A metastasis‐associated pannexin‐1 mutant (Panx11‐89) forms a minimalist ATP release channel

Junjie Wang; Noah J Levi; Maykelis Diaz‐Solares; Carsten Mim; Gerhard Dahl; Rene Barro‐Soria

doi:10.1111/febs.70060

. 2025 Mar 14;292(13):3378–3396. doi: 10.1111/febs.70060

A metastasis‐associated pannexin‐1 mutant (Panx1^1‐89) forms a minimalist ATP release channel

Junjie Wang ¹, Noah J Levi ², Maykelis Diaz‐Solares ², Carsten Mim ³, Gerhard Dahl ^1,^✉, Rene Barro‐Soria ^2,^✉

PMCID: PMC12220848 PMID: 40087867

Abstract

A truncated form of the ATP release channel pannexin 1 (Panx1), Panx1^1–89, is enriched in metastatic breast cancer cells and has been proposed to mediate metastatic cell survival by increasing ATP release through mechanosensitive Panx1 channels. However, whether Panx1^1‐89 on its own [without the presence of wild‐type Panx1 (wtPanx1)] mediates ATP release has not been tested. Here, we show that Panx1^1‐89 by itself can form a constitutively active membrane channel, capable of releasing ATP even in the absence of wtPanx1. Our biophysical characterization reveals that most basic structure–function features of the channel pore are conserved in the truncated Panx1^1‐89 polypeptide. Thus, augmenting extracellular potassium ion concentrations enhances Panx1^1‐89‐mediated conductance. Moreover, despite the severe truncation, Panx1^1‐89 retains sensitivity to most wtPanx1 channel inhibitors. Therefore, Panx1 blockers may be of therapeutic value to combat metastatic cell survival. Our study both provides a mechanism for ATP release from cancer cells and suggests that Panx1^1‐89 might aid in the structure–function analysis of Panx1 channels.

Keywords: ATP release, cancer, metastasis, Pannexin1 channel, Pannexin1‐89 polypeptide

A truncation mutant of Panx1 (Panx1^1‐89), enriched in metastatic breast cancer cells, is capable of forming a large ATP‐permeable membrane channel despite containing only 20% of the amino‐terminal amino acids of the wild‐type channel protein (wtPanx1). However, cells expressing Panx1^1‐89 alone die, whereas co‐expression of wtPanx1 is protective. In metastatic cells, mutant and wtPanx1 are typically co‐expressed and have been shown to facilitate metastasis via purinergic signaling.

graphic file with name FEBS-292-3378-g005.jpg

Abbreviations

ANOVA: analysis of variance
BB FCF: brilliant blue for coloring food
BzATP: 2′(3′)‐O‐(4‐benzoylbenzoyl)adenosine‐5′‐triphosphate tri(triethylammonium) salt
CBX: carbenoxolone
CFTR: cystic fibrosis transmembrane conductance regulator
DAPI: 4′,6‐diamidino‐2‐phenylindole
His AB: anti‐6× His tag® antibody
MPB: 3‐(N‐maleimidopropionyl)biocytin
MTSET: 2‐(trimethylammonium)ethyl methanethiosulfonate
OR: oocyte Ringer solution
Panx1: panexin 1
Prob: probenecid
TCEP: tris(2‐carboxyethyl)phosphine
WGA: wheat germ agglutinin

Introduction

Extracellular ATP and its catabolites generated by ecto‐ATPases are critical players in cancer biology [1, 2, 3, 4, 5]. Thus, as an ATP release channel, Pannexin1 (Panx1) has gained considerable attention as a mediator of metastasis [6, 7, 8, 9, 10, 11, 12, 13]. Based on the direct correlation of the expression of a truncated form of the Panx1 channel (Panx1^1‐89) with the metastatic potential of breast cancer cells, a mechanism for metastatic cell survival in the microvasculature has been proposed [14]. A key aspect of this mechanism is that the release of ATP is augmented by the expression of Panx1^1‐89. Indeed, experimental evidence indicates that released ATP facilitates the transit of metastatic cells through the vessel wall and protects them from the mechanical stress endured during transit [14]. However, given that the Panx1^1‐89 polypeptide truncates the protein from 426 to the amino‐terminal 89 amino acids, it is still unclear how Panx1^1‐89 (alone or in combination with wild‐type Panx1) mediates ATP release.

While it is generally accepted that wtPanx1 forms an ATP release channel, presently available structural data do not support this function [15, 16]. In all published structures of wt Panx1 or its caspase cleavage product, a constriction with a radius of 4–4.5 Å is prominent at the extracellular entry to the channel pore [17, 18, 19, 20, 21, 22, 23]. Since ATP has an Einstein‐Stokes radius of 7 Å, it should therefore be excluded from passage through the Pannexin 1 channel. Instead, the 4–4.5 Å radius is consistent with the chloride selective conformation of the Panx1 channel. This apparent contradiction can be explained by a two‐channel hypothesis [24]: depending on the stimulus modality, the Panx1 channel adopts two distinct conformations [25]. Thus, although the voltage‐activated or caspase‐activated channel is highly selective for chloride ions, certain physiological or pathological stimuli induce the ATP‐permeable, large‐pore channel conformation [24, 25, 26, 27, 28, 29]. However, the structure of the latter conformation is unknown.

In this study, we show that Panx1^1‐89 by itself can form a constitutively active membrane channel, capable of releasing ATP even in the absence of wtPanx1. Since Panx1^1‐89 forms a constitutively open channel with a pore dimension that allows ATP conduction, structural studies on Panx1^1‐89 might show the large pore conformation of the Panx1 channel, which to date has remained elusive in all 8 publications of cryo‐EM structures of wtPanx1. Furthermore, our biophysical characterization shows that most basic structure–function features of the channel pore are conserved in Panx1^1‐89. Despite the severe truncation, Panx1^1‐89 retains most pharmacological properties of the wtPanx1 channel and can thus be targeted. Consequently, Panx1 blockers have the potential to be of therapeutic value to combat metastatic cell survival.

Results

Exclusive expression of Panx1^1‐89 results in cell death

Panx1^1‐89 is co‐expressed in metastatic cells with wtPanx1. It has been suggested that the two proteins interact, probably by assembling into hetero‐oligomers, which mediate constitutive ATP release [14]. To assess the biophysical properties of such hetero‐oligomers, Xenopus oocytes were co‐injected with mRNA for Panx1^1‐89 and wtPanx1 proteins at a 1:1 mRNA ratio. As controls, oocytes were injected with either Panx1^1‐89 or wtPanx1 alone. Oocytes expressing wtPanx1 alone or co‐expressing Panx1^1‐89 and wt Panx1 survived the 24–48 h (or even 72 h in some cases) incubation period after mRNA injection required to observe robust Panx1 channel currents (Fig. 1a). In contrast, oocytes expressing Panx1^1‐89 alone typically did not survive longer than 12 h after mRNA injection (Fig. 1b). These oocytes could not be voltage‐clamped, and their appearance was distinctly abnormal. Their pigmentation at the animal pole was spotted, and the cell surface exhibited indentations. There was, however, a time window of 3–8 h when ^1‐89Panx1‐expressing oocytes appeared normal and could easily be voltage‐clamped.

Fig. 1 — Expression of Panx1^1‐89 in *Xenopus* oocytes induces constitutive membrane currents. (A, B) Photographs of *Xenopus laevis* oocytes (A) (uninjected, top), co‐injected with wtPanx1 and Panx1^1‐89 mRNA (75 ng each in a volume of 60 nL, 24 h prior, bottom), and (B) Panx1^1‐89 mRNA 24 h prior with (bottom) and without (top) 100 μM CBX. Arrowhead in (A) shows an injection scar. Scale bar represents 0.2 mm. (C, H) Representative voltage ramp current traces from oocytes expressing wtPanx/Panx1^1‐89 in OR (5:3 ratio, black trace in C) or gluconate (orange trace in C) solutions, wtPanx1 alone (blue traces in C and H), and Panx1^1‐89 (red trace in H). Substitution of Cl⁻ by Gluconate⁻ shifted the reversal potential to less negative potentials (inset in C). While the currents carried by wtPanx1 exhibited the typical outward rectification, the currents mediated by Panx1^1‐89 were prominent throughout the voltage range from −100 to +100 mV. The inset in (H) shows the area where the currents reversed from inward to outward currents. The ^1‐89Panx1‐mediated currents reversed at a more positive potential than the wtPanx1 currents. (D, I) Quantitative analysis of the reversal potential for (D) co‐injected wtPanx1/Panx1^1‐89 and (I) wtPanx1 (blue) and Panx1^1‐89 (red). Current records were obtained from oocytes 3–5 h after injection of Panx1^1‐89 mRNA and 24–48 h after injection of wtPanx1 mRNA. For reference, data from Wang and Dahl 2018 for oocytes expressing wtPanx1 alone are shown in (E). (F) Representative current traces from uninjected (control), wtPanx1, Panx1^1‐89, and CFTR‐expressing oocytes for the indicated voltage protocol (inset). Currents from Panx1^1‐89 (or wtPanx1 and CFTR) expressing oocytes were recorded 4 h (or 24 h) after injection of mRNA. Data in (D, G, I) are mean ± SEM. Currents from “n” oocytes over “N” batches (n/N); in (D) n/N = 5/2 (wtPanx1/Panx1^1‐89); in (I) n/N = 5/2 (wtPanx1) and n/N = 7/3 (Panx1^1‐89); and in (G) n/N = 12/3 (control), n/N = 11/3 (wt Panx1), n/N = 13/3 (Panx1^1‐89), and n/N = 6/2 (CFTR). Statistical significance was determined using Student's t‐test.

Oocytes co‐injected with wtPanx1 and Panx1^1‐89 mRNA 24 h prior to voltage clamp analysis exhibited currents with reversal potentials and outward rectification that were similar to those observed in cells expressing wtPanx1 alone. Even with a 5:3 ratio (Panx1^1‐89/wtPanx1) to promote an effect with an excess of Panx1^1‐89, the currents matched those of cells expressing wtPanx1 (Fig. 1c). Next, we performed an ion substitution experiment to test whether channels from oocytes co‐expressing Panx1^1‐89 and wtPanx1 were as selective for chloride ions as those expressing the wtPanx1 channel are. Figure 1c,d shows that replacing Cl⁻ with gluconate⁻ (Glu) shifted the reversal potential from ~−28 mV to ~−5 mV. This assay revealed a subtle difference between wtPanx1 channels and those from oocytes co‐expressing Panx1^1‐89 and wtPanx1. As shown previously [24, 26], the same chloride ion substitution resulted in currents reversing at positive potentials in cells expressing only wtPanx1, rather than the negative potentials found here for the co‐expressing cells (Fig. 1d,e). Thus, the negative reversal potential in cells co‐expressing wt Panx1 and Panx1^1‐89 in gluconate indicates different permeability of the heteromeric and/or mix of homomeric channels as compared to the homomeric wtPanx1 channels with a positive reversal potential under identical conditions.

The cell death observed with longer incubation periods of cells expressing Panx1^1‐89 alone indicated that the expression of Panx1^1‐89 alone was toxic to the cells. This toxicity could have occurred at different cellular sites, such as the synthetic pathway for proteins or at the cell membrane. If the latter, the Panx1^1‐89 might have created a nonspecific membrane leak or could have associated with endogenous membrane proteins (channel or transporter) to confer cell toxicity. Alternatively, the Panx1^1‐89 polypeptide might have assembled in the membrane to form a channel with some resemblance to the wtPanx1 channel. The Panx1^1‐89 polypeptide contains only one of the four transmembrane helices of wt Panx1. Although assembly of ion channels by proteins/polypeptides with only one transmembrane segment is unusual, there is precedence for it. The pioneering work by the Montal laboratory [30] has shown that single‐pass membrane polypeptides formed by synthetic peptides can induce channel activity in lipid bilayers. Subsequently, it has been found that such a phenomenon is also found in a biological setting as several viroporins are oligomerized single‐pass membrane polypeptides of a similar size range to Panx1^1‐89. Various viroporin polypeptides oligomerize to tetramers, pentamers, or hexamers to form patent channels [31, 32, 33, 34]. It should be noted, however, that not all viroporins may function as membrane channels. For example, the SARS‐Covid protein Orf3a is involved in membrane trafficking rather than forming a channel, as once thought [35].

Studies using electrophysiology in combination with site‐directed mutagenesis and cysteine or alanine scans have shown that the first 89 amino acid residues within the wtPanx1 channel contain key elements of the channel, including pore‐lining amino acids and binding sites for several modulators of channel activity (Fig. S1). For example, structural and functional studies showed that amino acids W74 and R75 are involved in channel inhibition by ATP and carbenoxolone (CBX), as well as for channel activation by extracellular potassium ions ([K⁺]_o) [17, 36, 37, 38, 39]. Many of these findings were subsequently verified by determination of the Panx1 structure by cryo‐EM [17, 18, 19, 20, 21, 22, 23]. Considering that some viroporins with a single transmembrane segment form channels and that key features of the wtPanx1 channel are present in the Panx1^1‐89 sequence, it was imperative to know the biophysical properties of cells expressing Panx1^1‐89 heterologously.

Expression of Panx1^1‐89 induces constitutively active membrane channels

We used a two‐electrode voltage clamp to measure current from ^1‐89Panx1‐expressing oocytes (Fig. 1f,g). Compared to uninjected oocytes or oocytes expressing the chloride channel cystic fibrosis transmembrane conductance regulator (CFTR), ^1‐89Panx1‐expressing cells produced a robust current as early as 3 h after mRNA injection. This is in contrast to wtPanx1, which required >12 h incubation for currents to become detectable. Consistent with their outward rectifying properties, wtPanx1‐expressing oocytes exhibited moderately increased currents when clamped at −60 mV compared with uninjected oocytes. Application of a voltage ramp to oocytes expressing wtPanx1 (blue trace in Fig. 1h) showed the typical pronounced outward rectification of wtPanx1. In contrast, as shown by the large inward current at negative potentials, the channels induced by Panx1^1‐89 rectified less (red trace in Fig. 1h). In this respect, the Panx1^1‐89 channels are similar to the wtPanx1 channels activated by caspase cleavage at position 378 [40]. However, while the currents mediated by wtPanx1 cleaved by caspase or truncated at position 378 reverse at the same potential as voltage‐activated wtPanx1 does [16, 25, 41], the Panx1^1‐89–induced channels exhibited a reversal potential shifted to a more positive potential (inset Fig. 1h). Figure 1i shows that this shift was significant. Ion substitution experiments have shown that the voltage‐activated or caspase‐cleaved wtPanx1 channels are selective for chloride ions. Consequently, the currents reverse at the chloride equilibrium potential [25, 26, 27, 40]. Therefore, the shift of the reversal potential in Panx1^1‐89‐expressing oocytes indicates a different permeability for Panx1^1‐89‐induced channels than the ⁻Cl‐selective property of wt and caspase‐cleaved Panx1 channels.

The Panx1^1‐89‐mediated current is sensitive to wtPanx1 channel blockers

Panx1^1‐89 contains amino acids within the binding sites for various Panx1 blockers [17, 36, 37, 38, 39]. To test whether these elements were sufficient to affect the currents mediated by Panx1^1‐89 expression, we applied several wtPanx1 blockers (Fig. 2). Application of 1 mM probenecid (Prob) greatly attenuated the currents, although it did not completely inhibit the currents as it does in wtPanx1 channels [42] (Fig. 2a,b). Other inhibitors of wtPanx1 currents, when applied in excess of their IC₅₀, including 100 μM CBX, the food dye BB FCF (10 μM), and BzATP (100 μM), also attenuated the currents in oocytes expressing Panx1^1‐89 (Fig. 2b–d). We chose these inhibitors for their common effect on Panx1 while exhibiting no overlapping effects on other targets. As probenecid, these Panx1 inhibitors affected Panx1^1‐89 to a lesser extent than wtPanx1. Since Panx1^1‐89 only contains part of the binding site(s) for the various inhibitors, such attenuation was to be expected. The inhibition of the currents was observed over a wide voltage range, as shown in Fig. 2e for the most specific of the Panx1 inhibitors, BB FCF. Together, these data indicate that Panx1^1‐89 retains a similar pharmacological profile as the wtPanx1 channel.

Fig. 2 — Effect of Panx1 channel inhibitors on membrane currents from oocytes injected with Panx1^1‐89 mRNA. (A, C, D) Current traces from oocytes expressing Panx1^1‐89 before and after the application of (A) 1 mM probenecid, (c) 100 μM of the food dye BB FCF, and (D) 100 μM of BzATP followed by 100 μM carbenoxolone (CBX) as induced by the voltage protocol (0.1 Hz from −60 to +60 mV). Recordings were done 3–5 h after Panx1^1‐89 mRNA injection. (B) Quantitative analysis of % inhibition of ^1‐89Panx1‐induced currents by carbenoxolone (CBX, 100 μM), BzATP (10 μM), probenecid (1 mM) and brilliant blue for coloring food (BB FCF, 10 μM). (E) Voltage ramp current traces from oocytes expressing Panx1^1‐89 in the absence (black) or presence of the Panx1 inhibitor BB FCF (blue). BB FCF at 10 μM attenuated the currents throughout the voltage range from −100 to +100 mV. Data represent mean ± SEM. Currents from “n” oocytes over “N” batches (n/N); In (B) n is indicated above the bars: n/N = 7/3 (CBX), n/N = 4/2 (BzATP), n/N = 3/2 (Prob), n/N = 4/2 (BB FCF) and in (E) n/N = 3/1.

Although 100 μM CBX attenuated Panx1^1‐89 current, it did not prevent death from oocytes expressing Panx1^1‐89 (Fig. 1b). For this reason, we performed all subsequent experiments within 4–8 h after RNA injection where Panx1^1‐89 activity was prevalent in the absence of leaks. Note that a similar finding was also reported for the caspase‐cleaved wtPanx1 [43], in which a constitutively active current was attenuated by CBX without preventing cell death. Intriguingly, some constitutively active Panx1 mutants can be protected from cell death by Panx1 blockers such as CBX [44], while others are not, despite blockage of channel activity.

Next, we set out to examine the subcellular localization of Panx1^1‐89 by introducing a His6‐tag following the COOH‐terminal Q89 (Panx1^1‐89‐His6). To ensure that the addition of the His6 tag did not interfere with the channel formation by Panx1^1‐89, the Panx1^1‐89‐His6 version was expressed in oocytes, and currents were recorded using two‐electrode voltage clamp electrophysiology. As shown in Fig. 3a, Panx1^1‐89‐His6 yielded similar CBX‐sensitive currents as untagged Panx1^1‐89. Application of voltage steps from −60 mV to +60 mV induced currents averaging 2.09 ± 0.2 μA (n = 8), which were inhibited by 100 μM CBX by 64.0 ± 3.2% (n = 4) (Fig. 3a). The channels formed by Panx1^1‐89‐His6 were sensitive to increased extracellular [K⁺] like those formed by wtPanx1 and untagged polypeptide (Fig. 3b). Perfusion of oocytes voltage clamped at −60 mV responded to a potassium solution (K⁺ replacing Na⁺) with an inward current averaging 2.57 ± 0.44 μA (n = 4), and the mean conductance at the peak of the K⁺ response was 75.5 ± 23.6 μS. Immunohistochemistry with an antiHis6 antibody revealed co‐localization of Panx1^1‐89‐His6 with the membrane stain wheat germ agglutinin (Fig. 3c). Interestingly, Panx1^1‐89‐His6 was mainly restricted to the oocyte's vegetal pole. This asymmetric distribution mimics that of connexins expressed in oocytes [45]. As a negative control, oocytes expressing untagged wtPanx1 for 24 h were subjected to the same staining protocol (Fig. 3d).

Fig. 3 — Expression of 6His‐tagged Panx1^1‐89 in oocytes. (A, B) Current traces from oocytes expressing 6His‐tagged Panx1^1‐89 before and after application of (A) 100 μM of the Panx1 channel blocker carbenoxolone (CBX) or (B) high K⁺ (85 mM) for the indicated voltage protocols (right of each trace). Extracellular application of 85 mM K⁺ resulted in an inward current and an increase in membrane conductance. (C, D) Confocal images of 40 μm thick cryosections of oocytes expressing (C) 6His‐tagged Panx1^1‐89 (8 h prior fixation) and (D) untagged wtPanx1. The sections were immunostained for His6 (green), DAPI (blue), and the membrane marker wheat germ agglutinin (WGA, red). The merged images show co‐localization of the His‐tagged Panx1^1‐89 with the membrane marker. Scale bar = 0.2 mm. n = 3 oocytes from 2 different batches.

To test whether Panx1^1‐89 also forms constitutively active channels in the plasma membrane of mammalian cells, we expressed Panx1^1‐89‐His6 in HEK293T cells. Immunostaining (4 h after transfection) with an antiHis6 antibody also revealed co‐localization of Panx1^1‐89‐His6 with the membrane stain wheat germ agglutinin (Fig. 4a). However, 24 h after transfection, Panx1^1‐89‐His6/YFP tagged cells were not further observed (Fig. 4b,c). Only small fluorescent specs (likely remnants of dead cells expressing Panx1^1‐89‐His6/YFP, green arrow in Fig. 4c) were detected at this time, and mostly untagged cells (probably untransfected cells) occupied the culture dish. Whole‐cell patch clamp analysis showed that, compared with the typical outward rectification of wtPanx1 current (Fig. 4d, blue, see also Fig. 1h, blue), expression of Panx1^1‐89‐His6 in HEK293T cells yielded large inward current at negative potentials (did not rectify, Fig. 4d, red), similar to Panx1^1‐89 currents shown in oocytes (Fig. 1). Moreover, both wtPanx1 and Panx1^1‐89‐His6 currents were blocked by 100 μM CBX (Fig. 4d,e), similar to the inhibition observed in oocytes (Fig. 2).

Fig. 4 — Expression of 6His‐tagged Panx1^1‐89 in HEK293T cells. (A) Immunohistochemistry of HEK293T cells expressing 6His‐tagged Panx1^1‐89 using confocal microscopy. 4 h after transfection, a subset of cells expressed the His tag (red). Staining of the plasma membrane with fluorescently labeled WGA (yellow) shows co‐localization of 6His‐tagged Panx1^1‐89 with the membrane marker. The most intense label (for example cell center top) was observed in compromised cells. (B, C) Immunohistochemistry of HEK293T cells 8 (B) or 24 (C) hours after co‐transfection of 6His‐tagged Panx1^1‐89 with green fluorescent protein (YFP). High intensity of the YFP signal was associated with irregular cell shape, suggesting cell damage. 24 h after transfection, the YFP fluorescence appeared only as puncta, likely representing cell debris (green arrow). Staining of the cell membrane with fluorescent WGA (white) revealed that most cells had a healthy appearance. DAPI staining is shown in blue. Scale bar = 10 μM. (C) Representative whole‐cell current traces from wtPanx1 (blue) and Panx1^1‐89‐His6 (red) expressing HEK293T cells in the absence (control) or presence of 100 μM CBX. Cells were held at −80 mV for 1 s followed by a slow ramp protocol from −100 to +100 mV for 30 s before returning to the‐80 mV holding potential. Note that similar to Panx1^1‐89 currents shown in Fig. 1, Panx1^1‐89‐His6 currents did not rectify, indicating constitutive activity of the channels over a wide voltage range. Currents from wtPanx1 and Panx1^1‐89‐His6 expressing HEK293T cells were recorded between 24 and 30 h and at 4 h after transfection, respectively. (D) Summary data of experiments shown in (C) for currents at +100 mV. Both wtPanx1 and Panx1^1‐89‐His6 expressed in HEK293T cells displayed similar CBX‐sensitive currents as those observed in oocytes (Fig. 1). Data represent mean ± SEM. Statistical significance was determined using one‐way ANOVA and Bonferroni's *post hoc* test to compare the CBX‐induced change in current amplitude. P < 0.05; Currents from “n” cells over “N” replicates (n/N); n/N = 5/3 (wt Panx1), n/N = 7/4 (Panx1^1‐89).

Panx1^1‐89 responds more strongly to increased extracellular potassium ion concentration than wtPanx1

The Panx1^1‐89 polypeptide not only contains binding sites for Panx1 blockers but also amino acids involved in the activation of the Panx1 channel at negative potentials by extracellular K⁺ [38]. To test whether K⁺‐activation was still retained by the truncated protein, oocytes clamped at −60 mV were perfused with a solution containing 85 mM KCl. Because Ca²⁺ attenuates the K⁺‐activation of wtPanx1 [38], no calcium was added. Figure 5 shows the responses of uninjected oocytes and wtPanx1 or ^1‐89Panx1‐expressing oocytes. As reported previously, uninjected oocytes respond to extracellular high [K⁺] with a small inward current, which is increased in wtPanx1‐expressing cells (Fig. 5a). In contrast, the response to increased [K⁺]_o by Panx1^1‐89 expressing cells was on average >5 times larger (Fig. 5a,b). These values are comparable to those observed for channels formed by the tagged Panx1^1‐89‐His6 polypeptide for the same parameters (Fig. 3b). For both wtPanx1 and Panx1^1‐89, a linear relationship was observed between K⁺‐induced currents and voltage‐induced currents (Fig. 5c). The latter were a result of voltage steps from −60 to +60 mV and reflected the expression level of the channel proteins. The slope was about 5 times steeper for Panx1^1‐89 than for wtPanx1. The linear correlations between the voltage‐activated currents and the K⁺‐activated currents from oocytes expressing either the wtPanx1 or the Panx1^1‐89 channels suggest that the observed K⁺‐induced currents are generated by wtPanx1 or Panx1^1‐89 channels, respectively, rather than from endogenous channels.

Fig. 5 — Effect of increased extracellular potassium ion concentration (85 mM, with K⁺ replacing Na⁺) on membrane currents in oocytes. (A) Oocytes were voltage clamped at −60 mV and 12 mV depolarizing test pulses for 5 s at 0.1 Hz for conductance measurements. Traces for uninjected oocytes (top), oocytes expressing wtPanx1 (middle) or Panx1^1‐89 (bottom) are shown. K⁺ was applied as indicated by the horizontal bars. Measurements for wtPanx1 and Panx1^1‐89 were performed 24 h and 3–5 h after injection of mRNA, respectively. (B) Average calculated conductances for uninjected (gray), wtPanx1 (blue) and Panx1^1‐89 (red) expressing oocytes are shown. Data are presented as mean ± SEM. Currents from “n” oocytes over “N” batches (n/N); n/N = 10/3 (uninjected), n/N = 11/3 (wtPanx1), and n/N = 10/3 (Panx1^1‐89). (C) Relationship between K⁺‐induced currents and currents induced by voltage steps from −60 to +60 mV as a measure of expression levels. Only oocytes expressing wtPanx1 or Panx1^1‐89 beyond the threshold of 1 μA for voltage‐induced currents were analyzed. The lines are based on regression analysis with the formula y = a + b*x yielding slopes of 0.62 ± 0.09 (red) and 0.07 ± 0.01 (blue) and R‐squares of 0.92 and 0.86, respectively. Data are presented as mean ± SEM. n/N = 6/3 (wt Panx1), n/N = 5/3 (Panx1^1‐89). Statistical significance was determined using Student's t‐test.

Panx1^1‐89 expression yields a permeation pathway with large pore properties

The observation that the channels appearing in response to the injection of mRNA for Panx1^1‐89 had a different reversal potential than the channels formed by wtPanx1 (Fig. 1h,i) suggested that the two channels differed in their ion permeabilities. Since wtPanx1 can assume different conformations, one with chloride selectivity or one with ATP permeability [24, 25, 28], an ATP release function of Panx1^1‐89 in the absence of wtPanx1 was tested. Figure 6a shows that after an incubation period of 30 min, the ATP content of the supernatant of ^1‐89Panx1‐expressing cells was significantly higher than that of uninjected cells. This suggests that the Panx1^1‐89 channel was in the large pore conformation even in the absence of a stimulus. ATP release was also enhanced when ^1‐89Panx1‐expressing cells were incubated in high [K⁺] (Fig. 6b). Notably, the Panx1 blocker carbenoxolone inhibited the ^1‐89Panx1‐mediated ATP release (Fig. 6b).

Fig. 6 — ATP release by Panx1^1‐89. (A) ATP release by uninjected oocytes (gray) and oocytes expressing Panx1^1‐89 (7 h after injection of mRNA, red) incubated in Ringer solution (OR) in the absence of extracellular K⁺ stimulation. (B) ATP release by uninjected oocytes (gray) and oocytes expressing Panx1^1‐89 (7 h after injection of mRNA, red) incubated in 85 mM K⁺ solution with and without 100 μM CBX. All data were normalized to the ATP content of the supernatant of unstimulated control oocytes. Data from 3 oocytes (2 from the Panx1^1‐89 OR and 1 from the K⁺ + CBX data set) were excluded from the analysis due to visible cell damage. Data are presented as mean ± SEM, “n” oocytes over “N” batches (n/N); n/N = 12/3 (control, OR uninjected), n/N = 11/3 (Panx1^1‐89 in OR), n/N = 12/3 (control “uninjected” + K⁺), n/N = 13/3 (Panx1^1‐89 + K⁺), and n/N = 10/3 (Panx1^1‐89 + K⁺ + CBX). Statistical significance was determined using (A) Student's t‐test and (B) one‐way ANOVA and Bonferroni's *post hoc* test. P < 0.05.

Cells release ATP via exocytosis, via large pore channels, or in response to cell lysis. Since ^1‐89Panx1‐expressing cells have a limited life span, a contribution of lytic release of ATP cannot be ruled out. Therefore, as an independent method, we tested whether the ^1‐89Panx‐induced channel was able to pass other molecules in the size range of ATP, such as cAMP.

To test cAMP permeability, we used the CFTR channel, which is activated by high levels of intracellular cAMP. We reasoned that if the pore formed by Panx^1‐89 expression is large enough to pass cAMP, then it would provide a path for the influx of extracellular cAMP such that the levels of intracellular cAMP increase and, in turn, activate CFTR‐mediated Cl⁻ flux out of cells co‐expressing both Panx1^1‐89 and CFTR channels (Fig. 7a). Indeed, we found that the application of extracellular cAMP activated a large inward current and a corresponding increase in membrane conductance in cells co‐expressing CFTR and Panx1^1‐89 (Fig. 7b). The induced current and conductance reversed after the washout of extracellular cAMP, making a leak pathway for the entry of cAMP unlikely. As a control, oocytes expressing either CFTR alone or Panx^1‐89 alone did not respond to 1 mM extracellular cAMP, despite a strong Forskolin‐mediated inward current indicating robust membrane expression of CFTR (Fig. 7c,d). Figure 7e,f shows a quantitative analysis of the cAMP effects on currents and conductances. In summary, these data indicate that the large pore mediated by Panx1^1‐89 expression allows sufficient entry of extracellular cAMP into the cell to reversibly activate CFTR.

Fig. 7 — Uptake of extracellular cAMP. (A) The schematic diagram shows the voltage clamp arrangement and the deduced uptake of cAMP through ^1‐89Panx1‐mediated channels and subsequent activation of CFTR‐mediated chloride currents. Also shown is the activation of CFTR by Forskolin. Current traces are shown for oocytes co‐expressing Panx1^1‐89 with CFTR (B), expressing CFTR alone (C) and expressing Panx1^1‐89 alone (D). 1 mM cAMP was applied extracellularly as indicated. To test for the expression of CFTR, 20 μM Forskolin was applied as indicated in (C). Panel (E) shows how inward currents (orange arrow) and membrane conductance (red arrow) were determined. A quantitative analysis of cAMP‐induced inward membrane currents and calculated conductances is shown in (F). The oocytes were voltage clamped at −60 mV, and 12 mV depolarizing voltage steps were applied at a frequency of 0.1 Hz for conductance measurements. mRNAs for CFTR and for Panx1^1‐89 were injected 48 h and 5 h prior to the records, respectively. Data are presented as mean ± SEM. Currents from “n” oocytes over “N” batches (n/N); n/N = 5/2. Statistical significance was determined using the Mann–Whitney test.

Panx1^1‐89 may provide the pore lining of the induced channel

While the similarities in channel properties between wtPanx1 and the Panx1^1‐89– induced channels (i.e. inhibition of membrane currents by several Panx1 channel blockers, sensitivity to extracellular K⁺ and ATP permeability) suggest that Panx1^1‐89 by itself is sufficient to form the ion‐conducting pore of the channel, it cannot be excluded that Panx1^1‐89 could associate with an endogenous protein and impose Panx1‐like properties on it. To test whether Panx1^1‐89 provided the pore lining of the induced channel by itself, we generated a cysteine substitution at position T62 (T62C) in Panx1^1‐89 (Fig. 8). It has previously been shown that amino acid 62 is pore lining in the wtPanx1 channel [44], a conclusion confirmed by the cryo‐EM structure [17, 18, 19, 20, 21, 22, 23]. Figure 8 shows the effect of the thiol reagent MTSET (1 mM) on currents from oocytes expressing the cysteine replacement mutant Panx1^1‐89‐T62C. As observed for the equivalent mutation in wtPanx1, the same concentration of MTSET attenuated the currents carried by Panx1^1‐89‐T62C but not the currents carried by Panx1^1‐89 (Fig. 8). Interestingly, the effect of MTSET reversed after washout (Fig. S2), as had also been observed for Panx1‐T62C/C426S [24]. In most other ion channels, the effect of thiol reagents such as MTSET is irreversible unless reducing agents are employed. However, it has been shown for large pore channels, such as channels formed by connexins, that the effect of thiol reagents reverses spontaneously upon withdrawal of the thiol reagent [24, 46, 47, 48]. This reversibility has been explained by the ability of cytoplasmic reducing agents, in particular glutathione, to enter the large pore and gain access to the modified cysteine in the channel pore. Thus, the reversibility of the MTSET effect on Panx1^1‐89‐T62C is a further indication that the channel formed by this polypeptide is in a large pore conformation, similar to the one observed in wtPanx1 in response to selected stimuli.

Fig. 8 — Effect of (2‐(trimethylammonium)ethyl)MethaneThioSulfonate (MTSET) on membrane currents of the cysteine replacement mutant Panx1^1‐89‐T62C. (A) 3–5 h after injection of Panx1^1‐89‐T62C mRNA, oocytes were held at −60 mV and stepped to +60 mV. Application of 1 mM MTSET was followed by perfusion of 100 μM CBX to verify that the currents were mediated by Panx1^1‐89. (B) As a control, 1 mM MTSET was applied to oocytes expressing Panx1^1‐89. (C) Quantitative analysis of the inhibition of membrane currents by MTSET in oocytes expressing Panx1^1‐89‐T62C (green) or Panx1^1‐89 (blue). Data are presented as mean ± SEM. Currents from “n” oocytes over “N” batches (n/N); n/N = 6/2 (Panx1^1‐89‐T62C, green), n/N = 4/2 (Panx1^1‐89‐T62C, blue). Statistical significance was determined using Student's t‐test.

Effect of reducing agents on membrane currents induced by Panx1^1‐89 expression

Wild‐type Panx1 contains four conserved extracellular cysteines, two in each of the two extracellular loops. Mutation of any of these cysteines leads to the loss of channel function, suggesting a key structural/functional role of these moieties [49]. Furthermore, the reducing agent TCEP has been shown to attenuate wtPanx1 currents [50]. In Panx1^1‐89, two of these four cysteines are retained. In wtPanx1, these two cysteines are in close proximity. To test whether these cysteines are disulfide bonded, the reducing agent TCEP was applied. As shown in Fig. 9a,b, 10 mM of TCEP attenuated the Panx1^1‐89 currents by 45.7 ± 5.6%. Subsequent addition of the thiol reagents MPB (100 μM) or MTSET (1 mM) did not further affect the currents (Fig. 9b), indicating that the cysteines (C66 and C84) do not contribute to the pore lining. The spontaneous ATP release by Panx1^1‐89 expressing oocytes was nearly abolished by the addition of TCEP (Fig. 9c). The large effect on ATP release with a more than 50% residual conductance after TCEP application suggests a change in permeability, so that the flux of larger molecules is more affected than that of smaller ions (Fig. S3). The published cryo‐EM structures of wild‐type hPANX1 show the cysteines disulfide bonded between the two extracellular loops (C66 to C265 and C84 to C246) [21, 22]. Such bonding is not possible in Panx1^1‐89. However, the two cysteines in the first extracellular loop are in close proximity in wt Panx1. Thus, in Panx1^1‐89, the two cysteines appear to be well positioned for disulfide bonding.

Fig. 9 — Effect of the reducing agent TCEP on Panx1^1‐89‐mediated membrane currents and ATP release. (A) Representative trace of membrane current induced by voltage steps from −60 to +60 mV (as indicated) in an oocyte injected with Panx1^1‐89 mRNA 5 h prior recording. Application of 10 mM TCEP was followed by perfusion of 1 mM MTSET. (B) Membrane currents recorded from 9 oocytes expressing Panx1^1‐89 before (OR) and after application of TCEP. In 3 oocytes TCEP was followed by application of the thiol reagent MPB (100 μM, blue) and in another 3 oocytes by MTSET (1 mM, green). (C) Spontaneous ATP release of oocytes expressing Panx1^1‐89 incubated in oocyte Ringer solution (OR) or in OR supplemented with TCEP. Since TCEP attenuates the luciferase reaction used for ATP determination, the supernatant from the oocytes in OR were spiked with an equimolar amount of TCEP. Uninjected oocytes served as control and the mean values were subtracted from the values of Panx1^1‐89 expressing oocytes. In (B, C) data present mean ± SEM, Currents from “n” oocytes over “N” batches (n/N); in (B) n/N = 9/3 and in (C) n/N = 8/2. Statistical significance was determined using Student's t‐test.

The reducing effect of TCEP not only attenuated Panx1^1‐89 currents and ATP release, but also abolished the response of Panx1^1‐89 to increased extracellular [K⁺]. Figure 10 shows that in the presence of TCEP, application of 85 mM [K⁺] did not increase the holding current or the membrane conductance, in contrast to what was typically observed in the non‐reduced Panx1^1‐89 channel (compare to Fig. 5). These data indicate that reduction of the disulfide bonds in Panx1^1‐89 results in a major conformational change of the extracellular channel entry (Fig. S3), which affects currents, permeability, and activation by K⁺.

Fig. 10 — Loss of activation of Panx1^1‐89 by extracellular K⁺ in response to reduction by TCEP. (A) Current traces of an oocyte expressing Panx1^1‐89 4 h after injection of mRNA. The membrane potential was clamped at – 60 mV and voltage steps to +60 mV (voltage protocol shown on the left) were applied at a frequency of 0.1 Hz. 10 mM TCEP was applied as indicated by the bar. At the time point indicated by the downward arrow, the pulse protocol was changed to 12 mV pulses from a −60 mV holding potential (voltage protocol from −60 mV to −48 mV, inset) to test the effect of K⁺ on membrane conductance. 85 mM K⁺ was applied in the presence of TCEP as indicated by the bar. Scale: 1 μA for 120 mV pulses and 0.5 μA for 12 mV pulses. (B) Quantitative analysis of the effect of TCEP on 120 mV voltage step‐induced membrane currents (left) and membrane conductance during the 12 mV voltage step period (right). In contrast to non‐reducing conditions (compared to Fig. 4), K⁺ did not activate Panx1^1‐89. The attenuation in membrane conductance during the application of K⁺, could be due to a continued reduction of Panx1^1‐89 activity by TCEP or a negative effect of K⁺ on membrane conductance under reducing conditions. Data represent mean ± SEM. Currents from “n” oocytes over “N” batches (n/N); n/N = 5/2. Statistical significance was determined using Student's t‐test.

Discussion

Building on the discovery that metastatic breast cancer cells express Panx1^1‐89, our results showed that established Panx1 inhibitors reduced the currents of oocytes expressing Panx1^1‐89. Also, like wtPanx1 channels, they were hyperactivated by K⁺, and their sensitivity to reducing agents and the reactivity of the cysteine replacement mutant (Panx1^1‐89‐T62C) to thiol reagents were all consistent with Panx1^1‐89 forming a membrane channel rather than modulating an endogenous channel.

The correlation between the expression of a severely truncated Panx1 polypeptide, Panx1^1‐89, in metastatic breast cancer cells and the well‐established role of extracellular ATP in the metastasis of several types of cancers led Furlow et al. [14] to formulate an intriguing hypothesis for a Panx1 role in metastatic cell survival in the microvasculature. They suggested that the Panx1^1‐89 polypeptide amplifies the release of ATP by the membrane channel formed by wtPanx1 [14]. Since measurements of ATP release were the sole basis for this conclusion, we set out to perform a more detailed characterization of the biophysical properties of these presumably heteromeric channels in the Xenopus oocyte expression system. However, control oocytes expressing exclusively Panx1^1‐89 exhibited a membrane conductance as soon as 3 h after injection of the mRNA. This conductance was inhibited by a series of inhibitors of the wtPanx1 channel, including probenecid, carbenoxolone, BB FCF, and BzATP. Thus, it appears that Panx1^1‐89 may not require the presence of wtPanx1 to render a membrane conductance. To do this, Panx1^1‐89 could have associated with an endogenous membrane channel, or it may have formed a channel by itself, probably as an oligomer. That latter interpretation is more likely, since similar results were obtained in HEK293T cells, where the His‐tagged Panx1^1‐89 polypeptide appeared as early as 4 h after transfection and yielded CBX‐sensitive currents similar to those observed in oocytes.

It is unlikely, if not impossible, that a polypeptide (like Panx1^1‐89) with a single transmembrane segment forms a membrane pore as a monomer. Instead, much as viroporins with a single transmembrane segment form pores by oligomerizing 4, 5, or 6 subunits [31, 32, 33, 34], Panx1^1‐89 may oligomerize. Based on cryo‐EM data, the wtPanx1 channel assembles as a heptamer [17, 18, 19, 20, 21, 22, 23]. Presently, it is not known whether the channel formed by Panx1^1‐89 assumes any defined oligomeric state. It is possible that Panx1^1‐89 assembles into various oligomeric states since elements controlling the oligomerization of the wt channel possibly are missing in the Panx1^1‐89 polypeptide. Thus, different oligomeric states may co‐exist.

When co‐expressed with wtPanx1, as in cancer cells, it may be that various heteromeric and homomeric assemblies form. In metastatic cancer cells, the ratio of wtPanx1 to Panx1^1‐89 mRNA was about 3:1¹⁴. The observation that in oocytes, Panx1^1‐89 channels appeared much earlier than wt Panx1 channels, suggests a higher translation rate of the shorter polypeptide. Consequently, the distribution of the various heteromers and homomers is unknown. Nevertheless, it is likely, based on co‐injection, that a certain (albeit small) percentage of co‐expressing cells contain homomeric Panx1^1‐89 channels. Cells expressing Panx1^1‐89 alone died, while co‐expression with wtPanx1 allowed the cells to survive. The protective effect of wtPanx1 apparently applies to both oocytes and HEK293T cells. This observation might indicate that in cancer cells there is a selection for co‐expressors. A sufficient number of Panx1^1‐89 channels may be required to release enough ATP to facilitate metastasis. A larger number, however, would promote cell death due to excessive activity of channels that would run down membrane gradients, including that for ATP. Thus, there might be an optimal ratio between the expression of Panx1^1‐89 and wtPanx1 in cancer cells.

The mechanism by which wtPanx1 protects cells from the deleterious effect of Panx1^1‐89 is unclear and needs further investigation. Several processes at different levels are conceivable. For example, the protection could occur during translation by competition between the wt Panx1 mRNA and the Panx1^1‐89 mRNA. The longer amino acid chain of wtPanx1 will occupy the translational machinery for a longer time and thereby could reduce the translation rate of Panx1^1‐89. Incorporation of part of the Panx1^1‐89 polypeptides into heteromers with wtPanx1 subunits could limit the number of potentially lethal Panx1^1‐89 homomeric channels. Alternatively, albeit less likely in the absence of organizing domains in Panx1^1‐89, all Panx1^1‐89 subunits could be incorporated into heteromers with wtPanx1, yielding channels with low open probability and/or low ATP permeability. Finally, heteromerization of wtPanx1 and Panx1^1‐89 could destine these channels for accelerated degradation.

Although the truncation deletes 80% of the Panx1 protein, it is plausible that Panx1^1‐89 could form a functional membrane channel. The 1–89 sequence contains most or all of the pore‐lining amino acids of the wt channel [17, 18, 19, 20, 21, 22, 44]. In addition, the “binding sites” for channel regulators located in the first extracellular loop of the wt channel are still contained in the Panx1^1‐89 polypeptide [17, 36, 37, 38, 39, 51]. While Panx1^1‐89 may associate with a channel endogenous to oocytes, the finding that the conductance induced by Panx1^1‐89 is inhibited by thiol reagents in a cysteine mutant, Panx1^1‐89‐T62C, similar to the inhibition of Panx1‐T62C/C426S by the same thiol reagent, suggests that Panx1^1‐89 by itself forms the channel pore. There is precedence for single membrane‐spanning polypeptides to form functional membrane channels, as shown for viroporins or alamethicin [31, 32, 33, 34, 52, 53].

In their initial paper on Panx1^1‐89 found in metastatic cells, Furlow et al. did not observe ATP release by cells expressing Panx1^1‐89 alone [14]. In the present study, we found that the survival time of oocytes was limited to about 12 h after injection of mRNA for Panx1^1‐89. Similarly, in cultures of HEK293T cells co‐transfected with Panx1^1‐89 and YFP, the YFP signal was lost within 24 h after transfection. These observations provide a plausible explanation for the absence of ATP release by Panx1^1‐89 when expressed alone in HEK293T cells. HEK293T cells that did express Panx1^1‐89 at high levels might not have survived 24 h post transfection, and accordingly, the HEK293T cells surviving transfection might not have expressed Panx1^1‐89 or might have done so at a reduced level.

In contrast to mammalian cell cultures, in the oocyte expression system every cell is accounted for in terms of time of injection, appearance of a new membrane conductance, pharmacology, and cell survivability. The observation that oocytes co‐injected with Panx1^1‐89 and wtPanx1 mRNAs survived several days suggests that wtPanx1 is protective against the effects of Panx1^1‐89 and that Panx1^1‐89 and wtPanx1 proteins might interact with each other. For example, the channels formed by these proteins could be heteromeric and exhibit a lower open probability and/or a smaller pore size.

There is ample evidence in the literature that constitutive activity of the Panx1 channel can be due to various mechanisms, including not only truncations but also single amino acid changes without any simple recognizable pattern [44, 49, 54, 55, 56, 57]. These studies have shown that the truncation site needs to be precise, i.e., truncations a few amino acids upstream or downstream of the site do not yield the same consequences.

A well‐documented truncation involves the carboxyterminal 47 amino acids by caspase cleavage of Panx1, which renders the channel constitutively active [40]. In this respect, the removal by caspase of 48 (mouse) or 47 (human) amino acids and the severe truncation in Panx1^1‐89, with 337 amino acids removed, have similar consequences. However, while the caspase‐cleaved wtPanx1 channel, in the absence of an additional stimulus, is a highly selective Cl⁻ channel without ATP permeability [16, 25, 41], the channel formed by Panx1^1‐89 has a different permeability, as indicated by the shift of the reversal potential versus wtPanx1, the release of ATP, and the uptake of cAMP without an additional stimulus. Thus, the Panx1^1‐89, in the absence of a stimulus, is both constitutively active and resembles the large pore formed by wtPanx1 in response to a variety of physiological or pathological stimuli.

Another similarity between wtPanx1 and Panx1^1‐89 channels is their response to increased extracellular potassium ion concentration. Several amino acids involved in the activation of wtPanx1 are still present in the truncated Panx1^1‐89. However, compared to wtPanx1, the response of Panx1^1‐89 to extracellular potassium was considerably stronger, indicating that constraints on K⁺ activation are absent in the truncated Panx1^1‐89 polypeptide. A boost to the K⁺ response has been shown for two alanine replacement mutants of the intact Panx1 protein (D241A and L266A), consistent with such a loss of constraints [38]. Although the concentration of K⁺ required to activate wtPanx1 is beyond the physiological range, it is conceivable that under pathological conditions, such as the penumbra of a stroke lesion [58], sufficiently high [K⁺]_o is reached. For cells expressing Panx1^1‐89, even slight elevations of [K⁺] may result in a large release of ATP.

The present findings do not rule out that in metastatic cells, wtPanx1 and Panx1^1‐89 form heteromeric channels that are constitutively active and permeant to ATP, as proposed [14]. But the present results do show that Panx1^1‐89 can form functional membrane channels and mediate ATP release even in the absence of wtPanx1. Most importantly, the homomeric Panx1^1‐89 channels are inhibited by the same drugs that inhibit wtPanx1 channels. Thus, compounds such as probenecid should be considered for complementary use to standard treatments of metastatic cancers if the Panx1^1‐89 polypeptides are present.

Another aspect of the present findings is that Panx1^1‐89 may simplify the structure–function analysis of Panx1 channels. Since Panx1^1‐89 retains most basic functional motifs of the wt channel, the pore structure analysis should be facilitated. In addition, because Panx1^1‐89 forms constitutively open channels of pore dimensions that allow ATP conduction, cryo‐EM studies on Panx1^1‐89 might capture the long‐sought large‐pore conformation of Panx1 channels.

Materials and methods

Materials

Mouse pannexin1 was kindly provided by Dr. Rolf Dermietzel (University of Bochum). Human CFTR was kindly provided by Dr. Seth Alper (Harvard Medical School). ATP, BzATP, Brilliant Blue FCF (BB FCF), cAMP, carbenoxolone (CBX), Forskolin, and tris(2‐carboxyethyl)phosphine (TCEP) were purchased from Sigma Aldrich. Probenecid was obtained from Alfa Aesar. The thiol reagents MTSET and MPB were purchased from Toronto Research Chemicals and Sigma Aldrich, respectively.

Mutagenesis

The Panx1 mutations were engineered with the QuickChange II site‐directed mutagenesis kit (Stratagene) according to the manufacturer's specifications. The purified mutant plasmids were sequenced by Genewiz.

The following primers were used:

Stop codon following Q89:

5′‐GCTGCTGTACAGTAGAAGAGCTCCCTGC‐3′

5′‐GCAGGGAGCTCTTCTACTGTACAGCAGC‐3′

Cysteine replacement T62C:

5′‐GGAGATCTCCATCGGTTGTCAGATAAGCTGC‐3′

5′‐GCAGCTTATCTGACAACCGATGGAGATCTCC‐3′

His tag at COOH terminal:

5′‐GCTGGGCTGCTGTACAGCATCATCACCACCATCACTAAGAATTCAAGG‐3′

5′‐CCTTGAATTCTTAGTGATGGTGGTGATGATGCTGTACAGCAGCCCAGC‐3′

Preparation of oocytes

All procedures were approved by the University of Miami Institutional Animal Care and Use Committee (# 20‐046‐LF) and conducted in accordance with the Guiding Principles for Research Involving Animals and Human Beings of the American Physiological Society. Study details are in accordance with ARRIVE guidelines for the use of laboratory animals. Ovaries were harvested from adult female Xenopus laevis. Oocytes were prepared following procedures previously described in [25]. Briefly, ovaries were cut into small pieces and incubated in collagenase (2.5 mg·mL⁻¹; Worthington) in calcium‐free oocyte Ringer (OR) solution, stirring at one turn per second at room temperature, until the follicle separates from the oocyte (typically ~3 h). After thorough washing with regular OR (82.5 mM NaCl, 2.5 mM KCl, 1 mM Na₂HPO₄, 1 mM MgCl₂, 1 mM CaCl₂, and 5 mM HEPES), oocytes devoid of follicle cells and having a uniform pigmentation were selected and stored in OR at 18 °C for 18 h to 3 days before electrophysiological analysis at room temperature.

Preparation of mRNA and electrophysiology

We prepared mRNAs and performed electrophysiological recordings following procedures previously described in [25]. Briefly, the plasmid containing mouse pannxin1 or its mutants in pCS2 was linearized with Not I. The CFTR cDNA was linearized with XhoI. In vitro transcription was performed with SP6 (Panx1) or T7 (CFTR) polymerase, using the Message Machine kit (Ambion, Austin, TX). mRNAs were quantified by absorbance (260 nm), and the proportion of full‐length transcripts was checked by agarose gel electrophoresis. In vitro‐transcribed mRNAs (60 nL for wt and 32 nL for Panx1^1‐89 at ~1 μg·μL⁻¹) were injected into Xenopus oocytes. Cells were kept in regular Ringer solution with the antibiotic streptomycin (10 mg·mL⁻¹). Whole‐cell membrane currents of oocytes were measured using a two‐electrode voltage clamp (Gene Clamp 500B, Axon Instruments/Molecular Devices) under constant perfusion according to the protocols described in the Figures. Glass pipettes were pulled using a P‐97 Flaming/Brown puller (Sutter). Both voltage‐measuring and current‐passing microelectrodes were filled with 3 m KCl. Three electrophysiological protocols were usedto determine membrane current and conductance. (1) the membrane potential was held at −60 mV, and small test pulses lasting 6 s to −48 mV were applied at a frequency of 0.1 Hz. Conductance was calculated by dividing the step‐induced currents by 12. (2) To determine the effect of drugs, test pulses from −60 to +60 mV were applied at a frequency of 0.1 Hz. % inhibition was calculated by (i _max − i _min) × 100/i _max. i _min was determined when the inhibition reached a steady state level. (3) Alternatively, voltage ramps lasting 35 s were applied, typically from −100 mV to +100 mV. OR containing elevated extracellular potassium was made by substituting K⁺ for equivalent concentrations of Na⁺.

Because of the abbreviated lifetime of oocytes expressing Panx1^1‐89, only a short time window of 4 to 8 h after mRNA injection was available for electrical recording and determination of ATP release. In contrast, adequate expression of CFTR required ~48 h after the injection of mRNA. For co‐expression of CFTR and Panx1^1‐89, CFTR mRNA was injected first, followed by the injection of Panx1^1‐89 48 h later. After 5–7 h, co‐expressing cells were subjected to voltage clamp analysis.

Patch clamp electrophysiology

Whole‐cell patch clamp experiments were performed in HEK293T cells (ATCC cat # CRL‐3216, RRID: CVCL_0063). The cells were routinely authenticated by STR (short tandem repeat) analysis, G‐band karyotyping, and validated to be mycoplasma‐free using a PCR‐based assay. Recordings were made using an inverted Nikon microscope (Eclipse TE2000‐U) custom‐made for fluorescence and equipped with a 20× 0.8NA and 40× 0.8NA objectives. Cells were continuously perfused with a bath solution (in mM): 135 NaCl, 5 KCl, 1 MgCl₂, 1 CaCl₂, 2.8 Na Acetate, 10 ‘4‐(2‐hydroxyethyl)‐1‐piperazineethanesulfonic acid’ (HEPES), pH 7.4. Records were made at room temperature ~22 °C. Patch pipettes were pulled from borosilicate glass (outer diameter, 1.5 mm; inner diameter, 0.86 mm, Sutter Instruments) using a horizontal puller (P‐97 Sutter Instruments) and fire‐polished prior to use. Electrode resistances were between 2 and 5 MΩ when filled with the intracellular solution (in mM): 130 KCl, 5 EGTA, 1 MgCl₂, 4 Na₂ATP, 0.1 GTP‐tris, 10 HEPES, pH 7.2. Voltage‐clamp recordings were acquired using an Axopatch 200B amplifier (Molecular Devices, San Jose, CA, USA), digitized at 10 kHz (filtered at 5 kHz, Axon Digidata 1550; Molecular Devices), and collected using pClamp 11 (Axon Instruments, San Jose, CA, USA). Cells were held at −80 mV for 1 s followed by a slow ramp protocol from −100 to +100 mV for 30 s before returning to the −80 mV holding potential.

Immunohistochemistry

Oocytes were injected with 41.1 nL of mRNA of Panx1‐wt (3.4 μg·μL⁻¹ mRNA) or Panx1^1‐89 His6 (2.2 μg·μL⁻¹ mRNA). After injection (4 and 24 h for wt Panx1 and 4 and 8 h for Panx1^1‐89‐His6) Oocytes were incubated in OR3 ND96 solution (96 mM NaCl, 2 mM KCl, 1 mM MgCl2, 1.8 mM CaCl2, 5 mM HEPES; pH = 7.5) at 18 °C and embedded in OTC compound (Tissue‐Tek) and frozen at −80 °C. Oocytes were cut into 40 μm sections transverse to the animal–vegetal (A–V) axis using a cryostat. For immunohistochemistry, slices were washed 2 times with PBS for 5 min and incubated in blocking solution (Universal Blocker Reagent; Biogenex, San Ramon, CA, USA) for 1 h. Thereafter, slices were incubated overnight at 4 °C with primary antibody diluted in blocking solution (rabbit Anti‐6x His tag^® antibody, ab213204, Abcam, Cambridge, UK, 1:60). After three washes with PBS, Goat anti‐Rabbit IgG (Heavy chain) Superclonal™ Secondary Antibody, Alexa Fluor™ 488 (A27034, Invitrogen, Carlsbad, CA, USA, 1:1000) was used as the secondary antibody, and samples were incubated for 2 h at room temperature. In the second hour of incubation, Wheat Germ Agglutinin (WGA) Alexa Fluor Plus 568 conjugates (W56133, Invitrogen, 5 μg·mL⁻¹) were added to label the plasma membrane. Finally, the slices were washed 3 times with PBS, and cell nuclei were stained with DAPI (4′,6‐diamidino‐2‐phenylindole, D1306, Invitrogen 1:1000).

HEK293T cells were cultured at 37 °C in 5% CO₂ in DMEM/12 containing 10% fetal bovine serum and 1% Penicillin/Streptomycin. Cells were grown on 8‐well Ibidi Chamber (5 × 10⁵ cells·well⁻¹, 300 μL·well⁻¹) one day before transfection so that cells were approximately 50–80% confluent on the day of transfection. Cells were transfected with 600 ng of plasmid mix 300 ng of Panx1 or PanX1^1‐89 His6 and 300 ng of YFP using FuGENE^® HD Transfection Reagent (Promega, Madison, WI, USA) according to the manufacturer's instructions. 24 h (for wtPanx1) or 4 and 8 h (for PanX1^1‐89 His6) after transfection, cells were washed twice with PBS and fixed in a 4% formaldehyde PBS solution for 15 min at 4 °C. Cells were washed 3x in PBS and incubated in blocking solution (Universal Blocker Reagent; Biogenex, San Ramon, CA, USA) for 1 h. Subsequently, cells were incubated overnight at 4 °C with primary antibody diluted in blocking solution (rabbit Anti‐6X His tag® antibody, ab213204, Abcam, 1:50). After three washes with PBS, Goat anti‐Rabbit IgG Secondary Antibody, Alexa Fluor™ 647 (A56570, Invitrogen, 1:1000 in PBS) was used as the secondary antibody, and samples were incubated for 2 h at room temperature. In the second hour of this incubation, Wheat Germ Agglutinin (WGA) Alexa Fluor Plus 568 conjugates (W56133, Invitrogen, 5 μg·mL⁻¹) were added to label the plasma membrane. Finally, the cells were washed 3 times with PBS and nuclei were stained with DAPI (D1306, Invitrogen, Carlsbad, CA, USA, 1:1000).

Oocyte sections and HEK293T cells were mounted with Fluoromount‐G mounting medium (Southern Biotech, Birmingham, AL, USA) and imaged on an inverted laser‐scanning confocal microscope (Leica TCS SP5; Leica Microsystems, Deerfield, IL, USA) with LAS AF software using a 63× oil immersion objective (NA 1.4). Images were analyzed using imagej software (National Institute of Health, ImageJ.net).

ATP release assay

ATP flux was determined by luminometry. Oocytes were analyzed 4 h after injection of Panx1^1‐89 mRNA or 2 days after injection of wtPannexin1. Ninety microliters of the oocyte supernatant were added to 40 μL of luciferase‐luciferin solution (Promega, Madison, WI, USA) for assaying luciferase activity. Data from oocytes with visible damage were excluded from the analysis.

Statistics

All experiments were repeated 3 or more times from at least three batches of oocytes and/or cells. Pairwise comparisons were achieved using either ANOVA and Bonferroni's post hoc test or Student's t‐test, as indicated in each Figure. Data are represented as mean ± SEM (standard error of mean) and “n” represents the number of experiments.

Author contributions

GD and RB‐S conceived the project and supervised all research. CM, GD, and RB‐S designed the experiments. JW, NJL, GD, and MD‐S performed the experiments. JW, NJL, GD, and RB‐S analyzed the data. JW, GD, and RB‐S designed and performed the statistical analysis. GD and RB‐S wrote the manuscript. All authors reviewed the manuscript.

Conflict of interest

The authors declare no conflict of interests.

Peer review

The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer‐review/10.1111/febs.70060.

Supporting information

Fig. S1. Membrane topology of wtPanx1 and Panx1^1‐89 channels.

Fig. S2. Reversibility of the effect of MTSET on membrane channels in oocytes expressing Panx1^1‐89‐T62C.

Fig. S3. Schematic illustrating the effect of the reducing agent TCEP on the pore conformation of Panx1^1‐89.

FEBS-292-3378-s001.pdf^{(589.3KB, pdf)}

Acknowledgements

This work was supported by the National Institutes of Health, NINDS (1R01NS110847) to RB‐S. We thank Drs. H. Peter Larsson and Kenneth Muller for critically reading the manuscript.

Contributor Information

Gerhard Dahl, Email: gdahl@med.miami.edu.

Rene Barro‐Soria, Email: rbarro@med.miami.edu.

Data availability statement

The authors declare that all data supporting the findings of this study are available from the corresponding authors upon reasonable request.

References

1. Vultaggio‐Poma V, Sarti AC & Di Virgilio F (2020) Extracellular ATP: a feasible target for cancer therapy. Cells 9, 2496. [DOI] [PMC free article] [PubMed] [Google Scholar]
2. Brenet M, Martinez S, Perez‐Nunez R, Perez LA, Contreras P, Diaz J, Avalos AM, Schneider P, Quest AFG & Leyton L (2020) Thy‐1 (CD90)‐induced metastatic cancer cell migration and invasion are beta3 integrin‐dependent and involve a Ca(2+)/P2X7 receptor signaling Axis. Front Cell Dev Biol 8, 592442. [DOI] [PMC free article] [PubMed] [Google Scholar]
3. Martin S, Dudek‐Peric AM, Garg AD, Roose H, Demirsoy S, Van Eygen S, Mertens F, Vangheluwe P, Vankelecom H & Agostinis P (2017) An autophagy‐driven pathway of ATP secretion supports the aggressive phenotype of BRAF(V600E) inhibitor‐resistant metastatic melanoma cells. Autophagy 13, 1512–1527. [DOI] [PMC free article] [PubMed] [Google Scholar]
4. Kepp O, Bezu L, Yamazaki T, Di Virgilio F, Smyth MJ, Kroemer G & Galluzzi L (2021) ATP and cancer immunosurveillance. EMBO J 40, e108130. [DOI] [PMC free article] [PubMed] [Google Scholar]
5. Ledderose C, Woehrle T, Ledderose S, Strasser K, Seist R, Bao Y, Zhang J & Junger WG (2016) Cutting off the power: inhibition of leukemia cell growth by pausing basal ATP release and P2X receptor signaling? Purinergic Signal 12, 439–451. [DOI] [PMC free article] [PubMed] [Google Scholar]
6. Penuela S, Gyenis L, Ablack A, Churko JM, Berger AC, Litchfield DW, Lewis JD & Laird DW (2012) Loss of pannexin 1 attenuates melanoma progression by reversion to a melanocytic phenotype. J Biol Chem 287, 29184–29193. [DOI] [PMC free article] [PubMed] [Google Scholar]
7. Freeman TJ, Sayedyahossein S, Johnston D, Sanchez‐Pupo RE, O'Donnell B, Huang K, Lakhani Z, Nouri‐Nejad D, Barr KJ, Harland L et al. (2019) Inhibition of pannexin 1 reduces the tumorigenic properties of human melanoma cells. Cancers Basel 11, 102. [DOI] [PMC free article] [PubMed] [Google Scholar]
8. Sayedyahossein S, Huang K, Li Z, Zhang C, Kozlov AM, Johnston D, Nouri‐Nejad D, Dagnino L, Betts DH, Sacks DB et al. (2021) Pannexin 1 binds beta‐catenin to modulate melanoma cell growth and metabolism. J Biol Chem 296, 100478. [DOI] [PMC free article] [PubMed] [Google Scholar]
9. Bao L, Sun K & Zhang X (2021) PANX1 is a potential prognostic biomarker associated with immune infiltration in pancreatic adenocarcinoma: a pan‐cancer analysis. Channels 15, 680–696. [DOI] [PMC free article] [PubMed] [Google Scholar]
10. Zuzul M, Lozic M, Filipovic N, Canovic S, Didovic Pavicic A, Petricevic J, Kunac N, Soljic V, Saraga‐Babic M, Konjevoda S et al. (2022) The expression of connexin 37, 40, 43, 45 and pannexin 1 in the early human retina and choroid development and tumorigenesis. Int J Mol Sci 23, 5918. [DOI] [PMC free article] [PubMed] [Google Scholar]
11. Liu H, Yuan M, Yao Y, Wu D, Dong S & Tong X (2019) In vitro effect of pannexin 1 channel on the invasion and migration of I‐10 testicular cancer cells via ERK1/2 signaling pathway. Biomed Pharmacother 117, 109090. [DOI] [PubMed] [Google Scholar]
12. Shi G, Liu C, Yang Y, Song L, Liu X, Wang C, Peng Z, Li H & Zhong L (2019) Panx1 promotes invasion‐metastasis cascade in hepatocellular carcinoma. J Cancer 10, 5681–5688. [DOI] [PMC free article] [PubMed] [Google Scholar]
13. Penuela S, Gehi R & Laird DW (2013) The biochemistry and function of pannexin channels. Biochim Biophys Acta 1828, 15–22. [DOI] [PubMed] [Google Scholar]
14. Furlow PW, Zhang S, Soong TD, Halberg N, Goodarzi H, Mangrum C, Wu YG, Elemento O & Tavazoie SF (2015) Mechanosensitive pannexin‐1 channels mediate microvascular metastatic cell survival. Nat Cell Biol 17, 943–952. [DOI] [PMC free article] [PubMed] [Google Scholar]
15. Syrjanen J, Michalski K, Kawate T & Furukawa H (2021) On the molecular nature of large‐pore channels. J Mol Biol 433, 166994. [DOI] [PMC free article] [PubMed] [Google Scholar]
16. Mim C, Perkins G & Dahl G (2021) Structure versus function: are new conformations of pannexin 1 yet to be resolved? J Gen Physiol 153, e202012754. [DOI] [PMC free article] [PubMed] [Google Scholar]
17. Michalski K, Syrjanen JL, Henze E, Kumpf J, Furukawa H & Kawate T (2020) The Cryo‐EM structure of pannexin 1 reveals unique motifs for ion selection and inhibition. eLife 9, e54670. [DOI] [PMC free article] [PubMed] [Google Scholar]
18. Deng Z, He Z, Maksaev G, Bitter RM, Rau M, Fitzpatrick JAJ & Yuan P (2020) Cryo‐EM structures of the ATP release channel pannexin 1. Nat Struct Mol Biol 27, 373–381. [DOI] [PubMed] [Google Scholar]
19. Jin Q, Zhang B, Zheng X, Li N, Xu L, Xie Y, Song F, Bhat EA, Chen Y, Gao N et al. (2020) Cryo‐EM structures of human pannexin 1 channel. Cell Res 30, 449–451. [DOI] [PMC free article] [PubMed] [Google Scholar]
20. Mou L, Ke M, Song M, Shan Y, Xiao Q, Liu Q, Li J, Sun K, Pu L, Guo L et al. (2020) Structural basis for gating mechanism of pannexin 1 channel. Cell Res 30, 452–454. [DOI] [PMC free article] [PubMed] [Google Scholar]
21. Qu R, Dong L, Zhang J, Yu X, Wang L & Zhu S (2020) Cryo‐EM structure of human heptameric pannexin 1 channel. Cell Res 30, 446–448. [DOI] [PMC free article] [PubMed] [Google Scholar]
22. Ruan Z, Orozco IJ, Du J & Lu W (2020) Structures of human pannexin 1 reveal ion pathways and mechanism of gating. Nature 584, 646–651. [DOI] [PMC free article] [PubMed] [Google Scholar]
23. Zhang S, Yuan B, Lam JH, Zhou J, Zhou X, Ramos‐Mandujano G, Tian X, Liu Y, Han R, Li Y et al. (2021) Structure of the full‐length human Pannexin1 channel and insights into its role in pyroptosis. Cell Discov 7, 30. [DOI] [PMC free article] [PubMed] [Google Scholar]
24. Wang J, Ambrosi C, Qiu F, Jackson DG, Sosinsky G & Dahl G (2014) The membrane protein Pannexin1 forms two open‐channel conformations depending on the mode of activation. Sci Signal 7, ra69. [DOI] [PMC free article] [PubMed] [Google Scholar]
25. Wang J & Dahl G (2018) Pannexin1: a multifunction and multiconductance and/or permeability membrane channel. Am J Physiol Cell Physiol 315, C290–C299. [DOI] [PubMed] [Google Scholar]
26. Romanov RA, Bystrova MF, Rogachevskaya OA, Sadovnikov VB, Shestopalov VI & Kolesnikov SS (2012) The ATP permeability of pannexin 1 channels in a heterologous system and in mammalian taste cells is dispensable. J Cell Sci 125, 5514–5523. [DOI] [PMC free article] [PubMed] [Google Scholar]
27. Ma W, Compan V, Zheng W, Martin E, North RA, Verkhratsky A & Surprenant A (2012) Pannexin 1 forms an anion‐selective channel. Pflugers Arch 463, 585–592. [DOI] [PubMed] [Google Scholar]
28. Dahl G (2015) ATP release through pannexon channels. Philos Trans R Soc Lond B Biol Sci 370, 20140191. [DOI] [PMC free article] [PubMed] [Google Scholar]
29. Dahl G (2018) The Pannexin1 membrane channel: distinct conformations and functions. FEBS Lett 592, 3201–3209. [DOI] [PubMed] [Google Scholar]
30. Montal MS, Blewitt R, Tomich JM & Montal M (1992) Identification of an ion channel‐forming motif in the primary structure of tetanus and botulinum neurotoxins. FEBS Lett 313, 12–18. [DOI] [PubMed] [Google Scholar]
31. Nieva JL, Madan V & Carrasco L (2012) Viroporins: structure and biological functions. Nat Rev Microbiol 10, 563–574. [DOI] [PMC free article] [PubMed] [Google Scholar]
32. Surya W, Li Y, Verdia‐Baguena C, Aguilella VM & Torres J (2015) MERS coronavirus envelope protein has a single transmembrane domain that forms pentameric ion channels. Virus Res 201, 61–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
33. Largo E, Queralt‐Martin M, Carravilla P, Nieva JL & Alcaraz A (2021) Single‐molecule conformational dynamics of viroporin ion channels regulated by lipid‐protein interactions. Bioelectrochemistry 137, 107641. [DOI] [PMC free article] [PubMed] [Google Scholar]
34. OuYang B & Chou JJ (2014) The minimalist architectures of viroporins and their therapeutic implications. Biochim Biophys Acta 1838, 1058–1067. [DOI] [PMC free article] [PubMed] [Google Scholar]
35. Miller AN, Houlihan PR, Matamala E, Cabezas‐Bratesco D, Lee GY, Cristofori‐Armstrong B, Dilan TL, Sanchez‐Martinez S, Matthies D, Yan R et al. (2023) The SARS‐CoV‐2 accessory protein Orf3a is not an ion channel, but does interact with trafficking proteins. eLife 12, e84477. [DOI] [PMC free article] [PubMed] [Google Scholar]
36. Qiu F & Dahl G (2009) A permeant regulating its permeation pore: inhibition of pannexin 1 channels by ATP. Am J Physiol Cell Physiol 296, C250–C255. [DOI] [PMC free article] [PubMed] [Google Scholar]
37. Qiu F, Wang J & Dahl G (2012) Alanine substitution scanning of pannexin1 reveals amino acid residues mediating ATP sensitivity. Purinergic Signal 8, 81–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
38. Wang J, Jackson DG & Dahl G (2018) Cationic control of Panx1 channel function. Am J Physiol Cell Physiol 315, C279–C289. [DOI] [PubMed] [Google Scholar]
39. Michalski K & Kawate T (2016) Carbenoxolone inhibits Pannexin1 channels through interactions in the first extracellular loop. J Gen Physiol 147, 165–174. [DOI] [PMC free article] [PubMed] [Google Scholar]
40. Chekeni FB, Elliott MR, Sandilos JK, Walk SF, Kinchen JM, Lazarowski ER, Armstrong AJ, Penuela S, Laird DW, Salvesen GS et al. (2010) Pannexin 1 channels mediate ‘find‐me’ signal release and membrane permeability during apoptosis. Nature 467, 863–867. [DOI] [PMC free article] [PubMed] [Google Scholar]
41. Chiu YH, Ravichandran KS & Bayliss DA (2014) Intrinsic properties and regulation of pannexin 1 channel. Channels 8, 1–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
42. Silverman W, Locovei S & Dahl G (2008) Probenecid, a gout remedy, inhibits pannexin 1 channels. Am J Physiol Cell Physiol 295, C761–C767. [DOI] [PMC free article] [PubMed] [Google Scholar]
43. Jackson DG, Wang J, Keane RW, Scemes E & Dahl G (2014) ATP and potassium ions: a deadly combination for astrocytes. Sci Rep 4, 4576. [DOI] [PMC free article] [PubMed] [Google Scholar]
44. Wang J & Dahl G (2010) SCAM analysis of Panx1 suggests a peculiar pore structure. J Gen Physiol 136, 515–527. [DOI] [PMC free article] [PubMed] [Google Scholar]
45. Levine E, Werner R, Neuhaus I & Dahl G (1993) Asymmetry of gap junction formation along the animal‐vegetal axis of xenopus oocytes. Dev Biol 156, 490–499. [DOI] [PubMed] [Google Scholar]
46. Slavi N, Rubinos C, Li L, Sellitto C, White TW, Mathias R & Srinivas M (2014) Connexin 46 (cx46) gap junctions provide a pathway for the delivery of glutathione to the lens nucleus. J Biol Chem 289, 32694–32702. [DOI] [PMC free article] [PubMed] [Google Scholar]
47. Stridh MH, Tranberg M, Weber SG, Blomstrand F & Sandberg M (2008) Stimulated efflux of amino acids and glutathione from cultured hippocampal slices by omission of extracellular calcium: likely involvement of connexin hemichannels. J Biol Chem 283, 10347–10356. [DOI] [PMC free article] [PubMed] [Google Scholar]
48. Tong X, Lopez W, Ramachandran J, Ayad WA, Liu Y, Lopez‐Rodriguez A, Harris AL & Contreras JE (2015) Glutathione release through connexin hemichannels: implications for chemical modification of pores permeable to large molecules. J Gen Physiol 146, 245–254. [DOI] [PMC free article] [PubMed] [Google Scholar]
49. Bunse S, Schmidt M, Hoffmann S, Engelhardt K, Zoidl G & Dermietzel R (2011) Single cysteines in the extracellular and transmembrane regions modulate pannexin 1 channel function. J Membr Biol 244, 21–33. [DOI] [PubMed] [Google Scholar]
50. Bunse S, Locovei S, Schmidt M, Qiu F, Zoidl G, Dahl G & Dermietzel R (2009) The potassium channel subunit Kvbeta3 interacts with pannexin 1 and attenuates its sensitivity to changes in redox potentials. FEBS J 276, 6258–6270. [DOI] [PubMed] [Google Scholar]
51. Wang J, Jackson DG & Dahl G (2013) The food dye FD&C Blue No. 1 is a selective inhibitor of the ATP release channel Panx1. J Gen Physiol 141, 649–656. [DOI] [PMC free article] [PubMed] [Google Scholar]
52. Gordon LG & Haydon DA (1972) The unit conductance channel of alamethicin. Biochim Biophys Acta 255, 1014–1018. [DOI] [PubMed] [Google Scholar]
53. Pressman BC (1968) Ionophorous antibiotics as models for biological transport. Fed Proc 27, 1283–1288. [PubMed] [Google Scholar]
54. Bunse S, Schmidt M, Prochnow N, Zoidl G & Dermietzel R (2010) Intracellular cysteine 346 is essentially involved in regulating Panx1 channel activity. J Biol Chem 285, 38444–38452. [DOI] [PMC free article] [PubMed] [Google Scholar]
55. Engelhardt K, Schmidt M, Tenbusch M & Dermietzel R (2015) Effects on channel properties and induction of cell death induced by c‐terminal truncations of pannexin1 depend on domain length. J Membr Biol 248, 285–294. [DOI] [PubMed] [Google Scholar]
56. Dourado M, Wong E & Hackos DH (2014) Pannexin‐1 is blocked by its C‐terminus through a delocalized non‐specific interaction surface. PLoS One 9, e99596. [DOI] [PMC free article] [PubMed] [Google Scholar]
57. Zhou J, Mao R, Wang M, Long R, Gao L, Wang X, Jin L & Zhu L (2024) A novel heterozygous missense variant of PANX1 causes human oocyte death and female infertility. J Ovarian Res 17, 1–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
58. Sick TJ, Feng ZC & Rosenthal M (1998) Spatial stability of extracellular potassium ion and blood flow distribution in rat cerebral cortex after permanent middle cerebral artery occlusion. J Cereb Blood Flow Metab 18, 1114–1120. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Fig. S1. Membrane topology of wtPanx1 and Panx1^1‐89 channels.

Fig. S2. Reversibility of the effect of MTSET on membrane channels in oocytes expressing Panx1^1‐89‐T62C.

Fig. S3. Schematic illustrating the effect of the reducing agent TCEP on the pore conformation of Panx1^1‐89.

FEBS-292-3378-s001.pdf^{(589.3KB, pdf)}

Data Availability Statement

The authors declare that all data supporting the findings of this study are available from the corresponding authors upon reasonable request.

[febs70060-bib-0001] 1. Vultaggio‐Poma V, Sarti AC & Di Virgilio F (2020) Extracellular ATP: a feasible target for cancer therapy. Cells 9, 2496. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0002] 2. Brenet M, Martinez S, Perez‐Nunez R, Perez LA, Contreras P, Diaz J, Avalos AM, Schneider P, Quest AFG & Leyton L (2020) Thy‐1 (CD90)‐induced metastatic cancer cell migration and invasion are beta3 integrin‐dependent and involve a Ca(2+)/P2X7 receptor signaling Axis. Front Cell Dev Biol 8, 592442. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0003] 3. Martin S, Dudek‐Peric AM, Garg AD, Roose H, Demirsoy S, Van Eygen S, Mertens F, Vangheluwe P, Vankelecom H & Agostinis P (2017) An autophagy‐driven pathway of ATP secretion supports the aggressive phenotype of BRAF(V600E) inhibitor‐resistant metastatic melanoma cells. Autophagy 13, 1512–1527. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0004] 4. Kepp O, Bezu L, Yamazaki T, Di Virgilio F, Smyth MJ, Kroemer G & Galluzzi L (2021) ATP and cancer immunosurveillance. EMBO J 40, e108130. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0005] 5. Ledderose C, Woehrle T, Ledderose S, Strasser K, Seist R, Bao Y, Zhang J & Junger WG (2016) Cutting off the power: inhibition of leukemia cell growth by pausing basal ATP release and P2X receptor signaling? Purinergic Signal 12, 439–451. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0006] 6. Penuela S, Gyenis L, Ablack A, Churko JM, Berger AC, Litchfield DW, Lewis JD & Laird DW (2012) Loss of pannexin 1 attenuates melanoma progression by reversion to a melanocytic phenotype. J Biol Chem 287, 29184–29193. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0007] 7. Freeman TJ, Sayedyahossein S, Johnston D, Sanchez‐Pupo RE, O'Donnell B, Huang K, Lakhani Z, Nouri‐Nejad D, Barr KJ, Harland L et al. (2019) Inhibition of pannexin 1 reduces the tumorigenic properties of human melanoma cells. Cancers Basel 11, 102. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0008] 8. Sayedyahossein S, Huang K, Li Z, Zhang C, Kozlov AM, Johnston D, Nouri‐Nejad D, Dagnino L, Betts DH, Sacks DB et al. (2021) Pannexin 1 binds beta‐catenin to modulate melanoma cell growth and metabolism. J Biol Chem 296, 100478. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0009] 9. Bao L, Sun K & Zhang X (2021) PANX1 is a potential prognostic biomarker associated with immune infiltration in pancreatic adenocarcinoma: a pan‐cancer analysis. Channels 15, 680–696. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0010] 10. Zuzul M, Lozic M, Filipovic N, Canovic S, Didovic Pavicic A, Petricevic J, Kunac N, Soljic V, Saraga‐Babic M, Konjevoda S et al. (2022) The expression of connexin 37, 40, 43, 45 and pannexin 1 in the early human retina and choroid development and tumorigenesis. Int J Mol Sci 23, 5918. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0011] 11. Liu H, Yuan M, Yao Y, Wu D, Dong S & Tong X (2019) In vitro effect of pannexin 1 channel on the invasion and migration of I‐10 testicular cancer cells via ERK1/2 signaling pathway. Biomed Pharmacother 117, 109090. [DOI] [PubMed] [Google Scholar]

[febs70060-bib-0012] 12. Shi G, Liu C, Yang Y, Song L, Liu X, Wang C, Peng Z, Li H & Zhong L (2019) Panx1 promotes invasion‐metastasis cascade in hepatocellular carcinoma. J Cancer 10, 5681–5688. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0013] 13. Penuela S, Gehi R & Laird DW (2013) The biochemistry and function of pannexin channels. Biochim Biophys Acta 1828, 15–22. [DOI] [PubMed] [Google Scholar]

[febs70060-bib-0014] 14. Furlow PW, Zhang S, Soong TD, Halberg N, Goodarzi H, Mangrum C, Wu YG, Elemento O & Tavazoie SF (2015) Mechanosensitive pannexin‐1 channels mediate microvascular metastatic cell survival. Nat Cell Biol 17, 943–952. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0015] 15. Syrjanen J, Michalski K, Kawate T & Furukawa H (2021) On the molecular nature of large‐pore channels. J Mol Biol 433, 166994. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0016] 16. Mim C, Perkins G & Dahl G (2021) Structure versus function: are new conformations of pannexin 1 yet to be resolved? J Gen Physiol 153, e202012754. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0017] 17. Michalski K, Syrjanen JL, Henze E, Kumpf J, Furukawa H & Kawate T (2020) The Cryo‐EM structure of pannexin 1 reveals unique motifs for ion selection and inhibition. eLife 9, e54670. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0018] 18. Deng Z, He Z, Maksaev G, Bitter RM, Rau M, Fitzpatrick JAJ & Yuan P (2020) Cryo‐EM structures of the ATP release channel pannexin 1. Nat Struct Mol Biol 27, 373–381. [DOI] [PubMed] [Google Scholar]

[febs70060-bib-0019] 19. Jin Q, Zhang B, Zheng X, Li N, Xu L, Xie Y, Song F, Bhat EA, Chen Y, Gao N et al. (2020) Cryo‐EM structures of human pannexin 1 channel. Cell Res 30, 449–451. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0020] 20. Mou L, Ke M, Song M, Shan Y, Xiao Q, Liu Q, Li J, Sun K, Pu L, Guo L et al. (2020) Structural basis for gating mechanism of pannexin 1 channel. Cell Res 30, 452–454. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0021] 21. Qu R, Dong L, Zhang J, Yu X, Wang L & Zhu S (2020) Cryo‐EM structure of human heptameric pannexin 1 channel. Cell Res 30, 446–448. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0022] 22. Ruan Z, Orozco IJ, Du J & Lu W (2020) Structures of human pannexin 1 reveal ion pathways and mechanism of gating. Nature 584, 646–651. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0023] 23. Zhang S, Yuan B, Lam JH, Zhou J, Zhou X, Ramos‐Mandujano G, Tian X, Liu Y, Han R, Li Y et al. (2021) Structure of the full‐length human Pannexin1 channel and insights into its role in pyroptosis. Cell Discov 7, 30. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0024] 24. Wang J, Ambrosi C, Qiu F, Jackson DG, Sosinsky G & Dahl G (2014) The membrane protein Pannexin1 forms two open‐channel conformations depending on the mode of activation. Sci Signal 7, ra69. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0025] 25. Wang J & Dahl G (2018) Pannexin1: a multifunction and multiconductance and/or permeability membrane channel. Am J Physiol Cell Physiol 315, C290–C299. [DOI] [PubMed] [Google Scholar]

[febs70060-bib-0026] 26. Romanov RA, Bystrova MF, Rogachevskaya OA, Sadovnikov VB, Shestopalov VI & Kolesnikov SS (2012) The ATP permeability of pannexin 1 channels in a heterologous system and in mammalian taste cells is dispensable. J Cell Sci 125, 5514–5523. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0027] 27. Ma W, Compan V, Zheng W, Martin E, North RA, Verkhratsky A & Surprenant A (2012) Pannexin 1 forms an anion‐selective channel. Pflugers Arch 463, 585–592. [DOI] [PubMed] [Google Scholar]

[febs70060-bib-0028] 28. Dahl G (2015) ATP release through pannexon channels. Philos Trans R Soc Lond B Biol Sci 370, 20140191. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0029] 29. Dahl G (2018) The Pannexin1 membrane channel: distinct conformations and functions. FEBS Lett 592, 3201–3209. [DOI] [PubMed] [Google Scholar]

[febs70060-bib-0030] 30. Montal MS, Blewitt R, Tomich JM & Montal M (1992) Identification of an ion channel‐forming motif in the primary structure of tetanus and botulinum neurotoxins. FEBS Lett 313, 12–18. [DOI] [PubMed] [Google Scholar]

[febs70060-bib-0031] 31. Nieva JL, Madan V & Carrasco L (2012) Viroporins: structure and biological functions. Nat Rev Microbiol 10, 563–574. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0032] 32. Surya W, Li Y, Verdia‐Baguena C, Aguilella VM & Torres J (2015) MERS coronavirus envelope protein has a single transmembrane domain that forms pentameric ion channels. Virus Res 201, 61–66. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0033] 33. Largo E, Queralt‐Martin M, Carravilla P, Nieva JL & Alcaraz A (2021) Single‐molecule conformational dynamics of viroporin ion channels regulated by lipid‐protein interactions. Bioelectrochemistry 137, 107641. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0034] 34. OuYang B & Chou JJ (2014) The minimalist architectures of viroporins and their therapeutic implications. Biochim Biophys Acta 1838, 1058–1067. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0035] 35. Miller AN, Houlihan PR, Matamala E, Cabezas‐Bratesco D, Lee GY, Cristofori‐Armstrong B, Dilan TL, Sanchez‐Martinez S, Matthies D, Yan R et al. (2023) The SARS‐CoV‐2 accessory protein Orf3a is not an ion channel, but does interact with trafficking proteins. eLife 12, e84477. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0036] 36. Qiu F & Dahl G (2009) A permeant regulating its permeation pore: inhibition of pannexin 1 channels by ATP. Am J Physiol Cell Physiol 296, C250–C255. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0037] 37. Qiu F, Wang J & Dahl G (2012) Alanine substitution scanning of pannexin1 reveals amino acid residues mediating ATP sensitivity. Purinergic Signal 8, 81–90. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0038] 38. Wang J, Jackson DG & Dahl G (2018) Cationic control of Panx1 channel function. Am J Physiol Cell Physiol 315, C279–C289. [DOI] [PubMed] [Google Scholar]

[febs70060-bib-0039] 39. Michalski K & Kawate T (2016) Carbenoxolone inhibits Pannexin1 channels through interactions in the first extracellular loop. J Gen Physiol 147, 165–174. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0040] 40. Chekeni FB, Elliott MR, Sandilos JK, Walk SF, Kinchen JM, Lazarowski ER, Armstrong AJ, Penuela S, Laird DW, Salvesen GS et al. (2010) Pannexin 1 channels mediate ‘find‐me’ signal release and membrane permeability during apoptosis. Nature 467, 863–867. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0041] 41. Chiu YH, Ravichandran KS & Bayliss DA (2014) Intrinsic properties and regulation of pannexin 1 channel. Channels 8, 1–7. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0042] 42. Silverman W, Locovei S & Dahl G (2008) Probenecid, a gout remedy, inhibits pannexin 1 channels. Am J Physiol Cell Physiol 295, C761–C767. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0043] 43. Jackson DG, Wang J, Keane RW, Scemes E & Dahl G (2014) ATP and potassium ions: a deadly combination for astrocytes. Sci Rep 4, 4576. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0044] 44. Wang J & Dahl G (2010) SCAM analysis of Panx1 suggests a peculiar pore structure. J Gen Physiol 136, 515–527. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0045] 45. Levine E, Werner R, Neuhaus I & Dahl G (1993) Asymmetry of gap junction formation along the animal‐vegetal axis of xenopus oocytes. Dev Biol 156, 490–499. [DOI] [PubMed] [Google Scholar]

[febs70060-bib-0046] 46. Slavi N, Rubinos C, Li L, Sellitto C, White TW, Mathias R & Srinivas M (2014) Connexin 46 (cx46) gap junctions provide a pathway for the delivery of glutathione to the lens nucleus. J Biol Chem 289, 32694–32702. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0047] 47. Stridh MH, Tranberg M, Weber SG, Blomstrand F & Sandberg M (2008) Stimulated efflux of amino acids and glutathione from cultured hippocampal slices by omission of extracellular calcium: likely involvement of connexin hemichannels. J Biol Chem 283, 10347–10356. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0048] 48. Tong X, Lopez W, Ramachandran J, Ayad WA, Liu Y, Lopez‐Rodriguez A, Harris AL & Contreras JE (2015) Glutathione release through connexin hemichannels: implications for chemical modification of pores permeable to large molecules. J Gen Physiol 146, 245–254. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0049] 49. Bunse S, Schmidt M, Hoffmann S, Engelhardt K, Zoidl G & Dermietzel R (2011) Single cysteines in the extracellular and transmembrane regions modulate pannexin 1 channel function. J Membr Biol 244, 21–33. [DOI] [PubMed] [Google Scholar]

[febs70060-bib-0050] 50. Bunse S, Locovei S, Schmidt M, Qiu F, Zoidl G, Dahl G & Dermietzel R (2009) The potassium channel subunit Kvbeta3 interacts with pannexin 1 and attenuates its sensitivity to changes in redox potentials. FEBS J 276, 6258–6270. [DOI] [PubMed] [Google Scholar]

[febs70060-bib-0051] 51. Wang J, Jackson DG & Dahl G (2013) The food dye FD&C Blue No. 1 is a selective inhibitor of the ATP release channel Panx1. J Gen Physiol 141, 649–656. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0052] 52. Gordon LG & Haydon DA (1972) The unit conductance channel of alamethicin. Biochim Biophys Acta 255, 1014–1018. [DOI] [PubMed] [Google Scholar]

[febs70060-bib-0053] 53. Pressman BC (1968) Ionophorous antibiotics as models for biological transport. Fed Proc 27, 1283–1288. [PubMed] [Google Scholar]

[febs70060-bib-0054] 54. Bunse S, Schmidt M, Prochnow N, Zoidl G & Dermietzel R (2010) Intracellular cysteine 346 is essentially involved in regulating Panx1 channel activity. J Biol Chem 285, 38444–38452. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0055] 55. Engelhardt K, Schmidt M, Tenbusch M & Dermietzel R (2015) Effects on channel properties and induction of cell death induced by c‐terminal truncations of pannexin1 depend on domain length. J Membr Biol 248, 285–294. [DOI] [PubMed] [Google Scholar]

[febs70060-bib-0056] 56. Dourado M, Wong E & Hackos DH (2014) Pannexin‐1 is blocked by its C‐terminus through a delocalized non‐specific interaction surface. PLoS One 9, e99596. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0057] 57. Zhou J, Mao R, Wang M, Long R, Gao L, Wang X, Jin L & Zhu L (2024) A novel heterozygous missense variant of PANX1 causes human oocyte death and female infertility. J Ovarian Res 17, 1–12. [DOI] [PMC free article] [PubMed] [Google Scholar]

[febs70060-bib-0058] 58. Sick TJ, Feng ZC & Rosenthal M (1998) Spatial stability of extracellular potassium ion and blood flow distribution in rat cerebral cortex after permanent middle cerebral artery occlusion. J Cereb Blood Flow Metab 18, 1114–1120. [DOI] [PubMed] [Google Scholar]

PERMALINK

A metastasis‐associated pannexin‐1 mutant (Panx11‐89) forms a minimalist ATP release channel

Junjie Wang

Noah J Levi

Maykelis Diaz‐Solares

Carsten Mim

Gerhard Dahl

Rene Barro‐Soria

Abstract

Abbreviations

Introduction

Results

Exclusive expression of Panx11‐89 results in cell death

Fig. 1.

Expression of Panx11‐89 induces constitutively active membrane channels

The Panx11‐89‐mediated current is sensitive to wtPanx1 channel blockers

Fig. 2.

Fig. 3.

Fig. 4.

Panx11‐89 responds more strongly to increased extracellular potassium ion concentration than wtPanx1

Fig. 5.

Panx11‐89 expression yields a permeation pathway with large pore properties

Fig. 6.

Fig. 7.

Panx11‐89 may provide the pore lining of the induced channel

Fig. 8.

Effect of reducing agents on membrane currents induced by Panx11‐89 expression

Fig. 9.

Fig. 10.