Isolation and characterization of some of the proteins that shuttle between cytoplasm and nucleus in Amoeba proteus

The Rapidly Migrating Proteins (RMP) which shuttle nonrandomly between nucleus and cytoplasm and equilibrate in approximately equal amounts in each compartment, were isolated from Amoeba proteus by implanting ³H-protein containing nuclei into unlabeled cells and some time later extracting the labeled material from the cytoplasms of such cells. The labeled material was subsequently fractionated by gel filtration in Sephadex G-100 columns. The RMP are soluble in dilute salt solutions and appear as a heterogenous group of molecules, one component of which seems to be a single species of protein accounting for ca. one-third of the RMP fraction. Because of its distinctness this component, called the LR fraction, received the major attention in this study. LR was found to comprise ca. 17% of the aqueous-soluble proteins of the nucleus and ca. 3–4% of the total cell protein. LR has a very low molecular weight as determined, e.g., by its elution from a Sephadex G-100 column. Because of its low molecular weight, LR could be purified by taking advantage of the fact that LR is (1) soluble in a saturated Solution of ammonium sulfate and (2) insoluble in butanol, diethyl ether, and 10% trichloroacetic acid. LR migrates toward the anode as a single band when subjected to electrophoresis on “standard disc” and SDS polyacrylamide gels. It does not enter a gel designed to separate basic proteins (at pH 4.0). When subjected to Sephadex G-25 gel filtration LR migrates through the gel as a single band and elutes from the gel at a position in the middle of the linear separation range that indicates its molecular weight is ca. 2300. The only N-terminal amino acid found in the LR fraction is proline. Evidence is presented to show that LR is not the product of a non-specific breakdown of protein produced during its isolation, but the possibility that it results from the cleavage of a single chemical bond of a larger polypeptide, has not been eliminated. When injected into non-labeled amebae, purified radioactive LR concentrates in the nucleus — just as radioactive RMP concentrates in a recipient cell nucleus when an amino acid-labeled nucleus is implanted into an unlabeled cell.     

Proteins in nucleocytoplasmic interactions. I. The fundamental characteristics of the rapidly migrating proteins and the slow turnover proteins of the Amoeba proteus nucleus

By the transplantation of amino acid-³H-labeled nuclei between cells and the subsequent isolation of nuclei for quantitative assay, we have confirmed that all the nuclear proteins of Amoeba proteus are divisible into two classes that are sharply defined by their physiological behavior. About 40% of the proteins in the nucleus rapidly migrates back and forth between the nucleus and the cytoplasm. These rapidly migrating proteins (RMP) are 25–50 times more concentrated in the nucleus than in the cytoplasm, and migration into the nucleus therefore occurs against a high concentration differential. The remaining 60% of nuclear proteins has been classified as slow turnover proteins (STP) since (as reported in a following paper) virtually all of them ultimately undergo turnover. Turnover in this context means loss of label from the nucleus, by either protein breakdown or protein migration to the cytoplasm. Isolation of nuclei in the detergent Triton X-100 results in a 20% loss of nuclear proteins but conclusions about RMP and STP were not found to be significantly affected by this loss.     

Proteins in nucleocytoplasmic interactions. 3. Redistributions of nuclear proteins during and following mitosis in Amoeba proteus

The behavior of nuclear proteins in Amoeba proteus was studied by tritiated amino acid labeling, nuclear transplantation, and cytoplasmic amputation. During prophase at least 77% (but probably over 95%) of the nuclear proteins is released to the cytoplasm. These same proteins return to the nucleus within the first 3 hr of interphase. When cytoplasm is amputated from an ameba in mitosis (shen the nuclear proteins are in the cytoplasm), the resultant daughter nuclei are depleted in the labeled nuclear proteins. The degree of depletion is less than proportional to the amount of cytoplasm removed because a portion of rapidly migrating protein (a nuclear protein that is normally shuttling between nucleus and cytoplasm and is thus also present in the cytoplasm) which would normally remain in the cytoplasm is taken up by the reconstituting daughter nuclei. Cytoplasmic fragments cut from mitotic cells are enriched in both major classes of nuclear proteins, i.e. rapidly migrating protein and slow turn-over protein. An interphase nucleus implanted into such an enucleated cell acquires from the cytoplasm essentially all of the excess nuclear proteins of both classes. The data indicate that there is a lack of binding sites in the cytoplasm for the rapidly migrating nuclear protein. The quantitative aspects of the distribution of rapidly migrating protein between the nucleus and the cytoplasm indicate that the distribution is governed primarily by factors within the nucleus.     

THE CYTONUCLEOPROTEINS OF AMEBAE : I. Some Chemical Properties and Intracellular Distribution

Autoradiographs of whole Amoeba proteus host cells fixed after the implantation of single nuclei from A. proteus donors labeled with any one of 8 different radioactive amino acids showed that the label had become highly concentrated in the host cell nucleus as well as in the donor nucleus and that the cytoplasmic activity was relatively low. When these amebae were sectioned, the radioactivity was found to be homogeneously distributed throughout the nuclei. The effect of unlabeled amino acid "chaser," the solubility of the labeled material, and the long-term behavior of the labeled material gave evidence that the radioactivity was in protein. At equilibrium, the host cell nucleus contained approximately 30 per cent of the radioactivity distributed between the two nuclei. This unequal nuclear distribution is attributed to the presence of two classes of nuclear proteins: a non-migratory one that does not leave the nucleus during interphase, and a migratory one, called cytonucleoprotein, that shuttles between nucleus and cytoplasm in a non-random manner. It is estimated that between 12 per cent and 44 per cent of the cytonucleoproteins are present in the cytoplasm of a binucleate cell at any one moment. Nuclei of Chaos chaos host cells also concentrated label acquired from implanted radioactive A. proteus nuclei.     

THE CYTONUCLEOPROTEINS OF AMEBAE : II. Some Aspects of Cytonucleoprotein Behavior and Synthesis

Radioactivity, apparently in cytonucleoproteins, from an amino acid-labeled nucleus implanted into a non-radioactive cell appeared in the host nucleus within 10 minutes, and the typical equilibrium ratio 70:30 donor nucleus radioactivity:host nucleus radioactivity was reached in 4 to 5 hours at 25°C. If such binucleates grew and divided, no localization of radioactivity was observable in cells fixed during mitosis, but the protein label remained concentrated in the daughter interphase nuclei for at least 4 generations. Continued migration of cytonucleoproteins was observed if these daughter nuclei were transplanted to other unlabeled cells. The Q₁₀ (19° to 29°C) of the migration rate of radioactive cytonucleoproteins was ca. 1.3, suggesting that passage through the cytoplasm occurred by diffusion. Both non-migratory nuclear proteins and cytonucleoproteins appear to be synthesized in the cytoplasm.     

STUDIES ON THE ORIGIN OF RIBOSOMES IN AMOEBA PROTEUS

The origin of cytoplasmic RNA and ribosomes was studied in Amoeba proteus by transplantation of a radioactive nucleus into an unlabeled cell followed by examination of the cytoplasm of the recipient for the presence of label. When a RNA-labeled nucleus was used, label appeared in the ribosomes, ribosomal RNA, and soluble RNA. Since the kinetics of appearance of labeled RNA indicates that the nucleus was not injured during the transfer, and since the transferred nuclear pool of labeled acid-soluble RNA precursors is inadequate to account for the amount of cytoplasmic RNA label, it is concluded that cytoplasmic ribosomal RNA is derived from acid-insoluble nuclear RNA and is probably transported as an intact molecule. Likewise, cytoplasmic soluble RNA probably originated in the nucleus, although labeling by terminal exchange in the cytoplasm is also possible. The results were completely different when a protein-labeled nucleus was grafted into an unlabeled host. In this case, label was found only in soluble proteins in the host cell cytoplasm, and there were no (or very few) radioactive ribosomes. This suggests that the nuclear pool of ribosomal protein and ribosomal protein precursors is relatively small and perhaps nonexistent (and, furthermore, shows that there was no cytoplasmic ribosomal contamination of the transferred nucleus).     

Distribution of proteins between nucleus and cytoplasm of amoeba proteus

By transplanting nuclei between labeled and unlabeled cells, we determined the localization of the major proteins of amebas and described certain features of their intracellular distributon. We identified approximately 130 cellular proteins by fluorography of one-dimensional polyacrylamide electrophoretic gels and found that slightly less than half of them (designated NP, for nuclear proteins) are almost exclusively nuclear. About 95 percent of the other proteins (designated CP for cytoplamsic proteins) are roughly equally concentrated in nucleus and cytoplasm, but—because the cytoplasm is 50 times larger than the nucleus—about 98 percent of each of the latter is in the cytoplasm. Of the CP, roughly 5 percent are not detectable in the nucleus. Assuming that these are restricted to the cytoplasm only because, for example, they are in structures too large to enter the nucleus and labeled CP readily exit a nucleus introduced into unlabeled cytoplasm, we conclude that the nuclear envelope does not limit the movement of any nonstructural cellular protein in either direction between the two compartments. Some NP are not found in the cytoplasm (although ostensibly synthesized there) presumably because of preferential binding within the nucleus. Almost one half of the protein mass in nuclei in vivo is CP and apparently only proteins of that group are lost from nuclei when cells are lysed. Thus, while an extracellular environment allows CP to exit isolated nuclei, the nuclear binding affinities for NP are retained. Further examination of NP distribution shows that many NP species are, in fact, detectable in the cytoplasm (although at only about 1/300 the nuclear concentration), apparently because the nuclear affinity is relatively low. These proteins are electrophoretically distinguishable from the high-affinity NP not found in the cytoplasm. New experiments show that an earlier suggestion that the nuclear transplantation operation causes an artifactual release of NP to the cytoplasm is largely incorrect. Moreover, we show that cytoplasmic “contamination” of nuclear preparations is not a factor in classifying proteins by these nuclear transplantation experiments. We speculate the no mechanism has evolved to confine most CP to the cytoplasm (where they presumably function exclusively) because the cytoplasm’s large volume ensures that CP will be abundant there. Extending Bonner’s idea of “quasi-functional nuclear binding sites” for NP, we suggest that a subset of NP usually have a low affinity for available intranuclear sites because their main function(s) occurs at other intranuclear sites to which they bind tightly only when particular metabolic conditions demand. The other NP (those completely absent from cytoplasm) presumable always are bound with high affinity at their primary functional sites.     

Intracellular Amounts of Nucleic Acids and Protein During Pollen Grain Growth in Tradescantia

The rapid growth, large organelles, and synchronous development of T. paludosa pollen grains make them ideal subjects for cytochemical analysis. A microphotometric study of the nucleoli, chromosomes, and cytoplasm fixed at daily intervals during pollen grain maturation indicated that: 1. DNA (Feulgen) synthesis in the generative nucleus occurred during the first third of interphase, while the DNA content of the vegetative nucleus remained unchanged. 2. Throughout development, changes in RNA (azure B) content, in general, paralleled changes in protein (NYS¹, Millon) content in each organelle of the vegetative cell. Initially, the RNA and protein of all organelles increased up to mid interphase, when chromosomal and nucleolar fractions began to decline despite a continued increase in cytoplasmic RNA and protein. At least 24 hours before anthesis, the vegetative nucleolus had disappeared and chromosomal protein and RNA of the vegetative nucleus were apparently in rapid decline. Such a system offered an opportunity to study the role of the nucleus, especially the nucleolus, in RNA and protein metabolism in the cytoplasm, by noting what cytoplasmic processes could and could not continue at a time when nuclear mechanisms were absent or minimal. It was found that at least 2 fundamental processes continued during this period: both RNA and protein accumulated in the cytoplasm at a rapid rate. It was concluded that the nucleus is not the sole source of cytoplasmic RNA, for the data suggest that there are at least 2 separate and independent, or remotely dependent synthesizing systems, one nuclear and the other cytoplasmic. It is evident that nuclear influence on cytoplasmic synthesis need be neither direct nor immediate.     

Stable nuclear RNA returns to post-division nuclei following release to the cytoplasm during mitosis

When nuclei from ³H-RNA-containing amebae (A. proteus), chased for as many as 8 cell generations, are implanted into unlabeled enucleate cells, the nuclei retain 30% or more of the cellular ³H-RNA (or at least 15 times the cytoplasmic concentration of ³H-RNA). After such cells divide, the daughter nuclei retain approximately the same proportion of total cellular ³H-RNA—although all (or almost all) of the nuclear RNA is liberated to the cytoplasm during mitosis. Thus, we conclude that RNA stably associated with the interphase nucleus has a particular affinity for the nucleus despite the fact it is in the cytoplasm when the chromosomes are condensed and the nuclear envelope is not intact.     

Human Regulatory Subunit RIβ of cAMP-Dependent Protein Kinases: Expression, Holoenzyme Formation and Microinjection into Living Cells

The human regulatory subunit RIβ of cAMP-dependent protein kinases was expressed in Escherichia coli as a fusion protein with glutathione S -transferase. Purification was performed by affinity chromatography on glutathione-agarose beads after cleavage with thrombin. The human recombinant Riff protein migrated at 55 kDa on SDS-PAGE and displayed immunoreactivity with an anti-human RIβ antiserum. Furthermore, the purified recombinant RIβ protein was shown to exist as a dimer that was able to form holoenzyme with the catalytic subunit Cα. The rate of RIβ₂Cα₂ holoenzyme formation was faster in the presence than in the absence of MgATP. The kinase activity measured before and after adding cAMP to the holoenzyme showed that the presence of cAMP resulted in holoenzyme dissociation and release of active Cα-subunit, due to cAMP binding to RIβ. Compared to a RIα₂Cα₂ holoenzyme, the RIβ₂Cα₂ holoenzyme exhibited a more than twofold higher sensitivity to cAMP. The subcellular localization of Riff was analyzed in quiescent REF-52 fibroblasts and Wistar rat thyroid (WRT) cells after microinjection of fluorescently labeled proteins into the cytoplasm. A cytoplasmic distribution was observed when free RIβ was injected, whereas free Cα injected into the cytoplasm appeared in the nucleus. When holoenzymes with labeled Riff and unlabeled Cα, or unlabeled RIβ and labeled Cα, were injected, unstimulated cells showed fluorescence in the cytoplasm of both cell types. REF-52 cells stimulated with 8-bromo-cAMP (8-Br-cAMP) and WRT cells treated with thyrotropin (TSH) showed fluorescence mainly in the cytoplasm when RIβ was the labeled subunit of the in vivo dissociated bioenzyme. In contrast, nuclear fluorescence was evident from the release and translocation of labeled Cα from the holoenzyme complex after stimulation with 8-Br-cAMP or TSH.     

Chicken erythrocyte membranes: comparison of nuclear and plasma membranes from adults and embryos

These experiments were designed to determine whether the migration of RNA molecules from an implanted nucleus to the host cytoplasm and from there into the host cell nucleus against a concentration gradient might reflect an artefact induced by the process of nuclear transplantation. That is, are RNA molecules, as previously shown for certain nuclear proteins, caused to artefactually leave a manipulated nucleus and then move into the host cell nucleus (as well as return to the grafted nucleus) during the recovery period?A variety of experiments involving different kinds of manipulative sequences and different numbers of nuclear transplantations suggest—but do not prove—that no artefact is involved in the migration of RNA from one nucleus to another but two experiments strongly support the view that the shuttling activity is a normal physiological process. One of the latter involved a determination of the rate of egress of ³H-RNA from an implanted nucleus and reveals that that rate, in contrast with the equivalent rate of egress for labeled proteins which is clearly abnormal after micromanipulation, is totally consonant with the rate of movement of RNA from nucleus to cytoplasm established from experiments that do not involve micromanipulation. The other experiment involves comparison of (1) the amount of radioactivity acquired by an unlabeled nucleus present in the cell at the time a labeled nucleus is implanted with (2) the amount of radioactivity acquired by an unlabeled nucleus implanted after a labeled nucleus had been implanted and had time to recover from any possible operation-induced trauma. With ³H-protein nuclei the host nuclei of (1) acquired much more label than the host nuclei of (2) because in (1) the host nuclei were able to acquire much of the artefactually-released ³H-protein. For the ³H-RNA experiments, however, little difference was found between (1) and (2) in the amount of label acquired by the host cell nuclei. It can be concluded that little, if any, of the non-random shuttling activity of RNA molecules can be a reflection of an artefact.     

The presence of actin in nuclei: a critical appraisal.

To assess the significance of actin associations with nuclei, we have examined Amoeba proteus nuclei for the presence of labeled actin under a variety of circumstances without (in most instances) isolating nuclei or breaking up cytoplasms prior to the extraction of proteins.We first established that: the 42,000 dalton proteins (presumed to be actin) present in cytoplasm and non-isolated nuclei are identical electrophoretically; the putative actin of amebas has the same size and almost the same isoelectric point as rat muscle actin; and the peptide “fingerprints” of putative ameba actin and rat actin are very similar after tryptic digestion. We therefore concluded that the 42,000 dalton protein of ameba is actin.We determined that: the concentrations of actin in the cytoplasm and nucleus of amebas are the same; actin is readily lost from nuclei that are released from lysed cells; shortly after a ³⁵S-labeled nucleus is transplanted into unlabeled cytoplasm, or an unlabeled nucleus is transplanted into ³⁵S-labeled cytoplasm, the concentration of ³⁵S-actin in nucleus and cytoplasm is the same; and when cells containing ³⁵S-actin are subjected to long chase periods on unlabeled food, the concentrations of ³⁵S-actin in nucleus and cytoplasm fall in parallel. These observations taken together suggest that actin is not tightly associated with nuclei. Rather, actin may associate with nuclei for the trivial reason that the nuclear envelope is no barrier to free movement of that protein between the two compartments.We conclude that the mere presence of actin in nuclei is insufficient grounds for assuming that it has any role in nuclear functions, such as, for example, chromosome condensation.     

Cse1p Is Involved in Export of Yeast Importin α from the Nucleus

Proteins bearing a nuclear localization signal (NLS) are targeted to the nucleus by the heterodimeric transporter importin. Importin α binds to the NLS and to importin β, which carries it through the nuclear pore complex (NPC). Importin disassembles in the nucleus, evidently by binding of RanGTP to importin β. The importin subunits are exported separately. We investigated the role of Cse1p, the Saccharomyces cerevisiae homologue of human CAS, in nuclear export of Srp1p (yeast importin α). Cse1p is located predominantly in the nucleus but also is present in the cytoplasm and at the NPC. We analyzed the in vivo localization of the importin subunits fused to the green fluorescent protein in wild-type and cse1-1 mutant cells. Srp1p but not importin β accumulated in nuclei of cse1-1 mutants, which are defective in NLS import but not defective in NLS-independent import pathways. Purified Cse1p binds with high affinity to Srp1p only in the presence of RanGTP. The complex is dissociated by the cytoplasmic RanGTP-binding protein Yrb1p. Combined with the in vivo results, this suggests that a complex containing Srp1p, Cse1p, and RanGTP is exported from the nucleus and is subsequently disassembled in the cytoplasm by Yrb1p. The formation of the trimeric Srp1p-Cse1p-RanGTP complex is inhibited by NLS peptides, indicating that only NLS-free Srp1p will be exported to the cytoplasm.     

The fate and origin of the nuclear envelope during and after mitosis in Amoeba proteus. I. Synthesis and behavior of phospholipids of the nuclear envelope during the cell life cycle

The synthesis and behavior of Amoeba proteus nuclear envelope (NE) phospholipids were studied. Most NE phospholipid synthesis occurs during G2 and little during mitosis or S. (A. proteus has no G1 phase). Autoradiographic observations after implantation of [3-H] choline nuclei into unlabeled cells reveal little turnover of NE phospholipid during interphase but during mitosis all the label is dispersed through the cytoplasm. Beginning at telophase all the label is dispersed through the cytoplasm. Beginning at telophase all the NE phospholipid label returns to the daughter NEs. This observation, along with the finding that no NE phospholipid synthesis occurs during mitosis or S, indicates that no de novo NE phospholipid production is required for newly forming NEs. Similarlyemetine, at concentrations that inhibit 97 percent of protein synthesis, does not prevent the post mitotic formation of NEs, suggesting that previously manufactured proteins are used in making new NEs. If a nucleus containing labeled NE phospholipids is transplanted into an unlabeled nucleate cell and the cell is allowed to grow and divide, the resultant four nuclei are equally labeled. This finding supports, but does not prove (see next paragraph), the conclusion that there probably is no continuity of the A. proteus NE during mitosis. When a phospholipid-labeled nucleus is implanted into a cell in mitosis, the grafted nucleus is not induced to enter mitosis. There is, however, a marked increase in the turnover of that nucleus's NE phospholipids with no apparent breakdown of the NE; this indicated that the mitotic cytoplasm possesses a factor that stimulates NE phospholipid exchange with the cytoplasm. That enhanced turnover is not accompanied by visible structural alteration makes less certain the earlier conclusion that no NE continuity exists during mitosis. Perhaps the most important finding in this study is that there are present, at restricted times in the cell cycle, factors capable of inducing accelerated exchange of structural components without microscopically detectable disruptions of structure.     

Dissociable and non-dissociable cytoplasmic protein-thyroid hormone interactions

Dog kidney cytosol contains a high molecular weight (50 000–70 000) and a low molecular weight (approx. 6000) thyronine-binding protein. Low molecular weight cytosol thyronine-binding protein has not been previously recognized in cytoplasm. Binding of thyroxine (tetraiodothyronine, T₄) by the low molecular weight protein has a half-time of association of more than 24 h and accounts for 32% of bound cytoplasmic tetraiodothyronine after 48 h of incubation. Binding of labeled tetraiodothyronine and triiodothyronine by this moiety is non-dissociable in the presence of 1 · 10^?5 M unlabeled tetra- or triiodothyronine. The low molecular weight protein exists in a dispersed and apparently aggregated form; the latter elutes in the void volume on Sephadex G-100 and its generation is minimized by 2 mM Ca²⁺. This binding protein elutes in a fraction which has a high A_260nm : A_280nm ratio, is pentose enriched (orcinol method) and which, because of these characteristics and low susceptibility to digestion by nuclease, is postulated to be a ribosylated cytoplasmic protein or polypeptide.Binding of tetra- and triiodothyronine by the high molecular weight protein has a half-time of association of 2 h and is saturable. Displacement of labeled triiodothyronine from this cytosol thyronine-binding protein is more readily effected with excess unlabeled tetra- than with triiodothyronine, indicating the absence of a triiodothyronine-specific cytosol thyronine-binding protein site. 3,3′,5′-Triiodothyronine (reverse triiodothyronine) is bound with low avidity. Uptake of high molecular weight protein by isolated kidney cell nuclei cannot be demonstrated.Binding of tetraiodothyronine by cytosol proteins is independent of pH in the pH range 6.8–8.9, but binding of triiodothyronine is minimized at pH 7.4 and enhanced at alkaline pH to the point of equivalency of tetra- and triiodothyronine binding at pH 8.9.At concentrations of tetraiodothyronine calculated to exist intracellularly, essentially all soluble fraction tetraiodothyronine is bound to cytosol thyronine-binding protein, restricting access of this iodothyronine to binding sites in nucleus and mitochondria. Cytosol removes labeled tetra- and triiodothyronine previously reacted in vitro with isolated cell nuclei; such removal is a linear function of cytosol protein concentration and is blocked by saturation of cytosol thyronine-binding protein with unlabeled iodothyronines. Only the high molecular weight protein accounts for unbinding by cytosol of nuclear hormone.     

Nuclear protein synthesis in animal and vegetal hemispheres of Xenopus oocytes

Experiments were conducted to determine if nuclear proteins are preferentially synthesized in the vicinity of the nucleus, a factor which could facilitate nucleocytoplasmic exchange. Using Xenopus oocytes, animal and vegetal hemispheres were separated by bisecting the cells in paraffin oil. It was initially established that protein synthesis is not affected by the bisecting procedure. To determine if nuclear protein synthesis is restricted to the animal hemisphere (which contains the nucleus), vegetal halves and enucleated animal halves were injected with [3H]leucine and incubated in oil for 90 min. The labeled cell halves were then fused with unlabeled, nucleated animal hemispheres that had been previously injected with puromycin in amounts sufficient to prevent further protein synthesis. Thus, labeled polypeptides which subsequently entered the nuclei were synthesized before fusion. Three hours after fusion, the nuclei were isolated, run on two-dimensional gels, and fluorographed. Approximately 200 labeled nuclear polypeptides were compared, and only 2 were synthesized in significantly different amounts in the animal and vegetal hemispheres. The results indicate that nuclear protein synthesis is not restricted to the cytoplasm adjacent to the nucleus.     

EVIDENCE FOR CYTOPLASMIC DNA IN ROOT CELLS OF NICOTIANA

Sterile root cultures from Nicotiana tabacum were grown with H³-thymidine added to the medium for various intervals. Incorporation of the labeled nucleoside into nuclear DNA occurred in a fraction of the nuclei which increased with time. In addition, the cytoplasm of all cells incorporated enough tritium to be readily detected by autoradiography. The tritium was not removed by hydrolysis in 1 N HCl at 60°C for 10 minutes, but was removed by digestion in a DNase solution which also removed nuclear DNA. The amount of tritium in the cytoplasm increased during the first 2 hours, but did not appear to increase significantly during the following 5 hours. If the roots were transferred to unlabeled medium after 2 hours, the label was diluted faster than expected by growth without turnover of the labeled component. If FUdR was added to the unlabeled medium, the depletion occurred faster during the first 6 hours, but later appeared to level off so that at 10 hours these cultures did not differ from those incubated without FUdR. However, the addition of an excess of unlabeled carrier had no effect on the rate of depletion of the cytoplasmic label. Actinomycin D, which inhibited the incorporation of H³-cytidine into RNA in the root tips, had no effect on the incorporation of H³-thymidine into the cytoplasmic component. However, Mitomycin C or a high concentration of deoxyadenosine inhibited the incorporation of H³-thymidine into the cytoplasmic component as well as into the nuclear DNA. It is concluded that H³-thymidine is incorporated into a cytoplasmic fraction which has the characteristics of DNA, with a measurable rate of turnover. This fraction is synthesized regardless of whether or not the nucleus is synthesizing DNA. Although the function of cytoplasmic fraction is not yet known, it does not appear to be that of supplying precursors for the synthesis of the nuclear DNA.