Connexin-32: A Protein That Forms Channels through One or Two

Downloaded by UNIV OF CALIFORNIA SAN DIEGO on December 29, 2015 | http://pubs.acs.org Publication Date: May 5, 1994 | doi: 10.1021/ba-1994-0235.ch010

10 Connexin-32: A Protein That Forms Channels through One or Two Membranes Andrew L. Harris Thomas C. Jenkins Department of Biophysics, The Johns Hopkins University, 3400 North Charles Street, Baltimore, MD 21218

Connexin protein, which forms gated channels through closely apposed cell membranes ("gap junction channels"), also forms channels in single membranes. The mechanisms by which single- and double-membraneconnexin channels are formed, gated, and regulated are of biophysical interest, yet are largely unknown. Biophysical studies have been hindered by inaccessibility of cellular connexin channels and by the unique problems of connexin reconstitution. Connexin-32 from isolated gap junctions and from monoclonal immunoaffinity purification from plasma membrane forms large, dynamically gated channels in liposomes and planar bilayers. These connexin channels may be single hemichannels—the subunits that span a single membrane in situ—and are accessible for detailed biophysical study. Implications and possibilities for future studies of permeation, gating, and modulation are discussed.

T F H E STUDY O F PROTEIN FUNCTION is increasingly focused o n the interac tions between p r o t e i n structure, electrochemical potentials, a n d molecular m o t i o n . F o r m e m b r a n e proteins, insertion through a p h o s p h o l i p i d bilayer imparts a defined orientation to the p r o t e i n a n d a specified relation between the macroscopic electrical field a n d the axis o f the molecule n o r m a l to the membrane. These constraints can b e exploited to obtain detailed information about m e m b r a n e p r o t e i n function a n d dynamics, particularly f o r proteins that f o r m aqueous pores through membranes. T h e ability to r e c o r d currents that 0065-2393/94/0235-0197$09.26/0 © 1994 American Chemical Society

In Biomembrane Electrochemistry; Blank, Martin, et al.; Advances in Chemistry; American Chemical Society: Washington, DC, 1994.

198

BIOMEMBRANE ELECTROCHEMISTRY


flow through individual channel molecules [in cell membranes ( I ) a n d i n planar p h o s p h o l i p i d bilayers ( 2 , 3)] a n d so to observe directly transitions between conductance states provides precise functional information that is unavailable f o r other proteins. M e m b r a n e i o n channels, therefore, afford a u n i q u e opportunity to explore h o w physical a n d electrochemical forces inter act w i t h the s t r u c t u r e - f u n c t i o n o f proteins, particularly w h e n amino a c i d sequence a n d transmembrane topology are k n o w n . T h i s chapter reviews some o f the molecular biophysical questions that are raised b y the properties o f o n e class o f c h a n n e l - f o r m i n g proteins, connexin, a n d that may b e addressed through the study o f connexin i n reconsti t u t e d m e m b r a n e systems. T h e first section introduces issues o f biophysical interest a n d provides b a c k g r o u n d information about connexin. T h e second section discusses the prospects f o r utilizing reconstituted systems to study the key questions, f o l l o w e d b y a b r i e f review o f data f r o m o u r laboratory. T h e final sections evaluate the findings a n d discuss future studies.

The Gap Junction Channel, Connexin Protein, and Their Importance for Electrochemical Studies I n tissues, connexin p r o t e i n composes structures called gap junction channels. E a c h junctional channel forms a w i d e , dynamically gated aqueous pore through two closely apposed c e l l membranes ( F i g u r e 1) (4-6). Solutes u p to ~ 14 Â across c a n diffuse f r o m c e l l to c e l l through the pores without leakage to the extracellular environment. T h e junctional channel is f o r m e d b y t w o end-to-end subunits (called hemichanneh o r connexons) that each span a single c e l l m e m b r a n e . T h e h e m i c h a n n e l is c o m p o s e d o f six connexin monomers arranged around a central aqueous pore. E a c h connexin m o n o m e r spans the plasma m e m b r a n e f o u r times, w i t h C - a n d N - t e r m i n i i n the cytoplasm a n d t w o extracellular loops (7-9; reviewed i n reference 10). Connexins are a family o f closely related proteins o f w h i c h a steadily increasing n u m b e r have b e e n c l o n e d a n d sequenced ( I I , 12). T h e amino acid sequences o f connexins are highly homologous to one another i n transmembrane a n d extracellular domains, b u t highly variable i n cytoplasmic domains ( I I ) . G a p j u n c t i o n channels provide pathways f o r direct current flow between cells a n d f o r regulated intercellular movement o f important cellular signaling molecules, i n c l u d i n g cyclic nucleotides ( 5 , 13-17). Consequently, gap j u n c tion channels are thought to play crucial regulatory roles i n cell biology, development,,and physiology (18-22). C o n n e x i n p r o t e i n is u n i q u e i n f o r m i n g a gated pore through two m e m branes. T w o questions particular to this structure are (1) h o w is the junctional channel assembled (e.g., h o w d o the t w o hemichannels find each other a n d interact to f o r m a stable structure) a n d (2) what is the interaction between



10.

HARRIS

Connexin-32

199

Figure 1. Schematic drawings of connexin channels. The upper figure is an edge-on view of the junctional channel in situ that spans two plasma membranes (gray areas) across extracellular space. The middle figure shows a single hemichannel, the subunit of the junctional channel that spans each membrane. The lower figure is an end-on view of a hemichannel that shows the arrangement of six connexin monomers around a central, water-filled pore. the two hemichannels d u r i n g gating transitions a n d regulatory processes? Analogous questions apply to other p r o t e i n complexes that mediate signaling across d o u b l e - m e m b r a n e structures, such as the ryanodine r e c e p t o r - d i h y d r o pyridine receptor complex i n muscle ( 2 3 ) . T h e s e questions, a n d others that pertain specifically to i o n channels, are i n t r o d u c e d i n succeeding text. Junctional C h a n n e l Structure(s). A t regions o f c e l l contact, j u n c tional channels are f o u n d i n two-dimensional arrays that c a n contain t h o u sands o f channels. T h e arrays vary f r o m loose collections o f channels to pseudocrystalhne two-dimensional lattices. A p o r t i o n o f the most regular arrays c a n be isolated b y techniques that select for m e m b r a n e structures that are resistant to severe chaotropic treatments such as ionic detergents o r prolonged exposure to highly alkaline conditions ( p H 12.5) (24-27). T h e highly regular arrays o f channels obtained have b e e n studied i n detail w i t h electron a n d X - r a y diffraction techniques (reviewed i n reference 28), a n d


200



recently b y atomic force microscopy (29), to reveal structural features. These studies established the hexameric structure o f the h e m i c h a n n e l a n d p r o v i d e d two possible structural correlates o f gating. I n one view, the pore is o c c l u d e d at each cytoplasmic e n d b y a p r o t e i n mass that can lift away f r o m the p o r e — a trap door system ( 3 0 ) . A n o t h e r v i e w is that the connexin monomers are rigid rods that undergo a torsional tilt, sliding against each other a n d twisting to close the l u m e n o f the pore at the cytoplasmic e n d — s o m e w h a t analogous to an iris diaphragm (31, 32).

Functional Hemichannels in Cells. A g r o w i n g b o d y o f c i r c u m stantial evidence indicates that single hemichannels are functional channels i n plasma membranes. T h e r e is indirect evidence for the existence o f large channels f o r m e d b y connexin i n the plasma m e m b r a n e o f a macrophage c e l l line (33). Physiological data strongly suggest that hemichannels mediate a gated plasma m e m b r a n e conductance i n teleost horizontal cells (34). W h e n Xenopus oocytes are injected w i t h m R N A c o d i n g for connexin-46 (but not other connexins), a voltage-gated conductance a n d a permeability to the dye L u c i f e r Y e l l o w are i n d u c e d i n the plasma m e m b r a n e (35, 36). T h e s e findings indicate that plasma m e m b r a n e connexin channels have n o r m a l physiological functions, a n d they also may afford access to hemichannels for certain types of electrophysiological studies. Permeability. P e r m e a t i o n through most i o n channels is described b y barrier models based o n E y r i n g rate theory ( 3 7 ) . P e r m e a t i o n through the junctional pore, however, might be more accurately described b y electrodiffusion (38) because the junctional pore is sufficiently w i d e that atomic ions may pass one another. L a r g e r permeants may interact w i t h the walls o f the pore (39, 40) a n d their deviation f r o m electrodiffusive fluxes may reveal details about the structure o f the pore. Voltage-clamp studies o f c o u p l e d cells typically show a b r o a d distribution o f unitary c h a n n e l conductances, ranging f r o m ~ 20 pS to several h u n d r e d picosiemens (cf. references 41-44; for review, see reference 45). Some o f this variability is u n d o u b t e d l y due to the presence o f varying a n d m u l t i p l e connexins, but even i n junction-incompetent cells transfected w i t h c D N A for a single connexin, more than one size can be observed (46). U n i t a r y conductance might be m o d u l a t e d b y factors i n the cellular m i l i e u i n c l u d i n g protein kinases (47). A relation between differences i n unitary conductance a n d selectivity a m o n g large permeants has yet to be established. Transjunctional Voltage Sensitivity. I n most tissues w h e r e j u n c tional conductance has b e e n examined, it is sensitive to transjunctional voltage, membrane potential, or b o t h . F u r t h e r m o r e , i n a single j u n c t i o n there can be more than one kinetically distinct process for a given type o f voltage sensitivity (6, 45). I n general, for transjunctional voltage sensitivity, junctional



10.

HARRIS

Connexin-32

201

conductance is highest w h e n the transjunctional voltage is zero a n d decreases w i t h transjunctional voltage o f either sign (48, 49; b u t see reference 50). I n a well-characterized case o f steep transjunctional voltage sensitivity, w h e n the polarity o f the transjunctional voltage was reversed, the junctional channels o p e n e d transiently (51). T h e kinetics o f the transient were m o d e l e d as i f the channel was gated b y two oppositely oriented voltage sensors a n d the gate that was closed b y the initial voltage polarity h a d to o p e n before t h e other gate c o u l d sense the n e w voltage. A p r o p o s e d mechanism for this "contingent gating" located the voltage-sensing moieties w i t h i n the l u m e n o f the pore. F o r this mechanism, w h e n the l u m e n is o c c l u d e d b y one gating structure, the b u l k o f the m e m b r a n e field drops across it, leaving little o f the field to b e sensed b y the voltage-sensing moiety i n series. T h i s mechanism is a specific example o f a state-dependent alteration o f field across a protein (52). A n alternative explanation is that the interaction between the gating mechanisms is mediated not b y changes i n the field, b u t b y allosteric interactions between the t w o hemichannels. Recent studies that u t i l i z e d molecular genetic tech niques showed that d o m a i n swaps involving the extracellular domains o f differing connexins c a n affect transjunctional voltage sensitivity (45, 53-55) [there is n o obvious analog o f the S-4 region thought to b e crucial for voltage sensitivity i n other channels (56)]. These studies also indicate that the end-to-end interactions o f hemichannels c a n dramatically affect voltagedependent behavior. Indirect evidence implies that i n one c e l l type, such end-to-end modulatory interactions d o not take place (34). Studies o f the voltage dependence are further c o m p l i c a t e d b y observations o f d i f f e r i n g voltage sensitivities f o r the same connexin expressed i n different cells (see reference 45). T h e study o f h e m i c h a n n e l voltage sensitivity u n d e r simple and defined conditions w o u l d b e useful.

Ligand Sensitivity. Rigorous identification o f cellular ligands that act directly o n junctional channels has b e e n difficult because access to the physiological modulatory sites o f the junctional channel is only via cytoplasm (and it has b e e n difficult to identify junctional channels i n patch-clamp studies even w h e r e they are likely to exist; see reference 57). Experiments i n w h i c h junctional c o u p l i n g is m o d u l a t e d b y exposure o f cells to various agents can b e informative w i t h regard to aspects o f cellular regulation o f junctional channels, b u t cannot precisely define w h i c h cytoplasmic factors act o n the channel p r o t e i n itself to effect a change. Connexins and Channel Composition. T h e molecular diversity among connexins is likely to p r o d u c e differences i n the properties o f the channels they f o r m (cf. reference 58). T h e presence o f naturally o c c u r r i n g connexin variants may b e useful i n elucidating the role o f specific domains o f the protein. T h e physiology o f junctional channels i n situ varies dramatically f r o m tissue to tissue (see reference 6), and some o f the variability is n o doubt



202


due to the fact that different combinations of connexins are f o u n d i n each tissue. I n at least one tissue, one connexin predominates sufficiently that it must be capable o f f o r m i n g homo-oligomerie hemichannels ( 5 9 - 6 1 ) . I n tissues w h e r e there are comparable amounts o f different connexins, two different connexins can be f o u n d i n the same array o f junctional channels (62), a n d recent b i o c h e m i c a l data indicate that h e m i c h a n n e l structures can be f o r m e d o f m o r e than one connexin (63). E x p e r i m e n t s i n w h i c h m R N A s that code for different connexins are expressed i n Xenopus oocytes show that junctional channels can be f o r m e d b y two hemichannels, each o f w h i c h is w h o l l y f o r m e d b y different connexins (cf. references 53, 54, 64, 65). These experiments a n d others (48, 66, 6 7 ) indicate that each h e m i c h a n n e l contains discrete gating mechanisms (that may interact strongly w i t h those of the h e m i c h a n n e l w i t h w h i c h it is i n series). T h u s , properties o f junctional channels may be d e t e r m i n e d b y the variable composition o f each h e m i c h a n n e l (hemichannel-specific properties) as w e l l as the variable effects o f end-toe n d h e m i c h a n n e l cooperative interactions. T h e ability to define h e m i c h a n n e l properties independently o f the effects o f their interactions w i t h each other w o u l d be o f value.

Lateral Interactions between Hemichannels. C o n n e x i n chan nels display a p r o f o u n d tendency to aggregate laterally, w h i c h gives rise to the large arrays o f channels previously m e n t i o n e d . N o " h e l p e r " p r o t e i n or cytosketal organizing influence has b e e n identified. A l t h o u g h the functional role o f this aggregation is unclear, it c o u l d favor the formation o f junctional channels b y accumulation o f precursors to localized m e m b r a n e domains, or perhaps it facilitates cooperative gating interactions between adjacent j u n c tional channels [as some researchers have reported (68, 69)], even though adjacent hemichannels i n tightly p a c k e d arrays are separated b y l i p i d (30, 31). W h e n hemichannels are isolated f r o m j u n c t i o n a l membranes b y b i o c h e m i c a l techniques, they can f o r m linear structures, regular two-dimensional arrays, a n d three-dimensional stacks (70, 71). A m o d e l o f the interchannel spacing i n the arrays i n c e l l membranes l e d to the proposal that the interactions between the channels are inherently repulsive, a n d that they aggregate to m i n i m i z e repulsive force between the apposed membranes (72). T h i s mechanism may not operate i n the previously m e n t i o n e d example o f i n vitro aggregation o f p u r i f i e d hemichannels. Junctional Channel Formation. T h e mechanism o f formation o f junctional channels is not k n o w n . A l t h o u g h direct data are lacking, b y analogy w i t h other oligomerie m e m b r a n e proteins, connexin is t h o u g h to be organized into oligomerie (probably hemichannel) structures i n intracellular m e m b r a n e compartments (73, 74). T w o extreme views are (1) that hemichannels are inserted into the plasma m e m b r a n e solely at sites o f j u n c t i o n formation a n d (2) that hemichannels are already present i n the m e m b r a n e a n d nucleate at


10.

HARRIS

Connexin-32

203

the time o f j u n c t i o n formation (75). T h e reality may involve b o t h processes. E a r l y immunofluorescence studies failed to reveal nonjunctional connexin i n plasma m e m b r a n e ( 5 9 , 76), b u t dispersed p r o t e i n w o u l d have b e e n difficult to detect. M o r e recent data argue strongly that i n n o r m a l cells, connexin precursors (presumably hemichannels)

are present nonjunctionally i n the

plasma m e m b r a n e and, furthermore, their assembly into plaques o f junctional channels is associated w i t h a phosphorylation ( 7 7 ) . F o r a connexin expressed Downloaded by UNIV OF CALIFORNIA SAN DIEGO on December 29, 2015 | http://pubs.acs.org Publication Date: May 5, 1994 | doi: 10.1021/ba-1994-0235.ch010

i n oocytes, indirect data argue that connexin precursors are accessible to extracellular reagents p r i o r to formation o f junctional channels (78). T h e formation o f junctional channels requires end-to-end

association

between the extracellular domains o f hemichannels. T h i s interaction is not likely to b e covalent because junctional channels can b e split into h e m i c h a n nels b y alkaline urea treatments (7, 79). H o w e v e r , the formation o f junctional channels between

m R N A - i n j e c t e d oocytes is critically dependent o n the

presence o f the three cysteines i n each connexin extracellular l o o p (78). These cysteines apparently d o not f o r m i n t e r h e m i c h a n n e l disulfide bonds (80-82),

b u t m a y serve to stabilize the extracellular domains d u r i n g the

h o m o p h i l i e b i n d i n g reaction w i t h the apposing h e m i c h a n n e l . H e m i c h a n n e l s i n apposing membranes must b e brought sufficiently close for t h e m to interact. E a c h h e m i c h a n n e l protrudes extracellularly only ~ 10 Â f r o m the plasma m e m b r a n e ( 3 0 , 31), w h i c h is a very short distance c o m p a r e d w i t h usual intercellular separation. T h e typical m e m b r a n e - m e m b r a n e

separa

tion w i t h i n a gap j u n c t i o n is ~ 20 Â ( 1 9 ) [intriguingly close to the e q u i l i b r i u m distance between uncharged p h o s p h o l i p i d surface (83)]. Strong evi dence indicates that cell adhesion molecules play a crucial role i n b r i n g i n g c e l l membranes close enough for the hemichannels to interact (84, 8 5 ) . T h e r e are significant physiological, structural, a n d b i o c h e m i c a l data w i t h regard to the junctional channel. T h i s b o d y o f knowledge provides a basis f o r exploration o f certain issues w i t h respect to the physics o f channel-forming proteins a n d o f molecular dynamics specific t o junctional channels that may have implications f o r other macromolecular structures. These issues i n c l u d e the location o f voltage-sensing charges, the roles o f specific p r o t e i n domains, testing o f specific mechanisms

o f channel closure, interactions

between

conductance state a n d electric field, cooperative effects between h e m i c h a n nels, mechanism o f permeation, a n d the forces i n v o l v e d i n formation o f double-membrane p r o t e i n structures.

Prospects for Biophysical Study of Channels Formed by Connexin W h a t are the possibilities f o r addressing these issues? I n most cases, the r e q u i r e d data are difficult to obtain d u e to (1) the inaccessibility o f the


204


junctional channel i n situ or (2) the complications o f the presence o f two functional hemichannels i n series, w h i c h furthermore interact i n u n k n o w n ways, or a combination o f b o t h . F o r example, it is difficult to obtain a selec tivity sequence w h e n the ionic environment at neither e n d o f the pore can be changed w i t h impunity. Changes i n the connexin amino acid sequence can affect the nature o f the h e m i c h a n n e l - h e m i c h a n n e l interactions simultane ously w i t h the gating, kinetic, a n d permeability properties that are o f p r i m a r y Downloaded by UNIV OF CALIFORNIA SAN DIEGO on December 29, 2015 | http://pubs.acs.org Publication Date: May 5, 1994 | doi: 10.1021/ba-1994-0235.ch010

interest. These factors

might be deconvoluted b y parametric studies

of

oocytes that express hemichannels f o r m e d b y differing connexins (cf. refer ence 54), but straightforward studies o f the properties o f single hemichannels w o u l d be o f clear value. T h e Xenopus oocyte expression system can be used to study macroscopic junctional conductances, b u t not to make single-channel measurements. D u a l voltage c l a m p o f c o u p l e d cells u n d e r conditions o f r e d u c e d junctional c o n ductance can reveal single-channel transitions (41, 46), but the problems o f accessibility a n d series hemichannels r e m a i n . A w e l l - d e f i n e d accessible system is r e q u i r e d . Reconstitution into l i p o somes a n d planar p h o s p h o l i p i d bilayers can provide the accessibility a n d has b e e n used to great advantage for many i o n channels (86). T h e disadvantages o f reconstitution are that the protein may be damaged a n d that it may lack c h e m i c a l factors

r e q u i r e d for physiological function. I n addition,

u n i q u e problems arise for reconstitution o f connexin channels

several

(87).

T h e gap j u n c t i o n channel spans two membranes. A l t h o u g h it is desirable to reconstitute d o u b l e - m e m b r a n e connexin channels at the interface o f two closely apposed bilayers, at present a well-characterized a n d reliable system o f this type does not exist (but see reference 88). O n the other h a n d , the junctional channel is c o m p o s e d o f discrete single-membrane structures that, as previously mentioned, may function normally as channels i n nonjunctional plasma membrane. Therefore, it may be possible to study single h e m i c h a n nels w i t h standard reconstitution techniques. D a t a f r o m structural studies a n d f r o m cellular physiology indicate that each h e m i c h a n n e l has a f u l l c o m p l e ment o f gating structures a n d sensitivities (e.g., each can respond to voltage, p H , etc.) ( 3 1 ,

48,

64-67, 89).

E v e n though some aspects o f junctional

channel operation are likely to involve end-to-end h e m i c h a n n e l - h e m i c h a n n e l interactions, it w o u l d be a major step to fully characterize the biophysics o f a single h e m i c h a n n e l . I n fact, this must be achieved i n order to determine h o w the cooperative end-to-end interactions affect the function o f each h e m i c h a n nel. A second p r o b l e m is the identification o f a bilayer channel as a relevant connexin channel. M o s t channels i n bilayers are identified b y defining physi ology (e.g., ionic selectivity, toxin effects). F o r the junctional channel, the only certain defining properties are permeability to large molecules and, i n most cases, voltage sensitivity. O t h e r factors that affect junctional conduc-


10.

HARRIS

205

Connexin-32

tance between cells (e.g., p H , c a l c i u m , octanol) may not act o n the channel molecule itself. Because bilayers report the activity o f a very small n u m b e r o f molecules (typically < 5 f o r single-channel recordings), standard b i o c h e m i c a l criteria alone cannot be used to identify a connexin channel i n a bilayer because the sampling error is too great. F o r example, 10 μ g o f p r o t e i n a d d e d to a bilayer chamber corresponds to approximately 3 X 1 0

1 3

channels o f the

molecular weight o f a hemichannel; even an extremely m i n o r contaminant Downloaded by UNIV OF CALIFORNIA SAN DIEGO on December 29, 2015 | http://pubs.acs.org Publication Date: May 5, 1994 | doi: 10.1021/ba-1994-0235.ch010

c o u l d produce significant a n d misleading channel activity, especially i f consid eration is given to (1) differences

i n efficiency o f incorporation into the

bilayer a n d (2) lack o f certainty that an incorporated connexin channel w i l l function u n d e r a given set o f conditions. F o r the same reason, the use o f antisera to identify bilayer channels places exceedingly stringent requirements o n antisera purity, because only a f e w antibody molecules w i t h u n d e s i r e d specificity c o u l d misidentify the small n u m b e r o f channel molecules active i n a bilayer. A related p r o b l e m is h o w to obtain connexin suitable f o r reconstitution. P u r i t y w o u l d not b e necessary i f there were an easy w a y to specifically activiate o r select for reconstituted connexin channels. T h e preparations o f isolated junctional m e m b r a n e

described previously y i e l d condensed j u n c

tional membranes a n d involve denaturing conditions k n o w n to affect con nexin secondary structure ( 9 0 ) . T h e preparations also contain nonconnexin protein, because it is the junctional m e m b r a n e structure that is purified, not specifically the channel-forming protein. Several studies report channel activ ity f r o m such preparations, o r f r o m detergent-solubilized material f r o m these preparations, c o m p a r i n g the channel activity w i t h physiology, pharmacology, a n d antisera effects i n f e r r e d f r o m cellular c o u p l i n g studies (91-96).

O u r own

early w o r k w i t h this material selected specifically f o r reconstituted channels that were permeable to large molecules ( 9 7 ) . Recently, preparations for connexin that may provide m o r e

suitable

material f o r reconstitution have b e e n reported. Immunoaffinity purification o f two different connexins, u n d e r nondenaturing conditions, f r o m membranes solubilized w i t h nonionic detergent has b e e n reported: one purification used a monoclonal antibody against a specific connexin ( 9 8 a n d R h e e ,

S. K . ;

Harris, A . L . , unpublished) a n d the other used a sequence-specific polyclonal antibody ( 9 9 ) . A technique

that c o m b i n e d alkaline conditions, nonionic

detergent, a n d chromatography y i e l d e d morphologically p u r e hemichannels (e.g., not junctional membranes) f r o m liver (70). Detergent solubilization o f lens fiber c e l l membranes y i e l d e d single hemichannels, though this prepara tion contains nonconnexin protein (71). F u n c t i o n a l reconstitution o f channels f r o m the hemichannels p u r i f i e d f r o m liver has b e e n reported (100). O u r studies w i t h junctional membranes a n d connexin p u r i f i e d w i t h m o n o c l o n a l antibody (87,

101-103,

a n d Rhee, S. K . ; Harris, A . L . , unpublished) are

briefly s u m m a r i z e d i n the following section.


206



Development of a Reconstitution System for Study of Channels Formed by Connexin-32 T h e studies o u t l i n e d i n this section describe the ways w e have addressed the foregoing problems o f connexin reconstitution b y utilizing connexin-32, the predominant f o r m o f connexin i n rat liver. O u r goals were to establish unambiguously that connexin-32 f o r m e d channels i n liposome membranes, to identify connexin channels i n planar M a y e r s , a n d to study their properties. T w o methods were used to identify reconstituted channels f o r m e d b y c o n nexin-32. I n one method, p r o t e i n was solubilized f r o m preparations o f junctional m e m b r a n e a n d incorporated into unilamellar liposomes. C o n n e x i n 32 was identified as a c h a n n e l - f o r m i n g p r o t e i n b y its specific enrichment i n liposomes that were permeable to sucrose. I n the other m e t h o d , connexin-32 was affinity-purified ( w i t h a m o n o c l o n a l antibody directed specifically against connexin-32) directly f r o m octylglucoside-solubilized plasma membranes. L i posomes f o r m e d w i t h such material were permeable to sucrose a n d L u c i f e r Y e l l o w . Sucrose-permeable liposomes f r o m each m e t h o d were fused w i t h planar bilayers to study the properties o f connexin channels.

Connexin from Isolated Junctions Forms Large Channels in Liposomes (87, 101, 102). P r o t e i n solubilized i n octylglucoside f r o m preparations o f junctional membranes was incorporated into unilamellar p h o s p h o l i p i d liposomes b y g e l filtration (104). L i p o s o m e s that contained large o p e n channels were separated f r o m those that d i d not using a p e r m e ability-specific liposome fractionation protocol (105). I n brief, the liposomes were f o r m e d i n a solution that contained several h u n d r e d m i l l i m o l a r urea a n d were spun o n a n isoosmotic density gradient f o r m e d f r o m the urea buffer a n d a buffer i n w h i c h the urea was osmotically replaced b y sucrose (sucrose is m o r e dense than urea). L i p o s o m e s that are i m p e r m e a b l e to sucrose a n d urea move a short distance into the gradient to a steady-state position. Liposomes that contain large open pores exchange urea f o r sucrose a n d move to a characteristic lower position (higher density) i n the gradient ( F i g u r e 2 A - C ) . I n calibration studies using V D A C (voltage-dependent anion channel) (106), the movement o f the liposomes to the lower (sucrose-permeable) position was correlated w i t h a loss o f a large intraliposomal marker (calcein; molecular weight M W = 660 D a ) (105). T h i s m e t h o d is reliable f o r separation o f liposomes that are permeable to large molecules f r o m liposomes that are not. Sucrose c a n permeate gap j u n c t i o n channels, b u t not most other m e m b r a n e channels. A significant fraction o f the liposomes that contain junctional p r o t e i n were sucrose-permeable ( F i g u r e 2 D a n d E ) . T h i s result shows a correlation between p r o t e i n f r o m junctional m e m b r a n e a n d sucrose permeability, b u t does not, b y itself, p e r m i t the conclusion that connexin-32 forms the sucrosepermeable pathway.



Β

C

D

0

5

Ε

10 Fraction Number

Density Gradient

15

20

Figure 2. Sucrose permeability in liposomes induced by protein from isolated gap junctions. Unilamellar liposomes are separated on the basis of permeability to urea and sucrose on iso-osmotic density gradients as described in the text and in reference 107. Liposomes containing large open channels exchange urea for sucrose and increase in density (move to a lower position in the gradient), whereas those that are not remain at a lighter density nearer the top of the gradient. A—C, Diagram of the transport-specific liposome fractionation technique. D, Sucrose permeability induced by protein from isolated gap junctions. Liposomes containing junctional protein segregated into two characteristic populations. A fluorescently labeled lipid (rhodamine-PE) was a marker for the liposomes in this experiment. The upper arrow indicates the band formed by the sucrose-impermeable liposomes, and the lower arrow indicates the band formed by the sucrose-permeable liposomes. Liposomes formed without junctional protein or with control proteins remained in a single band at the position of the upper arrow (see Ε). E, Graph of liposome fractionation. Liposomes without protein (open bars) and with junctional protein (filled bars) were separated as in D (in this experiment, a radioactive lipid was used to follow liposome position). The tops of the gradients are at the left. Liposomes formed without protein were in a single band at fractions 6 and 7. Liposomes formed with junctional protein separated into two bands: one at the upper position defined by the sucrose-impermeable liposomes and one approximately threefold larger at the higher density sucrose-permeable position (fractions 15 and 16). The lower relative position of the sucrose-impermeable liposomes compared with those in D results from a longer centrifuge spin (8 h vs. 3 h); with longer spins the upper band tends to drift slowly to lower positions, a possible reflection of nonspecific liposome permeability (105). (Reproduced with permission from reference 102. Copyright 1992 Elsevier.)

A

30

40,



208


T o identify the p r o t e i n responsible for the permeability, proteins i n the liposomes w e r e separated b y standard denaturing electrophoresis, blotted, and stained for total p r o t e i n a n d for connexin-32. W e s t e r n blots o f unfractionated liposomes f o r m e d i n the presence o f p r o t e i n solubilized f r o m isolated gap junctions are shown i n F i g u r e 3 A . T h e blots show the m o n o m e l i c , d i m e r i c , a n d trimerie forms o f connexin-32 c o m m o n l y observed i n s o d i u m dodecyl sulfate ( S D S ) gels o f isolated junctions ( 5 9 - 6 2 , 107). T h e y also show the presence o f a c o m m o n l y seen proteolytic fragment o f connexin-32 (59, 61, 107) (better seen i n F i g u r e 3 B ) , w h i c h contributes to the broadness o f the staining b e l o w the m o n o m e r a n d d i m e r bands. T h e liposomes typically contained no detectable nonconnexin p r o t e i n . W e s t e r n blots o f transport-selected liposomes ( F i g u r e 3 B ) show specific selection for full-length connexin-32 b y sucrose permeability: the ratios o f connexin-32 to its proteolytic fragment were very different i n the sucrosepermeable a n d sucrose-impermeable liposomes. T h e ratio was m u c h greater i n the sucrose-permeable liposomes than i n the sucrose-impermeable lipo somes. D i g i t a l integration o f lanes 5 a n d 6 (areas u n d e r the peaks i n the tracings) show that the ratio o f full-length connexin-32 peak to its fragment was 1.6 for the sucrose-permeable lipsomes a n d 0.52 for the sucrose-imper meable liposomes. T h u s , the sucrose permeability essentially e n r i c h e d for full-length connexin-32, a n d sucrose i m p e r m e a b i l i t y selected for the c o n nexin-32 fragment. F u l l - l e n g t h connexin-32 was almost totally absent f r o m the sucrose-impermeable liposomes. A simple explanation is that full-length connexin-32 can f o r m a sucrose-permeable pore (and the fragment cannot). T h e argument against sucrose permeability b e i n g d u e to a nonconnexin protein is as follows: I f the sucrose permeability was due to a hypothetical nonconnexin protein, a n d the proteins were distributed independently o f one another i n the liposomes, then connexin-32 and its fragment w o u l d be present i n the same ratio i n the two liposome populations, each distributed independently o f sucrose permeability. Because they are not, it is reasonable to conclude that connexin-32 causes a permeability to sucrose. T h e presence o f some fragment i n the sucrose-permeable liposomes is accounted for b y its presence i n liposomes that also contained full-length (functional) connexin. T h e conclusion that connexin-32 forms a pore does not require that all nonconnexin p r o t e i n be e x c l u d e d f r o m the sucrose-permeable liposomes or that connexin-32 account for all o f the sucrose permeability. It relies o n the positive a n d specific correlation between sucrose permeability a n d e n r i c h ment for full-length connexin-32.

Affinity-Purified Connexin-32 Forms a Large Pore in Liposomes (103). C o n n e x i n - 3 2 solubilized i n oetylglucoside f r o m crude plasma m e m b r a n e was affinity-purified u n d e r nondenaturing conditions using a monoclonal antibody specific for connexin-32 (7) that was attached to a b e a d matrix (98 a n d R h e e , S. K . ; H a r r i s , A . L . , unpublished). O v e r l o a d e d denatur-



10.

HARRIS Connexin-32

1

2

209

3

4

5

6

Figure 3. Sucrose permeability selects for liposomes containing full-length connexin-32. A, Western blots of liposomes containing protein solubilized from isolated gap junctions. Left lane: stained for protein with colloidal gold. Right lane: stained with monoclonal antibody against connexin-32 (7), visualized with an alkaline phosphatase conjugated secondary antibody. (Connexin-32 is a 32-kD protein that runs anomalously at the 27-kD position in these 13% acrylamide gels (126). Gels were intentionally overloaded so the positions of multimeric and proteolytic connexin-32 could be identified.) R, Western blots of transport-selected liposomes show specific selection for full-length connexin-32 by sucrose permeability. Lanes 3 and 5 are blots of sucrose-permeable liposomes, and lanes 4 and 6 are of sucrose-impermeable liposomes. The left pair (lanes 3 and 4) are stained with monoclonal antibody against connexin-32 as in part A. Therightpair (lanes 5 and 6) are digital reconstructions of blots stained for protein with colloidal gold. Tracings are integrations of blots stained for protein with colloidal gold in the monomer region; right to left on the tracings correspond to moving down the lanes of the blots. The lighter and darker tracing lines correspond to the sucrose-permeable and sucrose-impermeable lanes, respectively, and the right and left peaks correspond to the full-length and proteolytic fragment positions, respectively. Each pair of blots (3 and 4; 5 and 6) shows that sucrose permeability selects for a liposome population containing more full-length connexin-32, relative to its fragment, than do sucrose-impermeable liposomes. Sucrose impermeability strongly selects against the full-length form, as it is almost absent in the sucrose-impermeable liposomes. The proteolyzed fragment contributes to the dimeric connexin-32, producing a lower molecular weight form of the dimer (open arrowhead). (All four lanes are from the same starting population of liposomes and the same density gradient spin. To demonstrate the difference in distribution of the two bands in the transport-selected populations, the liposome population used in this example contained more of the connexin-32 fragment than the full-length monomer. Absorbance of gold-stained blots was digitized and background absorbance subtracted. Staining density of the full lane width was integrated along its length. Peaks were approximated by overlapping Gaussian curves, represented at densities in lanes 5 and 6. Differences in staining between the left andrightpairs of lanes occur because the antibody method (lanes 3 and 4) stains the connexin-32 fragment darker than the full-length form and stains the multimeric forms with decreasing intensity, perhaps because of increasing inaccessibility of the antigenic site.) (Reproduced with permission from reference 102. Copyright 1992 Elsevier.)



210


i n g gels showed no detectable contaminants. T h e y i e l d o f connexin-32 was considerably greater than that obtained b y junctional m e m b r a n e isolation, and the connexin-32 was i n a soluble f o r m . G e l filtration o f the material i n d i cated that the connexin-32 was p r e d o m i n a n t l y o f a size consistent w i t h hexanieric connexin. T h e affinity-purified connexin-32 was incorporated into unilamellar l i p o somes as before. T h e connexin-32 i n d u c e d a sucrose permeability i n l i p o somes, as assayed b y the density-shift technique, a n d gave results essentially identical to those i n F i g u r e 2. L i p o s o m e s that were sucrose-permeable d i d not retain the dye L u c i f e r Y e l l o w (retained b y the sucrose-impermeable liposomes), w h i c h is near the u p p e r size-permeability limit for gap j u n c t i o n channels (19, 108, 109). T h e fraction o f the liposomes that were permeable to sucrose decreased b y a factor o f 4 w h e n the p H i n the gradients was changed f r o m 7.5 to 6.0. T h i s effect was partially reversible. T a k e n together, the two p r e c e d i n g liposome studies provide a robust demonstration that connexin-32 can be successfully reconstituted into u n i lamellar p h o s p h o l i p i d membranes, w h e r e it forms pores w i t h permeabilty similar to that o f junctional channels. T h e data are consistent w i t h the c o n d u c t i n g unit b e i n g the h e m i c h a n n e l .

Bilayer Channels from Sucrose-Permeable Liposomes (87, 101 102). Sucrose-permeable liposomes f r o m the foregoing studies were fused w i t h planar p h o s p h o l i p i d M a y e r s (110, 111). T h e data f r o m the liposomes that contain connexin f r o m isolated junctions d i f f e r e d i n some respects f r o m the data obtained w i t h the affinity-purified connexin. A t the present time, the data f r o m the affinity-purified material is not fully charac terized. Therefore, most o f the data described i n the following text are f r o m liposomes that contain connexin-32 f r o m isolated junctions; exceptions are noted. y

F u s i o n into the bilayer o f sucrose-permeable liposomes resulted i n three unitary conductance levels, two o f w h i c h c o u l d be attributed to connexin-32. F i g u r e 4 A shows typical channel bilayer activity at ± 1 0 0 m V . A t +100 m V , one 130-pS channel a n d at least two 20-pS channels undergo gating transi tions. A t —100 m V , three 45-pS channels are seen. T h e channel activity is more fully characterized b y the voltage r a m p shown i n F i g u r e 4 B . T h e membrane conductance is h i g h a n d linear, without channel gating activity f r o m —25 to +25 m V . A t more negative voltages, four 45-pS channels t e n d e d to close. A t more positive voltages, a 130-pS channel t e n d e d to close a n d a 20-pS channel t e n d e d to open. I n this bilayer, the entire m e m b r a n e c o n d u c tance was accounted for b y the s u m o f the observed unitary conductance transitions. T h e 20-pS channel was also seen i n sucrose-impermeable l i p o somes a n d was closed at 0 m V (i.e., it was not likely to mediate a liposome density shift), so it cannot be assigned an identity based o n the density shift studies.


10.

HARRIS

211

Connexin-32

-100

mV +100

JITWI 20pAJ


*20

mV -100

sec

10

sec

+ 50

+ 100

8mV

Figure 4. Channel activity following fusion of transport-selected (sucrose-permeable) connexin-32 containing liposomes with planar bilayers. A, Record showing the three channel sizes observed. Sucrose-permeable liposomes were added to one side of a bilayer chamber and fused with a planar phospholipid bilayer (1:1, PS:PE) by an osmotic stress technique (110, 111). Positive bilayer voltages correspond to inside-liposome positive potentials. R, Voltage ramp showing character of the channel activity. A slow voltage ramp from —100 to +100 mV over 8.5 min was delivered to the same membrane as in part A. At positive voltages, a 130-pS channel tends to close and a 20-pS channel tends to open. At negative voltages, several 45-pS channels tend to close. (Adapted with permission from reference 87. Copyright 1991 CRC Press.)

T h e t w o larger channel transitions (those at 4 5 - 5 5 p S a n d at 1 2 0 - 1 4 0 pS), o n the other hand, were found only w i t h sucrose-permeable liposomes a n d therefore c a n be attributed to connexin. T h e t w o conductance levels c o u l d indicate either two distinct channels o r the presence o f a subconductance state. T h e f o r m e r interpretation is favored because the two levels w e r e not f o u n d i n constant ratio i n different membrances, a n d because (as i n F i g u r e 4 B ) the entire conductance o f the bilayer c o u l d often b e accounted for b y the s u m m e d conductances o f the transitions, as i f each c o n t r i b u t e d independently to the bilayer conductance. T h e key properties o f the connexin channels were that they were o p e n at zero bilayer voltage a n d exhibited p r o f o u n d l y asymmetric sensitivity to voltage. T h e larger channel t e n d e d to close at positive voltages a n d r e m a i n open at negative voltages. T h e smaller one h a d a voltage sensitivity o f opposite sign, t e n d i n g to close only at negative potentials. T h e voltage sensitivity o f the reconstituted channels was w e l l character i z e d b y a f o r m o f the B o l t z m a n n relation that contained a voltage-dependent energy t e r m , indicating that the voltage-sensitive transitions are first order and that the energy difference between the o p e n and closed states is a linear function o f voltage (49, 112). T h e n u m b e r o f equivalent gating charges f o r


212


the larger channel ranged f r o m 1.5 to 2.5, a n d the voltage at w h i c h the o p e n probability is one-half ( V ) is 0

~ 110 m V . Overall, the smaller channel h a d

approximately one-half the voltage sensitivity o f the larger channel a n d a similar V . A description o f a possible pitfall i n determination o f voltage sen 0

sitivity follows. A c o m m o n m e t h o d to determine voltage sensitivity is shown i n F i g u r e 5 A , i n w h i c h the data for the larger channel f r o m one experiment are fitted to Downloaded by UNIV OF CALIFORNIA SAN DIEGO on December 29, 2015 | http://pubs.acs.org Publication Date: May 5, 1994 | doi: 10.1021/ba-1994-0235.ch010

a f o r m o f the B o l t z m a n n relation that c a n b e plotted as a linear f u n c t i o n o f

•

81

A

1

ln(Po/Pc)

40

60

80

100

120

140

120

140

VOLTAGE (mV)

40

60

80

100

VOLTAGE (mV) Figure 5. Voltage dependence of Ρ for the larger (130-pS) connexin-32 channels. A, Fit to ln(F /P ) as a function of voltage. B, Fit to P as a function of voltage. The solid line is calculated with the Boltzmann relation (SSE = 0.007) and the dashed line is an independent fit (SSE = 0.001). σ

Q

c

0


10.

HARRIS

213

Connexin-32

voltage: ln( P / P ) = nqV/kT, 0

c

where P

is the o p e n probability, P

Q

c

is the

closed probability calculated as 1 — P , q is t h e electronic charge, η is the G

n u m b e r o f charges that move t h r o u g h the bilayer voltage V , k is B o l t z m a n n s constant, a n d Γ is temperature (49, 106, 113). T h e fit appears

reasonable

( r = 0.99) a n d evaluates η as 2.1, w h i c h corresponds to a n e-iold change i n the o p e n - c l o s e d ratio for every ~ 12 m V . T h e V

0

is 109 m V .

W h e n these parameters are used to predict the P versus V relation a n d Q

the exponential f o r m o f the B o l t z m a n n relation ( P = 1/{1 + exp[nq(V

—


0

is used, a reasonable fit to the data is also seen (solid line i n F i g u r e

V )/kT]}) 0

5 B ) . H o w e v e r , fitting this f o r m o f the B o l t z m a n n relation directly to the P

0

versus V data (dashed line i n F i g u r e 5 B ) gives a statistically i m p r o v e d fit (F-test; Ρ < 0.05) a n d particularly w e l l describes the region o f the greatest slope. T h e independent fit gives a greater voltage sensitivity (n = 2.8, w h i c h corresponds to an e-fold change for ~ 9 m V ) a n d approximately the same V

0

(107 m V ) . T h e reason f o r the difference i n the results f r o m the two fitting procedures is that, d u e to the mathematical transformation f r o m

P

Q

to

ln( P / P ) , the uncertainty f o r values d e r i v e d f r o m P s that approach 1 is Q

C

Q

greater than the uncertainty d e r i v e d f r o m smaller P s [e.g., a ± 0 . 0 1 uncer Q

tainty i n a P o f 0.95 gives deviations o f + 0 . 2 3 4 and - 0 . 1 9 2 i n l n ( P / P ) , 0

G

C

whereas the same ± 0.01 deviation i n a P o f 0.5 gives a deviation o f ± 0.04]. G

Because o f this difference, the lower values o f ln( P / P ) s h o u l d b e w e i g h t e d Q

C

more than the higher values d u r i n g the curve fitting. T h i s weighting c a n b e achieved b y use o f a w e l l - d e f i n e d mathematical technique (114). W h e n the values are appropriately weighted, the curves generated b y fitting to the logarithmic f o r m a n d the exponential f o r m o f the B o l t z m a n n relation are statistically indistinguishable a n d give similar values o f η a n d V V

0

0

(n = 2.6 and

= 107 m V for the w e i g h t e d logarithmic fit). This observation is o f some

interest because curve

fitting

to ln( P / P ) is a long-standing technique. G

C

F a i l u r e to appropriately weight the data i n the fitting procedure may c o n tribute to some o f the variability i n voltage sensitivity reported f o r many conductances a n d channels, i n c l u d i n g junctional conductances a n d connexins. T h e channels h a d d w e l l times that ranged f r o m several h u n d r e d millisec onds to several seconds. C h a n n e l s that responded to positive a n d negative voltages i n a symmetric manner were rarely seen. A t voltages near V , the 0

larger channel h a d a n exponential distribution o f o p e n times, w h i c h indicates a single open state ( F i g u r e 6 A ) . A t lower voltages, the fit to a single exponential was not as good, though reasonable, suggesting the contribution o f a m i n o r , slower second exponential, w h i c h c o u l d indicate closures f r o m a second o p e n state. T h e p r i m a r y component o f the m e a n o p e n times h a d an exponential relation to voltage ( F i g u r e 6 B ) . T h e apparent voltage sensitivity of the closing rate corresponds to the movement o f ~ 1.6 charges through the bilayer field. T h i s suggests that most o f the voltage sensitivity o f the steady-state P arises f r o m the closing rate. Q


214



# EVENTS

SEC

τ (sec)

VOLTAGE (mV) Figure 6. Voltage-dependent kinetics of the larger connexin-32 channels. A, Open time histogram of the larger channel. The open dwell times at 100 mV (near V ) were well-fit by a single exponential, suggesting that the channel closes from a single open state. Bin width is 1.4 s. (Reproduced with permission from reference 87. Copynght 1991 CRC Press). B, Open time versus voltage for the larger channel. The mean open times for the channel are plotted as a function of voltage, and well-fit by an exponential. The voltage sensitivity of the apparent closing rate corresponds to the movement of ~ 1.6 charges through the bilayer field. 0

T h e reversal potentials f o r bilayers containing the channels were mea sured under asymmetric K C l a n d N a C l conditions, a n d polarity was c o n f i r m e d i n the K C l experiments b y addition o f the potassium-selective ionophore v a l i n o m y c i n . T h e reversal potentials indicated approximately a three-fold selectivity f o r chloride over either potassium o r s o d i u m ions.


10.

HARRIS

Connexin-32

215


A 20-pS channel was occasionally observed along w i t h these channels, but because it was closed at zero m e m b r a n e voltage, it was u n l i k e l y to mediate the density shift. Because it was also f o u n d i n sucrose-impermeable liposomes, the 20-pS channel c o u l d not b e assigned a n identity based o n sucrose permeability. T h i s c h a n n e l may b e a m i n o r contaminant o r a partially denatured f o r m o f connexin-32. L i p o s o m e s that contained i m m u n o p u r i f i e d connexin-32 p r o d u c e d macro scopic bilayer conductances that w e r e highly asymmetric w i t h regard to voltage. H o w e v e r , clean single-channel transitions were only occasionally observed, even at l o w ( 1 0 0 - 2 0 0 - p S ) macroscopic m e m b r a n e conductance, because the bilayer currents w e r e noisy a n d unsteady. S u c h recordings often characterize " d i r t y " bilayers o r detergent-induced conductances, but without the p r o f o u n d asymmetry w i t h respect to voltage observed here. A d d i t i o n o f control liposomes made u n d e r the same conditions without p r o t e i n p r o d u c e d no detectable conductances, a n d addition o f octylglucoside eventually p r o d u c e d only symmetric increases i n conductance noise (the liposome m e m branes s h o u l d not contain appreciable detergent i n any case). W h e n these records are filtered extensively, the transitions are seen to cluster at intervals o f ~ 125 p S ( F i g u r e 7 A ) . A l t h o u g h these currents are difficult to interpret, the findings might b e accounted f o r b y h i g h conduc tance channels that fluctuate rapidly between many subconductance levels. P r e l i m i n a r y higher resolution bilayer experiments ( i n collaboration w i t h S. M . Bezrukov; F i g u r e 7 B ) revealed discrete conductance transitions w i t h largemagnitude current fluctuations d u r i n g the o p e n states, supporting this idea, but further studies are r e q u i r e d .

Discussion of the Major Findings W h a t d o these liposome and bilayer data i m p l y about connexin-32 channels?

Structural Form of the Reconstituted Channels. T h e experi ments w i t h connexin-32 f r o m isolated junctions d o not directly address the issue o f structural f o r m . H o w e v e r , the asymmetry o f the voltage dependence o f channels f o r m e d b y connexin-32 f r o m b o t h sources suggests a n asymmetric structure, w h i c h rules out a n o r m a l l y configured junctional channel. T h e sizing data suggest that the p u r i f i e d connexin is isolated p r e d o m i n a n t l y as single hemichannels. A l s o , the L u c i f e r Y e l l o w permeability suggests that the aqueous pore is not significantly narrower than the pores o f connexin channels i n situ. These considerations, plus the fact that the channels exist i n what are macroscopically single membranes, place the weight o f the evidence o n the reconstituted connexin-32 b e i n g i n the f o r m o f single hemichannels. T h i s conclusion is consistent w i t h voltage-clamp studies o f cells, w h i c h suggest that each h e m i c h a n n e l contains a gating mechanism sensitive to a single voltage polarity ( 3 5 , 48, 5 i , 54).


216



A

10 pA 50 msec

Figure 7. Bilayer conductance induced by affinity-purified connexin-32. A , Sucrosepermeable liposomes formed with affinity-purified connexin-32 were fused with planar phospholipid bilayers as described. Highlyfiltered(5-Hz corner frequency) currents show unstable conductances, but large, rapidfluctuationsthat cluster around multiples of about 125 pS may be discerned (arrowheads). The bilayer voltage was 50 mV. B, Higher resolution recording of channels from affinity-purified connexin-32. Records show discrete gating conductance transitions, but with a high rate and amplitude of currentfluctuationsthrough the open channels. Unitary conductance is difficult to determine, but is near 200 pS. The bilayer voltage was 100 mV.


10.

HARRIS

Connexin-32

217


Effects of the Two Methods of Obtaining Connexin-32. T h e p H sensitivity o f the sucrose permeability o f liposomes that contain affinity-purified connexin suggests that p H c a n act directly o n connexin channels. This contrasts w i t h the lack o f p H sensitivity o f the connexin-32 channels obtained f r o m isolated junctional membranes [our o w n studies (J02) a n d references 93 a n d 94]. T h e apparent difference i n single channel behavior between the t w o preparations m a y also b e a function o f the different isolation procedures. These differences c o u l d result f r o m harsher conditions involved i n j u n c t i o n isolation [known to alter secondary structure o f connexin i n w h i c h case the physiology o f the affinity-purified material may more accurately reflect the physiology o f the native f o r m o f the protein. I n addition, the two isolation methods c o u l d select for connexin that originated i n d i f f e r i n g populations o f junctional structures. T h i s is almost certainly the case because the affinity-purified preparation consists only o f readily soluble plasma m e m b r a n e connexin (not obtained i n junctional m e m brane preparations).

(90)1,

Open Channels in Single Membranes. F o r future reconstitution studies, the most significant finding is simply that connexin-32 can f o r m large, open, regulated channels i n single p h o s p h o l i p i d membranes. Several factors argue that these channels represent functionally relevant channels a n d not structures that are overly denatured, damaged, o r i n other ways artifactual. T h e aqueous pathway through the molecule is as w i d e as that o f junctional channels, a n d the channels are regulated b y voltage i n a n asymmetric m a n n e r that is consistent w i t h t h e m b e i n g single hemichannels. T h e p H effect o n permeability o f the channels f o r m e d b y affinity-purified connexin-32 is also supportive, as are the data w i t h regard to the size o f the p u r i f i e d structures a n d the finding o f asymmetric voltage dependence. F r o m a biological perspective, the finding o f functioning hemichannels seems counterintuitive. T h e r e is an understandable bias that such large channels, i f they were open, w o u l d rapidly k i l l cells b y destruction o f the selective permeability o f the plasma m e m b r a n e . H o w e v e r , the plasma m e m brane o f macrophages a n d mast cells c a n b e c o m e permeable to L u c i f e r Y e l l o w [possibly through connexin channels (33)] for many minutes w i t h o u t lethal effect (115, 116). T h e finding o f o p e n connexin channels i n single membranes suggests that either single hemichannels d o not exist i n plasma membranes (for very long) or that i f they exist, they are kept closed most o f the t i m e b y cytoplasmic, extracellular, o r m e m b r a n e factors not present i n the reconstituted system [as may occur i n the teleost horizontal cells (34) a n d does not occur i n oocytes that express connexin-46 ( 3 5 , 36)]. I n the f o r m e r case, w h e n intercellular channels are f o r m e d , the hemichannels may b e inserted i n apposing m e m branes i n a coordinated fashion, so that there is only a very b r i e f o p e n i n g to the outside d u r i n g channel formation. I n the latter case, the reconstituted


218



system offers a way to precisely identify the factors that modulate hemiehann e l gating.

Interactions between Hemichannels. W h a t does the isolation o f single hemichannels i m p l y about the e n d - t o - e n d interactions o f h e m i c h a n nels? This is an important consideration for possible reconstitution o f j u n c tional channels f r o m individual hemichannels. Unfortunately, the data s u m m a r i z e d above are not informative o n this point. H a r s h conditions are r e q u i r e d to " s p l i t " apart isolated junctional membranes (7, 79). H o w e v e r , any junctional structures that were easily solubilized, as i n the affinity-purified preparation, w o u l d not have b e e n recovered as isolated junctional m e m branes a n d so might not require such harsh conditions for splitting. T h e starting material for the affinity purification was predominantly plasma m e m b r a n e f r o m presumably w e l l - c o u p l e d cells; therefore, the c o n nexin that is isolated is most likely f r o m junctional structures. Because the p u r i f i e d connexin-32 is predominantly i n structures the size o f single hemichannels, a simple conclusion is that octylglucoside can disrupt the end-to-end interactions o f junctional channels. H o w e v e r , there is the caveat that the junctions may split prior to solubilization, that is, i n the preparation o f the crude plasma m e m b r a n e fraction. Junctions can split w h e n a tissue is perfused w i t h hyperosmotic solutions (117), so a " c e l l u l a r " splitting c o u l d precede disruption o f junctional channels b y octylglucoside.

Future Studies T h e fundamental problems regarding reconstitution o f connexin channels have b e e n overcome: connexin forms channels i n unilamellar liposomes a n d planar bilayers. Size, permeability, and gating behavior are consistent w i t h conducting units that are single h e m i c h a n n e l s — t h e structures that span a single c e l l m e m b r a n e a n d f o r m one-half o f the junctional channel. C o n n e x i n can be obtained b y affinity purification u n d e r nondenaturing conditions. T h u s , channels f o r m e d by a single connexin can be studied i n a w e l l - d e f i n e d a n d accessible system. W h a t are some o f the important biophysical questions that can be addressed? T h e permeability a n d selectivity o f the channel can be character i z e d i n detail. T h e selectivity to large, charged molecules can be assessed. A l s o the nature o f permeation (electrodiffusive versus barrier models) can be assessed for various permeants. T h e voltage sensitivity can be explored, particularly w i t h regard to the location o f the dipoles or charges involved. I n the single-membrane channels, the voltage sensitivity can be explored without the c o n f o u n d i n g presence o f another gate i n series, as i n the junctional channel (see reference 51). It may be possible to determine the change i n v o l u m e i n the l u m e n o f the pore d u r i n g gating transitions (118, 119), a n d so



10.

HARRIS

Connexin-32

219

to distinguish the two models o f gating proposed f r o m structural data (30, 31). M o d u l a t i o n o f t h e channels c a n b e studied b y exposing the bilayer channels to agents that regulate junctional channels (e.g., kinases, cyclic nucleotides) a n d other agents that elucidate the function o f specific domains or regulatory sites (e.g., specific peptidases, sequence-specific antibodies, group-specific reagents). It w i l l b e important f o r biophysical a n d biological studies to establish i n the bilayer system precisely w h i c h cellular components or covalent modifications alter the rapidly fluctuating transitions p r o d u c e d b y the affinity-purified connexin. Presumably other connexins c a n b e similarly p u r i f i e d a n d studied i n bilayers to provide a solid basis o f data o n the relation between specific amino acid sequences a n d single-channel physiology. Variants o f t h e connexins p r o d u c e d b y the techniques o f molecular biology c a n also b e reconstituted, and careful comparison o f the channel behavior i n bilayers w i t h that seen i n cellular expression systems w i l l b e fruitful. It is h o p e d that such studies o f single hemichannels w i l l b e c o m p l e mented b y studies o f reconstituted junctional channels: t h e d o u b l e - m e m brane f o r m . D e v e l o p m e n t o f a stable, well-characterized, a n d w e l l - c o n t r o l l e d double-membrane system is a challenging prospect. T h e literature o n osmotic control o f fusion o f apposed b u l g e d bilayers m a y b e h e l p f u l i n this regard (120, 121). S u c h a system w o u l d p e r m i t exploration o f the forces i n v o l v e d i n the assembly o f junctional channels, w h i c h w o u l d b e o f interest f r o m biophys ical a n d cellular perspectives. F o r example, does the space between t w o membranes n e e d to b e dehydrated f o r hemichannels t o interact (122)? D o the hemichannels find each other b y r a n d o m interactions o r does dielectric attraction (123-125) play a role? O n c e junctional channels are f o r m e d , h o w reversible is the interaction between t h e m a n d what forces t e n d to stabilize it? M o s t important, h o w are t h e permeability, gating, a n d modulation o f single hemichannels altered b y interactions w i t h each other i n the d o u b l e m e m b r a n e f o r m ? These a n d other considerations make t h e exploration o f connexin channels i n reconstituted systems o f p r o f o u n d interest a n d p r o m i s ing prospects.

Acknowledgments T h e various experiments s u m m a r i z e d here were carried out over the last f e w years i n collaboration w i t h C a r v i l l e Bevans, Sergei Bezrukov, D a n i e l G o o d enough ( N a t i o n a l Institutes o f H e a l t h G r a n t G M 1 8 9 7 4 ) , D a v i d P a u l ( N a tional Institutes o f H e a l t h G r a n t G M 3 7 7 5 1 ) , Seung K . R h e e ( A m e r i c a n L i v e r F o u n d a t i o n Student Research Fellowship), A n n e W a l t e r , a n d Joshua Z i m merberg. T h e i r experimental a n d intellectual contributions are most grate fully appreciated. T h e author expresses appreciation to V . A . Parsegian for his


220



unwavering support a n d his contagious spirit o f i n q u i r y a n d exploration. Research was supported b y Office o f N a v a l Research G r a n t N00014-90-J-1960 a n d N a t i o n a l Institutes o f H e a l t h G r a n t G M 3 6 0 4 4 to the author a n d N a t i o n a l Institutes o f H e a l t h B i o m e d i c a l Research Support G r a n t S 0 7 R R 0 7 0 4 1 to Johns H o p k i n s University. Research was p e r f o r m e d w i t h equipment a n d sup plies p r o v i d e d , i n part, b y the M i l l i p o r e C o r p o r a t i o n .

References 1. Neher, E.; Sakmann, B. Nature (London) 1976, 260, 799-802. 2. Mueller, P.; Rudin, D. O. J. Theor. Biol. 1968, 18, 222-258. 3. Bean, R. C.; Shepherd, W. C.; Chan, H.; Eichner, J. J. Gen. Physiol. 1969, 53, 741-757. 4. Spray, D. C.; Bennett, M. V. L. Annu. Rev. Physiol. 1985, 47, 281-303. 5. Gap Junctions; Hertzberg, E. L.; Johnson, R. G., Eds.; Alan R. Liss: New York, 1988. 6. Biophysics of Gap Junction Channels, Peracchia, C., E d . ; CRC Press: Boca Raton, FL, 1991. 7. Goodenough, D. Α.; Paul, D. L.; Jesaitis, L. J. Cell Biol. 1988, 107, 1817-1824. 8. Milks, L. C.; Kumar, N. M.; Houghton, R.; Unwin, N.; Gilula, Ν. Β. EMBO J. 1988, 7, 2967-2975. 9. Yeager, M.; Gilula, Ν. B. J. M o l . Biol. 1992, 223, 929-948. 10. Stauffer, K. A.; Unwin, N. Semin. Cell Biol. 1992, 3, 17-20. 11. Beyer, E. C.; Paul, D. L.; Goodenough, D. A. J. Membr. Biol. 1990, 116, 187-194. 12. Kumar, N. M.; Gilula, Ν. B. Semin. Cell Biol. 1992, 3, 3-16. 13. Lawrence, T. S.; Beers, W. H.; Gilula, Ν. B. Nature (London) 1978, 272, 501-506. 14. Bennett, M. V. L. In Handbook of Physiology: The Nervous System; Kandel, E., E d . ; Williams and Wilkins: Baltimore, MD, 1977; Vol 1, Sec. I, pp 357-416. 15. Lasater, E. M.; Dowling, J. E. Proc. Natl. Acad. Sci. U.S.A. 1985, 82, 3025-3029. 16. Fraser, S. Ε.; Green, C. R.; Bode, H. R.; Gilula, Ν. B. Science (Washington, D.C.) 1987, 237, 49-55. 17. Saez, J. C.; Connor, J. A.; Spray, D. C.; Bennett, M. V. L. Proc. Natl. Acad. Sci. U.S.A. 1989, 86, 2708-2712. 18. Pitts, J. D. In Intercellular Junctions and Synapses. Receptors and Recognition; Feldman, J.; Gilula, N. B.; Pitts, J. D., Eds.; Chapman and Hall: London, 1978; Series B, V o l . 2, pp 63-79. 19. Bennett, M. V. L.; Goodenough, D. Α., Eds. Neurosci. Res. Prog. Bull. 1978, 16, 373-486. 20. Bennett, M. V. L.; Spray, D. C.; Harris, A. L. In Perspectives in Differentiation and Hypertrophy; Anderson, W. Α.; Sadler, W., Eds.; Elsevier: New York, 1982; pp 103-119. 21. Loewenstein, W. R.; Rose, B. Semin. Cell Biol. 1992, 3, 59-79. 22. Warner, A. Semin. Cell Biol. 1992, 3, 81-91. 23. Tanabe, T.; Beam, K. G.; Adams, Β. Α.; Niidome, T.; Numa, S. Nature (London) 1990, 346, 567-569. 24. Goodenough, D. A.; Stoeckenius, W. J. Cell Biol. 1972, 54, 646-656. 25. Hertzberg, E. L.; Gilula, Ν. B. J. Biol. Chem. 1979, 254, 2138-2147.



10.

HARRIS

Connexin-32

221

26. Hertzberg, E. L. J. Biol. Chem. 1984, 259, 9936-9943. 27. Baker, T. S.; Caspar, D. L. D.; Hollingshead, C. J.; Goodenough, D. A. J. Cell Biol. 1983, 96, 204-216. 28. Sosinsky, G. E. Electr. Microsc. Rev. 1992, 5, 59-76. 29. H o h , J. H.; L a l ,R.;John, S. Α.; Revel, J.-P.; Arnsdorf, M. F. Science (Washing ton, D.C.) 1991, 253, 1405-1408. 30. Makowski, L.; Caspar, D. L. D.; Phillips, W. C.; Goodenough, D. A. J. Mol. Biol. 1984, 174, 449-481. 31. Unwin, P. N. T.; Zampighi, G. Nature (London) 1980, 283, 545-549. 32. Unwin, P. N. T.; Ennis, P. D. J. Cell Biol. 1984, 97, 1459-1466. 33. Beyer, E. C.; Steinberg, T. H. J. Biol. Chem. 1991, 266, 7971-7974. 34. DeVries, S. H.; Schwartz, E. A. J. Physiol. 1992, 445, 201-230. 35. Paul, D. L.; Ebihara, L.; Takemoto, L. J.; Swenson, K. I.; Goodenough, D. A. J. Cell Biol. 1991, 115, 1077-1089. 36. Ebihara, L.; Steiner, E. Biophys. J. 1992, 61, A408. 37. Lauger, P. Biochim. Biophys. Acta 1973, 311, 423-441. 38. Lauger, P. Biochim. Biophys. Acta 1976, 445, 493-509. 39. Flagg-Newton, J. L.; Loewenstein, W. R. J. Membr. Biol. 1979, 50, 65-100. 40. Brink, P. R.; Dewey, M. M. Nature (London) 1980, 285, 101-102. 41. Neyton, J.; Trautmann, A. Nature (London) 1985, 317, 331-335. 42. Veenstra, R. D.; DeHaan, R. L. Science (Washington, D.C.) 1986, 233, 972-974. 43. Burt, J. M.; Spray, D. C. Proc. Natl. Acad. Sci. U.S.A. 1988, 85, 3431-3434. 44. Somogyi, R.; Kolb, H.-A. Pflugers Arch. 1988, 412, 54-65. 45. Bennett, M. V. L.; Verselis, V. K. Semin. Cell Biol. 1992, 3, 29-47. 46. Eghbali, B.; Kessler, J. Α.; Spray, D. C. Proc. Natl. Acad. Sci. U.S.A. 1990, 87, 1328-1331. 47. Moreno, A. P.; Fishman, G. I.; Spray, D. C. Biophys. J. 1992, 62, 51-53. 48. Bennett, M. V. L.; Verselis, V.; White, R. L.; Spray, D. C. In Modern Cell Biology; Hertzberg, E. L.; Johnson, R. G., Eds.; Alan R. Liss: New York, 1988; Vol. 7, pp 287-304. 49. Spray, D. C.; Harris, A. L.; Bennett, M. V. L. J. Gen. Physiol. 1981, 77, 75-94. 50. Jaslove, S. W. In Biophysics of Gap Junction Channels; Peracchia, C., E d . ; CRC Press: Boca Raton, FL, 1991; pp 145-162. 51. Harris, A. L.; Spray, D. C.; Bennett, M. V. L. J. Gen. Physiol. 1981, 77, 95-120. 52. Finkelstein, Α.; Peskin, C. S. Biophys. J. 1984, 46, 549-558. 53. Barrio, L. C.; Suchyna, T.; Bargiello, T. Α.; X u , L. X.; Roginski, R. S.; Bennett, M. V. L.; Nicholson, B. J. Proc. Natl. Acad. Sci. U.S.A. 1991, 88, 8410-8414. 54. Rubin, J. B.; Verselis, V. K.; Bennett, M. V. L.; Bargiello, T. A. Biophys. J. 1992, 62, 183-195. 55. Rubin, J. B.; Verselis, V. K.; Bennett, M. V. L.; Bargiello, T. A. Proc. Natl. Acad. Sci. U.S.A. 1992, 89, 3820-3824. 56. Catterall, W. A. Science (Washington, D.C.) 1988, 242, 50-61. 57. Brink, P. R.; Fan, S. Biophys. J. 1990, 56, 579-593. 58. Veenstra, R. D.; Wang, Η. Z.; Westphale, Ε. M.; Beyer, E. C. Circ. Res. 1992, 71, 1277-1283. 59. Paul, D. L. J. Cell Biol. 1986, 103, 123-134. 60. Kumar, N. M.; Gilula, Ν. B. J. Cell Biol. 1986, 103, 767-776. 61. Nicholson, B. J.; Zhang, J. In Modern Cell Biology; Hertzberg, E. L.; Johnson, R. G., Eds.; Alan R. Liss: New York, 1988; Vol. 7, pp 207-218. 62. Traub, O.; Look, J.; Dermietzel, R.; Brummer, F.; Hulser, D.; Willecke, K. J. Cell Biol. 1989, 108, 1039-1051. 63. Kordel, M.; Nicholson, B. J.; Harris, A. L. Biophys. J. 1993, 64, A192.



222

BIOMEMBRANE

ELECTROCHEMISTRY

64. Werner, R.; Levine, E.; Rabadan-Diehl, C.; Dahl, G. Proc. Natl. Acad. Sci. U.S.A. 1989, 86, 5380-5384. 65. Swenson, Κ. I.; Jordan, J. R.; Beyer, E. C.; Paul, D. L. Cell 1989, 57, 145-155. 66. Spray, D. C.; Bennett, M. V. L.; Campos de Carvalho, Α.; Eghbali, B.; Moreno, A. P.; Verselis, V. In Biophysics of Gap Junction Channels; Peracchia, C., Ed.; CRC Press: Boca Raton, FL, 1991; pp 97-116. 67. Verselis, V. K.; White, R. L.; Spray, D. C.; Zavilowitz, J.; Bennett, M. V. L. J. Cell Biol. 1987, 105, 307a. 68. Chen, Y. H.; DeHaan, R. L. J. Membr. Biol. 1992, 127, 95-111. 69. Manivannan, K.; Ramanan, S. V.; Mathias, R. T.; Brink, P. R. Biophys. J. 1992, 61, 216-227. 70. Stauffer, K. A.; Kumar, Ν. M.; Gilula, Ν. B.; Unwin, N. J. Cell Biol. 1992, 115, 141-150. 71. Lampe, P. D.; Kistler, J.; Hefti, Α.; Bond, J.; Muller, S.; Johnson, R. G.; Engel, A. J. Struct. Biol. 1991, 107, 281-290. 72. Abney, J. R.; Braun, J.; Owicki, J. C. Biophys. J. 1987, 52, 441-454. 73. Ceriotti, Α.; Coleman, A. J. Cell Biol. 1990, 111, 409-420. 74. Paulson, H. L.; Ross, A. F.; Green, W. N.; Claudio, T. J. Cell Biol. 1991, 113, 1371-1384. 75. Loewenstein, W. R. Physiol. Rev. 1981, 61, 829-912. 76. Beyer, E. C.; Kistler, J.; Paul, D. L.; Goodenough, D. A. J. Cell Biol. 1989, 108, 595-605. 77. Musil, L. S.; Goodenough, D. A. J. Cell Biol. 1991, 115, 1357-1374. 78. Dahl, G.; Werner, R.; Levine, E.; Rabadan-Diehl, C. Biophys. J. 1992, 62, 172-182. 79. Manjunath, C. K.; Goings, G. E.; Page, E. J. Membr. Biol. 1984, 78, 147-155. 80. Dupont, Ε.; el Aoumari, Α.; Briand, J. P.; Fromaget, C.; Gros, D. J. Membr. Biol. 1989, 108, 247-252. 81. John, S. Α.; Revel, J.-P. Biochem. Biophys. Res. Commun. 1991, 178, 1312-1318. 82. Rahman, S.; Evans, W. H. J. Cell Sci. 1991, 100, 567-578. 83. Lis, L. J.; McAlister, M.; Fuller, N.; Rand, R. P.; Parsegian, V. A. Biophys. J. 1982, 37, 657-666. 84. Mege, R. M.; Matsuzaki, F.; Gallin, W. E.; Goldberg, J. I.; Cunningham, Β. Α.; Edelman, G. M. Proc. Natl. Acad. Sci. U.S.A. 1988, 85, 7274-7278. 85. Musil, L. S.; Cunningham, Β. Α.; Edelman, G. M.; Goodenough, D. A. J. Cell Biol. 1990, 111, 2077-2088. 86. Miller, C., E d . Ion Channel Reconstitution; Plenum: New York, 1986. 87. Harris, A. L. In Biophysics of Gap Junction Channels; Peracchia, C., E d . ; CRC Press: Boca Raton, FL, 1991; pp 373-389. 88. Brewer, G. J. In Biophysics of Gap Junction Channels; Peracchia, C., E d . ; CRC Press: Boca Raton, FL, 1991; pp 301-316. 89. Makowski, L.; Caspar, D. L. D.; Phillips, W. C.; Goodenough, D. A. J. Cell Biol. 1977, 74, 629-645. 90. Cascio, M.; Gogol, E.; Wallace, B. A. J. Biol. Chem. 1990, 265, 2358-2364. 91. Lynch, E.; Harris, A. L.; Paul, D. Biophys. J. 1984, 45, 61a. 92. Spray, D. C.; Saez, J. C.; Brosius, D.; Bennett, M. V. L.; Hertzberg, E. L. Proc. Natl. Acad. Sci. U.S.A. 1986, 83, 5494-5497. 93. Young, J. D.; Cohn, Ζ. Α.; Gilula, Ν. B. Cell 1987, 48, 733-743. 94. Campos de Carvalho, A. C.; Eiras, L. Α.; Waltzman, M.; Hertzberg, E. L.; Spray, D. C. Braz. J. Med. Biol. Res. 1992, 25, 81-92. 95. Donaldson, P.; Kistler J. J. Membr. Biol. 1992, 129, 155-165.



10.

HARRIS

Connexin-32

223

96. Mazet, J. L.; Jarry, T.; Gros, D.; Mazet, F. Eur. J. Biochem. 1992, 210, 249-256. 97. Walter, Α.; Zimmerberg, J.; Paul, D.; Harris, A. L. Soc. Neurosci. Abstr. 1988, 12, 1191. 98. Rhee, S. K.; Paul, D. L.; Harris, A. L. FASEB J. 1989, 3, A602. 99. Dupont, Ε.; el Aoumari, Α.; Fromaget, C.; Briand, J.-P.; Gros, D. Eur. J. Biochem. 1991, 200, 263-270. 100. Ehring, G. R.; Kumar, N. M.; Stauffer, K. A.; Unwin, P. N. T.; Gilula, Ν. B.; Hall, J. E. Biophys. J. 1992, 61, A409. 101. Harris, A. L.; Park, J.; Balakrishnan, K.; Bevans, C.; Rhee, S.; Paul, D. Biophys. J. 1988, 53, 507a. 102. Harris, A. L.; Walter, Α.; Paul, D.; Goodenough, D. Α.; Zimmerberg, J. J. Mol. Brain Res. 1992, 15, 269-290. 103. Rhee, S. K.; Harris, A. L. Biophys. J. 1991, 59, 439a. 104. Mimms, L. T.; Zampighi, G.; Nozaki, Y.; Tanford, C.; Reynolds, J. A. Biochem istry 1981, 20, 833-840. 105. Harris, A. L.; Walter, Α.; Zimmerberg, J. J. Membr. Biol. 1989, 109, 243-250. 106. Schein, S. J.; Colombini, M.; Finkelstein, A. J. Membr. Biol. 1976, 30, 99-120. 107. Hertzberg, E. L.; Skibbens, R. V. Cell 1984, 39, 61-69. 108. Stewart, W. W. Cell 1978, 14, 741-759. 109. Bennett, M. V. L.; Spira, M. E.; Spray, D. C. Dev. Biol. 1978, 65, 114-125. 110. Cohen, F. S.; Zimmerberg, J.; Finkelstein, A. J. Gen. Physiol. 1980, 75, 251-270. 111. Zimmerberg, J.; Cohen, F.C.; Finkelstein, A. J. Gen. Physiol. 1980, 75, 241-250. 112. Ehrenstein, G.; Lecar, Η. Q. Rev. Biophys. 1977, 10, 1-34. 113. Labarca, P.; Coronado, R.; Miller, C. J. Gen. Physiol. 1980, 76, 397-424. 114. Bevington, P. R. Data Reduction and Error Analysis for the Physical Sciences; McGraw-Hill: New York, 1969; pp 180-182. 115. Buisman, H. P.; Steinberg, T. H.; Fischbarg, J.; Silverstein, S. C.; Vogelzang, S. Α.; Ince, C.; Ypey, D. L.; Leijh, P. C. J. Proc. Natl. Acad. Sci. U.S.A. 1988, 85, 7988-7992. 116. Tatham, P. E. R.; Lindau, M. J. Gen. Physiol. 1990, 95, 459-476. 117. Goodenough, D. Α.; Gilula, Ν. B. J. Cell Biol. 1974, 61, 575-590. 118. Zimmerberg, J.; Parsegian, V. A. Nature (London) 1986, 323, 36-39. 119. Zimmerberg, J. Α.; Bezanilla, F.; Parsegian, V. A. Biophys. J. 1990, 57, 1049-1064. 120. Fisher, L. R.; Parker, N. S. Biophys. J. 1984, 46, 253-258. 121. Lucy, J. A.; Ahkong, Q. F. FEBS Lett. 1986, 199, 1-11. 122. Rand, R. P.; Parsegian, V. A. Biochim. Biophys. Acta. 1989, 988, 351-376. 123. Lifshitz, Ε. M. Sov. Phys. 1956, 2, 73-83. 124. Parsegian, V. A.; Ninham, B. W. Nature (London) 1969, 224, 1197-1198. 125. Ninham, B. W.; Parsegian, V. A. Biophys. J. 1970, 10, 646-663. 126. Green, C. R.; Harfst, E.; Gourdie, R. G.; Severs, N. J. Proc. R. Soc. Lond. Ser. Β 1988, 233, 165-174. RECEIVED for review F e b r u a r y 20, 1991. ACCEPTED revised manuscript January 26, 1993.


Connexin-32: A Protein That Forms Channels through One or Two

Recommend Documents