Exploiting nonresonant background signal of Multiplex-CARS as an intrinsic contrast mechanism for mapping muscle proteins with M-CARS spectroscopy

The discrimination between protein families within their native biological environment without the use of labeling strategies remains a major challenge in biological imaging. In this study, we introduce an original approach that takes the advantage of the non-resonant background (NRB) signal in Multiplex Coherent Anti-Stokes Raman Scattering (M-CARS) spectroscopy. Such a signal is usually treated as a perturbative signal. By exploiting the contrast between resonant and non-resonant components in hyperspectral M-CARS datasets, our method enables the label-free identification and spatial mapping of distinct protein families. As a proof of principle, we successfully discriminate actin and myosin filaments in muscle tissues. Myosin localization is determined through its intrinsic second harmonic generation (SHG) signal, while pixel-wise analysis of resonant to non-resonant ratios reveals the distribution of actin filaments, consistent with SHGbased expectations. This work demonstrates that the often-overlooked silent spectral region contains valuable information for protein discrimination, and underlines the potential of nonlinear optical contrast to expand the analytical capabilities of M-CARS for complex biological systems.