Mod:GFP Mutation Project
Introduction
In 2017, Banerjee et al. performed exhaustive single point mutations on seven residues surrounding the chromophore of Green Fluorescent Protein[1]. In the unfolded state, proteins with the chromophore do not fluoresce. The rigid structure of the environment in the folded prevents non-radiative decay via a rotation around the free double bond in the chromophore that would otherwise occur in the unfolded state.[2] This would suggest that mutations that disrupt the environment or allow greater freedom for the chromophore would disrupt fluorescence. Despite this, GFP was found to be surprisingly tolerant of these mutations, even in highly conserved residues.
Four mutants were found by Banerjee to be totally dark, but even these displayed some degree of fluorescence when stored, suggesting a slow chromophore maturation (CM) rate. Low relative fluorescence (RF) in mutants that wasn't due to a slower CM rate were often due to an increased mobility of the chromophore. This would reduce the barrier to non-radiative decay.
In this study we aim to correlate computational results with Banerjee's experimental work, with particular focus on mutants that change the emission and excitation wavelength.
Existing Data
Interactive Model
The model below shows the seven residues tested by Banerjee in red, and the two residues previously tested in blue. The ESPT pathway is shown with dashed green lines.
Residues
S65-G67-Y66
Chromophore
The chromophore is formed by the autocatalytic cyclisation and oxidation of residues 65-67, in a process known as chromophore maturation (CM).
This process requires a positively charged residue near the carbonylic oxygen of Y66, which stabilises the enolic tautomer during CM[3]. This positive charge usually comes from R96.
The chromophore only fluoresces in its protein environment. Outside of this, a rotation about the doubly bond in Y66 forms the lowest energy route from the excited state, and is non-radiative[2].
R96
Arginine
Previously characterised
Highly conserved residue[3].
Two mutations on this residue, R96A and R96M, which would otherwise switch off fluorescence can be compensated with mutation Q183R, which is close enough to the Y66 carbonyl oxygen to replace R96's role.
T203
Threonine
Previously characterised
Stabilises the phenolate form of chromophore Y66[4].
E222
Glutamic Acid
Previously characterised
Highly conserved.
E222K displays wild-type fluorescence[1].
Serves as the control for Banerjee's work, and was in agreement with Nakano and co-workers' findings[5].
Could help catalyse CM. Proton sink for proposed ESPT. Despite this, a mutation with a positively charged side-chain is brightly fluorescent (E222K). E222R has reduced fluorescence. E222A, E222P and E222W have no fluorescence.
Q69
Glutamine
Characterised by Banerjee et al[1].
Brightest RF: 1.01 (Q69L) Dimmest RF: 0.56 (Q69Y)
Tolerant of all mutations. Hydrogen bonds with Q183 and has space to accept any side chain.
Q69 is part of the central helix that folds early. Despite this, no mutations eliminate nor increase fluorescence. Generally they cause a reduction in RF.
Q94
Glutamine
Characterised by Banerjee et al[1].
Brightest RF: 1.25 (Q94P) Dimmest RF: 0.46 (Q94G) Dark: Q94D
Packs with R96. Q94D (dark) allows R96 to adopt a non-native rotamer, slowing or halting CM. Conformation changes during CM.
F145
Phenylalanine
Characterised by Banerjee et al[1].
Brightest RF: 1.22 (F145M) Dimmest RF: 0.18 (F145W)
Constrains rotation of chromophore. F145W does not impeded CM, but has a lower QY and glows yellow.
In strand 7 that folds last according to Banerjee. This side chain is tightly packed, and is in close contact with S205/T205 (T205 in precursor structures) and H169.
H148
Histidine
Characterised by Banerjee et al[1].
Brightest RF: 1.18 (H148W) Dimmest RF: 0.57 (H148Y)
Hydrogen bonds with chromophore phenolate, but tolerates all mutations. Conformation changes during CM.
In strand 7 that folds last according to Banerjee. Surface residue with space to accommodate mutations.
F165
Phenylalanine
Characterised by Banerjee et al[1].
Brightest RF: 1.12 (F165W) Dimmest RF: 0.51 (F164G)
Constrains rotation of chromophore. Tolerates all mutations, but Banerjee suggests packing may affect relative fluorescence.
Unusual rotamer (p χ1 = 60°) against V150 due to lack of space. Mutations here that have high relative energies at this rotamer are likely to pack in such a way to increase mobility around the chromophore, reducing QY. Check rotamers at mutation at this point
H181
Histidine
Characterised by Banerjee et al[1].
Brightest RF: 1.03 (H181M) Dimmest RF: 0.36 (H181W) Dark: H181D
Packs with R96. H181D (dark) allows R96 to adopt a non-native rotamer. Tightly packed, and H181W clashes with A179 making the chromophore more mobile.
Q183
Glutamine
Characterised by Banerjee et al[1].
Brightest RF: 1.08 (Q183N) Dimmest RF: 0.50 (Q183R) Dark: Q183L Dark: Q183P
Hydrogen bonds with Q69 and packs against R96. Q183L (dark) interferes with packing of side-chains. Q183P (dark) creates suboptimal backbone conformation.
Dark Mutants
These mutants expressed little or no fluorescence out of the 140 mutants. Of the four totally dark mutants, all acquired some degree of fluorescence after storage at 4ºC for 3 months.[1]
H181D
Histidine to Aspartic Acid
RF = 0[1]
"A larger negative ellipticity near 217 nm ... suggesting more β-sheet character in these mutants than in the wild-type protein"[1]
Forms a stable salt bridge with R96, bending R96 away from the chromophore and reducing rate of CM.
Q183P
Glutamine to Proline
RF = 0[1]
"A smaller positive ellipticity near 200 nm ... indicating a lower degree of hydrogen bonding"[1]
"A larger negative ellipticity near 217 nm ... suggesting more β-sheet character in these mutants than in the wild-type protein"[1]
Q183L
Glutamine to Leucine
RF = 0[1]
"A larger negative ellipticity near 217 nm ... suggesting more β-sheet character in these mutants than in the wild-type protein"[1]
Q94D
Glutamine to Aspartic Acid
RF = 0[1]
CM is slowed or halted in this mutant, so it might not be possible to study it here.
The negative charge stabilises the protonated state of the chromophore. Excitation maximum is 385 nm.
Forms a stable salt bridge with R96, bending R96 away from the chromophore and reducing rate of CM.
F145W
Phenylalanine to Tryptophan
RF = 0.18[1]
"A smaller positive ellipticity near 200 nm ... indicating a lower degree of hydrogen bonding"[1]
"A larger negative ellipticity near 217 nm ... suggesting more β-sheet character in these mutants than in the wild-type protein"[1]
This mutant fluorescences intensely yellow, but possibly has a low QY. F145 is already tightly packed, and increasing the size of the group to tryptophan adds to this steric clash with S205/T205 and H169. F145M on the other hand has relieved sterics and is brighter.
Emission is blue shifted by 2 nm - a phenomenon seen in red fluorescent proteins. The modelled protein's tryptophan forms a hydrogen bond with the chromophore's phenolate oxygen, allowing it to accommodate a larger negative charge, increasing the gap between the ground and excited states.
Banerjee suggests that an S205G or H169 -> shorter side chain mutation could potentially raise the QY. Mini Project?
Special Cases
F145M
Phenylalanine to Methionine
RF = 1.22[1]
In contrast to F145W, this mutant is brighter and has a 2 nm red shift in its excitation spectrum. Supposedly, more water will have access to the cavity which could help stabilise the excited state (increased Stoke shift?) Check will additional water in the cavity on mutation
Possible Projects
Changes in Emission and Absorption Profiles
Banerjee's paper contains a huge amount of data. Their main finding is that there is little change in the emission and absorption profiles between most mutations. We should be able to show this lack of change with QM calculations provided there is not much structural change in these mutants.
Computational Methods
Using a small set of reference geometries, we will first compare the performance of several hybrid DFT functionals (TD/B3LYP, TD/CAM-B3LYP, TD/wB97XD, TD/APFD) in producing vertical excitation energies and emissions from the excited state.
Once a suitable method has been chosen, we will begin the process of screening many mutants including the exceptional mutants listed above.
Non-Covalent Interactions
We can check how sterics play a role in holding the chromophore rigid by performing an NCI calculation[6]. We will be able to compare strong and weak attractive and repulsive forces between different mutations, giving insight as to how packing is influenced. Other effects may become visible after NCI analysis.
Computational Methods
The JMol interface provides a graphical NCI analysis using the electronic densities produced from QM calculations in Gaussian (cube files).
References
- ↑ 1.00 1.01 1.02 1.03 1.04 1.05 1.06 1.07 1.08 1.09 1.10 1.11 1.12 1.13 1.14 1.15 1.16 1.17 1.18 1.19 1.20 1.21 Banerjee S., Schenkelberg C.D., Jordan T.B., Reimertz J.M., Crone E.E., Crone D.E., Bystroff C., Biochemistry, 7 Feb 2017, 56(5), 736-747, DOI:10.1021/acs.biochem.6b00800 .
- ↑ 2.0 2.1 Zhang, Q., Chen, X., Cui, G., Fang, W.-H., and Thiel, W. Angewandte Chemie International Edition, 2014, 53, 8649-8653, DOI:10.1002/anie.201405303
- ↑ 3.0 3.1 Wood, T. I., Barondeau, D. P., Hitomi, C., Kassmann, C. J., Tainer, J. A., and Getzoff, E. D. Biochemistry, 2005, 44, 16211-16220, DOI:10.1021/bi051388j
- ↑ Tsien, R. Y. Annual Review Biochemistry, 1998, 67, 509-544, DOI:10.1146/annurev.biochem.67.1.509
- ↑ Nakano, H., Okumura, R., Goto, C., and Yamane, T. Biotechnology and Bioprocess Engineering, 2002, 7, 311-315, DOI:10.1007/BF02932841
- ↑ Johnson E.R., Keinan S., Mori-Sánchez P., Contreras-García J., Cohen A.J., Yang W. JACS, 15 Mar 2010, 132, 6498-6506, DOI:10.1021/ja100936w .