Bioorthogonal Engineering of Bacterial Effectors for Spatial-Temporal Modulation of Cell Signaling.

The complicated and entangled cell signaling network is dynamically regulated by a wide array of enzymes such as kinases. It remains desirable but challenging to specifically modulate individual, endogenous kinases within a cell, particularly in a spatial-temporally controlled fashion. Current strategies toward regulating the intracellular functions of a kinase of interest either lack specificity or require genetic engineering that may perturb its physiological activity. Herein, we harnessed a bacterial effector OspF for optical and chemical modulation of the endogenous mitogen-activated protein kinase (MAPK) cascade in living cells and mice. The phospho-lyase OspF provided high specificity and spatial resolution toward the desired kinase such as the extracellular signal-regulated kinase (ERK), while the genetically encoded bioorthogonal decaging strategy enabled its temporal activation in living systems. The photocaged OspF (OspF*) was applied to dissect the subcellular signaling roles of ERK in nucleus as opposed to cytoplasm, while the chemically caged OspF (OspFc) was introduced into living mice to modulate ERK-mediated gene expression. Finally, our spatially and chemically controlled OspFc was further used to precisely tune immune responses in T cells. Together, our bioorthogonal engineering strategy on bacterial effectors offers a general tool to modulate cell signaling with high specificity and spatial-temporal resolution.

Introduction

Eukaryotic cells have evolved a diverse repertoire of enzymes for catalyzing potent chemical modifications on proteins that dictate diverse signaling events.[1,2] For example, nearly 600 kinases exist in human cells to control phosphorylation, and dysregulation of this complex and interconnected signaling network is often linked to diseases such as cancer.[3] The endogenous mitogen-activated protein kinase (MAPK) cascade is one of the central signaling pathways that regulates numerous cellular processes.[4,5] As two essential nodes at the end of the MAPK cascade, extracellular signal-regulated kinase (ERK) and p38 receive upstream signals and shuttle between cytoplasm and nucleus to phosphorylate more than 150 substrates participating in regulation of gene expression, cell proliferation, as well as diverse cellular responses.[6,7] In addition, ERK can directly phosphorylate almost all the upstream components of the receptor tyrosine kinase (RTK)-RAS-MAPK cascade to mediate the negative feedback regulation.[8] Misfunction of MAPK cascade is considered as a hallmark in cancer cells including melanoma and colorectal cancers,[9,10] and has also been connected to immune diseases.[11] Nevertheless, methods for precise tuning of the endogenous MAPK cascade are highly challenging, particularly in a spatial–temporally controlled fashion.[12] For example, although small-molecule modulators have been developed for targeting the MAPK pathway, they often have certain off-target effects, particularly among the isoforms of MAPK family enzymes.[3,13,14] Meanwhile, optogenetics and chemical genetic strategies require genetic manipulation and overexpression of the kinase of interest that may perturb its native cellular functions.[15−18]

Diverse effector proteins have been evolved by bacteria to modulate signaling pathways inside host cells with high specificity.[19,20] For example, OspF is a phospho-lyase from Shigella spp. that can be secreted through a Type III secretion system (T3SS) into host cells to specifically dephosphorylate phosphothreonine at residue 202 on ERK and residue 180 on p38, respectively (Scheme and Figure S1).[21,22] The resulting dehydrobutyrine cannot be rephosphorylated by its upstream kinases and thus permanently abrogated ERK and p38 activity (Scheme B).[21] Inspired by the high specificity of OspF toward ERK and p38 as opposed to other kinases including other members of the MAPK family, we decided to employ this phospho-lyase to modulate the endogenous MAPK signaling cascade. However, since the constitutively active form of OspF will irreversibly inhibit ERK and p38 activity and thus permanently turn off the MAPK signaling cascade, temporally controlled activation of OspF is highly desired. Previous methods relying on controlled expression of OspF protein have a poor temporal resolution,[23] and are exceedingly difficult for applications in live animals. We envisioned that our recently developed genetically encoded bioorthogonal decaging strategy would enable precise and temporally controlled activation of OspF under living conditions. Herein, by bioorthogonal engineering and targeting of the bacterial effector OspF to different subcellular compartments (e.g., cytosol versus nucleus), we report spatial–temporally controlled modulation of the endogenous MAPK signaling cascade in living systems by small-molecule triggers or light.

Scheme 1

Bioorthogonal Engineering of a Bacterial Effector OspF for Optical and Chemical Modulation of Endogenous Signal Transductions

(A) Bioorthogonal caging of the phospho-lyase OspF can be used to specifically modulate the endogenous ERK and p38 activity by light or chemical triggers. (B) Structure of caged OspF bearing ONBK (o-nitrobenzyloxycarbonyl-Nε-lysine, left) or TCOK-a (axial isomer of trans-cyclooctene lysine, right) at residue 134. The catalytic center of OspF is highlighted. The presence of 365 nm light or the chemical trigger (Me2-Tz) would decage ONBK or TCOK-a, and rescue the phospho-lyase activity of OspF. The in situ rescued OspF can specifically remove the phosphate group on phosphothreonine at residue 202 on ERK and residue 180 on p38, respectively. The resulting dehydrobutyrine at these sites can no longer be rephosphorylated, leading to permanently abrogated ERK and p38 activity.

Results and Discussion

Engineering an Optically Controlled OspF (OspF*) for Living Cells

We started by engineering an optically controlled OspF in living cells based on the genetic code expansion system.[24] The photocaged lysine analogues such as ONBK[25,26] and other o-nitrobenzyl caged[27−29] or coumarin lysines[30] have been previously developed for controlling lysine-dependent protein activity inside living cells. As lysine 134 (K134) is the catalytic residue for OspF’s phospho-lyase activity,[22] we used ONBK to replace OspF’s K134 residue (Scheme B). The resulting protein OspF-K134ONBK (termed OspF*) completely lost its phosphothreonine lyase activity similar to the catalytically inactive OspF-K134A mutant until the 365 nm light-mediated photo-decaging (Scheme B). The expression of OspF* was first verified by multiple methods including immunoblotting, flow cytometry analysis, as well as mass spectrometry analysis (Figure A and Figures S2 and S3A,D), followed by examining its photoactivation property in live HEK293T cells. Before photolysis, we used PMA (phorbol 12-myristate 13-acetate), a protein kinase C (PKC) activator, to stimulate the endogenous ERK/p38 phosphorylation and activate the MAPK signaling cascade. Interestingly, the phosphorylation levels of endogenous ERK/p38 were completely eliminated in cells harboring OspF* after 365 nm light irradiation for 5 min (Figure A,B and Figure S4), whereas no variation of ERK/p38 phosphorylation was observed in cells without photoactivation (Figure A,B and Figure S4). Furthermore, no phosphorylation change was detected on ERK/p38 when the same batch of cells without OspF* were treated with the same dosage of light, which confirmed that the low dosage of light we used did not cause the phosphorylation change on ERK, p38, as well as their substrates inside cells (Figure S5). Finally, we found that expression of OspF* did not affect the phosphorylation status of JNK (another MAPK that is not OspF’s substrate) as well as other kinases we tested including IκBα and Akt before and after photoactivation (Figure A and Figure S6), which further confirmed the high specificity of OspF toward ERK/p38. We also constructed a stable HEK293T cell line that constitutively expresses the MmPylRSNBK-1/tRNAPyl CUA pair, the key component for UAA incorporation, for production and optical rescue of OspF*. Immunoblotting results confirmed the expression and photoactivation of OspF*, and flow cytometric analysis further showed the expression efficiency in cells (Figures S3G,J and S7).

Figure 1

Optical modulation of the endogenous MAPK signaling cascade by OspF* in living cells. (A) Immunoblotting analysis of expression and activation of OspF* in living cells. OspF activity was detected by phosphorylation of ERK and p38 using specific antibodies. The phosphorylation of JNK and IκBα was also analyzed to confirm the specificity of OspF. The expression of OspF* was verified by using anti-Myc tag antibody. OspF-WT and actin were used as controls. (B) Variation of photoactivation time from 0 to 5 min was performed on cells harboring OspF*, and a decreasing of ERK/p38 phosphorylation was detected with increasing of photoactivation time. Cells with no ONBK supplementation were used as a control. (C) Dual-luciferase analysis of activation of OspF* in living cells. The relative luminescence activity is proportional to the endogenous phosphorylation level of ERK. Cells transfected with empty vector or OspF-WT were used as controls. Data are presented as mean ± SD (n = 3). (D) RT-qPCR analysis of IL-8 mRNA transcription in cells harboring OspF*. Data are presented as mean ± SD (n = 3).

We then applied OspF* for temporal modulation of the MAPK pathway. First, we showed that, upon photoactivation of OspF* inside cells, the attenuation of ERK/p38 phosphorylation was observed as fast as 5 min (Figure S8) while the protein level of OspF* was not changed even 60 min after photoactivation. We then employed a luciferase reporter, SRE-luc, to monitor OspF activity in living cells.[31] SRE-luc reporter responds to ERK/MAPK signaling, allowing the luciferase expression level correlated with the endogenous ERK activity. For example, the reactivated OspF* attenuated ERK activity constitutively and thus reduced luciferase expression and the bioluminescence signal. Indeed, an obvious decrease of bioluminescence signal was detected in PMA stimulated cells harboring OspF* after 5 min of light treatment, which was similar to the effect of OspF-WT (Figure C). In contrast, no bioluminescence variation was observed in the same batch of cells without light treatment (Figure C). We further examined the effect of OspF* on gene transcription. Previous studies demonstrated that inactivation of ERK/p38 would block the immune response mediated by the MAPK pathway such as the transcription and secretion of cytokine including interleukin-2 (IL-2) and interleukin-8 (IL-8).[23,32] Indeed, upon photoactivation of OspF* in cells prestimulated by PMA, a dramatic reduction on transcription and secretion of IL-8 was observed (Figure D and Figure S9). Taken together, our engineered photoactivatable OspF can modulate the endogenous MAPK signaling in living cells with high temporal resolution.

Spatial–Temporal Modulation of OspF* by Light

ERK is known to shuttle between the nucleus and cytoplasm and phosphorylate a series of substrates with different physiological outputs.[6] Activated ERK can translocate into the nucleus to phosphorylate transcription factors such as Myc proto-oncogene protein (c-MYC) to induce gene expression and cell proliferation. Meanwhile, phosphorylated ERK remains in the cytoplasm to phosphorylate over 50 substrates and mediate biological processes such as negative feedback regulation. Modulating ERK’s activity with spatial resolution may help to dissect its functions in different subcellular regions. Small-molecule ERK inhibitors will block all of ’s activity that lacksspatial resolution. In contrast, subcellular-targeted OspF* may allow modulation of ERK’s activity in the nucleus as opposed to cytoplasm with high spatial resolution. We first generated nucleus-located OspF* (nu-OspF*) by fusing a bipartite nuclear localization sequence (NLS) from nucleoplasmin to the N-terminal of OspF*, while two K to A mutations were introduced to this NLS to render the resulting OspF* exclusively cytosol-located (cyt-OspF*) (Figure A).[33,34] Subcellular fractionation and immunoblotting analysis confirmed the proper subcellular localization as well as the similar expression levels of nu-OspF* and cyt-OspF* in the nucleus and cytoplasm of HEK293T cells (Figure S10). Flow cytometric analysis verified the expression of both nu-OspF* and cyt-OspF* in most cells (Figure B and Figure S3B,C,E,F), while immunofluorescence analysis also verified the adequate expression as well as the proper subcellular localization of nu-OspF* and cyt-OspF* in the majority of cells (Figure C and Figure S11). Next, a significantly decreased ERK phosphorylation was observed in cells after photoactivation of either nu-OspF* or cyt-OspF* (Figure S12). Furthermore, we analyzed the extent of ERK phosphorylation in the nucleus and cytoplasm after activation of subcellular-targeted OspF*. Immunoblotting results showed that phosphorylation of ERK in the nucleus was significantly decreased after nu-OspF* activation, but negligible variation was detected on cytosolic ERK. In contrast, a significant decrease of ERK phosphorylation in the cytoplasm was detected after optical rescue of cyt-OspF* inside cells (Figure S10). These results proved that our subcellular-targeted OspF* could modulate the extent of ERK phosphorylation in different cellular compartments.

Figure 2

Optical control of subcellular-targeted OspF* enables spatial–temporal modulation of the MAPK signaling cascade. (A) Diagram of the nucleus- and cytosol-located OspF* (nu-OspF* and cyt-OspF*). (B) Immunofluorescent staining and flow cytometry quantification of the expression of subcellular-located OspF*. Cells expressing nu-OspF* or cyt-OspF* were stained using an anti-Myc antibody followed by an Alexa fluor 488-conjugated antimouse IgG before being evaluated by flow cytometry. Histogram showing percentage of transfected cells that harbor full-length nu-OspF* or cyt-OspF*. Data were analyzed by FlowJo software. (C) Immunofluorescence imaging of the subcellular-located OspF*. Cells expressing nu-OspF* or cyt-OspF* were fixed and evaluated by immunofluorescence imaging using an anti-Myc antibody followed by an Alexa fluor 488-conjugated antimouse IgG. The nucleus was stained with Hoechst 33342. Overlay of fluorescence image and DIC image is shown, and all fluorescence images are shown in Figure S11. Scale bars represent 10 μm. (D, E) Schematic illustration and immunoblotting results of the effects on ERK substrates after rescue of nu-OspF*. Phosphorylation of c-MYC was attenuated within 5 min in nu-OspF* expressed cells after photoactivation, followed with dramatic protein degradation. No obvious variation on p90RSK phosphorylation was detected. (F, G) Schematic illustration and immunoblotting results of the effects on ERK substrates after rescue of cyt-OspF*. Phosphorylation of p90RSK was attenuated within 5 min in cyt-OspF* expressed cells after photoactivation. Less decrement of c-MYC phosphorylation and protein abundance was detected at the same time point (5 min, 60 min) in comparison with nu-OspF* harboring cells. Data shown in parts E and G are representative of at least three independent experiments.

To demonstrate the spatial modulation capability of our subcellular-located OspF*, we monitored the phosphorylation states of ERK’s nuclear substrate c-MYC as well as the cytoplasmic substrate p90 ribosomal S6 kinase (p90RSK), respectively. As an oncogenic transcriptional factor, the amplification of c-MYC has been shown as a hallmark in cancer and malignant tumor generation.[35] Phosphorylation of c-MYC on residue 62 mediated by ERK promotes its stability and transcriptional activity.[36] The cytoplasmic substrate p90RSK plays an important role in transcription and cell growth regulation.[37] Once p90RSK is phosphorylated on residues T359/S363 by ERK in cytoplasm, it would phosphorylate many targets in both the cytoplasm and nucleus. Photoactivation on cells harboring nu-OspF* led to a significant decrease of c-MYC phosphorylation within 5 min and degradation of c-MYC in 60 min, while no obvious variation on p90RSK was observed (Figure D,E and Figure S13). In contrast, upon photoactivation on cells expressing cyt-OspF*, the phosphorylation of p90RSK was inhibited within 5 min (Figure F,G). In addition, less decrement of c-MYC phosphorylation (5 min) and protein abundance (60 min) was detected at the same time point in comparison with cells harboring nu-OspF* (Figure D,E and Figures S13 and S14A). Moreover, we used the MEK1/2 inhibitor U0126 to inhibit ERK phosphorylation in cytosol and investigated the variation of c-MYC phosphorylation and abundance. No alternation of c-MYC phosphorylation and protein abundance was observed within 30 min, and a decreased phosphorylation was detected at 45 min (Figure S14B). This is similar to the results acquired from cells harboring cyt-OspF* that inactivated ERK in cytosol. Therefore, specific targeting of nucleus-located ERK can downregulate c-MYC activity and stability without interfering with its cytosolic substrate p90RSK.

Similarly, activation of cytosol-located OspF* did not affect nuclear ERK and c-MYC activity in a short time (within ∼30 min) (Figure G and Figure S14A). After a longer incubation time (60 min or above), no more cytosol-activated ERK can be translocated into the nucleus because of the inhibition of ERK phosphorylation in cytosol, which caused the decrease of c-MYC’s phosphorylation level as well as protein abundance (Figure G and Figures S13 and S14A). In addition, we also verified our subcellular specific photoactivation strategy in HEK293T cells stably expressing the MmPylRSNBK-1/tRNAPyl CUA pair. Either nu-OspF* or cyt-OspF* was expressed in this stable HEK293T cell line, and the expression of full-length OspF* was analyzed by flow cytometry and immunoblotting (Figures S3H,I,K,L and S15). Similar to the cells transiently transfected with nu-OspF* and cyt-OspF*, ERK phosphorylation was attenuated 5 min after light irradiation in both cells. Alternation on p90RSK phosphorylation was only observed in cells harboring cyt-OspF*, and a decrease of c-MYC phosphorylation and protein abundance was detected only in the nu-OspF* expressed cells within 5 min after optical rescue (Figure S15). Taken together, our engineered OspF can modulate endogenous ERK functions with high spatial–temporal resolution.

Engineering a Chemically Controlled OspF (OspFc) for Living Animals

We have shown that our engineered OspF* could be applied for modulation of endogenous MAPK signaling cascade in living cells, which provided high spatial–temporal resolution desired for mechanistic study in vivo. Meanwhile, small-molecule-mediated chemical-rescue strategies possess higher penetration capability in deep tissue and are thus more compatible for kinase signaling modulation at the tissue and animal level. Since we have recently developed the small-molecule-mediated chemical-decaging strategy to rescue lysine-dependent protein activity in living animals,[38−41] we created a chemically rescued OspF (termed OspFc) for applications in living animals. The inverse electron demand Diels–Alder reaction (iDA) between TCOK-a and Me2-Tz (3,6-dimethyl-1,2,4,5-tetrazine) has been shown as a rapid and efficient bioorthogonal cleavage reaction for intracellular protein activation in both prokaryotic and eukaryotic systems;[38,39,42] we site-specifically incorporated the chemically caged lysine-TCOK-a into OspF at residue K134, and the resulting OspFc (OspF-K134TCOK) exhibited no phospho-lyase activity (Scheme B and Figure S1). The addition of the decaging reagent Me2-Tz would regenerate free lysine at K134 with its phospho-lyase activity rescued (Scheme B and Figure S1). Indeed, our immunoblotting results showed that the lyase activity of OspFc against ERK and p38 can be effectively rescued after treatment with 100 μM Me2-Tz (Figure A). We also performed a time course study to show the stability of OspFc and its effect on endogenous MAPK signaling after chemical decaging. Immunoblotting results showed that an obviously attenuated phosphorylation level of ERK and p38 was observed within 10 min after Me2-Tz treatment and can sustain for over 60 min (Figure S16). Meanwhile, no apparent change of OspFc abundance was detected even 60 min after chemical decaging (Figure S16). We then employed the aforementioned luciferase reporter, SRE-luc, to monitor OspFc activity in living cells. Indeed, an obvious decrease of bioluminescence signal was detected in PMA stimulated cells harboring OspFc after 100 μM Me2-Tz treatment, which was similar to the effect of OspF-WT (Figure B). In contrast, no bioluminescence variation was observed in the same batch of cells without Me2-Tz treatment (Figure B). These results showed that our chemical-rescue strategy can rescue OspFc for MAPK modulation in living cells.

Figure 3

Chemical rescue of OspFc in living cells and mice. (A) Immunoblotting analysis of expression and activation of OspFc in HEK293T cells. After expression, cells were incubated with 100 μM Me2-Tz for another 1 h. OspF activity was detected by phosphorylation of ERK and p38 using specific antibodies. The phosphorylation of JNK was also analyzed to confirm the specificity of OspF. Cells expressing OspF-WT or transfected with empty vector were used as controls, and actin was used as loading controls. (B) Dual-luciferase analysis of activation of OspFc in living cells. The relative luminescence activity is proportional to the endogenous phosphorylation level of ERK. Cells transfected with empty vector or OspF-WT were used as controls. Data are presented as mean ± SD (n = 3). (C) Schematic flow showing the chemical rescue of OspFc in living mice. HEK293T cells harboring OspFc were subcutaneously injected into mice followed by tail vein injection of Me2-Tz (50 μL, 300 mM). Mice were fed for another 24 h for luciferase expression, and luciferin was injected 10 min before bioluminescence imaging. OspFc by Me2-Tz treatment would attenuate ERK activity and thus reduce luciferase expression and bioluminescence signal. (D) Representative images of rescued OspFc activity as measured by bioluminescence after Me2-Tz treatment. An obvious decrease of bioluminescence was observed only in the leg injected with cells harboring OspFc after Me2-Tz treatment.

We next pursued the chemical rescue of OspFc in living mice. A previously developed xenograft model was used to demonstrate the chemical activation of OspFcin vivo.[39] Stimulated HEK293T cells expressing OspFc and SRE-luc reporter were injected into living mice subcutaneously, followed by tail vein injection of Me2-Tz (50 μL, 300 mM, equal to 66 mg/kg body weight). As a control, a noncleavable lysine analogue, CbzK (benzyloxycarbonyl caged lysine), was incorporated into OspF at the same residue K134 to generate an inactivated counterpart (OspF-K134CbzK). Cells harboring OspF-K134CbzK and SRE-luc were subcutaneously injected into the other leg on the same mouse (Figure C). Bioluminescence imaging showed that the lyase activity of OspFc can be effectively rescued with Me2-Tz treatment, as made evident by the significantly attenuated bioluminescence signal in mice upon Me2-Tz addition (Figure D and Figure S17). In contrast. a similar bioluminescence level remained in the leg injected with OspF-K134CbzK-expressing cells with and without Me2-Tz treatment. Therefore, our chemical-decaging strategy on OspF can be applied to living animals.

Precise Tuning of T Cell Responses by Chemically Controlled, Subcellular-Targeted OspFc

We further expanded our chemically rescued OspFc to tune immune responses. T cell receptor (TCR) stimulation can lead to T cell activation with a series of immune responses such as cytokine production and release (Figure A).[43−45] As a key component in TCR signaling, activated ERK will translocate into the nucleus and regulate the expression of cytokines such as IL-2 and IL-8.[46] Meanwhile, ERK also phosphorylates and inhibits the activity of a panel of upstream components in cytosol including MEK1, BRAF, SOS, and Lck to avoid overstimulation and hyperactivation of TCR signaling (Figure B).[47,48] For immunotherapeutic strategies such as adoptive T cell therapy, overstimulation of the MAPK cascade on native or engineered T cells may lead to cytokine release syndrome (also called cytokine storm), a life-threatening immune response.[11] To this end, specific silencing of ERK-induced cytokine expression in the nucleus without affecting the cytosolic signaling loops may become a viable strategy for suppressing cytokine release without affecting T cell activity. We envision that our subcellular-targeted OspFc can be used to specifically inhibit ERK’s activity in the nucleus with the cytosolic feedback regulation unaffected (Figure B).

Figure 4

Precise tuning of T cell activation and immune responses by subcellular-targeted, chemically controlled OspFc. (A) Stimulation on T cells will induce immune responses and cytokine release, while overstimulation on T cells can induce cytokine release syndrome. (B) Schematic illustration of the effects on cytokine expression and ERK-mediated feedback regulations after nu-OspFc activation in T cells upon TCR activation. (C) Immunoblotting analysis of expression and activation of nu-OspFc in Jurkat T cells. After expression, cells were incubated with 100 μM Me2-Tz for another 1 h. OspF activity was detected by phosphorylation of ERK using specific antibodies, and actin was used as loading controls. No obvious variation on BRAF phosphorylation and Lck was observed after nu-OspFc activation. (D) Schematic illustration and ELISA results on cytokine secretion in Jurkat cells after chemical rescue of nu-OspFc. Cells transfected with nu-OspFc were stimulated by PMA and ionomycin (time = 0 h), followed by the addition of Me2-Tz at time = 10 min (red) or time = 3 h (purple) for nu-OspFc activation. Secretion of IL-2 and IL-8 was detected at indicated time points by ELISA assay. Data are presented as mean ± SD (n = 3).

We first surveyed the cytotoxicity of our chemical-decaging method on T cells. Jurkat cells treated with different concentrations of TCOK-a (0.5–2 mM) or Me2-Tz (100–500 μM) were subject to the MTS assay, and no apparent cell death was observed (Figure S18). These results proved that our chemical-rescue strategy is compatible with immune cells. Next, we modulated the endogenous ERK activity by our nucleus-located OspFc (nu-OspFc) in T cells. To acquire a better expression efficiency, we introduced the plasmid encoding nu-OspFc into Jurkat cells by electroporation, which were then subject to PMA and ionomycin stimulation before chemical rescue of OspFc. The immunoblotting results showed that the lyase activity of nu-OspFc against ERK can be effectively rescued after the treatment with 100 μM Me2-Tz (Figure C). We also investigated the impact of nu-OspFc-mediated subcellular ERK inhibition on cytosolic signaling loops, which showed that rescue of nu-OspFc in stimulated Jurkat cells did not change the feedback regulation on BRAF phosphorylation (Figure C). We further monitored the phosphorylation state of lymphocyte-specific protein tyrosine kinase (Lck). Activation of T cells can induce an ERK-mediated feedback regulation and phosphorylation on Lck, which would interfere with the recruitment of Src homology region 2 domain-containing phosphatase-1 (SHP-1) and sustain TCR signaling.[48] It has been previously reported that this ERK-dependent phosphorylation on Lck resulted in its conversion from p56 Lck (p56) to p59 Lck (p59).[49] Indeed, whereas the treatment of a MEK1/2 inhibitor-U0126 attenuated the formation of p59 Lck (Figure S19),[50] no band-shift was observed on Lck in the cells bearing nu-OspFc before and after chemical decaging (Figure C). Taken together, we showed that our nucleus-located, chemically controlled OspFc allowed precise modulation of nuclear ERK’s activity without affecting its cytosolic feedback regulations.

Finally, we monitored the controlled cytokine secretion by our engineered, subcellular-located OspFc in T cells. We first confirmed that our chemical-decaging reagents TCOK-a and Me2-Tz did not interfere with cytokine secretion in Jurkat cells (Figure S20). The nu-OspFc-expressing Jurkat cells were then stimulated by PMA and ionomycin for 10 min, followed by the addition of 100 μM Me2-Tz to rescue nu-OspFc activity. The secretions of IL-2 and IL-8 were found completely inhibited by chemically rescued nu-OspFc even 6 h after PMA and ionomycin stimulation (Figure D, red line), whereas a continuous IL-2 and IL-8 secretion was detected in cells without nu-OspFc activation (Figure D, blue line). Furthermore, the same batch of Jurkat cells was stimulated by PMA and ionomycin for 3 h before the addition of 100 μM Me2-Tz, which also led to a decrease of IL-2 and IL-8 secretion levels (Figure D, purple line). Taken together, we showed that our OspFc-enabled spatial–temporal modulation of ERK-signaling would allow the in situ suppression of cytokine secretion without affecting other functions of T cells.

Conclusion and Outlook

Bacterial effector proteins have been previously applied for modulation of cellular signaling pathways but often lack desired spatial–temporal resolution in living systems.[23] In this work, we engineered optically and chemically controlled phospho-lyase OspF (OspF* and OspFc) by masking its key lysine residue K134 with photocaged and chemically caged lysine analogue. Subcellular targeting of these bioorthogonal-engineered OspF variants allowed us to specifically modulate the endogenous ERK and p38 activity and thus the native MAPK signaling cascade with high spatial–temporal resolution. While the nucleus- and cytosol-located OspF* allowed modulation and dissection of endogenous ERK activity in space and time, the chemically caged OspFc enabled modulation of endogenous ERK activity in living mice, making our strategy compatible with various living systems. Given the versatility of bacterial effectors that have been evolved to manipulate diverse host signaling pathways, our bioorthogonal engineering strategy on OspF may be extended to its homologues (e.g., SpvC) and other effectors to modulate or rewire cell signaling with high specificity and spatial–temporal resolution.

Dysregulation of the MAPK signaling cascade occurs frequently in melanoma, colorectal, and breast cancers. Off-target activation or overstimulation on MAPK cascade also induces the overproduction of proinflammatory cytokines which contributes to autoimmune diseases. Inhibitors against components in this signaling pathway, particularly RAF and MEK, have been developed or are undergoing clinical trials.[12] As key nodes of the MAPK signaling cascade, ERK and p38 receive all upstream signals and regulate a number of substrates involved in different cellular processes including gene expression, cell proliferation, and immune response. Selective and spatially specific inhibition of ERK and p38 provides an attractive choice for targeting the MAPK cascade that may have lower adverse effects or can overcome adaptive drug resistance.[51,52] Given the central role of ERK and p38 at the terminal of the MAPK cascade, our engineered OspF variants offer a powerful toolset for perturbing endogenous MAPK activity with high selectivity, subcellular specificity, as well as minute-scale temporal resolution. Finally, by introducing our nucleus-located OspFc variant into T cells, we selectively disabled ERK’s nuclear function on stimulating interleukin expression without interrupting its feedback regulation on the cytosolic signaling network. This chemical-decaging strategy may provide promising feasibilities to precisely tune the timing and strength of interleukin secretion in activated T cells under clinically relevant settings.