Translational control of Nrf2 within the open reading frame

2.1 Recombinant constructs

A plasmid containing the cDNA of Nrf2 was obtained from Thermo fisher (accession no. BC011558 clone ID: 4548874) and was used as a template for PCR reactions. Also the plasmid pLVTHM (addgene.org clone 12247) was used as a template for eGFP PCR reactions.

All the recombinant constructs described in this work were cloned in the plasmid PLEX-MCS (Thermo fisher) that was modified to include in the C-term of the recombinant proteins, a strep tag II and a His 6X tag [13]. The recombinant constructs were created with the following primer sets, and contained, in the forward primer, a restriction site for BamHI (Underlined) plus a kozak sequence (lower case), and in the reverse primer a restriction site for AgeI (Underlined); the integrity of all the construct described was confirmed by sequencing. Nrf2 F: 5’ CGG GAT CCg ccg cca ccA TGA TGG ACT TGG AGC TGC C 3’ R: 5’ TCC CAC CGG TGT TTT TCT TAA CAT CTG GCT TCT T 3’; Δ17–32 Nrf2 F: 5’ CGG GAT CCg ccg cca ccA TGA TGG ACT TGG AGC TGC CGC CGC CGG GAC TCC CGT CCC AGC AGG ACA GTC GAG AAG TAT TTG ACT TCA GTC A 3’; Segment 1 F: 5’ CGG GAT CCg ccg cca ccA TGA TGG ACT TGG AGC TGC C 3’ R: 5’ TCC CAC CGG TCT CAA CCA GCT TGT CAT TTT CA 3’; Segment 2 F: 5’ CGG GAT CCg ccg cca ccAT GAC TAC CAT GGT TCC AAG TCC AG 3’ R: 5’ TCC CAC CGG TTC CAG GGG CAC TAT CTA GCT CTT 3’; Segment 3 F: 5’ CGG GAT CCg ccg cca ccA TGA GTG TCA AAC AGA ATG GTC CTA AA 3’ R: 5’ TCC CAC CGG TGT TTT TCT TAA CAT CTG GCT TCT T 3’; Segment1–2 F: 5’ CGG GAT CCg ccg cca ccA TGA TGG ACT TGG AGC TGC C 3’ R: 5’ TCC CAC CGG TTC CAG GGG CAC TAT CTA GCT CTT 3’; Segment 2–3 F: 5’ CGG GAT CCg ccg cca ccAT GAC TAC CAT GGT TCC AAG TCC AG 3’ R: 5’ TCC CAC CGG TGT TTT TCT TAA CAT CTG GCT TCT T 3’. All these PCR products were gel-purified (Promega), digested with BamHI and AgeI (Fermentas) and ligated into PLEX-MCS previously digested with the same enzymes.

The creation of the constructs containing eGFP fused to Segment 2 and Segment 3 was performed in three steps: First, a PCR product for eGFP containing a C-term His 6X followed by two stops codons and a KpnI recognition site was created with the primer set F: 5’ CGG GAT CCg ccg cca ccA TGG TGA GCA AGG GCG AG 3’ R: 5’ TCC CAC CGG TGG TAC CTT ACT AAT GAT GGT GAT GGT GGT GTC GAG ATC TGA GTC CGG ACT T 3’. This PCR product contained the recognition sites for BamHI and AgeI and was cloned into PLEX-MCS as described above to over express eGFP with C-term His tag. The same PCR product was used to create the fusion constructs eGFP-Segment 2 and eGFP-Segment 3 by using the KpnI recognition site. Second, a PCR product for Segment 2 and Segment 3 containing a KpnI recognition site in the 5’ was obtained with the following set of primers: KpnI-Segment 2 F: 5’ GGG GTA CCAC TAC CAT GGT TCC AAG TCC AG 3’ R: the primer described above for Segment 2; KpnI-Segment 3 F: 5’GGG GTA CCA GTG TCA AAC AGA ATG GTC CTA AA 3’ R: the primer described above for Segment 3. Third, the PCR products for eGFP, KpnI-Segment 2 and KpnI-Segment 3 were digested with KpnI and a ligation was performed between eGFP and Segment 2 and Segment 3. These ligations were used as templates to obtain the fusion clones eGFP-Segment 2 and eGFP-Segment 3 by using the Forward primer to amplify eGFP and the Reverse primers for Segment 2 and Segment 3 described above. These PCR products were digested with BamHI and AgeI and were cloned into PLEX-MCS.

The generation of the modified Segment 3 with all codons substituted with synonym codons was performed manually using a standard codon table as a reference. The construct containing the modified sequence was synthesized by IDT DNA technologies using gblock technology. This synthetic construct was fused with an eGFP PCR product obtained from the eGFP His construct described above using the primers F: 5’ ATA GAA GAC ACC GAC TCT ACT AGA GGA TCC GCC GCC ACC ATG GTG AG 3’ and R: 5’ GGG CGT CTT TGG GCC GTT TTG CTT AAC CGA TTA CTA ATG ATG GTG ATG GTG GT 3’. The gibson assembly method [14] with the gibson enzyme mix from New England Biolabs was used, and was then cloned into PLEX-MCS previously digested with BamHI-AgeI. The sequence of the modified Nrf2 segment 3 is as follows: TCGGTTAAGCAAAACGGCCCAAAGACGCCCGTCCACTCGTCAGGTGACATGGTCCAGCCACTGTCCCCCTCGCAAGGACAAAGTACGCATGTACACGACGCTCAGTGCGAAAATACCCCCGAAAAGGAGCTACCCGTGTCCCCCGGGCACAGAAAGACGCCCTTTACGAAGGATAAGCACTCCTCCAGGTTAGAAGCCCACCTAACGCGCGACGAGCTCCGAGCGAAGGCGTTACACATACCCTTTCCCGTGGAGAAGATAATAAATTTGCCGGTAGTCGATTTTAATGAGATGATGAGTAAGGAACAATTTAACGAGGCCCAGCTAGCGTTGATAAGGGACATCAGACGCCGAGGAAAAAACAAGGTAGCAGCGCAAAACTGTCGGAAGCGGAAGTTAGAGAACATCGTGGAGCTCGAACAGGACCTCGACCACCTAAAGGACGAGAAGGAGAAGCTCCTAAAGGAGAAGGGGGAGAACGATAAGTCATTGCATTTGCTAAAGAAGCAGTTGTCGACTTTGTACTTAGAGGTATTTTCTATGTTGCGGGACGAGGACGGCAAGCCCTACTCGCCCTCAGAGTATTCGCTCCAACAGACCCGAGACGGTAACGTCTTTCTAGTCCCTAAGTCCAAAAAACCCGACGTGAAAAAGAAT

To create a full length Nrf2 with the mutated segment 3 described above, two PCR fragments corresponding to a product containing the wild type sequence for segment 1 and 2 and the other containing the sequence of the mutated segment 3 were fused together using the gibson assembly mix (New England biolabs). The fragment containing segments 1–2 was obtained with the primer set: F: 5’ ATA GAA GAC ACC GAC TCT ACT AGA GGA TCC GCC GCC ACC ATG ATG GAC T 3’ R: 5’ GGG CGT CTT TGG GCC GTT TTG CTT AAC CGA TCC AGG GGC ACT ATC TAG CTC TT 3’ and the mutant segment 3 with the set: F 5’ TCG GTT AAG CAA AAC GGC 3’ R: 5’ GTT GGC GCA GCA GCC GGG GCA GCA ACC GGT ATT CTT TTT CAC GTC GGG TTT 3’. The fusion of these PCR fragments was performed in the same vector as described above.

2.2 Cell culture and transfections

HEK-293T cells were obtained from ATCC and were grown in Dulbecco’s minimal essential medium supplemented with 10% fetal bovine serum. The recombinant plasmids reported in this work were extracted using the Pureyield Maxiprep system from Promega and were transfected using Jetprime (Polyplus) following the manufacturer's recommendations without modifications. The protein expression levels were evaluated 48 hours after transfection with Western blotting or fluorescence laser scanning.

2.3 Western blotting

Transfected HEK-293T cells were lysed by using M-PER mammalian protein extraction reagent (Thermo scientific) containing Halt protease inhibitor cocktail (Thermo scientific). Lysates were centrifuged at 16,000 × g at 4 °C for 15 min. The supernatants (40 micrograms) from each sample were separated by SDS-PAGE and transferred into nitrocellulose membranes. The following antibodies were used: mouse monoclonal Anti strep-tagII labeled with HRP (Genescript Cat. A01742-100), mouse monoclonal anti His C-term labeled with HRP (Life technologies cat. R931-25) and mouse monoclonal anti beta actin labeled with HRP (Santa Cruz Biotechnologies Cat. sc-47778 HRP). All antibodies were diluted 1:5000 in 1× PBS with 0.1% tween 20 and 5% nonfat dry milk.

2.4 Fluorescence laser scanning

For fluorescence imaging, HEK-293T cells were grown and transfected in 48 well tissue culture plates. NucRed Live 647 (Life technologies) was added to label the cell nuclei following the manufacturer's recommendations. Cells fluorescence was determined using a Fujifilm FLA-5000 Laser scanner. The 473nm laser and the LPB filter was used for eGFP detection and the 635nm laser in combination with the LPR filter was used to detect nuclei fluorescence. Densitometry measurements were obtained with the Fujifilm image analysis software Multi Gauge.

2.5 Software Analysis

The codon adaptation Index (CAI) was calculated using the application developed by Puigbo et al. [15] available at http://genomes.urv.es/CAIcal/

3.1 The translation of the open reading frame of Nrf2 is low despite having a good codon usage frequency

The codon adaptation index (CAI) [16] is a measurement of codon bias that allows the comparison of the codons present in a specific gene versus a reference codon usage set from the organism in which the protein is expressed. This index ranges from 0 to 1 and correlates with protein translation efficiency. An index of 1 indicates that a gene uses the most common codons for a particular amino acid in the set. We found a CAI of 0.73 for Nrf2, suggesting a codon composition that is expected to be highly expressed.

In agreement with previous reports [9], we also found that while Nrf2 can be detected by western blot (Fig 1A), the expression is low, and is only slightly elevated if a degradation-resistant Nrf2 mutant previously described (Δ17-32aa) [17] is used for overexpression (Fig 1A). This low Nrf2 expression is more evident when compared to the recombinant expression with the same vector and transfection conditions of Grp78 (HSPA5), a protein that has a similar size and a similar CAI (0.77) (Fig 1B). These results suggest that the low expression is due the presence of an unidentified Keap-1 independent mechanism regulating the expression of Nrf2 within the ORF.

3.2 Nrf2 expression is regulated by a translational control mechanism within the open reading frame

Because there was no previous information suggesting the location of potential regulatory elements for protein translation within the ORF of Nrf2, we decided to explore the translation potential by dividing the whole transcript into three segments in order identify a segment with repressed translation. The Nrf2 ORF is 1815 bp excluding the stop codon and therefore the three segments were composed of the following base pairs: Segment 1=1–627bp, Segment 2=628–1158bp and Segment 3=1159–1815bp (Fig. 2A). Their length was selected according to the possibility of designing good primers pairs for PCR amplification. We also verified that the 3 segments have similar CAI (Segment1=0.71, Segment 2=0.75 and Segment 3=0.73), which indicated that their ability to be efficiently translated was similar.

To exclude the possibility of poor protein detection by fast proteosomal degradation, the constructs were overexpressed with and without the proteasome inhibitor MG132. We first verified that the three constructs were efficiently transcribed (Fig. 2B bottom panel). Next, we determined the expression levels of the 3 segments of Nrf2 by western blot with anti strep tag II antibody. We found that the expression of segment 1 was low (Fig. 2B lane 1), but was rescued with the use of the proteasomal inhibitor. This result is as expected because segment 1 contains the amino acids sequence that interacts with Keap1 to promote proteasomal degradation [9,17]. In contrast, the expression of segment 2 was elevated and was independent of the proteasomal degradation (Fig. 2B lane 2). Surprisingly, the expression of segment 3 could not be detected (Fig. 2B lane 3), even after the use of proteasomal inhibitor, suggesting the presence of an unknown mechanism preventing the expression of this segment. To corroborate this finding, we decided to create other constructs to evaluate the effect on protein expression by fusing segment 2, which we found to be highly over expressed, with segment 3. As a control, we also evaluated the translation of segment 1 fused together with segment 2. The expression of all the constructs was evaluated with and without the use of a proteasomal inhibitor. We found that while segment 1 dramatically reduced the expression of the fused segment 2 (Fig 2C lane 1), the expression can be rescued with the use of the proteasomal inhibitor. On the other hand we confirmed that segment 3 prevented the expression of segment 2 even with the inhibition of the proteasomal degradation (Fig 2C lane 3). Collectively, these results suggest that segment 3 contains a novel translational repressor mechanism that regulates the expression of Nrf2.

3.3 The regulation of the expression of Segment 3 is dependent on the mRNA sequence and not by the amino acids encoded by the sequence

To confirm that the mRNA sequence of segment 3 contains regulatory elements for protein translation, and to exclude the possibility that an unknown mechanism was promoting protein degradation by targeting amino acids present in the segment 3, we evaluated the effect of fusing eGFP with the mRNA sequences of segment 3. The experimental design included two stop codons in between the sequences of eGFP and segment 3 to prevent the translation of the amino acids encoded by segment 3 (Fig 3A). As a control, we generated a similar construct by fusing eGFP with segment 2 (Fig. 3A). The constructs were transfected into HEK-293T cells and eGFP was detected by western blot using an anti 6X-His tag included in the C-term of eGFP.

We found that the mRNA sequence of segment 2 did not alter the expression of eGFP (Fig. 3B lane 2). On the other hand, we verified that the segment 3 mRNA sequence dramatically reduced the translation of eGFP (Fig. 3B lane 3), even if the translation of the amino acids of segment 3 did not occur. Our results suggest that the mechanism inhibiting the translation of segment 3, alone or fused to other sequences is not by an unidentified protein degradation process.

3.4 Synonym mutations of Segment 3 reverse the translational repression

Next, we asked whether the translational repression of Segment 3 could be reversed by a mutant with synonymous substitutions of all the codons present in Segment 3. The experimental design was to fuse this mutant in a similar way to the eGFP study described above (Fig. 3A). We were able to create a Segment 3 sequence with synonym mutations that was only 61% similar to the wild type sequence. After transfection of this construct into HEK-293T cells, it was confirmed that the translational repression regulating the expression of Segment 3 could be completely reversed by changing the sequence of the mRNA without requiring alteration in the amino acids encoded by the sequence (Fig. 4A well in lane 3 vs well in lane 2). These mutations were also able to revert the translational repression when they were included into the full length Nrf2 open reading frame (Fig. 4B) and promoted a 6 fold increase in translation when compared to the wild type ORF sequence.