J. Embryol. exp. Morph. Vol. 35, 2, pp. 355-367, 1976 3 5 5

Printed in Great Britain

Isoenzyme transitions of creatinephosphokinase, aldolase and phosphoglycerate

mutase in differentiating mouse cells

By EILEEN D. ADAMSON1

From the Department of Zoology, University of Oxford

SUMMARYExtracts of embryonic mouse tissues (skeletal, cardiac and smooth muscle, and brain) were

analysed by Cellogel electrophoresis, for their isoenzymic distributions of three enzymes,creatine phosphokinase, aldolase and phosphoglycerate mutase. Embryonic tissues from the12th day to the end of gestation were examined for isoenzyme transitions, and it was foundthat the adult forms of these enzymes appeared during gestation. Extracts from clonedteratocarcinoma cells were similarly examined in order to determine their degree of bio-chemical differentiation. Undifferentiated embryonal carcinoma cells contained only theearly embryonic forms of all three enzymes, while differentiated cells formed in vivo, and insome cases in vitro, started to express the adult types of creatine phosphokinase and aldolase.Thus, biochemical parallels have been demonstrated between developing embryonic tissuesand teratocarcinoma cells differentiating in vitro.

INTRODUCTION

It is known that cells of embryoid bodies and derived cell lines can differentiateboth in vivo and in vitro (reviewed by Martin, 1975). In most cases the differentia-tion has been followed by histology and they form recognizable muscle, nerve,cartilage, pigmented cells and keratinizing and glandular epithelium. Bio-chemical studies have shown that during the differentiation of embryoid bodiesthe specific activity of alkaline phosphatase and protease declines; these areenzymes characteristic of the embryonal carcinoma cells (Bernstine, Hooper,Grandchamp & Ephrussi, 1973; Hall et al 1975). As the differentiation proceedsso the specific activity of acetylcholinesterase and creatine phosphokinaseincreases and one study has suggested the formation of nervous tissue (Levine,Torosian, Sarokhan & Teresky, 1974) while another showed striated muscleformation (Gearhart & Mintz, 1974). The extent to which these enzymaticproperties can be taken as markers of particular cell differentiation was notestablished.

Here I have compared the appearance of three enzymes in developing mouse

1 Author's address: Department of Zoology, University of Oxford, South Parks Road,Oxford 0X1 3PS, U.K.

23-2

356 E. D. ADAMSON

embryos and in teratocarcinomas. Particular patterns of the isoenzymic forms ofthese enzymes are characteristic of muscle and brain tissues in adult mouse andtheir appearance in developing embryonic tissues is recorded. The enzymes arecreatine phosphokinase (CPK, EC 2.7.3.2), fructosediphosphate aldolase(EC 4.1.2.13), and phosphoglycerate mutase (PGM, EC 2.7.5.3). Theseenzymes were chosen because studies on other species have shown that thedistribution of their isoenzymes changes during the development of skeletalmuscle and brain.

CPK is a dimeric enzyme composed of sub-units M and B in homo- or hetero-polymers MM, BB and MB. In the brains of a wide variety of species, only BB ispresent, while in skeletal muscle only the MM form is found. Since the type ofCPK found in the early embryos of several species is usually exclusively BB,there must be a transition of isoenzymic forms during the development ofskeletal muscle (Cao, de Virigilis, Lippi & Coppa, 1971; Turner & Eppen-berger, 1973). PGM is a dimeric enzyme similar to CPK in the form anddistribution of its isoenzymes in several mammalian species. In addition, itsisoenzymic transitions are similar during human muscle development (Omenn &Hermodson, 1975).

Aldolase is a tetrameric enzyme composed of distinct sub-units A, B or C.Adult skeletal muscle usually has only one isoenzyme of aldolase, A4; both Aand B sub-units are found in liver and kidney; A and C sub-units occur in brain(Rutter et al. 1968). In rat embryos the predominant aldolase isoenzyme is A4 andthere is no change of type during the development of cardiac and skeletal muscle.There is, however, a change in developing brain tissue, resulting finally inapproximately equal amounts of five isoenzymes, A4, A3C, A2C2, AC3 and C4

(Turner & Eppenberger, 1973).The transitions of the isoenzymes of CPK and aldolase that occur in develop-

ing skeletal muscle in vivo also occur during the maturation and differentiationof chick myoblasts in vitro (Morris, Cooke & Cole, 1972; Turner, Maier &Eppenberger, 1974) and rat myoblasts in vitro (Yaffe & Dym, 1972). I have there-fore investigated the possibility of using isoenzymic analyses to determine thetype and the degree of differentiation of cultured teratocarcinoma cells.

MATERIALS AND METHODS

(1) Biological materials

Several stocks of mice were used in this study and no differences were detectedbetween them. PO mice were obtained from the Pathology department of theUniversity of Oxford, U.K., CFLP mice from Carworth Europe, Alconbury,Hunts., U.K. The day of detection of the copulation plug was designated the firstday of gestation. Pregnant female mice were killed by cervical dislocation andthe embryos were dissected in ice-cold solution A of Dulbecco & Vogt (1954)(PBS). Embryonic tissues were collected into centrifuge tubes, drained free ofmedium and either frozen at - 20 °C, or processed immediately.

Isoenzyme transitions in differentiating mouse cells 357

Cultured teratoma cells and solid teratomas were obtained from 129/J andC3H derived growths (Papaioannou, McBurney, Gardner, & Evans, 1975).This material was kindly provided by M. McBurney of this laboratory.

OC15 SI is a cloned line of cells derived from a transplantable teratocarcinomaof strain 129/J. Three other cloned cell lines from transplantable teratocarci-noma of strain C3H mice were also examined. These lines were maintained ashomogeneous populations of embryonal carcinoma cells. All the cell lines coulddifferentiate under suitable conditions of culture (McBurney, in preparation).

(2) Tissue extractions

Adult tissues were homogenized with four volumes of either 0*25 M sucrose,0-025 M Tris-HCl, pH 7-4, 2-5 DIM disodium salt of EDTA (for brain tissue andalso for tissue culture cells), or four volumes of 0*25 M sucrose, 0-025 M Tris-HCl, pH 7-4, 0-025 M magnesium acetate, 0-2 % (v/v) 2-mercaptoethanol (for allother tissues).

Embryonic tissues were suspended in a volume of the above solutions approxi-mately equal to that of the tissue and were sonicated for two bursts of 10 seceach or until disintegrated.

Homogenates were centrifuged at 100000 g for 30 min at 2-4 °C in an M.S.E.High-Speed 65 centrifuge. Supernatants were frozen at - 20 °C in small aliquots.

(3) Electrophoretic analyses

Three to five microlitres of tissue extracts were analysed electrophoretically onCellogel strips (Reeve Angel Scientific Ltd, Whatman Labsales Ltd, Maidstone,Kent, U.K.), using a micro-scale applicator. The electrophoretic buffer and stripsoaking solution was 0-06 M barbitone buffer, pH 8-6, containing 0-lmM2-mercaptoethanol (and also 5 HIM EDTA for aldolase determinations).Electrophoreses (at 4 °C) were run at 250 V (25 V/cm) for 60 min in the case ofCPK separations, 60-75 min for aldolase and 3 | h for PGM on a ShandonModel U77 electrophoresis apparatus (Shandon Scientific Company Ltd, 65Pound Lane, London NW10).

Creatine phosphokinase isoenzymes were stained using an agar overlayessentially as described by Dawson & Eppenberger (1970). Also included in thereaction mixture was 1 HIM adenosine-5'-monophosphate in order to inhibit theenzyme myokinase which also stains with this reaction mixture. Controls wereperformed by staining duplicate Cellogel strips in a reaction mixture which didnot contain the substrate creatine phosphate.

Aldolase isoenzymes were stained similarly by the reaction mixture, essentiallyas described by Penhoet, Rajkumar & Rutter (1966). A modification introducedby Lebherz & Rutter (1969) was used in order to reduce alcohol dehydrogenasestaining. Control strips were incubated with the same agar overlay except thatfructose diphosphate was omitted. Cellogel/agar strips were incubated at 37 °Cin the dark for a period of 10-30 min or until a good formazan colour had

358 E. D. ADAMSON

developed. The strips were processed and 'whitened' by the procedure recom-mended by the manufacturers before being photographed.

Phosphoglycerate mutase isoenzymes were made visible on Cellogel strips bya fluorescent method essentially as described by Omenn & Hermodson (1975).After incubating 1-3 h at 37°, black spots (NAD produced by the enzyme)could be seen against a background of fluorescent NADH in ultra-violet light(365 nm). These were photographed at /3-5 for 45 sec on FP 4 film (Ilford Ltd,Ilford, Essex, U.K.) using a dark green filter.

(4) Chemical materials

General chemicals were analytical reagent grade from Fisons ScientificApparatus, Loughborough, Leics., U.K., or from British Drug Houses, Poole,Dorset, U.K. Enzymes and substrates were obtained from Koch-Light Labora-tories Ltd, Colnbrook, Bucks., U.K.: from Sigma (London) Chemical Company,Kingston-upon-Thames, Surrey, U.K.; or from the Boehringer Corporation(London) Ltd, Lewes, E. Sussex, U.K.

RESULTS

Table 1 shows the distribution of the isoenzymes of CPK and aldolase inadult mouse tissues. It is similar to that of rat and other mammals (Masters,1968; Turner & Eppenberger, 1973). CPK B sub-unit homopolymer is the onlyisoenzyme present in adult brain tissue, and is the predominant form present invery early embryonic cells (see Table 2). Thus the tissues which must transformtheir CPK isoenzyme type at some stage in development are skeletal (striatedmuscle), cardiac muscle, and the smooth muscle of the intestine and bladder.On the other hand, for aldolase the transition must occur in brain, kidney andliver tissues since the major embryonic form is A4 (see Table 3).

(1) CPK isoenzymes in ontogeny

Table 2 shows the results of isoenzymic analyses of four kinds of muscle atseveral stages of development. By the 15th day of development a significantamount of MM activity above control (see below) is detected for the first time inskeletal muscle. It is probable that both the M and B sub-units of the enzyme arebeing synthesized because the heteropolymer MB is visible (Fig. 1 a, track 4) andthis would not appear unless either there was simultaneous production of bothsub-units in the same cytoplasm or continuous dissociation and reassociation ofthe sub-units of the enzyme. The embryonic heart is highly developed morpho-logically by the 12th day, and this correlates with the very early production ofMB (Table 2).

Notice that in Fig. 1 there is a significant amount of staining activity at theposition of MM in early embryonic extracts. This activity is found in extracts ofall early embryonic cells as well as in teratocarcinoma cells. However, since the

Tab

le 1

. D

istr

ibut

ion

of C

PK

and

ald

olas

e is

oenz

ymes

in

vari

ous

adul

t ti

ssue

s of

mou

se

CP

KA

ldol

ase

Tis

sue

MM

MB

BB

B4

B,A

B

,A9

BA

aA

4A

C3

c 4

5 v as* I a »

••. S' I'

Skel

etal

mus

cle

Car

diac

mus

cle

Smoo

th m

uscl

e(a

) U

teru

s(6

) St

omac

h(c

) B

ladd

erB

rain

Liv

erK

idne

ySp

leen

( + )

+ 4-

The

int

ensi

ty o

f enz

ymat

ic r

eact

ion

is in

dica

ted

by th

e nu

mbe

r of

+ s

igns

.(+

) m

eans

a v

ery

fain

t re

actio

n.*

Sim

ilar

activ

ity w

as fo

und

on th

e co

ntro

l or

bla

nk s

trip

.

360 E. D. ADAMSON

Table 2. Changes of CPK isoenzyme distribution in developingmouse muscle tissues

Day of development MM MB BB

A. Skeletal muscle (hind limbs)9th day whole embryo ( + )* . + +

12th day whole bodies +* . + + +12th day skeletal muscle +* . + +13th day skeletal muscle +* . + +14th day skeletal muscle + * . + + +15th day skeletal muscle + + + + + +16th day skeletal muscle + + + + + + +17th day skeletal muscle + + + + +19th day skeletal muscle + + +New-born skeletal muscle + + + +

B. Skeletal muscle (tongue)13th day tongue + * . + + +14th day tongue +*15th day tongue + +16th day tongue + + +17th day tongue + + + +

C. Cardiac muscle12th day ( + )13th day (+) +14th day + +15th day + + + +16th day + + + + +17th day + + + + •19th day + + + +•

D. Smooth muscle (intestine)12th and 13th day bodies +*13th day smooth muscle +*14th day smooth muscle +*15th day smooth muscle (+)*17th day smooth muscle (+)Adult intestine + + + + + +

See notes to Table 1 for explanation of signs.

control gels (with no creatine phosphate in the agar overlay reaction mixture)showed identical staining (Fig. 1 b) at this position as well as at the myokinaseposition, it is likely that this activity is caused by an enzyme activity other thanCPK. In Table 2 therefore, this kind of stained band is denoted by an asterisk.

The necessity for control incubations made at the same time as reactionincubations is demonstrated by Fig. l(b). The spurious stained band at theposition of MM in some of the CPK reaction gels was not identified. It mighthave been MM which also stained in the control gel because of the presence ofcreatine phosphate in the cell extract. This would have been possible only if

Isoenzyme transitions in differentiating mouse cells 361

CPK

I

2

3

4

5

6

Origin- |

"i 4

• | 4

t f tMM MB

1

+

444

BB

Myokinase

(o)

Origin- \

* *i l

•

I

4

€

lb) Control

Fig. 1. Cellogel electrophoretic analysis of CPK at various stages in the developmentof skeletal muscle in the hind-limb of mouse, (a) Reaction strips incubated in thepresence of substrate, (b) control strips corresponding to (a). Sample 7 was run at adifferent time to samples 1 to 6. 1, 12th day of the gestational period; 2, 13th day;3, 14th day; 4,15th day; 5, 16th day; 6,17th day; 7, adult skeletal muscle. Note thesimilar amount of staining at the position of MM in (a) samples 1 to 4, and in (b) thecontrol samples. This enzymatic activity was assumed not to be MM (see text fordetails). For experimental details see the Materials and Methods section.

Table 3. Changes of aldolase isoenzyme distribution indeveloping mouse brain tissue

Day of development A4 A3C AC3

12th day bodies12th day brain13th day brain14th day brain15th day brain16th day brain17th day brain18th day brainAdult brain

See notes to Table 1 for explanation of signs.

cieatine phosphate co-electrophoresed with the MM form of the enzyme. Thiswas tested by adding increasing amounts of creatine phosphate to an adultskeletal muscle extract, electrophoresing these mixtures as usual, and stainingas for a control gel. None of the samples stained at the MM position. Thus theunknown band is unlikely to be MM. Myokinase, however, which has a mobilityintermediate between that of MM and MB, was present as a stained band whoseintensity increased with increasing concentrations of creatine phosphate in the

362 E. D. ADAMSON

(a) Origin (P)

2

3

4

5

6

7

8

1

k1 1

fit

A

•

A2

i ttfo*f i<!

c4c2

Origin

Spurious

Aldolase

li 1 Adult brain| 2 Adult cardiac muscle

? 2 Control/bands

Fig. 2. Cellogel electrophoretic analysis of aldolase (a) in developing mouse brain atthe increasing times in the gestational period 1, 12th day; 2, 13th day; 3, 14th day;4, 15th day; 5, 16th day; 6, 17th day; 7, 18th day; 8, adult brain tissue. Not allsamples were run at the same time, (b) Samples stained for aldolase sometimesrevealed a contaminating set of stained bands. The lower two tracks are the corres-ponding control strips run in parallel with 1, adult brain tissue; 2, adult cardiacmuscle. Note the control stips are similarly stained and that the mobilities of thespurious bands are slightly different from true aldolase isoenzymes.

original sample. This suggested that creatine phosphate was co-electrophoresingwith myokinase rather than with the MM form of CPK.

(2) Aldolase isoenzymes in ontogeny

Table 3 depicts the aldolase isoenzyme transition in brain tissue. By the 14thday of development well-defined heteropolymers of A and C sub-units arepresent, though these are faintly visible also in the 12th and 13th day brain(Fig. Id). About 95 % of all aldolase activity in early embyonic cells is that ofA4 with 5 % A3C. During the development of all types of muscle, this remainsabout the same.

In a small proportion of electrophoretic separations of aldolase isoenzymes,a set of five faint bands moving faster than aldolase was visible in both the reac-tion gel and the control gel (Fig. 2b). The slower-moving bands of this set couldeasily be identified erroneously as A3C, A2C2, etc., and illustrates the necessity forcontrol gels run in parallel with the reaction strips.

(3) Ontogeny of PGM isoenzymes

The data for PGM isoenzyme patterns is shown in Table 4. In the developingskeletal muscle of the hind limb, a transition from BB to MM occurs andthe first appearance of MM is observed at the 15th day of gestation as it iswith CPK. In contrast to the latter, the BB isoenzyme of PGM remains the

Isoenzyme transitions in differentiating mouse cells 363

Table 4. Changes of PGM isoenzyme distribution in developingmouse muscle tissues

Day ofdevelopment MM MB BB

12th day13th day14th day15th day16th day17th dayAdult

13th day14th day15th day16th day17th dayAdult

12th day13th day14th day15th day16th day17th dayAdult

A. Skeletal muscle (hind limbs)

B. Skeletal muscle (tongue)

C. Cardiac muscle

PGM

t t tBB MB MM

Fig. 3. Cellogel electrophoretic analysis of PGM in developing hind-limb muscle ofmouse. 1, Skeletal muscle at the 12th day of gestation; 2, 13th day; 3, 14th day; 4,15th day; 5,16th day; 6,17th day; 7, adult skeletal muscle. The adult form (MM) isjust visible in 15th day muscle but is only clearly stained in the 17th day extracts.See the Materials and Methods section for experimental details.

364 E. D. ADAMSON+ -

Origin Origin

1 mycMM

(")

MB

CPK

tBB

A

(b)

tq<4 A

1<

2c2

Aldolase

Fig. 4. Isoenzyme analyses in teratocarcinoma cells, (a) CPK, (b) aldolase. Track 1,embryonal carcinoma cells of clone OC15 SI; 2, differentiated OC15 SI; 3, solidteratoma tissue. This was produced in vivo from the same teratocarcinoma line(OTT 6050) which was cloned to give OC15 SI.

predominant form to a much later time in development and the Cellogel reactionstrips have to be rather over-incubated to detect the low proportion of the MMform of very early embryonic muscle (Fig. 3). As in the case of CPK, the firstappearance of MM in tongue and heart muscle is earlier (14th and 13th dayrespectively) than in the skeletal muscle of the hind-limb. MM does not becomethe predominant form of PGM until after the 17th day of development in alltypes of muscle.

(4) Are these isoenzymes and their transitions detectable in teratocarcinoma cells?

Solid teratomas consisting of several different types of tissues gave the patternsof CPK and aldolase isoenzymes expected of differentiated tissues. Fig. 4 (a) and(b), track 3, shows that both muscle and brain type CPK and aldolase werepresent in one extract. A cell line (OC15 SI) derived from this teratoma (seeMaterials and Methods) and cultured under conditions to maintain it in anembryonal carcinoma (undifferentiated) form, gave only the embryonic formsof CPK (BB) (together with the spurious band at MM) and aldolase (A^. Whenthis cell line was allowed to differentiate and both nervous- and epidermal-likecells were visible in the culture dish, the aldolase pattern had changed to a 5-banded type typical of well-differentiated brain tissue (Fig. 4b, track 2). On theother hand no change was detected in the CPK pattern (Fig. 4a), or in the PGMpattern. ;

This result has been confirmed for several cell lines which were derived fromdifferent sources of teratocarcinoma. Two lines were from embryoid bodies of aC3H-strain tumour and one was from a solid tumour of C3H origin. Embryoidbodies, like embryonal carcinoma cells in culture, contained only the BB form ofCPK and A4 aldolase. In all cases, after partial differentiation to nerve-type cellsduring in vitro culture, the aldolase pattern had transformed into a 3- to 5-banded type typical of brain tissue. In addition, in one dish, a moderate amountof smooth muscle tissue appeared (visible by phase-contrast optics), and when

Isoenzyme transitions in differentiating mouse cells 365

the CPK isoenzyme pattern was examined a very small amount of MB hetero-polymer was observed. There was no detectable MB or MM form of PGM inany of the differentiated cell line extracts.

DISCUSSION

(1) Isoenzyme transitions in development

The study of isoenzymic transitions in developing mouse tissues gave resultssummarized in Tables 2, 3, and 4. These are intended to serve as normal tables ofsome of the changing biochemical events during differentiation. The adult formsof CPK and PGM appear at different times in different muscle tissues; they areapparent in cardiac muscle at the 12th day, in tongue by the 14th day, in mostskeletal muscle by the 15th day, and in some kinds of smooth muscle at the17th day. Developing brain tissue acquires the typical adult-type isomers ofaldolase on the 14th day of gestation. In summary, CPK and PGM would appearto be markers of some types of differentiating muscle cells, and aldolase of braincells.

Although the proportion of the adult form of PGM in early embryonic muscletissues was lower than that of CPK, the MB and MM forms of PGM appearedat the same time in development. This was equally true of hind-limb skeletalmuscle, tongue and cardiac muscle, but the adult forms appeared at a differenttime in each tissue. The degree of morphological development of the tissuescould be roughly correlated with the time of appearance of adult isoenzymicforms. Possibly the parallel appearance of the differentiated forms of CPK andPGM denotes a linked control system such as that suggested by Turner &Eppenberger (1973). These authors argue that a number of proteins such asCPK, aldolase and myosin may be co-ordinately regulated during muscledevelopment. It is interesting that striations are not obvious in sections of 15thday hind-limb examined in the light microscope and so the above biochemicalchanges occur in advance of the more obvious histological signs of differentiation.

Although the appearance of adult isomers of CPK and PGM was simultaneousin all types of developing muscle, the latter enzyme could prove to be a morereliable marker of the biochemical differentiation of muscle. This is becauseextracts of all tissues at all stages in development showed no staining on thecontrol gel. There is therefore no ambiguity in the interpretation of the results.

(2) Isoenzyme transitions in teratocarcinoma cells

This is the first report of isoenzyme analyses of cloned teratocarcinoma celllines. Figure 4 shows that undifferentiated embryonal carcinoma cells are verylike early embryonic cells in having only the early embryonic isoenzymes,namely, the BB form of CPK, the BB form of PGM (not shown), and the A4

form of aldolase. The patterns for embryoid bodies formed in vivo were similar.When undifferentiated embryonal carcinoma cells of four different cloned cell

366 E. D. ADAMSON

lines produced in this laboratory were cultured under conditions to promotetheir differentiation, they formed morphologically recognisable nerve, epi-dermal and muscle cells. Extracts from a cloned line (OC15 SI) which producesa large proportion of nerve-type cells gave a highly differentiated brain-typepattern of aldolase isoenzymes. One cloned cell line which differentiated to givesome patches of smooth-muscle-like cells as well as nerve cells, produced thecorresponding isoenzyme patterns, that is, a band (rather faint) of the MB formof CPK and a set of four bands of A and C aldolases.

For mouse teratocarcinomas to be useful as an alternative to embryos in thestudy of cell differentiation, they have to be shown to have properties whichclosely parallel those of embryos. The stem cells of the tumours (embryonalcarcinoma cells) have been shown to be pluripotent (Kleinsmith & Pierce,1964), and similar to early embryonic cells, both ultrastructurally (Damjanov,Solter & Skreb, 1971) and antigenically (Artzt et al. 1973). Histochemical studieshave shown that the distribution of alkaline phosphatase follows the patternin embryos (Bernstine et al. 1973; Solter, Damjanov & Skreb, 1973). Somebiochemical properties of in vivo embryoid bodies allowed to differentiate invitro have been determined. Gearhart & Mintz (1974) followed the increasingacetylcholinesterase specific activities of such cultures and identified striatedmuscle fibres which eventually formed in them. Levine et al. (1974) showed thatspecific activities of both acetylcholinesterase and CPK increase during cultureof in vivo embryoid bodies, but they correlated these with the production ofnerve cells on the basis of histology and the identification of the BB type ofCPK. Since embryoid bodies were shown (above) to contain the BB form ofCPK before differentiation, there was no transition of isoenzyme type duringdifferentiation.

It is important to show that clonal teratocarcinoma cells differentiating underdefined conditions in vitro reflect the orderly processes which occur during thedevelopment of the embryo. Martin & Evans (1975) showed that one suchclonal line produced embryoid bodies in vitro, and these contained the correctdistribution of alkaline phosphatase. Here I have described the distribution inembryos of the isoenzymes of three enzymes, and have shown that differentiatingteratocarcinoma cell clones change their patterns similarly.

I wish to thank Miss S. E. Ayers for skilful technical assistance. I am grateful to Dr C. F.Graham for helpful discussions and criticisms, and to Dr M. McBurney for cloned terato-carcinoma cell lines. This work was supported by the Cancer Research Campaign.

REFERENCES

ARTZT, K., DUBOIS, P., BENNETT, D., CONDAMINE, H., BABINET, C. & JACOB, F. (1973).Surface antigens common to mouse cleavage embryos and primitive teratocarcinoma cellsin culture. Proc. natn. Acad. Sci. U.S.A. 70, 2988-2992.

BERNSTINE, E. G., HOOPER, M. L., GRANDCHAMP, S. & EPHRUSSI, B. (1973). Alkaline phos-phatase activity in mouse teratoma. Proc. natn. Acad. Sci. U.S.A. 70, 3899-3903.

Isoenzyme transitions in differentiating mouse cells 367CAO, A., DE VIRIGILIS, S., LTPPI, C. & COPPA, G. (1971). Serum and muscle creatine-kinase

isoenzymes and serum aminotransferase isoenzymes in progressive muscular dystrophy.Enzyme 12, 49-62.

DAMJANOV, I., SOLTER, D. & SKREB, N. (1971). Enzyme histochemistry of experimentalembryo-derived teratocarcinomas. Z. Krebsforsch. 76, 249-256.

DAWSON, D. M. & EPPENBERGER, H. M. (1970). Creatine phosphokinase. In Methods inEnzymology, 17 (ed. H. Tabor & C. W. Tabor) pp. 995-1002. New York: Academic Press.

DULBECCO, R. & VOGT, M. (1954). Plaque formation and isolation of pure lines with polio-myelitis viruses. / . exp. Med. 99, 167-182.

GEARHART, J. D. & MINTZ, B. (1974). Contact-mediated myogenesis and increased acetyl-cholinesterase activity in primary cultures of mouse teratocarcinoma cells. Proc. natn.Acad. Sci. U.S.A. 71, 1734-1738.

HALL, J. D., MARSDEN, M., RIFKIN, D., TERESKY, A. K. & LEVINE, A. J. (1975). The in vitrodifferentiation of embryoid bodies produced by a transplantable teratoma of mice. InTeratomas and Differentiation (ed. M. I. Sherman & D. Solter), pp. 251-267. New York:Academic Press.

KLEINSMITH, L. J. & PIERCE, G. B., Jr. (1964). Multipotentiality of single embryonal car-cinoma cells. Cancer Res. 24, 1544-1551.

LEBHERZ, H. G. & RUTTER, W. J. (1969). Distribution of fructose diphosphate aldolasevariants in biological systems. Biochemistry, N.Y. 8, 109-121.

LEVINE, A. J., TOROSIAN, M., SAROKHAN, A. J. & TERESKY, A. K. (1974). Biochemicalcriteria for the in vitro differentiation of embryoid bodies produced by a transplantableteratoma of mice. The production of acetylcholinesterase and creatine phosphokinase byteratoma cells. / . cell. Physiol. 84, 311-318.

MARTIN, G. R. (1975). Teratocarcinoma as a model system for the study of embryogenesisand neoplasia: a review. Cell 5, 229-244.

MARTIN, G. R. &EVANS, M. J. (1975). Differentiation of clonal lines of teratocarcinoma cells:formation of embryoid bodies in vitro. Proc. natn. Acad. Sci. U.S.A. 72, 1441-1448.

MASTERS, C. J. (1968). The ontogeny of mammalian fructose-l,6-diphosphate aldolase.Biochim. biophys. Acta 167, 161-171.

MORRIS, G. E., COOKE, A. & COLE, R. J. (1972). Isoenzymes of creatine phosphokinase duringmyogenesis in vitro. Expl Cell Res. 74, 582-585.

OMENN, G. S. & HERMODSON, M. A. (1975). Human phosphoglycerate mutase: isozymemarker for muscle differentiation and for neoplasia. In Isoenzymes, vol. Ill (ed. C. L.Markert), pp. 1005-1018. New York: Academic Press.

PAPAIOANNOU, V. E., MCBURNEY, M. W., GARDNER, R. L. & EVANS, M. J. (1975). The fateof teratocarcinoma cells injected into early mouse embryos. Nature, Lond. 258, 70-73.

PENHOET, E., RAJKUMAR, T. & RUTTER, W. J. (1966). Multiple forms of fructose diphosphatealdolase in mammalian tissues. Proc. natn. Acad. Sci. U.S.A. 56, 1275-1282.

RUTTER, W. J., RAJKUMAR, T., PENHOET, E., KOCHMAN, M. & VALENTINE, R. (1968). Aldolasevariants: structure and physiological significance. Ann. N.Y. Acad. Sci. 151, 102-117.

SOLTER, D., DAMJANOV, I. & SKREB, N. (1973). Distribution of hydrolytic enzymes in earlyrat and mouse embryos. Z. Anat. EntwGesch. 139, 119-126.

TURNER, D. C. & EPPENBERGER, H. M. (1973). Developmental changes in creatine kinase andaldolase isoenzymes and their possible function in association with contractile elements.Enzyme 15, 224-238.

TURNER, D. C, MAIER, V. & EPPENBERGER, H. M. (1974). Creatine kinase and aldolaseisoenzyme transitions in cultures of chick skeletal muscle cells. Devi Biol. 37, 63-89.

YAFFE, D. & DYM, H. (1972). Gene expression during differentiation of contractile musclefibres. Cold Spring Harb. Symp. quant. Biol. 37, 543-547.

(Received 24 October 1975)