Discovery Shows How Fungi Pump Up Defense

Modern medicine was in for a dramatic change that September day in 1928 when Alexander Fleming discovered penicillin. Here was a substance that could literally tear down the walls of bacteria that had besieged humans for millennia. The news almost a century later is that bacteria are, once again, gaining the upper hand. Some have found clever ways to defend their borders. One strategy is to equip the cell wall with tiny pumps capable of expelling not just one but a wide variety of antibiotics soon after they enter.

Microbes belonging to another phylogenetic kingdom—the fungi—have adopted a similar defense. Though not as well publicized, multidrug-resistant versions of Candida and other fungi have been on the rise over the past few decades, invading the mucosal linings and bloodstreams of people whose immune systems have been ravaged by AIDS, chemotherapy, or transplantation procedures. A team of researchers drawn from across HMS and other institutions has made a discovery that could help turn the tables on this emerging microbial threat.

Like bacteria, multidrug-resistant fungi act swiftly—some have been caught transcribing pump-making genes only minutes after exposure to antifungal drugs. Researchers have uncovered some of the transcriptional proteins involved, but what has not been clear is how the microbes detect the toxin in the first place.

“Nobody had really figured out exactly how they can respond to activating chemicals,” said Anders Naar, HMS assistant professor of cell biology. Some wondered whether the drugs might activate sentinel proteins lying below the surface of the cell.

It now appears that the drugs cut a more direct path. Jitendra Thakur, Haribabu Arthanari, and Fajun Yang, working with Naar, Gerhard Wagner, and colleagues, exposed two species of yeast—one harmless and the other pathogenic—to radiolabeled ketoconazole, a common antifungal. The researchers discovered that the drug travels right to the inner sanctum of the cell, the nucleus. Once inside, it binds to a protein sitting on the DNA, setting in motion a series of gene-activating steps. First, the protein, Pdr1, undergoes a conformational change, which allows it to interact with a huge piece of transcriptional machinery, the Mediator complex. This interaction then leads to the expression of the efflux-pump–making gene. The findings appear in the April 3 Nature.

What has researchers marveling is not just the simplicity of the arrangement but the feeling that they have seen it before. And they have, in a very different setting. Human cells detect hormones, as well as toxins, through a similar set of steps. Exogenous substances travel directly to the nucleus, where they bind to a class of transcription factors, known as nuclear hormone receptors. But these receptors were thought to have evolved much more recently.

“One thinks of nuclear hormone receptors as very mammalian-specific or higher eukaryote factors because the things a lot of them respond to are very specialized hormones,” said Kevin Struhl, the David Wesley Gaiser professor of biological chemistry and molecular pharmacology at HMS. “They’re probably derived from an ancient family that was really just set up to interact with various small molecules. To me that’s quite unexpected and interesting.”

What excites Naar and colleagues is the possibility that targeting Pdr1—essentially thwarting its pump-producing activities—might render once-resistant fungi vulnerable. Preliminary experiments carried out in a worm model developed by Eleftherios Mylonakis suggest this might be so.

It’s a long road from the worm lab to the human clinic, but even the possibility of a new drug target is welcome news. Though Candida can be a nuisance when it enters the mucosal linings, causing vaginal infections or oral thrush, it can become devastating when it enters the bloodstream. Unchecked by the immune system, it may develop filaments that poke through blood vessels and enter tissues, eventually killing the host. Mortality rates can go as high as 40 or 50 percent for such systemic infections. To avoid such devastation, doctors may give immunosuppressed patients prophylactic doses of fluconazole, one of the handful of antifungals that exist. But the practice can promote the appearance of resistant strains.

“In our efforts to help patients, we’re ‘birthing’ another class of resistant fungi,” said Mylonakis, HMS assistant professor of medicine at Massachusetts General Hospital.

Outsmarting deadly microbes was not on the Naar lab to-do list when they began their experiments. Naar, along with Yang and Thakur, were primarily interested in understanding the mechanics of transcription. In 2006, Yang, HMS instructor in medicine at MGH, identified a small domain on a subunit of the human and yeast transcriptional co-activator, Mediator. He wanted to know how the domain, dubbed KIX, worked and set out to explore the question in a common species of yeast, Saccharomyces cerevisiae.

He began fishing for partners for KIX, exposing it to thousands of yeast proteins. Remarkably, only one stuck. “It’s almost unheard of to find only one. And it was Pdr1,” Naar said. It turns out, nuclear hormone receptors—the subject of Naar’s graduate thesis—do occasionally have exclusive binding partners. Naar knew that one such receptor, PXR, binds directly to foreign substances such as toxins and that the binding leads to the transcription of a gene for a class of pump proteins called ATP-binding cassette (ABC) transporters. Could Pdr1 be acting in a similar way? Could it be binding the drug and then, through its interaction with KIX, be turning on pump-making genes?

To find out, Thakur tried exposing purified yeast Pdr1 to radiolabeled ketoconazole. Sure enough, the drug bound to Pdr1. Thakur, HMS research fellow in medicine at MGH, showed that the binding caused Pdr1 to undergo a conformational change, exposing a KIX-binding domain. The interaction between Pdr1 and KIX was further illuminated by Arthanari, HMS research fellow in BCMP, and Wagner, the Elkan Blout professor of biological chemistry and molecular pharmacology, using NMR spectroscopy.

S. cerevisiae is resistant to drugs but harmless to humans. Working with Brendan Cormack at Johns Hopkins, the researchers deleted the genes for Pdr1 and the KIX-containing subunit, Gal11, from a pathogenic multidrug-resistant fungal species, Candida glabrata. “We found you need both Pdr1 and Gal11 to stimulate multidrug resistance in the Candida,” said Naar.

In a final series of experiments, the researchers, working with Mylonakis and colleagues, introduced the PDR1- and the GAL11-deleted strains of C. glabrata separately into the worm species C. elegans. Normally, C. glabrata kills worms in six days, even in the presence of antifungals. The PDR1-deleted and the GAL11-deleted versions of C. glabrata turned out to be much more vulnerable adversaries. In the presence of fluconazole, a potent antifungal, the worms survived much longer, suggesting the yeast were no longer resistant to the drug.

“So you could cure the worm patient or at least decrease lethality,” said Naar. Though it is not possible to delete the PDR1 gene from yeast prior to human infection, the Pdr1 protein might be inactivated by small molecules after the yeast enter the body. There is a well-known precedent for this kind of approach. The breast cancer drug tamoxifen targets a nuclear hormone receptor, essentially locking it in an inactive position.

Attacking Pdr1 while at the same time delivering a standard antifungal such as fluconazole could help turn the tide in the battle against multidrug-resistant candida and other fungi. Naar and Wagner are currently working with ICCB-Longwood to explore the possibility. “Our finding proves the concept of co-therapeutics. So you can take the standard azole that no longer works because the bug is multidrug resistant, and then you just add a small molecule that takes out this system,” he said. “Now it’s sensitive again—it prevents all the pumps from being expressed. This is the premise of our approach at ICCB.”