Genetically Encoded Lysine Analogues with Differential Light Sensitivity for Activation of Protein Function

Abstract

Genetically encoded unnatural amino acids are versatile tools for controlling protein function, but options for regulating multiple proteins in a single experiment are limited. Here, we report the genetic encoding of two new photocaged lysine derivatives, 1-(2-nitrophenyl)-ethyl lysine and nitrodibenzylfuranyl lysine, for sequential light-activation of protein function in live cells. Nitrodibenzylfuranyl (NDBF) caging groups have a redshifted absorbance maximum and high sensitivity to light compared to the 1-(2-nitrophenyl)-ethyl group (NPE), enabling selective decaging and protein activation. We characterized the responses of these new caged amino acids by optically triggering nuclear localization and firefly luciferase activity. The ability to selectively activate distinct proteins through simple light titration makes this a useful approach with broad applications.

Introduction

Modern biological experiments demand responsive tools for controlling multiple effectors at once. As nature tightly regulates where and when biochemical reactions take place, tools to probe key steps are highly useful when deployed with spatiotemporal control.^[1–4] To study proteins, we have developed several unnatural amino acids (UAAs) containing light-removable protecting groups to study protein function with light. These UAAs can be installed site-specifically into proteins in live cells using orthogonal translational machinery.^[5]

Photocaged UAAs containing photocleavable protecting groups are versatile tools for probing protein function because light allows for noninvasive decaging that is orthogonal to biochemical reactions found in cells. Steric encumbrance provided by bulky, photolabile protecting groups can be introduced via genetic code expansion to strategically block activity until activation with excellent spatiotemporal control. Several examples of photocaged cysteine, lysine, and tyrosine derivatives have been reported and have been used to control cell signaling,^[6] enzyme function,^[7] intein splicing,^[8] protein subcellular localization,^[9,10] and virus-host cell interactions.^[11–13] Early photocaged UAAs used the o-nitrobenzyl (oNB) group,^[14] a favored photocaging moiety in biological systems since cAMP was caged in 1977.^[15] While the oNB motif has proven useful in caging many different functionalities, modifications of the oNB scaffold have significantly improved its photophysical properties. Improvements on oNB characteristics include redshifting λ_max toward less toxic wavelengths (e.g., 405 nm), improving absorptivity (ε) and quantum yield (ϕ) for higher efficiency decaging, and improving the two-photon cross section for activation using near-infrared radiation deep inside tissues with high axial resolution.^[16]

Strategies to improve caging groups include substituting the α-position and elaborating the π-electron system. α-Substitution more than doubles the quantum yield^[17,18] and improves decaging kinetics and efficiency.^[19] An α-methylated nitrobenzyl caging group (1-(2-nitrophenyl)-ethyl; NPE) has quantum yields of 0.49–0.65 (Figure S1),^[17,18,20] as compared to 0.08 for the unsubstituted oNB group.^[21] The molar extinction coefficient for the NPE group is ~500 M⁻¹ cm⁻¹ at UV wavelengths,^[22] which is not improved from the oNB group. Further, the nitrodibenzylfuranyl (NDBF) group has an expanded π-electron system,^[22] thus improving the quantum yield to 0.67–0.73 and molar extinction coeffcient to 18,400 M⁻¹ cm⁻¹.^[22] The absorbance spectrum of NDBF^[22,23] is also red-shifted, allowing for decaging with common, nontoxic, blue/violet light sources, such as 405 nm lasers. Several syntheses of NDBF and applications on various bioactive compounds have been reported.^[24–28] Further, NDBF has been used to photocage the chemokine CCL5, yielding a tool for studying lymphocyte chemosensing.^[29] The enhanced photochemical properties of NDBF make it highly desirable for controlling protein function with spatiotemporal precision. Photocaged lysines have been used to optically control protein function in multiple ways, with key UAAs including photocaged lysine (PCK),^[9,10,30] hydroxycoumarin lysine (HCK),^[31] o-nitrobenzyl lysine (ONBK),^[32] and aminocoumarin lysine (ACK),^[33] among others. We believe the improved photochemical properties of NDBFK will provide a valuable addition to this toolset.

Results and Discussion

We hypothesized that NDBF can be activated independently of NPE due to its significantly higher absorptivity and redshifted absorbance, and that this pair of caging groups could be applied to control protein function by using genetic code expansion. To test this, we synthesized two caged lysines, NPEK and NDBFK (Figure 1A–B). The caging groups were connected via carbamate linkers as they are well-validated leaving groups for photochemical release,^[21] and N^ε-carbamylated lysine analogs are efficient substrates for PylRS enzymes used for genetic code expansion.^[34]

Figure 1.

Syntheses of NPEK and NDBFK. A) Preparation of NPEK and NPEK-OMe via acylation of Boc-Lys-OH or Boc-Lys-OMe with the p-nitrophenyl carbonate 1. B) Preparation of NDBFK via acylation of Boc-Lys-OH with the NHS carbonate 9. C) Absorbance spectra of NDBFK (blue) and NPEK (purple), as well as the ratio of NDBFK:NPEK absorbance. pNP: para-nitrophenyl; Boc: tert-butyloxycarbonyl; DMAP: 4-dimethylaminopyridine; TFA: trifluoroacetic acid; Et₃SiH: triethylsilane; DSC: N,N’-disuccinimidyl carbonate; Su: succinimidyl.

NPEK has been reported as a synthetic intermediate,^[35] and was synthesized according to this procedure (Figure 1A). Commercially available NPE-OH was activated as the p-nitrophenyl carbonate 1 in quantitative yield using p-nitrophenyl chloroformate in the presence of DMAP and Et₃N in DCM. The carbonate 1 was reacted with Boc-Lys-OH in a water/dioxane mixture using Na₂CO₃ as a base, affording the carbamate 2 in modest yield.^[35] The Boc group was removed using TFA in DCM, with Et₃SiH added as a carbocation scavenger, affording the product (NPEK^[35]) in 78% yield. The methyl ester (NPEK-OMe) was also prepared to test the impact on incorporation efficiency (Figure 1A), as methyl esters can have improved uptake and solubility.^[36–38] To prepare the methyl ester, Boc-Lys-OMe was reacted with the p-nitrophenyl carbonate 1 to afford the carbamate 3 in 13% yield, which was Boc-deprotected to NPEK-OMe in 83% yield using HCl in MeOH (generated from AcCl in MeOH^[39]).

NDBFK was prepared in 8 steps (Figure 1B). First, the NDBF caging group was synthesized based on our reported procedure,^[24] which was modified as follows for improved yields. Commercially available 4-fluoro-2-nitrotoluene was treated with N,N-dimethylformamide dimethyl acetal followed by oxidation with NaIO₄ to afford aldehyde 4 in 87% yield. The distal ring of NDBF was installed by a copper-catalyzed Ullmann coupling to ortho-iodophenol, generating the diphenyl ether 5 in 67% yield. Subsequent methylation of the carbonyl group using trimethylaluminum furnished the secondary alcohol 6 in nearly quantitative yield. Next, a Heck reaction was used to construct the dibenzofuran ring system. Previously, yields were modest,^[24] which we attributed to interference by the hydroxyl group. As silyl ethers are well-tolerated in the Heck reaction,^[40] 6 was treated with TMSCl in the presence of Et₃N to generate the silyl ether 7 in 90% yield. Heating 7 in the presence of catalytic Pd(OAc)₂ (10 mol%) and Cs₂CO₃ created the desired NDBF alcohol 8 in 81% yield (the TMS ether was cleaved during workup). To create the desired carbamate linkage to lysine, we activated alcohol 8 as the NHS carbonate 9 in modest yield (31%).^[41,42] Subsequent reaction with N^α-Boc-Lys-OH afforded 10 in 97% yield. A final Boc-deprotection with TFA in the presence of Et₃SiH provided NDBFK in 83% yield.

With both UAAs in-hand, we next estimated their relative decaging efficiencies (ε × ϕ). To determine extinction coefficients, we obtained UV-Vis absorbance spectra (Figure 1C). The absorbance of NDBFK is approximately one order of magnitude higher than that of NPEK at wavelengths commonly used for decaging. The measured extinction coefficients of NDBFK at both 365 and 405 nm are 2,535 and 287 M⁻¹ cm⁻¹, respectively, while that of NPEK are substantially lower, measuring 203 and 32 M⁻¹ cm⁻¹ for each respective wavelength (Figure S1). The quantum yield of NDBFK is expected to be 0.67–0.73,^[22] which is 10–30% greater than NPEK (0.49–0.65,^[17,18,20] see Figure S1). Thus, we reasoned that we could achieve exceptional selectivity for NDBFK over NPEK using either 365 nm or 405 nm light sources (Figure 2A).

Figure 2.

UAAs for activation of protein function in live cells. A) Structures of the UAAs and schematic of activation leveraging NDBFK’s greater absorptivity to UV and violet light. B) Reporter construct (mCherry-TAG-EGFP-HA) for genetic code expansion in mammalian cells. C) Epifluorescence micrographs demonstrating genetic encoding of NPEK and NDBFK, as evidenced by UAA-dependent EGFP expression. HEK293T cells were co-transfected using linear polyethylenimine (LPEI) with p(IPYE)HCKRS-PylT₄ and pmCherry-TAG-EGFP-HA-PylT₄ and grown in the presence or absence of NPEK or NDBFK (0.25 mM) for 48 h before imaging. Micrographs demonstrating NPEK-OMe incorporation are available in Figure S4. D) Western blots of HEK293T cells grown in the presence or absence of NPEK or NDBFK (0.25 mM) or NPEK-OMe (0.1 mM) for 48 h probing for the C-terminal HA tag to confirm UAA-dependent fusion protein expression.

To genetically encode the UAAs, we used a pyrrolysyl-tRNA synthetase (PylRS) and the corresponding tRNA^Pyl (PylT).^[34] We used the MbPylRS/PylT pair from Methanosarcina barkeri, an archaeal species; this system is orthogonal in both prokaryotes and eukaryotes including bacterial, yeast, and mammalian cells,^[10,43,44] as well as fruit flies, nematodes, frogs, mice, and zebrafish.^[45–49] This pair functions orthogonally to native tRNAs and aminoacyl-tRNA synthetases. The anticodon of PylT, CUA, enables genetic code expansion by suppressing in-frame amber stop codons (UAG).^[50] We screened multiple PylRS mutants to determine if any could accept NPEK or NDBFK as substrates.^[51] Despite previous encoding of a nitrobenzyl-based photocaged lysine and extensive applications in bacterial systems, we found that screening of PylRS mutants (Table S1) and subsequent expression of proteins containing NPEK and NDBFK failed in E. coli (Figure S2). Based on reports of potential nitro reduction^[52] and cleavage of similar nitrobenzyl groups in E. coli hosts,^[53] we instead performed the synthetase mutant screen and subsequent experiments exclusively in mammalian cells.^[52] PylRS-Y271A-L274 M, also called HCKRS or BhcKRS (Table S1), was found to be the most efficient synthetase and was previously reported to be polyspecific for sterically demanding lysine carbamates.^[31] We created an efficient mammalian version of this enzyme^[42,54] based on a reported optimized N-terminal domain with four mutations (“IPYE”)^[55] and used this for further studies.

We transfected HEK293T cells with one plasmid containing HCKRS and four copies of PylT (p(IPYE)HCKRS-PylT₄; “p” here denotes “plasmid”), and one dual reporter plasmid containing mCherry and EGFP bearing an in-frame TAG codon between the two fluorescent proteins and a C-terminal HA tag for simplified immunoblotting and four additional copies of PylT, called pmCherry-TAG-EGFP-HA-PylT₄ (Figure 2B). This two-plasmid system has previously demonstrated robust UAA incorporation.^[54] Fluorescence microscopy showed that mCherry was expressed independently of UAA addition, while EGFP was expressed conditionally in the presence of NPEK or NDBFK (Figure 2C). UAA-dependent expression of the full-length reporter protein was confirmed via western blot (Figure 2D). Testing of NPEK-OMe revealed that the UAA was cytotoxic at or above 0.25 mM (Figure S3), but follow-up testing at 0.1 mM showed good incorporation efficiency (on par with 0.25 mM free acid, Figure 2D, Figure S4). When compared directly, NPEK is incorporated with greater efficiency (at least 3-fold) as determined by the amount of EGFP fluorescence relative to mCherry fluorescence (Figure S5). This variation in incorporation efficiency is likely due to the diminished ability of the HCKRS active site to accommodate the steric bulk of NDBFK. For NDBFK, a 0.25 mM concentration is optimal and there is no added benefit to using a higher concentration (Figure S6).

Once we established conditions for genetically encoding NPEK and NDBFK, we next tested their ability to control protein function in living cells. We selected a well-validated nuclear translocation reporter, OptoNLS,^[9,42,56] which consists of EGFP linked to mCherry via a nuclear localization sequence (NLS) derived from the transcription factor SATB1. This NLS has an essential lysine residue (K29)^[57] that is required for activity that we mutagenized to NPEK or NDBFK to control localization using light (Figure 3A, Figure S7). In contrast to the previous expression reporter (red ± green), with this construct, EGFP is always expressed, and the caged NLS and mCherry are translated downstream when amber suppression is successful (green ± red). Irradiation and subsequent quantification of nuclear versus cytoplasmic mCherry fluorescence gives a functional readout of lysine decaging, as nuclear translocation can only occur with the free lysine.^[51] The full-length EGFP-mCherry construct has sufficient molecular weight to prevent passive diffusion through nuclear pores.^[9] Free EGFP (from early termination) diffuses freely throughout the cytosol and nucleus and shows the outline of the cell.

Figure 3.

Light-triggered activation of the Opto-NLS. A) Cartoon illustrating optically triggered NLS activation of an mCherry-NLS-EGFP fusion protein (OptoNLS). B, C) Time-course images showing a representative cell expressing OptoNLS-K29NPEK after irradiation (365 nm, 30 s). At 14 min, the construct is completely excluded from the nucleus, but it localizes to the nucleus within 100 min after irradiation. Scale bar, 50 μm. D, E) Kinetics of nuclear translocation for OptoNLS-K29NDBFK and OptoNLS-K29NPEK, respectively. The ratio of nuclear mCherry signal intensity (I_n) cytosolic mCherry signal intensity (I_c) was calculated, along with fits to the data using nonlinear regression with an exponential model. Mean±SEM (n=4; from different fields of view) is shown. F) Comparison of 365 nm and 405 nm irradiation times for NPEK (red) and NDBFK (green) decaging with respect to OptoNLS translocation 112 min after irradiation, with fold-change between respective UAAs shown. Mean±SEM (n=4) is shown.

HEK293T cells were co-transfected with the OptoNLS reporter in a pCDH vector^[9] and pAG-HCKRS-PylT₄, as this synthetase plasmid has been used extensively with this reporter.^[42,56] After 48 h, cells were irradiated with 365 nm or 405 nm light and were imaged every 8 minutes for 2 hours in an on-stage incubator at 37°C. Following irradiation, we observed a gradual translocation of the full-length fusion protein (marked by mCherry) to the nucleus (Figure 3B–C; see Figure S8). The response was quantified in multiple cells from different fields of view. In each condition, a plot of the natural logarithm of cytosolic fluorescence intensity (ln(I_c)) over time was linear (Figure S9), suggesting first-order kinetics, which is consistent with published results.^[58] Translocation for OptoNLS-K29NDBFK was rapid, with a t_½ of 35–47 min (Figure 3C–E, Table S2). This outperformed hydoxycoumarin lysine (HCK), a previously reported photocaged lysine with a t_½ of 54 min.^[56] All tested light exposure times gave near-quantitative nuclear translocation for OptoNLS caged with NDBFK, including a very short 1 s exposure to 365 nm irradiation (Figure 3D). Both a 15 s and 30 s exposure to 405 nm light also resulted in a complete translocation response. Overall, this shows that NDBFK decaging generates a rapid functional response in live cells and that it decages significantly faster than previously encoded caged lysines.^[9,56]

OptoNLS-K29NPEK was significantly less sensitive to irradiation (Figure 3D–E). While the extent of translocation was tunable using 30 s of 365 nm irradiation, no response was observed after a 1 s irradiation. With 405 nm irradiation, extended light exposure (over 4 min) was required for any translocation response (Figure 3E). Differential nuclear translocation responses with NDBFK vs. NPEK of 21-fold and 27-fold were achieved using 365 nm and 405 nm light, respectively (Figure 3F). Overall, the OptoNLS translocation following 365 nm or 405 nm irradiation of NPEK was slower and weaker when compared to the high sensitivity of NDBFK at either wavelength. The requirement of several-fold longer irradiation times to achieve the same translocation response for NPEK-versus NDBFK-caged proteins indicates their compatibility for sequential optical activation.

To further explore applications of these new genetically encoded UAAs, we tested whether they could control enzyme activity. Luciferases provide a sensitive readout of enzymatic activity that is quantifiable with a dynamic range spanning several orders of magnitude.^[59] We used a validated method in which we expressed a fusion protein comprising firefly luciferase (Fluc) and Renilla luciferase (Rluc).^[60] Fluc has been expressed bearing a variety of oxycarbonyl-protected lysine analogues at K529, including oNB-lysine,^[61] p-azidobenzyllysine,^[62] and TCO-lysine.^[63] This position has also been masked with enzymatically cleavable acetyl- and crotonyl-lysine.^[64] Given the versatility of the K529 position for caging via UAA mutagenesis, we chose to install NPEK and NDBFK at this position (Figure 4A; based on PDB ID 2D1S^[65]).

Figure 4.

Optical control of luciferase activity with NPEK and NDBFK. A) Model of the Fluc active site in which the K529NDBFK mutation blocks the entry of ATP until photolysis. B, C) HEK293T cells in white 96-well plates were co-transfected with p(IPYE)HCKRS-PylT₄ and pCS2-Fluc-K529TAG-Rluc using FuGENE HD. After 48 h of incubation in the presence of NDBFK (0.25 mM) or NPEK (0.25 mM), cells were washed and irradiated with a B) 365 nm or C) 405 nm LED. Cells were lysed and assayed for Fluc and Rluc activity. Mean±SEM is plotted (n=3).

To test optical control of Fluc, we doubly transfected HEK293T cells with p(IPYE)HCKRS-PylT₄ and pFluc-K529TAG-Rluc in media supplemented with NPEK or NDBFK (each 0.25 mM). After 24 h, the cells were washed and irradiated with a 365 nm or 405 nm LED. Fluc-K529NDBFK-Rluc is highly sensitive to 365 nm light, with noticeable activation after a 0.1 s exposure and over 50% activation after just 0.25 s of irradiation (Figure 4B). Irradiating the cells for 3 s was sufficient to activate the protein maximally. The response of Fluc-K529NDBFK-Rluc to 405 nm irradiation, as expected based on the lower molar extinction coefficient compared to 365 nm (Figure 1C, Figure S1), is less sensitive, requiring at least 60 s for 50% activation (Figure 4C). The lower restoration of Fluc function compared to Opto-NLS reflects the 75% lower power used for irradiation in this experiment. As we aimed to find the range over which enzyme activity was tunable, we demonstrated the lower limit of irradiation power required thus showcasing the sensitivity of NDBFK. These experiments were performed in white plates to maximize photon reflection toward the luminometer during the luciferase assay, which is also expected to increase the amount of light reaching samples during irradiation for decaging.

Fluc-K529NPEK-Rluc was much less sensitive to light (Figure 4B–C). To achieve only 50% of maximal activation, irradiation times for 365 nm and 405 nm were 12 s and 120 s, respectively. With only 1–3 s of 365 nm irradiation Fluc-K529NDBFK-Rluc was decaged and subsequently activated, but Fluc-K529NPEK-Rluc demonstrated negligible decaging or activation in that time frame. Further, 405 nm irradiations ranging from 30–60 s thoroughly decaged NDBFK and minimally affected NPEK. Additional testing with further redshifted light sources showed that selective NDBFK activation at 415 nm or 425 nm is also possible (Figure S10). As expected from the diminishing absorptivity at these longer wavelengths (Figure 1C, Figure S1), selectivity for NDBFK is not as high as it is for 365 nm or 405 nm illumination (only 3- to 6-fold as compared to over 10-fold).

Conclusions

New photocaged lysine derivatives NPEK, NPEK-OMe, and NDBFK were synthesized and genetically encoded in mammalian cells using an optimized HCKRS mutant. We applied these caged UAAs to control protein nuclear localization and enzyme function. We showed that the extent of protein activation may be tuned by varying the duration or intensity of irradiation dependent on the UAA selected.

The successful genetic encoding of both UAAs further establishes the versatility of the HCKRS mutant of PylRS,^[31,42,54] expanding the substrate scope to include a sterically demanding tricyclic UAA (NDBFK). This large UAA may be used to cage challenging targets, such as large active sites and large surfaces defining protein-protein interactions, more effectively than existing UAAs. Compared to its predecessors, NDBFK decaging activates protein function with the fastest reported kinetics.^[9,31] As it has the highest absorptivity and extinction coefficient yet reported for a genetically encoded photocaged UAA, it decreases the amount of light required for these experiments. We expect this will minimize photodamage, make the approach more practical in deeper or more opaque tissues, and will enable studies of very fast biological processes. Further, the high two-photon cross section for the NDBF group^[22] suggests that labs may activate NDBFK with near-infrared sources for enhanced axial resolution and deeper tissue penetration,^[4] adding to the limited repertoire of two such UAAs.^[31]

NPEK and NPEK-OMe incorporation was improved over that of NDBFK, but required longer irradiation times for sufficient decaging and protein activation. While not directly improving on the properties of other photocaged lysines (HCK, ACK, mNPK, etc.), this novel UAA adds an additional tool for use in complex protein systems that require precise, spatiotemporal activation. Further, NPEK is favored over NPEK-OMe, as equivalent expression was observed in the presence of either UAA at 0.25 mM, but the increased cytotoxicity at high concentrations of NPEK-OMe limited its application. NPEK’s efficient and robust incorporation efficiency and distinct spectral sensitivities from NDBFK and other UAAs make NPEK useful on its own or in combination with other UAAs.

Due to its enhanced photophysical properties compared to most reported photocaged UAAs, we anticipate that NDBFK can be activated sequentially with other photocaged lysine, tyrosine, cysteine, and serine analogues.^[11] The broad range of wavelengths for optical triggering, good photochemical properties, and ability to control multiple protein targets sequentially makes these UAAs useful for further studies in cells and animals.

Experimental

Genetic encoding of NPEK and NDBFK in mammalian cells (mCherry-TAG-EGFP-HA reporter) – imaging.

HEK293T cells (American Type Culture Collection) were seeded at 20,000 per well into a PdK-treated (MP Biomedicals, 70–150 kDa) black transparent-bottom 96-well plate (Greiner Bio-One μClear). Cells were grown (37°C, 5% CO₂) in Dulbecco’s Modified Eagle’s Medium (DMEM; GE Life Sciences) supplemented with fetal bovine serum (FBS, 10% (v/v), Sigma-Aldrich), penicillin (100 UmL⁻¹, Corning Cellgro), and streptomycin (100 μgmL⁻¹, Corning Cellgro) until reaching ~75% confluency (16–24 h). Cell culture media was replaced with DMEM (180 μL, no antibiotics) supplemented with FBS (10% (v/v)) in the presence or absence of UAA (NPEK (0.25–1 mM), NDBFK (0.25 mM), or NPEK-OMe (0.1 mM)). To transfect the cells, 1 mg mL⁻¹ linear polyethylenimine (LPEI, Polysciences^[66]) was diluted to 0.33 mg mL⁻¹ with prewarmed Opti-MEM media (Gibco). The resulting solution (2 μL/well) was added to solutions of plasmid DNA (100 ng of each plasmid) in Opti-MEM media (18 μL). pmCherry-TAG-EGFP-HA-PylT₄^[37] was used as the reporter. p-(IPYE)HCKRS-PylT₄^[42,54] was used to provide the necessary synthetase and additional suppressor tRNA. The mixture was pipetted up and down five times, and was incubated for 15 minutes at ambient temperature. The transfection reagent (20 μL) was then added dropwise to each well by dispensing a small droplet of the reagent onto the pipette tip, then gently touching it to the surface of the media in the center of the well. The cells were grown for 48 h (37°C, 5% CO₂) after transfection. Media was removed and replaced with prewarmed Live Cell Imaging Solution (Molecular Probes/Invitrogen). The fusion protein was visualized by epi-fluorescence microscopy. For all imaging, one of two Zeiss Axio Observer Z1 microscopes was used with a 20× objective (numerical aperture 0.8 Plan-Apochromat M27). For NPEK and NDBFK, a mercury arc lamp (HBO 100 W) was used with the following filters: for mCherry, excitation (ex), BP550/25, emission (em), BP605/70; and for EGFP ex, BP470/40, em, BP525/50. Images were captured with an AxioCam MRm camera on this microscope. For NPEK-OMe, a microscope with an Excelitas X–Cite 120 LED Boost was used with the following filters: for GFP, ex ET 470/40, em ET 525/50; and for mCherry, ex BP 550/25, em 605/70. Images were captured with an Andor Zyla 4.2 camera on this microscope.

Genetic encoding of NPEK and NDBFK – western blot.

HEK293T cells (American Type Culture Collection) were seeded at 160,000 per well into PdK-coated (MP Biomedicals, 70–150 kDa) clear 6-well plates (Greiner Bio-One) and grown for 18 h (37°C, 5% CO₂) to ~80% confluency in DMEM (2 mL, no antibiotics) supplemented with FBS (10% (v/v), Sigma-Aldrich), penicillin (100 UmL⁻¹, Corning Cellgro), and streptomycin (100 μgmL⁻¹, Corning Cellgro). Cell culture media was replaced with DMEM (1.8 mL) supplemented with FBS (10% (v/v)) in the presence or absence of the UAA (NPEK (0.25 mM), NPEK-OMe (0.1 mM), or NDBFK (0.25 mM)). To transfect the cells, 1 mgmL⁻¹ linear polyethylenimine (LPEI, Polysciences) was diluted to 0.33 mgmL⁻¹ with prewarmed Opti-MEM media (Gibco). The resulting solution (10 μL) was added to solutions of plasmid DNA (1.5 μg of each plasmid) in Opti-MEM media (200 μL). pmCherry-TAG-EGFP-HA-PylT₄^[54] was used as the reporter. p-(IPYE)HCKRS-PylT₄^[42,54] was used to provide the necessary synthetase and additional suppressor tRNA. The mixture was pipetted up and down five times, and was incubated for 15 minutes at ambient temperature. The transfection reagent mixture was then added to the cells (180 μL per well). The reagent mixture was added by dispensing a small droplet out of the tip of the pipette tip and touching it to the surface of the media over the cells. The reagent droplets were added to the surface all around the well to distribute it evenly. Once the reagent was added to all wells, the plate was gently swirled for 5 seconds and placed in the incubator (37°C, 5% CO₂). After 48 h, cells were cooled on ice and washed with ice-cold PBS (2×1 mL). The cells were lysed in cold mammalian protein extraction buffer (250 μL) (GE Life Sciences) supplemented with Halt Protease Inhibitor Cocktail (Thermo Scientific) on ice with orbital shaking for 20 min. The lysed cells were scraped from the plates using P1000 pipette tips. Lysates were pipetted into micro-centrifuge tubes, and were clarified by centrifugation (21,000 rcf, 20 min, 4°C).

For the immunoblots, cell lysate supernatants (30 μL) were mixed with 6× Laemmli sample buffer (6 μL) in PCR tubes and heated to 95°C for 5 min. They were loaded (32 μL/well) into 1.5 mm 10% polyacrylamide gels alongside PageRuler molecular weight marker (Thermo Scientific). Empty wells were filled with PBS (32 μL). Gels were run at 60 V for 20 min followed by 120 V for 70 min. Proteins were transferred to PVDF membranes (Immobilon-P, EMD Millipore) for 2 h and 45 min at 75 V at 0°C. The membranes were blocked with 5% (w/v) powdered nonfat milk in TBS-T for 1 h at room temperature. Membranes were rinsed with TBS-T (3×5 min) and incubated in anti-GAPDH (Santa Cruz Biotechnology, Dallas, TX, sc365062, mouse mAb, diluted 1:2,000 in 5% (w/v) milk in TBS-T) or anti-HA (Cell Signaling Technology, Danvers, MA, C29F4, rabbit mAb, diluted 1:1,500 in 5% (w/v) BSA in TBS-T) antibody solutions overnight at 4°C. Membranes were washed with TBS-T (5×5 min) and incubated with secondary antibodies for 1 h at ambient temperature (GAPDH: goat anti-mouse IgG-HRP, Santa Cruz Biotechnology, sc-2005, diluted 1:20,000 in TBS-T; HA: goat anti-rabbit IgG-HRP, Santa Cruz Biotechnology, 7074S, diluted 1:5,000 in TBS-T). Membranes were washed with TBS-T (5×5 min) and were imaged with West Pico Chemiluminescence Substrate (Thermo Scientific) using a ChemiDoc XRS+ system with Image Lab 6.0 software (Bio-Rad) in high resolution mode (2 × 2 binning).

Optically triggered SATB1 Translocation – imaging.

HEK293T cells (American Type Culture Collection) were seeded at 20,000 per well into a PdK-treated (MP Biomedicals, 70–150 kDa) black μClear transparent-bottom 96-well plate (Greiner Bio-One). Cells were grown for 18 h (37°C, 5% CO₂) in DMEM (180 μL, no antibiotics) supplemented with FBS (10% (v/v), Sigma-Aldrich), penicillin (100 UmL⁻¹, Corning Cellgro), and streptomycin (100 μgmL⁻¹, Corning Cellgro) until reaching ~75% confluency. Media was then changed to DMEM +10% FBS supplemented with FBS (10% (v/v)) and NPEK (0.25 mM) or NDBFK (0.25 mM). To transfect the cells, 1 mgmL⁻¹ linear polyethylenimine (LPEI, Polysciences) was diluted to 0.33 mgmL⁻¹ with prewarmed Opti-MEM media (Gibco). The resulting solution (2 μL/well) was added to solutions of plasmid DNA (200 ng total per well, with equal amounts of each plasmid) in Opti-MEM media (18 μL). The cells were transfected with pCDH-EGFP-OptoNLS-K29TAG-SATB1-mCherry^[9] and pAG-HCKRS-4PylT (also referred to using “BhcKRS”).^[31] After 40 h of incubation, cells were incubated in 3 changes of FluoroBrite DMEM (Gibco) supplemented with l-glutamine (Frontier Scientific, 2 mM) in a 37°C/5% CO₂ incubator to wash excess UAAs out of the cells (200 μL each time, 2 × 30 min wash, then a 10 h incubation overnight). The cells were then placed on a WSKM Stage Top Incubator (Tokai Hit) on a Zeiss Axio Observer Z1 microscope where they were maintained at 37°C in a humidified, 5% CO₂ atmosphere. Cells were irradiated with a 365 nm or 405 nm LED at 100% power through an aluminum foil mask with a circular cutout the size of one well. A 40x air objective (numerical aperture 0.6 Plan-Neofluar) and Excelitas X–Cite 120 LED Boost light source was used with the following filters: for GFP, ex ET 470/40, em ET 525/50; for mCherry, ex BP 550/25, em 605/70. Images were captured with an Andor Zyla 4.2 camera.

Representative fields of view were set as positions in Slidebook 6 software (Intelligent Imaging Innovations/3i). Cells were irradiated while in the on-stage incubator through an aluminum foil mask and plate cover. The LEDs were shined through an aluminum foil mask with a cutout the size of a single well on top of the plate cover. The incubator was maintained at 37°C with the lid open and CO₂ was temporarily turned off to avoid wasting gas. The 365 nm LED was used at 0.60 A current (max. pulse length: 10 s), and the 405 nm LED at 0.65 A current (max. pulse length: 30 s). LEDs were held 1–2 mm above the mask, which was directly on top of the plate cover. Wells were irradiated one pulse at a time in a sequence so that each well would have ≥1 min between pulses to prevent overheating. Time course images were acquired (every 8 minutes for 2 hours) using a Zeiss Definite Focus unit. Images were analyzed as described in the Supplementary Methods.

Optical triggering of luciferase activity with photocaged lysine analogues.

HEK293T cells (American Type Culture Collection) were seeded at 20,000 per well into a PdK-coated (MP Biomedicals, 70–150 kDa) white transparent-bottom (Greiner Bio-One μClear) 96-well plate. Cells were grown for 18 h (37°C, 5% CO₂) in DMEM (GE Life Sciences) supplemented with FBS (10% (v/v), Sigma-Aldrich), penicillin (100 UmL⁻¹, Corning Cellgro), and streptomycin (100 μgmL⁻¹, Corning Cellgro) until reaching ~80% confluency. Media was replaced with DMEM+10% FBS (130 μL, no antibiotics) supplemented with NPEK (0.25 mM) or NDBFK (0.25 mM). To transfect the cells, a transfection reagent mixture was prepared using FuGENE HD (Fugent LLC). FuGENE HD (0.4 μL/well) was added to a mixture of p(IPYE)HCKRS-PylT₄^[42,54] (55 ng/well) and pCS2-Fluc-K529TAG-Rluc^[33] (55 ng/well) in Opti-MEM (Gibco, 5 μL/well). The mixture was incubated at room temperature for 30 min before adding it dropwise to the cell culture media (5 μL/well).

Cells were incubated for 48 h (37°C, 5% CO₂) after transfection. Cells were washed by incubating them in 3 washes of FluoroBrite DMEM (Gibco, 200 μL, 3×30 min). Cells were irradiated with a 365 nm or 405 nm LED at 25% power (0.15 A or 0.17 A, respectively). A black notebook was placed under the plate to minimize light reflection into unintended wells. The LEDs were shined through an aluminum foil mask with a cutout the size of a single well on top of the plate cover from the top. The 365 nm LED were used in ≤15 s pulses in all experiments to avoid overheating the cells. The 405 nm LED, which generated less heat, were used in ≤30 s pulses. Wells were irradiated with pulses in sequence so that each well could cool for ≥1 minute between pulses. To lyse the cells, media was removed and replaced with warm PBS (200 μL/well). The PBS was removed and cells were lysed by adding freshly diluted 1× Passive Lysis Buffer (Promega, 20 μL/well) with a repeating pipette (Eppendorf Repeater Model 4780). The plate was placed into a Tecan Infinite M1000Pro plate reader with iControl software (Tecan) and was equilibrated to 30°C for 10 min using the “Heating” function. Luminescence Assay Reagent II from a Dual-Luciferase Reporter Assay System kit (Promega E1980) was added using the instrument’s “Inject” function, one well at a time (35 μL, 100 μL/s injection speed), and Fluc luminescence was read 2 seconds later (200 ms integration time, automatic attenuation mode). After the Fluc measurements were completed, Stop-and-Glo reagent from the kit was added using the instrument’s “Inject” function, one well at a time (35 μL, 250 μL/s injection speed for rapid mixing), and luminescence was read 2 seconds later (200 ms integration time, automatic attenuation mode). Fluc/Rluc ratios were calculated in Microsoft Excel. Data were plotted using Graphpad Prism 9. Significant outliers were excluded by Grubb’s test (GraphPad Quickcalcs, https://www.graphpad.com/quickcalcs/grubbs1/, α=0.05).

Supplementary Material

Supporting information for this article is available on the WWW under https://doi.org/10.1002/cptc.202300312

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.