香京julia种子在线播放

Dimethyl sulfoxide (DMSO, cat. D2650), ammonium chloride (NH₄Cl, cat. A9434), nigericin (cat. N7143), progesterone (cat. P8783), and concanavalin A (cat. C2010) were purchased from Sigma–Aldrich. Pluronic F-127 (cat. P6867), SNARF-5F AM (5-(and-6)-carboxylic acid, acetoxymethyl ester) (cat. S23923), and 2′, 7′-Bis-(2-carboxyethyl)-5-(and-6)-carboxyfluorescein, acetoxymethyl ester (BCECF AM) (cat. B1170) were obtained from ThermoFisher Scientific.

Fluorescence spectra of SNARF-5F were determined with a Perkin-Elmer LS 55 (Perkin-Elmer, Waltham, MA, United States) fluorescence spectrometer using the software FL WinLab version 4.00.03 for data acquisition (Figure 1). SNARF-5F non-permeable and AM versions were used at 20 μM final concentration in 50 mM potassium phosphate buffer (see table in Supplementary Figure S1). Multiple spectra were acquired using various excitation wavelengths (405, 440, 465, 488, 510, 532, and 543 nm) and different pH_e (5.5, 6.0, 6.4, 6.8, 7.0, 7.2, 7.4, 7.8, and 8.2) (Supplementary Figure S1).

Motile mouse/human spermatozoa (10 × 10^6/ml) were incubated with 20 μM SNARF-5F AM in the presence of 0.1% pluronic F-127 during 90 min at 37°C with 5% CO₂ in the dark. The cells were washed once by centrifugation at 200 × g for 5 min and resuspended with fresh medium. To obtain fluorescence spectra of SNARF-5F incorporated into spermatozoa, human spermatozoa loaded with SNARF-5F AM were diluted in the media of different pH_e as described above and treated with 0.1% Triton-X 100 detergent. Fluorescence spectra of the lysed spermatozoa were acquired at excitation wavelengths 488 and 532 nm. For single cell recordings, spermatozoa were attached to Concanavalin A-treated coverslips for 2–3 min and mounted in recording chambers.

For BCECF experiments, motile mouse/human spermatozoa (10 × 10^6/ml) were incubated with 1 μM BCECF AM without pluronic acid during 15 min at 37°C with 5% CO₂ in the dark and the unincorporated dye was removed by centrifugation as the case of SNARF-5F.

Single cell images were acquired using two different setups: (1) Olympus iX71 LED-light source epifluorescence microscope and (2) Olympus iX81 laser widefield/total internal reflection fluorescence (TIRF) microscope (Olympus, Japan). The LED setup was equipped with a PlanApo N 60X/1.42 oil objective and a 3A 532 nm LED coupled to Opto-LED light controller (Cairn Research, United Kingdom). The laser setup was equipped with an Apo N (TIRF) 60X/1.49 oil objective and a 488 nm laser with high speed imaging shutter. To acquire dual emission images of SNARF-5F, an image splitter OptoSplit II (Cairn Research, United Kingdom) was used for both setups, LED (with band-pass filter ET 530/30X) and laser. To acquire the images with SNARF-5F, a wide band-pass filter ET 575/50 M (channel 1) and a band-pass filter ET 640/20 M (channel 2) were employed as dual emission filters combined with a dichroic mirror DC 610lp (Chroma Technology, United States) (Figure 2). For BCECF experiments, a 3.15 A 465 nm LED (Luminus Devices, Woburn, MA, United States) with bandpass filter HQ 480/40X was used for excitation light, combined with a dichroic mirror (Q505lp) and an emission filter (HQ 535/50M) (Chroma Technology, United States). Each setup has a 512 × 512 Andor iXon 3 EMCCD camera (model X3 DU897E-CS0) (Oxford Instruments, United Kingdom).

Images were acquired with the software Andor iQ version 2.9.1 (LED set-up) (Oxford Instruments, United Kingdom) and Xcellence version 1.2 (laser set-up) (Olympus, Japan). Fluorescence images of both SNARF-5F and BCECF were taken with 1 × 1 binning, 5 images/s (5 ips), with an exposure time of 10 ms for the LED setup and 30 ms (laser potency 35%) for the laser setup. Images were analyzed with ImageJ version 1.52n (NIH, United States), obtaining mean fluorescence intensities, selecting heads and flagellum as regions of interest.

To convert fluorescence data to pH values in vitro, the following equation is commonly used:

where R is the ratio Fλ₁/Fλ₂ of fluorescence intensities (F) measured at two wavelengths λ₁ and λ₂ and the subscripts A and B represent the values at the acidic and basic conditions, respectively (Whitaker et al., 1991). However, it is difficult to maintain live spermatozoa in highly acidic and alkaline condition to obtain F_B(λ2)/F_A(λ2) values from the same cells. Therefore, we performed in vivo pH_i calibration by fixing external pH (pH_e) between 6.0 and 8.0 as reported previously (Grillo-Hill et al., 2014). Briefly, spermatozoa suspensions were incubated for 15 min with calibration medium (in mM: 120 KCl, 25 HEPES, 1 MgCl₂, and 0.01 nigericin, at different pH_e: 6.0, 6.5, 7.0, 7.5, and 8.0, adjusted with KOH). We measured the fluorescence intensity at two emissions, 575 (channel 1) and 640 nm (channel 2), always subtracting the fluorescence background value in each channel. As fluorescence ratio values (R_F640/F575) have a lineal relation with pH_i values (between 6.0 and 8.0) (Figures 2D,E), we used the following equations to estimate pH_i from the fluorescence ratio values, for human: pH_i = (R_F640/F575 + 6.96)/1.22 and for mouse: pH_i = (R_F640/F575 + 7.52)/1.32.

The results are expressed as the mean ± SEM of at least three independent experiments (three different donors or mice), with a minimum of 200 cells per condition. The data were analyzed by a comparison test between the groups, using the non-parametric Mann–Whitney U-test with 95% statistical significance. The paired tests were carried out comparing the head and the flagellum of the same cell. Additionally, the Bonferroni correction was used when multiple comparisons were made.

To perform ratiometric fluorescence measurements with a good signal to noise ratio (S/N ratio), it is important to acquire bright fluorescence images in both channels. In other words, if we detect dim fluorescence signals in one channel, the S/N ratio of the dual-emission ratio values become undesirably low even when we detect bright signals in the other channel. As spermatozoa possess a quite reduced cytoplasm, it is crucial to use an appropriate excitation wavelength and emission filters. Therefore, we first determined the fluorescence spectra of SNARF-5F at several pH_e values (5.5–8.2), exciting with various wavelengths (405–543 nm). As shown in Supplementary Figure S1, the longer excitation wavelength (longer than 465 nm) gives the higher fluorescence intensities in all emission wavelengths we explored (550–750 nm). Namely, 543 nm produced the highest fluorescence values.

Figure 1 illustrates the fluorescence spectra of SNARF-5F at different pH_e (5.5–8.2) excited at 488 and 532 nm. At both exciting wavelengths, the fluorescence intensities around 575 nm (the first peak) decrease when the pH increases, while those of around 640 nm (the second peak) increase at the same conditions. When exciting at 532 nm, the relative fluorescence intensities within the first peak at different pH_e are much smaller than those within the second peak, as reported in the original article of SNARF-5F (Liu et al., 2001). Conversely, the relative fluorescence intensities within the two peaks became almost equal when 488 nm was used as excitation light (Figure 1A). In spite of this favorable feature, their absolute fluorescence intensities are small. Considering these characteristics, we selected 532 nm as the best compromise between brightness and peak balance for this study.

To evaluate the incorporation of the membrane permeant dye SNARF-5F AM into the spermatozoa, we incubated human spermatozoa with 20 μM of this dye for 90 min. After the excess dye was washed out by centrifugation, the cells were lysed with 0.1% Triton X-100 and the fluorescence spectra were acquired (Figure 1B). The spectra of the dye incorporated into human spermatozoa were not identical to SNARF-5F in vitro, suggesting that SNARF-5F AM was not completely hydrolyzed in the cell and/or some of the dye was bound to certain molecules of the cell. Nevertheless, SNARF-5F AM incorporated into the cell responded to pH_e changes similarly to SNARF-5F AM.

A conventional dual-emission fluorescence imaging setup usually is composed of an epi-fluorescence microscope, a CCD camera, and a filter wheel, which exchanges two emission filters alternatively. In this type of setup, there is always a time lag between the image in one channel and the image in the other channel. Since spermatozoa are small and motile cells, the presence of a time lag between two images of each channel is undesirable. Therefore, we used an image splitter (Kinosita et al., 1991) that allows the simultaneous capture of the images from the two channels with a single camera (Figure 2A). In a common configuration of dual-emission ratiometric imaging, emission lights are divided into two components (two channels) at the isosbestic point, around 595 nm in the case of SNARF-5F AM excited by 532 nm. However, because the fluorescence intensity of the first peak (575 nm) is lower than the second peak (640 nm) as described previously (Figure 1), we separated the emission light at 610 nm (about 15 nm longer than the isosbestic point) by a dichroic mirror. Consequently, we collected a wide range of wavelengths, 550–600 nm, as the fluorescence signals in the first channel (channel 1). Then, we collected 630–650 nm wavelengths as the longer wavelength fluorescence (channel 2). This configuration gives us comparable fluorescence intensities from the two channels without the insertion of a neutral density filter (Figure 2A).

Figure 2B shows fluorescence images (gray scale) of SNARF-5F in media at different pHs, clearly demonstrating the opposite changes of fluorescence intensity between Channel 1 and Channel 2. Since fluorescence spectra of SNARF-5F and SNARF-5F AM incorporated into human spermatozoa show a slight difference (Figures 1A,B), we performed in vivo calibration using human spermatozoa to convert the ratio fluorescence values into pH_i values. To perform in vivo calibration, spermatozoa pH_i was equilibrated to the pH_e using high K⁺ (120 mM) media in the presence of 10 μM nigericin (an ionophore that facilitates K⁺/H⁺ exchange across the lipid bilayer). Figure 2C shows fluorescence images (pseudo color) of human spermatozoa, whose pH_i was fixed at different pH_e (6.0–8.0).

The mean ratio values of fluorescence intensities of the two channels (F640/F575) in distinct pH_i are summarized in Figure 3C and these ratio values are plotted as a function of pH_i (Figure 2D). The ratio values increase proportionally to pH_i between 6.0 and 8.0 with excellent linearity (human spermatozoa: R_F640/F575 = 1.22 × pH_i – 6.96, R² = 0.99; mouse spermatozoa: R_F640/F575 = 1.32 × pH_i – 7.52, R² = 0.98) (Figure 2E).

BCECF is known to be quite phototoxic to spermatozoa and this effect can be easily detected as flagellar beat attenuation and as a decrease in the fluorescence intensity (photo-bleaching) during the intense exposure of excitation light (Nishigaki et al., 2006; González-Cota et al., 2015). In this study, we confirmed the phototoxic effect of BCECF on sperm using the same setup utilized for SNARF-5F (Supplementary Figure S2). Particularly, the 488 nm laser excitation attenuated the flagellar beat of human and mouse spermatozoa after around 60 and 20 s illuminations, respectively. Subsequently, notable photobleaching of BCECF was observed in both human and mouse spermatozoa (Supplementary Figures S2A,C). On the other hand, LED illumination caused less photobleaching in human spermatozoa (Supplementary Figure S2B), but certain level of photobleaching was still observed in 40% of mouse spermatozoa (Supplementary Figure S2D). This result suggests that mouse sperm are more susceptible to oxidative stress than human spermatozoa.

In contrast, SNARF-5F incorporated into spermatozoa is much less toxic to the cells than BCECF (Figure 3). The fluorescence intensities excited by LED and 488 nm laser (epi-fluorescence mode) did not cause photo-bleaching of the dye during our experimental periods (5 ips for 250 s) (Figures 3A,B). However, we observed a slight photobleaching of SNARF-5F when excited by the 488 nm laser in the TIRF configuration (Figure 3C). This photobleaching was negligible when we reduced the frequency of image acquisition from 5 to 2.5 ips (data not shown).

In our previous study of pH_i imaging (epi-fluorescence mode) of sea urchin spermatozoa using BCECF, fluorescence intensities of the flagellum were much lower than those of the heads and their fluorescence signals were noisy with a poor S/N ratio (González-Cota et al., 2015). We thought that the use of TIRF would improve this aspect, avoiding the saturation of fluorescence signal in the head. However, the difference of SNARF-5F images between the two configurations (epi-fluorescence and TIRF) was relatively small in both human (Figure 4A) and mouse spermatozoa (Figure 4B). Particularly, in mouse spermatozoa, the TIRF fluorescence signals in the head and the flagellum (primarily mid piece) are very similar to the epi-fluorescence images. This result is probably due to the thin hook-like shape of the mouse sperm head. As a consequence, an important difference between the two systems (epi-fluorescence and TIRF) is not significant for mouse spermatozoa.

Using the epi-fluorescence configuration with the LED as a light source, we acquired fluorescence images upon pH_i manipulations. During image acquisition, we added HTF or TYH medium as control in human and mouse spermatozoa, respectively. As additional control, 10 mM NH₄Cl and 5 mM HCl were added to increase and reduce the pH_i, respectively. The upper panels of Figures 5A,B show human sperm fluorescence signals from the two channels in the head and the flagellum, respectively. Changes of fluorescence intensities were observed during the additions even in control conditions (indicated with arrows). Also, fluorescence signals from some cells are noisy probably due to the continuous movement associated to the flagellar beat. Once the dual emission signals were converted into the ratio and pH_i values (Figures 5A,B, lower panels), the problems of addition artifacts and movement were eliminated in the both regions, demonstrating the advantage of the dual-emission ratiometric imaging. Additionally, the effects of NH₄Cl and HCl can be clearly observed as an increase and a decrease of the ratio and the pH_i, respectively. Figure 6 basically demonstrates the same results as Figure 5 but using mouse spermatozoa. In this experiment, the fluorescence signal of the flagellum arises mainly from the mid piece since mouse spermatozoa flagellum is much longer than that of human spermatozoa and therefore it is difficult to capture the image of the entire flagellum of mouse spermatozoa. In these experiments, the average pH_i of non-capacitated human and mouse spermatozoa was 6.72 ± 0.19 (SEM) and 6.63 ± 0.23 (SEM), respectively; n = 3.

To obtain new insights of mammalian spermatozoa pH_i regulation, we determined the effect of HCO3- (10 and 25 mM) on pH_i in non-capacitated human and mouse spermatozoa using our dual-emission imaging system (Figure 7). In these experiments, we confirmed that HCO3- increases the pH_i of human (Figure 7A) and mouse (Figure 7B) spermatozoa, in a concentration-dependent manner. We did not observe statistical differences in the pH_i increase induced by HCO3- between human and mouse spermatozoa (Figure 7C). However, we found a significant difference in the kinetics of the pH_i increase between the two species. Namely, HCO3- rapidly increases the pH_i of mouse spermatozoa, and the time to reach 50% of the maximum pH_i increase (t₅₀) was around 10 s. In contrast, HCO3- increases human sperm pH_i gradually with a longer t₅₀ (40 s) in our experimental conditions (Figure 7D). Moreover, the pH_i increase in human spermatozoa was slightly, but significantly, slower in the flagellum compared to the head with 10 and 25 mM HCO3- additions.

In the literature, there is some controversy about the effect of progesterone on human sperm pH_i. A decrease (Garcia and Meizel, 1996; Cross and Razy-Faulkner, 1997), no change (Fraire-Zamora and González-Martínez, 2004) or a slow increase (Hamamah et al., 1996) in pH_i have been reported by different groups in response to this hormone. Therefore, we determined the effect of different progesterone concentrations on pH_i in human spermatozoa. Figure 8 shows that progesterone at 500 nM (I), 1 μM (II), and 10 μM (III) did not change pH_i neither in the head (Figure 8A) nor in the flagellum (Figure 8B) of these cells. As a control we tested 1 μM monensin, a Na⁺ ionophore that exchanges Na⁺/H⁺ (Babcock, 1983). As anticipated, this ionophore alkalized pH_i in these cells in both head (Figure 8A, IV) and flagellum (Figure 8B, IV).

In this study we established a dual-emission ratiometric imaging system using SNARF-5F AM, which has negligible photo-toxicity compared to BCECF (Figure 3 and Supplementary Figure S2). Our system allows determining mammalian spermatozoa (head and flagellum) pH_i with minimum artifacts associated to cell movements and focus alteration upon addition or exchange of bath solutions (Figures 5–7). Commonly, the ratio of dual fluorescence signals utilizes the dye isosbestic point (595 nm in our condition). However, the first peak fluorescence intensity (575 nm) of SNARF-5F excited at 532 nm is much smaller than the second peak (640 nm). Therefore, we divided the fluorescence at 610 nm, 15 nm longer than the isosbestic point, and employed a wide band-pass filter (550–600 nm) for Channel 1. In this configuration, fluorescence signals of the two channels are comparable (Figure 2), which is a critical point to obtain the ratio values with a good S/N ratio. This type of optical filter configuration (division of fluorescence signals not at the isosbestic point) could be applied to other dual-emission indicators such as GEM-GECO (Zhao et al., 2011) and Asante Calcium Red (Hyrc et al., 2013) because their dual-emission signals are quite asymmetric.

We observed a slight difference between the fluorescence spectra of SNARF-5F in vitro and inside human spermatozoa. Therefore, in order to convert the fluorescence emission values into the pH_i, we performed an in vivo pH_i calibration with human and mouse spermatozoa using a high K⁺ solution combined with nigericin in order to equal pH_i to the pH_e. This protocol is based on the assumption that the cytoplasmic K⁺ concentration is 120 mM in human and mouse spermatozoa, as determined in bovine spermatozoa (Babcock, 1983). Therefore, depending on the real cytoplasmic K⁺ concentration in human and mouse spermatozoa, the absolute pH_i values could be different. In our conditions, we determined that the pH_i value in non-capacitated human and mouse spermatozoa is 6.72 ± 0.19 and 6.63 ± 0.23, respectively. These values were measured in the head, but no significant differences were observed in the flagellum (see below). There are several reports of pH_i determinations (most of them in cell population experiments and a few using single cell determination) of non-capacitated spermatozoa from distinct mammalian species: 6.24 (Parrish et al., 1989a) and 6.7 (Vredenburgh-Wilberg and Parrish, 1995) in bovine spermatozoa, 6.55 (Balderas et al., 2013), 6.54 (Zeng et al., 1996), and 6.8 (Carlson et al., 2007) in mouse sperm, and 6.7 (Hamamah et al., 1996; Fraire-Zamora and González-Martínez, 2004) and 6.94 (Cross and Razy-Faulkner, 1997) in human spermatozoa. Our findings that pH_i values of non-capacitated mammalian spermatozoa are >6.5 are consistent with the

report that detergent-demembranated bovine spermatozoa do not exhibit motility at pH 6.5, although they are highly motile at pH 7.0 (Ho et al., 2002).

We did not observe significant differences between the head and the flagellum in the basal pH_i in non-capacitated spermatozoa, although the head pH_i tends to be slightly higher than the flagellar pH_i both in human (6.72 ± 0.19 and 6.69 ± 0.24, respectively) and mouse spermatozoa (6.63 ± 0.23 and 6.60 ± 0.26, respectively). Our results are similar to those reported in bovine spermatozoa (Vredenburgh-Wilberg and Parrish, 1995). In general, the epifluorescent signal from an indicator incorporated into the sperm head is generally much higher than in the flagellum, independently of the species. Therefore, we examined if TIRF microscopy would reduce the fluorescence difference between the head and the flagellum. However, we did not observe significant differences nor advantages of TIRF microscopy compared to epifluorescence microscopy (Figure 4) either in mouse or human spermatozoa for measuring pH_i. With these data, we can conclude that epifluorescence microscopy with SNARF-5F AM allows performing reliable single spermatozoa pH_i imaging with a good S/N ratio in both spermatozoa head and flagellum (mid piece of flagellum in the case of mouse).

HCO3- is an essential ion for mammalian sperm to acquire the ability to fertilize the oocyte (Lee and Storey, 1986). In fact, the HCO3- concentrations in rabbit uterine and tubal fluids are approximately twice as high as in the blood plasma, which results in pH values of 7.4 and 8.1–8.3, respectively (Vishwakarma, 1962). In rhesus monkeys, the pH and HCO3- concentration in the oviduct lumen change during the menstrual cycle. Namely, these values are similar to those of the blood plasma during the follicular phase, but they suddenly increase concomitantly with ovulation (Maas et al., 1977). This observation supports the importance of HCO3- for fertilization in mammals. The principal role of cytoplasmic HCO3- in mammalian spermatozoa is considered to be the activation of the soluble adenylyl cyclase, which increases cAMP (Okamura et al., 1985; Buck et al., 1999; Chen et al., 2000), leading to protein kinase A (PKA) stimulation. The enhanced PKA activity increases flagellar beat frequency (Wennemuth, 2003) and elevates CatSper activity (Carlson et al., 2003; Orta et al., 2018), among many other things.

In this work, we observed that HCO3- elevates the pH_i in both human and mouse spermatozoa (Figure 7D). In contrast, Carlson et al. (2007) reported that HCO3- did not induce a pH_i increase in mouse spermatozoa. Interestingly, we found a difference in the kinetics of HCO3--induced pH_i increase between the two species (Figure 7D), namely a faster increase in mouse compared to human spermatozoa, but of similar magnitude (Figure 7C). So far, several mechanisms have been reported for HCO3- influx, such as the NBC (Demarco et al., 2003), Cl^–/HCO3- exchangers (Chavez et al., 2012), and the CFTR channel (Hernández-González et al., 2007; Xu et al., 2007). However, the physiological relevance of each transporter is unclear as well as differences between the two species. In addition to HCO3- transporters, CO₂ diffusion with subsequent hydration by intracellular CA contributes to an increase in cytoplasmic HCO3- concentration (Carlson et al., 2007; Wandernoth et al., 2010; José et al., 2015). Curiously, a general CA inhibitor, ethoxyzolamide, potently affects human but not mouse sperm motility (José et al., 2015), suggesting a difference in the involvement of CAs in the motility of the two species. Another explanation is that human spermatozoa may have higher pH buffering capacity than mouse spermatozoa. This might be correlated to the time required for capacitation (>6 h in human compared to 1–2 h in mouse spermatozoa). Indeed, the pH_i of mammalian spermatozoa studied so far increases around 0.14–0.4 units during capacitation (Parrish et al., 1989b; Vredenburgh-Wilberg and Parrish, 1995; Hamamah et al., 1996; Zeng et al., 1996; Cross and Razy-Faulkner, 1997; Fraire-Zamora and González-Martínez, 2004; Balderas et al., 2013). A significant part of this pH_i change can be attributed to the HCO3- influx into the cell. Therefore, further studies are required for a better understanding of the mechanism of HCO3--induced pH_i increase during capacitation. The pH_i imaging system established in this study should contribute to this issue.

Progesterone increases [Ca²⁺]_i in human spermatozoa at concentrations as low as 300 nM, through CatSper activation (Tesarik et al., 1992; Harper et al., 2003; Achikanu et al., 2018). Recently it was described that the progesterone receptor in these cells is a α/β hydrolase domain-containing protein (ABHD2), which depletes the endocannabinoid 2-arachinoylglycerol (2AG) from membrane and removes CatSper inactivation (Miller et al., 2016; Mannowetz et al., 2017). In contrast, there is inconsistency regarding how progesterone affects pH_i. Some groups suggest that this hormone acidifies, others that it alkalizes or does not induce pH_i changes (Garcia and Meizel, 1996; Hamamah et al., 1996; Cross and Razy-Faulkner, 1997; Fraire-Zamora and González-Martínez, 2004). In the present work, progesterone did change pH_i in human spermatozoa even at concentrations as high as 10 μM (Figure 8). Our result supports that progesterone activates CatSper in a pH-independent manner, possibly exclusively via ABHD2-2AG.

Progesterone-induced Ca²⁺ influx through CatSper may affect the activity of transporters and enzymes that can affect pH_i such as PMCA (Wennemuth et al., 2003; Okunade et al., 2004) and NOX5 (Baker and Aitken, 2004; Musset et al., 2012). Both PMCA and NOX5 may acidify pH_i when they are activated, namely when the [Ca²⁺]_i is high. However, the pH_i acidification together with the membrane potential depolarization caused by Ca²⁺ influx through CatSper and electron efflux through NOX5 could activate Hv1 channel (Lishko et al., 2010; Berger et al., 2017) and may rapidly neutralize the acidification (alkalize the pH_i). Depending on the experimental conditions, one activity (acidifying or alkalizing) may exceed the other when progesterone stimulates human CatSper. This may account for part of the discrepancies regarding human sperm pH_i responses to progesterone. Further studies are required to confirm this hypothesis.