Figure 1.
Sequence conservation between diverse vertebrate Panx1 channels.
(A) Circles represent individual amino acids of mouse Panx1 from 1 to 426. Coloring represents sequence homology among a diverse sample of vertebrate Panx1 channel amino acid sequences. The sequences used in this analysis are mouse, rat, horse, human, cat, dog, cow, anole (Anolis carolinensis), zebrafish, collared flycatcher (Ficedula albicollis), and frog (Xenopus tropicalis). Colors were assigned as follows: dark blue (0–1 different residues present among these species), light blue (2 different residues), yellow (3 different residues), orange (4 differences), and red (5 or more different residues). Regions of low conservation among this sample of vertebrate species are part of the first extracellular loop (residues 90–100), part of the loop between TM2 and TM3 (residues 160–180), and the c-terminus after the caspase cleavage site (indicated). (B) A multiple sequence lineup shows the sequences of the c-terminus region in detail.
Figure 2.
Unlike full-length mouse Panx1, Panx1 truncated at amino acid 379 (Δ379) is constitutively active.
(A) HEK293T cells transfected with Δ379, full-length (FL) Panx1, or pCDNA3.1 vector alone (mock) were assayed for Yo-Pro influx (example images shown). Scale bars are 100 µm. (B) Time-lapse fluorescence microscopy was used to collect images of the Yo-Pro fluorescence once per minute over a period of 1 hour following addition of 1 µM Yo-Pro-1 to the cells. (B-inset) The functional activity of the channels is represented as the slope of the Yo-Pro uptake over the initial 30 minutes. A one-way ANOVA comparing all pairs of mean slopes with Tukey-Kramer correction for multiple comparisons was used to show that Δ379 has significantly higher Yo-Pro influx compared to mock and FL (p<0.01), while FL was not different than mock. Full-length and Δ379 Panx1 are expressed on the plasma membrane as determined by (C) ICC and (D) surface biotinylation-Western blot. A differential interference contrast (DIC) image (C-right) is shown side-by-side with the ICC of FL Panx1. Scale bars for ICC are 10 µm. Both the isolated (D-left) surface fraction as well as the (D-right) cytoplasmic fraction are shown with loading controls transferrin receptor (TfR) and actin respectively. Such Western blots show the three forms of full-length Panx1 (non-glycosylated gly0, partially-glycosylated gly1, and fully glycosylated gly2 as indicated) while only gly0 and gly1 forms are present in the case of Δ379. Typical whole-cell patch clamp recordings of cells expressing either (E-left, black) Δ379 or (E-right, black) FL Panx1 relative to empty plasmid (red in both left and right) are shown following leak subtraction with 100 µM carbenoxolone. The current-voltage (I–V) curves shown here were generated by ramping the membrane voltage from −100 mV to +60 mV in 500 msec (from a holding voltage of −20 mV).
Figure 3.
Panx1 channels with double-alanine substitutions within the c-terminus did not show detectable Yo-Pro-1 influx.
(A) HEK293T cells were transfected with a series of double-alanine substitutions and assayed for Yo-Pro influx (example images shown). Scale bars are 100 µm. The Δ379 truncation mutant is used as a positive control for the assay. (B) All the substitution mutants were expressed on the membrane surface as shown by surface biotinylation-Western. (C and D) Pairwise Student's T-test comparisons show that there was no detectable increase in Yo-Pro influx over time (relative to mock transfected cells) for any of the double-alanine substitutions tested.
Figure 4.
C-terminal truncations of mouse Panx1 shorter than Δ407 result in channels with detectable constitutive activity.
(A) HEK293T cells were transfected with a series of Panx1 truncations and assayed for Yo-Pro influx with FL and Δ379 as negative and positive controls respectively (example images shown). Scale bars are 100 µm. (B) All truncations were expressed on the membrane surface except for Δ327 as shown by surface biotinylation-Western. (C) Yo-Pro influx was measured using fluorescence microscopy and (D) normalized slopes (normalized to the Δ379 control) of the initial 30 min were plotted. Statistical analysis was carried out for each individual experiment using a one-way ANOVA comparing the average Yo-Pro fluorescence between 3000 and 4000 sec for each mutant against the FL Panx1 control with Tukey-Kramer correction for multiple comparisons. The Yo-Pro fluorescence of each truncation mutant shorter than Δ407 and longer than Δ365 (as well as Δ347) was found to be significantly greater than that of FL control-transfected cells in all experiments (p<0.05). The length required for half-maximal Yo-Pro influx was determined to be 400.9 amino acids by fitting the data to a 3-parameter logistic function. (E) The c-terminal sequence of the truncated mutants are as shown.
Figure 5.
Poly-alanine (pAla) substitutions did not result in a constitutively active Panx1 channel.
(A) HEK293T cells were transfected with a series of poly-alanine mutant Panx1 channels and assayed for Yo-Pro influx (example images shown). Scale bars are 100 µm. (B) All the poly-alanine mutants were expressed at the cell surface except for the outer tail-end substitution pAla-5, which did not express as determined by surface biotinylation-Western. (C and D) None of the poly-alanine mutants tested displayed any measurable Yo-Pro influx over time when compared with the Δ379 mutant. In particular, pairwise Student's T-test comparisons showed no statistically significant difference between mock, FL, and any of the poly-alanine mutants. (E) The c-terminal sequences of the poly-alanine mutants are as shown.
Figure 6.
Activity of longer poly-alanine/asparagine Panx1 mutants as well as Panx1 mutants with scrambled c-termini.
C-termini containing long poly-alanine repeats or completely scrambled c-termini were expressed and tested for Yo-Pro influx. As usual, FL and Δ379 served as negative and positive controls respectively. (A and B) All mutant channels were expressed on the cell surface as shown by surface biotinylation-Western. (C and D) Yo-Pro influx was measured using video microscopy and (E and F) normalized slopes of the initial 30 min were plotted. For each set of experiments, a one-way ANOVA with Tukey-Kramer multiple-comparison correction was used to show that pAlaExt, pAlaExt-2, Scr2, and Scr3 were not significantly different from mock or FL while pAlaExt-3 was constitutively active (p<0.05) and not different from Δ379. (G) Panx1 c-terminal sequences of tested mutants are as shown.
Figure 7.
Presence of the pAlaExt mutation causes a 10-amino acid shift in the function-vs-length curve.
C-terminal truncations of the pAlaExt mutation were tested for Yo-Pro influx side-by-side with truncations of the wild-type Panx1 channel. (A) All such mutant channels were expressed on the membrane surface as shown by surface biotinylation-Western. (B) The Yo-Pro influx of these mutants was measured using fluorescence microscopy and normalized slopes were calculated as before. (C) The presence of the pAlaExt mutation shifted the function-vs-length-dependence of c-terminal truncations by 10.5 amino acids, as determined by fitting both sets of data to a 3-parameter logistic function. The parameters were obtained as follows: for the wild-type channel, max = 1.0, xc = 400.9, and slope factor = −0.297. For the pAlaExt mutant channel, max = 1.31, xc = 411.4, and slope factor(k) = −0.52. The logistic function used was y = max/(1+exp(−k*(x-xc))). (D) Sequences of the tested mutant Panx1 channel c-termini are as shown.