The conserved Trp114 residue of thioredoxin reductase 1 has a redox sensor-like function triggering oligomerization and crosslinking upon oxidative stress related to cell death

The selenoprotein thioredoxin reductase 1 (TrxR1) has several key roles in cellular redox systems and reductive pathways. Here we discovered that an evolutionarily conserved and surface-exposed tryptophan residue of the enzyme (Trp114) is excessively reactive to oxidation and exerts regulatory functions. The results indicate that it serves as an electron relay communicating with the FAD moiety of the enzyme, and, when oxidized, it facilitates oligomerization of TrxR1 into tetramers and higher multimers of dimers. A covalent link can also be formed between two oxidized Trp114 residues of two subunits from two separate TrxR1 dimers, as found both in cell extracts and in a crystal structure of tetrameric TrxR1. Formation of covalently linked TrxR1 subunits became exaggerated in cells on treatment with the pro-oxidant p53-reactivating anticancer compound RITA, in direct correlation with triggering of a cell death that could be prevented by antioxidant treatment. These results collectively suggest that Trp114 of TrxR1 serves a function reminiscent of an irreversible sensor for excessive oxidation, thereby presenting a previously unrecognized level of regulation of TrxR1 function in relation to cellular redox state and cell death induction.

complete EDTA-free protease inhibitor cocktail (Roche, USA). Lysates were cleared by centrifugation at 16,000 × g at 4 °C for 10 min. Protein concentration was determined by the Bradford method (Bio-Rad, USA) using bovine serum albumin (BSA, Sigma, USA) as a standard.

Transient knockdown of TrxR1 and 75 Se labeling of cellular proteins
HCT116 cells (5 × 10 4 cells/well) were seeded in 6-well plates. For small interfering RNA (siRNA) transfection, cells were incubated with 10 nM siRNA duplexes for 24 h, TrxR1 siRNA sequences (Siseq1 and Siseq2) and the scrambled control (mock) used. Transient transfection procedures have been previously described in (Eriksson et al, 2009). The indicated drugs were added 48 h post siRNA transfection. However, for the 75 Se-incorporation experiment, approx. 1 × 10 5 cells/well were plated in the presence of 1.5 µCi [ 75 Se]-selenite. After 24 h siRNA transfection, the medium was removed and fresh medium without 75 Se was added. Cells were subsequently incubated with the indicated drugs.

Endpoint insulin assay of cellular samples
The end-point Trx-dependent insulin assay (Arnér & Holmgren, 2001) was modified and applied to 96-well plates to measure the thioredoxin reductase activity of cell lysates. Briefly, 10 µl of fractions was incubated with 20 µM of wild type human Trx1, 297 µM insulin, and 1.3 mM NADPH, in presence of 85 mM HEPES buffer containing 13 mM EDTA (pH 7.5) at 37 °C for 90 min, in a total volume of 50 µl. The reaction was stopped by addition of 200 µl 7.2 M guanidine-HCl in 200 mM Tris-HCl, pH 8.0, containing 1 mM DTNB. The Trx-dependent newly formed thiols in the reduced insulin products were then determined at 25 °C by measuring the absorbance at 412 nm (extinction coefficient of 13,600 M -1 cm -1 ) using a VersaMax spectrophotometer (Molecular Devices, USA) with a background absorbance reference for each samples containing all components except Trx, incubated and treated in the same manner.

Site-directed mutagenesis of recombinant rat TrxR1
Wild-type TrxR1 was generated as previously described using a recombinant system that provides for incorporation of the active site Sec (Böck et al, 1991). Several previously characterized sitedirected mutants of TrxR1 that have been studied earlier, such as C59S (Cenas et al, 2004), C64S (Cenas et al, 2004), C59S/C64S (Anestal et al, 2008), U498C (Zhong et al, 2000), Y116I and The reverse primer was C-term-r, 5'-gcggtctcggactggaggatgtctcc-3'. Phusion ® High-Fidelity PCR Master Mix (Thermo Fisher Scientific, Waltham, USA) was utilized for the inverse PCRs and all molecular cloning was performed as described (Xu & Arner, 2012). In brief, together with the addition of 20 pmol of each primer and 10 ng of template plasmid in a total volume of 50 µl, PCR reactions were initiated at 98 °C for 2 min, followed by 30 cycles of amplification (98 °C, 10 s; 65 °C, 45 s; 72 °C, 3 min) and extended at 72 °C for additional 10 min. The resulting PCR products were analyzed by 1% agarose electrophoresis and the targeted PCR band was cut and then purified using the Gel Extraction Kit (Qiagen, Germany), digested with Eco31 I and Dpn I (Thermo Fisher Scientific, Waltham, USA) at 37 °C for 30 min, re-purified with the PCR purification kit (Qiagen, Germany), and ligated using T4 DNA ligase (Thermo Fisher Scientific, Waltham, USA) at 22 °C for 2 h. Finally, the DNA constructs were then transformed into E. coli BL21 (DE3) gorcompetent cells (Tet + ), and then DNA sequencing (GATC Biotech, Konstanz, Germany) verified the mutations.

Electrophoresis
TrxR1 samples were analyzed by SDS PAGE and native PAGE using either NuPAGE TM 4-12% Bis-Tris SDS gel or NativePAGE TM 4-16% Bis-Tris gel (Life Technologies, USA). The protein bands were stained using PhastGel TM Brilliant Blue R250 (GE Healthcare Life Sciences, Uppsala, Sweden), destained with 10% acetic acid and 30% methanol and then documented using a Bio-Rad ChemiDoc XRS scanner (Bio-Rad, USA). Band intensity analysis was performed using the 1-D Image Analysis software (Bio-Rad, USA).

Western blotting
Protein bands in the PAGE gel were electroblotted to a nitrocellulose membrane (Millipore, USA) at 30 V at 4 °C for 3 h. The membrane was blocked in 5% fat-free milk at 4 °C for 1 h and incubated in 1% BSA containing the anti-TrxR1 19A1 monoclonal antibody (dilution 1:1000, v/v) for 2 h. It was then washed with PBS-T solution (PBS, 0.1% Tween-20, pH 7.5) for 3 times of 5 min each, and then transferred into 1% BSA containing secondary goat anti-mouse IgG conjugated to horseradish peroxidase (dilution 1:500, v/v) for 1 h. The membrane was washed with PBS-T solution 3 times (5 min each). Finally, antigen-antibody binding was detected using the Western Lightning TM Chemiluminescence Reagent kit (Perkin Elmer, USA).

Se Autoradiography
Radiolabeling using 1 µCi [ 75 Se]-selenite was utilized for visualization of Sec incorporation into recombinantly expressed protein, using previously described protocols (Böck et al, 1991). 0.5 mM IPTG-induced protein samples were prepared and used for analyses with reducing SDS-PAGE gel and autoradiography with documentation using a GE Typhoon TM FLA 7000 Biomolecular Imagers and ImageQuant TM TL software version 7.0 (GE Healthcare Life Sciences, Uppsala, Sweden).

TrxR activity assays using 96-well microtiter plate
Enzymatic activities of purified TrxR1 variants were determined using four different assays in 96well microtiter plate format. a) Insulin-coupled Trx reduction assay (Arnér & Holmgren, 2001).

Circular Dichroism (CD) Spectrum Analysis
Circular dichroism spectrum analysis was carried out with a 1 cm quartz cuvette at 20 °C using a Jasco TM J-810 Spectropolarimeter with Jasco TM PTC-423S single position Peltier thermostatted cell holder (Jasco, Japan). Each sample (800 µl total volume) contained 100 nM TrxR1 variants in 1 mM Tris-HCl buffer containing 40 nM EDTA (pH 7.5). The CD spectra scanning were scanned from 195 nm to 260 nm with a step size of 1 nm. The buffer spectrum was subtracted from the sample spectra.

Thermostability Analysis
Thermostability of TrxR1 variants was measured in parallel by applying a light-scattering-based methodology (Vedadi et al, 2006) with a multi-well format StarGazer-384 TM instrument (Harbinger Biotechnology and Engineering Corp., Toronto, Canada). TrxR1 samples in flat-bottomed 384-well black plates were subjected to gradually increased temperature from 25 °C to 80 °C at 1 °C per min. 30 µl enzyme solution was placed inside of each well and then mixed with 20 µl TE buffer (pH 7.5).
Mineral oil (40 µl) (M1180, Sigma, USA) was added to the wells to protect sample evaporation.
Images of scattered light were taken every 30 s to monitor protein aggregation.

Selenium Determination
Selenium contents of TrxR1 variants were determined by elemental analysis (ALS Scandinavia AB, Luleå, Sweden), using two separate samples per enzyme. The results were used for normalization of catalytic turnover numbers with regards to selenium content as described in Table E4.

References for the supplementary methods section:
Anestal K, Prast-Nielsen S, Cenas N, Arner ES (2008)  Zhong L, Arner ES, Holmgren A (2000) Structure and mechanism of mammalian thioredoxin reductase: the active site is a redox-active selenolthiol/selenenylsulfide formed from the conserved cysteine-selenocysteine sequence. Proc Natl Acad Sci U S A 97: 5854-5859 Scheme S1. Scheme S1. Dimers and oligomers of TrxR1 as modulated by the Trp114 residue. The different forms of TrxR1 characterized in this study are here schematically summarized. The native form of TrxR1 is mainly present as a dimer, with its two subunits non-covalently associated in a head-to-tail configuration and having the Trp114 residue fully surfaceexposed, although a minor fraction of the enzyme species can be seen as non-covalently associated into a tetramer in solution (top of the figure, equilibrium shifted towards the dimeric form). Upon oxidation, as seen either in vitro or in cells treated with RITA but prevented by NDGA, the Trp114 residue easily becomes oxidized, which shifts equilibrium towards the tetramer (lower part of the figure) and can subsequently lead to covalent crosslinking between two modified Trp114 residues (lower right, crosslink indicated with a cartoon of the electron density found in the tetramer crystal structure). The resulting forms of TrxR1 detected in solution or in reducing denaturing SDS-PAGE are indicated below the scheme. TrxR1 oligomers together with the regular dimer were used to study the DTT-resistance. After incubation with DTT (5 to 100 mM) for 2 h at room temperature, the samples were mixed with 4 × NativePAGE™ Sample Buffer and then load onto non-reducing NativePAGE™ Novex ® 4-16% Bis-Tris gels (Life Technologies, USA). The results show that dissociation of TrxR1 from larger oligomers to smaller oligomers, and finally to dimers, is concentration-dependent and occurs in a step-wise manner.
Supplementary Figure S3 Supplementary Figure S3. Partial sequence alignment of animal TrxRs. The TrxRs used for the sequence alignment were as follows: rat TrxR1 (Swiss-Prot accession no.O89049), mouse TrxR1 (no.Q9JMH6), human TrxR1 (no.Q16881), pongo TrxR1 (no.Q5NVA2), pig TrxR1 (no.Q9MYY8), bovine TrxR1 (no.O62768), horse TrxR1 (NP_001131081; UPI000185554C), dog TrxR1 (NP_001116145; UPI00016DBCE0), rabbit TrxR1 (NP_001243891; UPI000252888A), Rhesus monkey TrxR1 (NP_001243204; UPI000246789D), white-tufted-ear marmoset TrxR (NP_001243348; UPI00024A9A3E), chimpanzee TrxR1 (NP_001116141; UPI00016DBCDF), platypus TrxR1 (no.F6RSS8), chicken TrxR1 (NP_001025933; UPI00015294EF), zebra finch bird TrxR1 (NP_001257893; UPI00027A7715), zebrafish TrxR1 (no.A8WGN7), C. elegans TrxR1 (no.Q17745), rat TrxR2 (no.Q9Z0J5), mouse TrxR2 (no.Q9JLT4), human TrxR2 (no.Q9NNW7), pig TrxR2 (no.D2K6G0), bovine TrxR2 (no.Q9N2I8), mouse TGR (no.Q99MD6), human TGR (no.Q86VQ6), dog TGR (NP_001116250.1; UPI000170CD7D), chicken TGR (NP_001116249; UPI000170CD7C), and zebrafish TGR (Q7T2CB). Fully conserved amino acid residues are highlighted in red, such as His 96 , Trp 98 , Gln 106 , His 108 , Ser 111 , Leu 112 , Asn 113 , Trp 114 , Gly 115 , Arg 117 , Leu 120 and Val 125 . Among those amino acid residues found in all organisms, Trp 114 (marked by a black arrow) is clearly highly conserved. The red arrow indicates His 116 that is found in mitochondrial TrxR2 instead of Tyr 116 present in cytosolic TrxR1 or testis-specific TGR (blue arrow).  Figure S4. Circular dichroism (CD) spectrum analysis. Circular dichroism spectrum analysis was carried out with a 1 cm quartz cuvette at 20 °C using a Jasco TM J-810 Spectropolarimeter with Jasco TM PTC-423S single position Peltier thermostated cell holder (Jasco, Japan). Each sample (800 µl total volume) contained 100nM TrxR1 variants in 1mM Tris-HCl buffer containing 40 nM EDTA (pH 7.5). The CD spectra were scanned from 195 nm to 260 nm with a step size of 1 nm. The buffer spectrum was subtracted from the sample spectra. Secondary structure calculations were subsequently done using the DICHROWEB (http://dichroweb.cryst.bbk.ac.uk/) online tool based on the K2d method, with the results shown in the Table below the graphs. For comparison, the top row of the table lists the secondary structure parameters of our previously determined crystal structure of the wild-type dimeric enzyme (PDB entry 3EA0). All variants analyzed here except W114E displayed good agreement with the secondary structure found in the wildtype dimeric crystal structure, although all displayed normalized root mean square deviation (NRMSD) values above 0.1, thus suggesting some discrepancy between the predicted secondary structures and calculated values. This discrepancy was most pronounced for the W114E variant, suggesting that the calculated secondary structure elements for this variant were less reliable than for the other proteins. n.a.=not applicable  IPTG-induced samples with either the same total protein amount (a), or the same total DTNB reductase activities (b), or the same total Trx-linked insulin reduction activities (c) were loaded onto separate reducing SDS-PAGE gels. W114E showed a strong radioactive signal when compared to w.t. TrxR1 and its variants, especially for the gel loaded on the basis of equivalent total activity in the insulin assay. These results indicate that the W114E mutant had high Sec incorporation at its C-terminus, but that the significant loss of TrxR1 function is mainly due to the W114E substitution. Seen as a band in both the protein lysate and the selenium labeling is also the chloramphenicol acetyl transferase (CAT), which is a product of the pSUABC plasmid used for efficient Sec incorporation.

Supplementary Figure S4
Supplementary Figure S8 Supplementary Figure S8. PAGE analyses of Trp114 mutants of rat TrxR1. After 0.5 mM IPTG induction at 24 °C for 24 h, four Trp114 mutants of rat TrxR1 were expressed in E. coli BL21 (DE3) gorhost strain and further analyzed by reducing SDS-PAGE (a) and non-reducing native PAGE (b) In (a), the 55 kDa band represents the size of the TrxR1 subunit polypeptide. In (b), wild-type rat TrxR1 exhibited clear aggregation behaviour and formed non-covalently linked oligomers in addition to the dimer. Replacement of Trp114 with Arg, Glu or Gly significantly lowered the propensity to form oligomers as in the analyses on non-reducing native PAGE. When Trp was replaced by the aromatic residue Phe, the oligomerization tendency of the enzyme was intermediate.
Supplementary Figure S9 Supplementary Figure S9. 3D fluorescence excitation emission spectra of various TrxR1 variants and free FAD in solution, as well as 2nd harmonic of EPR X-band spectra. The indicated TrxR1 variants (see on-line methods for full descriptions) were dissolved in TE buffer (pH 7.5) and protein concentrations were adjusted to 6.8 µM based on FAD absorbance at 463 nm. In 200 µl (total volume) the TrxR1 samples (or free FAD at a concentration of 2 µM) were loaded into a flat-bottomed 96-well black plate (Thermo Fisher Scientific, USA) for fluorescence measurements using the EnSpire TM Fluorescence Analyzer (Perkin Elmer, USA). The emission spectrum ranging from 250 nm to 650 nm was obtained by exiting the samples first at the excitation wavelength of 230 nm and subsequently the excitation wavelength was increased by 10-nm intervals up to 510 nm. The data obtained for all excitation wavelengths were subsequently used to plot the fluorescence emission spectra in 3D format, with 'X-axis' as the excitation wavelength (Ex), 'Y-axis' as the emission wavelength (Em), and the 'Z-axis' as fluorescence intensity. The TE buffer 3D spectrum was subtracted from all samples. The lower right figure displays the 2 nd harmonic of the EPR X-band spectra of wtTrxR, W114R, and Truncated (Sec-minus) TrxR1 variants, where all enzymes were reduced with NADPH to the EH4 state and the EPR signals normalized to the same intensity before the 2 nd harmonic was generated so that spectral features could be directly compared. Tables   Table S1. Mass spectrometry analysis of protein bands immunoprecipitated with TrxR1 specific antibodies after treatment of HCT116 cells with RITA. HCT116 cells were treated with 1 µM RITA for 8 h, following which protein in 540 µl eluate was pulled down from cell extracts using immunoprecipitation with anti-TrxR1 antibodies and analyzed on reducing SDS-PAGE, whereupon bands at 52 kDa (band 1), 55 kDa (band 2), 60 kDa (band 3) and 110 kDa (band 4), visualized by Coomassie staining (see Fig. 1a). These were subjected to in-gel tryptic digestion and analyses by mass spectrometry using MALDI, which confirmed the presence of TrxR1 in all samples and showed excellent peptide coverage for the applicable peptide mass range 200 Da). The peptide containing the Trp114 residue could not be found in band 4 and was present in several oxidation states in the other bands. See main text for further details. 1 Amino acid residues of human TrxR1 are re-numbered according to the rat TrxR1 sequence (also see Fig. S3). 2 Theoretical peptide masses (MH+) were obtained in using the online MS-digest software from UCSF (http://prospector.ucsf.edu/prospector/). 3 M is Met, methionine residue; W is Trp, tryptophan residue; ox is oxidation; 1 or 2 indicate that the peptides were modified once or twice, respectively. 4 Detected peptide masses (MH+) in MADLI and ESI-TOF are matching to human cytoplasmic thioredoxin reductase 1 (Swiss-Port accession no. Q16881). 5 Samples analyzed in MALDI corresponding to the protein bands of these apparent molecular weights as identified in Fig. 1a. 6 Lack of peptide mass number shows that the corresponding mass could not be found in the indicated sample. Residues in disallowed regions 0 Figures in parentheses correspond to the highest resolution shell * 5% of reflections have been used for the monitoring of the refinement Table S3. Mass spectrometry analysis of tryptophan oxidation in TrxR1 tetramer and dimer. The dimer and tetramer bands of TrxR1 were cut from the non-reducing native PAGE gel and followed by ingel tryptic digestion, subsequently used for mass spectrometry analyses. Both MALDI-TOF and ESI-TOF mass spectra were combined to identify the distinct peptide printing from the TrxR1 tetramer and dimer preparation. Trp114 oxidation was observed in the TrxR1 tetramer but not in the dimer.  Table S4. Normalization of turnover of TrxR1 variants for selenium content. The turnover numbers of the TrxR1 variants (panel a) were adopted form Table 1 (main text) to be normalized for their selenium content, determined in duplicate using elemental analysis (results shown as percentages in panel a).

Supplementary
Thereafter the selenium content values were used to normalize the turnover numbers as shown in panel b. The selenium contents of the purified TrxR1 variants were determined by elemental analysis at ALS Scandinavia AB, Luleå, Sweden and were calculated as the estimated percentages of the TrxR1 subunits in each TrxR1 variant that contained selenium.