T-kininogen, the major kininogen in rat plasma, releases Ile-Ser-bradykinin (T-kinin) when incubated with trypsin, but is not a substrate for tissue kallikrein. Enzymes able to release T-kinins from T-kininogen have been found in the rat submandibular gland, but precise identification of these enzymes and their possible relationship to kallikrein-like enzymes has not been established. We studied T-kininogenase activity in fractionated submandibular gland homogenate. The main T-kininogen catalytic enzyme was purified and characterized, and found to be identical to antigen gamma, a kallikrein-like enzyme which we have previously characterized. Of other identified kallikrein-like enzymes only tonin showed weak T-kininogenase activity, which was about 0.25% of that of antigen gamma. No other T-kininogen catalytic enzymes were observed. Antigen gamma released a kinin which was identified as T-kinin by reverse-phase h.p.l.c. The T-kininogenase activity of antigen gamma had a Km of 29 +/- 4 microM and a kcat/Km of 140 M-1.s-1, and was comparable with its high and low molecular mass-kininogenase activity (7.4 and 10 micrograms of kinin/h per mg respectively). In contrast, tissue kallikrein released 0.2 and 42,200 micrograms of kinin/h per mg respectively. Thus antigen gamma is a weak kininogenase. The isoelectric point of antigen gamma, but not its molecular mass, differed from that of other kallikrein-like enzymes. Isoelectrofocusing in flat-bed gels combined with immunostaining was therefore a convenient method for identification. The kallikrein-like nature of antigen gamma was demonstrated by its immunological similarity to tissue kallikrein and tonin and by 91% and 87% amino acid sequence similarity with tonin and kallikrein respectively (67 amino acids sequenced). Complete identity was also not observed with other sequenced kallikrein genes, mRNAs or proteins.

The submandibular gland of the rat contains several enzymes belonging to the kallikrein family. These include tissue kallikrein, antigen gamma (T-kininogenase), esterase B and tonin. In the present study, a new member of this family, which we have named KLP-S3, was identified and purified from the submandibular gland. KLP-S3 was classified as a kallikrein-like enzyme on the basis of its immunological similarity to other kallikrein-like enzymes and its showing 70% and 73% identity in partial amino acid sequence with tissue kallikrein and tonin respectively. Furthermore, the 44 sequenced amino acid residues showed complete correspondence to the mRNA S3 of the kallikrein gene family, which was the rationale for the name kallikrein-like protein (KLP) S3. KLP-S3 consisted of three isoenzymes with pI 6.75, 6.90 and 6.95, which significantly differed from those of other kallikrein-like enzymes. In conjunction with its immunological relationship to kallikrein, this parameter (pI) was considered robust enough to identify the enzyme during purification, since a specific physiological substrate for KLP-S3 has yet to be identified. In SDS/PAGE the three isoenzymes ran as one band with a molecular mass of 25,800 Da, which after reduction with 2-mercaptoethanol was split into two chains with molecular masses of 16,500 and 13,300 Da. In common with other kallikrein-like enzymes, KLP-S3 was inhibited by phenylmethanesulphonyl fluoride, and was thus classified as a serine protease. It was also inhibited by soya-bean trypsin inhibitor but not by aprotinin. It showed weak reactivity against the chromogenic substrates S2288, S2266, S2366 and S2302 (D-Ile-Pro-Arg 4-nitroanilide, D-Val-Leu-Arg 4-nitroanilide, Glu-Pro-Arg 4-nitroanilide and D-Pro-Phe-Arg 4-nitroanilide respectively) and did not cleave rat T-kininogen or dog high-molecular-mass/low-molecular-mass kininogen. Its specific angiotensin II-generating activity (angiotensin I as substrate) was 0.04% of that of rat tonin. KLP-S3 (1-100 nM) induced a statistically significant angiotensin-independent contraction of isolated rat aorta rings. The maximum contraction was 15% of the response to the alpha-adrenoceptor agonist phenylephrine (1 microM). The concentration of KLP-S3 in the rat submandibular gland was by single radial immunodiffusion estimated to be 47 +/- 3 micrograms/mg of protein.

The signal pathway for bradykinin-induced relaxation followed by contraction in the isolated rat duodenum was investigated by comparing the effect of blocking agents on the response to bradykinin and acetylcholine. The phospholipase C inhibitor U-73122 inhibited the relaxation induced by bradykinin, but had no effect on the contraction to either bradykinin or acetylcholine. The same response pattern was observed when the tissues were pre-treated with thapsigargin, a selective inhibitor of microsomal Ca2+ pumps. An inhibitor of non-voltage-dependent Ca2+ influx, SK&F 96365, inhibited the relaxant response to bradykinin and the contraction induced by acetylcholine, but not the contraction induced by bradykinin. In Ca2+-free Krebs-Henseleit buffer, the tissues failed to respond when they were exposed to either bradykinin or acetylcholine. When the tissues were partly depolarized (30 mM KCI), both bradykinin and acetylcholine induced contraction, while the relaxant response to bradykinin was almost completely abolished. Apamin (an antagonist of low-conductance calcium-activated K+ channel) together with charybdotoxin (CTX, an antagonist of large-conductance calcium-activated K+ channel) and CTX alone inhibited the relaxant but not the contractile response to bradykinin. We conclude that the biphasic response in isolated rat duodenum to bradykinin involves two distinct pathways. We propose that the relaxant component is induced indirectly via inositol-mediated increase in cytosolic Ca2+ in non-muscle cells with subsequent signals to the smooth muscle cells, whereas the contractile response is induced by direct effect on the smooth muscle cells.

The rat submandibular gland contains several members of the kallikrein family. In the present study we purified and raised an antiserum against one of these enzymes, i.e., esterase B, which was first described by Khullar et al. in 1986. N-terminal amino acid analysis revealed complete homology between esterase B and the kallikrein family gene RSKG-7. For characterization of the antiserum, flat-bed isoelectrofocusing with immunoblotting was superior to immunoelectrophoresis and double immunodiffusion in detecting and identifying crossreacting proteins. This was due to the fact that kallikrein-like enzymes were readily separated by isoelectrofocusing, and immunoreactivity was easily detected by the sensitive peroxidase-anti-peroxidase staining after blotting onto nitrocellulose membrane. Immunohistochemical controls were carried out accordingly, including homologous as well as crossreacting antigens. In the submandibular gland, esterase B was detected exclusively in all granular convoluted tubular cells, co-localized with tissue kallikrein and tonin. Some staining was also observed in striated duct cells; however, this staining reaction was induced by cross-reactivity with kallikrein, since staining was abolished by addition of kallikrein as well as esterase B to the primary antiserum. It was therefore concluded that like tonin and antigen gamma, but unlike kallikrein, esterase B was not detected in the striated ducts of the submandibular, parotid, or sublingual glands. This separation in anatomic distribution between esterase B and kallikrein may indicate that prokallikrein activation is not the only biological function of esterase B.

The signal pathway for bradykinin-induced contraction of the uterine smooth muscle was investigated by comparing the effect of blocking agents on bradykinin and oxytocin induced contractions of the isolated rat uterus in organ bath. The phospholipase C inhibitor U-73,122 abolished the effect of both bradykinin and oxytocin. Inhibition of non-voltage-dependent Ca2+ influx by SK & F 96,365 reduced the contraction induced by both agonists to about 20% of control. The tissues failed to contract when they were exposed to bradykinin or oxytocin in Ca(2+)-free Krebs-Henseleit buffer with 2 mM EDTA. Both bradykinin and oxytocin induced further contraction when the tissues were partially depolarized and partially contracted by 30 mM KCl. These observations suggest that bradykinin, like oxytocin, activates phospholipase C which generates IP3 with a subsequent release of Ca2+ from intracellular stores followed by store-operated Ca2+ influx. Thus, membrane potential independent steps appear to be important in bradykinin-induced contraction in the rat uterus.