Size-Dependent Fluorescence Kinetics Reveal Contributions of Intrinsic Quenching and Singlet-Triplet Annihilation during LHCII Aggregation

Abstract

Aggregation of the main antenna complex of higher plants, Light-Harvesting Complex II (LHCII), is widely used as an in vitro model for energy-dependent quenching (qE), yet fluorescence reduction in aggregates is frequently interpreted without a quantitative separation of intrinsic quenching from excitation-induced annihilation. Here, we address this ambiguity by directly correlating aggregate size, concentration, steady-state fluorescence intensity, and decay kinetics during controlled, incremental aggregation of isolated LHCII. By combining fluorescence correlation spectroscopy (FCS) with TCSPC in a unified experimental framework, we monitored structural and photophysical changes in real time as detergent removal drives biphasic aggregation. We quantified the aggregate composition from the particle concentrations, enabling direct scaling of the absorption cross-section with aggregate size. The average fluorescence lifetime decreased semi-logarithmically with increases in hydrodynamic radius, whereas steady-state fluorescence intensities deviated strongly from this trend. Intensitydependent measurements and steady-state kinetic modeling reveal that singlet-triplet annihilation (STA) emerges at moderate excitation intensities and rapidly becomes the dominant contributor to fluorescence quenching, even for relatively small aggregates. In contrast, intrinsic quenching increases more gradually with aggregate size. By quantitatively disentangling intrinsic excitation quenching from annihilation processes, this work demonstrates that STA can govern the apparent photophysical response of aggregated LHCII across excitation regimes commonly considered non-annihilating. The size-dependent mechanistic framework presented here provides a basis for distinguishing intrinsic quenching from annihilation effects in aggregation-based studies of photosynthetic antenna complexes.

Figures (6)

• Figure 1: A. Fluorescence intensity traces of freely-diffusing LHCII recorded with SPAD 1 after increasing intervals of detergent removal. The colours indicate different samples extracted from the same solution incubated with Bio-Beads at 5-min intervals. B. Examples of fluorescence decays recorded in the same measurements using SPAD 1 (0.01 ns bin size). Biexponential fits are shown in black. C. Corresponding examples of raw autocorrelation curves of samples extracted, measured within 10 -- 50 minutes after extraction. $G(\tau)$-fits of the autocorrelation curves are shown in black; one or two 3D diffusion components were used. The inset shows an example of a $G(\tau)$-fit for very large LHCII aggregates. Fitting residuals are shown (bottom).
• Figure 2: A. Average fluorescence intensities measured with SPAD 1 over detergent removal time of three repetitions of the incremental aggregation experiment, corresponding with examples shown in Fig. \ref{['fig:LHCII_decay_and_autocorrelation_dynamics']}A. B. Corresponding average fluorescence lifetimes from biexponential fits of decays corresponding with Fig. \ref{['fig:LHCII_decay_and_autocorrelation_dynamics']}B. C. Average background-corrected particle numbers corresponding with example autocorrelation curves shown in Fig. \ref{['fig:LHCII_decay_and_autocorrelation_dynamics']}C. D. Corresponding hydrodynamic radii obtained from autocorrelation diffusion times. Data points from three independent repetitions of the experiment can be distinguished by shape, viz., circles, squares, and triangles. E. Fluorescence lifetime components resolved by biexponential fitting. F. Normalized amplitudes of the two lifetime components in D. averaged over the three repetitions of the experiment, with $A_1$ corresponding with $\tau_1$ and $A_2$ with $\tau_2$. G. Average fluorescence lifetimes compared with the average hydrodynamic radii. A semi-logarithmic linear regression (base $e$) with 95% confidence interval is shown, with a gradient of $-0.59 \pm 0.04$ and $r^2 = 0.9$.
• Figure 3: A. Fluorescence decays (0.01 ns bin sizes) of LHCII at 10-minute-long intervals of detergent removal time recorded with SPAD 2 ($\sim$38 ps IRF) at a lower excitation intensity (31 W$\cdot$cm$^{-2}$). B. Fluorescence decays recorded with SPAD 2 at a higher intensity (258 W$\cdot$cm$^{-2}$). C. Normalized amplitudes of the three resolved lifetime components at the higher excitation intensity, with corresponding average lifetime components indicated. D. Normalized average fluorescence intensities (left axis, blue) and average fluorescence lifetimes (right axis, red) of the same samples measured at 31 and 258 W$\cdot$cm$^{-2}$.
• Figure 4: A. Time-dependent increase in the average number of LHCII trimers per aggregate, $M$, calculated through the relative inverse concentration (Eq. \ref{['eq:aggregate_composition']}), for three repetitions of detergent removal with Bio-Beads. Each shape represents a different experimental run. I and II refer to the slow and fast aggregation phases, respectively. B. Hydrodynamic radius against $M$ for the three repetitions in A (data points) along with a double logarithmic linear regression (solid line) with a corresponding 95% confidence interval (shaded area). The gradient is 0.63 $\pm$ 0.04, with $r^2$ = 0.88, giving a fractal dimension of $d_f = 1.6 \pm 0.1$. The linear regression for the small aggregates ($0 < M < 10$, dashed line) gives $d_f = 2.3 \pm 0.4$. The dotted outlines indicate the approximate borders between Phases I and II of the aggregation process.
• Figure 5: A. Relative fluorescence yield of LHCII trimers (blue circles) at increasing excitation intensities on a logarithmic scale, effectively increasing $n_0$ from 0.05 to 5 for $M=1$, calculated theoretical fluorescence yield taking into account only SSA (dotted line, Eq. \ref{['eq:SSA_FL_yield']}), only STA (dashed line, Eq. \ref{['eq:STA_FL_yield']}), and both SSA and STA (solid line), and amplitude-averaged fluorescence lifetimes (red circles). B. Relative fluorescence intensity ($F/F_0$, blue symbols), equivalent to the relative effective yield of fluorescence after incubation with Bio-Beads started ($\Phi^{eff}/\Phi^{eff}_0$), and average two-component fluorescence lifetime ($\tau^{avg}$, red symbols), displayed against $M$. A linear regression fit of $\tau^{avg}$ (red line) approximates $\Phi/\Phi_0$ based on quenching, Q, giving a gradient of $-0.49 \pm$ 0.03 ns/ln(M) and an $r^2$-coefficient of 0.94. Calculated $\Phi^{eff}/\Phi^{eff}_0$ based on SSA (Eq. \ref{['eq:SSA_FL_yield']}) and Q (Eq. \ref{['eq:Q']} via the $\Phi/\Phi_0$ regression fit) (dotted line), based on STA (Eq. \ref{['eq:STA_FL_yield']}) and Q (dashed line), and Q, SSA, and STA combined (solid line).