**2. Results and Discussion**

#### *2.1. Hydrodynamic Behavior of the Histidine Tagged Forms Assessed by Size Exclusion Chromatography*

The secondary and tertiary structures of a protein involve non-covalent interactions. The more the amino-acid residues will interact, the more compact (globular) the protein will be. IDRs will involve fewer interactions between residues and lower compaction. Consequently, the hydration sphere of an intrinsically disordered protein is often larger than what would be expected for globular proteins of a similar molecular weight. A simple way to assess the degree of compactness of a protein in solution is to measure its hydrodynamic radius (*Rh*). A commonly used method for measuring *Rh* is size exclusion chromatography, SEC (Figure 1).

When submitted to SEC (Figure 1A), the hydrodynamic behavior of His6 eIF4E, *M* 28,550 Da (MALDI-TOF spectrometry of the purified recombinant protein), was in agreement with that of a 30 kDa globular protein. *Rh* values were deduced from experimentally determined apparent molecular weights (*Mapp*) (see Section 4). The elution profile of purified recombinant His6 VPg, *M* 26,250 Da (mass spectrometry), featured two populations, with the major one (69%) suggesting that of a 40 kDa globular protein, and the minor species (27%) averaging 90 kDa. Clearly, His6 VPg did not behave like a globular protein. The trypsin hydrolysis kinetics of His6 VPg shows a moderate proteolytic resistance profile comparable to that of α-casein, a disordered protein (Figure S1). A method was developed to classify IDPs according to the relationship between their apparent molecular density (*ρ*) and their true molecular weight (*M*) [22]. From Figure 1C,D it can be deduced that His6 VPg shares hydrodynamic features of molten and pre-molten globules. The binary complex His6 VPg-His6 eIF4E showed a more complex elution profile (Figure 1A). Its major component (peak 2) displayed a *Mapp* of 56kDa, which is close to the weight (*M* 54,772 Da) calculated by adding the two partners molecular weights. SDS-PAGE analysis showed that this elution fraction contained equal amounts of His6 VPg and His6 eIF4E indicating a possible compaction of VPg upon its association with eIF4E. The major component eluted under peak 1, *Mapp* = 110 kDa, corresponds to oligomeric forms of His6 VPg (Figure 1B). The VPg propensity to aggregate was previously described [10].

Using PONDR-VLXT, intrinsic disorder was predicted for the tagged and untagged proteins. The His6 tag potentially brings disorder to the N-terminus of all proteins. After the His6 tag removal, the first 46 amino acids segment of the native eIF4E was still predicted as unstructured. This prediction was validated by previously reported structural data showing that the first 40 amino acids of eIF4E were intrinsically disordered and fold upon binding with eIF4G [23]. In addition, upon His6 tag removal, the first 25 amino acids of VPg were predicted as disordered (Figure S2). This region was recently reported as being a conformational switch [24]. Therefore, the contribution of these regions, predicted as disordered, to the hydrodynamic properties of the proteins was assessed.

The hydrodynamic behavior of VPg, eIF4E and their binary complex was deduced from fluorescence anisotropy measurements.

**Figure 1.** Size exclusion chromatography of His6 tagged VPg, His6 tagged 4E (eIF4E) and His6 VPg-His6 eIF4E, their binary complex. (**A**) Separated runs of the purified monomers (1 mL of 0.7–1 mg/mL) were performed. Vertical dashed lines refer to the elution volumes expected for globular proteins of 28,549 Da (upper panel) and 26,137 Da (middle panel), the molecular weights of recombinant His6 eIF4E and His6 VPg respectively (MALDI-TOF spectrometry determinations). The His6 VPg-His6 eIF4E binary complex was pre-formed by mixing His6 eIF4E (0.7 mg/mL with an excess of His6 VPg (1.2 mg/mL) and loaded up to the column (lower panel). A mass of 54,772 Da was estimated for the binary complex (summing His6 VPg and His6 eIF4E molecular weights) vertical dashed line. For comparison, absorbance values were standardized to the maximum value of each peak. (**B**) Distribution of the various molecular species through the size exclusion chromatography (SEC) separation of a His6 VPg-His6 eIF4E mix. Upon elution, fractions 1. 2 and 3 were recovered and submitted to SDS-PAGE analysis. (**C**) Determination of the three molecular species apparent molecular weights (*Mapp*) deduced from standard calibration with a set of known globular proteins. *Kav* is a mean value determined from at least three independent SEC runs. (**D**) Apparent molecular densities (*ρ*) of the three molecular species were deduced from their experimentally determined hydrodynamic radius (*Rh*) values (see material and methods). For each protein, the intersection between log*ρ* and log(*M*), *M* being the true molecular mass, allows to deduce the conformational families to which they belong to.

Because SEC leads to proteins diluting and to complexes partly dissociating during the chromatography process, it required substantial amount of proteins at concentration above 0.5 mg/mL. We experienced difficulties to obtain isolated untagged VPg and eIF4E at the concentrations suited for SEC experiments. Indeed, in vitro enzymatic tag cleavage resulted in a mixture of molecular species, which after separation (0.1–0.2 mg/mL), were not concentrated enough for SEC. Consequently, hydrodynamic parameters were deduced from fluorescence anisotropy measurements, which return a more detailed analysis of hydrodynamic behavior and can be operated at much lower

concentrations. Anisotropy measurements give access to the rotational correlation time (*θ*) of the proteins. This parameter is strongly related to their hydrodynamic properties, as it depends on the protein shape and it is linked to the effective solvent shell accompanying the protein rotational diffusion. For that purpose, a fluorescent probe, *N*-acetyl-*N* -(5-sulfo-1-naphtyl)ethylenediamine (AEDANS) was linked to the VPg single cysteine with an efficiency of 0.9 AEDANS moiety per VPg. In another set of experiments, AEDANS was also coupled to His6 eIF4E, eIF4E and eIF4EΔ1–46 to evaluate the first 1–46 disordered residues contribution to eIF4E compaction and also, more generally, the effect of the His6 tag on the compaction of monomeric forms. There are four cysteine residues within lettuce eIF4E, among which two are strictly conserved in plant orthologues. The modification resulted in a mean of 1.7 AEDANS moieties per molecule. The addition of the fluorophore did not alter the binding properties of His6 eIF4E, eIF4E, and eIF4EΔ1–46 to VPg (Figure S3). This result was not surprising as wheat eIF4E, either reduced, oxidized, or with a cysteine-to-serine mutation do not undergo structural changes and are functional, all binding m7GTP in a similar and labile manner [25]. In addition, in the structure of pea eIF4E, the two sulfur atoms are in close proximity but are clearly not bridged [26].

#### *2.2. His Tagging Modulates Proteins Hydrodynamic Parameters*

As expected, upon addition of His6 eIF4E, the fluorescence anisotropy of His6 VPg\*, increased proportionally to the amount of His6 VPg\*-His6 eIF4E complex formed. It reached a plateau value indicating a saturation, Figure 2A inset.

**Figure 2.** Fluorescence anisotropy of the various molecular forms of eIF4E and VPg. (**A**) The fluorescence anisotropy of His6 VPg\* (300 nM, open circles) and His6 VPg\*-His6 eIF4E complex (mix of 300 nM His6 VPg\* and 2 μM His6 eIF4E filled circles) was recorded as a function of the viscosity increase at 25 ◦C. The reciprocal of the emitted light anisotropy (Perrin's plot), 1/*A*, is plotted as a function of T/*η*. *V*, the apparent molecular volume of the proteins and their complexes and *A*0, the fundamental anisotropy were obtained respectively from the slope and intercept at infinite viscosity. Inset, the fluorescence anisotropy increase upon association of His6 eIF4E with His6 VPg\* (300 nM). (**B**) Perrin's plots of untagged VPg\* (open circles) and VPg\*-eIF4E complex (filled circles) in the same conditions as (**A**). (**C**) Perrin's plots of His6 VPg\* (open squares), VPg\* (filled circles) and D. His6 eIF4E\* (filled circles), eIF4E\* (open squared) and eIF4E<sup>Δ</sup>1–46\* (filled triangles). Experimental conditions were the same as in B. All measurements were obtained using VPg from lettuce mosaic virus (LMV) AF199 strain.

The modification by the probe did not significantly change the binding strength as comparable dissociation constants (*K*D) values were found for probed and unprobed proteins. A *K*<sup>D</sup> value of 63 nM could be extracted from the data, in agreement with intrinsic fluorescence measurements (Figure S3). The hydrodynamic molar volume (*V*) and the rotational correlation time (*θexp*) of the various molecular species were determined from steady state fluorescence anisotropy measurements as described in the experimental procedure section.
