Generation of Photocaged Nanobodies for Intracellular Applications in an Animal Using Genetic Code Expansion and Computationally Guided Protein Engineering

Abstract Nanobodies are becoming increasingly popular as tools for manipulating and visualising proteins in vivo. The ability to control nanobody/antigen interactions using light could provide precise spatiotemporal control over protein function. We develop a general approach to engineer photo‐activatable nanobodies using photocaged amino acids that are introduced into the target binding interface by genetic code expansion. Guided by computational alanine scanning and molecular dynamics simulations, we tune nanobody/target binding affinity to eliminate binding before uncaging. Upon photo‐activation using 365 nm light, binding is restored. We use this approach to generate improved photocaged variants of two anti‐GFP nanobodies that function robustly when directly expressed in a complex intracellular environment together with their antigen. We apply them to control subcellular protein localisation in the nematode worm Caenorhabditis elegans. Our approach applies predictions derived from computational modelling directly in a living animal and demonstrates the importance of accounting for in vivo effects on protein‐protein interactions.


C. elegans maintenance
C. elegans strains were maintained under standard conditions unless otherwise stated. [1,2]

Plasmid generation
All plasmids used for transgenesis of C. elegans are described in Supplementary  Table 1. Plasmids for expression in C. elegans were generated using 3 Fragment Multisite Gateway Cloning (Thermo Fisher Scientific). Details on cloning are listed in Supplementary Table 2. All primers and synthetic genes (gBlocks) were synthesised by IDT. All genes were optimised for expression in C. elegans using the online C. elegans codon adaptor. [3] PCRs were carried out using Q5 2x Hot-Start Master Mix (New England Biolabs). PCR and digestion products were recovered from agarose gels following electrophoresis using the Zymogen Gel Recovery Kit (Zymo Research). DNA assembly was performed using NEBuilder 2x Master Mix (New England Biolabs). Standard ligation of restriction enzyme digest products was performed using the Roche Rapid Ligation Kit (Merck Life Sciences Life Sciences). All reagents were used in accordance with the manufacturer's specifications.
Entry and expression plasmids were grown in NEB-5α cells (New England Biolabs). Destination plasmids were grown in One Shot ccdB survival cells (Thermo Fisher Scientific). Transformations were performed in accordance with the manufacturer's specifications. Transformed cells were grown overnight on TB agar containing the appropriate selection antibiotics. Plasmids were recovered from bacteria using the QIAprep Spin Miniprep kit (Qiagen).

C. elegans transgenesis
Transgenic C. elegans strains were generated by biolistic bombardment into either the N2 or smg-6(ok1794) genetic backgrounds, with Hygromycin B resistance used as the selection marker as previously described. [4] Spermidine (Merck Life Sciences Life Sciences) was used to precipitate DNA onto gold particles of 0.3-3µm diameter (ChemPur). 900psi rupture discs (Bio-Rad) and macro carrier discs (Inbio Gold) were used in a PDS1000/He Biolistic Particle Delivery System (Biorad). Transgenic strains were maintained on NGM agar supplemented with 0.3mg/mL Hygromycin B (Formedium). All strains used in this study are described in Supplementary Table 3.

C. elegans feeding of photocaged amino acids
The amino acid NPY and the dipeptide K-NPY were custom synthesised (NewChem Technologies), and ONBY was purchased (Fluorochem). K-NPY was dissolved in water and added to molten NGM agar to make the desired concentration. NPY and ONBY were dissolved in 5M NaOH before addition to molten NGM agar, followed by neutralisation by HCl.
Before feeding with ncAAs, worms were grown on NGM agar plates seeded with a lawn of E. coli OP50 bacteria. Animals were grown until the food was depleted and the population contained a large number of age synchronised L1 larvae. Synchronised populations of L1 animals were washed off starved plates using M9 buffer and transferred to ncAA-NGM agar plates. 40uL of dissolved freeze-dried OP50 (LabTIE) was added to the plates as food. Animals were grown on ncAA plates at 20°C for 24-48 hours before uncaging followed by imaging or western blotting.

Uncaging of photocaged nanobodies
Animals were uncaged under a 365nm LED with the output set to 30% (10mW/cm 2 ). Animals were transferred from ncAA-NGM agar plates to NGM-only plates before uncaging. Animals were imaged immediately after uncaging.

C. elegans lysis and Western blotting
Synchronised C. elegans populations were grown on ncAA-NGM agar plates at 20°C for 24-48 hours, then washed off plates using M9 buffer (supplemented with 0.001% Triton-X100 to prevent animals from sticking to pipette tips). Worms were settled by gravity, the supernatant was removed, and worms were resuspended in lysis buffer at a 1:2 volume ratio of worms to lysis buffer. The lysis buffer consisted of a 4:1 mix of 4X Bolt LDS Sample Buffer (Life Technologies) and NuPAGE Sample Reducing Agent (Thermo Fisher Scientific) respectively. Lysis was performed by a freeze/thaw cycle of overnight freezing at -80°C followed by 15 min incubation at 95°C while shaking.
Samples were run on precast Bolt 4 to 12% gels (Thermo Fisher Scientific) for 18 min at 200V. Proteins were transferred from the gel onto a nitrocellulose membrane using an iBlot2 device (Thermo Fisher Scientific).
After transfer, the membrane was blocked with 5% milk powder in PBST (1xPBS supplemented with 0.1% Tween-20) for 1h at room temperature. Incubation with primary antibodies was carried out in PBST + 5% milk powder at 4C overnight. Blots were washed 4 x 5 minutes with PBST + 5% milk powder before incubation with secondary antibody for 1h at room temperature.
The primary antibodies used were mouse anti-GFP (clones 7.1 and 13.1) (Roche) at a dilution 1:5000 for SGR57 and SGR58, rat anti-HA clone 3F10 (Roche) at a dilution of 1:1000 for SGR57, and mouse anti-mCherry-Tag Monoclonal (Elabscience) at a dilution of 1:1000 for SGR58. The secondary antibodies used were Horse anti-mouse IgG HRP (Cell Signalling Technology) at a dilution of 1:5000 for anti-GFP and 1:3000 for anti-mCherry-Tag Monoclonal, and Goat anti-Rat IgG (H+L) HRP (Thermo Fisher Scientific) 1:5000. Pierce ECL Western Blotting Substrate (Thermo Fisher Scientific) or SuperSignal West Femto chemiluminescent Substrate (Thermo Fisher Scientific) were used as detection agent.

Imaging of C. elegans
All microscopy imaging was carried out on a Zeiss M2 imager. Animals were mounted on 2.5% agar pads on glass slides. To immobilise, animals were picked into drop of 50mM NaN3 (Thermo Fisher Scientific) on the pad. The NaN3 was diluted from a 100mM stock using M9. Animals picked into the drop were left for 3-4 minutes before imaging.

Measurement of mCherry nuclear:cytoplasmic ratio
Measurement of mCherry nuclear/cytoplasmic ratio in animals expressing photocaged nanobody variants was performed using ImageJ software. Nuclear average brightness was measured in a region of interest within the nucleus avoiding interference from the nucleolus, from which nanobody::mCherry fusions appeared excluded. Cytoplasmic average was measured from the cytoplasmic area proximal to the nuclei, as per the method described by Kelley and Paschal. [5] The ImageJ "threshold" function was used to generate regions of interest encompassing the nuclei in the GFP channel. These "small" ROIs were used to generate a mask which was then enlarged by applying the ImageJ "dilate" function three times to generate the "large" ROIs which extended beyond the nucleus. The area and average brightness of these "small" and "large" ROIs were measured, and cytoplasmic brightness was acquired by the following formulae:

Statistical analysis of mCherry nuclear:cytoplasmic ratio measurements
Statistical significance of subcellular localisation results was determined using a twotailed Welch's T-test. Regressions were fit using a one-phase linear decay interpolation. Calculations were performed using GraphPad Prism version 9.0.2 for Windows (GraphPad Software).

Parameterisation of photocaged amino acids for molecular dynamics
To prepare input files for MD simulations of eNB Y37ONBY and eNB Y37NPY it was necessary to generate parameters for the ONBY and NPY caging groups. These parameters were derived from the Generalised Amber Force Field (GAFF) [6] and supplemented with quantum mechanical parameters for the nitro-group. [7] A detailed description of the process, including files, code, and Python environment used to generate the input files for these simulations, can be found at https://github.com/wells-wood-research/oshea-j-interfaceengineering-2021.

Molecular dynamics of nanobody/GFP complexes
Files were prepared for molecular dynamics using AmberTools18 [6] and Open Babel. [8] Molecular dynamics simulations were performed using OpenMM. [9] The nonbonding interactions were modelled by PME, and the cut-off for non-bonding interactions was 1nm. Bonds involving hydrogen atoms were constrained in length. Simulations were run at 1bar and 300K. Pressure was maintained by a Monte Carlo barostat and temperature was maintained by the Langevin integrator. The frictional constant for the Langevin integrator was 1ps -1 . A timestep of 2fs was used. A detailed description of the simulations, including the files, code, and Python environment used to perform them, can be found at https://github.com/wells-wood-research/oshea-j-interface-engineering-2021.

Alanine scanning of nanobody/GFP complexes
Alanine scanning was performed using the open-source program BUDE Alanine Scan. A detailed description of the process, including the environment, files, and code used, can be found at https://github.com/wells-wood-research/oshea-j-interface-engineering-2021.

Cloning plasmids for E.coli periplasmic protein expression
For expression of proteins in the periplasm of E. coli, constructs were cloned into the pSANG10-3F plasmid, and these plasmids were transformed into BL21(DE3) cells (New England Biolabs).

Expression and purification of GST::GFP and GST
Plasmids for expression of GST::GFP or GST were transformed into BL21(DE3) cells (New England Biolabs) under conditions specified by the manufacturer. The transformation was grown on TB agar plates supplemented with carbenicillin (200ug/mL) at 37°C overnight. The next day, a colony was picked to inoculate 3mL of TB supplemented with carbenicillin (200ug/mL) at 37°C with shaking at 220rpm overnight. The next day, this culture was used at 1000x to inoculate AIM-2YT Broth Base including trace elements (Formedium) supplemented with carbenicillin (400ug/mL). For GST::GFP fusion or GST alone, cultures of 150mL were used. For nanobody mutants, cultures of 1L were used. The culture was grown at 37°C for 2 hours, then 30°C overnight, shaking at 220rpm.
For purification of GST::GFP fusion or GST alone, the culture was split between 10 15mL falcon tubes and cells were centrifuged (4000G. 4°C, 10 minutes). Pellets were stored at -20°C for up to 2 weeks. When protein was required, cells were resuspended in BugBuster Extraction Reagent (Merck Life Sciences) supplemented with lysozyme (25ng/mL, Thermo Fisher Scientific) cOmplete EDTA-free protease inhibitor cocktail tablets (1 tablet per 50mL, Merck Life Sciences), and DNaseI (2U/mL, New England Biolabs). Protein was recovered as per the method described for BugBuster protein extraction described by the manufacturer.

Expression and purification of nanobody mutants
Plasmids for expression of given protein was transformed into BL21(DE3) cells (New England Biolabs) under conditions specified by the manufacturer. The transformation was grown on TB agar plates supplemented with kanamycin (100ug/mL) at 37°C overnight. The next day, a colony was picked to inoculate 3mL of TB supplemented with kanamycin (100ug/mL) at 37°C with shaking at rpm overnight. The next day, this culture was used at 1000x to inoculate 1L AIM-2YT Broth Base including trace elements (Foremedium) supplemented with kanamycin (100ug/mL). The culture was grown at 37°C for 2 hours shaking at 220rpm, then 30°C overnight shaking at 160rpm.
Cultures were centrifuged (4000G, 4°C, 10 minutes). The supernatant was discarded. The cells were resuspended in 5mL lysis buffer per gram of pellet. The lysis buffer consisted of BugBuster Extraction Reagent (Merck Life Sciences) supplemented with lysozyme (25ng/mL, Thermo Fisher Scientific), cOmplete EDTA-free protease inhibitor tablets (1 tablet per 50mL, Merck Life Sciences), and DNaseI (2U/mL. New England Biolabs). The lysis was incubated at room temperature for 30 minutes with gentle shaking, then centrifuged (16,000G, 4°C, 20 minutes). The supernatant was recovered at stored at 4°C. The supernatant was supplemented with 1mL Ni-NTA resin and 10mM imidazole, then incubated at 4°C while rolling for 1 hour. Ni-NTA beads were centrifuged briefly, the supernatant discarded, and the beads resuspended in 5mL of ice-cold wash buffer (20mM imidazole in 1xPBS). This wash was performed a total of 3 times. After the final wash, the beads were resuspended in 1mL of ice-cold elution buffer (500mM imidazole in 1xPBS), left gently rocking for 10 minutes at room temperature, then briefly centrifuged, supernatant recovered and stored at 4°C, and the beads resuspended again in 1mL elution buffer. This elution was performed a total of 4 times, and all recovered supernatants were assayed by SDS-PAGE. 5uL of sample was diluted 1:1 with a 4:1 mix of 4X Bolt LDS Sample Buffer (Life Technologies) and NuPAGE Sample Reducing Agent (Thermo Fisher Scientific) respectively. The mix was shaken at 95°C for 15 minutes to denature, then electrophoresis was performed on precast Bolt 4 to 12% gels (Thermo Fisher Scientific) for 18 min at 200V. The gel was incubated in Coomassie reagent shaking at room temperature for 15 minutes before imaging. The elute fractions with little contamination and significant amount of nanobody were combined, transferred into 12-14,000 Da MWCO dialysis tubing (Scientific Laboratories Supplies) and dialysed against 10L of 1xPBS at 4°C overnight, then 2L of 1xPBS at 4°C for 1 hour. The dialysed fraction was recovered, and presence of desired protein was determined by SDS-PAGE.
The dialysed fraction was concentrated using Amicon Ultra 0.5mL 3kDa Molecular Weight Cut Off columns (Merck Life Sciences) by centrifugation at 14,000 G. Protein concentration was determined by Bradford assay against a BSA standard.

ELISA of nanobody/GFP interactions
GST-containing protein (either GST::GFP for main experiments or GST alone for the non-specific binding control) was diluted in PBS to a final concentration of 3μg/mL. 100 μl of GST-containing protein was then added to wells of a Pierce™ glutathione coated white 96well plate (Thermo Fisher Scientific), and allowed to bind for 1 hour at room temperature, with gentle shaking. For each nanobody concentration to be assayed 3 wells were prepared, as well as two wells for a negative control, without nanobody. Unbound GST-containing protein was removed from the wells by a single downward shaking motion, by hand. Any liquid remaining in the plate was then removed by striking the upturned plate onto tissue paper, 10 times. Each well was then washed with 100 μl of PBST, for 60 seconds. The PBST was removed from the wells as described above, and this wash was performed 4 times.
100 μl of diluted nanobody was then added to each well, diluted in PBS. Each nanobody concentration was assayed in duplicate during assay optimisation, and in triplicate for final measurements. Negative control wells were included in which only PBS was added. Nanobodies were allowed to bind to GFP for 1 hour at room temperature, with gentle shaking. Excess nanobody was then removed, and wells were washed 4 times as before.
100 μl of mouse anti-FLAG M2 antibody (Merck Life Sciences) was then added to the wells, diluted 1:1000 (v/v) in PBST-BSA (3%). This was allowed to bind for 1 hour at room temperature, with gentle shaking. Antibody was then removed, and wells washed as before.
100 μl of secondary HRP-conjugated rabbit anti-mouse P026002 antibody (Agilent) was then added, diluted 1:1000 (v/v) in PBST-BSA (3%). This was allowed to bind for 1 hour at room temperature, with gentle shaking. This was then removed and wells were washed, as before.
Following the final washes, 100 μl of ECL reagent (luminol) was added to each well, and the plate was then immediately scanned for luminescence using a Fluoroskan Ascent FL plate reader (Thermo Fisher Scientific).
The mean of the background luminescence was calculated from wells where PBS was added instead of nanobody, and then subtracted from each measured value to obtain background-adjusted luminescence. Due to variations in absolute luminescence values for different nanobodies, the values were then normalised vs the lowest and highest luminescence values for each ligand titration. Curves were then fit to the data using GraphPad Prism 9.1, using the 'One site -specific binding' pre-set, thus calculating affinity values and 95% CI.

Supplementary Tables
Supplementary Table 1