Discovery of Spilanthol Endoperoxide as a Redox Natural Compound Active against Mammalian Prx3 and Chlamydia trachomatis Infection

Chlamydia trachomatis (Ct) is a bacterial intracellular pathogen responsible for a plethora of diseases ranging from blindness to pelvic inflammatory diseases and cervical cancer. Although this disease is effectively treated with antibiotics, concerns for development of resistance prompt the need for new low-cost treatments. Here we report the activity of spilanthol (SPL), a natural compound with demonstrated anti-inflammatory properties, against Ct infections. Using chemical probes selective for imaging mitochondrial protein sulfenylation and complementary assays, we identify an increase in mitochondrial oxidative state by SPL as the underlying mechanism leading to disruption of host cell F-actin cytoskeletal organization and inhibition of chlamydial infection. The peroxidation product of SPL (SPL endoperoxide, SPLE), envisioned to be the active compound in the cellular milieu, was chemically synthesized and showed more potent anti-chlamydial activity. Comparison of SPL and SPLE reactivity with mammalian peroxiredoxins, demonstrated preferred reactivity of SPLE with Prx3, and virtual lack of SPL reaction with any of the reduced Prx isoforms investigated. Cumulatively, these findings support the function of SPL as a pro-drug, which is converted to SPLE in the cellular milieu leading to inhibition of Prx3, increased mitochondrial oxidation and disruption of F-actin network, and inhibition of Ct infection.


Introduction
Chlamydia trachomatis (Ct) is one of the most prevalent sexually transmitted infections in the world with close to 130 million new cases arising each year [1]. Ct infections are responsible for a host of diseases including pelvic inflammatory disease, infertility, ectopic pregnancy disease [2], and causing a potential 200-300% increased chance of HIV contraction [3]. Ct infection has also been linked to an increased risk of cervical cancer [4,5]. Although the molecular mechanisms underlying this activity against Ct infection. Studies with recombinant antioxidant proteins, revealed that the SPL E but not SPL reacts preferentially with the mitochondrial protein peroxiredoxin 3 (Prx3), an activity likely to be responsible for the shift to a more oxidative state of host cell mitochondria. Overall, the studies presented here identify the reliance of Ct infection on mitochondrial redox state as the Achilles' heel to be exploited for prevention or treatment of this disease, establish SPL or its SPL E derivative as natural compounds with anti-chlamydial activity, and SPL E as a scaffold for the development of selective inhibitors against mitochondrial Prx3. The data demonstrate for the first time the reaction of peroxiredoxins with endoperoxides and suggest a novel oxidizing activity distinguishing Prx2 and Prx3 from Prx1.

Chemical Synthesis of SPL E
Ethyl (E)-6-oxohex-2-enoate (3). To a solution of 2 [23] (1 g, 6.32 mmol) in CH 2 Cl 2 (50 mL) was added pyridinium chlorochromate (PCC) (1.5 eq., 2.04 g, 9.48 mmol) and silica (2 g). The reaction stirred at 25 • C for 6 h and the solvent was removed in vacuo. The resulting brown powder was filtered through a short silica pad using 5:1 hexane:ethyl acetate to give 3 (0.75 g, 4.8 mmol, 76%), which was used without further purification. NMR data matched that reported in [24]. The NMR spectra of 3 and other chemical products are included in the Supporting Information. 1  Antioxidants 2020, 9, x FOR PEER REVIEW 3 of 17 potent activity against Ct infection. Studies with recombinant antioxidant proteins, revealed that the SPL E but not SPL reacts preferentially with the mitochondrial protein peroxiredoxin 3 (Prx3), an activity likely to be responsible for the shift to a more oxidative state of host cell mitochondria.
Overall, the studies presented here identify the reliance of Ct infection on mitochondrial redox state as the Achilles' heel to be exploited for prevention or treatment of this disease, establish SPL or its SPL E derivative as natural compounds with anti-chlamydial activity, and SPL E as a scaffold for the development of selective inhibitors against mitochondrial Prx3. The data demonstrate for the first time the reaction of peroxiredoxins with endoperoxides and suggest a novel oxidizing activity distinguishing Prx2 and Prx3 from Prx1.

Cell Culture
HeLa cells were purchased from ATCC, USA and cultured in antibiotic-free DMEM containing 10% FBS at 37 °C and 5% CO2.

Formulation of Chlamydia Trachomatis Elementary Body (EB) Stock Solution
HeLa cells were grown to 80% confluency followed by infection with Ct serovars A or D in the presence of 1 μg/mL cycloheximide (Sigma-Aldrich, St. Louis, MO, USA) and 6% glucose (sterile) for 48 h. Following infection, cells were lysed by centrifuging with autoclaved glass beads for ten minutes at 2,000 rpm at 4 °C. The supernatant was collected and subjected to further centrifuging for one hour at 16,000 rpm at 4 °C. The resulting supernatant was discarded. The remaining bacterial pellet was resuspended in ice-cold SPG, a crystallization stock solution made up of succinic acid (S), sodium dihydrogen phosphate (P), and glycine (G). Suspended EBs were aliquoted and stored at −80 °C until use.

Chlamydia Trachomatis Infection
HeLa cells were infected with Ct EBs serovars A or D using MOI 2 in the presence of 6% glucose in antibiotic-free DMEM. Cells were further treated as described below for each individual condition and then processed for analysis. Vehicle control for Ct infection was either SPG or heat-inactivated Ct EBs (56 C, 30 min).

SPL, SPL E , and MitoPQ (MitoParaquat) Treatment of Infected Cells
HeLa cells were incubated with SPL or SPL E at the concentrations indicated in the respective figure legends at the time of Ct infection or at 24 h post infection (hpi) at a MOI of 2. Treatment with vehicle solution (e.g., 0.1% DMSO) was used as control. Treatment with MitoPQ (20 μM) [14] or narciclasine (5 nM) [27] was used to validate the mitochondrial mechanisms.

Cell Proliferation Assay
Cell proliferation was calculated by imaging the cells over time using Essen Bioscience IncuCyte ZOOM live-cell imager and extracting the % confluence from imaging data.

Confocal Imaging
HeLa cells were seeded on 8-well EZ glass slides (Millipore, Temecula, CA, USA) in antibioticfree DMEM. After Ct infection and treatment, the cells were washed with PBS and fixed in 4% paraformaldehyde for 15 min, washed three times with PBS and permeabilized with 0.1% TritonX-100 for 10 min at 25 °C. Cells were further washed with PBS and blocked with 1% BSA for one hour at 25 °C. Cells were incubated with goat anti-Ct EB (Meridian Life Sciences, Saco, ME, USA) targeting the major outer membrane protein (MOMP) overnight at 4 °C, followed by addition of the rhodamine-conjugated anti-goat secondary antibody (Jackson Laboratories, West Grove, PA, USA) for one hour at 25 °C. F-actin was detected with Alexa Fluor 488-phalloidin (green). The slides were then mounted with Vecta shield mounting medium containing DAPI for fluorescence staining of host cell nuclei (Vector Laboratories, Burlingame, CA, USA) and #1.5 glass coverslips. Confocal imaging was performed using a Zeiss LSM880 with a 40×/0.95 objective at 405 nm (DCP-NEt2C, [28]), 488 nm (F-actin), and 561 nm (Ct inclusions) wavelengths. The percent area occupied by Ct inclusions was quantified with ImageJ [29] and data were fitted to calculate EC50.

Analysis of Mitochondrial Protein Sulfenylation by Flow Cytometry
HeLa cells were purchased from ATCC, USA and cultured in antibiotic-free DMEM containing 10% FBS at 37 • C and 5% CO 2 .

Formulation of Chlamydia Trachomatis Elementary Body (EB) Stock Solution
HeLa cells were grown to 80% confluency followed by infection with Ct serovars A or D in the presence of 1 µg/mL cycloheximide (Sigma-Aldrich, St. Louis, MO, USA) and 6% glucose (sterile) for 48 h. Following infection, cells were lysed by centrifuging with autoclaved glass beads for ten minutes at 2,000 rpm at 4 • C. The supernatant was collected and subjected to further centrifuging for one hour at 16,000 rpm at 4 • C. The resulting supernatant was discarded. The remaining bacterial pellet was resuspended in ice-cold SPG, a crystallization stock solution made up of succinic acid (S), sodium dihydrogen phosphate (P), and glycine (G). Suspended EBs were aliquoted and stored at −80 • C until use.

Chlamydia Trachomatis Infection
HeLa cells were infected with Ct EBs serovars A or D using MOI 2 in the presence of 6% glucose in antibiotic-free DMEM. Cells were further treated as described below for each individual condition and then processed for analysis. Vehicle control for Ct infection was either SPG or heat-inactivated Ct EBs (56 • C, 30 min).

SPL, SPL E , and MitoPQ (MitoParaquat) Treatment of Infected Cells
HeLa cells were incubated with SPL or SPL E at the concentrations indicated in the respective figure legends at the time of Ct infection or at 24 h post infection (hpi) at a MOI of 2. Treatment with vehicle solution (e.g., 0.1% DMSO) was used as control. Treatment with MitoPQ (20 µM) [14] or narciclasine (5 nM) [27] was used to validate the mitochondrial mechanisms.

Cell Proliferation Assay
Cell proliferation was calculated by imaging the cells over time using Essen Bioscience IncuCyte ZOOM live-cell imager and extracting the % confluence from imaging data.

Confocal Imaging
HeLa cells were seeded on 8-well EZ glass slides (Millipore, Temecula, CA, USA) in antibiotic-free DMEM. After Ct infection and treatment, the cells were washed with PBS and fixed in 4% paraformaldehyde for 15 min, washed three times with PBS and permeabilized with 0.1% TritonX-100 for 10 min at 25 • C. Cells were further washed with PBS and blocked with 1% BSA for one hour at 25 • C. Cells were incubated with goat anti-Ct EB (Meridian Life Sciences, Saco, ME, USA) targeting the major outer membrane protein (MOMP) overnight at 4 • C, followed by addition of the rhodamine-conjugated anti-goat secondary antibody (Jackson Laboratories, West Grove, PA, USA) for one hour at 25 • C. F-actin was detected with Alexa Fluor 488-phalloidin (green). The slides were then mounted with Vecta shield mounting medium containing DAPI for fluorescence staining of host cell nuclei (Vector Laboratories, Burlingame, CA, USA) and #1.5 glass coverslips. Confocal imaging was performed using a Zeiss LSM880 with a 40×/0.95 objective at 405 nm (DCP-NEt 2 C, [28]), 488 nm (F-actin), and 561 nm (Ct inclusions) wavelengths. The percent area occupied by Ct inclusions was quantified with ImageJ [29] and data were fitted to calculate EC50.

Analysis of Mitochondrial Protein Sulfenylation by Flow Cytometry
HeLa cells were seeded on 6-well culture plates and allowed to reach 80% confluency. Cells were then treated with SPL alone (120 µM) or SPL and Ct (SPL 120 µM, Ct MOI of 2). At 24 hpi, the cells were incubated with 50 µM DCP-NEt 2 C for thirty-minute at 37 • C in complete media to label mitochondrial sulfenylated proteins [28]. Cells were then trypsinized, washed with ice-cold PBS, and fixed in 100% methanol for ten minutes on ice. The cells were washed then three times with FACS wash buffer and finally resuspended in 300 µL FACS buffer for analysis. Flow cytometry analysis was performed at 405 nm wavelength using a BD FACSCanto II flow cytometer.

Transmission Electron Microscopy Imaging
HeLa cells were treated with Ct alone (MOI 2) and Ct in combination with SPL (120 µM). At 24 hpi, the cells were washed with PBS and fixed with 2.5% glutaraldehyde in 0.1 M Millonig's buffer (pH 7.3) for one hour followed by addition of 1% osmium tetroxide in 0.1 M Millonig's buffer for thirty minutes. The cells were washed three times with 0.1 M Millonig's buffer and dehydrated with 50% and 100% ethanol. The cells were then scraped and centrifuged to obtain a pellet. The pellets were processed in propylene oxide two times fifteen minutes each and further infiltrated with 1:1 solution of Spurr's resin and propylene oxide for one hour, followed by a 2:1 solution of resin and propylene oxide for two hours, and finally into a pure solution of resin for one hour. The resin infiltrated pellets were kept overnight in a 70 • C oven and the pellets were sectioned into 90 nm thin sections using a Reichart-Jung ultramicrotome. Sections were stained with uranyl acetate and lead citrate and viewed under FEI Tecnai BioTwin electron microscope (HV = 80.0 kV; direct magnification 9300×).

Recombinant Proteins and Electrospray Ionization Time-of-Flight Mass Spectrometry (ESI-TOF MS)
Human wild-type recombinant Prx1, Prx2 and Prx3, His-tag Prx3, and sulfiredoxin (Srx) were expressed in Escherichia coli and purified essentially as described earlier [30][31][32]. The proteins (100 µM) were reduced with 10 mM DTT in 50 mM ammonium bicarbonate pH 8.0 for one hour at room temperature. Reduced Prx was then passed through a Bio-Spin Bio-Gel P-6 equilibrated with 50 mM ammonium bicarbonate pH 8.0 to remove DTT. An aliquot of the resulting protein was saved as control, and two other aliquots were incubated at room temperature for one hour with 1 mM of either SPL or SPL E . The reaction mixture was further passed through a Bio-Spin Bio-Gel P-6 column equilibrated with 50 mM ammonium bicarbonate pH 8.0, and the protein was analyzed along with the reduced protein by ESI-TOF MS. ESI-TOF MS analysis was performed on an Agilent 6120 MSD-TOF system operating in positive ion mode with the following settings: capillary voltage of 3.5 kV, nebulizer gas pressure of 30 psig, drying gas flow of 5 L/min, fragmentor voltage of 175 V, skimmer voltage of 65 V, and dry gas temperature of 325 • C. Samples were introduced via direct infusion at a flow rate of 20 µL/min using a syringe pump (KD Scientific). Mass spectra were acquired over the range of 500-3200 m/z, then averaged, deconvoluted, and ion abundance quantified using Agilent MassHunter Workstation software v B.02.00.

Data Analysis
Statistical analyses for cell proliferation were performed using Microsoft Excel and all results are displayed as mean ± SEM. The significance of the difference of means was determined by student's t-test and p values of <0.05 were considered significant. P values were further confirmed using the GraphPad QuickCalcs. Flow cytometry data were analyzed as described previously [28,33]. Briefly, effect size was calculated by finding the difference in the means between the control and treatment conditions and dividing by the larger SD of the two groups. Effect sizes were classified as very small (x < 0.2), small (0.2 < x < 0.4), medium (*) (0.4 < x < 0.7), large (**) (0.7 < x < 1.2), and very large (***) (x > 1.3). Data analysis for EC50 calculation was performed in GraphPad and data are shown as ±SEM based on three biological replicates.

SPL Inhibits Intracellular Ct Development
Given our earlier findings of sulforaphane promoting the intracellular development of Ct [14], we started our investigations by testing the pro-or anti-chlamydial (Ct serovar D) activity of SPL in HeLa cells, a widely accepted and validated in vitro model for mechanistic studies of sexually transmitted infections. The effects of SPL were monitored by confocal imaging of intracellular inclusions, the protective vacuoles within which Ct amplifies. The results revealed inhibition of Ct infection by SPL regardless of whether the compound was added to HeLa cells before Ct infection (Figure 1a) or at 24 h post-infection (Figure 1b). The confocal imaging results were further confirmed by monitoring chlamydial heat shock protein 60 (HSP60), an indicator of chlamydial load (Figure 1c), and transmission electron microscopy showing smaller inclusions with a lower number of EBs and RBs ( Figure 1d). Importantly, SPL did not impact host cell proliferation demonstrating it targets Ct-induced events (Figure 1e). The activity of SPL was then further tested using infection of HeLa cells with Ct serovar A, responsible for ocular infections and trachoma (Figure 1f). The results showed comparable anti-chlamydial effects, demonstrating that the activity of SPL was not dependent on a specific Ct serovar. These initial results were very promising as they revealed SPL's potential utilization for both the prevention or treatment of Ct infections. All subsequent studies described in the next sections were performed with Ct serovar D and HeLa cells.

SPL Increases Mitochondrial Oxidative State and Disrupts the F-Actin Ring Supporting the Ct Inclusion
Considering the critical function of mitochondria in the regulation of cellular redox processes, we have focused our subsequent studies on the role of mitochondria redox state in Ct infection and the impact of SPL on mitochondrial protein oxidation. For these studies we have used selective labeling of protein sulfenylation at 24hpi with fluorescent mitochondria-targeted DCP-NEt 2 C probe [28] and performed confocal imaging ( Figure 2a) and flow cytometry analysis (Figure 2b) of cells treated with Ct and SPL alone or in combination. While Ct induced a statistically significant but low increase in mitochondrial protein oxidation, a much higher protein oxidation level was noted in cells treated with SPL (Figure 2a,b). Interestingly, Ct was able to significantly suppress the increase in mitochondrial protein sulfenylation induced by SPL, to a level closer to that observed with Ct treatment in the absence of SPL (Figure 2a,b). In a previous publication, we utilized MitoPQ, a generator of mitochondrial ROS to increase protein sulfenylation in HeLa cells [14] and showed this to be detrimental to Ct infection. Here we sought to further investigate if the increase in mitochondrial ROS and protein sulfenylation by MitoPQ or SPL could disrupt the F-actin cytoskeletal organization critical to the formation of the F-actin ring sustaining the Ct inclusions. First, HeLa cells were treated with MitoPQ in the absence of SPL, and we monitored both the inclusions (red, Figure 2c) and the host cell F-actin (green, Figure 2c). Consistent with previously reported data [14], the results show inhibition of Ct infection, suggesting that the main mechanism of Ct inhibition by SPL is through disruption of mitochondrial redox metabolism. Also notable in the MitoPQ data was the loss of cytoskeletal F-actin organization (Figure 2c). To investigate if this is also a feature of SPL treatment, the same imaging analysis was performed in cells infected with Ct, with and without SPL treatment (Figure 2d), and using narciclasine to rescue mitochondrial metabolism [27] (Figure 2e). Indeed, the results show a general disruption of F-actin organization by SPL and prevention of this activity along with the rescue of Ct intracellular development by narciclasine. Further confocal imaging analysis focusing on the F-actin ring surrounding the intracellular inclusion demonstrates the damaging effects of SPL on this structure when added either at the time of infection or after infection (Figure 2f). induced events (Figure 1e). The activity of SPL was then further tested using infection of HeLa cells with Ct serovar A, responsible for ocular infections and trachoma (Figure 1f). The results showed comparable anti-chlamydial effects, demonstrating that the activity of SPL was not dependent on a specific Ct serovar. These initial results were very promising as they revealed SPL's potential utilization for both the prevention or treatment of Ct infections. All subsequent studies described in the next sections were performed with Ct serovar D and HeLa cells.

The Endoperoxide Derivative of SPL (SPL E ) Is a More Potent Inhibitor of Ct Infections and Similarly Interferes with the Organization of the F-Actin Ring
Given the findings of increased mitochondrial oxidation state with SPL, we proceeded to investigate potential mechanisms leading to these effects. A starting hypothesis consistent with these observations was the possibility of SPL conversion to an endoperoxide derivative SPL E in the mammalian cellular milieu by trapping singlet oxygen ( 1 O 2 ) species similar to what has been proposed in Acmella Ciliata [21] (referred to as dioxacmellamide in [21]). First, we proceeded with the chemical synthesis of SPL E as described in Materials and Methods. This new route to SPL was designed to maximize the ability to prepare derivatives with structural variation in the diene tail and carboxyl substituent in the future. SPL was obtained as a mixture of geometric isomers of which approximately 40% was 2E, 6Z, 8E, the predominant isomer in commercial source of SPL used in this study. The mixture of SPL isomers was then converted to SPL E by singlet oxygenation using Rose Begal as photosensitizer. Purification by TLC gave rac-SPL E , a racemic mixture of the 6R,9S-and 6S,9R-cis enantiomers (a single diastereomer). NMR and mass spectrometry data (Figure 3a; 0.009 delta mass between observed and calculated values) were consistent with this assignment (NMR data described in Materials and Methods and spectra shown in Supporting Information). Next, we tested the activity against Ct infection using the same set of assays as presented for SPL. SPL E was more potent than SPL at inhibiting Ct infection

The Endoperoxide Derivative of SPL (SPL E ) Is a More Potent Inhibitor of Ct Infections and Similarly Interferes with the Organization of the F-Actin Ring
Given the findings of increased mitochondrial oxidation state with SPL, we proceeded to investigate potential mechanisms leading to these effects. A starting hypothesis consistent with these observations was the possibility of SPL conversion to an endoperoxide derivative SPL E in the mammalian cellular milieu by trapping singlet oxygen ( 1 O2) species similar to what has been proposed in Acmella Ciliata [21] (referred to as dioxacmellamide in [21]). First, we proceeded with the chemical synthesis of SPL E as described in Materials and Methods. This new route to SPL was designed to maximize the ability to prepare derivatives with structural variation in the diene tail and carboxyl substituent in the future. SPL was obtained as a mixture of geometric isomers of which approximately 40% was 2E, 6Z, 8E, the predominant isomer in commercial source of SPL used in this study. The mixture of SPL isomers was then converted to SPL E by singlet oxygenation using Rose Begal as photosensitizer. Purification by TLC gave rac-SPL E , a racemic mixture of the 6R,9S-and 6S,9R-cis enantiomers (a single diastereomer). NMR and mass spectrometry data (Figure 3a; 0.009 delta mass between observed and calculated values) were consistent with this assignment (NMR data Figure 2. SPL disrupts the formation of F-actin ring and stability of Ct inclusions by increasing mitochondrial oxidative state. (a) Confocal imaging analysis of mitochondrial protein sulfenylation induced by Ct and SPL (120 µM) alone or when used in combination. Mitochondrial sulfenylated proteins were imaged with DCP-NEt 2 C; (b) Flow cytometry analysis of DCP-NEt 2 C labeled cells to further validate and quantify the effects of Ct and/or SPL treatments. Statistical significance is indicated by the effects size as medium (*, 0.4 < x < 0.7) or large (**, 0.7 < x < 1.2) (n = 2); (c) Disruption of F-actin cytoskeletal organization and Ct inclusions by MitoPQ, a mitochondria-targeted ROS generator. F-actin was detected with Alexa Fluor 488-phalloidin (green) and chlamydial inclusions were detected using anti-MOMP antibody (red); (d-f) The same imaging analysis monitoring Ct inclusions and F-actin was performed to investigate the effect of SPL on Ct inclusions and F-actin network (d), the rescue of SPL detrimental effects by narciclasine (e), and to demonstrate that SPL disrupts the F-actin organization and the ring structure surrounding the Ct inclusion regardless of the time of addition to cells relative to Ct infection (control, Ct infection 24 hpi, 0.1% DMSO). All imaging data are representative of a minimum two biological replicates and were collected at 40x magnification. White arrows point to the F-actin ring.

SPL E Reacts Predominantly with Mitochondrial Antioxidant Protein Peroxiredoxin (Prx3)
Peroxiredoxins are key intracellular enzymes involved in detoxification of oxidants (e.g., H 2 O 2 ) and oxidized species (e.g., lipid peroxides), connecting phosphorylation and redox signaling [34]. These enzymes can be inactivated at high levels of H 2 O 2 or lipid peroxides during turnover leading to the formation of sulfinic acid at the catalytic peroxidatic cysteine (hyperoxidation) or by posttranslational modifications. The hyperoxidized inactive species can be repaired by sulfiredoxin (Srx), restoring their antioxidant activity. Amongst the 6 mammalian isoforms, Prx3 is located in the mitochondria and is the key regulator of mitochondrial H 2 O 2 . Prx1 and Prx2 belong to the same family of 2-Cys peroxiredoxins as Prx3, but are predominantly located in the cytosol. An increase in overall mitochondrial protein sulfenylation (Figure 2a,b) with SPL treatment could ensue if Prx3 is inactivated during SPL/SPL E treatment by either hyperoxidation at the catalytic peroxidatic cysteine residue or by covalent adduct formation with SPL or SPL E . Covalent adduct formation of Prx3 with SPL or SPL E was deemed feasible considering the nucleophilic properties of the peroxidatic cysteine residue (Cys108 in Prx3) and the presence of α,β-unsaturated carbonyl in SPL and SPL E amenable to Michael addition, and the reactive endoperoxide in SPL E . Western blot analysis did not indicate substantial changes in Prx3 hyperoxidation during the time course of SPL treatment of Ct infected cells (Figure 4a). To investigate the possibility of SPL or SPL E reaction with Prx3 and adduct formation resulting in inhibition of Prx3 activity, we performed side-by-side mass spectrometry analysis of recombinant Prx1, Prx2, and Prx3 reaction with SPL and SPL E (Figure 4b). Interestingly, SPL E but not SPL reacted with mitochondrial Prx3, providing an explanation for increased mitochondrial protein sulfenylation and anti-chlamydial activity of SPL E . Further experiments were conducted with a His-tagged version of Prx3, which previous studies using size exclusion chromatography multiple angle light scattering (SEC-MALS) showed to reside in concatenated (two intertwined) dodecamers compared to the dodecameric wild-type reduced Prx3 [35]. Although there was an adduct observed with SPL E , this was of lower abundance compared with wild-type Prx3 (Figure 4c). Lastly, we investigated the reaction of SPL E with sulfiredoxin (Srx), both because of the Srx function in the repair of hyperoxidized inactive 2-Cys peroxiredoxins, and also as another protein containing a reactive cysteine to gauge the relative selectivity of SPL E toward Prx3. SPL E did not react with Srx, the trace amount of adduct observed corresponded to a non-covalently bound species (Figure 4d). Based on the average delta mass for the covalent adduct of Prx3 with SPL E (+251.1 Da) and the lack of reaction with SPL, we propose a reaction mechanism by which the peroxidatic cysteine in Prx3 attacks the endoperoxide, forming a sulfenate adduct (sulfenate ester), which seems to be further oxidized by an unknown mechanism. Interestingly, although the Prx2 adduct also corresponded to a mass shift of +251.1 Da, for Prx1 this was +252.9 (~2 protons). Based on these observations we envisioned a chemical path (Figure 4e), which we further rationalize in the Discussion.

Discussion
Regulation of redox metabolism is one of the cell's primary defense response to bacterial infection [36,37]. Indeed, Ct infections have been reported to induce an increase in intracellular reactive oxygen species (ROS) [38][39][40] (1)(2)(3)(4) in HeLa cells showing that SPL does not increase the level of hyperoxidized Prx in Ct infected cells. Protein stain is shown to illustrate equal loading of cell lysates; (b) ESI-TOF MS data of Prx1, Prx2 and Prx3 after reduction with DTT (black upper spectra), and after the reaction with SPL (red middle spectra) or SPL E (green lower spectra). The results show lack of reaction with SPL and preferred reaction of SPL E with Prx3, as well as formation of a more oxidized adduct with Prx2 and Prx3 (delta mass +251.1) compared with Prx1 (delta mass +252.9); (c,d) The same reaction as in (b) but using a His-tagged construct of Prx3 (c), showing the dependence of the reaction on the oligomeric structure of native Prx3, and using Srx (d) showing lack of reaction with this protein; (e) Potential mechanisms for the reaction of SPL E with reduced Prx1, Prx2, and Prx3.

Discussion
Regulation of redox metabolism is one of the cell's primary defense response to bacterial infection [36,37]. Indeed, Ct infections have been reported to induce an increase in intracellular reactive oxygen species (ROS) [38][39][40], in part mediated by transient activation of NADPH oxidase (NOX) [39,41]. By examining disulfide bond formation and cross-linking in the chlamydial outer membrane, it has been determined that Ct exists in a more reduced environment during early infection, gradually shifting to a more oxidized but non-apoptotic state in later stages of the developmental cycle [42]. Interestingly, there are also reports of chlamydial infections actually protecting cells from H 2 O 2 -induced apoptosis [43].
These findings have led to the exploration of compounds with redox-altering properties for prevention and/or treatment of chlamydial infections (e.g., ascorbic acid [38]). Our efforts focused on examining the potential anti-chlamydial activity of natural compounds with reported anti-inflammatory and antioxidant activities. Surprisingly, the first compound we investigated, sulforaphane-a key compound in broccoli and other cruciferous vegetables, promoted chlamydial infection by suppressing mitochondrial protein sulfenylation and inducing activation of complement C3 in mammalian host cells [14]. Given these intriguing findings, we proceeded to test the anti/pro-chlamydial activity of other natural plant products with reported anti-inflammatory activity. We selected to work with SPL (described in the Introduction) because of its established antimicrobial activity against several intracellular pathogens, unexplored activity against Ct, and its anti-inflammatory and redox-regulatory properties.
The very first studies ( Figure 1) revealed a diametrical opposite anti-chlamydial effect of SPL compared to the infection-promoting activity of sulforaphane [14]. Importantly, the studies showed that this activity was not serovar-dependent being effective against both Ct serovars A and D, and not dependent on the time of treatment suppressing Ct infections when added either at the time of treatment or after the establishment of intracellular inclusions. The data also showed that these effects are not due to the killing of host cells as SPL treatment did not alter cell growth.
As sulforaphane activity was linked in our previous studies to a decrease in mitochondrial protein sulfenylation, the next experiments were directed to quantify the effects of SPL on these species (Figure 2a,b). Indeed, confocal imaging using a mitochondria-targeted chemical probe selective for protein sulfenylation showed significantly increased oxidation with SPL treatment, which then Ct was able to suppress but down to levels that were still higher than those induced by Ct alone. This analysis reveals the very precise regulation of host cell mitochondrial redox metabolism by Ct to a level that is just sufficiently high to support its intracellular development and not detrimental to host cell. The data also shows that Ct can regulate mitochondrial protein sulfenylation both up from basal levels in host cells and down from the SPL-induced levels upon infection. The critical function of mitochondrial metabolism and redox state in Ct infection was further validated with the MitoPQ and narciclasine studies. The induction of mitochondrial ROS and protein sulfenylation by MitoPQ proved detrimental to Ct infection consistent with our earlier findings [14]. Here we show that MitoPQ also induced a loss of F-actin cytoskeletal network and prevented the organization of the F-actin ring around the chlamydial inclusion (Figure 2c), similar to the effects of SPL (Figure 2d). Narciclasine, a natural compound with mitochondrial ROS suppressing activity, rescued the SPL-induced damage to F-actin organization and Ct infection (Figure 2e). Importantly, when SPL was added after establishment of chlamydial inclusions (e.g., at 18 and 20 hpi), it was still able to disrupt the inclusion and inhibit intracellular Ct expansion.
The increase in mitochondrial protein sulfenylation with SPL was puzzling as SPL and the plant extracts from which it originates are generally considered as antioxidants [44]. A hypothesis to resolve this apparent contradiction was that within the cellular environment SPL may convert to an endoperoxide derivative (SPL E ) by trapping 1 O 2 , which would be responsible for its perceived antioxidant activity, and then the SPL E (an oxidant) would induce the increase in protein sulfenylation. The first part of this hypothesis is supported by the identification of SPL E in plant extracts from Acmella ciliata [21], and the chemical synthesis of SPL E from SPL by in situ generation of 1 O 2 described in Materials and Methods. Indeed, SPL E showed stronger activity against Ct infection resulting in disruption of F-actin cytoskeletal organization similar to the effects of SPL (Figure 3).
Finally, we wanted to go one step further and investigate potential targets of SPL or SPL E that would lead to increased mitochondrial protein sulfenylation. The main mitochondrial antioxidant enzyme responsible for the suppression of H 2 O 2 , a reactive oxygen species leading to protein sulfenylation, is peroxiredoxin isoform 3 (Prx3). This enzyme belongs to the 2-Cys class of peroxiredoxins cycling between the reduced, sulfenylated upon reaction with H 2 O 2 or lipid peroxides, and intermolecular disulfide states during the catalytic cycle. The disulfide oxidized protein is reduced back to the active state by the mitochondrial thioredoxin 2/thioredoxin reductase 2/NADPH system. Amongst mammalian Prx isoforms, Prx3 is the most resistant to inactivation by hyperoxidation (Prx-C P -SO 2 H; C P denotes the peroxidatic/reactive cysteine), which can nevertheless occur at high levels of peroxide (H 2 O 2 or lipid peroxides) [30,31,45]. The possibility of repair exists, and it is catalyzed by sulfiredoxin allowing Prx to re-enter the catalytic cycle. Interestingly, side-by-side monitoring of SPL or SPL E adduct formation with recombinant Prx3 and the cytosolic Prx1 and Prx2, revealed that SPL did not react at any significant extent with any of the Prx isoforms, and SPL E reacted preferentially with Prx3. This profile is consistent with the hypothesis of SPL to SPL E conversion within the cell or perhaps more selectively within the mitochondria, and partial inhibition of Prx3 activity leading to the observed increase in host cell mitochondrial sulfenylation. To our knowledge, this is the first report of a peroxiredoxin reaction with endoperoxides. However, the chemical nature of the adducts and the reaction path to these remain to be elucidated. Given the lack of reaction of the peroxiredoxins with SPL, the Michael addition at the α,β-unsaturated carbonyl is unlikely although possibly the binding and/or reaction of the endoperoxide group could position the unsaturated carbonyl for reaction with a protein nucleophile. Assuming the reaction occurs similar to the reaction with H 2 O 2 (H-O-O-H) or lipid peroxides (R-O-O-H), the peroxidatic cysteine could attack at either one of the two oxygens in the R-O-O-R sequence of the endoperoxide leading to a sulfenate ester and a hydroxyl (C P -S-O-R~R-OH). This is close in mass to the adduct observed in Prx1. However, the adduct observed in Prx2 and Prx3 is 2 Da smaller than this suggesting a potential subsequent oxidation event or a different chemical path. We considered the attack of C P -S-O-R~R-OH by a nucleophile at the α,β-unsaturated carbonyl within the Prx monomer, but this reaction does not result in additional loss of 2 Da. This very intriguing observation remains to be explored in future studies.

Conclusions
A large drawback for treatment of Ct infections is the absence of a vaccine and of low-cost drugs, as many patients particularly in third world countries cannot afford the cost of antibiotic treatment. In addition, there are also other problems with antibiotic treatment in developing nations, including the prevalence of expired antibiotics being donated to these countries [46][47][48], unhygienic conditions arising from lack of infrastructure and overpopulation [49][50][51], and the increasing concerns with antibiotic resistance. Overall, the studies presented here identify the reliance of Ct infection on mitochondrial redox state as the Achilles' heel to be exploited for prevention or treatment of this disease, establish SPL or its SPL E derivative as natural compounds with anti-chlamydial activity, and SPL E as scaffold for development of selective inhibitors against mitochondrial Prx3. Funding: This research was funded by Wake Forest Center for Redox Biology and Medicine (AWT, HBH), and NIGMS R01 GM072866 to WTL and CMF. We also want to acknowledge the Kimbrell family for the support of mass spectrometry instrumentation in CMF and AWT's laboratories, Wake Forest University Biology Microscope Imaging Core, and the Wake Forest Baptist Comprehensive Cancer Center (NIH/NCI P30 CA012197) for support of shared resource facilities (Cellular Imaging Shared Resource, Flow Cytometry Shared Resource, and the Proteomics and Metabolomics Shared Resource). Confocal imaging experiments were conducted at the Wake Forest University Microscopic Imaging Core, supported by the Center for Molecular Signaling. The authors declare no competing financial interests.