Computational protein function prediction: are we making progress?

The computational prediction of gene and protein function is rapidly gaining ground as a central undertaking in computational biology. Making sense of the flood of genomic data requires fast and reliable annotation. Many ingenious algorithms have been devised to infer a protein's function from its amino acid sequence, 3D structure and chromosomal location of the encoding genes. However, there are significant challenges in assessing how well these programs perform. In this article we explore those challenges and review our own attempt at assessing the performance of those programs. We conclude that the task is far from complete and that a critical assessment of the performance of function prediction programs is necessary to make true progress in computational function prediction.

The status of structural genomics defined through the analysis of current targets and structures.

Structural genomics--large-scale macromolecular 3-dimenional structure determination--is unique in that major participants report scientific progress on a weekly basis. The target database (TargetDB) maintained by the Protein Data Bank (http://targetdb.pdb.org) reports this progress through the status of each protein sequence (target) under consideration by the major structural genomics centers worldwide. Hence, TargetDB provides a unique opportunity to analyze the potential impact that this major initiative provides to scientists interested in the sequence-structure-function-disease paradigm. Here we report such an analysis with a focus on: (i) temporal characteristics--how is the project doing and what can we expect in the future? (ii) target characteristics--what are the predicted functions of the proteins targeted by structural genomics and how biased is the target set when compared to the PDB and to predictions across complete genomes? (iii) structures solved--what are the characteristics of structures solved thus far and what do they contribute? The analysis required a more extensive database of structure predictions using different methods integrated with data from other sources. This database, associated tools and related data sources are available from http://spam.sdsc.edu.

Glimmers in the midnight zone: characterization of aligned identical residues in sequence-dissimilar proteins sharing a common fold.

Sequence comparison of proteins that adopt the same fold has revealed a large degree of sequence variation. There are many pairs of structurally similar proteins with only a very low percentage of identical residues at structurally aligned positions. It is not clear whether these few identical residues have been conserved just by coincidence, or due to their structural and/or functional role The current study focuses on characterization of STructurally Aligned Identical ResidueS (STAIRS) in a data set of protein pairs that are structurally similar but sequentially dissimilar. The conservation pattern of the residues at structurally aligned positions has been characterized within the protein families of the two pair members, and mutually highly and weakly conserved positions of STAIRS could be identified About 40% of the STAIRS are only moderately conserved, suggesting that their maintenance may have been coincidental. The mutually highly conserved STAIRS show distinct features that are associated with protein structure and function: a relatively high fraction of these STAIRS are buried within their protein structures. Glycine, cysteine, histidine, and tryptophan are significantly over-represented among the mutually conserved STAIRS. A detailed survey of these STAIRS reveals residue-specific roles in the determination of the protein's structure and function.

Recombinant in vitro tools to predict drug metabolism and safety.

Drug metabolism determines several pharmacological and toxicological properties of pharmaceuticals and is catalysed by drug metabolizing enzymes. Prediction of drug metabolism in humans based on animal experiments is complicated by species differences in the catalytic properties of these enzymes. This review describes and evaluates the use of recombinant models that contain human drug metabolizing enzymes to facilitate the prediction of pharmacokinetic properties of candidate drugs in humans.

Evaluation of PSI-BLAST alignment accuracy in comparison to structural alignments.

The PSI-BLAST algorithm has been acknowledged as one of the most powerful tools for detecting remote evolutionary relationships by sequence considerations only. This has been demonstrated by its ability to recognize remote structural homologues and by the greatest coverage it enables in annotation of a complete genome. Although recognizing the correct fold of a sequence is of major importance, the accuracy of the alignment is crucial for the success of modeling one sequence by the structure of its remote homologue. Here we assess the accuracy of PSI-BLAST alignments on a stringent database of 123 structurally similar, sequence-dissimilar pairs of proteins, by comparing them to the alignments defined on a structural basis. Each protein sequence is compared to a nonredundant database of the protein sequences by PSI-BLAST. Whenever a pair member detects its pair-mate, the positions that are aligned both in the sequential and structural alignments are determined, and the alignment sensitivity is expressed as the percentage of these positions out of the structural alignment. Fifty-two sequences detected their pair-mates (for 16 pairs the success was bi-directional when either pair member was used as a query). The average percentage of correctly aligned residues per structural alignment was 43.5+/-2.2%. Other properties of the alignments were also examined, such as the sensitivity vs. specificity and the change in these parameters over consecutive iterations. Notably, there is an improvement in alignment sensitivity over consecutive iterations, reaching an average of 50.9+/-2.5% within the five iterations tested in the current study.

Secondary dimerization between members of the epidermal growth factor receptor family.

Growth factor receptors of the epidermal growth factor (EGF) receptor family play pivotal roles in the regulation of cell proliferation and differentiation and are involved in the development of human cancers. It has been well documented that these receptors undergo growth factor-stimulated homo- and heterodimerization as a first step in the initiation of signaling cascades. Here we provide evidence for a new mechanism for growth factor-stimulated receptor dimer formation, designated secondary dimerization. The growth factor-induced dimerization and ensuing receptor trans-autophosphorylation results in the dissociation of the original (primary) receptor dimer. Each phosphorylated receptor monomer then interacts with a new (nonphosphorylated) receptor to form a secondary dimer. Treatment of cells with EGF yields Neu-ErbB3 secondary dimers, and heregulin treatment induces the formation of Neu-EGF receptor (secondary) dimers. The ability of EGF and heregulin to stimulate a cascade of dimerization events points to a novel mechanism by which multiple signaling activities and diverse biological responses are initiated by members of the EGF receptor family.

Submitochondrial distribution of three key steroidogenic proteins (steroidogenic acute regulatory protein and cytochrome p450scc and 3beta-hydroxysteroid dehydrogenase isomerase enzymes) upon stimulation by intracellular calcium in adrenal glomerulosa cells.

In adrenal glomerulosa cells, angiotensin II (Ang II) and potassium stimulate aldosterone synthesis through activation of the calcium messenger system. The rate-limiting step in steroidogenesis is the transfer of cholesterol to the inner mitochondrial membrane. This transfer is believed to depend upon the presence of the steroidogenic acute regulatory (StAR) protein. The aim of this study was 1) to examine the effect of changes in cytosolic free calcium concentration and of Ang II on intramitochondrial cholesterol and 2) to study the distribution of StAR protein in submitochondrial fractions during activation by Ca2+ and Ang II. To this end, freshly prepared bovine zona glomerulosa cells were submitted to a high cytosolic Ca2+ clamp (600 nM) or stimulated with Ang II (10 nM) for 2 h. Mitochondria were isolated and subfractionated into outer membranes, inner membranes (IM), and contact sites (CS). Stimulation of intact cells with Ca2+ or Ang II led to a marked, cycloheximide-sensitive increase in cholesterol in CS (to 143 +/- 3. 2 and 151.1 +/- 18.1% of controls, respectively) and in IM (to 119 +/- 5.1 and 124.5 +/- 6.5% of controls, respectively). Western blot analysis revealed a cycloheximide-sensitive increase in StAR protein in mitochondrial extracts of Ca2+-clamped glomerulosa cells (to 159 +/- 23% of controls). In submitochondrial fractions, there was a selective accumulation of StAR protein in IM following stimulation with Ca2+ (228 +/- 50%). Similarly, Ang II increased StAR protein in IM, and this effect was prevented by cycloheximide. In contrast, neither Ca2+ nor Ang II had any effect on the submitochondrial distribution of cytochrome P450scc and 3beta-hydroxysteroid dehydrogenase isomerase. The intramitochondrial presence of the latter enzyme was further confirmed by immunogold staining in rat adrenal fasciculata cells and by immunoblot analysis in MA-10 mouse testicular Leydig cells. These findings demonstrate that under acute stimulation with Ca2+-mobilizing agents, newly synthesized StAR protein accumulates in IM after transiting through CS. Moreover, our results suggest that the import of StAR protein into IM may be associated with cholesterol transfer, thus promoting precursor supply to the two first enzymes of the steroidogenic cascade within the mitochondria and thereby activating mineralocorticoid synthesis.

Combined effects of ATP and its analogs on the membrane permeability in transformed mouse fibroblasts.

Extracellular ATP (0.6 mM) induces a marked decrease in the membrane potential, followed by an increase in cell membrane permeability in transformed mouse fibroblasts. The effects of the ATP analogs, p[CH2]ppA and p[NH]ppA (0.6 mM), on the membrane potential and permeability are much less pronounced. ATP at 0.05 mM has no effect by itself, but markedly increases the analog-induced membrane potential dissipation and permeability. The data suggest that ATP-induced membrane permeation is composed of two processes: One is common to ATP and its analogs and appears to be a receptor-mediated process. The second is unique for ATP, effective even at low concentration (0.05 mM), and might be mediated by cell surface enzymes, for which ATP, but not its analogs, serves as a substrate.

Activation of cell growth inhibitor by ectoprotein kinase-mediated phosphorylation in transformed mouse fibroblasts.

Our previous studies have shown that exogenous ATP induces cell growth inhibition in transformed mouse fibroblasts, 3T6 cells, whereas the growth of their nontransformed counterparts, Swiss 3T3 cells, is only slightly affected. In this study a similar selective, ATP-induced growth inhibition was found in Balb/c SV40-3T3 cells and in primary cultures of adenovirus-transformed murine fibroblasts. The inhibitory activity was found in the conditioned media of ATP-treated cultures. Several lines of evidence have shown that ectoprotein kinase (ecto-PK) plays a major role in the ATP-induced growth inhibition. (a) There is a good correlation between the activity of ecto-PK and the ability of ATP to induce cell growth inhibition. (b) The removal of the ecto-PK from the cell surface prevents the ATP-induced growth inhibition. (c) Addition of the removed enzyme to the cell culture reconstitutes the ability of ATP to induced growth inhibition. (d) Serum-containing, or serum-free, conditioned media from untreated cultures gain an inhibitory activity after their phosphorylation, and dephosphorylation of conditioned media from ATP-treated cultures results in the loss of the inhibitory activity. (e) Growth medium by itself does not inhibit cell proliferation after its phosphorylation. The findings described in d and e indicate, as well, that the ATP-induced growth inhibitor is produced by the cells. The putative inhibitor was found to be a protein, with an apparent molecular mass of 13 kDa. The selectivity of the inhibition for transformed cells is due to the higher level of ecto-PK in these cells, as well as to their higher susceptibility to the inhibitor, as compared with their non-transformed counterparts.

Growth inhibition of breast cancer cells induced by exogenous ATP.

Addition of ATP (> 0.1 mM) to cultures of human breast cancer T47D cells resulted in an inhibition of cell proliferation. The inhibition was found to be specific for ATP, and dependent on its concentration. Growth inhibition continued for at least three days, although ATP and its hydrolysis products were metabolized within one day. Conditioned medium from ATP-treated cultures (CM+) was found to inhibit the growth of cells that were not exposed to ATP. This is an indication that extracellular factors, besides ATP, are involved in the inhibition process. The inhibition was maintained after dialysis of the CM+, using an 8 kDa cut-off membrane. Conditioned medium from untreated cultures (CM-), however, only slightly affected cell growth. The data suggest that the CM(+)-induced cell growth inhibition is mediated by an ATP-activated growth inhibiting factor. Flow microfluorometry and thymidine incorporation experiments have shown that the growth arrest is mainly due to the elongation of the S-phase of the cell cycle.

Spheroplast-derived membrane vesicles from Rhodobacter capsulatus cells catalyzing nucleotide transport.

Rodobacter capsulatus cells, which were cultured anaerobically in high light intensity, had fewer foldings in the cytoplasmic membrane than those which were grown in lower light intensities. Spheroplast-derived membrane fractions obtained from cells cultured under high light intensity contained a high yield of large right-side-out membrane vesicles. The right-side-out vesicles catalyzed reversible light-induced proton efflux as did intact cells. Nucleotide transport activity was also catalyzed by these membrane vesicles. This activity was indirectly monitored by measurement of photophosphorylation or hydrolysis of externally added diphospho- and triphosphonucleosides. These enzymatic activities occur inside the cytoplasmic membrane of spheroplasts and membrane vesicles and therefore require the transport of the externally added reagents. The indirect measurements of transport were complemented by the demonstration of direct uptake of radiolabeled nucleotides into the membrane vesicles. These data support the suggestion that a nucleotide transporter located in the cytoplasmic membrane of R. capsulatus bacteria mediates these activities.

Effect of ionophore A23187 on the membrane permeability in mouse fibroblasts.

Addition of the divalent cation ionophore A23187 to transformed mouse fibroblasts (3T6) resulted in an increase in the cell membrane permeability to normally impermeant solutes (e.g., nucleotides). The membrane permeability was assessed by following the efflux of prelabeled adenine nucleotides, the influx of p-nitrophenyl phosphate in cells attached to plastic dishes and reconstitution of intracellular protein synthesis in the presence of exogenously added normally impermeant factors required for macromolecular synthesis. The permeability change of 3T6 cells was found to be dependent on the specific presence of external calcium ion. The permeabilization was found to occur preferably in alkaline pH and specific to certain transformed cells. It is preceded by rapid efflux of K+, influx of Na+ and partial hydrolysis of cellular nucleotides in 3T6 cells. Similar ion fluxes were previously found to precede cell permeabilization by electrogenic ionophores for monovalent ions and by exogenous ATP. Our data suggest that a calcium dependent process caused the K+ release and excess Na+ entry, causing dissipation of the membrane potential and subsequent formation of aqueous channels.

Two distinct receptors for ATP can be distinguished in Swiss 3T6 mouse fibroblasts by their desensitization.

The stimulation of calcium efflux from Swiss 3T6 mouse fibroblasts by extracellular ATP was studied. It was found that the cells could be desensitized to ATP by a previous exposure to the nucleotide, lending support to the theory that this is a receptor mediated process. Another ATP-receptor mediated process in Swiss 3T6 cells, that is also subject to desensitization, causes the permeabilization of the plasma membrane to nucleotides and other normally impermeant compounds [Gonzalez et al., J. Cell. Physiol. 139:109 (1989)]. Here we demonstrate that selective desensitization of the ATP-dependent calcium mobilization pathway can be achieved without affecting ATP-induced permeabilization. Data are presented in support of the existence of multiple ATP-receptors (purinoceptors) in Swiss 3T6 cells.

ATP-resistant variants of transformed mouse fibroblasts.

Addition of ATP to cultures of transformed mouse fibroblasts, 3T6 cells, resulted in cell growth inhibition, whereas the growth of the non-transformed counterparts, 3T3 cells, was only slightly affected. The inhibition was found to be specific for adenine nucleotides, and concentration dependent. At relatively low concentrations (e.g., 1.0 mM) the effect of ATP was cytostatic, whereas at higher concentrations (e.g., 1.0 mM) a cytotoxic effect was exerted. ATP-resistant variants of 3T6 cells were selected by exposure of cultures to gradually elevated concentrations of ATP. The variants were found to resemble the non-transformed counterparts, 3T3 cells, more than the 3T6 parent cells, by the following criteria: ATP-induced alterations in the membrane potential, changes in membrane permeability, cell growth inhibition, and colony formation on soft agar. The data indicate that long exposure of the transformed cells to external ATP results in redifferentiation and reduction in their tumorigenicity.

Bombesin, diacylglycerols, and phorbol esters rapidly stimulate the phosphorylation of an Mr = 80,000 protein kinase C substrate in permeabilized 3T3 cells. Effect of guanine nucleotides.

We have used digitonin permeabilization to study the mechanism of bombesin-induced activation of protein kinase C in Swiss 3T3 cells. Protein kinase C-mediated phosphorylations in permeabilized cells were identified using phorbol esters and diacylglycerols. Addition of phorbol 12,13-dibutyrate (PDBu) in the presence of [gamma-32P]ATP and digitonin caused a marked and rapid time- and dose-dependent increase in the phosphorylation of an Mr 80,000 cellular protein (maximum stimulation = 12.6 +/- 1.6-fold after 1 min, EC50 = 27 nM). 1-oleoyl-2-acetylglycerol substituted for PDBu in stimulating the phosphorylation of Mr 80,000 protein (EC50 = 13 microM). Bombesin also caused a striking increase in the phosphorylation of Mr 80,000 protein with a time course similar to that observed with PDBu. This phosphorylation was mimicked by mammalian bombesin-like peptides and blocked by the bombesin antagonists [D-Arg1,D-Phe5,D-Trp7,9,Leu11]substance P and [Leu13 psi (CH2NH)Leu14]bombesin. Down-regulation of protein kinase C in intact cells by prolonged exposure to PDBu prevented Mr 80,000 protein phosphorylation upon subsequent bombesin addition in digitonin-permeabilized cells. Comigration on one- and two-dimensional gel electrophoresis and phosphopeptide mapping confirmed that the Mr 80,000 protein phosphorylated in permeabilized cells was indistinguishable from the Mr 80,000 protein which is the major protein kinase C substrate in intact cells. The GDP analogue guanosine-5'-O-(2-thiodiphosphate) (GDP beta S) caused a 70% inhibition of the bombesin-induced phosphorylation of Mr 80,000 protein but had no effect on the phosphorylation induced by PDBu. Bombesin stimulated Mr 80,000 protein phosphorylation in permeabilized cells in a dose-dependent manner (EC50 = 4 nM), and GDP beta S shifted the bombesin dose response curve to higher bombesin concentrations (EC50 = 14 nM). These results demonstrate for the first time a growth factor receptor-mediated activation of protein kinase C in permeabilized cells and provide functional evidence for the involvement of a G protein in the transmembrane signaling pathway that mediates the stimulation of protein kinase C by bombesin in Swiss 3T3 cells.

Growth inhibition of transformed mouse fibroblasts by adenine nucleotides occurs via generation of extracellular adenosine.

The growth of transformed mouse fibroblasts (3T6 cells) in medium containing 5% fetal bovine serum was inhibited after treatment with concentrations greater than 50 microM ATP, ADP, or AMP. Adenosine, the common catabolite of the nucleotides, had no effect on cell growth at concentrations below 1 mM. However, the following results indicate that the toxicity of ATP, ADP, and AMP is mediated by serum- and cell-associated hydrolysis of the nucleotides to adenosine. 1) ADP and AMP, but not ATP, were toxic to 3T6 cells grown in serum-free medium or medium in which phosphohydrolase activity of serum was inactivated. Under these conditions, the cells exhibited cell-associated ADPase and 5'-nucleotidase activity, but little ecto-ATPase activity. 2) Inhibition of adenosine transport in 3T6 cells by dipyridamole or S-(p-nitrobenzyl)-6-thioinosine prevented the toxicity of ATP in serum-containing medium and of ADP and AMP in serum-free medium. 3) A 16-24-h exposure to 125 microM AMP or ATP was needed to inhibit cell growth under conditions where serum- and cell-associated hydrolysis of the nucleotides generated adenosine in the medium continuously over the same time period. In contrast, 125 microM adenosine was completely degraded to inosine and hypoxanthine within 8-10 h. Furthermore, multiple doses of adenosine added to the cells at regular intervals over a 16-h period were significantly more toxic than an equivalent amount of adenosine added in one dose. Treatment of 3T6 cells with AMP elevated intracellular ATP and ADP levels and reduced intracellular UTP levels, effects which were inhibited by extracellular uridine. Uridine also prevented growth inhibition by ATP, ADP, and AMP. These and other results indicate that serum- and cell-associated hydrolysis of adenine nucleotides to adenosine suppresses growth by adenosine-dependent pyrimidine starvation.

Heat-induced alterations in cell membrane permeability and cell inactivation of transformed mouse fibroblasts.

Hyperthermia, has recently been extended in many permutations as a modality of anticancer treatment, but the mechanisms underlying heat-induced cell inactivation are poorly understood. In this study, the role of the cell permeability barrier in the process of heat cytotoxicity are examined. Changes in cell membrane permeability were determined by following the efflux of normally impermeant metabolites, e.g. nucleotides, in cultures of Swiss mouse 3T3 cells, and their transformed derivatives, 3T6 cells. The increase in cell membrane permeability as a function of temperature and exposure duration was found to be characterized by a sigmoid curve, with a threshold value, above which the permeability markedly increased. A correlation was found between cell membrane permeabilization and cell inactivation. Both heat-induced permeabilization and heat cytotoxicity were more pronounced in the transformed cells, as compared to their untransformed counterparts. The temperature-dependent permeabilization was more effective in the presence of the ionophore amphotericin B. The data suggest that heat-induced lesion in the cell membrane has a major role in hyperthermia cytotoxicity.

ATP analogues induce membrane permeabilization in transformed mouse fibroblasts.

The mechanism underlying ATP-induced permeabilization of transformed mouse fibroblasts was studied by using nonhydrolyzable analogues of ATP. Incubation of 3T6 cells with 0.6 mM of either ATP, 5'-adenylyl imidodiphosphate (p[NH]ppA) or adenosine 5'-[beta, gamma-methylene]triphosphate (p[CH2]ppA) resulted in an increase of 17-, 8- or 5-times, respectively, in the cell membrane permeability, measured by the efflux of normally impermeant metabolites from the cells. The induced cell permeabilization was preceded by a reduction in the membrane potential (delta psi), determined according to the distribution of the cation tetraphenylphosphonium (TPP+) between the cells and the medium. Reduction of 26, 18 and 13 mV in delta psi was exerted by 0.6 mM of either ATP, p[NH]ppA or p[CH2]ppA, respectively. In 3T3 cells the untransformed counterparts of 3T6 cells, neither reduction of delta psi, nor alterations in membrane permeability were exerted by either ATP or by its analogues. The data indicate that the dissociation of the beta, gamma-phosphate bond is not essential for membrane permeabilization by external ATP, implying that the binding of ATP to the cell surface of transformed cells is sufficient to initiate the permeabilization process. The data also suggest that delta psi is involved in the control of membrane permeability.