A second expressed kininogen gene in mice

Edward G. Shesely, Chun-Bo Hu, François Alhenc-Gelas, Pierre Meneton, Oscar A. Carretero


We isolated PCR, RNA ligase-mediated rapid amplification of cDNA ends (RLM-RACE-PCR)-, and RT-PCR-generated clones from mouse kininogen family transcripts. DNA sequencing indicated that the clones were from two distinct genes. One set (K1) is from the previously reported mouse kininogen gene. The second set (K2) has an open reading frame, is 93% identical to K1 in the overlapping nucleotide sequence, and, unlike T-kininogens in the rat, encodes a bradykinin motif identical to K1. We discovered that K2 exists with two different 5′ ends. We used RT-PCR to determine the distribution and relative abundance of K1 and K2 mRNA in mouse tissues. K2 is transcribed and K1 and K2 are generally both expressed in the same tissues; however, they differ in their regulation of the alternative splicing event that yields either low-molecular-weight kininogen (LMWK) or high-molecular-weight kininogen (HMWK). For example, in the liver K1 is expressed as both HMWK and LMWK, whereas K2 is only expressed as LMWK. Conversely, in the kidney K2 is strongly expressed as both HMWK and LMWK, whereas K1 is not expressed as HMWK and expressed only very weakly as LMWK.

  • polymerase chain reaction
  • DNA sequencing
  • alternative splicing

kininogens are protein substrates for tissue and plasma kallikrein. When cleaved, they release kinin peptides that function as vasodilators and are involved in the cardioprotective effect of angiotensin-converting enzyme inhibitors. Mammals produce two types of kininogens, termed high-molecular-weight kininogen (HMWK) and low-molecular-weight kininogen (LMWK). Both HMWK and LMWK mRNA are encoded by the same gene via 1) alternative RNA splicing in exon 10 and 2) the use of different polyadenylation sites, resulting in the generation of two distinct types of kininogen mRNA.

Bovine, human, rat, and mouse LMWK and HMWK cDNAs have been cloned and sequenced (4, 5, 9). The sequences of the two additional kininogen family genes unique to rats, called T-kininogens, have also been determined (2). T-kininogens exist only as LMWK; they are not substrates for tissue and plasma kallikrein and do not have a standard bradykinin motif, but rather T-kinin (Ile-Ser-bradykinin). In addition, T-kininogens have a remnant noncoding DNA sequence that is homologous to the exon 10 HMWK-specific region of rat kininogen.

Previously, only a single kininogen gene has been identified in all species studied except for rats. Although a previous report indicated that mice likely have only a single kininogen gene (10), we (3) and others (1) have reported the discovery of a second expressed kininogen family gene in this species. Here we further examine this second kininogen gene's sequence and expression.


Initial generation of mouse HMWK-specific PCR fragments.

PCR primers specific to the mouse kininogen exon 10 HMWK region (primer set A) were chosen by homology among rat, mouse, and human HMWK DNA; where they differed, the mouse sequence was chosen. Primer set A includes 5′ primer AGAACTGTAAGTCCACCCTACA and 3′ primer GGATATCAGGGATCAAATCATC, with a predicted band size of 630 bp based on the published mouse HMWK-specific cDNA sequence. Another primer set (primer set B) was chosen, including the 5′ primer from the region common to both HMWK and LMWK just inside exon 10 (ACTGAAATGGCAAGAAGGCC) and the same 3′ primer used in primer set A, with a predicted band size of 703 bp. The 5′ primer in primer set B is in a region of poor sequence conservation between species. PCR was performed by adding both forward and reverse primers (each at 0.2 μM) and 100 ng of inbred mouse strain C57BL/6J genomic DNA template to a final volume of 25 μl in PCR buffer [mM: 10 Tris·HCl (pH 9.0) at 25°C, 50 KCl, and 1.5 MgCl2, with 0.1% Triton X-100] containing 0.5 U of Taq DNA polymerase and 0.2 mM dNTPs. The reaction was carried out with 35 cycles of denaturation at 94°C for 30 s, annealing at 60°C for 30 s, and extension at 72°C for 1 min with a PTC-100 programmable thermal controller (MJ Research, Waltham, MA). Agarose gel electrophoresis was performed with the above PCR amplifications; the resulting PCR fragments were visualized by ethidium bromide staining, and the amplified DNA was isolated and purified from nonstained lanes of the gels.

Cloning of PCR and RT-PCR fragments.

The 576-bp K2 amplification product and amplified DNA fragments for each RT-PCR described below were isolated from agarose gels (without ethidium bromide staining) with a QIAEX II gel extraction kit (Qiagen, Valencia, CA). These were subcloned into the pGEM-T Easy vector system (Promega, Madison, WI) per the manufacturer's protocol and transformed into competent Escherichia coli, and plasmid DNA was isolated with a Miniprep Kit (Qiagen).

RNA ligase-mediated rapid amplification of cDNA ends PCR.

To obtain clones with the K2 5′ end, we used the GeneRacer kit (Invitrogen, Carlsbad, CA) per the manufacturer's protocol. In this procedure, an RNA oligonucleotide of known sequence is ligated to the 5′ ends of decapped mRNA, which is then used in RT-PCR. One to five micrograms of total RNA is treated with calf intestinal phosphatase to remove 5′ phosphates. This has no effect on capped full-length mRNA but will remove non-mRNA or truncated mRNA from the ligation reaction. The sample is treated with tobacco acid pyrophosphatase, which removes the 5′ cap from full-length mRNA while leaving a 5′ phosphate required for ligation. The mRNA has the GeneRacer RNA oligomer ligated to the 5′ end with T4 RNA ligase, which provides a known priming site for the GeneRacer PCR primers. A reverse transcription reaction is then performed, using the 5′ oligomer-ligated mRNA as template for the GeneRacer oligo(dT) primer and avian myeloblastosis virus reverse transcriptase. A touchdown PCR amplification of the resulting reverse transcription is done with the proofreading Platinum Pfx DNA polymerase, the GeneRacer 5′ primer (specific to the GeneRacer RNA oligomer derived sequences), and either primer GGAGGCCTCCTTCGGATAGG (specific to K2 exon 10 HMWK and LMWK common region; primer set C) or primer ACATCACTGGATTCTTCTGC (specific to exon 11 LMWK2; primer set D). Cycling conditions follow the manufacturer's protocol. After successful amplification is confirmed by agarose gel electrophoresis of an aliquot of the PCR product, a second round of nested PCR amplification is done as above, except that the GeneRacer 5′ nested primer is used in place of the GeneRacer 5′ primer. The amplified PCR products are isolated from an agarose gel by excising the band, loading it into a SNAP column, and spinning it in a microcentrifuge. They are then cloned into plasmid vector pCR4-TOPO and transformed into competent One Shot TOP10 cells with a TOPO TA Cloning kit (part of the GeneRacer kit package). The transformed cells are plated on LB + amp plates, and plasmid DNA is isolated from the colonies with standard procedures.

To obtain K2 clones with the 3′ end of the transcript, we use the GeneRacer kit as above but start at the reverse transcription step with the GeneRacer oligo(dT) primer. Nested PCR is then done, using as 3′ primers first the GeneRacer 3′ primer and then the GeneRacer 3′ nested primer. The 5′ primer is either AAAGCCACATTTAGTAACAG (specific to the HMWK2 exon 10; primer set E) or ACTCCTGCTGACTTTAACAC (specific to LMWK2 exon 11; primer set F).

RT-PCR of kininogen family amplification products.

Reverse transcription (RT) reactions were carried out on mouse total RNA with a Moloney murine leukemia virus reverse transcriptase (MMLV-RT) system. Two micrograms of total RNA from various Swiss Webster mouse tissues including the liver, kidney, thymus, embryo, heart, lung, brain, testis, ovary, and spleen (Ambion, Austin, TX) was mixed with 0.5 μg of random primers in H2O to a final volume of 13 μl. This was incubated at 70°C for 5 min to melt secondary structures. Then 5 μl of 5× MMLV-RT reaction buffer, 5 μl of dNTPs (each at 2.5 mM), 40 U of rRNasin ribonuclease inhibitor, and 200 U of MMLV-RT were added, and samples were incubated at 37°C for 60 min. All non-RNA components were obtained from Promega.

PCR was performed as described above except that 2 μl of the above cDNAs was used as template, and the annealing temperature varied with the primer set. Primers used for PCR amplification and the resulting band sizes are as follows: K2 primer set G: 5′ primer CAAAGTCTCGACTGCAATGCT, 3′ primer CCATGTGGTTTCTGGTGTCC (amplifies a 331-bp K2-specific band); K1 primer set H: 5′ primer CAGCCCAGAGCTGAAGGAGG, 3′ primer CCTGGAGGCCTTCTTGCCAT (amplifies a 329-bp K1-specific band); LMWK1 primer set I: 5′ primer TGCCAAGCATTAGATATGACTGAA, 3′ primer CGCTCCACACATCCCAGG (amplifies a 248-bp fragment specific to K1 LMWK); LMWK2 primer set J: 5′ primer GCATTAGATAAGACTATTCCTATCCG, 3′ primer ACATCACTGGATTCTTCTGC (amplifies a 237-bp fragment specific to K2 LMWK). HMWK primer set A (above) amplifies a 630-bp band specific to K1 HMWK and a 576-bp band specific to K2 HMWK. Agarose gel electrophoresis was performed on the RT-PCR amplification products, which were visualized by UV light after ethidium bromide staining.

DNA sequencing.

All PCR, RT-PCR, and rapid amplification of cDNA ends (RACE)-RT-PCR clones were sequenced with SP6, T3, or T7 primers (as appropriate to the vector) and an automated DNA sequencing system (ABI 377 DNA sequencer, Applied Biosystems, Foster City, CA). PCR fragments were sequenced with the above automated system and HMWK primer set A.


PCR detection of a second kininogen family gene in mice.

We used PCR to determine whether there is more than one kininogen gene in mice. PCR primers were designed based on kininogen cDNA exon 10 HMWK-specific sequences conserved among human, bovine, and mouse DNA (primer set A; locations of all PCR primer sets are illustrated in Fig. 1). PCR amplification using this primer set and mouse genomic DNA as template produced the 630-bp fragment expected based on the mouse kininogen cDNA sequence and also amplified an additional 576-bp fragment, suggesting that a second kininogen gene exists in mice (Fig. 2). DNA sequence analysis of the two PCR fragments obtained with primer set A indicated that they were both derived from distinct kininogen genes. The 630-bp fragment is clearly a portion of the mouse kininogen cDNA described by Takano et al. (10) and is referred to as K1. The 576-bp fragment appears to be from a second kininogen family gene, which we call K2.

Fig. 1.

Location of the PCR primer sets used in this manuscript, showing a generic schematic of K1 and K2 RNA/cDNA. The bradykinin motif is indicated as a box. The low-molecular-weight kininogen (LMWK)-specific portion represents the result of LMWK-specific exon 11 being joined to the high-molecular-weight kininogen (HMWK)-LMWK common portion by the use of the alternative splice site in exon 10 (alt splice). If the alternative splice site is not used, the HMWK-specific remainder of exon 10 is retained, forming a HMWK transcript. The added 5′ and 3′ RNA oligomers for RNA ligase-mediated rapid amplification of cDNA ends PCR are indicated by dotted lines. Primers are described in methods and are indicated as connected arrowed lines. Lower case indicates a HMWK-specific primer set; upper case and underlined indicates a primer set common to both HMWK and LMWK; and upper case only is specific to LMWK. Primer sets C, D, E, and F are nested primer sets shown as 2 primers for simplicity. Figure is not to scale.

Fig. 2.

HMWK-specific PCR from mouse genomic DNA. We used primer set A to amplify mouse genomic DNA by PCR. Agarose gel electrophoresis of the amplification products yielded 2 bands. One is the expected 630-bp band, based on the published HMWK sequence, and the other is 576 bp.

Genomic Southern blotting for kininogen genes.

Takano et al. (10) suggested that mice likely have only a single kininogen gene. To demonstrate that our PCR fragments originate from different genomic templates, we conducted Southern blots of mouse genomic DNA. As probes, we used both K1- and K2-specific HMWK PCR fragments (from primer sets A and B). The restriction enzymes we used (EcoRI and BamHI) do not have sites within the area covered by the probes (and thus generate only 1 hybridizing band per gene, as the restriction sites must flank the probe on either side). We found two strongly hybridizing bands, consistent with two mouse kininogen family genes (Fig. 3).

Fig. 3.

Southern blots of mouse genomic DNA digested with either BamHI or EcoRI. The probes were either the 703-bp K1 fragment generated by PCR primer set B or the 576-bp K2 fragment generated by primer set A.

K2 also exists in LMWK form.

The sequence data we had for K1 and K2 all corresponded to regions of the genes that were either common to both HMWK and LMWK or specific to HMWK. To obtain LMWK2 sequence data, we used our existing K1 and K2 sequence data and searched GenBank for homologous sequences. We found a number of mouse DNA partial liver and kidney cDNA sequences generated by automated sequencing that were originally identified as being homologous to human kininogen (for example, GenBank accession nos. A1255556 and A1314053). Examination showed that these sequences could be divided into two groups. One was derived from K1 (both HMWK and LMWK forms); the other was from LMWK2, which we were able to verify by the overlapping regions in common with our HMWK2 sequence.

Cloning and sequencing of K2 cDNAs.

We used RNA ligase-mediated (RLM) and oligo-capping RACE PCR with a proofreading polymerase cocktail to obtain cDNA clones with the K2 5′ end from mouse kidney total RNA. The 3′ K2-specific primers were either from the LMWK2 region of exon 11 or the region common to both HMWK and LMWK2 in exon 10 (primer sets B and C). Both primer sets yielded amplification products, which were cloned and sequenced. In all cases, the 3′ sequence corresponded to our previously identified K2, which we were now able to extend to the 5′ beginning of the transcript. Sequencing the new 5′ K2 clones allowed them to be divided into two sets. In one set, the sequence of K2 remained highly homologous to the previously identified kininogen gene (K1) until the reported 5′ end of the K1 transcript. In the second set, although the 3′ sequence was highly homologous to K1, the 5′ K2 end strongly diverged from the first ∼300 bp of the start of K1 transcription. Interestingly, the two sets of K2-specific 5′ ends strongly diverged from each other just before translation initiation.

We obtained cDNA of the K2 3′ end with RT-PCR, in which the reverse transcription uses an oligo(dT) sequence with a unique extension to allow specific primer binding sites for PCR. The 5′ PCR primer is located in either HMWK2 exon 10 or LMWK2 exon 11 (primer sets D and E). Combining the sequence data allowed us to generate a consensus for the entire K2 message. Comparing the common region of K2 to the published K1 cDNAs, we find 93% and 92% sequence identity for HMWK and LMWK, respectively.

Examination of the sequence data (Fig. 4) established that the bradykinin motifs of K1 and K2 encode identical nonapeptides. The nucleotide sequence inside the bradykinin motif has one nucleotide substitution between K1 and K2, but it is a silent substitution, leaving the predicted amino acids unchanged. However, just outside the 5′ and 3′ ends of the bradykinin motif there are numerous base pair substitutions. There are also two gaps in exon 10 of HMWK2 compared with HMWK1, one measuring 18 bp and the other 33 bp. These and further sequence comparisons are commented on in discussion.

Fig. 4.

Comparison of the nucleotide sequence of HMWK1 and HMWK2, comparing overlapping sequence data we generated for HMWK1 and HMWK2 exon 10 and the region immediately 5′ of this exon. The top continuous sequence represents HMWK1, with HMWK2 indicated at bottom at points where the sequence differs from HMWK1. Deletions are indicated by dashes. Sequence areas of note are boxed, from top to bottom: the bradykinin motif, which contains only one silent nucleotide difference; a K1-specific insertion of unknown function, not found in kininogen genes of other species studied; a region in both K1 and K2 that is highly homologous to an insertion in the rat T-kininogen exon 10 HMWK remnant and is not found in kininogen genes of other species studied; and a histidine/glycine repeat found in both K1 and K2 and thought to be involved in the coagulation cascade. The start of exon 10 and the alternative splice donor site are indicated by arrows.

K2 has two transcriptional start sites.

Using the two different K2 5′ ends, the rest of our K2 sequence, and the K1 cDNA sequence, we searched the draft version of the mouse genome database. As shown by the database, both K1 and K2 are in the same area of the genome on chromosome 16 in a head-to-tail orientation. Examination of the K2 5′ region of the genome revealed both alternative 5′ ends. The K2 transcript that remains homologous to the 5′ end of K1 has strong intron/exon similarity to K1. Again, as shown by the mouse genome database the other K2 transcript is unusual in that it begins in the classic “intron” 2. Transcription of this mRNA is contiguous with and continues into exon 3 and is thereafter similar in intron/exon structure to the other kininogen transcripts (Fig. 5). Thus in the mouse kidney K2 has two different transcriptional start sites resulting in different 5′ untranslated regions of mRNA.

Fig. 5.

Schematic of the organization of the K2 alternative 5′ ends. The exons of the K2 transcript with 5′ homology to the K1 transcript are indicated by filled boxes, the transcriptional start site is indicated as Start 1, and the splice pattern is indicated by arrowed lines. The alternate K2 start of transcription is indicated as Start 2, with the open box being exon material unique to this transcript that then continues into exon 3. Introns indicated as horizontal lines are 1/3 scale.

Expression pattern of K1 and K2 mRNA.

Although RLM-RACE PCR indicated that K2 is transcribed in the kidney, we investigated the tissue specificity of K2 expression, examined whether the transcripts exist in both HMWK and LMWK2 forms, and compared the expression pattern to K1 by RT-PCR.

First we used our sequence data to design PCR primers specific to the region common to both HMWK and LMWK for the newly discovered K2 (primer set G, from exon 7 to the 5′ end of exon 10) and another similar set for K1 (primer set H). RT-PCR of mouse liver (the most abundant source of kininogen) amplified the predicted bands for both K1 (331 bp) and K2 (329 bp) (Fig. 6). Representatives of these bands were cloned, and DNA sequencing established that they were K1- and K2 specific, confirming that both genes are actively transcribed in the liver.

Fig. 6.

K1- and K2-specific RT-PCR amplification in the liver. On the basis of our sequence data, we designed primers specific to K1 and K2 in the region common to both LMWK and HMWK transcripts. RT-PCR was performed with total liver RNA, followed by agarose gel electrophoresis of the amplification products. We used K1-specific primer set H to amplify liver cDNA and obtained the expected 331-bp band (K1), whereas a buffer control (K1C) generated no bands. We used K2-specific primer set G to amplify liver cDNA and obtained the expected 329-bp band (K2), whereas a buffer control (K2C) generated no bands.

We further examined the expression of HMWK1, HMWK2, LMWK1, and LMWK2 mRNA in various mouse tissues, using RT-PCR and primers specific to each mRNA. The tissues examined were the heart, embryo, ovary, testis, kidney, brain, liver, spleen, lung, and thymus. To examine HMWK1 and HMWK2 expression, we employed primer set A, previously used to amplify both K1 and K2 HMWK-specific bands in exon 10 from genomic DNA. RT-PCR using this primer set amplified both HMWK1 (630 bp)- and HMWK2 (576 bp)-specific bands that were visualized by agarose gel electrophoresis. We also designed specific primers for LMWK1 (primer set I, predicted band size 248 bp) and LMWK2 (primer set J, predicted band size 237 bp). Specificity of PCR-amplified products was confirmed by DNA sequencing after cloning representatives of the amplified products (data not shown). Although both K1 and K2 were generally transcribed in the same tissues, there were tissue-specific differences between K1 and K2 with regard to use of the alternative splice donor that generates HMWK and LMWK (Fig. 7). In the liver, embryo, and ovary, K1 existed as both HMWK and LMWK, whereas K2 in the liver and embryo was only LMWK, and K2 in the ovary was only HMWK. Conversely, in the kidney, lung, and thymus, K2 existed as both HMWK and LMWK, whereas K1 in the kidney was not present as HMWK and only very weakly present as LMWK (the kininogen transcripts present in the kidney are predominantly from K2) and K1 in the lung and thymus was only HMWK. In the heart, testis, and brain, we found HMWK for both K1 and K2. Neither was expressed in the spleen.

Fig. 7.

Expression of HMWK1, HMWK2, LMWK1, and LMWK2 mRNA in mouse tissues. A: HMWK expression pattern of K1 and K2. We used primer set A to amplify mouse total RNA from the indicated tissues by RT-PCR and obtained 630-bp HMWK1 and 576-bp HMWK2 bands in several tissues. B: LMWK1 expression pattern. We used LMWK1-specific primer set I to amplify mouse total RNA by RT-PCR and obtained the 248-bp LMWK1-specific band in several tissues. C: LMWK2 expression pattern. We used LMWK2-specific primer set J to amplify mouse total RNA by RT-PCR and obtained a 237-bp LMWK2-specific band in several tissues. All reactions used β-actin as an amplification control and were visualized by agarose gel electrophoresis. The buffer control lane is labeled on left.


Previously, all mammals except rats were thought to have only a single kininogen gene, and this was considered true of mice as well (10). However, our preliminary report (3) as well as that of Cardoso et al. (1) have established the existence of two kininogen genes. We have further examined this by PCR and by obtaining various cDNA clones. DNA sequence analysis of the cDNA clones and PCR fragments clearly shows that there are two kininogen genes in mice. One is the previously identified kininogen gene (10), and the other is 93% identical in the HMWK form and 92% identical in the LMWK form; it has an open reading frame and an identical bradykinin motif. An unexpected observation is that, at least in the kidney, K2 transcripts have two different 5′ ends. This conclusion is based on our use of RLM-RACE PCR to isolate clones with two distinct 5′ ends for the kininogen transcripts. Identification of the K2 gene in the mouse genome via the mouse genome database allowed us to locate these 5′ ends. One is similar to the K1 transcript, whereas the other is completely divergent and has been identified as part of the classic intron 2, with the transcript continuing into exon 3 without a splice event. Furthermore, unlike the situation with the renin 2 gene, in which some mouse strains have the gene and others do not, the second kininogen gene has been found in all the Mus musculus strains we investigated, namely, C57BL/6J and Swiss Webster reported here and DBA/2J, 129P3/J, and BALB/cJ (data not shown).

Takano et al. (10) suggested that mice likely had only a single kininogen gene, based on the fact that they obtained HMWK and LMWK clones derived from only one gene from a mouse liver cDNA library, and they measured hybridizing bands on Southern blots with a kininogen cDNA probe. However, Southern blots could be misleading if bands at or near the same size are only being counted once. We have obtained Southern blots of mouse genomic DNA, using the K1- and K2-specific HMWK PCR fragment as a probe and hybridizing to genomic DNA digested with restriction enzymes not present in the PCR fragment. We found two strongly hybridizing bands, consistent with two mouse kininogen family genes. We do not know whether other species besides rats have more than one kininogen gene.

Although we have no direct evidence that the K2 (or K1) gene product is cleaved by either plasma or tissue kallikrein and releases bradykinin, this seems possible, as the base pair difference between K1 and K2 inside the bradykinin motif is silent, implying selective pressure for maintenance of a functional bradykinin motif. In this case, mice would have two sources of vasodilator kinins. The area just outside the bradykinin motif is poorly conserved between K1 and K2 (and between species), which may influence the ability of either tissue or plasma kallikrein to cleave either K1 or K2. At this point we have no information regarding the ability of either K1 or K2 to serve as a substrate for any specific kininogenase. However, there is an Arg-Arg motif at the NH2-terminal cleavage site in both K1 and K2 that is critical for bradykinin release by kallikrein. Unlike humans, mice do not release Lys-bradykinin (kallidin) after cleavage by tissue kallikrein.

We found two major differences in the HMWK-specific region of exon 10 between K1 and K2. The first involves the 3′ region of the exon, in the histidine-rich area thought to play a role in the intrinsic clotting cascade that provides recipient sites for prekallikrein or factor XI in human, bovine, and rat HMWK (5, 6, 8). In K1 we found the unique perfect 15-unit His-Gly repeat previously identified by Takano et al. (10) In K2 we identified a similar repeat; however, it was only 12 units and was imperfect (Fig. 2), with 4 amino acid substitutions. In K2 this region may still be functional, as the equivalent human, bovine, and rat regions also show a smaller number of imperfect repeats compared with K1. We do not know whether K2 retains another classic function of kininogen by acting as a cysteine protease inhibitor. In addition, there is a 33-bp insertion into the 5′ HMWK region of exon 10 in K1 (Fig. 4) not present in K2, rat K- and T-kininogen, or human or bovine kininogen. We have no information as to the functional significance of this insertion.

Comparison of the sequences of K1 and K2 to rat K-kininogen and T1- and T2-kininogen [first described by Okamoto and Greenbaum (7)] resulted in an interesting observation. Although rat T-kininogens do not form HMWK, they do contain a remnant of the HMWK region of kininogen exon 10 that exhibits a number of mutations, including generation of termination codons (4). The T-kininogens also have three characteristic insertions in the HMWK-specific region of exon 10 that are not present in rat, bovine, or human kininogen (4). The first of these insertions, comprising 30 bases, is also found in both K1 and K2, with K1 identical to T-kininogen at 27 bases and K2 identical to T-kininogen at 25 (Fig. 4). This suggests that the gene duplication that resulted in the first rat T-kininogen, as well as the 30-bp insertion, occurred before the evolutionary separation of mice and rats. However, as one of the authors previously determined that inducing inflammation in mice by the application of turpentine did not increase plasma kinins (Alhenc-Gelas F, personal observation), it appears unlikely that either K1 or K2 is an acute-phase protein as is the case of the rat T-kininogen genes.

We also investigated whether or not both kininogen family genes are expressed in the mouse, in their expected location, and whether they are expressed as LMWK, HMWK, or both. Using RT-PCR, we examined the RNA of a number of different tissues for expression of HMWK1, LMWK1, HMWK2, and LMWK2. We found that if a tissue expressed one kininogen family gene it also expressed the other. However, the splice forms differed from one tissue to another. For example, in the liver K1 mRNA was present as both HMWK and LMWK, whereas K2 was present only as LMWK [in contrast to Cardoso et al. (1), who did detect HMWK2 as well]. In contrast, in the kidney K2 mRNA was present as both HMWK and LMWK, whereas K1 was present only very weakly as LMWK. Thus whether or not the splice donor in exon 10 is used is not a random event but is regulated by some unknown mechanism. There are several ways that the use of the alternative splice site could vary between tissues. There could be a mechanism capable of distinguishing between K1 and K2 transcripts, with the splice site for each transcript dependent on the cell type. It is also possible that if we divided organs into specific segments the organ-specific splice patterns could prove to be a composite of two or more cell types with differing control of the splice site. Such a mechanism may have functional significance, as plasma kallikrein can only cleave HMWK, whereas tissue kallikrein can cleave both HMWK and LMWK.

We found some minor discrepancies between our RT-PCR data and those of Cardoso et al. and Takano et al. regarding tissue expression of the new and previously identified kininogen genes. As mentioned above, Cardoso et al. did detect HMWK2 in the liver; they also detected a number of alternative splice variants that we did not observe. In contrast, we observed an alternative transcription start site in K2 intron 2 that Cardoso et al. did not report. Although we and Takano et al. observed LMWK1 in the liver and kidney, we did not find it in the heart, testis, or brain, unlike Takano et al. (albeit very weakly, and then only after Southern blot of the PCR products using mouse LMWK cDNA as a probe). Likewise, whereas we both observed high levels of HMWK1 expression in the liver, Takano et al. did not observe consistent expression in the heart, testis, brain, and lung, but rather only very faintly (again by Southern blot) and only after extending the number of amplifications from 30 to 40. One reason for these contradictory findings may be the different RT-PCR primers and methodology, which may have affected sensitivity to low levels of gene expression. Another possibility is that kininogen expression may vary between the Swiss Webster mice we used and the ICR and FVB/N mice studied by Takano et al. (10) and Cardoso et al. (1), respectively.

The mouse, in part because of the ability to manipulate the genome by generation of transgenic and gene-knockout mice, is presently a major model organism for use in the study of human pathology. In this regard it is important to know the areas in which this model varies genetically from humans. There are known differences in some genes whose products are involved in blood pressure regulation, such as the angiotensin II type 1b receptor and renin 2 genes, which are present in mice (in some strains in the case of renin 2) but not in humans. The determination that there are three kininogen genes joins these known genetic differences. At this point it is unknown what impact, if any, the existence of two kininogen genes with differing expression patterns has on the suitability of using the mouse as a model of human blood pressure regulation or other pathways in which kininogen is involved. In addition, the presence of a second kininogen gene complicates any planned genetic manipulation of the mouse kininogen gene by gene targeting. It is not clear which of the genes to disrupt, whether the remaining kininogen gene could compensate, or whether both would need to be targeted. A better understanding of the specific roles of the two mouse kininogen genes would help in the interpretation of data from the mouse as a model for studies in which kininogen is involved.

In summary, we have identified and sequenced a second kininogen gene in the mouse, which we have designated K2. K2 is highly homologous to the previously identified kininogen gene (K1), with an open reading frame and an intact bradykinin motif. K2 is transcribed and its mRNA exists in both HMWK and LMWK forms. K2 transcripts can have two different 5′ ends, at least in the kidney. Interestingly, whether or not the splice donor site generates LMWK (if used) or HMWK (if not used) differs according to both cell type and whether the transcript is K1 or K2. This implies that use of the splice donor site is specifically regulated by some unknown mechanism.


This work was funded by National Heart, Lung, and Blood Institute Grant HL-028982-24 and National Kidney Foundation of Michigan Grant F40010.


  • Article published online before print. See web site for date of publication (http://physiolgenomics.physiology.org).

    Address for reprint requests and other correspondence: E. G. Shesely, Henry Ford Hospital, 7043 E&R Bldg., 2799 West Grand Blvd., Detroit, MI 48202 (e-mail: eshesel1{at}hfhs.org).


View Abstract