Protein photo-cross-linking in mammalian cells by site-specific incorporation of a photoreactive amino acid

We report a method of photo-cross-linking proteins in mammalian cells, which is based on site-specific incorporation of a photoreactive amino acid, p-benzoyl-L-phenylalanine (pBpa), through the use of an expanded genetic code. To analyze the cell signaling interactions involving the adaptor protein Grb2, pBpa was incorporated in its Src homology 2 (SH2) domain. The human GRB2 gene with an amber codon was introduced into Chinese hamster ovary (CHO) cells, together with the genes for the Bacillus stearothermophilus suppressor tRNA(Tyr) and a pBpa-specific variant of Escherichia coli tyrosyl-tRNA synthetase (TyrRS). The Grb2 variant with pBpa in the amber position was synthesized when pBpa was included in the growth medium. Upon exposure of cells to 365-nm light, protein variants containing pBpa in the positions proximal to the ligand-binding pocket were cross-linked with the transiently expressed epidermal growth factor (EGF) receptor in the presence of an EGF stimulus. Cross-linked complexes with endogenous proteins were also detected. In vivo photo-cross-linking with pBpa incorporated in proteins will be useful for studying protein-protein interactions in mammalian cells.     

Electron‐deficient p‐benzoyl‐l‐phenylalanine derivatives increase covalent chemical capture yields for protein–protein interactions

The photoactivatable amino acid p‐benzoyl‐l ‐phenylalanine (pBpa) has been used for the covalent capture of protein–protein interactions (PPIs) in vitro and in living cells. However, this technique often suffers from poor photocrosslinking yields due to the low reactivity of the active species. Here we demonstrate that the incorporation of halogenated pBpa analogs into proteins leads to increased crosslinking yields for protein–protein interactions. The analogs can be incorporated into live yeast and upon irradiation capture endogenous PPIs. Halogenated pBpas will extend the scope of PPIs that can be captured and expand the toolbox for mapping PPIs in their native environment.     

Cross-linking of mRNA to initiation factor eIF-3, 24 kDa cap binding protein and ribosomal proteins S1, S3/3a, S6 and S11 within the 48S pre-initiation complex

Native small ribosomal subunits from rabbit reticulocytes contain all initiation factors necessary for the formation of the mRNA-containing 48S pre-initiation complex. The complex formed in the presence of Met-tRNAf and 125I-labelled globin mRNA was cross-linked with diepoxybutane, and the covalent mRNA-protein complexes were isolated under denaturating conditions. The proteins of the covalent complex were identified as the 110, 95 and 66/64 kDa subunits of eIF-3. In addition, the 24 kDa cap binding protein and the ribosomal proteins S1, S3/3a, S6 and S11 were found covalently linked to the mRNA. Ribosomal proteins S3/3a and S6 were also involved in the ribosomal mRNA-binding domain of reticulocyte polysomes.     

Cross-linking of initiation factor IF3 to proteins of the Escherichia coli 30 S ribosomal subunit

Complexes of 30 S subunits and [¹⁴C]IF3 were allowed to react with the protein cross-linking reagents, N,N′-p-phenylenedimaleimide or dimethylsuberimidate. Non-cross-linked IF3 was removed from the complex by centrifugation in a buffer containing a high salt concentration, and the total protein was extracted from the pelleted particles. The mixture of cross-linked products was analyzed by radioimmunodiffusion with antisera prepared against all of the individual 30 S ribosomal proteins. Radioactivity was found in the precipitin bands formed with antisera against ribosomal proteins S1, S11, S12, S13, S19 and S21. The results show that IF3 was present in covalent cross-linked complexes containing those 30 S ribosomal proteins and imply that they comprise or are near the binding site for initiation factor IF3.     

Structure of the oncoprotein gankyrin in complex with S6 ATPase of the 26S proteasome

Gankyrin is an oncoprotein commonly overexpressed in most hepatocellular carcinomas. Gankyrin interacts with S6 ATPase of the 19S regulatory particle of the 26S proteasome and enhances the degradation of the tumor suppressors pRb and p53. Here, we report the structure of gankyrin in complex with the C-terminal domain of S6 ATPase. Almost all of the seven ankyrin repeats of gankyrin interact, through its concave region, with the C-terminal domain of S6 ATPase. The intermolecular interactions occur through the complementary charged residues between gankyrin and S6 ATPase. Biochemical studies based on the structure of the complex revealed that gankyrin interacts with pRb in both the presence and absence of S6 ATPase; however, the E182 residue in gankyrin is essential for the pRb interaction. These results provide a structural basis for the involvement of gankyrin in the pRb degradation pathway, through its association with S6 ATPase of the 26S proteasome.     

Identification of the ribosomal proteins present in the vicinity of globin mRNA in the 40S initiation complex

The interaction of ribosomal proteins with mRNA in the 40S initiation complex was examined by chemical cross-linking. 40S initiation complexes were formed by incubating rat liver [(3)H]Met-tRNAi, rat liver 40S ribosomal subunits, rabbit globin mRNA, and partially purified initiation factors of rabbit reticulocytes in the presence of guanylyl(beta, gamma-methylene)-diphosphonate. The initiation complexes were then treated with 1,3-butadiene diepoxide to introduce crosslinks between the mRNA and proteins. The covalent mRNA-protein conjugates were isolated by chromatography on an oligo(dT) cellulose column in the presence of sodium dodecyl sulfate, followed by sucrose density gradient centrifugation. Proteins cross-linked to the mRNA were labeled with Na(125)I, extracted by extensive ribonuclease digestion, and analyzed by two-dimensional and diagonal polyacrylamide gel electrophoresis. Three ribosomal proteins, S6, S8, and S23/S24, together with small amounts of S3/S3a, S27, and S30, were identified as the protein components cross-linked to the globin mRNA protein complex, and were shown to attach directly to the mRNA. It is suggested that these proteins constitute the ribosomal binding site for mRNA in the 40S initiation complex.     

Effects of two photoreactive spermine analogues on peptide bond formation and their application for labeling proteins in Escherichia coli functional ribosomal complexes.

The effect of two photoreactive analogues of spermine, N(1)-azidobenzamidino- (ABA-) spermine and N(1)-azidonitrobenzoyl- (ANB-) spermine, on ribosomal functions was studied in a cell-free system derived from Escherichia coli. In the dark, both analogues stimulated the binding of AcPhe-tRNA to poly(U)-programmed ribosomes, enhanced the stability of the ternary complex AcPhe-tRNA.poly(U).ribosome (complex C), and caused stimulatory and inhibitory effects on peptidyltransferase activity. ABA-spermine exhibited more pronounced effects than ANB-spermine. Each photoprobe was covalently attached after irradiation to both ribosomal subunits and also to free rRNA isolated from 70S ribosomes. Photolabeled complex C showed a reactivity toward puromycin, similar to that exhibited by complex C reacting reversibly with photoprobes free in solution. The distribution of the incorporated radioactivity among the ribosomal components was determined under two experimental conditions, one stimulating and the other inhibiting peptidyltransferase activity. Under both conditions, ABA-spermine was the strongest cross-linker. Upon stimulatory conditions, 14% of ABA-[(14)C]spermine cross-linked to complex C was bound to the protein fraction. The proteins primarily labeled were identified as S3, S4, L2, L3, L6, L15, L17, and L18. Upon inhibitory conditions, a higher percent of the incorporated radioactivity was found in ribosomal proteins, while the pattern of protein labeling was characterized by a remarkable decrease of cross-linked proteins L2, L3, L6, L15, L17. and L18 and by an increase of cross-linked proteins S9, S18, L1, L16, L22, L23, and L27. On the basis of these results and literature data, the involvement of spermine in the conformation and important functions of ribosomes is discussed.     

Tandem photoaffinity labeling of a target protein using a linker with biotin,alkyne and benzophenone groups and a bioactive small molecule with an azide group

A novel linker containing biotin, alkyne and benzophenone groups (1) was synthesized to identify target proteins using a small molecule probe. This small molecule probe contains an azide group (azide probe) that reacts with an alkyne in 1 via an azide–alkyne Huisgen cycloaddition. Cross-linking of benzophenone to the target protein formed a covalently bound complex consisting of the azide probe and the target protein via 1. The biotin was utilized via biotin–avidin binding to identify the cross-linked complex. To evaluate the effectiveness of 1, it was applied in a model system using an allene oxide synthase (AOS) from the model moss Physcomitrella patens (PpAOS1) and an AOS inhibitor that contained azide group (3). The cross-linked complex consisting of PpAOS1, 1 and 3 was resolved via SDS–PAGE and visualized using a chemiluminescent system. The method that was developed in this study enables the effective identification of target proteins.     

Structure of T4 pyrimidine dimer glycosylase in a reduced imine covalent complex with abasic site-containing DNA

The base excision repair (BER) pathway for ultraviolet light (UV)-induced cyclobutane pyrimidine dimers is initiated by DNA glycosylases that also possess abasic (AP) site lyase activity. The prototypical enzyme known to catalyze these reactions is the T4 pyrimidine dimer glycosylase (T4-Pdg). The fundamental chemical reactions and the critical amino acids that lead to both glycosyl and phosphodiester bond scission are known. Catalysis proceeds via a protonated imine covalent intermediate between the alpha-amino group of the N-terminal threonine residue and the C1' of the deoxyribose sugar of the 5' pyrimidine at the dimer site. This covalent complex can be trapped as an irreversible, reduced cross-linked DNA-protein complex by incubation with a strong reducing agent. This active site trapping reaction is equally efficient on DNA substrates containing pyrimidine dimers or AP sites. Herein, we report the co-crystal structure of T4-Pdg as a reduced covalent complex with an AP site-containing duplex oligodeoxynucleotide. This high-resolution structure reveals essential precatalytic and catalytic features, including flipping of the nucleotide opposite the AP site, a sharp kink (approximately 66 degrees ) in the DNA at the dimer site and the covalent bond linking the enzyme to the DNA. Superposition of this structure with a previously published co-crystal structure of a catalytically incompetent mutant of T4-Pdg with cyclobutane dimer-containing DNA reveals new insights into the structural requirements and the mechanisms involved in DNA bending, nucleotide flipping and catalytic reaction.     

Intramolecular cross-linking evaluated as a structural probe of the protein folding transition state

We examine the utility of intramolecular covalent cross-linking to identify the structure present in the folding transition state. In mammalian ubiquitin, cysteine residues located across two beta-strands are cross-linked with dichloroacetone. The kinetic effects of these covalent cross-links in ubiquitin, and engineered disulfide bonds in src SH3 (Grantcharova, V. P., Riddle, D. S., and Baker, D. (2000) Proc. Natl. Acad. Sci. U.S.A. 97, 7084-7089), are compared to the results of psi-analysis where strand association is stabilized by metal ion binding to engineered bihistidine sites (Krantz, B. A., Dothager, R. S., and Sosnick, T. R. (2004) J. Mol. Biol. 337, 463-75) at the same positions. The results for the two methods agree at some of the sites. The cross-linking phi crosslink-values agree with their corresponding psi-values when they have both have values of zero or one, which represent the absence and presence of native structure, respectively. When phi crosslink > psi, the apparent inconsistency is rationalized by the difference between each method's mode of stabilization; cross-linking reduces the configurational entropy of the unfolded state whereas metal binding directly stabilizes the native state. However, when the cross-linking phi-values are smaller than their corresponding psi-values, the apparent underestimation of structure formation is difficult to rationalize while retaining the assumption that the cross-link exclusively affects the entropy of the unfolded state. The interpretation also is problematic for data on cross-links located across strands which are not hairpins, and hence, these sites are likely to be of limited utility in folding studies. We conclude that cross-linking data for sites on hairpins generally report on the amount of structure formed within the enclosed loop while the metal binding data report on the amount structure formed at the site itself.     

Conformational detours during folding of a collapsed state

The protein S6 is a useful model to probe the role of partially folded states in the folding process. In the absence of salt, S6 folds from the denatured state D to the native state N without detectable intermediates. High concentrations of sodium sulfate induce the accumulation of a collapsed state C, which is off the direct folding route. However, the mutation VA85 enables S6 to fold from C directly to N through the transition state TS(C). According to the denaturant dependence of this reaction, TS(C) and C are equally compact, but the data are difficult to deconvolute. Therefore, I have measured the heat capacities (DeltaC(p)) for the D-->C and C-->TS(C) transitions. The DeltaC(p)-values suggest that C needs to increase its surface area in order to fold directly to N. This underlines that it is a misfolded state that can only fold by at least partial unfolding. In contrast to the C-state formed by S6 wildtype, the VA85 C-state is just as compact as the native state, and this may be a prerequisite for direct folding. Individual "gatekeeper" residues may thus play a disproportionately large role in guiding proteins through different folding pathways.     

Covalent cross-linking of proteins without chemical reagents

A facile method for the formation of zero-length covalent cross-links between protein molecules in the lyophilized state without the use of chemical reagents has been developed. The cross-linking process is performed by simply sealing lyophilized protein under vacuum in a glass vessel and heating at 85 degrees C for 24 h. Under these conditions, approximately one-third of the total protein present becomes cross-linked, and dimer is the major product. Chemical and mass spectroscopic evidence obtained shows that zero-length cross-links are formed as a result of the condensation of interacting ammonium and carboxylate groups to form amide bonds between adjacent molecules. For the protein examined in the most detail, RNase A, the cross-linked dimer has only one amide cross-link and retains the enzymatic activity of the monomer. The in vacuo cross-linking procedure appears to be general in its applicability because five different proteins tested gave substantial cross-linking, and co-lyophilization of lysozyme and RNase A also gave a heterogeneous covalently cross-linked dimer.     

F protein-F protein interaction within the Sendai virus identified by native bonding or chemical cross-linking

The spatial arrangement of the F protein spike in the Sendai virus was studied after purifying the protein and reconstituting it in lipid vesicles (Sechoy, O., Philippot, J. R., and Bienvenue, A. (1986) Biochim. Biophys. Acta 857, 1-12). The different components of the F protein spikes were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis under various conditions of treatment, i.e. at different temperatures and sodium dodecyl sulfate concentrations, using different detergents for F protein solubilization (Triton X-100 and octyl glucoside), by fast protein liquid chromatography analysis, and by chemical cross-linking between subunits with bifunctional agents such as dimethyl adipimidate and dithiobis(succinimidyl propionate). The F protein spike appeared to be a structurally stable complex, composed of a noncovalent association of four homooligomers, each consisting of two peptides, F1 and F2, linked by a disulfide bond. Octyl glucoside and Triton X-100 solubilized the F protein, preserving the tetramer, which is probably the native form. Using chemical cross-linking, a covalent bond was formed between two monomers. We hypothesize that the tetrameric form of the F protein in its native form (spike) consists of two identical dimers that can be chemically cross-linked in a stable complex.     

The formation of a disulfide cross-link between the two subunits demonstrates the dimeric structure of the mitochondrial oxoglutarate carrier

Isolated oxoglutarate carrier (OGC) can be cross-linked to dimers by disulfide-forming reagents such as Cu²⁺-phenanthroline and diamide. Acetone and other solvents increase the extent of Cu²⁺-phenanthroline-induced cross-linking of OGC. Cross-linked OGC re-incorporated in photeoliposomes fully retains the oxoglutarate transport activity. The amount of cross-linked OGC calculated by densitometry of scanned gels depends on the method of staining, since cross-linked OGC exhibits a higher sensitivity to Coomassie brilliant blue as compared to silver nitrate. Under optimal conditions the formation of cross-linked OGC dimer (stained with Coomassie brilliant blue) amounts to 75% of the total protein. Approximately the same cross-linking efficiency was evaluated from Western blots. Cross-linking of OGC is prevented by SH reagents and reversed by SH-reducing reagents, which shows that it is mediated by disulfide bridge(s). The formation of SS bridge(s) requires the native state of the protein, since it is suppressed by SDS and by heating. Furthermore, the extent of cross-linking is independent of OGC concentration indicating that disulfide bridge(s) must be formed between the two subunits of native dimers. The number and localization of disulfide bridge(s) in the cross-linked OGC were examined by peptide fragmentation and subsequent cleavage of disulfide bond(s) by β-mercaptoethanol. Our experimental results show that cross-linking of OGC is accomplished by a single disulfide bond between the cysteines 184 of the two subunits and suggest that these residues in the putative transmembrane helix four are fairly close to the twofold axis of the native dimer structure.     

Gankyrin is an ankyrin-repeat oncoprotein that interacts with CDK4 kinase and the S6 ATPase of the 26 S proteasome.

A yeast two-hybrid screen with the human S6 (TBP7, RPT3) ATPase of the 26 S proteasome has identified gankyrin, a liver oncoprotein, as an interacting protein. Gankyrin interacts with both free and regulatory complex-associated S6 ATPase and is not stably associated with the 26 S particle. Deletional mutagenesis shows that the C-terminal 78 amino acids of the S6 ATPase are necessary and sufficient to mediate the interaction with gankyrin. Deletion of an orthologous gene in Saccharomyces cerevisiae suggests that it is dispensable for cell growth and viability. Overexpression and precipitation of tagged gankyrin from cultured cells detects a complex containing co-transfected tagged S6 ATPase (or endogenous S6) and endogenous cyclin D-dependent kinase CDK4. The proteasomal ATPases are part of the AAA (ATPases associated with diverse cellular activities) family, members of which are molecular chaperones; gankyrin complexes may therefore influence CDK4 function during oncogenesis.     

Immunoglobulin M possesses two binding sites for complement subcomponent C1q, and soluble 1:1 and 2:1 complexes are formed in solution at reduced ionic strength

Soluble complexes were formed between C1q, a subunit of the first component of human complement, and four different Waldenstr?m IgM proteins at reduced ionic strengths. The equilibria between these complexes and the free proteins were studied in the ultracentrifuge. Complex formation was found to be a very sensitive function of the salt concentration, and at physiological ionic strength complex formation was negligible. The complexes were cross-linked with a water-soluble carbodiimide and separated by sucrose gradient centrifugation. Both 22 S 1:1 and 26 S 2:1 C1q X IgM complexes were formed; stoichiometry was established by cross-linking 125I-C1q with 131I-IgM and determining the ratios of the specific activities of the gradient-purified materials. The association process was studied as a function of protein concentration and was analyzed by Scatchard and Hill plots to yield stoichiometry, association constant, and degree of cooperativity. The results indicated that IgM has two identical and independent binding sites for C1q. The intrinsic association constant was found to vary between 10(6) M-1 at 0.084 M ionic strength to 10(4) M-1 at physiological ionic strength; the slope of the log-log plot gave a value of -6.0. The cross-linked complexes were examined by electron microscopy, and the C1q appeared to be attached to the IgM through the C1q heads, implying that the biologically significant binding sites were involved in this interaction. For the 2:1 complexes, the two C1q appeared to attach to opposite surfaces of the IgM, suggesting the presence of a pseudo-2-fold axis lying in the plane of the IgM disk.     

Identification of variant molecules of Bacillus thermoproteolyticus ferredoxin: crystal structure reveals bound coenzyme A and an unexpected [3Fe-4S] cluster associated with a canonical [4Fe-4S] ligand motif

During the purification of recombinant Bacillus thermoproteolyticus ferredoxin (BtFd) from Escherichia coli, we have noted that some Fe-S proteins were produced in relatively small amounts compared to the originally identified BtFd carrying a [4Fe-4S] cluster. These variants could be purified into three Fe-S protein components (designated as V-I, V-II, and V-III) by standard chromatography procedures. UV-vis and EPR spectroscopic analyses indicated that each of these variants accommodates a [3Fe-4S] cluster. From mass spectrometric and protein sequence analyses together with native and SDS gel electrophoresis, we established that V-I and V-II contain the polypeptide of BtFd associated with acyl carrier protein (ACP) and with coenzyme A (CoA), respectively, and that V-III is a BtFd dimer linked by a disulfide bond. The crystal structure of the BtFd-CoA complex (V-II) determined at 1.6 A resolution revealed that each of the four complexes in the crystallographic asymmetric unit possesses a [3Fe-4S] cluster that is coordinated by Cys(11), Cys(17), and Cys(61). The polypeptide chain of each complex is superimposable onto that of the original [4Fe-4S] BtFd except for the segment containing Cys(14), the fourth ligand to the [4Fe-4S] cluster of BtFd. In the variant molecules, the side chain of Cys(14) is rotated away to the molecular surface, forming a disulfide bond with the terminal sulfhydryl group of CoA. This covalent modification may have occurred in vivo, thereby preventing the assembly of the [4Fe-4S] cluster as observed previously for Desulfovibrio gigas ferredoxin. Possibilities concerning how the variant molecules are formed in the cell are discussed.     

Evidence for a thiol ester in duck ovostatin (ovomacroglobulin)

The structure and the mechanism for proteinase inhibition of the egg white protein ovostatin (ovomacroglobulin) are similar to those of plasma alpha 2-macroglobulin, but previous studies have shown that chicken ovostatin lacks a reactive thiol ester (Nagase, H., and Harris, E. D., Jr. (1983) J. Biol. Chem. 258, 7490-7498). Here we show that duck ovostatin has conserved such a thiol ester and is capable of inhibiting both metallo- and serine proteinases stoichiometrically. Evidence for thiol esters was established by the following results with duck ovostatin: 1) autolysis into fragments of Mr = 123,000 and 60,000 occurred by heating in sodium dodecyl sulfate, but was prevented by treatment with CH3NH2; 2) covalent linkages were formed with proteinases on complex formation; 3) reaction with CH3NH2 generated 3.6 SH groups/mol, and 3.9 mol of [14C]CH3NH2 were incorporated per mol of protein; and 4) saturation with a proteinase liberated 3.8 SH groups/mol of the inhibitor. Conformational rearrangement of duck ovostatin upon reacting with CH3NH2 or proteinases was demonstrated by an increased mobility of the protein in polyacrylamide gel electrophoresis. CH3NH2-treated duck ovostatin was able to bind and inhibit proteinases without forming covalent bonds, but, unlike unmodified ovostatin, its inhibitory activity was destroyed by freezing and thawing. Complexes formed between CH3NH2-treated duck ovostatin and a proteinase were not dissociable except under denaturing conditions. These results and other evidence indicate that covalent bond formation through reaction with a thiol ester is a separate process from the trapping and inhibition of proteinases by this family of proteins.     

Structural basis for the recognition between the regulatory particles Nas6 and Rpt3 of the yeast 26S proteasome

The 26S proteasome-dependent protein degradation is an evolutionarily conserved process. The mammalian oncoprotein gankyrin, which associates with S6 of the proteasome, facilitates the degradation of pRb, and thus possibly acts as a bridging factor between the proteasome and its substrates. However, the mechanism of the proteasome-dependent protein degradation in yeast is poorly understood. Here, we report the tertiary structure of the complex between Nas6 and a C-terminal domain of Rpt3, which are the yeast orthologues of gankyrin and S6, respectively. The concave region of Nas6 bound to the alpha-helical domain of Rpt3. The stable interaction between Nas6 and Rpt3 was mediated by intermolecular interactions composed of complementary charged patches. The recognition of Rpt3 by Nas6 in the crystal suggests that Nas6 is indeed a subunit of the 26S proteasome. These results provide a structural basis for the association between Nas6 and the heterohexameric ATPase ring of the proteasome through Rpt3.