Results

Overview

After carefully designing and planning the experiment, we initiated the experimental part of our project. We primarily focused on validating the activation of the downstream pathway through two inputs: the occurrence of chain-exchange reactions in the CC domain and the efficiency of the positive feedback loop, as well as the performance of the three outputs. We then collected our data and analyzed it according to the Engineering cycle (See page Engineering) to improve our experiment.

Detailed information regarding how we implement it and the record kept in the wet lab can be found in the Experiments part and the Notebook part.

Split Protein as Output

To evaluate whether CC-mediated molecular interactions could produce measurable functional signals, we designed three CC-coupled split protein systems: split luciferase (Luc), split green fluorescent protein (GFP), and split trehalase (TreA). We first validated the feasibility of using these split proteins as output modules, since their reconstitution leads to luminescence, fluorescence, or enzymatic activity recovery, respectively. In each design, the complementary CC domains AP4(C') and P3(C) were fused to the corresponding fragments of the split reporter protein, enabling structural reassembly and activity restoration upon CC heterodimerization (Fig.1).

Fig.1 CC-coupled split protein output systems. (a) Split luciferase complementation. (b) Tripartite split GFP reconstitution. (c) Split trehalase complementation.

We next examined whether these CC-driven split protein systems could successfully reassemble and generate measurable outputs in E. coli or in vitro. The performance of each system was characterized by monitoring its specific signal - luminescence, fluorescence, or enzymatic activity - following CC-induced reconstitution.

Split Luciferase

Both fusion proteins, nLuc_AP4(C') (BBa_25DSNIG7) and P3(C)_cLuc (BBa_2527MQP2), were expressed in E. coli BL21(DE3) and purified using Ni-NTA affinity chromatography. SDS-PAGE confirmed successful purification (Figs.2, 3).

Fig.2 SDS-PAGE analysis of nLuc_AP4(C') purification using Ni-NTA affinity chromatography. Lane M: protein molecular weight marker; Lane U: whole cell lysate before IPTG induction; Lane I: whole cell lysate after IPTG induction; Lane P: insoluble fraction after sonication (pellet); Lane S: soluble fraction after sonication (supernatant); Lane FT: flow-through from Ni-NTA column; Lane E1-E4: elution fractions with 20, 50, 100, and 200 mM imidazole, respectively.

Fig.3 SDS-PAGE analysis of P3(C)_cLuc purification using Ni-NTA affinity chromatography. Lane designations are the same as in Fig.2.

When equal molar amounts of nLuc_AP4(C') and P3(C)_cLuc were co-incubated and measured using Bac-Lumi™ luciferase assay, a clear luminescent signal at around 590 nm was observed shortly after reagent addition (Fig.4). However, the luminescence intensity decayed rapidly, returning to baseline within 5 minutes, in contrast to the 30-minute stability expected from the manufacturer’s protocol. This may result from the lower structural stability of the reconstituted split luciferase compared with the intact enzyme, as the complex likely dissociates after catalysis. Nonetheless, the immediate luminescence upon reagent addition clearly demonstrated that our CC-mediated split luciferase system successfully reconstituted and restored enzymatic activity.

Fig.4 Luminescence of CC-coupled split luciferase in different concentrations. Data were background-subtracted and the blank control therefore corresponds to a constant baseline of zero.

Split GFP

Similarly, two fusion proteins, GFP10_AP4(C') (BBa_25EVV9VN) and P3(C)_GFP11 (BBa_25EVV9VN), were co-expressed with GFP1-9 in E. coli BL21(DE3).

Fluorescence at 528 nm (excitation 485 nm) was continuously monitored for 16 hours (Fig.5). A marked increase in fluorescence was detected within 2 hours after induction, indicating successful reconstitution of the split GFP complex mediated by the paired Coiled-Coil domains.

Fig.5 Normalized fluorescence of CC-coupled tripartite split GFP. Data were background-subtracted using the fluorescence and OD600 of blank LB medium.

Due to time constraints, we did not purify the proteins and repeat the experiment in a cell-free system. However, considering that our construct responds directly to molecular interactions rather than protein accumulation, it is expected to exhibit a detectable fluorescence signal within a shorter timeframe under cell-free conditions.

Split TreA

Both fusion proteins, MBP_nTreA_AP4(C') (BBa_25MD5KSC, abbreviated as nTreA) and P3(C)_cTreA (BBa_25V4VNKA, abbreviated as cTreA), were expressed in E. coli BL21(DE3). Due to time constraints, we have not yet performed protein purification. Instead, we used the supernatant from ultrasonic disruption of E. coli as the crude enzyme solution.

We employed the Trehalase (THL) Activity Assay Kit (BoxBio) to detect trehalase activity, characterised by OD505 (Fig.6, 7). Compared with adding induced cTreA or nTreA supernatants only, those samples both adding induced cTreA and nTreA supernatants exhibited significantly higher trehalase activity. This demonstrates that our CC-mediated split trehalase was successfully reconstituted and restored to enzymatic activity.

Fig.6 After adding the detection reagent, measure the absorbance at 505 nm every minute for 15 minutes. For the nTreA + cTreA group, 30 μL of the supernatant containing nTreA and 30 μL of the supernatant containing cTreA were added; For the cTreA + cTreA and nTreA + nTreA groups, 60 μL of the corresponding supernatant was added, respectively; The negative control group received the negative control solution provided in the assay kit. All OD505 readings were background-subtracted against the baseline absorbance of the buffer solution at 505 nm.

Fig.7 Statistical analysis of the data of Fig.6 at 15 min for each group. * p<=0.05; ** p<=0.01; *** p<=0.001.

Cleavage Efficiency of AvrRpt2

According to previous reports (Chisholm et al., 2005), the AvrRpt2 protease specifically recognizes and cleaves its native plant substrate RIN4 at a conserved heptapeptide motif (VPxFGxW). Chisholm et al. (2005) qualitatively tested the cleavage efficacy of AvrRpt2 on a series of mutants based on the canonical RIN4 cleavage site (Fig.8), which providing strong evidence that AvrRpt2 can indeed cleave the RIN4-derived sequence within our system.

Fig.8 AvrRpt2-dependent cleavage of mutants based on the canonical RIN4 cleavage site. Putative targets do not accumulate in the presence of AvrRpt2. The indicated epitope-tagged proteins were expressed in N. benthamiana leaves alone (-), with AvrRpt2:FLAG (lane A), or with catalytically deficient mutants AvrRpt2:FLAG(C122A) (lane C). X, noninfiltrated leaf. Putative cleavage site amino acid sequences are given in parentheses. Protein extracts were resolved by denaturing PAGE and used for immunoblot analyses with αT7. Approximate positions of molecular mass standards are shown. (Chisholm et al., 2005)

To investigate the cleavage efficiency of AvrRpt2, we selected a series of substrate variants (VPKFGNW, VPKFGDW, VPAFGSW, VPAFGGW) based on the canonical RIN4 cleavage site, aiming to obtain variants with improved cleavage efficiency. In order to quantitatively characterize the proteolytic activity of AvrRpt2, we employed the CC-coupled Split Luciferase system previously validated in the Split Protein as Output section. In this system, different AvrRpt2 cleavage sequences were inserted between the CC domain and the nLuc region of the split luciferase protein nLuc_AP4(C'). Only the uncleaved nLuc_AP4(C') can reconstitute an active luciferase upon interaction with P3(C)_cLuc, leading to measurable luminescence (Fig.9).

Fig.9 Verification of AvrRpt2 cutting efficiency. (a) Prior to AvrRpt2 cleavage, split luciferase exhibits normal luminescence. (b) After AvrRpt2 cleavage, split luciferase fails to exhibit luminescence.

By co-incubating AvrRpt2 with nLuc_AP4(C') constructs containing different cleavage sites and comparing the resulting luminescence intensity with that of the control group (without AvrRpt2), we could quantify the relative cleavage efficiency of AvrRpt2 toward each substrate variant.

AvrRpt2 was expressed in E. coli BL21(DE3) and purified using Ni-NTA affinity chromatography. SDS-PAGE confirmed successful purification (Fig.8). P3(C)_cLuc was successfully purified in the Split Protein as Output section (Fig.10).

Fig.10 SDS-PAGE analysis of AvrRpt2 purification using Ni-NTA affinity chromatography. Lane designations are the same as in Fig.2.

As for the nLuc_AP4(C')_pET-15b construct containing the AvrRpt2 recognition site, colony PCR results indicated that transformation into E. coli BL21(DE3) was unsuccessful. We plan to repeat the transformation and subsequently express this protein in E. coli BL21(DE3). Since this construct differs from the original nLuc_AP4(C') only by the insertion of an AvrRpt2 cleavage site—and the latter was successfully expressed in a soluble form—we believe that the modified protein can also be purified using a similar approach.

In any case, although we have not yet been able to verify the cleavage efficiency of AvrRpt2 at our selected recognition sites, the results reported by Chisholm et al. (2005) confirm that these sites are recognized and cleaved by AvrRpt2.

Interaction of COI1 and JAZ1

The COI1 and JAZ1 proteins constitute the second input pathway in our system. According to previous studies (Sheard et al., 2010), JAZ1 interacts with COI1 through a short peptide segment (Glu200-Val220) and mediates the sensing of Jasmonic Acid-Isoleucine (JA-Ile). In our design, this peptide segment (ELPIARRASLHRFLEKRKDRV) was used to enable the specific interaction between JAZ1 and COI1. Coronatine (COR) is a structural analogue of JA-Ile and can similarly activate the COI1-JAZ1 complex.

Thines et al. (2007) conducted in vitro pull-down assays (Fig.11a) using the functional c-Myc-tagged tomato COI1 (COI1-Myc) and full-length JAZ1, demonstrating that JA-Ile induces COI1-JAZ1 interaction in vitro. Also, Sheard et al. (2010) validated via in vitro radioligand binding assays that our selected JAZ1 short peptide (ELPIARRASLHRFLEKRKDRV) could bind to COI1 in the presence of COR(Fig.b). These results collectively indicate that COR or JA-Ile can successfully mediate the binding between COI1 and the JAZ1 short peptide within our system design.

Fig.11 Previous studies on the interaction between COI1 and JAZ1. (a) Y2H protein-protein interaction assay. Sl, Solanum lycopersicum; At, Arabidopsis thaliana. Blue colonies indicate interaction, white colonies indicate no interaction. Colonies of a positive-control strain, pLexA-53/pB42AD-T, are shown (top row). (b) Saturation binding of COI1-ASK1 and the JAZ1 +5 degron peptide, with a Kd of 108±29 nM. All results are the mean ± S.E. of two to three experiments performed in duplicate. (Thines et al., 2007; Sheard et al., 2010)

To verify whether COR or JA-Ile can act as input signals to activate our split protein system fused with COI1 and JAZ1, and to determine their activation threshold, we employed the CC-coupled Split GFP system, previously validated in the Split Protein as Output section (Fig.12). We constructed fusion proteins COI1-GFP11 and GFP10-JAZ1, and co-incubated them with GFP1-9. By adding a series of JA-Ile concentrations and measuring fluorescence intensity, we determined the minimal threshold concentration required for JA-Ile to activate the COI1-JAZ1 pair as a signal switch. (Due to funding reasons, we only use JA-Ile in our experiments to validate the functionality of this signalling module.)

Fig.12 Verification of COI1 and JAZ1 interaction.

Initially, we attempted to purify GFP10-JAZ1 directly using Ni-NTA affinity chromatography. However, due to its small molecular weight (8.8 kDa), the protein was difficult to visualize on SDS-PAGE and could not be effectively concentrated by ultrafiltration. Therefore, we constructed a fusion protein MBP-GFP10-JAZ1 (BBa_254RZXUU) containing a TEV protease cleavage site between the MBP tag and GFP10. The MBP tag increased the molecular weight of the fusion protein to 49.5 kDa, facilitating visualization and enrichment. MBP-GFP10-JAZ1 was expressed in E. coli BL21(DE3) and purified via Ni-NTA affinity chromatography. SDS-PAGE analysis confirmed successful purification (Fig.13).

Fig.13 SDS-PAGE analysis of MBP-GFP10-JAZ1 purification using Ni-NTA affinity chromatography. Lane M: protein molecular weight marker; Lane FT: flow-through from Ni-NTA column; Lane E1-E4: elution fractions with 20, 50, 100, and 200 mM imidazole, respectively; Lane S: soluble fraction after sonication (supernatant); Lane P: insoluble fraction after sonication (pellet); Lane I: whole cell lysate after IPTG induction; Lane U: whole cell lysate before IPTG induction.

We also attempted to express COI1-GFP11 directly in E. coli BL21(DE3), but even under optimized induction conditions (12 °C, 0.2 mM IPTG, 20 h), no soluble protein was detected in the supernatant (Fig.14). Therefore, we constructed MBP-COI1-GFP11 to enhance the solubility of the COI1-GFP11 fusion protein through the MBP tag. However, soluble MBP-COI1-GFP11 was still not obtained under the tested conditions (Fig.15). We plan to switch to a yeast expression system or try the inclusion body refolding method to obtain soluble MBP-COI1-GFP11 proteins.

Fig.14 SDS-PAGE analysis of COI1-GFP11 purification using Ni-NTA affinity chromatography. Lane designations are the same as in Fig.2.

Fig.15 SDS-PAGE analysis of MBP-COI1-GFP11 purification using Ni-NTA affinity chromatography. Lane M: protein molecular weight marker; Lane U: whole cell lysate before IPTG induction; Lane I: whole cell lysate after IPTG induction; Lane P: insoluble fraction after sonication (pellet); Lane S: soluble fraction after sonication (supernatant); Lane FT: flow-through from Ni-NTA column; Lane E1-E4: elution fractions with 20, 50, 100, and 200 mM imidazole, respectively. (a) Induction conditions: 30°C, 0.5 mM IPTG, 200 rpm, 6 hours. (b) Induction conditions: Lane 2-5 22°C, 0.5 mM IPTG, 180 rpm, 16 hours; Lane 7-10 16°C, 0.3 mM IPTG, 160 rpm, 16 hours.

CC domains

The chain exchange reaction between CC domains forms the core of our system design. To confirm the occurrence of this reaction can occur and to evaluate its efficiency, we selected CC domains A', B, and A as representative components. We employed the CC-coupled Split GFP system, previously validated in our Split Protein as Output section, to characterize the chain exchange reaction between these domains.

We designed two fusion proteins, GFP10-A-TEVs-B and A'-GFP11. In the presence of TEV protease, the cleaved GFP10-A-TEVs-B is expected to undergo chain exchange with A'-GFP11, enabling the reconstitution of an active Split GFP complex upon binding to GFP1-9 (Fig.16).

Fig.16 Verification of the chain exchange reaction between CC domains.

To construct the experimental system, the coding sequences of GFP10-A-TEVs-B, A'-GFP11, and GFP1-9 were cloned into the pET-28a(+) vectors. The expression of GFP10-A-TEVs-B and A'-GFP11 was driven by an rhaB-inducible promoter, while GFP1-9 was expressed under the constitutive J23119 promoter (Fig.17).

Fig.17 Plasmid map of A'_B-A_GFP_pET-28a(+) construct.

However, the presence of multiple CC domains and repetitive sequences, such as double terminators, made this plasmid construct was difficult to assemble successfully. In future experiments, we plan to express these proteins separately using individual vectors, followed by in vitro assays after purification. Reflections and design improvements related to this construct are discussed in detail in our Engineering section.

Although we were unable to successfully construct the A'_B-A_GFP_pET-28a(+) vector, our design employed the same logic gate configuration as that described in the literature (Fink et al., 2019). The binding strengths between the CC domains we used were comparable to those of the CC domains they employed. They characterized the rearrangement between the two CC domains in a system similar to ours, A'-B-A and C-D-E-C', using cleaved luciferase and obtained significant luminescence results after 30 minutes (Fig.18).

Fig.18 Experimental analysis of the indicated SPOC logic function designs in HEK293T cells with introduced genetic circuits. The design and expected response to all input combinations is schematically shown below the graphs with experimental luciferase-based results. Input signals are combinations of two orthogonal proteases, TEVp and PPVp, and the output signal is split luciferase activity. Values are the means from two (a-b) cell cultures ±s.d. and are representative of two independent experiments. Significance was tested by one-way analysis of variance (ANOVA) with Tukey’s comparison (values of confidence intervals, degrees of freedom, F and P are indicated). (Fink et al., 2019)

Given that our system employs an in vitro approach, eliminating the need for protein accumulation through cellular expression, we anticipate that a shorter observation period will suffice to detect a pronounced luminescent signal. Moreover, the signal intensity remains unaffected by interference from complex intracellular environments.

Positive Feedback Loop

To verify whether our positive feedback pathway could amplify the AvrRpt2 signal and lower the limit of detection, we designed a series of fusing protein (Fig.19), including cTEVp*-B1-AvrRpt2s-TEVs-B1-nTEVp (BBa_25LMVZ4Y, abbreviated as nTEV), MBP-A'-cTEVp (BBa_258NBWZH, abbreviated as cTEV) and cXTEVp-B1-AvrRpt2s-B1-nTEVp (BBa_25LOPU94, abbreviated as nonTEV). The cTEV and nTEV protein are designed to achieve positive feedback, while nonTEV, in which TEV protease cleavage site was deleted, serves as a negative control lacking the function of positive feedback.

Fig.19 Schematic diagram of the positive feedback loop section. (a) Protein nTEV and cTEV exhibit positive feedback. (b) Protein nonTEV and cTEV fail to exhibit positive feedback

The fusion proteins nTEV, cTEV and nonTEV were expressed in E. coli BL21(DE3) and purified using Ni-NTA affinity chromatography. SDS-PAGE confirmed successful purification (Figs.20a, b and c). We also purified AvrRpt2 protease (Fig.8) as the input of the positive feedback loop.

Fig.20 SDS-PAGE analysis of nTEV (a), cTEV (b) and nonTEV (c) purification using Ni-NTA affinity chromatography. Lane M: protein molecular weight marker; Lane U: whole cell lysate before IPTG induction; Lane I: whole cell lysate after IPTG induction; Lane P: insoluble fraction after sonication (pellet); Lane S: soluble fraction after sonication (supernatant); Lane FT: flow-through from Ni-NTA column; Lane E1-E4: elution fractions with 20, 50, 100, and 200 mM imidazole, respectively.

We firstly co-incubated AvrRpt2 with nonTEV and nTEV at room temperature for two hours to preliminarily determine whether our purified AvrRpt2 had cleavage activity, thereby enabling the initiation of a positive feedback reaction. Theoretically, AvrRpt2-mediated cleavage of nTEV or nonTEV should yield two bands, each with a molecular weight of approximately 20 kDa. However, SDS-PAGE analysis (Fig.21) did not reveal a distinct band around 20 kDa in the co-incubated samples, indicating that AvrRpt2 did not perform effective cleavage to initiate the intended reaction.

Fig.21 SDS-PAGE analysis of AvrRpt2 and nTEV/nonTEV co-incubation. The molecular weight of AvrRpt2 is 30.3 kDa, while that of nTEV is 42.4 kDa and nonTEV is 41.2 kDa.

Through literature review (Chisholm et al., 2005; Jin, Wood, Wu, Xie, & Katagiri, 2003; Mudgett & Staskawicz, 1999), we found that the full-length AvrRpt2 (30 kDa) lacks cleavage activity. Upon entering plant cells, it is activated by Host Cleavage Factor(s), causing AvrRpt2 to cleave off its own N-terminal 70 amino acids, thereby generating an active AvrRpt2 with a molecular weight of approximately 25 kDa. We will subsequently attempt to replicate the work of Jin, Wood, Wu, Xie, & Katagiri (2003) by activating AvrRpt2 using yeast extract and employing the activated AvrRpt2 for further experiments.

References

1. Brechun, K. E., Arndt, K. M., & Woolley, G. A. (2019). Selection of Protein-Protein Interactions of Desired Affinities with a Bandpass Circuit. Journal of Molecular Biology, 431(2), 391-400.
2. Chisholm, S. T., Dahlbeck, D., Krishnamurthy, N., Day, B., Sjolander, K., & Staskawicz, B. J. (2005). Molecular characterization of proteolytic cleavage sites of the Pseudomonas syringae effector AvrRpt2. Proc Natl Acad Sci U S A, 102(6), 2087-2092.
3. Fink, T., Lonzarić, J., Praznik, A., Plaper, T., Merljak, E., Leben, K., ... Jerala, R. (2019). Design of fast proteolysis-based signaling and logic circuits in mammalian cells. Nat Chem Biol, 15(2), 115-122.
4. Jin, P., Wood, M. D., Wu, Y., Xie, Z., & Katagiri, F. (2003). Cleavage of the Pseudomonas syringae type III effector AvrRpt2 requires a host factor(s) common among eukaryotes and is important for AvrRpt2 localization in the host cell. Plant Physiol, 133(3), 1072-1082.
5. Mudgett, M. B., & Staskawicz, B. J. (1999). Characterization of the Pseudomonas syringae pv. tomato AvrRpt2 protein: demonstration of secretion and processing during bacterial pathogenesis. Molecular Microbiology, 32(5), 927-941.
6. Sheard, L. B., Tan, X., Mao, H., Withers, J., Ben-Nissan, G., Hinds, T. R., ... Zheng, N. (2010). Jasmonate perception by inositol-phosphate-potentiated COI1-JAZ co-receptor. Nature, 468(7322), 400-405.