Peter Cresswell's work has generally focused on how the immune system responds to microbes, rather than how microbes infect their hosts. But these two processes are inextricably linked, as he learned from witnessing the food-borne pathogen Listeria monocytogenes co-opting its opponent for its own good. He and his colleagues now report on Listeria's trickery in subverting its host's immune response.

Listeria produces a protein called LLO that injects pores into membranes, allowing the bacterium to escape from intracellular compartments charged with its destruction, and to replicate and spread with abandon. To become active, LLO must be chemically reduced, which the bacterium cannot achieve alone. Cresswell, an immunobiologist at Yale University School of Medicine, and his colleagues report that the host provides the pathogen with a reduction service actioned by the enzyme GILT. In mice in which the gene encoding GILT has been knocked out, Listeria clearance is speedier and the bacterium's growth in cells is drastically impaired (see page 1244).

Cresswell has worked in the field of antigen presentation for decades. Certain immune cells unfold bacterial proteins by exposing them to increasingly acidic environments in intracellular compartments called phagosomes that form when the cell engulfs a bacterium, and through reduction by GILT, which breaks disulphide bonds. This unfolding reveals antigenic 'epitopes' — portions of the protein that are then 'presented' on the cell surface and used as recognition sites by the immune system. In earlier work, Cresswell's group had developed a mouse strain lacking the GILT gene to investigate the enzyme's role in antigen processing and presentation. In a 2001 paper published in Science, the team found that antigen processing was significantly diminished in animals lacking GILT.

Cresswell and his colleagues wondered what else they might learn from the Gilt-knockout mice. “We thought we could look at bacterial infections in the knockout mice and ask if we could see any differences,” says Cresswell. “It was really as simple as that, a real 'look–see' experiment.” If anything, he expected that GILT, which is normally secreted in response to infection, would be missed in the knockout mice, and that their immune responses would be weakened. But the presence or absence of GILT didn't seem to make much difference to infection by several pathogens, including viruses, that the researchers tested — until they came to Listeria.

Cresswell was initially surprised that the Gilt-knockout mice fared better than their normal counterparts against Listeria infection. But he quickly realized what the likely mechanism was — that the reducing enzyme GILT was activating the bacterial toxin LLO in phagosomes. This conclusion relied on two connections. First, Cresswell had worked with SLO, another member of the lysin protein family to which LLO belongs. “We knew that in order to activate SLO, you had to reduce it,” Cresswell explains. Second, because of the work of a former colleague at Yale, he also knew that Listeria used a similar protein to “pop its way out” of phagosomes and into the cytosol where it replicates.

Thus, without GILT, LLO cannot readily escape the phagosome. Instead, it gets killed, digested and processed for priming subsequent immune responses, says Cresswell. As a result, GILT may represent a new therapeutic target for developing a Listeria-specific antibiotic.