Search

We have examined the correlation between the extents of methylation and expression of pepsinogen A genes in normal human tissues. Expression of pepsinogen A mRNA was detected only in the fundic mucosa of the stomach and both CCGG and GCGC sites in the genes region were less methylated in the fundic mucosa than in other non-expressing tissues. Thus, there was an inverse correlation between the extents of methylation and expression of pepsinogen A genes and the role of DNA methylation in the regulation of pepsinogen A genes expression during normal differentiation was suggested.

Only two types of pepsinogens, which are products of the Pg-1 locus, are present in rat urine. In gastric mucosa, however, additional pepsinogen isozymes are expressed. We have found a polymorphism for rat gastric mucosa pepsinogen using agarose gel electrophoresis. Some inbred rat strains expressed a pepsinogen band, while others did not. The trait was found to be controlled by a single autosomal locus. We tentatively designated the locus as Pg-2 with two alleles, Pg-2a for the one controlling presence of the band and Pg-2o for the one controlling absence. Linkage analysis using BN and TM strains revealed that Pg-2 was closely linked to Pg-1 (3.7 +/- 1.8 cM), and that it did not belong to LG I (Hbb and p), LG II (Acon-1 and Mup-1), LG IV (Hao-1 and Svp-1), LG V (Es-1 and Es-3), LG VI (Gc and h), LG IX (RT1), LG X (Fh and Pep-3), nor a LG containing Ahd-2 (as yet undetermined).

Previous studies in rats have shown that stomach carcinogens cause a decrease of total pepsinogen content and a decrease of the fastest pepsinogen isozymogen in gastric mucosa before the appearance of carcinoma. In man the presence of Pg 5, and in particular a strong Pg 5, appears to be associated with gastric cancer. Low serum PG I levels were found in patients with gastric cancer, atrophic gastritis, intestinal metaplasia and gastric polyp. In this chapter we review the literature and report our results of PG I isozymogen and serum PG I determinations in 500 patients with various gastric disorders. It is concluded that premalignant changes of the stomach are associated with the presence of a strong Pg 5 isozymogen and with a low serum PG I level.

A pepsinogen (Pg 1) polymorphism was observed in gastric mucosal samples of Wistar rats by polyacrylamide slab gel electrophoresis. Three phenotypes were observed: S (a slow-moving band), F (a fast-moving band), and FS (both bands). Pg 1 variation was controlled by a pair of codominant alleles at a single autosomal locus. Interstrain variation of Pg 1 was studied in five strains of laboratory rats. Genetic variation of Pg 1 was detected only in the Wistar strain.

Poly(A)-containing RNA was isolated from total RNA of swine gastric mucosa by oligo(dT)-cellulose chromatography, and was translated in a wheat germ cell-free system. Analysis of the translation products by sodium dodecyl sulfate-polyacrylamide gel electrophoresis revealed that pepsinogen was a major translation product and was synthesized as two different molecular forms with apparent molecular weights of 45,000 and 43,000. The pepsinogen translated in the in vitro translation system had no autocatalytic activity. The pepsinogen mRNA was further purified by sucrose density gradient centrifugation, where the mRNA activity for pepsinogen was located around 15 S. At this stage, the translation product of the pooled fractions appeared to be almost exclusively pepsinogen. The peptide maps of the pepsinogen which was translated in vitro and digested by alpha-chymotrypsin and by Staphylococcus aureus V8 protease were nearly identical with the corresponding peptide maps of standard pepsinogen. The double-stranded complementary DNA which had been synthesized from the partially purified mRNA by avian myeloblastosis virus reverse transcriptase was cloned in Escherichia coli chi 1776, using plasmid pBR322 as a cloning vector. A colony carrying pepsinogen cDNA sequence was identified by in situ colony hybridization using the cDNA synthesized from the partially purified mRNA as a probe and further by a positive hybridization-translation assay. One of the recombinant clones (pSPcA1) had an insert of about 850 base pairs, and the nucleotide sequence analysis revealed that pSPcA1 codes for swine pepsinogen.

The evidence for a genetic component in peptic ulcer disease has been based on twin, family, and blood group studies. A polygenic model for the inheritance of peptic ulcers has been displaced by a genetic heterogeneity model based on several lines of evidence, some of the most powerful being recent work using subclinical markers. One marker in particular, an elevated level of serum pepsinogen I (PG I), a pepsin precursor produced by the gastric mucosa, secreted into the stomach lumen and also appearing in the bloodstream, has been found to be associated with a subgroup of duodenal ulcer patients. Segregation analysis of elevated serum PG I in duodenal ulcer sibships demonstrates familial aggregation consistent with autosomal dominant inheritance. Elevated PG I is also accompanied by gastric hyperacidity and presumably indicates those individuals with an increased mass of chief and parietal cells, and thus an increased capacity for peptic activity, an important element in the pathogenesis of ulcer disease. An evolutionary hypothesis based on selection for peptic activity and acidity is offered to explain several of the epidemiologic and genetic elements of this group of chronic diseases.

Pepsinogens (PGA) are the inactive precursors of pepsin, the major acid protease found in the stomach. Highly polymorphic variation of these proteins has been demonstrated in several populations, and comparison of the DNA restriction fragment patterns obtained from informative pepsinogen phenotypes suggest that the polymorphism results from chromosomal haplotypes containing variable numbers of pepsinogen genes. In order to isolate the three most common PGA haplotypes (A, B, and C) and to unambiguously demonstrate their relationship to the observed protein heterogeneity, we constructed mouse X human somatic cell hybrids from individuals heterozygous for PGA and INS (insulin). Here, we describe analysis of hybrid cell lines that segregated human chromosomes containing the PGA genes and thereby provided for the parasexual discrimination of the different haplotypes on chromosome 11 determining the corresponding heterozygous phenotypes. These studies demonstrate that the A, B, and C haplotypes contain three, two, and one PGA genes, respectively. This unusual polymorphism of genomic DNA encoding very similar proteins probably reflects recent evolution by gene duplication.

A clone, pSPcA2, which carries the full-length swine pepsinogen cDNA was isolated. The coding sequence comprised the signal peptide [15 amino acids (aa)], the activation peptide segment (44 aa) and mature pepsin (327 aa). The deduced amino acid sequence agrees with the published sequence with two exceptions. Asparagine instead of aspartate is present at aa positions 19 and 308. Two types of plasmids, pAS and pUCtacSPc series, were constructed for expressing swine pepsinogen cDNA. These plasmids directed the synthesis of polypeptides which were detected by employing an antibody to swine pepsinogen. However, all the polypeptides formed aggregates and showed no acid protease activity. Only the protein directed by pAS5 regained the acid protease activity after renaturation procedures. The activity was completely inhibited by pepstatin. Furthermore, the renatured pAS5 protein was spontaneously converted to pepsin under acidic conditions. The presence of Arg-8 in the activation peptide segment appears important for the stabilization of the pepsinogen molecule.

Japanese monkey progastricsin was shown to be activated to gastricsin exclusively by a two-step process through an intermediate form. The occurrence of this process was substantiated by the isolation of the intermediate form and released peptides. By NH2-terminal sequence analyses of these protein and peptide species, the amino acid sequence of the 43-residue activation segment (propart) was determined to be as follows: (Formula: see text) The NH2-terminal 26-residue peptide was released first, resulting in generation of the intermediate form. The subsequent release of peptides, residues Nos. 27-40 and 27-43, generated two gastricsins as the final products. This two-step process of activation of Japanese monkey progastricsin is in striking contrast to the one-step activation process occurring exclusively for pepsinogen A of the same monkey species. The course of molecular evolution of pepsinogens including progastricsins was deduced from the amino acid sequences of their activation segments by constructing phylogenic trees. The trees divided pepsinogens into 3 clusters, i.e., pepsinogens A, progastricsins and prochymosin, showing that these three groups diverged from one another very early on in the course of the evolution of pepsinogens.

To unravel the genetic basis for the pepsinogen A (PGA) protein polymorphism, we have isolated and characterized a number of PGA genes, distinguishable by polymorphic EcoRI fragments of 12.0, 15.0, and 16.6 kb. Using a HindIII or AvaII polymorphism, we can discriminate between different 15.0 (15.0 and 15.0*) and 12.0 (12.0s and 12.0l) genes, respectively. The coding sequences of a 15.0 and a 16.6 gene were determined, together with considerable stretches of the 5'- and 3'-flanking regions and introns. The genes were demonstrated to encode Pg5 and Pg4, respectively. Because substitutions in codons 43 and 207 appeared to be critical in the determination of the encoded proteins, we sequenced only these regions in the two 12.0 genes and the 15.0* gene. On the basis of these partial sequences, we assume that these genes encode Pg3. In the evolutionary model of the PGA gene cluster presented here, the 12.0 genes arose by an unequal, but homologous crossover. The results of sequence analysis of the second intron of the 12.0s, 12.0l, 15.0, and 16.6 genes suggest that the two 12.0 genes have arisen from two different crossover events.

The role of pepsin in the pathogenesis of peptic ulcer has been the subject of intense study and debate for many years. Two difficulties inherent in distinguishing between the role of acid alone vs acid and pepsin are that a) acid-containing gastric juice always contains pepsin, and, b) that hydrogen ion concentration (pH) is a major determinant of the activity of pepsin. However, studies in animal models of peptic ulcer indicate clearly that pepsin, in combination with acid, produces much more severe and more extensive mucosal damage than acid alone. Recent interest in pepsin and its precursor, pepsinogen, has stemmed from the finding that each is remarkably heterogeneous, and that the heterogeneity has a genetic basis. Results of studies using radioimmunoassays specific for the 2 major forms of pepsinogen, pepsinogen I and pepsinogen II, have shown that serum levels of pepsinogen I and pepsinogen II, and the ratio of pepsinogen I to pepsinogen II, can be used as noninvasive probes of gastric mucosal structure and function, indicators of the genetics and heterogeneity of duodenal ulcer, and as markers of increased risk for duodenal ulcer and gastric ulcer.

Two immunologic subclasses of human pepsinogen I alpha-PG I and beta-PG I, have been identified based on their reactivity toward a murine monoclonal antibody that recognizes an epitope on the alpha-PG I isozymogens. The antibody was used to purify the major alpha- and beta-isozymogens from gastric mucosa and to determine their contributions to the previously described genetic polymorphism of PG I. The alpha-epitope was localized to the pepsin region of the molecules. The two major alpha-PG I isozymogens (Pg 3 alpha and Pg 5 alpha) and the major beta-PG I isozymogen (Pg 4 beta) were demonstrated to contain net charge differences located in the respective pepsin and activation peptide regions. We propose that the alpha- and beta-subclasses contain net charge amino acid substitutions encoded by the corresponding pepsinogen genes: PGA3, PGA4, and PGA5.

Electrophoretic pepsinogen A patterns were determined in gastric fundic mucosa biopsies from 601 patients with various gastric disorders and 25 healthy volunteers. Pepsinogen A patterns with an intense fraction 5 appeared to be associated with gastric cancer and premalignant changes of the stomach (p less than 10(-9)). In 60 individuals pepsinogen A patterns were determined in normal mucosa from different parts of the stomach. No differences were found between these patterns. In 29 out of 59 gastric cancer patients pepsinogen A could be demonstrated in the macroscopically malignant tissue. In two cases a different pattern compared with uninvolved fundic mucosa was observed. During a follow up study, major changes in the pepsinogen A pattern were observed in 7 out of 56 patients. In 8.6% of the examined patients urinary pepsinogen A patterns differed considerably as compared with the pattern observed in the gastric fundus. The results suggest that the highly significant association between intense Pg5 (the product of the D gene) and gastric cancer or its precursors may be caused by genetic as well as non-genetic factors.

A new genetic model is proposed to explain the inheritance of the urinary pepsinogen (PG1) polymorphism. Each main fraction, 3, 4 and 5, in the multibanded electrophoretic pattern, is determined by its own specific gene, B, C and D respectively. The intensity ratio of the fractions is principally determined by the number of gene copies. Accordingly, the PG1 phenotypes are determined by gene combinations, haplotypes, some of which may be identical to alleles in previous one locus models. Some critical families, not interpretable using previous genetic models, are presented to support the hypothesis. Preliminary population data from the Netherlands are described. The molecular background of this polymorphism and its relevance for gastric (pre)malignancy is discussed.

cDNA of rat pepsinogen C (PGC) hybridizes to, among others, a 3.2-kb band in Southern blot analysis of BamHI-cleaved human genomic DNA. This property was employed to localize the human PGC gene. Use of flow-sorted human chromosomes and 12 human x mouse somatic cell hybrid lines demonstrated that the gene is located on chromosome 6.

Serum PG I levels were measured by an enzyme linked immuno sorbent assay (ELISA) in 567 blood donors and 171 patients from a routine gastroscopy program to study the relationship between PG I phenotypes and serum levels. In normal subjects no association was found between PG I phenotypes and serum levels. A low serum PG I level in patients was associated with phenotypes characterized by intense PG I fraction 5. This probably reflects the high frequency of these phenotypes in patients with atrophic gastritis or gastric cancer. The mean (+/- SD) serum PG I level was 45.8 +/- 17.7 micrograms/1 in control subjects. Serum PG I levels were lower in females (41.4 +/- 17.5) than in males (47.3 +/- 17.7) and increased with advancing age up to 65 years.

A nearly full-length rat pepsinogen cDNA was isolated from a rat gastric mucosa cDNA library and its nucleotide sequence was determined. The cDNA comprises 1370 base pairs (bp), including the 5'-non-coding region (60 bp), the coding nucleotide sequence (1176 bp) and the 3'-non-coding region (131 bp). The predicted amino acid sequence of rat prepepsinogen (392 residues) contains a 16-residue signal sequence followed by the pepsionogen moiety of 376 residues. Rat pepsinogen has an amino acid composition characteristic of C-type pepsinogens and is much more homologous in amino acid sequence with C-type pepsinogens than A-type pepsinogens. These results indicate that the major form of rat pepsinogen can be classified as a C type pepsinogen.

Embryonic chicken pepsinogen gene was isolated from a chicken genomic library. This gene occupied approximately 3.5 kb of the genomic DNA and was separated into nine exons by eight introns. The positions of exon-intron junctions coincided with those in the human pepsinogen A gene and the bovine prochymosin gene. Southern blot analysis of chicken genomic DNA revealed that the structure of the isolated gene reflects the original structure in the chicken chromosome. At the same time, the presence of another copy of embryonic pepsinogen gene was suggested. 5'-flanking region of the isolated ECPg gene was analyzed.

In order to carry out studies on structure and function relationships of porcine pepsinogen using site-directed mutagenesis approaches, the cDNA of this zymogen was cloned, sequenced, expressed in Escherichia coli, and the protein refolded, and purified to homogeneity. Porcine pepsinogen cDNA, obtained from a lambda gt10 cDNA library of porcine stomach contains 1364 base pairs. It contains leader, pro, and pepsin regions of 14, 44, and 326 residues, respectively. In addition, it also contains 5'- and 3'-untranslated regions. Four differences are present between the sequence deduced from the cDNA and the pepsinogen sequence determined previously by protein chemistry methods. Residues P19 (in the pro region) and 263 are asparagines in the cDNA sequence instead of aspartic acids. Isoleucine 230 is not present in the cDNA sequence and residue 242 is a tyrosine in the cDNA instead of an aspartic acid. Porcine pepsinogen cDNA was placed under the control of a tac promoter in a plasmid and expressed in E. coli. The synthesis of pepsinogen was optimized to about 50 mg/liter of culture. The recombinant (r-) pepsinogen, which was insoluble, was recovered by centrifugation, washed, dissolved in 6 M urea in Tris-HCl, pH 8, and refolded by rapid dilution. r-pepsinogen was purified to homogeneity after chromatography on Sephacryl S-300 and fast protein liquid chromatography on a monoQ column. r-pepsinogen contains an additional methionine residue at the NH2 terminus as compared to native (n-) pepsinogen. However, r- and n-pepsinogens are indistinguishable in their intramolecular activation constants. After activation, r- and n-pepsins have the same NH2-terminal sequences as well as Km values. Based on these data, r-pepsinogen was judged suitable for mutagenesis studies. A mutant pepsinogen (D32A) with the active site aspartic acid changed to an alanine was produced and purified. D32A-pepsinogen did not convert to pepsin in acid solution but it bound to pepstatin with an apparent KD of about 5 x 10(-10) M. D32A-pepsinogen possesses no detectable proteolytic activity. These results indicate that (i) intramolecular pepsinogen activation is accomplished by the pepsin active site, and (ii) unlike subtilisin (Carter, P., and Wells, J. A. (1988) Nature 332, 564-568), the active site mutant of pepsin is not enzymically active.

A human cosmid library was screened with a pepsinogen A (PGA) cDNA probe, yielding 18 clones with (parts of) one, two or three PGA genes. By aligning these cosmids a restriction map of a PGA gene quadruplet was obtained in which the four genes are arranged in a highly ordered fashion in a head-to-tail orientation. Using the length in kilobases of the large polymorphic EcoRI fragment of the PGA genes, this quadruplet can be described as 15.0-12.0-12.0-16.6. An AvaII polymorphism allowed us to identify the two PGA haplotypes of the individual whose DNA had been cloned in the cosmid library to be a gene triplet and a gene quadruplet. By comparing the restriction maps of the central 12.0 genes in these multiplets to those of the flanking 15.0 and 16.6 genes, we postulate that these central genes arose from unequal but homologous crossing over between two 15.0-16.6 gene pairs. This hypothesis provides for the creation of a variety of haplotypes by additional cross overs and mutations. Southern blots of family and population material supports the existance of at least five common PGA haplotypes, including a single-gene haplotype, giving rise to a large number of different EcoRI patterns. The single PGA gene is probably the reciprocal crossing over product. Comparison between the DNA and protein polymorphisms suggests further micro-heterogeneity in the different PGA haplotypes.

Changes have been studied in human and rat pepsinogen phenotypes induced by N'-methyl-N'nitro-N-nitrosoguanidine (MMNG) in in vitro rat experiments and in vivo cultures of human and rat isolated gastric chief cells. In vivo the fastest migrating electrophoretic band decreased or disappeared as early as 3 weeks after the start of MNNG treatment. The changes, observed in 17 of 32 rats receiving MNNG, were permanent and consistently associated with pronounced histopathologic changes seen 10 months later (17 of 17). Comparable phenotypic changes were observed after 7 days only in MNNG-treated rat chief cell cultures. In human chief cell cultures a decrease of the Pg3 band, which is consistent with the "carcinogenic" phenotype, was observed in two of six preparations treated with MNNG. This early preceding change in phenotype preceding tumor formation may be useful as a diagnostic tool for the onset of gastric cancer.

Embryonic chicken pepsinogen is an aspartyl proteinase that is specifically secreted during the embryonic period in the chicken proventriculus (glandular stomach). To learn the phylogeny of this pepsinogen, we isolated a cDNA clone by screening a lambda gt11 library of embryonic proventricular cDNAs with an antiserum to the embryonic chicken pepsinogen. We obtained a 200-base pair cDNA clone which encoded 18 amino acids that had high sequence homology with the carboxyl termini of other pepsinogens. Northern blot analysis revealed that this cDNA clone hybridized to a mRNA of 1,600 bases in the embryonic proventriculus but not to the mRNA in the adult proventriculus. The almost complete nucleotide sequence of embryonic chicken pepsinogen-cDNA was determined by sequencing longer cDNAs obtained by screening the same library with the 200-base pair cDNA and primer extension with a synthetic primer. The cDNA consisted of 1,281 nucleotides and encoded 383 amino acids for prepepsinogen. The predicted amino acid sequence was compared with the sequences of other aspartyl proteinases: pepsinogen A of human, monkey, pig, and chicken, progastricsin of monkey and rat, and bovine prochymosin. The phylogenetic tree constructed for them indicates the possibility that embryonic chicken pepsinogen diverged from prochymosin, after prochymosin and pepsinogen A had diverged from each other.

Human pepsinogen A (PGA) displays highly polymorphic isozymogen patterns after polyacrylamide gel electrophoresis and activity staining. The patterns differ with respect to the presence and the relative intensity of the individual fractions. Family studies strongly suggest that these isozymogen patterns are encoded by allelic haplotypes, encompassing different numbers and types of PGA genes. In this paper, we confirm the essential features of this multigene model. We establish the relationship between the haplotypes and the corresponding isozymogen patterns by determination of the PGA polymorphism at both the DNA and the protein level in 117 Dutch individuals, 60 of whom were unrelated. The combination of HindIII and EcoRI restriction fragment length polymorphisms (RFLPs) has enabled us to define different haplotypes, which are shown to segregate within families. Most genes are characterized by their specific EcoRI fragments. The HindIII RFLP is in strong linkage disequilibrium with PGA genes showing strong expression of the relevant isozymogen. Although a general picture of the relationship between genotypes and phenotypes is emerging, there are exceptions, suggesting that rare haplotypes evolve by unique crossover events.

A study of 948 Norwegians including 118 matings with 429 children provided evidence that the Pg I group of pepsinogens must be coded for by more than one gene locus. At the Pg5 locus the alleles Pg5N, Pg5F and Pg5S with frequencies 0.644, 0.059 and 0.004, each code for a single electrophoretic isozyme band responsible for Pg phenotypes Pg 5, intense Pg (4) and Pg 5S, and a null-allele Pg5o with frequency 0.293. The Pg 2, Pg 3 and weak Pg 4 bands are not coded for by alleles at the Pg5 locus and differ from the Pg5 gene products in the oligopeptides split off by pepsinogen-pepsin conversion. The Pg5 alleles differ from each other in the pepsin-coding part of the gene. Intensity variations of Pg 5 have no simple genetic explanation.

Total human pepsinogen (PG) was isolated from gastric fundic mucosa and PGA (formerly called PGI) from urine, using standard ion-exchange and gel filtration techniques. Gastric PGA was separated from PGC (formerly called PGII) either by immunoaffinity or high resolution ion-exchange chromatography (fast protein liquid chromatography, Pharmacia, Uppsala, Sweden). The individual PGA isozymogens 2, 3, 4 and 5 could be isolated to homogeneity with the aid of the same ion-exchanger. Evidence was obtained for the existence of secondary modifications of the PGA fractions 3, 4 and 5, electrophoretically overlapping the primary (genetic) isozymogens.

A 0.9kb cDNA fragment, corresponding to a large part of Rhesus monkey pepsinogen A mRNA, was used as probe for the chromosomal localization of the human pepsinogen A gene(s) using human-rodent somatic cell hybrids. Southern blot analysis of 14 human-Chinese hamster and three human-mouse cell hybrids, strongly indicates that the human PGA locus is on chromosome 11. The human-mouse hybrids, containing a translocation involving chromosome 11, allow sublocalization to the region q12-pter.

Overlapping recombinant clones that appear to encompass the entire renin gene, named Ren 1, have been isolated from a library of BALB/c mouse genomic DNA fragments. Based on restriction endonuclease mapping and DNA sequence analysis, Ren 1 spans 9.6 kb and contains nine exons interrupted by eight intervening sequences of highly variable size. The first exon, encoding the signal peptide of preprorenin, is separated from the eight following exons by a 3-kb intron. These eight exons are organized into two clusters of four separated by a 2-kb intron. DNA stretches encoding the aspartyl residues, which are part of the active site of renin, are located at homologous positions in both clusters. Our results show that aspartyl protease genes have arisen by duplication and fusion of an ancestral gene containing five exons. The estimated date of the duplication event of the mouse renin genes Ren 1 and Ren 2 is discussed.

A genetic analysis of human urinary pepsinogen isozymes is presented. Nine discrete phenotypes were identified in a population survey of 215 unrelated Caucasian individuals. The phenotypes were characterized by differences among the staining intensities of the activated group I pepsinogens, Pg 5, Pg 4, Pg 3, and Pg 2. The genetic studies demonstrated that the codominant expression of four alleles, PgA, PgB, PgC and PgD, at a single genetic locus determined the nine phenotypes identified. Linkage analysis excluded close linkage of the Pg locus with the chromosome 6 markers HLA, GLO1, and Bf.

An 1110-base-pair cDNA clone for human cathepsin D was obtained by screening a lambda gt10 human hepatoma G2 cDNA library with a human renin exon 3 genomic fragment. Poly(A)+ RNA blot analysis with this cathepsin D clone demonstrated a message length of about 2.2 kilobases. The partial clone was used to screen a size-selected human kidney cDNA library, from which two cathepsin D recombinant plasmids with inserts of about 2200 and 2150 base pairs were obtained. The nucleotide sequences of these clones and of the lambda gt10 clone were determined. The amino acid sequence predicted from the cDNA sequence shows that human cathepsin D consists of 412 amino acids with 20 and 44 amino acids in a pre- and a prosegment, respectively. The mature protein region shows 87% amino acid identity with porcine cathepsin D but differs in having nine additional amino acids. Two of these are at the COOH terminus; the other seven are positioned between the previously determined junction for the light and heavy chains of porcine cathepsin D. A high degree of sequence homology was observed between human cathepsin D and other aspartyl proteases, suggesting a conservation of three-dimensional structure in this family of proteins.

The entire human pepsinogen C gene has been isolated from a cosmid genomic library. The nucleotide sequences of all the exons and the 5'- and 3'-flanking regions of the gene are presented. The organization of the gene is fundamentally compatible with those of other aspartic proteinases, allowing us to conclude that the genes of these aspartic proteinases including pepsinogen C are derived from a common ancestral gene. The predicted 388-residue amino acid sequence of human pepsinogen C consists of a signal sequence of 16-amino acid residues, an activation peptide of 43 residues, and the mature pepsin of 329 residues containing the two active-site aspartic acids. In the light of present notions about eukaryotic gene expression, possible regulatory roles of the oligonucleotide DNA sequences in the promoter region of the gene are discussed.

The pepsinogen gene is hypomethylated in the stomach, in which it is expressed. For demonstration that this hypomethylation of the pepsinogen gene in the stomach reflects pepsinogen-producing cells, we analyzed fractions of dispersed mucosal cells with various contents of pepsinogen-producing cells prepared from guinea pig stomach by centrifugal elutriation. mRNA expression and the extent of hypomethylation of the pepsinogen gene in each fraction was closely correlated with the content of pepsinogen-producing cells. These results suggested hypomethylation of the pepsinogen gene in pepsinogen-producing cells and differential pepsinogen gene methylation in cell subpopulations in the stomach.

Urinary pepsinogen A (PGA or PG I) phenotypes and serum PGA levels were studied in MZ and DZ twins and their parents. In 45 out of 48 MZ twin pairs PGA patterns were completely identical, while 3 MZ twin pairs showed minor differences in the relative intensity of the Pg5 isozymogen. This suggests that the intensity of this isozymogen may be influenced by nongenetic factors. There was little difference in the interclass correlations of serum PGA levels between MZ and DZ twins, indicating a large contribution of common environmental factors to serum PGA levels. This is in contrast with previous studies.

The genes coding for human pepsinogen (PGA3, PGA4, and PGA5) were assigned to chromosome region 11q13 by in situ hybridization. Previously we localized the PGA gene complex to a centromeric region of chromosome 11 (p11----q13) by Southern blot analysis of mouse-human somatic cell hybrids. Our in situ hybridization results confirm this assignment and further localize the genes to a smaller region on the long arm.

The entire rat pepsinogen C gene has been isolated from a rat genomic library, using the rat pepsinogen C cDNA as a probe. Southern blot analysis showed that there exists at least two rat pepsinogen C genes. The nucleotide sequences of the coding regions and the 5'- and 3'-flanking regions of one of the rat pepsinogen C genes have been determined. This gene is split into 9 exons interrupted by eight intervening sequences. The 5'-flanking region is similar to that of the human pepsinogen C gene, but only the former has the core sequence of the Sp1 binding site. The amount of transcripts of the rat pepsinogen C genes was found to increase during development, and a similar increase was shown to be induced by injection of hydrocortisone. As a candidate of a factor which regulates the transcription, we found a 25-kDa protein by Southwestern blotting. It binds to a specific site in the 5'-flanking region of the gene only in the presence of Mg2+ ion, and it is present in the nuclear fraction of the gastric mucosa but not of the liver.

The total RNAs were extracted from human, swine, rat, and calf gastric mucosae, and translated in vitro in the presence of radiolabeled amino acids using a wheat germ cell-free system. Upon sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis of the translation products, a protein band with a molecular weight of about 43,000 was obtained in each case as one of the major products. These products could be specifically immunoprecipitated with a corresponding anti-pepsinogen or anti-chymosin antiserum. Radiosequence analysis of these translation products purified by SDS-polyacrylamide gel electrophoresis showed that each of them is a precursor form, i.e., prepepsinogen or preprochymosin, having an amino-terminal extension peptide (signal sequence) comprising 15 (human and swine) or 16 (rat and calf) amino acid residues. The primary structures of these signal sequences were determined to be as follows: (Sequence: see text). These signal sequences share common characteristics with those of other pre-secretory proteins, i.e., the presence of positive charges in the NH2-terminal region, hydrophobic amino acid clusters in the interior part, and amino acids with short side chains at the site of cleavage by the signal peptidase.

Serum pepsinogen 1, serum gastrin, ABO blood groups, secretor status of ABH blood group substances and behavioral factors were studied in 15 patients with duodenal ulcer and 61 their relatives affected and unaffected to duodenal ulcer. Duodenal ulcer patients had hyperpepsinogenemia 1 either with or without a positive family history of duodenal ulcer. Serum gastrin level was higher in duodenal ulcer patients and unaffected relatives with a positive family history of duodenal ulcer than those with a negative family history. Non secretor status was frequently observed in duodenal ulcer patients with a positive family history. There was no difference in behavioral factors between duodenal ulcer patients and unaffected relatives with a positive family history. It is concluded that genetically determined variables such as hyperpepsinogenemia 1 and non secretor status play an important role on the susceptibility to duodenal ulcer in subjects with a positive family history, and hypergastrinemia may be subclinical marker of familial aggregation of duodenal ulcer.

Three human cosmid clones containing pepsinogen A (PGA) encoding sequences were isolated from a genomic bank derived from a single individual. One cosmid contains two PGA genes in tandem in a head-to-tail orientation, while the other two cosmids each contain a single PGA gene. The three cosmids were characterized by restriction mapping and sequence analysis (exons 1 and 2 and flanking regions). As judged from these data, three of the four PGA genes isolated appear to be nearly identical, but one of the tandem genes is clearly different from the other genes. The first exon of all four genes codes for the same amino acid sequence. However, in the second exon of one of the tandem genes we found a nucleotide substitution giving rise to a Glu----Lys substitution of the 43rd amino acid residue of the activation peptide, leading to a charge difference of the corresponding isozymogens. The presence of two distinct PGA genes in the isolated gene pair conclusively proves the multigene structure of the PGA system. These genes might be responsible for at least part of the electrophoretic polymorphism at the protein level.