Proton motive force in plant photosynthesis dominated by ΔpH in both low and high light

Abstract

The proton motive force (pmf) across the thylakoid membrane couples photosynthetic electron transport and ATP synthesis. In recent years, the electrochromic carotenoid and chlorophyll absorption band shift (ECS), peaking ∼515 nm, has become a widely used probe to measure pmf in leaves. However, the use of this technique to calculate the parsing of the pmf between the proton gradient (ΔpH) and electric potential (Δψ) components remains controversial. Interpretation of the ECS signal is complicated by overlapping absorption changes associated with violaxanthin de-epoxidation to zeaxanthin (ΔA505) and energy-dependent nonphotochemical quenching (qE; ΔA535). In this study, we used Arabidopsis (Arabidopsis thaliana) plants with altered xanthophyll cycle activity and photosystem II subunit S (PsbS) content to disentangle these overlapping contributions. In plants where overlap among ΔA505, ΔA535, and ECS is diminished, such as npq4 (lacking ΔA535) and npq1npq4 (also lacking ΔA505), the parsing method implies the Δψ contribution is virtually absent and pmf is solely composed of ΔpH. Conversely, in plants where ΔA535 and ECS overlap is enhanced, such as L17 (a PsbS overexpressor) and npq1 (where ΔA535 is blue-shifted to 525 nm) the parsing method implies a dominant contribution of Δψ to the total pmf. These results demonstrate the vast majority of the pmf attributed by the ECS parsing method to Δψ is caused by ΔA505 and ΔA535 overlap, confirming pmf is dominated by ΔpH following the first 60 s of continuous illumination under both low and high light conditions. Further implications of these findings for the regulation of photosynthesis are discussed.

Electrochromic shift absorption kinetics show the steady-state transthylakoid proton motive force in plants is dominated by the proton concentration gradient under both low and high light conditions.

Introduction

Photosynthesis relies upon many interconnected bioenergetic and biochemical processes. Within the chloroplast thylakoid membrane, light energy is used to drive charge separation in the photosynthetic reaction centers, photosystem I (PSI) and photosystem II (PSII). These photochemical reactions, and the subsequent operation of the Q-cycle within cytochrome b₆f (cytb₆f), result in the movement of electrons and protons across the span of the thylakoid membrane bilayer, generating an electrical potential (Δψ) and a chemical gradient of protons (ΔpH) (Kramer et al., 2003; Malone et al., 2021). This electrochemical gradient is known as the proton motive force (pmf) and is utilized by the thylakoid ATP synthase to drive the endergonic synthesis of ATP in the chloroplast stroma (Nelson and Junge, 2015). According to Mitchell’s chemiosmotic theory, Δψ and ΔpH are thermodynamically and kinetically equivalent components of the pmf (Mitchell, 1961; Hangarter and Good, 1982) that can be expressed as follows:

$$\mathit{pmf}=\Delta {\psi }_{i-o}-\frac{2.3\text{RT}}{\mathrm{F}}\cdot \Delta {\text{pH}}_{o-i}$$

where Δψ_i−o is the electrical gradient (lumen-minus-stroma), R is the ideal gas constant, T is the temperature, F is the Faraday constant, and ΔpH_o−i is the proton gradient (stroma-minus-lumen).

In addition to its central role in cellular energy conservation, the Δψ and ΔpH components of the pmf also play important roles in the regulation of photosynthesis (Armbruster et al., 2017). Increased ΔpH acts as the trigger for the major rapidly reversible component of nonphotochemical quenching (known as “qE”), which protects PSII from photooxidative damage (Ruban and Wilson, 2020) and for “photosynthetic control”, which protects PSI from overreduction in excess light by regulating the rate of plastoquinol (PQH₂) oxidation at the cytb₆f complex (Suorsa et al., 2013). Increased Δψ, in contrast, has been shown to cause photodamage in thylakoids by promoting charge recombination between the primary and secondary radical pairs in the PSII RC, chlorophyll triplet formation, and thus generation of singlet oxygen (Bennoun, 1994; Davis et al., 2016, 2017). Consistent with these contrasting effects, a wide range of experimental approaches, including microelectrodes, pH sensitive dyes, and radiolabeling, concluded that the vast majority of pmf in chloroplasts is stored as ΔpH due to rapid compensatory counterion movements that dissipate Δψ (Dilley and Vernon, 1965; Bulychev et al., 1972; Rottenberg et al., 1972; Schuldiner et al., 1972; Barber et al., 1974; Pick et al., 1974; Chow et al., 1976; Vredenberg and Bulychev, 1976; Slovacek and Hind, 1981; Bulychev, 1984; Van Kooten et al., 1986; Remiš et al., 1986; Vredenberg, 1997). The preference of chloroplasts for ΔpH was in contrast to the situation in mitochondria where pmf is stored mainly as Δψ, due to the low ion permeability of the mitochondrial inner membrane, with a ΔpH contribution of only ∼0.5 U, ∼25% of the total mitochondrial pmf (Mitchell, 1961, 2011; Lambert and Brand, 2004; Wolf et al., 2019). These differences were rationalized on the basis that since mitochondria utilize chemical reductants, such as NADH and succinate, and consume oxygen through respiration, avoiding charge recombination is unnecessary.

The consensus view that the steady-state transthylakoid pmf consists primarily of ΔpH, built largely on work with isolated chloroplasts, was later challenged by the emergence of the electrochromic shift (ECS) signal as an in vivo probe of the pmf in intact leaves (Kramer and Sacksteder, 1998; Cruz et al., 2001; Kramer et al., 2003). The Δψ component induces an ECS in the Soret peak absorption of chlorophylls and carotenoids in the thylakoid membrane (Witt, 1971, 1979; Vredenberg, 1997; Bailleul et al., 2010). This results in the formation of a transient absorption peak ∼515 nm upon illumination of leaves (Witt, 1971). Since a significant proportion of the ECS signal persisted during continuous illumination, Kramer et al. suggested that in vivo, a larger fraction of pmf is stored as Δψ than was suggested by the earlier in vitro work (Kramer and Sacksteder, 1998; Cruz et al., 2001; Kramer et al., 2003). Interestingly, they found that the parsing of the pmf between the Δψ and ΔpH, implied by the ECS measurements, was affected by light intensity and CO₂ availability (Kanazawa and Kramer, 2002; Takizawa et al., 2007). More recently, the generation of mutants deficient in thylakoid-associated ion channels involved in counterion movements, such as the Cl⁻ channel VCCN1 and the H⁺/K⁺ antiporter KEA3, have highlighted the crucial importance of pmf composition to plant fitness (Carraretto et al., 2013; Armbruster et al., 2014; Duan et al., 2016; Herdean et al., 2016a, 2016b).

However, while the ECS signal has proven itself a useful probe of the pmf amplitude, proton flux, and conductivity in leaves, its suitability for probing pmf parsing has been questioned (Johnson and Ruban, 2014). The complicating issue is the congested nature of the spectral region where the ECS absorption changes are observed. Overlapping light-driven absorption changes include those due to the de-epoxidation of violaxanthin to zeaxanthin, which produces a large positive band at ∼505 nm, hereafter ΔA505 (Yamamoto et al., 1971; Bilger et al., 1989: Ruban et al., 1993) and the qE-related absorption changes approximately 525–540 nm, often called ΔA535 (Bilger and Björkman, 1990). Of these, the qE-related absorption changes are the most problematic since they form and relax relatively rapidly and are thus more difficult to distinguish from the ECS signal. Whilst the cyt f redox changes occur on similarly rapid timescales, the related absorption peak is narrow and centered at ∼554 nm, with little-to-no overlap with the ECS, Δ505, or Δ535 (Nishio and Whitmarsh, 1993; Metzger et al., 1997). Initially attributed to light scattering changes caused by altered thylakoid structure provoked by the ΔpH formation (Murakami and Packer, 1970b; Murakami and Packer, 1970a; Duniec and Thorne, 1977), they were later shown by resonance Raman spectroscopy to reflect an absorption change in a sub-population of zeaxanthin, which required the presence of photosystem II subunit S (PsbS; Ruban et al., 2002). Theoretical work later showed that ΔA535 may arise from zeaxanthin J-dimers formed at the interface of aggregated LHCII proteins in the qE state (Duffy et al., 2010). Interestingly, when zeaxanthin formation is blocked by inhibitors or through the absence of violaxanthin de-epoxidase, the qE-related absorption peak shifts from 535 to 520–525 nm, increasing its overlap with the ECS signal (Crouchman et al., 2006; Johnson et al., 2009). These observations led Johnson and Ruban (2014) to use the ECS method to assess the parsing of pmf in the lut2npq1 mutant of Arabidopsis (Arabidopsis thaliana), which fails to synthesize zeaxanthin and is deficient in qE, removing much of the signal contamination from the ECS absorbance window. The data demonstrated that components of the ECS signal could be separated by their differing temporal, Δψ, and ΔpH dependence. In wild-type (WT) leaves, the 515 nm signal shows a sharp rise as illumination commences before decaying to <50% of its initial amplitude over the next 30 s, this was then followed by a slower secondary rise which stabilized at approximately 60%–70% of the initial amplitude, and according to the ECS parsing method, is attributed to steady-state Δψ. The secondary rise of the ECS signal was completely absent in lut2npq1 and could be eliminated using the H⁺/K⁺ antiporter nigericin, which collapses ΔpH. These observations led Johnson and Ruban (2014) to propose that the steady-state Δψ in the WT was caused by the overlapping qE-related absorption change.

In the following study, we widened our investigation into the origin of the steady-state ECS signal to include a range of Arabidopsis plants with altered xanthophyll cycle and PsbS content. Unlike our previous measurements, these data were obtained on the widely used Walz Dual-PAM device with the P515 emitter/detector modules (Klughammer et al., 2013). The results support the original view in the literature that the steady-state Δψ contribution to the pmf in vivo is negligible (< 10%), and that the secondary rise in the ECS signal reflects the contribution of the overlapping qE-related absorption changes.

Results

Characterization of the ECS, xanthophyll cycle, and qE-related signals in WT Arabidopsis leaves

According to the ECS parsing method (Kramer et al., 2003) the light-to-dark transients of the 550–515 nm absorption difference signal provide information on the relative contributions of the Δψ and ΔpH to the pmf. However, this section of the absorption spectrum is heavily congested with light-induced absorption changes. Figure 1A shows a selection of such changes. De-epoxidation of violaxanthin to zeaxanthin causes the appearance of a large positive band at ∼505 nm, whilst the PsbS-dependent red-shift of a sub-population of zeaxanthin during qE causes an absorption peak at ∼535 nm (Ruban et al., 2002; Johnson et al., 2009; Johnson and Ruban, 2009). The ECS-related peak is formed within microseconds of illumination and has its peak at ∼515 nm. The qE-related peak forms in seconds to minutes depending on the pre-illumination history of the leaf and can vary in magnitude and position, according to the xanthophyll content of the leaf, as shown in Figure 1B (Johnson et al., 2009). While the WT peak appears at 534 nm, in the absence of zeaxanthin in the npq1 mutant, this peak becomes blue-shifted, here shown to be at 523 nm. In the npq2 mutant, where zeaxanthin is constitutively present, the qE peak becomes red-shifted relative to WT, with its peak appearing at 538 nm. The PsbS-overexpressor, L17, possesses a much greater qE response, and this is reflected in the larger magnitude of its qE-related peak at 532 nm.

Figure 1.

Difference spectra and the light-to-dark transition in the ECS signal. A, Difference spectra of qE (black) 5 min light-minus-5 min dark recovery; zeaxanthin synthesis (red) dark adaptation-minus-5 min light; ECS (blue) 15 s light-minus-5 min dark recovery. B, qE spectra (5 min light-minus-5 min dark recovery) for a range of Arabidopsis transformants. WT (black); npq1 (blue); npq2 (red); L17 (gray). C, ECS kinetic signal (ΔA 550–515 nm) measured on a WT Arabidopsis leaf. The magnitude of the trough formed upon cessation of illumination is the ECS_t. The difference between the minima of the ECS_t and the steady-state signal in the dark is termed the ECS_inv. Thus, the difference between ECS_t and ECS_inv is termed the ECS_t-inv. AL, actinic light.

Figure 1C shows an expanded and annotated view of a light-to-dark transition in the ECS signal. After the cessation of illumination, an initial sharp trough forms, which slowly relaxes (∼30 s) to a pseudo-baseline in the dark. The total amplitude of the initial trough has been assumed to be proportional to the total pmf, as here is termed ECS_t (Klughammer et al., 2013). In WT leaves, the post relaxation pseudo-baseline is at a level between the maximal light signal and the minima of the ECS_t. According to the ECS parsing method, the difference between the pseudo-baseline and the ECS_t will be representative of the total ΔpH and is hereafter termed ECS_inv, where the subscript denotes a transient inverse Δψ generated when the continuing efflux of protons through the ATP is not rapidly matched by the movement of other ions upon cessation of illumination (Cruz et al., 2001; Kramer et al., 2003). Finally, the difference between the pseudo-dark baseline and the steady-state level of the ECS signal just prior to the cessation of illumination is attributed to the Δψ and is termed ECS_t-inv.

To investigate this further, WT Arabidopsis leaves were exposed to eight steps of 3-min illumination followed by 3 min of darkness at intensities of 71, 151, 308, 417, 548, 708, 1,128, 1,396 µmol photons m⁻² s⁻¹. Here, the ECS (ΔA 550–515 nm) and ΔA535 signals can be measured in parallel, as previously described (Klughammer et al., 2013). Figure 2A shows a representative ECS kinetic trace of the light titration. Under continuous light flux below the growth light intensity (i.e. < 190 µmol photons m⁻² s⁻¹; the first two steps), the steady-state ECS signal rises to a maximum level after ∼60 s, before relaxing to a level above the subsequent pseudo-dark baseline (Figure 2A;  Supplemental Figure S1). This kinetic behavior is also observed in the ΔA535 signal, shown in Figure 2B. During these initial two light stages, the ECS_t reaches a level up to ∼50% of its maxima, whilst the ECS_inv and ECS_t-inv remain in approximately a 1:1 stoichiometry. At 71 µmol photons m⁻² s⁻¹, the ECS_inv accounts for 67 ± 13% of the total ECS_t, while the ECS_t-inv accounts for 33 ± 31% (P > 0.05, Student’s t test). At 158 µmol photons m⁻² s⁻¹, the ECS_inv accounts for 61 ± 7% of the total ECS_t, while the ECS_t-inv accounts for 39 ± 8% (P > 0.05, Student’s t test). According to the ECS parsing method, this would imply an approximately equal partitioning of the pmf between ΔpH and Δψ. At 308 µmol photons m⁻² s⁻¹ and above, the ΔA535 signal ceases to relax in the light phase, as does the steady-state ECS in the light, which shows a stark upward rise in the light. Between 308 and 548 µmol photons m⁻² s⁻¹, the ECS_t also reaches its maxima. Again, here the ECS_inv and ECS_t-inv components remain at similar levels (each ∼50% of the maximum ECS_t) with no significant differences between the two (P > 0.05; Student’s t test). At light intensities of 708 µmol photons m⁻² s⁻¹ and higher, the ECS_t starts to decrease, with the minimum under high light being achieved at 1,396 µmol photons m⁻² s⁻¹, with an ECS_t 84.9 ± 1.7% of the maximum. Furthermore, as the light intensity increases, the ΔA535 signal continues to rise to a maximum level, 42.9% higher at 1,396 µmol photons m⁻² s⁻¹ than at 308 µmol photons m⁻² s⁻¹. Under high light, the ECS_inv proportion continues to rise with respect to the ECS_t-inv, as shown in Figure 2C. However, it is worth noting that even under 1,396 µmol photons m⁻² s⁻¹, the ECS_t-inv is still 29.3 ± 5% of the ECS_t, implying a substantial Δψ contribution to pmf even under high light in the WT.

Figure 2.

ECS and 535 nm measurements on WT leaves. A, Representative ECS (ΔA 550–515 nm) and (B) ΔA 535 nm kinetics. Each assay consisted of eight steps of 3-min illumination (white bars) and 3-min darkness (black bars). The illumination increased at each step using 0, 71, 151, 308, 417, 548, 708, 1,128, and 1,396 µmol photons m⁻² s⁻¹. C, ECS_t, ECS_inv, and ECS_t-inv at each light intensity. Here, measurements are normalized to the maximum ECS_t. Error bars represent standard error of the mean (n = 7).

It is worth noting the overall “signal drift” of the ECS kinetics recorded on the WT leaves, with the overall ECS signal rising to a maximum at around 548 µmol photons m⁻² s⁻¹, approximately half-way through the assay. This has been proposed to be due to the o verlap of the relatively slowly forming ΔA505 signal (half time 6–8 min) with the ECS (Johnson et al., 2009; Klughammer et al., 2013; Wilson and Ruban, 2020).

Disentangling the impact of xanthophyll cycle activity on the ECS signal changes

To disentangle the impact of the xanthophyll cycle on the ECS signal, npq1, a mutant lacking violaxanthin de-epoxidase activity was measured. This mutant is unable to synthesize zeaxanthin, and lacks the corresponding ΔA505 absorption increase (Niyogi et al., 1998; Johnson et al., 2009). Interestingly, the ECS and ΔA535 signals for npq1 show sharp differences with respect to the WT (Figure 3, A and B). First, the ECS signal contains no general upward signal drift, confirming this feature is related to the ΔA505 change. Furthermore, ΔA535 absorption change in npq1 is greatly diminished, consistent with the fact that in the absence of zeaxanthin the qE-related absorption changes are smaller and now peak at 525 nm (Figure 1B; Johnson et al., 2009; Ilioaia et al., 2011; Johnson and Ruban, 2014). Under light intensities lower than the growth intensity (<190 µmol photons m⁻² s⁻¹), there is an initial sharp rise in the ECS signal, which, after ∼60 s, decays to a level slightly above the following dark pseudo-baseline, similar to WT (Figure 3A;  Supplemental Figure S1). However, in npq1, there is little to no upward drift in the light or downward signal drift in the dark away from the baseline. The similar levels and kinetics of the ΔA535 signal between WT and npq1 at these low light intensities, particularly at 71 µmol photons m⁻² s⁻¹, suggests that signal drift is therefore likely associated with zeaxanthin synthesis, and not the wavelength of the qE-related peak. At more moderate light intensities (308–548 µmol photons m⁻² s⁻¹), the ECS_t again reaches its maximum. However, the balance between ECS_inv and ECS_t-inv differs from the observed behavior in WT leaves. After illumination at 548 µmol photons m⁻² s⁻¹, the ECS_t-inv is ∼50% higher than in WT leaves (P < 0.01). Interestingly, there is a maintained offset of the npq1 ECS_t-inv of about 50% throughout the rest of the light titration, relative to the WT ECS_t-inv. According to the ECS parsing method, this would imply that in the absence of zeaxanthin, Δψ becomes the dominant component of the pmf under light intensities ≥308 µmol photons m⁻² s⁻¹ in npq1. Alternatively, the blue shift of the qE-related peak to 525 nm and the lack of a ΔA505 absorption change is responsible for the skewing of the ECS_inv and ECS_t-inv kinetics relative to WT. It is interesting to note in npq1 that the ECS_t follows a nearly identical relationship to light intensity as in the WT, as shown in Figure 3C. This is in agreement with studies showing that absence of violaxanthin de-epoxidation in npq1 has no effect on the total pmf amplitude or ΔpH relative to the WT (Crouchman et al., 2006; Johnson et al., 2012).

Figure 3.

ECS and 535 nm measurements on npq1 leaves. A, Representative ECS (ΔA 550–515 nm) and (B) ΔA 535 nm kinetics. Each assay consisted of eight steps of 3-min illumination (white bars) and 3 min darkness (black bars). The illumination increased at each step using 0, 71, 151, 308, 417, 548, 708, 1,128, and 1,396 µmol photons m⁻² s⁻¹. C, ECS_t, ECS_inv, and ECS_t-inv at each light intensity. Here, measurements are normalized to the maximum ECS_t. Error bars represent standard error of the mean (n = 8).

We next examined the npq2 mutant lacking the zeaxanthin epoxidase. Since npq2 constitutively accumulates zeaxanthin during development in place of violaxanthin, it lacks the light-induced ΔA505 (Niyogi et al., 1998; Pérez-Bueno et al., 2008; Johnson et al., 2009). Consistent with this, the baseline of the ECS signal shows no upward drift during illumination cycles as seen in the WT (Figure 4A). The qE-related ΔA535 remains in this mutant (Figure 4B), though it is red-shifted, peaking at 540 nm (Figure 1B; Johnson et al., 2009). If amplitude of the ECS_t-inv signal is influenced by the degree of overlap with the qE-related absorption change then it should be affected in this mutant. In Figure 4C, this effect is observed. At light intensities up to 151 µmol photons m⁻² s⁻¹, the ECS signal forms a larger ECS_t-inv component than the WT (Figure 4A)_. Again similar to WT, the partitioning between the ECS_inv and ECS_t-inv is approximately 1:1 at 308 µmol photons m⁻² s⁻¹ (P > 0.05). Interestingly, at the top three light intensities used (708, 1,128, and 1,396 µmol photons m⁻² s⁻¹), the ECS_inv signal rises to a level where it now exceeds the ECS_t (Figure 4, A and C). This effect can be explained by increased positive contribution of the qE-related absorption change (Figure 1B; ∼540 nm in npq2) to the 550 nm signal that is used for calculation of the ECS signal (ΔA 550–515 nm). The result is that using the ECS parsing method, at light intensities ≥417 µmol photons m⁻² s⁻¹, virtually all pmf in npq2 is present as ΔpH (Figure 4C).

Figure 4.

ECS and 535 nm measurements on npq2 leaves. A, Representative ECS (ΔA 550–515 nm) and B, ΔA 535 nm kinetics. Each assay consisted of eight steps of 3-min illumination (white bars) and 3-min darkness (black bars). The illumination increased at each step using 0, 71, 151, 308, 417, 548, 708, 1,128, and 1,396 µmol photons m⁻² s⁻¹. C, ECS_t, ECS_inv, and ECS_t-inv at each light intensity. Here, measurements are normalized to the maximum ECS_t. Error bars represent standard error of the mean (n = 6).

PsbS-mediated modulation of qE and its effect on the ECS signal

The amplitude and kinetics of qE also depend on the levels of the PsbS protein, which interacts with LHCII altering its ΔpH sensitivity by promoting its aggregation (Li et al., 2002; Crouchman et al., 2006; Johnson and Ruban, 2011; Sacharz et al., 2017). The npq4 mutant, which lacks PsbS, still displays the ΔA505 associated with zeaxanthin synthesis but lacks qE (Horton et al., 2000; Li et al., 2000). In line with this, we find the slow rise of the baseline of the ECS signal during illumination is still present in npq4 (Figure 5A), though the ΔA535 is greatly diminished at all light intensities measured (Figure 5B). In line with the virtual absence of the ΔA535 signal, the ECS_t-inv in npq4 is smaller at light intensities ≥308 µmol photons m⁻² s⁻¹ compared to the WT (Figure 5A). According to the ECS parsing method at 417 µmol photons m⁻² s⁻¹, 73.61 ± 5% of the total maximum ECS_t is present as ΔpH (ECS_inv) in npq4, versus just 50.43 ± 4% in the WT plants (P < 0.01). Indeed, at the maximum light intensity tested here (1,396 µmol photons m⁻² s⁻¹), the ECS_inv reaches 94.13 ± 6% of the ECS_t in npq4, compared to 70.67 ± 3% in the WT (P < 0.01).

Figure 5.

ECS and 535 nm measurements on npq4 leaves. A, Representative ECS (ΔA 550–515 nm) and (B) ΔA 535 nm kinetics. Each assay consisted of eight steps of 3-min illumination (white bars) and 3-min darkness (black bars). The illumination increased at each step using 0, 71, 151, 308, 417, 548, 708, 1,128, and 1,396 µmol photons m⁻² s⁻¹. C, ECS_t, ECS_inv, and ECS_t-inv at each light intensity. Here, measurements are normalized to the maximum ECS_t. Error bars represent standard error of the mean (n = 8).

To further test our hypothesis that the ECS signal is polluted by the qE-related ΔA535, we investigated the PsbS-overexpressor plants, L17, which show a two-fold larger qE-response compared to the WT (Li et al., 2002; Crouchman et al., 2006). Increased qE in L17 should result in a larger ΔA535 signal and a corresponding increase in the extent of the overlap with the ECS signal. Consistent with this, ECS and ΔA535 kinetics in L17 display stark differences compared to the WT (Figure 6, A and B). In L17, the ΔA535 signal is ∼2.15 times that of the WT and ∼10 times that of npq4 at 1,396 µmol photons m⁻² s⁻¹. The larger amplitude of the ΔA535 signal in L17 results in a much larger overlap with the ECS signal and therefore a much larger ECS_t-inv signal persists at the highest light intensities used compared to the WT (Figure 6, A and C). Therefore, according to the ECS parsing method, Δψ (ECS_t-inv) in L17 comprises 77.0 ± 0.04% of the total pmf at 1,396 µmol photons m⁻² s⁻¹ (Figure 6C), the reverse of the situation described above for npq4 (Figure 5C).

Figure 6.

ECS and 535 nm measurements on L17 leaves. A, Representative ECS (ΔA 550–515 nm) and (B) ΔA 535 nm kinetics. Each assay consisted of eight steps of 3-min illumination (white bars) and 3-min darkness (black bars). The illumination increased at each step using 0, 71, 151, 308, 417, 548, 708, 1,128, and 1,396 µmol photons m⁻² s⁻¹. (C) ECS_t, ECS_inv, and ECS_t-inv at each light intensity. Here, measurements are normalized to the maximum ECS_t. Error bars represent standard error of the mean (n = 8).

The nature of the ECS signal in the absence of PsbS and zeaxanthin

While the ECS_t-inv signal in npq4 is lower than that observed in the WT under moderate and high illumination (≥308 µmol photons m⁻² s⁻¹; Figure 5C), it is still not completely absent. One possibility is that the remaining signal reflects the gradual rise of the ECS baseline due to the ΔA505 associated with zeaxanthin synthesis. To test this idea further we investigated the npq1npq4 double mutant that lacks both zeaxanthin synthesis and qE (Li et al., 2000). In the npq1npq4 mutant, the ECS and ΔA535 kinetics (Figure 7, A and B) display very different behavior to WT leaves. Up to 417 µmol photons m⁻² s⁻¹, there appears to be a small contribution (<10%) of ECS_t-inv to the total ECS_t, which may reflect the true contribution of Δψ to the total pmf, free from overlapping ΔA505 and ΔA535 signals. However, as the ECS_t signal reaches its maxima at 308 µmol photons m⁻² s⁻¹, the ECS_inv accounts for 89.7 ± 8.3% of the total pmf rising to 100% above this intensity. These ECS kinetics closely match those previously reported for the lut2npq1 mutant that also lacks both ΔA505 and ΔA535 (Johnson and Ruban, 2014). Therefore, in the absence of these overlapping signals the ECS parsing method would conclude that the majority of the steady-state pmf comprised ΔpH at all light intensities measured and that any Δψ contribution to the pmf is dissipated within 60 s.

Figure 7.

ECS and 535 nm measurements on npq1npq4 leaves. A, Representative ECS (ΔA 550–515 nm) and (B) ΔA 535 nm kinetics. Each assay consisted of eight steps of 3-min illumination (white bars) and 3-min darkness (black bars). The illumination increased at each step using 0, 71, 151, 308, 417, 548, 708, 1,128, and 1,396 µmol photons m⁻² s⁻¹. C, ECS_t, ECS_inv, and ECS_t-inv at each light intensity. Here, measurements are normalized to the maximum ECS_t. Error bars represent standard error of the mean (n = 5).

Discussion

Overlap, origin of absorption changes

Over the past 20 years, much work has been focused on trying to measure the amplitude, kinetics, and parsing of the pmf noninvasively in leaves using the ECS signal peaking at 515 nm (Kramer et al., 1999). This method has been widely adopted by the photosynthesis research community and has provided a useful tool for comparing the pmf phenotypes of a wide-range of photosynthetic mutants. Nonetheless from the inception of the ECS method it was recognised that the overlapping absorption changes associated with qE (ΔA535) and zeaxanthin formation (ΔA505) could influence the ECS signal. Since these absorption changes form and relax relatively slowly (ΔA505, minutes timescale; ΔA535, seconds to minutes timescale), they will have relatively little effect on parameters calculated from the rapid (ms) light-to-dark transition changes in the ECS signal (e.g. proton conductivity (gH⁺), proton flux (vH⁺), and total pmf (ECS_t)). Indeed, the constancy of the total pmf across the different mutants used in this study supports this view. However, it is worth noting that rapid fluxes of ions on ms timescale may affect the ECS signal and cause a potential underestimation of the total pmf. In contrast, the parsing of ECS is calculated based on the relatively slower ECS signal changes occurring in the time following the first 60 s illumination or in the 60 s following cessation of illumination where clearly the overlap presents more of an issue. Early work attributed the ΔA535 to selective light scattering, and thus Kramer et al. attempted to remove its contribution through pre-scattering the incident light (Kramer and Sacksteder, 1998; Cruz et al., 2001). However, later work using resonance Raman spectroscopy showed that the ΔA535 arose from a genuine absorption change and hence this approach fails (Ruban et al., 2002; Duffy et al., 2010; Ilioaia et al., 2011). Johnson and Ruban (2014) highlighted the potential extent of this overlap issue by showing that any ECS_t-inv signal is missing from the lut2npq1 mutant that lacks the ΔA505 and ΔA535 changes. Since lut2npq1chloroplasts showed identical quenching of 9-aminoacridine (9AA) fluorescence compared to the WT, this suggested that pmf is entirely composed of ΔpH and that the ECS_t-inv signal arises from the overlap with the ΔA505 and ΔA535 in the WT. This idea was further corroborated by the fact that the ECS_t-inv signal in the WT could be abolished with an uncoupler. Nevertheless, perhaps because the Johnson and Ruban (2014) study was carried out using a mutant with quite divergent carotenoid composition compared to the WT, and only at a single high light intensity (700 µmol photons m⁻² s⁻¹) where the contribution of the Δψ to the pmf has been argued to be small (Klughammer et al., 2013), this work has been largely overlooked and the ECS parsing method has remained in widespread use.

In the current study, we lay bare the full extent of the overlap issue across the full range of light intensities from low (72 µmol photons m⁻² s⁻¹) to high (1,396 µmol photons m⁻² s⁻¹) using a wide range of Arabidopsis plants with altered ΔA535, ΔA505, or both absorption changes. From our data, it is apparent that the ECS_t-inv signal corresponds closely with the extent of ΔA535. As more and more zeaxanthin is synthesized as the light intensity increases, the qE-related signal shifts from 523 nm toward 535 nm and hence the extent of the overlap with the ECS signal is reduced (Figure 1B; Johnson and Ruban, 2014). Consistent with this in npq1, which lacks zeaxanthin, the ECS_t-inv remains large under high light intensities unlike in the WT since the qE-related signal remains “stuck” at 523 nm (Figure 1B; Johnson et al., 2009). A similar situation is seen in the PsbS overexpressor L17, where zeaxanthin is present, but the greatly increased amplitude of the ΔA535 absorption change increases the extent of overlap with the ECS, resulting in a large ECS_t-inv contribution, particularly at high light. Application of the ECS parsing method to these mutants would suggest a greatly enhanced Δψ and diminished ΔpH contribution to the total pmf. If true, such a large Δψ would lead to significant photoinhibition of PSII through promotion of charge recombination (Bennoun, 1994; Davis et al., 2016, 2017). Moreover, the extremely small ΔpH would preclude the formation of the large qE observed in L17 and the normal photosynthetic control observed in both (Roach and Krieger-Liszkay, 2012; Tikkanen et al., 2015). Indeed, previous studies have shown that the level of 9AA quenching and so ΔpH in isolated chloroplasts of L17 is unchanged compared to the WT (Crouchman et al., 2006). Likewise, the results from the npq4 mutant highlight that when qE is inhibited by the absence of PsbS, the qE-related absorption changes are largely lost and then ECS_t-inv contribution seen in the WT is accordingly greatly diminished. The residual ECS_t-inv in npq4 can be largely attributed to the ΔA505 absorption change and its elimination in the npq1npq4 mutant allows us to see that pmf is predominantly composed of ΔpH once the steady-state has been established via counter ion-movements in the 10–60 s that follow illumination. Once again, the ECS parsing method would suggest that ΔpH is enhanced in the npq4 and npq1npq4 compared to the WT, yet the photosynthetic control phenotypes of these mutants confirm it is unchanged (Horton et al., 2000; Roach and Krieger-Liszkay, 2012; Tikkanen et al., 2015). Our conclusion of a dominant ΔpH contribution to pmf is in agreement with the recent theoretical model of Lyu and Lazár (2017), which suggested a steady-state Δψ of just 14 mV, which is ∼10%–15% of the total pmf value required to drive ATP synthesis given a H⁺/ATP of 4–4.67, as suggested by both functional (Steigmiller et al., 2008; Petersen et al., 2012) and structural studies (Daum et al., 2010; Hahn et al., 2018). Furthermore, Lyu and Lazár (2017) show that whilst high light intensities promote pmf storage as ΔpH, even under low light intensities the contribution of Δψ remains small, again in agreement with the measurements here on the npq1npq4 mutant.

How might our conclusion that pmf is dominated by ΔpH in both low and high light be reconciled with the work carried out in the last decade on thylakoid ion channels? To date, three classes of thylakoid ion channels have been reported, the TPK3 K⁺ transporter (Carraretto et al., 2013), the VCCN1 Cl⁻ transporter (Duan et al., 2016; Herdean et al., 2016a, 2016b) and the KEA3 H⁺/K⁺ antiporter (Armbruster et al., 2014). The partial absence of counterion channels in the thylakoid would be anticipated to alter the WT situation, where ΔpH dominates, leaving a larger Δψ that would diminish qE. Indeed, the vccn1 and tpk3 mutants show lower qE and an increase in ECS_t-inv, while the extent of total pmf is similar (Carraretto et al., 2013; Herdean et al., 2016a, 2016b). In contrast, overexpressors of VCCN1 (oeVCCN1) show a complete absence of ECS_t-inv, and therefore 100% ΔpH, which the authors used as an argument for the existence of a steady-state Δψ component in the WT (Herdean et al., 2016b). However, the oeVCCN1 plants also showed an increase in the total pmf and zeaxanthin synthesis, both of which reduce the overlap of qE-related absorption changes with the ECS as seen here for npq2, and in our previous study (Johnson and Ruban, 2014). Plants lacking KEA3 (kea3) show slower recovery from qE upon dark-to-low light or high light-to-low light transitions, and a corresponding penalty in terms of PSII efficiency and CO₂ fixation (Armbruster et al., 2014, 2016). ECS_t-inv is decreased in kea3 compared to the WT suggesting some steady-state Δψ in the latter. However, the lower ECS_t-inv could also be explained by reduced overlap between ΔA535 and the ECS due to increased zeaxanthin formation in kea3 (Armbruster et al., 2014, 2016). Upon high-to-low transitions in light intensity, a sudden drop in proton-coupled electron transfer, but continued H⁺ efflux through the ATPase leads to a transient inverse Δψ (Kramer et al., 2003). The inverted field limits the rate of H⁺ efflux and therefore qE relaxation, thus, an electroneutral antiporter, such as KEA3, would allow more rapid dissipation of pmf than would be possible by the ATPase alone. The slower counterion movements would then subsequently restore the steady-state domination of ΔpH at a new lower level of pmf. To our knowledge, Arabidopsis mutants lacking Ca²⁺/H⁺ or Mg²⁺/H⁺ antiporters are yet to be generated and characterized, despite evidence of both being identified in the thylakoid membrane (Barber et al., 1974; Ettinger et al., 1999). In the future, crossing the ion-channel mutants into the npq1npq4 background has the potential to clarify their net contributions to the dissipation of Δψ in the steady state.

Domination of pmf in low and high light by the ΔpH and its apparent saturation at 308 µmol photons m⁻² s⁻¹ in the npq1npq4 mutant raises a series of interesting issues for the regulation of photosynthesis. If pmf is saturated at moderate light (between 308 and 548 µmol photons m⁻² s⁻¹) in the WT (Figure 2C), why then is qE (and ΔA535) still seen to increase gradually up to the maximum light intensity used of 1,396 (Figure 2B)? A similar early saturation of the ΔpH formation is observed in isolated chloroplasts using 9AA (Schuldiner et al., 1972; Oxborough and Horton, 1988; Ruban and Horton, 1999; Evron and McCarty, 2000; Johnson and Ruban, 2011, 2014; Roach and Krieger-Liszkay, 2012; Yamamoto and Shikanai, 2020). This discrepancy can be explained by the relatively slow synthesis of zeaxanthin, which shows a half-time of approximately 6–8 min under high illumination (Bilger et al., 1989; Johnson et al., 2009; Townsend et al., 2018; Wilson and Ruban, 2020). Thus, despite the saturation of pmf at moderate light, qE continues to increase since de-epoxidation of violaxanthin to zeaxanthin shifts the pKa of the qE response from ∼5.0 to 6.0 (Horton et al., 1991, 2000; Ruban et al., 2012). This type of allosteric control is particularly crucial since it allows maximal rates of LET and qE to co-exist. Similar to qE, measurements of photosynthetic control, using the proxy of P700⁺ accumulation, suggest it reaches a maximum at high rather than moderate light intensities (Suorsa et al., 2013). However, there is evidence that photosynthetic control is also regulated by the redox state of the NADP⁺/NADPH pool, with reducing conditions increasing pH sensitivity of the cytb₆f complex (Johnson, 2003; Hald et al., 2008). Redox regulation may, therefore, work synergistically with ΔpH to regulate photosynthetic control as the xanthophyll cycle regulates qE. Such complex regulation of photosynthesis is necessary because, otherwise, the excessively large ΔpH that would be required to give the requisite downregulation of cytb₆f, and increase in qE, would lead to the inhibition of the oxygen-evolving complex of PSII (Krieger and Weis, 1993; Spetea et al., 1997; Zaharieva et al., 2011; Wilson and Ruban, 2019).

Conclusion

Our data show that the slow secondary rise of the ECS signal during illumination in the WT is caused by the strongly overlapping absorption changes associated with zeaxanthin synthesis and qE. In Arabidopsis mutants lacking an active xanthophyll cycle or qE activity, where ΔA505 and ΔA535 signals are absent it is clear that the Δψ component of the pmf is dissipated almost completely (<10% contribution) within 60 s of illumination. The data here are therefore in agreement with the wide range of existing experimental data in the literature derived from microelectrodes, pH-sensitive dyes, and radiolabelling experiments (Dilley and Vernon, 1965; Bulychev et al., 1972; Rottenberg et al., 1972; Schuldiner et al., 1972; Barber et al., 1974; Pick et al., 1974; Chow et al., 1976; Vredenberg and Bulychev, 1976; Slovacek and Hind, 1981; Bulychev, 1984; Van Kooten et al., 1986; Remiš et al., 1986; Vredenberg, 1997) all of which show a dominant ΔpH contribution to pmf in both low and high light.

Materials and methods

Plant growth conditions

WT Arabidopsis (Arabidopsis thaliana; Col-0), the violaxanthin de-epoxidase knockout (npq1; Niyogi et al., 1998), the PsbS knockout (npq4; Li et al., 2000), the PsbS overexpressor (L17; Li et al., 2002), and the violaxanthin de-epoxidase and PsbS double-knockout mutant (npq1npq4; Havaux and Niyogi, 1999) were used in this study. Seeds were sterilized in 50% (v/v) ethanol and 0.1% (v/v) Triton-X 100 and were stored for 48 h at 4°C before being sown on a 6:6:1 ratio of Levington M3 compost, John Innes No. 3 soil, and Perlite (Scotts UK, Ipswich, UK). All measurements were carried out on 4- to 5-week-old plants, grown at 190 µmol photons m⁻² s⁻¹ with a 10-h photoperiod at 22°C. Plants were grown in a Percival AR-75L3 plant growth cabinet (Percival Scientific Inc., USA), equipped with Phillips MASTER TL-D Super 80 36 W/840 bulbs, which emit a cool white light (Koninklijke Philips N.V., Netherlands). Before each measurement, plants were dark adapted for 30 min.

Absorption measurements in leaves

ECS and 535 nm absorption kinetics were measured in parallel on attached leaves on a Walz DUAL-PAM-100 (Walz, Germany) and its P515/535 emitter-detector modules (Schreiber and Klughammer, 2008), with the measuring light set to a frequency of 1,000 Hz. To calibrate each measurement to account for varying leaf thickness and chlorophyll content, the ECS signal wavelengths (ΔA 550–515 nm) were balanced using the inbuilt software and the ECS kinetics from a single-turnover pulse. Leaves were illuminated for 3 min, followed by 3 min darkness over a total of eight steps of increasing red actinic light (λ = 635 nm). The actinic light intensities used were 0, 71, 151, 308, 417, 548, 708, 1,128, 1,396 µmol photons m⁻² s⁻¹. ECS_t, ECS_inv, and ECS_t-inv were calculated as previously described (Sacksteder and Kramer, 2000; Klughammer et al., 2013), and as shown in Figure 1C.

Absorption spectra in the 410–560 nm region were measured using a SLM DW2000 dual-wavelength spectrophotometer (Olis Inc., USA), as previously described (Johnson et al., 2009). Whole Arabidopsis leaves were detached from plants dark-adapted for 30 min and the petioles wrapped in moist filter paper. The leaves were inserted into a 1 cm² transparent cuvette at 45° to the DW2000 measuring light path. An optic fiber, at 90° to the DW2000 measuring light, delivered red actinic light (700 μmol photons m⁻² s⁻¹) illuminating the leaf at 45°, and was defined using a Corning 2-58 filter. The photomultiplier was protected using a Corning 4-96 filter and an OCLl Cyan T400-570 mirror. The instrument slit-width was 5 nm and the scan rate was 4 nm s⁻¹. The sample compartment was water-cooled to maintain the leaf temperature at 22°C.

Accession numbers

The sequence data from this article can be found in The Arabidopsis Information Resource database (https://www.arabidopsis.org/) under the following accession numbers: npq1 (AT1G08550); npq4/L17 (AT1G44575).

Supplemental data

The following material is available in the online version of this article.

Supplemental Figure S1. Expanded view of low and high illumination effect on the ECS signal in each plant line.

 

M.P.J., A.V.R., and S.W. designed the research, S.W. and M.P.J. carried out the experimental work, and all authors analyzed and discussed the data and contributed to the writing of the manuscript.

The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (https://academic.oup.com/plphys/pages/general-instructions) is: Alexander V. Ruban (a.ruban@qmul.ac.uk).