Colour-Switching Glow Lets Plants Report Gene Activity Despite Metabolic Noise

Bioluminescent plants that glow on their own, without any added chemicals, have become one of the most striking tools in modern plant biology. A self-sustained luminescence system lets researchers watch molecular events unfold inside a living organism without destroying it, simply by recording light over time. But the technology has long carried a frustrating flaw: the brightness of the glow depends not only on how active a gene is, but also on the metabolic state of the tissue doing the glowing. A plant under heat stress, infection or shifting energy conditions may dim for reasons that have nothing to do with the gene being studied, muddying the quantitative interpretation of the signal. A team of researchers reporting in Nature Plants has now engineered a way around this problem, using two luciferases that shine at different colours from the same substrate, so that the ratio between the two colours reveals gene activity while metabolic fluctuations cancel out.

The core idea is ratiometric imaging, a strategy familiar from fluorescence microscopy but difficult to implement in autoluminescent systems. The researchers co-express an engineered reference luciferase variant alongside the reporter luciferase. Both enzymes use the same fungal luciferin substrate, produced and recycled inside the plant by a dedicated metabolic pathway, but each emits light at a different wavelength. The reference luciferase is driven by a constitutive promoter, so its output tracks the overall metabolic capacity of the tissue: how much substrate is available, how healthy the cells are, how efficiently the light-producing reaction proceeds. The reporter luciferase, driven by a promoter that responds to a specific biological signal, adds its own variation on top. By dividing the signal at one wavelength by the signal at the other, frame by frame, the shared metabolic contribution largely drops out of the measurement.

The fungal bioluminescence system underlying the work comes from a pathway whose genes have been characterised and optimised for expression in plants. The researchers built their constructs with Golden Gate assembly, codon-optimising the bioluminescence genes for Nicotiana benthamiana and Arabidopsis thaliana. The pathway includes enzymes for producing the fungal luciferin precursor and for recycling caffeoylpyruvic acid, the product of the light-emitting oxidation, back into the substrate. This recycling loop is what makes the luminescence self-sustained: the plant continuously regenerates the fuel for its own glow. A ‘luciferase-less’ master line of Arabidopsis, designated AT9865, already carried all the genes needed for fungal luminescence except the luciferase itself, allowing the team to cross in different luciferase reporters and generate stable lines expressing defence-related or jasmonic-acid-responsive reporters.

The colour variants at the heart of the ratiometric design are engineered luciferases named Luz533, Luz593, Luz608 and Luz640, named for their emission wavelengths in nanometres. In their validation experiments, the researchers paired a reporter such as Luz533, emitting greenish light, with a reference such as Luz608, emitting at a longer wavelength. Image analysis was performed in Fiji (ImageJ) and custom Python scripts: images were converted to 32-bit format, a 580-nanometre stack was thresholded to remove background, and the 520-nanometre stack was divided by the 580-nanometre stack frame by frame, producing a pseudocoloured ratio image using the Turbo lookup table. The ratio was then calculated for regions of interest, giving a quantitative readout that is far less sensitive to the overall brightness of the tissue than either channel alone.

To test how well the colour ratio resists metabolic disturbance, the team used cell packs of Nicotiana tabacum BY-2 cultured cells, transiently transformed with mixtures of Agrobacterium strains carrying the bioluminescence genes. They applied two kinds of stress: infiltration with buffer alone and heat shock at 37 degrees Celsius for 30 minutes. For both fungal and bacterial bioluminescence colour variants, they assessed the stability of the wavelength ratio under these treatments, comparing ratiometric and non-ratiometric configurations. The experiments were imaged over 48 hours using a consumer-grade Sony camera, a detail that underscores how accessible the technique has become: no specialised ultra-sensitive equipment is required for many applications, though the infection experiments were also imaged on an IVIS Spectrum CT imager with 520- and 580-nanometre filters.

A further technical question concerned the stability of the luciferase enzymes themselves. Because the ratio method assumes that the two luciferases behave comparably over time, differences in protein degradation could distort the measurement. The researchers ran cycloheximide chase assays on BY-2 cell packs constitutively expressing Luz533 or Luz608, blocking new protein synthesis and supplying excess fungal luciferin so that the decay of luminescence reflected enzyme turnover alone. Signals measured at 0, 5.5, 10, 24 and 54 hours were fitted to a mono-exponential decay model, and half-lives were calculated from the decay constants. Structural predictions for Luz593, Luz608 and Luz640 were performed with AlphaFold3, helping to rationalise the engineered colour shifts.

With the principle validated, the team built functional ratiometric reporters for plant signalling. One used the TCSv2 promoter, which responds to cytokinin, a hormone controlling cell division and growth. In BY-2 cell packs, the cytokinin reporter was challenged with a range of thidiazuron (TDZ) concentrations from 0.01 nanomolar to 1 micromolar, and signal variation was evaluated by comparing luminescence at the response peak, 7 hours after treatment, with the signal at 72 hours. In N. benthamiana leaves, the ratiometric TCSv2-Luz533 reporter was co-expressed with constitutive Luz608, and leaves were treated with TDZ at 0, 10 and 100 micromolar and 1 millimolar concentrations, then imaged for up to 76 hours. Reverse-transcription quantitative PCR confirmed that the luminescent output tracked the reporter transcript levels.

The most demanding test came in stable transgenic Arabidopsis. Using T1 plants, the researchers monitored defence-related and early jasmonic-acid-responsive reporters during infection with Pseudomonas savastanoi, a hemibiotrophic pathogen in the Pseudomonas syringae group. Leaves were infiltrated with bacterial solution at an optical density of 0.4, with buffer alone as a mock control, and luminescence was recorded with 520- and 580-nanometre filters. In stable ratiometric reporter lines based on the WRKY70 promoter, a regulator of salicylic-acid-mediated defence, plants sprayed with 1 millimolar acetylsalicylic acid showed the expected induction, and the team sampled tissue at 0, 16 and 24 hours post treatment to relate luminescence to Luz533 transcript levels by RT-qPCR, using the housekeeping gene EF1a as a reference. The ratiometric readout tracked the hormonal response even as the infection and the spray treatments altered the metabolic landscape of the tissue.

The practical significance of this work lies in what it makes possible for whole-organism physiology. Imaging was performed in a dark chamber with a Sony Alpha ILCE-7M3 camera and a fast 35-millimetre lens, with exposure times of 5 to 30 seconds and ISO settings from 400 to 20,000, capturing frames roughly every 30 minutes over 48-hour sessions. This means a laboratory can follow hormone signalling, pathogen responses and gene expression dynamics in intact, growing plants over days, with quantitative confidence that was previously unattainable with single-channel autoluminescence. The red-shifted luciferase variants have been deposited with Addgene, plasmid maps are available in GenBank format, raw imaging data are hosted on Figshare, and the analysis scripts are published on GitHub, making the full toolkit reproducible and adaptable.

Limitations remain, as the authors’ own methods make clear. The ratiometric approach requires that both luciferases share the same substrate pool and that their relative stability be characterised; the cycloheximide chase data and the enzyme half-life modelling address exactly this dependency. The reported validations were carried out in tobacco cell cultures, N. benthamiana leaves and Arabidopsis seedlings and leaves, and the statistical analyses relied on non-parametric Mann-Whitney U tests and one-way ANOVA with Tukey’s correction, with sample numbers reported per experiment. Extending the colour-ratio principle to other organisms and other signalling pathways will require similar calibration for each new promoter pair and luciferase combination. Even so, the demonstration that a metabolic reference channel can be built into the autoluminescent system itself marks a substantial step toward quantitative, non-invasive imaging of molecular events in whole living organisms.

Subject of Research: Ratiometric autoluminescent reporters for metabolically robust imaging of gene expression in plants

Article Title: Metabolically robust ratiometric autoluminescent reporters

Article References: Morozov, V. V., Balakireva, A. V., Gao, F., Fleiss, A., Mishina, N. M., Perfilov, M. M., Mitiouchkina, T. Y., Palkina, K. A., Gorbachev, D. A., Gavrikov, A. S., Kotik, A. A., Mislavskiy, S. M., Wood, K. V., Yampolsky, I. V., Sarkisyan, K. S., & Mishin, A. S. (2026). Metabolically robust ratiometric autoluminescent reporters. Nature Plants. https://doi.org/10.1038/s41477-026-02437-y

Image Credits: AI Generated

DOI: 10.1038/s41477-026-02437-y

Keywords: bioluminescence, ratiometric reporter, fungal luciferase, plant imaging, Arabidopsis thaliana, jasmonic acid, cytokinin, gene expression, luciferin recycling, Nicotiana benthamiana, plant defence, whole-organism imaging