The loss of phenotypic traits by differentiated cells. VI. Behavior of the progeny of a single chondrocyte.

A single, functional, mitotically quiescent chondrocyte may be induced to reenter the mitotic cyde, and produce a progeny of over 10(11) cells. Sessile, adherent, polygonal cells deposit matrix, whereas amoeboid, dispersed, flattened fibroblastic cells do not. The prior synthetic history of a cell is of greater importance in determining whether the characteristic chondrogenic phenotype will be expressed, rather than growth in "permissive" or "nonpermissive" medium. Clonal conditions select for stem-like cells, some of whose progeny may become polygonal chondrocytes. The retention of the characteristic chondrogenic phenotype in vitro is favored by pruning the dedifferentiated chondrocytes which arise in these cultures. Dedifferentiated chondrocytes interfere with the deposition and synthesis of chondroitin sulfate by neighboring functional chondrocytes. Possible mechanisms are proposed to explain this type of cell-cell or cell exudate interference. If the progeny of a single, genetically programmed chondrocyte may or may not synthesize chondroitin sulfate, then extragenic sites in the cytoplasm or cell surface must influence the decision as to which cluster of "luxur" molecules the cell will synthesize.

The loss of phenotypic traits by differentiated cells. 3. The reversible behavior of chondrocytes in primary cultures.

Observations were made on the behavior of chondrocytes grown under various conditions in vitro. The chondrocytes in 10-day embryonic chick vertebrae were grown as cultures of intact vertebrae, as pellets of chondrocytes liberated from their matrix, and as monodispersed cells plated out on plasma clots. Cartilage matrix was stained metachromatically with toluidine blue. Radioautographs were made of incorporated H(3)-thymidine, H(3)-proline, and S(35)-sulfate to determine the extent of DNA synthesis, collagen synthesis, and chondroitin sulfate synthesis, respectively. Chondrocytes in intact vertebrae or in pellets are rounded and actively synthesizing chondroitin sulfate and collagen. There is little DNA synthesis by cells in either vertebrae or pellets. Chondrocytes grown as monodisperse cells rapidly cease synthesizing cytologically detectable chondroitin sulfate and are induced to synthesize DNA and divide. There is a change in the shape of these chondrocytes from a rounded to a more stellate condition which accompanies the shift in metabolic activity. Conversely, when the cells attain a certain cell density, they reacquire a rounded shape, cease dividing, and again synthesize chondroitin sulfate. Clusters of chondrocytes synthesize more chondroitin sulfate than isolated chondrocytes. It is concluded that most chondrocytes synthesizing chondroitin sulfate do not concurrently synthesize DNA. Interaction between associated chondrocytes is important in inducing and maintaining chondroitin sulfate synthesis in genetically determined chondrocytes. Failure of interaction between chondrocytes leads to DNA synthesis and cell multiplication.

The effect of mitotic inhibitors on myogenesis in vitro.

The effect of colchicine and other antimitotic drugs was studied in cultures of 11-day chick embryo breast muscle. Exposure of such cultures to 10(-6)M colchicine results in fragmentation of the elongate myotubes into rounded, cytoplasmic sacs (myosacs) containing various numbers of nuclei. Comparison of the dose-response relation between myotube fragmentation and metaphase arrest suggests that the underlying mechanism may be similar in both cases. Low temperature does not duplicate the effects of colchicine. Glycerinated myotubes are not affected by the mitotic inhibitors. The effect of colchicine on myotubes is reversible. Myosacs elongate within several days after removal from colchicine. However, the regenerated myotubes fail to incorporate additional mononucleated cells. Colchicine does not interfere with the process of fusion itself, but the metaphase block prevents cells from entering that phase of the cell cycle during which fusion can occur. Cells arrested in mitosis by colchicine do not recover when incubated in normal medium. Colcemid-induced arrest is reversible and does not prevent subsequent fusion of the cells.

Response of myogenic and fibrogenic cells to cytochalasin B and to colcemid. I. Light microscope observations.

Cytochalasin B (CB) induces a biphasic retraction is some cell types. The rapid response that peaks in 30 min leads to the "dendritic" condition. Replicating myogenic and fibrogenic cells, as well as postmitotic myoblasts and myotubes, participate in this reaction. This is followed by a slower phase that requires 40 h for stabilization and leads to the fully "absorized" state. Only replicating myogenic and fibrogenic cells participate in this reaction. Postmitotic myoblasts and myotubes do not arborize but round up and float off into the medium. Pretreatment with Colcemid does not block the rapid response to CB, but does block arborization. CB-arborized cells exposed to Colcemid while in the presence of CB develop sufficient tension to pull themselves apart. If CB depolymerizes actin-like filaments, and if such filaments constitute the only contractile system in the cell, then it is difficult to visualize how cells in CB develop such tension. Colcemid induces twisting, birefringent bands in interphase- and metaphase-arrested myogenic and fibrogenic cells, and in postmitotic myotubes. Such bands are more evident when CB-arborized cells are removed from CB and allowed to relax in Colcemid. These birefringent bands assemble in the prescence of cycloheximide, and may constitute 20% of the volume of the cell.

Inhibition of myoblast fusion after one round of DNA synthesis in 5-bromodeoxyuridine.

The thymidine analogue 5-bromodeoxyuridine (BUdR) has a differential effect on the synthesis of tissue-specific products and molecules required for growth and division. Proliferating myogenic cells cultured in BUdR fail to fuse and fail to initiate the synthesis of contractile protein filaments. Conversely, BUdR has but a minor effect on cell viability and reproductive integrity. Low concentrations of BUdR result in an enhancement of cell number relative to the controls; higher concentrations are cytotoxic. Suppression of myogenesis is reversible after at least 10 cell generations of growth in the analogue. Cells that do not synthesize DNA, such as postmitotic myoblasts and myotubes, are not affected by BUdR. Incorporation of BUdR for one round of DNA synthesis was accomplished by first incubating myogenic cells, prior to fusion, in 5-fluorodeoxyuridine (FUdR) to block DNA synthesis and collect cells in the presynthetic phase. The cells were then allowed to synthesize either normal DNA or BU-DNA for one S period by circumventing the FUdR block with BUdR or BUdR plus thymidine (TdR). The cultures were continued in FUdR to prevent dilution of the incorporated analogue by further division. After 3 days, the cultures from the FUdR-BUdR series showed the typical BUdR effect; the cells were excessively flattened and few multinucleated myotubes formed. Cells in the control cultures were of normal morphology, and multinucleated myotubes were present. These results were confirmed in another experiment in which BUdR-(3)H was added to 2-day cultures in which myotubes were forming. Fusion of thymidine-(3)H-labeled cells begins at 8 hr after the preceding S phase. In contrast, cells which incorporate BUdR-(3)H for one S period do not fuse with normal myotubes.

Mitosis and the processes of differentiation of myogenic cells in vitro.

The relation between the mitotic cycle and myoblast fusion has been studied in chick skeletal muscle in vitro. The duration of the cell cycle phases was the same in both early and late cultures. By tracing a cohort of pulse-labeled cells, it was found that myoblast fusion does not occur in S, G(2), or M. Cell surface alterations required for fusion are dependent upon the position of the cell in the division cycle. In early cultures, fusion takes place only after a minimum delay of 5 hr from the time the cell has entered G(1). The mitosis preceding fusion may condition the cell for the abrupt shift in synthetic activity that occurs in the subsequent G(1). In older cultures fusion of labeled cells is diminished. Two factors account for the cessation of fusion in older cultures. First, the number of myogenic stem cells declines, but these cells do not disappear as the cultures mature. Their persistence was demonstrated by labeling dividing mononucleated cells in older cultures and challenging them with nascent myotubes. Some of these labeled cells were incorporated into the forming myotubes. Second, a block to fusion develops during myotube maturation. Well developed myotubes challenged with labeled competent myogenic cells failed to incorporate the labeled nuclei.

Mitosis and intermediate-sized filaments in developing skeletal muscle.

A new class of filaments intermediate in diameter between actin and myosin filaments has been demonstrated in skeletal muscle cells cultured from chick embryos. These filaments, which account for the majority of free filaments, average 100 A in diameter. They may run for more than 2 micro in a single section and can be distinguished in size and appearance from the thick and thin filaments assembled into myofibrils. The 100-A filaments are seen scattered throughout the sarcoplasm at all stages of development and show no obvious association with the myofibrils. The 100-A filaments are particularly conspicuous in myotubes fragmented by the mitotic inhibitors, colchicine and Colcemid. In addition, filaments similar in size and appearance to those found in myotubes are present in fibroblasts, chondrocytes, and proliferating mononucleated myoblasts. The 100-A filaments are present in cells arrested in metaphase by mitotic inhibitors. Definitive thick (about 150 A) or thin (about 60 A) myofilaments are not found in skeletal myogenic cells arrested in metaphase. Myogenic cells arrested in metaphase do not bind fluorescein-labeled antibody directed against myosin or actin. For these reasons, it is concluded that not all "thin" filaments in myogenic cells are uniquely associated with myogenesis.

Separation of precursor myogenic and chondrogenic cells in early limb bud mesenchyme by a monoclonal antibody.

We have addressed the problem of the segregation of cell lineages during the development of cartilage and muscle in the chick limb bud. The following experiments demonstrate that early limb buds consist of at least two independent subpopulations of committed precursor cells--those in (a) the myogenic and (b) the chondrogenic lineage--which can be physically separated. Cells obtained from stage 20, 21, and 22 limb buds were cultured for 5 h in the presence of a monoclonal antibody that was originally isolated for its ability to detach preferentially myogenic cells from extracellular matrices. The detached limb bud cells were collected and replated in normal medium. Within 2 d nearly all of the replated cells had differentiated into myoblasts and myotubes; no chondroblasts differentiated in these cultures. In contrast, the original adherent population that remained after the antibody-induced detachment of the myogenic cells differentiated largely into cartilage and was devoid of muscle. Rearing the antibody-detached cells (i.e., replicating myogenic precursors and postmitotic myoblasts) in medium known to promote chondrogenesis did not induce these cells to chondrify. Conversely, rearing the attached precursor cells (i.e., chondrogenic precursors) in medium known to promote myogenesis did not induce these cells to undergo myogenesis. The definitive mononucleated myoblasts and multinucleated myotubes were identified by muscle-specific antibodies against light meromyosin or desmin, whereas the definitive chondroblasts were identified by a monoclonal antibody against the keratan sulfate chains of the cartilage-specific sulfated proteoglycan. These findings are interpreted as supporting the lineage hypothesis in which the differentiation program of a cell is determined by means of transit through compartments of a lineage.

Formation of arrowhead complexes with heavy meromyosin in a variety of cell types.

Heavy meromyosin (HMM) forms characteristic arrowhead complexes with actin filaments in situ. These complexes are readily visualized in sectioned muscle. Following HMM treatment similar complexes appear in sectioned fibroblasts, chondrogenic cells, nerve cells, and several types of epithelial cells. Thin filaments freshly isolated from chondrogenic cells also bind HMM and form arrowhead structures in negatively stained preparations. HMM-filament complexes are prominent in the cortex of a variety of normal metaphase and Colcemid-arrested metaphase cells. There is no detectable binding of HMM with other cellular components such as microtubules, 100-A filaments, tonofilaments, membranes, nuclei, or collagen fibrils. The significance of HMM-filament binding is discussed in view of the finding that arrowhead complexes form in types of cells not usually thought to contain actin filaments.

Primitive erythropoiesis in early chick embryogenesis. I. Cell cycle kinetics and the control of cell division.

The primitive line of embryonic chick blood cells develop as a relatively homogeneous cohort of cells. Using an analysis based on the continuous uptake of thymidine-(3)H, we have established the generation time, G1, S, and G2 for progressively more mature generations of these immature erythroblasts. The data indicate that after the initiation of hemoglobin synthesis, the average cell will yield six generations of hemoglobin producing erythroblasts. The older generations of erythroblasts exhibit a longer generation time, G1, S, and G2 than the earlier generations of erythroblasts. Other methods of analysis corroborated these findings. One of these methods, an estimate of total erythrocyte productivity from the primitive stem cells (hematocytoblasts), led to the conclusion that the erythroblast cell lineage might be initiated as early as the sixth or seventh division following fertilization. In addition, primitive erythroblasts characterized by one set of cell cycle parameters, when grown in serum associated with erythroblasts of different parameters, showed no alteration in mitotic behavior. These results suggest the presence of programmed cell division not immediately cued by extracellular influence.

Some properties of embryonic myosin.

Myosins from the following sources were purified by diethylaminoethyl-Sephadex chromatography: moytubes grown in vitro for 7-8 days, prepared from pectoralis muscles of 10-day old embryos, and breast and leg muscles from 16-day old embryos. The adenosine triphosphatase activities of these myosins were close to that of adult m. pectoralis myosin. The light chains of the embryonic myosins had the same mobilities in sodium dodecyl sulfate electrophoresis as those in adult pectoralis muscle myosin and were clearly distinguishable from those in myosin from tonic muscle m. latissimus dorsi anterior. The fastest light chain in embryonic muscle myosin-apparent mol wt 16,000-was present in smaller amounts than in adult myosin. The negative staining pattern of paracrystals of embryonic light meromyosin (LMM) was indistinguishable from that of adult fast muscle LMM. The significance of these results for differentiation of various muscle types has been discussed.

DNA polymerase activity in muscle cultures.

Nuclei within myotubes do not synthesize DNA for replication. Accordingly, cultures of myotubes display low levels of DNA polymerase activity. The coincidental decline in DNA polymerase activity and increased formation of multinucleated myotubes during culture does not prove that the loss of capacity to synthesize DNA is a consequence of fusion. Tne experiments described demonstrate that myogenic cells prevented from fusing have low levels of DNA polymerase activity. This is consistent with the notion that, in myogenic cultures, there is a population of mononucleated cells, the myoblasts, which have withdrawn from the mitotic cycle before fusion.

Role of stress fiber-like structures in assembling nascent myofibrils in myosheets recovering from exposure to ethyl methanesulfonate.

When day 1 cultures of chick myogenic cells were exposed to the mutagenic alkylating agent ethyl methanesulfonate (EMS) for 3 d, 80% of the replicating cells were killed, but postmitotic myoblasts survived. The myoblasts fused to form unusual multinucleated "myosheets": extraordinarily wide, flattened structures that were devoid of myofibrils but displayed extensive, submembranous stress fiber-like structures (SFLS). Immunoblots of the myosheets indicated that the carcinogen blocked the synthesis and accumulation of the myofibrillar myosin isoforms but not that of the cytoplasmic myosin isoform. When removed from EMS, widely spaced nascent myofibrils gradually emerged in the myosheets after 3 d. Striking co-localization of fluorescent reagents that stained SFLS and those that specifically stained myofibrils was observed for the next 2 d. By both immunofluorescence and electron microscopy, individual nascent myofibrils appeared to be part of, or juxtaposed to, preexisting individual SFLS. By day 6, all SFLS had disappeared, and the definitive myofibrils were displaced from their submembranous site into the interior of the myosheet. Immunoblots from recovering myosheets demonstrated a temporal correlation between the appearance of the myofibrillar myosin isoforms and the assembly of thick filaments. The assembly of definitive myofibrils did not appear to involve desmin intermediate filaments, but a striking aggregation of sarcoplasmic reticulum elements was seen at the level of each I-Z-band. Our findings suggest that SFLS in the EMS myosheets function as early, transitory assembly sites for nascent myofibrils.

Redistribution of intermediate filament subunits during skeletal myogenesis and maturation in vitro.

The distribution of intermediate filament (IF) subunits during maturation of skeletal myotubes in vitro was examined by immunofluorescence, using antibodies against two different types of chick IF subunits: (a) 58-kdalton subunits of fibroblasts (anti-58K), and (b) 55-kdalton subunits of smooth muscle (anti-55K). Anti-58K bound to a filament network in replicating presumptive myoblasts and fibroblasts, as well as in immature myotubes. The distribution in immature myotubes was in longitudinal filaments throughout the cytoplasm. With maturation, staining of myotubes by anti-58K diminished and eventually disappeared. Anti-55K selectively stained myotubes, and the fluorescence localization underwent a drastic change in distribution with maturation--from dense, longitudinal filaments in immature myotubes to a cross-striated distribution in mature myotubes that was associated with the I--Z region of myofibrils. However, the emergence of a cross-striated anti-55K pattern did not coincide temperally with the emergence of striated myofibrils, but occurred over a period of days thereafter.

Synthesis of myosin heavy and light chains in muscle cultures.

The weight ratio of myosin/actin, the myosin heavy chain content as the percentage of total protein (wt/wt), and the kinds of myosin light chains were determined in (a) standard muscle cultures, (b) pure myotube cultures, and (c) fibroblast cultures. Cells for these cultures were obtained from the breast of 11-day chick embryos. Standard cultures contain, in addition to myotubes, large numbers of replicating mononucleated cells. By killing these replicating cells with cytosine arabinoside, pure myotube cultures were obtained. The myosin/actin ratio (wt/wt) for pure myotube, standard muscle, and fibroblast cultures average 3.1, 1.9, and 1.1 respectively. By day 7, myosin in myotube cultures represents a minimum of 7% of the total protein, but about 3% in standard cultures and less than 1.5% in fibroblasts cultures. Myosin from standard cultures contains light chain LC1, LC2, and LC3, with a relative stoichiometry of the molarity of 1.0:1.9:0.5 and mol wt of 25,000, 18,000 and 16,000 daltons, identical to those in adult fast muscle. Myosin from pure myotubes exhibits light chains LC1 and LC2, with a molar ratio of 1.5:1.6. Myosin from fibroblast cultures possesses two light chains with a stoichiometry of 1.8:1.8 and mol wt of 20,000 and 16,000 daltons. Clearly, the faster migrating light chain, LC3, found in standard cultures is synthesized not by the myotubes but ty the mononucleated cells. In myotubes, both the assembly of the sarcomeres and the interaction between thick and thin filaments required for spontaneous contraction occur in the absence of light chain LC3. One set of structural genes for the myosin light and heavy chains appears to be active in mononucleated cells, whereas another set appears to be active in multinucleated myotubes.

Desmin/vimentin intermediate filaments are dispensable for many aspects of myogenesis.

An expression vector was prepared containing a cDNA coding for a truncated version of the intermediate filament (IF) protein desmin. The encoded truncated desmin protein lacks a portion of the highly conserved alpha-helical rod region as well as the entire nonhelical carboxy-terminal domain. When transiently expressed in primary fibroblasts, or in differentiating postmitotic myoblasts and multinucleated myotubes, the truncated protein induces the complete dismantling of the preexisting vimentin or desmin/vimentin IF networks, respectively. Instead, in both cell types vimentin and desmin are packaged into hybrid spheroid bodies scattered throughout the cytoplasm. Despite the complete lack of intact IFs, myoblasts and myotubes expressing truncated desmin assemble and laterally align normal striated myofibrils and contract spontaneously in a manner indistinguishable from that of control myogenic cells. In older cultures the spheroid bodies shift from a longitudinal to a predominantly transverse orientation and loosely align along the I-Z-I-regions of striated myofibrils (Bennett, G.S., S. Fellini, Y. Toyama, and H. Holtzer. 1979. J. Cell Biol. 82:577-584), analogous to the translocation of intact desmin/vimentin IFs in control muscle. These results suggest the need for a critical reexamination of currently held concepts regarding the functions of desmin IFs during myogenesis.

Titin and myosin, but not desmin, are linked during myofibrillogenesis in postmitotic mononucleated myoblasts.

Monoclonal antibodies specific for the muscle protein titin have been used in conjunction with muscle-specific antibodies against myofibrillar myosin heavy chains (MHCs) and desmin to study myogenesis in cultured cells. Desmin synthesis is initiated in replicating presumptive myoblasts, whereas the synthesis of titin and MHC is initiated simultaneously in their progeny, the postmitotic, mononucleated myoblasts. Both titin and MHC are briefly localized to nonstriated and thereafter to definitively striated myofibrils. At no stage during myofibrillogenesis is either protein observed as part of a sequence of mini-sarcomeres. Titin antibodies bind to the A-I junction, MHC antibodies to the A bands in nascent, maturing, and mature myofibrils. In contrast, desmin remains distributed as longitudinal filaments until well after the definitive myofibrils have aligned laterally. This tight temporal and topographical linkage between titin and myosin is also observed in postmitotic, mononucleated myoblasts and multinucleated myotubes when myofibrillogenesis is perturbed with Colcemid or taxol. Colcemid induces elongating postmitotic mononucleated myoblasts and multinucleated myotubes to round up and form Colcemid myosacs. The myofibrils that emerge in these rounded cells are deployed in convoluted circles. The time required for their nonstriated myofibrils to transform into striated myofibrils is greatly protracted. Furthermore, as Colcemid induces immense desmin intermediate filament cables, the normal spatial relationships between emerging individual myofibrils is distorted. Despite these disturbances at all stages, the characteristic temporal and spatial relationship observed in normal myofibrils between titin and MHC is observed in myofibrils assembling in Colcemid-treated cells. Newly born postmitotic mononucleated myoblasts, or maturing myotubes, reared in taxol acquire a star-shaped configuration and are induced to assemble "pseudo-striated myofibrils." Pseudo-striated myofibrils consist of laterally aggregated 1.6-micron long, thick filaments that interdigitate, not with thin filaments, but with long microtubules. These atypical myofibrils lack Z bands. Despite the absence of thin filaments and Z bands, titin localizes with its characteristics sarcomeric periodicity in pseudo-striated myofibrils. We conclude that the initiation and subsequent regulation of titin and myosin synthesis, and their spatial deployment within developing sarcomeres are tightly coupled events. These findings are discussed in terms of a model that proposes interaction between two relatively autonomous "organizing centers" in the assembly of each sarcomere.

Immunological characterization of the major chick cartilage proteoglycan and its intracellular localization in cultured chondroblasts: a comparison with Type II procollagen.

Polyclonal antibodies were raised in a rabbit against the major proteoglycan of chick sternal cartilage. A total of six antisera was obtained, three after the first booster injection (A1, A2, and A3) and three after the second booster injection (A4, A5, and A6). The A1 antiserum, which was characterized in most detail, immunoprecipitated native as well as chondroitinase ABC-digested or chondroitinase ABC/keratanase-digested cartilage proteoglycan synthesized by cultured chick chondroblasts, but failed to immunoprecipitate the major proteoglycan synthesized by chick skin fibroblasts. This antiserum was also able to immunoprecipitate the cartilage proteoglycan core protein newly synthesized by cultured chondroblasts, but no other major cell protein. However, the late bleed antisera obtained from the same rabbit after a second booster injection reacted with a new chondroblast-specific polypeptide(s) of approximately 60,000 mol wt in addition to the cartilage proteoglycan. By immunofluorescence procedures, the A1 antiserum stained the extracellular proteoglycan matrix of cultured chondroblasts but not that of skin fibroblasts. Following enzymatic removal of the extracellular matrix and cell membrane permeabilization, this antiserum stained primarily a large, juxtanuclear structure. Additional radioautographic evidence suggests that this structure represents the Golgi complex. Similar immunofluorescent staining with antibodies to the cartilage-characteristic Type II collagen revealed that type II procollagen was localized in numerous cytoplasmic, vacuole-like structures which were scattered throughout most of the chondroblast cytoplasm but were notably scanty in the Golgi complex area. In conclusion, our data suggest the transit of the major cartilage proteoglycan through the Golgi complex of cultured chondroblasts and possible differences in the intracellular distribution of newly synthesized cartilage proteoglycan and Type II procollagen.

Taxol induces postmitotic myoblasts to assemble interdigitating microtubule-myosin arrays that exclude actin filaments.

Taxol has the following effects on myogenic cultures: (a) it blocks cell replication of presumptive myoblasts and fibroblasts. (b) It induces the aggregation of microtubules into sheets or massive cables in presumptive myoblasts and fibroblasts, but not in postmitotic, mononucleated myoblasts. (c) It induces normally elongated postmitotic myoblasts to form stubby, star-shaped cells. (d) It reversibly blocks the fusion of the star-shaped myoblasts into multinucleated myotubes. (e) It augments the number of microtubules in postmitotic myoblasts, and these are assembled into interdigitating arrays of microtubules and myosin filaments. (f) Actin filaments are largely excluded from these interdigitating microtubule-myosin complexes. (g) The myosin filaments in the interdigitating microtubule-myosin arrays are aligned laterally, forming A-bands approximately 1.5 micrometers long.

Primitive erythropoiesis in early chick embryogenesis. II. Correlation between hemoglobin synthesis and the mitotic history.

Primitive erythroblasts in the circulating blood of the chick embryo continue to divide while synthesizing hemoglobin (Hb). Hb measurements on successive generations of erythroblasts show that there is a progressive increase in the Hb content of both interphase and metaphase cells. Furthermore, for any given embryo the Hb content of metaphase cells is always significantly greater than that of interphase cells. The distribution of Hb values for metaphase cells suggests that there are six Hb classes corresponding to the number of cell cycles in the proliferative phase. The location of erythroblasts in the cell cycle was determined by combining Feulgen cytophotometry with thymidine radioautography on the same cells. Measurements of the Hb content for erythroblasts in different compartments of the cell cycle (G1, S, G2, and M) show a progressive increase through the cycle. Thus, the amount of Hb per cell is a function of the number of cell divisions since the initiation of Hb synthesis and, to a lesser degree, the stage of the cell cycle. Earlier generations of erythroblasts synthesize Hb at a faster rate than the terminal generation. Several models have been proposed to explain these findings.