Enamel matrix serine proteinase 1: stage-specific expression and molecular modeling.

Enamel proteins are cleaved by proteinases soon after their secretion by ameloblasts. Intact proteins concentrate in the outer enamel at or near the growing tips of the enamel crystallites while cleavage products accumulate in the deeper enamel. In the transition and early maturation stages there is a dramatic increase in proteolytic activity. This activity, coupled with the diminished secretory and increased reabsorptive functions of ameloblasts, leads to a precipitous fall in the amount of enamel protein in the matrix. Recently we have cloned and characterized an mRNA encoding a tooth-specific serine proteinase designated enamel matrix serine proteinase 1 (EMSP1) [Simmer et al., JDR (1998) 77: 377]. EMSP1 can be detected in the inner enamel during the secretory stage and its activity increases sharply during the transition stage. Stage-specific Northern blot analysis demonstrates this increase is accompanied by a parallel increase in the amount EMSP1 mRNA. A 3-dimensional computer model of EMSP1, based upon the crystal structure of bovine trypsin, has been generated and analyzed. All six disulfide bridges as well as the active site are conserved. Changes in the peptide binding region and the specificity pocket suggest that interaction of the proteinase with protein substrates is altered, potentially causing a shift in substrate specificity. The calcium binding region of trypsin is thoroughly modified suggesting that the calcium independence of EMSP1 activity is due to an inability to bind calcium. The three potential N-linked glycosylation sites, N104, N139 and N184, are in surface accessible positions away from the active site.

The catalytic mechanism of the amidotransferase domain of the Syrian hamster multifunctional protein CAD. Evidence for a CAD-glutamyl covalent intermediate in the formation of carbamyl phosphate.

The multifunctional protein CAD catalyzes the first three steps in pyrimidine biosynthesis in mammalian cells, including the synthesis of carbamyl phosphate from bicarbonate, MgATP and glutamine. The Syrian hamster CAD glutaminase (GLNase) domain, a trpG-type amidotransferase, catalyzes glutamine hydrolysis in the absence of MgATP and bicarbonate (Km = 95 microM and kcat = 0.14 s-1). Unlike E. coli carbamyl phosphate synthetase (Wellner, V.P., Anderson, P.M., and Meister, A. (1973) Biochemistry 12, 2061-2066), a stable thioester intermediate did not accumulate when the mammalian enzyme was incubated with glutamine. However, a covalent adduct could be isolated when the protein was denatured in acid. The steady state concentration of the intermediate increased with increasing glutamine concentration to nearly one mole per mole of enzyme with half saturation at 105 microM, close to the Km value for glutamine. The adduct formed at the active site of the glutaminase domain. The rate of breakdown of the intermediate (k4), determined directly, was 0.17 s-1 and the rate of formation (k3) was estimated as 0.52 s-1. In the absence of MgATP and bicarbonate, k4 = kcat indicating that the decomposition of the intermediate is the rate-limiting step. The intermediate was chemically and kinetically competent, and the glutamine dissociation constant (330 microM) and rate constants were consistent with steady state kinetics and accurately predicted the steady state concentration of the intermediate. These studies suggest a mechanism similar to the cysteine proteases such as recently proposed by Mei and Zalkin (Mei, B., and Zalkin, H. (1989) J. Biol. Chem. 264, 16613-16619) who identified a catalytic triad in glutamine phosphoribosyl-5'-pyrophosphate amidotransferase, a purF-type enzyme. MgATP and bicarbonate increased kcat of the glutaminase reaction 14-fold by accelerating both the rate of formation and the rate of breakdown of the intermediate, and prevented the accumulation of the intermediate; however, the Km value for glutamine was not significantly altered. The instability of the thioester intermediate leads to appreciable hydrolysis of glutamine in the absence of the other substrates. However, bicarbonate alone spares glutamine by increasing the Km and Ks of glutamine to 600 and 8960 microM, respectively, thus reducing kcat/Km 3-fold when MgATP is limiting. In the absence of MgATP and bicarbonate, ammonia decreased the rate of hydrolysis and the accumulation of the thioester intermediate indicating that ammonia had direct access to the thioester at the GLNase domain active site.(ABSTRACT TRUNCATED AT 400 WORDS)

Intracellular location of the multidomain protein CAD in mammalian cells.

The first three steps of mammalian de novo pyrimidine biosynthesis are catalyzed by the multifunctional protein CAD, consisting of glutamine-dependent carbamylphosphate synthetase, aspartate transcarbamylase, and dihydroorotase. The intracellular distribution of CAD in two hamster cell lines, BHK 21 and BHK 165-23 (a strain in which the CAD gene was selectively amplified), was determined by differential centrifugation and by two different cytochemical immunolocalization methods. Ammonia-dependent carbamylphosphate synthetase I was found in both cell types at a concentration of 0.01% of the total cell protein, so its distribution was also determined as a control for possible cross-reactivity of the CAD antibody probes and as a mitochondrial marker. CAD was localized in the cytoplasmic compartment and almost completely excluded from the nucleus. A punctate staining pattern suggested that it was not uniformly dispersed throughout the cytosol (unlike typical soluble proteins) but was associated with subcellular organelles. Although there was a slight tendency for CAD to be localized in the vicinity of the nuclear envelope, the amount of staining was much less than expected from differential centrifugation, which showed that 30% of the protein was found in the nuclear fraction. No interactions with other subcellular components could be detected by centrifugation. It is possible, however, that CAD is associated with subcellular structures that cosediment with the nuclei. Despite a 150-fold increase in CAD concentration in the over-producing cells, the distribution of the protein was unaltered. CAD was not concentrated near the mitochondria where the next enzyme of the de novo pathway, dihydroorotate dehydrogenase, is localized, which indicates that the intermediate dihydroorotate is not channeled, but rather dissociates from CAD and diffuses through the bulk cellular fluid.