Isolation of a Mammalian Homologue of a Fission Yeast Differentiation Regulator

S. pombe strains and media.

The S. pombe strains used in this study have the genotype h⁻ pat1-114 leu1-32 and h⁹⁰ ura4-D18 nrd1::ura4⁺. Media were prepared as described previously (10, 23, 45, 48, 50).

Isolation of multicopy suppressors.

trans-Complementation cloning of the ROD1 gene was performed as described previously (50) with h⁻ pat1-114 leu1-32 cells as a cloning host. A rat kidney fibroblast (NRK-49F) cDNA library and a human fibroblast cDNA library were constructed with the pcD2 vector (6) and transfected into the mutant yeast together with the pAL19 transducing vector (50). Cells were spread on minimum medium agar (MMA) plates, incubated at 23°C for 24 h, and then further incubated at 32.2°C for 4 to 5 days to select rescued cells. The colonies that formed on MMA plates were isolated and subjected to an instability test to distinguish authentic transformants from phenotypic revertants. Plasmid cDNA clones were recovered in Escherichia coli from the colonies that passed the instability test and confirmed for their suppressor activities by subsequent transfection into the host strain. DNA sequencing was performed by the dideoxynucleotide method (61) after being subcloned into M13-derived vectors and pBluescript II KS(+) (Stratagene). The sequence was confirmed by sequencing both strands.

RNA binding analysis.

A nitrocellulose membrane was washed with RNase-free distilled water and then dried. The washed membrane was spotted with 40 μg of poly(A), poly(U), poly(C), and poly(G) and then dried. The RNA homopolymers were then immobilized to the membrane by UV cross-linking. The membrane was incubated for 30 min in 5% nonfat dried milk in 10 mM Tris-HCl (pH 7.2) containing 150 mM NaCl and 0.05% Tween 20 (M-TBST). A glutathione S-transferase (GST)-fused rat Rod1 protein was produced in E. coli from the pGEX2T vector (Pharmacia) containing the Dra1 fragment of the rat ROD1 cDNA. The fusion protein was purified with glutathione-agarose. A 0.5-ml M-TBST solution containing GST-rat Rod1 fusion protein at 20 μg/ml was laid on the membrane and incubated at room temperature for 2 h. The membrane was then washed twice with M-TBST for 10 min and incubated with anti-human Rod1 polyclonal antibody for 1 h. After two washes with M-TBST, the membrane was incubated with horseradish peroxidase-conjugated anti-rabbit immunoglobulin (diluted 1:1,000) (Amersham), and signals were detected by enhanced chemiluminescence (Amersham). A negative control experiment was carried out with the same amount of GST protein, followed by detection with anti-GST polyclonal antibody (MBL).

Antibody production.

The anti-human Rod1 rabbit polyclonal antibody was generated against the whole GST-fused human Rod1 protein produced in E. coli with the pGEX2T vector (Pharmacia) containing the full-length human ROD1. The anti-human Rod1 monoclonal antibody was also generated against the whole GST-fused human Rod1 protein. Both were obtained from MBL.

Conjugation assay.

The mating frequency of the h⁹⁰ ura4-D18 nrd1::ura4⁺ strain was assayed as follows. Cells were grown to mid-log phase in pombe minimum (PM) medium (2% glucose), washed with sterile water, and inoculated in low-glucose (0.5%) PM, NH₄Cl-free PM, or NH₄Cl-free low-glucose (0.5%) PM medium at a density of 5 × 10⁶ cells/ml followed by incubation at 30°C. After incubation for the indicated times, 1 ml of cell suspension was removed and sonicated gently, and the numbers of zygotes were counted under the microscope. The percent mating frequencies were calculated by dividing the number of zygotes (one zygote was counted as two cells) by the number of total cells.

Northern blot analysis.

Total RNA was prepared (13), and Northern blot analysis was performed as described earlier (47). The DNA probes used are the 1.3-kb PvuII fragment of ste11⁺ (65), the 3.2-kb ClaI fragment of mei2⁺ (75), the 1.9-kb cDNA fragment of rep1⁺ (66), and the 0.7-kb HindIII fragment of sxa2⁺ (29).

Cell culture, DNA transfection, and proliferation assay.

The human leukemia cell lines K562 (JCRB 0019) (40), KG-1 (JCRB 9051) (35), U937 (JCRB 9021) (67), Jurkat (ATCC TIB 152) (78), and Daudi (JCRB 9071) (34) were cultured in 5% CO₂ at 37°C in RPMI 1640 medium containing l-glutamine (Irvine Scientific) supplemented with 10% heat-inactivated fetal calf serum (Gibco). HL60 (JCRB 0085) (8) was maintained in the same medium but with 20% fetal calf serum.

The differentiation of cells was induced as follows. K562, HL60, KG-1, U937, Daudi, and Jurkat cells were incubated for the indicated times with the specified culture medium containing 10 nM 12-O-tetradecanoylphorbol-13-acetate (TPA), 10 μM all-trans-retinoic acid, 20 nM TPA, 32 nM TPA, 20 nM TPA, or 2 μg of phytohemagglutinin plus 20 ng of TPA per ml, respectively.

pCMV-human ROD1 contains a full-length human ROD1 cDNA in the cytomegalovirus (CMV) promoter base expression vector (27). Stable transfection was carried out by electroporation. pCMV-human ROD1 (5 μg) and pcD2neo (0.5 μg) were electroporated into 10⁶ K562 cells suspended in 100 μl of phosphate-buffered saline (PBS). After being left on ice for 10 min, cells were incubated in 10 ml of serum-supplemented RPMI 1640 medium for 2 days for recovery, adjusted to 2.5 × 10⁴ cells/ml, and selected for 1 week in the presence of G418 sulfate at a final concentration of 1.5 mg/ml. The selected cells were then diluted 1,000-fold, and 0.2-ml portions of each were transferred to microtiter plate wells and further selected in G418 for 3 to 5 weeks. G418-resistant cells that grew (at frequencies of 1 in 5 or 6 wells) were expanded and analyzed for protein expression and differentiation ability.

Detection of Rod1 protein.

Approximately 10⁶ cells were washed twice with ice-cold PBS and incubated in 0.5 ml of ice-cold 10% trichloroacetic acid (TCA) solution for 30 min. Denatured cells were pelleted by centrifugation at 15,000 rpm for 5 min at 4°C with a Microfuge. The pelleted cells were lysed by gentle sonication in 80 μl of 9 M urea containing 2% Triton X-100 and 1% dithiothreitol (DTT). After the addition of 20 μl of 10% lithium dodecyl sulfate (LiDS) and 10 μl of 1 M Trizma base, cell lysates were sonicated again until their viscosity disappeared and then centrifuged to remove the insoluble material. The cell extracts (5 μl each) derived from 5 × 10⁴ cells were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and analyzed by Western blot as described previously (31).

Various tissues of embryonic and postnatal rats were dissected and quickly frozen in liquid N₂. These tissues were weighed and ground into powder while frozen in liquid N₂. Then, 1 ml of 10% TCA solution was added to 0.1 g (wet weight) of ground tissues, and the mixture was incubated on ice for 30 min. The suspension was transferred into Microfuge tubes and centrifuged at 15,000 rpm for 10 min at 4°C to collect the insoluble material. Protein was extracted from the insoluble material as described above. To 0.1 g each of the tissues, 200 μl of 9 M urea solution containing 2% Triton X-100, 1% DTT, 40 μl of 10% LiDS, and 20 μl of 1 M Trizma base was added. Then, 20 μg each of the extracted protein was loaded into each slot and electrophoresed on SDS gels. Western blot analysis was then carried out with anti-human Rod1 antibody as described earlier (31). For β-actin detection, the blotted membranes used for Rod1 detection were incubated at 50°C for 30 min in stripping buffer containing 62.5 mM Tris-HCl (pH 6.7), 100 mM 2-mercaptoethanol, and 2% SDS. The membranes were then rinsed with a TBS-T solution of 10 mM Tris-HCl (pH 7.4), 0.15 M NaCl, and 0.05% Tween 20 for 30 min at room temperature with three buffer changes and used for Western detection with anti-β-actin monoclonal antibody (Sigma).

Analysis of megakaryocytic differentiation of K562 cells.

TPA (Sigma) was stored at −20°C as a stock solution in dimethyl sulfoxide and diluted with RPMI medium just before use. K562 cells exponentially growing at 10⁵ cells/ml were incubated with culture medium containing 10 nM TPA for up to 3 days. Their megakaryocytic differentiation was monitored by measuring the expression of three independent differentiation markers. The expression of the platelet-specific cell surface glycoprotein IIb/IIIa (CD61) was determined by fluorescence-activated cell sorter (FACS) analysis (Becton Dickinson) and by scanning with an argon laser microscope after immunostaining. The cells were cultured for the indicated times or for 3 days in the growth medium containing the indicated concentrations of TPA. The cells were then harvested, washed twice with cold PBS, and incubated for 30 min on ice with 1:10-diluted fluorescein-labeled mouse anti-human CD61 monoclonal antibody (Dako Corp.) in PBS. Stained cells were washed with cold PBS and fixed in 1% (wt/vol) paraformaldehyde in PBS. Mock-treated cells were similarly analyzed as a negative control. Cells with signals higher than the negative control were considered CD61 positive, and the percent megakaryocytic differentiation was calculated by dividing the population of the CD61-positive cells by the total cell population. The expression of the α₂ integrin protein (CD49b) was assessed by argon laser microscopy after staining with 1:10-diluted fluorescein-labeled mouse anti-human CD49b antibody (Immunotech).

The expression of platelet-derived growth factor B (PDGF-B) was detected as follows. Cells were treated with 10 nM TPA and harvested at the indicated times as described above. Cells (2 × 10⁶) were lysed in 200 μl of lysis buffer containing 10 mM Tris-HCl (pH 7.8), 0.15 M NaCl, 0.1% SDS, 2 mM sodium orthovanadate, 100 mM sodium fluoride, 10 μg of aprotinin per ml, 1 μg of pepstatin per ml, 1 μg of leupeptin per ml, and 0.1 ng of phenylmethylsulfonyl fluoride per ml. Cell extracts (30 μg of protein), quantified by the BCA protein assay method (Pierce), were electrophoresed on an SDS–15% gel and transferred to an Immobilon-P transfer membrane (Millipore) by using the semidry electroblotting apparatus. PDGF-B was detected by Western blot by using the anti-human PDGF-B polyclonal antibody (Santa Cruz Biotechnology) as described earlier (31).

Analysis of erythroid cell differentiation.

n-Butyric acid sodium salt (NaB) (Sigma) was dissolved in PBS, adjusted to pH 7.2 with NaOH, adjusted to a concentration of 1 M, and filter sterilized. Erythroid differentiation of K562 cells was induced by culturing them for 7 days in medium containing 1.25 mM NaB. The cells were harvested, washed twice with PBS, and pelleted. The cell pellets were lysed by vortexing in distilled water (100 μl per 10⁶ cells) containing 0.01% Nonidet P-40 at room temperature and then placed on ice. The protein amounts of the extracts were quantified by the Bradford method (Bio-Rad) and adjusted to 2.5 mg/ml with ice-cold water. The diaminofluorene (DAF) colorimetric reagent (0.6 ml) was added to each 0.2 ml of the adjusted cell extracts, followed by incubation at room temperature for 5 min. Within 10 min, the optical densities were measured with a Bio-Spec 1600 (Shimadzu) spectrophotometer at 610 nm. The DAF colorimetric reagent was made of 0.1 ml of DAF stock solution (1% [weight per volume] of DAF in 90% acetic acid), 0.1 ml of 30% H₂O₂, and 10 ml of 0.1 M Tris-HCl (pH 7.0) containing 6 M urea. For each photograph, 50 μl of each of the cell lysates was mixed with 150 μl of the DAF colorimetric reagent in the wells of a flat-bottom 96-well plate.

Isolation of ROD1 gene.

To search for mammalian genes involved in differentiation control, rat and human fibroblast expression cDNA libraries were screened for genes that suppress the lethality of the pat1-114^ts mutant. By using this screening strategy, two mutually related cDNA clones were obtained. As described below, they were the rat and human counterparts of the same gene and were named ROD1 (regulator of differentiation 1). Both cDNAs rescued the pat1 lethality at restrictive temperatures of up to 34°C.

The rat ROD1 cDNA contains an open reading frame capable of encoding a 523-amino-acid protein with a calculated molecular mass of 56,715 Da and with a localized amino acid homology to polypyrimidine tract RNA binding proteins (4, 19, 20, 55). The predicted human Rod1 protein is two amino acids shorter than (with 96% amino acid identity to) the predicted rat Rod1 (Fig. 1A and B). Like Nrd1, the predicted Rod1 protein has four repeats of typical RNA binding domains containing two semiconserved sequences called RNP1 and RNP2 (Fig. 1C). However, the relative locations of the four RNA binding domains and the intervening spaces were not identical between Rod1 and Nrd1, and their mutual amino acid homology was low (<20% identity). Rod1 has one Cdc2 kinase and two MAP kinase phosphorylation consensus sites in the first RNA binding domain (Fig. 1A). As indicated by its structure, Rod1 has an RNA binding activity. In an in vitro binding assay, Rod1 preferentially bound both poly(G) and poly(U), whereas Nrd1 bound poly(U) (Fig. 1D) (70).

ROD1 shares a functional similarity with nrd1⁺.

The fact that the ROD1 gene was isolated by the same screening strategy as that used for nrd1⁺ and that it encodes an RNA binding protein similar to Nrd1 prompted us to investigate the functional similarity between ROD1 and nrd1⁺. Cells deleted for nrd1⁺ commit conjugation without nutrient starvation (70). In nitrogen-rich low-glucose (0.5%) medium or in nitrogen-poor high-glucose (2%) medium, nrd1 disruptants efficiently conjugate, whereas nrd1⁺ cells remain uncommitted to conjugation or else conjugate poorly. Moreover, in nitrogen-poor low-glucose (0.5%) medium, the standard condition for inducing mating of the fission yeast, the nrd1 disruptants conjugate much earlier and to a higher extent than did nrd1⁺ cells. We compared the ability of ROD1 with that of nrd1⁺ to suppress the phenotype of the nrd1 disruptants under these three culture conditions. The ROD1 and nrd1⁺ coding sequences were inserted in the pcL vector and introduced into a homothallic h⁹⁰nrd1 disruptant and exposed to the three culture conditions for the induction of mating. As shown in Fig. 2, in these assays, both human and rat ROD1 behaved like nrd1⁺ in their ability to suppress the mating-proficient phenotype of the nrd1 disruptant.

One clear function of Nrd1 is to repress Ste11-regulated genes, particularly those that also require the mating pheromone signal for their induction, until the cell reaches a critical level of starvation (70). sxa2⁺ encoding a protease that degrades mating pheromone (29) and rep1⁺ encoding a factor essential for the onset of premeiotic DNA synthesis (66) are among those genes. In the cells lacking nrd1⁺, these genes are derepressed in low-glucose (0.5%) medium as reported previously (70).

To further investigate the functional similarity between nrd1⁺ and ROD1, ROD1 was compared to nrd1⁺ in its ability to repress the sxa2⁺ and rep1⁺ genes derepressed in the low-glucose-exposed nrd1 disruptant. nrd1 disruptants harboring an empty vector, or ROD1 or nrd1⁺ inserted into the pcL expression vector, were transferred into 0.5% glucose PM medium. In the empty vector-harboring disruptant, both rep1⁺ and sxa2⁺ were rapidly induced. The induction of these genes was markedly repressed by ROD1 and nrd1⁺ to the same extent until 9 h posttransfer (Fig. 3). At hour 9, both genes started to be induced, which is at least partly due to a loss of the vector plasmid. Thus, ROD1 was indistinguishable from nrd1⁺ in this ability. All of these results taken together indicate that ROD1 acts in a way functionally similar to nrd1⁺ in the fission yeast.

Rod1 protein is expressed in hematopoietic organs from the early stages of development.

The structural and functional similarity of ROD1 to nrd1⁺ suggested that ROD1 might play a role controlling differentiation in mammals. We investigated this possibility. Because of high amino acid sequence conservation between human and rat Rod1 (Fig. 1B), we first generated polyclonal and monoclonal antibodies against human Rod1 and used them to analyze by Western blotting the tissue-specific expression of Rod1 in rats. Lysates from various tissues of embryonic and postnatal rats were examined. As shown in Fig. 4, Rod1 protein was detected as two bands, of 57 and 50 kDa, by SDS-PAGE. Other anti-Rod1 monoclonal and polyclonal antibodies reacting with different epitopes detected both bands, indicating that both bands were ROD1 gene products. Consequently, one or both of these double bands were likely to be generated by modification or slight truncation of the original Rod1 protein. In 4- to 7-week-old adult rats, Rod1 protein was detected specifically in the spleen, thymus, lungs, and bone marrow (Fig. 4A). In this assay, β-actin protein was used as a control for detection in nonmuscle tissues. The relatively low signal of β-actin, at least in the pancreas, may partly be due to the relatively high expression of muscle-type actins in these organs (18). At early stages of development, Rod1 was expressed in a variety of tissues, even in brain, muscle, and kidney, but it was expressed predominantly in the thymus and liver, where hematopoiesis occurs at these stages (9, 58) (Fig. 4B). We concluded that Rod1 is predominantly expressed in hematopoietic organs throughout development.

Rod1 protein was also expressed in all of the hematopoietic cell lines examined (Fig. 5). K562 is a human pluripotent hematopoietic leukemia cell line and differentiates to megakaryocytes in response to TPA or to erythroids in response to sodium butyrate (5, 68). HL-60 is derived from an acute promyelocytic leukemia patient and differentiates to mature granulocytes upon treatment with retinoic acid (3). KG-1 is also derived from an acute myelogenic leukemia patient and consists of myeloblasts at different stages of maturation (36). Both KG-1 and U937, the latter derived from a generalized histiocytic lymphoma patient, differentiate to mature macrophages upon TPA treatment (43). Daudi is a Burkitt lymphoma-derived human lymphoblastoid cell line and differentiates to mature plasma cells (14). Jurkat is derived from an acute lymphoblastic leukemia patient and differentiates to interleukin-2-producible T cells by costimulation with TPA and phytohemagglutinin (PHA) (21). In these hematopoietic cell lines, the level of Rod1 protein was unchanged or slightly increased during differentiation induced by the appropriate stimuli (Fig. 5).

Overexpression of ROD1 inhibits TPA-induced megakaryocytic differentiation of K562 cells.

Overexpression of nrd1⁺ in fission yeast cells inhibits their conjugation and meiosis without affecting the growth property (70). We therefore examined whether ROD1 exerts a similar activity in mammalian cells. For this experiment, we took the pluripotent hematopoietic human leukemia cell line K562 as a model because this cell line expressed ROD1, as already shown, and differentiates to megakaryocytes upon TPA treatment. Differentiation of this cell can be monitored by the induction of platelet-specific genes, morphological changes, and increased cell-cell and cell-substrate adhesions.

The ROD1 cDNA was inserted into a CMV promoter-based expression vector and was stably transfected in K562 cells, together with the neo marker gene for selection. As shown in Fig. 6A, stable K562 transformants expressing higher levels of Rod1 were obtained. In fission yeast cells, overexpression of nrd1⁺ does not influence the cell’s growth properties (70). Similarly, overexpression of ROD1 did not affect the growth properties of K562 cells. The four highest overexpressors and one modest ROD1 overexpressor, Rod1-C2, Rod1-C3, Rod1-C7, and Rod1-C9 and Rod1-C1, respectively, showed the same growth rates as had the empty vector-transfected K562 cells in a wide range from logarithmic growth to saturation (Fig. 6B).

The ability of the ROD1 overexpressors to perform megakaryocytic differentiation in response to TPA was then examined. In this experiment, differentiation was monitored by semiquantifying the expression of PDGF, CD61, and CD49b. PDGF is a growth factor specifically produced in platelets (7). CD61 (gpIIIa), a component of the fibrinogen receptor complex IIb-IIIa antigen, is a glycoprotein complex associated with platelets and megakaryocytes (5, 81). CD49b (α₂ integrin) forms a complex with β₁ integrin and serves as a cell surface receptor for collagen (5). Among these three, CD61 and PDGF are induced early, whereas CD49b is induced late in megakaryocytic differentiation. The expression of these markers was measured by FACS analysis, Western blotting, or immunostaining. The ROD1 overexpressors and empty-vector-transfected cells were treated with TPA and examined for differentiation. As indicated by the emergence of a positive peak in the FACS analysis, in a majority of the control empty-vector-transfected cells, CD61 was expressed within 24 h of TPA treatment, and further treatment did not significantly increase the number of positive cells (Fig. 7A). By contrast, no obvious positive peak was observed for either one of the two ROD1 overexpressors, although the entire population slightly drifted to the right, a result perhaps due to increased staining backgrounds that were caused by TPA treatment. The authenticity of the positive cells in the FACS analysis was confirmed by microscopic examination of stained cells, which revealed specific staining of cell membranes where CD61 was supposed to be present (Fig. 7B). Again, there were few specifically stained cells in the ROD1 overexpressors.

Similar results were obtained with the other two differentiation markers. In the ROD1 overexpressors, PDGF expression was markedly reduced, as shown by Western blot analysis (Fig. 7C). Moreover, there were much fewer CD49b-expressing cells in the overexpressors (Fig. 7D). These results indicate that overexpression of Rod1 blocked the entire megakaryocytic differentiation of K562 cells rather than merely inhibited the expression of one or two specific genes.

To gain insights into the function of Rod1, the dose response of the overexpressors to TPA was examined in comparison to the control cells. The concentration of TPA examined ranged from <1 to 100 nM. For control cells, differentiation was induced by as low as 0.3 nM TPA, and the maximum differentiation (70%) was obtained with 0.9 nM TPA (Fig. 8B). The two overexpressors showed much lower responses to TPA, which were almost invisible in the FACS patterns, but a calculation demonstrated that there was a linear increase in the CD61-positive cell population at up to 1.2 nM TPA. Further increases in TPA at up to 10 nM or more did not elevate the differentiation frequencies in either Rod1 overexpressor or control cells. These results show that Rod1 overproduction strongly inhibited TPA-induced megakaryocytic differentiation of K562 cells at all of the TPA concentrations tested.

ROD1 overexpression also inhibits sodium butyrate-induced erythroid differentiation of K562 cells.

The ability of Nrd1 to inhibit fission yeast differentiation is not specific to a particular nutrient starvation signal (70). We therefore examined whether the function of Rod1 is specific to particular differentiation signals or even to particular differentiation phenotypes. Treatment with NaB or hemin induces K562 to differentiate to erythroid cells (1, 60). The four ROD1 overexpressors (Rod1-C2, Rod1-C3, Rod1-C7, and Rod1-C9) and the four control clones were treated with NaB for 7 days, and erythroid differentiation was monitored by colorimetrically measuring the amount of synthesized hemoglobin, which is stained blue with DAF. As shown in Fig. 9, NaB-induced erythroid differentiation was also strongly blocked by Rod1 overproduction. Thus, Rod1 action was not specific to particular differentiation signals or differentiation phenotypes, which is the property expected for Rod1 on the basis of its similarity to Nrd1.