Muscular Dystrophy Alters the Processing ofLight Acetylcholinesterase but notButyrylcholinesterase Forms in Liver ofLama2dy Mice

J.L. Gomez,1,2 M.S. Garcıa-Ayllon,1 F.J. Campoy,1 and C.J. Vidal1*1Departamento de Bioquımica y Biologıa Molecular-A, Universidad de Murcia, Espinardo, Murcia, Spain2Departamento de Ciencias de la Salud, Universidad Autonoma Metropolitana, Unidad Iztapalapa, Mexico, DF,Mexico

In order to know whether the histopathological changesof liver, which accompany muscular dystrophy, affect thesynthesis of cholinesterases, the distribution and glyco-sylation of acetylcholinesterase (AChE) and butyrylcho-linesterase (BuChE) forms in normal (NL) and dystrophicLama2dy mouse liver (DL) were investigated. About half ofliver AChE, and 25% of BuChE were released with asaline buffer (fraction S1), and the rest with a saline-Brij96 buffer (S2). Abundant light (G2

A and G1A) AChE (87%)

and BuChE (93%) forms, and a few G4H and G4

A ChEspecies were identified in liver. The dystrophic syndromehad no effect on solubilization or composition of ChEforms. Most of the light AChE and BuChE species(.95%) were bound by octyl-Sepharose, while most lightAChE forms (80%), but not BuChE isoforms (15%), wereretained in phenyl-agarose. About half of the AChE dimerslost their amphiphilic anchor with phosphatidylinositol-specific phospholipase C (PIPLC), and the fraction ofPIPLC-resistant species increased in DL. AChE T and Rtranscripts were detected by reverse transcriptase-polymerase chain reaction (RT-PCR) of liver RNA. ChEcomponents of liver, erythrocyte, and plasma were dis-tinguished by their amphiphilic properties and interactionwith lectins. The dystrophic syndrome increased the livercontent of the light AChE forms with Lens culinaris ag-glutinin (LCA) reactivity. The abundance of ChE tetramersin plasma and their small amount in liver suggest thatafter their assembly in liver they are rapidly secreted,while the light species remain associated to hepaticmembranes. J. Neurosci. Res. 62:134–145, 2000.© 2000 Wiley-Liss, Inc.

Key words: congenital muscular dystrophy; cholinester-ase components; lectin binding; hydrophobic chroma-tography, GPI anchor

Muscular dystrophies form a heterogeneous group ofneuromuscular disorders, which are characterized by mus-cle weakness (for review see Roberts, 1995; Culligan etal., 1998; Emery, 1998; Ozawa et al., 1998). Laminin-2 is

the predominant muscle form of laminin, and consists ofa2 (merosin), b1, and g1 chains. About half the patientswith the classical form of congenital muscular dystrophy(CMD) have a deficiency of the laminin a2 chain (Tome,1999). In humans, the a2 gene maps at 6q2 (Naom et al.,1997), and mutations in this gene have been reported(Tome, 1999). The mouse a2 chain gene (Lama2) maps atchromosome 10, close to the dystrophia muscularis (dy)locus (Sunada et al., 1995). Murine dystrophia muscularisis an autosomal recessive disease leading to muscle degen-eration and developmental dysmyelination of peripheralnerves (Campbell, 1995). Homozygous dystrophin-positive Lama2dy mice (formerly dy mice) show a defect ofthe laminin a2 chain in skeletal muscle, heart, and periph-eral nerves (Matsumura et al., 1997).

Acetylcholinesterase (AChE, EC 3.1.1.7) and bu-tyrylcholinesterase (BuChE, EC 3.1.1.8) hydrolyze ace-tylcholine and other cholinesters. Accordingly, excit-able tissues, such as the central nervous system (CNS),peripheral nervous system (PNS), and striated musclepossess ChEs. However, the enzymes also exist in non-excitable tissues, such as blood cells (Marcos et al.,1998; Chan et al., 1998), plasma (Altamirano and Lock-ridge, 1999; Garcıa-Ayllon et al., 1999), spinal menin-ges (Ummenhofer et al., 1998), and liver (Perelman etal., 1990). The occurrence of ChEs in many tissuessuggests that they may be involved in functions whichare unrelated to their catalytic activities (Layer andWillbold, 1995; Chan et al., 1998; Sternfeld et al.,1998; Brimijoin and Koenigsberger, 1999; Johnson andMoore, 1999; Weitnauer et al., 1999).

Contract grant sponsor: Fondo de Investigacion Sanitaria de la SeguridadSocial; Contract grant number: 98/0442; Contract grant sponsor: Funda-cion Seneca de la Comunidad Autonoma de Murcia; Contract grant num-ber: 230-CV-97.

*Correspondence to: C.J. Vidal, Departamento de Bioquımica y BiologıaMolecular-A, Edificio de Veterinaria, Universidad de Murcia, Apdo. 4021,E-30071 Espinardo, Murcia, Spain. E-mail: cevidal@um.es

Received 1 May 2000; Revised 26 June 2000; Accepted 28 June 2000

Journal of Neuroscience Research 62:134–145 (2000)

© 2000 Wiley-Liss, Inc.

AChE and BuChE differ in their preferred substrateand sensitivity to inhibitors (Massoulie et al., 1993). Bothenzymes occur as polymers of catalytic subunits, and theyare classified into asymmetric (A12, A8, A4) and globular(G4 , G2, G1) forms (Chatonnet and Lockridge, 1989;Schwarz et al., 1995). AChE and BuChE exist asmembrane-bound (amphiphilic) or soluble (hydrophilic)globular forms in neural and non-neural tissues (Massoulieet al., 1993; Saez-Valero and Vidal, 1996).

Three AChE mARNs are produced by alternativesplicing (Schwarz et al., 1995; Grisaru et al., 1999). Theprincipal transcript in brain and muscle is called AChE-T (for“tailed”) and consists of E1-E2-E3-E4-E6, yielding hydro-philic or amphiphilic globular and asymmetric forms. Thetranscript AChE-H (for “hydrophobic”) contains E1-E2-E3-E4-E5-E6 and generates a polypeptide, which, upon specificcleavage, adds glycosylphosphatidylinositol (GPI). The tran-script AChE-R (for “readthrough”) consists of E1-E2-E3-E4-I4-E5-E6, and is expressed in embryonic and tumor cells.A single BuChE transcript has been identified (Jbilo et al.,1994). In addition, ChEs are glycoproteins, and the mousecDNAs encoding AChE or BuChE subunits contain three orseven N-glycosylation sites (Rachinsky et al., 1990).

The skeletal muscle of the Lama2dy mouse shows adeficit of G4 AChE forms (Gisiger and Stephens, 1988;Cabezas-Herrera et al., 1994a), a distinct reactivity of G4AChE and BuChE forms with the lectin RCA (Cabezas-Herrera et al., 1994a, 1997), and a higher level of BuChEin internal muscle membranes (Moral-Naranjo et al.,1999). These changes do not occur in dystrophic mouseheart (Gomez et al., 1999), brain (Moral-Naranjo et al.,1996), or plasma (Cabezas-Herrera et al., 1994a).

As regards to liver, histopathological changes such ascellular swelling and necrosis have been observed in dys-trophic cats (Carpenter et al., 1989). The liver pathologyis produced by the congestion and hypoxia caused byimpaired cardiac function. In our search for a possiblerelationship between liver pathology and ChE synthesis,the properties of enzyme forms in liver of normal ordystrophic mice were investigated. In addition, the liverorigin of plasma ChEs was tested by analyzing structuralhomologies and differences between the enzymes in thetwo sources.

MATERIALS AND METHODS

Materials

Acetylthiocholine iodide, butyrylthiocholine iodide, 5,5’-dithio-bis-2-nitrobenzoic acid (DTNB), 1,5-bis(4-allyldimethyl-ammoniumphenyl)-pentan-3-one dibromide (BW284c51), tetrai-sopropyl pyrophosphoramide (Iso-OMPA), N-[2-Hydroxyethyl]-piperazine-N9-[2-ethanesulfonic acid] (HEPES), proteinase inhib-itors, polyoxyethylene-10-oleyl ether (Brij 96), enzyme markers forsedimentation analysis (catalase and alkaline phosphatase), phenyl-agarose, octyl-Sepharose, Sepharose 4B, agarose-bound lectins(concanavalin A, Con A; Lens culinaris agglutinin, LCA; wheatgerm (Triticum vulgaris) agglutinin, WGA; Ricinus communis agglu-tinin, RCA120), MMLV retrotranscriptase, and Taq polymer-ase were all from Sigma Chemicals Co. (St. Louis, MO).

Phosphatidylinositol-specific phospholipase C (PIPLC) from Bacil-lus thuringiensis was from Glyko (Novato, CA). Trizol and poly-merase chain reaction (PCR) primers were from Gibco (GrandIsland, NY). Human placenta ribonuclease inhibitor (HPRI) wasfrom Amersham (Arlington Heights, IL), and dNTPs were fromRoche Diagnostics (Barcelona, Spain).

Animals and Tissue Preparation

Phenotypically normal 129B6F1/J (1/?) and dystrophicLama2dy mice (dystrophin-positive, merosin-negative) werepurchased from Jackson Memorial Laboratories (Bar Harbor,ME), bred and kept at the University of Murcia animal care unitaccording to the ethical animal care guidelines. Animals, 4–5months old, were ether-anesthetized before opening the tho-racic cavity. The blood of heart and liver was removed byextensive perfusion of heart with 5.4 mM EDTA, 155 mMNaCl, pH 7.4, for 10 min. The mice were then beheaded, andthe liver was removed, washed with 1 M NaCl, 50 mM MgCl2,15 mM HEPES, pH 7.0 (HEPES-saline buffer, HSB), rinsed,blotted, and weighed before homogenization.

AChE and BuChE Solubilization and Assay

Liver (1 g) from normal (NL) or dystrophic (DL) mice washomogenized (10% w/v) with HSB, to which a fresh mixture ofproteinase inhibitors (Moral-Naranjo et al., 1996) had beenadded. After centrifugation at 100,000 3 g for 1 hr at 4°C, thefraction rich in soluble or loosely bound ChEs (S1) was saved.Then, the pellet was extracted again with HSB supplementedwith 1% w/v Brij 96 and antiproteinases. After centrifugation asabove, the membrane-bound ChEs were released in the super-natant S2. Inhibition of BuChE by the mixture of salt and TritonX-100 (Moral-Naranjo et al., 1996) was prevented by using Brij96 instead of Triton X-100 to release membrane-bound ChEs.

For experiments with ChEs from erythrocyte and plasma,blood was taken from ether-anesthetized mice by heart punc-ture. Blood clotting was prevented with 1 mg/ml EDTA(Garcıa-Ayllon et al., 1999). Plasma was separated from bloodcells by centrifugation and saved for further use. Blood cellswere extensively washed with 154 mM NaCl, 5.4 mM EDTAin 5 mM phosphate buffer, pH 7.5, and white cells wereremoved to avoid the release of proteases. Erythrocyte mem-branes were lysed by incubation with 5 mM phosphate pH 7.5,and membrane-bound AChE was released with 1% w/v TritonX-100 in HSB.

ChE activity in the homogenates and supernatants wasmeasured by the Ellman method, as described earlier (Moya-Quiles et al., 1992). AChE was assayed with 1 mM acetylthio-choline (AcTCh) and 50 mM Iso-OMPA, while BuChE wasmeasured with 1 mM butyrylthiocholine (BuTCh) and 10 mMBW284c51; Iso-OMPA and BW284c51 are specific inhibitorsof BuChE and AChE, respectively (Massoulie et al., 1993).

Nonspecific esterase activity in liver extracts was routinelymeasured by adding both BW284c51 and Iso-OMPA to theassays. Although it accounted only for 5–10% of ChE activity,the amount of substrate (AcTCh or BuTCh) hydrolyzed byesterases was always discounted for estimating true AChE orBuChE activity.

One unit (U) of ChE activity is the amount of enzymewhich hydrolyzes one mmol of the substrate per hr at 37°C.

Changes of ChEs in Dystrophic Mouse Liver 135

AChE and BuChE activities in the fractions collected fromsucrose gradients were determined by a microtiter assay (Moya-Quiles et al., 1992), ChE activity being expressed in arbitraryunits, one unit of activity referring to an increase of 0.001absorbance units per microliter of sample, and per minute.Protein content was measured by a modified Lowry assay (Dul-ley and Grieve, 1975).

Sedimentation Analysis

Molecular forms of AChE and BuChE in extracts S1 or S2were separated by velocity sedimentation, and identified accord-ing to their sedimentation coefficients. Samples (0.3–0.5 ml),with added sedimentation markers, were centrifuged on 5–20%(w/v) linear sucrose gradients made in 1 M NaCl, 50 mMMgCl2, and 10 mM Tris, pH 7.0, with Triton X-100 (0.5%w/v) or Brij 96 (0.5% w/v; Moral-Naranjo et al., 1996). Thetubes with gradients were centrifuged at 150,000 3 g in a SW41Ti Beckman rotor, for 18 hr at 4°C. About 40 fractions werecollected and assayed for AChE, BuChE, and the sedimentationmarkers (catalase, 11.4S, and alkaline phosphatase, 6.1S). Sedi-mentation coefficients for the ChE forms were calculated asreported before (Cabezas-Herrera et al., 1994b). After sevenruns with the supernatants S1 and S2 from NL, and six from DL,mean values and standard deviation for sedimentation coeffi-cients of ChE species were obtained.

Separation of Hydrophilic and AmphiphilicCholinesterase Components

The amphiphilic behavior of AChE and BuChE formswas established by chromatography on phenyl-agarose andoctyl-Sepharose. Each gel (5 ml) was poured into a glass columnand washed with HSB. The extracts S1 and S2 (1 ml each) weremixed and passed through the gels. Fractions with nonretainedChE activity, consisting of hydrophilic or weakly amphiphilicforms, were collected, while bound ChE activity was elutedwith Triton X-100 (2% w/v) in 15 mM HEPES, pH 7.0.Hydrophilic or amphiphilic ChE forms were characterized bytheir sedimentation coefficients.

Removal of Amphiphilic Domains in AChE Forms byIncubation With PIPLC

The existence of glycosylphosphatidylinositol (GPI) resi-dues in ChE forms was tested by incubating a mixture ofsupernatants S1 and S2 (0.5 ml each) without (control), and withPIPLC (3 U/ml), 2 hr at 37°C. The PIPLC-converted hydro-philic ChE forms were identified by sedimentation analysis.

RNA Isolation and Analysis of AChE Transcripts

AChE mRNAs in mouse liver and other tissues wereidentified by RNA isolation followed by reverse transcriptase-polymerase chain reaction (RT-PCR) with selected primers.Total RNA was isolated with Trizol, according to the manu-facturer’s instructions. The RNA content was measured byOD260, and the yield for liver was about 8 mg RNA/mg.Primers were designed to identify the various AChE transcriptsby RT-PCR, according to the mouse AChE gene sequence(Rachinsky et al., 1990; Li et al., 1991). The selected primerswere: primer P1 (in exon E3), 5’-CACGCAGGAGAG-GATCTTTG-3’; primer P10 (in exon E5), 5’-ACGGAACAG-

GTCGGGTAGTG-3’; and primer P4 (in exon E6), 5’-GAGCTTAGCCCAAGACATGC-3’.

After mRNA processing, all the exons downstream of thetranscript-specific sequence are maintained. Following the in-variable exon E4, transcripts R contain I4, E5, and E6; tran-scripts H contain E5 and E6, and transcripts T contain only E6.Therefore, all three transcripts contain the exon 6, and theircDNAs can all be obtained with the primer P4. In the followingPCR reactions, two pairs of primers were used, each pairconsisting of a forward primer in the invariable exon E3 (P1,common to the two pairs), plus a reverse primer, either P10 (inE5) or P4 (in E6). Both pairs of primers can amplify the cDNAsfor several transcripts, but the RT-PCR products can be iden-tified according to their sizes. Thus, primers P1 1 P4 canamplify the cDNA for the T transcripts (giving fragments with524 bp), but also transcripts H (1,188 bp) and R (1,304 bp); pairP1 1 P10 amplifies transcripts H (864 bp) and R (with a 980 bpRT-PCR product). The use of a forward primer in exon E3would help identify amplified fragments due to immature RNAor contaminant DNA, which would be much longer because ofintron I3 (of about 1.2 Kb).

First-strand cDNAs were obtained by reverse transcrip-tion in a final volume of 20 ml. The reverse primer (P4, 2 pmol)and the RNA (5 mg) were mixed, heated at 70°C for 10 min andcooled on ice. A mixture containing buffer, DTT, dNTPs, andhuman placenta ribonuclease inhibitor (HPRI, 20 units) wasadded, and all was incubated at 37°C for 2 min. Moloneymurine leukemia virus (MMLV) reverse transcriptase (5 units)was added, and the reactions were carried out at 37°C for60 min. Samples were then heated at 85°C for 10 min and keptfrozen. The cDNA was amplified by PCR, using the commonforward primer P1 plus P4 or P10 as reverse primer. Aliquots(1/15th) of the cDNA samples were incubated with Taq poly-merase (2.5 U) in the presence of primers (15 pmol each) anddNTPs, in a total volume of 50 ml. PCR reactions were madein a PTC-150 Minicycler (MJ Research Inc., Watertown, MA),starting with 3 min at 95°C, followed by 35 cycles of 1 min at94°C, 1 min at 56°C, and 70 seconds at 72°C; with a final 72°Cstep of 10 min. PCR products were separated in 1.5% agarosegels, using a Tris-acetate-EDTA buffer, and visualized withethidium bromide.

Binding of Cholinesterase Formsto Immobilized Lectins

Possible differences in sugar residues of ChE forms innormal and dystrophic liver were investigated by their interac-tion with lectins (Cabezas-Herrera et al., 1994a). For lectinbinding, the extracts S1 and S2 (0.5 ml each) were mixed andincubated with a lectin-free Sepharose 4B (control), or withCon A-, LCA-, WGA- or RCA-agarose. After overnight in-cubation at 4°C, ChE-lectin complexes were separated by cen-trifugation, and the unbound ChE forms identified by theirsedimentation coefficients.

Statistical Analyses

The values of AChE and BuChE activity in membranesuspensions and soluble fractions of mouse liver, sedimentationcoeficients of AChE and BuChE forms, and percentages ofinteraction between lectins and ChE forms are presented as

136 Gomez et al.

means 6 S.D. Statistical significance between data obtainedfrom normal and dystrophic mouse liver was assessed by appli-cation of the nonparametric Mann-Whitney test.

RESULTSSolubilization of CholinesterasesFrom Mouse Liver

AChE and BuChE activities in membrane suspen-sions and soluble fractions of NL and DL are given inTable I. Mouse liver was very rich in BuChE, its activityin membrane suspensions (672 U/g) being more than100-fold that of AChE activity (4.5 U/g). BuChE activitywas much higher in mouse liver than in brain (17 U/g;Moral-Naranjo et al., 1996), muscle (9 U/g; Cabezas-Herrera et al., 1994a,b), or heart (21 U/g; Gomez et al.,1999). Conversely, AChE was lower in liver than in brain(383 U/g), skeletal muscle (48 U/g), or heart (10 U/g).

The percentages of ChEs in S1 and S2, and therelative proportions of AChE and BuChE forms in mouse

liver are shown in Table II. About half of AChE and25–30% of BuChE were released with a saline buffer (S1),and the rest with detergent (S2). Sedimentation analysesrevealed a similar composition of AChE and BuChE formsin the supernatants S1 and S2. No significant differenceswere observed regarding ChE activity in NL or DL or theextent of enzyme solubilization (Tables I and II).

AChE Components in Mouse LiverSedimentation analyses of salt-soluble (S1) and

detergent-soluble (S2) ChEs revealed abundant 3.3 6 0.5S(Svedberg units) and 4.6 6 0.3S AChE forms, along withlesser 8.4 6 0.2S and 10.7 6 0.2S components, providingthat Brij 96 was added to the sucrose gradients (Fig. 1). Ingradients with Triton X-100, the above AChE formsmigrated at 4.0 6 0.2S, 5.2 6 0.3S, 9.2 6 0.4S, and10.6 6 0.4S (profile not shown). The enzyme specieswere assigned to amphiphilic G1

A, G2A and G4

A, and hydro-philic G4

H AChE according to: (1) their sedimentation

TABLE I. Solubilization of AChE and BuChE From Normal (NL) and Dystrophic (DL) Mice Liver*

AChE activity BuChE activity Protein content

U/ml U/mg protein U/g wet tissue U/ml U/mg protein U/g wet tissue mg/ml mg/g wet tissue

Normal liverH 0.38 6 0.05 0.025 6 0.007 4.5 6 1.1 67 6 18 4.5 6 0.8 672 6 176 15.6 6 3.2 159 6 34S1 0.18 6 0.05 0.020 6 0.007 3.0 6 1.3 20 6 6 2.0 6 0.4 208 6 60 11.1 6 2.9 116 6 31S2 0.22 6 0.07 0.033 6 0.011 2.8 6 0.9 50 6 17 8.2 6 2.0 549 6 164 6.4 6 0.8 70 6 7

Dystrophic liverH 0.51 6 0.11 0.035 6 0.007 5.6 6 1.8 59 6 8.5 3.8 6 1.3 538 6 124 14.4 6 2.0 164 6 46S1 0.20 6 0.05 0.020 6 0.005 2.3 6 0.6 21 6 6.0 3.6 6 2.2 147 6 59 10.0 6 1.3 113 6 24S2 0.27 6 0.04 0.053 6 0.015 3.2 6 0.6 40 6 8.3 8.6 6 1.4 451 6 88 5.3 6 1.1 62 6 16

*The hearts of ether-anesthetized mice were perfused as described in Materials and Methods. One liver from a normal (NL) or dystrophic (DL) mousewas homogenized in HEPES-saline buffer (HSB), containing proteinase inhibitors. After centrifugation, the supernatant S1 was recovered. The pellet wasextracted with HSB plus 1% w/v Brij 96 and antiproteinases to obtain the supernant S2. The S1 and S2 fractions were rich in soluble and weakly bound,and firmly bound cholinesterases (ChEs), respectively. Aliquots of membrane suspensions and supernatants were assayed for ChE activity and proteincontent. Acetylcholinesterase (AChE) and butyrylcholinesterase (BuChE) activities are given in mmol of acetyl- or butyrylthiocholine hydrolyzed per hr(U). Results are mean 6 S.D. from seven and six preparations made with NL and DL, respectively.

TABLE II. Molecular Forms of AChE and BuChE in Liver of Normal and Dystrophic Mice*

% AChE forms % BuChE forms

G4H G4

A G2A 1 G1

A G4H G4

A G2A 1 G1

A

Normal liverSaline extract (S1) 13 (7) 5 (3) 82 (43) 8 (2) 2 (,1) 90 (24)Brij 96 extract (S2) 5 (2) 3 (1) 92 (44) 5 (4) 1 (,1) 94 (69)Total (9) (4) (87) (6) (1) (93)

Dystrophic LiverSaline extract (S1) 19 (8) 2 (1) 79 (33) 8 (2) 1 (,1) 91 (23)Brij 96 extract (S2) 9 (5) 2 (1) 89 (52) 7 (5) 1 (,1) 92 (69)Total (13) (2) (85) (7) (1) (92)

*Percentages of cholinesterase (ChE) solubilization refer to the activity recovered in the supernatants S1 plus S2. These were as follows: acetylcholinesterase(AChE) in S1 and S2, 52% and 48% for liver of control mice, and 42% and 58% for liver of dystrophic mice; butyrylcholinesterase (BuChE) in S1 and S2,27% and 73% for normal mice, and 25% and 75% for dystrophic mice. Percentages of the various ChE forms considering the solubilization yield in S1 andS2 are given in parentheses. Relative content of enzyme forms and their hydrophilic (H) or amphiphilic (A) behavior reflect the results of sedimentationanalysis in gradients with Brij 96 or Triton X-100. Percentages of ChE forms in successive experiments remained fairly constant; the deviations were within10% of the mean values shown.

Changes of ChEs in Dystrophic Mouse Liver 137

values (Moral-Naranjo et al., 1996; Gomez et al., 1999);(2) the difference in their migration with the detergentadded to the gradient; and (3) the adsorption to hydro-phobic matrices (see later). As a whole, the above findings

show that mouse liver is very rich in G1A and G2

A AChE,which, together, accounted for nearly 90% of the enzymeactivity (Table II), and that it also contains a small amountof G4

H AChE and G4A forms. Although AChE dimers and

monomers are hardly resolved by sedimentation analysis, ahigher proportion of dimers than monomers was consis-tently observed in mouse liver (Figs. 1, 2). No significantdifferences in the composition of AChE forms in NL orDL were observed.

Characterization of BuChE FormsSedimentation profiles of the soluble S1 and S2 frac-

tions obtained from mouse liver revealed principal 3.1 60.3S, and 4.7 6 0.5S BuChE components, and minor9.16 0.2S, and 11.4 6 0.5S enzyme species, when Brij 96was added to the sucrose gradients (Fig. 1). In gradientswith Triton X-100, the BuChE forms sedimented at4.1 6 0.2S, 5.5 6 0.3S, 9.7 6 0.1S, and 11.5 6 0.5S(profile not shown). Although the light (G1

A and G2A)

AChE and BuChE species migrate to nearby positions, theuse of the specific BuChE inhibitor and the reduction ofactivity in sucrose gradients with Triton X-100 supportthe assignment of the peaks to BuChE species. Accordingto their sedimentation values (Moral-Naranjo et al., 1996;Gomez et al., 1999), the different migration patterns ac-cording to the detergent added, and their retention inoctyl-Sepharose (see later), the peaks of BuChE activitywere attributed to G1

A, G2A, G4

A, or G4H enzyme forms.

The results revealed that, as in the case of AChE species,G1

A and G2A BuChE forms accounted for 90% of the

Fig. 1. Identification of cholinesterase forms in normal mouse liver.The tissue was extracted with HEPES-saline buffer (HSB) and antipro-teinases, and after centrifugation, the soluble and loosely bound cho-linesterases (ChEs) were recovered in the supernatant S1. The pellet wasreextracted with HSB containing antiproteinases and 1% w/v Brij 96and, after centrifugation, the tightly bound cholinesterases were savedin S2. ChE forms in S1 and S2 were analyzed by centrifugation on5–20% linear sucrose gradients containing 0.5% Brij 96. ChE activity isexpressed in arbitrary units, one unit referring to an increase of 0.001units of absorbance per ml, and per min. Sedimentation coefficients ofChE forms were calculated by reference to the internal sedimentationmarkers (arrows) catalase (C, 11.4S) and alkaline phosphatase (P, 6.1S).Peaks of enzyme activity were assigned to individual ChE forms ac-cording to previous data (Moral-Naranjo et al., 1996; Gomez et al.,1999). Note the much higher content of butyrylcholinesterase(BuChE) activity, the small contribution of acetylcholinesterase(AChE) or BuChE tetramers (insets), and the higher proportions of G2

A

AChE and G1A BuChE species in the extracts. A comparison of the

sedimentation patterns obtained with samples of normal and dystrophicmouse liver showed the absence of significant differences in the distri-bution of ChE forms.

Fig. 2. Detachment of the amphiphilic anchor in AChE forms withphosphatidylinositol-specific phospholipase C (PIPLC). After separa-tion of ChE species from normal or dystrophic mouse liver by sedi-mentation analysis, the fractions rich in G2

A and G1A AChE were pooled,

dialyzed in 10 mM Tris, pH 7.4, incubated without (control, opencircles) or with PIPLC (3 U/ml) for 2 hr at 37°C (filled circles), andfinally analyzed on sucrose gradients with Brij 96. Internal sedimenta-tion markers as in Figure 1. Note the decreased content of G2

A and G1A

forms, and the appearance of a new peak at 6.1S (Svedberg units) insamples of normal liver incubated with PIPLC. The profiles on theright show that amphiphilic AChE dimers and monomers of dystrophicmouse liver are more resistant to PIPLC than the corresponding formsof control mice.

138 Gomez et al.

enzyme activity in mouse liver, the remaining fractionbeing accounted for by G4

A and G4H BuChE species

(Table II). Although BuChE dimers and monomerscoexist in liver extracts, the latter are more abundant (Fig.1), revealing that the homologous AChE and BuChEforms are differently distributed in mouse liver. As for theAChE forms, no significant differences in the proportionof BuChE species in NL or DL were observed.

Amphiphilic or Hydrophilic Cholinesterase Formsin Mouse Liver

The differential adsorption of ChEs in hydrophobicgels such as phenyl-agarose and octyl-Sepharose providesvaluable information regarding the nature of the am-phiphilic domain in AChE and BuChE species (Gomez etal., 1999; Garcıa-Ayllon et al., 1999). A high fraction(78%) of AChE activity in a mixture of the S1 and S2extracts, and a lower amount (17%) of BuChE activity,were retained in phenyl-agarose (Table III). Sedimenta-tion analyses showed that, despite the common capacity ofthe G1

A and G2A AChE or BuChE components to interact

with detergent, most of the amphiphilic AChE forms, butonly a few BuChE species, were bound to phenyl-agarose.The experiments with octyl-Sepharose revealed thatG1

A 1 G2A AChE or BuChE forms were fully retained

(.95%) in the matrix, and that G4H AChE or BuChE

species failed to bind to the gel.The results confirmed that the ChE components

identified as G1A, G2

A, and G4A are truly amphiphilic, and

G4H, hydrophilic. However, the fact that the light AChE

forms were bound to phenyl- or octyl-Sepharose, and that

the BuChE isoforms were adsorbed to octyl-, but not tophenyl-agarose, clearly indicated that the light AChEforms differed from the homologous BuChE species intheir amphiphilic domains.

For assessing whether hepatic AChE dimers andmonomers consist of H subunits (type I molecules, withlinked GPI) or T subunits (type II, free of GPI), the peakfractions rich in G1

A1G2A AChE obtained after various

runs of gradient centrifugation were pooled, dialyzed, andincubated with PIPLC. The mixture was then analyzed ingradients with Brij 96. The sedimentation patterns (Fig. 2)showed a new peak at 6.1S (G2

H; Saez-Valero and Vidal,1995), and a decrease of those at 4.6S (G2

A) and 3.6S (G1A),

demonstrating that some amphiphilic AChE dimers andmonomers were converted into their hydrophilic variantsby cleavage of the GPI anchor with PIPLC. Unfortu-nately, the sedimentation coefficients of G2

A and G1H

AChE coincide (Legay et al., 1993), and this prevents theextent of the G1

A conversion into G1H from being calcu-

lated. Nevertheless, from the difference of peak areas inthe sedimentation profiles, it was roughly estimated thatabout half of the AChE dimers and monomers in NL lostthe amphiphilic domain with the phospholipase. Further-more, the extent of the GPI cleavage was clearly lower forthe G2

A and G1A AChE species released from DL (Fig. 2),

which indicated that the processing of the GPI anchor wasmodified by dystrophy.

The occurrence of PIPLC-sensitive and PIPLC-resistant AChE species in liver suggests heterogeneity inthe composition of the GPI anchor, a property alreadydescribed for AChE dimers of mouse erythrocytes (Tou-tant et al., 1991), and heart (Gomez et al., 1999). Theexistence of abundant GPI-linked AChE dimers, alongwith some enzyme tetramers, demonstrated that, as in thecase of heart, mouse liver had the capacity to produceAChE H and T subunits. This point was confirmed byidentifying the AChE transcripts in mouse liver.

Analysis of AChE Transcripts in Mouse TissuesTo investigate the origin of the AChE molecules

found in the liver extracts, assays were made to detectAChE messenger RNAs in mouse liver. Total RNA wasisolated from this and other tissues (as positive controls).The RNA was treated with retrotranscriptase in the pres-ence of primer P4, which can produce first-strand cDNAfor the three AChE transcripts, and afterwards cDNAswere amplified by PCR with one of the reverse primers(P4 or P10) plus the common forward primer P1.

The products obtained are shown in the agarose gelof Figure 3. The lanes on the left correspond to primersP1 1 P4, which were designed to detect the T transcripts.A band of ;520 bp was observed in brain, liver, spleen,and testis. This band clearly derives from the T transcript(524 bp expected product), which gives rise to the synapticAChE subunit, and is especially abundant in brain(Rachinsky et al., 1990). Using these primers, H and Rtranscripts should provide bands of 1,188 bp and 1,304 bp,but such bands were not detected.

TABLE III. Adsorption of Mouse Liver ChE Components inHydrophobic Gels*

Molecular form

% Adsorption

Phenyl-agarose Octyl-Sepharose

AChETotal 78 89G4

H 20 18G4

A 85 94G2

A 1 G1A 81 96

BuChETotal 17 91G4

H 7 6G4

A 51 75G2

A 1 G1A 17 98

*Equal volumes of the soluble fractions S1 and S2 from normal mouse liverwere mixed, and aliquots were poured into small columns of phenyl-agarose or octyl-Sepharose. Unbound cholinesterase (ChE) forms elutedfreely, and bound molecules were eluted with 2% Triton X-100 in salt-free N-[2-Hydroxyethyl]piperazine-N9-[2-ethanesulfonic acid] (HEPES)buffer. Acetylcholinesterase (AChE) and butyrylcholinesterase (BuChE)forms in unbound and bound fractions were identified by centrifugation ingradients with Brij 96. Percentages of binding refer to the activity recoveredin unbound plus bound fractions (together 95–105% of the activity appliedonto the columns). The level of interaction of each molecular form wascalculated from the sedimentation profiles. Similar results were obtained inexperiments made with extracts of dystrophic mouse liver.

Changes of ChEs in Dystrophic Mouse Liver 139

Lanes on the right side show the products with P1 1P10. With these primers, R transcripts should produce a980-bp fragment, and a band of this size was seen in thefour tissues, being more intense in spleen and weaker inbrain, liver, and testis. A much weaker band of ;870 bp,which was attributed to H transcripts (864-bp product),was observed in samples of spleen, brain, and liver. There-fore, the RT-PCR experiments allowed us to detect T, R,and H AChE mRNAs in mouse liver, although they seemto be scarce in this organ, especially the H transcript. Thelow content of AChE messengers in mouse liver agreeswith previous observations made in rabbit liver, wherethey were detected by Northern hybridizations after verylong exposures (Jbilo et al., 1994). Our observations dem-onstrate that liver has the capacity to produce the full set ofAChE subunits.

Interaction of AChE From MouseLiver With Lectins

Despite the identification of AChE mRNAs inmouse liver and its exhaustive perfusion, it might besuspected that a certain amount of the AChE activityassayed in liver comes from blood. To investigate this, thelevel of interaction between lectins and AChE in liver,erythrocyte, and plasma were compared, a study whichcould also shed light on possible alterations in the process-ing of ChE forms in liver of dystrophic mice, as occurs inmuscle (Cabezas-Herrera et al., 1994a, 1997).

The liver AChE activity from normal mice was fullybound to Con A, 50–60% to LCA, 50–60% to WGA,and 20–30% to RCA; the erythrocyte enzyme was 90–

100% fixed by LCA, 70–80% by WGA, and 20–30% byRCA. In addition, the hepatic AChE reacted with LCA toa lower extent (50–60%) than did the plasma enzyme(90–100%). Consequenly, it is very unlikely that theAChE activity measured in liver extracts derives fromblood contamination.

Since glycosylation of ChEs depends on the tissueand the particular molecular form, a possible relationshipbetween hepatic and blood ChEs was investigated bymeasuring the lectin absorption of individual enzymeforms in liver, erythrocytes, and plasma. In addition, thebinding of lectins to ChEs from normal and dystrophicmouse liver was analyzed to check whether the oligogly-cans linked to ChE forms were altered by dystrophy.

About half of the liver AChE dimers and monomersand a high amount of tetramers were bound to LCA orWGA, and less to RCA (Table IV, Fig. 4), suggesting thatan important fraction of the AChE species bears oligogly-cans with both terminal mannose and fucose in the “fu-cosylation core” (as revealed by the LCA reactivity) andterminal NAcGlc/sialic acid (WGA reactivity), whereasfew AChE species contain terminal galactose (RCA reac-tivity). Nevertheless, the much higher adsorption ofAChE tetramers to lectins than dimers or monomersclearly indicates that tetrameric and light AChE compo-nents are separately processed. Unexpectedly, the amountof dimers or monomers with LCA reactivity notably in-creased in the liver of dystrophic mice (Fig. 4), and thisrevealed that the processing of oligoglycans linked to thelight AChE forms was disturbed by the pathology.

Fig. 3. Analysis of AChE messenger RNAs in mouse liver. Total RNAfrom various tissues was isolated with Trizol and subjected to reversetranscriptase-polymerase chain reaction (RT-PCR) for AChE. PrimerP4 was used for obtaining cDNAs, and PCR was made with P1 plus thereverse primer indicated, either P4 or P10. After cDNA amplification,PCR products were separated in agarose gels and visualized withethidium bromide. The expected positions for amplified fragments of

each transcript are marked in the margins. With P4, the four tissuesshowed a band of about 524 bp, derived from T transcripts. In addition,a PCR product of 980 bp, corresponding to R transcripts, was pro-duced with P10. Using this primer, a very weak H product (864 bp)was observed for spleen, brain, and liver, but it is hardly distinguishedin the picture. DNA size markers denoted by mk1 and mk2 aremultiples of 100 bp, the band with the black star having a size of 1,000 bp.

140 Gomez et al.

AChE dimers in mouse liver or erythrocytes weredistinguished by their different interaction with lectins;only 40–50% of the hepatic dimers (the most abundantforms in the tissue) bound to LCA or WGA (Table IV),while the majority of dimers in the red blood cell mem-branes were recognized by LCA (90–100%) or WGA(70–80%). Important differences were also observed inthe percentage of binding of LCA or WGA with the liverAChE dimers (40–50% with either lectin), or the plasmaisoforms (90–100% with LCA, and 70–80% with WGA).

Adsorption of Mouse Liver and PlasmaBuChE to Lectins

Since it has long been admitted that liver is theprincipal source of plasma BuChE (Chatonnet and Lock-ridge, 1989; Kambam et al., 1994), similar patterns oflectin interaction with homologous BuChE forms fromthe two sources should be expected. However, the plasmaBuChE species bound to a higher extent to RCA than thehomologous liver forms (Table IV). Thus, despite thecommon amphiphilic behavior of liver (see above) orplasma BuChE dimers (Garcıa-Ayllon et al., 1999), theirdistinct interaction with RCA supports the notion that the

liver BuChE dimers (and possibly tetramers) were made tobe exported or to remain in the tissue differ in glycosyla-tion. It is noteworthy that, in contrast to what is observedfor the light AChE forms, no differences were found in theextent of lectin binding, whether the light BuChE com-ponents were released from liver of control or dystrophicmice (Fig. 4).

DISCUSSIONThe results reported here show that mouse liver

contains an important level of BuChE and much lessAChE activity, which contrasts with the similar content ofboth enzymes in rat hepatocytes (Perelman and Brandan,

TABLE IV. Percentages of Interaction of Lectins With AChEand BuChE Forms Extracted From Liver, Erythrocytes, andPlasma of Normal Mice*

Sample Molecular form

% Lectin binding

LCA WGA RCA

AChELiver G4

H 1 G4A 80–90 70–80 50–60

G2A 1 G1

A 40–50 40–50 10–20Erythrocytes G2

A 90–100 70–80 20–30Plasma G4

H 90–100 90–100 30–40G2

H 90–100 70–80 10–20G*1 90–100 30–40 0–10

BuChELiver G4

H 1 G4A 80–90 80–90 10–20

G2A 1 G1

A 70–80 60–70 10–20Plasma G4

H 90–100 90–100 90–100G2

A 90–100 70–80 60–70G1

A 90–100 30–40 30–40

*One ml of liver extracts (S1 1 S2), Triton-solubilized erythrocyte mem-branes or plasma were left to react with 0.5 ml of Sepharose 4B (control),concanavalin A (Con A)-, Lens culinaris agglutinin (LCA)-, wheat germagglutinin (WGA)-, and Ricinus communis agglutinin (RCA)-agarose. Afterovernight incubation, at 4°C, the enzyme-lectin complexes were sedi-mented by centrifugation, and the unbound cholinesterase (ChE) formsidentified by sedimentation analysis. The peaks of the individual forms inthe supernatant after incubation with Sepharose 4B were taken as the 100%values. The results are given as the percentage range of separated molecularforms which were fixed by lectins. G*1 denotes acetylcholinesterase (AChE)monomers whose sedimentation value is not modified by detergents butbind to octyl-Sepharose (Garcıa-Ayllon et al., 1999). Both AChE andbutyrylcholinesterase (BuChE) in the various sources were almost com-pletely bound (80–100%) by Con A, and therefore sedimentation analysis ofthe Con A supernatants was not attempted. The values were obtained fromat least four separate experiments.

Fig. 4. Interaction of lectins with ChE forms of mouse liver. Equalvolumes of the supernatants S1 and S2 extracted from liver of normal ordystrophic mice were mixed, and 1.0-ml aliquots were added to 0.5 mlof lectin-free Sepharose 4B (control), concanavalin A (Con A)-, Lensculinaris agglutinin (LCA)-, wheat germ (Triticum vulgaris) agglutinin(WGA)- or Ricinus communis agglutinin (RCA)-agarose. After incuba-tion overnight at 4°C, the mixtures were centrifuged to separateprotein-lectin complexes. The unbound ChE forms were identified bysedimentation analysis. The histograms show the percentages of inter-action of individual ChE forms with the lectins. ChE activity in controlexperiments was considered as the 100% value, and the extent of lectinassociation was calculated from the difference in peak areas of controland lectin sedimentation profiles. Percentages of binding of ChE formsto Con A-agarose are not represented because the enzyme activity inthe supernatant (10–15% for AChE, and 0–10% for BuChE) was toolow for sedimentation analyses. Results are means 6 S.D. of eightseparate experiments made with normal and seven with dystrophicmice liver; Asterisk denotes statistical significance (P , 0.002).

Changes of ChEs in Dystrophic Mouse Liver 141

1989; Perelman et al., 1990). Abundant G2A and G1

A AChE(together 87%), and BuChE isoforms (93%), along withminor G4

H and G4A enzyme species are identified in mouse

liver. Although sucrose gradient centrifugation did notpermit a complete separation of G2

A and G1A, a close look

at the sedimentation profiles (Fig. 1) shows that G2A AChE

and G1A BuChE predominate in liver extracts. Important

differences exist in the composition of ChEs forms ofmouse liver and plasma, in which G1

A (60%), G4H (25%),

and G2H (15%) AChE forms, and G4

H (80%), G2A, and G1

A

(together 20%) BuChE species have been identified(Garcıa-Ayllon et al., 1999).

Although AChE and BuChE dimers and monomersof mouse liver display amphiphilic properties, as shown bytheir interaction with detergent and octyl-Sepharose, theretention of the AChE species in phenyl-agarose and thelow adsorption of the BuChE forms in the gel demonstratethat the hydrophobic domains involved in binding to thehydrophobic resin differ in AChE and BuChE subunits.This feature can be successfully exploited to separate lightAChE from BuChE components, by means of phenyl-agarose.

As in mouse heart (Gomez et al., 1999), liver AChEdimers and monomers possess GPI domains for membraneanchorage, and this coincides with previous observationsmade in rat hepatocytes (Perelman and Brandan, 1989;Perelman et al., 1990). The GPI residues are responsiblefor the amphiphilic properties of the light AChE forms.The liver G2

A and G1A BuChE forms remain amphiphilic

after incubation with PIPLC, and this agrees with theproposal that their amphiphilicity is caused by an exposedamphipatic a-helix, which is made of seven conservedaromatic amino acids, located at the C-terminus of thesubunit, and involved in tetramerization (Altamirano andLockridge, 1999). Since BuChE monomers of mouseliver, plasma (Garcıa-Ayllon et al., 1999), brain (Moral-Naranjo et al., 1996), heart (Gomez et al., 1999), orskeletal muscle (Moral-Naranjo et al., 1999) show am-phiphilic properties, it seems that in mouse tissues andplasma at least, the amphiphilic behavior of BuChE dimersand monomers should be considered as the rule rather thanas the exception. The occurrence of abundant hydrophilicBuChE tetramers in mouse serum and of amphiphilicvariants in brain raises various intriguing questions, such ashow the BuChE subunits are folded in each case, whichproteins are involved in the formation of amphiphilic orhydrophilic tetramers, and what physiological factors de-cide the specific pattern of folding.

While BuChE dimers and monomers of liver andplasma share amphiphilic properties, the homologousAChE forms differ in amphiphilicity. Thus, although thelight AChE forms from both sources bind to octyl-Sepharose, the migration of the liver monomers varieswith the detergent added to the sucrose gradient, a prop-erty not exhibited by the isoforms of plasma (Garcıa-Ayllon et al., 1999). The hydrophobic GPI anchor in liverAChE monomers and the lack of the anchor in the plasma

isoforms might explain their distinct hydrodynamic be-havior.

As regards the origin of liver AChE, variable levels ofT, H, and R AChE transcripts are detected by RT-PCRof liver RNA, and this indicates that liver may produce thefull set of AChE subunits. Nevertheless, the presence ofGPI residues in AChE dimers of liver and erythrocytes,and the common amphiphilic properties of the lightBuChE forms in liver and plasma prompted us to inves-tigate whether some of the ChE activity measured in liverextracts could come from blood.

The results concerning the composition of ChEforms in liver, erythrocytes, and plasma, the differences inamphiphilicity of AChE components, and the distinctlevel of interaction of ChE forms in the three sources withlectins demonstrate that mouse liver preparations are freeof significant blood contamination. This conclusion isbased on the following observations: (1) in spite of theoverwhelming abundance of BuChE tetramers in mouseplasma, they exist as minor species in liver; (2) in contrastto the liver AChE monomers and dimers, the plasmacomponents are devoid of GPI residues, and even thedimers display hydrophilic behavior (Garcıa-Ayllon et al.,1999); (3) the AChE dimers in liver and erythrocytes differin the much higher association of the latter with LCA; (4)while a small fraction of the hepatic BuChE tetramers areadsorbed by RCA, those in plasma bind completely to thelectin; (5) similarly, a small part of the light BuChE formsof liver react with RCA, and the percentage of bindingincreases for the plasma components.

It has been suggested that plasma AChE comes fromsecretion of various tissues as well as from erythrocyte,after cleavage of the GPI anchor of dimers with plasmaphospholipases (Chatonnet and Lockridge, 1989). Theincapacity of both plasma AChE dimers and the erythro-cytic GPI-devoid isoforms (obtained after incubation withPIPCL) to be bound in octyl-Sepharose, and the similarpatterns of interaction with lectins (Table IV) support theabove suggestion. The abundance of GPI-anchored AChEmonomers and dimers in mouse liver, along with someAChE tetramers, and the identification of H and T tran-scripts, demonstrate that the organ has the capacity toproduce H subunits for building GPI-linked AChE com-ponents, and T subunits for tetrameric species (from theprecursor light forms). This shows that type I G1

A and G2A

AChE (made of H subunits) coexist with type II isoforms(made of T subunits). Assuming that liver is a source ofplasma AChE, the lack in this organ of measurableamounts of GPI-devoid AChE dimers or monomers (withamphiphilic properties like those of plasma) and the smallamount of tetramers may indicate that, in contrast to theGPI-linked (type I) molecules, the GPI-free (type II)AChE subunits are rapidly assembled into tetramers andsecreted perhaps through the constitutive pathway.

As regards BuChE molecules in liver and plasma, fewBuChE tetramers of liver (10–20%) are recognized byRCA, whereas most of the plasma variants (90–100%) arefixed by the lectin. Several alternatives may explain this

142 Gomez et al.

difference: (1) liver may produce two pools of BuChEtetramers, one addressed to hepatic compartments and theother to plasma; probably those identified in liver residepermanently in the tissue to fulfill certain functions (suchas collaborating in detoxification), while those destined toplasma are not detected in liver because of rapid exocy-tosis; and (2) the BuChE tetramers are in transit for de-livery, in which case, prior to release, they should acquiregalactose and other sugars in the distal Golgi regions.

As regards the physiological role of ChEs in liver,their identification in intestine epithelial cells (Sine et al.,1992; L’Hermite et al., 1996), meninges (Ummenhofer etal., 1998), bronchial cells (Taisne et al., 1997), and hepa-tocytes (Perelman and Brandan, 1989; Perelman et al.,1990) strongly suggests that these enzymes may function inepithelial cells as a hydrolytic filter for the withdrawal ofblood-circulating acetylcholine (through AChE activity)and other aliphatic or aromatic esters (through BuChEactivity). In addition, the occurrence of GPI-linked AChEdimers and monomers in mouse liver and rat hepatocytes(Perelman et al., 1990), the abundance of GPI-anchoredproteins in caveolae-rich membrane domains (Masserini etal., 1999), and the identification of caveolin-1, the struc-tural component of caveolae, on the surface membrane ofhepatocytes and sinusoidal plasma membrane (Pol et al.,1999), raise the interesting possibility that AChE forms areconcentrated in caveolae in the sinusoidal space. This

location of AChE forms on the blood-facing plasma mem-brane may facilitate clearance of acetylcholine, while theirassociation to caveolae may permit their rapid adjustmentto satisfy particular physiological demands of the liver andother epithelial tissues.

The dystrophic phenotype caused by a genetic defi-ciency of the laminin a2 chain in mice does not modifythe content or the composition of ChE forms in liver.Nevertheless, the fraction of PIPLC-resistant AChEmonomers, and probably dimers, increases in liver ofdystrophic mice. This observation and the much highercontent of light AChE forms with LCA reactivity indystrophic liver strongly suggest that both the GPI anchorand the oligosaccharides linked to AChE subunits arealtered by the pathology (Fig. 5). In spite of the patho-logical changes reported in feline muscular dystrophy(Carpenter et al., 1989), little information is availableregarding the effects of muscular dystrophy on liver me-tabolism. Nevertheless, the decreased content of cyto-chrome P450 in liver of dystrophic mdx mice, and theincreased level of lactate dehydrogenase, aspartate transam-inase, and cholesterol in the mice serum (Brazeau et al.,1992) demonstrate a relationship between dystrophy andliver pathology. Whether the abnormal liver metabolismalters glycosyl transferases action remains to be elucidated.

In summary, our results show that mouse liver is richin light (G2

A 1 G1A) AChE and BuChE forms, a charac-

Fig. 5. Biosynthetic relationship between AChE components empha-sizing the changes observed in dystrophic mouse liver. AChE pre-mRNA is alternatively spliced at its 3’ end to produce three maturemRNAs encoding protein products with three distinct C-termini. Theprotein products correspond to R (for “readthrough”), H (for “hydro-phobic”), and T (for “tailed”) AChE subunits. The AChE mRNAs aretranslated by ribosomes associated to rough endoplasmic reticulum(RER). The product for the R transcript has not yet clearly beenidentified. AChE H subunits become amphiphilic after incorporationof glycosylphosphatidylinositol (GPI) residues at the C-terminus ofsubunits, which produces GPI-bearing G1

A and G2A light AChE forms.

Some GPI-linked proteins lose the GPI domain by incubation withPIPLC, but not all GPI anchors are PIPLC-sensitive; some GPI anchors

are PIPLC-resistant probably because the anchors have an extra palmi-tate residue at the inositol ring (Toutant et al., 1991). Before insertionin the plasma membrane, the glycans linked at the subunits of the lightAChE forms should be remodelled at the Golgi system. In dystrophicmouse liver, the light AChE forms show an increased interaction withthe lectin LCA, and their GPI residues are more resistant to PIPLCtreatment, features that may reflect an abnormal processing of the lipidanchor and N-linked oligosaccharides. After assembly of T subunits atthe RER, G4

A and G4H AChE species transit the Golgi system where

their N-linked glycans are processed before entering the exocyticpathway. The distinct patterns of lectin interaction of the liver AChElight forms and tetramers suggest that GPI-linked components andGPI-free tetramers are separately processed at the Golgi system.

Changes of ChEs in Dystrophic Mouse Liver 143

teristic which makes it a good source for the isolation andfurther study of their functional and structural properties,as well as for gaining insights into the mechanism involvedin tetramerization of BuChE subunits. The detection ofnoticeable amounts of AChE transcripts in liver indicatesthat this organ produces AChE molecules. Moreover, thecomposition of liver AChE and BuChE forms, their am-phiphilic properties, and the patterns of interaction withlectins confirm that the enzymes arise from the tissue andnot from blood contamination. The increased content ofPIPLC-resistant and LCA-reactive AChE forms in liver ofdystrophic mice suggests that the hepatic pathology, whichis secondary to muscular dystrophy, probably affects thebiosynthesis of the lipid anchor and the oligoglycans. Theabnormal processing of GPI-bearing glycoproteins mayhave functional consequences for the liver.

ACKNOWLEDGMENTSWe greatly thank Dr. Victor Garre for his help and

discussion of results regarding identification of AChEmRNAs, and Dr. S. Torres and J. C. Gomez for the useof the thermocycler. J.L.G. and M.S.G-A. were recipientsof predoctoral fellowships from the Instituto de Coopera-cion Iberoamericana, and the University of Murcia, re-spectively. F.J.C. was supported by the Ministerio deEducacion y Cultura and by the Fundacion Seneca.

REFERENCESAltamirano CV, Lockridge O. 1999. Conserved aromatic residues of the

C-terminus of human butyrylcholinesterase mediate the association oftetramers. Biochemistry 38:13414–13422.

Brazeau GA, Mathew M, Entrikin RK. 1992. Serum and organ indices ofthe mdx dystrophic mouse. Res Commun Chem Pathol Pharmacol77:178–179.

Brimijoin S, Koenigsberger C. 1999. Cholinesterases in neural develop-ment: new findings and toxicologic implications. Environ Health Persp107:59–64.

Cabezas-Herrera J, Moral-Naranjo MT, Campoy FJ, Vidal CJ. 1994a. G4

forms of acetylcholinesterase and butyrylcholinesterase in normal anddystrophic mouse muscle differ in their interaction with Ricinus communisagglutinin. Biochim Biophys Acta 1225:283–288.

Cabezas-Herrera J, Campoy FJ, Vidal CJ. 1994b. Amphiphilic properties ofmolecular forms of acetylcholinesterase in normal and dystrophic muscle.J Neurosci Res 38:505–514.

Cabezas-Herrera J, Moral-Naranjo MT, Campoy FJ, Vidal CJ. 1997.Glycosylation of acetylcholinesterase forms in microsomal membranesfrom normal and dystrophic Lama2dy mouse muscle. J Neurochem 69:1964–1974.

Campbell KP. 1995. Three muscular dystrophies: loss of cytoskeleton-extracellular matrix linkage. Cell 80:675–679.

Carpenter JL, Hoffman EP, Romanul CA, Kunkel LM, Rosales RK, MaNSF, Dasbach JJ, Rae JF, Moore FM, McAfee MB, Pearce LK. 1989.Feline muscular dystrophy with dystrophin deficiency. Am J Pathol135:909–919.

Chan RYY, Adatia FA, Krupa AM, Jasmin BJ. 1998. Increased expressionof acetylcholinesterase T and R transcripts during hematopoietic differ-entiation is accompanied by parallel elevations in the levels of theirrespective molecular forms. J Biol Chem 273:9727–9733.

Chatonnet A, Lockridge O. 1989. Comparison of butyrylcholinesterase andacetylcholinesterase. Biochem J 260:625–634.

Culligan KG, Mackey AJ, Finn DM, Maguire PB, Ohlendieck K. 1998.Role of dystrophin isoforms and associated proteins in muscular dystrophy(review). Int J Mol Med 2:639–648.

Dulley JR, Grieve PA. 1975. A simple technique for eliminating interfer-ence by detergents in the Lowry method of protein determination. AnalBiochem 64:136–141.

Emery AEH. 1998. The muscular dystrophies. Brit Med J 317:991–995.Garcıa-Ayllon MS, Gomez JL, Vidal CJ. 1999. Amphiphilic properties of

acetylcholinesterase monomers in mouse plasma. Neurosci Lett 265:211–214.

Gisiger V, Stephens HR. 1988. Localization of the pool of G4 acetylcho-linesterase characterizing fast muscles and its alteration in murine musculardystrophy. J Neurosci Res 19:62–78.

Gomez JL, Moral-Naranjo MT, Campoy FJ, Vidal CJ. 1999. Character-ization of acetylcholinesterase and butyrylcholinesterase forms in normaland dystrophic Lama2dy mouse heart. J Neurosci Res 56:295–306.

Grisaru D, Sternfeld M, Eldor A, Glick D, Soreq H. 1999. Structural rolesof acetylcholinesterase variants in biology and pathology. Eur J Biochem264:672–686.

Jbilo O, L’Hermite Y, Talesa V, Toutant JP, Chatonnet A. 1994. Acetyl-cholinesterase and butyrylcholinesterase expression in adult rabbit tissuesand during development. Eur J Biochem 225:115–124.

Johnson G, Moore SM. 1999. The adhesion function on acetylcholinest-erase is located at the peripheral anionic site. Biochem Biophys ResCommun 258:758–762.

Kambam JR, Franks JJ, Mets B, Janicki PK, Hickman R, Van der Watt M.1994. The effect of hepatectomy and plasma cholinesterase inhibition oncocaine metabolism and cardiovascular responses in pigs. J Lab Clin Med124:715–722.

L’Hermite A, Sine JP, Colas B. 1996. Ultraestructural localization ofbutyrylcholinesterase in epithelial cells of rat intestine. Eur J Histochem40:299–304.

Layer PG, Willbold E. 1995. Novel functions of cholinesterases indevelopment, physiology and disease. Progr Histochem Cytochem29:1–93.

Legay C, Bon S, Vernier P, Coussen F, Massoulie J. 1993. Cloning andexpression of a rat acetylcholinesterase subunit: generation of multiplemolecular forms and complementarity with a Torpedo collagenic subunit.J Neurochem 60:337–346.

Li Y, Camp S, Rachinsky TL, Getman D, Taylor P. 1991. Gene structureof mammalian acetylcholinesterase: alternative exons dictate tissue-specific expression. J Biol Chem 266:23083–23090.

Marcos MR, Sanchez-Yague J, Hernandez-Hernandez A, Llanillo M.1998. Amphiphilic and hydrophilic forms of acetylcholinesterase fromsheep platelets. Biochim Biophys Acta 1415:163–173.

Masserini M, Palestini P, Pitto M. 1999. Glycolipid-enriched caveolae andcaveolae-like domains in the nervous system. J Neurochem 73:1–11.

Massoulie J, Pezzementi L, Bon S, Krejci E, Vallette FM. 1993. Molecularand cellular biology of cholinesterases. Prog Neurobiol 41:31–91.

Matsumura K, Chiba A, Yamada H, Fukuta-Ohi H, Fujita S, Endo T,Kobata A, Anderson LVB, Kanazawa I, Campbell KP, Shimizu T. 1997.A role of dystroglycan in schwannoma cell adhesion to laminin. J BiolChem 272:13904–13910.

Moral-Naranjo MT, Cabezas-Herrera J, Vidal CJ. 1996. Molecular formsof acetyl- and butyrylcholinesterase in normal and dystrophic mousebrain. J Neurosci Res 43:224–234.

Moral-Naranjo MT, Campoy FJ, Cabezas-Herrera J, Vidal CJ. 1999.Increased butyrylcholinesterase levels in microsomal membranes of dys-trophic Lama2dy mouse muscle. J Neurochem 73:1138–1144.

Moya-Quiles MR, Villalba-Sanchez J, Munoz-Delgado E, Vidal CJ. 1992.Alkaline treatment of muscle microsomes releases amphiphilic and hy-drophilic forms of acetylcholinesterase. Biochim Biophys Acta 1121:88–96.

144 Gomez et al.

Naom I, D’Alessandro M, Topaloglu H, Sewry C, Ferlini A, Helbling-Leclerc A, Guicheney P, Weissenbach J, Schwartz J, Bushby K, Philpot J,Dubowitz V, Muntoni F. 1997. Refinement of the laminin a2 chainlocus to human chromosome 6q2 in severe and mild merosin deficientcongenital muscular dystrophy. J Med Genet 34:99–104.

Ozawa E, Noguchi S, Mizuno Y, Hagiwara Y, Yoshida M. 1998. Fromdystrophinopathy to sarcoglycanopathy: evolution of a concept of mus-cular dystrophy. Muscle Nerve 21:421–438.

Perelman A, Brandan E. 1989. Different membrane-bound forms of ace-tylcholinesterase are present at the cell surface of hepatocytes. Eur J Bio-chem 182:203–207.

Perelman A, Abeijon C, Hirschberg CB, Inestrosa NC, Brandan E. 1990.Differential association and distribution of acetyl- and butyrylcholinest-erase within rat liver subcellular organelles. J Biol Chem 265:214–220.

Pol A, Calvo M, Lu A, Enrich C. 1999. The “early-sorting” endocyticcompartment of rat hepatocytes is involved in the intracellular pathway ofcaveolin-1 (VIP-21). Hepathology 29:1848–1857.

Rachinsky TL, Camp S, Li Y, Ekstrom TJ, Newton M, Taylor P.1990. Molecular cloning of mouse acetylcholinesterase: tissue distributionof alternatively spliced mRNA species. Neuron 5:317–327.

Roberts RG. 1995. Dystrophin, its gene, and the dystrophinopathies. AdvGenet 33:177–231.

Saez-Valero J, Vidal CJ. 1995. Monomers and dimers of acetylcholinester-ase in human meningioma are anchored to the membrane by glyco-sylphosphatidylinositol. Neurosci Lett 195:101–104.

Saez-Valero J, Vidal CJ. 1996. Biochemical properties of acetyl- andbutyrylcholinesterase in human meningioma. Biochim Biophys Acta1317:210–218.

Schwarz M, Glick D, Loewenstein Y, Soreq H. 1995. Engineering of

human cholinesterase explains and predicts diverse consequences ofadministration of various drugs and poisons. Pharmacol Ther 67:283–322.

Sine JP, Toutant JP, Weigel P, Colas B. 1992. Amphiphilic forms ofbutyrylcholinesterase in mucosal cells of rat intestine. Biochemistry 31:10893–10900.

Sternfeld M, Ming GL, Song HJ, Sela K, Timberg R, Poo MM, Soreq H.1998. Acetylcholinesterase enhances neurite growth and synapse devel-opment through alternative contribution of its hydrolytic capacity, coreprotein, and variable C termini. J Neurosci 18:1240–1249.

Sunada Y, Bernier SM, Utani A, Yamada Y, Campbell KP. 1995. Identi-fication of a novel mutant transcript of laminin a-2 chain gene responsiblefor muscular dystrophy and dysmyelination in dy2j mice. Hum Mol Genet4:1055–1061.

Taisne C, Norel X, Walch L, Labat C, Verriest C, Mazmanian GM, BrinkC. 1997. Cholinesterase activity in pig airways and epithelial cells. Fun-dam Clin Pharmacol 11:201–205.

Tome FMS. 1999. The saga of congenital muscular dystrophy. Neurope-diatrics 30:55–65.

Toutant JP, Krall JA, Richards MK, Rosenberry TL. 1991. Rapid analysisof glycolipid anchors in amphiphilic dimers of acetylcholinesterase. CellMol Neurobiol 11:219–230.

Ummenhofer WC, Brown SM, Bernards CM. 1998. Acetylcholinesteraseand butyrylcholinesterase are expressed in the spinal meninges of monkeysand pigs. Anesthesiology 88:1259–1265.

Weitnauer E, Ebert C, Hucho F, Robitzki A, Weise C, Layer PG. 1999.Butyrylcholinesterase is complexed with transferrin in chicken serum. JProtein Chem 18:205–214.

Changes of ChEs in Dystrophic Mouse Liver 145