Generalized bioluminescent platform to observe and track cellular interactions

Abstract

Cell-to-cell communications are critical to biological processes ranging from embryonic development to cancer progression. Several imaging strategies have been developed to capture such interactions, but many are challenging to deploy in thick tissues and other complex environments. Here, we report a platform termed Luminescence to Observe and Track Intercellular Interactions (LOTIIS). The approach features split fragments of a luciferase enzyme that reassemble when target cells come into proximity. One fragment is secreted by “sender” cells, and the complementary piece is secreted by “receiver” cells. Split reporter assembly is facilitated by a single chain variable fragment (scFv)-peptide interaction on the receiver cell, resulting in localized light production. We demonstrate that LOTIIS can rapidly label cells in close proximity in a time- and distance-dependent fashion. The platform is also compatible with bioluminescence resonance energy transfer probes for multiplexed imaging. Collectively, these data suggest that LOTIIS will enable a variety of cellular interactions to be tracked in biological settings.

Graphical Abstract

INTRODUCTION

Cellular interactions underlie diverse biological processes, ranging from tissue development to disease progression.^1–3 Understanding how cells communicate will thus not only advance our understanding of organismal biology, but also provide new avenues for therapeutic intervention. Information can be exchanged between cells via direct contact or over greater distances via secreted small molecules and other bioactive ligands.^4,5 The information flow varies depending on the cell types involved and their location within an organism. Decoding cellular communication requires methods to identify when and where cells interact. However, tools to illuminate cellular contacts within complex networks, particularly within heterogeneous cultures, are limited.

To date, several systems for visualizing cellular contacts have been developed that rely on fluorescent protein fusions. Included in this group are methods that leverage split reporters,⁶ such as the green fluorescent protein (GFP) reconstitution across synaptic partners^7,8 (GRASP) approach. GRASP features fragments of split GFP that are fused to truncated transmembrane domains, anchoring them to the surfaces of defined cell populations. Complementation of the split reporter occurs when the cells of interest are in direct contact. Light emission thus provides a readout on the interacting cells. Recently, a new variant of this system (known as GPI anchored reconstitution-activated proteins highlight intercellular connections or GRAPHIC),⁹ has been reported. GRAPHIC relies on generic anchoring proteins and a brighter split GFP probe. While GRASP and GRAPHIC have enabled cell contacts to be profiled, they can perturb endogenous cell function and motility. The membrane-bound anchors can fuse interacting cells together for extended time periods. Moreover, while these fluorescence-based platforms enable high-resolution imaging, they can be difficult to employ for longitudinal studies due to phototoxicity and photobleaching effects.

We aimed to address these limitations by developing a bioluminescent-based reporter of cellular interactions. Bioluminescence can be advantageous for monitoring cellular interactions over time, as it does not rely on excitation light for photon production.^10,11 Rather, bioluminescent light derives from a chemical reaction: the luciferase-catalyzed oxidation of small molecule luciferins. Luciferase-luciferin pairs have been widely used to image cell populations on a macro scale in tissues and whole organisms. The lack of required excitation light mitigates photobleaching effects that often occur during serial imaging sessions with fluorescent molecules. Thus, bioluminescent probes have been used extensively to visualize cellular events over time, such as cell proliferation and motility.^12–14 Split luciferase reporters have also been developed and used for examining cellular interactions.¹⁵ However, these probes were not localized to target cells and could only indirectly report on interacting cells.

Here we report a more general and traceable split luciferase strategy for monitoring cell-cell contacts, known as Luminescence to Observe and Track Intercellular Interactions (LOTIIS). LOTIIS relies on the split-NanoLuc platform (NanoBiT, NBiT)¹⁶ that comprises a small BiT (SmBiT) peptide and large BiT (LgBiT) fragment. NBiT has been widely used to monitor protein-protein and other biomolecule interactions.^10,17–19 We aimed to repurpose this technology for reporting on cell-cell interactions. The split reporter was localized and reconstituted on target cell surfaces without forcing prolonged interactions, which is critically important for visualizing contacts and downstream behaviors as they occur. We leveraged the multimeric SunTag system to further enhance signal production.²⁰ SunTag features 24 copies of a GCN4 leucine zipper motif from yeast²¹; each GCN4 unit can be recognized by a specific single chain variable fragment (scFv), enabling multiple copies of a detectable probe to be recruited to the tag. SunTag has previously been used to enhance signal outputs for single molecule fluorescence imaging among other applications.^22,23 For LOTIIS, we co-opted the GCN4 motifs to localize NBiT to the surface of “receiver” cells after interacting with “sender” cells, and amplify the overall signal. The platform was optimized based on the localization and expression of the requisite engineered fragments, and the intensity of bioluminescent outputs. We also showcased the generality of LOTIIS for multiplexed imaging and tracking a range of interactions over time.

RESULTS AND DISCUSSION

Design of LOTIIS

As a starting point for tagging and tracking cellular interactions via LOTIIS, we aimed to prepare “sender” and “receiver” cells featuring the scFv-epitope pair from the SunTag system. The sender cells were envisioned to secrete scFv fused to the SmBiT fragment. The receiver cells would secrete the complementary LgBiT piece, in addition to the multimerized GCN4 epitope (localized to the cell membrane, Figure 1A). In this scenario, NBiT assembly was hypothesized to occur in the extracellular space, with reporter complementation driven by the inherent affinity of the two components. LOTIIS would thus leverage the high affinity interaction of the scFv to localize NBiT to the surface of a receiver cell, with NBiT reconstitution driven by the high local concentration of LgBiT and SmBiT. Therefore, cellular labeling would result from a two-step process dictated by the binding affinity of the scFv, which remains constant, and the complementation affinity of the split reporter. The two-step process would ensure that signal remains low in the absence of cellular interactions.

Figure 1.

Labeling and tracking cell-cell interactions with bioluminescence. A) Strategy for proximity tagging using NBiT and SunTag. scFv-SmBiT fusions are secreted from sender cells and localize to multimerized epitope scaffolds on receiver cell surface. High local concentrations of LgBiT secreted from receiver cells drive NBiT complementation. NBiT catalyzes light emission with luciferin addition. B) Parameters evaluated in LOTIIS: Effect of self-complementation affinity (K_D values from sub-nM to μM) on signal intensity and labeling efficiency; adapting platform for use with other split reporters (see Figure S3 for split GFP design); multimerization of the peptide scaffold to enhance signal; developing BRET constructs for multiplexed imaging.

We envisioned that LOTIIS would have several beneficial features (Figure 1B). First, the strength of the signal could be tuned via split reporter binding affinity. A variety of SmBiT peptides (with varying K_D values) have been reported¹⁶ and could potentially be used to tune signal intensity. Lower affinity peptides would provide the most stringent distance-dependent signal; higher affinity interactions would provide greater coverage. Second, because reporter complementation is independent of the scFv-peptide interaction, LOTIIS could be used in conjunction with other split-reporters for modular imaging. Third, the system is amenable to multimerization to improve signal-to-noise ratios. Last, to visualize complex multicellular interactions, a multiplexed platform is required. We envisioned that such a platform could be readily achieved via LOTIIS, using fluorescent protein (FP) fusions for bioluminescence resonance energy transfer (BRET). This strategy would enable multiple cell populations to be observed simultaneously via spectrally resolved BRET signatures.

Identification of an optimal LOTIIS configuration

We first investigated the optimal sender/receiver configuration for LOTIIS based on the placement of NBiT fragments (Figure 2A). Both an epitope tag (HA) and linker (G₄S_x4, 20 a.a.) were inserted between the scFv and NBiT components for downstream confirmation of fragment localization and secretion. The labeled cell, termed “receiver”, localizes the multimerized GCN4 peptide-scaffold (two tandem copies) on the cell surface via a secretion signal. The scaffold is anchored to the membrane through a truncated CD4 transmembrane domain. A FLAG tag was inserted between the epitopes and CD4 transmembrane domain to verify expression and localization of the probe. In one configuration, LgBiT was secreted from the receiver cells (GCN4-LgBiT) via fusion to a synthetic secretion tag.²⁴ A P2A peptide was also used to separate the NBiT component through co-translational cleavage.²⁵ Correspondingly, SmBiT was fused to the scFv (scFv-SmBiT) and secreted from sender cells. The second configuration involved secretion of SmBiT from receivers (GCN4-SmBiT) with scFv-LgBiT secreted from senders. In both configurations, when a sender is in proximity of a receiver, scFv–peptide binding should occur, localizing one fragment to the receiver cell surface. High local concentrations of the cognate NBiT fragment secreted by the receiver drive the assembly of a functional luciferase, providing detectable signal.

Figure 2.

Determining the optimal LOTIIS configuration. A) Two possible strategies were envisioned as shown. The sender component comprised a signal peptide (SP), scFv, HA-tag, and NBiT fragment. The receiver component comprised a SP, GCN4 epitopes, FLAG-tag, truncated CD4 transmembrane (TM) domain and a secreted NBiT fragment. B) Sender and receiver cells were plated separately or co-cultured for 24 h. Media was removed, and cells were washed with PBS prior to lifting and imaging with furimazine. C) Media was collected from each sender and receiver cell line and mixed in a 1:1 ratio as indicated. Media was then imaged with furimazine. D) Sender and receiver cells were plated in equal numbers and incubated for 48 h. Cells were then lifted and lysed using a small gauge needle. Cell lysates were mixed in a 1:1 ratio as indicated. Lysates were then imaged with furimazine. For C-D, error bars represent the SEM for n = 9 replicates. ****, p < 0.0001

HEK293T cells stably expressing each sender-receiver pair (GCN4-LgBiT/scFv-SmBiT and GCN4-SmBiT/scFv-LgBiT) were generated. GCN4 expression was verified by flow cytometry, staining receiver cells with α-FLAG antibodies (Figure S1A). GCN4-SmBiT cells expressed higher levels of the binding epitope, likely due to the overall smaller construct size. scFv–peptide interactions on the cell surface were evaluated by adding media from individually cultured senders onto either HEK293T cells or receivers. Media was removed and cells were labeled with α-HA conjugates, then analyzed by flow cytometry (Figure S1B). Fluorescent signal was enhanced only in the case of receiver cells. Additionally, flow cytometry analyses revealed increased binding of scFv to GCN4-SmBiT receivers, likely due to increased expression of the membrane epitopes. We next evaluated the ability of both sender-receiver pairs to provide bioluminescent signal upon cellular interaction. Sender and receiver cells were cultured individually or together, then assayed for NBiT activity. A significant signal turn-on (>130-fold) was observed with GCN4-LgBiT/scFv-SmBiT co-cultures whereas minimal signal (< 3-fold) was observed with GCN4-SmBiT/scFv-LgBiT co-cultures, despite the latter exhibiting increased membrane epitope expression (Figure 2B).

We hypothesized that the discrepancy in cell labeling could be due to poor secretion of the SmBiT peptide from GCN4-SmBiT receivers. To examine this possibility, cell media was collected from each sender and receiver line and mixed in a 1:1 ratio (Figure 2C). When media from scFv-SmBiT-expressing cells was mixed with media from either GCN4-LgBiT- or scFv-LgBiT-expressing cells, robust signal was observed. However, no signal was observed when media samples from GCN4-LgBiT- and GCN4-SmBiT-expressing cells were mixed. These results suggest little to no functional SmBiT peptide was being produced by GCN4-SmBiT receiver cells. To identify if the lack of functional SmBiT peptide was due to poor secretion or overall low expression, each sender and receiver cell line was collected and lysed using a 30-gauge needle. Cell lysates were then mixed in a 1:1 ratio in all combinations (Figure 2D). Similar to the mixed cell media experiment, robust signal was observed when lysate from scFv-SmBiT cells was mixed with that from GCN4-LgBiT or scFv-LgBiT cells. However, no complementation was observed when GCN4-LgBiT cell lysate was mixed with GCN4-SmBiT lysate. These data indicate that SmBiT peptide was being poorly expressed and that the scFv-LgBiT/GCN4-SmBiT sender-receiver pair is not suitable for imaging cellular interactions. Free SmBiT peptide could be prone to proteolysis in a cellular environment and unable to freely complement with LgBiT. By contrast, the scFv-SmBiT/GCN4-LgBiT sender-receiver pair provided robust signal enhancement in the presence of interacting partners, providing a viable platform.

LOTIIS reports on interactions of variable affinity

Moving forward, we optimized the scFv-SmBiT/GCN4-LgBiT pair by modulating the binding properties of the NBiT fragments. LOTIIS relies on the inherent affinities of the split pieces to drive complementation. We evaluated reporter complementation using previously described SmBiT peptide fragments. High affinity peptides would be expected to complement at lower concentrations and shorter time scales, but potentially lead to higher levels of background complementation in free solution. Lower affinity peptides, by contrast, would provide the most stringent readout on cellular interactions. Higher local concentrations of such peptides would be required to drive complementation - concentrations that would only be achieved by cells in close proximity. If the affinity is too low, though, no complementation and poor signal would result.

To examine the above scenarios, we generated two additional sender cell lines expressing SmBiT peptides with different affinities for LgBiT. In addition to the previously tested “high-affinity” sender (K_D=0.7 nM, sender^high), we generated medium-affinity (K_D=180 nM, sender^med) and low-affinity (K_D =190 μM, sender^low) senders. Each sender cell line was co-cultured with receiver cells and assayed for NBiT activity after 24 h. As shown in Figure 3A, robust signal was observed only on receiver cells co-cultured with sender^high or sender^med cells. Similar results were observed when media from each sender cell line was added to receiver cells. We further showed that LOTIIS with sender^high cells is capable of illuminating small populations of cells (Figure S2). Larger populations of receiver cells can also be labeled using a comparatively small population of senders. This feature could prove useful when investigating interactions among rare cell populations.

Figure 3.

Reporter affinity modulates signal intensity and labeling rate A) Receiver cells were cultured with either sender cells stably expressing high, medium, or low affinity SmBiT-scFv fusions, or media from the sender cells. Co-cultured cells were incubated for 24 h prior to imaging. Cells treated with media samples were incubated for 1 h prior to imaging. B) Receiver cells were co-cultured with sender^high or sender^med cells for 0-48 h. Media was removed and re-plated in 96 black-well plates prior to addition of furimazine and imaging. Total flux values for the co-cultured cells over time are plotted. The inset highlights the 0-5 h region. C) High, medium, or low affinity sender cells were co-cultured with receiver cells stably expressing 2x, 5x, or 10x copies of the GCN4 epitope on the cell surface. For A-C, error bars represent the SEM for n = 9 replicates. ****, p < 0.0001; ***, p < 0.001; **, p < 0.01; *, p < 0.05

Cellular interactions can occur over a range of timescales, with some being very short-lived. Methods that can rapidly report on cell proximity are therefore desirable. To evaluate LOTIIS in this context, we examined the effect of complementation affinity on the rate of signal production. Sender^high cells and receivers (50,000 each) were co-cultured, and the samples were assayed for bioluminescence over time. Robust signal was observed 1 h post-plating (Figure 3B). Signal increased over time, before tapering off after 8 h. By contrast, when sender^med cells were co-cultured with receivers, signal production was delayed. Detectable photon outputs were only observed 4 h post-plating. Prolonged incubation was likely necessary to build up sufficient concentrations of the lower-affinity fragments to drive complementation.

To showcase the generality of LOTIIS, we exchanged NBiT with the split-GFP system (Figure S3A).⁶ GFP signal was visualized only on receivers when co-cultured with senders. However, very few receiver cells were visibly labeled using the split-GFP system (Figure S3B). Similar observations have been made with other platforms, such as GRASP and GRAPHIC, likely due to the low binding affinity of split-GFP.²⁶ These results suggest that while LOTIIS is compatible with other reporters, additional tuning or dimerizing partners will likely be necessary for split variants with larger (~micromolar) K_D values.

We further recognized that interaction-dependent signal could be enhanced via multimerization. In previous work, SunTag was used to amplify intracellular targets via 24 copies of the GCN4 epitope.²⁰ We examined whether this same concept could be applied extracellularly to amplify the NBiT luminescent signal. Receiver cell lines stably expressing up to 10 additional copies of GCN4 were generated and co-cultured with sender cells. Interestingly, signal amplification was not observed with additional GCN4 copies (Figure 3C). The inverse correlation between GCN4 copies and overall luminescence could be a result of several factors: low expression, poor secretion and surface localization, or improper folding in the extracellular space, among others. To evaluate these factors, receiver lines were labeled with α-FLAG antibodies and analyzed via flow cytometry. In line with the bioluminescent data, α-FLAG signal decreased with increasing GCN4 copies (Figure S4A). A similar trend was observed when receiver lines were incubated with sender^high media and subsequently immunostained for HA epitopes (from receiver-bound scFv) or co-stained (α-FLAG, α-HA) for both GCN4 scaffold and bound scFv (Figure S4B). These results suggest that the larger GCN4 constructs are poorly secreted or surface localized. The SunTag scaffold, which was optimized for expression and folding intracellularly, thus requires additional engineering for extracellular multimerization.

LOTIIS labels cell in a time and distance dependent manner

A platform for tracking cellular interactions should provide signal when two cells of interest come into proximity. Since LOTIIS relies on fragment secretion and concentration-dependent assembly, we reasoned that signal production would be both distance- and time-dependent. To test this hypothesis, we seeded sender^high cells and receivers at defined distances using biocompatible stencils (Figure 4A). Cells were placed in the cutout wells and allowed to adhere. Cells were then washed with PBS to remove existing SmBiT and LgBiT fragments secreted into the wells. Fresh media was subsequently added to the cultures, and interactions were monitored. Robust bioluminescent signal was observed in co-cultured cells after 1 h of incubation (Figure 4B-C). By contrast, there was a delay in signal production when sender^high cells and receivers were separated at defined distances. Even at the closest distance of 1 mm, there was a delay of 1 h in signal production compared to co-cultured cells. Given enough time, LOTIIS was able to track cells up to 7 mm apart. It is key to note that the signal on labeled receiver cells decreased rapidly in the absence of senders supplying a necessary fragment (Figure S5). Additionally, co-cultured cells were still brighter than cells separated at a distance. Together, these data indicate that LOTIIS is optimal for labeling transient cellular interactions on the hour time scale. LOTIIS is also not limited to tracking direct cellular interactions. The platform can be applied to visualizing cells within a given proximity to a population of interest. Such a feature could prove useful in probing tumor-immune interactions, for example, to understand if immune cell populations have directly penetrated the tumor or simply reside at the periphery.

Figure 4.

LOTIIS is distance and time dependent. A) Schematic of procedure using PDMS stencils to plate cells at defined distances. Sender and receiver cells (50,000) were seeded in the left and right wells, respectively. Cells were incubated for 1 h to adhere to the plate. Cells were washed with PBS (600 μL) before addition of media (400 μL) to report on cell interactions. B) Sender^high cells and receivers were co-cultured or seeded 1-, 3-, or 7 mm apart. Samples were incubated for 0-4 h post-media addition. Media was removed and cells washed with PBS before addition of furimazine and imaging. Representative bioluminescent images are shown. C) Quantification of bioluminescent outputs from (B). Error bars represent the SEM for n = 9 replicates. ****, p < 0.0001; **, p < 0.01; *, p < 0.05

Development of split-BRET probes for multiplexed imaging

One advantage of the LOTIIS platform is that it is highly modular and amenable to tracking multiple cell populations. Such platforms could be applicable in heterogenous environments, such as the tumor microenvironment. We sought to multiplex LOTIIS via bioluminescence resonance energy transfer (BRET), resulting in different colors of emission from NBiT conjugates. BRET reporters comprising FPs fused to the N-terminus of NanoLuc (Nluc), have been engineered to enable both multiplexed imaging and red-shifted emission.^27–30 Building on this concept, we aimed to substitute Nluc with NBiT, which is split close to the C-terminus of the luciferase. We fused FPs from reported BRET constructs to LgBiT and preserved the FP truncations and linkers (Figure 5A). We generated receiver lines expressing derivatives of CeNL, GeNL, and YeNL from the enhanced NanoLantern series²⁸ as well as a derivative of LumiScarlet.²⁹ These receivers were termed CeLOTIIS, GeLOTIIS, YeLOTIIS, and SeLOTIIS, respectively.

Figure 5.

BRET probes can be adapted for LOTIIS. A) Schematic diagram of designed split-BRET receivers. Numbers represent the relative amino acid position in the original construct. B) Each receiver line was co-cultured with sender cells in a 1:1 ratio. Complementation of LgBiT and SmBiT from sender-receiver interaction results in resonance energy transfer and shifting of NBiT emission based on the fused FP. C) Split-BRET receivers were co-cultured with sender^high cells (50,000 cells each) and emission spectra collected on a luminometer. Photon outputs were normalized to the maximum emission wavelength and averaged (n = 3). D) Total flux of split-BRET receivers co-cultured with sender^high cells (50,000 cells each). Error bars represent the SEM for n = 9 replicates for co-cultured cells, and n = 6 replicates for receiver cells only.

Each receiver line was co-cultured with sender^high cells and imaged. Both the overall signal intensity and emission spectra were measured (Figure 5B). Spectral shifts were observed with the split-BRET systems (Figure 5C). Importantly, robust signals were observed when cells were co-cultured together indicating that LOTIIS can be performed with a variety of BRET probes (Figure 5D). BRET receivers should enable tracking of complex cellular interactions, including more than two populations of interest in heterogeneous environments. While the emission spectra are quite broad and overlapping, multiplexed imaging is possible using spectral unmixing or bioluminescent phasor analysis.^28,31

CONCLUSIONS

We developed a novel luciferase-based platform to label and monitor cellular interactions. This system (termed LOTIIS) combines the high signal-to-noise ratio of NBiT with the high-affinity scFv-peptide pair from the SunTag system to label receiver cells of interest. We demonstrated that although LOTIIS can function in two different orientations (scFv-SmBiT/GCN4-LgBiT or scFv-LgBiT/GCN4-SmBiT), only the scFv-SmBiT/GCN4-LgBiT pair is viable. We also found that expression of the GCN4 epitope was greatly impacted by gene product size, with large copy numbers resulting in weaker photon outputs. Optimizing the overall levels or display of GCN4 in the extracellular space could potentially enhance bioluminescent signals. Even without these improvements, though, the current iteration provides >130-fold turn-on upon sender-receiver cell interaction.

LOTIIS differs from other established methods for tracking cell-cell interactions. One notable feature is its reliance on both a secreted element and one that is anchored to the cell surface. Most other platforms anchor both components to the cell populations of interest.^7–9,32,33 While such strategies can reliably report on direct cell-cell contacts, they can result in perturbations to cell motility and function due to reporter complementation. Secreted tags can mitigate many of these issues. With the highest affinity NBiT pair (K_D = 0.7 nM) we were able to achieve robust light emission within an hour of sender-receiver incubation, with signal persisting for at least two hours post-interaction. We attribute light production to extracellular complementation of the luciferase fragments, and media transfer experiments suggest that this mechanism is feasible. However, we cannot rule out signal generation from intracellular scFv binding and/or luciferase reassembly, and additional experiments are necessary to clarify this point. The net result is the same in either case, though, with signal localized to the receiver cell and dependent on the proximity of sender cells. We further showcased the generality of LOTIIS for other split-reporter systems and developed a split-BRET platform for multiplexed imaging.

While presenting some advantages, LOTIIS is not without limitation. Although compatible with other split reporters, LOTIIS requires that the fragments exhibit relatively high complementation affinities. Additionally, the high affinity binding of the scFv (K_D = 44 pM)²¹ complicates attempts to modulate the labeling radius. Within a static media environment, the concentration of secreted scFv quickly reaches a threshold whereby signal is observed even at long distances. Moreover, both fragments of NBiT are secreted, resulting in background complementation within the media and an erosion of the distance-dependent nature of the signal. These effects can potentially be mitigated in studies with continuous flow or in vivo. Indeed, previous studies have shown that sample agitation increases the stringency on interacting cells, requiring direct contact for efficient assembly of secreted reporters.¹⁵ It should also be noted that there is appreciable signal in the absence of fragment complementation. Background photon production can result from LgBiT itself (which retains some residual catalytic activity despite being a split fragment) and FRZ auto-oxidation, which is highly context dependent.³⁴ Alternative split fragments or FRZ analogs could potentially reduce such signals and provide improved signal-to-noise outputs.^29,35

We anticipate that additional engineering of LOTIIS will enable more refined tracking of interactions over a range of distances. The distance dependence of LOTIIS can be controlled by tuning the complementation affinity among the NBiT fragments. Multimerization of the SmBiT peptides can further enhance signal output. We intend to examine these parameters in future studies using the wide range of SmBiT affinities reported. Future work will also apply LOTIIS to multiplexed and orthogonal imaging using the split-BRET reporters featured in this study and related variants.^28–36 Collectively, such probes will be able to capture both transient and long-lasting interactions across varying distances, improving our understanding of complex multicellular networks.

Materials and Methods

General information

Q5 DNA polymerase, restriction enzymes, and all buffers were purchased from New England BioLabs. dNTPs were purchased from Thermo Fisher Scientific. Luria-Bertani medium (LB) was purchased from Genesee Scientific. All plasmids and primer stocks were stored at −20 °C unless otherwise noted. Primers were purchased from Integrated DNA Technologies and plasmids were sequenced by Genewiz. Sequencing traces were analyzed using Benchling. Nanopure water was used for all biological methods unless otherwise noted.

General cloning methods

The SunTag and scFv scaffolds were amplified from pHRdSV40-NLS-dCas9-24xGCN4_v4-NLS-P2A-BFP-dWPRE (Addgene plasmid #60910) and pHRdSV40-scFv-GCN4-sfGFP-VP64-GB1-NLS (Addgene plasmid #60904), respectively. These plasmids were gifts from Ron Vale. Venus and mTurqoise2 fluorescent proteins were amplified from YeNL/pcDNA3 (Addgene plasmid #85201) and CeNL/pcDNA3 (Addgene plasmid #85199), which were gifts from Takeharu Nagai. mScarlet was amplified from pcDNA3-LumiScarlet, a gift from Huiwang Ai (Addgene plasmid #126623). Polymerase chain reaction (PCR) was used to prepare genes of interest, and products were analyzed via gel electrophoresis. Products were excised and purified. Amplified genes were ligated into destination vectors via Gibson assembly ³⁷. All PCR reactions were performed in a BioRad C3000 Thermocycler using the following conditions: 1) 95 °C for 3 min, 2) 95 °C for 30 s, 3) −1.2°C per cycle starting at 72 °C for 30 s, 4) 72 °C for 30 sec, repeat steps 2–4 ten times, 5) 95 °C for 3 min, 6) 95 °C for 30 s, 7) 60 °C for 30 s, 8) 72 °C for 2 min repeat steps 6–8 twenty times, then 72 °C for 5 min, and hold at 12 °C until retrieval from the thermocycler. Gibson assembly conditions were 50 °C for 60 min and held at 12 °C until retrieval from the thermocycler. Ligated plasmids were transformed into TOP10 E. coli cells using the heat shock method. After incubation at 37 °C for 18–24 h, colonies were picked and expanded overnight in 5 mL LB broth supplemented with ampicillin (100 μg/mL) or kanamycin (100 μg/mL). DNA was extracted from colonies using a Zymo Research Plasmid Mini-prep Kit. DNA was subjected to restriction enzyme digest to confirm gene insertion. Positive hits were further sequenced.

General cell culture methods

HEK293T cells (HEK, ATCC) and stable cells lines derived from HEK293T cells were cultured in complete media: DMEM (Corning) containing 10% (v/v) fetal bovine serum (FBS, Life Technologies), penicillin (100 U/mL), and streptomycin (100 μg/mL, Gibco). Cell lines stably expressing sender or receiver plasmids were generated via transfection of HEK293T with Lipofectamine 2000 and the following three plasmids: AAVS1 donor plasmid containing the relevant gene of interest, Cas9 (Addgene #41815), and AAVS1 sgRNA (Addgene #53370). Transfected cells were further cultured with puromycin (20 μg/mL) to preserve gene incorporation. Cells were incubated at 37 °C in a 5% CO₂ humidified chamber. Cells were serially passaged using trypsin (0.25 % in HBSS, Gibco).

General flow cytometry methods

Cells were treated with trypsin for 5 min at 37 °C and then neutralized with complete media. Cells were transferred to Eppendorf tubes and pelleted (500 xg, 5 min) using a tabletop centrifuge (Thermo Fischer Sorvall Legend Micro 17). The resulting supernatants were discarded, and cells were washed with PBS (2×400 μL). Cells were immunostained with anti-FLAG or anti-HA antibodies (see staining protocol), then immediately analyzed on a Novocyte flow cytometer (ACEA BioSciences Inc). Live cells were gated and, in some cases, cells expressing FLAG or HA were further gated. For each sample, 10,000 events were collected on either the “Live cell” gate, “FLAG⁺” or “HA⁺” gate. AlexFlour488 or 647 fluorescence was analyzed and quantified using FlowJo v10.8.0 software (FlowJo, LLC).

General bioluminescence imaging

All analyses were performed in black 96-well plates (Greiner Bio-One) or clear 12-well plates (Genesee Scientific). In some cases, cells were lysed via passage through a 30-guage needle. Furimazine (FRZ, Promega) was added to all samples using a 1:100 dilution of the commercial stock. Plates were imaged in a dark, light-proof chamber using an IVIS Lumina (PerkinElmer) CCD camera chilled to −90 °C. The stage was kept at 37 °C during imaging and the camera was controlled using Living Image software. Exposure times were set to 1 min and binning levels were set to medium. Regions of interest were selected for quantification and total flux values were analyzed using Living Image software. All data were exported to Microsoft Excel or PRISM (GraphPad) for further analysis.

General fluorescence imaging

Fluorescence microscopy images were acquired on a Keyence BZ-X800 microscope using a 10X or 40X objective. Appropriate filters were used to image TagBFP2, mNeptune and sfGFP expression. Exposure times were selected below threshold for camera saturation.

General data analyses

For all bar graphs, standard errors of the mean were calculated using n=3 or n=9 replicates. To determine statistical significance, unpaired t-tests were performed using GraphPad Prism 9, and the following notation was used: ****p<0.0001, ***p<0.001, **p<0.01, *p<0.05, n.s. = non-significant

Cell staining

Cells were stained with either AlexaFluor488 conjugated anti-FLAG antibodies (1:200, BioLegend) or AlexaFluor647 conjugated anti-HA antibodies (1:200, BioLegend). Antibodies were diluted in PBS with 10% FBS and 0.1% sodium azide. Cells were incubated with antibodies for 30 min on ice. Cells were then washed 3 times with PBS with 10% FBS and 0.1% sodium azide. Cells were then analyzed through flow cytometry.

Stencil experiments

Biocompatible stencils¹⁵ were used to separate the cells. Prior to cell addition, the stencils were sterilized with 100% ethanol in 12-well plates (1–7 mm stencils) and allowed to dry uncovered in a biosafety cabinet for several hours. Stable sender and receiver cells were immediately plated (50,000 cells/well) and allowed to settle over 1 h. Media (400 μL) was slowly added to each well to avoid lifting cells. Cells were then incubated up to 4 h before imaging on the IVIS. Prior to imaging, media was removed, and cells washed with PBS (600 μL). FRZ solution (100 μL of a 1:100 dilution in PBS) was added to the cells, and samples were incubated for 1 min prior to imaging.

Bioluminescent spectra

Receiver cells (50,000 cells stably expressing LOTIIS, CeLOTIIS, GeLOTIIS, YeLOTIIS, or SeLOTIIS) were co-cultured with 50,000 sender^high cells for 24 h. Cells were collected and replated into 96-well black plates. FRZ was added and a luminescence scan was performed on a TECAN Spark Microplate Reader (398-653 nm, bandwidth = 25, step size = 15, 1s integration). Measurements were then normalized based on the minimum and maximum values and plotted on PRISM.

Supplementary Material

Amino acid sequences for LOTIIS components, flow cytometry analyses, cell labeling analyses, LOTIIS adaptation for split GFP, analysis of GCN4 copy number, LOTIIS signal longevity (Figures S1–S5).