the kyphoscoliosis (ky) mouse is deficient in hypertrophic

© 2001 Oxford University Press Human Molecular Genetics, 2001, Vol. 10, No. 1 9–16

ARTICLE

The kyphoscoliosis (ky) mouse is deficient inhypertrophic responses and is caused by a mutationin a novel muscle-specific proteinGonzalo Blanco1,+, Gary R. Coulton2, Andrew Biggin2, Christopher Grainge2, Jill Moss3,Michael Barrett3, Anne Berquin4, Georges Maréchal4, Michael Skynner5, Peter van Mier6,Athena Nikitopoulou1, Manfred Kraus1, Chris P. Ponting7, Roger M. Mason2 andSteve D.M. Brown1

1MRC Mammalian Genetics Unit and UK Mouse Genome Centre, Harwell, Oxon OX11 ORD, UK, 2MolecularPathology Group, Division of Biomedical Sciences and 3Division of Investigative Sciences, Imperial College School ofMedicine, Imperial College of Science, Technology and Medicine, London SW7 2A, UK, 4Departément dePhysiologie, Faculté de Médicine, Université de Louvain 5540, 1200 Brussels, Belgium, 5Pfizer Global Research andDevelopment, University of Cambridge Forvie Site, Robinson Way, Cambridge CB2 2QB, UK, 6Department ofAnatomy and Neurobiology, Washington University School of Medicine, St Louis, MO 63110, USA and 7MRCFunctional Genetics Unit, University of Oxford, Department of Human Anatomy and Genetics, South Parks Road,Oxford OX1 3QX, UK

Received 4 October 2000; Revised and Accepted 6 November 2000 DDBJ/EMBL/GenBank accession no. AJ293727

The ky mouse mutant exhibits a primary degenerative myopathy preceding chronic thoraco-lumbar kyphoscolio-sis. The histopathology of the ky mutant suggests that Ky protein activity is crucial for normal muscle growthand function as well as the maturation and stabilization of the neuromuscular junction. Muscle hypertrophy inresponse to increasing demand is deficient in the ky mutant, whereas adaptive fibre type shifts take place. Theky locus has previously been localized to a small region of mouse chromosome 9 and we have now identified thegene and the mutation underlying the kyphoscoliotic mouse. The ky transcript encodes a novel protein thatis detected only in skeletal muscle and heart. The identification of the ky gene will allow detailed analysis ofthe impact of primary myopathy on idiopathic scoliosis in mice and man.

INTRODUCTION

As many as 1–2% of the population at 16 years of age showabnormal spinal curvature, with ∼0.6% of these requiringorthotic or surgical correction (1). Concordance for adolescentidiopathic scoliosis (AIS) in monozygotic twins has beenfound to be significant but not complete, supporting a geneticaetiology for a multifactorial disorder (2,3). The inheritance ofsevere forms of this disease (degrees II–IV) can be describedby a model that assumes a dominant major gene effect (4),although support for any specific mode of inheritance islimited. Fundamental understanding of the primary causes ofAIS is hampered because patients usually present in earlypuberty with already well developed curves. Moreover,without the identification of specific gene defects from linkagestudies, the early pathological processes in man are impossibleto investigate. One approach to dissecting the genetic basis ofAIS is to identify mouse mutants showing scoliosis andappraise them as candidates for models of AIS (5,6).

Currently no mouse mutant fully satisfies all requirementsfor acceptance as a model for AIS; however, the autosomalrecessive mouse mutant kyphoscoliosis (ky) is a useful systemfor examining the neuromuscular and skeletal mechanisms bywhich spinal curvature develops (7). Chronic thoraco-lumbarkyphoscoliosis in homozygous ky/ky (Fig. 1a) is preceded by asingle postnatal phase of muscle fibre degeneration and regenera-tion within postural slow contracting muscles (e.g. soleus) (8).The serum enzyme levels of pyruvate kinase and total creatinekinase have been measured in ky/ky mice and found to bewithin normal limits (8). All ky/ky muscles are smaller, slowercontracting and weaker than controls (9). Biochemical analysishas shown that in muscles of ky/ky mice there are dramaticshifts to the expression of contractile protein isoforms typicalof slow muscle, including myosin heavy chains (MHCs) andmyosin light chains (MLCs) (10).

In order to gain insights into the mechanisms leading to scoliosis,the pathological changes provoked by the ky mutation have

+To whom correspondence should be addressed. Tel: +44 1235 834393; Fax: +44 1235 824542; Email: [email protected]

10 Human Molecular Genetics, 2001, Vol. 10, No. 1

Human Molecular Genetics, 2001, Vol. 10, No. 1 11

been further characterized. Also, in the present study, the kygene has been identified by positional cloning and the causativemutation unveiled.

RESULTS

MHC isoform expression shifts in ky/ky mice

The expression shifts in the MHC isoforms in the mutant back-ground have been further characterized by immunolocalizationof MHC1 in both wild-type BDL and ky/ky muscle fibres(Fig. 1b–d). For example, ky/ky soleus muscle shows aprogressive and eventually complete shift to type I fibresexpressing only MHC1 that is not due to selective fast musclefibre death. The distribution of two other MHCs (MHC2A and2B) showed complementary changes with complete loss of fastmyosin expression (data not shown).

Response to compensatory overload

Whereas the body weight of ky/ky mice is ∼80% of that of age-matched BDL controls, both soleus and extensor digitorumlongus (EDL) muscle weights are as low as 25% (10). Thus,we predicted that ky/ky muscles might be exposed to an endog-enously generated overload arising from the disproportionatebody mass. If normal muscles are overloaded in vivo they canrespond by adapting the type of MHC and MLC expressed andalso by undergoing hypertrophy to generate increased force.We hypothesized that the small size of ky/ky muscles mightreflect a defect in their ability to adapt to load and that theextreme shifts to a slow contractile phenotype were generatedin order to increase their resistance to fatigue. We thereforeexamined the effect of surgical compensatory overloadimposed on the EDL muscle in vivo in order to induce musclefibre hypertrophy. The EDL was chosen as it does not undergonecrosis and regeneration in ky/ky and thus any changes couldbe separated from this event. There is a complete absence of ahypertrophic response to 5 weeks of compensatory overload inthe ky EDL, whereas the mean fibre diameters of type 2A and2B fast muscle fibres increase up to 17.8% in wild-type BDLmice (Fig. 1f).

Neuromuscular junction remodelling

ky/ky muscles such as soleus, whose muscle fibres necrose andregenerate, show motoneuron sprouting in the absence ofmotoneuron loss (8). We have now examined the distributionof acetylcholine receptors in neuromuscular junctions (NMJs)of ky/ky soleus and show that they are abnormally arranged innumerous small patches rather than the well ordered structurestypical of wild-type NMJs (Fig. 2a). The EDL muscle isnormal in this respect. Moreover, in vivo imaging shows that

NMJs within sternomastoid, a muscle that in ky/ky mice doesnot undergo necrosis and regeneration, are persistentlyunstable showing both synapse loss and formation withinindividual adult muscles (Fig. 2b). Thus, NMJ disorganizationin ky/ky muscles is not a simple consequence of muscle fibreremodelling. Although this may reflect a secondary adaptivechange to a primary muscle abnormality, it does not rule out arole for Ky in NMJ stabilization.

Positional cloning and identification of the ky mutation

We have undertaken a positional cloning effort to identify themouse ky mutation. High resolution genetic and physicalmapping of the ky locus has led to the identification of a putativemuscle-specific transcription unit from the ky non-recombinantinterval (11). Using a combination of rapid amplification ofcDNA ends and muscle cDNA library screening, a 5.3 kbcomposite of the candidate gene transcript was assembled. Aperfect Kozak’s consensus sequence (12) surrounds the putativeATG, indicating the beginning of an open reading frame(ORF) encoding a 73 kDa protein. Although 3′-untranslatedregion (3′-UTR) probes showed additional bands on northernblots, possibly identifying alternatively spliced products,probes from the predicted coding region identified only asingle band in skeletal and heart muscles (Fig. 3a). Semi-quantitative RT–PCR analysis revealed that this transcript ispresent at significantly lower levels in ky/ky mice comparedwith wild-type (Fig. 3b). Northern blot analysis also demonstrateda very low level of expression of this muscle-specific gene inky/ky mice (Fig. 3b). Sequencing of the candidate gene in ky/kymice identified a GC deletion affecting codon 24 downstreamof the putative first ATG (Fig. 4). Other mouse strains,including the parental wild-type strain BDL, C57BL/6, Musspretus, 129, C3H and BALB/c showed the expected wild-typesequence across this region. The GC deletion alters the readingframe leading to a premature stop codon at position 105.

The Ky protein belongs to a new family oftransglutaminase-like proteins

The human ky orthologue, identified through expressedsequence tag matches and mapped using the Genebridge4 radiationhybrid panel, falls into a conserved region of synteny on 3q21,between markers D3S3086 and D3S3108 (data not shown).Two rounds of PSI-BLAST (13) database searches using anE-value inclusion threshold of 10–3 revealed significantsimilarity to five predicted proteins: Drosophila melanogasterCG13435, Caenorhabditis elegans K02C4.4, Saccharomycescerevisiae Ydl117wp, Schizosaccharomyces pombe C9G1.06Cand Synechocystis sp. sll1681. Our own studies (Fig. 4) andthose of others (14) demonstrate that these five proteins eachcontain a domain homologous to human transglutaminase.

Figure 1. Phenotypic analysis of ky/ky mice. (a) ky/ky mice show inability to carry out a placing response (stretching and lifting of forelimbs and head to grab aclose edge) and visible scoliosis on X-ray analysis (white arrow). (b–f) Immunohistochemical localization of MHC1 (slow) in wild-type and ky/ky muscles.(b) BDL soleus muscle (40-day-old) showing normal checkerboard of biochemically differentiated fibre types. (c) Post-regenerative (40 days) soleus muscle of ky/kymutant mice, showing wide variation in levels of MHC1 between muscle fibres. Variation is also seen for other isoforms (data not shown). (d) BDL soleus at100 days, with normal checkerboard pattern of fibres. (e) ky/ky soleus at 100 days, muscle fibres express only MHC1. Bars: (a–d) 100 µm. (f) Histogram showingchanges in mean EDL muscle (not affected by necrosis) fibre diameters of type 1, 2A and 2B muscle fibres in response to 5 weeks of compensatory overloadin vivo. The mean diameters of type 2A and 2B muscle fibres increases significantly (P < 0.0001) in the EDL muscle of BDL mice (no significant change in typeI fibres). ky/ky mice, on the other hand, make no hypertrophic response.


With the exception of C.elegans K02C4.4, which containssubstitutions of key active site residues, these transglutaminasehomologues are likely to possess protease functions (14). TheSaccharomyces cerevisiae Ky homologue Ydl117wp hasrecently been renamed CYK3 because of its role in cyto-kinesis. This novel cytokinetic factor possibly contributes tosecretory targeting or septal deposition during yeast cell division.Furthermore, genetic interactions also argue for CYK3 havinga central role in cell division (15).

DISCUSSION

The mutation identified in this paper has unveiled a novelmuscle-specific protein that plays a key role in skeletalmuscle’s normal response to mechanical load. Our resultsshow that the signalling mechanisms controlling changes incontractile protein biochemistry are intact in ky/ky. Thecomplete shift to type I fibres expressing only MHC1 thatoccurs in ky/ky mice probably reflects an adaptive change ofpostural muscles to the endogenous overload caused by thelow muscle-to-body weight ratio (10). Adaptive change is alsoobserved in AIS where muscles from the convex and concavesides of the curvature show different fibre type proportions(16–18), probably reflecting response to changing demand(19).

Fibre conversions and hypertrophy are part of the normalsynchronized response of skeletal muscle to functional overloador chronic increase in demand. Interestingly, there is acomplete absence of a hypertrophic response to compensatoryoverload in the ky EDL. To our knowledge ky is the onlypublished mouse mutant to exhibit the absence of a hyper-trophic response to overload in vivo. As such, our results showthat the Ky protein is an essential component of the musclehypertrophy pathway. Given that fast-to-slow fibre transfor-mations are normal in ky, the ky mouse provides genetic

evidence for the independent regulation of these two cellularprocesses.

We have also shown that the NMJs of ky/ky mice are notnormally differentiated. Muscles undergoing fibre remodelling(e.g. soleus) show a striking pattern of acetylcholine receptordistribution with no typical endplate localization, althoughclustering still occurs in many isolated patches. The uninterruptedloss and formation of adult sternomastoid NMJs shown byin vivo imaging demonstrates that NMJ instability is not aconsequence of muscle fibre remodelling, as sternomastoid isspared the necrosis and regeneration processes affecting othermuscle groups in ky/ky mice. These data are most consistentwith the changes seen at ky/ky NMJs being an adaptiveresponse to intense muscle activity; similar changes have beenobserved in intensely exercised rats (20). Nevertheless, ourresults do not rule out the possibility that Ky plays a role in themaintenance of stability in maturing NMJs.

Although proximal muscle weakness is often found inhuman scoliotic patients (21,22) there are no detailed functionalstudies to date that demonstrate a similar muscle dysfunctionin humans to the one provoked by the ky mutation in themouse. Nevertheless, in the light of the results shown in thispaper, we hypothesize that mutations in the human homologueof ky will lead to weakness of postural muscles and therefore itis a candidate for a group of familial AIS. Moreover, furtherelucidation of the function of this protein may shed light on themolecular pathogenesis leading to scoliosis in humans.

MATERIALS AND METHODS

Production of ky mutant stock

The ky recessive mutation arose in the BDL strain of laboratorymice. Homozygote ky/ky mice were produced from hetero-zygous parents and were readily identified because muscleweakness and subsequent chronic scoliosis prevents ky/ky

Figure 2. Analysis of the structure of NMJs in ky/ky mice. (a) Postnatal maturation of neuromuscular junctions in ky/ky and BDL soleus and EDL muscles. NMJsin the soleus muscle of ky/ky mice become disorganized into isolated patches of AChRs at ∼30 days. Disorganization persists at least up to 360 days. (b) In vivoimaging of AChRs at NMJs within sternomastoid muscles of 10-week-old ky/ky mice. Two examples of progressive loss of AChRs from NMJs over a 33 day period(white arrows). Bars: (a) 10µm; (b) 20 µm.


mutant mice from executing a normal placing reflex. In addition,each ky/ky mouse was autopsied to confirm its muscle pheno-type. Muscle fibres from the whole of the posterior compart-ment of the lower hind limb were examined. The ky/kyphenotype was characterized by the presence of active musclefibre necrosis, regeneration as indicated by the presence ofsmall diameter fibres that are highly basophilic and the retentionof central nucleation (8).

Expression analysis

Multiple tissue northern blots (Clontech) were probed with kycDNA 3′-UTR and ORF probes (Fig. 3a) following the manu-facturer’s recommendations. The 3′-UTR probe was generatedwith primers forward 5′-AAGCAGATTAGCCCAAGCAG-3′and reverse 5′-GGCTGTGTCTAGCGAGTGG-3′. The ORFprobe was generated with primers forward 5′-TGAGGAC-CCAGAAGACCAAC-3′ and reverse 5′-AGCTTCATCCCGT-TCTTCCT-3′. For expression analysis, samples were obtainedfrom 40- to 50-day-old ky/ky and wild-type BDL mice.Northern blots were made using 500 µg of total RNA isolatedfrom the whole of the posterior compartment of the lower hind

limb using the Trizol method (Sigma). Semi-quantitative RT–PCR analysis was carried out on first strand cDNA generatedwith Superscript II (Gibco BRL) from 500 ng of poly(A)+

RNA purified from total RNA using mini-columns (Qiagen),according to the manufacturer’s recommendations. RT–PCRreactions for ky (forward 5′-GGGATGAAGCTGGAGGTGTA-3′,reverse 5′-ATTAGGGTGTCCTGGGCTCT-3′), desmin (forward5′-GACGCTGTGAACCAGGAGTT-3′, reverse 5′-ACGAG-CTAGAGTGGCTGCAT-3′) and myogenin (forward 5′-CCT-GCCCTGAGTTGAGAGAG-3′, reverse 5′-CGGCAGCTTT-ACAAACAACA-3′) were performed at increasing numbers ofcycles (31, 33, 35, 37, 39 and 41 cycles) and the optimalnumber of cycles reflecting the concentration of each transcriptin the first strand cDNA template was empirically determinedon agarose gels.

Histology

Tissues were fixed by immersion in 10% neutral bufferedformalin for 24 h, dehydrated in ethanol, cleared in xylene andembedded in paraffin wax. Sections of 4 mm thickness were

Figure 3. Identification and analysis of the ky transcript. (a) (Top) A diagram of the ky composite cDNA drawn to scale. Black solid bars depict the position andlength of probes used in northern blots below. (b) (Left) Northern blot using total RNA from BDL and homozygous ky/ky mice probed with ky and Desmin as aninternal loading control; (right) combined semi-quantitative RT–PCR for ky and two controls (desmin and myogenin). RT–PCR reactions were independentlycalibrated for each gene (see Materials and Methods). To facilitate comparison, the PCR products for ky, desmin and myogenin have been combined and loadedtogether.


dewaxed, brought to water via an ethanol series and stainedwith haematoxylin and eosin.

Immunohistochemistry

Tissues were frozen in isopentane cooled in liquid nitrogenprior to crysosectioning. Standard indirect immunohistochemistrywas employed using primary mouse monoclonal antibodiesspecific to MHC isoforms: BAD5 for MHC1, SC-71 forMHC2A and BF-F3 for MHC2B (23). Primary antibodies werelocalized using an FITC-conjugated polyclonal anti-mouse IgG(Sigma) and visualized using an Olympus PROVIS microscope.

Whole mount muscle histochemistry

Whole mounts of muscles for confocal microscopic imagingwere prepared, following dissection, by pinning out under

slight tension onto layers of Sylgard polymer. Muscles werestained in 20 µg/ml α-bungaroxin–tetrarhodamine isothiocyanate(α-BTX–TRITC) (Molecular Probes) dissolved in phosphate-buffered saline (PBS) (pH 7.4, 15–30 min room temperature).Muscles were washed three times in PBS (15 min) with agitation.Muscles were fixed in 1% freshly prepared phosphate-bufferedparaformaldehyde (pH 7.4, 15 min) prior to three further PBSwashes (15 min each). They were then teased into smallbundles containing ∼100–150 fibres. Bundles were immediatelymounted in Vectashield UV-free aqueous mountant under1 µm thick coverslips laid onto Vectabond-coated slides.

Confocal microscopy

Confocal imaging was performed using two different systems.Initially, images were captured using an MRC500 confocal

Figure 4. The ky mutation and homologues of the Ky protein. The chromatogram shows the parental and homozygous ky sequences. The mutation, a GC deletion,is highlighted. An arrow indicates the position of the ky mutation in the predicted Ky protein, disrupting amino acid 24. Conserved domains in Ky homologues areindicated as colour-filled boxes: blue, LIM domain; green, SH3 domain; red, transglutaminase domain. A multiple alignment of the Ky transglutaminase domainwith both Ky homologues and the transglutaminase domain from human factor XIIIa of known tertiary structure (PDB code 1GGT) is shown. This alignmentcorresponds to that shown by Meier et al. (19), ProDom domain PD018376 and SMART domain TGc [http://smart.embl-heidelberg.de/ (28,29)]. Active site resi-dues in factor XIIIa are marked with a red asterisk. The known secondary structures of factor XIIIa are shown beneath the alignment: E, β-strand; H, α-helix.


visualization system (BioRad). Subsequently images werecaptured using a DVC-250 Viewscan confocal microscope(BioRad). The configuration, means of alignment and means ofregulating the laser input of the MRC 500 confocal microscopehave been described in detail elsewhere (24). Samples wereilluminated with either the 488 nm or the 515 nm line of theargon ion laser, selected with an appropriate band pass filter ofa 10 nm bandwidth. The images were collected with a 565 nmlong pass filter placed before the detector. For the MRC500images were collected for an average of 160 scans of the laserand the number of scans used by the Viewscan was empiricallydetermined for each sample. Images from several planes offocus 0.6–1 µm apart were collected and presented in a singleimage where the vertical pixel that exhibited the maximumbrightness in the data set was displayed.

In vivo microscopy

Sternomastoid muscles were isolated from mice and visualizedas described previously (25). Essentially, 10- to 15-week-oldky/ky and BDL (+/+) mice were anaesthetized with a singleintraperitoneal injection of chloral hydrate (0.5–0.6 mg/kgbody wt) or a mixture (5.0 ml/kg body wt) containing 0.17 mgof ketamine (Ketaset; Aveco) and 1.7 mg of xylazine (Anased;Lloyd Laboratories) per millilitre of 0.9% sodium chloridesolution. The anaesthetized mouse was placed on its back onthe stage of a modified microscope so that the upper partswung away. The mouse was then intubated and mechanicallyventilated for the duration of the experiment. A midline incisionwas made from the sternum to the apex of the mandible, andthe left sternomastoid muscle was exposed by lateral deflectionof the skin and salivary glands. The muscle was gently lifted ona small platform (23). Nerve terminals and muscle fibres werestained for 3 min with a 4–10 µM solution of 4-Di-2-Asp(Molecular Probes), a fluorescent vital mitochondrial marker(23,26), in Ringer’s solution. Post-synaptic acetylcholinereceptors (AChRs) were visualized by direct application in thesame Ringer’s solution of non-blocking concentrations (2–4 µg/ml) of α-BTX–TRITC. After staining, the wound was washedwith lactated Ringer’s solution and a coverslip was lowered onthe surface of the muscle. With the microscope swung backinto its normal position, the NMJs were visualized withconventional epifluorescence and video microscopy. Using alow-light-level SIT camera (Dage/MTI) with a Trapix digitalimage processor (Recognition Concepts) connected to a micro-VAX (23), video images were digitized, averaged (32–64 frames)and stored on optical disc. Image processing was done withIMAGR (27).

Surgical compensatory overload

Under deep anaesthesia, the tibialis anterior muscle in the frontof the right hindlimb was removed by severing the tendons.The wound was then sutured, animals were provided with anal-gesics and allowed to recover under observation. After5 weeks, mice were killed by cervical dislocation and the EDLmuscles were dissected and prepared for immunohistochemistry.All procedures were performed in accordance with the UKAnimals (Scientific Procedures) Act 1986 and European Unionregulations.

Morphological and statistical analysis

The least diameter of 200 muscle fibres randomly sampledfrom sections stained with antisera to MHC1, MHC2A andMHC2B was measured from digitized images using NIHImage software. These data were analysed statistically usingStudent’s t-test to determine significant differences betweenloaded and unloaded muscles.

Accession numbers

GenBank accession numbers are as follows: D.melanogasterCG13435 (7302360), C.elegans K02C4.4 (3878111), Saccharo-myces cerevisiae Ydl117wp (6320086), Schizosaccharomycespombe C9G1.06C (3183389) and Synechocystis sp. sll1681(1651866); M.musculus ky (AJ293727).

ACKNOWLEDGEMENTS

We wish to thank Dr Leslie Bridges (University of Leeds) forthe image of the normal neuromuscular junction and Dr AlanEntwistle (Ludwig Institute, London) for assistance withconfocal imaging. This research was supported by grants fromthe MRC, SERC, BBSRC and the Charing Cross andWestminster Hospital Trustees.

REFERENCES

1. Rogala, E.J., Drummond, D.S. and Gurr, J. (1978) Scoliosis: incidenceand natural history. A prospective epidemiological study. J. Bone JointSurg. Am., 60, 173–176.

2. Kesling, K.L. and Reinker, K.A. (1997) Scoliosis in twins. A meta-analysis ofthe literature and report of six cases. Spine, 22, 2009–2014.

3. Inoue, M., Minami, S., Kitahara, H., Otsuka, Y., Nakata, Y., Takaso, M.and Moriya, H. (1998) Idiopathic scoliosis in twins studied by DNAfingerprinting: the incidence and type of scoliosis. J. Bone Joint Surg. Br.,80, 212–217.

4. Axenovich, T.I., Zaidman, A.M., Zorkoltseva, I.V., Tregubova, I.L. andBorodin, P.M. (1999) Segregation analysis of idiopathic scoliosis:demonstration of a major gene effect. Am. J. Med. Genet., 86, 389–394.

5. Giampietro, P.F., Raggio, C.L. and Blank, R.D. (1999) Synteny-definedcandidate genes for congenital and idiopathic scoliosis. Am. J. Med.Genet., 83, 164–177.

6. Blank, R.D., Raggio, C.L., Giampietro, P.F. and Camacho, N.P. (1999) Agenomic approach to scoliosis pathogenesis. Lupus, 8, 356–360.

7. Mason, R.M. and Palfrey, A.J. (1984) Intervertebral disc degeneration inadult mice with hereditary kyphoscoliosis. J. Orthop. Res., 2, 333–338.

8. Bridges, L.R., Coulton, G.R., Howard, G., Moss, J. and Mason, R.M.(1992) The neuromuscular basis of hereditary kyphoscoliosis in themouse. Muscle Nerve, 15, 172–179.

9. Marechal, G., Coulton, G.R. and Beckers-Bleukx, G. (1995) Mechanicalpower and myosin composition of soleus and extensor digitorum longusmuscles of ky mice. Am. J. Physiol., 268, C513–519.

10. Marechal, G., Beckers-Bleukx, G., Berquin, A. and Coulton, G. (1996)Isoforms of myosin in growing muscles of ky (kyphoscoliotic) mice. Eur.J. Biochem., 241, 916–922.

11. Blanco, G., Nikitopoulou, A., Kraus, M., Mason, R.M., Coulton, G.R. andBrown, S.D. (1998) A STS content physical and transcription map acrossthe ky, kyphoscoliosis, nonrecombinant region. Genomics, 54, 415–423.

12. Kozak, M. (1996) Interpreting cDNA sequences: some insights from studieson translation. Mamm. Genome, 7, 563–574.

13. Altschul, S.F., Madden, T.L., Schaffer, A.A., Zhang, J., Zhang, Z., Miller,W. and Lipman, D.J. (1997) Gapped BLAST and PSI-BLAST: a newgeneration of protein database search programs. Nucleic Acids Res., 25,3389–3402.

14. Makarova, K.S., Aravind, L. and Koonin, E.V. (1999) A superfamily ofarchaeal, bacterial, and eukaryotic proteins homologous to animaltransglutaminases. Protein Sci., 8, 1714–1719.


15. Korinek, W.S., Bi, E., Epp, J.A., Wang, L., Ho, J. and Chant, J. (2000)Cyk3, a novel SH3-domain protein, affects cytokinesis in yeast. Curr.Biol., 10, 947–950.

16. Bylund, P., Jansson, E., Dahlberg, E. and Eriksson, E. (1987) Muscle fibertypes in thoracic erector spinae muscles. Fiber types in idiopathic andother forms of scoliosis. Clin. Orthop., 222–228.

17. Chiu, J.C. (1988) Morphological studies on the erector spinae muscle insixty consecutive scoliotic patients. Nippon Seikeigeka Gakkai Zasshi, 62,1163–1175.

18. Gonyea, W.J., Moore-Woodard, C., Moseley, B., Hollmann, M. andWenger, D. (1985) An evaluation of muscle pathology in idiopathicscoliosis. J. Pediatr. Orthop., 5, 323–329.

19. Meier, M.P., Klein, M.P., Krebs, D., Grob, D. and Muntener, M. (1997)Fiber transformations in multifidus muscle of young patients withidiopathic scoliosis. Spine, 22, 2357–2364.

20. Deschenes, M.R., Maresh, C.M., Crivello, J.F., Armstrong, L.E.,Kraemer, W.J. and Covault, J. (1993) The effects of exercise training ofdifferent intensities on neuromuscular junction morphology. J.Neurocytol., 22, 603–615.

21. Pecak, F., Trontelj, J.V. and Dimitrijevic, M.R. (1980) Scoliosis inneuromuscular disorders. Int. Orthop., 3, 323–328.

22. Kumano, K. (1980) Congenital non-progressive myopathy, associated withscoliosis–clinical, histological, histochemical and electron microscopic stud-ies of seven cases. Nippon Seikeigeka Gakkai Zasshi, 54, 381–402.

23. Lichtman, J.W., Magrassi, L. and Purves, D. (1987) Visualization ofneuromuscular junctions over periods of several months in living mice. J.Neurosci., 7, 1215–1222.

24. Entwistle, A. and Noble, M. (1992) The quantification of fluorescentemission from biological samples using analysis of polarization. J.Microsc., 165, 347–365.

25. van Mier, P. and Lichtman, J.W. (1994) Regenerating muscle fibersinduce directional sprouting from nearby nerve terminals: studies in livingmice. J. Neurosci., 14, 5672–5686.

26. Magrassi, L., Purves, D. and Lichtman, J.W. (1987) Fluorescent probesthat stain living nerve terminals. J. Neurosci., 7, 1207–1214.

27. Turney, S.G., Culican, S.M. and Lichtman, J.W. (1996) A quantitativefluorescence-imaging technique for studying acetylcholine receptor turnoverat neuromuscular junctions in living animals. J. Neurosci. Methods, 64,199–208.

28. Schultz, J., Milpetz, F., Bork, P. and Ponting, C.P. (1998) SMART, asimple modular architecture research tool: identification of signalingdomains. Proc. Natl Acad. Sci. USA, 95, 5857–5864.

29. Schultz, J., Copley, R.R., Doerks, T., Ponting, C.P. and Bork, P. (2000)SMART: a web-based tool for the study of genetically mobile domains.Nucleic Acids Res., 28, 231–234.

the kyphoscoliosis (ky) mouse is deficient in hypertrophic

Documents