Tài liệu Báo cáo khoa học: Acetylcholinesterase from the invertebrate Ciona intestinalis is capable of assembling into asymmetric forms when co-expressed with vertebrate collagenic tail peptide doc

Pezzementi, Department of Biology, Birmingham-Southern College, Box 549022, Birmingham, AL 35254, USA Fax: +1 205 226 3078 Tel: +1 205 226 4806 E-mail: lpezzeme@bsc.edu Website: http://f

Ciona intestinalis is capable of assembling into

asymmetric forms when co-expressed with vertebrate

collagenic tail peptide

Adam Frederick1, Igor Tsigelny2, Frances Cohenour1, Christopher Spiker1, Eric Krejci3,

Arnaud Chatonnet4, Stefan Bourgoin1, Greg Richards1, Tessa Allen1, Mary H Whitlock1and

Leo Pezzementi1

1 Department of Biology, Birmingham-Southern College, Birmingham, AL, USA

2 Department of Chemistry and Biochemistry, San Diego Supercomputer Center, University of California at San Diego, La Jolla, CA, USA

3 Institut National de la Sante´ et de la Recherche Me´dicale U686, Universite´ Paris Descartes, Biologie des Jonctions Neuromusculaires, Paris, France

4 Institut National de la Recherche Agronomique, Montpellier, France

Keywords

acetylcholinesterase; asymmetric forms;

butyrylcholinesterase; Ciona intestinalis;

evolution

Correspondence

L Pezzementi, Department of Biology,

Birmingham-Southern College, Box 549022,

Birmingham, AL 35254, USA

Fax: +1 205 226 3078

Tel: +1 205 226 4806

E-mail: lpezzeme@bsc.edu

Website: http://faculty.bsc.edu/lpezzeme/

Database

The nucleotide sequence and derived amino

acid sequence data reported for the AChE

from Ciona intestinalis are available in the

Third Party Annotation Section of the

DDBJ ⁄ EMBL ⁄ GenBank databases under the

accession no TPA: BK006073 The

align-ment used to determine the phylogenetic

tree for vertebrate and invertebrate

cholines-terases presented here is deposited at the

EMBL-ALIGN database as ALIGN_001208

(Received 14 November 2007, revised 7

January 2008, accepted 15 January 2008)

doi:10.1111/j.1742-4658.2008.06292.x

To learn more about the evolution of the cholinesterases (ChEs), acetylcho-linesterase (AChE) and butyrylchoacetylcho-linesterase in the vertebrates, we investi-gated the AChE activity of a deuterostome invertebrate, the urochordate Ciona intestinalis,by expressing in vitro a synthetic recombinant cDNA for the enzyme in COS-7 cells Evidence from kinetics, pharmacology, mole-cular biology, and molemole-cular modeling confirms that the enzyme is AChE Sequence analysis and molecular modeling also indicate that the cDNA codes for the AChET subunit, which should be able to produce all three globular forms of AChE: monomers (G1), dimers (G2), and tetramers (G4), and assemble into asymmetric forms in association with the collagenic subunit collagen Q Using velocity sedimentation on sucrose gradients, we found that all three of the globular forms are either expressed in cells or secreted into the medium In cell extracts, amphiphilic monomers (G1) and non-amphiphilic tetramers (G4na) are found Amphiphilic dimers (G2) and non-amphiphilic tetramers (G4na) are secreted into the medium Co-expression of the catalytic subunit with Rattus norvegicus collagen Q produces the asymmetric A12form of the enzyme Collagenase digestion of the A12AChE produces a lytic G4form Notably, only globular forms are present in vivo This is the first demonstration that an invertebrate AChE is capable of assembling into asymmetric forms We also performed a phylo-genetic analysis of the sequence We discuss the relevance of our results with respect to the evolution of the ChEs in general, in deuterostome inver-tebrates, and in chordates including vertebrates

Abbreviations

a

, amphiphilic; AChE, acetylcholinesterase; AChE H, splice variant H; AChE T, splice variant T; ATCh, acetylthiocholine; BTCh,

butyrylthiocholine; BuChE, butyrylcholinesterase; ChE, cholinesterase; ColQ, collagen Q; DEPQ,

7-[(diethoxyphosphoryl)oxy]-1-methylquinolinium iodide; DTNB, 5-(3-carboxy-4nitro-phenyl)disulfanyl-2-nitro-benzoic acid; GPI, glycophosphatidylinositol; HIS buffer, high ionic strength buffer; IC 50, half maximal inhibitory concentration; LBA, long branch attraction;na, non-amphiphilic; PPII, polyproline II; PRAD, proline-rich attachment domain; PRiMA, proline-rich membrane anchor; WAT, tryptophan (W) amphipathic tetramerization domain.

Gnathostome vertebrates have two evolutionarily

related cholinesterases (ChEs), acetylcholinesterase

(AChE; EC 3.1.1.7) and butyrylcholinesterase (BuChE;

EC 3.1.1.8) AChE rapidly hydrolyzes the

neurotrans-mitter acetylcholine at cholinergic synapses BuChE

appears to act as a scavenger of cholinergic toxins, but

may also play a role in synaptic transmission [1,2]

These two enzymes appear to be the result of a gene

duplication event early in vertebrate evolution [3]

Both enzymes have a 20 A˚ deep catalytic gorge lined

with aromatic amino acids [4] AChE has fourteen

aro-matic residues lining the gorge; in BuChE, aliphatic

amino acids replace six of the aromatic moieties In

particular, smaller non-aromatic residues in BuChE

replace the two phenylalanines of the acyl pocket of

AChE (Phe288 and Phe290 in Torpedo californica

AChE), a subsite of the enzyme that plays an

impor-tant role in substrate specificity Amino acid position

numbers appearing in parentheses represent the

homo-logous positions in mature AChE from Torpedo

cali-fornica; residues of the T peptides of AChET from

different species are numbered from 1 to 48 to

facili-tate comparisons As a result, BuChE can

accommo-date larger and more diverse substrates and inhibitors

compared to AChE [5] By contrast to the

dichoto-mous acyl pocket situation of ChEs in the vertebrates,

invertebrates have a wider diversity in the structure of

this subsite In approximately 90% of invertebrate

ChEs, the acyl pocket is formed in a fundamentally

different way [6] Instead of Phe288 and Phe290

form-ing the pocket, it is formed by phenylalanines at

posi-tions homologous to Phe290 and Val400 For example,

in ChE2 from the cephalochordate amphioxus, which

is very specific for the substrate acetylthiocholine

(ATCh), the acyl pocket is composed of Phe312

(Phe290) and Phe422 (Val400) [6] One of the

excep-tions to this invertebrate pattern is found in the

sequence for a putative AChE from the urochordate

Ciona intestinalis [7,8], a deuterostome invertebrate

that is a close relative to the vertebrates In this

enzyme, phenylalanines homologous to those of the

acyl pocket of vertebrates appear to form the acyl

pocket [6] Previously, based on substrate and inhibitor

specificity, it was reported that C intestinalis possesses

an AChE in vivo [9–11] However, that work was

con-ducted before the techniques of molecular biology were

available, precluding the definitive identification of the

enzyme

Another difference between vertebrate and

inverte-brate ChEs is that verteinverte-brates possess both globular

and asymmetric forms of the enzymes, but

inverte-brates apparently possess only globular forms The

globular forms of ChEs are monomers (G1), dimers

(G2), and tetramers (G4) of catalytic subunits The asymmetric forms are comprised of one (A4), two (A8), or three (A12) tetramers attached to a triple-stranded collagenic tail (collagen Q; ColQ) [12,13] The asymmetric forms associate with the basal lamina [14]

Alternative splicing of the AChE gene in the verte-brates produces a number of carboxyl termini [15], resulting in the multiple molecular forms mRNAs containing the H-terminus (AChEH) are translated into glycophosphatidylinositol-membrane-anchored (GPI)

G2forms of AChE By contrast, transcripts containing the alternatively spliced T-terminus (AChET) are capable of forming all globular forms: amphiphilic monomers (G1 ), amphiphilic dimers (G2 ), and non-amphiphilic tetramers (G4na), but not GPI-membrane-anchored G2 More importantly, AChET, via its tryptophan (W) amphipathic tetramerization domain (WAT) sequence [16], can associate with the proline-rich attachment domain (PRAD) of the collagenic sub-unit ColQ to form asymmetric enzyme [17,18] or with the proline-rich membrane anchor (PRiMA) protein [19]

AChET appears to be rare in invertebrates, where AChEH predominates AChET has been reported for AChE1 from the nematodes Caenorhabditis spp [20,21] and Meloidogyne spp [22], where it forms G1 and a G4form that may associate with a structural subunit [20,21] Meedel reported that C intestinalis lar-vae have G1, G2, and G4forms of AChE, implying the presence of AChET in the invertebrate, but did not find any asymmetric forms [11]

The cloning, in vitro expression, and characterization

of this putative AChE from C intestinalis should iden-tify the nature of the enzyme and provide additional information about the evolution of the ChEs, including the origins of the acyl pocket, the T exon, and the asymmetric forms of ChE in the vertebrates

Results The sequence of the ChE from C intestinalis suggests that the enzyme is an AChET The sequence for C intestinalis AChE contains 618 amino acids (see supplementary Fig S1) The mem-bers of the catalytic triad of AChE are found as Ser

229, Glu 356, and His 471 The three pairs of con-served cysteine residues involved in intrachain disul-fide bonding are also found as Cys 94–Cys 121, Cys 293–Cys 297, and Cys 431–Cys 562 Another cysteine (Cys 616) near the carboxyl terminus of the sequence probably mediates interchain disulfide bonding Of

the fourteen aromatic amino acids that line the cata-lytic gorge of vertebrate AChE, 13 are conserved in the C intestinalis AChE (AChE1; Table 1) The sequence shows 41% identity with the AChE from

T californica

The formation of the acyl pocket of C intestinalis AChE may more closely resemble that of vertebrate AChE rather than invertebrate AChE (Fig 1; see molecular modeling below) However, the acyl pockets

of C intestinalis and T californica are clearly not iden-tical because, as is the case for other invertebrates, there

is a deletion in the region of the acyl pocket of C intes-tinalis compared to the vertebrate enzyme (Fig 1) Additionally, the carboxyl terminus of the C

intestinal-is AChE appears to be coded for by a T exon: six of the seven aromatic residues of the T californica AChE WAT domain are conserved; there is a 74% sequence similarity with T californica AChE, and the domain has ability to form an amphipathic helix, characteristic

of the T sequence A cysteine that mediates interchain disulfide bonds is also conserved (Fig 2A,B) We found

no evidence in the genomic sequence of an upstream

H exon in the C intestinalis AChE gene

A second gene for AChE in C intestinalis has been proposed [8] (Genbank accession no AK112482; cioin-acche2 in ESTHER; AChE2; Table 1) [23] However,

Table 1 Aromatic amino acids in the catalytic gorge of putative

AChEs from C intestinalis and AChE from T californica Numbering

for C intestinalis AChE2 starts at first methionine residue in the

sequence Conserved aromatic residues are shown in bold

Desig-nations of AChE1 and AChE2 are from the ESTHER database [23]

to distinguish the AChE described in the present study (AChE1)

and another sequence proposed to be an AChE from C intestinalis

(GenBank accession no AK112482).

Subsite

C intestinalis AChE1

C intestinalis AChE2

T californica AChE

Choline binding site

and hydrophobic

site

a

There is a deletion in the CLUSTALW alignment in this region of the

sequence for AChE2.

Fig 1 Amino acid residues surrounding the acyl pocket of some vertebrate and invertebrate acetylcholinesterases This figure illustrates the differences between the construction of the acyl pocket in vertebrate and invertebrate AChEs The line separates the vertebrate and invertebrate AChEs The numbers at the top of the figure correspond to the amino acids in T californica In the vertebrates, the acyl pocket

is composed of Phe288 and Phe290 In the invertebrates, the acyl pocket phenylalanines homologous to the Phe290 and Val400 positions form the acyl pocket ( CLUSTALW aligned the amino acid sequences [58]) The alignment of the sequence for C intestinalis AChE was slightly adjusted manually (the QE sequence) to emphasize the similarity with vertebrate AChEs The GenBank accession nos are: Homo sapiens (M55040), Bos taurus (BC123898), Mus musculus (X56518), R norvegicus (S50879), Gallus gallus (U03472), Bungarus fasciatus (U54591),

T californica (X03439), Myxine glutinosa (U55003), C intestinalis (TPA: BK006073), B floridae (U74381), S purpuratus (XM_777020; pre-dicted similar to AChE), Drosophila melanogaster (X05893), Anopheles stephensi (228651), C elegans (X75332), Meloidogyne incognita (AF075718), Loligo opalescens (AF065384), and Boophilus microplus (AJ223965).

the derived amino acid sequence shows only 28%

iden-tity with the AChE from T californica, and only 30%

homology with the C intestinalis AChE described in

the present study Although the three pairs of

con-served cysteine residues involved in intrachain disulfide

bonding in AChEs are found in the sequence, only two

members of the catalytic triad are present: serine and

glutamate The third residue, histidine, is replaced by a

cysteine This replacement would probably inactivate

the enzyme; in T californica and human AChEs,

respectively, H440Q and H447Q mutants lack activity

[24,25] Additionally, of the fourteen aromatic amino

acids that line the catalytic gorge of vertebrate AChE,

only six are conserved in the sequence (Table 1);

how-ever, the sequence shows the invertebrate acyl pocket

conformation, which provides a seventh aromatic

resi-due in the gorge Nevertheless, in BuChE, eight of the

residues are conserved [1] Particularly important is the

absence of the tryptophan of the choline-binding site

In human AChE, a W86A mutation increases Kmby

660-fold [26] Finally, the sequence clearly does not

have a carboxyl terminus coded for by an AChETexon,

as only one of the seven aromatic residues is preserved

It is highly unlikely that this protein is an active AChE because it is missing a member of the catalytic triad and the main aromatic residue for binding of substrate Additionally, the protein would not be expected to pro-duce all three globular forms because it does not con-tain a WAT domain However, it could represent a GPI-anchored protein because it has a putative signal sequence, and a putative hydrophobic C-terminus and cleavage site (not shown) What, if any, role the protein may play in the organism has not yet been determined; although it shows highest homology with ChEs and not other ChE-like adhesion molecules

Kinetic characterization of recombinant ChE from

C intestinalis expressed in vitro and native enzyme expressed in vivo indicates the enzyme

is AChE

To determine the nature of the cholinesterase activity

of the recombinant C intestinalis enzyme and to com-pare it with the native AChE, we assayed the

hydroly-Fig 2 Amino acid sequences of T peptides found in vertebrates and deuterostome and protostome invertebrates Top, alignment of

T amino acid sequences: Vertebrates, H sapiens (M55040), M musculus (X56518), T californica (X03439); deuterostome invertebrates,

C intestinalis (TPA: BK006073), S purpuratus (XM_775310); protosome invertebrates, Apysia californica (AASC01147222.1), C elegans (X75332) An alternatively spliced exon codes for the T peptide and numbering starts at the first amino acid of the peptide The six conserved aromatic amino acids of the WAT domain are indicated by ; the one nonconserved aromatic residue by h The S purpuratus sequence is associated with a putative AChE; the A californica sequence has not been associated with AChE Sequences aligned with

CLUSTALW Bottom: helical wheel representation of the WAT domain organized as an amphipathic a-helix [62] The conserved aromatic, hydrophobic (green diamonds) cluster at the top of the wheel The arrow points to the nonconserved Tyr in the WAT domain of C intestinalis AChE Green and yellow residues are hydrophobic Red, blue, and orange residues are hydrophilic.

sis of ATCh and butyrylthiocholine (BTCh) by enzyme

that was secreted into the medium by the COS-7 cells,

enzyme extracted from the cells, and enzyme extracted

from adult C intestinalis Only ATCh is hydrolyzed

appreciably, as indicated by the low values of

VmaxBTCh⁄ VmaxATCh It proved difficult to determine

accurate kinetic parameters for BTCh hydrolysis given

the low activity that the enzyme showed for the

sub-strate, and it was not possible to detect BTCh

hydroly-sis by extracts of adult organisms; nevertheless, the

kinetic parameters determined are in reasonable

agree-ment The enzymes also show substrate inhibition

(i.e lower enzyme activity at high substrate

concentra-tions, and bparameter values of < 1) (Fig 3;

Table 2) The selective hydrolysis of ATCh is

charac-teristic of AChE

Pharmacological characterization of the

recombinant ChE from C intestinalis expressed

in vitro and native enzyme expressed in vivo

confirms that the enzyme is AChE

To determine further the nature of the cholinesterase

activity of the recombinant enzyme, we determined half

maximal inhibitory concentration (IC50) values of the

enzymes for the inhibitors

(3aS-cis)-1,2,3,3a,8,8a-hexa-hydro-1,3a,8-trimethylpyrrolo[2,3-b]indol-5-ol

methyl-carbamate (physostigmine), which inhibits all

cholinesterases;

[4-[5-[4-(dimethyl-prop-2-enyl-ammo-

nio)phenyl]-3-oxo-pentyl]phenyl]-dimethyl-prop-2-enyl-azanium dibromide (BW284c51), which inhibits AChE

preferentially; and 10-(2-diethylaminopropyl)

phenothi-azine hydrochloride (ethopropphenothi-azine) and

N-[bis(pro-

pan-2-ylamino)phosphoryloxy-(propan-2-ylamino)phos-phoryl]propan-2-amine (iso-OMPA), which inhibit

BuChE at low concentrations Physostigmine and

BW284c51 inhibit the enzymes at lm concentrations;

by contrast, much higher concentrations of

ethoprop-azine and iso-OMPA are required for inhibition

(Fig 4; Table 3) This pattern is characteristic of AChE

COS-7 cells transfected with cDNA for AChE

of C intestinalis produce all three globular molecular forms of AChE

To determine the molecular forms of AChE produced

in vitro by COS-7 cells transfected with the catalytic subunit for C intestinalis AChE, we performed vel-ocity sedimentation on sucrose gradients in the pres-ence and abspres-ence of Triton X-100 Cell extracts have

G1 and G4na because the G1form shifts to a higher sedimentation coefficient in the absence of detergent The forms of AChE secreted into media are G2 and

Substrate (M)

10–6 10–5 10–4 10–3 10–2 10–1

10–7

0 200 400 600 800 1000 1200

Fig 3 Representative experiment showing concentration depen-dencies for ATCh and BTCh hydrolysis by an extract of COS-7 mon-key cells expressing recombinant C intestinalis AChE cDNA Transfected COS-7 cells producing C intestinalis AChE were extracted in HIS buffer and assayed with ATCh (d) or BTCh (s) as described in the Experimental procedures.

Table 2 Kinetic parameters for recombinant and native AChE from C intestinalis Data are the mean ± SE of four or more determinations Sources of enzyme: medium, enzyme secreted into the medium, usually 12 mL; cells, enzyme extracted with 5 mL of HIS buffer from the COS-7 cells as described in the Experimental procedures; and organism, enzyme extracted from adult C intestinalis, as described in the Experimental procedures.

Source

VmaxATCh

(mAb ⁄ min)

KmATCh

(l M )

KssATCh

(m M ) bATCh

VmaxBTCh

(mAb ⁄ min)

KmBTCh

(m M )

KssBTCh

(m M ) bBTCh V maxBTCh⁄ V maxATCh

a

Values of b less than 0.02 are indistinguishable from zero.bHigh concentrations of endogenous reducing compounds in the adult tissue increased the background in the Ellman’s assay and, despite correction, obscured whatever low levels of BTCh hydrolysis there may have been; kinetic parameters for BTCh hydrolysis could not be obtained.

G4na because the sedimentation coefficient of the

G2form also increases in the absence of detergent For

extracts of adult C intestinalis, G1 and G4na are seen

on the gradients In both extracts and media, the

sedimentation coefficient of the G4 form remains

unchanged (Fig 5; Table 4)

COS-7 cells co-transfected with cDNAs for the

catalytic subunit of C intestinalis and ColQ from

the rat produce the A12form of AChE

To determine whether the catalytic subunits of C

in-testinalis AChE catalytic subunits could assemble into

asymmetric forms of AChE in the presence of a

colla-genic tail, we co-transfected COS-7 cells with cDNAs

for the C intestinalis catalytic subunit and for R

nor-vegicus ColQ, and analyzed cell extracts on sucrose

gradients In addition to peaks corresponding to G1

and G4, a peak of enzyme activity appears at

approxi-mately 16S, which is characteristic of the A12form of AChE Collagenase digestion at 37C converts the putative A12form to a lytic G4; a shoulder of residual undigested A12is visible (Fig 6; Table 5) We have not found genes for ColQ or PRiMA in the C intestinalis genome

Molecular modeling of C intestinalis AChE also indicates that the catalytic subunit can assemble into asymmetric forms

Molecular modeling, in addition to sequence analysis, also indicates that the catalytic gorge of C intestinalis AChE is similar to that of vertebrate AChEs, showing

Fractional AChE activity

1.00

0.75

0.50

0.25

0.00

Inhibitor ( M )

10 –8 10 –7 10 –6 10 –5 10 –4 10 –3

Fig 4 Representative experiment showing concentration

depen-dencies for inhibition of ATCh hydrolysis by recombinant AChE

from C intestinalis Media from transfected COS-7 cells secreting

C intestinalis AChE was collected and incubated with inhibitors for

20 min prior to being assayed for activity The inhibitors used were

BW284c51 (d), physostigmine (s), ethopropazine (.), and

iso-OMPA (,).

Table 3 IC50values (l M ) for inhibition of recombinant and native

AChE from C intestinalis Data are the mean ± SE of three or

more determinations Sources of enzyme are the same as in

Table 1.

Source Physostigmine BW284c51 Ethopropazine Iso-OMPA

Medium 5.09 ± 0.66 0.93 ± 0.17 768 ± 203 > 3000

Cells 7.35 ± 0.28 1.91 ± 0.01 650 ± 93 > 3000

Organism 14.1 ± 0.76 1.23 ± 0.76 741 ± 60 > 3000

0 5 10 15 20 25

0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16

Sedimentation coefficient

0 5 10 15 20 25

0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14

Fig 5 Velocity sedimentation analysis of the globular molecular forms of C intestinalis AChE produced in vitro and in vivo Medium from COS-7 cells transfected with cDNA for the catalytic subunit for C intestinalis AChE, total HIS (d); extracts of the transfected cells (h) and total HIS extracts of adult C intestinalis tissue ( ) were analyzed on sucrose gradients in the presence (top) and absence (bottom) of Triton X-100 as described in the Experimental procedures.

an AChE-like acyl pocket, a hydrophobic patch

(including the choline binding site), and an oxyanion

hole (see supplementary Fig S2) The distance between

the acyl pocket phenylalanines, Phe317 and Phe319, is

3.7 A˚, the same as for T californica AChE However,

the volume of the catalytic gorge for C intestinalis AChE is 780 A˚3, whereas the volume of the gorge for

T californicaAChE is 986 A˚3 Molecular modeling of monomeric C intestinalis AChE catalytic subunits with the PRAD domain

of ColQ indicates that the WAT domain of the

C intestinalis AChE is capable of organizing the sub-units into a tetramer through interaction with the PRAD domain of ColQ The [AChET]–ColQ com-plex model was built based on the PRAD–WAT interaction; inter-subunit interactions involving the catalytic domains were considered secondary [27] As

a result, the complex has a quasi-four-fold axis of symmetry (Fig 7A,B) The four WAT domains of the tetramer form a-helices and coil around a single antiparallel PRAD domain, which approximates a left-handed polyproline II (PPII) helical conforma-tion The three tryptophans of the WAT domain ori-ent inwards to interact with ColQ, and come into close contact and stack with the prolines of the PRAD domain (Fig 7C,D)

Phylogenetic analysis of AChE sequences supports a classical phylogeny for deuterostome invertebrates

A phylogenetic analysis of vertebrate and deutero-stome and protostome invertebrate ChEs places

C intestinalis AChE intermediate between the echino-derms and the cepalochordate amphioxus (see supple-mentary Fig S3) This placement is consistent with conventional phylogenetic trees based primarily on morphological data [28] Note, however, that the branch length for C intestinalis AChE is the longest in the tree, and the bootstrap value for the branching between amphioxus and C intestinalis is one of the weakest in the tree

Discussion

We have expressed in vitro a synthetic recombinant ChE from the urochordate C intestinalis Based on

Table 4 Sedimentation coefficients of recombinant and native forms of AChE from C intestinalis Data are the mean ± SE of ‡ 8 determi-nations for recombinant enzyme and three and four determidetermi-nations for enzyme extracted from adult C intestinalis in the presence and absence of Triton X-100, respectively Sources of enzyme are the same as in Table 1.

Conditions

Sedimentation coefficients

Sedimentation coefficient

Fractional activi

ty on gradient

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.16

Fig 6 Velocity sedimentation analysis of globular and asymmetric

forms of AChE produced by cotransfection with cDNAs for C

intes-tinalis catalytic subunit and rat ColQ Total HIS cell extracts were

digested with collagenase and analyzed on sucrose gradients as

described in the Experimental procedures Control (d); collagenase

digestion (s).

Table 5 Sedimentation coefficients of C intestinalis AChE

cata-lytic subunit co-expressed with ColQ with and without digestion by

collagenase Data are the mean ± SE of ‡ 7 determinations.

Conditions Sedimentation coefficients

)Collagenase 5.10 ± 0.07 11.48 ± 0.10 16.09 ± 0.14

+Collagenase 5.16 ± 0.12 11.54 ± 0.08 15.65 ± 0.20 a

a Estimated from residual activity b Amphiphilic or non-amphiphilic

forms are not designated because the appropriate velocity

sedi-mentation experiments on sucrose gradients in the presence and

absence of Triton X-100 were not performed The forms are

assumed to be G 1aand G 4na.

substrate and inhibitor specificity, the enzyme is

AChE The AChE is AChET because transfected

COS-7 cells produce G1, G2, and G4forms

Co-expression of C intestinalis AChE catalytic subunit

and rat collagenic tail, ColQ, results in the assembly

of the A12asymmetric form Sequence analysis and

molecular modeling support both of these

conclu-sions In some respects, the AChE from C

intestinal-is more closely resembles the AChE of the

vertebrates than any other invertebrate AChE and

provides information about the evolution of the

ChEs

The ChE from the invertebrate C intestinalis is

an AChE that resembles vertebrate AChE

Our kinetic data are consistent with those of Fromson

and Whittaker [9] and Meedel and Whittaker [10],

who investigated ChE activity in extracts of larval

C intestinalis, and also concluded that the activity is

due to AChE They found that the hydrolysis of

BTCh was 4.5% of that for ATCh at 25 mm [8], and

that high concentrations of ATCh produced substrate

inhibition [10] They do not show a hydrolysis curve

for BTCh and, in the present study, we were unable

to detect hydrolysis of BTCh by extracts of adult

C intestinalis Our estimates of their values for Km

(approximately 100 lm) and Kss (approximately

100 mm) for ATCh hydrolysis data are comparable to our own [10]

Our pharmacological results are also consistent with previous studies of C intestinalis demonstrating that physostigmine and BW284c51 were effective inhibitors

of the activity, but that iso-OMPA was not [9,10] The only IC50that can be obtained from these data is for BW284c51 (approximately 1 lm) [9], which is vir-tually identical to that found in the present study Not only does the congruence of the kinetic and phar-macologic data indicate that C intestinalis possesses AChE, but it also argues that the cDNA expressed

in vitro in the present study corresponds to the gene expressed in vivo

Sequence analysis of important residues in the cata-lytic gorge also supports the assertion that the enzyme

is AChE Only one of the 14 aromatic amino acids that line the catalytic gorge of T californica and most other vertebrate AChEs is missing in C intestinalis AChE, Tyr70, a member of the peripheral site, which

is replaced by Ile97 The Kss of C intestinalis AChE for ATCh is rather high and this substitution could contribute to this value [29]

More interesting is the nature of the acyl pocket In all vertebrate AChEs, the acyl pocket is comprised of two phenylalanines close to one another in the pri-mary sequence By contrast, for approximately 90%

of invertebrate AChEs, the acyl pocket is composed

Fig 7 Modeled structures of C intestinalis AChE [AChET]–ColQ complex (A, B) The [AChET]-ColQ complex modeled on the basis of the [WAT] 4 –PRAD structure, from the side and bottom respectively Each cata-lytic subunit is shown in a different color (purple, yellow, blue and orange), as is ColQ (green) (C) Hydrophobic interactions between WAT and PRAD helices The view

is down and into the PRAD helix in the center of the figure The four WAT helices are shown colored as in (A) and (B) The magenta space-filled residues are the Trps

of the WAT domains, which all face inward and surround the PRAD (D) Cut away view showing the Trps (in space-filling format) of two WAT domains (colored as above) inter-acting with the PRAD PPII helix The Trp side-chains zipper into the grooves of the PPII helix.

of two phenylalanines far apart in the primary

sequence, corresponding to Phe290 and Val400 in

T californica AChE The only known exception in

this subset of invertebrate AChEs is C intestinalis

AChE, where the acyl pocket phenylalanines are

homologous to those of the vertebrates, suggesting

that the C intestinalis acyl pocket is ancestral to that

of the vertebrates This conclusion is confounded by

the acyl pocket conformations of the two

acetyl-cholinesterases from the cephalochordate amphioxus

(Branchiostoma floridae); the cephalochordates have

long been considered to be the sister group of the

ver-tebrates ChE2 from this organism shows the typical

invertebrate acyl pocket structure, whereas ChE1

apparently has a novel acyl pocket, unlike the typical

vertebrate or invertebrate conformations [6,30,31] The

echinoderms, urochordates (tunicates, C intestinalis),

cephalochordates (amphioxus, B floridae), and

verte-brates are members of the deuterostome branch of the

animal kingdom, with the echinoderms generally

con-sidered as the most basal of the groups, the

urochor-dates intermediate, and the cephalochorurochor-dates closest

to the vertebrates [28,32,33] However, recent data

from metaphylogenies and phylogenomics have

chal-lenged this view, with Blair and Hedges [34], Delsuc

et al [35], and Vienne and Pontarotti [36] proposing

that the urochordates are actually the closest living

relatives of the vertebrates, with the cephalochordates

intermediate to the echinoderms Our phylogenetic

analysis of deuterostome AChEs supports the classical

phylogeny and is similar to the phylogenetic tree for

AChE of various vertebrates and deuterostome

inver-tebrates provided by Vienne and Pontarotti [36] Note,

however, that the branch length for C intestinalis

AChE is the longest in the tree; this long branch

length is typical of many C intestinalis genes and is a

result of rapid evolution in the species [34,37] This

rapid evolution and the resultant long branch length

gives rise to an artifact called long branch attraction

(LBA), which has a number of effects Most

impor-tantly in this case, LBA results in the grouping of two

sequences that evolve more rapidly than the others

do: C intestinalis AChE and a putative AChE from

the echinoderm Strongylocentrotus purpuratus LBA is

also a problem in metaphylogenies, but can be

cor-rected for more easily, and a consensus is forming

around the revised deuterostome phylogeny, with the

urochordates actually being the sister group to the

vertebrates [28,34–36] Not only does LBA

compro-mise our AChE phylogeny, but also the bootstrap

value for the branching between amphioxus and C

in-testinalis AChEs is one of the weakest in the tree,

indicating its uncertainty If it is assumed that the

urochordates are the closest living relative of the ver-tebrates, the acyl pocket of C intestinalis may in fact

be ancestral to that of the vertebrates What may have been responsible for the shift in acyl pocket structure during the transition from invertebrates to vertebrates,

or nonchordates to chordates and vertebrates, remains

a matter of speculation

The AChE from C intestinalis is AChETand is able to assemble into asymmetric forms organized by vertebrate ColQ

Analysis of the carboxyl terminus sequence indicates that the C intestinalis AChE is AChET, which should

be capable of forming the three globular forms: G1,

G2 , and G4na When the catalytic subunit of AChE from C intestinalis was expressed in vitro, G1 , G2 , and

G4naforms of enzyme were produced The

amphiphilici-ty of G1and G2is due to the exposure of the hydropho-bic T peptide of their carboxyl termini, which interact with detergent micelles on the gradients; while the

T peptide of G4 is sequestered away from solvent and unable to interact with detergent [38] Extracts of adult

C intestinalis contained G1 and G4naforms By con-trast, it was reported that extracts of the larvae produce all three globular forms, possibly indicating a develop-mental difference in AChE assembly between the larvae and adults [11] Nevertheless, all three G forms pro-duced in vivo are also propro-duced in vitro

Inspection of the T peptide sequence shows that all

of the tryptophans of the WAT domain are conserved

in the C intestinalis sequence However, one of the seven aromatic amino acids, Tyr20, is replaced by Ser20 In Torpedo marmorata AChE, the mutations Y20A and Y20P decrease the amphipathic nature of the T peptide a-helix and abolish the assembly of secreted tetramers when catalytic subunits are co-expressed in the presence or absence of a truncated, soluble version of ColQ [39] In the [WAT]4–PRAD model of Dvir et al [18] (PDB ID code 1VZJ), there is

an edge-on p-p interaction between the edge of Phe14

in WAT strand A and the face of Tyr20 in chain D This interaction is not observed for the other Phe14-Tyr20 combinations The Y20A and Y20P mutations would disrupt this interaction and apparently destabi-lize the tetramer Clearly, this is not the case for the AChE tetramer of C intestinalis, which forms tetra-mers in the absence and presence of ColQ The WAT– PRAD interaction of our tetrameric molecular model

is in good agreement with the corresponding structure

of Dvir et al [18], and indicates that the side-chain of Ser20 in strand D is oriented towards and in close proximity to the edge of Phe14 in strand A One

possibility is that the tetramer is stabilized by the

for-mation of a weak C–HÆO hydrogen bond between the

hydroxyl oxygen of Ser20 and a slightly polar C–H

group of the aromatic ring of Phe14 Such bonds were

first proposed in 1982 for phenylalanines in proteins

by Thomas et al [40], and have received considerable

attention in recent years [41–44]

Co-expression of C intestinalis AChE catalytic

sub-unit with rat ColQ resulted in the production of the

A12asymmetric form of AChE These results confirm

our molecular modeling, which indicated that the

appropriate interactions between the WAT domain of

the catalytic subunit and the PRAD domain of ColQ

were present to assemble the catalytic tetramers of the

asymmetric forms The A12form consists of three such

tetramers attached to the triple-stranded helix of ColQ

This result is the first demonstration of the assembly

of catalytic subunits of an invertebrate AChE into

asymmetric forms

The evolution of the T peptide and tetrameric

forms of AChE

However, one question arises: what, if anything,

assembles the C intestinalis G4 tetramers in the

absence of ColQ in vivo or in vitro? T peptide

sequences have been identified in vertebrates;

deutero-stome invertebrates, the urochordate C intestinalis and

the echinoderm S purpuratus; and in protostome

invertebrates, the mollusk Aplysia californica and

vari-ous nematodes, including Caenorhabditis elegans,

sug-gesting that the peptide is widespread in nature The

presence of the T peptide in both branches of the

ani-mal kingdom indicates that it may be as old as and

conserved for ‡ 900 million years because it would

have had to evolve prior to the

protostome-deutero-stome split [34] Interestingly, all of the phyla that

have the T sequence also have G4AChE, and use

ace-tylcholine as a neurotransmitter at their neuromuscular

junctions, suggesting both are a prerequisite for

effi-cient synaptic transmission at the junctions Given the

recent research on the interaction between WAT

domains of AChE catalytic subunits and the PRAD

domains of ColQ and PRiMA [38,39,45,46], the fact

that G4AChE interacts with a noncatalytic subunit in

nematodes [19,20], the recent finding of small

PRAD-containing polypeptides associated with soluble

tetramers of vertebrate BuChE [47], and the apparent

ubiquity of the T domain, we propose that

PRAD-containing proteins mediate tetramerization of AChE

throughout evolution, with ColQ and PRiMA of the

vertebrates comprising just two of the many examples

of such proteins

Experimental procedures Materials

DMEM, fetal bovine serum, and OptiMEM medium were purchased from Invitrogen (Carlsbad, CA, USA) FuGene was obtained from Roche (Indianapolis, IN, USA) ATCh, BTCh, BW284c51, 5-(3-carboxy-4nitro-phenyl)disulfanyl-2-nitro-benzoic acid (DTNB), iso-OMPA, ethopropazine, and physostigmine were purchased from Sigma (St Louis, MO, USA) Type-3 collagenase was obtained from Worthington (Lakewood, NJ, USA) 7-[(diethoxyphosphoryl)oxy]-1-methylquinolinium iodide (DEPQ) was a gift from Yacov Ashani Adult specimens of C intestinalis were purchased from The Marine Biological Laboratory (Woods Hole,

MA, USA) We thank Andrew Gannon for help with the

C intestinalisdissection

Gene synthesis and analysis

The ci0100132088 gene from the urochordate C intestinalis

is now identified in the Department of Energy Joint Gen-ome Institute (DOE JGI) Database (http://genGen-ome.jgi-psf org/Cioin2/Cioin2.home.html) as an AChE gene The sequence for this gene is embedded in the C intestinalis genome sequence (GenBank accession no AABS01000124) [7,8] We spliced out the intronic sequences and translated the coding exonic sequences in silico Nucleotide sequence and derived amino acid sequence data reported are avail-able in the Third Party Annotation Section of the DDBJ⁄ EMBL ⁄ GenBank databases under the accession no TPA: BK006073 These sequence data are also available on the DOE JGI Database The amino acid sequence for the protein has also been deposited in the Esther database as cioin-acche1 (http://bioweb.ensam.inra.fr/ESTHER/general? what=index) [23] A BLAST search was conducted at NCBI with the translated sequence, and it was found to be similar to many AChE amino acid sequences in that data-base, showing 72% homology with the AChE of Ciona sav-ignyi GenScript Corporation (Piscataway, NJ, USA) synthesized and subcloned a cDNA for the protein into pcDNA3.1 (Invitrogen) after linker sequences containing EcoRI and XbaI restriction sites were added to the 5¢- and 3¢-ends of the cDNA, respectively, for ligation of the cDNA into the expression plasmid Double-strand DNA sequenc-ing confirmed the sequence The recombinant plasmid was then used to transform competent Escherichia coli (XL1-Blue; Stratagene, La Jolla, CA, USA) Qiagen maxi-preps (Qiagen, Valancia, CA, USA) were used to obtain plasmid DNA for transfections

In vitro expression and extraction of enzymes

COS-7 monkey cells (American Type Culture Collection, Manassas, VA, USA) were grown in DMEM containing