Chemosensory Ca2+ dynamics correlate with diverse behavioral phenotypes in human sperm.

In the female reproductive tract, mammalian sperm undergo a regulated sequence of prefusion changes that “prime” sperm for fertilization. Among the least understood of these complex processes are the molecular mechanisms that underlie sperm guidance by environmental chemical cues. A “hard-wired” Ca 2 (cid:2) signaling strategy that orches-trates specific

In the female reproductive tract, mammalian sperm undergo a regulated sequence of prefusion changes that "prime" sperm for fertilization. Among the least understood of these complex processes are the molecular mechanisms that underlie sperm guidance by environmental chemical cues. A "hard-wired" Ca 2؉ signaling strategy that orchestrates specific motility patterns according to given functional requirements is an emerging concept for regulation of sperm swimming behavior. The molecular players involved, the spatiotemporal characteristics of such motility-associated Ca 2؉ dynamics, and the relation between a distinct Ca 2؉ signaling pattern and a behavioral sperm phenotype, however, remain largely unclear. Here, we report the functional characterization of two human sperm chemoreceptors. Using complementary molecular, physiological, and behavioral approaches, we comparatively describe sperm Ca 2؉ responses to specific agonists of these novel receptors and bourgeonal, a known sperm chemoattractant. We further show that individual receptor activation induces specific Ca 2؉ signaling patterns with unique spatiotemporal dynamics. These distinct Ca 2؉ dynamics are correlated to a set of stimulus-specific stereotyped behavioral responses that could play vital roles during various stages of prefusion sperm-egg chemical communication.
On their journey to locate the oocyte, navigating mammalian sperm depend on both chemical (1) and physical (2,3) cues to regulate flagellar motion, switch flagellar beat modes, and, thus, direct their movement. How such environmental signals are detected and translated into changes in motility, however, remains largely unknown.
Sperm acquire motility upon vaginal deposition (4). This initial swimming behavior is referred to as activated motility and characterized by a relatively low amplitude/high frequency sinusoidal flagellar motion (5). Motile sperm that pass the cervical mucus barrier must locate and enter the oviduct. Only a fraction of sperm present in the uterus (ϳ10% in humans (6)) accomplishes this task. Within the oviductal isthmus, sperm frequently attach to the highly convoluted epithelial surface (3,7,8), forming a sperm reservoir in close proximity to the fertilization site. Upon ovulation, capacitated sperm that adopt a hyperactivated state (i.e. displaying asymmetrical and relatively high amplitude/low frequency flagellar beating) (4), generate sufficient propulsion force to detach from the epithelium, enter the ampulla, and eventually penetrate the cumulus and zona pellucida surrounding the egg (9).
Given the small fraction of sperm that reaches the oviduct, a "competitive race" scenario has fallen out of favor in recent years. Instead, effective sperm guidance mechanisms are required to assure a synchronized arrival and encounter of both gametes at the fertilization site (1). Current models propose a complex multistep process of sperm navigation along thermal (2) and chemical (10) gradients. Accordingly, chemical guidance cues are secreted by both the oocyte and the surrounding cumulus cells after ovulation (11). In addition to these chemoattractants, further unidentified chemosignals may participate in other prefusion processes that prime sperm for fertilization (12).
In contrast to the far more advanced knowledge about marine invertebrate sperm chemotaxis (13)(14)(15), the ligands and receptors involved in mammalian sperm-oocyte communication are largely unknown (1,3,7). In this context, we and others (16 -20) recently proposed a functional role of sperm chemosensory receptors that, based on sequence homology analysis, are members of the G protein-coupled receptor superfamily of odorant receptor (OR) 4 proteins (21,22). Many laboratories have consistently reported OR expression in mammalian male reproductive tissues, pre-and postmeiotic germ cells, and mature spermatozoa (16 -19, 23-36). In recent microarray surveys, up to 83 different human OR genes were found expressed in testes (28,36), and several of these testicular ORs appear to have evolved under stronger evolutionary constraint than OR genes expressed exclusively in the olfactory epithelium (36,37). So far, however, only two such receptors (OR1D2 (17) in humans and Olfr16 (19) in mice, respectively) have been attributed physiological functions as mediators of sperm chemotaxis. OR1D2 is activated in vitro by the synthetic odor molecule bourgeonal but antagonized by the unrelated aldehyde undecanal (17, 38 -40). When human sperm are exposed to an ascending bourgeonal gradient, a fraction of cells displays robust chemotactic behavior (1,41), which is abolished by undecanal (16,17). Analogous to peptide-mediated sperm chemotaxis in marine invertebrates (15), bourgeonal-triggered cytosolic Ca 2ϩ signals are believed to control changes in flagellar beating that underlie chemotactic swimming behavior in human sperm.
Playing a pivotal role in the vast majority of cellular signaling processes, the cytosolic Ca 2ϩ concentration is subject to tight spatiotemporal control (42,43) in both somatic and germ cells. In sperm, the restricted cytoplasmic volume and highly polarized morphology pose a unique challenge for the cell to orchestrate discrete Ca 2ϩ -sensitive responses and maintain Ca 2ϩ signaling specificity (3,44). In addition, developing sperm shed most intracellular membranous organelles during spermiogenesis, retaining only the acrosome, the redundant nuclear envelope, and the mitochondria as potential Ca 2ϩ stores (45). Thus, some components of the somatic cell Ca 2ϩ signaling "tool kit" are not available to control Ca 2ϩ dynamics and coordinate the flagellar waveform. However, sperm appear to employ a "hardwired" Ca 2ϩ signaling strategy in which a given stimulus generates a distinct Ca 2ϩ signal in separate microdomains and little if any input integration occurs (44). This way, specific Ca 2ϩregulated motility patterns (e.g. chemotactic turns, hyperactivation, etc.) are initiated according to the cell's functional requirements. The spatiotemporal characteristics of such motility-associated Ca 2ϩ dynamics in human sperm, however, remain unclear.
Here, we report functional characterization of two human sperm chemoreceptors. We identify receptor-specific agonists and comparatively describe physiological Ca 2ϩ responses in sperm to these novel stimuli and bourgeonal, a known sperm chemoattractant. We further show that individual receptor activation induces Ca 2ϩ signals with unique spatiotemporal dynamics. These distinct Ca 2ϩ mobilization patterns are correlated to a set of stimulus-specific stereotyped behavioral responses that could mediate chemical communication during various prefusion stages.

Quantitative Real-time RT-PCR Analysis
Quantitative TaqMan RT-PCR analysis was performed using the Applied Biosystems PRISM 7900 sequence detection system. Human tissue mRNA probes were obtained from Ambion, Inc. (Austin, TX), BioChain Institute, Inc. (Hayward, CA), Clontech Laboratories (Mountain View, CA), and Stratagene (La Jolla, CA). DNase I digestion was performed to remove residual genomic DNA, and mRNAs were reverse transcribed using random hexamer primers. Specific odorant receptor probes were carefully designed, and comparable probe efficiencies were assured by amplification of genomic DNA. Results were normalized to ␤-actin controls, and relative expression (arbitrary units) was calculated according to the following formula: relative expression ϭ 2 (18 Ϫ (Ct(probe) Ϫ Ct(actin))) . The threshold parameter Ct is defined as the cycle number at which the amplification plot passes a fixed threshold above base line. The primers and fluorescent probes used are described in supplemental Table S1.

Sperm Preparation
Motility Analysis Using Computer-assisted Video Motion Analysis (CAVMA)-Frozen semen samples of healthy donors were prepared for bioassays using a two-wash protocol (48), according to World Health Organization guidelines. After bringing suspensions to 37°C over 10 min, each sperm sample was adjusted to 10 6 cells/ml, placed in HEPES-buffered human tubal fluid (HTF; Irvine Scientific, Santa Ana, CA), and incubated at 37°C in a 5% CO 2 atmosphere for 30 min.
Motility Analysis Using Laser Tweezers-Frozen human semen samples from several healthy donors were prepared for bioassays as described above. Sperm was suspended in HEPESbuffered HTF (osmolarity ϳ280 mosmol/liter, pH 7.4). A dilution of ϳ30,000 sperm/ml was transferred into a Rose tissue culture chamber and mounted on a microscope stage. The sample was kept at 37°C using an air curtain incubator (ASI 400 Air Stream Incubator, NEVTEK, Burnsville, VA). A thermocouple was attached to the Rose chamber to ensure temperature stability.

Single Cell Ca 2؉ Imaging Analysis
Glass-bottom dishes, containing either human sperm adhered to concanavalin A coating or cultured HEK293 cells (loaded with fura-2/AM; 3 M; 30 min), were transferred to the stage of an inverted microscope equipped for ratiometric live cell imaging (Olympus IX71) with a 150-watt xenon arc lamp, a motorized fast change filter wheel illumination system (MT20, Olympus) for multiwavelength excitation, a 12-bit 1376 ϫ 1032-pixel charge-coupled device (CCD) camera (F-View II, Olympus), and cellR imaging software (Olympus). Cells in a randomly selected field of view were viewed at ϫ60 magnification using an oil immersion objective (UPLSAPO; numerical aperture ϭ 1.35; Olympus) and illuminated sequentially at 340 and 380 nm at variable cycle times (sperm, 200 or 1000 ms, respectively; HEK293, 1000 ms). The average pixel intensity within user-selected regions of interest was digitized and stored on a PC. Ca 2ϩ -dependent fluorescence signals at 510 nm were calculated as the F 340 /F 380 intensity ratio. Stimuli were applied using a custom-made pressure-driven multibarrel perfusion system that allows for instantaneous solution change and focal application. Single sperm imaging experiments were repeated multiple times using cells from at least four different donors. In recombinant OR characterization experiments that tested for a specific stimulus mixture or a single agonist, Ca 2ϩ imaging measurements were performed repeatedly (n Ն 30) using cells from at least three independent transfections (in parallel to untransfected controls).

Microcapillary Bioassays
Microcapillary bioassays were performed as described previously (6,17) to distinguish between mechanisms of chemotaxis, chemokinesis, and other processes that might cause sperm accumulation (1). Briefly, two opposing wells containing fresh sperm suspensions (10 6 cells/ml) were connected by a 6-l microcapillary tube. Sperm were exposed in parallel experiments to (a) an ascending chemical concentration gradient (capillaries containing stimuli, wells filled with HTF), (b) a uniform concentration (stimulus present in both capillaries and wells), or (c) a descending gradient (wells filled with stimuli, capillaries containing HTF). At least four replicate trials were performed for every test, using sperm from different human donors. After incubation (60 min, 37°C), the contents of each capillary were transferred to toxoplasmosis slides, heat-fixed, and stained with 0.1% acridine orange. During trials, O 2 microelectrode measurements showed no change in oxygen levels inside and outside of the capillaries. Cell counts were determined using an inverted phase-contrast microscope (Olympus BH2).

CAVMA
Microcapillary bioassays were combined with CAVMA as described previously (17,49) to determine behavioral/motility effects induced by exposure to odor gradients. For each stimulus concentration (10 Ϫ8 to 10 Ϫ5 M), four replicate trials were performed in a 4 (concentration) ϫ 6 (stimuli) factor design. Sperm from different donors were used in each trial. Swimming speeds and directions of individual cells were videotaped over 15 s, beginning 10 -15 min after the start of each microcapillary bioassay. Fluid dynamic theory predicts that the concentration gradient within 300 m of the capillary tip should reach a steady state within ϳ10 min. In addition, the presence of a steep, ascending gradient was confirmed by quantifying fluorescent marker release (rhodamine) through fluorescence imaging. Rhodamine concentration (a) remained highest at the tip, decreasing as approximately the three-halves power of time and (b) decreased geometrically as a function of 1/distance from the tip (consistent with mass transport via molecular diffusion). Using non-motile sperm as tracers for flow visualization, fluid motion (due to convection, flow from the pipette, and Chemoreceptor-guided Sperm Motility MAY 13, 2011 • VOLUME 286 • NUMBER 19 JOURNAL OF BIOLOGICAL CHEMISTRY 17313 mixing) was found to be insignificant (Ͻ5 m/s) when compared with swimming speeds of live cells (17). Sperm images were recorded on magnetic tape as described (49), using a cooled CCD camera (6415-2000, Cohu (San Diego, CA)) mounted on an inverted light microscope (Olympus IX70, Melville, NY) at ϫ90 magnification. A region within 300 m of the microcapillary tip was monitored at ϳ100 m depth of field. According to fluid dynamic theory, drag forces show effects on flagellar motion within a distance of ϳ10 sperm body lengths from a microscope slide surface. To avoid potential artifacts, we assayed sperm motility in response to chemical stimuli at ϳ2 mm (i.e. ϳ35 sperm body lengths) from the nearest chamber wall. Images were digitized at 30-Hz frame rates and processed using a CAVMA system (VP320; Motion Analysis Corp., Santa Rosa, CA) and custom software interfaced with a Sun SPARC2 computer work station (50). To avoid parallax, we discarded paths in which apparent sperm head size changed between frames by more than 20%. Swimming speed was determined on a frame-by-frame basis and averaged over each path. The angle of sperm orientation () was measured with respect to an origin (0°), defined as the shortest tangent between each cell and the capillary tip. Using circular statistics, a mean vector (r) was calculated to describe the average direction of movement of a cell population (vector length ϭ 1 indicates that all cells swim in a single, common direction; vector length ϭ 0 indicates random motion).

Flagellar Beating Analysis
Beat frequency of an individual sperm flagellum was determined as an adjunct to analysis of swimming behavior (see above). Microcapillary assays were repeated for each test or control treatment (at 37°C and 5% CO 2 atmosphere). Sperm, however, were exposed only to ascending concentration gradients. After a gradient had been established (10 min), motile sperm were photographed at 30 Hz as described. Beat patterns were determined for 10 -24 cells/test or control, respectively. A random sampling of cells was used in each analysis, with the following caveats: (a) cells were motile; (b) they remained within the focal plane of the camera for one continuous second or longer (i.e. 30 images or more); and (c) they were positioned less than 150 m from the capillary tip. National Institutes of Health ImageJ software was used to calculate the average beat frequency of a sperm's flagellum as the reciprocal of total time elapsed over three complete beat cycles, divided by 3. This calculation gave highly reproducible results.

Acrosome Reaction Assay
The ability of sperm to undergo the acrosome reaction was determined from an independent series of microcapillary assays. 4-l samples of cells residing either in wells or in capillaries were collected after a 1-h incubation (37°C) in ascending gradients of bourgeonal, Myrac, or PI-23472 (versus HTF controls; 5 replicates/treatment/control). Each sample was split into equal volumes (2 l) and treated as follows. The first volume was cooled on ice for 10 min, added to 95% methanol (2 l) for 30 s, and then reacted for 30 min with 5 l of the acrosomal marker fluorescein isothiocyanate-conjugated Pisum sativum agglutinin (FITC-PSA; 1 mg/ml). After wash (HTF), the sample was centrifuged (300 ϫ g, 5 min), transferred to toxoplasmosis slides, and air-dried. Using fluorescence microscopy (ϫ1000; Olympus BX40), the percentage of cells that had spontaneously undergone the acrosome reaction was determined for the first 200 sperm encountered. Reacted sperm retained a unique equatorial stain, as opposed to exhibiting uniform fluorescence across the entire head. In the second volume, cells were challenged for 30 min with 5 l of phorbol 12-myristate 13-acetate (5 M) prior to incubating with FITC-PSA and counting. Inducing the acrosome reaction, phorbol 12-myristate 13-acetate affects only sperm in a capacitated state. The combined percentage of sperm exhibiting either spontaneous or induced acrosome reactions indicates the total population of cells capable of binding to and penetrating the zona pellucida (51).

Motility Analysis Using Laser Tweezers
The laser tweezers optical design, hardware, sperm tracking software, experimental procedure, and analysis of sperm have been described in detail elsewhere (52,53). Briefly, a single point gradient trap was generated using an Nd:YVO 4 continuous wave 1064-nm wavelength laser (BL-106C, Spectra Physics, Mountain View, CA), coupled to a Zeiss Axiovert S100 microscope equipped with a phase III, oil immersion objective (ϫ40 magnification; numerical aperture ϭ 1.3). Phase-contrast images of swimming sperm were acquired at a 40-Hz frame rate using a CCD camera (Cohu 7800). Individual sperm were randomly selected and automatically tracked and trapped using the real-time automated tracking and trapping system (RATTS). In an operational mode analogous to conventional computer-assisted sperm analysis, the swimming speed (curvilinear velocity (m/s)) was calculated based on the pixel coordinates (x/y) of the sperm's swimming trajectory. Once a cell was automatically trapped by RATTS, the laser power was continually attenuated by rotating a linear polarizer set in a stepper-motor-controlled rotating mount (PR50PP, Newport Corp., Irvine, CA). The laser power at which the sperm was capable of escaping the trap was recorded (escape power, P esc (milliwatts)) and converted to swimming force in piconewtons by the equation, F ϭ n 1 QP/c (54). The odorant stimulus solution was delivered through 1 ⁄ 32-inch inner diameter Tygon tubing connected to a 1 1 ⁄ 2-inch 21-gauge needle shaft. After a specimen was loaded onto the microscope, the needle was inserted through the silicon gasket and positioned to the center of the chamber. Immediately after insertion, 10 l of the odorant solution were injected, and calculations of the curvilinear velocity and escape laser power began. Only sperm in close vicinity to the needle exit were analyzed. This analysis lasted for 5 min. Odorant and non-odorant chambers were alternated during the span of the experiment.

ELISA-based Cyclic Nucleotide Assays
For quantitative determination of intracellular cAMP/ cGMP, a commercial kit was used according to the manufacturer's instructions (Enzo Life Sciences, Loerrach, Germany). Briefly, pooled sperm samples (4 -5 donors) were prepared as described above, and cell density was adjusted to ϳ4 ϫ 10 8 cells/ml (20 l). Equal volumes of stimulus solution (IBMX (200 M) and either bourgeonal, Myrac, PI-23472 (10 M each), NaHCO 3 (30 mM), or SNP (1 M)) were added. After 30 s, reactions were stopped. Desiccated lysates (5 h; 37°C) were reconstituted in 0.1 M HCl, and 100-l samples were transferred to a microtiter plate. cAMP/cGMP concentrations were measured according to the manufacturer's instructions with the following changes. To exclude differences in antibody binding efficiency, all samples and standards were adjusted to identical ionic conditions. For each tested condition and standard, three replicate experiments were performed, and results were averaged.

Statistical Analysis
All data were obtained from sperm samples of at least three (usually five or more) different donors. If not stated otherwise, results are presented as means Ϯ S.E., and statistical analyses were performed using paired or unpaired Student's t tests (as dictated by data distribution and experimental design). To determine if cell movement within a chemical gradient was non-random during CAVMA, each mean direction was compared with a uniform circular distribution using Rayleigh's test. Data from microcapillary bioassays were analyzed statistically by a two-way ANOVA using stimulus concentration and gradient type as factors, followed by post hoc Scheffé tests (17,49). The curvilinear velocity and escape laser power distributions resulting from the experiments testing the effects of the three odorants were statistically compared using the non-parametric Wilcoxon rank sum test (based on 5% significance). The distributions are displayed graphically in box plot form (Fig. 6C). Each box plot displays the following parameters for a given distribution:

Chemicals and Reagents
Unless stated otherwise, all chemicals and reagents were obtained from Sigma-Aldrich (Schnelldorf, Germany).

Distinct Agonist Profiles of Three Human Testicular
Chemoreceptors-To understand the functional link between chemoreceptor activation, Ca 2ϩ signaling, and swimming behavior in human sperm, we aimed to identify stimuli that specifically activate orphan testicular ORs. We concentrated on two distinct receptors, OR4D1 and OR7A5, which had previously been found expressed at high levels in human testicular tissue and/or fertile human sperm by various groups (24,25,32,35). First, we confirmed testicular expression of OR4D1 and OR7A5 by RT-PCR (Fig. 1A, inset) using OR1D2 as a positive control (17). We next examined whether this expression pattern is tissue-specific or a result of "leaky" and therefore tissue-independent low level OR transcription. In a series of quantitative real-time PCR experiments, we comparatively surveyed relative mRNA expression levels of OR4D1, OR7A5, and OR1D2 in nine different human tissues (Fig. 1A). Our results indicate no (or negligi-ble) OR gene transcription in heart, lung, liver, stomach, aorta, brain, skeletal muscle, and prostate. By contrast, all three receptors were expressed at significant levels in testicular tissue samples.
Chemoreceptor Agonists Mediate Ca 2ϩ Signals in Human Sperm-Next, we asked whether the same agonists that activate recombinant testicular chemoreceptors also stimulate individual sperm. Brief (10-s) exposure to bourgeonal, Myrac, or PI-23472 triggered Ca 2ϩ transients in single cell fluorescence imaging recordings of fura-2-loaded sperm ( Fig. 2A). As observed for recombinantly expressed receptors (supplemental Fig. S1), undecanal incubation abolished Ca 2ϩ responses to bourgeonal but lacked an inhibitory effect on both Myrac-and PI-23472-induced signals. These data suggest that, analogous to OR1D2 activation in sperm (1,16,17,41), Myrac as well as PI-23472 selectively activate sperm chemoreceptors, probably OR7A5 and OR4D1. Because a hallmark of OR expression in olfactory sensory neurons is their monogenic expression according to a one-cell/one-receptor model (22), we investigated whether bourgeonal, Myrac, and PI-23472 activate receptor-specific sperm subpopulations or if an individual cell can be stimulated by all three chemoreceptor agonists. When sequentially applied in random order at interstimulus intervals of Ն60 s, all three stimuli induce repetitive Ca 2ϩ responses in the great majority of sperm (Fig. 2, B and D, and supplemental Fig. S2A). We could, however, exclude a nonspecific activation of sperm by chemical cues per se. When sperm populations were challenged in plate reader-based Ca 2ϩ assays (17) (17,57,59,60)), vertoprenal (a structural analog of Myrac that we found inactive at OR7A5 (Fig. 1B)), or mustard oil (a potent TRPA1 channel activator in sensory neurons (61)), Ca 2ϩ signals were essentially negligible (Fig. 2E). Concentration dependence of bourgeonal, Myrac, and PI-23472 response probability/strength was assessed by calculating normalized quasi-dose-response curves from single cell Ca 2ϩ imaging experiments (Fig. 2F). As previously reported (17), initial Ca 2ϩ responses to bourgeonal were recorded in the low nanomolar range (Fig. 2F), whereas both Myrac and PI-23472 response thresholds were right-shifted by approximately 2 orders of magnitude. At stimulus concentra-tions of Ն30 M, the maximum percentage of cells was activated by all three stimuli. Independent of the stimulus, responses were critically dependent on the extracellular Ca 2ϩ concentration (supplemental Fig. S2, B and C). Together, human sperm respond to defined ligands of different chemoreceptors with some differences in ligand sensitivity but without any obvious preference for a specific stimulus at saturating concentrations.
What is the nature of the signaling cascade(s) activated downstream receptor-ligand interaction? Our previous results suggested a role of particulate adenylate cyclase and cAMP in OR1D2 signaling (16). In pharmacological single cell assays, we   MAY Fig. S3, A-G). Ca 2ϩ responses to bourgeonal, Myrac, and PI-23472 were strongly diminished by two adenylate cyclase inhibitors (MDL-12,330A and SQ-22536), whereas KH7, a described blocker of the soluble adenylate cyclase (62), had no effect. A nonspecific effect of SQ-22536 on general Ca 2ϩ signal capacity was excluded by stimulation with bicarbonate or elevated extracellular K ϩ (supplemental Fig. S3D). A battery of agents that target a variety of different signaling molecules showed no significant response inhibition (supplemental Fig. S3G). Surprisingly, we did not measure an increase in cAMP upon sperm exposure to bourgeonal, Myrac, or PI-23472 (5 M) (supplemental Fig. S3H); nor did we detect a rise in cGMP levels (supplemental Fig. S3I). As positive controls, the soluble adenylate cyclase activator bicarbonate and the NO donor sodium nitroprusside (SNP) triggered a significant increase in cAMP and cGMP, respectively.

Chemoreceptor-guided Sperm Motility
Agonist-dependent Ca 2ϩ Response Patterns-Ca 2ϩ signaling is the common denominator of most physiological processes that occur in sperm, and complex Ca 2ϩ dynamics selectively control individual sperm motility patterns (3,44). Thus, we set out to describe the Ca 2ϩ mobilization characteristics in response to bourgeonal, Myrac, and PI-23472, respectively. We used saturating stimulus concentrations (50 M each) (Fig. 2F) to avoid data interpretation errors that could occur when lower agonist concentrations match different stages within the dynamic range of a given receptor/ligand dose-response curve (Fig. 2F). For comparative semiquantitative Ca 2ϩ measurements both within different subcompartments of the same cell and between different sperm cells, we opted for a fura-2-based ratiometric imaging approach. Using F 340 /F 380 measurements corrects for unequal dye loading, bleaching, focal plane shift, dye leakage, and (most important) differences of both optical path length and accessible cytosolic volume (63,64). We distinguished Ca 2ϩ signals in six different sperm compartments (i.e. the most anterior area marked by the acrosome cap, the posterior head region, the midpiece, and three different regions of the principal piece (proximal, intermediate, and distal flagellum with respect to the sperm head)) ( Fig. 3). For each compartment and stimulus, we analyzed a combination of Ca 2ϩ response parameters, including the signal onset, the response amplitude, the rise time (20 -80% of amplitude), the rising speed (rise time (s)/amplitude 20 -80% (ratio units)), and the Ca 2ϩ extrusion kinetics (time constant , calculated from monoexponential fits of the Ca 2ϩ decay phase).
First, we investigated the origin of the Ca 2ϩ signal and its potential propagation through different sperm compartments (Fig. 3, A and B). For both bourgeonal and PI-23472, significantly increased Ca 2ϩ concentrations (⌬F 340 /F 380 Ͼ 2 ϫ S.D. base line ) were first observed in the proximal part of the flagellum, whereas Myrac responses originated in the midpiece. For all stimuli, we recorded bilateral Ca 2ϩ propagation patterns from the signal origin along a cell's longitudinal axis toward its anterior and posterior end, respectively. With few exceptions, Ca 2ϩ signal onsets differed significantly between adjacent compartments (Fig. 3A). Maximum delays of 3.16 Ϯ 0.31 s (bourgeonal), 2.48 Ϯ 0.27 s (Myrac), and 2.5 Ϯ 0.39 s (PI-23472) were observed between the first and last detected Ca 2ϩ response onset (corresponding to an average longitudinal distance of ϳ10 -30 m). Moreover, Ca 2ϩ signals spread significantly faster toward the sperm head when cells were stimulated with Myrac (Fig. 3, A and D). These data show that different chemoreceptor agonists elicit Ca 2ϩ signals in human sperm that all originate in the proximal area of the principal piece but spread at different speed through the cell, indicating that receptor/ ligand-specific mechanisms govern signal propagation.
When Ca 2ϩ response amplitudes were analyzed, PI-23472 induced significantly larger signals than bourgeonal (flagellar regions) or Myrac (all compartments) (Fig. 3, D and E). Moreover, bourgeonal triggered larger Ca 2ϩ signals in the head than observed in Myrac-exposed sperm. Another distinctive parameter of stimulus-dependent Ca 2ϩ signals is their individual rising speed. Myrac-mediated Ca 2ϩ transients rise significantly slower in all compartments. Differences between bourgeonal and PI-23472 responses are also significant although less pronounced. Plotting both stimulus-and compartment-dependent Ca 2ϩ response amplitudes as a function of rising speed, data points group in separated clusters, and stimulus-specific response profiles become apparent (Fig. 3E).

Chemoreceptor-guided Sperm Motility
Together, our data indicate that all three chemosignals trigger characteristic Ca 2ϩ mobilization patterns in human sperm. These are probably established by receptor/ligand-specific signaling mechanisms and might control distinct motility phenotypes.
Agonist-dependent Motility Profiles-To investigate whether stimulus-dependent sperm Ca 2ϩ dynamics determine unique behaviors, we performed microcapillary accumulation assays that were complemented by CAVMA studies (16,17) that measured single cell swimming speed and trajectories as well as flagellar beating frequency as a function of sperm exposure to different chemical concentration gradients.
As described previously (16,17), sperm exposed to an ascending gradient of Ն10 Ϫ7 M bourgeonal swam faster (Fig. 5A), strongly oriented with respect to the stimulus source (capillary tip) (Fig. 5D and supplemental Fig. S4), and accumulated inside the capillary at higher cell densities than in control experiments (Fig. 5B). These effects on sperm behavior were dose-dependent (supplemental Fig. S4A), peaked at a stimulus concentration of 10 Ϫ6 M, and declined at concentrations of Ն10 Ϫ5 M (data not shown), exhibiting a bell-shaped dose-response curve. In contrast, sperm exposed to an ascending gradient of Myrac or PI-23472 (Ն10 Ϫ7 M), displayed a random orientation pattern (Fig. 5D). Both chemostimuli, however, significantly elevated sperm swimming speed (Fig. 5A). Cell densities within the capillaries were only significantly elevated in the presence of Myrac, with this increase being minor (62%) relative to bourgeonal ( Fig. 5B; p Ͻ 0.05). To examine whether this Myrac-dependent cell accumulation indicates a chemotactic response (which appeared unlikely given the lack of orientation in single cell motion analysis (Fig. 5D and supplemental Fig. S4)), we assayed the cell density in microcapillaries when stimulus concentrations were uniform (Fig. 5C). In sharp contrast to bourgeonal-mediated effects, sperm accumulation in presence of uniform Myrac or PI-23472 concentrations did not significantly differ from densities observed in an ascending gradient, suggesting that the Myrac-induced increase in density is not a chemotactic but rather a chemokinetic effect (1,7). Neither Myrac-nor PI-23472-mediated responses were affected in the presence of undecanal (Fig. 5,  A and B, and supplemental Fig. S4B), whereas this described OR1D2 antagonist (17,40) abolished bourgeonal-dependent behavior. Ascending gradients of undecanal alone (up to 10 Ϫ5 M) did not induce any behavioral response (Fig. 5, A and B, and supplemental Fig. S4B). These data indicate that all three identified chemoreceptor agonists trigger distinct and stimulus-specific motility patterns in human sperm.

Chemoreceptor-guided Sperm Motility
Previous work has shown that the population of chemotactic sperm is correlated with the percentage of capacitated cells (1). We therefore determined the capacitative state among sperm populations showing either chemosensitive (capillary recruitment) or -insensitive (residing in wells) behavior in an independent series of accumulation assays (supplemental Fig. S5A). Using an FITC-Pisum sativum agglutinin staining assay, the fraction of spontaneously acrosome-reacted sperm was similar in wells and capillaries, independent of the experimental conditions. By contrast, the proportion of sperm in which the acrosome reaction can be induced (phorbol 12-myristate 13-acetate; 5 M; 30 min), representing sperm in a capacitated state (51), is significantly increased in cells found in capillaries after stimulus exposure. These findings suggest that the ability of a sperm not only to detect a chemotactically active stimulus but to translate this chemosignal into a meaningful behavior is somewhat dependent on attainment of the capacitative state.
Given that sperm alter flagellar motion as dictated by their environment (3), we next asked whether flagellar beating is changed when cells are exposed to ascending agonist gradients (supplemental Fig. S5, B-D). Frame-by-frame video motion analysis (30 Hz) revealed that both bourgeonal and Myrac induced a significant increase in the average flagellar beat frequency, whereas beating remained essentially unchanged in the presence of PI-23472 as well as undecanal (supplemental Fig.  S5B). As observed for swimming directionality and speed (supplemental Fig. S4A), changes in beat frequency were dose-dependent (supplemental Fig. S5C). In general, both bourgeonal and Myrac elevated flagellar beat frequencies without changing the overall pattern of relatively low amplitude sinusoidal flagellar motion (supplemental Fig. S5D). Together, our results indicate that the chemokinetic effects of both bourgeonal and Myrac result, at least in part, from an increased flagellar beat frequency. Beating symmetry, however, remains unaltered, thus showing that neither stimulus induces hyperactivated motility.
Next, we employed laser tweezers as a non-invasive biophysical tool to examine the power at which swimming sperm escape from an optical trap (52,54). This value relates directly to the swimming force (in piconewtons) of individual sperm (54). RATTS allows the escape laser power for an individual cell to be measured simultaneously with its swimming speed. As predicted from CAVMA results ( Fig. 5 and supplemental Fig.  S4), bourgeonal, PI-23472, and Myrac significantly elevated sperm curvilinear velocity (Fig. 6B). This effect on swimming speed, however, was only correlated with a significant increase in escape laser power when sperm were exposed to bourgeonal (Fig. 6, A and B). Neither PI-23472 nor Myrac induced a significant increase in escape laser power. In summary, we identified three sperm chemoreceptor agonists that induce both spatially and temporally distinct and stimulus-specific Ca 2ϩ mobilization patterns that correlate with individually unique behaviors.

DISCUSSION
Cells invest substantial energy to control cytosolic Ca 2ϩ levels (43,65,66). In sperm, localized Ca 2ϩ signals control flagellar beating via Ca 2ϩ -sensitive axonemal motor proteins, affect sperm capacitation, and mediate the acrosome reaction, yet our conceptual understanding of how external chemosignal detection, spatiotemporally discrete Ca 2ϩ signaling, and distinct behavioral responses are linked in mammalian sperm is still rudimentary. Here, we identify agonists for two orphan human ORs (OR4D1 and OR7A5) that were previously described in testis and/or mature sperm (24,25,32,35). In concert with the recently identified human sperm chemoreceptor OR1D2 (17), all three recombinant receptors display broadly tuned but nonoverlapping receptive fields. Analogous to OR1D2, agonists of OR4D1 and OR7A5 induce dose-dependent low threshold Ca 2ϩ responses in mature human sperm. In contrast to ORspecific tuning profiles of olfactory sensory neurons, a given sperm cell is activated independently by all three agonists. We show, however, that each chemostimulus triggers a characteristic Ca 2ϩ response that correlates with a stimulus-specific motility pattern. Our results thus indicate multiple stages of chemical communication between human sperm and its environment, each of which might utilize a separate receptor-ligand combination to activate an adequate context-specific sperm behavior.
Human Sperm Chemoreceptors-In most mammals, including humans, ORs constitute the largest gene superfamily, with ϳ400 apparently functional members in humans and ϳ1200 in mice (22). Despite intensive efforts, the vast majority of mammalian ORs are yet to be deorphanized (67)(68)(69). When Parmentier et al. (23) first proposed human OR expression in male germ cells, their finding sparked a controversial discussion about the potential function of such "ectopic" ORs. Today, numerous groups have reported mammalian OR expression on the transcript and/or protein level in both spermatogenic cells and mature sperm (16 -19, 23-36). In addition, functional evidence for a physiological role of "ectopic" chemoreceptors is accumulating (17-20, 55, 71-73). In the present study, we focused on two receptors, OR4D1 and OR7A5, for which different groups had independently reported medium to high level expression in human testicular tissue and/or fertile human sperm (24,25,32,35). Our quantitative PCR data suggest a high level of tissue specificity because relative expression levels in eight other human tissues (pooled high purity samples from up to 32 donors) proved negligible.
To identify specific stimuli for OR4D1 and OR7A5, we used the same experimental strategy that we had successfully employed in the past to match human OR-ligand pairs (17,38,(55)(56)(57)74). Given the poor OR surface expression rate typically observed in heterologous cells (47), careful evaluation of single cell Ca 2ϩ responses is imperative, making this an "ultra-low throughput" assay that is neither applicable to the full repertoire of human ORs 5 nor suitable for large scale ligand screening. When successful, however, this approach has provided robust results that were later confirmed by different laboratories (39,40,59,60). Recombinant OR4D1, OR7A5, and OR1D2 each respond to an apparently non-overlapping set of structurally similar low molecular weight compounds (Fig. 1, B and C). Such relatively broad yet odor subset-selective tuning profiles have also been reported for other ORs (69,75). Thus, testicular ORs appear to follow mechanistic principles similar to those of "conventional" nasal ORs. In a recent high throughput screening of 245 human ORs, including OR4D1, OR7A5, and OR1D2, against a panel of 93 structurally different odor molecules (69), a total of 10 ORs were matched to specific ligands. The three testicular ORs identified here, however, did not respond to any of the 93 test molecules, none of which bears a clear structural resemblance to bourgeonal, Myrac, or PI-23472.
For OR1D2, we could previously confirm its expression in sperm by immunocytochemistry (20). Despite several attempts, however, we failed to generate specific antibodies against both OR4D1 and OR7A5, probably because of the large size of the gene family and substantial sequence homology among its members. The testis-specific OR expression pattern we found in quantitative PCR experiments as well as both the effective Ca 2ϩ mobilization and the distinct motility patterns induced in sperm by the identified OR agonists, however, suggest functional expression of OR4D1 and OR7A5 (in addition to OR1D2) in male human gametes. Activation of an individual cell by all three receptor-specific stimuli, however, suggests different transcription control mechanisms in spermatogenic cells as compared with olfactory sensory neurons in which monogenic expression of only a single type of OR is tightly controlled (22).
Chemosensory Ca 2ϩ Signaling-Complex spatiotemporal Ca 2ϩ dynamics orchestrate sperm motility (3). The flagellar motor, however, is not the only Ca 2ϩ -dependent molecular machinery in sperm. Most vital physiological processes, such as capacitation or the acrosome reaction, are controlled by cytosolic Ca 2ϩ (44,76,77). Therefore, functional separation of individual signaling pathways with little, if any, input integration is critical.
Using semiquantitative ratiometric Ca 2ϩ imaging, we determined compartment-specific response characteristics when sperm were exposed to saturating concentrations of bourgeonal, Myrac, or PI-23472 (Fig. 3). In a given flagellar compartment, Ca 2ϩ rise times were significantly different for all stimuli. Vice versa, considerably faster rising speeds were observed in flagellar versus head regions for each stimulus. In addition, PI-23472 response amplitudes were substantially increased, and the Ca 2ϩ "wave" triggered by Myrac propagated significantly faster than bourgeonal-or PI-23472-mediated signals.
Despite their individually unique spatiotemporal features, Ca 2ϩ responses to bourgeonal, Myrac, and PI-23472 share some commonalities. First, each characteristic response originated in proximal areas of the principal piece and showed a wave-like propagation along the cell's longitudinal axis (Fig.  3A). Second, for all three chemoreceptor agonists, sperm responses depend on extracellular Ca 2ϩ (supplemental Fig. S2, B and C). This finding does not exclude a secondary function of Ca 2ϩ storage organelles, such as the acrosome vesicle (78) or the redundant nuclear envelope (45) in response amplification/ propagation, but it clearly implicates plasma membrane Ca 2ϩ channels as the primary targets of all three transduction pathways. Third, Ca 2ϩ clearance kinetics displayed essentially no stimulus-specific differences in a given region (Fig. 3C). We conclude that sperm Ca 2ϩ extrusion mechanisms are functionally uncoupled from the Ca 2ϩ influx machinery and, thus, stimulus-independent. Using synthetic Ca 2ϩ dyes, such as fura-2, it is impossible to distinguish cytosolic from organelle-specific Ca 2ϩ signals. By loading cells with membrane-permeable AM esters, some dye molecules will inevitably be hydrolyzed within, for example, the mitochondrial matrix or the acrosome. Thus, we cannot rule out a role of organelle-specific signaling in chemoreceptor Ca 2ϩ cascades.
The apparent tail-to-head Ca 2ϩ propagation in response to Myrac is significantly faster (ϳ0.9 s) than recorded for both bourgeonal and PI-23472 (ϳ3.2 s (bourgeonal) and ϳ2.5 s (PI-23472)) (Fig. 3A). The simplest explanation for this difference is that Myrac-sensitive chemoreceptors are more broadly distributed along the tail-to-head axis, whereas receptors for bourgeonal and PI-23472, respectively, are more strictly confined to the midpiece and proximal flagellum. An analog distribution pattern was recently shown for OR1D2 by immunocytochemistry (20). Pronounced antibody staining is found in the midpiece and, in a more patchy pattern, in proximal flagellar areas. Upon bourgeonal activation, ␤-arrestin2 mediates OR1D2 translocation from the plasma membrane. Immunofluorescent signals in the head equatorial region, however, proved essentially unaffected (20). To segregate and/or modulate head and tail Ca 2ϩ signals, the sperm annulus, a membrane diffusion barrier between the principal piece and midpiece (79), and the densely packed helical array of mitochondria in the midpiece are ideally positioned to restrict lateral chemoreceptor diffusion and function as a Ca 2ϩ buffer (3), respectively. This way, chemoreceptor signaling might also stimulate oxidative phosphorylation and, thus, regulate ATP generation and homeostasis.
In sperm, the identities of the signaling cascades activated downstream individual chemoreceptors remain a matter of debate. Consistent with data reported previously (16), our pharmacological experiments suggest a common pathway that involves adenylate cyclase activation. To serve distinct physiological functions, this common cascade has to split and trigger receptor-specific pathways. Such specificity could, for example, be established by expression of individual chemoreceptor types within distinct transducisomes (70). The lack of a detectable cAMP increase, however, is puzzling. Future experiments will have to address whether potential changes in cAMP occur below the detection threshold of our assay or if the drugs we used, in addition to their described function, target a yet unknown signal transduction component.
Chemosignals, Ca 2ϩ , and Sperm Behavior-The control of flagellar shape and movement by Ca 2ϩ is an evolutionarily conserved concept established from marine invertebrate to mammalian sperm (4,15,43). Flagellar (a)symmetry, amplitude, and beat frequency change as a function of Ca 2ϩ mobilization, thus discretely generating straight or circular swimming paths, directed (chemotactic) turns, random (exploratory) turning, hyperactivation, or reversible motility arrest (3) according to the prevailing physicochemical conditions. Our observation that activation of different sperm chemoreceptors resulted in discrete Ca 2ϩ dynamics led us to investigate whether an individual Ca 2ϩ response could be linked to a distinct behavior. Microcapillary assays (comparing cell responses to an ascending chemical concentration gradient, a descending gradient, and a uniform chemical concentration field; Fig. 5) faithfully measure sperm accumulation and allow the distinction between positive or negative chemotaxis, chemokinesis, and trapping (1). Single cell motion analysis and path tracking ( Fig.  5D and supplemental Figs. S4 and S5) in an ascending threedimensional chemical gradient capture a complementary "snapshot" of the mechanisms underlying cell accumulation patterns and, thus, provide information on directionality, swimming speed, and flagellar waveform/beat frequency. In addition, laser tweezers provide another non-invasive tool to examine swimming force by determining the laser power required for individual sperm cells to escape from an optical trap. When used with RATTS, sperm curvilinear velocity can be correlated with the escape laser power (52). Together, each tool in this analytical "package" adds valuable information on chemoreceptor-mediated sperm behavior. None of the assays, however, measures percentages of chemosensitive cells because each bioassay is time-critical (e.g. prior recruitment of chemotactic sperm during gradient establishment). In fact, we expect the actual population of chemotactic cells to correlate closely with the fraction of sperm that has adopted the capacitated state at any given point in time (i.e. ϳ10 -15% (supplemental Fig. S5A)). As might be expected from the observed stimulus-specific Ca 2ϩ dynamics, distinct behavioral "output" patterns were generated by bourgeonal, Myrac, and PI-23472, respectively. Among the three chemoreceptor agonists, only bourgeonal functions as a chemoattractant. All three stimuli increase flagellar beating frequency and thus velocity, but a significant increase in propulsion force only occurs in response to bourgeonal. None of the stimuli caused asymmetric "whiplike" beating, indicating that neither agonist caused hyperactivation.
How do these findings complement our current concepts of sperm navigation in the female reproductive tract? First, our data strengthen the notion that sperm guidance might be much more elaborate than originally thought (1). Second, we propose that individual chemosensory receptors and signaling pathways are utilized at different prefusion stages during a sperm's journey.
The chemoreceptor agonists identified in this study all represent synthetic compounds. Thus, a major effort in future research will have to be directed toward identifying both the chemical nature of the endogenous "genuine" ligands of sperm chemoreceptors and their individual secretion sources.