Deep learning of a bacterial and archaeal universal language of life enables transfer learning and illuminates microbial dark matter.

The majority of microbial genomes have yet to be cultured, and most proteins identified in microbial genomes or environmental sequences cannot be functionally annotated. As a result, current computational approaches to describe microbial systems rely on incomplete reference databases that cannot adequately capture the functional diversity of the microbial tree of life, limiting our ability to model high-level features of biological sequences. Here we present LookingGlass, a deep learning model encoding contextually-aware, functionally and evolutionarily relevant representations of short DNA reads, that distinguishes reads of disparate function, homology, and environmental origin. We demonstrate the ability of LookingGlass to be fine-tuned via transfer learning to perform a range of diverse tasks: to identify novel oxidoreductases, to predict enzyme optimal temperature, and to recognize the reading frames of DNA sequence fragments. LookingGlass enables functionally relevant representations of otherwise unknown and unannotated sequences, shedding light on the microbial dark matter that dominates life on Earth.

mebipred: identifying metal-binding potential in protein sequence.

MOTIVATION: metal-binding proteins have a central role in maintaining life processes. Nearly one-third of known protein structures contain metal ions that are used for a variety of needs, such as catalysis, DNA/RNA binding, protein structure stability, etc. Identifying metal-binding proteins is thus crucial for understanding the mechanisms of cellular activity. However, experimental annotation of protein metal-binding potential is severely lacking, while computational techniques are often imprecise and of limited applicability. RESULTS: we developed a novel machine learning-based method, mebipred, for identifying metal-binding proteins from sequence-derived features. This method is over 80% accurate in recognizing proteins that bind metal ion-containing ligands; the specific identity of 11 ubiquitously present metal ions can also be annotated. mebipred is reference-free, i.e. no sequence alignments are involved, and is thus faster than alignment-based methods; it is also more accurate than other sequence-based prediction methods. Additionally, mebipred can identify protein metal-binding capabilities from short sequence stretches, e.g. translated sequencing reads, and, thus, may be useful for the annotation of metal requirements of metagenomic samples. We performed an analysis of available microbiome data and found that ocean, hot spring sediments and soil microbiomes use a more diverse set of metals than human host-related ones. For human microbiomes, physiological conditions explain the observed metal preferences. Similarly, subtle changes in ocean sample ion concentration affect the abundance of relevant metal-binding proteins. These results highlight mebipred's utility in analyzing microbiome metal requirements. AVAILABILITY AND IMPLEMENTATION: mebipred is available as a web server at services.bromberglab.org/mebipred and as a standalone package at https://pypi.org/project/mymetal/. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Computational predictors fail to identify amino acid substitution effects at rheostat positions.

Many computational approaches exist for predicting the effects of amino acid substitutions. Here, we considered whether the protein sequence position class - rheostat or toggle - affects these predictions. The classes are defined as follows: experimentally evaluated effects of amino acid substitutions at toggle positions are binary, while rheostat positions show progressive changes. For substitutions in the LacI protein, all evaluated methods failed two key expectations: toggle neutrals were incorrectly predicted as more non-neutral than rheostat non-neutrals, while toggle and rheostat neutrals were incorrectly predicted to be different. However, toggle non-neutrals were distinct from rheostat neutrals. Since many toggle positions are conserved, and most rheostats are not, predictors appear to annotate position conservation better than mutational effect. This finding can explain the well-known observation that predictors assign disproportionate weight to conservation, as well as the field's inability to improve predictor performance. Thus, building reliable predictors requires distinguishing between rheostat and toggle positions.

Building a genome analysis pipeline to predict disease risk and prevent disease.

Reduced costs and increased speed and accuracy of sequencing can bring the genome-based evaluation of individual disease risk to the bedside. While past efforts have identified a number of actionable mutations, the bulk of genetic risk remains hidden in sequence data. The biggest challenge facing genomic medicine today is the development of new techniques to predict the specifics of a given human phenome (set of all expressed phenotypes) encoded by each individual variome (full set of genome variants) in the context of the given environment. Numerous tools exist for the computational identification of the functional effects of a single variant. However, the pipelines taking advantage of full genomic, exomic, transcriptomic (and other) sequences have only recently become a reality. This review looks at the building of methodologies for predicting "variome"-defined disease risk. It also discusses some of the challenges for incorporating such a pipeline into everyday medical practice.

Full-field optical coherence microscopy is a novel technique for imaging enteric ganglia in the gastrointestinal tract.

BACKGROUND: Noninvasive methods are needed to improve the diagnosis of enteric neuropathies. Full-field optical coherence microscopy (FFOCM) is a novel optical microscopy modality that can acquire 1 µm resolution images of tissue. The objective of this research was to demonstrate FFOCM imaging for the characterization of the enteric nervous system (ENS). METHODS: Normal mice and EdnrB(-/-) mice, a model of Hirschsprung's disease (HD), were imaged in three-dimensions ex vivo using FFOCM through the entire thickness and length of the gut. Quantitative analysis of myenteric ganglia was performed on FFOCM images obtained from whole-mount tissues and compared with immunohistochemistry imaged by confocal microscopy. KEY RESULTS: Full-field optical coherence microscopy enabled visualization of the full thickness gut wall from serosa to mucosa. Images of the myenteric plexus were successfully acquired from the stomach, duodenum, colon, and rectum. Quantification of ganglionic neuronal counts on FFOCM images revealed strong interobserver agreement and identical values to those obtained by immunofluorescence microscopy. In EdnrB(-/-) mice, FFOCM analysis revealed a significant decrease in ganglia density along the colorectum and a significantly lower density of ganglia in all colorectal segments compared with normal mice. CONCLUSIONS & INFERENCES: Full-field optical coherence microscopy enables optical microscopic imaging of the ENS within the bowel wall along the entire intestine. FFOCM is able to differentiate ganglionic from aganglionic colon in a mouse model of HD, and can provide quantitative assessment of ganglionic density. With further refinements that enable bowel wall imaging in vivo, this technology has the potential to revolutionize the characterization of the ENS and the diagnosis of enteric neuropathies.

The neuroprotective adenosine-activated signal transduction pathway involves activation of phospholipase C.

We have demonstrated before that exposure of neuronal cultures to poisoning by iodoacetic acid (IAA) followed by "reperfusion" (IAA-R insult), results in severe cytotoxicity, which could be markedly attenuated by prior activation of the adenosine A1 receptors. We also have demonstrated that adenosine activates a signal transduction pathway (STP), which involves activation of PKC epsilon and opening of KATP channels. Here, we provide proof for the involvement also of phospholipase C (PLC) in the neuronal protective adenosine-activated STP. R-PIA, a specific A1 adenosine receptor agonist, was found to enhance neuronal PLC activity and protect against the IAA-R insult. The PLC inhibitor U73122, abrogated both R-PIA-induced effects. These results demonstrate that activation of PLC is a vital step in the neuronal protective adenosine-induced STP.

Clinical and biochemical manifestations and molecular characterization of the mutation HPRT Jerusalem.

A novel point mutation (I137T) was identified in the hypoxanthine-guanine phosphoribosyltransferase (HPRT) encoding gene, in a patient with partial deficiency of the enzyme. The mutation, ATT to ACT (substitution of isoleucine to threonine), occurred at codon 137, which is within the region encoding the binding site for 5-phosphoribosyl-1-pyrophosphate (PRPP). The mutation caused decreased affinity for PRPP, manifested clinically as a Lesch-Nyhan variant (excessive purine production and delayed acquisition of language skills). The partial HPRT deficiency could be detected only by measuring HPRT activity in intact fibroblasts (uptake of hypoxanthine into nucleotides).

Kelley-Seegmiller syndrome due to a unique variant of hypoxanthine-guanine phosphoribosyltransferase: reduced affinity for 5-phosphoribosyl-1-pyrophosphate manifested only at low, physiological substrate concentrations.

A male child, who presented at the age of 3.5 years with acute renal failure, was diagnosed as having partial deficiency of hypoxanthine-guanine phosphoribosyltransferase (HPRT; EC 2.4.2.8). The underlying HPRT mutation was unique in that the specific activity of HPRT in erythrocyte and in fibroblast lysates was normal, but the rate of uptake of hypoxanthine into nucleotides of intact cultured fibroblasts was markedly reduced (23% of normal). The low functioning of HPRT in the intact fibroblasts was associated with decreased utilization of endogenously generated hypoxanthine and with decreased utilization of the cosubstrate 5-phosphoribosyl-1-pyrophosphate (PRPP). The non-utilized hypoxanthine was excreted into the incubation medium. The accumulation of PRPP was indicated by the 2.3-fold increase in the rate of uptake of adenine into intact cell nucleotides and by the 7. 5-fold enhancement of the rate of de novo purine synthesis. Kinetic studies of HPRT activity in fibroblast lysates revealed reduced affinity of the enzyme for PRPP (apparent K(m) 500 microM in comparison to 25 microM in control lysates), manifested in low activity at low (physiological), but not at high PRPP concentrations. The apparent K(m) for hypoxanthine was normal (23 microM in comparison to 14.2 microM in control lysates). With allopurinol treatment, our patient has had no problems since presentation, and is developing normally at 5 years of age.

EHD1--an EH-domain-containing protein with a specific expression pattern.

A cDNA that is a member of the eps15 homology (EH)-domain-containing family and is expressed differentially in testis was isolated from mouse and human. The corresponding genes map to the centromeric region of mouse chromosome 19 and to the region of conserved synteny on human chromosome 11q13. Northern analysis revealed two RNA species in mouse. In addition to the high levels in testis, expression was noted in kidney, heart, intestine, and brain. In human, three RNA species were evident. The smaller one was predominant in testis, while the largest species was evident in other tissues as well. The predicted protein sequence has an EH domain at its C-terminus, including an EF, a Ca2+ binding motif, and a central coiled-coil structure, as well as a nucleotide binding consensus site at its N-terminus. As such, it is a member of the EH-domain-containing protein family and was designated EHD1 (EH domain-containing 1). In cells in tissue culture, we localized EHD1 as a green fluorescent protein fusion protein, in transferrin-containing, endocytic vesicles. Immunostaining of different adult mouse organs revealed major expression of EHD1 in germ cells in meiosis, in the testes, in adipocytes, and in specific retinal layers. Results of in situ hybridization to whole embryos and immunohistochemical analyses indicated that EHD1 expression was already noted at day 9.5 in the limb buds and pharyngeal arches and at day 10.5 in sclerotomes, at various elements of the branchial apparatus (mandible and hyoid), and in the occipital region. At day 15.5 EHD1 expression peaked in cartilage, preceding hypertrophy and ossification, and at day 17.5 there was no expression in the bones. The EHD1 gene is highly conserved between nematode, Drosophila, mouse, and human. Its predicted protein structure and cellular localization point to the possibility that EHD1 participates in ligand-induced endocytosis.

Adenine nucleotide metabolism in primary rat neuronal cultures.

The metabolism of adenine nucleotides (AdRN) has been studied previously in whole brains, brain slices and brain extracts, containing mixed populations of neurons and glia. The availability of primary neuronal cultures enables us to study these pathways in almost pure neuronal preparations. The aim of the present study was to characterize the relative importance of the pathways of AdRN metabolism in the neurons. The metabolic fate of (8-14C) adenine and of AdRN prelabeled with (8-14C)adenine were studied in immature and mature primary rat neuronal cultures. Specific inhibitors were used to clarify the various metabolic fluxes, which were evaluated based on the time-related changes in the distribution of label (the cellular nucleotide content did not change during incubation). The turnover rate of AdRN was found to reflect mainly conversion of label to acid insoluble derivatives (AID) and partly degradation to hypoxanthine. The turnover was faster in the immature neurons. The combined addition of 2'-deoxycoformycin (2'-dCF) and of 5'-amino-5'-deoxyadenosine, inhibiting adenosine metabolism, resulted in both cultures in enhanced loss of label from AdRN, mainly to adenosine and adenine. This finding indicates the activity of the futile cycle AMP-->adenosine-->AMP. In both cultures, in the presence of these inhibitors, the ratio (hypoxanthine + inosine)/(adenine + adenosine) was 1.1, indicating that the fluxes through AMP deamination and AMP dephosphorylation are about equal. Addition of L-alanosine, inhibiting the conversion of IMP to AMP, resulted in both cultures, but especially in the mature neurons, in enhanced loss of label from AdRN to hypoxanthine and inosine. This finding indicates the functioning of the adenine nucleotide cycle (AMP-->IMP-->adenylosuccinic acid-->AMP). Under conditions of enhanced degradation of ATP (induced by iodoacetate and antimycin A), addition of 2'-dCF resulted in the immature cultures in lowering the ratio (hypoxanthine + inosine + IMP)/(adenine + adenosine) to 0.62, indicating a shift in favor of AMP dephosphorylation.

Developmental changes in purine nucleotide metabolism in cultured rat astroglia.

The present study was conducted in order to clarify the role of the glia in brain purine metabolism. This, in connection with the clarification of the etiology of the neurological manifestations associated with some of the inborn errors of purine metabolism in man. Purine nucleotide content, the capacity for de novo and salvage purine synthesis and the activity of several enzymes of purine nucleotide degradation, were assayed in primary cultures of rat astroglia in relation to culture age. The capacity of the intact cells to produce purine nucleotides de novo exhibited a marked decrease with the culture age, but the activity of hypoxanthine-guanine phosphoribosyltransferase (HGPRT), catalyzing salvage nucleotide synthesis, increased. Aging was also associated with a marked increase in the activity of the degradation enzymes AMP deaminase, purine nucleoside phosphorylase (PNP) and guanine deaminase (guanase). The activity of adenosine deaminase and of AMP-5'-nucleotidase, increased markedly during the first 17 days in culture, but decreased thereafter. The results indicate that purine nucleotide metabolism in the cultured astroglia is changing with aging to allow the cells to maintain their nucleotide pool by reutilization of preformed hypoxanthine, rather than by de-novo production of new purines. Aging is also associated with increased capacity for operation of the adenine nucleotide cycle, contributing to the homeostasis of adenine nucleotides and to the energy charge of the cells. In principle, the age-related alterations in purine metabolism in the astroglia resemble those occurring in the maturating neurons, except for the capacity to produce purines de novo, which exhibited inverse trends in the two tissues. However, in comparison to the neurons, the cultured astroglia possess the capacity for a more intensive metabolism of purine nucleotides.

Metabolism of guanine and guanine nucleotides in primary rat cardiomyocyte cultures.

The metabolic fate of labeled guanine and of prelabeled guanine nucleotides (GuRN) was studied in cultured rat cardiomyocytes. Special attention was given to guanine salvage in comparison to degradation; to the contribution of GuRN to adenine nucleotides (AdRN); to the fluxes from GMP to IMP and from IMP to GMP; and to the degradation pathways of GuRN. In accordance with the 3- to 4-fold higher activity of guanine deaminase (guanase), in comparison to that of hypoxanthine-guanine phosphoribosyltransferase (HGPRT), the rate of guanine deamination to xanthine exceeded that of guanine incorporation into nucleotides (at 4 microM) by 13.2-fold. The label from guanine incorporated into nucleotides was found mainly (81%) in GuRN, but also in IMP and AdRN. The prelabeled GuRN lost 43% of the label in 4 h, reflecting mainly degradation to xanthine (and uric acid) and synthesis of nucleic acids. Blocking nucleoside degradation was associated with a marked accumulation of label in guanosine and inosine (guanosine/inosine labeling ratio is 1.25). The results indicate that in the myocardium guanine is a poor substrate for salvage synthesis of GuRN and that its contribution to the homeostasis of adenine nucleotides is negligible; that GMP degradation to xanthine proceeds through both guanosine and IMP; and that the cardiomyocytes contain the activity of GMP reductase and of the enzymes converting IMP to GMP.

The GDP-bound form of the small G protein Rac1 p21 is a potent activator of the superoxide-forming NADPH oxidase of macrophages.

Phagocytes produce superoxide by the assembly of a multicomponent complex that utilizes NADPH for the reduction of molecular oxygen (NADPH oxidase). The components participating in the assembly are a membrane-bound flavocytochrome and three cytosolic proteins, one of which was shown to be a dimer of the small GTP-binding protein (G protein) Rac1 p21 or Rac2 p21 with GDP dissociation inhibitor for Rho (Rho GDI). We determined the identity and quantity of the nucleotide bound to Rac1 p21 by high performance anion exchange chromatography of extracts prepared from highly purified Rac1 p21-Rho GDI, isolated from guinea pig macrophage cytosol. Rac1 p21 contained only GDP at a ratio of close to 1 mol of GDP per mol of G protein. The GDP-bound form of Rac1 p21 complexed to Rho GDI functioned as a potent activator of NADPH oxidase in a cell-free system that contained no free GTP or ATP. We propose that the GDP-bound form of Rac1 p21 might be the physiological activator of NADPH oxidase in macrophages, following its dissociation from Rho GDI, and that nucleotide exchange or conversion to GTP is not necessarily involved.

Effects of differentiation-inducing agents on purine nucleotide metabolism in an ovarian cancer cell line.

The effects of the differentiation-inducing agents sodium butyrate (NaOBt), dimethylsulfoxide (DMSO) and mycophenolic acid (MA), on purine nucleotide metabolism, was studied in an ovarian carcinoma cell line (GZL-8). Exposure to these agents inhibited cell proliferation, but did not affect cell viability. Three hours following exposure, NaOBt and DMSO moderately decelerated purine synthesis de novo, but MA accelerated it three-fold, this being associated with a two-fold increase in the excretion of hypoxanthine and xanthine into the incubation medium. NaOBt and DMSO did not affect the cellular nucleotide content, but MA caused a 73% decrease in GTP content and about a 50% increase in the cellular content of UTP. The following alterations in cellular enzyme activity were observed 72 h following exposure: NaOBt decreased the activity of hypoxanthine-guanine phosphoribosyltransferase and increased the activity of IMP and of AMP 5'-nucleotidases, DMSO increased the activity of IMP 5'-nucleotidase, and MA increased the activity of the two nucleotidases. The results suggest that, in the carcinoma cell line studied, the differentiation process induced by NaOBt and DMSO may be associated with a general shift in the direction of purine metabolism from anabolism to catabolism, whereas that induced by MA is associated with a specific decrease in the production of GTP.

Characterization of the alterations in purine nucleotide metabolism in hypoxanthine-guanine phosphoribosyltransferase-deficient rat neuroma cell line.

A rat neuroma cell line (B103 4C), deficient of hypoxanthine-guanine phosphoribosyltransferase (HGPRT), was utilized as a model tissue in search for the biochemical basis of the Lesch-Nyhan syndrome (LNS). The HGPRT-deficient neurons exhibited the following properties: an almost complete absence of uptake of guanine and of hypoxanthine into intact cell nucleotides (0.92% and 0.69% of normal, respectively); a significant increase in the availability of 5'-phosphoribosyl-1-pyrophosphate; a three- to fourfold acceleration of the rate of de novo nucleotide synthesis; a normal excretion of xanthine, but 15-fold increase in the excretion of hypoxanthine into the culture media; a normal cellular purine nucleotide content, including the absence of 5-amino-4-imidazole carboxamide nucleotides (Z-nucleotides), but enhanced turnover of adenine nucleotides (loss of 86% of the radioactivity of the prelabeled pool in 24 h, in comparison to 73% in the normal line), and an elevated UTP content. The results suggest that, under physiological conditions, guanine salvage does not occur in the normal neurons, but that hypoxanthine salvage is of great importance in the homeostasis of the adenine nucleotide pool. The finding of the normal profile of purine nucleotides in the HGPRT-deficient neurons indicates that the lack of hypoxanthine salvage is adequately compensated by the enhanced de novo nucleotide synthesis. These results did not furnish evidence in support of the possibility that GTP or ATP depletion, or Z-nucleotide accumulation, occurs in HGPRT-deficient neurons and that these are etiological factors causing the neurological abnormalities in LNS.(ABSTRACT TRUNCATED AT 250 WORDS)

Metabolic fate of hypoxanthine and inosine in cultured cardiomyocytes.

The metabolic fate of labeled hypoxanthine and inosine, degradation products of adenine nucleotides, was studied in cultured beating cardiomyocytes, in order to assess the physiological significance of their contribution to salvage nucleotide synthesis in the heart. Inosine and hypoxanthine were found to be incorporated into nucleotides by a similar rate, but in the presence of 8-aminoguanosine, a potent inhibitor of purine nucleoside phosphorylase (EC 2.4.2.1), the rate of inosine incorporation into nucleotides was markedly reduced (by 75%), indicating that inosine incorporation to IMP (inosinic acid) occurs following its degradation to hypoxanthine. The proportion of hypoxanthine converted to IMP by hypoxanthine-guanine phosphoribosyltransferase (EC 2.4.2.8) is markedly greater than that degraded to xanthine and uric acid by xanthine oxidase (EC 1.3.2.3). However, close to 50% of the IMP formed was degraded to inosine by IMP 5'-nucleotidase (EC 3.1.3.5). The results demonstrate the activity of the following futile cycle in the cardiomyocytes: hypoxanthine----IMP----inosine----hypoxanthine. The rational for the activity of this energy consuming cycle is yet unclear.

Developmental changes in the activity of enzymes of purine metabolism in rat neuronal cells in culture and in whole brain.

The activities (Vmax) of several enzymes of purine nucleotide metabolism were assayed in premature and mature primary rat neuronal cultures and in whole rat brains. In the neuronal cultures, representing 90% pure neurons, maturation (up to 14 days in culture) resulted in an increase in the activities of guanine deaminase (guanase), purine-nucleoside phosphorylase (PNP), IMP 5'-nucleotidase, adenine phosphoribosyltransferase (APRT), and AMP deaminase, but in no change in the activities of hypoxanthine-guanine phosphoribosyltransferase (HGPRT), adenosine deaminase, adenosine kinase, and AMP 5'-nucleotidase. In whole brains in vivo, maturation (from 18 days of gestation to 14 days post partum) was associated with an increase in the activities of guanase, PNP, IMP 5'-nucleotidase, AMP deaminase, and HGPRT, a decrease in the activities of adenosine deaminase and IMP dehydrogenase, and no change in the activities of APRT, AMP 5'-nucleotidase, and adenosine kinase. The profound changes in purine metabolism, which occur with maturation of the neuronal cells in primary cultures in vitro and in whole brains in vivo, create an advantage for AMP degradation by deamination, rather than by dephosphorylation, and for guanine degradation to xanthine over its reutilization for synthesis of GMP. The physiological meaning of the maturational increase in these two ammonia-producing enzymes in the brain is not yet clear. The striking similarity in the alterations of enzyme activities in the two systems indicates that the primary culture system may serve as an appropriate model for the study of purine metabolism in brain.

Prenatal diagnosis of Lesch-Nyhan syndrome: experience with three fetuses at risk.

Our experience with the prenatal detection of the Lesch-Nyhan syndrome (LNS; hypoxanthine-guanine phosphoribosyltransferase (HGPRT) deficiency) in three fetuses at risk is reported. Enzyme activities were measured in cultured amniocytes in two pregnancies, and in tissues and cultures obtained from chorionic villus sampling (CVS) in a third pregnancy. In all tissues the specific activities of HGPRT and adenine phosphoribosyltransferase (APRT) were determined and APRT/HGPRT ratios were calculated. In addition to the enzyme assays, the rate of purine synthesis de novo was assessed in the two amniocyte cultures, and the rate of [14C]hypoxanthine incorporation into nucleotides and sensitivity to azaguanine were measured in one of the amniocyte cultures. We report the diagnosis of normal fetuses by study of amniocytes in two pregnancies and of LNS using CVS in one pregnancy. In all three cases the diagnosis was confirmed.