Sulforaphane promotes chlamydial infection by suppressing mitochondrial protein oxidation and activation of complement C3

Abstract

Sulforaphane (SFN), a phytochemical found in broccoli and other cruciferous vegetables, is a potent antioxidant and anti‐inflammatory agent with reported effects in cancer chemoprevention and suppression of infection with intracellular pathogens. Here we report on the impact of SFN on infection with Chlamydia trachomatis (Ct), a common sexually transmitted pathogen responsible for 131 million new cases annually worldwide. Astoundingly, we find that SFN as well as broccoli sprouts extract (BSE) promote Ct infection of human host cells. Both the number and size of Ct inclusions were increased when host cells were pretreated with SFN or BSE. The initial investigations presented here point to both the antioxidant and thiol alkylating properties of SFN as regulators of Ct infection. SFN decreased mitochondrial protein sulfenylation and promoted Ct development, which were both reversed by treatment with mitochondria‐targeted paraquat (MitoPQ). Inhibition of the complement component 3 (complement C3) by SFN was also identified as a mechanism by which SFN promotes Ct infections. Mass spectrometry analysis found alkylation of cysteine 1010 (Cys1010) in complement C3 by SFN. The studies reported here raise awareness of the Ct infection promoting activity of SFN, and also identify potential mechanisms underlying this activity.

Abbreviations

4‐HNE

4‐hydroxynonenal

BSE

broccoli sprouts extract

complement C3

complement component 3

Ct

Chlamydia trachomatis

DCP‐NEt₂C

coumarin‐tagged mitochondria‐targeted protein sulfenylation probe

EB

elementary body

hpi

hours post‐infection

HPLC

high‐pressure liquid chromatography

HSP60

heat shock Protein 60

KEAP1

Kelch‐like ECH associated Protein 1

MitoPQ

mitochondria‐targeted paraquat

MOMP

Ct major outer membrane protein

MS

mass spectrometry

NRF2

NF‐E2‐related Factor 2

RB

reticulate bodies

ROS

reactive oxygen species

SFN

sulforaphane

TEM

transmission electron microscopy

Introduction

Chlamydia trachomatis (Ct) is the most commonly reported sexually transmitted infection in the United States.1 Approximately 131 million people are infected with Ct every year with women being diagnosed three times more often than men.1, 2 Ct is obligate intracellular bacteria with a biphasic developmental cycle.3, 4, 5, 6, 7 Elementary bodies (EBs, 0.3 μm size) are the infectious, metabolically repressed form of Ct that differentiates in the host cell into the metabolically active reticulate bodies (RBs, 1 μm size). RBs multiply within protective intracellular vacuoles called inclusions. At approximately 24 h post‐infection (hpi), RBs asynchronously re‐differentiate back to infectious EBs, which are then released from the host cells through extrusion or cell lysis for the next round of infection.8 Under certain conditions (nutrient deprivation, antibiotic treatment) the RBs can also differentiate into a form called persistent or aberrant bodies (ABs, 1.2–1.5 mm size).3, 4, 5, 6, 7 The entire process of Ct development from attachment/entry to extrusion/lysis is regulated by an arsenal of Ct and host cell proteins.9, 10, 11, 12 The overarching theme of our research is the discovery of chemical agents with activity against Ct infections. The current treatment of Ct infections consists of azithromycin or doxycycline antibiotics and costs $10–$20 per patient. However, recent reporting points to emerging concerns of resistance to antibiotic treatment in clinic,13 bringing about the need for new, cost‐effective, and affordable therapeutics.

Naturally occurring compounds constitute a rich and diverse resource and have been used extensively to identify compounds with activity against bacterial infections, cancer, and other diseases.14, 15, 16 Here we focused on sulforaphane (SFN), a phytochemical abundantly present in various cruciferous plants such as broccoli, cabbage, and Brussels sprouts.17 Epidemiological studies over the last decades indicate that consumption of dietary supplements from cruciferous vegetables reduces the risk of prostate, breast, lung, and colorectal cancers.18 In vitro and in vivo studies have also demonstrated inhibition of HIV,19 HCV,20 and Helicobacter pylori 21, 22, 23 infections by SFN. The most widely known mechanism of SFN action is through the activation of KEAP1/NRF2 system triggered by SFN reaction with key regulatory cysteine residues in KEAP1.24 KEAP1 inactivation through this mechanism leads then to increased nuclear localization of NRF2 transcription factor and upregulation of a number of antioxidant proteins and phase II detoxification enzymes that are under the transcriptional control of NRF2.25

Based on the reported anti‐HIV, ‐HCV, and ‐H. pylori effects of SFN, our starting hypothesis for this study was that SFN would repress Ct infection. This hypothesis was also supported by the knowledge that Ct relies on thiol‐mediated chemistry to enter the host cell26, 27 and regulate its developmental cycle,28 and that Ct upregulates host cell reactive oxygen species (ROS), which we and others thought to be needed for its intracellular development.11, 29 Thus, given both the antioxidant and thiol‐alkylating properties of SFN, we aimed to investigate the potential therapeutic capacity of SFN against Ct infections using pure SFN and SFN‐containing broccoli sprouts extract (BSE) [Fig. 1(a)] to mimic dietary intake.

Figure 1.

Sulforaphane and broccoli sprouts extract stimulate chlamydial infection. (a) Chemical structure of sulforaphane (SFN), commonly found in cruciferous vegetables, and image of broccoli sprouts used to generate the SFN‐containing broccoli sprouts extract (BSE). (b) HPLC chromatogram of SFN quantification in BSE. Top: overlay chromatogram of SFN standards and standard curve. Bottom: chromatogram of SFN in BSE. (c) Confocal microscopy imaging of cells infected with Chlamydia trachomatis (Ct) only at 24 h post infection (hpi) or pretreated with either SFN or BSE for 24 h and then infected with Ct for 24 h. Chlamydial inclusions were visualized with anti‐chlamydial major outer membrane protein (MOMP) antibody (red) and nuclei were stained with DAPI (blue). Quantification analysis in the right graph shows statistically significant increases in both inclusion number and size in SFN and BSE pretreated cells compared to cells treated with a vehicle control at 24 hpi (n = 3) and 36 hpi (n = 2) (*0.01 < P < 0.05, **0.001 < P < 0.01). (d) Transmission electron microscopy (TEM) imaging (4800×) comparing the inclusions of Ct infected cells with and without 24 h pretreatment with SFN or BSE (0.5 and 0.8 μM SFN, respectively). The small dense features are the elementary bodies (marked with EB in the middle panel) while the larger lighter gray features are the reticulate bodies (marked with RB in the middle panel). Quantification analysis in the right graph shows statistically significant increases in the occupancy of inclusions in SFN and BSE pretreated cells compared to cells treated with a vehicle control at 24 hpi (n = 2) (*0.01 < P < 0.05, **0.001 < P < 0.01). (e) Western blot analysis and quantification of Ct HSP60 protein showing again that SFN pretreatment results in increased bacterial load in infected cells (*0.01 < P < 0.05; n = 3). (f) Inside–out assay of Ct infected cells pretreated with either SFN or a vehicle for 24 h and imaged at 15, 30, and 60 min post‐infection as well as 2.5 hpi (no statistically significant differences noted at any of the time points post infection between vehicle and SFN treated cells; n = 2).

The data showed that contrary to our hypothesis, SFN actually stimulated Ct infection and that this activity was dependent on both the mitochondria‐localized antioxidant properties of SFN as well as on its reactivity with a key cysteine thiol in the complement component 3 (complement C3) protein. This latter finding also points to a potential mechanism by which SFN exerts its anti‐cancer properties through inhibition of the pro‐tumorigenic complement system, which is a new discovery with broad implications in cancer prevention.

Results

SFN and BSE promote Ct infection

High‐pressure liquid chromatography (HPLC) was used to quantify the amount of SFN in BSE obtained from 4 ounces of broccoli sprouts, a typical serving size in US markets [Fig. 1(b)]. This was found to correspond to 4.7 mg SFN/pound of broccoli sprouts or 1.4 mg SFN/mL BSE (0.8 μM SFN). Thus, the follow‐up studies utilized low micromolar concentrations of SFN (0.5 μM) to examine the effect of SFN on Ct infection. This concentration is lower than what is being typically used to mimic nutritional levels (2–15 μM)30 and within the levels typically measured in blood and urine (100 nM–2.5 μM).30, 31

To elucidate first the consequence of SFN treatment on Ct infection, HeLa cells were pretreated with either SFN or BSE for 24 h, followed by a 24 h infection with Ct. Pretreatment was used to mimic everyday SFN exposure brought on by supplementation with SFN and/or diet. At 24‐ and 36 h post‐infection (hpi), the cells were imaged using an anti‐chlamydial major outer membrane protein (MOMP) antibody to visualize and quantify the number and size of Ct inclusions. As shown in Figure 1(c), there was a significant increase in both the size and number of Ct inclusions when the host cells were treated with SFN or BSE compared with control cells at both time points selected for analysis. Similarly, transmission electron microscopy (TEM) analysis showed bigger and more densely populated Ct inclusions in cells treated with either SFN or BSE compared to non‐treated cells [Fig. 1(d)], a finding further confirmed by Western blot analysis of chlamydial HSP60 protein [Fig. 1(e)]. To further identify if the observed effects on inclusion size and number are derived from SFN regulation of EBs attachment and entry into the host cells, we performed an inside‐out assay using confocal imaging at short timepoints of infection. The results showed no statistically significant difference in the ratio of attached to entered EBs at any of the time points investigated (0.25, 0.5, 1, and 2.5 hpi) while the expected decrease in this ratio was observed within each group over time as more EBs entered the cells [Fig. 1(f)].

Mitochondria redox state drives Ct infection

We and others have previously reported an increase in intracellular ROS with Ct infection,11, 29 leading to the initial thought that Ct‐induced ROS may be needed for Ct infection and intracellular development. As the antioxidant activity of SFN has been soundly established in previous literature,32, 33, 34 and the data in Figure 1 clearly show SFN promoting Ct infection of HeLa cells, raise the possibility that the increase in intracellular ROS by Ct infection was in fact due to the host cell response to Ct invasion and may in fact be detrimental to intracellular Ct development. The revised hypothesis was then that host cell ROS (perhaps partly suppressed by Ct to enable growth and prevent host cell apoptosis during development), actually slows down intracellular Ct development. Thus, antioxidants like SFN would release the brakes on Ct development, while a pro‐oxidant would inhibit this process. To investigate this revised hypothesis, we focused our studies on mitochondria given the prior literature showing NRF2‐dependent mechanism of mitochondria protection by SFN when cells were exposed to H₂O₂,35 and the known function of mitochondria in triggering cell apoptosis. First, we sought to determine whether SFN does indeed decrease mitochondrial protein oxidation in the HeLa cells used for Ct infection. Second, we wanted to investigate the effects of increasing mitochondrial ROS and protein oxidation on Ct infection. Mitochondria‐targeted paraquat (MitoPQ36) and a newly reported mitochondria‐targeted chemical probe for protein sulfenylation (DCP‐NEt₂C37), were used to increase mitochondrial ROS and monitor protein oxidation, respectively. Using these chemical tools, we first set out to establish SFN as a mitochondrial antioxidant with MitoPQ as a positive control. To do this, cells were treated with either 20 μM MitoPQ, 0.5 μM SFN, or both. Untreated cells were used as control [Fig. 2(a)]. As expected, SFN induced a decrease in mitochondrial protein sulfenylation, while MitoPQ increased the level of these species within the mitochondria. When used in combination, SFN was able to suppress the oxidation induced by MitoPQ, suggesting that it would be capable of suppressing potential oxidation induced by Ct. Upon establishment of the mitochondrial redox effects of both compounds, we then determined their impact on chlamydial infection. HeLa cells were infected with Ct alone, pretreated with SFN and then infected with Ct, and infected with Ct in the presence of MitoPQ. Imaging results clearly show increased inclusion size and number by Ct infection, consistent with data in Figure 1. Using cells only infected with Ct as a reference, we quantified the effects of MitoPQ on Ct infection. Indeed, the increased protein sulfenylation resulting from MitoPQ treatment was associated with attenuated chlamydial infection marked by a drastic reduction in both the number and size of the inclusions [Fig. 2(b)]. The MitoPQ effect on extracellular EBs was also assessed by pretreating the cells with MitoPQ and removing it prior to infection with Ct. The results showed comparable levels of Ct suppression based on inclusion number and size leading to the conclusion that the effects of MitoPQ are indeed intracellular [Fig. 2(c)].

Figure 2.

Chlamydial infection is regulated by mitochondrial redox state. (a) Confocal microscopy imaging of mitochondrial protein sulfenylation with DCP‐NEt₂C (cyan) in cells treated with either vehicle control, SFN (0.5 μM), MitoPQ (20 μM), or a combination of SFN and MitoPQ for 24 h. Quantification of fluorescence signal (graph below) found significant differences between all treatments and the vehicle control (*0.01 < P < 0.05, **0.001 < P < 0.01, *** P < 0.001). SFN treatment decreased mitochondrial protein sulfenylation while MitoPQ increased mitochondrial protein sulfenylation, an effect that was countered by concomitant treatment with SFN and MitoPQ. (b) Confocal microscopy imaging of mitochondrial protein sulfenylation using DCP‐NEt₂C (cyan) in cells infected with Ct for 24 h (inclusions were visualized with anti‐MOMP antibody and are shown in pink in the overlay image), pretreated with SFN for 24 h and then infected with Ct for 24 h and treated/infected with MitoPQ/Ct at the same time. Clearly SFN decreased protein sulfenylation and promoted Ct infection consistent with data in Figures 1 and 2(a) (quantification not shown), while MitoPQ treatment increased mitochondrial protein sulfenylation consistent with Figure 2(a) and decreased Ct inclusion size and number as shown in the graph below (*0.01 < P < 0.05, **0.001 < P < 0.01; n = 3). (c) Quantification of confocal imaging data for inclusion size and number comparing conditions when MitoPQ was removed prior to Ct infection or maintained during infection. Both conditions yielded significant reduction in inclusion number and size, with no statistically significant difference between sustained treatment and removal (*0.01 < P < 0.05, **0.001 < P < 0.01; n = 2).

SFN also exerts its anti‐Ct effects by alkylation and inhibition of complement C3

Previous redox proteomics studies in our group on a related project focusing on the mechanisms by which SFN exerts anti‐cancer activity have shown a decrease in complement C3 sulfenylation when cells were treated with SFN and labeled with BP1, a biotin‐tagged chemical probe for protein sulfenylation38 (data not shown). The complement system is a key component of innate immunity and the first line of defense against pathogens though more recent research has also shown a function of the complement system in adaptive immunity.39 Under basal conditions, complement C3 cycles in serum between an intramolecular thioester state (Cys998 and Gln991 residue based on Ref. 40; Cys1010 and Gln1013 based on UniProtKB – P01024), and hydrolyzed state [C3(H₂O)]. The presence of pathogens triggers the complement cascade, which starts with the activation of complement C1 leading to the formation of the C3 convertase (classical pathway). C3 convertase acts on the thioester state of complement C3 breaking it down into the C3a and C3b fragments, but not on the C3(H₂O) state.41 C3b is further needed to cleave C5 into C5a and C5b, which together with other proteins form pores in the bacterial membranes causing the cells to swell and burst (alternative pathway). The thioester bond in C3b is surface exposed enabling binding of C3b to the surface of invading pathogens by reacting with protein amine (‐NH₂), hydroxyl (‐OH), or thiol groups (‐SH) to form covalent amides, esters, or new thioester bonds, and release free ‐SH on Cys1010.

The role of the complement system, specifically C3 in other Chlamydia species has been well established (e.g., C3‐dependent upregulation of B and T cell response was shown to attenuate infection in a murine model of Chlamydia psittaci 42), but studies have focused primarily on complement activity in serum. Less research has been done in general to understand the intracellular function(s) and mechanism(s) of the complement system, and when available this is limited to immune cells.43 Given the reliance of the complement system on thiol chemistry and the known reactivity of SFN with thiols, we hypothesized that SFN may exert its Ct promoting activity by targeting complement C3 within the infected cells. SFN mechanism of action may be through alkylation of reactive Cys1010 in C3(H₂O), which could trap this cysteine blocking thioester formation and consequently C3 proteolysis into the C3a and C3b forms.

To determine first whether there is a function of complement C3 in Ct infection, we treated the HeLa cells with compstatin, a complement C3 inhibitor, in increasing concentrations followed by incubation with Ct for 24 h. As shown in Figure 3(a), both the number and size of Ct inclusions were significantly increased in compstatin‐treated cells compared to Ct alone, whereas treatment with complement C3 showed significantly reduced Ct inclusion size and number compared to control Ct infection [Fig. 3(b)].

Figure 3.

Complement C3 suppresses chlamydial infection. (a) Confocal microscopy imaging of cells infected with Ct (24 hpi) following 24 h pretreatment with increasing concentrations of complement C3 inhibitor compstatin. Statistically significant increase in both inclusion number and size was found in cells treated with compstatin (*0.01 < P < 0.05; n = 3). (b) Confocal microscopy imaging of HeLa cells infected with Ct following 24 h pretreatment with increasing concentrations of human complement C3. Statistically significant decrease in both inclusion number and size was found in cells treated with complement C3 (*0.01 < P < 0.05, **0.001 < P < 0.01; n = 3). In both panels, chlamydial inclusions were visualized with anti‐MOMP antibody (red) and host cell nuclei with DAPI (blue).

Upon finding with these two complementary assays that indeed complement C3 inhibits Ct infection, we sought to elucidate a potential mechanistic interaction between SFN and C3. First, we investigated if pretreatment of host cells with SFN or BSE would relieve the inhibitory effect of complement C3. Indeed, SFN as well as BSE increased both the number and size of Ct inclusions in infected cells when these were also treated with complement C3 [Fig. 4(a)]. To investigate the mechanisms of SFN interference with the activation of complement C3, we performed Western blot analysis of cell infected with Ct, treated with SFN, and treated with SFN prior to Ct infection. As shown in Figure 4(b), Ct suppressed the overall levels of complement C3 and also induced proteolysis to a fragment resembling the size or iC3b, the inhibitory form of C3b which cannot participate in the formation of the membrane‐associated attack complex, a key event in the destruction of infectious pathogens by the complement system. SFN pretreatment did not significantly impact this profile but yet it did suppress the anti‐Ct effects of complement C3 [Fig. 4(a)]. As mentioned above, one of the mechanisms by which complement C3 exerts its function is through cleavage into C3a and C3b by C3 convertase, a process dependent on the presence of the thioester bond. The Cys1010 SFN alkylated state would not be susceptible to proteolytic cleavage as it cannot form the critical thioester with Gln1013 and would effectively block the activation of the complement cascade at C3. To investigate this potential mechanism, we analyzed the modification of C3 by SFN using mass spectrometry. The analysis identified Cys1010 as modified by SFN [Fig. 4(c)], suggesting SFN functions by trapping the Cys1010 thiol group, thus inactivating the unprocessed form of complement C3.

Figure 4.

Sulforaphane inhibits complement C3 activation by alkylation of reactive Cys1010 residues. (a) Pretreatment of cells with SFN and BSE (24 h) overcomes the suppression of Ct infection by complement C3 (*0.01 < P < 0.05, **0.001 < P < 0.01; n = 3; complement C3 was added at the time of infection with Ct). Similar to other figures, chlamydial inclusions were visualized with anti‐MOMP antibody (red) and host cell nuclei with DAPI (blue). (b) Western blot analysis of complement C3 using an antibody against the alpha chain shows overall suppression of complement C3 and conversion to iC3b in Ct infected cells that was not impacted by SFN. (c) Mass spectrometry analysis identifies SFN alkylation of Cys1010 in complement C3. MS/MS spectrum is shown for the alkylated Cys1010 containing peptides with representative b and y fragment ions shown in red and blue, respectively.

Figure 5.

Proposed mechanism of SFN effects on chlamydial infections. SFN alkylation of Cys1010 could trap C3 or C3b in a state that cannot react with chlamydial cell surface proteins.

Discussion

In a previous study, we have shown that Ct infection activates epidermal growth factor receptor (EGFR), a hallmark of cancers associated with sexually transmitted infections (e.g., head and neck cancer, and cervical cancer), and increases intracellular level of ROS.11 We proposed that these findings complemented by other observations such as induction of systemic inflammatory and innate immune response, induction of centrosome abnormalities and genome duplication, and the finding that Ct encodes its own HSP60, which interferes with host cell apoptosis and cellular senescence,44, 45, 46, 47 support a potential role of Ct infections in cancer etiology. We also anticipated that effective therapies against Ct could also work as cancer chemopreventive agents for populations at risk for sexually transmitted infections.

With this study, we set out to determine the effects of a naturally occurring compound, SFN, with documented activity in cancer prevention, on chlamydial infections using a Chlamydia trachomatis (Ct) model. Based on the mechanistically connected antioxidant and thiol alkylating activities of SFN, we hypothesized that SFN would inhibit Ct infection. This study demonstrates that on the contrary, dietary levels of SFN and BSE promote Ct infection acting primarily on the intracellular development of Ct (Fig. 1). Thus, despite the inhibitory effects of SFN on certain bacteria and viruses, SFN has opposite impact on Ct infection. Then, we went on to determine the mechanisms by which SFN was promoting chlamydial infection by focusing on the known biological activity and chemistry of SFN as antioxidant and electrophile reacting with nucleophilic thiols, respectively. Consistent with previous literature,35 SFN demonstrated mitochondrial antioxidant properties, which were linked here to its Ct promoting activity. This conclusion was further supported by the studies with the mitochondria‐targeted pro‐oxidant MitoPQ, which exerted an inhibitory effect on Ct infection (Fig. 2). The anti‐ and pro‐oxidant activities of SFN and MitoPQ, respectively, were monitored in Figure 2 with the mitochondria‐targeted protein sulfenylation probe, DCP‐NEt₂C.37 This probe was previously shown to not interfere with mitochondrial respiration, a key property to avoid potential artifacts. Efforts to monitor selectively H₂O₂ levels in mitochondria with Orp1‐roGFP48 or HyPer49 genetically encoded probes were not successful due to the interference of SFN reaction with the reactive thiols in these reporter proteins.

Next, rather than focusing the studies on the well‐established KEAP1/NRF2 pathway, we wanted to explore other potential protein targets of SFN. As described in the Results, we were guided toward the complement system and in particular complement C3 as a potential mechanistic target by other redox proteomics studies in our group. The initial experiments with a known complement C3 inhibitor, compstatin, or by treating cells with complement C3 confirmed the function of complement C3 in the partial suppression of Ct infection (Fig. 3). While these experiments did not distinguish between extracellular and intracellular complement C3, the data in Figure 4(b) show that these effects are at least partly mediated by the intracellular complement C3. Further mass spectrometry analysis identified the alkylation of reactive Cys1010 in complement C3 by SFN, presumably blocking the activation of the complement cascade at C3. Since in C3b the thioester bond is further exposed and susceptible to hydrolysis,50 it is also possible that in vivo SFN could block the complement cascade at this step. Numerous questions remain to be addressed such as the detailed mechanisms by which Ct itself suppresses the activation of complement C3, whether SFN exerts its activity at other stages of complement activation. Noteworthy, thiol‐disulfide chemistry is key to all stages of Ct infection and Ct expresses a number of thiol‐disulfide oxidoreductases.26, 51 It has been reported that within cells, proteins on the surface of Ct are primarily in reduced state.52 Blocking of these thiols by SFN could also prevent opsonization by the complement system, though other amine or hydroxyl groups could react with the exposed thioester in C3b. It is also possible that Ct would release its own proteases to target the complement system as has been suggested by recent studies.53 Based on our results, we suggest that would be also possible to inhibit activation of the complement cascade at C3 with thiol‐alkylating electrophiles such as SFN. This mechanism will be further explored in future studies.

Is the Ct‐promoting mechanism of SFN relevant to population health? SFN absorption in the human body is determined by source, contents, formulations, and doses.54 Only a relatively small proportion of dietary SFN is absorbed in tissue with nearly 80% of this being secreted in urine.55 Given that Ct is a sexually transmitted disease, this may place people at a higher risk of Ct infection. The levels of SFN and its metabolites in blood and urine were measured in several studies and ranged from ~100 nM to 2.5 μM (e.g., Refs. 30, 31), depending on the type of intervention, individual's metabolism, and to some extent the analytical methods used to measure these compounds. Thus, the concentration of SFN utilized in our studies (0.5 μM) is meaningful from a nutritional standpoint. Knowledge of the SFN effects on Ct infection may result in recommendations to limit the dietary consumption of high SFN‐containing vegetables or supplements at a minimum during the antibiotic treatment of Ct infections.

From a broader perspective, there is also the important question of whether there is a relationship between the antioxidant effects of SFN, mitochondria redox state and complement C3. Several studies have established a correlation between ROS levels and complement C3 expression. Knockout of mitochondrial SOD2 in mice showed increased complement C3 protein in the brain under basal conditions, which was further enhanced after induction of transient focal cerebral ischemia.56 This was also associated with increased 4‐HNE adducts, biomarkers of lipid peroxidation often associated with activation of apoptosis signaling.57 Interestingly, a study analyzing the effects of complement C3 knockout in mice in relation to retinal alterations with aging found decreased 4‐HNE adducts, increased catalase and glutathione reductase antioxidant proteins, increased anti‐apoptotic proteins survinin and Mcl1/Bak, and decreased pro‐apoptotic Caspase 3 activity.58 Thus, these two lines of evidence suggest a bidirectional relationship between intracellular ROS, specifically mitochondrial ROS, and the complement system. Novel inhibitors of the complement system and in particular C3 are highly sought after for numerous diseases including alleviation of inflammatory damage associated with aging, cancer development, and ischemia–reperfusion injury affecting heart, liver and brain functions.59 New analogs of compstatin continue to be developed against complement C3 as well as other inhibitors or biologics targeting other proteins in the complement cascade.60 This makes our discovery of SFN covalent inhibition of the complement system significant to the broader body of studies on the complement system.

Materials and Methods

Reagents and chemicals

Antibodies were obtained from the following sources: goat anti‐Chlamydia trachomatis MOMP (Meridian Life Sciences, Saco, ME), rhodamine red X conjugated anti‐goat secondary antibody (Jackson Laboratories, West Grove, PA), anti‐C3 (ABCAM, Cambridge, United Kingdom), HRP conjugated anti‐rabbit secondary antibody (Cell Signaling Technologies, Danvers, MA), anti‐HSP60 (Santa Cruz Biotechnology, Santa Cruz, CA), and HRP conjugated anti‐mouse secondary antibody (Cell Signaling Technologies, Danvers, MA). Fetal bovine serum (FBS) (Atlanta Biologicals, Flowery Branch, GA), Dulbecco's modified Eagle medium (DMEM)/F12 (Invitrogen, Grand Island, NY) and phosphate buffered saline (PBS) were purchased from Lonza, Walkersville, MD. Sulforaphane (SFN) was obtained from Sigma‐Aldrich, St. Louis, MO and Bicinchoninic acid (BCA) used for measurement of protein concentration was purchased from Thermo Scientific, Rockford, IL. Ethyl acetate, ammonium formate, and dichloromethane were obtained from ACROS Organics (Pittsburgh, PA) and 3,3′‐diindolylmethane was from Santa Cruz Biotechnology, Santa Cruz, CA. Acetonitrile and water used in the experiment were HPLC grade from Fisher Scientific and Sep‐Pak Vac 3 cc (200 mg) certified silica cartridges were purchased from Waters, USA. Compstatin was purchased from R&D Systems, Minneapolis, MN. Complement C3 was purchased from Abcam, Cambridge, UK.

Cell culture

Chlamydia trachomatis (Ct) serovar D strain UW‐3/Cx and HeLa cells were purchased from ATCC, USA. HeLa cells were cultured in DMEM containing 10% FBS and maintained at 37°C and 5% CO₂. Ct EBs were purified as previously described.61 Briefly, HeLa cells were grown to 80% confluency and infected with Ct for 48 h in the presence of cycloheximide. Cells were then ruptured with autoclaved glass beads. Cell lysates were separated by centrifugation. The supernatant was then extracted and centrifuged at 20,000 g and 4°C for 30 min. The resulting bacterial pellet was suspended in ice‐cold SPG, a crystallization stock solution consisting of succinic acid, sodium dihydrogen phosphate, and glycine, aliquoted and stored at −80°C until use.

Preparation of broccoli sprouts extract and quantification of SFN content

Broccoli sprouts were purchased from Sunny Creek Farm in Tryon, NC. Total 60 ounces of fresh broccoli sprouts were homogenized and filtered through 0.22 μm nylon filters (VWR, Radnor, PA) to obtain the broccoli sprouts extract (BSE). BSE was then lyophilized and stored at‐80°C for further use.

To measure SFN content, BSE was further purified using silica SPE (solid phase extraction) cartridge. The silica cartridge was conditioned with 4 mL of dichloromethane, 200 μL of BSE was loaded through the silica cartridges, then the silica cartridge was washed with 4.0 mL of ethyl acetate and the BSE was eluted with 4 mL of dichloromethane. The obtained extracts were evaporated to dryness in a vacuum oven at 45°C for 2 h, and re‐dissolved with 100 μL of 20 mM ammonium formate. The basic reverse‐phase chromatography was conducted using a Waters2695 HPLC and Xbridge C18, 3.5um, 4.6 × 150 mm column and a non‐linear gradient of solvent A (2% acetonitrile, 5 mM ammonium formate, pH 10) and solvent B (90% acetonitrile, 5 mM ammonium formate, pH 10) at a flow rate of 0.5 mL/min: 0–6.5 min: 2–48% B; 6.5–7.5 min: 48–95% B; 7.5–9.0 min: 95% B; 9.0–10.0 min: 95–48% B; 10.0–11.0 min: 48–2% B; 11.0–15.0 min: 2% B. The total running time was 15 min and SFN was monitored by UV detection at 220 nm.

Infection and treatment of HeLa cells

HeLa cells were treated with: 0.5 μM SFN, 4 μL of BSE (equivalent to 0.8 μM SFN), MitoPQ (20 μM), compstatin (10 μM, 20 μM, or 30 μM), or complement C3 (5 nM, 10 nM, or 15 nM) and then infected with Ct with multiplicity of infection of 2 (2 MOI). The treatment with SFN, BSE, MitoPQ, compstatin or complement C3 was maintained during the infection with Ct. Controlled conditions included the cells infected with Ct alone or pretreated with the corresponding vehicle for SFN (PBS), BSE (PBS), MitoPQ (DMSO), compstatin (PBS), and complement C3 (PBS). After completion of the incubation period (before Ct infection or at 24 h post‐infection, hpi), the cells were processed for imaging or Western blot analysis as described below.

Confocal microscopy

HeLa cells were seeded on 8‐well EZ glass slide (Millipore, Temecula, CA) in 10% FBS containing DMEM (no antibiotics). Cells were pretreated and/or infected with Ct as indicated in each figure legend. Cells were washed with PBS and fixed in 4% paraformaldehyde for 15 min, washed three times with ice‐cold PBS and permeabilized with 0.1% TritonX‐100 for 10 min. Cells were further washed with PBS and blocked with 1% BSA for 1 h at room temperature. Cells were incubated with goat anti‐Ct EB (Meridian Life Sciences, Saco, ME) 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) for 1 h at room temperature. The slides were mounted with Vecta shield mounting medium containing DAPI for fluorescence staining of host cell nuclei (Vector Laboratories, Burlingame, CA) and #1.5 glass coverslips. Confocal imaging was performed using a Zeiss 880 scanning confocal inverted microscope at 405 nm (DAPI and DCP‐NEt₂C) and 561 nm (Ct inclusions) wavelengths. For the inside‐out assay, staining of attached versus internalized EBs was done as described previously.11 Briefly, HeLa cells were infected with Ct at different time points in the presence of either SFN or a vehicle control. Cells were then fixed in formaldehyde and blocked with BSA. Following blocking, a FITC conjugated anti‐Ct was used to stain attached EBs. Cells were then washed and permeabilized with TritonX‐100. Following permeabilization, cells were blocked again and probed with anti‐MOMP for internalized EBs as described above.

Transmission electron microscopy

HeLa cells pretreated with SFN or BSE for 24 h were infected with Ct as described above. 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 1 h followed by addition of 1% osmium tetroxide in 0.1 M Millonig's buffer for 30 min. The cells were washed three times with 0.1 M Millonig's buffer and dehydrated with 50% and 100% ethanol. The cells were scraped and centrifuged to get a pellet. The pellets were processed in propylene oxide (ppo) two times 15 min each and further infiltrated with 1:1 solution of Spurr's resin and ppo for 1 h, followed by a 2:1 solution of resin and ppo for 2 h and finally into a pure solution of resin for 1 h. 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.

Mass spectrometry analysis of complement C3 alkylation by SFN

Complement C3 (1.5 μM) was incubated with SFN at a molar ratio of 1:10 (C3/SFN) in 100 μL PBS buffer (pH 7.2) for 2 h at 37°C prior to in‐solution digestion and MS analysis. In‐solution digestion was carried out using trypsin (Thermo Scientific) at a ratio of 1:50 (w/w) at 37°C overnight. The reaction was stopped by adding formic acid at a final concentration of 1%. Prior to MS analysis, the peptides were desalted using Thermo Scientific Pierce C18 spin columns. Eluted peptides were dried using a vacuum centrifuge and then resuspended in 1% formic acid, 5% acetonitrile.

The resuspended peptides were analyzed using a Thermo Q Exactive HF Orbitrap mass spectrometer coupled with a Dionex Ultimate 3000 nanoLC system, which was equipped with an Acclaim PepMap100 Nano Trap Column, and an Acclaim PepMap RSLC chromatography nanocolumn. The peptides were separated using a flow rate of 300 nL/min and the following gradient: 0–30 min: 2–10% B; 30–60 min: 10–18% B; 60–80 min: 18–33% B; 80–80.1 min: 33–95% B; 80.1–83 min: 95% B; 83–83.1 min, 95–2% B; 83.1–90 min, 2% B (solvent A: 95% H₂O, 5% acetonitrile, 0.1% formic acid; solvent B: 20% H₂O, 80% acetonitrile, 0.1% formic acid). The spray voltage was set to 2.0 kV and the temperature of the capillary was 250°C. Full MS scans were acquired in the Orbitrap mass analyzer (m/z 350–1,800, the resolution was 60,000) in positive ion mode. Twenty most intense peaks were fragmented in HCD collision cell with normalized collision energy 35% (resolution for MS/MS scans was 15,000). Data were acquired using the XCalibur v.4.0 software. Identification of SFN alkylated cysteines was performed using Proteome Discoverer v.2.1 with SEQUEST for database searching. FASTA file of the complement C3 sequence was downloaded from UniProt (http://www.uniprot.org/) and imported into the software. Search criteria were as follows: enzyme: trypsin; maximum missed cleavage sites: 2; variable modification: oxidation (M), SFN modification (C); search mode: MS/MS ion search with decoy database search included; precursor mass tolerance: 10 ppm; fragment mass tolerance: 0.02 Da; target false discovery rate (FDR): 0.01.

Statistical analysis

ImageJ62 was used for quantification of all imaging data (confocal and Western blot). For confocal imaging of inclusion size and number, the data is based on minimum two biological replicates, five fields per well, and between 40 and 110 inclusions. Statistical analysis was performed using Microsoft Excel and all results are displayed as mean ± standard deviation. The significance of the difference of means was determined by paired t‐test and P values < 0.05 were considered significant.