OPTICAL ROTATORY DISPERSION AND CIRCULAR DICHROISMOF THE ,3-FORM OF SILK FIBROIN IN SOLUTION*

BY EISAKU IIZUKAt AND JEN Tsi YANG

CARDIOVASCULAR RESEARCH INSTITUTE AND DEPARTMENT OF BIOCHEMISTRY,

UNIVERSITY OF CALIFORNIA (SAN FRANCISCO)

Communicated by Julius H. Comroe, Jr., March 11, 1966

In the past decade ORD has become a powerful tool in characterizing the con-formations of proteins and polypeptides and estimating their helical contents.This "overemphasis" on the helical conformation, however, merely points up ourlack of knowledge about other structures in a protein molecule. As early as 1957we knew that the oligomers of y-benzyl-L-glutamate favored the formation of(-aggregates and displayed positive visible rotatory dispersion that resembledneither an a-helix nor a random coil.' The Cotton effects characteristic of the(3-form, however, have escaped detection until now, mainly because in most in-stances the (-forms can be found only in organic solvents that absorb stronglybelow 250 m1A. We have studied silk fibroin, which is known to possess the (3-con-formation in the solid state and therefore has the potential of retaining it in solutionunder favorable conditions. We found that the Cotton effects of the (-form of silkfibroin in mixed solvents of water with dioxane or methanol differ markedly fromthose of the helical and coiled conformations.Experimental.-Bombyx mori L. silk fibroin was dissolved in 9.3 M LiBr at 370C

for 2 hr, dialyzed thoroughly against water until an AgNO3 test could detect notrace of bromide ion, and clarified by centrifugation. The stock solution was thenmixed with an appropriate amount of dioxane or methanol and its apparent pHadjusted to 7.3.

All ORD was measured with a Cary 60 spectropolarimeter at 270C and the CDwith a Jasco ORD/UV-5 with a CD attachment at room temperature, using cellsof light paths between 0.05 and 100 mm. The data were expressed as reducedmean residue rotation, [m'], and mean molar ellipticity, [0], which equals 3,300

- Er),2 using a mean residue molecular weight of 78. The infrared spectra of thedeuterated solutions and cast films (from aqueous solutions) were measured withPerkin-Elmer 21 and 337 spectrophotometers.Results.-The ORD of silk fibroin in aqueous solution changes drastically when

more than 30 per cent (v/v) dioxane or methanol is present in solution. Thesechanges are time-dependent: the larger the amount of organic solvent, the fasterthey reach a limiting plateau. The Cotton effects of silk fibroin in the two mixedsolvents (Fig. 1) display a trough at 229-230 my and a peak at 205 m/,u and anothertrough near 190 my. (Because of solvent absorption, we could not extend measure-ments to 190 m.i for dioxane-water solution, but similarity with the ORD ofmethanol-water solution at longer wavelengths argues for the existence of thistrough in dioxane-water solution.) They contrast sharply with those of a coiledconformation, which shows a 205-mA trough and a 190-mu peak instead.4: Theyalso differ from those characteristic of an a-helix, which has a 233-mu trough and a198-mM peak with a shoulder near 215 mM.4, 5 Recently, the a-helix measurementswere extended toward 182 my, and a second trough was found near 184 my (S. D.

1175

1176 BIOCHEMISTRY: IIZUKA AND YANG PROC. N. A. S.

0~~~~~~~~~~~~~~~~~~~~~~~0-41.5 +3

.41 2

0 +2~~~~~~~~~~~~~~~~

+0.

0

0.5

180 200 220 240 2607WAVELENGTH, mpu

FIG. 1.-Reduced mean residue rota- 180 200 220 240 260tions of silk fibroin in mixed solvents WAVELENGTH, mp(v/v). Curves: 1, 93% methanol; 2, FIG. 2.-Mean molar ellipticity of silk50% methanol; 3, 50% dioxane; and 4 fibroin in mixed solvents. Symbols samewater only. as in Fig. 1.

Duke, Durrum Instrument Corp., private communication). Note that the averageof 184 and 198 mg is very close to the positive dichroic band at 190 mg.The circular dichroism of the same solutions also suggests a conformation which

is neither the a-helix nor the coiled conformation. Instead of two minima near 222and 206 mgA and a maximum at 190 mu, characteristic of the helix,' Figure 2 showstwo well-defined dichroic bands, one negative at 218 mg and the other positive at197 mg in methanol-water solution. An almost identical negative band wasobserved in dioxane-water solution, but the measurements were limited to wave-lengths above 210 mju. In striking contrast, the CD of the coiled conformationactually has a larger negative band near 199 mgA and a very small positive band near230 mgA.6 Note also that the 197-mg band is close to the 194-mg absorptionmaximum of the 3-form, just as the a-helix has a dichroic band at 190 mg1 and anabsorption maximum at the same wavelength.4 (On the other hand, the coiledconformation has a dichroic band at 200 mgA, but an absorption maximum at 192mg.4) Furthermore, the averages of the 229-230-mg trough and 205-mg peak andalso the 205-mg peak and 190-mg trough are very close to the 218- and 197-mg bands.

Next, we studied the amide I and V bands of the infrared spectra. In dioxane-D20 or CH30D-D20 solutions, silk fibroin shows a 1620-cm-1 as well as a 1650-cm-' band, and its cast films (from aqueous solutions) reveal two bands at 690 and640 cm-' (Fig. 3). Both the 690- and 1620-cm-1 bands are characteristic of the 13-form and are absent from silk fibroin in its coiled form (the other two bands in Fig. 3are identified with the coiled conformation) .7, 8 The amide I band near 1690 cm-'also suggests the presence of the 1-form, probably of the antiparallel type. FromFigure 3 we conclude that not all the coiled form was converted into 13-form in themixed solvents. By resolving graphically the amide I band into three (for the two 13and one coiled bands) and using a method similar to that of Wada et!al.,9 we esti-mated about 50 per cent 1-form in the mixed solvents. [Since the concentrations

VOL. 55, 1966 BIOCHEMISTRY: IIZUKA AND YANG 1177

0oo 700 600 500 40010 o +3

\ I- ' -+

2' ~~~~~~~~~200220 240 260/''sWe4' ~~~~~WAVELENGTH, my

(J) /\0_

/IFIG. 4. alculated rotations of hy-le 00 1700 1600 1500 1400 pothetical mixed conformations of

WAVE NUMBER, cm-1 a-helix (H), #-form. (ft), and coil (C).Solutions: (a) helical poly-Iglutamic

FIG. 3.-Infrared spectra of silk fibroin. acid in 1: 1 dioxane-water (pH 7.3);Bottom, in mixed solvents (v/v); curves: (b) coils of (a) in water only; (c) 50%I ', 93: 7 CHO0D-D20 * 2', 1: 1 CH30D-D20;* tform. of sil fibroin in 1: 1 methanol-8/, 1: 1 dioxane-D20; and 4', D20 only. water; and (d) coils of (c) in 8 M urea.Top, cast films; curves 1-4 refer to the Curves: 1, 35H-15#-50C (35:30:35corresponding solutions 1'~-4' (bottom), ex- a-c-d) 2, 1H-9C (from a and b); 2',cept that H20 is used instead of D20. 1H-96 (fr'om a and d); and 3, 1fl-9C

(fromc and d).

(1-2%o) used for infrared studies were usually more than 10 times those for theORD measurements, we made several test runs using the same concentrated solu-tions for both experiments and found that the ORD remained essentially the sameas those shown in Fig. 1. ] Such estimates are by necessity very crude, but they doprovide the order of magnitude of the Cotton effects for a 100 per cent ,8-form.Table 1 summarizes the known Cotton effects of all three polypeptide conforma-

tions, but their magnitudes as reported in the literature are still uncertain and aretherefore not listed. Table 2 shows the Cotton effects of the #-form of silk fibroin inmixed solvents, and also the Moffitt parameters" for the visible rotatory dispersion.inr4 Figures 1 and 2 also reveal another feature which cautions us against any over-emphasis on the quantitative interpretation of the ,8-form at this stage. The mag-nitude of the 229-230-mg trough actually decreased and that of the 205-muA peak

TABLE 1COTTON EFFECTS OF I=POLYPEPTIDES IN VARIOUS CONFORMATIONS

, ~Optical Rotatory Dispersion, Mmu-- Dichroic Bands, mA----Cross- Cross-

Conformation Trough over Peak Negative over Positives-Form 229-230 -220 205 218 -208 197

-190 -196a-Helix 233 -220 198-199 222

0184 190 2060209 202 190-192Coil 238 (small) 228 (small)

204205 -198190 199-200

1178 BIOCHEMISTRY: IIZUKA AND YANG PROC. N. A. S.

increased when the per cent of methanol was increased from 50 to 93 (Fig. 1), eventhough infrared spectra suggest a somewhat higher per cent of 3-form in the moreconcentrated methanol solution. A similar upward shift is also evident in the CDprofile (Fig. 2). It is difficult to ascertain whether this is purely a solvent effect, orthe result of variation in the 3-forms, or both. It would therefore be premature totake the present numerical values too literally, especially when the discovery of theCotton effects of the ,3-form will surely be applied to the study of other proteins.The ORD of silk fibroin in the visible region obeys a one-term Drude equation,

but can also comply with the Moffitt equation. In the two mixed solvents studied,the bo's of the 3-form were close to zero. In highly aggregated solutions,3 however,the bo did become positive, but usually the Moffitt equation could only be applied inthese cases over a narrow wavelength range, say, above 400 mjA, or the plots were nolonger linear.

Since the Rs97 has a larger magnitude than the R229 (Table 2), the two dichroicbands would thus yield a dextrorotation in the visible region. It is highly suggestivethat another negative band exists below 180 mjA for the 1-form which would com-pensate for the dextrorotation. This band, if present, could even account for thesmall or zero magnitude of bo of the 1-form, when the three Drude terms arisingfrom the dichroic bands are represented in the Moffitt equation. 10 With the data inFigure 2, we can fit the visible rotatory dispersion of silk fibroin in 93:7 methanol-water with three Drude terms (expressed in terms of mean residue rotation):

[mIX = 1580(197)2/(X2 - 1972) - 450(218)2/(X2 -2182)-1500(177)2/(X2 - 1772). (1)

The last term on the right side of equation (1) can be further split into two Drudeterms,

1500(177)2/(X2 - 1772) = kl/(X2 - X12) + k2(X2 - X22), (2)

one representing the third dichroic band (negative) of the 13-form and the other thebackground rotations. The latter also includes the rotations due to the aromaticCotton effects near and below 280 m,4, which are too small to be measured accuratelyand are not shown distinctly in Figures 1 and 2.3 Since the disordered form of silkfibroin is levorotatory and usually has a Xc, [or X2 in equation (2)1 above 200 myt, Ximust then be smaller than 177 mIA, if both k's are negative.11 For example, withX2 = 209 m1A and k2 =-270(209)2 based on the data in 8M urea, X1 and ki were 151

TABLE 2COTTON EFFECTS OF THE fl-FoRm OF SILK FIBROIN IN MIXED SOLVENTS (V/V)*

_________________-_______________Solvents1:1 Dioxane-H20 1:1 Methanol-H20 93:7 Methanol-H20

50% , 100% 46% 100% , 52% ,8 100%[Mi]229 X 10-3 -3.5 -6 -2.8 -5 -1.7 -3[M'1205 X 10-3 +9.0 +20 +9.9 +24 +15.0 +27[Mi']190 X 10-3 -6.8 -17 -6.7 -16R218 X 1040 -5.0 -10 -5.8 -13 -4.9 -9R197 x 1040 +8.9 +22 +17.2 +33aot -260 Nonlinear -40bet +90 - +30 -

* Symbols: The first column under each mixed solvent is the experimental values, and the secondcolumn refers to the extrapolated values to 100% ,8-form.

t Based on the Moffitt equation (with No = 212 mp)):10 [m'] = aoeo2/(X2 - X52) + bosx4/(X2 - Xe2)2over the wavelength range 360-600 mAs.

VOL. 55, 1966 BIOCHEMISTRY: IIZUKA AND YANG 1179

m1A and - 1390(151)2, respectively. Similar analyses can be made for data in othermixed solvents. With four unknowns in equation (2), however, the solution is notunique.'2 Furthermore, if the 197-mu band is due to an exciton effect in the NV1region of the 1-form, one would expect a negative band only slightly below 190 m'Arather than (or in addition to) the one at much shorter wavelengths, e.g., 151 m/A(R. W. Woody, private communication). Very probably the Xi-term in equation(2) again represents the approximation of several Drude terms. Any definiteconclusion must then await a detailed theoretical treatment.Discussion.-The 1-form, if present in a protein molecule, would obviously com-

plicate an ORD analysis which assumes a linear model having only the helical anddisordered forms. Since the magnitude of [m'] at 229-230 and 205 mjA for the 1-form is less than one half that of an a-helix at the same wavelengths, the Cottoneffects of the former may easily escape detection, if both conformations coexist in theprotein molecule. The close proximity of their troughs near 229-233 m/A makes itdifficult to separate them experimentally, and the Cotton effects of chromophoresother than the peptide linkages might further mask such distinctions. The situa-tion is slightly better for the 198-mi peak, which undergoes a red shift when a sig-nificant amount of the 3-form is present. In addition, the is and coiled- forms sur-prisingly have opposing peaks and troughs below 210 mM; this would partially canceltheir rotatory contributions. To illustrate, we calculate the rotations of severalhypothetical mixed conformations. Curve 1 in Figure 4 has a distribution of variousconformations close to that found in lysozyme by the X-ray method.'3 The 233-mntrough and 198-miA peak are characteristic of helices, even though about 15 per cent13-form is present. The profile of 10 per cent helix-90 per cent coil, based on thedata on poly-L-glutamic acid (curve 2), differs significantly from that of the 13-form.Curves 2 and 2' both represent 10 per cent helix-90 per cent coil mixtures. Herewe encounter the problem of choosing the correct reference values. Poly-L-glu-tamic acid has been widely used as a model compound, but we have recently shown5that the coiled form of this polymer contracts in the presence of salt, and its rota-tions change accordingly. In comparison, the disordered form of silk fibroin hasunusually small rotations, even when one takes into consideration that this proteinhas more than 30 per cent glycine residues. For instance, the 205-mit trough inFigure 1 is only about 1/7 that of coiled poly-L-glutamic acid in water. Hydro-dynamic measurements of silk fibroin also suggest a rather compact molecule evenin 8 M urea,3 unlike the extended conformation of poly-L-glutamic acid in 8 M urea.(The possibility of a small amount of the 13-form in aqueous solution of silk fibroincannot be completely ruled out, but its existence seems very unlikely in 8 M urea.)Curve 3 clearly illustrates the partial cancellation of the rotations of the 13- and coiledforms.For the 233-miA trough and 198-mA peak (or the 229-mn trough and 205-mM peak)

we can write (assuming additivity of the optical rotations)

Im Ix= f ,[m'L] + fWm'] + fW[m'], (3)with 2f = 1. In principle, the three conformations can be determined with twosimultaneous equations, say, at 233 and 198 m1A. But the reference values forvarious forms are still very uncertain. In addition, the Cotton effects due to thearomatic groups and disulfide linkages can modify the peak and trough rotations.

1180 BIOCHEMISTRY: IIZUKA AND YANG PROC. N. A. S.

Thus, quantitative calculations would appear premature at present. But it can beshown from equation (3) that

fa = [([m'] - [m'c) - f,([m'], - [mC)]/([m ]a- [mkC)or

fGi = fa(apparent) - f,([m']6 - [m']c)/([m']a - [m']c). (4)Since both ([mm']c-[i' ]c) and ([n' ], - [m' ]c) have the same sign at either 233 or198 mjA, fa,, will always be less than f,,(app), which neglects the presence of the 3-form. Consider a hypothetical case of 35 per cent helix-15 per cent (3-50 per centcoil, using [m'L -16,000, [m']6 _-6,000, and [m']c__-2,000 at 233 m/A. Theapparent fcc turns out to be about 0.40 instead of the 0.35 when the fl-form is neg-lected. Thus, a moderate amount of the f-form, if present in a protein molecule,would raise the estimated helical content only slightly.With one or two exceptions, the bo of the Moffitt equation'0 for the fl-form in the

literature appears to be close to zero, whereas the corresponding ao differs from thatfor the coiled form. If future investigations substantiate these findings, we canthen estimate helical content with the Moffitt equation even in the presence of thef-form. So far the ao parameter seems to have been unjustifiably neglected, prob-ably because of the popular use of the bo value (there are instances in the literaturewhere the ao values are not even listed). We believe that the a0 will deservedly playan increasing role in our understanding of the fi and other conformations.

In view of the foregoing findings, we note with interest the recent work of Jirgen-sons,'4 who observed that the Cotton effects of many nonhelical proteins such as alacid glycoprotein, a-chymotrypsin, deoxyribonuclease, glucose oxidase, and serumy-globulin display a shallow minimum at 220-235 myA and a flat maximum at 200-205 miA. The bo's of these proteins are close to zero. It is therefore very temptingto suggest that these proteins exist partly in the f-form, which would be manifest innonexistent, or rather low, helical contents. oy-Globulin and fi-lactoglobulin havelong been suspected to contain f-form or other structures. 15 f-Lactoglobulin showsone trough near 230 mu and a peak at 200 m1A;16 now its infrared spectra indicates aprevalence of the fl-form.'7 The infrared spectra of 'y-globulin, however, are stillinconclusive, but K. Nakamura of this laboratory detects a negative dichroic bandat 218 m1A and a positive one near 200 miu, suggesting the presence of the f-form.By the same criterion, he found that the light chains of 7y-globulin have more f-formthan the heavy chains. Bovine serum albumin in 60 per cent n-prol)anol, heated to100'C for 10 min, was reported to have the fl-conformation;18 we have found thatits Cotton effects show a trough near 235 m1A and a peak at 205 mA.

Unlike the a-helix and coiled form, the various fl-forms (intermolecular paralleland antiparallel pleated sheets and intramolecular cross-f) may all have theircharacteristic Cotton effects. Neither are we certain how the extent of aggregationwill affect the ORD and CD of the f-form. The chain lengths of the segments inthe case of the cross-fl-form can also become a problem. So far all the experimentalevidence seems to suggest one type of Cotton effect for the fl-form; but recent theo-retical analysis predicts that the rotational strengths of the parallel and antiparallelfl-forms are opposite in sign and, further, that their magnitudes are smaller thanthose calculated for the a-helix (R. W. Woody, private communication). In the

VOL. 55, 1966 BIOCHEMISTRY: IIZUKA AND YANG 1181

absence of a detailed theory, we may tentatively identify the negative 218-mACD band with an n-7r* transition, and the positive 197-miA band with one of the'X-7r* transitions.As regards silk fibroin in solution, the "jumping cracker" (or "Chinese cracker")

model of Astbury et al.,'9 in which the polypeptide chains are transversely foldedwith glycine-containing bends, could very likely be partly present, since this proteinhas a high percentage of glycine and alanine residues. The formation of additionalhydrogen bonds between adjacent bends can also explain the aggregation of the silkfibroin molecules, which, however, does not significantly change the Cotton effectsof the (3-form.3While this work was in progress, P. K. Sarkar (on leave from University of

California, San Francisco) began to study the 3-form of poly-L-lysine20 in ProfessorP. Doty's laboratory.2' Similar work was carried out by Davidson et al.22 TheORD and CD of poly-L-lysine are very similar to those of silk fibroin.Summary.-Silk fibroin undergoes a coil-to-j transition when dioxane or methanol

is added to its aqueous solution. The Cotton effects of the ,8-form show (1) a 229-230-mu trough, a 205-mA peak, and another trough near 190 mjA; (2) a negativedichroic band at 218 miA, and another larger positive band at 197 m/A. The magni-tude of these Cotton effects is much smaller than that of the a-helix Cotton effects;thus, the (-form will be overshadowed by the a-helix if both conformations coexistin a protein molecule. The bo of the M~offitt equation for the d-form is close to zero;this, if substantiated by future investigations, suggests that the estimation of helicalcontent by the bo method will not be altered significantly in the presence of the (3-form.

The authors thank Professors J. C. Craig and L. A. Strait for the use of their Jasco ORD/UV-5with CD attachment and Perkin-Elmer 337 spectrophotometer, and Dr. K. Nakamura for hisvaluable help in interpreting the infrared spectra. They are indebted to Professor M. F. Moralesand Dr. L. Peller for critical comments.

Abbreviations used in this work: ORD, optical rotatory dispersion; CD, circular dichroism;Ri, rotational strength.

* This work was aided by grants from the U.S. Public Health Service (GM-K3-3441, GM-10880, HE-06285). Presented at the 50th Annual Meeting of the Federation of American Societiesfor Experimental Biology, Atlantic City, April 1966 [Federation Proc., 25, 411 (1966)].

t Present address: The Sericultural Experiment Station, Ministry of Agriculture and Forestry,Tokyo, Japan.

1 Yang, J. T., and P. Doty, J. Am. Chem. Soc., 79, 761 (1957).2 Moscowitz, A., in Optical Rotatory Dispersion, ed. C. Djerassi (New York: McGraw-Hill,

1960), chap. 12.3 Iizuka, E., and J. T. Yang, to be published.4Blout, E. R., I. Schmier, and N. S. Simmons, J. Am. Chem. Soc., 84, 3193 (1962).6 izuka, E., and J. T. Yang, Biochemistry, 4, 1249 (1965).6 Holzwarth, G., and P. Doty, J. Am. Chem. Soc., 87, 218 (1965).7 Miyazawa, T., and E. R. Blout, J. Am. Chem. Soc., 83, 712 (1961).8Miyazawa, T., Y. Masuda, and K. Fukushima, J. Polymer Sci., 62, S62 (1962).'Wada, A., M. Tsuboi, and E. Konishi, J. Phys. Chem., 65, 1119 (1961).'0Moffitt, W., and J. T. Yang, these PROCEEDINGS, 42, 596 (1956).11 Iizuka, E., and J. T. Yang, Biochemistry, 3, 1519 (1964).12 Yang, J. T., these PROCEEDINGS, 53, 438 (1965).13 Phillips, D. C., lecture presented at the IOPAB Symposium on Some Biological Systems at

the Molecular Level, Naples, September 1965.

1182 BIOCHEMISTRY: KLEINSMITH ET AL. PROc. N. A. S.

14 Jirgensons, B., J. Biol. Chem., 240, 1064 (1965); ibid., 241, 147 (1966).15 Yang, J. T., Tetrahedron, 13, 143 (1961); in Polyamino Acids, Polypeptides, and Proteins, ed.

M. A. Stahmann (Madison, Wisconsin: Univ. of Wisconsin Press, 1962), p. 226.16 Timasheff, S. N., and R. Townend, Biochem. Biophys. Res. Commun., 20, 360 (1965).

71 Timasheff, S. N., and H. Susi, J. Biol. Chem., 241, 249 (1966).8 Imahori, K., Biochim. Biophys. Acta, 37, 336 (1960).19 Astbury, W. T., E. Beighton, and K. D. Parker, Biochim. Biophys. Acta, 35, 17 (1959).20 Rosenheck, K., and P. Doty, these PROCEEDINGS, 47, 1775 (1961).21 Sarkar, P. K., and P. Doty, these PROCEEDINGS, 55, 981 (1966).22 Davidson, B., N. Tooney, and G. D. Fasman, Biochem. Biophys. Res. Commun., in press.

PHOSPHOPROTEIN METABOLISM IN ISOLATED LYMPHOCYTENUCLEI*

BY LEWIS J. KLEINSMITH, VINCENT G. ALLFREY, AND ALFRED E. MIRSKY

THE ROCKEFELLER UNIVERSITY

Communicated March 16, 1966

The experiments to be described are concerned with the synthesis and functionof phosphorylated proteins in cell nuclei, and deal specifically with the nature ofphosphate incorporation and "exchange" on serine and threonine residues in nuclearproteins.The existence of phosphoprotein fractions which rapidly incorporate P32-phos-

phate has been known in a variety of cell types for some time. 1-4 Recently, the occur-rence and formation of phosphoproteins in the cell nucleus have come under intensiveinvestigation, notably in the work of T. A. Langan and F. Lipmann.6 They havepresented convincing evidence for the nuclear localization of a protein fraction con-taining 1.0-1.2 per cent phosphorus, mainly in the form of phosphoserine.The presence of highly phosphorylated proteins in the cell nucleus is of special

interest because it raises the possibility that regions of high negative charge densityin phosphoproteins may modify DNA-histone interactions and perhaps influencethe template activity of the chromatin in RNA synthesis. Two lines of evidencesupport this view: (1) recent experiments by Langan and Smith have shown thatphosphoproteins can interact with histones in vitro, and that complex formationwith phosphoprotein diminishes the inhibitory effects of added histones on DNA-dependent RNA synthesis ;6 and (2) several analyses by T. A. Langan of chromatinfractions derived from thymocyte nuclei by the method of Frenster, Allfrey, andMirsky7 indicate that "phosphoprotein" concentrations in chromatin fractionswhich are active in RNA synthesis greatly exceed those observed in chromatinfractions which are relatively inactive in RNA synthesis.6 Thus, phosphoproteinsappear not only to be localized in chromatin but also to be preferentially bound toits "diffuse" or "active" state. 8

In order to obtain further information on the behavior of nuclear phosphopro-teins, we have employed tracer techniques to study the pathways of phosphate in-corporation, the nature of the linkage between phosphate and protein, and themetabolic stability of the phosphate previously incorporated. It will be shown