecule 1, which homodimerizes only by 59. its V domains and has a C2 category domain (42), than to those 8. Boehm, T., N. McCurley, Y. Sutoh, M. Schorpp, M. Kasahara, and M. D. Cooper. found in TCRs and BCRs, which are V-C1–organized molecules 2012. VLR-based adaptive immunity. Annu. Rev. Immunol. 30: 203–220. 9. Glusman, G., L. Rowen, I. Lee, C. Boysen, J. C. Roach, A. F. Smit, K. Wang, that use their V domains, as well as their C1-category domains, for B. F. Koop, and L. Hood. 2001. Comparative genomics of the human and mouse dimer formation. In invertebrates, a few C1-like sequences have T cell receptor loci. Immunity 15: 337–349. 10. Eason, D. D., J. P. Cannon, R. N. Haire, J. P. Rast, D. A. Ostrov, and

been found (41), but whether they are truly related to C1 is de- by guest on October 1, 2021 G. W. Litman. 2004. Mechanisms of antigen receptor evolution. Semin. Immu- batable, and their structures remain to be determined. For now, nol. 16: 215–226. 11. Cannon, J. P., R. N. Haire, and G. W. Litman. 2002. Identification of diversified with the elucidation of the AmpIgVJ-C2 structure, it can be con- genes that contain immunoglobulin-like variable regions in a protochordate. Nat. cluded that important features of the VJ domain of BCRs and Immunol. 3: 1200–1207. TCRs were already present in early chordates, whereas the timing 12. Watson, F. L., R. Pu¨ttmann-Holgado, F. Thomas, D. L. Lamar, M. Hughes, of the emergence of C1-domain features needs further investiga- M. Kondo, V. I. Rebel, and D. Schmucker. 2005. Extensive diversity of Ig- superfamily proteins in the immune system of insects. Science 309: 1874–1878. tion. We, like other investigators, deem it likely that the vertebrate 13. Meijers, R., R. Puettmann-Holgado, G. Skiniotis, J. H. Liu, T. Walz, J. H. Wang, BCR and TCR molecular lineage originated from primitive ad- and D. Schmucker. 2007. Structural basis of Dscam isoform specificity. Nature hesion molecules (42) or AgRs (43), which did not have the V -C1 449: 487–491. J 14. Hernańdez Prada, J. A., R. N. Haire, M. Allaire, J. Jakoncic, V. Stojanoff, structure. J. P. Cannon, G. W. Litman, and D. A. Ostrov. 2006. Ancient evolutionary origin Under selective pressure, Ig molecules in the gnathostome AIS of diversified variable regions demonstrated by crystal structures of an immune- type receptor in amphioxus. Nat. Immunol. 7: 875–882. evolved to distinguish between commensal microorganisms and 15. Rast, J. P., L. C. Smith, M. Loza-Coll, T. Hibino, and G. W. Litman. 2006. pathogenic invaders in the primitive gut (44). Certain TCRs and Genomic insights into the immune system of the sea urchin. Science 314: 952– MHC class I/II molecules in mammals exert an innate-like immune 956. 16. Huang, S., S. Yuan, L. Guo, Y. Yu, J. Li, T. Wu, T. Liu, M. Yang, K. Wu, H. Liu, function in the gut mucosa, indicating that the gut is an ancient et al. 2008. Genomic analysis of the immune gene repertoire of amphioxus re- immune organ that existed prior to the acquisition of diverse reveals extraordinary innate complexity and diversity. Genome Res. 18: 1112– ceptors (45, 46). The digestive tract is also an important immune 1126. 17. Putnam, N. H., T. Butts, D. E. Ferrier, R. F. Furlong, U. Hellsten, T. Kawashima, defense in amphioxus and other invertebrates. ALP1 and ALP2, M. Robinson-Rechavi, E. Shoguchi, A. Terry, J. K. Yu, et al. 2008. The am- which are amphioxus apextrin-like proteins and mediate bacterial phioxus genome and the evolution of the chordate karyotype. Nature 453: 1064– 1071. recognition, are most highly expressed in the hepatic cecum (47). 18. Huang, S., X. Wang, Q. Yan, L. Guo, S. Yuan, G. Huang, H. Huang, J. Li, VCBPs of Ciona also bind to bacteria in the gut lumen (48). In M. Dong, S. Chen, and A. Xu. 2011. The evolution and regulation of the mucosal response to a bacterial challenge, the gut epithelial cells of am- immune complexity in the basal chordate amphioxus. J. Immunol. 186: 2042– 2055. phioxus exhibit active phagocytosis (39). The amphioxus hepatic 19. Yu, C., M. Dong, X. Wu, S. Li, S. Huang, J. Su, J. Wei, Y. Shen, C. Mou, X. Xie, cecum, which is connected to the gut, also has antibacterial func- et al. 2005. Genes “waiting” for recruitment by the adaptive immune system: the tions, such as induction of inflammation (49). In addition to its insights from amphioxus. J. Immunol. 174: 3493–3500. 20. Zhang, Y., K. Xu, A. Deng, X. Fu, A. Xu, and X. Liu. 2014. An amphioxus diversified IgVJ domain and potential binding sites (Fig. 1), RAG1-like DNA fragment encodes a functional central domain of vertebrate core RAG1. Proc. Natl. Acad. Sci. USA 111: 397–402. AmpIgVJ-C2 was functionally located in the gut, hepatic cecum, 21. Huang, S., X. Tao, S. Yuan, Y. Zhang, P. Li, H. A. Beilinson, Y. Zhang, W. Yu, and gill. Thus, in addition to other IgVJ-C molecules, AmpIgVJ-C2 P. Pontarotti, H. Escriva, et al. 2016. Discovery of an active RAG transposon may play a defensive role in the amphioxus digestive tract. illuminates the origins of V(D)J recombination. Cell 166: 102–114. The Journal of Immunology 13

22. Zhang, Q. J., Y. Sun, J. Zhong, G. Li, X. M. Lu¨, and Y. Q. Wang. 2007. Con- guidance receptors that exhibit isoform-specific homophilic binding. Cell 118: tinuous culture of two lancelets and production of the second filial generations in 619–633. the laboratory. J. Exp. Zoolog. B Mol. Dev. Evol. 308: 464–472. 37. Marchalonis, J. J., I. Jensen, and S. F. Schluter. 2002. Structural, antigenic and 23. Liu, J., X. Qian, Z. Chen, X. Xu, F. Gao, S. Zhang, R. Zhang, J. Qi, G. F. Gao, evolutionary analyses of immunoglobulins and T cell receptors. J. Mol. Rec- and J. Yan. 2012. Crystal structure of cell adhesion molecule nectin-2/CD112 ognit. 15: 260–271. and its binding to immune receptor DNAM-1/CD226. J. Immunol. 188: 5511– 38. Ely, K. R., J. N. Herron, M. Harker, and A. B. Edmundson. 1989. Three- 5520. dimensional structure of a light chain dimer crystallized in water. Conforma- 24. Schneider, T. R., and G. M. Sheldrick. 2002. Substructure solution with tional flexibility of a molecule in two crystal forms. J. Mol. Biol. 210: 601–615. SHELXD. Acta Crystallogr. D Biol. Crystallogr. 58: 1772–1779. 39. Han, Y., G. Huang, Q. Zhang, S. Yuan, J. Liu, T. Zheng, L. Fan, S. Chen, and 25. McCoy, A. J., R. W. Grosse-Kunstleve, P. D. Adams, M. D. Winn, L. C. Storoni, A. Xu. 2010. The primitive immune system of amphioxus provides insights into and R. J. Read. 2007. Phaser crystallographic software. J. Appl. Cryst. 40: 658– the ancestral structure of the vertebrate immune system. Dev. Comp. Immunol. 674. 34: 791–796. 26. Cowtan, K. D., and P. Main. 1996. Phase combination and cross validation in 40. Liu, Y., X. Li, J. Qi, N. Zhang, and C. Xia. 2016. The structural basis of chicken, iterated density-modification calculations. Acta Crystallogr. D Biol. Crystallogr. swine and bovine CD8aa dimers provides insight into the co-evolution with 52: 43–48. MHC I in endotherm species. Sci. Rep. 6: 24788. 27. Perrakis, A., M. Harkiolaki, K. S. Wilson, and V. S. Lamzin. 2001. ARP/wARP 41. Kasahara, M., T. Suzuki, and L. D. Pasquier. 2004. On the origins of the adaptive and molecular replacement. Acta Crystallogr. D Biol. Crystallogr. 57: 1445– immune system: novel insights from invertebrates and cold-blooded vertebrates. 1450. Trends Immunol. 25: 105–111. 28. Emsley, P., and K. Cowtan. 2004. Coot: model-building tools for molecular 42. Stanfield, R. L., H. Dooley, M. F. Flajnik, and I. A. Wilson. 2004. Crystal graphics. Acta Crystallogr. D Biol. Crystallogr. 60: 2126–2132. structure of a shark single-domain antibody V region in complex with lysozyme. 29. Murshudov, G. N., A. A. Vagin, and E. J. Dodson. 1997. Refinement of mac- Science 305: 1770–1773. romolecular structures by the maximum-likelihood method. Acta Crystallogr. D 43. Dermody, T. S., E. Kirchner, K. M. Guglielmi, and T. Stehle. 2009. Immuno- Biol. Crystallogr. 53: 240–255. globulin superfamily virus receptors and the evolution of adaptive immunity. 30. Adams, P. D., R. W. Grosse-Kunstleve, L. W. Hung, T. R. Ioerger, A. J. McCoy, PLoS Pathog. 5: e1000481. N. W. Moriarty, R. J. Read, J. C. Sacchettini, N. K. Sauter, and T. C. Terwilliger. 44. McFall-Ngai, M. 2007. Adaptive immunity: care for the community. Nature 445: 2002. PHENIX: building new software for automated crystallographic structure 153. determination. Acta Crystallogr. D Biol. Crystallogr. 58: 1948–1954. 45. Morita, C. T., E. M. Beckman, J. F. Bukowski, Y. Tanaka, H. Band, B. R. Bloom, Downloaded from 31. Laskowski, R. A., M. W. MacArthur, D. S. Moss, and J. M. Thornton. 1993. D. E. Golan, and M. B. Brenner. 1995. Direct presentation of nonpeptide prenyl PROCHECK: a program to check the stereochemical quality of protein struc- pyrophosphate antigens to human gamma delta T cells. Immunity 3: 495–507. tures. J. Appl. Cryst. 26: 283–291. 46. Gold, M. C., and D. M. Lewinsohn. 2013. Co-dependents: MR1-restricted 32. Yokoyama, W. M., M. Christensen, G. D. Santos, and D. Miller. 2006. Pro- MAIT cells and their antimicrobial function. Nat. Rev. Microbiol. 11: 14–19. duction of monoclonal antibodies. Curr. Protoc. Immunol. Chapter 2: Unit 2.5. 47. Huang, G., S. Huang, X. Yan, P. Yang, J. Li, W. Xu, L. Zhang, R. Wang, Y. Yu, 33. Campbell, K. S., B. T. Ba¨ckstro¨m, G. Tiefenthaler, and E. Palmer. 1994. CART: S. Yuan, et al. 2014. Two apextrin-like proteins mediate extracellular and in- a conserved antigen receptor transmembrane motif. Semin. Immunol. 6: 393– tracellular bacterial recognition in amphioxus. Proc. Natl. Acad. Sci. USA 111:

410. 13469–13474. http://www.jimmunol.org/ 34. Pancer, Z., W. E. Mayer, J. Klein, and M. D. Cooper. 2004. Prototypic T cell 48. Dishaw, L. J., S. Giacomelli, D. Melillo, I. Zucchetti, R. N. Haire, L. Natale, receptor and CD4-like coreceptor are expressed by lymphocytes in the agnathan N. A. Russo, R. De Santis, G. W. Litman, and M. R. Pinto. 2011. A role for sea lamprey. Proc. Natl. Acad. Sci. USA 101: 13273–13278. variable region-containing chitin-binding proteins (VCBPs) in host gut-bacteria 35. Zhang, S. M., C. M. Adema, T. B. Kepler, and E. S. Loker. 2004. Diversification interactions. Proc. Natl. Acad. Sci. USA 108: 16747–16752. of Ig superfamily genes in an invertebrate. Science 305: 251–254. 49. Wang, Y., and S. Zhang. 2011. Identification and expression of liver-specific 36. Wojtowicz, W. M., J. J. Flanagan, S. S. Millard, S. L. Zipursky, and genes after LPS challenge in amphioxus: the hepatic cecum as liver-like organ J. C. Clemens. 2004. Alternative splicing of Drosophila Dscam generates axon and “pre-hepatic” acute phase response. Funct. Integr. Genomics 11: 111–118. by guest on October 1, 2021