(443e) A Split Enzyme-Based Self Amplification System for Ultrasensitive Detection of Proteins and Small Molecules at the Point of Care

Introduction: Rapid, inexpensive detection of biomarkers at the point of care is vital for many clinical purposes. While the lateral flow assay (LFA) has allowed rapid diagnostic information to be gathered without equipment for many conditions, limitations in the LFA platform have prevented its application to detecting protein biomarkers at low concentrations, forcing clinicians to rely either on potentially inaccurate empirical diagnosis or expensive, slow lab tests to make treatment decisions.¹ Sensitive detection of nucleic acid targets has been readily achieved by amplifying signals, but there has been a lack of innovation for detection of low concentration antigens and small molecules at the point of care. Biology, on the other hand, has evolved intricate mechanisms for rapidly amplifying protein signals in vivo via post-translational modification and protein based signaling networks. We have developed a novel sensing platform that leverages split enzyme engineering and a novel autocatalytic feedback loop made up of engineered protein components that results in rapid, bistable, ultrasensitive in vitro sensing of biomarkers.

Experimental Materials and Methods: The detection platform uses split adenylate cyclase to detect the presence of an analyte of interest in vitro. The analyte mediates interactions of the two enzyme component halves fused to binding domains (i.e., a sandwich assay in solution), resulting in cAMP accumulation. Proof of concept demonstration has been performed by detecting rapamycin, a small molecule, via the FKBP12-rapamycin-FRB complex; we have engineered protein constructs with split halves of adenylate cyclase (nAC and cAC) fused to either the FRB or FKBP12 protein domains, such that in the presence of rapamycin, adenylate cyclase is reconstituted. Detection of cAMP is achieved via a FRET-based cAMP biosensor that shifts fluorescence upon cAMP binding.

Experimental Results: When the split adenylate cyclase-rapamycin sensing fusions were mixed with a serial dilution of rapamycin, the system transitioned from an OFF state to an ON state between the concentrations of 10 and 100nM (Fig. 1a). The OFF state had comparable activity levels to the split adenylate cyclase negative controls (split adenylate cyclase fragments with no binding domains), and the ON state had comparable activity to the positive controls (split adenylate cyclase fragments fused to self-assembling coiled-coil domains). The biosensor has been experimentally validated and shows an immediate shift in fluorescence in the presence of cAMP (Fig. 1b).

Theoretical Introduction of Bistability: We have further designed a theoretical autocatalytic feedback loop that when incorporated with the split adenylate cyclase sensing fusions and the FRET-based cAMP biosensor gives rise to bistable response dynamics. Bistability generally requires two network conditions: (1) signal ultrasensitivity and (2) positive and/or double-negative feedback loops. To enable a positive feedback motif, we have designed a feed-forward system composed of split adenylate cyclase fused to cAMP receptor protein (CRP) and small DNA fragments that act as scaffolds. CRP cooperatively binds two cAMP molecules, driving it to bind to DNA at specific sequences. Short DNA fragments with multiple repeats of the promotor region sequence for CRP act as a scaffold for multiple CRP fusions to bind, reconstituting additional adenylate cyclase, driving increased production of cAMP, and further reconstituting adenylate cyclase fused with CRP. Activation of the system can be tuned to trigger predetermined analyte concentrations by the inclusion of phosphodiesterase (PDE), degrading cAMP to inactive AMP. Therefore, the system triggers only when the rate of cAMP generation reliably exceeds that of degradation by PDE.

Computational Results: We have developed an ordinary differential equation-based model to simulate the dynamics of the protein-protein interactions in this assay using available kinetic parameters of PDE and in vivo expression of split adenylate cyclase. This integrated model combines the initial antigen sensing and FRET-based sensing with the cAMP autocatalytic feedback loop. ON/OFF behavior across a range of the analyte concentrations was achieved by tuning PDE concentration (Fig. 1c). Importantly, the modeled system also showed bistability, as illustrated by the hysteresis in the modeled dose response curve of activated biosensor across antigen concentrations (Fig. 1d); once the system is activated at the “trigger” concentration, the antigen concentration would need to decrease by an order of magnitude for the system to revert to the OFF state.

Conclusions: Here, we have demonstrated proof-of-concept for a novel sensing platform for protein and small molecule biomarkers. Because of the rapid turnover of cAMP synthesis to amplify the final signal, detection is not limited by 1-to-1 protein analyte-to-color change agent as in LFAs, but rather directly by the K_D­ of the analyte and binding domains with amplification through enzyme turnover. This work is significant because our system bypasses many of the time and infrastructure barriers preventing the implementation of molecular diagnostics at the point of care, while providing sensitive detection of proteins and small molecules in less than a minute in a simple to measure fluorometric output. Further, this system will be broadly applicable for low-concentration antigen detection and can be used to detect a wide range of target analytes with known antibody binding domains. As such, this system could be used a platform for the detection of many analytes currently unable to be rapidly detected at the point of care.

Figure 1. (A) Rapamycin sensing split adenylate cyclase fusions (nAC-FRB and cAC-FKBP, solid line) were incubated at 37°C across a dilution of rapamycin concentrations. All reactions were run in triplicate for 30 minutes. Accumulated cAMP was measured using a cAMP ELISA. Dashed line: positive control fusions (split adenylate cyclase fragments fused to self-assembling coiled-coil domains). Dotted line: negative control (split adenylate cyclase fragments with no binding domains). (B) The FRET-based cAMP biosensor shows an immediate shift in fluorescence emission wavelength (excitation: ECFP=430nm, emission: ECFP=475nm and EYFP=529nm) upon addition of cAMP (solid line) compared to fluorescence emission prior to cAMP addition (dashed line). (C) Simulated results using kinetic parameters measured in vivo show that as the concentration of PDE increases, higher concentrations of antigen (rapamycin) are necessary to turn the system ON, as shown by the proportion of biosensor that is activated. This demonstrates critical “trigger” concentration tunability. (D) The simulated dose response curve shows hysteresis, indicating bistability, a key design goal.

Acknowledgements: Research reported in this abstract was supported by the National Institute of Biomedical Imaging and Bioengineering of the National Institutes of Health under Award Number F32EB031608. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.