This review focuses on the methods used to study ion channels, but resembles more closely a work in progress, because the various methods for examining the structure and function of ion channels are still in the developmental stage.   Often, established methods are applied to the study of ion channels, but they also have been developed specifically for them.   One of the first methods "borrowed" from the already established field of protein chemistry and applied to ion channel research was called "limited proteolysis" (in the late 1960s).   Limited proteolysis was also one of the first methods to definitively show that ion channels were constructed of proteins.   Unfortunately, many other established protein chemistry methods, unlike limited proteolysis, have not been so easy to apply, due in large part by the simple fact that ion channels by their very nature are integral membrane proteins; meaning they sit in a membrane in order to function normally.    This necessity for being in a membrane introduces many fundamental problems into a potential experiment.   For example, when using light spectroscopy-based methods, the problem of noise due to light scattering can become dominant over signal because of the size of the membrane.   Detergents are often necessary for the solubilization and purification of relatively hydrophobic ion channels from membranes.   This hydrophobicity of ion channels often needs to be taken into account when using techniques borrowed from molecular biology such as SDS-PAGE, because hydrophobic proteins may tend to run differently on a protein gel compared to regular proteins.     Perhaps most troubling of all concerning the study of ion channels is the often limited availability of the ion channels themselves.  It is therefore not surprising that the first ion channels to be studied in purified form, the nAChR ion channels, were available in relatively large amounts and found in easily accessible tissue sources compared to many other interesting, but hard to obtain ion channels (like CFTR).    

I have divided up this methods review into three main sections, even though not all techniques fit so neatly into a single one.    The first section is called Electronics and Pharmacology Techniques and involves taking advantage of the fact that ion channels transmit a "current" of ions which can be detected with highly sensitive instruments.   A current the nature of which can often be probed by the use of various toxins and drugs.    The second section is based on the use of light waves to probe the macromolecular nature of ion channels and is therefore called Spectroscopic and Crystallographic Techniques.   These methods are often useful for gaining structural information about ion channel proteins.   It is important to know the correct structure of a protein for several reasons, perhaps the most important of which is that the structure can then be used as a "framework" on which one can place information gained from other techniques.   Without enough valid structural information, all data gleaned from traditional protein chemistry techniques can become simply conjecture.     The third major category concerns the methods borrowed from Molecular Biology.    In many ways, these techniques were developed not only as a way to discover new ion channels but also to provide ion channel material in larger quantities, making them more convenient to work with.     And finally, the last category is a compilation, or listing,  of techniques that have been recently used by researchers and presented in various journals or at scientific meetings.      


Some Background: The galvanometer was invented around 1820 by Schweigger, and was a necessary instrument for detecting currents in Galvanii's famous twitching frog leg experiment.  Without a galvanometer, no fundamental understanding into the mechanisms of electrical phenomena in animal tissue could ever have been uncovered.   The invention also coincided with the discovery by Faraday of electric and magnetic forces being tied together.  "...although it could not detect bioelectrical currents, it led the way to a succession of contrivances for the detection and recording of them, invented over the next century and a half, and ranging from the astatic galvanometer, to the reflecting-mirror galvanometer, the capillary electrometer, the string galvanometer, the cathode-ray oscilloscope, and eventually to the sophisticated apparatus used today."  from the book entitled Nineteenth-Century Origins of Neuroscience Concepts  by Clarke.    Nobili in 1825 devised an ingenious instrument in which he made use of the astatic needle that had been invented in 1820 by Ampere.   It neutralized, or greatly reduced the background effects of terrestrial magnetism, and with the addition of his own modifications, he constructed an improved Schweigger astatic galvanometer that provided a sensitivity denied earlier workers.  (note:  "although early multipliers, or galvanoscopes, could detect the presence of a current, they still could not measure it" as they were too insensitive for quantitation)   Nobili's instrument was still too insensitive to register the presence of the brief and minute electrical signal produced in nerve conduction.   But later he would detect the presence in the frog of an electric current in 1827.     Matteucci used the "rheoscopic frog" to detect currents from the membranes of injured muscle cells.   This current detector consisted of a frog's leg in a jar.   The nerve in the leg acted as a simple readout device.   In the period of approximately 150 years, from 1800s to 1970s, electrophysiologists would progress from these methods to devices so sensitive, that it is now possible to routinely detect the tiny current emanating from a single ion channel of a cell.  

Voltage-Clamp:   Invented in the late 1940s by Kenneth Cole, voltage clamping involves placing a second glass electrode inside the cell in order to "voltage clamp" the interior of the cell.   Voltage clamping made it possible for the first time to keep constant the membrane potential on the interior of the cell even during the sodium influx during the action potential.    Researchers from that point on could distinguish the voltage effects caused by influx of sodium or efflux of potassium from changes those made deliberately by the experimenter.   Voltage clamp essentially means to control the potential across a cell membrane.   Typically, a current is applied and changes in the cell membrane voltage potential are recorded.   Applied current flows locally across the cm both as ionic and capacity current.     A voltage is applied and the current is measured.   The measure of ionic current is the ionic movements thru ion channels.     Most methods use two intracellular electrodes.  One to record voltage and the other to send the current.    It is important to measure cell membrane potential changes at the cell membrane itself and not just at the electrodes.     It wasn't until voltage clamping was invented that a quantitative measurement of ionic currents was possible.    

Voltage clamping is often needed due to the capacitance-caused current (the cell membrane can be thought of as a capacitor).  This can happen whenever a change in voltage occurs across a membrane.    I(total) = I(ion channels) + Cm(dV/dt)     When doing voltage clamping, (which was invented by Marmont and Cole and more fully developed by Huxley and Hodgkin in 1952)  a current is injected which is equal in amplitude but opposite in sign to that which flows across cell membrane.  Therefore, there is no NET current across a cell membrane and the membrane potential is therefore kept constant.   By measuring the current that has to be injected to clamp potential, one also measures current flowing across cell membrane. 

Patch Clamp:  A very versatile and powerful technique which can be used to study individual ion channel activity in an isolated patch of cell membrane.   The invention of the patch clamp in the late 1970s allowed for the first time, the underlying kinetics of ion channels to be measured.    Patch clamping is undoubtedly the most important technique we have to study ion channels, and is often used in conjunction with other techniques such as mutagenesis in order to determine what effects these changes made in an ion channel have on its activity.  From these types of studies, structural and functional detail can be obtained about ion channels.   

Procedure: The cell membrane is first clamped so the voltage potential across it isn't changed.    The amount of electric current needed to maintain the cell membrane at a constant voltage is used as a measure to quantify the movement of ions thru the ion channel, or patch.       Patch clamping requires filtering of random high-frequency noise.   This noise arises primarily from charging currents inherent to the voltage clamping of a membrane.   Patch clamp has 3 main configurations: 1) inside out (pipette simply touches outside of membrane), 2) whole cell (where membrane is broken. Allows "summed" activity of all channels to be measured), 3) outside out, where pipette is pulled out (as in 2) and outside of cm side is now outside patch clamp. 4) like 1 but membrane is permeabilized by antibiotics).      An important difference between patch clamp and the two-electrode voltage clamp method is that patch clamp uses a single electrode both to control membrane potential and to measure current.    Another difference is that the patch clamp amplifier is highly sensitive and is able to resolve the tiny currents (pA) flowing thru single channels.

With a patch clamp, one can also measure 1) synaptic transmission between neurons in brain, 2) monitor change in cell membrane area during secretion.  Note: there are two critical noise sources: seal and current-to-voltage amplifier. Patch clamp is now sensitive enough to measure capacitance change due to a single vesicle fusion event.    Note also that single-channel recording (i.e. patch clamping) gives only 2 basic pieces of information: 1) current amplitude, and 2) dwell times.      One of the drawbacks of using patch clamp when looking for new channels is that it will bias the discovery process towards higher conducting channels, while lower ones (less than 1 pS) will be underrepresented.    The frequency response of most electronics is ~10 microseconds, however the actual transition (opening and closing) of ion channel gating is less than that, and will therefore remain invisible to these techniques.   

Voltage and Patch clamp facts, etc:   

Differing results have been obtained by various groups studying CFTR depending on the exact type of equipment or the cells used for protein expression.     Standardization is very important when using these techniques.   Whether or not an open channel is able to pass ions may depend on such factors as the absolute voltage, the concentrations of the ions, and the type of ions present on each side.    This means that agreement on basis for comparison is important.

When using artificial bilayers and purified ion channels, it is important to remember that these bilayers are usually more noisy.    However, with artificial bilayers, one can vary more details during the experiment, for example: lipid type, toxins, etc.

In theory, the two most direct ways to investigate ion channels is to record ion current that flows directly thru them (patch-clamping) or to measure a change in membrane potential that ion flow makes (voltage-clamping).       



Spectroscopic experiments  are useful for gaining structural as well as functional information about ion channels (like CFTR).   Many have been used in place of x-ray crystal structures.  Others to complement them.  Due to difficulties inherent in working with transmembrane proteins,  it has been difficult to obtain good structural information.   For example, these difficulties have often necessitated the use of peptide fragments rather than the whole ion channel protein in many of these studies.   In short, good reliable methods for obtaining structures of ion channels is still lacking.      

X-Ray Crystallization:   With this method, it is possible to construct models of protein structure to give information as to the location of nearly all of the atoms in a protein.    For this reason, it is considered a very powerful technique when it has been possible to apply it.    The "high-water mark" to date in ion channel research has been the solved structure of the potassium ion channel KscA from bacteria.   It was completed in 1998.   In the near future, it may be possible to improve crystallization methods enough to obtain many more ion channel structures.  (The company GenomiX has received 13 million dollars from the Cystic Fibrosis Foundation to solve the crystal structure of CFTR).  

Electron Microscopy:  (EM):   A technique involving the use of short-wavelength electrons to obtain extremely high resolution images of sub-microscopic structures.   It can be used in the "cryo" form to visualize individual membrane proteins without the need for harsh contrasting agents, or need to put sample directly into a vacuum.   EM is relatively easy to perform, and it can yield large amounts of information with low amounts of protein sample.   TEM (transmission electron microscopy) is probably the most applicable type of EM for ion channels.   STEM (scanningTEM) is primarily used to determine mass by recording elastic annular dark-field images (an angular detector specifically retrieves elastically scattered electrons at large angles, i.e. dark images) which represent a 2-D distribution of the projected mass of the conventional TEM.     It's Cryo-version (Cryo-TEM) is used to retrieve high resolution structural information (less than 10A) from ordered two-dimensional crystal arrays.   This helps avoid staining artifacts, which can lend significant problems for interpretation of data.     In cryoEM, one often takes advantage of the fact that high resolution amplitude information can be more readily obtained from electron diffraction patterns, which can then be combined with the phase information from the image.  This approach is "electron crystallography".      2-D crystals have been known to occur naturally following a simple reconstitution procedures, however this has proven to be the exception rather than the rule.    Using single particles, the resolution is rarely less than 20A (channels in membranes).   This has worked for proteins over 250 KD in mass (1995).     In the case of ion channels (due smaller masses of most of them) it is therefore necessary to use Cryo-TEM (which gives ~20A resolution) and an absolute requirement for highly resolved (less than 10A) cryoTEM to grow 2-D crystals.    The major advantage of 2-D crystals over single particles is due to fact that all structural motifs are already precisely aligned.      Natural state  (unstained, hydrated) means no stains or vacuum used.  Therefore, molecules will not be in vacuum and will not be as labile to ionizing radiation.  Therefore one needs to minimize the electron dose and decrease temperatures by lowering it to less than -160 C.   This allows for a significant lowering in free radicals, a better energy dissipation, and increased conductance, leading to increased radiation resistance.   To this end, 2-D crystals are especially useful since electron dose can be distributed over a large number of identical motifs.        The most limiting parameter in increased resolution EM is the quality of specimen.    Once 2D crystals that diffract to less than 20A in negative stain can be obtained reliably, the specimen-relevant resolution should be reassessed under cryo-conditions.  

*More about CRYO-EM:   The following is from a 1999 article entitled "Analysis of macromolecule structure and dynamics by cryoem" by Kuhlbrandt:   Many 2D crystals yield intermediate (6-8A) resolution and some are sufficiently well-ordered for atomic resolution structure determination.   Novel methods of 2D crystallization are being developed, including those that exploit the presence of affinity tags on overexpressed molecules and/or propensity of them to concentrate and crystallize at surfaces (air-water). Functionalized lipids that promote specific interactions are exploited in both lipid nanotubes, which serve as substrates for helical crystallization and lipids that form monolayers at air/water interface. 
Alpha/beta tubulin was recently resolved to 3.7A using cryo-EM.  Until recently, only naturally abundant membrane proteins (photosynthesis, respiration) were available in quantities sufficient for crystallization.   New overexpression methods are changing this.   For example, NhaA, a Na/H transporter that was homologusly overexpressed in E.Coli yielded a 4A projection map and a 3-D map at 7A is underway. (they formed highly ordered 3-D crystals).    2D crystals of the neurospora cell membrane proton pump H+-ATPase in the open, E1, conformation were grown directly on the surface of carbon-coated grids by precipitant-induced surface crystallization.   The crystals proved to be double layers. Yielded a map at 8A resolution in plane and 22A out of plane. Also, tubular crystals of the SR pump Ca++-ATPase yielded an 8A 3D map of the closed E2 conformation.    A 3D map at 7.5A resolution in plane and 21A out.   It was derived from 2D crystals of a truncated "cardiac gap junction channel', an ion channel that mediates communication between cells.    This is the first structure of a recombinant eukaryotic membrane protein.    Only micrograms were available, however. 
Photosystem II, which has 25 membrane proteins, subcore complex contains subunits D1,2,CP47, cyt.b559 and several minor proteins.   It was possible to visualize 23 transmembrane helices.
The near-perfect 2D crystals of bacteriorhodopsin continue to yield exciting new results on structure... A refined model at 3A has revealed the loops on the surface, including a small beta-sheet not seen in earlier cryoEMs.    Light-driven conformational changes during its photocycle have been examined by the calculation of electron diffraction difference maps at 3.5 A between ground state and photocycle intermediates.   They found two main conformations with the main change occurring between M1 and M2 intermediates.    In contrast to sheet-like 2D crystals, it is not necessary to record images of tilted tubular crystals (actin, microtubules), as all views are present in a single image (an important advantage).   Single-particle reconstruction offers a powerful means of visualizing structure and dynamics aspects of large macromolecules. The obvious advantage is that crystals are not required.    Moreover, structures of an extremely broad size range are suitable, with the lowest theoretical size limit at present being ~100 KD, determined by the requirement for accurately aligning individual particles.   But for molecules that don't have a lot of internal symmetry (note: viruses have a lot and therefore can get very good images. Highest is less than 15A for E. coli ribosomes, clathrin cages at around 21A and complex I, a 42 subunit, 890 kda membrane protein from bovine heart mitochondria at 22A.    The Future: As image processing methods for correcting electron optical effects are refined and developed, it should soon become possible to achieve resolutions that reveal secondary structure and even near-atomic detail in unsymmetric particles as well.    The full-length Shaker potassium channel was analyzed to 25A resolution by Sokolova et al. (Brandeis University, 2001).   They confirmed the location of the T1 domain.     EM studies can provide evidence for structure and stoichiometry of subunits.

Immunoelectron Microscopy:  uses metal-labeled antibodies specific to ion channels to determine location of ion channels in cell membranes.   It has been found with this technique that some types of channels tend to cluster together, for example nAChRs.    

Circular Dichroism (CD) and Fourier Transform Infra-Red Spectroscopy (FTIR):  

CD or FTIR are the best techniques when a particular protein sample can be obtained in high amounts. Gives the best results for approximation of the proportion of a-helix and b-sheet structure of a protein.   Potentially, NMR could be useful, but as of 1995 it's limited to peptides or small proteins in organic solvents.    There are 2 main types of spectroscopy: 1) for direct structure information, FTIR, CD, NMR can be used.   2) more indirect methods include fluorescence and ESR.    ESR and FTIR are well-adapted for membrane studies since they are undisturbed by large systems where light scattering is a problem.    High resolution NMR, however, is applicable only to relatively small peptide structures, especially CD, and are potentially susceptible to scattering artifacts.    FTIR is especially suited for the determination of secondary structure for peptides both in solution and in opaque and highly scattering (membrane) samples.    This technique uses vibration of the amide I bonds of the protein backbone, and it can also resolve certain side chains.   The rates of slow exchange of amide protons for deuterium can be determined either from band shifts in the amide I region, or from the loss of intensity in the amide II region.   FTIR can also give tertiary information as well.    For example, a membrane sample can be oriented on a substrate, and the orientation of the peptide backbone relative to the membrane surface can be determined from the dichroism of the amide I band with the polarized IR radiation. 

Dichroism is done at ~180-250 nm.     Alpha-helices usually give more definite results than beta-sheets.  Complications can arise because the  peptide bond may be affected by the interaction with neighboring groups, especially aromatics.   Membranes cause considerable problems due to scattering and nonuniform distribution of chromophores in light beam.  

Electron Spin Resonance (ESR):  is a sensitive technique which is able to resolve molecular mobility.    ESR can be used with opaque samples   The time-scale of sensitivity of the spectra from nitroxide spin labels is well-matched to the motions of lipids in membranes.     It is possible to find rotational rates for membrane proteins (which are related to cross-sectional area of the rotating species), and perimeter of channel assemblies.    In saturation studies, accessing of spin-labeled residues to paramagnetic ions and to molecular oxygen can be used to determine location, orientation and conformation of channel-peptide segments in membrane.    Perozo et al, using EPR,  trapped the K+ ion channel KscA in both open and closed positions and analyzed difference in EPR signal (they put in cystines and spin-labeled them.  Nitroxide spin labels were used ot analyze change in spin-label mobility and intersubunit (it has 4) spin-spin coupling as channel changed due to pH.   They found that transmembrane helices 1 and 2 underwent conformational changes as the pore opened.  The displacement of these helices was calculated to increase the diameter of pore and the position where the transmembrane helices converged.  Science 7/2/99

Information into protein folding can be among the most difficult types of information to obtain.    It's usually best to have EM or x-ray data.    But even at low resolution,  structural information can be very valuable.  For example, it can be used to provide evidence for quaternary structure and/or its overall shape.      

NMR:  (Nuclear Magnetic Resonance Spectroscopy):  NMR is valued for its high resolution spectra, but is only obtainable from relatively small molecules (or aggregates) for which there is rapid molecular rotation in solution.   This means that NMR is often limited to the structure determination of short, transmembrane peptides sequences placed in organic solvents or incorporated into liposomes or micelles.     Usually, the larger the aggregate (for example, membrane proteins incorporated into membranes) the broader the spectral lines.  Therefore, solid-state NMR is becoming more popular.    NMR has been used in the form of "2-D proton NMR" to determine structure of a 21 residue S4 segment of the sodium channel in deuterated organic solvent.       

A typical experiment is done the following way:  1) obtain sequence-specific assignments of all resonances, especially amide protons.   COSY (2-D) is used to assign thru-bond  connectivities (up to 3 bonds).  2) Thru-space (2-D NOE) bonds are done by NOESY NMR.     2.5 to 3.5 angstrom resolution can be obtained for strong and medium intensity cross-peaks.     3) Structural information is obtained from NOE which is used to determine the inter-proton distance constraints, including those of longer-range (less than 5 A for weak cross-peaks).     Mixing time is an important parameter.   (~100-400 msec for NOESY)   The mixing time depends on protein's size and mobility, and the mixing time limits the range of distances over which NOEs may be obtained, by limiting spin diffusion to more remote nuclei.    4) Secondary structure such as alpha helices and beta sheets can be seen using inter-residue NOESY cross peaks.      5)  When they can be measured, additional information comes from spin-spin coupling constant between the alpha-carbon and amide protons, which are approximately 4 HZ for an alpha-helix and 9 for beta-sheet.   6) The 3-D structure comes from a consistent interpretation of all NOE distance constraints.     For proteins of less than 100 amino acids, this usually results in a family of closely related structures that can be further refined by energy minimization and restrained molecular dynamics.  Additional conformational constraints e.g. dihedral angles from coupling constants, can be introduced into these calculations to improve precision.     Transfer-NOE is becoming useful in the study of ion channels. 

Fluorescence:  Involves use of light given off (fluorescence) following stimulation of a probe by light of a lower wavelength.    Both the spectrum and the quantum yield of the light signal are sensitive indicators of environmental effects like polarity.   (Fluorescence is much more sensitive to these factors than the conventional absorption spectrum).   Quenching molecules can be used to help determine the structure of ion channels, as only sites tagged with probes will be accessible.   Fluorescence can also be used in RET (resonance energy transfer) mode to give distance information.    Fluorescence spectra polarization can also be sensitive to the rotational dynamics that take place with in fluorescence lifetime of the probe.    Main drawbacks are: light scattering by particulate suspensions.     GFP fusion mutants are highly useful for localization studies of intact proteins.   Fluroescent probes which are halide specific, or change fluorescence upon changes in membrane potential are available for monitoring transport thru ion channels.   For example, the fluorophore SPQ can often be used to monitor chloride transport thru CFTR because it looses it's ability to fluoresce signal when chloride is present.    Osmotic swelling and resultant light scattering can also be used for activity measurements.    

FRAP:  Fluorescence Recovery After Photobleaching.    Used to help determine mobility of membrane proteins.   Involves use of fluorophore-tagged proteins which are temporarially bleached by intense laser light.   The time it takes for the proteinss outside the path of the laser beam (which are still fluorescent) to move back into the area is a measure of mobility of the proteins (protein diffusion coefficient) in that specific area.   For these channels in embryonic muscle cell membranes, they diffuse at about 10^-10 cm2/sec, which is considered typical for membrane proteins.   When they become immobile, they have coefficients two orders of magnitude less.   This can occur upon neuron contact with the muscle cell.   

Atomic Force Microscopy (AFM):  a "surface-imaging technique" invented in the 1980s.   While not a true "spectroscopic technique", AFM can be used for two-dimensional crystals of membrane proteins (like bacteriorhodopsin) or single ion channels in membranes.   Some advantages include: no dyes needed, it can be performed in aqueous buffers, and only small amounts of material often needed.   


These techniques can often be used to obtain relatively large amounts of ion channels and are therefore useful when techniques requiring large amounts of material are to be tried.   They can also help in the gaining of functional and structural information of ion channels, for example "site-directed mutagenesis".     Mutants often still need to be characterized by activity assays based on use of patch clamp method.  

Some background:    The discovery of restriction enzymes were central to molecular biology, as was the discovery of the enzyme reverse transcriptase (RT).  This is because RNA itself is very difficult to work with.  RT is able to change RNA such as mRNA into DNA (cDNA).     Historically two of the methods used to find channels have been 1) expression cloning, and 2) positional cloning.   An example of the use of expression cloning was done in 1993 by Kubo.   He first isolated mRNA from cardiac muscle, and found the ion channel Kir2.1, one of the first K+ channels that were inwardly rectifying.   He went about this by fractionating the mRNA from the cardiac muscle by size and then placed them into pools.  He then tested each pool of mRNAs in oocytes. subdivided over and over again until only one was left.  Then, homology screening of cDNA libraries was done to find the sequence of the gene.     The first positiononal cloning of a gene (ultrabiothorax) was the bithorax complex of homeobox genes in flies, and was done in 1979.    Positional cloning was made possible because of newly created overlapping segments of chromosomal DNA cloned in bacteriophage lambda covering over 200 kb were constructed using chromosome walking (1978).  An inversion that linked this region to the bithorax complex was done.  Note that, by late 1980, many mutant alleles had been located on the restriction map of the complex and shown to be the result of chromosomal breakage or transposable element insertion.  As early as 1975, cloned libraries of Drosophila existed for the entire genome.

ABC transporter proteins have been discovered by hybridization studies, degenerate PCR, as well as DNA sequence inspection.   Screening cDNA libraries can result in missing of homologus proteins due to tissue-specificity, while genomic libraries have the drawback of containing introns.   EST databases can be searched and overlapping sequences can be obtained to construct contigs over 1000 bases.   It's also possible to use antibodies which cross-react for screening expression libraries.   Heterodimers may coprecipitate.  

PCR:   Polymerase Chain Reaction has been used to find a mutation in a codon of a sodium channel gene by the use of appropriate primers.  The disease this mutation causes is called PAM (potassium-aggravated myotonia), and causes muscle stiffness.     Other techniques for gene characterization include SSCP:  single-stranded conformational polymorphisms.   CFTR, PAM (3 different mutations in the above codon discovered).  

Antibody generation to peptide regions can help in structure determination, but Abs don't often recognize native protein.  Monoclonal antibodies  (mAbs) have been made to specific regions, or peptides, of ion channels and then injected into cells such as nerve axons, revealing how they may sometimes alter the property of the ion channel, such as gating.     Polyclonal antibodies (pAbs) and mAbs were used in 1992 to localize CFTR to cytosol of CF epithelia.  Confocal microscopy to visualize it.  

Enzyme engineering into TM loop regions (betalactamase, PhoA), and monitoring whether it was expressed on intracellular or extracellular side. Can be done for each loop region.

Cystein Scanning: cystein engineering, radiolabels, fluorescent or spin-labeled SH reagents can pinpoint location of residues. "considered by many as the most definitive topographic probes. Can reveal nearest-neighbor interactions via monitoring formation of disulfide bonds. Distance relationships can be calculated by fluorescence quenching measurements as well. 

Coprecipitation experiments:  were used to find that CFTR binds a sodium channel.   Here, an antibody for one protein or the other is used in amounts necessary for precipitation out of solution.  Copurification is similar technique, often done unintentionally.  This is how it was discovered that a sodium ion channel interacts in rat brain neurons with the cytoskeletal protein ankyrin.     (Ankrin is believed to be the primary way membrane proteins are attached to the underlying cytoskeleton) 

Peptide models, in 1992, showed in vitro that the delta-F508 mutant CFTR is more sensitive to denaturing conditions than the wild-type.    Synporins are synthetic peptides derived from individual transmembrane helices of ion channels and may form ion channels in lipid bilayers.  

Reverse Pharmacology: makes use of chemicals which block the pore.   Pore-blocking drugs are uses as probes of pore structure.   Mutagenesis is combined.    Drug is assumed to bind in pore.   Can be a problem.  TEA, and peptides like charybdotoxin have been used.    This technique is used in conjunction with patch clamp.      Ligand-substitution rates can give information about pore diameters. 

In Vitro Translation:  Involves producing a protein from its mRNA in vitro (in a test tube without the cell being present)  It has been done for CFTR and was used to find uncleavable signal sequences.  It was found that CFTR's membrane targeting and insertion may depend on SRP and SRP receptor.  

Inhibitors:  Useful small molecules which reduce activity of a particular protein or complex of proteins.   For example, CFTR was shown to have it's degradation kinetics affected by inhibitors of the cytosolic proteosome, which caused accumulation of polyubiquinated forms of immature CFTR.     Also, nonhydrolyzable analogs of ATP have been used with CFTR in transport studies involving use of patch clamp.     In addition, inhibitors of microtubule assembly (nocodazole and colchicine) in T84 cells have been shown to reduce cAMP-stimulated chloride transport but not calcium-activated halide permeability.   Cytochalasin D, a disruptor of microtubule (actin) polymerization seems to stimulate CFTR in mouse fibroblasts.    Purified actin has also been shown by others to activate CFTR directly in excised patches.  

Overexpression:   When amounts of the protein are expressed by a cell over and above what is normal for the cell to make.  For example,  dF508-CFTR was overexpressed under control of a MT promoter.  It was found that some was able to escape from the endosome. 

Antisense Where oligonucleotides (two 18 mers) complementary to mRNA are injected into a cell to shut down translation of the mRNA into protein.  This technique was used for CFTR mRNA to show how CFTR is activated.  

Partial Proteolysis Fingerprinting:  Some enzyme proteases are able to cleave other proteins at specific places, and the peptide fragments this creates can be used to identify and/or characterize a protein of interest.   It can be used to find similarities between proteins.   Limited proteolysis was used by Armstrong in 1973 to prove for the first time that ion channels, specifically voltage-gated potassium channels, are made out of proteins.   

Electrophoresis Mobility Studies  Based on the way a protein moves thru a gel when under the influence of an electric field.   Some proteins will migrate at different rates depending on the conformation of the structure.  It has been used to detect conformational changes in the purified R-subunits of CFTR.   

SDS-PAGE   Electrophoresis of proteins which is done under denaturing conditions.  SDS detergent denatures nearly all proteins.   It can show differences in glycosylation patterns, as well as different translation initiation sites.  

Pulse-Chase When a radiolabeled isotope of an amino acid (for example, S-35 methionine) is used to indicate translation of a protein.  It can be used to follow ion channel maturation in the ER and Golgi.   Pulse chase studies showed that wild-type CFTR is inefficiently processed thru the ER. Endogenous proteases degrade 50 to 80% of it.   The half-life was determined to be 35 minutes.     

Radioactive toxin (ligand) binding to channels.  This method can give an estimate of channel densities in membranes.  For example, it has been used to determine that every square micrometer of an axon membrane has 110 sites for the toxin STX (STX binds directly to an ion channel).   Iodine-125 labeled alpha-bungarotoxin has been used to follow the time course of degradation of the nAChR ion channels.  It was found that the receptors had a half-life at the cell surface of 17-23 hours.     Radioactive toxins can also be used to study the mobility of ion channel receptors in membranes by following their distribution in the membrane as a function of time.   .   

Biotinylation (addition of a biotin moiety) of carbohydrate groups on the outside of ion channels allows monitoring of CFTR internalization and intracellular trafficking.    Lectin wheat germ agglutinin has also been used to label these "glycoproteins".   

Mutagenesis This is a method by which the sequence of amino acids in proteins can be changed very specifically.  One of the most interesting facts about proteins themselves is that often changing just a single amino acid can have profound effects on the protein function as a whole.    Mutagenesis has been used with CFTR to verify where the intracellular and extracellular loops are.   Glycosylation sites were engineered into the loops of CFTR one at a time and changes in the electrophoretic mobility was measured.     Mutagenesis is considered a very powerful technique when the data from it is analyzed correctly.    In general, mutations of amino acids that affect conductance or the binding of blocking molecules which only bind to the open state of the channel probably are involved in forming the pore of the ion channel.    Mutations that affect gating rather than conductance are most likely in regulatory regions of the ion channel.   

Counter-flow  Where a radioactive substrate is used to study movement of molecules thru the pores of ion channels.    This method was used to study selectivity of the channel  GlpF aquaporin.   It was found to transport a glycerol substrate.   With this method, a small amount of radioactive glycerol was placed on one side of membrane, while other was much higher in regular glycerol.   Since no counter-flow was discovered, it was concluded that this channel is highly specific for glycerol.    Radioisotope transport, for example chloride-36 for CFTR, has a long history in ion channel research.   

Protein-Reactive Reagents: Some small molecules will react specifically or nonspecifically with proteins.   Often, the location of the reaction can be determined in order to obtain some structural information about the protein.   Chlorpromazie reacts with protein side chains after an intense flash of ultraviolet light.   Sometimes, radioactive [H3]Chlorpromazie is used to help in identification of the residues or subunits involved following separation by gel electrophoresis.    The cystein-scanning method above relies on reagents which react specifically with sulfhydral groups.   Other specific molecules used include quinacrine, 

Cross-Linking: Often used with ion channels to determine if the channels form multimers.   For example, the mechanosensitive ion channel MscL forms a homopentamer.    This was shown by cross-linking the subunits together using the reagent disuccinimidyl suberate (DSS).   The MscL protein produced five bands during gel electrophoresis, suggesting a pentameric association.    It was later verified by x-ray crystallography in 1998.     

RECENT METHODS USED WHICH WERE REPORTED AT BIOPHYSICAL SOCIETY MEETING IN FEB. 2001   (this list includes methods used to characterize membrane proteins in addition to ion channels)

Infrared Dichroism:  used to deduce orientations of alpha helices and beta strands of membrane proteins.  

Site-directed spin labeling: used to reveal first step in transport pathway of a TONB-dependent transporter.  

Various solvents and NMR to characterize structure of peptides of membrane proteins.

Solid State NMR was used to investigate structure and dynamics of membrane bound colicin 1a.      "Solid-state NMR is a developing technique for the determination of there-dimensional molecular structures, and has particular promise for structure determination of membrane proteins.  Using samples that have uniformly aligned molecules in a lipid bilayer, the measurements of dipolar and quadrupolar couplings and chemical shifts provide orientational restraints of the molecular frame of atomic sites with respect to the magnetic field and a unique molecular axis.   

Antibody Imprinting was used to help map sites of tertiary structure of light-adapted rhodopsin.    

Cross-linking studies have been used to characterize active sites, protein-protein interactions, etc.

Mass Spectroscopy (MALDI) was used with picomol levels of rhodopsin to characterize fragments created by cyanogen bromide.   Some fragments were sequenced to verify.    No chromatography step was needed to separate peptides.  

Chemical Modification Fingerprinting was done on the erythrocyte calcium pump and the renal sodium pump.  Here, various reagents known to interact with His or Cys were added in presence of various possible inhibitors to infer structure/function.

FT-IR was used to infer oxidation states of various tyrosines in photosystem 2.   Vibrational spectra was used to obtain structural information about the tyrosine residues in both oxidized and neutral form.   Tyrosyl radicals were generated by UV light and photolysis was followed using Difference FT-IR spectra.   EPR was also used in order to measure spin density distribution in each sample.    In another study, FT-IR was used to follow cysteine structural changes via SH stretching.   Monitored hydrogen bonding and conformation changes as well as electric field effects.   

AFM was used to probe lipid-protein interactions of AB peptides in Alzheimers disease.   The surface topography of outer leaflet was imaged.  

Solvent Accessibility Method was used to study activation gating conformational changes in the ion channel KCSA.   

SDS-PAGE  and perflurooctonate-PAGE (which is more tolerant of protein-protein interactions) has been used to determine hydrogen-bonding between TM helices takes place because it probably forms a hairpin secondary structure in the gel and therefore moves faster.

AQP1 human water channel was solved to 3.7A resolution in-plane and ~7A perpendicular to bilayer using 2D crystals of AQP1 and CRYO-Crystallography.   Backbones for loops between helices were clear, but side-chains less so.    Note: Negative Stain EM has been used to discern the major features of the 300 KD volt. gated sodium channel (Yutaka Ueno) along w/cryo-EM to see diameter (100A) and height (140A).   

Tryptophan fluorescence quenching was used to follow NBD and R domain interactions when expressed as soluble domains.   "ATP binding to NBDs quenches Trp fluorescence.   As well as when domains were mixed together. 

Freeze-Fracture and Lectin-Gold-labeling:  Note: FFR alone indicated in 1998 CFTR exists as dimers.   Lectin-gold labeling of the single glycosylaton site in P-glycoprotein (at 25A resolution) indicated a monomeric form. 

3 methods to count ion channels:  1) neurotoxins like TTX as labels Need to know specifics of binding stoichiometry,  2) useful only for steeply voltage-gated channels starts with gating -current measurements.   Requires isolating the gating current for one type of channel and knowing how many gating charges to attribute to each channel molecule.  3) measurements of single-channel conductance.     Ex:  "autoradiographic studies on muscles of frog, lizard, and mouse show that the site density in the extrajunctional region (away from the endplate) averages 6 to 50 sites/um2, while at the top of the junctional folds opposite the active zones, it is as high as 20,000/um2.   


Various methods to conceptualize how of ion channels function can be done with computer programs, each with varying degrees of accuracy (agreement with experimental data).    Monte Carlo and Brownian dynamics, as well as treatment of solvents as simple force fields have all been tried.   Molecular dynamics can be used to simulate movements on nanosecond time scales which is fast enough to model how phospholipids in membrane bilayers move.  With molecular dynamics simulation, it is possible to assign all atoms of a system such as an ion channel in solvent and ions as discrete points in motion.  Newton's laws of motion are then solved for the system.  A drawback is that the three-dimensional structure of the ion channel needs to be known.   

Sequence info can work for hydropathy as well as determining where post-transl. modifications occur (glycosylation, phosphroylaton).    Hydropathy plots have a problem with distinguishing alpha helices from beta sheets.   Looks for a run of 20 or more hydrophobic amino acids.  Each amino acids is assigned a numerical score related to the hydrophobicity of its side chain.   Doesn't always work since some globular proteins have hydrophobic regions inside protein, also hydrophilic amino acids are needed in TM segments for pore or h-bonding to other helices, 

Hydropathy Analysis:  When the primary amino acid sequence, along with algorithyms, is used to predict which amino acids contribute to the transmembrane domains, and which are probably cytoplasmic.    When the MscL ion channel was solved by x-ray crystallography in 1998, the sequence positions of the two transmembrane helices roughly agreed with the predictions of the hydropathy analysis done in 1994.    



Ion Channels:  Molecules in Action     Aidley     1996

Ionic Channels of Excitable Membranes   Bertil Hille    Sinauer Associates Inc.  Sunderland, MA  1992