H+ antiporter which use similar transport and H+ coupling mechanisms

J. theor. Biol. (1991) 150, 239-249

Proposed Partial fl-structures for Lac Permease and the N a + / H + Antiporter which use Similar Transport and H + Coupling Mechanisms WILSON RADD1NGt

Department of Physiology and Biophysics, University of Alabama at Birmingham, Birmingham, Alabama 35294, U.S.A. (Received on 29 June 1990, Accepted in revised form on 30 October 1990) Most antiporters, symporters, and transporters have been represented as containing ten to 14 transmembrane helices, primarily on the basis of hydropathy plots. However, multihelix systems provide no obvious mechanism of transport and no simple way of distinguishing substrates. The models oflac permease and the Na+/H ÷ antiporter presented here postulate that/3-structures are involved in the transport of substrate, and in following this postulate arrive at readily understandable mechanisms for transport and for substrate specificity. The percentage of /3-structures necessary for these models is low enough that it is not in conflict with prior physical evidence for secondary structures. Immunological data also cannot rule these/3structure mechanisms invalid. In lac permease the new model is obtained by formal representation of the C-terminal amino acids 243-405 as /3-strands. This formal representation nets two interchangeable/3-barrels which provide a simple mechanism for sugar transport. The alternating barrel system may comprise as little as 1/5 the entire permease. In one configuration the barrel forms a pocket with hydrogen bonding residues oriented to the outside of the cell. In the other configuration the barrel forms an analogous pocket oriented towards the inside. Six particular amino acids participate in the substrate hydrogen bonding schemes of both forms, providing a mechanism to shuttle lactose from the outside to the inside or vice versa. A trigger for change of forms which could couple the /3-barrel to H+-transport is easily devised, and it involves the apparently critical His322-Glu325 charge relay system. The N a + / H + antiporter can be organized similarly with an interchanging/3-barrel-/3clamshell structure attached to 7-transmembrane helices. Charged amino acid sidechains form the basis of an ionic shuttle which is analogous to the lactose shuttle. In this case, too, coupling of Na ÷ transport to H ÷ transport may be accomplished by a histidine-glutamate charge relay system. 1. Introduction Antiporters, symporters and transporters share similar hydropathy plots which are g e n e r a l l y i n t e r p r e t e d as m e a n i n g t h a t t h e y h a v e ten to 14 t r a n s m e m b r a n e helices. T h e p u t a t i v e helix b u n d l e s a r e a s s u m e d to b e i n v o l v e d in t r a n s p o r t , b e c a u s e the r e m a i n i n g h y d r o p h i l i c r e g i o n s a r e v a r i a b l e , o f t e n s m a l l , a n d f r e q u e n t l y have h o m o l o g y to f u n c t i o n a l entities w h i c h m o d u l a t e m o l e c u l a r b e h a v i o r such as p h o s p h a t e a c c e p t i n g r e g i o n s o r n u c l e o t i d e b i n d i n g subunits. H o w e v e r , t h e helix m o d e l d o e s n o t give a s i m p l e , r e a s o n a b l e m o l e c u l a r m e t h o d for d i s c r i m i n a t i n g s u b s t r a t e s t Present address: Department of Pathology, University of Alabama at Birmingham, Birmingham, Alabama 35294, U.S.A. 0022-5193/91/100239+ 11 $03.00/0

© 1991 Academic Press Limited

240

w. RADDING

or for transporting them. Further, the physical, antibody based, and site directed mutagenesis derived evidence does not conclusively demonstrate that helix bundles completely describe the hydrophobic structures of these molecules. This article demonstrates how g-structures can provide reasonable mechanisms for substrate binding and transport in lac permease and the Na+/H ÷ antiporter. An underlying similarity of mechanism can be found, but the detailed techniques for binding and transport are different for the sugar and the Na ÷. The biochemical basis for considering symporters, antiporters and transporter members of the same superfamily (Sardet et ai., 1989) is not without flaw. Among these molecules there are some homology-related families and some families related by function, but there are also transport involved proteins with the usual hydropathy plots which lie outside these families. For example, one homology based family of sugar mobilizing proteins includes facilitative transporters, ATP driven transporters and hydrogen ion transmembrane potential linked transporters (Szkutnicka et al., 1989; Maiden et al., 1987). Another functionally organized family substantially overlapping the sugar transporters includes nucleotide phosphate hydrolysis driven transporters and their relatives (Riordan et al., 1989) which appear similar to the multidrug resistance p-glycoprotein (Gros et aL, 1986; Gerlach et al., 1986) and the yeast pheromone exporter (Kuchter et al., 1989; McGrath & Varshavsky, 1989). Interestingly, when homology is the final criterion, lac permease, one of the best characterized symporters (Buchel et aL, 1980; Kaback, 1986, 1988), does not appear to belong to these families, nor does the Na÷/glucose symporter (Hediger et al., 1987), the H+/melibiose symporter (Yazyu et al., 1984), the C I - / H C O ; antiporter (Kopito & Lodish, 1985) or the Na÷/H ÷ antiporter (Sardet et al., 1989). The existence of these molecules, which are not demonstrably of an antiporter, symporter or transporter family, reinforces the impression given by the diversity of substrates, that the hydropathy representation, this "pictographic molecular paradigm" (Luisi & Thomas, 1990), might be misleading researchers by emphasizing possible similarity in structure while obscuring important detail. Other authors have questioned the tendency of researchers to conclude that transmembrane protein hydrophobic regions of the appropriate length are helices (Ferenci, 1989; Fasman & Gilbert, 1990). They are supported by the fact that there are transmembrane non-transport proteins which are formed primarily of g-structure (Vogel & Jahnig, 1986; Charbit et aL, 1988; Ferenci et al., 1988) and by the fact that fourier transform infra-red data on the human erythrocyte glucose transporter, a member of the sugar transporter family, give unequivocal evidence of substantial g-structure (Alvarez et aL, 1987). Therefore, it would seem worthwhile to investigate how putative helical arrays might be transformed into g-structures. As few as six amino acids in a g-strand are necessary to cross a membrane (Paul & Rosenbusch, 1985) and ten is certainly enough. Consequently almost any putative transmembrane helix, embedded in a hydrophobic region of 20 or more amino acids, can also represent two g-strands plus a loop to take care of excess chain length. That hydrophilic g-turns might be predicted to interleave with the hydrophobic regions does not alter this possible g-strand interpretation, because the hydrophilic sequence may be short enough and contain few enough charged residues

/B-STRUCTURES

IN PERMEASE

AND

ANTIPORTER

241

to have only marginal influence on the hydropathy plot. Even if the hydrophobicitypredicted helices are also predicted to be helices by C h o u - F a s m a n type criteria (Chou & Fasman, 1974; Garnier et al., 1978), the/3-strand alternative may still be a possibility, because there are so few transmembrane proteins in the data base which underlies these predictions. A putative 12-helix system might therefore actually be a 7-helix system plus a 10-stranded/3-structure, most likely a barrel or clamshell. Application of this reasoning to lac permease and to the N a ÷ / H ÷ antiporter has resulted in the models below, which are consistent with one another, which do not flagrantly contradict present structural data, and which have obvious reasonable mechanisms for coupled transport. 2. Lac Permease

Structural data for lac permease predict that it is between 50% and 85%-helix (Foster et al., 1983; Vogel et al., 1985). If lac permease is more nearly 50%- than 85%-helix, there would be plenty of molecule from which to form an active fl-structure. Circular dichroism studies which indicated 85%-helix did not show the standard double minimum at 208 and 222 nm (Foster et al., 1983), and so they were based on negative ellipticity at 223 nm (Bewley et al., 1964). When a molecule contains a large number of phenylalanines, as lac permease does (Buchel et al., 1980), this technique can lead to a serious overestimate of helical content, because in low dielectric media the phenylalanine sidechain can produce strong negative ellipticity in the 215-223 nm region (Radding, 1988). Consequently this high estimate o f helix may not convey the actual percentage. The Raman secondary structure determination (Vogel et al., 1985) estimated "ordered helix", the largest component and probably the easiest to estimate, at 53%. However, from preparation-to-preparation of this transmembrane protein percentages assigned to any substructure varied by as much as 10% (Vogel et al., 1985). Thus, in a second preparation " o r d e r e d " helix was estimated at 42%. Since the "ordered helix" category ignores two residues on each end of a helix, the second preparation could assign as few as 51% of the residues to a set of helices, each long enough to cross the membrane. In lac permease only 110 amino acids, 26% of the molecule, are necessary to form the ll-stranded fl-structure proposed below. The Raman data do not rule this possibility out. In fact the 42%-"ordered helix" data could be considered to support it, since nearly 50% of the molecule is available for something other than a transmembrane polyhelical array. The other categories of secondary structure are assigned relatively low percentages, and are therefore more subject to uncertainty derived from noise in the original spectrum. It should also be noted that the data base on which these estimates rest (Williams, 1983) does not include an antiparallel /3-barrel such as that proposed below. Thus, neither CD nor Raman spectra rule out the model proposed here. Of the many documented fl-structures (Richardson, 1981) the antiparallel /3-clamshell and antiparallel fl-barrel would appear to be most able to transport lactose through the membrane. A/3-clamshell could transport sugar in two ways: it could open to one side of the membrane, receive substrate, and then turn to the

242

w. RADDING

other side and release substrate, or it could be oriented with its opening essentially in the membrane, receiving substrate in one side of its mouth and releasing it out the other. The first mechanism requires a polypeptide idler arm reminiscent of the linkage between the steering wheel shaft and the front end assembly of many motor vehicles. It seems unlikely. The second system requires the perpendicular/3-strands of the clamshell to rearrange in a very complex way in order to sucessfully transport the sugar. It, too, seems unlikely. The/3-barrel structure is simpler, and results in a readily visualized, realistic mechanism. For the purpose of introducing/3-structure as a possible alternative to helix, the amino icids 243-405 are drawn in antiparallel strands in Fig. l(a) and (b). The strands are all more than long enough to cross a membrane, a fact which is important,

(o)

(b)

Out

Out

rK'~ ~'~

f-h

F

Q

fA'~

F

L

V

L

C

A

L

G

,,,T ,, f / T M

V

C

L

F

f-h

H

M

Y

C

V

L

L

F

i

F

L

G

T~'/'E> T

F

Y

L V

,I,

E

L

F

G

L

FR~

A

M

N

V

I

fF~

F

M

G

F

V

G

I

M K

A

K

A

N

T

A

S

I

M/'/E>

C

1

L

T

Y

E

T

S

L

P

S

Q

G

A

v

F

V

L

~,

G

L

G

V

I

F

F

L

Q

L

M

L

I

F

F

L

A

R

V

L

R

A

F

G

T

K

T

V

M

V

Y

I

L

N

F

A

G

T

F( T s-M)o @

f,~,~

~

F

V

1

I

S

L

P

L

T"

Y

M"

T

S"

T

F

Y

S

G

F

G

F

K

L

I

F

V

A

L

T

I

V

A

A

R

P

L

S

V

G

S

T

T

L

L

A

V

C

F

N

G

S

S

E

F

K

F

G

G

I

A

L

E

F

R

M

I

G

F

G

F

G

' A '/

E

O

~

S

P

F

T

A

G

G

k'A/)

P

I

L

R

S

N

G

A

I

N

K

V

Y

N

K

Y

E

~"E)

R

G

I

F

I

G

T

O

k._J In

A

P

I

\s j

k,,.j)

~

L

I) In

FIG. 1. Diagram of the C-terminal amino acids 243-405 o f lac permease as antiparallel fl-sheet. (a) the /3-structure organized to interact with sugar on the outside of the membrane. (b) The alternate configuration which would interact with sugar on the inside of the membrane. Note that in forming a barrel the right-hand strand in the figure will lie beside the left-hand one. Starred amino acids are those which remain in contact with the sugar in both configurations. Solid lines forming ovals encircle hydrogen bonding domains. Diamonds show hydrophobic residues which can form a molecular stop. The rectangle outlines interacting residues of the Glu-His trigger in its two major recta-stable forms.

/3-STRUCTURES

IN PERMEASE

AND

ANTIPORTER

243

because antiparallel barrels need loops to allow irregular twist between strands (Chothia et al., 1977; Salemme, 1983). Most interesting, however, is the fact that this purely formal representation of the antiparallel B-barrel provides a mechanism for the transport of lactose. The B-barrel can take two forms, one with Glu325-His322 mating [Fig. l(a)] and one with Glu325-Lys319 mating [Fig. l(b)]. In the former [Fig. 1(a)] the strands are arranged so that there are two bands of hydrogen bonding residues. Since there is an even number of amino acids from band-to-band, both sets of hydrogen bonding sidechains can be inside the barrel. Deeper in the barrel is an almost complete band of hydrophobic residues, again in register with the hydrogen bonding sidechains so that all can be inside the barrel. Thus, there is formed a pocket with 13 sticky fingers which can contact the lactose, and the pocket has an impenetrable bottom necessary to keep lactose from merely diffusing into the cell. All the important features of this model can be obtained with nine B-strands as well as 11 B-strands, reducing the percentage of the permease necessarily devoted to B-strands to even less than the 26% mentioned in the introduction. As would be essential for a transport system, the second form [Fig. l(b)] recreates the same hydrogen bonding bands in the same orientation and again achieves a hydrophobic band in the middle of the B-barrel. Six of the hydrogen bonding residues are identical in both configurations. Hydrogen bonding residues which are not in this set of six come as pairs, an odd number of amino acids apart. This arrangement ensures that these alternate hydrogen bonding amino acids must release their contact with the lactose when the configuration changes, because these sidechains will presumably point outside the barrel, while the new hydrogen bonding sidechains will be inside. Recent molecular biological studies showing that lac permease with Tyr382 replaced by phenylalinine catalyses lactose transport at a reduced rate (Roepe & Kaback, 1989) support this picture. This system has the potential for confounding antibody approaches for determining secondary structure and topology. For example, one of the antibodies used to try to establish helicity for amino acid sequence 320-344 was raised against FAFAGAYAQA (Bieseler et al., 1986). In this model there are two ways that a hexapeptide of this hapten can be approximated by closely juxtaposed tripeptides: FCF--.GAY [Fig. 1(b)] or FAN..-GAY [Fig. 1(b)]. Cross-reactivity with a polyclonal antibody raised against the initial peptide seems likely. A similar argument pertains to the second helicity determining antibody, LGSYISAVR (Bieseler et al., 1986), where on one loop of the B-barrel the seqnence is YI-SA--VR. In fact, because of the proposed mobility of the residues, a great number of antibodies are likely to react positively, but not well, in both secondary structure and topological determinations. A virtue of this mechanism is that it incorporates a Glu-His charge relay as a trigger for the configurational change. A schematic of states (Fig. 2) shows how the lactose and proton transport can be linked, if proton transport is regarded as the province of the seven helical array except at one critical point where the Glu-His relay intervenes. In this mechanism protons from outside the membrane will force the glutamate out of Lysll9 apposition and into His322 apposition, if sugar is in the outside pocket. This change will trigger the configurational change, pulling the

244

RADDING

w.

sugar into the inside pocket. This scheme can also account for sequential release of sugar, then protein, to the outside (Kaback, 1986), because, when sugar is in the outside pocket but His325 is not protonated, the histidine is outside the barrel in a useless position. The final transition to a totally relased state (top of Fig. 2) faces a large free energy barrier which is not present if the release of sugar is accomplished first. When lactose approaches the inside pocket where hydrogen ion does not interact directly with the sugar pocket, the two substrates can enter or leave their

H S ~ - H

No sugor No H+

/E-'= tfK ÷ / /

,.-" E, x

.osogOro K+ out of pocket

k x \ H* \

H E-

S

Sugoro ÷ No Ho

K+ ~. f.~Sugor

Sugoro _.

H+ k

\

\ \k

K÷

~

K+ position intermediote but mostly in pocket

\

k

%____

H

Sugor

",\ ) t ¢,, -~I .,', /"

Sugorz H÷

/

( K+."

K+in pocket r

";m°sug°r'.

H

S

~K~

Sugarz No H~-

H H÷

S

// 1 / K," ~. / k /

9

NO sugor

NO H*

FIG. 2. Diagram of possible metastable states for lac permease. Only the immediate surroundings o f the G l u - H i s relay are s h o w n to emphasize its importance in allowing or predisposing transport to proceed. H + s t a n d s for proton except in two cases in the body o f the diagram where it represents protonated histidine. The subscript I stands for the configuration with a pocket toward the interior, the subscript o for the configuration with the pocket toward the exterior. Arrows indicate major kinetic pathways. Close sidechains are circled with solid line and fairly close sidechains are circled with dotted lines. This set o f states a n d reasonable a s s u m p t i o n s about barriers between them can account for most o f lac permease transport behavior. For example, when sugar is being transported from the inside to the outside, the initial sequence o f binding, s u g a r - p r o t o n or p r o t o n - s u g a r will not make a perceptible difference. However, release to the outside can only occur in one sequence, s u g a r - p r o t o n .

/3-STRUCTURES

IN

PERMEASE

AND

ANTIPORTER

245

binding sites in either order. This mechanism is congruent with molecular biological evidence which shows that replacement of Lys319 with leucine produces effects similar to replacement of Glu325 with alanine (Roepe et al., 1990). This kind of mechanism has important implications for spectroscopy and crystallography. The model in Fig. 2 predicts as many as seven conformational isomers with similar free energies. It would not be surprising if two or three were wellpopulated in any system containing protons and galactoside. To investigate the exact nature of the system it may therefore be necessary to arrange conditions so that there is a suicide sugar substitute substrate on one side of the membrane and low pH on the other. In this way it might be possible to force the system to one state.

3. The N a ÷ / H + Antiporter

The N a + / H + antiporter can also be portrayed as a /3-structure attached to a 7-transmembrane polyhelix. In this case the /3-structure is at the N-terminal end and the helical array at the C-terminal. A configurational change which could easily allow gated transport of a cation is exactly analogous to the rearrangement of /3-strands in the preceding lac permease model. It may also be triggered by a His-Glu charge relay. In this case, however, it appears that the /3-structure alternates from /3-clamshell (Chothia et al., 1982) to/3-barrel and back. An " a p o " (no ion attached) form of the /3-structure is portrayed schematically in Fig. 3(a). If it takes ten residues to cross the membrane in a straight line perpendicular to the plane of the membrane, 14 residues will cross it at a 45 ° angle. All the strands of this apo configuration are 16-18 amino acids long, providing ample space for turns between each. Most important, however, is the fact that the tip of each loop at the extracellular side has a negatively charged resiude. When the strands are folded into a clamshell form, these charged residues can form the points of a diamond. The resultant charge potential at a reasonable distance should be effective at attracting any hydrated cation. However, this ion cannot pass while the antiporter is in this configuration, because just below but within the membrane is a layer of positive charge reminiscent of the hydrophobic band in lac permease. As a hydrated cation approaches, charge-charge interactions dictate that it will tend to shed the hydration layer, interacting ever more directly with the negative sidechains at the surface. These sidechains in turn wil be tending to pull the expanded clamshell form into a narrower /3-barrel. If the transport mechanism here is analogous to that of lac permease, the permissive trigger of configuration change will be the alteration of Glu66-Arg72 apposition to Glu66-His76 apposition. The latter requires protonation of His76, presumably critically positioned relative to the 7-transmembrane helical array. Then the/3-clamshell can become a/3-barrel (Fig. 3b) with a negative charge layer within the membrane and a positive charge layer at the surface. Three of the initial four negatively charged sidechains remain in contact with the ion, while the fourth, E70, which may take part in the extended charge relay system, is replaced by D95. The/3-barrel with charged interior would be ideal for size exclusion of larger cations. While very satisfying about entry of

246

w.

RADDING

ta) /EI84.., fF-EI04 f G - EISI T N /..,.ETO~, P I V T L P S T S G P F G T R72 RIO0 L Kt47 P Q T V P V W I F RISO I E66 V H98 I L L L L P N T L G O P t P H76 Y L G S L S D95 A V DI59 F F A V I CII3 L V Y A T T G L L F G V T DSO L M G F A V V HSI V K116 V L D172 G D59 G P I V F L T G

Out

M

F

G

A P K87 ~'~ R86 ""

F HI20

I

K84

S R55

I V

I

L

L

L5 3

1

L PI67

I PI68

L

W N

sCI

p T

A200

El31 P I

V

S S I

(b) Out

/R72\ S E70 P T

P V N H76

V H98 T Y D95

RIO0 T P F El04

t L G G V

K147 G V G EI5t

/.R 180\ L Q P F L T F EI84

V E66 P P

S V T DBO

I G L V

I S L W

L L G V

T P P F

Y G A DI72

A

H81 G M K84 P

P F A K R86

T T V D59

1 L L A CI I3 ML KII6

V I L L CI55 SEISI p

L L O I S I DI59 P V PI67 F L F L V LF I SS

I

G FHI20 V i pT

N L G T l L I F AI95

t

FIG. 3. Diagram of the N-terminal amino acids 55-200 of the N a + / H + antiporter. (a) The non-ion b o u n d state oriented toward Na + outside the m e m b r a n e . An extended chain o f the molecule, the H 120.-.C 133 loop, would allow it to become an example o f orthogonal/3-sheets, or a clamshell, although it might be a distorted one, since clamshells generally need more strands. (b) Either the ion b o u n d state with Na + in the m e m b r a n e or non-ion b o u n d state oriented toward Na + inside. Note the possibility of locking this configuration with a C113-C133 disulfide bridge.

~-STRUCTURES IN PERMEASE AND ANTIPORTER

247

the ion and its gating, this model seems less specific about its release to the interior. From the negative band of charge found four residues in from the exterior it is another four residues to a pair of negative charges D80 and D172. Were only six residues necessary to cross the membrane, the transport ion would be well inside the interior surface of the m e m b r a n e here. Two more residues suffice to total ten residues from the exterior and two more reach an area where the charged residues average to 0 or slightly positive. This arrangement is adequate for the release of uptake of ions at the interior surface of the membrane. It is clear, however, that the binding and kinetics at the interior surface will be very different from those at the exterior.

4. Conclusions and Speculation The active regions of two transmembrane transporters, lac permease and the N a + / H ÷ antiporter, can be modeled by similar overall structures, that is, they both fit a seven membered polyhelix plus a/3-structure. These models are further similar in their transport mechanisms. Both a c c o m m o d a t e G l u - H i s charge relay triggers or interlocks between lactose transport or Na ÷ and H + transport. Both achieve transport by shifting the /3-strands relative to one another. This shift can also be viewed as a translation along the polypeptide of lops in the/3-structure coupled with length changes in those loops. This mechanism is very simple and generalizable. The tryptophan synthase multienzyme, which has been crystallized, would a p p e a r to use a similar mechanism, transporting a substrate between two sites while the substrate is enclosed between two/3-sheets (Hyde & Miles, 1990). If the mechanism holds, antibody approaches to molecular topology in the/3-structure regions may easily be confounded. Furthermore, spectroscopic techniques will also be confounded by multiple states and crystals will be difficult to obtain. The fact that both these molecules fit models containing seven m e m b e r e d polyhelices indicates that these molecules may be G-protein linked. While these molecules are themselves feedback controlled by exquisite sensitivity to concentration levels of their substrates, perhaps the cell could make use of a G-protein linkage to keep an accounting of major fluxes, rather than just concentration. Such a m e m b r a n e flux accounting system could be useful in speedy identification of temporal gradients. At the least such information would be useful in avoiding concentration overshoots, but it could also provide the cell with warnings of stress or indications of the presence of storable fuel. Permease or antiporter associated G-proteins could thus be the source o f vital cues for cellular homeostasis.

Timely completion of this work has only been possible with the support of Dr Jimmy Neill, Chairman of the Department of Physiology and Biophysics, UAB. Discussions with Dr Paul Roepe of Memorial-Sloan Kettering, New York, were very helpful. The comparison of a possible His-Glu charge relay in the Na+/H ' antiporter with the putative His-Glu charge relay in lac permease was first suggested by Professor David Warnock's laboratory at the

248

w. RADDING

University o f A l a b a m a , B i r m i n g h a m ( M c D a n i e l et al., 1990). I greatly a p p r e c i a t e their help a n d suggestions.

REFERENCES

ALVAREZ, J., LEE, D. C., BALDWIN, S. A. & CHAPMAN, D. (t987). Fourier transform infrared spectroscopic study of the structure and conformational changes of the human erthrocyte glucose transporter. J. biol. Chem. 262, 3502-3509. BEWLEY, T. A., BROVETTO-CRuz, J. & LI, C. H. (1964). Human pituitary growth hormone. Physiochemical investigations of the native and reduced-alkylated protein. Biochemistry 8, 4701-4708. BIESELER, B., PRINZ, H. & BEYREUTHER, K. (1986). Topological studies of lactose permease of Escherichia coil by protein sequence analysis. Ann. IV. Y. Acad. Sci. 309-334. BUCHEL, D. E., GRONENBORN, B. & MULLER-HILL, B. (1980). Sequence of the lactose permease gene. Nature, Lond. 283, 541-545. CHARBIT, A., GEHRING, K., NIKAIDO, H., FERENCI, T. & HOFNUNG, M. (1988). Maltose transport and starch binding in phage-resistant point mutants of maltoporin..,I, molec. Biol. 201, 487-493. CHOTHIA, C. & JANIN, J. (1982). Orthogonal packing of/3-pleated sheets in proteins. Biochemistry 21, 3955-3956. CHOTHIA, C., LEVI'f'F, M. & RICHARDSON, D. (1977). Structure of proteins: Packing of a-helices and pleated sheets, proc. nam. Acad. Sci. U.S.A. 74, 4130-4134. CHOU, P. Y. & FASMAN, G. D. (1974). Prediction of protein conformation. Biochemistry 13, 222-225. FASMAN, G. D. & GILBERT, W. A. (1990). The prediction of transmembrane protein sequences and their conformation: an evaluation. T1BS 15, 89-92. FERENCI, Y. (1989). Is "'hydrophobicity analysis" sufficient to predict topography of membrane proteins? TIBS 14, 96. FERENCI, T., SAURIN, W. & HOFNUNG, M. (1988). Folding of maltoporin with respect to the outer membrane. J. molec. BioL 201, 493-496. FOSTER, D. L., BOUBLIK, M. & KABACK, H. R. (1983). Structure of the lac carrier protein of Escherichia coli. J. bioL Chem. 25g, 31-34. GARNIER, J., ASGUTHORPE, D. J. & ROBSON, B. (1978). Analysis of the accuracy and implications of simple methods for predicting the secondary structure of globular proteins. J. molec. Biol. 120, 97-120. GERLACH, .I.n., ENDICOTF, J. m., JURANKA, P. F., HENDERSON, G., SARANGI, F., DEUCHARS, K. L. & LING, V. (1986). Homology between P-glycoprotein and a bacterial heamolysin transport protein suggests a model for muttidrug resistance. Nature, Lond. 324, 485-489. GROS, P., GROOP, J. & HOUSMAN. (1988). Mammalian multidrug resistance gene: complete cDNA sequence indicates strong homology to bacterial transport proteins. Cell 47, 371-380. HEDIGER, M. A., COADY, M. J., IKEDA, T. S. & WRIGHT, E. M. (1975). Expression cloning and cDNA sequencing of the Na+/glucose co-transporter. Nature, Lond. 330, 379-381. HYDE, C. C. & MILES, E. W. (1990). The tryptophan synthase multienzyme complex: exploring structure-function relationships with X-ray crystallography and mutagenesis. Biotechnology 8, 27-32. KABACK, H. R. (1986). Active transport in Escherichia coli: passage to permease. Am. Rev. Biophys. biophys. Chem. 15, 279-319. KABACK, H. R. (1988). Site-directed mutagenesis and ion-gradient driven active transport: on the path of the proton. Ann. Rev. Physiol. 50, 243-256. KOP1TO, R. R. & LOD1SH, H. F. (1985). Primary structure and transmembrane orientation of the murine anion exchange protein. Nature, Lond. 234-238. KUCHLER, K., STERNE, R. E. & THORNER, J. (1989). Saccharomyces cerevisiae STE6 gene product: a novel pathway for protein export in eukaryotic cells. Embo J. 8, 3973-3984. LUISl, P.-L & THOMAS, R. M. (1990). The pictographic molecular paradigm. Naturwissenchaften 77, 67-74. MALDEN, M. C. J., DAVIS, E. O., BALDWIN, S. m., MOORE, D. C. M. & HENDERSON, P. J. F. (1987). mammalian and bacterial sugar transport proteins are homologous. Nature, Lond. 325, 641-643. MCDANIEL, H. B., HUANG, Z.-Q., COOK, W. J. & WARNOCK, D. G. (1990) Molecular modeling of the Na/H antiporter: the charge relay hypothesis. Kidney Int. 37, 232a. MCGRATH, J. P. & VARSHAVSKY, A. (1989). The yeast STE6 gene encodes a homologue of the mammalian multidrug resistance P-glycoprotein. Nature, Lond. 340, 400-404. PAUL, C. & ROSENBUSCH, J. P. (1985). Folding patterns of porin and bacteriorhodopsin. Embo. J. 4, i 593-1597.

/3-STRUCTURES

IN P E R M E A S E

AND

ANTIPORTER

249

RADDI NG, W. (1988). Water molecules can control the side-chain rotamer distribution of an aryl peptide in a nonpolar environment. Biophys. Chem. 30, 193-197. RICHARDSON, J. S. (1981). The anatomy and taxonomy of protein structure. Adv. Protein Chem. 34, 167-339. RIORDAN, J. R., ROMMENS, J. M., KEREM, B.-S. ALON, N., ROZMAHEL, R., GRZELCZAK, Z., Z1ELENSKI, J., LOK, S., PLAVSIC, N., CHOU, J . - L , DRUMM, M. L., IANNUZZI,M. C., COLLINS, F. S. & TSUI, L.-C. (1989). Identification of the cystic fibrosis gene: cloning and characterization of complementary DNA. Science 245, 1067-1080. ROEPE, P. D., CONSLER, T. G., MENEZES, M. E. & KABAC'K, H. R. (1990). The lac permease of Escherichia coli: Site directed mutagenesis studies on the m e c h a n i s m o f / 3 - g a l a c t o s i d e / H ÷ symport. Res. Microbiol. 141, 290-308. SALEMME, F. R. (1983). Structural properties of protein D-sheets. Prog. Biophys. molec. Biol. 42,95-133. SARDET, C., ERANCHI, A. & POUYSSEGUR, J. (1989). Molecular cloning, primary structure, and expression of the h u m a n growth factor-activatable N a + / H + antiporter. Cell 56, 271-280. SZKUTNIC'KA, K., TSC'HOPP, J. F., ANDREWS, L. & CIRILLO, V. P. (1989). Sequence and structure of the yeast galactose transporter. J. Bact. 171, 4486-4493. VOGEL, H. ~/. JAHNIG, E. (1986). Models for the structure of outer-membrane proteins of Escherichia coli derived from R a m a n spectroscopy and prediction methods. J. molec. Biol. 190, 191-199. VOGEL, H., WRIGHT, J. K. • JAHNIO, F. (1985). The structure of the lactose permease derived from R a m a n spectroscopy and prediction methods. EMBO../. 4, 3625-3631. WILLIAMS, R. W. (1983). Estimation of protein secondary structure from the laser Raman amide I spectrum. J. molec. Biol. 166, 581-603. YAZYU, H., SHIOTA-NI1YA,S., SHIMAMOTO,T., KANAZAWA,H., FUTA1,M. & TSUCHIYA,T. (1984). Nucleotide sequence of the melB gene and characteristics of deduced amino acid sequence of the melibiose carrier in Escherichia coll. £ biol. Chem. 259, 4320-4326.