Aromatic interactions, usually described as π-π interactions, are ubiquitous in nature and are involved in many biological processes like in the antigen-binding of immunoglobulins (Padlan, 1990), stability of duplex DNA (Kool, 2001) or stabilizing protein tertiary structures (Mitchell et al., 1994, Singh and Thornton, 1990). The delocalized electrons of the benzene ring are the basis for the these interactions (Kryger et al., 1998, Kryger et al., 1999, Obst et al., 1997).
Contribution of aromatic interactions in binding affinity and ligand selectivity in the S3/
S4 pocket of bovine trypsin, human factor Xa and chimeric S3/S4 mutants have been studied (Di Fenza et al., 2007). The aromatic character of this pocket increases from trypsin (only Trp215) to factor Xa (Trp215, Tyr99, Phe174).
The results show that the establishment of favourable directional aromatic-aromatic interactions in the S3/S4 pocket with a bound ligand will increasingly contribute to binding affinity and will thus determine selectivity (Di Fenza et al., 2007). Factor Xa is thus more selective with respect to bovine trypsin for ligands which opportunely interact with the fully established aromatic box in the S3/S4 subsite.
The effect of tyrosine sulfation on trypsin
Modification of aromatic tyrosine residue by sulfation provides it with highly polarizable electrons and makes it even more electronegative. Thus, proteins and peptides become more interactive by this PTM (Lyon et al., 2000, Sasaki et al., 1999, Woods et al., 2007). Tyrosine sulfate has been shown to be involved in protein-protein interactions (Costagliola et al., 2002, Stone and Hofsteenge, 1986, Wilkins et al., 1995, Woods et al., 2007) and proteolytic activity (Michnick et al., 1994).
In trypsinogen and trypsin, Tyr154 is located in the S’2 subsite within the primary substrate binding pocket (Gaboriaud et al., 1996, Katona et al., 2002). Thus, it is likely that sulfation of Tyr154 in trypsin contributes to more efficient substrate binding. Indeed, autoactivation of sulfated trypsinogen-1 was shown to be faster than that of the nonsulfated recombinant form (Sahin-Tóth et al., 2006). (Sulfated) pancreatic trypsin-1 and -2 were shown to be more effective activators of pro-uPA than (non-sulfated) TAT-1 and -2, respectively (Koivunen et al., 1989). Furthermore, modified (sulfated) trypsin-1 was shown to be more efficiently inhibited by PSTI than the non-modified form (Sahin-Tóth et al., 2006, Szilagyi et al., 2001).
On contrary to the findings of Sahin-Tóth et al. (Sahin-Tóth et al., 2006) and Szilagyi et al. (Szilagyi et al., 2001), (non-sulfated) TATs were somewhat more efficiently inhibited by TATI and soybean trypsin inhibitor than the pancreatic trypsins, whereas pancreatic trypsin-1 was more efficiently inhibited by limabean trypsin inhibitor than TAT-1 in a study of Koivunen et al. (Koivunen et al., 1989). As described above, there is evidence supporting more efficient substrate binding for the sulfated trypsin forms as compared to the non-sulfated ones. TATI has been found to be heterogenous (Huhtala et al., 1982, Kikuchi et al., 1985) so the PSTI/TATI preparations used in these studies may not necessarily be
comparable. It is also shown that substrate binding is not only determined by the primary substrate binding site, but several distal interactions are also involved (Hedstrom et al., 1992, Hedstrom et al., 1994b). These distal binding interactions and the possible differences in the PSTI/TATI preparations used could explain the discrepancy between these results.
The enzymatic parameters of native (sulfated) pancreatic and non-modified trypsins from pancreatic juice, ovarian cyst fluid and recombinant trypsin expressed in Escherichia coli have been determined using p-nitroanilide peptide substrates. The kinetic constants of pancreatic and tumor-associated trypsins were similar for one substrate (S-2222), but for two substrates (S-2444 and S-2251) the kcat for pancreatic trypsin-2 was lower and the kcat/ Km higher than that for TAT-2 (Koivunen et al., 1989) indicating for more efficient substrate
binding by (sulfated) pancreatic trypsin-2.
In other studies using different p-nitroanilide peptide substrates, the catalytic activity of native (sulfated) pancreatic trypsin-1, native modified pancreatic trypsin-1 and non-modified recombinant trypsin-1, respectively, was found to be practically identical (Sahin-Tóth et al., 2006, Szilagyi et al., 2001). The nonexistent or modest differences reported in the enzymatic parameters between (sulfated) pancreatic trypsins and (non-sulfated) tumor-associated or recombinant trypsins are likely to result from the structure of the p-nitroanilide peptide substrates used. The Tyr154 residue in the S’2 subsite interacts with the leaving group side of the scissile bond, not the acyl group side. In the chromogenic substrates used there is no P’2 residue, only the acyl group side (P1 to P4) with an arginine or lysine as the P1 residue. Thus, the influence of S’2 site on substrate binding when using p-nitroanilide peptide substrates is unlikely.
Conclusions
The most important result of this study was the development of specific MAbs and TR-IFMAs to trypsinogen-1 and -2. With these we could show that:
1) in acute pancreatitis serum trypsinogen-2 is elevated 50-fold, whereas serum trypsinogen-1 is elevated 15-fold, suggesting that trypsinogen-2 could be a diagnostic marker for acute pancreatitis.
2) TAT-2 is the predominant form in ovarian cyst fluids and its concentrations correlate with malignancy of these tumors. Thus, TAT is likely to be involved in ovarian tumor dissemination and breakage of tissue barriers.
3) serum samples from pancreatectomized patients contain immunoreactive trypsinogen isoenzymes. These results indicate that trypsinogen is not exclusively expressed in the pancreas and certain tumors, but that it may also be produced by normal extrapancreatic tissues.
4) two forms of trypsinogen-1 and -2, respectively, can be found in human sera and ovarian cyst fluids.
Finally, we confirmed by ESI-MS analysis that pancreatic trypsinogen-1 and -2 are sulfated and not phosphorylated at Tyr154, whereas tumor-associated trypsinogen-2 is not. We suggest that this modification may explain the previously observed differences between pancreatic and tumor-associated trypsin.