Definition of a third VLR gene in hagfishJianxu Li, Sabyasachi Das, Brantley R. Herrin, Masayuki Hirano, and Max D. Cooper1

Emory Vaccine Center and Department of Pathology and Laboratory Medicine, Emory University, Atlanta, GA 30322

Contributed by Max D. Cooper, August 1, 2013 (sent for review July 3, 2013)

Jawless vertebrates (cyclostomes) have an alternative adaptiveimmune system in which lymphocytes somatically diversify theirvariable lymphocyte receptors (VLR) through recombinatorial useof leucine-rich repeat cassettes during VLR gene assembly. Threetypes of these anticipatory receptors in lampreys (VLRA, VLRB, andVLRC) are expressed by separate lymphocyte lineages. However,only two VLR genes (VLRA and VLRB) have been found in hagfish.Here we have identified a third hagfish VLR, which undergoessomatic assembly to generate sufficient diversity to encode a largerepertoire of anticipatory receptors. Sequence analysis, structuralcomparison, and phylogenetic analysis indicate that the uniquehagfish VLR is the counterpart of lamprey VLRA and the previouslyidentified hagfish “VLRA” is the lamprey VLRC counterpart. Thedemonstration of three orthologous VLR genes in both lampreysand hagfish suggests that this anticipatory receptor system evolvedin a common ancestor of the two cyclostome lineages around480 Mya.

adaptive immunity | molecular evolution | antigen receptor | agnatha

Phylogenetic studies of immunity indicate the emergence oftwo types of recombinatorial adaptive immune systems

(AISs) in vertebrates (1, 2). All of the extant jawed vertebratesgenerate a vast repertoire of Ig-domain–based T- and B-cellantigen receptors primarily by the recombinatorial assembly of IgV-(D)-J gene segments and somatic hypermutation (3). Theextant jawless vertebrates, lampreys and hagfish, instead have analternative AIS that is based on variable lymphocyte receptors(VLRs), the diversity of which is generated through recombi-natorial use of leucine-rich repeat (LRR) cassettes (4–6). Thegerm-line VLR genes are incomplete in that they only containcoding sequences for the leader sequence, incomplete amino-and carboxyl-terminal LRR subunits (LRRNT and LRRCT) andthe stalk region (4, 7, 8). However, each germ-line VLR gene isflanked by hundreds of different LRR-encoding sequences,which can be used as templates to add LRR sequences during theassembly of a mature VLR gene (4, 8–11). This gene conversion-like process is postulated to involve the activation-induced cyti-dine deaminase (AID) orthologs, cytidine deaminases 1 and 2(CDA1 and CDA2) (8, 12). The combinatorial VLR assemblycan generate a vast repertoire of anticipatory receptors compa-rable in diversity to the repertoire of Ig-domain–based antigenreceptors in jawed vertebrates (8, 9).Three VLR genes have been identified in lampreys (VLRA,

VLRB, and VLRC), but only two (VLRA and VLRB) have beenidentified so far in hagfish (4, 7, 8, 13). Lamprey VLRB-expressingcells can respond to immunization by undergoing lymphoblastoidtransformation, clonal expansion, and secretion of their antigen-specific VLRB antibodies (9, 14). VLRA- and VLRC-bearingcells also proliferate in response to antigen stimulation, but do notdifferentiate into antibody secreting cells; instead they maintaincell surface expression of their receptors, while increasing theexpression of proinflammatory cytokines, macrophage migrationinhibitory factor (MIF), and interleukin-17 (IL-17) (12). CDA1-expressing progenitors assemble their VLRA and VLRC genes tobecome VLRA+ and VLRC+ lymphocytes in a thymus-equivalentregion of the gills termed the thymoid (15, 16). Conversely, VLRBassembly coincides with CDA2 expression during VLRB+ lym-phocyte development in hematopoietic tissues (15). The T- and

B-like characteristics of the lamprey lymphocytes imply that jaw-less vertebrates have humoral and cellular arms of adaptive im-munity comparable to those of jawed vertebrates.In the present study, we sought to determine whether or not

hagfish have a third VLR gene. In comparing the sequences ofthe two known hagfish VLRs with those of the three lampreyVLRs, we noticed that the highly variable inserts in the LRRCTmodule of the currently designated hagfish “VLRA” are moresimilar to those of lamprey VLRC than to those of lampreyVLRA (13, 17, 18). This led us to hypothesize that a third hagfishVLR, if present, would be the true counterpart of lampreyVLRA. Through a similarity search against the hagfish database,we identified a fragment of a potential third VLR gene, whichwas in turn used to clone and sequence a previously uncharac-terized VLR gene. Here, we report the characterization of thisthird VLR type in pacific hagfish (Eptatretus stoutii) and itsphylogenetic and structural relationship with previously identi-fied VLRs. Our findings suggest a modified nomenclature for thehagfish VLRs.

ResultsIdentification and Characterization of a third VLR Gene in Hagfish.Based on the hypothesis that, if hagfish have a third VLR, itwould be orthologous to lamprey VLRA, we used the signalpeptides, LRRNT, LRRCT, and the C-terminal regions from 30mature lamprey VLRAs as queries to perform a tBLASTn-basedsimilarity search of the hagfish EST database. A peptide frag-ment encoded in the third open reading frame (ORF) of an ESTclone (GenBank accession no. BAI66885) was found to share75% identity with the 5′LRRCT region of one of the query se-quences (GenBank accession no. ABO21305). The hagfish ESTclone containing this peptide fragment was retrieved and hypo-thetically translated into three ORFs. Whereas a nonproductive

Significance

The jawless vertebrates (hagfish and lampreys) possess an al-ternative adaptive immune system in which variable lympho-cyte receptors (VLRs) constructed of leucine-rich repeats areused to recognize foreign antigens. Three VLR genes have beenidentified in lampreys (VLRA, VLRB, and VLRC), but only two(VLRA and VLRB) have been found in hagfish. Here, we iden-tified and characterized a third hagfish VLR gene. Our analysisindicates that the third hagfish VLR is the ortholog of lampreyVLRA, while the previously identified hagfish “VLRA” is thecounterpart of lamprey VLRC. The demonstration of threeorthologous VLR genes in hagfish and lampreys suggests thatthis anticipatory receptor system evolved in a common ances-tor of the two jawless vertebrate lineages ∼480 Mya.

Author contributions: J.L. and M.D.C. designed research; J.L., S.D., B.R.H., and M.H. per-formed research; J.L., S.D., B.R.H., M.H., and M.D.C. analyzed data; and J.L., S.D., B.R.H.,M.H., and M.D.C. wrote the paper.

The authors declare no conflict of interest.

Data deposition: The sequences reported in this paper have been deposited in theNational Center for Biotechnology Information database (GenBank accession nos.KF314046–KF314110).1To whom correspondence should be addressed. E-mail: mdcoope@emory.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1314540110/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1314540110 PNAS | September 10, 2013 | vol. 110 | no. 37 | 15013–15018

IMMUNOLO

GY

Dow

nloa

ded

by g

uest

on

Dec

embe

r 14

, 202

0

VLR-like sequence in ORF1 lacked a connecting peptide (CP)and intact LRRCT motif, the corresponding region in ORF3could encode a novel CP-5′LRRCT module (Fig. S1). When thisCP-5′LRRCT peptide sequence was blasted in the NationalCenter for Biotechnology Information (NCBI) database, the top50 BLAST hits were all agnathan VLRs; the highest similarity(68% amino acid identity) was with the sea lamprey (Petromyzonmarinus) VLRA clone PmVLRA.D4. Accordingly, a reverse primerwas designed for 5′-RACE extension to clone the upstream re-gion of this gene using a hagfish leukocyte cDNA library. Afterobtaining the 5′-region, a primer against the hypothetical signalpeptide region was used to perform 3′-RACE to clone its down-stream region. We thus obtained a cDNA clone encoding aprotein with the identifying characteristics of currently definedVLRs in jawless vertebrates (Fig. 1). The predicted protein con-sists of a signal peptide (SP) of 21 residues, a 36-residue LRRNTdomain, an 18-residue LRR1 followed by four diverse 24-residue

LRRVs, a 13-residue CP, a 62-residue LRRCT module, and a 77-residue C-terminal domain including a threonine/proline-richstalk region and a hydrophobic tail. Notably, the signal peptideand the C-terminus of this protein have only weak sequencesimilarity to previously identified hagfish VLRs (Fig. 1). Thisunique hagfish VLR was provisionally named the 3rdVLR.

Structure of the Germ-Line 3rdVLR. The germ-line structure of thehagfish 3rdVLR gene was determined by RT-PCR in combinationwith genomic PCR. Erythrocyte DNA was used as template forthe genomic PCR because these cells maintain their VLR genesin germ-line configuration. As is the case for hagfish “VLRA”and VLRB (7), our genomic PCR analysis suggested that thehagfish 3rdVLR is a single-copy gene. The hagfish germ-line3rdVLR is also incomplete, but distinctive from previously iden-tified VLRs (Fig. 2). It consists of two coding regions separatedby a 173-bp intervening sequence (Fig. 2A). The first coding

Es3rdVLR M - - E R W F S R I T M F W M L F L G E V - - - - - - - - Q S Q D W T K - C - - - - - S T G - C T C D G N K M S - - - - V T C - - N G L K N V P N DPm VLRA M - G L L S S S P P P P H L I L I L L - - V L A S S R P G S A S - W - K T C - - - E T V T G - C T C N E G K K E - - - - V D C Q S K D L T S V P P GLc VLRA M - G R L S S S P P T P Y I I L I L L L I V L A S S R P A S A S - W - K T C - - - E K D T F - C T C N E T K K E - - - - V D C Q S K G L Q A V P S GEb "VLRA" M - - - - - M G P V L A A C L L I I L S T A W I S Q A N E - A L - - - - - C - - - K K D G G V C S C N N N K N S - - - - V D C S S K R L T A I P S NEs "VLRA" M - - - - - M G P V L A A C L L I I L S T A W I S Q A N G - A T - - - - - C - - - K K D G G V C T C - - - - N D Q T K N V D C S G K M L T A I P S NPm VLRC M - - - - - - G F V V A L L V L G A W C G S C S - - - - - - A Q - R Q R A C V E A G K S D V - C I C S S A T D S S P E T V D C S S K T L A D V P T GLc VLRC M - - - - - - G F V V A L L V L G A W C G S C S - - - - - - A Q G Q R R A C L A V G K D D I - C T C S N K T D S S P E T V D C S S K K L T A V P T GLp VLRC M - - - - - - G F V V A L L V L G A W C G S C S - - - - - - A Q G R E R A C F A A G K D D L - C T C S N K T E S S P E T V D C S S K K L T T V P T GPm VLRB M - W I K W I A T L V A F G A L V Q S A V - - - - - - - - - A - - - - - - C - - - - P S Q - - C S C - - - - - D Q T T - V D C R N K R F S S V P A GLc VLRB M - W I K W I A T L V A F G A L V Q S A V - - - - - - - - - A - - - - - - C - - - - P A Q - - C S C - - - - - S S T L - V N C Q S R S L A S V P A GEs VLRB M K F A L R G T C V L L A L L L C C R - - - - - - - - N G K A - - - - - - C - - - - P S R - - C S C - - - - - S G T T - V N C R S K S F T S F P S GEb VLRB M K F A L R G T C V L L A L L L C C R - - - - - - - - N G K A - - - - - - C - - - - P S R - - C S C - - - - - S G T E - V Y C N S K G L T S V P T G

Es3rdVLR I Q K D A K K L D L K Y N S L S K L S P T A F H G L N K L T I L Y L R G N K L Q T L P A G V F D E L K N L E T L W L E Q N Q L T S L P P G V F D H LPm VLRA I P T D T E K L D L R Y N A F T Q L S S N A F Q G L T K L T F L N L E D N Q L Q A L T A D I F H P L N E L R T L S F A R N Q V S A L P L G V F D R LLc VLRA I P A D T E K L D L N S N S L A T L S D T A F R G L T K L T W L N L Q Y N A L Q T L P P G V F D Q L R E L K D L Y L G Q N Q L K S L P E R V F D R LEb "VLRA" I P T D T E N L K L G R N K L S S L S A K A F H S L S S L T Y L S L F N N K L K S L P E G V F D Q L V N L T D L R L Y Q N Q L T S L P S G I F D K LEs "VLRA" I P V D T D R L V L Q G N K L S S L P H T A F H G M K E L T Y L G L E G N R L Q T L P P G V F D H L V T L E T L G L S N N Q L K P L P R G V F D K LPm VLRC I P A S T E R L Q L H Y N Q L T S I P V N A F R A L T Q L T Y L N L D N N K L Q S L P A G V F D G L S E L D R L S L Q D N Q L K S L P P G V F D K LLc VLRC I P A N T E K L Q L D F N Q L A N I P A E A F H G L T R L T Y L A L D Y N Q L Q S L P V G V F D Q L N N L N E L R L Q D N Q L T S L P P G V F D S LLp VLRC I P A S T E R L E L Q Y N Q L T S I P D K A F H G L A R L T Y L G L S N N Q L Q S L P V G V F D Q L K N L N E L R L S S N K L K S L P P A V F D S LPm VLRB I P T D R Q N L W L N N N Q I T K L E P G V F D R L A Q L T G L D L S H N Q F T A L P A Q V F D R L V N L Q K L W L N S N K L T A I P A G V F D K LLc VLRB I P S T T Q Y L G L S S N Q I T K L E P G V F D R L A Q L T Q L D L S H N Q F T A L P A R V F D R L V N L Q K L S L S V N Q L Q A L P T G V F N K LEs VLRB I P S R T T V L Y L D G N K L Q S I P S G V F D K L T Q L T R L E L D R N Q L K S L P M G I F D K L T K L T W L E L Y T N Q L Q S L P M G I F D K LEb VLRB I S A S T T Y L N L N T N Q L Q S L P S G V F D K L T Q L T K L Y L H Y N Q L Q S L P S G V F D K L T Q L T I L Y L H N N K L Q S L P S G L F D N L

Es3rdVLR T K L T Y L G L Y E N K L Q S L P H G L F D K L T Q L E T L Y L R N N Q L R S V P D K A F D K L S K L E T I F L T G N D W D C A S C - D I L Y L S QPm VLRA I N L D K L Y L N M N Q L K S L P P R V F D S L S K L T Y L S V G Q N E L Q S V P H G A F D S F G K L E T I T L N T N N W D C S N C - T I L Y L S DLc VLRA T K L T L L D L Y N N Q L Q S V P H G A F D - - - - - - - - - - - - - - - - - - - - - - - - R L T N L Q T I L L Y A N Q W N C S S C - G I L Y L S EEb "VLRA" T K L T D L R L S E N K L Q S L P H G V F D K L T E L K T L Y L F N N Q L K S V P E G A F N S L E K L A L L Q L Q S N P W D C S - C K D I L Y L R DEs "VLRA" T K I T Y L D L Y E N K L Q S L P H G V F D K L T N L K E L W L R N N Q L R R V P E G A F N F L E K L T R L Q L E E N P W D C S - C N D I L Y L R DPm VLRC T K L T R L D L N N N Q L Q S I P A G V F D K L T N L D R L E L S T N Q L Q S V P D G A F D S L T N L E T M N L F N N P W D C A - C S D I I Y L R TLc VLRC T K L T Y L T L S Q N Q L Q S I P A G V F D K L T N L N R L E L S T N Q L Q S V P H G A F D S L V N L E T L H L E L N P W D C A - C S D I I Y L R TLp VLRC T Q L T W L D L R E N Q L Q S I P E K V F D K L T Q L Q Q L Y L N N N Q L P S I P A G V F D S L A N L E T M N L F D N P W D C E - C A S I I Y F V NPm VLRB T E L T Y L N L N T N Q L T A L P E G V F D K L P K L T H L V L H T N Q L T S I P R G A F D N L K S L T H I W L F D N P W D C A - C S D I L Y L S RLc VLRB T Q L T H L S L H T N Q L K S I P R G A F D - - - - - - - - - - - - - - - - - - - - - - - - N L K S L T H I W L F N N P W D C E - C S D I L Y L K NEs VLRB T K L T R L E L Y S N Q L Q S L P M G I F D K L T K L T R L E L Y S N Q L K S V P D G I F D R L T S L Q K I W L H T N P W D C S - C P R I D Y L S REb VLRB S K L K E L Y L Y S N K L Q S L P S G L F D K L T Q L T R L E L Y S N Q L K S V P D G I F D R L T S L Q S I Y L Y S N P W D C T - C P G V D Y L S R

Es3rdVLR W I P V N A E K V K N F L T G Q N D L P D P D G V T C - Q G T M E S V N K I T K E N T F Q A G C P T T T Q Q M A T S P A S T T T T E T M T V - - - -Pm VLRA W I G Q N A N K V K R L A G G E Y - V E E P D S V T C - - S D G K V V R T V T N E T L K Y E - C P F V E S K N F T D G P S S A S E N K K Q S S - - -Lc VLRA W I G A H G P T V K S A D G K - - - H S A P D G V T C - - S D G K V V R T V T N E T L K Y V - C P F V E S K N F T G D S S S E S E N K I Y - - - - -Eb "VLRA" W I D D N K D K V T - - - - - - - - - - G A Q D A A C G D Q Q S K A V L E I T E K D A A S D - C V S P N T T T A I P - - I G T M T P A S V I Y D D IEs "VLRA" W I N D N K D K V T - - - - - - - - - - G A Q D A A C G D Q Q S K A V L E I T E E E A A E D - C V Y P N T T T A I P T T I I T - T L A S S N D D D IPm VLRC F I A K N T D K I S - - - - - - - - - - G I E S A Q C - N G T S T A V K D V N T E L I E D V P C K H E I P T P K M T A S P P N T A T S V F T T - E LLc VLRC F I A K N T D K I S - - - - - - - - - - G M E S A Q C - N G T S T A V K D V N T E K I K N V T C N H V Y P T S K I T A S S P T P A T S I F I K K - LLp VLRC W L K E N P K H D S - - - - - - - - - - G A - S - - C E K P A G T A V K D V K T E P I K N V P C K H V Y P T P K I T A S S P T P A T P I F I - P E LPm VLRB W I S Q H P G V V R K D E A G - Y - P V D P D S A R C - S G T N T P V R A V T E A S T S P S K C P G Y V A T T T T P T T T T - P E F I P E T T T S PLc VLRB W I V Q H A S I V N P E G H G - - - - - G V D N V K C - S G T N T P V R A V T E A S T S P S K C P G Y V A T T T T P T T T T - P E I I P E T T T L PEs VLRB W L N K N S Q K E Q - - - - - - - - - - G - - S A K C - S G S G K P V R S I I - - - - - - - - C P T T T T T T T T T T T T M - P - - - - - T T T T LEb VLRB W L H T N S K K E T - - - - - - - - - - - S T S A K C - S G S G K P V R S I I - - - - - - - - C P T T T T T T T T T T - - M - P - - - - - T T T T L

# # # # # # # # # # # # #

Es3rdVLR - R E K Q T K P R M K I F Q E E L - F S K A S I F F - - - - - - - - - - - C Q V I F T Q H V S L V V I H I L A E I P L L H I V C V V R K P - - -Pm VLRA - P - L A V S A P T - E F W S L L K L S C D V A A F - - - - G S L - S L - C A A H L L L A V T V I - - S L I - S S L V - K M L - - - - P S L S -Lc VLRA - P - A A V S A P T - E F W S L L K L S C D V A T F - - - - G S L - S L - C A A H L L L A V T V I - - S L I - S S L V - K M L - - - - P S L P -Eb "VLRA" - H E I - - K V P Q - E - N - F Q K F L - G Y - Q E P - - D H L P T Q P Q C L M S I S G Y L G L M M - S L M L T S A A I L Y V I H F L K K A - -Es "VLRA" - P E L - - P V P Q - E - N - F Q K F L - G Y - Q E P - - D H L P T Q P Q C L M S I S G Y L G L M M - S L V L T S A A I L Y V I H F L K K A - -Pm VLRC N S - - T T Y - P - - - - N - A T H E H - T - - - - - - - D V - - - - - - C N M - P F V S H I C L L F C N L F S T C S L C F I I - - K P L H R YLc VLRC N S - - T T N L - - - - - N - A I H E H R T H - T - - - - D V - - - - - - C N M - P F V S H M C L L F C N L F S T C S L C F I I - - K P L H R YLp VLRC N S - - T T N L - - - - - N - A I H E H R T H - T - - - - D V - - - - - - C N M - P F V T H M C L L F C N L F S T C S L C F I I - - K P L H R YPm VLRB Q P V I T T Q K P K P L W N - F - N - - C T S I Q E R K N D G G D - - - - C G K - P A C T T L L N C A N F L S C L C S T C A L C - - R K - - R -Lc VLRB Q P V I T T Q K P R S L M N - F - N - - C S S I Q E R K N D G G D - - - - C G K - P A C T T L L N C A N F L S C L C S T C A L C - - K K - - R -Es VLRB - P - - T T T K M S M V K V P L V P P - - E - A F G R V M N A - - - - - - C A Y F P S Y I F L H L V H G L A A V P L A Y L V - C H A - - S Q L LEb VLRB - P - - T T T K M S M V K V P L V P P - - E - A F G R V M N A - - - - - - C A Y F P S Y I F L H L V H G L A A V P L V Y L I - C H A - - S Q L L

> SP < > LRRNT

< > LRR1 < > LRRV1 < > LRRV2 < >

LRRV3 < > LRRV4 < > CP < >

LRRCT < >

C-terminus <

Fig. 1. Sequence comparison between a productive hagfish 3rdVLR and representatives of other types of VLRs. Deduced amino acid sequences of VLRs from Pacifichagfish (Es, Eptatretus stoutii), inshore hagfish (Eb, Eptatretus burgeri), sea lamprey (Pm, Petromyzon marinus), Arctic lampery (Lc, Lethenteron camtschaticum), andEuropean brook lamprey (Lp, Lampetra planeri) are shown in the alignment. Conserved cysteines are highlighted in light green and other conserved residues inyellow, whereas “−” indicates a gap and “#” represents the highly variable inserts in LRRCT. The fifth LRRVmodule of LpVLRC is not shown to simplify the alignmentfor illustrative purposes. GenBank accession numbers of representative VLR sequences: Es3rdVLR, KF314050; PmVLRA, ABO27114; LcVLRA, BAJ14924; Eb“VLRA”,ABB59067; Es“VLRA”, ABB59097; PmVLRC, KC244079; LcVLRC, BAJ14926; LpVLRC, AGD98752; PmVLRB, AAT70348; LcVLRB, BAJ14925; EsVLRB, ABB59059; andEbVLRB, BAI66964.

15014 | www.pnas.org/cgi/doi/10.1073/pnas.1314540110 Li et al.

Dow

nloa

ded

by g

uest

on

Dec

embe

r 14

, 202

0

region encodes the signal peptide, LRRNT, and LRR1 modules,wheras the second coding region encodes the 3′-end of LRRCTand the C-terminus domain (Fig. 2B).

Hagfish 3rdVLR Diversity. To evaluate the diversity of hagfish3rdVLRs, 64 mature cDNA clones from two Pacific hagfish weresequenced. Sequence analysis indicated that none of these clonescoded for an identical protein sequence (Fig. S2). The 3rdVLRsequences contain three to six 24-residue LRRV modules (av-erage, four), which is very similar to the hagfish “VLRA” modulecomposition (average, four; range from two to six) but largerthan the average module composition of hagfish VLRB (average,2.8; range from one to six). Interestingly, this relationship is alsotrue for the three lamprey VLRs (Fig. S3). Of the 256 LRRVmodules encoded by the 64 hagfish 3rdVLR cDNA clones, 94%(241/256) had unique sequences. Similarly, 42% (27/64) of theLRR1 modules and 45% (29/64) of the CP modules had uniquesequences. Of note, even though the germ-line 3rdVLR encodesa complete LRR1 module, this sequence can be replaced bysequences of divergent LRR1-coding genomic donor cassettes.Like lamprey VLRA and VLRB and hagfish VLRB (4, 7–9), theLRRCT region of the hagfish 3rdVLRs displays a high level ofsequence diversity (72% unique LRRCTs, 46/64), due to use ofunique 5′-LRRCT donor cassettes during the assembly process.The sequence diversity of the 3rdVLR of hagfish thus appearscomparable to that of previously known VLRs.

Hagfish 3rdVLR Is a Lamprey VLRA Ortholog. To examine the evo-lutionary relationship of hagfish 3rdVLR with the known VLRsin jawless vertebrates, we conducted phylogenetic analyses us-ing the maximum likelihood (ML) and neighbor-joining (NJ)methods (19, 20). The bootstrap support for the hagfish “VLRA”

and lamprey VLRC-containing cluster was low (64%) in the MLtree (Fig. 3A) and below 60% in the NJ phylogenetic analysis

(Fig. 3B). However, the hagfish 3rdVLR clustered with lampreyVLRA with 90% bootstrap support in the ML tree, indicatingan orthologous relationship (Fig. 3A), which is also supportedby their clustering together with high bootstrap support in theNJ phylogenetic tree (Fig. 3B). Similarly, hagfish VLRB clus-tered with lamprey VLRB with 99% bootstrap support inthe ML tree (Fig. 3A) and 98% bootstrap support for the NJtree (Fig. 3B).

Hagfish 3rdVLR and Lamprey VLRA Share Common LRRCT Features.Allof the VLRs contain four cysteines that form two sets of disulfidebridges in both the LRRNT and LRRCT. These serve to cap thehydrophobic core of the solenoid structure of VLRs (21). Withthe exception of lamprey VLRA, however, the LRRCT of pre-viously defined VLRs belongs to the class 1 according to theC-terminal cysteine-containing flanking domain classification (22)in that they share a characteristic four-cysteine motif C1-X-C2-Xm-C3-Xn-C4 (where X stands for any amino acid other thancysteine and m and n stand for variable numbers) (Fig. S4). Theexception is LRRCT of lamprey VLRA, which contains a dis-tinctive four cysteine motif in which an extra serine or glycine isinserted between the first and second cysteines, C1-(S/G)-X-C2-Xm-C3-Xn-C4. Interestingly, the four-cysteine motif in the LRRCTof the hagfish 3rdVLR (C1-X-S-C2-Xm-C3-Xn-C4) matches with

AT GGAACGAT GGT T CAGCAGAAT AACCAT GT T T T GGAT GCT CT T CCT GGGT GAAGT CCAGT CT CAGGACT GGACAAAAT GCT CAACAGGCT GCACGT GT GAT GGAAACAAM E R W F S R I T M F W M L F L G E V Q S Q D W T K C S T G C T C D G N K

GAT GAGT GT GACCT GCAAT GGCCT GAAGAACGT ACCAAACGAT AT T CAAAAGGAT GCCAAGAAACT CGACCT T AAGT ACAACT CACT GAGCAAGCT CGACCCCACGGCATM S V T C N G L K N V P N D I Q K D A K K L D L K Y N S L S K L D P T A

T T CACGGCCT GAGT GCAAAGGT GT GAAGAAGT GCACAGGACT AT CGAT GCAT T CGT T T T GT T GCAGAGT GGGAT GGT GAGGGT AT AAT T AT GAAGT GT GAAGT GT GT GAGF H

AGGAAAAT AAAGAT GAGAGGAT AAT GAGGT T T AACAT CT CAT T T AAAACACT GAAAACACCCCCCAAAACT GAT CCAGACGGT GT AACCT GCCAAGGT ACGAT GGAAT CTT D P D G V T C Q G T M E S

GT AAAT AAAAT T ACCAAAGAGAACACAT T T CAAGCT GGAT GCCCAACAACAACT CAACAGAT GGCAACCT CGCCAGCT T CAACAACAACAACT GAGACCAT GACAGT GCGV N K I T K E N T F Q A G C P T T T Q Q M A T S P A S T T T T E T M T V R

AGAAAAACAGACCAAGCCAAGAAT GAAAAT AT T T CAGGAAGAGCT T T T CT CAAAGGCAT CGAT AT T T T T T T GCCAGGT T AT T T T CACT CAGCAT GT CT CT CT T GT GGT GAE K Q T K P R M K I F Q E E L F S K A S I F F C Q V I F T Q H V S L V V

T T CACAT T T T AGCGGAGAT T CCGCT T T T GCACAT T GT GT GT GT T GT T CGCAAACCAT AAI H I L A E I P L L H I V C V V R K P *

> < >

< >

< >

< >

< >

<

Signal peptide LRRNT

LRR1

Intervening sequence

3'LRRCT

3rdVLR-SF

C-terminus

3rdVLRstop-R

Lamprey

VLRA

VLRB

VLRC

Hagfish

C-terminus3’LRRCTSP

LRRNTSP C-terminusCT3’LRRSP

LRRNT C-terminus3’LRRCTSP

C-terminus3’LRRCT

5’LRRCT

5’LRRNTSP

CP LRRCTLRRNT C-terminusSPSP“VLRA”

VLRB

LRRNTSP C-terminusCT3’LRR3rdVLR LRR1

A

B

Fig. 2. Germ-line gene of Pacific hagfish 3rdVLR and its comparison withpreviously described agnathan VLR genes. (A) Germ-line gene sequence ofthe hagfish 3rdVLR (GenBank accession no. KF314110). The gene (719nucleotides in length) possesses two coding regions separated by an in-tervening sequence. Arrowed lines indicate the forward and reverse primerregions used for 3′-RACE and cDNA cloning. (B) Schematic depiction of thehagfish and lamprey VLR genes (not drawn to scale). Note the distinctivestructure of the hagfish 3rdVLR germ-line gene. E

s 3rd

VLR

Es

3rdV

LRP

m V

LRA

Es

VLR

BP

m V

LRA

Pm

VLR

CE

s “V

LRA

”P

m V

LRB

Pm

VLR

CE

s “V

LRA

”E

s V

LRB

Pm

VLR

B

90

100

99

64

100

100

99

99

92

98

99

100

81

100

99

100

100

A B

Fig. 3. Phylogenetic relationship of Pacific hagfish 3rdVLR and other agna-than VLR genes. The phylogenetic tree is constructed by the (A) maximumlikelihood and (B) neighbor-joining methods using amino acid sequences ofthe diversity region that could be reliably aligned (LRRNT, LRR1, terminalLRRV, CP, and LRRCT). The tree is condensed at the 60% bootstrap value level.The bootstrap supports for interior branches are shown. GenBank accessionnumbers. of VLR sequences are: hagfish 3rdVLR, KF314046–KF314055; hagfish“VLRA”, ABB59089–ABB59098; hagfish VLRB, ABB59051–ABB59060; lampreyVLRA, ABO21293–ABO21302; lamprey VLRC, KC244100–KC244109; and lam-prey VLRB, ABO15209–ABO15218.

Li et al. PNAS | September 10, 2013 | vol. 110 | no. 37 | 15015

IMMUNOLO

GY

Dow

nloa

ded

by g

uest

on

Dec

embe

r 14

, 202

0

lamprey VLRA (Fig. S4) in that the first and second cysteines areseparated by two amino acids instead of the single amino acidspacing found in other VLRs.In the lamprey VLRAs and VLRBs and in hagfish VLRBs, the

LRRCT regions contain a highly variable insert that can forma protruding loop. The importance of this loop in antigen rec-ognition has been elucidated by several solved VLR-antigenstructures (18, 23–25). However, the variable inserts in lampreyVLRCs (median length, 2 residues) and hagfish “VLRA”s(median length, 3 residues) are too short to form this protrusion(Fig. S5) (13, 17). Our 3D structure modeling predicts that thehagfish 3rdVLR adopts a solenoid structure like the other VLRsand has the potential to form a protruding loop at its LRRCTportion (Fig. 4) due to its relatively large LRRCT insert (medianlength, 10 residues). Analyses of the inserts indicate that the3rdVLR insert lengths range from 9 to 13 residues (Fig. S5),which is comparable to the insert lengths in lamprey VLRAs(from 10 to 13 residues) (8, 24).

Expression Patterns of the Three Hagfish VLRs. We used qPCR toexamine the transcript expression profiles for hagfish 3rdVLR,“VLRA”, and VLRB in different tissues (Fig. S6). Hagfish3rdVLR and the previously identified “VLRA” have similar ex-pression patterns, wherein their transcripts were most abun-dant in blood leukocytes, followed by liver, intestine, and skin;only trace levels were found in kidneys and gills. The expres-sion levels of hagfish VLRB were generally higher, especially inliver and intestine; expression was also detected in blood andskin samples.

DiscussionThis study identifies a third hagfish VLR gene whose germ-lineconfiguration and sequence are unique. Our analysis of cDNAclones indicates that the sequence diversity of the newly iden-tified hagfish VLRs is comparable to that observed for otherVLRs (8, 9). Together with the previously described “VLRA”and VLRB genes, the third hagfish VLR is transcribed by cells inthe blood, liver, intestine, and skin, suggesting a potential role inimmune defense.Previous studies have shown that the hagfish VLRB gene is

orthologous to the lamprey VLRB gene (7). Our analysis of thecysteine configuration in the LRRCT motif, length variations ofthe LRRCT inserts, and phylogenetic relationships of the thirdhagfish VLR indicate that it is orthologous to lamprey VLRA,whereas the previously designated hagfish “VLRA” is moreclosely related to lamprey VLRC. Based on our comparativeanalysis of the three types of cyclostome VLR genes, we proposea modified nomenclature for the hagfish VLRs in which theunique third VLR in hagfish is the true lamprey VLRA coun-terpart, and the previously identified “VLRA” in hagfish isorthologous to the lamprey VLRC.Structural comparison of the three VLRs in lamprey and hag-

fish also support the revised nomenclature proposal. VLRs useamino acid variations on their concave surfaces, which are com-posed of the LRRNT, LRR1, LRRVs, and CP motifs, along withthe highly variable LRRCT loops to bind antigens (18). Togetherwith the results of previous studies, our analysis reveals structuraldifferences among the three VLR isotypes, which imply special-ization in terms of ligand recognition (18, 23–25). The VLRAsand VLRCs in lampreys and hagfish have a higher averagenumber of LRRV modules than do the VLRBs, thus indicatingthey have a larger concave surface for antigen binding. Anotherstructural difference in the three VLR types relates to the highlyvariable LRRCT inserts. The VLRAs and VLRBs typically havea protruding loop because of their relatively large LRRCTinserts, whereas the LRRCT inserts in VLRCs are too short toform this loop (refs. 13 and 17 and this study). Furthermore, thelength variation of LRRCT inserts in hagfish and lampreyVLRAs is less pronounced than in the VLRBs. These structuraldifferences may suggest different antigen-binding modes for thethree VLR isotypes. The single example in which a structuralcomparison was conducted of hen egg white lysozyme (HEL)binding by a VLRA.R2.1 and by a VLRB.2D favors this in-terpretation (24), although further analysis of the antigen-bindingmodes of the different VLRs is clearly needed.Our recent studies indicate that lampreys have two primordial

T-like lineages of lymphocytes that express VLRA and VLRC,respectively, and a prototypic B-like lineage of lymphocytes,much like the αβ T, γδ T, and B cells in jawed vertebrates. In

Hagfish

“VLRA”

VLRB

3rdVLR

VLRA

VLRB

VLRC

Lamprey

Fig. 4. Comparison of the predicted 3D structure of hagfish 3rdVLR (Gen-Bank accession no. KF314090) and other agnathan VLRs. The crystal struc-tures of hagfish “VLRA” (PDB ID: 2O6Q), hagfish VLRB (PDB ID: 2O6S),lamprey VLRA (PDB ID: 3M18), and lamprey VLRB (PDB ID: 3E6J) were re-trieved from Protein Data Bank (PDB). The 3D structure of lamprey VLRC(GenBank accession no. KC244064) is also predicted for comparison. β-Sheetand α-helix structures are shown in blue and green colors, respectively. Loopslocated in the LRRCT portion are indicated in red. Note that the hagfish3rdVLR, lamprey VLRA, hagfish VLRB, and lamprey VLRB all possess a pro-truding loop in the LRRCT region, whereas hagfish “VLRA” and lampreyVLRC do not.

Common ancestor

Hagfish Lamprey

VLRA VLRC VLRB

VLRA VLRC VLRBVLRA(3rd VLR)

VLRC(“VLRA”) VLRB

Fig. 5. Evolutionary scenario of VLR genes in the two cyclostome lineages.The revised nomenclature for hagfish “VLRA” and VLRC is indicated.

15016 | www.pnas.org/cgi/doi/10.1073/pnas.1314540110 Li et al.

Dow

nloa

ded

by g

uest

on

Dec

embe

r 14

, 202

0

some ways the lamprey VLRA- and VLRC-expressing cells re-semble the αβ and γδ T cells of jawed vertebrates, whereas theVLRB-bearing cells are more morphologically and functionallysimilar to B cells (12, 14, 16). Hagfish VLRC (“VLRA” in the oldnomenclature) and VLRB are also expressed by two distinctlymphocyte populations (26), and the hagfish VLRB protein issecreted (27). The future characterization of the cells that ex-press the three types of hagfish VLRs described here promises toprovide deeper insight into the evolution of this tripartitelymphocytic differentiation pattern. Our findings indicate thatthe three VLR isotypes existed in the common ancestor oflampreys and hagfish (Fig. 5).The relatively primitive anatomy of hagfish compared with the

lampreys initially led to their classification as more basal jawlessvertebrates (28). However, molecular sequence data suggestsa monophyletic relationship between hagfish and lampreys (29)and recent studies of hagfish embryos suggest that their mor-phological divergence may reflect a loss of features as cyclo-stomes divergently evolved under different selective pressures(30). Whereas the presence of three related VLR types in thesecyclostome representatives suggests that a common ancestor ofthe hagfish and lamprey lineages possessed three VLR-basedanticipatory receptors, many differences can be anticipated be-tween the immune systems of hagfish and lampreys, given the>400 My of independent evolution of the two lineages (31).

Materials and MethodsAnimals. Pacific hagfish E. stoutii (30–60 cm long) were purchased fromMarinus and maintained at 14–17 °C in artificial sea water (Oceanic Systems).All experiments were approved by the Institutional Animal Care and UseCommittee at Emory University.

Data Mining and Similarity Search. The signal peptides, LRRNT, LRRCT, and 3′-terminus portions from 30 mature lamprey VLRAs (GenBank accession nos.ABO21280–ABO21309 and ABO27114) were used as queries for tBLASTnsearch against the hagfish EST database (http://transcriptome.cdb.riken.go.jp/vtcap/blast/blast.html) (32). The correct ORF was identified by alignment ofthe query sequence and the hypothetical translation of the hit sequence. Thesignal peptide region was determined using SignalP (33). The sequenceconservation was assessed by WebLogo (http://weblogo.berkeley.edu/logo.cgi) (34).

cDNA Library Construction and Cloning. Hagfish were anesthetized in MS-222(Sigma) before collecting blood from the tail sinus (7). Blood samples wereloaded onto a 10% (wt/vol) OptiPrep gradient (Axis-Shield) and centrifuged at800 × g for 15 min to obtain white blood cells in the buffy coat. Total RNA

extracted from buffy coat cells using the RNeasy kit (Qiagen) was used toconstruct a 5′- and 3′-RACE cDNA library (Clontech). Full-length hagfish3rdVLR was cloned from the 5′- and 3′-RACE cDNA library by a two-step PCR.The first step was to clone the 5′-region of the hagfish 3rdVLR by usinga 3rdVLR-R1 primer (Fig. S1 and Table S1) against the partial sequenceidentified from the hagfish EST database in combination with the 5′-RACEadaptor primer included in the kit. In step two, the 3′-RACE PCR was per-formed to clone the 3′-region of hagfish 3rdVLR. The 3′-RACE adaptor primertogether with a 3rdVLR-SF primer (forward primer against the signal peptideregion of hagfish 3rdVLR as indicated in Fig. 2A and Table S1) was used in thisstep. The germline and mature hagfish 3rdVLRs were cloned using the pri-mers 3rdVLR-SF and 3rdVLRstop-R (reverse primer according to the positionof stop codon as indicated in Fig. 2A and Table S1).

Phylogenetic Analysis. Sequences were aligned by using the CLUSTALWprogram (35) with adjustment by manual correction where necessary. NJ (20)and ML (19) trees were constructed using MEGA software 5.0 with thepairwise deletion option (36). The evolutionary distances were computedusing the Jones–Taylor–Thornton (JTT) matrix-based method (37). Bootstrapvalues were generated from 1,000 replicates.

Three-Dimensional Structure Modeling. The crystal structure files of lampreyand hagfish “VLRA”, VLRB were extracted from Protein Data Bank (PDB).Three-dimensional models of lamprey VLRC and hagfish 3rdVLR were pre-dicted through PHYRE2 Protein Fold Recognition Server (21) and visualizedby Pymol software (http://www.pymol.org/).

VLR Transcription Analysis. Expression of 3rdVLR, “VLRA”, and VLRB was ex-amined by quantitative real-time PCR (qPCR) analysis of mRNA from cells invarious tissues using RNeasy kit with on-column DNA digestion by DNase I(Qiagen). First-strand cDNA was synthesized with random hexamer primersand SuperScript III (Invitrogen). qPCR was conducted with SYBR Green ona 7900HT ABI Prism detection system (Applied Biosystems). Cycling con-ditions were 1 cycle at 50 °C for 2 min, 1 cycle at 95 °C for 10 min, followedby 40 cycles of denaturation at 95 °C for 15 s and annealing/extension at 60 °Cfor 1 min. The values for the VLR genes were normalized to the expression ofGAPDH. Primers used in this experiment are listed in Table S1.

Statistical Analysis. Statistical analyses were performed using GraphPad Prismsoftware 5.0. The Kruskal–Wallis with post hoc Dunn’s tests was used toevaluate the LRRV module difference between different VLRs. Statisticalsignificance was set up to P < 0.05.

ACKNOWLEDGMENTS. We thank Baohua Wu and Yueju Su for theirassistance in statistical analysis. J.L., S.D., B.R.H., M.H. and M.D.C. weresupported by National Institutes of Health Grants R01 AI072435 and R01GM100151 and the Georgia Research Alliance.

1. Litman GW, Anderson MK, Rast JP (1999) Evolution of antigen binding receptors.Annu Rev Immunol 17:109–147.

2. Hirano M, Das S, Guo P, Cooper MD (2011) The evolution of adaptive immunity invertebrates. Adv Immunol 109:125–157.

3. Dudley DD, Chaudhuri J, Bassing CH, Alt FW (2005) Mechanism and control of V(D)Jrecombination versus class switch recombination: Similarities and differences. AdvImmunol 86:43–112.

4. Pancer Z, et al. (2004) Somatic diversification of variable lymphocyte receptors in theagnathan sea lamprey. Nature 430(6996):174–180.

5. Herrin BR, Cooper MD (2010) Alternative adaptive immunity in jawless vertebrates.J Immunol 185(3):1367–1374.

6. Boehm T, et al. (2012) VLR-based adaptive immunity. Annu Rev Immunol 30:203–220.7. Pancer Z, et al. (2005) Variable lymphocyte receptors in hagfish. Proc Natl Acad Sci

USA 102(26):9224–9229.8. Rogozin IB, et al. (2007) Evolution and diversification of lamprey antigen receptors:

Evidence for involvement of an AID-APOBEC family cytosine deaminase. Nat Immunol8(6):647–656.

9. Alder MN, et al. (2005) Diversity and function of adaptive immune receptors ina jawless vertebrate. Science 310(5756):1970–1973.

10. Nagawa F, et al. (2007) Antigen-receptor genes of the agnathan lamprey are as-sembled by a process involving copy choice. Nat Immunol 8(2):206–213.

11. Das S, et al. (2013) Organization of lamprey variable lymphocyte receptor C locus andrepertoire development. Proc Natl Acad Sci USA 110(15):6043–6048.

12. Guo P, et al. (2009) Dual nature of the adaptive immune system in lampreys. Nature459(7248):796–801.

13. Kasamatsu J, et al. (2010) Identification of a third variable lymphocyte receptor in thelamprey. Proc Natl Acad Sci USA 107(32):14304–14308.

14. Alder MN, et al. (2008) Antibody responses of variable lymphocyte receptors in thelamprey. Nat Immunol 9(3):319–327.

15. Bajoghli B, et al. (2011) A thymus candidate in lampreys. Nature 470(7332):90–94.16. Hirano M, et al. (2013) Evolutionary implications of a third lymphocyte lineage in

lampreys. Nature, 10.1038/nature12467.17. Kim HM, et al. (2007) Structural diversity of the hagfish variable lymphocyte re-

ceptors. J Biol Chem 282(9):6726–6732.18. Velikovsky CA, et al. (2009) Structure of a lamprey variable lymphocyte receptor in

complex with a protein antigen. Nat Struct Mol Biol 16(7):725–730.19. Felsenstein J (1981) Evolutionary trees from DNA sequences: A maximum likelihood

approach. J Mol Evol 17(6):368–376.20. Saitou N, Nei M (1987) The neighbor-joiningmethod: A newmethod for reconstructing

phylogenetic trees. Mol Biol Evol 4(4):406–425.21. Kelley LA, Sternberg MJ (2009) Protein structure prediction on the Web: A case study

using the Phyre server. Nat Protoc 4(3):363–371.22. Kajava AV (1998) Structural diversity of leucine-rich repeat proteins. J Mol Biol 277(3):

519–527.23. Han BW, Herrin BR, Cooper MD, Wilson IA (2008) Antigen recognition by variable

lymphocyte receptors. Science 321(5897):1834–1837.24. Deng L, et al. (2010) A structural basis for antigen recognition by the T cell-like

lymphocytes of sea lamprey. Proc Natl Acad Sci USA 107(30):13408–13413.25. Kirchdoerfer RN, et al. (2012) Variable lymphocyte receptor recognition of the im-

munodominant glycoprotein of Bacillus anthracis spores. Structure 20(3):479–486.26. Kishishita N, et al. (2010) Regulation of antigen-receptor gene assembly in hagfish.

EMBO Rep 11(2):126–132.27. Takaba H, et al. (2013) A major allogenic leukocyte antigen in the agnathan hagfish.

Sci Rep 3:1716.

Li et al. PNAS | September 10, 2013 | vol. 110 | no. 37 | 15017

IMMUNOLO

GY

Dow

nloa

ded

by g

uest

on

Dec

embe

r 14

, 202

0

28. Janvier P (1996) Early Vertebrates (Oxford Univ Press, Oxford).29. Heimberg AM, Cowper-Sal-lari R, Sémon M, Donoghue PC, Peterson KJ (2010)

microRNAs reveal the interrelationships of hagfish, lampreys, and gnathostomesand the nature of the ancestral vertebrate. Proc Natl Acad Sci USA 107(45):19379–19383.

30. Ota KG, Kuraku S, Kuratani S (2007) Hagfish embryology with reference to theevolution of the neural crest. Nature 446(7136):672–675.

31. Kuraku S, Ota KG, Kuratani S (2009) Jawless fishes (Cyclostomata). The Timetree ofLife, eds Hedges SB, Kumar S (Oxford Univ Press, New York), pp 317–319.

32. Takechi M, et al. (2011) Overview of the transcriptome profiles identified in hagfish,shark, and bichir: Current issues arising from some nonmodel vertebrate taxa. J ExpZoolog B Mol Dev Evol 316(7):526–546.

33. Bendtsen JD, Nielsen H, von Heijne G, Brunak S (2004) Improved prediction of signalpeptides: SignalP 3.0. J Mol Biol 340(4):783–795.

34. Crooks GE, Hon G, Chandonia JM, Brenner SE (2004) WebLogo: A sequence logogenerator. Genome Res 14(6):1188–1190.

35. Thompson JD, Higgins DG, Gibson TJ (1994) CLUSTAL W: Improving the sensitivity ofprogressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res 22(22):4673–4680.

36. Tamura K, et al. (2011) MEGA5: Molecular evolutionary genetics analysis usingmaximum likelihood, evolutionary distance, and maximum parsimony methods. MolBiol Evol 28(10):2731–2739.

37. Jones DT, Taylor WR, Thornton JM (1992) The rapid generation of mutation datamatrices from protein sequences. Comput Appl Biosci 8(3):275–282.

15018 | www.pnas.org/cgi/doi/10.1073/pnas.1314540110 Li et al.

Dow

nloa

ded

by g

uest

on

Dec

embe

r 14

, 202

0