Thirtieth Annual Meeting 9–13 February 1986 Brooks Hall/Convention Center, San Francisco, California

1986 Biophysical Journal  
We wish to test the following related hypotheses: (1) human color vision is mediated by a family of rhodopsin-like proteins encoded by the corresponding members of a family of genes, and (2) the common inherited forms of color vision deficiency ("color blindness") are caused by mutations in the members of this hypothetical gene faminly. By molecular cloning of this gene family we have proven both hypotheses and deduced from the gene sequences the amino acid sequences of the four human visual
more » ... ments (rhodopsin and the three cone pigments). These sequences show that some of the pigments differ with respect to charged amino acids that are predicted to contact 11-cis retinal. These differences are consistant with a model in which the visual pigments tune the absorption spectrum of 11-cis retinal by altering its electrostatic environment. Analysis of red and green pigment genes from different color defective individuals reveals a simple genetic mechanism responsible for their defects. Homologous but unequal recombination among members of a tandem array of red and green pigment genes produces both changes in gene number and hybrid genes. These data account for several long-standing psychophysical observations. Na channels purified from rat brain consist of three glycoprotein subunits [a (260 kD), P1 (36 kD), and 12 (33 kD)] in 1:1:1 stoichiometry in a complex of 320 kD. The P2 subunit is linked to a by disulfide bonds while 12 is noncovalendy attached. The a subunit is a transmembrane protein which binds neurotoxins on its exuacellular surface and can be phosphofylated by protein kinases on its intraceliular surface. Incorporation of the purified Na channel into phospholipid vesicles of approprate composition restores neurotoxin action at 3 receptor sites and selective ion flux. Fusion with planar bilayers and treatment with batrachotoxin results in single channels of 23 pS with the ion selectivity, toxin sensitivity, and voltage dependence of batrachotoxin-modified native Na channels. Selective removal of the 131, but not the 12, subunit from the purified complex causes loss of saxitoxin binding and neurotoxin-activated ion flux. Tetrodotoxin prevents dissociation of the 131 subunit indicating strong energy coupling between toxin binding and association of a and P1 subunits. Evidently, a complex of a and P1 subunits is necessary to retain the functional properties of purified neuronal sodium channels. cDNA's encoding the a subunit of the neuronal sodium channel have been identified, cloned, and sequenced. They recognize mRNA's of 7.3 and 8.3 kb in rat brain. As in electroplax (Noda et al, Nature 3l 121, 1984), four intemally homologous transmembrane domains are observed. Each contains a highly positively charged segment which is nearly 100% conserved from eel electroplax to rat brain (Auld et al, J. Gen. Physiol. X, 10a, 1985). A model of sodium channel structure is developed in which the transmembrane pore is formed by a square array of the 4 homologous transmembrane domains of the a subunit. The highly conserved segments of concentrated positive charge are postulated to traverse the membrane and form the voltage sensing elements of the Na channel according to a Sliding Helix model of voltage-dependent gating (Catterall, Trends Neurosci., in press). Whole cell and single-channel giga-ohm seal patch voltage clamp experiments were performed at 9.50C on myoblasts and myotubes grown from neonatal rat thigh muscle in DMEM (10% horse serum, 1% chick embryo extract). Pipettes contained (mM) 140 CsF, 10 NaCl, 5 EGTA, 10 HEPES. The bath contained 160 NaCl, 2 CaC12, 1 MgC12, 5 glucose, 10 HEPES. Whole cell recordings showed that Na channels with both high and low affinity for TTX coexisted. In the absence of TTX, the amplitude distribution of open events from single channel recordings had two peaks and was fit by maximum likelihood to the sum of two Gaussians. For test pulses to -40 mV, the population of large events had a mean open channel current PL=1.37 pA (±0.05 pA, SEM, n=8 experiments), while the population of small events had a mean, vs=0.97 pA (±0.04 pA, n=8). A plot Of IL and PS versus membrane potential (-60 to 0 mV) showed 1) no overlap Of PL and pS at any voltage, even between different experiments, and 2) the slope conductances were respectively, 12.3 pS and 8.0 pS for the large and small events, and were significantly different (p<O.Ol). In outside-out patch experiments, the observed number of large events decreased to 0 with increasing [TTX](5-156 nM), while the appearance of small events was essentially unchanged. PL and pS were not affected by TTX. [TTX] = 2-13 PM reversibly blocked all channels. The different conductances of the two types of Na channels is consistent with a different configuration of carboxyl oxygens or other charged groups near the mouths of the channels. The binding of monovalent tetrodotoxin (TTX) and divalent saxitoxin (STX) to batrachotoxinmodified sodium channels in lipid bilaTers is altered by changes in [Na+], [TEA+], and [Zn++]. The interaction between TTX binding and [Na ] could formally be described as a competitive interaction at a neutral binding site. The interaction between STX and [Na+] deviated, however, from this description; the Na+ dependence of STX binding was larger than predicted. Similarly, TEAe and Zn++ reduced the binding of STX to a much greater extent than the binding of TTX. These findings appear to result from changes in surface potential at the toxin binding site. KD(STX)/KD(TTX) vs [Na+1 was used to fit a model based on the Gouy-Chapman theory of the diffuse double layer 2nd competition between toxin and Na+. The surface charge density estimated from the fit is 1/300 Is. Zn , like Ca++ (Yamamoto et al. Biouhvs. J., 45:337, 1984), causes a voltage-dependent block of the Na current and appears to bind at an electrical distance of 0.20 -0.25 into the channel from the Deextracellular" side. In 20 mM Na+, 0.8 mM Zn++ reduced the single-channel current by 80% at -60 mV and by 30% at 60 mV. If the permeation path is close to the toxin binding site, electrostatic repulsion between Zn++ and STX should affect the voltage-dependence of STX binding. The STf dissociation rate constant was unaffected by Zn++ (-60 mV < V S 60 mV), while the association rate was decreased by a constant factor at all potentials. ZnF+ in the permeation path therefore does not interact with STX at its binding site, suggesting that the toxin binding site is separated from the permeation path by at least one Debye length (22 1 in 20 mM Na+). In order to map functionally important extracellular carboxyl groups that modulate ion permeation and tetrodotoxin (TTX) binding, batrachotoxin-modified sodium channels from canine forebrain were incorporated into planar lipid bilayers and exposed to membrane-impermeant carbodiimides (1 -10 mM of 1-cyclohexyl-3-(2-morpholinoethyl)-carbodiimide or 1-ethyl-3-(4-azonia-4,4-dimethyl-pentyl)-carbodiimide) at pH 7.4, in the absence or presence of a nucleophile. I/V characteristics and TTX block were investigated between -40 and +60 mV. In 500 mM NaCl, two discrete conductance decreases, either 11 -13% or 23 -25%, were seen after addition of carbodiimide and nucleophile. The I/V characteristics were linear before and after modification. Under similar conditions, there were no changes in the conductance of valine gramicidin A channels; the conductance decreases do not result from modification of the phospholipids. The conductance decrease in 100 mM NaCl was '40%. We have not been able to abolish TTX block concommitantly with a change in single-channel conductance. The single-channel conductance changes are qualitativoly consistent with the modification of carboxyl groups close to the channel entrance, where the conductance changes would result from a decrease in the surface charge "density". We cannot exclude, however, that we have modified carboxyl groups that are spatially separate from the permeation path and that the conductance changes result from conformational changes. The results suggest nevertheless that the ion permeation pathway and the neurotoxin binding site are spatially separate. SODIUM CHANNELS I Previous electrophysiological and planar bilayer experiments showed that p -conotoxin GIIIA, a peptide purified from Conus geographus venom, blocked Na-channels from skeletal muscle but not those of nerve or brain (L.J. Cruz et al 1985, JBC 260:9280). The effect of GMA on binding of 3H-saxitoxin (STX) to membrane preparations from various tissues was investigated to characterize the possible interaction of GMA with the external receptor site for guanidinium toxins. GIIA is a potent inhibitor of STX binding to membranes from rat skeletal muscle and eel electroplax but has no effect on STX binding to rat heart and lobster axon at 10 p M GIIA and only 20% inhibition of STX binding to rat brain at 40 P M GUIA. In eel electroplax, analysis of the displacement of STX binding by GUIA titrations and Scatchard plot analysis of STX binding suggests a purely competitive interaction with KD's of 1 nM for STX and 50 nM for GIIA at 00C and 0.2 M choline Cl. In a Ttubule preparation from rat skeletal muscle, GIIA displaced about 80% of specific STX binding, but the remaining 20% of high affinity STX sites were unaffected by GHIA. Results with rat muscle preparations are consistent with two classes of Na-channels, both having similar high affinity for STX (KD 1 nM); but one being insensitive to GHIA, while the other has a KD of about 20 nM for GIIIA. These results and studies of tetrodotoxin-insensitive Na-channels from canine heart in planar bilayers (Uehara and Moczydlowski, in preparation) can be used to classify three tissue-specific isochannels of the voltage-activated Na-channel: I. nerve and brain, II. muscle and electroplax, Im. heart-specific channels that also appear to be expressed in denervated skeletal muscle. (Suported by AHA, MDA, NIH AM35128 and Searle Scholars Program/The Chicago Community Trust.) Neosaxitoxin (neoSTX) differs from saxitoxin (STX) only in having the N-1 -H replaced by an -OH. In STX, the 7,8,9, but not the 1,2,3 guanidinium, as well as the hydrated ketone on C-12 have been found to be crucial for channel-blockade. On single frog skeletal muscle fibers voltage-clamped by the vaseline-gap method, we studied the actions of STX and neoSTX to look for a role of the 1,2,3 group. At pH's 6.50, 7,25 and 8.25, the concentrations for halving the maximum INa (ED50) are respectively 5.22, 5.07 and 8.90 nM for STX, and 1.46, 2.07 and 16.80 nM for neoSTX. The relative potencies of STX at these pH's coincide with the relative abundance of the protonated 7,8,9 guianidinium (pKa=8.25). The relative potencies of neoSTX deviate sharply from the relative abundance of the 7,8 9 group (pKa=8.65), but follow better that of the deprotonated N-1 group (pKa=6.75). In constantratio mixtures of STX and neoSTX, the two toxins compete for the same hinding site. At pH 6.50, the observed and expected ED50's are 0.73 and 0.74 nM of neoSTX, both toxins acting at full efficacy (e =1.0). At pH 8.25, the observed ED50 (6.04 nM of STX) agrees with that expected (6.05 nM) only if E of STX is 1.0 and that of neoSTX is 0.75, as if 25% of the collisions of neoSTX with the receptor fail to block the channel. In the absence of any chemical data on the ketone-ketone hydrate equilibrium for neoSTX, the differences in the potencies of STX and neoSTX could be due to the presence of an anionic site in the receptor close to the 1,2,3 group, distinct from that around the C-12 -OH's. Such a site could hydrogen-bond the protonated M-l of neoSTX, and charge-repel the deprotonated form, thus obviating a need to postulate a covalent bonding at C-12. A subclone of the Neuro-2a mouse neuroblastoma cell line, N2AB-1, was treated with tunicamycin (TM) to determine its effect on the number of functional sodium channels. Tunicamycin is an inhibitor of asparagine-linked protein glycosylation. A decrease in number of sodium channels could be expected in light of recent evidence showing that the sodium channel is a glycoprotein. Cells were grown in standard tissue culture medium at 370C. Sodium current was measured in single cells using the patch clamp whole cell recording configuration. Any possible current through K and Ca channels was blocked by using Cs+, F-, and EGTA in the patch pipet. After treatment of cells with lug/ml TM for 24 hours, the maximum inward sodium current was reduced to around one third of the value measured in untreated cells. There was no further reduction in sodium current over the succeeding 24 hours. The effect of TM was at least partially reversible. Sodium currents increased to almost 80% of control values after 24 hours of recovery from a two day treatment with TM. The voltage dependence of inactivation was not significantly altered by TM. Ensemble fluctuation analysis of the currents shows that the decrease in sodium channel current due to TM was primarily the result of a decrease in the number of channels in the cell membrane and not an alteration in single channel current. Total cell protein (Peterson's modification of the Lowry assay) measured in cells treated with TM for 24 hours was not significantly different from values obtained from untreated cells. These results indicate that TM decreased the number of functional sodium channels in the cell membrane. There is no evidence for incorporation of functional under-glycosylated channels into the surface membrane with TM treatment. SODIUM CHANNELS I 42a Trimethyloxonium (TMO) was applied to the extracellular side of single batrachotoxin-activated sodium channels from rat brain, incorporated into planar lipid bilayers. TMO, which methylates protein carboxyl resdiues, abolished block of the channels by tetrodotoxin and saxitoxin (STX), reduced the unit conductance by about one third, and greatly reduced the voltage-dependent block of the channels by extracellular calcium. These three effects of TMO always occurred concommittantly, suggesting that all three effects are consequences of a single hit by TMO. Calcium competitively inhibited block of the sodium channels by STX and the binding of 3H-STX to brain membranes. Both calcium and STX protected the channels from TMO modification of the STX binding site. These results suggest that, in the absence of STX, sodium ions interact with the negatively-charged carboxyl residue at the STX binding site as they enter the channel pore. Calcium ions also interact with this site as they enter the pore to block the inward flow of sodium. Methylation of this carboxyl group by TMO appears to slow access of both sodium and calcium to one or more sites, deeper in the pore, which determine selectivity of the channel. A single occupancy, rate-theory model for ion permeation through the channel can account, quantitatively, for these conclusions if it is assumed that TMO changes the energy profiles for sodium and calcium only by raising the levels of the outermost barrier and the adjacent well, which is located very close to the external solution. Internal Cs (Csi) as compared to Ki slows time to peak Na current, slows its decline from peak and increases steady state to peak current ratio, INa 0'/INa peak, in internally perfused Myxicola giant axons. Neither activation nor deactivation kinetics are significantly affected by Csi. INa rising phases, times to half maximum and tail current time courses are the same in Csi and Ki. Inactivation time constants (both one, Th, and two, Tcp pulse methods) are also the same in Csi and Ki. All effects of Csi on INa time course are due to an increased INa */INa peak' Csi selectively decreases steady state Na inactivation, preventing some fraction of inactivation gates from closing at all, the rest apparently closing normally. This suggests an inactivation blocking site, located internally as Cs is not permeant through the Na channel in Myicola (Ebert and Goldman, . 68:327, 1976), which when occupied prevents inactivation gate closure. Site occupancy is affected by both current magnitude and direction. INa o/INa peak in Csi increases with increasingly positive potential proportionately with the decrease in INa peak seen over this potential range. In Ki, 'Na "/IN eak is always small, generally 0 to 0.15, for inward Na current, but well larger, about 0.6 to 6.9, for outward Na channel current. Both observations argue for a location of the inactivation blocking site in the current pathway, possibly into the inner mouth of the channel. This site could mediate the normal operation of the inactivation gate. A possible mechanism for inactivation gate closure would involve a positively charged group moving to associate with a negative site at the inner channel mouth. The pyrethroid deltamethrin has been shown to prolong whole cell sodium current of neuroblastoma NlE-115 cells due to an increase in sodium channel open time (Chinn, K. and Narahashi, T., Neurosci. Abstr., 11: 784, 1985). As expected from whole cell data, in the presence of deltamethrin (10 tM) there were modified channels whose open times were greatly prolonged as well as channels having short open times which mav have derived from unmodified channels. Also as expected, modified channels open at the end of a pulse remained open for up to 30 s after repolarization. Single channel conductance was not affected by deltamethrin. Several unexpected properties of modified channels were also revealed. Deltamethrin reduced the number of active channels. In addition, a subconducting state and a state of persistent flickering were observed in the presence of the drug. Finally, even in the absence of channel openings during a depolarizing step, some channels opened after repolarization. No spontaneously active channels were detected at the holding potential. These observations are compatible with the hypothesis that deltamethrin stabilizes many different states of the sodium channel, besides an open state. To discriminate specific roles of activation and inactivation gates in the use-dependent block of Na channels produced by local anesthetics, we compared the QX-314 block of Na currents before and after removal of Na inactivation in internally perfused and voltage clamped squid axons. Pronase and chloramine-T were used to remove Na inactivation because they had little effect on the activation process. The use-dependent block (UDB) was characterized by its voltage dependence. of the pathway of nucleotide triphosphate hydrolysis. This work was supported by AM31986 (Ivan Rayment)Profilin is a 12-16 kDa protein that inhibits nucleation and suppresses elongation of actin polymers by seauestering G-actin. We purified profilin from the Schneider 3 Drosophila cell line by chromotography on DEAE, hydroxylapatite, and Sephadex G-75. At a 1:1 molar ratio of profilin to actin, turbidity assays demonstrate that this 16 kDa polypeptide increases the lag phase of actin polymerization by -3 fold and decreases its maximum rate by,4 fold in 2mM MgCl , 0.2mM ATP, 5mM KPO pH7.5, 0.5mM DTT. Preliminary results with rabbit antibodies to D. melano-
doi:10.1016/s0006-3495(86)83674-8 fatcat:eycpjiszenaqbbnnf6f55ow23e