research papers\(\def\hfill{\hskip 5em}\def\hfil{\hskip 3em}\def\eqno#1{\hfil {#1}}\)

Journal logoSTRUCTURAL
BIOLOGY
ISSN: 2059-7983

A small step towards an important goal: fragment screen of the c-di-AMP-synthesizing enzyme CdaA

crossmark logo

aDepartment of Molecular Structural Biology, Institute of Microbiology and Genetics, GZMB, Georg-August-University Göttingen, Justus-von-Liebig-Weg 11, 37077 Göttingen, Germany, bMacromolecular Crystallography, Helmholtz-Zentrum Berlin, Albert-Einstein-Strasse 15, 12489 Berlin, Germany, cInstitut für Nanostruktur- und Festkörperphysik, Universität Hamburg, Luruper Chaussee 149, 22761 Hamburg, Germany, and dCluster of Excellence `Multiscale Bioimaging: From Molecular Machines to Networks of Excitable Cells' (MBExC), University of Göttingen, 37075 Göttingen, Germany
*Correspondence e-mail: pneuman2@uni-goettingen.de

Edited by A. Berghuis, McGill University, Canada (Received 22 February 2024; accepted 16 April 2024; online 29 April 2024)

CdaA is the most widespread diadenylate cyclase in many bacterial species, including several multidrug-resistant human pathogens. The enzymatic product of CdaA, cyclic di-AMP, is a secondary messenger that is essential for the viability of many bacteria. Its absence in humans makes CdaA a very promising and attractive target for the development of new antibiotics. Here, the structural results are presented of a crystallographic fragment screen against CdaA from Listeria monocytogenes, a saprophytic Gram-positive bacterium and an opportunistic food-borne pathogen that can cause listeriosis in humans and animals. Two of the eight fragment molecules reported here were localized in the highly conserved ATP-binding site. These fragments could serve as potential starting points for the development of antibiotics against several CdaA-dependent bacterial species.

1. Introduction

Multidrug-resistant bacteria constitute a major global health threat in modern healthcare (Cassini et al., 2019[Cassini, A., Högberg, L. D., Plachouras, D., Quattrocchi, A., Hoxha, A., Simonsen, G. S., Colomb-Cotinat, M., Kretzschmar, M. E., Devleesschauwer, B., Cecchini, M., Ouakrim, D. A., Oliveira, T. C., Struelens, M. J., Suetens, C., Monnet, D. L., Strauss, R., Mertens, K., Struyf, T., Catry, B., Latour, K., Ivanov, I. N., Dobreva, E. G., Tambic Andraševic, A., Soprek, S., Budimir, A., Paphitou, N., Žemlicková, H., Schytte Olsen, S., Wolff Sönksen, U., Märtin, P., Ivanova, M., Lyytikäinen, O., Jalava, J., Coignard, B., Eckmanns, T., Abu Sin, M., Haller, S., Daikos, G. L., Gikas, A., Tsiodras, S., Kontopidou, F., Tóth, Á., Hajdu, Á., Guðlaugsson, Ó., Kristinsson, K. G., Murchan, S., Burns, K., Pezzotti, P., Gagliotti, C., Dumpis, U., Liuimiene, A., Perrin, M., Borg, M. A., de Greeff, S. C., Monen, J. C., Koek, M. B., Elstrøm, P., Zabicka, D., Deptula, A., Hryniewicz, W., Caniça, M., Nogueira, P. J., Fernandes, P. A., Manageiro, V., Popescu, G. A., Serban, R. I., Schréterová, E., Litvová, S., Štefkovicová, M., Kolman, J., Klavs, I., Korošec, A., Aracil, B., Asensio, A., Pérez-Vázquez, M., Billström, H., Larsson, S., Reilly, J. S., Johnson, A. & Hopkins, S. (2019). Lancet Infect. Dis. 19, 56-66.]). Recent reports from the World Health Organization (WHO) indicate that antimicrobial resistance contributes to approximately 700 000 deaths globally each year, with concerns about a potential significant rise in fatalities without effective intervention (an estimated prediction of ten million deaths annually by 2050; Asokan et al., 2019[Asokan, G. V., Ramadhan, T., Ahmed, E. & Sanad, H. (2019). Oman Med. J. 34, 184-193.]). The misuse and overuse of antibiotics in recent decades have fueled the relentless evolution of bacterial pathogens, severely limiting our effective antimicrobial arsenal. Consequently, there is an urgent need for innovative antibacterial strategies to combat this escalating global health crisis. In this context, the bacterial cyclic di-adenosine monophosphate (c-di-AMP) signaling pathway, which is not present in humans (Rosenberg et al., 2015[Rosenberg, J., Dickmanns, A., Neumann, P., Gunka, K., Arens, J., Kaever, V., Stülke, J., Ficner, R. & Commichau, F. M. (2015). J. Biol. Chem. 290, 6596-6606.]), has emerged as a promising antibacterial target (Corrigan & Gründling, 2013[Corrigan, R. M. & Gründling, A. (2013). Nat. Rev. Microbiol. 11, 513-524.]; Römling, 2008[Römling, U. (2008). Sci. Signal. 1, pe39.]; Schuster et al., 2016[Schuster, C. F., Bellows, L. E., Tosi, T., Campeotto, I., Corrigan, R. M., Freemont, P. & Gründling, A. (2016). Sci. Signal. 9, ra81.]). Functioning as a secondary messenger, c-di-AMP intricately regulates essential bacterial cellular processes, including osmolyte and potassium ion homeostasis, biofilm formation, cell-wall integrity, DNA integrity scanning, modulation of metabolism, virulence and stress response (Gundlach et al., 2016[Gundlach, J., Rath, H., Herzberg, C., Mäder, U. & Stülke, J. (2016). Front. Microbiol. 7, https://doi.org/10.3389/fmicb.2016.00804.]; Mehne et al., 2013[Mehne, F. M. P., Gunka, K., Eilers, H., Herzberg, C., Kaever, V. & Stülke, J. (2013). J. Biol. Chem. 288, 2004-2017.]; Rørvik et al., 2020[Rørvik, G. H., Liskiewicz, K. A., Kryuchkov, F., Naemi, A., Aasheim, H., Petersen, F. C., Küntziger, T. M. & Simm, R. (2020). Microorganisms, 8, 1269.]; Whiteley et al., 2017[Whiteley, A. T., Garelis, N. E., Peterson, B. N., Choi, P. H., Tong, L., Woodward, J. J. & Portnoy, D. A. (2017). Mol. Microbiol. 104, 212-233.]). Diadenylate cyclases (DACs), which are c-di-AMP-synthesizing enzymes, contribute significantly to this intricate signaling network by catalysing the conversion of two adenosine triphosphate (ATP) molecules into one c-di-AMP molecule (Fig. 1[link]) in a metal-ion-dependent manner (Witte et al., 2008[Witte, G., Hartung, S., Büttner, K. & Hopfner, K. (2008). Mol. Cell, 30, 167-178.]). Hence, DACs are essential in the bacterial cell. Remarkably, they exhibit extensive structural conservation across diverse species (mainly Gram-positive bacteria), highlighting the profound importance of DACs and rendering them attractive candidates for drug design.

[Figure 1]
Figure 1
Crystal structure of L. monocytogenes Δ100CdaA. (a) Biochemical reaction of DAC. (b) Overall structure of the truncated monomeric Δ100CdaA. The c-di-AMP molecule, depicted in ball-and-stick mode, marks the active site. Color coding is according to conservation score: green, low; white, medium; purple, high. (c) Cartoon representation of the catalytically active Δ100CdaA dimer in its postcatalytic state with c-di-AMP bound in the active site (PDB entry 6hvl).

Five different classes of DACs are known to date: CdaA, DisA, CdaS, CdaM and CdaZ (Blötz et al., 2017[Blötz, C., Treffon, K., Kaever, V., Schwede, F., Hammer, E. & Stülke, J. (2017). Front. Microbiol. 8, 1328.]; Commichau et al., 2019[Commichau, F. M., Heidemann, J. L., Ficner, R. & Stülke, J. (2019). J. Bacteriol. 201, 1-14.]; Corrigan & Gründling, 2013[Corrigan, R. M. & Gründling, A. (2013). Nat. Rev. Microbiol. 11, 513-524.]; Römling, 2008[Römling, U. (2008). Sci. Signal. 1, pe39.]), of which the first three have been structurally characterized. They all share the highly conserved DAC domain (Fig. 1[link]) accompanied by regulatory domains (Commichau et al., 2019[Commichau, F. M., Heidemann, J. L., Ficner, R. & Stülke, J. (2019). J. Bacteriol. 201, 1-14.]). The DAC domain of CdaA reveals an overall globular fold, with a central seven-stranded mixed β-sheet flanked by five α-helices. The well conserved active site is located between α-helix 4, β-strand 1 and β-strand 5 as well as several loops connecting secondary-structure elements (Fig. 1[link]). The synthesis of c-di-AMP requires the transient dimerization of two ATP-loaded CdaA molecules in a face-to-face orientation, as seen in the hexagonal CdaA–c-di-AMP complex structure in the postcatalytic state (PDB entry 6hvl; Fig. 1[link]; Heidemann et al., 2019[Heidemann, J. L., Neumann, P., Dickmanns, A. & Ficner, R. (2019). J. Biol. Chem. 294, 10463-10470.]; Rosenberg et al., 2015[Rosenberg, J., Dickmanns, A., Neumann, P., Gunka, K., Arens, J., Kaever, V., Stülke, J., Ficner, R. & Commichau, F. M. (2015). J. Biol. Chem. 290, 6596-6606.]; Witte et al., 2008[Witte, G., Hartung, S., Büttner, K. & Hopfner, K. (2008). Mol. Cell, 30, 167-178.]). Most bacteria that are known to synthesize c-di-AMP possess only one class of DAC: either DisA or CdaA. The latter has been described to be the most prevalent DAC domain-containing protein in several bacterial species (Commichau et al., 2019[Commichau, F. M., Heidemann, J. L., Ficner, R. & Stülke, J. (2019). J. Bacteriol. 201, 1-14.]), including human pathogens such as Listeria monocytogenes, Borrelia turicatea, Enterococcus faecalis, Staphylococcus aureus, Mycobacterium tuberculosis and Streptococcus pneumoniae (Bai et al., 2012[Bai, Y., Yang, J., Zhou, X., Ding, X., Eisele, L. E. & Bai, G. (2012). PLoS One, 7, e35206.]; Corrigan et al., 2011[Corrigan, R. M., Abbott, J. C., Burhenne, H., Kaever, V. & Gründling, A. (2011). PLoS Pathog. 7, e1002217.]; Jackson-Litteken et al., 2021[Jackson-Litteken, C. D., Ratliff, C. T., Kneubehl, A. R., Siletti, C., Pack, L., Lan, R., Huynh, T. N., Lopez, J. E. & Blevins, J. S. (2021). Infect. Immun. 89, e00787-20.]; Woodward et al., 2010[Woodward, J. J., Iavarone, A. T. & Portnoy, D. A. (2010). Science, 328, 1703-1705.]; Zarrella et al., 2020[Zarrella, T. M., Yang, J., Metzger, D. W. & Bai, G. (2020). J. Bacteriol. 202, https://doi.org/10.1128/jb.00691-19.]). They possess CdaA as the sole DAC and the last three mentioned belong to the 12 bacterial species that pose the greatest threat to human health according to the WHO (Asokan et al., 2019[Asokan, G. V., Ramadhan, T., Ahmed, E. & Sanad, H. (2019). Oman Med. J. 34, 184-193.]). This renders CdaA a particularly intriguing target for broad-spectrum therapeutics, especially against multidrug-resistant bacteria, as DACs are absent in humans (Rosenberg et al., 2015[Rosenberg, J., Dickmanns, A., Neumann, P., Gunka, K., Arens, J., Kaever, V., Stülke, J., Ficner, R. & Commichau, F. M. (2015). J. Biol. Chem. 290, 6596-6606.]). The recently reported biochemical and structural details of L. monocytogenes (Lm) CdaA provide a solid foundation for the development of a potent CdaA inhibitor using fragment-based drug discovery (FBDD; Erlanson et al., 2016[Erlanson, D. A., Fesik, S. W., Hubbard, R. E., Jahnke, W. & Jhoti, H. (2016). Nat. Rev. Drug Discov. 15, 605-619.]; Wollenhaupt et al., 2021[Wollenhaupt, J., Barthel, T., Lima, G. M. A., Metz, A., Wallacher, D., Jagudin, E., Huschmann, F. U., Hauss, T., Feiler, C. G., Gerlach, M., Hellmig, M., Förster, R., Steffien, M., Heine, A., Klebe, G., Mueller, U. & Weiss, M. S. (2021). J. Vis. Exp., e62208.]), a modern technique that is playing an increasingly important role in the development of new clinical candidates. This method uses small-molecule probes or fragments, which generally possess weak binding affinities to their target (Giordanetto et al., 2019[Giordanetto, F., Jin, C., Willmore, L., Feher, M. & Shaw, D. E. (2019). J. Med. Chem. 62, 3381-3394.]), that allow a thorough search of the available chemical space and provide starting points for drug development. In combination with X-ray crystallography, this enables structure-based drug discovery from the outset and can lead to the development of superior molecules.

Here, we present details of the crystallographic fragment-screening campaign for the bacterial target Lm CdaA utilizing the previously commercially available Frag Xtal Screen (Huschmann et al., 2016[Huschmann, F. U., Linnik, J., Sparta, K., Ühlein, M., Wang, X., Metz, A., Schiebel, J., Heine, A., Klebe, G., Weiss, M. S. & Mueller, U. (2016). Acta Cryst. F72, 346-355.]; Jena Bioscience, MiTeGen) and eight fragment crystal structures derived from it. The target protein crystallized in the ortho­rhombic system with a noncatalytic `back-to-back' homodimer in the asymmetric unit. The two active sites (one per monomer) are exposed to solvent, which makes them accessible to small-molecule ligands and favors soaking experiments. Two of the identified fragments that have been localized in the ATP-binding site of CdaA could serve as potential starting points for antibiotic drug discovery and for `growing' the computationally designed lead candidate comprising two substituted thiazole rings (compound 7; Neumann et al., 2023[Neumann, P., Kloskowski, P. & Ficner, R. (2023). microLife, 4, uqad021.]). This in silico-designed compound showed inhibitory potential based on ITC measurements and is the first structurally characterized CdaA inhibitor to date.

2. Experimental procedures

2.1. Bacterial strains and growth conditions

Escherichia coli BL21 (DE3) cells, which were used for overexpression of the protein, were cultivated in 2×YT medium comprising 1.6%(w/v) tryptone, 1.0%(w/v) yeast extract and 0.5%(w/v) NaCl. The transformed cells were selected on lysogeny broth medium plates containing ampicillin (100 µg ml−1).

2.2. Protein expression and purification

Protein expression and purification were performed as described previously (Heidemann et al., 2019[Heidemann, J. L., Neumann, P., Dickmanns, A. & Ficner, R. (2019). J. Biol. Chem. 294, 10463-10470.]; Neumann et al., 2023[Neumann, P., Kloskowski, P. & Ficner, R. (2023). microLife, 4, uqad021.]). Briefly, the plasmid pGEXpBP33, originating from the pGEX-6P-1 (Cytiva) vector and encoding a truncated Lm CdaA protein (Δ100CdaA) with an N-terminal GST tag, was transformed into E. coli BL21 (DE3) cells. The resulting cell cultures were grown in 2×YT medium at 37°C. Protein expression was induced at an OD600 of ∼0.6 by the addition of 1 mM isopropyl β-D-1-thiogalactopyranoside and the cultures were incubated overnight at 16°C. The harvested cells were disrupted with a microfluidizer (M-110S Microfluidizer, Microfluidics) in 20 mM Tris–HCl pH 7.5, 10 mM EDTA, 1 M NaCl and then centrifuged for 30 min at 4°C. The retained lysate containing the GST-tagged target protein was loaded onto a Glutathione Sepharose column (Cytiva) and eluted with 40 mM reduced GSH. The tag was proteolytically cleaved with 1:100(w:w) PreScission protease overnight at 4°C in cellulose tubing placed in dialysis buffer (300 mM NaCl, 20 mM Tris–HCl pH 7.5). The remaining impurities and the tag were removed using a Superdex 75 column (Cytiva) coupled to a Glutathione Sepharose column in 20 mM Tris–HCl pH 7.5, 300 mM NaCl. The protein was concentrated to 20 mg ml−1.

2.3. Crystallization

The sitting-drop vapor-diffusion method was applied for crystallization. The previously reported crystallization condition for apo CdaA (Heidemann et al., 2019[Heidemann, J. L., Neumann, P., Dickmanns, A. & Ficner, R. (2019). J. Biol. Chem. 294, 10463-10470.]) was optimized in order to yield crystals that were more suited for fragment screening, i.e. had better diffraction properties (Neumann et al., 2023[Neumann, P., Kloskowski, P. & Ficner, R. (2023). microLife, 4, uqad021.]). Crystallization trials were performed at 20°C using a concentration of 6 mg ml−1 Δ100CdaA in 2 µl droplets and a 1:1 protein:reservoir ratio. In order to facilitate crystal growth, microseeding was performed. Thin crystal plates were obtained overnight in 3.7 M NaCl, 0.1 M Na HEPES pH 8.5, 3% DMSO. The ideal DMSO concentration had been determined by previous stability tests.

2.4. Fragment soaking, data collection and structure determination

The Frag Xtal Screen fragment-screening library (Jena Bioscience, MiTeGen), comprising 96 fragments, each present in a dried form in two lenses of the respective well of a 96-well crystallization plate, was used. Solubilization of the dried fragments was achieved by pipetting 0.5 µl of the crystallization reservoir into one of the two lenses, resulting in a nominal fragment concentration of 100 mM. The second lens was supplemented with cryoprotecting buffer (crystallization reservoir saturated with sucrose; Heidemann et al., 2019[Heidemann, J. L., Neumann, P., Dickmanns, A. & Ficner, R. (2019). J. Biol. Chem. 294, 10463-10470.]). The crystallization plates were sealed and left overnight at 20°C. Subsequently, two to four apo CdaA crystals were transferred to each well with the solubilized fragment. All crystals were soaked overnight, captured in SPINE-standard cryo-loops and cryoprotected in the second drop prior to plunging them into liquid nitrogen.

Diffraction images were collected on EMBL beamline P13 at PETRA III, DESY, Hamburg, Germany and the MASSIF-3 beamline at ESRF, Grenoble, France. All diffraction images were processed using a custom script (Neumann & Tittmann, 2014[Neumann, P. & Tittmann, K. (2014). Curr. Opin. Struct. Biol. 29, 122-133.]) utilizing the XDS package (Kabsch, 2010[Kabsch, W. (2010). Acta Cryst. D66, 125-132.]). Structure solution was performed using DIMPLE (Wojdyr et al., 2013[Wojdyr, M., Keegan, R., Winter, G. & Ashton, A. (2013). Acta Cryst. A69, s299.]) employing programs from the CCP4 suite (Murshudov et al., 2011[Murshudov, G. N., Skubák, P., Lebedev, A. A., Pannu, N. S., Steiner, R. A., Nicholls, R. A., Winn, M. D., Long, F. & Vagin, A. A. (2011). Acta Cryst. D67, 355-367.]; Agirre et al., 2023[Agirre, J., Atanasova, M., Bagdonas, H., Ballard, C. B., Baslé, A., Beilsten-Edmands, J., Borges, R. J., Brown, D. G., Burgos-Mármol, J. J., Berrisford, J. M., Bond, P. S., Caballero, I., Catapano, L., Chojnowski, G., Cook, A. G., Cowtan, K. D., Croll, T. I., Debreczeni, J. É., Devenish, N. E., Dodson, E. J., Drevon, T. R., Emsley, P., Evans, G., Evans, P. R., Fando, M., Foadi, J., Fuentes-Montero, L., Garman, E. F., Gerstel, M., Gildea, R. J., Hatti, K., Hekkelman, M. L., Heuser, P., Hoh, S. W., Hough, M. A., Jenkins, H. T., Jiménez, E., Joosten, R. P., Keegan, R. M., Keep, N., Krissinel, E. B., Kolenko, P., Kovalevskiy, O., Lamzin, V. S., Lawson, D. M., Lebedev, A. A., Leslie, A. G. W., Lohkamp, B., Long, F., Malý, M., McCoy, A. J., McNicholas, S. J., Medina, A., Millán, C., Murray, J. W., Murshudov, G. N., Nicholls, R. A., Noble, M. E. M., Oeffner, R., Pannu, N. S., Parkhurst, J. M., Pearce, N., Pereira, J., Perrakis, A., Powell, H. R., Read, R. J., Rigden, D. J., Rochira, W., Sammito, M., Sánchez Rodríguez, F., Sheldrick, G. M., Shelley, K. L., Simkovic, F., Simpkin, A. J., Skubak, P., Sobolev, E., Steiner, R. A., Stevenson, K., Tews, I., Thomas, J. M. H., Thorn, A., Valls, J. T., Uski, V., Usón, I., Vagin, A., Velankar, S., Vollmar, M., Walden, H., Waterman, D., Wilson, K. S., Winn, M. D., Winter, G., Wojdyr, M. & Yamashita, K. (2023). Acta Cryst. D79, 449-461.]). The resulting atomic models were refined using a customized self-written refinement pipeline making use of the Phenix package (phenix.refine and phenix.real_space_refine; Afonine et al., 2018[Afonine, P. V., Poon, B. K., Read, R. J., Sobolev, O. V., Terwilliger, T. C., Urzhumtsev, A. & Adams, P. D. (2018). Acta Cryst. D74, 531-544.]; Liebschner et al., 2017[Liebschner, D., Afonine, P. V., Moriarty, N. W., Poon, B. K., Sobolev, O. V., Terwilliger, T. C. & Adams, P. D. (2017). Acta Cryst. D73, 148-157.], 2019[Liebschner, D., Afonine, P. V., Baker, M. L., Bunkóczi, G., Chen, V. B., Croll, T. I., Hintze, B., Hung, L.-W., Jain, S., McCoy, A. J., Moriarty, N. W., Oeffner, R. D., Poon, B. K., Prisant, M. G., Read, R. J., Richardson, J. S., Richardson, D. C., Sammito, M. D., Sobolev, O. V., Stockwell, D. H., Terwilliger, T. C., Urzhumtsev, A. G., Videau, L. L., Williams, C. J. & Adams, P. D. (2019). Acta Cryst. D75, 861-877.]). The refined structural models were subjected to PanDDA (Pearce et al., 2017[Pearce, N. M., Krojer, T. & von Delft, F. (2017). Acta Cryst. D73, 256-266.]) to facilitate the identification of bound fragment molecules. Selected structures with well defined fragment molecules were manually inspected in Coot (Emsley et al., 2010[Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. (2010). Acta Cryst. D66, 486-501.]) and refined in Phenix using the default restrained refinement strategy (isotropic ADP refinement for all non-H atoms, automated TLS group assignment for protein molecules, optimization of the bulk-solvent mask and weights for stereochemical and ADP restraints). Ligand libraries were generated with phenix.elbow. Omit electron-density maps were calculated using phenix.polder (Liebschner et al., 2017[Liebschner, D., Afonine, P. V., Moriarty, N. W., Poon, B. K., Sobolev, O. V., Terwilliger, T. C. & Adams, P. D. (2017). Acta Cryst. D73, 148-157.]; Fig. 2[link]). The identified small-molecule fragments are listed in Table 1[link]. Data-collection, processing and refinement statistics are summarized in Tables 2[link] and 3[link].

Table 1
Fragment molecules

Plate ID Fragment SciFinder name CAS ID MW (g mol−1)
B04
[Scheme 1]
Acetamide, 2-[(cyanomethyl)methylamino]-N-(6-methyl-2-pyridinyl)- 1311649-76-9 218.25
B06
[Scheme 2]
2-(4-Fluorophenyl)acetohydrazide 34547-28-9 168.17
C08
[Scheme 3]
Methyl 4-(aminomethyl)benzoate hydrochloride (1:1) 6232-11-7 201.65
C11
[Scheme 4]
1H-Indole-3-ethanamine, N-[(1-methyl-1H-pyrrol-2-yl)methyl]- 289487-79-2 253.34
D07
[Scheme 5]
1H-Pyrazole-4-acetamide, 1,3,5-trimethyl-N-2-pyridinyl- 1171575-61-3 244.29
E01
[Scheme 6]
(2-Chlorophenyl)methyl carbamimidothioate 14122-38-4 200.69
E05
[Scheme 7]
(1S,2S,3S,4R,5R)-2-Amino-4-phenylsulfanyl-6,8-dioxabicyclo[3.2.1]octan-3-ol 1212574-75-8 253.32
H04
[Scheme 8]
Butanedioic acid, 1-(2,2-dimethylhydrazide) 1596-84-5 160.17

Table 2
Data-collection statistics

Values in parentheses are for the highest resolution shell.

Structure D07 E01 E05 B06 B04 H04 C11 C08
PDB code 8s47 8s46 8s45 8s49 8s4a 8s4c 8s48 8s4b
Beamline P13, PETRA III MASSIF-3, ESRF P13, PETRA III MASSIF-3, ESRF MASSIF-3, ESRF MASSIF-3, ESRF P13, PETRA III P13, PETRA III
Wavelength (Å) 0.97625 0.96770 0.97625 0.96770 0.96770 0.96770 0.97625 0.97625
Space group P212121 P212121 P212121 P212121 P212121 P212121 P212121 P212121
a, b, c (Å) 44.22, 65.07, 130.33 39.79, 64.65, 130.00 41.75, 64.71, 129.04 41.66, 64.75, 129.06 41.70, 64.46, 128.69 43.74, 64.11, 129.79 41.33, 64.68, 128.65 41.49, 64.72, 128.39
α, β, γ (°) 90, 90, 90 90, 90, 90 90, 90, 90 90, 90, 90 90, 90, 90 90, 90, 90 90, 90, 90 90, 90, 90
Resolution range (Å) 46.04–2.11 (2.16–2.11) 45.84–1.99 (2.03–1.99) 39.72–1.69 (1.71–1.69) 45.71–1.70 (1.72–1.70) 45.54–2.23 (2.30–2.23) 45.61–1.80 (1.83–1.80) 35.74–1.65 (1.67–1.65) 45.58–2.15 (2.25–2.15)
Total reflections 295034 194674 330217 311980 123142 288916 337480 149891
Unique reflections 41687 (2736) 44268 (2726) 74809 (2708) 73647 (2829) 31404 (2885) 65236 (2875) 79766 (2847) 36215 (4603)
Multiplicity 7.1 4.4 4.4 4.2 3.9 4.4 4.2 4.1
Completeness (%) 99.76 (99.45) 99.64 (99.67) 98.91 (97.10) 99.24 (98.64) 96.06 (95.72) 99.81 (99.62) 99.51 (98.44) 99.80 (99.08)
Mean I/σ(I) 10.08 (0.92) 11.58 (0.71) 13.10 (0.80) 10.06 (0.98) 8.04 (0.98) 13.12 (0.85) 10.42 (0.80) 5.68 (1.26)
Wilson B factor (Å2) 47.56 42.80 31.64 34.57 39.00 37.16 31.89 46.01
Rmerge (%) 10.8 (265.1) 7.6 (258.9) 5.0 (205.5) 5.5 (204.6) 12.9 (148.6) 5.0 (173.9) 5.5 (180.8) 14.9 (136.5)
Rmeas (%) 11.7 (286.1) 8.6 (293.3) 5.7 (235.0) 6.4 (232.9) 14.9 (174.1) 5.7 (197.1) 6.3 (206.3) 17.1 (157.3)
CC1/2 99.9 (50.0) 99.9 (40.1) 99.9 (52.1) 99.8 (53.6) 99.6 (37.0) 99.8 (50.4) 99.9 (55.8) 99.3 (62.0)
Rmeas is the redundancy-independent multiplicity-weighted R factor for comparing symmetry-related reflections (Diederichs & Karplus, 1997[Diederichs, K. & Karplus, P. A. (1997). Nat. Struct. Mol. Biol. 4, 269-275.]).

Table 3
Crystallographic refinement statistics

Values in parentheses are for the highest resolution shell.

Structure D07 E01 E05 B06 B04 H04 C11 C08
PDB code 8s47 8s46 8s45 8s49 8s4a 8s4c 8s48 8s4b
Resolution range (Å) 46.04–2.11 45.84–1.99 39.72–1.69 45.71–1.70 45.54–2.23 45.61–1.80 35.74–1.65 45.58–2.15
Reflections used, refinement 41687 (2736) 44268 (2726) 74809 (2708) 73647 (2829) 31404 (2885) 65236 (2875) 79766 (2847) 36122 (2755)
Reflections used for Rfree 2083 (139) 2219 (146) 3748 (137) 3698 (142) 1572 (142) 3278 (147) 3989 (145) 1801 (126)
Rwork 0.2059 (0.4197) 0.2247 (0.4747) 0.1964 (0.5552) 0.1974 (0.5124) 0.2008 (0.3377) 0.2015 (0.4860) 0.1908 (0.5033) 0.2400 (0.4334)
Rfree 0.2402 (0.5195) 0.2587 (0.4941) 0.2357 (0.6358) 0.2207 (0.4554) 0.2436 (0.3930) 0.2268 (0.5125) 0.2187 (0.5334) 0.2774 (0.4832)
No. of non-H atoms
 Total 2495 2418 2604 2532 2485 2576 2674 2415
 Macromolecules 2366 2344 2416 2383 2332 2414 2428 2326
 Ligands 38 13 17 14 33 11 19 12
 Solvent 91 61 171 135 120 151 227 77
Protein residues 308 305 310 308 304 314 311 302
R.m.s.d., bond lengths (Å) 0.002 0.002 0.010 0.009 0.003 0.004 0.008 0.002
R.m.s.d., angles (°) 0.42 0.49 0.93 0.97 0.54 0.70 0.83 0.42
Ramachandran favored (%) 98.68 99.34 99.35 98.68 97.67 99.35 98.37 96.64
Ramachandran allowed (%) 1.32 0.66 0.65 1.32 2.33 0.65 1.63 3.36
Ramachandran outliers (%) 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Rotamer outliers (%) 2.65 1.14 1.11 1.12 0.00 0.37 1.84 2.30
Clashscore 1.65 2.52 2.45 3.10 2.31 1.22 2.84 1.06
Average B factor (Å2)
 Overall 62.68 60.48 43.22 47.51 50.42 54.82 44.81 53.12
 Macromolecules 62.40 60.57 43.07 47.22 50.21 54.48 44.45 53.31
 Ligands 77.62 65.78 41.50 53.99 64.16 80.38 48.22 45.93
 Solvent 63.74 56.01 45.50 51.97 50.88 58.50 48.36 48.49
[Figure 2]
Figure 2
Detailed views of the binding sites of the eight identified fragments. Protein residues within 4 Å are depicted as sticks; hydrogen bonds are marked as yellow dashed lines. The polder omit map (mFoDFc) is shown at either the 3.3σ or the 3.9σ level. (a) D07, (b) B04, (c) C08, (d) B06, (e) C11, (f) E05, (g) E01, (h) H04. For fragments bound in the active sites (a) and (b), an ATP molecule is shown for reference (depicted as a green stick model, PDB entry 8c4o).

2.5. Binding-site assignment of fragment molecules located between symmetry mates

The binding sites of five small-molecule fragments from the Frag Xtal Screen (B06, C08, C11, E01 and E05) were clearly localized between a CdaA molecule occupying the asymmetric unit (chain A) and a symmetry mate of the second CdaA molecule: chain Bsym (Fig. 3[link]). Initially, the inspected fragment molecules were modeled distant from the ATP-binding site and were therefore assigned to the CdaA molecule labeled as chain B (model 1, Fig. 3[link]). The refined atomic models of these five CdaA complexes (model 1) were slightly manipulated to generate alternative structural models from the fragment perspective. In these generated models, the fragment molecules were placed in closer proximity to the ATP active site of the CdaA molecule labeled chain A (model 2), as shown in Fig. 3[link]. Repositioning of the fragment molecules was performed in Coot using a selected symmetry mate as a reference. Subsequently, the two alternative structural models (models 1 and 2) were decomposed into two components: ligand (fragment) and receptor (protein). The receptor, which is identical for both models, was converted into PDBQT format using the MGL tools (Sanner, 1999[Sanner, M. (1999). J. Mol. Graph. Model. 17, 57-61.]). For each receptor–ligand pair, the ligand pose (the position of the fragment within the putative binding site) was reassessed using the molecular-docking program Gnina (McNutt et al., 2021[McNutt, A. T., Francoeur, P., Aggarwal, R., Masuda, T., Meli, R., Ragoza, M., Sunseri, J. & Koes, D. R. (2021). J. Cheminform, 13, 43.]) utilizing an ensemble of convolutional neural networks (CNNs) as a scoring function. In addition, three frequently used scoring functions, AutoDock 4 (AD4), Vina and Vinardo, which provide a computationally predicted binding energy (affinity), were also tested. For this purpose, we used both Vina (version 1.2.5; Eberhardt et al., 2021[Eberhardt, J., Santos-Martins, D., Tillack, A. F. & Forli, S. (2021). J. Chem. Inf. Model. 61, 3891-3898.]; Trott & Olson, 2010[Trott, O. & Olson, A. J. (2010). J. Comput. Chem. 31, 455-461.]) and the Gnina program, the latter without CNN rescoring, thus simulating the Vina run. It should be noted that the commonly used scoring functions are likely to give non-identical results in terms of different predicted free energies of binding (predicted binding affinities) due to the way that they were developed: Vinardo is an empirical (regression-based) scoring function, AutoDock 4 is a physics-based scoring function and Vina is a hybrid scoring function. The final assignment of the CdaA molecule, with the estimated dominant contribution to the binding event, was based on the difference of the calculated CNN score and the lower predicted binding affinity, providing a comprehensive and robust basis for the selection of the most appropriate fragment-binding site. The proposed binding-site evaluation is computationally inexpensive (less then 5 s per fragment) and in addition it can easily be performed and scripted, as Gnina and Vina do not require any dictionaries or additional parameters for the fragment molecules.

[Figure 3]
Figure 3
Binding-site assignment of fragment molecules located between symmetry-related molecules. (a) The asymmetric unit of the ortho­rhombic CdaA crystal (blue semi-transparent surface) with two neighboring symmetry-related dimers (cartoon representation, white semi-transparent surface). A close-up view shows a detailed representation of the fragment interactions (C08). (b) A 180° rotated view of (a). The alternative assignments of the C08 fragment molecule are depicted as model 1 (distant from the active site) and model 2 (closer proximity to the ATP-binding site).

3. Results

3.1. Identification of bound fragments

L. monocytogenes (Lm) CdaA is a membrane-bound protein consisting of an N-terminal transmembrane domain (residues 1–80), a 20-amino-acid flexible linker and a DAC domain (residues 101–255, abbreviated as Δ100CdaA; the numbering corresponds to that of L. monocytogenes). In the orthorhombic CdaA crystals, the asymmetric unit consists of two Δ100CdaA monomers forming a noncatalytic dimer with two outward-facing and solvent-accessible active sites. Therefore, it is expected that certain fragments can bind to both monomers. Using the Frag Xtal Screen, a 96-fragment library, fragment screening was performed and resulted in the collection of approximately 200 data sets (two to three crystals per fragment, together with some apo crystals). Remarkably, about 15% of the soaked crystals did not diffract or dissolved during the soaking experiments, while approximately 69% of all measured crystals diffracted to between 1.7 and 2.5 Å resolution. The macromolecular crystallography pipeline for refinement and ligand screening (DIMPLE) suggested only one structure with an unmodeled electron-density map (bound fragment molecule) as reported by the `Find Blob' function using the default settings. Therefore, further analysis of all structures was performed with PanDDA. Manual inspection of the DIMPLE-based electron-density maps of the PanDDA hits revealed that all of them were placed in prominent blobs of the mFoDFc electron-density map at the +3σ level that went undetected by DIMPLE. A total of eight fragments (Table 1[link] and Fig. 4[link]) were identified. The polder omit maps for all fragments are shown in Fig. 2[link]. Five fragments (B06, C08, C11, E01 and E05) were bound between one CdaA molecule (chain A) occupying the asymmetric unit and the symmetric counterpart of the second molecule (chain Bsym; Fig. 3[link]). This led to ambiguity in assignment of the fragment-binding site to a single protein molecule (chain) within the crystal lattice. To resolve this ambiguity, the choice of binding-site assignment to a particular CdaA molecule was aided by a molecular docking-based approach using an ensemble of convolutional neural networks (CNNs) as well as predicted binding affinities to assess which protein molecule makes the largest contribution to the binding event. The binding sites of the remaining three fragments (C11, D07 and H04) could be unambiguously assigned, as the first two were located in the active site of CdaA and the last one was located on the solvent-exposed surface of a single CdaA molecule.

[Figure 4]
Figure 4
The eight fragment molecules (identified binding events) mapped on the surface of the CdaA monomer and colored according to the surface-conservation score. The individual views (90° increments) represent different views of the CdaA monomer. Two fragment molecules (B04 and D07) were localized in the highly conserved region (active site, marked with the AMP molecule colored yellow), while the remaining six fragments (B06, C08, C11, E01, E05 and H04) were identified in the nonconserved regions.

3.2. Molecular docking-based assessment of the binding-site localization

Molecular docking is a computational procedure that predicts the conformation of a small molecule (ligand) binding noncovalently to a receptor (protein). This prediction outputs both the conformation and the evaluated fitness of the pose, usually in the form of the binding affinity of the small molecule in its predicted minimum-energy state. These features have enabled the widespread use of this method for the virtual screening of large libraries of compounds. Recently, a new docking program, Gnina (McNutt et al., 2021[McNutt, A. T., Francoeur, P., Aggarwal, R., Masuda, T., Meli, R., Ragoza, M., Sunseri, J. & Koes, D. R. (2021). J. Cheminform, 13, 43.]), that uses a CNN scoring function to re-evaluate the output poses has been released and made available to the scientific community. It outperforms other docking programs on redocking and cross-docking tasks when the binding pocket is defined (McNutt et al., 2021[McNutt, A. T., Francoeur, P., Aggarwal, R., Masuda, T., Meli, R., Ragoza, M., Sunseri, J. & Koes, D. R. (2021). J. Cheminform, 13, 43.]). This inspired us to use Gnina to assess the fitness of the crystallographically refined positions of five fragment molecules (B06, C08, C11, E01 and E05) identified based on prominent spots of the mFoDFc electron-density map at the +3σ level. These fragments were located between two symmetry-related CdaA (receptor) molecules within the orthorhombic crystal lattice. Consequently, these fragment molecules can be formally assigned to two alternative positions on the CdaA surface, as shown in Fig. 3[link]: close to the ATP active site (potentially significant for drug development; model 2) or far away from it (probably of lesser importance; model 1). Our primary objective was to evaluate and compare the fitness of these two alternative assignments (model 1 versus model 2) by simulating the idealized scenario in which the ligand is bound exclusively by one of the two CdaA molecules (no influence of the crystal lattice). We calculated the convolutional neural network (CNN) scores with Gnina and predicted binding affinities (Gnina and Vina; AD4, Vina and Vinardo scoring functions) for both ligand positions (model 1 and model 2) using the same receptor molecule (two CdaA monomers occupying the asymmetric unit), and compared the results (Fig. 5[link]). The results obtained clearly show that in the case of four ligands (B06, C08, E01 and E05) our original assignment (model 1) was correct considering only the CNN scoring. In the case of fragment C11, the difference in the CNN score of 0.026 indicates a slightly larger contribution of the binding event originating from the region close to the ATP-binding site. These results are in very good agreement with the assessed differences in predicted binding affinity calculated using three scoring functions with the Gnina and Vina programs (Fig. 5[link]). The affinity-based assessment clearly shows that the initial binding-site assignment (model 1) was correct for all five ligands studied, regardless of the program and scoring function used. Remarkably, both tested programs provided different calculated affinities when the Vinardo and AD4 scoring functions were employed. This is most likely due to different default values of weights applied to certain energetic terms, for example repulsion, hydrogen bonding, hydrophobic interactions and van der Waals, that are used by both tested programs. Despite the formal assignment of the fragment molecules to a specific CdaA molecule (chain A or B), the observed ΔCNN scores and computed binding affinities (Fig. 5[link]) indicate a non-negligible contribution of both involved CdaA molecules to the binding of inspected fragments. This could imply that the probability of observing these binding events in other crystal forms or in solution might be relatively low.

[Figure 5]
Figure 5
Molecular docking-based assessment of binding-site localization for five fragments located at the interface between symmetry-related CdaA molecules. (a) Differences between the Gnina-calculated CNN scores assessed for model 2 and model 1. The negative difference indicates that the calculated CNN score for model 1 is higher than that for model 2. (b) Differences in predicted binding affinity. Rescoring of fragment positions was calculated with the Gnina (G*) and Vina (V*) programs using three scoring functions: AD4, Vinardo and Vina. A positive difference indicates that the estimated free energy of binding (binding affinity) for model 1 is lower (more negative) than that for model 2 (less negative).

3.3. Prospects for fragment-based drug development for Lm CdaA

Using an assessment approach based on molecular docking, we aimed to achieve the unbiased assignment of a single protein molecule within the crystal lattice that predominantly contributes to the binding event. This was a prerequisite for evaluating the usability of the identified fragment molecules for the future development of a lead compound that will potentially inhibit several of the bacterial species that pose a major global health threat: in particular, those species that possess CdaA as the only DAC and constitute the group of 12 bacteria that pose the greatest threat to human health according to the WHO. Therefore, we mapped the eight identified binding events on the surface of the CdaA monomer colored according to the surface-conservation score (Fig. 4[link]). Six of the eight localized fragments (B06, C08, C11, E01, E05 and H04) were identified in nonconserved regions on the CdaA surface. Moreover, the first five of these were located at nonbiologically relevant interfaces that are formed between two CdaA monomers and result from crystal packing. This unfavored positioning makes them unsuitable for the design of a general lead compound that should target several bacterial species that possess CdaA as the only enzyme synthesizing c-di-AMP. In contrast, only two fragment molecules (B04 and D07) localized in the highly sequence-conserved region (active site) are promising candidates for drug design and for further development (`growing') of the recently published CdaA inhibitor (Neumann et al., 2023[Neumann, P., Kloskowski, P. & Ficner, R. (2023). microLife, 4, uqad021.]). These two fragments (B04 and D07) share the N-(2-pyridyl) moiety (Table 1[link]), possessing the structural pattern —N=C—NH2, which has two adjacent N atoms that are structurally similar to N1 and N6 of the adenine ring (Figs. 2[link] and 6[link]). This chemical and structural similarity leads to the formation of two hydrogen bonds similar to those observed for the adenine moiety (adenine N1–Leu188 N and adenine N6–Leu188 O). These two conserved hydrogen bonds have been observed in several CdaA complex structures (ADP, ATP and c-di-AMP), regardless of the postcatalytic or precatalytic state, the well described Tyr187 backrub motion and its ππ stacking interactions with aromatic rings of bound ligands (Fig. 6[link]; Neumann et al., 2023[Neumann, P., Kloskowski, P. & Ficner, R. (2023). microLife, 4, uqad021.]). Superposition of ATP-bound CdaA (PDB entry 8c4o) with the structure containing fragment D07 (Figs. 2[link] and 6[link]) clearly shows the structural and positional equivalence of the —N=C—NH2 pattern. Its importance and its relevance for binding is further confirmed by fragment B04 (Figs. 2[link] and 6[link]), which shows that even the presence of a hindering methyl group at position 6 [the N-(6-methyl-2-pyridyl) moiety] cannot prevent the formation of these two conserved hydrogen bonds. Remarkably, the aforementioned structural pattern was also used as a strong donor and acceptor constraint for the development of the bi-thiazole inhibitor (compound 7; Fig. 6[link]), which binds to CdaA in the micromolar range and with a Kd ∼8-fold lower than that of the natural substrate ATP (Neumann et al., 2023[Neumann, P., Kloskowski, P. & Ficner, R. (2023). microLife, 4, uqad021.]). Superposition of the crystal structure of the CdaA–compound 7 complex (PDB entry 8c4p) with those of the B04 and D07 fragment complexes reported here (Fig. 6[link]) provides a solid background for the further development (`growing') of compound 7. This common strategy starting from prepositioned fragments and molecules has already been successfully used in structure-based design to identify high-affinity inhibitors for a variety of targets (Kick et al., 1997[Kick, E. K., Roe, D. C., Skillman, A. G., Liu, G., Ewing, T. J. A., Sun, Y., Kuntz, I. D. & Ellman, J. A. (1997). Chem. Biol. 4, 297-307.]; Liebeschuetz et al., 2002[Liebeschuetz, J. W., Jones, S. D., Morgan, P. J., Murray, C. W., Rimmer, A. D., Roscoe, J. M. E., Waszkowycz, B., Welsh, P. M., Wylie, W. A., Young, S. C., Martin, H., Mahler, J., Brady, L. & Wilkinson, K. (2002). J. Med. Chem. 45, 1221-1232.]).

[Figure 6]
Figure 6
Comparison of the ATP-binding mode with fragments B04 and D07 and the in silico-designed CdaA inhibitor compound 7. The structural —N=C—NH2 pattern with two N atoms responsible for hydrogen-bond formation with the main-chain atoms of Leu188 is marked. The ππ stacking interactions are shown as yellow sticks. Tyr187 and Leu188 are shown as sticks. (a) ATP-binding mode in the active site of CdaA. (b) Superposition of CdaA structures with bound D07 and ATP molecules. (c) Superposition of the CdaA structures with bound B04 and ATP molecules. (d) Superposition of the structure of compound 7 with the structures of B04 and D07 provides a solid background for the development (`growing') of compound 7.

4. Discussion

To facilitate the development of antibacterial therapeutics targeting Lm CdaA, we recently determined several CdaA crystal structures in apo (1.45 Å resolution), substrate ATP-bound (1.97 Å resolution; Neumann et al., 2023[Neumann, P., Kloskowski, P. & Ficner, R. (2023). microLife, 4, uqad021.]) and post-catalytic c-di-AMP-bound states (Heidemann et al., 2019[Heidemann, J. L., Neumann, P., Dickmanns, A. & Ficner, R. (2019). J. Biol. Chem. 294, 10463-10470.]), as well as a complex structure with the in silico-designed CdaA inhibitor (compound 7) at 1.2 Å resolution (Neumann et al., 2023[Neumann, P., Kloskowski, P. & Ficner, R. (2023). microLife, 4, uqad021.]). These structures provide insight into the metal-dependent catalytic mechanism and confirmed the druggability of the highly conserved active site (Figs. 1[link] and 4[link]). In order to explore the possibility of identifying additional binding pockets and to probe the available chemical space, a crystallographic fragment screen with 96 fragments was performed. Our main goal was to explore the possibility of obtaining CdaA complexes with non-ATP-like molecules that can be used for structure-guided optimization and the design of new compounds, with the future prospect of the rational design of antimicrobial therapeutics. We opted for soaking experiments, even though the crystallization conditions were not optimal (the buffer contained 3.6 M NaCl) and probably affected the solubilization of several of the compounds tested. Consequently, the hit rate of the fragment-screening campaign (a fragment binding to a particular site of the protein) was relatively low (∼8%). Nevertheless, despite the high-salt soaking conditions, the localized fragment molecules showed high occupancies and strong and well defined electron-density maps (Fig. 2[link]). The two fragments identified in the conserved active site of CdaA, which share the N-(2-pyridyl) moiety, underline the importance of the —N=C—NH2 structural pattern known from the adenine ring, which seems to be crucial for binding. This chemical and structural constraint could be beneficial for the future development of an antibacterial therapeutic for several bacteria that should target CdaA from different species, for example Listeria monocytogenes, Borrelia turicatea, Entero­coccus faecalis, Staphylococcus aureus, Mycobacterium tuberculosis and Streptococcus pneumoniae. The first step towards this goal is the optimization of the abovementioned in silico-designed CdaA inhibitor (compound 7) with the fragment molecules (B04 and D07) reported here, which share the N-(2-pyridyl) moiety. The optimization should be accompanied by the structural and biochemical characterization of CdaA from other bacterial species, preferably multidrug-resistant E. faecalis and S. aureus. Taken together, this would be a small step towards an important goal: the expansion of our effective antimicrobial arsenal.

In recent years, especially since the beginning of the COVID-19 pandemic, several crystallographic fragment-screening campaigns have been conducted and made available to the scientific community to facilitate the development of new antiviral therapeutics (Boby et al., 2023[Boby, M. L., Fearon, D., Ferla, M., Filep, M., Koekemoer, L., Robinson, M. C., Chodera, J. D., Lee, A. A., London, N., von Delft, A., von Delft, F., Achdout, H., Aimon, A., Alonzi, D. S., Arbon, R., Aschenbrenner, J. C., Balcomb, B. H., Bar-David, E., Barr, H. & Zvornicanin, S. N. (2023). Science, 382, eabo7201.]). X-ray crystallography is an excellent technique that provides structural information that enables the rapid and efficient assessment of hits that can subsequently be used for fragment-based drug discovery. However, the binding events must be carefully evaluated to account for artifacts such as fragment binding between symmetry-related protein interfaces. Therefore, the influence of the crystal lattice on fragment binding should not be neglected in both co-crystallization and soaking experiments. In some cases, when the target macromolecule forms oligomers that are either present in the asymmetric unit or formed by the crystallographic symmetry, the contribution of the lattice is beneficial. In many other cases, however, the binding site shared by the molecules indicates that the binding event is a product of the contribution of more than one protein molecule. It is beyond the scope of this manuscript to discuss the utility of such binding events for fragment-based drug design. Regardless of this, these events are also probes of the protein surface and could therefore provide useful insights into preferred chemical groups, undiscovered side-chain flexibility etc. Therefore, evaluating which receptor molecule is the major contributor to the observed binding event is crucial for successful fragment-based drug design relying on the most reliable fragment-binding region as a starting point.

Footnotes

These authors contributed equally to this work.

Acknowledgements

We would like to thank ESRF, Grenoble and PETRA III, Hamburg for the beamtime allocation. We would further like to thank Daniel Weinrich for his general technical assistance. Open access funding enabled and organized by Projekt DEAL. Author contributions were as follows. Funding acquisition and resources: RF. Experimental design: RF, JH, PN. Protein preparation, crystallization and soaking experiments: JH, AD. Data acquisition: JH, PN, MA. Data processing and refinement: PN. Data analysis: PN, JH, JW, RF, MW. Manuscript writing: PN. Manuscript review: all co-authors.

Funding information

This work was funded by the Deutsche Forschungsgemein­schaft (DFG) via Program SPP1879 (to RF) and INST186/1117 (to RF).

References

First citationAfonine, P. V., Poon, B. K., Read, R. J., Sobolev, O. V., Terwilliger, T. C., Urzhumtsev, A. & Adams, P. D. (2018). Acta Cryst. D74, 531–544.  Web of Science CrossRef IUCr Journals Google Scholar
First citationAgirre, J., Atanasova, M., Bagdonas, H., Ballard, C. B., Baslé, A., Beilsten-Edmands, J., Borges, R. J., Brown, D. G., Burgos-Mármol, J. J., Berrisford, J. M., Bond, P. S., Caballero, I., Catapano, L., Chojnowski, G., Cook, A. G., Cowtan, K. D., Croll, T. I., Debreczeni, J. É., Devenish, N. E., Dodson, E. J., Drevon, T. R., Emsley, P., Evans, G., Evans, P. R., Fando, M., Foadi, J., Fuentes-Montero, L., Garman, E. F., Gerstel, M., Gildea, R. J., Hatti, K., Hekkelman, M. L., Heuser, P., Hoh, S. W., Hough, M. A., Jenkins, H. T., Jiménez, E., Joosten, R. P., Keegan, R. M., Keep, N., Krissinel, E. B., Kolenko, P., Kovalevskiy, O., Lamzin, V. S., Lawson, D. M., Lebedev, A. A., Leslie, A. G. W., Lohkamp, B., Long, F., Malý, M., McCoy, A. J., McNicholas, S. J., Medina, A., Millán, C., Murray, J. W., Murshudov, G. N., Nicholls, R. A., Noble, M. E. M., Oeffner, R., Pannu, N. S., Parkhurst, J. M., Pearce, N., Pereira, J., Perrakis, A., Powell, H. R., Read, R. J., Rigden, D. J., Rochira, W., Sammito, M., Sánchez Rodríguez, F., Sheldrick, G. M., Shelley, K. L., Simkovic, F., Simpkin, A. J., Skubak, P., Sobolev, E., Steiner, R. A., Stevenson, K., Tews, I., Thomas, J. M. H., Thorn, A., Valls, J. T., Uski, V., Usón, I., Vagin, A., Velankar, S., Vollmar, M., Walden, H., Waterman, D., Wilson, K. S., Winn, M. D., Winter, G., Wojdyr, M. & Yamashita, K. (2023). Acta Cryst. D79, 449–461.  Web of Science CrossRef IUCr Journals Google Scholar
First citationAsokan, G. V., Ramadhan, T., Ahmed, E. & Sanad, H. (2019). Oman Med. J. 34, 184–193.  CrossRef CAS PubMed Google Scholar
First citationBai, Y., Yang, J., Zhou, X., Ding, X., Eisele, L. E. & Bai, G. (2012). PLoS One, 7, e35206.  CrossRef PubMed Google Scholar
First citationBlötz, C., Treffon, K., Kaever, V., Schwede, F., Hammer, E. & Stülke, J. (2017). Front. Microbiol. 8, 1328.  PubMed Google Scholar
First citationBoby, M. L., Fearon, D., Ferla, M., Filep, M., Koekemoer, L., Robinson, M. C., Chodera, J. D., Lee, A. A., London, N., von Delft, A., von Delft, F., Achdout, H., Aimon, A., Alonzi, D. S., Arbon, R., Aschenbrenner, J. C., Balcomb, B. H., Bar-David, E., Barr, H. & Zvornicanin, S. N. (2023). Science, 382, eabo7201.  CrossRef PubMed Google Scholar
First citationCassini, A., Högberg, L. D., Plachouras, D., Quattrocchi, A., Hoxha, A., Simonsen, G. S., Colomb-Cotinat, M., Kretzschmar, M. E., Devleesschauwer, B., Cecchini, M., Ouakrim, D. A., Oliveira, T. C., Struelens, M. J., Suetens, C., Monnet, D. L., Strauss, R., Mertens, K., Struyf, T., Catry, B., Latour, K., Ivanov, I. N., Dobreva, E. G., Tambic Andraševic, A., Soprek, S., Budimir, A., Paphitou, N., Žemlicková, H., Schytte Olsen, S., Wolff Sönksen, U., Märtin, P., Ivanova, M., Lyytikäinen, O., Jalava, J., Coignard, B., Eckmanns, T., Abu Sin, M., Haller, S., Daikos, G. L., Gikas, A., Tsiodras, S., Kontopidou, F., Tóth, Á., Hajdu, Á., Guðlaugsson, Ó., Kristinsson, K. G., Murchan, S., Burns, K., Pezzotti, P., Gagliotti, C., Dumpis, U., Liuimiene, A., Perrin, M., Borg, M. A., de Greeff, S. C., Monen, J. C., Koek, M. B., Elstrøm, P., Zabicka, D., Deptula, A., Hryniewicz, W., Caniça, M., Nogueira, P. J., Fernandes, P. A., Manageiro, V., Popescu, G. A., Serban, R. I., Schréterová, E., Litvová, S., Štefkovicová, M., Kolman, J., Klavs, I., Korošec, A., Aracil, B., Asensio, A., Pérez-Vázquez, M., Billström, H., Larsson, S., Reilly, J. S., Johnson, A. & Hopkins, S. (2019). Lancet Infect. Dis. 19, 56–66.  CrossRef PubMed Google Scholar
First citationCommichau, F. M., Heidemann, J. L., Ficner, R. & Stülke, J. (2019). J. Bacteriol. 201, 1–14.  CrossRef Google Scholar
First citationCorrigan, R. M., Abbott, J. C., Burhenne, H., Kaever, V. & Gründling, A. (2011). PLoS Pathog. 7, e1002217.  Web of Science CrossRef PubMed Google Scholar
First citationCorrigan, R. M. & Gründling, A. (2013). Nat. Rev. Microbiol. 11, 513–524.  CrossRef CAS PubMed Google Scholar
First citationDiederichs, K. & Karplus, P. A. (1997). Nat. Struct. Mol. Biol. 4, 269–275.  CrossRef CAS Web of Science Google Scholar
First citationEberhardt, J., Santos-Martins, D., Tillack, A. F. & Forli, S. (2021). J. Chem. Inf. Model. 61, 3891–3898.  Web of Science CrossRef CAS PubMed Google Scholar
First citationEmsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. (2010). Acta Cryst. D66, 486–501.  Web of Science CrossRef CAS IUCr Journals Google Scholar
First citationErlanson, D. A., Fesik, S. W., Hubbard, R. E., Jahnke, W. & Jhoti, H. (2016). Nat. Rev. Drug Discov. 15, 605–619.  Web of Science CrossRef CAS PubMed Google Scholar
First citationGiordanetto, F., Jin, C., Willmore, L., Feher, M. & Shaw, D. E. (2019). J. Med. Chem. 62, 3381–3394.  CrossRef CAS PubMed Google Scholar
First citationGundlach, J., Rath, H., Herzberg, C., Mäder, U. & Stülke, J. (2016). Front. Microbiol. 7, https://doi.org/10.3389/fmicb.2016.00804CrossRef Google Scholar
First citationHeidemann, J. L., Neumann, P., Dickmanns, A. & Ficner, R. (2019). J. Biol. Chem. 294, 10463–10470.  CrossRef CAS PubMed Google Scholar
First citationHuschmann, F. U., Linnik, J., Sparta, K., Ühlein, M., Wang, X., Metz, A., Schiebel, J., Heine, A., Klebe, G., Weiss, M. S. & Mueller, U. (2016). Acta Cryst. F72, 346–355.  Web of Science CrossRef IUCr Journals Google Scholar
First citationJackson-Litteken, C. D., Ratliff, C. T., Kneubehl, A. R., Siletti, C., Pack, L., Lan, R., Huynh, T. N., Lopez, J. E. & Blevins, J. S. (2021). Infect. Immun. 89, e00787-20.  CAS PubMed Google Scholar
First citationKabsch, W. (2010). Acta Cryst. D66, 125–132.  Web of Science CrossRef CAS IUCr Journals Google Scholar
First citationKick, E. K., Roe, D. C., Skillman, A. G., Liu, G., Ewing, T. J. A., Sun, Y., Kuntz, I. D. & Ellman, J. A. (1997). Chem. Biol. 4, 297–307.  CrossRef CAS PubMed Google Scholar
First citationLiebeschuetz, J. W., Jones, S. D., Morgan, P. J., Murray, C. W., Rimmer, A. D., Roscoe, J. M. E., Waszkowycz, B., Welsh, P. M., Wylie, W. A., Young, S. C., Martin, H., Mahler, J., Brady, L. & Wilkinson, K. (2002). J. Med. Chem. 45, 1221–1232.  CrossRef PubMed CAS Google Scholar
First citationLiebschner, D., Afonine, P. V., Baker, M. L., Bunkóczi, G., Chen, V. B., Croll, T. I., Hintze, B., Hung, L.-W., Jain, S., McCoy, A. J., Moriarty, N. W., Oeffner, R. D., Poon, B. K., Prisant, M. G., Read, R. J., Richardson, J. S., Richardson, D. C., Sammito, M. D., Sobolev, O. V., Stockwell, D. H., Terwilliger, T. C., Urzhumtsev, A. G., Videau, L. L., Williams, C. J. & Adams, P. D. (2019). Acta Cryst. D75, 861–877.  Web of Science CrossRef IUCr Journals Google Scholar
First citationLiebschner, D., Afonine, P. V., Moriarty, N. W., Poon, B. K., Sobolev, O. V., Terwilliger, T. C. & Adams, P. D. (2017). Acta Cryst. D73, 148–157.  Web of Science CrossRef IUCr Journals Google Scholar
First citationMcNutt, A. T., Francoeur, P., Aggarwal, R., Masuda, T., Meli, R., Ragoza, M., Sunseri, J. & Koes, D. R. (2021). J. Cheminform, 13, 43.  CrossRef PubMed Google Scholar
First citationMehne, F. M. P., Gunka, K., Eilers, H., Herzberg, C., Kaever, V. & Stülke, J. (2013). J. Biol. Chem. 288, 2004–2017.  CrossRef CAS PubMed Google Scholar
First citationMurshudov, G. N., Skubák, P., Lebedev, A. A., Pannu, N. S., Steiner, R. A., Nicholls, R. A., Winn, M. D., Long, F. & Vagin, A. A. (2011). Acta Cryst. D67, 355–367.  Web of Science CrossRef CAS IUCr Journals Google Scholar
First citationNeumann, P., Kloskowski, P. & Ficner, R. (2023). microLife, 4, uqad021.  CrossRef PubMed Google Scholar
First citationNeumann, P. & Tittmann, K. (2014). Curr. Opin. Struct. Biol. 29, 122–133.  Web of Science CrossRef CAS PubMed Google Scholar
First citationPearce, N. M., Krojer, T. & von Delft, F. (2017). Acta Cryst. D73, 256–266.  Web of Science CrossRef IUCr Journals Google Scholar
First citationRömling, U. (2008). Sci. Signal. 1, pe39.  Web of Science PubMed Google Scholar
First citationRørvik, G. H., Liskiewicz, K. A., Kryuchkov, F., Naemi, A., Aasheim, H., Petersen, F. C., Küntziger, T. M. & Simm, R. (2020). Microorganisms, 8, 1269.  PubMed Google Scholar
First citationRosenberg, J., Dickmanns, A., Neumann, P., Gunka, K., Arens, J., Kaever, V., Stülke, J., Ficner, R. & Commichau, F. M. (2015). J. Biol. Chem. 290, 6596–6606.  CrossRef CAS PubMed Google Scholar
First citationSanner, M. (1999). J. Mol. Graph. Model. 17, 57–61.  Web of Science PubMed CAS Google Scholar
First citationSchuster, C. F., Bellows, L. E., Tosi, T., Campeotto, I., Corrigan, R. M., Freemont, P. & Gründling, A. (2016). Sci. Signal. 9, ra81.  CrossRef PubMed Google Scholar
First citationTrott, O. & Olson, A. J. (2010). J. Comput. Chem. 31, 455–461.  Web of Science CrossRef PubMed CAS Google Scholar
First citationWhiteley, A. T., Garelis, N. E., Peterson, B. N., Choi, P. H., Tong, L., Woodward, J. J. & Portnoy, D. A. (2017). Mol. Microbiol. 104, 212–233.  CrossRef CAS PubMed Google Scholar
First citationWitte, G., Hartung, S., Büttner, K. & Hopfner, K. (2008). Mol. Cell, 30, 167–178.  CrossRef PubMed CAS Google Scholar
First citationWojdyr, M., Keegan, R., Winter, G. & Ashton, A. (2013). Acta Cryst. A69, s299.  Web of Science CrossRef IUCr Journals Google Scholar
First citationWollenhaupt, J., Barthel, T., Lima, G. M. A., Metz, A., Wallacher, D., Jagudin, E., Huschmann, F. U., Hauss, T., Feiler, C. G., Gerlach, M., Hellmig, M., Förster, R., Steffien, M., Heine, A., Klebe, G., Mueller, U. & Weiss, M. S. (2021). J. Vis. Exp., e62208.  Google Scholar
First citationWoodward, J. J., Iavarone, A. T. & Portnoy, D. A. (2010). Science, 328, 1703–1705.  Web of Science CrossRef CAS PubMed Google Scholar
First citationZarrella, T. M., Yang, J., Metzger, D. W. & Bai, G. (2020). J. Bacteriol. 202, https://doi.org/10.1128/jb.00691-19CrossRef Google Scholar

This is an open-access article distributed under the terms of the Creative Commons Attribution (CC-BY) Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original authors and source are cited.

Journal logoSTRUCTURAL
BIOLOGY
ISSN: 2059-7983
Follow Acta Cryst. D
Sign up for e-alerts
Follow Acta Cryst. on Twitter
Follow us on facebook
Sign up for RSS feeds