Alpha-Beta_Branden-Tooze

Alpha-Beta_Branden-Tooze - The most frequent of the domain...

Info iconThis preview shows pages 1–13. 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
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: The most frequent of the domain structures are the alpha/beta (cc/[3) domains, which consist of a central parallel or mixed [3 sheet surrounded by or helices. All the glycolytic enzymes are 11/13 structures as are many other enzymes as well as proteins that bind and transport metabolites. In all} domains, bind- ing crevices are formed by loop regions. These regions do not contribute to the structural stability of the fold but participate in binding and catalytic action. Parallel B strands are arranged in barrels or sheets There are three main classes of all} proteins. In the first class there is a core of twisted parallel [3 strands arranged close together, like the staves of a bar- rel. The (I helices that connect the parallel [3 strands are on the outside of this barrel (Figure 4.1a). This domain structure is often called the TIM barrel from the structure of the enzyme triosephosphate isomerase, where it was first observed. The second class contains an open twisted [3 sheet surrounded by or helices on both sides. A typical example is shown in Figure 4.1b, a nucleotide-binding domain sometimes calied the Rossman fold after Michael Rossman, Purdue University, who first discovered this fold in the enzyme lactate dehydrogenase in 1970. The third class is formed by amino acid sequences that contain repetitive regions of a specific pattern of leucine residues, so-called leucine-rich motifs, which form or helices and B strands. The 13 strands form a curved parallel [3 sheet with all the or helices on the out- side. The structure of one member of this class, a ribonuclease inhibitor (illus- trated in Figure 4.11), is shaped like a horseshoe, and consequently this class is called the horseshoe fold. Barrels, open sheets, and horseshoe structures are all built up from B-or—B motifs. To illustrate how they differ, let us consider two fi-a—B motifs: [ll-du- fiz and Bg-oag-Bir linked together by helix 0523. There are two fundamentally different ways these two motifs can be connected into a B sheet of four par— allel strands, as shown in Figure 4.2. Strand 33 can be aligned adjacent either to strand [32, giving the strand order 1 2 3 4, or to strand Bl, giving the strand order 4 3 1 2. In the first case the two [ii-0H3 motifs are joined with the same Orientation. Since the B-ot—B unit is almost always a right-handed structure, all three or helices (one from each motif and the joining helix) are on the same side, above the [3 sheet (Figure 4.2a). In barrel and horseshoe structures the {i- ot-B motifs are linked in this way and consist of consecutive B-o—fi units, all in the same orientation. 47 In the second case we must turn the second motif around in order to align [3 strands 1 and 3. As a result of the right-handed structure of the B—u—{i motif, its a helix is on the other side of the [3 sheet (Figure 4.2m. In open twisted B-sheet structures there are always one or more such alignments and therefore there are Ct helices on both sides of the [5 sheet. These geometric rules apply because virtually all B-d—B motifs are right-handed. As pointed out in Chapter 2, this is an empirical rule that almost always applies, although no convincing explanation has been found. ‘ Alpha/beta barrels occur in many difi‘erent enzymes In ofB structures where the strand order is 1 2 3 4, all connections are on the same side of the [3 sheet. An open twisted [3 sheet of this sort with four or more parailel 13 strands would leave one side of the parallel B sheet exposed to the solvent and the other side shielded by the a helices. Such a domain structure is rarely observed, except in the horseshoe structure or as part of more complex structures where loop regions, extra o helices, or additional B sheets cover the exposed side of the fi sheet. Instead, a closed barrel of twisted B strands is formed with all the connecting U. hetices on the outside of the barrel, as shown in Figure 4.1a. However, more than four [3 strands are needed to provide enough staves to form a closed barrel, and almost all the Closed o/B barrels observed to date have eight parallel l3 strands. These are arranged such that B strand 8 is adjacent and hydrogen-bonded to B strand 1. In a few cases the barrels do not have eight parallel l3 strands; there are aiso barrels that contain ten parallel [3 strands and some that contain eight paral- lel and two antiparallel 13 strands. In almost all cases the cross-connections between the parallel $3 strands are a helices; in addition, there is usually an a helix after the last [3 strand. The eight-stranded (xiii-barre] structure is one of the largest and most reg- ular of all domain structures. A minimum of about 200 residues are required to form this structure. it has been found in many different proteins, most of which are enzymes, with completely different amino acid sequences and 48 Figure 4.1 Aiphaf’beta domains are found in many proteins. They occur in different classes, two of which are shown here: (a) a closed barrel exemplified by schematic and topological diagrams of the enzyme triosephosphate isomerase and (b) an open twisted sheet with helices on both sides, as in the coenzyme- binding domain of some dehydrogenases. Both classes are buitt up from B-ot—B motifs that are linked such that the B strands are parallel. Rectangles represent or helices, and arrows represent [3 strands in the topological diagrams. [(a) Adapted from J. Richardson. (b) Adapted from B. Fumgren.] mam-mm.” W_ m “reams-AM- "M..- different functions. Superimposing the structures of these proteins shows that around 160 residues are structurally equivalent. These residues form the B strands and or helices. The remaining residues form the loop regions that connect the B strands with the a helices. These loops have quite different lengths and conformations in the different proteins. This reflects the fact that the [i strands and or helices form the structural framework of the enzyme, whereas the loops contain the amino acids responsible for its catalytic cherri- istry. In some cases the loops are very long and form independent domains in the overall subunit structure. Branched hydrophobic side chains dominate the core of rig/)3 barrels in barrels the hydrophobic side chains of the (r helices are packed against hydrophobic side chains of the [3 sheet. The d helices are antiparallel and adjacent to the B strands that they connect. Thus the barrel is provided with a shell of hydrophobic residues from the OE helices and the B strands. Since the side chains of consecutive amino acids of a B strand are on opposite sides of the [3 sheet, every second residue of the B strands contributes to this hydrophobic shell. The other side chains of the B strands point inside the barrel to form a hydrophobic core; this core is therefore comprised exclu- sively of side chains of B-strand residues (Figure 4.3). The packing interactions between or helices and B strands are dominated by the residues Val (V), He (1), and Len (L), which have branched hydropho- bic side chains. This is reflected in the amino acid composition: these three amino acids comprise approximately 40% of the residues of the B strands in parallel [3 sheets. The important role that these residues play in packing tr helices against [3 sheets is particularly obvious in uni-barrel structures, as shown in Table 4.1. Figure 4.3 in most (xiii-barrel structures the eight [3 strands of the barrel enclose a tightly packed hydrophobic core formed entirely by side chains from the i3 strands. The core is arranged in three layers, with each layer containing four side chains from alternate f3 strands. The schematic diagram shows this packing arrangement in the all} barrel of the enzyme glycolate oxidase, the stmcture of which was determined by Carl Branden andcolleagues in Uppsala, Sweden. Figure 4.2 A {in-[i motif is a right-handed structure. Two such motifs can be joined into a four~stranded parallel {3 sheet in two different ways. They can be aligned with the a helices either on the same side of the [3 sheet (a) or on opposite sides (b). in case (a) the last [3 strand of motif 1 (red) is adjacent to the first [3 strand of motif 2 (blue), giving the strand order 1 2 3 4. The motifs are aligned in this way in barrel structures (see Figure 4.13) and in the horseshoe fold (see Figure 4.11). in case (b) the first [i strands of both motifs are adjacent, giving the strand order 4 3 1 2. Open twisted sheets (see Figure 4.1b) contain at least one motif alignment of this kind. In both cases the motifs are joined by an or helix (geen). 49 WW Table 4.1 The amino acid residues of the eight parallel [5 strands in the barrel structure of the enzyme trlosephosphate isomerase from chicken muscle Positions Strand no. Residue no. 1 2 3 4 5 1 6—10 Phe Val Gly Gly Asn 2 37—41 Glu Val Val Cys Gly 3 59—63 Gly Val Ala Ala Gin 4 3993 Trp Vat lle Leu Gly 5 121—125 Gly Val lle Ala Cys 6 158-162 Lys Val Val Leu Ala 7 204—208 Arg Ile lie Tyr Gly 8 227—231 Gly Phe Leu Val Giy The sequences are aligned so that residues in positions 1, 3, and 5 point into the bar- rel and residues in positions 2. and 4 point toward the i1 helices on the outside and are involved in the hydrophobic interactions between the i3 strands and the a helices. Bulky hydrophobic residues from positions 1, 3, and 5 of the 13 strands fitl the interior of the barrel and form a tightly packed hydrophobic core (see Figure 4.3). Note from Table 4.1 that some of these residues are Lys, Arg, or Gin, which have a polar end group (see Panel 1.1, pp. 6—?) terminating a chain of hydrophobic —CH2 groups. These chains are in the hydrophobic interior and traverse part of the barrel; their polar end groups are on the top or bottom surface of the barrel and are in contact with the aqueous environ- ment. By this arrangement even amino acids that are classified as polar can participate in the formation of hydrophobic cores of compact globular domains through the hydrophobic parts of their side chains. There is one exception to the rule that requires bulky hydrophobic residues to fill the interior of eight—stranded d/B barrels in order to form a tightly packed hydrophobic core. The coenzyme Blg—dependent enzyme methylmalonylwcoenzyme A mutase, the x-ray structure of which was deter~ mined by Phil Evans and colleagues at the MRC Laboratory of Molecular 50 Figure 4.4 Schematic diagram of the structure of the [xiii-barrel domain of the enzyme methylmalonyl—coenzyme A mutase. Alpha helices are red, and [5 strands are blue. The inside of the barrel is lined by small hydrophilic side chains (serine and threonine) from the B strands, which creates a hole in the middle where one of the substrate molecules, coenzyme A (green), binds along the axis of the barrel from one end to the other. (Adapted from a computer~generated diagram provided by P. Evans.) Biology, in Cambridge, UK, has a hole in the middle of its (xiii-barrel domain (Figure 4.4). The serine and threonine side chains that project from the B strands into the interior of the barrel are small and polar, and therefore do not fill up the space available inside the barrel. The resulting tunnel through the barrel provides an idea] environment for the catalytic reaction and is suf— ficiently large for the substrate molecule, methylmalonyl—coenzyme A, to bind. Many enzyme-catalyzed reactions, including the reaction catalyzed by this mutase, require that the reactive part of the substrate molecule be shielded from solvent during the catalytic reaction. When the substrate is bound in the tunnel inside the barrel, it is shielded from the outside world. in all} barrels with a hydrophobic core, the substrate binds at the surface of the barrel and conformational changes of loop regions shield the substrate from the solvent. Pymvatte kinase contains several domains, one of which is on ofli bowel All known eight-stranded d/B-barrel domains have enzymatic functions that include isomerization of small sugar molecules, oxidation by flavin co- enzymes, phosphate transfer, and degradation of sugar polymers. In some of these enzymes the barrel domain comprises the whole subunit of the protein; in others the polypeptide chain is longer and forms several additional domains. An enzymatic function in these multidornain subunits, however, is always associated with the barrel domain. For example, each subunit of the dimeric glycolytic enzyme triosephos- phate isomerase (see Figure 4.1a) consists of one such barrel domain. The polypeptide chain has 2.48 residues in which the first B strand of the barrel starts at residue 6 and the last or helix of the barrel ends at residue 246. In contrast, the subunit of the glycolytic enzyme pyruvate kinase (Figure 4.5), which was solved at 2.6 A resolution in the laboratory of Hilary Muirhead, Bristol University, UK, is folded-into four different domains. The polypeptide chain of this cat muscle enzyme has 530 residues. In Figure 4.5, residues 1—42 Figure 4.5 The polypeptide chain of the enzyme pyruvate kinase folds into several domains, one of which is an cr/B barrel (red). One of the loop regions in this barre] domain is extended and comprises about 100 amino acid residues that fold into a separate domain (blue) built up from anfiparallel l3 strands. The C-terminai region of about 140 residues forms a third domain (green), which is an open twisted all} structure. 51 H cmoz-I Hc¢CN / | il H+ T Cl)— Hct‘ C H \C/ \N { \c5—wo—P=o H “II 0 l | H / \ 0— C: H H C4 I I I/| H fZ—‘EB H OH OH PRAI if .._.. co 1&1 /C—OH 2 HC¢ \lc l | m i i T i" “ |\cl=c2—c3——c4—c5-—o—pm H | | l | | f_ H OH OH OH H 0 H20? [ops i” T” T T— H—CB—Cd—CSWO—on H A J. .I. l— Hcy \c/K ll /C1 Hck /c\N c H I H form a small domain (yellow) involved in subunit contacts in the tetrameric molecule; residues 43—1 15 and 224—387 form an OMB-barrel domain (red) that binds substrate and provides the catatytic groups; residues 116—223 loop out from the end of B-strand number 3 in the barrel domain and are folded into a separate domain consisting of an antiparallel [3 sheet (blue); and finally, residues 388—530 form an open twisted all} domain (green). This structure illustrates perfectly how a long polypeptide chain can be arranged in do— mains of different structural types. Double barrels have occurred by gene firsion PRA-isoruerase:IGP-synthase, a bifunctional enzyme from E. coli that we aiyzes two reactions in the synthesis of tryptophan (Figure 4.6), has a polypeptide chain that forms two crib barrels. The structure of this enzyme, solved at 2.8 A in the laboratory of Hans Jansonius in Basel, Switzerland, showed that residues 48—254 form one barrel with IGP-synthase activity, while residues 255—450 form the second barrel with PRAvisomerase activity (Figure 4. 7). In llncillus subtilis these two reactions are catalyzed by two separate enzymes that have amino acid sequences homologous to the corresponding regions of the bifunctional enzyme from E. coli, and thus each forms a barrel 52 Figure 4.6 The bifunctional enzyme PRA— isomerase (PRAI):EGP—synthase (IGPS) catalyzes two sequential reactions in the biosynthesis of tryptophan. In the first reaction (top half)r which is catalyzed by the C-terminal PRAl domain of the enzyme, the substrate N- (5'—phosphoribosyl) anthranilate (PEA) is converted to 1-(o-carboxyphenylamino)—l- deoxyribulose 5-phosphate (CdRP) by a rearrangement reaction. The succeeding step (bottom half), a ring closure reaction from CdRP to indole-S-glycerol phosphate (1GP), is catalyzed by the N—terminal IGPS domain. Figure 4.7 Two of the enzymatic activities involved in the biosynthesis of tryptophan in E. coli, phosphoribosyl anthranilate (PRA) isomerase and indoleglycerol phosphate (1GP) synthase, are performed by two separate domains in the polypeptide chain of a bifunctional enzyme. Both these domains are ot/[i-barrel structures, oriented such that their active sites are on opposite sides of the molecule. The two catalytic reactions are therefore independent of each other. The diagram shows the IGP—synthase domain (residues 48—254) with dark colors and the PRA-isomerase domain with light colors. The or helices are sequentially labeled a—h in both barrel domains. Residue 255 (arrow) is the first residue of the second domain. (Adapted from J.P. Priestle et a]., Proc. Natl. Affld. Sci. USA 84-: 5690—5694, 1987.) structure. There is no obvious functional advantage to E. coli in having these two enzymatic activities in one polypeptide chain, since the active sites of the two barrels are on opposite sides of the molecule facing away from each other and the two reactions are thus independent of each other. A third organism, Neurospora crassa, has an enzyme with three catalytic activities within the same polypeptide chain; here two domains similar to those of the E. coli enzyme are linked to a third domain that has yet another enzymatic function in the same biosynthetic pathway. These differences between species reflect different ways to organize the genome. DNA sequences that code for protein domains with different functions are organized into separate genes in one organism and fused into a single gene in another. Although the three-dimensional structures of these enzymes in B. subtilis and N. crasso have not been solved by crystailography, we can be certain that they are or/B- f barrel domains because of their sequence homologies to the E. coli proteins. The active site is formed by loops at one end of the or/fi barrel In all these nib-barrel domains the active site is in a very similar position. It is situated in the bottom of a funnel—shaped pocket created by the eight loops that connect the carboxy end of the B strands with the amino end of the or helices (Figure 4.8). Residues that participate in binding and catalysis are in Figure 4.8 The active site in all all} barrels is in a pocket formed by the loop regions that connect the carboxy ends of the [5 strands with the adiacent or helices, as shown schematicaily in (a), where only two such loops are shown. (b) A view from the top of the barrel of the active site of the enzyme RuBisCo (ribulose bisphosphate carboxylase), which is involved in C02 fixation in plants. A substrate analog (red) binds across the barrel with the two phosphate groups, P1 and P2, on opposite sides of the pocket. A number of charged side chains (blue) from different loops as well as a Mg2+ ion (yellow) form the substrate-binding site and provide catalytic groups. The structure of this 500 kD enzyme was determined to 2.4 A resolution in the laboratory of Carl Branden, in Uppsala, Sweden. (Adapted from an originai drawing provided by Bo Furugren.) 53 these loop regions. In other words, these enzymes are modeled on a common stable scaffold of eight parallel 13 strands surrounded by eight or helices. in each case the specific enzymatic activity is determined by the eight loop regions at the carboxy end of the B strands, which do not contribute to the structural stability of the scaffold. In some cases an additional loop region from a second domain or a different subunit comes close to this active site and also participates in binding and catalysis. Alpha/beta barrels provide examples of evolution of new enzyme activities How do new enzyme activities evolve? Are new enzymes formed from ran- dom sequences generated by recombination and other genetic rearrange- ments or do they arise by divergent evolution from a preexisting set of enzymes. Greg Petsko at Brandeis University has provided strong evidence for the latter case from studies of Dc/B-barrel enzymes in a rare metabolic path way, conversion of rnanclelate to benzoate. This rare metabolic pathway is thought to be of recent evolutionary origin, since it is present in only a few pseudomonad species. The first enzyme in this pathway, mandelate racemase, catalyzes the interconversion of the two optical isomers of mandelate (Figure 4.9a). The key step in this reaction is proton abstraction from a carbon atom, produc— ing an enolic intermediate. Petsko found that the three-dimensional struc- ture of this enzyme, including its (Jr/[3 barrel, is very similar to that of a quite different enzyme, muconate lactonizing enzyme, which catalyzes a different chemical reaction (Figure 4.9b) but which also involves the formation of an intermediate by proton abstraction. The amino acid sequences of the 350 residues of these enzymes showed 26% sequence identity, which clearly demonstrates that they are evolutionarily related. By comparing these two structures in detail Petsko found significant similarities in the region of the active site that catalyzes proton abstraction and intermediate formation but substantial differences in those regions of the active site that confer substrate specificity. These results are compatible with an evolutionary history in Which the new enzyme activity of rnandelate racemase has evolved from a preexisting enzyme that catalyzes the basic chemical reaction of proton abstraction and formation of an intermediate. Subsequent mutations have modified the «11.. . substrate —" each: intermediate m“ product HC:CH HC: CH HC=CH / \ / \ / \ HC CH HC CH HC CH I_\I\C // O \\ // \\ // O —- c He — c - HC -— // \ /O C\ // c = c EH E c _ c / \ _ _ OH 0 g \ \o H OH 0 \\ /C — CH 18 E \ 0“ \CH 0* c/ \fn 0“: c/ \ch 1’ ‘\ \ HC 0‘ 0 —CH - o — CH 0 \\ I? \ x0 \ // / c —~ c\ "E— C = c EH E c W c _ / \ _ H o H o ,f \ \o‘ H H 54 Figure 4.9 Mechanisms of the reactions catalyzed by the enzymes mandelate racemase (a) and muconate lactonizing enzyme (b). The two overall reactions are quite different: a change of configuration of a carbon atom for mandelate racemase versus ring ciosure for the lactonizing enzyme. However, one cruciai step (pink) in the two reactions is the same: addi- tion of a proton (blue) to an intermediate of the substrate (pink) from a lysine residue of the enzyme (E) or, in the reverse direction, formation of an intermediate by proton abstraction from the carbon atom adiacent to the carboxylate group. .mw‘rrw-Mwmamm—sw ‘ s i l l l 1—5 strand —-— loop 4—— ot helix Figure 4.10 Consemus amino acid sequence and secondary structure of the leucine-rich 2 5 7 l2 17 20 24 - _ — ‘ _ _ . a d _ ~ _ _ _ _ _ _ - _ _ _ _ _ _ _ _ _ _ _ ‘ motifs of type A and type B. “X” denotes any “WEN NH: x L E x L x L X X C X L T x x X C x x L X x a I“ X x x X amino acid; "a" denotes an aliphatic amino “mm B} NHZ _X_L_R-E_L_x_L_X”X_N_X_L_G_D_X_G_fl_X_X_L_X_X_X_L'X_X_P_X_X, acid. 1Eonserved restdues are shown in bold in pe . substrate specificity while preserving the ability to catalyze the basic chemical reaction. Chemistry is the important factor to preserve during evolution of new enzymes, while specificity can be modified. it would therefore seem that relatively nonspecific enzymes, which may have existed earlier in evolution or which may arise occasionally through random genetic rearrangements, are the clay from which nature sculpts new enzymes. To preserve the original enzyme activity and at the same time allow divergence, the precursor gene for the enzyme must first be duplicated at some point. Further strong evidence that proteins with new functions evolve by the process of gene duplication and subsequent modification by mutation has recently been found by examination of the genome sequence of the bacter- I ium Hnemoplrilus influenzae, the first free-living organism to be completely sequenced. Sequence comparisons showed that, of the 1680 identified pro- teins coded by this genome, at least one third are related to one or more other proteins within the genome and therefore have arisen from processes that involve gene duplications. Alpha/beta barrels are particularly well suited to such an evolutionary strategy, since substrate specificity and catalytic function reside in loop regions which are separated from the residues of the a helices and B strands that contribute to the structural stability of these domain structures. Such enzymes should also be excellent targets for genetic redesign in vitro. By 5 changing the lengths and specific residues of the active—site loop regions, it ‘ might be possible to produce novel substrate specificities without affecting the stability of the structural framework and therefore the enzyme. Leucr’rze—rich motifs fomr an or/fiJIorseshoe fold r Leucine—rich motifs—tandem homologous amino acid sequences of about 20—30 residues—whave been identified from sequence studies in over 60 dif- ferent proteins, including reteptors, cell adhesion molecules, bacterial viru- lence factors, and molecules involved in RNA splicing and DNA repair. The x-ray structure of one member of this class of proteins, a ribonuclease inhibitor, has been determined by Johann Deisenhofer and colleagues at the University of Texas, Dallas. The ‘156 amino acids of the polypeptide chain are i arranged in 15 tandem leucine-rich motifs of two types that alternate along the chain: type A, with 29 residues, and type B, with 28 residues. In addition there are two short regions with nonhomologous sequences at the termini of the chain. The consensus sequence of these homologous repeats (Figure 4.10) indicates that both types of repeat contain a characteristic pattern of leucine residues that play an important structural role, as we will see. Each repeat forms a right-handed B—lDOp—{l structure similar to those found in the two other classes of trip structures described earlier. Sequential [Hoop—cc repeats are joined together in a similar way to those in the Lilli-bar- rel structures. The B strands form a parallel [3 sheet, and all the u helices are on one side of the {3 sheet. However, the [3 strands do not form a closed bar- rel; instead they form a curved open structure that resembles a horseshoe with or helices on the outside and a [3 sheet forming the inside wall of the horseshoe (Figure 4.11). One side of the [3 sheet faces the a helices and par- ticipates in a hydrophobic core between the Cut helices and the [3 sheet; the other side of the [3 sheet is exposed to solvent, a characteristic other [1/]?! structures do not have. The leucine residues in this leucine-rich motif form a hydrophobic core between the [3 sheet and the or helices. Leucine residues 2, 5, and 7 (see Figure l i ii l. l 3%. 55 4.10) from the B strand of the motif pack against leucine residues 20 and 24 of the or helix to form the main part of the hydrophobic region. Leucine residue 12 from the loop region is also part of this hydrophobic core, as is residue 17 from the or helix, which is usually hydrophobic (Figure 4.12). Leucine residues 2, 5, 7, 12., 2.0, and 24 of the motif are invariant in both type A and type B repeats of the ribonuclease inhibitor. An examination of more than 500 tandem repeats from 68 different proteins has shown that residues 20 and 24 can be other hydrophobic residues, whereas the remaim ing four leucine residues are present in all repeats. On the basis of the crystal structure of the ribonuclease inhibitor and the important structural role of these leucine residues, it has been possible to construct plausible structural models of several other proteins with leucine-rich motifs, such as the extra- cellular domains of the thyrotropin and gonadotropin receptors. Alpha/beta twisted open-sheet structures contain or helices on both sides of the [3 sheet In the next class of MB structures there are or helices on both sides of the B sheet. This has at least three important consequences. First, a closed barrel cannot be formed unless the B strands completely enclose the a helices on one side of the B sheet. Such structures have never been found and are very unlikely to occur, since a large number of B strands would be required to enclose even a single a helix. Instead, the B strands are arranged into an open twisted B sheet such as that shown in Figure 4.11). Second, there are always two adjacent B strands (B1 and B3 in Figure 4.2b) in the interior of the B sheet whose connections to the flanking B strand are on opposite sides of the B sheet. One of the loops from one of these two B strands goes above the B sheet, whereas the other loop goes below. This cre- ates a crevice outside the edge of the B sheet between these two loops (Figure 4.13). Almost all binding sites in this class of (it/B proteins are located in crevices of this type at the carboxy edge of the B sheet, as we discuss in detail Figure 4.12 Schematic diagram illustrating the role of the conserved leucine residues (green) in the leucine-rich motif in stabilizing the B-loop—or structural module. In the ribonuclease inhibitor, leucine residues 2, 5, and 7 from the B strand pack against leucine residues 17, 20, and 24 from the or helix as well as leucine residue 12 from the loop to form a hydrophobic core between the B strand and the (1 helix. 56 Figure 4.11 Schematic diagram of the structure of the ribonuclease inhibitor. The molecule, which is built up by repetitive B-loop-ot motifs, resembles a horseshoe with a 17-stranded parallel B sheet on the inside and 16 a: helices on the outside. The B sheet is light red, or helices are blue, and loops that are part of the B-loop-ot motifs are orange. (Adapted from B. Kobe et al., Nahtre 366: 751—756, 1993) loop B strand l Figure 4.13 (a) The active site in open twisted a/B domains is in a crevice outside the carboxy ends of the [3 strands. This CIEVice is formed by two adjacent loop regions that connect the two strands with or helices on opposite sides of the [3 sheet. This is illustrated by the curled fingers of two hands (b), where the top halves of the fingers represent loop regions and the bottom halves represent the [i strands. The rod represents a bound molecule in the binding crevice. crevice later. (We define the carboxy edge of the sheet as the edge that is formed by the carboxy ends of the parallel [3 strands in the sheet.) Third, in open-sheet structures the or helices are packed against both sides of the [3 sheet. Each [3 strand thus contributes hydrophobic side chains to pack against or helices in two similar hydrophobic core regions, one on each side of the [3 sheet. Open fi~sheet structures have a variety of topologies We have seen that all members of the (rift—barrel domain structures have the same basic arrangement of eight or helices and eight [3 strands. Within the open ou'j} sheets, however, there is much more variation in structure, as is obvious from purely geometric considerations. Since the B strands form an open [3 sheet, there are no geometric restrictions on the number of strands involved. In fact, the number varies from four to ten. Furthermore, the two is strands joined by a crossover connection need not be adjacent in the [3 sheet, although the [in-B motif where the two [i strands are adjacent is a pre- ferred structural building block. In addition, there can be mixed B sheets in which hairpin connections give rise to some antiparallel B strands mixed with the parallel [3 strands. All these variations occur in actual structures, some of which are illustrated in Figure 4.14am. There are thus many varia- tions on the regular arrangement of six parallel B strands (see Figure 4.11)). The positions of active sites can be predicted in (Jr/,6 stmctm‘es We have described a general relationship between structure and function for the offs-barrel structures. They all have the active site at the same position with respect to their common structure in spite of having different functions as Well as different amino acid sequences. We can now ask if similar rela- tionships also occur for the open olB-sheet structures in spite of their much greater variation in structure. Can the position of the active sites be predict- ed from the structures of many open-sheet all} proteins? In almost every one of the more than 100 different known unfit structures of this class the active site is at the carboxy edge of the [3 sheet. Functional residues are provided by the loop regions that connect the carboxy end of the [3 strands with the amino end of the a helices. In this one respect a fun- damental similarity therefore exists between the u/B—barre] structures and the open trill—sheet structures. The general shapes of the active sites are quite different, however. Open Ell/l3 structures cannot form funnel-shaped active sites like the barrel struc- tures. Instead, they form crevices at the edge of the [3 sheet. Such crevices occur when there are two adjacent connections that are on opposite sides of the [3 sheet. One of the loop regions in these two connections goes out from 57 58 crevice crevice Figure 4.14 Exam- ples of different types of open twisted all} structures. Both schematic and topo- logical diagrams are given. In the topological diagrams, arrows denote strands of [3 sheet and i rectangles denote a helices. (a) The FMN— bincling redox protein flavodoxin. (b) The enzyme adenylate kinase, which catalyzes the reaction AMP + ATP v—z 2 ADP. The structure was determined to 3.0 A 5 resolution in the laboratory of Georg Schulz in Heidelberg, Germany. (c) The ATP— binding domain of the glycolytic enzyme hexokinase, which catalyzes the phospho- rylation of glucose I The structure was l i determined to 2.8 A resolution in the l laboratory of Tom E Steitz, Yale University. 1 (d) The glycolytic E enzyme phospho- glycerate mutase, which catalyzes ' transfer of a phos— é phoryl group from E carbon 3 to carbon 2 in phosphoglycerate. The structure was determined to 2.5 A resolution in the laboratory of Herman Watson, Bristol Uni- versity, UK. (Adapted from J. Richardson.) its 5 strand above the :3 sheet and the other below, creating a crevice between them (see Figure 4.13). The active site or part of it is usually found in such a crevice. The position of such crevices is determined by the topology of the 5 sheet and can be predicted from a topology diagram. The crevices occur when the strand order is reversed, and can be easily identified in a topology dia- gram as the place where connections from the carboxy ends of two adjacent [3 strands go in opposite directions, one to the left and one to the right. Let us examine the first two diagrams given in Figure 4.14. The first structure, flavodoxin (Figure 4.14a), has one such position, between strands 1 and 3. The connection from strand 1 goes to the right and that from strand 3 to the left. In the schematic diagram in Figure 4.14:: we can see that the corresponding or helices are on opposite sides of the [3 sheet. The loops from these two l3 strands, 1 and 3, to their respective or helices form the major part of the binding cleft for the coenzyme FMN (flavin mononu- cleotide). The second structure, adenylate kinase (Figure 4.1413), has two such posi— tions, one on each side ofli strand 1. The connection from strand 1 to strand 2 goes to the right, whereas the connection from the flanking strands 3 and 4 both go to the left. Crevices are formed between B strands 1 and 3 and between strands 1 and a. One of these crevices forms part of an AMP~binding site, and the other crevice forms part of an ATP-binding site that catalyzes the formation of ADP from AMP and ATP. Such positions in a topology diagram are called topological switch points. It was postulated in 1980 by Carl Branden, in Uppsala, Sweden, that the position of active sites could be predicted from such switch points. Since then at least one part of the active site has been found in crevices defined by such switch points in almost all new a/B structures that have been deter- mined. Thus we can predict the approximate position of the active site and possible loop regions that form this site in tit/[3 proteins. This is in contrast to proteins of the other two main classes—ct-helical proteins and antiparallel [3 proteins—where no such predictive rules have been found. We will now examine a few examples that illustrate the relationship between the topolo- gy diagrams of some all} proteins, their switch points, and the active-site residues. These examples have been chosen because they represent different types of out} open-sheet structures. Tyrosyl—tRNA synthetase has two different domains (at/,3 + or) One of the crucial steps in protein synthesis is performed by the group of enzymes called aminoacyl-tRNA synthetases. These enzymes connect each amino acid with its specific transfer RNA molecule in a two-step reaction. First the amino acid is activated by ATP to give an enzyme-bound amino acid adenyiate; then this complex is attacked by the tRNA to give the aminoacyl- -.' tRNA. The structure of the synthetase specific for the amino acid tyrosine was determined to 2.7 A resolution in the laboratory of David Blow in London. Figure 4.15 shows a schematic diagram of the first 320 residues of a single subunit of this dimeric molecule. The last 100 residues are disordered in the crystal and are not visible. There are essentially two different domains, one Figure 4.15 Schematic diagram of the enzyme tyrosyl-tRNA synthetase, which couples tyrosine to its cognate transfer RNA. The central region of the catalytic domain (red and green) is an open twisted o/B structure with five parallel [i strands. The active site is formed by the loops from the carboxy ends of [3 strands 2 and 5. These two adjacent strands are connected to 0‘. helices on opposite sides of the [3 sheet. Where more than one or helix connects two [i strands (for example, N between strands 4 and 5), they are represented as one rectangle in the topology diagram. (Adapted from IN. Bhat et al., 1. Mol. Biol. 158: 699—709, 1982.) 59 ...
View Full Document

Page1 / 13

Alpha-Beta_Branden-Tooze - The most frequent of the domain...

This preview shows document pages 1 - 13. Sign up to view the full document.

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