Orthogonal Electric Field Measurements near the Green Fluorescent Protein Fluorophore through Stark Effect Spectroscopy and pKa Shifts Provide a Unique Benchmark for Electrostatics Models

Measurement of the magnitude, direction, and functional importance of electric fields in biomolecules has been a long-standing experimental challenge. pK_a shifts of titratable residues have been the most widely implemented measurements of the local electrostatic environment around the labile proton, and experimental data sets of pK_a shifts in a variety of systems have been used to test and refine computational prediction capabilities of protein electrostatic fields. A more direct and increasingly popular technique to measure electric fields in proteins is Stark effect spectroscopy, where the change in absorption energy of a chromophore relative to a reference state is related to the change in electric field felt by the chromophore. While there are merits to both of these methods and they are both reporters of local electrostatic environment, they are fundamentally different measurements, and to our knowledge there has been no direct comparison of these two approaches in a single protein. We have recently demonstrated that green fluorescent protein (GFP) is an ideal model system for measuring changes in electric fields in a protein interior caused by amino acid mutations using both electronic and vibrational Stark effect chromophores. Here we report the changes in pK_a of the GFP fluorophore in response to the same mutations and show that they are in excellent agreement with Stark effect measurements. This agreement in the results of orthogonal experiments reinforces our confidence in the experimental results of both Stark effect and pK_a measurements and provides an excellent target data set to benchmark diverse protein electrostatics calculations. We used this experimental data set to test the pK_a prediction ability of the adaptive Poisson-Boltzmann solver (APBS) and found that a simple continuum dielectric model of the GFP interior is insufficient to accurately capture the measured pK_a and Stark effect shifts. We discuss some of the limitations of this continuum-based model in this system and offer this experimentally self-consistent data set as a target benchmark for electrostatics models, which could allow for a more rigorous test of pK_a prediction techniques due to the unique environment of the water-filled GFP barrel compared to traditional globular proteins.