THE JOURNALOF BIOLOGICALCHEMISTRY Vol. 264, No. Ieaue of January 5,, pp. 625-629,1989 0 1989 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.SA.

Acylation of Keratinocyte byPalmitic and Myristic Acids in the Membrane Anchorage Region*

(Received for publication, June 27, 1988)

Rupa Chakravarty and Robert H.Rice From the Charles A. Dana Laboratory of Toxicology, Harvard School of Public Health, Boston, Massachusetts 02115

The membrane-boundform of keratinocyte transglu- linking. To clarify the mechanism of envelope formation, taminase was found to be labeled by addition of [‘HI knowledge of critical structural featuresof the inter- acetic, [‘Hlmyristic, or [SH]palmiticacids to the culture actions with substrates and the membrane will be essential. medium of human epidermal cells. Acid methanolysis As a first step in this direction, the present work addresses and high performance liquid chromatography analysis the biochemical basis of the enzyme anchorage in the mem- of palmitate-labeled transglutaminase yielded only brane. methyl palmitate. In contrast, analysis of the myris- Generally, areanchored in membranes by stretches tate-labeled yielded approximately 40% of hydrophobic amino acids, by inositol phospholipid glycan methyl myristate and 60%methyl palmitate. Incorpo- moieties, or by fatty acids acylated directly to amino acid Downloaded from ration of neither label was significantly affected by residues (Sefton and Buss, 1987).In thelast category, proteins cycloheximide inhibition of protein synthesis. The im- portance of the fatty acid moiety for membrane an- modified by myristate may be found in cytosolic as well as chorage was demonstrated in three ways. First, the membrane compartments (Olson et al., 1985) or even encap- enzyme was solubilized from the particulate fraction sidated in nonenveloped viruses (Streuli and Griffin, 1987; of cell extractsby treatment with neutral 1 M hydrox- Chow et al., 1987; Paul et al., 1987). In contrast, palmitate ylamine, which was sufficient to release the fatty acid labeling occurs on proteins which are found almost exclusively www.jbc.org label. Second, solubilization of active enzyme from the in membranes, but exceptions include the secreted apolipo- particulate fraction upon mild trypsin treatment re- protein A-I (Hoeg et al., 1986), which binds lipid, and sulted in a reduction in size by approximately 10 kDa from Dictyostelium discoideum (Stadler et al., 1985). Keratin-

and removal of the fatty acid radiolabels. Third, the ocyte transglutaminase is found not only in the particulate at unknown institution, on January 18, 2010 small fraction of soluble transglutaminase in cell ex- fraction of cultured human epidermal cell extracts, but also tracts was found almost completely to lack fatty acid to a small extent in thesoluble fraction, and can be released labeling. Keratinocyte transglutaminase translated from membranes by mild trypsin treatment (Thacher and from poly(A+) RNA in a reticulocyte cell-free system Rice, 1985). The present work reveals that the membrane- was indistinguishable in size from the native enzyme, bound form incorporates bothpalmitate and myristate, suggesting anchorage requires only minor post-trans- whereas the native and trypsin-released soluble forms show lational processing. Thus, the data are highly compat- little if any fatty acid labeling. ible with membrane anchorage by means of fatty acid acylation within 10 kDa of the NH2 orCOOH terminus. MATERIALSAND METHODS Cell Culture-Keratinocytes from normal human epidermis were cultured in 10-cm dishes with support from feeder layers of lethally irradiated 3T3 cells according to standard methods (Allen-Hoffman Keratinocyte transglutaminase participates in the remark- and Rheinwald, 1984) in a 3:l mixture of Dulbecco-Vogt Eagle’sand able maturation process of cells in the epidermis and related Ham’s F-12media supplemented with fetal bovine serum (5%), hy- epithelia. Activated by flux into the cytoplasm, this drocortisone (0.4 pg/ml), epidermal growth factor (10ng/ml), adenine (0.18mM), insulin (5 pg/ml), transferrin (5 pg/ml), triiodothyronine enzyme cross-links substrates by isopeptide bonding into en- (20 p~),and antibiotics. The cells were inoculated in the presence of velopes that contribute to the cohesiveness of the terminally cholera toxin (10 ng/ml) and held at confluence for 3-5 days prior to differentiated cells (Green, 1979). Primarily membrane-bound harvesting for isolation of poly(A+)RNA or up to 10 days for protein (Lichti et al., 1985; Simon and Green, 1985; Thacher and isolation. Rice, 1985), the enzyme is known to cross-link the soluble Cell-free Translation-Typically, 10 cultures were dissolved with protein involucrin and several membrane proteins into par- homogenization in buffered 6 M guanidine thiocyanate, and theRNA was pelleted through CsCl by standard procedures (Chirgwin et al., ticulate structures (Rice and Green, 1979; Simon and Green, 1979). Poly(A+) RNA was then isolated on oligo(dT)-cellulose (type 1984, 1985). The processes bywhich thesestructures are 111, Collaborative Research) essentially as described (Aviv and Leder, assembled and become esterified with w-hydroxyacylsphin- 19721, precipitated with ethanol, and used to program rabbit reticu- gosines (Swartzendruber et al., 1987) remain to be elucidated. locyte extracts (Promega Biotec). Translations were performed in Moreover, participating substrates (Michel et al., 1987; Nagae final volumes of 90 ~1 containingup to 9 pgof poly(A+) RNA et al., 1987) and the resulting ultrastructure Warhol et al., (sufficient for maximal incorporation), 60 rl of reticulocyte extract pretreated with micrococcal nuclease, and 6 units of RNasin (Pro- 1985) reportedly vary with the physiological state or deriva- mega Biotec). The relatively low level of radioactivity incorporated tion of the cell and the method or speed of inducing cross- into transglutaminase was not increased by addition of a mixture of required amino acids lacking methionine or by use of [36S]cysteineor * This work was supported by Grant AR 27130 from the National [3H]leucine. Institutes of Health. The costs of publication of this article were Fatty Acid Labeling and Immunoprecipitation-Cultures were in- defrayed in part by the payment of page charges. This article must cubated for 4.5 h with 1 mCi of [3H]palmitic, [3H]myristic, or [3H] therefore be hereby marked “aduertisement” in accordance with 18 acetic acid in medium supplemented with 10% delipidized (Rothblat U.S.C. Section 1734 solely to indicate this fact. et al., 1976) fetal bovine serum and 5 mM sodium pyruvate (Olson et 625 626 TransglutaminaseFatty Acid Acylation al., 1985). (To maximize incorporation of acetate label into fatty ab def acids, cultures were preincubated overnight in this medium prior to OlnmCn, addition of [3H]acetic acid.) Each culture was then rinsed several times with isotonic neutral saline, scraped from the dish, storedfrozen overnight, and homogenized in 6 ml of 50 mM Tris-C1 (pH 8.0), 1 mM EDTA. The particulate material was isolated by high-speed centrif- ugation (100,000 X g for 1 h), resuspended in 6 ml of 20 mM Tris-C1 (pH 8.0), 1 mM EDTA, 0.3% Emulgen 911 nonionic detergent and stirred at 4 "C for 2 h. (Alternately, the transglutaminase was solu- bilized from the isolated particulate material by suspension for 5 min FIG. 1. Coomassie Blue-stained transglutaminase immuno- in 2 ml of 20 mM Tris-C1 (pH 8.0) containing 50 pgof L-l-tosylamido- precipitates. Samples were solubilized by Emulgen 911 nonionic 2-phenylethyl chloromethyl ketone-trypsin (Worthington), followed detergent (a, c) or mild trypsin (b) treatment in one experiment and by addition of 150 pg of soybean trypsin inhibitor.) The solubilized by Emulgen (d, e) or 1 M hydroxylamine (f) in a second experiment. material was clarified by high-speed centrifugation and to thesuper- In each case, negative control immunoprecipitates without added natant were added first 2 pg of B.C1 monoclonal (Thacher B.C1 mouse monoclonal (c, d) permitted identification of and Rice, 1985) and NaCl to 0.2 M, and then (after1 h at 4 "C), 4 pg the transglutaminase bands (indicated by arrowheads) in a, b, e, and of rabbit anti-mouse IgG (Cappel) and 2 mg of protein A-Sepharose. f. In each lane, the heavily stained band of greater mobility represents After 45 min at room temperature, the immune complexes were the heavy chains of added B.C1 and rabbit IgGs. recovered by centrifugation and rinsed three times in immunoprecip- itation buffer. In most experiments the transglutaminase was re- a bcde solubilized in SDS,' electrophoresed in 10% polyacrylamide gels ., , ~ (Laemmli, 1970), stained with Coomassie Blue, treated with EN3HANCE (New England Nuclear), and submitted to autoradiog- raphy for 1-2 weeks. For quantitation of relative amounts of protein Downloaded from and label, the stained gel and autoradiogram were scanned by laser densitometry. Analysis of Incorporated Fatty Acids by High Performance Liquid Chromatography-Immunoprecipitates of labeled protein were ex- tracted several times with chloroform-methanol untilno further radioactivity was obtained in the washes. The dry residue was then FIG. 2. Autoradiography of fatty acid-labeled particulate transglutaminase. Cells were labeled with [3H]acetic acid (a), [3H] incubated in 1 ml of 2 N HC1,83% methanol at 95 "C for up to 60 h www.jbc.org in vocuo and extracted with petroleum ether. The organic extract was myristic acid (b, c), or [3H]palmitic acid (d, e). Prior to immunopre- dried under nitrogen gas, dissolved in 80% acetonitrile,and submitted cipitation, transglutaminase was solubilized with Emulgen nonionic to reverse-phase chromatography in this solvent (Olson et al., 1985) detergent except for c, and d, which were released by mild trypsin using Cl8 columns from Waters. treatment. Transglutaminuse Assay-Enzymatic activity was measured by at unknown institution, on January 18, 2010 incorporation of [3H]putrescine (15 pM, 125 Ci/mol) into 2 mg/ml abcdefg h reductively methylated (Means and Feeney, 1968) a-casein (Wor- thington) in 100 mM Tris-C1 (pH 8.3), 4 mM CaC12, 0.4 mM EDTA, 5 mM dithioerythritol. Samples (typically 10-20 pl containing approx- imately 10 pg of protein) were incubated in final volumes of 0.26 ml at 35 "C for 30 min, in the linear range of the assay (Thacher and Rice, 1985).

RESULTS Under the extraction conditions used in the present work, FIG. 3. Properties of fatty acid labeling. [3H]Palmitate- (a, b) an improvement over the more rapid method previously em- or [3H]myristate- (c, d) labeled transglutaminase treated with 0 (a, ployed (Thacher and Rice, 1985), approximately 90% of the d) or 1 M (b, c) hydroxylamine in the gel for 16h at room temperature transglutaminase activity in keratinocyte particulate material prior to autoradiography is shown. Transglutaminase from cells la- typically was solubilizedwith the nonionic detergent Emulgen beled with [3H]palmitate (e, f) or [3H]myristate (g, h) in thepresence 911. Initial experiments showed that the enzyme was solubi- (e, h) or absence (f, g) of 10 pg/ml cycloheximide in the culture lized nearly as well by treatment of particulate material with medium is also shown. 1M hydroxylamine at neutral pHfor 2 h at room temperature. In these experiments, efficacy was judged not only by trans- was lost (typically 80%) upon treatment of the palmitate- glutaminase assays after removal of hydroxylamine (showing (compare lanes a and b) or myristate-labeled (lanes c and d) the enzyme retains activity), but also by the amount of Coo- transglutaminasewith hydroxylamine after immunopre- massie Blue-stained enzyme visible upon electrophoresis of cipitationand gel electrophoresis. In addition, when the immunoprecipitates (Fig. 1, e and f). As illustrated, the ap- myristate-labeled cell particulates were treated with 1 M parent molecular weights of the enzyme solubilized by hy- hydroxylamine for 2 h and the released enzyme was then droxylamine or detergent were indistinguishable. immunoprecipitated and examined by SDS gel electrophore- The efficacy of hydroxylamine solubilization suggested that sis, virtually no radioactivity was detected upon alltoradiog- the enzyme was anchored by esterified or thioesterified fatty raphy (Fig. 4d). acids. To confirm this interpretation, keratinocyte cultures Keratinocyte transglutaminase canalso be solubilized from were incubated in the presence of [3H]acetic acid in the the particulate fraction of cell extracts by mild trypsin diges- medium for 4.5 h (Towler and Glaser, 1986), after which the tion (Thacher and Rice, 1985). This treatment results in a transglutaminase was immunoprecipitated and examined by reduction in size from approximately 92 to 89 kDa (Fig. 1, autoradiography. A low degree of labeling was observed, as lanes a and b). When palmitate- or myristate-labeled enzyme shown in Fig. 2a. Labeling under similar conditions with [3H] was released from the membrane by trypsin and immunopre- myristate or [3H]palmitate was considerably more effective cipitated, no radioactivity was evident in theautoradiograms (lanes b and e), with the latter giving the most incorporation (Fig. 2, c and d). In such experiments, similar amounts of of radioactivity. As illustrated in Fig. 3, most of the radiolabel Coomassie Blue-stained protein (quantitated by laser densi- tometry) were electrophoresed as in the parallel samples ' The abbreviation used is: SDS, sodium dodecyl sulfate. which were solubilized with nonionic detergent and which TransglutaminaseFatty Acid Acylation 627

ab cde ab W

FIG.4. Fatty acid acid labelingof soluble transglutaminase. Similar amounts of particulate (a, d, e) and soluble (b, c) forms of transglutaminase from cells labeled with [3H]palmitate (a, b) or [3H] FIG.6. Autoradiography of cell-free translation products. myristate (c, d, e)were applied to thegels. In d, the myristate-labeled Immunoprecipitates of material synthesized in reticulocyte extracts enzyme was solubilized from the particulate fraction by treatment programmed with keratinocyte poly(A+) RNA were prepared with (a) with 1 M hydroxylamine and immunoprecipitated. or without (b)B.C1 antitransglutaminase monoclonal antibody. The arrowhead shows the migration of native transglutaminase electro- 600 phoresed in parallel and detected by Coomassie Blue staining. a P A small fraction (typically 5%) of the keratinocyte trans- glutaminase in crude extracts is detected as a soluble protein

of the same apparent molecular weight as theparticulate form Downloaded from in SDS gels (Thacher and Rice, 1985). In thepresent experi- ments, the relative palmitic and myristic acid labeling of the two forms were compared. As shown in Fig. 4, the particulate form contained far greater (at least 30-fold more) labeling than the soluble form when the same amount of Coomassie

Blue-stained transglutaminase was electrophoresed in adja- www.jbc.org cent lanes. The little radioactivity evident in the autoradi- ogram of the soluble enzyme in lane b may even exceed the actual amount, since a slight contamination with the partic-

ulate form is difficult to avoid. at unknown institution, on January 18, 2010 Poly(A+) RNA isolated from cultured keratinocytes was translated in a rabbitreticulocyte cell-free system. As shown in Fig. 6a, polyacrylamide gel electrophoresis of the transla- tion products with subsequent autoradiography led to the Fraction (1 ml) detection of a band of material specifically immunoprecipi- FIG.5. Analysis of methyl esterified fatty acids released tated by antitransglutaminase monoclonal antibody. This from radiolabeled transglutaminase.The cultures were labeled band matched in mobility the native transglutaminase elec- with palmitic acid (a)or myristic acid in the absence (b)or presence trophoresed in theadjoining lane and detectedby Coomassie (c) of cycloheximide (10 pg/ml). The elution positions of methyl Blue staining (indicated by the arrow). Thus, major post- myristate (M)and methyl palmitate (P)were determined using [3H] translational modification of the translatedform to yield the myristic and [3H]palmitic acids esterified and chromatographed in native protein, such as the proteolytic activation of the clot- parallel. The elution of methyl palmitate in a differs from that in b, and e, since the former and latterchromatograms were obtained from ting factor XI11 catalytic subunit (Schwartz et al., 1973), was different columns. not detectable in thisfashion. Several other translation prod- ucts, presumed to be primarily the insoluble keratins which gave substantial labeling (lanes b and e). Thus, unless fatty are synthesized in abundance in keratinocytes, were observed acids were removed fromboth endsof the protein, the labeling to precipitate nonspecifically (shown in lane b), but they did was restricted to ananchorage region at one terminus. not interfere with the interpretation. These results are con- The degree of transglutaminase labeling by tritiated myris- sistent with the finding that post-translational fatty acid tate or palmitate was not reduced (and may actually have acylation suffices to anchor the enzyme in the membrane. been stimulated slightly) by including 10 pg/ml cycloheximide in the culture medium (Fig. 3, e-h), which inhibits protein DISCUSSION synthesis by 96% in these cells (Rice and Green, 1978). To be The observed post-translational addition of fatty acid pro- sure this finding represented a lack of dependence on protein vides a molecular basis for membrane anchorage of the ma- synthesis of the incorporation of each fatty acid, the identity jority of keratinocyte transglutaminase and is consistentwith of the protein-bound radioactivity was analyzed. Thus, the the minimal fatty acid labeling of the small fraction found in immunoprecipitated tritium-labeled protein was subjected to a soluble form. This enzyme, then, resembles the folate- acid methanolysis in parallel with fatty acid standards and binding protein in cultured KB cells, which is expressed in the products were analyzed by high performance liquid chro- membrane-bound and soluble forms that differ detectably matography. The profile from labeling with either fatty acid only in the lack of covalently bound fatty acid in the latter was unchanged by cycloheximidetreatment of the cells. When (Luhrs et al., 1987). In the present case, release of a soluble the protein was labeled with [3H]palmitic acid, methyl pal- transglutaminase from cell particulates by neutral hydroxyl- mitate was obtained in high yield and was the only major strongly indicates fatty acid acylation is required for product obtained (Fig. 5a). In contrast, when the protein was anchorage. If the soluble form does not differ from the acyl- labeled with [3H]myristi~acid, approximately 40% of the ated form in other respects, its relative amount could reflect esterified fatty acid eluted from the column at theposition of the efficiency of acylation after translation. No information methyl myristate and 60% with methyl palmitate (Fig. 4b). is available as towhich amino acid residues are acylated and 628 Acylation Acid TransglutaminmeFatty whether single enzyme molecules contain both fatty acids. been successful (Rice et al., 1988). This observation by itself Isolation and sequencing of fatty acid-containing peptides does not reflect fatty acid recycling, however, since the solu- from the myristate-labeled protein, for example, are antici- bilized enzyme exhibits a trypsin-like ~leavage,~which would pated to clarify this point. Conclusive evidence concerning a in any case result in release from the membrane. requirement for myristate or palmitate or bothfor anchorage The incorporation of two different fatty acids into keratin- may be generated by alteration of the appropriate acylated ocyte transglutaminase raises the interesting possibility that residue(s) by site-specific mutagenesis. This approach has each hasa distinct function. For example, one might be shown, for example, that the ras-related yeast YPTl protein sufficient for anchorage, whereas the other interacts with has 2 COOH-terminal subject to palmitoylation substrates. The possibility that acylated fatty acids could have (Molenaar et al., 1988), whereas only 1 of 2 neighboring a structural function other than simply anchorage has been cysteines in the human transferrin receptor becomes palmi- suggested by the finding that the transferrin receptor (Jing toylated (Jing andTrowbridge, 1987). and Trowbridge, 1987) and the histocompatibility antigen Myristoyl linkages in proteins largely are resistant to hy- HLA-D are palmitoylated post-translationally, despite their droxylamine treatment and occur cotranslationally (Olson et anchorage by hydrophobic amino acid segments. The fatty al., 1985;Magee and Courtneidge, 1985; McIlhinney et al., acid modification can have marked structural consequences, 1985; wilcox et al., 1987), leading to observation of N-myris- since, in thelatter, it reportedly prevents bond toyl-glycine at theamino terminus as the predominant form formation (Koch and Hammerling, 1986). The apparentspec- of this modification (Paul et al., 1987). However, a clear ificity of labeling of cellular proteins with palmitateand example of post-translational N-myristoylation has now been myristate (Olson et al., 1985; Magee and Courtneidge, 1985; reported in the B-lymphocyte membrane immunoglobulin McIlhinney et al., 1985) presumably reflects (i) thepaucity of heavy chain (Pillai and Baltimore, 1987). In the keratinocyte proteins containing bothN-acyl myristate and S- (or 0-)acyl Downloaded from transglutaminase, myristoylation (and palmitoylation) has palmitate and (ii) the marked fatty acid specificities of the been localized to a 10-kDa anchorage region at the amino or transacylases involved. In cell-free systems, the transacylases carboxyl terminus of the protein. This modification also is forming hydroxylamine-sensitive linkages may showa decided clearly post-translational but is not resistant to hydroxyl- preference for incorporating palmitate, but myristate can be amine. In one survey, 30-40% of myristate incorporated into incorporated as well (Schmidt, 1984; Riendeau and Guertin, protein is releasable by hydroxylamine (Olson et al., 1985). 1986). Thus, whether keratinocyte transglutaminase hasmore www.jbc.org Some myristate labeling could appear hydroxylamine-sensi- than a single acylation site remains to be established. In any tive if the added fatty acid actually acylated had first been case, the soluble guinea pig liver tissue transglutaminase has elongated to palmitate, as can occur in cells (Schmidt, 1984). 17 free sulfhydryls (Folk and Cole, 1966; Ikura et al., 1988); a

Although this elongation did occur to a considerable extent similar number on the keratinocyte enzyme would provide at unknown institution, on January 18, 2010 in the present case, it is clear that myristate and palmitate more than enough sites for the observed fatty acid acylation. were both incorporated in hydroxylamine-sensitive (ester or Indeed, comparison of the structures of the two ) linkages. Several attempts to obtain amino acid expressed by a given cell type such as human keratinocytes sequence information from an estimated 20 nmol of the (Thacher and Rice, 1985) might prove valuable for under- trypsin-released 80-kDa form electroeluted from gels or blot- standing the specificity of the acylating machinery leading to tedonto Immobilon polyvinylidene difluoride membrane membrane anchorage only of one of them. (Matsudaira, 1987) gave negative results, suggesting that its NH2 terminus isblocked.' Although not unique, the most REFERENCES plausible interpretation of these results is that the protein is Aderem,A. A., Albert, K. A,, Keum, M. M., Wang, J. K. T.,Greengard, anchored by fatty acid at thecarboxyl terminus andis blocked P., and Cohn, Z. A. (1988) Nature 332, 362-364 at the NH, terminus but not by myristate. We note that the Allen-Hoffman, B. L., and Rheinwald, J. G. (1984) Proc. Natl. Acad. tissue transglutaminase also has a blocked NH, terminus, Sci. U. S. A. 81, 7802-7806 Aviv, H., and Leder, P. (1972) Proc. Natl. Acad. Sci. U. S. A. 69, which may be pyroglutamic acid (Connellan et al., 1971) or 1408-1412 acetylalanine (Ikura et al., 1988). Our current approach of Chirgwin, J. M., Przybyla, A. E., MacDonald, R. J., and Rutter, W. localizing fatty acid acylation sites in peptides isolated from J. (1979) Biochemistry 18,5294-5299 proteolytic digests, in concert with the primary structure Chow, M., Newman, J. F. E., Filman, D., Hogle, J. M., Rowlands, D. deduced from the cDNA sequence, should permit J., and Brown, F. (1987) Nature 327,482-486 more definitive analysis of the orientation of the anchorage Connellan, J. M., Chung, S. I., Whetzel, N. K., Bradley, L. M., and Folk, J. E. (1971) J. BWl. Chem. 246, 1093-1098 region in theprotein. Folk, J. E., and Cole, P. W. (1966) J. Biol. Chem. 241, 5518-5525 Palmitate acylation has been found to occur in the absence Green, H. (1979) Harvey Lect. 74, 101-139 of protein synthesis in general (Magee and Courtneidge, 1985; Hoeg, J. M., Meng, M. S., Ronan, R., Fairwell, T., and Brewer, H. B., McIlhinney et al., 1985) and in the specific cases of ankyrin Jr. (1986) J. Biol. Chem. 261,3911-3914 (Staufenbiel and Lazarides, 1986) and N-ras (Magee et al., Ikura, K., Nasu, T., Yokota, H., Tsuchiya, Y., Sasaki, R., and Chiba, 1987) as well as in the present study. Although exogenous H. (1988) Biochemistry 27,2898-2905 Jing, S., and Trowbridge, I. S. (1987) EMBO J. 6, 327-331 stimulation of hydroxylamine-resistant myristoylation in Koch, N., and Hammerling, G. J. (1986) J. BWl. Chem. 261, 3434- cells such as macrophages (Aderem et al., 1988) most likely 3440 becomes manifest during synthesis of the modified proteins, Laemmli, U. K. (1970) Nature 227,680-685 post-translational recycling of esterified fatty acids could Lichti, U., Ben, T., and Yuspa, S. H. (1985) J. Biol.Chem. 260, serve a modulatory role during the functional life of the 1422-1426 protein. Indeed, the effect of turnover at two sites might even Luhrs, C. A., Pitiranggon, P., da Costa, M., Rothenberg, S. P., Slomiany, B. L., Brink, L., Tous, G. I., and Stein, S. (1987) Proc. be greater than for a single site. The solubility of the enzyme Natl. Acad. Sci. U. S. A. 84, 6546-6549 or its interaction with protein substrates could plausibly be Magee, A. I., and Courtneidge, S. A. (1985) EMBO J. 4,1137-1144 altered in this fashion. Attempts to demonstrate endogenous Magee, A. I., Gutierrez, L., McKay, I. A., Marshall, C. J., and Hall, solubilization of the enzyme from particulate extracts have A. (1987) EMBO J. 6, 3353-3357

R. Chakravarty and R. H. Rice, unpublished data. X. Rong and R. H. Rice, manuscript in preparation. TransglutaminmeFatty Acid Acylation 629 Matsudaira, P. (1987) J. Biol. Chem. 262, 10035-10038 F., eds) Vol. 231, pp. 51-61, Plenum Press, New York. McIlhinney, R. A. J., Pelly, S. J., Chadwick, J. K., and Cowley, G. P. Rothblat, G. H., Arbogast, L. Y., Ouellette, L., and Howard, B.V. (1985) EMBO J. 4, 1145-1152 (1976) In Vitro (Rockville) 12, 554-557 Means, G. E., and Feeney, R. E. (1968) Biochemistry 7,2192-2201 Schmidt, M. F. G. (1984) EMBO J. 3, 2295-2300 Michel, S., Schmidt, R., Robinson, S. M., Shroot, B., and Reichert, Schwartz, M. L., Pizzo, S. V., Hill, R. L., and McKee, P. A. (1973) J. U. (1987) J. Invest. Dermatol. 88, 301-305 Biol. Chern. 248,1395-1407 Molenaar, C. M. T., Prange, R., and Gallwitz, D. (1988) EMBO J. 7, Sefton, B. M., and Buss, J. E. (1987) J. Cell Bid. 104, 1449-1453 971-976 Simon, M., and Green, H. (1984) Cell 36,827-834 Nagae, S., Lichti, U., DeLuca, L. M., and Yuspa, S. H. (1987) J. Simon, M., and Green, H. (1985) Cell 40,677-683 Invest. Dermatol. 89, 51-58 Stadler, J., Gerisch, G., and Deppert, W. (1985) EMBO J. 4, 1153- Oslon, E. N., Towler, D. A., and Glaser, L. (1985) J. Biol. Chem. 260, 1156 3784-3790 Staufenbiel, M., and Lazarides, E. (1986) Proc. Natl.Acud. Sci. Paul, A. V., Schultz, A., Pincus, S. E., Oroszlan, S., and Wimmer, E. U. S. A. 83,318-322 (1987) Proc. Natl. Acud. Sci. U. S. A. 84, 7827-7831 Streuli, C. H., and Griffin, B. E. (1987) Nature 326, 619-622 Pillai, S., and Baltimore, D. (1987) Proc. Natl. Acud. Sci.U. S. A. 84, Swartzendruber, D. C., Wertz, P. W., Madison, K. C., and Downing, 7654-7658 D. T. (1987) J. Inuest. Dermatol. 88, 709-713 Reindeau, D., and Guertin, D. (1986) J. Biol. Chem. 261, 976-981 Thacher, S. M., and Rice, R. H. (1985) Cell 40,685-695 Rice, R. H., and Green, H. (1978) J. Cell Biol. 76, 705-711 Towler, D., and Glaser, L. (1986) Biochemistry 25,878-884 Rice, R. H., and Green, H. (1979) Cell 18,681-694 Warhol, M. J., Roth, J., Lucocq, J. M., Pinkus, G. S., and Rice, R. H. Rice, R. H., Chakravarty, R., Chen, J., O’Callahan, W. O., and Rubin, (1985) J. Histochem. Cytochern. 33, 141-149 A. L. (1988) in Advancesin Post-Translational Modifications of Wilcox, C., Hu, J.-S., and Olson, E. N. (1987) Science 238, 1275- Proteins and Aging (Zappia, V., Galleti, P., Porta, R., and Wold, 1278 Downloaded from www.jbc.org at unknown institution, on January 18, 2010