Expression of the isiA gene is essential for the survival of the cyanobacterium Synechococcus sp. PCC 7942 by protecting photosystem II from excess light under iron limitation


Youn-Il Park. E-mail; Tel. (+82) 42 821 5493; Fax (+82) 42 822 9690.


Iron deficiency is known to suppress primary productivity in both marine and freshwater ecosystems. In response to iron deficiency, certain cyanobacteria induce a chlorophyll (Chl)–protein complex, CP43′, which is encoded by the isiA gene. The deduced amino-acid sequence of CP43′ predicts some structural similarity to the CP43 polypeptide of photosystem II, but the function of CP43′ remains uncertain. In order to assess its physiological role, the isiA gene of a cyanobacterium, Synechococcus sp. PCC7942, was inactivated by insertion mutagenesis (giving isiA cells). Compared with isiA cells, under iron deprivation, wild-type cells showed both lower rates of photosystem II-mediated O2 evolution at limiting light irradiances and decreased yields of room temperature Chl fluorescence at various irradiances. These observations strongly suggest that the decreased photosystem II activity in wild-type cells with CP43′ is attributable to increased non-radiative dissipation of light energy. In agreement with this hypothesis, isiA cells were more susceptible to photoinhibition of photosynthesis than wild-type cells, resulting in much slower growth rates under iron limitation. Based on these results, we suggest that CP43′ functions as a non-radiative dissipator of light energy, thus protecting photosystem II from excessive excitation under iron-deficient conditions.


Iron is an essential redox component for many critical cellular processes, including photosynthesis, respiration, nitrogen fixation, ribonucleotide and haem syntheses. Although iron is known as the most abundant transition metal in the earth's crust, its availability from aquatic photoautotrophs is normally limited because of its low solubility in aerobic ecosystems (Braun et al., 1990). Consequently, availability of iron may be an important factor determining the structure and function of the photosynthetic apparatus (Ferreira and Straus, 1994) and, hence, primary productivity (Behrenfeld et al., 1996). However, cyanobacteria have developed various means to cope with conditions of severe iron deficiency, including the production of siderophore-based iron-scavenging systems (Straus, 1994).

Among the most interesting changes triggered by iron deficiency in many cyanobacteria such as Synechococcus, Synechocystis and Anabaena is the induction of an iron stress-induced Chl-containing protein, CP43′, which is encoded by the isiA gene of the isi operon (Laudenbach and Straus, 1988; Burnap et al., 1993; Leonhardt and Straus, 1994; Falk et al., 1995). The ferric uptake regulation (fur ) gene is known to regulate this isiAB operon (Ghassemian and Straus, 1996).

Based on the homology of predicted amino-acid sequences between CP43′ and CP43, a Chl protein complex of photosystem II (PSII), three possible physiological functions of CP43′ have been proposed: it may act as (i) an alternative antenna complex for PSII to compensate for the partial loss of phycobilisomes (Pakrasi et al., 1985); (ii) a functional replacement for CP43 in iron-deficient cells (Burnap et al., 1993); or (iii) a Chl reservoir to facilitate recovery from iron deficiency (Troyan et al., 1989; Burnap et al., 1993). However, we favour the view that CP43′ has a unique role other than CP43, as CP43 is such a conserved functional unit in oxygenic photosynthesis. Further, storing Chl under iron-deficient conditions may be wasteful and harmful rather than beneficial, as it would be easily destroyed by the inevitable absorption of light. In fact, cyanobacteria survive and even form massive blooms at concentrations of soluble iron well below the level required to induce isiA expression (Ferreira and Straus, 1994).

The main objective of this study was to determine the physiological role of CP43′ in an iron-starved cyanobacterium, Synechococcus PCC 7942. Clarifying the exact role of CP43′ is necessary for our understanding of aquatic ecology with respect to the global carbon cycle, as cyanobacteria together with prochlorophytes are the dominant species throughout most of the open ocean (Falkowski et al., 1994), where productivity is often limited by iron deficiency (Behrenfeld et al., 1996).


Physiological characteristics of isiA cells under iron stress

Cells of Synechococcus sp. PCC 7942 grown under iron deficiency showed characteristic symptoms such as chlorosis, a short wavelength shift of the red Chl a absorption peak and a dominant 77K Chl fluorescence emission spectrum peak at 685 nm (Table 1 and Fig. 1B[link]), confirming previous work (Öquist, 1974; Guikema and Sherman, 1983). Among these changes, both the short wavelength shift of Chl absorption and the dominant 685 nm emission peak at 77K are directly related to the expression of the isiA gene encoding for CP43′, as isiA cells under iron stress did not show such phenomena (Table 1 and Fig. 1B[link]). The isiA inactivation strain used in the present study differs from other mutant cells with an inactivated isi operon (Burnap et al., 1993; Falk et al., 1995). To exclude the possible pleiotropic effect caused by the absence of the isiB gene in previous isiAB inactivation strains (Burnap et al., 1993), we put the isi promoter immediately after the kanamycin resistance cassette interrupting the isiA gene, allowing the isiB gene to be expressed normally under iron deficiency. Consistent with previous work (Laudenbach and Straus, 1988; Burnap et al., 1993; Falk et al., 1995), iron deficiency severely inhibited the growth of isiA cells compared with wild-type cells (Fig. 1A). Furthermore, the isiA cells failed to induce both the short wavelength shift of the red Chl absorption peak (Table 1) and the dominant 685 nm peak in the 77K Chl fluorescence emission (Fig. 1B), when grown under iron deficiency.

Table 1. . Characteristics of Synechococcus sp. PCC 7942 wild-type cells (WT) and a CP43′-lacking mutant (isiA) grown in iron-sufficient (+) and -deficient (−) BG-11 media. Chlorophyll (Chl) and phycocyanin (PC) contents are expressed as relative content (nmol/A750). Amax is the room temperature absorption peak of Chl in the red region. Number of functional PSII (mmol PSII mol−1 Chl) was obtained from repetitive flash O2 yield measurement. Fo is the minimal Chl fluorescence in the dark-adapted cells. Each value is the mean of 5–10 independent experiments (± SD).Thumbnail image of
Figure 1.

. A. Growth curves on a Chl basis of Synechococcus wild-type (WT; ▪, □) and isiA (○) cells after transfer to iron-sufficient (▪) and -deficient (□, ○) conditions. B. The 77K Chl fluorescence emission spectra of Synechococcus wild-type (WT) and isiA cells grown under iron-sufficient (+Fe) and -deficient (−Fe) conditions. Data for isiA cells grown under iron-sufficient conditions are not presented, as they are very similar to WT cells.

PSII-mediated O2 evolution and room temperature Chl fluorescence

Light-response curves of O2 evolution mediated by PSII (H2O → DCBQ) on a Chl basis showed that, compared with cells grown in iron-sufficient conditions, iron deficiency decreased the rate of photosynthetic O2 evolution at all light regimes from limiting, through to saturating and excess lights (Fig. 2A, insert). In particular, the slope for the assessment of photon yield of O2 evolution at limiting light (Φ) in iron-deficient cells (Φ = 0.49, relative units) was half that of the control cells (Φ = 0.99, relative units). However, iron-deficient isiA cells showed a similar apparent quantum yield to wild-type cells grown in iron-sufficient medium, although maximal PSII activity was similar to that of wild-type cells grown under iron deficiency. Iron deficiency hardly changed the dark-adapted Chl fluorescence yield (Fo, minimal Chl fluorescence yield when all PSIIs are open) on a Chl a basis in wild-type and isiA cells (Table 1), which allowed us to normalize the Chl fluorescence yields induced by actinic light (F) and saturating light (Fm′, maximal Chl fluorescence yield when all PSIIs are closed) during steady-state photosynthesis (Fig. 2B). Upon illumination of wild-type cells, both F and Fm′ increased rapidly to maxima at low irradiances. As the irradiance exceeds the limiting values for photosynthetic O2 evolution, both parameters decreased slightly. However, iron-deficient isiA cells lacking CP43′ showed much higher F and Fm′ levels than wild-type cells at all light irradiances, suggesting that CP43′ is involved in the quenching of Chl fluorescence.

Figure 2.

. A. Light–response curves of photosynthetic O2 evolution mediated by PSII in Synechococcus wild-type (WT; ▪, □) and isiA (○) cells grown under iron-sufficient (▪) and -deficient (□, ○) conditions. B. Light–response curves of room temperature Chl fluorescence yields induced by various actinic (F; ▴, ▵) and saturating (Fm′; ▾, ▿) light during steady-state photosynthesis in wild-type (▵, ▿) and isiA (▴, ▾) cells grown under iron-deficient conditions. Data for WT and isiA cells grown under iron-sufficient conditions are not presented, as they are very similar to isiA cells.

Photosynthetic performance in iron-limited cells under high light treatment

The proposed role of CP43′ as an excitation energy quencher made us check whether cells with CP43′ have increased resistance to high light stress, i.e. to photoinhibition of photosynthesis. For this purpose, cells enclosed in the O2 electrode chamber were illuminated under saturating light (750 μmol m−2 s−1), and the rate of O2 exchange was measured continuously as a function of time of illumination. Lincomycin (1 mg ml−1) was included to block the repair of photodamaged D1 protein of PSII (Fig. 3). Photoinhibition of photosynthesis was seen as a decrease in O2 evolution. Wild-type cells grown under iron deficiency showed a clear lag phase after the onset of inhibitory high light exposure, whereas iron-sufficient wild-type cells showed no lag in the onset of photoinhibition. Clearly, iron-deficient wild-type cells containing CP43′ are much more resistant to photoinhibition of photosynthesis than isiA cells lacking CP43′. The somewhat higher resistance after 10 min illumination observed in both wild-type and isiA cells grown in iron-deficient conditions than in wild-type cells grown in iron-sufficient conditions could be attributable to the smaller physical antenna size of PSII (i.e. a decrease in the phycobilin content; Table 1), as photoinactivation of PSII would depend on the antenna size (Park et al., 1997).

Figure 3.

. Changes in the photosynthetic O2 evolution as a function of saturating light treatment (750 μmol m−2 s−1) in Synechococcus wild-type (▪, □) and isiA (○) cells grown under iron-sufficient (▪) and -deficient (□, ○) conditions. To block the repair process of damaged reaction centre protein D1, lincomycin (1 mg ml−1) was added before the onset of illumination. Data for isiA cells grown under iron-sufficient condition are not presented, as they are very similar to WT cells.

Changes in pigment contents, Chl a absorption peak and 77K emission peak area at 685 nm during iron-induced recovery

In order to test the proposed role of CP43′ as a storage protein of Chl during iron stress (Troyan et al., 1989; Burnap et al., 1993), iron recovery was conducted in the presence of a Chl synthesis inhibitor, carboxymethoxylamine. After the addition of iron, iron-limited cells showed rapid recovery from iron deficiency after a 3 h lag period, as estimated by the increase in the Chl content, restoration of the peak position of the red Chl absorption peak, decrease in the 685 nm 77K fluorescence peak and an increased phycocyanin/Chl ratio (Fig. 4). As expected, carboxymethoxylamine treatment together with iron blocked Chl synthesis and caused cells to maintain the characteristics of the iron-deficient state even 24 h after the addition of iron. Therefore, it is unlikely that the primary function of CP43′ is to serve as a Chl reserve.

Figure 4.

. Time course for the recovery of the Chl a absorption peak wavelength position (A), the Chl a content (B), the ratio of phycobilin (PC)/Chl (C) and the fraction of emission peaking at 685 nm (F685) deconvoluted from the 77K emission spectra (D) after the addition of iron to iron-limited WT cells with (□) or without (▪) the Chl biosynthesis inhibitor, carboxymethoxylamine (CMA; 10 μM).


Iron deficiency represents a unique case of nutrient deficiency for cyanobacteria such as Synechococcus, Synechocystis and Anabaena (Ferreira and Straus, 1994; Straus, 1994). We found that deficiency of nitrate, sulphate or phosphate induces losses in the contents of Chl and phycobilin (data not shown). However, only the iron deficiency-induced chlorosis was accompanied by the characteristic short-wavelength shift of the red Chl a absorption peak and a dominant 77K fluorescence peak at 685 nm, all changes that have been reported before (Öquist, 1974; Guikema and Sherman, 1983). In the present study, using an isiA insertional mutagenesis strain, we confirm the earlier suggestion that these characteristic symptoms of the thylakoid membranes in Synechococcus sp. PCC 7942 subjected to iron deficiency are caused by the accumulation of the isiA gene message and its corresponding protein CP43′ (Burnap et al., 1993; Falk et al., 1995). These two studies and our data also showed that inactivation of the isiA gene severely diminishes cell growth, indicating that this gene determines the viability of cells under iron stress.

CP43′ is not an alternative to CP43 during iron stress

The 50% lower relative quantum yield of PSII in iron-deficient wild-type cells compared with isiA cells (Fig. 2A) indicates that there is less efficient utilization of absorbed light by PSII in the presence of CP43′. The utilization efficiency of light energy by PSII is determined by both the antenna size of PSII and the trapping efficiency. The physical antenna sizes of PSII in the wild-type and isiA cells are comparable, as the Chl and phycobilin contents per PSII are similar (data not shown). Thus, the lower photochemical efficiency is attributable to decreased trapping efficiency. The reduction in trapping efficiency may have its origins in either the water oxidation process or the energy transfer process (Falkowski et al., 1994). The reaction centre quenching should lead to a reduction in both the limiting and the saturating levels of PSII-mediated O2 evolution activity. This prediction is, however, contradicted by the result that PSII with CP43′ is fully active at saturating irradiances for wild-type and isiA cells under iron deficiency (Fig. 2A). Instead, the idea of antenna quenching-induced reduction of trapping efficiency fits well with both the similar maximal activities of PSII at saturating light for iron-deficient wild-type and isiA cells and the lowered photochemical efficiencies at low light of iron-deficient wild-type cells containing CP43′. Our interpretation is that there will be competition between thermal dissipation in CP43′ and excitation energy for the PSII reaction centre. Thus, we propose that the decrease in PSII activity caused by CP43′ is caused by losses of excitation in the pigment bed rather than by malfunction in the reaction centre, resulting in the decreased availability of photons for PSII photochemistry. Therefore, it is unlikely that CP43′ can replace CP43 functionally as a proximal PSII antenna component during iron deficiency as discussed by Burnap et al. (1993).

CP43′ is not likely to be a Chl reservoir for iron recovery

The iron recovery experiments in the presence of the Chl synthesis inhibitor carboxymethoxylamine (Fig. 4) show that CP43′ does not provide Chl molecules to other Chl–protein complexes synthesized during recovery from iron stress. This so-called Chl reservoir hypothesis of CP43′ is supported by the observation of a tiny accumulation of Chl proteins, such as CP43 and CP47, in the presence of the Chl synthesis inhibitor gabaculine (Troyan et al., 1989). These results clearly differ from those of the present study and from those of Guikema (1984), who used laevulinic acid to inhibit Chl synthesis. Maybe the site for gabaculine inhibition of Chl synthesis differs from that of higher plants, as gabaculine-treated Anacystis nidulans showed a substantial accumulation of 5-aminolaevulinic acid (ALA), which cannot easily be understood in view of gabaculine inhibiting ALA formation in higher plants (Guikema et al., 1986).

CP43′ functions as a photon dissipator protecting PSII from light stress

Light absorbed by PSII will be used for photochemical work or dissipated, either as emitted light (fluorescence) or as heat (Horton et al., 1996). The lowered photochemical efficiency of cells containing CP43′ is accompanied by lowered room temperature Chl fluorescence yields (Fig. 2B). Hence, light energy reaching PSII with CP43′ will be dissipated as heat to a greater extent than in PSII with CP43. This implies that CP43′ functions as a non-radiative heat dissipator of excitation energy, resulting in reductions in both Chl fluorescence yields and photochemistry of PSII. If this is so, the PSII complexes with CP43′ should be more resistant to light stress than PSII complexes with CP43, simply because photoinactivation of PSII depends on the total number of photons absorbed by PSII reaction centres (Park et al., 1995). As expected, PSII complexes in iron-starved cells are more resistant to light stress (Fig. 3), supporting the hypothesis that CP43′ has a role as a non-radiative dissipator.

The exact mechanism of how CP43′ causes antenna quenching of excitation energy is as yet unknown. Although further experiments are certainly required, it seems likely that the CP43′-induced antenna quenching results from changes in PSII structure through direct and/or indirect interactions with CP43′. Based on the fact that the major difference between CP43 and CP43′ lies in the length of the largest luminal loop (Burnap et al., 1993), we assume that CP43′ may sit in the position occupied by CP 43 under iron stress. Under such conditions, PSII with CP43′ may undergo structural changes caused by the truncation of the large hydrophilic domain, resulting in inefficient energy transfer to the reaction centre while fully preserving the water-evolving capacity. Alternatively, CP43′ may sit in the thylakoid membranes close to PSII and compete with reaction centres for excitation energy. The finding that CP43′ could not restore CP43 function under iron stress (Ferreira and Straus, 1994) is consistent with our view of CP43′ as an excitation energy quencher.


Cyanobacterial CP43′ induced under iron deficiency functions as a non-radiative dissipator of light energy, protecting photosynthesis from light stress. This process is thought to be a major regulatory mechanism and is likely to have great significance. Damage to the reaction centre occurs under all light conditions, being an inevitable consequence of PSII function and, hence, requires regulation of light harvesting. A substantial fraction of the excited state energy harvested in antenna pigments would, thus, be dissipated by quenching before it is ever transferred to the reaction centre. It is much easier and safer for cyanobacterial cells under iron stress to dispose energy in this way, before initiating the photochemical processes in the reaction centre than it would be for them to repair the substantial photooxidative damage that would result from excess light. Interaction of CP43′ with PSII seems to facilitate this regulation process, but further work is required to assess if, and how, these changes in PSII configuration are related to the mechanism of quenching. This will be of great importance for understanding cyanobacterial photosynthesis and, hence, primary production in oxic ecosystems in which the biological availability of iron is often severely limited.

Experimental procedures

Cell strains and growth conditions

Cells of Synechococcus sp. PCC 7942 were grown axenically in BG-11 inorganic medium (Rippka et al., 1979), supplemented with 10 mM 3-(N-morpholino) propanesulphonic acid, pH 7.5, and bubbled with 5% CO2 in air at 37°C under continuous illumination of 50 mmol photons m−2 s−1 of white light (Philips, TLD 18W/95O). Iron limitation of cells was imposed by culturing cells in BG-11 media lacking iron citrate. Recovery was initiated by adding 30 μM ferric ammonium citrate to the iron-deficient cultures. In some experiments, 10 μM carboxymethoxylamine, an inhibitor of the formation of aminolaevulinic acid, was added to inhibit Chl synthesis during the recovery process.

Construction of isiA inactivation plasmid and transformation

An insertional inactivation of the isiA gene of the isiAB operon was made using the polymerase chain reaction (PCR) technique (GeneAmp PCR system 2400, Perkin-Elmer). A kanamycin resistance cassette was amplified as a selection marker from the plasmid pUC-4K. The construct was cloned into plasmid pUC19 (designated pISI219) with inactivated isiA and transformed into competent E. coli DH5α cells. The positive transformants were selected on LA plates containing 50 μg ml−1 kanamycin. Wild strains of Synechococcus sp. PCC 7942 were transformed (Van der Plas et al., 1990) with plasmid pISI219 and subjected to selection on BG-11 plates containing 5 μg ml−1 kanamycin. When total DNA was isolated from homozygous transformants, digested with HindIII and analysed by Southern blotting (data not shown), the restriction fragments had the sizes expected from a homologous, double recombination event. DNA from the isiA strains hybridized to a slightly smaller fragment than wild-type DNA. The slightly larger size of the fragment from isiA that hybridizes to the isiB probe also confirms the double recombination event.

DNA isolation and hybridization

Southern blot analysis was carried out as described by Ausubel et al. (1989). isiA- and isiB-specific DNA probes radiolabelled with [α-32P]-dCTP were made by PCR amplification of the coding region of isiA (Burnap et al., 1993) and the isiB-specific oligonucleotides used in making plasmid pISI219 respectively.

Room temperature Chl fluorescence and photosynthetic O2 evolution

Chl a fluorescence yield and oxygen evolution were measured simultaneously using a pulse amplitude modulated fluorometer (PAM chlorophyll fluorometer; Walz) with the PAM 103 accessory and a Schott KL 1500 lamp, compatible with a system of cuvette, magnetic stirrer, oxygen electrode and Bjökman type actinic lamp (Hansatech) (Campbell and Öquist, 1996). To measure PSII-mediated O2 evolution in intact cells, 5 mM K3Fe(CN)6 and 2 mM 2,5-dichloro-1,4-benzoquinone (DCBQ) were added sequentially 2 min before illumination. A home-built cuvette system was used to measure O2 evolution resulting from a series of repetitive single turnover flashes (2.5 μs half bandwidth, 10 Hz frequency) supplemented with far-red light (Park et al., 1995). Various flash intensities were obtained by changing the applied voltage to flash tube (type FX2000, EG&G Electro Optics).

Measurements of pigment contents, absorption spectra and 77K fluorescence emission spectra

Absorption spectra of intact cells were recorded with a Shimadzu spectrophotometer (MPS-50L), and the Chl and phycocyanin contents were determined (Myers et al., 1980). 77K Chl fluorescence emission spectra of intact cells were measured using a fibreoptic-based fluorometer as described previously (Ögren and Öquist, 1984).


We thank J. M. Anderson and A. Ivanov for their stimulating discussions. This work was supported by the Swedish Natural Science Research Council and the Swedish Research Council for Engineering Sciences.