key: cord-0861479-d0zhx2zl
authors: e Costa, Rute A.P.; Granato, Daniela C.; Trino, Luciana D.; Yokoo, Sami; Carnielli, Carolina M.; Kawahara, Rebeca; Domingues, Romênia R.; Pauletti, Bianca Alves; Neves, Leandro Xavier; Santana, Aline G.; Paulo, Joao A.; Aragão, Annelize Z.; Heleno Batista, Fernanda Aparecida; Migliorini Figueira, Ana Carolina; Laurindo, Francisco R.M.; Fernandes, Denise; Hansen, Hinrich P.; Squina, Fabio; Gygi, Steven P.; Paes Leme, Adriana F.
title: ADAM17 cytoplasmic domain modulates Thioredoxin-1 conformation and activity
date: 2020-09-24
journal: Redox Biol
DOI: 10.1016/j.redox.2020.101735
sha: b7b1e0a2f4cfd42c34e436c763b07e66e1133c26
doc_id: 861479
cord_uid: d0zhx2zl

The activity of Thioredoxin-1 (Trx-1) is adjusted by the balance of its monomeric, active and its dimeric, inactive state. The regulation of this balance is not completely understood. We have previously shown that the cytoplasmic domain of the transmembrane protein A Disintegrin And Metalloprotease 17 (ADAM17cyto) binds to Thioredoxin-1 (Trx-1) and the destabilization of this interaction favors the dimeric state of Trx-1. Here, we investigate whether ADAM17 plays a role in the conformation and activation of Trx-1. We found that disrupting the interacting interface with Trx-1 by a site-directed mutagenesis in ADAM17 (ADAM17cyto(F730A)) caused a decrease of Trx-1 reductive capacity and activity. Moreover, we observed that ADAM17 overexpressing cells favor the monomeric state of Trx-1 while knockdown cells do not. As a result, there is a decrease of cell oxidant levels and ADAM17 sheddase activity and an increase in the reduced cysteine-containing peptides in intracellular proteins in ADAM17cyto overexpressing cells. A mechanistic explanation that ADAM17cyto favors the monomeric, active state of Trx-1 is the formation of a disulfide bond between Cys(824) at the C-terminal of ADAM17cyto with the Cys(73) of Trx-1, which is involved in the dimerization site of Trx-1. In summary, we propose that ADAM17 is able to modulate Trx-1 conformation affecting its activity and intracellular redox state, bringing up a novel possibility for positive regulation of thiol isomerase activity in the cell by mammalian metalloproteinases.

The activity of Thioredoxin-1 (Trx-1) is adjusted by the balance of its monomeric, active and its dimeric, inactive state. The regulation of this balance is not completely understood. We have previously shown that the cytoplasmic domain of the transmembrane protein A Disintegrin And Metalloprotease 17 (ADAM17cyto) binds to Thioredoxin-1 (Trx-1) and the destabilization of this interaction favors the dimeric state of Trx-1. Here, we investigate whether ADAM17 plays a role in the conformation and activation of Trx-1. We found that disrupting the interacting interface with Trx-1 by a site-directed mutagenesis in ADAM17 (ADAM17cyto F730A ) caused a decrease of Trx-1 reductive capacity and activity. Moreover, we observed that ADAM17 overexpressing cells favor the monomeric state of Trx-1 while knockdown cells do not. As a result, there is a decrease of cell oxidant levels and ADAM17 sheddase activity and an increase in the reduced cysteine-containing peptides in intracellular proteins in ADAM17cyto overexpressing cells. A mechanistic explanation that ADAM17cyto favors the monomeric, active state of Trx-1 is the formation of a disulfide bond between Cys 824 at the C-terminal of ADAM17cyto with the Cys 73 of Trx-1, which is involved in the dimerization site of Trx-1. In summary, we propose that ADAM17 is able to modulate Trx-1 conformation affecting its activity and intracellular redox state, bringing up a novel possibility for positive regulation of thiol isomerase activity in the cell by mammalian metalloproteinases.

Trx-1 is a key component of the cellular redox system, also comprising NADPH and thioredoxin reductase (TrxR) that provides electrons to various enzymes [1] . This system maintains the redox state of the cell, in concert with other enzymatic and non-enzymatic antioxidant systems, such as superoxide dismutase (SOD1 and SOD2), and glutathione. The evolutionarily conserved metalloproteinase ADAM17, originally discovered as TNFαconverting enzyme (TACE), is the major sheddase for EGF receptor ligands and various surface proteins [2] .

Both ADAM17 and Trx-1 are involved in many physiological and pathological processes in the cell. ADAM17 is particularly implicated in inflammation processes [3, 4] and the protection of the central nervous system [5] . Also, inhibitors of ADAM17 and matrix metalloproteases (MMP) serve as beneficial adjuvants in bacterial meningitis therapy. Besides, the shedding of growth factors, ADAM17 is necessary for tumor progression and growth, and also contributes to inflammation in tumors [5, 6] .

Another interesting mechanism that has recently been proposed for ADAM17 concerning its role in inflammatory diseases, is the possible connection with SARS-CoV-2 effects by cytokine storm generation. It has been proposed that the possible molecular mechanism that leads to the acute respiratory distress syndrome (ARDS) in COVID-19 patients is caused by the massive release of proinflammatory cytokines, known as cytokine storm [7] . The major contributors for cytokine storm are the proinflammatory cytokines IL-6 and TNF-α [8, 9] . ADAM17 also known as TACE (Tumor necrosis factor-alpha converting enzyme) can be one of the main players involved in this process. It is well known that ADAM17 affects the biology of TNFα and IL-6 [10] [11] [12] [13] , leading to the speculation that inhibition of ADAM17 might present a positive response in autoimmune diseases [11] . With that, some studies have suggested that potential treatments for SARS-CoV-2 most destructive complications could help improve the life expectancy of the infected patients [7, 8, [14] [15] [16] . Taking this rationale into account, ADAM17 inhibitors, such as Trx-1, could be promising strategies for future studies to minimize the ADAM17 extracellular role in shedding and possibly cytokine storm effect observed in patients infected with SARS-CoV-2.

Furthermore, regarding the inflammation process, Trx-1 is an oxidative stress-limiting protein with anti-apoptotic properties and shows an ambiguous function in terms of inflammation. It has been shown that Trx-1 might exert atheroprotective effects by promoting macrophage differentiation into the M2 antiinflammatory phenotype [17] . The ability of Trx-1 to promote differentiation of macrophages into an alternative, anti-inflammatory phenotype may explain its protective effects in cardiovascular diseases, suggesting a link between oxidative stress and cardiovascular diseases. However, it has also been shown that the Trx-1 pathway enables nuclear NF-κB DNA-binding and thereby pro-inflammatory responses in monocytes and dendritic cells [18] . Moreover, independent of this activity, Trx-1 is critical for NLRP3 inflammasome activation and IL-1β production in macrophages by detoxifying excessive ROS levels [19] [20] [21] . Inflammation often complicates diseases associated with oxidative stress. It has been shown that inflammatory macrophages release proteins with specific forms of cysteine oxidation to disulfides, particularly glutathionylation [18, 22, 23] .

Redox proteomics identified peroxiredoxin 2 (PRDX2) as a protein released in glutathionylated form by inflammation both in vivo and in vitro, and extracellular PRDX2 then triggers the production of TNF-α, one of the main substrates of ADAM17 [22] . Together, these studies indicate that redoxdependent mechanisms, in an oxidative cascade, can induce inflammation and that targeting Trx1 may be exploited to treat inflammatory diseases, in the same manner that has been suggested for ADAM17. This leads to the suggestion that by targeting the Trx-1/ADAM17 interaction it is possible to modulate the downstream inflammation scenario promoted by different diseases. However, in the past, unspecific elimination of ROS by the use of low molecular mass antioxidant compounds was not successful in counteracting disease initiation and progression in clinical trials [18] .

Trx-1 is conserved among species and contains 105 amino acid residues. Of its five cysteine residues, two are part of the catalytic site (Cys 32 and Cys 35 ) and the remaining three (Cys 62 , Cys 69 , and Cys 73 ) are available for physiological modification. Cys 62 and Cys 69 are potential sites for S-nitrosylation but can also form a disulfide bond with each other. Cys 73 is localized on a hydrophobic patch of the protein surface and can contribute to an intermolecular disulfide bond (Trx-1 homodimer) but is also a site for glutathionylation (GS-ylation).

Homodimerization (Cys 73 -Cys 73 ) suppresses the functionality of Trx-1, since the dimer hides the catalytic site of the enzyme, rendering it no longer available as a substrate for thioredoxin reductase [24] . The oxidation or reduction of this bridge is dependent on the redox state of the cell [25] . Thus, the Trx-1 monomer-dimer balance is important to regulate the catalytic activity of Trx-1 [26] .

Previously, we described the interaction between ADAM17 and Trx-1 [27] , which has a negative effect upon ADAM17 activity. This interaction is in the opposite location of the catalytic site of Trx-1. A recent work from our group [28] has shown that ADAM17/Trx-1-interacting interface was destabilized by a mutation in Trx-1 (Trx-1 K72A ), favoring the Cys 73 -Cys 73 ' homodimeric form of Trx-1, which is the inactive state of Trx-1. Therefore, we hypothesized whether the interface of ADAM17/Trx-1 may also play a role in the Trx-1 activity.

Through different biochemical strategies, we have demonstrated in this study that i) ADAM17cyto WT affects Trx-1 activity by increasing its disulfide reductase activity, its reduced state and its monomer state; ii) the overexpression of ADAM17cyto WT or ADAM17 fulllength increases the monomeric state of recombinant and endogenous Trx-1, respectively; iii) lower cell oxidant levels and ADAM17 activity are observed in the presence of ADAM17cyto WT cells; and finally, iv) ADAM17cyto may indirectly regulate other intracellular proteins. These data bring new insights into the role of ADAM17 in modulating indirectly the cell redox processes through Trx-1.

ADAM17cyto mutant (ADAM17cyto F730A ) showed a decrease in the interaction with Trx-1, an increase in the oxidant levels, and a decrease in cysteine reactivity and redox activity when compared to ADAM17cyto WT .

Considering the evidences of the dimer and monomer forms of Trx-1 [24, 28] , we studied whether ADAM17 contributes with these states according to the described experimental design (Fig. 1) .

In order to evaluate the role of ADAM17 interaction in the Trx-1 activity in the cell, we have generated a mutation of ADAM17 (ADAM17cyto F730A ) in which the amino acid substitution was predicted by in silico analysis to be important for the maintenance of the interacting interface [27] . According to the solid-phase binding assay, the interaction of Trx-1 decreased in the presence of ADAM17cyto F730A compared to ADAM17cyto WT ( Fig.  2A) .

To evaluate whether the interaction between Trx-1 and cytoplasmic domain of ADAM17 affected Trx-1 activity, several experiments were performed to address this effect. Firstly, we observed by a DTNB assay more reduced cysteines in Trx-1 in the presence of ADAM17cyto WT compared to ADAM17cyto F730A (Fig.  2B) . Also, we observed that ADAM17cyto WT and ADAM17cyto F730A presented cysteine reactivity, showing higher reactivity in the former. Although these results indicate that ADAM17cyto WT may present more exposed cysteines than the mutant, we performed circular dichroism and J o u r n a l P r e -p r o o f it was not observed any significant changes in the secondary structure and folding properties of these proteins (Supplementary Figure 1) . This difference between ADAM17cyto WT and ADAM17cyto F730A alone can be due to the protein flexibility [27] .

Besides, we evaluated if ADAM17cyto WT could directly increase Trx-1 reducing activity upon Di-E-GSSG substrate. We first incubated Trx-1 with different ADAM17cyto WT and ADAM17cyto F730A protein concentrations (1, 2.5, and 5µM) and the fluorescence resulted from the substrate reduction was measured. Figure 2C shows that the concentration of ADAM17cyto WT above 2.5µM was able to promote an increase in the substrate reduction (p<0.05, Student´s t-test), while ADAM17cyto F730A did not have this same effect on Trx-1 reducing activity. The results showed that in the presence of ADAM17cyto WT , the reaction with the substrate was increased, compared to Trx-1 alone and Trx-1+ADAM17cyto F730A , strongly suggesting that ADAM17cyto WT interaction with Trx-1 led the thiol reductase enzyme to a more active form.

To investigate whether ADAM17 binding was able to affect the catalytic and reductase activity of Trx-1, a cysteine reactivity assay was performed with ADAM17cyto WT and its mutant. To define the best concentration of H 2 O 2 and Trx-1 used in this assay, a standard curve was performed ( Supplementary  Figure 2A- Figure 2C) . After incubation of the recombinant proteins with H 2 O 2 and Amplex Red ® probe (Supplementary Figure 2) , we detected more H 2 O 2 consumption in the presence of Trx-1+ADAM17cyto WT compared to Trx-1+ADAM17cyto F730A or Trx-1 alone (Fig. 2D) . Also to evaluate whether ADAM17cyto WT modulated Trx-1 cysteine reactivity by interfering with the oxidation-reduction state of the catalytic site (Cys 32 -Cys 35 ) or the dimeric state (Cys 73 -Cys 73 ), the cysteine reactivity assay was performed with ADAM17cyto WT and Trx-1 mutants (Trx-1 K72A and Trx-1 C32/35S ). As expected, the activity of catalytic site mutant was not changed, but we showed that Trx-1 K72A , a catalytically inactive mutant due to its dimeric form [28] , could have its cysteine reactivity increased by the incubation with ADAM17cyto WT , suggesting the ability of ADAM17cyto WT in the increase of Trx-1 activity acting upon the site of dimerization (Cys 73 -Cys 73 ) (Fig. 2D) .

ADAM17cyto WT directly increased the monomeric state levels of Trx-1.

To evaluate whether ADAM17cyto WT could increase the Trx-1 monomeric state, we next incubated recombinant Trx-1-HA protein with increasing concentrations of ADAM17cyto WT (0, 3, 6, 12µM). Then, the proteins were separated on a non-reducing SDS-PAGE and Coomassie Blue stained. Afterward, the bands were digested with trypsin and evaluated by LC-MS/MS to perform the relative semiquantification of peptides in the Trx-1 monomer and dimer bands (Fig.3A) . This result indicated that upon the increase of ADAM17cyto WT concentration, there was an increase in Peptide Spectrum Matches (PSM) of Trx-1 in the respective Trx-1 monomer band, while decreased PSMs of Trx-1 in the related Trx-1 dimeric band (Fig.  3A, Supplementary Table 1) .

To evaluate the propensity to dissociate the dimer in the presence of ADAM17cyto WT , both recombinant proteins were incubated in a 1:1 ratio and analyzed by analytical ultracentrifugation.

The sedimentation assays show that Trx-1 alone is formed mainly by dimers (~25kDa, with large peak width), but in the presence of ADAM17cyto WT , two distinct peaks could be observed, indicating monomeric Trx-1 and homo-and heterodimers of Trx-1 and ADAM17cyto WT complex ( Fig. 3B Table 3 ). The results showed that the Cys 73 containing peptides increased in the presence of ADAM17cyto WT , in a concentration-dependent manner, suggesting an increase in Trx-1 monomeric state ( Fig. 3C and Supplementary Table 4 ). We observed that ADAM17cyto F730A followed a similar tendency, but with lower intensity due to the lower ability to bind Trx-1, as previously demonstrated ( Fig. 2A) .

ADAM17cyto WT modulated the Trx-1 dimer and monomer ratio.

We next evaluated if the interaction between Trx-1 and ADAM17cyto WT can be dependent on disulfide bond formation. For that, we evaluated the 30kDa-dimer band indicated in Fig. 3A by datadependent acquisition (DDA) at the four different ADAM17cyto WT concentrations. We identified linked Trx-1 and cytoplasmic domain peptides (Fig. 4A) . The Trx-1 peptide 73 CMPTFQFFK 81 was linked with one trypsin missed cleavage ADAM17cyto peptide 816 VDSKETEC 824 (m/z 813.83, +3) and also with full-cleavage ADAM17cyto peptide 820 ETEC 824 (m/z 813.83, +2). Note that Cys 824 is the last amino acid in the ADAM17 sequence and Cys 73 is the amino acid involved in Trx-1 dimer formation. All peptides involved in the same region, suggesting a disulfide bond shuffle between these proteins, were identified in the same band shown in Table  1 , Supplementary Table 5 Furthermore, these heterodimer peptides were also monitored by SRM. The results indicated that the intensity of both peptides 73 We further evaluated if the interaction between Trx-1 and ADAM17cyto WT was only dependent on the disulfide bond. For that, Trx-1 and ADAM17cyto WT were treated with DTT, or DTT, and IAA or only IAA before recombinant protein incubations, followed by solid-phase binding assay. We observed higher interactions of ADAM17cyto WT and Trx-1 in the absence of DTT, either in the absence or presence of IAA (Fig. 4D) . It indicates that the interaction may happen independently of the disulfide bond. The SDS-PAGE performed in reducing and non-reducing conditions confirms the states of Trx-1 and ADAM17cyto upon the DTT treatment (Fig. 4E) .

ADAM17cyto WT decreased cell redox state and ADAM17 shedding activity and increased Trx-1 monomeric levels in mammalian cells.

To investigate the effect of the ADAM17 modulation of Trx-1 in dimeric and monomeric forms on reactive oxygen species (ROS), we measured H 2 O 2 levels by Amplex Red ® probe in cells. Lower oxidant levels were detected when ADAM17cyto WT and Trx-1 WT were co-expressed in HEK293 cells compared to ADAM17cyto F730A and Trx-1 (Fig. 5A) . Furthermore, we evaluated in the same conditions the consumption of H 2 O 2 with the addition of 10 µM H 2 O 2 into the system, besides Trx-1 alone. Similar results were found, in which the presence of ADAM17cyto WT had higher consumption when compared to ADAM17cyto F730A (Fig. 5B) . The ADAM17 sheddase activity was also evaluated in the presence of ADAM17cyto WT or ADAM17cyto F730A .

Under PMA stimulation, in the presence of recombinant Trx-1 or Trx-1 + ADAM17cyto WT , the shedding was reduced and the rescue was observed in the presence of Trx-1+ ADAM17cyto F730A (Fig. 5C) .

To investigate the conformational state of Trx-1 in cell lines in the presence of ADAM17, firstly, we co-overexpressed ADAM17cyto WT and Trx-1 or ADAM17cyto F730A and Trx-1 in HEK293 cells, pre-treated with 1 mM of H 2 O 2 and evaluated in reducing and non-reducing conditions. The results showed a significant increase in the monomeric form of Trx-1 in the presence of ADAM17cyto WT , indicated by the densitometry analysis (Fig. 5D) .

The increase in the monomeric state of Trx-1 was also confirmed under endogenous levels of Trx-1. For this, we used HEK293/Flp-In ADAM17-HA cells overexpressing ADAM17 and we showed in Figure 5E a pattern of higher abundance of the monomeric form of endogenous Trx-1. We also demonstrated in A431 ADAM17 knockdown cells the opposite results, indicating less abundance of the monomeric form of endogenous Trx-1 in ADAM17 knockdown cells (Fig.  5F ).

The transfection performed in all assays shown in Figure 5 were performed following the same procedure as shown in Supplementary Figure 5A .

ADAM17cyto WT was involved in the reduced state of intracellular proteins.

Considering the indirect or direct effect of Trx-1 monomer or dimer state in the global redox state, we evaluated the status of the reduced cysteine residues in proteins of HEK293 cells transiently expressing either ADAM17cyto WT or ADAM17cyto Table 7) .

2,148 cysteine-containing peptides were identified, with 821 quantifiable peptides and 574 corresponding protein groups (filtered out after quality control). The results from Supplementary Figure  5B -C showed the performance indicators, such as CV% indicating a reasonable quality control performance of the iodoTMT assay. The clustering analysis of iodoTMT data revealed that the peptide abundance of free cysteine residues was also able to separate the samples into two main clusters based on the expression of ADAM17cyto WT or ADAM17cyto F730A (Fig. 6A) .

The volcano plot indicated the abundance of free cysteine residues, quantified by iodoTMT labeling, in which the majority of quantified peptides were more abundant in the Cys-SH state in Trx-1 + ADAM17cyto WT compared to Trx-1+ ADAM17cyto F730A by presenting the Cysred/Cysox fold change around 1.5 to 3 ( Fig.  6B and Supplementary Figure 5D ). Among the statistically significant abundant peptides, the identified peptides from proteins TXD17, CBR1, FAS, and PRDX1 showed peptides with reduced cysteines in the presence of ADAM17cyto. These proteins are particularly interesting due to their involvement in redox processes (Supplementary Table 8 Table 9 and Supplementary Table  10) .

Finally, based on the present results, we propose that ADAM17 can regulate the balance between monomeric and dimeric states of Trx-1 with consequences in the intracellular and extracellular redox state of the cell and in the free cysteines of intracellular proteins (Fig. 7) .

Thioredoxin-1 and -2 are important antioxidant proteins that facilitate the reduction of proteins through cysteine thiol-disulfide exchange. It belongs to a system of redox molecules, including NADPH and the flavoprotein thioredoxin reductase (TrxR1). Trx-1 exists either in the reduced form Trx-(SH) 2 containing a dithiol, or in its oxidized form with an intramolecular disulfide bond between the Cys residues (Trx-S 2 ). Thus, Trx-1 participates in redox reactions through reversible oxidation of its active site. TrxR1 catalyzes the reduction of oxidized Trx-S 2 by NADPH using FAD and its active redox disulfide site. Then, Trx-(SH) 2 is capable to reduce the disulfide bond of various substrate proteins [29] .

The structures of the reduced and oxidized human Trx-1 proteins (Trx-red and Trx-ox), and of the two mutant forms of the protein (Cys73/C73S; Cys32/Cys35) have been reported previously [24] showing dimeric thioredoxin in all four structures. Thioredoxin is thought to function as a monomer in redox reactions [24] . The role of disulfide bond formation in thioredoxin catalytic activity and oligomerization in non-redox conditions was assessed by previous studies [24] . In addition, an inactive dimeric form of human thioredoxin has previously been noted, leading to the suggestion that dimer formation may function in a regulatory sensor. That is plausible considering that oligomeric organization in crystallized proteins is not uncommon (eg. packing in E. coli thioredoxin crystals [30] . Previous studies have been shown J o u r n a l P r e -p r o o f that Trx-1 noncatalytic cysteine residues were oxidized followed dimerization and inactivation of the enzyme, indicating that Cys73 is critical for the regulation of the enzyme activity via formation of intra-and inter-disulfide bonds [31] , that were reduced by the enzyme with kinetics which supported a process of autoactivation. This maybe an important control mechanism for the activity of the bovine thioredoxin system and possibly other ortholog thioredoxins [32] . This reduced activity of Trx caused by dimerization provides a mechanism by which Trx activity is transiently inhibited under the conditions of oxidative stress, providing more time for sensing and transmission of oxidative signals. This can be seen by redox regulation of cellular PTEN activity, which is restrained by the oxidation of active-site cysteine by reactive oxygen species (ROS) [33] . Recovery of its enzymatic activity predominantly depends on the availability of cellular thioredoxin (Trx) and peroxiredoxins (Prx), both are important players in cell signaling. Trx and Prx undergo redox-dependent conformational changes through the oxidation of cysteine residues at their active sites. Their dynamics are essential for protein functionality and regulation. The same study shows a tight association of the redox regulation of PTEN with Trx dimerization and Prx hyperoxidation, providing guidance for the identification of novel therapeutic targets [33] . In the literature, there is an evidence of a dimeric Trx-1 structure, but the existence of a quaternary structure of this protein has been debated. In humans and other eukaryotes, the presence of a cysteine residue at the crystallographic symmetry axis points to the relevance of dimer formation in solution and in vivo. Crystallographic data for shrimp thioredoxin (LvTrx) obtained under different redox conditions reveal a dimeric arrangement mediated by a disulfide bond through residue Cys73 and other hydrophobic interactions located in the crystallographic interface, as reported for human Trx [34] . Through the analysis of five mutants located at the crystallographic interface, the study provides structural and biochemical evidence for the existence in solution of monomeric and dimeric populations of wildtype LvTrx and five mutants [34] . Based on the results of biochemical assays, SAXS studies, and the crystallographic structures it is clear that the Cys73 residue is essential for dimerization. Therefore, there are evidence of similar behavior in human thioredoxin, which shares a Cys at position 73 with LvTrx, a structural feature that is also present in some Trxs from vertebrates and crustaceans [34] . However, how protein folding drives disulfide bond formation is poorly understood despite the role such proteins play in variety of extracellular and intracellular functions [35] .

It is known that disulfide bonds are important stabilizing protein features as well as they can be considered labile cross-links promoted by disulfide shifts [36] . Although disulfide rearrangements through a thiol-disulfide exchange is recognized to play an important role in different physiological processes, its molecular determinants are still unknown. The reactivity of the thioldisulfide exchange has been studied by computational models and in the case of thioredoxins, the J o u r n a l P r e -p r o o f hydrophobic pocket around the CXXC motif, the geometry, dynamics, and electrostatic environment are key elements to decide on the redox potential and kinetics of this family of enzymes [37] . The accessibility of the disulfide by the attacking thiol depends on the highly dynamic and steric protein composition, with or without an evolutionarily designed active site [38] .

Besides PDI, it is already known that extracellular [39] and intracellular [27] thiol isomerases, such as, Trx-1 are potential regulators of ADAM17 [27, 39] . We previously showed that Trx-1 interacts with the cytoplasmic domain of ADAM17 [27] , where the interface of interaction is opposite to the catalytic site of Trx-1 and Trx-1 negatively modulates the activity of ADAM17. It is also known that Trx-1 can be modulated by a change in its dimeric/monomeric state since the dimeric form is inactive and does not expose the catalytic site [24] . Also, we imply that the effect that ADAM17cyto plays on Trx-1 dimerization can occur by promoting this conformational change and rendering the cysteines less accessible to dimerization and autoregulation.

In the in silico ADAM17cyto and Trx-1 complex modeling, the residue F730 of ADAM17 was previously shown to be spatially located near the amino acid residue K72 of Trx-1, residue that was also shown to contribute to the binding interface and modeling of Trx-1 dimers (Cys 73 -Cys 73 ' disulfide bond in direct vicinity to K72) [27, 28] . Therefore, we have interrogated if the interaction with ADAM17cyto WT could modulate or influence the oligomerization state of Trx-1 and, thus, affect the oxidation/reduction state of the cell.

Herein, we have generated an ADAM17 mutation (ADAM17cyto F730A ) by site-directed mutagenesis to interfere in the interaction between ADAM17cyto and Trx-1. The residue F730 was chosen for mutation because in silico modeling revealed that it contributes to the complex binding energy and is predicted to be an important residue to maintain the complex [27] .

Considering that the Trx-1 dimer is inactive and the Trx-1 Cys 73 mutation is less prone to dimerization, Trx-1 activity could be regulated by modulating its dimeric/monomeric ratio [24] . It has been demonstrated that when the Cys 73 residue is mutated, more Trx-1 homodimers are formed and the activity of Trx-1 is decreased [24] . Therefore, it is possible that this interaction interface may also regulate the balance between Trx-1 dimers and monomers, since a higher proportion of dimers compared to monomers could explain the lower activity of Trx-1. We propose that this regulation is performed by ADAM17cyto WT or ADAM17 modulating the Trx-1 conformation and favoring the Trx-1 monomer state.

We indicated that the interaction of ADAM17cyto WT with Trx-1 favors a monomeric form of Trx-1, with consequent higher activity and resulting in lower cell oxidant levels. In addition, we showed that one of the mechanisms of ADAM17cyto WT can involve the disulfide bond formation between the proteins. Most importantly, the levels of the monomeric form of Trx-1 were also increased in ADAM17 overexpressing cells and decreased in ADAM17 knockdown cells.

On the other hand, as a control, ADAM17cyto F730A was used as a consequence of its lower Trx-1 binding ( Fig. 2A) , and it was not able to increase the Trx-1 reductase activity (Fig. 2C) , its presence promoted a higher hydrogen peroxide concentration (Fig. 2D and  5B ) resulting in the increase of ADAM17

shedding activity compared with the wild type (Fig.  5C) . The results indicated a constitutive shedding reduction by expression of recombinant Trx-1 and the rescue of shedding by ADAM17cyto F730A .

Also, these results reinforced that ADAM17cyto WT directly and positively affected the Trx-1 redox activity. Besides, it corroborated with previous results showing that Trx-1 and reactive oxygen species modulated ADAM17 activity [28, 40] .

Besides, considering the role of the Trx-1 redox state, we evaluated the effect of ADAM17 on intracellular proteins, by iodoTMT analyses, using HEK293 cells cotransfected with Trx-1 and ADAM17cyto WT or Trx-1 and ADAM17cyto F730A (Fig. 6) . The experiments suggested that the presence of ADAM17cyto WT resulted in more free cysteine-containing enzymes and their substrates, involved in the redox pathway. For example, in the presence of ADAM17cyto WT , peroxiredoxin-1 was identified in a more reduced form (LNCQVIGASVDSHFCHLAWVNTP K, m/z 550.46, +6, Supplementary Table 7) . Also, TrxR1, known to reduce Trx-1, presented more free cysteines (CDYENVPTTVFTPLEYGACGLSE EK, m/z 856.67, +4, Supplementary Table 7) , making it possible to regenerate Trx-1 in the presence of the complex with ADAM17cyto WT .

The modulation of thiol reductase activity by ADAM17cyto WT reflected in the increase of the reduction state of several cysteine-containing proteins mainly associated with oxidoreductase function (Fig. 6C and Supplementary Table 8 ).

In the case of Trx-1, the Cys 73 containing peptide was the unique peptide identified using iodoTMT labeling, and it was with reduced cysteine. However, the mass spectrometry analysis was not able to detect the differences in its abundance between the conditions, ADAM17cyto WT or ADAM17cyto F730A . We can suggest that although the iodoTMT labeling protocol performed an enrichment of peptides containing cysteines, possibly due to the high complexity of samples and untargeted approach, we could not quantify such small differences in cells (Supplementary Table 7) .

On the other hand, the Cys 73 containing peptide was identified by DDA and quantified by SRM (Fig.  3C ) when we used recombinant proteins Trx-1 and ADAM17cyto. Moreover, we confirmed an increase in the free Cys 73 containing peptides in the presence of concentrationdependent manner ADAM17cyto WT (Supplementary Table  4 ), explaining the higher monomeric levels and active Trx-1. Corroborating with this result, these overall effects were reduced in the presence of ADAM17cyto F730A . Together with this increase of the free Cys 73 containing peptides, the heterocomplex of ADAMcyto and Trx-1, linked by a disulfide bond ( Fig.4B-C) , also increased in a concentration-dependent manner of ADAM17cyto WT . Unfortunately, we were unable to confidently quantify the Cys 73 containing homodimer peptides, due to the lower intensity of transitions resulted probably from the low accessibility of trypsin enzyme during digestion performed in non-reducing conditions.

The effect of ADAM17cyto WT favoring the Trx-1 monomer state affects other free cysteine residue containing proteins, mostly involved in nucleotide binding and ligase activity, oxidoreductase activity, and peroxiredoxin activity (Supplementary Table 9 14) , is a disulfide reductase with two active sites containing Cys residues in its WCPDC motif and it is known that Cys residues are also reduced by thioredoxin reductase (TRR) such as Trx-1 [41] . TRP14 is known to inhibit the TNF-α-induced NF-κB activation to a greater extent than Trx-1 [42] . Herein, we detected an increase in the reduced state level of TRP14 in the presence of ADAM17cyto WT . This fact reinforces the complex nature of communication between ADAM17, Trx-1, and other TRR substrates, in addition to the cellular oxidant levels.

The signaling dependent on redox sensors in the cell is also known for other proteins. In nonoxidatively stressed cells, one of the MAPKKK (or MAP3K) enzymes named apoptosis signal-regulated kinase 1 (ASK-1) is kept inactive by association with reduced thioredoxin. Upon formation of a disulfide bond with thioredoxin via peroxiredoxin-catalyzed oxidation of two critical cysteines by H 2 O 2 , the oxidized thioredoxin dissociates from ASK-1 allowing its oligomerization and activation. The covalent binding of reduced Trx-1 to reduced ASK1 impairs its downstream activation of the kinases p38 and Jun NH2-terminal kinase (JNK). On the other hand, the oxidation of ASK1 promotes rapid multimerization [43] . H 2 O 2 may also prolong MAP kinase signals by directly inactivating MAP kinase phosphatases via cysteine oxidation, and in growth factor signaling, oxidation of the phosphatase PTEN allows peptide growth factor (e.g. PDGF or EGF)-induced H 2 O 2 production to mediate sustained proliferation signals [44] .

Finally, it is interesting to note that the MS data-dependent acquisition indicated a disulfide shuffling within N-terminus and Cterminus of ADAM17cyto, and also within Cys 73 of Trx-1 responsible for its homodimer ( Table 1 and Supplementary Figure 4) . A similar shuffling of disulfide bond patterns was previously observed in the membrane-proximal domain of ADAM17, controlled by the protein disulfide isomerase (PDI), resulting in an active open state and an inactive closed conformation [45, 46] .

Altogether, this study suggests a dynamic relationship between ADAM17 and Trx-1. Besides Trx-1 can negatively regulate ADAM17 activity [27] , ADAM17cyto WT and full-length ADAM17 can positively modulate Trx-1, thereby revealing a new and unexpected function for the cytoplasmic domain of ADAM17 promoting a dual regulation of the thiol reductase and as consequence an autoregulation of the metalloproteinase activity (Fig. 7) . It has been suggested that strategies for increasing the antioxidant capacity of anti-tumor T cells, by activating Trx-1, can modulate their immune-metabolic phenotype J o u r n a l P r e -p r o o f leading to improved immunotherapeutic control of established tumors [47] . In this context, ADAM17cyto can be promoting a more antioxidant state, but there is also an increase of ADAM17 shedding.

These findings create a new paradigm for the influence of ADAM17 in regulating the redox state of the cells, implicating in novel mechanisms of ADAMs and their roles in redox regulation of the tumor microenvironment.

All chemicals were analytical grade obtained from commercial sources. Isopropyl-β-dthiogalactopyranoside (IPTG, 99%) was purchased from Promega. Tris(hydroxymethyl)-aminomethane (Tris, NH 2 C(CH 2 OH) 3 , 99.8%) was purchased from Affymetrix. The cocktail Protease Inhibitor was purchased by Roche.

The TMT10plex™ isobaric label reagent set was purchased from Thermo Fisher Scientific. The antibody anti-Trx (LF-PA0187) was purchased from AbFrontier and the anti-actin (ab8227) was purchased from Abcam.

Culture and Transient Transfections HEK293 cells were cultured in DMEM medium supplemented with 10% FBS, 100 IU/mL penicillin, and 100 mg/mL streptomycin, at 37 o C in a humidified atmosphere with 5% CO 2 . HEK293) stably expressing alkaline phosphatase reporter (HB-EGF-AP) were cultured in DMEM medium supplemented with 10% FBS, 100 IU/mL penicillin, and 100 mg/mL streptomycin at 37 o C in a humidified atmosphere with 5% CO 2 . The geneticin 418 (G418) antibiotic was added to the culture medium (1 mg/mL final concentration) to maintain the selection of the clone. HEK293/Flp-In ADAM17-HA cells were cultured in DMEM medium supplemented with 10% FBS, 100 IU/mL penicillin, and 100 mg/mL streptomycin, at 37 o C in a humidified atmosphere with 5% CO 2 . Tetracycline was added to the culture medium (1 µg/mL final concentration) to induce the expression of the ADAM17-HA recombinant gene for 24 hours at 37 o C.

The A431 ADAM17 knockdown cells) were cultured in RPMI-1640 medium supplemented with 10% FBS, 100 IU/mL penicillin, and 100 mg/mL streptomycin, at 37 o C in a humidified atmosphere with 5% CO 2 . The geneticin 418 (G418) antibiotic was added to the culture medium (1 mg/mL final concentration) to maintain the selection of resistant cells. Cells were cultured until 70-80% of confluency.

Mutagenesis of ADAM17cyto (ADAM17cyto F730A ) was performed with the mutagenesis kit (Stratagene, San Diego, CA) and specific primers were used (Forward: 5´-GTTCGCATTATCAAACCCGCTCC TGCGCCCCAG-3´ and Reverse: 5´-CTGGGGCGCAGGAGCGGGTTTG ATAATGCGAAC-3´), following the manufacturer´s protocol. ADAM17cyto F730A was cloned in pET28a (+) and pcDNA 3.1+, for protein expression in E. coli and mammalian cells, respectively. Trx-1, ADAM17cyto WT and Trx1 mutants containing plasmids for protein expression in E. coli and mammalian cells have been previously described [27] . For E. coli expression, genes for Trx-1 and Trx-1 mutants (Trx-1 K72A and Trx-1 C32/35S ) were cloned with a HIS6 tag in the N-terminal and an HA tag in the C-terminal; genes for ADAM17cyto WT and ADAM17cyto F730A were cloned with a HIS6 and FLAG tags in the Nterminal.

For expression in mammalian cells, genes for Trx-1 and Trx-1 mutants (Trx-1 K72A and Trx-1 C32/35S ) were cloned with an HA tag in the C-terminal; genes for ADAM17cyto WT and ADAM17cyto F730A were cloned with a FLAG tag in the N-terminal. HEK293 was transiently transfected by Lipofectamine, following the manufacturer´s protocol (Invitrogen, Carlsbad, CA) or using polyethylenimine (Polysciences Inc., Warrington, PA) for 48 h.

Trx-1, Trx-1 K72A , Trx-1 C32/35S , ADAM17cyto WT and ADAM17cyto F730A proteins were expressed in BL21 (DE3) cells at 37°C for 4 h after induction with 0.5 mM IPTG in LB. The harvested cells were resuspended in lysis buffer (20 mM Tris, pH 7.5, containing 1 mM CaCl 2 and 1 mM PMSF) and disrupted by lysozyme treatment (100 µg/mL for 30 min on ice), followed by sonication (Vibracell VCX 500; Sonics & Materials, Inc., Newtown, CT). The suspensions were centrifuged at 20,000 x g for 10 min at 4°C. The supernatant was loaded onto nickel-charged 5-ml His Trap Chelating columns (GE Healthcare, Chicago, IL) using a flow rate of 1 mL/min in buffer A (20 mM Tris, pH 7.5, containing 1 mM CaCl 2 and 1 mM PMSF). Proteins were eluted using a linear gradient of 0-1 M imidazole. All the purified fractions were combined in a 3-kDa cut-off Amicon filter (Millipore, Burlington, MA) and the buffer was J o u r n a l P r e -p r o o f changed into phosphate-buffered saline (PBS). PBS was prepared with NaCl (137 mM), KCl (2.7 mM), Na 2 HPO 4 (10 mM), and KH 2 PO 4 (1.8 mM), the pH was adjusted to 7.2 with HCl for 1 L of an aqueous solution. The purified fractions were separated by 15% SDS-PAGE under denaturing conditions to evaluate the quality of the protein purification. Trx-1 K72A and Trx-1 C32/35S were purified as previously reported [28] . Final protein concentration was determined by a Bicinchoninic acid (BCA) protein assay kit (Thermo Fisher Scientific, Waltham, MA). At least three independent preparations of purified protein were obtained for further experiments.

One µg of purified Trx-1 was immobilized into a 96-well polystyrene High Bind microtiter plate (Corning Glass, Corning, NY) in a 0.1 M sodium carbonate buffer, pH 9.6, with gentle agitation overnight at 4°C, as described previously [48] . The wells were washed three times with wash solution (PBS supplemented with 0.05% Tween 20) and then blocked with PBS containing 5% nonfat dry milk for 2 h at room temperature. After blocking, the wells were washed again and the purified proteins, ADAM17cyto WT and ADAM17cyto F730A , were added in an increasing concentration (0.5-8 µM), diluted in PBS supplemented with 0.05% Triton X-100 and 0.1% BSA. The plates were incubated for 2 h at room temperature and after that, the wells were washed three times. We incubated the complexes for 1 h at room temperature with anti-FLAG peroxidase-linked antibody, diluted at 1:5000 for 1 h at room temperature. After incubation with the antibody, the wells were washed three times and a buffer containing 2,2´-azinobis- ( Also, the DTNB assay was performed to evaluate the reactivity of Trx-1 recombinant cysteines in the presence of the complex with ADAM17cyto as described previously [28] . For that, recombinant Trx-1 was incubated with DTT (10mM) and excess DTT was removed by buffer exchange in Microcon ® Filter (Merck, Darmstadt, DE). 5 µl of 10mM DTNB (at a final concentration of 1mM) was added to 100 µM of recombinant, purified, and previously reduced Trx-1 WT together with ADAM17cyto WT or ADAM17cyto F730A in 43 µl of 100mM of glycine buffer at pH 8.3. The reaction product was measured at 450 nm. Three independent experiments were performed. Besides, to evaluate if the mutation in ADAM17cyto was able to change its secondary structure, circular dichroism analysis was performed with both proteins (ADAM17cyto WT and ADAM17cyto F730A ) as previously described [49] . One independent experiment was performed.

To test whether ADAM17cyto WT could affect Trx-1 reducing activity and determine the concentration of ADAM17cyto WT and ADAM17cyto F730A that should be used in the assay, increasing concentration of the proteins (1, 2.5 and 5µM) were incubated with 500nM of previously reduced Trx-1 and 150nM of Di-E-GSSG in 0.1 M potassium phosphate buffer with 1 mM EDTA pH 7.5 (PE buffer) at 25°C. Later, fluorescence was measured at λex of 525 nm and λem of 545 nm for all assays over time. The experiment was performed three times independently in duplicate.

To analyze the monomeric state of Trx-1, recombinant purified Trx-1 (15µM) and ADAM17cyto WT (3, 6, 12, and 24 µM) proteins were first incubated together for 30 min on ice. Furthermore, sample buffer (without β-mercaptoethanol) was added to the protein complex and monomer and dimer forms of the recombinant protein Trx-1 alone or incubated with ADAM17cyto WT was analyzed on a non-reducing 15% SDS-PAGE gel. This experiment was performed two times.

In parallel, to perform the relative semiquantitative analysis of recombinant Trx-1 protein in the monomer and dimer forms the peptide spectrum matches (PSMs) were obtained. For that, gel bands were analyzed from a non-reducing 15% SDS-PAGE gel and the bands of monomer and dimer forms in Trx-1 alone or incubated with ADAM17cyto WT were first alkylated (14 mM iodoacetamide, 30 min at room temperature in the dark), and then digested with trypsin (Promega, Madison, WI). The samples were dried in a vacuum concentrator and reconstituted in 20 µl of 0.1% formic acid. Four and a half microliters of the resulting peptide mixture were analyzed on an LTQ Orbitrap Velos mass spectrometer (Thermo Fisher Scientific, Walthan, MA) coupled with LC-MS/MS by an EASY-nLC system (Proxeon Biosystem, Odense, DK) through a Proxeon nanoelectrospray ion source according to [27] .

Peak lists (msf) were generated from the raw data files using Proteome Discoverer version 1 To evaluate the effect of ADAM17cyto WT and ADAM17cyto F730A on Trx-1 Cys 73 reduced state, the Trx-1 peptide ( 73 CMPTFQFFK 81 ), which contains the cysteine residue responsible for Trx-1 homodimerization, was submitted to relative quantification using SRM as previously described in [28] , with modifications. After the proteins´ incubations (Trx-1+ ADAM17cyto WT and Trx-1+ ADAM17cyto F730A , under ADAM17cyto concentrations of 3, 6, 12, and 24µM) during 15 min at 25 o C, all reactions were stopped by denaturing condition using 4M urea, in non-reducing conditions. The mixture of proteins was alkylated and submitted to trypsin digestion (1:50).

The experiment was performed in three technical replicates.

Proteotypic peptides for Trx-1 and ADAM17cyto proteins were selected based on our DDA analysis and following the rules as previously described [53, 54] . Briefly, a total of 3 proteotypic peptides were selected to be monitored label-free, including the disulfide-linked peptide at position Cys 73 , and a total of 2 proteotypic peptides were selected to be monitored label-based, including the carbamidomethylated peptide at position Cys 73 with and without methionine oxidation (Supplementary Table 3) . At least five transitions were monitored for the light and heavy peptides, with a total of 90 transitions. A mixture of 2pmol of the heavy isotopic labeled peptide was added to the samples prior to desalinization performed using Sep-pak C18 cartridge (Waters, Milford, MA). Internal retention time standards (iRT, Pierce Peptide Retention Time Calibration Mixture, Thermo Fisher Scientific, Watlham, MA) were spiked in all samples at 50 fmol/µl, before sample injection and four peptides with 12 transitions were monitored. Also, two trypsin auto-lysis peptides were monitored to evaluate the trypsin digestion efficiency.

Samples were analyzed on a Xevo TQ-XS triple quadrupole mass spectrometer (Waters, Milford, MA) equipped with an electrospray ion source (Ion Key; Waters) and using MassLynx software (version 4.2) as previously described [28] . Visualization and inspection of peaks were manually performed in Skyline. The label-free and labelbased quantification had the light peptides checked regarding the quality of data by observing coelution of all five transitions and relative intensity correlation with spectral library (dotp) and stable isotope-labeled peptides (SIL) (rdotp).

For label-based quantification, the light peptides were normalized using SIL (Supplementary Table 4 ) and for label-free quantification, the light peptides were normalized to the intensity of two iRT peptides (average intensity value was used as denominator) (Supplementary Table 6 ). iRT peptides intensities were also evaluated among all runs to determine the quality of the LC. In addition, in the case of methioninecontaining peptide, the intensity of the two forms, oxidized or nonoxidized, were summed (Supplementary Table 4 and 6) . In addition, for label-free disulfide bond peptides, the transitions were selected based on the inspection of MS/MS spectra provided by DDA.

To measure the effect of ADAM17cyto WT in hydrogen peroxide levels present in the extracellular medium of mammalian cells, Amplex Red Hydrogen Peroxide/Peroxidase Assay Kit ® (Molecular Probes, Invitrogen, Carlbad, CA) was used according to the manufacturer´s protocol and the reaction was read in a fluorimeter at 37°C, at λex of 530 nm and λem of 590 nm. Briefly, HEK293 cells stably transfected with HB-EGF-AP were seeded (0.3x10 6 ) into 6-well plates (Corning, Corning, NY) and starved for 24 h. Next, the cells were transiently co-transfected for 6 h with the encoding vectors Trx-1/ADAM17cyto WT or Trx-1/ADAM17cyto F730A . After, cells were activated with PMA (50 ng/mL) for 30 min in DMEM (without FBS), they were harvested, washed with cold PBS and trypsinized. Approximately, 1x10 4 cells were placed into a 96well microplate and to the cells was added 50 µL of the reaction buffer (working solution of 20 μM Amplex ® Red reagent, 0.2 U/mL HRP and 1x reaction buffer) to each well. The cells were incubated in the dark for 30 min and the reaction was measured in the fluorimeter. Three independent experiments in duplicate were performed.

To measure ADAM17 shedding activity in HEK293 cells, upon the presence of Trx-1 WT /ADAM17cyto WT complex, HEK293 cells stably transfected with HB-EGF-AP [27] were seeded (0.3x10 6 ) into 6-well plates (Corning, Corning, NY). The following day, the cells were transiently transfected for 24 h with vectors encoding Trx-1, ADAM17cyto WT, and ADAM17cyto F730A for co-expression of Trx-1/ADAM17cyto WT or Trx-1/ADAM17cyto F730A . After 24 h, the cells were trypsinized, counted, and 3×10 5 cells were seeded into 24-well plates (Corning, Corning, NY). The following day, the cells were starved for 4 h. Following starvation, the cells were activated with PMA (50 ng/mL) [54] for 1 h in a phenol-free medium. The conditioned media was removed from the culture plate and the shedding of HB-EGF-AP was measured by alkaline phosphatase enzymatic assay. Briefly, 100 µL of conditioned media was collected from each well and added to individual wells of a 96-well plate containing 100 µL of AP buffer (0.5M Tris-HCl, pH 9.5, containing 5 mM pnitrophenyl phosphate disodium, 1 mM diethanolamine, 50 µM MgCl 2 , 150 mM NaCl, 5 mM EDTA). After the enzymatic reaction was completed the product was measured at an absorbance wavelength (Aλ) of 405 nm. Three independent experiments were performed in duplicate for nonoxidative and oxidative conditions.

To analyze if ADAM17cyto WT is able to alter Trx-1 conformation in mammalian cells, Trx-1 with ADAM17cyto WT or ADAM17cyto F730A were co-expressed in HEK293 cells for 48 h, followed by cell lysis with lysis buffer (50 mM Tris-HCl pH 7,4, 150 mM NaCl, 1 mM EDTA, 1% Triton and protease inhibitor cocktail tablet with the addition of 100 mM iodoacetamide) [43] and addition of sample buffer with or without βmercaptoethanol. Forty micrograms of total protein were loaded on a 15% SDS-PAGE gel, followed by anti-HA Western Blot (Sigma-Aldrich, 1:2000) to visualize the monomeric form of Trx-1. This experiment was performed once.

To study if ADAM17 fulllength can influence the oligomer The iodoTMT experiment was performed first co-transfecting HEK293 cells with Trx-1 plus ADAM17cyto WT (Trx-1 / ADAM17cyto WT ) or Trx-1 plus ADAM17cyto F730A (Trx-1/ADAM17cyto F730A ). The cells were lysed with modified RIPA buffer (25mM HEPES, pH 8.5 containing 150mM NaCl, 1mM EDTA, and 1% Triton X-100) supplemented with complete protease inhibitor cocktail (Roche, Basileia, CH). Afterwards, protein quantification was performed by BCA method and 500 µg of the lysate was mixed to 0. For peptide enrichment, the samples were lyophilized and redissolved in 600 μL of TBS buffer (25 mM Tris-HCl, pH 7.5, 0.15 M NaCl). 100 μL of settled immobilized anti-TMT antibody resin was employed for the enrichment of every 1 mg labeled sample at 4°C. After the supernatant was removed, the resin was washed three times with one column volume of TBS and three times with one column volume of deionized water. Peptides were finally eluted with four column volumes of elution buffer [10 mM cis-2,6-dimethylpiperidine (2,6-DMPp)/50/500 mM tetraethylammonium bicarbonate (TEAB), pH 8.5 [55] , dried on vacuum concentrator, resuspended in 25% acetonitrile /0.1% formic acid and injected directly (~2 µg/µL) onto the Orbitrap Fusion mass spectrometer. Three independent experiments were performed.

All spectra were acquired using an Orbitrap Fusion mass spectrometer (Thermo Scientific, Watlham, MA) in line with an Easy-nLC 1000 (Thermo Fisher Scientific, Watlham, MA) ultra-high pressure liquid chromatography pump. Peptides (∼1µg) were separated in a column packed in-house with a 100 µm inner diameter containing 1 cm of Magic C4 resin (5 μm, 100 Å, Michrom Bioresources, Auburn, CA) followed by ∼30 cm of GP-18 resin (1.8 μm, 200 Å; Sepax, Newark, DE) with a gradient consisting of 9-30% acetonitrile containing 0.125% formic acid over 180 min at ~250 nL/min.

For all LC-MS/MS experiments, the mass spectrometer was operated in the data-dependent mode.

For the Synchronous-Precursor-Selection (SPS-MS 3 ) analysis with the Orbitrap Fusion Classic mass spectrometer (Thermo Scientific, Watlham, MA) the scan sequence began with an MS 1 spectrum (resolution 120,000; mass range 400−1400 m/z; AGC target 5×10 5 ; maximum injection time 250 ms). Precursors for MS 2 /MS 3 analysis were selected using a Top10 method. MS 2 analysis consisted of collision-induced dissociation (CID); AGC 1.8×10 4 ; normalized collision energy (NCE) of 35%; maximum injection time of 120 ms; and isolation window of 0.7 Da. Following the acquisition of each MS 2 spectrum, we collected an MS 3 spectrum using the recently described method in which multiple MS 2 fragment ions (10 MS 2 fragment ions) were captured in the MS 3 precursor population using isolation waveforms with multiple frequency notches (multi-notch MS 3 ) (McAlister et al., 2014). MS 3 precursors were fragmented by higher-energy collisional dissociation (HCD) and analyzed in the Orbitrap (NCE 55%; AGC 1.0×10 5 ; maximum injection time of 120 ms, resolution was 60,000 at 400 Th).

Searches were performed with Sequest using a 50 ppm precursor ion tolerance and 1Da for fragment [56] . Only  peptide  sequences  with  both  termini  consistent with the protease specificity of trypsin were considered in the database search, and up to two missed cleavages were accepted. IodoTMT tags on cysteine residues (+329.22659 Da) was set as static modifications, while methionine oxidation (+15.99492 Da) was set as a variable modification.

An MS 2 spectral assignment false discovery rate of less than 1% was achieved by applying the targetdecoy strategy [57] . Filtering was performed using an in-house linear discrimination analysis algorithm described previously [58] to create one composite score from the following peptide ion and MS 2 spectra properties: XCorr, ΔCn score, peptide ion mass accuracy, peptide length and missedcleavages.

The resulting discriminant scores were used to sort peptides prior to filtering to a 1% FDR. Following spectral assignment, peptides were assembled into proteins and further filtered based on the combined probabilities of their constituent peptides to a final FDR of 1%. In cases of redundancy, shared peptides were assigned to the protein sequence with the most matching peptides, thus adhering to principles of parsimony [59] .

A compilation of in-house software was used to convert the Thermo RAW files to mzXML format, as well as to adjust monoisotopic m/z measurements and erroneous peptide charge state assignments. Using Sequest, MS2 spectra were searched against the Human UniProt database (April 2 nd, 2014, 71,882 total sequences), including sequences of common contaminating proteins. This database was followed forward by a decoy component. To evaluate the proteins that were significantly different in the presence of ADAM17cyto WT and ADAM17cyto F730A , we used the twotailed t-test with unequal variance. A false discovery rate (FDR) of 1% was considered using the Benjamin-Hochberg method with the in-house development tool. All the 821 proteins that were quantified in the multi-notch MS 3 data, considering single (only one) and composite (two or more) cysteine sites in the iodoTMT experiment, were analyzed and presented in the Supplementary Table 7-8.

GO annotation of molecular function for the differential proteins was performed using BinGO plugin 3.0.3 with Cytoscape 3.7.1, with the significance threshold set at P-value <0.05 using the Hypergeometric test. The whole human proteome Gene Ontology (GO) annotation file was used as reference set.

Overrepresented GO terms are represented as bar plots according to the P-value of the enrichment analysis (Supplementary Table 9 ). The fold enrichment was calculated between the enriched terms for the upregulated proteins and the downregulated proteins (Supplementary Table 10) .

For the statistical analysis of solid-phase binding assay, Trx-1 reductive activity, ADAM17 activity by AP reporter activity, measurement of H 2 O 2 production, and reactivity of cysteines by AmplexRed and DTNB, we performed Student´s t-test or oneway ANOVA followed by Tukey´s or Bonferroni's posttest. The significance level was stated at p<0.05 (GraphPad Prism version 5 for Windows, California, U.S.A).

The mass spectrometry proteomics data have been deposited in the ProteomeXchange Consortium (http://proteomecentral.proteomexch ange.org) via the PRIDE partner repository [60] with the dataset identifier PXD015431 (https://www.ebi.ac.uk/pride/archive/ projects/PXD015431). The SRM analyses for the seven measured proteins are available through the Panorama repository at the following link (https://panoramaweb.org/ 7j9Lbk.url). Skyline exported data for all quantified peptides are available in Supplementary Tables 4 and 6 .

All other data supporting the findings of this study are available from the corresponding author on request. Three independent experiments were performed in duplicate (two-way ANOVA followed by Bonferroni´s posttest, p<0.05). The plus (+) and asterisks (*) represent the statistics comparing the results from the no coated samples with Trx-1 coated+ADAM17cyto WT , and Trx-1 coated+ADAM17cyto F730A , respectively. The comparison between the samples Trx-1 coated + ADAM17cyto WT , and Trx-1 coated+ ADAM17cyto F730A is represented by the minus symbol (-). Data are represented as mean ± SD. B) Trx-1 in the presence of mutant ADAM17cyto F730A had less free cysteines than in the presence of ADAM17cyto WT . DTNB assay was performed by incubating 20 µg of purified Trx-1 protein (previously reduced) with ADAM17cyto WT or ADAM17cyto F730A in the presence of DTNB. Three independent experiments were performed in duplicate (one-way ANOVA followed by Tukey´s test p<0.05). Data are represented as mean ± SD. C) Trx-1 in the presence of 2.5µM of ADAM17cyto WT had more reduced activity over Di-E-GSSG substrate. Di-E-GSSG substrate (150 nM) was incubated with 500 nM of purified and reduced Trx-1 alone, Trx-1/ ADAM17cyto WT or Trx-1/ ADAM17cyto F730A in a concentration-dependent manner, in 0.1 M potassium phosphate buffer with 1 mM EDTA pH 7.5 at 25°C and the fluorescence (λex 545 nm and λem 520 nm) was measured over time and the slope of the curve was obtained. Three independent experiments were performed (Student´s t-test p<0.05). Data are represented as mean ± SD. D) Trx-1 in the presence of mutant ADAM17cyto F730A had less cysteine activity than in the presence of ADAM17cyto WT , while Trx-1 K72A mutant, that is inactive due to its dimeric form, in the presence of ADAM17cyto WT had more cysteine reactivity. A cysteine reactivity assay was performed by incubating 20 µg of purified Trx-1 or Trx-1 mutant (Trx-1 K72A and Trx-1 C32/35S ) proteins with ADAM17cyto WT and Trx-1 with ADAM17cyto F730A and 20 µM Amplex red and 1 U/mL peroxidase, followed by fluorescence measurements (λex 530 nm and λem 590 nm) and the addition of 10 µM of H 2 O 2 . Three independent experiments were performed in duplicate (one-way ANOVA followed by Tukey´s test p<0.05). Data are represented as mean ± SD.

J o u r n a l P r e -p r o o f Mass spectrometry analyses revealed an increase in the relative intensity of Trx-1 in the presence of increasing amounts of ADAM17cyto WT (3, 6, 12µM). The experiment was performed twice, with two different protein preparations. B) Analytical ultracentrifugation data showed that the presence of ADAM17cyto WT favored the monomeric state of Trx-1 WT . This experiment was performed four times with two different protein preparations. The x-axis below represents the molecular weight (MW) data from each sample and peak. C) The carbamidomethylated peptide of Trx-1 C 73 (C 73 MPTFQFFK, m/z 574.76, +2 and C 73 M ox PTFQFFK, m/z 603.27, +2) was quantified by SRM. The peptide was measured in the absence of ADAM17cyto and in the presence of ADAM17cyto WT and ADAM17cyto F730A at concentrations of 3, 6, 12, and 24µM. The proteins were submitted to in solution digestion in non-reduced conditions, with iodoacetamide treatment. For normalization of Trx-1 Cys 73 peptide, the synthetic heavy labeled counterpart peptide was used. Data are represented as mean ± SD. (3, 6, 12 , and 24µM) analyzed in nonreducing SDS-PAGE. The 30kDa-band corresponding to the Trx-1 dimer size and/or Trx-1-ADAM17cyto WT complex were excised, trypsin digested, and analyzed by mass spectrometry using DDA. The data were searched using MassMatrix and pLINK2 software. Two independent experiments were performed with 4 different ADAM17cyto concentrations and the disulfide bond peptide was identified in 2 conditions of each independent experiment in the presence of ADAM17cyto WT . This experiment was repeated with two different sample preparations, and the same disulfide peptides were confirmed. The validation of the disulfide bond is shown. The symbol (-) represents the cysteines involved in disulfide bond formation. B-C) The disulfide bonds of the heterodimer Trx-1 and ADAM17cyto WT (CMPTFQFFK-VDSKETEC, m/z 818.83, +3 and CMPTFQFFK-ETEC, m/z 813.83, +2) were quantified by mass spectrometry using SRM. The peptide corresponding to the disulfide bond was measured in the absence or presence of ADAM17cyto WT and ADAM17cyto F730A (3, 6, 12, and 24µM). The experiment was performed in three technical replicates. Data are represented as mean ± SD. D) ADAM17cyto WT bound more to Trx-1 in the absence of DTT and the presence of IAA treatments. A solidphase binding assay was conducted by adding 0.5 and 1µM of ADAM17cyto WT Figure 5 . ADAM17cyto WT increased the redox activity in cells and stimulated HB-EGF shedding while the Trx-1 monomer form was favored in ADAM17cyto WT and in ADAM17 overexpressing cells. A) Lower H 2 O 2 levels were observed in ADAM17cyto WT overexpressing cells compared to ADAM17cyto F730A . HEK293 cells stably expressing AP-HB-EGF were transiently co-transfected with Trx-1/ADAM17cyto WT or Trx-1/ADAM17cyto F730A and stimulated with PMA. Cells were incubated with 20 µM Amplex Red and 1 U/mL of peroxidase, followed by fluorescence measurement (λex 530 nm and λem 590 nm). Three independent experiments were performed in duplicate (One-way ANOVA, followed by Tukey´s multiple comparison posttest (p<0.05)). Data are represented as mean ± SD. B) HEK293 cells overexpressing ADAM17cyto WT had higher cysteine reactivity than overexpressing ADAM17cyto F730A . A cysteine reactivity assay was performed using supernatants from HEK293 co-expressing Trx-1/ADAM17cyto WT or Trx-1/ADAM17cyto F730A , and incubating the supernatants with 20 µM Amplex red and 1 U/mL peroxidase, followed by fluorescence measurements (λex 530 nm and λem 590 nm) and hydrogen peroxide. Three independent experiments were performed in duplicate (Oneway ANOVA followed by Tukey´s test, p<0.05). Data are represented as mean ± SD. C) Overexpression of ADAM17cyto WT showed an inhibitory effect on ADAM17 shedding activity compared to ADAM17cyto F730A . HEK293 cells stably expressing AP-HB-EGF were transiently co-transfected with Trx-1 and ADAM17cyto WT or Trx-1 and ADAM17cyto F730A , starved for 4 h and then stimulated with PMA, followed by measurement of alkaline phosphatase activity at 405 nm. Three independent experiments were performed in duplicate (Oneway ANOVA followed by Tukey´s test, p<0.05). Data are represented as mean ± SD. D) The intracellular monomeric form of Trx-1 increased in the presence of Hierarchical clustering was performed in the R environment using the Euclidean distance with the method Abscorrelation average. B) Volcano plot demonstrated that in the presence of ADAM17cyto WT / Trx-1, the abundance of ninety-one peptides with freecysteines, from a total of 2,147 iodoTMT identified peptides, were increased compared to ADAM17cyto F730A / Trx-1. Three independent experiments were performed (Student´s t-test, p<0.05). C) The 10 molecular functions that better represented the group of differentially reduced/oxidized proteins between Trx-1/ADAM17cyto WT and Trx-1/ADAM17cyto F730A were protein binding, NAD or NADH binding, oxidoreductase activity, peroxiredoxin activity, phosphoglycerate dehydrogenase activity, cofactor and coenzyme binding. The model showed that ADAM17cyto WT positively regulated Trx-1 enzyme activity by favoring the Trx-1 monomeric state and it resulted in the increase of Trx-1 redox activity, promoting reduced state of intracellular proteins and lower H 2 O 2 in the extracellular environment. One of the mechanisms that explained this dimer-monomer modulation by ADAM17cyto was the formation of a heterodimer with the disulfide bond formed between Cys 824 of ADAM17cyto and Cys 73 of Trx-1. Cys 73 is known to be the site of Trx-1 dimerization [24] .

150221/2014-2 to DCG)

Measurement of Thioredoxin and Thioredoxin Reductase

ADAM-17: The enzyme that does it all

Role of Cytokines in the Pathogenesis of Inflammatory Bowel Disease

In pneumococcal meningitis a novel water-soluble inhibitor of matrix metalloproteinases and TNF-&#x3b1; converting enzyme attenuates seizures and injury of the cerebral cortex

Tumorigenicity of cortical astrocyte cell line induced by the protease ADAM17

Decoding the enigma of antiviral crisis: Does one target molecule regulate all?

The COVID-19 Cytokine Storm; What We Know So Far

Inflammatory skin and bowel disease linked to ADAM17 deletion

ADAM17 as a therapeutic target in multiple diseases

ADAM17: a molecular switch to control inflammation and tissue regeneration

Tumor necrosis factor-α convertase (ADAM17) mediates regulated ectodomain shedding of the severe-acute respiratory syndrome-coronavirus (SARS-CoV) receptor, angiotensin-converting enzyme-2 (ACE2)

Modulation of TNF-αconverting enzyme by the spike protein of SARS-CoV and ACE2 induces TNF-α production and facilitates viral entry

The pathogenesis and treatment of the `Cytokine Storm' in COVID-19

Cytokine storm syndrome in coronavirus disease 2019: A narrative review

COVID-19 in the heart and the lungs: could we "Notch" the inflammatory storm?

Thioredoxin-1 promotes anti-inflammatory macrophages of the M2 phenotype and antagonizes atherosclerosis

Reactive oxygen species (ROS) as pleiotropic physiological signalling agents

Thioredoxin-1 distinctly promotes NF-kB target dna binding and nlrp3 inflammasome activation independently of txnip

Anti-oxidative, anticancer and anti-inflammatory actions by thioredoxin 1 and thioredoxinbinding protein-2

Thioredoxininteracting protein links oxidative stress to inflammasome activation

Linkage of inflammation and oxidative stress via release of glutathionylated peroxiredoxin-2, which acts as a danger signal

Reactive aldehyde modification of thioredoxin-1 activates early steps of inflammation and cell adhesion

Crystal structures of reduced, oxidized, and mutated human thioredoxins: Evidence for a regulatory homodimer

Site-specific and redox-controlled S-nitrosation of thioredoxin

Thioredoxin 1 is inactivated due to oxidation induced by peroxiredoxin under oxidative stress and reactivated by the glutaredoxin system

Identification of novel interaction between ADAM17 (a disintegrin and metalloprotease 17) and thioredoxin-1

Thioredoxin-1 Negatively Modulates ADAM17 Activity Through Direct Binding and Indirect Reductive Activity

Crystal structure of thioredoxin from Escherichia coli at 1.68 A resolution

Mutagenesis of structural half-cystine residues in human thioredoxin and effects on the regulation of activity by selenodiglutathione

Redox regulation of tumor suppressor PTEN in cell signaling

Is dimerization a common feature in thioredoxins? The case of thioredoxin from {\it Litopenaeus vannamei}

Protein folding guides disulfide bond formation

Evidence for intramolecular disulfide bond shuffling in the folding of mutant human lysozyme

How Thioredoxin Dissociates Its Mixed Disulfide

Accessibility explains preferred thiol-disulfide isomerization in a protein domain

Thiol isomerases negatively regulate the cellular shedding activity of ADAM17

Oxidative stress activates ADAM17/TACE and induces its target receptor shedding in platelets in a p38-dependent fashion

Roles of TRP14, a Thioredoxin-related Protein in Tumor Necrosis Factor-α Signaling Pathways

Thioredoxin-related protein 14, a new member of the thioredoxin family with disulfide reductase activity: implication in the redox regulation of TNF-alpha signaling., Free Radic

Disulfide Bondmediated multimerization of Ask1 and its reduction by thioredoxin-1 regulate H(2)O(2)-induced c-Jun NH(2)-terminal kinase activation and apoptosis

Redox control of cancer cell destruction

A disintegrin and metalloprotease 17 dynamic interaction sequence, the sweet tooth for the human interleukin 6 receptor

Membrane-proximal domain of a disintegrin and metalloprotease-17 represents the putative molecular switch of its shedding activity operated by protein-disulfide isomerase

Thioredoxin-1 improves the immunometabolic phenotype of antitumor T cells

New insights into the structural elements involved in the skin haemorrhage induced by snake venom metalloproteinases

RXR agonist modulates TR: corepressor dissociation upon 9-cis retinoic acid treatment

Optimization of Search Engines and Postprocessing Approaches to Maximize Peptide and Protein Identification for High-Resolution Mass Data

Intramolecular cross-linking experiments on cytochrome c and ribonuclease A using an isotope multiplet method

MS2Assign, automated assignment and nomenclature of tandem mass spectra of chemically crosslinked peptides

Size-Distribution Analysis of Macromolecules by Sedimentation Velocity Ultracentrifugation and Lamm Equation Modeling

ADAM17 is regulated by a rapid and reversible mechanism that controls access to its catalytic site

Proteomic quantification and site-mapping of S -nitrosylated proteins using isobaric iodoTMT reagents

A probabilitybased approach for high-throughput protein phosphorylation analysis and site localization

Target-decoy search strategy for increased confidence in large-scale protein identifications by mass spectrometry

McCulloch-Pitts Brains and Pseudorandom Functions

ProteomeXchange provides globally coordinated proteomics data submission and dissemination

ANOVA, analysis of variance

FDR, false discovery rate; GO, gene ontology

heparin-binding epidermal growth factor in fusion with alkaline phosphatase; iodoTMT, iodo-Tandem Mass Tags

PMA, phorbol 12-myristate 13-acetate

ROS, reactive oxygen species

Trx-1 + ADAM17cyto WT (-DTT +IAA), and the statistics are represented by asterisk (*), plus (+) and minus (-) signals, respectively. Four independent experiments were performed. Data are represented as mean ± SD. E) 3µg of recombinant proteins (Trx-1 and ADAM17cyto WT complex and ADAM17cyto WT alone) was loaded in a 15% SDS-PAGE gel followed by western blot anti-ADAM17cyto

Western blot under non-reducing conditions of HEK293 cells co-transfected with Trx-1 and ADAM17cyto WT or Trx-1 and ADAM17cyto F730A and treated with 10 mM H 2 O 2 for 10 min

E) Reducing and non-reducing SDS-PAGE followed by Western Blots of endogenous Trx-1 in samples from HEK293/Flp-In cells overexpressing fulllength ADAM17 showed an increase in the monomeric form of Trx-1. The monomeric form was normalized by the total amount of reduced Trx-1. This experiment was performed once. F) Reducing and non-reducing Western Blots of endogenous Trx-1 in scrambled and knockdown ADAM17 A431 cells showed an increase in the monomeric form of Trx-1 in A431 scrambled cells

We appreciated the suggestions in enzyme kinetics from Dr. André Damásio and Carolline Ascensão. We acknowledge the Mass Spectrometry Laboratory, Spectroscopy and Colorimetry Laboratory, and Laboratory of

Protein purification at Brazilian Biosciences National Laboratory, CNPEM, Campinas, Brazil for their support with the analyses.

The authors had no conflicts of interest concerning the topic under consideration in this article. 

The authors had no conflicts of interest concerning the topic under consideration in this article.