Carboxypeptidases in disease: insights from peptidomic studies

Abstract

Carboxypeptidases (CPs) perform many diverse physiological functions by removing C-terminal amino acids from proteins and peptides. Some CPs function in the degradation of proteins in the digestive tract while other enzymes play biosynthetic roles in the formation of neuropeptides and peptide hormones. Another set of CPs modify tubulin by removing amino acids from the C-terminus and from polyglutamyl side chains, thereby altering the properties of microtubules. This review focuses on three CPs: carboxypeptidase E, carboxypeptidase A6, and cytosolic carboxypeptidase 1. Naturally occurring mutations in all three of these enzymes are associated with disease phenotypes, ranging from obesity to epilepsy to neurodegeneration. Peptidomics is a useful tool to investigate the relationship between these mutations and alterations in peptide levels. This technique has also been used to define the function and characteristics of CPs. Results from peptidomics studies have helped to elucidate the function of CPs and clarify the biological underpinnings of pathologies by identifying peptides altered in disease states. This review describes the use of peptidomic techniques to gain insights into the normal function of CPs and the molecular defects caused by mutations in the enzymes.

Keywords: Carboxypeptidase; Metallopeptidase.

Figure 1. Subfamiles and Domain Structure of M14 Metallocarboxypeptidases

The M14 family of metallocarboxypeptidases is divided into four subfamilies differentiated by domain structure and homology. The enzymatic CP domain is shown with grey shading; this domain has amino acid sequence homology between various members of the family, whereas all other domains are only conserved within the subfamily and not between the different subfamilies. All active members of the M14 family contain the six conserved residues shown, which coordinate the central zinc ion (H69, E72, H196) and participate in substrate binding (R145, Y248) or catalysis (E270). As per convention in the field, the numbering system of these residues is based on their position in the mature form of bovine CPA1. All members of the A/B and N/E subfamilies contain an N-terminal signal peptide of ~20–30 amino acids (indicated by black boxes); none of the CCP or AA subfamily members contain this domain. All but one member of the CPA/B subfamily contain an inactivating prodomain of ~90 residues (indicated by shading with vertical lines), located N-terminal to the carboxypeptidase domain. This region is folded into a beta sheet-rich structure. Members of the N/E subfamily do not have this prodomain, but have C-terminal transthyretin-like domain thought to be involved in protein folding; although no sequence or structural homology to the prodomain of most A/B subfamily members, the transthyretin-like domain of the N/E subfamily members folds into a beta sheet-rich structure (indicated by vertical shading). There are a total of 8 members of the N/E subfamily, although only 5 of these have been shown to have carboxypeptidase activity; three other family members are not active as carboxypeptidases and lack one or more of the critical residues indicated in the figure. Members of the CCP family have several additional domains of unknown function; the length of these domains vary by family member. CCP1 has a large N-terminal domain and a shorter C-terminal domain. All CCPs contain an N-terminal domain that folds into a beta sheet-rich structure (indicated by vertical lines). Aminoacylases have a C-terminal shielding domain that limits accessibility to the active site. The C-terminal domain folds into a beta sheet-rich structure (indicated by vertical lines). In the aminoacylases, the residue that is functionally equivalent to Tyr 248 is located within this C-terminal domain, but in the three-dimensional structure folds into a conformation with an active site very similar to those found in other subfamilies.

Figure 2. Quantitative peptidomic approaches to study carboxypeptidases

A. Chemical reagents used to label free amines on peptides. TMAB-NHS, trimethylammoniumbutyryl-N-hydroxysuccinimide; Me, methyl; d, deuterium. The five different isotopic forms of this compound are named D0, D3, D6, D9, and D12, although the latter compound contains 9 deuterium atoms, and one 13C for a 12 Dalton mass difference from the D0 form. B. Representative experiment to examine which native peptides can be cleaved by a peptidase. In this example, peptides are extracted from the biological sample and then divided into five aliquots, each of which is incubated with a different concentration of enzyme. Following the reaction, the peptides are labeled with the TMAB-NHS isotopic tags, pooled, and analyzed by liquid chromatography/mass spectrometry (LC/MS). C. A representative experiment to examine which peptides are altered by the absence of an enzyme in knockout (KO) animals. A similar approach can be used with knockdown or overexpression techniques. In the example shown, three wild-type animals (WT) are compared to two KO animals. Peptides are extracted from the tissues, labeled with the TMAB-NHS tags, pooled and analyzed by LC/MS. D. Representative data showing a peptide that is neither a substrate nor product of the enzyme. This spectrum was from an experiment testing different concentrations of carboxypeptidase D with the cellular peptidome from HEK293T cells. The peptide was subsequently identified by tandem mass spectrometry as an internal fragment of 60S acidic ribosomal protein P2 (sequence LDSVGIEADDDRLNKV) with two TMAB tags incorporated and 1 proton (total charge 3+). Although this spectrum was from the scheme shown in panel B, a similar pattern would be observed for studies using the scheme shown in panel C for peptides that are neither substrates nor products. E. Representative data showing a peptide that is a substrate for carboxypeptidase D, from the same experiment as in panel D, using the scheme shown in Panel B. The peptide was subsequently identified by tandem mass spectrometry as a C-terminal fragment of 40S ribosomal protein S28 (sequence KGPVREGDVLTLLESEREARRLR) with 2 TMAB tags and 4 protons (total charge 6+). Note the dramatic decrease in peak height with the highest concentration of enzyme (D9) and a moderate decrease in peak height with the second highest concentration of enzyme (D6). F. Representative data showing a peptide that is elevated in the striatum of adult mice lacking active CCP1 due to a naturally occurring mutation (pcd^3J). This peptide was not identified by tandem mass spectrometry. In the original report describing the changes in the peptidome in the adult pcd mouse brain, peptides that were elevated were thought to be substrates of CCP1 [75]. However, studies testing purified CCP1 with substrates showed a specificity for C-terminal Glu residues and no activity against peptides with C-terminal hydrophobic residues, whereas most of the peptides elevated in the peptidomics analysis of pcd mice contained C-terminal aliphatic residues [–13]. Further peptidomic analysis of pcd mouse tissues showed that levels of peptides were not altered in the brains of young animals, taken prior to the onset of neurodegeneration [81]. Thus, while the result in panel F is consistent with the change expected for a CCP1 substrate, it is more likely an indirect change that results from the neurodegeneration of cerebellar Purkinje cells. This emphasizes the need to perform both types of studies; those testing the activity of purified enzymes with mixtures of peptides (as in Panels B, D, and E) and those testing animals lacking the enzyme (as in Panels C and F).