The cost of the circadian desynchrony on the Leydig cell function.

The increased frequency of different lifestyles that disrupts circadian rhythms, together with a trend in the accretion of male idiopathic infertility, imposes the necessity to understand the contribution of circadian rhythms disruption to fertility regulation. In this study, the effects of circadian desynchrony (CD) on the steroidogenic capacity of adult Leydig cells were studied. Adult rats were housed under a disturbing light regime (2 days of constant light, 2 days of continual dark, and 3 days of 12:12 h light:dark schedule) designed to mimic shiftwork in humans. CD was characterized by changed and decreased rhythmic locomotor activity and reduced blood testosterone. In the Leydig cells changed transcription of the clock genes (Bmal1, Clock, Cry1 and Reverba/b increased while Per1/2 reversed phase) was detected. This was followed by reduced transcription of genes (Star, Cyp11a1, and Hsd3b1/2) primarily involved in mitosteroidogenesis. In parallel, mitochondrial membrane potential (Δψi) and ATP production declined losing their characteristic oscillatory pattern. Also, the main markers of mitochondrial biogenesis (Ppargc1a, Nrf1, Tfam, Cytc), fusion (Mfn2), and mitophagy (Pink1 and Tfeb) were disturbed. Collectively, CD targets mitochondria in Leydig cells by reducing mitosteroidogenesis, mitoenergetics, and disturbing mitochondrial dynamics. These changes contribute to testosterone decline compromising androgen-dependent functions, including reproduction.

Trade-offs between effectiveness and cost in bifunctional enzyme circuit with concentration robustness.

A fundamental trade-off in biological systems is whether they consume resources to perform biological functions or save resources. Bacteria need to reliably and rapidly respond to input signals by using limited cellular resources. However, excessive resource consumption will become a burden for bacteria growth. To investigate the relationship between functional effectiveness and resource cost, we study the ubiquitous bifunctional enzyme circuit, which is robust to fluctuations in protein concentration and responds quickly to signal changes. We show that trade-off relationships exist between functional effectiveness and protein cost. Expressing more proteins of the circuit increases concentration robustness and response speed but affects bacterial growth. In particular, our study reveals a general relationship between free-energy dissipation rate, response speed, and concentration robustness. The dissipation of free energy plays an important role in the concentration robustness and response speed. High robustness can only be achieved with a large amount of free-energy consumption and protein cost. In addition, the noise of the output increases with increasing protein cost, while the noise of the response time decreases with increasing protein cost. We also calculate the trade-off relationships in the EnvZ-OmpR system and the nitrogen assimilation system, which both have the bifunctional enzyme. Similar results indicate that these relationships are mainly derived from the specific feature of the bifunctional enzyme circuits and are not relevant to the details of the models. According to the trade-off relationships, bacteria take a compromise solution that reliably performs biological functions at a reasonable cost.

Contribution to Carbapenem Resistance and Fitness Cost of DcuS/DcuR, RcsC/RcsB, and YehU/YehT Two-Component Systems in CTX-M-15-Producing Escherichia coli.

Alteration in two-component systems (TCSs), which are signal transduction pathways in prokaryotes, can result in antibiotic resistance. Recently, it has been shown that the overexpression, using a multicopy cloning vector, of the dcuR, rcsB, and yehT genes, which code for the response regulator (RR) part of TCSs, enhanced the minimal inhibitory concentrations (MICs) of carbapenems in Escherichia coli K-12 derivative KAM3. Herein, the contribution to carbapenem resistance of the DcuS/DcuR, RcsC/RcsB, and YehU/YehT TCSs was assessed in E. coli K-12 derivative BW25113 (A phylogroup) and 536 (B2 phylogroup) recipient strains in combination with extended-spectrum ß-lactamase that exhibit a weak carbapenemase activity. The genes encoding both the sensor kinase (SK) and the RR, on the one hand, and the genes encoding the SK only, on the other hand, of these regulating pathways were disrupted. Subsequently, the mutants and their parental strains were transformed by a recombinant plasmid encoding the CTX-M-15 gene, before testing their susceptibility to carbapenems and their fitness. Results showed a trade-off between enhanced MICs for ertapenem, which remained above the clinical resistance breakpoint, and decreased growth rate, specifically for the 536 strain SK mutants. In conclusion, mutations in dcuS/dcuR, rcsC/rcsB, and yehU/yehT genes may be a pivotal first-step event in the development of carbapenem resistance.

At the Intersection of Chemistry, Biology, and Medicine.

After an undergraduate degree in biology at Harvard, I started graduate school at The Rockefeller Institute for Medical Research in New York City in July 1965. I was attracted to the chemical side of biochemistry and joined Fritz Lipmann's large, hierarchical laboratory to study enzyme mechanisms. That work led to postdoctoral research with Robert Abeles at Brandeis, then a center of what, 30 years later, would be called chemical biology. I spent 15 years on the Massachusetts Institute of Technology faculty, in both the Chemistry and Biology Departments, and then 26 years on the Harvard Medical School Faculty. My research interests have been at the intersection of chemistry, biology, and medicine. One unanticipated major focus has been investigating the chemical logic and enzymatic machinery of natural product biosynthesis, including antibiotics and antitumor agents. In this postgenomic era it is now recognized that there may be from 105 to 106 biosynthetic gene clusters as yet uncharacterized for potential new therapeutic agents.

A general method for rapid and cost-efficient large-scale production of 5' capped RNA.

The eukaryotic mRNA 5' cap structure is indispensible for pre-mRNA processing, mRNA export, translation initiation, and mRNA stability. Despite this importance, structural and biophysical studies that involve capped RNA are challenging and rare due to the lack of a general method to prepare mRNA in sufficient quantities. Here, we show that the vaccinia capping enzyme can be used to produce capped RNA in the amounts that are required for large-scale structural studies. We have therefore designed an efficient expression and purification protocol for the vaccinia capping enzyme. Using this approach, the reaction scale can be increased in a cost-efficient manner, where the yields of the capped RNA solely depend on the amount of available uncapped RNA target. Using a large number of RNA substrates, we show that the efficiency of the capping reaction is largely independent of the sequence, length, and secondary structure of the RNA, which makes our approach generally applicable. We demonstrate that the capped RNA can be directly used for quantitative biophysical studies, including fluorescence anisotropy and high-resolution NMR spectroscopy. In combination with (13)C-methyl-labeled S-adenosyl methionine, the methyl groups in the RNA can be labeled for methyl TROSY NMR spectroscopy. Finally, we show that our approach can produce both cap-0 and cap-1 RNA in high amounts. In summary, we here introduce a general and straightforward method that opens new means for structural and functional studies of proteins and enzymes in complex with capped RNA.

In silico screening of antifolate based novel inhibitors from Brucea mollis Wall. ex kurz against quadruple mutant drug resistant PfDHFR.

Plasmodium falciparum is the most lethal form of the genus Plasmodium which causes malaria, a 'disease of antiquity'. Globally it affects the health and socio-economic development of a large population especially in Sub-Saharan Africa and Southeast Asia. The Plasmodium falciparum dihydrofolate reductase-thymidylate synthase (PfDHFR-TS) is an important target of antimalarial drugs. Mutations at the active site of PfDHFR have resulted in decrease drug binding affinity of DHFR-inhibitors. In the present study we selected ten compounds of Brucea mollis Wall. Ex kurz and checked for their drug likeness using various computational tools and potential interactions with PfDHFR by molecular docking study. Soulameanone, a quassinoid of Brucea mollis Wall. Ex kurz showed better binding affinity when compared to pyrimethamine for both wild and quadruple mutant drug resistant PfDHFR. In addition, similar isomers of soulameanone were screened for their drug likeness and to study their interactions with PfDHFR. Twenty three compounds showed better binding affinity compared to soulameanone.

Assessment of BAK1 activity in different plant receptor-like kinase complexes by quantitative profiling of phosphorylation patterns.

Plant receptor-like kinases (RLKs) constitute a large family of receptors coordinating developmental programs with adaptation to environmental stresses including immune defenses. BRI1-ASSOCIATED KINASE 1 (BAK1), a member of the plant RLK family, forms receptor complexes with multiple RLK proteins including BRI1, FLS2, EFR and BIK1 to regulate responses to growth hormones or PAMPs. RLK activation and signal initiation involve protein complex formation and phosphorylation/dephosphorylation between BAK1 and its interacting partners. To gain new insight into how phosphorylation contributes to BAK1-mediated signaling specificity, we first mapped the phosphorylation patterns of BAK1 associated with different RLK partners (BRI1, FLS2, EFR and BIK1). Quantitative phospho-pattern profiling by label-free mass spectrometry revealed that differential phosphorylation patterns of RLK partners resulted from altered BAK1 phosphorylation status. More interestingly, the study of two BAK1 mutants (T450A and C408Y) both showing severe defect in immune defense yet normal growth phenotype suggested that varied phosphorylation patterns of RLK partners by BAK1 could be the molecular basis for selective regulation of multiple BAK1-dependent pathways. Taken together, this phospho-pattern profiling strategy allowed for explicit assessment of BAK1 kinase activity in different RLK complexes, which would facilitate elucidation of BAK1 diverse functions in plant development, defense, and adaptation. BIOLOGICAL SIGNIFICANCE: BAK1 is a functionally important co-receptor known to interact with different receptor-like kinases (RLKs) to coordinate plant development and immune defenses. Our study first mapped the phosphorylation patterns of BAK1 associated with four RLK partners (BRI1, FLS2, EFR and BIK1), and further revealed that differential phosphorylation patterns of multiple RLK partners resulted from altered BAK1 phosphorylation status. More interestingly, the study of two BAK1 mutants suggested that varied phosphorylation patterns of RLK partners by BAK1 could be the basis for selective regulation of signaling pathways. Taken together, this phospho-pattern profiling strategy allowed for explicit assessment of BAK1 kinase activity in different RLK complexes, which would facilitate elucidation of BAK1 diverse functions in plant development, defense, and adaptation.

Type I pyridoxal 5'-phosphate dependent enzymatic domains embedded within multimodular nonribosomal peptide synthetase and polyketide synthase assembly lines.

BACKGROUND: Pyridoxal 5'-phosphate (PLP)-dependent enzymes of fold type I, the most studied structural class of the PLP-dependent enzyme superfamily, are known to exist as stand-alone homodimers or homotetramers. These enzymes have been found also embedded in multimodular and multidomain assembly lines involved in the biosynthesis of polyketides (PKS) and nonribosomal peptides (NRPS). The aim of this work is to provide a proteome-wide view of the distribution and characteristics of type I domains covalently integrated in these assemblies in prokaryotes. RESULTS: An ad-hoc Hidden Markov profile was calculated using a sequence alignment derived from a multiple structural superposition of distantly related PLP-enzymes of fold type I. The profile was utilized to scan the sequence databank and to collect the proteins containing at least one type I domain linked to a component of an assembly line in bacterial genomes. The domains adjacent to a carrier protein were further investigated. Phylogenetic analysis suggested the presence of four PLP-dependent families: Aminotran_3, Beta_elim_lyase and Pyridoxal_deC, occurring mainly within mixed NRPS/PKS clusters, and Aminotran_1_2 found mainly in PKS clusters. Sequence similarity to the reference PLP enzymes with solved structures ranged from 24 to 42% identity. Homology models were built for each representative type I domain and molecular docking simulations with putative substrates were carried out. Prediction of the protein-protein interaction sites evidenced that the surface regions of the type I domains embedded within multienzyme assemblies were different from those of the self-standing enzymes; these structural features appear to be required for productive interactions with the adjacent domains in a multidomain context. CONCLUSIONS: This work provides a systematic view of the occurrence of type I domain within NRPS and PKS assembly lines and it predicts their structural characteristics using computational methods. Comparison with the corresponding stand-alone enzymes highlighted the common and different traits related to various aspects of their structure-function relationship. Therefore, the results of this work, on one hand contribute to the understanding of the functional and structural diversity of the PLP-dependent type I enzymes and, on the other, pave the way to further studies aimed at their applications in combinatorial biosynthesis.

Comparison of reversible-jump Markov-chain-Monte-Carlo learning approach with other methods for missing enzyme identification.

Computational identification of missing enzymes plays a significant role in accurate and complete reconstruction of metabolic network for both newly sequenced and well-studied organisms. For a metabolic reaction, given a set of candidate enzymes identified according to certain biological evidences, a powerful mathematical model is required to predict the actual enzyme(s) catalyzing the reactions. In this study, several plausible predictive methods are considered for the classification problem in missing enzyme identification, and comparisons are performed with an aim to identify a method with better performance than the Bayesian model used in previous work. In particular, a regression model consisting of a linear term and a nonlinear term is proposed to apply to the problem, in which the reversible jump Markov-chain-Monte-Carlo (MCMC) learning technique (developed in [Andrieu C, Freitas Nando de, Doucet A. Robust full Bayesian learning for radial basis networks 2001;13:2359-407.]) is adopted to estimate the model order and the parameters. We evaluated the models using known reactions in Escherichia coli, Mycobacterium tuberculosis, Vibrio cholerae and Caulobacter cresentus bacteria, as well as one eukaryotic organism, Saccharomyces Cerevisiae. Although support vector regression also exhibits comparable performance in this application, it was demonstrated that the proposed model achieves favorable prediction performance, particularly sensitivity, compared with the Bayesian method.

Enhanced polyamine catabolism alters homeostatic control of white adipose tissue mass, energy expenditure, and glucose metabolism.

Peroxisome proliferator-activated receptor gamma coactivator 1 alpha (PGC-1 alpha) is an attractive candidate gene for type 2 diabetes, as genes of the oxidative phosphorylation (OXPHOS) pathway are coordinatively downregulated by reduced expression of PGC-1 alpha in skeletal muscle and adipose tissue of patients with type 2 diabetes. Here we demonstrate that transgenic mice with activated polyamine catabolism due to overexpression of spermidine/spermine N(1)-acetyltransferase (SSAT) had reduced white adipose tissue (WAT) mass, high basal metabolic rate, improved glucose tolerance, high insulin sensitivity, and enhanced expression of the OXPHOS genes, coordinated by increased levels of PGC-1 alpha and 5'-AMP-activated protein kinase (AMPK) in WAT. As accelerated polyamine flux caused by SSAT overexpression depleted the ATP pool in adipocytes of SSAT mice and N(1),N(11)-diethylnorspermine-treated wild-type fetal fibroblasts, we propose that low ATP levels lead to the induction of AMPK, which in turn activates PGC-1 alpha in WAT of SSAT mice. Our hypothesis is supported by the finding that the phenotype of SSAT mice was reversed when the accelerated polyamine flux was reduced by the inhibition of polyamine biosynthesis in WAT. The involvement of polyamine catabolism in the regulation of energy and glucose metabolism may offer a novel target for drug development for obesity and type 2 diabetes.

c-Cbl-deficient mice have reduced adiposity, higher energy expenditure, and improved peripheral insulin action.

Casitas b-lineage lymphoma (c-Cbl) is an E3 ubiquitin ligase that has an important role in regulating the degradation of cell surface receptors. In the present study we have examined the role of c-Cbl in whole-body energy homeostasis. c-Cbl-/- mice exhibited a profound increase in whole-body energy expenditure as determined by increased core temperature and whole-body oxygen consumption. As a consequence, these mice displayed a decrease in adiposity, primarily due to a reduction in cell size despite an increase in food intake. These changes were accompanied by a significant increase in activity (2- to 3-fold). In addition, c-Cbl-/- mice displayed a marked improvement in whole-body insulin action, primarily due to changes in muscle metabolism. We observed increased protein levels of the insulin receptor (4-fold) and uncoupling protein-3 (2-fold) in skeletal muscle and a significant increase in the phosphorylation of AMP-activated protein kinase and acetyl-CoA carboxylase. These findings suggest that c-Cbl plays an integral role in whole-body fuel homeostasis by regulating whole-body energy expenditure and insulin action.

Protein complex prediction via cost-based clustering.

MOTIVATION: Understanding principles of cellular organization and function can be enhanced if we detect known and predict still undiscovered protein complexes within the cell's protein-protein interaction (PPI) network. Such predictions may be used as an inexpensive tool to direct biological experiments. The increasing amount of available PPI data necessitates an accurate and scalable approach to protein complex identification. RESULTS: We have developed the Restricted Neighborhood Search Clustering Algorithm (RNSC) to efficiently partition networks into clusters using a cost function. We applied this cost-based clustering algorithm to PPI networks of Saccharomyces cerevisiae, Drosophila melanogaster and Caenorhabditis elegans to identify and predict protein complexes. We have determined functional and graph-theoretic properties of true protein complexes from the MIPS database. Based on these properties, we defined filters to distinguish between identified network clusters and true protein complexes. CONCLUSIONS: Our application of the cost-based clustering algorithm provides an accurate and scalable method of detecting and predicting protein complexes within a PPI network.

Assessment of impaired proteasomal function in a cellular model of polyglutamine diseases.

A protein marked for degradation by the ubiquitin-proteasome pathway (UPP) is attached to multiple molecules of ubiquitin, a 76-amino-acid protein that targets the protein for rapid hydrolysis by 26S proteasome. Impaired function of UPP results in accumulation of misfolded and ubiquitinated proteins and has been implicated in the pathogenesis of various neurodegenerative diseases, including polyglutamine diseases. Impaired function of UPP can be evaluated either by assaying the proteasome's protease activity or the accumulation of ubiquitinated proteins.

A hierarchical mixture of Markov models for finding biologically active metabolic paths using gene expression and protein classes.

With the recent development of experimental high-throughput techniques, the type and volume of accumulating biological data have extremely increased these few years. Mining from different types of data might lead us to find new biological insights. We present a new methodology for systematically combining three different datasets to find biologically active metabolic paths/patterns. This method consists of two steps: First it synthesizes metabolic paths from a given set of chemical reactions, which are already known and whose enzymes are co-expressed, in an efficient manner. It then represents the obtained metabolic paths in a more comprehensible way through estimating parameters of a probabilistic model by using these synthesized paths. This model is built upon an assumption that an entire set of chemical reactions corresponds to a Markov state transition diagram. Furthermore, this model is a hierarchical latent variable model, containing a set of protein classes as a latent variable, for clustering input paths in terms of existing knowledge of protein classes. We tested the performance of our method using a main pathway of glycolysis, and found that our method achieved higher predictive performance for the issue of classifying gene expressions than those obtained by other unsupervised methods. We further analyzed the estimated parameters of our probabilistic models, and found that biologically active paths were clustered into only two or three patterns for each expression experiment type, and each pattern suggested some new long-range relations in the glycolysis pathway.

Presentation assessment of minor histocompatibility antigens by predictive proteasomal cleavage analysis.

Minor histocompatibility peptides (mHps) derived from polymorphic segments of endogenous proteins are thought to be targets for graft-versus-host and graft-versus-leukemia reactions after HLA-identical stem cell transplantation. A great majority of antigenic peptides is generated by fragmentation of proteins in the course of proteasomal processing. An algorithm was recently developed to predict cleavage sites during proteasomal processing. We tested the accuracy of the algorithm to predict mHps using 18 amino acid (AA) sequences of minor histocompatibility antigens (mHags) encoded by autosomal genes representing single nucleotide polymorphisms or by Y-chromosomal genes. The algorithm correctly predicted the C-termini of 11 of 13 experimentally confirmed mHps: 1) Correct prediction of C- and N-termini, e.g., for HA-1(H); 2) Correct prediction of C- and N-termini while anticipating intra-epitope cleavage sites, e.g., for SMCY-A*0201; 3) Correct prediction of C-termini and N-terminal extensions, e.g., for HA-8(R/V); and 4) Correct prediction of C-termini and N-terminal extensions while anticipating intra-epitope cleavage sites, e.g., for UTY-B8. Analysis of experimentally unconfirmed allelic counterparts of four autosomal mHags showed that AA substitutions either led to the insertion of an epitope-destroying cleavage site (e.g., in HA-1(R)) or abolished the correct C-terminus (e.g., in HA-2(M)). The proteasomal processing algorithm provides reliable data on the generation of mHps and forecasts their presence or absence. Combined with MHC class I ligand prediction, it can be a useful tool for the prediction of generation and presentation of new CTL epitopes derived from minor histocompatibility antigens.

Genome-based analysis of biosynthetic aminotransferase genes of Corynebacterium glutamicum.

Due to broad and overlapping substrate specificities, aminotransferases remain the last uncharacterized enzymes from most amino acid biosynthetic pathways in Corynebacterium glutamicum. We report here a complete description of all aminotransferases participating in the biosynthesis of the branched-chain amino acids and phenylalanine in C. glutamicum. We used methods of profile analysis on the newly available genome sequence to systematically search for and characterize members of the four known aminotransferase classes. This led to the discovery of sixteen new, potential aminotransferase encoding genes in the C. glutamicum genome, eleven of which were subsequently characterized experimentally with respect to their participation in different amino acid biosynthetic pathways. Disruption by insertion mutagenesis of ilvE, encoding a branched-chain amino acid aminotransferase, confirmed its function in leucine and isoleucine biosynthesis. Two double mutants lacking both ilvE and genes classified as class I aminotransferases exhibited additional auxotrophic requirements for valine and phenylalanine, respectively. In C. glutamicum the branched-chain amino acid aminotransferase thus participates in four amino acid biosynthetic pathways, for which in case of valine and phenylalanine biosynthesis two additional enzymes with overlapping substrate specificity exist. The novel protein with aminotransferase activity in valine biosynthesis belongs to the very recently described MocR subfamily of GntR-type helix-turn-helix transcriptional regulators, is located upstream of a potential operon of a newly described pyridoxine biosynthetic pathway and when disrupted, gives rise to a pyridoxine auxotrophy. The theoretical and experimental data we present should further provide a solid platform for ongoing research and understanding of the network of aminotransferases which participate in amino acid biosynthesis in C. glutamicum.

Assessment of proteasome activity in cell lysates and tissue homogenates using peptide substrates.

The ubiquitin-proteasome pathway is a major route of degradation of cell proteins. It also plays an essential role in maintaining cell homeostasis by degrading many rate-limiting enzymes and critical regulatory proteins. Alterations in proteasome activity have been implicated in a number of pathologies including Parkinson's disease, Alzheimer's disease and diabetes. The eukaryotic proteasome is a multicatalytic protease characterized by three activities with distinct specificities against peptide substrates. Although substrates were identified which could selectively measure the individual activities in the purified proteasome little data is available on how specific those substrates are for proteasomal activity when used with biological samples which may contain many other active peptidases. Here we examine the three major peptidase activities in lysates of two cell types and in a liver cytosol fraction in the presence of specific proteasome inhibitors and after fractionation by gel permeation chromatography. We demonstrate that other proteinases present in these preparations can degrade the commonly used proteasome substrates under the standard assay conditions. We develop a simple method for separating the proteasome from the lower molecular weight proteases using a 500kDa molecular weight cut-off membrane. This allows proteasome activity to be accurately measured in crude biological samples and may have quite broad applicability. We also identify low molecular weight tryptic activity in both the cell and tissue preparations which could not be inhibited by the proteasome inhibitor epoxomycin but was inhibitable by two cysteine proteinase inhibitors and by lactacystin suggesting that lactacystin may not be completely proteasome specific.

Assessment of proteasomal cleavage probabilities from kinetic analysis of time-dependent product formation.

Proteasomes are multicatalytic cellular protease complexes that degrade intracellular proteins into smaller peptides. Proteasomal in vitro digests have revealed that the various peptide bonds of a given substrate are cleaved in a highly selective manner. Regarding the key role of proteasomes as the main supplier of antigenic peptides for MHC class I-mediated antigen presentation, it is important to know to what extent these preferences for specific peptide bonds may vary among proteasomes of different cellular origin and of different subunit composition. Here, we quantify such cleavage rates by means of a kinetic proteasome model that relates the time-dependent changes of the amount of any generated peptide to the rates with which this peptide can be either generated from longer precursor peptides or degraded into smaller successor peptides. Numerical values for these rates are estimated by minimizing the distance between simulated and measured time-courses. The proposed method is applied to kinetic data obtained by combining HPLC fractionation and mass spectrometry (MS) to trace the degradation of two model peptides (pp89-25mer and LLO-27mer) by either the constitutive (T2) or immunoproteasome (T2.27). To convert the intensity of the MS signals into the respective peptide amounts, we use two methods leading to similar results: experimental calibration curves and theoretically determined linear scaling functions based on a novel approach using mass conservation rules. Comparison of the cleavage probabilities and procession rates obtained for the two types of proteasomes reveals that the striking differences between the time-dependent peptide profiles can be accounted for mainly by a generally higher turnover rate of the immunoproteasome. For the pp89-25mer, there is no significant change of the cleavage probabilities for any of the ten observed cleavage sites. For the LLO-27mer, there appears to be a significant change in the cleavage probabilities for four of the nine observed cleavage sites when switching from the constitutive to the immunoproteasome.

Large-scale isolation and crystallization of epothilone D from Myxococcus xanthus cultures.

The introduction of the epothilone polyketide synthase (PKS) into Myxococcus xanthus has enabled the heterologous production of epothilone D (1) on a large scale. To isolate this valuable product from the fermentation medium, an economical, scalable, and high-yielding purification process was developed. With the crystallization of 1 from a binary solvent system that consisted of ethanol and water, the product was recovered as white crystals with a final purity of > or =97% (w/w). This is the first reported crystallization of 1.

QSARs for 6-azasteroids as inhibitors of human type 1 5alpha-reductase: prediction of binding affinity and selectivity relative to 3-BHSD.

Quantitative structure-activity relationships (QSARs) are developed to describe the ability of 6-azasteroids to inhibit human type 1 5alpha-reductase. Models are generated using a set of 93 compounds with known binding affinities (K(i)) to 5alpha-reductase and 3beta-hydroxy-Delta(5)-steroid dehydrogenase/3-keto-Delta(5)-steroid isomerase (3-BHSD). QSARs are generated to predict K(i) values for inhibitors of 5alpha-reductase and to predict selectivity (S(i)) of compound binding to 3-BHSD relative to 5alpha-reductase. Log(K(i)) values range from -0.70 log units to 4.69 log units, and log(S(i)) values range from -3.00 log units to 3.84 log units. Topological, geometric, electronic, and polar surface descriptors are used to encode molecular structure. Information-rich subsets of descriptors are identified using evolutionary optimization procedures. Predictive models are generated using linear regression, computational neural networks (CNNs), principal components regression, and partial least squares. Compounds in an external prediction set are used for model validation. A 10-3-1 CNN is developed for prediction of binding affinity to 5alpha-reductase that produces root-mean-square error (RMSE) of 0.293 log units (R(2) = 0.97) for compounds in the external prediction set. Additionally, an 8-3-1 CNN is generated for prediction of inhibitor selectivity that produces RMSE = 0.513 log units (R(2) = 0.89) for the external prediction set. Models are further validated through Monte Carlo experiments in which models are generated after dependent variable values have been scrambled.