Search

F0F1 ATP synthase harnesses a transmembrane electrochemical gradient for the production of ATP. When operated in reverse, this multiprotein complex catalyzes ATP hydrolysis. In bacteria, the ε subunit is involved in regulating this ATPase activity. Also, ε is essential for coupling ATP hydrolysis (or synthesis) to proton translocation. The ε subunit consists of a β sandwich and two C-terminal helices, α1 and α2. The protein can switch from a compact fold to an alternate conformation where α1 and α2 are separated, resulting in an extended structure. ε from the thermophile Bacillus PS3 (Tε) binds ATP with high affinity such that this protein may function as an intracellular ATP level sensor. ATP binding to isolated Tε triggers a major conformational transition. Earlier data were interpreted in terms of an ATP + Tεextended → ATP·Tεcompact transition that may mimic aspects of the regulatory switching within F0F1 (Yagi et al. (2007) Proc. Natl. Acad. Sci. U.S.A., 104, 11233–11238). In this work, we employ complementary biophysical techniques for examining the ATP-induced conformational switching of isolated Tε. CD spectroscopy confirmed the occurrence of a large-scale conformational transition upon ATP binding, consistent with the formation of stable helical structure. Hydrogen/deuterium exchange (HDX) mass spectrometry revealed that this transition is accompanied by a pronounced stabilization in the vicinity of the ATP-binding pocket. Surprisingly, dramatic stabilization is also seen in the β8−β9 region, which is remote from the site of ATP interaction. Analytical ultracentrifugation uncovered a previously unrecognized feature of Tε: a high propensity to undergo dimerization in the presence of ATP. Comparison with existing crystallography data strongly suggests that the unexpected β8−β9 HDX protection is due to newly formed protein–protein contacts. Hence, ATP binding to isolated Tε proceeds according to 2ATP + 2Tεextended → (ATP·Tεcompact)2. Implications of this dimerization propensity for the possible role of Tε as an antibiotic target are discussed.

Dynactin, a large multiprotein complex, binds with the cytoplasmic dynein-1 motor and various adaptor proteins to allow recruitment and transportation of cellular cargoes toward the minus end of microtubules. The structure of the dynactin complex is built around an actin-like minifilament with a defined length, which has been visualized in a high-resolution structure of the dynactin filament determined by cryo-electron microscopy (cryo-EM). To understand the energetic basis of dynactin filament assembly, we used molecular dynamics simulation to probe the intersubunit interactions among the actin-like proteins, various capping proteins, and four extended regions of the dynactin shoulder. Our simulations revealed stronger intersubunit interactions at the barbed and pointed ends of the filament and involving the extended regions (compared with the interactions within the filament), which may energetically drive filament termination by the capping proteins and recruitment of the actin-like proteins by the extended regions, two key features of the dynactin filament assembly process. Next, we modeled the unknown binding configuration among dynactin, dynein tails, and a number of coiled-coil adaptor proteins (including several Bicaudal-D and related proteins and three HOOK proteins), and predicted a key set of charged residues involved in their electrostatic interactions. Our modeling is consistent with previous findings of conserved regions, functional sites, and disease mutations in the adaptor proteins and will provide a structural framework for future functional and mutational studies of these adaptor proteins. In sum, this study yielded rich structural and energetic information about dynactin and associated adaptor proteins that cannot be directly obtained from the cryo-EM structures with limited resolutions.

A large number of cellular processes are mediated by protein-protein interactions, often specified by particular protein binding modules. PDZ domains make up an important class of protein-protein interaction modules that typically bind to the C-terminus of target proteins. These domains act as a scaffold where signaling molecules are linked to a multiprotein complex. Human glutaminase interacting protein (GIP), also known as tax interacting protein 1, is unique among PDZ domain-containing proteins because it is composed almost exclusively of a single PDZ domain rather than one of many domains as part of a larger protein. GIP plays pivotal roles in cellular signaling, protein scaffolding, and cancer pathways via its interaction with the C-terminus of a growing list of partner proteins. We have identified novel internal motifs that are recognized by GIP through combinatorial phage library screening. Leu and Asp residues in the consensus sequence were identified to be critical for binding to GIP through site-directed mutagenesis studies. Structure-based models of GIP bound to two different surrogate peptides determined from nuclear magnetic resonance constraints revealed that the binding pocket is flexible enough to accommodate either the smaller carboxylate (COO(-)) group of a C-terminal recognition motif or the bulkier aspartate side chain (CH(2)COO(-)) of an internal motif. The noncanonical ILGF loop in GIP moves in for the C-terminal motif but moves out for the internal recognition motifs, allowing binding to different partner proteins. One of the peptides colocalizes with GIP within human glioma cells, indicating that GIP might be a potential target for anticancer therapeutics.

Ubiquitin (Ub) modifications are transduced by receptor proteins that use Ub-binding domains (UBDs) to recognize distinct interaction faces on the Ub surface. We report the nuclear magnetic resonance (NMR) solution structures of the A20-like zinc finger (A20 Znf) UBD of the Ub receptor ZNF216, and its complex with Ub, and show that the binding surface on Ub centered on Asp58 leaves the canonical hydrophobic Ile44 patch free to participate in additional interactions. We have modeled ternary complexes of the different families of UBDs and show that while many are expected to bind competitively to the same Ile44 surface or show steric incompatibility, other combinations (in particular, those involving the A20 Znf domain) are consistent with a single Ub moiety simultaneously participating in multiple interactions with different UBDs. We subsequently demonstrate by NMR that the A20 Znf domain of ZNF216 and the UBA domain of the p62 protein (an Ile44-binding UBD), which function in the same biological pathways, are able to form such a Ub-mediated ternary complex through independent interactions with a single Ub. This work supports an emerging concept of Ub acting as a scaffold to mediate multiprotein complex assembly.

The bulk of the red blood cell membrane proteins are partitioned between two multiprotein complexes, one associated with ankyrin R and the other with protein 4.1R. Here we examine the effect of phosphorylation of 4.1R on its interactions with its partners in the membrane. We show that activation of protein kinase C in the intact cell leads to phosphorylation of 4.1R at two sites, serine 312 and serine 331. This renders the 4.1R-associated transmembrane proteins GPC, Duffy, XK, and Kell readily extractable by nonionic detergent with no effect on the retention of band 3 and Rh, both of which also interact with 4.1R. In solution, phosphorlyation at either serine suppresses the capacity of 4.1R to bind to the cytoplasmic domains of GPC, Duffy, and XK. Phosphorylation also exerts an effect on the stability in situ of the ternary spectrin-actin-4.1R complex, which characterizes the junctions of the membrane skeletal network, as measured by the enhanced competitive entry of a β-spectrin peptide possessing both actin- and 4.1R-binding sites. Thus, phosphorylation weakens the affinity of 4.1R for β-spectrin. The two 4.1R phosphorylation sites lie in a domain flanked in the sequence by the spectrin- and actin-binding domain and a domain containing the binding sites for transmembrane proteins. It thus appears that phosphorylation of a regulatory domain in 4.1R results in structural changes transmitted to the functional interaction centers of the protein. We consider possible implications of our findings for the altered membrane function of normal reticulocytes and sickle red cells.

S100A13 and acidic fibroblast growth factor (FGF1) are involved in a wide array of important biological processes, such as angiogenesis, cell differentiation, neurogenesis, and tumor growth. Generally, the biological function of FGF1 is to recognize a specific tyrosine kinase on the cell surface and initiate the cell signal transduction cascade. Amlexanox (2-amino-7-isopropyl-5-oxo-5H-[1]benzopyrano[2,3-b]pyridine-3-carboxylic acid) is an antiallergic drug that binds S100A13 and FGF1 and inhibits the heat shock induced release of S100A13 and FGF1. In the present study, we investigated the interaction of amlexanox with S100A13 using various biophysical techniques, including isothermal titration calorimetry, fluorescence spectrophotometry, and multidimensional NMR spectroscopy. We report the three-dimensional solution structure of the S100A13-amlexanox complex. These data show that amlexanox binds specifically to the FGF1-S100A13 interface and prevents the formation of the FGF1-releasing complex. In addition, we demonstrate that amlexanox acts as an antagonist of S100A13 by binding to its FGF1 binding site and subsequently inhibiting the nonclassical pathway of these proteins. This inhibition likely results in the ability of amlexanox to antagonize the angiogenic and mitogenic activity of FGF1.

In humans, nine aminoacyl-tRNA synthetases form a stable multiprotein complex with the three auxiliary proteins p18, p38, and p43. The N-terminal moiety of p43 is involved in its anchoring to the complex, and its C-terminal moiety has a potent tRNA binding capacity. The p43 component of the complex is also the precursor of p43(ARF), an apoptosis-released factor, and of p43(EMAPII), the endothelial-monocyte activating polypeptide II. Here we identified a new translation product of the gene of p43, which contains nine additional N-terminal amino acid residues. This gene product is targeted to the mitochondria and accounts for 2% of p43 expressed in human cells. The cytoplasmic and mitochondrial species of p43 are produced from the same mRNA by a mechanism of leaky scanning of the AUG codon at position -27, which is in an unfavorable sequence context for translation initiation. The finding that a mitochondrial species of p43 exists in human cells further exemplifies the multifaceted implications of p43 and opens new perspectives for the understanding of the role of p43 in the apoptotic cell.

Members of the genus Trypanosoma, which include the pathogenic species Trypanosoma brucei and Trypanosoma cruzi, edit their post-transcriptional mitochondrial RNA via a multiprotein complex called the editosome. In T. brucei, the RNA is nicked prior to uridylate insertion and deletion. Following editing, nicked RNA is religated by one of two RNA-editing ligases (TbREL). This study describes a recent 70 ns molecular dynamics simulation of TbREL1, an ATP-dependent RNA-editing ligase of the nucleotidyltransferase superfamily that is required for the survival of T. brucei insect and bloodstream forms. In this work, a model of TbREL1 in complex with its full double-stranded RNA (dsRNA) substrate is created on the basis of the homologous relation between TbREL1 and T4 Rnl2. The simulation captures TbREL1 dynamics in the state immediately preceding RNA ligation, providing insights into the functional dynamics and catalytic mechanism of the kinetoplastid ligation reaction. Important features of RNA binding and specificity are revealed for kinetoplastid ligases and the broader nucleotidyltransferase superfamily.

Polycomb group (PcG) proteins are required for maintaining cell identity and stem cell self-renewal. RING1B and Polycomb (Pc) are two components of a multiprotein complex called polycomb repression complex 1 (PRC1) that is essential for establishing and maintaining long-term repressed gene states. Here we characterize the interaction between the C-terminal region of RING1B (C-RING1B) and the Pc cbox domain. The C-RING1B-cbox interaction displays a 1:1 stoichiometry with dissociation constants ranging from 9.2 to 180 nM for the different Pc orthologues. NMR analysis of C-RING1B alone reveals line broadening. However, when it is in complex with the cbox domain, there is a striking change to the NMR spectrum indicative of conformational tightening. This conformational change may arise from the organization of the C-RING1B subdomains. The C-terminal regions of all PcG RING1 proteins are composed of two stretches of conserved sequences separated by a variable linker sequence. While the entire C-RING1B region is required for cbox binding, the N- and C-terminal halves of C-RING1B can be separated and are able to interact, suggesting the presence of an intramolecular interaction within C-RING1B. The flexibility within the C-RING1B structure allowing transitions between the intramolecular bound and unbound states may cause the broadened peaks of the C-RING1B NMR spectrum. Binding the cbox domain stabilizes C-RING1B, whereby broadening is eliminated. The presence of flexible regions could allow C-RING1B to bind a variety of different factors, ultimately recruiting RING1B and its associated PcG proteins to different genomic loci.

Hemojuvelin is a recently-identified iron-regulatory protein that plays an important role in affecting the expression of hepcidin, a key iron regulatory hormone. Although the underlying mechanism of this process is not clear, several hemojuvelin-binding proteins, including the cell surface receptor neogenin and bone morphogenetic protein (BMP) cytokines, have been identified. The ectodomain of neogenin is composed of four immunoglobulin-like (Ig) domains followed by six fibronectin type III-like (FNIII) domains. Here we report expression of soluble versions of hemojuvelin and neogenin for biochemical characterization of their interaction and the interaction of HJV with BMP-2. Hemojuvelin normally undergoes an autocatalytic cleavage, and as in vivo, recombinant hemojuvelin exists as a mixture of cleaved and uncleaved forms. Neogenin binds to cleaved and non-cleaved hemojuvelin, as verified by its binding to an uncleaved mutant hemojuvelin. We localized the hemojuvelin binding site on neogenin to the membrane-proximal fifth and sixth FNIII domains and the juxtamembrane linker, and showed that a fragment containing only this region binds 2-3 orders of magnitude more tightly than the entire neogenin ectodomain. Binding to the most membrane-proximal region of neogenin may play a role in regulating the levels of soluble and membrane-bound forms of hemojuvelin, which in turn would influence the amount of free BMP-2 available for binding to its receptors and triggering transcription of the hepcidin gene. Our finding that BMP-2 and neogenin bind simultaneously to hemojuvelin raises the possibility that neogenin is part of a multi-protein complex at the hepatocyte membrane involving BMP, its receptors, and hemojuvelin.

F 1 F o -ATP synthase is a large multiprotein complex, including at least 10 subunits in the membrane-bound F o -sector. One of these F o proteins is subunit e (Su e), involved in the stable dimerization of F 1 F o -ATP synthase, and required for the establishment of normal cristae membrane architecture. As a step toward enabling structure-function studies of the F o -sector, the Su e transmembrane region was structurally characterized in micelles. Based on a series of NMR and CD (circular dichroism) studies, a structural model of the Su e/micelle complex was constructed, indicating Su e is largely helical, and emerges from the micelle with Arg20 near the phosphate head groups. Su e only adopts this folded conformation in the context of the micelle, and is essentially disordered in DMSO, water or trifluoroethanol/ water. Within the micelle the C-terminal Ala10-Arg20 stretch is helical, while the region N-terminal may be transiently helical, based on negative CSI (chemical shift index) values. The Ala10-Arg20 helix contains the G 14 XXXG 18 motif, which has been proposed to play an important role in dimer formation with another protein from the F o -sector. The Gly on the C-terminal end of this motif (Glyl8) is slightly more mobile than the more buried Gly 14, based on NMR order parameter measurements (Gly 14 S 2 = 0.950; Glyl8 S 2 = 0.895). Only one Su e transmembrane peptide is bound per micelle, and micelles are 22-23 A in diameter, composed of 51 ±4 dodecylphosphocholine detergent molecules. Although there is no evidence for Su e homodimerization via the transmembrane domain, potentially synergistic roles for N-terminal (membrane) and C-terminal (soluble) domain interactions may still occur. Furthermore, the presence of a buried charged residue (Arg7) suggests there may be interactions with other F o -sector protein(s) that stabilize this charge, and possibly drive the folding of the N-terminal 9 residues of the transmembrane domain.

Human acidic fibroblast growth factor (hFGF-1) is a potent mitogen and is involved in the regulation of key cellular process such as angiogenesis, differentiation, and morphogenesis. hFGF-1 is a signal peptide-less protein that is released into the extracellular compartment as a multiprotein complex consisting of S100A13, synaptotagmin (Syt1), and a hFGF-1 homodimer. Cu(2+) is known to play an important role in the formation of the multiprotein release complex. The source of Cu(2+) required for the formation of the multiprotein release complex is not clear. In this study, we show that the cytoplasmic C2A domain of synaptotagmin binds to Cu(2+) ions with high affinity. Results from the isothermal calorimetry (ITC), near-UV circular dichroism (CD), and absorption spectroscopy experiments suggest that four Cu(2+) ions bind per molecule of C2A domain. Far-UV CD and limited trypsin digestion analysis reveal that the C2A domain undergoes a mild conformational change upon binding to Cu(2+). Competition experiments monitored by ITC and fluorescence resonance energy transfer indicate that Cu(2+) and Ca(2+) ions share common binding sites on the C2A domain. Cu(2+) ions compete with and replace Ca(2+) ions bound to the C2A domain. Two-dimensional nuclear magnetic resonance spectroscopy data clearly show that Cu(2+) ions bind to the Ca(2+) binding sites in the loops (loops 1-3) located at the apex of the structure of the C2A domain. In addition, there is a unique Cu(2+) binding site located in the loop connecting beta-strands 7 and 8. It appears that the C2A domain provides the Cu(2+) ions required for the formation of the multiprotein FGF release complex.

Protein S, the cofactor of activated protein C (APC), also expresses anticoagulant activity independent of APC by directly inhibiting prothrombin activation via interactions with factor Xa, factor Va, and phospholipids. In different studies, however, large variations in APC-independent anticoagulant activities have been reported for protein S. The investigation presented here shows that within purified protein S preparations different forms of protein S are present, of which a hitherto unrecognized form (5% of total protein S) binds with high affinity to phospholipid bilayers (K(d)1 nM). The remaining protein S (95%) has a low affinity (K(d) = 250 nM) for phospholipids. Using their different affinities for phospholipids, separation of the forms of protein S was achieved. Native polyacrylamide gel electrophoresis demonstrated that the form of protein S that binds to phospholipids with low affinity migrated as a single band, whereas the high-affinity protein S exhibited several bands that migrated with reduced mobility. Size-exclusion chromatography revealed that the slower-migrating bands represented multimeric forms of protein S. Multimeric protein S (5% of total protein S) appeared to have a 100-fold higher APC-independent anticoagulant activity than the abundant form of protein S. Comparison of purified protein S preparations that exhibited a 4-fold difference in APC-independent anticoagulant activity showed that the ability to inhibit prothrombin activation correlated with the content of multimeric protein S. Multimeric protein S could not be identified in normal human plasma, and it is therefore unlikely that this form of protein S contributes to the APC-independent anticoagulant activity of protein S that is observed in plasma.

SecB is a cytosolic tetrameric chaperone in Escherichia coli, which maintains polypeptides, destined for export in a translocation competent state. The thermodynamics of unfolding of SecB was studied as a function of protein concentration, by using high sensitivity-differential scanning calorimetry and spectroscopic methods. The thermal unfolding of tetrameric SecB is reversible and can be well described as a two-state transition in which the folded tetramer is converted directly to unfolded monomers. Increasing the pH decreases the stability of the tetramer significantly, the $T_m$ changing from 341.3 K at pH 6.5 to 332.6 K at pH 9.5. The value of $\Delta$Cp obtained from measurements of $\Delta$Hm as a function of $T_m$ was 10.7 $\pm$ 0.7 kcal $mol^{-1} K^{-1}$. The value of $\Delta$Cp is among the highest measured for a multimeric protein. At 298 K, pH 7.4, the $\Delta\hspace{2mm}G^0\hspace{2mm}_u$ for the SecB tetramer is 27.9 $\pm$ 2 kcal $mol^{-1}$. Denaturant-mediated unfolding of SecB was found to be irreversible. The reactivity of the four solvent-exposed free thiols in tetrameric SecB is salt dependent. The kinetics of reactivity suggests that these four cysteines are in close proximity to each other and that these residues on each monomer are in chemically identical environments. The thermodynamic data suggest that SecB is a stable, well-folded, and tightly packed tetramer and that substrate binding occurs at a surface site rather than at an interior cavity.

The multisubunit Smc5-Smc6 holocomplex (Smc5/6) plays a critical role in chromosome stability maintenance, DNA replication, homologous recombination, and double-stranded DNA damage repair. Smc5 and Smc6 form the core of the holocomplex, along with six non-SMC elements, for which most functions are not yet understood. Mms21 (Nse2), the relatively well-studied subunit in Smc5/6, contains a SP-like-RING finger motif on the C-terminus and was identified as a SUMO E3 ligase. Deletion of Mms21 is lethal; however, while deficient in DNA damage repair, SUMO ligase mutants remain viable. These functions of Mms21 in Smc5/6 are hard to address without understanding the interaction between Smc5 and Mms21. Previously, we systematically examined the architecture of Saccharomyces cerevisiae Smc5/6 and, using yeast two-hybrid methods, found that Mms21 interacts with the coiled-coil of Smc5. Later, crystallographic studies revealed the molecular arrangement of Mms21 with Smc5/6. For this study, we use a combination of limited proteolysis, mass spectrometry, and N-terminal sequencing to precisely define the interaction region of Smc5 with Mms21. In addition, using isothermal titration calorimetry, we find that Mms21 interacts with Smc5 in a 1:1 ratio with a K(d) of 0.68 μM. This combination of methods would be useful in examining the structure of any large multiprotein complex.

Frataxin, a conserved nuclear-encoded mitochondrial protein, plays a direct role in iron-sulfur cluster biosynthesis within the ISC assembly pathway. Humans with frataxin deficiency have Friedreich's ataxia, a neurodegenerative disorder characterized by mitochondrial iron overload and disruption in Fe-S cluster synthesis. Biochemical and genetic studies have shown frataxin interacts with the iron-sulfur cluster assembly scaffold protein (in yeast, there are two, Isu1 and Isu2), indicating frataxin plays a direct role in cluster assembly, possibly by serving as an iron chaperone in the assembly pathway. Here we provide molecular details of how yeast frataxin (Yfh1) interacts with Isu1 as a structural module to improve our understanding of the multiprotein complex assembly that completes Fe-S cluster assembly; this complex also includes the cysteine desulfurase (Nfs1 in yeast) and the accessory protein (Isd11), together in the mitochondria. Thermodynamic binding parameters for protein partner and iron binding were measured for the yeast orthologs using isothermal titration calorimetry. Nuclear magnetic resonance spectroscopy was used to provide the molecular details to understand how Yfh1 interacts with Isu1. X-ray absorption studies were used to electronically and structurally characterize how iron is transferred to Isu1 and then incorporated into an Fe-S cluster. These results were combined with previously published data to generate a structural model for how the Fe-S cluster protein assembly complex can come together to accomplish Fe-S cluster assembly.

The spindle pole body (SPB) is a multiprotein complex that organizes microtubules in yeast. Due to its large size and association with the nuclear membrane, little is known about its detailed structure. In particular, although many SPB components and some of the interactions between them have been identified, the molecular details of how most of these interactions occur are not known. The prevalence of predicted coiled-coil regions in SPB proteins suggests that some interactions may occur via coiled coils. Here this hypothesis is supported by biochemical characterization of isolated coiled-coil peptides derived from SPB proteins. Formation of four strongly self-associating coiled-coil complexes from Spc29, Spc42, and Spc72 was demonstrated by circular dichroism (CD) spectroscopy and a fluorescence resonance energy transfer (FRET) assay. Many weaker self- and heteroassociations were also detected by CD, FRET, and/or cross-linking. The thermal stabilities of nine candidate homooligomers were assessed; six unfolded cooperatively with melting temperatures ranging from11 to50 degrees C. Solution studies established that coiled-coil peptides derived from Spc42 and Spc72 form parallel dimers, and this was confirmed for Spc42 by a high-resolution crystal structure. These data contribute to a growing body of knowledge that will ultimately provide a detailed model of the SPB structure.

TolR is a part of the Pal/Tol system which forms a five-member, membrane-spanning, multiprotein complex that is conserved in Gram-negative bacteria. The Pal/Tol system helps to maintain the integrity of the outer membrane and has been proposed to be involved in several other cellular processes including cell division. Obtaining the structure of TolR is of interest not only to help explain the many proposed functions of the Pal/Tol system but also to gain an understanding of the TolR homologues ExbD and MotB and to provide more targets for antibacterial treatments. In addition, the structure may provide insights into how colicins and bacteriophages are able to enter the cell. Here we report the solution structure of the homodimeric periplasmic domain of TolR from Haemophilus influenzae, determined with conventional, NOE-based NMR spectroscopy, supplemented by extensive residual dipolar coupling measurements. A novel method for assembling the dimer from small-angle X-ray scattering data confirms the NMR-derived structure. To facilitate NMR spectral analysis, a TolR construct containing residues 59-130 of the 139-residue protein was created. The periplasmic domain of TolR forms a C 2-symmetric dimer consisting of a strongly curved eight-stranded beta-sheet, generating a large deep groove on one side, while four helices cover the other face of the sheet. The structure of the TolR dimer together with data from the literature suggests how the periplasmic domain of TolR is most likely oriented relative to the cytoplasmic membrane and how it may interact with other components of the Pal/Tol system, particularly TolQ.

Major histocompatability complex class II (MHCII) molecules are an essential component of the mammalian adaptive immune response. The expression of MHCII genes is regulated by a cell-specific multiprotein complex, termed the MHCII enhanceosome. The heterotrimeric RFX complex is the key DNA-binding component of the MHCII enhanceosome. The RFX complex is comprised of three proteins, RFXB, RFXAP, and RFX5, all of which are required for DNA binding and activation of MHCII gene expression. Static light scattering and chemical cross-linking of the three RFX proteins show that RFXB and RFXAP are monomers and that RFX5 dimerizes through two separate domains. One of these domains, the oligomerization domain, promotes formation of a dimer of dimers of RFX5. In addition, we show that the RFX complex forms a 2:1:1 complex of RFX5.RFXAP.RFXB, which can associate with a further dimer of RFX5 to form a 4:1:1 complex through the oligomerization domain of RFX5. On the basis of these studies, we propose DNA-binding models for the interaction between the RFX complex and the MHCII promoter including a DNA looping model. We also provide direct evidence that the RFX5(L66A) point mutation prevents dimerization of the RFX complexes and propose a model for how this results in a loss of MHCII gene expression.

The human cellular factor (HCF) is a multidomain protein that is implicated in processes of cell cycle progression, and it is recruited into a multicomponent assembly that triggers the expression of the herpes simplex virus genome. The amino-terminal domain of HCF has been proposed to form a "kelch" type beta-propeller fold, and the carboxy-terminal domain contains a repeat of a fibronectin-like motif. We describe the expression, purification, and characterization of the domains from the human HCF and of the full-length HCF from Caenorhabditis elegans. The purified recombinant C. elegans HCF can substitute for the human HCF in efficiently forming a multiprotein complex on a herpes simplex virus promoter element. As noted in earlier studies, a segment of human HCF encompassing the human kelch domain forms a stable complex on a viral promoter element. The purified fibronectin domain can also be recruited into this complex, but not into the stable complex formed with the minimal kelch domain. These results suggest that the fibronectin domain can interact with HCF in the transcriptional activating complex and that the association requires a region outside the putative beta-propeller.

Shiga-like toxin 1 (SLTx), produced by enterohemorrhagic strains of Escherichia coli (EHEC), belongs to a family of structurally and functionally related AB(5) protein toxins that are associated with human disease. EHEC infection often gives rise to hemolytic colitis, while toxin-induced kidney damage is one of the major causes of hemolytic uremic syndrome (HUS) and acute renal failure in children. As such, an understanding and analysis of the noncovalent interactions that maintain the quaternary structure of this toxin are fundamentally important since such interactions have significant biochemical and medical implications. This paper reports on the analysis of the noncovalent homopentameric complex of Shiga-like toxin B chain (SLTx-B(5)) using electrospray ionization (ESI) triple-quadrupole (QqQ) mass spectrometry (MS) and tandem mass spectrometry (MS/MS) and the analysis of the noncovalent hexameric holotoxin (SLTx-AB(5)) using ESI time-of-flight (TOF) MS. The triple-quadrupole analysis revealed highly charged monomer ions dissociate from the multiprotein complex to form dimer, trimer, and tetramer product ions, which were also seen to further dissociate. The ESI-TOFMS analysis of SLTx-AB(5) revealed the complex remained intact and was observed in the gas phase over a range of pHs. Theses findings demonstrate that the gas-phase structure observed for both the holotoxin and the isoloated B chains correlates well with the structures reported to exist in the solution phase for these proteins. Such analysis provides a rapid screening technique for assessing the noncovalent structure of this family of proteins and other structurally related toxins.

TFIIH is a multiprotein complex that plays a central role in both transcription and DNA repair. The subunit p62 is a structural component of the TFIIH core that is known to interact with VP16, p53, Eralpha, and E2F1 in the context of activated transcription, as well as with the endonuclease XPG in DNA repair. We used limited proteolysis experiments coupled to mass spectrometry to define structural domains within the conserved N-terminal part of the molecule. The first domain identified resulted from spontaneous proteolysis and corresponds to residues 1-108. The second domain encompasses residues 186-240, and biophysical characterization by fluorescence studies and NMR analysis indicated that it is at least partially folded and thus may correspond to a structural entity. This module contains a region of high sequence conservation with an invariant FWxxPhiPhi motif (Phi representing either tyrosine or phenylalanine), which was also found in other protein families and could play a key role as a protein-protein recognition module within TFIIH. The approach used in this study is general and can be straightforwardly applied to other multidomain proteins and/or multiprotein assemblies.

Formylglycinamide ribonucleotide amidotransferase (FGAR-AT) catalyzes the conversion of formylglycinamide ribonucleotide (FGAR), ATP, and glutamine to formylglycinamidine ribonucleotide (FGAM), ADP, P(i), and glutamate in the fourth step of the purine biosynthetic pathway. PurL exists in two forms: large PurL (lgPurL) is a single chain, multidomain enzyme of about 1300 amino acids, whereas small PurL (smPurL) contains about 800 amino acids but requires two additional gene products, PurS and PurQ, for activity. smPurL contains the ATP and FGAR binding sites, PurQ is a glutaminase, and the function of PurS is just now becoming understood. We determined the structure of Bacillus subtilis PurS in two different crystal forms P2(1) and C2 at 2.5 and 2.0 A resolution, respectively. PurS forms a tight dimer with a central six-stranded beta-sheet flanked by four helices. In both the P2(1) and the C2 crystal forms, the quaternary structure of PurS is a tetramer. The concave faces of the PurS dimers interact via the C-terminal region to form a twelve-stranded beta-barrel with a hydrophilic core. We used the structure of PurS together with the structure of lgPurL from Salmonella typhimurium to construct a model of the PurS/smPurL/PurQ complex. The HisH (glutaminase) domain of imidazole glycerol phosphate synthetase was used as an additional model of PurQ. The model shows stoichiometry of 2PurS/smPurL/PurQ using a PurS dimer or 4PurS/2smPurL/2PurQ using a PurS tetramer. Both models place key conserved residues at the ATP/FGAR binding site and at a structural ADP binding site. The homology model is consistent with biochemical studies on the reconstituted complex.

The mechanism of the electron transfer from the soluble protein plastocyanin to the multiprotein complex of photosystem I from spinach has been studied in detail. The two kinetic components of P700+ reduction by plastocyanin after a laser flash, showing a constant half-life of 11 microseconds and a variable half-life of the second-order reaction, respectively, are used to monitor the electron transfer from bound and soluble plastocyanin. The effect of increasing concentration of reduced plastocyanin on both of these kinetic components and the competition by oxidized plastocyanin is used to estimate the individual dissociation constants of the complex between the proteins in each of its oxidized and reduced state. The dissociation constant of oxidized plastocyanin is about six times larger than that of 7 microM found for reduced plastocyanin and purified PSI. Consistent with this result the midpoint redox potential of plastocyanin bound to photosystem I either in equilibrium with soluble plastocyanin or after cross-linking to photosystem I is found to be 50-60 mV higher than that of soluble plastocyanin. It is concluded that the driving force of the intracomplex electron transfer is decreased in favor of an optimized turnover of photosystem I. Double-flash excitation shows that oxidized plastocyanin has to leave the complex after the electron transfer before a new reduced plastocyanin molecule can bind to photosystem I. This release of oxidized plastocyanin with a half-life of about 60 microseconds limits the turnover of photosystem I. All data are consistently described by a model including the formation of a complex at a single binding site of photosystem I. Differences in the rate and binding constants are discussed with respect to the structure and the electrostatic and hydrophobic interactions stabilizing the complex as well as their modification by the membrane environment in situ.

The conformational requirements for the efficient assembly of bovine brain tubulin into microtubules have been investigated by using near-UV circular dichroism. Microtubule protein was prepared by the assembly-disassembly method of Shelanski et al. [Shelanski, M. L., Gaskin, F., & Cantor, C. R. (1973) Proc. Natl. Acad. Sci. U.S.A. 70, 765-768]. Tubulin dimer, isolated from this multiprotein complex by phosphocellulose ion-exchange chromatography in the presence and absence of Mg2+, was compared with tubulin dimer (WT dimer) prepared by the method of Weisenberg & Timasheff [Weisenberg, R. C., & Timasheff, S. N. (1970) Biochemistry 9, 4110-4116]. The tubulin from both dimer preparations showed identical electrophoretic patterns in which high molecular weight protein was undetectable. However, reproducible and significant differences were found in the near-UV CD spectra. Phosphocellulose-treated tubulin resembles the original microtubule protein more closely than does WT dimer, although this latter material has been widely accepted as being representative of the native protein. The phosphocellulose-treated tubulin and WT dimer are not readily interconvertible by simple physical or chemical treatments. The assembly capability of the various tubulin dimer preparations was compared by measuring the enhancement by tubulin dimer of assembly of ring fraction (isolated from microtubule protein by gel filtration of Sepharose 6B). Again phosphocellulose-treated tubulin is found to have more like native microtubule protein than does WT dimer.