Báo cáo khoa học: Non-enzymatic developmental functions of acetylcholinesterase – the question of redundancy docx

Non-enzymatic developmental functions of acetylcholinesterase – the question of redundancy Glynis Johnson, Chrisna Swart and Samuel W. Moore Divisions of Paediatric Surgery ⁄ Molecular Biology and Human Genetics, University of Stellenbosch, Tygerberg, South Africa Acetylcholinesterase (AChE) is defined by its enzy- matic role in the hydrolysis of the neurotransmitter acetylcholine (ACh) in the synapse and neuromuscular junction. It is also expressed in cells and tissues that lack cholinergic innervation, for example, in the early embryo [1]. This has suggested that AChE may have non-classical functions, which may be broadly defined as any function outside the context of the synapse or neuromuscular junction. Such functions could be either cholinergic (enzymatic) or non-cholinergic (presumably mediated by structural sites). The latter possibility was supported by the discovery that AChE is homologous to a number of non-enzymatic cell adhesion and sig- nalling molecules that are active in neural development [2]. Evidence for non-cholinergic functions has been sought, and it has been found that AChE is capable of promoting cell adhesion and neurite outgrowth [3], amyloidosis [4] and apoptosis [5] in vitro. Interactions with a number of proteins and peptides have been reported; these include laminin-111 [6,7], collagen IV [6], fibronectin [8], the nicotinic acetylcholine receptor [9], the prion protein [10] and the amyloid beta-peptide Keywords acetylcholinesterase; heparan sulfate; laminin; neuroligin; perlecan Correspondence G. Johnson, Divisions of Paediatric Surgery ⁄ Molecular Biology and Human Genetics, Faculty of Health Sciences, University of Stellenbosch, PO Box 19063, Tygerberg 7505, South Africa Fax: +27 21 933 7999 Tel: +27 21 938 9422 E-mail: gjo@sun.ac.za (Received 13 June 2008, revised 13 August 2008, accepted 14 August 2008) doi:10.1111/j.1742-4658.2008.06644.x Despite in vitro demonstrations of non-enzymatic morphogenetic functions in acetylcholinesterase (AChE), the AChE knockout phenotype is milder than might be expected, casting doubt upon the relevance of such functions in vivo. Functional redundancy is a possible explanation. Using in vitro findings that AChE is able to bind to laminin-111, together with detailed information about the interaction sites, as well as an epitope analysis of adhesion-inhibiting anti-AChE mAbs, we have used molecular docking and bioinformatics techniques to explore this idea, investigating structurally similar molecules that have a comparable spatiotemporal expression pat- tern in the embryonic nervous system. On this basis, molecules with which AChE could be redundant are the syndecans, glypicans, perlecan, the receptor tyrosine kinase Mer, and the low-density lipoprotein receptor. It is also highly likely that AChE may be redundant with the homologous neuroligins, although there is no evidence that the latter are expressed before synaptogenesis. AChE was observed to dock with Gas6, the ligand for Mer, as well as with apolipoprotein E3 (but not apolipoprotein E4), both at the same site as the laminin interaction. These findings suggest that AChE may show direct functional redundancy with one or more of these molecules; it is also possible that it may itself have a unique function in the stabilization of the basement membrane. As basement membrane molecules are characterized by multiple molecular interactions, each contributing cumulatively to the construction and stability of the network, this may account for AChE’s apparently promiscuous interactions, and also for the survival of the knockout. Abbreviations ACh, acetylcholine; AChE, acetylcholinesterase; ApoE, apolipoprotein E; BChE, butyrylcholinesterase; BM, basement membrane; ECM, extracellular matrix; FGF, fibroblast growth factor; HSPG, heparan sulfate proteoglycan; LDL, low-density lipoprotein; PRiMA, proline-rich membrane anchor. FEBS Journal 275 (2008) 5129–5138 ª 2008 The Authors Journal compilation ª 2008 FEBS 5129 [11]. A number of structural sites on AChE that medi- ate these interactions have also been described [9,11,12]. The addition of anti-AChE mAbs to neural cells was found to ablate cell adhesion and neurite out- growth [13,14] and also to induce apoptosis [14], which suggested that the site on AChE recognized by the antibodies is necessary for both adhesion and survival. By contrast, and seemingly contrary to the in vitro evidence, the AChE knockout mouse survives. It was found that the related cholinesterase, butyrylcholin- esterase (BChE), compensates, to some degree, for the lack of AChE in synapses and neuromuscular junc- tions [15]. BChE, however, has not generally been observed to promote non-cholinergic cell adhesion [14,16], and so is unlikely to replace AChE in this context. The knockout has severe abnormalities: it is largely immobile, with deficiencies in muscle structure and function [17], requires a liquid diet in order to sur- vive, has significant behavioural abnormalities, and has nervous system defects, in particular, in the develop- ment of the eye [18]. The last-named, in particular, suggests the presence of non-classical functions, and that AChE is indispensable in this context. However, another in vivo study, using a catalytically inactive, but otherwise structurally intact, AChE in the zebrafish (which does not have BChE) showed little evidence for non-cholinergic developmental functions [19]. The zebrafish study did, however, show evidence of choli- nergic non-classical functions. Neurotransmitters are known to have morphogenetic activity; ACh in partic- ular, inhibits cell adhesion and neurite outgrowth [20]. AChE, by hydrolysing and thus removing ACh, is indirectly able to promote the opposite effect, namely, the stimulation of cell adhesion and neurite outgrowth. Nevertheless, the lack of indisputable in vivo evidence for non-cholinergic functions has led to a questioning of their relevance [21]. There is therefore a discrepancy between the in vivo evidence of the knockouts and the documented effects and interactions in vitro. Presumably, AChE is indeed capable of producing the effects seen in vitro. A possi- ble explanation, and one that is able to reconcile both sides of the debate, is that of functional redun- dancy. Redundancy appears to be fairly common in higher organisms, as suggested by the number of knockouts with no apparent phenotype. It seems to be more frequent in proteins expressed in develop- mental, rather than ‘housekeeping’, contexts. This may be due to the tendency for developmental proteins to be expressed in precise spatiotemporal patterns with a relatively smaller margin for error; redundancy may promote robustness by providing a backup or fail-safe device. Analysis of AChE in embryonic development sug- gests that there are two discrete phases of expression: the morphogenetic, corresponding to the migration and differentiation of neural crest cells; and the syn- aptogenetic, corresponding to synapse formation [1]. The start of differentiation is characterized by an increase in AChE expression. This involves the assem- bly of largely intracellular monomeric forms into tetra- mers, which are initially secreted. Concomitant with neurite outgrowth is a shift in expression from secreted to membrane-associated tetramers [22]. The tetramers are anchored in the membrane by an association with the proline-rich membrane anchor (PRiMA), a type I integral membrane protein [23]. The PRiMA has a full transmembrane domain, as well as a short 31-residue cytoplasmic domain; it is not known, how- ever, whether the PRiMA interacts with cytoplasmic molecules or with the cytoskeleton. Cells transfected with AChE cDNA show high AChE immunoreactivity on the outer margins of cell bodies and growth cones [24]. A number of in vitro studies has shown that pro- viding AChE in the culture medium or as a plate-coat- ing induces neurite outgrowth [14,25,26]. This suggests the possibility that AChE may be able to exert a mor- phogenetic effect from a location exterior to the cell as well, perhaps corresponding to the secreted tetrameric forms of early differentiation. The extracellular matrix (ECM) is a web-like net- work of proteins and proteoglycans that provides the cell with both structural support and information about its environment. The basement membrane (BM) is the layer of specialized ECM immediately surround- ing the cell. The most abundant components of the BM are laminins (laminin-111 in the developing ner- vous system) and collagen IV, which interact and self- associate to form a scaffold to which other ECM components, such as nidogen ⁄ entactin, fibronectin and perlecan, bind [27]. ECM molecules tend to be large, and the majority are modular, many with domains resembling those of other ECM molecules; this results in a multiplicity of binding sites and interactions, pro- ducing a strong and resilient molecular web which is furthermore anchored to the cell by interactions with cell-surface receptors, in particular, the integrins and a-dystroglycan [28]. The BM and the ECM in general provide the milieu through which growth factors dif- fuse, neural precursors migrate and through which the developing axons, or neurites, grow. For these pro- cesses to occur, there must be informative associations between the cell, via its surface receptors, and the BM. It is well documented that BM components are involved in such associations and promote cell migra- tion and neurite outgrowth [29]. Acetylcholinesterase redundancy G. Johnson et al. 5130 FEBS Journal 275 (2008) 5129–5138 ª 2008 The Authors Journal compilation ª 2008 FEBS We have recently identified the epitopes of seven adhesion-inhibiting anti-AChE mAbs by long synthetic peptides, and also by a microarray of short overlap- ping conformationally constrained peptides [12]. These antibodies showed a common epitope, centred on the 40 PPMGPRRFL and 90 RELSED sequences, which are linked by a salt bridge between 46 R and 94 E. Docking of the mouse AChE and laminin structures showed the major interaction site on AChE to be 90 RELSED, with contributions from 40 PPM, 46 R and 60 VDATT (Fig. 1A). The interaction site on laminin was also conformational, consisting of a number of clusters: 2718 VRKRL, 2738 YY, 2789 YIKRK and 2819 RK in the alpha1 G4 domain (Fig. 1B). In this study, we used bioinformatics and in silico docking to explore the possibility of functional redun- dancy. We took the premise that, when the mAbs ablate adhesion and induce apoptosis in neuroblas- toma cells in vitro, they may be interacting, not only with AChE, but also with another molecule or mole- cules that have similar sites. AChE may function as a backup to these molecules in vivo. We limited our investigation specifically to the AChE–laminin interac- tion, for which we have detailed information, and to those molecules expressed during the migration and differentiation ⁄ neurite outgrowth stages of neural development. This would be a preliminary step to defining ways in which AChE may indeed function non-cholinergically in vivo. Results and Discussion Clues from the laminin site The site on laminin to which AChE binds overlaps with the heparin-binding site [30]. This site was previ- ously identified with the peptide AG73 (which also binds AChE), which forms part of the site [31]. AChE competes with heparan sulfate for binding to laminin [12], suggesting that AChE may be redundant with heparin-containing molecules. Many proteoglycans are expressed during neural development, both in the ECM and on the cell surface [32]. Although much of our knowledge of proteoglycan expression patterns and function is sketchy to say the least, there is accumulating evidence that they play important roles in development, promoting cell adhe- sion, cell–cell interactions and growth factor signalling [33,34]. The protein core may be decorated with hepa- ran, chondroitin, or less frequently, dermatan sulfate, alone or in combination. Intermolecular interactions have been shown to occur both by the sugars and the protein core [32,33]; variations in sugar composition and length, together with the diversity of proteins to which they are attached, provide a multiplicity of potential binding and signalling structures. Heparan A B Fig. 1. Interacting sites of mouse AChE and the mouse laminin alpha1 chain. (A) Laminin-binding site on AChE. Detail of the mouse AChE dimer (1J06.pdb) showing the peripheral anionic site residues in yellow, with the arrow indicating the direction of the active site gorge. The laminin-binding residues are shown in magenta. (B) AChE-binding site on laminin alpha1. Detail of the G4 domain of the mouse laminin alpha1 G4-5 domain pair (2JD4.pdb). Residues inter- acting with AChE are shown in cyan. G. Johnson et al. Acetylcholinesterase redundancy FEBS Journal 275 (2008) 5129–5138 ª 2008 The Authors Journal compilation ª 2008 FEBS 5131 sulfate proteoglycans (HSPGs) in the developing nervous system have been observed to bind to a heter- ogeneous group of molecules, including proteins (NCAM, slit proteins, laminin, fibronectin and the thrombospondins) and growth factors [members of the fibroblast growth factor (FGF), Wnt, transforming growth factor b and Hedgehog families and pleiotro- phin] [33]. Binding may occur exclusively by the heparan sulfate chains, or there may be contributions from the protein core as well. HSPGs that show similar spatiotemporal expression to AChE are the membrane-associated syndecans, glypicans and testicans, and the extracellular molecules perlecan, agrin and collagen XVIII (Table 1). The syndecans are a family of four transmembrane receptors that are expressed in a variety of tissues and appear to have multiple biological functions [35]. Syndecans carry both heparan and chondroitin sul- fate chains, and their extracellular domains may be shed as functional molecules into the matrix [36]. Although all four syndecans are expressed in the developing nervous system, there are differences in their spatiotemporal distribution, and it is likely that they have different functions. Syndecans have been observed to bind various growth factors, as well as ECM molecules and LDL [37]. Knockouts of synde- cans show no obvious phenotypes [32]. Syndecan-1 has been shown to bind laminin-111 through interac- tion of the heparan sulfate with the AG73 site in the LG4 domain [38]. The glypicans are glycosylphosphatidylinositol- linked membrane HSPGs that appear to play impor- tant roles in cell growth and differentiation [39]. Like the syndecans, individual glypicans also show differ- ences in their developmental expression patterns, sug- gesting distinct functions; glypicans also modulate growth factor signalling through their heparan sulfate chains. Glypican-2 is exclusive to the nervous system, and has been shown to bind laminin-111 in vitro [40], as has glypican-1. There is no documentation of lami- nin binding by other members of the glypican family. The testicans are a subgroup of the BM-40 ⁄ SPARC ⁄ osteonectin family of modular proteins. There is no documented evidence of them binding laminin. They have an inhibitory effect on neurite outgrowth [41], which would suggest they are unlikely to show redundancy with AChE. Perlecan is a large multidomain HSPG that cross- links many cell-surface and ECM components. Apart from its role in the formation of the basement Table 1. HSPGs in the developing nervous system. The information is taken from [32–34], unless otherwise indicated. Notes A–F indicate the relevant references. Family HSPG Expression Ligands Reference Growth factors Proteins Syndecans Syndecan-1 Cell surface FGF family transforming growth factor b family pleiotrophin Laminin, fibronectin, tenascin-C a , LDL ba [39] b [55] Syndecan-2 Cell surface FGF2 Laminin, synbindin c , fibronectin c [59] Syndecan-3 Cell surface FGF2, midkine, pleiotrophin Laminin, EGFR dd [60] Syndecan-4 Cell surface FGF family Laminin, synbindin c , fibronectin c [59] Glypicans GLP-1 Cell surface FGF family Laminin, slit-1 slit-2 GLP-2 Cell surface FGF2 Laminin, thrombospondin GLP-3 Cell surface IGFII e FGF2 e [61] GLP-4 Cell surface FGF family GLP-5 Cell surface GLP-6 Cell surface Testicans Testican-1 Cell surface Testican-2 Cell surface Testican-3 Cell surface Perlecan ECM Laminin, collagen, fibronectin, tenascin, amyloid precursor protein f f [45] Agrin ECM FGF family Laminin, NCAM, tenascin-C, thrombospondin Collagen XVIII ECM Laminin, heparan sulfate Acetylcholinesterase redundancy G. Johnson et al. 5132 FEBS Journal 275 (2008) 5129–5138 ª 2008 The Authors Journal compilation ª 2008 FEBS membrane, it is believed to support various biological functions including cell adhesion, growth-factor bind- ing and apoptosis [42]. It is expressed from very early stages of development. Perlecan consists of five domains including domain I which contains the heparan sulfate attachment sites, domain II contain- ing LDL receptor repeats and domain V which is homologous to the laminin G domains. Perlecan is known to bind laminin through the AG73 site [43]. Perlecan also binds AChE, interacting with the ColQ collagen-like tail associated with the asymmetric AChE isoforms in the neuromuscular junction [44]. The ColQ-containing isoforms are not, however, expressed during the earlier stages of neural develop- ment. A recent study [45] observed colocalization of AChE and perlecan near membrane protrusion sites in fibroblasts and astrocytes, with results suggesting the possibility of interactions with amyloid precursor protein. Colocalization may indicate the presence of functional redundancy. Agrin is a multidomain HSPG that is best known for its role in the clustering of ACh receptors during synaptogenesis. It binds various molecules, including laminin-111, by both heparan sulfate-dependent and -independent means [32,33]. The interaction site on laminin, however, does not correspond to the heparin- binding site AG73 in the LG4 domain [46]. Further- more, agrin, when used as a substrate, inhibits rather than enhances neurite outgrowth. Both these factors argue against agrin as functionally redundant with AChE. Collagen XVIII, and its cleavage product endostatin, are components of the BM with structural characteris- tics of both proteoglycans and collagen. Collagen XVIII acts as a ligand for neural receptor tyrosine phosphatases, an interaction that modulates axon growth [47]. It binds laminin, albeit not at the heparin- binding site, and also itself binds heparan sulfate on the cell surface [47]. These factors suggest that collagen XVIII is an unlikely candidate as an AChE-redundant molecule. Clues from the AChE site Homologous proteins AChE belongs to the a ⁄ b hydrolase fold family of pro- teins, which includes the cholinesterases (AChE and BChE), the cholinesterase-domain proteins (the neuro- ligins, neurotactin, glutactin, gliotactin, the Dictyosteli- um crystal protein and thyroglobulin), as well as the carboxylesterases and lipases [2]. Neurotactin, glutactin and gliotactin are invertebrate proteins, whereas the neuroligins are expressed in vertebrates. The AChE site 90 RELSED falls partly within a carboxylesterase type b signature 2 (signature sequence EDCLYLNVWTP; ProSite pattern PS00941). This signature is strongly conserved throughout the a ⁄ b hydrolase fold family and occurs in the sequence surrounding a cysteine involved in a disulfide bond. This sequence conservation implies that at least part of the 90 RELSED site is conserved in the cholinesterase- domain proteins also. Additional conserved residues are found in the 40 PPMGPRRFL sequence where R46 is conserved, as it forms a salt bridge with E94. Prolines 40 and 41 are also conserved (Fig. 2). Although BChE is closely homologous to AChE (70% identity), it does not promote cell adhesion [14,16]. It also does not bind laminin in vitro [6], nor do Fig. 2. Sequence alignment of neuroligins 1–4, AChE and BChE. All sequences are human. Conserved residues are indicated by asterisks, and conservative replacements by dots. The residues (and their equivalents) forming the laminin-binding site in AChE are shown in bold. Alignment was carried out using CLUSTALW. G. Johnson et al. Acetylcholinesterase redundancy FEBS Journal 275 (2008) 5129–5138 ª 2008 The Authors Journal compilation ª 2008 FEBS 5133 the proteins dock. The neuroligins are a group of four transmembrane proteins, located in the postsynaptic membrane [48]. They form an adhesion complex with b-neurexins in the presynaptic membrane, promoting the formation of the synapse. The extracellular domain of neuroligin-1 shows 34% homology to AChE, with clear resemblances in both the 90 RELSED ( 147 QDQSED) and 40 PPMGPRRFL ( 88 PPTFERRFQ) sequences (Fig. 2). It has been proposed that neuroligin- 1 and AChE may be functionally redundant [49], with both binding to b-neurexin. Although this was not confirmed in a subsequent study [50], the neurexins show considerable alternative splicing, and it is possible that isoforms other than those tested bind. The neuro- ligins, however, do not appear to be expressed before synaptogenesis, so would not be capable of redundancy with AChE at earlier stages of development. Searches for similar motifs in other proteins Searches for the 40 PPMGPRRFL sequence (and equiva- lents with conservative replacements) showed only vari- ous AChEs and neuroligins from a number of species. Searches for the 90 RELSED sequence (and equiva- lents with conservative replacements) in neural mole- cules yielded the syntaxins, ligatin, proto-oncogene receptor tyrosine kinase Mer, perlecan and the LDL receptor. Of these, only Mer, perlecan and LDL recep- tor are expressed during migration and differentiation. Searches for the subsidiary 60 VDATT motif also involved in AChE’s interaction with laminin yielded a large number of candidates. The subset of developmen- tally associated neural proteins with both the 90 REL- SED and 60 VDATT motifs was considerably smaller: perlecan and Mer. It would appear from the position of the two motifs in the perlecan sequence that they may be situated far apart. Unfortunately, the structure of perlecan has not been solved, so this cannot be verified. Although the structure of the relevant part of Mer has also not been solved, it appears from the sequence that the motifs may be relatively close. Mer belongs to the Ax1 ⁄ Sky ⁄ Mer family of receptor tyrosine kinases, and is expressed in both embryonic and mature nervous tissue [51]. Mer appears to induce both cell adhesion and flattening, and, in combination with interleukin-3, promotes differentiation. Unlike many receptor tyrosine kinases, it does not appear to stimu- late proliferation [52]. Gas6, the product of the growth arrest-specific gene 6, is a ligand for Mer, as well as for Ax1 and Sky. It contains two laminin-like G domains, in which the receptor-binding site is located. We investigated docking of Gas6 with AChE. It was observed that Gas6 docks with AChE in the same position as lami- nin (Fig. 3A). The AChE 90 RELSED motif lies within 2A of Gas6 residues 296-298 (YLG) and 306-309 (VIRL). This site is essentially identical to that described for Gas6 binding to Ax1 [53]. The AChE 40 PPMGPRRFL peptide lies within 2A ˚ of Gas6 A B Fig. 3. Docking of AChE with Gas6 and apolipoprotein E3. (A) Detail of the docking of the mouse AChE dimer (1J06.pdb) with human Gas6 (1H30.pdb). AChE is shown in grey, and Gas6 in cyan. On AChE, the peripheral anionic site residues are shown in yellow, and the residues 40–42, 46, 60–64 and 90–95 in black. (B) Detail of the docking of the mouse AChE dimer (1J06.pdb) with human apolipoprotein E3 (1LPE.pdb). AChE is coloured grey, and apoE3, cyan. Peripheral anionic site residues are shown in yellow, and the residues 40–42, 46, 60–64 and 90–95 in magenta. Acetylcholinesterase redundancy G. Johnson et al. 5134 FEBS Journal 275 (2008) 5129–5138 ª 2008 The Authors Journal compilation ª 2008 FEBS residues 339–345 (GMQDSW) as well as F428 and D432, and the 60 VDATT motif within 2A ˚ of G298 and R299, as well as 329–332 (DPEG), 350–351 (LR) and 437–440 (IPR). It would thus appear to be a pos- sibility that AChE and Mer may be functionally redundant. The LDL receptor pentapeptide DGSDE Low-density lipoprotein domain repeats are found in a number of molecules, including perlecan. Their func- tion is not known. In the LDL receptor itself, these regions have been identified as involved in the binding of LDL. A sequence that is especially important is the conserved pentapeptide DGSDE [54]. This sequence is remarkably similar to the AChE 91 ELSED sequence. Furthermore, LDL, the ligand that binds to DGSDE, is also known to bind heparin, and it appears that the AChE site resembles heparin as both bind to the same site on laminin. It has been reported [55] that LDL binds syndecan-1. Lipoproteins are implicated in neurite outgrowth and plasticity, as well as in the pathology of Alzhei- mer’s disease, where the presence of the apoE4 allele is associated with increased risk and earlier age of onset of the disease [56]. Apolipoprotein E3 (ApoE3) pro- motes neurite outgrowth, whereas apoE4 inhibits it [57]; however, the mechanisms by which this occurs are unclear. ApoE binds the amyloid beta-peptide and colocalises with amyloid deposits. Although both iso- forms have been observed to bind, apoE4 binds with greater avidity. We investigated docking of the apoE isoforms with AChE (1LPE.pdb and 1LE4.pdb; apoE3 and apoE4, respectively). We found that AChE docked with apoE3 (Fig. 3B) again via the same site that binds laminin. ApoE residues lying within 2A ˚ of the AChE 90 RELSED motif were R142, K143, R145 and K146, while those within 2A ˚ of the 40 PPMGPRR sequence were W34, R38, R145 and L149. Those within 2A ˚ of the 60 VDATT sequence were L43, Q48, W118, E131, L133, R134, V135 and R136. The receptor-binding region of apoE has been localized between residues 135 and 151 [58], which is the same region that docks with the AChE site. Many of these residues are basic. By contrast, apoE4 does not dock with AChE. Conclusions Redundancy would explain the apparent inconsistency between the in vitro findings detailing non-cholinergic functions and the evidence from the knockout models. In this study, we have concentrated on the develop- mental functions attributed to AChE during neural crest cell migration and differentiation and on the AChE–laminin-111 interaction. Candidate molecules on the cell surface are the syn- decans and glypicans, by virtue of their heparan sulfate chains. The lack of comprehensive information on the developmental expression and interactions of HSPGs makes it difficult to narrow the field of possibilities. The neuroligins are also strong candidates, based on their homology with AChE, although there is no documented evidence that they are expressed before synaptogenesis. Another cell-surface receptor is the receptor tyrosine kinase Mer, which has similar peptide motifs to AChE, as does the LDL receptor. The only candidate for redundancy in the ECM is perlecan, by the double virtue of its heparan sulfate chains and sequence similarity to both AChE and the LDL receptor. Extracellular matrix molecules characteristically demonstrate multiple interactions, by means of various sites. Many are modular with several types of domains; a number have laminin G-like domains, thus re- sembling the laminin site with which AChE interacts in vitro. These include agrin, pentraxin, slit, serum amyloid P component, Gas6 and b-neurexin. Docking results indicate that AChE may bind Gas6, and an interaction of AChE with b-neurexin has been postu- lated on the basis of the AChE-neuroligin homology [49]. AChE has been found to interact with a number of ECM molecules: laminin, collagen IV and fibro- nectin, as well as the amyloid beta-peptide, and also appears to have a number of interaction sites itself. AChE thus, in its potential for multiple interactions, resembles ECM molecules. Interestingly, a majority of the molecules – the amyloid beta-peptide, laminin, collagen, fibronectin, perlecan, various HSPGs, apoE, agrin, serum amyloid P component – with which AChE interacts or may interact, are found in amyloid deposits. From an evolutionary perspective, the cholinesterases, cholinesterase-domain proteins and ACh appear to have been around for a very long time: AChE and ACh, in particular, are found in bacteria, algae and protozoa as well as, as far as is known, throughout the plant and animal kingdoms. Cholinesterase-domain cell adhesion molecules have been described not only in mammals and insects, but also in the slime mould Dictyostelium, suggesting that the split between enzymes and non-enzymes may have occurred in the earliest life-forms. Presumably, the common ancestor had both enzymatic and adhesive characteristics, and the cholinesterase-domain protein branch of the family G. Johnson et al. Acetylcholinesterase redundancy FEBS Journal 275 (2008) 5129–5138 ª 2008 The Authors Journal compilation ª 2008 FEBS 5135 specialized in cell adhesion and signalling, losing their catalytic function. AChE, however, retained and perfected its enzymatic ability, while apparently at the same time retaining its adhesive capability. Although early organisms did not have nervous systems, it is possible that ACh and AChE may have become involved in morphogenesis. It is thus possible that AChE’s morphogenetic functions, both enzymatic and non-enzymatic, may be more ancient than its synaptic role. Such presumed antiquity suggests there may have been a distinct selective advantage in retaining these functions, that they fulfill a definite role and are neither trivial nor fortuitous. The bioinformatic evidence presented here indicates that AChE might mimic functions of the syndecans, glypicans, Mer or the LDL receptor on the cell mem- brane, or of perlecan in the BM. This could result in functional redundancy in the strict sense of the word, with one molecule substituting directly for another. This is supported by the findings with antibodies, where incubation of cells with antibodies resulted in a loss of adhesion followed by apoptosis, indicating the blocking of an essential site. It could also, however, be something less precise: that AChE, through its ability (demonstrated and postulated) to interact with a vari- ety of matrix molecules, simply acts to enhance the stability of the BM. This would be advantageous, so presumably would have been retained by natural selec- tion, and would also account for AChE’s ability to promote cell adhesion and neurite outgrowth, both of which depend heavily on a favourable matrix. It is also entirely possible that AChE may function in both ways, as a direct backup molecule and as an enhancer of BM stability. The findings also suggest that AChE, through its multiple interactions, may play a significant role in amyloidosis. Experimental procedures The identification of similar structures was carried out on ProSite (http://www.expasy.ch/tools/scanprosite). The sequences of mouse (NP 033729) and human (P22303) AChE were used. Other sequences used were mouse laminin (NP 032506), human neuroligin 1 (NP 055747), mouse neu- roligin 1 (NP 619607), human receptor tyrosine kinase Mer (NP 006334), human perlecan (P98160), mouse perlecan (Q05793) and human LDL receptor (NP 000518). Docking was performed by hex 4.5. This program uses rigid-body docking, and spherical polar Fourier correla- tions to accelerate docking. Structures were downloaded from the Protein Data Bank (http://www.rcsb.org/pdb/): mouse AChE dimer (1J06.pdb), mouse laminin alpha1 G4-5 domain pair (2JD4.pdb), C-terminal LG domain pair of human Gas6 (1H30.pdb) and the LDL receptor binding domain of human apolipoprotein E3 (1LPE.pdb) and human apolipoprotein E4 (1LE4.pdb). Acknowledgements We thank the National Research Foundation, the Medical Research Council of South Africa, and the Harry Crossley Foundation of the University of Stel- lenbosch for financial support. References 1 Layer PG & Willbold E (1995) Novel functions of cholinesterases in development, physiology and disease. Prog Histochem Cytochem 29, 1–95. 2 Holmquist M (2000) Alpha ⁄ beta-hydrolase fold enzymes: structures, functions and mechanisms. Curr Protein Peptide Sci 1, 209–235. 3 Paraoanu LE & Layer PG (2008) Acetylcholinesterase in cell adhesion, neurite outgrowth and network forma- tion. FEBS J 275, 618–625. 4 Inestrosa NC, Dinamarca MC & Alvarez A (2008) Amyloid–cholinesterase interactions: implications for Alzheimer’s disease. FEBS J 275, 625–632. 5 Zhang XJ, Yang L, Zhao Q, Caen JP, He HY, Jin QH, Guo LH, Alemany M, Zhang LY & Shi YF (2002) Induction of acetylcholinesterase expression during apoptosis in various cell types. Cell Death Differ 9, 790–800. 6 Johnson G & Moore SW (2003) Human acetylcholines- terase binds to mouse laminin-1 and collagen IV by an electrostatic mechanism at the peripheral anionic site. Neurosci Lett 337, 37–40. 7 Paraoanu LE & Layer PG (2004) Mouse acetylcholines- terase interacts in yeast with the extracellular matrix component laminin-1 beta. FEBS Lett 576, 161–164. 8 Giordano C, Poiana G, Augusti-Tocco G & Biagioni S (2007) Acetylcholinesterase modulates neurite out- growth on fibronectin. Biochem Biophys Res Commun 356, 398–404. 9 Greenfield SA, Day T, Mann EO & Bermudez I (2004) A novel peptide modulates alpha7 nicotinic receptor responses: implications for a possible trophic:toxic mechanism within the brain. J Neurochem 90, 325–331. 10 Pera M, Roman S, Ratia M, Camps P, Munoz-Torrero D, Colombo L, Manzoni C, Salmona M, Badia A & Clos MV (2006) Acetylcholinesterase triggers the aggre- gation of PrP 106–126. Biochem Biophys Res Commun 346, 89–94. 11 De Ferrari GV, Canales MA, Shin I, Weiner LM, Silman I & Inestrosa NC (2001) A structural motif of acetylcholinesterase that promotes amyloid beta-peptide fibril formation. Biochemistry 40, 10447–10457. Acetylcholinesterase redundancy G. Johnson et al. 5136 FEBS Journal 275 (2008) 5129–5138 ª 2008 The Authors Journal compilation ª 2008 FEBS 12 Johnson G, Swart C & Moore SW (2008) Interaction of acetylcholinesterase with the G4 domain of the laminin alpha1 chain. Biochem J 411, 507–514. 13 Sharma KV & Bigbee JW (1998) Acetylcholinesterase antibody treatment results in neurite detachment and reduced outgrowth from cultured neurons: further evidence for a cell adhesive role for neuronal acetyl- cholinesterase. J Neurosci Res 53, 454–464. 14 Johnson G & Moore SW (2000) Cholinesterases modulate cell adhesion in human neuroblastoma cells in vitro. Int J Dev Neurosci 18, 781–790. 15 Xie W, Stribley JA, Chatonnet A, Wilder PJ, Rizzino A, McComb RD, Taylor P, Hinrichs SH & Lockridge O (2000) Postnatal developmental delay and supersensi- tivity to organophosphate in gene-targeted mice lacking acetylcholinesterase. J Pharmacol Exp Ther 292, 896– 902. 16 Mack A & Robitzki A (2000) The key role of butyry- lcholinesterase during neurogenesis and neural disor- ders: and antisense-5¢-butyrylcholinesterase-DNA study. Prog Neurobiol 60, 607–628. 17 Vignaud A, Fougerousse F, Mouisel E, Bertrand C, Bonafas B, Molgo J, Ferry A & Chatonnet A (2008) Genetic ablation of acetylcholinesterase alters muscle function in mice. Chem Biol Interact doi: 10.1016/ j.cbi.2008.04.035. 18 Bytyqi AH, Lockridge O, Duysen E, Wang Y, Wolfrum U & Layer PG (2004) Impaired formation of the inner retina in an AChE knockout mouse results in degenera- tion of all photoreceptors. Eur J Neurosci 20, 2953– 2962. 19 Behra M, Cousin X, Bertrand C, Vonesch JL, Biell- mann D, Chatonnet A & Strahle U (2002) Acetylcholin- esterase is required for neuroanl and muscular development in the zebrafish embryo. Nat Neurosci 5, 111–118. 20 Biagioni S, Tata AM, De Jaco A & Augusti-Tocco G (2000) Acetylcholine synthesis and neuron differentia- tion. Int J Dev Biol 44, 689–697. 21 Cousin X, Strahle U & Chatonnet A (2005) Are there non-catalytic functions of acetylcholinesterases? Lessons from mutant animal models. Bioessays 27, 189–200. 22 Ferrand C, Clarons D, Delteil C & Weber M.J (1986) Cellular localization of the molecular forms of acetyl- cholinesterase in primary cultures of rat sympathetic neurons and analysis of the secreted enzyme. J Neuro- chem 46, 349–358. 23 Perrier AL, Massoulie J & Krejci E (2002) PriMA: the membrane anchor of acetylcholinesterase in the brain. Neuron 33, 275–285. 24 Karpel R, Sternfeld M, Ginzberg D, Guhl E, Graess- mann A & Soreq H (1996) Overexpression of alterna- tive human acetylcholinesterase forms modulates process extensions in cultured glioma cells. J Neurochem 66, 114–123. 25 Bigbee JW, Sharma KV, Chan EL & Bogler O (2000) Evidence for the direct role of acetylcholinesterase in neurite outgrowth in primary dorsal root ganglion nerves. Brain Res 861, 354–362. 26 Olivera S, Rodriguez-Ithurralde D & Henley JM (2003) Acetylcholinesterase promotes neurite elongation, syn- apse formation and surface expression of AMPA recep- tors in hippocampal neurones. Mol Cell Neurosci 23, 96–106. 27 Timpl R & Brown JC (1996) Supramolecular assembly of basement membranes. Bioessays 18, 123–132. 28 Talts JF, Andac Z, Gohring W, Brancaccio A & Timpl R (1999) Binding of the G domains of laminin alpha1 and alpha2 chains and perlecan to heparin, sulfatides, alpha-dystroglycan and several extracellular matrix proteins. EMBO J 18, 863–870. 29 Perris RM & Perissinotto D (2000) Role of the extracel- lular matrix during neural crest cell migration. Mech Dev 95 , 3–21. 30 Harrison D, Hussain SA, Combs AC, Ervasti JM, Yurchenco PD & Hohenester E (2007) Crystal structure and cell surface anchorage sites of laminin alpha1 G4-5. J Biol Chem 282, 11573–11581. 31 Hoffman MP, Engbring JA, Nielsen PK, Vargas J, Steinberg Z, Karmand AJ, Nomizu M, Yamada Y & Kleinman HK (2001) Cell type-specific differences in glycosaminoglycans modulate the biological activity of a heparin-binding peptide (RKRLQVQLSIRT) from the G domain of the laminin alpha1 chain. J Biol Chem 276, 22077–22085. 32 Hartman U & Maurer P (2001) Proteoglycans in the nervous system – the quest for functional roles in vivo. Matrix Biol 20, 23–35. 33 Bandtlow CE & Zimmerman DR (2000) Proteoglycans in the developing brain: new conceptual insights for old proteins. Physiol Rev 80, 1267–1290. 34 Selleck SB (2000) Proteoglycans and pattern formation: sugar biochemistry meets developmental genetics. Trends Genet 16, 206–212. 35 Tkachenko E, Rhodes JM & Simons M (2005) Synde- cans: new kids on the signaling block. Circ Res 96, 488– 500. 36 Kim CW, Goldberger OA, Gallo RL & Bernfield M (1994) Members of the syndecan family of heparan sul- fate proteoglycans are expressed in distinct cell-, tissue-, and development-specific patterns. Mol Biol Cell 5, 797–805. 37 Bernfield M & Sanderson RD (1990) Syndecan, a devel- opmentally regulated cell surface proteoglycan that binds extracellular matrix and growth factors. Phil Trans R Soc Lond B Biol Sci 327, 171–186. 38 Hoffman MP, Nomizu M, Roque E, Lee S, Jung DW, Yamada Y & Kleinman HK (1998) Laminin-1 and laminin-2 G-domain synthetic peptides bind syndecan-1 and are involved in acinar formation of a human G. Johnson et al. Acetylcholinesterase redundancy FEBS Journal 275 (2008) 5129–5138 ª 2008 The Authors Journal compilation ª 2008 FEBS 5137 submandibular gland cell line. J Biol Chem 273, 28633– 28641. 39 Bernfield M, Gotte M, Park PW, Reizer O, Fitzgerald ML, Lincecum J & Zako M (1999) Functions of cell surface heparan sulfate proteoglycans. Annu Rev Bio- chem 68, 729–777. 40 Stipp CS, Litwack ED & Lander AD (1994) Cerebro- glycan: an integral membrane heparan sulfate proteo- glycan that is unique to the developing nervous system and expressed specifically during neuronal differentia- tion. J Cell Biol 124, 149–160. 41 Schnepp A, Komp Lindgren P, Hulsmann H, Krogel S, Paulsson M & Hartmann U (2005) Mouse testican-2. Expression, glycosylation and effects on neurite out- growth. J Biol Chem 280, 11274–11280. 42 Farach-Carson MC & Carson DD (2007) Perlecan – a multifunctional extracellular proteoglycan scaffold. Glycobiology 17, 897–905. 43 Brown JC, Sasaki T, Gohring W, Yamada Y & Timpl R (1997) The C-terminal domain V of perlecan promotes beta1 integrin-mediated cell adhesion, binds heparin, nidogen and fibulin-2 and can be modi- fied by glycosaminoglycans. Eur J Biochem 250, 39–46. 44 Kimbell LM, Ohno K, Engel AG & Rotundo RL (2004) C-terminal and heparin-binding domains of collagenic tail subunit are both essential for anchoring acetylcholinesterase at the synapse. J Biol Chem 279, 10997–11005. 45 Anderson AA, Ushakov DS, Ferenczi MA, Mori R, Martin P & Saffell JL (2007) Morphoregulation by acetylcholinesterase in fibroblasts and astrocytes. J Cell Physiol 215, 82–100. 46 Denzer AJ, Schulthess T, Fauser C, Schumacher B, Kammerer RA, Engel J & Ruegg MA (1998) Electron microscopic structure of agrin and mapping of its bind- ing site in laminin-1. EMBO J 17, 335–343. 47 Marneros AG & Olsen BR (2005) Physiological role of collagen XVIII and endostatin. FASEB J 19, 716–728. 48 Chubykin AA, Liu X, Comoletti D, Tsigelny L, Taylor P & Sudhof TC (2005) Dissection of synapse induction by neuroligins: effect of a neuroligin mutation associ- ated with autism. J Biol Chem 280, 22365–22374. 49 Grifman M, Galyam N, Seidman S & Soreq H (1998) Functional redundancy of acetylcholinesterase and enuroligin in mammalian neuritogenesis. Proc Natl Acad Sci USA 95, 13935–13940. 50 Comoletti D, Flynn R, Jennings LL, Chubylin A, Mat- sumura T, Hasegawa H, Sudhof TC & Taylor P (2003) Characterisation of the interaction of a recombinant soluble neuroligin-1 with neurexin-1 beta. J Biol Chem 278, 50497–50505. 51 Graham DK, Bowman GW, Dawson TL, Stanford WL, Earp HS & Snodgrass HR (1995) Cloning and developmental expression analysis of the murine c-mer tyrosine kinase. Oncogene 10, 2349–2359. 52 Guttridge TD, Luft JC, Dawson TL, Koslowska E, Mahajan NP, Varnum B & Earp HS (2002) Mer receptor tyrosine kinase ignaling: prevention of apop- tosis and alteration of cytoskeletal architecture without stimulation of proliferation. J Biol Chem 277, 24057– 24066. 53 Sasaki T, Knysev PG, Cheburkin Y, Gohring W, Tisi D, Ulrich A, Timpl R & Hohenester E (2002) Crystal structure of a C-terminal fragment of growth arrest-spe- cific protein Gas6. Receptor tyrosine kinase activation by laminin G-like domains. J Biol Chem 277, 44164– 44170. 54 Noonan DM, Fulle A, Valente P, Cai S, Horigan E, Sasaki M, Yamada Y & Hassell JR (1991) The complete sequence of perlecan, a basement membrane heparan sulfate proteoglycan, reveals extensive similarity with laminin A chain, low density lipoprotein- receptor, and the neural cell adhesion molecule. J Biol Chem 266, 22939–22947. 55 Fuki IV, Kuhn KM, Lomasov IR, Rothman VL, Tus- zynski GP, Iozzo RV, Swenson TL, Fisher EA & Willi- mans KJ (1997) The syndecan family of proteoglycans: novel receptors mediating internalisation of atherogenic lipoproteins in vitro. J Clin Invest 100, 1611–1622. 56 Strittmatter WJ, Saunders AM, Schmechel D, Pencak- Vance M, Enghild J, Salveson GS & Roses AD (1993) Apolipoprotein E: high avidity binding to beta-amyloid and increased frequency of type 4 allele in late-onset familial Alzheimer’s disease. Proc Natl Acad Sci USA 90, 1977–1981. 57 Beffert U, Stolt PC & Herz J (2002) Functions of lipo- protein receptors in neurons. J Lipid Res 45, 403–409. 58 Datta G, Chaddha M, Garber DW, Chung B-H, Tytler EM, Dashti WA, Bradley SH, Gianturco SH & Anan- tharomaiah GM (2000) The receptor binding domain of apolipoprotein E, linked to a model class A amphi- pathic helix, enhances internalisation and degradation of LDL by fibroblasts. Biochemistry 39, 213–220. 59 Ethell IM, Hagihara K, Miura Y, Irie F & Yamaguchi Y (2000) Synbindin, a novel syndecan-2-binding protein in neuronal dendritic spines. J Cell Biol 151, 53–68. 60 Hienola A, Tumova S, Kulesskiy E & Rauvala M (2006) N-syndecan deficiency impairs neural migration in brain. J Cell Biol 174, 569–590. 61 Song HH, Shi W, Xiang YY & Filmus J (2005) The loss of glypican-3 induces changes in Wnt signaling. J Biol Chem 280, 2116–2125. Acetylcholinesterase redundancy G. Johnson et al. 5138 FEBS Journal 275 (2008) 5129–5138 ª 2008 The Authors Journal compilation ª 2008 FEBS . Non-enzymatic developmental functions of acetylcholinesterase – the question of redundancy Glynis Johnson, Chrisna Swart and Samuel W. Moore Divisions of Paediatric Surgery. by other members of the glypican family. The testicans are a subgroup of the BM-40 ⁄ SPARC ⁄ osteonectin family of modular proteins. There is no documented evidence of them binding laminin. They. virtue of their heparan sulfate chains. The lack of comprehensive information on the developmental expression and interactions of HSPGs makes it difficult to narrow the field of possibilities. The