Comprehensive Review of Drug-Drug Interaction Prediction Based on Machine Learning: Current Status, Challenges, and Opportunities.

Detecting drug-drug interactions (DDIs) is an essential step in drug development and drug administration. Given the shortcomings of current experimental methods, the machine learning (ML) approach has become a reliable alternative, attracting extensive attention from the academic and industrial fields. With the rapid development of computational science and the growing popularity of cross-disciplinary research, a large number of DDI prediction studies based on ML methods have been published in recent years. To give an insight into the current situation and future direction of DDI prediction research, we systemically review these studies from three aspects: (1) the classic DDI databases, mainly including databases of drugs, side effects, and DDI information; (2) commonly used drug attributes, which focus on chemical, biological, and phenotypic attributes for representing drugs; (3) popular ML approaches, such as shallow learning-based, deep learning-based, recommender system-based, and knowledge graph-based methods for DDI detection. For each section, related studies are described, summarized, and compared, respectively. In the end, we conclude the research status of DDI prediction based on ML methods and point out the existing issues, future challenges, potential opportunities, and subsequent research direction.

Structural identification and comprehensive comparison of saponin-related impurities present in the three different compound glycyrrhizin tablets.

Compound Glycyrrhizin Tablet (CGT) is a glycyrrhizin-containing (monoammonium glycyrrhizate, MAG) preparation, which has been widely used in clinical treatment of chronic liver diseases, eczema, atopic dermatitis and other conditions. However, the impurity profile of CGT has not yet been completely elucidated. In this study, eight main saponin-related impurity compounds were initially isolated and identified. Thereafter, based on the characteristic MS/MS fragmentation pathways analysis of the isolated compounds, a novel strategy for characterization and identification of saponin-related impurities was proposed. Then, a total of 41 saponin-related impurities were identified or tentatively characterized in CGTs. Furthermore, principal component analysis (PCA), Wayne diagram and heatmap analysis revealed that the process-related impurity profile in CGTs from three different manufacturers was significantly different. Overall, our findings provided additional technological support for evaluating saponin-related impurities, thereby laying a solid foundation to develop strategies for future product quality improvement.

[Determination of absolute configuration of a new triterpenic acid in leaves of Ilex hainanensis].

This study aimed to explore the triterpenic acid components in leaves of Ilex hainanensis. Alkaline water extraction, macroporous resin adsorption, and high performance liquid chromatography were used to separate and purify the triterpenic acid components in leaves of I. hainanensis. The physical and chemical property analysis, MS, NMR spectroscopy, and literature comparison were performed to identify the structures, and a new triterpene acid compound was discovered:(3S, 4R, 5R, 8R, 9R, 10R, 14S, 17S, 18S, 19R)-3,19-dihydroxyursa-12,20(30)-diene-24,28-dioic-acid, and named ilexhainanin F. In addition, according to its structural characteristics, the ~(19)F-NMR Mosher method was further employed to study its absolute configuration. By comparison of the ~(19)F-NMR chemical shifts of Mosher esters, it was determined that the absolute configuration of the 3-position chiral center of the compound was the S configuration.

Evolution of mixing width induced by general Rayleigh-Taylor instability.

Turbulent mixing induced by Rayleigh-Taylor (RT) instability occurs ubiquitously in many natural phenomena and engineering applications. As the simplest and primary descriptor of the mixing process, the evolution of mixing width of the mixing zone plays a notable role in the flows. The flows generally involve complex varying acceleration histories and widely varying density ratios, two dominant factors affecting the evolution of mixing width. However, no satisfactory theory for predicting the evolution has yet been established. Here a theory determining the evolution of mixing width in general RT flows is established to reproduce, first, all of the documented experiments conducted for diverse (i.e., constant, impulsive, oscillating, decreasing, increasing, and complex) acceleration histories and all density ratios. The theory is established in terms of the conservation principle, with special consideration given to the asymmetry of the volume-averaged density fields occurring in actual flows. The results reveal the sensitivity or insensitivity of the evolution of a mixing front of a neighboring light or heavy fluid to the degree of asymmetry and thus explain the distinct evolutions in two experiments with the same configurations.

Glucuronoyl esterases are active on the polymeric substrate methyl esterified glucuronoxylan.

Alkali extracted beechwood glucuronoxylan methyl ester prepared by esterification of 4-O-methyl-D-glucuronic acid side residues by methanol was found to serve as substrate of microbial glucuronoyl esterases from Ruminococcus flavefaciens, Schizophyllum commune and Trichoderma reesei. The enzymatic deesterification was monitored by (1)H NMR spectroscopy and evaluated on the basis of the decrease of the signal of the ester methyl group and increase of the signal of methanol. The results show for the first time the action of enzymes on polymeric substrate, which imitates more closely the natural substrate in plant cell walls than the low molecular mass artificial substrates used up to present.

Hypoglycemic and anti-depressant effects of Zuogui Jiangtang Jieyu formulation in a model of unpredictable chronic mild stress in rats with diabetes mellitus.

This study aimed to investigate the hypoglycemic, lipid-lowering and antidepressant effects of Zuogui Jiangtang Jieyu formulation (ZGJTJY) in a model of unpredictable chronic mild stress (UCMS) in rats with diabetes mellitus (DM; the UCMS-DM model). Sixty rats were randomly divided into blank control, vehicle (model plus vehicle), positive control (model plus metformin and fluoxetine), and high, medium and low dose ZGJTJY (model plus high, medium and low doses of ZGJTJY, respectively) groups. Following establishment of DM by a high-fat diet with intraperitoneal injection of streptozotocin (38 mg/kg), the depression model was established by application of UCMS for 28 days. The behavioral scores of the rats were detected in an open field test and Morris water maze test. The levels of blood glucose, glycosylated hemoglobin (HbA1c) and blood lipids were assayed. The total scores of the open field test and the space exploration times (SETs) in the Morris water maze test were significantly lower and the escape latency (EL) times in the Morris water maze test were significantly longer in the vehicle group compared with those in the blank control group. In addition, in the vehicle group, the levels of blood glucose, HbA1c, total cholesterol (TC), triglycerides (TGs) and low-density lipoprotein cholesterol (LDL-C) were significantly higher and the levels of high-density lipoprotein cholesterol (HDL-C) were significantly lower compared with those in the blank control group. The high dose of ZGJTJY decreased the locomotor activity levels in the open field test, the EL times of the model on day 4, the SETs in the Morris water maze test and the HDL-C levels, and reduced the blood glucose, HbA1c, TC, TG and LDL-C levels compared with those in the model group. Thus, ZGJTJY is a potential candidate for the prevention and treatment of the comorbidity of depression with DM.

Trichoderma reesei CE16 acetyl esterase and its role in enzymatic degradation of acetylated hemicellulose.

BACKGROUND: Trichoderma reesei CE16 acetyl esterase (AcE) is a component of the plant cell wall degrading system of the fungus. The enzyme behaves as an exo-acting deacetylase removing acetyl groups from non-reducing end sugar residues. METHODS: In this work we demonstrate this exo-deacetylating activity on natural acetylated xylooligosaccharides using MALDI ToF MS. RESULTS: The combined action of GH10 xylanase and acetylxylan esterases (AcXEs) leads to formation of neutral and acidic xylooligosaccharides with a few resistant acetyl groups mainly at their non-reducing ends. We show here that these acetyl groups serve as targets for TrCE16 AcE. The most prominent target is the 3-O-acetyl group at the non-reducing terminal Xylp residues of linear neutral xylooligosaccharides or on aldouronic acids carrying MeGlcA at the non-reducing terminus. Deacetylation of the non-reducing end sugar may involve migration of acetyl groups to position 4, which also serves as substrate of the TrCE16 esterase. CONCLUSION: Concerted action of CtGH10 xylanase, an AcXE and TrCE16 AcE resulted in close to complete deacetylation of neutral xylooligosaccharides, whereas substitution with MeGlcA prevents removal of acetyl groups from only a small fraction of the aldouronic acids. Experiments with diacetyl derivatives of methyl ß-d-xylopyranoside confirmed that the best substrate of TrCE16 AcE is 3-O-acetylated Xylp residue followed by 4-O-acetylated Xylp residue with a free vicinal hydroxyl group. GENERAL SIGNIFICANCE: This study shows that CE16 acetyl esterases are crucial enzymes to achieve complete deacetylation and, consequently, complete the saccharification of acetylated xylans by xylanases, which is an important task of current biotechnology.

Mode of action of acetylxylan esterases on acetyl glucuronoxylan and acetylated oligosaccharides generated by a GH10 endoxylanase.

BACKGROUND: Substitutions on the xylan main chain are widely accepted to limit plant cell wall degradability and acetylations are considered as one of the most important obstacles. Hence, understanding the modes of action of a range of acetylxylan esterases (AcXEs) is of ample importance not only to increase the understanding of the enzymology of plant decay/bioremediation but also to enable efficient bioconversion of plant biomass. METHODS: In this study, the modes of action of acetylxylan esterases (AcXEs) belonging to carbohydrate esterase (CE) families 1, 4, 5 and 6 on xylooligosaccharides generated from hardwood acetyl glucuronoxylan were compared using MALDI ToF MS. Supporting data were obtained by following enzymatic deacetylation by (1)H NMR spectroscopy. CONCLUSIONS: None of the used enzymes were capable of complete deacetylation, except from linear xylooligosaccharides which were completely deacetylated by some of the esterases in the presence of endoxylanase. A clear difference was observed between the performance of the serine-type esterases of CE families 1, 5 and 6, and the aspartate-metalloesterases of family CE4. The difference is mainly due to the inability of CE4 AcXEs to catalyze deacetylation of 2,3-di-O-acetylated xylopyranosyl residues. Complete deacetylation of a hardwood acetyl glucuronoxylan requires additional deacetylating enzyme(s). GENERAL SIGNIFICANCE: The results contribute to the understanding of microbial degradation of plant biomass and outline the way to achieve complete saccharification of plant hemicelluloses which did not undergo alkaline pretreatment.

Mach-number-invariant mean-velocity profile of compressible turbulent boundary layers.

A series of Mach-number- (M) invariant scalings is derived for compressible turbulent boundary layers (CTBLs), leading to a viscosity weighted transformation for the mean-velocity profile that is superior to van Driest transformation. The theory is validated by direct numerical simulation of spatially developing CTBLs with M up to 6. A boundary layer edge is introduced to compare different M flows and is shown to better present the M-invariant multilayer structure of CTBLs. The new scalings derived from the kinetic energy balance substantiate Morkovin's hypothesis and promise accurate prediction of the mean profiles of CTBLs.

Structure of the catalytic domain of glucuronoyl esterase Cip2 from Hypocrea jecorina.

The structure of the catalytic domain of glucuronoyl esterase Cip2 from the fungus H. jecorina was determined at a resolution of 1.9 Å. This is the first structure of the newly established carbohydrate esterase family 15. The structure has revealed the residues Ser278-His411-Glu301 present in a triad arrangement as the active site. Ser278 is present in the novel consensus sequence GCSRXG reported earlier in the members of CE-15 family. The active site is exposed on the surface of the protein which has implications for the ability of the enzyme to hydrolyze ester bonds of large substrates. Efforts are underway to obtain crystals of Cip2_GE complexed with inhibitor and synthetic substrates. The activity of the glucuronoyl esterase could play a significant role in plant biomass degradation as its expected role is to separate the lignin from hemicelluloses by hydrolysis of the ester bond between 4-O-methyl-D-glucuronic acid moieties of glucuronoxylans and aromatic alcohols of lignin.

Action of xylan deacetylating enzymes on monoacetyl derivatives of 4-nitrophenyl glycosides of ß-D-xylopyranose and α-L-arabinofuranose.

Measurements of esterase activity by enzyme-coupled assays on monoacetates of 4-nitrophenyl ß-D-xylopyranoside and 4-nitrophenyl α-L-arabinofuranoside showed that acetylxylan esterases of families 1, 4 and 5 produced by Trichoderma reesei and Penicillium purpurogenum have a strong preference for deacetylation of position 2 in xylopyranosides. The acetylxylan esterases exhibit only weak activity on acetylated arabinofuranosides, with 2-acetate as the best substrate. Acetyl esterases of family 16 produced by the same two fungi deacetylate in xylopyranosides preferentially positions 3 and 4. Their specific activity on arabinofuranosides is also much lower than on xylopyranosides, however, substantially greater than that in the case of typical acetylxylan esterases.

Multilocus phylogenetic analyses, pullulan production and xylanase activity of tropical isolates of Aureobasidium pullulans.

Aureobasidium pullulans is the source of the commercially valuable polysaccharide pullulan and the enzyme xylanase. Isolates are typically off-white to pale pink or black on solid media, while some tropical isolates have been described as 'color variants' with bright pigments of red, yellow or purple. We sequenced 5 loci (internal transcribed spacer, intergenic spacer 1, translation elongation factor-1 alpha, beta tubulin, and RNA polymerase II) from 45 new isolates from Thailand. Based on the phylogenetic analyses, isolates were classified into 12 clades. Each clade showed different colors on different culture media including two clades with 'color variants' and some clades exhibited high levels of pullulan production or xylanase activity. Colony characteristics do not correlate perfectly with DNA sequence phylogeny or the physiological characters, but DNA sequence differences rapidly identify isolates with genetic novelty.

First example of highly stereoselective synthesis of 1,2,3-trisubstituted cyclopropanes via chiral selenonium ylides.

As the first example of the application of chiral selenonium ylides in asymmetric cyclopropanation, a new strategy for the highly stereoselective synthesis of chiral 1,2,3-trisubstituted cyclopropanes via selenonium ylides is described. The reaction afforded three stereoisomers of chiral 1,2,3-trisubstituted cyclopropanes with good yields, excellent diastereoselectivities, very high enantioselectivities (up to >99% ee), good generality and recyclability of chiral selenides. The possible pathways and models of the asymmetric cyclopropanation via the chiral selenonium ylides were also proposed to rationalize the excellent enantioselectivities and diastereoselectivities.

Novel family of carbohydrate esterases, based on identification of the Hypocrea jecorina acetyl esterase gene.

Plant cell walls have been shown to contain acetyl groups in hemicelluloses and pectin. The gene aes1, encoding the acetyl esterase (Aes1) of Hypocrea jecorina, was identified by amino-terminal sequencing, peptide mass spectrometry, and genomic sequence analyses. The coded polypeptide had 348 amino acid residues with the first 19 serving as a secretion signal peptide. The calculated molecular mass and isoelectric point of the secreted enzyme were 37,088 Da and pH 5.89, respectively. No significant homology was found between the predicated Aes1 and carbohydrate esterases of known families, but putative aes1 orthologs were found in genomes of many fungi and bacteria that produce cell wall-degrading enzymes. The aes1 transcript levels were high when the fungal cells were induced with sophorose, cellulose, oat spelt xylan, lactose, and arabinose. The recombinant Aes1 produced by H. jecorina transformed with aes1 under the cellobiohydrolase I promoter displayed properties similar to those reported for the native enzyme. The enzyme hydrolyzed acetate ester bond specifically. Using 4-nitrophenyl acetate as substrate, the activity of the recombinant enzyme was enhanced by D-xylose, D-glucose, cellobiose, D-galactose, and xylooligosaccharides but not by arabinose, mannose, or lactose. With the use of 4-nitrophenyl-beta-D-xylopyranoside monoacetate as substrate in a beta-xylosidase-coupled assay, Aes1 hydrolyzed positions 3 and 4 with the same efficiency while the H. jecorina acetylxylan esterase 1 exclusively deacetylated the position 2 acetyl group. Aes1 was capable of transacetylating methylxyloside in aqueous solution. The data presented demonstrate that Aes1 and other homologous microbial proteins may represent a new family of esterases for lignocellulose biodegradation.

Lycotoxin-1 insecticidal peptide optimized by amino acid scanning mutagenesis and expressed as a coproduct in an ethanologenic Saccharomyces cerevisiae strain.

New methods of safe biological pest control are required as a result of evolution of insect resistance to current biopesticides. Yeast strains being developed for conversion of cellulosic biomass to ethanol are potential host systems for expression of commercially valuable peptides, such as bioinsecticides, to increase the cost-effectiveness of the process. Spider venom is one of many potential sources of novel insect-specific peptide toxins. Libraries of mutants of the small amphipathic peptide lycotoxin-1 from the wolf spider were produced in high throughput using an automated integrated plasmid-based functional proteomic platform and screened for ability to kill fall armyworms, a significant cause of damage to corn (maize) and other crops in the United States. Using amino acid scanning mutagenesis (AASM) we generated a library of mutagenized lycotoxin-1 open reading frames (ORF) in a novel small ubiquitin-like modifier (SUMO) yeast expression system. The SUMO technology enhanced expression and improved generation of active lycotoxins. The mutants were engineered to be expressed at high level inside the yeast and ingested by the insect before being cleaved to the active form (so-called Trojan horse strategy). These yeast strains expressing mutant toxin ORFs were also carrying the xylose isomerase (XI) gene and were capable of aerobic growth on xylose. Yeast cultures expressing the peptide toxins were prepared and fed to armyworm larvae to identify the mutant toxins with greatest lethality. The most lethal mutations appeared to increase the ability of the toxin alpha-helix to interact with insect cell membranes or to increase its pore-forming ability, leading to cell lysis. The toxin peptides have potential as value-added coproducts to increase the cost-effectiveness of fuel ethanol bioproduction.

Enzyme characterization for hydrolysis of AFEX and liquid hot-water pretreated distillers' grains and their conversion to ethanol.

Dried distillers' grains with solubles (DDGS), a co-product of corn ethanol production, was investigated as a feedstock for additional ethanol production. DDGS was pretreated with liquid hot-water (LHW) and ammonia fiber explosion (AFEX) processes. Cellulose was readily converted to glucose from both LHW and AFEX treated DDGS using a mixture of commercial cellulase and beta-glucosidase; however, these enzymes were ineffective at saccharifying the xylan present in the pretreated DDGS. Several commercial enzyme preparations were evaluated in combination with cellulase to saccharify pretreated DDGS xylan and it was found that adding commercial grade (e.g. impure) pectinase and feruloyl esterase (FAE) preparations were effective at releasing arabinose and xylose. The response of sugar yields for pretreated AFEX and LHW DDGS (6wt%/solids) were determined for different enzyme loadings of FAE and pectinase and modeled as a response surfaces. Arabinose and xylose yields rose with increasing FAE and pectinase enzyme dosages for both pretreated materials. When hydrolyzed at 20wt%/solids with the same blend of commercial enzymes, the yields were 278 and 261g sugars (i.e. total of arabinose, xylose, and glucose) per kg of DDGS (dry basis, db) for AFEX and LHW pretreated DDGS, respectively. The pretreated DDGS's were also evaluated for fermentation using Saccharomyces cerevisiae at 15wt%/solids. Pretreated DDGS were readily fermented and were converted to ethanol at 89-90% efficiency based upon total glucans; S. cerevisiae does not ferment arabinose or xylose.

Identification of genes encoding microbial glucuronoyl esterases.

One type of covalent linkages connecting lignin and hemicellulose in plant cell walls is the ester linkage between 4-O-methyl-D-glucuronic acid of glucuronoxylan and lignin alcohols. An enzyme that could hydrolyze such linkages, named glucuronoyl esterase, occurs in the cellulolytic system of the wood-rotting fungus Schizophyllum commune. Here we report partial amino acid sequences of the enzyme and the results of subsequent search for homologous genes in sequenced genomes. The homologous genes of unknown functions were found in genomes of several filamentous fungi and one bacterium. The gene corresponding to the cip2 gene of Hypocrea jecorina (Trichoderma reesei), known to be up-regulated under conditions of induction of cellulolytic and hemicellulolytic enzymes, was over-expressed in H. jecorina. The product of the cip2 gene was purified to homogeneity and shown to exhibit glucuronoyl esterase activity.

Variation in relative substrate specificity of bifunctional beta-D-xylosidase/alpha-L-arabinofuranosidase by single-site mutations: roles of substrate distortion and recognition.

To probe differential control of substrate specificities for 4-nitrophenyl-alpha-l-arabinofuranoside (4NPA) and 4-nitrophenyl-beta-d-xylopyranoside (4NPX), residues of the glycone binding pocket of GH43 beta-d-xylosidase/alpha-l-arabinofuranosidase from Selenomonas ruminantium were individually mutated to alanine. Although their individual substrate specificities (kcat/Km)(4NPX) and (kcat/Km)(4NPA) are lowered 330 to 280,000 fold, D14A, D127A, W73A, E186A, and H248A mutations maintain similar relative substrate specificities as wild-type enzyme. Relative substrate specificities (kcat/Km)(4NPX)/(kcat/Km)(4NPA) are lowered by R290A, F31A, and F508A mutations to 0.134, 0.407, and 4.51, respectively, from the wild type value of 12.3 with losses in (kcat/Km)(4NPX) and (kcat/Km)(4NPA) of 18 to 163000 fold. R290 and F31 reside above and below the C4 OH group of 4NPX and the C5 OH group of 4NPA, where they can serve as anchors for the two glycone moieties when their ring systems are distorted to transition-state geometries by raising the position of C1. Thus, whereas R290 and F31 provide catalytic power for hydrolysis of both substrates, the native residues are more important for 4NPX than 4NPA as the xylopyranose ring must undergo greater distortion than the arabinofuranose ring. F508 borders C4 and C5 of the two glycone moieties and can serve as a hydrophobic platform having more favorable interactions with xylose than arabinofuranose.

Structure-function relationships of a catalytically efficient beta-D-xylosidase.

Beta-D-Xylosidase from Selenomonas ruminantium is revealed as the best catalyst known (kcat, kcat/Km) for promoting hydrolysis of 1,4-beta-D-xylooligosaccharides. 1H nuclear magnetic resonance experiments indicate the family 43 glycoside hydrolase acts through an inversion mechanism on substrates 4-nitrophenyl- beta-D-xylopyranoside (4NPX) and 1,4-beta-D-xylobiose (X2). Progress curves of 4-nitrophenyl-beta-D-xylobioside, xylotetraose and xylohexaose reactions indicate that one residue from the nonreducing end of substrate is cleaved per catalytic cycle without processivity. Values of kcat and kcat/Km decrease for xylooligosaccharides longer than X2, illustrating the importance to catalysis of subsites -1 and +1 and the lack there of subsite +2. Homology models of the enzyme active site with docked substrates show that subsites beyond -1 are blocked by protein and subsites beyond +1 are not formed; they suggest that D14 and E186 serve catalysis as general base and general acid, respectively. Individual mutations, D14A and E186A, erode kcat and kcat/Km by <103 and to a similar extent for substrates 4NPX and 4-nitrophenyl-alpha-L-arabinofuranoside (4NPA), indicating that the two substrates share the same active site. With 4NPX and 4NPA, pH governs kcat/Km with pKa values of 5.0 and 7.0 assigned to D14 and E186, respectively. kcat(4NPX) has a pKa value of 7.0 and kcat(4NPA) is pH independent above pH 4.0, suggesting that the catalytically inactive, "dianionic" enzyme form (D14-E187-) binds 4NPX but not 4NPA.

Expression of an AT-rich xylanase gene from the anaerobic fungus Orpinomyces sp. strain PC-2 in and secretion of the heterologous enzyme by Hypocrea jecorina.

The catalytic domain encoded by an adenine-thymine (AT)-rich xylanase gene (xynA) of the anaerobic fungus Orpinomyces was expressed in Hypocrea jecorina under the control of the cel7A promoter and terminator. No XynA protein was detected in H. jecorina culture supernatants when the original sequence was fused to the H. jecorina cel5A region coding for its signal peptide, carbohydrate-binding module, and hinge. Replacing the xynA (56% AT content) with a synthetic sequence containing lower AT content (39%) supported the extracellular production (150 mg l(-1)) of the fusion xylanase by H. jecorina. Northern analysis revealed that successful production after the decrease in AT content was related to higher levels of the xylanase-specific mRNA. Another construct with an RDKR-coding sequence inserted between the cel5A linker and the xynA catalytic domain allowed production of the fully processed active xylanase catalytic domain. Both the fusion (40 kDa) and the fully processed (28 kDa) forms displayed enzymatic properties of family 11 xylanases. Both the R and the Kex2-like KR sites were recognized during secretion, resulting in a mixture of two amino termini for the 28-kDa xylanase. The work demonstrated for the first time that glycoside hydrolases derived from anaerobic fungi can be produced by H. jecorina.