Species-specific differences in translational regulation of dihydrofolate reductase

Abstract

We have observed that rodent cell lines (mouse, hamster) contain approximately 10 times the levels of dihydrofolate reductase as human cell lines, yet the sensitivity to methotrexate (ED(50)), the folate antagonist that targets this enzyme, is similar. Our previous studies showed that dihydrofolate reductase protein levels increased after methotrexate exposure, and we proposed that this increase was due to the relief of feedback inhibition of translation as a consequence of methotrexate binding to dihydrofolate reductase. In the current report, we show that unlike what was observed in human cells, dihydrofolate reductase (DHFR) levels do not increase in hamster cells after methotrexate exposure. We provide evidence to show that although there are differences in the putative mRNA structure between hamster and human mRNA in the dihydrofolate reductase binding region previously identified, "hamsterization" of this region in human dihydrofolate reductase mRNA did not change the level of the enzyme or its induction by methotrexate. Further experiments showed that human dihydrofolate reductase is a promiscuous enzyme and that it is the difference between the hamster and human dihydrofolate reductase protein, rather than the DHFR mRNA, that determines the response to methotrexate exposure. We also present evidence to suggest that the translational up-regulation of dihydrofolate reductase by methotrexate in tumor cells is an adaptive mechanism that decreases sensitivity to this drug.

Fig. 1.

a, Western blot analysis of total cell lysates from human and rodent cell lines after MTX exposure. Human (C85), hamster (CHO), and mouse (3T6) cells were exposed to l μM MTX or media without MTX for 24 and 48 h. b, increase in EGFP fluorescence of various transfectants after 1 μM MTX treatment for 24 h. EGFP and wt human DHFR-EGFP transfectants are negative and positive control, respectively. c, human and hamster DHFR-EGFP transfectants of DG44 cells were exposed to either 1 or 10 μM MTX in the presence of hypoxanthine and thymidine to prevent cell death due to MTX treatment for 24 and 48 h. d, DHFR-EGFP transfectants were exposed to 1 μM MTX, trimetrexate (TMTX), raltitrexed (RTX), and pemetrexed (PTX), respectively. a to c, 50 μg of total cell lysate was loaded and probed with either an anti-DHFR antibody (a) or EGFP antibody (b and c). Equal loading was determined by either α-tubulin or GAPDH, respectively.

Fig. 2.

a, comparison of the primary sequence of binding region (BR) of human DHFR to hamster DHFR. b, secondary structures of human and hamster DHFR within the binding region were obtained using the mfold program. c and d, hamsterized human DHFR-EGFP protein levels increase upon exposure to MTX. Batches (c) and four individual clones (d) from DG44 cells transfected with hamsterized DHFR-EGFP. Cells were exposed to 1 μM MTX or media without MTX for 24 and 48 h. Total cell lysate (50 μg) was loaded and probed with an anti-EGFP antibody. Equal loading was determined by Ponceau S staining (data not shown).

Fig. 3.

a, experimental design for fully mutated binding region (FBMR) and half-mutated binding regions (BR1 and BR2). Human DHFR-EGFP with FMBR and human DHFR-EGFP with a half-mutated binding regions demonstrate increases in fusion protein levels upon exposure to MTX. b, batches from DG44 cells transfected with an FMBR, BR1, and BR2 human were exposed to 1 μM MTX or media without MTX for 24 and 48 h. Total cell lysate (50 μg) was loaded and probed with an anti-EGFP antibody. Equal loading was determined by Ponceau S staining (data not shown).

Fig. 4.

The only ATG within the binding region was mutated to ACG, which resulted in M140T. After exposure to 0.1 and 1 μM MTX for 24 and 48 h, human DHFR variant M140I resulted in increased levels of DHFR-EGFP fusion protein. Western blots were performed on total cell lysates from DG44 cells transfected with M140T. Cells were exposed to MTX or media without MTX in the presence of hypoxanthine and thymidine to prevent cell death due to MTX treatment. Total cell lysate (50 μg) was loaded and probed with an anti-EGFP antibody. Equal loading was determined by GAPDH antibody.

Fig. 5.

wt human DHFR restores the lack of feedback regulation in hamster and human DHFR S118A variant, both of which are not responsive to MTX-induced up-regulation. a, DG44 cells stably transfected with DHFR S118A-EGFP were exposed to 1 μM MTX for 24 h, demonstrating the lack of up-regulation of DHFR by MTX. DHFR S118A-EGFP levels were detected using antibody against EGFP. Equal loading was determined with GAPDH antibody. b, DG44 cells transfected with DHFR S118A-EGFP were transfected once more with Flag-tagged wt human DHFR. Although Flag-tagged wt human DHFR protein levels were detected with anti-Flag antibody, mutant DHFR S118A-EGFP levels were detected with anti-EGFP antibody. Five stable clones of doubly transfected cell lines were exposed to 1 μM MTX for 48 h, and DHFR protein was detected using Western blotting. Graph depicts the quantitative analysis of the data in b. The intensity of DHFR S118A-EGFP protein bands was normalized to the intensity of GAPDH. c, C85 cells transfected with hamster DHFR-EGFP were exposed to 0.l and 1 μM MTX for 48 h. Total cell lysate (50 μg) was loaded and probed with either DHFR antibody for endogenous levels of wt human DHFR and EGFP antibody for wt hamster-EGFP fusion protein. To control for equal loading, blots were stripped and reprobed with an α-tubulin antibody.

Fig. 6.

a, synthesis of wt human, hamster, and variants of human DHFR-EGFP using an in vitro reticulocyte translation assay. b, DHFR-EGFP protein levels of hamster is higher than the wt human DHFR-EGFP. Serial dilutions of hamster DHFR-EGFP lysates were compared with 30 μg of wt human DHFR-EGFP transfected cells. Western blots were probed with an anti-EGFP antibody. Loading was controlled using GAPDH antibody. c, DHFR activity of hamster and human DHFR variants that are not up-regulated by MTX are higher than the wt DHFR-EGFP. d, cytotoxicity of MTX to DG44 cells transfected with wt human and hamster DHFR-EGFP.

Fig. 7.

Model for divergent translational regulation of rodent and human DHFR. wt human DHFR mRNA translation is feedback-regulated through the interaction of DHFR protein with its cognate mRNA. The cis-acting regulatory elements on human DHFR mRNA were localized within the coding region, and the trans-acting regulatory domain was suggested to be within the NADPH binding site of DHFR. wt human DHFR protein has at least two conformations, one of which is bound preferentially to NADPH, and the other is bound to DHFR mRNA. These two conformers are in equilibrium and can interconvert. Binding of MTX to the binary complex of DHFR-NADPH shifts the equilibrium toward the NADPH-bound conformer, releasing DHFR mRNA to be translated. However, three variants of human DHFR protein (S118A, L22R, and E30A) and hamster are unable to bind their own cognate mRNA, which leads to increased synthesis of these proteins. Therefore, although binding of MTX to the wt human DHFR-mRNA complex leads to resumption of DHFR synthesis, hamster and three variants of human DHFR proteins are not induced by MTX.