Predicting splice variant from DNA chip expression data.

Alternative splicing of premessenger RNA is an important layer of regulation in eukaryotic gene expression. Splice variation of a large number of genes has been implicated in various cell growth and differentiation processes. To measure tissue-specific splicing of genes on a large scale, we collected gene expression data from 11 rat tissues using a high-density oligonucleotide array representing 1600 rat genes. Expression of each gene on the chip is measured by 20 pairs of independent oligonucleotide probes. Two algorithms have been developed to normalize and compare the chip hybridization signals among different tissues at individual oligonucleotide probe level. Oligonucleotide probes (the perfect match [PM] probe of each probe pair), detecting potential tissue-specific splice variants, were identified by the algorithms. The identified candidate splice variants have been compared to the alternatively spliced transcripts predicted by an EST clustering program. In addition, 50% of the top candidates predicted by the algorithms were confirmed by RT-PCR experiment. The study indicates that oligonucleotide probe-based DNA chip assays provide a powerful approach to detect splice variants at genome scale.

Expression profiling and identification of novel genes in hepatocellular carcinomas.

Liver cancer is the fifth most common cancer worldwide and unlike certain other cancers, such as colon cancer, a mutational model has not yet been developed. We have performed gene expression profiling of normal and neoplastic livers in C3H/HeJ mice treated with diethylnitrosamine. Using oligonucleotide microarrays, we compared gene expression in liver tumors to three different states of the normal liver: quiescent adult, regenerating adult, and newborn. Although each comparison revealed hundreds of differentially expressed genes, only 22 genes were found to be deregulated in the tumors in all three comparisons. Three of these genes were examined in human hepatocellular carcinomas and were found to be upregulated. As a second method of analysis, we used Representational Difference Analysis (RDA) to clone mRNA fragments differentially expressed in liver tumors versus regenerating livers. We cloned several novel mRNAs that are differentially regulated in murine liver tumors. Here we report the sequence of a novel cDNA whose expression is upregulated in both murine and human hepatocellular carcinomas. Our results suggest that DEN-treated mice provide an excellent model for human hepatocellular carcinomas.

Assessment of the sensitivity and specificity of oligonucleotide (50mer) microarrays.

To examine the utility and performance of 50mer oligonucleotide (oligonucleotide probe) microarrays, gene-specific oligonucleotide probes were spotted along with PCR probes onto glass microarrays and the performance of each probe type was evaluated. The specificity of oligonucleotide probes was studied using target RNAs that shared various degrees of sequence similarity. Sensitivity was defined as the ability to detect a 3-fold change in mRNA. No significant difference in sensitivity between oligonucleotide probes and PCR probes was observed and both had a minimum reproducible detection limit of approximately 10 mRNA copies/cell. Specificity studies showed that for a given oligonucleotide probe any 'non-target' transcripts (cDNAs) >75% similar over the 50 base target may show cross-hybridization. Thus non-target sequences which have >75-80% sequence similarity with target sequences (within the oligonucleotide probe 50 base target region) will contribute to the overall signal intensity. In addition, if the 50 base target region is marginally similar, it must not include a stretch of complementary sequence >15 contiguous bases. Therefore, knowledge about the target sequence, as well as its similarity to other mRNAs in the target tissue or RNA sample, is required to design successful oligonucleotide probes for quality microarray results. Together these results validate the utility of oligonucleotide probe (50mer) glass microarrays.

Yeast homolog of human SAG/ROC2/Rbx2/Hrt2 is essential for cell growth, but not for germination: chip profiling implicates its role in cell cycle regulation.

In an attempt to understand the signaling pathway mediating redox-induced apoptosis, we cloned SAG, an evolutionarily conserved zinc RING finger gene that, when overexpressed, protects cells from apoptosis induced by redox agents. Here we report functional characterization of SAG by the use of yeast genetics approach. Targeted disruption of ySAG, yeast homolog of human SAG, and subsequent tetrad analysis revealed that ySAG is required for yeast viability. Complementation experiment showed that the lethal phenotype induced by the ySAG deletion is fully rescued by wildtype SAG, but not by several hSAG mutants. Complementation experiment has also confirmed that ySAG is essential for normal vegetative growth, rather than being required for sporulation. Furthermore, cell death induced by SAG deletion was accompanied by cell enlargement and abnormal cell cycle profiling with an increased DNA content. Importantly, SAG was found to be the second family member of Rbx (RING box protein) or ROC (Regulator of cullins) or Hrt that is a component of SCF E3 ubiquitin ligase. Indeed, like ROC1/Rbx1/Hrt1, SAG binds to Cul1 and SAG-Cul1 complex has ubiquitin ligase activity to promote poly-ubiquitination of E2/Cdc34. This ligase activity is required for complementation of death phenotype induced by ySAG disruption. Finally, chip profiling of the entire yeast genome revealed induction of several G1/S as well as G2/M checkpoint control genes upon SAG withdrawal. Thus, SAG appears to control cell cycle progression in yeast by promoting ubiquitination and degradation of cell cycle regulatory proteins. Oncogene (2000) 19, 2855 - 2866

Functional similarities between HIV-1 Tat and DNA sequence-specific transcriptional activators.

The Tat regulatory protein encoded by human immunodeficiency virus type 1 (HIV-1) induces high levels of transcription from the viral long terminal repeat (LTR) promoter element after interacting with a promoter proximal RNA target sequence. In the wild-type HIV-1 LTR, this activation is facilitated by the synergistic interaction of Tat with the NF-kappa B and, particularly, SP1 regulatory proteins that bind to DNA sequences within the LTR promoter element. Using a synthetic Tat responsive indicator construct, we here demonstrate that NF-kappa B and SP1 are not uniquely or even unusually competent to synergize with HIV-1 Tat. Instead, these proteins can be functionally replaced by several, but not all, of the heterologous cellular and viral transcriptional activators tested. Tat therefore shares the ability to functionally synergize with a range of transcriptional activators, which is characteristic of DNA-sequence-specific regulatory proteins.

Mutational analysis of the transcription activation domain of RelA: identification of a highly synergistic minimal acidic activation module.

The potent C-terminal activation domain of the RelA (p65) subunit of the cellular transcription factor NF-kappa B is shown to contain several discrete acidic activation modules. These short, approximately 11-amino-acid modules were able to give rise to only a low level of transcription activation when fused to the GAL4 DNA-binding domain as monomers. However, dimers and higher-order multimers activated the transcription of minimal promoter elements as effectively as the full-length RelA or VP16 activation domain. Therefore, this 11-amino-acid RelA-derived acidic module appears to contain all of the sequence information required to fully activate a target promoter element as long as it is presented in a form that permits functional synergy. Critical primary sequence requirements for acidic activation module function included a core phenylalanine residue and flanking bulky hydrophobic residues. Overall negative charge was necessary but not sufficient for function. While dimeric forms of the 11-amino-acid acidic activation module bound to either TFIIB or TATA-binding protein efficiently in vitro, a similarly charged peptide lacking the core phenylalanine residue failed to interact. Overall, these data demonstrate that the biological activity of the RelA activation domain is dependent on acidic activator sequences that are closely comparable to those detected in the activation domain of the viral VP16 regulatory protein. We hypothesize that the ability of these acidic activators to specifically interact with multiple components of the transcription initiation complex likely underlies the dramatic functional synergy exhibited by this class of activation domains in vivo.

Sequence requirements for Rev multimerization in vivo.

Multimerization of the human immunodeficiency virus type 1 (HIV-1) Rev protein is believed to be critical to its biological activity. However, the precise protein sequence requirements for Rev multimerization in vivo, and whether multimerization is facilitated by specific RNA binding or vice versa, has remained controversial. In this report, we describe a sensitive in vivo assay for the multimerization of HIV-1 Rev on its cognate RRE primary RNA binding site. Using this assay, we demonstrate that an intact Rev arginine-rich domain, while critical to specific RNA binding, is dispensable for multimerization on the RRE. Mutations introduced into Rev sequences that flank this basic domain produce a partial multimerization phenotype in vivo even though these mutations are known to block Rev multimerization in vitro. Similarly, mutations introduced into the leucine-rich activation domain of Rev, which appear to have no effect on in vitro multimerization, also markedly inhibit multimerization of Rev on the RRE in vivo. Overall, these data appear consistent with the hypothesis that in vivo formation of the multimeric Rev:RRE ribonucleoprotein complex is facilitated by both the RRE RNA substrate and, as first proposed by Bogerd and Greene U. Virol. 67, 2496-2502, 1993), by bridging by a cellular cofactor for Rev that likely interacts with multiple Rev activation domains.

Functional analysis of interactions between Tat and the trans-activation response element of human immunodeficiency virus type 1 in cells.

Transcriptional trans-activation of the human immunodeficiency virus type 1 long terminal repeat requires that the virally encoded Tat effector interacts with its target trans-activation response element (TAR) RNA stem-loop. Although the arginine-rich region of Tat from amino acids 49 to 59 is sufficient to bind to TAR RNA in vitro, the RNA-binding domain of Tat has not been defined in vivo. Human immunodeficiency virus type 1 also encodes the Rev protein, which acts through an RNA stem-loop called the Rev-response element to transport unspliced and singly spliced viral RNA species from the nucleus to the cytoplasm. To map the RNA-binding domain of Tat, we performed assays that relied on Rev function using the heterologous RNA-tethering mechanism of Tat and the TAR. By examining the effects of selected targeted mutations of Tat on the abilities of hybrid Tat/Rev proteins to rescue the expression of unspliced mRNA via the TAR, we demonstrated that residues throughout the N-terminal 59 amino acids of Tat are required for binding of Tat and TAR RNA in vivo.

Genetic analysis of the cofactor requirement for human immunodeficiency virus type 1 Tat function.

The Tat protein of human immunodeficiency virus type 1 is a potent transcriptional trans activator of the viral long terminal repeat promoter element. Tat function requires the direct interaction of Tat with a cis-acting viral RNA target sequence termed the trans-activation response (TAR) element and has also been proposed to require at least one cellular cofactor. We have used a genetic approach to attempt to experimentally define the role of the cellular cofactor in Tat function and TAR binding. Our data suggest that neither Tat nor the cellular cofactor binds to TAR alone in vivo and indicate, instead, that the interaction of Tat with its cellular cofactor is a prerequisite for TAR binding. The known species tropism of lentivirus Tat proteins appears to arise from the fact that not only Tat but also the cellular cofactor can markedly influence the RNA sequence specificity of the resultant protein complex. These data also suggest that the Tat cofactor is likely a cellular transcription factor that has been highly conserved during vertebrate evolution. We hypothesize that the primary function of Tat is to redirect this cellular factor to a novel viral RNA target site and to thereby induce activation of viral gene expression.

The VP16 transcription activation domain is functional when targeted to a promoter-proximal RNA sequence.

Among eukaryotic transcription trans-activators, the human immunodeficiency virus type 1 (HIV-1) Tat protein is exceptional in that its target site TAR is an RNA rather than a DNA sequence. Here, we confirm that fusion of Tat to the RNA-binding domain of the HIV-1 Rev protein permits the efficient activation of an HIV-1 long terminal repeat (LTR) promoter in which critical TAR sequences have been replaced by RNA sequences derived from the HIV-1 Rev response element (RRE). An RRE target sequence as small as 13 nucleotides is shown to form an effective in vivo target for Rev binding. More important, a fusion protein consisting of Rev attached to the VP16 transcription activation domain was also observed to efficiently activate the HIV-1 LTR from this nascent RNA target. These data demonstrate that trans-activation of transcription by acidic activation domains does not require a stable interaction with the promoter DNA and suggest that VP16, like Tat, can act on steps subsequent to the formation of the HIV-1 LTR preinitiation complex. The finding that the activation domains of VP16 and Tat are functionally interchangeable raises the possibility that these apparently disparate viral trans-activators may nevertheless act via similar mechanisms.

Precursors of U4 small nuclear RNA.

The processing and ribonucleoprotein assembly of U4 small nuclear RNA has been investigated in HeLa cells. After a 45-min pulse label with [3H]uridine, a set of apparently cytoplasmic RNAs was observed migrating just behind the gel electrophoretic position of mature U4 RNA. These molecules were estimated to be one to at least seven nucleotides longer than mature U4 RNA. They reacted with Sm autoimmune patient sera and a monoclonal Sm antibody, indicating their association with proteins characteristic of small nuclear ribonucleoprotein complexes. The same set of RNAs was identified by hybrid selection of pulse-labeled RNA with cloned U4 DNA, confirming that these are U4 RNA sequences. No larger nuclear precursors of these RNAs were detected. Pulse-chase experiments revealed a progressive decrease in the radioactivity of the U4 precursor RNAs coincident with an accumulation of labeled mature U4 RNA, confirming a precursor-product relationship.

Eukaryotic small ribonucleoproteins. Anti-La human autoantibodies react with U1 RNA-protein complexes.

Anti-La sera from patients with autoimmune disorders precipitate a set of nuclear and cytoplasmic small RNA-protein complexes. Up to now, it has been thought that the La antigen is associated only with RNAs transcribed by RNA polymerase III, including precursors of tRNA and 5 S ribosomal RNA. Here we report that anti-La sera also react with ribonucleoprotein particles containing small nuclear RNA U1, which is transcribed by RNA polymerase II. Anti-La sera from 12 out of 12 patients tested were found to precipitate U1 RNA-protein complexes from HeLa cell nuclear extracts, under conditions where nonimmune sera do not. Ribonucleoprotein particles containing a second small nuclear RNA, U2, do not react appreciably with anti-La sera although they are present in HeLa cell nuclei at the same concentration as U1 RNA. Anti-La sera also react with U1 RNA-protein complexes in mouse and frog cells, but not in Drosophila or Chironomus, two organisms which lack the La antigen. Hybridization of cloned U1 DNA with anti-La-reactive RNA from HeLa cell nuclear extracts reveals mature U1 RNA, whereas anti-La-reactive cytoplasmic RNA contains a series of hybridizing bands that represent molecules 1-7 nucleotides longer than U1 and which may include precursors of nuclear U1 RNA (Madore, S. J., Wieben, E. D., and Pederson, T. (1984) J. Cell Biol., 188-192). Pulse-chase experiments suggest that the association of La antigenicity with these cytoplasmic U1 RNA molecules is transient. These results are discussed in relation to the presence of uridylate-rich sequences in the 3' termini of U1 RNA precursors and mature U1 RNA, which are similar to La antigen binding sites in several RNAs transcribed by RNA polymerase III.

Intracellular site of U1 small nuclear RNA processing and ribonucleoprotein assembly.

We have investigated the intracellular site and posttranscriptional immediacy of U1 small nuclear RNA processing and ribonucleoprotein (RNP) assembly in HeLa cells. After 30 or 45 min of labeling with [3H]uridine, a large amount of U1-related RNA radioactivity in the cytoplasm was found by using either hypotonic or isotonic homogenization buffers. The pulse-labeled cytoplasmic U1 RNA was resolved as a ladder of closely spaced bands running just behind mature-size U1 (165 nucleotides) on RNA sequencing gels, corresponding to a series of molecules between one and at least eight nucleotides longer than mature U1. They were further identified as U1 RNA sequences by gel blot hybridization with cloned U1 DNA. The ladder of cytoplasmic U1 RNA bands reacted with both RNP and Sm autoimmune sera and with a monoclonal Sm antibody, indicating a cytoplasmic assembly of these U1 RNA-related molecules into complexes containing the same antigens as nuclear U1 RNP particles. The cytoplasmic molecules behave as precursors to mature nuclear U1 RNA in both pulse-chase and continuous labeling experiments. While not excluding earlier or subsequent nuclear stages, these results suggest that the cytoplasm is a site of significant U1 RNA processing and RNP assembly. This raises the possibility that nuclear-transcribed eucaryotic RNAs are always processed in the cell compartment other than that in which they ultimately function, which suggests a set of precise signals regulating RNA and ribonucleoprotein traffic between nucleus and cytoplasm.

U1 small nuclear ribonucleoprotein studied by in vitro assembly.

The small nuclear RNAs are known to be complexed with proteins in the cell (snRNP). To learn more about these proteins, we developed an in vitro system for studying their interactions with individual small nuclear RNA species. Translation of HeLa cell poly(A)+ mRNA in an exogenous message-dependent reticulocyte lysate results in the synthesis of snRNP proteins. Addition of human small nuclear RNA U1 to the translation products leads to the formation of a U1 RNA-protein complex that is recognized by a human autoimmune antibody specific for U1 snRNP. This antibody does not react with free U1 RNA. Moreover, addition of a 10- to 20-fold molar excess of transfer RNA instead of U1 RNA does not lead to the formation of an antibody-recognized RNP. The proteins forming the specific complex with U1 RNA correspond to the A, B1, and B2 species (32,000, 27,000, and 26,000 mol wt, respectively) observed in previous studies with U1 snRNP obtained by antibody-precipitation of nuclear extracts. The availability of this in vitro system now permits, for the first time, direct analysis of snRNA-protein binding interactions and, in addition, provides useful information on the mRNAs for snRNP proteins.

Protein binding sites are conserved in U1 small nuclear RNA from insects and mammals.

To gain insight into the ribonucleoprotein (RNP) structure of small nuclear RNAs, HeLa cell poly(A)+ mRNA was translated in a reticulocyte lysate, and the in vitro binding of 35S-labeled proteins to individual small nuclear RNA species was examined by using human autoimmune antibodies. A Mr 32,000 protein binds to U1 RNA but not to U2, U4, U5, or U6. The resulting U1 RNP complex is recognized both by Sm and RNP antibodies. U2 RNA also forms a complex with protein, which is recognized by Sm antibody. Thus, the lack of binding of the Mr 32,000 protein to U2 RNA is not due to a failure of U2 to bind specific proteins in the in vitro system. Similar translation-assembly experiments with Drosophila poly(A)+ mRNA reveal that a Mr 26,000 protein identified previously in Drosophila U1 RNP [Wieben, E. D. & Pederson, T. (1982) Mol. Cell. Biol. 2, 914-920] also binds to U1 RNA in vitro. When the translation products of HeLa or Drosophila mRNA are presented with U1 RNA of the other species, the Mrs 32,000 and 26,000 proteins recognize binding sites on the heterologous U1 and, in both cases, form complexes recognized by RNP antibody. These results establish that a Mr 32,000 protein is unique to U1 RNA in human cells and that the U1 RNA binding sites for this and a Mr 26,000 homologue have been highly conserved in evolution. These sites may be the identical 13 nucleotides at the 5' ends of human and Drosophila U1 RNA or a highly conserved aspect of U1 secondary structure.