The Ca²⁺-activated K⁺ current of human sperm is mediated by Slo3

[Editors’ note: the author responses to the first round of peer review follow.]

While our manuscript was under review, a competing manuscript by Mannowetz et al. (2013) entitled “Slo1 is the principal potassium channel of human spermatozoa” appeared in eLife. Each manuscript reaches entirely different conclusions regarding the molecular identity of the K⁺ channel underlying I_KSper current in human sperm: we conclude that it is Slo3, whereas Mannowetz et al. propose Slo1; both reports cannot be right at the same time.

The two groups agree that the primary human sperm K⁺ current is activated by Ca²⁺.

The major points of difference are: (1) we observe that the human sperm K⁺ current also retains some, albeit modest pH dependence, consistent with the modest pH dependence of heterologously expressed hSlo3; and (2) conflicting pharmacological and biochemical results concerning the molecular identity of the channel.

Concerning the pharmacology, we show that Slo3 inhibitors, but not Slo1 inhibitors, suppress I_KSper and depolarize human sperm. In contrast, Mannowetz et al. report that Slo1 inhibitors suppress I_KSper. Although we do not know the reason for this discrepancy, there are differences in experimental conditions, which we mention in the Discussion of our revised manuscript. Moreover, to ascertain that human Slo3 is insensitive to Slo1 inhibitors, we studied the action of IBTX, CTX, and paxilline on heterologous hSlo3 + LRRC52 channels, see Author response image 1 below). We find that human Slo3 is insensitive to IBTX, CTX, and paxilline. As positive control, we used the inhibition of heterologous hSlo1 by IBTX.

Author response image 1

Download asset Open asset

In our revised manuscript, we include additional data, showing that progesterone inhibits hSlo3 with a K_i of 17 μM, consistent with the inhibition of I_KSper in human sperm (Mannowetz et al. 2013). We further show that Slo1 is largely insensitive to progesterone. Moreover, we show single-channel recordings from human sperm that reveal a Ca²⁺-activated K⁺ channel of ∼ 60-70 pS conductance, i.e., similar to the conductance of human Slo3, but inconsistent with a Slo1 conductance of ∼ 240 pS. These new results further strengthen our conclusion that Slo3 underlies I_KSper in human sperm, but are inconsistent with Slo1 properties.

To implement the new data and to improve readability, we rearranged Figures 1-5 and we include two new Figures.

Reviewer 1:

The authors present compelling evidence that that Ksper current in human sperm is carried by the Slo3 potassium channel, as in mouse sperm, but the channel is gated by calcium, unlike in mouse where it is gated by protons. Heterologous expression studies confirm that human Slo3 seems to be calcium gated.

The results suggest that the Slo3 channel is gated by different ligands in different species. It is puzzling that the authors agree with the results of the Lishko lab that human Ksper is activated by calcium, but disagree as to which channel it is, as the Lishko paper argues that the mouse Ksper is in fact the (normally calcium gated) Slo1. I find no reason to doubt either study's conclusions and I expect the controversy to be settled with future work.

I believe the paper to be appropriate for publication in eLife, especially as the Lishko paper appeared there also.

We thank the referee for the positive comments on our manuscript.

Reviewer 2:

However, mechanistically and conceptually the advances of the current manuscript are limited. The salient role of Slo3 in mouse sperm has recently been highlighted by several groups. It is well documented in the literature that mouse Slo3 is rather insensitive to calcium and that human Slo3 coexpressed with LRRC52 is less sensitive to alkaline pH as compared to mouse. The authors confirm these observation by others by direct electrophysiological recordings in human sperm. However, no information is provided on the molecular basis underlying these functional differences. Thus, the present manuscript still leaves central mechanistic questions unanswered and the paper is confirmatory to quite some extent.

The referee’s impression seems to be that the main purpose of our manuscript was to establish that the K⁺ current in human sperm is regulated by pH, like mouse KSper and Slo3. If that were the case, we would certainly agree that the manuscript contributes little new information. However, the focus of the manuscript is in establishing that the human sperm K⁺ current is, in fact, regulated by Ca²⁺, with weaker regulation by pH. The prototypical Ca²⁺-activated and pH-activated members of this channel family are Slo1 and Slo3, respectively. Thus, the important question is whether the human sperm K⁺ current is carried by Slo1 or Slo3. Our results demonstrate that human I_KSper current arises from Slo3 and not Slo1. Furthermore, we show that heterologously expressed human Slo3 is Ca²⁺ regulated. These observations are entirely new and it is surprising that two orthologous proteins, mSlo3 and hSlo3, have such different regulation by cytosolic ligands.

Identification of the site on the hSlo3 gating ring underlying regulation by Ca²⁺ and H⁺ is highly interesting, but is beyond the scope of this manuscript. However, we acknowledge that the extensive introductory paragraph on Slo gating rings in our original manuscript might have led the referee astray. Therefore, we shortened this paragraph and moved it to the Discussion.

Reviewer 3:

This referee raised several concerns about the reliability of our electrical recordings from human sperm, in particular, the activation of I_KSper by Ca²⁺ and pH_i. This is somewhat surprising given that the referee might have also seen the Mannowetz et al. manuscript; at least in the case of Ca²⁺, this is not a point of contention between the two groups. Moreover, by manifold controls, we scrutinized the reliability of our recordings:

• To ascertain that the voltage-, pH_i-, and Ca²⁺-activated currents are carried by K⁺ channels, we determined the reversal potential V_rev of currents at high and low [K⁺]_o and [Cl^-]_o.

• To distinguish experimentally I_KSper from monovalent _ICatSper, we substituted intracellular K⁺ by Cs⁺ and used 1 mM Ca²⁺ in the extracellular medium.

• We show activation of I_KSper by alkalization both by superfusion with NH₄Cl and by using different pH values in the pipette. We studied I_Ksper at low and high [Ca²⁺]_i in one and the same cell by photorelease of Ca²⁺ from caged Ca²⁺, while monitoring [Ca²⁺]_i with a fluorescent indicator.

• Ca²⁺ activation of I_KSper was studied at 6 different intracellular Ca²⁺ concentrations, ranging from 2.5 to 1000 μM.

• pH_i- and Ca²⁺ activation of I_KSper was confirmed by current-clamp experiments.

• We tested Slo1 inhibitors both in voltage- and current clamp.

• We tested Slo3 inhibitors both in voltage- and current clamp.

• We studied both Slo1 and Slo3 expression in human sperm by targeted mass spectrometry.

• Finally, heterologous expression of Slo3 in both oocytes and CHO cells corroborated our electrophysiological and pharmacological conclusions drawn from sperm recordings.

Altogether, the identification of the hSlo3 protein in human sperm by two independent techniques, the pharmacological fingerprint of human I_KSper and hSlo3, and the Ca²⁺-activation of heterologous hSlo3 both in oocytes and CHO cells provide compelling support for our conclusion that the Ca²⁺-activated I_KSper in human sperm arises from hSlo3, but not Slo1.

There are three major conclusions in this paper. First, the human K⁺ current I_Ksper is activated by intracellular Ca²⁺, rather than pH increase as previously established in mice. Second, Ca²⁺ increase in human sperm under “physiological” conditions (presumably through Catsper) is sufficient to activate I_Ksper. Third, heterologously expressed human Slo3 (hSlo3) is activated by high [Ca²⁺]_i, but is less sensitive to pH. These findings, if confirmed, represent a major step forward in our understanding of sperm membrane potential regulation and fertilization. The third conclusion is very well supported by careful and well-controlled analysis using heterologous expression in both Xenopus oocytes and CHO cells, with or without the auxiliary subunit LRRC52. I find that the first two conclusions, though quite interesting, are not convincing at their present form.

We provide plenty of experimental evidence that Ca²⁺ activates I_KSper (see our general comment above). Therefore, we disagree with this statement. The issue concerning the activation of I_KSper by physiological Ca²⁺ concentrations is addressed in more detail below.

The major data supporting the conclusion that human I_Ksper is potentiated by Ca²⁺ but not pH_i is in Figures 2 & 3.

There seems to be a misunderstanding. We do not claim that human I_KSper is exclusively controlled by Ca²⁺. In fact, we show that at pipette pH 6.2, alkalization by NH₄Cl enhances currents by about 2-fold. Similarly, increasing pipette pH from 6.2 to 7.3 enhances human _IKSper by about 2-fold.

In Figure 2 where the authors want to show what Ca²⁺, not pH, controls V_m in human sperm, two [Ca²⁺]_i conditions are compared: 0 Ca²⁺ and 1 mM Ca²⁺. The 0 Ca²⁺ pipette had 1 mM EGTA with no added Ca²⁺, which results a free [Ca²⁺] of perhaps < 10 nM. The pipette solution contains no ATP or GTP. Such condition may be fine for studying the biophysical properties of a channel with voltage-clamping, but is too “unphysiological” for studying the regulation of membrane potential. It is hard to interpret the apparent lack of large effect on V_m by pH_i change under this condition. The authors also cited the paper by Navarro (PNAS 104:7688-7692) to compare the difference between human and mouse sperm in the effects of pH_i on V_m. However, there is large difference between the pipette solutions used in the two studies (Navarro et al. included ATP and GTP in the pipette and had apparently different free Ca²⁺ concentration). The authors should at least test several Ca²⁺ concentrations close to the physiological levels.

Both Ca²⁺ and pH do affect human I_KSper and alkalization does hyperpolarize sperm; yet, Ca²⁺ evokes a more pronounced hyperpolarization. To make comparison of current-clamp and voltage-clamp experiments meaningful, we used the same intracellular solution (without ATP/GTP) throughout the manuscript. Of note, Mannowetz et al. used a similar intracellular solution (also without ATP/GTP) to study human I_KSper. Under these conditions, our current clamp and voltage-clamp data are consistent, e.g., we observe modest pH_i control, but stronger Ca²⁺ control of _IKSper and V_m. Current-clamp experiments are always to a certain degree nonphysiological, no matter whether intracellular ATP/GTP is present or not: the cell’s cytoplasm is exchanged by artificial intracellular solutions and, moreover, experiments are hampered by the high input resistance of sperm (discussed in Zheng et al. 2013, JGP). Therefore, we have tempered our assertions that Ca²⁺ rather than pH_i controls V_m of human sperm to our recording conditions. Moreover, we now include a paragraph in the discussion stating that it will be an interesting question for future work to determine the relative contributions of cytosolic pH and [Ca²⁺] to the regulation of human sperm function.

In the current-clamp experiments, we used a pipette [Ca²⁺] of either a few nanomolar or 1 mM to study the action of Ca²⁺. As requested, we now include current clamp data at 1, 30, and 70 μM Ca²⁺, showing that the effect of 1 μM Ca²⁺ is miniscule, whereas 30 μM evokes a sizeable hyperpolarization that saturates at ≤ 70 μM. Thus, the concentration dependence of Ca²⁺ action is similar in current- and voltage-clamp experiments.

In addition, all the pipette solutions have 50 mM HEPES. It's not clear whether 10 mM NH₄Cl in the bath is able to significantly increase pH_i when 50 mM HEPES is included in the pipette, and this is critical to the conclusion that pH_i increase doesn't have much effect on V_m.

In fact, 10 mM NH₄Cl increases pH_i by about 1 unit under the conditions used (see response to Reviewer 2). The modest action of NH₄Cl on I_KSper and V_m is consistent with the modest changes of I_KSper and V_m observed when the pipette pH was changed from pH 6.2 to 7.3. Moreover, at 50 mM intracellular HEPES, we routinely observe NH₄Cl-induced CatSper activation (see for example Strünker et al. 2011, Brenker et. al. 2012). Altogether, these results show unequivocally that the modest action of pH_i on I_KSper and V_m is not due to the specifics of our experimental conditions.

Finally, the numbers of cells (n = 3,4) are too small to draw statistically significant conclusions, especially when measuring V_m of primary cells.

We take issue with this comment. First, the small s.d. shows that the V_m values are highly reproducible. Second, the current-clamp and voltage-clamp results are consistent with each other in all aspects. Third, we tackled each issue by at least 2 different experimental approaches, e.g., we study Ca²⁺ activation of I_KSper by both photorelease of Ca²⁺ and at different Ca²⁺ concentrations in the pipette; we examine pH_i activation of I_KSper by both superfusion with NH₄Cl and by different pipette pH; we study the pharmacology of Slo3 and Slo1 inhibitors in voltage- and current clamp; we study Slo3 currents both in oocytes and CHO cells. These different approaches all lead to the same conclusions. Finally, we note that for most experiments n≥4.

The native I_Ksper's Ca²⁺ sensitivity is not that strong, at least not until [Ca²⁺]_i reaches 100 μM (Figure 3I). Based on Figure 6, the authors conclude that physiologically stimulated [Ca²⁺] increases can reach the levels sufficient to activate I_Ksper. The Ca²⁺ measurements in Figure 6 with high concentration of progesterone (2 μM) do not have calibration (though with indicators of various Ca²⁺ affinity) and, more importantly, are from whole sperm. The majority of the signal is likely from sperm head where Slo3 is not localized (Figure 1H). Therefore, it's hard to make meaningful correlation between the Progesterone-induced Ca²⁺ increase and I_Ksper activation.

Figure 3 (now also Figure 2) and Figure 6 (former Figure 5, heterologous expression of Slo3) show that [Ca²⁺] (not 100 μM as this referee states) enhances both I_KSper and hSlo3 currents. We agree with the referee on potential limitations of the Ca²⁺ indicator experiments in terms of defining absolute concentrations. However, although Figure 8 (the former Figure 6) does not show calibrated changes in [Ca²⁺]_i, it clearly argues (in particular the responses recorded with the low-affinity indicator Fluo-5N, K_d = 90 μM) that [Ca²⁺]_i increases globally μM. In fact, it has been shown that progesterone-induced Ca²⁺ signals originate in the flagellum and propagate in a tail-to-head direction (Servin-Vences et al., 2013, Reproduction). Therefore, the local [Ca²⁺] in the flagellum is predicted to be even higher than in the head. However, we state now in the text that the interplay between CatSper and Slo3 requires further experimentation.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Unlike mouse KSper, the principal K channel in human sperm is primarily activated by internal Ca²⁺, but only modestly by alkaline pH. This agrees with Mannowetz et al.

We trust that the reviewer means that both studies show that human sperm K⁺ current is activated by Ca²⁺. The two papers reach different conclusions regarding the role of alkaline pH. Mannowetz et al. claim that human I_KSper is completely insensitive to pH_i, whereas we and a recent paper by Mansell et al. (Mol. Hum. Reprod. 2014) find that human I_KSper is enhanced by alkalization.

The primary disparity between the papers is that Mannowetz et al found that the specific Slo1 blocker, Iberiotoxin, blocked human KSper, while Brenker et al find that it does not. The experimental solutions used in the two studies may underlie this difference.

Although the pharmacological differences between the two studies remain difficult to explain, we note that the intracellular and extracellular [Ca²⁺] used in the pharmacological experiments differ. Overall, our pharmacological profile is consistent with the biochemical evidence for Slo3, the single-channel fingerprint, and the progesterone sensitivity of human I_KSper.

The following issues need to be addressed in the revision:

1) The authors presume that the rapid evolution of gamete genes results in gating ring mutations that confer these changes to human Slo3. It would be of interest to readers to show alignment of mouse, intervening species, and humans to narrow in on the changes most likely to mediate these properties. We do not suggest that the authors sort out the changes by mutagenesis, as this would be a separate work in itself and may take several years to complete.

We include the requested alignment. Unfortunately, the alignment does not provide insights regarding the molecular determinants underlying the different ligand sensitivities of human versus mouse Slo3. Except for the Ca²⁺ bowl residues, the determinants of the RCK1 high- affinity site and the lower-affinity site involve multiple residues scattered throughout RCK1. For lower-affinity sites, recognition of any “site” by sequence gazing would be remarkable.

2) The mass spec identification supporting the conclusion is not shown. To support the findings, indicators of merit (such as number of peptides and protein coverage etc.) for all the proteins identified and mentioned in the text should be provided as a Table in Supplementary Information.

We now include a Supplementary file with all indicators of merit for the mass spectrometric results

3) Although we understand the reasons for using 1 mM Ca²⁺ in some internal solutions to maximize the response, many readers will question the validity of this condition, and it will likely be a source of contention. This should be at least addressed fully in the discussion. Better would be to show data at intervening 10, 30, 100 µM Ca²⁺. However, the reviewers are not worried by the use of 1 mM Ca²⁺. The inclusion of data with other concentrations is optional.

Please note that, in Figure 2, we do, in fact, show current clamp data for 1, 30, 70, and 1,000 µM [Ca²⁺]_i. In Figure 3 and Figure 3–figure supplement C, we show I-V curves for I_KSper at 2.5, 10, 40, 100, and 1,000 µM [Ca²⁺]_i. Moreover, in Figure 3, we show original current traces at 0, 40, and 1,000 µM [Ca²⁺]_i. We feel this data should address the concerns of the reviewer.

There are three experiments where only 1 mM [Ca²⁺]_i was used:

• 1) to study the pharmacology of I_KSper,

• 2) to show that at high [Ca²⁺]_i, V_m is independent of [Cl^-]_o,

• 3) to show that at high [Ca²⁺]_i, I_KSper is still potentiated by alkalization.

For experiment 1, we provide the rationale for using 1 mM [Ca²⁺]_i in the results section and in the discussion. For experiments 2 and 3, it is also advisable if not required to use high [Ca²⁺]_i. Accordingly, we do not expect the high concentrations a potential source of contention. Therefore, we prefer to leave the Discussion as it is.

4) Introduction: Navarro et al definitively identified P2X2 channels, not P2X1, in mouse sperm through KO studies. The Brenker study does not identify the P2X channel subtype. After writing a whole paper on the mis-identification of KSper, it is not reassuring that you misidentify P2X in your own summary!

Although we assume that the reviewer made this comment in jest, our work was not about mis-identification of KSper, but simply reports our experiments showing that human KSper is regulated by both cytosolic calcium and alkalization, and that properties of heterologously expressed hSlo3 are consistent with human KSper. Moreover, we thank the referee pointing out the typo regarding P2X channels; we have made some clarifying changes to the respective sentence: “purinergic P2X channels are functional in mouse (Navarro et al., 2011), but not in human sperm (Brenker et al., 2012)”.