Development 101, 673-684 (1987)Printed in Great Britain © The Company of Biologists Limited 1987

673

Nerve-dependent accumulation of myosin light chain 3 in developing

limb musculature

PETER A. MERRIFIELD* and IRWIN R. KONIGSBERG

Department of Biology, University of Virginia, Charloltesville, Virginia 22901, USA

* Present address for correspondence: Department of Anatomy. Medical Sciences Building, University of Western Ontario. London.Ontario. Canada N6A 5C1

Summary

Myosin alkali light chain accumulation in developingquail limb musculature has been analysed on immuno-blots using a monoclonal antibody which recognizes anepitope common to fast myosin light chain 1 (MLC|f)and fast myosin light chain 3 (\ILC3r). The limbmuscle of early embryos (i.e. up to day 10 in ovo) has aMLC profile similar to that observed in myotubescultured in vitro; although MLC)r is abundant, MLC3r

cannot be detected. MLC3f is first detected in 11-dayembryos. To determine whether this alteration inMLCir accumulation is nerve or hormone dependent,limb buds with and without neural tube were culturedas grafts on the chorioallantoic membrane of chickhosts. Although differentiated muscle develops in both

aneural and innervated grafts, innervated grafts con-tain approximately three times as much myosin asaneural grafts. More significantly, although aneuralgrafts reproducibly accumulate normal levels ofMLC if, they fail to accumulate detectable levels ofMLCjf. In contrast, innervated grafts accumulateboth MLCif and MLC3f, suggesting that the presenceof neural tube in the graft promotes the maturation, aswell as the growth, of muscle tissue. This is the firstpositive demonstration that innervation is necessaryfor the accumulation of MLC,r that occurs duringnormal limb development in vivo.

Key words: CAM grafting, myosin light chains, musclematuration, limb buds, quail.

Introduction

Muscle contractile proteins are the products of multi-gene families. The number of genes in each familydepends upon the organism, but it has been estimatedthat there are between 7 (Robbins, Freyer, Chisholm& Gilliam, 1982) and 31 distinct myosin heavy chainsequences (Kropp, Gulick & Robbins, 1986) and atleast 10 actin-like sequences (Schwartz & Rothblum,1980) in the chicken genome. Since striated, smoothand cardiac muscles all express different contractileprotein isoforms, the expression of some of thesefamily members is obviously tissue specific. Recently,it has been suggested that some isoforms are develop-mentally regulated within a single tissue. Forexample, several groups have demonstrated the exist-ence of embryonic and neonatal isoforms of myosinheavy chain which are present for only brief periods

during embryonic and early post-hatch developmentof the chick and which differ from the adult isoforms(Rushbrook & Stracher, 1979; Bandman, Matsuda &Strohman, 1982; Bader, Masaki & Fischman, 1982;Winkelman, Lowey & Press, 1983). Analysis of themRNA populations of embryonic muscle usingcDNA probes to myosin has confirmed this obser-vation (Umeda etal. 1983). Isoform switching duringearly chick development has also been described forcr-actin (Paterson & Eldridge, 1984), tropomyosin(Roy, Sreter & Sarkar, 1979; Montarras, Fiszman &Gros, 1982; Matsuda, Bandman & Strohman, 1983),troponin (Matsuda, Obinata & Shimada, 1981) andmyosin light chain subunits (Chi, Fellini & Holtzer,1975; Rubinstein, Pepe & Holtzer, 1977; Merrifield &Konigsberg, 1986; Crow, Olson & Stockdale, 1983).

In the avian embryo, multiple forms of myosin lightchains are coexpressed in the muscles of embryos

674 P. A. Merrifield and I. R. Konigsberg

younger than 10 days, including the adult fast (If, 2f),adult slow (Is, 2s) and embryonic-specific (le) lightchains (Stockdale, Raman & Baden, 1981; Takano-Ohmuro et al. 1985). Remarkably, alkali light chain 3(MLC3f) is not expressed at these early stages and isfirst detected in developing fast muscle prior tohatching at about the stage at which adult slow andembryonic light chains disappear (Crow et al. 1983;Takano-Ohmuro etal. 1985). In cultured muscle cells,however, the coexpression of fast and slow isoformspersists and MLC3f does not accumulate to detectablelevels (Keller & Emerson, 1980). Thus, both thedisappearance of slow and embryonic MLC isoformsand the activation of MLC3f synthesis in developingfast muscle have been taken as markers for thematuration of muscle in the avian embryo.

Although the timing of isoform switching seemsspecific for each contractile protein, it has beensuggested that many of these changes may be orches-trated by some common factor in the developingembryo. In light of the well-documented effect ofhormones (Johnson et al. 1980; Gambke et al. 1983;Butler-Browne, Herlicoviez & Whalen, 1984; Wha-len, Toutant, Butler-Browne & Watkins, 1985;Izumo, Nadal-Ginard, & Mahdavi, 1986) and nerve(Buller, Eccles & Eccles, 1960; Gutmann, 1976;Jolesz & Sreter, 1981) upon adult mammalian muscle,it is possible that either influence may be playing arole in affecting isoform switches. This possibility issupported by the observation that muscle cells cul-tured in vitro differentiate into embryonic muscle butdo not undergo the switches seen in the maturingembryo.

In order to examine the role of hormonal andnervous influences upon the activation of MLC3f

synthesis in the avian embryo, we have used theapproach of grafting both aneural and innervatedlimb buds onto the chorioallantoic membrane (CAM)of chick embryo hosts and then analysing the devel-opmental potential of these grafts using a monoclonalantibody that can detect the developmental switch inmyosin fast alkali light chains. The technique of CAMgrafting was first used to study the developmentalcapacity of early limb buds by Murray & Huxley(1925), who observed that such grafts could attain 'aform which is extraordinarily close to the normal'.Murray (1926) concluded that limb buds from 3- to 5-day-old chick embryos are 'self-differentiating' andthat morphogenesis of the limb is independent ofinnervation and function. These observations weresubsequently confirmed and extended by others(Hunt, 1932; Eastlick, 1943; Bradley, 1970; Kenny-Mobbs & Hall, 1983). While these latter studies havebeen concerned with the ability of aneural andinnervated limb buds to produce differentiatedmuscle when grafted onto the CAM, this study is

largely concerned with the ability of these grafts todevelop mature muscle, i.e. muscle that expresses theadult-specific contractile protein isoform pattern.Using low ionic strength precipitation of myosin andimmunoblot analysis, we have been able to demon-strate that the inclusion of neural tube in the graftedlimb bud influences not only the growth but also theexpression of MLC3f in muscle tissue of the develop-ing graft.

Materials and methods

Fertile quail eggs (Coturnix cotumix japonica) were ob-tained from our own breeding stock and used to obtaindonor limb buds. Chicken eggs (Gallus domesticus,Hubbard x Hubbard strain) obtained from HeatwoleHatchery, Harrisonberg, Virginia were used as hosts. Bothchick and quail embryos were incubated in a forced draft,humidified incubator at 37-6°C and staged according to thecriteria of Hamburger & Hamilton (1951) and Sato,Hoshino, Mizuma & Nishida (1971), respectively.

Chorioallantoic grafting procedureGrafting of limb buds from 3-day quail embryos onto theCAM of 7-day chick hosts was carried out as described byHamburger (1942). This choice of experimental protocolwas largely dictated by technical considerations. Inchoosing a donor limb bud, we wanted initially to obtainbuds that had not yet become innervated. Since nerve fibresfirst penetrate the limb bud of the chick at Hamilton andHamburger (H.H.) stage 23 (approximately day 4 ofincubation) (Fouvet, 1973), 3-day-old (H.H. stage 20)chicks were the oldest embryos from which we couldreproducibly obtain truly 'aneural' limb buds. Since theCAM of the chick embryo is not sufficiently developed toreceive grafts before day 7 of development and because theblood vessels of the CAM that support the graft begin todegenerate on day 19 (Hall, 1978), limb buds from 3-day-old chick embryos could only be incubated as grafts to atotal chronological age of 15 days. This approach would beineffective, since it has been shown that MLC3f accumu-lation cannot be detected in the chick embryo until day 16of development. To circumvent this difficulty, we decidedto obtain limb buds from 3-day quail rather than chickdonors. Since quail embryos develop faster (16 versus 21days to hatching) and activate MLC3f earlier than the chick(day 11 versus day 16), these aneural limb buds were graftedonto a day-7 chick host and incubated to a chronologicalage of 15 days. It has been known for some time that quailsomitic muscle precursor cells transplanted into chick hostsdevelop into cytologically normal muscle (Christ, Jacob &Jacob, 1977). Thus, even if the grafts were somewhatretarded in their growth (as is often the case) activation ofMLC3f in the graft would most likely be detected, should itoccur.

Quail embryos were isolated sterilely, placed in How-ard's saline (Howard, 1953) and dissected free of all extra-embryonic membranes using electrolytically sharpenedtungsten needles (Dossel, 1958). Aneural limb buds weredissected to contain the somatopleure lateral to the somites

MLC3f accumulation is nerve dependent 675

at the level of the limb. Innervated limb buds were excisedto include both somite and neural tube (see Fig. 3). Afterpuncturing the shell above the air space, a 3 mm squarewindow was cut into the chick egg, the shell membraneremoved and the graft placed onto the chorioallantoicmembrane (using a Spemann pipette) at a site where twoblood vessels bifurcated. The shell was replaced on thewindow, sealed with paraffin and the hosts were incubated(without rolling) for 12 days. The grafts were removed,freed of any adherent membrane, weighed, photographedand extracted for electrophoresis and immunoblot analysis.

Preparation and analysis of limb myosinNormal limbs and grafts were homogenized in 10 vol. (w/v)of high salt extraction buffer [0-6M-NaCl, 15 mM-Tris-HCl(pH7-4) containing 1 mm-PMSF] in a ground-glass hom-ogenizer and extracted for 15 min on ice (Burridge & Bray,1975). Insoluble material was pelleted by centrifugation at12 800 g for 15 min and the supernatant was made 50 % inglycerol and stored at -20°C. Extracts were either analyseddirectly or diluted by dialysis to lower the ionic strength ofthe extract and precipitate out a crude myosin fraction. Toobtain this fraction, glycerinated extracts were dialysedagainst 50 vol. of cold distilled water containing lmM-PMSF for 6h at 4°C using Spectaphor 6000 Mr cut-offdialysis tubing. Myosin was pelleted at 10 000 g for 15 min ina Sorvall HS-4 rotor, solubilized in 0-5 % SDS and assayedfor protein concentration (Lowry, Rosebrough, Fair &Randall, 1951). Samples to be run on gels were mixed with5x Laemmli sample buffer and run on 12-5% acrylamidegels as described by Laemmli (1970) and either stained withCoomassie Blue (Fairbanks, Steck & Wallach, 1971) orelectrophoretically transferred onto 0-1/iM-nitrocellulosepaper (Schleicher & Schuell, PH 79) overnight at 10 wattsconstant power (Towbin, Staehelin & Gordon, 1979).

Detection of myosin alkali light chains onimmunoblotsA monoclonal antibody to adult quail breast muscle myosinwas used to detect the presence of MLClf and MLC3f incrude myosin samples transferred to nitrocellulose. Theantibody (QBM-2) is specific for fast alkali light chains(Merrifield, Payne & Konigsberg, 1983) and recognizes anepitope common to both MLClf and MLC3f. Followingelectrophoretic transfer, the nitrocellulose replica wasrinsed with PBS, blocked for 2h at 37°C with 5 % BSA inPBS and then incubated overnight with shaking at 4°C withmonoclonal antibody QBM-2. The blot was then washedwith several changes of PBS (45 min) and incubated with a1/1000 dilution of peroxidase-conjugated goat anti-mouseIgG (TAGO) in 0 1 % BSA in PBS for 2h at 37°C.Following a final wash, peroxidase activity was localizedusing o-dianisidine HC1 in 10 mM-Tris-HCl buffer (pH7-4)containing 0033 % hydrogen peroxide as a substrate.

Results

Myosin light chains during normal limb developmentWhen total high-salt-soluble extracts of limb budsfrom different-aged quail embryos are analysed on

immunoblots for their alkali light chain content, weobserve that limb buds from 9-day embryos accumu-late detectable levels of MLC,f but not MLC3f

(Fig. IB). The presence of MLC]f in this preparation,as well as the presence of myosin heavy chain andother muscle-specific proteins in the stained gel(Fig. 1A), demonstrate that differentiated muscle iscertainly present in 9-day limbs, and the absence ofMLC3f suggests that the muscle fibres that differen-tiate first in the developing limb are significantlydifferent from adult fibres. The alkali light chaincontent of the early muscle fibres in the embryo is, infact, very similar to that previously described forquail myotubes cultured in vitro (Merrifield et al.

MHC

9 10111213141516 Mt 9 10111213141516Days in ovo

Fig. 1. Myosin alkali light chain content of embryonicquail limb muscle as determined by Coomassie Bluestaining and immunoblot analysis with monoclonalantibody QBM-2. Total high salt extracts (A,B) orprecipitated myosin (C,D) from limbs of 9- to 16-day-oldembryos were analysed on 12-5 % polyacrylamide-SDSgels by staining (A,C) and duplicate gels wereelectrophoretically transferred onto nitrocellulose andprobed with antibody QBM-2 (B,D). Although MLC,f isdetected in both total extracts and myosin precipitatesfrom all ages and in myosin precipitated from culturedmyotubes (Mt), MLC3f is first detected in total extracts atday 12 and in precipitated myosin at day 11. MHC,myosin heavy chain; A, actin; MLQf, myosin alkali lightchain 1; MLC3f, myosin light chain 3; Mt, myosinprecipitated from cultured myotubes. Each lane contains25 fig of protein.

676 P. A. Merrifield and 1. R. Konigsberg

1983). MLC3f is also absent from the limb muscu-lature of day-10 and -11 embryos, and is first detectedby immunoblot analysis of total limb extracts in 12-day embryos. The MLC3f content increases in olderembryos until, by day 16, MLC3f is detectable,although faintly, by Coomassie Blue staining of totalprotein.

In determining the precise developmental stage atwhich MLC3f can first be detected in the limb, wethought it important to consider other changes takingplace in this dynamic structure. Because the mass ofdifferentiated muscle changes dramatically duringembryonic limb development relative to other non-muscle tissues (McLennan, 1983a; Butler, Cosmos &Brierley, 1982), we used a low ionic strength precipi-tation to isolate crude myosin fractions from the high-

salt-soluble extracts of developing limbs. A plot ofthe amount of precipitated myosin (as a percentage oftotal protein) versus developmental age thus providesa biochemical representation of how muscle masschanges during limb development (see Fig. 5), andconfirms the histological evidence presented byothers. Preliminary analysis of limb extract from day-12 embryos showed that precipitation of myosin wascomplete by 6h and that this precipitation wasindependent of myosin concentration in the extractover a 10-fold range (Fig. 2). This method is alsoreproducible since, in three different ionic strengthprecipitations of extracts from 12-day limbs, a meanvalue of 22-1% was obtained with a standard devi-ation of ±1-6. Although the myosin precipitated inthis way contains other proteins of the contractile

S1 S 3 S 6 S1 2S2 4

6 7 8 9 10Precipitation time (h)

11 12 77 23 24

Fig. 2. Precipitation kinetics of myosin during dilution of a high salt extract of limb by dialysis against 50 vol. of coldwater containing 1 mM-PMSF. A crude extract from 12-day embryos was diluted 1:1 with glycerol and either dialyseddirectly ( • ) or diluted 1:10 with extraction buffer and then dialysed (•) . At each time point, triplicate samples werecentrifuged to precipitate myosin and the protein content of the pellet was determined using the Lowry assay. In aseparate experiment (see inset), the crude extract (E) and the precipitates (P) and supematants (S) at 1, 3, 6, 12 and24 h were analysed on SDS gels. Each lane contains 25 fig protein stained with Coomassie Blue; MHC, myosin heavychain.

accumulation is nerve dependent 677

apparatus, the supernatant fraction is essentiallydevoid of myosin (Fig. 2). Consequently, by analys-ing a constant amount of precipitated myosin fromembryos of different ages, we can compensate for thedifference in muscle mass between individual em-bryos. Using precipitated myosin in our immunoblotanalysis, rather than total limb extract, we observe asimilar developmental appearance of MLC3f. exceptthat MLC3f is first detected in 11- (rather than 12-)day embryos (Fig. ID). Thus, ionic strength precipi-tation represents a convenient and reproduciblemethod for concentrating and quantifying myosinprior to gel electrophoresis. and provides for agreater degree of sensitivity in our immunoblot assayfor MLC3f.

Since it is impossible to prove a negative, wecannot absolutely rule out the possibility that MLC3f

is synthesized in myotubes cultured in vitro and indeveloping limbs prior to 11 days in ovo. However,we do know that our immunoassay is sufficientlysensitive to detect both alkali light chains in a samplecontaining as little as 1 fig of myosin precipitated fromadult quail muscle (results not shown). Such samplescontain 15 ng of each light chain. Since we routinelyapply 25 pig of precipitated myosin to each lane of ourgels prior to electrophoresis and immunoblot analysis(i.e. 375 ng of each light chain, assuming identicalstoichiometry to adult myosin) MLC3f, if it occurs atall before 11 days of incubation, is present at levelslower than 4 % of that found in adult muscle. This isconsistent with a study of incorporation of [35S]meth-ionine into the light chains of avian muscle cultures(Keller & Emerson, 1980) in which MLC3f was shownto represent approximately 4 % (±3 %) of total lightchain synthesis. Thus, even if MLC3f is present atthese very low levels, its contribution to the physio-logical state of the muscle is of little significance. Thequestion which is a natural consequence of theseobservations concerns the nature of the signal whichis responsible for the activation of MLC3f synthesisand accumulation to physiologically significant levels.

Characteristics of quail limbs grafted ontochorioallantoic membranes of chick hostsIn order to assess whether humoral or neural influ-ences are involved in the expression of MLC3f at day11, we grafted limb buds from 3-day quail embryosonto the chorioallantoic membrane of 7-day-old chickhosts and allowed them to develop for 12 days prior toanalysis for MLC3f content. We reasoned that ifhumoral influences (like hormones) were the stimulusfor MLC3f activation, then aneural grafts wouldproduce mature muscle with normal accumulation ofMLC3f. If, on the other hand, innervation were thestimulus, then only the grafts containing neural tubewould contain detectable levels of MLC3f.

Of a total of 93 grafts performed. 43 of the chickhosts survived until the end of the graft incubationperiod (day 19). Of those chick hosts that survived. 23supported the successful growth and morphologicaldevelopment of grafts. These grafts ranged in weightfrom 22 to 293 mg but only those grafts that attained amass greater than 45 mg were extracted for furtheranalysis (Table 1). Although grafts which includedneural tube seemed to survive better, their averagesize and morphological development were very simi-lar to aneural grafts (Fig. 3). Morphologically, thegrafts ranged from feathered tissue masses of ap-proximately 2 mm in diameter to normal but stuntedlimbs of 13-15 mm in length with normal scale,feather and digit formation (Fig. 4). They differedsignificantly, however, in the amount of myosinprecipitated per mass of total protein in the extract.In the 12 aneural grafts analysed, myosin as apercentage of total protein ranged from 7-5 to 18-8 %with a mean of 12-0% (Table 1). In contrast, themyosin content of innervated grafts ranged from 12-6to 58-8% with a mean value of 31-6 - almost threetimes the amount obtained in aneural grafts. Therange of values for innervated grafts was similar tothat observed in limbs of 9- to 16-day embryos(Fig. 5), with one of the grafts (G33) containing alevel of myosin comparable to that seen in a normallimb from an embryo of the same age (i.e. 15 days).The myosin content of aneural limb bud grafts was in

Table 1. Graft characteristics

Graftno.

G3G20G21G22G24G26G27G29G31G43G45G46

G14G15G16G17G32G33G35G40G41

N.A.

Neural Weighttube (mg)

591002661129952

118995746

111103

+ 60+ 69+ 209+ 136+ 94+ 293+ 156+ 138+ 80

= Not available.

Myosinas %total Mean

protein MLC1 MLC3 ± S.D.

N.A. +10-2 +8-8 +

14-0 +7-5 +

N.A. +112 +18 8 +110 +12-3 +13-8 +11-9 +

22-9 + +12-626-0 + +26-9 + +36-3 + +58-8 + +41-1 + +340 + +25-8 + +

12-0±3-l

316 ± 13 1

678 P. A. Merrifield and I. R. Konigsberg

Fig. 3. Aneural and innervated limb buds of a 3-day quail embryo before (A) and after (B,C) incubation on thechorioallantoic membrane. The area contained by the solid line in panel A represents a typical dissection of a limb budwhich excluded somites and neural tube (i.e. an aneural graft like G27) while the broken line indicates all of the tissuewhich was typically included in an 'innervated' graft like G17. Two representative gTafts which resulted from a 12-dayincubation on the CAM are shown in panels B (G17) and C (G27). G17 was subsequently shown to contain both MLClf

and MLC3f while G27 contains only MLC]f (Table 1). The bar represents 1 mm in each respective field.

all cases below the value obtained for a 9-day em-bryo.

Although the crude myosin prepared by a singleionic strength precipitation contains some noncon-tractile proteins, it is nevertheless a good approxi-mation of the amount of muscle protein per mass oftotal protein. Since all grafts were extracted in thesame way, this value is an estimate of the amount ofdifferentiated muscle tissue contained in each graft.Clearly, those grafts that contained neural tube con-sistently contained more muscle than aneural grafts.

Myosin light chains in aneural and innervated graftsThe most striking difference between aneural andinnervated grafts was the myosin light chain 3 con-tent. When a constant amount of myosin from eachgraft was analysed with QBM-2 following immuno-blotting, MLCif

w a s always present in detectableamounts. MLC3r, however, was never detected in theaneural grafts. In this respect, these myosin prep-arations had the same isoform content as thoseprepared from cultured myotubes or from the limbs

of normal quail embryos sampled prior to 11 days inovo (Fig. 6). In contrast, most of the innervated graftscontained detectable levels of both MLC,f and MLC3f

when analysed under the same conditions (Fig. 7;Table 1). This contrast is most apparent when similaramounts of actomyosin from an innervated (G17) andan aneural (G3) graft are analysed on the sameimmunoblot along with myosin from normal limbs(Fig. 8). MLC3f cannot be detected in the aneuralgraft extract while the amount of MLC3f in theinnervated graft is comparable to that seen in normallimbs from day-11 to -12 embryos. Of the nineinnervated grafts analysed for MLC content, only one(G15) did not contain detectable levels of MLC3f

(Table 1). However, this particular graft also did notcontain MLC)f, suggesting that for some reason it wascompletely devoid of muscle. These results suggestthat the quality of the muscle, as well as the quantity,differs in limb buds grafted with and without neuraltube. Innervated grafts that contained differentiatedmuscle always contained detectable levels of MLC3f.

MLC3f accumulation is nerve dependent 679

mm

mm

'33

Fig. 4. Potential of the CAM grafting procedure.Aneural (G3) and innervated (G33) grafts sometimesattained morphological development comparable to thatof normal limbs from a 16-day embryo (L16). Althoughstunted, G3 contains bone (at arrow), feathers, scales,rudimentary digits and a well-formed ankle joint. G33was the largest graft obtained and contained the greatestmuscle mass (see Table 1). Subsequent immunoblotanalysis revealed that the muscle in G3 contained onlyMLClf whereas G33 contained both MLC]f and MLC3f.

Discussion

Although it has previously been demonstrated thatlimb buds from chick embryos older than stage 18 candevelop cytologically normal muscle when graftedonto the chorioallantoic membrane (Eastlick, 1932;Kenny-Mobbs, 1985), the present study is the first touse immunological approaches to analyse the con-tractile protein isoforms of muscle from CAM grafts

to assess its degree of maturation. Using a mono-clonal antibody that can detect the developmentallyregulated accumulation of MLC3f, we have demon-strated that the muscle that forms in aneural limb budgrafts expresses MLC,f but not MLC3f, in spite of thefact that it has a chronological age of 15 days. In thisrespect, it is embryonic-like and has not maturednormally. Since these grafts had to be exposed to theblood supply of the normal host in order to survive,this observation suggests that circulating hormonesdo not play a major role in activating the accumu-lation of MLC3f. If hormonal influences play a role, itis strictly permissive and not sufficient to allow MLC3f

accumulation to reach detectable levels.Our observation that grafts containing neural tube

and cultured on the CAM under similar conditionsaccumulate both MLCif and MLC3f strongly suggeststhat some influence of the neural tube is responsiblefor the accumulation of MLC3f that occurs normallyin developing limb muscle in vivo. At present, theprecise nature of this neural influence is unknown.However, experiments by others on cultured chickmuscle cells have demonstrated that MLC3f accumu-lation can be promoted by electrostimulation (Srihari& Pette, 1981) and is positively correlated with a highlevel of contractile activity (Moss, Micou-Eastwood& Strohman, 1986). Recently, it has been reportedthat, at least in the case of troponin C, coculture withnerve or culture in the presence of nerve extractresults in normal switching of tissue-specific isoforms(Toyota & Shimada, 1983). In addition, denervationexperiments on newly hatched chicks have demon-strated that the expression of MLC3f can be repressedby denervating the muscle and rendering it paralysed(Saitoh, Kitani & Obinata, 1983). This latter exper-iment should be evaluated carefully, however, sincedenervation is known to promote degeneration andsubsequent regeneration of muscle fibres. While it isdifficult to quantify the spontaneous contractile ac-tivity of the limb grafts described in this study, it isimportant to note that several of the innervated graftswere observed to twitch spontaneously during re-moval from the CAM. Aneural grafts, on the otherhand, were never observed to twitch. Experimentsare currently in progress to determine whether treat-ment of developing embryos and innervated CAMgrafts with curare will affect the normal accumulationof MLC3f in the functionally denervated muscles.

The increased myosin content per gram tissue thatwe observe in our innervated grafts may be a result ofthe same trophic influence that is responsible forpromoting MLC3f accumulation, or may be promotedindependently by the neural tube. Although thetrophic effect of nerve on muscle formation in limbbud grafts has been recognized since 1943 (Eastlick,1943; Bradley, 1970; Kenny-Mobbs & Hall, 1983), the

680 P. A. Merrifield and I. R. Konigsberg

70-

60-

Q.

40-

C

so

30 G41G14

G16

G15,.

10 11 12Embryonic age (days)

13 14 15 16

Fig. 5. Myosin content of developing limbs and innervated grafts as a measure of the relative mass of muscle. Myosinwas precipitated from limb extracts (in triplicate ±S.D. • ) and grafts (see Table 1 for characteristics), quantified usingthe Lowry assay and expressed as a % of the total protein present in the extract. Aneural grafts all contained lessmyosin (as a relative %) than limbs of 9-day-old embryos and were not plotted on the graph. Innervated grafts containamounts of myosin comparable to that of normal limbs from 9- to 15-day embryos.

MHC

MLd

i' - » m MLC3 — ^6 9 111316 G3 Al Ab 6 91113 16G3AIAb

Fig. 6. Myosin alkali light chain content of an aneuralgraft in relation to normal limb muscle. Total limbextracts from limbs of 6-, 9-, 11-, 13- and 16-day-old quailembryos and from an aneural graft (G3) wereelectrophoresed on 12-5 % SDS gels and either stained(A) or immunoblotted for analysis with antibody QBM-2(B). Al, extract of adult limb muscle; Ab, purifiedmyosin from adult breast muscle; MHC, myosin heavychain; MLC^ myosin light chain 1; MLC3, myosin lightchain 3.

mechanism for this induction has not been elucidated.While the migration of muscle cell precursors fromthe somite has been implicated in affecting musclemass in some of the CAM-grafting experiments doneby others (Kenny-Mobbs, 1985), this is clearly not thecase in our study since limb buds were dissected fromstage-20 embryos after the migration of somitic cellshad been completed (Chevallier, 1978). In a verycareful study, Kenny-Mobbs (1985) has confimed thatmyogenesis in H.H. stage 18 and later stage wingbuds is not affected by the presence or absence of thesomitic tissues.

In related studies, McLennan (1983i») has shownthat the functional denervation of chick hindlimbmuscles with curare inhibits the normal formation ofsecondary myotubes in developing chick embryos,with a concomitant reduction in muscle mass. Simi-larly, others have shown a dramatic decrease in thewing muscle cross-sectional area (Butler et al. 1982)and myotube number (Phillips & Bennett, 1984) indenervated chick embryos relative to control birds.Although there is a correlation between these twoparameters of growth, it is obvious that the failure ofnerve-dependent secondary myotubes to differentiatenormally cannot account for all of the deficit in the

MLC3f accumulation is nerve dependent 681

32 33 31 43 45 46 32 33 31 43 45 46nt -nt +nt

Fig. 7. Myosin alkali light chain content of representativeaneural and innervated grafts. Myosin precipitated frominnervated grafts (32, 33) and aneural grafts (31, 43, 45,46) were electrophoresed on 12-5 % SDS gels and stained(A) or immunoblotted (B) for analysis with antibodyQBM-2. MLQtf can only be detected in the innervatedgrafts, in spite of the fact that grafts 45 and 46 werelarger than the innervated graft 32 and constant amountsof precipitated myosin (25 fig) from each sample wereanalysed.

-11 -12 "17 M

Fig. 8. Immunoblot analysis of the myosin alkali lightchain content of aneural (G3) and innervated (G17)grafts in relation to normal limb muscle. Myosin wasprecipitated from limb extracts of day-11 and -12 embryos(Lll, L12) as well as from grafts G3 and G17 andcompared by immunoblot analysis. 50 fig of precipitatedmyosin from each graft or normal limb and 10 fig ofactomyosin (A) and purified myosin (M) were appliedper well. Only the light chain portion of the blot isshown. The amount of MLC3f in G17 is comparable tothat seen in day-11 to -12 limbs from developing embryos.

size of aneural muscles (McLennan, 19836). Sinceboth the myotube number and the muscle cross-sectional area actually decrease in denervated em-bryos after stage 42, it is clear that, as both Hunt(1932) and Eastlick (1943) observed, nerve plays arole in maintaining muscle once it has formed. Thus,the decreased muscle mass that we observe in ouraneural grafts may result from the absence of second-ary myotube formation, an increased rate of muscledegeneration or perhaps from a combination of thesetwo factors. We have no way, at present, to evaluatethe relative importance of hyperplasia and hypertro-phy in this trophic effect.

The precise relationship between increased musclemass and MLC3f accumulation in innervated grafts isnot clear. While it is interesting to speculate that bothMLC3f accumulation and increased muscle mass maybe associated with the appearance of secondarymyotubes, there is no evidence at present to supportthis hypothesis. Experiments are currently underway, however, to examine whether MLC3f accumu-lates preferentially in secondary myotubes duringnormal limb development and in innervated limbbuds grafted onto the CAM.

Our demonstration that nervous influence plays arole in the relative expression of MLClf and MLC3f isespecially important in light of recent evidence thatthese two proteins are actually transcribed from thesame gene (Nabeshima, Fujii-Kuriyama, Muramatsu& Ogata, 1984). Since MLClf and MLC3f mRNAsmay be produced by differential splicing and/or theuse of different initiation sites of the same gene, ourresults suggest that nerve-muscle interactions canregulate gene expression at the level of RNA process-ing. It is not yet known if other developmentallyregulated contractile protein isoforms (i.e. a--cardiacand (^-skeletal actin; embryonic, neonate and adultmyosin heavy chain) are regulated by neural influ-ences in a similar way to MLC3f-. Experiments onnerve and hormonal regulation of myosin heavy chainisoform switching in mammals have, to date, beenhighly contradictory. While there is good evidencethat neonatal and adult isoforms of myosin heavychain are induced in cultured mouse muscle fibres bycoculture with embryonic mouse spinal cord (Ecob-Prince, Jenkison, Butler-Browne & Whalen, 1986),the accumulation of adult myosin heavy chains hasalso been reported in pure cultures of differentiatedmouse C2Ci2 cells in the absence of nerve (Silber-stein, Webster, Travis & Blau, 1986). Evidence fromin vivo experiments have also demonstrated thatthyroid hormone can have a profound effect on theexpression of myosin heavy chain isoforms in embry-onic and adult rats (Johnson etal. 1980; Gambke etal.1983; Butler-Browne, Herlicoviez & Whalen, 1984;Whalen etal. 1985; Izumo et al. 1986). Denervation,

682 P. A. Merrifield and I. R. Konigsberg

on the other hand, does not seem to affect the normalaccumulation of adult myosin heavy chain in newbornrats (Butler-Browne et al. 1982). These reports mayindicate that different contractile protein multigenefamilies are regulated differently. Alternatively, therole of nervous influences in the process of musclematuration may be much more important in birdsthan in mammals.

Our successful use of innervated limb bud grafts tomimic normal nerve-muscle interactions in the intactembryo is entirely consistent with Eastlick's (1943)earlier demonstration that the normal interneuronalconnections within the spinal column are not requiredfor muscle maintenance by a neural tube segment. Incoelomic grafts of limb buds containing neural tube,those that became attached loosely to the mesenteriesseemed to contain as much striated muscle as thelimbs that were grafted to the lateral body wall andwere thus innervated by intact trunk or limb nerves.Similarly, Weiss (1950) demonstrated that neuraltube transplanted together with an embryonic am-phibian limb into the tail fin of a larva establishedfunctional connections with the limb and dischargedimpulses into it in the absence of external stimulation.In addition, we now know that the uncoordinatedmovements of the chick limb which can be observedfrom day 5 to ca. 19 days of development in ovo arereflexogenic and thus are independent of the brain(Hamburger & Balaban, 1963). Although the rate ofspontaneous movements in the embryo is reduced byisolating the spinal cord from the brain, they docontinue at a significant rate (Hamburger, Wenger &Oppenheim, 1966). It is not clear at this point,however, whether normal synaptic transmission andgeneration of an action potential in the muscle isnecessary for muscle maturation or whether someother trophic factor from the nerve is responsible forthe maturational events which we have observed inour innervated grafts and which are known to occurduring normal development.

We would like to thank Anita Mentzer for her excellenttechnical assistance and Dr Michael Payne for generatingthe QBM-2 antibody used in these studies. This work wassupported by grants from the NIH (07083) and the Muscu-lar Dystrophy Association of America (MDA) to IRK. Partof this work was carried out while PAM was a postdoctoralfellow of the Muscular Dystrophy Association of Americaand was presented during a poster session at the UCLASymposium on the Molecular Biology of Muscle Develop-ment held in Park City, Utah; March 15-22, 1985.

References

BADER, D., MASAKI, T. & FISCHMAN, D. (1982).

Immunochemical analysis of myosin heavy chain during

avian myogenesis in vivo and in vitro. J. Cell Biol. 95.763-770.

BANDMAN, E., MATSUDA, R. & STROHMAN, R. C. (1982).

Developmental appearance of myosin heavy and lightchain isoforms in vivo and in vitro in chicken skeletalmuscle. Devi Biol. 93, 508-518.

BRADLEY, S. J. (1970). An analysis of self-differentiationof chick limb buds in chorioallantoic grafts. J. Anat.107, 479-490.

BULLER, A. J., ECCLES. J. C. & ECCLES, R. M. (1960).

Interactions between motoneurons and muscles inrespect of the characteristic speeds of their responses.J. Physiol., Lond. 150, 417-439.

BURRIDGE, K. & BRAY, D. (1975). Purification and

structural analysis of myosins from brain and othernon-muscle tissues. J. molec. Biol. 99, 1-14. (1971).

BUTLER, J., COSMOS, E. & BRIERLEY, J. (1982).

Differentiation of muscle fiber types in aneurogenicbrachial muscles of the chick embryo. J. exp. Zool. 224,65-80.

BUTLER-BROWNE, G., HERLICOVIEZ, D. & WHALEN, R. G.

(1984). Effects of hypothyroidism on myosin isozymetransitions in developing rat muscle. Fed. Eur.Biochem. Soc. 166, 71-75.

BUTLER-BROWNE, G. S., BUGAISKY, L. B., CUENOUD, S.,

SCHWARTZ, K. & WHALEN, R. G. (1982). Denervationof newborn rat muscles does not block the appearanceof adult fast myosin heavy chain. Nature, Lond. 299,830-833.

CHEVALLIER, A. (1978). Migration of somitic cells into thesomatopleure mesoderm of the limb anlage. WilhelmRoux Arch, devl Biol. 184, 57-73.

CHI, J. C , FELLINI, S. A. & HOLTZER, H. (1975).

Differences among myosins synthesized in non-myogenic cells, presumptive myoblasts, and myoblasts.Proc. natn. Acad. Sci. U.S.A. 72, 4999-5003.

CHRIST, B., JACOB, H. J. & JACOB, M. (1977).

Experimental analysis of the origin of the wingmusculature in avian embryos. Anal. Embryol. 150,171-J86.

CROW, M. T., OLSON, P. S. & STOCKDALE, F. E. (1983).

Myosin light-chain expression during avian muscledevelopment. J. Cell Biol. 96, 736-744.

DOSSEL. W. E. (1958). Preparation of tungsten micro-needles for use in embryological research. Lab. Invest.7, 171-173.

EASTLICK, H. L. (1943). The development of muscle innerveless and innervated grafts. J. exp. Zool. 93, 27-45.

ECOB-PRINCE, JENKISON, M., BUTLER-BROWNE, G. &

WHALEN, R. G. (1986). Neonatal and adult myosinheavy chain isoforms in a nerve-muscle culture system.J. Cell Biol. 103, 995-1005.

FAIRBANKS, G., STECK, T. L. & WALLACH, D. F. H.

(1971). Electrophoretic analysis of the majorpolypeptides of the human erythrocyte membrane.Biochem. J. 10, 2606-2616.

FOUVET, B. (1973). Innervation et Morphoge'nese de lapatte chez l'embryon de poulet. I. Mise en place deI'innervation normale. Archs Anat. Microsc. 62,281-291.

MLC3f accumulation is nerve dependent 683

GAMBKE, B.. LYONS, G. E.. HASELGROVE. J.. KELLY. A. &

RUBENSTEIN. N. A. (1983). Thyroidal and neuralcontrol of myosin transitions during development of ratfast and slow muscles. Fed. Eur. Biochem. Soc. 156,335-339.

GUTMANN. E. (1976). Neurotrophic relations. A. Rev.Physiol. 38, 177-216.

HALL, B. K. (1978). Grafting of organs and tissues to thechorioallantoic membrane of the embryonic chick. TCAManual 4, 881-883.

HAMBURGER, V. (1942). Chorioallantoic grafts. In AManual of Experimental Embryology' (ed. ViktorHamburger) pp. 143-149. University of Chicago Press.

HAMBURGER, V. & BALABAN, M. (1963). Observationsand experiments on spontaneous rhythmical behaviorin the chick embryo. Devi Biol. 7, 533-545.

HAMBURGER, V. & HAMILTON, H. L. (1951). A series ofnormal stages in the development of the chick embryo.J. Morph. 88, 49-92.

HAMBURGER, V., WENGER, E. & OPPENHEIM, R. W.

(1966). Motility in the chick embryo in the absence ofsensory input. J. exp. Zool. 162, 133-160.

HOWARD, E. (1953). Some effects of NaCl concentrationon the development of early chick blastoderms inculture./, cell. comp. Physiol. 41, 237-259.

HUNT, E. (1932). The differentiation of chick limb budsin chorio-allantoic grafts, with special reference to themuscles. J. exp. Zool. 62, 57-91.

IZUMO, S., NADAL-GINARD, B. & MAHDAVI, V. (1986).

All members of the MHC multigene family respond tothyroid hormone in a highly tissue-specific manner.Science 231, 597-600.

JOHNSON, M. A., MASTAGLIA, F. L., MONTGOMERY, A.

G., POPE, B. & WEEDS, A. G. (1980). Changes inmyosin light chain in the rat soleus afterthyroidectomy. Fed. Eur. Biochem. Soc. 110, 230-235.

JOLESZ, F. & SRETER, F. A. (1981). Development,innervation, and activity pattern induced changes inskeletal muscle. A. Rev. Physiol. 43, 531-552.

KELLER, L. R. & EMERSON, C. P., JR (1980). Synthesis ofadult myosin light chains by embryonic musele cultures.Proc. natn. Acad. Sci. U.S.A. 77, 1020-1024.

KENNY-MOBBS, T. (1985). Myogenic differentiation inearly chick wing mesenchyme in the absence of thebrachial somites. J. Embryol. exp. Morph. 90, 415-436.

KENNY-MOBBS, T. & HALL, B. K. (1983). Muscledifferentiation in cultures and chorio-allantoic grafts ofpre-wing territories. In Limb Development andRegeneration, (ed. A. Caplan) Part B, pp. 323-332.New York: Alan R. Liss, Inc.

KROPP, K., GULICK, J. & ROBBINS, J. (1986). A canonicalsequence organization at the 5'-end of the myosinheavy chain genes. J. biol. Chem. 261, 6613-6618.

LAEMMLI, U. K. (1970). Cleavage of structural proteinsduring the assembly of the head of the bacteriophageT4. Nature, Lond. 227, 680-685.

LOWRY, O. H., ROSEBROUGH, N. J., FARR, A. L. &

RANDALL, R. J. (1951). Protein measurement with theFolin phenol reagent. J. biol. Chem. 193, 265-275.

MATSUDA, R., BANDMAN, E. & STROHMAN, R. C. (1983).

Regional differences in the expression of myosin light

chains and tropomyosin subunits during developmentof chicken breast muscle. Devi Biol. 95. 484-491.

MATSUDA, R., OBINATA, T. & SHIMADA. Y. (1981). Typesof troponin components during development of chickenskeletal muscle. Devi Biol. 82, 11-19.

MCLENNAN. I. S. (1983rt). The development of thepattern of innervation in chicken hindlimb muscles:evidence for specification of nerve-muscle connections.Devi Biol. 97, 229-238.

MCLENNAN, I. S. (1983b). Neural dependence andindependence of myotube production in chickenhindlimb muscles. Devi Biol. 98, 287-294.

MERRIFIELD, P. A. & KONIGSBERG, I. R. (1986).

Reflective densitometry of Western blots to quantitatethe developmentally regulated accumulation of myosinalkali chain 3. Biochem. biophys. Res. Comnmn. 135.778-784.

MERRIFIELD, P. A., PAYNE, M. R. & KONIGSBERG, I. R.

(1983). Isoform specificity of monoclonal hybridomaantibodies to quail skeletal muscle myosin subunits.Biochem. biophys. Res. Commun. 113, 407-417.

MONTARRAS, D., FISZMAN, M. Y. & GROS, F. (1982).

Changes in tropomyosin during development of chickembryonic skeletal muscles in vivo and duringdifferentiation of chick muscle cells in vitro. J. biol.Chem. 257, 545-548.

Moss. P., MICOU-EASTWOOD. J. & STROHMAN, R. C.

(1986). Altered synthesis of myosin light chains isassociated with contractility in cultures ofdifferentiating chick embryo breast muscle. Devi Biol.114,311-314.

MURRAY, P. D. F. (1926). An experimental study of thedevelopment of the limbs of the chick. Linnean Soc.New South Wales Proc. 51, 186-263.

MURRAY, P. D. F. & HUXLEY, J. S. (1925). Self-differentiation in the grafted limb bud of the chick.J. Anat. 59, 379-385.

NABESHIMA, Y., FUJII-KURIYAMA, Y., MURAMATSU, M. &

OGATA, K. (1984). Alternative transcription and twomodes of splicing result in two myosin light chains fromone gene. Nature, Lond. 308, 333-338.

PATERSON, B. M. & ELDRIDGE, J. D. (1984). a- Cardiacactin is the major sarcomeric isoform expressed inembryonic avian skeletal muscle. Science 224,1436-1438.

PHILLIPS. W. D. & BENNETT, M. R. (1984).

Differentiation of fiber types in wing muscles duringembryonic development: effect of neural tube removal.Devi Biol. 102, 112-123.

ROBBINS, J., FREYER, G. A., CHISHOLM, D. & GILLIAM, T.

C. (1982). Isolation of multiple genomic sequencescoding for chicken myosin heavy chain protein. J. biol.Chem. 257, 549-556.

ROY, R. K., SRETER, F. A. & SARKAR, S. (1979). Changesin tropomyosin subunits and myosin light chains duringdevelopment of chicken and rabbit striated muscles.Devi Biol. 69, 15-30.

RUBINSTEIN, N. A., PEPE, F. A. & HOLTZER, H. (1977).

Myosin types during development of embryonicchicken fast and slow muscles. Proc. natn. Acad. Sci.U.S.A. 74, 4524-4527.

684 P. A. Merrifield and I. R. Konigsberg

RUSHBROOK, J. I. & STRACHER, A. (1979). Comparison ofadult, embryonic and dystrophic myosin heavy chainsfrom chicken muscle by sodium dodecyl sulfatepolyacrylamide gel electrophoresis and peptidemapping. Proc. natn. Acad. Sci. U.S.A. 76, 4331-4334.

SAITOH, O., KITANI, S. & OBINATA, T. (1983). Transitionof muscle protein isoform caused by denervation. Dev.Growth Diff. 25, 416.

SATO, K., HOSHINO, T., MIZUMA, Y. & NISHIDA, S.

(1971). A series of normal stages in the earlydevelopment of the Japanese quail, Coturnix cotumixjaponica, embryo. Tohoku J. Agri. Res. 22, 80-95.

SCHWARTZ, R. J. & ROTHBLUM, K. (1980). Regulation ofmuscle differentiation: Isolation and purification ofchick actin messenger ribonucleic acid and quantitationwith complementary deoxyribonucleic acid probes.Biochem. 19, 2506-2514.

SILBERSTEIN, L., WEBSTER, S. G., TRAVIS, M. & BLAU, H.

M. (1986). Developmental progression of myosin geneexpression in cultured muscle cells. Cell 46, 1075-1081.

SRIHARI, T. & PETTE, D. (1981). Myosin light chains innormal and electrostimulated cultures of embryonicchicken breast muscle. Fed. Eur. Biochem. Soc. 123,312-314.

STOCKDALE, F., RAMAN, N. & BADEN, H. (1981). Myosinlight chains and the developmental origin of fastmuscle. Proc. natn. Acad. Sci. U.S.A. 78, 931-935.

TAKANO-OHMURO, H., OBINATA, T., KAWASHIMA, M.,

MASAKI, T. & TANAKA, T. (1985). Embryonic chicken

skeletal, cardiac and smooth muscle expresses acommon embryo-specific myosin light chain. J. CellBiol. 100, 2025-2030.

TOWBIN, H., STAEHELIN, T. & GORDON, J. (1979).Electrophoretic transfer of proteins frompolyacrylamide gels to nitrocellulose sheets: procedureand some applications. Proc. natn. Acad. Sci. U.S.A.76, 4350-4354.

TOYOTA, N. & SHIMADA, Y. (1983). Isoform variants oftroponin in skeletal and cardiac muscle cells culturedwith and without nerves. Cell 33, 297-304.

UMEDA, P. K., KAVINSKY, C. J., SINHA, A. M., Hsu, H.J. JAKOVCIC, S. & RABINOWTTZ, M. (1983). ClonedmRNA sequences for two types of embryonic myosinheavy chains from chick skeletal muscle. 11.Expression during development using SI nucleasemapping. J. biol. Chem. 258, 5206-5214.

WEISS, P. (1950). The deplantation of fragments ofnervous system in amphibians. /. exp. Zool. 113,397-461.

WHALEN, R. G., TOUTANT, M., BUTLER-BROWNE, G. &WATKINS, S. C. (1985). Heriditary pituitary dwarfism inmice affects skeletal and cardiac myosin isozymetransitions differently. J. Cell Biol. 101, 603-609.

WlNKELMANN, D. A., LOWEY, S. & PRESS, J. L. (1983).Monoclonal antibodies localize changes on myosinheavy chain isozymes during avian myogenesis. Cell 34,295-306.

(Accepted 3 August 1987)