Evidence for the presence of a kininogen-like species in a case of total deficiency of low and high molecular weight kininogens

Low and high molecular weight kininogens (LK and HK), containing 409 and 626 amino acids with masses of ~65 and 120 kDa after glycosylation, respectively, are coded by a single gene mapped to the human chromosome 3 by alternative splicing of the transcribed mRNA. The NH2-termini Glu1-Thr383 region, identical in LK and HK, contains bradykinin (BK) moieties Arg363-Arg371. LK, HK and their kinin products Lys-BK and BK are involved in several biologic processes. They are evolutionarily conserved and only 7 patients, all apparently normal, have been reported to lack them. In one of these patients (Williams trait), a codon mutation (Arg178 → stop) has been blamed for the absence of LK and HK. However, using Western blots with 2 monoclonal anti-HK antibodies, one that recognizes the region common to LK and HK and the other that recognizes only HK, I detected ~110-kDa bands in the plasma of this LK/HK-deficient patient vs ~120-kDa bands in normal human and ape plasmas. With polyclonal anti-Lys-BK antibody, which strongly detects BK cleaved at its COOHterminus in purified HK, I detected ~110-kDa bands in the normal and the deficient plasmas. Western blots with a monoclonal antiprekallikrein (PK) antibody showed that surface activation of PK and distribution of PK activation products, both dependent on HK, were similar in these plasmas. These findings suggest that a mutant gene yielded a kininogen-like species possibly involving aberrant mRNA splicing structurally different from normal HK, but apparently with the capacity to carry out seemingly vital HK functions. Correspondence


Introduction
Kininogens and their cleavage products kinins are multifunctional molecules well perserved in evolution (1,2).In humans, low (LK) and high molecular kininogens (HK) (~65 kDa and ~120 kDa, respectively) are single chain glycoproteins, which contain kinin moieties.Specific hydrolyses of LK and HK, catalyzed by tissue and plasma kallikreins, release Lys-bradykinin (Lys-BK) and BK, respectively, and divide each kininogen into 2 fragments, heavy and light chains, bridged by a disulfide bond.Both LK and HK stem by alternative splicing of the mRNA transcribed from a single gene mapped to chromosome 3q26-qter in humans that contains 11 exons (3,4).The identical sequences of LK and HK heavy chains -362 amino acids (aa) -are encoded by exons 1-10 (5).Exon 10 also encodes Lys-BK and BK (10 and 9 aa, respectively), part of LK light chain (12 aa) and the entire HK light chain (255 aa), and exon 11 encodes the 26 aa of the COOH-terminal sequence of LK light chain.The homology among the genes that encode human, cow and rat LK and HK, rat T-I and T-II kininogens and human stephins and cystatins shows that they originate from the same ancestor (5,6).It also shows that the function of these proteins as cysteine protease inhibitors, with activities located in LK/HK heavy chains, has been conserved through evolution (6)(7)(8).In addition, activation of the contact system of coagulation, as assayed in vitro, requires binding of HK light chain to prekallikrein (PK) and factor XI. The fact that kininogens and kinins have been conserved through evolution, participate in multiple biologic processes including inflammation, regulation of blood pressure and vascular permeability, cardioprotection and pain modulation (9)(10)(11)(12)(13)(14)(15)(16)(17), and that kinin receptors are ubiquitous in mammalian tissues, including brain, suggests that HK/LK and kinins are essential to normal biologic functions.Nevertheless, 7 persons have been reported to be totally deficient in HK/LK.In each case they appeared to be physiologically normal.Genetic analysis done in only one of these patients (Williams trait (18)) by Cheung et al. (19) attributes this deficiency to a codon mutation (Arg 178 → stop) in exon 5.These authors suggest that their failure to detect the expected 21.5-kDa truncated protein could be explained by stability or transport anomalies of the defective protein.
However, the unexpected detection of a ~110-kDa band on Western blots of Williams trait plasma with the monoclonal anti-HK light chain antibody mAb 371-28 (20), in our laboratory, indicated the presence of a kininogen-like species in this plasma.This finding was confirmed, as described in the present report, by comparing Western blots of Williams plasma with those of normal human and ape plasmas, which show intact ~120-kDa HK with mAb 371-28 and the anti-LK/HK heavy chain mAb 11-2 (21), and cleaved ~110-kDa HK with anti-Lys-BK (22) and anti-BK (23) antibodies.In addition, HK function was tested by comparing PK activation on a surface and distribution of its activation products, which depend on HK (24,25), in the deficient and normal plasmas (preliminary report, Ref. 26).

Sources of chemicals, plasmas and antibodies
The source and grade of chemicals have been reported before (27).Solutions of purified human plasma HK (Enzyme Research Laboratories, South Bend, IN) were adjusted to a neutral pH with Tris before the assay.SDS-PAGE analysis of the reduced and the nonreduced neutralized HK showed only a ~120-kDa band.The following plasmas were obtained from George King Biomedicals (Overland Park, KS): pooled, normal human plasma; PK-deficient plasma with PK undetectable by clotting, amidolytic and Western blot assays, and kininogen-deficient, Williams trait plasma lot GK1602-316E1.Chimpanzee (Pan troglodytes) and orangutan (Pongo pygmaeus) plasmas were kindly supplied by the Yerkes Regional Primate Research Center (Atlanta, GA).Blood from a female (D) and a male (T) was drawn into tubes containing 3.8% sodium citrate and 2.5% dextrose (9 blood:1 citrate/dextrose mixture, v:v).Blood was centrifuged for 5 min at 2,000 g, at 23 o C. One sample of plasma T was frozen and thawed several times, which resulted in marked activation.Only plasticware or siliconized glassware was used for blood or plasma samples.
Monoclonal anti-human plasma heavychain PK antibody (mAb 13G11) (28) -patents: U.S. #5,444,156, European # 0210029, and Canadian #1312564, inventors D.C. Veloso and R.W. Colman -was purified from ascites produced at Temple University.and was provided by that University.The following antibodies were generous gifts: mAbs 371-28 and 11-2 from Dr. Sesha Reddigari and Dr. Allen P. Kaplan, and rabbit polyclonal anti-BK and anti-Lys-BK antibodies from Dr. David Proud and Dr. A. Guillermo Scicli, respectively.Blotting detection kit for mouse antibodies (Amersham kit) was supplied by Amersham Corp. (Arlington Heights, IL), and alkaline phosphatase affinity purified goat anti-rabbit IgG by Cappel, Organon Teknika Corp., West Chester, PA.Plasmas and antibodies were stored at < -70 o C.

Amidolytic and clotting activities
Amidolytic activity was determined as reported by Veloso et al. (27).Briefly, activation was effected by mixing plasmas with an equal volume of a kaolin suspension (1 mg/ml 0.15 M NaCl) for 5 s in a Vortex, followed by a 55-s incubation at 23 o C. Initial velocities of D-Pro-Phe-Arg-p-nitroanilide hydrolysis (0.8 mM at zero times) in the activated plasmas (5-15 µl) were measured at 37 o C and 405 nm with a Beckman DU-7 spectrophotometer.Clotting activity was determined by the method of Proctor and Rapaport (29) using the GK1602-316E1 plasma as source of proteins.

Western blots
Plasmas (8 µl) with or without kaolin activation were subjected to SDS/7.5, 8.0 or 10%-PAGE according to Laemmli (31) using a standard size Hoefer apparatus as described by Veloso et al. (27).Briefly, plasma was activated by mixing with an equal volume of a kaolin suspension (5 mg/ml 0.15 M NaCl) in a Vortex for 5 s followed by incubation at 4 o C for 2 h, 23 o C for 20 min or 37 o C for 15 min, and centrifugation (1,000 g, 3 min).For immunoblotting of the resulting electrophoretogram, the method of Towbin et al. (32), as modified by Burnette (33), was used, except that electrotransfer to nitrocellulose membranes was done at 10 o C for 15-17 h with a Transphor model TE50 (Hoefer Scientific Instruments, San Francisco, CA).Free sites were blocked with 5% casein adjusted to pH 7.4 with Tris (2-h incubation, 23 o C).The same casein solution containing 0.05% (w/v) NONIDET P-40 was used for washing the membranes (4-5 times followed by rinsing with H 2 O) after their incubation (2 h, 23 o C) with the iodinated or diluted antibodies as follows: 125 I-mAb 13G11 (~1 µg, 6 x 10 2 Bq/ml) for PK/kallikrein detection as reported (34); mAb 371-28 (2.5 µg/ ml) or mAb 11-2 (3 µg/ml) for kininogen binding followed by detection with sheep anti-mouse biotinylated antibody (the secondary antibody) and a streptavidin-biotinylated phosphatase complex as the signal generating system, as described by the Amersham kit manufacturers, and Prouds antibody (diluted 20,000-fold) or Sciclis freeze-dried antibody (1 ml of H 2 O added before a 2,500-fold dilution) for binding of the COOH-termini of BK or Lys-BK fol-lowed by binding to goat anti-rabbit antibody conjugated to alkaline phosphatase (diluted 1,500-fold).Antibodies were diluted with the casein solution.Nitro-blue tetrazolium and 5-bromo-4-chloro-3-indoyl phosphate dissolved in diethanolamine containing 5 mM MgCl 2 were the substrates in the alkaline phosphatase reactions.For comparison of proteins detected by different antibodies, after free site blocking, the membrane was split and then processed separately.

Recognition of kininogens or kininogen-like proteins in Williams' trait plasma
A ~110-kDa band was recognized by mAb 371-28 on Western blots of the HK/ LK-deficient plasma GK1602-316E1 (Williams trait plasma).Like normal human and orangutan plasmas, Williams plasma showed only the ~85-and 88-kDa PK variants seen in normal plasmas (Figure 1A).PK concentrations determined by densitometry of the autoradiographs of the Western blots were 48% those in normal human plasma, a value comparable to that of 45% first reported for Williams plasma (18).Figure 1B shows that, unlike PK-deficient plasma and normal human and chimpanzee plasmas, a ~110-kDa band(s), and not the 120-kDa HK band observed in normal plasmas, was detected in the deficient plasma with mAb 371-28.An additional ~110-kDa band, also observed in the PK-deficient plasma, is probably an HK degradation product.

Recognition of the HK heavy and light chains by monoclonal antibodies in normal and Williams' plasmas
Similar to mAb 371-28 (Figure 2A), mAb 11-2, which specifically recognizes the heavy chain shared by HK and LK, also recognized the ~110-kDa band in the Williams plasma and the 120-kDa band in normal human and chimpanzee plasmas (Figure 2B).The 90-103-kDa bands probably correspond to cleaved HK. mAb 11-2 barely detected LK on Western blots of the deficient or normal plasmas.

Detection of a kinin antigen(s) in Williams' plasma
mAb 371-28 recognized ~120-kDa bands in purified HK at a position similar to that in normal human and chimpanzee plasmas and also the ~110-kDa band in the Williams plasma (Figure 3A).Reactivity of these proteins was also compared with those processed under similar conditions using Sciclis polyclonal anti-Lys-BK antibody (Figure 3B), which specifically recognizes BK cleaved at its COOH-terminus in HK or HK fragments (35).Sciclis antibody recognized bands at similar positions in the deficient and normal human plasmas (Figure 3B) but not in purified HK, apparently because the purified HK was intact.Bands of ~110 kDa were also detected on blots of the Williams and the D and T plasmas with Prouds polyclonal anti-BK antibody (Figure 4B), while immunoblots run simultaneously with mAb 371-28 showed the expected 110-and the 120-kDa bands in their plasmas (Figure 4A).Intense bands (90 to 103 kDa) were detected with both mAb 371-28 and Prouds antibody on blots of the T plasma activated by several freeze-thaw cycles (lanes 3 of Figure 4A and B).As expected, an increase of cleaved HK in this activated plasma was accompanied by a decrease of intact, 120-kDa HK (Figure 4A, lane 3).Curved, fine bands (~50 kDa), visible on some of the kininogen blots, have not yet been identified.

Discussion
This is the first report that provides evidence for the presence of kininogen-like species in a patient plasma, Williams plasma, reported to lack LK and HK and not to be affected by their absence.To reconcile the apparently essential role of kinins and their kininogen precursors to normal biologic functions with their deficiency in apparently normal people, I hypothesized (36) that a compensatory biologic mechanism or the produc-tion of one or more mutant unstable proteinsundetectable by the procedures then available, but potentially operative in vivo -might be present in the LK/HK deficient patients.In contrast, Cheung et al. (19) claimed that an Arg 178 → stop mutation in exon 5 of the Williams kininogen gene was responsible for total LK/HK deficiency.Their efforts to detect the truncated 21.5-kDa protein on Western blots with the anti-LK/HK heavy chain mAb 2B5 were unsuccessful, and a compensatory mechanism was not apparent.However, as shown in the present report (Figures 1-4), the anti-HK light chain mAb 371-28 detected a ~110-kDa band on Western blots of the deficient plasma vs the expected 120-kDa bands of normal human and ape plasmas; bands at these positions were also seen on Western blots with the anti-LK/HK heavy chains mAb 11-2 (Figure 2).mAb 11-2 did not detect LK on Western blots of Williams or normal plasmas although it recognized purified LK and kininogen heavy chain, and normal human plasma with ELISA in microtiter plates (20,21).Because this ELISA does not differentiate LK from HK in plasmas, one cannot rule out that the failure of mAb 11-2 to detect LK on Western blots was probably due to an effect of one or more plasma components on epitope availability.The lower intensity of the 110-kDa bands recognized by mAb 371-28 and 11-2 on Williams plasma -an apparent quantitative deficiency -might be due to an effect of the mutation on protein synthesis, transport or resistance to protease degradation.As a follow-up up step, the presence of kinins was investigated with Prouds anti-BK and Sciclis anti-Lys-BK antibodies.As expected from the study by Tayeh et al. (35), Sciclis antibody did not detect or barely detected the intact ~120-kDa HK in purified HK and normal plasma.Also, as expected from the study by Tayeh et al. (35), this antibody strongly detected kinin in normal plasma with HK cleaved at its COOH-terminus, which has a higher electrophoretic mobility than intact, purified HK, and kinin in Williams plasma.The similar position of the bands detected with mAbs 371-28, 11-2 and anti-kinin antibodies in Williams plasma in combination with the stronger detection by the anti-kinin antibodies than that by the mAbs suggests the presence of only kininogen cleaved at its BK COOH-terminuswith no effect on electrophoretic mobility of the kininogen-like species -or epitope availability in the intact molecule due to structural changes.The failure of Sciclis (or Prouds) antibody to detect kinin in LK of normal or Williams plasmas was probably due to LK intactness or structural changes affecting epitope detection.Further evidence for the presence of a kininogen-like species in the deficient plasma comes from the following: 1) the position of the bands (90 to 103 kDa) detected by mAb 371-28 and Sciclis and Prouds antibodies in both normal and deficient plasmas was comparable to that of the bands detected in purified HK cleaved by kallikrein by mAb 371-28 (20) and Sciclis antibody (35), and 2) the intensity of these bands was stronger in the plasma subjected to activation with several freeze-thaw cycles.Differences between band positions in normal and deficient plasmas could 45 not be evaluated because, in such a complex milieu like that of plasma, the degree of activation (not known in these plasmas) is expected to affect band numbers and positions.
A functional test based on the requirement of HK for optimum activation of plasma PK on a surface showed results similar to those obtained with Williams and normal plasmas, except that a lower activation rate was apparent in the deficient plasma probably due to a lower amount of the protein in this plasma.Like normal plasmas, the HK/ LK-deficient plasma showed only the 85and 88-kDa PK bands (Figure 1A).Also, the fact that the pattern of PK activation and distribution of the kallikrein formed among its inhibitors is similar in the Williams and normal plasmas, although HK affects kallikrein binding to α 2 M but not to C1 inhibitor (25), further indicates the presence of a functional kininogen-like species.In addition, the similarity between the Williams and normal plasma patterns at 4, 23 and 37 o C (Figure 5), under conditions in which binding of kallikrein to α 2 M or C1 inhibitor depends on inhibitor concentrations and incubation temperatures (37)(38)(39), further indicates the presence of a kininogen-like function in the Williams plasma.That kallikrein bound to α 2 M can hydrolyze D-Pro-Phe-Arg-p-nitroanilide, while C1 inhibitor rapidly inhibits this kallikrein activity at 37 o C, may explain why, at 4 o C, Williams plasma showed 8% of the amidolytic activity of that in normal plasma (21,40) while at 37 o C its amidolytic and clotting activities were undetectable.The negative results were probably due to rapid inactivation of the small amounts of kallikrein formed in the deficient plasma by C1 inhibitor, the main inhibitor of kallikrein at 37 o C. Concentrations of free kallikrein or kallikrein bound to α 2 M would be too low at 37 o C to show clotting and amidolytic activities in Williams plasma under conditions (37 o C) used for plasma assays.
To explain my finding of a ~110-kDa kininogen-like species in the plasma reported by Cheung et al. (19) to be totally deficient in LK/HK due to the mutation in exon 5, I suggest an alternative consequence of this mutation: aberrant mRNA splicing with skipping of exon 5 yields the 110-kDa kininogen-like species.In fact, exon 5 encodes 36 amino acids that, after glycosylation, might correspond to the ~10-kDa fragment missing in the deficient plasma.The bases for this suggestion are the following: 1) a mutation in an exon may result in skipping of that exon (41); 2) in a patient, hereditary protein S deficiency has been found to be associated with skipping of an exon that harbored a Ser 62 → stop mutation in that exon (42); 3) Val-Val-Ala-Gly, a tetrapeptide in the reactive site of papain inhibitors (7), is encoded by both exons 5 and 8, and 4) exon 5 encodes 36 of the 222 aa of domain 2, while exon 8 encodes 29 of the 116 aa of domain 3, homologous domains that exhibit similar functions as cysteine protease inhibitors (7,8).Therefore, deletion of exon 5 should not greatly affect HK size or kinin and kininogen functions.Questions were raised but not answered in this study due, at least in part, to the discontinuity of Williams plasma supplythe only plasma reported to have a total LK/ HK deficiency attributed by Cheung et al. (19) to premature protein termination.However, it is desirable and probably feasible, in future research, to identify the putative aberrant mRNA in stored, small amounts of platelet-rich Williams plasma by using a transcription/polymerase chain reaction/cycle sequencing procedure, such as that described by Yamazaki et al. (43) and Okamoto et al. (42).For a better understanding of gene/ protein structures and LK/HK function correlates, more studies are needed in newly discovered Williams trait cases or in genes mimicking that of Williams case.For example, mimic genes might be built by molecular means from normal LK/HK genes with the codon mutation (Arg 178 → stop) in exon 5 or with exon 5 deletion.Knock-out mice might also help to elucidate whether or not LK/HK functions are crucial to a normal life.

Figure 4 -
Figure 4 -Western blots of HK and HK-like species in Williams' and normal human plasmas.SDS/PAGE and Western blots of plasmas were done as indicated in the legend to Figure 1 with the following differences: A, HK or HK-like species in Williams' (lane 1), normal human female (D) (lane 2), and normal human male (T) plasmas with and without further activation (lanes 3 and 4, respectively) were probed with mAb 371-28; B, kinins in HK, HK-like species or their cleavage products in plasmas applied to gel lanes as indicated in (A) were probed with Proud's anti-BK antibody.Molecular masses of HK (120 kDa) or HK-like species (110 kDa), cleaved kininogen (~90 kDa) and an unidentified component (~50 kDa) are indicated.