Rottlerin Inhibits Chlamydial Intracellular Growth and Blocks Chlamydial Acquisition of Sphingolipids from Host Cells

Abstract

We report that rottlerin, a plant-derived compound known to inhibit various mammalian kinases, profoundly inhibited chlamydial growth in cell culture with a minimal inhibition concentration of 1 μM. The inhibition was effective even when rottlerin was added as late as the middle stage of chlamydial infection cycle, against multiple Chlamydia species, and in different host cell lines. Pretreatment of host cells with rottlerin prior to infection also blocked chlamydial growth, suggesting that rottlerin targets host factors. Moreover, rottlerin did not alter the chlamydial infection rate and did not directly target chlamydial protein synthesis and secretion. The rottlerin-mediated inhibition of chlamydial replication and inclusion expansion correlated well with the rottlerin-induced blockade of host cell sphingolipid trafficking from the Golgi apparatus into chlamydial inclusions. These studies not only allowed us to identify a novel antimicrobial activity for rottlerin but also allowed us to uncover a potential mechanism for rottlerin inhibition of chlamydial growth.

The genus Chlamydia represents a diverse group of obligate intracellular bacteria and consists of multiple species, each causing various health problems in humans and animals. C. trachomatis infection in the human urogenital tract is a leading cause of sexually transmitted diseases (22), while respiratory infection with C. pneumoniae can lead to serious airway symptoms in immune-compromised individuals and is also associated with cardiovascular diseases (4, 16). Other chlamydiae, such as C. muridarum, C. caviae, and C. psittaci, infect mainly animals. Despite the difference in tissue tropism, all chlamydiae share a common biphasic growth cycle, with the infectious elementary bodies (EBs) gaining entry into and exiting out of host cells and the metabolically active but noninfectious reticulate bodies (RBs) replicating inside cytoplasmic vacuoles (also called inclusions) of the host cells (13). Since Chlamydia has to complete EB-to-RB differentiation, biosynthesis, and RB replication, as well as RB-to-EB conversion within the inclusions, the expansion of a chlamydial inclusion, starting with a single EB at the beginning of infection and ending with thousands of organisms at the conclusion of a growth cycle, is considered a hallmark of chlamydial growth. Chlamydia has to acquire both nutrients and energy from host cells for maintaining its own biosynthesis, replication, and inclusion expansion. Analysis of chlamydial genome sequences has revealed that Chlamydia has short-circuited many of its biosynthetic pathways and acquired the ability to obtain metabolic intermediates and macromolecules from the host cells (24). It is known that Chlamydia has to import lipids from host cells to maintain its growth (3, 5, 15, 25). For example, the chlamydial acquisition of sphingolipids from host cells was elegantly demonstrated by monitoring the trafficking of sphingomyelin synthesized from fluorochrome-labeled ceramide in the Golgi apparatus into chlamydial inclusions (14, 15).

Despite the fact that chlamydial infection can be readily treated with antibiotics, the search for nonantibiotic inhibitors that can be applied topically to block chlamydial growth at the site of infection remains a priority. Various small molecules have been studied to determine their abilities to block sexually transmitted infections; these molecules include Toll-like receptor agonists (1), metalloprotease inhibitors (2), and inhibitors of bacterial type III secretion systems (30). These molecules exhibit antimicrobial activity either by activating the host innate immunity receptor-mediated host defense responses or by directly targeting microbial components. For example, the Toll-like receptor 9 ligand CpG DNA-enhanced Th1-dominant and chlamydial antigen-specific host response can prevent chlamydial infection-induced pathology in a mouse urogenital infection model (8). Here we report that rottlerin [5,7-dihydro-2,2-dimethyl-6-(2,4,6-trihydroxy-3-methyl-5-acetylbenzyl)-8-cinnamoyl-1,2-chromene], a small molecule isolated from the pericarps of Mallotus philippinensis and known to have anticarcinogenic properties (7, 26) and to inhibit various kinases (10), also profoundly inhibited chlamydial intracellular growth. The chlamydial growth inhibition also correlated with rottlerin-induced blockade of sphingolipid trafficking from the host cell Golgi apparatus into chlamydial inclusions.

To determine the dose of rottlerin for inhibition of chlamydial growth, HeLa 229 cells were treated with rottlerin at concentrations ranging from 0.1 to 10 μM at the time of chlamydial infection. The chlamydial inclusions were monitored 48 h after infection using an immunofluorescence microscope assay as previously described (11, 18). As shown in Fig. 1A, there were only scattered chlamydial cells but no clear inclusions in the cell samples treated with rottlerin at a concentration of 1 μM or higher, suggesting that there was complete inhibition of intracellular chlamydial growth. At a concentration of 0.25 or 0.5 μM, rottlerin partially inhibited chlamydial growth, as indicated by smaller inclusions. However, at a concentration of 0.1 μM, rottlerin did not affect chlamydial growth. It is worth noting that rottlerin did not affect the infection rate even at a concentration of 10 μM, suggesting that rottlerin mainly restricted chlamydial replication inside cells without affecting chlamydial entry into host cells. The scattered chlamydial particles in HeLa 229 cells treated with rottlerin at a concentration of 1 μM or higher did not colocalize with the lysosomal marker LAMP-1 (Fig. 1, panel q). The lack of association with LAMP-1 may suggest that the organisms were still viable. However, most of these scattered organisms were no longer infectious since <100 inclusion-forming units per well was recovered (by titrating the homogenates on fresh monolayers), while the control well contained >10⁷ inclusion-forming units (data not shown). These observations suggest that rottlerin may be able to induce a chlamydial inhibition/killing mechanism independent of phagolysosomal fusion, for example, by blocking chlamydial nutrient uptake. It is not clear why the noninfectious organism-laden vacuoles failed to fuse with LAMP-1 vesicles after 48 h of incubation in the presence of rottlerin. This lack of fusion may have been partially due to the fact that vacuoles containing nonviable chlamydiae were shown to fuse slowly with LAMP-1 vesicles in HeLa cells (21). In addition, rottlerin may also have inhibited the vesicle fusion events via its kinase inhibition activity, which could further delay the fusion between chlamydial vacuoles and LAMP-1 vesicles.

FIG. 1.

Inhibition of chlamydial growth by rottlerin. (A) HeLa monolayer cells grown on coverslips were treated with rottlerin at the concentrations indicated at the top. The cultures were simultaneously infected with C. trachomatis serovar L2 at a multiplicity of infection of 1. The cell samples were fixed and permeabilized 48 h after infection for immunofluorescence staining. Chlamydial cells and inclusions were visualized by using a rabbit antichlamydial antibody plus a goat anti-rabbit immunoglobulin G antibody conjugated with Cy2 (green), while the host cell cytoplasm and nucleus were visualized using an F-actin-specific dye (Alexa-Fluor 568 Phalloidin, red) and a DNA dye (Hoechst 33345, blue), respectively. Note that no inclusions were found in samples treated with rottlerin at a concentration of 1 μM or higher. The scattered chlamydial cells (panel q, green) in the sample treated with 1 μM rottlerin were also costained with a mouse anti-human CD107a (LAMP-1) antibody plus a goat anti-mouse immunoglobulin G antibody conjugated with Cy3 (red). Note that there was no obvious colocalization of the chlamydia-laden vacuoles and LAMP-1-positive vesicles despite the lack of chlamydial replication. (B) Rottlerin (at a final concentration of 1 μM) was added to HeLa cell cultures at various times after chlamydial infection, as indicated at the top and bottom. Forty-eight hours after infection, the cell samples were processed as described above for immunofluorescence imaging. Note that the inclusions were significantly smaller in samples treated with rottlerin beginning 0 to 24 h after infection than in the control sample (48 h without rottlerin). (C) Five different Chlamydia species, as indicated at the top, were used to infect HeLa cells at a multiplicity of infection of 1, and rottlerin was added at a final concentration of 1 μM at the time that the cells were infected (panels a to f) (beginning cycle or 0 h) or at the middle cycle (panels g to l) (16 h for serovars L2 and D, MoPn, GPIC, and 6BC and 48 h for AR39). At various times after infection, as indicated at the bottom, the cell samples were processed for image analysis as described above for panel A. The arrows indicate intracellular chlamydiae. Note that rottlerin completely inhibited chlamydial replication when it was added at beginning cycle and severely restricted inclusion expansion when it was added at the middle cycle. (D) C. trachomatis serovar L2 was used to infect five different human cell lines, as indicated at the top, and rottlerin was added to the cultures as described above for panel C. The cells were processed at 48 h after infection for immunofluorescence microscope analysis. Note that rottlerin inhibited chlamydial replication and inclusion expansion in all cells.

When the antichlamydial activity of rottlerin was evaluated at different time points after chlamydial infection (Fig. 1B), we found that rottlerin treatment significantly reduced the chlamydial inclusion size even when rottlerin was applied as late as the middle cycle (up to 24 h after infection), suggesting that rottlerin can rapidly block chlamydial inclusion expansion. We also demonstrated that the rottlerin inhibitory effect was not unique to one strain or species of Chlamydia and that all chlamydial species tested were inhibited (Fig. 1C), suggesting that rottlerin inhibited a common requirement for chlamydial replication. Finally, we found that the rottlerin inhibition of chlamydial intracellular growth was not restricted to a given host cell line and that infection with chlamydial serovar L2 was inhibited in all cell types tested regardless of the time when rottlerin was applied, at the beginning or middle cycle of infection (Fig. 1D).

Given rottlerin's profound inhibitory effects on chlamydial growth, we next evaluated whether rottlerin directly targeted chlamydial biosynthesis or acted on host reactions that could indirectly affect chlamydial replication. We pretreated HeLa cells with 1 μM rottlerin for 4 h, and the treated cells were thoroughly washed with fresh medium to remove any residual extracellular rottlerin prior to chlamydial infection. To our surprise, this procedure resulted in complete inhibition of chlamydial replication (Fig. 2A, panel a), which suggests that rottlerin may irreversibly inhibit host cell machinery or molecules required for chlamydial intracellular replication. As a control, the C₁ compound, which is known to target the chlamydial type III secretion apparatus (30), did not affect chlamydial growth when it was applied similarly to host cells 4 h before infection (Fig. 2, panel g), but it suppressed chlamydial growth when it was added 2 or 16 h after infection (Fig. 2, panels h and i). The C₁ compound seemed to induce aberrant growth of chlamydiae, especially when the drug was added 16 h after infection (Fig. 2, panels f and i). The enlarged-body morphology is very similar to the morphology observed in persistent cultures by electron microscopy (9). However, more experiments are necessary to further characterize the C₁ compound-induced chlamydial growth at the molecular level. It will be interesting to test whether shutting down the type III secretion pathways is a mechanism used by Chlamydia cells to convert themselves into a persistent state during natural infection. Rottlerin rapidly blocked chlamydial inclusion expansion since the sizes of inclusions in samples treated with rottlerin at various time points after infection were similar to the sizes of inclusions visualized at the corresponding time points when rottlerin was added (Fig. 2B). Despite the small inclusions in rottlerin-treated samples, the accumulation of IncA (inclusion membrane protein A encoded by C. trachomatis open reading frame CT119, known to be secreted to the inclusion membrane), CPAF (chlamydial protease/proteosome-like activity factor encoded by CT858, secreted into host cell cytosol), and HSP60 (heat shock protein 60) and the secretion of IncA and CPAF were noticeable, as revealed by antibody staining, suggesting that rottlerin did not actively inhibit chlamydial protein biosynthesis and secretion. Together, the observations described above led us to conclude that rottlerin may target host cell factors or machinery required for chlamydial replication and inclusion expansion. Since Chlamydia is known to depend on host lipids for meeting its own biosynthesis requirements, we evaluated the effect of rottlerin on host lipid trafficking into chlamydial inclusions (Fig. 3). A fluorochrome-labeled ceramide trafficked to the Golgi apparatus when it was added exogenously to a HeLa cell culture (Fig. 3A, panel a). It has been shown that a fluorochrome-labeled ceramide can be incorporated into sphingomyelin in the Golgi apparatus (12). Consistent with results reported previously (15), we found that the inclusions were highly enriched in the labeled sphingomyelin (Fig. 3, panels b to d), suggesting that sphingomyelin incorporating the labeled ceramide trafficked from the Golgi apparatus to the chlamydial inclusions. More importantly, treatment of the cultures with rottlerin completely prevented the sphingomyelin accumulation in the chlamydial inclusions, while the accumulation of the sphingolipid in the Golgi apparatus of the same infected or adjacent normal cells was not affected by the rottlerin treatment at all (Fig. 3B, panels b and f). In controls, brefeldin A (BFA) partially inhibited chlamydial acquisition of the sphingolipid, while nocodazole did not inhibit chlamydial acquisition of the sphingolipid, although both compounds affected host Golgi apparatus morphology (Fig. 3B, panels c, d, g, and h).

FIG. 2.

Rottlerin inhibits chlamydial growth by targeting host cells. (A) HeLa cells grown on coverslips were infected with C. trachomatis serovar L2, and at various time points after infection or 4 h prior to infection, as indicated on the left, either rottlerin or the C₁ compound was added to the cultures at the concentrations indicated at the top. The cell samples pretreated with inhibitors for 4 h were thoroughly washed to remove residual drugs prior to infection. At 48 h after infection, all cell samples were processed for immunofluorescence staining. The chlamydiae (green) and nuclear DNA (blue) were visualized as described in the legend to Fig. 1, and a mouse anti-IncA monoclonal antibody (clone BB2) plus a goat anti-mouse immunoglobulin G antibody conjugated with Cy3 were used to visualize chlamydial IncA protein (red). Note that pretreatment with rottlerin but not pretreatment with the C₁ compound inhibited chlamydial growth, although both inhibitors blocked chlamydial growth when they were added after infection. (B) HeLa cells infected with C. trachomatis serovar L2 were either treated with 1 μM rottlerin (panels a to d, I to l, and q to t) or processed for the immunofluorescence assay (panels e to h, m to p, and u to x) at various time points after infection, as indicated at the top. All rottlerin-treated samples were processed for antibody staining at 48 h after infection. The chlamydiae (green) and nuclear DNA (blue) were visualized as described in the legend to Fig. 1. Mouse anti-IncA (clone BB2), anti-CPAFct (clone 100a), and antichlamydial HSP60 (clone BC7.1) monoclonal antibodies plus a goat anti-mouse immunoglobulin G antibody conjugated with Cy3 (red) were used to visualize chlamydial IncA, CPAF, and HSP60, respectively. Note that although all the rottlerin-treated samples were fixed and observed at 48 h after infection, the inclusion sizes were similar to those in samples observed at the time when rottlerin was added.

FIG. 3.

Rottlerin blocks sphingomyelin trafficking from host Golgi apparatus to chlamydial inclusions. HeLa cells grown on coverslips were infected or not infected with C. trachomatis serovar L2 at a multiplicity of infection of 1. (A) At various time points after infection, the cultures were pulsed with BODIPY-FL-C5-ceramide [N-(4,4-difluoro-5,7-dimethyl-4-bora-a,4a-diaza-s-indcene-3-pentanoyl)sphingosine] (green) at a final concentration of 5 μM at 4°C for 30 min, which was followed by a 1-h back-exchange with 0.34% defatted bovine serum albumin at 37°C. After the coverslips were washed with Hanks’ balanced salt solution, they were immediately observed with a fluorescence microscope. Note that BODIPY fluorescence (green) accumulated in both the host Golgi apparatus and chlamydial inclusions during the entire infection course, suggesting that BODIPY-FL-C5-ceramide trafficked to the Golgi apparatus and that the sphingomyelin synthesized from BODIPY-FL-C5-ceramide was recruited into chlamydial inclusions. (B) At 16 h postinfection, the inhibitors rottlerin (1 μM), BFA (1 μg/ml), and nocodazole (10 μg/ml), as indicated at the top, were added to the cultures. After growth for an additional 8 h at 37°C, all samples were subjected to BODIPY-FL-C5-ceramide labeling and visualization with a fluorescence microscope as described above. Note that rottlerin blocked the accumulation of BODIPY-FL-sphingomyelin in the chlamydial inclusions without affecting the trafficking of BODIPY-FL-C5-ceramide into the host cell Golgi apparatus. The pound sign indicates the host cell nucleus, the arrowhead the host cell Golgi apparatus, asterisks the chlamydial inclusions enriched with the labeled sphingomyelin, and arrows the chlamydial inclusions lacking the labeled sphingomyelin.

Rottlerin has been evaluated in both animal models (32) and human cells (7, 26, 33) to determine its ability to inhibit various pathological conditions. However, its antimicrobial activity has not been described previously. Here we present convincing evidence showing that rottlerin is a powerful inhibitor of Chlamydia that targets host reactions required for intracellular chlamydial replication. First, rottlerin can completely inhibit chlamydial growth at a concentration as low as 1 μM, which is more than 10 times lower than the cytotoxic concentration. In fact, no significant cytotoxicity was observed when rottlerin was used at a concentration of 10 μM in the current study (data not shown). Second, rottlerin has a wide time window for inhibiting chlamydial growth, from prior to infection to the middle cycle during infection, which makes this compound relevant as a topically applied antichlamydial agent. Third, rottlerin is effective against all Chlamydia species and in different cell lines. This is probably due to the rottlerin-mediated inhibition of the shared requirement of all chlamydiae to acquire lipids from host cells to maintain chlamydial growth. The rottlerin antimicrobial effect may also be very specific to the genus Chlamydia since chlamydiae seem to be the only type of microbial pathogens that are known to acquire host lipids by intercepting the Golgi apparatus lipid exocytic pathway (14). Finally, rottlerin can selectively block sphingolipid trafficking from the Golgi apparatus to chlamydial inclusions without affecting ceramide trafficking to and sphingomyelin synthesis in the Golgi apparatus, as shown by the normal accumulation of fluorochrome-labeled sphingomyelin in the Golgi apparatus of rottlerin-treated cells regardless of infection. It is unlikely that rottlerin blocked the normal trafficking of sphingomyelin from the Golgi apparatus to cytoplasmic membrane of host cells since rottlerin-treated cells showed normal cytoplasmic membrane accumulation of the fluorochrome-labeled sphingomyelin and a prolonged rottlerin treatment (up to 4 days) was not toxic at the concentration at which chlamydial acquisition of host sphingolipids was completely inhibited.

The next question is what mechanism(s) rottlerin uses to block chlamydial acquisition of host sphingolipids. Although BFA also inhibits chlamydial uptake of sphingolipids (15), BFA alters Golgi apparatus membrane trafficking by targeting the small GTPases, such as Arf1 (31), and no kinase inhibition activity has ever been described for BFA. However, rottlerin is a known kinase inhibitor, and its targets include protein kinase C delta (PKCδ), calmodulin-dependent protein kinase III, and p38-regulated/activated kinase (6, 10). Clearly, rottlerin may use a different mechanism to block chlamydial uptake of lipids. Rottlerin can also act as a mitochondrial uncoupler (23), induce apoptosis of human tumor cell lines, and potentiate chemotherapy-induced cytotoxicity (7, 26). A recent study showed that PKCδ, one of the major kinases targeted by rottlerin, was selectively recruited around chlamydial inclusions (27). Although the inclusion recruitment of PKCδ may prevent PKCδ from exerting its proapoptotic activity in the mitochondria (20, 27), the localization of PKCδ near the inclusions may also serve as a platform for Chlamydia to utilize the PKCδ activity to facilitate chlamydial acquisition of host lipids. This hypothesis is supported by the finding that PKCδ plays an essential role in various exocytosis processes, including antigen receptor-induced lytic degranulation (19), insulin secretion (28), and sphingomyelin transportation from the Golgi apparatus to the plasma membrane (29). In addition, PKCδ can also activate the Raf-MEK-ERK-cPLA2 signaling pathway (17), which is required for chlamydial acquisition of glycerophospholipids from host cells (25). Inhibition of PKCδ by rottlerin may represent an effective way of shutting down chlamydial acquisition of host lipids.

Although we correlated the rottlerin-induced blockage of ceramide trafficking with the inhibition of chlamydial growth, the blockade of chlamydial uptake of ceramide alone cannot explain the robust inhibition of chlamydial intracellular growth in rottlerin-treated cells. This is because inhibition of ceramide trafficking into inclusions by BFA is not sufficient to block production of chlamydial infectious progeny (14). Rottlerin may alter multiple host pathways required for intracellular chlamydial growth. For example, rottlerin may simultaneously block CD63⁺ multivesiclar body trafficking, a redundant pathway utilized by Chlamydia to acquire lipids (3). Finally, it is also possible that rottlerin may exert direct killing effects on chlamydiae in addition to its alteration of host cell signaling pathways. Testing of these and other hypotheses is under way in order to pinpoint the precise mechanisms of the rottlerin antichlamydial activity.