The authors frame the problem as a fundamental dialectic. Synthetic probes deliver sensitivity, molecular specificity, and compatibility with advanced microscopy, allowing ions, metabolites, and reactive species to be visualized in real time with subcellular resolution. Yet the very features that make probes bright and targetable, including high-affinity binding, lipophilic partitioning, organelle trapping, and excited-state reactivity, constitute measurement back-action. Label-free imaging, which draws contrast from intrinsic properties such as autofluorescence, refractive index, and molecular vibrations, emerged partly as a response to this paradox, but it carries its own burden of spectral congestion and mechanistic ambiguity.

Endogenous fluorophores occupy a conceptually privileged position in this debate. Aromatic amino acids such as tryptophan and tyrosine absorb in the ultraviolet and emit fluorescence whose properties are modulated by protein folding, binding interactions, and local polarity. Metabolites of the kynurenine pathway extend this landscape into the near-visible UV-A range, while the redox cofactors NAD(P)H and flavins provide spatially resolved metabolic readouts whose optical redox ratio and fluorescence lifetime report on cellular state. Because these molecules are constitutive components of biochemistry, their optical properties are a consequence of function rather than a precondition for it, placing native fluorescence at a lower bound on perturbation against which every exogenous labeling strategy must be benchmarked.

That benchmark is not a gold standard. NAD(P)H autofluorescence cannot distinguish NADH from NADPH contributions, and fluorescence lifetimes shift with pH, binding partners, and local dielectric properties, so the same sensitivity that encodes microenvironmental information also produces non-specificity. Kynurenine pathway activity varies with inflammation, immune activation, aging, and disease, meaning changes in fluorescence may reflect metabolite concentrations rather than the biological process under investigation. The authors conclude that label-free imaging is not a replacement for probes but a diagnostic, defining a lower bound on perturbation and an upper bound on interpretability challenges.

Nowhere is probe back-action more transparent than in calcium imaging. Classical indicators such as quin-2, fura-2, and indo-1 are derivatives of high-affinity chelator chemistry that can reach millimolar intracellular concentrations, directly competing with every endogenous calcium-binding protein for the same pool of free ions. To sense calcium with sufficient signal-to-noise, the indicator must bind it; to bind it competitively, it must buffer it. There is no regime in which a high-affinity chelator reports calcium without reshaping the calcium landscape it reports on, making the indicator both the ruler and a force applied to the object being measured.

Experiments in Xenopus oocytes demonstrated that buffer kinetics, not merely affinity, reshape signaling patterns. The slow buffer EGTA accelerated calcium signals and fragmented propagating waves into localized puff-like events, while the fast buffer BAPTA slowed responses and promoted spatially uniform signals. These differential effects arose through modulation of calcium-induced calcium release feedback rather than generic calcium clearance, and signals evoked by influx through voltage-gated channels were substantially less affected. The implication is profound: the probe actively participates in the decision-making machinery of calcium signaling, and spatial microdomains encoding biological decisions can be reclassified by the presence of the indicator itself.

Organelle-targeted dyes present a different but equally fundamental problem. Lipophilic cations accumulate in mitochondria because the negative inner membrane potential pulls them in, yet at imaging concentrations those same cations can depolarize the membrane, impair electron transport, and stimulate uncoupled respiration. Weak-base amines accumulate in lysosomes because acidic compartments protonate and trap them, but in doing so they titrate the very proton gradient that defines lysosomal identity, raising intralysosomal pH and altering organelle morphology and enzymatic activity. The concept of dark toxicity extends this further: penetrating cations can facilitate proton transport, react with respiratory-chain redox centers, and inhibit metabolic enzymes entirely in the absence of light, meaning their biological footprint precedes any imaging session. The authors propose that the boundary between imaging agent and pharmacological effector should be defined empirically, through organelle morphology, functional parameters, and viability assessed over the full experimental duration, not by chemical structure alone.

Phototoxicity, meanwhile, is not merely a dosimetry problem but a mechanistically predictable consequence of excited-state chemistry. Every fluorophore that undergoes intersystem crossing to a triplet state is transiently a photosensitizer, and every triplet state that encounters molecular oxygen is a potential source of singlet oxygen and downstream reactive oxygen species. Quantitatively, singlet oxygen quantum yields span orders of magnitude, from less than 0.01 for silicon-rhodamine dyes in their cell-permeable form to roughly 0.76 for rhodamine B in organic solvents, and halogenation can push BODIPY derivatives above 0.5, the regime of photodynamic therapy agents. Molecular rotor BODIPYs engineered for intramolecular motion suppress triplet formation and reduce phototoxicity three- to ten-fold, offering a template for next-generation probes. Yet the boundary between safe intrinsic contrast and reactive exogenous labeling is one of degree rather than kind: endogenous kynurenine metabolites are bona fide photosensitizers, and in the aging eye lens they can shift from photoprotective UV filters to mediators of protein crosslinking and cataract formation.

Middle-ground strategies redistribute rather than resolve the paradox. Chemigenetic tags such as HaloTag and SNAP-tag replace thermodynamic partitioning with covalent, genetically encoded specificity, and fluorogenic polymethine dyes deliver turn-on responses exceeding one hundred-fold with no-wash imaging. But self-labeling tags are large fusion proteins that can alter target localization, activity, and interactions, translocating the observer effect from dye chemistry to fusion biology. Bioorthogonal fluorogenic chemistry, including tetrazine-trans-cyclooctene cycloadditions, reduces the concentration of unbound non-specifically localized dye, but metabolic handle incorporation competes with endogenous biosynthetic pathways and can trigger stress responses. Stimulated Raman scattering microscopy pushes further toward minimal perturbation by detecting isotope-enriched amino acids and alkynyl vibrational tags that preserve native biochemical behavior, though it remains constrained by instrumental complexity, moderate sensitivity, and the photothermal effects of high-intensity near-infrared irradiation.

On the quantitative side, ratiometric imaging and fluorescence lifetime imaging microscopy each correct specific confounds while introducing new ones. FLIM is largely independent of fluorophore concentration and can separate free from protein-bound NADH and FAD, but it replaces concentration artifacts with photon-statistics artifacts, instrument response distortions, and pile-up effects that can bias lifetime estimates in ways mimicking genuine biological signals. The authors propose six design principles for zero-perturbation probes: tune affinity to the minimum required for signal fidelity, minimize the intrinsic biological activity of targeting motifs, engineer away from triplet-state formation, exploit fluorogenicity to lower effective probe dose, favor covalent over non-covalent localization, and co-optimize photophysics and biological activity in the same synthetic iteration. Their central conclusion is that measurement is never separable from intervention, and that the next frontier in biological imaging is not optical but epistemological, demanding perturbation assessment as a first-class experimental variable so that imaging data are interpreted not as a picture of the cell, but as a picture of the cell as perturbed by the act of looking at it.

Subject of Research: Perturbations caused by small-molecule fluorescent probes in live-cell imaging and label-free alternatives

Article Title: To label or not to label? The paradox of small-molecule probes in biological characterization

Article References: Cuevas-Galindo, M. E., & Jiménez-Sánchez, A. (2026). To label or not to label? The paradox of small-molecule probes in biological characterization. Frontiers in Photonics, 7, Article 1858154. https://doi.org/10.3389/fphot.2026.1858154

Image Credits: AI Generated

DOI: 10.3389/fphot.2026.1858154

Keywords: fluorescent probes, observer effect, calcium imaging, autofluorescence, FLIM, phototoxicity, reactive oxygen species, mitochondria, lysosomes, bioorthogonal chemistry, HaloTag, stimulated Raman scattering