858
Biochemical Society Transactions (2005) Volume 33, part 4
Role of electron-transfer quenching of chlorophyll
fluorescence by carotenoids in non-photochemical
quenching of green plants
A. Dreuw*1 , G.R. Fleming†‡ and M. Head-Gordon†‡
*Institute for Physical and Theoretical Chemistry, Johann Wolfgang Goethe-University Frankfurt, Marie Curie-Str. 11, 60439 Frankfurt am Main, Germany,
†Department of Chemistry, University of California, Berkeley, CA 94720-1460, U.S.A., and ‡Chemical Sciences and Physical Biosciences Division, Lawrence
Berkeley National Laboratory, Berkeley, CA 94720-1460, U.S.A.
Abstract
NPQ (non-photochemical quenching) is a fundamental photosynthetic mechanism by which plants protect
themselves against excess excitation energy and the resulting photodamage. A discussed molecular mechanism of the so-called feedback de-excitation component (qE) of NPQ involves the formation of a quenching
complex. Recently, we have studied the influence of formation of a zeaxanthin–chlorophyll complex on the
excited states of the pigments using high-level quantum chemical methodology. In the case of complex
formation, electron-transfer quenching of chlorophyll-excited states by carotenoids is a relevant quenching
mechanism. Furthermore, additionally occurring charge-transfer excited states can be exploited experimentally to prove the existence of the quenching complex during NPQ.
Introduction
Under normal light conditions, regular photosynthesis occurs in green plants, and the absorbed sunlight is almost exclusively used for photochemical charge separation in the
reaction centre. However, as soon as the excitation energy
present in the photosystem exceeds the capacity of the reaction centre, dangerous by-products, i.e. triplet states and
singlet oxygen accumulate and can lead to severe photodamage. One important process besides triplet quenching and
singlet oxygen scavenging, by which green plants protect
themselves against that damage, is generally termed NPQ
(non-photochemical quenching) [1–7]. NPQ can generally
be subdivided into three components: feedback de-excitation
(qE), photoinhibition (qI) and state transition (qT) [8]. In the
context of the present study, qE is of interest, since it is
the only component connected with changes induced by
high-light exposure on a time scale of a few minutes to seconds
[5]. In contrast with triplet quenching and singlet oxygen
scavenging, qE relies on the direct quenching of chlorophyll
(Chl)-excited states before the generation of harmful reactive
intermediates.
Today, it is well known that qE is regulated by the magnitude of the transmembrane pH gradient (pH) across the
thylakoid membrane [3,5]. When pH becomes larger than
that under regular light conditions, the enzyme violaxanthin
Key words: carotenoid, chlorophyll, electron transfer, excitation energy transfer, feedback deexcitation component, non-photochemical quenching.
Abbreviations used: Car, carotenoid; Chl, chlorophyll; CT, charge transfer; ET, electron transfer;
EET, excitation energy transfer; NPQ, non-photochemical quenching; TDDFT, time-dependent
density functional theory; Vio, violaxanthin; Zea, zeaxanthin.
1
To whom correspondence should be addressed (email [email protected]).
C 2005
Biochemical Society
(Vio) de-epoxidase is activated, and Vio is reversibly converted into zeaxanthin (Zea) (Figure 1) through the so-called
xanthophyll cycle [9]. Following this conversion of Vio, qE
manifests itself by a change in the absorption spectrum at
535 nm (A535 ) [9], which is in the region of the carotenoid
(Car) S0 → S2 absorption. Furthermore, a substantial decrease
of the Chl fluorescence lifetime from approx. 2.0 to approx.
0.3 ns is observed [10]. It is also well known that qE requires
the presence of a functional version of the PsbS protein [11],
which is a Photosystem II subunit with a molecular mass of
22 kDa [12–16]. The functional role of PsbS, however, is still
elusive.
In summary, qE represents a complicated biological process comprising many different aspects of the plant physiology. However, the key step of qE is the dissipation of excess
excitation energy. At present, neither a detailed molecular
mechanism of Chl fluorescence quenching during qE nor the
exact location of the quenching site has been established. In
the following paragraphs the state of research with respect to
these aspects will be briefly reviewed.
Early experiments have proved that the presence of Zea
is necessary for qE in chloroplasts [9]. However, its role has
been disputed. On one hand, it has been suggested that the
conversion from Vio to Zea has an indirect influence on
the structure and organization of the photosystem allowing
for efficient quenching [3,17–19]. On the other hand, a
direct influence of Zea as the quencher of Chl fluorescence
has been discussed [20,21]. Recent elaborate femtosecond
pump–probe experiments on intact spinach thylakoids have
shown that Zea may in fact directly be involved in the
quenching process [22], being probably the quencher of Chl
fluorescence. This is in line with the so-called ‘molecular gear
Mechanisms of Bioenergetic Membrane Proteins
Figure 1 Vio (top) possesses nine conjugated double bonds. De-epoxidation of Vio by the so-called xanthophyll cycle converts it into
Zea (bottom) exhibiting 11 conjugated double bonds
shift’ model [20], which relies on the fact that the conversion
of Vio to Zea leads to a substantial decrease of the S1 energy of
the Cars, which is related to the fact that Zea exhibits
11 conjugated double bonds in contrast with nine of Vio.
Although the S1 energy of Vio is assumed to lie above the
Qy state of Chl a and thus allows for EET (excitation energy
transfer) from Vio to Chl a, i.e. light harvesting, the S1 state of
Zea should be below the Qy state. This switches the direction
of the EET process and, thereby, makes quenching of Chl
fluorescence by Zea possible [20]. However, a considerable
uncertainty still exists about the actual position of the S1
energy levels of the Cars, and indeed, a recent experiment
actually suggests that the S1 states of Vio and Zea are both
energetically below the Qy state of Chl a [23], arguing against
such a model.
Besides its direct influence, Zea can still have an indirect
influence on qE meaning that conversion from Vio to Zea
can still lead to structural rearrangements necessary for efficient quenching of excess excitation energy. One discussed
model is the formation of a quenching complex consisting
of Zea and Chl a during the induction time of qE [10,24].
This would increase the coupling strengths between the
relevant excited states and may lead to additional dissipation
channels. Recent experimental and theoretical investigations
of linked carotenoporphyrins suggested that quenching of
electronically excited Qy states of porphyrins can in principle
also proceed by ET (electron transfer) quenching [25–27],
which corresponds to non-radiative decay into an appropriate
energetically low-lying CT (charge transfer) excited state.
Fungo et al. pointed out that ET quenching is possible in
linked carotenoporphyrins, when the Car exhibits at least
eight conjugated double bonds [25].
We performed high-level quantum chemical calculations
to illuminate the possible quenching pathways for excited
Qy states of Chl a by Vio and Zea depending on whether
a quenching complex of these pigments is formed or not
[28,29]. On the basis of our calculations, we predicted that
EET as well as ET quenching is possible if a quenching
complex is formed during qE. Moreover, we were able to
suggest a pump–probe experiment to corroborate the existence of such a complex. The experiment has recently been
performed and confirmed our theoretical predictions [30].
Computational methods
Before the study of Car–Chl complexes, the structures of Vio,
Zea and a Chl a model complex were separately optimized
at the level of density functional theory (DFT) using the hybrid B3LYP [31] exchange-correlation functional with the
3–21G basis set as implemented in the Q-Chem [32] package
of programs. This approach usually yields geometries with
errors of approx. 0.02 Å (1 Å = 10−10 m) in bond lengths and
1–5◦ in bond angles compared with experimental data. In
this case also, the agreement is very much favourable [28,29].
Since the side chains of Chl a have only a minor influence on
the excited states considered here, a model could be used in
which they are neglected.
For the calculation of the excited states of the composed
complexes, TDDFT (time-dependent density functional
theory) [33] has been used, which is well known to yield
excitation energies for π –π ∗ excited states with an accuracy
of 0.1–0.5 eV. In the present study, the Q band of Chl a and the
lowest states of the Cars have been calculated with the Tamm–
Dancoff approximation [34] to TDDFT using the BLYP [35]
functional. In all calculations the 3–21G basis set is used,
since larger basis sets are at present not applicable due to
the enormous molecular size of the considered complexes.
Calculations of the excited states of individual Zea and Chl
a have shown that in both cases, the excitation energies
are overestimated by approx. 0.2 eV compared with known
experimental values resulting in very good relative energies.
Although TDDFT methods yield reliable results for these
states, the excitation energies of CT states are usually underestimated by approx. 1 eV and CT potential energy curves
do not exhibit the correct 1/r asymptote [36,37]. Therefore,
a hybrid approach that combines Tamm–Dancoff approximation and CIS (configuration interaction singles) as suggested in [36,37] has been employed for the calculation of
the CT states to obtain self-interaction corrected potential
C 2005
Biochemical Society
859
860
Biochemical Society Transactions (2005) Volume 33, part 4
energy curves and excitation energies along an intermolecular
separation coordinate.
Results and discussion
Since qE is a dynamic process and neither the location nor
a possible structure of the discussed quenching complex is
known, the assumption of a reasonable relative arrangement
of the Cars and Chl a has been the starting point of our
study. We tested several arrangements [29], but here we focus
on a co-facial-middle arrangement, which turned out to be
the energetically most favourable one in the gas phase. From
this perspective, it is a good candidate to be formed in the
‘quenched’ state. Of course, within the protein environment
of Photosystem II, the possible geometrical arrangements are
strongly sterically restricted.
In this model structure, the Chl a model complex is placed
in the middle of the Cars such that the π -systems are parallel
to each other (see inset of Figure 2). The calculation of
the low-lying excited states of such Zea–Chl and Vio–Chl
complexes along a separation coordinate using the theoretical
methodology described above resulted in the potential energy
curves displayed in Figure 2. The separation coordinate is
simply the distance between the planes of the two π -systems.
The nature of the excited states was analysed by means of the
molecular orbitals corresponding to the electronic transitions.
In the case of the Zea–Chl complex (lower part of Figure 2),
the energy of the S1 state of Zea (blue curve) is at all distances
lower than the energy of the Chl Qy state (green curve)
allowing for EET at all distances. Going to shorter distances
a CT excited state (red curve) decreases rapidly in energy, and
at a Zea–Chl distance of 5.5 Å the CT state crosses the Qy
state. Then Chl fluorescence quenching through ET from Zea
to Chl becomes energetically accessible. At distances smaller
than 5.2 Å, the CT state is even lower in energy than the S1
state of Zea. The additional electrostatic attraction between
the Zea radical cation and the Chl radical anion in the CT
state is the reason why it drops in energy more rapidly than
the S2 , Qy and S1 states (Figure 2) and that the CT state finally
becomes the energetically lowest state.
Comparing with the analogous cofacial Vio–Chl complex
(upper part of Figure 2), the S1 energy of Vio is clearly
above the Qy state of the Chl along the complete distance
range, i.e. the quenching channel through EET is not
accessible. However, at short distances, smaller than approx.
4.8 Å, a CT state analogous to the one of the Zea–Chl
complex becomes lower in energy than the Qy state. As
a result, a quenching channel for the Vio–Chl complex
through ET quenching is in principle accessible at these short
distances.
Based on our calculation, a model for energy dissipation
can be deduced, which is schematically sketched in Figure 3.
When Vio is present in Photosystem II (left side of Figure 3), EET from Qy to S1 is not possible independent of
whether a quenching complex is formed or not. At short
distances, however, when a Vio–Chl complex is formed, ET
quenching is possible, which disagrees with experimental
C 2005
Biochemical Society
Figure 2 Potential energy surfaces of the ground (magenta) and
excited states (red, CT states; green, Chl a states; blue, Car states)
of the Vio–Chl a complex (upper panel) and of the Zea–Chl a
complex (lower panel) along the distance coordinate between the
Cars and the Chl
The structure of the model complex is shown in the inset of the lower
panel.
findings. The CT state decays back into the ground state
by charge recombination.
In the case of Zea, EET from Qy to S1 is possible at all
distances, although this is more efficient at short distances
than at large ones. Furthermore, if a quenching complex is
formed, the Qy state can also decay into the identified lowlying CT state, i.e. by ET quenching. Since in a complex the
CT state is also lower in energy than the S1 state of Zea,
the S1 state can also, in principle, decay through ET into the
respective CT state, which finally decays non-radiatively into
the ground state by thermal dissipation and back-transfer of
an electron. The roles of Zea in this scenario are both indirect and direct: on one hand structurally to form the complex and on the other hand electronically to either accept and
Mechanisms of Bioenergetic Membrane Proteins
Figure 3 Schematic sketch of the possible non-radiative decay
channels in a Car–Chl quenching complex when either Vio or Zea is
present
In the case of Vio (left side), excitation energy located on Chl can only
decay through ET quenching into a low-lying CT excited state, which decays back to the ground state by charge recombination (RC). If Zea is
present, the Qy excited state of Chl can be quenched in analogy to the
Vio case by ET with subsequent RC, but also by EET to the S1 state of Zea,
which finally decays non-radiatively by internal conversion (IC) back to
investigated. It can be expected that inclusion of protein
interactions will shift the CT states to lower energies due
to polarity of the protein and the huge dipole moment of
these states compared with the other π –π * states.
A.D. gratefully acknowledges financial support by the Deutsche
Forschungsgemeinschaft as an ‘Emmy-Noether’ Fellow.
the electronic ground state.
References
dissipate excess excitation energy through EET or to provide
the electron in the ET process. In summary, two competing
Chl fluorescence-quenching pathways, ET and EET, are in
principle accessible, if a Zea–Chl complex with an intermolecular distance of less than 5.5 Å is formed during qE.
The proposed molecular mechanism for Chl fluorescence quenching by Zea through EET and ET has been
found under the assumption that a Zea–Chl complex is
formed during the induction time of qE (see above). Although
we could not draw a final conclusion based on our calculations whether a quenching complex is formed during qE or
not, nevertheless, we could suggest a pump–probe experiment to obtain a conclusive answer to this question. Since
the low-lying CT state consisting of a Zea radical cation and
a Chl radical anion is only energetically accessible when a
complex is formed, it can serve as a spectroscopic signature of
a quenching complex in the quenched state of Photosystem II.
Thus we suggested to excite the Qy state of Chls a of Photosystem II while NPQ is active and to probe the characteristic electronic absorption of the Car radical cation.
Guided by our calculations and predictions, the suggested
pump–probe experiment has recently been performed.
Indeed, a strong transient absorption band has been identified
at 1000 nm and has been interpreted as Zea radical cation
absorption. This band is only present when qE is active, i.e.
Zea is present. This is an experimental proof for the existence
of a Zea–Chl quenching complex during qE and our proposed
ET quenching mechanism.
In future work, similar calculations will be performed for
a geometry of Vio–Chl and Zea–Chl complexes, which has
been identified in the recently published X-ray structure
of LHC-II [38,39] to illuminate further the mechanism of
qE. Also the influence of the protein environment, which
is neglected within the presented calculations, shall be
1 Demmig, B., Winter, K., Krüger, A. and Czygan, F.-C. (1987) Plant Physiol.
84, 218
2 Barber, J. and Andersson, B. (1992) Trends Biochem. Sci. 17, 61–66
3 Horton, P., Ruban, A.V. and Walters, R.G. (1996) Annu. Rev. Plant Physiol.
Plant Mol. Biol. 47, 665
4 Niyogi, K.K. (1999) Annu. Rev. Plant Physiol. Plant Mol. Biol. 50,
333–359
5 Müller, P., Li, X.-P. and Niyogi, K.K. (2001) Plant Physiol. 125,
1558–1566
6 Kühlheim, C., Ågren, S. and Jansson, S. (2002) Science 297, 91–93
7 Holt, N.E., Fleming, G.R. and Niyogi, K.K. (2004) Biochemistry 43,
8281–8289
8 Horton, P. and Hague, A. (1988) Biochim. Biophys. Acta 932, 107–115
9 Gilmore, A.M. (1997) Physiol. Plant 99, 197–209
10 Gilmore, A.M., Hazlett, T.L. and Govindjee (1995) Proc. Natl. Acad.
Sci. U.S.A. 92, 2273–2277
11 Li, X.-P., Björkman, O., Shih, C., Grossman, A.R., Rosenquist, M.,
Jannson, S. and Niyogi, K.K. (2000) Nature (London) 403, 391–395
12 Ljungberg, U., Akerlund, H.E. and Andersson, B. (1986) Eur. J. Biochem.
158, 477–482
13 Ghanotakis, D.F., Waggoner, C.M., Bowlby, N.R., Demetrious, D.M.,
Babcock, G.T. and Yocum, C.F. (1987) Photosynth. Res. 14, 191–199
14 Wedel, N., Klein, R., Ljungberg, U., Andersson, B. and Herrmann, R.G.
(1992) FEBS Lett. 314, 61–66
15 Kim, S., Sandusky, P., Bowlby, N.R., Aebersold, R., Green, B.R.,
Vlahakis, S., Yocum, C.F. and Pichersky, E. (1992) FEBS Lett. 314,
67–71
16 Li, X.-P., Müller-Moulé, P., Gilmore, A.M. and Niyogi, K.K. (2002)
Proc. Natl. Acad. Sci. U.S.A. 99, 15222–15227
17 Ruban, A.V., Young, A.J., Pascal, A.A. and Horton, P. (1994) Plant Physiol.
104, 227–234
18 Ruban, A.V., Young, A.J. and Horton, P. (1994) Biochim. Biophys. Acta
1186, 123–127
19 Gilmore, A.M. and Yamamoto, H.Y. (1992) Proc. Natl. Acad. Sci. U.S.A.
89, 1899–1903
20 Frank, H.A., Cua, A., Chynwat, V., Young, A.J., Gosztola, D. and
Wasielewski, M.R. (1994) Photosynth. Res. 41, 389–395
21 Young, A.J. and Frank, H.A. (1996) J. Photochem. Photobiol. B: Biol. 36,
3–15
22 Ma, Y.-Z., Holt, N.E., Li, X.-P., Niyogi, K.K. and Fleming, G.R. (2003)
Proc. Natl. Acad. Sci. U.S.A. 100, 4377–4382
23 Polivka, T., Herek, J.L., Zigmantas, D., Akerlund, H. and Sundström, V.
(1999) Proc. Natl. Acad. Sci. U.S.A. 96, 4914
24 Van Amerongen, H. and van Grondelle, R. (2001) J. Phys. Chem. B 105,
604–617
25 Fungo, F., Otero, L., Durantini, E., Thompson, W.J., Silber, J.J., Moore, T.A.,
Moore, A.L., Gust, D. and Sereno, L. (2003) Phys. Chem. Chem. Phys. 5,
469
26 Cardoso, S.L., Nicodem, D.E., Moore, T.A., Moore, A.L. and Gust, D. (1996)
J. Braz. Chem. Soc. 7, 19
27 Hermant, R.M., Lidell, P.A., Lin, S., Alden, R.G., Kang, H.K., Moore, A.L.,
Moore, T.A. and Gust, D. (1993) J. Am. Chem. Soc. 115, 2080–2081
28 Dreuw, A., Fleming, G.R. and Head-Gordon, M. (2003) J. Phys. Chem. B
107, 6500–6503
29 Dreuw, A., Fleming, G.R. and Head-Gordon, M. (2003) Phys. Chem.
Chem. Phys. 5, 3247–3256
30 Holt, N.E., Zigmantas, D., Valkunas, L., Li, X.-P., Niyogi, K.K. and
Fleming, G.R. (2005) Science 307, 433–436
31 Becke, A.D. (1993) J. Chem. Phys. 98, 5648
C 2005
Biochemical Society
861
862
Biochemical Society Transactions (2005) Volume 33, part 4
32 Kong, J., White, C.A., Krylov, A.I., Sherrill, D., Adamson, R.D., Furlani, T.R.,
Lee, M.S., Lee, A.M., Gwaltney, S.R., Adams, T.R. et al. (2000)
J. Comput. Chem. 21, 1532–1548
33 Casida, M.E. (1995) in Recent Advances in Density Functional Methods,
Part I (Chong, D.P., ed.), pp. 155–192, World Scientific, Singapore
34 Hirata, S. and Head-Gordon, M. (1999) Chem. Phys. Lett. 314, 291
35 Becke, A.D. (1988) Phys. Rev. A 38, 3098
36 Dreuw, A., Weisman, J.L. and Head-Gordon, M. (2003) J. Chem. Phys.
119, 2943–2946
C 2005
Biochemical Society
37 Dreuw, A. and Head-Gordon, M. (2004) J. Am. Chem. Soc. 126,
4007–4016
38 Liu, Z., Yan, H., Wang, K., Kuang, T., Zhang, J., Gui, L., An, X. and
Chang, W. (2004) Nature (London) 428, 287–292
39 Standfuss, J., Terwisscha van Scheltinga, A., Lamborghini, M. and
Kühlbrandt, W. (2005) EMBO J. 24, 919–924
Received 29 March 2005