| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||
PERSPECTIVES |
1 Department of Neuroscience, Johns Hopkins School of Medicine, Baltimore, MD 21205, USA Email:jreisert{at}jhmi.edu
2 Laboratoire de Physiologie Cérébrale, Université Paris 5, CNRS UMR 8118, 75006 Paris, France Email:jonathan.bradley{at}university-paris5.fr
Ion channels that are gated by cyclic nucleotides (CNG channels) have been best studied for their roles in visual and olfactory sensory transduction. Sensory stimulation of vertebrate photoreceptors by light and olfactory receptor neurones (ORNs) by odorants activates 7-transmembrane receptors, and triggers, via a G protein, changes in the intracellular concentration of second messenger cyclic nucleotides. In the case of photoreceptors, the G protein Gt stimulates phosphodiesterase leading to the hydrolysis of cGMP, which in turn promotes closure of cGMP-gated channels and hyperpolarization. In the case of ORNs, the G protein Golf stimulates adenylyl cyclase to produce cAMP, which promotes opening of cAMP-gated channels and depolarization. The ability of CNG channels to be gated by cyclic nucleotides is therefore crucial for organisms to monitor their visual and odorant environments.
In native cells CNG channels are heteromeric tetramers, comprising subunits of two phylogenetically distinct subfamilies, CNGA and CNGB, in different stoichiometries, (for review see Kaupp & Seifert, 2002; Bradley et al. 2005). In heterologous expression systems (HEK cells or Xenopus oocytes) CNGA1, 2 or 3 can be expressed as homomeric tetramers, but the channels of these tetramers differ considerably from those of heteromers (e.g. EC50 for cyclic nucleotides, ion permeation and selectivity, and modulation). All together, CNG channels contain as many as four binding sites for cyclic nucleotide. However, the mechanism of gating by cyclic nucleotide is not understood.
What is the relationship between ligand binding and channel gating, and what roles do the individual subunits play in this process? These questions have been addressed over the years and several constantly evolving models have been proposed (for review see Li et al. 1997; Kaupp & Seifert, 2002). In a linear model, cyclic nucleotides bind sequentially, and the channel opens only once it is fully liganded. This model emerged from studying CNG channels in inside-out membrane patches from rod photoreceptors, under non-equilibrium conditions, using rapid application of cGMP (achieved by photo-liberating cGMP from caged cGMP) (Karpen et al. 1988). Subsequently it was found that CNG channels can open (with low open probability) in the absence of ligand (Picones & Korenbrot, 1995; Ruiz & Karpen, 1997; Tibbs et al. 1997; Kleene, 2000), which is inconsistent with the linear model. The cyclic allosteric model (MWC) (Monod et al. 1965), where the channel can open whether or not it is partly or fully liganded, addressed this issue, but failed to predict the experimentally observed open probabilities for partially liganded CNG channels (Ruiz & Karpen, 1997). Liu et al. (1998) then proposed a coupled dimer (CD) model where pairs of CNG channel subunits functionally couple within the channel protein. The CD and MWC models differ significantly in the factor by which the binding of cyclic nucleotide increases channel open probability.
In the current issue of The Journal of Physiology, Nache et al. (2005) revisit these questions, expressing CNGA2 homomers or CNGA2/A4/B1b heteromers heterologously in Xenopus oocytes, and propose another model. In this elegant study the authors use recently available versions of caged cAMP and cGMP (Hagen et al. 2001) which are resistant to hydrolysis in aqueous solution and, importantly, allow release of high concentrations of cyclic nucleotides so that one can cover the whole range of cyclic nucleotide sensitivity. Thus, recording from the activity of many channels simultaneously, they could study the activation kinetics of the channels in unprecedented detail. A main finding of Nache et al. (2005) is that there is not a monotonic acceleration of activation in the direction of higher agonist concentration, as would be predicted if agonist binds to equivalent sites and the binding limits the activation. Instead, they find CNG channels open more quickly at low and at high compared to intermediate agonist concentrations. Also, the current was found to rise bi-exponentially, which could be matched neither by the MWC nor CD models, which both predict a more sigmoidal rise in current. In addition, Nache et al. (2005) cleverly exploited the difference in cAMP and cGMP sensitivity of the expressed CNG channels to address which step (binding or the conformational change) is rate limiting for channel activation. They found that, for a given open probability, whether activated by cAMP or cGMP, both agonists have similar profiles of activation kinetics, suggesting that the conformational change is the rate-limiting step. Furthermore, channels that are CNGA2 homotetramers, or CNGA2/A4/B1b heterotetramers, display similar activation kinetics for a given open probability, hinting that the auxiliary subunits contribute similarly to channel activation.
The authors tested a series of models and concluded that one with three binding steps, where the ligand can only bind to the closed channel, best described their data. To use the authors' words, the conclusion is ‘remarkably simple’. Of the three binding steps, the first and third have a high ligand affinity, while the second binding step has a low affinity, but only this step switches the channels' open probability from low to high.
With the new experimental tools put in place by Nache et al. (2005) to study the kinetics of CNG channel activation it will be interesting to see if the olfactory CNG channel is unique in its activation mechanism, considering it consists of three not two different subunits, as do rod and cone CNG channels. Also, physiologically, olfactory transduction, which leads to opening of CNG channels, may require a higher cyclic nucleotide concentration than visual transduction, which leads to closing of the CNG channels. In addition in visual transduction, where cGMP concentrations are always quite low, a model of single channel activity that includes channel substates might be needed (Ruiz & Karpen, 1997). Without doubt, however, the modulatory effects of point mutations, phosphorylation, and desensitization by Ca2+–calmodulin can now be addressed in more kinetic detail.
References
Bradley J et al. (2005). Curr Opin Neurobiol 15, 343–349.[CrossRef][Medline]
Hagen V et al. (2001). Angew Chem Int Ed Engl 40, 1045–1048.[CrossRef]
Karpen JW et al. (1988). Cold Spring Harb Symp Quant Biol 53, 325–332.[Medline]
Kaupp
UB
&
Seifert
R (2002). Physiol Rev
82, 769–824.
Kleene SJ (2000). J Membr Biol 178, 49–54.[CrossRef][Medline]
Li J et al. (1997). Q Rev Biophys 30, 177–193.[CrossRef][Medline]
Liu DT et al. (1998). Neuron 21, 235–248.[CrossRef][Medline]
Monod J et al. (1965). J Mol Biol 12, 88–118.[Medline]
Nache
V
et al. (2005). J Physiol
569, 91–102.
Picones A & Korenbrot JI (1995). J Physiol 485, 699–714.[Medline]
Ruiz ML & Karpen JW (1997). Nature 389, 389–392.[CrossRef][Medline]
Tibbs GR et al. (1997). Nature 386, 612–615.[CrossRef][Medline]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
