The sequence coding for Arabidopsis PIN1At (At2g18040) was cloned between the NdeI and BamHI restriction sites of pET12a and pET15b, enabling the expression of untagged and N-terminal His-tagged recombinant proteins, respectively. Cys69 residue was substituted into serine, aspartate, or asparagine using the pET15b-PIN1At plasmid as a template, mutagenic oligonucleotides, and the QuikChange site-directed mutagenesis kit (Agilent Technologies). The corresponding variants were named PIN1At C69S, C69D, and C69N. All primers used are listed in Supplementary Table S1.

Protein expression was performed in the E. coli BL21 (DE3) strain, containing the pSBET plasmid, which allows expression of the tRNA recognizing AGG and AGA codons. Cell cultures were progressively amplified up to 2.4L in LB medium supplemented with 50μg/mL of required antibiotics (ampicillin and kanamycin) and grown at 37°C. Protein expression was induced at exponential phase by adding 100μM isopropyl β-D-thiogalactopyranoside for 4h at 37°C. After centrifugation (20min at 6,380×g), the cell pellets were resuspended in about 20mL of 30mM Tris-HCl pH 8.0, 1mM EDTA, and 200mM NaCl (TE NaCl buffer) for the untagged PIN1At protein or 50mM Tris-HCl pH 8.0, 300mM NaCl, and 10mM imidazole for His-tagged versions of PIN1At and conserved at −20°C.

Cell lysis was performed by sonication (3×1min with intervals of 1min), and the soluble and insoluble fractions were separated by centrifugation for 30min at 27,216×g. The purification of untagged PIN1At protein was carried out in three steps. The soluble fraction was precipitated by ammonium sulfate in two steps (from 0 to 40% and then to 80% of the saturation) separated by centrifugation for 30min at 27,216×g. The 40–80% ammonium sulfate-precipitated fraction was subjected to gel filtration chromatography (ACA44 gel, Biosepra) equilibrated with TE NaCl buffer. After dialysis of the fractions of interest against TE buffer and concentration, the sample was applied to a DEAE-Sephacel ion exchange column equilibrated with TE buffer. The recombinant PIN1At protein passed through the DEAE column and was concentrated and dialyzed by ultrafiltration (Amicon cells, YM10 membrane) under nitrogen pressure and stored in TE buffer at −20°C. Since untagged PIN1At cysteinic variants were insoluble, they were expressed as 6-Histidine-tagged proteins. For His-tagged proteins, the soluble fraction was directly loaded on Ni²⁺ affinity columns (Sigma-Aldrich). After extensive washing, the recombinant proteins were eluted by adding 50mM Tris-HCl pH 8.0, 300mM NaCl, and 250mM imidazole. The recombinant proteins were concentrated and dialyzed as above and stored in TE buffer at −20°C. Protein concentrations were determined spectrophotometrically using a molecular extinction coefficient at 280nm of 6,990M⁻¹ cm⁻¹. Other recombinant proteins used in this work such as Arabidopsis NTRB and poplar TRXh1 and h3 have been purified as described previously (Jacquot et al., 1994; Behm and Jacquot, 2000).

The sequences coding for PIN1At and its C69S, C69D, and C69N variants were cloned between the BamHI and XhoI restriction sites of a p426-TDH3 plasmid using primers listed in Supplementary Table S1 (Mühlenhoff et al., 2010). The S. cerevisiae BY4741 (MATa; ura3Δ0; leu2Δ0; his3Δ1; met15Δ0) and ess1^H164R (MATa; ura3Δ0; leu2Δ0; his3Δ1; met15Δ0; ess1-H164R:kanMX) strains were obtained from Euroscarf. Empty or recombinant p426-TDH3 plasmids were introduced in yeast cells by the LiAc method (Gietz and Schiestl, 2007). Cells were grown on synthetic dropout medium with 2% glucose but lacking uracil as a selection marker. For functional complementation assay, tenfold serial dilutions of yeast cell cultures starting from an initial 0.05 O.D. at 600nm were spotted onto synthetic dropout solid medium lacking uracil. Plates were incubated at 20°C for 8days or 37°C for 4days. Complementation of the ess1^H164R strain, which expresses a catalytically deficient enzyme that restricts growth at 37°C, was visualized by the restored ability of transformed cells to grow at this temperature.

The PPIase activity measurements were performed using the chymotrypsin-coupled method described previously with slight modifications (Behrsin et al., 2007). The synthetic substrate N-succinyl-Ala-Glu-Pro-Phe-p-nitroanilide (suc-AEPF-p-NA; Bachem) was dissolved at a stock concentration of 10mM in 450mM LiCl/trifluoroethanol. Reactions were performed at 5°C in a final volume of 500μL TE buffer with 10 to 500μM Suc-AEPF-p-NA in the presence or in the absence of 2μM of His-tagged PIN1At. After 1min of incubation, 25μg of chymotrypsin was added and the velocity of the reaction catalyzed by chymotrypsin onto trans conformers of the substrate was followed at 390nm. The concentration of released p-nitrolamine was determined using a molar extinction coefficient at 390nm of 13,300M⁻¹ cm⁻¹. Rates of PIN1At-catalyzed isomerization were obtained by subtracting the chemical isomerization measured prior to adding PIN1At from the combined chemical and enzyme-catalyzed rate obtained after addition of PIN1At. Apparent k_cat and K_m values were calculated by nonlinear regression using the Michaelis–Menten equation in GraphPad Prism.

To assess the impact of hydrogen peroxide treatment on PIN1At isomerase activity, 200μM PIN1At protein was reduced by 1mM DTT(red) for 1h in TE buffer. Reducing agent was removed by desalting (G25) in TE buffer. Control and oxidation mixtures were performed as follows: 20μM reduced proteins were incubated for 30min in the absence or in the presence of 500μM or 2mM H₂O₂ in TE buffer. The reversibility of treatment with 2mM H₂O₂ was tested by subsequently incubating treated protein with 10mM DTT(red) for 30min. Then, 2μM of untreated or H₂O₂-treated proteins were used, as described above, with 200μM Suc-AEPF-p-NA for isomerase activity. The results were expressed as a percentage of the activity of the untreated PIN1At protein.

Around 10mg of PIN1At was reduced using 30mM DTT(red) in 500μL of 30mM Tris-HCl pH 8.0 for 1h at 25°C. The reduced proteins were then desalted on a G25 column pre-equilibrated with 30mM Tris-HCl pH 8.0 buffer. The measurement of the pK_a of Cys69 was taken at 25°C by incubating 10μM of reduced protein with 200μM 5-iodoacetamido fluorescein (fluorescein IAM) in 200μL of sodium citrate or phosphate buffer ranging from pH 2.0 to 7.5 as described previously (Zannini et al., 2017). Values were transformed into percentage of thiolate and fitted to the following nonlinear regression: % thiolate=Bottom+(Top−Bottom)/[1+10^(Log pK_a – pH×HillSlope)].

In a final volume of 150μL of 30mM Tris-HCl pH 8.0 buffer, 150μM of reduced PIN1At was incubated or not in the presence of 500μM H₂O₂ for 30min at 25°C. After extensive dialysis against 10mM Tris-HCl pH 8.0, the untreated and H₂O₂-treated samples were analyzed by electrospray ionization mass spectrometry (ESI-MS) as described previously.

The sensitivity of PIN1At to H₂O₂ was evaluated by incubating reduced PIN1At (10μM) for 30min at 25°C in a final volume of 400μL of 30mM Tris-HCl pH 8.0 buffer in the presence of 0 to 1mM H₂O₂. Subsequent labeling of protein-free thiol groups by monobromobimane (mBBr) and fluorescence measurement were performed as described previously (Zannini et al., 2019).

300μL reaction mixtures containing 76μM (300μg of protein) reduced PIN1At proteins were incubated for 30min in 30mM Tris-HCl pH 8.0, 200mM NaCl without or with H₂O₂ at concentrations ranging from 100μM to 2.5mM. Reaction mixtures were then injected on Superdex200 10/300 column (GE Healthcare) connected to an ÄKTA-Purifier FPLC (GE Healthcare) and coupled to a MALS detector miniDAWN TREOS II (Wyatt) and a differential refractometer Optilab T-rEX (Wyatt). Data were integrated with ASTRA software to determine the oligomerization state and exact molecular masses.

Oxidized PIN1At was prepared by incubating 1mM reduced protein with 16.7mM H₂O₂ for 30min in 300μL of HEPES 100mM pH 8.0 at 25°C before desalting on G25 column. The formation of disulfide-bridged PIN1At homodimer was then verified by ESI-MS analysis as described above. Oxidation–reduction titration by incubating 10μM of oxidized PIN1At for 2h in mixtures of oxidized and reduced DTT at 2mM was carried out as described previously (Selles et al., 2012).

Reduced PIN1At (10μM) was incubated at 25°C for 30min in the presence of 0 to 2.5mM H₂O₂ in 50μL of 30mM Tris-HCl pH 8.0 buffer. Protein-free thiol groups were alkylated with methoxy-PEG (mPEG)-maleimide of 2kDa as described previously before separating the protein mixture on non-reducing 15% SDS-PAGE (Zannini et al., 2017).

Reduced PIN1At (16μM) was incubated at 25°C for 30min with 500μM H₂O₂ in 20μL of 30mM Tris-HCl pH 8.0 buffer. Then, the whole reconstituted TRX reducing system (200μM NADPH, 100nM Arabidopsis NTRB and 1μM poplar TRXh1 or h3) was added to the 20μL reaction and further incubated 30min at 25°C. The reaction was stopped by addition of non-reducing Laemmli buffer and protein mixtures were separated on non-reducing 15% SDS-PAGE.

It was previously reported that plant parvulins were able to complement the yeast ess1 strains even though they lack the N-terminal WW domain (Metzner et al., 2001; Yao et al., 2001). However, the importance of Cys69 for the physiological role of PIN1At was not examined. To this end, we tested the ability of three PIN1At variants to complement the temperature sensitive phenotype of ess1^H164R strain (Wu et al., 2009). In these variants, the Cys69 was substituted to an aspartate, mimicking the residue present in a few isoforms, but also to serine and asparagine to compare with the work conducted on human PIN1 (Behrsin et al., 2007). Only PIN1At and its C69D variant restored the ability of ess1^H164R strain to grow at the 37°C restrictive temperature (Figure 2).

To investigate further the role of Cys69 in vitro, the variants were expressed as 6-His-tagged recombinant proteins in E. coli, purified to homogeneity and the PPIase activity measured using Suc-AEPF-pNA peptide as a substrate since it was the preferred substrate of the Lotus japonicus PIN1 ortholog (Kouri et al., 2009). The steady-state catalytic parameters were determined for each recombinant protein using increasing substrate concentrations (0 to 500μM; Table 1; Supplementary Figure S3). All proteins exhibited PPIase activity with the Suc-AEPF-pNA peptide. While the turnover numbers were not much affected, the proteins differed in the apparent K_m values for the substrate. This resulted in similar catalytic efficiencies for PIN1At and its C69D variant (1.0×10⁴ and 1.2×10⁴M⁻¹ s⁻¹, respectively) which are 7 to 18-fold higher than the ones of the C69N and C69S variants (1.4×10³ and 5.8×10²M⁻¹ s⁻¹, respectively; Table 1). From these results, we can conclude that an aspartic acid can fulfill the function of the PIN1At cysteine residue at least on the synthetic substrate used here but also in a cellular context as exemplified by the yeast complementation.

We sought to assess next the effect of an H₂O₂ treatment on the PPIase activity of PIN1At. The activity was measured after treating a reduced protein with a 25- or 100-fold excess H₂O₂. A two-fold decrease was observed with the largest excess (Figure 3A). The subsequent treatment of H₂O₂-treated protein with DTT restored the full PPIase activity of PIN1At suggesting that a reversible redox modification was responsible for the inhibition of protein activity. The redox state of protein used for activity assays was analyzed by non-reducing SDS-PAGE after alkylation of the remaining free thiol groups with 2kDa mPEG-maleimide that enables to discriminate reduced and oxidized forms of protein (Figure 3B). At a 25-fold H₂O₂ excess, a covalent homodimer was preferentially observed as oxidized form with an oxidized monomeric form and a reduced monomeric form as well, while at a 100-fold H₂O₂ excess, PIN1At was fully oxidized with an equivalent amount of overoxidized monomeric and disulfide-bridged dimeric forms. A subsequent treatment with DTT reduced the dimeric form leading to the appearance of a reduced monomer in addition to the oxidized monomeric form that was not affected by DTT addition as expected (Figure 3B). These results reveal a relationship between inhibition of protein activity and its overoxidation and suggest that the disulfide-bridged dimer may be still active although less efficient than reduced PIN1At.

Further experiments were performed to specifically follow cysteine oxidation. Hence, 10μM pre-reduced PIN1At was incubated with increasing H₂O₂ concentrations (from 0 to 1mM, i.e., up to a 100-fold excess) and then reacted with mBBr, a fluorescent molecule that alkylates thiol groups. We observed a progressive decrease of the fluorescence as the H₂O₂ concentration increases which reflects an increase of PIN1At oxidation (Figure 4A). From these results, we determined a S_0.5 value of 52±5μM that corresponds to the H₂O₂ concentration necessary to reach half-maximal protein oxidation. The redox state of PIN1At after treatment with 10- to 250-fold H₂O₂ excess was further analyzed by non-reducing SDS-PAGE after alkylation with 2kDa mPEG-maleimide (Figure 4B). At the lowest H₂O₂ excess (10- to 20-fold corresponding to 0.1 and 0.25mM, respectively), a covalent homodimer was preferentially observed as oxidized form, while an oxidized monomeric form became predominant from a 50-fold excess (0.5mM). PIN1At was almost fully oxidized in the presence of 100-fold H₂O₂ excess and mostly under oxidized monomeric form. It does presumably correspond to sulfinic and/or sulfonic acid forms, as observed for human PIN1 (Chen et al., 2015).

To validate this assumption, we performed ESI-MS analysis of the reduced PIN1At (150μM) treated or not with a~3-fold H₂O₂ excess (500μM; Table 2; Supplementary Figures S4, S5). A single species was obtained for the reduced protein with a mass of 12,882.7Da. The decrease of approximately 131Da compared to the theoretical molecular mass corresponds clearly to the cleavage of the first methionine as expected from the presence of an alanine at the second position. For the H₂O₂-treated protein, we detected three monomeric forms (molecular masses of 12,882.8, 12,913.9Da, and 12,930.9Da) and a disulfide-bridged dimer (molecular mass of 25,765Da; Table 2; Supplementary Figure S5). Among monomeric forms, the increase of 31.2 and 48.2Da compared to the reduced monomer suggests the addition of two and three oxygen atoms, respectively, consistent with the formation of sulfinic and sulfonic acids. Overall, these results demonstrate that PIN1At forms disulfide-bridged dimers when H₂O₂ excess is low. This may prevent the irreversible oxidation of the sole cysteine residue present in PIN1At that is observed at higher H₂O₂ concentrations.

These observations were completed by Size Exclusion Chromatography Coupled With Multi-Angle Light Scattering (SEC-MALS) analysis that enables to determine the oligomeric state and the molecular mass of proteins in solution. In these experiments, reduced PIN1At was treated with 1- to 30-fold excess H₂O₂ (0 to 2.5mM) and subsequently analyzed by SEC-MALS (Figure 5). Under reduced form, PIN1At eluted as a single peak with a calculated molecular mass of 12.3kDa (Figure 5). After H₂O₂ treatment, PIN1At eluted as two peaks with calculated molecular masses ranging from 12.2 to 12.5kDa and from 24.9 to 25.5kDa, thus corresponding to monomer and dimer forms, respectively. In the presence of 100μM H₂O₂ (1x excess), PIN1At remains mostly under a monomeric form whereas a dimeric form became predominant at 250μM H₂O₂ (3x excess). At higher H₂O₂ concentrations, the proportion of dimers did not increase further as already noticed in the cysteine alkylation experiments (Figure 4B). Moreover, the elution profile of the monomeric form is modified which suggested that cysteine overoxidation affected the globularity/hydrodynamic features of PIN1At (Figure 5). Hence, altogether these results indicate that a small excess of H₂O₂ promotes cysteine sulfenylation and subsequent formation of disulfide-bridged PIN1At dimer.

We sought to determine the pK_a value of the conserved Cys69 of PIN1At using an alkylation method relying on fluorescent IAM (fluorescein IAM) that reacts with thiolate but not thiol groups. The pH-dependent reaction of reduced PIN1At with fluorescein IAM was followed in different buffers ranging from 2.0 to 7.5. From the titration curve, we obtained a pK_a value of 4.9±0.1 (Figure 6A) indicating that Cys69 of PIN1At will be predominantly present as thiolate at physiological pH and thus susceptible to redox modifications pending it is exposed.

Next, we aimed at determining the redox midpoint potential (E_m) of the intermolecular disulfide bridge formed upon PIN1At dimerization. We thus performed titration with mixtures of reduced and oxidized DTT and determined an E_m value of −313±3mV at pH 7.0 (Figure 6B). This negative value is in the range of those determined for TRX partners and this prompted us to evaluate the ability of a reconstituted physiological TRX system comprising NADPH, Arabidopsis NTRB, and poplar TRXh1/h3 to catalyze the in vitro reduction of the disulfide-bridged PIN1At dimer. Adding catalytic amounts of each TRX isoform successfully reduced PIN1At dimers (Figure 6C). This suggests the existence of a protective mechanism by which the formation of this disulfide prevents the overoxidation of the sole cysteine residue of PIN1At under adverse physiological conditions where H₂O₂ is formed. Reduction by the TRX system might then constitute the second part of this protective mechanism for a rapid enzyme reactivation.

The major difference between multidomain (human PIN1 and yeast ESS1) and monodomain (plant isoforms) parvulins is the presence of the so-called WW domain (Landrieu et al., 2000; Yao et al., 2001). While the WW domain of human PIN1 is important for the complementation of the ess1^H164R strain, the observed complementation by monodomain parvulins from plants lacking the WW domain suggests that plant parvulin proteins adopt a different substrate recognition mechanism (Landrieu et al., 2000; Zhou et al., 2000; Yao et al., 2001). It was proposed that a four-amino-acid insertion within the PPIase domain of monodomain parvulins might substitute for the need of the WW domain as in multidomain parvulins (Yao et al., 2001). In human PIN1, it was reported that the domain interactions increase protein affinity toward peptide ligands (Matena et al., 2013) and that substrate recognition via the WW domain induces dynamic changes that promote PPIase domain activity (Campitelli et al., 2018; Lee and Liou, 2018).

Another question relates to the presence of a cysteine or an aspartate in the active site of parvulins and the role played by these residues. We have thus investigated the role of Cys69 of PIN1At by performing yeast functional complementation with PIN1At variants (C69D, C69S, and C69N) mimicking those of human PIN1 (C113D, C113S, and C113N variants) used in a previous study (Ranganathan et al., 1997; Behrsin et al., 2007). Heterologous expression of PIN1At C69S and C69N did not rescue the growth defects of the yeast strain. It seems that the lower PPIase activity of both C69N and C69S variants is not sufficient to support yeast growth (Figure 2; Table 2) as already observed with the human PIN1 C113N and C113S variants, the latter displaying a 20- to 50-fold decrease of PPIase activity in vitro (Ranganathan et al., 1997; Behrsin et al., 2007).

Both PIN1At C69D and human PIN1 C113D variants are functional and restore the growth of the ess1^H164R mutant in restrictive conditions (Figure 2; Behrsin et al., 2007). Nevertheless, replacing Cys69 by an aspartate residue did not affect the substrate affinity and activity of PIN1At while the human PIN1 C113D variant displays a 6-fold decreased isomerase activity but apparently remains functional enough (Gemmill et al., 2005; Behrsin et al., 2007). This raises the question about the conservation of the catalytic mechanism and/or structural differences between monodomain and multidomain parvulins. Structural analysis of human PIN1 C113S and C113A variants have demonstrated that the dynamics in hydrogen bond network in the active site involving two conserved histidine residues and a threonine residue (Mueller et al., 2011) is modified by the nature of residue present at the position where the cysteine is (Barman and Hamelberg, 2014; Ikura et al., 2019). The effect of C113D mutation in human PIN1 has also been attributed to the putative disruption of the hydrogen bond between His59 and His157 which most probably affects the binding of the phosphate moiety of the substrates (Barman and Hamelberg, 2014; Xu et al., 2014; Wang et al., 2015; Chang et al., 2016). Since these two His are also present in PIN1At, it may be that the presence of the WW domain which helps in the positioning of the substrate in multidomain parvulins is at the origin of the different effects caused by cysteine substitution (Landrieu et al., 2002; Barman and Hamelberg, 2014; Xu et al., 2014). In any case, the fact that a Cys to Asp permutation is allowed at this position supports a mechanism where the Cys is rather required for maintaining the hydrogen-bond network than for playing a direct role in catalysis.

In plants, several members of the PPIase superfamily such as cyclophilin 20–3 (Laxa et al., 2007), FKBP20-2 (Lima et al., 2006) or FKBP13 (Gopalan et al., 2004; see Vasudevan et al., 2015 for review) are subject to redox post-translational modifications (PTMs). In previous studies, PIN1At has also been identified as sulfenylated in response to H₂O₂ treatments (Liu et al., 2014; Wei et al., 2020) and persulfidated in standard growing conditions (Aroca et al., 2017). Similarly, the Cys113 of human PIN1 is prone to redox PTMs such as persulfidation and irreversible oxidation (sulfinylation and sulfonylation), in addition to other PTMs such as phosphorylation, acetylation, and SUMOylation (Chen et al., 2015; Innes et al., 2015; Longen et al., 2016; Chen et al., 2020). Of interest, the irreversible oxidation of human PIN1 was associated with various pathologies as mild cognitive impairment or Alzheimer’s disease (Butterfield et al., 2006; Sultana et al., 2006) and induces a decrease of reactivity and mislocalization impairing the physiological function of the protein (Chen et al., 2015).

Unlike human PIN1, H₂O₂ did not inhibit completely the PPIase activity of PIN1At using Suc-AEPF-pNA peptide as substrate (Figure 3A; Chen et al., 2015). In vitro investigation revealed that oxidation of human PIN1 induced protein instability and aggregation (Bayer et al., 2003). While the protein stability was not investigated here, SEC-MALS analysis did not identify an oxidation-dependent aggregation of PIN1At after an H₂O₂ treatment. Furthermore, the previous study of PIN1At 3D structure by NMR revealed that Cys69 undergoes very little changes upon addition of the small phosphorylated Cdc25 peptide AcEQPLpTPVTDL (Landrieu et al., 2002). According to the activity assays, it appears that an overoxidized monomer is inactive, whereas the disulfide-bridged dimer is still active in vitro and does not abolish the interaction with the small Suc-AEPF-pNA peptide (Figure 3). Nevertheless, we consider possible that it might prevent protein–protein interactions with larger molecules such as Arabidopsis auxin transporter PIN1, and MADS-domain transcription factors, AGL24 and SOC1 (Wang et al., 2010; Xi et al., 2016) and affect its physiological function.

The results obtained with both PIN1At and human PIN1 suggest that the sulfenylation formed in response to H₂O₂ may be a general behavior of cysteine-containing parvulins. In the case of PIN1At, this H₂O₂ sensitivity correlates with the low pK_a value of 4.9 of Cys69 as it was proposed for human PIN1 (Behrsin et al., 2007) but not experimentally confirmed. Given the rather instable and reactive nature of sulfenic acids, this is only an intermediate toward the formation of a disulfide-bridged dimer in PIN1At. Whether this is a unique feature among parvulins remains to be investigated further because it was not described for human PIN1 which thus becomes overoxidized. Overoxidation of PIN1At occurs also in vitro when the H₂O₂ excess increases. Hence, disulfide bond formation could represent a mechanism to protect the sole thiol group of PIN1At against irreversible oxidation by H₂O₂. Alternatively, the cysteine of human PIN1 and PIN1At may be protected by the formation of persulfides which would occur if for instance hydrogen sulfide reacts with the sulfenic acid (Cuevasanta et al., 2015). The fact that protein persulfidation protects against cysteine oxidation has been already illustrated for several proteins (Wedmann et al., 2016; Moseler et al., 2021). For a fully protective and reversible mechanism, reduction of these oxidized species should occur. The in vitro data presented here suggest that TRX may be the physiological reductant of PIN1At dimer, the redox potential of the intermolecular disulfide being in the range of those usually reported for TRX targets.