Computational redesign of a fluorogen activating protein with Rosetta

Abstract

The use of unnatural fluorogenic molecules widely expands the pallet of available genetically encoded fluorescent imaging tools through the design of fluorogen activating proteins (FAPs). While there is already a handful of such probes available, each of them went through laborious cycles of in vitro screening and selection. Computational modeling approaches are evolving incredibly fast right now and are demonstrating great results in many applications, including de novo protein design. It suggests that the easier task of fine-tuning the fluorogen-binding properties of an already functional protein in silico should be readily achievable. To test this hypothesis, we used Rosetta for computational ligand docking followed by protein binding pocket redesign to further improve the previously described FAP DiB1 that is capable of binding to a BODIPY-like dye M739. Despite an inaccurate initial docking of the chromophore, the incorporated mutations nevertheless improved multiple photophysical parameters as well as the overall performance of the tag. The designed protein, DiB-RM, shows higher brightness, localization precision, and apparent photostability in protein-PAINT super-resolution imaging compared to its parental variant DiB1. Moreover, DiB-RM can be cleaved to obtain an efficient split system with enhanced performance compared to a parental DiB-split system. The possible reasons for the inaccurate ligand binding pose prediction and its consequence on the outcome of the design experiment are further discussed.

Fig 1. Rosetta Design results.

(A) The summary of the mutations observed in the 50 best (sorted by interface score) models output of the Rosetta Design protocol. The frequency of observation of the given amino acid (y-axis) at the given position (x-axis) in the analyzed variants is indicated by color (light grey–less often, dark grey–more often). Amino acid identities in the parental protein DiB1 are indicated as green squares. Amino acid substitutions in the selected for experimental evaluation protein DiB-RM are indicated as cyan squares. (B) Cartoon representation of the lipocalin Blc (PDB ID 1QWD, chain A). The positions in which mutations were observed among the 50 best (sorted by interface score) models output of the Rosetta Design protocol are shown as sticks. The positions which were mutated in the variant selected for experimental testing (DiB-RM) are colored cyan and labeled. (C) A fragment of the alignment of the DiB1 amino acid sequence (top line) and ten best design sequences (sorted by interface score). There are no mutations in any of the proteins outside the shown region. Dots indicate the same amino acid in the given position as in DiB1. The sequence selected for the experimental testing (DiB-RM) is highlighted in light blue.

Fig 2. Experimental characterization of new proteins in cellulo.

(A-D) Representative widefield fluorescence images of living HEK293 cells transiently transfected with H2B-TagBFP-DiB1 (A, B) or H2B-TagBFP-DiB-RM (C, D) constructs in presence of 0.5 μM M739 in green (A, C) and blue (B, D) detection channels; scale bars are 15 μm. (E) Green to blue fluorescence signal ratio distribution in nuclei of living HEK293 cells transiently transfected with H2B-DiB1-TagBFP or H2B-DiB-RM-TagBFP in presence of 0.5 μM M739. Represented are data for 307 cells transiently transfected with H2B-DiB1-TagBFP and 241 cells transiently transfected with H2B-DiB-RM-TagBFP. (F-G) Confocal fluorescence microscopy imaging (excitation: 488 nm, emission: 520–560 nm) of living HeLa cells transiently transfected with DiB-RM constructs in presence of 0.25 μM M739; (F) vimentin-DiB-RM, scale bar 10 μm; (G) LifeAct-TagBFP-DiB-RM, scale bar 20 μm. (H-I) Widefield fluorescence images of living HEK293 cells transiently cotransfected with H2B-DiB-RM-splitN_1-109 and DiB-RM-splitC_110-177-TagBFP constructs in presence of 0.1 μM M739 in green (H) and blue (I) detection channels; scale bars are 15 μm.

Fig 3. DiB1, DiB-RM, and DiB-RM-split performance in super-resolution microscopy of living cells.

HeLa cells transiently transfected with vimentin-DiB1 (A, D, E), vimentin-DiB-RM (B, F, G), and transiently co-transfected with vimentin-DiB-RM-splitN_1-109 + DiB-RM-splitC_110-177-TagBFP constructs (C, H, I) in the presence of 20 nM M739. Imaging conditions: 1.1 kW cm^-2 of 488 nm laser light, 30Hz acquisition frequency. (A-C) Super-resolution reconstruction from 10 000 frames, scale bars are 5 μm. (D, F, H) Average projection of 1 000 frames and super-resolution reconstructions from 10 000 frames; scale bars are 0.5 μm. (E, G, I) Normalized intensity profiles between yellow arrows shown on the images (D, F, H); black curve–widefield and red curve–super-resolution. (J-L) Comparison of DiB1, DiB-RM, and DiB-RM-split performance in a localization microscopy setup, average values for 7 cells. (J) Photostability in the localization microscopy setup. (K) The number of detected photons per single-molecule event; vertical lines represent median values. (L) Localization precision per single-molecule event; vertical lines represent median.

Fig 4. DiB-RM crystal structures analysis.

Crystal structures of the full-length DiB-RM protein crystallized in the presence of DDM (A-C) or without it (D-E), and their comparison (F-G). Cartoon representation of the content of the asymmetric unit (A, D), the overlay of two protein chains from the asymmetric unit (B, E), and binding sites of the DDM molecules (C). The difference in the protein conformation between two crystals (F, DiB-RM protein structure obtained in the presence of DDM is colored blue, without DDM–colored orange) and in chain’s packing in the asymmetric unit (G). (H) Crystal structure of the DiB-RM-split protein. The N terminus part of the split protein is colored pink, the C terminus part of the split is colored blue. (I) The lipocalin fold is well preserved in DiB-RM-split protein as showed by the overlay of DiB-RM-split protein (both fragments colored pink) with the full-length DiB-RM (colored yellow).

Fig 5. Comparison of the initial docking and modeling results with the later obtained DiB1:M739 co-crystal structure.

(A) Binding pockets in the apo Blc structures showed as green (PDB ID 1QWD) or orange (PDB ID 3MBT) mesh. (B) Position of vaccenic acid (shown as blue sticks) and M739 (shown as yellow sticks) in the binding pocket of the lipocalin Blc (shown as a blue cartoon, PDB ID 2ACO) or DiB1 (shown as a yellow cartoon, PDB ID 6UBO), correspondingly. (C) The side and the top views on the overlay of the DiB1 complex (shown in yellow) with the DiB-RM apo structure (shown in light gray). The side chains of the amino acids in the positions that differ between DiB1 and DiB-RM as well as M739 are shown as sticks. Positions of the ligand in 50 top-scored structures after redocking of the chromophore M739 into (D) DiB1 crystal structure or (E) the DiB1 crystal structure-based DiB-RM model starting from the position of the ligand found in the DiB1:M739 co-crystal structure. The M739 chromophore from the co-crystal structure is shown as yellow sticks. All docked chromophores are shown as green lines.