Vanderbilt Institute of Chemical Biology



Discovery at the VICB







Harnessing Light to Kill Bacteria


By: Carol A. Rouzer, VICB Communications
Published:  August 9, 2017



A compound that activates heme biosynthesis sensitizes bacteria to light-dependent toxicity.



Infections of skin and soft tissues impart a very high burden on our healthcare system, leading to 14 million ambulatory care visits and affecting 10% of hospitalized patients in the United States each year. Gram-positive bacteria, including Staphylococcus aureus, Staphylococcus epidermidis, Propionibacterium acnes, and Bacillus anthracis, are the cause of most of these infections, and the rise in antibiotic resistance renders them more difficult to treat. Thus, a search for new antimicrobial targets is ongoing, leading Vanderbilt Institute of Chemical Biology members Eric Skaar, Gary Sulikowski, and Borden Lacy along with their laboratories to investigate the use of a novel small molecule stimulator of heme biosynthesis in conjunction with photodynamic therapy to combat Gram-positive skin and soft tissue infections [M. C. Surdel et al., (2017) Proc. Natl. Acad. Sci. U.S.A., published online July 24, DOI:10.1073/pnas.1700469114].


For many years, scientists believed that heme biosynthesis occurs through the same pathway in all species. However, recent research demonstrated that the pathway in Gram-positive bacteria diverges from that in other species following the synthesis of coproporphyrinogen III (CPGIII, Figure 1A). In Gram positive bacteria, oxidation of CPGIII to copropophyrin III (CPIII) by coproporphyrinogen oxidase (CgoX) is the next step, whereas in other organisms CPGIII is converted to protoporphyrin IX by different enzymes. This opens the possibility of designing antimicrobial agents that work by selectively targeting CgoX or enzymes that catalyze later unique steps in Gram-positive bacteria. The Skaar lab had an early advantage in this endeavor due to their prior search for small molecules that activate the HssRS heme-sensing two-component system in S. aureus. Activation of HssRS occurs upon accumulation of excess intracellular heme, which is toxic to living cells. The result is expression of the HrtAB transporter, which serves as a heme efflux pump. The Skaar lab had used a high-throughput assay for HssRS activation to screen a small molecule library, leading to the discovery of VU0038882 ('882, Figure 1B). Further studies demonstrated that '882 activates HssRS by promoting heme biosynthesis.




FIGURE 1. (A) Heme biosynthetic pathway showing steps thought to be found in all organisms (blue) and steps limited to Gram-positive bacteria (red). (B) Structure of '882. (C) Design of a suicide strain of bacteria to find bacteria resistant to '882. The two genes encoding HrtAB are replaced with genes encoding the E. coli RelE toxin.(D) Mutant bacteria were designed that either could not produce HrtAB (top) or that produced the relE toxin under control of the hrtAB promoter (bottom). Neither heme nor '882 markedly affected growth in the absence of HrtAB expression, but both compounds suppressed the growth of the toxin-producing cells. (E) the T183K mutant of CgoX supports bacterial growth as well as the wild-type (WT) enzyme, whereas deletion of the enzyme leads to substantial growth suppression. Image reproduced by permission from M. C. Surdel et al., (2017) Proc. Natl. Acad. Sci. U.S.A., published online July 24, DOI:10.1073/pnas.1700469114. Copyright 2017, M. C. Surdell, et al.



To understand how '882 affects heme biosynthesis, the investigators designed a suicide strain of S. aureus in which the genes encoding HrtAB were replaced by genes encoding the RelE toxin from E. coli (Figure 1C). Thus, these bacteria produced the toxin in response to high levels of intracellular heme or any other inducer of the HssRS system such as '882, resulting in marked growth suppression (Figure 1D). The researchers incubated these bacteria in the presence of '882 and isolated the small number of surviving colonies. Total genome sequencing revealed that the resistant bacteria expressed a T183K mutant CgoX enzyme. Further studies demonstrated that the bacteria could synthesize heme normally, but they did not respond to '882 with increased heme production (Figure 1E).


These findings suggested that '882 modulates heme biosynthesis, likely at the step catalyzed by CgoX. Support for this hypothesis came from an analysis of heme biosynthetic intermediates in wild-type S. aureus exposed to '882. These bacteria exhibited high levels of CPIII, the product of CgoX. Notably, CPIII is the only fluorescent heme biosynthetic intermediate in Gram-positive bacteria, and the '882-treated cells were, indeed, highly fluorescent. Further studies demonstrated that '882 directly stimulates the activity of CgoX in vitro but fails to modulate the otherwise normal activity of the T183K mutant. Medicinal chemistry efforts identified three compounds that more potently activated CgoX, but they showed no advantage over '882 in intact cells, likely due to poor bioavailability.


Examination of the crystal structure of CgoX from B. subtilis revealed that T183 is located on a short helix that faces a cleft leading to the active site. Nearby on the same loop is N186 (Figure 2). Both N186Y and N186F mutations led to increased CgoX activity, suggesting that this loop plays a role in modulation of the enzyme. These two enzymes and one bearing an N186A mutation were all resistant to the effects of '882. Unfortunately, the investigators were not able to obtain a crystal structure of CgoX complexed with '882, so they used computational approaches to dock the molecule into the region close to T183 and N186. The results identified additional residues that might contribute to the CgoX-'882 interaction, so they created mutants of these residues to evaluate their potential roles. Of these, V146M, Y171A, F184A, and D450Y were unstable proteins. However, M167F and F187W proved to be active enzymes, and both were resistant to the effects of '882.



FIGURE 2. (Top) Model of the CgoX from B. subtilis showing the location of the proposed binding site for '882. (Bottom) close up of the '882 binding site revealing the helical loop bearing T183 and N186. Image reproduced by permission from M. C. Surdel et al., (2017) Proc. Natl. Acad. Sci. U.S.A., published online July 24, DOI:10.1073/pnas.1700469114. Copyright 2017, M. C. Surdell, et al.



Exposure to '882 is not toxic to most wild-type bacteria. However, the investigators hypothesized that they could harness the CPIII accumulation that occurs in the presence of the compound to develop an antimicrobial strategy. Specifically, in the presence of light at 395 nm, CPIII undergoes photochemical reactions that produce toxic reactive oxygen species (Figure 3). Thus, such light exposure should suppress the growth of or kill '882-treated bacteria. The investigators tested this hypothesis in multiple Gram-positive species, including S. haemolyticus, S. lugdunensis, B. anthracis, P. acnes, and four strains of S. aureus. In all cases, '882 plus light was toxic to the bacteria. Including aminolevulinic acid, a precursor to CPIII, further increased the toxic effects.




Figure 3. Use of '882 in combination with photodynamic therapy as an antimicrobial agent. Exposure to '882 results in increased production of CPIII by CgoX. Interaction of light with CPIII leads to reactions that generate toxic reactive oxygen species (ROS) in the bacteria. Image reproduced by permission from M. C. Surdel et al., (2017) Proc. Natl. Acad. Sci. U.S.A., published online July 24, DOI:10.1073/pnas.1700469114. Copyright 2017, M. C. Surdell, et al.



To determine if '882-mediated photodynamic therapy might be effective as an antibacterial agent in vivo, the researchers tested it in two mouse models of skin infection, one using S. aureus and the other, P. acnes. In both cases, photodynamic therapy with '882 substantially reduced the bacterial burden at sites of infection, and the effects were augmented by inclusion of aminolevulinic acid (Figure 4).




FIGURE 4. (Left) Photomicrograph of skin from a mouse infected with S. aureus. The epidermis is heavily infiltrated with leukocytes, and colonies of bacteria (inset) can be observed. (Right) Photomicrograph of skin from a mouse infected with S. aureus but treated with '882, aminolevulinic acid, and photodynamic therapy. Leukocyge infiltration is markedly reduced, and obvious bacterial colonies are not observed. Image reproduced by permission from M. C. Surdel et al., (2017) Proc. Natl. Acad. Sci. U.S.A., published online July 24, DOI:10.1073/pnas.1700469114. Copyright 2017, M. C. Surdell, et al.



The investigators conclude that '882 is an important prototype for a new approach to the photodynamic therapy of skin and soft tissue infections caused by Gram-positive bacteria. The compound offers a distinct advantage over prior approaches that depend on increasing fluorescent heme biosynthetic intermediate concentrations by addition of aminolevulinic acid alone. As this also increases the intermediates in host cells, substantial toxicity can result. '882 has no effect on mammalian heme biosynthetic enzymes, so therapeutic specificity, and thereby safety can be substantially improved. Further work will be directed towards developing better ways to deliver light into deeper tissues and, perhaps, to further improve on the pharmacological properties of '882.




View PNAS article: Antibacterial photosensitization through activation of coproporphyrinogen oxidase.







The Vanderbilt Institute of Chemical Biology, 896 Preston Building, Nashville, TN 37232-6304, phone 866.303 VICB (8422), fax 615 936 3884
Vanderbilt University is committed to principles of equal opportunity and affirmative action. Copyright © 2014 by Vanderbilt University Medical Center