Sequence databases: integrated information retrieval and data submission.

This unit describes the NCBI's Entrez database browser. Entrez integrates DNA and protein sequence data, three dimensional structures, and taxonomic information with its associated abstracts and citations contained in PubMed (MEDLINE). It is possible to search the Entrez information space using conventional search queries (authors, gene names, map location) as well as by bibliographic associations (articles that are related to one another) and sequence homology. Also described are the procedures for submission of new data, updates, and corrections to the sequence databases.

Sequence databases: integrated information retrieval and data submission.

This unit provides an overview of biomedical information resources, focusing on sequence data, structure information, and the associated literature, and also discusses how nucleotide sequence data gets into the databases in the first place. Some specific databases covered here are MEDLINE, GenBank, and Entrez.

An effective approach for analyzing "prefinished" genomic sequence data.

Ongoing efforts to sequence the human genome are already generating large amounts of data, with substantial increases anticipated over the next few years. In most cases, a shotgun sequencing strategy is being used, which rapidly yields most of the primary sequence in incompletely assembled sequence contigs ("prefinished" sequence) and more slowly produces the final, completely assembled sequence ("finished" sequence). Thus, in general, prefinished sequence is produced in excess of finished sequence, and this trend is certain to continue and even accelerate over the next few years. Even at a prefinished stage, genomic sequence represents a rich source of important biological information that is of great interest to many investigators. However, analyzing such data is a challenging and daunting task, both because of its sheer volume and because it can change on a day-by-day basis. To facilitate the discovery and characterization of genes and other important elements within prefinished sequence, we have developed an analytical strategy and system that uses readily available software tools in new combinations. Implementation of this strategy for the analysis of prefinished sequence data from human chromosome 7 has demonstrated that this is a convenient, inexpensive, and extensible solution to the problem of analyzing the large amounts of preliminary data being produced by large-scale sequencing efforts. Our approach is accessible to any investigator who wishes to assimilate additional information about particular sequence data en route to developing richer annotations of a finished sequence.

A transcript map for the 2.8-Mb region containing the multiple endocrine neoplasia type 1 locus.

Multiple endocrine neoplasia type 1 (MEN 1) is an inherited cancer syndrome in which affected individuals develop multiple parathyroid, enteropancreatic, and pituitary tumors. The locus for MEN1 is tightly linked to the marker PYGM on chromosome 11q13, and linkage analysis places the MEN1 gene within a 2-Mb interval flanked by the markers D11S1883 and D11S449. Loss of heterozygosity studies in MEN 1 and sporadic tumors suggest that the MEN1 gene encodes a tumor suppressor and have helped to narrow the location of the gene to a 600-kb interval between PYGM and D11S449. Focusing on this smaller MEN1 interval, we have identified and mapped 12 transcripts to this 600-kb region. A precise ordered map of 33 transcripts, including 12 genes known to map to this region, was generated for the 2.8-Mb D11S480-D11S913 interval. Fifteen candidate genes (of which 10 were examined exhaustively) were evaluated by Southern blot and/or dideoxy fingerprinting analysis to identify the gene harboring disease-causing mutations.

A 2.8-Mb clone contig of the multiple endocrine neoplasia type 1 (MEN1) region at 11q13.

Multiple endocrine neoplasia type 1 (MEN1) is an autosomal dominant disorder that results in parathyroid, anterior pituitary, and pancreatic and duodenal endocrine tumors in affected individuals. The MEN1 locus is tightly linked to the marker PYGM on chromosome 11q13, and linkage analysis has placed the MEN1 gene within a 2-Mb interval flanked by D11S1883 and D11S449. As a step toward cloning the MEN1 gene, we have constructed a 2.8-Mb clone contig consisting of YAC and bacterial clones (PAC, BAC, and P1) for the D11S480 to D11S913 region. The bacterial clones alone represent nearly all of the 2.8-Mb contig. The contig was assembled based on a high-density STS-content analysis of 79 genomic clones (YAC, PAC, BAC, and P1) with 118 STSs. The STSs included 22 polymorphic markers and 20 transcripts, with the remainder primarily derived from the end sequences of the genomic clones. An independent cosmid contig for the 1-Mb PYGM-SEA region was also generated. Support for correctness of the 2.8-Mb contig map comes from an independent ordering of the clones by fiber-FISH. This sequence-ready contig will be a useful resource for positional cloning of MEN1 and other disease genes whose loci fall within this region.

Characterization and expression of NAD(H)-dependent glutamate dehydrogenase genes in Arabidopsis.

Two distinct cDNA clones encoding NAD(H)-dependent glutamate dehydrogenase (NAD[H]-GDH) in Arabidopsis thaliana were identified and sequenced. The genes corresponding to these cDNA clones were designated GDH1 and GDH2. Analysis of the deduced amino acid sequences suggest that both gene products contain putative mitochondrial transit polypeptides and NAD(H)- and alpha-ketoglutarate-binding domains. Subcellular fractionation confirmed the mitochondrial location of the NAD(H)-GDH isoenzymes. In addition, a putative EF-hand loop, shown to be associated with Ca2+ binding, was identified in the GDH2 gene product but not in the GDH1 gene product. GDH1 encodes a 43.0-kD polypeptide, designated alpha, and GDH2 encodes a 42.5-kD polypeptide, designated beta. The two subunits combine in different ratios to form seven NAD(H)-GDH isoenzymes. The slowest-migrating isoenzyme in a native gel, GDH1, is a homohexamer composed of alpha subunits, and the fastest-migrating isoenzyme, GDH7, is a homohexamer composed of beta subunits. GDH isoenzymes 2 through 6 are heterohexamers composed of different ratios of alpha and beta subunits. NAD(H)-GDH isoenzyme patterns varied among different plant organs and in leaves of plants irrigated with different nitrogen sources or subjected to darkness for 4 d. Conversely, there were little or no measurable changes in isoenzyme patterns in roots of plants treated with different nitrogen sources. In most instances, changes in isoenzyme patterns were correlated with relative differences in the level of alpha and beta subunits. Likewise, the relative difference in the level of alpha or beta subunits was correlated with changes in the level of GDH1 or GDH2 transcript detected in each sample, suggesting that NAD(H)-GDH activity is controlled at least in part at the transcriptional level.

Eighteen new polymorphic markers in the multiple endocrine neoplasia type 1 (MEN1) region.

Multiple endocrine neoplasia type 1 (MEN1) is an autosomal dominant disorder in which affected individuals develop tumors primarily in the parathyroids, anterior pituitary, endocrine pancreas, and duodenum. The locus for MEN1 is tightly linked to the marker PYGM on chromosome 11q13, and linkage analysis has previously placed the MEN1 gene within a 2-Mb interval flanked by markers D11S1883 and D11S449. Loss of heterozygosity (LOH) studies in MEN1 and sporadic tumors have helped narrow the location of the gene to a 600-kb interval between PYGM and D11S449. Eighteen new polymerase chain reaction (PCR)-based polymorphic markers were generated for the MEN1 region, with ten mapping to the PYGM-D11S449 interval. These new markers, along with 14 previously known polymorphic markers, were precisely mapped on a 2.8-Mb (D11S480-D11S913) high-density clone contig-based, physical map generated for the MEN1 region.

Identification and expression of a cDNA from Daucus carota encoding a bifunctional aspartokinase-homoserine dehydrogenase.

Aspartokinase (EC 2.7.2.4) and homoserine dehydrogenase (EC 1.1.1.3) catalyze steps in the pathway for the synthesis of lysine, threonine, and methionine from aspartate. Homoserine dehydrogenase was purified from carrot (Daucus carota L.) cell cultures and portions of it were subjected to amino acid sequencing. Oligonucleotides deduced from the amino acid sequences were used as primers in a polymerase chain reaction to amplify a DNA fragment using DNA derived from carrot cell culture mRNA as template. The amplification product was radiolabelled and used as a probe to identify cDNA clones from libraries derived from carrot cell culture and root RNA. Two overlapping clones were isolated. Together the cDNA clones delineate a 3089 bp long sequence encompassing an open reading frame encoding 921 amino acids, including the mature protein and a long chloroplast transit peptide. The deduced amino acid sequence has high homology with the Escherichia coli proteins aspartokinase I-homoserine dehydrogenase I and aspartokinase II-homoserine dehydrogenase II. Like the E. coli genes the isolated carrot cDNA appears to encode a bifunctional aspartokinase-homoserine dehydrogenase enzyme.

A genetic map of soybean (Glycine max L.) using an intraspecific cross of two cultivars: 'Minosy' and 'Noir 1'.

Genetic markers were mapped in segregating progeny from a cross between two soybean (Glycine max (L.) Merr.) cultivars: 'Minsoy' (PI 27.890) and 'Noir 1' (PI 290.136). A genetic linkage map was constructed (LOD [Symbol: see text] 3), consisting of 132 RFLP, isozyme, morphological, and biochemical markers. The map defined 1550cM of the soybean genome comprising 31 linkage groups. An additional 24 polymorphic markers remained unlinked. A family of RFLP markers, identified by a single probe (hybridizing to an interspersed repeated DNA sequence), extended the map, linking other markers and defining regions for which other markers were not available.

Identification and expression of a cDNA clone encoding aspartate aminotransferase in carrot.

A full-length cDNA clone encoding aspartate aminotransferase (AAT) has been identified from a carrot root cDNA library. Degenerate oligo primers were synthesized from the known amino acid sequence of AAT form I from carrot (Daucus carota L. cv Danvers). These primers were utilized in a polymerase chain reaction to amplify a portion of a carrot AAT gene from first strand cDNA synthesized from poly(A)(+) RNA isolated from 5-d-old cell suspension cultures. The resulting 750-bp fragment was cloned, mapped, and sequenced. The cloned fragment, mpAAT1, was used as a probe to identify a full-length cDNA clone in a library constructed from poly(A)(+) RNA isolated from carrot roots. A 1.52-kb full-length clone, AAT7, was isolated and sequenced. AAT7 has 54% nucleotide identity with both the mouse cytoplasmic and mitochondrial AAT genes. The deduced amino acid sequence has 52 and 53% identity with the deduced amino acid sequences of mouse cytoplasmic and mitochondrial AAT genes, respectively. Further analysis of the sequence data suggests that AAT7 encodes a cytoplasmic form of carrot AAT; the evidence includes the (a) absence of a transit or signal sequence, (b) lack of "m-residues," or invariant mitochondrial residues, in the carrot AAT sequence, and (c) high degree of sequence similarity with the amino acid sequence previously obtained for form I of carrot, a cytoplasmic isoenzyme. High- and low-stringency hybridizations to Southern blots of carrot nuclear DNA with AAT7 show that AAT is part of a small multigene family. Northern blot analysis of AAT7 suggests that AAT is expressed throughout cell culture up to 7 d and is highly expressed in roots but not in leaves.

Isoforms of soybean seed oil body membrane protein 24 kDa oleosin are encoded by closely related cDNAs.

We have characterized two cDNA clones for 24 kDa soybean oleosin, the seed oil body membrane protein. Differences in the predicted amino acid sequences of the two clones and the presence of a doublet on immunoblots indicate that 24 kDa oleosin exists in at least two isoforms in soybean. The predicted amino acid sequence also contains a unique carboxy terminal region that is dominated by a series of different tandem amino acid repeats.

The promoter of the recA gene of Escherichia coli.

The growth defect of a lambda phage carrying a recA-lacZ fusion was used to select mutations that reduced recA expression. Nine single base changes in the recA promoter were isolated that reduced both induced and basal (repressed) levels of expression. Deletion analysis of the promoter region and mapping of transcripts indicated that there is one main promoter responsible for both basal and induced expression. Some of the mutants displayed a lowered induction ratio, raising the possibility that there is a second, weak promoter that is not regulated by the SOS response. When one of the mutants was examined, it showed normal affinity for LexA repressor binding to the operator site. Binding of RNA polymerase to this mutant promoter, however, was much reduced. Further binding experiments suggested that LexA does not block RNA polymerase binding to the recA promoter, but inhibits a later step in initiation.

Molecular cloning of a protein associated with soybean seed oil bodies that is similar to thiol proteases of the papain family.

A 34,000-Da protein (P34) is one of the four major soybean oil body proteins observed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis of isolated organic solvent-extracted oil bodies from mature seeds. P34 is processed during seedling growth to a 32,000-Da polypeptide (P32) by the removal of an amino-terminal decapeptide (Herman, E.M., Melroy, D.L., and Buckhout, T.J. (1990) Plant Physiol, in press). A soybean lambda ZAP II cDNA library constructed from RNA isolated from midmaturation seeds was screened with monoclonal antibodies directed against two different epitopes of P34. The isolated cDNA clone encoding P34 contains 1,350 base pairs terminating in a poly(A)+ tail and an open reading frame 1,137 base pairs in length. The open reading frame includes a deduced amino acid sequence which matches 23 of 25 amino-terminal amino acids determined by automated Edman degradation of P34 and P32. The cDNA predicts a mature protein of 257 amino acids and of 28,641 Da. The open reading frame extends 5' from the known amino terminus of P34 encoding a possible precursor and signal sequence segments with a combined additional 122 amino acids. Prepro-P34 is deduced to be a polypeptide of 42,714 Da, indicating that the cDNA clone apparently encodes a polypeptide of 379 amino acids. A comparison of the nucleotide and deduced amino acid sequences in the GenBank Data Bank with the sequence of P34 has shown considerable sequence similarity to the thiol proteases of the papain family. Southern blot analysis of genomic DNA indicated that the P34 gene has a low copy number.

Mutations at the cysteine codons of the recA gene of Escherichia coli.

Each of the three cysteine residues in the Escherichia coli RecA protein was replaced with a number of other amino acids. To do this, each cysteine codon was first converted to a chain-terminating amber codon by oligonucleotide-directed mutagenesis. These amber mutants were then either assayed for function in different suppressor strains or reverted by a second round of mutagenesis with oligonucleotides that had random sequences at the amber codon. Thirty-three different amino acid substitutions were obtained. Mutants were tested for three functions of RecA: survival following UV irradiation, homologous recombination, and induction of the SOS response. It was found that although none of the cysteines is essential for activity, mutations at each of these positions can affect one or more of the activities of RecA, depending on the particular amino acid substitution. In addition, the cysteine at position 116 appears to be involved in the RecA-promoted cleavage of the LexA protein.

Temporal control of colicin E1 induction.

The expression of the gene encoding colicin E1, cea, was studied in Escherichia coli by using cea-lacZ gene fusions. Expression of the fusions showed the same characteristics as those of the wild-type cea gene: induction by treatments that damage DNA and regulation by the SOS response, sensitivity to catabolite repression, and a low basal level of expression, despite the presence of the fusion in a multicopy plasmid. Induction of expression by DNA-damaging treatments was found to differ from other genes involved in the SOS response (exemplified by recA), in that higher levels of DNA damage were required and expression occurred only after a pronounced delay. The delay in expression following an inducing treatment was more pronounced under conditions of catabolite repression, indicating that the cyclic AMP-cyclic AMP receptor protein complex may play a role in induction. These observations also suggest a biological rationale for the control of cea expression by the SOS response and the cyclic AMP-cyclic AMP receptor protein catabolite repression system.

Direct selection of mutations reducing transcription or translation of the recA gene of Escherichia coli with a recA-lacZ protein fusion.

When a recA-lacZ protein fusion was cloned into phage lambda, the resulting transducing phage grew normally on wild-type Escherichia coli, but its growth was severely inhibited in lexA(Def) mutant strains that express recA constitutively at high levels. Mutants of the transducing phage that grew on the lexA(Def) strains were isolated and were found to affect production of the RecA-beta-galactosidase hybrid protein. Most mutants, including a number of nonsense mutants, were phenotypically LacZ-. LacZ+ mutants were also isolated; most of these expressed lower basal and induced levels of beta-galactosidase activity. DNA sequence analysis revealed that some of the LacZ+ mutations were in the recA promoter. One of these was found to prevent induction. Unexpectedly, three of the mutations that reduced expression were located in the recA structural gene, at codons 10, 11, and 12. Further analysis of the codon 10 mutant showed that it most likely affected translation since it had little effect on transcription as measured by beta-galactosidase synthesis from a recA-lacZ operon fusion. This expression defect was not limited to the protein fusion, since the codon 10 mutation also reduced synthesis of RecA protein when present in a complete recA gene. Analysis of the recA DNA sequence in the fusion revealed that each of the mutations at codons 10, 11, and 12 increases the homology between this region of the mRNA and a sequence found at codons 1 to 4. Thus, the secondary structure of the mutant recA mRNAs may be affecting translation.

Measurement of in vivo expression of the recA gene of Escherichia coli by using lacZ gene fusions.

A recA-lacZ protein fusion was constructed in vivo by using bacteriophage Mu dII301(Ap lac). The fusion contained the promoter and first 47 codons of the recA mutant, as determined by DNA sequence analysis. The fusion was cloned and used to construct a recA-lacZ operon fusion at the same site within the recA gene. These fusions were introduced into the Escherichia coli chromosome at the lambda attachment site either as complete or cryptic lambda prophages. Synthesis of beta-galactosidase from these fusions was inducible by UV radiation. As the UV dose was increased, induction became slower and persisted for a longer period of time. At low doses of UV radiation, more beta-galactosidase was produced in a uvrA mutant than in a wild-type strain; however, at high doses, no induced synthesis of beta-galactosidase occurred in a uvrA mutant. recA+ strains carrying either the protein or operon fusion on a multicopy plasmid showed reduced survival after UV irradiation. This UV sensitivity was not exhibited by strains containing a single copy of either fusion, however; hence, the fusions provide a reliable measure of recA expression.

Lambda placMu: a transposable derivative of bacteriophage lambda for creating lacZ protein fusions in a single step.

We isolated a plaque-forming derivative of phage lambda, lambda placMu1 , that contains sequences from bacteriophage Mu enabling it to integrate into the Escherichia coli chromosome by means of the Mu transposition system. The Mu DNA carried by this phage includes both attachment sites as well as the cI, ner (cII), and A genes. Lambda placMu1 also contains the lacZ gene, deleted for its transcription and translation initiation signals, and the lacY gene of E. coli, positioned next to the terminal 117 base pairs from the S end of Mu. Because this terminal Mu sequence is an open reading frame fused in frame to lacZ, the phage can create lacZ protein fusions in a single step when it integrates into a target gene in the proper orientation and reading frame. To demonstrate the use of this phage, we isolated lacZ fusions to the malB locus. These showed the phenotypes and regulation expected for malB fusions and could be used to isolate specialized transducing phages carrying the entire gene fusion as well as an adjacent gene (malE). They were found to be genetically stable and rarely (less than 10(-7] gave rise to secondary Lac+ insertions. We also isolated insertions into high-copy-number plasmids. The physical structure of these phage-plasmid hybrids was that expected from a Mu-dependent insertion event, with the lambda placMu prophage flanked by the Mu attachment sites. Lac+ insertions into a cloned recA gene were found at numerous positions and produced hybrid proteins whose sizes were correlated with the position of the fusions in recA.