Quantifying Biomolecular Interactions in High-Conductivity Samples With Capillary Electrophoresis

Abstract

Capillary electrophoresis (CE) is a powerful tool for studying biomolecular interactions due to its high speed, low sample consumption, and adaptability. However, challenges arise when sample buffers possess higher conductivity than the background electrolyte (BGE), leading to peak distortions and reduced measurement accuracy in binding assays such as affinity probe CE and nonequilibrium CE of equilibrium mixtures. This study investigates these effects using a combination of simulation and experiment, focusing on aptamer-protein interactions. Moderate conductivity mismatches (e.g., sample buffer = 2 × tris glycine, BGE = 30 mM tricine) led to peak splitting artifacts, whereas large mismatches (e.g., sample buffer = phosphate-buffered saline, BGE = 30 mM tricine) produced broad, indistinct peaks, obscuring free and bound species. Simulations revealed that these artifacts arise from analyte ions trapped in high-conductivity sample plugs and are exacerbated by longer injection times. Experimental results confirmed that reducing plug length and selectively excluding artifact peaks during analysis improves quantification accuracy. When traditional separation fails under high-conductivity conditions, we propose an alternative method based on quantifying the "de-stacked" fraction of aptamers escaping the sample zone. This approach yielded values for the dissociation constant (K_d) and Hill coefficient (n) comparable to those obtained using fluorescence anisotropy, demonstrating its viability. The method was further validated by measuring the binding of an integrin-targeting aptamer (S10yh2) to human serum albumin. Overall, this work provides practical guidelines and analytical strategies for accurate quantification of binding interactions in CE under nonideal conductivity conditions, broadening the applicability of CE for bioanalytical research.

Keywords: albumin; aptamer; conductivity; simulation; thrombin.

FIGURE 1

Electropherograms of 75 nM FAM‐labeled 29mer (upper trace) and 110 nM thrombin + 75 nM FAM‐labeled 29mer (lower trace) samples dissolved in different sample buffers and separated in different run buffers. In all cases, samples were injected at 0.3 psi for 3 s: (A) sample buffer = 2× TG, run buffer = 2× TG; (B) sample buffer = 2× TG, run buffer = 30 mM tricine at pH 7.5; and (C) sample buffer = 1× PBS, run buffer 30 mM tricine at pH 7.5. The red star in 2B labels the extra peak.

FIGURE 2

Simul 6 simulations of adenosine monophosphate (AMP) peak profiles when dissolved in TG buffer and separated in 30 mM tricine (pH 7.5): (A) AMP concentration profile is shown at different time intervals (10 s) of the simulation along the capillary length; (B) AMP concentration profile at a set time point (t3) at different TG sample buffer concentrations that produce different conductivity gaps between TG sample buffer and 30 mM tricine (pH 7.5) run buffer.

FIGURE 3

(A) Simulated concentration profile of adenosine monophosphate (AMP) along the capillary length at a set time (10 s) point during the separation. Sample was dissolved in 2× TG buffer and separated in 30 mM tricine (pH 7.5). The initial AMP plug length is decreased (from top trace to the bottom). All other separation conditions were kept constant. (B) Experimental electropherograms of 75 nM FAM‐labeled 29mer + 110 nM thrombin when the injection time is decreased from 9 s (uppermost trace) to 2 s (lowest trace). Separation conditions: voltage = 30 kV, L _t = 30 cm, L _d = 10 cm, injection pressure 0.3 psi.

FIGURE 4

(A) Two possible peak integration methods when an extra peak is present in the electropherogram. (B) Electropherogram of FAM‐labeled 29mer incubated with increasing thrombin concentrations (from top to bottom) in 2× TG buffer and separated in 30 mM tricine buffer (pH 7.5) injected for 5 s at 0.3 psi. Arrow points to the extra peak due to splitting. (C) Binding curves obtained by plotting the complex peak areas calculated in two different methods and plotted against thrombin concentrations. (D) Injection time is changed to 3 s. Rest of the conditions are similar to part (C). (E) Binding curves obtained for the measurements shown in part (D). (F) Binding curve measured using fluorescence anisotropy (FA).

FIGURE 5

(A) The proposed model for de‐stacking of the aptamer in the presence of different thrombin concentrations in the sample. (B) Simul 6 simulation of a hypothetical analyte is dissolved in PBS and separated in 30 mM tricine buffer at pH 7.5. The mobility of the analyte is increased from top trace to the bottom to observe the increase of de‐stacking.

FIGURE 6

(A) Electropherogram of FAM‐labeled 29mer incubated with decreasing thrombin concentrations (from top to bottom) in 1× PBS buffer and separated in 30 mM tricine buffer (pH 7.5) injected for 3 s at 0.3 psi. (B) De‐stacked peak area percentage is plotted against the thrombin concentration. Error bars show standard deviation (n = 3). The inset shows the binding curve measured using FA. (C) The inserted table showing the comparison between the K _d and n obtained from FA and CE for the same set of 29mer + thrombin samples with the concentrations indicated in part (A). CE, capillary electrophoresis; FA, fluorescence anisotropy.

FIGURE 7

(A) Electropherogram of FAM‐labeled S10yh2 aptamer incubated with decreasing HSA concentrations (from top to bottom) in 1× PBS buffer and separated in 30 mM Tris–HCl buffer (pH 8.5) injected for 3 s at 0.3 psi. (B) De‐stacked peak area percentage is plotted against the HSA concentration. The inset shows the binding curve measured using FA. (C) The inserted table showing the comparison between the K _d and n obtained from FA and CE for the same samples. CE, capillary electrophoresis; FA, fluorescence anisotropy; HAS, human serum albumin.