Crystal Structure of Bright Fluorescent Protein BrUSLEE with Subnanosecond Fluorescence Lifetime; Electric and Dynamic Properties

The rapid development of new microscopy techniques for cell biology has exposed the need for genetically encoded fluorescent tags with special properties. Fluorescent biomarkers of the same color and spectral range and different fluorescent lifetimes (FLs) became useful for fluorescent lifetime image microscopy (FLIM). One such tag, the green fluorescent protein BrUSLEE (Bright Ultimately Short Lifetime Enhanced Emitter), having an extremely short subnanosecond component of fluorescence lifetime (FL~0.66 ns) and exceptional fluorescence brightness, was designed for FLIM experiments. Here, we present the X-ray structure and discuss the structure-functional relations of BrUSLEE. Its development from the EGFP (enhanced green fluorescent proteins) precursor (FL~2.83 ns) resulted in a change of the chromophore microenvironment due to a significant alteration in the side chain conformations. To get further insight into molecular details explaining the observed differences in the photophysical properties of these proteins, we studied their structural, dynamic, and electric properties by all-atom molecular-dynamics simulations in an aqueous solution. It has been shown that compared to BrUSLEE, the mobility of the chromophore in the EGFP is noticeably limited by nonbonded interactions (mainly H-bonds) with the neighboring residues.


Introduction
The GFP-like fluorescent proteins (FPs) have become essential noninvasive tools for visualization and monitoring of the biochemical processes within cells or whole organisms, and the range of their application is expanding continuously [1][2][3][4][5]. In the past decade, advances in fluorescence lifetime imaging grew from fundamental biological studies to advanced clinical diagnostics. Fluorescence lifetime-imaging microscopy (FLIM) is now effectively used in cell biology to monitor dynamic signaling events in the living cell [6,7]. It became an important research tool that provides for a new way to detect, visualize, and investigate the structure and function of biological systems, especially in the studies of protein-protein interactions using fluorescence resonance energy transfer (FRET) [8]. The fluorescence lifetime (FL) of the fluorophore, rather than its intensity, is used to create an image in the FLIM in time-resolved spectroscopy, demonstrating high sensitivity to the local microenvironment of the fluorophore. It produces spatially resolved images providing Int. J. Mol. Sci. 2023, 24, 6403 2 of 10 another dimension of information for visualizing fluorophores. FLIM can separate probes of the same color but different FL, providing an approach for multiparameter imaging.
Most GFP-like FPs available to date have a narrow range of FLs of 2.3-3.5 ns, which limits their application in a multiparameter FLIM. FPs with subnanosecond lifetimes remain virtually unexplored. The main issue of the method is that the shortening of the protein FL correlates with a decrease in its fluorescence quantum yield (FQY), and hence brightness. Until recently, green FPs with subnanosecond lifetimes were represented by variants with low-fluorescence brightness, which complicated their application in multiparameter FLIM.
BrUSLEE (Bright Ultimately Short Lifetime Enhanced Emitter) is a new green FP probe (λex/ λem 487/509 nm) with two reliably detected time-resolved FL components: a short one with a subnanosecond FL and a τ of~0.66 ns, and a long one with a τ~1.5 ns (Table 1) [9,10]. The protein was designed from EGFP by introducing three critical mutations, namely Thr65Gly, Tyr145Met, and Phe165Tyr ( Figure 1). The first two mutations sharply decreased the FL, and the third increased the FQY. BrUsLEE shows an exceptional fluorescence brightness of the short FL component, reaching~78% of that of the green EGFP precursor. The BrUsLEE demonstrates high performance in the multiparameter FLIM experiments and allows both a reliable detection of the probe and the recording of distinct FL signals in the presence of spectrally similar GFPs. Here, we present the X-ray structure of BrUSLEE and supporting results from the structure-based molecular dynamics (MD) simulation. is used to create an image in the FLIM in time-resolved spectroscopy, demonstrating high sensitivity to the local microenvironment of the fluorophore. It produces spatially resolved images providing another dimension of information for visualizing fluorophores. FLIM can separate probes of the same color but different FL, providing an approach for multiparameter imaging.
Most GFP-like FPs available to date have a narrow range of FLs of 2.3-3.5 ns, which limits their application in a multiparameter FLIM. FPs with subnanosecond lifetimes remain virtually unexplored. The main issue of the method is that the shortening of the protein FL correlates with a decrease in its fluorescence quantum yield (FQY), and hence brightness. Until recently, green FPs with subnanosecond lifetimes were represented by variants with low-fluorescence brightness, which complicated their application in multiparameter FLIM.
BrUSLEE (Bright Ultimately Short Lifetime Enhanced Emitter) is a new green FP probe (λex/ λem 487/509 nm) with two reliably detected time-resolved FL components: a short one with a subnanosecond FL and a τ of ~0.66 ns, and a long one with a τ~1.5 ns (Table 1) [9,10]. The protein was designed from EGFP by introducing three critical mutations, namely Thr65Gly, Tyr145Met, and Phe165Tyr ( Figure 1). The first two mutations sharply decreased the FL, and the third increased the FQY. BrUsLEE shows an exceptional fluorescence brightness of the short FL component, reaching ~78% of that of the green EGFP precursor. The BrUsLEE demonstrates high performance in the multiparameter FLIM experiments and allows both a reliable detection of the probe and the recording of distinct FL signals in the presence of spectrally similar GFPs. Here, we present the X-ray structure of BrUSLEE and supporting results from the structure-based molecular dynamics (MD) simulation. Figure 1. Alignment of the amino-acid sequences of BrUSLEE and its precursor EGFP. Differences are shown in red. The chromophore-forming triad is highlighted in gray. Residues from the chromophore nearest environment (<4 Å), which stabilize the chromophore by H-bonds (direct or via water) or hydrophobic interactions, are highlighted in yellow. The identical residues are marked with *, and the similar ones with :. | | | | BrUSLEE HYQQNTPIGDGPVLLPDNHYLSTQSALSKDPNEKRDHMVLLEFVTAAGITLGMDELYK EGFP HYQQNTPIGDGPVLLPDNHYLSTQSALSKDPNEKRDHMVLLEFVTAAGITLGMDELYK ********************************************************** Figure 1. Alignment of the amino-acid sequences of BrUSLEE and its precursor EGFP. Differences are shown in red. The chromophore-forming triad is highlighted in gray. Residues from the chromophore nearest environment (<4 Å), which stabilize the chromophore by H-bonds (direct or via water) or hydrophobic interactions, are highlighted in yellow. The identical residues are marked with *, and the similar ones with:.

Overall Structure
The crystal asymmetric unit of BrUSLEE contains three dimers AC, BD, and EF. Each dimer comprises two monomers related by a noncrystallographic twofold symmetry axis with a side-to-side packing at~120 • . The interface between the dimer subunits has a contact area of~2900 Å 2 . It is stabilized by nine direct hydrogen bonds and the hydrophobic cluster of six residues (see Table 2 for an example of EF interface). The principal structural fold of the BrUSLEE monomer, shared with all other members of the GFP family, is an 11-stranded β-barrel with loop caps from both sides. Table 2. H-bonds and hydrophobic interactions at the E-F dimer interface.

H-Bonds * d (Å)
3.62 Hydrophobic cluster * Phe123 + Leu221 + Ala206 * The numbers of four H-bonds and the residues in the hydrophobic cluster are doubled due to a 2-fold crystallography symmetry axis.

Structural Features of the Chromophore Area
The principal structural fold of the BrUSLEE monomer, shared with all other members of the GFP family, is an 11-stranded β-barrel with loop caps from both sides. The nearest shell of the Gly65-Tyr66-Gly67 chromophore is composed of 19 residues, forming an extensive H-bond network around the chromophore (Figures 2 and 3). The position of the chromophore is stabilized by direct H-bonds with five amino-acid residues and two water molecules mediating interaction with six other residues. Seven hydrophobic residues provide additional chromophore stabilization through the hydrophobic interactions. affect its absorbance and emission spectra but greatly impacted its FL, which predominantly depends on the chromophore's local microenvironment.     Three mutations in the EGFP chromophore and its nearest environment (Thr65Gly, Tyr145Met, and Phe165Tyr) resulted in BrUSLEE. The substitutions increased the free volume around the chromophore, causing a 0.7 Å shift of the hydroxyl group of Tyr66. The mutation of the chromophore Thr65 to Gly shifted the H-binding of catalytic Glu222 from Thr65 to the imidazolinone nitrogen, and the replacement of Phe165 with Tyr led to the formation of an H-bonded chain-Tyr165····Arg96····O=C-imidazolinone. The replacement of Tyr145 by a relatively flexible met facilitated the conformational change of Thr203, accompanied by a disruption of its H-bond with the Tyr66 hydroxyl. An increase in the distance between the His148 side chain and the Tyr66 hydroxyl from 2.89 Å to 3.52 Å indicated the disruption of another H-bond, typically stabilizing the position of the chromophore. These rearrangements in the nearest chromophore environment did not affect its absorbance and emission spectra but greatly impacted its FL, which predominantly depends on the chromophore's local microenvironment.

Electric and Dynamic Properties of the Chromophore-Binding Site
We assumed that a significant 3.5-fold decrease in the BrUSLEE fluorescence lifetime compared to the EGFP could be a consequence of the changes in the structural, electrostatic, and dynamic properties of the chromophore nearest environment. To explore this issue in detail, and to get further insight into the molecular origin of the observed differences in the photophysical properties of these proteins, we performed all-atom molecular-dynamics (MD) simulations of BrUSLEE and EGFP in water. The results of the MD revealed that, in both FPs, the chromophore preserves well the initial conformation observed in the crystal structures. In particular, its two-pi-electron systems almost do not change their mutual orientation-the angles between the planes of phenolic and imidazolinone rings remain at 7 • for both chromophores.
On the other hand, the MD data clearly demonstrate increased flexibility of the side chains around the chromophore (within 5 Å) in BrUSLEE compared to the EGFP precursor, thus indicating their greater conformational freedom in the mutant. To evaluate the mobility of the residues forming the nearest chromophore environment, we calculated the values of their root-mean-square fluctuation (RMSF) at the equilibrium part of the MD trajectories ( Figure 4). The RMSF values for the residues in both structures show significant differences.
On the other hand, the MD data clearly demonstrate increased flexibility of the side chains around the chromophore (within 5 Å) in BrUSLEE compared to the EGFP precursor, thus indicating their greater conformational freedom in the mutant. To evaluate the mobility of the residues forming the nearest chromophore environment, we calculated the values of their root-mean-square fluctuation (RMSF) at the equilibrium part of the MD trajectories ( Figure 4). The RMSF values for the residues in both structures show significant differences. As seen in Figure 4, the mobility of residues Leu42, Gln94, Glu95, Arg96, Asn121, Met145, Ser147, and His148 (see Figure 1) around the chromophore is noticeably higher in BrUSLEE than in EGFP (the reverse picture is observed only for Gln69 and Ile167). Thus, in general, the conformational flexibility of residues in the chromophore-binding site of EGFP is strongly limited. To elucidate the nature of this effect for both proteins, we performed a detailed analysis of interactions between the chromophore and its nearest amino-acid environment at the later stages of MD trajectories. It revealed that EGFP has a significantly higher average number of the protein-chromophore H-bonds and π-π stacking interactions per unit MD time (Table 3).  As seen in Figure 4, the mobility of residues Leu42, Gln94, Glu95, Arg96, Asn121, Met145, Ser147, and His148 (see Figure 1) around the chromophore is noticeably higher in BrUSLEE than in EGFP (the reverse picture is observed only for Gln69 and Ile167). Thus, in general, the conformational flexibility of residues in the chromophore-binding site of EGFP is strongly limited. To elucidate the nature of this effect for both proteins, we performed a detailed analysis of interactions between the chromophore and its nearest amino-acid environment at the later stages of MD trajectories. It revealed that EGFP has a significantly higher average number of the protein-chromophore H-bonds and π-π stacking interactions per unit MD time (Table 3). Compared to BrUSLEE, the parental EGFP chromophore forms~1.5-2 times more H-bonds with the protein environment, and these bonds are characterized by a greater lifetime. The values of the probability of occurrence of the H-bonds and π-π interactions between the chromophore and its nearest protein environment in EGFP and BrUSLEE are listed in Table 4 (see Figures 1-3). The Gln94, Arg96, and His148 in BrUSLEE demonstrate the highest variation of RMSF values relative to the parental EGFP precursor (Figure 4).
Since the photophysical properties of the studied protein directly depend on the electric field in the chromophore area, we also used MD simulations to get the relevant information. As a result, we delineated two main electronic states in EGFP and BrUSLEE, differing in the orientation of the dipole moment (DM) vector of the chromophore ( Figure 5). The most populated state is characterized by the DM of~5.0 ± 0.3 D with an orientation of 50 • to the CG2-CZ(OH) axis of the chromophore phenolic ring. The minor state is much more diffuse, having a DM of~2.5 ± 0.4 D with a corresponding orientation of~25 • . These states correspond to different orientations of the O-H and C=O groups of the chromophore phenolic and imidazolinone rings. The hydroxyl proton has either a trans-(major state) or cis-(minor state) configuration with respect to the oxygen in the carbonyl group. Since the photophysical properties of the studied protein directly depend o electric field in the chromophore area, we also used MD simulations to get the rel information. As a result, we delineated two main electronic states in EGFP and BrUS differing in the orientation of the dipole moment (DM) vector of the chromophore (F 5). The most populated state is characterized by the DM of ~5.0 ± 0.3 D with an or tion of ~50° to the CG2-CZ(OH) axis of the chromophore phenolic ring. The minor s much more diffuse, having a DM of ~2.5 ± 0.4 D with a corresponding orientation of These states correspond to different orientations of the O-H and C=O groups o chromophore phenolic and imidazolinone rings. The hydroxyl proton has either a (major state) or cis-(minor state) configuration with respect to the oxygen in the car group.

Conclusive Summary
The green-fluorescent biomarker BrUSLEE, characterized by a short fluorescent lifetime (FL) of~0.66 ns has been studied by X-ray and molecular dynamics methods. The protein demonstrates an exceptional fluorescence brightness, reaching~78% of that of the green EGFP precursor (FL~2.83 ns). BrUSLEE allows the recording of a distinct FL signal in the presence of spectrally similar GFPs in multiparameter FLIM experiments. Its development from the EGFP precursor by three mutations in the chromophore area (Thr65Gly, Tyr145Met, and Phe165Tyr) resulted in a change of the chromophore microenvironment due to a significant rearrangement in the side chain conformations.
The mutations increased the free volume around the chromophore, causing a shift of the Tyr66 hydroxyl group of the chromophore by 0.7 Å. The following increase in the conformational freedom of the residues in the chromophore area was confirmed by a molecular dynamics (MD) simulation. Compared to the EGFP precursor, which is characterized by an increased packing density in the chromophore area, in BrUSLEE the stabilizing contacts of the chromophore with the environment were noticeably weaker. The average number of protein-chromophore H-bonds and stacking interactions was 1.5 times less (Tables 3 and 4). This, in turn, resulted in higher mobility of its neighboring residues; for some of them, the increase in the corresponding RMSF values reached 40%, (Figure 4). The aforementioned reorganization of the chromophore-binding site in BrUSLEE as compared to its precursor also changed the electric properties of the chromophore and its microenvironment. This significantly shifted the populations of the two main electronic states with different dipole moments observed in the course of MD simulations. The major state was characterized by DM~5.0 D with a~50 • orientation to the CG2-CZ(OH) axis of the chromophore phenolic ring, while the more diffuse minor state had a DM~2.5 D with a corresponding~25 • orientation.

Crystallization
BrUSLEE was dialyzed against 20 mM Tris pH 8.0 200 mM NaCl buffer and concentrated to 40 mg/mL. Crystals suitable for data collection were obtained from 6.4% Tacsimate pH 5.0, 16% PEG 3350 by hanging drop experiment in which 2 µL of the protein were mixed with 2 µL of the reservoir solution and incubated against the same reservoir at 20 • C for two weeks. The chemicals for crystallization were obtained from Hampton Research (Aliso Viego, CA, USA).

X-ray Data Collection, Structure Solution, and Crystallographic Refinement
The X-ray experiment was collected at the SER-CAT BM line at APS. Prior to data collection, the crystal was briefly dipped into the cryoprotecting solution containing the reservoir solution plus 30% glycerol and flash frozen in a 100 K nitrogen stream. All diffraction images were processed with HKL2000 [13]. Crystal structures were solved by the molecular replacement method with MOLREP [14,15] using the coordinates of EGFP (PDB ID: 4EUL) as a search model. Crystallographic refinement was performed with REFMAC5 [16], alternating with manual revision of the model with COOT [17]. Crystallographic data and refinement statistics are given in Table 5. The coordinates of BrUSLEE were deposited in the Protein Data Bank under accession code PDB_ID: 8BVG.

Molecular Dynamics Simulation
The starting models of EGFP and BrUSLEE, corresponding to their crystal structures, were taken from the Protein Data Bank (PDB entries 2Y0G and 8BVG, respectively). Both proteins were placed in rhombic dodecahedron boxes (A = B = C = 100 Å, α = β = γ = 60 • ) and solvated with explicit TIP3P water [18]. Na + ions [19] were added to neutralize the total charge of the system. Molecular dynamics (MD) simulations were carried out with the GROMACS software package version 2021.5 [20] using the CHARMM36 force field [21]. An integration time step of 2 fs was used, and 3D periodic boundary conditions were imposed. Simulations were performed with a constant temperature (325 K) and pressure (1 bar) maintained using the V-rescale [22] and the Parrinello-Rahman [23] algorithms, respectively. The isotropic pressure coupling was used in the simulations. The 12 Å cutoff radius was defined for the Coulombic and van der Waals interactions. Electrostatic interactions were evaluated using the particle-mesh Ewald (PME) summation [24] (real-space cutoff of 12 Å). Protein and solvent molecules were coupled separately. Partial charges and bonded parameters of the chromatophore were obtained from Reuter et al. [25]. Missing values were transferred from the protein [21] or CGenFF [26] CHARMM36 parameter sets by analogy.
The simulated systems were first equilibrated in several stages: 3000 steps of steepest descent minimization followed by heating from 5 K to 325 K during a 100-ps MD run, in which coordinates of the protein heavy atoms were restrained to permit solvent relaxation. In the next step, the systems were subjected to 50 ns of MD equilibration with fixed protein Cα atoms. Then, 100 ns long production-MD simulations without restraints were performed. Atomic coordinates from the MD trajectories were analysed with a timestep of 100 ps using original GROMACS and in-house utilities. Lastly, 75 ns of production MD run were used in the analysis. The per-residue values of the root-mean-square fluctuations (RMSF) were calculated using the rmsf utility from the GROMACS package. The dipole moment (DM) of the chromatophore part (imidazolone ring, aryl-alkene bond, and phenol ring) was calculated using the CHARMM36 force field partial charges. The intermolecular contacts (including H-bonds), DM parameters, and the geometrical characteristics of the chromophore were calculated using in-house software.
Author Contributions: Writing-original draft preparation, E.G. and I.A.; molecular dynamics calculations, R.E. and N.K.; data curation, A.B. and A.M.; X-ray data collection, S.P.; protein structure solution, N.P.; conceptualization, supervision, V.P.; correction final paper version, R.E. All authors have read and agreed to the published version of the manuscript.
Funding: This work was supported by a Grant from the Russian Science Foundation (project № 23-24-00011). Supercomputer MD calculations were supported within the framework of the HSE University Basic Research Program. Access to computational facilities of the Supercomputer Center "Polytechnical" at the St. Petersburg Polytechnic University is gratefully appreciated.
Institutional Review Board Statement: Not applicable.