Antibacterial Flavonoids from Medicinal Plants Covalently Inactivate Type III Protein Secretion Substrates

Abstract

Traditional Chinese Medicines (TCMs) have been historically used to treat bacterial infections. However, the molecules responsible for these anti-infective properties and their potential mechanisms of action have remained elusive. Using a high-throughput assay for type III protein secretion in Salmonella enterica serovar Typhimurium, we discovered that several TCMs can attenuate this key virulence pathway without affecting bacterial growth. Among the active TCMs, we discovered that baicalein, a specific flavonoid from Scutellaria baicalensis, targets S. Typhimurium pathogenicity island-1 (SPI-1) type III secretion system (T3SS) effectors and translocases to inhibit bacterial invasion of epithelial cells. Structurally related flavonoids present in other TCMs, such as quercetin, also inactivated the SPI-1 T3SS and attenuated S. Typhimurium invasion. Our results demonstrate that specific plant metabolites from TCMs can directly interfere with key bacterial virulence pathways and reveal a previously unappreciated mechanism of action for anti-infective medicinal plants.

Figure 1

Baicalein from Scutellaria baicalensis (Huángqín) extract inhibits SPI-1 T3SS. (A) Scheme for SopE2-CPG2-HA:Glu-CyFur reporter system for monitoring type III protein secretion. SopE2-CPG2-HA (SPI-1 T3SS) reporter encodes an enzyme activity of carboxypeptidase G2 (CPG2) that when fused to the C-terminus of a known S. Typhimurium T3SS bacterial effector (SopE2) can be secreted and specifically report on type III protein secretion through cleavage of fluorogenic substrates (Glu-CyFur). (B) Screen of TCM activity at 5 mg/mL using SopE2-CPG2-HA:Glu-CyFur reporter system. TCMs were incubated with bacteria for 4 hr and growth media was collected and analyzed for CPG2 activity using Glu-CyFur. CyFur fluorescence was monitored at 610 nm with excitation at 563 nm. (C) Structure of baicalein. (D) Dose-dependent activity of baicalein on SPI-1 T3SS (SopE2-CPG2-HA) reporter in S. Typhimurium growth media and cell lysate. (E) Dose-dependent effect of baicalein on the levels of SPI-1 T3SS secreted proteins (SipA, SipB, SopB, SipC, and SipD) and flagella components levels in S. Typhimurium growth media. The identities of the polypeptides were confirmed by in-gel digestion and LC-MS/MS analysis.

Figure 2

Survey of naturally occuring flavonoids against type III protein secretion reporter. (A) Structures of six major classes of flavonoid family: I, flavone (showing the positions of A, B, and C ring), II, flavonol, III, flavanone, IV, isoflavone, V, flavanonol, and VI, proanthocyanidin. Profile of T3SS inhibitory activity of 28 compounds representing 6 different classes of flavonoids at 25 μM using SopE2-CPG2-HA reporter system. Mean ± s.d., n = 3. (B) Chemical structures and IC₅₀ values of most active T3SS inhibitors measured with the SopE2-CPG2-HA reporter fluorescence assay. Mean ± s.d., n = 3.

Figure 3

Active flavonoids inhibit SPI-1 T3SS-mediated bacterial invasion of epithelial cells. (A) Left: Flow cytometry analysis of flavonoids at 100 μM on S. Typhimurium invasion of HeLa cells judged by anti-S. Typhimurium antibody staining. MOI = 10, 30 minute infection. Experiment was done in triplicate and similar results were seen in two independent runs. Right: Baicalein and quercetin blocked the invasion of S. Typhimurium into HeLa cells in a dose-dependent manner (darker color: 100 μM, lighter color: 50 μM). Mean ± s.d., n = 3. The values have been normalized to those of DMSO treated, which were considered 100 %. (B) Growth of S. Typhimurium in quercetin abolishes bacterial invasion of cultured epithelial cells. S. Typhimurium was grown in the presence of quercetin (labeled “wt grown in quercetin”) or, alternatively, quercetin was added shortly (30 min) before bacteria were added to cultured epithelial cells (labeled “wt in quercetin). The ability of S. Typhimurium to enter into cultured epithelial cells was evaluated by the gentamicin protection assay as described in the Supplemental Methods. Values are the mean ± standard deviation of three independent measurements and represent the % of the inoculum that survive the gentamicin treatment because of their intracellular location. The values have been standardized to those of wild type, which were considered 100 %. The indicated p values were determined by the standard Student t test. * indicates p <0.001 (change panel C for the revised version, which includes *). (C) Immunofluorescence analysis of intracellular S. Typhimurium in HeLa cells with 100 μM flavonoids. Scale bar = 10 μm.

Figure 4

Chemical proteomic analysis of alk-baicalein labeled proteins in S. Typhimurium. (A) Alkyne-flavonoid probes for activity-based protein profiling studies. (B) Effect of alkyne-flavonoid probes on the levels of SPI-1 T3SS secreted proteins in S. Typhimurium culture growth media. (C) In-gel fluorescence analysis of alkyne-flavonoid probe labeling of S. Typhimurium proteins following click chemistry reaction with az-rho fluorescence dye.

Figure 5

Validation of alk-baicalein labeled SPI-T3SS substrate, SipA. (A) Summary of S. Typhimurium proteins that were labeled by alkyne-flavonoid probes. The total number of peptide-sequence matches (PSM) between alk-baicalein and the inactive probes, alk-trimethylbaicalein, alk-oroxylin A were compared to DMSO control. The graph shows the 16 proteins that were labeled by alk-baicalein with more than 5-fold PSM enrichment compared to control samples. (B) Alk-baicalein labeling of SipA-HA in wild-type (wt) and T3SS-deficient (invA) S. Typhimurium. SipA-HA was immunopurified from alk-baicalein treated S. Typhimurium, reacted with az-rho and analyzed by in-gel fluorescence and anti-HA western blot. Coomassie blue staining of bacterial culture supernatants was performed in parallel to assay on protein secretion. (C) SipA-HA and SipA(1-270)-HA was immunopurified from alk-baicalein treated S. Typhimurium and analyzed as described in (B).

Figure 6

Binding and circular dichroism analysis (CD) of SipA-flavonoid interactions. (A) SipA domains. (B) Spectra of SipA-GST fusion purified from the E. coli treated with flavonoids. Quercetin and baicalein treated SipA-GST had a shoulder at the 370 nm region, while the inactive flavonoid chrysin and untreated SipA did not show this spectral feature. (C) Spectra of the N- and C-terminal domains of SipA purified with and without quercetin. Quercetin-treated SipA N-terminal construct exhibits a peak at 370 nm, while the signal sequence mutant (27-271) and C-terminal domain-GST fusion did not. (D) CD spectra of SipA untreated and expressed in the presence of flavonoids. CD spectra of SipA-1-271-GST expressed without the presence of flavonoids showed a characteristic α-helical pattern. Treatment with 50 μM quercetin and baicalein induced significant spectra change at 208 nm and 222 nm. However, addition of quercetin to purified SipA-1-271-GST in vitro did not induce changes in CD spectra.