connect LHCII to the core, whereas the struc-
tural studies now imply that this is only the
case for CP29 in C 2 S 2 M 2 of plants, which acts
as the main“bridge”between M-LHCII and
the core. Analyses of plant mutants lacking
individual antenna complexes show that, al-
though the absence of CP26 has a negligible
effect on the transfer of energy in PSII, the im-
pact of the absence of CP29, and especially of
CP24, is very large, leading to functional discon-
nection of the antenna from the core ( 12 , 78 ).
Taken together, the structural and functional
results suggest that some of the minor antennas
perform a structural role in maintaining the or-
ganization of the supercomplex. However, they
play a limited role in light harvesting in the
sense that they contain relatively few pigments
and do not connect LHCII to the core, with the
exceptionofCP29inplantsand,toafarlesser
extent, inC. reinhardtii. The view that the
minor complexes have a structural role rather
than an important direct role in light har-
vesting is compatible with the situation in
C. reinhardtii,inwhichCP24isabsentand
CP29 therefore appears to be less strongly
anchored in the supercomplex. Indeed, in this
green alga, CP29 can migrate to PSI together
with LHCII to rebalance the excitations over
the two photosystems in the light conditions
that are favorable for PSII, a process known as
state transitions ( 79 ).
It is important to note that the organization
of EET along parallel (only weakly intercon-
nected) pathways is not favorable for the
robustness and speed of transfer toward the
RC, for which interconnected pathways are
in principle better. The presence of alternative
transfer pathways enhances the robustness
and, because two-dimensional excitation dif-
fusion is faster than one-dimensional diffusion
( 76 ), interacting EET pathways would lead to
faster excitation trapping in the RC and thus
increased quantum efficiency. Because of the
presence of only weakly interconnected path-ways, it seems more likely that evolution has
favored the current design for regulatory rea-
sons, which will still have to be unveiled. In
principle, it might also be possible that in the
stacked thylakoid membranes of, for example,
plant chloroplasts, EET between different mem-
brane layers would occur, thereby invalidating
the above reasoning, but a comparison be-
tween time-resolved fluorescence measure-
ments on stacked and unstacked grana has
shownthatalmostnointermembranetransfer
takes place ( 80 ).PSI versus PSII
When we compare the plant supercomplexes
of PSI and PSII, it appears that the different
trapping kinetics in the two photosystems can
to a large extent be traced back to differences
in their core complexes. As stated above, for
the PSI core, the trapping time is ~20 ps,
whereas for the PSII core, average trapping
times up to 100 ps and above have beenCroceet al.,Science 369 , eaay2058 (2020) 21 August 2020 6of9
Fig. 5. Structures of PSII from various organisms.[For original data, please see ( 28 ), ( 29 ), ( 35 ), ( 39 ), ( 40 ), and ( 110 ).] For most complexes, the number of Chls
and the average time of trapping of excitations in the reaction center are provided ( 66 , 73 , 111 ). Chlais shown green, Chlsbin blue, and carotenoids in orange.
RESEARCH | REVIEW