Small-angle X-ray scattering experiments of monodisperse intrinsically disordered protein samples close to the solubility limit

Abstract

The condensation of biomolecules into biomolecular condensates via liquid-liquid phase separation (LLPS) is a ubiquitous mechanism that drives cellular organization. To enable these functions, biomolecules have evolved to drive LLPS and facilitate partitioning into biomolecular condensates. Determining the molecular features of proteins that encode LLPS will provide critical insights into a plethora of biological processes. Problematically, probing biomolecular dense phases directly is often technologically difficult or impossible. By capitalizing on the symmetry between the conformational behavior of biomolecules in dilute solution and dense phases, it is possible to infer details critical to phase separation by precise measurements of the dilute phase thus circumventing complicated characterization of dense phases. The symmetry between dilute and dense phases is found in the size and shape of the conformational ensemble of a biomolecule-parameters that small-angle X-ray scattering (SAXS) is ideally suited to probe. Recent technological advances have made it possible to accurately characterize samples of intrinsically disordered protein regions at low enough concentration to avoid interference from intermolecular attraction, oligomerization or aggregation, all of which were previously roadblocks to characterizing self-assembling proteins. Herein, we describe the pitfalls inherent to measuring such samples, the experimental details required for circumventing these issues and analysis methods that place the results of SAXS measurements into the theoretical framework of LLPS.

Keywords: Coflow; IDP; IDR; Intrinsically disordered protein; LLPS; Phase separation; SAXS; Self-association; Size-exclusion chromatography.

Figure 1:

Single chain IDR dimensions report on emergent properties. (A) Conditions that lead to dilute chain compaction similarly drive phase separation, while conditions that result in expansion maintain disperse protein. (B) The resolution of a SAXS experiment. The parameter D represents the resolution or distances within the protein that are primarily contributing to the SAXS curve at a defined point. The fractal dimension, d, is defined as the slope of log(I(q)) vs log(q) in the higher angles.

Figure 2:

SEC-SAXS enables the separation of monomeric IDR from unwanted oligomers and aggregates. The baseline can be chosen from statistically similar frames, where no part of the curve deviates from expected random noise, near the sample elution. The desired sample data is averaged from a region where the R_G is not concentration dependent.

Figure 3:

The R_G as a function of elution time is diagnostic of intermolecular interactions. The elution from the SEC column is monitored by UV absorption. The magnitude of the absorption is directly related to the protein concentration. The R_G is calculated from a rolling average along the protein elution. If the R_G is flat across all elution concentrations, there are no inter-monomer interactions and the sample is monodispersed. A negative slope indicates that larger species are separated by the column and are eluting first. A domed elution profile indicates that the protein self-associates in a concentration dependent manner after elution from the column.

Figure 4:

Primary SAXS data transformations. (A) Raw SAXS data on IDRs presented in log-log format highlights the low q data, power law regimes and correlation length. Binned data is shown as circles. (B) The Guinier transformation of low q (qR_G < 1) data is used to calculate R_G. Lower plot shows the residuals. (C) The Kratky transformation enables visual inspection of the fractal dimension as indicated by the slope of the data at high q relative to Gaussian chain and globule references. On a dimensionless Kratky plot, the intersection of red lines indicates the location of the maximum for a globule. Deviation of the maximum from this point indicates flexibility.

Figure 5:

The coflow sample chamber enables maximizing signal-to-noise while minimizing radiation damage. (A) A schematic view of the coflow sample chamber. The top of the chamber contains buffer inlet and an outlet to remove excess buffer. The sample is injected into the middle of the capillary where a constant flow is maintained by the total flow outlet pump. The result is the sample ‘coflowing’ with the buffer in the laminar regime and not mixing. (B) The coflow sample chamber installed at the BioCAT beamline (18ID-D) at the Advanced Photon Source. (C) Low q data shown in Guinier format indicating that identical R_Gs can be calculated from 0.05 mL samples injected onto a 3 mL SEC column at concentrations ranging from 13.5 to 0.4 mg/mL (12.5 kDa protein).

Figure 6:

Model-dependent fitting of SAXS data on a phase-separating IDR. Fits using the Gaussian Chain, Swollen Gaussian Chain and the Empirical Molecular FF to experimental data on a phase-separating IDR are shown in (A) raw and (B) Kratky format. The Swollen Gaussian Chain and Empirical Molecular FF provide good fits to the data and near equal R_G. However, the scaling exponents differ significantly.

Figure 7:

Changing sequence features that influence the driving force for phase separation are reflected in the protein dimensions under dilute conditions. (A) Raw data on a series of IDR variants with different numbers of aromatic amino acids shows the predicted decrease in radius with increased propensity to phase separate. (B) The scaling exponent increases with the removal of aromatic amino acids indicating more favorable solvation. Dashed lines are fits to the Empirical Molecular FF.