Quantum Dots Assembled with Photosynthetic Antennae on a Carbon Nanotube Platform: A Nanohybrid for the Enhancement of Light Energy Harvesting

The construction of artificial systems for solar energy harvesting is still a challenge. There needs to be a light-harvesting antenna with a broad absorption spectrum and then the possibility to transfer harvested energy to the reaction center, converting photons into a storable form of energy. Bioinspired and bioderivative elements may help in achieving this aim. Here, we present an option for light harvesting: a nanobiohybrid of colloidal, semiconductor quantum dots (QDs) and natural photosynthetic antennae assembled on the surface of a carbon nanotube. For that, we used QDs of cadmium telluride and cyanobacterial phycobilisome rods (PBSr) or light-harvesting complex II (LHCII) of higher plants. For this nanobiohybrid, we confirmed composition and organization using infrared spectroscopy, X-ray photoelectron spectroscopy, and high-resolution confocal microscopy. Then, we proved that within such an assembly, there is a resonance energy transfer from QD to PBSr or LHCII. When such a nanobiohybrid was further combined with thylakoids, the energy was transferred to photosynthetic reaction centers and efficiently powered the photosystem I reaction center. The presented construct is proof of a general concept, combining interacting elements on a platform of a nanotube, allowing further variation within assembled elements.

where I(t) is the number of photon counts in the time t, I0 is the tail offset (background), Ai is the amplitude of the i-th component associated with the τi lifetime, and t0 is the time offset.Then, the lifetime values obtained from the fit were used to fit the images -the photon distribution for each pixel was fitted to the equation above with the lifetime values fixed.As the result, four amplitude images, each associated with a different lifetime component, were calculated.
The next steps of the analysis were performed using a self-written Python script.Four calculated lifetimes and the amplitude values corresponding to each image pixel were used to calculate the amplitude-weighted mean lifetime (τav) image, according to the equation: The set of two-pixel matrices -the emission intensity image (recorded by the confocal system) and the calculated mean lifetime image -were used to extract the ROI areas.Using the threshold corresponding to 50-70% of the maximum intensity value (the brightest pixel of the intensity image; the threshold was set depending on the background noise level), the mask was created to remove the pixels corresponding to the low-intensity background.The remained image consisted of irregular patches which were detected by the contour-finding algorithm (imported from the scikit-image library [1]) and sorted according to the surface area (in pixels).50-100 patches of the largest area were selected as individual ROIs in each FLIM image.
The spatial overlaps for each ROI were calculated as the fraction: overlap = number of ROI pixels overlapping with ROIs from the other channel ROI area [pixels] The overlap with the pairs of the channels (e.g.QD530+PBS) was calculated using the pixels shared by both channels (the intersection of both sets of pixels).Absorption and emission spectra normalized to 1 at maximum.Fluorescence was excited at 405 nm.
For the titration curves, emission at 630 nm was followed.For the explanation of "transmittance effect" see Figure S2 and main manuscript.In comparison, thylakoids illuminated with optimal light range (bandpass filter 650-670 nm).The samples contained the number of thylakoids corresponding to 20 μg/ml chlorophyll and ~ 400 μg/ml nanohybrid in the presence of a redox system (1 mM sodium ascorbate + 0.1 mM DCPIP + 0.5 mM methyl viologen).

Figure S1 .
Figure S1.Transmittance changes in the dependence on the CNT concentration.(A) Original transmittance scan recorded for given CNT suspensions and (B) change in the transmittance at 650 nm as the function of CNT concentration.

Figure
Figure S2.Absorption (A) and emission (B) spectra of Zn-mesoprohyrine, free (red) or bound to HP7 (green) (1:2, protein:ligand ratio) as well as the titration of HP7-ZnMP with MWCNT at different temperatures (C).For titration with MWCNTs, a protein concentration was 2 μM in a final 1 ml of 25 mM Hepes/NaOH pH 7.5, MWCNTs were added in 1-2 μM aliquots from concentrated stocks solution.Absorption and emission spectra normalized to 1 at maximum.Fluorescence was excited at 405 nm.For the titration curves, emission at 630 nm was followed.For the explanation of "transmittance effect" see FigureS2and main manuscript.

Figure
Figure S3.FT-IR analysis of the CNT conjugation process with LHCII.(A) OH and CH vibration range and (B) amide vibration range, recorded for oxidized version of MWCNTs decorated with BSA, BSA-LHCII and BSA-LHCII-QD570.LHCII and QD570 spectra are shown for reference.No normalization was applied.

Figure S4 .
Figure S4.The background for FRET reaction in the studied nanohybrids: emission spectra of donors (QD530 or QD570) and absorption spectra of acceptors (LHCII and PBS).

Figure
Figure S5.(on the previous page) The representative example of the FLIM analysis for MWCNT-QD530-PBSr (~300 μg/ml) and thylakoid mixture (100 μg/ml of chlorophyll).(A, B, C) FLIM images of QD530, PBSr, and thylakoid, respectively.Averaged lifetimes [ns] for selected ROIs are indicated.The excitation wavelength was 500 nm, and the emission ranges for each channel listed.(D) The overlap image of ROIs extracted from FLIM images.The contours of the individual ROIs are colored according to their lifetime.(E, F, G) The dependence of the QD530, PBS and chlorophyll lifetime on the spatial overlap with the other components of the FRET system.The points represent individual ROIs, colored according to the lifetime.The dotted line shows the averages from the points in the overlap value ranges spanning 10 percentage points each.

Figure S6 .
Figure S6.The representative example of the FLIM analysis for MWCNT-QD530-PBSr hybrid (~300 μg/ml).(A, B) FLIM images of QD530 and PBSr.Averaged lifetimes [ns] for selected ROIs are indicated.The excitation wavelength was 500 nm, and the emission ranges for each channel were listed.(C) The

Figure
Figure S7.CLSM images of MWCNT-QD570-LHCII hybrids and their mixture with Synechocystis PCC 6803 thylakoids.The excitation wavelength was 470 nm, and the emission ranges for each channel are indicated.

Figure S8 .
Figure S8.Fluorescence emission spectra recorded with CLSM for selected ROI on MWCNT-BSA-QD530 (A) or MWCNT-BSA-QD570-LHCII (B) preparation.Please note that points represent actual data of the summarized region pixel intensities and the bold line works as an eye-guide only.Emission was excited with 405 nm (A) or 470 nm (B) lase lines.Spectral resolution was set at 3 nm, maximum possible with the system.Arrows indicate position of emission maxima of QD, PBSr and LHCII.Note, the intensity of particular bands do not necessarily represent the amount of fluorophore, as the excitation wavelength may not be optimal for maximum fluorescence.

Figure S9 .
Figure S9.The representative example of the FLIM analysis for MWCNT-QD530 hybrib (~300 μg/ml) mixed with thylakoids.(A) FLIM images of a mixture with QD530 ROIs andalysed and (B), chlorophyll emission ROIs analysis for the same mixture.The excitation wavelength was 500 nm, and the emission ranges were 515-554 and 702-770, respectively.(C, D) The dependence of the QD530 and PBS lifetime on the spatial overlap with the other component of the FRET system.The points represent individual

Figure S10 .
Figure S10.Atomic force microscopy of BSA-coated (A, B) and fully functionalized (C, D) single-walled carbon nanotubes.Images show topography (A, C) and respective deflection image (B, D).

Table S2
. Atomic concentrations [%], estimated by XPS, in CNT and its nanohybrids.Determination has been made based on the survey spectra of analyzed samples using CasaXps software.