reported, but the number of pigments is almost
a factor of three lower. The reason for this dif-
ference seems to be the number of pigments
connecting the antenna and the RC, which is
lower for PSII. They are also at larger distances
with respect to each other, presumably to avoid
oxidation of the antenna pigments by the pri-
mary donor with its high oxidation potential
( 81 ). An immediate consequence of the high
efficiency of the PSI core is that it can accommo-
date substantially more LHCs around it. In the
PSI-LHCI supercomplex ofC. reinhardtii,with
its ~240 Chls, the reported trapping time is
50 ps and the quantum efficiency is still higher
than that of the PSII core, with its 35 pigments.
The recently observed structures also suggest
that there is a far larger variability in size,
composition, and organization in PSI than in
PSII. In PSI, the antennas can be associated
to the core in many different positions and
in several layers completely surrounding it.
Photosystems in the membrane
All of the structures discussed above were ob-
tained for detergent-isolated complexes. Are
these complexes representative of the in vivo
situation, where they may be larger or in a
different functional state? It is known that the
number of LHCII trimers in the membrane
changes with the growth conditions of the
plant or alga, increasing in low light and
decreasing in high light, and can be far larger
than in the purified supercomplexes ( 82 – 84 ).
Although in plants, the C 2 S 2 M 2 complex is
the prominent form in low light and C 2 S 2 is
relatively more abundant in high light, the
largest adjustments concern a pool of LHCII
(called L), which are loosely bound to the
supercomplexes and are lost during purifi-
cation ( 84 ). At present, only a low-resolution
projection map of spinach PSII complexes
containing one L-trimer is available ( 85 ). Func-
tionally, the L-trimers are well connected to
the PSII core, providing excitation energy ef-
ficiently, albeit more slowly, than the S- and
M-trimers ( 84 ). A similar picture is now emerg-
ing for PSI, suggesting that in plants, more than
one LHCII trimer can be associated with each
PSI ( 59 , 86 – 88 ).
Experimental data clearly show that the
antenna of PSI and PSII can be enlarged with
extra L-LHCIIs. But by how much? Is there a
maximal antenna size for the photosystems?
When the size of the antenna becomes larger,
more photons are absorbed and the light-
harvesting capacity goes up, but the trapping
time also increases, thereby lowering the quan-
tum yield. These counteracting effects lead to
an optimal (maximal) antenna size in low light
of 200 to 250 Chlamolecules per PSII RC in
A. thaliana( 84 ). This number is far higher for
PSI, as already discussed above, which seems
to be largely due to the extremely short trapping
time in the PSI core. It can thus be expected
that PSI complexes with far larger antenna size
can be found in organisms living in shaded
environments. Is there also a minimal antenna
size? InA. thaliana, even in very high light,
there are still almost two LHCII trimers present
per PSII ( 84 ), although this is not needed to
have sufficient light-harvesting capacity. How-
ever, mutants without outer antennas are more
sensitive to photodamage ( 89 ), suggesting that
a minimal antenna is needed for photoprotec-tion. At first sight, this may seem counter-
intuitive, but it should be realized that the
functional antenna size can also be regulated
by changing the excited-state lifetime ( 33 ). This
is particularly important for sudden changes in
light intensity for which protein synthesis and
degradation is not an option. As a short-term
response, the antenna switches between a light-
harvesting state, characterized by a long excited-
state lifetime, and a quenched state, in which
the excited-state lifetime is strongly reduced
and the functional antenna size is decreased
( 90 ). In plants, this mechanism, called non-
photochemical quenching, requires the pres-
ence of the protein PsbS ( 91 ), the location and
action mechanism of which are still enigmatic.
It has been proposed that PsbS binds in the
cleft between CP24 and the core in the C 2 S 2 M 2
supercomplex ( 28 ), but this proposal still needs
validation. In green algae, the main protein re-
sponsible for nonphotochemical quenching is
instead LHCSR3 ( 92 ), which inC. reinhardtii
was suggested to be located at the periphery
of the C 2 S 2 complex in direct contact with the
outer antenna ( 93 ).Outlook
The high-resolution structures of the super-
complexes revealed recently represent an excel-
lent starting point for studying energy flow
in detail through advanced modeling. This
will contribute to the basic understanding of
light harvesting in oxygenic photosynthesis,
but will also furnish the basis for a rational
redesign of the photosynthetic apparatus that
should lead to an increase in crop productivity
[e.g., regulated reduction of the antenna size
and extension of the photosynthetic-active-
radiation regime ( 3 )]. However, it is important
to realize that the correct assignment of the
Chl identities species is crucial to identifying
EET pathways from the structures. Moreover,
whereas complexes such as plant PSI have been
studied in detail using spectroscopy, little is
known about the functional behavior of most
of the algal supercomplexes. Therefore, spectro-
scopic measurements on these complexes are
now required to relate structure to functionality.
It is expected that cryo-EM will continue
to provide new structures, e.g., of the much
wanted PSII-PsbS complex, and will also permit
exploration of the flexibility of photosynthesis
by studying complexes of different organisms,
especially the ones living in extreme environ-
ments. The structures of PSI from cyanobac-
teria adapted to far-red light ( 94 , 95 )ortoiron
deficiency ( 96 ) are good examples of this.
The most important next step, however, is
a molecular understanding of light harvesting
and regulation in physiologically relevant con-
ditions. This requires studying the complexes
in their natural environment: the thylakoids.
How are the complexes assembled in the
membrane? How are they organized? How doCroceet al.,Science 369 , eaay2058 (2020) 21 August 2020 7of9
Zea mays Pisum sativumEET between subcomplexes
15 ps
30 ps
75 psPSI-LHCI-LHCII PSII-C 2 S 2 M 2 PSII-C 2 S 2 M 2 N 2
C. reinhardtiiM-LM-LM-LM-LM-LM-LM-LM-LM-LM-LM-LM-MM-LMM-L-LHCHCIHCIHCHCIHCHCIHCIHCIHCIHCIHCICCCCCIIIIIIIIIIIIIII S-LS-LSSSS-S-LSSSSS-LSS-LSSLHCIHCHCHHCHHHCHCIHCIHCHHHHCIIII
CCCCP2CP2CP2CP2CP2CP2CPCPCPCP2CP2CP2CCCCCCCP2CP2CCP2CCCC 222266666666666666666666CP2CCP2CP2CP2CP2CCCCCP2CCCCCCP2CCCCC 44444444444
PSPSIPSPSPSIPPSPSIPSIPSPSIPSIPSPSPSPSSISISSISISSIIIIIIIIIIIIIIIccccccooooooorrrrrreeeeeeeeeeeeeeeeCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCP2P2P2P2P2P2P2PPPPPPPPPPPPPP2PP2P2P2P2PPP2P2PP2P2P2P2P2PP2P2P2P2P2P2P2PP2P2P2P2PP2PPPPPPPPPPP 2222222222299999999999999999999999M-LHCII S-LHCII
CP26CP24
PSII coreL-LHLHCICI IPSIPSI corecoreL-LHCIILhca2PSI coreLhca320 ps100
psM-LHCIIM-LHCII S-LHCIIS-LHCII
CP26CP26CP29CP29PSII corePSII corePSII corePSII coreN-LHCIIN-LHCIIM-LHCII S-LHCII
CP26CP29PSII corePSII corePSIPSPPSIPSIPSIPSIPPPSIIcIcI cIcIIccoreoreoreoreoreoreooreoreoreorerrererererereeeeee PSII coreCP29
N-LHCIILhca1
Lhca4100
ps
100
ps100
psFig. 6. Major EET pathways in different supercomplexes.The colors of the arrows represent approximate
effective transfer times from one complex to another, as indicated in the inset. The semicircular red
arrow indicates the main pathway for EET from N-LHCII through M-LHCII and S-LHCII to the core in the upper
half of the supercomplex and replaces three orange arrows between the individual trimers and from
S-LHCII to the core. The most interesting features of these structures is that there are many“parallel”EET
pathways from the outer antenna to the core that are relatively weakly connected to each other, and that
in most cases, the minor complexes do not play a major role in connecting LHCII to the core.
RESEARCH | REVIEW