Crystal structures of cholera toxin in complex with fucosylated receptors point to importance of secondary binding site

Subject terms: X-ray crystallography, Glycobiology

Table 1.

Data collection and refinement statistics.

Table 2.

K_d values for CTB and CT variants and glycans GM1os, Le^x triaose and tetraose determined by SPR. For comparison, CTB binds to Le^y tetraose with a K_d of 1.05 ± 0.04 mM²¹.

Crystal structures

Protein interaction by SPR

Le^x and similar fucosylated structures bind to the secondary binding site of CT

We set out to explore the molecular interaction of CT with fucosylated receptors, in particular Le^x. Already twenty years ago, CTB was shown to bind fucose⁴⁵. More recently, evidence was presented that HBGAs may act as functional toxin receptors for the related LTB^46,47, and fucosylated structures were shown to be functionally active and cause cellular uptake of CT^30,32–34. For example, CTB binding to jejunal epithelial cells can be blocked by the Le^x trisaccharide or a monoclonal antibody against Le^x 32. Crosslinking and immunoprecipitation of CTB-bound cellular proteins revealed that CTB binds to glycoproteins modified with Le^x 32. It was, however, unclear how Le^x binds to CT. Since G33D, a primary site CTB variant with greatly reduced affinity to GM1^17,48, showed weaker binding to Le^x 32, it was suggested that also Le^x might bind to the primary binding site³². To explore this possibility, Cervin et al. performed competition experiments with plate-bound GM1 or Le^x and CTB pre-incubated with the free sugars³². Competition experiments were also performed with human granulocytes, T84 and Colo205 cells. Despite the fact that these cell lines only express low levels of GM1, pre-incubation of CTB with GM1os had an inhibitory effect. Taken together, these results were interpreted to indicate that GM1 competitively blocks fucosylated sugars from binding to the CT primary binding site³². However, it should be noted that even at high concentrations, GM1os could not block CTB binding completely and Le^x could not block binding to plate-bound GM1.

Here we show that Le^x binds solely to the secondary binding site of CTB, and even l-fucose alone (when co-crystallized in 500-fold molar excess) did not bind to the primary binding site. Moreover, SPR experiments with toxin variants confirmed that disruption of the secondary binding site, but not the primary binding site, lowered the affinity to Le^x (Fig. 4, Table 2). We also found that Le^x bound more weakly to CTB than Le^y or A-Le^y, which is in good agreement with inhibitor studies identifying Le^y as the most potent small-molecule inhibitor of CTB³³. The binding mode of Le^x is highly similar to that of Le^y tetraose and A-Le^y pentaose²¹, which both feature an additional fucose residue (Fig. 2c). The partial competition of GM1 and Le^x observed by Cervin et al.³² is most likely due to allosteric cross-talk, in line with recent findings^49,50. Likewise, we found that the CT secondary binding site variant H18A exhibits increased GM1os binding.

Comparison of ligand structures to structure–activity relationship data

l-fucose has been shown to block CTB binding in cell culture, serving as a promising starting point for the design of novel cholera inhibitors³³. Recently, Wands et al. described important molecular determinants of l-fucose by competition binding assays of CTB to various intestinal epithelial cells³³. Their structure–activity relationship (SAR) studies showed that the stereochemical positioning of the terminal fucose is crucial for CTB binding. While the removal of hydroxyl groups on C3 or C4 (OH3, OH4) had a significant effect on CTB binding, substitution of the hydroxyl group at C2 (OH2) did not result in reduced binding (compare Fig. 2c). The addition of a hydroxyl group to C6 was tolerated, but the removal of C6 resulted in a complete loss of CTB binding suggesting the importance of a hydrophobic patch within the binding pocket. Addition of a methyl group to OH1, locking the anomeric hydroxyl group in the β-configuration, resulted in less efficient CTB binding compared to l-fucose or α-locked fucose. Finally, the SAR study showed increased binding of α1,2-linked fucose compared to α1,3-linked fucose³³.

All crystal structures of CTB with Le^x (6HJD; this work) and related molecules, such as Le^y or A-Le^y (5ELB²¹ and 5ELD²¹), show that fucose OH3 can form one or two hydrogen bonds to His94 and to a conserved water molecule (not shown in Fig. 2), explaining why removing OH3 in the SAR study caused reduced CTB binding. OH4 was reported to contribute the strongest to CTB binding³³, which is in good agreement with the fact that OH4 forms two hydrogen bonds to backbone atoms of residues 47 and 94 (Fig. 2d). The crystal structures also identify a match for the proposed key hydrophobic residues interacting with C6 in Phe48, possibly together with Ala46 (Fig. 2d).

CTB binds l-fucose with its free anomeric hydroxyl group in the β-configuration (Fig. 3b), thus free fucose is not limited to bind in the α-configuration. However, locking the fucose in the β-configuration by the addition of a methyl group would cause steric clashes with residues Thr47 and Gly45, explaining why α-linked fucose bound stronger to CTB than β-linked fucose³³.

According to Wands et al., OH2 does not contribute significantly as a hydrogen bond acceptor³³, whereas we observe a hydrogen bond to Gln3# from the adjacent B-subunit (Fig. 2d). However, their conclusion was based on a compound with a fluorine replacing the OH2 group (rather than with OH2 removed), and it is still under debate if fluorine can form hydrogen bonds^51,52.

Finally, Wands et al. found increased inhibition of CTB binding by fucosyllactose with α1,2-linked compared to α1,3-linked fucose residues³³. The latter showed no difference compared to l-fucose alone. The crystal structure of CTB in complex with Le^y (5ELB²¹) clearly shows that both α1,2-linked and α1,3-linked fucose can bind to CT. However, the preferred orientation appears to be highly influenced by the precise details of ligands and toxins. For example, the structure of a CTB-LTB chimera in complex with a human milk oligosaccharide similar to A-Le^y exhibited exclusive binding of Fucα2 in the fucose binding pocket of the toxin²⁵.

In conclusion, the structural data presented here are a very good match to the reported structure–activity relationship data, pointing to the same binding site identified in both studies, i.e. the secondary binding site of the toxin.

Roles of GM1 and fucosylated receptors

A large body of research points to GM1 and fucosylated structures facilitating toxin binding and uptake^{12,16–20,30,32–34}, however, little is known about the exact localization and expression of specific fucosylated structures on healthy intestinal cells and tissues^32,53–57. It is not clear which fucosylated structures are the major cellular attachment sites besides GM1. However, all evidence points towards glycoproteins with type-2 core structures similar to Le^x, Le^y and related HBGAs^{21,27–30,32,33}. In particular, Le^x has been shown to allow CTB binding to intestinal cells³².

Whether the binding of GM1 and fucosylated structures happens simultaneously or sequentially is also unknown, nor what triggers or causes toxin uptake. Recent crystal structures from our lab suggest that CTB can bind sugars at both binding sites at the same time, specifically Le^y tetraose and galactose²¹. However, it has not yet been shown if this also holds true for GM1os. A recently published hetero-multivalent binding model suggests that CTB first binds to a high-affinity ligand such as GM1. Subsequent binding occurs much more readily (up to 10,000-fold faster) also for low-affinity ligands like GM2, since binding is then confined to the 2D membrane surface⁵⁸. In fact, cooperativity was found to be enhanced for a heterogeneous mixture of ligands⁵⁸. In a biological context, it seems more likely that the toxins first bind to the broadly available fucosylated structures on the cell surface or mucus layer until they can bind to the few available GM1 receptors, preventing the flow in the small intestine to remove unbound toxins from the intoxication site²¹. In addition, V. cholerae has other virulence factors that may influence receptor distribution and availability, e.g. neuraminidase VcN⁵⁹. It has been shown recently that VcN can remodel intestinal polysialylated gangliosides into GM1 to provide an increased numbers of cellular receptors for CT⁵⁹, and it is well conceivable that fucosylated glycoconjugates serve as intermediate anchoring points until VcN unveils enough GM1 receptors. Indeed, several studies showed that the exogenous addition of GM1 to cells markedly increases CT binding and intoxication^12,34,60,61. Remarkably, a new study by Sethi et al. indicates that fucosylated glycoconjugates play a functional role in the process, in the presence or absence of GM1³⁴.

Correlation with epidemiological data: Why are secretors protected?

Many diseases show blood group association⁶², and cholera is the textbook example. Individuals with blood group O experience the most severe cholera symptoms^35–38. Similarly, individuals with non-secretor phenotype are at increased risk of severe cholera, whereas so-called ‘secretors’, who secrete HBGAs and display these structures in their small intestinal mucosa, enjoy some protection^63,64. Recently, progress has been made in the understanding of cholera ABO(H) blood group dependence²¹, however, it remains unclear why secretors are protected from severe cholera compared to non-secretors. The work presented here for Le^x represents the first relevant structural and affinity data for a non-secretor HBGA, bringing us in a unique position to shed light on this phenomenon.

Non-secretors have a non-functional Se (FUT2) gene, which codes for an α1,2-fucosyltransferase that is important for the generation of HBGAs on epithelial and endothelial cells and most body fluids³⁹. Non-secretors therefore display a distinct set of receptors on their intestinal cells that is enriched in Lewis^a (Le^a) and Le^x antigens, and characterized by limited fucosylation^39,55,65. Interestingly, when probing jointly for secretor status and ABO(H) blood group, only secretors with blood group A or B showed less severe cholera symptoms, whereas the disease burden for blood group O individuals was equally high independent of secretor status⁶⁴, which could be explained if the H-antigen characteristic of blood group O itself is a risk factor for severe cholera.

The H-antigen exists in two types, type-1 and type-2, with different linkages, both of which contain an α1,2-linked fucose. Addition of a second fucose residue (Fucα4 or Fucα3) then gives rise to Le^b/y determinants, respectively. If the same fucose is added to HBGA precursors instead, this yields Le^a/x epitopes, which in contrast to H-antigens and Le^b/y cannot be further modified by A or B glycosyltransferases, since they lack the α1,2-linked fucose. So far, all evidence points towards only type-2 antigens being able to bind to the CT^13,21,25, focusing the attention on Le^y and Le^x. The lower affinity of Le^x compared to the other HBGAs makes it unlikely that Le^x binding is the cause for more efficient CT uptake in non-secretors. There are only limited data regarding the expression of fucosyltransferases and the distribution of glycoconjugates in human tissues^{32,53–56,66}. Moreover, it is unknown which fucosylated CT receptors are relevant in vivo. The available data, however, suggest that compared to the villi, the deeper parts of the intestinal membranes are unaffected by the non-functional FUT2 of non-secretors, due to expression of an alternative 1,2-fucosyltransferase encoded by the H gene^65,67. Non-secretors cannot secrete soluble ABO(H) glycans, but might still present suitable docking sites (i.e., Le^y) on these deeper membrane regions, independent of their secretor status, and cooperative binding is expected to be enhanced for hetero-multivalent binding including GM1⁵⁸. Indeed, Cervin et al. observed tighter binding of CTB to the crypts than to the villi³². Due to the lack of soluble HBGAs, non-secretors would moreover lack competition by soluble and mucin-bound CT receptors, leading to more efficient CT uptake and more severe cholera symptoms.

Conclusions and perspective

The importance of fucosylated sugars for cholera intoxication has recently been in the limelight^21,27–34, but there has been limited structural information on their interaction with the CT. Here we show that fucosylated receptors, except fucosyl-GM1, bind solely to the secondary CT binding site, in contrast to recent suggestions³². We provide detailed information on how Le^x and l-fucose bind to CTB, and corroborate our findings with quantitative binding data of holotoxin variants. In addition, we present the first CT variant – H18A – with reduced affinity to fucosylated structures, but slightly increased affinity to GM1os. This not only lends further support to a possible cross-talk between the primary and secondary CT binding sites, but may also have practical value. Currently CT is widely used as a marker of GM1 and lipid rafts⁶⁸. We suggest to replace wild-type CTB with variant H18A as more specific GM1 marker, to limit the number of false-positive results caused by the interaction with secondary binding site receptors. Furthermore, CT variants H18A and H18AH94A, being devoid of a functional secondary binding site, are predicted to facilitate the analysis of cellular uptake in cellular and organoid models^31,69 by allowing discrimination of the effects caused by interaction with primary and secondary sites.

Mutagenesis

Vector pARCT5, which was kindly provided by Professor Randall K. Holmes, contains an arabinose-inducible CT operon, with signal sequences derived from the LT-IIb B gene⁷⁰. Site-directed mutagenesis was performed using the manufacturer’s protocol (Q5® Site-Directed Mutagenesis Kit, NEB) and DNA oligo nucleotides (Eurofins Genomics) shown in Table 3. Successful mutagenesis was verified by Sanger sequencing using primers Seq 1–3 (Eurofins Genomics, GATC).

Production of classical CTB, standard protocol

Expression was performed essentially as described previously²¹. Briefly, the gene for CTB (Uniprot: Q57193; classical biotype) was heterologously expressed in E. coli BL21 (DE3) using a CTB-pET21b+ construct. For protein production, cells were grown at 37 °C in LB medium containing ampicillin until OD_{600 nm} of 0.5 was reached. The temperature was reduced to 25 °C and isopropyl-β-d-thiogalactopyranoside (IPTG) was added to a final concentration of 0.5 mM to start CTB production. Cells were harvested after 14–18 h by centrifugation (6900 × g, 20 min, 4 °C) and the pellet was re-suspended in ice-cold sucrose buffer (20 mM Tris/HCl, 25% (w/v) sucrose, 5 mM EDTA at pH 8.0). After 15 min on ice, the solution was centrifuged (8000 × g, 20 min, 4 °C) and the pellet was re-suspended in periplasmic lysis buffer (5 mM MgCl₂, 150 μg/mL lysozyme, DNase). The solution was kept cold for 30 min, centrifuged (8000 × g, 20 min, 4 °C) and dialyzed for several hours against PBS in a Snakeskin tube (Thermo Scientific, 3500 MWCO).

Production of CTB and CT holotoxins for TALON chromatography

The gene for CTB (Uniprot: Q57193) was heterologously expressed in E. coli BL21 (DE3) using a CTB-pET21b +construct. For protein production, cells were grown at 37 °C in LB medium containing ampicillin until OD_{600 nm} of 0.5 was reached. The temperature was reduced to 25 °C and IPTG was added to a final concentration of 0.5 mM to start CTB production.

The genes for CT and CT variants (W88K, H18A, H18AH94A) were heterologously expressed in OverExpress™ C43 (DE3) cells (Sigma) using pARCT5 or pARCT5 derivatives. For protein production, cells were grown at 37 °C in TB medium containing chloramphenicol until OD_{600 nm} of 2.0 was reached. l-arabinose was added to a final concentration of 0.2% (w/v) to start holotoxin production.

Cells were harvested after 14–18 h (CTB) or 3 h (holotoxin) by centrifugation (6900 × g, 20 min, 4 °C) and the pellet was re-suspended in 1/60^th volume of TALON A buffer (50 mM sodium phosphate, 300 mM NaCl, pH 8) with 1 mg of polymyxin B (Sigma Aldrich) per mL, cOmplete™ Protease Inhibitor (Roche) and benzonase (EMD Millipore) and shaken at 37 °C for 15 min. Insoluble debris and cells were removed from the periplasmic extracts by centrifugation (8000 × g, 20 min, 4 °C). The filtered supernatant was directly applied to the TALON affinity column.

Purification of CTB using D-galactose affinity chromatography

The protein solution was loaded onto a d-galactose-sepharose affinity column (Thermo Scientific) and eluted using 300 mM galactose in PBS. Fractions containing pure CTB were pooled and concentrated using Vivaspin 20 mL concentrator tubes (5000 MWCO, PES membrane, Sartorius). The protein was subjected to size-exclusion chromatography using PBS and a Superdex 75 column mounted on an ÄKTA purifier (GE Healthcare). Fractions with toxin were pooled, dialyzed over night against Tris buffer (20 mM Tris/HCl, 200 mM NaCl, pH 7.5), concentrated to 3–9 mg/mL, snap-frozen in liquid nitrogen and stored at −80 °C.

Purification of CTB and CT holotoxins using TALON affinity chromatography

As alternative to the d-galactose-sepharose affinity purification (to preclude contamination with galactose), the periplasmic extract of CTB or CT (wild-type and variants of classical biotype) was loaded onto a pre-equilibrated HiTrap TALON crude column (GE Healthcare) and eluted with TALON B buffer (50 mM sodium phosphate, 300 mM NaCl, 50 mM imidazole at pH 8) in order to avoid residual D-galactose bound to the protein. The protein solution was concentrated using Vivaspin concentrator tubes (10000 MWCO, PES membrane, Sartorius) and loaded onto a Superdex 75 (CTB) or Superdex 200 column (holotoxins) mounted on an ÄKTA purifier (GE Healthcare) equilibrated with Tris buffer (20 mM Tris/HCl, 200 mM NaCl at pH 7.5) or PBS. Fractions with CTB pentamer (Tris buffer) were pooled, concentrated to 3 mg/mL, snap-frozen in liquid nitrogen and stored at −80 °C. Fractions with holotoxins (PBS) were pooled and stored at 4 °C.

CD spectroscopy and MS

Correct folding of all CT variants was verified by CD spectroscopy (J-810 CD spectrometer, JASCO). Prior to the measurement, buffer exchange to 10 mM sodium phosphate buffer pH 7.5 was performed using Vivaspin concentrator tubes. All CT variants were characterized by SDS-PAGE analysis, tryptic digestion and mass spectrometry (performed in house). The amino acid substitution H94A was only confirmed indirectly by DNA sequencing, since tryptic digestion and MS did not yield a suitable C-terminal peptide.

Co-crystallization of CTB complexes

CTB (Gal-affinity) and Le^x triaose (Gly049, Elicityl) or CTB (TALON) and fucosyl-GM1os (GLY103, Elicityl) were mixed at a molar ratio of 1:10 (B-subunit to ligand). CTB (TALON) and l-fucose (F2252, Sigma Aldrich) were mixed at a molar ratio of 1:500 (B-subunit to ligand), corresponding to a final concentration of 0.14 M of l-fucose, for a protein concentration of 3.3 mg/ml. All samples were incubated at RT for 2 h. Initially, crystals were obtained by sitting-drop vapour-diffusion using a crystallization robot (Oryx4, Douglas Instruments) at 20 °C in the Morpheus screen⁷¹ condition A12 (Molecular Dimensions). For CTB and Le^x triaose another crystal form was obtained in Morpheus screen condition A4. Crystals from condition A12 were optimized for all ligands by several rounds of hanging-drop vapour-diffusion experiments. Final crystals for the CTB complex with Le^x were obtained by mixing protein-ligand solution (1 µL, 3.3 mg/mL) 1∶1 with crystallization buffer (0.1 M Tris (base)/BICINE pH 8.5, 8% PEG1000, 8% PEG3350, 8% MPD, 0.03 M MgCl_2, 0.03 M CaCl₂) and 0.2 µL seed solution (microseeding, Seed bead kit, Hampton Research). CTB l fucose co-crystals were obtained by mixing protein-ligand solution (1 µL, 3 mg/mL) 1:1 with crystallization buffer (0.1 M Tris (base)/BICINE pH 8.7, 10% PEG1000, 10% PEG3350, 10% MPD, 0.03 M MgCl₂, 0.03 M CaCl₂). Final crystals for the CTB complex with fucosyl-GM1os were obtained by mixing protein-ligand solution (1.5 µL, 1.5 mg/mL) 1:1 with crystallization buffer (0.1 M Tris (base)/BICINE at pH 8.7, 6% PEG1000, 6% PEG3350, 6% MPD, 0.03 M MgCl₂, 0.03 M CaCl₂) and cryoprotection was achieved by transferring this crystal to crystallization buffer with higher PEG and MPD concentrations (0.1 M Tris (base)/BICINE pH 8.7, 10% PEG1000, 10% PEG3350, 10% MPD, 0.03 M MgCl₂, 0.03 M CaCl₂. Crystals were harvested, mounted in loops, and flash-cooled in liquid nitrogen (CTB l fucose, CTB fucosyl GM1os) or a nitrogen cryo-stream (CTB-Le^x), respectively.

X-ray data collection and refinement

Synchrotron data collection for the CTB complex with Le^x was performed at ID23–1, ESRF, Grenoble, France (100 K, 0.9770 Å). Data collection for CTB complexes with l-fucose or fucosyl-GM1os was performed at BioMAX, Max IV, Lund, Sweden (100 K, 0.9795 Å and 0.9808 Å, respectively). Data were processed with XDS⁷² and NEGGIA/DECTRIS (CTB l fucose, CTB fucosyl GM1os) or xia2 and DIALS (CTB-Le^x)^73,74 and AIMLESS from the CCP4 software suite^{75,76 (Table 1)}. Diffraction data cut-offs were chosen based on the assessment of the anisotropic CC_1/2 and the quality of the electron density. Both crystal forms of CTB-Le^x belong to space group P2₁2₁2₁ and contain two B-pentamers in the asymmetric unit, but have different cell parameters. For the CTB-Le^x model, we used the highest resolution data set (1.5 Å) from an optimized Morpheus A12 condition, which crystallized under similar conditions, and with the same space group and pentamer arrangement (“top-to-top”) as the CTB complex with Le^y tetraose reported previously (5ELB²¹). CTB l fucose and CTB fucosyl GM1os crystals belong to space group C2 and contain two B-pentamers in the asymmetric unit. The structures were solved by molecular replacement using Phaser⁷⁷ from the CCP4 software suite^75,76 and search model 5ELB²¹, from which ligands and water molecules had been manually removed. To avoid potential model bias, five cycles of refinement including two cycles with simulated annealing (starting temperature of 5000 K) were carried out with the Phenix software suite⁷⁸. The final model was obtained after several cycles of manual building with Coot⁷⁹, followed by refinement with REFMAC5⁸⁰. Initial refinement steps involved local NCS restraints, while final refinement steps involved TLS parameterization (REFMAC5, automatic, 5 cycles)⁸¹. Water molecules were placed using COOT:Find_waters and then manually inspected for several criteria, including distances from hydrogen-bond donors/acceptors and quality of the electron-density. Most of the disulfide bridges are partly reduced, due to minor radiation damage. Le^x triaose, GM1os and fucosyl-GM1os were built using MAKE LIGAND (AceDRG)⁸² from the CCP4 software suite^75,76 and isomeric SMILES strings. The restraints for a Thr-Fuc bond were generated using JLigand⁸³. Le^x triaose, fucosyl-GM1os and α-l-fucose or β-l-fucose were included last to avoid model bias. To improve the density for the terminal fucose residue, GM1os was included prior to fucosyl-GM1os. For the CTB complex with fucose, additional elongated electron density was found in two of the primary binding sites, however, the origin of the density could not be identified, even with Polder⁸⁴ maps calculated with the Phenix software suite⁷⁸ (the density was clearly not compatible with a sugar ring). PDB_REDO⁸⁵ was used to evaluate the models before final refinement steps. Occupancies were refined by evaluating the difference Fourier maps and by comparing the B-factors of the ligands with interacting protein atoms (exception: fucose residues in the CTB-fucosyl-GM1os secondary sites, which were modelled at full occupancy). The final models were analysed using the Analyse geometry task of the CCP4 software suite^75,76. The percentages of amino acid residues occupying the favoured, allowed and outlier regions in the Ramachandran plot are 97.5/2.5/0.0% for CTB-Le^x, 97.4/2.4/0.2% for CTB l fucose, and 97.7/2.3/0.0% for CTB-fucosyl-GM1os, respectively. Figures were generated using PyMol (Schrödinger LLC), α-helices and β-strands were assigned using STRIDE⁸⁶.

Surface plasmon resonance spectroscopy

SPR analyses were performed using Series S CM5 sensor chips and a Biacore T100 biosensor system (Biacore Life Sciences, GE Healthcare). Due to the high cost of the oligosaccharides combined with the relatively low affinity typical for carbohydrate-protein interactions, experiments were performed in duplicates or triplicates. Proteins were immobilized on the surface of the sensor chip to yield a signal of approximately 4200 response units, as described previously²¹. Le^x tetraose (GLY050, Elicityl) and triaose were dissolved in the running buffer (10 mM HEPES/NaOH, 140 mM NaCl, 3 mM EDTA, 0.005% P20, pH 7.4) and injected in concentrations from 78.125 μM up to 40 mM for 45 s. Dissociation was monitored for additional 15 s. The regeneration was not needed for Le^x tetraose, whereas for Le^x triaose, a short pulse (6 s) of 3 mM NaOH was injected to completely remove it from the surface. Binding of 3.125–200 nM GM1os (GM1a, GLY096, Elicityl) was monitored for 240 s, with additional 420 s for dissociation. The regeneration of the chip was the same as for Le^x triaose. All binding steps were performed at 30 μL/min and 25 °C. The data were evaluated using Biacore T100 Evaluation software. For GM1os, data were fitted to the Langmuir 1:1 interaction model, since responses did not reach the equilibrium during the association phase. Other binding affinities were determined from plots of steady-state analyte binding levels against the concentration.

• 1.Sack DA, Sack RB, Nair GB, Siddique AK. Cholera. Lancet. 2004;363:223–233. doi: 10.1016/S0140-6736(03)15328-7. [DOI] [PubMed] [Google Scholar]
• 2.WHO. Ending cholera: a global roadmap to 2030, http://www.who.int/cholera/publications/global-roadmap-summary.pdf (Date of access: 22/11/2018) (2018).
• 3.The Lancet Gastroenterology & Hepatology Health catastrophe: the toll of cholera in Yemen. The Lancet Gastroenterology & Hepatology. 2017;2(9):619. doi: 10.1016/S2468-1253(17)30224-8. [DOI] [PubMed] [Google Scholar]
• 4.WHO. Yemen, Annual report 2017, https://reliefweb.int/report/yemen/world-health-organization-yemen-annual-report-2017 (Date of access: 29/09/2018) (2018).
• 5.Sanchez J, Holmgren J. Cholera toxin - a foe & a friend. Indian J Med Res. 2011;133:153–163. [PMC free article] [PubMed] [Google Scholar]
• 6.Heggelund, J. E., Bjørnestad, V. A. & Krengel, U. Vibrio cholerae and Escherichia coli heat-labile enterotoxins and beyond. In The Comprehensive Sourcebook of Bacterial Protein Toxins (eds Alouf, J., Landant, D. & Popoff, M. R.) 195–229 (Elsevier, 2015).
• 7.Merritt EA, Hol WGJ. AB5 toxins. Curr Opin Struct Biol. 1995;5:165–171. doi: 10.1016/0959-440X(95)80071-9. [DOI] [PubMed] [Google Scholar]
• 8.Chinnapen DJ, Chinnapen H, Saslowsky D, Lencer WI. Rafting with cholera toxin: endocytosis and trafficking from plasma membrane to ER. FEMS Microbiol Lett. 2007;266:129–137. doi: 10.1111/j.1574-6968.2006.00545.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 9.Gill DM, Meren R. ADP-ribosylation of membrane proteins catalyzed by cholera toxin: basis of the activation of adenylate cyclase. Proc Natl Acad Sci USA. 1978;75:3050–3054. doi: 10.1073/pnas.75.7.3050. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 10.Field M, Rao MC, Chang EB. Intestinal electrolyte transport and diarrheal disease (1) N Engl J Med. 1989;321:800–806. doi: 10.1056/NEJM198909213211206. [DOI] [PubMed] [Google Scholar]
• 11.Holmgren J. Comparison of the tissue receptors for Vibrio cholerae and Escherichia coli enterotoxins by means of gangliosides and natural cholera toxoid. Infect Immun. 1973;8:851–859. doi: 10.1128/iai.8.6.851-859.1973. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 12.Holmgren J, Lönnroth I, Mansson J, Svennerholm L. Interaction of cholera toxin and membrane GM1 ganglioside of small intestine. Proc Natl Acad Sci USA. 1975;72:2520–2524. doi: 10.1073/pnas.72.7.2520. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 13.Ångström J, et al. Novel carbohydrate binding site recognizing blood group A and B determinants in a hybrid of cholera toxin and Escherichia coli heat-labile enterotoxin B-subunits. J Biol Chem. 2000;275:3231–3238. doi: 10.1074/jbc.275.5.3231. [DOI] [PubMed] [Google Scholar]
• 14.Turnbull WB, Precious BL, Homans SW. Dissecting the cholera toxin-ganglioside GM1 interaction by isothermal titration calorimetry. J Am Chem Soc. 2004;126:1047–1054. doi: 10.1021/ja0378207. [DOI] [PubMed] [Google Scholar]
• 15.Merritt EA, et al. Crystal structure of cholera toxin B-pentamer bound to receptor GM1 pentasaccharide. Protein Sci. 1994;3:166–175. doi: 10.1002/pro.5560030202. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 16.Fishman PH, Moss J, Vaughan M. Uptake and metabolism of gangliosides in transformed mouse fibroblasts. Relationship of ganglioside structure to choleragen response. J Biol Chem. 1976;251:4490–4494. [PubMed] [Google Scholar]
• 17.Jobling MG, Holmes RK. Analysis of structure and function of the B subunit of cholera toxin by the use of site-directed mutagenesis. Mol Microbiol. 1991;5:1755–1767. doi: 10.1111/j.1365-2958.1991.tb01925.x. [DOI] [PubMed] [Google Scholar]
• 18.Wolf AA, et al. Ganglioside structure dictates signal transduction by cholera toxin and association with caveolae-like membrane domains in polarized epithelia. J Cell Biol. 1998;141:917–927. doi: 10.1083/jcb.141.4.917. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 19.Badizadegan K, et al. Floating cholera toxin into epithelial cells: functional association with caveolae-like detergent-insoluble membrane microdomains. Int J Med Microbiol. 2000;290:403–408. doi: 10.1016/S1438-4221(00)80052-1. [DOI] [PubMed] [Google Scholar]
• 20.Fujinaga Y, et al. Gangliosides that associate with lipid rafts mediate transport of cholera and related toxins from the plasma membrane to endoplasmic reticulm. Mol Biol Cell. 2003;14:4783–4793. doi: 10.1091/mbc.e03-06-0354. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 21.Heggelund JE, et al. High-resolution crystal structures elucidate the molecular basis of cholera blood group dependence. PLoS Pathog. 2016;12:e1005567. doi: 10.1371/journal.ppat.1005567. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 22.Morita A, Tsao D, Kim YS. Identification of cholera toxin binding glycoproteins in rat intestinal microvillus membranes. J Biol Chem. 1980;255:2549–2553. [PubMed] [Google Scholar]
• 23.Monferran CG, Roth GA, Cumar FA. Inhibition of cholera toxin binding to membrane receptors by pig gastric mucin-derived glycopeptides: differential effect depending on the ABO blood group antigenic determinants. Infect Immun. 1990;58:3966–3972. doi: 10.1128/iai.58.12.3966-3972.1990. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 24.Balanzino LE, Barra JL, Galvan EM, Roth GA, Monferran CG. Interaction of cholera toxin and Escherichia coli heat-labile enterotoxin with glycoconjugates from rabbit intestinal brush border membranes: relationship with ABH blood group determinants. Mol Cell Biochem. 1999;194:53–62. doi: 10.1023/A:1006971913175. [DOI] [PubMed] [Google Scholar]
• 25.Holmner Åsa, Lebens Michael, Teneberg Susann, Ångström Jonas, Ökvist Mats, Krengel Ute. Novel Binding Site Identified in a Hybrid between Cholera Toxin and Heat-Labile Enterotoxin. Structure. 2004;12(9):1655–1667. doi: 10.1016/j.str.2004.06.022. [DOI] [PubMed] [Google Scholar]
• 26.Holmner Å, Askarieh G, Ökvist M, Krengel U. Blood group antigen recognition by Escherichia coli heat-labile enterotoxin. J Mol Biol. 2007;371:754–764. doi: 10.1016/j.jmb.2007.05.064. [DOI] [PubMed] [Google Scholar]
• 27.Heggelund JE, et al. Both El Tor and classical cholera toxin bind blood group determinants. Biochem Biophys Res Commun. 2012;418:731–735. doi: 10.1016/j.bbrc.2012.01.089. [DOI] [PubMed] [Google Scholar]
• 28.Mandal PK, et al. Towards a structural basis for the relationship between blood group and the severity of El Tor cholera. Angew Chem Int Ed Engl. 2012;51:5143–5146. doi: 10.1002/anie.201109068. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 29.Vasile F, et al. Comprehensive analysis of blood group antigen binding to classical and El Tor cholera toxin B-pentamers by NMR. Glycobiology. 2014;24:766–778. doi: 10.1093/glycob/cwu040. [DOI] [PubMed] [Google Scholar]
• 30.Wands AM, et al. Fucosylation and protein glycosylation create functional receptors for cholera toxin. Elife. 2015;4:e09545. doi: 10.7554/eLife.09545. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 31.Kuhlmann FM, et al. Blood group O-dependent cellular responses to cholera toxin: parallel clinical and epidemiological links to severe cholera. Am J Trop Med Hyg. 2016;95:440–443. doi: 10.4269/ajtmh.16-0161. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 32.Cervin J, et al. GM1 ganglioside-independent intoxication by cholera toxin. PLoS Pathog. 2018;14:e1006862. doi: 10.1371/journal.ppat.1006862. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 33.Wands Amberlyn M., Cervin Jakob, Huang He, Zhang Ye, Youn Gyusaang, Brautigam Chad A., Matson Dzebo Maria, Björklund Per, Wallenius Ville, Bright Danielle K., Bennett Clay S., Wittung-Stafshede Pernilla, Sampson Nicole S., Yrlid Ulf, Kohler Jennifer J. Fucosylated Molecules Competitively Interfere with Cholera Toxin Binding to Host Cells. ACS Infectious Diseases. 2018;4(5):758–770. doi: 10.1021/acsinfecdis.7b00085. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 34.Sethi A, et al. Cell type and receptor identity regulate cholera toxin subunit B (CTB) internalization. Interface Focus. 2019;9:20180076. doi: 10.1098/rsfs.2018.0076. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 35.Glass RI, et al. Predisposition for cholera of individuals with O blood group. Possible evolutionary significance. Am J Epidemiol. 1985;121:791–796. doi: 10.1093/oxfordjournals.aje.a114050. [DOI] [PubMed] [Google Scholar]
• 36.Swerdlow DL, et al. Severe life-threatening cholera associated with blood group O in Peru: implications for the Latin American epidemic. J Infect Dis. 1994;170:468–472. doi: 10.1093/infdis/170.2.468. [DOI] [PubMed] [Google Scholar]
• 37.Harris JB, et al. Blood group, immunity, and risk of infection with Vibrio cholerae in an area of endemicity. Infect Immun. 2005;73:7422–7427. doi: 10.1128/IAI.73.11.7422-7427.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 38.Harris JB, et al. Susceptibility to Vibrio cholerae infection in a cohort of household contacts of patients with cholera in Bangladesh. PLoS Negl Trop Dis. 2008;2:e221. doi: 10.1371/journal.pntd.0000221. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 39.Henry S, Oriol R, Samuelsson B. Lewis histo-blood group system and associated secretory phenotypes. Vox Sang. 1995;69:166–182. doi: 10.1111/j.1423-0410.1995.tb02591.x. [DOI] [PubMed] [Google Scholar]
• 40.Breimer ME, Hansson GC, Karlsson KA, Larson G, Leffler H. Glycosphingolipid composition of epithelial cells isolated along the villus axis of small intestine of a single human individual. Glycobiology. 2012;22:1721–1730. doi: 10.1093/glycob/cws115. [DOI] [PubMed] [Google Scholar]
• 41.Dertzbaugh MT, Cox LM. The affinity of cholera toxin for Ni2+ ion. Protein Eng. 1998;11:577–581. doi: 10.1093/protein/11.7.577. [DOI] [PubMed] [Google Scholar]
• 42.Furth AJ. Methods for assaying nonenzymatic glycosylation. Anal Biochem. 1988;175:347–360. doi: 10.1016/0003-2697(88)90558-1. [DOI] [PubMed] [Google Scholar]
• 43.Masserini M, Freire E, Palestini P, Calappi E, Tettamanti G. Fuc-GM1 ganglioside mimics the receptor function of GM1 for cholera toxin. Biochemistry. 1992;31:2422–2426. doi: 10.1021/bi00123a030. [DOI] [PubMed] [Google Scholar]
• 44.O’Neal CJ, Amaya EI, Jobling MG, Holmes RK, Hol WG. Crystal structures of an intrinsically active cholera toxin mutant yield insight into the toxin activation mechanism. Biochemistry. 2004;43:3772–3782. doi: 10.1021/bi0360152. [DOI] [PubMed] [Google Scholar]
• 45.Mertz JA, McCann JA, Picking WD. Fluorescence analysis of galactose, lactose, and fucose interaction with the cholera toxin B subunit. Biochem Biophys Res Commun. 1996;226:140–144. doi: 10.1006/bbrc.1996.1323. [DOI] [PubMed] [Google Scholar]
• 46.Galvan EM, Diema CD, Roth GA, Monferran CG. Ability of blood group A-active glycosphingolipids to act as Escherichia coli heat-labile enterotoxin receptors in HT-29 cells. J Infect Dis. 2004;189:1556–1564. doi: 10.1086/383349. [DOI] [PubMed] [Google Scholar]
• 47.Galvan EM, Roth GA, Monferran CG. Functional interaction of Escherichia coli heat-labile enterotoxin with blood group A-active glycoconjugates from differentiated HT29 cells. FEBS J. 2006;273:3444–3453. doi: 10.1111/j.1742-4658.2006.05368.x. [DOI] [PubMed] [Google Scholar]
• 48.Merritt EA, et al. Structural studies of receptor binding by cholera toxin mutants. Protein Sci. 1997;6:1516–1528. doi: 10.1002/pro.5560060716. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 49.Hatlem D, Heggelund JE, Burschowsky D, Krengel U, Kristiansen PE. 1H, 13C, 15N backbone assignment of the human heat-labile enterotoxin B-pentamer and chemical shift mapping of neolactotetraose binding. Biomol NMR Assign. 2017;11:99–104. doi: 10.1007/s12104-017-9728-9. [DOI] [PubMed] [Google Scholar]
• 50.Heggelund Julie, Heim Joel, Bajc Gregor, Hodnik Vesna, Anderluh Gregor, Krengel Ute. Specificity of Escherichia coli Heat-Labile Enterotoxin Investigated by Single-Site Mutagenesis and Crystallography. International Journal of Molecular Sciences. 2019;20(3):703. doi: 10.3390/ijms20030703. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 51.Howard JAK, Hoy VJ, O’Hagan D, Smith GT. How good is fluorine as a hydrogen bond acceptor? Tetrahedron. 1996;52:12613–12622. doi: 10.1016/0040-4020(96)00749-1. [DOI] [Google Scholar]
• 52.Dalvit C, Invernizzi C, Vulpetti A. Fluorine as a hydrogen-bond acceptor: experimental evidence and computational calculations. Chem. Eur. J. 2014;20:11058–11068. doi: 10.1002/chem.201402858. [DOI] [PubMed] [Google Scholar]
• 53.Björk S, Breimer ME, Hansson GC, Karlsson KA, Leffler H. Structures of blood group glycosphingolipids of human small intestine. A relation between the expression of fucolipids of epithelial cells and the ABO, Le and Se phenotype of the donor. J Biol Chem. 1987;262:6758–6765. [PubMed] [Google Scholar]
• 54.Finne J, et al. Novel polyfucosylated N-linked glycopeptides with blood group A, H, X, and Y determinants from human small intestinal epithelial cells. J Biol Chem. 1989;264:5720–5735. [PubMed] [Google Scholar]
• 55.Ravn V, Dabelsteen E. Tissue distribution of histo-blood group antigens. APMIS. 2000;108:1–28. doi: 10.1034/j.1600-0463.2000.d01-1.x. [DOI] [PubMed] [Google Scholar]
• 56.Henry SM. Molecular diversity in the biosynthesis of GI tract glycoconjugates. A blood-group-related chart of microorganism receptors. Transfus Clin Biol. 2001;8:226–230. doi: 10.1016/S1246-7820(01)00112-4. [DOI] [PubMed] [Google Scholar]
• 57.Robbe C, et al. Evidence of regio-specific glycosylation in human intestinal mucins: presence of an acidic gradient along the intestinal tract. J Biol Chem. 2003;278:46337–46348. doi: 10.1074/jbc.M302529200. [DOI] [PubMed] [Google Scholar]
• 58.Krishnan P, et al. Hetero-multivalent binding of cholera toxin subunit B with glycolipid mixtures. Colloids Surf B Biointerfaces. 2017;160:281–288. doi: 10.1016/j.colsurfb.2017.09.035. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 59.Alisson-Silva F, et al. Human evolutionary loss of epithelial Neu5Gc expression and species-specific susceptibility to cholera. PLoS Pathog. 2018;14:e1007133. doi: 10.1371/journal.ppat.1007133. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 60.Cuatrecasas P. Gangliosides and membrane receptors for cholera toxin. Biochemistry. 1973;12:3558–3566. doi: 10.1021/bi00742a032. [DOI] [PubMed] [Google Scholar]
• 61.Moss J, Fishman PH, Manganiello VC, Vaughan M, Brady RO. Functional incorporation of ganglioside into intact cells: induction of choleragen responsiveness. Proc Natl Acad Sci USA. 1976;73:1034–1037. doi: 10.1073/pnas.73.4.1034. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 62.Heggelund JE, Varrot A, Imberty A, Krengel U. Histo-blood group antigens as mediators of infections. Curr Opin Struct Biol. 2017;44:190–200. doi: 10.1016/j.sbi.2017.04.001. [DOI] [PubMed] [Google Scholar]
• 63.Chaudhuri A, DasAdhikary CR. Possible role of blood-group secretory substances in the aetiology of cholera. Trans R Soc Trop Med Hyg. 1978;72:664–665. doi: 10.1016/0035-9203(78)90031-7. [DOI] [PubMed] [Google Scholar]
• 64.Arifuzzaman M, et al. Individuals with Le(a+b−) blood group have increased susceptibility to symptomatic Vibrio cholerae O1 infection. PLoS Negl Trop Dis. 2011;5:e1413. doi: 10.1371/journal.pntd.0001413. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 65.Glynn LE, Holborow EJ, Johnson GD. The distribution of blood-group substances in human gastric and duodenal mucosa. Lancet. 1957;273:1083–1088. doi: 10.1016/S0140-6736(57)90115-0. [DOI] [PubMed] [Google Scholar]
• 66.Torrado J, Blasco E, Cosme A, Gutierrez-Hoyos A, Arenas JI. Expression of type 1 and type 2 blood group-related antigens in normal and neoplastic gastric mucosa. Am J Clin Pathol. 1989;91:249–254. doi: 10.1093/ajcp/91.3.249. [DOI] [PubMed] [Google Scholar]
• 67.Mollicone R, Bara J, Le Pendu J, Oriol R. Immunohistologic pattern of type 1 (Lea, Leb) and type 2 (X, Y, H) blood group-related antigens in the human pyloric and duodenal mucosae. Lab Invest. 1985;53:219–227. [PubMed] [Google Scholar]
• 68.Chiricozzi Elena, Mauri Laura, Ciampa Maria Grazia, Prinetti Alessandro, Sonnino Sandro. On the use of cholera toxin. Glycoconjugate Journal. 2018;35(2):161–163. doi: 10.1007/s10719-018-9818-7. [DOI] [PubMed] [Google Scholar]
• 69.Zomer-van Ommen DD, et al. Functional characterization of cholera toxin inhibitors using human intestinal organoids. J Med Chem. 2016;59:6968–6972. doi: 10.1021/acs.jmedchem.6b00770. [DOI] [PubMed] [Google Scholar]
• 70.Jobling MG, Palmer LM, Erbe JL, Holmes RK. Construction and characterization of versatile cloning vectors for efficient delivery of native foreign proteins to the periplasm of Escherichia coli. Plasmid. 1997;38:158–173. doi: 10.1006/plas.1997.1309. [DOI] [PubMed] [Google Scholar]
• 71.Gorrec F. The MORPHEUS protein crystallization screen. J Appl Crystallogr. 2009;42:1035–1042. doi: 10.1107/S0021889809042022. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 72.Kabsch W. XDS. Acta Crystallogr D Biol Crystallogr. 2010;66:125–132. doi: 10.1107/S0907444909047337. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 73.Parkhurst JM, et al. Robust background modelling in DIALS. J Appl Crystallogr. 2016;49:1912–1921. doi: 10.1107/S1600576716013595. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 74.Winter G, et al. DIALS: implementation and evaluation of a new integration package. Acta Crystallogr D Struct Biol. 2018;74:85–97. doi: 10.1107/S2059798317017235. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 75.Collaborative Computational Project, N The CCP4 suite: programs for protein crystallography. Acta Crystallogr D Biol Crystallogr. 1994;50:760–763. doi: 10.1107/S0907444994003112. [DOI] [PubMed] [Google Scholar]
• 76.Potterton L, et al. CCP4i2: the new graphical user interface to the CCP4 program suite. Acta Crystallogr D Struct Biol. 2018;74:68–84. doi: 10.1107/S2059798317016035. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 77.McCoy AJ, et al. Phaser crystallographic software. J Appl Crystallogr. 2007;40:658–674. doi: 10.1107/S0021889807021206. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 78.Adams PD, et al. The Phenix software for automated determination of macromolecular structures. Methods. 2011;55:94–106. doi: 10.1016/j.ymeth.2011.07.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 79.Emsley P, Lohkamp B, Scott WG, Cowtan K. Features and development of Coot. Acta Crystallogr D Biol Crystallogr. 2010;66:486–501. doi: 10.1107/S0907444910007493. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 80.Murshudov GN, et al. REFMAC5 for the refinement of macromolecular crystal structures. Acta Crystallogr D Biol Crystallogr. 2011;67:355–367. doi: 10.1107/S0907444911001314. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 81.Winn MD, Isupov MN, Murshudov GN. Use of TLS parameters to model anisotropic displacements in macromolecular refinement. Acta Crystallogr D Biol Crystallogr. 2001;57:122–133. doi: 10.1107/S0907444900014736. [DOI] [PubMed] [Google Scholar]
• 82.Long F, et al. AceDRG: a stereochemical description generator for ligands. Acta Crystallogr D Struct Biol. 2017;73:112–122. doi: 10.1107/S2059798317000067. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 83.Lebedev AA, et al. JLigand: a graphical tool for the CCP4 template-restraint library. Acta Crystallogr D Biol Crystallogr. 2012;68:431–440. doi: 10.1107/S090744491200251X. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 84.Liebschner D, et al. Polder maps: improving OMIT maps by excluding bulk solvent. Acta Crystallogr D Struct Biol. 2017;73:148–157. doi: 10.1107/S2059798316018210. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 85.Joosten RP, Long F, Murshudov GN, Perrakis A. The PDB_REDO server for macromolecular structure model optimization. IUCrJ. 2014;1:213–220. doi: 10.1107/S2052252514009324. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 86.Heinig M, Frishman D. STRIDE: a web server for secondary structure assignment from known atomic coordinates of proteins. Nucleic Acids Res. 2004;32:W500–502. doi: 10.1093/nar/gkh429. [DOI] [PMC free article] [PubMed] [Google Scholar]

Supplementary Materials

Data Availability Statement

The coordinates and structure factors have been deposited with the Protein Data Bank with accession codes: 6HJD, 6HMW and 6HMY.