Accurate prediction of enzyme mutant activity based on a multibody statistical potential

MOTIVATION: An important area of research in biochemistry and molecular biology focuses on characterization of enzyme mutants. However, synthesis and analysis of experimental mutants is time consuming and expensive. We describe a machine-learning approach for inferring the activity levels of all unexplored single point mutants of an enzyme, based on a training set of such mutants with experimentally measured activity. RESULTS: Based on a Delaunay tessellation-derived four-body statistical potential function, a perturbation vector measuring environmental changes relative to wild type (wt) at every residue position uniquely characterizes each enzyme mutant for model development and prediction. First, a measure of model performance utilizing area (AUC) under the receiver operating characteristic (ROC) curve surpasses 0.83 and 0.77 for data sets of experimental HIV-1 protease and T4 lysozyme mutants, respectively. Additionally, a novel method is introduced for evaluating statistical significance associated with the number of correct test set predictions obtained from a trained model. Third, 100 stratified random splits of the protease and T4 lysozyme mutant data sets into training and test sets achieve 77.0% and 80.8% mean accuracy, respectively. Next, protease and T4 lysozyme models trained with experimental mutants are used to predict activity levels for all remaining mutants; a subsequent search for publications reporting on dozens of these test mutants reveals that experimental results are matched by 79% and 86% of predictions, respectively. Finally, learning curves for each mutant enzyme system indicate the influence of training set size on model performance. AVAILABILITY: Prediction databases at http://proteins.gmu.edu/automute/     

Increased plasma cholesterol esterification by LCAT reduces diet-induced atherosclerosis in SR-BI knockout mice

LCAT, a plasma enzyme that esterifies cholesterol, has been proposed to play an antiatherogenic role, but animal and epidemiologic studies have yielded conflicting results. To gain insight into LCAT and the role of free cholesterol (FC) in atherosclerosis, we examined the effect of LCAT over- and underexpression in diet-induced atherosclerosis in scavenger receptor class B member I-deficient [Scarab(−/−)] mice, which have a secondary defect in cholesterol esterification. Scarab(−/−)×LCAT-null [Lcat(−/−)] mice had a decrease in HDL-cholesterol and a high plasma ratio of FC/total cholesterol (TC) (0.88 ± 0.033) and a marked increase in VLDL-cholesterol (VLDL-C) on a high-fat diet. Scarab(−/−)×LCAT-transgenic (Tg) mice had lower levels of VLDL-C and a normal plasma FC/TC ratio (0.28 ± 0.005). Plasma from Scarab(−/−)×LCAT-Tg mice also showed an increase in cholesterol esterification during in vitro cholesterol efflux, but increased esterification did not appear to affect the overall rate of cholesterol efflux or hepatic uptake of cholesterol. Scarab(−/−)×LCAT-Tg mice also displayed a 51% decrease in aortic sinus atherosclerosis compared with Scarab(−/−) mice (P < 0.05). In summary, we demonstrate that increased cholesterol esterification by LCAT is atheroprotective, most likely through its ability to increase HDL levels and decrease pro-atherogenic apoB-containing lipoprotein particles.     

Cellular pregnenolone esterification by acyl-CoA:cholesterol acyltransferase

Pregnenolone (PREG) can be converted to PREG esters (PE) by the plasma enzyme lecithin: cholesterol acyltransferase (LCAT), and by other enzyme(s) with unknown identity. Acyl-CoA:cholesterol acyltransferase 1 and 2 (ACAT1 and ACAT2) convert various sterols to steryl esters; their activities are activated by cholesterol. PREG is a sterol-like molecule, with 3-β-hydroxy moiety at steroid ring A, but with much shorter side chain at steroid ring D. Here we show that without cholesterol, PREG is a poor ACAT substrate; with cholesterol, the V(max) for PREG esterification increases by 100-fold. The binding affinity of ACAT1 for PREG is 30-50-fold stronger than that for cholesterol; however, PREG is only a substrate but not an activator, while cholesterol is both a substrate and an activator. These results indicate that the sterol substrate site in ACAT1 does not involve significant sterol-phospholipid interaction, while the sterol activator site does. Studies utilizing small molecule ACAT inhibitors show that ACAT plays a key role in PREG esterification in various cell types examined. Mice lacking ACAT1 or ACAT2 do not have decreased PREG ester contents in adrenals, nor do they have altered levels of the three major secreted adrenal steroids in serum. Mice lacking LCAT have decreased levels of PREG esters in the adrenals. These results suggest LCAT along with ACAT1/ACAT2 contribute to control pregnenolone ester content in different cell types and tissues.     

Comprehensive mutagenesis of HIV-1 protease: a computational geometry approach

A computational geometry technique based on Delaunay tessellation of protein structure, represented by C(alpha) atoms, is used to study effects of single residue mutations on sequence-structure compatibility in HIV-1 protease. Profiles of residue scores derived from the four-body statistical potential are constructed for all 1881 mutants of the HIV-1 protease monomer and compared with the profile of the wild-type protein. The profiles for an isolated monomer of HIV-1 protease and the identical monomer in a dimeric state with an inhibitor are analyzed to elucidate changes to structural stability. Protease residues shown to undergo the greatest impact are those forming the dimer interface and flap region, as well as those known to be involved in inhibitor binding.     

Efficiency of extension of mismatched primer termini across from cisplatin and oxaliplatin adducts by human DNA polymerases beta and eta in vitro

DNA polymerases beta and eta are among the few eukaryotic polymerases known to efficiently bypass cisplatin and oxaliplatin adducts in vitro. Our laboratory has previously established that both polymerases misincorporated dTTP with high frequency across from cisplatin- and oxaliplatin-GG adducts. This decrease in polymerase fidelity on platinum-damaged DNA could lead to in vivo mutations, if this base substitution were efficiently elongated. In this study, we performed a steady-state kinetic analysis of the steps required for fixation of dTTP misinsertion during translesion synthesis past cisplatin- and oxaliplatin-GG adducts by pol beta and pol eta. The efficiency of translesion synthesis by pol eta past Pt-GG adducts was very similar to that observed for this polymerase when the template contains thymine-thymine dimers. This finding suggested that pol eta could play a role in translesion synthesis past platinum-GG adducts in vivo. On the other hand, translesion synthesis past platinum-GG adducts by pol beta was much less efficient. Translesion synthesis by pol eta is likely to be predominantly error-free, since the probability of correct insertion and extension by pol eta was 1000-2000-fold greater than the probability of incorrect insertion and extension. Our results also indicated that for pol eta the frequency of misincorporation is the same across from the 3'G and the 5'G of the platinum-GG adducts for both cisplatin and oxaliplatin adducts. On the other hand, pol beta is more likely to misinsert at the 3'G of the adducts and misinsertion occurs at higher frequency for oxaliplatin-GG than for cisplatin-GG adducts.     

Sequence context-dependent replication of DNA templates containing UV-induced lesions by human DNA polymerase iota

Humans possess four Y-family polymerases: pols eta, iota, kappa and the Rev1 protein. The pivotal role that pol eta plays in protecting us from UV-induced skin cancers is unquestioned given that mutations in the POLH gene (encoding pol eta), lead to the sunlight-sensitive and cancer-prone xeroderma pigmentosum variant phenotype. The roles that pols iota, kappa and Rev1 play in the tolerance of UV-induced DNA damage is, however, much less clear. For example, in vitro studies in which the ability of pol iota to bypass UV-induced cyclobutane pyrimidine dimers (CPDs) or 6-4 pyrimidine-pyrimidone (6-4PP) lesions has been assayed, are somewhat varied with results ranging from limited misinsertion opposite CPDs to complete lesion bypass. We have tested the hypothesis that such discrepancies might have arisen from different assay conditions and local sequence contexts surrounding each UV-photoproduct and find that pol iota can facilitate significant levels of unassisted highly error-prone bypass of a T-T CPD, particularly when the lesion is located in a 3'-A[T-T]A-5' template sequence context and the reaction buffer contains no KCl. When encountering a T-T 6-4PP dimer under the same assay conditions, pol iota efficiently and accurately inserts the correct base, A, opposite the 3'T of the 6-4PP by factors of approximately 10(2) over the incorporation of incorrect nucleotides, while incorporation opposite the 5'T is highly mutagenic. Pol kappa has been proposed to function in the bypass of UV-induced lesions by helping extend primers terminated opposite CPDs. However, we find no evidence that the combined actions of pol iota and pol kappa result in a significant increase in bypass of T-T CPDs when compared to pol iota alone. Our data suggest that under certain conditions and sequence contexts, pol iota can bypass T-T CPDs unassisted and can efficiently incorporate one or more bases opposite a T-T 6-4PP. Such biochemical activities may, therefore, be of biological significance especially in XP-V cells lacking the primary T-T CPD bypassing enzyme, pol eta.     

poliota-dependent lesion bypass in vitro

Based upon phylogenetic relationships, the broad Y-family of DNA polymerases can be divided into various subfamilies consisting of UmuC (polV)-like; DinB (polIV/polκ)-like; Rev1-like, Rad30A (polη)-like and Rad30B (polι)-like polymerases. The polIV/polκ-like polymerases are most ubiquitous, having been identified in bacteria, archaea and eukaryotes. In contrast, the polV-like polymerases appear restricted to bacteria (both Gram positive and Gram negative). Rev1 and polη-like polymerases are found exclusively in eukaryotes, and to date, polι-like polymerases have only been identified in higher eukaryotes. In general, the in vitro properties of polymerases characterized within each sub-family are quite similar. An exception to this rule occurs with the polι-like polymerases, where the enzymatic properties of Drosophila melanogaster polι are more similar to that of Saccharomyces cerevisiae and human polη than to the related human polι. For example, like polη, Drosophila polι can bypass a cis-syn thymine–thymine dimer both accurately and efficiently, while human polι bypasses the same lesion inefficiently and with low-fidelity. Even in cases where human polι can efficiently insert a base opposite a lesion (such as a synthetic abasic site, the 3′T of a 6-4-thymine–thymine pyrimidine–pyrimidone photoproduct or opposite benzo[a]pyrene diol epoxide deoxyadenosine adducts), further extension is often limited. Thus, although polι most likely arose from a genetic duplication of polη millions of years ago as eukaryotes evolved, it would appear that polι from humans (and possibly all mammals) has been further subjected to evolutionary pressures that have “tailored” its enzymatic properties away from lesion bypass and towards other function(s) specific for higher eukaryotes. The identification of such functions and the role that mammalian polι plays in lesion bypass in vivo, should hopefully be forthcoming with the construction of human cell lines deleted for polι and the identification of mice deficient in polι.     

Roles of ferritin and iron in ischemic preconditioning of the heart

Iron and copper play major roles in biological systems, catalyzing free radical production and consequently causing damage. The relatively high levels of these metals, which are mobilized into the coronary flow following prolonged ischemia, have been incriminated as key players in reperfusion injury to the heart. In the present communication we investigated other roles of iron – providing protection to the ischemic heart via preconditioning (PC).PC was accomplished by subjecting isolated rat hearts to three episodes of 2 min ischemia separated by 3 min of reperfusion. Prolonged ischemia followed the PC phase. PC hearts (group I) were compared to hearts subjected to normal perfusion (group II, no ischemia) and to ischemia without PC (group III). Group I showed a marked improvement in the recovery of hemodynamic function vs. group III. Biochemical parameters further substantiated the PC protection provided to group I against prolonged ischemia. Correspondingly, group I presented markedly lower re-distribution and mobilization of iron and copper into the coronary flow, following prolonged ischemia, as evinced from the decrease in total levels, and in the 'free' fraction of iron and copper.During the PC phase no loss of cardiac function was observed. A small wave of re-distribution and mobilization of iron (typically less than 4–8% of the value of 35 min ischemia) was recorded. The cellular content of ferritin (Ft) measured in the heart was significantly higher in group I than in group III (0.90 and 0.54 g/mg, respectively). Also, iron-saturation of Ft was significantly lower for PC hearts, compared to both groups II and III (0.22 vs. 0.32 and 0.31 g/mg, for 35 min ischemia, respectively). These findings are in accord with the proposal that intracellular re-distribution and mobilization of small levels of iron, during PC, cause rapid accumulation of ferritin – the major iron-storage protein.It is proposed that iron play a dual role: (i) It serves as a signaling pathway for the accumulation of Ft following the PC phase. This iron is not involved in cardiac injury, but rather prepares the heart against future high levels of 'free' iron, thus reducing the degree of myocardial damage after prolonged ischemia. (ii) High levels of iron (and copper) are mobilized following prolonged ischemia and cause tissue damage.     

Hepatic ABCG5 and ABCG8 overexpression increases hepatobiliary sterol transport but does not alter aortic atherosclerosis in transgenic mice

The individual roles of hepatic versus intestinal ABCG5 and ABCG8 in sterol transport have not yet been investigated. To determine the specific contribution of liver ABCG5/G8 to sterol transport and atherosclerosis, we generated transgenic mice that overexpress human ABCG5 and ABCG8 in the liver but not intestine (liver G5/G8-Tg) in three different genetic backgrounds: C57Bl/6, apoE-KO, and low density lipoprotein receptor (LDLr)-KO. Hepatic overexpression of ABCG5/G8 enhanced hepatobiliary secretion of cholesterol and plant sterols by 1.5-2-fold, increased the amount of intestinal cholesterol available for absorption and fecal excretion by up to 27%, and decreased the accumulation of plant sterols in plasma by approximately 25%. However, it did not alter fractional intestinal cholesterol absorption, fecal neutral sterol excretion, hepatic cholesterol concentrations, or hepatic cholesterol synthesis. Consequently, overexpression of ABCG5/G8 in only the liver had no effect on the plasma lipid profile, including cholesterol, HDL-C, and non-HDL-C, or on the development of proximal aortic atherosclerosis in C57Bl/6, apoE-KO, or LDLr-KO mice. Thus, liver ABCG5/G8 facilitate the secretion of liver sterols into bile and serve as an alternative mechanism, independent of intestinal ABCG5/G8, to protect against the accumulation of dietary plant sterols in plasma. However, in the absence of changes in fractional intestinal cholesterol absorption, increased secretion of sterols into bile induced by hepatic overexpression of ABCG5/G8 was not sufficient to alter hepatic cholesterol balance, enhance cholesterol removal from the body or to alter atherogenic risk in liver G5/G8-Tg mice. These findings demonstrate that overexpression of ABCG5/G8 in the liver profoundly alters hepatic but not intestinal sterol transport, identifying distinct roles for liver and intestinal ABCG5/G8 in modulating sterol metabolism.     

Changes in cytokine production and impaired hematopoiesis in patients with anorexia nervosa: the effect of refeeding

The changes in cytokines and hormones involved in hematopoiesis were studied in the serum of 7 girls with anorexia nervosa, 15-24 yr old, on admission and after 5% and 10% weight gain. Hematopoiesis was studied by in-vitro culturing of circulating granulocyte-macrophage colony forming cells and erythroid burst forming cells. Nutritional status was studied by anthropometric measurements and resting energy expenditure. On admission, granulocyte-macrophage colony forming cells and erythroid burst forming cells were significantly lower than in age-matched controls and increased significantly along weight gain. Blood leptin and erythropoietin levels increased significantly with weight gain. TNF-alpha levels tended to decrease while IL-1beta levels were lower than in the controls on admission (p <0.05) and did not change significantly during weight gain. IL-3, GM-CSF and IL-6 were undetected on admission or along weight gain. The changes in granulocyte-macrophage colony forming cells and erythroid burst forming cells positively correlated with changes in resting energy expenditure and fat free mass. These results may suggest that undernutrition affects hematopoiesis as indicated by the reduction of hematopoietic progenitor cells before treatment and the significant increase with weight gain. The changes in the levels of hormones and cytokines known to be involved in hematopoiesis along refeeding may suggest a role for these factors in anorexia nervosa.