Introduction

The Nano Luciferase (NanoLuc), introduced in 2013 by Promega, is a new member of the luciferase reporter gene/protein family. It can transfer the Furimazine (FMZ) and oxygen into Furimamide and carbon dioxide, generating a glow-type signal as output. NanoLuc has a smaller molecular weight (19.1kDa), brighter signal output (150-fold than conventional luciferases), longer signal half-life (about 120 min), and higher signal-to-noise ratio. So many advantages make the NanoLuc has excellent application prospects in bioluminescence imaging, exploring gene regulation and cell signaling, and investigating protein-protein interaction. Here we develop a new protocol for the use of NanoLuc, with a low-cost device for NanoLuc concentration measurement, a software with rear-ends separation structure data processing and visualization, and new kinetics which has higher generalizability. The new kinetics are not only suitable for a low-cost device like the camera but also the microplate readers proposed by Promega's protocol.

Protocol

Protocol for low-cost device

Materials

1. Several pieces of filter paper
2. Known concentration of NanoLuc luciferase
3. Nano-Glo® Luciferase Assay Substrate (keep on ice after take out from -20 ℃ refrigerator )
4. 1x PBS
5. Raspberry Pi Camera
6. Light-proof box

Preparation

Dilute NanoLuc luciferase with 1x PBS gradiently

Protocol

1. Place the pieces of filter paper equidistantly in the box and keep them within the camera's view.
2. Add 0.2 µl of Nano-Glo® Luciferase Assay Substrate to each piece.
3. Add 10 µl of NanoLuc luciferase of different concentrations simultaneously to each paper.
4. Put the lid on immediately and start taking videos.

Protocol for microplate readers

Materials ＆ Apparatus

1. 384-well plate microtiter plates
2. TECAN® Infinite M200 Pro Microplate Reader
3. Known concentration of NanoLuc luciferase
4. 1x PBS
5. Nano-Glo® Luciferase Assay Substrate (keep on ice after take out from -20 ℃ )

Preparation

Dilute NanoLuc luciferase with 1x PBS gradiently

Protocol

1. Add 0.2 µl of Nano-Glo® Luciferase Assay Substrate to the 384-well plate.
2. Add 10 µl of different concentrations of NanoLuc luciferase to the 384-well plate.
3. The relative luminescence unit (RLU) is measured by microplate reader (460 nm) with 120 cycles for every 5 seconds.

Experomental Result

Low-cost hardware

To derive a model for explaining the relationship between the luminescence decay kinetics and NanoLuc concentration, Video A is created as the learning data using purified NanoLuc (147.12 μg/mL), which has three NanoLuc concentrations (100%, 50%, and 10%). Each sample in video A contains 0.2μL FMZ and 10μL NanoLuc solution of defined concentration. Video A is attached as follows.

The data extracted using the back-end algorithm in the software is shown as follows (Fig. 1).

Fig. 1 The data extracted from Video A.
Video B has four NanoLuc concentrations (100%, 50%, 25%, and 10%), while Video C has two NanoLuc concentrations (100% and 30%). Each sample in the video contains 0.2μL FMZ and 10μL NanoLuc solution of defined concentration, which keeps the same conditions as Video A. After processing in the software, the data is shown in Fig. 2.

Fig. 2 Data collected using hardware. a Data of Video B. b Data of Video C.

Microplate Reader

The purified NanoLuc was diluted from 10³ times to 10⁸ times to gain different NanoLuc concentrations from nM to pM. The same protocols that remove buffer were used, while the monitor device was changed to the microplate reader since the NanoLuc concentration is too small to be detected by the camera. All the data collected is shown in Fig. 3.

Fig. 3 The data collected in the the microplate reader with different NanoLuc concentrations from nM to pM.
The data in Fig. 3 shows that the luminescence decay kinetics exist in the different NanoLuc concentrations from 7.7 nM to 0.77 pM.
The same exploration is also implemented for the FMZ. As for a detection protocol, the substrate in excess is an important condition. The limitation of FMZ concentration of our new protocol is also one of the key points. The FMZ is diluted using PBS in different ratios, including 1:1, 1:3, 1:7, 1:15, and 1:31. Using the same protocol and test in the microplate reader, the data is plotted in Fig. 4. With the decreasing of FMZ, the reaction rate is also decreasing, and the peak is becoming observable.

Fig. 4 The data collected in the the microplate reader with different FMZ concentrations.

Kinetic Equations

To interface the concentration of NanoLuc for our new protocol, a new ODE equation is derived using machine learning and genetic programming in the Model part. The formula of the NanoLuc luminescence decay kinetics is shown as following: $$\frac{du}{dt}=\alpha \cdot u^{1+\beta }$$ , where $\alpha $ is the learnable parameters related to enzyme activity and $\beta $ is the normalized NanoLuc concentration.

The ODE equation is discovered using data from Video A. Fig. 5 shows the simulation of Video A using the ODE above and its gap compared to experimental data in Video A.

The ODE equation is tested using data from Video B and Video C by the NanoLuc concentration inference. The inference result is shown in the following table. All the test samples have the error within 5%. Considering the data collected by the Hardware is noisy, the ODE equation is reliable.

The model's generalizability is tested in the data collected in the microplate reader with large fold change in the concentration of both NanoLuc and FMZ. By regarding all the sample as the standard sample (100% NanoLuc), the linearity between the numerical differentiation calculated from light intensity (du/dt) and the squared light intensity data (u^2) is checked in the Fig. 6 (data in Fig. 3) and Fig. 7 (data in Fig. 4).

When the concentration of NanoLuc decreases, the linearity is not guaranteed, the reason may be as follows: (1) when the light intensity is close to the LOD of the microplate reader, the data collected is oscillating, which can be seen in Fig. 3. (2) The numerical differentiation is not so reliable, especially for the oscillating data. A data filter will be reconmend in future data processing to deal with oscillation.

The linearity is not maintained in the high dilution rate due to the lack of FMZ.

All in all, our ODE equations is quite reliable with our new protocol, which expanded usable range of NanoLuc.