(Received for publication, October 28, 1996, and in revised form, January 10, 1997)
From the Research Institute for Food Science, Kyoto University, Uji, Kyoto 611, Japan
The cDNA encoding a novel isoform of human trypsinogen was identified. The isoelectric points of the proenzyme and active forms calculated from the deduced amino acid sequence are consistent with those of mesotrypsin(ogen), known to be an inhibitor-resistant trypsin isoform. The cDNA attached with a bacterial signal peptide sequence was expressed in Escherichia coli. The recombinant proenzyme purified from periplasm showed enterokinase-dependent activation similar to a major isoform of human trypsinogen. The enzyme was far less inhibited by trypsin inhibitors such as soybean trypsin inhibitor, aprotinin, or pancreatic secretory trypsin inhibitor than the control trypsin. A gel filtration assay showed that the enzyme and aprotinin did not form a stable complex. It is noteworthy that the amino acid at position 198, which is in close vicinity to the active Ser, is Arg while those of other major trypsins are all Gly. It is concluded that the cloned cDNA encodes human mesotrypsinogen, a unique isoform of trypsinogen with inhibitor resistance.
Pancreatic trypsin is a key enzyme, which leads to activation of a number of pancreatic digestive proenzymes including chymotrypsinogens, procarboxypeptidases, proelastases, and prophospholipase A2 as well as trypsinogens themselves. Three different trypsinogens, which show unique isoelectric points (pI), have been described in human pancreatic juice using two-dimensional isoelectric focusing/SDS-PAGE.1 These trypsinogens have been designated as trypsinogen 1 (pI = 4.9), 2 (pI = 5.7), and 3 (pI = 6.2); the isoforms having been numbered from anode to cathode in accordance with the IUPAC-IUB Commission on biochemical nomenclature of multiple forms of enzymes (1).
Trypsinogen 1 and 3 represent 23.1 and 16.0%, respectively, of total pancreatic secretory proteins in humans and play a major role in the proteolytic digestion (1). Trypsinogen 2, alternatively called mesotrypsinogen (based on its intermediate pI relative to the other two major trypsinogens) accounts for only a trace amount (<0.5%) and displays a specific characteristic best demonstrated when challenged by naturally occurring trypsin inhibitors such as soybean trypsin inhibitor or aprotinin; mesotrypsin is far less inhibited by these inhibitors than are other isoforms of trypsin (1, 2).
Researchers have found that levels of mesotrypsinogen are increased in chronic alcoholics and decreased in patients with pancreatitis (3). These findings raise the question of whether inhibitor-resistant activity of mesotrypsin has a role in the pathogenesis of pancreatic diseases (4). However, the limited availability of mesotrypsinogen has made determination of its biological properties and physiological significance difficult.
Two major trypsinogen cDNAs, TRYI and TRYII, which correspond to trypsinogen 3 (pI = 6.2) and 1 (pI = 4.9), respectively, based on estimated pI values, have been identified (5). Another human pancreatic trypsinogen cDNA, trypsinogen III, has also been described (6), which has been proposed to correspond to trypsinogen 2. However, at present, little is known of trypsinogen III, and the estimated pI based on the deduced amino acid sequence (pI = 5.94 for trypsinogen form; pI = 7.0 for active trypsin form) is different from those experimentally obtained in native mesotrypsinogen (pI = 5.7 for trypsinogen form; pI = 6.4 for active trypsin form) (1, 2).
Here we describe the cDNA encoding a novel isoform of human trypsinogen. The deduced amino acid sequence displays a pI that is in accordance with the biochemical data, and furthermore, characterization of the heterologously expressed product shows that the cDNA encodes the human trypsin isoform, which resists trypsin inhibitors, suggesting that it could be a human trypsinogen 2/mesotrypsin gene.
-DNA modification enzymes
and restriction enzymes were purchased from standard commercial sources
and used according to the manufacturers' instructions. PCR was
performed as described (7) except that Taq polymerase was
used with Pfu polymerase (1:1 unit ratio) to increase the
fidelity of the template amplification. The recombinant plasmids
constructed as described below were confirmed by a combination of
restriction enzyme mapping and DNA sequencing. The
oligonucleotides used are as follows: F1,
5-GATGACAAGCTTGTTGGGGGCTACA-3
; F2, 5
-AAACAGCTATGACCATG-3
; GT10F,
5
-CGGGATCCGCTGGGTAGTCCCCACCTTT-3
; GT10R,
5
-GCTCTAGACTTATGAGTATTTCTTCCAGGGTA-3
.
Expression of mesotrypsin in bacteria
was performed by ligating the cDNA-coding sequences for
mesotrypsinogen to the DNA sequence encoding a bacterial signal peptide
at the HindIII site in pFlag-1 (IBI). The coding region of
mesotrypsinogen with HindIII sites at both ends was
generated by using PCR with synthetic oligonucleotide primers, F1 and
F2, from a clone carrying the mesotrypsinogen cDNA. The
PCR-derived mesotrypsin cDNA used in the expression vector was
sequenced, and its identity with the mesotrypsin cDNA sequence
derived from
clones was confirmed. The vector is hereafter referred
to as pFlag-meso. In this vector, the native signal peptide of
premesotrypsinogen is replaced with the signal peptide of OmpA to
ensure efficient translocation of the expressed protein into the
periplasmic space of bacteria. The probability that the protein would
fold with correct disulfide formation was also expected to be increased
in the oxidizing environment of the periplasm. Use of the vector allows
for regulated control of mesotrypsinogen expression by adding IPTG. An
expression vector was also constructed in the same manner (with a
cDNA-encoding human trypsinogen 1 (anionic trypsinogen, pI = 4.9)) and utilized as a control. This vector is referred to as
pFlag-T1.
Escherichia coli strain JM109 was
transformed with pFlag-meso or pFlag-T1 and grown to log phase. IPTG
was added to a final concentration of 0.5 mM, and the
cultures were further incubated in the presence of 5 mM
reducing glutathione, which is known to accelerate expression and
correct folding of secretory proteins in E. coli (8).
IPTG-induced expression was analyzed with SDS-PAGE. A major induced
band (about 20% of total protein) was shown to have a molecular size
corresponding to trypsinogen (24 kDa). Cells were collected by
centrifugation at 6000 × g for 20 min, washed with a
Tris-HCl buffer, pH 7.5, containing 1 mM EDTA, and then treated with lysozyme (0.1 mg/ml) in the same buffer at 4 °C for 30 min. Periplasmic proteins were obtained as supernatant fractions after
centrifugation at 10,000 × g for 20 min. Expressed
trypsinogens were further purified with differential sedimentation with
ammonium sulfate (9), and the resulting material was further purified by high pressure liquid chromatography with a TSK3000 gel filtration column. Elution was monitored at 280 nm absorbance and then by SDS-PAGE. The fractions containing the 24-kDa band were pooled and
stored at 20 °C until use in the presence of 0.5 mM
HCl to prevent the autoactivation of zymogens.
The recombinant mesotrypsinogen or trypsinogen 1 (0.1 mg/ml) was activated with enterokinase (1 µg/ml) (New England Biolabs) at 37 °C for 40 min, and then benzoyl arginine p-nitroanilide (BApNA, Wako Pure Chemicals, 2 mM) was added as a substrate according to the method described (10). The mixtures were incubated at 37 °C for 15 min, and then the reactions were stopped by addition of acetic acid (20%). p-Nitroanilides liberated by tryptic activities were measured with a spectrophotometer (Beckman DU-650) at 410 nm. Under the experimental conditions, linearity between increments of p-nitroanilide and incubation time was maintained up to 40 min. Thus the reaction was routinely incubated for 15 min, and then trypsin activity was calculated as the rate of A410/min unless specified.
Degradation of Macromolecule by TrypsinsProteolytic activities of trypsins were examined using azocasein (Sigma) as substrate according to the method described by Hofsten and Renhammar (11). Hydrolysis of azocasein was monitored in increments at A366.
Trypsin Inhibitory AssayThe recombinant mesotrypsinogen or
trypsinogen 1 (0.1 mg/ml) was activated with enterokinase (1 µg/ml)
as described above, and then various proteinaceous or synthetic trypsin
inhibitors were added at given concentrations as shown in the figures.
The reaction mixtures were incubated at 22 °C for 15 min to allow inhibitor/enzyme interaction, after which BApNA was added. The reaction
was carried out at 37 °C for 15 min and stopped as described above.
The residual trypsin activity was measured and then expressed as
percentages of the activity in control reactions without inhibitors. Soybean trypsin inhibitor (SBTI) was from Sigma.
Bovine pancreatic trypsin inhibitor (BPTI), phenylmethylsulfonyl
fluoride,
N-p-tosyl-L-lysine
chloromethyl ketone (TLCK), and diisopropyl fluorophosphate were from
Wako Pure Chemicals.
N,N-dimethylcarbamoylmethyl-4-(4-guanidinobenzoyloxy)phenylacetate methane sulfate (FOY-305) was a gift from Ono Pharmaceutical Co., Ltd.
Human pancreatic secretory trypsin inhibitor (PSTI) was obtained from
Shionogi Pharmaceutical Company, Japan. Rat PSTI was purified according
to the method described (12).
Trypsin/inhibitor interactions were observed with a gel filtration. A Sephadex G-100 column was equilibrated with a buffer containing 0.1 M NaCl, 50 mM Tris-HCl, pH 7.5, and 1 mM CaCl2, and then calibrated with trypsin (24 kDa) and BPTI (aprotinin) (6 kDa), separately. Activated recombinant mesotrypsin or trypsin 1 (0.1 mg/ml) was incubated with BPTI (250 µg/ml) (about equal molecular ratio) at 22 °C for 30 min, and then the mixture was loaded onto the column. Elution profiles were monitored with a spectrophotometer (Beckman DU-650) at 280 nm.
-A human pancreatic cDNA library
constructed in -gt10 was screened with a rat trypsinogen cDNA
fragment as a probe. About 200 positive clones were obtained. Each
DNA with a cDNA insert was extracted from a portion of phage stock
solution and utilized as a template for subsequent PCR. A set of
primers (GT10F and GT10R) was designed to hybridize to the flanking
regions of the cDNA insert. Amplified PCR products that were
sufficient to encode full-length trypsinogen (more than 750 base pairs)
were selected and then systematically analyzed with a series of
restriction enzymes to categorize them. About 80% of clones were
categorized into predicted patterns based on published TRYI
(trypsinogen 3, cationic) or TRYII (trypsinogen 1, anionic) DNA
sequences. Of the remaining 20% a category comprising 10 clones shared
the same pattern, but this pattern was different from those of the
known trypsin cDNAs. We selected these clones for further analysis. Five individual cDNA inserts in
DNAs from this category were subcloned into pBluescript and then the DNA sequences were analyzed. All five clones showed identical DNA sequences as shown in Fig. 1.
Structure of Mesotrypsin cDNA
-The open reading frame of
the cDNA encodes a protein of 247 amino acids (from 741 nucleotides), exhibiting a high, although not identical, similarity to
the previously reported human trypsinogens as shown in Fig.
2. The deduced amino acid sequence contains the major
features characteristic of trypsinogens; a negatively charged activation peptide with a typical enterokinase cleavage site (DDDDKI), three conserved catalytic triad residues (His, Asp, Ser), and 10 Cys
residues in conserved positions. The calculated isoelectric point based
on the deduced amino acid sequence is pI = 5.68 for the
trypsinogen form and pI = 6.55 for the active trypsin form, which
is remarkably close to experimentally obtained human mesotrypsinogen (pI = 5.7 for trypsinogen form and pI = 6.4 for active
trypsin form), as summarized in Table I. Therefore we
designate this clone as the human mesotrypsinogen cDNA. Further
confirmation of its identity using recombinant protein appears below.
The amino acid sequence of the cDNA has 86, 88, and 97% homology
with those of TRYI (trypsinogen 3), TRYII (trypsinogen 1), and
trypsinogen III, respectively.
|
-E. coli transformed with pFlag-meso
carrying mesotrypsinogen cDNA produced mesotrypsinogen protein at
about 20% of total protein upon IPTG induction (Fig.
3). A large portion (>90%) of the expressed protein
remained insoluble in the bacteria even after lysis with 1% Triton
X-100. A small portion (about 10%) of the expressed mesotrypsinogen
was obtained as soluble protein in the periplasm space after lysozyme
treatment. The apparent molecular size of the soluble mesotrypsinogen
separated in SDS-PAGE was smaller by about 2000 Da (24 kDa total) than
insoluble protein (26 kDa total), indicating that the soluble form of
mesotrypsinogen was translocated to the periplasm and a signal peptide
of OmpA (its predicted size is 2046 Da; 21 amino acids) was released
resulting in the proenzyme (mesotrypsinogen) form. The apparent low
efficiency of translocation may be due to saturation of E. coli translation machinery with high levels of exogenous protein
expression.
Further purification was needed to obtain stable mesotrypsinogen, which can be used for enzymatic studies. Activation of trypsinogen to trypsin was observed in some preparations without addition of enterokinase. An affinity chromatography procedure using anti-Flag monoclonal antibody M1 (IBI) was not successful in purifying mesotrypsinogen. Antibody M1 is designed to bind Flag, an N-terminal activation peptide of trypsinogen (DDDDK), when it is expressed as a tag with a foreign protein. However, our results suggest that antibody M1 is not able to access and thus bind the Flag peptide when it is expressed as a portion of the authentic structure of trypsinogen. Chemical bonds that are supposed to be required for Flag peptide/antibody interaction would be already occupied by intramolecular interactions within trypsinogen. A series of differential sedimentations with ammonium sulfate and a gel filtration were employed to obtain purified mesotrypsinogen. The recombinant mesotrypsinogen showed weak cross-reactivity with anti-anionic and cationic trypsinogen antibodies (provided from Dr. Scheele) indicating that mesotrypsinogen shares structural homology to other major trypsinogens (data not shown).
Characterization of Recombinant Mesotrypsinogen-Recombinant
mesotrypsinogen showed enterokinase-dependent activation in
a time-related manner, which was very similar to that shown by
recombinant trypsinogen 1 (anionic human trypsinogen) (Fig.
4). In the absence of enterokinase neither
mesotrypsinogen nor trypsinogen 1 were activated (Fig. 4). A
control preparation similarly purified from E. coli
periplasm transformed with vector alone did not show any significant
trypsin activity in the presence or absence of enterokinase (data not
shown) indicating efficient separation of endogenous E. coli
proteases by the purification. Activated mesotrypsin was completely
inactivated by 5 mM diisopropyl fluorophosphate, which is
sufficient to inactivate ordinary serine proteases, confirming that
mesotrypsin is a serine protease. The proteolytic activity of
mesotrypsin to azocasein was the same for those of human trypsin 1 and
bovine trypsin (Fig. 5).
When recombinant mesotrypsin was mixed with various concentrations of
protein inhibitors, it showed significant resistance to those
inhibitors. When equal molar amounts of SBTI (Mr = 24,000, 0.1 mg/ml) was added to the recombinant trypsin 1 (prepared
as a control using the same procedure) more than 95% of the activity was inhibited while mesotrypsin was hardly affected (Fig.
6A). Although increasing the amount of
inhibitor had a dose-dependent effect on inhibition, about
30% of trypsin activity remained even with a 1000-fold molar excess of
SBTI. A marked difference was obtained when mesotrypsin or trypsin 1 was mixed with human PSTI (Mr = 6,000), which is
a physiological inhibitor coexisting with trypsinogen in the secretory
granule in the pancreas and may serve to prevent premature activation
of trypsins (12). With a 4-fold molar excess of human PSTI (0.1 mg/ml)
trypsin 1 was 100% inhibited while almost full activity of mesotrypsin
remained (Fig. 6B). In the presence of a 40-fold excess of
human PSTI (1 mg/ml), about 80% of the activity of mesotrypsin
survived. Similar results were obtained using purified rat PSTI (data
not shown). These profiles were also observed in the presence of BPTI
(aprotinin), another type of protein inhibitor purified from bovine
pancreatic tissue (Fig. 6C). On the other hand, mesotrypsin
was sensitive to low molecular weight, synthesized serine protease
inhibitors including phenylmethylsulfonyl fluoride (Fig.
6D), TLCK (Fig. 6E), and FOY-305 (Fig.
6F). The sensitivity of mesotrypsin to these inhibitors was
similar to that shown by human trypsin 1. For each inhibitor, mesotrypsin was almost 100% inhibited at a concentration great enough
to completely inhibit trypsin 1. Such characteristics are consistent
with the results found in native mesotrypsin (2). In the case of
FOY-305, however, mesotrypsin showed slightly higher resistance when
intermediate concentrations of the inhibitor were used (Fig.
6F).
Interaction of Mesotrypsin and a Trypsin Inhibitor
-Interaction of mesotrypsin with a trypsin inhibitor
was investigated by gel filtration chromatography (Fig.
7). When control human trypsin 1 (24 kDa) was
pre-incubated with an equal molar amount of BPTI (6 kDa) and the
mixture loaded onto a Sephadex G-100 column, a peak of apparent
molecular size of 30 kDa, corresponding to a trypsin/inhibitor complex,
was observed (Fig. 7, middle). On the other hand, when
mesotrypsin and BPTI were mixed and loaded, two peaks with an apparent
molecular size of 24 and 6 kDa were observed, neither of which
corresponded to a complex (Fig. 7, lower) indicating that
mesotrypsin did not bind to the inhibitor stably.
It is known that human trypsin, when compared with bovine trypsin, is less inhibited by naturally occurring trypsin inhibitors such as chicken ovomucoid (13). The occurrence of two forms (anionic and cationic) of trypsinogen has been reported by a number of researchers (3, 14-17). The cationic form of human trypsin, which represents the major part of trypsin activity of the whole pancreatic juice, is shown to be less inhibited by ovomucoid and soybean trypsin inhibitor than the anionic trypsin (18, 19). Both human trypsins are completely inhibited by aprotinin or human PSTI, a physiological inhibitor co-localized with trypsins in the zymogen granules, at a stoichiometric enzyme-to-inhibitor ratio of one to one (18). More recently, a third minor trypsinogen in human pancreatic juice, named trypsinogen 2 (1) or mesotrypsinogen (2), has been reported (1, 2) that shows almost no inhibition by either naturally occurring trypsin inhibitors or human PSTI. (Thus, anionic and cationic trypsinogens are referred to as trypsinogen 1 and 3 according to their pI values as summarized in Table I.)
In this study, after extensive screening of a human pancreatic cDNA library, we identified and characterized a cDNA-encoding mesotrypsinogen. The identification was based on the following points: 1) deduced amino acid sequence shows a unique isoform containing the major features characteristic of trypsinogens, 2) the calculated isoelectric point is consistent with that obtained from native mesotrypsinogen, and 3) recombinant protein expressed from the cDNA shows resistance against protein inhibitors, which has been observed in native mesotrypsin.
It has been predicted, based on Southern analysis, that the human genome carries at least ten trypsinogen genes (5). Recently, a large scale DNA sequencing of the human beta T receptor locus has revealed that eight trypsinogen genes (denoted as T1 through T8) are intercalated in the locus (20); among them, three (T4, T6 and T8) are apparently functional while the rest are pseudo or relic trypsinogen genes. Data base analysis reveals that T4 corresponds to TRYI and T8 to TRYIII (see Table I), two major trypsinogens. However, T6 does not correspond to any of the reported trypsinogen cDNAs. We examined whether mesotrypsinogen cDNA corresponds to T6 or any other trypsinogens found in this locus but found no match.
Analysis of the mesotrypsinogen cDNA in the data bases reveals that there are several trypsinogen-related cDNAs from the pancreas and brain, which show close similarities to the mesotrypsinogen sequence described here. Relatively high similarity between the mesotrypsinogen and trypsinogen III (6) genes might suggest these are allelic divergences in the human genome. However, we thought that this possibility is unlikely because differences in sequences (Thr167-Gln168, Tyr175, Cys196-Gln197 versus Arg167-Glu168, Cys175, Trp196-Lys197) include major substitutions of charged amino acids in the opposite direction and replacement of conserved Cys, which are quite unusual for regular allelic divergences, and partly because our intensive search for the trypsinogen III sequence in the screened clones failed. This suggests that trypsinogen III is a very minor component, or else it is highly up-regulated upon stimulation, similar to the expression of rat trypsinogen isoform P23 (21, 22). Products of trypsinogen III cDNA and the T6 gene remain to be further investigated.
Another mesotrypsin-like sequence found in the data bases is D23456, which has been identified in the human brain using a PCR technique (23, 24). The D23456 and our mesotrypsinogen sequences have about 200 amino acids in common at their C-terminals but the amino acid sequences at the N-terminals are quite different; the brain sequence lacks any signal sequence characteristic of pancreatic trypsinogens. It was suggested that tissue specific alternative splicing might generate the brain-type (cystosolic) and the pancreas-type (secretory) trypsinogens (23). Such a hypothesis is quite attractive; however, how a trypsinogen lacking a signal peptide forms correct disulfide bonds to generate an enzymatically active structure in the cystosolic environment that is reductive should be further investigated. Moreover, the alternative (secretory) form expressed in the pancreas was not detected with PCR (23). We did not find a cystosolic form of trypsinogen even after intensive screening of the pancreatic cDNA library. A reason for the failure of previous studies could be partly explained by the fact that they designed the PCR primer to detect a secretory partner of the brain-type trypsinogen in the pancreas based on trypsinogen III (6), the nucleotide sequence of which differs by more than 3% from the sequence of mesotrypsinogen cDNA described here. Mesotrypsinogen in the pancreas could be a better candidate for an alternative spliced form of a brain-type trypsinogen gene if the splicing model is correct.
Although elucidation of the molecular basis of the inhibitor resistance of mesotrypsin will require site-directed mutagenesis and x-ray crystallography, both of which are currently underway in our laboratory, it is noteworthy to point out that amino acid substitutions are unique in mesotrypsinogen when compared with other major isoforms of trypsinogen that are inhibitor sensitive, in particular at and around the reactive center that interacts with inhibitors. Between mesotrypsin and the other two major human trypsinogens, the catalytic triad of His63, Asp107, and Ser200 as well as an obligatory Asp at position 194 are all conserved. However, at amino acid 198 of mesotrypsinogen an Arg residue takes the place of a Gly residue. The Gly residue is normally found in the active center of the known rat and human sequences and participates in the interactions with trypsin inhibitors such as BPTI (aprotinin) (25). Substitution by a positively charged amino acid at Gly198 position might interfere with the trypsin/trypsin inhibitor interaction because a site in the trypsin inhibitor that interacts with trypsin also contains Arg. In a large number of Kunitz- or Kazal-type trypsin inhibitors the consesus sequence has been found to be Arg/Lys-Ile (26). By utilizing x-ray crystallography data of rat trypsin and BPTI complex (24) and superimposing the replacement of Gly by Arg at position 198 using the software Quanta (Molecular Simulations), the Van der Waals radius of both Arg198 in mesotrypsin and Arg17 in BPTI was found to overlay at any possible rotation. This suggests that two Arg residues are mutually exclusive in the complex. In fact two molecules could not bind tightly to each other. The model partly explains why mesotrypsin and BPTI did not form a stable complex in the gel filtration assay shown in Fig. 7.
The fact that mesotrypsin has a strong inhibitor resistance suggests several physiologically relevant functions. First, mesotrypsin may serve as the inhibitor-resistant trypsin for effective digestion when certain diets rich in naturally occurring trypsin inhibitors such as SBTI are ingested. Second, if levels of mesotrypsinogen are increased with a certain stimulus or by pathogenic conditions the inhibitor-resistant trypsin may cause potentially deleterious effects on autoactivation within pancreatic tissues in particular in the case of pancreatitis; autoactivated mesotrypsin can further activate trypsinogen even in the presence of PSTI or other endogenous trypsin inhibitors. The work presented in this paper provides essential information for future studies designed to reveal the physiological significance of mesotrypsin and its underlying molecular basis.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) D45417[GenBank].