doi: 10.1053/ejpn.2000.0438 available online at http://www.idealibrary.com on I1[~|" European Journal of Paediatric Neurology 2001 ; 5(Suppl. A): 69-72

O R I G I N A L A R T I C L E

The substrate range of tripeptidyl-peptidase I

FRANCESCA BERNARDINI, MICHAEL J WARBURTON Department of Histopathology, St George's Hospital Medical School, London

Tripeptidyl-peptidase I (TPP-I) is an exopeptidase which removes tripeptides from the N-terminus of peptides. Mutations in TPP-I are responsible for late infantile neuronal ceroid lipofuscinosis (CLN2). The nature of the physiological substrates and the range and specificity of the enzyme are unclear. Previous experiments suggest that the enzyme can degrade small peptides but not proteins. Digestion of a range of peptides of different size by TTP-I suggests that the enzyme will degrade small peptides with an extended N-terminal domain but not structured peptides. In general, this cut-off occurs between masses of 4.5 kDa and 6 kDa. Reference to the structures of other peptidases suggests a mechanism for this size selectivity.

Keywords: Tripeptidyl-peptidase I. Late infantile neuronal ceroid lipofuscinosis. Peptide degradation.

Introduction

There is now abundant evidence that the lysosomal enzyme mutated in classical late infantile neuro- nal ceroid lipofuscinosis (CLN2) is tripeptidyl- peptidase I (TPP-I). 1,2 Human fibroblasts and brain tissue from CLN2 patients are devoid of TPP-I activity 2 and CLN2 fibroblasts are defective in the metabolism of small peptides. 1 Tripeptidyl- peptidase I is an exopeptidase that sequentially removes tripeptides from the N-terminus of pep- tides provided that the N-terminus is not blocked. 3 A small degree of homology with a family of bacterial enzymes suggests that the enzyme might be classified as a pepstatin-insensitive carboxyl proteinase. 4 Tripeptidyl-peptidase I was originally believed to be involved in the degradation of collagen because it is able to release Gly-X-Y triplets from synthetic collagen-like peptides and because inhibitors of TPP-I decrease the rate of bone resorption, where type I collagen is the major protein being degraded, in an in vitro assay, s,6 However, such a role for TPP-I remains uncertain.

Tripeptidyl-peptidase I appears to be able to degrade small peptides at least up to the size of glucagon (Mr 3500) 3 but is unable to degrade proteins. The size range of peptides which are substrates for TPP-I and the mechanism for excluding proteins from the active site are unclear. We now report data on the degradation of a range of peptides by TPP-I and propose a mechanism for its substrate size selectivity.

Materials and methods

The TPP-I used in these experiments was purified from pig kidneys by the method of Vines and Warburton. 3 The specific activity of the purified enzyme was 18.2 units/mg. Peptide digestions were carried out in a volume of 200/fl containing 30#mol of peptides, 0.5#g of TPP-I in 25mM sodium acetate pH4.0 for 5 h at 37°C. The reactions were terminated by adding trifluoroacetic acid (TFA) to a final concentration of 1%. Digests were chromatographed on C18 reversed phase high performance liquid chromatography (HPLC)

Correspondence: Dr Michael Warburton, Department of Histopathology, St George's Hospital Medical School, Cranmer Terrace, London SW17 ORE, UK e-mail: mwarburt@sghms.ac.uk

1090-3798/01/05/A069+4 $35.00 © 2001 European Paediatric Neurology Society

70 Original article: F Bernardini, M J Warburton

Table 1 Degradation of polypeptides by TPP-I

Rate of Peptide Size (Da) degradation (%)

Discussion

Insulin-A 2500 87 Insulin-B 3500 91 Insulin 6000 0 Glucagon 3500 100 PTH-(1-34) 4000 81 Neuropeptide Y 4500 60 IGF-1 7500 0 Ubiquitin 86 000 0 Cytochrome c 12 400 0 Lysozyme 14 300 0

Peptides (30 #mol) were incubated with TPP-I (0.5 #g) for 5 h at 35°C in the presence of 25raM sodium acetate pH 4.0. The digest was chromatographed on a C18 reversed phase HPLC column. Rates of degradation were estimated by measuring the decrease in the peak area of the peptide. Rates were compared with the degradation of glucagon (100%).

column (4.6 m m x 250 mm, Dynamax C18-300) using a 0-40% water-acetonitrile/0.1% TFA gradi- ent over 60rain. The rate of degradation of the peptide was estimated by measuring the decrease in its peak area after the digestion. Peptide structures are taken from The Protein Data Bank (ht tp: / /www.rcsb.org/pdb/) . Peptides were pur- chased from Sigma (Poole, UK) or Bachem (St Helens, UK).

Results

Some of the peptides studied were substrates for TPP-I (insulin A and insulin B chains, glucagon, parathyroid hormone (PTH) (1-34) and neuro- peptide Y), whereas other peptides were not degraded (insulin A/B dimer, IGF-I, ubiquitin, cytochrome c and lysozyme) (Table 1). We have previously demonstrated that TPP-I is able to completely degrade glucagon to the level of tripeptides. 3 About 40% of glucagon was degraded under the experimental conditions used. Insulin A and B chains and PTH (1-34) were degraded at similar rates and neuropeptide Y at about half the rate of glucagon degradation. There appears to be a cut-off between Mrs of 4250 or 6000 which marks Whether or not a peptide is a substrate for TPP-I. All peptides above the size of insulin were resistant to degradation by the enzyme. Mass spectrometric analysis of insulin and IGF-I before and after digestion indicated that there had been no decrease in the mass of these peptides.

The accumulation of proteinaceous material within the lysosomes of most tissues in CLN2 patients and cells grown in vitro has, for some time, suggested that the defective enzyme was a protease. The cloning of the mutated gene and its identification as being related to a family of bacterial pepstatin- insensitive carboxyl proteinases supported this view. 4 The CLN2 protease was subsequently demonstrated to be identical to TPP-I, a lysosomal exopeptidase. 1 Trepeptidyl-peptidase I was origin- ally believed to be involved in the latter stages of collagen degradation, s,6 However, such a role has not yet been substantiated. Earlier studies demon- strated that TPP-I degraded small peptides up to the size of glucagon but was inactive on small proteins such as cytochrome c. 3 In this study, we have shown that the cut-off point, between degrad- ation and resistance to TPP-I, lies between masses of 4500 Da and 6000 Da. Neuropeptide Y (4500 Da) is readily degraded whereas insulin (6000Da) is completely resistant to TPP-I. Tripeptidyl- peptidase I has been reported to cleave synthetic peptides which are substrates for bacterial endo- peptidases 7 although any action on proteins remains to be demonstrated. One of the major proteins which accumulates in the lysosomes of CLN2 cells is subunit c of mitochondrial ATP synthase. 8 Subunit c is a very hydrophobic protein with a mass of 7500 Da although it forms aggre- gates with a mass > 1 million Da. Co-incubation of extracts from normal and CLN2 fibroblasts results in the degradation of subunit c which suggests it is a substrate for TPP-I. 9

The ability of TPP-I to degrade the individual A and B chains of insulin but not the disulphide bonded AB dimer suggests a possible mechanism for the size selectivity of TPP-I. Inspection of the three-dimensional structures of the peptide back- bones reveals that all the peptides that are substrates for TPP-I have extended N-terminal domains (Fig. 1). Indeed, they have little second- ary or tertiary structure and the entire molecule can adopt an extended conformation. Larger peptides, such as IGF-I have considerable struc- ture and adopt a globular conformation and the N-terminus does not extend away from the molecule. The individual A and B chains of insulin fulfil the criteria as substrates for TPP-I in that they have extended structures. However, when the two chains are disulphide bonded together, the overall conformation is globular. A hypothetical model for the active site of TPP-I i s

presented in Fig. 2.

Original article: Substrate range of tripeptidyl-peptidase I 71

A

G

Tk

D

insulin chains can penetrate into the tunnel whereas the entry of pept ides wi th globular structures, e.g. insulin and IGF-I, is restricted. Clearly, any globular protein in which the N- terminus extends away f rom the bulk of the protein might also be a substrate for TPP-I.

An example of an enzyme displaying such a mechanism is prolyl oligopeptidase. TM The active site is located in a large interior cavity which the substrate enters through a nar row tunnel. The central cavity could accommodate a pept ide of about 30 amino acids in length, however the rate of hydrolysis decreases with the size of the peptide. As proposed for TPP-I, any pept ide wi th tertiary s tructure would be unable to traverse the nar row tunnel and gain access to the active site.

Fig. 1. Three-dimensional structures of the backbones of peptides which are, or are not, substrates for TPP-I. (A) glucagon (Brookhaven code 1BH0), (B) insulin, A and B chains (1A7F), (C) PTH (1BQT), (D) neuropeptide Y (1RON). The asterisk * marks the N-terminus. Glucagon, individual A and B chains of insulin, PTH and neuropeptide Y have extended N-terminal regions and are substrates for TPP-I whereas insulin A/B dimer has a globular structure and is not substrate for TPP-I.

Acknowledgement This work was suppor ted by a grant f rom Action Research.

Fig. 2. Schematic diagram of a hypothetical model of the active site of TPP-I. Amino acids in the active site may be located in a narrow tunnel which extends from the surface to the interior of the enzyme. The dimensions of the tunnel permit the entrance of extended amino-terminal domains but exclude structured peptides.

The active site of TPP-I may be located in a tunnel which extends from the surface of the enzyme into its interior. Alternatively, the active site ma y lie in an interior cavity which can only be accessed via a na r row tunnel. The active site residues, here shown as two aspartates, are located some distance into the tunnel. Only extended N- terminal domains can enter the tunnel to a sufficient dep th to encounter the active site. Thus, the N-termini of glucagon and the individual

References =

1 Vines DJ, Warburton MJ. Classical late infantile neuronal ceroid lipofuscinosis fibroblasts are defi- cient in lysosomal tripeptidyl peptidase-I. FEBS Lett 1999 443: 131-135.

2 Warburton MJ, Bernardini F. Tripeptidyl peptidase-I deficiency in classical late infantile neuronal ceroid lipofuscinosis brain tissue. Evidence for defective peptidase rather than proteinase activity. J Inherit Metab Dis 2000; 23: 145-154.

3 Vines DJ, Warburton MJ. Purification and character- isation of a tripeptidyl aminopeptidase-I from rat spleen. Biochim Biophys Acta 1998; 1384: 233--242.

4 Sleat DE, Donnelly RJ, Lackland H et al. Association of mutations in a lysosomal protein with classical late infantile neuronal ceroid lipofuscinosis. Science 1997; 277: 1802-1805.

5 McDonald JK, Hoisington AR, Eisenhauer DA. Partial purification and characterisation of an ovarian tri- peptidyl peptidase: a lysosomal exopeptidase that sequentially releases collagen-related (Gly-Pro-X) triplets. Biochim Biophys Res Commun 1985; 126: 63--71.

6 Page AE, Fuller K, Chambers TJ, Warburton MJ. Purification and characterization of a tripeptidyl peptidase from human osteoclastomas: evidence for its role in bone resorption. Arch Biochem Biophys 1993; 306: 354--359.

7 Ezaki J, Takeda-Ezaki M, Oda K, Kominarni E. Characterization of endopeptidase activity of tri- peptidyl peptidase-I/CLN2 protein which is

72 Original article: F Bernardini, M J Warburton

deficient in classical late infantile neuronal ceroid lipofuscinosis. Biochem Biophys Res Commun 2000; 268: 904-908. Kominami K, Ezaki J, Muno D et al. Specific storage of subunit c of mitochondrial ATP synthase in lysosomes of neuronal ceroid lipofuscinosis (Batten's disease). J Biochem 1992; 111: 278-282.

9 Ezaki J, Wolfe LS, Kominami E. Decreased lysosomal subunit c-degrading activity in fibroblasts from patients with late infantile neuronal ceroid lipofus- cinosis. Neuropediatrics 1997; 28" 53-55

10 Fulop V, Bocskei Z, Polgar L. Prolyl oligopeptidase: an unusual ]/-propeller domain regulates pro- teolysis. Cell 1998; 94: 161-170.