FaRLiP involves remodeling of PSI complexes to form trimers (Fig. 1-3) and large-scale proteome alterations to adjust levels of light-absorbing components, reaction center and electron transfer (Fig. 4 ). Focus on changes at the whole cell level and investigate any heterogeneity in the cellular distribution of Chl a and Chl f, as well as FRL phycobilisomes, live cells cultured in WL and acclimated to FRL C. thermalis 7203 were immobilized on agarose gel pads and imaged under near-physiological conditions. Samples were characterized by spectral and fluorescence lifetime imaging in a home-built FLIM instrument optimized for the far-red and near-infrared ranges of the spectrum. False-color fluorescence images of WL (Fig. 5A) and FRL (Fig. 5B) samples, acquired in epifluorescence mode and illuminated by the light-emitting diode (LED) source at 470 nm, show the distribution of cells on the surface of the sample. Figure 5C shows the emission spectrum of a single WL cell, with an emission maximum at 680 nm, while two FRL cells (Fig. 5C) have an additional emission peak, of variable amplitude with respect to on emission at 680 nm, at 735 nm. By switching to scanning confocal mode and using the 485 nm pulsed laser as the excitation light source, we were able to obtain spectral maps of the cells, in which a complete emission spectrum is recorded for each pixel of the picture. Depending on cell size, we collect 300-350 data points, in this case emission spectra, per cell, sufficient to map the distribution of emitters, with additional specificity conferred by selective excitation and tightly filtered emission ± 6 nm around the central wavelength selected by the monochromator. Spatial distributions of emission intensities for cells cultured under WL and FRL are shown in Fig. 5 (D and E, respectively), for emission intensities at 680 nm (left) and 735 nm (right), corresponding to Chl a and Chl f, respectively. The spectral imaging in Figure 5 (D and E) mirrors the emission spectra obtained for single cells in Figure 5C. FRL cells acquire an additional Chl f emission band at 735 nm (Fig. 5E right) that emits fluorescence at this wavelength appears to be more evenly distributed around the cell than the more peripheral emission at 680 nm (Fig. 5, D and E, left). The emission signal at 735 nm will contain a contribution from FRL phycobiliproteins, which fluoresce at this wavelength in Leptolyngbya JSC-1 and Synechococcus 7335 (59). FRL phycobiliproteins potentially represent the dominant source of emission at 735 nm in our spectral imaging, but the 485 nm excitation wavelength used strongly favors Chl a and Chl f over FRL-specific phycobiliproteins. The emission spectra of two individual FRL cells (designated as cells 2 and 3 in Fig. 5B) have two maxima, at 680 and 735 nm, but the relative intensity of the peak differs between the cells, highlighting the variability of FRL acclimation at the single cell level. The images in Fig. 5 (D and E) also reflect this cell-to-cell variability.
Simultaneously with spectral imaging, fluorescence lifetime maps of immobilized cells were recorded. The photon fluence for all lifetime measurements was about 1.0×1014 photon pulse-1 cm– 2, which is low enough to minimize excitonic annihilation in the antenna. As with spectral imaging, we collected 300–350 data points per cell; transient fluorescence decays were limited by the instrumental response of the system, which is approximately 130 ps. A double-exponent amplitude-averaged lifetime image of WL cells, recorded at 680 nm, is shown in Figure 5F; the distribution of average lifetimes in amplitude on the image displayed in the histogram in Figure 5I shows lifetime values in the range of 800 to 1600 ps with a mean and standard deviation of 1315 ± 121 ps . Longer lifetimes (orange/red) tend to be found at the cell periphery in WL cells, following the same distribution for emission at 680 nm as in Figure 5D (left). Lifetime images of FRL cells recorded at emission wavelengths of 680 and 735 nm with pulsed laser excitation at 485 nm (Fig. 5, G and H) show markedly different lifespans, with a mean and standard deviation of 1090 ± 93 ps and 964 ± 55 ps, respectively (Fig. 5, J and K), where there is almost no amplitude in WL cells. Thus, acclimation to FRL involves not only the assembly of an antenna containing Chl a and Chl f but also the FRL phycobiliproteins (see above) which tend to be coupled to a population of faster traps than those found in WL. Longer lifetimes associated with 680 nm emission (Fig. 5G) tend to be in more peripheral regions of WL and FRL cells (Fig. 5, F and G), while somewhat shorter lives follow the Chl f emission pattern in Fig. 5E and have a more uniform distribution (Fig. 5H).
Finally, negatively stained thin sections of WL (Fig. 5, L and M) and FRL (Fig. 5, N and O) cells show that the FaRLiP response involves tighter packing of the thylakoid membranes probably due to phycobiliprotein complexes more compact which are synthesized by FRL cells. Large membrane-extrinsic hemidiscoid phycobilisomes synthesized under WL normally impose separations of approximately 100 nm between thylakoids, as shown in Figure 5M. Although the structure of phycobiliprotein complexes assembled during FRL acclimatization in C. thermalis 7203 have not yet been determined, they are likely smaller than those synthesized under WL conditions, which allow for closer association of adjacent thylakoid membranes (25, 59–62), consistent with the images in Fig. 5 (N and W).