A novel cyclic nucleotide-gated ion channel enriched in synaptic terminals of isotocin neurons in zebrafish brain and pituitary.

Cyclic nucleotide-gated (CNG) channels are nonselective cation channels opened by binding of intracellular cyclic GMP or cyclic AMP. CNG channels mediate sensory transduction in the rods and cones of the retina and in olfactory sensory neurons, but in addition, CNG channels are also expressed elsewhere in the CNS, where their physiological roles have not yet been well defined. Besides the CNG channel subtypes that mediate vision and olfaction, zebrafish has an additional subtype, CNGA5, which is expressed almost exclusively in the brain. We have generated CNGA5-specific monoclonal antibodies, which we use here to show that immunoreactivity for CNGA5 channels is highly enriched in synaptic terminals of a discrete set of neurons that project to a subregion of the pituitary, as well as diffusely in the brain and spinal cord. Double labeling with a variety of antibodies against pituitary hormones revealed that CNGA5 is located in the terminals of neuroendocrine cells that secrete the nonapeptide hormone/transmitter isotocin in the neurohypophysis, brain, and spinal cord. Furthermore, we show that CNGA5 channels expressed in Xenopus oocytes are highly permeable to Ca(2+), which suggests that the channels are capable of modulating isotocin release in the zebrafish brain and pituitary. Isotocin is the teleost homolog of the mammalian hormone oxytocin, and like oxytocin, it regulates reproductive and social behavior. Therefore, the high calcium permeability of CNGA5 channels and their strategic location in isotocin-secreting synaptic terminals suggest that activation of CNGA5 channels in response to cyclic nucleotide signaling may have wide-ranging neuroendocrine and behavioral effects.

Beta-ionone activates and bleaches visual pigment in salamander photoreceptors.

Vision begins with photoisomerization of 11-cis retinal to the all-trans conformation within the chromophore-binding pocket of opsin, leading to activation of a biochemical cascade. Release of all-trans retinal from the binding pocket curtails but does not fully quench the ability of opsin to activate transducin. All-trans retinal and some other analogs, such as beta-ionone, enhance opsin's activity, presumably on binding the empty chromophore-binding pocket. By recording from isolated salamander photoreceptors and from patches of rod outer segment membrane, we now show that high concentrations of beta-ionone suppressed circulating current in dark-adapted green-sensitive rods by inhibiting the cyclic nucleotide-gated channels. There were also decreases in circulating current and flash sensitivity, and accelerated flash response kinetics in dark-adapted blue-sensitive (BS) rods and cones, and in ultraviolet-sensitive cones, at concentrations too low to inhibit the channels. These effects persisted in BS rods even after incubation with 9-cis retinal to ensure complete regeneration of their visual pigment. After long exposures to high concentrations of beta-ionone, recovery was incomplete unless 9-cis retinal was given, indicating that visual pigment had been bleached. Therefore, we propose that beta-ionone activates and bleaches some types of visual pigments, mimicking the effects of light.

Mutation of a single residue in the S2-S3 loop of CNG channels alters the gating properties and sensitivity to inhibitors.

We previously found that native cyclic nucleotide-gated (CNG) cation channels from amphibian rod cells are directly and reversibly inhibited by analogues of diacylglycerol (DAG), but little is known about the mechanism of this inhibition. We recently determined that, at saturating cGMP concentrations, DAG completely inhibits cloned bovine rod (Brod) CNG channels while only partially inhibiting cloned rat olfactory (Rolf) channels (Crary, J.I., D.M. Dean, W. Nguitragool, P.T. Kurshan, and A.L. Zimmerman. 2000. J. Gen. Phys. 116:755-768; in this issue). Here, we report that a point mutation at position 204 in the S2-S3 loop of Rolf and a mouse CNG channel (Molf) found in olfactory epithelium and heart, increased DAG sensitivity to that of the Brod channel. Mutation of this residue from the wild-type glycine to a glutamate (Molf G204E) or aspartate (Molf G204D) gave dramatic increases in DAG sensitivity without changing the apparent cGMP or cAMP affinities or efficacies. However, unlike the wild-type olfactory channels, these mutants demonstrated voltage-dependent gating with obvious activation and deactivation kinetics. Interestingly, the mutants were also more sensitive to inhibition by the local anesthetic, tetracaine. Replacement of the position 204 glycine with a tryptophan residue (Rolf G204W) not only gave voltage-dependent gating and an increased sensitivity to DAG and tetracaine, but also showed reduced apparent agonist affinity and cAMP efficacy. Sequence comparisons show that the glycine at position 204 in the S2-S3 loop is highly conserved, and our findings indicate that its alteration can have critical consequences for channel gating and inhibition.

Mechanism of inhibition of cyclic nucleotide-gated ion channels by diacylglycerol.

Cyclic nucleotide-gated (CNG) channels are critical components in the visual and olfactory signal transduction pathways, and they primarily gate in response to changes in the cytoplasmic concentration of cyclic nucleotides. We previously found that the ability of the native rod CNG channel to be opened by cGMP was markedly inhibited by analogues of diacylglycerol (DAG) without a phosphorylation reaction (Gordon, S.E., J. Downing-Park, B. Tam, and A.L. Zimmerman. 1995. Biophys. J. 69:409-417). Here, we have studied cloned bovine rod and rat olfactory CNG channels expressed in Xenopus oocytes, and have determined that they are differentially inhibited by DAG. At saturating [cGMP], DAG inhibition of homomultimeric (alpha subunit only) rod channels was similar to that of the native rod CNG channel, but DAG was much less effective at inhibiting the homomultimeric olfactory channel, producing only partial inhibition even at high [DAG]. However, at low open probability (P(o)), both channels were more sensitive to DAG, suggesting that DAG is a closed state inhibitor. The Hill coefficients for DAG inhibition were often greater than one, suggesting that more than one DAG molecule is required for effective inhibition of a channel. In single-channel recordings, DAG decreased the P(o) but not the single-channel conductance. Results with chimeras of rod and olfactory channels suggest that the differences in DAG inhibition correlate more with differences in the transmembrane segments and their attached loops than with differences in the amino and carboxyl termini. Our results are consistent with a model in which multiple DAG molecules stabilize the closed state(s) of a CNG channel by binding directly to the channel and/or by altering bilayer-channel interactions. We speculate that if DAG interacts directly with the channel, it may insert into a putative hydrophobic crevice among the transmembrane domains of each subunit or at the hydrophobic interface between the channel and the bilayer.

Perfusion system components release agents that distort functional properties of rod cyclic nucleotide-gated ion channels.

In switching from studying native cyclic nucleotide-gated (CNG) ion channels in rod cells to studying the corresponding cloned channels expressed in Xenopus oocytes, we changed our perfusion system to a more efficient one. This change involved replacing culture flasks and a small plexiglass/glass chamber with plastic syringes, metal needles, and plastic petri dishes. We now report that these new perfusion system components release agents that distort or obscure measured functional properties of rod CNG channels. The magnitude and time course of appearance of the artifacts vary widely among individual components (e.g. from syringe to syringe). The effects most resemble voltage-dependent block of the channels, giving a decrease in current at positive potentials, and producing distortions of the kinetics and voltage dependence of channel activation.

Diacylglycerol analogs inhibit the rod cGMP-gated channel by a phosphorylation-independent mechanism.

The electrical response to light in retinal rods is mediated by cyclic nucleotide-gated, nonselective cation channels in the outer segment plasma membrane. Although cGMP appears to be the primary light-regulated second messenger, cellular levels of other substances, including Ca2+ and phosphatidylinositol-4,5-bisphosphate, are also sensitive to the level of illumination. We now show that diacylglycerol (DAG) analogs reversibly suppress the cGMP-activated conductance in excised patches from frog rod outer segments. This suppression did not require nucleoside triphosphates, indicating that a phosphorylation reaction was not involved. DAG was more effective at low than at high [cGMP]: with 50 microM 8-Br-cGMP, the DAG analog 1,2-dioctanoyl-sn-glycerol (1,2-DiC8) reduced the current with an IC50 of approximately 22 microM (Hill coefficient, 0.8), whereas with 1.2 microM 8-Br-cGMP, only approximately 1 microM 1,2-DiC8 was required to halve the current. DAG reduced the apparent affinity of the channels for cGMP: 4 microM 1,2-DiC8 produced a threefold increase in the K1/2 for channel activation by 8-Br-cGMP, as well as a threefold reduction in the maximum current, without changing the apparent stoichiometry or cooperativity of cGMP binding. Inhibition by 1,2-DiC8 was not relieved by supersaturating concentrations of 8-Br-cGMP, suggesting that DAG did not act by competitive inhibition of cGMP binding. Furthermore, DAG did not seem to significantly reduce single-channel conductance. A DAG analog similar to 1,2-DiC8--1,3-dioctanoyl-sn-glycerol (1,3-DiC8)--suppressed the current with the same potency as 1,2-DiC8, whereas an ethylene glycol of identical chain length (DiC8-EG) was much less effective. Our results suggest that DAG allosterically interferes with channel opening, and raise the question of whether DAG is involved in visual transduction.

Modulation of the cGMP-gated ion channel in frog rods by calmodulin and an endogenous inhibitory factor.

1. Outer segment patches excised in the light were used to investigate the effects of exogenous calmodulin and an endogenous inhibitory factor on the cGMP-gated channel of frog rods. 2. Calmodulin shifted to the right the dose-response relation for activation of the channels by 8-Br-cGMP, but did not change the maximum current or the form of the relation. Reversal of this effect by removal of calmodulin was accelerated by brief exposure to saturating [8-Br-cGMP]. Inhibition by calmodulin required calcium and gave as much as a 5-fold decrease in current for an [8-Br-cGMP] functionally comparable to the presumed physiological [cGMP]. 3. Exposure to low [Ca2+]i (tens of nanomolar) appeared to irreversibly remove or inactivate an endogenous channel inhibitory factor from the patches, increasing the current at low [8-Br-cGMP]. Like calmodulin, this factor slowed the voltage-dependent channel-gating kinetics and did not change the maximum current. However, unlike calmodulin, the endogenous factor remained stably associated with the patches at high [Ca2+]i (1 microM), even with exposure to saturating [8-Br-cGMP]. 4. After the low-Ca2+ treatment increased the current, calmodulin reduced the current to about the same level as it had before the low-Ca2+ treatment, giving a larger fractional suppression. Furthermore, patches with high initial sensitivity to 8-Br-cGMP had small low-Ca2+ effects and large calmodulin effects, while the reverse was true for patches with low initial agonist sensitivity. 5. Application of trypsin to the intracellular surface of the patch prevented the responses to calmodulin and to low [Ca2+]i, suggesting involvement of a cytoplasmic portion of the channel. However, trypsin also reduced the total agonist-induced patch current. 6. Our results are consistent with a model in which calmodulin and an endogenous calcium-binding protein compete for the same site, inhibiting channel opening or cGMP binding. The tight association of the endogenous factor with the channel even at relatively low [Ca2+]i suggests that in the transducing rod it may inhibit the channels most of the time in darkness and in dim light, preventing any potential inhibitory effects of calmodulin. The endogenous factor would be expected to leave the channel only in bright or prolonged light, when the [Ca2+]i is thought to be very low.

Cyclic nucleotide gated channels.

Recent studies have revealed that cyclic nucleotide gated channels have a variety of forms and functions. These channels are now thought to be heteromultimers of at least two kinds of subunits and to undergo functional modulation. Ion permeation involves at least two ion-binding sites, and recent work on the alpha subunit suggests that many structural regions are involved in the control of channel gating. The continued use of both molecular and physiological approaches promises to further our understanding of how these channels work and how they are involved in cellular function.

Heat-shock 70 messenger RNA levels in human brain: correlation with agonal fever.

Systematic review of antemortem clinical information on randomly selected Alzheimer disease (AD) patients revealed that approximately 40% of the patients had a recorded fever of > or = 39.2 degrees C at or near death. Using isolation and quantitation techniques appropriate for analysis of human brain mRNAs, we found that low levels of inducible heat-shock protein 70 (hsp70) mRNAs were present in cerebellum of afebrile AD patients and that mRNA levels were usually lower in two brain regions affected in AD, i.e., hippocampus and temporal cortex. Levels of hsp70 mRNAs were increased three- to 33-fold in cerebellum of febrile patients compared with levels in patients whose recorded temperatures were < or = 37.5 degrees C. Levels of hsp70 mRNAs were also increased in hippocampus and cortex of these febrile patients, but to a lesser extent than cerebellum. Heat-shock cognate 70 (hsc70) mRNAs were present at highest levels in afebrile cerebellum and were also present in the other brain regions. In cerebellum of patients with the highest temperatures, hsc70 mRNAs were induced severalfold over basal levels. Although there was a low and variable induction of hsc70 mRNAs in temporal cortex of these patients, there was no evidence for any induction in hippocampus. Increased heat-shock 70 mRNA levels did not correlate with hypoxia, coma, hypertension, hypoglycemia, seizures, or medication. These results indicate that a specific agonal stress, namely fever, can increase the levels of heat shock 70 mRNAs in AD brain; however, there is no evidence to suggest that affected regions of AD brain have higher overall levels of these mRNAs. Failure to obtain adequate agonal state information could result in inaccurately identifying short-term stress-related changes in postmortem brain as neuropathology characteristic of a chronic disease state.

Selective postmortem degradation of inducible heat shock protein 70 (hsp70) mRNAs in rat brain.

1. Altered mRNA levels in postmortem brain tissue from persons with Alzheimer's disease (AD) or other neurological diseases are usually presumed to be characteristic of the disease state, even though both agonal state (the physiological state immediately premortem) and postmortem interval (PMI) (the time between death and harvesting the tissue) have the potential to affect levels of mRNAs measured in postmortem tissue. Although the possible effect of postmortem interval on mRNA levels has been more carefully evaluated than that of agonal state, many studies assume that all mRNAs have similar rates of degradation postmortem. 2. To determine the postmortem stability of inducible heat shock protein 70 (hsp70) mRNAs, themselves unstable in vivo at normal body temperature, rats were heat shocked in order to induce synthesis of the hsp70 mRNAs. hsp70 mRNA levels in cerebellum and cortex were then compared to those of their heat shock cognate 70 (hsc70) mRNAs, as well as to levels of 18S rRNAs, at 0 and at 24 hr postmortem. 3. Quantiation of northern blots after hybridization with an hsp70 mRNA-specific oligo probe indicated a massive loss of hsp70 mRNA signal in RNAs isolated from 24-hr postmortem brains; quantitation by slot-blot hybridization was 5- to 15-fold more efficient. Even using the latter technique, hsp70 mRNA levels were reduced by 59% in 24-hr-postmortem cerebellum and by 78% in cortex compared to mRNA levels in the same region of 0-hr-postmortem brain. There was little reduction postmortem in levels of the hsp70 mRNAs or of 18S rRNAs in either brain region. 4. In situ hybridization analysis indicated that hsp70 mRNAs were less abundant in all major classes of cerebellar cells after 24 hr postmortem and mRNAs had degraded severalfold more rapidly in neurons than in glia. There was no corresponding loss of intracellular 18S rRNA in any cell type. 5. We conclude from these results that the effect of postmortem interval on mRNA degradation must be carefully evaluated when analyzing levels of inducible hsp70 mRNAs, and perhaps other short-lived mRNAs, in human brain.

Protein phosphatases modulate the apparent agonist affinity of the light-regulated ion channel in retinal rods.

Ion channels directly activated by cGMP mediate the light response in retinal rods. Several components of the enzyme cascade controlling cGMP concentration are regulated, but there are no accepted mechanisms for modulation of the response of the channel to cGMP. Here we report evidence that in excised patches an endogenous protein phosphatase converts the channel from a state with low cGMP sensitivity to a state with almost 3 orders of magnitude higher sensitivity in the predicted physiological range of cGMP concentration. The action of this endogenous phosphatase was blocked by specific serine/threonine phosphatase inhibitors (microcystin-LR, okadaic acid, and calyculin A). An increase in apparent agonist affinity also was produced by addition of purified protein phosphatase 1. In contrast, protein phosphatase 2A decreased apparent agonist affinity, suggesting that two phosphorylation sites may regulate the agonist sensitivity of the channel in a reciprocal manner. This regulation may be involved in fine-tuning the light response or in light or dark adaptation.

Cation interactions within the cyclic GMP-activated channel of retinal rods from the tiger salamander.

1. The ionic dependence of current through the 3',5'-cyclic guanosine monophosphate (cyclic GMP)-activated channels of salamander rods was studied in excised inside-out membrane patches from isolated outer segments. Voltage-clamp experiments on transducing rods were performed so that the channels in intact cells could be compared with those in excised patches. 2. The reversal potential of the cyclic GMP-induced patch current was close to the Na+ equilibrium potential when the concentration of NaCl on the cytoplasmic surface of a patch was varied at constant external NaCl concentration. Fitting the Goldman-Hodgkin-Katz equation indicated that the apparent ratio of permeabilities for Na+ and Cl- was at least 50. This confirms a previous report that the channel's Na+ permeability is much larger than its Cl- permeability. 3. Na+ currents through the channel did not obey the independence principle. The outward patch current at large positive potential began to saturate with increasing concentrations of internal Na+, as if permeation required Na+ to bind to a site with an apparent dissociation constant around 180 mM. 4. In symmetrical NaCl solutions containing very low concentrations of divalent cations the current-voltage relation measured from excised patches 50 microseconds after switching the voltage showed mild outward rectification. By 1 ms the rectification was more pronounced. The rectification at 50 microseconds is attributed to voltage dependence of Na+ permeation. The additional rectification at later times is attributed to voltage dependence of the channel's probability of being open, depolarization favouring the open state. 5. In symmetrical Mg2+ solutions the cyclic GMP-induced patch currents were smaller and the outward rectification was more pronounced. 6. Addition of Mg2+ or Ca2+ to an internal Na+ solution blocked the cyclic GMP-induced Na+ current through the channels, as if by occupying a single binding site with an affinity in the 0.1-2 mM range. Block by Mg2+ was voltage dependent, suggesting that the binding site was within the channel's transmembrane electric field. Raising the Mg2+ concentration on the external surface of the patch increased the apparent dissociation constant of block by internal Mg2+, as expected if external and internal Mg2+ compete for the same binding site. 7. Block by internal Ca2+ had an opposite and weaker voltage dependence than block by internal Mg2+. 8. In symmetrical solutions containing both Na+ and Mg2+ the outward rectification was more pronounced than in solutions containing Na+ alone. In solutions thought to be close to physiological the outward patch current increased e-fold for a depolarization of 24-30 mV.(ABSTRACT TRUNCATED AT 400 WORDS)

Hindered diffusion in excised membrane patches from retinal rod outer segments.

Excised inside-out membrane patches are useful for studying the cGMP-activated ion channels that generate the electrical response to light in retinal rod cells. We show that strong ionic current across a patch changes the driving force on the current by altering the ionic concentration near the surface membrane, an effect somewhat like that first described by Frankenhaeuser and Hodgkin (1956) in squid axons. The dominant concentration change occurs in the solution adjacent to the cytoplasmic (inner) surface of the membrane, where diffusion is impaired by intracellular material that adheres to the patch during excision. The magnitude and time course of the ionic changes are consistent with the expected volume of this material and with an effective diffusion coefficient about an order of magnitude less than that in free solution. Methods are described for correcting current transients observed in voltage clamp experiments, so that channel gating kinetics can be obtained without contamination by changes in driving force. We suggest that restricted diffusion may occur in patches excised from other types of cells and influence rapid kinetic measurements.

Gating kinetics of the cyclic-GMP-activated channel of retinal rods: flash photolysis and voltage-jump studies.

The gating kinetics of the cGMP-activated cation channel of salamander retinal rods have been studied in excised membrane patches. Relaxations in patch current were observed after two kinds of perturbation: (i) fast jumps of cGMP concentration, generated by laser flash photolysis of a cGMP ester ("caged" cGMP), and (ii) membrane voltage jumps, which perturb activation of the channel by cGMP. In both methods the speed of activation increased with the final cGMP concentration. The results are explained by a simple kinetic model in which activation involves three sequential cGMP binding steps with bimolecular rate constants close to the diffusion-controlled limit; fully liganded channels undergo rapid open-closed transitions. Voltage perturbs activation by changing the rate constant for channel closing, which increases with hyperpolarization. Intramolecular transitions of the fully liganded channel limit the kinetics of activation at high cGMP concentrations (greater than 50 microM), whereas at physiological cGMP concentrations (less than 5 microM), the kinetics of activation are limited by the third cGMP binding step. The channel appears to be optimized for rapid responses to changes in cytoplasmic cGMP concentration.

Cyclic GMP-sensitive conductance of retinal rods consists of aqueous pores.

The surface membrane of retinal rod and cone outer segments contains a cation-selective conductance which is activated by 3',5'-cyclic guanosine monophosphate (cGMP). Reduction of this conductance by a light-induced decrease in the cytoplasmic concentration of cGMP appears to generate the electrical response to light, but little is known about the molecular nature of the conductance. The estimated unitary conductance is so small that ion transport might occur via either a carrier or a pore mechanism. Here we report recordings of cGMP-activated single-channel currents from excised rod outer segment patches bathed in solutions low in divalent cations. Two elementary conductances, of approximately 24 and 8 pS, were observed. These conductances are too large to be accounted for by carrier transport, indicating that the cGMP-activated conductance consists of aqueous pores. The dependence of the channel activation on the concentration of cGMP suggests that opening of the pore is triggered by cooperative binding of at least three cGMP molecules.

Interaction of hydrolysis-resistant analogs of cyclic GMP with the phosphodiesterase and light-sensitive channel of retinal rod outer segments.

cGMP opens cation-selective channels when applied to the cytoplasmic side of excised patches of membrane from retinal rod outer segments (ROS). If the light-sensitive channel in intact rods is gated only by cGMP, it should be possible to find a hydrolysis-resistant analog of cGMP that blocks the normal response to light by holding the channel open independent of the degree of illumination. We have studied the interaction of 8-bromo-cGMP (8-Br-cGMP) and the SP and RP phosphorothioate derivatives of cGMP [(Sp)-cGMP[S] and (RP)-cGMP[S]) with the cGMP phosphodiesterase (PDEase) of ROS, the cGMP-sensitive channel of excised ROS patches, and the light-sensitive channel of intact rods. All three analogs were hydrolyzed by PDEase much more slowly than was cGMP. The maximal rates of hydrolysis of 8-Br-cGMP, (SP)-cGMP[S], and (RP)-cGMP[S] were 7.3, 3.7, and less than 0.2 s-1, respectively, compared with 4000 s-1 for cGMP. These analogs are effective competitive inhibitors of the PDEase, with Ki values of 48, 25, and 90 microM, respectively. The nucleotide-activated conductances of excised patches were half-maximal at concentrations of 1.6, 210, and 1200 microM, respectively, compared with 17 microM for cGMP. Thus, 8-Br-cGMP is a highly potent channel agonist. The effects of these analogs on the dark current and photoresponses of intact rod cells were also measured. A suction electrode monitored membrane current across the ROS, while a patch electrode sealed on the inner segment was used to introduce a cGMP analog and to control membrane potential. All three analogs increased the dark current and markedly slowed the response to light flashes. 8-Br-cGMP increased the dark current of the outer segment as much as 48-fold. After the concentration of this analog had risen sufficiently, little of the current could be shut off by light, as expected of a direct effect on the light-sensitive channel of the plasma membrane. These results are consistent with the notions that (i) the light-sensitive channel of rods is controlled solely by the instantaneous concentration of cGMP and (ii) the cGMP-sensitive channel of excised patches is identical to the light-sensitive channel of intact rods.

Permeability properties of cell-to-cell channels: kinetics of fluorescent tracer diffusion through a cell junction.

We have analyzed the intracellular and cell-to-cell diffusion kinetics of fluorescent tracers in the Chironomus salivary gland. We use this analysis to investigate whether membrane potential-induced changes in junctional permeability are accompanied by changes in cell-to-cell channel selectivity. Tracers of different size and fluorescence wavelength were coinjected into a cell, and the fluorescence was monitored in this cell and an adjacent one. Rate constants, kj, for cell-to-cell diffusion were derived by compartment model analysis, taking into account (i) cell-to-cell diffusion of the tracers; (ii) their loss from the cells; (iii) their binding (sequestration) to cytoplasmic components; and (iv) their relative mobility to cytoplasm, as determined separately on isolated cells. In cell pairs, we compared a tracer's kj with the electrical cell-to-cell conductance, gj. At cell membrane resting potential, the kj's ranged 3.8-9.2 X 10(-3) sec-1 for the small carboxyfluorescein (mol wt 376) to about 0.4 X 10(-3) sec-1 for a large fluorescein-labeled sugar (mol wt 2327). Cell membrane depolarization reversibly reduced gj and kj for a large and a small tracer, all in the same proportion. This suggests that membrane potential controls the number of open channels, rather than their effective pore diameter or selectivity. From the inverse relation between tracer mean diameter and relative kj we calculate an effective, permeation-limiting diameter of approximately 29 A for the insect cell-to-cell channel. Intracellular diffusion was faster than cell-to-cell diffusion, and it was not solely dependent on tracer size. Rate constants for intracellular sequestration and loss through nonjunctional membrane were large enough to become rate-limiting for cell-to-cell tracer diffusion at low junctional permeabilities.