Identification of a Testicular Odorant Receptor Mediating Human Sperm Chemotaxis

See allHide authors and affiliations

Science  28 Mar 2003:
Vol. 299, Issue 5615, pp. 2054-2058
DOI: 10.1126/science.1080376


Although it has been known for some time that olfactory receptors (ORs) reside in spermatozoa, the function of these ORs is unknown. Here, we identified, cloned, and functionally expressed a previously undescribed human testicular OR, hOR17-4. With the use of ratiofluorometric imaging, Ca2+ signals were induced by a small subset of applied chemical stimuli, establishing the molecular receptive fields for the recombinantly expressed receptor in human embryonic kidney (HEK) 293 cells and the native receptor in human spermatozoa. Bourgeonal was a powerful agonist for both recombinant and native receptor types, as well as a strong chemoattractant in subsequent behavioral bioassays. In contrast, undecanal was a potent OR antagonist to bourgeonal and related compounds. Taken together, these results indicate that hOR17-4 functions in human sperm chemotaxis and may be a critical component of the fertilization process.

More than a decade ago, about a thousand vertebrate genes were found that code for olfactory receptor proteins. These receptors are coupled to complex signaling pathways, and despite their name, they also reside in tissues other than those involved in olfaction. Several distinct ORs are expressed predominantly or exclusively in human spermatogenic cells (1, 2). Immunocytochemistry indicates that receptor proteins are localized to the sperm flagellar midpiece (2). These observations have led to speculation that ORs function in chemosensory signaling pathways, and hence in direct sperm chemotaxis (3–5). Here, we determined the ligand specificity and functional importance of hOR17-4, a newly identified testicular OR. Sequential studies were conducted at the molecular, cellular, and behavioral levels to establish the role of this OR in human sperm physiology and behavior.

ORs constitute 17 different gene families and occur on all human chromosomes except 20 and Y (6). Initially, we analyzed testicular expression patterns for most members of the OR gene cluster on human chromosome 17 [17p13.3 (7)] by means of reverse transcription polymerase chain reaction (RT-PCR) with testis cDNA (8). Of all candidate receptors tested, only hOR17-2 and hOR17-4 [for nomenclature, see (9)] were found to be expressed in human testis (Fig. 1A). The expression of hOR17-2 in testis had been demonstrated by Parmentier and colleagues (1), whereas hOR17-4 represents an undescribed human testicular odorant receptor.

Figure 1

hOR17-4 is expressed in testicular tissue and can be activated by odorants. (A) RT-PCR detection of OR expression in human testis. OR17-2 and -4 were detected in testicular cDNA (upper panel). Genomic DNA served as positive control (lower panel); mRNA before cDNA synthesis did not yield any PCR product (size in bp) (20). (B and C) In a randomly selected field of view, the complex odorant mixture Henkel 100 induced transient Ca2+ signals in transfected HEK293 cells (B, red traces). Cells not transfected or not expressing hOR17-4 did not respond to cyclamal (black traces). Adenosine triphosphate (ATP) served as a control of HEK cells' excitability. The integrated fluorescence ratio (f 340/f 380) for various fura-2–loaded cells is shown as a function of time. (C) Ca2+ changes in individual cells are indicated in pseudocolors. Both Henkel 100 (1:1000) and ATP (200 μM) were applied for 5 s.

The putative binding pocket of an OR occurs within the third to sixth transmembrane (TM III–VI) domain (10). In the present study, we inserted a cloned TM II–VII segment of hOR17-4 into a shuttle expression vector (pOLF-express) backbone based on the N- and C-terminal regions of the recently described helional receptor hOR17-40 (8, 11) (fig. S1). The resulting chimeric protein and full-length hOR17-4 should thus have the same ligand specificity.

We and other groups previously reported the functional expression of recombinant OR proteins of different species (11–14). Accordingly, we transiently expressed hOR17-4 in HEK293 cells (8). Receptor activation in the recombinant system leads to a transient inositol 1,4,5-trisphosphate (IP3)–mediated increase in cytosolic Ca2+ that can be detected by ratiofluorometric imaging techniques (8, 11).

A given OR can theoretically respond to an extensive stimulus repertoire. Thus, a complex mixture of 100 compounds (Henkel 100), including aromatic and short-chain aliphatic hydrocarbons, was used for initial ligand screening. Henkel 100, diluted 1:1000 in Ringer's solution, induced transient Ca2+ responses in 2 to 5% of all cells tested (Fig. 1B). This induction rate is typical for transient OR transfection (11). In untransfected HEK cells, Ca2+ signals were not observed, even with a factor of 10 increase in mixture concentration. By subdivision into smaller fractions, we ultimately found only one active ligand in the Henkel 100 mixture: cyclamal (fig. S2) (8). The molecular receptive field of hOR17-4 was determined using cyclamal as a template. Several structurally related molecules were tested singly at 500 μM for Ca2+ induction response (Fig. 2A) (8).

Figure 2

The molecular receptive field of hOR17-4. (A) Structurally related molecules were individually tested for their ability to activate hOR17-4, using cyclamal as a template. Effective ligands, collectively characterized by the presence of an aldehyde group connected to an aromatic ring via a carbon chain of two to four carbons, are shown in a green field; inactive compounds are in a magenta field [3-PPA, 3-phenylpropionic aldehyde; 4-PBA, 4-phenylbutyraldehyde; 5-PVA, 5-phenylvaleraldehyde; PAA, phenylacetaldehyde; 3-PBA, 3-phenylbutyraldehyde; 4-BPAA, (4-tert-butylphenoxy) acetaldehyde]. (B toD) Representative ratiofluorometric recordings of pOR17-4–transfected HEK293 cells. The cytosolic Ca2+ level of different fura-2–loaded cells is depicted as the integrated fluorescence ratio (f 340/f 380) and viewed as a function of time. If not indicated by an application bar, all tested compounds (500 μM) and ATP (200 μM) were applied for 5 s. (B) Henkel Mix H 10/2, including bourgeonal and undecanal, failed to activate cells expressing hOR17-4. (C) Undecanal (preincubation for 50 s) inhibited hOR17-4 receptor activation by bourgeonal. (D) Henkel Mix H 10/2 lacking undecanal activated hOR17-4–expressing HEK293 cells.

Surveying the range of active and inactive odorants, we conclude that the presence of an aldehyde group connected to an aromatic ring via a carbon chain of defined length (two to four carbons) is a key determinant for an effective hOR17-4 ligand. The presence or absence of nonpolar methyl groups is largely tolerated, whereas addition of polar structures and double bonds can effectively decrease the ability of a ligand to activate the receptor.

A rank ordering of the effective odorants (figs. S3 and S4) shows that ligands stimulate the smallest Ca2+ response amplitudes when they lack a molecular group in the para position. Notably, bourgeonal, canthoxal, and lilial were more potent than cyclamal as agonists to transfected HEK cells. It is a common feature of ORs that they are activated by multiple ligands with different receptor affinities (15), and testicular ORs appear to be no exception.

Identification of a receptor antagonist emerged from studies of Ca2+ responses to odorant mixtures. Bourgeonal was part of Henkel Mix H 10/2 (fig. S2), one of the 10 compound mixtures that failed to activate hOR17-4 (Fig. 2B). Coincubation of bourgeonal with one constituent (undecanal), but not the other eight H 10/2 compounds, suppressed the Ca2+ response (Fig. 2, C and D). Because application of undecanal alone did not stimulate a Ca2+response, inhibitory effects on bourgeonal may result from competition for the binding pocket.

To rule out the possibility that the above responses were associated only with the chimeric construct, we cloned full-length hOR17-4 (8) and tested its sensitivity to bourgeonal and undecanal. Whereas bourgeonal activated full-length hOR17-4 (fig. S5A), this response was inhibited by undecanal (fig. S5B). These results validate the receptor TM II–VII region as a ligand-binding pocket. In contrast, experiments with HEK cells expressing hOR17-40 (11) showed no effects of bourgeonal or undecanal (fig. S6). In their opposing actions, bourgeonal and undecanal therefore uniquely targeted hOR17-4.

Having established the molecular receptive fields of cloned chimera and full-length hOR17-4, we next explored receptor function in a natural system: human sperm. Human testicular OR may control a Ca2+-mediated physiological process, for example, by causing an acrosome reaction, capacitation, or hyperactivated motility (16–19). Bourgeonal induced repeatable Ca2+responses in 36% of examined cells [43 of 119 spermatozoa in 12 experiments (8)]. This compound was the most potent ligand tested (fig. S7), and prolonged application led to sustained Ca2+ signals (Fig. 3A) (8). A ranking of ligands according to magnitude of Ca2+ response was similar for spermatozoa and hOR17-4–expressing HEK cells (20). Moreover, spermatozoa detected bourgeonal, and other ligands, at very dilute concentrations (≥10−8 M), about two orders of magnitude lower than those required by the recombinant receptor (Fig. 3D). Coapplication of equimolar undecanal and bourgeonal completely blocked Ca2+transients (Fig. 3B, n = 10 cells, three experiments). Together, these data strongly suggest that hOR17-4 is expressed on the surface of human spermatozoa and that signal transduction is a Ca2+-mediated process.

Figure 3

Representative ratiofluorometric recordings of the head and midpiece of individual human spermatozoa. The cytosolic Ca2+ level of various fura-2–loaded cells is depicted as the integrated fluorescence ratio (f 340/f 380) and viewed as a function of time. Unless indicated by an application bar, MDL-12,330A (50 μM) and all other components (5 μM) were applied for 5 s. (A) Bourgeonal induces repeatable and stimulus-correlated Ca2+ responses in spermatozoa. (B) Coapplication of undecanal suppresses responses to bourgeonal. (C) The bourgeonal-induced Ca2+ signal is cAMP-mediated as MDL-12,330A blocks bourgeonal sensitivity in more than 70% of all spermatozoa. (D) The percentage of sperm showing Ca2+ signals in response to bourgeonal is dose-dependent [mean (±SEM) populations of responsive sperm]. Inset shows a fluorescence image of a fura-2–loaded spermatozoon. (E) Accumulation of human sperm in microcapillary tubes presenting an ascending, uniform, or descending chemical concentration gradient. Trials were performed using bourgeonal concentrations of 10−9 to 10−5 M and HTF as the control. Plotted values are mean (±SEM) cell densities.

Pharmacological studies were then used to investigate the primary signal transduction pathway activated by bourgeonal. Inhibitors of adenylate cyclase [MDL-12,330A (50 μM)] or phospholipase C [U73122 (50 μM)] were coapplied with bourgeonal. Results revealed that adenylate cyclase, but not phospholipase, plays a key role in hOR17-4–mediated signaling in human sperm. Most (72%) spermatozoa responses to bourgeonal were inhibited by MDL-12,330A (Fig. 3C), and reactions required the presence of extracellular Ca2+(20). Numerous functions are regulated by adenosine 3′,5′-monophosphate (cAMP)–dependent mechanisms (19), so these data forecast a physiological role of hOR17-4 expression in human spermatozoa.

To determine the behavioral implications of our findings, we investigated the effects of bourgeonal on human sperm swimming. An initial experiment tested for the accumulation of cells in microcapillaries placed within ascending, uniform, or descending chemical gradients (8). Results were used to distinguish between mechanisms of chemotaxis (directed movement with respect to a chemical concentration gradient), chemokinesis (change in swim speed), and cell trapping (arrest of swimming) (21). In both the ascending and uniform gradients, cell accumulation within capillaries increased significantly as a function of dose (Fig. 3E) [two-way analysis of variance (ANOVA): F 4, 281 = 30.01, P < 0.0001]. Moreover, densities of cells in capillaries within an ascending gradient were significantly higher than those within a uniform or descending gradient (two-way ANOVA:F 2, 281 = 164.28, P < 0.0001). Finally, at each effective concentration (10−8 to 10−5 M), accumulation of cells in capillaries was ranked in the following order, according to chemical gradient: ascending ≫ uniform > descending. These collective results are consistent only with a chemotaxis mechanism—that is, directed movement toward a region of locally elevated bourgeonal concentration (21).

During a second series of microcapillary assays involving an ascending attractant gradient, sperm swimming speeds and paths were measured by computer-assisted video motion analysis [CAVMA (8)]. This combination of bioassays and motion analysis clearly showed that sperm cells were navigating within a chemical gradient and toward the capillary tip. Human sperm responded to bourgeonal in a dose-dependent manner, exhibiting elevated swim speeds (Fig. 4A), directed movements (Fig. 4A), and higher cell densities in capillaries (fig. S8). As in our pharmacological studies of hOR17-4 function, behavioral responses to bourgeonal required extracellular Ca2+ (fig. S9). Bourgeonal concentrations of 10−8 M or higher evoked significant responses relative to the human tubal fluid (HTF) control (Figs. 3E and 4A; P < 0.05). As an alternative way of distinguishing chemotaxis from chemokinesis, this experiment thus reveals the dual role of bourgeonal as both a potent chemoattractant and an effective swimming stimulant to navigating sperm cells.

Figure 4

Swimming behavior of human sperm imaged near a microcapillary tip (inside diameter, 316 μm) by CAVMA. Sperm were placed in an ascending gradient of (A) 10−9 to 10−5 M bourgeonal (or HTF as control), or (B) mixtures that combined 10−7 M bourgeonal with 10−9, 10−8, or 10−7 M undecanal. Controls tested HTF, 10−7 M bourgeonal, and 10−7 M undecanal. For brevity, data are not shown here for sperm responses to 10−5 M bourgeonal or to a mixture of 10−7 M bourgeonal and 10−7 M undecanal. Open circles correspond to video images captured at intervals of 0.033 s; arrowheads indicate directions of travel for individual cells. A minimum of 35 paths were analyzed for each test or control solution. To eliminate selection bias, we used a random number generator to choose representative paths from the entire pool. In contrast, complete data sets—not representative paths—were used to calculate mean (±SEM) swimming speeds (w), mean angles (θ), and mean vector lengths (r). The angle of sperm orientation was measured with respect to an origin (0°) defined as the shortest tangent between each individual cell and the capillary tip. A mean vector length of 1 indicates that all cells swim in a single, common direction; a vector length of 0 indicates random motion. To determine whether cell movement within a chemical gradient was significant (P), we used Rayleigh's test to compare each mean direction against a uniform circular distribution.

Undecanal was tested as a potential inhibitor of bourgeonal. When 10−7 M bourgeonal was combined with 10−8 or 10−7 M undecanal, swim speeds (Fig. 4B), directed movements (Fig. 4B), and cell densities (fig. S8) were not significantly elevated relative to the HTF control (P> 0.50). Moreover, undecanal (10−7 M) alone had no measurable impact on any aspect of sperm behavior. Thus, undecanal strongly inhibited bourgeonal effects on sperm navigation and swim speeds.

In conclusion, these data suggest that hOR17-4 signaling potentially governs chemical communication between sperm and egg. Thus, the hOR17-4 signaling system could be used to manipulate fertilization, with important consequences for contraception and procreation.

Supporting Online Material

Materials and Methods

SOM Text

Figs. S1 to S9


  • * To whom correspondence should be addressed. E-mail: hanns.hatt{at}


View Abstract

Stay Connected to Science

Navigate This Article