Recursive partitioning for tumor classification with gene expression microarray data

Precise classification of tumors is critically important for cancer diagnosis and treatment. It is also a scientifically challenging task. Recently, efforts have been made to use gene expression profiles to improve the precision of classification, with limited success. Using a published data set for purposes of comparison, we introduce a methodology based on classification trees and demonstrate that it is significantly more accurate for discriminating among distinct colon cancer tissues than other statistical approaches used heretofore. In addition, competing classification trees are displayed, which suggest that different genes may coregulate colon cancers.

The C-terminal SET domains of ALL-1 and TRITHORAX interact with the INI1 and SNR1 proteins, components of the SWI/SNF complex

The ALL-1 gene was discovered by virtue of its involvement in human acute leukemia. Its Drosophila homolog trithorax (trx) is a member of the trx-Polycomb gene family, which maintains correct spatial expression of the Antennapedia and bithorax complexes during embryogenesis. The C-terminal SET domain of ALL-1 and TRITHORAX (TRX) is a 150-aa motif, highly conserved during evolution. We performed yeast two hybrid screening of Drosophila cDNA library and detected interaction between a TRX polypeptide spanning SET and the SNR1 protein. SNR1 is a product of snr1, which is classified as a trx group gene. We found parallel interaction in yeast between the SET domain of ALL-1 and the human homolog of SNR1, INI1 (hSNF5). These results were confirmed by in vitro binding studies and by demonstrating coimmunoprecipitation of the proteins from cultured cells and/or transgenic flies. Epitope-tagged SNR1 was detected at discrete sites on larval salivary gland polytene chromosomes, and these sites colocalized with around one-half of TRX binding sites. Because SNR1 and INI1 are constituents of the SWI/SNF complex, which acts to remodel chromatin and consequently to activate transcription, the interactions we observed suggest a mechanism by which the SWI/SNF complex is recruited to ALL-1/trx targets through physical interactions between the C-terminal domains of ALL-1 and TRX and INI1/SNR1.

Sucrose is a signal molecule in assimilate partitioning

The proton–sucrose symporter mediates the key transport step in the resource distribution system that allows many plants to function as multicellular organisms. In the results reported here, we identify sucrose as a signaling molecule in a previously undescribed signal-transduction pathway that regulates the symporter. Sucrose symporter activity declined in plasma membrane vesicles isolated from leaves fed exogenous sucrose via the xylem transpiration stream. Symporter activity dropped to 35–50% of water controls when the leaves were fed 100 mM sucrose and to 20–25% of controls with 250 mM sucrose. In contrast, alanine symporter and glucose transporter activities did not change in response to sucrose treatments. Decreased sucrose symporter activity was detectable after 8 h and reached a maximum by 24 h. Kinetic analysis of transport activity showed a decrease in Vmax. RNA gel blot analysis revealed a decrease in symporter message levels, suggesting a drop in transcriptional activity or a decrease in mRNA stability. Control experiments showed that these responses were not the result of changing osmotic conditions. Equal molar concentrations of hexoses did not elicit the response, and mannoheptulose, a hexokinase inhibitor, did not block the sucrose effect. These data are consistent with a sucrose-specific response pathway that is not mediated by hexokinase as the sugar sensor. Sucrose-dependent changes in the sucrose symporter were reversible, suggesting this sucrose-sensing pathway can modulate transport activity as a function of changing sucrose concentrations in the leaf. These results demonstrate the existence of a signaling pathway that can control assimilate partitioning at the level of phloem translocation.

The Yeast Dynactin Complex Is Involved in Partitioning the Mitotic Spindle between Mother and Daughter Cells during Anaphase B

Although vertebrate cytoplasmic dynein can move to the minus ends of microtubules in vitro, its ability to translocate purified vesicles on microtubules depends on the presence of an accessory complex known as dynactin. We have cloned and characterized a novel gene, NIP100, which encodes the yeast homologue of the vertebrate dynactin complex protein p150glued. Like strains lacking the cytoplasmic dynein heavy chain Dyn1p or the centractin homologue Act5p, nip100Δ strains are viable but undergo a significant number of failed mitoses in which the mitotic spindle does not properly partition into the daughter cell. Analysis of spindle dynamics by time-lapse digital microscopy indicates that the precise role of Nip100p during anaphase is to promote the translocation of the partially elongated mitotic spindle through the bud neck. Consistent with the presence of a true dynactin complex in yeast, Nip100p exists in a stable complex with Act5p as well as Jnm1p, another protein required for proper spindle partitioning during anaphase. Moreover, genetic depletion experiments indicate that the binding of Nip100p to Act5p is dependent on the presence of Jnm1p. Finally, we find that a fusion of Nip100p to the green fluorescent protein localizes to the spindle poles throughout the cell cycle. Taken together, these results suggest that the yeast dynactin complex and cytoplasmic dynein together define a physiological pathway that is responsible for spindle translocation late in anaphase.

sqv-3, -7, and -8, a set of genes affecting morphogenesis in Caenorhabditis elegans, encode enzymes required for glycosaminoglycan biosynthesis

sqv (squashed vulva) genes comprise a set of eight independent loci in Caenorhabditis elegans required zygotically for the invagination of vulval epithelial cells and maternally for normal oocyte formation and embryogenesis. Sequencing of sqv-3, sqv-7, and sqv-8 suggested a role for the encoded proteins in glycolipid or glycoprotein biosynthesis. Using a combination of in vitro analysis of SQV enzymatic activities, sqv+-mediated rescue of vertebrate cell lines, and biochemical characterization of sqv mutants, we show that sqv-3, -7, and -8 all affect the biosynthesis of glycosaminoglycans and therefore compromise the function of one specific class of glycoconjugates, proteoglycans. These findings establish the importance of proteoglycans and their associated glycosaminoglycans in epithelial morphogenesis and patterning during C. elegans development.

Delineating developmental and metabolic pathways in vivo by expression profiling using the RIKEN set of 18,816 full-length enriched mouse cDNA arrays

We have systematically characterized gene expression patterns in 49 adult and embryonic mouse tissues by using cDNA microarrays with 18,816 mouse cDNAs. Cluster analysis defined sets of genes that were expressed ubiquitously or in similar groups of tissues such as digestive organs and muscle. Clustering of expression profiles was observed in embryonic brain, postnatal cerebellum, and adult olfactory bulb, reflecting similarities in neurogenesis and remodeling. Finally, clustering genes coding for known enzymes into 78 metabolic pathways revealed a surprising coordination of expression within each pathway among different tissues. On the other hand, a more detailed examination of glycolysis revealed tissue-specific differences in profiles of key regulatory enzymes. Thus, by surveying global gene expression by using microarrays with a large number of elements, we provide insights into the commonality and diversity of pathways responsible for the development and maintenance of the mammalian body plan.

Raft-partitioning of the ubiquitin ligases Cbl and Nedd4 upon IgE-triggered cell signaling

The high affinity receptor for IgE, FcɛRI on mast cells and basophils plays an essential role in immunological defense. Upon multivalent antigen binding, FcɛRI becomes phoshorylated by the protein-tyrosine kinase Lyn, as a result of receptor clustering in lipid rafts. FcɛRI has been shown to be ubiquitinated. Ubiquitination can lead to degradation by proteasomes, but it can also act as a sorting signal to internalize proteins destined to the endosomal/lysosomal pathway. We have analyzed whether FcɛRI ubiquitination takes place within rafts. We report biochemical and imaging evidence in rat basoleukemia cells for the presence of ubiquitinated FcɛRI in clustered rafts upon receptor activation. Moreover, we demonstrated that the ubiquitin ligases Cbl and Nedd4 colocalize with FcɛRI patches and showed that both ligases become associated with lipid rafts after activation of IgE signaling. Because Cbl is known to interact with the FcɛRI signaling complex, ubiquitination is likely to be an important parameter regulating IgE-triggered signaling occurring in rafts.

A Novel Alkaline α-Galactosidase from Melon Fruit with a Substrate Preference for Raffinose1

The cucurbits translocate the galactosyl-sucrose oligosaccharides raffinose and stachyose, therefore, α-galactosidase (α-d-galactoside galactohydrolase, EC 3.2.1.22) is expected to function as the initial enzyme of photoassimilate catabolism. However, the previously described alkaline α-galactosidase is specific for the tetrasaccharide stachyose, leaving raffinose catabolism in these tissues as an enigma. In this paper we report the partial purification and characterization of three α-galactosidases, including a novel alkaline α-galactosidase (form I) from melon (Cucumis melo) fruit tissue. The form I enzyme showed preferred activity with raffinose and significant activity with stachyose. Other unique characteristics of this enzyme, such as weak product inhibition by galactose (in contrast to the other α-galactosidases, which show stronger product inhibition), also impart physiological significance. Using raffinose and stachyose as substrates in the assays, the activities of the three α-galactosidases (alkaline form I, alkaline form II, and the acid form) were measured at different stages of fruit development. The form I enzyme activity increased during the early stages of ovary development and fruit set, in contrast to the other α-galactosidase enzymes, both of which declined in activity during this period. In the mature, sucrose-accumulating mesocarp, the alkaline form I enzyme was the major α-galactosidase present. We also observed hydrolysis of raffinose at alkaline conditions in enzyme extracts from other cucurbit sink tissues, as well as from young Coleus blumei leaves. Our results suggest different physiological roles for the α-galactosidase forms in the developing cucurbit fruit, and show that the newly discovered enzyme plays a physiologically significant role in photoassimilate partitioning in cucurbit sink tissue.

Modification of Carbon Partitioning, Photosynthetic Capacity, and O2 Sensitivity in Arabidopsis Plants with Low ADP-Glucose Pyrophosphorylase Activity1

Wild-type Arabidopsis plants, the starch-deficient mutant TL46, and the near-starchless mutant TL25 were evaluated by noninvasive in situ methods for their capacity for net CO2 assimilation, true rates of photosynthetic O2 evolution (determined from chlorophyll fluorescence measurements of photosystem II), partitioning of photosynthate into sucrose and starch, and plant growth. Compared with wild-type plants, the starch mutants showed reduced photosynthetic capacity, with the largest reduction occurring in mutant TL25 subjected to high light and increased CO2 partial pressure. The extent of stimulation of CO2 assimilation by increasing CO2 or by reducing O2 partial pressure was significantly less for the starch mutants than for wild-type plants. Under high light and moderate to high levels of CO2, the rates of CO2 assimilation and O2 evolution and the percentage inhibition of photosynthesis by low O2 were higher for the wild type than for the mutants. The relative rates of 14CO2 incorporation into starch under high light and high CO2 followed the patterns of photosynthetic capacity, with TL46 showing 31% to 40% of the starch-labeling rates of the wild type and TL25 showing less than 14% incorporation. Overall, there were significant correlations between the rates of starch synthesis and CO2 assimilation and between the rates of starch synthesis and cumulative leaf area. These results indicate that leaf starch plays an important role as a transient reserve, the synthesis of which can ameliorate any potential reduction in photosynthesis caused by feedback regulation.

Nitric oxide partitioning into mitochondrial membranes and the control of respiration at cytochrome c oxidase

An emerging and important site of action for nitric oxide (NO) within cells is the mitochondrial inner membrane, where NO binds to and inhibits members of the electron transport chain, complex III and cytochrome c oxidase. Although it is known that inhibition of cytochrome c oxidase by NO is competitive with O2, the mechanisms that underlie this phenomenon remain unclear, and the impact of both NO and O2 partitioning into biological membranes has not been considered. These properties are particularly interesting because physiological O2 tensions can vary widely, with NO having a greater inhibitory effect at low O2 tensions (<20 μM). In this study, we present evidence for a consumption of NO in mitochondrial membranes in the absence of substrate, in a nonsaturable process that is O2 dependent. This consumption modulates inhibition of cytochrome c oxidase by NO and is enhanced by the addition of exogenous membranes. From these data, it is evident that the partition of NO into mitochondrial membranes has a major impact on the ability of NO to control mitochondrial respiration. The implications of this conclusion are discussed in the context of mitochondrial lipid:protein ratios and the importance of NO as a regulator of respiration in pathophysiology.

Drying Increases Intracellular Partitioning of Amphiphilic Substances into the Lipid Phase1 : Impact on Membrane Permeability and Significance for Desiccation Tolerance

Previously we proposed that endogenous amphiphilic substances may partition from the aqueous cytoplasm into the lipid phase during dehydration of desiccation-tolerant organ(ism)s and vice versa during rehydration. Their perturbing presence in membranes could thus explain the transient leakage from imbibing organisms. To study the mechanism of this phenomenon, amphiphilic nitroxide spin probes were introduced into the pollen of a model organism, Typha latifolia, and their partitioning behavior during dehydration and rehydration was analyzed by electron paramagnetic resonance spectroscopy. In hydrated pollen the spin probes mainly occurred in the aqueous phase; during dehydration, however, the amphiphilic spin probes partitioned into the lipid phase and had disappeared from the aqueous phase below 0.4 g water g−1 dry weight. During rehydration the probes reappeared in the aqueous phase above 0.4 g water g−1 dry weight. The partitioning back into the cytoplasm coincided with the decrease of the initially high plasma membrane permeability. A charged polar spin probe was trapped in the cytoplasm during drying. Liposome experiments showed that partitioning of an amphiphilic spin probe into the bilayer during dehydration caused transient leakage during rehydration. This was also observed with endogenous amphipaths that were extracted from pollen, implying similar partitioning behavior. In view of the fluidizing effect on membranes and the antioxidant properties of many endogenous amphipaths, we suggest that partitioning with drying may be pivotal to desiccation tolerance, despite the risk of imbibitional leakage.

A Fiberless Seed Mutation in Cotton Is Associated with Lack of Fiber Cell Initiation in Ovule Epidermis and Alterations in Sucrose Synthase Expression and Carbon Partitioning in Developing Seeds1

Fiber cell initiation in the epidermal cells of cotton (Gossypium hirsutum L.) ovules represents a unique example of trichome development in higher plants. Little is known about the molecular and metabolic mechanisms controlling this process. Here we report a comparative analysis of a fiberless seed (fls) mutant (lacking fibers) and a normal (FLS) mutant to better understand the initial cytological events in fiber development and to analyze the metabolic changes that are associated with the loss of a major sink for sucrose during cellulose biosynthesis in the mutant seeds. On the day of anthesis (0 DAA), the mutant ovular epidermal cells lacked the typical bud-like projections that are seen in FLS ovules and are required for commitment to the fiber development pathway. Cell-specific gene expression analyses at 0 DAA showed that sucrose synthase (SuSy) RNA and protein were undetectable in fls ovules but were in abundant, steady-state levels in initiating fiber cells of the FLS ovules. Tissue-level analyses of developing seeds 15 to 35 DAA revealed an altered temporal pattern of SuSy expression in the mutant relative to the normal genotype. Whether the altered programming of SuSy expression is the cause or the result of the mutation is unknown. The developing seeds of the fls mutant have also shown several correlated changes that represent altered carbon partitioning in seed coats and cotyledons as compared with the FLS genotype.

A minimal gene set for cellular life derived by comparison of complete bacterial genomes.

The recently sequenced genome of the parasitic bacterium Mycoplasma genitalium contains only 468 identified protein-coding genes that have been dubbed a minimal gene complement [Fraser, C.M., Gocayne, J.D., White, O., Adams, M.D., Clayton, R.A., et al. (1995) Science 270, 397-403]. Although the M. genitalium gene complement is indeed the smallest among known cellular life forms, there is no evidence that it is the minimal self-sufficient gene set. To derive such a set, we compared the 468 predicted M. genitalium protein sequences with the 1703 protein sequences encoded by the other completely sequenced small bacterial genome, that of Haemophilus influenzae. M. genitalium and H. influenzae belong to two ancient bacterial lineages, i.e., Gram-positive and Gram-negative bacteria, respectively. Therefore, the genes that are conserved in these two bacteria are almost certainly essential for cellular function. It is this category of genes that is most likely to approximate the minimal gene set. We found that 240 M. genitalium genes have orthologs among the genes of H. influenzae. This collection of genes falls short of comprising the minimal set as some enzymes responsible for intermediate steps in essential pathways are missing. The apparent reason for this is the phenomenon that we call nonorthologous gene displacement when the same function is fulfilled by nonorthologous proteins in two organisms. We identified 22 nonorthologous displacements and supplemented the set of orthologs with the respective M. genitalium genes. After examining the resulting list of 262 genes for possible functional redundancy and for the presence of apparently parasite-specific genes, 6 genes were removed. We suggest that the remaining 256 genes are close to the minimal gene set that is necessary and sufficient to sustain the existence of a modern-type cell. Most of the proteins encoded by the genes from the minimal set have eukaryotic or archaeal homologs but seven key proteins of DNA replication do not. We speculate that the last common ancestor of the three primary kingdoms had an RNA genome. Possibilities are explored to further reduce the minimal set to model a primitive cell that might have existed at a very early stage of life evolution.

The C-terminal domain of the largest subunit of RNA polymerase II interacts with a novel set of serine/arginine-rich proteins.

Although transcription and pre-mRNA processing are colocalized in eukaryotic nuclei, molecules linking these processes have not previously been described. We have identified four novel rat proteins by their ability to interact with the repetitive C-terminal domain (CTD) of RNA polymerase II in a yeast two-hybrid assay. A yeast homolog of one of the rat proteins has also been shown to interact with the CTD. These CTD-binding proteins are all similar to the SR (serine/arginine-rich) family of proteins that have been shown to be involved in constitutive and regulated splicing. In addition to alternating Ser-Arg domains, these proteins each contain discrete N-terminal or C-terminal CTD-binding domains. We have identified SR-related proteins in a complex that can be immunoprecipitated from nuclear extracts with antibodies directed against RNA polymerase II. In addition, in vitro splicing is inhibited either by an antibody directed against the CTD or by wild-type but not mutant CTD peptides. Thus, these results suggest that the CTD and a set of CTD-binding proteins may act to physically and functionally link transcription and pre-mRNA processing.