Synechocystis and its mutants were grown in BG-11 medium on a shaker (130 rpm) at 30 °C under continuous white light (20 µmol photons m^-2s^-1 for LL) [14]. BG-11 medium with 1.2% (w/v) agar containing 0.3% (w/v) sodium thiosulfate was solidified for plate cultures. The photomixotrophic growth was achieved by diluting an aliquot from a photoautotrophic culture in the log phase (OD₇₃₀ ~1.0) to an OD₇₃₀ ~0.05 with BG-11 containing 10 mM glucose. The growth and cell densities were monitored at OD₇₃₀ on a Beckman DU800 spectrophotometer. The growth was evaluated based on the average of three parallel experiments. Chlorophyll a (Chl a) [15] or phycocyanin (PC) [16] concentrations were determined by absorption spectroscopy and calculated per cell [17].

All genetic manipulations were carried out according to standard protocols [18]. The apcE gene locus (slr0335) in Synechocystis encodes the L_CM protein, which consists the following domains: the N-terminal loop(80–150), Arm1(241–249), Rep1(250–400), Arm2(401–534), Rep2(535–685), Arm3(686–714), and Rep3(715–860). To delete these domains and/or motifs, mutation plasmids were constructed as follows.

2.2.1 Plasmid for Construction of $\mathrm{\Delta }$(rep3)

A 0.8-kbp DNA fragment including arm2-rep2 and a 0.6-kbp DNA fragment including arm3-rep3 of apcE was amplified from Synechocystis genomic DNA by PCR using primers P1–P4 (Supplementary Table 1), as upstream and downstream targeting arms, respectively. The restriction sites XhoI plus EcoRV and EcoRV plus XbaI were introduced in the targeting arms, respectively, for cloning the fragment into pBluescript. A DNA fragment containing streptomycin resistance cassette was excised from plasmid pHB45 via the restriction site SmaI and then inserted in the plasmid containing the targeting arms via the restriction site EcoRV, thereby yielding pBlue-$\mathrm{\Delta }$(rep3)-str (Fig. 1C, Ref. [19]).

Fig. 1.

Construction and characterization of ApcE-mutants in Synechocystis PCC 6803. (A) is the domain location in the ApcE protein and (B) is the $\mathrm{\Delta }$PBL_𝐶𝑀 mutant strain reported recently (Zlenko et al., 2019 [19]). In $\mathrm{\Delta }$(rep3) (C) Rep3 is deleted, in $\mathrm{\Delta }$(arm2/rep3) (D) Arm2 and Rep3 are deleted, in $\mathrm{\Delta }$(rep2/rep3) (E) Rep2 and Rep3 are deleted, in $\mathrm{\Delta }$(arm2/rep2/rep3) (F) Arm2, Rep2 and Rep3 are deleted from L_CM of Synechocystis by inserting the spectinomycin resistance cassette. The other Arm2-truncated mutants were constructed based on $\mathrm{\Delta }$(rep3), and some amino acids deleted from Arm2 are shown in (A). Km^R stands for kanamycin resistance cassette, Cm^R indicates chloramphenicol resistance cassette, and Sm^R is spectinomycin resistance cassette.

The mutant strains were generated by transforming WT Synechocytis cells with the respective plasmids via homologous recombination [20]. Transformants were selected on BG-11 plates containing 12.5 µg/mL spectinomycin, with segregation ensured throughsuccessive rounds of selection at 25 µg/mL spectinomycin. The genotypes of all final mutant strains were confirmed by PCR using respective primer sets (Supplementary Fig. 2, Supplementary Table 1).

The room temperature absorption spectra of the cell samples and the isolated PBS fractions were recorded on a UV-9000S spectrophotometer with a slit width of 1 nm in a 2-mm cuvette with high scattering property. At the corresponding peak wavelengths, molar extinction coefficients of 1185 mM^-1cm^-1 and 770 mM^-1cm^-1 were used for a trimer of PC (620 nm) and APC (650 nm), respectively. According to the absorption spectra of CPC [21], the molar extinction coefficient of 266 mM^-1cm^-1 was used for a trimer of PC (650 nm), while the molar extinction coefficient of 295 mM^-1cm^-1 was used for a trimer of APC in 620 nm according to the absorption spectra of APC [22]. The concentrations of CPC and APC were calculated according to established methods [16, 23]. The equations used were:

$${\mathrm{A}}_{620}={\epsilon }_{\mathrm{CPC}}^{620}{\mathrm{C}}_{\mathrm{cpc}}+{\epsilon }_{\mathrm{APC}}^{620}{\mathrm{C}}_{\mathrm{APC}}$$ $${\mathrm{A}}_{650}={\epsilon }_{\mathrm{CPC}}^{650}{\mathrm{C}}_{\mathrm{CPC}}+{\epsilon }_{\mathrm{APC}}^{650}{\mathrm{C}}_{\mathrm{APC}}$$

Spectra were acquired using an F-320 Fluorolog spectrofluorimeter, with slit widths set at either 5 or 10 nm. Synechocystis WT and mutant strains were cultivated under normal light until reaching an optical density of A730~0.8. For measurement, the chlorophyll a (Chl a) concentration of the cell suspensions was adjusted to either 3 or 5 µg/mL, corresponding to the excitation wavelengths of 580 nm or 430 nm, respectively. Subsequently, 800 µL aliquots of the adjusted cells were dark-adapted for 5 minutes in fresh BG11 medium prior to spectral recording. All experiments were performed with at least three independent biological replicates.

The 77-K fluorescence emission spectra were monitored using a Horiba Fluorolog spectrofluorimeter with a slit width of 8 nm. The cells at CChl = 3 µg/mL (580 nm excitation) were collected by centrifugation and suspended in an equivalent volume of fresh BG11 containing 40% (v/v) glycerol and 25 mM HEPES-NaOH (pH 7.5). In all cases, whole cells were dark-adapted for 15 min before the measurements. Then, the spectra were recorded corresponding to State II. For State I spectra, cells were illuminated with 55 µmol photons m^-2s^-1 of blue light for 5 min. Then, 800 µL suspensions were quickly frozen in quartz tubes by immersion in liquid nitrogen for 10 s. The excitation was measured at 580 nm, and emission was scanned at 600–800 nm [24, 25]. Date are from at least three biological replicates.

The light response curves, NPQ, and state transitions were monitored on a PAM fluorometer (PAM 2500; Walz, Effelrich, Germany) [26, 27]. Mutant and WT cells were grown under similar conditions (A730 = 0.6) and estimated on dark-adapted (15 min) whole cells at a chlorophyll concentration of 3 mg/L. F0 is a minimal fluorescence level determined by illuminating dark-adapted cells with a low intensity of red-modulated light (pulses of 1 s, 1.6 kHz, 0.024 µmol photons m²s^-1). For the measurements of state transitions [28], cells were dark-adapted, irradiated with blue (55 µmol photons m²s^-1) or orange light (20 µmol photons m²s^-1), and subjected to saturating pulses (2000 µmol photons m²s^-1, 30 s) to measure the F_m^′ levels (maximum fluorescence under illumination).

All NPQ induction and recovery experiments were carried out in the presence of chloramphenicol (30 g/mL), a protein synthesis inhibitor, to inhibit protein synthesis [26]. For the measurements of the light response curve, cells (CChl = 20–30 mg/L) were irradiated with various intensities (0–1400 µmol photons m²s^-1) after dark adaptation (15 min) and then subjected to saturating pulses (2000 µmol photons m²s^-1, 30 s) to measure the F_m^′ levels. The initial slope of the rapid light response curve (alpha), the maximal rate of electron transfer (ETRmax), and the minimal saturated light intensity at semi-saturated light intensity (Ik) were calculated from the PAM data using the analysis software [29, 30]. Data are presented as mean $\pm$ standard deviation (SD).

Protein concentration was quantified using the Bradford method [31] with bovine serum albumin (BSA) as a standard. The proteins from sucrose-containing fractions were precipitated by 50% (NH4)2SO4 and collected by centrifugation to remove the sucrose. Then, the samples were diluted with distilled water and precipitated by 10% trichloroacetic acid (TCA) to remove the ammonia. The remaining TCA was removed by acetone extraction. Subsequently, the obtained samples were dried, solubilized in the sample buffer [32], and analyzed by SDS-PAGE using Laemmli buffer system [33]. Finally, the proteins were stained with Coomassie brilliant blue [34].

PBSs were isolated from Synechocystis cells by an ultracentrifugation-based protocol [35]. Briefly, cells were lysed with 2% (v/v) Triton X-100, and the lysate was clarified by centrifugation to remove unbroken cells, debris, and chlorophyll. The resulting supernatant (2 mL) was then layered onto sucrose step-gradients for ultracentrifugation. The gradients, prepared in 12-mL tubes with 0.8 M KPB (pH 7.2), consisted of 1 mL of 2.0 M, 2.5 mL of 1.0 M, 2 mL of 0.75 M, 2 mL of 0.5 M, and 1.5 mL of 0.25 M sucrose solutions. Centrifugation was performed at 230,000 $\times$g for 13 h at 18 ℃ using a P40ST rotor (Hitachi CP80-WX ultracentrifuge). The distinct blue bands containing PBSs were collected and stored at 4 ℃ in sucrose solution for analysis within 48 hours. PBS isolations were performed from at least three independent cultures per mutant.

Electron microscopy was performed on a HITACHI H-7650 transmission electron microscope operated at 100 kV. The images were recorded with a 1024 $\times$ 1024 CCD camera at a magnification of 100,000$\times$. PBS preparations were negatively stained with 2% uranyl acetate by droplet method [7]. Drops of about 5 µL, containing the samples at OD at 620 nm [$\mathrm{\Delta }$(arm2(6-36)/rep3) was 0.02 OD, $\mathrm{\Delta }$(rep3) was 0.05 OD] in 0.8 M potassium phosphate (pH 7.2) were deposited on the surface of glow-discharge and carbon-coated copper grids. After 1–3 min, the droplets were stained with 2% uranyl acetate.

Arm2 in apo-protein of L_CM (ApcE) from Synechocystis constitutes 401–534 aa. Thus, we re-numbered Arm2 as 1–134 aa (Fig. 1A). Strikingly, the amino acid sequence of Arm2 of the red alga Griffithsia pacifica was highly similar to that of Synechocystis (Supplementary Fig. 1A). Based on the 3.5 Å PBS structure from the red alga Griffithsia pacifica [8], the conservative motif of Arm2(6–36) in Griffithsia pacifica formed the super-secondary structure of $\alpha$-helix/loop/$\alpha$-helix (Supplementary Fig. 1B) and the rest of the molecule forms random coils. Arm2 in Synechocystis also forms this super-secondary structure, i.e., a loop (Arm2(18–28)) linking two $\alpha$-helices (Arm2(6–17) and Arm2(29–36)), followed by random coils Supplementary Fig. 1C). To investigate the functions of Arm2, we generated a series of Synechocystis mutants-$\mathrm{\Delta }$(rep3), $\mathrm{\Delta }$(arm2/rep3), $\mathrm{\Delta }$(rep2/rep3), and $\mathrm{\Delta }$(arm2/rep2/rep3)-by selectively deleting the Rep3, Arm2/Rep3, Rep2/Rep3, and Arm2/Rep2/Rep3 domains, respectively, from the ApcE protein (Fig. 1). In order to locate the functional motifs in Arm2, a set of mutants with partial truncations within Arm2 were constructed by truncating an amino acid fragment of Arm2 based on $\mathrm{\Delta }$(rep3) mutant (Fig. 1). Complete segregation of all mutants was confirmed by PCR analysis (Supplementary Fig. 2). The traces of the 1.5- to 2.5-kbp fragments were detected in WT but not the mutants, indicating complete segregation.

The assembly of PBSs in the absences of Rep3 remains a subject of debate. The classic model of PBSs of Synechocystis suggests that Pep3 is essential for Cylinder 3 attachment, with its lost resulting in a simple, two-cylinder core (Cylinder1 and 2) [7]. In contrast, the structural study on PBS on the red alga Griffithsia pacifica [8], Cylinder3 is lost in the absence of Rep3, and then Cylinder1 and 2 are disconnected. Given the symmetry of PBSs, Cylinder1 and 2 are same, and we would only isolate a simple PBS of Cylinder1 or 2 in the core. Next, we isolated the PBSs from $\mathrm{\Delta }$(rep3): the ultracentrifugation of PBSs retrieved three fractions in a molar ratio of 1.47:1:7.17 (Fig. 2A and Supplementary Table 2). The major fraction showed maximal absorption at 620 nm and maximal fluorescence at 668 nm at room temperature and at 684 nm at 77 K. Similarly, the two minor fractions absorbed maximally at 654 nm and exhibited fluorescence at 664–668 nm at room temperature and at 680–684 nm at 77 K. The data indicated the following: (1) while dismantled CPC was barely detected, a small part of the dismantled PBS cores could be isolated. In the isolated cores, the energy absorbed by APC could be efficiently transferred to the two terminal emitters, such that the cores were not damaged. (2) In the isolated PBSs, the energy absorbed by CPC could efficiently be transferred to the two terminal emitters, such that the PBSs remained intact. The SDS-PAGE (Supplementary Fig. 3A lane 3) showed that the major fraction PBSs of $\mathrm{\Delta }$(rep3) assembled intact similar to WT (Supplementary Fig. 3I lane 1), except for a truncated ApcE. The photosynthetic capacity parameters (Supplementary Fig. 4, Supplementary Tables 3,4) and cell fluorescence spectra at room temperature (Supplementary Fig. 5A,D and Supplementary Table 5) and 77 K (Supplementary Fig. 6A,D and Supplementary Table 5) of $\mathrm{\Delta }$(rep3) were similar to those of WT, further supporting the intactness and functionality of the PBSs. However, the absorption spectra of $\mathrm{\Delta }$(rep3) showed that this mutant had 1.3 times less phycobiliprotein per chlorophyll than WT (Supplementary Fig. 7A and Supplementary Table 5), indicating that the deletion of Rep3 of ApcE might affect the assembly of PBSs or phycobiliprotein biosynthesis [19]. Considering the smaller size observed by TEM (Supplementary Fig. 8) and the structural model from Liu et al. [12], we interpret these results as being most consistent with the isolation of a simple, one-cylinder PBS.

Fig. 2.

Characterization of the PBSs of $\mathrm{\Delta }$(rep3), $\mathrm{\Delta }$(arm2(6–36)/rep3), $\mathrm{\Delta }$(arm2(6–17)/rep3), and $\mathrm{\Delta }$(arm2(29–36)/rep3) mutants from Synechocystis PCC 6803. Sucrose gradient ultracentrifugation (A), absorption (B,E) and fluorescence emission spectra at room temperature (C,F) and 77 K (D,G) of the PBSs of the four mutants. The samples (B–D) of $\mathrm{\Delta }$(rep3) (Zone1, black), $\mathrm{\Delta }$(arm2(6–36)/rep3) (Zone4, blue), $\mathrm{\Delta }$(arm2(6–17)/rep3) (Zone6, cyan), and $\mathrm{\Delta }$(arm2(29–36)/rep3) (Zone8, violet) were obtained from 0.25 M sucrose gradient. The samples (E–G) of $\mathrm{\Delta }$(rep3) (Zone2 and Zone3, solid and dash black), $\mathrm{\Delta }$(arm2(6–36)/rep3) (Zone5, blue), $\mathrm{\Delta }$(arm2(6–17)/rep3) (Zone7, cyan), and $\mathrm{\Delta }$(arm2(29–36)/rep3) (Zone9, violet) were obtained from 0.5–0.75 M or 0.75 M sucrose gradient. The absorption spectra were normalized at 620 nm, while the fluorescence emission spectra were normalized at their main peak position. All the fluorescence emission spectra were measured by excitation at 580 nm. The sucrose gradient ultracentrifugation of the $\mathrm{\Delta }$(arm2(18–29)/rep3) mutant was not carried out due to its liability to reverse mutation.

We next analyzed PBSs from the double mutant $\mathrm{\Delta }$(arm2(6–36)/rep3). Ultracentrifugation on a sucrose gradient (Fig. 2A) indicated that the PBSs of $\mathrm{\Delta }$(arm2(6–36)/rep3) showed a major fraction at 0.25 M and a minor fraction suspended at 0.5–0.75 M sucrose gradient; the two fractions were in 5.59:1 molar ratio (Supplementary Table 2). The major fraction had an absorbance of 620 nm and a fluorescence maxima at 658 nm at room temperature, while the 77 K fluorescence peak shifted to 666 nm (Fig. 2B–D). The minor fraction had the typical absorption and fluorescence maxima of CPC at 620 nm and 642 nm, respectively, and the 77 K fluorescence shifted to 654 nm (Fig. 2E–G). Importantly, the PBSs of $\mathrm{\Delta }$(arm2(6–17)/rep3) and $\mathrm{\Delta }$(arm2(29–36)/rep3) presented characteristics at the same sucrose gradient similar to those of $\mathrm{\Delta }$(arm2(6–36)/rep3) (Fig. 2). These results indicated that (1) a specific amount of the CPC rods was dismantled from the PBS cores in these mutants. These findings could question the integrity of PBSs in these mutants; (2) the major fraction contained a simple, partially dismantled PBS consisting of only CPC and APC in 1.66 $\pm$ 0.29 molar ratio, but the energy absorbed by CPC in this simple PBS was transferred to APC. Hence, it could be speculated that in the simple PBSs, one APC hexamer binds three CPC trimers, or two APC hexamers (disks) bind three CPC hexamers (disks); (3) for the three mutants, the truncation of the helix-loop-helix of 6–36 aa, the helix of 6–17, or the helix of 29–36 from Arm2 presented a similar phenotype of PBS, i.e., the helix-loop-helix, or one of the helix exerted a similar role in PBS assembly, consistent with our previous report [11].

According to the PBS structure from the red alga Griffithsia pacifica [8], the Arm2 element of helix-loop-helix is inserted between one normal APC disk and two APC disks, one of which has a L_CM terminal emitter and the other has AP-B terminal emitter in one basal cylinder (i.e., Cylinder1 or 2). When one helix or both helices of Arm2 were truncated, a simple PBS containing only APC/CPC was generated such that the other longitudinal half PBS with the APC disks harboring the terminal emitters (L_CM and ApcD) was lost. The cell spectra showed a high fluorescence of these mutants at 660 nm (Supplementary Fig. 5A–C) that disappeared completely at 683 nm (Supplementary Fig. 5D–F). Upon excitation at 430 nm and at 580 nm, the corresponding fluorescence confirmed that the energy transfer from APC to the terminal emitters of PBSs was disrupted [19]. Moreover, due to the absence of terminal emitters, the energy transfer from the PBSs to PSs was disrupted (Supplementary Fig. 5). The photosynthetic capacities (Supplementary Table 3) were much lower than those of $\mathrm{\Delta }$(rep3) and WT. Consequently, the mutants $\mathrm{\Delta }$(arm2/rep3), $\mathrm{\Delta }$(arm2(6–36)/rep3), $\mathrm{\Delta }$(arm2(6–17)/rep3), and $\mathrm{\Delta }$(arm2(29–36)/rep3) of the damaged Arm2 element grew poorly under photoautotrophic conditions but are capable of growing photomixotrophically in the culture media supplemented with 10 mM glucose (Supplementary Fig. 4). Supplementary Table 4 shows that when the chlorophyll content is maintained, the level of F₀ in these mutants is 2- to 5-fold higher than that of $\mathrm{\Delta }$(rep3), which is consistent with the high fluorescence intensity of these mutants (Supplementary Fig. 5). Reportedly, a specific number of uncoupled PBSs may be one of the reasons for increased F₀ in cyanobacteria [19, 36, 37]. The results suggested Arm2 as a putative candidate for PBSs binding thylakoid membranes [13]. Also, the generation of small PBSs containing only APC/CPC effectuated a weak association between PBSs and thylakoid membranes.

Therefore, the Arm2 element is responsible for the longitudinal assembly of PBSs. The loss of the element results in the dissociation of half of the PBS harboring the terminal emitters, while the other half, i.e., the simple PBS with only APC and CPC is retained in the cells. The cell spectra showed the absence of 77 K fluorescence at 684–685 nm corresponding to the terminal emitters (Supplementary Fig. 6, Supplementary Table 5). Hence, the dissociated parts of the PBSs could be degraded via specific phycobiliprotein degradation routes [38, 39]. One terminal emitter, ApcD, could undergo de-chromophorylation for its degradation [22]. The other terminal emitter, L_CM, degraded slowly, and hence, a little L_CM (the small peak at ~670 nm in Fig. 2 and Supplementary Fig. 6) was detected in the isolated PBSs from some mutants of the damaged Arm2 element.

To further investigate the roles of the motif of Arm2(6–36), we isolated the PBSs of $\mathrm{\Delta }$(arm2/rep3), $\mathrm{\Delta }$(rep2/rep3), and $\mathrm{\Delta }$(arm2/rep2/rep3). Their ultracentrifugation (Supplementary Fig. 9) showed a similar fraction pattern compared to $\mathrm{\Delta }$(arm2(6–36)/rep3), and the spectra of the fractions were also similar to those of $\mathrm{\Delta }$(arm2(6–36)/rep3). The coincidence between $\mathrm{\Delta }$(arm2/rep3) and $\mathrm{\Delta }$(arm2(6–36)/rep3) confirmed Arm2(6–36) as the key element of Arm2. According to the PBS structure of the red alga Griffithsia pacifica [8] and a PBS core conformation similar to that suggested by Liu et al. [12], the truncation of Rep2 and 3 dismantles the APC disk of Cylinder3 and basal cylinders and retains the APC disk with L_CM of basal cylinders. Although the spectra showed 77 K fluorescence at ~680 nm (Supplementary Fig. 9D), which would be L_CM in a trimer [40], the fluorescence degree was too small compared to that at 666 nm of APC. Thus, the energy absorbed by APC could only slightly be transferred to L_CM, indicating that (1) the simple PBSs having only APC and CPC were predominant and/or (2) the APC disks with and without L_CM were not well connected.

The ultracentrifugation of $\mathrm{\Delta }$(arm2(37–67)/rep3) generated a fraction pattern similar to that of $\mathrm{\Delta }$(arm2(6–36)/rep3). On the other hand, the minor fraction at 0.25 M sucrose gradient showed similar absorption and fluorescence at 680 nm fluorescence (Supplementary Fig. 10), the fraction suspended at the 0.5–0.75 M sucrose gradient showed maximal absorption at 620 nm, fluorescence at 650 nm at room temperature and at 656 nm at 77 K, which corresponded to CPC, as verified by SDS-PAGE (Supplementary Fig. 3J). The molar ratio of minor:major fraction was 1:1.57 (Supplementary Table 6), indicating that CPC rods mainly dissociated from the cores in $\mathrm{\Delta }$(arm2(37–67)/rep3) and that the 36–67-aa motif, although in random coils, might be responsible for the attachment of CPC rods at the cores (Supplementary Fig. 1).

Conversely, $\mathrm{\Delta }$(arm2(68–98)/rep3) and $\mathrm{\Delta }$(arm2(99–129)/rep3) showed an ultracentrifugation pattern different from the mutants of the damaged Arm2 but similar to that of $\mathrm{\Delta }$(rep3). Among the three fractions, the two minor fractions at 0.25 and 0.5 M sucrose gradient absorbed maximally at 654 nm and 620 nm with an obvious 654 nm peak that exhibited fluorescence at 661–672 nm at room temperature and at 683–684 nm at 77 K, respectively (Supplementary Fig. 10B–G). In addition, the major fraction at 0.75 M sucrose gradient had similar characteristic absorption and fluorescence at room temperature and 77 K fluorescence spectra as the PBSs from intact $\mathrm{\Delta }$(rep3) and WT (Supplementary Fig. 10H–J). Therefore, the energy absorbed by CPC could be transferred efficiently to the terminal emitters. The cell spectra also showed an efficient energy transfer in vivo from CPC to AP-B (Supplementary Fig. 6, Supplementary Table 5). Furthermore, due to the truncation, the fluorescence at 684–685 nm assigned to the terminal emitter, AP-B, decreased gradually according to the order of $\mathrm{\Delta }$(rep3), $\mathrm{\Delta }$(arm2(99–129)/rep3), and $\mathrm{\Delta }$(arm2(68–98)/rep3) (Supplementary Fig. 10J).

Cyanobacterial NPQ and state transitions affect the balance between energy transfer and dissipation within the PBSs and the distribution of excitation energy from the PBSs to the PSs [41]. Experimentally, states I and II are induced by blue LL and orange LL irradiation, respectively, such that the energy is preferentially transferred to photosystem I (PSI) and from PBS to PSII [42]. When state transitions occur at a low irradiation intensity, the high irradiation intensity quenches the non-photochemical PBSs, triggered by OCP [19, 43].

Pulse amplitude-modulated fluorometer (PAM) fluorescence technique is often used to study the energy transfer between PBSs and PSs. In PAM measurements, dark-adapted cells have a low dark maximal fluorescence (F_md), but when they are continuously irradiated by blue LL (55 µmol photons m²/s), the F_m^′ increases rapidly and arrives at a maximal F_m^′ fluorescence level (F_mb^′), indicating that the cells have undergone the transition to state I. Thus, the ratio of F_vb/F_vd (F_V = F_m – F₀) indicated the ability of mutants to transform state II$\to$state I. When the Blue LL light-adapted cells are irradiated with orange light (20 µmol photons m²s^-1), the decrease in F_m^′ indicates the ability of mutants to transform from state I$\to$state II [24, 44, 45]. Herein, we compared the state transition performance of WT and mutants: the ratio of F_vb/F_vd was consistent (approximately 15), and the F_m^′ decreased rapidly in WT and $\mathrm{\Delta }$(rep3) mutant (Supplementary Fig. 11). However, the ratio of F_vb/F_vd in $\mathrm{\Delta }$(arm2/rep3) and $\mathrm{\Delta }$(arm2(6–36)/rep3) mutants was negative (Fig. 3C,D), and that in the mutants without the Arm2(6–17) or Arm2(29–36) motif was lower than that in $\mathrm{\Delta }$(rep3) and WT (Fig. 3E,F), implying a strongly-weakened ability of these mutants to perform state II$\to$state I. This phenomenon was supported by the results that the 77 K fluorescence emission spectra of blue-adapted WT and $\mathrm{\Delta }$(rep3) cells had a high PSII fluorescence band with respect to dark-adapted cells, while the PSII fluorescence peak of $\mathrm{\Delta }$(arm2/rep3) and $\mathrm{\Delta }$(arm2(6–36)/rep3) mutants or $\mathrm{\Delta }$(arm2(6–17)/rep3) and $\mathrm{\Delta }$(arm2(29–36)/rep3) mutants showed only a slight increase (Fig. 4) [19, 36, 46]. On the contrary, the ratio of F_vb/F_vd in the other mutants of the damaged Arm2 was higher than that in $\mathrm{\Delta }$(rep3) (Supplementary Fig. 11D–F). Interestingly, under orange LL, the F_m^′ of all mutants of the damaged Arm2 did not decrease, indicating that the transition from transition I to state II cannot occur (Fig. 3 and Supplementary Fig. 11). Based on these results, we concluded that (1) damaging the helix-loop-helix element of Arm2 produced simple APC/CPC PBSs, effectuating a weak association between PBSs and the thylakoid membrane. This prevents the state I transition from state II and the energy transfer between the two PSs; (2) damaged Arm2 disrupted the connection between APC disk and APC disk with terminal emitters, such that the energy transfer from the PBSs to PSs was perturbed and the transition from state I to state II was impaired. The terminal emitters are responsible for transferring the absorbed energy to the reaction centres and plays a crucial part in energy migration and affects state transitions [47].

Fig. 3.

State transitions in WT, $\mathrm{\Delta }$(rep3), $\mathrm{\Delta }$(arm2/rep3), $\mathrm{\Delta }$(arm2(6–36)/rep3), $\mathrm{\Delta }$(arm2(6–17)/rep3), and $\mathrm{\Delta }$(arm2(29–36)/rep3) mutants from Synechocystis PCC 6803. The fluorescence changes were followed with a PAM fluorometer. Dark-adapted cells containing 3 µg/mL Chl in BG11 were stimulated with blue LL (55 µmol photons m^-2s^-1) and orange (609 nm) LL (20 µmol photons m^-2s^-1). Typical experiments are shown for Synechocystis WT (A), $\mathrm{\Delta }$(rep3) (B), $\mathrm{\Delta }$(arm2/rep3) (C), $\mathrm{\Delta }$(arm2(6–36)/rep3) (D), $\mathrm{\Delta }$(arm2(6–17)/rep3) (E), and $\mathrm{\Delta }$(arm2(29–36)/rep3) (F) mutant cells. The saturating pulses were separated by 30 s with 2000 µmol photons m^-2s^-1. Traces are from one representative experiment and are consistent with the results from two other independent biological replicates (n = 3).

Fig. 4.

State transitions in WT, $\mathrm{\Delta }$(rep3), $\mathrm{\Delta }$(arm2/rep3), $\mathrm{\Delta }$(arm2(6–36)/rep3), $\mathrm{\Delta }$(arm2(6–17)/rep3), and $\mathrm{\Delta }$(arm2(29–36)/rep3) mutants from Synechocystis PCC 6803. 77 K fluorescence emission spectra of dark (solid, State II) and blue-light (dash, state I, 55 µmol photons m^-2s^-1) adapted WT (dark yellow, A), $\mathrm{\Delta }$(rep3) (black, B), $\mathrm{\Delta }$(arm2/rep3) (red, C), $\mathrm{\Delta }$(arm2(6–36)/rep3) (blue, D), $\mathrm{\Delta }$(arm2(6–17)/rep3) (cyan, E), and $\mathrm{\Delta }$(arm2(29–36)/rep3) (violet, F) cells. Normalization was carried out at 720 nm. The measurements were recorded at a chlorophyll concentration of 3 µg/mL. All fluorescence emission spectra were measured by excitation at 580 nm. Spectra are from one representative experiment and are consistent with the results from two other independent biological replicates (n = 3). APC, allophycocyanin; PSII, photosystem II; PSI, photosystem I; PC, phycocyanin.

Next, the NPQ performance of WT and mutants was compared (Fig. 5). The results showed strong blue light-induced PBS fluorescence quenching (about 14%) in WT, while that in $\mathrm{\Delta }$(arm2/rep3) or $\mathrm{\Delta }$(rep3) mutants was low (about 0.69% or 4.96%) (Fig. 5G). Interestingly, the mutants of the damaged helix-loop-helix element of Arm2 showed an NPQ performance similar to the WT (Fig. 5A–F), while the remaining mutants showed a phenotype similar to $\mathrm{\Delta }$(rep3) mutants (Supplementary Fig. 12C–E). When dark-adapted cells were exposed to white high light (HL) directly, Fluorescence quenching was reversible in all strains in the darkness (Fig. 5H and Supplementary Fig. 12F). Importantly, WT and $\mathrm{\Delta }$(arm2/rep3) recovered from NPQ a similar rate but 3-fold faster than $\mathrm{\Delta }$(rep3), while $\mathrm{\Delta }$(arm2(6–36)/rep3), $\mathrm{\Delta }$(arm2(6–17)/rep3), and $\mathrm{\Delta }$(arm2(29–36)/rep3) recovered faster than $\mathrm{\Delta }$(rep3). These results support the following findings: (1) the damaged helix-loop-helix element of Arm2 produced small APC/CPC PBSs, which had the more packed conformation increased the probability of energy transfer between different rods [47] and might allow OCP to approach the PBSs closely in some directions, such that OCP-dependent NPQ is significant; (2) the integrity of PBSs is crucial for strong binding of activated OCP. The activated OCP may be loosely bound by the small PBSs and easily released from them compared to WT PBSs.

Fig. 5.

Changes in fluorescence levels induced by different intensities of blue light and dark recovery of quenched fluorescence measured in a PAM fluorometer. Dark-adapted Synechocystis WT cells (dark yellow, A), $\mathrm{\Delta }$(rep3) (black, B), $\mathrm{\Delta }$(arm2/rep3) (red, C), $\mathrm{\Delta }$(arm2(6–36)/rep3) (blue, D), $\mathrm{\Delta }$(arm2(6–17)/rep3) (cyan, E) and $\mathrm{\Delta }$(arm2(29–36)/rep3) (violet, F) mutant cells were illuminated with strong blue light intensities (460 nm) at 127 (squares), 307 (circle), and 503 (triangle) µmol photons m^-2s^-1 to induce the quenched state. (G) Dark-adapted Synechocystis WT cells (dark yellow), $\mathrm{\Delta }$(rep3) mutant cells (black), and $\mathrm{\Delta }$(arm2/rep3) mutant cells (red) were illuminated successively with a low-intensity blue light (460 nm, 55 µmol photons m^-2s^-1) and a high-intensity blue light (460 nm, 503 µmol photons m^-2s^-1). (H) Dark-adapted cells of six mutants shown above containing 3 µg/mL Chl were illuminated directly with white HL (20 µmol photons m^-2s^-1) for 5 min and then kept in the darkness. Saturating pulses were separated by 30-s intervals with 2000 µmol photons m^-2s^-1. All the experiments were carried out in the presence of chloramphenicol. Traces are from one representative experiment and are consistent with the results from two other independent biological replicates (n = 3).