BETA-structures_Branden-Tooze

BETA-structures_Branden-Tooze - Antiparallel beta[3...

Info iconThis preview shows pages 1–15. Sign up to view the full content.

View Full Document Right Arrow Icon
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 2
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 4
Background image of page 5

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 6
Background image of page 7

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 8
Background image of page 9

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 10
Background image of page 11

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 12
Background image of page 13

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 14
Background image of page 15
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Antiparallel beta ([3) structures comprise the second iarge group of protein domain structures. Functionally, this group is the most diverse; it includes enzymes, transport proteins, antibodies, cell surface proteins, and virus coat proteins. The cores of these domains are built up by B strands that can vary in number from four or five to over ten. The [3 strands are arranged in a predominantly antiparallel fashion and usually in such a way that they form two [3 sheets that are joined together and packed against each other. The [5 sheets have the usual twist, and when two such twisted 13 sheets are packed together, they form a barrel—Eike structure (Figure 5.1). Antiparal— lel [3 structures, therefore, in general have a core of hydrophobic side chains inside the barrel provided by residues in the {i strands. The surface is formed by residues from. the loop regions and from the strands. The aim of this chapter is to examine a number of antiparallel B structures and demonstrate how these rather complex structures can be separated into smaller comprehensible motifs. Figure 5.1 The enzyme superoxide dismutase (SOD). SOD is a 13 structure comprising eight antiparallel [3 strands (a). In addition, SOD has two metal atoms, Cu and Zn (yellow circles), that participate in the catalytic action: conversion of a superoxide radical to hydrogen peroxide and oxygen. The eight [3 strands are arranged around the surface of a barrel, which is viewed atong the barrel axis in (b) and perpendicular to this axis in (c). [(a) Adapted from J.S. Richardson. The structure of SOD was determined in the laboratory of J.S. and DR. Richardson, Dulce University] In Chapter 2 we described the 24 different ways that two B-loop-B units can form a four-stranded B sheet. The number of possible ways to form antiparallel B—sheet structures rapidly increases as the number of strands increases. It is thus surprising, but reassuring, that the number of topologies actually observed is small and that most B structures fall into a few groups of common or similar topology. The three most frequently occurring groups— up-and-down barrels, Greek keys, and jelly roll barrels—~can all be related to simple ways of connecting antiparaliel B strands arranged in a barrel structure. Up-and-down barrels have a simple topology The simplest topology is obtained if each successive B strand is added adjacent to the previous strand until the last strand is joined by hydrogen bonds to the first strand and the barrel is closed (Figure 5.2). These are called up-and-down B sheets or barrels. The arrangement of B strands is similar to that in the DUB—barrel structures we have just described in Chapter 4, except that here the strands are antiparallel and all the connections are hairpins. The structural and functional versatility of even this simple arrangement will be illustrated by two examples. The retinal-binding protein binds retinal inside an up—and—down [3 barrel The first example is the plasma-borne retinal—binding protein, RBP, which is a single polypeptide chain of 182 amino acid residues. This protein is responsible for transporting the lipid alcohol vitamin A (retinol) from its storage site in the liver to the various vitamin—A-dependent tissues. It is a disposable package in the sense that each REP molecule transports only a single retinol molecule and is then degraded. REP is synthesized in the hepatocytes, where it picks up one molecule of retinol in the endoplasmic reticulum. Both its synthesis and its secretion from the hepatocytes to the plasma are regulated by retinol. in plasma, the 68 Figure 5.2 Schematic and topological diagrams of an up-and-down [3 barrel. The eight B strands are all antiparallel to each other and are connected by hairpin loops. Beta strands that are adjacent in the amino acid sequence are also adjacent in the three5 dimensional structure of up-and-down barrels. Figure 5.3 Schematic diagram of the structure of human plasma retinal-binding protein (REP), which is an up-and-down B barrel. The eight antiparallel B strands twist and curl such that the structure can also be regarded as two B sheets (green and blue) packed against each other. Some of the twisted B strands (red) participate in both B sheets. A retinoi molecule, vitamin A (yellow), is bound inside the barrel, between the two B sheets, such that its only hydrophiiic part (an 01-1 tail) is at the surface of the molecule. The topological diagram of this structure is the same as that in Figure 5.2. (Courtesy of Alwyn Jones, Uppsala, Sweden.) A“, RETINOL RBP-retinoi complex binds to a larger protein molecule, prealbumin, which further stabilizes it and prevents its loss via the kidney. Recognition of this complex by a cell—surface receptor causes RBP to release the retinol and, as a result, to undergo a conformational change that drastically reduces its affin— ity for prealbumin. The free RBP molecule is then excreted through the kidney glomerus, reabsorbed in the proximal tubule cells, and degraded. The structure of RBP with bound retinol has been determined in the laboratory of Alwyn Jones in Uppsala, Sweden to 2.0 A resolution. Its most striking feature is a B-barrel core consisting of eight up-and-down antiparallel B strands as shown in Figure 5.3. In addition, there is a four-turn or helix at the carboxy end of the polypeptide chain that is packed against the outside of the [3 barrel. The [3 strands are curved and twisted, and the barrel is wrapped around the retinol molecule. One end of the barrel is open to the solvent whereas the other end is closed by tight side—chain packing: the tail of the retinol molecule is at the open end of the barrel. The hydrophobic retinol molecule is packed against hydrophobic side chains from the [3 strands in the barrel’s core (Figure 5.4). The structure of the apo-forrn Where the retinol molecule is removed is also known. Surprisingly, there is almost no change in the barrel structure, and a large hole is left inside the barrel. Amino acid sequence reflects B smtcrure On a large part of the surface of RE? (the front face in Figure 5.3), side chains from residues in the B strands are exposed to the solvent. This is achieved by alternating hydrophobic with polar or charged hydrophilic residues in the Figure 5.4 The binding site for retinol inside the REF barrel is lined with hydrophobic residues. They provide a hydrophobic surrounding for the hydrophobic part of the retinol molecule. 69 strand residue air-tine acid sequence no. no. 2 4 M43 3 53—60 4 7l—78 amino acid sequences of the B strands; in other words, side chains of the B strands form the hydrophobic core of the barrel as well as part of the hydrophilic outer surface. Strands 2 3 4 of RBP clearly illustrate this arrange- ment (Figure 5.5, where the core residues are colored green). This structure also very clearly illustrates that an antiparailel 13 barrel is built up from two B sheets that are packed against each other (see Figure 5.3). Beta strands 1 2 3 4 5 6 (blue and red color) form one sheet, and strands 1 8 7 6 5 (green and red color) form the second sheet. Strands 1 5 6 thus contribute to both sheets by having sharp corners where they can turn over from one sheet to the other. The retinal-binding protein belongs to a superfamily of protein structures REP is one member of a superfamily of proteins with different functions, marginally homologous amino acid sequences, but similar three-dimensional structures. This superfamily also includes an insect protein that binds a blue pigment, biliverdin, and B-lactoglobuiin, a protein that is abundant in milk. All have polypeptide chains of approximately the same lengths that are wrapped into very similar up-and-down eight—stranded antiparallel B barrels. They all tightly bind hydrophobic ligands inside this barrel. There is a second family of small lipid-binding proteins, the P2 family, which include among others cellular retinol- and fatty acid—binding proteins as well as a protein, P2, from myelin in the peripheral nervous system. How- ever, members of this second family have ten antiparallel B strands in their barrels compared with the eight strands found in the barrels of the REF super- family. Members of the P2. family show no amino acid sequence homology to members of the RBP superfamily. Nevertheless, their three-dimensional structures have similar architecture and topology, being up-and-down B barrels. Netrmminidase folds into up-and-down [3 sheets A second example of up-and—down B sheets is the protein neuraminidase from influenza virus. Here the packing of the sheets is different from that in RBP. They do not form a simple barrel but instead six small sheets, each with four B strands, which are arranged like the blades of a six-bladed propeller. Loop regions between the B strands form the active site in the middle of one side of the propeller. Other similar structures are known with different numbers of the same motif arranged like propellers with different numbers of blades such as the G-proteins discussed in Chapter 13. Influenza virus is an RNA virus with an outer lipid envelope. There are two virai proteins anchored in this membrane, neuraminidase and hemag- glutinin. They are both transmembrane proteins with a few residues inside the membrane and a transmembrane region followed by a stalk and a head- piece outside the membrane. The heads are exposed on the surface of the viri~ on and thus provide the antigenic determinants of this epidemic virus. The function of hemagglutinin, which is glycosylated, is to mediate the binding of virus particles to host cells by recognizing and binding to sialic acid residues on glycoproteins of the cell membrane, as we shall discuss in more detail at the end of this chapter. The role of the viral neuraminidase, conversely, seems to be to facilitate the release of progeny virions from infected cells by cleaving sialic acid 70 Figure 5.5 Amino acid sequence of B strands 2 3 4 in human plasma retinal-binding pro- tein. The sequences are listed in such a way that residues which point into the barrel are aligned. These hydrophobic residues are arrowed and colored green. The remaining residues are exposed to the solvent. residues from the carbohydrate side chains both of the viral hemaggiutinin and of the glycosylated cellular membrane proteins. This helps prevent prog- eny virions from binding to and reinfecting the cells from whose surface they have just budded. From the point of view of the viruses, reinfecttng an already infected cell is, of course, a waste of time. The neuraminidase molecule is a homotetramer made up of four identi- cal polypeptide chains, each of around 470 amino acids; the exact number varies depending on the strain of the virus. If influenza virus is treated with the proteolytic enzyme pronase, the head of the neuraminidase, which is soluble, is cleaved off from the stalk projecting from the viral envelope. The soluble head, comprising four subunits of about 400 amino acids each, can be crystallized. Folding motifs form a propeller—like smtchtre in neuraminidase The structure of these tetrameric neuraminidase heads was determined in the laboratory of Peter Colman in Parkville, Australia to 2.9 A resolution. Each of the four subunits of the tetrarner is folded into a single domain built up from six closely packed, similarly folded motifs. The motif is a simple up-and— down antiparallel l3 sheet of four strands (Figure 5.6). The strands have a rather large twist such that the directions of the first and the fourth strands differ by 90°. To a first very rough approximation the six motifs are arranged within each subunit with an approximate sixfold symmetry around an axis through the center of the subunit (Figure 5.7a). These six :3 sheets are arranged like six blades of a propeller. {a} Figure 5.6 Schematic and topological diagrams of the folding motif in neuraminidase from influenza virus. The motif is built up from four antiparallel ii strands joined by hairpin loops, an up-and-down open3 sheet. Figure 5.7 The subunit structure of the neuraminidase headpiece (residues 34—469) from influenza virus is built up from six similar, consecutive motifs of four up-and— down antiparallel B strands (Figure 5.6). Each such motif has been called a propeller blade and the whole subunit structure a six-blade propeller. The motifs are connected by loop regions from [3 strand 4 in one motif to [3 strand 1 in the next motif. The schematic diagram (21) is viewed down an approximate sixfold axis that relates the centers of the motifs. Four such six-blade propeller subunits are present in each complete neuraminidase molecule (see Figure 5.8). In the topological diagram (b) the yellow loop that connects the N—terminal B strand to the first Bstrand of motif 1 is not to scale. In the folded structure it is about the same length as the other loops that connect the motifs. (Adapted from J. Varghese et 211., Nature 303: 35—40, 1983.) 71 In summary, the whole moiecule has almost 1600 amino acid residues. It is composed of four identical polypeptide chains, each of which is folded into a superbarrel with 24 B strands (Figure 5.8). These 24 B strands are arranged in six similar motifs, each of which contains four [3 strands that form the blades of a propeller—like structure. The active site is in the middle of one side of the propeller Not only are the topologies within the six [3 sheets in each subunit identical, but so are their connections to each other, with the exception of the last B sheet (see Figure 5.7b). The fourth strand of each h sheet is connected across the top of the subunit (seen in Figure 5.7a coming out of the page) to the first strand of the next sheet. The loop that connects strands 2 and 3 within the sheet is also at the top of the subunit. Furthermore, because of the approximate sixfold symmetry of the fi- sheet motifs, these 12-1oop regions, derived from the six [3 sheets, are on the same side of. the molecule, as can be seen in Figure 5.9a, where we see a single polypeptide chain (one of the four subunits) from the side of the propeller. The [3 sheets are arranged cyclically around an axis through the center of the molecule, The loop regions at the top of this barrel are exten- sive (Figure 5.9a) and together they form a wide funnei-shaped pocket containing the active site (Figure 5.91)). This is analogous to the active site formed by the loop regions at the top of the (xiii-barrel structures. Greek key motifs occur frequently in mitipamllei fl structures We saw in Chapter 2. that the Greek key motif provides a simple way to connect antiparallel Li strands that are on opposite sides of a barrel structure. We will now look at how this motif is incorporated into some of the simple antiparaliel B-barrel structures and show that an antiparaliel [3 sheet of eight strands can be built up only by hairpin and/or Greek key motifs, if the connections do not cross between the two ends of the [3 sheet. 72 Figure 5.8 Schematic view down the fourfold axis of the tetrameric molecule of neurarninidase as it appeared on the cover of Nature, May 5, 1983. r-er-Iwrgm." n Assume that we have eight antiparallel [i strands arranged in a barrel structure. We decide that we want to connect strand number n to an antipar— allel strand at the same end of the barrel. We do not want to connect it to strand number n + 1 as in the up-and-down barrels just described, nor do we want to connect it to strand number n w 1 which is equivalent to turning the up-and-down barrel in Figure 5.2 upside down. What alternatives remain? It is easy to see from Figure 5.10 that there are only two alternatives. We can connect it either to strand number n + 3 or to n — 3. Both cases require only short loop regions that traverse the end of the barrel. How do we now continue the connections? The simplest way to connect the strands that were skipped over is to join them by up—and-down connections, as illustrated in Figure 5.10- n+2 (a) - n+3 [5) 12—3 [9) Active site Figure 5.9 The six four-stranded motifs in a single subunit of neuraminidase form the six blades of a propeller-like structure. A schematic diagram of the subunit structure shows the propeller viewed from its side (a). An idealized propeller structure vieWed from the side to highlight the position of the active site is shown in (b). The loop regions that connect the motifs {red in b) in combination with the loops that connect strands 2 and 3 within the motifs (green in b) form a wide funnel-shaped active site pocket. [(51) Adapted from P. Colman et ai., Nahrre 326: 358—363, 1987.] Figure 5.10 Idealized diagrams of the Greek key motif. This motif is formed when one of the connections of four antiparallel {5 strands is not a hairpin connection. The motif occurs when strand number n is connected to strand n + 3 (a) or u - 3 (b) instead of n + 1 or n ~ 1 in an eight»stranded antiparallel [3 sheet or barrel. The two different possible connections give two different hands of the Greek key motif. In all protein structures known so far only the hand shown in (a) has been observed. 73 We have now connected four adjacent strands of the barrel in a simple and logical fashion requiring only short loop regions. The result is the Greek key motif described in Chapter 2, which is found in the large majority of antiparaliei [3 structures. The two cases represent the two possible different hands, but in all structures known to us the hand that corresponds to the case where {3 strand :1 is linked to [3 strand :1 + 3 as in Figure 5.10a is present. The remaining four strands of the barrel can be joined either by Lip-and- down connections before and after the motif or by another Greek key motif. We will examine examples of both cases. The Hrystallin molecule has two domains The transparency and refractive power of the lenses of our eyes depend on a smooth gradient of refractive index for visible light. This is achieved partly by a regular packing arrangement of the cells in the lens and partly by a smoothly changing concentration gradient of lens-specific proteins, the crystallins. There are at least three different classes of crystallins. The or and [3 are heterogeneous assemblies of different subunits specified by different genes, whereas the gamma (7) crystallins are monomeric proteins with a polypeptide chain of around 170 amino acid residues. The structure of one such y crystallin was determined in the laboratory of Tom Biundell in London to 1.9 A resolution. A picture of this molecule generated from a graphics display is shown in Figure 5.11. Let us now examine this molecule and dissect it into its structural components to see if we can understand how these are put together. We will reduce this rather complex, and at first sight bewildering, structure to its simplest representation as a series of motifs. This will help us to understand the structure and see its relationships to other structures. We can immediately discern from Figure 5.31 that the molecule is divided into two clearly separated domains that seem to be of similar size. For the next step we would need a stereopicture of the model or, much better, a graphics display where we could manipulate the model and look at it from different viewpoints. Here instead we have made a schematic diagram of one domain (Figure 5.12), which is normally not done until the analysis is completed and the structural principles are clear. The domain structure has a simple topology We will now follow the main polypeptide chain and trace out a topological diagram for this domain. We can immediately see from Figure 5.13 that the only secondary structure in the molecule is made up of [3 strands, which are arranged in an antiparallel fashion into two separate [3 sheets. Beta strands 1, 2., 4, and 7 form one antiparallel [5 sheet with the strand order 2 l 4 7. We thus draw the left four arrows in Figure 5.13 and connect strands 1 and 2. Similarly, we see that [3 strands 3, 5, 6, and 8 form another antiparallel B sheet with the strand order 6 5 8 3. We notice that strands 7 and 6 are adjacent although not hydrogen bonded to each other on the back side of the domain. We thus position strand 6 adjacent to strand 7 in the topology diagram but make a space between them to indicate that they belong to different [3 sheets. Alternatively, we could have positioned strands 2. and 3 adjacent to each other, which would have given a topologically identical diagram. We then connect the strands in consecutive order along the polypeptide chain. Two Greek key motifs form the domain The topological diagram of Figure 5.13 has been drawn to reflect the observation that the two [5 sheets are separate: i3 strands 2 and 3 are not hydrogen bonded to each other, nor are strands 6 and 7. The connections 74 Figure 5.11 A computerwgena'ated diagram of the structure of yerystallin comprising one polypeptide chain of 170 amino acid residues. The diagram illustrates that the polypeptide chain is arranged in two domains (blue and red). Only main chain (N, (1', Cu) atoms and no side chains are shown. Figure 5.12 Schematic diagram of the path of the polypeptide chain in one domain (the blue region in Figure 5.11) of the y—crystaliin molecule. The domain structure is built up from two {3 sheets of four antiparallel [3 strands, sheet 1 from £3 strands 1, 2, 4, and 7 and sheet 2 from strands 3, 5, 6, and 8. Figure 5.13 A preliminary topological diagram of the structure of one domain of 'y crystallin shown in Figure 5.12, illustrating that the two [3 sheets are separate within the domain. :EMWMAVHW‘GNAWWWUb-fifihfl‘r .«v « wummmmmnmnmwnw ; t > t 'i <. look unnecessarily complicated, but notice from the schematic diagram of the domain in Figure 5.12 that the two B sheets are packed against each other so that they form a distorted barrel. To see if the diagram can be simplified, we idealize the barrel and plot the strands along the surface of the barrel as shown in Figure 5.14. It is then immediately obvious that strands 1, 2, 3, and 4 form a Greek key motif, as do strands 5, 6, 7, and 8. These two motifs are joined by a loop across the bottom of the barrel, between strands 4 and 5. On the basis of this new insight we can draw the topology diagram shown in the left half of Figure 5.15b. What is the difference between this and the previous topological diagram we made? The only changes we have made are to move [3 strand 3 from the right edge to the left edge of the domain topology and to close the gap between strands 7 and 6. We have changed neither the strand order nor the connections between the strands; thus the two diagrams are topologically identical. The two domains have identical topology Using a graphics display, we could do the same thing for the second domain and arrive at the full topology diagram in Figure 5.1513. From this diagram it is obvious that the two domains have identical topology and thus in all probability similar structures. This realization is not at all trivial. To be able Figure 5.14 The eight B strands in one domain of the crystallin structure in this idealized diagram are drawn along the surface of a barrel. From this diagram it is obvious that the [3 strands are arranged in two Greek key motifs, one (red) formed by strands 1% and the other (green) by strands 5—8. Notice that the [i strands that form one motif contribute to both 13 sheets as shown in Figure 5.12. Figure 5.15 Schematic diagram (a) and topol- ogy diagram (13) for the y-crystallin molecule. The two domains of the complete molecule have the same topology; each is composed of two Greek key motifs that are joined by a short loop region. [(a) Adapted from T. Blundell et al., Nature 289: 771—777, 1931.] 75 to see it when one looks at the structure on the display requires considerable experience because the two domains are in different orientations in the molecule. The brain therefore has to store the image of one domain while examining different orientations of the second domain. A topology diagram, on the other hand, immediately reveals similarities in domain structures. This illustrates one very important use of topology diagrams—namely, to reduce a complicated pattern to a simpler one, from which conclusions can be drawn that are also valid for the complicated pattern. The two domains have sir-nilar structures A relevant question to ask at this stage is, do the topological identities displayed in the diagram reflect structural similarity? We can now see that topologically the polypeptide chain is divided into four consecutive Greek key motifs arranged in two domains. How similar are the domain structures to each other, and how similar are the two motifs within each domain? Tom Blundell has answered these questions by superposing the CrI atoms of the two motifs within a domain with each other and by superposing the Cu atoms of the two domains with each other. As a rule of thumb, when two structures superpose with a mean deviation of less than 2 it they are consid~ ered structurally equivalent. For each pair of motifs Blundell found that 40 Cu atoms superpose with a mean distance of 1.4 A. These 40 C,1L atoms within each motif are therefore structurally equivalent. Since each motif comprises only 43 or 44 amino acid residues in total, these comparisons show that the structures of the complete motifs are very similar. Not only are the individual motifs similar in structure, but they are also pairwise arranged into the two domains in a similar way since superposition of the two domains showed that about 80 C.I atoms of each domain were structurally equivalent. This structural similarity is also reflected in the amino acid sequences of the domains, which show 40% identity. They are thus clearly homologous to each other. The motif structures within the domains superpose equally well but their sequence homology is less, being around 30% between motifs 1 and 2 and 20% between 3 and 4. This study, however, clearly shows that the topo— logical description in terms of four Greek key motifs is also valid at the structural and amino acid sequence levels. The Greek key motifs in yaysmllin are evolutionarin related These comparisons strongly suggest that the four Greek key motifs are evolutionarily related. We can guess from the amino acid sequence compari- son that this protein evolved in two stages, beginning with the duplication of a primordial gene coding for one motif of about 40 amino acid residues, followed by fusion of the duplicated genes to give a single gene encoding one domain. The gene for this domain, we may imagine, later duplicated in turn and fused to give the full gene for the present—day y—crystallin polypeptide. The evidence that this was the second step lies in the fact that the amino acid sequence homology is greater between the domains than between the motifs within each domain. There is some circumstantial evidence in the organization of the crystallin gene for the evolutionary history that we have reconstructed. The amino acid sequence of a mouse £3 crystallin is homologous to that of y crystallin and shows the same four homologous motifs. Its coding sequence is in separate DNA sequences (exons) interrupted by noncoding DNA sequences (introns). Walter Gilbert at Harvard University suggested in 1978 that genes for large proteins might have evolved by the accidental juxtaposition of exons coding for specific functions. In [3 crystallin the three introns are positioned at the junctions between the four motifs, supporting Gilbert‘s ideas. These introns couid, therefore, be evolutionary remnants of the gene duplication and fusion events. 76 M:~(‘W_nv:v.v-.\:v m, iwwwmemquwmm :.w7“1.w:¢rvv: m . mmmwrwwwmw :(v-r Frame. ‘:::_-_r 7.x: .gnw «.m- .;-.- 1 I'rm Hwilxutw'fin‘hldnwa i l l l i The Greek key motifs can form jelly roll barrels In antiparallel barrel structures with the Greek key motif one of the connections in the motif is made across one end of the barrel. Such connec- tions can be made several times in a 13 barrel giving different variations and combinations of the Greek key motif. In the structure of crystallin there are two consee‘utive Greek key motifs that form a barrel with two such COI‘lIlGO tions. There is a different but frequently occurring motif, the jelly roll motif, in which there are four connections of this type. It is called jelly roll, or Swiss roll, because the polypeptide chain is wrapped around a barrel core like a jelly roll. This motif has been found in a variety of different structures including the coat proteins of most of the spherical viruses examined thus far by x-ray crystallography, the plant lectin concanavalin A and the hemagglutinin protein from the influenza virus. The jelly roll motif is wrapped around a barrel To illustrate how this rather complicated structure is built up, We will start by wrapping a piece of string around a barrel as shown in Figure 5.16. The string goes up and down the barrel four times, crosses over once at the bottom and twice at the top of the barrel. This configuration is the basic pattern for the jelly roll motif. Let us do the same thing with a strip of paper the width of which is approximately one-eighth of the circumference of the top of the barrel. We imagine that a polypeptide chain follows the edges of this strip, starting at the bottom right corner of the strip and ending at the bottom left corner (Figure 5.17a). The polypeptide chain has eight straight sections, [3 strands, interrupted by loop regions. The B strands are arranged in a long antiparallel hairpin such that strand 1 is hydrogen bonded to strand 8, strand 2 to strand 7, and so on. We now wrap the strip around the barrel following the path of the string in Figure 5.16 and in such a way that the [3 strands go along the sides of the barrel and the loop regions form the connections at the top and bottom of the barrel (Figure 5.17.13). Figure 5.16 A diagram of a piece of string wrapped around a barrel to illustrate the basic pattern of a ielly roll motif. Figure 5.17 A simple illustration of the way eight [3 strands are arranged in a jelly roll motif. (a) The eight 53 strands are drawn as arrows along two edges of a strip of paper. The strands are arranged such that strand 1 is opposite strand 8, etc. The [3 strands are separated by loop regions. (b) The strip of paper in (a) is wrapped around a barrel in the same way as the string in Figure 5.16, such that the [l strands follow the surface of the barrel and the loop regions (gray) provide the connections at both ends of the barrel. The B strands are now arranged in a jelly roll motif. 77 The hydrogen-bonded antiparallel [3 strand pairs 1:8, 2:7, 3:6, and 4:5 are now arranged such that 13 strand 1 is adjacent to strand 2, 7 is adjacent to 4, 5 to 6, and 3 to 8. These can also form hydrogen bonds to each other. All adjacent jl strands are antiparallel. This is the basic jelly roll [l-barrel structure for eight I} strands (Figure 5.18a}. Most such barrels have eight strands, but any even number of strands greater than four can form a jelly roll barrel. In eight—stranded barrels there are two connections across the top of the barrel and two across the bottom. In addition, there are two connections between adjacent j3 strands at the top and one at the bottom. A topological diagram of this fold is given in Figure 5.18b. The jelly roll barrel is thus conceptually simple, but it can be quite puzzling if it is not considered in this way. Discussion of these structures will be exemplified in this chapter by hemagglutinin and in Chapter 16 by viral coat proteins. The jelly roll barrel is usually divided into two sheets The barrels we have used to illustrate both the Greek key and the jelly roll structures provide topological descriptions, as defined in Chapter 2. A topo- logical description accurately represents the connectivity and the strand order around the barrel and thus is very useful in the same way that a subway map tells you how stations are interconnected. However, when one analyzes the pattern of hydrogen bonds between the B strands of such barrels, one finds that they usually form two sheets with few if any hydrogen bonds between strands that belong to the different [5 sheets, as we saw in the crystallin structure. The barrel is distorted and adjacent [3 strands are separated from each other in two places across the barrel. The division of [3 strands into these two sheets does not necessarily follow the division into topological motifs. The {3 strands in jelly roll barrels are also usually arranged in two sheets that are packed against each other. This does not, however, change either the topology or the usefulness of the description of these structures as barrels as long as one keeps in mind that these barrels are distorted and flattened. Figure 5.19 Schematic picture of a single subunit of influenza virus hemagglutinin. The two polypeptide chains HA1 and HA; are held together by disulfide bridges. Figure 5.18 Topological diagrams of the jelly roll structure. The same color scheme is used as in Figure 5.17. 13.5 rim outside msrde HA2 uuufiwwmwamfltrmammquWm-mmwymmu‘”w”mww'wm Figure 5.28 Schematic diagrams of the two-sheet [in helix. Three com- plete coils of the helix are shown in (a). The two parallel [3 sheets are colored green and red, the loop regions that connect the B strands are yellow. (b) Each structural unit is composed of 18 residues forming a B-loop-B-loop structure. Each loop region contains six residues of sequence Gly—Gly—X-Giy-X-Asp where X is any residue. Calcium ions are bound to both loop regions. (Adapted from F. Jurnak et al., Curr. Opiu. Struct. Biol. 4: 802—306, 1994.) polypeptide chain, folded into a stable conformation. On cleavage to form the HA] and HA2 chains of the mature molecules the free ends snap 20 it apart to give the metastable high pH structure. This molecule is like a set trap: lowering the endosomal pH springs the trap, setting in motion a series of events which starts with a large conformational change to bring the fusion peptide into a position where it can engage the target membrane. Parallel li—helix domains have a novel fold In the first edition of this book this chapter was entitled “Antiparallel Beta Structures" but we have had to change this because an entirely unexpected structure, the [3 helix, was discovered in 1993. The B helix, which is not related to the numerous antiparaliel [El structures discussed so far, was first seen in the bacterial enzyme pectate lyase, the structure of which was deter- mined by the group of Frances Jurnak at the University of California, River- side. Subsequently several other protein structures have been found to contain B helices, including extracellular bacterial proteinases and the bacteriophage P22 tailspike protein. In these [El—helix structures the polypeptide chain is coiled into a wide helix, formed by B strands separated by loop regions. In the simplest form, the two-sheet [3 helix, each turn of the helix comprises two [3 strands and two loop regions (Figure 5.28). This structural unit is repeated three times in extracellular bacterial proteinases to form a right-handed coiled structure which comprises two adjacent three-stranded parallel [3 sheets with a hydrophobic core in between. This structural organization has striking similarities to that of all} proteins, the difference being that the loop-u helix-loop that connects the parallel [i strands in (:43 structures is substituted by a loop-[3 strand-loop in these li-helix the CUB structures described in the previous chapter, B—helix structures have two parallel [3 sheets. In wfi structures a twist of about 20° between adjacent [l strands is imposed by the packing requirements of the u helices: in order to pack ridges into grooves, as described in Chapter 3, the a helices have to be twisted with respect to each other and this forces the [3 strands also to be twisted. In fl—heiix structures no such constraint is present and therefore the sheets are almost planar and form straight walls (Figure 5.28:1). The basic structural unit of these two-sheet B helix structures contains 18 amino acids, three in each {3 strand and six in each loop. A specific amino acid sequence pattern identifies this unit; namely a double repeat of a nine— residue consensus sequence Gly-Gly-X-Gly-X-Asp—X-U-X where X is any amino acid and U is large, hydrophobic and frequently leucine. The first six residues form the loop and the last three form a [3 strand with the side chain of U involved in the hydrophobic packing of the two 5 sheets. The loops are stabilized by calcium ions which bind to the Asp residue (Figure 5.28). This sequence pattern can be used to search for possible two-sheet [3 structures in databases of amino acid sequences of proteins of unknowri structure. A more complex B helix is present in pectate lyase and the bacteriophage P22 tailspike protein. In these [i helices each turn of the helix contains three short [3 strands, each with three to five residues, connected by loop regions. The B helix therefore comprises three parallel [3 sheets roughly arranged as the three sides of a prism. However, the cross-section of the $3 helix is not quite triangular because of the arrangement of the 13 sheets. Two of the sheets are 84 arranged adjacent to each other as in the two-sheet 5 helix, and the third |3 sheet is almost perpendicular to the other two (Figure 5.29). The [3 strands in these three parallel sheets are connected by three loop regions. One loop (loop a in Figure 5.29) is short and always formed by only two residues which have invariant conformations. The other two loops are much longer and vary in size and conformation. These long loops protrude from the (3 sheets and probably form the active site regions on the external surface of the protein. Since the long loop regions vary in size, the number and type of amino acids in each turn of the Iii-helix structures vary. Consequently, no specific amino acid sequence pattern has been identified for the three-sheet B-helix structures. The number of helical turns in these structures is larger than those found so far in two-sheet [3 helices. The pectate lyase [3 helix consists of seven complete turns and is 34 A long and 17—27 fl in diameter (Figure 5.30) while the iii-helix part of the bacteriophage P22 tailspike protein has 13 complete turns. Both these proteins have other structural elements in addition to the B-helix moiety. The complete tailspike protein contains three intertwined, identical subunits each with the threensheet l3 helix and is about 200 A long and 60 A Wide. Six of these trimers are attached to each phage at the base of the icosahedral capsid. The interior of the B-helix structure in pectate lyase is completely filled with side chains leaving no room for a channel. Interestingly, these side chains are not limited to hydrophobic groups but include polar and charged groups which are all neutralized either by hydrogen bonding or by electro- static interactions. All the side chains in the interior of the helix are stacked such that the side chains of adjacent turns of the helix have a linear arrange- ment parallel to the helix axis. These internal stacks fall into different classes,- polar stacks of Asn or Ser side chains, aliphatic stacks of Ala, Val, Leu or Ile side chains and aromatic stacks of Phe or Tyr side chains. Pectate lyase is an unusually stable protein and this stacking arrangement no doubt contributes to its stability. Conclusion Antiparallel [3 structures comprise the second maior class of protein conformations. In these the antiparallel B strands are usually arranged in two [3 sheets that pack against each other and form a distorted barrel structure, the core of the molecule. Depending on the way the B strands around the barrel are connected along the polypeptide chain—in other words, depending on the topology of the barrels—~they are divided into three main groups: up- and-down barrels, Greek key barrels, and jelly roll barrels. The number of possible ways to form antiparallel (3 structures is very large. The number of topologies actually observed is small, and most [3 structures fall into these three major groups of barrel structures. The last two groups—the Greek key and gelly roll barrels—include proteins of quite diverse function, where functional variability is achieved by differences in the loop regions that connect the B strands that build up the common core region. Up—and-down barrels are the simplest structures. Each [3 strand is connected to the next strand by a short loop region. Eight [3 strands arranged Figure 5.29 Schematic diagrams of the three-sheet [3 helix. (:1) The three sheets of parallel l3 strands are colored green, blue and yellow. Seven complete coils are shown in this diagram but the number of coils varies in different structures. Two of the $3 sheen (blue and yellow) are parallel to each other and are perpendicular to the third (green). 0)) Each structural unit is composed of three [3 strands connected by three loop regions (labeled a, b and c). Loop a (red) is invariably composed of only two residues, whereas the other two loop regions vary in length. (Adapted from F. Jurnak et al., Curr. Opin. Elmer. Biol. 4: 802—806, 1994.) (a) 85 in this way form the core of a family of proteins that includes the plasma retinol-binding protein in mammals, biliverdin-binding proteins in insects, and B-lactoglobulin from milk. Members of this family, as well as of the relat- ed P2 family with 10 [3 strands, bind large hydrophobic ligands inside the barrel. The barrel seems to be particularly suited to act as a container for chemically quite diverse ligands. Diversity in ligand binding is achieved by differences in the size of the barrel and in the amino acids that also participate in building up the common core. Most of the known antiparallel [3 structures, including the immunoglob— ulins and a number of different enzymes, have barrels that comprise at least one Greek key motif. An example is y crystallin, which has two consecutive Greek key motifs in each of two barrel domains. These four motifs are homologous in terms of both their three-dimensional structure and amino acid sequence and are thus evolutionarily related. The jelly roll barrels are found in a variety of protein molecules, including viral coat proteins and hemagglutinin from influenza virus. This structure looks complicated but, in principle, is very simple if one thinks of the analogy of wrapping a strip of paper around a barrel, like a jelly roll. The hemagglutinin receptor-binding domain forms such a jelly roil barrel of eight I3 strands, where the receptor binding site is at one end of the barrel. During the membrane fusion process of influenza virus infection the hemaggluh’nin molecule undergoes a major structural change in which the fusion peptide moves about 100 A to a position close to the receptor binding site. The second protein in the membrane of influenza virus, neuraminidase, does not belong to any of these three groups of barrel structures. Instead, it forms a propellerwlike structure of 24 B strands, arranged in six similar motifs that form the six blades of the propeller. Each motif is a [3 sheet of 4 up-and- down—connected B strands. The enzyme active site is formed by loop regions on one side of the propeller. In addition to the antiparallel [Li—structures, there is a novel fold called the £3 helix. In the fi-helix structures the polypeptide chain is folded into a wide helix with two or three [i strands for each turn. The B strands align to form either two or three parallel t3 sheets with a core between the sheets completely filled with side chains. 86 Figure 5.30 Schematic diagrams of the structure of the enzyme pectate lyase C, which has a three-sheet parallel B-heiix topology. (a) Idealized diagram highlighting the helical nature of the path of the polypeptide chain which comprises eight helical turns. Dotted regions indicate positions where large external loops have been removed for clarity. (b) Ribbon diagram of the polypeptide chain. The predominant secondary structural elements are three parallel B sheets which are colored green, blue and yellow. Each [5 sheet is composed of 7-10 parallel B strands with an average length of four to five residues in each strand. The short loop regions of two residues length are shown in red. (Adapted from MD. Yoder et al., Science 260: 1503—1507, 1993.) l l r ...
View Full Document

Page1 / 15

BETA-structures_Branden-Tooze - Antiparallel beta[3...

This preview shows document pages 1 - 15. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online