In Xenopus oocytes, COS-7 cells, neuroblastoma cells, and muscle cells, the expression of PRiMA has been identified as a limiting factor in organizing G₄ AChE and targeting it to the cell membrane (Perrier et al., 2002; Xie et al., 2007), as well as in directing its membrane raft localization (Xie et al., 2010a). In the absence of PRiMA, the G₄ AChE, or the G₄ BChE, could not be formed. In the brain, the catalytic subunit contained in G₄ AChE is AChE_T (Inestrosa et al., 1994). The expression of PRiMA mRNA and protein are increased with an increment of G₄ AChE during the development of brain and spinal cord (Leung et al., 2009; Xie et al., 2010b). The PRiMA-linked AChE is also present at the nmj, in which both motor neuron and muscle are the major suppliers. In rat muscles, the protein expression of PRiMA and AChE_T, as well as the G₄ AChE, are dramatically increased during development (Leung et al., 2009). The parallel expression of PRiMA and G₄ AChE during the development strengthens the importance of PRiMA in directing, and/or regulating the formation of G₄ enzyme. The PRiMA-linked G₄ BChE however has not been well studied, and therefore the expression profile of which in brain or muscle are not fully revealed.

Asymmetric forms of AChE and BChE are characterized by the presence of collagen-like tail, which is formed by the triple helical association of three ColQ subunits (Feng et al., 1999). The cDNA encoding ColQ has been cloned in Torpedo (Krejci et al., 1991), rat (Krejci et al., 1997), human (Ohno et al., 1998), quail, and chicken (Ruiz and Rotundo, 2009a). The presence of the collagen-tailed forms of ChEs has been found in all classes of vertebrates, but not in invertebrates. They are specifically expressed in muscles and regulated by physiological activity (Sketelj and Brzin, 1985; Deprez et al., 2003; Lau et al., 2008). In human and rat, COLQ gene has two transcripts, ColQ-1 and ColQ-1a: they are differentiatedly expressed in slow-twitch (ColQ-1) and fast-twitch (ColQ-1a) muscles (Ohno et al., 1998; Krejci et al., 1999). The differentiated expression patterns may account for the synaptic and non-synaptic expression profile of ColQ-linked AChE in fast slow-twitch and fast-twitch muscles, respectively (Lee et al., 2004; Crne-Finderle et al., 2005; Choi et al., 2007).

Most of our knowledge concerning ChEs derives from the studies on the classical AChE and BChE homogenous oligomers. However, the sequence similarity between AChE_T and BChE_T has been further emphasized regarding the existence of hybrid A₁₂ forms in new-born chicken muscle (Tsim et al., 1988a), as well as a hybrid G₄ from in the brain (Chen et al., 2010) and embryonic muscles of chicken (Chen et al., 2009). In these hybrid enzymes, both AChE_T and BChE_T are attached to the same anchoring protein, ColQ for A₁₂ form and PRiMA for G₄ form (Figure 1). The hybrid enzyme exists as a single type, with equivalent number of AChE and BChE catalytic subunits in a single molecule. Interestingly, the expression of these hybrid molecules in chicken brain and muscle was found to be developmentally regulated. In 1-day-old chicken muscle, the predominant form of AChE is A₁₂ hybrid form; however, the proportion of BChE subunits in the hybrid molecules progressively disappear during the muscle development from embryonic, hatching to adult, and the homogeneous asymmetric AChE becomes the sole form upon the muscle maturation (Tsim et al., 1988b). On the other hand, a continuous increase of AChE–BChE G₄ hybrid expression was observed in chicken brain, while an obvious decrease was found in the leg muscles. To date, we do not have conclusive idea about the underlining regulatory mechanism of these hybrid enzymes. We believe that AChE–BChE hybrid enzymes could take part in cholinergic functions, which includes (i) they carry the catalytic activities of AChE and BChE in hydrolyzing ACh; (ii) they are associated with the anchoring protein, PRiMA or ColQ, which can anchor the hybrid enzymes onto the plasma membrane in the brain or nmj in the muscles; and (iii) the level of BChE_T indirectly regulates the expression of homogenous G₄ AChE.

In fact, the notion of having the existence of AChE–BChE hybrid enzyme is not new. An abnormal ChE species in the serum of patients suffering from carcinomas was reported (Zakut et al., 1988). This abnormal ChE species in human serum was inhibited by both AChE inhibitor BW284c51 and BChE inhibitor iso-OMPA. In addition, a collagen-tailed asymmetric hybrid AChE has been found relatively abundant in young chicken muscle (Tsim et al., 1988b), but which tends to disappear at the adult stage (Tsim et al., 1988b). Moreover, a hybrid tetramer having AChE and BChE activity was also found in cyst fluids derived from a human astrocytoma (García-Ayllón et al., 2001). However, none of the physiological function of these abnormal ChEs has been elucidated.

AChE_T and BChE_T have a catalytic domain of approximately 500 amino acids, followed by a C-terminal t-peptide of 40 and 41 residues, respectively. The t-peptide on AChE_T and BChE_T presents a considerable sequence similarity, with 24 identical residues, including seven aromatic residues and one cysteine near the C-terminus, which are conserved in human, cat, rabbit, mouse, cow, rat, chicken, and Torpedo (Massoulié et al., 1993).

The t-peptide was reported to play an important role in the biosynthesis of ChEs, particularly in the protein folding and exportation. The presence of aromatic residues in the t-peptide induces the misfolding of newly synthesized AChE_T polypeptides, and this effect depends on the hydrophobic character of these residues, because the same effect occurs when they are replaced by leucines (Falasca et al., 2005). In the absence of a proline-rich attachment domain (PRAD)-containing anchoring protein, the t-peptide enhanced a pool of AChE_T molecules toward endoplasmic reticulum (ER)-associated degradation (Belbeoc’h et al., 2003).

The major function of t-peptide is directing the assembly of tetramers of AChE_T (Bon and Massoulié, 1997) and BChE_T (Blong et al., 1997), as well as the association with the structure proteins, ColQ and PRiMA (Krejci et al., 1997; Perrier et al., 2002; Bon et al., 2004). The t-peptide is also named as tryptophan (W) amphiphilic tetramerization (WAT) domains, which contains a sector with seven aromatic residues that are strictly conserved between AChE_T and BChE_T. The association between AChE_T or BChE_T catalytic subunits and anchoring proteins, ColQ and PRiMA, is mostly based on the interaction between four WAT domains on the t-peptides and a PRAD on ColQ or PRiMA (Bon et al., 1997; Dvir et al., 2004; Noureddine et al., 2007). A crystallographic analysis of the complex of synthetic t-peptide and PRAD peptide indicated that four α-helical t-peptides form coiled-coil structure around the PRAD, which is arranged in a poly-proline II helix (Dvir et al., 2004). In addition, the formation of disulfide bonds through the cysteine residues near the end of the t-peptides stabilizes the quaternary association, and in fact this association appears to be critical in the case of AChE dimer formation. Dimers could be hardly observed after mutagenesis of the C-terminal cysteine residues in H or T-peptides of Torpedo and rat AChE (Morel et al., 2001; Chen et al., 2010).

The importance of t-peptide in the assembly of AChE_T and BChE_T was also reported by DNA mutagenesis studies. The truncated mutants, AChE_ΔT and BChE_ΔT, in which the t-peptides were deleted from the catalytic subunits, produced only monomers (Duval et al., 1992). AChE_T and AChE_BChE-T, BChE_T and BChE_AChE-T, in which the catalytic domain of each enzyme was swapped with the t-peptide of each other, presented similar assembly ability to form oligomers (Liang et al., 2009; Chen et al., 2010).

Although the nature of AChE_T and BChE_T oligomers depends on the presence of the t-peptides, the catalytic domains also influence the oligomerization patterns. The X-ray crystallography studies of Torpedo AChE dimers showed that the contact zone between two AChE_H subunits could be a “four-helix bundle” (FHB), formed by two α helices from each catalytic domain (Sussman et al., 1991). In addition, rat AChE_T was also demonstrated to dimerize through FHB inter-subunit contact zone (Morel et al., 2001). Based on these findings, we compared the predicted FHB sequences that are responsible for the dimeric contact zone of AChE_T and BChE_T from different species. Indeed, FHB domains are highly conserved across different species for either AChE or BChE, including human, mouse, rat, chicken, and Torpedo (Figure 3). On the other hand, the similarity between FHB domains of AChE_T and BChE_T is very low. An inter-species hybrid dimer could be formed between human AChE_T and chicken AChE_T, but not between mammalian AChE_T and BChE_T, in transfected HEK293T cells (Chen et al., 2010). The selectivity of dimerization seems to be based on the feature that the FHB domains of AChE are highly conserved among different species of vertebrates, but are distinguished from vertebrate BChE_Ts. Moreover, another hybrid dimer, between human AChE_BChE-T and chicken AChE_T, which contained the similar FHB but different t-peptides, was formed when they were co-expressed together in HEK293T cultured cells (Chen et al., 2010). This further confirmed that the catalytic domains, possibly the FHB domains, should play a critical role in the selection of subunits during the dimerization of ChEs.

According to our current knowledge, the oligomerization of AChE_T or BChE_T with their associated anchoring proteins could rely on three types of interactions (Figure 4): (i) the FHB interaction for the formation of dimer through hydrophobic interaction; (ii) intercatenary disulfide bonds between the t-peptide of AChE_T or BChE_T subunits; and (iii) tight hydrophobic interaction between the WAT domains of AChE_T or BChE_T subunits and the PRAD on PRiMA or ColQ.

AChE_T and BChE_T are well-known highly glycosylated enzymes, which carry various amounts of N-linked carbohydrate side chains attached to their core polypeptides (Liao et al., 1992; Kolarich et al., 2008). Mature human AChE_T monomer possesses three potential N-linked glycosylation sites (Soreq et al., 1990; Velan et al., 1993). Mature human BChE_T carries nine potential N-linked glycosylation sites (Lockridge et al., 1987). These N-linked glycosylation sites, both in AChE and BChE, are highly conserved in mammals, which implies the physiological importance of these glycans for ChEs.

Glycosylation is proposed to be used as a marker for the progression of ChE forms through different subcellular compartments, since it is known that the glycans added in ER are remodeled and matured in Golgi apparatus. In chicken muscles, the assembly of catalytically active dimer and tetramer occurred in the rough ER, with a subset of tetramers being further assembled with ColQ in Golgi apparatus into asymmetric forms (Rotundo, 1984). Once assembled, these catalytically active AChE oligomers were stable, acquired complex oligosaccharides in Golgi apparatus, and were transported to plasma membrane or secreted into medium (Rotundo, 1984). All of these exported AChE molecules contain complex oligosaccharides, because they could bind to lectins such as wheat germ agglutinin and ricin, and were endoglycosidase H (Endo H) resistant (Rotundo et al., 1989). In contrast, 70–80% of the newly synthesized AChE polypeptide chains in chicken muscle appeared to be catalytically inactive and Endo H sensitive, and they were degraded intracellularly with a half-life of about 1.5 h (Rotundo, 1988). Recently, Ruiz and Rotundo (2009a,b) reported that the expression of AChE in quail muscles is regulated by muscle activity through post-translational controls: the over-expression of ER molecular chaperons, such as calnexin, ER protein 72, and protein disulfide isomerase results in an increase of catalytic active ColQ-linked AChE in quail muscles.

The biological function of the glycans on ChEs was elucidated by site mutagenesis studies. Elimination of N-glycosylation sites did not interfere with the ability of AChE_T to form a soluble dimer (Velan et al., 1993), or to assembly with PRiMA to form a PRiMA-linked AChE tetramer (Chen et al., 2011). It appears therefore that the oligosaccharide side chains do not affect the structural elements that are responsible for the interaction of different subunits. Indeed, none of the N-glycosylation sites on AChE is close to or within the FHB domain implied in dimerization of AChE subunits (Sussman et al., 1991; Morel et al., 2001), or the t-peptide that allows the oligomerization of AChE_T with the anchoring proteins, e.g., ColQ and PRiMA (Bon et al., 2004). Moreover, the glycosylation of AChE_T can greatly affects the protein folding and membrane trafficking. When the glycosylation is eliminated, the folding of AChE_T fails, leading to a severe lost of the enzymatic activity (Chen et al., 2011). In the absence of glycosylation, the secreted G₁ and G₂ AChE are dramatically reduced in transfected cells, and the PRiMA-linked G₄ AChE is retained in ER and fails to be exported to plasma membrane (Chen et al., 2011).

The importance of N-glycosylation in the biosynthesis of AChE could explain the abnormality of glycosylation status in some pathological conditions. Proper glycosylation of AChE is important for normal brain function. Accumulation of molecular forms of AChE with altered patterns of glycosylation has been observed in the brain and cerebrospinal fluid of Alzheimer’s patients (Sáez-Valero et al., 1999, 2000). Moreover, characteristics of AChE found in the senile plaques are different from those in normal brain with a higher degree of glycosylation, which is proposed to be one of the factors facilitating formation of amyloid fibrils in the senile plaques (Mimori et al., 1997). These abnormalities in the glycosylation of AChE are very specific for Alzheimer’s disease and are not detected in other dementia illness, which suggests that glycosylation of AChE may have a diagnostic value.

To summarize the assembly and membrane processing of G₄ ChEs, a model is being proposed as Figure 5. After the mRNAs are translated into peptides, AChE_T and BChE_T polypeptides undergo initial glycosylation, presumably in a co-translation manner. Shortly after the synthesis, the glycosylated AChE_T and BChE_T are assembled into homodimers spontaneously. When these homodimers encounter PRiMA, they form PRiMA-linked G₄ AChE, or G₄ BChE, or AChE–BChE G₄ hybrid in ER. Afterward, PRiMA targets these G₄ complexes to Golgi apparatus where AChE_T and BChE_T subunits have further glycosylation. This trafficking allows the G₄ enzymes to be fully functional and finally become anchored onto the plasma membrane. During the whole process, the proper glycosylation of AChE_T is the key point for the membrane targeting. Without glycosylation, AChE_T polypeptides cannot fold properly, resulting in inactive AChE molecules. These un-glycosylated and inactive AChE_T molecules can still form PRiMA-linked G₄ AChE and AChE–BChE G₄ hybrid, but both of them are retained in ER, failing to be exported to Golgi apparatus, and finally they are subjected to degradation most possibly.