Fluorescence resonance energy transfer glucose sensor from site-specific dual labeling of glucose/galactose binding protein using ligand protection

Abstract

Background: Site-selective modification of proteins at two separate locations using two different reagents is highly desirable for biosensor applications employing fluorescence resonance energy transfer (FRET), but few strategies are available for such modification. To address this challenge, sequential selective modification of two cysteines in glucose/galactose binding protein (GGBP) was demonstrated using a technique we call "ligand protection."

Method: In this technique, two cysteines were introduced in GGBP and one cysteine is rendered inaccessible by the presence of glucose, thus allowing sequential attachment of two different thiol-reactive reagents. The mutant E149C/A213C/L238S was first labeled at E149C in the presence of the ligand glucose. Following dialysis and removal of glucose, the protein was labeled with a second dye, either Texas Red (TR) C5 bromoacetamide or TR C2 maleimide, at the second site, A213C.

Results: Changes in glucose-dependent fluorescence were observed that were consistent with FRET between the nitrobenzoxadiazole and TR fluorophores. Comparison of models and spectroscopic properties of the C2 and C5 TR FRET constructs suggests the greater rigidity of the C2 linker provides more efficient FRET.

Conclusions: The ligand protection strategy provides a simple method for labeling GGBP with two different fluorophores to construct FRET-based glucose sensors with glucose affinity within the human physiological glucose range (1-30 mM). This general strategy may also have broad utility for other protein-labeling applications.

Figure 1

Glucose/galactose binding protein is known to exhibit a hinge-twist conformational change upon binding of ligand.^, Here we suggest that, when two cysteine sites (SH) are available on the protein, one may be protected by the presence of ligand, allowing the remaining cysteine site to be selectively labeled

Figure 2

Structures of the cysteine-reactive dyes used in the labeling studies. Atoms indicated with dots were used in modeling studies to measure distances between NBD and TR

Figure 3

Absorbance spectra for the NBD absorption spectral region of NBD-labeled single GGBP mutants E149C (solid curve) and A213C (long dashed curve) together with the CCS-2NBD (dotted curve), NBD-TR-C2 CCS (dash-dot-dot curve), and NBD-TR-C5 CCS (short dashed curve). For visual comparison, these spectra have been normalized to 1 at their maximum NBD absorbance

Figure 4

Fluorescence response of (A) NBD-TR C2 CCS and (B) NBD-TR C5 CCS with increasing amounts of glucose. The NBD fluorescence peak occurs at 540 nm, and the TR fluorescence peak occurs at 610 nm

Figure 5

Titration curves of fluorescence response versus glucose concentration for NBD-TR C2 CCS and NBD-TR C5 CCS. The apparent dissociation constant, K_apparent, was fit using Equation (1)

Figure 6

Results of conformational analysis for NBD-TR C2 CCS (A) open form and (B) closed form and for NBD-TR C5 CCS (C) open form and (D) closed form. Texas Red at A213C is on the upper left and NBD at E149C is on the lower left in each structure. The protein is viewed with the C-terminal domain in front and the N-terminal domain in the back