**5. Analysis of NADH Autofluorescence by FLIM**

It has been known for several decades that NADH emits autofluorescence and, in contrast, NAD+ does not [26]. It is important to notice that, as the spectral properties of NADH fully overlaps with those of NADPH, it is common to measure the fluorescent contribution of both components and denominate them as NAD(P)H. Conversely, reduced flavin adenine dinucleotide (FADH2) does not produce autofluorescence compared to its oxidized version (FAD) [130]. This inverse relationship has been used to measure a "redox ratio" defined as the total fluorescence intensity of FAD divided by the total fluorescence intensity of NADH [131]. As such, under relatively constant FAD, lower levels of NAD(P)H may indicate a larger redox ratio and may correlate with a more oxidative cellular environment.

Complementing the classical intensity-based fluorescence methods, the time-resolved decay of fluorescence by FLIM provides unique information about the environment of fluorophores, including changes in pH, viscosity, or binding state to enzymes [132–135]. Importantly, at least two configurations and fluorescence lifetimes of NADH can be distinguished with this approach, namely free NADH and protein-bound NADH [32]. This is possible because the fluorescence decay of NADH in solution markedly differs when binding to different proteins, i.e., enzymes. As such, when NADH is in solution (free NADH) it exists in a folded configuration, which causes quenching of the reduced nicotinamide by the adenine group and shortening of its fluorescent lifetime. On the contrary, protein-bound NADH has an extended configuration, favoring a prolonged decay of its fluorescence. As such, the reported lifetime of free NADH in solution is significantly lower (~0.4 ns) than the protein-bound conformation (the lifetime of NADH bound to LDH is 3.4 ns) [136]. Furthermore, taking advantage of their binding to different metabolic enzymes, it has been possible to measure the particular contribution of NADH and NADPH separately by FLIM [99,137]. This may constitute a great diagnostic tool to monitor oxidative stress as NADPH is an element directly involved in redox management.

Different methods can be used to calculate the fluorescence lifetime. For this purpose, data can be fitted into a single-exponential or multi-exponential decay function where the exponential factor tau (τ) corresponds to the fluorescence lifetime of the fluorophore. Nevertheless, it is often not possible to determine the best method to fit the data a priori. As a way to circumvent this limitation, data can be also analyzed by phasor approach. Phasor analysis is a fit-free technique in which the fluorescence decay from each pixel is transformed into a point in a two-dimensional (2-D) phasor space. As such, it works on the unbiased raw data without any approximation, and it does not require a priori knowledge

of the sample being imaged, giving instantaneous results. Importantly, FLIM is compatible with confocal or multiphoton laser scanning microscopy as well as wide-field illumination. To obtain more details from each methodology, readers may refer to the following excellent publications [138–141].
