Neurokinin Bs and neurokinin B receptors in zebrafish-potential role in controlling fish reproductionJakob Birana, Ori Palevitcha, Shifra Ben-Dorb, and Berta Levavi-Sivana,1

aDepartment of Animal Sciences, The Robert H. Smith Faculty of Agriculture, Food, and Environment, Hebrew University of Jerusalem, Rehovot 76100, Israel;and bDepartment of Biological Services, Weizmann Institute of Science, Rehovot 76100, Israel

Edited by John E. Halver, University of Washington, Seattle, WA, and approved April 27, 2012 (received for review December 2, 2011)

The endocrine regulation of vertebrate reproduction is achieved bythe coordinated actions of several peptide neurohormones, tachy-kinin among them. To study the evolutionary conservation andphysiological functions of neurokinin B (NKB), we identified tachy-kinin (tac) and tac receptor (NKBR) genes from many fish species,and cloned two cDNA forms from zebrafish. Phylogenetic analysisshowed that piscine Tac3s and mammalian neurokinin genes arisefrom one lineage. High identity was found among different fishspecies in the region encoding the NKB; all shared the common C-terminal sequence. Although the piscine Tac3 gene encodes for twoputative tachykinin peptides, the mammalian ortholog encodes foronly one. The second fish putative peptide, referred to as neuro-kinin F (NKF), is unique and found to be conserved among the fishspecieswhen tested in silico. tac3awas expressed asymmetrically inthe habenula of embryos, whereas in adults zebrafish tac3a-ex-pressing neurons were localized in specific brain nuclei that areknown to be involved in reproduction. Zebrafish tac3amRNA levelsgradually increased during thefirst fewweeks of life and peaked atpubescence. Estrogen treatment of prepubertal fish elicited in-creases in tac3a, kiss1, kiss2, and kiss1ra expression. The syntheticzebrafish peptides (NKBa, NKBb, and NKF) activated Tac3 receptorsvia both PKC/Ca2+ and PKA/cAMP signal-transduction pathwaysin vitro. Moreover, a single intraperitoneal injection of NKBa andNKF significantly increased leuteinizing hormone levels in maturefemale zebrafish. These results suggest that the NKB/NKBR systemmay participate in neuroendocrine control of fish reproduction.

gonadotropin-releasing hormone | kisspeptin | teleost | gonadotropin

Reproduction is a highly integrated and complex function thatrequires synchronized production of gametes by both sexes at

an optimum time for offspring survival. Fish show an enormousvariety of reproductive strategies (1), and were recently chosen asmodels for the study of growth, metabolism, and human diseases.The hypothalamic regulation of gonadotropin secretion in fish isdifferent from that of mammals, from both endocrinal and ana-tomical aspects. In teleosts, the pituitary is innervated directly byneurons projecting to the vicinity of the pituitary gonadotrophs(2). Among the neuropeptides released by these nerve endings aregonadotrophin-releasing hormones (GnRHs) and dopamine,which act as stimulatory and inhibitory factors on the release ofluteinizing hormone (LH) and follicle-stimulating hormone (3).However, new actors have recently entered the field of re-productive physiology: kisspeptins, neurokinin, and dynorphinhave all been implicated in controlling GnRH (4).Topaloglu et al. (5) found that humans bearing loss-of-func-

tion mutations of the genes encoding either neurokinin B (NKB)or its cognate receptor, neurokinin receptor 3 (NKBR, Tac3r)displayed hypogonadotropic hypogonadism; this seminal reportimplicated NKB signaling as an essential factor in the onset ofpuberty and control of gonadotropin secretion in mammals.Recent studies provided evidence that, in mammals, a group ofneurons in the hypothalamic arcuate nucleus (ARC) are steroid-responsive and coexpress NKB, kisspeptin, dynorphin, NKBRand estrogen receptor α(6). Compelling evidence indicates thatthese neurons function in the hypothalamic circuitry regulatingGnRH secretion. The main objective of the present study, using

zebrafish (Danio rerio) as a model, was to examine the in-volvement of NKB in fish reproduction.NKB is a member of the tachykinin (TK) family of peptides.

TKs are characterized by a common carboxyl-terminal amino acidsequence of FXGLM-NH2 (where X is a hydrophobic residue),and include substance P, neurokinin A (NKA) and NKB, as wellas neuropeptide K, neuropeptide-γ, and hemokinin-1 (7). NKB isthe only TK synthesized from the preprotachykinin-B gene (8),which is currently designated as TAC3 in mammals, except forrodents, where it was named Tac2. Because there are differentnames for the gene encoding NKB in different species (TAC3 orTac2), in this article we will refer to mRNA products of this geneas tac3 mRNA and to the peptides as NKB. The receptor thatbinds NKB, which is termed NKBR in humans, will be termedtac3r at the mRNA level and Tac3r at the protein level.Until now, NKB was not cloned from any fish species, nor was

the NKB/NKBR system shown to be involved in reproduction orpuberty. We report here the identification of previously un-identified fish NKB/NKBR genes and their possible involvementin the control of reproduction.

Results and DiscussionCloning Two Types of tac3 and tac3r and Their Phylogenetic Analysis.As the first step toward examining the involvement of the NKB/NKBRs (tac3r) in the control of reproduction in fish, we reporthere the identification of the full-length tac3a and tac3b cDNAfrom zebrafish brain using real-time PCR with specific primers(Table S1). Tac3a contains the decapeptide sequence EMH-DIFVGLM (Fig. S1A) (accession no. JN392856), whereas tac3bcontains a 24-aa peptide (STGINREAHLPFRPNMNDIFVGLL)(Fig. S1B) (accession no. JN392857), both with the TK signaturemotif (FXGLM-NH2) flanked by potential dibasic cleavage sitesand an adjacent glycine at the C terminus for amidation (9). Typ-ically, following prohormone convertase action, a carboxypepti-dase removes theC-terminal dibasic residues, and a peptidylglycinea-amidating enzyme converts the exposed glycine into a C-terminalamide (10). At the protein level, the resulting zfNKBa (Tac3a)hormone precursor displayed around 25% identity with human ormouse TAC3, 55% with putative salmon Tac3a, and 52% withmedaka Tac3 identified in this study (accession nos. BK008102 andBK008114, respectively). zfNKBb (Tac3b) showed only around18% identity with human andmouse TAC3, 40% and 36% identitywith salmon and medaka Tac3b, respectively. The zebrafish tac3sshared only a 36% identity (Fig. S1 and Table S2).In our search for the identification of Tac3 sequences con-

taining the NKB peptide sequence in fish mRNAs and ESTsknown to date, we encountered 30 previously unidentified

Author contributions: J.B. and B.L.-S. designed research; J.B. and O.P. performed research;J.B., S.B.-D., and B.L.-S. analyzed data; and J.B., O.P., and B.L.-S. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Data deposition: The sequences reported in this paper has been deposited in the GenBankdatabase [accession nos. JN392856 (zftac3a), JN392857 (zftac3b), JF317292 (zftac3ra),JF317293 (zftac3rb), and BK008087–BK008126].1To whom correspondence should be addressed. E-mail: sivan@agri.huji.ac.il.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1119165109/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1119165109 PNAS | June 26, 2012 | vol. 109 | no. 26 | 10269–10274

AGRICU

LTURA

LSC

IENCE

S

Dow

nloa

ded

by g

uest

on

July

5, 2

020

piscine Tacs. We generated a phylogenetic tree of all availablevertebrate neurokinin genes (Fig. 1A and Fig. S2A); it showedthat the vertebrate neurokinin genes fall into several distinctlineage groups. The identified Tac3 precursors from fish wereclustered with all other previously cloned or predicted Tac3sequences from mammals, frogs, and alligators (Fig. 1A); a sec-ond lineage included Tac1 from both mammals and fish thatwere identified in the present study, and the third lineage in-cluded mammalian Tac4 and a unique piscine group, now namedTac4 (Fig. S2A). No precursors containing the exact NKB se-quence were found in invertebrate species (11).We cloned the full-length tac3ra and tac3rb cDNA from

zebrafish brain by PCR with specific primers (Table S1). Thepredicted tac3ra and tac3rb N termini have features consistentwith a signal peptide, as defined by SignalP program analysis(Fig. S1). Sequence analysis of the two types of zebrafishreceptors identified distinct potential sites for N-glycosylation,

phosphorylation by protein kinase C, protein kinase A, caseinkinase II, tyrosine kinase, and N-myristoylation (Fig. S3). The Nand C termini are, as in other G protein-coupled receptors, themost divergent regions. In our bioinformatic search for NKBreceptors we found four additional TK receptor genes in zebra-fish. To enable assignment of the genes to the three TK receptorsubfamilies already defined in mammals (12), we identified familymembers in additional species, particularly in other fish. Thephylogenetic tree containing vertebrate TK receptors form threeclearly separable groups that corresponded to TAC3R, TAC1R,and TAC2R (Fig. S2B). Putative orthologs of NK receptormembers were also identified in several nonvertebrate species,Caenorhabditis elegans, ciona, and octopus; these served as out-group sequences in determining the root of these three groups,which was located between TAC2R and the others, indicatingthat these groups split early in the family evolution. The treeshows that TAC3R and TAC1R are closest to each other, sug-gesting that their separation was a more recent evolutionary event(Fig. 1B and Fig. S2B). Fish have one gene of Tac2r, and twogenes each of Tac1r and Tac3r. Surprisingly, three Tac3r werefound in zebrafish, the only species to have a third receptor of anyone type. However, we were unable to clone this gene from brainmRNA, so it was excluded from further biological analysis. Thesimilarity and identity among the various TAC3 receptors isshown in Table S3.We found two forms of tac3 genes in zebrafish and salmon, but

more evolved fish contained only one tac3 ortholog; however, all fishspecies exhibit two forms of NKB receptors, suggesting that thepiscine NKB/NKBR can provide an excellent model for under-standing the molecular coevolution of the peptide/receptor pairs.

Gene Organization of tac3 and Chromosomal Synteny of Tac3 andTac3 Receptor. The in silico analyses of fish genomic structureverified that the zftac3 consists of seven exons (Fig. 1C). Inmammals the tac3 gene contains seven exons, five of which aretranslated to form the prepro-NKB protein (11). Notably, theNKBa peptide sequence was encoded in the fifth exon [like inmammals (13, 14)], whereas NKBb spans exons 3–5 (Fig. 1C).Surprisingly, unlike in mammalian NKBs (11), the deducedamino acid sequences of both zftac3 genes encoded an additionalputative TK sequence flanked by a Gly C-terminal amidationsignal, and typical endoproteolytic sites at both termini, sug-gesting that additional TK peptides (YNDIDYDSFVGLM-NH2and YDDIDYDSFVGLM-NH2, spliced from Tac3a and Tac3b,respectively) (Fig. 1C and Fig. S1) are produced by the sameprecursors. Intriguingly, we found this additional peptide in tac3not only in zebrafish but in all other fish species identified in thisstudy (11 species), but not in chicken, lizard, or alligator. Thesepeptides possess an N-terminal dibasic cleavage site with po-tential to release the peptide, and the common NKB motifFVGLM at their C terminal; therefore, we termed this uniquepeptide neurokinin F (NKF) because it has only been found infish species to date. As Page et al. (11) anticipated, the verte-brate TAC3 gene encoded an additional TK in exon 3, ina similar position to substance P in TAC1, and endokinin A/B inTAC4. This TK (NKF) still exists in fish but was lost from otherspecies during evolution. Interestingly, in Tac4 there is a similarloss of one active peptide in mammals (the C-terminal peptide inTac4 as opposed to the N-terminal peptide in Tac3), whereasmost fish species retain putative active peptides in both locations.Chromosome syntenic analysis revealed that the locus of tac3

is highly conserved between teleosts (Fig. S4). zftac3a is locatedon chromosome 23 and tac3b on chromosome 6. The only tac3found in medaka is located on chromosome 7 (Fig. S4B). For thezftac3 gene, the nearest neighboring gene (c1galt1a) is non-syntenic, whereas the next nearest ones (b4galnt1a and slc6a1)were found in inverse order in humans (Fig. S4A). Despite nearlyperfect preservation of synteny, we found substantial shuffling ofgene order along corresponding chromosome arms betweenzebrafish and human. The neighborhoods of gene loci of tac3awere conserved in the zebrafish and medaka (Fig. S4B). We then

A

B

C

Fig. 1. Unrooted phylogenetic tree of neurokinin (A) or neurokinin receptor(B) sequences generated with neighbor-joining (ClustalW 2.1) and maximumlikelihood (Phylip 3.69, ProML) on the basis of alignments performed both byClustalW and Muscle (3.8.31), and visualized with FigTree 1.3.1. The Tac1 andTac4 (A) and Tacr1 and Tacr2 (B) were grouped (full trees in Fig. S2). Thesequences identified in this study are marked in bold. Gene nomenclature hasbeen standardized to Tac3, and species are indicated for illustration andcomparison. Numbers at nodes indicate the bootstrap values from 1,000 rep-licates. (Scale bar: the substitution rate per residue.) GenBank accession num-bers are detailed in the legend to Fig. S2. (C) Gene organization of zebrafishtac3a and tac3b. Each gene is consists of seven exons and encode bothNKF andNKB. SP, signal peptide; ATG and TGA, start and stop codon, respectively.

10270 | www.pnas.org/cgi/doi/10.1073/pnas.1119165109 Biran et al.

Dow

nloa

ded

by g

uest

on

July

5, 2

020

explored the genomic locations of tac3 receptors in humans andvarious fish species (Fig. S4 C and D). In human, TAC3R islocated on chromosome 4, whereas in zebrafish, fugu (Takifugurubripes), medaka (Oryzias latipes), and tetraodon (Tetraodonnigroviridis) tac3ra are located on chromosome 1, unplaced(UN), 1, and 18, respectively. The nearest neighboring gene(cnga2) of the zebrafish tac3ra gene is nonsyntenic with human,but syntenic with the medaka, fugu, and tetraodon. The next-upstream neighboring genes (bdh2, nhedc2, and cisd2) werefound in similar locations in all analyzed species (Fig. S4C). Thehuman genome lacks tac3rb, which is present in tetraodon, fugu,and medaka. The next-upstream neighboring genes in thezebrafish (acy3.1, acy3.2, cldnd, and glb1) were found in reverseorder in the tetraodon (Fig. S4D). Tac3rc had no discerniblesynteny to any other family members. The presence of two formsof NKB [in zebrafish and salmon (salmo salar)] and two forms ofcognate receptor genes in lower-vertebrate species supports thehypothesis of two rounds of genome duplication followed bydegeneration and complementation of the genes. We concludethat the synteny is better within fish than with mammals; more-over, fish tac3 and tac3r have conserved synteny with the humangenome, consistent with orthology.

Tissue Distribution of tac3 and tac3 Receptors in Zebrafish. To elu-cidate the physiological roles of the NKB/NKBR signaling sys-tem, we next examined the tissue distribution of both ligands(tac3a, tac3b) and receptors (tac3ra, tac3rb) mRNAs in zebrafishby means of real-time PCR analysis, according to Biran et al.(15). We dissected the zebrafish brain into three parts, of whichthe anterior part contains the telencephalon, the midbrain con-tains the optic tectum, diencephalon, and hypothalamus and thehindbrain, the medulla oblongata and cerebellum. We mostlydetected tac3a mRNA in the midbrain and tac3b mRNA wasfound mainly in the forebrain. Both tac3a and tac3ra wereexpressed in the pituitary (Fig. 2A), collaborating findings inmammals where NKB and NKBR were expressed in the medianeminence, which is missing in fish (6). tac3rb was expressed in theforebrain and was highest in the ovary. Different types of tac3and tac3r were expressed in the ovary and testis (Fig. 2A). Dif-ferent levels of expression of tac3 and tac3r types were found inextrabrain tissues (Fig. S5). The expression patterns of thesegenes in the brain-pituitary-gonad axis further support the po-tential role of the NKB system in fish reproduction.

Gene Expression of the NKB/NKBR System During Sexual Maturation.Because it is known that the mammalian NKB/NKBR system isinvolved in reproduction, and especially in puberty initiation(16), we aimed to examine whether the piscine system fulfillsa similar role. We used real-time PCR to evaluate the expressionprofiles of the zfNKBs and their receptor mRNAs in the brainduring several development stages. Expression of tac3a mRNAwas low at 2–4 wk postfertilization (wpf) (Fig. 2B); it thengradually increased, peaked at 8 wpf, when the zebrafish gothrough puberty, and subsequently decreased by 12 wpf (Fig.2B), when the gonads contained clear, well-developed oocytes or

spermatozoa (15). The expression of the tac3b, tac3ra, and tac3rbmRNAs in the zebrafish brain was low and did not change duringsexual maturation. The increase of tac3a mRNA toward puberty,consistent with kisspeptin signaling initiating puberty (15, 17),may indicate a possible involvement of the NKB/NKBR systemin controlling puberty.

Localization of Embryonic tac3a Cells by Whole-Mount in Situ Hybridi-zation. The first appearance of tac3a expression was detected at 3d postfertilization (dpf) in the right habenula nuclei and the mid-brain (Fig. 3 A, F, and K). The earliest stage that the signal wasobserved in the left habenula was at 4–5 dpf, when the signal in-tensity in the right habenula and midbrain increased, probablyreflecting increased cell numbers (Fig. 3 B, C, G, H, L, and M). Inaddition, there was expression in the hindbrain. Analysis of elderlylarvae (7 and 9 dpf) revealed decreased tac3a signal intensity (Fig. 3D, E, I, J,N, andO), and at 12 dpf it was barely detected. No signalwas observed at any stage by use of the tac3a sense riboprobe.During embryogenesis tac3awas dominantly expressed in the right-habenula nuclei, but in adults it was expressed in both habenularlobes (Fig. 3P). This asymmetrical expression of tac3a in thehabenula is consistent with previous findings that the habenula inzebrafish displayed left-right asymmetries in gene expression (18,19). Interestingly, it was shown that neurons appear sooner in theleft than in the right habenula (20). Neuronal organization asym-metries in the epithalamus (i.e., habenular nuclei and pineal com-plex) are well known among vertebrates (21).

Localization of tac3a mRNA in the Brain of Adult Zebrafish. By usingin situ hybridization (ISH) techniques to determine the localizationof tac3a mRNA in the zebrafish brain, we detected tac3a mRNA-expressing neurons in the habenula (Fig. 3P), where kisspeptin 1(kiss1) was previously shown to be expressed (22). Tac3a was alsodetected along the periventricular hypothalamus (Fig. 3 Q and R),in the periventricular nucleus of the posterior tuberculum (Fig. 3Q),and in the posterior tuberal nucleus (Fig. 3R). These brain nucleiwere previously found to express other important neuropeptidesthat regulate reproduction (17, 22), metabolism (23), and stress(24). In the mammalian ARC, KISS1 neurons that coexpress NKBand dynorphin were hypothesized to be a central node throughwhich potential stress, metabolic and photoperiodic signals regulateGnRH release (25). The nuclear lateralis tuberis is considered as thepiscine structure homologous to the mammalian ARC (23, 26). Thelocalization of tac3a, kisspeptin 2 (kiss2), kiss1 receptor b (kiss1rb),leptin receptor, melanin-concentrating hormone (MCH) 2 and twoMCH1 receptors, Urotensin I, corticotropin-releasing factor, andcorticotropin-releasing factor-binding protein to the ventral zone ofthe periventricular hypothalamus (Fig. 3Q) (17, 22–24, 26) mightsuggest that not all neuropeptide pathways are as conserved asformerly thought. This suggestion is supported by the recent findingsthat kiss2 is not expressed in the zebrafish nuclear lateralis tuberis,and that kiss2 neurons of the periventricular hypothalamus do notdirectly innervate the zebrafish pituitary (22). Finally, the closesimilarity between the expression patterns of zebrafish kisspeptingenes and zebrafish tac3a suggests a possible interaction between

A BFig. 2. Expression of zebrafish tac3a, tac3b,tac3ra, or tac3rb mRNA in various parts of thebrain (A) and changes in zebrafish tac3a at vari-ous ages toward puberty (B) as determined byreal-time PCR. The relative abundance of themRNAs was normalized to the amount of elon-gation factor 1 α (ef1α) by the comparativethreshold cycle method; the comparative thresh-old reflects the relative amount of the transcript.Results are means ± SEM (n = 11–15). Meansmarked with different letters differ significantly(P < 0.05).

Biran et al. PNAS | June 26, 2012 | vol. 109 | no. 26 | 10271

AGRICU

LTURA

LSC

IENCE

S

Dow

nloa

ded

by g

uest

on

July

5, 2

020

these two systems, as previously shown in mammals; however, thisneeds further investigation.

Pharmacological Analysis and Signal Transduction Pathways of zfTac3Receptors. We used functional expression analysis with COS-7cells to evaluate the response, binding selectivity, and signaltransduction pathways of the unique TK receptors to their ago-nists. We previously showed the specificity of the reporter serumresponsive element (SRE)-Luc and cAMP responsive element(CRE)-Luc, to activation of PKC/Ca2+ and PKA/cAMP signaltransduction pathways, respectively (15). Graded concentrationsof the zfTKs (NKBa, NKBb, and NKF), and of hNKB and itsagonist senktide were applied to COS-7 cells that expressedhNKBR, zfTac3ra, or zfTac3rb. The EC50 values of TKs for eachreceptor are summarized (Table S4). Both human and piscineTKs induced concentration-dependent increases in both SRE-Luc and CRE-Luc activity (Fig. 4 A–F). For hNKBR, in bothsignal transduction systems, zfNKBa and NKF showed high po-tency, very similar to human NKB, but zfNKBb peptide exhibitedrelatively low potency (Fig. 4 A–D). For both zfTac3 receptors,zfNKBa and NKF (both derived from tac3a) were similarlyhighly potent in both signal transduction pathways (Fig. 4).These results confirmed that both zfNKBa and NKF were en-dogenous ligands of Tac3 receptors, and it is noteworthy that thisreport of activation of TK receptors by a second peptide derivedfrom the NKB gene (zfNKF) is unique. However, NKBb was lesseffective than the other forms in eliciting luciferase activity byboth signal transduction pathways. It was previously shown un-equivocally that hNKBR potentially can couple directly to bothphospholipase C and adenylate cyclase, and stimulate bothphosphoinositides hydrolysis and cAMP formation (27). Indeed

both zftac3 receptors, as well as the hNKBR posses both PKCand PKA phosphorylation sites (Fig. S3). These findings confirmthat the unique zebrafish receptors relay their signal throughboth PKC and PKA transduction pathways.

Ligand Models. Fig. 4G is a ribbon representation of the zfNKBstructural model prediction, compared with the hNKB (PDB ID1p9f). Intriguingly, although the three zfNKBs vary in size—10,24, and 13 aa for NKBa, NKBb, and NKFa, respectively—all ofthe predicted peptides yielded high-resolution, high-qualitystructures with typical globular folding. These structures com-prised α-helix-loop motifs (highlighted in red in Fig. 4G). Allthree zfNKBs model structures approximated a binding-compe-tent conformation similar to the human NKB. Our results cor-roborate previous ones found for mammals that the formation ofa helical conformation in the mid region of each of the TKsappears to be crucial for TK-receptor activation (28). Moreover,mammalian NKB were found to form a helical structure in thepresence of dodecylphosphocholine micelles (29).

In Vivo Effect of Estradiol. Our next aim was to test the in-volvement of the NKB system in reproduction. In fish, clear ev-idence exists regarding the important role estradiol plays duringthe period of reproduction (1). Moreover, ISH was recently usedto show that estradiol increased kisspeptin cell number in theventral hypothalamus, where the zfkiss2 neurons are homologousto the estrogen-sensitive kiss1 hypothalamic neurons in medaka(17, 22, 30). Similarly to mammals, zebrafish have two forms ofGnRH: GnRH2 is localized to the midbrain tegmentum, andGnRH3 (considered to be the hypophysiotropic form) is locatedat both the olfactory bulb terminal nerve and the preoptic area(31). Because in mammals kisspeptin and NKB are expressed inthe same neurons in the hypothalamic ARC, and play a key rolein physiological regulation of GnRH neurons (25), we tested theeffect of estradiol on the expression of genes along the GnRH-kisspeptin system. Estradiol treatment of prepubertal zebrafishenhanced expression of key genes involved in reproduction(gnrh3, kiss2, and kiss1) concomitantly with significantly increasedtac3a (Fig. 5A). In parallel, we also found significantly increasedexpression of tac3ra, tac3rb, and kiss1ra (Fig. 5B). Both Tac3receptors bound the zfNKBs (Fig. 4 A–F), and Kiss1ra was shownpreviously to bind Kiss2, considered to be the more importantform, with higher affinity than Kiss1 (32). In mammals there isstrong evidence of sexual dimorphism of NKB neurons: largernumbers of NKB neurons have been identified in ARC of ewesthan of rams (33). The transcription of NKB could be directlyaltered by estrogen receptors, as sequences corresponding to theestrogen responsive element and imperfect palindromic estrogenresponsive element have been reported upstream of the TAC3gene transcriptional start site (8). In fish estradiol is involved inboth early oogenesis and the beginning of the first wave of vi-tellogenesis that precede puberty (34), collaborating our findingthat tac3a expression peaked in prepubertal fish (Fig. 2B).Moreover, increased levels of estradiol are a required charac-teristic of both follicular growth and final oocyte maturation infish (3, 35), pointing toward the involvement of the piscine NKBsystem in control of reproduction, probably in concert withkisspeptin and GnRH.

In Vivo Effect of NKBs. We next tested the in vivo biologicalfunction of zebrafish NKB peptides. Single intraperitoneal in-jection of zfNKBa or zfNKF elicited significant LH secretion insexually mature female zebrafish (Fig. 5C). The magnitude of theinduced LH discharge was comparable with that observed inresponse to GnRH. LH response to the hNKBR agonist, senk-tide, or zfNKBb was less pronounced (Fig. 5C), in a similar wayto the order of potency obtained in the transactivation assay (Fig.4). Our findings corroborate previous findings that activation ofNKB receptors evoked potent LH-secretory responses inrodents, sheep, and monkey (4, 36–38).

Fig. 3. (A–O) Localization of tac3a during early stages of development, asdetected by whole-mount ISH. tac3a is dominantly expressed in the righthabenular nuclei,midbrain, andhindbrain. Dorsal viewof larva heads, anterioris to the left (A–E). High magnification of boxed area inUpper panel. Note theunilateral expression of tac3a to right of the midline (dotted line) in thehabenula (F–J), lateral view of larva heads (K–O). rHbn, right habenula; MB,midbrain; HB, hindbrain. (P–R) Localization of tac3a in adult zebrafish brain asindicated by ISH (nomenclature according to ref. 40). Tac3a mRNA-expressingcells were observed in the ventral (Hav) and medial (Ham) habenula (P) tac3amRNA-expressing cells detected in the periventricular nucleus of posteriortuberculum (TPp), dorsal (Hd), and ventral zone (Hv) of periventricular hypo-thalamus (Q) tac3a mRNA-expressing cells observed in the posterior tuberalnucleus (PTN) and central zone (Hc) of periventricular hypothalamus (R).(Magnification: A–E and K–O, 40×; F–J, 120×).

10272 | www.pnas.org/cgi/doi/10.1073/pnas.1119165109 Biran et al.

Dow

nloa

ded

by g

uest

on

July

5, 2

020

In conclusion, we provided detailed information on the orga-nization of the NKB systems in teleosts, including the splice tac3(NKF). We also show that the zftac3a, unlike zftac3b, expressionincreased toward puberty, increased in response to estradioltreatment, caused increase in LH secretion, and was abundantlyexpressed in the brain, notably in the hypothalamus. Tac3receptors were expressed in the pituitary and gonads, suggestingthat these ligand-receptor pairs are likely involved in the controlof reproductive functions, during puberty or during diverse stepsof reproduction.

Materials and MethodsAnimals.Wild-type zebrafish were purchased from a commercial supplier (A &H). All experimental procedures were approved by the Hebrew UniversityAdministrative Panel for Laboratory Animal Care.

Data Mining, Phylogenetic Analysis, and Chromosomal Synteny. The putativeTac3 gene sequences were isolated from zebrafish by using a stepwiseevolutionary strategy, with the mouse Tac2 protein (NP_033338.2) as first-step input, and extension to other genomes and ESTs, as described in the SIMaterials and Methods. Details of sequences, phylogenetic analyses andsynteny are presented in SI Materials and Methods.

Isolation of Zebrafish tac3 Ligands and Receptors. We designed specific pri-mers for cloning the putative zfTac3 ligands and receptors (Table S1). Thefragments were PCR-amplified from an adult zebrafish brain cDNA libraryand were cloned into pCRII-TOPO vector (Invitrogen).

Tissue Distribution and Expression Profiles. Tissue distributions of zftac3a,zftac3b, zftac3ra, and zftac3rbweredeterminedby real-time PCR as previouslydescribed (15). Tissue samples were collected from sexually mature post-vitellogenic female and milt-producing male zebrafish, total RNA extraction,and cDNA samples were prepared as previously described (15). To study geneexpression of the NKB/NKBR system at different ages, 15 fish were randomlysampled at ages 2, 4, 6, 8, and 12wpf. The brainwas removed, and thepubertalstage classification was determined as described previously (15). Elongationfactor 1α (39) was used as a reference gene. The primer sequences, R2 values,and slopes of the real-time PCR analyses, calculated by linear regression, arepresented in Table S1.

ISH Analysis of Embryos and Adults. ISH was conducted on embryos and adultsaccording to Mitani et al. (30) and Palevitch et al. (40) respectively, as de-tailed in SI Materials and Methods.

Peptide Synthesis. Zebrafish NKBa (EMHDIFVGLM-NH2), NKBb (STGIN-REAHLPFRPNMNDIFVGLLEMHDIFVGLM-NH2), and NKF (YNDIDYDSFVG-LM-NH2) were synthesized by the automated solid-phase method by

-11 -10 -9 -8 -7 -6 -5. 0

0

1

2

3

4

-11 -10 -9 -8 -7 -6 -5. 0

0

1

2

3

4

SR

E-L

uc activity

(fo

ld

in

du

ctio

n)

-11 -10 -9 -8 -7 -6 -5. 0

0

1

2

3

4

-11 -10 -9 -8 -7 -6 -5. 0

0

1

2

3

-11 -10 -9 -8 -7 -6 -5. 0

0

2

4

6

8

10

peptide (logM)

CR

E-L

uc activity

(fo

ld

in

du

ctio

n)

-11 -10 -9 -8 -7 -6 -5. 0

0

2

4

6

8

10

12

peptide (logM)

-11 -10 -9 -8 -7 -6 -5. 0

0

2

4

6

8

peptide (logM)

A B C

D E F

GhNKBR zfTac3ra zfTac3rb

zfNKBa

zfNKBb

zfNKF

hNKB

Senktide

zfNKBa

zfNKBb

zfNKF

hNKB

Senktide

Fig. 4. Ligand selectivity of the NKB receptors. Human NKBR (A and D) zebrafish Tac3ra (B and E) or zebrafish Tac3rb (C and F), each together with SRE-Luc(A–C) or CRE-Luc (D–F). The cells were treated with various concentrations of human (hu) NKB: DMHDFFVGLM-NH2; zebrafish (zf) NKBa: EMHDIFVGLM-NH2;zfNKBb: STGINREAHLPFRPNMNDIFVGLL-NH2; or zfNKF: YNDIDYDSFVGLM-NH2. Data are expressed as the increase in luciferase activity over basal activity.Each point was determined in quadruplicate and is given as a mean ± SEM. (G) Ribbon representation of human and zebrafish NKBs structural model. ThePDB ID for the human structure is 1p9f.

A B C

Fig. 5. Exposure to estradiol (18 nM) by immersion caused a significant increase in mRNA expression of ligands (A) or receptor (B) genes, in the brain ofjuvenile zebrafish. sGnRHa, zfNKBa, zfNKBb, zfNKF, or senktide (see legend to Fig. 4 for details), were injected intraperitoneally to mature zebrafish andblood was collected 6 h thereafter (C). Hormone values are means ± SEM. Statistical significance vs. corresponding control values: **P < 0.01; *P < 0.05.

Biran et al. PNAS | June 26, 2012 | vol. 109 | no. 26 | 10273

AGRICU

LTURA

LSC

IENCE

S

Dow

nloa

ded

by g

uest

on

July

5, 2

020

applying Fmoc active-ester chemistry, purified by HPLC to >95% purity(GeneMed), and the carboxyl terminus of each peptide was amidated.

Receptor Transactivation Assay and Protein Structure Modeling. To study thesignaling pathways of the recently identified zfNKBs, the entire codingregions of zftac3ra and zftac3rb were inserted into pcDNA3.1 (Invitrogen).The cDNA clone for hNKBR was obtained from the Missouri S&T cDNA Re-source Center (www.cdna.org), and the luciferase assay was conducted (15).Protein structures of zebrafish NKBa, NKBb, or NKF were predicted on theI-Tasser server (41, 42).

In Vivo Experiments. Two-month-old prepubertal zebrafish were exposed toE2 (Sigma) at a concentration of 5 μg/L (18 nM), or to vehicle (ethanol 100%to a final concentration of 33 μL/L), by immersion for 3 d. This relatively lowconcentration was used because the natural E2 concentration in the plasmaof adult vitellogenic female zebrafish was determined as 3–4 ng/mL (43).

Each group of 10 fish was kept in a 3-L aquarium with water maintained at28 °C and replaced daily. After 3 d, fish were anesthetized and brains wereremoved. RNA extraction, reverse transcription of RNA and real-time PCRwere carried out as previously described (15).

Adult female zebrafish were injected intraperitoneally with 20 pmol/gbody weight of either saline, salmon GnRH analog [(D-Ala6,Pro9-Net)-mammalian GnRH], zfNKBa, zfNKBb, zfNKF, or senktide (n = 8 fish pergroup). Six hours postinjection the fish were bled (44). Blood wascentrifuged at 970 × g 30 min and the plasma was separated and stored at−20 °C. Plasma from two animals were pooled (within each treatment)and analyzed for LH using ELISA for other Cyprinidae, carp, according toAizen et al. (35, 45).

ACKNOWLEDGMENTS. This research was funded by United States-IsraelBinational Science Foundation Grant 2005096, and by Binational Agricul-tural Research and Development Fund MD-8719-08.

1. Yaron Z, Levavi-Sivan B (2006) Fish reproduction. Physiology of Fishes, eds Evans DH,Claiborne JB (CRC, New York), 3rd Ed, pp 345–388.

2. Ball JN (1981) Hypothalamic control of the pars distalis in fishes, amphibians, andreptiles. Gen Comp Endocrinol 44:135–170.

3. Levavi-Sivan B, Bogerd J, Mañanós EL, Gómez A, Lareyre JJ (2010) Perspectives on fishgonadotropins and their receptors. Gen Comp Endocrinol 165:412–437.

4. Navarro VM, et al. (2011) Interactions between kisspeptin and neurokinin B in thecontrol of GnRH secretion in the female rat. Am J Physiol Endocrinol Metab 300:E202–E210.

5. Topaloglu AK, et al. (2009) TAC3 and TACR3 mutations in familial hypogonadotropichypogonadism reveal a key role for Neurokinin B in the central control of re-production. Nat Genet 41:354–358.

6. Rance NE, Krajewski SJ, Smith MA, Cholanian M, Dacks PA (2010) Neurokinin B andthe hypothalamic regulation of reproduction. Brain Res 1364:116–128.

7. Page NM (2004) Hemokinins and endokinins. Cell Mol Life Sci 61:1652–1663.8. Page NM, Woods RJ, Lowry PJ (2001) A regulatory role for neurokinin B in placental

physiology and pre-eclampsia. Regul Pept 98:97–104.9. Seidah NG, Prat A (2002) Precursor convertases in the secretory pathway, cytosol and

extracellular milieu. Essays Biochem 38:79–94.10. Eipper BA, Stoffers DA, Mains RE (1992) The biosynthesis of neuropeptides: Peptide

alpha-amidation. Annu Rev Neurosci 15:57–85.11. Page NM, Morrish DW, Weston-Bell NJ (2009) Differential mRNA splicing and pre-

cursor processing of neurokinin B in neuroendocrine tissues. Peptides 30:1508–1513.12. Maggi CA (1995) The mammalian tachykinin receptors. Gen Pharmacol 26:911–944.13. Bonner TI, Affolter H-U, Young AC, Young WS, 3rd (1987) A cDNA encoding the

precursor of the rat neuropeptide, neurokinin B. Brain Res 388:243–249.14. Kotani H, Hoshimaru M, Nawa H, Nakanishi S (1986) Structure and gene organization

of bovine neuromedin K precursor. Proc Natl Acad Sci USA 83:7074–7078.15. Biran J, Ben-Dor S, Levavi-Sivan B (2008) Molecular identification and functional

characterization of the kisspeptin/kisspeptin receptor system in lower vertebrates.Biol Reprod 79:776–786.

16. Navarro VM, et al. (2012) Role of neurokinin B in the control of female puberty and itsmodulation by metabolic status. J Neurosci 32:2388–2397.

17. Kitahashi T, Ogawa S, Parhar IS (2009) Cloning and expression of kiss2 in the zebrafishand medaka. Endocrinology 150:821–831.

18. Aizawa K, et al. (2007) Responses of embryonic germ cells of the radiation-sensitiveMedaka mutant to gamma-irradiation. J Radiat Res (Tokyo) 48:121–128.

19. Bianco IH, Carl M, Russell C, Clarke JD, Wilson SW (2008) Brain asymmetry is encodedat the level of axon terminal morphology. Neural Dev 3:9.

20. Roussigné M, Bianco IH, Wilson SW, Blader P (2009) Nodal signalling imposes left-right asymmetry upon neurogenesis in the habenular nuclei. Development 136:1549–1557.

21. Concha ML, Wilson SW (2001) Asymmetry in the epithalamus of vertebrates. J Anat199:63–84.

22. Servili A, et al. (2011) Organization of two independent kisspeptin systems derivedfrom evolutionary-ancient kiss genes in the brain of zebrafish. Endocrinology 152:1527–1540.

23. Berman JR, Skariah G, Maro GS, Mignot E, Mourrain P (2009) Characterization of twomelanin-concentrating hormone genes in zebrafish reveals evolutionary and physi-ological links with the mammalian MCH system. J Comp Neurol 517:695–710.

24. Alderman SL, Bernier NJ (2007) Localization of corticotropin-releasing factor, ur-otensin I, and CRF-binding protein gene expression in the brain of the zebrafish,Danio rerio. J Comp Neurol 502:783–793.

25. Lehman MN, Coolen LM, Goodman RL (2010) Minireview: Kisspeptin/neurokinin B/dynorphin (KNDy) cells of the arcuate nucleus: A central node in the control of go-nadotropin-releasing hormone secretion. Endocrinology 151:3479–3489.

26. Liu Q, et al. (2010) Expression of leptin receptor gene in developing and adult ze-brafish. Gen Comp Endocrinol 166:346–355.

27. Nakajima Y, Tsuchida K, Negishi M, Ito S, Nakanishi S (1992) Direct linkage of threetachykinin receptors to stimulation of both phosphatidylinositol hydrolysis and cyclicAMP cascades in transfected Chinese hamster ovary cells. J Biol Chem 267:2437–2442.

28. Almeida TA, et al. (2004) Tachykinins and tachykinin receptors: Structure and activityrelationships. Curr Med Chem 11:2045–2081.

29. Mantha AK, Chandrashekar IR, Baquer NZ, Cowsik SM (2004) Three dimensionalstructure of mammalian tachykinin peptide neurokinin B bound to lipid micelles.J Biomol Struct Dyn 22:137–148.

30. Mitani Y, Kanda S, Akazome Y, Zempo B, Oka Y (2010) Hypothalamic Kiss1 but notKiss2 neurons are involved in estrogen feedback in medaka (Oryzias latipes). Endo-crinology 151:1751–1759.

31. Steven C, et al. (2003) Molecular characterization of the GnRH system in zebrafish(Danio rerio): Cloning of chicken GnRH-II, adult brain expression patterns and pitui-tary content of salmon GnRH and chicken GnRH-II. Gen Comp Endocrinol 133:27–37.

32. Lee YR, et al. (2009) Molecular evolution of multiple forms of kisspeptins and GPR54receptors in vertebrates. Endocrinology 150:2837–2846.

33. Cheng G, Coolen LM, Padmanabhan V, Goodman RL, Lehman MN (2010) The kiss-peptin/neurokinin B/dynorphin (KNDy) cell population of the arcuate nucleus: Sexdifferences and effects of prenatal testosterone in sheep. Endocrinology 151:301–311.

34. Lubzens E, Young G, Bobe J, Cerdà J (2010) Oogenesis in teleosts: How eggs areformed. Gen Comp Endocrinol 165:367–389.

35. Aizen J, Kasuto H, Levavi-Sivan B (2007) Development of specific enzyme-linked im-munosorbent assay for determining LH and FSH levels in tilapia, using recombinantgonadotropins. Gen Comp Endocrinol 153:323–332.

36. Ramaswamy S, et al. (2010) Neurokinin B stimulates GnRH release in the male monkey(Macaca mulatta) and is colocalized with kisspeptin in the arcuate nucleus. Endocri-nology 151:4494–4503.

37. Billings HJ, et al. (2010) Neurokinin B acts via the neurokinin-3 receptor in the ret-rochiasmatic area to stimulate luteinizing hormone secretion in sheep. Endocrinology151:3836–3846.

38. García-Galiano D, et al. (2012) Kisspeptin signaling is indispensable for neurokinin B,but not glutamate, stimulation of gonadotropin secretion in mice. Endocrinology153:316–328.

39. Tang R, Dodd A, Lai D, McNabb WC, Love DR (2007) Validation of zebrafish (Daniorerio) reference genes for quantitative real-time RT-PCR normalization. Acta BiochimBiophys Sin (Shanghai) 39:384–390.

40. Palevitch O, et al. (2007) Ontogeny of the GnRH systems in zebrafish brain: In situhybridization and promoter-reporter expression analyses in intact animals. Cell TissueRes 327:313–322.

41. Zhang Y (2008) I-TASSER server for protein 3D structure prediction. BMC Bio-informatics 9:40.

42. Roy A, Kucukural A, Zhang Y (2010) I-TASSER: A unified platform for automatedprotein structure and function prediction. Nat Protoc 5:725–738.

43. Heiden TK, Carvan MJ, 3rd, Hutz RJ (2006) Inhibition of follicular development, vi-tellogenesis, and serum 17beta-estradiol concentrations in zebrafish followingchronic, sublethal dietary exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin. Toxicol Sci90:490–499.

44. Detrich HW, Westerfield M, Zon LI (2011) The Zebrafish: Cellular and DevelopmentalBiology (Elsevier, Amsterdam), Vol 101, p 92.

45. Aizen J, et al. (2012) Steroidogenic response of carp ovaries to piscine FSH and LH de-pends on the reproductive phase. Gen Comp Endocrinol 178:28–36.

10274 | www.pnas.org/cgi/doi/10.1073/pnas.1119165109 Biran et al.

Dow

nloa

ded

by g

uest

on

July

5, 2

020