Characterization of the expression-linked gene copies of variant surface glycoprotein 118 in two independently isolated clones of Trypanosoma brucei.

We have analysed the gene for variant surface glycoprotein 118 in eight independent clones of Trypanosoma brucei, two of which express the 118 gene. Expression of this gene is strictly coupled to the presence of an extra copy of the gene. In both clones the expression-linked copy is transposed to the same (or a very similar) expression site elsewhere in the genome, but the length of the sequences flanking the transposed segment in the expression site differs markedly. By means of S1 nuclease protection experiments we demonstrate that the 3'-ends of the messenger RNAs for variant surface glycoproteins 118a and 118b are different, in agreement with the hypothesis that the generation of an expression-linked copy involves a recombination between the 3' segment of the basic gene copy and a homologous region present in the expression site.     

Comparison of the expression-linked extra copy (ELC) and basic copy (BC) genes of a trypanosome surface antigen

A recombinant clone of an expression-linked extra copy (ELC) gene of a trypanosome-variable surface glycoprotein was sequenced. In addition the sequences of the corresponding cDNA and portions of the two basic copy genes were determined. Comparison of these sequences reveals that the 5' boundary of the ELC-transposed segment (2.2 kb) occurs within a repetitive sequence about 700 bp upstream from the start codon of the coding sequence. This sequence does not contain internal symmetries and is not homologous with the repetitive sequence at the 3' boundary. The first 35 nucleotides of the cDNA are different than the corresponding ELC sequence and presumably were transcribed from another genomic location. A restriction fragment containing predominantly sequences outside of the 5' boundary hybridizes to a Pst I fragment whose length is variable in different trypanosome clones. This hybridization pattern is similar to that observed using probes for surface glycoprotein genes that are expressed via the nonduplication-associated (NDA) mechanism rather than the ELC mechanism. This indicates that there is a sequence correlation between these two DNA rearrangement mechanism.     

The inactivation and reactivation of an expression-linked gene copy for a variant surface glycoprotein in Trypanosoma brucei.

We have previously shown that the gene for variant surface glycoprotein 118 of Trypanosoma brucei (strain 427) is activated by a duplicative transposition to a telomeric expression site. In chronically-infected animals, this expression-linked copy is lost when the 118 gene is replaced at the expression site by another variant surface glycoprotein gene. We show here that expression of the 118 gene can also be switched off without loss of the extra expression-linked copy. In two variants, called 1.8b and 1.8c, we find expression of the variant surface glycoprotein 1.8 gene, notwithstanding the continued presence of the 118 expression-linked copy. The 1.8 gene activated has a telomeric location, like the 118 expression-linked copy. In variant 1.8b, activation is accompanied by duplication of the 1.8 gene, resulting in an extra telomeric gene copy; in variant 1.8c it is not. Variants 1.8b and 1.8c both switch back preferentially to expression of the 118 gene. The 5'-flanking regions of the active, inactive and reactivated versions of the 118 expression-linked copy are indistinguishable by restriction mapping up to 28 kb. We conclude that there are at least two separate telomeric expression sites in our T. brucei strain. How these are switched on and off is unclear. The ability to retain expression-linked copies in inactive form may allow the trypanosome to re-programme the order in which variant surface glycoprotein genes are expressed.     

At least two transposed sequences are associated in the expression site of a surface antigen gene in different Trypanosome clones

The expression of several trypanosome surface antigen genes proceeds by duplication of a basic copy (BC) of the gene and transposition of the expression-linked copy (ELC) into an expression site. This site, which seems to be the same for different genes of the same repertoire, is located near a chromosome end. In the AnTat 1.1 antigen gene expression site, the ELC is found associated with another sequence that we have called the “companion.” We found that this companion is the transposed copy of another sequence also located in an unstable DNA terminus, and that it is conserved in the expression site of AnTat 1.10 and AnTat 1.1B, two clones successively derived from AnTat 1.1. The companion sequence is not part of the surface antigen gene, but we may infer from extensive homologies with another ELC sequence (IoTat 1.3, J. E. Donelson, personal communication) that it represents a 5′ residual fragment of a former ELC. In three other AnTat 1.1-like clones, the companion sequence was not found associated with the ELC. It is concluded that the expression-linked duplicative transposition of variable antigen genes is a flexible mechanism, which can apply to variably sized stretches of the same BC.     

A DNA fragment hybridizing to a nif probe in Rhodobacter capsulatus is homologous to a 16S rRNA gene

We have sequenced the Rhodobacter capsulatus nifH and nifD genes. The nifH gene, which codes for the dinitrogenase reductase protein, is 894 bp long and codes for a polypeptide of predicted Mr 32,412. The nifD gene, which codes for the alpha subunit of dinitrogenase, is 1,500 bp long and codes for a protein of predicted Mr 56,113. A 776-bp BglII-XhoI fragment containing only nif sequences was used as a hybridization probe against R. capsulatus genomic DNA. Two HindIII fragments, 11.8 kb and 4.7 kb in length, hybridize to this probe. Both fragments have been cloned from a cosmid library. The 11.8-kb fragment contains the nifH, D and K genes, as previously demonstrated (Scolnik and Haselkorn, 1984). In this paper we present evidence that suggests that the 4.7-kb HindIII fragment contains a gene coding for 16S rRNA, and that although homology between nif and this fragment can be observed in filter hybridization experiments, a second copy of the nif structural genes seems not to be present in this region.     

Novel use of a chimpanzee pseudogene for chromosomal mapping of human cytochrome c oxidase subunit IV

We have isolated a chimpanzee processed pseudogene for subunit IV of cytochrome c oxidase (COX; EC 1.9.3.1) by screening a chimpanzee genomic library in lambda Charon 32 with a bovine liver cDNA encoding COX subunit IV (COX IV), and localized it to a 1.9-kb HindIII fragment. Southern-blot analysis of genomic DNA from five primates showed that DNAs from human, gorilla, and chimpanzee each contained the 1.9-kb pseudogene fragment, whereas orangutan and pigtail macaque monkey DNA did not. This result clearly indicates that the pseudogene arose before the divergence of the chimpanzee and gorilla from the primate lineage. By screening Chinese hamster x human hybrid panels with the human COX4 cDNA, we have mapped COX4 genes to two human chromosomes, 14 and 16. The 1.9-kb HindIII fragment containing the pseudogene, COX4P1, can be assigned to chromosome 14, and by means of rearranged chromosomes in somatic cell hybrids, to 14q21-qter. Similarly, the functional gene, COX4, has been mapped to 16q22-qter.     

The rabbit alpha-like globin gene cluster is polymorphic both in the sizes of BamHI fragments and in the numbers of duplicated sets of genes

The alpha-like globin gene cluster in rabbits contains embryonic zeta- globin genes, an adult alpha-globin gene, and theta-globin genes of undetermined function. The basic arrangement of genes, deduced from analysis of cloned DNA fragments, is 5'-zeta 0-zeta 1-alpha 1-theta 1- zeta 2-zeta 3-theta 2-3'. However, the pattern of restriction fragments containing zeta- and theta-globin genes varies among individual rabbits. Analysis of BamHI fragments of genomic DNA from 24 New Zealand white rabbits revealed eight different patterns of fragments containing zeta-globin genes. The large BamHI fragments containing genes zeta 0 and zeta 1 are polymorphic in length, whereas a 1.9-kb fragment containing the zeta 2 gene and the 3.5-kb fragment containing the zeta 3 gene do not vary in size. In contrast to this constancy in the size of the restriction fragments, the copy number of the zeta 2 and zeta 3 genes does vary among different rabbits. No length polymorphism was detected in the BamHI fragments containing the theta-globin genes, but again the copy number varies for restriction fragments containing the theta 2 gene. The alpha 1- and theta 1-globin genes are located in a nonpolymorphic 7.2-kb BamHI fragment. The combined data from hybridization with both zeta and theta probes shows that the BamHI cleavage pattern does not vary within the region 5'-alpha 1-theta 1- zeta 2-zeta 3-theta 2-3', but the pattern genomic blot-hybridization patterns for the progeny of parental rabbits with different zeta-globin gene patterns shows that the polymorphic patterns are inherited in a Mendelian fashion. Two different haplotypes have been mapped based on the genomic blot-hybridization data. The variation in the alpha-like globin gene cluster in the rabbit population results both from differences in the copy number of the duplication block containing the zeta-zeta-theta gene set and from the presence or absence of polymorphic BamHI sites.      

The use of incomplete genes for the construction of a Trypanosoma equiperdum variant surface glycoprotein gene.

The expression of Trypanosoma equiperdum variant surface protein (VSG) 78 is accomplished by the duplicative transposition of silent basic copy (BC) genes into a telomer-linked expression site to form an expression-linked copy (ELC). In two independent isolates expressing VSG 78, the ELC is a composite gene. The analysis of VSG 78 cDNA clones from these two Bo Tat 78 isolates and the respective BC genes revealed that both ELCs were constructed from the same three BC genes, a 3' BC which donated the last 255 bp of each ELC and two closely related 5' BCs. Although sequences of both 5' BC genes were found in each ELC, the junction with the 3' BC was provided by the same 5' BC in both cases. This 5' BC is an incomplete gene with insufficient open reading frame to code for a complete VSG and thus can only be used when joined to a competent 3' end. Furthermore, both 5' BC genes lack a conserved 14 nucleotide sequence found on all VSG mRNAs. These results support a model in which composite gene formation plays a role in the determination of the order of VSG expression. They also illustrate similarities between immunoglobulin gene and VSG gene construction.     

Human carboxypeptidase A identifies a BglII RFLP and maps to 7q31-qter.

A genomic clone for human carboxypeptidase has been isolated with a probe for rat CPA1 cDNA. A 1.7-kb HindIII/EcoRI fragment from the 3' flanking region of human carboxypeptidase detects a DNA polymorphism with BglIII. Multipoint linkage analysis with an established map of chromosome 7 markers shows that the most likely location of carboxypeptidase is at 7q31-qter, between D7S87 and D7S93. All other placements can be excluded with odds greater than 100:1. These and other markers confirm that carboxypeptidase lies distal to the locus for cystic fibrosis, at a distance of approximately 12 centimorgans. The regions containing identity to the rat gene were sequenced and shown to be 82% identical to exons 9 and 10 of the rat gene. The presence of a codon for isoleucine at the residues corresponding to codon 255 of rat CPA1 cDNA strongly suggests that the A form of human carboxypeptidase has been isolated.     

Coincident multiple activations of the same surface antigen gene in Trypanosoma brucei

Trypanosomes with a coat of variant surface glycoprotein (VSG) 118, consistently appear around day 20 when a rabbit is infected with Trypanosoma brucei strain 427. There is a single chromosome-internal gene for VSG 118 and this is activated by duplicative transposition to a telomeric expression site. We show here that the expression-linked extra copy of VSG gene 118 in a day 18 population of a chronic infection is heterogeneous, and we infer that the population is not monoclonal but is the result of multiple independent activations of the 118 gene. We show that the heterogeneity of expression-linked extra copies is also present in other trypanosome populations expressing chromosome-internal VSG genes. We present a model for the timing of VSG gene activation during chronic infection that emphasizes two features: the relative activation and inactivation frequencies of different expression sites, and the degree of homology of the sequences flanking VSG genes with expression sites.     

DNA rearrangements and antigenic variation in Trypanosoma equiperdum: expression-independent DNA rearrangements in the basic copy of a variant surface glycoprotein gene.

Antigenic variation in Trypanosoma equiperdum is associated with the sequential expression of variant surface glycoprotein (VSG) genes in a process which involves gene duplication and transposition events. In this paper we present evidence that the genomic environment of the VSG-1 basic copy gene, the template for duplicated, expression-linked VSG-1 genes, differs in every trypanosome clone examined. This variation is thus independent of the expression of the VSG-1 gene, and it also appears to be restricted to the 3' genomic environment. It is also demonstrated that the DNA located 3' to the VSG-1 basic copy gene is moderately sensitive to digestion when the nuclei of either expressor or non-expressor trypanosomes are treated with DNase I.     

Organization of the biosynthesis genes for the peptide antibiotic gramicidin S.

A recombinant bacteriophage containing the intact Bacillus brevis gene for gramicidin S synthetase 1, grsA, and the 5' end of the gramicidin S synthetase 2 gene, grsB, was identified by screening an EMBL3 library with anti-GrsA antibodies. This clone, EMBL315, has a 14-kilobase (kb) insert that hybridizes to the previously isolated 3.9-kb fragment of the grsB gene, which encodes the 155-kilodalton ornithine-activating domain of gramicidin S synthetase 2. Deletion and subcloning experiments with the 14-kb insert located the grsA structural gene and its putative promoter on a 4.5-kb PvuII fragment which encoded the full-length 120-kilodalton protein in Escherichia coli. In addition, hybridization analysis revealed that the 5' end of the grsB gene is located approximately 3 kb from the grsA structural gene. Furthermore, these studies indicated that grsA and grsB are transcribed in opposite orientations.     

Telomere conversion in trypanosomes.

Activation of the gene coding for variant surface glycoprotein (VSG) 118 in Trypanosoma brucei proceeds via a duplicative transposition to a telomeric expression site. The resulting active expression-linked extra copy (ELC) is usually flanked by DNA that lacks sites for most restriction enzymes and that is thought to interfere with the cloning of the ELC as recombinant DNA in Escherichia coli. We have circumvented this problem by cloning an aberrant 118 ELC gene, flanked at the 3'-side by at least 1 kb DNA, that contains restriction enzyme sites. Our analysis shows that this DNA and the 3'-end of the 118 ELC gene are derived from another VSG gene (1.1006) that is permanently located at a telomeric position. We propose that the 3'-end of the 1.1006 gene and (all of) its 3' flanking sequence moved to the expression site by a telomere conversion. Such a telomere conversion can also account for the appearance of an extra copy of the 1.1006 gene detected in a sub-population of our trypanosome strain.     

The human calbindin 27-kDa gene: structural organization of the 5' and 3' regions, chromosomal assignment, and restriction fragment length polymorphism

The 5' and 3' regions of the human gene coding for calbindin 27 kDa were cloned and sequenced. Structural features of the 5' region included the presence of an Alu repeat and two elements regularly associated with eukaryotic promoters: an alternating purine-pyrimidine element and a homopurine-homopyrimidine box. The 3' region contained a second Alu family member and a degenerate 1.4-kb L1 repeat. A comparison with the chicken promoter was made in order to define regions conserved in evolution and potentially important in gene expression regulation. The greater similarity is located around the TATA box, but strongly conserved elements were not found. The gene was assigned to chromosome 8 by using human-rodent hybrid cell lines. Two restriction fragment length polymorphisms (HindIII and SacI) were detected with a cDNA probe recognizing the 3' end of the gene.     

Two coding regions closely linked to the rat apolipoprotein E gene: nucleotide sequences of rat apolipoprotein C-I and ECL cDNA.

Restriction fragments isolated from a 17-kb rat genomic DNA clone containing the gene for apolipoprotein (apo) E were radiolabeled and used to screen a rat liver cDNA library. A cDNA clone hybridizing to a 6-kb genomic DNA fragment was isolated and the nucleotide sequence of the cDNA insert determined. The sequence was homologous to the sequence for human apo C-I and was used to derive the corresponding amino acid sequence. Unlike human apo C-I, mature rat apo C-I contains histidine, lacks valine, and has alanine at the C terminus and aspartate as the N terminus. Screening the rat liver cDNA library with a radiolabeled 1.9-kb restriction fragment from the genomic DNA clone containing the rat apo E gene identified another cDNA clone (ECL cDNA). Nucleotide sequencing yielded a derived 75-amino-acid sequence for the ECL protein with a hydrophobicity profile similar to that of rat apo C-I. Northern analysis demonstrated a 0.50-kb band for ECL mRNA. The tissue-specific expression of the gene is similar to that of rat apo C-I. This study indicates that the rat apo C-I and ECL genes are closely linked, about 4.5 and 12 kb downstream of the apo E gene, respectively.     

cDNA for the carboxyl-terminal region of a rat intestinal mucin-like peptide.

When subjected to thiol reduction, purified intestinal mucins have been shown to undergo a decrease in molecular mass and to liberate a 118-kDa glycopeptide (Roberton, A. M., Mantle, M., Fahim, R. E. F., Specian, R., Bennick, A., Kawagishi, S., Sherman, P., and Forstner, J. F. (1989) Biochem. J. 261, 637-647). The latter has been called a putative "link" component because it is assumed to be important for disulfide bond-mediated mucin polymerization. Controversy exists as to whether the putative link is an integral mucin component or a separate mucin-associated glycopeptide. In the present study both NH2-terminal and internal amino acid sequences of the 118-kDa glycopeptide of rat intestinal mucin were used to generate opposing oligonucleotide primers for polymerase chain reaction. A specific 1.2-kilobase (kb) product was obtained, from which a 0.5-kb HindIII fragment was used as a probe to screen a lambda ZAP II cDNA library of rat intestine. A 2.6-kb cDNA (designated MLP 2677) was sequenced and revealed an open reading frame of 2.5 kb encoding 837 amino acids. The deduced amino acid sequence showed that the putative link peptide is equivalent to the carboxyl-terminal 689 amino acids of a larger peptide. Northern blots revealed a mRNA size of approximately 9 kb. Computer searches revealed no sequence homology with other proteins, but similarities were seen in the alignment of cysteine residues in the link and in several domains of human von Willebrand factor, as well as cysteine-rich areas of bovine and porcine submaxillary mucins and a frog skin mucin designated FIM-B.1. In keeping with earlier demonstrations of the presence of mannose in the 118-kDa glycopeptide, there were several (13) consensus sequences for attachment of N-linked oligosaccharides within the link domain. Further sequencing of MLP 2677 in a direction 5' to the codon specifying the NH2-terminal proline of the link has revealed a coding region for 148 amino acids, including a unique 75-amino acid domain rich in cysteine and proline, and a region containing 4.5-variable tandem repeats (each 11-12 amino acids) rich in serine, threonine, and proline. The presence of mucin-like tandem repeats suggests that the entire cysteine-rich link peptide represents the carboxyl-terminal region (75.5 kDa) of a mucin-like peptide (MLP). The latter is estimated to have a molecular mass of approximately 300 kDa.