Introduction

Evolution of PS2 is currently believed to have begun in sea water about 3 billion years ago (Fournier et al. 2021). This has been sufficient time for evolution to produce and refine a redox enzyme that is a model of (almost) foolproof simplicity. If this were not so, life on earth would likely be a good deal more anaerobic than it is today. This overview encompasses about four decades dating from the appearance of active, highly resolved PS2 preparations from spinach; what we know now is overwhelming in comparison to the information that was available at the time when PS2 was named the “Inner Sanctum” of photosynthesis by Kok and Cheniae (1966). Advances achieved with thylakoid membranes, green algae and cyanobacteria should not be forgotten. The period 4 oscillations in oxygen yield and the S-state model (Joliot et al. 1969; Kok et al. 1969), and the demonstration of ATP synthesis in darkness, driven by an acid–base transition without electron transfer (Jagendorf and Uribe 1966), are among the important discoveries in research on molecular bioenergetics, two of the many significant discoveries made using thylakoid membranes or intact algae. Now PS2 and its OEC (O2 evolving complex) are center stage, as documented by many reviews (for example Barber 2016; Cox et al. 2020; Shen 2015; Vinyard and Brudvig 2017). A comprehensive review by Junge (2019) adds an important historical component to an assessment of scientific progress. The author of this overview will comment on discoveries about PS2 made across a number of years, but no attempt will be made to match the depth of coverage afforded by Prof. Junge’s more extensive review.

Probing the inner sanctum

The author’s generation of PS2 researchers were confronted by a series of experimental results that were, in some instances, puzzling. For example, in the early 1980’s, the simplicity with which pure, active PS2 preparations could be obtained (Berthold et al. 1981; Kuwabara and Murata 1982) went against the opinions of a number of senior investigators who believed this couldn’t be accomplished. (A historical note: the BBY and K&M notations for these preparations were not an invention of the authors of these papers on PS2 isolation (see Dunahay et al. 1984)). Another very intriguing result at the same time was the discovery that the lumenal side of inside-out thylakoid vesicles, thought to be a hydrophobic domain, contained H2O soluble polypeptides associated with PS2 activity (Åkerlund et al. 1982). This observation led to identification of the important extrinsic PS2 subunits now known as PsbO, P and Q in eukaryotes and PsbO, U, and V (a cytochrome, C550) in cyanobacteria (reviewed by Bricker et al. 2012). Using purified PS2 from spinach, extraction of PsbP and Q inhibited O2 evolution and created another puzzle. A general rule in biochemistry is that if extraction of a protein inhibits an activity, reconstitution of that protein should restore activity. No such luck in the case of PS2: repeated reconstitutions of these proteins were ineffective in restoring O2 evolution, but addition of Ca2+, a redox inert metal that would not have been a first choice at the time, was effective (Ghanotakis et al. 1984a). This was an odd result to say the least, and finding its cause led to the discovery that the PsbP and PsbQ extrinsic proteins form a barrier, along with PsbO, to prevent Ca2+ loss (Ghanotakis et al. 1984b) and also to protect the OEC from reductant induced Mn2+ formation and loss from the OEC (Ghanotakis et al. 1984c). Now it is well-established that PsbP has the key role in Ca2+ retention (Ifuku et al. 2005).

The early 1980’s also brought the discovery, first in thylakoid membranes, that the S2 state produced a multiline EPR signal (Dismukes and Siderer 1981), detected at liquid helium temperatures, which is now found in other S-states as well (Haddy 2007). Characterization of effects of Cl− on reversible amine inhibition of steady state O2 evolution activity revealed the presence two sites in the OEC where NH3 inhibits O2 evolution; inhibition at one of the sites is insensitive to interference by Cl−, whereas at the second site both NH3 and Tris (tris-(hydroxymethyl)-aminomethane) inhibitions are reversed by Cl− (Sandusky and Yocum 1984). Another unexpected result was that a tyrosine radical (YD·) is the source of the EPR detectable Signal II species (Barry and Babcock 1987); site-directed mutagenesis identified two redox active residues as Y161(YZ) of PsbA (D1), and Y160 (YD) of PsbD (D2) (Debus et al. 1988; Vermaas et al. 1988). These discoveries joined many others in expanding research on PS2 in a number of areas. Electron paramagnetic resonance spectroscopy (EPR) was used to characterize a number of PS2 signals (Miller and Brudvig 1991). Optical absorbance changes from Mn oxidation state advancements were reported (Dekker et al.1984), and X-ray absorption spectroscopy became an important technique for probing Mn oxidation states and Mn–Mn, Mn-Ca2+ and Mn-ligand distances (Sauer et al. 2005). The identification of Ca2+ as a cofactor in the oxygen evolving reaction and a characterization of the Cl− requirement for activity led to research to establish the stoichiometries of these ions in PS2. The results indicated that a single Ca2+ atom (Ádelroth et al. 1995) and a Cl− atom with high and low affinity binding behavior (Lindberg et al. 1993, 1996) are present in active PS2 preparations. Large scale projects to provide DNA sequences of many species, but of cyanobacteria in particular, provided the background for application of site directed mutagenesis to identify other amino acids in addition to the redox-active tyrosines that are involved in PS2 reactions (Debus 2008). The techniques of molecular biology were also important factors in elucidating the steps in the mechanism for repair of reaction centers damaged by photoinhibition (Järvi et al. 2015) and for defining, in eukaryotes, which genes reside in the chloroplast, and which are found in the nucleus, such as genes for the extrinsic subunits and light harvesting antenna subunits, whose products are imported into chloroplasts (Theg 2018).

Additional research on components and reactions involved in O2 evolution was facilitated by a variety of biophysical techniques. X-ray absorption spectroscopy was applied to characterizations of Mn oxidation states in PS2, but the requirement for a variety of model Mn compounds to fit to the experimentally observed edge spectra made this a demanding research problem. The “high oxidation state” model for the S-states shown below has gained considerable experimental support (see, for example, Cheah et al. 2020).

SO, [3Mn(III)/Mn(IV)]: S1, [2Mn(III)/2Mn(IV)]: S2, [Mn(III)/3Mn(IV)]: S3, [4Mn(IV)]: and S4 [?].

Measurements of H+ release during S-state advancement revealed a stoichiometry of 1,0,1,2 (Junge et al. 2002) for the Mn oxidation state transitions.

Additional X-ray absorption experiments (EXAFS (Extended X-ray Absorption Fine Structure)) provided information on the positioning of the Mn atoms with respect to one another and to putative ligands, using comparisons of the experimental data from PS2 samples to data from Mn model compounds. Generally agreed distances were: 1.8Å (Mn–O); 2.7Å (Mn – Mn); 3.3–3.4 Å (Mn–Mn, Mn-Ca) (Sauer et al. 2005). Results of ENDOR (Electron Nuclear Double Resonance) spectroscopy on the S2 state by Peloquin et al. (2000) produced a new proposal for the structure of the Mn cluster in PS2, a trimer of Mn atoms and a 4th atom, termed a “dangler”, separated from the trimer. The combined results from biochemical and biophysical characterizations of PS2 at this point defined the stoichiometries of components of the OEC and provided important information on their possible arrangements, and Mn oxidation states, while data from site-directed mutagenesis experiments cited above, in cyanobacteria, identified amino acid residues of D1 and D2 that were involved in formation of a functional active site of water oxidation.

Viewing the inner sanctum

The first attempts to crystallize PS2, from spinach, were successful, but the crystals were not suitable for extended X-ray crystallographic analysis (Fotinou et al. 1993). A major breakthrough in PS2 research started in the new millennium with reports that PS2 isolated from thermophilic cyanobacteria (Thermosynochococcus elongatus and T. vulcanus) had been crystallized, and their structures were being determined (Shen and Kamiya 2000; Zouni et al. 2001). Viewed from a distance, these structures fit existing topological data. The OEC was located towards the lumenal side of the structure, the plastoquinones on the opposite (stromal) side and the 6 Chla and 2 Pheoa, bound to D1 and D2 were located as one would have predicted from the structure of the bacterial reaction center. A major advance in understanding reactions of the OEC was the report by Ferreira et al. (2004) of a more detailed structure of the metal cluster, one in which the model of Peloquin et al. (2000) was confirmed and expanded to 3 dimensions, with the “dangler” Mn atom appended to a distorted cube in which the Ca2+ atom replaced a Mn in the 4th corner of the cube near YZ. A significant problem has been overcome since then. Reduction of the Mn oxidation states by radiation exposure has been avoided with the “diffract and destroy” technique, using the XFEL (X-ray Free Electron Laser) technique (Suga et al. 2014), and at least 2 groups using the XFEL technique are in the process of elucidating the structures of the individual S-states (Young et al. 2016; Suga et al. 2017; Suga et al. 2019; Ibrahim et al. 2020).

The new millennium has also seen an increased contribution to research on PS2 structure and function from the application of computational methods (Vinyard and Brudvig 2017). New models for substrate H2O binding have appeared (Wang et al. 2017), and with all of the data now available or emerging, it is reasonable to expect that key steps in the mechanism of H2O oxidation are, and will be, based on data from crystals of PS2, and from experiments that utilize crystal structures for their design. The activities of thermophilic PS2 preparations in the crystalline and soluble forms (Ananyev et al. 2019; Ibrahim, et al. 2020), and structural correlations between metal–metal and metal–ligand distances from the crystals and from characterizations by X-ray absorption spectroscopy provide confidence that the current crystal structures present an accurate representation of the native structure and organization of the polypeptides of cyanobacterial PS2, and of the inorganic ion constituents of the OEC.

Remaining questions

If PS2 crystals are revealing a great deal more about one of the most important biological reactions on earth, should there be a concern over what, if anything, remains to be discovered? What follows offers some observations about the properties of PS2, clearly exposing the author’s biases, that might profit from additional attention. The first of these is whether it will one day be possible to obtain high quality crystals from a eukaryotic source like spinach, where the majority of biochemical and biophysical characterizations have been carried out. The successful crystallization of eukaryotic PSI (Amunts and Nelson 2010) and the early results of Fotinou et al. (1993) suggest that eukaryotic PS2 crystals might one day be available. It would also be interesting to explore the origins of differences between thermophilic and mesophilic PS2 in cyanobacteria. The T. volcanus PS2 preparation requires much higher concentrations of NH2OH (up to 50 mM) for Mn2+ release (Zhang et al. 2017) than does the mesophile Synechocystis sp PCC 6830 (1 mM; Anton P. Avramov et al. 2020). It seems unlikely that NH2OH redox chemistry would be species dependent, so there may be significant differences in channels that lead to the OEC between thermophiles and mesophiles. What can be learned about structural properties of PS2 in these species might also account for the differences in access of a small reductant like NH2OH to the OEC. CryoEM offers the prospect of an alternate technology that can be used to address this question, as it has in the cases of photoactivation of the OEC (Gisriel et al. 2020) and the structure of channels and other features of PS2 from plants, cyanobacteria, and green algae (Sakashita et al. 2017; Gisriel et al. 2022; Nathan Nelson Personal Communication).

There are a number of other properties of the OEC that might deserve additional experimental scrutiny. Starting with the S0 state, there’s a difficulty on account of its instability; S0 is slowly oxidized to S1 in a reaction catalyzed by YD· (see Rutherford et al. 2004). Therefore, obtaining highly resolved structural information on a homogeneous S0 sample would be difficult. This is the OEC oxidation state where initial binding of substrate H2O may occur, and any proposed mechanism for the oxidation-induced reduction of S4 should account for the resulting S0 structure. The core inorganic structure of S0 does not appear to differ drastically from that of the other S-states (Ibrahim et al. 2020); Pantazis (2018) presents an in-depth analysis of possible arrangements of ions in S0, and these observations might be a starting place to seek a better understanding of this S-state.

The S1 state has received considerable attention and is certain not to be ignored in the future. The unique properties of S2, including the appearance of the EPR multiline signal, has caused researchers to sit up and take notice. Early on, Frasch and Cheniae (1980), using thylakoid membranes, showed that irreversible inhibition of O2 evolution by Tris buffer under their conditions was maximum in S2, and Velthuis (1975) had shown that NH3 blocked electron transfer beyond S2. Addition of NH3 along with Tris blocked irreversible Tris inhibition of PS2 activity (Frasch and Cheniae 1980). Setting aside other results on amine binding behavior of the OEC for the time being, what new property of S2, not observed in S1, is responsible for increased Tris sensitivity? A subtle change in structure (a channel?) that allows a large amine to gain access to a site of inhibition, the presence of another Mn(IV), or another factor, not discovered, that increases inhibition kinetics under the experimental conditions (0.8 M Tris in the reaction mixture)?

The mechanism of Ca2+ extraction from the OEC is also an interesting question related to S2. Miyao and Murata (1986) showed that Ca2+ depletion of the OEC is accelerated by illumination in concert with release of extrinsic polypeptides. Subsequently, Boussac and Rutherford (1988) identified S2 and S3 as susceptible states for Ca2+ loss from PS2 samples exposed to high salt (1.2 M NaCl) to remove PsbP and PsbQ. The result of Ca2+ removal under illumination creates a modified multiline EPR signal and an unusually stable S2 state that decays upon addition of Ca2+ (Boussac et al. 1989). Illuminated Ca2+-depleted samples also produce an EPR signal that was later shown to originate from YZ (Gilchrist et al. 1995). These results show that loss of Ca2+ from S2 interferes with electron transfer from the Mn cluster to YZ and raise a number of questions. What is the mechanism of the accelerated release of Ca2+ from its site in the distorted Mn3Ca cube? Is it possible that Ca2+is displaced by another monovalent cation from the solvent during extraction? Potassium and Cs+ are effective competitors for the Ca2+ site in PS2, but Na+ is not (Ono et al. 2001). High concentrations of NaCl (1.2 M), used for polypeptide extraction, would be difficult to test for competition with Ca2+ binding. Is it possible that Ca2+ release in S2 and S3 is due to electrostatic repulsion arising from formations of additional positive charges (Mn(IV) in the Mn3Ca cube? The last, and most complex issue is the question of why Ca2+ release blocks electron transfer from S2 to YZ; the answer could be informative in achieving a complete understanding of the mechanism of O2 evolution.

It’s also curious that the S2 state, after Ca2+ extraction, exhibits a robust stability (Boussac et al. 1989). Positive charges on the metals in dark-stable S1 state and in a putative Ca2+-depleted, monovalent cation substituted S2 state might, in theory, be the same (+ 16) in both states. A major structural reorganization of the metal cluster after Ca2+ removal seems an unlikely cause of stability, given the rapid recovery of activity upon Ca2+ reconstitution in depleted samples during assays of steady state activity (Homann 1988b). Finally, Ca2+ extraction and replacement of the vacant site with Sr2+ slows the steady state rate of O2 evolution, produces changes to the S2 multiline signal, and slows the decay of YZ∙on the S3 → SO transition (Boussac et al. 2004). Does the crystal structure of Sr2+ substituted PS2 (Koua et al. 2013) provide any clues to this observation, which may be linked to the difference in Lewis acidities of Ca2+ and Sr2+ (Vrettos et al. 2001)? Finally, the role of Ca2+ in formation of the Mn4Ca cluster is still being examined; Anton P. Avramov et al. (2020) have characterized the Ca2+- Mn competition in photoactivation and present a model for how the stability of high affinity Mn-binding sites depend on the presence of Ca2+.

The S2 state is probably as good a place as any to consider the Cl− questions(s) and amine inhibition of the OEC. Chloride is not required for formation of S2 in terms of Mn oxidation, but it is required to observe the S2 multiline signal (Ono et al. 1986). Chloride is required for advancement beyond S2 to S3, and for the S3 → S0 transition (Wincencjusz et al. 1997). The majority of biochemical data on Cl− function has been generated using thylakoids and isolated PS2 from spinach and other plants, while structures and the effects of site-directed mutants on Cl− function come from cyanobacteria. Chloride is an essential cofactor for the O2 evolving reaction in eukaryotic PS2; its KM, from reconstitution experiments, is 0.9 mM in thylakoids (Kelly and Izawa 1978), and lower in PS2 preparations depending on assay conditions, such as the presence or absence of extrinsic polypeptides (Homann 1988a) or the assay pH (Baranov and Haddy 2017). Linear double reciprocal plots are characteristic of a single site of activation of O2 evolution by Cl−. Unfortunately, buffer and sucrose solutions contain residual Cl– contaminations, which are likely to be utilized by PS 2 (see, for example, Kelley and Izawa (1978) and Ishida et al. (2008)). In cyanobacterial and plant PS2, evidence for Cl− involvement in the O2 evolving reaction also comes from detection of a slowing of the S3 → S0 decay by biosynthetic replacement of Cl− with Br− in T. elongatus (Ishida et al. 2008) or with Br−, I−, or NO32− in spinach PS2 by incubation reconstitution (Wincencjusz et al. 1999). In PS2 crystals, Cl− is detected at two sites, Cl1 near a narrow channel leading to the dangler Mn4 in the OEC, and Cl2, in a wider channel (Kawakami et al. 2009; Ibrahim et al. 2020). The distances from the metal cluster (6–7 Å) place Cl− in the outer shell around the cluster. In the case of Cl1, the PS2 structure has been used to guide directed mutagenesis experiments using Synechocystis 6803. The mutant D2K312A negatively affects O2 evolution activity, appears to abolish a large fraction of Cl− dependent activity, but leaves behind activity at low rates and a multiline signal that is weaker than the wildtype signal (Pokhrel et al. 2013). Is this signal a result of retention of low levels of Cl−? Otherwise, the hypothesis that Cl1is a structural component in PS2, organizing a hydrogen bonding water network leading away from the Mn/Ca cluster of the OEC (Kawakami et al. 2009; Rivalta et al. 2011)) is consistent with the results obtained with D2K312A.

The use of inhibitors is often one sign that the biochemist using them is unable to vary the substrate concentration, as in the case of PS2 with 55.4 M H2O. Inhibition of O2 evolution activity by primary amines involves Cl− (Sandusky and Yocum 1984, 1986). The Cl− insensitive site binds NH3, and is a Mn atom (Britt et al. 1989) that has been identified by Oyala et al. (2015) as Mn4 in the OEC, near Cl1. Ammonia binding displaces an H2O from the dangler (Marchiori et al 2018). Manganese binding and displacement of a possible substrate H2O would account for NH3 inhibition of O2 evolution and the identity of S2 as the NH3-sensitive state (Velthuis 1975). The problem with respect to Cl− is that all amines, from NH3 and CH3NH2 through Tris to t-butylamine, also inhibit steady state activity (Ghanotakis et al 1983); the inhibition constant (I50) is proportional to the amine pKa, suggesting a metal (Lewis acid) as the binding site of the free bases of these amines (Angelici, 1973). Characterization of this inhibition also showed that all amines compete with Cl− for the second binding site, and again, a linear relationship between inhibition constants and amine pKa values was obtained (Sandusky and Yocum 1986).

What is the identity of the site where all amines tested compete with Cl− and inhibit OEC activity under illumination in the steady state? Is the assumption valid that this site is a Lewis acid? There doesn’t appear to be any EPR evidence for direct binding of larger amines to the Mn cluster, and the steady-state inhibition experiments would indicate that Cl− and amines must exchange rapidly with the second binding site. There is another issue. At room temperature NH3 binding to illuminated thylakoid samples produced the YZ EPR signal (Yocum and Babcock 1981), whose amplitude increased with increasing microwave power up to 200 mW. This was interpreted to arise from a spin–spin interaction between the organic radical and Mn in the OEC. Additional experiments (Ghanotakis et al. 1983) showed that larger amines that do not irreversibly inhibit activity also elicited the YZ signal under room temperature illumination, but the spin–spin interaction was altered; signal saturation occurred at ~ 25 mW for both CH3NH2 and 2-ethyl-2 amino propanediol. Effects of Cl− on amine binding to the OEC were unknown at the time these experiments were conducted and additional experiments were not pursued.

The data on amine-Cl− interactions in eukaryotic PS2 are difficult to rationalize with the crystal structures. Ammonia binding to the dangler Mn4 has been discussed above. The Cl2 site is distant from the metal cluster and YZ, but the environment around YZ includes space that might accommodate both NH3 and larger amines to affect spin–spin interactions between the Mn cluster and the tyrosine radical. This would imply that at least under turnover conditions, the Lewis acid for which amines and Cl− compete could be Ca2+, which is already proposed to be involved in substrate H2O binding to the OEC (Vrettos et al. 2001); the substrate analog CH3OH is reported to bind to Ca2+ in S2 as well (Oyala et al. 2014). In the case of Cl− and amines, these species would have to be involved as inner shell ligands, but not to Mn. This is certain to be an unpopular suggestion because none of the available crystal structures support Cl− binding to Ca2+. However, EPR and ESEEM (Electron Spin Echo Envelope Modulation) results with acetate, which also competes with Cl−, are interpreted to show binding of the former ligand near YZ (Szalai et al. 1998; Klemens et al. 2002). If, as these data might suggest, Cl− binds to Ca2+, what is its role in the mechanism of H2O oxidation? Structural, as part of a substrate H2O binding site on Ca2+? Displacement of a non-substrate H2O? Catalytic, by affecting the Lewis acidity of Ca2+? The rate of O2 evolution activity is sensitive to the Lewis base present: Cl− > Br− > I− ~ NO32− (Wincencjusz et al. 1999).

Determination of the mechanism of the S3 → S0 reaction will reveal how PS2 oxidizes 2H2O to O2. A question here is the identity of the electron donors to S3YZ∙; Cox et al (2020) present a thorough discussion of this issue. In any case the reaction involves an oxidation-induced reduction of both YZ and 3 Mn(IV). In the meantime, Pantazis’s review (2018) provides an in-depth analysis of various remaining issues regarding H2O oxidation, and the crystal structures identify the plentiful supply of potential substrates.

A final OEC issue concerns one of the many site-directed mutations in the mesophilic cyanobacterium Synechocystis 6803. These mutants have proven crucial in identification of essential residues involved in Mn and/or Ca2+ ligation, for example, and have been confirmed by crystal structures. One residue, D1E181 is clearly the donor of oxo ligands, one each to Mn and to Ca2+ in the OEC. An unusual result was the discovery that substitution of a positively charged lysine (E189K) produced an active OEC. This result has been discussed in detail by Kim and Debus (2020), who present evidence from FTIR spectroscopy supporting normal functioning of H2O molecules in the mutant OEC. An alternate, less likely explanation would be that another residue near E189 has substituted for E189K. This has been observed in a H2O-soluble metalloenzyme with highly conserved Fe-ligating cys residues. A conserved cysteine residue (C20A) was mutagenized in a 4Fe/4S cluster in Ferredoxin I from Azotobacter vinelandii (Martin et al. 1991). The mutation retained its 4Fe/4S cluster and X-ray crystallography revealed that a nearby Cys residue (C24) had replaced the ligand function of the mutagenized residue. This seems to be a rare occurrence in site-directed mutagenesis experiments and is not likely to explain E189K’s properties.

Moving away from the OEC to the protein structure of PS2, the effects of removing the extrinsic proteins on PS2 activity raise some questions. In the case of the PsbP and Q subunits, extraction is remediated by additions of Ca2+ and Cl−. The main problem here, and it may be a minor one, is that the activity recovered by this approach does not completely restore the original activity; elevated concentrations of inorganic ions do not compensate completely for activity loss caused by polypeptide extraction (Miyao and Murata 1985). This could be due to a failure to overcome the release of these cofactors in the light (such as expulsion of Ca2+), or to some subtle structural changes induced by the treatments used to remove the polypeptides. In the case of PsbO, it was once known as Manganese Stabilizing Protein (Burnap and Sherman 1991) because its removal caused a slow loss of Mn(II) and activity. Miyao and Murata (1984) showed that this denaturation event could be attenuated by incubation of extracted samples with Cl−, and it has further been shown that the residual activity after PsbO extraction requires high Cl− concentrations (Miyao and Murata 1985). Is this due to structural changes to channels, caused by PsbO extraction? The effect appears to be specific to Cl−. It has also been shown that PsbO is tightly bound to PS2 (Leuschner and Bricker (1996)) and extraction of this subunit from PS2 requires exposure to urea/NaCl or 1 M CaCl2, neither of which can be classified as gentle treatments; extraction procedures might contribute to the properties of PsbO-depleted PS2. The majority of information on the interactions between extrinsic subunits and PS2 come from experiments with eukaryotic PS2. It is difficult to predict what differences might emerge from research on the comparative properties of the extrinsic proteins found in cyanobacteria (Gisriel, et al. 2022). Lastly, a very interesting observation was reported by Ettinger and Theg (1991), who showed that thylakoid membranes from pea and spinach contained unassembled PsbO, P and Q subunits. One cause of this is proposed to be that these proteins are waiting to come to the rescue of photodamaged PS2 reaction centers. Could overexpression of these subunits in vivo provide additional stability to PS2 when plants are under stressful conditions that might destabilize binding of the extrinsic subunits?

Coming finally to the other function of PS2, and to thylakoid membranes as well, it is still unclear how electrons get from QBH2 to the cytochrome b6f complex, which reduces plastocyanin, the mobile carrier that reduces P700+ (Hӧhner et al. 2020). This question has been raised by Bill Cramer in annual conversations (now by Zoom, unfortunately) with the author, who has no answers. What is the nature of interactions between the b6f complex and the reducing side of PS2 that promote rapid transfers of reducing equivalents to the cytochrome complex? Special channels? Diffusion? This is certainly not the only issue left to explore, and the current cohort of PS2 investigators, including the authors contributing to these special issues, with new ideas and techniques can provide substantial additions to any list of needed experiments, along with new answers. That this is the case is a positive sign of the health of photosynthesis research overall. The challenge now is to get back together in person at meetings and international congresses. In the meantime, the author apologizes for losing track of all of the important questions that he’s overlooked in this overview.