Stanier and Van Niel (1962) were the first to have the opinion that blue-green algae must be classified under bacteria. This led Roger Stanier to write a letter to Peter Raven in 1970 (November 5, 1970, National Archives of Canada, MG 31, accession J35, vol. 6) where he used the term “cyanobacteria” as a replacement for “blue-green algae.” By 1974, cyanobacteria became “official” since it was included as such in Bergey's Manual of Bacteriology (Buchanan and Gibbons, 1974). It took several more years before the name “blue-green algae” was abandoned, in favor of cyanobacteria, in the photosynthesis literature (for historical issues, see Sapp, 2005a,b).

It was about 2.4 billion (Ga) years ago that oxygen began to accumulate in significant amounts on our Earth and oxygenic life became possible leading to the evolution of life, as we know it today (Falkowski, 2006; Kump, 2008). However, there is ample evidence that almost a billion years earlier, morphological forms of cyanobacteria were present in stromatolites (Allwood et al., 2006; Bosak et al., 2009; Figure 1). In all likelihood, oxygen produced by those cyanobacteria was sequestered in iron deposits, and thus not released into the atmosphere. There is also the possibility that some of the earliest cyanobacteria behaved like anoxygenic photosynthetic bacteria. A major difference between cyanobacteria and algae is that the former are prokaryotic (i.e., they lack membranes around their organelles) and that respiratory and photosynthetic complexes share a common membrane. Almost all oxygenic photosynthesizers, with the exception of Acaryochloris, a cyanobacterium, use chlorophyll (Chl) a (see an overview by Björn et al. (2009) on why Chl a was the chosen one). For evolutionary diversity of cyanobacteria, see Tomitani et al. (2006) and for images of various cyanobacteria, search for “cyanobacteria, images” using “Google.”

Most cyanobacteria, in addition to some carotenoids, contain Chl a, phycocyanin and allophycocyanin as pigments for photosynthesis (Govindjee and Mohanty, 1972; Govindjee and Braun, 1974; Papageorgiou, 1996; Govindjee, 1999; Figure 2). Several cyanobacteria, in addition, contain phycoerythrin (Figure 2). Some (e.g., Prochlorococcus) contain Chl b, and Acaryochloris sp. are known to contain Chl d (Swingley et al., 2008). Further, some Chl a-containing cyanobacteria can be transformed to contain Chl b as well (Vavilin et al., 2003). This shows the great flexibility of cyanobacteria. Cyanobacteria have a unique position in life as they are the precursors of chloroplasts of algae, and, thus, of all plants (Björn and Govindjee, 2009). An interesting phenomenon, recently discovered, is that certain viruses carry either photosystem I (PSI) genes (Sharon et al., 2009) or photosystem II (PSII) genes (Mann et al., 2003; Lindell et al., 2005). This may have interesting implications for the past as well as future evolution. In addition, some differences between cyanobacteria and higher plants are also found on the water-oxidizing side (oxygen-evolving complex, OEC) of PSII. Thus, the water-splitting Mn₄CaO₅ cluster (Umena et al., 2011) is stabilized by three extrinsic proteins on the lumenal side of PSII. In the case of higher plants, these extrinsic proteins are PsbO (33 kDa), PsbP (23 kDa), and PsbQ (17 kDa) (Figure 3A); in cyanobacteria (Figure 3B), the two smaller proteins are substituted by cytochrome (Cyt) c550 (PsbV, 17 kDa) and PsbU (12 kDa) (Roose et al., 2007).

As mentioned above, cyanobacteria are prokaryotes. Like photosynthetic eukaryotes (algae and plants), they transfer electrons, linearly, from water to NADP, utilizing PSII and PSI, and a Cyt b₆f complex that sits in – between these two photosystems. The result is evolution of oxygen, reduction of NADP, and the formation of ATP. In addition, they also cycle electrons around PSI to drive extra ATP production (Bendall and Manasse, 1995; Van Thor et al., 2000).

Literature on cyanobacteria is extensive. We refer the readers to books edited by Desikachary (1972), Carr and Whitton (1982), Packer and Glazer (1988), Bryant (1994), Herrero et al. (2008), Gault and Marler (2009), and Peschek et al. (2011) and to several web sites:

For basic information on photosynthesis, we refer the readers to books by Rabinowitch and Govindjee (1969; available at http://www.life.illinois.edu/govindjee/photosynBook.html), Ke (2001), Blankenship (2002), and by Falkowski and Raven (2007). In addition, Scientific American articles are useful in understanding the basic steps (Rabinowitch and Govindjee, 1965; Govindjee and Govindjee, 1974; Govindjee and Coleman, 1990).

In this personal perspective, we will focus mainly on the adventures of one of us (G) that relate to his work on cyanobacteria with many graduate students, postdoctoral associates and researchers around the World. A perspective of this nature is essential to appreciate the Frontiers in Plant Physiology. We present here a summary of discoveries on specific cyanobacteria that will include new spectral forms of Chl a and P750 (in Anacystis (A.) nidulans), the two-light effect and the two-pigment system, excitation energy transfer that seems to follow Förster's Resonance Energy Transfer theory (from phycobilins to Chl a), the first observation of fast fluorescence changes, the first demonstration of the role of bicarbonate on the electron-acceptor side of PSII, the relationship of slow fluorescence changes with changes that are unrelated to the quencher “Q_A,” but related to “conformational” changes and/or “state changes” (energy distribution between the two photosystems), and regulation of this energy distribution by various factors, and most recently, the observation of light-induced oxygen evolution in PSI-minus mutants in mixotrophic cultures and in the presence of glucose that demonstrates the flexibility of cyanobacteria.

Unlike anoxygenic photosynthetic bacteria (Blankenship et al., 1995; Hunter et al., 2009), cyanobacteria, like algae and plants, perform oxygenic photosynthesis (Mimuro et al., 2008). For that they use two-pigment systems and two-light reactions. Emerson and Chalmers (1958) discovered that, in A. nidulans, light absorbed in phycobilins enhanced the yield of oxygen evolution when added on top of light absorbed in Chl a. Thus, they proposed that in cyanobacteria, the “short-wave” photosystem is run by phycobilins and the “long-wave” photosystem is run by Chl a (also see Emerson and Rabinowitch, 1960). Kok (1959) showed that the long-wave (red light) system oxidized P700 (P700⁺; a special Chl a dimer; the reaction center, RC, of PSI), whereas the short-wave (orange light) system reduced P700⁺ and he related these results to Emerson's 2-light effect. Govindjee and Rabinowitch (1960a) found that a small peak for Chl a was present in the same system that phycobilins were. This solved the dilemma how phycobilins could run the primary photochemical reactions themselves since Duysens (1952) had found that, in cyanobacteria, phycobilins transferred energy very efficiently to Chl a. Cederstrand et al. (1966) discovered, using an integrating dodecahedron, at least two spectral absorption bands in the Chl a region: Chl a 670 and Chl a 680: the idea of several different Chl a spectral absorption bands was already known, from derivative absorption spectroscopy (see Brown and French, 1961). In the green alga Chlorella and in the diatom Navicula, Chl a 670 belonged to the same system as to where Chl b (or Chl c) belonged (Govindjee and Rabinowitch, 1960b).

Amesz and Duysens (1962) provided experimental data on redox changes in the Cyt and NADP that supported the following picture of electron transfer path in cyanobacteria:

(Scheme 1)

Here, Q refers to a plastoquinone Q_A, P700 to RC of PSI, and X to ferredoxin.

For a historical perspective on the evolution of the Z-Scheme of electron transport in photosynthesis, see Govindjee and Björn (2011). In addition to the linear electron transport from water to NADP, there is also a cyclic electron transport around PSI. See Figure 4 for a current version of the Z-scheme of electron transport in cyanobacteria. Early measurements on ATP production in cyanobacteria, using the luciferin–luciferase assay, were made by Bedell and Govindjee (1973) in A. nidulans and by Lubberding and Schroten (1984) in Synechococcus sp.

The flexibility of cyanobacteria was demonstrated recently (see an abstract by Govindjee et al., 2011; Q. J. Wang, A. Singh, H. Li, L. Nedbal, L. A. Sherman, J. Whitmarsh and Govindjee, unpublished) when cells of a PSI-minus mutant of mixotrophically grown Synechocystis sp. PCC 6803 were shown to evolve oxygen (∼25% of that of wild type) for 30 min in the presence of glucose. Studies are under way to find the mechanism for this observation. Perhaps, it involves uphill electron transfer using Δ pH made by PSII.

In addition to the Emerson enhancement effect, an unusual observation was that in several algae and in A. nidulans, far red light (720–760 nm range) caused inhibition (de-enhancement) of photosynthesis when added to short-wavelength light (Govindjee and Rabinowitch, 1960; Govindjee et al., 1960; Rabinowitch et al., 1960). Soon thereafter, Owens and Hoch (1963) using ¹⁸O-labeling and mass spectrometry showed that this de-enhancement was due to the effect of light on respiration (O₂ uptake).

In search of pigments, which absorb in this wavelength region, Govindjee et al. (1961) discovered a new pigment, in A. nidulans, that had an absorption band at 750 nm (also, see Govindjee, 1963a; Figure 5). P750 had an emission band at 760 nm (Govindjee, 1963b) that was confirmed later by Goedheer and Hammans (1975). Unfortunately, no correlation of this band with photosynthesis was found (Gassner, 1962). This is an example of dead-end research as far as the laboratory of one of us (G) was concerned; however, Van Baalen (1965) found that it was involved in photo-oxidation of uric acid in A. nidulans; and Fischer and Metzner (1969) stated that P750 may be an open chain tetrapyrrole, but it not a bacteriopheophytin, and not a chlorin.

Öquist (1974) found that iron-deficiency in A. nidulans increased fluorescence at 755 nm, perhaps, due to an increase in the concentration of P750. Goedheer and Hammans (1975) observed excitation energy transfer from several pigments to P750; and Hammans et al. (1977), however, established that P750 actually initiated oxygen uptake in A. nidulans. Further, Hammans (1978) showed that P750 sensitized photo-oxidation of several endogenous reductants in A. nidulans; however, these reductants did not serve as electron donors to PSI, nor were they oxidized in any Chl-sensitized reactions; there was, however, a hint that excitation of both P750 and Chl may cause the formation of singlet oxygen. Murata et al. (1981) showed that P750 was present in the cell envelope, not in the thylakoids, where there was also a small amount of pheophytin (Pheo)-like pigment with absorption maximum at 673 nm, and even some carotenoids. This observation explained why P750 had nothing to do with photosynthesis. Nultsch et al. (1983) found that P750 was absent in Anabaena and that it was not involved in phototaxis. Worcester et al. (1986) related an in vitro micellar system of aggregated Chls to P750; a correlation was suggested with the chlorosomes of the green bacteria. Later, Gombos et al. (1987) showed that P750, an aggregated form of Chl a, was formed during nitrogen starvation of cyanobacteria at ∼39°C, and that this process was reversible. Interestingly, however Shubin et al. (1991) found that, in Spirulina platensis, redox titration shows that F758 (fluorescence band at 758 nm) intensity has the same midpoint potential as P700; they concluded that the fluorescence quantum yield of F758 is proportional to the concentration of P700. Thus, in contrast, to other observations, they suggested a link of P750 to photosynthesis, at least in this organism. Further research is needed to understand fully the role of P750, i.e., where exactly it occurs in the cyanobacterial cells, what function or functions it performs, and possibly its importance in the evolutionary tree.

Chl a fluorescence (FL) and its relation to photosynthesis has been reviewed by many (see chapters in Govindjee et al., 1986, and in Papageorgiou and Govindjee, 2004). For historical perspectives on this subject, see Govindjee (1995) and Govindjee (2004). In order to fully appreciate this topic, we refer the readers to Wydrzynski and Satoh (2005) and Govindjee et al. (2010) to obtain a background on PSII since it is the system that has the most relationship with Chl a FL. For PSI, see Golbeck (2006).

When plants, algae, and cyanobacteria are illuminated and then quickly frozen to very low temperatures (e.g., 77 K) in darkness, and then warmed in dark, they emit light, and, this light emission is the afterglow, called thermoluminescence (TL). Arnold and Sherwood (1957) were the first to observe this phenomenon in dried chloroplasts. TL has been used extensively to monitor the water oxidation cycle (referred to as the S-state or the Kok cycle) as well as the redox status of the primary and secondary electron acceptors (Q_A and Q_B, respectively) of PSII since TL involves recombination reactions of the electron donor and acceptor sides of PSII. The so-called “B” bands of TL are due to recombination of S₂/S₃ states with reduced Q_B, whereas the “Q” band is due to recombination of the S₂ state with reduced Q_A (Sane and Rutherford, 1986; Demeter and Govindjee, 1989; Vass and Govindjee, 1996; Vass, 2005). Excited Chl a (Chl*), which is needed to give light, is produced when electrons on reduced Q_A and Q_B move up by thermal energy, and positive charges on the S-states provide the “holes” to the ground state of Chl a, with which electrons can recombine. In their theory for TL, DeVault et al. (1983) and DeVault and Govindjee (1990) included temperature-dependent equilibrium between the various steps in the electron transport pathways and provided a generalized theory to explain the major TL bands in the photosynthetic systems. For TL bands, due to reduced Q_B, Rose et al. (2008) did show that the theory of DeVault and Govindjee (1990) is essential for explaining all the TL bands, although, for TL band due to reduced Q_A, simpler models may be sufficient (Rappaport and Lavergne, 2009).

Many investigators have observed TL from cyanobacteria; these have provided important information about their photosynthesis characteristics. Some examples are:

Cyanobacteria, unlike other anoxygenic photosynthetic bacteria, require bicarbonate anion for their PSII activity. Bicarbonate (HCO3−) is a traditional term, but “hydrogen carbonate” is now recommended for it since IUPAC uses the latter; we can abbreviate it as HC. For reviews on the bicarbonate effect and PSII, see Govindjee and Van Rensen (1978), Stemler (1982), Vermaas and Govindjee (1982), Blubaugh and Govindjee (1988), Diner et al. (1991), Govindjee (1991, 1993), Govindjee and Van Rensen (1993), Govindjee et al. (1997), Van Rensen et al. (1999), Stemler (2002), Van Rensen and Klimov (2005), and McConnell et al. (2011). For the absence of bicarbonate effect in the photosystems of anoxygenic photosynthetic bacteria, see Shopes et al. (1989), and Wang et al. (1992).

HC has been suggested to function both on the electron acceptor as well as on the electron donor side of PSII; there is no effect on PSI, however (Khanna et al., 1977; Eaton-Rye and Govindjee, 1984). Data on higher plant thylakoids suggest that the diffusing species is CO₂ and the binding species is HCO3−/CO32− (Sarojini and Govindjee, 1981; Blubaugh and Govindjee, 1986). Using thylakoids, the HC effect has been located on the electron-acceptor side, close to the herbicide-binding side, particularly on the electron transport and protonation reactions at the secondary quinone Q_B and near the non-heme iron (Govindjee et al., 1976, 1984, 1991a, 1992; Jursinic et al., 1976; Khanna et al., 1980, 1981; Van Rensen and Vermaas, 1981; Eaton-Rye and Govindjee, 1988a,b). Several studies have established that the HC effect can be observed in intact cells and in leaves (Garab et al., 1988; El-Shintinawy and Govindjee, 1990; El-Shintinawy et al., 1990; Govindjee et al., 1993). To study the HC effect, very often formate (HCO2−) is used to obtain HCO3−/CO2-free system to which HCO3− anions are added to measure the HC effect (Figure 9A). Nugent et al. (1988) showed clear effects of removal of HC (by formate addition) directly on the Q_A-Fe-Q_B signals in the PSII of the cyanobacterium Phormidium laminosum (Figure 9B). Both the ESR bands related to Q_A and Q_B (in the g = 1.8 and g = 1.6 regions) as well as to non-heme region (g = 6 region; shown on spinach samples; Figure 9C) were clearly affected upon the addition of formate (100 mM). The release of CO₂ by formate addition was shown first by Govindjee et al. (1991b; 1997) using mass spectrometry and infrared spectroscopy; this has been recently confirmed by Shevela et al. (2008a,b).

From the early work of Blubaugh and Govindjee (1988), we have known that there are at least two binding sites of HC. Thus, besides the site at non-heme iron between Q_A and Q_B, there has to be another HCO3−-binding site, which, however, was not seen by the recent crystallographic studies of PSII (Guskov et al., 2010; Umena et al., 2011). Most probably, at the second binding site, may be loosely bound to PSII, and may have escaped detection due possibly to its loss caused by X-ray irradiation as well as by the treatments required during the X-ray measurement.

In view of a large body of experimental evidence for HC effect on the PSII acceptor side, and the latest PSII crystal structure at the resolution of 1.9 Å that shows the clear presence of HC on the non-heme iron between Q_A and Q_B in a thermophilic cyanobacterium Thermosynechococcus vulcanus (Umena et al., 2011), there is a renewed interest in understanding the mechanism of HC effect on the PSII acceptor side. There are clear evidences that HC is not a tightly bound constituent of the OEC of PSII (Aoyama et al., 2008; Shevela et al., 2008b; Ulas et al., 2008; Umena et al., 2011); however, there are numerous indications for the action of HC on the water-splitting side of PSII. For the HC effects on the donor side of PSII, see reviews by Stemler (2002) and Van Rensen and Klimov (2005).

Cao and Govindjee (1988, 1990) were the first to show the effect of HC on PSII in transformable Synechocystis sp. PCC 6803. Govindjee et al. (1993) confirmed this effect in other cyanobacterial cells. Through Chl a FL measurements, the effect was located on the electron-acceptor side of PSII. In view of the fact that in chloroplasts, HC effect was suggested to be near the herbicide-binding site, Govindjee et al. (1990) examined herbicide-resistant D1 mutants of Synechocystis 6714. Chl a FL and O₂ evolution measurements revealed a differential sensitivity to HCO3−-reversible formate inhibition; resistance to formate treatment followed the order (highest to lowest): [double mutant] A251V/F211S (AZ V) > [single mutant] F211S (AZ 1) > wild type > [single mutant] S264A (DCMU II-A). These results are in agreement with the concept that there may be overlapping herbicide and bicarbonate niches in cyanobacteria as well.

Soon thereafter, the concept arose that even D2 protein may be involved with the HC effect, particularly its specific arginines (Cao et al., 1991). Measurements of oxygen evolution showed that the D2 mutants R233Q (arginine-233 ∼ glutamine) and R251S (arginine-251 ∼ serine) were 10-fold more sensitive to formate than the wild type of Synechocystis sp. PCC 6803, suggesting their possible role in HC binding. However, further, studies on D1-S264 and D1-F255 mutants of Synechococcus sp. PCC 6942 led to the conclusion that the equilibrium dissociation constant for HC is increased in the mutants, while that of the formate remains unchanged. The hierarchy of the equilibrium dissociation constant for HC (highest to lowest, ±2 mM) was: D1-F255L/S264A (46 mM) > D1-F255Y/S264A (31 mM) – D1-S264A (34 mM) – D1-F255Y (33 mM) > wild type (25 mM). These data suggested the importance of D1-S264 and D1-F255 in the HCO3−-binding niche. A possible involvement of HC and these two residues in the protonation of QB−, the reduced secondary plastoquinone of PSII, in the D1 protein was discussed (Cao et al., 1992).

A picture began to emerge that we may not be dealing with specific binding site of HC that is involved in the reactions, but that conformation changes may be responsible for some of our observations in cyanobacteria. This became rather obvious when Vernotte et al. (1995) used several single and double mutants in the d–e interhelical loop of D1 protein in Synechocystis sp. PCC 6714. They showed a differential sensitivity of formate inhibition on the acceptor side of PSII in the following order: S264A-F255L > S264A = N266T > N266T-S264A > F255L = N266D = WT > > N266T-A251V = A251V. On the basis of the absence of additivity of the effects between single and double mutants, Vernotte et al. (1995) concluded that the couples S264 and N266, N266 and A251, and S264 and F255 interact with each other in formate binding. These results are best explained to be due to the modification, to some extent, of the “overall conformation of the D1 protein which, in turn, modified the HCO2−/HCO3−-binding at its site in the same protein.” In a way, “long-distance” (conformational?) effect on the redox properties of the Q_A and Q_B seems to support such a thesis since mutations on the d-e loop affect S2QA− and S2QB− and their recombination (Maenpaa et al., 1995; Mulo et al., 1997). Maenpaa et al. (1995) found that formate which inhibits the Q_A to Q_B reaction was several-fold less effective in the CA1 mutant (where three glutamates at 242, 243, and 244 positions were deleted and glutamine 241 was changed to His); on the other hand D1-E229D and D1-E243K mutants were like the wild type Synechocystis sp. PCC 6803. Mulo et al. (1997) demonstrated, also in Synechocystis sp. PCC 6803, that deletion of either the PEST like sequence (ΔR225-F239) or the putative cleavage region (ΔG240-V249, ΔR225-V249) of the D1 protein resulted in severe perturbations on the function of the Q_B electron acceptor of PSII.

Xiong et al. (1997), considering all the results available till then, presented a model of PSII, based on homology with the bacterial system, and introduced the hypothesis of how HCO3− may function in PSII. Obviously, only the basic points are known: HC was suggested in the literature to bind to the non-heme iron that sits between Q_A and Q_B. To accommodate other experiments, particularly those on Chlamydomonas mutants (Hutchison et al., 1996; Xiong et al., 1997), specific arginines were implicated as well as conformational changes. What is known from the work of Eaton-Rye and Govindjee (1988a,b) on spinach thylakoids is that HC functions not only in the reduction of Q_B, but also in its protonation. Recent studies have suggested that D2-K264 (Cox et al., 2009), but more likely D1-Y246 and D2-Y244 (Takahashi et al., 2009) are directly involved in HC binding. Figure 10 shows the binding of HCO3− to the electron-acceptor side of PSII from the 1.9 Å resolution structure of PSII (Umena et al., 2011). Here the importance of D1-Y246 and D2-Y244 in the HC effect is clearly obvious. Komenda et al. (2002) discovered that a mutant of Synechocystis sp. PCC 6803, that lacked PsbH protein, had a much weakened binding of HC on the PSII acceptor side; this observation needs to be understood in light of the recent PSII crystal structure (Umena et al., 2011).

Answer to the HC problem lies in future studies on the mutants of these and some other amino acids. So, we would say that a new hope has arisen and we can hardly wait to get the final answers!

We end this review with remarks on evolution (for recent reviews, see Buick, 2008; Björn and Govindjee, 2009; Blankenship, 2010; Hohmann-Marriott and Blankenship, 2011). Figure 11 shows a simplified model of the origin and evolution of life (tree of life) based on small-subunit RNA analysis. Despite the simple representation of the origin of the first oxygenic photosynthesizers in this Figure, the earliest steps in the origin of photosynthesis are not known, and are the subject of vivid discussions in the literature (Blankenship et al., 2007; Buick, 2008 and references therein).

The authors stated that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.