Translation of Silk Fibroin Messenger RNA in an Ehrlich Ascites Cell-free Extract

RNA identified by its base composition and T, RNase oligonucleotide pattern as the message for silk fibroin was purified from mature posterior silk glands of’ Bombyx mori larvae and used to direct polypeptide synthesis in an Ehrlich ascites cell-free extract. Fibroin mRNA stimulated [3H]alanine incorporation about 3- to 4-fold in the presence of 80 mM K+ and 4 mM Mg2+. The stimulation was reduced in the presence of 5 x 10m6 to 1O-4 M aurintricarboxylic acid, an inhibitor of the initiation of protein synthesis. The cell-free products were heterogeneous in size, including peptides as large as 100,000 daltons. They co-precipitated with carrier fibroin sequences after digestion with trypsin. A large fraction of the polypeptides synthesized in response fibroin was precipitated by directed against amino acid sequences in noncrystalline polypeptides


Translation
of Silk Fibroin Messenger RNA in an Ehrlich Ascites Cell-free Extract (Received for publication, December 3, 1974)  was reduced in the presence of 5 x 10m6 to 1O-4 M aurintricarboxylic acid, an inhibitor of the initiation of protein synthesis. The cell-free products were heterogeneous in size, including peptides as large as 100,000 daltons. They co-precipitated with carrier fibroin sequences after digestion with trypsin. A large fraction of the polypeptides synthesized in response to fibroin mRNA was precipitated by antiserum directed against amino acid sequences in noncrystalline region polypeptides of fibroin. Furthermore, after digestion with chymotrypsin, a major fraction of the cell-free products specifically co-precipitated with crystalline region sequences of native fibroin. The size and amino acid composition of the fibroin crystalline region polypeptides isolated from the cell-free products were similar to those from native fibroin.
Several messenger RNA species have been identified from their cell-free translation products (e.g. l-5). In contrast, the mRNA for silk fibroin of Bombyx mori has been identified solely by its characteristic base composition, and by certain unusual properties of its nucleotide sequence (6). Purified fibroin mRNA has been shown to stimulate amino acid incorporation in mouse Ehrlich ascites cell extracts (7), but the unusual properties of fibroin protein made impractical the application to the cell-free products of several of the common techniques for protein identification.
Silk fibroin has a molecular weight of 3.5 x lo5 (8). Since cell-free synthesis of a complete molecule that is so large seemed unlikely, determination of the molecular weight of the cell-free products by polyacrylamide gel electrophoresis could not be used as a criterion of fibroin synthesis. Moreover, tryptic peptide mapping is not a practical method for the identification of fibroin, since digestion with trypsin results in precipitation of oligopeptides representing 90% of the molecule (9). More than 90% of the amino acids in fibroin are glycine, alanine, serine, and tyrosine, in a ratio of 44.5:29.4:12.1:5.2. The molecule is organized into many repeats of two kinds of sequences termed crystalline region and noncrystalline region polypeptides.
Chymotrypsin digestion of fibroin releases crystalline region polypeptides, which crystallize and precipitate. Crystalline region polypeptides contain 60% of the amino acids in the protein, and have a sequence proposed by Lucas et al. (10) to be Gly-Ala-Gly-Ala-Gly[Ser-Gly-(Ala-Gly),l,-Ser-Gly-Ala-Ala-Gly-Tyr *Present address, New England Nuclear, 575 Albany Street, Boston, Massachusetts 02118.
where n is usually 2 and has an average value of 2.
The major noncrystalline region sequences remain soluble after chymotrypsin digestion of fibroin, yet except for the more frequent occurrence of tyrosine, resemble those sequences which precipitate.
Antibodies elicited by fibroin in rabbits seem to be directed against certain of these soluble peptides (11)(12)(13).
For a review of the structural aspects of fibroin protein, see Lucas and Rudall (14).
We have taken advantage of the unusual features of this protein to detect the cell-free synthesis of fibroin polypeptides by specific precipitation after digestion with trypsin and by a chymotrypsin digestion, co-crystallization technique which specifically detects fibroin crystalline region polypeptides. An immune precipitation procedure which detects peptides in the noncrystalline region of fibroin has also been used.

MATERIALS AND METHODS
Isolation and Identification of Silk Fibroin mRNA-Diapause eggs of Bombyx mori were purchased from Turtox General Biological Supply House, Chicago, Illinois, and stored at 4" until used. Eggs hatched after about 2 weeks at 25". Animals were raised at 25". and fed local mulberry leaves. On the 8th or 9th day of the fifth instar (just prior to cocoon spinning) larvae were immobilized on ice and dissected under ice-cold SSC.' The posterior silk glands were dissected out, rinsed in cold SSC, and frozen in liquid nitrogen. Storage of frozen

Isolation
and Identification of Fibroin mRNA-Fibroin mRNA was extracted from posterior silk glands removed from B. mori larvae just prior to spinning, and purified by separation procedures based on its large size (6). Fig. la shows the fractionation obtained when whole posterior silk gland RNA is passed over a column of Sepharose 2B, which excludes polynucleotides larger than 2 x lo6 molecular weight. Thus, fibroin mRNA (5.5 to 6.0 x lo6 molecular weight) (23) is largely freed of other RNA in this single step. The RNA shown by the brackets in Fig. la was collected and heated in a solution of low ionic strength to disrupt aggregates before sedimentation in a sucrose gradient (Fig. lb). Because they contain an internal nick, contaminating 28 S rRNA molecules sediment as two discrete fragments together with 18 S rRNA after denaturation (24, 25). Only small amounts of rRNA were removed by a second heating'and sucrose gradient sedimentation (Fig. 1~ Optimum stimulation was obtained at a fibroin mRNA concentration of 100 Kg/ml in the presence of 80 mM K+ and 4 mM Mg'+. Amino acid incorporation was linear for 40 min and continued for 120 min (Fig. 2).
The size of the reduced and alkylated product was determined by polyacrylamide gel electrophoresis to be heterogeneous, and to include polypeptides estimated relative to reovirus structural protein markers to be of up to lo5 molecular weight (Fig. 3A). No polypeptides as large as native fibroin (3.5 x 105) (8) were detected in 10% gels. Glycine and alanine are enriched relative to leucine and valine in polypeptides greater than 20,000 molecular weight, a result consistent with the amino acid composition of fibroin, and the fact that the smaller peptides are synthesized in the same amount in the absence of added mRNA (Fig. 3B).
The protein synthesis inhibitor, aurintricarboxylic acid (ATA), inhibits the stimulation of amino acid incorporation significantly at concentrations between 5 x lo-@ M and lo-' M, but has little effect on endogenous incorporation (   EMC viral RNA are poorly precipitated (Table III). Thus, nonfibroin polypeptides are soluble after a treatment which precipitates both authentic fibroin polypeptides and the cellfree polypeptides directed by fibroin mRNA. Despite exhaustive washing .and hot TCA extraction, in four such experiments, more than 100% of the hot TCA-precipitable radioactivity initially present was found in the tryptic precipitate of fibroin mRNA-directed cell-free products. Some of this precipitated radioactivity presumably reflects trapping of unincorporated labeled amino acids in the gelatinous precipitate formed from fibroin under these conditions (9).

Detection of Crystalline
Region Polypeptides-The specific precipitation of fibroin crystalline region polypeptides after chymotrypsin digestion provided another method for the identification of fibroin sequences in the cell-free products. Fibroin mRNA was used to stimulate protein synthesis in the presence of [3H]alanine (34% of the crystalline region polypeptide amino acids are alanine) and ["Clleucine (absent from crystalline region polypeptides). Carrier fibroin was then added, and a precipitate formed by digestion with chymotrypsin. One-third of the incorporated labeled alanine precipitates together with the authentic fibroin crystalline region polypeptides, compared to only 3% of the leucine (Table IV). When the same experiment was performed with polypeptides formed in response to added EMC viral RNA, less than 2% of the incorporated alanine or leucine was precipitated. This demonstrates that a significant fraction of the cell-free synthesis consists of alanine-containing polypeptides which are similar enough to fibroin crystalline region polypeptides to co-crystallize and precipitate with them. From the amino acid composition of intact fibroin and fibroin crystalline region, the fraction of the total amount of each amino acid in fibroin which precipitates with chymotrypsin can be calculated. This criterion was applied to the cell-free products by using fibroin mRNA to stimulate amino acid incorporation in a series of 12 tubes, each containing a different tritiated amino acid, but otherwise identical.
In each case, total incorporation and the fraction of it that was precipitated after digestion were determined and compared to the prediction for native fibroin.
The results indicate that cell-free polypeptides similar in amino acid composition to the authentic fibroin crystalline region polypeptides are precipitated by this technique (Table V). The eight amino acids not occurring in native fibroin crystalline region polypeptides precipitate poorly, but the fraction of incorporated glycine, alanine, serine, and tyrosine precipitated was close to the predicted value. This result suggests that polypeptides with an amino acid composition resembling that of authentic fibroin crystalline region polypeptides compose about the same fraction of the cell-free synthesis stimulated by fibroin mRNA as would be found if native fibroin were the primary product.
The actual amino acid composition of the cell-free crystalline region polypeptides could not be reliably calculated from the data of Table V because different amounts of products could have been synthesized in each tube. Consequently, the relative quantities of glycine, alanine, serine, and tyrosine present in the crystalline region polypeptides synthesized in a single large reaction were determined.
The specific activity of each precursor amino acid was determined by amino acid analysis of a hot acid extract of a sample of the reaction mixture taken before protein synthesis began. In order to prevent incorporation of unlabeled amino acids from endogenous charged tRNAs prior to incorporation of labeled amino acids, the ascites extract was freed of endogenous tRNA and supplemented with deacylated ascites cell tRNA (see under "Experimental Procedure"). After synthesis had proceeded for 120 min, the cell-free products were digested with chymotrypsin. The resulting precipitate of crystalline region polypeptides was hydrolyzed and the incorporated radioactivity in glycine, alanine, serine, and tyrosine was quantitated.
The amino acid composition of the cell-free crystalline region polypeptides was calculated using the precursor specific activities. Table VI shows that the polypeptides precipitated from the cell-free product by chymotrypsin digestion are unusual peptides. More than 40% of the amino acids in these polypeptides are glycine, and [3H]tyrosine as precursors in a single 1.2.ml reaction. The.initial specific activity of each amino acid was determined on a hot acid extract of an aliquot taken before synthesis began. After incubation for 120 min, the cell-free products were precipitated with chymotrypsin. The cell-free crystalline region polypeptides were hydrolyzed, and the radioactivity in each amino acid was quantitated. Results were calculated in terms of picomoles by using the specific activities, and then were converted to per cent values. The authentic per cent total was calculated from the total amino acid composition of crystalline region polypeptides prepared from cocoon fibroin (data not shown). crystalline region polypeptides prepared from authentic fibroin. Serine and tyrosine are also present in approximately the same proportions as in authentic crystalline region polypeptides.
Size of the Cell-free Crystalline Region Polypeptides-Fibroin crystalline region polypeptides are insoluble in water and most common protein solvents, but can be solubilized in 10 M LiBr. Thus, to estimate the relative sizes, labeled cell-free crystalline region polypeptides and unlabeled authentic crysand more than an additional 30% are alanine, as in the case for talline region polypeptides were co-chromatographed on a  Fig. 1, and passed over a 3.ml oligo(dT)-cellulose column in 0.5 M NaCl, 0.01 M Tris-Cl, pH 7.5. Unbound mRNA was that which was not retained by the column after three passages, bound mRNA was eluted with 0.01 ~Tris, pH 7.5. Cell-free protein synthesis, chymotrypsin digestion and precipitation, and antiserum precipitation were as described under "Experimental Procedure." Fibroin mRNA was fractionated on oligo(dT)-cellulose and tested in the ascites cell-free system. The bound and unbound mRNA were not distinguished by their total stimulation of amino acid incorporation or in the differential stimulation of alanine over leucine incorporation, and both mRNA fractions exhibited the same dependence of stimulation of alanine incorporation upon mRNA concentration (Table VII). The bound and unbound molecules stimulated the synthesis of fibroin polypeptides equally well. Each led to the synthesis of similar amounts of fibroin crystalline region polypeptides, as shown by the co-crystallization assay, and of noncrystalline region fibroin polypeptides, as shown by specific precipitation with antifibroin serum (Table VII). FIG. 4. Chromatography of chymotryptic precipitates on Sephadex G-50. Cell-free crystalline region polypeptides labeled with [3H]alanine (A---A) and unlabeled crystalline region polypeptides prepared from cocoon fibroin (-) were co-chromatographed in 10 M LiBr on a 3 x 40 cm column of Sephadex G-50 fine. A,,,, was monitored, and 4-ml fractions were counted in 7 ml of Aquasol. Molecular weight markers, run separately, were: insulin, 5,800; RNase A,13,700;chymotrypsinogen,25,000;cocoon fibroin,60,000 to 200,000,excluded. column of G-50 fine Sephadex at room temperature in 10 M LiBr. About 75% of the labeled polypeptides eluted in a major peak slightly ahead and 15% in a shoulder which chromatographed slightly behind the authentic crystalline region polypeptides (Fig. 4). About 10% of the labeled material is excluded from the column, and thus, had a molecular weight greater than 30,000.
The average molecular weight for the unlabeled crystalline region polypeptides was estimated to be about 6800. The sequence postulated by Lucas et al. (10) has a calculated molecular weight of 5087. While the molecular weight markers were chromatographed separately, and thus, the estimate is not extremely accurate, it appears that the actual molecular weight of the average authentic crystalline region polypeptide may be somewhat larger than that proposed by Lucas et al. (10). The apparent molecular weight of the major peak of cell-free crystalline region is 8200. Chromatography-Fibroin mRNA is not retained by either cellulose nitrate filters or glass filters impregnated with poly (U) under conditions in which poly (A) is bound (7). However, about 40% of the 47 S mRNA molecules bind to 5119 Evidently, mRNA molecules exhibiting differential affinity for oligo(dT)-cellulose do not differ significantly in their translation in the ascites cell-free system. Similar results have been reported for globin mRNA translation in a cell-free system (31). Recent studies, in fact, suggest that mRNA poly(A) sequences may not be involved in the initial round(s) of protein synthesis, but rather act in an unknown way to preserve the long translational efficiency of the mRNA (32, 33). The mechanism of this protection is unknown.

DISCUSSION
The messenger RNA for silk fibroin was the first mRNA to be identified by chemical means. Its characteristic base composition and unusual nucleotide sequence properties were both predicted from the structure of fibroin protein (6). Support for this identification of fibroin mRNA is provided by the demonstration of fibroin polypeptide synthesis directed by the mRNA in a mouse Ehrlich ascites cell extract. sL. P. Gage, unpublished observations. Three separate techniques have been used to demonstrate the cell-free synthesis of fibroin peptides in response to added mRNA. An immunoprecipitation technique detected noncrystalline region fibroin peptides. Two other tests were based on the tendency of fibroin peptides to form specific precipitates when digested with proteases. Cebra (13) analyzed the structure of the antifibroin-binding site by quantitative inhibition studies employing fibroin peptides. He concluded that the binding site for rabbit antifibroin antibodies was closely approximated by or was identical to a dodecapeptide composed of glycine and alanine and some tyrosine, much like the noncrystalline region sequences left in solution after chymotrypsin digestion of fibroin (12). Cebra's results led to an estimate of about 20 antibody binding sites/molecular weight of 400,000, and there are just a sufficient number of these noncrystalline region peptides per monomer (34) to account for the number of antigenic sites. The rabbit antifibroin serum used in the present work was shown to be identical to a sample of Cebra's antifibroin serum by an Ouchterlony double diffusion test against fibroin. Thus, formation of specific antiserum-precipitable material is evidence for the cell-free synthesis of noncrystalline region fibroin peptides. When a solution of fibroin is digested with trypsin, a gelatinous precipitate forms which contains 90% of the amino acids (9). This procedure precipitates cell-free polypeptides directed by fibroin mRNA, but not those directed by EMC virus RNA, or those synthesized in the absence of added RNA. This unusual behavior upon trypsin digestion demonstrates the similarity of authentic fibroin polypeptides and the cellfree polypeptides directed by fibroin mRNA. Furthermore, the behavior of both is distinct from that of polypeptides with a more typical amino acid composition such as those synthesized in the presence of EMC virus RNA or on endogenous templates.
As shown by their different sizes, the cell-free crystalline region polypeptides and the average crystalline region polypeptide produced from authentic fibroin are not identical. The in vitro synthesized crystalline region polypeptide has an estimated average molecular weight of 8200, and thus, contains about 88 amino acids. The apparent average molecular weight of the authentic crystalline region polypeptides is 6800, corresponding to 79 amino acids. Thus, the major cell-free crystalline region polypeptide is nine amino acids longer than the average authentic crystalline region polypeptide. Lucas et al. (10) have suggested that the length of the crystalline region polypeptides may not be rigidly fixed. There are at least 30 copies of this polypeptide per fibroin molecule, but only about one-quarter of the entire molecule was synthesized in the cell-free extract. If initiation occurs properly, the slightly larger size of the crystalline region polypeptides synthesized in vitro may indicate a bias toward larger units of this repeated sequence near the NH,-terminal end of the fibroin molecule.
The chymotrypsin-digested, precipitated material excluded from the G-50 column may represent the initial crystalline region polypeptide, substantially larger than the others because it contains a part of the amino-terminal sequence which lacks tyrosine.
Qualitatively, of 12 amino acids polymerized in fibroin mRNA-directed assays, only glycine, alanine, serine, and tyrosine were found to co-precipitate with authentic crystalline region polypeptides.
The quantitative amino acid composition of the cell-free chymotryptic precipitate was found to be very authentic crystalline region polypeptides are approximately 80 to 90 amino acids long, and contain more than 35 glycine residues, and more than 25 alanine residues.
Specific inhibition by ATA suggests that the stimulation of polypeptide synthesis by added fibroin mRNA is initiationdependent.
In the experiments described in Table II, 44.4% of  the hot TCA-precipitable [ SH]alanine was precipitated by antifibroin serum. The stimulation of amino acid incorporation above the endogenous level in this experiment was 3.8-fold. Since endogenous synthesis appears to continue in the presence of added mRNA (Fig. 2), the results indicate that at least 60% of the [3H]alanine specifically incorporated in response to fibroin mRNA is incorporated into fibroin sequences. In the experiment described in Table IV, 32.4% of the hot TCA-precipitable [SH]alanine was precipitated after digestion with chymotrypsin.
In this experiment, the stimulation of alanine incorporation above the endogenous level was 3.3-fold. Assuming that the endogenous synthesis also occurs in the presence of added mRNA, and that the radioactivity precipitable with chymotrypsin represents 67% of the total alanine incorporated into fibroin, 70% of the [3H]alanine incorporation in response to fibroin mRNA is incorporated into fibroin polypeptides.
The average value for four similar experiments was that 79% of the [3H]alanine incorporated in response to fibroin mRNA is incorporated into fibroin polypeptides. Thus, cell-free synthesis was usually initiated in frame on fibroin mRNA, and elongated correctly for the duration of synthesis.
Since globin mRNA is often used for cell-free protein synthesis at a concentration of approximately 5 fig/ml (35, 36), the requirement for 100 pug/ml of fibroin mRNA for maximal stimulation of polypeptide synthesis may seem unusually large. However, the molecular weight of globin p chain mRNA is only 2.27 x lo5 (37), compared to 5.5 to 6.0 x lo6 for fibroin mRNA (23). Thus, if there is one initiation site per molecule, 100 pg of fibroin mRNA contain the same number of initiation sites as 4 fig of globin /I chain mRNA.
However, whereas complete globin chains are synthesized by various cell-free systems, only partial fibroin chains are synthesized. In' rabbit and duck reticulocyte cell-free systems, approximately two mouse globin chains are synthesized per mRNA molecule in 30 min (38). Assuming that the crystalline region polypeptides described in Table VI represent 60% of the total fibroin synthesized in 1.2 ml, a similar calculation indicates that one fibroin chain of 50,000 molecular weight is synthesized/350 mRNA molecules in 30 min. Thus, globin chains are initiated about 700.fold more frequently per mRNA molecule than are fibroin chains.
In the mature silk gland, a single cell synthesizes about 1 pug of fibroin in 30 min (39), about 400-fold more than 1 ml of mouse Ehrlich ascites cell-free system. However, since a single silk gland cell contains only 170 ng of fibroin mRNA (40), the cell is more than 105-fold more efficient than a l-ml cell-free system.
It should be noted that synthesis of this unusual protein by the ascites cell extract occurred without addition of any specific initiation factors or tRNAs from the silk gland. The specific translation of an insect mRNA in a mammalian cell-free system again demonstrates the universality of the genetic code and the mechanisms of its expression.