Structural aspects of actin-binding proteins.

The three-dimensional structures of myosin subfragment 1 (S1), gelsolin segment 1 complexed with alpha-actin, villin fragment 14T, Acanthamoeba profilin-I, and bovine profilin complexed with beta-actin were completed last year. Together, they expand our understanding of the structural organization of actin-binding proteins. In addition, the segment 1 and bovine profilin complexes provide atomic-level descriptions of their interfaces with actin.

Molecular dynamics simulation of a phospholipid micelle.

The dynamic character of phospholipid aggregates limits conventional structural studies to the determination of average molecular features. In order to develop more detailed descriptions of phospholipid structure for comparison with experiment, the molecular dynamics of a hydrated lysophosphatidylethanolamine (LPE) micelle, incorporating 85 LPE and 1591 water molecules, have been simulated. Comparison of the initial and equilibrated micelles shows substantial differences both in LPE hydrocarbon chain conformation and polar head-group-solvent interactions. Although these changes produce only subtle effects on the averaged structural properties of the system, the alterations in hydrocarbon chain packing and head-group solvation appear to mimic a polymorphic pretransition from a spherical toward a cylindrical micelle structure.

The enhancement of PCR amplification by low molecular weight amides.

Amplification of a DNA target by the polymerase chain reaction (PCR) often requires laborious optimization efforts. In this regard, the use of certain organic chemicals such as dimethyl sulfoxide, polyethylene glycol, betaine and formamide as cosolvents has been found to be very helpful. Unfortunately, very little is known about the precise structural features that make these additives effective and, accordingly, the number of such chemicals currently known to enhance PCR is limited. In order to address these issues, we decided to focus on formamide and undertook an extensive study of low molecular weight amides as a class to see how changing the substituents in the amide structure influences its effect on PCR. We describe here the results of this study, which involved 11 different amides, and present observations that provide a cohesive picture of structure-activity relations in this group of additives. We found several of these amides to be exceptionally effective and introduce them as novel PCR enhancers.

The crystal structure of a major allergen from plants.

BACKGROUND: Profilins are small eukaryotic proteins involved in modulating the assembly of actin microfilaments in the cytoplasm. They are able to bind both phosphatidylinositol-4,5-bisphosphate and poly-L-proline (PLP) and thus play a critical role in signaling pathways. Plant profilins are of interest because immunological cross-reactivity between pollen and human profilin may be the cause of hay fever and broad allergies to pollens. RESULTS: The determination of the Arabidopsis thaliana profilin isoform I structure, using multiwavelength anomalous diffraction (MAD) to obtain structure-factor phases, is reported here. The structure of Arabidopsis profilin is similar to that of previously determined profilin structures. Conserved amino acid residues in profilins from plants, mammals, and lower eukaryotes are critically important in dictating the geometry of the PLP-binding site and the overall polypeptide fold. The main feature distinguishing plant profilins from other profilins is a solvent-filled pocket located in the most variable region of the fold. CONCLUSIONS: Comparison of the structures of SH3 domains with those of profilins from three distinct sources suggests that the mode of PLP binding may be similar. A comparison of three profilin structures from different families reveals only partial conservation of the actin-binding surface. The proximity of the semi-conserved actin-binding site and the binding pocket characteristic of plant profilins suggests that epitopes encompassing both features are responsible for the cross-reactivity of antibodies between human and plant profilins thought to be responsible for type I allergies.

Evidence that the methylesterase of bacterial chemotaxis may be a serine hydrolase.

CheB, the methylesterase of chemotactic bacteria, catalyzes the hydrolysis of glutamyl-methyl esters in bacterial chemoreceptor proteins. The two cysteines predicted by the amino acid sequence of CheB were replaced by alanine residues. The resulting mutants, Cys207-Ala, Cys309-Ala and a double cysteine mutant Cys207-Ala/Cys309-Ala, retained methylesterase activity, indicating that sulfhydryls are not crucial for CheB mediated catalysis. A homology search revealed a conserved serine active-site region between residues 162 and 166 which is homologous to the active-site region of acetylcholine esterases, suggesting that Ser164 of CheB is the active-site nucleophile. Oligonucleotide-directed mutagenesis was used to change the serine to a cysteine. This Ser164-Cys mutant had less than 2% of the wild-type activity. Unlike the serine proteinases which utilize a 'catalytic triad' mechanism, CheB does not have the conserved histidine and aspartic acid residues located in positions N-terminal to the active-site serine. In addition, CheB is not labeled with di-isopropylfluorophosphate, a potent inhibitor of other serine hydrolases. A novel mechanism is proposed for CheB involving substrate-assisted catalysis to account for these apparent anomalies.

The use of poly(L-proline)-Sepharose in the isolation of profilin and profilactin complexes.

In the purification of proline hydroxylase by affinity chromatography on poly(L-proline)-Sepharose it was found earlier that two other components, profilin and the complex profilin-actin, also bind with high affinity to this matrix. We have exploited this observation to develop a rapid procedure for the isolation of profilin and profilin-actin complexes in high yields directly from high-speed supernatants of crude tissue-extracts. Through an extensive search for elution conditions, avoiding poly(L-proline) as the desorbant, we have found that active proteins can be recovered from the affinity column with a buffer containing 30% dimethyl sulphoxide. Subsequent chromatography on hydroxylapatite separates free profilin and the two isoforms of profilactin, profilin-actin beta and profilin-actin gamma. The profilin-actin complexes produced this way have high specific activities in the DNAase-inhibition assay, give rise to filaments on addition of Mg2+, and can be crystallized. From the isolated profilin-actin complexes the beta- and gamma-actin isoforms of non-muscle cells can easily be prepared in a polymerization competent form. Pure profilin is either obtained from an excess pool present in some extracts or by dissociation of profilin-actin complexes and removal of the actin.

The structure of an open state of beta-actin at 2.65 A resolution.

The structure of an "open state" of crystalline profilin:beta-actin has been solved to 2.65 A by X-ray crystallography. The open-state crystals, in 1.8 M potassium phosphate, have an expanded unit cell dimension in the c direction of 185.7 A compared with 171.9 A in the previously solved ammonium sulphate-stabilized "tight-state" structure. The unit cell change between the open and the tight states is accompanied by large subdomain movements in actin. Furthermore, the nucleotide in the open state is significantly more exposed to solvent, and local conformational changes in the hydrophobic pocket surrounding cysteine 374 occur during the transition to the tight state. Significant changes were observed at the N terminus and in the DNase-I binding loop. Neither the structure of profilin nor its contact with beta-actin are affected by the changes in the unit cell. Applying osmotic pressure to profilin:beta-actin crystals brings about a collapse of the unit cell comparable with that seen in the open to tight-state transition, enabling an estimate of the work required to cause this transformation of beta-actin in the crystals. The slight difference in energy between the open and collapsed states explains the extreme sensitivity of profilin:beta-actin crystals to changes in chemical and thermal environment.

Domain motions in actin.

Previous crystallographic investigations have shown that actin can undergo large conformational changes, even when complexed to the same actin binding protein. We have conducted a formal analysis of domain motions in actin, using the four available crystal structures, to classify the mechanism as either hinge or shear and to quantify the magnitude of these changes. We demonstrate that actin consists of two rigid cores, a semi-rigid domain and three conformationally variable extended loops. Confirming predictions about the nature of the domain rotation in actin based on its structural similarity to hexokinase, we show, using an algorithm previously used only to identify protein hinges, that residues at the interface between the two rigid cores undergo a shear between alternative conformations of actin. Rotations of less than 7 degrees in the torsion angles of five residues in the polypeptides that connect the rigid cores enable one actin conformation to be transformed into another. Because these torsion angle changes are small, the interface between the domains is maintained. In addition, we show that actin secondary structure elements, including those outside the rigid cores, are conformationally invariant among the four crystal structures, even when actin is complexed to different actin binding proteins. Finally, we demonstrate that the current F-actin models are inconsistent with the principles of actin conformational change identified here.

Crystallization and preliminary X-ray analysis of Gc, the vitamin D-binding protein in serum.

The vitamin D-binding protein, Gc, was purified from human serum and crystallized using the hanging-drop method. The best crystals were grown from 28% polyethylene glycol 400 in 50 mM-sodium acetate at pH 4.8. These crystals diffract to 3.4 A and the observed diffraction is consistent with orthorhombic space groups P4(1) and P4(3). The unit cell parameters were determined to be a = b = 135.5 A and c = 75.6 A.

Molecular packing in profilin: actin crystals and its implications.

Analysis of profilin: actin crystals reveals an extensive intermolecular network, rather than a discrete "monomeric complex", comprising stacked actin ribbons held in place by columns of profilin molecules, wedged in between neighboring actin subunits and running perpendicular to the ribbons. Comparison with data from electron microscopy, X-ray diffraction, spectroscopy, and biochemistry of actin suggests that a simple transformation relates the ribbon to f-actin. The crystals exhibit unusual polymorphic properties, which strengthens the view that movements within the actin monomer are important for force generation.

Preliminary X-ray crystallographic analysis of holotoxin from Bordetella pertussis.

Pertussis (whooping cough) is a serious infectious disease caused by the bacterium Bordetella pertussis. One of the major virulence factors is a protein known as pertussis toxin, which is composed of six subunits, with a total molecular weight of 106,000. Enzymatic transfer of ADP-ribose from NAD to a family of GTP-binding proteins is effected by the largest subunit (S1 or the A monomer), while binding of host cells and entry of S1 to the interior is a function of the other subunits (the B oligomer). The holotoxin crystallizes in the orthorhombic space group P2(1)2(1)2(1), with unit cell dimensions a = 98.4 A, b = 164.2 A and c = 195.2 A. The crystals are suitable for high-resolution X-ray diffraction analysis.

X-ray structure determination of human profilin II: A comparative structural analysis of human profilins.

Human profilins are multifunctional, single-domain proteins which directly link the actin microfilament system to a variety of signalling pathways via two spatially distinct binding sites. Profilin binds to monomeric actin in a 1:1 complex, catalyzes the exchange of the actin-bound nucleotide and regulates actin filament barbed end assembly. Like SH3 domains, profilin has a surface-exposed aromatic patch which binds to proline-rich peptides. Various multidomain proteins including members of the Ena/VASP and formin families localize profilin:actin complexes through profilin:poly-L-proline interactions to particular cytoskeletal locations (e.g. focal adhesions, cleavage furrows). Humans express a basic (I) and an acidic (II) isoform of profilin which exhibit different affinities for peptides and proteins rich in proline residues. Here, we report the crystallization and X-ray structure determination of human profilin II to 2.2 A. This structure reveals an aromatic extension of the previously defined poly-L-proline binding site for profilin I. In contrast to serine 29 of profilin I, tyrosine 29 in profilin II is capable of forming an additional stacking interaction and a hydrogen bond with poly-L-proline which may account for the increased affinity of the second isoform for proline-rich peptides. Differential isoform specificity for proline-rich proteins may be attributed to the differences in charged and hydrophobic residues in and proximal to the poly-L-proline binding site. The actin-binding face remains nearly identical with the exception of five amino acid differences. These observations are important for the understanding of the functional and structural differences between these two classes of profilin isoforms.

Preliminary X-ray crystallographic analysis of a modified basic fibroblast growth factor.

Human basic fibroblast growth factor (hbFGF) has been modified, with Ala3 and Ser5 substituted by glutamic acid, and the purified recombinant protein has been crystallized. The crystals are triclinic (space group P1) with unit cell parameters a = 31.0 A, b = 33.6 A, c = 34.7 A, alpha = 88 degrees, beta = 85 degrees, gamma = 76 degrees, and they diffract to at least 2 A.

Crystal structure of the oligomerization domain of NSP4 from rotavirus reveals a core metal-binding site.

During the maturation of rotaviral particles, non-structural protein 4 (NSP4) plays a critical role in the translocation of the immature capsid into the lumen of the endoplasmic reticulum. Full-length NSP4 and a 22 amino acid peptide (NSP4(114-135)) derived from this protein have been shown to induce diarrhea in young mice in an age-dependent manner, and may therefore be the agent responsible for rotavirally-induced symptoms. We have determined the crystal structure of the oligomerization domain of NSP4 which spans residues 95 to 137 (NSP4(95-137)). NSP4(95-137) self-associates into a parallel, tetrameric coiled-coil, with the hydrophobic core interrupted by three polar layers occupying a and d-heptad positions. Side-chains from two consecutive polar layers, consisting of four Gln123 and two of the four Glu120 residues, coordinate a divalent cation. Two independent structures built from MAD-phased data indicated the presence of a strontium and calcium ion bound at this site, respectively. This metal-binding site appears to play an important role in stabilizing the homo-tetramer, which has implications for the engagement of NSP4 as an enterotoxin.

Crystallization and structure determination of bovine profilin at 2.0 A resolution.

Profilin regulates the behavior of the eukaryotic microfilament system through its interaction with non-filamentous actin. It also binds several ligands, including poly(L-proline) and the membrane phospholipid phosphatidylinositol 4,5-bisphosphate (PtdIns(4,5)P2). Bovine profilin crystals (space group C2; a = 69.15 A, b = 34.59 A, c = 52.49 A; alpha = gamma = 90 degrees, beta = 92.56 degrees) were grown from a mixture of poly(ethylene glycol) 400 and ammonium sulfate. X-ray diffraction data were collected on an imaging plate scanner at the DORIS storage ring (DESY, Hamburg), and were phased by molecular replacement, using a search model derived from the 2.55 A structure of profilin complexed to beta-actin. The refined model of bovine profilin has a crystallographic R-factor of 16.5% in the resolution range 6.0 to 2.0 A and includes 128 water molecules, several of which form hydrogen bonds to stabilize unconventional turns. The structure of free bovine profilin is similar to that of bovine profilin complexed to beta-actin, and C alpha atoms from the two structures superimpose with an r.m.s. deviation of 1.25 A. This value is reduced to 0.51 A by omitting Ala1 and the N-terminal acetyl group, which lie at a profilin-actin interface in crystals of the complex. These residues display a strained conformation in crystalline profilin-actin but may allow the formation of a hydrogen bond between the N-acetyl carbonyl group of profilin and the phenol hydroxyl group of Tyr188 in actin. Several other actin-binding residues of profilin show different side-chain rotomer conformations in the two structures. The polypeptide fold of bovine profilin is generally similar to those observed by NMR for profilin from other sources, although the N terminus of Acanthamoeba profilin isoform I lies in a distorted helix and the C-terminal helix is less tilted with respect to the strands in the central beta-pleated sheet than is observed in bovine profilin. The majority of the aromatic residues in profilin are exposed to solvent and lie in either of two hydrophobic patches, neither of which takes part in an interface with actin. One of these patches is required for binding poly(L-proline) and contains an aromatic cluster comprising the highly conserved residues Trp3, Tyr6, Trp31 and Tyr139. In forming this cluster, Trp31 adopts a sterically strained rotamer conformation.(ABSTRACT TRUNCATED AT 400 WORDS)

The enhancement of PCR amplification by low molecular-weight sulfones.

DNA amplification by polymerase chain reaction (PCR) is frequently complicated by the problems of low yield and specificity, especially when the GC content of the target sequence is high. A common approach to the optimization of such reactions is the addition of small quantities of certain organic chemicals, such as dimethylsulfoxide (DMSO), betaine, polyethylene glycol and formamide, to the reaction mixture. Even in the presence of such additives, however, the amplification of GC-rich templates is often ineffective. In this paper, we introduce a novel class of PCR-enhancing compounds, the low molecular-weight sulfones, that are effective in the optimization of high GC template amplification. We describe here the results of an extensive structure-activity investigation in which we studied the effects of a series of six different sulfones on PCR amplification. We identify two sulfones, sulfolane and methyl sulfone, that are especially potent enhancers of high GC template amplification, and show that these compounds often outperform DMSO and betaine, two of the most effective PCR enhancers currently used. We conclude with a brief discussion of the role that the sulfone functional group may play in such enhancement.

A new perspective on muscle contraction.

Recent experimental findings suggest that the myosin cross-bridge theory may no longer be adequate to account for certain basic facts concerning muscle contraction. A newly-proposed mechanism based on length changes in actin filaments might be the basis for a simpler explanation for how the free energy of ATP hydrolysis can be transduced into work by muscle fibers.

The effect on actin ATPase of phalloidin and tetramethylrhodamine phalloidin.

Actin polymerization has been studied in the absence of excess nucleotide. Using G-actin ATP monomers, it was shown that mechanical shearing stimulates ATP hydrolysis. The procedures used enabled the detection of differential effects of phalloidin and tetramethylrhodamine-phalloidin, on the P(i)-release step of the actin ATPase. It is concluded that tetramethylrhodamine, in contrast to phalloidin, accelerates P(i)-release from actin filaments.

Covalent binding of ATPgammaS to the nucleotide-binding site in S14C-actin.

We have recently reported on the characterization of beta-actin carrying the mutation S14C in one of the phosphate-binding loops. The present paper describes the attachment of the adenosine 5'-[gamma-thio]-triphosphate (ATPgammaS) to actin containing this mutation. Treatment of S14C-actin with ATPgammaS blocked further nucleotide exchange and raised the thermal stability of the protein, suggesting the formation of a covalent bond between the sulfhydryl on the terminal phosphate of ATPgammaS and cysteine-14 of the mutant actin. The affinity of the derivatized G-actin for DNase I as compared to wild-type ATP-actin was lowered to a similar extent as that of ADP.AlF(4)-actin. The derivatized actin polymerized slower than ATP-actin but faster than ADP-actin. Under these conditions the bound ATPgammaS was hydrolyzed, suggesting the formation of a state corresponding to the transient ADP.P(i)-state. ATPgammaS-actin interacted normally with profilin, whereas the interaction with actin depolymerizing factor (ADF) was disturbed, as judged on the effects of these proteins on actin polymerization.

Mutagenesis of human profilin locates its poly(L-proline)-binding site to a hydrophobic patch of aromatic amino acids.

The actin-binding protein, profilin, contains a src-homology (SH) 3-like fold (Schutt, C.E. et al., submitted), and its tight interaction with poly(L-proline) is reminiscent of the binding activity exhibited by SH3-domains. Here we demonstrate that replacements of aromatic amino acids in a hydrophobic patch on the surface of the profilin molecule abolish its poly(L-proline)-binding capacity. However, the location of this hydrophobic patch is found in another region of the molecule than that displaying structural similarities with SH3 domains.