JBC

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Originally published In Press as doi:10.1074/jbc.M402596200 on April 21, 2004

J. Biol. Chem., Vol. 279, Issue 26, 27098-27107, June 25, 2004
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
279/26/27098    most recent
M402596200v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Lee, H. H. C.
Right arrow Articles by Tsim, K. W. K.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Lee, H. H. C.
Right arrow Articles by Tsim, K. W. K.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

Transcriptional Regulation of Acetylcholinesterase-associated Collagen ColQ

DIFFERENTIAL EXPRESSION IN FAST AND SLOW TWITCH MUSCLE FIBERS IS DRIVEN BY DISTINCT PROMOTERS*

Henry H. C. Lee{ddagger}§, Roy C. Y. Choi{ddagger}§, Annie K. L. Ting{ddagger}, Nina L. Siow{ddagger}||, Joy X. S. Jiang{ddagger}, Jean Massoulié**, and Karl W. K. Tsim, Held a visiting professorship at Ecole Normale Supérieure in 2001 and 2003{ddagger}{ddagger}{ddagger}

From the {ddagger}Departments of Biology and Molecular Neuroscience Center, The Hong Kong University of Science and Technology, Clear Water Bay Road, Hong Kong SAR, China and **Laboratoire de Neurobiologie Cellulaire et Moléculaire, CNRS UMR 8544, Ecole Normale Supérieure, 75005 Paris France

Received for publication, March 8, 2004 , and in revised form, April 16, 2004.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 
The presence of a collagenous protein (ColQ) characterizes the collagen-tailed forms of acetylcholinesterase and butyrylcholinesterase at vertebrate neuromuscular junctions which is tethered in the synaptic basal lamina. ColQ subunits, differing mostly by their signal sequences, are encoded by transcripts ColQ-1 and ColQ-1a, which are differentially expressed in slow and fast twitch muscles in mammals. Two distinct promoters, pColQ-1 and pColQ-1a, were isolated from the upstream sequences of human COLQ gene; they showed muscle-specific expression and were activated by myogenic transcriptional elements in cultured myotubes. After in vivo DNA transfection, pColQ-1 showed strong activity in slow twitch muscle (e.g. soleus), whereas pColQ-1a was preferably expressed in fast twitch muscle (e.g. tibialis). Mutation analysis of the ColQ promoters suggested that the muscle fiber type-specific expression pattern of ColQ transcripts were regulated by a slow upsteam regulatory element (SURE) and a fast intronic regulatory element (FIRE). These regulatory elements were responsive to a calcium ionophore and to calcineurin inhibition by cyclosporine A. The slow fiber type-specific expression of ColQ-1 was abolished by the mutation of an NFAT element in pColQ-1. Moreover, both the ColQ promoters contained N-box element that was responsible for the synapse-specific expression of ColQ transcripts. These results explain the specific expression patterns of collagen-tailed acetylcholinesterase in slow and fast muscle fibers.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 
Acetylcholinesterase (AChE1; EC 3.1.1.7 [EC] ) is a highly polymorphic enzyme (1). A single ACHE gene produces several types of catalytic subunits by alternative splicing, but a single splice variant called type T (AChET) is expressed in adult mammalian muscles. AChET catalytic subunits produce amphiphilic monomers and dimers and nonamphiphilic homotetramers as well as heteromeric associations with anchoring proteins, ColQ and PRiMA, which allow their functional localization in cholinergic synapses (2). Collagen ColQ thus characterizes the collagen-tailed forms (A forms) of AChE and butyrylcholinesterase (3–5), which are localized in the basal lamina at neuromuscular junctions (nmjs) of vertebrates (6, 7); in these molecules (A4, A8, A12), one, two, or three tetramers of catalytic subunits are disulfide-linked to the strands of a triple helix of ColQ collagen. The cDNAs encoding ColQ have been cloned in Torpedo and mammals (4, 8, 9). The primary structure of ColQ comprises a signal peptide, an N-terminal domain containing a proline-rich motif (PRAD) that associates with four C-terminal T peptides of AChET or butyrylcholinesterase subunits, a collagenous domain with characteristic repeats of glycines every three residues, and a C-terminal region (10, 11).

In rodents the expression and distribution of AChE forms in muscles depend on age, on the fast or slow type of muscle, on the contact with the motor nerve, and on contractile activity (1). In rat, the appearance of collagen-tailed forms coincides with the establishment of nerve-muscle contacts (12, 13), and their level decreases dramatically after denervation (14, 15). In addition, the composition and the distribution of these molecules differ between fast and slow muscles, which vary with activity (12, 16–18). In fast muscles, the A12 form is largely predominant, and it is exclusively localized at the nmjs. In slow muscles, the A8 and A4 forms are relatively abundant, and they are found in extra-junctional regions of muscle fibers (12, 19). These differences can be explained by the distribution of AChE and ColQ transcripts in junctional and extra-junctional regions of the muscles (20); ColQ is only expressed at the nmjs in fast twitch muscles, but it is uniformly expressed in extra-junctional and junctional regions of slow twitch muscles (21).

In human and rat the COLQ gene (~50 kb in length) possesses two origins of transcription located ~23 kb apart, generating alternative exons 1 and 1a which are followed by 17 constitutive exons (21, 22). Exon 1 and exon 1a encode essentially the signal peptides so that the mature ColQ-1 and ColQ-1a subunits possess nearly identical sequences, in particular the PRAD. In rat muscles, the ColQ-1 and ColQ-1a transcripts are expressed differentially in slow and fast twitch muscles. To understand the regulation of collagen-tailed AChE at the nmjs, we analyzed the factors that control transcription of the COLQ gene in slow and fast muscles; we show that the two independent promoters preferentially drive the expression of ColQ-1 transcript in slow muscle and of ColQ-1a in fast muscle. We find that the fiber type-specific expression patterns of ColQ-1 and ColQ-1a are controlled by specific transcription elements, which strongly resemble those previously identified in promoters of other muscle proteins (23–25); that is, a slow upsteam regulatory element (SURE) and a fast intronic regulatory element (FIRE), which are active, respectively, in slow and fast twitch muscles. In addition, we identified elements that control the synapse-specific expression of ColQ.


   EXPERIMENTAL PROCEDURES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
RESULTS
 DISCUSSION
 REFERENCES
 
Cell Cultures—Eggs of New Hampshire chicks were purchased from a local farm and hatched in the Animal Care Facility at Hong Kong University of Science and Technology (HKUST). For primary chick myotube cultures, hindlimb muscles dissected from 11-day-old chick embryos were minced and then dissociated by trypsinization, stirring, and centrifugation as described previously (26). Muscle cells were routinely cultured in Eagle's minimal essential medium supplemented with 10% heat-inactivated horse serum, 2% chick embryo extract, 1 mM L-glutamine, 100 units/ml penicillin, and 100 µg/ml streptomycin at 37 °C in a water-saturated incubator of 95% air and 5% CO2. All cell culture reagents were from Invitrogen. Myotubes were treated with 10-5 M cytosine arabinoside at day 3 after plating. Mouse C2C12 muscle cells were obtained from the American Type Culture Collection (ATCC; Manassas, VA). Undifferentiated C2C12 myoblasts were maintained in Dulbecco's modified Eagle's medium supplemented with 20% fetal bovine serum, 100 units/ml penicillin, and 100 µg/ml streptomycin and incubated at 37 °C in a water-saturated incubator of 95% air and 5% CO2. Myogenic differentiation was induced by serum reduction according to Siow et al. (27). Mouse neuroblastoma x rat glioma NG108–15 cells, human embryonic kidney cells 293, and COS-7 cells were cultured in 100-mm culture dishes as described previously (28).

Promoter-Reporter Constructs and cDNAs—Human genomic DNA was extracted from human embryonic kidney cells 293 cells using the phenol-chloroform method (29), and genomic PCR was performed using Pfx (Invitrogen) to clone the ColQ-1 promoter (pColQ-1, GenBankTM accession number AF057036 [GenBank] ) with primers 5'-AA CCC GGG AAA ATT CAT ATT CTG ACA CC-3' (-2076 to -2057) and 5'-AA AGA TCT TGG ATT CAG GAC AAC CAT GC-3' (+18 to -2) and the ColQ-1a promoter (pColQ-1a, GenBankTM accession number AF229118 [GenBank] ) with primers 5'-AA CCC GGG GTT ATA GAT AGA GAG ACA CA-3' (-2252 to -2233) and 5'-AA AGA TCT GAA ATG ATG AGC AGA TGA GC-3' (+44 to +25). Restriction sites were added at the 5' end as indicated (not underlined). The PCR products were subcloned into the upstream of firefly luciferase in a pGL3-Basic vector (Promega, Madison, WI) to form pColQ-1-Luc for ColQ-1 promoter and pColQ-1a-Luc for ColQ-1a promoter. The promoter sequences were analyzed by the TF search program to identify potential binding sites for transcriptional factors. Truncations were constructed using different sense primers, 5'-AAC CCG GGA TTT TCA AGA AAA GGC AAA A-3' (-1487 to -1468), 5'-AAC CCG GGC CAG TAA CTA GCA GAC TGT T-3' (-985 to -966) and 5'-AAC CCG GGA ATC AGG AGA GTT GTT TCT-3' (-499 to -481) for the ColQ-1 promoter and primers 5'-AAC CCG GGA TAT GTC AAT AAG AGA GCC C-3' (-1389 to -1370), 5'-AAC CCG GGA AAG TGA GGT TGT AAC GG-3' (-895 to -878) and 5'-AAC CCG GGG AGT CAA GTC GGT TTT T-3' (-387 to -371) for the ColQ-1a promoter. Deletion of the SURE region in the ColQ-1 promoter was performed with primers 5'-AA GAA TTC GTT TCT GCC CAG TTC C-3' (-486 to -471) and 5'-AA GAA TTC ACC TTT TGA TTC AAT GCA AG-3' (-687 to -706) to form pColQ-1{Delta}SURE-Luc, and deletion of the FIRE region in the ColQ-1a promoter was performed with primers 5'-AA CTC GAG GGC TGC CTA CCC ATT-3' (-209 to -195) and 5'-AA CTC GAG TGC TCT GGT AAG CCT CAA AG-3' (-410 to -429) to form pColQ-1a{Delta}FIRE-Luc. The minimum regions corresponding to SURE of rat troponin I slow isoform promoter and FIRE from quail troponin I fast isoform promoter were generated by PCR according to published sequences (30); they were subcloned into pTA-Luc vector (BD Biosciences Clontech) to form pSURETnls-Luc and pFIRETnlf-Luc, respectively, as control plasmids. The SURE region from ColQ-1 promoter was obtained by PCR using the primers 5'-AA GGT ACC TGC AGC TTT TGG AGA GAG-3' (-686 to -669) and 5'-AA AGA TCT AAC TCT CCT GAT TCA AAC TG-3' (-487 to -506), and the FIRE region from ColQ-1a promoter was obtained by PCR with primers 5'-AA GGT ACC GTC AAT ACT GCT GTG GTA-3' (-525 to -508) and 5'-AA AGA TCT CTG CTC CAG CTG CCC-3' (-340 to -326); these SURE/FIRE regions were subcloned into pTA-Luc to form pSUREColQ-1-Luc and pFIREColQ-1a-Luc, respectively. Mouse ColQ-1 and ColQ-1a cDNAs including exon 1 and 1a were generated from the published sequences (8) or from genomic clones. Mouse myoD cDNA (Ref. 31; GenBankTM accession number M84918 [GenBank] ) and mouse growth-associated-binding protein (GABP) {alpha} and {beta} cDNAs (GenBankTM accession numbers NM008065 and NM016654) generated by RT-PCR from mouse C2C12 myotubes were subcloned into pcDNA3 for transfection. All cloned sequences were verified before the use. Vectors expressing the luciferase reporter either constitutively under a SV40 promoter (pSV40-Luc) or driven by NFAT elements (p3x(NFAT)-Luc) and expressing wild type Raf-1 (RafWT) and constitutively active Raf (RafCAAX) were purchased from BD Biosciences Clontech. Myogenin cDNA and luciferase reporter for E-box element (p4xRE-Luc) were described previously (27). An ~2-kb rat AChR {epsilon}-subunit promoter was tagged with luciferase to form pAChR{epsilon}-Luc (32), and a vector expressing luciferase under N-box elements (p3x(N-box)-Luc) was also constructed by subcloning into pTA-Luc (33).

Semiquantitative RT-PCR and Northern Blots—Total RNA was prepared by the LiCl method (29) from myotubes. RT-PCR was used to amplify regions flanking mouse ColQ exon 1 (sense primer, 5'-GGT GGT CCT GAA TCC AAT G-3') and exon 1a (sense primer, 5'-CTT CTC CTC ATC ATT TCG G-3') to common exon 3 (antisense primer, 5'-GAA GGT TCT TCA TGT CTG G-3'), with sizes of ~245 bp for exon 1-exon 3 and ~191 bp for exon 1a-exon 3, and mouse glyceraldehyde-3-phosphate dehydrogenase (5'-AAC GGA TTT GGC CGT ATT GG-3' and 5'-CTT CCC GTT CAG CTC TGG G-3'), with size ~650 bp. All methods for northern blotting were described by Siow et al. (27), with detection by a ~1.4-kb mouse ColQ subunit cDNA 32P-labeled probe (generated by RT-PCR from known sequence) and a ~1.8-kb probe corresponding to mouse AChE catalytic subunit (27). Samples of 20 µg of RNA were loaded per gel lane. The consistency of the RNA loading in every lane was confirmed by ethidium bromide staining of the ribosomal RNAs. Ethidium bromide-labeled and 32P-labeled bands were quantified using calibration curves from the same gel (26).

Site-directed Mutagenesis of NFAT Binding Site—The NFAT binding site was mutated by using overlapping PCR with mutagenic primers (32). Overlapping PCR was performed with two PCR fragments containing six nucleotides to be mutated (in bold). The first fragment was developed using primers 5'-AA CCC GGG AAA ATT CAT ATT CTG ACA CC and 5'-AGG AAG GAA GTA GTT TAT GAA TTC AAA CAG CTT TCT CTG TGA AAA A-3'; the second fragment was developed using primers 5'-TTT TTC ACA GAG AAA GCT GTT TGA ATT CAT AAA CTA CTT CCT TCC T-3' and 5'-AA AGA TCT TGG ATT CAG GAC AAC CAT GC-3' to give pColQ-1{Delta}NFAT-Luc.

Intramuscular DNA Injection—Intramuscular DNA injection was performed according to the method described by Gramolini et al. (34). In brief, 100 µg of luciferase-reporter plasmid and 50 µg of CMV promoter-driven {beta}-galactosidase plasmid in 100 µl of 0.9% NaCl was used in each injection. The SV40-driven luciferase, pSV40-Luc, was a control plasmid. The DNAs were intramuscularly injected into the soleus or tibialis muscles of anesthetized 2-month-old SD rats. Rats were from the Animal Care Facility at HKUST. All procedures conformed to the Guidelines of the Animal Research Panel of HKUST for the use and care of laboratory animals in research. A week after injection, the muscles were collected and homogenized (4 ml of ice-cold luciferase lysis buffer/g of tissue) using a Polytron homogenizer (Ultra-Turrax T25; Labortecknik) 3 times at 24,000 rpm for 1 min followed by centrifugation at 14,000 rpm for 10 min at 4 °C. The supernatants were collected for luciferase and {beta}-galactosidase assays.

cDNA Transfection and Drug Treatment—The purified DNAs encoding various constructs were purified by Maxiprep DNA purification columns (BD Biosciences Clontech), and they were transfected into cultured C2C12 myoblasts or chick myotubes by LipofectAMINE Plus transfection (Invitrogen). G418-resistant stable cell lines were selected in medium containing 400 µg/ml G418 (Geneticin) and used in analysis of expression profile during myogenic differentiation. For drug treatments myoblasts were transfected as above, allowed to fuse to myotubes, and treated with cyclosporin A (CsA; Calbiochem), Ca2+ ionophore A23187 [GenBank] (Sigma), recombinant neuregulin (Upstate Biotechnology Inc., Lake Placid, NY), or 2-MeSADP (Sigma) for 2 days. For treatment of cultured myotubes with 2-MeSADP, we took care to eliminate free ATP/ADP from the cultures, as described previously (32). Cultures were collected to perform luciferase, {beta}-galactosidase, and protein assays.

Gel Mobility Shift Assay—A nuclear extract of C2C12 myotubes was prepared and used to perform gel shift assay as described by Choi et al. (32). In brief, nuclear extract (2–5 µg of protein plus 2.5 µg of poly(dI-dC) per sample) was preincubated in binding buffer containing 2.5 mM dithiothreitol, 20% glycerol, 5 mM MgCl2, 2.5 mM EDTA, 250 mM NaCl, and 50 mM Tris-HCl (pH 7.5) for 20 min at room temperature. The samples were further incubated for 20 min with 32P-labeled double-stranded oligonucleotides (20-mers; 0.1 pmol) carrying the control NFAT binding site (NFATCon, is underlined) that was synthesized according to the elements found in the promoters of troponin I and myoglobin (23), 5'-GGC GTG GAG GAA AAA CTG TT-3' and 5'-CCG CAC CTC CTT TTT GAC AA-3', the NFAT site of the ColQ-1 promoter (NFATColQ-1, underlined), 5'-AAA AAA TTG GAA ATA TAA AA-3' and 5'-TTT TTT AAC CTT TAT ATT TT-3' (-642 to -630), and the mutated NFAT site ({Delta}NFATColQ-1, underlined), 5'-AAA AAA TTG AAT TCA TAA AA-3' and 5'-TTT TTT AAC TTA AGT ATT TT-3' (-642 to -630). To obtain a supershift the samples were incubated with an anti-NFATc1 antibody (Santa Cruz, with 1:100 dilution) in binding buffer for 20 min before the addition of the radioactive probe. Finally, the reaction mixtures were separated by electrophoresis in 4% polyacrylamide gels, which were then dried and subjected to autoradiography.

Sucrose Density Gradients—Separation of various molecular forms of AChE was performed by sucrose density gradient analysis, as previously described (3). Sucrose gradients (5–20%) were prepared on a 0.5-ml cushion of 60% sucrose in 12-ml polyallomer tubes. Sucrose gradients were in 10 mM HEPES (pH 7.5), 0.5% Triton X-100, 1 M NaCl. Cell extract (~0.1 ml) was loaded onto the gradient. Centrifugation was run for 16 h at 38,000 rpm in a Sorvall TH 641 rotor at 4 °C. Sedimentation markers (catalase, 11.4 S; {beta}-galactosidase, 16 S) were used for calibration of the gradient. About 40 fractions were collected for the determination of AChE enzymatic activity.

Other Assays—For Western blot analysis proteins were extracted from cultures and tissues by homogenization; protein (20 µg) was used to perform SDS-PAGE and Western blot analysis (32). ECL detection was carried out according to the manufacturer. Anti-troponin I slow isoform and fast isoform antibodies (1:500) were from Santa Cruz (Delaware Avenue, CA); anti-{alpha}-tubulin monoclonal antibody (1:5,000) was from Sigma. Protein concentrations were measured routinely by the Bradford method (35) with a kit from Bio-Rad. Luciferase assay was performed with a commercial kit (Tropix Inc., Bedford, MA). In brief, cell cultures were washed with phosphate-buffered saline and resuspended in 0.2% Triton X-100, 1 mM dithiothreitol, and 100 mM potassium phosphate buffer (pH 7.8). 30 µl of lysate per sample were used in luciferase assay. The luminescent reaction was quantified in a Tropix TR717 microplate luminometer, and the activity was expressed as absorbance (up to 560 nm) per mg of protein. The activity of AChE was assayed in medium containing 0.1 mM tetraisopropylpyrophosphoramide (iso-OMPA) as an inhibitor of butyrylcholinesterase (36). The luciferase activity was normalized by {beta}-galactosidase activity in the same amount of protein in each sample. Statistical tests were made by the PRIMER program; differences from basal or control values (as shown in the plots) were classified as highly significant (**) where p < 0.001.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
DISCUSSION
 REFERENCES
 
Two Distinct Promoters Drive the Differential Expression of ColQ Transcripts—The human COLQ gene produces two transcripts (ColQ-1 and ColQ-1a) from two distinct initiation sites; these transcripts differ by their first exons (exons 1 and 1a), which only encode signal sequences of the corresponding ColQ variants (21, 22). The distance between these two transcripts (~23 kb) and the fact that their expressions differ in their tissue dependence suggest that they are driven from two distinct promoters. To analyze the regulation of ColQ expression, ~2-kb sequences upstream of the ATG initiation codons of exon 1 and exon 1a were cloned from human genomic DNA using appropriate primers (Fig. 1A). The length of ColQ promoters was arbitrarily chosen to include putative N-box elements. These DNAs were sequenced and showed no difference from those of the corresponding regions of the COLQ gene deposited in GenBankTM. The upstream sequences of exon 1 and 1a were subcloned into pGL3 vectors upstream of a luciferase reporter, generating constructs pColQ-1-Luc and pColQ-1a-Luc, respectively (Fig. 1B). Sequence analysis indicated no sequence similarity between the upstream sequences of exon 1 and 1a. We identified consensus sequences for transcriptional factors E-box, NFAT, c-Ets, Elk-1, N-box, and MEF2, which are known to play a role in muscle-specific transcriptional activity (Fig. 1B). In addition, sequence alignment of these identified motifs with the known gene of troponin I revealed putative SURE and FIRE motifs in pColQ-1 and pColQ-1a, respectively.



View larger version (23K):
[in this window]
[in a new window]
  		FIG. 1.
Isolation of two ColQ promoters, pColQ-1 and pColQ-1a, and its putative regulatory motifs. A, human COLQ gene contains exon 1 and 1a, which drive the expression of ColQ-1 and ColQ-1a transcripts. Primers were designed to clone ~2-kb regions upstream of exon 1 and exon 1a, representing two distinct ColQ promoters, pColQ-1 and pColQ-1a. The 5'-untranslated region is shaded. B, the cloned sequences were analyzed by the TF search program, and putative binding motifs of transcription factors were identified: E-protein (E-box; CAN NTG), NFAT (GGA AA), c-Ets transcription factor (c-Ets; (C/A)GG A(A/T)), Elk-1, N-Box (CCG GAA), and MEF2 (CTA AAA ATA A). The putative SURE and FIRE motifs are indicated. The DNAs were subcloned upstream of the luciferase gene in pGL3 vectors as pColQ-1-Luc and pColQ-1a-Luc.

 
The promoter activities of pColQ-1-Luc and pColQ-1a-Luc were tested in transfected cells. In non-muscle cells, the promoter-driven luciferase activity was as low as that obtained with a pGL3 control construct. In contrast, luciferase was robustly expressed in mouse C2C12 and chick myotubes with either pColQ-1-Luc or pColQ-1a-Luc (Fig. 2A). In addition, the promoter activity was significantly higher for pColQ-1 than for pColQ-1a in the transfected myotubes, particularly in C2C12 myotubes. The differential expression of ColQ promoters in the myotubes could be due to the fiber type of the muscle cultures. Fig. 2B shows a Western blot indicating the levels of slow and fast isoforms of troponin I in rat soleus (a mostly slow muscle), tibialis (a mostly fast muscle), C2C12 myotubes, and chick C2C12 myotubes, expressed predominantly in the slow isoform of troponin I (~21 kDa), whereas chick myotube expressed both fast and slow isoforms at equal levels. Thus, C2C12 myotubes can be considered to be predominantly of the slow twitch muscle fiber type, and chick myotubes can be considered to be a mixture of slow and fast twitch muscle fibers. Thus, the differential expression of pColQ-1 and pColQ-1a in C2C12 and chick myotubes may be related to their fiber typing.



View larger version (32K):
[in this window]
[in a new window]
  		FIG. 2.
Expression of ColQ promoters in muscle by in vitro and in vivo transfections. A, constructs pColQ-1-Luc and pColQ-1a-Luc were transiently transfected into cultured cells. After 2 days, cells were collected for luciferase activity assay. Both ColQ promoters showed high expressions in myotube cultures but not in non-muscle cultures. MB, myoblast; MT, myotube. B, fiber type analysis of different myotube cultures. Total proteins (20 µg) from adult rat soleus, tibialis, C2C12 myotubes, and chick myotubes were analyzed by Western blot using antibodies against slow and fast isoforms of troponin I; both had a size of ~21 kDa. C, 100 µg of promoter plasmid and 50 µg of CMV promoter-driven {beta}-galactosidase plasmid resuspended in 0.9% NaCl were injected into soleus (slow fiber) or tibialis (fast fiber) of 2-month-old rats intramuscularly. After a week, the muscles were excised, and the luciferase activities were assayed. Deleted promoter constructs as shown here, pColQ-11500->1-Luc (A), pColQ-11000->1-Luc (B), pColQ-1500->1-Luc (C), pColQ-1a1500->1-Luc (D), pColQ-1a1000->1-Luc (E), pColQ-1a500->1-Luc (F), were generated and used similarly in the muscle DNA injection. Expression of SV40 promoter-driven luciferase reporter (pSV40-Luc) served as the control. Values are normalized to the activity of {beta}-galactosidase. In this and further figures showing luciferase activity, values are expressed as the ratio to the basal activity (obtained after parallel transfection with pGL3 vector alone) in a final lysate. Values are the means ± S.E. for five independent experiments, each with triplicate samples.

 
The two promoter constructs were also expressed in vivo by DNA transfection in rat fast muscle (e.g. tibialis) and slow muscle (e.g. soleus). A week after DNA injection, the entire muscle was collected for luciferase assay. Fig. 2C shows that the two ColQ promoters had distinct expression patterns in fast twitch and slow twitch muscles. The pColQ-1 construct was expressed considerably more actively in soleus than in tibialis, where expression was very low, so that the ratio between the two muscles could be as high as ~60-fold. In contrast, expression of the pColQ-1a-Luc construct was ~12-fold higher in tibialis than in soleus, where it was very low. The activity of pColQ-1 in slow twitch muscle was ~5-fold higher than that of pColQ-1a in fast twitch muscle; this difference is consistent with the fact that endogenous ColQ transcripts are more abundant in slow muscle than in fast muscle (21). The distinct expression of pColQ-1-Luc and pColQ-1a-Luc was not due to a difference in transfection efficiency because all expression levels were normalized to that of a control vector containing the {beta}-galactosidase gene downstream of a constitutively active CMV promoter.

To analyze the promoters we introduced a set of nested deletions upstream of a luciferase gene in pColQ-1-Luc and pColQ-1a-Luc, as shown in Fig. 2C. Transfection of these constructs was performed in soleus and tibialis by in vivo DNA injection. The deleted constructs showed a reduced promoter activity and a reduced difference between slow and fast muscles compared with the original constructs. This analysis showed that the ~2-kb ColQ promoters were sufficient to drive specific expression of ColQ transcripts according to the fiber type both in vivo and in vitro. Thus, the preferential expressions of ColQ-1 transcript in slow muscle and of ColQ-1a transcript in fast muscle appear to have resulted from the differential activities of the corresponding promoters, pColQ-1 and pColQ-1a, in slow twitch and fast twitch muscles.

Expression of ColQ Promoters during Myogenic Differentiation—To distinguish ColQ-1 and ColQ-1a transcripts in C2C12 myotubes, the sequences of exon 1 and 1a were cloned from mouse genomic DNA; they presented a strong homology with rat and human sequences (Fig. 3A). Specific primers flanking exons 1 and 1a were then used for RT-PCR, generating fragments of ~245 bp for exon 1, ~191 bp for exon 1a, and ~650 bp for glyceraldehyde-3-phosphate dehydrogenase, which was used as a control. We found that ColQ-1 and ColQ-1a transcripts were both low at the myoblast stage (Fig. 3B, upper panel). The ColQ-1 transcript markedly increased by >5-fold upon fusion to myotubes. In contrast, the ColQ-1a transcript remained low during the myogenic differentiation of C2C12 cells (Fig. 3B, lower panel).



View larger version (55K):
[in this window]
[in a new window]
  		FIG. 3.
Expression profiles of ColQ-1 and ColQ-1a transcripts during the myogenic differentiation of C2C12 cells. A, exon 1 and exon 1a of the COLQ gene were cloned from mouse muscle cDNAs. The predicted coding sequences of exon 1 and exon 1a as well as part of common exon 2 are shown for comparison with human and rat sequences. Putative signal peptides are in lowercase. B, the endogenous transcripts of ColQ-1 and ColQ-1a in C2C12 myoblast and myotube were determined by RT-PCR. The PCR product of glyceraldehyde-3-phosphate dehydrogenase served as the control. The lower panel shows the results of quantitation of such blots by densitometry. Values are in arbitrary units, relative to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) levels. C, total RNAs (20 µg) from C2C12 cells at different stages of myotube formation (Day 0 to Day 7) were used in each lane. ColQ-1 cDNA was used as probe in the Northern analysis. Transcripts encoding the AChE catalytic subunit and ribosomal RNA loading markers are shown. The lower panel shows the results of quantitation of such blots by densitometry. Values are expressed as ratios to the basal level (Day 0 of culture). D, sedimentation analysis of AChE molecular forms during myogenic differentiation of C2C12 cells. Sedimentation markers are indicated. In all cases, values are the means ± S.E. for five independent experiments, each with triplicate samples.

 
The expression of transcripts encoding ColQ-1 and ColQ-1a was further determined by Northern blot during the myogenic differentiation of C2C12 cells in vitro. A mouse cDNA corresponding to the entire coding sequence of ColQ-1, which did not distinguish ColQ-1 and ColQ-1a transcripts, was used as a probe to determine the total expression of the COLQ gene during the formation of C2C12 myotubes. Analysis by RT-PCR indicated that the ColQ-1 species was predominantly expressed in C2C12 cells. In Northern analysis, ColQ expression was undetectable in the early stages of myotube formation. A single transcript of ~4.2 kb became detectable 3 days after the myoblasts began to fuse into myotubes, and it reached a maximum at day 6 (Fig. 3C, upper panel). This profile of expression during myotube formation appears to parallel the increase in expression of AChE catalytic subunits, which are encoded by two transcripts of ~2.5 and ~3.6 kb (Fig. 3C). Besides, the expression of ColQ during myogenesis also explains the observed changes of AChE molecular forms. In sucrose density gradient analysis, the A forms of AChE were not detected until a late stage of myotube formation, when ColQ expression reached a maximum (Fig. 3D).

To further analyze the activity of the two ColQ promoters during muscle differentiation, pColQ-1-Luc and pColQ-1a-Luc were stably transfected into C2C12 myoblasts, which were induced to differentiate. The expression of luciferase driven by the pColQ-1 promoter was low in myoblasts, started to increase at day 4, and reached a maximum at a late stage of myotube formation (Fig. 4A). In contrast, expression driven by pColQ-1a remained low throughout the differentiation process. The distinct expression patterns observed for pColQ-1 and pColQ-1a in C2C12 cells are, therefore, consistent with the levels of endogenous ColQ transcripts, as revealed by RT-PCR. Moreover, the expression profile of pColQ-1-Luc showed a close similarity with that of a construct driven by the AChE promoter, pAChE-Luc (Fig. 4A; see also Ref. 27).



View larger version (43K):
[in this window]
[in a new window]
  		FIG. 4.
The activities of ColQ promoters during muscle differentiation. A, stably transfected C2C12 cells expressing pColQ-1-Luc, pColQ-1a-Luc, and pAChE-Luc (~2.2 kb of human AChE catalytic subunit promoter in pGL3 vector; see Ref. 28) were generated, and they were induced to differentiate. Values are expressed as ratios to the basal level (Day 0 of culture). B, deleted promoter constructs of pColQ-1 and pColQ-1a (the nomenclature is shown in Fig. 2) were used to transfect C2C12 cells. Luciferase activities were compared for different constructs at myoblast and myotube stages, and they were normalized to {beta}-galactosidase activity. C, in C2C12 myotubes, pColQ-1-Luc or pColQ-1a-Luc was co-transfected with equal amounts of myoD and myogenin cDNAs. After 2 days, cell lysates were collected for luciferase assay. Values are expressed as ratios to the basal activity. The construct p4xRE-Luc served as a positive control. In all cases, values are the means ± S.E. for five independent experiments, each with triplicate samples. **, p < 0.001 as compared with the control pcDNA3 only.

 
Promoters of different lengths were also tested during C2C12 myotube formation, as shown in Fig. 4B; pColQ-1-Luc showed a higher level of expression than pColQ-1a-Luc, and both ~2-kb constructs showed a more significant increase during fusion of myoblasts to myotubes than shorter constructs. A comparison of the expression of endogenous ColQ transcripts with that of ColQ promoters in C2C12 cells suggests that the ~2-kb upstream fragments used in pColQ-1 and pColQ-1a reproduce the actual expression of the two ColQ transcripts in myotubes.

To determine the role of muscle-specific regulatory elements in the ColQ promoters, cDNAs encoding myoD and myogenin were co-transfected with pColQ-1-Luc or pColQ-1a-Luc in cultured C2C12 myotubes. A luciferase construct driven by four E-box-responsive elements (p4xRE-Luc), used as a positive control, was activated at least 10–20-fold by the co-expression with either myoD or myogenin. The activities of pColQ-1-Luc and pColQ-1a-Luc were induced by at least 2–4-fold when they were co-transfected with either myoD or myogenin (Fig. 4C). MyoD and myogenin activated pColQ-1-Luc more strongly than pColQ-1a-Luc, indicating a difference between the two promoters.

Fiber Type-specific Expression Patterns of ColQ Transcripts—Regulatory motifs named SURE and FIRE have been shown to control the transcription of genes encoding contractile proteins (e.g. troponin I) in slow and fast twitch skeletal muscles, respectively. SURE and FIRE consist of different conserved DNA regulatory elements. We examined whether such elements exist in the ColQ promoters and might explain their fiber type specificity. Analysis of the promoter sequences revealed putative SURE in pColQ-1 and FIRE in pColQ-1a (Fig. 5A). The role of these regulatory elements in the fiber type-specific expression of ColQ transcripts was analyzed by deletion. The resulting constructs, pColQ-1{Delta}SURE-Luc and pColQ-1a{Delta}FIRE-Luc, were injected in vivo into rat soleus and tibialis muscles. These deleted constructs did not present the differential expression patterns observed with pColQ-1 and pColQ-1a in slow and fast muscles (Fig. 5B).



View larger version (25K):
[in this window]
[in a new window]
  		FIG. 5.
Putative SURE and FIRE motifs in ColQ promoters direct the fiber type-specific expression pattern in rat muscles. A, a sequence search revealed putative SUREColQ-1 and FIREColQ-1a motifs in pColQ-1 and pColQ-1a, respectively. Key regulatory elements are indicated. Similar sequences from the promoters of slow (SURETnls) and fast (FIRETnlf) isoforms of troponin I are shown for comparison. B, mutants, which were deleted in either SURE or FIRE motifs of the ColQ promoters (pColQ-1{Delta}SURE-Luc or pColQ-1a{Delta}FIRE-Luc), were generated. The promoter plasmid and CMV promoter-driven {beta}-galactosidase plasmid were co-injected into soleus or tibialis as in Fig. 2. Values are expressed as the ratios to the basal activity (transfection with pGL3 vector) in the final lysate and in the means ± S.E. for five independent experiments, each with triplicate samples.

 
The DNA regions corresponding to SURE and FIRE were amplified by PCR from the human ColQ promoters, and fragments corresponding to rat or quail troponin I promoter were generated. These DNA fragments were placed upstream of a luciferase gene, generating the following constructs: pSUREColQ-1-Luc containing SURE from pColQ-1, pSURETnls-Luc containing SURE from the promoter of the slow isoform of troponin I, pFIREColQ-1a-Luc containing FIRE from pColQ-1a, and pFIRETnlf-Luc from the promoter of troponin I fast isoform (Fig. 6A). These constructs were used for in vivo DNA transfection in soleus and tibialis muscles to determine their fiber type-specific expression patterns. The SURE-containing constructs showed a high expression in soleus but not in tibialis, whereas FIRE-containing constructs showed opposite responses (Fig. 6A). This was true for constructs derived from both troponin I and ColQ promoters.



View larger version (35K):
[in this window]
[in a new window]
  		FIG. 6.
The SURE and FIRE motifs of ColQ promoters are sufficient to drive muscle fiber type-specific expression. A, approximately 200-bp regions corresponding to SURE and FIRE motifs from pColQ-1 and pColQ-1a, respectively, were cloned into pTA-Luc vectors upstream of a luciferase reporter gene to generate pSUREColQ-1-Luc and pFIREColQ-1a-Luc. In parallel, similar constructs were generated from the SURE and FIRE elements of troponin I slow (rat) and fast (quail) isoforms as pSURETnIs-Luc and pFIRETnIf-Luc. The plasmids containing SURE or FIRE were co-injected with a plasmid expressing {beta}-galactosidase under the CMV promoter into soleus or tibialis as in Fig. 2. pSV40-Luc served as the control. B, SURE-containing plasmids, pColQ-1-Luc, pSUREColQ-1-Luc, and pSURETnIs-Luc were transfected transiently into cultured chick myotubes. Two days later the transfected myotubes were treated for 16 h with the Ca2+ ionophore A23187 [GenBank] (100 µM) alone or with calcineurin inhibitor (CsA; 250 nM). C, FIRE-containing plasmids pColQ-1a-Luc, pFIREColQ-1-Luc, and pFIRETnIf-Luc were transfected transiently into cultured chick myotubes. Drugs were applied as in B but in a different order. In all cases values are expressed as the ratio to the basal activity (transfection with pGL3 vector or no drug treatment) in a final lysate; the values are the means ± S.E. of five independent experiments, each with triplicate samples.

 
We also studied the responsiveness of these constructs to calcineurin, which characterizes the muscle fiber type-specific expression pattern of contractile proteins; calcineurin selectively activates slow fiber type-specific promoters, whereas inhibition of calcineurin activity by administration of CsA promotes the transformation of slow fiber to fast fiber. In contrast, the application of Ca2+ ionophore A23187 [GenBank] increased the intracellular Ca2+ level, then subsequently activated the calcineurin. The Ca2+-induced fast to slow transformation of the cultured muscle cells is comparable with the effect of low frequency stimulation on fast skeletal muscle in vivo and causes the changes of gene expression (37). In transfected cultured chick myotubes expressing pColQ-1-Luc, the Ca2+ ionophore A23187 [GenBank] stimulated the expression by ~3-fold, and this effect was blocked by CsA (Fig. 6B). We observed similar responses to A23187 [GenBank] and CsA in chick myotubes expressing pSUREColQ-1-Luc or pSURETnls-Luc, which was used as a control. In contrast, the promoter activity of pColQ-1a was induced by CsA and blocked by A23187 [GenBank] . As expected, transfected myotubes expressing pFIRETnlf-Luc, used as a control, or pFIREColQ-1a-Luc showed the same responses to A23187 [GenBank] and CsA (Fig. 6C).

NFAT is known as a CsA-sensitive transcription element that mediates calcineurin induction of gene expression in slow muscle fibers. NFAT exists as four isoforms: NFATc, NFATp, NFAT4/x/c3, and NFAT3 (24, 38, 39); skeletal muscles express NFATc, NFATp, and NFAT4/x/c3 (40). A search in pColQ-1 revealed a possible NFAT binding site within the SURE motif (see Figs. 1B and 7A). To authenticate the NFAT binding site located in pColQ-1, we synthesized an oligonucleotide containing the corresponding element, called NFATColQ-1 (Fig. 7A). The NFAT sequence derived from troponin I (NFATCon) was used as a control. Nuclear extracts from cultured C2C12 myotubes were probed with 32P-labeled double-stranded NFATColQ-1 or NFATCon in gel mobility shift assays. We observed a concentration-dependent binding of these probes (Fig. 7B). Nonspecific binding was avoided by (i) always including a large excess of a nonspecific competitor, double-stranded poly(dI-dC), and (ii) demonstrating a progressive increase of the binding with increasing the amounts of the nuclear extract (Fig. 7B). To further test the specificity of NFAT binding, the NFATColQ-1 sequence was mutated to change six consecutive nucleotides to form {Delta}NFATColQ-1, as specified in Fig. 7A. The {Delta}NFATColQ-1-mutated oligonucleotide showed no gel mobility shift with a myotube nuclear extract, even at high concentration (Fig. 7B). A portion of the binding complex that contained the nuclear extract-bound 32P-labeled NFATColQ-1 and NFATCon was "supershifted" by incubating with an anti-NFAT antibody (Fig. 7C), which suggested that part of the bound nuclear extract was a member of the NFAT protein family. However, the identity of the remaining bound nuclear extract, which did not show any supershift by the anti-NFAT antibody, was not determined. The possibilities of other transcriptional factors that bound to the synthetic DNAs, therefore, could not be eliminated.



View larger version (41K):
[in this window]
[in a new window]
  		FIG. 7.
Identification of NFAT binding site in human pColQ-1. A, potential binding sites of NFAT, as indicated by boxes, was found within the SURE motif of the ~2.0-kb human pColQ-1 DNA. An oligonucleotide corresponding to the NFAT sequence was synthesized as NFATColQ-1. The mutated sequence {Delta}NFATColQ-1 and the consensus sequence of NFAT (shaded) from troponin I, denoted as NFATCon, were synthesized as controls. B, nuclear extract (Nu. Ex.) of C2C12 myotube was incubated with 32P-labeled double-stranded oligonucleotides (0.1 pmol/sample) covering NFATCon, NFATColQ-1,or {Delta}NFATColQ-1 in gel mobility shift assays. Assays were performed by including increasing amounts of the nuclear extracts (from 2 to 6 µg of protein, indicated above the gels). The positions of the free probe and of the bound (shifted) probe are indicated by arrowheads. C, anti-NFAT antibody (Ab; 1:100 dilution) was preincubated before the assay as in B, with the highest amount of nuclear extract used in B to further supershift the specific complexes.

 
Mutation of the NFAT element was performed in pColQ-1, to give the construct pColQ-1{Delta}NFAT-Luc (Fig. 8A). Expression of pColQ-1{Delta}NFAT-Luc in rat muscles by in vivo DNA injection showed that this mutation completely abolished the fiber type-specific expression pattern of pColQ-1-Luc (Fig. 8A). We found that a luciferase construct driven by three repeated NFAT binding sites (p3x(NFAT)-Luc), used as a control, was expressed in slow twitch muscle but not in fast twitch muscle and that a construct driven by the constitutively active SV40 promoter was expressed equally in both muscles (Fig. 8A). In addition, we examined the responsiveness of pColQ-1{Delta}NFAT-Luc to Ca2+ ionophore A23187 [GenBank] in transfected chick myotubes, whereas the activation was markedly reduced compared to pColQ-1-Luc or p3x(NFAT)-Luc, indicating a crucial role of NFAT in directing expression of ColQ-1 in slow muscle fibers (Fig. 8B). The roles of NFAT elements outside the SURE of pColQ-1 have not been determined; however, they may not involve in the fiber type-specific gene expression.



View larger version (18K):
[in this window]
[in a new window]
  		FIG. 8.
Mutation of the NFAT motif of pColQ-1 blocks its fiber type-specific expression pattern in muscle. A, the ~2.0-kb pColQ-1 DNA was mutated at its NFAT binding sites, as shown in Fig. 7A, and tagged with luciferase reporter gene as pColQ-1{Delta}NFAT-Luc. Plasmids pColQ-1-Luc and pColQ-1{Delta}NFAT-Luc were co-injected with CMV promoter-driven {beta}-galactosidase plasmid into soleus or tibialis as in Fig. 2. B, similar promoter constructs were transfected into cultured chick myotubes. Two days later A23187 [GenBank] (100 µM) was applied to the transfected myotubes for 16 h. Values are expressed as the ratios to the basal activity (transfection with pGL3 vector or no drug treatment) and represent the means ± S.E. of five independent experiments, each with triplicate samples. **, p < 0.001 as compared with the response of pColQ-1-Luc.

 
Synapse-specific Expression of ColQ Transcripts—The A forms of AChE, which are associated with ColQ, are restricted to the nmjs in fast twitch muscles but not in slow twitch muscles (2, 21). We studied whether the N-box element and the c-Ets-related transcription factor GABP might be involved in the synapse-specific expression of ColQ. Sequence analysis revealed the existence of putative N-box elements in both pColQ-1 and pColQ-1a (Fig. 9A), suggesting that these promoters could be activated by GABP and neuregulin. In cultured chick myotubes co-transfection of cDNAs encoding GABP {alpha} or {beta} with the ColQ promoter constructs induced a 3–5-fold up-regulation of the luciferase activity produced by pColQ-1-Luc, an induction similar to that of the AChR {epsilon}-subunit promoter (pAChR{epsilon}-Luc). Deleted mutants of pColQ-1-Luc were tested for their response to GABP activation. The mutant pColQ-11500->1-Luc, which was deleted by ~500 bp from its 5' end but retained the N-box element, showed an increased induction when it was co-expressed with GABP {alpha} or {beta} (>10–20-fold), suggesting a suppressive role of the deleted region (Fig. 9A). Shorter mutants from which the N-box element was deleted, such as pColQ-11000->1-Luc and pColQ-1500->1-Luc, were not activated by GABP.



View larger version (38K):
[in this window]
[in a new window]
  		FIG. 9.
Activation of ColQ promoters by synaptic signaling molecules. A, deleted promoter constructs of pColQ-1-Luc and pColQ-1a-Luc (the nomenclature is shown in Fig. 2) were generated and used in DNA transfection. In cultured chick myotubes these constructs were co-transfected with a vector expressing mammalian GABP {alpha} or {beta}. The promoter construct of AChR {epsilon}-subunit (pAChR{epsilon}-Luc) served as the control. After 2 days of transfection luciferase assay was performed. Values are expressed as ratios to the basal activity (co-transfection with pcDNA 3 vector) in a final lysate. B, different promoter constructs were transfected into cultured chick myotubes for 2 days. The cells were treated with neuregulin (NRG; 3 nM) or 2-MeSADP (50 µM) for 2 days. The cell lysates were collected for luciferase assay. C, promoter constructs were co-transfected with RafCAAX (constitutively active mutant of Raf-1) or RafWT (wild type) cDNA for 2 days. Lysates were collected as above. Values are expressed as ratios to the basal activity (no drug or co-transfection with a control vector) in a final lysate and represent the means ± S.E. for five independent experiments, each with triplicate samples. **, p < 0.001 as compared with the basal activity.

 
In parallel experiments we found that GABP activated pColQ-1a-Luc ~8-fold, more strongly than pColQ-1. Again, deleted mutants without a N-box element, such as pColQ-1a1500->1-Luc, pColQ-1a1000->1-Luc, and pColQ-1a500->1-Luc, showed no activation by GABP (Fig. 9A). These results suggest that an N-box element activates the expression of both ColQ-1 and ColQ-1a at synapses, although the response of pColQ-1 was lower than that of pColQ-1a. We also tested the response of ColQ promoters to neuregulin, which induces synapse-specific gene expression at nmjs. In chick myotubes transfected with the promoter constructs, the application of neuregulin (3 nM) activated the promoter of AChR {epsilon}-subunit (pAChR{epsilon}-Luc) by ~5-fold as well as a construct in which luciferase expression was driven by three repeated N-box elements (p3x(N-box)-Luc) (Fig. 9B). In parallel, neuregulin induced the activities of pColQ-1-Luc and pColQ-1a-Luc by ~2.5-fold.

We challenged ColQ promoters with another synapse-specific inducer, ATP, which was demonstrated to regulate the promoter activity of AChE catalytic subunit by activating P2Y receptors on the post-synaptic membrane of the nmjs (26, 32). When chick myotubes expressing pColQ-1-Luc or pColQ-1a-Luc were treated with a specific agonist of the P2Y1 nucleotide receptor, 2-MeSADP, the promoter activity was increased more than 2-fold (Fig. 9B). As in the case for neuregulin, pColQ-1a was more sensitive to 2-MeSADP than pColQ-1.

Raf-1 is a downstream effector of neuregulin and/or of P2Y receptor signaling (32), and we also determined its role in activation of ColQ promoters. When the constitutively active and membrane-targeted mutant RafCAAX (41) was co-expressed with different promoter constructs in chick myotubes, it activated pColQ-1-Luc ~3-fold, pColQ-1a-Luc ~10-fold, and the control construct pAChR{epsilon}–Luc ~8-fold (Fig. 9C). In contrast, wild type Raf-1 (RafWT) did not significantly activate expression of these constructs. Neuregulin and GABP activated pColQ-1a, markedly stronger than pColQ-1; these results suggest that neuregulin exerts a stronger effect in fast twitch muscle than in slow twitch muscle.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
REFERENCES
 
The present findings show that specific regulatory elements in the two ColQ promoters, pColQ-1 and pColQ-1a, explain the expression of the ColQ protein and, therefore, of collagen-tailed AChE forms in slow and fast muscles. We find that these two independent promoters mediate the muscle fiber type-specific expression pattern of ColQ transcripts; pColQ-1 drives the expression of the ColQ-1 transcript in slow twitch muscles, whereas pColQ-1a is responsible for expression of the ColQ-1a transcript in fast twitch muscles. Both ColQ promoters contain E-box elements that control their muscle-specific expression and explain that it increases during the myogenic differentiation of C2C12 cells. This muscle fiber type-specific expression pattern is regulated by DNA sequences that present homology with the SURE and FIRE motifs, which control the muscle fiber type-specific expression of several muscular contractile proteins. These unique regulatory elements are also responsible for the calcineurin-mediated regulation of ColQ expression in muscle fibers. In particular, the presence of a NFAT element within the SURE motif of pColQ-1 is crucial in directing the slow fiber type-specific expression of ColQ-1; mutation of this element in pColQ-1 completely abolished the expression of ColQ-1 transcript in slow twitch muscle and its sensitivity to CsA in transfected myotubes. Finally, the synapse-specific expression pattern of ColQ transcripts is mediated by N-box element, as previously demonstrated for the synaptic expression of AChR subunits (42) and AChE (43) at nmjs; both ColQ promoters contain N-box elements and are activated by neuregulin in cultured chick myotubes.

SURE and FIRE motifs consist of closely grouped DNA regulatory elements, CAGG, CCAC, MEF2, and E-box (23–25); in addition, the SURE motif contains an NFAT element between MEF2 and E-box (23). These unique arrangements of DNA regulatory elements are sufficient to determine muscle fiber type-specific expression patterns (23, 30). Calcineurin, a Ser/Thr phosphatase activated by Ca2+-calmodulin, is also known to be involved in the maintenance of the slow muscle fiber-specific gene expression, and suppression of its activity causes a transformation from slow to fast fiber type-specific gene expression (38, 39). We identified SURE and FIRE motifs in the pColQ-1 and pColQ-1a promoters, respectively. Although these motifs present some differences with those previously reported and, in particular, do not contain CAGG and CCAC sequences upstream of MEF2 (see Fig. 5A), we showed that they controlled the fiber type-specific expressions of ColQ-1 and ColQ-1a and their response to calcineurin.

We did not analyze post-transcriptional events, which may also contribute to the muscle fiber type-specific expression patterns of ColQ subunits. The levels of AChE in fast muscle (18) and of utrophin (34) largely depend on the stability of their transcripts; AChE transcripts are stabilized by activation of calcineurin during the formation of C2C12 myotubes (44).

Physiologically, the hydrolysis of acetylcholine at nmjs depends on AChE collagen-tailed forms (5). Several lines of evidence indicate that the unique expression pattern of these molecular forms in muscle depends primarily on the regulation of ColQ expression. In fast twitch muscles, the collagen tail of the A12 AChE form is composed of ColQ-1a subunits, and it is exclusively localized at the nmjs. In contrast, slow twitch muscles contain A8 and A4 forms that possess ColQ-1 subunit and are relatively abundant in the extra-junctional regions of muscle fibers as well as at the nmjs (2, 21). The assembly, processing, and externalization of collagen-tailed AChE forms do not depend on a specialized post-synaptic apparatus and can occur in non-specialized cells (4, 45). The ColQ-1 and ColQ-1a proteins differ essentially in their signal sequences and a few amino acids in their N termini (see Fig. 3A), which may influence the production of the protein and the formation of AChE oligomers, perhaps by targeting the ColQ-1 and ColQ-1a subunits to different ER sub-compartments; this might explain that in mRNA-injected Xenopus oocytes, the co-expression of AChE with ColQ-1a produced a higher level of collagen-tailed molecules than with ColQ-1 (21). In addition, the upstream sequence of PRAD was shown to affect the assembly of AChE oligomers (10).

The ratio of AChE and ColQ subunits probably determines the composition of collagen-tailed forms in slow and fast twitch muscles (17, 21). A relatively high abundance of ColQ subunits compared with catalytic AChE subunits, thus, explains the formation of A8 and A4 AChE in slow twitch muscles. In contrast, the production of the A12 form is favored in fast twitch muscle because of an excess of AChE catalytic subunits over ColQ subunits. This assumes that the assembly of the two types of subunits is a random process, which is not influenced by the different signal peptides of ColQ-1 and ColQ-1a.

The N-box element of the ColQ-1a promoter could play a key regulatory role in directing the production of ColQ-1a and, therefore, of A12 AChE at nmjs of fast twitch muscles in the same way as such elements control the synaptic expression of other proteins (42). Although the ColQ-1 promoter, which drives the expression of ColQ-1 in slow muscles, also contains N-box element, the ColQ-1-containing A8 and A4 form AChE are found extra-junctionally in slow muscles. Two possible reasons could account for this discrepancy. First, despite its N-box element, the ColQ-1 promoter is much less responsive to the synaptic activators, GABP, neuregulin, and 2-MeSDAP, than the ColQ-1a promoter. As shown in our deletion analysis, the ~500-bp region located upstream of the N-box element in the ColQ-1 promoter appears to contain a suppressive element(s) that reduces the effect of synaptic inducers. In line with this hypothesis, the neuregulin-mediated activation of a kinase transduction pathway (p70S6 kinase) has been proposed to be responsible for the repression of extrasynaptic AChR at the nmjs (46). Secondly, activation of the Ca2+/calmodulin-activated protein kinase II (47) and Ser/Thr kinase (48) have been shown to mediate the activity-dependent extrasynaptic repression of AChR, although the regulatory elements responsible for this effect have not been identified. This extrasynaptic repression mediated by nerve activity may be less effective for ColQ-1 than for ColQ-1a. This possible regulatory mechanism now deserves further exploration.

Our results show that the expressions of ColQ and AChE proteins are regulated in parallel during myogenic differentiation and during the formation of nmjs. The regulatory elements of AChE promoters have been analyzed in detail in human, rodent, and Torpedo (see Ref. 28 and references therein). Several regulatory elements that have been found to control AChE expression also exist in the ColQ promoters, suggesting that they may contribute to a co-regulation of AChE and ColQ in muscles under different physiological states. E-box elements, which are present in AChE, ColQ-1, and ColQ-1a promoters may be responsible for the muscle-specific expression of these proteins as well as for the increase of transcriptional activity observed during myogenesis; we show here that the activity of the ColQ-1 promoter increases during formation of C2C12 myotube in parallel with the expression of AChE. The AChE and ColQ-1a promoters contain N-box elements. The presence of N-box elements in the promoter of rodent AChE suggests that it could be activated by neuregulin, although neuregulin did not increase the level of AChE transcripts in chick myotubes (49).

ATP, which is released at synapses, activates transcription of the AChE gene in cultured myotubes through P2Y receptors and mitogen-activated protein kinase signaling (26); this activation has been shown to involve Elk-1 regulatory elements located on the human AChE promoter (32). Because the ColQ-1 and ColQ-1a promoters also contain several possible Elk-1 binding elements, these may also mediate gene activation by synaptic ATP.


   FOOTNOTES
 
* This work was supported by the Research Grants Council of Hong Kong (HKUST 6098/02M and 6283/03M) (to K. W. K. T.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Back

§ Both authors contributed equally to this study. Back

Supported by a post-doctoral matching fund from Hong Kong University of Science and Technology. Back

|| Holds a Croucher Foundation Scholarship. Back

{ddagger}{ddagger} To whom correspondence should be addressed: Dept. of Biology, The Hong Kong University of Science and Technology, Clear Water Bay Rd., Kowloon, Hong Kong SAR, China. Tel.: 852-2358-7332; Fax: 852-2358-1559; E-mail: botsim{at}ust.hk.

1 The abbreviations used are: AChE, acetylcholinesterase; NFAT, nuclear factor of activated T cells; HKUST, Hong Kong University of Science and Technology; MEF2, myocyte enhancer factor-2; WT, wild type; SURE, slow upsteam regulatory element; FIRE, fast intronic regulatory element; nmjs, neuromuscular junctions; kb, kilobase(s); GABP, growth-associated-binding protein; RT, reverse transcription; CMV, cytomegalovirus; iso-OMPA, tetraisopropylpyrophosphoramide; CsA, cyclosporin A; 2-MeSADP, 2-methylthioadenosine 5'-diphosphate. Back


   ACKNOWLEDGMENTS
 
We are grateful to Dr. Tina Dong and H. Y. Choi from our laboratory for expert technical assistance to Dr. Suzanne Bon of Ecole Normale Supérieure for fruitful comments during the study.



   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Massoulié, J., Pezzementi, L., Bon, S., Krejci, E., and Vallette, F.-M. (1993) Prog. Neurobiol. (N.Y.) 41, 31-91
  2. Massoulié, J. (2002) Neurosignals 11, 130-143[CrossRef][Medline] [Order article via Infotrieve]
  3. Tsim, K. W. K., Randall, W. R., and Barnard, E. A. (1988) Proc. Natl. Acad. Sci. U. S. A. 85, 1262-1266[Abstract/Free Full Text]
  4. Krejci, E., Coussen, F., Duval, N., Chatel, J.-M., Legay, C., Puype, M., Vandekerckhove, J., Cartaud, J., Bon, S., and Massoulié, J. (1991) EMBO J. 10, 1285-1293[Medline] [Order article via Infotrieve]
  5. Feng, G., Krejci, E., Molgo, J., Cunningham, J. M., Massoulié, J., and Sanes, J. R. (1999) J. Cell Biol. 144, 1349-1360[Abstract/Free Full Text]
  6. McMahan, U. J., Sanes, J. R., and Marshall, L. M. (1978) Nature 271, 172-174[CrossRef][Medline] [Order article via Infotrieve]
  7. Peng, H. B., Xie, H., Rossi, S. G., and Rotundo, R. L. (1999) J. Cell Biol. 145, 911-921[Abstract/Free Full Text]
  8. Krejci, E., Thomine, S., Boschetti, N., Legay, C., Sketelji, J., and Massoulié, J. (1997) J. Biol. Chem. 272, 22840-22847[Abstract/Free Full Text]
  9. Donger, C., Krejci, E., Serradell, A. P., Eymard, B., Bon, S., Nicole, S., Chateau, D., Gary, F., Fardeau, M., Massoulié, J., and Guicheney, P. (1998) Am. J. Hum. Genet. 63, 967-975[CrossRef][Medline] [Order article via Infotrieve]
  10. Bon, S., Coussen, F., and Massoulié, J. (1997) J. Biol. Chem. 272, 3016-3021[Abstract/Free Full Text]
  11. Deprez, P., Inestrosa, N. C., and Krejci, E. (2003) J. Biol. Chem. 278, 23233-23242[Abstract/Free Full Text]
  12. Sketelj, J., and Brzin, M. (1985) J. Neurosci. Res. 14, 95-103[CrossRef][Medline] [Order article via Infotrieve]
  13. Legay, C., Huchet, M., Massoulié, J., and Changeux, J.-P. (1995) Eur. J. Neurosci. 7, 1803-1809[CrossRef][Medline] [Order article via Infotrieve]
  14. Silman, I., Lyles, J. M., and Barnard, E. A. (1978) FEBS Lett. 94, 166-170[CrossRef][Medline] [Order article via Infotrieve]