The nerve of ovulation-inducing factor in semenMarcelo H. Rattoa, Yvonne A. Leducb, Ximena P. Valderramac, Karin E. van Straatend, Louis T. J. Delbaereb,1,Roger A. Piersone, and Gregg P. Adamsf,2

aFaculty of Veterinary Sciences and cFaculty of Agricultural Sciences, Universidad Austral de Chile, Valdivia, Chile; and Departments of bBiochemistry,dChemistry, fVeterinary Biomedical Sciences, and eObstetrics Gynecology and Reproductive Science, University of Saskatchewan, Saskatoon, SK,Canada S7N 5B4

Edited* by George E. Seidel, Colorado State University, Fort Collins, CO, and approved July 16, 2012 (received for review April 13, 2012)

A component in seminal fluid elicits an ovulatory response and hasbeen discovered in every species examined thus far. The existence ofan ovulation-inducing factor (OIF) in seminal plasma has broad impli-cations and evokes questions about identity, tissue sources, mecha-nism of action, role among species, and clinical relevance in infertility.Most of these questions remain unanswered. The goal of this studywas to determine the identity ofOIF in support of thehypothesis thatit is a single distinct and widely conserved entity. Seminal plasmafrom llamas and bulls was used as representative of induced andspontaneous ovulators, respectively. A fraction isolated from llamaseminal plasma by column chromatography was identified as OIF byeliciting luteinizing hormone (LH) release and ovulation in llamas.MALDI-TOF revealed a molecular mass of 13,221 Da, and 12–23 aasequences of OIF had homology with human, porcine, bovine, andmurine sequences of β nerve growth factor (β-NGF). X-ray diffractiondata were used to solve the full sequence and structure of OIF asβ-NGF. Neurite development and up-regulation of trkA in phaeochro-mocytoma (PC12) cells in vitro confirmed NGF-like properties of OIF.Western blot analysis of llama and bull seminal plasma confirmedimmunorecognition of OIF using polyclonal mouse anti-NGF, and ad-ministration of β-NGF from mouse submandibular glands inducedovulation in llamas. We conclude that OIF in seminal plasma isβ-NGF and that it is highly conserved. An endocrine route of actionof NGF elucidates a previously unknown pathway for the direct in-fluence of the male on the hypothalamo–pituitary–gonadal axis ofthe inseminated female.

neurotrophins | hypothalamus | fertility | neuroendocrine

In a monograph nearly 50 y ago, ThaddeusMann summarized thenatural properties of seminal plasma as a vehicle for sperm

transport, a controller of sperm motility and capacitation, and asa stimulant of uterine contractility (1). Notwithstanding Mann’sadmonishment to resist the temptation “to assign to every newlydiscovered chemical constituent of semen a major role in theprocess of fertilization,” recent isolation of a protein factor inseminal plasma (2–4) suggests an additional role of the ejaculate—as an inducer of ovulation.The role of the fluid portion of the ejaculate, and the male

accessory glands responsible for producing it, has been enig-matic. From an evolutionary perspective, it has been suggestedthat the male accessory glands likely originated as the machineryfor producing a copulatory plug, which has the “chastity effect”of preventing the sperm of other males from entering the femaletract, as well as minimizing sperm loss after insemination (5). Ifthis is so, then the persistence of an elaborate accessory glandsystem in many species in which plug formation does not occurmay be viewed as nothing more than an evolutionary vestige.The first reports of an ovulation-inducing factor (OIF) in se-

men resulted from the observation that ovulation occurred afterintravaginal or intramuscular administration of Bactrian seminalplasma to female Bactrian camels (6, 7). The existence of OIF inseminal plasma was confirmed in later studies involving llamasand alpacas, New World relatives of camels (2). The results ofthis and subsequent studies have led to the discoveries that OIFexists in the semen of camelids (induced ovulators), is a potentstimulator of luteinizing hormone (LH) secretion, has a dose-

dependent effect on ovulation and the form and function of thecorpus luteum and acts via a systemic rather than a local pathwayat physiologically relevant doses (2, 8–10).It is not surprising that the discovery of OIF in seminal plasma

wasmade in species categorized as induced ovulators (e.g., rabbits,camelids, and koalas) because factors influencing the occurrenceof ovulation can be studied without the confounding effects ofspontaneous ovulation. However, results of recent studies supportthe hypothesis that OIF in seminal plasma is conserved amongspecies, including those considered to be spontaneous ovulators(e.g., cattle, horses, pigs, and mice) (9, 11). Furthermore, OIF inseminal plasma influenced ovarian function in species consideredto be spontaneous ovulators. That is, OIF induced ovulation ina prepubertal mouse model (11) and altered ovarian follicularwave dynamics in cows (12).To characterize the biochemical nature of OIF, attempts were

made to ablate the bioactivity of seminal plasma by molecularmass cutoff filtration, treatment with charcoal or heat, and en-zymatic digestion with proteinase K or pronase E (3). An in vivollama ovulation bioassay was used to confirm that OIF is nota steroid, prostaglandin, or gonadotropin releasing hormone(GnRH); rather it is a protein molecule that is resistant to heatand enzymatic digestion with proteinase K and has a molecularmass of more than about 30 kDa. Protein fractions of llamaseminal plasma were subsequently isolated and purified by col-umn chromatography and tested for ovulation-inducing bio-activity using the in vivo llama ovulation bioassay (4).The primary goal of the following series of studies was to

determine the identity of OIF in seminal plasma in support ofthe hypothesis that OIF is a single distinct and widely conservedconstituent of seminal plasma. Seminal plasma from llamas andbulls was used as that of species representative of induced andspontaneous ovulators, respectively.

ResultsBiochemical Purification and Protein Identification of OIF. Using anapproach similar to that described in our previous report (4),three protein fractions of llama seminal plasma were identifiedusing hydroxylapatite column chromatography (fractions A, B,and C) and a prominent protein band with a mass of ∼14 kDa wasidentified by SDS/PAGE of fraction C. Fraction C was loaded intoa sephacryl gel filtration column for further purification resulting intwo distinct subfractions, C1 and C2. Fraction C2 contained a highly

Author contributions: M.H.R., R.A.P., and G.P.A. designed research; M.H.R., Y.A.L., X.P.V.,K.E.v.S., and G.P.A. performed research; L.T.J.D. and G.P.A. contributed new reagents/analytictools; M.H.R., Y.A.L., X.P.V., K.E.v.S., and G.P.A. analyzed data; and M.H.R., Y.A.L., K.E.v.S.,and G.P.A. wrote the paper.

The authors declare no conflict of interest.

*This Direct Submission article had a prearranged editor.

Data deposition: The atomic coordinates have been deposited in the Protein Data Bank,www.pdb.org (PDB ID code 4EFV).1Deceased October 5, 2009.2To whom correspondence should be addressed. E-mail: gregg.adams@usask.ca.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1206273109/-/DCSupplemental.

15042–15047 | PNAS | September 11, 2012 | vol. 109 | no. 37 www.pnas.org/cgi/doi/10.1073/pnas.1206273109

Dow

nloa

ded

by g

uest

on

Nov

embe

r 6,

202

0

purified isolate of a 14-kDa protein (Fig. 1) that was identifiedpreviously as OIF by the effect of eliciting a preovulatory LH surgefollowed by ovulation and corpus luteum formation in >95% ofllamas after intramuscular administration (4).Analysis of the purified nontrypsinated fraction C2, cut from the

SDS/PAGEgel, byMALDI-TOF revealed amajor peakwith amassof 13,221Da (Fig. 1). Four tryptic peptides of 12–17 aa, derived fromthe 14-kDa band of fraction C2 by capillary LC-MS/MS, had peptidesequence homology ranging from 80 to 100% with human, por-cine, bovine, and murine sequences of type β nerve growth factor(β-NGF) listed in the database of the National Center for Bio-technology Information (NCBI; http://blast.ncbi.nlm.nih.gov/Blast.cgi?PAGE=Proteins), and aligned using the database of the Eu-ropeanBioinformatics Institute (EBI;www.ebi.ac.uk/Clustal; TablesS1 and S2). A sample of nondenatured C2 fraction was analyzed byEdman degradation to confirm the sequence obtained by LC-MS/MS; a 23-residue segment of C2 bore sequence homology rangingfrom 73 to 86% of human, porcine, bovine, and murine sequencesfor NGF listed in the EBI database (Tables S1 and S2 and Fig. 2).The molecular mass of fraction C2 of llama seminal plasma,

determined by SDS/PAGE and MALDI-TOF analyses, corre-sponded to that of themonomeric subunit of β-NGF (i.e., 13 kDa).A narrow peak appeared at 16.755 mL of elution volume afterloading nondenatured fraction C2 into a calibrated superdex gelfiltration column. The peak corresponded to themolecularmass ofβ-NGF homodimer (i.e., 26 kDa; Fig. 1). The reducing conditionsof SDS/PAGE used in this and the previous study (4) apparentlyrendered the homodimer into monomers.

X-Ray Sequence and Crystal Structure of OIF from Seminal Plasma.The four tryptic peptides andEdman chemistry sequence providedonly partial sequencing of OIF. X-ray crystallography was used toderive full OIF sequence information based on electron density. ABLAST search in the database of the National Center for Bio-technology Information (NCBI) with the tryptic peptides revealedhigh sequence identity with β-NGF (Fig. 2). The structure of OIFwas solved by molecular replacement using murine β-NGF (PDBcode 1bet) (13) as a search model, and subsequently refined to2.3-Å resolution. Simulated annealing was used throughout therefinement process to eliminate model bias and generate unbiased

omit maps (Table S3). Most of the OIF residues (residues 9–117)could be assigned on the basis of shape of the electron density, thepartial OIF sequence, identity with homologous β-NGF proteins(Fig. 2), and the chemical environment of the residues. Some resi-dues at theN terminus and loop region (residues 61–66) showedpooror no density. Residues 1–8 were omitted from the structure becausethere was no density; however, they were assigned to the OIF se-quence on the basis of the Edman chemistry sequence. Residue 9showed only enough density to build in Ala, but it is most likely Arg,basedon theLC-MS/MSsequence. Similarly, residue12was assignedas Leu, but is likely Phe, based on the LC-MS/MS sequence. Residue65wasbuilt in asAla, butNGFhomologs haveAsp,Glu, or Ser at thisposition (Fig. 2). The final OIF sequence (residues 1–117; Fig. 2) hasa calculated molecular weight of 13,039 Da, which is slightly smallerthan the experimental value of 13,221Da obtained byMALDI-TOF.The difference may be attributed to missing residues at the C ter-minus (i.e., no observed electron density in the crystal structure dueto flexibility) or some ambiguities from crystallographic sequencing.The overall structure of OIF consists of a noncovalently as-

sociated biological homodimer related by a noncrystallographictwofold rotation axis (Fig. 3). It exists as a dimer in solution (Fig.1D) and crystallizes as a dimer. Each monomer has four loops(L1–L4) and eight β-strands joined by three disulfide bridges.High structural similarity exists between OIF and NGF of mice(PBD codes 1btg and 1bet) (13) and humans (PDB codes 1www,1sg1, and 1sgf) (14). Superposition of the NGF structures ontoOIF revealed root-mean-square deviations ranging from 0.8 to1.2 Å for all overlapping Cα atoms (Fig. 3). This is consistentwith the very high sequence identity between OIF and NGFshown in Fig. 2. Thus, the crystal structure and sequence derivedfrom X-ray crystallography confirmed that OIF is β-NGF.

Confirmation by PC12 Bioassay. Immortalized rat phaeochromocy-toma cells (PC12) have trkA receptors on their cell surface thatbind specifically with NGF, and neurite growth in PC12 cells inresponse to treatment is used as an in vitro bioassay for NGF(15). Treatment of PC12 cells with purified OIF (fraction C2 ofllama seminal plasma) induced neurite outgrowth in a fashionsimilar to that of recombinant mouse NGF after 6 d of in vitroculture (Fig. 4). Further, OIF up-regulated mRNA expression

Fig. 1. Separation and molecular mass of ovulation-inducing factor (OIF) in llama seminal plasma. (A)Separation of fraction C was done using sephacryl gelfiltration fast protein liquid chromatography andisocratic elution with PBS. Fraction C was isolatedpreviously by hydroxylapatite column chromatogra-phy. (B) Protein band at about 14 kDa on denaturing12% SDS/PAGE (circled) was the major constituent offraction C2. (C) Mass spectra (mass-to-charge ratio; m/z)of OIF (fraction C2 isolated from llama seminal plasma)by MALDI-TOF analysis showing a single peak at 13,221Da, corresponding to the monomeric subunit of thehomodimer of β-NGF. (D) Narrow peak appeared at16.755 mL of elution volume after loading a non-denatured sample of fraction C2 into a calibratedsuperdex gel filtration column, corresponding to themolecular mass of β-NGF (i.e., homodimer of 26 kDa).

Ratto et al. PNAS | September 11, 2012 | vol. 109 | no. 37 | 15043

PHYS

IOLO

GY

Dow

nloa

ded

by g

uest

on

Nov

embe

r 6,

202

0

for trkA in PC12 cells similar to that of recombinant mouse NGFafter 4 and 6 d of in vitro culture (Fig. 4).

Immunoblot Analysis of Purified OIF and Whole Seminal Plasma.Western immunoblot analysis using a commercial polyclonal an-tibody against NGF revealed the similarity in immunorecognitionbetween NGF and OIF (fraction C2 from llama seminal plasma;Fig. 5). Further, samples of whole seminal plasma of llamas andbulls displayed a similar staining pattern with a distinct band at∼13 kDa (i.e., that of the NGF monomer). A less distinct bandappeared at just over 60 kDa in immunoblots of whole seminalplasma and was interpreted as pro-NGF (16).

Ovulation-Inducing Effect of NGF. In replicate 1, the proportion ofllamas that ovulated in response to intramuscular treatment with

OIF (250 μg fraction C2 of llama seminal plasma), β-NGF (250 μgfrom mouse submandibular glands), or saline (negative control)was 4/4, 2/4, and 0/4, respectively. An i.v. route of administrationof the same treatments in replicate 2 resulted in an ovulation rateof 4/5, 4/5, and 0/5, respectively. Combined among replicates, theproportion of llamas that ovulated was similar in the OIF- andNGF-treatment groups, both of which were higher than in thesaline-treated group (8/9, 6/9, 0/9; P < 0.01).

DiscussionNerve growth factor belongs to a family of neurotrophins thatincludes brain-derived neurotrophic factor (BDNF), neuro-trophin-3 (NT-3), and neurotrophin-4 (NT-4). All of the neuro-trophins exist in nature as homodimers with a molecular mass of

Fig. 2. Structure-based sequence alignment of ovulation-inducing factor (OIF) from seminal plasma and nerve growth factor (NGF). Sequence alignments ofmature OIF (determined by X-ray crystallography sequencing) with NGF from different species (mouse, PDB 1bet; Saimiri boliviensis, Q5ISB0.2; chimpanzee,BAA90438.1; gorilla, Q9N2F0.1; human, PDB 1sg1; rat, P25427.2; guinea pig, P19093.1; dog, AAY16195.1; goat, AFA52664.1; cow, NP_001092832.1; pig,Q29074.1; chicken, P05200.1; snake venom, Q5YF90.1; turtle, ACY72443.1; and frog, P21617.2). Tryptic is the tryptic peptide sequence of llama OIF (resi-dues:10–23, 35–51, 89–99, and 100–113) obtained by LC-MS/MS and PepED is the N-terminal amino acid sequence of llama OIF obtained by internal Edmandegradation. Residues are numbered in black and follow the sequence of mouse NGF (PDB 1bet). Solid black arrows indicate β-strands. TT indicates tightturns. Strictly conserved residues are indicated by white letters on a red background. Conservatively substituted residues are indicated by red letters ona white background. The three disulfide bridges between Cys15 and Cys80, Cys58 and C108, and C68 and Cys110 are shown in green numbers. Loops arelabeled L1–L4 (blue). Stars indicate flexible loop 3.

15044 | www.pnas.org/cgi/doi/10.1073/pnas.1206273109 Ratto et al.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 6,

202

0

26–27 kDa (17). That OIF is NGF explains the paradoxical resultsof two previous studies regarding the molecular mass of thebioactive fraction of OIF. Seminal plasma filtered to fractionscontaining less than ∼30 kDa molecules failed to induced anovulatory response in llamas (3), yet the fraction digested to lessthan about 19 kDa by proteinase K (as determined by denaturingSDS/PAGE) retained ovulation-inducing activity (4). The apparentcontradiction may be attributed to the breakage of the homodimerinto monomers less than 19 kDa, by the denaturing conditions ofthe SDS-PAGE. Given the retention of bioactivity, it is unlikely thatproteinase K actually rendered seminal NGF into its monomers.Originally discovered in mouse sarcoma, cobra venom, and

submandibular salivary glands of adult mice, NGF has beencharacterized classically by its role in promoting survival andgrowth of sensory (dorsal root) and sympathetic neurons, andcells of the adrenal medulla (18). However, NGF has sub-sequently been identified in a variety of nonneuronal cells in-cluding tissues of both the male and female reproductive organs.Early purification experiments revealed that bovine seminalplasma is a rich source of NGF (19) and is likely produced pri-marily by the vesicular glands (20). It has also been detected inthe prostate gland of guinea pigs, rabbits, and bulls (19).

Identification of NGF and its receptor in the testis and epidid-ymis of mice, and in human and bovine ejaculated spermatozoahave led to the suggestion that NGF may play a role in spermviability and fertilizing ability (21, 22).The NGF protein and its receptor trkA have also been iden-

tified in the ovary of rodents (23), sheep (24), cattle (25), andhumans (26). Synthesis of NGF and its receptor were found inboth granulosa and theca cells in cows and women (25, 27).Vascular cell proliferation in cultured neonatal rat ovaries (ei-ther directly or through synthesis of vascular endothelial growthfactor) was induced by treatment with NGF (28), suggestinga role in the maintenance of follicular and luteal vasculature. Aswell, NGF stimulated estradiol secretion from cultured humangranulosa cells (directly and by increasing the formation of FSHreceptors (27)). Up-regulation of trkA and NGF mRNA wasdetected in granulosa and theca cells, respectively, at the firstpreovulatory surge of gonadotropins at puberty in rats, and theuse of NGF antibodies or a trkA blocker at the time of thepreovulatory LH surge inhibited ovulation (23). An increase inNGF and trkA in follicular cells after the preovulatory LH surgewas associated with disruption of gap junctions among follicularcells preceding ovulation (29).The abundance of NGF in seminal plasma and the effects of

seminal plasma on ovarian function strongly support the idea ofan endocrine mode of action (i.e., systemic distribution withdistant target tissues). This is in sharp contrast to the notion thatNGF is produced by target cells and has paracrine/autocrineactions (described above). Support for an endocrine mechanismis found in the following observations: (i) intramuscular admin-istration of seminal plasma resulted in ovulation in 33/35 (94%)llamas and alpacas compared with 0/35 (0%) given saline overfour separate in vivo studies (2, 8); (ii) ovulation rate was highestafter intramuscular administration of seminal plasma, in-termediate after intrauterine treatment with endometrial curet-tage (to mimic postcoital endometrial inflammation) (8), lowerafter intrauterine administration without curettage, and nil afterintrauterine administration of saline with or without curettage(93, 67, 24, and 0%, respectively) (2, 8); (iii) intramuscular orintrauterine administration of seminal plasma resulted in a surgein circulating concentrations of LH beginning within 1 h andpeaking within 3 h of treatment, followed by ovulation at 30 h(2); (iv) administration of purified OIF from seminal plasma hada dose-dependent effect on LH release, ovulation, and corpusluteum (CL) form and function, and the effect was evident at

Fig. 3. Protein structure of ovulation-inducing factor (OIF) from seminalplasma. (A) Monomer of OIF. β-Strands are labeled (β1–β8). Loops are labeledL1–L4. The three disulfide bridges between Cys15 and Cys80, Cys58 andC108, and C68 and Cys110 are shown in stick representation. (B) Biologicaldimer of OIF, rotated 90° with respect to A. OIF monomers are colored redand blue. (C) Cα superpositions of OIF (blue), mouse NGF (PDB code 1btg;red), and human NGF (1sg1; green). Superpositions of the three dimers re-veal high structural similarity between OIF and NGF. Loops of one of the twomonomers are labeled L1–L4.

Fig. 4. Purified OIF (fraction C2 isolated from llamaseminal plasma) induced neurite growth and NGF-specific receptor trkA in immortalized rat phaeo-chromocytoma (PC12) cells in vitro. PC12 cells after6 d of in vitro culture with (A) no treatment (nega-tive control), (B) treatment with 50 ng/mL ofrecombinant mouse NGF (positive control), or (C)treatment with 50 ng/mL of purified OIF (fraction C2

from llama seminal plasma). (D) Transcript expres-sion for the NGF-specific receptor trkA in PC12 cellsafter 6 d of in vitro culture with 50 ng/mL ofrecombinant mouse NGF or 50 ng/mL of purifiedOIF. MM: DNA ladder 100–1100 bp. Lanes 1 and 2:Control template without primers. Lane 3: trkAproduct in PC12 cells treated with rNGF. Lane 4: trkAproduct in PC12 cells treated with OIF. (E) Responsein mRNA for trkA in PC12 cells after 6 d of in vitroculture with 50 ng/mL recombinant mouse NGF(white bar) or 50 ng/mL of purified OIF (black bar).Mean ± SEM of three independent experiments. (aand b) Values with different superscripts are differ-ent (P < 0.01).

Ratto et al. PNAS | September 11, 2012 | vol. 109 | no. 37 | 15045

PHYS

IOLO

GY

Dow

nloa

ded

by g

uest

on

Nov

embe

r 6,

202

0

physiologically relevant doses (i.e., as little as 1/100th of thatpresent in an ejaculate) (10); (v) trkA receptors are present ingonadotrophs (30) and OIF (NGF) purified from seminal plasmainducedLH release from cultured pituitary cells of llamas and cows(31); and (vi) the in vivo LH- and ovulation-inducing effects ofOIFpurified from seminal plasma were ablated by pretreatment offemales with a GnRH antagonist (i.e., the hypothalamus is a targetfor OIF/NGF) (32).A mechanism of action that involves GnRH neurons in the

hypothalamus (32) implies that OIF/NGF in seminal plasmacrosses the blood–brain barrier of the inseminated female. Theimplication is consistent with results of studies in which I125-labeled NGF was detected in the hypothalamus after i.v. ad-ministration in mice (33), and in which serum NGF activatedthe hypothalamo–pituitary–adrenal axis in rats by acting withinthe CNS (34). The responsiveness of GnRH neurons to NGF isapparently unknown. The anterior pituitary is not protected bythe blood–brain barrier and support for the notion of a directeffect of OIF/NGF at the level of the pituitary is found in studiesin which LH release was elicited by seminal plasma treatment ofin vitro cultures of anterior pituitary cells of rats (35), llamas, andcattle (31). The NGF protein and its trkA receptor were detectedin 75 and 44%, respectively, of LH-containing gonadotrope cellsof the rat anterior pituitary (30). This is consistent with a pitui-tary route of action; however, it remains unclear whether the LHresponse of pituitary gonadotropes is the result of increasedsynthesis of LH, increased release, or both. Binding of OIF/NGFto trkA receptors induces phosphorylation of phospholipase C(PLC)-γ and Shc leading to activation of phosphatidylinositol3-kinase and mitogen-activated protein kinase, in turn leading toactivation of early growth response protein-1 (Egr-1). Expressionof Egr-1 (originally identified as a nerve growth factor responsegene product) (36) within the pituitary gland is limited to thegonadotrope and somatotrope subpopulation (30). As a memberof the gene family that recognizes the GC-rich nucleotide se-quence in the promoter region of the LHβ gene, Egr-1 has

a profound effect on LHβ gene expression (37), thus providinga plausible pathway for OIF/NGF regulation of LH secretion.Nerve growth factor is highly conserved among species (Fig.

2), consistent with results of recent studies that demonstrate thatOIF in seminal plasma is conserved among species (9, 11, 38).Not only was ovulation induced in llamas by conspecific andheterospecific seminal plasma (camelid, bull, stallion, boar, andrabbit), but administration of llama seminal plasma affectedovarian function in females of unrelated species (mice and cows)(11, 21). However, compared with llama seminal plasma, that ofbulls, stallions, and boars induced a lower incidence of ovulationin female llamas. This may be due to differing quantities orisoforms of OIF present in each species (Fig. 5) or to slightvariations in amino acid sequences affecting receptor binding(Figs. 2 and 3).Two isoforms of NGF have been isolated from mouse sub-

maxillary glands. One is a dimer of two identical subunits linkedtogether by noncovalent bounds with a molecular mass of 26kDa, known as 2.5S-NGF or β-NGF (18). The second is a highermolecular mass isoform known as 7S-NGF complex was origi-nally reported to have a molecular mass of 140 kDa, and iscomposed of three different protein structures termed α, β, and γ(39). The 2.5S-NGF is a product of proteolytic cleavage of the βsubunit from the larger 7S-NGF complex and is biologically ac-tive in stimulating neurite outgrowth. Early observations thathigher molecular weight forms of NGF also have biological ac-tivity (18) have been confirmed more recently. Isoforms of NGFhave been identified with molecular masses ranging from 16to 60 kDa (16, 40). Using human and mouse monoclonalantibodies against β-NGF, a 60-kDa pro-NGF was identified inseveral different commercial β-NGF preparations, as well as inrecombinant NGF and in dorsal root ganglia of rats (16). In thepresent study, a less distinct band appeared at just over 60 kDa inimmunoblots of whole seminal plasma of llamas and bulls(representative of two general categories of mammals consid-ered to be induced and spontaneous ovulators) and was inter-preted as an unprocessed proform of NGF. Proforms of NGFare most prominent in the central and peripheral nervous sys-tem; that is, little or no β-NGF is found in these tissues (41). Therelevance of a given NGF isoform and its relative stability,binding properties, and physiological effects on the reproductivesystem remain to be determined.

ConclusionsOvulation-inducing factor (OIF) is nerve growth factor (NGF)in seminal plasma. It is surprising that its effects in the femalewere not identified earlier, given the abundance of NGF inseminal plasma. This realization helps justify the existence ofthe elaborate male accessory gland system as more than anevolutionary vestige among species. Nerve growth factor ishighly conserved, and identification of OIF as NGF explainsrecent discoveries of the effects of seminal plasma on gonado-tropin release and ovarian function in a variety of species. Thepurification yield of the mature protein, β-NGF, from llamaseminal plasma is high, and purification of large quantitiespermitted the discovery that NGF from the ejaculate has animportant role in regulating gonadotropin release and ovarianfunction in the female. The notion of an endocrine route ofaction of NGF on reproductive function is unique and eluci-dates a direct pathway for the influence of the male on thehypothalamo–pituitary–gonadal axis of the female. Identifica-tion of OIF as NGF in llama seminal plasma representsa unique sequence for the family Camelidae and the onlyseminal plasma-derived NGF to be isolated, characterized, andfully sequenced in any species. Crystallographic study of purifiedOIF from seminal plasma provided the opportunity to de-termine the structure of natural NGF (described previously onlyin the mouse, human, and cow) and has laid the foundation for

Fig. 5. Immunoblot analysis of the seminal plasma of llamas and bulls witha polyclonal mouse anti-NGF. Negative control: cytochrome C (300 ng).Positive control: recombinant mouse NGF (300 ng). Fraction C2: OIF purifiedfrom llama seminal plasma (300 ng). Whole llama seminal plasma (800 ngtotal protein). Whole bovine seminal plasma (0.8, 1.6, 3.2, 4.8, and 6.4 μgtotal protein, respectively).

15046 | www.pnas.org/cgi/doi/10.1073/pnas.1206273109 Ratto et al.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 6,

202

0

examining structure–function relationships of proforms of NGFand receptor recognition sites. Purification of the mature andproforms of the protein in quantity from seminal plasma willenable in vivo study of the role of NGF in reproductive andneurologic health and disease.

Materials and MethodsEjaculates were collected from llamas by artificial vagina and from bulls byelectroejaculation. The seminal plasmawas separated by centrifugation fromthe cellular portion of the ejaculate and samples were stored frozen. Seminalplasma samples were pooled within species after thawing for the purposes ofdetermining the identity of OIF. Llama seminal plasma was loaded on a hy-droxylapatite chromatography column and three protein fractions (desig-nated A, B, and C) were eluted. In fraction C, a major 14-kDa protein detectedby SDS/PAGEwas subsequently loaded on a gel filtration column yielding twosubfractions, C1 and C2. A purified protein band cut from SDS/PAGE offraction C2 was used to determine molecular mass by MALDI-TOF and pro-tein identification by LC-MS/MS. Samples of the nondenatured fraction of C2

were loaded on a calibrated Superdex-75 HiLoad 16/60 gel filtration columnto confirm the molecular mass. Sequences of tryptic (LC-MS/MS) and non-denatured peptides (Edman degradation) derived from fraction C2 werecompared with known protein sequences listed in the database of the Eu-ropean Bioinformatics Institute. To determine the protein structure,

a sample of OIF was screened against 1,536 mixture solutions, and conditionswere optimized to produce crystals for X-ray diffraction. Diffraction datawere collected on beamline 08ID-1 and the structure was solved by molec-ular replacement using murine β-NGF structure as the search model. Neuritedevelopment in immortalized rat PC12 was used as an in vitro bioassay forNGF-like properties of OIF. Up-regulation of an NGF-specific receptor (i.e.,mRNA for trkA) was compared in PC12 cells cultured in vitro with NGF vs. OIF.Samples of fraction C2 of llama seminal plasma, as well as whole seminalplasma from llamas and bulls were examined by Western blot analysis usinga polyclonal mouse anti-NGF as the primary antibody. Lastly, the ovulatoryresponse of female llamas was examined by ultrasonography after treat-ment with OIF (fraction C2 of llama seminal plasma), β-NGF (from mousesubmandibular glands), or saline. (An expanded version of Materials andMethods can be found in SI Materials and Methods.)

ACKNOWLEDGMENTS. We thank Orleigh Bogle, Manuel Palomino, andMiriam Cervantes for assistance with animal data collection and Lata Prasadfor assistance with X-ray diffraction data collection. This research wasperformed at the University of Saskatchewan and supported by grants fromthe Natural Sciences and Engineering Research Council of Canada, the AlpacaResearch Foundation, the Chilean National Science and Technology ResearchCouncil (Fondecyt 1120518), the Saskatchewan Health Research Foundation,and the Canadian Institutes of Health Research. X-ray crystallography wasdone at the Canadian Light Source, Saskatoon, Saskatchewan.

1. Mann T (1964) Biochemistry of Semen and of the Male Reproductive Tract (Butler,Tanner Ltd, Frome, UK), p 493.

2. Adams GP, Ratto MH, Huanca W, Singh J (2005) Ovulation-inducing factor in theseminal plasma of alpacas and llamas. Biol Reprod 73:452–457.

3. Ratto MH, Huanca W, Adams GP (2010) Ovulation-inducing factor: A protein com-ponent of llama seminal plasma. Reprod Biol Endocrinol 8:44.

4. Ratto MH, Delbaere LTJ, Leduc YA, Pierson RA, Adams GP (2011) Biochemical isolationand purification of ovulation-inducing factor (OIF) in seminal plasma of llamas. Re-prod Biol Endocrinol 9:24.

5. Bedford JM (2004) Enigmas of mammalian gamete form and function. Biol Rev CambPhilos Soc 79:429–460.

6. Chen BX, Yuen ZX, Pan GW (1985) Semen-induced ovulation in the bactrian camel(Camelus bactrianus). J Reprod Fertil 74:335–339.

7. Xu YS, Wang HY, Zeng GQ, Jiang GT, Gao YH (1985) Hormone concentrations beforeand after semen-induced ovulation in the bactrian camel (Camelus bactrianus).J Reprod Fertil 74:341–346.

8. Ratto MH, Huanca W, Singh J, Adams GP (2005) Local versus systemic effect of ovu-lation-inducing factor in the seminal plasma of alpacas. Reprod Biol Endocrinol 3:29.

9. Ratto MH, Huanca W, Singh J, Adams GP (2006) Comparison of the effect of ovula-tion-inducing factor (OIF) in the seminal plasma of llamas, alpacas, and bulls. Ther-iogenology 66:1102–1106.

10. Tanco VM, Ratto MH, Lazzarotto M, Adams GP (2011) Dose-response of female llamasto ovulation-inducing factor from seminal plasma. Biol Reprod 85:452–456.

11. Bogle OA, Ratto MH, Adams GP (2011) Evidence for the conservation of biologicalactivity of ovulation-inducing factor in seminal plasma. Reproduction 142:277–283.

12. Tanco VM, Van Steelandt MD, Ratto MH, Adams GP (2012) Effect of purifiedllama ovulation-inducing factor (OIF) on ovarian function in cattle. Theriogenology,10.1016/ j.theriogenology.2012.03.036.

13. McDonald NQ, et al. (1991) New protein fold revealed by a 2.3-A resolution crystalstructure of nerve growth factor. Nature 354:411–414.

14. Bax B, Blundell TL, Murray-Rust J, McDonald NQ (1997) Structure of mouse 7S NGF: Acomplex of nerve growth factor with four binding proteins. Structure 5:1275–1285.

15. Greene LA, Tischler AS (1976) Establishment of a noradrenergic clonal line of ratadrenal pheochromocytoma cells which respond to nerve growth factor. Proc NatlAcad Sci USA 73:2424–2428.

16. Reinshagen M, et al. (2000) Commercial recombinant human beta-nerve growthfactor and adult rat dorsal root ganglia contain an identical molecular species ofnerve growth factor prohormone. J Neurochem 74:2127–2133.

17. Kolbeck R, Jungbluth S, Barde YA (1994) Characterisation of neurotrophin dimers andmonomers. Eur J Biochem 225:995–1003.

18. Angeletti RH, Bradshaw RA (1971) Nerve growth factor from mouse submaxillarygland: Amino acid sequence. Proc Natl Acad Sci USA 68:2417–2420.

19. Harper GP, Glanville RW, Thoenen H (1982) The purification of nerve growth factorfrom bovine seminal plasma. Biochemical characterization and partial amino acidsequence. J Biol Chem 257:8541–8548.

20. Hofmann HD, Unsicker K (1982) The seminal vesicle of the bull: A new and very richsource of nerve growth factor. Eur J Biochem 128:421–426.

21. Ayer-LeLievre C, Olson L, Ebendal T, Hallböök F, Persson H (1988) Nerve growth factormRNA and protein in the testis and epididymis of mouse and rat. Proc Natl Acad SciUSA 85:2628–2632.

22. Li C, et al. (2010) Detection of nerve growth factor (NGF) and its specific receptor

(TrkA) in ejaculated bovine sperm, and the effects of NGF on sperm function. Ther-

iogenology 74:1615–1622.23. Dissen GA, et al. (1996) A role for trkA nerve growth factor receptors in mammalian

ovulation. Endocrinology 137:198–209.24. Barboni B, et al. (2002) Preovulatory rise of NGF in ovine follicular fluid: Possible in-

volvement in the control of oocyte maturation. Microsc Res Tech 59:516–521.25. Dissen GA, et al. (2000) Direct effects of nerve growth factor on thecal cells from

antral ovarian follicles. Endocrinology 141:4736–4750.26. Abir R, et al. (2005) Presence of NGF and its receptors in ovaries from human fetuses

and adults. Mol Hum Reprod 11:229–236.27. Salas C, et al. (2006) Nerve growth factor-dependent activation of trkA receptors in

the human ovary results in synthesis of follicle-stimulating hormone receptors and

estrogen secretion. J Clin Endocrinol Metab 91:2396–2403.28. Julio-Pieper M, Lara HE, Bravo JA, Romero C (2006) Effects of nerve growth factor

(NGF) on blood vessels area and expression of the angiogenic factors VEGF and

TGFbeta1 in the rat ovary. Reprod Biol Endocrinol 4:57.29. Mayerhofer A, et al. (1996) Involvement of nerve growth factor in the ovulatory

cascade: trkA receptor activation inhibits gap junctional communication between

thecal cells. Endocrinology 137:5662–5670.30. Patterson JC, Childs GV (1994) Nerve growth factor and its receptor in the anterior

pituitary. Endocrinology 135:1689–1696.31. Bogle OA, Ratto MH, Adams GP (2012) Ovulation-inducing factor (OIF) induces LH

secretion from pituitary cells. Anim Reprod Sci, 10.1016/j.anireprosci.2012.06.006.32. Silva ME, et al. (2011) Cetrorelix suppresses the preovulatory LH surge and ovulation

induced by ovulation-inducing factor (OIF) present in llama seminal plasma. Reprod

Biol Endocrinol 9:74.33. Loy R, Taglialatela G, Angelucci L, Heyer D, Perez-Polo R (1994) Regional CNS uptake

of blood-borne nerve growth factor. J Neurosci Res 39:339–346.34. Taglialatela G, Perez-Polo JR (1994) Developmental profile of the hypothalamo-pi-

tuitary-adrenal axis response to nerve growth factor. Neurosci Lett 182:231–234.35. Paolicchi F, Urquieta B, Del Valle L, Bustos-Obregón E (1999) Biological activity of the

seminal plasma of alpacas: Stimulus for the production of LH by pituitary cells. Anim

Reprod Sci 54:203–210.36. Milbrandt J (1987) A nerve growth factor-induced gene encodes a possible tran-

scriptional regulatory factor. Science 238:797–799.37. Brown P, McNeilly AS (1999) Transcriptional regulation of pituitary gonadotrophin

subunit genes. Rev Reprod 4:117–124.38. Silva ME, et al. (2011) Is an ovulation-inducing factor (OIF) present in the seminal

plasma of rabbits? Anim Reprod Sci 127:213–221.39. Varon S, Normura J, Shooter EM (1968) Reversible dissociation of the mouse nerve

growth factor protein into different subunits. Biochemistry 7:1296–1303.40. Paoletti F, et al. (2011) Conformational plasticity of proNGF. PLoS ONE 6(7): e22615.

Doi:10.1371/journal.pone.0022615.41. Pedraza CE, et al. (2005) Pro-NGF isolated from the human brain affected by Alz-

heimer’s disease induces neuronal apoptosis mediated by p75NTR. Am J Pathol 166:

533–543.

Ratto et al. PNAS | September 11, 2012 | vol. 109 | no. 37 | 15047

PHYS

IOLO

GY

Dow

nloa

ded

by g

uest

on

Nov

embe

r 6,

202

0