Microdrop screening: A rapid method to optimize solvent conditions for NMR spectroscopy of proteins

Determining appropriate solvent conditions is a crucial first step for carrying out NMR spectroscopy of proteins, but rapid and efficient methods for doing so are currently lacking. Microdrop screening examines a large number of different solvent conditions using very small amounts of protein and minimal labor. Starting from one initial buffer condition, small aliquots of protein solution are combined with an array of solutions in which concentration, pH, buffer type, and added stabilizers are systematically varied. The protein concentration of each microliter-sized test drop (microdrop) is gradually changed using vapor diffusion, and the solubility of the protein is determined by visual examination. A variety of analytical techniques may be applied to the contents of the microdrops to monitor enzymatic activity, aggregation, ligand binding, and protein folding.     

Use of machine learning algorithms to classify binary protein sequences as highly-designable or poorly-designable

Background

By using a standard Support Vector Machine (SVM) with a Sequential Minimal Optimization (SMO) method of training, Naïve Bayes and other machine learning algorithms we are able to distinguish between two classes of protein sequences: those folding to highly-designable conformations, or those folding to poorly- or non-designable conformations.

Results

First, we generate all possible compact lattice conformations for the specified shape (a hexagon or a triangle) on the 2D triangular lattice. Then we generate all possible binary hydrophobic/polar (H/P) sequences and by using a specified energy function, thread them through all of these compact conformations. If for a given sequence the lowest energy is obtained for a particular lattice conformation we assume that this sequence folds to that conformation. Highly-designable conformations have many H/P sequences folding to them, while poorly-designable conformations have few or no H/P sequences. We classify sequences as folding to either highly – or poorly-designable conformations. We have randomly selected subsets of the sequences belonging to highly-designable and poorly-designable conformations and used them to train several different standard machine learning algorithms.

Conclusion

By using these machine learning algorithms with ten-fold cross-validation we are able to classify the two classes of sequences with high accuracy – in some cases exceeding 95%.

     

Methodological approaches and insights on protein aggregation in biological systems

Introduction: The proper folding of native proteins is critical and dynamic, but inherently unstable. Therefore, proteins eventually end up adopting misfolded conformations which compromise their function and may even trigger aggregation. Risk factors for neurodegenerative, metabolic and heart diseases compromise cellular protein quality-control systems, promoting protein aggregation. Multiple protein post-translational modifications dynamically regulate protein aggregation and disaggregation in a very complex, intricate and delicate balance.

Areas covered: Herein, we overview the more promising techniques and approaches for the elucidation of the biological implications of protein aggregation. The particular insights provided by different techniques were discriminated and several examples of post-translational modifications together with their targets were pooled and critically discussed, representing promising future therapeutic targets.

Expert commentary: In the years to come, differences between physiological and pathological protein aggregation will certainly become easier to determine. Techniques such as hydrogen/deuterium exchange, circular dichroism spectroscopy and novel mass spectrometry-based approaches are being optimized and are expected to introduce inhibitors of protein aggregation into the clinic. However, protein aggregation is not an isolated phenomenon, but rather influenced by multiple cellular components which complete knowledge is still far.     

Conformational Targeting of Fibrillar Polyglutamine Proteins in Live Cells Escalates Aggregation and Cytotoxicity

Background

Misfolding- and aggregation-prone proteins underlying Parkinson''s, Huntington''s and Machado-Joseph diseases, namely α-synuclein, huntingtin, and ataxin-3 respectively, adopt numerous intracellular conformations during pathogenesis, including globular intermediates and insoluble amyloid-like fibrils. Such conformational diversity has complicated research into amyloid-associated intracellular dysfunction and neurodegeneration. To this end, recombinant single-chain Fv antibodies (scFvs) are compelling molecular tools that can be selected against specific protein conformations, and expressed inside cells as intrabodies, for investigative and therapeutic purposes.

Methodology/Principal Findings

Using atomic force microscopy (AFM) and live-cell fluorescence microscopy, we report that a human scFv selected against the fibrillar form of α-synuclein targets isomorphic conformations of misfolded polyglutamine proteins. When expressed in the cytoplasm of striatal cells, this conformation-specific intrabody co-localizes with intracellular aggregates of misfolded ataxin-3 and a pathological fragment of huntingtin, and enhances the aggregation propensity of both disease-linked polyglutamine proteins. Using this intrabody as a tool for modulating the kinetics of amyloidogenesis, we show that escalating aggregate formation of a pathologic huntingtin fragment is not cytoprotective in striatal cells, but rather heightens oxidative stress and cell death as detected by flow cytometry. Instead, cellular protection is achieved by suppressing aggregation using a previously described intrabody that binds to the amyloidogenic N-terminus of huntingtin. Analogous cytotoxic results are observed following conformational targeting of normal or polyglutamine-expanded human ataxin-3, which partially aggregate through non-polyglutamine domains.

Conclusions/Significance

These findings validate that the rate of aggregation modulates polyglutamine-mediated intracellular dysfunction, and caution that molecules designed to specifically hasten aggregation may be detrimental as therapies for polyglutamine disorders. Moreover, our findings introduce a novel antibody-based tool that, as a consequence of its general specificity for fibrillar conformations and its ability to function intracellularly, offers broad research potential for a variety of human amyloid diseases.     

Semi-mature IL-12 secreting dendritic cells present exogenous antigen to trigger cytolytic immune responses

Dendritic cells (DC) are candidates for antigen-presenting cells that present exogenous antigen on MHC class I molecules to cytotoxic T lymphocytes (CTL), a process referred to as cross-priming. We triggered interleukin (IL)-12 release from DC, which was limited to the first day after maturation induction, by a combination of lipopolysaccharide (LPS) and interferon (IFN)-. To stimulate T lymphocytes, we used soluble protein derived from lysis of Epstein-Barr virus (EBV)-transformed lymphoblastoid cell lines (LCL) or ovalbumin loaded onto DC. Co-culture was initiated 2–6 or 48 h after maturation corresponding to semi-mature actively IL-12-secreting type 1 DC (sm-DC1) or a fully mature DC1 that had lost the ability to release IL-12 (fm-DC1), respectively. IL-12-secreting sm-DC1 but not fm-DC1 efficiently triggered cytolytic activity in autologous T lymphocytes. The combination of IL-1, IL-6, TNF-, and prostaglandin E2 generated type 2 DC that did not secrete IL-12 (DC2) and could not prime T-cell cytolytic activity. However, supplementation of cultures using DC2 with IL-12 resulted in CTL activity while the presence of anti-IL-12 monoclonal antibodies in cultures using IL-12 secreting sm-DC1 suppressed CTL activity. Thus, actively IL-12-secreting sm-DC1 are necessary and sufficient for the antigen-specific expansion of CTL in response to exogenously provided soluble antigen.     

Effect of short- and long-range interactions on protein folding

The relative importance of short- and long-range interactions is examined using a Monte Carlo simulation of protein folding on bovine pancreatic trypsin inhibitor. The model of the protein and the interaction energies were parametrized using X-ray structures of 30 native proteins. A nearest neighbor Ising model is used to determine the conformational state at each stage of the Monte Carlo procedure. Long-range interactions are simulated by contact free energies which become effective as two residues, separated by four or more residues along the chain, approach each other, and by disulfide-bond energies. Short-range interactions for residues separated by one, two, or three residues along the chain are also modeled by contact free energies and by -helical hydrogen bonds. A hard-sphere model is used to represent repulsive interactions. The ratios of short- to long-range interactions studied are 1:1, 2:1, 1:2, 0:1, and 1:0; e.g., for the 2:1 ratio, short-range interactions are weighted twice as much as long-range interactions, and for the 1:0 ratio, long-range interactions are omitted. For each ratio of short- to long-range interactions, a native conformation is found by a Monte Carlo procedure, a segment of 11 residues (residue numbers 1–11) is then rotated away from the rest of the molecule [breaking the 5–55 native disulfide bond, and moving this segment so that the distance between the sulfur atoms of the 5 and 55 cystine side chains (averaged for all native conformations) increases from 3.9 to 7.3 Å], and the Monte Carlo simulation is carried out (allowing the conformation of the whole molecule to change) until equilibrium is attained. For each ratio, the refolded conformation is compared to the native one using triangular distance maps and differential geometry distance criteria. With ratios of short- to long-range interaction energies of 1:1 and 0:1, the native disulfide bond could be re-formed; with ratios of 2:1 and 1:2 it did not; and with the 1:0 ratio, even a stable native conformation was not achieved. Therefore, long-range interactions (in addition to short-range ones) are required to bring remote parts of the protein together and to stabilize its native conformation.NIH Postdoctoral Fellow, 1977–1978.     

Vaginal candidosis

Enzyme linked immunosorbent assay was found to be a convenient method for the investigation of antibodies in mice immunized with Candida albicans ribosomes. Antibodies against the ribosomal antigen were detected in all the sera of mice (ICR and BALB/ c) immunized with ribosomes and incomplete Freund's adjuvant and in some of the sera of mice immunized with ribosomes only; the titer of antibodies varied from 1320 to 110 240. Vaccination of mice with ribosomal protein and IFA resulted in a high titer of antiribosomal antibodies. Treatment of ribosomes with pronase abrogated the capacity of the ribosomes to elicit anti ribosomal humoral responses, suggesting that the antibodies detected were directed against the protein moiety of the ribosomes. The presence of antibodies in sera of immunized mice could not be correlated with the protection afforded by the ribosomal vaccination.     

Conformational stability and aggregation of therapeutic monoclonal antibodies studied with ANS and thioflavin T binding

Characterization of aggregation profiles of monoclonal antibodies (mAb) is gaining importance because an increasing number of mAb-based therapeutics are entering clinical studies and gaining marketing approval. To develop a successful formulation, it is imperative to identify the critical biochemical properties of each potential mAb drug candidate. We investigated the conformational change and aggregation of a human IgG1 using external dye-binding experiments with fluorescence spectroscopy and compared the aggregation profiles obtained to the results of size-exclusion chromatography. We show that using an appropriate dye at selected mAb concentration, unfolding or aggregation can be studied. In addition, dye-binding experiments may be used as conventional assays to study therapeutic mAb stability.Key words: therapeutic monoclonal antibody, protein aggregation, conformational change, stability and shelf-life prediction, accelerated studiesMonoclonal antibodies (mAbs) have emerged as a novel class of protein drugs and are utilized for a variety of mostly incurable and debilitating diseases such as cancer and rheumatoid arthritis.^1–4 For treatment of chronic diseases, it is desirable for these drugs to be administered subcutaneously, in which case high protein concentrations (>100 mg/mL) are generally needed.^5,6 Protein-based drugs containing mAbs must contain minimum amounts of aggregation and fragmentation and conserve their structural integrity during storage because degraded or aggregated protein may induce immunogenicity or reduce efficacy. Currently, size-exclusion chromatography-high performance liquid chromatography (SEC-HPLC) is the most commonly used method to characterize mAb aggregation profiles;⁷ however it is time consuming, expensive and requires expertise. SEC-HPLC cannot be used to obtain accurate biophysical profiles of mAbs at high concentrations because dilution during the experiment might lead to reversible aggregation. Furthermore, the potential interaction of aggregates with surfaces, e.g., needle, tubing, column, will lead to the loss of sample and thus an inaccurate analysis.^8,9 Additional drawbacks of the technique are that different conformations such as partially unfolded monomers also cannot be distinguished by SEC-HPLC and large aggregates may be totally excluded during the injection into the column.External dye binding assays have been used to characterize protein stability and aggregation,^10–12 and studies involving biopharmaceuticals have been reported recently, e.g., for thermostability screening¹⁰ and detection of aggregation.^11–14 These methods are not limited by protein quantity and are more sensitive because they are fluorescence-based. We studied the accelerated unfolding of an IgG1 mAb with the hydrophobic dye 1-anilino-8-naphthale-nesulfonate (ANS), and its accelerated aggregation with aggregate specific Thioflavin T (ThT). We have also conducted accelerated aggregation studies with SEC-HPLC7 and compared the findings to the ThT binding results. We hypothesize that key structures formed during mAb aggregation can be probed selectively by the appropriate dyes (Fig. 1) with specific mAb concentrations.Open in a separate windowFigure 1Key structures of the mAb probed by fluorescent dyes. N and U are native and unfolded monomers, respectively. “n” reactive monomers form aggregates.     

Evidence for four nonallelic structural genes for the γ chain of human fetal hemoglobin

An abnormal human fetal hemoglobin not only may be either a^Gγ- or an^Aγ-chain variant but also may be present in a different proportion of the total fetal hemoglobin.^Gγ-Chain variants contribute either about one-fourth or one-eighth to the total production of HbF in the heterozygote, whereas the^Aγ-chain variants approximate either one-eighth or one-sixteenth of the total HbF. These observations may indicate the presence of four nonallelic Hb_γ structural genes (termed ) which produce γ chains in an approximate ratio of 4 : 2 : 2 : 1. HbF Malta I is considered to be the product of a mutant of the locus, an undefined HbF_x that of the locus, HbF Hull and HbF Jamaica products of mutated loci, and the newly discovered HbF Malta II a mutant of the gene. This work was supported in part by grants HL-05168 and HL-02558 from the National Institutes of Health, U.S. Public Health Service.     

Antigen forks: bispecific reagents that inhibit cell growth by binding selected pairs of tumor antigens

Bispecific antibodies of a new category, termed antigen forks, were constructed by crosslinking antibodies that recognized pairs of distinct tumor cell surface antigens. At concentrations of 1–100 nM, several such forks inhibited the growth of human tumor cell lines bearing both relevant antigens. The same cells were not inhibited by unconjugated component antibodies, and the active conjugates did not inhibit the growth of human cell lines that expressed lower levels of relevant antigens. The three most active antigen forks all contained monoclonal antibody 454A12, which recognizes human transferrin receptor. This antibody was conjugated respectively to antibodies 113F1 (against a tumor-associated glycoprotein complex), 317G5 (against a 42-kDa tumor-associated glycoprotein), or 520C9 (against the c-erbB-2 protooncogene product). The 317G5-454A12 fork strongly inhibited the HT-29 and SW948 human colorectal cancer cell lines, while the 113F1-454A12 fork was also effective against SW948. By designing forks against antigens of incompatible function that are co-expressed at high levels on tumor cells but not on normal tissues, it may be possible to generate reagents that inhibit tumor growth with enhanced selectivity.     

A replica exchange Monte Carlo algorithm for protein folding in the HP model

Background  

The ab initio protein folding problem consists of predicting protein tertiary structure from a given amino acid sequence by minimizing an energy function; it is one of the most important and challenging problems in biochemistry, molecular biology and biophysics. The ab initio protein folding problem is computationally challenging and has been shown to be -hard even when conformations are restricted to a lattice. In this work, we implement and evaluate the replica exchange Monte Carlo (REMC) method, which has already been applied very successfully to more complex protein models and other optimization problems with complex energy landscapes, in combination with the highly effective pull move neighbourhood in two widely studied Hydrophobic Polar (HP) lattice models.     

Scavenging of soluble organic matter from the prebiotic oceans

The existence of hot or cold nutrient broth or primeval soup is challenged on the basis of the recent geochemistry of soluble organic carbon in the oceans. Most of the dissolved organic carbon is recycled quickly by organisms, but the residual, biologically refractive, organic matter is efficiently scavenged from the oceans (residence time of 1000 to 3500 years) by nonbiologically mediated chemical and physical processes, such as adsorption on sinking minerals, polymerization and aggregation to humic type polymers or by aggregation to particulate matter through bubbling ans sinking of this material to the ocean bottom. Since there is no reason to believe that such nonbiological scavenging was not operative in the prebiotic oceans as well, then the prolonged existence of organic soup is very doubtful. The question of the orgin of life is thus assumed to be related to solid-liquid interfacial activity, and the answer may be associated with sediment-water interaction rather than with solution chemistry.     

A Comparison of Experimental and Computational Methods for Mapping the Interactions Present in the Transition State for Folding of FKBP12

The folding pathway of FKBP12, a 107 residue / protein, has been characterised in detail using a combination of experimental and computational techniques. FKBP12 follows a two-state model of folding in which only the denatured and native states are significantly populated; no intermediate states are detected. The refolding rate constant in water is 4 s^-1 at 25 °C. Two different experimental strategies were employed for studying the transition state for folding. In the first case, a non-mutagenic approach was used and the unfolding and refolding of the wild-type protein measured as a function of experimental conditions such as temperature, denaturant, ligand and trifluoroethanol (TFE) concentration. These data suggest a compact transition state relative to the unfolded state with some 70% of the surface area buried. The ligand-binding site, whichis mainly formed by two long loops, is largely unstructured in the transition state. TFE experiments suggest that the -helix may be formed in the transition state. The second experimental approach involved using protein engineering techniques with -value analysis. Residue-specific information on the structure and energetics of the transition state can be obtained by this method. 34 mutations were made at sites throughout the protein to probe the extent of secondary and tertiary structure in the transition state. In contrast to some other proteins of this size, no element of structure is fully formed in the transition state, instead, the transition state is similar to that found for smaller, single-domain proteins, such as chymotrypsin inhibitor 2 and the SH3 domainfrom -spectrin. For FKBP12, the central three strands of the -sheet (2, 4 and 5), comprise the most structured region of the transition state. In particular Val 101, which is one of the most highly buried residues and located in the middle of the central -strand,makes approximately 60% of its native interactions. The outer -strands, and the ends of the central -strands are formed to a lesser degree. The short -helix is largely unstructured in the transition state as are the loops. The data are consistent with a nucleation-condensation model of folding, the nucleus of which is formed by side chains within -strands 2, 4 and 5 and the C-terminus of the -helix. These residues are distant in the primary sequence, demonstrating the importance of tertiary interactions in the transition state. High-temperature molecular dynamic simulations on the unfoldingpathway of FKBP12 are in good agreement with the experimental results.     

Indications of a common folding pattern for VDAC channels from all sources

Previous research on the mitochondrial channel VDAC from the yeastS. cerevisiae had identified protein strands forming the wall of VDAC's aqueous pore. Here we report the results of analyzing the primary sequences of VDAC from various sources to see if the transmembrane folding pattern identified from this yeast is conserved for VDAC of different species. We analyzed the primary sequences of VDAC from higher plants, fungi, invertebrates, and vertebrates and found that all have a very similar -partern profile with 12–15 peaks indicating potential sided beta strands that are candidates for protein strands forming the wall of the aqueous pore. All these VDAC sequences can be put into the 13 transmembrane strand folding pattern previously identified for yeast VDAC. These folding patterns agree with available experimental data: both electrophysiological and protease digestion data. Although the primary sequences of VDAC from very diverse organisms show low homology, sequence similarity in the proposed corresponding 13 transmembrane strands is substantial. Competing proposals utilizing 16 transmembrane strands are in conflict with electrophysiological experimental observations and violate the constraints on such strands, such as no charged amino acids facing the phospholipid membrane and sufficient number of residues to span the membrane.     

Blastomycos1s: Specificity of antigens reflecting the mating types of Ajellomyces dermatitidis

In a study of sera from patients with proven or suspected blastomycosis, positive immunodiffusion tests were obtained in all active cases when fresh sera were tested with a cell sap (CS) antigen. False negatives occurred on occasion when an ethanol precipitate (EPF) antigen was used alone. No false positives were found. The CS antigen from the (+) mating type had in common two lines of identity with the (CS) antigen of the (–) type. In addition, other lines were present when the patient was infected with the same mating type as was used for the preparation of the antigen. No differences in the electrophoretic patterns of the enzymes leucine amino peptidase or phosphatase were noted when preparations from the two mating types were compared. However, a distinct pattern was noted when the esterases of the (+) and (–) mating types were examined. Specific esterase antibodies were present in patients' sera.     

Lateral diffusion of PDGF β-receptors in human fibroblasts

When platelet-derived growth factor (PDGF) binds to its receptors a number of biochemical reactions are elicited in the cell. Several models have been presented for the effects of ligand-induced receptor conformation and aggregation on signal transduction but little is known about the direct effects on receptor diffusion. This study concerns the lateral mobility of PDGF receptors in fibroblasts. It was assessed with fluorescence recovery after photobleaching (FRAP), using rhodaminated receptor antibodies or Fab-fragments of the antibody as ligands. The aims of the investigation were: (a) to compare the lateral mobility of membrane receptors of human fibroblasts labelled with either antibodies against the PDGF receptor or Fab-fragments of the same antibodies, and (b) to study the effects of serum or PDGF on the mobility of the receptors. Human foreskin fibroblasts (AG 1523) were grown on coverslips either under standard or under serum-free conditions yielding normal and starved cells, respectively. Two parameters of the diffusion were evaluated; the diffusion coefficient (D) and the mobile fraction (R) of the receptors. We found that normal fibroblasts had a smaller diffusion coefficient and a lower mobile fraction compared to starved cells using antibodies for receptor labelling. The addition of PDGF, just before the measurement, increased the D and R for normal cells, while starved cells, showing higher initial values, displayed slightly reduced values of D and R. After the addition of serum, D increased and R remained low for normal cells, whereas for starved cells both D and R increased to upper limits of 11.0×10^–10 cm²s^–1 and >90% respectively. In general, the D and R values, both in normal and starved cells, were higher for cells labelled with Fab-fragments than for antibody-labelled cells. The results are discussed in relation to the natural complexity of the receptor, and how PDGF, serum, antibodies and Fab-fragments might interfere with receptor structure, aggregation state and membrane diffusion characteristics.     

Purification of cytoplasmic antigens from the mycelial phase of Candida albicans: Possible advantages of its use in Candida serology

The serology of candidiasis is complicated by the use of poorly defined antigens. Total extracts of the yeast phase have been commonly used as cytoplasmic antigen, without regard to the significant amounts of carbohydrate that may contaminate such preparations. This is particularly true in the case of commercially available antigens that have been used as cytoplasmic antigens but actually are richer in carbohydrate than in protein. Affinity chromatography in concanavalin A — Sepharose provides a simple procedure to separate carbohydrates, mainly mannan, from protein antigens in whole Candida extracts. By using mannan-poor antigens, the specificity of serological reactions can be increased considerably, since both the positive reactions seen in asymptomatic donors and the cross-reactions seen in patients infected with other fungi are due to anti-mannan antibodies. In contrast, both anti-mannan and anti-cytoplasmic antigen antibodies can be detected in patients suspected of systemic candidiasis. On the other hand, absolute specificity may never be achieved for systemic candidiasis. We have found antibodies against cytoplasmic antigen in a patient allergic to C. albicans, in whom the microorganism was isolated from fecal material. It appears that, under favorable conditions, mucosal sensitization may also trigger a systemic reaction directed against both mannan and cytoplasmic antigens.Publication no. 341 from The Department of Basic and Clinical Immunology and Microbiology, Medical University of South Carolina.     

Molecular mapping of human band 3 aging antigenic sites and active amino acids using synthetic peptides

An aging antigen, senescent cell antigen appears on old cells and marks them for death by initiating the binding of IgG autoantibody and subsequent removal by phagocytes. This antigen is derived from the major anion transport protein, protein band 3, that is involved in respiration and acid base balance. We use synthetic peptides from the transmembrane, anion transport segment of band 3 to walk band 3 to identify potential aging antigenic sites. A competitive inhibition assay with affinity purified IgG autoantibody from senescent red cells was used. Results indicate that: aging antigenic sites reside on human band 3 residues 538–554, 593–601, and 812–830; and that the smallest residues which act as aging antigenic sites are 593–601 and 813–818. The contribution of lysine and/or arginine to antigenicity is examined by synthesizing peptide analogs in which glycines or arginines are substituted for lysines or arginines. Substitution of neutral glycine for the positively charged amino acids arginine or lysine or both arginine and lysine did not result in a significant difference in antigenicity between the analog and the native band 3 peptide. Substitution of the positively charged arginine for the positively charged lysine resulted in a significant reduction in antigenicity. The chicken sequence of band 3 peptides 538–554 and 812–827 differs from that of the human peptides at several sites. Antigenicity of these chicken analogs were tested and compared to the human peptides. The data suggest that the three-dimensional configuration of band 3 segments plays a dominant role in defining the antigenic determinants reactive with senescent cell IgG autoantibodies.