We introduce a class of string kernels, called mismatch kernels, for use with support vector machines (SVMs) in a discriminative approach to the problem of protein classification and remote homology detection. These kernels measure sequence similarity based on shared occurrences of fixed-length patterns in the data, allowing for mutations between patterns. Thus, the kernels provide a biologically well-motivated way to compare protein sequences without relying on family-based generative models such as hidden Markov models. We compute the kernels efficiently using a mismatch tree data structure, allowing us to calculate the contributions of all patterns occurring in the data in one pass while traversing the tree. When used with an SVM, the kernels enable fast prediction on test sequences. We report experiments on two benchmark SCOP datasets, where we show that the mismatch kernel used with an SVM classifier performs competitively with state-of-the-art methods for homology detection, particularly when very few training examples are available. Examination of the highest-weighted patterns learned by the SVM classifier recovers biologically important motifs in protein families and superfamilies.

TOR (Target of Rapamycin) is a highly conserved protein kinase and a central controller of cell growth. TOR is found in two functionally and structurally distinct multiprotein complexes termed TOR complex 1 (TORC1) and TOR complex 2 (TORC2). In the present study, we developed a two-dimensional liquid chromatography tandem mass spectrometry (2D LC-MS/MS) based proteomic strategy to identify new mammalian TOR (mTOR) binding proteins. We report the identification of Proline-rich Akt substrate (PRAS40) and the hypothetical protein Q6MZQ0/FLJ14213/CAE45978 as new mTOR binding proteins. PRAS40 binds mTORC1 via Raptor, and is an mTOR phosphorylation substrate. PRAS40 inhibits mTORC1 autophosphorylation and mTORC1 kinase activity toward eIF-4E binding protein (4E-BP) and PRAS40 itself. HeLa cells in which PRAS40 was knocked down were protected against induction of apoptosis by TNFα and cycloheximide. Rapamycin failed to mimic the pro-apoptotic effect of PRAS40, suggesting that PRAS40 mediates apoptosis independently of its inhibitory effect on mTORC1. Q6MZQ0 is structurally similar to proline rich protein 5 (PRR5) and was therefore named PRR5-Like (PRR5L). PRR5L binds specifically to mTORC2, via Rictor and/or SIN1. Unlike other mTORC2 members, PRR5L is not required for mTORC2 integrity or kinase activity, but dissociates from mTORC2 upon knock down of tuberous sclerosis complex 1 (TSC1) and TSC2. Hyperactivation of mTOR by TSC1/2 knock down enhanced apoptosis whereas PRR5L knock down reduced apoptosis. PRR5L knock down reduced apoptosis also in mTORC2 deficient cells. The above suggests that mTORC2-dissociated PRR5L may promote apoptosis when mTOR is hyperactive. Thus, PRAS40 and PRR5L are novel mTOR-associated proteins that control the balance between cell growth and cell death.

TOR is a structurally and functionally conserved Ser/Thr kinase found in two multiprotein complexes that regulate many cellular processes to control cell growth. Although extensively studied, the localization of TOR is still ambiguous, possibly because endogenous TOR in live cells has not been examined. Here, we examined the localization of green fluorescent protein (GFP) tagged, endogenous TOR1 and TOR2 in live S. cerevisiae cells. A DNA cassette encoding three copies of green fluorescent protein (3XGFP) was inserted in the TOR1 gene (at codon D330) or the TOR2 gene (at codon N321). The TORs were tagged internally because TOR1 or TOR2 tagged at the N or C terminus was not functional. The TOR1 D330-3XGFP strain was not hypersensitive to rapamycin, was not cold sensitive, and was not resistant to manganese toxicity caused by the loss of Pmr1, all indications that TOR1-3XGFP was expressed and functional. TOR2-3XGFP was functional, as TOR2 is an essential gene and TOR2 N321-3XGFP haploid cells were viable. Thus, TOR1 and TOR2 retain function after the insertion of 748 amino acids in a variable region of their noncatalytic domain. The localization patterns of TOR1-3XGFP and TOR2-3XGFP were documented by imaging of live cells. TOR1-3XGFP was diffusely cytoplasmic and concentrated near the vacuolar membrane. The TOR2-3XGFP signal was cytoplasmic but predominately in dots at the plasma membrane. Thus, TOR1 and TOR2 have distinct localization patterns, consistent with the regulation of cellular processes as part of two different complexes.

Metal ions are important for all living cells. In man, metal ion deficiency leads to anemia (1), whereas metal ion overload is toxic and leads to hemochromatosis (3), Menkes' disease (2), Wilson's disease (4), and neurodegenerative diseases (5-7). Metal ions such as iron, manganese, zinc, and cobalt are involved in many catalytic reactions, gene regulation, and signal transduction pathways (8 -10). An adequate supply of metal ions to cells is important and is provided by specialized transporters.The recently cloned mammalian metal ion transporter DCT1 (11, 12), originally named Nramp2 (natural resistance-associated macrophage protein 2) (13-15), is present in both plasma membranes and endosomal vesicles for translocation, via transferrin-dependent and -independent pathways, of metal ions into the cytoplasm of cells and for maintenance of systemic metal ion homeostasis. It has been found that a mutation in DCT1 at position 185 (G185R) causes microcytic anemia in mk Ϫ/Ϫ mice and Belgrade rats (12). This mutation was subsequently shown to result in loss of Fe 2ϩ transport ability (16). DCT1-mediated iron absorption in the intestine depends on the body iron status, which is regulated in part by the hemochromatosis gene HFE, a major histocompatibility complex gene (17, 18). A single point mutation in HFE (C282Y) results in iron overload in hemochromatosis patients (3). SMF1, SMF2, and SMF3 are yeast homologues of the Nramp proteins with 51-54% identity in amino acid sequence to each other and 33-36% identity to DCT1. SMF1 was originally thought to be localized in the yeast mitochondrial membrane (19) and was named SMF, which stands for suppressor of mitochondria import function. However, more recent studies using an antibody demonstrated that SMF1 is located in the yeast plasma membrane, where it is thought to mediate uptake of Mn 2ϩ and Zn 2ϩ into the cytoplasm (20). There was indirect evidence that other divalent metal ions such as Cd 2ϩ , Co 2ϩ , and Cu 2ϩ are also substrates of SMF1 (21). In analogy to HFE in mammalian cells, the product of the yeast BSD2 (bypass superoxide dismutase deficiency gene 2), localized in the endoplasmic reticulum, regulates metal ion absorption by exerting a negative control on SMF1 activity (21,22). Despite these findings, a functional characterization of SMF1 has not yet been reported.In the present study, we expressed SMF1, SMF2, and SMF3 in Xenopus oocytes and used both a radiotracer approach and the two-microelectrode voltage-clamp technique to investigate the function of these proteins. We show that SMF1 mediates H ϩ -dependent Fe 2ϩ transport and uncoupled Na ϩ currents. SMF2 also mediates significant H ϩ -coupled Fe 2ϩ transport and uncoupled Na ϩ currents, which are much smaller than those mediated by SMF1. SMF3 exhibited no detectable activities when expressed in oocytes. Because Na ϩ inhibited metal ion uptake in oocytes expressing SMF1, we investigated the effect of Na ϩ on yeast growth. EXPERIMENTAL PROCEDURESOocyte Preparation-Yeast SMF1, SMF2, and SMF3 cDNAs were subclo...

Null mutations in genes encoding V-ATPase subunits in Saccharomyces cerevisiae result in a phenotype that is unable to grow at high pH and is sensitive to high and low metal-ion concentrations. Treatment of these null mutants with ethylmethanesulfonate causes mutations that suppress the V-ATPase null phenotype, and the mutant cells are able to grow at pH 7.5. The suppressor mutants were denoted as svf (suppressor of V-ATPase function). The frequency of svf is relatively high, suggesting a large target containing several genes for the ethylmethanesulfonate mutagenesis. The suppressors' frequency is dependent on the individual genes that were inactivated to manifest the V-ATPase null mutation. The svf mutations are recessive, because crossing the svf mutants with their corresponding V-ATPase null mutants resulted in diploid strains that are unable to grow at pH 7.5. A novel gene family in which null mutations cause pleiotropic effects on metal-ion resistance or sensitivity and distribution of membrane proteins in different targets was discovered. The family was defined as VTC (Vacuolar Transporter Chaperon) and it contains four genes in the S. cerevisiae genome. Inactivation of one of them, VTC1, in the background of V-ATPase null mutations resulted in svf phenotype manifested by growth at pH 7.5. Deletion of the VTC1 gene (⌬VTC1) results in a reduced amount of V-ATPase in the vacuolar membrane. These mutant cells fail to accumulate quinacrine into their vacuoles, but they are able to grow at pH 7.5. The VTC1 null mutant also results in a reduced amount of the plasma membrane H ؉ -ATPase (Pma1p) in membrane preparations and possibly mistargeting. This observation may provide an explanation for the svf phenotype in the double disruptant mutants of ⌬VTC1 and ⌬VMA subunits. Null mutations in genes encoding vacuolar H1 subunits are likely to be lethal for most eukaryotic cells, because energization of the vacuolar system by this enzyme drives vital secondary transport processes across membranes of vacuolar-derived organelles (1, 2). Disruption of genes encoding V-ATPase subunits in Neurospora and Drosophila melanogaster caused lethality (3, 4). On the other hand, mutant Saccharomyces cerevisiae (yeast) cells can survive the lack of acidification that results from disruption of genes encoding V-ATPase subunits (5). With the exception of VPH1 and STV1, which encode homologous proteins (6, 7), all genes encoding subunits of the V-ATPase are present as a single copy in the yeast genome (1, 8). Disruption of each of the single-copy genes yields a similar phenotype in which cells cannot grow at a pH higher than 7 and are sensitive to low and high calcium or metal ion concentrations in the medium (5, 9 -11). Mutant S. cerevisiae (yeast) cells can survive the lack of acidification that results from disruption of genes encoding V-ATPase subunits by taking up acidic external fluid via endocytosis (5, 12). However the precise metabolic junction that prevents growth of V-ATPase null mutants at high pH is not known. Moreover the l...