Chemical Biology of Infectious Diseases: TB, Toxins, and Biofilms

Overview

Conventional conceptions underestimate bacteria, characterizing them as self-contained and independent organisms lacking communicative ability. But research into the nature of bacterial communities is challenging convention with surprising discoveries of intercellular signaling mechanisms. These evolutionarily fascinating and biologically elegant mechanisms also impinge directly on human health. This theme defined the February 9, 2012, meeting of the Chemical Biology Discussion Group, which explored the Chemical Biology of Infectious Diseases: TB, Toxins, and Biofilms.

Bacteria in nature often form continuous colonies on solid substrates with the cells held together by a secreted matrix of polysaccharides. These "biofilms" show remarkable patterns of coordinated growth in response to a broad range of environmental conditions, including the nature of local food sources and competitors. The experimentally reproducible and sometimes stunning architectural features of biofilms—involuted trenches, serpentine folds, and symmetric spokes—imply that the constituent cells possess sophisticated methods of communication. Infections often involve biofilms, sometimes with multiple species working together to better exploit the host. Bacteria living as biofilms are generally more resistant to antibiotic treatment than their unattached counterparts, at least in part because of the presence of the extracellular matrix. But it is the cells' ability to communicate, called "quorum sensing" (QS) that both poses a medical challenge and may prove the best therapeutic target for many infectious diseases.

All the speakers expressed concern over the state of antibiotic research: resistant bacterial strains are appearing with growing frequency, and the pharmaceutical industry is hesitant to invest in potential drugs that may become ineffective before they return a profit. Kim Janda suggested that researchers should target quorum sensing, an approach he described as a "stealth attack." Janda's laboratory at the Scripps Research Institute dissects the chemistry of QS pathway molecules. He described QS as simply bacteria secreting signaling molecules which they can also detect. For this reason they are called "autoinducers." The local concentration of signal functions as a proxy for cell density, with defined threshold concentrations triggering a variety of responses in bacterial gene expression. QS-regulated processes include not only biofilm formation, but also changes to metabolism and virulence factor secretion, for example.

Janda's talk addressed the characterization of two sets of organic molecules crucial to certain types of QS. First, he reported recent work by his group into Autoinducer-2 (AI-2), a ubiquitous chemical signal derived from 4,5-dihydroxy-2,3-pentanedione, or DPD, and thought to dominate inter-species bacterial communication. DPD interconverts rapidly between linear and cyclic forms, thwarting straightforward analyses of AI-2–utilizing systems. Janda's group began by chemically synthesizing DPD and testing its activity using a standard assay: bioluminescence induction in Vibrio harveyi. Various alkylated DPD derivatives were also evaluated for activity with a β-galactosidase reporter assay in Salmonella typhimurium, yielding both agonistic and synergistically agonistic (with native DPD) products and thereby revealing the chemical properties best suited for further analysis and manipulation in the search for AI-2 inhibitory modulators. Another approach was the synthesis of DPD with chemical "handles," or dendrimers. The resultant molecules are tolerated by bacterial cells and can function as probes. Janda's group optically tracked DPD cellular uptake by fluorescently labeling it via dendrimers: a new technique to assay the transport of signal molecules.

Gram-negative bacteria employ N-acylhomoserine lactones (AHLs) for QS, whereas gram-positive strains use peptides. By chemically synthesizing 3-oxo-C₁₂-homoserine lactone, Janda's group revealed an unexpected degradation product, a tetramic acid, that proved toxic to gram-positive strains but not to gram-negative strains. The tetramic acid also coordinates iron and may shuttle the metal. The in vivo relevance of these findings was confirmed by the detection of the tetramic acid, along with the parent AHL, in human sputum samples. Synthesis of an isomer, 3-hydroxy-homoserine lactone, led to the discovery that Acinetobacter baumannii and Pseudomonas aeruginosa respond to both stereoisomers of the molecule. Janda speculated that this versatility in response to different chemical structures may be evolutionarily advantageous, allowing interspecies cross-talk.

The Janda laboratory brings considerable expertise in antibody research to its QS projects. A major long-term goal of this work is the development of therapeutic antibodies against targets involved in QS signaling. Taking steps towards that goal and building on a solid understanding of the underlying chemistry of QS, the group has identified an antibody that inhibits AHL-mediated green fluorescent protein expression in Pseudomonas aeruginosa at relatively low doses.

Keeping with the notion that exploring chemistry can guide therapeutics, Janda and colleagues have identified molecules that can attenuate the effects of the potent toxin, botulinum neurotoxin. Botulinum neurotoxin, known commercially as "Botox" holds great therapeutic promise, but it has also been identified by the CDC as a category A hazardous agent. The toxin inhibits neurotransmitter release at synapses with a protease that catalytically cleaves "SNARE" proteins, which in this instance convey neurotransmitter-loaded vesicles. There are no available therapeutics to counteract botulinum intoxication, but Janda's group is hunting for inhibitors of the toxin's protease activity. An initial high-throughput screen identified scaffolding containing hydroxamic acid warheads as potential lead inhibitors of the protease, and further modifications based on the molecule's interaction with the botulinum active site pocket may yield a much needed therapeutic to counteract this deadly toxin.

Elizabeth Boon of Stony Brook University studies a seemingly simple substance: nitric oxide (NO). Though the chemical properties of this small molecule are well known, its biological functions are a fertile and relatively new area of study. High NO concentrations are toxic to most organisms, and so it seemed initially paradoxical that low concentrations figure in some eukaryotic signaling pathways. The heme group-harboring soluble guanylate cyclase (sGC) has been established as the eukaryotic NO receptor. Interestingly, recent genomic comparisons have identified bacterial homologues classified into the Heme Nitric oxide and/or OXygen binding (H-NOX) family of proteins. Research in the Boon laboratory is at the forefront of work illuminating the physiological relevance and biochemistry of H-NOX domain proteins.

Bacterial genomes tend to be sectionally organized, with related genes clustered together—often within the same operon, under coordinated transcriptional regulation. The Boon group looked for genes near H-NOX coding regions in Shewanella woodyi and found enzymes involved in cyclic di-GMP signaling. Cyclic di-GMP is a common bacterial secondary messenger, in this case a substrate for the opposing cyclase and phosphodiesterase activities of a diguanylate cyclase complex called swDGC. A high cyclic di-GMP concentration induces biofilm formation while a low cyclic di-GMP concentration induces growth of flagella, a characteristic of free-swimming bacteria. Boon hypothesized that NO signaling mediated by an H-NOX interaction with swDGC controls this interconversion. Her group found that NO suppressed cyclic di-GMP concentrations, but had no effect in an H-NOX mutant. Parsing the two opposing catalytic activities of swDGC, the group showed that cyclase activity was inhibited in the presence of H-NOX and NO and that phosphodiesterase activity was stimulated under the same conditions. These results demonstrate the link between NO, H-NOX, swDGC, and, ultimately, biofilm formation.

NO and H-NOX regulate biofilm formation through cyclic di-GMP signaling. (Image courtesy of Elizabeth Boon)

The Boon group has also demonstrated NO modulation of a classical QS system: the interconversion between biofilm formation and bioluminescence in Vibrio harveyi. This behavior is controlled by the well-characterized Lux proteins, and work in Boon's laboratory showed that LuxQ conveys a signal from H-NOX. First, NO was shown to induce the bioluminescent state but not in H-NOX mutants. Most aspects of Lux-mediated signaling involve phosphorylation, and NO bound to H-NOX was found to inhibit LuxQ phosphorylation. LuxQ was, in turn, shown to transfer phosphate to LuxU, a central protein in the bioluminescence pathway. The biochemical experiments in the Boon laboratory with swDGC and the Lux proteins methodically linked NO to QS processes via H-NOX interactions with established pathways. Boon and her colleagues have also undertaken structural studies and identified H-NOX residues crucial to heme stabilization, elucidating the role of the heme-coordinated iron ion. The observations suggest heme-mediated changes to H-NOX tertiary structure. Future work in this direction may yield a generalizable description of the H-NOX domain, quickly proving as important to signaling in prokaryotes as soluble guanylate cyclase is in eukaryotes.

In the symposium's final talk Kyu Rhee of Weill Cornell Medical College introduced the audience to a powerful and broadly applicable tool for profiling bacteria. Rhee emphasized that antibiotics are not designed but are empirically derived and that bacterial populations evolve quickly. Traditional approaches to drug identification fall short in the antibiotics arena, and Rhee sees an innovation gap. Reiterating the preceding speakers' remarks, he asserted that attempts to defining therapeutic targets must address chemical biology. But in order to do so effectively, researchers must be able to discern differences in bacterial populations quickly and comprehensively. Genomics may be a tempting approach, but it is poorly suited to this task: it may provide a comprehensive list of relevant genes, but functional annotations of those genes are severely limited. Furthermore, as Rhee explained, genomics ignores "chemistry, stochiometry, and kinetics."

Rhee and his colleagues have turned to "metabolomics" as a tool for studying infectious bacteria, especially Mycobacterium tuberculosis, the causative agent of TB. They aim to fine-tune their system for profiling cellular metabolites, with the extracted information showing the chemical state of a bacterial population. They invested significant initial effort in creating the proper analytical platform: a form of chromatography coupled to appropriate reference libraries of molecules that accurately reflect the breadth of the small molecular species exploited by bacteria. As proof of principle, the group has used the technique to elucidate the exploitation of multiple carbon sources by M. tuberculosis under different growth conditions. Other approaches would have been unlikely to uncover this trait. Rhee hopes to apply his metabolomic profiling to probe fundamental questions about the disease.

Use the tab above to find multimedia from this event.

Presentations available from:
Elizabeth Boon, PhD (Stony Brook University)
Kim D. Janda, PhD (Scripps Research Institute)
Kyu Rhee, MD, PhD (Weill Cornell Medical College)