In Vivo Thiophosphorylation of Chromosomal Proteins RECOVERY AND ANALYSIS OF HeLa HISTONES AND DERIVATIVE PHOSPHOPEPTIDES*

HeLa cells are shown to incorporate [36S]thiophos-phate (added to the medium as sodium thiophosphate) into the y-positions of ATP and GTP, and into other nucleotides. Under these conditions there is a transfer of radioactive thiophosphoryl groups to histones H1, H2A, H3, and H4. The newly thiophosphorylated chro- mosomal proteins can be recovered selectively by affinity chromatography on organomercurial-Sepharose columns. The thiophosphorylated histone H1 and its N H z - t e r m i n a l and COOH-terminal fragments were sub- jected to tryptic digestion and the sulfur-derivatized phosphopeptides were isolated by Hg-affinity chromatography prior to electrophoretic separation of differ- ent sites of modification. Thiophosphate was found to be incorporated into both serine and threonine residues of histone H1 in HeLa cells during logarithmic growth. When and were employed simultaneously as precursors, the thiophosphorylated H1 molecules retained on the mercury column also showed the presence of the [3H]threonine label. It follows that newly synthesized H1 molecules are subject to thiophosphorylation in the growing cell cultures. involves histones and nonhistone compa-nents of chromatin. Substrates for nuclear protein kinases all of the major histones of the nucleosome “core,”

The postsynthetic modification of nuclear proteins by phosphorylation involves both histones and nonhistone companents of chromatin. Substrates for nuclear protein kinases include all of the major histones of the nucleosome "core," histone H1, and its variant forms (1-8), histone H5 (9, lo), the nuclear high mobility group proteins 14 and 17 (11-14), and a multitude of other DNA-binding proteins, including major subunits of the RNA polymerases (15).
The functional significance of phosphorylation of chromosomal proteins is largely unknown but clearly complex. In the case of histones, as well as some nonhistone proteins, phosphorylation influences the DNA-binding properties of the polypeptide chain and might be expected to affect both the structure and function of the DNA template (16, 17). But even with well defined substrates, such as histone HI, the functional significance of phosphorylation cannot be simply deduced by correlations with overall levels of phosphate incorporation, because there are multiple sites of modification in different domains of the polypeptide chain. Certain phosphorylations of histone H1 correlate with gene activation (e.g. by cyclic AMP-dependent mechanisms (18-20)), while other phosphorylations are related to cell cycle progression (21-26) and to the mechanisms of chromatin compaction and assembly of higher orders of chromatin structure (27-31).
In analyzing these phenomena, it would be very helpful to have a method for the selective isolation of the recently phosphorylated protein molecules, and for the unequivocal identification of those sites which were most recently modified. One such method for the labeling of nuclear proteins in vitro (32), as well as other substrates (33-36), employs the ATP analogue, adenosine 5"0-(3-thiotriphosphate) (37) as a thiophosphoryl group donor in the appropriate kinase reaction. We have shown that thiophosphorylated histones can be selectively recovered after labeling with ATP-7-S' and cyclic AMP-dependent protein kinase by affinity chromatography on organomercurial-Sepharose columns (32). However, because nucleoside triphosphates such as ATP-y-S are not effective substrates for in vivo experiments, due to restricted permeation through the plasma membrane, what is needed is a sulfur-derivatized precursor that can readily enter living cells. We have employed a radioactive analogue of sodium orthophosphate, Na [35S]thiophosphate, and tested for its uptake and utilization by intact HeLa cells. The 35S-labeling of specific histones is reported here, together with affinity methods for the isolation of the thiophosphorylated histone molecules, sulfur-derivatized peptides, and thiophosphorylated serine and threonine residues. The entry of [35S]thiophosphate into intracellular nucleotide pools with the formation of radiolabeled ATP and GTP analogues is also demonstrated.

EXPERIMENTAL PROCEDURES
[35S]Thiophosphate and (3ZP)Orthophosphate Labeling of HeLa Cells-HeLa S-3 cells grown to log phase (3 X IO5 cells/ml) in 100-ml spinner bottles were harvested by centrifugation at 50 X g for 5 min and resuspended in 100 ml of phosphate-free medium A (Joklikmodified minimal essential medium without Na3P04 (Grand Island Biological Co.). After 45 min at 37 "C, the cells were collected by centrifugation and resuspended in 50 ml of medium A. Ten mCi of Na [35S]thiophosphate of specific activity 989 mCi/mmol, or carrierfree (32P)orthophosphate (New England Nuclear) were added and the cells were incubated for 3 h. More than 95% of the cells were viable at that time, as judged by trypan blue dye exclusion.
Labeling of ATP and GTP Pools by f5SJThiophosphate-HeLa S-3 cells (6 X IO' cells at a concentration of 6 X IO5 cells/ml) were incubated in phosphate-free medium A containing 2 mCi of sodium bicarbonate as described by Eckstein (39). Fractions (3 m l ) were monitored for ultraviolet absorption and "S activity, and those believed to contain [35S]ATP-y-S and [35S]GTP-y-S were pooled and concentrated in a rotary evaporator (to remove the buffer). Each residue was washed 3 times with methanol and further characterized by three different analytical methods: 1) chromatography on polyethyleneimine-impregnated cellulose; 2) high voltage electrophoresis, and 3) high pressure liquid chromatography. 1.) PEI-Thin Layer Chromatography-The "S-labeled nucleotide fractions separated on DEAE-Sephadex were redissolved in 1 m~ 2mercaptoethanol, and aliquots containing 2000-4000 cpm were applied to plates (20 X 20 cm) layered with polyethyleneimine-impregnated cellulose (Polygram CEL 300 PEI/UV2,; Boehringer Mannheim). Unlabeled AMP-a-S, ADP-P-S, ATP-$3, and GTP-y-S (100 pg each) were added as markers. The chromatograms were developed in 0.75 M potassium phosphate buffer, pH 3.5, for 5 h (39). The TLC plates were air-dried and the positions of the nucleotide markers were located by ultraviolet absorption. The positions of the '%-labeled nucleotides were then located by fluorography, after spraying the TLC plates with Enhance (New England Nuclear) and exposing to Kodak SB-5 preflashed x-ray film for 3-7 days.
2.) High Voltage Electrophoresis-The nucleoside thiotriphosphate spots separated by polyethyleneimine-thin layer chromatography were transferred to Eppendorf tubes, washed twice with 95% ethanol containing 1 m~ 2-mercaptoethanol, and extracted twice with 100-pl portions of 0.4 M triethylamine bicarbonate, pH 7.2. The extracts were lyophilized and redissolved in 1 m~ 2-mercaptoethanol. This step was repeated twice. The final solutions were analyzed by high voltage electrophoresis on Whatman No. 1 paper (46 x 57 cm) in 5% acetic acid/pyridine buffer, pH 3.5, for 2 h at 3000 V. Unlabeled nucleoside triphosphates (100 pg) were added as UV markers, while the positions of the %-labeled nucleotides were determined by fluorography.
3.) High Pressure Liquid Chr~matography-~~S-labeled nucleotide fractions separated by DEAE-Sephadex chromatography were analyzed on a column (3.9 mm X 30 cm) of pBondpak-NH2 (Waters Associates, Inc.). The nucleotides were eluted in 0.3 M potassium phosphate buffer, pH 3.5, at a flow rate of 3 ml/min. Unlabeled nucleoside thiotriphosphates were added as markers. Fractions (0.5 m l ) were collected for measurement of "S activity by scintillation spectrometry.
Thiophosphorylation of Newly Synthesized Histones-One hundred ml of HeLa cell suspensions (6 X IO5 cells/&) were centrifuged and the cells were resuspended in synthetic phosphate-free medium B (minimal essential medium containing phosphate-free balanced salts, vitamins, and essential amino acids (Grand Island Biological). After incubation at 37 "C for 45 min, the cells were harvested and resuspended in 100 ml of medium B containing 1.0 mM Na thiophosphate (unlabeled) and either 5 mCi of [3H]threonine of specific activity 309 mCi/mmol, or 3 mCi of [3H]serine of specific activity 11 Ci/mmol (Amersham Corp.). After 3-h incubation at 37 "C, the histones were prepared as described below.
Isolation of Nuclei-All steps were carried out at 4-10 "C. The labeled cells were collected by centrifugation and washed twice in buffer C (8 mM Na2HP04, 15 mM KHzP04,140 m~ NaC1,27 mM KCI). The washed cells were then lysed by vortexing in buffer D (0.4 M NaCI, 50 m~ MgC12, 4 mM CaC12, 0.01 M Tris-HC1, pH 7.4) containing 0.1% NP-40 (Particle Data Laboratories, Ltd., Elmhurst, IL). The nuclei were recovered by centrifugation for 10 min at 1100 X g and washed in buffer D without NP-40.
Extraction and Separation of Histones-Total histones were extracted from lo7 nuclei in 1 ml of 0.25 N HCl and precipitated in 5 volumes of cold acetone. For preparation of histone H1, the isolated nuclei were extracted successively with 1.0 and 0.5 ml of 5% HClO,. The extracts were combined and histone H1 was precipitated in 18% (w/v) trichloroacetic acid (40). Further purification of histone H1 was achieved by preparative electrophoresis in 15% polyacrylamide gels containing 0.1% sodium dodecyl sulfate (41), or by ion exchange chromatography on Bio-Rex 70 (42). Electrophoretically separated histone bands were identified in a thin section of the gel by staining with Coomassie blue dye, and the corresponding segments were excised for electrophoretic separation of the histones, using an ISCO Model 1750 Sample Concentrator. Organomercurial-Sepharose Column Chromatography-Thiophosphorylated histones were recovered by Hg-affinity chromatography as described by Sun et al. (32). Histones were first reduced in 1 ml of 0.1 M Tris-HC1, pH 8.6, containing 9 M urea and 5 mM dithiothreitol for 1 h at 40 " C . After dilution with an equal volume of H?O, the solution was brought to 18470 in trichloroacetic acid and the precipitated histones were washed 3 times with cold acidified acetone before drying in vacuo. The histones were redissolved in N2-saturated 10 mM Tris-HC1 buffer, pH 8.6, containing 2 m~ EDTA, 10 m~ NaCI, and 0.05% sodium dodecyl sulfate, and applied to AE-Gel501 columns (Bio-Rad). Adsorbed thiophosphorylated histone molecules were eluted in buffer containing 10 m~ dithiothreitol (32). In "S-labeling experiments, the radioactive histones were dialyzed extensively against water containing 0.05 m~ phenylmethylsulfonyl fluoride before lyophilization.
Separation of Thiophosphorylated Peptides from NH2-terminal and COOH-terminal Fragments of Histone H1-Histone H1 which had been 35S-thiophosphorylated in vivo was electrophoretically purified, separated by Hg-afinity chromatography, and cleaved with Nbromosuccinimide exactly as described by Hohmann et al. (43). The NHz-terminal and COOH-terminal fragments were separated by chromatography on Sephadex G-100, monitoring the elution by UV absorption at 218 nm (43) and by counting aliquots of the 0.8 ml fractions for 35S activity. The products of the N-bromosuccinimide cleavage of 35S thiophosphorylated HI were digested with trypsin according to Hohmann et al. (43) and the peptides in each digest were separated by high voltage electrophoresis on Whatman No. 1MM paper at pH 7.9. The resulting thiophosphopeptide bands were ana- Separation of Thiophosphorylated Threonine and Serine Residues-Histone H1 containing 35S-labeled thiophosphate groups (5000 cpm) was subjected to partial acid hydrolysis in 10 pl of 6 N HC1 at 98 "C for 60 min. After addition of 1 ml of H20, the solution was lyophilized, and the residue was redissolved and lyophilized repeatedly to ensure complete removal of HCI. Thiophosphoserine and thiophosphothreonine were then separated by high voltage electrophoresis (1200 V, 1 h, pH 1.9) (44). The positions of the amino acids were located by reaction with ninhydrin and compared with those of the standards: phosphoserine and phosphothreonine, and thiophos-ph~[~H]serine and thi~phospho[~H]threonine (prepared from acid hydrolysates of Hg-affinity purified histone H1 which had been labeled in vivo with the indicated amino acid and thiophosphate). These thiophosphorylated amino acids co-migrate with phosphoserine and phosphothreonine, respectively, under the electrophoretic conditions employed. Spots cut from the paper were analyzed for 35S activity by scintillation spectrometry.  Analysis of "S-labeled nucleotides by ascending thin layer chromatography on polyethyleneimine-impregnated cellulose. Fractions isolated by chromatography on DEAE-Sephadex were applied to PEI-thin layer plates with unlabeled nucleoside thiophosphates ( 5 0 pg each) added as markers. The positions of the markers were detected by UV absorption ( A ) while the positions of the %-labeled nucleotides in the various peaks were determined by fluorography (B). Note that peak I11 contains nucleotides which comigrate with ATP-y-S and GTP-y-S, while peak IV contains more of the GTP analogue. The nucleotide spots corresponding to ATP-y-S and GTP+ were recovered from the TLC plate for further analysis. detected in peak IV (Fig. 2). Peak IV also contained radioactive substances co-migrating with ATP-y-S and ADP-/3-S. These may represent the corresponding nucleotides containing more than one thiophosphoryl group, but we did not investigate this possibility further, due to the unavailability of the appropriate dithiophosphorylated nucleotide standards.

Entry of PSIThiophosphate into Intracellular Nucleotide
The identity of the 35S-labeled nucleotides was further confirmed by transfer of the spots from the TLC plates and analysis by high voltage electrophoresis. As can be seen in Fig. 3, 35S radioactivity in the spot identified as ATP-y-S in Fig. 2 co-migrated with the unlabeled ATP-y-S marker. Finally, the identity of the 35S-labeled nucleotides in the various peaks was confirmed by high pressure liquid chromatography.

In Vivo
Thiophosphorylation of Chromosomal Proteins peaks I11 and IV. It is clear that "S-labeled nucleotides eluting in the positions of authentic ATP-y-S and GTP-y-S can be identified by high pressure liquid chromatography. Thus, by a variety of analytical criteria, ["'Slthiophosphate enters the nucleotide pools of living cells in forms that should allow transfer of radioactive thiophosphoryl groups to proteins in ATP-or GTP-dependent reactions. Thiophosphorylation of Histones in Cells Exposed to YS] Thiophosphate-Nuclei were isolated from HeLa cells incubated for 3 h in media containing ["'Slthiophosphate or ("'P)orthophosphate, and the histones were extracted for electrophoretic separation and analysis. The distribution of ["' PI phosphate in the various histone fractions is shown in Fig. 5A, which demonstrates extensive phosphate uptake into histones H1, H2A, and H3, and a slight "P uptake into histone H4. The corresponding distribution of thiophosphorylated histones from cells incubated in the presence of ["S]thiophosphate is shown in Fig. 5B, which plots the 35S activity of the basic proteins a t different regions of the gel. [:"S]Thiophosphate incorporation into histones H1, H2A, H3, and H4 is clearly indicated. Note that the electrophoretic separation yields two prominent H1 bands, both of which are thiophosphorylated. However, the predominantly labeled band in [35S] thiophosphate labeling experiments is the faster moving H1 band (Fig. 5B), whereas in ["Plphosphate labeling experiments, it is the slow moving H1 band that is more radioactive (Fig. 5A). It follows that the relative rates of ['"P]phosphate uptake and turnover in individual nuclear proteins may not be accurately reflected when [""Slthiophosphate is employed as the precursor. However, the advantage of the thiophosphate label is the ability to recover selectively the recently thiophosphorylated molecules by Hg-affinity chromatography (32).
Affinity Purification of Thiophosphorylated Histone HI Molecules-For proteins which lack sulfhydryl groups, such as histone H1, the chromatographic separation of newly thiophosphorylated molecules can be achieved directly by affinity chromatography on organomercurial-Sepharose columns (32, 45). Thiophosphorylated histones are effectively retained by such columns but can be readily eluted from them with dithiothreitol (32). This procedure was used to isolate the thiophosphorylated H1 molecules from HeLa cells incubated in the presence of [3'S]thiophosphate. The histone H1 was extracted, purified electrophoretically, and then subjected to Hg-affinity chromatography on Affi-Gel 501. Fig. 6A shows the elution pattern of the "S-labeled histone, with retention and subsequent displacement of the thiophosphorylated molecules by dithiothreitol. Fig. 6B shows the results of a different experiment in which ["]threonine and nonradioactive thiophosphate were employed simultaneously as precursors. In this case, the thiophosphorylated H1 molecules retained on the Hg column also show the presence of the ['Hlthreonine label. A similar double labeling experiment, using ['Hlserine and nonradioactive thiophosphate, is shown in Fig. 6C. The presence of the newly incorporated amino acids in histone molecules which are retained by the column establishes that recently synthesized molecules are subject to thiophosphorylation in the growing cell cultures.
Recovery of Thiophosphorylated Histone HI Peptides-In a log phase HeLa culture, we would expect phosphorylation of histone H1 to take place a t multiple growth-associated sites. Langan has identified four of those sites in mammalian cells as threonines a t positions 16, 136, and 153, and serine at position 180 of the polypeptide chains (46). An additional site of phosphorylation at serine 37 is modified by a cyclic AMPdependent protein kinase (47). In mitotic cells, phosphorylations takes place on serine and threonine residues in the NHzterminal region, and on threonine residues in the COOHterminal region (43). Affinity purifkations of thiophosphorylated fragments of histone H1 should simplify the identification of which sites are subject to phosphorylation under different physiological conditions (or in different isolated enzyme systems).
We have tested the feasibility of this approach, as applied to histone H1 thiophosphorylation in log phase HeLa cells. Two methods were employed; in the first, ?3-labeled peptides were prepared from a limited tryptic digest of the entire H1 molecule; in the second, the H1 was first cleaved by treatment with N-bromosuccinimide, and peptides were prepared from the chromatographically separated NHz-terminal and COOHterminal fragments by tryptic digestion.
In the first procedure, we compared histone H1 molecules labeled in three separate cell cultures, using [35S]thiophosphate, ['zP]orthophosphate, or ["]threonine plus nonradio-  active thiophosphate, respectively, as precursors. In each case, histone H1 was purified electrophoretically, and the 35S-labeled and 3H-labeled histones were further fractionated by Hg-affinity chromatography to select the thiophosphorylated forms of the protein molecules. Each histone preparation was subjected to limited tryptic digestion and the resulting peptides were separated by electrophoresis in SDS-polyacrylamide gels. The separation of peptides derived from the affiiity purified 35S-labeled histone is shown in Fig. 7A. There are four prominent peaks of 35S activity separated by this procedure. The same peptide peaks are observed in the tryptic digest of the 32P-labeled histone (Fig. 7B), supporting the view that the thiophosphate modification can be used to probe the multiple sites of protein kinase action in uiuo.
The separation of thiophosphorylated peptides from newly synthesized H1 molecules is iUustrated in Fig. 7C. H1 was electrophoretically purified from each culture. The thiophosphorylated forms ( i n A and 0 were recovered by Hg-affinity chromatography. Each histone preparation was digested with trypsin in the presence of 100 pg of unlabeled H1, as described under "Experimental Procedures," and the fragments were separated by electrophoresis in 20% polyacrylamide-0.1% SDS slab gels at 17 mA for 6 h. After staining, each gel was sliced into 2-mm segments which were dissolved in 30% H202 and counted. Note the parallel distribution of [35S]thiophosphate-and [32P]phosphate-labeled peptides in A and B, respectively. A similar size distribution of labeled peptides is seen in the newly synthesized and thiophosphorylated histone HI molecules (C). tones were extracted from the isolated nuclei and separated by electrophoresis. Histone H1 bands were eluted from the gel and the thiophosphorylated molecules were recovered by Hg-affinity chromatography. They were subjected to limited tryptic digestion and the resulting peptide mixture was rechromatographed on the Hg-affinity column for selective recovery of the thiophosphorylated peptides. Subsequent separation of the 3H-labeled thiophosphopeptides by electrophoresis in SDS-polyacrylamide gels gave the pattern shown in Fig. 7C, with several peaks of [3H]threonine activity corresponding to the ?+labeled peptides (Fig. 7A) and the 32Plabeled peptides in Fig. 7B.
It is known that histone H1 has multiple phosphorylation sites involving serine residues (4, 6, 7 , 46). The incorporation of [35S]thiophosphate by HeLa cell cultures confirms the serine modification. The thiophosphorylated histone H1 was subjected to limited acid hydrolysis and the free amino acids were separated by high voltage electrophoresis. Both serine and threonine residues were found to be thiophosphorylated.
As judged by their respective 35S activities (62% of the total recovered counts in thiophosphothreonine and 38% of the total in thiophosphoserine), threonine residues represent the major sites of H1 phosphorylation under these conditions.
Chemical and enzymatic cleavages of thiophosphorylated histone H1 were combined to demonstrate that [35S]thiophosphate is incorporated into different amino acids in different regions of the molecule. The thiophosphorylated H1 molecules were purified by Hg-affinity chromatography and treated with N-bromosuccinimide. The resulting NH2-terminal and COOH-terminal fragments were separated chromatographically, and both were found to be labeled with [35S] thiophosphate (Fig. 8A). Each fragment was then digested with trypsin and the peptides separated by high voltage electrophoresis, as described by Hohmann et al. (43). The distribution of 35S activity in the different peptide bands is plotted in Fig. 8B. Particular attention is drawn to the 35Sthiophosphorylation of band I from the NHn-terminal fragment (indicated by the arrow in the top of Fig. 8B). This peptide band has been shown to contain only phosphoserine after 32P-labeling of Chinese hamster ovary cells (43)

DISCUSSION
To our knowledge, this is the fiist indication that inorganic thiophosphate can be used to label the nucleotide "pools" within living cells. It is known that ATP-y-S can be formed in vitro in reactions involving glyceraldehyde-3-phosphate dehydrogenase and 3-phosphoglycerate kinase (48). Isolated liver mitochondria have been reported to form a little ATPy-S from thiophosphate, but this is strongly inhibited by Pi (48). Isolated mitochondria incorporate [35S]thiophosphate in an acid-precipitable fraction which would have included proteins, but the sites of modification were not identified, and the uptake was found to be insensitive to inhibitors of oxidative phosphorylation such as aurovertin and oligomycin (49).
We have shown that formation of the thiophosphorylated analogues of ATP and GTP, as well as other nucleotides, takes place in intact cells. The nature of the products has been c o n f i i e d by DEAE-Sephadex chromatography, PEIthin layer chromatography, high voltage electrophoresis, and high pressure liquid chromatography. (Compounds identified as ATP-7-S and ADP-P-S could, of course, have more than one thiophosphoryl group, depending upon the balance between multiple nucleotide kinase reactions in the cell and their preference for phosphorylated, as compared to thiophosphorylated substrates.) One important consequence of the formation of ATP-y-S and GTP-y-S is the subsequent transfer of thiophosphoryl groups to phosphorylatable sites on histones and other phosphoproteins in uiuo. Although other mechanisms of protein modification by thiophosphate might be proposed, it has been clearly established that histones and other proteins are thiophosphorylated by ATP-y-S in kinase-mediated reactions (32-36), and we presume that this is the major route of protein thiophosphorylation in the living cell.
Because the thiophosphorylation of proteins permits their purification by affinity chromatography on organomercurial-

In Vivo Thiophosphorylation
Sepharose columns (32), the use of [j'Slthiophosphate (or nonradioactive thiophosphate) allows the selective recovery of histones and other protein molecules that have just been modified in vivo. This is an important extension of the previous method for the study of kinase-mediated protein phosphorylations in vitro (32). In both cases, the modification permits recovery of a variety of thiophosphorylated proteins. (Complications which might arise as a consequence of cysteine residues binding to the mercury column are reduced by pretreatment with iodoacetate before purification of the thiophosphorylated molecules (32)).
The present study extends the affinity purification method to the separation of thiophosphorylated peptides and thiophosphoamino acids produced by proteolytic or chemical cleavage of the modified proteins. One obvious advantage of this procedure is the elimination of many unmodified peptides. It also permits the separation of the recently modified thiophosphopeptides from similar peptides containing previously incorporated phosphate groups. Since only the sites of most recent modification are recovered, this should greatly simplify studies of site-specific phosphorylations in histone H1 (and other proteins) in response to cyclic nucleotides, hormones, divalent cations, and factors controlling progression through the cell cycle. The feasibility of analyzing the thiophosphorylations of different amino acids in different regions of the histone H1 molecule is indicated by the experiments summarized in Fig. 8. However, it is clear from the differences in rates of labeling of histone variants with [35S]thiophosphate and (3'P)orthophosphate (Fig. 5), that caution must be exercised in the interpretation of labeling differences in different phosphoproteins and in different regions of the same protein.
While this paper has emphasized the thiophosphorylation of histones in cultured HeLa cells using [35S]thiophosphate as a precursor, the procedure is not limited to cell cultures nor to components of the nucleus. Other experiments' have shown that [35S]thiophosphate is utilized for protein phosphorylation in intact animals; e.g. specific protein thiophosphorylations permit the recovery of a subset of membrane proteins from the "smooth' endoplasmic reticulum of rat liver.
Finally, the entry of thiophosphate into the nucleotide "pools" of living cells has other important consequences. Among them is the ability to label newly synthesized RNA and DNA molecules and to recover them selectively, excluding previously synthesized polynucleotides by Hg-affinity chromatography. Some applications of this technique will be described elsewhere.