Brief Communication | Published:

StaPLs: versatile genetically encoded modules for engineering drug-inducible proteins

Nature Methodsvolume 15pages523526 (2018) | Download Citation

Abstract

Robust approaches for chemogenetic control of protein function would have many biological applications. We developed stabilizable polypeptide linkages (StaPLs) based on hepatitis C virus protease. StaPLs undergo autoproteolysis to cleave proteins by default, whereas protease inhibitors prevent cleavage and preserve protein function. We created StaPLs responsive to different clinically approved drugs to bidirectionally control transcription with zinc-finger-based effectors, and used StaPLs to create single-chain, drug-stabilizable variants of CRISPR–Cas9 and caspase-9.

Access optionsAccess options

Rent or Buy article

Get time limited or full article access on ReadCube.

from$8.99

All prices are NET prices.

Additional information

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

  1. 1.

    Picard, D. Methods Enzymol. 327, 385–401 (2000).

  2. 2.

    Rakhit, R., Navarro, R. & Wandless, T. J. Chem. Biol. 21, 1238–1252 (2014).

  3. 3.

    Armstrong, C. M. & Goldberg, D. E. Nat. Methods 4, 1007–1009 (2007).

  4. 4.

    Liu, Y. C. & Singh, U. Int. J. Parasitol. 44, 729–735 (2014).

  5. 5.

    Gray, D. C., Mahrus, S. & Wells, J. A. Cell 142, 637–646 (2010).

  6. 6.

    Zetsche, B., Volz, S. E. & Zhang, F. Nat. Biotechnol. 33, 139–142 (2015).

  7. 7.

    McCauley, J. A. & Rudd, M. T. Curr. Opin. Pharmacol. 30, 84–92 (2016).

  8. 8.

    Butko, M. T. et al. Nat. Neurosci. 15, 1742–1751 (2012).

  9. 9.

    Lin, M. Z., Glenn, J. S. & Tsien, R. Y. Proc. Natl Acad. Sci. USA 105, 7744–7749 (2008).

  10. 10.

    Chung, H. K. et al. Nat. Chem. Biol. 11, 713–720 (2015).

  11. 11.

    Giacca, M. & Zacchigna, S. Gene Ther. 19, 622–629 (2012).

  12. 12.

    Ferrara, N. & Adamis, A. P. Nat. Rev. Drug. Discov. 15, 385–403 (2016).

  13. 13.

    Liu, P. Q. et al. J. Biol. Chem. 276, 11323–11334 (2001).

  14. 14.

    Chavez, A. et al. Nat. Methods 12, 326–328 (2015).

  15. 15.

    Zhou, X. X. et al. ACS Chem. Biol. 13, 443–448 (2018).

  16. 16.

    Straathof, K. C. et al. Blood 105, 4247–4254 (2005).

  17. 17.

    Gargett, T. & Brown, M. P. Front. Pharmacol. 5, 235 (2014).

  18. 18.

    Gao, Y. et al. Nat. Methods 13, 1043–1049 (2016).

  19. 19.

    Pu, J., Kentala, K. & Dickinson, B. C. ACS Chem. Biol. 13, 431–437 (2018).

  20. 20.

    Heinis, C. & Johnsson, K. Methods Mol. Biol. 634, 217–232 (2010).

  21. 21.

    Stennicke, H. R. et al. J. Biol. Chem. 274, 8359–8362 (1999).

  22. 22.

    Yang, Z. et al. J. Struct. Biol. 179, 269–278 (2012).

  23. 23.

    Soumana, D. I., Ali, A. & Schiffer, C. A. ACS Chem. Biol. 9, 2485–2490 (2014).

  24. 24.

    Romano, K. P. et al. PLoS. Pathog. 8, e1002832 (2012).

  25. 25.

    Jiang, F., Zhou, K., Ma, L., Gressel, S. & Doudna, J. A. A. Science 348, 1477–1481 (2015).

  26. 26.

    Tighe, A., Staples, O. & Taylor, S. J. Cell. Biol. 181, 893–901 (2008).

  27. 27.

    Schindelin, J. et al. Nat. Methods 9, 676–682 (2012).

  28. 28.

    Edelstein, A. D. et al. J. Biol. Methods 1, e10 (2014).

Download references

Acknowledgements

We thank T. Knaak and the Stanford Shared FACS Facility for their assistance with flow cytometry experiments, and the M. Kay lab for use of their real-time thermocycler. We are grateful to L.S. Qi (Stanford University, Stanford, CA, USA), A. Straight (Stanford University, Stanford, CA, USA), D. Spencer (Baylor College of Medicine, Houston, TX, USA), and G. Salvesen (Sanford Burnham Prebys Medical Discovery Institute, San Diego, CA, USA) for providing cell lines and plasmids. Flp-In T-REx HeLa cells were a gift from the S. Taylor lab (University of Manchester, Manchester, UK). We also thank V. Duong, Y. Geng, X. Zhou, H. Chung, L. Ning, Y. Yang, and other members of the Lin lab for advice and helpful discussions, and we thank S. Dixon for feedback. This work was supported by a Stanford Graduate Fellowship and NSF Graduate Research Fellowship (C.L.J.), a Stanford Bio-X Undergraduate Summer Research grant and Stanford UAR Major grant (R.K.B.), and an NIH/NIGMS EUREKA grant 5R01GM098734 (M.Z.L.).

Author information

Author notes

    • Ryan K. Badiee

    Present address: MD Program, University of California San Francisco School of Medicine, San Francisco, CA, USA

Affiliations

  1. Graduate Program in Biological Sciences, Stanford University, Stanford, CA, USA

    • Conor L. Jacobs
  2. Department of Neurobiology, Stanford University, Stanford, CA, USA

    • Conor L. Jacobs
    • , Ryan K. Badiee
    •  & Michael Z. Lin
  3. Department of Bioengineering, Stanford University, Stanford, CA, USA

    • Conor L. Jacobs
    • , Ryan K. Badiee
    •  & Michael Z. Lin
  4. Department of Biology, Stanford University, Stanford, CA, USA

    • Ryan K. Badiee
  5. Department of Pediatrics, Stanford University, Stanford, CA, USA

    • Michael Z. Lin

Authors

  1. Search for Conor L. Jacobs in:

  2. Search for Ryan K. Badiee in:

  3. Search for Michael Z. Lin in:

Contributions

M.Z.L. conceived of the study. C.L.J. designed orthogonal StaPL modules and StaPL-controlled proteins. R.K.B. and C.L.J. constructed plasmids and designed and carried out mammalian cell experiments. M.Z.L. and C.L.J. wrote the manuscript, with contributions from R.K.B.

Competing interests

M.Z.L., C.L.J., and R.K.B. have filed a provisional patent (Application 62/536307) with the US Patent and Trademark Office for compositions and methods for inducing protein function.

Corresponding author

Correspondence to Michael Z. Lin.

Integrated supplementary information

  1. Supplementary Figure 1 Development of orthogonal NS3 proteases.

    (a) Top, schematic of PSD95 fused to a SMASh cassette comprising an HCV NS3 protease cleavage site, NS3 protease, and a degron. Protease activity removes the degron, enabling PSD95 accumulation. Inhibition of the protease prevents degron removal. Center and below, NS3 with a T54A mutation is less well inhibited by 22 h TPV treatment than NS3 wildtype (wt) in transiently transfected HeLa cells, as evidenced by increased release of PSD95. Note the uncleaved protein has accelerated degradation but a small amount can be detected in some cases, as seen here. Data represent a single experiment. The two immunoblots are from the same experiment and are subsets of one membrane probed with the indicated antibodies. (b) Effects of single mutations and combinations of multiple mutations on the ability of TPV and ASV to block SMASh degron removal were tested by transient transfections of HEK293A cells in indicated drug for 22 h, and subsequent immunoblotting. Representative blots are shown. Data represent 3 separate wells for each condition, with transfections and immunoblots performed in parallel. (c) Mean inhibition of protease by varying doses of ASV and TPV was quantified from immunoblot experiments as in panel (b). NS3 activity is shown as the degree of escape from protease inhibition. For each construct, values were normalized to the DMSO (non-drug) condition. Error bars, s.e.m.

  2. Supplementary Figure 2 NS3AI and NS3TI protease variants can be regulated orthogonally by two different inhibitor drugs.

    (a) Dose-inhibition curves for NS3AI (V36M T54A S122G), NS3TI (F43L Q80K S122R D168Y), and wildtype NS3 with varying doses of ASV and TPV, quantified as degree of NS3 protease escape from drug inhibition in PSD95-SMASh variants transiently expressed in HEK293A cells. Values are normalized to the DMSO (non-drug) condition, and displayed as mean ± s.e.m. **P < .001, ***P < .0001 by 2-way ANOVA (see Online Methods for P values). Data including all single mutant variants are shown in Supplementary Figure 1b-c and represent 3 separate wells for each condition, with transfections and immunoblots performed in parallel. (b) HCV NS3 protease inhibitors used in this study. The asterisk marks the site of covalent bond formation between TPV and the catalytic serine. P2 moieties for each peptidomimetic compound are indicated. (c) Models of NS3AI and NS3TI were created by mutagenesis of published X-ray structures of NS3-ASV and NS3-TPV co-crystals (PDB entries 4WF8 and 3SV6), followed by energy minimization. (d) Sequences of original SMASh, SMAShAI, and SMAShTI tags. Tags intended for fusion at the C-terminus of a protein of interest are shown, such that the cleavage site (green) is at the N-terminus of the tag. Within the NS3 protease domain, the sequence found in the original SMASh tag (which contains a T54A mutation relative to wildtype NS3) is shown in black, mutations found in SMAShAI are in cyan, and mutations found in SMAShTI are in orange. Grey indicates artificial linker sequences. Dotted line above sequence indicates a segment required for degradation of SMASh tags.

  3. Supplementary Figure 3 StaPL sequence and initial validation.

    (a) Sequence of a StaPL cassette. Within the NS3 protease domain, wildtype sequences are shown in black. Mutations found in StaPLAI are in cyan, while mutations found in StaPLTI are in orange. Cleavage sites (red), derived from the NS4A/NS4B junction or the NS5A/NS5B junction, are shown at the beginning of the StaPL module, but can be placed at the C-terminus in addition or instead of the N-terminus. Green, NS4A-derived cofactor strand. Grey, artificial linker sequences. Brown, spacer sequence derived from the NS3 helicase domain. (b) StaPLAI and StaPLTI sequences are orthogonally controlled by ASV and TPV in HEK293A cells which transiently expressed tdYFP and/or tdRFP StaPL-NLS variants for 8 h. Expected cleavage patterns are observed, and StaPLAI/StaPLTI control by ASV/TPV is regulated equally well on tdYFP versus tdRFP. Data represent a single experiment. Corroborating data were obtained in an independent microscopy experiment (Supplementary Fig. 4).

  4. Supplementary Figure 4 Orthogonal StaPL modules allow for independent, simultaneous control of nuclear localization for two fluorescent proteins.

    Top, schematics of constructs used for small molecule control of nuclear localization. Tandem YFPs or RFPs were fused to a nuclear localization signal (NLS), with an intervening StaPL module and cleavage site governing preservation of the NLS. Constructs were transiently expressed in HEK293A cells and incubated 8-12 h in the indicated drugs. Left, for tdYFP-StaPLAI-NLS, concentrated nuclear YFP fluorescence is observed for ASV but not for TPV or DMSO vehicle. For tdRFP-StaPLTI-NLS, concentrated nuclear RFP fluorescence is observed in TPV but not for ASV or vehicle. Right, the opposite sensitivity to TPV and ASV is observed with tdYFP-StaPLTI-NLS and tdRFP-StaPLAI-NLS. Representative confocal images of live cells acquired 8-12 h after transfection are shown. Scale bars, 15 μm. Data represent a single experiment.

  5. Supplementary Figure 5 Optimization of ZFVEGFA-StaPL effectors.

    (a) Left top, initial design of the ZFVEGFA-StaPLAI transcriptional activator. StaPLAI links a zinc-finger domain recognizing the VEGFA promoter (ZFVEGFA) to the p65 activation domain. Right top, design of the ZFVEGFA-StaPLTI transcriptional repressor. StaPLTI links ZFVEGFA to the KRAB repressor domain. An HA tag on ZFVEGFA allows detection by immunoblot. Initial designs tested EDVVCC/H and DEMEEC/S substrate sequences. Left bottom, in HEK293A cells treated by drug or vehicle (DMSO) for 24 h after transfection, full-length ZFVEGFA-StaPLAI-tdYFP-p65 is preserved only in ASV, as expected, while cleavage occurs in TPV or DMSO. Both substrates worked well. Right bottom, EDVVCC/H is cleaved less efficiently than DEMEEC/S in the absence of drug, as seen in the full-length RFP-positive band. With DEMEEC/S, full-length ZFVEGFA- StaPLTI-tdRFP-KRAB is preserved only in TPV, as expected, and cleavage occurs in ASV or DMSO. Data are derived from a single experiment. (b) ZFVEGFA-StaPLAI-tdYFP-p65 and ZFVEGFA-StaPLTI-tdRFP-KRAB were expressed in HEK293A cells for 48 h and media supernatants were analyzed for secreted VEGF protein by ELISA. The repressor construct exhibited clear drug-dependent activity, whereas activator function was less clear. Measurements were performed in duplicate and values were averaged. VEGF concentrations were calculated relative to similarly treated cells transfected with empty vector (values were 385 pg/mL for DMSO, 294 pg/mL for ASV, and 645 pg/mL for TPV). To further test repression, CoCl2 was applied for the final 7 h to increase basal VEGF secretion. (c) Left top, final design for a ASV-dependent activator, ZFVEGFA-StaPLAI-YFP-VPR. Right top, final design for TPV-dependent repressor ZFVEGFA-StaPLTI-tdRFP-KRAB. Left bottom, in HEK293A cells treated by drug or DMSO for 48 h after transfection, full-length activator is preserved only in ASV, although it is weakly expressed. Upper and lower ZFVEGFA (anti-HA) panels are from the same blot, but intensity scaling for the upper panel is 4-fold tighter to show the full-length band more clearly. Right bottom, full-length repressor is preserved only in TPV. Data are derived from a single experiment.

  6. Supplementary Figure 6 Sequences of ZF-StaPL effectors with StaPL modules linking DNA-binding and transcriptional regulatory sequences.

    (a) ZFVEGFA-StaPLAI-YFP-VPR, and (b) ZFVEGFA-StaPLTI-tdRFP-KRAB. Grey, artificial linkers. Blue, HA tag. Purple, StaPL module, with the portion of the protease substrate N-terminal to the cleavage site underlined.

  7. Supplementary Figure 7 Internal loops in SpCas9 can accommodate StaPL modules, making its function drug-regulatable.

    (a) Two sites within dSpCas9 were tested for their ability to tolerate an inserted StaPLTI module, such that it would permit protein function in the presence of TPV, but abolish dSpCas9 function in either ASV or DMSO (vehicle). (b) HEK293 cells stably expressing the TRE3G-mCherry reporter cassette, and incubated in the indicated drugs, were fixed 48 h after transfection with VPR-dSpCas9 transcriptional activators ± sgRNA targeting the TRE3G locus. Increased expression of mCherry RFP indicates transcriptional activation in the TPV condition. Scale bars, 200 μm. Data represent a single experiment. Corroborating results were obtained in 2 independent RT-qPCR experiments (data not shown).

  8. Supplementary Figure 8 Sequence of a drug-controllable dSpCas9-based transcriptional activator, VPR-dSpCas9(StaPLTI).

    The StaPLTI module is inserted after residue 1246 of dSpCas9. Grey, artificial linkers. Blue, HA epitope tag. Purple, StaPL module, with the portion of the protease recognition substrate N-terminal to the cleavage site underlined.

  9. Supplementary Figure 9 Kinetics, reversibility, and dose responsiveness of StaPL-mediated transcriptional activators.

    (a) Activation time courses are similar between VPR-dCas9(StaPLTI) and constitutive VPR-dCas9 after transient transfection in HEK293-TRE3G-mCherry cells. Values were normalized to peak RFP fluorescence for each condition. Mean ± s.e.m. are plotted. Dots represent values from 3 independent experiments. (b) Kinetics of mCherry RFP transcriptional activation by VPR-dCas9(StaPLTI) in 10 μM TPV or in DMSO vehicle control, as measured by RT-qPCR. HEK293-TRE3G-mCherry cells expressed VPR-dCas9(StaPLTI) and sgRNA for indicated time. Gene expression values were normalized to those of drug- and time-matched cells that were transfected with empty vector. Mean ± s.e.m. are plotted. Dots represent values from 3 independent experiments. (c) Kinetics of VEGFA transcriptional activation by ZFVEGFA-StaPLAI-YFP-VPR in 1 μM ASV or in DMSO vehicle control. HEK293A cells expressed ZFVEGFA-StaPLAI-YFP-VPR for indicated time. Gene expression values were normalized to those of drug- and time-matched cells that were transfected with empty vector. Mean ± s.e.m. are plotted. Dots represent values from 3 independent experiments. (d,e) RFP transcriptional activation induced by VPR-dCas9(StaPLTI), sgRNA, and TPV in HEK293-TRE3G-mCherry cells (d) is not fully reversible after drug washout, while VEGFA transcriptional activation induced by ZFVEGFA-StaPLAI-YFP-VPR and ASV in HEK293A cells (e) is reversible, reflecting the different mechanisms of action of TPV and ASV. Values were normalized to drug- and time-matched cells that were transfected with empty vector alone. Mean ± s.e.m. are plotted. Dots represent values from 3 independent experiments. (f) Dose-response relationship between TPV and mCherry RFP transcriptional activation by VPR-dCas9(StaPLTI) and sgRNA. HEK293-TRE3G-mCherry cells expressed VPR-dCas9(StaPLTI) and sgRNA for 48 h in indicated drug condition. Gene expression values were normalized to those of drug-matched cells that were transfected with empty vector. Mean ± s.e.m. are plotted. Dots represent values from 3 independent experiments. (g) Dose-response relationship between ASV and VEGFA transcriptional activation by ZFVEGFA-StaPLAI-YFP-VPR. HEK293A cells expressed ZFVEGFA-StaPLAI-YFP-VPR for 48 h in indicated drug condition. Gene expression values were normalized to those of drug-matched cells that were transfected with empty vector. Mean ± s.e.m. are plotted. Dots represent values from 7 independent experiments.

  10. Supplementary Figure 10 Sequence of a drug-inducible initiator caspase, StaPLd-Casp9.

    Grey, artificial linkers. Blue, HA epitope tag. Purple, StaPLAI module, with the portion of protease recognition substrate N-terminal to the cleavage site underlined.

  11. Supplementary Figure 11 Initial characterization of the StaPLd-Casp9 suicide switch.

    (a) In HeLa Flp-In cells stably expressing catalytically inactive StaPLd-Casp9C287S for 24 h, full-length tandem dimer, revealed by HA blotting, is observed with ASV but not DMSO treatment. However, when cells expressing active StaPLd-Casp9 are treated with ASV, no signal is detected by blotting to HA, β-actin, or coexpressed RFP, suggesting cell loss. Immunoblots are from the same experiment and are subsets of one membrane. Data represent a single experiment. Corroborating results were obtained with wells treated and assayed in parallel (data not shown). (b) Visible rounding and lifting occurs after a 24 h in ASV for cells expressing StaPLd-Casp9 but not StaPLd-Casp9C287S. Scale bar, 500 μm. Data represent a single experiment. Similar results for live StaPLd-Casp9 were obtained in an independent experiment (data not shown). (c) Gating strategy used to process flow cytometry data. Events were first plotted as forward scatter amplitude (FSC-A) vs. side scatter amplitude (SSC-A) to exclude low-scatter debris-like events. Next they were plotted as FSC height (FSC-H) vs. FSC-A to exclude non-singlet events with an inconsistent FSC-H to FSC-A ratio. (d) Flow cytometry of live, annexin-stained stable StaPLd-Casp9-expressing Flp-In HeLa cells reveals extensive apoptosis after 24 h incubation in ASV, as evidenced by annexin signal. HeLa cells stably expressing a catalytically inactive C287S variant serves as a control, and coexpressed RFP reports StaPLd-Casp9 expression. Comparison to parental Flp-In HeLa cells confirms the stable cells were RFP positive. A representative experiment is shown. Comparable results were obtained in a second independent experiment (data not shown).

  12. Supplementary Figure 12 Apoptosis, as assessed by annexin staining, occurs in a dose-dependent manner in StaPLd-Casp9-expressing cells.

    (a) Representative flow cytometry plots testing responsiveness of StaPLd-Casp9 by asunaprevir (ASV) dose. Inactive StaPLd-Casp9C287S serves as a negative control for drug-only effects. Staurosporine (STS) is a robust inducer of apoptosis and serves as a positive control for annexin staining of apoptotic cells. RFP signal serves as an indicator for StaPLd-Casp9 expression. Cells were stained and fixed for flow cytometry analysis after 24 h in drug. Representative plots are shown. (b) Quantification of dose responsiveness of StaPLd-Casp9. Mean ± s.e.m. are graphed. Grey dots represent values from 3 independent experiments.

  13. Supplementary Figure 13 Comparison of StaPLd-Casp9 and iCasp9.

    StaPLd-Casp9 and iCasp9 are similarly effective at inducing cell death after a 24 h incubation in asunaprevir (ASV) and AP20187 (AP), respectively. Moreover, each suicide gene is exclusively activated by its own activating drug and not that of the other. Casp9C287S variants of each construct serve as negative controls for drug-only effects. Staurosporine (STS) is a robust inducer of apoptosis and serves as a positive control for annexin staining of apoptotic cells. RFP signal serves as an indicator for suicide gene expression. Cells were stained and fixed for flow cytometry analysis after 24 h in drug. Similar data were obtained in a second independent experiment (data not shown).

  14. Supplementary Figure 14 Time courses of apoptosis induced by StaPLd-Casp9 or iCasp9.

    (a) Representative flow cytometry plots showing time courses of apoptosis following drug activation of StaPLd-Casp9 or iCasp9. Parallel time courses were also performed with DMSO vehicle control to account for any cell death due to nutrient depletion over time. RFP signal serves as an indicator for suicide gene expression. Cells were stained and fixed for flow cytometry analysis after indicated time in drug. (b) Time course of cell death induced by activation of StaPLd-Casp9 or iCasp9. StaPLd-Casp9 takes 24 h to reach near-maximal apoptosis, whereas iCasp9 does so by 3 h. Mean ± s.e.m. are plotted. Dots represent values from 3 independent experiments.

  15. Supplementary Figure 15 Uncropped versions of immunoblots.

    All immunoblots were probed using fluorophore-conjugated secondary antibodies, for detection in the 800-nm and 700-nm channels. Each membrane is shown with its 800-nm scan on the left and its 700-nm scan on the right.

Supplementary information

  1. Supplementary Text and Figures

    Supplementary Figs. 1–15, Supplementary Notes 1–3 and Supplementary Tables 1–3

  2. Reporting Summary

About this article

Publication history

Received

Accepted

Published

DOI

https://doi.org/10.1038/s41592-018-0041-z

Further reading