the journal of biolcgical chemistry no. 28, … · dence of translational miscoding in uitro. fura...

THE JOURNAL OF BIOLCGICAL CHEMISTRY Vol. 268, No. 28, Issue of October 5, pp. 21438-21442, 1993 Printed in W.S.A.

Effects of 5-Fluorouracil Substitution on the RNA Conformation and in Vitro Translation of Thymidylate Synthase Messenger RNA*

(Received for publication, May 3, 1993)

Chris H. TakimotoSO, Donna B. VoellerS, John M. StrongTI, Larry Andersonn, Ed ChuS, and Carmen J. AllegraS From the Wational Cancer Institute-Navy Medical Oncology Branch, Division of Cancer Tbeatment, National Cancer Institute, National Naval Medical Center, Bethesda, Maryland 20889 and the 11Division of Clinical Pharmacology, Office of Research Resources, Center Drug Evaluation and Research, Food and Drug Administration, Rockville, Maryland 20850

In vitro transcribed thymidylate synthase (TS) mRNA which is lOoO/o substituted with 5-fluorouracil (FUra) was analyzed for changes in mRNA secondary structure, for alterations in translational efficiency, and for evidence of translational miscoding in uitro. FUra substitution in TS mRNA results in an altered migration pattern in non-denaturing RNA gels and in decreased hyperchromicity in RNA melting temperature studies, consistent with a change in mRNA secondary structure. However, no change in the translational efficiency of FUra-substituted TS mRNA is seen compared to control TS mRNA in either rabbit reticulocyte lysate or wheat germ extract in vitro translation systems. Analysis of the in vitro translation product of FUra-substituted TS mRNA by Western immunoblotting, isoelectric focusing, 5-fluoro-2’-deoxyuridine 5’-monophosphate binding, and TS catalytic activity experiments shows no difference compared to control TS mRNA We conclude that the in vitro translation products of FUra-substituted and control TS mRNA are identical. Our findings do not support the hypothesis that changes in the mRNA template are responsible for the RNA-directed cytotoxicity of FUra.

For the past 30 years, 5-fluorouracil (FUra)l has been the most active agent in the chemotherapy of colorectal cancer. Despite this experience, the exact mechanism of action of FUra cytotoxicity has not been precisely defined. The best characterized biochemical effect of FUra is the ability of its active me- tabolite, FdUMP, to inhibit the enzyme thymidylate synthase (TS; EC 2.1.1.45) in the presence of 5,lO-methylenetetrahydrofolate, thus, interfering with DNA synthesis. However, FUra can also be metabolized to 5-fluorouridine 5”triphosphate (FUTP) which can incorporate into RNA leading to alterations in RNA metabolism (1, 2). In some preclinical models, FUra cytotoxicity highly correlates with the degree of its incorporation into RNA, suggesting that an RNA-directed mechanism of cytotoxicity may be important for drug activity (3-6).

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

5 To whom correspondence should be addressed: NCI-Navy Medical Oncology Branch, National Naval Medical Center, Bldg. 8, Rm. 5101, Bethesda, MD 20889. Tel.: 301-496-0901; Fax: 301-496-0047.

synthase; FdUMP, 5-fluoro-2’-deoxyuridine 5‘-monophosphate; FUTP, The abbreviations used are: FUra, 5-fluorouracil; TS, thymidylate

5-fluorouridine 5”triphosphate; dUMP, 2”deoxyuridine 5’-monophosphate; bp, base pair; kb, kilobase; MOPS, 3-(N-morpholino)propanesul- fonic acid; PAGE, polyacrylamide gel electrophoresis; DHFR, dihydrofolate reductase; snRNA, small nuclear RNA.

Although the exact biochemical lesion responsible for this RNA-directed FUra cytotoxicity is not known, FUra can replace uracil in all types of RNA resulting in numerous changes in RNA metabolism (7-10). Alterations in mRNA translation may represent a potentially important mechanism of FUra cytotoxicity. Studies in both bacterial and mammalian systems suggest that FUra incorporation in mRNA may decrease translational efficiency (11) or cause miscoding during protein synthesis (12-14). The mechanisms by which FUra might cause these effects are not known. One possibility is that drug- induced translational errors are caused by base mispairing resulting from FUra-substitution in the mRNA template leading to the synthesis of altered or inactive proteins. Alterna- tively, FUra substitution in mRNA could potentially decrease the efficiency of translation requiring the cell to devote greater resources to protein synthesis. To examine these possibilities, we synthesized TS mRNA which was 100% substituted with FUra and analyzed it for changes in mRNA secondary structure, for alterations in translational eficiency, and for evidence of translational miscoding in vitro. TS was selected because it represents an important target enzyme for FUra activity which is essential for cell division. Furthermore, if FUra incorporation into TS mRNA alters the structure and function of newly synthesized TS protein, then this RNA-directed drug effect could potentially affect the ability of FdUMP to bind to and inhibit the TS enzyme.

MATERIALS AND METHODS Chemical~-[5-~H]dUMP (20 Cilmmol), [6-3H15-FdUMP (18 Ci/

mmol), and [3,8-3H]ATP (18 Cilmmol) were purchased from Moravek Biochemicals (Brea, CA). FUTP was synthesized by Sierra Biochemicals (Tucson, AZ). All other chemicals, unless otherwise specified, were re- agent grade and were purchased from Sigma.

Preparation of in Vitro Synthesized TS mRNA-Full-length TS mRNA ( 1525 bp) was transcribed from human TS cDNA using SP6 RNA polymerase (Promega, Madison, WI) according to the Promega protocol as described previously (15). Completely FUra-substituted TS mRNA was synthesized in an identical manner except that FUTP was used in place of UTP in the reaction mixture. The rate of the in vitro transcription reaction was measured by adding 3 pCi of [3,8-3H1ATP to the reaction mixture and subsequently removing 3-pl aliquots at various time points. The aliquots were precipitated by spotting on 1-cmz pieces of filter paper (Whatman No. 3, Maidstone, England) and boiling for 10 min in 10% (w/v) trichloroacetic acid, followed by two washes each in water, ethanol, and acetone. The filter papers were then air-dried and placed in 10 ml of 3a70B scintillation mixture (RPI, Mount Prospect, IL), and the tritium radioactivity was measured in a Beckman LS3801 scintillation counter. The synthesis of completely FUra-substituted TS mRNA was confirmed by hydrolysis of the RNA and subsequent analysis for uracil and FUra bases by gas chromatography/mass spectrometry (16).

RNA Gel Electrophoresis and RNA Melting Curves-TS mRNA (1 pg) was analyzed by electrophoresis on 12-cm horizontal 1.5% agarose gels using 20 m~ MOPS, 50 m~ sodium acetate, 1 mM EDTA, pH 7.0, as running buffer. Denaturing gels included 37% formaldehyde and were

21438

5-Fluorouracil-substituted Thymidylate Synthase mRNA 21439 run at 75 mA for 45 min. Overall, RNA migrated much more slowly in nondenaturing gels which were run at 75 mA for 2 h. Gels were stained with ethidium bromide and were photographed under W illumination using an Eagle Eye frame integrator (Stratagene, La Jolla, CAI.

Absorbance melting curve analysis of TS mRNA was performed using the method of Puglisi and Tinoco (17). Briefly, 1 pg of TS mRNA was dissolved in 1 ml of 50 mM sodium chloride, 10 m~ sodium phosphate, pH 7.0, and 0.1 m~ EDTA, pH 8.0, and the W absorbance was measured at 260 nm over a range of 28-70 "C using a Hewlett-Packard 8452A spectrophotometer equipped with a digitally controlled HP 89090A Peltier temperature control accessory cuvette holder.

In Vitro nanslation-Translation reaction mixtures (final volume, 30 pl) containing 10 pl of rabbit reticulocyte lysate, 66 m~ potassium acetate, 542 p~ magnesium acetate, 50 pCi of [35Slmethionine, and 13 pl of translation mixture, all supplied in an in vitro translation system kit (Du Pont-New England Nuclear), were incubated with TS mRNA at 37 "C for 60 min, and the products were analyzed by SDS-PAGE (12.5% acrylamide) as previously described (15). Unincorporated radioactive label was removed by washing the gel three times in 100 ml of 40% methanol, 10% acetic acid (v/v) for 20 min. The gel was then soaked in 250 ml of Enlightning (Du Pont-New England Nuclear) containing 3% glycerol for 2 h at room temperature. After drying for 2 h, the translation products were identified by autoradiography. Wheat germ extract translation reactions (final volume, 50 pl) contained 25 pl of wheat germ extract, 4 pl of an amino acid mixture minus methionine, 75 mM potassium acetate, 25 pCi of [35S]methionine, and 1 p1 of RNasin RNase inhibitor, all supplied in an in vitro translation system kit (Promega). The wheat germ reaction mixtures were incubated with TS mRNA for 25 "C for 60 min and then were analyzed in the same manner as the rabbit reticulocyte lysate reactions described above.

Western Immunoblot Analysis-In vitro translation protein products were resolved by SDS-PAGE (12.5% acrylamide), and the gel was electroblotted onto a nitrocellulose membrane with an LKB 2117 Multiphor I1 electrophoresis unit. Antibody staining was performed with a TS polyclonal antibody (18) at a 1:lOO dilution in Blotto buffer. Ahorserad- ish-conjugated antibody (Bio-Rad) at a dilution of 1:lOOO in Blotto buffer was used as the secondary antibody.

Isoelectric Focusing Analysis-Aliquots (10 pl) of the rabbit reticulocyte lysate in vitro translation reaction mixtures were analyzed on Ampholine PAGplates (1 mm thickness), pH range 3.5-9.5 (Pharmacia LKB Biotechnology Inc.) on a LKB Multiphor I1 isoelectric focusing apparatus set at a current of 50 mA for 1.5 h. After isoelectric focusing was completed, the proteins were visualized by autoradiography.

Thymidylate Synthase-FdUMP-Folate Complex Formation-The ability of the in vitro synthesized TS protein to bind FdUMP was measured by a modification of a previously described method (19). Aliquots (25 pl) of rabbit reticulocyte lysate in vitro translation reaction mixtures incubated with TS mRNA were combined with 75 p~ 5,lO-methylenetetrahydrofolate, 4.4 pmol of [6-3H15-FdUMP (18 Ci/mmol), 100 mM 2-mercaptoethanol, and 50 m~ potassium phosphate, pH 7.2, in a total volume of 200 1.11. Samples were incubated at 37 "C for 30 min, and then 1 ml of an albumin-coated charcoal slurry diluted 1:4 with 250 mM potassium phosphate, pH 7.2, was added. The albumin-coated charcoal slurry was prepared by mixing 10 g of acid-washed activated charcoal with 2.5 g of bovine serum albumin, 0.25 g of dextran, and 100 ml of ice-cold water. The mixtures were vortexed and allowed to stand at room temperature for 10 min and then centrifuged at 10,000 x g for 30 min. An 0.8-ml sample of the supernatant was removed and added to another 200 pl of undiluted charcoal slurry, vortexed, and resedimented at 10,000 x g for 30 min. An 0.8-ml sample of the supernatant was counted by liquid scintillation, and the results were corrected for the amount of [6-3H15-FdUMP bound nonspecifically (about 0.06 pmol) by rabbit reticulocyte lysates alone.

Thymidylate Synthase Catalytic Assay and Enzyme Kinetic Analysis -The catalytic activity of in vitro synthesized TS enzyme was measured by a modification of a previously described method (19). The assay was performed in a total volume of 200 pl containing 25 pl of the rabbit reticulocyte lysate in vitro translation reaction mixture, 100 mM 2-mercaptoethanol, 50 m~ potassium phosphate, pH 7.2, 75 p~ 5.10-methylenetetrahydrofolate, and concentrations of [B-"HIdUMP ranging from 2.5 to 25 p i . The reactions were incubated for 30 min at 37 "C and the reactions were stopped by adding 100 pl of ice-cold 20% trichloroacetic acid. Residual [5-3HldUMP was removed by adding 200 pl of an albumin-coated charcoal slurry, pH 7.2 (prepared by mixing 10 g of acid- washed activated charcoal with 2.5 g of bovine serum albumin, 0.25 g of dextran, and 100 ml of ice-cold water). The samples were vortexed and allowed to stand at room temperature for 10 min. The charcoal was sedimented by centrifugation at 10,000 x g for 30 min. A 300-pl sample

of the supernatant was combined with another 200 pl of albumin-coated charcoal slurry, vortexed, and immediately resedimented. A 300-pl sample of the supernatant was removed and assayed for L3H1H20 radioactivity by liquid scintillation. All assays were performed in dupli- cate. Enzyme kinetics were analyzed by fitting the experimental data to a Michaelis-Menten model using a weighted, nonlinear regression com- puter program designated "ENZYME as previously described (20). Es- timates ofK, and V,,, were obtained from three separate experiments, and statistical comparisons of mean values were made using the inde- pendent Student's t test.

RESULTS

Synthesis of FUra-substituted Thymidylate Synthase mRNA " I n vitro transcription of a plasmid containing full-length human TS cDNA using bacteriophage SP6 RNA polymerase results in a 1.5-kb transcript of human TS mRNA. FUra-substituted TS mRNA was synthesized by replacing the UTP with FUTP in the transcription reaction. The efficiency and the time course of the in vitro transcription reaction were quantified by adding [3,8-3H]ATP to the reaction mixture and measuring the tritium label accumulating in the acid-precipitable (RNA-containing) fraction over time (Fig. 1). The overall yield of TS mRNA is decreased to 82% when FUTP replaces UTP in the reaction mixture; however, the relative rates of the RNA polymerase reactions are similar with 50% of the maximal product formed after approximately 20 min for each reaction (Fig. 1). Complete hydrolysis of the in vitro transcribed TS mRNA and analysis of the pyrimidine bases by gas chromatography and mass spectrometry confirm the complete (>99.9%) replacement of uracil with FUra in the FUra-substituted TS mRNA. The fidelity of the in vitro transcription reaction utilizing FUTP was not directly examined; however, the results of our subsequent experiments on the translation of TS mRNA suggest an equiv- alent degree of fidelity of the SP6 RNA polymerase using either UTP or FUTP as substrate.

Altered Electrophoretic Migration of FUra-substituted TS mRNA in Nondenaturing Agarose Gels-Changes in RNA conformation and secondary structure of FUra-substituted TS mRNA were examined by gel electrophoresis in denaturing (formaldehyde-containing) and nondenaturing agarose gels. Under. denaturing conditions', the electrophoretic migration patterns of wild-type and FUra-substituted TS mRNA are identical (Fig. 2 A ) . In contrast, under nondenaturing conditions, FUra-substituted TS mRNA migrates more slowly than control TS mRNA (Fig. 2B) consistent with a difference in the RNA molecular conformation.

RNA Melting Curve Analysis-To further examine the difference in RNA secondary structure between FUra-substituted and control TS mRNA, RNA melting temperature experiments were performed. Heat denaturation of the ordered, base-paired

40

n z30 X -

.I // I

0 1 0 2 0 3 0 4 0 5 0 8 0 7 0 8 0 9 0

TIME (min)

FIG. 1. Efficiency of in uitro transcription of human TS eDNA Incorporation of [3,8-3H1ATP into TS mRNA was measured in reactions which utilize either UTP (B) or FUTP (e) as described under "Materials and Methods."

21440 5-Fluorouracil-substituted Thymidylate Synthase mRNA

A. B.

49.5 -

tuted Ts mRNA under denaturing and nondenaturing condi- FIG. 2. Agarose gel electrophoresis of control and FUra-substi-

tions. A, denaturing 1.5% agarose-formaldehyde gel electrophoresis of control TS mRNA rlane 1 ) and FUra-substituted TS mRNA (lane 2). Sizes are shown in kilobases. B, nondenaturing 1.5% agarose gel electrophoresis of control TS &A (lane I ) and FUra-substituted TS mRNA ( l a n e 2). Conditions of electrophoresis are as described under "Materials and Methods." TS RNA is indicated by the arrows.

'& i I

38 48 58 68 Temperature ("C)

FIG. 3. RNA mel t ing t empera tu re curves f o r control Ts mRNA (0) and FUra-substituted TS mRNA (W). Experimental conditions a re as described under 'Materials and Methods."

regions in an RNA molecule increases its UV absorption (hyperchromicity), and the change in UV absorbance is propor- tional to the number and type of base pairs broken (17). The melting curve of TS mRNA is markedly changed by FUra incorporation resulting in less hyperchromicity (6.3%) compared to wild-type TS mRNA (10.8%) (Fig. 3). This result is consistent with a decrease in RNA base pairing in FUra-containing TS mRNA, and it suggests that FUra-incorporation alters RNA secondary structure.

In Vitro Ranslation of FUra-substituted mRNA-The effects of FUra substitution on mRNA translation were examined by incubating TS mRNA in either rabbit reticulocyte lysates (Fig. 4A 1 or in wheat germ extracts (Fig. 4B ). Translation of FUra- substituted and control TS mRNA both yield a 35-kDa protein product in either in vitro translation system. The efficiency of translation in rabbit reticulocyte lysates was measured by de- termining the amount of 35S radioactivity accumulating in the trichloroacetic acid-precipitable fraction over time. Identical results are observed for both FUra-substituted and control TS mRNA (Fig. 5).

To further characterize the 35-kDa protein product of translation of FUra-substituted TS mRNA, a Western immunoblot analysis was performed using a rabbit polyclonal TS antisera

32.5 - 27.5

a

FIG. 4. Effect of FWra substitution on the in vitro translation of Ts mRNk Translation reactions containing either 10 pI of rabbit reticulocyte lysate ( A ) or 25 pl of wheat germ extract tB ) were incubated with no RNA, 2 pg of control TS mRNA, or 2 pg of FUra-substituted TS mRNA. Translation reactions were incubated a t 37 'C for 60 min and the protein products were analyzed by SDS-PAGE and autoradiography as described under "Materials and Methods." Sizes are shown in kilodaltons. TS protein is indicated by the arrow.

'" I I

* NoRNA * Control TS RNA -c FUra TS RNA 40

204 0 1 0 2 0 3 0 4 0 5 0 6 6

TIME (minutes)

FIG. 5. Efficiency of in uitro translation of FUra-substi tuted TS mRNA in rabbit reticulocyte lysates. Incorporation of ["Slmethio-

added RNA (W), control TS mRNA (O), or FUra-substituted TS mRNA nine into the protein products of translation was measured using no

(A) as described under "Materials and Methods."

(18). In vitro translation of both control and FUra-substituted TS mRNA yields a 35-kDa protein product which is recognized by rabbit polyclonal TS antisera (Fig. 6). No qualitative or quantitative differences between the FUra-substituted or control TS mRNA translational protein products are observed.

Isoelectric Focusing Analysis-Further analysis of the translational protein products was performed by isoelectric focusing. Aliquots of the rabbit reticulocyte lysate in vitro translation reaction mixture were placed on an isoelectric focusing apparatus, and the proteins were separated on the basis of their isoelectric points. The in vitro translation products of FUra- substituted and control TS mRNA resolve at the same isoelectric point, consistent with identical amino acid sequences for both proteins (Fig. 7).

FdUMP-binding Assay and TS Catalytic Activity "Immunologic and electrophoretic evidence suggests that the in vitro translational product of FUra-containing TS mRNA is identical to wild-type TS. However, amino acid substitutions might still be present which would be detected only by using more sensitive tests of enzyme structure or function. Therefore, we examined the ability of TS protein synthesized from FUra- containing TS mRNA to form a ternary complex with FdUMP and 5,lO-methylenetetrahydrofolate. Specific FdUMP-binding activity, approximately 2.5-3-fold higher than background, is detected only when either FUra-substituted or control TS

5-Fluorouracil-substituted 1

a i = P m c

49.5 -

32.5 - c

FIG. 6. Western immunoblot analysis of rabbit reticulocyte lysate in vitro translation products. Translation reactions included no exogenous RNA (Ianr I ), 2 pg of control TS mRNA (lane 2). or 2 pg of FUra-substituted TS mRNA (lane 3 ) . Translation reaction mixtures were incubated a t 37 "C for 60 min. and the protein products were resolved by SDS-PAGE. The gel was electroblotted onto a nitrocellulose membrane and antibody staining was performed as described under "Materials and Methods." TS protein is indicated by the arrow. Sizes are shown in kilodaltons.

Q

6.8 - 6.0 - 5.6 - 5.2 - 4.8 -

c

FIG. 7. Isoelectric focusing analysis of rabbit reticulocyte lysate in vitro translation products. Translation reactions included no added RNA (lane I ), 2 pg of control TS mRNA (lane 2). or 2 pg of FUra-substituted TS mRNA tlanr 3 ) . Translation reaction mixtures were incubated a t 37 "C for 60 min, and the protein products were analyzed by isoelectric focusing over a pH range of 9.5-3.5 as described under "Materials and Methods." TS protein is indicated by the arrow.

mRNA is added to rabbit reticulocyte lysates (Table I). The amount of ["H IFdUMP bound by the lysate mixtures containing FUra-substituted and control TS mRNA is not significantly different ( p = 0.58).

TS normally catalyzes the conversion of dUMP to dTMP by reductive methylation of the C-5 position.The activity of the newly synthesized TS enzyme was quantitated by adding [5-:3HldUMP and 5,lO-methylenetetrahydrofolate to the lysates and quantitating the amount of tritium released. TS catalytic activity 3.5-4-fold higher than background is seen only when either FUra-substituted or control TS mRNA is added to the rabbit reticulocyte lysates. The calculated Km values for dUMP as measured in the rabbit reticulocyte reaction mixture are not

'%,ymidylate Synthase mRNA 2 144 1

TAIILE I FdUMP hinding and TS nrtic8it.v in rahhif rrfircrlotvte 1.vsnfe.v

incuhnted with FlJrn-suhstitutrd TS mRNA Rabbit reticulocyte lysate in vitro translation rrnction mixturrs con-

taining 0.75 pg of either FUra-substitutrd or control TS mRNA wrrr incuhated with 75 p~ 5.10-methylenrtrtrahydrofolatr and rithrr 1"HIF- dUMP (4.4 pmol) for the FdUMP hinding measurrmrnts or with vary- ing concentrations of 1:'HIdUMP for TS catalytic activitv rxprrimrnts. FdUMP binding activity was quantified as thr amount of I '1IlFdUMP hound as described under "Materials and Methods." Thr thymidylatr synthase K,. for dUMP and V,,,, were drterminrd by the tritium rrleasr assay using a computerized nonlinear least-squarrs curvr fitting program as described under 'Materials and Mrthnds." Comparison hr- tween mean values were made using the indeprndent Studrnt's f trst. Values represent the mean = standard error of thrrr diffrrrnt rxprri- ments.

~ ~ ~ "" ~~

RNA I:'HIFdUW' ~ ~

hindmy: nrtlvtty Km,

TS ralnlvtlr nrtlvlty

V".... ~

pmol1.70 p l p u p m o l / m ~ n / . ' f ~ p l

~~ ~~ ~ ~~

of 1.vsnlr. of Ivsnr'.

Control TS mRNA 0.08 0.01 9.9 e 0.9 13.1 2 1.2 FUra TS mRNA 0.06 = 0.03" 9.2 n.4" 1n.3 e n.5" " No significant difference, p > 0.05.

significantly different for FUra-substituted TS mRNA (9.2 p ~ ) compared to control (9.9 p ~ ) ( p = 0.45) (Table I). These values are slightly higher than the previously reported TS enzyme K,, values for dUMP measured under optimized conditions (K,, = 3.4 pM) (21).

DISCUSSION

Evidence that FUra incorporation in mRNA causes miscoding during translation initially came from studies in bacteria. Nakada and Magasanik (12) found that FUra readily incorporated into Escherichia coli mRNA resulting in the synthesis of an altered, inactive p-galactosidase enzyme. This altered protein was postulated to result from the translation of FUra- containing mRNA. Additional evidence came from studies by Rosen et al. (13) on the amber alkaline phosphatase mutation in E. coli. The replacement of uracil by FUra changed the nonsense mutant amber codon. UAG, to FUra-AG, which was read during translation as CAG, resulting in the phenotypic reversal of the amber mutation. Evidence for translational miscoding in mammalian cells came from a series of experiments by Dolnick and Pink (14) in human KB cells which overexpress dihydrofolate reductase (DHFR). In this system. FUra exposure did not change the rate of synthesis of DHFR as measured by immunoprecipitation; however, a significant decrease in enzyme activity was observed. This apparent paradox was examined further by purifying DHFR mRNA from drug-treated cells and translating it in vitro in rabbit reticulocyte lysates. No change in the in vitro translational activity of the FUra-containing mRNA was seen. However, the DHFR enzyme isolated from these cells after FUra exposure had altered affinities for binding methotrexate and anti-DHFR antisera. These changes in enzyme structure and function were attrihuted to amino acid substitutions resulting from mispairing of FUra in mRNA during translation. Under the conditions used in this study, the amount of FUra substitution in total cellular RNA was esti- mated at 2%; however, the degree of substitution in the specific DHFR mRNA undergoing translation was not quantified. In addition, the protein product of the in rritro translation reaction was not characterized. Nonetheless, these observations sup- ported the theory that translational miscoding was an important mechanism of RNA-directed FUra cytotoxicity.

Our experiments do not demonstrate any change in mRNA function as a result of FUra incorporation in TS mRNA. No difference is seen in the translational eficiency of FUra-containing TS mRNA compared to wild-type TS mRNA, and both

2 1442 5-Fluorouracil-substituted Thymidylate Synthase mRNA yield similar 35-kDa protein products. Further analysis by Western immunoblots, isoelectric focusing, FdUMP-binding, and TS catalytic activity shows no difference in the protein products of translation of FUra-substituted and wild-type TS mRNA leading us to conclude that the proteins are identical. There is no evidence that FUra incorporation into the mRNA template causes miscoding during translation. Our findings are in agreement with others who have failed to observe FUra- induced alterations in mRNA translation. Grunberg-Manago and Michelson (11) found that the in vitro translation of polyfluorouridylic acid RNA resulted exclusively in the synthesis of polyphenylalanine, which is the same protein encoded by polyuridylic acid RNA. No evidence for translational miscoding was observed; however, in contrast to our findings, the efficiency of in vitro translation of polyfluorouridylic acid RNA was decreased by over 90% compared to polyuridylic acid RNA. In other experiments, Glazer and Hartman (22, 23) found that FUra incorporation into total cellular polyadenylated RNA only minimally affected its in vitro translational activity, even after exposure to toxic levels of FUra. They concluded that FUra- induced changes in mRNA were not important mechanisms of drug action.

However, our findings do not completely exclude FUra effects on mRNA as a mechanism of cytotoxicity. Potentially important drug-induced alterations in mRNA metabolism may still occur at the level of nuclear mRNA processing rather than during translation. FUra-substituted precursor mRNA can be blocked from reaching the cytoplasmic translation apparatus by a number of mechanisms, including impaired nuclear processing (241, increased nuclear degradation (25), or decreased nuclear trans- port (10). The best evidence for a FUra effect on mRNA processing comes from an elegant series of experiments by Doong and Dolnick (26). FUra substitution in precursor DHFR mRNA resulted in the formation of abnormal splicing products in an in vitro mRNA processing system. However, over 80% substitution with FUra was required before processing changes were observed. An alternative mechanism for FUra-induced miscoding has been proposed by Gleason and Fraenkel-Conrat (27). 5-Fluorocytosine, which can be generated in some cells by ami- nation of FUra, incorporated into RNA in the place of cytosine and was subsequently read as uracil during protein synthesis. Different cell lines varied in their ability to convert FUra to 5-fluorocytosine which may explain why translational miscoding was inconsistently observed. This would also account for the lack of translational miscoding seen in our in vitro system. A final possibility is that translational miscoding may result from FUra actions on other RNA species, such as rRNA (71, tRNA (28), or uracil-rich small nuclear RNA (snRNA) (10, 29). Although these mechanisms were not examined in our study, they should be pursued in further investigations.

The complete replacement of uracil with FUra in TS mRNA in our experiments results in an altered migration pattern in non-denaturing RNA gels and in decreased hyperchromicity in RNA melting temperature studies. These observations suggest that FUra substitution in TS mRNA induces a change in RNA secondary structure, a finding which is in agreement with other studies of FUra incorporation in RNA. Armstrong et al. (10) reported that FUra altered the secondary structure of snRNA as evidenced by the slower migration of certain snRNA bands on non-denaturing polyacrylamide gels. Likewise, Danenberg et al. (30) examined the effects of FUra-incorporation on the self-splicing activity of Tetrahymena rRNA. Substitution with 100% FUra moderately inhibited the autocatalytic, self-splicing reaction of this RNA, and it greatly increased the pH and temperature sensitivity of the process. This inhibition was at-

tributed to a conformational change in the RNA molecule resulting from weaker base pairing between FUra and adenine. The physiochemical differences between FUra and uracil in their ability to form base pairs are incompletely understood (31). Although the physical structure of the FUra-adenine base pair is similar to uracil-adenine, its rate of opening is enhanced (31), resulting in greater thermal instability (11). This difference has been attributed to the enhanced acidity of the N-3 hydrogen in FUra leading to a higher degree of ionization at physiologic pH; the pK, of FUra is 7.8 and uracil is 10.1 (30). However, high resolution nuclear magnetic resonance studies of oligodeoxynucleotide base pairs suggest that abnormal FUra base pairing is caused by decreased stacking of the FUra resi- dues and not by enhanced acidity (31).

In summary, we studied the effect of 100% replacement of uracil by FUra on the RNA conformation and in vitro translation of TS mRNA. Despite evidence that FUra incorporation alters RNA secondary structure, no difference in the translational efficiency of FUra-containing TS mRNA is observed. The protein product of translation of FUra-substituted TS mRNA is identical to control TS. These findings do not support the hypothesis that FUra-incorporation into mRNA causes miscoding during translation. Thus, it is unlikely that changes in the mRNA template are responsible for the RNA-directed cytotoxicity of FUra. Further studies are still necessary to evaluate whether FUra can alter protein translation by incorporation into other species of RNA, such as snRNA, rRNA, or tRNA.

REFERENCES 1. Grem, J. L. (1990) in Cancer Chem0therapy:Principles & Practice (Chabner, B.

A,, and Collins, J. M., eds) 2nd Ed., pp. 18&224, J. B. Lippincott Co., Philadelphia

2. Parker, W. B., and Cheng, Y. C. (1990) Pharmucol. & Ther. 48,381-395 3. Cadman, E. C., Heimer, R., and Davis, L. (1979) Science 206, 1134-1137 4. Spiegelman, S., Sawyer, R., Nayak, R., Stolfi, R., and Martin, D. (1980) Proc.

5. Kufe, D. W., and Major, P. P. (1981) J. Biol. Chem. 266, 9802-9805 6. Martin, D. S., Stolfi, R. L., Sawyer, R. C., and Spiegelman, S. (1983) Cancer

7. Wilkinson, D. S., and Pitot, H. C. (1973) J. Biol. Chem. 248, 63-68 8. Takimoto, C. H., Tann, Y. Y., Cadman, E. C., and Armstrong, R. D. (1987)

9. Greenhalgh, D. A,, and Parish, J. H. (1990) Br. J. Cancer 61, 415-419

Natl. Acad. Sci. U. S. A. 11, 496W970

Res. 43,2317-2321

Biochem. Pharmacol. 36,3243-3248

10. Annstrong, R. D., Takimoto, C. H., and Cadman, E. C. (1986) J. Biol. Chem.

11. GrunberE-Manago, M., and Michelson,A. M. (1964)Biochim. Biophys. Acta 87, 261.21-24

12. 13. 14. 15.

16.

17. 18.

19.

20.

21.

22. 23. 24. 25. 26. 27.

28. 29.

30.

31.

593-600 Nakada, D., and Magasanik, B. (1964) J. Mol. Biol. 8, 105-127 Rosen, B., Rothman, F., and Weigert, M. G. (1969) J. Mol. Biol. 44, 363-375 Dolnick, B. J., and Pink, J. J. (1985) J. Biol. Chem. 260, 30063014 Chu, E., Koeller, D. M., Casey, J. L., Drake, J. C., Chabner, B. A,, Elwood, P. C.,

Zinn, S., and Allegra, C. J. (1991) Proc. Natl. Acad. Sci. U. S. A. 86, 8977-

Uchida, J., Umeno, Y., and Takeda, S. (1992) Gan To Kagaku Ryoho 19,677- 8981

Puglisi, J. D., and Tinoco, I. (1989) Methods Enrymol. 180,304-325 683

Chu, E., Drake, J. C., Koeller, D. M., Zinn, S., Jamis-Dow, C. A., Yeh, G. C., and

Chu, E., Zinn, S., Boarman, D., and Allegra, C. J. (1990) Cancer Res. 60,

Allegra, C. J. , Chabner, B. A,, Drake, J. C., Lutz, R., Rodbard, D., and Jolivet,

Davisson, J. V., Sirawaraporn, W., and Santi, D. V. (1989) J. Biol. Chem. 264,

Glazer, R. I., and Hartman, K. D. (1981) Mol. Pharmucol. 19, 117-121 Glazer, R. I., and Hartman, K. D. (1983) Mol. Pharmacol. 23,540-546 Will, C. L., and Dolnick, B. J. (1987) J. Biol. Chem. 262, 5433-5436 Will, C. L., and Dolnick, B. J. (1989) J. Biol. Chem. 264, 21413-21421 Doong, S. L., and Donick, B. J. (1988) J. Biol. Chem. 263,44674473 Gleason. M. K.. and Fraenkel-Conrat, H. (1976) Proc. Natl. Acad. Sei. U. S. A.

Allegra, C. J. (1991) Mol. Pharmacol. 39, 136143

5834-5840

J. (1985) J. Biol. Chem. 260, 972&9726

9145-9148

Tseng, W. C., Medina, D., and Randerath, K. (1978) Cancer Res. 38,1250-1257 Sierakowska, H., Shukla, R. R., Dominski, Z., and Kole, R. (1989) J. Biol.

73, 1528-1531

Danenberg, P. V., Shea, L. C. C., and Danenberg, K. (1990) Cancer Res. 60,

Sowers, L. C., Eritja, R., Kaplan, B. E., Goodman, M. E, and Fazakerley, G. V.

Chem. 264, 19185-19191

1757-1763

(1987) J. Biol. Chem. 262, 1543G15442

the journal of biolcgical chemistry no. 28, … · dence of translational miscoding in uitro. fura...

Documents