|
Figure 3. Three cluster patterns are produced upon expression of a mixture of NR2 plus mutant NR1 subunits. Each panel shows example clusters and the amplitude histogram of all clusters from a single patch. (A) The Q/Q pattern (main = −8.0 pA, sub = −5.4 pA; see Fig. 1 B). (B) A novel pattern (NR2 hybrid) with three open channel levels (−8.3, −7.0, and −3.4 pA). This pattern was never observed in receptors with purely mutant or wild-type subunits and, therefore, arises from a hybrid receptor that has both N and Q NR2B subunits. (C) The Q/N pattern (main = −5.9 pA, sub = −0.8 pA; see Fig. 1 D). To show the sublevel more clearly, a dashed line marks the baseline, and the example traces showing of this pattern were taken from continuous recordings.
|
|
Figure 6. Six homogeneous cluster patterns are apparent upon expression of a mixture of NR1 plus mutant NR2 subunits. Each panel is from one patch in which one pattern predominated. The amplitude histograms pertain to all clusters in the patch; i.e., with no separation according to pattern. (A) The Q/Q pattern (see Fig. 1 B; n = 98 clusters). (B) Hybrid receptor pattern H1, with a main and sublevel similar to the Q/Q pattern but with a much higher relative occupancy in the sublevel (n = 44 clusters). (C) Hybrid receptor pattern H2, with a main level amplitude ∼15% lower than the Q/Q main level and with a preferential occupancy of the sublevel (n = 110 clusters). (D) Hybrid receptor pattern H3, with a main level amplitude ∼30% lower than the Q/Q main level, and nearly equal main and sublevel occupancies (n = 14 clusters). (E) Hybrid receptor pattern H4, with a main level of ∼40% lower than the Q/Q pattern, and a preferential occupancy of the sublevel (n = 13 clusters). (F) The N/Q pattern (see Fig. 1 C; n = 165 clusters). The amplitude, standard deviation, occupancy, and lifetime of the main and sublevel of each of these patterns is given in Table III.
|
|
Figure 5. Cluster sub- versus main level amplitudes for mixed NR2 subunit expression. Oocytes were injected with a mixture of wild-type and mutant NR2, plus pure mutant NR1 subunit RNAs. All valid clusters from 10 patches are plotted, irrespective of their classification according to pattern. Three groups are apparent. The bottom group is Q/Q receptors (see Fig. 2). The central group is NR2 hybrid receptors, composed of both wild-type and mutant NR2 subunits. There are three open channel current levels in this pattern; here, the smallest amplitude is plotted as a function of the largest amplitude. The top group is Q/N receptors. The number of clusters in this group is relatively small because during the automated analysis of clusters the sublevel was often not detected.
|
|
Figure 10. Cluster sub- versus main level amplitudes for mixed NR1 subunit expression. Oocytes were injected with a mixture of wild-type and mutant NR1, plus pure mutant NR2 subunit RNAs. All valid clusters from 10 patches are plotted, irrespective of their classification by pattern. Five groups, from six patterns, are apparent. The group at the lower left is Q/Q plus H1 receptors. The topmost group is N/Q receptors. The hybrid receptor groups (H1, H2, H3, and H4) fall along a diagonal that reflects the 1.4 main/sublevel amplitude ratio. Although there is significant scatter with each group, each population can be distinguished from its neighbors. The properties of these clusters is given in Table III; example clusters and amplitude histograms of each pattern are shown in Fig. 6.
|
|
Figure 2. Cluster sub- versus main level amplitudes for Q/Q and N/Q receptor compositions. Oocytes were injected with either mutant NR1 plus mutant NR2 (Q/Q receptors; •, five patches), or wild-type NR1 plus mutant NR2 (N/Q receptors; ▴, three patches). All valid clusters in all of the patches are plotted. A single pattern of current is present with these injections of unmixed subunits. The diagonal distribution of Q/Q amplitudes indicates that scatter in the main and sublevels is correlated.
|
|
Figure 8. Separation of clusters by pattern: a mixed NR1 subunit patch with three patterns. The oocyte was injected with a 1:1 mixture of wild-type and mutant NR1 subunits plus pure mutant NR2 subunits. Clusters with the H1, H3, and H4 patterns were apparent, and were separated using a log likelihood criterion. All 25 clusters in this patch are shown. (A) Amplitude histogram of all clusters, before separation. Only the closed channel component (rightmost) is a distinct peak. (B) Amplitude histograms of clusters after separation by pattern. (B, 1) H1 clusters (n = 14). (B, 2) H4 clusters (n = 3). (B, 3) N/Q clusters (n = 6). (bottom right) Rejected clusters (n = 2) that had a low likelihood because of a multiple opening, or the chance juxtaposition of an H1 and an N/Q cluster. For examples of homogeneous H1, H4, and N/Q patterns see Fig. 6.
|
|
Figure 4. Separation of clusters by pattern: mixed NR2 subunit expression. The oocyte was injected with pure mutant NR1 subunits, plus a 1:1 mixture of wild-type and mutant NR2 subunits. Clusters with the Q/Q and NR2 hybrid patterns were apparent, and were separated using a log likelihood criterion. (A) All-points amplitude histogram of all 134 clusters in the patch, before separation. (B) Amplitude histograms of clusters, after separation. (top) Histogram and examples of separated Q/Q clusters (n = 100, main = −7.9 pA, sub = −5.2 pA); (bottom) histogram and examples of NR2 hybrid clusters (n = 25, open channel amplitudes of −7.9, −6.6, and −3.1 pA). 10 clusters (7.5%) were rejected because of a low log likelihood (not shown). When summed, the Gaussian components of the separated histograms (dotted lines in A) plus that of the rejected clusters equal the combined histogram.
|
|
Figure 7. Separation of clusters by pattern: a mixed NR1 subunit patch with two patterns. The oocyte was injected with a 1:1 mixture of wild-type and mutant NR1 subunits, plus pure mutant NR2 subunits. Clusters with the H1 and H2 patterns were apparent, and were separated using a log likelihood criterion. (A) Amplitude histogram of all 94 clusters in the patch, before separation. (B) Amplitude histograms of clusters after separation. (top) H1 clusters (n = 38; main = −7.8 pA, sub = −5.6 pA; see Fig. 6 B). (bottom) H2 clusters (n = 56; main = −6.8 pA, sub = −4.9 pA; see Fig. 6 C). Three clusters were rejected because of a low likelihood (not shown). When summed, the Gaussian components of the separated histograms (A, dotted lines) plus the rejected clusters equals the combined histogram. For examples of homogeneous H1 and H2 patterns see Fig. 6.
|
|
Figure 9. Separation of clusters by pattern: a mixed NR1 subunit patch with four patterns. The oocyte was injected with a 1:1 mixture of wild-type and mu-tant NR1 subunits, plus pure mutant NR2 subunits, and the outside-out patch was superfused with 20 μM NMDA plus pure 10-μM glycine in divalent-cation-free solution. Clusters with the H1, H3, H4, and N/Q patterns were apparent and were separated using a log likelihood criterion. (A) Continuous 10-s trace at low time resolution. (B) Amplitude histogram from all 444 clusters in the patch, before separation. Only the baseline, and a component at −7.8 pA, are distinct peaks. (C) Amplitude histograms of clusters after separation by pattern. (top left) H1 clusters (n = 154), (top right) H3 clusters (n = 48), (bottom left) H4 clusters (n = 34), (bottom right) N/Q clusters (n = 164). Rejected clusters (n = 44) are not shown. For examples of homogeneous H1, H3, H4, and N/Q patterns see Fig. 6.
|
|
Figure 11. Pattern frequency and amplitudes are stable. Clusters of pattern H1, H3, H4, and N/Q were separated using a log likelihood criterion (same patch as in Fig. 9). The sub- (○) and main (•) level amplitudes are plotted for each cluster as a function of time. The results are consistent with four stable populations of receptor in the patch. Main and sublevel amplitudes of each pattern do not drift, and there is no apparent pattern switching or correlation among cluster frequencies of different patterns.
|
|
Figure 12. Q/Q and H1 receptors respond differently to extracellular Ca2+. Currents are from a single outside-out patch (Vm = −160 mV). The oocyte was injected with a 1:1 mixture of wild-type and mutant NR1, plus pure mutant NR2 subunits. Clusters with the Q/Q and H1 patterns were apparent. All traces are continuous. (A) Cluster patterns in Ca2+-free solutions. (bottom right) Amplitude histogram of all clusters. The main and sublevel peaks for the Q/Q and H1 patterns overlap in Ca2+-free solution. (B) Cluster patterns in 16.7 μM extracellular Ca2+. Because of channel block by Ca2+, the sublevel of the Q/Q pattern is reduced to −1.0 pA, but the sublevel of the H1 pattern is reduced only to −3.0 pA. (bottom right) Amplitude histogram of all clusters. Separate components for main and sublevels of Q/Q and H1 receptors are now resolved in the presence of Ca2+. Note that the H1 pattern appears to have three open channel amplitudes, similar to the NR2 hybrid pattern (Fig. 3). The differential effect of Ca++ on the Q/Q and H1 sublevel demonstrates that these patterns arise from distinct receptor populations.
|
|
Figure 13. Summary of proposed pentameric NMDA receptor stoichiometry. The results are consistent with the presence of three NR1 subunits (circles) and two NR2 subunits (squares) per receptor (black indicates a mutant subunit). It is assumed that both the number and position of the mutant sidechains influence the current pattern, and that compositions that superimpose upon rotation or reflection across a diameter are equivalent with respect to current pattern. The assemblies are drawn with the NR2 subunits apart, although the experiments do not distinguish whether they are adjacent or apart. (A) With three NR1 subunits, six nonequivalent compositions are possible when oocytes are injected with mixed NR1 subunits and pure NR2 subunits. (top) The all-mutant NR1 composition has the Q/Q pattern (left), and the all-wild-type NR1 composition has the N/Q pattern (right). (middle) Receptors with one wild-type and two mutant NR1 subunits can assemble with the mutant sidechains adjacent (left) or apart (right). (bottom) Receptors with two wild-type and one mutant NR1 subunit can assemble with the wild-type sidechains adjacent (left) or apart (right). Based on the results shown in Table IV, H1 and H2 patterns may have two mutant NR1 subunits, and H3 and H4 patterns may have one wild-type NR1 subunit. Example currents from each assembly are shown in Fig. 6. (B) With two NR2 subunits, only three nonequivalent compositions are possible when oocytes are injected with pure NR1 and mixed NR2 subunits. These compositions are associated with the Q/N pattern (top), the NR2 hybrid pattern (middle), and the Q/Q pattern (bottom). Example currents from each assembly are shown in Fig. 3.
|
