JBC Oz Biosciences

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Originally published In Press as doi:10.1074/jbc.M513634200 on March 23, 2006

J. Biol. Chem., Vol. 281, Issue 20, 14408-14416, May 19, 2006
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
281/20/14408    most recent
M513634200v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Sobko, A. A.
Right arrow Articles by Cramer, W. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Sobko, A. A.
Right arrow Articles by Cramer, W. A.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

Lipid Dependence of the Channel Properties of a Colicin E1-Lipid Toroidal Pore*

Alexander A. Sobko{ddagger}, Elena A. Kotova{ddagger}, Yuri N. Antonenko{ddagger}1, Stanislav D. Zakharov§, and William A. Cramer§

From the {ddagger}A. N. Belozersky Institute of Physico-Chemical Biology, Moscow State University, 119992 Moscow, Russia, the §Department of Biological Sciences, Purdue University, West Lafayette, Indiana 47907, and the Institute of Basic Problems of Biology, Russian Academy of Sciences, Puschino, Moscow Region 142290, Russia

Received for publication, December 22, 2005 , and in revised form, March 3, 2006.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 
Colicin E1 belongs to a group of bacteriocins whose cytotoxicity toward Escherichia coli is exerted through formation of ion channels that depolarize the cytoplasmic membrane. The lipid dependence of colicin single-channel conductance demonstrated intimate involvement of lipid in the structure of this channel. The colicin formed "small" conductance 60-picosiemens (pS) channels, with properties similar to those previously characterized, in 1,2-dieicosenoyl-sn-glycero-3-phosphocholine (C20) or thinner membranes, whereas it formed a novel "large" conductance 600-pS state in thicker 1,2-dierucoyl-sn-glycero-3-phosphocholine (C22) bilayers. Both channel states were anion-selective and voltage-gated and displayed a requirement for acidic pH. Lipids having negative spontaneous curvature inhibited the formation of both channels but increased the ratio of open 600 pS to 60 pS conductance states. Different diameters of small and large channels, 12 and 16 Å, were determined from the dependence of single-channel conductance on the size of nonelectrolyte solute probes. Colicin-induced lipid "flip-flop" and the decrease in anion selectivity of the channel in the presence of negatively charged lipids implied a significant contribution of lipid to the structure of the channel, most readily described as toroidal organization of lipid and protein to form the channel pore.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 
The lipid environment of membrane proteins that defines the local distribution of polarity, dielectric constant, and steric geometry has a central role in the determination of protein structure and function. For ion channels, lipids impose the following constraints: (i) structure-function is sensitive to the matching of hydrophobic thickness of the protein and lipid bilayer (1-10); (ii) opening of channels (e.g. mechano-sensitive, can be dependent on bilayer tension) (11), which manifests itself in a dependence on acyl chain length (12); (iii) spontaneous curvature (SC)2 of lipids linked to the lateral pressure of the bilayer can modulate ion channel activity (13), as demonstrated for the KcsA channel (14) and mechano-sensitive MscL channels (12, 15); and (iv) many channel proteins require anionic lipids for function (16-19).

Channel-forming proteins can require specific lipids, presumably because of a combination of charge and curvature parameters, to form active channels (e.g. anionic lipid bound between the transmembrane {alpha}-helices and necessary for gating of the KcsA channel (20, 21)). The lantibiotic nisin provides a well defined example of lipid participation in peptide pore formation, in which the peptidoglycan precursor lipid II is an intrinsic component of the pore formed by this antimicrobial peptide (22). The proposed structure of this pore contains 5-8 nisin molecules and an identical number of lipid II molecules (23). Participation of lipid in the formation of a channel wall was hypothesized for cholesterol-dependent cytolysins (24, 25).

Colicins are plasmid-encoded bacteriocins that are cytotoxic to Escherichia coli and related strains. Their modes of lethal action include enzymatic endoribonucleolytic and endodeoxyribonucleolytic cleavage of ribosomal RNA and DNA (26, 27) and voltage-gated ion channel or "pore" formation by colicins, including A, B, E1, Ia, Ib, and N, which depolarize the cytoplasmic membrane and de-energize the cell (26, 28-31). Colicin E1, which is the subject of the present study, unfolds on the lipid membrane surface to form a surface-bound two-dimensional flexible helical array that is a precursor state to the inserted channel (32, 33). The channel domain subsequently inserts into the membrane and forms ion-conducting pores that are voltage-dependent (29, 30, 34). These pore-forming colicins are remarkable for the fact that the formation of ion channels, as studied for colicin Ia, leads to voltage-gated translocation across the planar bilayer membrane of a large part of the colicin polypeptide (35, 36). The actual transmembrane structure of the colicin pores and the details of the translocation mechanism are not known at present. It has been inferred that the colicin Ia channel consists of three and four transmembrane (TM) helices, respectively, for the C-terminal pore-forming domain of colicin Ia and the intact colicin (37). A concern about this model is the structural sufficiency of three transmembrane helices for a stable and functional channel. An additional concern about all channel models of the pore-forming colicins is that the predicted length, 15-17 residues, of the two putative TM helices that form the hydrophobic hairpin anchor, two of the 3-4 helices in the channel structure, is appreciably shorter than the length of the average TM helix (~20 residues) of integral membrane proteins (38, 39).

A structural role of lipids in peptide-induced pore formation was implied by the sensitivity of the membrane-permeabilizing activity of the peptides to the lipid SC (40-42). Based on the SC dependence of colicin E1 macroscopic channel activity in bilayer lipid membranes, it has been proposed (43, 44) that the toroidal pore model, inferred for some {alpha}-helical antibacterial peptides (40, 45-48), may be applicable to the channel structure of the pore-forming colicins. This model assumes that the pore walls are formed not only by peptide {alpha}-helices but also by head-groups of the bilayer-forming phospholipids. Recently, the correlation between the pore forming potency of peptides and the SC of the lipid as a support for the toroidal model was questioned, based on similar correlations exhibited by melittin and alamethicin (49). Therefore, it was of importance to obtain additional lines of evidence for the toroidal pore formation in the case of colicin E1.


Figure 1
View larger version (24K):
[in this window]
[in a new window]
  		FIGURE 1.
A, single-channel trace of P178 in planar bilayer lipid membranes. The membrane was formed from a decane solution of DPhPC, as described under "Experimental Procedures." The buffer solution contained 10 mM potassium acetate and 1 M KCl, pH 4.0. A voltage of 80 mV, trans-negative, was applied to BLM. B, histogram of the current described in A. C, dwell time histogram for large channels. The solid line is a monoexponential fit of the data with a characteristic decay time of 200 ms. D, single channels of P178; changes in the current (solid line) upon switching the sign of the voltage (dashed line) applied to the BLM.

 
In the present work, the effect of lipid chain length and SC on single channels formed by colicin E1 was studied. A novel "large" channel state of the colicin E1 channel was described, whose conductance is greater by an order of magnitude than that of the "small" channel, many of whose properties have previously been reported in the literature (34, 50-54). Membrane thickness was found to be a major determinant of the ratio of the two ("low" and "high" conductance) states, since the open probability of the high conductance state increased with membrane thickness. A strong peptide-lipid interaction was also inferred from the observations of colicin E1-induced lipid flip-flop and the dramatic sensitivity of the channel conductance to lipid surface charge in 1 M KCl.


   EXPERIMENTAL PROCEDURES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
RESULTS
 DISCUSSION
 REFERENCES
 
Colicin E1—The 522-residue colicin E1 was prepared as previously described (55). The 178-residue C-terminal colicin E1 channel polypeptide, P178, was prepared by thermolysin proteolysis of intact colicin E1 (56).

Liposomes—Liposomes were prepared by evaporation under a stream of nitrogen of a 2% solution of a DPhPC/DPhPG mixture (70%/30%) in chloroform (Merck, Darmstadt, Germany) followed by hydration with a buffer solution containing 10 mM beta-alanine, 0.12 M KCl, pH 4.0. The mixture was vortexed, passed through a cycle of freezing and thawing, and extruded through 0.1-µm pore size Nucleopore polycarbonate membranes using an Avanti Mini-Extruder. The measurement of lipid flip-flop was performed as described by Müller et al. (57). To label vesicles with synthetic lipid probe 1-lauroyl-2-(1'pyrenebutyroyl)-sn-glycero-3-phosphocholine (pyPC) exclusively on the outer leaflet, 15 µM pyPC dissolved in ethanol was added to the buffer solution containing liposomes (final lipid concentration 300 µM). Incorporation of pyPC into the outer membrane leaflet was followed by measuring the ratio of fluorescence intensities of monomers (IM) and excimers (IE) at 395 and 495 nm, respectively. The fluorescence was excited at 344 nm.

Planar Bilayer Lipid Membranes (BLMs)—BLMs were formed from a 2% solution of lipid in squalene, hexadecane, or decane by the brush technique on a 0.5-mm diameter hole in a Teflon partition separating two compartments of a cell containing aqueous solutions of 1 M KCl, 10 mM CH3COOK, pH 4.0 (unless otherwise stated). All lipids used in the present work (1,2-dinervonoyl-sn-glycero-3-phosphocholine (DNvPC (C24:1)), 1,2-dierucoyl-sn-glycero-3-phosphocholine (DErPC (C22:1)), 1,2-dieicosenoyl-sn-glycero-3-phosphocholine (DEcPC (C20:1)), 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC (C18:1 cis)), 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE (C18)), 1,2-diphytanoyl-sn-glycero-3-phosphocholine (DPhPC (C16)), and 1,2-diphytanoyl-sn-glycero-3-(phospho-rac-(1-glycerol)) (DPhPG (C16))) were obtained from Avanti%20Polar%20Lipids">Avanti Polar Lipids (Alabaster, AL). The electrical current (I) was measured with a patch clamp amplifier (model BC-525C; Warner Instruments, Hamden, CT), digitized by a LabPC 1200 (National Instruments, Austin, TX), and analyzed by a personal computer with the help of Clampex (Axon) and WinWCP Strathclyde electrophysiology software designed by J. Dempster (University of Strathclyde). The current was low pass-filtered with a cut-off frequency of 100 Hz. In most experiments, a voltage of 80 mV (trans-negative) was applied to BLM to open the channels with Ag-AgCl electrodes placed directly into the cell. C-terminal colicin peptide (P178) or intact colicin E1 was added to, and defined, the cis side of the membrane. Agar-agar bridges connecting the electrodes and the membrane bathing solutions were used in selectivity measurements.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
DISCUSSION
 REFERENCES
 
Two States of the Colicin E1 Channel with a Very Different Conductance
Two conductance states of very different sizes are present in the recording of colicin E1 channel-forming domain (P178) in a DPhPC/decane membrane (Fig. 1A). Histograms of the channel current (Fig. 1B) and lifetime (Fig. 1C) show the high conductance state to be characterized by a single-channel conductance of 590 ± 10 pS (mean ± S.E., n = 13) and a lifetime of 120 ± 30 ms (single exponential fit, S.D.), whereas small channels have a conductance of 65 ± 2 pS (n = 13) and a lifetime of 4 ± 1 s (n = 18). The latter value was calculated as an average from a series of the recordings, because of the distortion of a histogram by simultaneous opening of several small channels. It is important to note that the high conductance state never appeared unless at least one small channel was already open (i.e. opening of the large (600-pS) channel required the formation and prior existence of the low conductance state). This implies that high conductance 600-pS channels are formed from 60-pS "precursor" channel. Both conductance states displayed the same voltage dependence; they were opened by a cis-positive voltage and closed when the sign of the voltage was reversed (Fig. 1D), which is characteristic for colicin-induced macroscopic current (51).

Conductance states of the two types, normal and high, with a single-channel conductance and lifetime similar to those of the 178-residue C-terminal channel-forming domain, P178, were also observed for intact colicin E1 (data not shown). It was shown previously that P178 or C-terminal peptides of similar size retain the channel-forming ability of the intact protein to form the "small" channels (50, 51, 58, 59).

Effect of Membrane Thickness on Colicin E1 Single Channels
Effect of Lipid Solvent—It is known that DPhPC/squalene membranes are solvent-free and substantially thinner than DPhPC/decane membranes (60, 61). As seen in Fig. 2A, a typical single-channel recording of P178 in a DPhPC/squalene membrane contains only low amplitude transitions corresponding to the 60-pS conductance state. This contrasts with the presence of 600-pS channels in DPhPC/decane membrane (Fig. 2B). Both recordings contained up to three active channels. Apparently, the activity of the 600-pS channel state observed in Fig. 2B was increased significantly compared with that reported in Fig. 1A. This can be attributed to the presence of three 60-pS channels in this recording, whereas only one channel was active in the data of Fig. 1A. Quantitative analysis of the open probability of the 600-pS conductance state (Po) under these conditions showed that Po was proportional to an average number of active 60-pS channels, m, estimated according to the equation, m = (1/{Lambda}) x ({Sigma}iIiMi/{Sigma}iMi), where Mi is the number of events with current Ii (Fig. 1B), and {Lambda} is the current amplitude of the 60-pS channel. This linear dependence (Po on m) confirms the conclusion that the opening of the 60-pS channel is a precondition for the opening of a high conductance state.

Effect of Lipid Chain Length—The appearance of a high conductance state in decane-containing membranes can be associated either with the presence of solvent or with an increase in bilayer thickness. To discriminate between these possibilities, single-channel measurements were performed with solvent-free membranes formed of lipids with fatty acyl tails of different lengths. As seen from the recordings depicted in Fig. 2 (C and D), both low and high conductance states of P178 appeared in DErPC (C22)/squalene membranes (Fig. 2D), whereas in DEcPC (C20)/squalene membranes (Fig. 2C) only the low conductance state was observed. These data support the idea that the formation of the high conductance state of colicin channels requires a sufficiently thick membrane. The thickness of a bilayer increases by 3-3.5 Å for each 2-carbon increase in the length of the hydrocarbon chains (62). The presence of a double bond with the cis orientation results in thinning of the bilayer by 2.5 Å. Table 1 summarizes the average conductance of the large and small colicin conductance states observed for different membrane lipid compositions. The channel activity of P178 was not observed in very thick DNvPC (C24)/squalene membranes (data not shown), which would confer a large hydrophobic mismatch.


View this table:
[in this window]
[in a new window]
  		TABLE 1
Average conductance (in pS) of the 60-pS/600-pS state in membranes of different composition

 


Figure 2
View larger version (28K):
[in this window]
[in a new window]
  		FIGURE 2.
Dependence of single-channel properties of P178 on the membrane solvent and the lipid chain length. The membranes were formed from a squalene solution of DPhPC (A), DEcPC (C), or DErPC (D) and from a decane solution of DPhPC (B), DOPC (E), or DEcPC (F). Other conditions were as in Fig. 1.

 
The presence of decane solvent down-shifted the acyl chain length boundary, from 20/22 to 18/20, between the small and large channels of P178. From the single-channel recordings of P178 in DOPC (C18)/decane (Fig. 2E) and DEcPC (C20)/decane (Fig. 2F) membranes and the corresponding data presented in Table 2, it is seen that the increase in membrane thickness led to an increase in both the open state probability of the high conductance state, from 0.25 ± 0.1 to 46 ± 4% and the single-channel lifetime from 100 ± 60 ms (n = 16) for C18 to 300 ± 200 ms (n = 12) for C20. No channel activity of the P178 channel domain was observed in DErPC (C22)/decane membranes (Table 2).


View this table:
[in this window]
[in a new window]
  		TABLE 2
Open state probability (in percent) of 600-pS channels in membranes of different lipid and solvent compositions The probability was normalized to the average number of open 60-pS channels (m) in the records, the number was estimated from the current histogram according to the equation, m = (1/{Lambda}) x ({Sigma}iIiMi/{Sigma}iMi) (see "Results"). Records were selected where m ranged from 1 to 3.

 
Although membranes formed from lipid dissolved in hexadecane contain solvent, the thickness of such solvent-containing membranes is intermediate between that of membranes formed from lipid solutions in squalene or decane (63, 64). Consistent with this finding, P178 showed the 600-pS high conductance state in DEcPC (C20)/hexadecane and DPhPC/hexadecane membranes. However, the open probability of this conductance state was smaller than in the corresponding decane membranes (Table 2). It should be noted that values of the open probability of the high conductance state presented in Table 2 were normalized to the mean number of open 60-pS channels, since it was established that the populations of the two kinds of conductance states are proportional to each other (data not shown). In summary, it was concluded that the occurrence of high conductance 600-pS channels of P178 is facilitated by increased membrane thickness.

Role of Spontaneous Curvature
Colicin E1-induced macroscopic current across BLM is sensitive to SC of the lipids comprising the membrane (43). Thus, it was important to determine how the SC regulates the colicin E1-induced currents on the single-channel level. Because preparation of planar membranes from lipids with a defined positive curvature is not experimentally feasible, comparisons of the effect of membranes of lipids with different curvature necessarily involve membranes with lipids of neutral or different degrees of negative curvature. By measuring the time course of the appearance of single channels after the addition of P178 to the bathing solution, we compared the kinetics of colicin insertion into membranes formed from lipids with different degrees of negative curvature (i.e. with relatively low (DOPC (C18)/decane) and high (DOPE (C18)/decane) negative SC (65)). The insertion proceeded much faster in the membrane with lipids having the smaller negative SC (data not shown). Because the binding of P178 to membranes is independent of SC (43), this result is ascribed to the increased channel-forming tendency of P178 in DOPC (C18)/decane membranes that have a more positive curvature compared with DOPE (C18)/decane membranes. Thus, negative SC inhibits the single-channel activity of colicin E1, in agreement with the previously observed effect of SC on the colicin-induced macroscopic current (43).

The open probability of 600-pS channels normalized to the mean number of open small channels is higher in DOPE than in DOPC (C18)/decane membranes (Table 2). Therefore, the increase in lipid-negative SC stimulates the opening of the 600-pS conductance state compared with that of 60 pS. The same conclusion can be made from the comparison of the recordings of P178 in DPhPC/decane membranes (high negative SC) (Fig. 2B) and DOPC/decane (low negative SC) (Fig. 2E) (66, 67). It should be noted that phosphatidylcholines (C16:0 and C18:1 cis) form membranes of similar thickness because of acyl chain bending caused by the presence of the double bond (62).

In contrast to the effect of the membrane thickness, the single-channel conductance and the lifetime (data not shown) of P178 channels are relatively insensitive to the SC. In line with these data, it was shown previously (43) that lysophosphatidylcholine and oleic acid, agents inducing positive and negative SC, respectively, do not produce a marked effect on the single-channel conductance and the lifetime of P178 channels.

Effect of Anionic Lipid on Small and Large Single Channels of P178
It is known that the macroscopic current across BLM induced by colicin E1 exhibits a bell-shaped dependence on the molar fraction of negatively charged lipids in the membrane and on surface potential (68). A study of the P178 single-channel parameters was carried out by varying the fraction of DPhPG in the membrane in the presence of a high ionic strength medium (1 M KCl) (Fig. 3, A and B). Whereas the conductance of the 60-pS state decreased 10-fold at 100% DPhPG compared with 100% DPhPC (Fig. 3A), the conductance of the 600-pS conductance state decreased only 2-fold under these conditions (Fig. 3B), The single-channel lifetime of the 600-pS state was 5-fold shorter when measured in 100% DPhPG than 100% DPhPC, and its open probability also decreased markedly.

Ion Selectivity of Colicin E1 Single Channels
The colicin E1-induced current across neutral planar bilayer membranes is known to exhibit anionic selectivity under acidic conditions (52, 69, 70). To characterize the properties of single P178 channels, their selectivity was measured through the zero current potential ({Delta}V) in the presence of a 10-fold gradient of KCl (1 M KCl and 0.1 M KCl on the cis and trans sides). Under these conditions, the permeability ratio for chloride and potassium ions, PCl/PK, can be calculated from the value of {Delta}V using the Goldman-Hodgkin-Katz equation. Fig. 4 presents the voltage dependence of single-channel conductance of both P178 states in DPhPC/hexadecane membranes in the presence of a 10-fold gradient of KCl. The crossing of the x axis by the linear fit to the data gives the value of {Delta}V. For the high conductance state, {Delta}V was equal to 37.1 ± 1.3 mV, and for the 60-pS channels, it was 25.6 ± 4.2 mV, which corresponds to permeability ratios, PCl/PK, of 7.2 ± 0.6 and 3.6 ± 0.8, respectively. Thus, the ion selectivities of 600- and 60-pS channel states are similar and also close to the literature data on the selectivity of the macroscopic current induced by colicin E1 across membranes formed from neutral lipids (52, 70). Small differences between the values of zero-current potentials could arise from the difference in the experimental conditions. For example, Raymond et al. (52) and Bullock (70) used buffer solutions containing several mM CaCl2.

Based on the nearly 10-fold difference between the values of the single-channel conductance of high and normal conductance states, one can expect a marked difference in their selectivity. An increase in channel size is often accompanied by a reduction in ion selectivity. It was shown that an increase in membrane thickness alters the swelling-dependent chloride channel selectivity from potassium to chloride ions (9). In the present case, both low and high P178 conductance states exhibit a similar small degree of chloride selectivity (Fig. 4).

As shown by Bullock (70), the ion selectivity of colicin E1 channels is sensitive to the pH of the bathing solution and the lipid content of the membrane. It was important to compare the ion selectivity in neutral membranes, described above, and in anionic membranes. Measurements of single-channel activity in DPhPG/hexadecane membranes in the presence of a 10-fold gradient of KCl were difficult due to the low conductance of small channels and the short lifetime of large channels under these conditions. However, it was possible to estimate the ion selectivity on the level of multichannel currents. {Delta}V for the macroscopic current induced by P178 across DPhPG/hexadecane membranes was equal to -25 ± 1 mV, which corresponds to a permeability ratio, PCl/PK, of 0.29 ± 0.02 (i.e. P178 channels showed cationic selectivity under these conditions). This is in qualitative agreement with the data of Raymond et al. (52), who observed that anionic lipids reduced the anionic preference of colicin E1 channels.


Figure 3
View larger version (14K):
[in this window]
[in a new window]
  		FIGURE 3.
Dependence of the single-channel conductance of small (A) and large (B) channels of P178 on the molar fraction of anionic (DPhPG) lipid. Membranes were formed from a decane solution of DPhPC/DPhPG. Other conditions were as in Fig. 1.

 


Figure 4
View larger version (22K):
[in this window]
[in a new window]
  		FIGURE 4.
Determination of zero current potentials of 60-pS (open circles) and 600-pS (closed circles) channels of P178. The cis- and trans-compartments contained 1.0 and 0.1 M KCl, respectively. Both solutions contained 10 mM potassium acetate, pH 4. The membrane was formed from a hexadecane solution of DPhPC. The inset shows the magnified region of small conductance values.

 
Estimation of the Size of 60- and 600-pS Channels
The diameter of an ion channel can be estimated from the dependence of the single-channel conductance on the size of nonelectrolytes present in the membrane bathing solutions (71-74). The effect of glycerol, PEG 300, and PEG 3350 on the conductance of small and large channels formed by P178 is shown (Figs. 5, A and B). The addition of glycerol to the bathing solution led to a decrease in the single-channel conductance of both large and small channels, whereas in the presence of PEG 300, only the conductance of large channels was diminished. PEG 3350 did not occlude the conductance of either large or small channels. From these data, it can be concluded that the diameter of the 600-pS channel is larger than that of the 60-pS channel. For quantitative estimation of the channel diameters, we measured the dependence on the hydrodynamic radius of nonelectrolyte molecules (R) of the ratio of the single-channel conductances measured in the presence of nonelectrolytes to the control conductance measured without nonelectrolytes ({Lambda}/{Lambda}0) (Fig. 5C). Curve 2, corresponding to the 600-pS channel, is shifted to the right with respect to curve 1, corresponding to the 60-pS channel, showing that the diameter of the higher conductance channel is larger. The maximal size of the channel is considered to be equal to the minimal size of nonelectrolytes that cannot enter the channel (75) and thus do not affect the single-channel conductance. Therefore, the maximum channel radius corresponds to the break in the slope of the dependence of {Lambda}/{Lambda}0 versus R. From curves 1 and 2 of Fig. 5C we obtained 12 and 16 Å for the diameters of 60- and 600-pS channels, respectively. The estimated diameter of the 60-pS channel is close to the diameter of the alamethicin channel, 12.6 Å, in its lowest conductance state (76).

Induction of Lipid "Flip-flop" by Colicin E1
The fluorescent phosphatidylcholine analog pyPC was employed to detect the occurrence of transbilayer lipid flip-flop. Membrane-incorporated pyPC displays two distinct peaks in the fluorescence spectrum, one originating from excited monomeric pyPC molecules and the other arising from excited dimeric (excimer) pyPC molecules (57). The transbilayer movement of pyPC can be estimated from the time course of the decrease in the ratio of the excimer to monomer fluorescence intensity signals (IE/IM). The addition of colicin E1 to DPhPC/DPhPG (70%/30%) liposomes containing pyPC localized exclusively in the external monolayer led to a decrease in the IE/IM ratio, reflecting the transfer of the fluorescent analog to the internal monolayer in a dose-dependent manner (Fig. 6, a-d). From the calibration dependence of the IE/IM ratio on the degree of pyPC transbilayer distribution (q) for 5 mol % pyPC (57), it can be calculated that the addition of 250 nM P178 led to a nearly equal distribution of pyPC between the two membrane leaflets (q close to 0.5). It is noteworthy that the transbilayer movement of pyPC was not detected if 3 µM gramicidin was added to the liposomes (data not shown). As shown previously, P178 at a similar protein/lipid ratio induces pore formation, leading to carboxyfluorescein leakage under identical conditions (43). Thus, these results imply that liposome permeabilization and lipid transbilayer relocation arising from insertion of P178 are mechanistically related phenomena.


Figure 5
View larger version (13K):
[in this window]
[in a new window]
  		FIGURE 5.
Effect of addition of glycerol and PEGs of different molecular weights on the currents through the small (A) and large (B) P178 channels. From left to right, control (no addition), glycerol, PEG 300, and PEG 3350. Note that B shows the difference between the currents through large and small channels. C, the dependence of the ratio of the single-channel conductance measured in the presence of a nonelectrolyte to that in the control ({Lambda}/{Lambda}0) for small and large P178 channels on the hydrodynamic radius (R) of a nonelectrolyte molecule. The buffer solution was 10 mM beta-alanine, 1.77 M KCl, and 20% (w/w) nonelectrolyte, pH 4.0. Other conditions were as in Fig. 1.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
REFERENCES
 
Two Conductance States of the Colicin E1 Channel: A "Giant" Conductance State Appears in Thick Bilayer Membranes—The present study demonstrated two stable open states of colicin E1 channel with single-channel conductances of 60 and 600 pS. The high conductance state was characteristic only for thicker membranes. Its open probability increased with increasing length of the fatty acyl chains. The activity of small colicin channels exhibited the opposite tendency. These data, together with the observation that large channels are dependent upon the prior appearance of small channels, imply that the colicin channel structure is dependent upon lipid thickness.

The observation of different conductance states of colicin channels might imply an oligomeric nature of the channel. However, there is a large body of evidence in the literature proving that colicin channels are formed by protein monomers (77-80). Based on the fact that the 600-pS channel opened when at least one 60-pS channel was open and also on the rapid kinetics of these transitions (opening and closing times were <10 ms, beyond the time resolution of the present instrumentation), it was inferred that the two kinds of channels do not reflect protein oligomerization but rather correspond to different structure states of a single colicin molecule. This conclusion is also supported by the difference in the diameters of small and large channels as determined by the PEG-induced reduction of the single-channel conductances (Fig. 5) as well as by the different selectivities of the two kinds of channels (Fig. 4).


Figure 6
View larger version (40K):
[in this window]
[in a new window]
  		FIGURE 6.
Time course of the transbilayer diffusion of pyPC in DPhPC/DPhPG liposomes induced by the addition (marked by an arrow) of 50 nM (b), 100 nM (c), and 250 nM (d) of P178. Curve a is the control. A decrease in the ratio of IE (intensity of excimer emission) to IM (intensity of monomer emission) reflects transbilayer diffusion of pyPC. The time course of the decrease in the IE/IM ratio measured after the addition of 500 nM P178 coincided with curve d.

 
The variation in the published values of colicin E1 single-channel conductance is centered around channels of small size (7-60 pS) (34, 50-54) and can be associated with the differences in anionic lipid content on which the small channel conductance is strongly dependent (Fig. 3A). Alternatively, this range of channel sizes for the C-terminal domain of colicin Ia has been attributed to channels made of three or four transmembrane helices (37). However, the absence of large channels in the studies cited can hardly be ascribed to an effect of anionic lipid, since the magnitude of single-channel conductance of large channels is only moderately sensitive to their presence (Fig. 3B).

A hypothesis that involves additional {alpha}-helices in the formation of the 600-pS conductance state (Fig. 7B) from the 60-pS state (Fig. 7A) would be consistent with the observed increase in the anionic selectivity of this high conductance state (Fig. 4). It is implied that the protein rearrangement would cause an increase in the pore wall of the number of positively charged amino acids, which are located on all {alpha}-helices of colicin E1 pore domain except for the hydrophobic helices VIII and IX. It should be pointed out that the selectivity measurements were performed with neutral lipids. Therefore, the involvement of negatively charged lipids in both the pore formation and the modulation of the near membrane concentrations of the chloride anions resulting from changes in the surface potential of lipid bilayer does not affect the conclusion that the 600-pS state has increased anionic selectivity.

Effect of Spontaneous Curvature—The inhibitory effect of lipids having negative SC on the activity of single channels of P178 confirms previous data on the influence of SC on the colicin-induced macroscopic current (43). It is of importance that the macroscopic measurements were made with thin, solvent-free membranes, where the activity of 600-pS channels is negligible, and 60-pS channels make the major contribution to the macroscopic current. Thus, it can be concluded that high negative SC inhibits the activity of the 60-pS state of colicin E1. The 600-pS state observed in thick membranes exhibited the opposite dependence on SC (Table 2). Its open probability with respect to the 60-pS state increased in membranes with higher negative SC.


Figure 7
View larger version (90K):
[in this window]
[in a new window]
  		FIGURE 7.
Models of the formation of 60-pS (A) and 600-pS (B) channel states by colicin E1 channel-forming domain. Lipids, which participate in the pore wall formation, are marked in green; two hydrophobic TM {alpha}-helices (VIII-IX) are marked in gray. The 600-pS channel is lengthened in the transmembrane direction as compared with the 60-pS channel due to an increase in the length of TM {alpha}-helices accompanied by a reduction of the number of {alpha}-helices (eight in B and nine in A).

 
The Dependence of the Single-channel Conductance on the Surface Charge of the Membrane—It might be suggested that the decrease in the single-channel conductance of the 60-pS channel with increasing DPhPG content could be accounted for by a decrease in chloride ion concentration at the membrane surface arising from the negative surface potential generated by anionic lipid. Actually the single-channel conductance of gramicidin A in 1 M CsCl is higher by 30% in the anionic phophatidylserine membrane as compared with that in neutral phosphatidylcholine membrane (81, 82). However, Aguilella and Bezrukov (83) have shown that the variation in the lipid surface charge does not change the conductance of alamethicin channels in 1 M NaCl. In addition, OmpF conductance is insensitive to the lipid surface charge in 1 M KCl (84), whereas that of {alpha}-hemolysin exhibits only a slight dependence on the surface charge under the high salt conditions (85).3 These data were rationalized by theoretical considerations taking into account the small Debye length at high ionic strength and the size of the lipid-free space occupied by the channel (83). Taking into account the estimated size of the 60-pS channel (see "Estimation of the Size of 60- and 600-pS Channels"), which is quite close to that of alamethicin (86), the observed strong dependence of the colicin E1 conductance on the lipid charge in 1 M KCl initially seems surprising.

This strong dependence can be understood if one assumes direct involvement of lipids in the pore wall formation, as suggested by a toroidal pore model for the colicin channel (Fig. 7, A and B). In fact, the appearance of one positive charge in the alamethicin peptide results in a substantial (about 20%) decrease in the conductance in 1 M KCl (87). Likewise, the conductances of OmpF (88) and {alpha}-hemolysin (89) channels are sensitive to the mutations introducing electrical charges into the wall region of the channels even under high ionic strength conditions. Therefore, the 10-fold decrease in the conductance of the 60-pS state (Fig. 3A) and the 2-fold decrease in the conductance of the 600-pS state (Fig. 3B) at 100% DPhPG compared with those at 100% DPhPC apparently result from the appearance in the channel walls of many negative charges that belong to lipid head groups.

As shown under "Results," P178 channels that are anion-selective in the DPhPC membrane became cation-selective in the membrane formed from 100% DPhPG. This difference in the selectivity of colicin channels in neutral and anionic membranes is also consistent with a model of direct involvement of lipids in the pore wall (90). In fact, the ion selectivity of alamethicin (87) and OmpF (87, 91) channels can be modulated by mutation of charged amino acid residues located in the pore walls, in agreement with the local electrostatics. The decrease in the conductance of both small and large channels upon increasing the anionic lipid content can be reconciled with the change in the ion selectivity if one assumes that channel permeability is reduced in anionic compared with neutral membranes due to alteration of channel geometry. Actually, intercalation of negatively charged DPhPG molecules between positively charged {alpha}-helices would decrease the electrostatic repulsion of the latter, thereby decreasing the pore diameter. This assumption is in agreement with calculations of Zemel et al. (92) showing that pore geometry can change considerably if charged species are involved in the pore formation.

Toroidal Pore Model and the Colicin E1 Channel—As discussed above, there is extensive documentation that the effect of lipid SC on pore formation can be a consequence of toroidal structure of the pore. Previously (43, 44), we proposed the applicability of a toroidal model for the mechanism of the colicin E1 pore formation, which was recently supported by the data of Musse et al. (93). Presently, this model is supported by four additional lines of evidence: (i) the ability of P178 to induce lipid flip-flop (Fig. 6); (ii) the essential dependence of the single-channel conductance on the lipid surface charge under high salt conditions (Fig. 3); (iii) the effect of lipid surface charge on the ion selectivity of the channels; and (iv) the observation of the "large" channels. As was first pointed out by Slatin (82), involvement of lipids in the pore formation of colicin E1 seems to be the way to overcome the problem of a monomeric channel having a rather large size. Thus, one of the consequences of the toroidal pore model for the mechanism of colicin channel formation is that it explains the surprising fact that a colicin channel having a diameter of at least 8 Å (52) could be formed by a smaller number of helices than would be required in the absence of lipid.

A common and newly proposed feature of the models in Fig. 7, A and B, is that they are both toroidal in the sense that lipids participate in their wall structure. Lipids with positive SC favor formation of the small 60-pS pore. We assume that the opposite SC dependence of the 600-pS channel (Table 2) does not contradict the toroidal pore model, because the stimulating effect of these lipids on the 600-pS channel state, due to hydrophobic mismatch (94), may overcome their unfavorable influence on the formation of toroidal lipidic pores. Moreover, there are examples of channels, believed to be of toroidal type, that demonstrate an increased activity caused by negative SC (e.g. see Ref. 95). A variable dependence of the pore-forming ability on SC has been also obtained for apoptotic proteins Bax and Bid (42, 90, 96-99).

The mechanism by which the number and length of TM {alpha}-helices are increased upon the formation of the 600-pS conductance state (Fig. 7B) from the 60-pS state (Fig. 7A) is not known. The experiments of Kienker et al. (37) may provide insight into this process. It is hypothesized that the voltage-induced translocation across the membrane of a large portion of the colicin Ia molecule (residues 474-541) also includes the possibility of transfer into the membrane bilayer of helices that are part of the translocated segment.

In summary, a "large" conductance state of colicin E1 channel has been observed and extensively characterized for the first time. Whereas this conductance state is anion-selective and has a voltage dependence similar to that of the "small" channel, the two channel states show opposite dependences on the membrane thickness and curvature. These differences between "large" and "small" channel states can be due to differences in the number, lengths, and arrangement of the transmembrane {alpha}-helices that define the channel, reflecting two distinct conformations of the colicin channel. The results of the present study support a toroidal mechanism of colicin E1 pore formation. Remarkably, our data showed that at least in some cases, a toroidal (protein-lipid) pore can be stable enough to exhibit well defined channel states, the properties of which are very sensitive to the lipid composition of the membrane. This stability may imply a unique structure of the colicin E1 channel and possibly those of pore-forming colicins, among membrane-active proteins forming toroidal pores.


   FOOTNOTES
 
* This work was supported by National Institutes of Health Fogarty Award TW01235 and Grant GM-18457 (to W. A. C.) and by Russian Foundation for Basic Research Grant 06-04-48523 (to Y. N. A.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Back

1 To whom correspondence should be addressed: Belozersky Institute of Physico-Chemical Biology, Moscow State University, Moscow 119992, Russia. Fax: 70-95-939-31-81; E-mail: antonen{at}genebee.msu.ru.

2 The abbreviations used are: SC, spontaneous curvature; P178, 178-residue C-terminal colicin E1 polypeptide, prepared by thermolysin proteolysis; BLM, bilayer lipid membrane; DNvPC (C24:1), 1,2-dinervonoyl-sn-glycero-3-phosphocholine; DErPC (C22:1) 1,2-dierucoyl-sn-glycero-3-phosphocholine; DEcPC (C20:1), 1,2-dieicosenoyl-sn-glycero-3-phosphocholine; DOPC (C18:1 cis), 1,2-dioleoyl-sn-glycero-3-phosphocholine; DOPE (C18), 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine; DPhPC (C16), 1,2-diphytanoyl-sn-glycero-3-phosphocholine; DPhPG (C16), 1,2-diphytanoyl-sn-glycero-3-[phospho-rac-(1-glycerol)] (sodium salt); PEG, polyethylene glycol; pyPC, 1-lauroyl-2-(1'pyrenebutyroyl)-sn-glycero-3-phosphocholine; TM, transmembrane; pS, picosiemens. Back

3 O. V. Krasilnikov, personal communication. Back


   ACKNOWLEDGMENTS
 
We are greatly indebted to M. Zhalnina for colicin preparation and Dr. P. Müller (Humboldt University, Berlin) for providing the pyrene-labeled lipid.



   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Killian, J. A. (1998) Biochim. Biophys. Acta 1376, 401-415[Medline] [Order article via Infotrieve]
  2. Andersen, O. S., Nielsen, C., Maer, A. M., Lundbaek, J. A., Goulian, M., and Koeppe, R. E. (1999) Methods Enzymol. 294, 208-224[Medline] [Order article via Infotrieve]
  3. Harroun, T. A., Heller, W. T., Weiss, T. M., Yang, L., and Huang, H. W. (1999) Biophys. J. 76, 3176-3185[Abstract/Free Full Text]
  4. Hamill, O. P., and Martinac, B. (2001) Physiol. Rev. 81, 685-740[Abstract/Free Full Text]
  5. Mobashery, N., Nielsen, C., and Andersen, O. S. (1997) FEBS Lett. 412, 15-20[CrossRef][Medline] [Order article via Infotrieve]
  6. Martinac, B., and Hamill, O. P. (2002) Proc. Natl. Acad. Sci. U. S. A. 99, 4308-4312[Abstract/Free Full Text]
  7. Williamson, I. M., Alvis, S. J., East, J. M., and Lee, A. G. (2002) Biophys. J. 83, 2026-2038[Abstract/Free Full Text]
  8. Yuan, C., O'Connell, R. J., Feinberg-Zadek, P. L., Johnston, L. J., and Treistman, S. N. (2004) Biophys. J. 86, 3620-3633[Abstract/Free Full Text]
  9. Garavaglia, M., Dopinto, S., Ritter, M., Furst, J., Saino, S., Guizzardi, F., Jakab, M., Bazzini, C., Vezzoli, V., Dossena, S., Rodighiero, S., Sironi, C., Botta, G., Meyer, G., Henderson, R. M., and Paulmichl, M. (2004) Cell. Physiol. Biochem. 14, 231-240[CrossRef][Medline] [Order article via Infotrieve]
  10. Sparr, E., Ash, W. L., Nazarov, P. V., Rijkers, D. T., Hemminga, M. A., Tieleman, D. P., and Killian, J. A. (2005) J. Biol. Chem. 280, 39324-39331[Abstract/Free Full Text]
  11. Betanzos, M., Chiang, C. S., Guy, H. R., and Sukharev, S. (2002) Nat. Struct. Biol. 9, 704-710[CrossRef][Medline] [Order article via Infotrieve]
  12. Perozo, E., Kloda, A., Cortes, D. M., and Martinac, B. (2002) Nat. Struct. Biol. 9, 696-703[CrossRef][Medline] [Order article via Infotrieve]
  13. Bezrukov, S. M. (2000) Curr. Opin. Colloid Interface Sci. 5, 237-243[CrossRef]
  14. van den Brink-van der Laan, Killian, J. A., and de Kruijff, B. (2004) Biochim. Biophys. Acta 1666, 275-288[Medline] [Order article via Infotrieve]
  15. Wiggins, P., and Phillips, R. (2004) Proc. Natl. Acad. Sci. U. S. A. 101, 4071-4076[Abstract/Free Full Text]
  16. Sunshine, C., and McNamee, M. G. (1992) Biochim. Biophys. Acta 1108, 240-246[Medline] [Order article via Infotrieve]
  17. Powl, A. M., East, J. M., and Lee, A. G. (2003) Biochemistry 42, 14306-14317[CrossRef][Medline] [Order article via Infotrieve]
  18. Park, J. B., Kim, H. J., Ryu, P. D., and Moczydlowski, E. (2003) J. Gen. Physiol. 121, 375-397[Abstract/Free Full Text]
  19. Tillman, T. S., and Cascio, M. (2003) Cell. Biochem. Biophys. 38, 161-190[CrossRef][Medline] [Order article via Infotrieve]
  20. Valiyaveetil, F. I., Zhou, Y., and MacKinnon, R. (2002) Biochemistry 41, 10771-10777[CrossRef][Medline] [Order article via Infotrieve]
  21. Alvis, S. J., Williamson, I. M., East, J. M., and Lee, A. G. (2003) Biophys. J. 85, 3828-3838[Abstract/Free Full Text]
  22. Hsu, S. T., Breukink, E., Tischenko, E., Lutters, M. A., de Kruijff, B., Kaptein, R., Bonvin, A. M., and van Nuland, N. A. (2004) Nat. Struct. Mol. Biol. 11, 963-967[CrossRef][Medline] [Order article via Infotrieve]
  23. Breukink, E., van Heusden, H. E., Vollmerhaus, P. J., Swiezewska, E., Brunner, L., Walker, S., Heck, A. J., and de Kruijff, B. (2003) J. Biol. Chem. 278, 19898-19903[Abstract/Free Full Text]
  24. Palmer, M., Harris, R., Freytag, C., Kehoe, M., Tranum-Jensen, J., and Bhakdi, S. (1998) EMBO J. 17, 1598-1605[CrossRef][Medline] [Order article via Infotrieve]
  25. Gilbert, R. J. (2005) Structure 13, 1097-1106[Medline] [Order article via Infotrieve]
  26. Braun, V., Pilsl, H., and Gross, P. (1994) Arch. Microbiol. 161, 199-206[Medline] [Order article via Infotrieve]
  27. James, R., Penfold, C. N., Moore, G. R., and Kleanthous, C. (2002) Biochimie (Paris) 84, 381-389
  28. Gould, J. M., and Cramer, W. A. (1977) J. Biol. Chem. 252, 5491-5497[Free Full Text]
  29. Schein, S. J., Kagan, B. L., and Finkelstein, A. (1978) Nature 276, 159-163[CrossRef][Medline] [Order article via Infotrieve]
  30. Lakey, J. H., and Slatin, S. L. (2001) Curr. Top. Microbiol. Immunol. 257, 131-161[Medline] [Order article via Infotrieve]
  31. Zakharov, S. D., and Cramer, W. A. (2002) Biochim. Biophys. Acta 1565, 333-346[Medline] [Order article via Infotrieve]
  32. Zakharov, S. D., Lindeberg, M., Griko, Y., Salamon, Z., Tollin, G., Prendergast, F. G., and Cramer, W. A. (1998) Proc. Natl. Acad. Sci. U. S. A. 95, 4282-4287[Abstract/Free Full Text]
  33. Lindeberg, M., Zakharov, S. D., and Cramer, W. A. (2000) J. Mol. Biol. 295, 679-692[CrossRef][Medline] [Order article via Infotrieve]
  34. Cramer, W. A., Heymann, J. B., Schendel, S. L., Deriy, B. N., Cohen, F. S., Elkins, P. A., and Stauffacher, C. V. (1995) Annu. Rev. Biophys. Biomol. Struct. 24, 611-641[CrossRef][Medline] [Order article via Infotrieve]
  35. Slatin, S. L., Qiu, X. Q., Jakes, K. S., and Finkelstein, A. (1994) Nature 371, 158-161[CrossRef][Medline] [Order article via Infotrieve]
  36. Jakes, K. S., Kienker, P. K., and Finkelstein, A. (1999) Q. Rev. Biophys. 32, 189-205[CrossRef][Medline] [Order article via Infotrieve]
  37. Kienker, P. K., Jakes, K. S., and Finkelstein, A. (2000) J. Gen. Physiol. 116, 587-598[Abstract/Free Full Text]
  38. Kyte, J., and Doolittle, R. F. (1982) J. Mol. Biol. 157, 105-132[CrossRef][Medline] [Order article via Infotrieve]
  39. Wimley, W. C., and White, S. H. (2000) Biochemistry 39, 4432-4442[CrossRef][Medline] [Order article via Infotrieve]
  40. Matsuzaki, K., Sugishita, K., Ishibe, N., Ueha, M., Nakata, S., Miyajima, K., and Epand, R. M. (1998) Biochemistry 37, 11856-11863[CrossRef][Medline] [Order article via Infotrieve]
  41. Malev, V. V., Schagina, L. V., Gurnev, P. A., Takemoto, J. Y., Nestorovich, E. M., and Bezrukov, S. M. (2002) Biophys. J. 82, 1985-1994[Abstract/Free Full Text]
  42. Basanez, G., Sharpe, J. C., Galanis, J., Brandt, T. B., Hardwick, J. M., and Zimmerberg, J. (2002) J. Biol. Chem. 277, 49360-49365[Abstract/Free Full Text]
  43. Sobko, A. A., Kotova, E. A., Antonenko, Y. N., Zakharov, S. D., and Cramer, W. A. (2004) FEBS Lett. 576, 205-210[CrossRef][Medline] [Order article via Infotrieve]
  44. Zakharov, S. D., Kotova, E. A., Antonenko, Y. N., and Cramer, W. A. (2004) Biochim. Biophys. Acta 1666, 239-249[Medline] [Order article via Infotrieve]
  45. Matsuzaki, K., Murase, O., Fujii, N., and Miyajima, K. (1996) Biochemistry 35, 11361-11368[CrossRef][Medline] [Order article via Infotrieve]
  46. Ludtke, S. J., He, K., Heller, W. T., Harroun, T. A., Yang, L., and Huang, H. W. (1996) Biochemistry 35, 13723-13728[CrossRef][Medline] [Order article via Infotrieve]
  47. Epand, R. M. (1998) Biochim. Biophys. Acta 1376, 353-368[Medline] [Order article via Infotrieve]
  48. Shai, Y. (1999) Biochim. Biophys. Acta 1462, 55-70[Medline] [Order article via Infotrieve]
  49. Lee, M. T., Hung, W. C., Chen, F. Y., and Huang, H. W. (2005) Biophys. J. 89, 4006-4016[Abstract/Free Full Text]
  50. Cleveland, M. V., Slatin, S., Finkelstein, A., and Levinthal, C. (1983) Proc. Natl. Acad. Sci. U. S. A. 80, 3706-3710[Abstract/Free Full Text]
  51. Bullock, J. O., Cohen, F. S., Dankert, J. R., and Cramer, W. A. (1983) J. Biol. Chem. 258, 9908-9912[Abstract/Free Full Text]
  52. Raymond, L., Slatin, S. L., and Finkelstein, A. (1985) J. Membr. Biol. 84, 173-181[CrossRef][Medline] [Order article via Infotrieve]
  53. Bishop, L. J., Cohen, F. S., Davidson, V. L., and Cramer, W. A. (1986) J. Membr. Biol. 92, 237-245[CrossRef][Medline] [Order article via Infotrieve]
  54. Liu, Q. R., Crozel, V., Levinthal, F., Slatin, S., Finkelstein, A., and Levinthal, C. (1986) Proteins 1, 218-229[CrossRef][Medline] [Order article via Infotrieve]
  55. Zhang, Y. L., and Cramer, W. A. (1992) Protein Sci. 1, 1666-1676[Abstract]
  56. Bishop, L. J., Bjes, E. S., Davidson, V. L., and Cramer, W. A. (1985) J. Bacteriol. 164, 237-244[Abstract/Free Full Text]
  57. Müller, P., Schiller, S., Wieprecht, T., Dathe, M., and Herrmann, A. (2000) Chem. Phys. Lipids 106, 89-99[CrossRef][Medline] [Order article via Infotrieve]
  58. Dankert, J. R., Uratani, Y., Grabau, C., Cramer, W. A., and Hermodson, M. (1982) J. Biol. Chem. 257, 3857-3863[Abstract/Free Full Text]
  59. Ohno-Iwashita, Y., and Imahori, K. (1982) J. Biol. Chem. 257, 6446-6451[Abstract/Free Full Text]
  60. White, S. H. (1978) Biophys. J. 23, 337-347[Abstract/Free Full Text]
  61. Uhrikova, D., Kucerka, N., Islamov, A., Kuklin, A., Gordeliy, V., and Balgavy, P. (2003) Biochim. Biophys. Acta 1611, 31-34[Medline] [Order article via Infotrieve]
  62. Lewis, B. A., and Engelman, D. M. (1983) J. Mol. Biol. 166, 211-217[Medline] [Order article via Infotrieve]
  63. Benz, R., Frohlich, O., Lauger, P., and Montal, M. (1975) Biochim. Biophys. Acta 394, 323-334[Medline] [Order article via Infotrieve]
  64. Chernyshev, A., Armstrong, K. M., and Cukierman, S. (2003) Biophys. J. 84, 238-250[Abstract/Free Full Text]
  65. Chen, Z., and Rand, R. P. (1997) Biophys. J. 73, 267-276[Abstract/Free Full Text]
  66. Hung, W. C., Chen, F. Y., and Huang, H. W. (2000) Biochim. Biophys. Acta 1467, 198-206[Medline] [Order article via Infotrieve]
  67. Haque, M. E., and Lentz, B. R. (2004) Biochemistry 43, 3507-351758[CrossRef][Medline] [Order article via Infotrieve]
  68. Zakharov, S. D., Rokitskaya, T. I., Shapovalov, V. L., Antonenko, Y. N., and Cramer, W. A. (2002) Proc. Natl. Acad. Sci. U. S. A. 99, 8654-8659[Abstract/Free Full Text]
  69. Shirabe, K., Cohen, F. S., Xu, S., Peterson, A. A., Shiver, J. W., Nakazawa, A., and Cramer, W. A. (1989) J. Biol. Chem. 264, 1951-1957[Abstract/Free Full Text]
  70. Bullock, J. O. (1992) J. Membr. Biol. 125, 255-271[Medline] [Order article via Infotrieve]
  71. Krasilnikov, O. V., Sabirov, R. Z., Ternovsky, V. I., Merzliak, P. G., and Muratkhodjaev, J. N. (1992) FEMS Microbiol. Immunol. 5, 93-100