REVIEW ARTICLE Mammalian transglutaminases Identification of substrates as a key to physiological function and physiopathological relevance Carla Esposito and Ivana Caputo Department of Chemistry, University of Salerno, Italy Mammalian transglutaminases and their catalytic activity Transglutaminases (TGs; EC 2.3.2.13) are encoded by a family of structurally and functionally related genes. Nine TG genes have been identified, eight of which encode active enzymes [1]. Only six TG enzymes have been isolated and characterized at the protein level. The TG enzyme family (Table 1) comprises: (a) the intracellular TG1, TG3 and TG5 isoforms, which are expressed mostly in epithelial tissue; (b) TG2, which is expressed in various tissue types and occurs in an intracellular and an extracellular form; (c) TG4, which is expressed in prostate gland; (d) factor XIII (FXIII), which is expressed in blood; (e) TG6 and TG7, whose tissue distribution is unknown; and (f) band 4.2, which is a component protein of the membrane that has lost its enzymatic activity, and serves to maintain erythrocyte membrane integrity [2]. In addition to diversity at the genetic level, TGs undergo a number of post-translational modifications, i.e. phosphoryla- tion, nitrosylation, fatty acylation and proteolytic clea- vage [2,3]. In most instances, TGs catalyse the post-transla- tional modification of proteins, a process that results in the formation of polymerized cross-linked proteins [3]. TGs catalyse the formation of isopeptide linkages between the c-carboxamide group of the protein- bound glutamine residue and the e-amino group of the protein-bound lysine residue, so that the reaction prod- uct results in stable, insoluble macromolecular com- plexes. In addition, TGs catalyse a number of distinct Keywords post-translational modification; protein substrates; proteomics; transglutaminase Correspondence C. Esposito Department of Chemistry, University of Salerno Via S. Allende, 84081 Baronissi, Salerno, Italy Fax: +39 089 965296 Tel: +39 089 965298 E-mail: cesposito@unisa.it (Received 27 July 2004, revised 3 November 2004, accepted 10 November 2004) doi:10.1111/j.1742-4658.2004.04476.x Transglutaminases form a large family of intracellular and extracellular enzymes that catalyse the Ca 2+ -dependent post-translational modification of proteins. Despite significant advances in our understanding of the biolo- gical role of most mammalian transglutaminase isoforms, recent findings suggest new scenarios, most notably for the ubiquitous tissue transglutami- nase. It is becoming apparent that some transglutaminases, normally expressed at low levels in many tissue types, are activated and⁄ or over- expressed in a variety of diseases, thereby resulting in enhanced concentra- tions of cross-linked proteins. As applies to all enzymes that exert their metabolic function by modifying the properties of target proteins, the iden- tification and characterization of the modified proteins will cast light on the functions of transglutaminases and their involvement in human dis- eases. In this paper we review data on the properties of mammalian trans- glutaminases, particularly as regards their protein substrates and the relevance of transglutaminase-catalysed reactions in physiological and dis- ease conditions. Abbreviations CE, cell envelope; ECM, extracellular matrix; FXIII, factor XIII; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; GTP, guanosine triphoshate; PAI-2, plasminogen activator inhibitor 2; SPR, small proline-rich protein; SV, seminal vesicle; TG, transglutaminase. FEBS Journal 272 (2005) 615–631 ª 2004 FEBS 615 reactions that lead to post-translational modification of a specific glutamine residue in the substrate [4]. The TG-catalysed reaction adds new properties to the pro- tein substrates, thereby enhancing substrate function, or more generally, altering it. The biochemical mechanism underlying the enzyme action involves a ‘ping-pong’ kinetics. The first, rate- limiting step is transamidation of the c-carboxamide group of a glutamine residue to form a thiol ester with an active site cysteine (resulting in the release of ammonia) followed by transfer of the acyl intermediate to a nucleophilic substrate, usually the e-amino group of a peptide-bound lysine residue (Fig. 1). This process results in the formation of an intermolecular isopeptide e-(c-glutamyl)lysine cross-link. However, the monomeric protein units themselves may become cross-linked internally [5]. Low molecular mass amines, especially polyamines, can replace lysines in transamidating reactions and result in the formation of N-mono(c- glutamyl)polyamine. In the presence of a second react- ive glutamine residue, the reaction may proceed to covalent cross-linking between two polypeptide chains via a N,N-bis(c-glutamyl)polyamine bridge. In the absence of suitable amines, water can act as a nucleo- phile and so cause deamidation of protein-bound glutamine residues [4]. The various TG gene products share a high degree of sequence similarity. The sequences around the active site are the most highly conserved (Fig. 2). Elucidation of the three-dimensional structure of FXIIIA and TG2 [6,7] revealed a cysteine proteinase-like active site comprising the catalytic triad cysteine, histidine and aspartic acid that is required for transamidation. A four-sequential domain arrangement is highly con- served in TG isoforms [2]. It consists of an N-terminal b-sandwich, a core (which contains a transamidation site and a Ca 2+ -binding site, and has a helices and b sheets in equal amounts), and two C-terminal b-bar- rel domains. It has been suggested that glutamyl sub- strates approach the enzymes from the direction of two b barrels, whereas lysyl substrates might approach the enzymes from the direction of the active site [2]. Although the relative positions of residues in the sub- strate-binding site region are highly conserved in TGs, the charge distribution differs among the various iso- enzymes. This difference may account for the different substrate specificities and hence the specialized func- tions of each isoenzyme. Intriguingly, TG2 and TG3 possess a site that binds and hydrolyses GTP even though the site lacks any obvious sequence similarity with canonical GTP-bind- ing proteins [7,8]. The primary sequence of TG5 contains a similar GTP-binding pocket, and TG5 transamidating activity is also inhibited by GTP in vitro [9]. It is noteworthy that TG2 intracellular GTPase activity, which is involved in the transduction of extracellular a 1 -adrenergic signals [10], occurs inde- pendently of cross-linking activity, but both activities are regulated by binding to GTP and Ca 2+ ([11] and references cited therein). GTP-hydrolyzing and tran- samidating activities are also regulated by enzyme translocation from the cytosol to the cell membrane. In fact, TG2 from the cytosolic compartment has higher cross-linking activity than membrane TG2, whereas the GTPase function of TG2 predominates when the enzyme is associated to cell membranes [12]. Substrate requirements for transglutaminases Although the mechanism governing the recognition of the target amino acids within the TG protein sub- strates is not known, some indications emerge from in vitro data. As regards glutamine specificity, two adjacent glutamine residues act as amine acceptors in a consecutive reaction, e.g. bA 3 -crystallin [13], sub- stance P [14], osteonectin [15] and insulin-like growth factor-binding protein 1 [16]. The spacing between the targeted glutamine and neighbouring residues is a crucial factor in the specificity of TGs. Positively charged residues flanking the glutamine residue dis- courage the TG reaction, at least in unfolded protein regions. In contrast, positively charged residues at two or four residues from the glutamine promote the reac- tion. Glycines and asparagines adjacent to the target Table 1. The mammalian transglutaminase family. Names Synonyms kDa Tissue Location TG1 TG K , keratinocyte TG, type 1 TG 90 Epithelia Cytosolic, membrane TG2 TG C , tissue TG, type 2 TG 80 Ubiquitous Cytosolic, nuclear, extracellular TG3 TG E , epidermal TG, type 3 TG 77 Epithelia Cytosolic TG4 TG P , prostate TG, type 4 TG 77 Prostate Extracellular TG5 TG X , type 5 TG 81 Epithelia Cytosolic TG6 TG Y , type 6 TG Unknown Unknown Unknown TG7 TG Z , type 7 TG 80 Ubiquitous Unknown FXIII Factor XIIIA, plasma TG, fibrin stabilizing factor 83 Blood plasma, platelets Extracellular Band 4.2 Erythrocyte protein band 4.2 77 Erythrocytes Membrane TGs and their substrates C. Esposito and I. Caputo 616 FEBS Journal 272 (2005) 615–631 ª 2004 FEBS glutamine may favour substrate accessibility [17,18]. Proline residues seem to be important in the recogni- tion of a given glutamine residue by the enzyme. In fact, a glutamine residue is not recognized as a sub- strate by the enzyme if it occurs between two proline residues [19]. Arentz-Hansen et al. examined the selectivity of human TG2 for glutamine residues, in gliadin peptides, in the generation of epitopes recognized by coeliac lesion CD4+ lymphocytes [20]. This was a challenging study because gliadin is an excellent TG2 substrate being comprised of  30–50 mol% of glutamine (Q), 15 mol% of proline (P) and 19 mol% of hydrophobic amino acids [21]. TG2 specifically deamidated Q65 (underlined) in the 57–68 peptide (QLQPFPQP QLPY) of A-gliadin. Therefore, in most cases the enzyme recognized QxP (where x represents a variable amino acid, and indicates the distance between glutamine and Fig. 1. TG-catalysed acyl transfer reactions. The c-carboxamide group of a glutamine residue (Q-donor) forms a thiol ester with the active site cysteine, and ammonia is released. (A) e-(c-Glutamyl)lysine cross-link formation; (B) N-mono(c-glutamyl)polyamine formation; (C) deamidation of protein-bound glutamine residue. Fig. 2. Comparison of the amino acid sequences of human TGs around the active site (black box). Dashes indicate gaps inserted to optimize sequence alignment. Boxed regions are regions in which amino acids are conserved in at least four gene products. Grey columns indicate the presence of conserved amino acids in all TGs. C. Esposito and I. Caputo TGs and their substrates FEBS Journal 272 (2005) 615–631 ª 2004 FEBS 617 proline), rather than QP or QxxP [22]. Moreover, to act as TG substrates, glutamine residues must be exposed at the surface of the protein or, more gener- ally, located in terminal extensions protruding from the compactly folded domains, where they can be accessible to covalent modification; N- and C-terminal glutamine residues are not recognized by the enzyme [19]. Therefore, it appears likely that the secondary and ⁄ or tertiary structure of the protein, rather than the location of the glutamine within the primary struc- ture itself, determines where cross-linking occurs [18]. This is supported by evidence that distinct TGs recog- nize distinct glutamine residues in the same protein; for instance, several typical FXIIIA substrates may also serve as substrates for TG2, albeit with a much lower affinity [23]. TGs are much less selective toward amine donor lysine residues than toward glutamine residues. For example, Lys148, 176, 183, 230, 413 and 457 in the Aa chain of fibrinogen cross-linked to only glutam- ines 83 and 86 in plasminogen activator inhibitor 2 (PAI-2) during cross-linking by TG2 and FXIIIA [24]. As in the case of TG recognition of glutamine residues, the nature of the amino acids directly pre- ceding the lysine may influence the latter’s reactivity [25]. Indeed, uncharged, basic polar and small ali- phatic residues enhance reactivity, whereas aspartic acid, glycine, proline and histidine residues reduce reactivity [26]. An exception to this rule is Lys191 in glyceraldehyde-3-phosphate dehydrogenase (GAPDH) which is preceded by a glycine that adversely affects the TG reaction [27]. Moreover, other GAPDH lysine residues are not amine donors even though they are located in regions with sequences that should enhance their reactivity. These observations suggest that the steric hindrance between enzyme and substrate prevents TG recognition of specific lysine residues. As a result, only a limited number of lysine residues in lysine-rich peptides ⁄ proteins are able to act as an amine donor for TG, e.g. one of five lysyl residues in b-endorphin [28], six of nine in seminal vesicle (SV) protein IV [5], and one of ten in both aB-crystallin [29,30] and S100A11 [31]. However, conformational changes in the native pro- tein induced by protein–protein interactions may affect the ability of some lysine residues to serve as TG substrates. Lys191, 268 and 331 of the 26 lysine residues in GAPDH are reactive amine donor sites that form cross-links with substance P, which bears the simplest Q n domain (n ¼ 2). Other GAPDH lysine residues (Lys248, 251, 256, 257 and 260) were recognized by TG2 in the presence of the polyQ 17 and polyQ 43 peptides, thus indicating that the polyQ n –GAPDH interaction makes GAPDH a better TG2 substrate in vitro [32]. Techniques for identifying transglutaminase substrates The intrinsic cross-linking activity of TGs tends to convert target proteins into massive, probably disor- dered, insoluble aggregates of multiple proteins. Con- sequently, it is difficult to identify individual protein substrates and to investigate alterations in their prop- erties. Nevertheless, biochemical and functional prote- omic studies in both in vitro and cellular systems have furthered our understanding of TG-modified proteins. However, although numerous TG substrates (both glu- tamine and lysine donors) have been identified in vitro, fewer have proved to be substrates in vivo. The detection of polymer formation by SDS ⁄ PAGE and ⁄ or western blot, and protein-to-protein cross-link- ing inhibition by amine- or glutamine-rich peptide incorporation is the most widely used indirect method of identifying TG protein substrates. Various proce- dures are used to identify TG substrates and the protein domains that function as acceptors in the cross-linking process, i.e. TG-catalysed labelling of iso- lated peptides ⁄ proteins with radioactive amines [33], monodansylcadaverine, fluoresceincadaverine [34] and 5-biotinamidopentylamine [35], or with dansylated or biotinylated glutamine-containing peptides such as dansyl-e-aminocaproyl-QQIV, -TVQQEL [29] and dan- syl-substance P [27]. The reactivity of TG to protein substrates in vitro does not necessarily mean that the proteins are sub- strates in vivo. Cross-linking in vivo can be evaluated by conducting in situ assays with whole cells ⁄ tissue. With an in situ assay it is possible not only to determine the amine acceptor ⁄ donor substrates in vivo, but also to assess the affinity of a TG for the interaction with the protein substrate in the presence of physiologically occurring alternative substrates. This procedure also yields information about the specific functions of a TG isoform, and about the physiological consequences of TG-catalysed post-translational modification of the protein substrate. It entails use of cell-penetrating syn- thetic TG substrates that do not interfere with normal cell processes. The donor-carrying reporter groups used are dansylated or biotinylated amines (e.g. 5-biotin- amidopentylamine, 3-[N a [N e -[-2¢,4¢-dinitrophenyl]- amino-n-hexanoyl-l-lysyl-amido]propane-1-ol) [36] or glutamine-containing peptides (e.g. penetratin-1-linked peptide) [37]. The advantage of this strategy is protein separation via affinity chromatography followed by identification of the labelled TG-reactive protein. TGs and their substrates C. Esposito and I. Caputo 618 FEBS Journal 272 (2005) 615–631 ª 2004 FEBS Depending on the probe used, the labelled substrates can be visualized by direct fluorescence microscopy, fluorography and western blot analysis, and identified by N-terminal sequencing or by MS. FAB ⁄ MS has yielded data on TG-mediated cross-links in the small purified monomeric proteins substance P [14], b-endo- rphin [28] and SV-IV [5]. Currently, TG protein substrates are identified using a procedure that combines gel electrophoresis separation with MS-based analyses. Tandem MS based on data-dependent analyses [38] has led to functional proteomic strategies in which TG protein substrates and the enzyme-sensitive amino acid site are identified in mixtures that have not undergone gel electrophoretic separation. Identification of protein substrates for transglutaminase-catalysed cross-linkage The recently created TRANSIT database (http:// crisceb.unina2.it/ASC/) lists  150 protein sequences that function as TG substrates [39]. The TRANSIT database also lists protein substrates from food, yeast and viruses. Our review focuses on mammalian TG pro- tein substrates. TG1, TG3 and TG5 Mammalian epidermis harbours at least four TG iso- forms (TG1, TG2, TG3 and TG5). These play con- secutive and complementary roles in the formation of a specialized structure known as the cornified cell envelope (CE) [40] on the intracellular surface of the plasma membrane of keratinocytes undergoing ter- minal differentiation. These TGs induce cross-linking of the various proteins that constitute the CE. TG2 is expressed only in the basal layer, whereas TG1, TG3 and TG5 are expressed in the upper layers [41]. Mem- brane-bound TG1 is the most abundant TG isoenzyme and is predominantly involved in epithelial differenti- ation [42]. Moreover, TG1 catalyses the ester linkage of specialized ceramides to CE proteins [43]. Numerous CE proteins are substrates cross-linked by TGs: invo- lucrin [41,44], loricrin [41,45], small proline-rich pro- teins (SPR) [46], cystatin a [47], trichohyalin [48], keratins [49], cornifin [50], sciellin [51], S100A11 [31], filaggrin [45], elafin [45,52], desmoplakin [45], envopla- kin [53], periplakin [48] and suprabasin intermediate filaments [45,54] (Table 2). In vitro, loricrin, SPR 1, -2 and -3, and trichohyalin functioned as complete sub- strates for TG1 and TG3 [48]. In addition, each Table 2. TG1, TG3 and TG5 protein substrates. IF, Intermediate filaments; SPRs, small proline-rich proteins. Protein substrates were identi- fied by functional proteomics. RL, radiolabelling; CL, cross-linking; P, proteolysis; L, labelling; S, sequencing; WB, western blot. Protein substrate Reactive Q a Reactive K a Method of identification Reference Cornifin ? ? Epidermal extract; WB [50] Cystatin a b —?L⁄ WB [47] Desmoplakin b 1646 ? Epidermal extract; P ⁄ S[45] Elafin 2, 59 6, 58, 60 Epidermal extract; P ⁄ S [45,52] Envoplakin ? ? Epidermal extract; WB [53] Epiplakin ? ? Hair follicle; P ⁄ S[48] Filaggrin 246, 247 ? Epidermal extract; P ⁄ S[45] IF b ? ? Epidermal extract; P ⁄ S[45] Involucrin 465, 496 468 RL ⁄ P ⁄ S [41,44] Keratins ? 9, 32, 71, 73, Epidermal extract; L ⁄ P ⁄ S[49] Loricrin 3, 6, 10, 153, 156 215, 216, 219, 225, 305, 306 4, 5, 88, 307, 315 Epidermal extract, L ⁄ P ⁄ S[41,45] Repetin ? ? Hair follicle; P ⁄ S[48] S100A7 b 5 ? Epidermal extract ⁄ WB [31] S100A10 b 4?L⁄ P ⁄ S S100A11 b 102 3 Sciellin ? ? Epidermal extract; WB [51] SPRs 4, 5, 16, 17, 18, 19, 87, 167 6, 21, 71, 164, 166, 168 Epidermal extract; P ⁄ S[46] Suprabasin ? ? CL ⁄ WB [54] Trichohyalin ? ? L ⁄ WB [48] a Q? and K? indicate that reactive glutamine and ⁄ or lysine are present but that the specific residue is not known; – indicates a lack of evidence for the presence of reactive glutamine. b Also in vitro TG2 substrate. C. Esposito and I. Caputo TGs and their substrates FEBS Journal 272 (2005) 615–631 ª 2004 FEBS 619 isoenzyme preferred selected reactive glutamine and lysine residues on the same substrate in vivo. However, like S100 proteins, which are a family of calcium- dependent signal transduction mediators, both TG1 and TG2 modify the same sites on S100A11 (i.e. Q102) and the rank order of reactivity of the three S100 proteins (A7, A10 and A11) is the same regard- less of which TG is involved [31]. Key substrates such as loricrin, involucrin and SPR3 are cross-linked by TG5 in the initial stage of epidermal differentiation. The small oligomers formed are cross-linked to the CE structure by the cytosolic TG3 isoenzyme and subse- quently by the membrane-bound TG1 enzyme [41]. Derangement of the mechanisms that lead to ter- minal keratinocyte differentiation might be involved in lamellar ichthyosis, in hyperkeratinization conditions such as psoriasis, and in some dermatitis disorders (e.g. herpetiform disorders through autoimmunity against TG3). Research is underway to develop drugs based on natural retinoids and synthetic retinoid-like agents that will regulate expression of TGs in the skin [55]. TG2 A large body of data is available for TG2. The results obtained in structural and functional proteomic studies are summarized in Tables 3 and 4, respectively. Both intracellular and extracellular proteins are recognized and post-translationally modified by TG2. Despite the lack of a leader sequence, TG2 is externalized from cells into the extracellular space where it has been implicated in the stabilization of the extracellular mat- rix (ECM) and in cell–ECM interactions by cross-link- ing matrix proteins [56]. Under ‘normal conditions’ TG2 externalized from cells becomes tightly bound to fibronectin and forms ternary complexes with collagens that function as a cementing substance in the ECM. This mechanism probably serves to clear TG2 from the circulation to prevent it inducing adverse effects. Fibronectin, a protein abundant in the extracellular space, is a major TG2 substrate in vitro and in vivo [57,58]. The other proteins involved in the assembly, remodelling and stabilization of the ECM are fibrino- gen ⁄ fibrin [24], von Willebrand factor [59], vitronectin [60], lipoprotein(a) [61], laminin and nidogen [17]. All have been identified as TG2 substrates in vitro (Table 3). The reversible interactions between mole- cules that form heteromeric complexes in the ECM of specific tissues, e.g. laminin–nidogen [17], fibronectin– collagen [62–64] and osteonectin–vitronectin [65], are stabilized by TG2 [42]. Perturbation of ECM forma- tion has been implicated in such diseases as liver, renal and pulmonary fibrosis, as well as atherosclerosis [66]. It is noteworthy that TG2 activity is increased and the number of e-(c-glutamyl)lysine cross-links is enhanced in all fibrotic disorders characterized by excessive scar tissue. Furthermore, TG2 contributes to the organiza- tion of the ECM by stabilizing the dermo-epidermal junction via cross-linking of the basement membrane components fibrillin-1, the major protein of micro- fibrils, microfibril-associated glycoprotein-1 and latent transforming growth factor binding protein [67,68]. Latent transforming growth factor binding protein-1 is particularly interesting because only after its TG2-cata- lysed linkage to the matrix does it release the active transforming growth factor b. Consequently, TG2 is presumed to be involved in the pathogenesis of chronic inflammatory diseases such as rheumatoid arthritis and osteoarthritis via regulation of the availability of this cytokine in the matrix [69]. In addition, extracellular TG2 might play a role in tissue mineralization by cata- lyzing the formation of the cross-linked clusters of the Ca 2+ -binding proteins osteonectin and osteopontin at the cell surface [70–72]. More intracellular proteins have been identified as TG2 acyl-donor and ⁄ or acyl-acceptor substrates in in vitro studies (Table 3) than in functional proteomic studies (Table 4). However, functional proteomics is a promising tool with which to identify differently labelled cellular proteins in relation to physiology and disease. Indeed, this technique allows one to explore the cross-linking pattern in such conditions as normal vs. neoplastic or metastatic cells, and normal vs. prolif- erating or necrotic ⁄ apoptotic cells, as well as to screen for differences in TG substrates between quiescent and differently stimulated cells. A large number of TG2 substrates are proteins involved in the organization of the cytoskeleton. In the cytoskeleton, the TG2 isoform colocalizes with stress fibres and, by virtue of its auto- catalytic activity, it cross-links to myosin. Upon activa- tion by Ca 2+ , TG2 contributes to the organization of the cytoskeleton by cross-linking various cytoskeletal proteins, i.e. microtubule protein tau [73–75], b-tubulin [76], actin [36,77], myosin [78], spectrin [78], thymo- sin b [77,79], troponin T [80,81] and vimentin [82]. This extensive polymerization, which occurs during the final steps of apoptosis, stabilizes the structure of the dying cells thereby preventing release of cell compo- nents that might give rise to inflammatory or auto- immune responses [83]. Interestingly, actin is a TG2 substrate during apoptosis in vivo [74]. Also the retino- blastoma gene product is a TG2 substrate during apoptosis in vivo and its polymerization has been indi- cated as a key signal for the initiation of apoptosis [84]. Moreover, nuclear proteins such as core histones TGs and their substrates C. Esposito and I. Caputo 620 FEBS Journal 272 (2005) 615–631 ª 2004 FEBS Table 3. TG2 protein substrates identified by structural proteomics. BHMT, betaine-homocysteine S-methyltransferase; EMP b-3, erythrocyte membrane protein band 3; ERM, ezrin–radixin–moesin binding phosphoprotein 50; KGDHC, a-ketoglutarate dehydrogenase; IGFBP-1, insulin- like growth factor-binding protein 1; MAGP-1, microfibril associated glycoprotein-1; MBP, myelin basic protein; NSB, nuclease sensitive ele- ment binding protein-1; PGD, phosphoglycerate dehydrogenase; PLA 2 , phospholipase A 2 ; Pro-CpU, procarboxypeptidase U; PSA, prostate- specific antigen; RAP, receptor-associated protein; ROCK-2, Rho-associated coiled-coil-containing protein kinase 2; UV RAD23, UV excision repair protein RAD23; VIP, vasoactive intestinal peptide. RL, radiolabelling; CL, cross-linking; P, proteolysis; MS, mass spectrometry; L, label- ling; S, sequencing; WB, western blot; M, mutagenesis. Protein substrate Reactive Q a Reactive K a Method of identification b Reference Actin 41, 167 ? CL ⁄ P ⁄ MS [77] Aldolase ? — L ⁄ P ⁄ S[35] Amyloid bA4 15 16, 28 L ⁄ P ⁄ S [86,87] BHMT ? ? RL [104] BiP protein ? — L ⁄ in gel P ⁄ MS [78] C1 inhibitor 453 — L ⁄ P ⁄ S [105] C-CAM ? ? CL ⁄ WB [106] Cementoin ? ? CL ⁄ WB [52] Chaperonin subunit 3 — ? L ⁄ in gel P ⁄ MS [78] Clathrin heavy chain ? — L ⁄ in gel P ⁄ MS [78] Collagen III, V, XI 14, ?, ? ? L ⁄ P ⁄ S [63,64] Crystallin bA3 23, 24 17 L ⁄ P ⁄ S ⁄ M [13,30] Crystallin bB2 9 — Crystallin bB3 21 — Crystallin aB? 175 Cythocrome c 42 — L ⁄ P ⁄ S[33] Dihydropyrimidinase-2 — ? L ⁄ in gel P ⁄ MS [78] DNase c ?— L⁄ MS [78] Elongation factor 1a —? L⁄ in gel P ⁄ MS [78] Elongation factor 1c —? L⁄ in gel P ⁄ MS [78] EMP b3 30 ? L ⁄ P ⁄ S [107] b-Endorphin 11 29 L ⁄ P ⁄ MS [28] ERM — ? L ⁄ in gel P ⁄ MS [78] Fatty acid synthase ? — L ⁄ in gel P ⁄ MS [78] F-box only protein — ? L ⁄ in gel P ⁄ MS [78] Fibrinogen A 366, 398, 399 148,176, 183, 457 CL ⁄ P ⁄ S[24] Galectin-3 ? ? L ⁄ WB [101] GAPDH ? 191, 248, 251, 256, 257, 260, 268, 331 L ⁄ P ⁄ MS [32] Glucagon 3, 20 — L ⁄ P ⁄ S [108] a 2 HS-glycoprotein ? ? L ⁄ P ⁄ S[72] Heat shock 60 kDa — ? L ⁄ in gel P ⁄ MS [78] Heat shock 70 kDa ? ? L ⁄ in gel P ⁄ MS [78] Heat shock 70 ⁄ 90 ? ? L ⁄ in gel P ⁄ MS [78] Heat shock 90 kDa ? ? L ⁄ in gel P ⁄ MS [78] Huntingtin ? — L ⁄ WB [90] Histone H1 ? ? L ⁄ P ⁄ S [85,92] Histone 2B 22, 95 — Histone 3B 5, 19, 125 — Histone 4B 27, 93 — Histone 2A 24, 104, 112 — IGFBP-1 66, 67 ? M ⁄ CL ⁄ WB [16] Importin b1 subunit — ? L ⁄ in gel P ⁄ MS [78] Insulin A chain 5, 15 — L ⁄ P ⁄ S [108] KGDHC — ? L ⁄ WB [93] Lipocortin I 19, 23 ? L ⁄ M[98] MAGP-1 20 — RL ⁄ M[68] MBP 74,122,146, 149 ?L⁄ WB [89] C. Esposito and I. Caputo TGs and their substrates FEBS Journal 272 (2005) 615–631 ª 2004 FEBS 621 are able to act as acyl-donor TG2 substrates during cell death [85]. Amyloid b-A4 peptide [86], a synuclein [86,87], the microtubule-associated tau protein [88] and myelin basic protein [89], which are all TG2 substrates in vitro, are major components of protein aggregates in the cytosol and nuclei, and in extracellular compartments in the brains of patients affected by degenerative neurological diseases. Consequently, TG-mediated cross-linking has been implicated in the pathogenesis of Alzheimer’s disease, Parkinson’s disease and in progres- sive suprabulbar palsy in which the abnormal accumu- lation of insoluble proteinaceous aggregates cause progressive neuronal death [66]. A body of evidence implicates TG2 in the aetiology of (CAG) n ⁄ Q n -diseases such as Huntington’s disease, i.e. elevated TG2 activity in the affected regions of diseased brains, colocalization of TG2 and proteinaceous complexes in cells expressing truncated huntingtin, c-glutaminyl-lysyl cross-links in nuclear inclusions in brain, and the finding that TG2 in vitro interacts with the polyglutamine domains to form cross-links with polypeptides containing lysyl groups [90–92]. Notably, GAPDH and a-ketoglutarate dehydrogenase, which are involved in energy metabo- lism, bind tightly to both huntingtin and several pro- teins involved in polyglutamine expansion disease [93]. This observation suggested that a slow decline in energy metabolism of neuronal cells may trigger the degenerative process that leads to cell death. TG2 is involved in the activation of members of the Rho-GTPase family [94–97]. In response to retinoic acid, TG2 causes transamidation of RhoA and Table 3. (Continued). Protein substrate Reactive Q a Reactive K a Method of identification b Reference Midkine 42, 44, 95 ? RL ⁄ M [100] Myosin — ? L ⁄ P ⁄ S [78] Nidogen 726 — L ⁄ P ⁄ S [17] NSB — ? L ⁄ in gel P ⁄ MS [78] Osteonectin 3, 4 — L ⁄ P ⁄ S [15] Osteopontin 34, 36 — L ⁄ P ⁄ S [71] Periplakin ? ? L ⁄ M [109] PGD — ? L ⁄ in gel P ⁄ MS [78] Phosphorylase kinase ? ? CL ⁄ S [110] PLA 2 410 CL⁄ WB [111] PSA ? — RL [113] RAP 26 ? L ⁄ P ⁄ S [117] RhoA 52, 63, 136 — L ⁄ P ⁄ S ⁄ MS [94] 40S Ribosomal SA — ? L ⁄ in gel P ⁄ MS [78] ROCK-2 ? — L ⁄ in gel P ⁄ MS [78] Sialoprotein ? ? L ⁄ P ⁄ S [72] Spectrin a ?— L⁄ in gel P ⁄ MS [78] Statherin ? ? CL ⁄ S [119] Substance P 5,6 — L ⁄ MS [14] Synapsin ? — RL [114] a-Synuclein 79 80 CL ⁄ S [87] Tau 351, 424 163, 174,180,190, 225, 234, 240 L ⁄ P ⁄ S [88] T-complex protein ? ? L ⁄ in gel P ⁄ MS [78] Thymosin b 4 23, 36 3, 18, 38 L ⁄ CL ⁄ P ⁄ MS [77,79] Thyroglobulin ? ? RL ⁄ CL ⁄ WB [118] b-Tubulin ? — RL ⁄ CL ⁄ WB [76] Tumour rejection ag-1 ? ? L ⁄ in gel P ⁄ MS [78] Uteroglobin ? 43 RL ⁄ CL [115] UV RAD23 — ? L ⁄ in gel P ⁄ MS [78] Valosin ? — L ⁄ in gel P ⁄ MS [78] Vigilin ? — L ⁄ in gel P ⁄ MS [78] Vimentin 453, 460 97, 104, 294, 439 L ⁄ P ⁄ S [82] VIP 16 21 L ⁄ P ⁄ MS [116] a Q? and K? indicate that reactive glutamine and ⁄ or lysine are present but that the specific residue is not known; – indicates that there is no evidence for the presence of reactive glutamine and ⁄ or lysine. TGs and their substrates C. Esposito and I. Caputo 622 FEBS Journal 272 (2005) 615–631 ª 2004 FEBS formation of the RhoA-Rho-associated coiled-coil-con- taining protein kinase 2, a complex that promotes the formation of stress fibres and focal adhesion com- plexes. RhoA-Rho-associated coiled-coil-containing protein kinase 2, like the ezrin ⁄ radixin ⁄ moesin intracel- lular signalling proteins and elongation factors that are critical for the assembly of junctional proteins and actin-cytoskeleton organization in intestinal epithelia, was shown to be a TG2 substrate [78]. These findings support the notion that TG2 acts as a signal transduc- tion protein by altering the function of signalling growth ⁄ differentiation factors such as the CD38 trans- membrane enzyme [96], dual leucine zipper-bearing kinase [97], insulin-like growth factor-binding protein 1 [16], lipocortin I [98] and the extracellular midkine [99–101] that are TG2 substrates in vivo (Table 4). Another interesting aspect of TG2 function is its involvement in receptor-mediated endocytosis in various cellular systems [102]. In vitro, valosin and clathrin, which are implicated in transport processes, are gluta- mine-donor substrates, whereas importin is a lysine donor [78]. Phosphoglycerate dehydrogenase and fatty acid synthase, which are involved in different metabolic processes, are TG2 substrates in vitro [78,103–119] (Table 3). Finally, the presence of autoantibodies against TG2 and its protein substrates in autoimmune diseases such as coeliac disease suggests that TG2 may cross-link potential autoantigens to itself and to other protein substrates so triggering the humoral response in autoimmune diseases [66,120]. In this scenario, TG2–protein complexes formed in vivo may function as hapten–carrier complexes [120]. An immune reaction was observed against the well- known TG2 substrates actin, myosin, tubulin, lipo- cortin I and histone H2B in patients with systemic lupus erythematosus, and against collagen and myelin basic protein in bullous pemphigoid and multiple sclerosis, respectively [66]. Besides its involvement in protein cross-linking, within the intracellular compartment, TG2 is more likely to catalyse the incorporation of polyamines into specific acyl-donor substrates especially when the con- centration of polyamines in the cell ⁄ tissue is in the millimolar range. Numerous proteins are covalently modified by polyamination in intact cells, and poly- amines can modulate the function and metabolism of the protein substrate. For example, TG2-catalysed polyamination of phospholipases A 2 increased activity of the enzyme in vitro [111], polyamination of micro- tubule-associated protein tau inhibits calpain-mediated proteolysis [73], and modification of substance P by spermine and spermidine incorporation protects the peptide against proteolysis [121]. Table 4. TG2 protein substrates identified by functional proteomics. AChE, acetylcholine esterase; GST, glutathione S-transferase; IGFBP-1, insulin-like growth factor-binding protein 1; LTBP-1, latent transforming growth factor-b binding protein-1; pRB, retinoblastoma; CL, cross-link- ing; WB, western blot; RL, radiolabelling; IP, immunoprecipitation; L, labelling; AC, affinity chromatography; S, sequencing. Protein substrate Function Localization Experimental model Method of identification Reference AchE a Neurotransmission Membrane ECM Myotubes CL ⁄ WB [103] Actin Cell morphology Cytosol HL-60 cells L ⁄ AC ⁄ S[74] CD38 a Signalling Membrane HL-60 cells RL ⁄ IP ⁄ WB [96] Collagen V ⁄ XI Fibrillar organization ECM A204 cells CL ⁄ WB [64] DLK Signalling Cytosol NIH 3T3 cells CL ⁄ WB [97] GST a Cytoprotection Cytosol Neuroblastoma cells IP ⁄ WB ⁄ S[75] Fibrillin-1 Myofibrils formation ECM Amniotic membranes CL ⁄ S[67] Fibronectin Matrix assembly Extracellular ECV304 cells L ⁄ WB [58] IGFBP-1 Signalling Extracellular Fibroblasts CL ⁄ WB [16] Lipocortin I Signalling Cytosol membrane A431 cells CL ⁄ WB [98] LTBP-1 Signalling ECM deposition Extracellular HT 1080 cells IP ⁄ WB ⁄ CL [69] Midkine Signalling EC Neurons L ⁄ WB [99] Osteonectin Bone mineralization ECM Organ culture tracheae L ⁄ WB [70] pRB Cell cycle control Nucleus U937 cells L ⁄ WB [84] RhoA Signalling Cytosol HeLa cells L ⁄ WB [95] Tau Cytoskeleton organization Cytoskeleton Brain tissue CL ⁄ WB [73] Troponin T a Cell morphology Cytoskeleton Isolated hearts perfused CL ⁄ WB [81] a Also an in vitro TG2 substrate. C. Esposito and I. Caputo TGs and their substrates FEBS Journal 272 (2005) 615–631 ª 2004 FEBS 623 TG4 TG4 is the only TG with prostate-specific and andro- gen-regulated expression. In rodents, TG4 is secreted by the anterior lobe of the prostate, also called ‘coagu- lating gland’, and induces the postmating formation of a vaginal coagulatory plug by cross-linking the major coagulating proteins, SV proteins I–V, which are secre- ted by the SV epithelium [122,123]. The SV I–V pro- teins are TG4 substrates, and SV IV was one of the first TG substrates in which glutamines and lysine resi- dues were identified by MS [5] (Table 5). TG4-cata- lysed polymeric forms of SV IV suppress epididymal sperm immunogenicity. Although no physiological function has yet been assigned to human TG4, the functions identified in the rat enzyme could apply to the human isoform because TG4 activity occurs both in human seminal plasma and on the spermatozoon surface [124]. Moreover, the major gel-forming pro- teins in human semen, semenogelin I and II, which correspond to rat SV proteins, are substrates for TG4 [125]. However, even though the rat and human enzymes are synthesized in the same organ and are unconventionally secreted, there are several differences between the rodent enzyme and the human homologue [126]. Human TG4 is expressed at a much lower level than the rat enzyme, and the two sequences share an amino acid identity of no more than 53%. Rat TG4 is very complex [127]. In fact, it is highly glycosylated and possesses a lipid anchor that is retained during enzyme apocrine secretion. It binds GTP, which acts as a negative modulator [128], and it is positively influ- enced by phosphatidic acids and SDS [127]. Finally, rat prostate secretion contains a kinesin-like protein able to act as an efficient acyl donor substrate for the enzyme in vitro. This protein substrate may be import- ant for the correct extrusion of TG4 from the coagula- ting gland [129]. FXIII Coagulation FXIII is a plasma TG, and circulates in blood as a heterotetramer consisting of two catalytic A (XIIIA) and two noncatalytic B (XIIIB) subunits Table 5. Protein substrates of TG4. SV IV, seminal vesicle I–V; CL, cross-linking; P, proteolysis; MS, mass spectrometry; L, labelling; WB, western blot. Substrate protein Reactive Q a Reactive K a Method of identification Reference Kinesin-like ? ? L ⁄ in gel P ⁄ MS [129] Semenogelin I–II ? ? CL ⁄ WB [125] SV II–IV–V ? ? L ⁄ CL [123] SV IV b 9, 86 2, 4, 59, 78, 79, 80 L ⁄ P ⁄ MS [5] a Q? and K? indicate that reactive glutamine and ⁄ or lysine are present but that the specific residue is not known. b Also an in vitro TG2 substrate. Table 6. Protein substrates of Factor XIII. Pro-CpU, procarboxypeptidase U. Proteins shown in bold have been identified by functional pro- teomics. RL, radiolabelling; CL, cross-linking; P, proteolysis; L, labelling; S, sequencing; WB, western blot. Substrate protein Reactive Q a Reactive K a Method of identification Reference a 2 -Antiplasmin b 2? CL⁄ P ⁄ MS [24] Collagen XVI ? ? RL ⁄ CL ⁄ WB [132] Fibronectin b 3— L⁄ P ⁄ S [57] Fibrinogen A 398, 399 148, 176, 208, 219, 224, 230, 413, 418, 427, 429, 448, 508, 539, 556, 580, 601, 606 CL ⁄ P ⁄ S [24,130] Filamin b ?? CL⁄ WB [134] Lipoprotein(a) b ?? RL⁄ CL ⁄ WB [61] a-Macroglobulin 670 — L ⁄ P ⁄ S [131] PAI-2 b 83, 84,86 ? CL ⁄ P ⁄ MS [133] Pro-CpU b 2, 5, 292 ? RL ⁄ P ⁄ MS [112] S19 ribosomal protein ? ? CL ⁄ Coomassie [137] Thrombospondin ? ? RL ⁄ CL [135] Vinculin ?? CL⁄ WB [134] Vitronectin b 73, 84, 86, 93 ? RL ⁄ CL ⁄ MS [60] VonWillebrand factor 313,509,560, 634 —L⁄ P ⁄ S [59] a Q? and K? indicate that reactive glutamine and ⁄ or lysine are present but that the specific residue is not known; indicate that there is no evidence for the presence of reactive glutamine and ⁄ or lysine. b Also an in vitro TG2 substrate. TGs and their substrates C. Esposito and I. Caputo 624 FEBS Journal 272 (2005) 615–631 ª 2004 FEBS [...]... Esposito and I Caputo (A2 B2) FXIII is a proenzyme that is activated by Ca2+ and thrombin generated in the final stage of the blood coagulation cascade The A subunit of coagulation FXIII (an FXIII A dimer) has been identified in the cytoplasm of platelets, megakaryocytes and monocytes–macrophages [2,3] FXIIIA catalyses the crosslinking of the fibrin(ogen) c-chains to form c–c dimers involving glutamine... 224 and 219 each accounted for between 2 and 5% (Table 6) TG2 can also cross-link the Aa chain of fibrinogen to PAI-2, but it does so by mediating cross-linking to Lys183 and Lys457 instead of Lys230 and Lys413 [24] FXIII plays an important role in haemostasis, wound healing and the maintenance of pregnancy It protects clots from plasminolysis by covalently linking a2 -antiplasmin and a2 -macroglobulin to. .. PJ & Conway E (1992) Identification of transglutaminase substrates in HT29 colon cancer cells: use of 5-biotinamidopentylamine as a transglutaminase-specific probe Biochim Biophys Acta 1136, 12–16 Nemes Z Jr, Adany R, Balazs M, Boross P & Fesus L (1997) Identification of cytoplasmatic actin as an abundant glutaminyl substrate for tissue transglutaminase in HL-60 and U937 cells undergoing apoptosis J Biol... Transglutaminase substrates: from test tube experiments to living cells and tissues Minerva Biotec 14, 149–153 Ruoppolo M, Orru S, D’Amato A, Francese S, Rovero ` P, Marino G & Esposito C (2003) Analysis of transglutaminase protein substrates by functional proteomics Protein Sci 12, 1290–1297 Facchiano AM, Facchiano A & Facchiano F (2003) Active sequences collection (ASC) database: a new tool to assign... Liu S, Cerione RA & Clardy J (2002) Structural basis for the guanine nucleotide-binding activity of tissue transglutaminase and its regulation of transamidation activity Proc Natl Acad Sci USA 99, 2743–2747 Ahvazi B, Boeshans KM, Idler W, Naxa U, Steinert PM & Rastinejad F (2004) Structural basis for the coordinated regulation of transglutaminase 3 by guanine nucleotides and calcium ⁄ magnesium J Biol... Muramatsu H, Suzuki Y, Kadomatsu K, Yoshizawa M, Hirose S, Kimura T, Sakakibara S & Muramatsu T (1997) Dimerization of midkine by tissue transglutaminase and its functional implication J Biol Chem 272, 9410–9416 101 Mahoney S -A, Wilkinson M, Smith S & Haynes LW (2000) Stabilization of neurites in cerebellar granule cells by transglutaminase activity: identification of midkine and galactin-3 as substrates. .. 141–155 102 Davies PJA, Davies DR, Levitzki A, Maxfield FR, Milhaud P, Willingham MC & Pastan IH (1980) Transglutaminase is essential in receptor-mediated endocytosis of a2 -macroglobulin and polypeptide hormones Nature 283, 162–167 103 Hand D, Dias D & Haynes LW (2000) Stabilization of collagen-tailed acethylcholinesterase in muscle cells through extracellular anchorage by transglutaminasecatalyzed cross-linking... Lens transglutaminase selects specific b-crystallin sequences as substrate Proc Natl Acad Sci USA 81, 7017–7020 Porta R, Esposito C, Metafora S, Pucci P, Malorni A & Marino G (1988) Substance P as a transglutaminase substrate: identification of the reaction products by fast atom bombardment mass spectrometry Anal Biochem 172, 499–503 Hohenadl C, Mann K, Mayer U, Timpl R, Paulsson M & Aeschlimann D (1995)... Williams-Ashman HG (1984) Transglutaminases and the clotting of mammalian seminal fluids Mol Cell Biochem 58, 51–61 Fawell SE & Higgins SJ (1987) Formation of rat copulatory plug: purified seminal vesicle secretory proteins serve as transglutaminase substrates Mol Cell Endocrinol 53, 49–52 Porta R, Esposito C, De Santis A, Fusco A, Iannone M & Metafora S (1986) Sperm maturation in human semen: role of transgutaminase-mediated... its function may depend on its subcellular and cellular localization and on access to proteins able to act as substrates Thus far, many more substrate proteins have been identified for TG2 than for the other TG isoforms However, TG2-mediated changes to a substrate protein have not always been linked to a given function In order to make this correlation, there is a need to identify as many in vivo TG substrates . REVIEW ARTICLE Mammalian transglutaminases Identification of substrates as a key to physiological function and physiopathological relevance Carla Esposito and Ivana Caputo Department of Chemistry,. human dis- eases. In this paper we review data on the properties of mammalian trans- glutaminases, particularly as regards their protein substrates and the relevance of transglutaminase-catalysed. University of Salerno, Italy Mammalian transglutaminases and their catalytic activity Transglutaminases (TGs; EC 2.3.2.13) are encoded by a family of structurally and functionally related genes. Nine