InsP 3 R channel gating altered by clustering? ARISING FROM T. Rahman, A. Skupin, M. Falcke & C. W. Taylor Nature 458, 655–659 (2009) The inositol trisphosphate receptor (InsP 3 R) forms a calcium channel that resides in the membrane of the endoplasmic reticulum and is activated by inositol trisphosphate (InsP 3 ). InsP 3 is a phosphorylated monosaccharide that is generated via hydrolysis of phosphatidylino- sitol 4,5-bisphosphate (PIP2), a phospholipid that is located in the plasma membrane, and activation of the InsP 3 R is involved in a broad range of biological processes, including cell division, apoptosis and development. Rahman et al. 1,2 reported that exposure to low concen- trations of InsP 3 induces rapid clustering of InsP 3 R Ca 21 release channels normally randomly distributed in endoplasmic reticulum/ outer nuclear membranes. Importantly, clustered channels gate dif- ferently from lone channels. Using similar protocols, we observed InsP 3 R channel clustering without exposure to InsP 3 (Fig. 1a), as we found in other systems 3–5 with protocols designed to avoid InsP 3 pre- exposure. More significantly, we find that clustering has no effect on InsP 3 R channel gating. For this reason, we believe that InsP 3 -induced channel clustering and modification of channel gating by clustering may not be universal phenomena. Rahman et al. 1,2 reported that in sub-optimal cytoplasmic free Ca 21 concentrations ([Ca 21 ] i ), clustered recombinant rat type 3 InsP 3 R (InsP 3 R3) channels expressed in InsP 3 R-deficient DT40-KO cells gated identically and independently, but with lower open probability (P o ) than lone channels, regardless of cluster size. In contrast, clus- tered channels had the same P o as lone channels in optimal ligand conditions, but gated with positive cooperativity. If broadly observed, these surprising findings have important implications for under- standing InsP 3 -mediated Ca 21 signals, and for quantitative analyses in single-channel InsP 3 R electrophysiology. To verify these observations, we examined the same InsP 3 R3 channels in the same DT40-KO cells using similar protocols and ligand conditions. Specifically, we used 5 mM (same as Rahman et al. 1 ) and 0.5 mM (more physiological) cytoplasmic free [ATP 4– ] ([ATP] f ). Records with #4 active channels were analysed with the same algorithm 1 . In addition, we similarly analysed nuclear patch- clamp records previously acquired under comparable ligand condi- tions for recombinant rat InsP 3 R3 expressed in Xenopus oocytes 6 , endogenous Xenopus type 1 InsP 3 R (InsP 3 R1) in oocytes 7 and endo- genous insect InsP 3 R in Sf9 cells 5 . For all channels examined in these various systems, we detected no statistical difference (P . 0.05, t-test) between P o in single- versus multi-channel patches in saturating [InsP 3 ] and sub-optimal [Ca 21 ] i (Fig. 1b), or in sub-saturating [InsP 3 ] and optimal [Ca 21 ] i (Fig. 1c). Furthermore, in two-channel records, similar channel gating patterns were detected in all [Ca 21 ] i (Fig. 2), with only a small fraction exhibiting positive cooperativity. Thus, our extensive data set reveals no effect of clustering on InsP 3 R channel gating in all ligand conditions. In constant ligand conditions, we consistently observed abrupt, stochastic, irreversible inactivation of InsP 3 R in on-nucleus or excised luminal-side-out nuclear patches, with mean activity durations of ,40 s for oocyte InsP 3 R (ref. 3), ,100 s for Sf9 InsP 3 R (ref. 5) and ,140 s for InsP 3 R from DT40-KO cells, whereas Rahman et al. reported no such inactivation 1 . Importantly, we analysed only current records long enough for the number of active channels to be counted with .99% confidence 1,5,6,8 . Because finite time elapsed between pipettes making contact with the outer nuclear membrane and gigaohm seal formation (,5 s for oocyte and Sf9 nuclei, ,10 s for DT40 nuclei), apparent single-channel patches possibly included a fraction (,11–26%) that actually contained multiple channels in a b c 100 1.0 0.8 0.6 Channel P o Channel [ATP] f (mM) r-3/DK 5 Channel [ATP] f (mM) r-3/DK 5 r-3/DK 0.5 r-3/Xo 0.5 X-1/Xo 0.5 iR/Sf9 0.5 0.4 0.2 0.0 1.0 Channel P o 0.8 0.6 0.4 0.2 0.0 Frequency Observed Poisson 80 60 40 20 0 0 1 2 Number of active channels, N A Records showing one active channel Records showing multiple active channels From ref. 1 From ref. 1 3 4 5 6 7 [InsP 3 ] = 3 μM [InsP 3 ] = 10 μM sub-optimal [Ca 2+ ] i [Ca 2+ ] i = 2 μM Figure 1 | InsP 3 R channels are clustered before exposure to InsP 3 , with gating properties unaltered by clustering. a, N A in nuclear membrane patches with no pre-exposure to InsP 3 obtained from InsP 3 R3 expressing DT40-KO cells. Note nonlinear square-root scale for frequency axis. b, P o observed under saturating [InsP 3 ] and sub-optimal [Ca 21 ] i in multi- and apparent single- channel current records for recombinant rat InsP 3 R3 (r-3) channels expressed in DT40-KO (DK) cells or Xenopus oocytes (Xo), endogenous Xenopus InsP 3 R1 (X-1) channels from Xenopus oocytes (Xo), and endogenous insect InsP 3 R (iR) channels from Sf9 cells. Concentrations of free ATP 4– ([ATP] f ) in the pipette solutions used are indicated. Mean P o with s.e.m. (as error bars) and P o for individual current records are shown, together with mean P o from Rahman et al. 1 c, P o of r-3 channels in DK cells in optimal [Ca 21 ] i and sub- saturating [InsP 3 ], ligand conditions not investigated in Rahman et al. 1 . Same symbols as in b are used. Channel [ATP] f (mM) 0.2 [InsP 3 ] = 10 μM [Ca 2+ ] i : Non-binomial Binomial Different P o Positive cooperativity Cooperativity index 0.1 0.0 –0.1 r-3/DK 5 r-3/DK 0.5 r-3/Xo 0.5 X-1/Xo 0.5 iR/Sf9 0.5 Optimal Sub-optimal Figure 2 | Distribution of cooperativity index for two-channel current records of different InsP 3 R channels in various systems in optimal and sub- optimal [Ca 21 ] i . Filled and open circles represent records with two channels exhibiting identical and independent, or non-binomial gating, respectively. Non-binomial records with cooperativity index, P 2 zP 1 =2 ð Þ 2 {P 2 , significantly greater than 0 (in yellow shaded region) had two channels gating with different P o , and those with cooperativity index significantly smaller than 0 (in blue shaded region) had two channels gating with positive cooperativity. The cooperativity indices have no correlation with the durations of the current records (data not shown) and therefore are unlikely to be significantly affected by current record durations limited by channel inactivation. BRIEF COMMUNICATIONS ARISING 13 OCTOBER 2011 | VOL 478 | NATURE | E1 Macmillan Publishers Limited. All rights reserved ©2011