Immunoglobulin fold and tandem repeat structures in proteoglycan N-terminal domains and link protein.

Detailed primary sequence and secondary structure analyses are reported for the hyaluronate binding region (G1 domain) and link protein of proteoglycan aggregates. These are based on six full or partial sequences from the chicken, pig, human, rat and bovine proteins. Determinations of a full pig and a partial human link protein sequence are reported in the Appendix. Five sequences at the N terminus in both proteins were compared with the structures of 11 variable immunoglobulin (Ig) fold domains for which crystal structures are available. Despite only modest sequence homology, a clear alignment could be proposed. Analysis of this shows that the equivalents of the first and second hypervariable segments are now significantly longer, and both proteins have N-terminal extensions that are up to 23 residues in length. Secondary structure predictions showed that these sequences could be identified with available crystal structures for the variable Ig fold. However the hydrophobic residues involved in interactions between the light and heavy chains in Igs are replaced by hydrophilic charged groups in both proteins. These results imply that both proteins are members of the Ig superfamily, but exhibit structural differences distinct from other members of this superfamily for which crystal structures are known. The proteoglycan tandem repeat (PTR) is a repeat of 99 residues that is found twice in the amino acid sequence of link protein and the proteoglycan G1 domain adjacent to the Ig fold, and also twice in the proteoglycan G2 domain. A total of 16 PTRs was available for analysis. Compositional analyses show that these are positively charged if these originate from link protein, and negatively charged if from the G1 or G2 domains. The 16 Robson secondary structure predictions for the PTRs were averaged to improve the statistics of the prediction, and checked by comparison with Chou-Fasman calculations. A strong alpha-helix prediction was found at residues 13 to 25, and several beta-strands were predicted. The overall content is 18% alpha-helix and 28% beta-sheet, with 44% of the remaining sequence being predicted as turns. These analyses show that both the proteoglycan G1 domain and link protein are constructed from two distinct globular components, which may provide the two functional roles of these proteins in proteoglycan aggregation.

Solution structure of human and mouse immunoglobulin M by synchrotron X-ray scattering and molecular graphics modelling. A possible mechanism for complement activation.

The pentameric 71-domain structure of human and mouse immunoglobulin M (IgM) was investigated by synchrotron X-ray solution scattering and molecular graphics modelling. The radii of gyration RG of human IgM Quaife and its Fc5, IgM-S, Fab'2 and Fab fragments were determined as 12.2 nm, 6.1 nm, 6.1 nm, 4.9 nm and 2.9 nm in that order. The RG values were similar for mouse IgM P8 and its Fab'2 and Fab fragments, despite the presence of an additional carbohydrate site. The IgM scattering curves, to a nominal resolution of 5 nm, were compared with molecular graphics models based on published crystallographic alpha-carbon co-ordinates for the Fab and Fc structures of IgG. Good curve fits for Fab were obtained based on the crystal structure of Fab from IgG. A good curve fit was obtained for Fab'2, if the two Fab arms were positioned close together at their contact with the C mu 2 domains. The addition of the Fc fragment close to the C mu 2 domains of this Fab'2 model, to give a planar structure, accounted for the scattering curve of IgM-S. The Fc5 fragment was best modelled by a ring of five Fc monomers, constrained by packing considerations and disulphide bridge formation. A position for the J chain between two C mu 4 domains rather than at the centre of Fc5 was preferred. The intact IgM structure was best modelled using a planar arrangement of these Fab'2 and Fc5 models, with the side-to-side displacement of the Fab'2 arms in the plane of the IgM structure. All these models were consistent with hydrodynamic simulations of sedimentation data. The solution structure of IgM can therefore be reproduced quantitatively in terms of crystallographic structures for the fragments of IgG. Putative Clq binding sites have been identified on the C mu 3 domain. These would become accessible for interaction with Clq when the Fab'2 arms move out of the plane of the Fc5 disc in IgM, that is, a steric mechanism exposing pre-existing Clq sites. Comparison with a solution structure for Clq by neutron scattering shows that two or more of the six globular Clq heads in the hexameric head-and-stalk structure are readily able to make contacts with the putative Clq sites in the C mu 3 domains of free IgM if if the Clq arm-axis angle in solution is reduced from 40 degrees-45 degrees to 28 degrees. This could be the trigger for Cl activation.

Secondary structure changes stabilize the reactive-centre cleaved form of SERPINs. A study by 1H nuclear magnetic resonance and Fourier transform infrared spectroscopy.

Proteinase inhibitor members of the SERPIN superfamily are characterized by the presence of a proteolytically sensitive reactive-site loop. Cleavage within this region results in a conformational transition from an unstable "stressed" native protein to a more stable "relaxed" cleaved molecule. In order to identify the principal molecular aspects of this transition, 1H nuclear magnetic resonance (n.m.r.) and FT-IR spectroscopy were applied to the study of four SERPINs. 1H n.m.r. spectra of approximately 20 high-field ring-current-shifted methyl signals exhibited slightly different chemical shifts in the native and cleaved forms of alpha 1-antitrypsin (alpha 1-AT), alpha 1-antichymotrypsin (alpha 1-ACT) and C1 inhibitor (C1-INH), but not ovalbumin, between 20 degrees C and 90 degrees C. Ring current calculations based on crystal co-ordinates for cleaved alpha 1-AT and alpha 1-ACT and native ovalbumin showed that these signals originate from highly localized interactions between different buried residues corresponding to alpha-helix and beta-sheet segments of the SERPIN fold. The small shift changes correspond to small relative conformational side-chain rearrangements of about 0.01 nm to 0.05 nm in the protein hydrophobic core, i.e. the tertiary structure interactions in the two forms of the SERPIN fold are well-preserved, and changes in this appear unimportant for the stabilization found after reactive centre cleavage. Fourier transform infrared (FT-IR) spectroscopic studies of the amide I band showed that the native and cleaved forms of alpha 1-AT, alpha 1-ACT and C1-INH contain 28% to 36% alpha-helix and 38% to 44% beta-sheet. Second derivative FT-IR spectra using H2O and 2H2O buffers revealed very large differences in the amide I band between the native and cleaved forms of alpha 1-AT, alpha 1-ACT and C1-INH, but not for ovalbumin. The alpha-helix band was most sensitive to 1H-2H exchange, while the beta-sheet bands were not, and greater amounts of antiparallel beta-sheet were detected in the cleaved form. 1H n.m.r. showed that polypeptide amide 1H-2H exchange was greater in the native forms of alpha 1-AT, alpha 1-ACT and C1-INH than in their cleaved forms, whereas for ovalbumin it was unchanged. The FT-IR and 1H-2H exchange data show that alterations in the secondary structure are central to the stabilization of the cleaved SERPIN structure.(ABSTRACT TRUNCATED AT 400 WORDS)

The quaternary structure in solution of human complement subcomponent C1r2C1s2.

C1r2C1s2 is a subcomponent of first component C1 of the complement cascade. Previously two distinct models for its structure have been described, in which C1r2C1s2 is either a linear rod-like assembly of the globular domains found in each of C1s and C1r, or these domains are arranged to form an asymmetric X-shaped structure. These two models were evaluated by using hydrodynamic simulations and neutron scattering. The data on C1s, C1s2 and C1r are readily represented by straight hydrodynamic cylinders, but not C1r2 or C1r2C1s2. Tests of the X-structure for C1r2 and C1r2C1s2 successfully predicted the experimental sedimentation coefficients, thus supporting this model. Neutron scattering analyses on C1s and C1r2 are consistent with a linear structure for C1s, but not for C1r2. An X-shaped structure for C1r2 was found to give a good account of the neutron data at large scattering angles. The total length of the C1s and C1r monomers was determined as 17-20 nm, which is compatible with electron microscopy. On the basis of the known sequences of C1r and C1s, this length is accounted for by a linear arrangement of a serine-proteinase domain (length 4 nm), two short consensus repeat domains (2 x 4 nm), and a globular entity containing the I, II and III domains (4-7 nm).

Molecular modeling of human complement component C4 and its fragments by X-ray and neutron solution scattering.

The solution structures of human complement component C4 and five derived fragments, C4u, C4(a + b), C4b, C4c, and C4d, were analyzed by synchrotron X-ray and neutron scattering. The X-ray radii of gyration RG for C4, C4u, and C4(a + b) in H2O buffers are similar at 5.23-5.28 nm, and likewise the cross-sectional radii of gyration RXS are similar at 2.48-2.52 nm. Molecular mass calculations using X-rays and neutrons show unexpectedly that C4c is dimeric; however, all the other forms are monomeric. C4c2 has an X-ray RG of 5.18 nm and an RXS of 2.89 nm. Neutron contrast variation gives RG values at infinite contrast of 4.87-4.93 nm for C4 and C4u, 4.79 nm for C4b, 4.94 nm for C4c2, and 2.69 nm for C4d. The RXS values at infinite contrast are 2.23-2.25 nm for C4 and C4u, 1.89 nm for C4b, and 2.62 nm for C4c2. These data show that a large conformational change occurs on going from C4 to C4b, but not on going from C4 to C4u, and this is attributed to the presence of the C4a moiety in C4u. Comparisons of the C4 and C4u scattering curves show that these are very similar out to a nominal resolution of 4 nm. Scattering-curve models were developed to account for the neutron scattering curves of C4, C4c2, and C4d in 2H2O buffers. The C4c monomer could be represented by a lamellar ellipsoid of size 8 nm x 2 nm x 18 nm. C4d was found to be 4 nm x 2 nm x 9 nm. The combination of these structures gave good accounts of the neutron data for C4, C4b, and C4c2 to resolutions of 5-6 nm. The C4 model was obtained by placing the long axis of C4d parallel to that of C4c such that the cross section is extended. C4b was best modeled by repositioning C4d relative to C4c such that this cross section becomes more compact. The C4 and C4b models are compared with possible structures for the C1 component of complement to show the importance of the surface accessibility of the protease domains and short consensus repeat domains in C1 for C4 activation.

Oligomeric domain structure of human complement factor H by X-ray and neutron solution scattering.

Factor H is a regulatory component of the complement system. It has a monomer Mr of 150,000. Primary structure analysis shows that the polypeptide is divided into 20 homologous regions, each 60 amino acid residues long. These are independently folding domains and are termed "short consensus repeats" (SCRs) or "complement control protein" (CCP) repeats. High-flux synchrotron X-ray and neutron scattering studies were performed in order to define its solution structure in conditions close to physiological. The Mr of factor H was determined as 250,000-320,000 to show that factor H is dimeric. This structure is maintained at concentrations between 1 and 11 mg/mL in the pH range 5-9. Zn2+ ions are an inhibitor of C3b cleavage by factor I, a reaction in which factor H acts as a cofactor. Additions of Zn2+ to factor H caused it to form oligomers containing 4-10 monomers. The radius of gyration RG of native factor H by X-rays or by neutrons in 0% or 100% 2H2O buffers is not measurable but is greater than 12.5 nm. Two cross-sectional radii of gyration RXS-1 and RXS-2 were determined as 3.0-3.1 and 1.8 nm, respectively. Analyses of the cross-sectional intensities show that factor H is composed of two distinct subunits. The RXS-1 corresponds to the cross-sectional properties of both subunits and exhibits an unusual radiation dependence on the X-ray flux. Since RXS-2 is close to the corresponding RXS of C4b binding protein (91% of which is formed from SCR/CCP domains), it is inferred that the SCR/CCP domains of factor H and C4b binding protein have similar solution structures. The use of hydrodynamic spheres to reproduce literature sedimentation coefficients of 5.5-5.6 S showed that these were compatible with a V-shaped arrangement of two rods (36 spheres each, length 87 +/- 5 nm) joined at an angle of 5 degrees. The use of a similar arrangement of 244 spheres arranged in two rods (length 77 nm) to fit the experimental X-ray and neutron scattering curves showed that the two rods are joined at an angle of 5 degrees. This model corresponds to an actual RG of 21-23 nm. The separation between each SCR/CCP in factor H is close to 4 nm. In the solution structure of factor H, the SCR/CCP domains are in a highly extended conformation.

Characterization of the structure of polydisperse human low-density lipoprotein by neutron scattering.

Low-density lipoproteins (LDL) in plasma are constructed from a single molecule of apolipoprotein B-100 (M(r) 512000) in association with lipid (approximate M(r) 2-3 x 10(6)). The gross structure was studied using an updated pulsed-neutron camera LOQ with an area detector to establish the basis for the interpretation of structural changes seen during dynamic studies of LDL oxidation. Neutron-scattering data for LDL in 100% 2H2O buffers emphasize their external appearance. Guinier analysis on a continuous-flux neutron camera D17 revealed pronounced concentration-dependences in the radius of gyration, RG, and the intensity of forward scattering, I(0) (equivalent to the M(r) of LDL) between 0.5 and 11 mg of LDL protein/ml. LDL preparations from different donors gave different RG values. When extrapolated to zero concentration, RG values ranged between 8.3 and 10.6 nm and were linearly correlated with M(r), which is consistent with a spherical structure. The distance-distribution function P(r) in real space showed a single maximum at 9.1-10.9 nm, which is just under half the observed maximum dimension of 23.1 +/- 1.2 nm expected for a spherical structure. The neutron radial-density function p(r) exhibited a plateau of high and featureless density at the centre of LDL. LDL can be modelled by a polydisperse assembly of spheres with two internal densities and a mean radius close to 10.0 nm in a normal distribution of radii with a standard deviation of 2.0 nm. The data are consistent with recent electron-microscopy and ultracentrifugation data.(ABSTRACT TRUNCATED AT 250 WORDS)

Structural homologies of component C5 of human complement with components C3 and C4 by neutron scattering.

The complement component C5 is one of a family of structurally related plasma proteins that includes components C3 and C4. Activation of C5 is the initial step in the formation of the membrane attack complex of complement. Analysis of the solution structure of C5 and comparisons with similar analyses of the structures of C3 and C4 are reported here. Neutron solution scattering gave an Mr for C5 of 201,000, which demonstrates that C5 is monomeric in solution. The radius of gyration RG of C5 at infinite contrast is 4.87 nm and corresponds to an elongated structure. The longest length of C5 was determined to be at least 15-16 nm from three calculations on the basis of the RG, the scattering intensity at zero angle I(0), and the indirect transformation of the scattering curve into real space. Comparison of the RG and contrast variation data and indirect transformations of the scattering curves for C3, C4, and C5 show that these have very similar structures. Comparisons of the C5 scattering curve with Debye small-sphere models previously employed for C4 and C3 show that good curve fits could be obtained. Unlike previous studies that have suggested significant differences, these experiments indicate that, while C5 differs from C3 and C4 in its activation and inactivation pathways, significant structural homology exists between the native proteins, as might be predicted from their high (and similar) sequence homology.

Neutron and X-ray solution-scattering studies of the ternary complex between proteoglycan-binding region, link protein and hyaluronan.

Proteoglycan aggregates of cartilage are stabilized by the formation of a ternary complex between the G1 domain at the N-terminus of the proteoglycan monomer (aggrecan), link protein and hyaluronan polysaccharide. Both the G1 domain and link protein contain similar three-domain structures formed from an immunoglobulin fold and two proteoglycan tandem repeats, the arrangement of which had been investigated by neutron and synchrotron X-ray scattering [Perkins, Nealis, Dunham, Hardingham & Muir (1991) Biochemistry 30, 10708-10716]. Here, solution scattering was used to investigate the ternary complexes formed between a proteolytic fragment of proteoglycan monomer containing G1 (termed binding region), link protein and hyaluronan oligosaccharides containing either 34 or 450 saccharide units (HA34 and HA450). The ternary complex with HA34 had a neutron radius of gyration, RG, at infinite contrast not exceeding 5.5 nm. The ternary complex with HA34 had an X-ray cross-sectional radius of gyration Rxs of 2.4 nm and a neutron Rxs at infinite contrast of 2.00 nm. Since both were similar or larger than the Rxs for binding region (X-rays, 2.04 nm; neutrons, 1.84 nm) and link protein (neutrons, 0.8 nm), analyses showed that the cross-sectional mean width of the ternary complex is greater than those in each of the free proteins, i.e. the two proteins associated side-by-side. Similar results were obtained with HA450 complexed with binding region and with both binding region and link protein. This structural model was verified by hydrodynamic simulations of the experimental sedimentation coefficient of 5.5 S, which showed that a compact ternary-complex structure was formed. Although scattering curve simulations using small spheres were limited for the ternary complex with HA34 because of its approximate RG value, the scattering data were compatible with the formation of a compact complex formed by side-by-side contacts between G1 and link protein.

Structural changes in oxidised low-density lipoproteins and of the effect of the anti-atherosclerotic drug probucol observed by synchrotron X-ray and neutron solution scattering.

The atherosclerotic properties of low-density lipoproteins (LDL) are thought to be strongly enhanced by oxidation. The lipid-lowering drug probucol reduces the susceptibility of LDL to oxidation. Synchrotron X-ray and high-flux neutron solution scattering curves were used to characterise the structural properties of human LDL, before and after modification by oxidation with Cu2+ and the addition of probucol, in order to evaluate these techniques. Analyses based on Guinier plots, simple two-shell spherical modelling, and the use of cubic splines and indirect transformation show that a 20-h incubation with Cu2+ ions (but not 6 h) causes some of the LDL to associate to form larger aggregated particles. Gel electrophoresis on Cu2+ -oxidised LDL shows a concomitant degradation of the apolipoprotein B-100 as well as the formation of high molecular mass forms. These experiments indicate that the apoprotein B-100 structure has been significantly disrupted by oxidation. The addition of probucol to LDL causes an increase in the polydispersity of LDL, as evidenced by small changes in the Guinier curves and some weakening of the minima in the X-ray scattering curves. No changes in the quasispherical shape of LDL are observed and gel electrophoresis indicates no changes. It is possible that probucol may exert its effect by increasing the range of sizes of LDL and that the lipid-lowering effect of probucol in vivo might be caused by the preferential catabolism of the higher molecular mass forms of LDL thus created.

Secondary structure in properdin of the complement cascade and related proteins: a study by Fourier transform infrared spectroscopy.

Six structural repeat motifs of 58 amino acids are found in the sequence of both mouse and human properdins. Twelve more examples of the motif are available from the sequences of thrombospondin, the terminal complement components, and the thrombospondin-related anonymous protein. The averaged Robson and Chou-Fasman secondary structure predictions show that there are 57-66% turn and 19-38% beta-sheet structures in the typical repeat motif. The high amount of turn structure is consistent with Gly, Pro, Cys, and Ser being the four most abundant amino acid residues in properdin. Comparisons with sequences found in the circumsporozoite protein from several species of malaria parasites show that their sequences and secondary structures strongly coincide only in a 18-residue segment. Further secondary structure analysis utilized Fourier transform infrared spectroscopy of human properdin in 2H2O buffers. These show a broad amide I band that, after second-derivative and deconvolution calculations, is shown to be composed of several components. Two at 1633 and 1683 cm-1 are strong evidence for beta-sheet structure, although overlap from beta-turns can also contribute. The presence of beta-turn structure is indicated by absorptions at 1662-1675 and 1645 cm-1. The properdin structure contains substantial quantities of beta-sheet and beta-turn structures, which is consistent with the secondary structure predictions and amino acid compositions. The length of the repeat motif is estimated as 3.3-4.3 nm, and an estimated 14-22% of nonexchanged amide protons reside in properdin. This is suggestive of a high degree of solvent accessibility in the structure.

Molecular modeling of the multidomain structures of the proteoglycan binding region and the link protein of cartilage by neutron and synchrotron X-ray scattering.

The interaction of proteoglycan monomers with hyaluronate in cartilage is mediated by a globular binding region at the N-terminus of the proteoglycan monomer; this interaction is stabilized by link protein. Sequences show that both the binding region (27% carbohydrate) and the link protein (6% carbohydrate) contain an immunoglobulin (Ig) fold domain and two proteoglycan tandem repeat (PTR) domains. Both proteins were investigated by neutron and synchrotron X-ray solution scattering, in which nonspecific aggregate formation was reduced by the use of citraconylation to modify surface lysine residues. The neutron and X-ray radius of gyration RG of native and citraconylated binding region is 5.1 nm, and the cross-sectional RG (RXS) is 1.9-2.0 nm. No neutron contrast dependence of the RG values was observed; however, a large contrast dependence was seen for the RXS values which is attributed to the high carbohydrate content of the binding region. The neutron RG for citraconylated link protein is 2.9 nm, its RXS is 0.8 nm, and these data are also independent of the neutron contrast. The scattering curves of binding region and link protein were modeled using small spheres. Both protein structures were defined initially by the representation of one domain by a crystal structure for a variable Ig fold and a fixed volume for the two PTR domains calculated from sequence data. The final models showed that the different dimensions and neutron contrast properties of binding region compared to link protein could be attributed to an extended glycosylated C-terminal peptide with extended carbohydrate structures in the binding region.(ABSTRACT TRUNCATED AT 250 WORDS)

Time-course studies by synchrotron X-ray solution scattering of the structure of human low-density lipoprotein during Cu(2+)-induced oxidation in relation to changes in lipid composition.

Low-density lipoproteins (LDLs) in plasma are constructed from a single molecule of apolipoprotein B-100 (apoB) (M(r) 512,000) in association with lipid [approximate M(r) (2-3) x 10(6)]. LDL oxidation is an important process in the development of atherosclerosis, and can be imitated by the addition of Cu2+ ions. Synchrotron X-ray scattering of LDL yields curves without radiation damage effects at concentrations close to physiological. The radius of gyration RG for preparations of LDL from different donors ranged between 12.1 and 16.0 nm, with a mean of 13.9 nm. At 4 degrees C, the distance distribution curve P(r) indicated a maximum dimension of 25-27 nm for LDL, a peak at 19.5 nm which corresponds to a surface shell of protein and phospholipid head groups in LDL, and submaxima between 1.7 and 13.5 nm, which correspond to an ordered lipid core in LDL. LDL from different donors exhibited distinct P(r) curves. For oxidation studies of LDL by X-rays, data are best obtained at 4 degrees C at a concentration of > or = 2 mg of LDL protein/ml together with controls based on non-oxidized LDL. LDL oxidation (2 mg of apoB/ml) was studied at 37 degrees C in the presence of 6.4, 25.6 and 51.2 mu of Cu2+/g of apoB. Large changes in P(r) were reproducibly observed in the inter-particle distance range between 13 and 16 nm shortly after initiation of oxidation. This corresponds to the phospholipid hydrocarbon in LDL, which has either increased in electron density during oxidation or become increasingly disordered. After 25 h, the structural changes subsequently spread to regions of the P(r) curves assigned to surface apoB and the central core of cholesteryl esters and triacyl-glycerols. Lipid analyses were carried out under the same solution conditions. The alpha-tocopherol and beta-carotene antioxidant contents of LDL were consumed within 1-2 h. Analyses of the formation of thiobarbituric acid-reactive substances and lipid hydroperoxides indicated that arachidonic acid was preferentially oxidized before the maximal formation of lipid hydroperoxides at 8-12 h after initiation of oxidation. High-performance TLC showed that phosphatidylcholine was continuously converted into lysophosphatidylcholine during oxidation, which is consistent with the early changes in the X-ray P(r) curves. The neutral core lipids became modified only after 12-15 h of oxidation. The combination of X-ray scattering structural analyses with biochemical analyses shows that the oxidation of LDL first affects the outer shell of surface phospholipid, then it spreads towards damage of apoB and the internal neutral lipid core of LDL.