Proteoliposomes are used to study purified channels and transporters reconstituted in a well-defined biochemical environment. An experimental procedure to measure efflux mediated by these proteins is illustrated. The steps to prepare proteoliposomes, perform the recordings, and analyze data to quantitatively determine the functional properties of the reconstituted protein are described.
The last 15 years have been characterized by an explosion in the ability to overexpress and purify membrane proteins from prokaryotic organisms as well as from eukaryotes. This increase has been largely driven by the successful push to obtain structural information on membrane proteins. However, the ability to functionally interrogate these proteins has not advanced at the same rate and is often limited to qualitative assays of limited quantitative value, thereby limiting the mechanistic insights that they can provide. An assay to quantitatively investigate the transport activity of reconstituted Cl– channels or transporters is described. The assay is based on the measure of the efflux rate of Cl– from proteoliposomes following the addition of the K+ ionophore valinomycin to shunt the membrane potential. An ion sensitive electrode is used to follow the time-course of ion efflux from proteoliposomes reconstituted with the desired protein. The method is highly suited for mechanistic studies, as it allows for the quantitative determination of key properties of the reconstituted protein, such as its unitary transport rate, the fraction of active protein and the molecular mass of the functional unit. The assay can also be utilized to determine the effect of small molecule compounds that directly inhibit/activate the reconstituted protein, as well as to test the modulatory effects of the membrane composition or lipid-modifying reagents. Where possible, direct comparison between results obtained using this method were found to be in good agreement with those obtained using electrophysiological approaches. The technique is illustrated using CLC-ec1, a CLC-type H+/Cl– exchanger, as a model system. The efflux assay can be utilized to study any Cl– conducting channel/transporter and, with minimal changes, can be adapted to study any ion-transporting protein.
In last two decades the ability to overexpress and purify membrane transport proteins has dramatically increased: ion channels, primary and secondary transporters are now routinely purified from heterologous expression systems as well as natural sources. New approaches to monitor expression, improve and facilitate the extraction and enhance stability of these proteins are constantly being developed 1-5. These technological breakthroughs have been instrumental in triggering the explosion of atomic-level structural information on membrane proteins which, in turn, enhanced our understanding of the structural bases of their function. In contrast, our ability to probe the functional properties of the purified proteins did not increase at the same rate, so that in some cases high resolution structural information is accompanied by qualitative functional data, thus limiting our ability to quantitatively test structure-based predictions. Hence, the development of quantitative and generalizable functional assays is a key step towards the elucidation of the mechanistic underpinnings of membrane protein function.
Here we describe an efflux assay that can be used to quantitatively determine key functional properties of purified and reconstituted Cl– channels and transporters. The principles underlying the assay can be generalized to a variety of transport systems, as well as to non ion-transporting proteins. Liposomes are reconstituted with purified Cl– channel/transporters in the presence of a large Cl– gradient (Figure 1A, B). Cl– efflux is initiated by the addition of an ionophore to allow for counter-ion flux, in our case the K+ ionophore valinomycin, which shunt the voltage established by the Cl– gradient and set the initial membrane potential to the equilibrium potential of K+6,7. Without the ionophore no significant net Cl– efflux occurs, as it is prevented by the generation of a transmembrane potential. The data is quantitatively described by two measurable parameters (Figure 1C): τ, the time constant of Cl– efflux, and f0, the fraction of liposomes not containing an active protein. From τ and f0 the unitary Cl– transport rate, the fraction of active proteins and the molecular mass of the active complex can be derived 8. The technique is illustrated here using proteoliposomes reconstituted with CLC-ec1, a well characterized CLC-type H+/Cl– exchanger of known structure and function. This assay is readily generalized to channels or transporters with different ionic selectivity or whose activity depends on the presence of voltage and/or ligands. Furthermore, this assay can be used to determine whether small molecules directly affect the reconstituted protein, to quantitate the effects of these compounds and how membrane composition or lipid-modifying reagents affect the function of the reconstituted channels and transporters.
1. Lipid Preparation
2. Proteoliposome Formation
NOTE: Several strategies can be employed to insert the detergent-solubilized protein into liposomes. For CLC-ec1 dialysis works well and is therefore the method of choice 6,9,10.
3. Recording Set-up
NOTE: The recording set-up (Figure 2A) consists of two chambers (flat bottomed cylinders, ~3-4 ml volume), a Cl– (see below) electrode, a pH meter with an analog or digital electrical output, a digitizer, and a computer with an appropriate acquisition software.
4. Preparation of the Cl– Electrode
5. Preparation of Unilamellar Vesicles
6. Efflux Measurement
7. Data Analysis
We describe a detailed and robust protocol to measure Cl– transport mediated by purified CLC-ec1, a prokaryotic CLC-type H+/Cl– exchanger, reconstituted in liposomes. A schematic representation of the experiment is shown in Figure 3. Proteoliposomes reconstituted with purified CLC-ec1 and containing high internal Cl– are immersed in a bath solution containing low Cl–. Under these conditions net Cl– efflux is prevented by the buildup of positive charge in the liposome (Figure 3A). Addition of Valinomycin allows K+ to move across the membrane thus shunting the electrical potential and initiating Cl– efflux (Figure 3B). The flux time course is monitored using a Cl–-selective electrode and the total amount of Cl– contained in the vesicles is directly measured by solubilizing them using detergent. From these bulk experiments it is possible to determine properties of the single reconstituted proteins, such as turnover rate, stoichiometry of the active complex and fraction of active protein. Using Equations 7-9 to fit the efflux time course and estimate the unitary transport rate of CLC-ec1 we find that τ(0.2 µg/mg) = 41 sec and f0(0.2 µg/mg) = 0.31, indicating that ~1/3 of the liposomes contains 0 active proteins. From these values we calculate that the unitary transport rate is γ~2,500 Cl– sec-1, in good agreement with published values 8,14.
Figure 1: The Cl– efflux assay. (A–B) Schematic representation of the Cl–-efflux assay for protein-free liposomes (A) or CLC-ec1 reconstituted vesicles (B). (C) Simulated time course of ionophore-initiated Cl– efflux from proteoliposomes (black) or protein-free liposomes (gray dashed line). Efflux is initiated by addition of ionophore (*) and terminated by the addition of detergent to dissolve liposomes (^). Red dashed line is an exponential fit to determine the time constant, τ, of Cl– efflux from liposomes containing at least 1 active copy of CLC-ec1 and f0 is the fraction of liposomes containing 0 active proteins. Please click here to view a larger version of this figure.
Figure 2: Recording set up. (A) The recording set up. (B) Close-up of the chambers. Please click here to view a larger version of this figure.
Figure 3: Sample trace and analysis. (A) Typical V(t) traces of efflux experiments from proteoliposomes reconstituted with WT CLC-ec1 at 0.2 μg/mg P/L. Arrows indicate the times of addition of the KCl Calibration pulse, Liposomes, valinomycin and β-OG. Dashed lines indicate the ΔV values at various stages. (B) The V(t) trace is converted into Cl(t) using Equation 5, normalized using Equation 6 and the efflux time course is fit to Equation 7 (red dashed line). (C) Normalized efflux time courses from proteoliposomes reconstituted with WT CLC-ec1 at 0.2 μg/mg P/L (black) or at 5 μg/mg P/L (gray). Dashed lines are best fits of the efflux time courses to Equation 7.
We have described a detailed protocol to measure Cl– transport mediated by purified anion-selective channels or transporters reconstituted in liposomes. The example used was the prokaryotic H+/Cl– exchanger CLC-ec1. However, the methodology can be readily adapted to study channels gated by ligands 12,13,15, voltage 11,12, or sporting different anionic selectivity 15,16 by replacing the Ag:AgCl electrode with one suitable for the ion under consideration. Electrodes selective for ions other than Cl–, such as H+, I– and F–, are commercially available.
A discussion of some of the critical steps and assumptions follow.
Limits of the Poisson dilution assumption
The derivation of the unitary rate (Equation 7-9) is valid only in a Poisson dilution regime, when most liposomes contain only one active protein so that the probability of having multiple copies in a given vesicle is low and the macroscopic efflux time constant of the vesicle population is well approximated by that of liposomes containing a single protein. Indeed, a comparison of the Cl– efflux time course mediated by liposomes reconstituted with CLC-ec1 at low (black, 0.2 µg/mg) and high P/L (gray, 5 µg/mg) (Figure 3C) shows that in the latter case the efflux kinetics become faster and f0 decreases, indicating that a smaller fraction of the total liposomes contains 0 proteins. The total amount of Cl– contained in both vesicle samples was comparable, ~0.7 and ~0.55 µmoles respectively, indicating that the total internal volumes of the vesicles in the two samples was similar and that the amount of protein reconstituted does not affect the size of the vesicles. Fitting the high P/L efflux time course to Equation 7 produces values of τ(5 µg/mg) = 8.8 sec and f0(5 µg/mg) = 0.06, indicating that only ~6% of the vesicles contain no active CLC-ec1 dimers in contrast to the ~31% found when P/L = 0.2 µg/mg. Thus, the estimate of the unitary turnover rate at high P/L is γ~450 Cl– sec-1, nearly 6-fold lower than the one obtained at low P/L and of published data 8,14. Therefore, analysis of efflux time courses from liposomes reconstituted at high P/L’s will underestimate the unitary transport rate of the reconstituted protein. Thus, to accurately determine the unitary transport rate of a novel protein it is of key importance to precisely determine the amount of protein reconstituted and to work at low P/L’s, in a Poisson dilution regime. This is ideally achieved by determining a full protein titration analysis to identify the optimal regime to work at.
Determination of the mass of the active complex
The analysis described above assumes that the stoichiometry of the active complex is known and the reconstituted protein is fully active. However, for new preparations these quantities might not be known a priori. In this case it is possible to directly determine these parameters by measuring the f0’s at different protein to lipid ratios
where p is the protein density,, ρ is the number of liposomes per mass of lipid, NA is Avogadro’s number, MP is the mass of the functional channel complex and φ is the fraction of active proteins. By fitting the experimentally determined f0’s at various protein to lipid ratios it is possible to determine p0, and therefore MP and φ.
A finite fraction of liposomes is refractory to incorporation
The derivation of the unitary transport rate assumes that at high P/L’s all liposomes will contain at least one active transport protein. However, in some cases this has been found not to be the case: at high P/L’s a significant fraction of liposomes remains refractory to reconstitution of some proteins 11-13,17. While the origin of this phenomenon are not clear it is conceivable that larger proteins might be excluded from vesicles with smaller radiuses thus reducing the number of available liposomes.
In this case Equation 10 needs to be modified to:
where θ is the fraction of liposomes refractory to protein incorporation. Importantly, ρ and φ are also affected by the reconstitution method since different procedures will have different recoveries for lipids and proteins during liposome reconstitution and formation. Assuming 100% yield for both will lead to a mis-estimation of γ. Therefore, to obtain a precise quantitation of the unitary turnover/conductance it is important to experimentally determine the fraction of lipid and protein lost during reconstitution 8,11.
Orientation of the reconstituted proteins
One of the assumptions made during the analysis is that all reconstituted proteins are functionally equivalent. However, many channels and transporters have preferred directions of transports, a phenomenon known as rectification. This functional asymmetry might lead to an estimation error in the turnover rate. Furthermore, the orientation of the reconstituted proteins is a priori random, so that the measured unitary turnover rate is the weighted average of the protein’s rates when operating forwards or backwards. Several methods can be used to determine the orientation of the reconstituted proteins, for example by measuring the fraction of a tag that is cleavable from the extraliposomal solution or by measuring the sidedness of the reactivity of a single cysteine residue introduced in a soluble-accessible region. Finally, it is possible to circumvent this problem by functionally silencing one of the two populations using sided inhibitors or by selectively activating only proteins in one orientation via the addition of activating compound to side only or by setting the transmembrane voltage to opportune values using K+ gradients, for channels or transporters that are ligand- or voltage-dependent.
Considerations on the formation of tight liposomes
One of the key factors enabling the use of the efflux assay described here is the formation of tight liposomes. Three important variables to be considered to this end are the choice of detergent used to solubilize the protein, the detergent-removal strategy and the choice of lipid composition of the liposomes. In the protocol presented here, CLC-ec1 is solubilized in Decyl-Maltopyranoside (DM), a detergent with a high critical micelle concentration (CMC) value which is therefore easily removed. However, a variety of synthetic and naturally occurring detergents have been used to purify proteins that were subsequently reconstituted in tight liposomes with no effects on the tightness of the vesicles. These detergents have a wide range of CMCs, as high as millimolar (such as DM or Digitonin) and as low as micromolar (such as Dodecyl- Maltopyranoside (DDM) or dioctylpropane-bis-maltopyranoside (DMNG)). It is important to note however, that the lipids used to form the liposomes are solubilized using CHAPS, or another mild detergent, and therefore the hybrid micelles resulting from the mixture of protein and lipids might have chemico-physical properties different from those of the parent detergents. It is therefore possible that in some cases the residual detergent removal that could destabilize the liposomes leading to more pronounced leaks. To circumvent this problem it is advisable to investigate alternative detergent removal strategies, such as biobeads 12,13, spin columns 15 or gel filtration 18. Finally, the lipid composition of the vesicles is an important parameter as specific mixtures might be leaky to different ions in certain conditions. For example, liposomes formed from E. coli polar lipid extract are tight to Cl– at acidic pH’s but become progressively more leaky as the pH increases 19 or liposomes formed from a 3:1 mixture of 1-palmitoyl-2-oleoyl phosphatidyl-ethanolamine (POPE) and 1-palmitoyl-2-oleoyl phosphatidylglycerol (POPG) display a much lower permeability to H+ than those formed from E. coli polar lipid extract 20. It is important to note however that the lipid composition of the vesicles might also affect the activity of the reconstituted protein 13,21,22. Thus, it is important to precisely determine the properties of the protein-free vesicles prepared in each condition and then to assess how this affects the properties of the reconstituted protein.
Generalization to non CLC-type channels and transporters
The efflux assay, as described here, can be directly utilized to determine the single molecule properties of any Cl–-selective channel or transporter, and for example has been used to characterize the properties of the Ca2+-activated Cl– channel TMEM16A 12. We now discuss how to adapt the efflux assay to investigate the properties of transporters or channels that are not Cl– selective and/or have unitary transport rates that are significantly different from the ~103 ion sec-1 of CLC-ec1.
The simplest case is that the protein under study is anion-selective. In this case the only necessary adjustment is to replace the Ag:AgCl electrode with a commercially available one of appropriate selectivity, as was done for the F–-selective Flucs and CLCs 16,23,24 or I– permeable channels such as CFTR 15. If, on the other hand, the reconstituted protein is selective for cations, an ionophore other than valinomycin has to be used to initiate efflux. Given the scarcity of validated Cl– ionophores a useful substitute is a H+ ionophore, such as Carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone (FCCP). In this case it is important to ensure that the intravesicular pH is maintained constant, for example by increasing the buffering capacity of the internal solution. An alternative strategy is to follow cation efflux through a reconstituted channel or transporter by using protons as a counterion and monitoring H+ transport using a pH meter 25 or a ratiometric pH-sensitive probe 26. However, the indirect nature of these measurements renders quantification of the transport properties of the reconstituted protein difficult. Finally, electrodes selective for a number of cations exist and are commercially available. A third scenario is that the protein under study is permeable to both anions and cations, for example a poorly selective channel 13 or a cation/Cl– cotransporter. In this case, the reconstituted liposomes do not maintain a KCl gradient during the solution exchange process (steps 5.4-5.5) so that they lose their salt content. Therefore in this case all kinetic information is lost. However, the fraction of liposomes containing at least one active protein, f0, can still be determined by adding detergent and measuring the residual trapped Cl– content (step 6.8). This allows for the determination of the molecular mass of the protein by carrying out a protein titration and using Equation 10.
Another aspect to consider is the possibility that unitary transport rate of the reconstituted protein is orders of magnitude different from that of CLC-ec1. For example, most transporters have turnover rates of 1-10 sec-1, 2-3 orders of magnitude lower than CLC-ec1, while most channels conduct ions at 106-107 sec-1, 3-4,000 fold faster than CLC-ec1. For slow transporters the rate of Cl– leakage from protein-free liposomes can become limiting, as its relative weight in the efflux process increases as the protein’s transport rate decreases. In our experience the leak rate varies somewhat between preparations and is usually in the order of a few ion sec-1. Therefore, when working with slow transporters it is advisable to prepare protein-free liposomes side by side to each reconstitution to accurately measure the leak corresponding to each batch of vesicles. The efflux assay has been successfully used to determine the unitary rate of slow transporters with turnover rates around 1-10 ion sec-1, such as a cyanobacterial CLC 27. The converse situation occurs when the reconstituted protein has a high conductance, for example an ion channel. In this case the efflux kinetics are too fast to be resolved as the mixing time (~1-2 sec) and the intrinsic response time of the recording electrode become rate-limiting. In this case, the kinetic information is lost but the mass of the active complex can still be determined 24. It is worth noting that other approaches, such as planar lipid bilayer recordings or patch clamping liposomes are more suitable to study purified ion channels as these techniques directly provide single molecule information with high time resolution.
The authors have nothing to disclose.
This work was supported by NIH grant GM085232 and an Irma T. Hirschl/ Monique Weill-Caulier Scholar Award (to A.A.).
Name of Reagent/ Equipment | Company | Catalog Number | Comments/Description |
Liposomicator, Avanti Polar Lipids Inc. | Avanti Polar Lipids Inc. | 610200 | |
IEC Centra CL2 Benchtop | Thermo Scientific | ||
Orion Research Model 701A Digital pH-mV meter | These can be found on Ebay. | ||
Non-functional pH probe | Any pH meter probe with silver wires will work. The glass/plastic coating needs to be removed and the wires cleaned. | ||
DI-710 Data Logger | DATAQ instruments | ||
WinDAQ acquisition software | DATAQ instruments | ||
Pierce Disposable Plastic Columns, Gravity-flow, 2ml | Pierce (Thermo Scientific) | 29922 | |
KIMAX Culture Tubes, Disposable, Borosilicate Glass | Kimble Chase | 73500-13100 | |
Extruder Set With Holder/Heating Block | Avanti Polar Lipids Inc. | 610000 | |
Computer |