Subscriber access provided by Kaohsiung Medical University
Molecular basis for the attachment of S-layer proteins to the cell wall of Bacillus anthracis David Sychantha, Robert Chapman, Natalie C Bamford, Geert-Jan Boons, P Lynne Howell, and Anthony John Clarke Biochemistry, Just Accepted Manuscript • DOI: 10.1021/acs.biochem.8b00060 • Publication Date (Web): 09 Mar 2018 Downloaded from http://pubs.acs.org on March 9, 2018
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 6 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Biochemistry
Molecular Basis for the Attachment of S-Layer Proteins to the Cell Wall of Bacillus anthracis David Sychantha†, Robert N. Chapman&, Natalie C. Bamford‡', Geert-Jan Boons&, P. Lynne Howell‡' and Anthony J. Clarke†* †
Dept. of Molecular and Cellular Biology, University of Guelph, Guelph, ON, N1G 2X1 Canada.
&
‡
Complex Carbohydrate Research Center, University of Georgia, 315 Riverbend Road, Athens, Georgia 30602, USA. Program in Molecular Medicine, The Hospital for Sick Children, Toronto, ON, M5G 0A4 Canada.
'
Dept. of Biochemistry, University of Toronto, ON, M5S 1A8 Canada.
ABSTRACT: Bacterial surface (S)-layers are paracrystalline arrays of protein assembled on the bacterial cell wall which serve as protective barriers and scaffolds for housekeeping enzymes and virulence factors. The attachment of S-layer proteins to the cell walls of the Bacillus cereus sensu lato, which includes the pathogen Bacillus anthracis, occurs through non-covalent interactions between their S-layer homology domains and secondary cell wall polysaccharides. To promote recognition for these interactions, it is presumed that the terminal N-acetylmannosamine (ManNAc) residues of the secondary cell wall polysaccharides must be ketalpyruvylated. For a few specific S-layer proteins, the O-acetylation of the penultimate N-acetylglucosamine (GlcNAc) is also required. Herein, we present the X-ray crystal structure of the SLH domain of the major surface array protein Sap from B. anthracis in complex with 4,6-O-ketal-pyruvyl-β-ManNAc-(1,4)-βGlcNAc-(1,6)-α-GlcN. This structure reveals for the first time that the conserved terminal SCWP unit is the direct ligand for the SLH domain. Furthermore, we identify key binding interactions which account for the requirement of 4,6-O-ketal-puruvlyl-ManNAc while revealing the insignificance of the O-acetylation on the GlcNAc residue for recognition by Sap.
The bacterial cell envelope plays a key role in a pathogen’s ability to evade host defense mechanisms. Like all Gram-positive bacteria, the cytoplasmic membrane of the Bacillus cereus group of pathogens, which includes Bacillus anthracis,1 is encased by a thick peptidoglycan sacculus. Secondary cell wall polysaccharides (SCWP) are covalently attached to peptidoglycan and extend out from the cell. The SCWPs form a scaffold onto which a surfacelayer (S-layer) self assembles.2 S-Layers are bi-dimensional paracrystalline arrays of one or more major proteins and a host of Slayer-associated proteins that assist in the pathogenesis of infections by providing protection from host defense factors.35 The two major S-layer proteins produced by B. anthracis are surface array protein (Sap) and extractable antigen 1 (EA1).6,7 The attachment of these and the 22 other S-layer associated proteins to the SCWPs of B. anthracis occurs through non-covalent interactions involving their S-layer homology (SLH) domains.8,9 Each of the B. anthracis S-layers proteins uses three consecutive repeats of the SLH domain for this binding.3 SLH domains are α-helical and their topological arrangement displays three-fold pseudosymmetry resembling a trefoil. Each domain forms a prong of the
trefoil, and the SCWPs are presumed to bind within the clefts of the three inter-prong grooves (IPG).10 The repeating trisaccharide unit of B. anthracis SCWP is composed of N-acetylglucosamine (GlcNAc) and N-acetyl-mannosamine (ManNAc) [→4)-β-ManNAc-(1→4)-β-GlcNAc-(1→6)-αGlcNAc-(1→].11 Strain-dependent galactosylations occur at O-3 of the β-GlcNAc residues and at O-3 and/or O-4 of the αGlcNAc12 (Figure 1A). Ketal-pyruvylation of the C-4 and C-6 hydroxyl groups of the terminal -ManNAc residue is thought to be required for recognition by the SLH domains of each S-layer protein,10 while O-acetylation of the β-GlcNAc appears to facilitate the attachment of only a few specific proteins, including EA1, BslO, and BslA.12 Both modifications occur on the same unit suggesting the modified SCWP terminus is the ligand recognized by SLH domains, however this has not been demonstrated directly. To provide structural insight into how the SLH domain of Sap (SapSLH) recognizes its ligand, we synthesized an analog of the terminal SCWP unit, 4,6-O-ketal-pyruvyl-β-ManNAc-(1,4)-βGlcNAc-(1,6)-α-GlcN with a pentylamino aglycone at the C1 oxygen of GlcN (MGG) (Figure 1A); the aglycone was needed as a linker for other investigations with MGG. Crystals of the SapSLH-MGG complex were obtained after a 20 min soaking time at pH 5.5. These crystals belonged to the space group P41212 and diffracted to 2.3 Հ resolution (Supplemental Table 1, Supporting Information). The general conformation of holo SapSLH was similar to the apo protein (RMSD = 0.60 Հ over 160 residues), where SLH1, SLH2, and SLH3 are configured in a pseudo-trimer with trefoil topology (Figure 1B). This domain configuration consequently forms three corresponding clefts, termed inter-prong grooves (IPG1–IPG3), which were previously postulated to function as SCWP binding sites. Positive electron density was observed within IPG2, between its two loops composed of Ile116–Gly122 in SLH2 (termed the B2-loop) and Glu160–Gly165 of SLH3 (termed the A3 loop). This density was best interpreted as the two non-reducing residues of MGG, as no electron density was observed for the α-1,6-linked GlcN residue at the reducing end after refinement (Figure 1C). Additionally, IPG3 contained electron density that best corresponded to a bound SO4 molecule previously not observed here in the apo structure.10 However, there is some undefined electron density within this region of the map originally deposited for the Apo enzyme (3PYW).
ACS Paragon Plus Environment
1
Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 2 of 6
FIGURE 1. Overall structure of SapSLH in complex with a terminal SCWP unit. (A) Terminal and repeat unit glycan structures are presented in extended form, where carbohydrate decorations and non-carbohydrate modifications are in blue and red, respectively. (B) Structural views of holo SapSLH in cartoon representation oriented at 90º relative to each other. (C) Structure of synthetic SCWP (MGG) unit. Top, post-refinement 2mFo-2Fc electron density map (blue) and pre-refinement mFo-Fc omit map (yellow) around MGG bound to SapSLH, contoured at 1.0 σ and 3.0 σ, respectively. Amino acid side chains and MGG are shown in stick representation. Bottom, chemical structure with pyruvyl ketal colored in red. Within IPG2, key interactions were identified with the 4,6-Oketal pyruvyl group of MGG, in which the methyl group is buried in a hydrophobic pocket while the carboxyl group formed an electrostatic interaction (2.6 Հ) with the NH2 of the guanidinium moiety of Arg72 (Figure 2A). A nearby oxygen atom, from the C-6 acetal linkage of -ManNAc, was also hydrogen bonded to Arg72, but through NH1 instead (3.0 Հ). Carboxyl group interactions of the ketal pyruvyl group further involved two hydrogen bonds from backbone atoms of the B2-loop, specifically though the backone amide and carbonyl oxygen of Lys117. The C-3 hydroxyl group of β-ManNAc was identified in a hydrogen bond to the carbonyl oxygen of Gly118, while the remaining internal B2-loop residues Thr199, Gly120 and Asn121 contributed to binding, but mainly through van der Waals contacts. Against the opposite face of the ligand, stacking interactions involving both sugar rings were observed with the indole ring of the conserved Trp164 (Figure 2B). The trefoil topology of the SapSLH domains suggests each IPG is capable of binding SCWP, although only IPG2 of the holo structure contains an MGG. Examination of crystal structure symmetry mates showed that while IPG2 is exposed, IPG1 is obstructed by an adjacent protomer and IPG3 is involved in packing. As MGG concentrations above a 1:1 protein/ligand ratio (calculated based on protein concentration in the crystallization drop) caused crystals to crack and dissolve, we speculate that the ligands inevitably interact and disrupt packing of SapSLH. Sulfate binding to IPG3 was likely possible because of its small size, avoiding perturba-
tion the crystal lattice. That sulfate is coordinated similarly to the pyruvate ketal within IPG2, and because crystals are sensitive to > 1:1 SapSLH /MGG ratios, SapSLH interactions with SCWP are most consistent with a multiple binding model. The intramolecular interactions that occur with the SO4 molecule bound within IPG3 closely resemble those involved in binding to the carboxyl group of 4,6-O-ketal-pyruvyl-β-ManNAc in IPG2. NH1 and NH2 from the guanidinum group of Arg131 are involved in electrostatic interactions with the O4 and O2 atoms of SO4, which occupy the positions equivalent to the pyruvate ketal. In addition, the backbone amide of Val179 is in position to hydrogen bond to O4 of SO4. The latter is facilitated by the minor backbone motions that occurred in Val179 and Gly180, relative to the apo structure (Figure 2C). Detailed comparison of Cα RMSD values of the apo and holo SapSLH structures revealed considerable rearrangement of the A and B-loops (Figure 3A). Minor displacements of 1.8 Հ and 1.1 Հ (Thr161 and Asn163, respectively) were observed in the A3-loop of IPG2, in which the side chain conformation of Trp164 was rotated relative to the apo structure. The most significant change was that of the B2-loop, where a maximum displacement of 4.6 Հ occurred in Thr119 Cα, which appears to be critical for properly positioning backbone residues to interact with MGG. As a result, the side chain of Thr119 is displaced by a maximum of 6.9 Հ (at the Cɣ), and the new positions of its backbone amide (2.8 Հ) and Oɣ (3.2 Հ) form hydrogen bonds to the carbonyl oxygen of Gly93. These latter hydrogen bonds likely help to stabilize the
ACS Paragon Plus Environment
2
Page 3 of 6 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Biochemistry
FIGURE 2. Characterization of ligands bound within the IPGs of SapSLH. (A) Surface cut-away of IPG2 containing bound MGG. (B) Detailed views of IPG2 showing intermolecular contacts with MGG oriented at 180 º relative to each other. (C) Detailed view of IPG3 showing intermolecular contacts with SO4 oriented at 180 º relative to each other. Amino acid side chains (white), MGG (green) and SO4 (yellow) are shown in stick representation. Hydrogen bonds are depicted as black dashed lines.
conformational state of the loop (Figure 3B). Similar structural deviations of 2.4 Հ and 1.4 Հ (Val179 and Gly180, respectively) were found within the B3-loop of IPG3. It is also worth noting that whereas residues Lys58 Met63 belonging to the B1-loop were modeled in the apo form, the electron density for these residues in the holo form were missing, suggesting a high degree of flexibility at this region. Whereas the previous structural characterization of apo SapSLH described a resting, or “loose” state of IPG2,10 association of the groove reveals that loop rearrangements are important for ligand interactions. The loop movement of IPG2 does not significantly affect the width of its opening around β-GlcNAc, but it narrows the interior cavity around 4,6-O-ketal-pyruvyl-β-ManNAc by 2 Հ, reducing the internal pocket volume by approximately 20 % (from 261 Հ3 to 211 Հ3) forming a “tight” state (Figure 3C). B-Loop movement appeared to have occurred though rotations in the dihedral angles of two glycine residues flanking the mobile region. The invariant Gly118 residue of the B2-loop (Gly57 and Gly180
in A1 and A3-loops, respectively) is rotated by ~ 50 ̊, accompanied by a corresponding rotation of Gly120. The amino acid composition of this loop, including its B1- and B3-loop equivalents, are variable amongst the 24 S-layer proteins in B. anthracis, however a consensus sequence of ΦΦxGxGxGxΦx (where Φ and x represent non-polar and polar residues, respectively) was identified to occur with the highest frequency (Figure 3B & D). The mobility of the B-loops is sharply limited at the C-terminal hydrophobic residue of this motif, as these loops appear to be secured by invariant Asp residues (Asp36/97/156). For example, in the B2-loop, Asp97 forms hydrogen bonds with the backbone amides of Phe123 and Glu124, respectively. The closing of IPG2 to a “tight” state appears to be driven by main chain interactions with the carboxyl group of the ketalpyruvyl moiety. As a result, the repositioned IPG2 latches the 4,6O-ketal-pyruvyl-β-ManNAc into position and encourages tight association of the sugar unit against Trp164. Whereas this looplatching mechanism involving the closing of a loop to sequester ligand from solvent is a common feature of many proteins, it is unusual for most of those involved in bacterial cell wall attachment. In this latter situation, ligands extend across the surfaces of binding domains. For example, the wall teichoic acid (WTA) binding domains of the choline-binding proteins produced by Streptococcus pneumoniae involve long multi-modular domains. It is speculated that the entire WTA polymer is wrapped around these domains in a helical fashion.13 With the LysM domains of the various enzymes that bind peptidoglycan,14 such as AtlA of Enterococcus faecalis, up to six inter-connected domains facilitate edge-on interactions with the glycans in shallow surface grooves.15 Of all of the bacterial cell wall binding domains studied to date, the CWB2 binding domain of Clostridium difficile surface proteins16 appears to be most similar structurally to the SLH domains of B. cereus group of bacteria (rmsd, 3.8 C over 96 residues), however the mechanism of ligand binding remains unknown. The holo structure of SapSLH provides structural rationale for earlier observations noting the importance of each of the three invariant basic residues (Arg72, Arg131, and Lys193) in their respective IPGs for ligand binding.10 Despite this, it is clear with IPG2 that Arg72 only contributes to a fraction of the total interactions, as significant binding contributions are also made by the B2loop and a hydrophobic pocket at the distal end of the groove. Comparable losses in binding were also obtained in SapSLH variants lacking the invariant Asp residues Asp36, Asp97, Asp158.10 These conserved Asp residues were previously proposed to be involved in peptidoglycan binding. However, analysis of the complex structure shows that they are more likely to be responsible (at least in part) for stabilizing the B-loops of the IPGs. Evidence for this is provided by the observation that Asp36 is missing at the N-terminus of the holo structure, and that the corresponding electron density of the B1-loop is absent due to high thermal motions. Therefore, the perturbation of SCWP binding in the triple Asp mutant could be the result of an inability to properly rearrange its B-loops. Non-stoichiometric O-acetylation of the terminal unit of B. anthracis SCWP occurs at the C-3 hydroxyl group of the βGlcNAc,12 which is catalyzed by PatB1.17 This modification is important for recognition of SCWP by EA1, but not for Sap. Whereas SapSLH interacts with the β-GlcNAc of MGG, its C-3 hydroxyl group does not contact the protein. Examination of the structure could not identify backbone or side chain atoms that could potentially either interact with or occlude the addition of an acetyl group at the C-3 position. Therefore, the binding interactions (or lack thereof) identified in the structure provide some explanation for why Sap can recognize terminal SCWP both with and without O-acetylation. Yet the question regarding the specificity of EA1 for O-acetylated SCWP still remains. It stands to reason that since the B2-loop undergoes extensive rearrangement
ACS Paragon Plus Environment
3
Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 4 of 6
FIGURE 3. Rearrangement of the B2-loop of IPG2 in the holo SapSLH structure. (A) Top: Mean normalized B-factor (Հ2) of C atoms from apo (blue) and holo (red) SapSLH. Bottom: Root-mean-square deviation (Հ) of Cα atoms from holo and apo SapSLH. Loop designations are labelled above. (B) Left, comparison of the SLH2 domain of apo (white) and holo (blue) SapSLH shown depicting B2-loop rearrangements. Right, 90 º rotation of the latter depicting the conserved hydrophobic region and loop-stabilizing hydrogen bonds. (C) Size and volume comparison of the IPG2 cavity of the apo (left) and holo (right) forms of SapSLH. (D) Identification of consensus motif within the B2-loop of SLH2 domains from the plasmid and chromosomally encoded S-layer proteins of B. anthracis Sterne. The xGxGxGxx motif applies to the B2-loops of all three SLH domains in the S-layer proteins. to latch onto MGG, maybe specificity determinants for O-acetyl SCWP could be conserved within its sequence. In comparing the B2-loop sequence of Sap with O-acetylation specific S-layer proteins, EA1, BslO, and BslA, Thr119 is replaced with an Asp. It is tempting to speculate that since Thr119 is the closest (4 Հ) residue to the C-3 of β-GlcNAc in SapSLH, then perhaps the longer side chain of Asp at this position in EA1 could protrude toward the Oacetyl group and promote an interaction. However, further structural evidence would be required to confirm this.
Information is available free of charge on the ACS Publications website at DOI:
AUTHOR INFORMATION Corresponding Author * E-mail:
[email protected]. Phone: +1-519-824-4120 ORCID Anthony Clarke: 0000-0003-4076-0488
Notes
ASSOCIATED CONTENT
The authors declare no competing financial interests.
Supporting Information
Author Contributions
Experimental procedures for production and purification of SapSLH, synthesis of MGG, crystallization and structural determination apo SapSLH and SapSLH in complex with MGG, and other analytical procedures, and Table S1, Summary of data collection and refinement statistics for SapSLH-MGG. The Supporting
DS and RNC conceived, and AJC supervised the study; DS and NCB designed and performed experiments; RNC and G-JB provided the MGG; DS and NCB analyzed the data; DS and AJC wrote the manuscript; PLH, NCB, and RNC made manuscript revisions.
ACS Paragon Plus Environment
4
Page 5 of 6 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Biochemistry Funding Sources These studies were supported in part by operating grants from GlycoNet (a Canadian National Centre of Excellence) to AJC, and the Canadian Institutes of Health Research (CIHR) to PLH (MOP 43998). PLH is the recipient of a Canada Research Chair.
ABBREVIATIONS S-Layer, surface layer; SCWP, secondary cell wall polysaccharide; ManNAc, N-acetylmannosamine; GlcNAc, N-acetylglucosamine; Sap, surface array protein; EA1, extractable antigen 1; SLH, S-layer homology; SapSLH, SLH domain of Sap; MGG; 4,6O-ketal-pyruvyl-β-ManNAc-(1,4)-β-GlcNAc-(1,6)-α-GlcN with a pentylamino aglycone; IPG, inter-prong grooves; WTA, wall teichoic acid.
REFERENCES (1) Liu, Y., Lai, Q., Göker, M., Meier-Kolthoff, J. P., Wang, M., Sun, Y., Wang, L., and Shao, Z. Sci. Rep. 2015, 5, 14082. (2) Fouet, A. Mol. Aspects Med. 2009, 30, 374–385. (3) Kern, J. W., and Schneewind, O. Mol .Microbiol. 2008, 68, 504-515. (4) Ruthel, G., Ribot, W. J., Bavari, S., and Hoover, T. J. Infect. Dis. 2004, 189, 13131316. (5) Missiakas, D., and Schneewind, O. Annu. Rev. Microbiol. 2017, 71, 7998. (6) Mesnage, S., Tosi-Couture, E., Mock, M., Gounon, P., and Fouet, A. Mol. Microbiol. 1997, 23, 11471155. (7) Etienne-Toumelin, I., Sirard, J. C., Duflot, E., Mock, M., and Fouet, A. J. Bacteriol. 1995, 177, 614620.
(8) Mesnage, S., Fontaine, T., Mignot, T., Delepierre, M., Mock, M., and Fouet, A. EMBO J. 2000, 19, 44734484. (9) Kern, J., Ryan, C., Faull, K., and Schneewind, O. Bacillus anthracis surface-layer proteins assemble by binding to the secondary cell wall polysaccharide in a manner that requires csaB and tagO. J. Mol. Biol. 2010, 401, 757775. (10) Kern, J., Wilton, R., Zhang, R., Binkowski, T. A., Joachimiak, A., and Schneewind, O. Structure of surface layer homology (SLH) domains from Bacillus anthracis surface array protein. J. Biol. Chem. 2011, 286, 2604226049. (11) Choudhury, B., Leoff, C., Saile, E., Wilkins, P., Quinn, C. P., Kannenberg, E. L., and Carlson, R. W. J. Biol. Chem. 2006, 281, 2793227941. (12) Forsberg, L. S., Abshire, T. G., Friedlander, A., Quinn, C. P., Kannenberg, E. L., and Carlson, R. W. Glycobiology 2012, 22, 11031117. (13) Voß, F., Abdullah, M. R., Galán-Bartual, S., Seinen, J., Sánchez-Murcia, P. A., Gago, .F, Bruix, M., Hammerschmidt, S., and Hermoso, J. A. Sci. Rep. 2016, 6, 38094. (14) Buist, G., Steen, A., Kok, J., and Kuipers, O. P. Mol. Microbiol. 2008, 68(4), 838847. (15) Mesnage, S., Dellarole, M., Baxter, N. J., Rouget, J. B., Dimitrov, J. D., Wang, N., Fujimoto, Y., Hounslow, A. M., Lacroix-Desmazes, S., Fukase, K., Foster, S. J., and Williamson, M. P. Nat. Commun. 2014, 30, 4269. (16) Willing, S. E., Candela, T., Shaw, H. A., Seager, Z., Mesnage, S., Fagan, R. P., and Fairweather, N. F. Mol. Microbiol. 2015, 96, 596608. (17) Sychantha, D., Little, D. J., Chapman, R. N., Boons, G,-J., Robinson, H., Howell, P. L., and Clarke, A. J. Nat. Chem. Biol. 2017, 14, 7985.
ACS Paragon Plus Environment
5
Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 6 of 6
For Table of Contents Use Only Molecular Basis for the Attachment of S-Layer Proteins to the Cell Wall of Bacillus anthracis David Sychantha, Robert N. Chapman, Natalie C. Bamford, Geert-Jan Boons, P. Lynne Howell and Anthony J. Clarke
ACS Paragon Plus Environment
6
