Genetic Variation/Evolution and Differential Host Responses Resulting from In-Patient Adaptation of Mycobacterium avium.

Members of the Mycobacterium avium complex (MAC) are characterized as nontuberculosis mycobacteria and are pathogenic mainly in immunocompromised individuals. MAC strains show a wide genetic variability, and there is growing evidence suggesting that genetic differences may contribute to a varied immune response that may impact the infection outcome. The current study aimed to characterize the genomic changes within M.avium isolates collected from single patients over time and test the host immune responses to these clinical isolates. Pulsed-field gel electrophoresis and whole-genome sequencing were performed on 40 MAC isolates isolated from 15 patients at the Department of Medical Microbiology at St. Olavs Hospital in Trondheim, Norway. Isolates from patients (patients 4, 9, and 13) for whom more than two isolates were available were selected for further analysis. These isolates exhibited extensive sequence variation in the form of single-nucleotide polymorphisms (SNPs), suggesting that M. avium accumulates mutations at higher rates during persistent infections than other mycobacteria. Infection of murine macrophages and mice with sequential isolates from patients showed a tendency toward increased persistence and the downregulation of inflammatory cytokines by host-adapted M. avium strains. The study revealed the rapid genetic evolution of M. avium in chronically infected patients, accompanied by changes in the virulence properties of the sequential mycobacterial isolates.

The EXIT Strategy: an Approach for Identifying Bacterial Proteins Exported during Host Infection.

Exported proteins of bacterial pathogens function both in essential physiological processes and in virulence. Past efforts to identify exported proteins were limited by the use of bacteria growing under laboratory (in vitro) conditions. Thus, exported proteins that are exported only or preferentially in the context of infection may be overlooked. To solve this problem, we developed a genome-wide method, named EXIT (exported in vivotechnology), to identify proteins that are exported by bacteria during infection and applied it to Mycobacterium tuberculosis during murine infection. Our studies validate the power of EXIT to identify proteins exported during infection on an unprecedented scale (593 proteins) and to reveal in vivo induced exported proteins (i.e., proteins exported significantly more during in vivo infection than in vitro). Our EXIT data also provide an unmatched resource for mapping the topology of M. tuberculosis membrane proteins. As a new approach for identifying exported proteins, EXIT has potential applicability to other pathogens and experimental conditions.IMPORTANCE There is long-standing interest in identifying exported proteins of bacteria as they play critical roles in physiology and virulence and are commonly immunogenic antigens and targets of antibiotics. While significant effort has been made to identify the bacterial proteins that are exported beyond the cytoplasm to the membrane, cell wall, or host environment, current methods to identify exported proteins are limited by their use of bacteria growing under laboratory (in vitro) conditions. Because in vitro conditions do not mimic the complexity of the host environment, critical exported proteins that are preferentially exported in the context of infection may be overlooked. We developed a novel method to identify proteins that are exported by bacteria during host infection and applied it to identify Mycobacterium tuberculosis proteins exported in a mouse model of tuberculosis.

High throughput crystallography of TB drug targets.

Tuberculosis (TB) infects one-third of the world population. Despite 50 years of available drug treatments, TB continues to increase at a significant rate. The failure to control TB stems in part from the expense of delivering treatment to infected individuals and from complex treatment regimens. Incomplete treatment has fueled the emergence of multi-drug resistant (MDR) strains of Mycobacterium tuberculosis (Mtb). Reducing non-compliance by reducing the duration of chemotherapy will have a great impact on TB control. The development of new drugs that either kill persisting organisms, inhibit bacilli from entering the persistent phase, or convert the persistent bacilli into actively growing cells susceptible to our current drugs will have a positive effect. We are taking a multidisciplinary approach that will identify and characterize new drug targets that are essential for persistent Mtb. Targets are exposed to a battery of analyses including microarray experiments, bioinformatics, and genetic techniques to prioritize potential drug targets from Mtb for structural analysis. Our core structural genomics pipeline works with the individual laboratories to produce diffraction quality crystals of targeted proteins, and structural analysis will be completed by the individual laboratories. We also have capabilities for functional analysis and the virtual ligand screening to identify novel inhibitors for target validation. Our overarching goals are to increase the knowledge of Mtb pathogenesis using the TB research community to drive structural genomics, particularly related to persistence, develop a central repository for TB research reagents, and discover chemical inhibitors of drug targets for future development of lead compounds.

FINDMOL: automated identification of macromolecules in electron-density maps.

Automating the determination of novel macromolecular structures via X-ray crystallographic methods involves building a model into an electron-density map. Unfortunately, the conventional crystallographic asymmetric unit volumes are usually not well matched to the biological molecular units. In most cases, the facets of the asymmetric unit cut the molecules into a number of disconnected fragments, rendering interpretation by the crystallographer significantly more difficult. The FINDMOL algorithm is designed to quickly parse the arrangement of trace points (pseudo-atoms) derived from a skeletonized electron-density map without requiring higher level prior information such as sequence information or number of molecules in the asymmetric unit. The algorithm was tested with a variety of density-modified maps computed with medium- to low-resolution data. Typically, the resulting volume resembles the biological unit. In the remaining cases the number of disconnected fragments is very small. In all examples, secondary-structural elements such as alpha-helices or beta-sheets are easily identifiable in the defragmented arrangement. FINDMOL can greatly assist a crystallographer during manual model building or in cases where automatic model building can only build partial models owing to limitations of the data such as low resolution and/or poor phases.

Antibody-ligand interactions: computational modeling and correlation with biophysical measurements.

Several new aspects of computer-assisted molecular modeling strategies and biophysical techniques, such as fluorescence spectroscopy, circular dichroism, and absorption spectroscopy, have proved useful in the analysis and description of antibody-ligand interactions. The molecular features involved in determining the specificity of antibody-ligand interactions, such as electrostatics (e.g. partial charges, salt bridges, p-cation motifs), hydrogen-bonds, polarization, hydrophobic interactions, hydration and solvation effects, entropy, and kinetics can be identified using a battery of biophysical techniques. An understanding of these parameters is essential to our use of antibodies as tools in high throughput screening of chemical libraries for the discovery of novel compounds.

Determining protein structure from electron-density maps using pattern matching.

TEXTAL is an automated system for building protein structures from electron-density maps. It uses pattern recognition to select regions in a database of previously determined structures that are similar to regions in a map of unknown structure. Rotation-invariant numerical values, called features, of the electron density are extracted from spherical regions in an unknown map and compared with features extracted around regions in maps generated from a database of known structures. Those regions in the database that match best provide the local coordinates of atoms and these are accumulated to form a model of the unknown structure. Similarity between the regions in the database and an uninterpreted region is determined firstly by evaluating the numerical difference in feature values and secondly by calculating the electron-density correlation coefficient for those regions with similar feature values. TEXTAL has been successful at building protein structures for a wide range of test electron-density maps and can automatically model entire protein structures in a few hours on a workstation. Models built by TEXTAL from test electron-density maps of known protein structures were accurate to within 0.6-0.7 A root-mean-square deviation, assuming prior knowledge of C(alpha) positions. The system represents a new approach to protein structure determination and has the potential to greatly reduce the time required to interpret electron-density maps in order to build accurate protein models.

Analysis of diversity of nucleotide and amino acid distributions in the VD and DJ joining regions in Ig heavy chains.

Nucleotide fill-in between the germ line V, D and J genes in the H3 loop of immunoglobulins contributes to the diversity of the antibody repertoire. This fill-in process is mediated by terminal deoxynucleotidyl transferase (TdT), which has been widely believed to insert nucleotides in a random fashion. Using a database of 2443 immunoglobulin sequences, we identified the regions of nucleotide fill-in between the V-D and D-J gene regions. We translated the fill-in nucleotides and measured the diversity within the two regions both at the nucleotide and amino acid level. We found that the nucleotide and amino acid distributions that resulted from nucleotide fill-in were in fact not random. Examination of the synonymous substitution rates of nucleotides revealed evidence suggesting that TdT plays a less significant role in generating antibody diversity than previously thought. We observed preferences for polar residues, which are more likely to encourage interaction with ligand than non-polar residues and are often found in loop regions in general. We also observed a preference for the insertion of smaller residues versus larger residues of similar biochemical properties, aiding in loop flexibility. We interpret these findings to reflect the significant influence of biochemical (i.e. folding) constraints and/or binding affinity constraints (at the cellular/selectional level) on the sequence diversity in the H3 region. These constraints act as a filter on the randomness generated by nucleotide addition by TdT, as well as other diversity generating processes such as recombination of VDJ gene segments and somatic mutation. The results of this study suggest that the antibody repertoire might be reduced from what is traditionally believed.

Conservation of cys-cys trp structural triads and their geometry in the protein domains of immunoglobulin superfamily members.

In almost all members of the immunoglobulin superfamily (IgSF) for which an experimental structure has been determined, a triad (C-CW) consisting of two cysteine residues that form a disulfide bond and a neighboring tryptophan can be found in the core of the protein fold. We analyzed the geometry of these C-CW triads among a database of 60 Fab crystal structures and found it to be remarkably conserved. We identified C-CW triads of a similar configuration in other members of the IgSF such as T cell receptor (TCR), major histocompatibility complex antigens (MHC), cell surface antigens CD4 and CD8, and cell-adhesion molecules. We used this C-CW pattern to search a database of non-IgSF proteins, and identified several proteins that contain a disulfide bridge associated with a tryptophan in a similar configuration. Examination of the distances and orientations between triads found in adjacent domains in Fab fragments and TCR also reveal a high degree of conservation, which reflects the invariance of the inter-chain domain packing. This high degree of conservation of the geometry of the C-CW triad in IgSF structures suggests that the Trp may contribute significantly to the stability of the disulfide bond. Knowledge of these geometric parameters may prove useful in the construction and validation of theoretical models of Ig, TCR, and other IgSF members.

TEXTAL: a pattern recognition system for interpreting electron density maps.

X-ray crystallography is the most widely used method for determining the three-dimensional structures of proteins and other macromolecules. One of the most difficult steps in crystallography is interpreting the electron density map to build the final model. This is often done manually by crystallographers and is very time-consuming and error-prone. In this paper, we introduce a new automated system called TEXTAL for interpreting electron density maps using pattern recognition. Given a map to be modeled, TEXTAL divides the map into small regions and then finds regions with a similar pattern of density in a database of maps for proteins whose structures have already been solved. When a match is found, the coordinates of atoms in the region are inferred by analogy. The key to making the database lookup efficient is to extract numeric features that represent the patterns in each region and to compare feature values using a weighted Euclidean distance metric. It is crucial that the features be rotation-invariant, since regions with similar patterns of density can be oriented in any arbitrary way. This pattern-recognition approach can take advantage of data accumulated in large crystallographic databases to effectively learn the association between electron density and molecular structure by example.

The context-dependence of amino acid properties.

One of the current limitations of using sequence alignments to identify proteins with similar structures is that some proteins with similar structures do not have significant sequence similarity by identity. One way to address this "hidden-homology" problem is to match amino acids based on their chemical and physical properties. However, the amino acid properties overlap, creating orthogonal dimensions of similarity, the relative strengths of which are ambiguous. It has been observed that the role an amino acid plays (and hence the property that is important) at a site in a protein depends on its secondary and tertiary environment. To approximate and take advantage of this dependence on context for improving the sensitivity of alignments of proteins whose structures are unknown, we propose a surrogate definition of context based on the pattern of hydropathy in a small window of contiguous neighbors surrounding each amino acid. We present the results of an experiment in which a search-based program iteratively tests and selects various properties in independent contexts, and incrementally increases the ability of sequence alignments to detect relationships among distantly-related proteins. The method is shown to perform better than using the MDM78 substitution table for partial match scores.

Constructive induction and protein tertiary structure prediction.

To date, the only methods that have been used successfully to predict protein structures have been based on identifying homologous proteins whose structures are known. However, such methods are limited by the fact that some proteins have similar structure but no significant sequence homology. We consider two ways of applying machine learning to facilitate protein structure prediction. We argue that a straightforward approach will not be able to improve the accuracy of classification achieved by clustering by alignment scores alone. In contrast, we present a novel constructive induction approach that learns better representations of amino acid sequences in terms of physical and chemical properties. Our learning method combines knowledge and search to shift the representation of sequences so that semantic similarity is more easily recognized by syntactic matching. Our approach promises not only to find new structural relationships among protein sequences, but also expands our understanding of the roles knowledge can play in learning via experience in this challenging domain.

Polymorphism at the self-incompatibility locus in Solanaceae predates speciation.

Sequences of 11 alleles of the gametophytic self-incompatibility locus (S locus) from three species of the Solanaceae family have recently been determined. Pairwise comparisons of these alleles reveal two unexpected observations: (i) amino acid sequence similarity can be as low as 40% within species and (ii) some interspecific similarities are higher than intraspecific similarities. The gene genealogy clearly illustrates this unusual pattern of relationships. The data suggest that some of the polymorphism at the S locus existed prior to the divergence of these species and has been maintained to the present. In support of this hypothesis, the number of shared polymorphic sites was found to exceed the number found in simulations with independent accumulation of mutations. Strictly neutral evolution is exceedingly unlikely to maintain the polymorphism for such a long time. The allele multiplicity and extreme age of the alleles is consistent with Wright's classic one-locus population genetic model of gametophytic self-incompatibility. Similarities between the plant S locus and the mammalian major histocompatibility complex are discussed.