Conjugating Berberine to aMultidrug Resistance Pump Inhibitor Creates an Effective Antimicrobial

In bacteria, multidrug-resistance pumps (MDRs) confer resistance to chemically unrelated amphipathic toxins. A major challenge in developing efficacious antibiotics is identifying antimicrobial compounds that are not rapidly pumped out of bacterial cells. The plant antimicrobial berberine, the active component of the medicinal plants echinacea and golden seal, is a cation that is readily extruded by bacterial MDRs, thereby rendering it relatively ineffective as a therapeutic agent. However, inhibition of MDR efflux causes a substantial increase in berberine antimicrobial activity, suggesting that berberine and potentially many other compounds could be more efficacious if an effective MDR pump inhibitor could be identified. Here we show that covalently linking berberine to INF55, an inhibitor of Major Facilitator MDRs, results in a highly effective antimicrobial that readily accumulates in bacteria. The hybrid molecule showed good efficacy in a Caenorhabditis elegans model of enterococcal infection, curing worms of the pathogen. E fflux by multidrug-resistance pumps (MDRs) is a universal mechanism by which microorganisms resist a broad variety of antimicrobials (1–4). Bacterial MDRs are found in all microorganisms and make up five distinct independently evolved protein families (5). The first MDR pump described was the humanABC (ATP binding cassette)-family P-glycoprotein transporter (6), which protects a number of tissues from xenobiotics and is an essential component of the blood–brain penetration barrier (7–9). Overexpression of P-glycoprotein plays an important role in tumor resistance to chemotherapeutic agents (10). Although a few bacterial P-glycoprotein homologues have been described, such as the LmrA MDR of Lactococcus lactis (11), bacterial ATP-dependent MDRs are uncommon, and efflux of clinically relevant compounds is due primarily to the drug/proton antiporters of the resistance nodulation cell division (RND) type MDRs of Gramnegative species (12) and major facilitator (MF) MDRs present in all groups of microorganisms (13). Some of the MF MDRs, such as the Staphylococcus aureus QacA pump, are carried on transmissible genetic elements (14). Unlike specialized transporters, MDRs recognize their substrates largely on the basis of polarity. In order to cross the lipid bilayer of the membrane, drugs must be amphipathic, containing both hydrophilic and hydrophobic components, whereas cytoplasmic compounds are hydrophilic, which prevents their escape from the cell (15, 16). Any amphipathic compound could potentially be a toxin, providing a simple basis for an MDR pump to discriminate self from harmful foreign mol*Corresponding authors, [email protected], [email protected]. Received for review June 5, 2006 and accepted September 14, 2006. Published online October 13, 2006 10.1021/cb600238x CCC: $33.50 © 2006 by American Chemical Society ARTICLE ACS CHEMICAL BIOLOGY • VOL.1 NO.9 www.acschemicalbiology.org 594 ecules. The crystal structure of the Escherichia coli RND pump AcrAB revealed an unusually large “binding site” capable of accommodating a vast variety of substrates (17). At the same time, MDRs have preferences, and the best substrates for all studied MDR groups are hydrophobic cations (16, 18). A positive charge allows a molecule to accumulate in the cell, driven by the transmembrane potential. An ability to accumulate up to 1000fold makes cations potentially highly toxic to the cells, and this threat may have been responsible for the origin of MDRs. Benzalkonium chloride is an example of a hydrophobic cation that is widely used as an antiseptic and whose efficiency is limited by MDR efflux (19). It was recently shown that creating a polymer of the hydrophobic cation hexyl pyridinium makes the compound insensitive to MDR efflux (20). Apparently, the pumps can extrude small molecules but not large polymers. This enables creation of effective “sterile surface” materials to prevent the spread of pathogens (21). At the same time, it would be very useful to have small molecules in our arsenal of potential pharmaceuticals that avoid MDR efflux. As described below, plants provide a natural example of a chemical strategy to block MDR efflux, thereby allowing antimicrobial compounds synthesized by the plant to inhibit the growth of microbial pathogens. The alkaloid berberine is a natural product and a hydrophobic cation that is the principal component of the medicinal plants golden seal (Hydrastis canadensis) and echinacea (Echinacea species). Berberine is a potentially excellent antimicrobial, because it accumulates in cells driven by the membrane potential (19) and hits two immutable targets, the membrane and DNA (22). Accumulation of hydrophobic cations in the membrane causes leaks, andberberine is also an excellent DNA intercalator (23). Resistance to berberine is thus unlikely to develop due to target modification. It was previously shown that resistance to berberine is based on MDRs (18); for example, it is readily pumped out of S. aureus cells by NorA, an MDR pump responsible for efflux of cationic antiseptics and fluoroquinolones. However, Berberis species of plants produce, in addition to berberine, 5=-methoxyhydnocarpin (5-MHC), an inhibitor of MF MDRs. 5-MHC strongly potentiates the action of berberine (24). The synergistic combination of an antimicrobial with an MDR inhibitor results in an effective antimicrobial that avoids bacterial resistance. Not surprisingly, the presence of MDRs has been an important impediment in the development of new synthetic antibiotics. One approach to solving the penetration problem has been to develop MDR inhibitors (25–27), similar to the natural strategy that plants use to combat microbial pathogens (24). A potential challenge of this approach, however, is to match pharmacokinetics and other properties of two unrelated molecules. We reasoned that the challenge of developing an efficacious MDR inhibitor could potentially be met by covalently linking an antimicrobial compound with a MDR inhibitor to create a well-penetrating molecule. Here, we report that combining berberine with the MDR inhibitor INF55 produces a novel hybrid antibacterial that is insensitive to MDR efflux. RESULTS ANDDISCUSSION To test the concept of an antimicrobial/MDR inhibitor hybrid, we synthesized a conjugate between berberine, a hydrophobic cation that is an excellentMDR substrate, and INF55, an inhibitor of MF family MDRs (25).