Cytoplasmic actin and myosin

TIBS- March 1976

55

PP; generated by anabolic processes and this amount will still be insufficient to permit continued glycolysis. Additional PP; must be available from other sources. One possible additional source of PP; is the reaction catalyzed by the newly found acetate kinase [18).

conserving the energy of the PP; bond. In a few instances it appears that such conservation steps are an essential feature in the organisms' metabolic processes. With the exception of acetate kinase (PP;) and PP;ase it appears that the biochemical reactions in which PP; participates are reversible under physiological conditions, thus these reactions may function in both glycolysis and gluconeogenesis. In a number of instances energy-conserving cycles are possible which equilibrate a nucleoside phosphate with PP;. A cell having such a cycle and also an active PP;ase must have developed some compartmentalization system which protects the adenylate energy charge of the cell from being dissipated by the PP;ase. Factors which have tended to cloud the metabolic role of PP; are the scarcity of reliable estimates ofPP; and Mg2 + concentration levels in most tissues and organisms, and the dearth of information on the compartmentalization of the metabolites, metal ions, and enzymes which are concerned with the catabolic reactions involving PP;. Enzymic assay methods [16,21] particularly when coupled with isotope procedures [I] now provide a tool for the estimation of PP; concentrations in the presence of 2 x 105 parts of P;, whereas purely chemical assay methods lose reliability in the presence of 50 parts of P;. It is in the area of free PP; and Mg2 + concentrations and the compartmentalization of metabolites and enzymes that the greatest need for new information exists.

acetyi-P + P;--+acetate+ PP;

This physiologically irreversible reaction could supply all the additional PP; required to support glycolysis provided all acetate produced by the ameba funnels through this channel. However, the last point is still uncertain. A further way by which the additional PP; required to support glycolysis could be gained by the ameba is through glycogen cycling. Cycling of glucose-1-P in and out of glycogen would produce the following reactions GIP+NTP--+NDPG+PP; NDPG + (glycogen)n--+ NDP + (glycogen)n + 1 (glycogen)n + 1 + Pi--+G I P +(glycogen)n Sum: NTP+P;--+NDP+PP;

The utilization of PP; generated by glycogen cycling may not properly be called conservation since the PP; arises at the expense of a nucleoside triphosphate bond. Perspectives

Peller's premise [20) that in vitro Polymer (RNA) synthesis should be coupled to PP; hydrolysis is valid, but all that IS required in vivo is that the PP; concentration be reduced to an acceptable level. This ~ay be accomplished by energy-conservIng reactions or by the relatively wasteful hydrolysis of PP;. Energy-poor anaerobes tend to fellow the former route, aerobes the latter. However, even in the aerobic organisms there are some reactions with c.onsiderable flux (e.g. the reverse of reaction a) which conserve the energy of PP;. How high the acceptable intracellular level of PP; may be is demonstrated, in an extreme instance, by the parasitic ameba which contains pp; concentrations calculated to be 2 x I0- 4 M [4]. This organism is eucaryotic, but it has a minimu.m of evident internal compartmentali~tJon. Since the anabolic processes occur rn the face of high PP; concentrations it may be presumed that low intracellular PP; is a luxury which only some organisms can afford. An alternative view is that the ~reater amount of information to be coded 1 ~to the nucleic acids and proteins of htgher organisms requires the extra thermodynamic leverage afforded by PP; con~ntrations in the JiM range. This would 1 ~Ply that the parasitic ameba has a primitive metabolism, a view which receives support from other directions. Mechanisms exist and function in a wide variety of organisms which are capable of

Bioi. Chern. 247, 3382-3392 4 Reeves, R. E., South, D. J., Blytt, H. J. and Warren, L. G. (1974) J. Biol. Chern. 249, 7737-7741 5 Siu, P. M. L. and Wood, H. G. (1962) J. Biol. Chem. 237, 3044-3051 6 Nordlie, R. C. (1974) Curr. Top. Cell. Regul. 8, 33-117 7 Nordlie, R. C., Arion, W. J. and Glende, Jr, E. A. (1965) J. Bioi. Chem. 240, 3479-3484 8 Hatch, M. D. and Slack, C. R. (1968) Biochem. J. 106, 141-146 9 Evans, H. J. and Wood, H. G. (1968) Proc. Nat. Acad. Sci. U.S.A. 61, 144&-1453 10 Re'eves, R. E. (1968) J. Biol. Chem. 243, 3202-3204 II Cagen, L. M. and Friedmann, H. C. (1968) Biochem. Biophys. Res. Commun. 33, 52&-533 12 O'Brien, W. E. and Bowien, S. (1975) Fed. Proc. 34,641 13 O'Brien, W. E., Bowien, S. and Wood, H. G. (in the press) 14 Keister, D. L. and Minton, N. J. (1971) Biochern. Biophys. Res. Commun. 42, 932-939 15 Buchanan, B. B. (1974) J. Bact. 119, 1066-1068 16 Lawson, J. W. R., Guynn, R. W., Cornell, N. and Veech, R. L. in Gluconeogenesis (Mehlman, M.A. and Hanson, R., eds), Wiley, New York (in the press); this chapter was reviewed from a prepublication copy of manuscript kindly sup. • plied by the authors 17 Gelderman, A. H., Keister, D. B., Bartgis, I. L. and Diamond, L. S. (1971) J. Protozoal. 57, 906-911 18 Reeves, R. E. and Guthrie, J. D. (1975) Biochern. Biophys. Res. Commun. (in the press) 19 Reeves, R. E. (1970) Biochim. Biophys. Acta 220, 346-349 20 Peller, L. (1975) Biochem. Biophys. Res. Commun. 63,912-916 21 Reeves, R. E. and Malin, L. K. (1969) Anal. Bio· chem. 28, 282-287

References I Flodgaard, H. and Fleron, P. (1974) J. Bioi. Chern. 249, 3465-3474 2 Wood, H. G., Davis, J. J. and Lochmuller, H. (1966) J. Bioi. Chern. 243, 5692-5704 3 Cagen, L. M. and Friedmann, H. C. (1972) J.

Cytoplasmic actin and myosin Edward D. Korn I

Actin and myosin provide the structural and enzymatic elements for essential motile systems in plant and animal cells that preceded in evolution i/,e adaptive use of actin and myosin for muscle contraction.

Three biochemically and morphologically distinct cellular motile mechanisms are known. Bacterial flagella are rigid helical protein structures that project through the cell wall and are rotated by a 'motor' located in the cytoplasmic membrane. Flagella of eukaryotic cells are completely different structures typically consisting of nine outer microtubule doublets surroundE.D.K. is arrhe Laboratory of Cell Biology, National Heart and Lung Institute, Bethesda, Md200!4, U.S.A. •

ing a pair of central microtubule doublets all surrounded by the plasma membrane of the cell. Dynein, a high molecular weight A TPase, is a component of arms that interconnect outer doublets. The conversion of chemical to mechanical energy (the motile process) occurs in the flagella. Cytoplasmic actin and myosin, the subjects of this article, form a third motile system responsible for such processes as ameboid movement, cytoplasmic stream-

TIBS- March 1976

56 ing, cell division, movement of microvilli and phagocytosis [1]. Since the proteins responsible for cytoplasmic motility are very similar to, and the evolutionary predecessors of, their muscle counterparts it will be easier to discuss them after reviewing briefly the better understood biochemistry and ultrastructure of skeletal muscle. Muscle sliding f:tlaments

The contractile unit of skeletal muscle is the sarcomere, the space between two Z-lines. Extending from the Z-lines are parallel arrays of thin filaments each consisting of two intertwined helical polymers of about 300-400 actin monomers (42000 dal'tons). A second protein, tropomyosin (68000 daltons), is associated with the actin filament (one tropomyosin molecule for every seven actin molecules) and one molecule of a third protein, troponin (80000 daltons), is associated with each molecule of tropomyosin. Finally, aactinin (95000 daltons) is a component of the Z-line perhaps involved in the attachment of the actin filament. Between the actin filaments (but not attached to the Z-line) lie thick filaments each consisting of several hundred thousand myosin molecules. Myosin is an A TPase of about 4 70 000 daltons containing, in rabbit skeletal muscle myosin, two heavy chains of 200000 daltons and four light chains: one 16000, two 18000 and one 20000 daltons. The two heavy chains are almost completely helical and wind around each other forming a super-coiled rod from which two globular heads project (one per heavy chain). Each head contains an A TPase site and an actin-binding site. The rod segments of myosin molecules aggregate to form the thick filament, a bipolar structure with a bare central region free of heads and extended regions at both ends from which heads project with opposite polarity. Pure muscle myosin is a K +-A TPase with about 10 O(o activity in the presence of Ca2+ and even less in the presence of Mg2 +. (In the enzyme assays, a controlled proteolytic degradation product of myosin, heavy meromyosin - the double stranded head region of myosin - is frequently used because it is more soluble.) Myosin forms a complex with actin, actomyosin, which is a Mg 2 +-A TPase of about equal specific activity to the K +-A TPase of myosin. The addition of tropomyosin and troponin converts the M g2+ -A TPase of muscle actomyosin into a Ca2 ·~, Mg2 +ATPase. Muscle contraction begins with the release ofCa2+ from th~ 'iarcoplasmic reticulum which ?.ll,Jw~ :1-;e ATP-myosin heads projer.ting fmm the ~hi~k fil:tmet'tf.

Fig. 1. Left: brush border isolated from chicken intestinal epithelium showing a core of actin filaments in each microvillus ( x 39000). Right: a demembranated microvillus, decorated with muscle myosin-S1 , illustrating the downward polarity oft he filaments ( x 116000). Micrographs by Drs Mooseker and Tilney.

to bind to the neighboring thin actin filaments. ATP is hydrolyzed. Chemical energy is transduced into movement of the attached myosin headgroup through an arc of 10 nm. Another molecule of A TP complexes to the myosin which dissociates from the actin and then undergoes a conformational change allowing recombination of the myosin-ATP and actin. Successive cycles pull the actin filaments associated with the bipolar myosin filaments towards each other shortening the sarcomere distance.

away from the microvillus tip) identical to that of muscle thin filaments with respect to the Z-line [6]. There is no well-documented example of cytoplasmic thick filaments in normal, non-muscle cells. Myosin is present in cells at much lower concentration than actin however, and the inability to detect thick filaments may be a statistical consequence of their low frequency [7]. On the other hand, the general observation that antimyosin antibodies react where thin filaments are seen [8), suggests that cytoplasmic myosin may be associated with actin Microfilaments filaments in a manner distinctly different Eukaryotic cells contain cytoplasmic from skeletal muscle. Actin has been purified from about ten filaments 5-7 nm wide and of indeterminate length underlying the plasma non-muscle cells and characterized by its membrane, in broad pseudopodal areas molecular weight and its ability to polyand within cell projections such as micro- merize into filaments that activate the spikes and microvilli. In about fifty animal Mg2 +-A TPase of muscle heavy meromyoand plant cells [2,3], these microfilaments sin [2,3]. (Actin usually accounts for about have been identified as actin filaments by l-4%,and frequently as much as 10-15%, their ability to react with heavy meromyo- of the total cell protein. Much of that actin sin under specific conditions to form is probably not filamentous at any given 'decorated' filaments indistinguishable time.) In five cases the cytoplasmic actins from those formed by the interaction of have been shown also to resemble muscle myosin with muscle actin. The location actin in their amino-acid compositions, inof these filaments in situ and their co-isola- cluding the presence of the rare amino acid tion with highly purified plasma mem- 3-methylhistidine, and the viscosity and branes [4,5] suggest their association with ultrastructure of the filamentous actins. the cytor,:a~ulic surface of the plasma Extensive homology, but not identity, of membrane. ln favorable situations, intesti-' the primary amino-acid sequences of nal microvilli, for example (Fig. 1), fila- several cytoplasmic actins and muscle ments der::orated with heavy meromyosin actin has been shown by peptide maps of show pohu:ty (3 11 arrowheads pointing proteolytic digests or by the composition

57

TIBS- March 1976 of the peptides produced by .cyanogen bromide cleavage at methionine bonds [2,3]. However, although actin has clearly changed very little during the long period of biological evolution, there are significant differences between different actins. Acanthamoeba actin, for example, differs qualitatively and quantitatively from muscle actin in its interactions with muscle heavy meromyosin (higher Kd;,), tropomyosin and troponins [9].

polymerizes upon stimulation to form a long cellular projection, the acrosomal process [14]. Actin also occurs in erythrocyte ghosts [15] which probably do not contain myosin. Other cytoplasmic motility proteins

Tropomyosin has been isolated from human platelets and rat brain, and Ca 2 + sensitivity has been observed for the actomyosin-A TPase of platelets, brain and Physarum suggesting that troponin-like Isolation of cytoplasmic myosins proteins are also present in these cells [2,3]. Myosin has been isolated from about Immunofluorescent techniques show that ten non-muscle cells [2,3]. All but one of the stress fibers of cultured fibr-oblastic the isolated cytoplasmic myosins have cells react not only with antibody prepared masses of about 450000 daltons and con- against smooth muscle actin but also with sist of heavy chains of about 200000 dal- anti-muscle tropomyosin and anti-muscle tons and a variable number of light chains a-actinin suggesting that proteins similar of different molecular weights. The one to the muscle proteins are associated with known exception is myosin isolated from the actin filaments [16]. Anti-a-actinin also Acanthamoeba castellanii [10]. It has a seems to react with intestinal microvilli mass of about 180000 daltons with a heavy membranes specifically at the regions of chain of ISO 000 daltons and at least two apparent filament-membrane junctions light chains. Except for Acanthamoeba [6]. However, tropomyosin and troponins myosin, which may lack a rod segment, are probably not present in all cells that the isolated cytoplasmic myosins are cap- contain actin and myosin since in many able of forming bipolar filaments resem- cases neither Ca 2 +-sensitivity nor direct bling those derived from muscle myosin evidence for the proteins have been found. but consistently shorter. Recently, large proteins of All of the cytoplasmic myosins have 200000-300000 daltons have been found higher Ca2+ -ATPase activity than Mg2+- associated with partially purified actins ATPase, as does muscle myosin, but, para- from several sources (acrosome [I 7], macdoxically, only the small Acanthamoeba rophage [18], fibroblast [19] and Acanthamyosin has been shown to have the high moeba [20]) where they are speculatively ~+-A TPase characteristic of muscle myo- assigned different roles: maintaining actin sm. The Mg2+ -A TPase activities of all of in a non-filamentous form, binding actin the cytoplasmic myosins are activated by filaments together, binding actin filaments actin but the extent of activation is usually to plasma membranes. In the erythrocyte, only two to five-fold not the fifty-fold or actin may be associated with the high greater activation found for skeletal molecular weight membrane protein specmuscle myosin. Again, the exception is trin [14]. Lower molecular weight proteins Acanthamoeba myosin for which actin may also be associated with cytoplasmic activation equals that for muscle myosin. actins. In the Limulus acrosomal process flowever, in this case it is necessary to proteins of 55 000 and 95 000 daltons are ~dd a 'co-factor' protein that has been par- the only components other than actin [17], tially purified from Acanthamoeba extracts in Dictyostelium a second protein co-iso111]. A similar requirement for 'co-factor' lates with actin [21] and gelled extracts of has recently been found for macrophage Acanthamoeba contain proteins of 70000 myosin [12] even though it, unlike Acan- and 50000 daltons [20]. thamoeba myosin, has a mass similar to that of muscle myosin. Thus, it is possible Understanding cell motility Actin and myosin are established as that undiscovered co-factors may be required for actin activation of the Mg2 +- significant, often the major, proteins in the ATPase activities of other cytoplasmic cytoplasm of non-muscle cells. In some myosins. A similar but different phenome- of these cells a-actinin, tropomyosin and non occurs with platelet myosin when troponin may also be present but these e.nzymatic phosphorylation of one of the 'regulatory' proteins probably do not ~ght. chains enhances actin activation [13]. always co-exist with cytoplasmic actin and bv1ously, although they share essential myosin. Actin can also occur without myof~atures, myosins from different sources sin. Other proteins, probably not present differ more from each other than do actins. in muscle cells although a thorough search In at least one well-documented case should be initiated, appear to be associated actin occurs in the absence of myosin. The with many of the cytoplasmic systems. ac.rosomal vesicle of Limulus sperm con- These auxiliary proteins include the cotams non-filamentous actin that rapidly factor proteins required for actin activa-

tion of the Mg 2 +-A TPase of Acanthamoeba and macrophage myosin and high and low molecular weight proteins associated with filamentous and non-filamentous actins. It is certain that cytoplasmic actin and myosin provide the basis for ameboid motility, endocytosis, exocytosis and cytokinesis. (Indeed, there is some evidence [22] for nuclear actin and myosin which may, in association with the mitotic spindle, provide the motile force for chromosome segregation.) The structural mechanisms by which these processes are accomplished are not known but it is easy to conceive models based on the sliding filament mechanism of muscle contraction in which the plasma membrane, cytoplasmic microtubules or other structural elements serve the role of the Z-line. However, the relatively low concentration of cytoplasmic myosin and the possible absence of thick filaments must be considered. Regulation of these cellular motile proce'sses may involve Ca2+ when proteins similar to tropomyosin and troponin are present but reversible polymerization of actin filaments and association of actin filaments into bundles and with membranes, through the intervention of other proteins, may provide other control mechanisms. Rapid redistribution of actin filaments occurs in such normal processes as phagocytosis [23] and cell division [24] and also when cells are virally t~;ansformed [25]. The understanding of cell motility is limited by the paucity of quantitative biochemical data. Of the many cytoplasmic motility proteins identified by direct and indirect means, only a very few have been purified and none has yet been studied extensively, alone or in recombination with other purified proteins. Only when 1this has been done will it be possible to reconstruct the biochemical and ultrastructural events of the essential motile processes that depend on cytoplasmic actin and myosin. 1

References l CihaFoundaliollSymposium(l973) Vol. l4, Elsevier, New York 2 Pollard, T. D. and Weihing, R. R. (1974) CRC Cril. Rei'. Biochem. 2, l-65 3 Weihing, R. R. (l976) in Cell Biology (Altman, P. L. and Dittner, D. S., eds) (in the press) 4 Korn, E. D. and Wright, P. L. (1973) J. Bioi. Cltem. 248, 439-447 5 Pollard, T. D. and Korn, E. D. (1973) J. Bioi. Chern. 248, 448-450 6 Mooseker, M.S. and Tilney, L. G. (1975) J. Cell Bioi. (in the press) 7 Niederman, R. a'nd Pollard, T. D. (1975) J. Cell Bioi. 67, 72-92 8 Painter, R. G., Sheetz, M. and Singer, S. J. ( 1975) Proc. Nal. Acad. Sci. USA 72, 1359-1363

TIBS- March 1976

58 9 Gordon, D., Yang, Y-Z., Eisenberg, E. and Korn, E. D. (1976) Cold Spring Harbor Conference on Cell Proliferation, Vol. 3 (in the press) 10 Pollard, T. D. and Korn, E. D. (1973) J. Bioi. Chern. 248, 4682-4690 II Pollard, T. D. and Korn, E. D. (1973) J. Bioi. Chern. 248,4691-4697 12 Hartwig, J. H. and Stossel, T. P. (1975) J. Bioi. Chem. 250, 5696-5705 13 Adelstein, R. S. and Conti, M.A. (1975) Nature 256, 597-598 14 Tilney, L. G., Hatano, S., Ishikawa, H. and Mooseker, M.S. (1973) J. Cell Bioi. 59, 109-126 15 Tilney, L. G. and Detmers, P. (1975) J. Cell Bioi. 66, 508-520 16 Lazarides, E. (1975) J. Cell Bioi. 65, 549-561 17 Tilney, L. G. (1975) J. Cell Bioi. 64, 289-310

18 Hartwig, J. H. and Stossel, T. P. (1975) J. Bioi. Chern. 250, 5696-5705 19 Weihing, R. R. (1976) Cold Spring Harbor Conference on Cell Proliferation, Vol. 3 (in the press) 20 Pollard, T. D. (1976) J. Cell Bioi. (in the press) 21 Spudich, J. A. and Cooke, R. (1975) J. Bioi. Chern. 250, 7485-7491 22 Sanger, J. W. (1975) Proc. Nat. Acad. Sci. USA 72,2451-2455 23 Korn, E. D., Bowers, B., Batzri, S., Simmons, S. R. and Victoria, E. J. (1974) J. Supramolecular Struct. 2, 517-528 24 Sanger, J. W. (1975) Proc. Nat. Acad. Sci. USA 72, 1913-1916 25 Pollack, R., Osborn, M. and Weber, K. (1975) Proc. Nat. Acad. Sci. USA 72, 994-998.

Lipid intermediates in protein glycosylation A. J. Parodi and L. F. Leloir

and its amount could be greatly increased by the addition of a lipid extract of liver. The acceptor contained in the extract was purified and its properties found to be identical to a sample of dolichol phosphorylated by a chemical method. It differed from undecaprenol phosphate in that it was acid stable, a difference that can be accounted for by the absence of the double bond in the first isoprene. The product formed by transfer of glucose from UOP-Glc to Ool-P, namely Ool-PGlc, was found to be acid labil~ and to give I ,6-anhydroglucose by alkaline treatment. This was taken as an indication that the glycosidic bond in Ool-P-Glc is ~ because the same anhydride is fonned by alkaline treatment of phenyl ~-glucosides. A transfer reaction was also detected with GOP-Man as donor (reaction 1.2) (9]. The product, Ool-P-Man, has been studied more thoroughly and compared with a synthetic specimen (10-I 1]. It is also the ~anomer.

The oligosaccharide moieties ofanimal glycoproteins that involve linkage between carbohydrate and asparagine are built up on dolichol pyrophosphate before incorporation into the protein. The use of dolichol pyrophosphate in place of the commoner nucleoside diphosphate as the sugar carrier may be because the former is more compatible with the hydrophobic environment of the membrane where the glycoprotein synthesis occurs. The glycosylation of proteins occurs in some cases by direct transfer of monosaccharides from sugar nucleotides to polypeptide chains, while in other cases it occurs by a mechanism involving lipidbound mono- and oligosaccharides as intermediates. Work on the latter mechanism is reported in this review (see also refs [l-3]). The pathway involving lipids seems to function in the biosynthesis of those glycoproteins which conrain oligosaccharides linked to asparagine residues. The glycoproteins in which the saccharide moieties are joined to hydroxylysine or to serine or threonine are usually believed to be built up by direct and sequential transfer from the sugar nucleotides but further work may uncover new facts(for reviews see refs [4,5]). It seems important to understand the mechanism of biosynthesis of glycoproteins because they are widely distributed in nature, from bacteria to animal tissues. They can be found among cytoplasmic, secreted or membrane-bound proteins and may function as enzymes, hormones, antibodies, etc. (reviewed in refs [5-7]). They are believed to have a key role in determining the specificity of cell surface recognition phenomena. A.J.P. and L.F.L. are at the lnstituto de Investigaciones Bioquimicas 'Fundaci6n Carnpornar' and Facultad de Ciencias Exactas y Naturales, Obligado 2490, Buenos Aires (28), Argentina

Polyprenols The role of lipid intermediates in sugar transfer was first detected in studies on the formation of some bacterial polysaccharides. The lipids involved are derivatives of polyprenols that have the general structure CH

H-

[

3

CH,-~=CH -CH,

]

n

OH.

The compound which acts as an intermediate in bacteria has n = 1I and two trans double bonds. It was isolated as the monoor diphosphate combined to sugars [3]. The lipid intermediates of animal tissues are derivatives of dolichol which is a very long polyprenol (n = 17-21). It has two trans double bonds and the isoprene which carries the -OH group is saturated. The compound was found partly free and partly esterified to fatty acids [3]. At present it is known to occur also as the monophosphate and the diphosphate, free or combined with sugars. DoJichol phosphate sugars Several reactions in which dolichol derivatives are involved have been described (see Table 1). Reaction l.l. which leads to the formation of Ool-P-Glc was first detected on incubation of glucose labeled UOP-Glc with liver microsomes and magnesium ions [8]. A radioactive compound soluble in organic solvents was formed,

The transfer of xylose to Ool-P (reaction 1.3) has also been detected [12]. Since xylose is identical to glucose except for the absence of the -CH 20H residue the reaction might be attributed to loose specificity of the glucose transfer reaction. Transfer of xylose occurs in the synthesis of the linkage region ofproteoglycans (serine-xylose), but the corresponding enzyme has been purified and no lipid requirement was detected [13]. DoJichol diphosphate sugars When UOP-GlcNAc is used as donor and Ool-P as acceptor a transfer of sugar phosphate takes place so that the product is Ool-P-P-GlcNAc (reaction 2.1) [14]. A second addition of acetylglucosamine can occur so that Ool-P-P-(GlcNAc) 2 is formed (reaction 4.1). There is evidence that the donor is UOP-GlcNAc and not Ool-P-P-GlcNAc [14]. The next reaction appears to be a transfer from GOP-Man so that a trisaccharide residue is formed (reaction 4.2) (15). This trisaccharide (~­ Manl-4~-GlcNAc-4-GicNAc) is of special interest because it occurs linked to asparagine residues in many glycoproteins such as immunoglobulins, ovalbumin, thyroglobulin, etc. However, there are exceptions since, for instance, immunoglobulin E contains an oligosaccharide with the sequence a-GlcNAcl-3~-Manl-4~­ GlcN Ac-~-asparagine [I 6]. The transfer of mannose residues from Ool-P-Man to the Ool-P-P linked trisaccharide leads to the fonnation of larger oligosaccharides (reaction 5.1) [15]. Such compounds can be detected by incubation of labeled GOP-Man or Ool-P-Man with liver or hen oviduct midosomes [17-18]. The reaction products, soluble in organic