Linkage of Crohn's disease-related serological phenotypes: NFKB1 haplotypes are associated with anti-CBir1 and ASCA, and show reduced NF-kappaB activation.

BACKGROUND AND AIMS: Genetics studies of the serum expression of antibodies to microbial antigens may yield important clues to the pathogenesis of Crohn's disease. Our aim was to conduct a linkage study using expression of anti-CBir1, anti-I2, anti-OmpC and ASCA as quantitative traits. METHODS: Expression of antibodies to microbial antigens was measured by enzyme-linked immunosorbant assay (ELISA) and a standard approximately 10 cM whole genome microsatellite study was conducted. Single nucleotide polymorphism genotyping was performed using either Illumina or TaqMan MGB technology. Nuclear factor Kappa B (NF-kappaB) activation in cells from Epstein-Barr virus (EBV)-transformed cell lines was assessed using an electrophoretic mobility shift assay and protein was measured using ELISA and western blotting. RESULTS: Evidence for linkage to anti-CBir1 expression was detected on human chromosome 4 (logarithm of odds (LOD) 1.82 at 91 cM). We therefore directly proceeded to test the association of haplotypes in NFKB1, a candidate gene. One haplotype, H1, was associated with anti-CBir1 (p = 0.003) and another, H3, was associated with ASCA (p = 0.023). Using cell lines from Crohn's disease patients with either H1 or H3, NF-kappaB activation and NF-kappaB p105 and p50 production were significantly lower for patients with H1 compared to patients with H3. CONCLUSIONS: These results suggest that NFKB1 haplotypes induce dysregulation of innate immune responses by altering NF-kappaB expression. The results also show the use of EBV-transformed lymphoblastoid cell lines to conduct phenotypic studies of genetic variation.

An IFNG SNP with an estrogen-like response element selectively enhances promoter expression in peripheral but not lamina propria T cells.

This study examines mucosa-specific regulatory pathways involved in modulation of interferon-gamma (IFN-gamma) in lamina propria T cells. Previous studies identified mucosa-specific CD2 cis-elements within the -204 to -108 bp IFNG promoter. Within this region, a single-site nucleotide polymorphism, -179G/T, imparts tumor necrosis factor-alpha stimulation of IFNG in peripheral blood lymphocytes, and is linked with accelerated AIDS progression. We discovered a putative estrogen response element (ERE) introduced by the -179T, which displays selective activation in peripheral blood mononuclear cells (PBMC) vs lamina propria mononuclear cells (LPMC). Transfection of PBMC with constructs containing the -179G or -179T site revealed CD2-mediated enhancement of the -179T compared to -179G allele, although, in LPMC, a similar level of expression was detected. Electrophoretic mobility shift assay (EMSA) analysis demonstrated CD2-mediated nucleoprotein binding to the -179T but not the -179G in PBMC. In LPMC, binding is constitutive to both -179G and -179T regions. Sequence and EMSA analysis suggests that the -179T allele creates an ERE-like binding site capable of binding recombinant estrogen receptor. Estrogen response element transactivation is enhanced by CD2 signaling, but inhibited by estrogen in PBMC but not in LPMC, although expression of estrogen receptor was similar. This is the first report to describe a potential molecular mechanism responsible for selectively controlling IFN-gamma production in LPMC.

Mucosa-specific targets for regulation of IFN-gamma expression: lamina propria T cells use different cis-elements than peripheral blood T cells to regulate transactivation of IFN-gamma expression.

Activation of lamina propria (LP) T cells via the CD2 pathway enhances IFN-gamma (IFN-gamma) secretion with further enhancement after CD28 coligation. The molecular mechanisms regulating IFN-gamma expression in LP T cells remain unknown. Previous studies in PBL and T cell lines identified cis- and trans-regulatory elements in TCR-mediated expression of IFN-gamma. This study examines CD2 and PMA/ionophore-responsive IFN-gamma promoter elements. Activation of LPMC via CD2-induced IFN-gamma secretion and a parallel up-regulation of mRNA expression. CD28 coligation enhanced mRNA stability without up-regulating transcription as measured by nuclear run-on. Transfection of a -2.7-kb IFN-gamma promoter-reporter construct into PBL and LP mononuclear cells (LPMC) revealed significant promoter activity after CD2 activation, with additional transactivation after CD2/CD28 costimulation in PBL, but not in LPMC. Functional analysis using truncated promoter fragments identified distinct cis-regulatory regions selectively transactivating IFN-gamma expression in PBL compared with LPMC. In PBL, CD2 activation elements reside within the -108- to +64-bp region. However, in LPMC the upstream region between -204 and -108 bp was essential. Transfection of the proximal and distal AP-1-binding elements, as well as TRE/AP-1 constructs, revealed functional activation of AP-1 subsequent to CD2 signaling, with activation critical in PBL but diminished in LPMC. Electromobility shift analysis using oligonucleotides encompassing the proximal, distal, and BED/AP-1-binding regions failed to demonstrate selective transactivation after CD2 signaling of LPMC. This report provides evidence that activation of LPMC results in transactivation of multiple promoter elements regulating IFN-gamma expression distinct from those in PBL.

Involvement of protein kinase Cepsilon (PKCepsilon) in thyroid cell death. A truncated chimeric PKCepsilon cloned from a thyroid cancer cell line protects thyroid cells from apoptosis.

The protein kinase C (PKC) family has been implicated in the regulation of apoptosis. However, the contribution of individual PKC isozymes to this process is not well understood. We reported amplification of the chromosome 2p21 locus in 28% of thyroid neoplasms, and in the WRO thyroid carcinoma cell line. By positional cloning we identified a rearrangement and amplification of the PKCepsilon gene, that maps to 2p21, in WRO cells. This resulted in the overexpression of a chimeric/truncated PKCepsilon (Tr-PKCepsilon) mRNA, coding for N-terminal amino acids 1-116 of the isozyme fused to an unrelated sequence. Expression of the Tr-PKCepsilon protein in PCCL3 cells inhibited activation-induced translocation of endogenous PKCepsilon, but its kinase activity was unaffected, consistent with a dominant negative effect of the mutant protein on activation-induced translocation of wild-type PKCepsilon and/or displacement of the isozyme to an aberrant subcellular location. Cell lines expressing Tr-PKCepsilon grew to a higher saturation density than controls. Moreover, cells expressing Tr-PKCepsilon were resistant to apoptosis, which was associated with higher Bcl-2 levels, a marked impairment in p53 stabilization, and dampened expression of Bax. These findings point to a role for PKCepsilon in apoptosis-signaling pathways in thyroid cells, and indicate that a naturally occurring PKCepsilon mutant that functions as a dominant negative can block cell death triggered by a variety of stimuli.

CD28 costimulation augments IL-2 secretion of activated lamina propria T cells by increasing mRNA stability without enhancing IL-2 gene transactivation.

The pathways leading to activation in lamina propria (LP) T cells are different from peripheral T cells. LP T cells exhibit enhanced IL-2 secretion when activated through the CD2 pathway. Coligation of CD28 leads to synergistic enhancement of IL-2 secretion. Previous studies have characterized the CD28 augmentation of TCR-mediated signaling in peripheral blood T cells through transcriptional activation of an IL-2 promoter CD28 response element (CD28RE), along with enhanced mRNA stability. This study characterized molecular events involved in CD28 costimulation of IL-2 production in LP mononuclear cells (LPMC). LPMC exhibited increased IL-2 production in response to CD28 costimulation, compared with cells activated through CD2 alone. IL-2 secretion was paralleled by increased expression of IL-2 mRNA, resulting from enhanced IL-2 mRNA stability. In contrast to transcriptional activation in PBMC, EMSA revealed that CD28 coligation of CD2-activated LPMC does not result in increased binding of trans-factors to the CD28RE, nor did Western blots detect changes in I-kappaBalpha or I-kappaBbeta levels following CD28 coligation. Furthermore, CD28 coligation fails to enhance IL-2 promoter-reporter or RE/AP construct expression in CD2-activated LPMC. The results reported herein indicate that the molecular mechanisms involved in CD28 cosignaling and regulation of IL-2 secretion in LP T cells are unique to that compartment and differ from those seen in peripheral blood T cells. These observations suggest a biological significance for different mechanisms of IL-2 activation in initiation and maintenance of the cytokine repertoire found in the mucosa.

From amplification to gene in thyroid cancer: a high-resolution mapped bacterial-artificial-chromosome resource for cancer chromosome aberrations guides gene discovery after comparative genome hybridization.

Chromosome rearrangements associated with neoplasms provide a rich resource for definition of the pathways of tumorigenesis. The power of comparative genome hybridization (CGH) to identify novel genes depends on the existence of suitable markers, which are lacking throughout most of the genome. We now report a general approach that translates CGH data into higher-resolution genomic-clone data that are then used to define the genes located in aneuploid regions. We used CGH to study 33 thyroid-tumor DNAs and two tumor-cell-line DNAs. The results revealed amplifications of chromosome band 2p21, with less-intense amplification on 2p13, 19q13.1, and 1p36 and with least-intense amplification on 1p34, 1q42, 5q31, 5q33-34, 9q32-34, and 14q32. To define the 2p21 region amplified, a dense array of 373 FISH-mapped chromosome 2 bacterial artificial chromosomes (BACs) was constructed, and 87 of these were hybridized to a tumor-cell line. Four BACs carried genomic DNA that was amplified in these cells. The maximum amplified region was narrowed to 3-6 Mb by multicolor FISH with the flanking BACs, and the minimum amplicon size was defined by a contig of 420 kb. Sequence analysis of the amplified BAC 1D9 revealed a fragment of the gene, encoding protein kinase C epsilon (PKCepsilon), that was then shown to be amplified and rearranged in tumor cells. In summary, CGH combined with a dense mapped resource of BACs and large-scale sequencing has led directly to the definition of PKCepsilon as a previously unmapped candidate gene involved in thyroid tumorigenesis.

Activation of the CD2 pathway in lamina propria T cells up-regulates functionally active AP-1 binding to the IL-2 promoter, resulting in messenger RNA transcription and IL-2 secretion.

The aim of this study was to identify molecular mechanisms involved in transcriptional regulation of IL-2 expression following CD2 and CD3 activation in lamina propria (LP) T cells. Studies used T cells from normal, ulcerative colitis, and Crohn's disease mucosa and freshly isolated PBMC, PBMC stimulated with IL-2 alone, and PBMC stimulated with IL-2 and cocultured with B cell lines (LP-like T cells). Electrophoretic mobility shift assays were performed with nuclear extracts from cells activated with either anti-CD2 or anti-CD3 Abs. CD2 signaling in LPMC and LP-like T cells led to a pattern of sustained up-regulation of AP-1-binding complexes, whereas CD3 activation resulted in only transient up-regulation. While the pattern of regulation of AP-1 binding observed in normal, uninflamed, or inflamed Crohn's disease LPMC is similar, differences in intensity of AP-1 binding were observed. Activation of LP-like T cells mimics the up-regulation of AP-1 with a kinetic profile similar to that observed with freshly isolated LPMC from Crohn's disease-inflamed tissue. The AP-1 complex formed following CD2 activation is composed of jun/fos heterodimers. The CD2-enhanced responsiveness is reflected in functional analysis experiments utilizing transfection of both multimeric-TRE or IL-2 promoter-luciferase constructs directly into normal, ulcerative colitis, or Crohn's disease LPMC. Our data suggest that activation of LP T cells from normal, ulcerative colitis, or Crohn's disease mucosa through the CD2 pathway leads to induction of AP-1 complexes that bind to the IL-2 promoter, and may play a pivotal role in modulating IL-2 production in the gut.

Identification of rapid turnover transcripts overexpressed in thyroid tumors and thyroid cancer cell lines: use of a targeted differential RNA display method to select for mRNA subsets.

The mRNAs of transiently expressed proteins such as cytokines and proto-oncogenes are commonly subject to rapid transcriptional activation and degradation. Transcript turnover is determined in part by association of certain proteins with consensus AU-rich motifs (AUUUA) in the 3'-untranslated region of the transcripts. Here we report a modification of differential RNA display (DRD) to detect differentially expressed rapid turnover mRNAs containing AU-rich motifs from thyroid cancer tissues and cell lines. RNA of normal and thyroid cancer tissues was differentially displayed using a 3'anchor primer to the poly(A) tail and an arbitrary 5'primer incorporating an AUUUA sequence. The appropriateness of the strategy was established by its ability to display known early response genes, such as c- fos, using partially degenerate primers. To test whether the novel cDNAs isolated coded for transcripts subject to rapid turnover, they were used as probes for Northern blots of RNA from clonal human thyroid carcinoma cell lines treated for varying periods with either cycloheximide or actinomycin D. A number of novel differentially expressed cDNA fragments were isolated from human papillary thyroid carcinoma tissues, among them a cDNA with zinc finger motifs and homology to other zinc finger proteins. Using this fragment to probe a cDNA library, a full-length cDNA (ZnF20) was isolated that was 4333 bp in length and contained an open reading frame of 1029 amino acids. The ZnF20 cDNA hybridized to multiple transcripts in a thyroid cancer cell line (8.0, 4.5 and 2 kb) that increased after cycloheximide treatment and decayed <2 h after addition of actinomycin D. The ZnF20 mRNA was overexpressed in three of six thyroid papillary carcinomas as compared with paired normal thyroid tissue controls. The data presented here support the use of a targeted DRD approach for the isolation of rapid turnover mRNAs, many of which may be interesting candidate oncogenes.

Reexpression of thyroid peroxidase in a derivative of an undifferentiated thyroid carcinoma cell line by introduction of wild-type p53.

Loss of function of p53 is believed to result in transformation through impairment of its properties as a transcription factor, which interferes with the regulation of the cell cycle and under certain conditions, with programmed cell death. We report that stable transfection of clonal undifferentiated thyroid carcinoma cell lines harboring endogenous p53 mutations with a wild-type p53 expression vector only rarely yields transfectants expressing authentic wild-type p53. Among these, most exhibited an increase in doubling time and an impairment of colony formation in soft agar. Only one clonal wild-type p53-overexpressing derivative of the NPA papillary carcinoma cell line was obtained, and these cells were found to reexpress thyroid peroxidase (TPO). This clone also demonstrated reexpression of the paired box domain transcription factor Pax-8, which specifically activates transcription of TPO. Wild-type p53 did not directly stimulate transcriptional activity of a TPO promoter construct. Although the low frequency of authentic wild-type p53 stable transfectants limits the power of this analysis, these data suggest that in addition to its role in malignant transformation, p53 may be significant in the determination or maintenance of cell differentiation in thyroid neoplasms.

Mutations of the adenomatous polyposis coli gene in sporadic thyroid neoplasms.

Several epidemiological studies have demonstrated an association between familial adenomatous polyposis coli (FAP) and thyroid neoplasms. Predisposition to FAP is conferred by mutations in the APC gene, located on chromosome 5q21. Somatic mutations of APC are also observed in about 60% of sporadic colorectal adenomas and carcinomas, suggesting that disruption of this putative tumor suppressor gene may play a role in both familial as well as acquired colorectal tumorigenesis. The APC gene is expressed in normal human thyroid, thyroid adenomas, and differentiated carcinoma tissues as well as in four clonal human thyroid carcinoma cell lines, as demonstrated by reverse transcriptase-polymerase chain reaction of a 388-base APC messenger ribonucleic acid fragment spanning exons 14 and 15, followed by hybridization to an exon 15-specific complementary DNA probe. Eighty human thyroid neoplasms were examined for loss of heterozygosity of the APC locus, using primers flanking a hypervariable dinucleotide (CA) repeat (CB26) immediately adjacent to the APC gene. Of 71% informative samples, 2 showed allelic loss: a follicular adenoma (FA) and a nodule from a multinodular goiter (MNG). The DNA of 83 benign and malignant thyroid neoplasms and 4 thyroid carcinoma cell lines was examined for mutations within a 1200-basepair stretch of exon 15 by single strand conformation polymorphism. Five sets of overlapping primers were used for PCR. The anaplastic thyroid carcinoma cell line (ARO) had 1 APC allele with an adenine insertion at codon 1556 (ACTA to AACTA), leading to a premature stop codon at 1558. An anaplastic carcinoma had a mutation of codon 1346 (TCA-CCA; Ser to Pro). In summary, the APC gene is expressed in normal and neoplastic human thyroid tissue and is a target for inactivating mutations in some thyroid tumors.

Down syndrome phenotypes: the consequences of chromosomal imbalance.

Down syndrome (DS) is a major cause of mental retardation and congenital heart disease. Besides a characteristic set of facial and physical features, DS is associated with congenital anomalies of the gastrointestinal tract, an increased risk of leukemia, immune system defects, and an Alzheimer-like dementia. Moreover, DS is a model for the study of human aneuploidy. Although usually caused by the presence of an extra chromosome 21, subsets of the phenotypic features of DS may be caused by the duplication of small regions of the chromosome. The physical map of chromosome 21 allows the molecular definition of the regions duplicated in these rare cases of partial trisomy. As a first step in identifying the genes responsible for individual DS features and their pathophysiology, a panel of cell lines derived from 16 such individuals has been established and the molecular break points have been determined using fluorescence in situ hybridization and Southern blot dosage analysis of 32 markers unique to human chromosome 21. Combining this information with detailed clinical evaluations of these patients, we have now constructed a "phenotypic map" that includes 25 features and assigns regions of 2-20 megabases as likely to contain the genes responsible. This study provides evidence for a significant contribution of genes outside the D21S55 region to the DS phenotypes, including the facies, microcephaly, short stature, hypotonia, abnormal dermatoglyphics, and mental retardation. This strongly suggests DS is a contiguous gene syndrome and augurs against a single DS chromosomal region responsible for most of the DS phenotypic features.

Molecular screening of pituitary adenomas for gene mutations and rearrangements.

Although pituitary tumors arise as benign monoclonal neoplasms, genetic alterations have not readily been identified in these adenomas. We studied restriction fragment abnormalities involving the GH gene locus, and mutations in the p53 and H-, K- and N-ras genes in 22 human GH cell adenomas. Twenty two nonsecretory adenomas were also examined for p53 and ras gene mutations. Seven prolactinoma DNA samples were tested for deletions in the multiple endocrine neoplasia-1 (MEN-1) locus, as well as for rearrangements in the hst gene, a member of the fibroblast growth factor family. Pituitary adenoma tissue and lymphocytes were obtained from patients at the time of transsphenoidal surgery. In DNA from GH-cell adenomas, identical GH restriction patterns were detected in both pituitary and lymphocyte DNA in all patients and in one patient with a mixed GH-TSH cell adenoma. Using polymerase chain reaction (PCR)-single stranded conformation polymorphism analysis, no mutations were detected in exons 5, 6, 7, and 8 of the p53 gene in GH cell adenomas nor in 22 nonsecretory adenomas. Codons 12/13 and 61 of H-ras, K-ras, and N-ras genes were also intact in GH cell adenomas and in nonsecretory adenomas. Site-specific probes for chromosome 11q13 including PYGM, D11S146, and INT2 were used in 7 sporadic PRL-secreting adenomas to detect deletions of the MEN-1 locus on chromosome 11. One patient was identified with a loss of 11p, and the remaining 6 patients did not demonstrate loss of heterozygosity in the pituitary 11q13 locus, compared to lymphocyte DNA. None of these patients, demonstrated hst gene rearrangements which also maps to this locus. These results show that p53 and ras gene mutations are not common events in the pathogenesis of acromegaly and nonsecretory tumors. Although hst gene rearrangements and deletions of 11q13 are not associated with sporadic PRL-cell adenoma formation, a single patient was detected with a partial loss of chromosome 11, including the putative MEN-1 site.

Binding of ATP and messenger RNA by the beta-subunit of eukaryotic initiation factor 2.

In addition to forming a ternary complex with Met-tRNA(f) and GTP, eukaryotic initiation factor 2 (eIF-2) recognizes a specific site in mRNA molecules. Both binding activities are regulated by ATP, which itself binds tightly and specifically to eIF-2. Denaturation of eIF-2 with urea leads to complete loss of Met-tRNA(f) binding activity, while mRNA binding activity is stable. Hence, distinct conformational features in eIF-2 are required for ternary complex formation and for binding of mRNA. Chromatography of eIF-2 over ATP-agarose, in denaturing conditions that induce polypeptide subunit dissociation, results in selective retention of the beta-subunit of eIF-2. Isolated beta-subunit is capable of binding mRNA as well as ATP. Cibacron blue 3G-A binds tightly to eIF-2 and inhibits the binding of mRNA. This inhibition is relieved upon addition of ATP, showing that Cibacron blue 3G-A competes with ATP for eIF-2. eIF-2 beta subunit, active in binding of mRNA, is recovered upon chromatography of eIF-2 in denaturing conditions over matrix-bound Cibacron blue 3G-A. These results show that the ability of eIF-2 to bind mRNA and its ability to bind ATP are both lodged within remarkably stable domains of its beta-subunit. During initiation of protein synthesis, the eIF-2 beta subunit may thus interact with three ligands important for translational control: Met-tRNA(f), mRNA and ATP.

Transforming DNA sequences present in human prolactin-secreting pituitary tumors.

Five human PRL-secreting pituitary tumors were tested for the presence of DNA-transforming sequences. After calcium phosphate transfection to NIH-3T3 mouse fibroblast cells, DNA samples derived from two prolactinomas induced foci of morphologically transformed cells which subsequently grew in soft agar. After retransfection of transformant DNA, resulting secondary transformants elicited rapidly growing solid tumors in nude mice. Southern analysis of transformant DNA revealed the integration of Alu-positive human DNA sequences into the mouse fibroblast NIH-3T3 cells, as judged by hybridization to a Blur-8 probe. The Alu signal became increasingly more difficult to detect with the multiple passaging (greater than 20) of transformant cells in culture. Alu polymerase chain reaction (PCR) was, therefore, used to selectively amplify human DNA sequences from the NIH-3T3 rodent background. PCR using a human Alu-specific primer resulted in amplification of an Alu-containing DNA region within these transformants. The transformant DNA did not hybridize to human genomic probes for genes known to evoke focus formation in this assay, including H-ras, K-ras, N-ras, trk, ret, ros, or met. Further identification of the Alu-containing region revealed that it contained sequences from the human hst gene, a member of the fibroblast growth factor family. The presence of human hst was demonstrated by strong hybridization to a 40-mer oligonucleotide probe to the second exon of hst, by amplification of this region with human hst-nested amplimers within the first and second introns, and finally by direct sequencing.(ABSTRACT TRUNCATED AT 250 WORDS)

Clonal origin of pituitary adenomas.

As the pathogenesis of pituitary adenomas remains unclear, the tumor clonal composition of these common neoplasms was studied. Clonality was determined in female patients by analysis of restriction fragment length polymorphisms of the X-chromosome genes hypoxanthine phosphoribosyl transferase and phosphoglycerate kinase in conjunction with their respective methylation patterns. Peripheral lymphocyte DNA was screened from 62 female patients undergoing transsphenoidal surgery for pituitary adenoma. Eleven patients were heterozygous for the BglI site on PGK, 4 for the BamHI site on HPRT, and 1 patient for both sites. Of these 16 patients, 3 had acromegaly, 4 had Cushing's disease, 7 had hyperprolactinemia, and 2 were clinically nonfunctional. After surgery, morphological study, including immunohistochemistry and electron microscopy of the pathological specimens, allowed a direct comparison between clonality and tumor cell type. Control fresh normal pituitary tissue was found to be polyclonal. The following tumors were monoclonal: all 3 somatotroph adenomas, 4 of 4 lactotroph tumors, 3 of 4 corticotroph cell adenomas, a gonadotroph adenoma, and a nonsecretory adenoma. A mixed plurihormonal adenoma was polyclonal, as were 2 tumors consisting of adenomatous lactotrophs interspersed with nontumorous adenohypophyseal pituitary tissue and one corticotroph adenoma mixed with normal pituitary tissue. Functional pituitary adenomas derived from somatotrophs, corticotrophs, or lactotrophs and nonsecretory tumors are monoclonal in nature, suggesting that somatic cell mutations precede clonal expansion of these cells and play a major role in pituitary tumorigenesis.

Binding of ATP to eukaryotic initiation factor 2. Differential modulation of mRNA-binding activity and GTP-dependent binding of methionyl-tRNAMetf.

Eukaryotic initiation factor 2 (eIF-2) is shown to bind ATP with high affinity. Binding of ATP to eIF-2 induces loss of the ability to form a ternary complex with Met-tRNAf and GTP, while still allowing, and even stimulating, the binding of mRNA. Ternary complex formation between eIF-2, GTP, and Met-tRNAf is inhibited effectively by ATP, but not by CTP or UTP. Hydrolysis of ATP is not required for inhibition, for adenyl-5'-yl imidodiphosphate (AMP-PNP), a nonhydrolyzable analogue of ATP, is as active an inhibitor; adenosine 5'-O-(thiotriphosphate) (ATP gamma S) inhibits far more weakly. Ternary complex formation is inhibited effectively by ATP, dATP, or ADP, but not by AMP and adenosine. Hence, the gamma-phosphate of ATP and its 3'-OH group are not required for inhibition, but the beta-phosphate is indispensible. Specific complex formation between ATP and eIF-2 is shown 1) by effective retention of Met-tRNAf- and mRNA-binding activities on ATP-agarose and by the ability of free ATP, but not GTP, CTP, or UTP, to effect elution of eIF-2 from this substrate; 2) by eIF-2-dependent retention of [alpha-32P]ATP or dATP on nitrocellulose filters and its inhibition by excess ATP, but not by GTP, CTP, or UTP. Upon elution from ATP-agarose by high salt concentrations, eIF-2 recovers its ability to form a ternary complex with Met-tRNAf and GTP. ATP-induced inhibition of ternary complex formation is relieved by excess Met-tRNAf, but not by excess GTP or guanyl-5'-yl imidodiphosphate (GMP-PNP). Thus, ATP does not act by inhibiting binding of GTP to eIF-2. Instead, ATP causes Met-tRNAf in ternary complex to dissociate from eIF-2. Conversely, affinity of eIF-2 for ATP is high in the absence of GTP and Met-tRNAf (Kd less than or equal to 10(-12) M), but decreases greatly in conditions of ternary complex formation. These results support the concept that eIF-2 assumes distinct conformations for ternary complex formation and for binding of mRNA, and that these are affected differently by ATP. Interaction of ATP with an eIF-2 molecule in ternary complex with Met-tRNAf and GTP promotes displacement of Met-tRNAf from eIF-2, inducing a state favorable for binding of mRNA. ATP may thus regulate the dual binding activities of eIF-2 during initiation of translation.

Relief of cytotoxicity and enhancement of interferon inducer activity of double-stranded RNA by eukaryotic initiation factor 2.

Double-stranded RNA (dsRNA) is a powerful interferon inducer but also possesses toxic properties. dsRNA was shown previously to inhibit translation by causing the inactivation of eukaryotic initiation factor 2 (eIF-2) and to bind tightly to this protein. The cytotoxicity of dsRNA was analyzed simultaneously with the induction of interferon in murine fibroblast cultures. Incubation of dsRNA with eIF-2 leads to a significant reduction in toxicity, concomitant with a marked stimulation of interferon induction. The enhancement of dsRNA-dependent interferon induction by eIF-2 is sensitive to a monoclonal antibody directed against the beta-subunit of eIF-2. When injected in combination with dsRNA, eIF-2 potentiates the dsRNA-dependent survival of mice infected with a lethal dose of Mengo virus.

The nature of the interaction of eukaryotic initiation factor 2 with double-stranded RNA.

In addition to binding messenger RNA molecules at specific sequences, eukaryotic initiation factor 2 (eIF-2) also binds to double-stranded RNA (dsRNA). The dsRNA is a powerful inhibitor of initiation of eukaryotic translation, causing the inactivation of eIF-2, but in the presence of certain mRNA templates, dsRNA fails to establish inhibition. Such mRNA templates bind to eIF-2 with higher affinity than does dsRNA, while globin mRNA, a template sensitive to inhibition, binds with lower affinity. Here, the nature of the interaction between dsRNA and eIF-2 was studied by examining both the binding of eIF-2 to Penicillium chrysogenum dsRNA molecules carrying 32P label at their 5' ends, and the ability of eIF-2 to protect such label against pancreatic ribonuclease digestion. The results reveal binding sites for eIF-2 at the 5' ends, as well as throughout internal regions of the dsRNA molecule. At least 15 molecules of eIF-2 can be accommodated on a 3000-base molecule of P. chrysogenum dsRNA. eIF-2 protects a 105-base-pair 5'-terminal fragment in dsRNA against digestion, but exhibits no noticeable preference for the 5' ends. By contrast, eIF-2 fails to protect label at the 5' ends of denatured dsRNA molecules, even though it binds to them at internal sites more avidly than to native dsRNA. Binding of eIF-2 to dsRNA is not restricted to specific sequences: eIF-2 binds with equal affinity to the synthetic dsRNA sequence, poly(rI . rC). The data support the interpretation that eIF-2 recognizes the A conformation in dsRNA rather than sequence. Apparently binding of eIF-2 at sites spaced 200 base pairs apart prevents relaxation of the intervening length of the double helix, thereby stabilizing the dsRNA molecule against ribonuclease attack. These results show that, even though dsRNA and mRNA compete in their binding to eIF-2, the structural features recognized by eIF-2 in these RNA species are distinct.