Regulation of folate reductase synthesis in sensitive and methotrexate-resistant sarcoma 180 cells. In vitro translation and characterization of folate reductase mRNA.

A highly specific assay for folate reductase mRNA activity from Sarcoma 180 cells was developed using the rabbit reticulocyte lysate protein synthesizing system. Quantitation of in vitro folate reductase synthesis was accomplished by direct immunoprecipitation from lysate reactions. The in vitro labeled folate reductase was synthesized in a linear response to a wide range of RNA concentrations, migrated as a single prominent radioactive species upon polyacrylamide gel electrophoresis, and was indistinguishable from authentic 14C-labeled folate reductase on the basis of molecular weight and immunotitration with anti-folate reductase gamma-globulin. The assay was used to quantitate folate reductase mRNA activity in various cell lines and under several conditions known to affect folate reductase synthesis. These included (a) sensitive and methotrexate-resistant Sarcoma 180 cells, (b) two lines of resistant cells having different relative rates of folate reductase synthesis, (c) growth of methotrexate-resistant cells in the absence of methotrexate, and (d) growth phase. The results indicate that the relative rate of folate reductase synthesis in each case can be explained solely by the level of translatable folate reductase mRNA. The use of poly(U)-Sepharose and sucrose gradient fractionation procedures indicated that folate reductase mRNA contains poly(A) and has a sedimentation coefficient of approximately 14 S. These two fractionation steps were combined to achieve an approximately 90-fold purification of folate reductase mRNA over total cytoplasmic RNA.

Samples were layered on isokinetic sucrose density gradients prepared in 1% (w/v) sodium dodecyl sulfate, 5 mM EDTA, and 10 mM Tris/Cl, pH 7.1, and centrifuged in an SW 41 rotor (Beckman) for 6.5 h at 41,000 rpm. Gradients were prepared as outlined by McCarty et al. (13) for the SW 41 rotor and run at 20". The initial concentration of sucrose at the top of the gradient (i.e. C,,, (13)) was 5% (w/v).

Denaturing Conditions-Samples
were prepared in 85% formamide, 10 mM Tris/Cl, pH 7.5, and 1 mM EDTA and heated for 2 min at 68". The RNA was applied to a linear sucrose gradient and centrifuged for 30 h in an SW 41 rotor using the conditions given by Macnaughton et al. (14). The apparent low level of stimulation of incorporation into folate reductase by RNA from sensitive cells (-8 cpm/pg of RNA) is quantitatively similar to that resulting from the addition of hen oviduct rRNA, i.e. poly(A) minus RNA, to lysate reactions (data not shown). This level of incorporation probably represents nonspecific radioactivity associated with the addition of a heterologous RNA to the lysate reaction. Therefore, a more sensitive comparison of the folate reductase mRNA activity of RNA from sensitive and methotrexateresistant cells was accomplished by electrophoretic analysis of immunoprecipitates from lysate reactions that received comparable amounts of RNA from either cell type. The radioactivity immunoprecipitated from lysate reactions prepared with RNA from resistant cells electrophoresed as one prominent species (Fig. 2~) that co-migrated with authentic "C-labeled folate reductase (Fig. 2~). However, immunoprecipitates from lysate reactions prepared with sensitive cell RNA revealed no significant radioactivity migrating in the region of folate reductase (Fig. 2b). Thus, folate reductase mRNA activity was not detectable in preparations of RNA from sensitive cells. The electrophoretic analysis of folate reductase immunoprecipitates ( Fig. 2) Fig. 3, the same relative percentage of the in vitro 3H-labeled and authentic '%-labeled folate reductase was immunoprecipitated at each y-globulin concentration. These results demonstrate that the anti-folate reductase -y-globulin was unable to distinguish between the in uivo and in vitro labeled enzymes, thus indicating their antigenic similarity.
Levels of Folate Reductase mRNA Activity in Cells with Different Rates of Folate Reductase Synthesis-An interesting feature of methotrexate-resistant Sarcoma 180 cells is that growth for extended periods of time in the absence of methotrexate results in a continual decline in the level of enzymatitally active and immunologically cross-reactive folate reductase that is correlated with a corresponding decrease in the rate of folate reductase synthesis (1). This property of resistant cells was characterized in more detail by comparing the rates of folate reductase synthesis with the levels of folate reductase mRNA activity in two lines of resistant cells (AT-3000 and a clonally derived subline, R-l) as a function of growth in the presence or absence of methotrexate (Table I). R-l cells grown in methotrexate have lower rates of folate reductase synthesis and lower levels of folate reductase mRNA activity than those observed for AT-3000 cells. However, growth of either line of resistant cells in the absence of methotrexate resulted in a continual drop in both folate reductase synthesis and mRNA activity to levels eventually approaching those of sensitive cells (Table I).
Another factor that influences the regulation of folate reductase synthesis in methotrexate-resistant cells, and sensitive cells as well. is growth phase (1) in AT-3000 cells is at least g-fold greater in the early logarithmic phase of cell growth than in stationary phase. In order to define the role of folate reductase mRNA activity in this process, we prepared cytoplasmic RNA from AT-3000 cells that were harvested from logarithmic and stationary phase cultures and determined the levels of folate reductase mRNA activity. A typical experiment is presented in Table II which shows that folate reductase synthesis in stationary phase cells had decreased more than B-fold compared to that of cells in logarithmic growth. Stationary phase cells had similar decline in the level of folate reductase mRNA activity (Table II). In other growth-phase experiments cultures exhibited intermediate levels of folate reductase synthesis and correspondingly intermediate folate reductase mRNA activities (see Fig. 7). eliminate most secondary structure and allow RNA species to sediment largely on the basis of molecular weight (14). Gradient fractions were assayed for folate reductase mRNA activity which, as shown in Fig. 4, was found to sediment the same distance relative to rRNA under either set of gradient conditions. From the gradient shown in Fig. 4a a sedimentation coefficient of 14 S for folate reductase mRNA was calculated using the procedure of McCarty et al. (13). Heating the RNA for 10 min at 68" to disaggregate RNA (5, 10-12) before sedimentation under nondenaturing conditions resulted in no substantial effect on the sedimentation properties of folate reductase mRNA activity. Actiuity-Cytoplasmic RNA from AT-3000 cells was fractionated by sedimentation through sucrose gradients that were prepared either in nondenaturing conditions with sodium dodecyl sulfate (Fig. 4a) or denaturing conditions with 85% formamide (Fig. 46). This latter set of conditions is expected to Chromatography-Cytoplasmic RNA from AT-3000 cells was fractionated by poly(U)-Sepharose chromatography into an unbound fraction and a fraction that was bound and subsequently eluted with formamide. The bound and eluted fraction accounted for 1 to 1.5% of the total cytoplasmic RNA and presumably included the poly(A)-containing mRNAs (22). This fraction was found to be highly enriched in folate reductase mRNA activity (Fig. 5) and comprised approximately twothirds of the activity recovered from the column (Table III). However, the total recovery of folate reductase mRNA activity in both the unbound and the bound and eluted fractions accounted for only 60% of the activity present in the original unfractionated RNA indicating that some of the activity either     was examined by sedimentation through sodium dodecyl sulfate sucrose gradients (Fig. 6). The RNA sedimented over a broad range of sedimentation values peaking in the region of 18 S, a size distribution similar to that seen for other animal cell poly(A)-containing RNA preparations (23,24). There was no appreciable quantity of small molecular weight RNA at the top of the gradient, suggesting that the mRNA preparation was relatively undegraded.
The RNA was essentially free of contamination by 28 S rRNA as judged by sedimentation analysis (Fig. 6); however, the presence of 18 S rRNA cannot be ruled out by this criterion. The gradient fraction having the highest folate reductase mRNA activity appeared in a region corresponding to a sedimentation coefficient of about 14 S (Fig. 6). This fraction represents a purification of approximately go-fold over total cytoplasmic RNA. DISCUSSION We have recently examined several factors affecting the regulation of folate reductase synthesis in Sarcoma 180 cells and found that in each case the level of folate reductase was determined solely by the rate of folate reductase synthesis (1).
In this report we have sought to define the role of folate reductase mRNA activity levels in these regulatory events. In order to carry out these studies, a convenient and highly specific assay for folate reductase mRNA was developed employing the rabbit reticulocyte lysate protein synthesizing system (7,8). On the basis of the following considerations we judged the assay to quantitate specifically the in vitro synthesis of folate reductase.
1. The in vitro labeled product, synthesized in a linear response to a wide range of RNA concentrations, was immunoprecipitated from lysate reactions by a highly specific antifolate reductase -y-globulin and migrated as a single radioactive species upon polyacrylamide gel electrophoresis.
2. The molecular weight of the in vitro product was indistinguishable from that of authentic "C-labeled folate reductase as judged by sodium dodecyl sulfate polyacrylamide gel electrophoresis.
3. Immunotitration of the in vitro product in the presence of "C-labeled folate reductase revealed no antigenic differences detectable by anti-folate reductase y-globulin.

The extent of [3H]leucine incorporation
into the irz vitro product as a function of added RNA was proportional to the level of folate reductase and the rate of folate reductase synthesis in the cells from which the RNA was derived (see Fig.  7 and below). Furthermore, in experiments not shown here, we were able to purify in vitro labeled folate reductase from lysate reactions by the use of methotrexate-Sepharose affinity chromatography. However, the usefulness of this procedure as a routine method for measuring folate reductase synthesis was hampered by poor recoveries.
Folate reductase constitutes as much as 6 to 8% of the soluble protein in mid-logarithmic phase methotrexate-resistant AT-3000 cells. This striking overproduction is entirely due to a specific increase in the relative rate of folate reductase synthesis by the resistant cells (1). We have presented evidence in this report that the high rate of folate reductase synthesis is largely, and probably entirely, due to a corresponding increase in the level of translatable folate reductase mRNA. Control experiments indicated that the low level of folate reductase mRNA activity in cytoplasmic RNA from sensitive cells did not result from the presence of an inhibitor of in uitro folate reductase synthesis or a potent ribonuclease activity that destroyed folate reductase mRNA activity. While this manuscript was in preparation, Chang and Littlefield reported the use of somewhat different techniques to provide evidence for an increase in the level of folate reductase mRNA activity in methotrexate-resistant baby hamster kidney cells (15). We have compared the rates of folate reductase synthesis with the levels of folate reductase mRNA activity under a variety of conditions known to affect the rate of folate reductase synthesis. These include not only (a) sensitive and methotrexate-resistant cells, but also (b) two lines of resistant cells having different rates of folate reductase synthesis, (c) growth of resistant cells in the absence of methotrexate, and (d) growth phase. The results of these independent determinations are compared in Fig. 7 where the relative rates of folate reductase synthesis in cells are plotted as a function of folate reductase mRNA activity assayed in the reticulocyte translation assay. The data clearly indicate that the major, and perhaps only, parameter governing the rate of folate reductase synthesis in each case is the level of translatable folate reductase mRNA.
The high levels of folate reductase mRNA activity present in the methotrexate-resistant cells, and the regulation of folate reductase mRNA activity by the factors mentioned above, presumably result from alterations in either the structure or metabolism of folate reductase mRNA. Such alterations may include those affecting the synthesis, degradation, utilization, or processing (including nuclear transport) of the message. In order to elucidate the role of these processes in the regulation of folate reductase mRNA activity, it will not only be necessary to have a reliable assay for the biological activity of the mRNA but, in addition, to have a specific sequence probe, i.e. a DNA complementary to folate reductase mRNA. These studies require the ability to prepare adequate quantities of folate reductase mRNA in pure form. We have therefore begun an initial characterization and partial purification of folate reductase mRNA.
Sedimentation analysis of folate reductase mRNA activity on isokinetic sucrose gradients (13) containing sodium dodecyl sulfate revealed a sedimentation coefficient of approximately 14 S. In addition, comparison with standards of known molecular weight (16) under conditions expected to eliminate most secondary structure (i.e. 85% formamide) suggested a molecular weight for folate reductase mRNA in the range of 4 x 105, a value consistent with that derived from equations relating sedimentation coefficient to molecular weight (17,18).
This corresponds to about 1,200 nucleotides and is therefore somewhat larger than required to code for folate reductase (Mr = 21,000). However, in view of the uncertainties involved with relating sedimentation behavior to molecular weight these values must be considered tentative until more reliable molecular weight determinations are made. The fact that the molecular weight of the in vitro synthesized folate reductase was indistinguishable from that of authentic 'Glabeled folate reductase suggests that extra nucleotides are not required to code for a higher molecular weight precursor, as observed for some secreted proteins (19)(20)(21). Some untranslated nucleotides are presumably accounted for by the presence of a poly(A) sequence on the mRNA (see below).
Poly(U)-Sepharose chromatography was used to separate folate reductase mRNA and other poly(A)-containing mRNAs (22) from rRNA. Sedimentation analysis of total poly(A)-containing mRNA showed that it was distributed over a broad range of sedimentation coefficients peaking in the region of 18 S. A similar size distribution of animal cell mRNAs has been observed by others (23,24). The combined use of poly(U)-Sepharose and sucrose gradient fractionation procedures resulted in an approximately go-fold purification of folate reductase mRNA activity over total cytoplasmic RNA. However, since folate reductase mRNA activity was not substantially separated from the bulk of the other mRNAs, and the sedimentation profile of poly(A)-containing mRNA revealed no peak of optical density in the region of folate reductase mRNA activity, fractionation of the mRNA on the basis of size will be of only marginal use as a step in its purification. Therefore, we are currently using the more specific indirect immunoprecipitation procedure previously employed in this laboratory for the purification of ovalbumin (25,26), albumin (26), and conalbumin' mRNAs.