Introduction

Since its early applications to biological systems in the 1950s, electron paramagnetic resonance (EPR) spectroscopy has been a valuable tool in understanding the chemistry of metal centers and organic radicals within proteins. It is an ideal technique for studying electron transfer systems such as those found in photosynthetic membranes because it involves the detection of unpaired electrons. As a result, a large body of the photosynthesis literature includes the routine application of EPR spectroscopy. This review focuses on the application of EPR spectroscopy to the manganese cluster (Mn4Ca) of photosystem II (PSII) and the numerous signals obtained from its various oxidation states.

EPR spectroscopy (also known as electron spin resonance or ESR spectroscopy) depends on the absorption of microwave radiation during electron spin-state transitions of one or more unpaired electrons (S ≥ ½). (For a general explanation of EPR spectroscopy, possible sources include Palmer (1985), Pilbrow (1990), Weil et al. (1994), and Brudvig (1995).) The spin state of the Mn4Ca cluster is defined by the coupling of electron spins of the four Mn ions, which are thought to be dominated primarily by Mn3+ and Mn4+ during the various S-states. As the Mn4Ca cluster proceeds through its catalytic cycle, the overall oxidation state of the cluster increases and the mode of coupling between Mn spins is likely to change, with the possibility of more than one coupling scheme in any one S-state. States with spin S > ½ may show zero field splitting (ZFS), i.e., separation of the MS energy levels in the absence of a magnetic field, which can lead to observed g-factors that are quite different from the “free electron” value of 2.0023 and that depend on the operating microwave frequency. EPR absorption lines may be split into multiple lines by hyperfine coupling of the electron spin with one or more nuclear spins, which in the case of a single 55Mn nucleus with I = 5/2 can lead to splitting into six lines. Both the g-factor and hyperfine coupling constants can show anisotropy that arises from inequivalence of principal axes about the spin center, indicating the type of symmetry of the molecular environment. These many possible features combined with the variety of spin states of the Mn4Ca cluster make it one of the most challenging problems in EPR spectroscopy found in the natural world.

At this point, researchers have obtained EPR signals from nearly all of the S-states of the Mn4Ca cluster, as well as from modified and intermediate states (Table 1). PSII prepared from higher plants, usually spinach, has been the standard for the study of EPR signals from the Mn4Ca cluster and this will be the source referred to below unless otherwise indicated. This review will describe those signals obtained from continuous wave (CW) EPR spectroscopy, in which the magnetic field is swept at a constant microwave frequency, usually X-band (9–10 GHz). Although spectra obtained by CW-EPR are influenced by ligands of the metal spin center, the ligand couplings are often not resolved. Interpretation of unresolved ligand couplings has relied on pulsed EPR methods such as electron spin echo envelop modulation (ESEEM) and electron-nuclear double resonance (ENDOR) (Britt et al. 2000). Previous reviews of the application of EPR spectroscopy in the study of the Mn4Ca cluster include one on the use of pulsed and parallel polarization EPR methods (Britt et al. 2000), an interpretation of results from CW-EPR, pulsed EPR, and X-ray absorption techniques (Åhrling et al. 2005), and a thorough survey of EPR signals from PSII discovered before 1991, including most of those from the S2 state (Miller and Brudvig 1991).

Table 1 Summary of EPR signals from the manganese cluster of PSII observed at X-band (9–10 GHz)a
Full size table

S2 state signals

The S2 state EPR signals are by far the best characterized of all signals from the Mn cluster of PSII because they have been known for the longest time (Dismukes and Siderer 1980, 1981; Casey and Sauer 1984; Zimmermann and Rutherford 1984). They include a multiline signal centered at g = 2 and a broad signal centered at g = 4.1 (Fig. 1, trace 1). The multiline signal shows a hyperfine pattern that is indicative of Mn nuclear couplings, thereby demonstrating the presence of the Mn cluster that had been long suspected of being at the catalytic center of oxygen evolution. Study of the S2 state signals was facilitated by the development of controlled illumination techniques that produced a single electron transfer from the Mn cluster (Brudvig et al. 1983).

Fig. 1
figure 1

Effect of ethanol and fluoride on the S2 state multiline and g = 4.1 EPR signals in PSII. PSII-enriched thylakoid membranes from spinach were either untreated, treated with 5% ethanol, or treated with 20 mM NaF, as indicated. For fluoride treatment, the PSII membranes were depleted of chloride by dialysis before NaF addition. Spectra shown are the difference between the states after illumination at 195 K minus the dark-adapted states. Spectra (X-band) were taken at 20 K using 20 mW microwave power

Full size image

The multiline signal shows 19 or more main hyperfine lines spaced by 80–90 G, indicating that it arises from an antiferromagnetically coupled mixed valence manganese cluster. It is now generally agreed that it arises from the S = ½ ground state of the Mn4Ca cluster probably in the MnIIIMnIV 3 valence state. The g-tensor and hyperfine pattern have been studied extensively (Dismukes and Siderer 1981; de Paula et al. 1987; Hansson et al. 1987; Haddy et al. 1989; Bonvoisin et al. 1992; Åhrling and Pace 1995; Zheng and Dismukes 1996; Hasegawa et al. 1998; Charlot et al. 2005), since their origin is related to the electronic and geometric structure of the Mn4Ca site. Studies at Q-band frequency (34 GHz) (Hansson et al. 1987; Smith et al. 1993; Haddy et al. 2004) showed that the signal is generally isotropic, with a central g-factor of 1.98. A recent W-band (94 GHz) study using a single crystal of PSII from thermophilic cyanobacteria revealed all three principal g-factors of 1.988, 1.981, and 1.965 (Matsuoka et al. 2006). A small amount of hyperfine anisotropy visible in the signal is likely to be contributed by MnIII, which typically shows Jahn–Teller distortion. The multiple shoulders that are evident on many of the main hyperfine lines at X-band were resolved in a study of the signal at S-band frequency (3.9 GHz) (Haddy et al. 1989), which revealed 40–50 separate hyperfine lines (Fig. 2). Hyperfine constants of the Mn ions contributing to the multiline signal have been accurately determined using 55Mn-ENDOR (Peloquin et al. 2000; Kulik et al. 2005), revealing significant contribution from all four Mn ions and axial symmetry for each. Study of the pH dependence of the S2 state multiline signal and the S-state transitions leading to and from it have been carried out (Geijer et al. 2000; Bernát et al. 2002), revealing pK as that are probably associated with the protonation state of side chains ligating the Mn4Ca cluster.

Fig. 2
figure 2

S-band (3.9 GHz) and X-band (9.4 GHz) spectra of the S2 state multiline signal. PSII samples were prepared with 2% ethanol and 100 μM DCMU for the S-band spectrum and 4% ethanol for the X-band spectrum. Both spectra are aligned at g = 1.98 and cover the same magnetic field width. The S-band spectrum was reproduced from Haddy et al. (1989)

Full size image

The signal at g = 4.1 has an isotropic appearance with width of 340–360 gauss and no resolved hyperfine structure. This appearance led to an early suggestion that the signal arose from rhombic Fe3+ in the S = 5/2 state (Casey and Sauer 1984), but later flash oscillation studies showed that it is associated with the S2 state of the OEC (Zimmermann and Rutherford 1986). Its origin in a Mn cluster was later clarified by the resolution of hyperfine structure with 36 G spacing in samples that were NH3 treated and oriented by partial dehydration (Kim et al. 1990, 1992). The signal arises from the middle Kramers doublet of an S = 5/2 spin state with zero field splitting parameters of D = 0.455 cm−1 and E/D = 0.25. The S = 5/2 spin state was demonstrated by multifrequency EPR studies (Fig. 3) (Haddy et al. 1992, 2004) and supported by pulsed EPR (Astashkin et al. 1994) and SQUID magnetization (Horner et al. 1998) studies. Use of higher-microwave frequencies generally improves the resolution of g-factors. At P-band (15 GHz) the signal shows partial resolution of two apparent g-factors at 4.7 and 3.85 (Fig. 3, middle spectrum) (Haddy et al. 1992). At Q-band (34 GHz) a broad signal corresponding to only one principal g-factor appears at about g = 3.1 (Fig. 3, bottom spectrum) (Haddy et al. 2004); the shift in g-factor and disappearance of two of the principal axis g-factors is indicative of ZFS energy that is comparable to the operating frequency. At X-band the signal shows a g-factor of 4.1 rather than 4.3, the position expected for rhombic S = 5/2 systems, because the Mn4Ca cluster is not completely rhombic, i.e., E/D = 0.25 rather than 0.33 which represents complete rhombicity. Another previously observed Q-band signal, with features at about g = 4.3 and g = 4.1, was identified as the S2 state g = 4.1 signal, leading to the suggestion that the signal arose from an S = 3/2 state (Smith et al. 1993; Smith and Pace 1996); however, the identity of this signal has been called into question (Haddy et al. 2004). The X-band g = 4.1 signal was thought not to form in cyanobacterial PSII (McDermott et al. 1988; Boussac et al. 1998a), however, it was recently shown that its presence correlates with the cytochrome c 550 (PsbV) content of the preparation (Lakshmi et al. 2002).

Fig. 3
figure 3

Q-band (34 GHz), P-band (15.5 GHz), and X-band (9.65 GHz) spectra of the S2 state g = 4.1 signal. The magnetic fields are displayed such that g-factor scales are aligned, with spectra centered at g = 4.1 (dashed line). The Q-band and X-band spectra are reproduced from Haddy et al. (2004). The identity of the Q-band signal at g = 4.6 is not clear, but see Haddy et al. (2004) for discussion. The P-band spectrum was previously unpublished, but is similar to those found in Haddy et al. (1992)

Full size image

Although arising from the same overall valence state, the two S2 state signals show characteristics that indicate they arise from two different spin states of the Mn4Ca cluster. This is indicated by the influence of the biochemical treatment on their relative intensities, as described in the following paragraphs. It is further supported by temperature-dependence studies of the signal intensities. These experiments, which have proven very difficult to carry out, indicate that both signals are associated with ground spin states (Hansson et al. 1987; Britt et al. 1992; Boussac and Rutherford 2000). The appearance of two different spin states of an exchange-coupled cluster in the same sample (S = ½ and S = 5/2 in this case) is typical of a state of “spin frustration,” in which the two spin states have a similar likelihood of occurring.

The two S2 state signals are frequently observed in the same sample, although their relative intensities are affected by the temperature of illumination and the presence of small alcohols, including methanol (3–5%), ethanol (3–5%), and the cryoprotectants glycerol (50%) and ethylene glycol (30%). PSII prepared in the absence of alcohols in sucrose buffer shows both S2 state signals after illumination at 200 K, whereas samples containing small alcohols do not show the g = 4.1 signal after illumination at 200 K (Fig. 1, trace 2) (Zimmermann and Rutherford 1986; de Paula et al. 1987; Pace et al. 1991). Samples either with or without small alcohols show the g = 4.1 signal but no multiline signal after illumination at 130–140 K, and upon warming to 200 K intensity is lost from the g = 4.1 signal with concurrent formation of the multiline signal (Casey and Sauer 1984; de Paula et al. 1985, 1987). The loss of g = 4.1 signal intensity due to the presence of methanol correlates closely with the appearance of multiline signal intensity with an apparent binding constant around 50–65 mM (Force et al. 1998; Deák et al. 1999). Methanol, ethanol, and propanol have been shown to bind at the manganese cluster using ESEEM measurements (Force et al. 1998).

Both S2 state signals are influenced by the presence of Cl−, which is required to fully activate oxygen evolution activity. Cl− is required for formation of the multiline signal and the signal height has been found to be correlated with oxygen evolution activity (Yachandra et al. 1986). Other activating anions (Br−, NO −3 , I−) have also been found to support multiline signal formation (Damoder et al. 1986; Yachandra et al. 1986; Ono et al. 1987; Lindberg and Andréasson 1996; Olesen and Andréasson 2003; Bryson et al. 2005). On the other hand, Cl− depletion or substitution of Cl− with anions such as F− or N −3 , which are competitors of Cl− activation, prevent multiline signal formation (Fig. 1, trace 3) (Casey and Sauer 1984; Damoder et al. 1986; Yachandra et al. 1986; Ono et al. 1987; van Vliet and Rutherford 1996; Haddy et al. 2000; Olesen and Andréasson 2003). In one study, the multiline signal was found to form after the addition of Cl− in the dark to Cl−-depleted PSII that had been flash treated (Ono et al. 1986), indicating that oxidizing equivalents to produce the S2 state had been accumulated before the multiline signal had formed. The g = 4.1 signal, on the other hand, is able to form whether or not PSII is Cl− depleted or Cl− substituted with inhibitory anions such as F− or N −3 (Casey and Sauer 1984; Ono et al. 1986, 1987; DeRose et al. 1995; Lindberg and Andréasson 1996; van Vliet and Rutherford 1996; Haddy et al. 2000; Olesen and Andréasson 2003). However, treatment with F− or N −3 , and possibly Cl− depletion, results in narrowing of the line width by about 10% (Casey and Sauer 1984; Haddy et al. 1992, 2000), indicating that Cl− influences the environment of the g = 4.1 signal spin center. Several studies have reported enhancement of the g = 4.1 signal by F−, which is partially explained by the narrower line width. An approximate inverse correlation was shown between the heights of the g = 4.1 signal and the multiline signal in response to F− concentration (DeRose et al. 1995).

A signal that is virtually identical to the g = 4.1 signal is produced by near IR illumination of samples in the S2 state at 140–150 K with concurrent loss of the multiline signal (Boussac et al. 1996, 1998b). The absorbance maximum of this transition is at 820 nm or 12,200 cm−1. The formation of the signal under these conditions is thought to be due either to a charge transfer within the Mn4Ca cluster from MnIII to MnIV or to a spin-state conversion of MnIII. An alternative proposal attributes the near IR absorption to a spin-allowed d-d transition in a single MnIII ion (Baxter et al. 1999). Since the g = 4.1 signal formed by near IR illumination relaxes back to the multiline signal at 200 K, the near IR component of broad light sources is not thought to be responsible for the g = 4.1 signal observed after illumination at 200 K. The near IR-induced conversion from multiline signal to g = 4.1 signal takes place via an intermediate S ≥ 5/2 spin state characterized by EPR signals at g = 6 and g = 10, which can be trapped by illumination at about 65 K (Boussac et al. 1998b). Similar intermediate signals at g = 5–9 have been observed in cyanobacteria (Boussac et al. 1998a), although the near IR-induced g = 4.1 signal has not.

A number of treatments produce modified versions of the S2 state multiline signal, most of which have been subsequently found to arise from centers in which the direct ligation to Mn has been altered. PSII treated with the inhibitor ammonia at pH 7.5 shows a multiline signal with line spacings of 65–70 G when binding takes place in the S2 state (Fig. 4, trace 4) (Beck and Brudvig 1986a, b; Andréasson et al. 1988; Ono and Inoue 1988; Boussac et al. 1990a). NH3 binds to two sites at the OEC, one of which is associated with Cl− activation. Binding of NH3 to the Cl− activation site causes narrowing of the g = 4.1 signal by about 10%, while binding to the non-Cl- site, which is possibly a water substrate site, modifies the multiline signal. NH3 bound at the latter site has been shown by ESEEM measurements to be directly ligated to the Mn4Ca cluster (Britt et al. 1989). Substitution of Sr2+ for Ca2+ at the Mn cluster results in a modified multiline signal with line spacing of 70–75 G (Fig. 4, trace 2) (Boussac and Rutherford 1988; Boussac et al. 1989; Ono and Inoue 1989; Tso et al. 1991; Latimer et al. 1995), similar to the effect of NH3. Sr2+ also causes an increase in the intensity of the g = 4.1 signal. Ca2+ depletion by several methods leads to a “dark stable” multiline signal with 25 or more lines of about 55 G spacing (Fig. 4, trace 3) (Boussac et al. 1989; Sivaraja et al. 1989; Ono and Inoue 1990). This signal persists at 0–20°C for up to hours in the dark after illumination, showing that the absence of Ca2+ from the Mn cluster stabilizes the S2 state. This signal has been closely associated with the split signal from the inhibited S2YZ· state (see below), which forms after illumination of the same samples.

Fig. 4
figure 4

Effect of Ca2+ depletion, Sr2+ treatment, and NH3 treatment on the S2 state multiline signal. For the control, Ca2+-depleted, and Sr2+-treated PSII samples, Ca2+ depletion was carried out using low-pH/citrate treatment, with 50 mM CaCl2 (control) or SrCl2 added as indicated; spectra are reproduced from Latimer et al. (1995). For the NH3-treated sample, the sample was prepared with 50 mM (NH4)2SO4 at pH 7.5 and illuminated at 273 K for 30 s to ensure binding of NH3 in the S2 state; the spectrum was reproduced from Andréasson et al. (1988)

Full size image

S0 state signal

The S0 state shows a multiline signal with 24–26 hyperfine lines with spacings of 80–90 G (Fig. 5) (Åhrling et al. 1997; Messinger et al. 1997a, b). The signal is produced by flashing forward through three S-states (Åhrling et al. 1997; Messinger et al. 1997b) or by chemical reduction with hydroxylamine to form the S0* state (Messinger et al. 1997a). The signal originates from an S = ½ ground state (Åhrling et al. 1998), which probably results from coupling within a MnIII 3MnIV or a MnIIMnIIIMnIV 2 cluster. The hyperfine coupling of the S0 multiline signal is observed only in the presence of methanol, although the broadened signal is observed in its absence. This is reminiscent of the effect of alcohols on improving the resolution of the hyperfine lines of the S2 state multiline signal. 55Mn-ENDOR has been used to determine the hyperfine constants of the Mn ions contributing to the S0 multiline signal (Kulik et al. 2005), favoring a MnIII 3MnIV assignment of the oxidation state. The pH dependence of the S0 state multiline signal, as well as its formation and loss, has been studied along with the S2 state multiline signal (Geijer et al. 2000; Bernát et al. 2002).

Fig. 5
figure 5

S0 state multiline signal compare with the S2 state multiline signal in PSII samples containing 1.5% methanol. The S0 state was prepared by flash advancement of the PSII sample, while the S2 state was produced by continuous illumination at 200 K. Both show difference spectra of the S0 or S2 state minus the S1 state. Spectra were reproduced from Messinger et al. (1997b)

Full size image

S1 state signals

Two EPR signals have been detected from the S1 state using parallel mode detection, in which the magnetic field component associated with the microwave radiation is polarized parallel to the applied magnetic field. Parallel mode EPR, as opposed to perpendicular mode used in conventional CW-EPR, is very useful for detecting partially allowed “ΔMS = 0” transitions in integer spin systems. The S1 state, with one more electron than the S2 state, is diamagnetic (S = 0) in the ground state (Koulougliotis et al. 1992), but integer spin excited states of the same spin manifold (S = 1, S = 2, etc.) are accessible.

The first report of an S1 state signal was of a featureless signal at g = 4.8–4.9 with width of about 600 G (Fig. 6a) (Dexheimer and Klein 1992; Yamauchi et al. 1997). This signal is thought to arise from an S = 1 state with ZFS parameters of D = −0.125 to −0.14 cm−1 and E/D = −0.20 to −0.25. The S = 1 state is separated from the ground state by about 2.5 K, implying an exchange coupling of J = −0.87 cm−1 for a two-spin system (Yamauchi et al. 1997). Methanol and ethylene glycol suppress the signal, as do Ca2+ or Cl− depletion (Yamauchi et al. 1997). The loss of the S1 state g = 4.8 signal upon proceeding to the S2 state appears to correlate with the formation of the S2 state multiline signal, but not with formation of the g = 4.1 signal (Dexheimer and Klein 1992; Yamauchi et al. 1997).

Fig. 6
figure 6

S1 state signals, observed using parallel mode detection: (a) Signal at g = 4.8 in spinach PSII; (b) Multiline signal at g = 12 in PSII from Synecocystis sp. 6803. Both show difference spectra of the S1 state minus the S2 state. Spectrum (a) was reproduced from Dexheimer and Klein (1992) and spectrum (b) was reproduced from Campbell et al. (1998b)

Full size image

Another S1 state EPR signal appears at about g = 12 and shows 18 or more hyperfine lines with an average splitting of 32 G (Fig. 6b) (Campbell et al. 1998a, b), features characteristic of a Mn cluster. The signal is detected only in higher plant PSII lacking the extrinsic PsbQ and PsbP subunits (17 and 23 kDa) (Campbell et al. 1998a) or in cyanobacterial PSII (Campbell et al. 1998b), which contains PsbU and PsbV (cytochrome c 550) as dissociable extrinsic subunits. The importance of the extrinsic subunits in observing the S1 state multiline signal indicates that they influence the magnetic properties of the Mn cluster.

S3 state signals

The S3 state (generated by two flashes of light or by illumination at 235 K) shows EPR signals using parallel and perpendicular mode EPR spectroscopy (Matsukawa et al. 1999; Ioannidis and Petrouleas 2000, 2002; Ioannidis et al. 2002). Parallel mode EPR reveals broad signals at g = 8 and g = 12 (or g = 17 using peak location) with widths of about 300 G and 200 G, respectively. Conventional perpendicular mode reveals a broad signal at g = 6.7 (or g = 10 using peak location). These signals are thought to arise from an S = 1 spin state with ZFS parameters D = ±0.435 cm−1 and E/D = −0.317 (Matsukawa et al. 1999). Very similar signals have also been observed in cyanobacterial PSII (Boussac et al. 2000).

S2′ state and S3′ state signals

Several signals have been observed from states of the Mn4Ca cluster that are evidently high-spin intermediates of the S-states (reviewed in Petrouleas et al. 2005). A signal is observed at about g = 5 (g = 4.7–4.8) in samples that are prepared in the S3 state and subsequently incubated in liquid N2 for several days (Fig. 7) or illuminated with near IR radiation at 50 K (Nugent et al. 1997; Ioannidis and Petrouleas 2000, 2002; Sanakis et al. 2001). The signal evidently arises from a modified S2 state, referred to as S2′, which is achieved by decay of the S3 state. The S2′ signal at g = 5 is thought to arise from an S = 7/2 spin state of the Mn4Ca cluster (Sanakis et al. 2001). The S2′ state also shows a second signal at g = 2.9 (Ioannidis et al. 2002), which arises from a different transition of the same spin manifold as the signal at g = 5. The signal at g = 5 also appears in samples at pH 8.1 after illumination of the S1 state at 243 K (Ioannidis and Petrouleas 2002), suggesting that the signal is associated with a proton-deficient form of the S2 state.

Fig. 7
figure 7

Appearance of the S2′ signal at g = 4.7 after decay of the S3 state. PSII that had been treated with 0.2 mM atrazine and 2 mM ferricyanide was illuminated at 243 K to induce a mixture of the S2 and S3 states (top), then incubated at 77 K for 11 days resulting in decay of the S3 state to the S2′ state (bottom). The top difference spectrum shows the state after 243 K illumination (S2/S3) minus the dark-adapted state (S1). The bottom difference spectrum shows the state after 77 K incubation (S2′) minus the top spectrum; the negative multiline signal shows decay of the S2 state. Spectra were taken at 10 K using 20 mW microwave power

Full size image

Additional EPR signals have been observed in samples prepared in the S3 state, illuminated with near IR radiation at 4 K, then warmed to 190 K, leading to a modified S3 state (S3′). These appear at g = 21 and g = 3.7 using perpendicular mode EPR spectroscopy (Ioannidis et al. 2002) and arise from an integer spin state of the Mn4Ca cluster.

Metalloradical signals

Several signals from the OEC appearing at g = 2 are attributed to the interaction between an organic free radical, usually identified as Tyr Z, and the Mn4Ca cluster (for a summary see Petrouleas et al. 2005). The first metalloradical signal of this type to be observed was a symmetrical 130–165 G wide “split” signal from Ca2+-depleted PSII (Boussac et al. 1989, 1990b; Sivaraja et al. 1989; Ono and Inoue 1990; Hallahan et al. 1992), which appears to consist of two lines (Fig. 8, trace 4). It is produced by illumination of samples showing the “dark stable” S2 state multiline signal (described above) at or above 250 K. The signal shows fast relaxation properties typical of a spin S = ½ radical interacting with a paramagnet, which in this case is the S = ½ spin state of the modified S2 state Mn cluster. Similar signals, with line widths depending on the treatment, are produced in PSII samples containing high concentrations of acetate (230 G wide) (MacLachlan and Nugent 1993; Szalai and Brudvig 1996), fluoride (160 G wide) (Baumgarten et al. 1990), or NH3 (100 G wide) (Andréasson and Lindberg 1992; Hallahan et al. 1992). In addition, a very similar signal that is 90–100 G wide has been produced in active PSII samples in the S3 state by elevation of the pH above 8.5 (Geijer et al. 2001). ESE-ENDOR and ESEEM studies of Ca2+-depleted and acetate-treated PSII have identified the radical as Tyr Z (Gilchrist et al. 1995; Peloquin et al. 1998), although a His radical has also been proposed (Boussac et al. 1990b). The signal has therefore been assigned to the S2YZ· state. Extensive study using multifrequency EPR spectroscopy has revealed details about the dipolar couplings and exchange interactions in the Ca2+-depleted and acetate-treated S2YZ· states (Lakshmi et al. 1998; Dorlet et al. 1999). The Ca2+-depleted system has under some conditions been suggested to include two signals in the g = 2 region, a symmetric doublet signal and an asymmetric singlet-like signal (Astashkin et al. 1997; Mino et al. 2000).

Fig. 8
figure 8

Metalloradical signals from the S0Y .Z state, S1Y .Z state, S2Y .Z state, and inhibited S2Y .Z (+acetate) state. The S0Y .Z and S1Y .Z signals correspond to the states produced by 4–10 K illumination of PSII in the S0 and S1 states, respectively. The S2Y .Z signal corresponds to the state produced after flash illumination of the S2 state at 190 K, followed by rapid cooling to 77 K, then 10 K. The S2Y .Z (+acetate) signal corresponds to the state produced by room temperature illumination of PSII in the presence of ∼500 mM acetate. The S0Y .Z , S1Y .Z , and S2Y .Z (+acetate) spectra are reproduced from Petrouleas et al. (2005); the S2Y .Z spectrum is reproduced from Ioannidis et al. (2006)

Full size image

Since Mn4Ca donates an electron to the reaction center via YZ, one would expect the presence of an intermediate involving the YZ radical on each S-state transition. Indeed, metalloradical signals have been observed for several S-states in uninhibited PSII with intact, Ca2+-containing Mn4Ca clusters. These signals decay within minutes even at liquid helium temperatures. Observation of those associated with the two lower S-states relies on illumination below 10 K to prevent oxidation of the Mn4Ca cluster. A signal attributed to the S1YZ· state has been observed after illumination of PSII samples in the S1 state or after near IR illumination of samples in the S2 state (Fig. 8, trace 2) (Nugent et al. 2002; Koulougliotis et al. 2003; Zhang and Styring 2003). The signal is typified by a peak at g = 2.035, but also includes a line at g = 2.0. The signal is produced by the interaction of the S = ½ organic radical with the S = 1 excited spin state of the S1 state Mn4Ca cluster, as demonstrated by simulation of EPR data from both X-band and W-band (94 GHz) (Koulougliotis et al. 2004). Another metalloradical signal attributed to the S0YZ· state has been observed at g = 2.0 after 5 K illumination of PSII samples in the S0 state (Fig. 8, trace 1) (Zhang and Styring 2003). This 160-G-wide signal has a symmetrical appearance similar to that of the “split” signal from the Ca2+-depleted S2YZ· state. Similar to the S2YZ· signal, it is produced by the interaction of an S = ½ organic radical with the S = ½ spin state of the S0 state Mn4Ca cluster. The S0YZ· and S1YZ· metalloradical signals have also been observed in cyanobacteria (Zhang et al. 2004). More recently, a signal attributed to the S2YZ· state has been observed in uninhibited PSII (Fig. 8, trace 3) (Ioannidis et al. 2006). This 116 G-wide signal is produced by flash illumination of the S2 state at 77–190 K, followed by rapid cooling to 77 K, then 10 K.

Summary

Through the use of EPR spectroscopy, nearly all of the S-states and many intermediate states have been characterized. The results support the basic picture of the OEC’s catalytic cycle as involving a cluster of four coupled Mn ions that increase in valence state as the S-states advance, with the S0 and S2 states present as half integer spin states and the S1 and S3 states present as integer spin states lying above zero spin ground states. Specific valences and coupling schemes of the S-states are supported by various simulation studies. The recent discovery of several intermediate state signals involving an interaction of the YZ radical with the intact Mn4Ca cluster correlates well with our understanding of YZ as the direct electron donor to Mn4Ca. We seem to be on the verge of correlating the parameters of the EPR signals with specific structural features of the OEC that are revealed by the recent X-ray crystallography studies (Zouni et al. 2001; Kamiya and Shen 2003; Biesiadka et al. 2004; Ferreira et al. 2004; Loll et al. 2005), made possible by the increasing resolution of those studies.

At a deeper level, the many signals discovered by the EPR studies indicate a more complicated picture that is so far only partially understood, but provides clues to the details of the catalytic mechanism. The presence of more than one signal in many of the states (e.g., the multiline and g = 4.1 signals in the S2 state, the low field multiline and g = 4.8 signals in the S1 state) is suggestive of a flexibility of coupling that may be important for a cluster that must move easily from one oxidation state to another. The influence of protonated states (pH), various ions, alcohols, and other affectors on many of the signals reveals points where a ligating amino acid is involved, an ion is required as cofactor, or a substrate H2O molecule has access to the Mn4Ca cluster. The recurrent theme of sensitivity to near IR radiation is suggestive of a key bond or point of electron transfer that is common to many of the S-states. These details will take time and much further study to interpret, but are the ones that will eventually lead to a full understanding of how the electron donor H2O is converted to O2.