Patch clamp studies of human sperm under physiological ionic conditions reveal three functionally and pharmacologically distinct cation channels

Whilst fertilizing capacity depends upon a K+ conductance (GK) that allows the spermatozoon membrane potential (Vm) to be held at a negative value, the characteristics of this conductance in human sperm are virtually unknown. We therefore studied the biophysical/pharmacological properties of the K+ conductance in spermatozoa from normal donors held under voltage/current clamp in the whole cell recording configuration. Our standard recording conditions were designed to maintain quasi-physiological, Na+, K+ and Cl− gradients. Experiments that explored the effects of ionic substitution/ion channel blockers upon membrane current/potential showed that resting Vm was dependent upon a hyperpolarizing K+ current that flowed via channels that displayed only weak voltage dependence and limited (∼7-fold) K+ versus Na+ selectivity. This conductance was blocked by quinidine (0.3 mM), bupivacaine (3 mM) and clofilium (50 µM), NNC55-0396 (2 µM) and mibefradil (30 µM), but not by 4-aminopyridine (2 mM, 4-AP). Progesterone had no effect upon the hyperpolarizing K+ current. Repolarization after a test depolarization consistently evoked a transient inward ‘tail current’ (ITail) that flowed via a second population of ion channels with poor (∼3-fold) K+ versus Na+ selectivity. The activity of these channels was increased by quinidine, 4-AP and progesterone. Vm in human sperm is therefore dependent upon a hyperpolarizing K+ current that flows via channels that most closely resemble those encoded by Slo3. Although 0.5 µM progesterone had no effect upon these channels, this hormone did activate the pharmacologically distinct channels that mediate ITail. In conclusion, this study reveals three functionally and pharmacologically distinct cation channels: Ik, ITail, ICatSper.


Introduction
Plasma membrane ion channels are central to the control of sperm function (Darszon et al., 1999;Lishko et al., 2011b) and, in particular, Ca 2+ entry via sperm cation channels (CatSper) is critical for several physiologically important processes, including hyperactivation, chemotaxis and the acrosome reaction (Lishko et al., 2011a, b;Strünker et al., 2011;Brenker et al., 2012). Like somatic cells, mouse and human spermatozoa normally display negative resting membrane potentials (V m ) that are dependent upon the activity of K + channels, and the magnitude of this potential exerts a strong influence over Ca 2+ influx since it determines the gating of CatSper and also sets the driving force for Ca 2+ entry through these channels. At least in mouse sperm, a negative shift in V m (hyperpolarization) is essential to capacitation, the acquisition of fertilizing ability that occurs within the female reproductive tract (Zeng et al., 1995;De la Vega-Beltran et al., 2012). Understanding the mechanisms that allow V m to be maintained is therefore central to our understanding of spermatozoon physiology.
Whilst protein and mRNA encoding several K + channel subtypes, including voltage-gated K + channels (KCNA5) (Felix et al., 2002), tandem pore domain K + channels (KCNK5) (Barfield et al., 2005a, b) and ATPgated K + channels (Acevedo et al., 2006;Martínez-Ló pez et al., 2009), is present in mouse and human sperm, the biophysical properties of K + channels in these cells are only just becoming clear. Electrophysiological studies of mouse sperm thus led to the identification of the sperm K + channel (KSper), a K + -permeable conductance whose activity was strongly enhanced by intracellular alkalinization (Navarro et al., 2007). KSper-dependent K + currents apparently flow via channels encoded by Slo3 (KCNMA3) (Navarro et al., 2007;Santi et al., 2010;Zeng et al., 2011), a gene expressed only in male germ cells (Schreiber et al., 1998;Santi et al., 2010;Yang et al., 2011;Zeng et al., 2011). Slo3-encoded channels resemble the endogenous mouse K + channels in their pharmacology, weak voltage sensitivity, low K + versus Na + selectivity and sensitivity to changes in intracellular pH (pH i ) (Schreiber et al., 1998;Zhang et al., 2006a, b;Martínez-Ló pez et al., 2009). Moreover, V m in mouse sperm is clearly dependent upon pH i , an observation consistent with a principal role for Slo3 in the mature spermatozoon (Navarro et al., 2007;Martínez-Ló pez et al., 2009). Slo3 gene deletion thus abolishes the hyperpolarization seen during capacitation and mimics the effects of K + channel blockade on sperm function (Santi et al., 2010;Zeng et al., 2011). Very recent studies of human sperm, on the other hand, suggest that the K + conductance of these cells is insensitive to changes in pH i but enhanced by high intracellular Ca 2+ (50 mM). These authors therefore suggested that the principal K + channel in human sperm is the large conductance, Ca 2+ -sensitive (BK) K + channel encoded by the Slo1 gene (Mannowetz et al., 2013). In neurons, these channels regulate excitability and control [Ca 2+ ] i by opening in response to increased [Ca 2+ ] i , causing a negative shift in membrane potential which 'switches off' voltagesensitive Ca 2+ channels (Hoshi, 2012).
Whilst the present study also uses the whole cell recording technique to characterize the K + channels in human sperm, our data suggest that K + currents flow via a population of channels that displays relatively poor ionic selectivity, a feature that is not consistent with a central role for Slo1 encoded channels. In addition, we identify a second, poorly selective, voltage-sensitive cation conductance whose activity is potentiated by progesterone but shows clear pharmacological difference to CatSper.

Preparation of spermatozoa
Semen samples were provided by volunteer donors with no known fertility problems after 48 -72 h of sexual abstinence. All donors were shown to produce normal semen (i.e. ≥ 32% progressive motility; ≥ 40% total motility; ≥15 × 10 6 cells ml 21 ) as defined by established criteria (see WHO, 2010). This procedure had the approval of the Tayside Committee of Medical Research Ethics (08/S1402/6) and written consent was obtained from each donor in accordance with the Human Fertilisation and Embryology Authority (HFEA) 8th Code of Practice. Each sample was allowed to liquefy at 378C for 30 min and the semen then added to a 50 ml Falcon tube containing 5 ml of HTF (see above). Since the aim was to separate motile spermatozoa from other components of the semen, this addition was undertaken gently to ensure that mixing was minimized and that the liquefied semen sample formed a distinct layer at the bottom of the tube. The tube was then inclined at 458 and incubated for 1 h at 378C. The overlying HTF was then aspirated carefully and the motile spermatozoa that had swam into the HTF then allowed to settle into a loose pellet (1 h at room temperature). The cells were re-suspended in capacitating media and maintained at 378C for 4 h (5% CO 2 ). Capacitated cells were then re-suspended in standard bath solution and allowed to adhere to glass coverslips that were transferred to an inverted microscope where they were superfused with standard bath solution.

Electrophysiology
The electrophysiological properties of individual spermatozoa were investigated using the whole cell recording technique (Hamill et al., 1981;Kirichok et al., 2006;Lishko et al., 2011a). The recording pipettes (10-18 MV) were fabricated from borosilicate glass and normally filled with standard pipette solution. Gigaohm seals were obtained by bringing the pipette tip into gentle contact with the cytoplasmic droplet, which lies just behind the sperm head, and the patch of membrane spanning the pipette tip then ruptured by applying suction in conjunction with 1 ms voltage pulses (see Lishko et al., 2010). Our standard recording conditions were designed to preserve physiologically relevant Na + , K + and Cl 2 gradients and V m was held (pClamp 10 Software, Axon Instruments) at a hyperpolarized value (292 mV) between test pulses. Initial experiments were undertaken by recording the membrane currents (I m ) evoked by ramping (250 ms) V m from 292 mV to 68 mV at 1 Hz. To analyse the results of such experiments, I m was first normalized to input capacitance (i.e. expressed as pA pF 21 ) to ensure that variations between the sizes of different spermatozoa did not contribute to the variability in the presented data. All cited values of V m were corrected for the liquid junction potential between the pipette/bath solutions (E L ), and for the voltage drop across the access resistance (R a , 62.8 + 0.8 MV, n ¼ 476 cells from 29 donors). The latter correction was applied retrospectively using the expression V m ¼ V Pip 2 R a . I m , where V Pip is the pipette potential. Since the bath was grounded via a 4% agar/3 M KCl, bridge, the bath solution changes imposed during the present study had negligible effects upon E L . Plots showing the relationship between I m and V m were constructed and, unless otherwise stated, cited values of membrane conductance (G m , pS pF 21 ) are derived by regression analysis (i.e. DI m /DV m ) of data recorded at positive potentials. Resting V m was either inferred from the reversal potential (V Rev , i.e. the value of V m at which I m is zero, voltage clamp experiments) or measured directly by monitoring (5 KHz, data low pass filtered at 3 KHz) the zero current potential (see Hamill et al., 1981). Experiments that quantified the responses to step changes in V m were undertaken using an experimental Sperm K + currents design that employed the standard features of pClamp software (V/4 protocol) to subtract leak/capacitive currents from all recorded data. The statistical significance of differences between control/experimental values were determined tested using Student's paired (repeated measurements on the same cells) or unpaired (comparison between different groups of cells) t-test. The results of experiments that followed more complex protocols were analysed by one way analysis of variance (ANOVA)/Dunnet's post hoc test. Data are cited as mean + s.e.m. and values of n refer to the number of spermatozoa in each group. All observations were confirmed using spermatozoa from at least three different donors.

Currents evoked by voltage ramps
Imposing depolarizing voltage ramps upon spermatozoa exposed to physiologically relevant Na + , K + and Cl 2 gradients (i.e. using standard pipette/baths solutions) consistently evoked noisy outward current. To characterize the conductance underlying this response, currents evoked by 10 successive voltage ramps were averaged ( Fig. 1A) and data derived from different cells pooled and plots showing the I m -V m relationship constructed. This analysis revealed small (1 -2 pA pF 21 ) inward currents at hyperpolarized potentials whilst 25-45 pA pF 21 of outward current became apparent once V m was depolarized past approximately 230 mV (Fig. 1B). Membrane conductance quantified at depolarized potentials (634 + 85 pS pF 21 ) was 15.7 + 2.0-fold greater than at hyperpolarized potentials ( Fig. 1B; P , 0.001). Since seal resistance was .20 GV, Ohm's Law predicts that ,5 pA of inward current will flow via this resistance at 2100 mV, and the magnitude of the current recorded at potentials below approximately 230 mV (Fig. 1) is therefore similar to the predicted magnitude of this 'leak current'. We therefore conclude that I m is too small to be measured when Vm is less than 230 mV. Switching to K + -rich bath solution (20 -30 s) depolarized resting V m by shifting the I m -V m relationship to the right ( Fig. 1C and D) whilst replacing pipette K + with Cs + virtually abolished the voltage-induced outward current and depolarized resting V m to 21.2 + 5.8 mV (P , 0.002; Fig. 1B). The K + -rich bath solution had no effect upon the currents recorded using Cs + -based pipette solution ( Fig. 1B and C) and this outward current must therefore be carried by K + . Figure 1A also shows that the recorded current consistently undershoots its basal value when V m is repolarized after each voltage ramp. Such 'tail currents' (I Tail ) imply the presence of voltagegated channels that become active during the depolarization but take a finite time to close when V m is repolarized.
Effects of altering internal pH (pH i ) Figure 2A shows I m -V m relationships quantified using internal (i.e. pipette) solutions adjusted to pH values between 6.2 and 8.0. These pipette solutions were buffered with 5 mM MES/5 mM HEPES rather than 10 mM HEPES (see Materials and Methods) but the data recorded at pH i 7.4 were virtually identical to the control data shown above and this modification thus has no effect upon the recorded current. These data therefore confirm that depolarization normally evokes outward current. Increasing pH i to 8.0 had no effect upon the I m -V m relationship ( Fig. 2A) and thus had no effect upon G m (Fig. 2B) or V m (Fig. 2C). Lowering pH i , below 6.8 reduced G m by 35% (Fig. 2B) but the residual conductance recorded under these conditions was still 10-fold greater than that quantified using Cs + -based pipette solutions (Fig. 1B). Moreover, lowering pH i had no statistically significant effect upon the currents (Subsequent experiments showed that 0.3 mM quinidine acted as effectively as 3 mM and so this drug was used at this lower concentration in all subsequent studies.) Quinidine and bupivacaine also depolarized resting V m (i.e. caused a rightward shift in reversal potential) and, whilst clofilium seemed to mimic this action, this effect was not statistically significant (Fig. 3C). Lidocaine and 4-AP, on the other hand, had no effect upon V m (Fig. 3C). Examination of the control data derived from this series of experiments showed that resting V m was normally 236.5 + 3.3 mV and regression analysis revealed a correlation between the magnitude of the experimentally induced fall in G m and the shift in V m (correlation coefficient ¼ 0.544, n ¼ 33 spermatozoa, P , 0.001). Since these data suggest that block of the hyperpolarizing K + current causes depolarization, we undertook further experiments in which resting V m was directly monitored under zero current clamp (see Materials and Methods). These studies (i) confirmed that high external K + (Fig. 4A), 0.3 mM quinidine (Fig. 4B) and 3 mM bupivacaine (Fig. 4C), but not 4-AP (Fig. 4E), depolarized V m and, (ii) verified the depolarizing effect of clofilium (Fig. 4D). These data therefore confirm that block of the human sperm K + conductance causes depolarization, but it was also clear that there were differences among the responses to the different agents tested. Clofilium thus depolarized resting V m to 0 mV ( 4D) and bupivacaine acted relatively slowly and the response to bupivacaine was biphasic (Fig. 4C). The physiological basis of these discrepancies was not investigated further.

Currents evoked by step depolarization
To investigate the biophysical properties of the human sperm conductance further, we characterized the currents evoked by step depolarizations using an experimental protocol that allowed us to subtract the background 'leak' currents that flow passively through voltage-independent ion channels or across the seal resistance itself (see Materials and Methods). The important point about this experimental design is that it enabled us to isolate the voltage-induced component of the membrane current. Figure 5A thus shows voltage-evoked currents induced by stepping V m to values between 252 and 68 mV. Depolarization consistently evoked outward current that developed over 300 ms (Fig. 5A) and analysis of the currents evoked by a step to 68 mV showed that the development of this current followed a time course that was accurately modelled as the sum of two exponential processes. The time constants Sperm K + currents associated with the fast (t Fast ) and slow (t Slow ) components of this response were 10 and 90 ms, respectively (Fig. 5B). Both parameters were independent of V m , and the kinetics of current activation are therefore independent of voltage ( Fig. 5B). The currents evoked by depolarization to 212 mV were too small to be accurately modelled in this way and this response was best described by a single exponential with a time constant of 70 ms (Fig. 5B). To quantify the effect of depolarization on membrane conductance we measured the voltageevoked currents flowing during the final 100 ms of each voltage pulse and used these data to quantify the voltage-induced increase in total membrane conductance (G V , i.e. I V /V m , Fig. 5C). Analysis of a solution to the Boltzmann Equation fitted to these data by non-linear regression showed that half-maximal activation occurred at 25 mV whilst the Boltzmann slope constant (k B ), which describes the channels' sensitivity to changes in voltage, was 20 mV 21 (Fig. 5C).
Whilst the control data in Fig. 6A confirm that depolarization evokes outward current, this figure also includes data recorded using a pipette solution modified by replacing K + with Na + . Whilst the response is smaller than normal, depolarization also induces outward current under these conditions. Separate experiments showed that this voltage-induced current was entirely abolished by replacing pipette K + with NMDG + (n ¼ 9) and, as the Na + -, K + -and NMDG + -based pipette solutions all contained identical concentrations of Cl 2 , the fact that we observed no voltage-induced current using the NMDG + -based solution shows that the voltage-induced currents in Fig. 6A must be carried by cations. Moreover, since V m was stepped to a value identical to the Na + equilibrium potential (E Na , i.e. 68 mV), the control currents in Fig. 6A can only be carried K + , and we could thus quantify the voltage-induced increase in K + conductance (G K ) using the equation . Similarly, the outward currents recorded using the Na + -rich pipette solution must be carried by Na + since this solution was entirely devoid of K + . We could thus quantify the voltage-induced increase in G Na using the equation G Na ¼ I m /DC Na . Although the voltage-induced current recorded using the Na + rich pipette solution was only 7.5% of that seen under control conditions (Fig. 6A), analysis of these data indicated that G Na was 15% of G K , and the apparent discrepancy between magnitudes of the recorded currents and the calculated conductance reflects the fact that C Na is smaller than C K . These experiments therefore show that G K /G Na in depolarized cells was 7 (Fig. 6B). Figure 6C shows data subsequently recorded from those cells stable enough to allow the recording to be repeated 20 -30 s after external Ca 2+ /Mg 2+ had been withdrawn (see Materials and Methods). It is clear that the currents recorded using either K + -based and Na + -based pipette solutions are larger than normal and further analysis showed that G Na /G K was now 1 (Fig. 6D). The modest degree of K + selectivity described above therefore depends upon external Ca 2+ /Mg 2+ . Pharmacological profile of the outward current recorded using Na 1 -based pipette solution The data presented in Fig. 7 confirm that depolarization evoked 2-5 pA pF 21 of outward current when Na + -based pipette solutions are used, whilst analysis of data recorded after 20-30 s exposure to putative blockers showed that 0.3 mM quinidine (Fig. 7A), 3 mM bupivacaine (Fig. 7B) and 50 mM clofilium (Fig. 7C) caused .80% block of this small Na + current. 4-AP was ineffective (Fig. 7D).
Quinidine, bupivacaine and clofilium, but not 4-AP, block CatSper As anticipated by earlier work (Lishko et al., 2011a;Strünker et al., 2011), inward and outward currents were recorded using bath and pipette solutions devoid of Ca 2+ /Mg 2+ that contained Cs + as the principal cation  5) and (E) 2 mM 4-AP (n ¼ 10) upon the zero current potential, which provides a read out of V m . The right-hand section in each panel; shows the pooled data (mean + s.e.m) derived from the entire series of experiments. Asterisks denote statistical significant differences between the control and experimental data (P , 0.001, Student's paired t-test).
Quindine-and clofilium-induced block of the K 1 current and CatSper Figure 10 shows the results of experiments that compared the effects of brief (1 min) exposure to 0.3 mM quinidine and 50 mM clofilium upon the current induced by repeated ramp depolarizations. Since we have shown that the voltage-induced K + current develops relatively slowly (Fig. 5), the voltage ramps used in the present studies were modified so that the cells were depolarized over 5 s. The mean current flowing during the final 200 ms of each voltage ramp was then quantified as a measure of the outward current (I Out ). The magnitude of I Out was normally 50 pA pF 21 and the data in Fig. 10A clearly show that exposure to quinidine rapidly (10 -15 s) inhibits this current, but that I Out quickly returns to its initial, control value once this drug is withdrawn. Figure 10A also includes pooled data that show I m -V m relationships constructed using the data recorded (i) under control conditions at the onset of the experiment; (ii) once the inhibitory effect of quinidine was fully developed and (iii) 3 min after the drug was washed from the bath. Analysis of these data confirmed that quinidine causes virtually complete (95.2 + 0.7%, P , 0.001, one way ANOVA/Dunnet's post hoc test) block of I Out and, as anticipated, this was accompanied by depolarization of V m (control: 219.3 + 4.0 mV; Quinidine: 21.0 + 0.01 mV, P , 0.001). Analysis of data recorded after this drug had been washed from the bath showed that I Out had virtually returned to its control value (97.3 + 4.1% recovery) and, similarly, V m had returned to a value (227.8 + 4.4 mV) that did not differ significantly from that measured at the start of the experiment. Figure 10B shows the results of (B) The responses to step depolarization consistently followed time courses that were very accurately modelled as the sum of two exponential processes. The time constants (t) for the fast and slow components of this response were calculated by non-linear regression and plotted against V Test . All data are mean + s.e.m. and leak/capacitative currents were subtracted on line in order to isolate the voltage-induced component of the total membrane current (I V ). (C) Steady-state values of I V were quantified over the final 50 ms of each voltage step, and used to quantify the voltage-induced increase in membrane conductance (G V ); the results of this analysis are plotted against V Test and the solid line shows a solution to the Boltzmann Equation fitted to these data by non-linear regression. The Boltzmann constant (k B ) and the voltage required for half-maximal activation (V 50 ) are presented. experiments that used the same method to explore the effects of 50 mM clofilium. As anticipated (see above) clofilium suppressed the recorded current although this block developed over 1 min and thus had a slower onset that the effects of quinidine (Fig. 10B). Analysis of data recorded once this effect was fully developed revealed substantial (91.4 + 0.3%, P , 0.001) inhibition of I Out and a clear depolarization (control: 231.9 + 4.1 mV, clofilium: 21.0 + 0.01 mV, P , 0.001). However, analysis of data recorded 5 min after this substance had been washed from the bath revealed negligible recovery of I Out (4.0 + 2.5% recovery) and no restoration of V m (21.0 + 0.01 mV). Whilst the quinidine-induced block of the voltage-induced outward current is fully reversible, the effects of clofilium do not reverse over the time scale of the present experiments. Moreover, an initial series of experiments (n ¼ 4) showed that progesterone increased the magnitude of the Cs + current flowing at both positive (68 -73 mV; control: 153 + 33 pA pF 21 ; progesterone: 285 + 33 pA pF 21 ; P , 0.001) and negative (283 to 285 mV; control: 256 + 19 pA pF 21 ; progesterone: 2148 + 34 pA pF 21 ; P , 0.01) potentials. Figure 10 also includes the results of experiments that used a directly analogous protocol to explore the effects of quinidine and clofilium upon the CatSper-dependent Cs + current that can be recorded under physiological conditions. These studies confirmed that 0.3 mM quinidine also causes substantial (91.0 + 1.6%, P , 0.001) block of CatSper (Fig. 10C). This block had a rapid (10-15 s) onset and was almost fully (87.1 + 9.3% recovery) reversible (Fig. 10C). Whilst clofilium (50 mM) blocked CatSper as effectively as quinidine (91.9 + 1.2% inhibition, P , 0.001), full inhibition developed over 1 min (Fig. 10D) and, although slight recovery was seen (Fig. 10D), the currents 5 min after the drug had been washed from the bath revealed only modest (23.2 + 4.3%) recovery. Indeed, the current recorded under these conditions did not differ significantly from the current measured in the presence of clofilium. This drug therefore causes essentially irreversible block of CatSper.
Biophysical properties of the channels underlying the voltage-induced 'tail current' To explore the conductive properties of the ion channels that underlie the 'tail' current shown in Fig. 1A, we initially held V m at a strongly Figure 6 Ionic selectivity of the voltage-induced conductance.
(A) Currents evoked by step depolarization to 68 mV were quantified both under control conditions (n ¼ 18) and using Na + -rich pipette solution (n ¼ 38) in cells exposed to the standard bath solution (SBS). (B) K + and Na + conductances quantified by analysis of data in (A). (C) Currents evoked using an identical voltage pulse that were subsequently recorded from the spermatozoa that were stable enough (standard pipette solution, n ¼ 3; Na + -based pipette solution, n ¼ 5) to allow the standard bath solution to be exchanged for a bath solution devoid of divalent cations (DVF). (D) K + and Na + conductances quantified by analysis of data in (C). All data are mean + s.e.m, asterisks denote statistically significant effect of replacing pipette K + with Na + (***P , 0.001, Student's t-test).

Figure 7
Effects of K + channel blockers on the outward Na + currents (I Na ). Currents evoked by step depolarization to 68 mV (top panel) were recorded using the Na + -rich pipette solution; left-hand panels show continuous recorded of I m whilst the right-hand panels shows mean currents quantified over the final 300 ms of the voltage pulse. In each experiment data were recorded during exposure to standard bath solution (control) and after 30 -60 s exposure to 0.3 mM quinidine (A, n ¼ 5); 3 mM bupivacaine (B, n ¼ 6); 50 mM clofilium (C, n ¼ 5) and 2 mM 4-AP (D, n ¼ 5). All data are mean + s.e.m.; asterisks denote statistically significant effects of the test substances (**P , 0.02; ***P , 0.001; Students paired t-test. Sperm K + currents depolarized value in order to activate the channels, and then stepped to a series of test values (V Test , Fig. 11A). Since leak/capacitive currents were subtracted (see Materials and Methods), the current measured immediately after the transition to V Test (I Tail , Fig. 11B) reflects current flow through channels opened by depolarization. Experiments undertaken under standard conditions showed that the I Tail -V m relationship was essentially linear (Fig. 11C) indicating that the channels do not display intrinsic rectification. Moreover, the channels cannot be K + selective since V Rev (244.9 + 2.9 mV) differed from E K (P , 0.0001, one sample t-test). As we do not observe Cl 2 current under the present conditions our subsequent analyses were based upon the assumption that these currents are carried by K + and Na + . The channels' fractional permeability to K + (P K ) and Na + (P Na ) were therefore assigned initial, arbitrary values that were used to predict V Rev from the Goldman -Hodgkin -Huxley (GHK) equation. The solution to this equation that best described the observed value of V Rev was then identified by reiteratively adjusting P K and P Na . This analysis showed that P K /P Na was 3.0 + 1.1. Figure 11D shows that brief (20 -30 s) exposure to K + -rich bath solution increased the magnitude of I Tail and depolarized V Rev to a value close to zero (P , 0.001). The observed shift in V Rev (44.1 + 1.9 mV) was virtually identical to that predicted by the GHK equation (45 mV) for a conductance with the degree of K + versus Na + reported above. Further experiments (n ¼ 5) used an analogous approach to measure the change in V Rev induced by lowering bath Na + to 11 mM by isoosmotically substituting NMDG + , a nominally impermeant ion. This response (215.7 + 1.5 mV) was also virtually identical to that predicted by the GHK equation (215 mV). The channels that underlie I Tail thus display modest ( 3-fold) K + versus Na + selectivity.
Whilst repolarization consistently induced I Tail ( Fig. 11A and B), this current was transient at hyperpolarized potentials and thus decayed rapidly to a stable value that was maintained throughout the remainder of the test pulse. Analysis of the 'steady-state' current (I Steady state ) recorded during the final few ms of each test pulse thus allows us to characterize the sustained voltage-induced outward current, and this analysis confirm that maintained depolarization induced a sustained outward current that is carried by K + .

Effects of K 1 channel blockers on the tail current
The control data in Fig. 12 confirm (i) that stepping V m to a series of test values (Fig. 12A) evokes sustained outward current (Fig. 12B) and (ii) that the subsequent repolarization induces I Tail (Fig. 12B). Since the protocol used here (Fig. 12A) implies that I Tail is always quantified at 292 mV the electrochemical driving forces on Na + and K + will be constant. The magnitude of I Tail will therefore depend upon the extent that the channels that underlie this current had become active during the preceding depolarization. Analysis of the I Tail -V Test relationship (Fig. 12E) therefore shows (i) that these channels normally become active at 0 mV, (ii) that the voltage needed for half-maximal activation (V 50 ) is 40 mV and The main panel shows mean currents evoked by a step depolarization to 68 mV recorded using the Na + -rich pipette solution whilst the right-hand panel shows mean currents quantified over the final 300 ms of the voltage pulse. Data were recorded during exposure to the standard bath solution (control) and after 20 -30 s exposure to 2 mM NNC55-0396. All data are mean + s.e.m and asterisks denote values that differed significantly from control (P , 0.001, Student's paired t-test).
(iii) that full activation occurs at 75 mV (Fig. 12E). Figure 12 also includes data recorded after 20-30 s exposure to 0.3 mM quinidine (Fig. 12C) and, as anticipated (see above), this substance abolished the voltage-induced outward current (Fig. 12D). However, despite this clear and consistent effect, repolarization did induce I Tail in quinidinetreated cells (Fig. 12C and E) and analysis of these data showed that this substance augmented this current but had no effect upon V 50 (Fig. 12E -G). Figure 13 shows data from a series of experiments that used an identical protocol to explore the effects of 2 mM 4-AP. These data confirm that 4-AP does not suppress the voltage-induced outward current (Fig. 13A -C) but, despite this clear finding, 4-AP did enhance I Tail without affecting V 50 . An additional series of experiments (not shown) which followed an identical approach confirmed that 50 mM clofilium abolished the voltage-induced outward current but this substance, in contrast to quinidine and 4-AP, also inhibited I Tail (Control: I Tail ¼ 218.7 + 3.7 pA pF 21 ; clofilium: I Tail ¼ 25.4 + 0.5 pA pF 21 , n ¼ 5, P , 0.02, Student's paired t-test).

Effects of progesterone on the tail current
Brief exposure to progesterone (0.5 mM, 2 -3 min) had negligible effect upon the voltage-induced outward currents (Fig. 14A-C) but enhanced I Tail both by augmenting the current induced by maximally effective voltage steps and by causing a leftward shift in the I Tail -V Test relationship and so that V 50 shifted from 40 to 20 mV ( Fig. 14A-C).

Discussion
The successful application of the whole cell recording technique (Hamill et al., 1981) to mouse (Kirichok et al., 2006;Santi et al., 2009Santi et al., , 2010Santi et al., , 2013 and human (Lishko et al., 2011a, b;Strünker et al., 2011;Orta et al., 2012;Mannowetz et al., 2013) sperm has allowed great progress to be made towards identifying and characterizing the ionic currents that flow across the membranes of these cells. In most instances, these studies have used recording conditions optimized for isolation and/or enhancement of specific currents, whereas the present experiments were undertaken using intracellular and extracellular salines that preserved physiologically relevant Na + , K + and Cl 2 gradients. Under these conditions the dominant membrane current was a voltage-gated cation conductance with low K + versus Na + selectivity ( 7:1) that allowed hyperpolarizing K + current to flow at potentials . 230 mV. This conductance was clearly important to the maintenance of resting V m since high external K + caused depolarization. The whole cell currents described here do, however, differ slightly from those reported in another recent study of human sperm (Orta et al., 2012). Although the principal aim of this study was to characterize the human sperm Cl 2 conductance, an initial series of experiments were undertaken using a K + -rich pipette solution in which [Ca 2+ ] i was buffered to a value that approximates to the normal resting level ( 0.1 mM). However, whilst our data consistently show an outwardly rectified current which reversed at approximately 230 mV, this earlier study described an essentially linear I m -V m relationship that reversed at approximately 212 mV with 200 pA of inward current at a potential of 2125 mV (Orta et al., 2012). However, our standard pipette solution contained only 30 mM Cl 2 and was slightly hypotonic whilst the pipette solution used in the earlier experiments contained 130 mM Cl 2 . We chose to work under these conditions since earlier studies of epithelial cells showed that isotonic pipette solutions containing high [Cl 2 ] promote cell swelling and this, in turn, can activate 'volume-sensitive' conductances for Cl 2 and K + (Worrell et al., 1989;Macri et al., 1993).

Sperm K + currents
Since such channels are present in human sperm (Yeung et al., 2005), their activation may explain the discrepancy between the two studies. Indeed in their subsequent experiments Orta et al. (2012) routinely used slightly hypertonic bath solutions to prevent the activation of such conductances.
Studies of human sperm using voltage-sensitive dyes suggest that V m is normally approximately 240 mV for non-capacitated cells (Blackmore et al., 1991;Linares-Hernandez et al., 1998) and approximately 250 mV for capacitated cells (Patrat et al., 2002) and, since similar values have been reported in mouse and bull (Zeng et al., 1995;Arnoult et al., 1999), capacitation seems to be associated with hyperpolarization. Although the cells used in the present study were exposed to capacitating medium, our estimates of V m are slightly less negative than those reported in earlier studies. Since it is now clear that several key components of the capacitation process are reversible (Bedu-Addo et al., 2005), it is possible that that the effects of incubation in capacitating conditions may not have been maintained during recording. Moreover, since low molecular weight substances (e.g. nucleotides, amino acids and sugars) are lost from the cytoplasm during whole cell recording (Hamill et al., 1981), we cannot exclude the possibility that such substances may be needed to maintain a fully polarized membrane potential. Pharmacological/biophysical properties of the human sperm K 1 conductance Although the hyperpolarizing K + current in human sperm was suppressed by acidification of pH i , this effect was modest and even at pH i 6.2, the residual K + conductance was large enough to maintain V m . In contrast, lowering pH i , 7.0 depolarizes V m of mouse sperm by inducing a profound fall in G K . The K + channels in mouse are thus more sensitive to changes in pH i than their human counterparts (Lishko et al., 2011b). The fact that G K displays such strict dependence upon pH i in mouse implies that the cells will hyperpolarize in response to cytoplasmic alkalinization and this provides a physiological basis for at least part of the hyperpolarizing shift in V m that occurs upon capacitation (Navarro et al., 2007;Martínez-Ló pez et al., 2009;Santi et al., 2010;Zeng et al., 2011). Whilst capacitation in human sperm also seems to involve hyperpolarization (Blackmore et al., 1991;Linares-Hernandez et al., 1998;Patrat et al., 2002), the present data show that the mechanisms that explain this process in mouse (Navarro et al., 2007;Martínez-Ló pez et al., 2009;Santi et al., 2010;Zeng et al., 2011) cannot necessarily be applied to humans.
Our data show that the human sperm K + conductance is blocked by quinidine, bupivacaine and, to a lesser extent, by lidocaine whilst 4-AP was ineffective. Moreover, experiments in which V m was directly monitored showed that quinidine, bupivacaine and clofilium, but not 4-AP, caused depolarization, and these data clearly confirm that K + channels are necessary for the maintenance of V m . These findings accord with data from mouse where the hyperpolarizing K + currents display a similar pharmacological profile (Navarro et al., 2007;Martínez-López et al., 2009;Santi et al., 2010;Zeng et al., 2011). However, rather than abolishing V m , quinidine and bupivacaine shifted this potential to a positive value. It is therefore interesting that these two compounds were the most effective blockers of the hyperpolarizing K + current and the fact that V m becomes positive when G K is blocked must indicate the presence a second conductance that mediates depolarizing current, and such current must be carried by Na + and/or Ca 2+ . Whilst the present data show that G K maintains V m under the physiological conditions, changes to the activity of this second conductance would allow control over this potential. In this context, it is interesting that in addition to the Slo3-encoded K + conductance, mouse sperm do appear to express epithelial Na channels (ENaC) that allow depolarizing Na + currents to influence V m . Indeed, inhibition of ENaC seems to contribute to the  Sperm K + currents hyperpolarizing shift in V m that is associated with capacitation (Hernandez-Gonzalez et al., 2007;Escoffier et al., 2012).
Experiments in which V m was stepped to a series of test potentials confirmed that depolarization evokes hyperpolarizing K + current in human sperm. However, though human spermatozoa do express protein and mRNA encoding 'classical' voltage-gated K + channels Cooper, 2001, 2008;Barfield et al., 2005a, Yeung et al., 2005, the currents reported here are not consistent with activity of these channels. Upon depolarization the current developed relatively slowly and the kinetics of current activation were independent of voltage. Half-maximal activation occurred at 25 mV whilst k B , which describes the channel's sensitivity to changes in voltage, was 20 mV 21 . Equivalent values for voltage-gated K + channels are approximately 220 and 6 mV 21 , respectively (reviewed by Wulff et al., 2009) and, in comparison, the K + channels in human spermatozoa thus display only very weak voltagedependence. These characteristics resemble those of the K + conductance found in mouse sperm (Navarro et al., 2007;Martínez-López et al., 2009;Santi et al., 2010;Zeng et al., 2011). Moreover, when pipette (cytoplasmic) K + was replaced by Na + depolarizing steps evoked a small outward current that displayed the same pharmacological profile as the K + current. The simplest explanation of these data is that this Na + current flows via the same population of ion channels as the K + current. Calculation of relative permeability for K + versus Na + gave a value for selectivity of 7. Again, this resembles the characteristics of the mouse K + conductance (Navarro et al., 2007;Martínez-Ló pez et al., 2009;Santi et al., 2010;Zeng et al., 2011).

Identity of the primary K 1 channel in mouse and human sperm
The hyperpolarizing K + currents in mouse sperm is believed to flow via channels encoded by Slo3 (KCNMA3) (Navarro et al., 2007;Santi et al., 2010;Zeng et al., 2011). These channels resemble the endogenous K + channels in mouse and human since (i) they are blocked by quinidine and clofilium but not by external 4-AP; (ii) are only weakly activated by depolarization and (iii) display poor K + /Na + selectivity (Schreiber et al., 1998;Martínez-Ló pez et al., 2009). Moreover, like the K + conductance and membrane potential of mouse sperm, Slo3-encoded channels are sensitive to changes in pH i, an effect that reflects altered channel  gating rather than an effect upon the permeability of the channel pore (Zhang et al., 2006a, b). Finally, Slo3 gene deletion abolishes the hyperpolarization seen during capacitation and mimics the effects of K + channel blockade (see Barfield et al., 2005a, b) by impairing progressive motility, suppressing the acrosome reaction and disrupting the control of cell volume (Santi et al., 2010;Zeng et al., 2011). However, despite these clear findings, heterologous expression studies show that Slo3-encoded K + channels are virtually inactive at potentials ,0 mV whilst it is abundantly clear that K + currents can be recorded from mouse (Navarro et al., 2007;Martínez-Ló pez et al., 2009;Santi et al., 2010;Zeng et al., 2011) and human at such potentials (see also Lishko et al., 2011b). This may reflect a requirement for interaction with the auxiliary subunit LRRC52 (leucine-rich repeat-containing protein no. 52) that is also found exclusively in male germ cells. Indeed, co-expression with of Slo3/LRRC52 modifies the behaviour of Slo3-encoded K + channels such that the current-voltage relationship more closely resembles that recorded from sperm themselves Yan and Aldrich, 2012).
Whilst these data are consistent with idea that the K + channels in mouse and human are encoded by Slo3, recent studies have shown that charybdotoxin, paxillin and iberiotoxin all block the human sperm K + conductance but have no effect upon the equivalent conductance in mouse (Mannowetz et al., 2013). Since these three substances are all thought to block the channels encoded by Slo1 and not Slo3 (Tang et al., 2010), these new data provide strong evidence that different K + channel subtypes underlie G K in mouse and human. Indeed, the fact that changes in pH i had only minor effects upon the K + current recorded from human sperm (see above) does tend to support this hypothesis since it is abundantly clear that the K + channels in mouse sperm are very sensitive to changes in pH i (Navarro et al., 2007;Santi et al., 2010;Zeng et al., 2011). However, recent experiments that directly compared the biophysical properties of mouse and human Slo3/ LRRC52 showed that the human channel complex could still pass hyperpolarizing K + current when pH i was ,7.0 (Leonetti et al., 2012), a result which accords well with the K + currents which we now describe in human spermatozoa themselves. The K + conductance associated with mouse Slo3/LRRC52, on the other hand, was essentially inactive under such conditions (Leonetti et al., 2012), a result that accords well with electrophysiological data derived from mouse spermatozoa (Navarro et al., 2007;Santi et al., 2010;Zeng et al., 2011). The K + channels encoded by human and murine Slo3 therefore display different biophysical properties. Moreover, we also show that the human sperm K + current is blocked by NNC55-0396 and mibefradil and, although these drugs are not usually considered to be K + channel blockers, they do seem to block Slo3 (Navarro et al., 2007;Zeng et al., 2011). Moreover, whilst Slo1 encoded K + channels display a very high degree of K + selectivity (Hoshi, 2012), the hyperpolarizing K + current seen during sustained depolarization flows via a population of ion channels that displayed only modest ( 7-fold) Na + versus K + selectivity.

Possible involvement of CatSper
Although CatSper forms a hormone-sensitive Ca 2+ channel under physiological conditions, this channel becomes freely permeable to monovalent cations if Ca 2+ /Mg 2+ are withdrawn and its activity has thus been assessed by monitoring Cs + current that can flow through the channel under divalent-free conditions (Kirichok et al., 2006;Lishko et al., 2011a). This characteristic of CatSper, which is a result of divalent cation binding within the channel pore, can explain earlier observations which showed that divalent cation-free (or depleted) medium causes enhanced Na + influx and depolarization of human sperm (Gonzáles-Martinez, 2003;Torres-Flores et al., 2011) and can also account for the loss of K + versus Na + selectivity that we observed in medium devoid of divalent cations medium. However, NNC55-0396 and mibefradil, structurally related compounds that block CatSper, suppressed the hyperpolarizing K + recorded under standard (physiological) conditions. Moreover, the effects of quinidine, bupivacaine, clofilium and 4-AP upon the CatSper-dependent Cs + current seen under divalentfree conditions were indistinguishable from their effects on the K + current recorded under standard conditions. It is therefore interesting that recordings of currents from sperm of mice null for Slo3 and/or CatSper1 show that hyperpolarizing K + current can flow though CatSper at potentials greater than 30 mV (Zeng et al., , 2013. Moreover, although mouse KSper and CatSper thus appear to share many pharmacological features, the clofilium-induced block of Slo3 was essentially irreversible whilst this drug's effect on CatSper reversed rapidly (Navarro et al., 2007;Zeng et al., 2011). Since this seems to provide a way of distinguishing between the two channel types , we undertook a detailed series of experiments that compared the effects of quinidine and clofilium upon the hyperpolarizing K + current and the CatSper-dependent Cs + currents in human sperm. Quinidine caused reversible block of both currents, consistent with data from mouse (Navarro et al., 2007;Zeng et al., 2011) but unlike the mouse, clofilium caused essentially irreversible block of both currents. Thus, in human sperm it is very difficult to distinguish K + -channel currents from monovalent CatSper on pharmacological grounds and, as in mouse (Zeng et al., , 2013, a part of the hyperpolarizing K + current may flow via CatSper. However, not all of our data were consistent with this hypothesis since progesterone augmented the CatSperdependent Cs + current (Kirichok et al., 2006;Lishko et al., 2011a;Smith et al., 2013) but had only a negligible effect upon the hyperpolarizing K + current. This result therefore suggests that most of the K + current must flows via a separate population of K + channels.

The tail current
Repolarization of V m after a test depolarization consistently induced transient inward current (I Tail ), and ionic substitution studies showed that these currents flowed via channels that were less K + selective than those underlying the sustained outward current. Furthermore, quinidine augmented I Tail despite causing full block of the hyperpolarizing K + current whilst 4-AP also augmented I Tail with no effect upon the sustained K + current. Clofilium, on the other hand, blocked both currents. There are therefore clear pharmacological and biophysical differences between the channel populations that underlie these two currents and depolarization must therefore activate at least two K + -permeable channel types. As far as we are aware, this is the first evidence that quinidine and 4-AP can activate any type of ion channel and these unusual responses could be highly significant since both substances can induce a 'hyperactive' pattern of motility (Barfield et al., 2005a;Alasmari et al., 2013b). Earlier studies have assumed that quinidine caused hyperactivation by blocking K + channels (Barfield et al., 2005a) whilst the effects of 4-AP have been attributed to changes in pH i and the mobilization of Ca 2+ from an intracellular store (Alasmari et al., 2013b). The fact that these substances can both activate ion channels raises the possibility that the channel underlying I Tail may contribute to the control of motility. Moreover, progesterone, which had no effect upon the sustained outward K + current but did augment I Tail , induces hyperactivation in a proportion of human sperm (Uhler et al., 1993;Fabbri et al., 1998;Teves et al., 2006;Sagare-Patil et al., 2012;Alasmari et al., 2013a, b). This response to progesterone may be critical for progress through the female tract and successful interaction with the egg and it is therefore interesting that spermatozoa from men with clinically identified fertility defects show that impaired activation by progesterone and 4-AP correlate well with reduced fertilization capacity (Alasmari et al., 2013a).
The ion channels that underlie I Tail displayed weak dependence upon V m and, since these channels are normally inactive at approximately 230 mV, current flow through these channels cannot contribute to the resting membrane potential under the conditions of the present experiments. However, it is possible that the activity of these channels may be modified by diffusible factors that would be lost from the cytoplasm once the whole cell recording configuration is established (Hamill et al., 1981) and we therefore cannot exclude the possibility that these channels may be important to the control of V m in intact spermatozoa. However, it is interesting that, as well as increasing the magnitude of I Tail , progesterone caused a hyperpolarizing shift in V 50 that allowed the I Tail to be activated by weaker depolarizations. Progesterone-induced activation of CatSper is now well documented and this response has been studied by quantifying changes to the Cs + current recorded under divalent-free conditions or to the current carried by Ba 2+ (Lishko et al., 2011a;Strünker et al., 2011). We believe that our data are the first to show a progesterone-induced change to the conductive properties of spermatozoa exposed to quasiphysiological ionic gradients. Whilst the biological significance of this novel response is presently unknown, the importance of progesterone to the control of motility makes it important to characterize the progesterone-sensitive channels more fully and to establish the extent to which other substances that control sperm motility can influence their activity.

Summary
Electrophysiological studies of mouse sperm (Kirichok et al., 2006;Navarro et al., 2007;Santi et al., 2010;Zeng et al., 2011) have led to the identification of two cation-permeable conductances. The first of these is a pH-sensitive K + conductance that sets the membrane potential and is almost certainly encoded by Slo3/LRRC52, whilst the second is a Ca 2+ -permeable channel encoded by members of the CatSper gene family (Ren et al., 2001). Although other cation-permeable conductances have been identified (see, for example, Felix et al., 2002;Acevedo et al., 2006;Martínez-López et al., 2009), studies of knockout mouse indicate that it is Slo3/CatSper that dominate the conductive properties of murine sperm (Zeng et al., 2013). The present electrophysiological studies of human sperm exposed to 'physiological' ionic conditions have identified a K + channel that is weakly activated by voltage (Fig. 15) and this conductance is broadly similar to that recently documented in separate studies (Mannowetz et al., 2013). However, whilst it has been suggested that this may flow via channels encoded by Slo1 (Mannowetz et al., 2013), the poor ionic selectivity and unusual pharmacological profile which we report are not consistent with this hypothesis. Moreover, we also show that depolarization activate a second voltage-dependent conductance that displays very poor K + selectivity and is subject to rapid inactivation. This current has a different pharmacological profile to both the sustained outward K + current and the CatSper-dependent Cs + current but, like CatSper, shows both stimulation and leftwardshift of I -V relationship in the presence of progesterone (Fig. 13). This previously undocumented conductance may thus play an important role in mediating the physiological effects of this hormone. experimental design. S.M.W., S.J.P. and C.L.R.B. were involved in the design of the study and obtained funding for the experiments. The initial funding was supported by grants from NHS Tayside, Infertility Research Trust (Barratt PI) and the Wellcome Trust (Publicover and Barratt PI). Additional funding was provided by MRC (MR/K013343/1, Wilson PI). S.M.W. and S.A.M. performed the detailed data analysis of the electrophysiological data. All authors contributed to the construction, writing and editing of the manuscript. All authors approved the final manuscript for submission.

Funding
This study was made possible by grants from the Wellcome Trust (086470), the Infertility Research Trust, NHS Tayside and the Medical Research Council (MR/K013343/1). Funding to pay the Open Access publication charges for this article was provided by Durham University.