Collagen-tailed Forms Research Articles

Localization of “non-extractable” acetylcholinesterase to the vertebrate neuromuscular junction.

Asymmetric forms of acetylcholinesterase (AChE) are thought to be the predominant forms of this enzyme at vertebrate neuromuscular junctions where they attach to the synaptic basal lamina via a collagen-like tail. High salt and heparin-containing buffers are capable of solubilizing asymmetric AChE molecules from skeletal muscle; however, detachment of AChE specifically from synaptic basal lamina using these procedures has not been demonstrated. To determine whether AChE can be solubilized from mature neuromuscular junctions, adult quail muscle fibers were extracted with buffered detergent solutions containing either 0.05 M NaCl, 1 m NaCl, 0.5-2 mg/ml heparin, 8 M urea, or 4 m guanidine HCl, and the remaining AChE molecules were localized by indirect immunofluorescence. Analysis of extracted AChE oligomeric forms showed that low salt buffers containing heparin and high salt buffers were capable of solubilizing substantial amounts of catalytically active collagen-tailed AChE, whereas none of these buffers were capable of detaching AChE from synaptic basal lamina. In contrast, digestion with purified collagenase detached asymmetric forms from the non-extractable fraction and removed the AChE from the neuromuscular junctions. Parallel experiments using rat gastrocnemius muscle and enzyme histochemistry to detect AChE gave similar results. These studies indicate that the junctional AChE molecules are firmly attached to the extracellular matrix and that all the conventional extraction buffers used to solubilize the asymmetric collagen-tailed forms of AChE are incapable of detaching this enzyme from the synaptic basal lamina.

Ryanodine induces maturation of embryonic acetylcholinesterase forms in cultured quail myotubes

[3H]Ryanodine is shown to specifically bind to cultured myotubes from 10 day quail embryo pectoralis. The binding of [3H]ryanodine increases in a time-dependent manner reaching 38 +/- 3 fmol/mg protein at 4 h. A level of theophylline (THEO; 5mM) that induces propagated wave-like contractures, doubles the capacity of the myotubes to bind [3H]ryanodine (78 +/- 7 fmol/mg protein at 4 h). Polycationic ruthenium red (100 microM) only partially inhibits (56%) [3H]ryanodine-binding, whereas the membrane permeable channel antagonist [2,6-dichloro-4-dimethyl-amino-phenyl]-isopropylamine (20 microM) inhibits occupancy > 80%. Ryanodine (10 microM) interferes with THEO-induced contractures. Pretreatment with micromolar ryanodine for 48 h, followed by washout for 48 h, causes a persistent decrease in [3H]ryanodine-binding sites. Persistent [3H]ryanodine receptor blockade coincides with a dramatic shift in AChE forms found in the myotubes. A transition from the embryonic 4S and 7S globular forms to the 20S collagen-tailed (adult) form is evident within 12 hr exposure to ryanodine and progresses after removal of the alkaloid from the culture medium, mimicking the transition that normally occurs during myocyte maturation in vivo. These results suggest that SR Ca++ movements and excitation-contraction coupling may, at least in part, contribute to AChE maturation.

Cloning and expression of a rat acetylcholinesterase subunit: generation of multiple molecular forms and complementarity with a Torpedo collagenic subunit.

We obtained a cDNA clone encoding one type of catalytic subunit of acetylcholinesterase (AChE) from rat brain (T subunit). The coding sequence shows a high frequency of (G+C) at the third position of the codons (66%), as already noted for several AChEs, in contrast with mammalian butyrylcholinesterase. The predicted primary sequence of rat AChE presents only 11 amino acid differences, including one in the signal peptide, from that of the mouse T subunit. In particular, four alanines in the mouse sequence are replaced by serine or threonine. In northern blots, a rat AChE probe indicates the presence of major 3.2- and 2.4-kb mRNAs, expressed in the CNS as well as in some peripheral tissues, including muscle and spleen. In vivo, we found that the proportions of G1, G2, and G4 forms are highly variable in different brain areas. We did not observe any glycolipid-anchored G2 form, which would be derived from an H subunit. We expressed the cloned rat AChE in COS cells: The transfected cells produce principally an amphiphilic G1a form, together with amphiphilic G2a and G4a forms, and a nonamphiphilic G4na form. The amphiphilic G1a and G2a forms correspond to type II forms, which are predominant in muscle and brain of higher vertebrates. The cells also release G4na, G2a, and G1a in the culture medium. These experiments show that all the forms observed in the CNS in vivo may be obtained from the T subunit. By co-transfecting COS cells with the rat T subunit and the Torpedo collagenic subunit, we obtained chimeric collagen-tailed forms. This cross-species complementarity demonstrates that the interaction domains of the catalytic and structural subunits are highly conserved during evolution.

Molecular architecture of acetylcholinesterase collagen-tailed forms; construction of a glycolipid-tailed tetramer.

Asymmetric forms of Torpedo acetylcholinesterase (AChE) are produced in COS cells by the simultaneous expression of collagenic subunits (Q) and catalytic T subunits (AChET). Truncated AChET delta subunits, from which most of the C-terminal peptide (TC) had been deleted by mutagenesis, did not associate with Q subunits. The TC peptide is therefore necessary for the association of the AChET and Q subunits. In order to determine the orientation of the Q subunit in the collagen-tailed forms, we have developed an antiserum against its non-collagenic C-terminal domain, expressed as a fusion protein in Escherichia coli. This antiserum, which recognized the Q subunit in Western blots, was found to react with intact asymmetric forms, but not with collagenase-treated forms, from which the distal part of the tail had been cleaved, suggesting that the N-terminal non-collogenic domain (QN) is responsible for the interaction with the AChET subunits. This was confirmed by creating a chimeric subunit (QN/HC), in which QN was linked to the C-terminal peptide of the H subunit of Torpedo AChE, which contains the glycophosphatidylinositol (GPI) cleavage/attachment signal: co-expression of AChET and QN/NC produced GPI-anchored tetramers, which were sensitive to PI-PLC and largely exposed to the external surface of the cells. We thus demonstrate that: (i) the HC peptide is sufficient to determine the addition of a glycolipid anchor and (ii) the QN domain is sufficient to bind a catalytic AChET tetramer by interacting with the TC peptide.

Neural influence on the expression of acetylcholinesterase molecular forms in fast and slow rabbit skeletal muscles

With the aim of investigating the roles of motor innervation and activity on muscle characteristics, we studied the molecular forms of acetylcholinesterase (AChE) in fast-twitch (semimembranosus accessorius; SMa) and slow-twitch (semimembranosus proprius; SMp) muscles of the rabbit. We have shown that SMa and SMp express different patterns and tissue distribution of AChE forms and that the effect of long denervation varies with age. Three principal findings concerning expression of AChE molecular forms emerge from these studies. (1) The activity of AChE and the pattern of its molecular forms are particularly altered in adult denervated SMa and SMp muscles. AChE activity increases by 10-fold in both muscles, but asymmetric forms disappear in SMa and increase by 20-fold in SMp muscles. A similar alteration of AChE is found after tenotomy of these muscles, showing that the effect of denervation may be partly due to suppression of muscle activity. (2) The different changes occuring in the composition of AChE molecular forms in adult denervated SMa and SMp muscles are consistent with fluorescent staining with anti-AChE monoclonal antibodies and with DBA or VVA lectins, which bind to AChE asymmetric, collagen-tailed forms. These lectins poorly stain denervated SMa muscle surfaces but intensely stain neuromuscular junctions and extrasynaptic areas in denervated SMp muscle. (3) In contrast with the adult, denervation of 1-day-old muscles does not markedly modify the total amount of AChE or the proportions of its molecular forms, despite dramatic effects on muscle structure. These results are supported by studies of labeling with fluorescent DBA: the lectin only slightly stains the muscle fiber surface of denervated 15-day-old SMp muscle. Taken together, these data show that denervated muscles escape physiological regulation, producing increased levels of AChE with highly variable cellular distribution and patterns of molecular forms, depending on the age of operation and on the type of muscle.

Primary structure of a collagenic tail peptide of Torpedo acetylcholinesterase: co-expression with catalytic subunit induces the production of collagen-tailed forms in transfected cells.

The asymmetric forms of cholinesterases are synthesized only in differentiated muscular and neural cells of vertebrates. These complex oligomers are characterized by the presence of a collagen-like tail, associated with one, two or three tetramers of catalytic subunits. The collagenic tail is responsible for ionic interactions, explaining the insertion of these molecules in extracellular basal lamina, e.g. at neuromuscular endplates. We report the cloning of a collagenic subunit from Torpedo marmorata acetylcholinesterase (AChE). The predicted primary structure contains a putative signal peptide, a proline-rich domain, a collagenic domain, and a C-terminal domain composed of proline-rich and cysteine-rich regions. Several variants are generated by alternative splicing. Apart from the collagenic domain, the AChE tail subunit does not present any homology with previously known proteins. We show that co-expression of catalytic AChE subunits and collagenic subunits results in the production of asymmetric, collagen-tailed AChE forms in transfected COS cells. Thus, the assembly of these complex forms does not depend on a specific cellular processing, but rather on the expression of the collagenic subunits.

Cholinesterases from flounder muscle

Flounder (Platichthys flesus) muscle contains two types of cholinesterases, that differ in molecular form and in substrate specificity. Both enzymes were purified by affinity chromatography. About 8% of cholinesterase activity could be attributed to collagen-tailed asymmetric acetylcholinesterase sedimenting at 17S, 13S and 9S, which showed catalytic properties of a true acetylcholinesterase. 92% of cholinesterase activity corresponded to an amphiphilic dimeric enzyme sedimenting at 6S in the presence of Triton X-100. Treatment with phospholipase C yielded a hydrophilic form and uncovered an epitope called the cross-reacting determinant, which is found in the hydrophilic form of a number of glycosyl-phosphatidylinositol-anchored proteins. This enzyme showed catalytic properties intermediate to those of acetylcholinesterase and butyrylcholinesterase. It hydrolyzed acetylthiocholine, propionylthiocholine, butyrylthiocholine and benzoylthiocholine. The Km and the maximal velocity decreased with the length and hydrophobicity of the acyl chain. At high substrate concentrations the enzyme was inhibited. The p(IC50) values for BW284C51 and ethopropazine were between those found for acetylcholinesterase and butylcholinesterase. For purified detergent-soluble cholinesterase a specific activity of 8000 IU/mg protein, a turnover number of 2.8 x 10(7) h-1, and 1 active site/subunit were determined.

A synapse-specific carbohydrate at the neuromuscular junction: association with both acetylcholinesterase and a glycolipid

With the aim of investigating the roles of carbohydrates in synapse formation, we have characterized a synapse-specific saccharide at the vertebrate neuromuscular junction. Two lectins of similar specificity (Dolichos biflorus agglutinin, DBA, and Vicia villosa-B4 agglutinin, VVA-B4) stain synaptic but not extrasynaptic regions of the rat muscle fiber surface and thus define a synapse-specific carbohydrate. Using these and other probes, we show that the carbohydrate moiety concentrated at the neuromuscular junction resembles N-acetylgalactosamine (GalNAc) linked in the beta-anomeric form to the termini of oligosaccharides. VVA-B4 also selectively stains neuromuscular junctions in human, mouse, rabbit, guinea pig, chick, frog, axolotl, snake, fish, and lamprey muscles, a phylogenetic conservatism that suggests a synapse-related role for GalNAc beta-terminal saccharides. AChE from muscle binds to DBA- and VVA-B4-agarose, and is thereby identified as a glycoconjugate bearing the synapse-specific carbohydrate. Assay of AChE isoforms reveals that asymmetric, collagen-tailed forms, known to be highly concentrated at the rat neuromuscular junction, bind DBA and VVA-B4, while globular forms, which are more widely distributed, do not. A second class of GalNAc-bearing glycoconjugates is demonstrable immunohistochemically with monoclonal antibodies to stage-specific embryonic antigen (SSEA)-3 (Shevinsky et al., 1982) and GM2 (Natoli et al., 1986), which recognize GalNAc-bearing glycolipids. These antibodies selectively stain neuromuscular junctions, where they recognize glycolipid-like molecules that bind VVA-B4 but are distinguishable from AChE. The association of a synapse-specific carbohydrate with at least 2 different synapse-specific molecules raises the possibility that the former plays a role in determining a property that the latter share, such as concentration at the synapse.

Isolation and characterization of acetylcholinesterase from Drosophila.

The purification and characterization of acetylcholinesterase from heads of the fruit fly Drosophila are described. Sequential extraction procedures indicated that approximately 40% of the activity was soluble and 60% membrane-bound and that virtually none (less than 4%) corresponded to collagen-tailed forms. The membrane-bound enzyme was extracted with Triton X-100 and purified over 4000-fold by affinity chromatography on acridinium resin. Hydrodynamic analysis by both sucrose gradient centrifugation and chromatography on Sepharose CL-4B revealed an Mr of 165,000 similar to that observed for dimeric (G2) forms of the enzyme in mammalian tissues. In contrast, the purified enzyme gave predominant bands of about 100 kDa prior to disulfied reduction and 55 kDa after reduction on polyacrylamide gel electrophoresis in sodium dodecyl sulfate, values that are significantly lower than those reported for purified G2 enzymes from other species. However, the presence of a faint band at 70 kDa which could be labeled by [3H]diisopropyl fluorophosphate prior to denaturation suggested that the 55-kDa band as well as a 16-kDa species arose from proteolysis. This was confirmed by reductive radiomethylation and amine analysis of the 70-, 55-, and 16-kDa bands. All three contained ethanolamine and glucosamine residues that are characteristic of a C-terminal glycolipid anchor in other G2 acetylcholinesterases. The catalytic properties of the enzyme were examined by titration with a fluorogenic reagent which revealed a turnover number for acetylthiocholine that was 6-fold lower than eel and 3-fold lower than human erythrocyte acetylcholinesterase. Furthermore, the Drosophila enzyme hydrolyzed butyrylthiocholine much more efficiently than these eel or human enzymes, an indication that the fly head enzyme has a substrate specificity intermediate between mammalian acetylcholinesterases and butyrylcholinesterases.

Identical N-terminal peptide sequences of asymmetric forms and of low-salt-soluble and detergent-soluble amphiphilic dimers of Torpedo acetylcholinesterase : Comparison with bovine acetylcholinesterase

We have determined partial N-terminal sequences of acetylcholinesterase (AChE) catalytic subnunits from Torpedomarmorata electric organs and from bovine caudate nucleus. We obtain identical sequences (23 amino acids) for the soluble (‘low-salt-soluble’ or LSS fraction) and particulate (‘detergent-soluble’, or DS fraction) amphiphilic dimers (G 2 form) and for the asymmetric, collagen-tailed forms (‘high-salt-soluble’, or HSS fraction, A 12 + A 8 forms). There are two amino acid differences, at position 3 (Asp/His) and 20 (Ile/Val), with the sequences obtained for T. californica by MacPhee-Quigley et al. [(1985) J. Biol. Chem. 260, 12185-12189] for the soluble G 2 form and the lytic G 4 form which is derived from asymmetric AChE. The bovine sequence (12 amino acids) presents an identity of 4 amino acids (Glu-Leu-Leu-Val) with that of Torpedo, at positions 5–8 ( Torpedo) and 7–10 (bovine). There is also a clear homology with the sequence of human butyrylcholinesterase [(1986) Lockridge et al. J. Biol. Chem., in press] indicating that these enzymes probably derive from a common ancestor.

Serum regulation of acetylcholinesterase in cultured myotubes

A large (20S) collagen-tailed form of acetylcholinesterase associated with the neuromuscular junction appears in cultures of chick embryo muscle cells when horse serum is withdrawn from the medium. In this report, 10-day-old cultures were incubated 2 days in serum-free medium or in medium containing either horse, bovine, fetal calf, chicken, heat-treated horse or chicken serum, low (<100K) or high (>100K) molecular weight fractions of horse serum, or fibronectin. Total acetylcholinesterase activity and activity of the 20S form increased in medium without serum, with fetal calf serum and with the low-molecular-weight fraction of horse serum. The largest increase occurred with fibronectin. The results suggest that a factor(s) greater than 100K in adult sera inhibits total acetylcholinesterase production and formation of the 20S form of the enzyme.

Action of veratridine on acetylcholinesterase in cultures of rat muscle cells

Studies on cultures of embryonic rat muscle cells have suggested that the presence of collagen-tailed forms may be correlated with spontaneous contractile activity: these forms disappear in the presence of tetrodotoxin which blocks the sodium channels involved in the propagation of action potentials. The effect of veratridine, a drug which maintains the sodium channels in the open state, was studied. It is shown here that in young cultures veratridine provoked a dramatic increase in total acetylcholinesterase activity and changed the distribution of the molecular forms of the enzyme, increasing the proportion and absolute amount of the A 12 form. In order to elucidate the mechanism of action of this drug, the effects of various ions, ionophores, or other agents that modify the ionic permeabilities of membranes were also investigated.

Molecular forms of acetylcholinesterase in synaptic and extrasynaptic regions of avian tonic muscle

Molecular forms of acetylcholinesterase and pseudocholinesterase were analyzed directly in the microdissected individual endplates of a slow-tonic chicken muscle. The major form in the endplate is the L 2(6.5 S) form, while the collagen-tailed H 2c (20 S) form, normally considered to be the synaptic form, is a very minor component, in contrast to its predominance at the chicken fast-twitch fibre endplate. The same is true for pseudocholinesterase at these endplates. Outside the tonic fibre endplates the same forms occur as at the endplates, but at a very much lower concentration. The enzyme at the tonic fibre endplate cannot be attached to the basal lamina by a collagen tail, but appears to have a hydrophobic attachment. Acetylcholinesterase is functional at tonic fibre endplates, but the absence of the collagen-tailed form may account for the lower efficiency of the enzymic removal of acetylcholine there.

Two classes of collagen-tailed, asymmetric molecular forms of acetylcholinesterase in skeletal muscle: differential effects of denervation

Skeletal muscles of different vertebrate species contain, as it is the case in other cholinergic tissues, two classes of collagen-tailed, asymmetric forms (A-forms) of acetylcholinesterase (AChE). Class I A-forms are readily brought into solution in the presence of high salt, while class II A-forms do additionally require a chelating agent, such as EDTA, for solubilization. All A-forms aggregate at low ionic strength but only class II A-forms are reaggregated by excess Ca(++), even in the presence of 1M NaCl. This Ca(++)-mediated aggregability of class II A-forms is slowly lost upon exposure to detergents such as Triton X-100. Although these two classes of AChE tailed forms seem to be present in endplate and non-endplate areas, and in both the extra- and intracellular compartments, class II A-forms are predominantly extracellular and endplate-specific, at least in the rat diaphragm. On the other hand, well-characterized fast- and slow-twitch muscles show no preference for either class of asymmetric AChE species. Upon denervation, class I A-forms are degraded faster and disappear earlier than their class II counterparts, which are still easily detectable 17 days after nerve section. Class I and class II AChE molecular species exist in similar relative proportions in many vertebrate muscles. Thus, collagen-tailed forms may be altogether more abundant, in skeletal muscle, than it was hitherto realized. It is expected that this further example of AChE polymorphism will contribute to a better understanding of cholinergic transmission in skeletal muscle and, more specially, of nerve-muscle interactions.

Relationship of collagen-tailed acetylcholinesterase with basal lamina components. Absence of binding with laminin, fibronectin, and collagen types IV and V and lack of reactivity with different anti-collagen sera.

In view of their supposed localization in extracellular structures, such as basal lamina, we have investigated the possible interactions of collagen-tailed forms of acetylcholinesterase from Electrophorus and bovine superior cervical ganglion with matrix proteins: laminin, fibronectin and types IV and V collagens. Using binding and sedimentation assays, with iodinated or non-radioactive matrix proteins, we have not observed any significant interaction, in conditions of high or low ionic strength. We also examined whether the collagen tail of acetylcholinesterase asymmetric forms possessed an immunological relationship with known collagen types (I, III, IV, V) from mammalian sources. We found no specific immunoreactivity with any of the 32 sera studied, either with the iodinated Electrophorus or with the native bovine enzyme. We conclude from these negative results that the collagen-like tail of acetylcholinesterase is clearly distinct from the classical types of collagen and that asymmetric forms of the enzyme do not interact specifically with the matrix proteins studied. This does not exclude the possibility of specific interactions with other components, remaining to be identified.

Heparin solubilizes asymmetric acetylcholinesterase from rat neuromuscular junction

We are interested in the factors involved in the anchorage of acetylcholinesterase (AChE) to the synaptic basal lamina. Here, we report studies showing that heparin, a sulfated glycosaminoglycan, specifically solubilized AChE from endplate regions of rat diaphragm muscle. Of the several molecular forms of AChE present in that region, heparin only released the asymmetric A 12 and A 8 forms of the enzyme. Our results strongly support the involvement of heparin-like macromolecules in the in vivo immobilization of the collagen-tailed forms of AChE to the basal lamina of the neuromuscular junction.

Recovery of acetylcholinesterase forms in quail muscle cultures after intoxication with diisopropylfluorophosphate

Studies of recovery of acetylcholinesterase (AChE, EC 3.1.1.7) after inhibition by organophosphates (OPs) have been hampered by the low number of in vitro systems with large collagen-tailed forms of AChE characteristic of motor end plates. Pectoral muscle cultures from Japanese quail with high levels of a large 20S form of AChE were used to study recovery of cells from acute treatment with diisopropylfluorophosphate (DFP), an irreversible AChE inhibitor. Low molecular weight AChE forms were synthesized more rapidly than the large 20S form following a 15-min treatment with 10 −4 M DFP. Most of the activities of the small forms, but only part of the activity of the 20S form, disappeared within 48 hr after cycloheximide was added to block protein synthesis. To the contrary, virtually all the activity of the 20S AChE that was newly synthesized after DFP treatment was lost within 24 hr after cycloheximide treatment. The results were generally consistent with the idea that the 20S AChE form is assembled inside the cell near to its surface and then is released to bind to its outside.

Collagen-tailed and globular forms of acetylcholinesterase in the developing chick visual system

We have carried out a comparative study of the developmental profiles of the enzyme acetylcholinesterase, and of its collagen-tailed and globular structural forms, solubilized in the presence of 1 M NaCl, 1% (w/v) sodium cholate and 2 mM EDTA, in the chick retina and optic lobes. The overall acetylcholinesterase activities, both per mg protein and per embryo or chick, are substantially higher in tectum than in retina, from embryonic day 16. The A(12) collagen-tailed form of the enzyme is present in similar amounts in the embryonic retina and optic tectum; however, while the A(12) activity increases significantly in retina after birth, both by percentage and in absolute terms, the tectal tailed enzyme follows a declining developmental profile, reaching a minimum after 6 months of life. On the other hand, the globular G(4) species shows developmental profiles, both in retina and tectum, rather similar to those obtained for the overall enzyme activity, while the G(2) and G(1) forms are present in comparable concentrations in both tissues. Besides, G(4) is the predominant globular form in the chick optic lobe after hatching, G(2) and G(1) being enriched in the embryonic tectum. In the case of retina, however, all the globular forms contribute more evenly to the total acetylcholinesterase activity, along the developmental period considered. The potential significance of some of the postnatal developmental profiles is discussed in terms of the progressive adjustment of retina and tectum to the requirements of visual function.

Acetylcholinesterase forms in fast and slow rabbit muscle.

In the skeletal muscles of vertebrates, the level of acetyl-cholinesterase (AChE, EC 3.1.1.7.) may be either decreased (in rat) or increased (in chicken1 and rabbit2,3) after denervation. This enzyme is very polymorphic, however, presenting molecular forms which differ in their cellular localization and physiological regulation. The different AChE molecules may be classified as globular forms (monomers G1, dimers G2 and tetramers G4) and asymmetric or collagen-tailed forms (containing one, two or three tetramers: A4, A8 and A12)4. The distribution of the collagen-tailed and globular forms along the muscle fibres varies according to the animal species and physiological state: in normal adult rat muscle, the A12 form, which represents the major collagen-tailed form, is localized exclusively in the endplate region4–6, whereas in rat embryo7 and human muscle8 this form is distributed over the entire muscle fibre. The presence of collagen-tailed AChE has, however, been considered as an indicator of neuromuscular interactions because its appearance during embryogenesis coincides with the establishment of neuromuscular contacts6,9 and because, after denervation, the A12 form disappears from rat5,6 and chicken10,11 muscles. Here we report that this form in fact increases markedly in a slow twitch oxidative muscle of the rabbit after denervation. The regulation of asymmetric—and particularly A12—forms of AChE thus depends dramatically on the species and on the nature of skeletal muscle.

Molecular forms of acetylcholinesterase in developing Torpedo embryos

We have studied the evolution of acetylcholinesterase molecular forms during the embryonic development of Torpedo marmorata, in the electric organs and in the electric lobes of the central nervous system. In the early stages of development (35 mm embryos, ‘myogenic phase’ of electric organ development), globular forms of acetylcholinesterase (G 4 and G 2) are abundant in both tissues and the collagen-tailed form A 12 is already present. In the electric organs, this form accumulates rapidly after the 55–60 mm stage (‘electrogenic phase’), when synapse formation first commences. Although the molecular characteristics of the collagen-tailed forms, and particularly their aggregation properties, do not appear to change during development, their solubilization requires higher concentrations of MgCl 2, as the electrocytes mature, suggesting that they become more tightly integrated in a better organized basal lamina. The smaller collagen-tailed form A 8 shows a transient increase which coincides approximately with the maximal accumulation of A 12, suggesting that it is an intermediate in its synthesis. The accumulation of the hydrophobic G 2, which eventually becomes predominent in the adult electric organs, lags behind that of A 12. The functional significance of this important fraction of acetylcholinesterase is therefore not that of a pool of precursor for the synthesis of A 12. In the electric lobes, the tetrameric form (G 4) is abundant during development, as well as G 2 and G 1 at certain stages, but the A 12 form is predominant in the adult.