Differential irradiation DEEP-STD NMR experiments

We first tested the protocol by analysing the effect of different irradiation frequencies on the binding of 2,7-anhydro-Neu5Ac to GH3317_{. We ran two STD NMR experiments}

irradiating (0.5 s) at 0.60 ppm (exp1) and 6.55 ppm (exp2). These frequencies are known to be centred in the aliphatic and aromatic protein proton spectral regions, respectively100_{. It is worth noticing that the selection of frequencies to irradiate different}

types of protein protons can be based on either known NMR spectral properties of the protein (if chemical shifts are assigned), on NMR databases (e.g., BMRB101_{), or on}

predictions if a 3D model of the protein is available (as in this study, Table 3.3101_).

The DEEP-STD factors (0.60 ppm/6.55 ppm) are shown in Figure 3.7a. Positive DEEP-STD factors report relative STD increases when irradiation was at 0.60 ppm (aliphatics); negative ones indicate increases when irradiation was at 6.55 ppm (aromatics). The

94 | P a g e resulting differential epitope map is shown in Figure 3.7b. The results clearly show how different protons of the ligand occupy distinct areas of the GH33 binding pocket lined by either aliphatic or aromatic residues. The positive ΔDEEP-STD for CH3 and H3a suggest

vicinity to aliphatic side chains, whereas negative ΔDEEP-STDs for H8, H9 and H9’ suggest vicinity to aromatic protons.

The DEEP-STD NMR results shown in Figure 3.7a are in excellent agreement with the published crystal structure of the complex between 2,7-anhydro-sialic acid and GH33 (Figure 3.7c), where the ligand sits between aliphatic (Ile258, Ile338, and Val502) and opposite aromatic patches (Tyr667 and Trp698). The ligand protons CH3 and H3a point

towards the aliphatic residues, while H8, H9, and H9’ are projected towards the aromatic side chains. Protons H3e, H4, H5, H6 and H7 sit in between these two regions in agreement with their negligible ΔDEEP-STD factors.

Figure 3.7. Left panel: Differential Epitope Mapping (0.60 ppm/6.55 ppm) of 2,7-anhydro-

Neu5Ac in complex with GH33. (a) ΔDEEP-STD histogram: positive ΔDEEP-STD (above the limit of +0.75) after aliphatic irradiation (0.60 ppm) are in cyan, and negative ΔDEEP-STD (below - 0.75) after aromatic irradiation (6.55 ppm) are in magenta. (b) Differential Epitope: ΔDEEP-STD map of the ligand. Cyan surfaces highlights ligand contacts with aliphatic side chains; magenta, contacts with aromatic side chains. The ligand polar protons have been omitted. (c) Crystal structure of the complex (PDB ID: 4X4A17_{). Protein protons are coloured as a function of their}

95 | P a g e 6.55 ppm (aromatics) in magenta (Table 3.3). Right panel: differential Epitope Mapping (2.25 ppm/0.60 ppm) of 3-nitrophenyl-α-D-galactopyranoside (3NPG) in complex with Cholera Toxin Subunit B (CTB). (d) ΔDEEP-STD histogram: protons with positive ΔDEEP-STD (above the limit of +0.75) after irradiation at 2.25 ppm are in orange. (e) Differential Epitope: ΔDEEP-STD map of the ligand. Orange surfaces indicate ligand contacts with protein side chains directly irradiated at 2.25 ppm. The ligand polar protons have been omitted. (f) Crystal structure of the complex (PDB ID: 1EEI88_{). Protein protons directly irradiated at 2.25 ppm are enclosed in orange surface.}

For both systems, protons were added using Schrodinger software102_{. Comparison of (b) and (c)}

highlights the excellent match of the differential epitope map of the ligand with the distribution of the residues in the binding pocket. Figure adapted from 87.

Further, we applied the novel protocol to the complex of CTB, a larger receptor (65 kDa), with 3NPG88_{. Unfortunately, the ligand contains an aromatic moiety, which precludes}

protein irradiation in this spectral region. However, in DEEP-STD NMR, it is possible to select other groups of protein protons for irradiation, providing that they are in spectral regions devoid of ligand signals. For CTB, we targeted protein resonances at 2.25 ppm, where no ligand protons showed signals, and protein signals were available for saturation. We predicted the chemical shifts of protons of CTB within 4 Å of the ligand in the X-ray structure, and the results indicated Glu51 and Gln56 protons as the ones likely to be directly irradiated (Figure 3.7f, and Table 3.3). Experiments conducted with differential frequencies (2.25 ppm/0.60 ppm), resulted in positive ΔSTD values for protons H4, H5, H6 and H6’ on the galactose, indicating an increase in relative STDs when irradiating at 2.25 ppm (Figure 3.7d). In contrast, negligible ΔSTD factors were observed for H1, H2, H3, and the aromatic protons at the opposite end of the molecule. The differential epitope map of 3NPG (Figure 3.7e) was found to be in perfect agreement with the published crystal structure of the complex between 3NPG and CTB (Figure 3.7f),[10] _{in which the galactose ring area of H4 to H6 is surrounded by the side chains of}

Glu51 and Gln56. In contrast, H1, H2, H3 and aromatic carbons are pointing far from those side chains in the binding pocket (Figure 3.7f). These results confirm that it is possible to identify the nature of the ligand-receptor contacts by means of differential protein irradiation. The spectra and the raw and processed data are shown in Figure 3.8, Tables 3.1 and 3.2.

96 | P a g e

Figure 3.8. Left panel: 1 mM of 2,7-anhydro-Neu5Ac in the presence of 50 M GH33 in

deuterated tris-d11 buffer, 293 K. (a) Reference spectra (x 1); (b) STD NMR spectrum with on-

resonance irradiation at 6.55 ppm (x 64); (c) STD NMR spectrum with on-resonance irradiation at 0.60 ppm (x 32). Right panel: 1 mM of 3NPG in the presence of 10 M CTB (50 M binding pockets) in deuterated Phosphate Buffer Saline (PBS buffer), 278 K. (a) Reference spectra (x 1);

(b) STD NMR spectrum with on-resonance irradiation at 2.25 ppm (x 8); (c) STD NMR spectrum

with on-resonance irradiation at 0.60 ppm (x 16). Figure from 87.

It is worth mentioning here that, which ΔSTD values should be considered significant will depend on the sizes of the STD factors for the protein-ligand system under study. Based on the results of the systems analysed here, we experimentally determined that ΔSTD greater than 0.75 in magnitude were significant.

97 | P a g e Proton 1_{H } (ppm)[a] STD % 0.60 ppm STD % 6.55 ppm Ratio STD 0.60 ppm/6.55 ppm ΔDEEP- STD CH3 1.90 3.66 0.94 3.89 1.68 H3a 1.89 2.87 0.91 3.15 0.94 H3e 2.04 2.42 1.05 2.30 0.09 H4 3.82 2.84 1.38 2.06 -0.16 H5 3.79 2.58 0.94 2.74 0.53 H6 4.41 2.68 1.18 2.27 0.06 H7 4.30 2.61 1.29 2.02 -0.19 H8 3.41 1.41 1.24 1.14 -1.08 H9 3.46 1.03 0.89 1.16 -1.06 H9' 3.63 1.15 0.82 1.40 -0.81

Sum Sum STD average

23.25 10.64 2.21

Table 3.1. DEEP-STD NMR using Differential Irradiation (0.60 ppm/6.55 ppm) of the GH33/2,7-

98 | P a g e Proton 1_{H } (ppm)[a] STD % 2.25 ppm STD % 0.60 ppm Ratio STD 2.25 ppm/0.60 ppm ΔDEEP- STD H1 5.50 6.73 3.66 1.84 -0.68 H2 3.73 12.51 4.95 2.53 0.01 H3 3.81 10.81 4.72 2.29 -0.23 H4 3.75 21.82 5.9 3.70 1.18 H5 3.73 15.21 4.6 3.31 0.78 H6/H6' 3.40 19.79 5.84 3.39 0.87 H cd 7.27 2.81 1.46 1.92 -0.60 Hb 7.69 5.24 2.9 1.81 -0.71 Ha 7.75 12.21 6.38 1.91 -0.61

Sum Sum STD average

107.13 40.41 2.52

Table 3.2. DEEP-STD NMR using Differential Irradiation (2.25 ppm/0.60 ppm) CTB/3NPG. [a]

Spectra for assignment acquired at 278 K. Table from 87.

3.3.4 Selection of irradiation frequencies and identification of directly irradiated residues