On the input side, proton deposition increases with irradiance because PSII water oxidation releases protons to the lumen and cytochrome b₆f transfers additional protons to the lumen during plastoquinol oxidation via the Q-cycle. In vivo, the resulting pmf rise is opposed by proton return through ATP synthase and by any dissipative leaks; thus, the instantaneous pmf reflects the net balance between proton deposition and proton efflux (Sacksteder et al., 2000; Davis et al., 2017). Importantly, the initial electrical component (Δψ) can rise rapidly at light onset or during an upshift, but counter-ion fluxes (anion influx and/or cation efflux) typically dissipate Δψ on short timescales, allowing continued proton deposition to manifest as lumen acidification (ΔpH) (Cruz et al., 2001; Davis et al., 2017).

Because many of the fastest photoprotective controls are ΔpH-dependent, not “pmf-dependent in general,” the capacity to convert Δψ into ΔpH is a key determinant of how effectively the proton circuit protects PSI during rapid light increases (Davis et al., 2017; Herdean et al., 2016b).

CEF around PSI increases pmf without net NADPH accumulation by recycling electrons from the PSI acceptor side back to the plastoquinone (PQ) pool, increasing proton deposition predominantly through additional b₆f turnover and (for the NDH branch) direct proton pumping (Shikanai, 2007; Nawrocki et al., 2019). Two routes are commonly distinguished: a PGR5/PGRL1-dependent pathway and an NDH-dependent pathway (Figure 2) (Shikanai, 2007; Nawrocki et al., 2019).

CEF returns electrons to PQ, its impact on PQ redox is conditional, reflecting the balance among electron input, b₆f throughput, PSI acceptor-side capacity, and downstream stromal sink activity. A more defensible framing is that CEF can support PSI integrity indirectly by (i) increasing pmf/ΔpH and thereby strengthening photosynthetic control at b₆f (slowing electron delivery to PSI when sinks lag) and/or (ii) improving ATP supply to stimulate stromal metabolism, which can accelerate electron consumption and relieve acceptor-side limitation (Figure 2) (Shikanai, 2007; Tikkanen et al., 2015; Davis et al., 2017; Degen and Johnson, 2024).

The key point for pmf partitioning is conditional: CEF increases the supply of pmf, but whether that increase appears mainly as ΔpH (versus Δψ) depends on counter-ion dissipation of Δψ and on ATP synthase proton efflux (g(H⁺)) at that moment (Cruz et al., 2001; Davis et al., 2017).

Chloroplast ATP synthase converts pmf into ATP, but its throughput is strongly regulated. Redox regulation of the γ-subunit (a thiol/disulfide switch controlled by thioredoxin) suppresses wasteful ATP hydrolysis in the dark and modulates enzyme activity in the light, thereby changing the effective rate at which protons return to the stroma (Hisabori et al., 2013). In vivo, pmf discharge kinetics are often parameterized as an effective proton conductivity, g(H⁺), estimated from the electrochromic shift decay after dark intervals (DIRK/ECS approaches) (Kiirats et al., 2009; Davis et al., 2017).

For dynamic partitioning, the logic is simple but must be stated precisely:

Crucially, g(H⁺) is not a “leaf constant”: it responds to redox regulation and also to metabolic state (ADP, Pi availability, and stromal demand), creating a feedback loop in which carbon metabolism can modulate how rapidly pmf is cashed out as ATP (Kiirats et al., 2009; Correa Galvis et al., 2020). Recent mechanistic work in algae further emphasizes that regulatory tuning of ATP synthase can differ across lifestyles and environmental regimes, reinforcing that ATP synthase control is a legitimate regulatory lever rather than a passive boundary condition (Rühle et al, 2024).

Because Δψ and ΔpH sum to the same pmf, ion movements that dissipate Δψ can increase the fraction of pmf expressed as ΔpH, allowing faster and deeper lumen acidification for a given net proton deposition. This “Δψ-relief to ΔpH-expression” logic underlies the importance of thylakoid anion and cation pathways for rapid photoprotection during light increases (Cruz et al., 2001; Davis et al., 2017).

A mechanistically consistent transient picture is as follows.

In the sections “upshift/downshift” logic to link pmf partitioning to rapid photoprotective engagement and recovery, and NPQ (especially qE) is often treated as a convenient kinetic readout of how rapidly lumen acidification is expressed and relaxed. However, NPQ is ΔpH-gated rather than ΔpH-determined: for an equivalent lumen pH trajectory, the sensitivity (pH threshold), amplitude, and relaxation of NPQ can differ substantially with genotype, pigment composition, and environmental or metabolic history. Conversely, differences in NPQ kinetics cannot be interpreted uniquely as differences in ΔpH timing unless pmf/partitioning (e.g., ECS-derived ΔpH fraction and g(H⁺)) and, where relevant, pigment state are assessed in parallel. These dependencies are especially important under fluctuating light, where “pre-conditioning” (previous sunflecks, stromal redox state, and temperature) can change the apparent speed and magnitude of NPQ independently of the underlying proton-circuit dynamics.

Table 2 summarizes major factors that shift NPQ induction and relaxation in intact leaves, providing a practical guide for interpreting NPQ transients without implicitly assuming a uniform activation model (Table 2).

At the level of first principles, pmf (ΔH⁺) equals the sum of Δψ and the chemical term (2.303 RT/F)· ΔpH. Conceptually, Δψ is shaped primarily by VCCN1, CLCe, and cation fluxes, whereas the ΔpH term is set by the balance between proton pumps and ATPase-mediated efflux, modulated by KEA3. Operationally, pumps (LEF/CEF) set the supply of pmf, ATP synthase (g(H⁺)) sets how fast pmf is spent as ATP, and VCCN1/CLCe/KEA3 (plus other ion fluxes) shape how much of that pmf is expressed as lumen acidification and how quickly it relaxes. This partitioning tunes the engagement of ΔpH-dependent photoprotective responses (qE and photosynthetic control at b₆f) and therefore the stability of PSI under fluctuating light (Cruz et al., 2001; Davis et al., 2017; Armbruster et al., 2016; Gollan et al., 2023) (Figure 3).

The dominant fast NPQ component in plants, qE, is triggered by lumen acidification and is strongly dependent on the small thylakoid protein PsbS (Li et al., 2000, 2002). PsbS is a member of the light-harvesting complex superfamily but does not act as the quencher itself, because it does not bind a stable set of antenna pigments in the way LHC proteins do. Instead, PsbS functions as a ΔpH-responsive regulatory module that alters the interaction network and conformational landscape of PSII antenna proteins, increasing the probability of forming quenched antenna states when lumen pH drops (Papageorgiou and Govindjee, 2014; Correa-Galvis et al., 2016).

Mechanistically, qE activation involves protonation of conserved lumen-exposed acidic residues in PsbS (often described as two key glutamates in Arabidopsis; residue numbering can vary depending on how the N-terminus is defined and processed). Structural and residue-level analyses support the view that protonation shifts PsbS toward conformations/interactions that favor antenna reorganization and quenching competence, rather than creating a pigment-localized quenching site within PsbS itself (Fan et al., 2015; Liguori et al., 2019). Residue-by-residue studies further suggest that additional local sequence features (e.g., lumen loop segments and hydrophobic patches on transmembrane helices) tune sensitivity and kinetics, which is important for explaining why qE can differ in amplitude and relaxation across genotypes and environments (Liguori et al., 2019; Correa-Galvis et al., 2016).

Because PsbS lacks pigments, the dominant quenching sites are thought to reside in LHCII and/or minor antenna complexes, and multiple quenching locations likely contribute depending on pigment composition and membrane state (Mozzo et al., 2008; Papageorgiou and Govindjee, 2014). Zeaxanthin is not a strict prerequisite for the fastest qE onset in all contexts, but it can substantially enhance and stabilize quenching states and alter relaxation, linking qE mechanistically to the xanthophyll cycle (Li et al., 2002; Demmig-Adams et al., 2020).

The npq4 phenotype demonstrates that loss of PsbS strongly suppresses qE and changes the dynamics of energy dissipation during light transitions (Li et al., 2000, 2002). However, PSI protection should not be attributed to PsbS/qE alone: electron-transfer control (e.g., PGR5-dependent regulation and ΔpH-dependent photosynthetic control at b₆f) can dominate PSI safety under many fluctuating-light regimes, so claims that “PsbS loss causes PSI photoinhibition” must be framed as conditional and supported by direct data (Tikkanen et al., 2015; Kono and Terashima, 2016; Kono et al, 2022). Conversely, engineering studies show that accelerating NPQ relaxation (often involving PsbS plus additional components) can improve carbon gain under fluctuating light, whereas overly persistent quenching can reduce yield when the light regime does not demand sustained dissipation (Kromdijk et al., 2016; Kromdijk and Walter, 2023).

The xanthophyll cycle couples lumen acidification to reversible changes in carotenoid composition that modulate both the magnitude and persistence of NPQ. Under acidic lumen pH, violaxanthin de-epoxidase (VDE) is activated on the lumenal side and uses ascorbate to convert violaxanthin to antheraxanthin and zeaxanthin, enriching the antenna with pigments that stabilize quenching states and improve photoprotection (Demmig-Adams et al., 2020). Importantly, VDE activity is not only pH-dependent but also lipid-dependent; evidence and models support recruitment/activation in MGDG-rich membrane environments, making local lipid organization a plausible contributor to heterogeneity in quenching competence (Goss and Latowski, 2020).

During lower light or recovery phases, zeaxanthin epoxidase (ZE) reconverts zeaxanthin toward violaxanthin on the stromal side, promoting relaxation. Under photoinhibitory stress, ZE abundance or activity can decrease, prolonging zeaxanthin retention and thereby extending the lifetime of quenching states—an effect consistent with ZE degradation under PSII photoinhibition conditions (Bethmann et al., 2019; Demmig-Adams et al., 2020).

Functionally, the division of labor is best stated explicitly:

Thus, changes in NPQ kinetics that are attributed solely to zeaxanthin should be interpreted carefully unless PsbS status and ΔpH dynamics are simultaneously measured.

Lumen acidification also enforces a rapid feedback at the main electron-transfer bottleneck: the cytochrome b₆f complex. When light harvesting transiently exceeds stromal capacity to consume ATP and reducing power, ΔpH can slow plastoquinol oxidation at the Qo site (“photosynthetic control”), restricting electron transfer from PSII toward PSI and thereby reducing the probability of PSI acceptor-side over-reduction during transients (Tikkanen et al., 2015; Kono and Terashima, 2016; Degen and Johnson, 2024).

This control is mechanistically distinct from qE:

Both mechanisms are triggered by ΔpH, the timing and amplitude of lumen pH changes—set upstream by proton deposition, ATP synthase g(H⁺), and counter-ion fluxes—determine how effectively these switches engage during a sunfleck and how quickly they relax during a shadefleck (Davis et al., 2017; Degen and Johnson, 2024).

Beyond qE, additional NPQ components operate on longer timescales and should be treated as mechanistically distinct rather than as “slower qE.”

The system performs well when ΔpH kinetics align with the response windows of the effectors:

Therefore, upstream partitioning controls—ATP synthase g(H⁺), KEA3-dependent ΔpH relaxation after downshifts, and Δψ-relieving anion pathways (VCCN1/CLCe)—indirectly determine when PsbS and VDE engage and how long quenching persists. This explains why thylakoid ion transport mutants can show large changes in NPQ amplitude and relaxation even when antenna proteins are unchanged (Davis et al., 2017; Armbruster et al., 2016; Dukic et al., 2019, 2022).

Light-driven proton pumping acidifies the lumen while alkalinizing the stroma, and this pH reset is not ancillary; it is required to bring several CBB enzymes into their high-activity regime and to coordinate them with the light reactions (Walker et al., 1986; Wang and Portis, 2006; Gregory et al., 2024). In vivo analyses and synthetic views converge on the point that stromal alkalization, with concomitant Mg²⁺ release into the stroma, promotes activation of FBPase, SBPase, and PRK, facilitates reductive activation by the thioredoxin system so that carbon-fixation capacity tracks light availability (Okegawa and Motohashi, 2015). Rubisco activase (RCA) links this protonic state to rubisco engagement, because its ATP/ADP- and redox-sensitive regulation makes it responsive to the chloroplast energy/redox state established by pmf and electron transport. Mechanistic and structural evidence further supports a model in which redox control tunes RCA nucleotide sensitivity, providing a molecular basis for light-dependent Rubisco activation (Wang and Portis, 2006). The proton circuit, therefore, does not merely supply ATP; it reconfigures the stromal microenvironment—pH, Mg²⁺, redox—so that the CBB cycle runs at the “right” set point for the current light and can be quickly dialed back when light fades (Werdan et al., 1975).

The “switch-on” logic is incomplete if treated only as pH/Mg²⁺/redox. ATP synthase requires Pi as a substrate, and its throughput is constrained by the return of ADP and Pi from stromal metabolism. When Pi availability is low (e.g., because Pi is sequestered in phosphorylated intermediates, export is restricted, or TPU limitation develops), proton efflux through ATP synthase can be throttled even if pmf is high, strengthening ΔpH-dependent regulation and slowing relaxation after downshifts (Kiirats et al., 2009; McClain et al., 2023). In this view, stromal alkalinization and enzyme activation set the 3.62.6 Contextual determinants of NPQ kinetics and amplitude for carbon fixation, whereas adenylate/Pi supply and sink turnover set the realized flux and strongly influence the kinetics of recovery after fluctuations.

By stoichiometry, LEF (Linear Electron Flow) provides ~1.3 ATP per NADPH, whereas C₃ carbon fixation plus photorespiration often demands more ATP per NADPH, especially in warm, variable light (Cruz et al., 2001). This mismatch necessitates pmf-boosting and ATP-biasing mechanisms—most notably CEF around PSI and NDH-dependent routes—which add protons, and thus ATP potential, without net NADPH production (Cruz et al., 2001; Shikanai, 2007; Buchanan, 2017). A recent review on “alternative power lines” argues that multiple auxiliary routes—CEF (Cyclic electron flow) via PGR5/PGRL1, NDH, and photorespiratory or respiratory couplings—dynamically rebalance ATP/NADPH to sustain carbon metabolism when LEF alone would undersupply ATP (Davis et al., 2017; Kedem et al., 2021). In C₄ leaves, where bundle-sheath ATP demand is higher due to the CO₂-concentrating cycle, NDH-mediated CEF has been quantified in vivo as the dominant PSI electron route in bundle-sheath chloroplasts, underscoring the centrality of proton-pumping routes for sustaining assimilation in these tissues (Ermakova et al., 2019). When reductant accumulates (high NADPH/low ATP), chloroplasts export reducing equivalents via the NADP-MDH “malate valve,” handing off redox to the cytosol or mitochondria and indirectly helping recycle NADP⁺ for PSI while supporting ATP production outside the chloroplast; this shuttle is repeatedly highlighted as a key buffer that couples the proton circuit to downstream metabolism (Figure 3) (Scheibe, 2004; Selinski et al., 2014; Selinski and Scheibe, 2019).

In natural light, assimilation rarely reaches steady state, and the timing of ΔpH formation and relaxation—set by pumps, the ATP synthase “valve,” and counter-ion routes—gates both protection and productivity on sub-minute scales (Werdan et al., 1975; Michelet et al., 2013; Savage et al., 2016; Walker et al., 2020). At sunfleck onset (low to high light), a rapid ΔpH spike triggers qE and photosynthetic control at b₆f, preventing PSI over-reduction while Rubisco and RCA catch up; because PSI lacks a rapid repair cycle, this prioritization averts damage that would depress CO₂ uptake for days. At shadefleck onset (high to low light), the rate of ΔpH relaxation limits recovery of PSII quantum yield and thus CO₂ assimilation; KEA3, a thylakoid K⁺/H⁺ antiporter whose C-terminus senses chloroplast energy state, accelerates qE and ΔpH relaxation and improves post-shadefleck efficiency. Under chronic fluctuation, insufficient biasing of ΔpH at upshifts—due to weak anion or cation counter-fluxes—leads to PSI photoinhibition, slow recovery, and assimilation constraints that far outlast the transient; recent syntheses and experiments place b₆f “photosynthetic control” at the center of this protective logic (Werdan et al., 1975; Avenson et al., 2005; Hisabori et al., 2013; Davis et al., 2017; Foyer and Noctor, 2020). Faster assimilation transients are observed when sink strength rises; for example, elevated CO₂ lowers PSI over-reduction during light transitions by pulling electrons into the CBB cycle more rapidly (Kono and Terashima, 2016).A critical addition is that Pi recycling can be rate-limiting in both directions. During rapid induction, Pi availability influences ATP synthase throughput and therefore the pace at which ATP supply rises to support CBB activation and RuBP regeneration (Kiirats et al., 2009). During recovery, Pi sequestration in phosphorylated intermediates and limitations in export/end-product use can slow ATP synthase flux and prolong ΔpH, delaying NPQ relaxation and suppressing assimilation even when incident light has decreased (McClain and Sharkey, 2019; McClain et al., 2023). These effects help explain why assimilation transients often reflect combined “biophysical timing” (ΔpH dynamics) and “biochemical timing” (Pi return and sink engagement), and why interpreting NPQ kinetics as a direct proxy for ΔpH timing can be misleading without concurrent assessment of pmf partitioning and Pi/export status.

ΔpH is the shared trigger for qE, xanthophyll cycling, and b₆f control, the same levers that set pmf partitioning, also tune CO₂-uptake dynamics (Michelberger et al., 2025). Lower ATP synthase conductance (g(H⁺)) preserves ΔpH for protection but can transiently restrict Φ_PSII and ATP supply, whereas higher conductance speeds ATP delivery and ΔpH collapse but risks PSI if electron sinks lag; the optimal g (H⁺) depends on the light-fluctuation pattern and downstream sink capacity (Michelet et al., 2013; Johnson et al., 2014; Davis et al., 2017). KEA3 modulates ΔpH relaxation in response to stromal energy state (ATP/ADP, pH) and thereby helps avoid “over-holding” qE after shadeflecks, reducing unnecessary losses in ΦPSII and boosting assimilation (Armbruster et al., 2016; Wang and Shikanai, 2019). Rapid Δψ relief through VCCN1 biases pmf to ΔpH at sunflecks, strengthening protective throttling at b₆f and preserving PSI, which maintains assimilation capacity over hours to days. The overarching design principle is that the proton circuit must be tuned not for maximal ΔpH or maximal ATP alone, but for the right ΔpH at the right time, preserving PSI while minimizing unnecessary qE carry-over that would tax ΦPSII and CO₂ uptake (Wilson et al., 2021; Hagino et al., 2022).

Downstream of the CBB cycle, the triose-phosphate/phosphate translocator (TPT) exchanges triose-P for inorganic phosphate (Pi), thereby coupling chloroplast export to Pi recycling for ATP synthase. Under high sugar build-up or low Pi return, leaves can enter triose phosphate utilization (TPU) limitation, in which assimilation is capped by end-product processing or export rather than by Rubisco activity or RuBP regeneration. Recent physiological and modeling studies treat TPU as a systems-level property of the source–sink network that becomes especially prominent under specific combinations of CO₂, temperature, and sink strength, rather than as a simple, static biochemical constant (Quick and Schaffer, 2017; McClain and Sharkey, 2019; Rogers et al., 2021). Because ATP synthase requires Pi as substrate, Pi availability—and thus TPT-mediated return—feeds back onto ATP production from pmf, providing another route by which carbon-export status influences how proton-motive energy is cashed out (Cruz et al., 2001; Golding and Johnson, 2003).

Beyond transport stoichiometry, source–sink status is coordinated by broader regulatory networks that couple carbohydrate availability, phloem transport capacity, and stress responses. While a full treatment is beyond the scope of this review, the central implication for a proton-circuit framework is that carbon-status signaling and sink strength can modulate the same proton-circuit variables that determine ΔpH timing, including g(H⁺), the persistence of ΔpH, and the degree to which photoprotective responses are maintained across repeated fluctuations (Savage et al., 2016; Quick and Schaffer, 2017). Likewise, photorespiration reshapes ATP demand and carbon flux partitioning, altering when ATP supply versus protection becomes the dominant constraint. These couplings motivate interpreting “proton-centric” diagnostics (ECS-based pmf/partitioning and g(H⁺), NPQ, P700) alongside carbon-status indicators (gas exchange, ACi/TPU signatures, and—where feasible—pigment state and Pi/nutrient context) to attribute limitations correctly in intact leaves (Walker et al., 2020; Rogers et al., 2021; McClain et al., 2023).

Plant thylakoids form a connected membrane network organized into appressed grana stacks and unappressed stroma lamellae, with grana margins providing curved transition regions between the two. This architecture is associated with lateral segregation of photosynthetic complexes: PSII–LHCII is enriched in grana cores, whereas PSI and chloroplast ATP synthase are largely excluded from grana cores and enriched in stroma-exposed membranes; cytochrome b ₆ f is distributed across domains and has frequently been reported to show enrichment toward grana margins (Daum and Kühlbrandt, 2011; Anderson et al., 2012; Kirchhoff, 2013, 2014).

A key point for proton dynamics is that the lumen is topologically continuous, so lateral equilibration of lumen protons is possible in principle. However, continuity does not imply instantaneous equilibration: electron-tomography and diffusion-based frameworks emphasize that narrow junctions at margins, tortuous geometry, and crowding can slow effective diffusion and create conditions where local proton chemistry differs transiently across domains (Daum and Kühlbrandt, 2011; Kirchhoff, 2014; Kirchhoff et al., 2008).

The simplest physical reason to expect heterogeneity is spatial separation of proton sources and sinks. PSII and cytochrome b6 f deposit protons into the lumen, while ATP synthase provides the dominant proton return (“valve”), and these components are not co-localized (Kirchhoff, 2013, 2014). During light transients, this source–sink separation can generate lateral gradients in proton concentration and local pmf partitioning (ΔpH vs Δψ), especially when counter-ion pathways that dissipate Δψ are locally limiting. In this view, domain-level heterogeneity is expected to be most pronounced on seconds-to-minutes timescales where pumping/efflux rates change quickly (Tikhonov, 2012; Davis et al., 2017).

A second contributor is the local balance of charge compensation. Experiments and theory show that ionic strength and counter-ion movements can change how pmf is expressed as Δψ versus ΔpH, even when total pmf is similar (Cruz et al., 2001; Davis et al., 2017). While most measurements of Δψ/ΔpH are spatially averaged, the same logic implies that regional differences in ion permeability (anion influx and cation efflux pathways) could yield domain-dependent partitioning during transients (Davis et al., 2017).

Thylakoid organization is not static. State transitions remodel antenna connectivity and excitation distribution on minute timescales through STN7-dependent LHCII phosphorylation, changing the coupling between PSII and PSI and altering the spatial pattern of excitation pressure (Minagawa, 2011; Grieco et al., 2012; Mekala et al., 2015). Recent work also indicates additional STN7-linked regulatory behaviors beyond the simplest “LHCII moves” cartoon, reinforcing that multiple remodeling routes can coexist (Węgrzyn et al., 2022). Because excitation distribution helps set where electron flow (and thus proton deposition) is strongest, these reorganizations are expected to influence proton budgets indirectly, even if pmf itself remains globally dominated by ΔpH. Modeling frameworks for state transitions and membrane interactions provide useful constraints on how these reorganizations can change effective connectivity and diffusion geometry (Wood and Johnson, 2020).

In parallel, lumen swelling and osmotic responses accompany illumination and lumen acidification. Water flux following ion movements changes lumen dimensions and can therefore change effective diffusion distances and buffering capacity, implying that proton equilibration kinetics are themselves dynamically modulated by the light state (Tikhonov, 2012; Kaňa et al., 2016).

ATP synthase is concentrated in stroma-exposed membranes and is largely excluded from grana cores, establishing spatially localized proton efflux zones that are separated from the strongest PSII proton sources in grana. This geometry alone makes transient lateral ΔpH gradients plausible unless fast charge-compensating ion fluxes minimize local Δψ buildup (Kirchhoff, 2013, 2014; Davis et al., 2017).

For ion pathways that reshape partitioning, molecular identity is now strong, but sub-domain localization is still a limiting uncertainty for spatial models. VCCN1 has been structurally resolved as a bestrophin-like homopentamer, but the in vivo distribution across grana margins versus stroma lamellae is not yet firmly mapped; because VCCN1 contributes to Δψ relief and supports rapid ΔpH expression during illumination, its placement will strongly influence whether ΔpH rises quickly at the relevant sites during upshifts (Herdean et al., 2016b; Dukic et al., 2019; Hagino et al., 2022). CLCe is genetically linked to chloroplast ion homeostasis and to regulatory responses that include LHCII phosphorylation behavior, but its spatial deployment across domains also remains uncertain, again limiting predictive modeling of Δψ dissipation in space (Herdean et al., 2016a; Dukic et al., 2022). KEA3 is best treated as a ΔpH relaxation module after downshifts; mapping where KEA3 is enriched (margins vs lamellae vs broader distribution) is essential for predicting where ΔpH relaxation initiates and how quickly it propagates (Armbruster et al., 2016; Wang and Shikanai, 2019).

A conservative, testable working hypothesis is that grana margins act as regulatory “exchange zones” because they juxtapose (i) a transition region between grana and lamella membranes, (ii) cytochrome b6 f distribution that often shows margin bias, and (iii) proximity to ATP synthase-rich stroma-exposed regions. If Δψ-relieving pathways (e.g., VCCN1/CLCe) are also enriched at margins, this would favor rapid ΔpH expression during upshifts and rapid engagement of ΔpH-dependent controls. This remains a hypothesis until localization and local-pH readouts are directly measured (Kirchhoff, 2013, 2014; Davis et al., 2017; Hagino et al., 2022).

Finally, the TPK3 localization debate provides a cautionary precedent: putative thylakoid channels should not be treated as established spatial players without convergent localization and functional evidence (Carraretto et al., 2013; Höhner et al., 2019).

At seconds-to-minutes timescales, leaf photosynthesis is well described by a five-node circuit organized into two nested feedback loops (Kiirats et al., 2009; Hohmann-Marriott and Blankenship, 2011). Proton-pumping modules, PSII, cytochrome b₆f, and CEF, raise the total pmf. Counter-ion pathways—notably KEA3 on the K⁺/H⁺ exchange side and VCCN1/CLCe on the anion side—reshape that pmf toward its chemical component (ΔpH) by relieving the electrical component (Δψ) (Dukic et al., 2022; Hagino et al., 2022). ATP synthase then spends pmf as ATP, and its effective proton conductance, gH+, sets the rate at which ΔpH collapses after a change in light or metabolic demand. pH-sensitive effectors read the ΔpH signal and throttle photochemistry in real time: PsbS mediates rapid, reversible qE; the VDE/ZE (violaxanthin de-epoxidase/zeaxanthin epoxidase) xanthophyll cycle establishes the longer-lived, zeaxanthin-dependent quenching state; and ΔpH-gated “photosynthetic control” at b₆f slows plastoquinol oxidation when sinks lag. Carbon metabolism (the Calvin–Benson–Bassham enzymes together with Rubisco activase and export/transport steps) draws down electrons and ATP and thereby feeds back on the upstream redox and energy state that determines how much pmf is built and how it is partitioned. In compact terms, the fast protective loop raises ΔpH, engages PsbS and b₆f control, and keeps PSI oxidized, whereas the productivity loop converts pmf into ATP, activates the CBB cycle, strengthens sinks, and relieves upstream over-reduction (Werdan et al., 1975; Zu Tittingdorf et al., 2004; Dall’Osto et al., 2012; Carmo-Silva and Salvucci, 2013; Hisabori et al., 2013; Wilson et al., 2021; Carbonera et al., 2022; Juhaszova et al., 2022).

The operational objective of the chloroplast proton-control network is conserved across genotypes and environments: to generate the right ΔpH at the right time. Too little ΔpH during a light upshift delays engagement of qE and cytochrome b₆f–dependent photosynthetic control, increasing the risk of acceptor-side over-reduction at PSI and long-lived photoinhibition; too much ΔpH—or ΔpH that relaxes too slowly after a downshift—carries unnecessary NPQ into shade and suppresses ΦPSII and CO₂ uptake (Zivcak et al., 2015; Allahverdiyeva et al., 2015; Tikhonov, 2018; Bethmann et al., 2019; Didaran et al., 2024). The same levers that shape pmf therefore set where the leaf sits on the photoprotection–productivity trade-off: anion flux pathways (e.g., VCCN1/CLCe) dissipate Δψ and thereby favor rapid ΔpH formation during light increases; KEA3 together with g(H+) sets the pace of ΔpH relaxation after transitions to lower light; and the balance between linear and cyclic electron flow determines how much pmf is generated in the first place. Because thylakoid architecture can impose spatial heterogeneity—microdomain ΔpH near b₆f/PSII versus ATP synthase–rich regions—local kinetics can shape how quickly protection is engaged and how quickly antennae reopen (Armbruster et al., 2016; Wang and Shikanai, 2019). Viewed this way, performance is a timing problem rather than a maximization problem: protection benefits from a prompt ΔpH pulse during light upshifts, whereas productivity benefits from timely ΔpH relaxation during downshifts so that quenching does not outlast the need for it (Werdan et al., 1975; Wilson et al., 2021).

The operating-regime map that follows turns the proton-centric model into an experimenter’s field guide. For each common context in intact leaves, it links how the pmf is partitioned between ΔpH and Δψ, which controllers and effectors are expected to respond (KEA3, VCCN1/CLCe, ATP-synthase gH+, PsbS-mediated qE, and ΔpH-gated control at cytochrome b₆f), and what carbon outcomes should result, including assimilation rate A, ΦPSII, and the redox status and safety of PSI (Wang and Shikanai, 2019; Bru et al., 2021). The map is organized around regimes that recur in practice: seconds-scale sunflecks and shadeflecks; dark-to-light induction; periods of high sink strength such as elevated CO₂; counter-ion–limited genotypes (vccn1, kea3, clce); situations with high ATP bias in photorespiring C₃ leaves; TPU/Pi limitation mediated by triose-phosphate export and phosphate return; state-transition–driven antenna remodeling; low-temperature, high-light conditions that permit qH; methodological pitfalls in pmf partitioning such as Δψ misassignment; and proton geography in which microdomain ΔpH governs local kinetics.

For each light condition, Table 3 specifies what balance of ΔpH and Δψ to expect, which protective and regulatory switches should engage, how net CO₂ assimilation (A) is likely to behave, and which measurements resolve remaining ambiguities. The minimal, comparable measurement set includes ECSt as a pmf proxy; the ΔpH fraction from multi-wavelength ECS; gH+ from DIRK analysis; NPQ and P700 kinetics; A_Ci curves; xanthophyll de-epoxidation state; and Pi/TPT status where TPU is suspected (Table 3).

An operating-regime map is most informative when anchored to realistic pH set points that translate directly into the chemical component of the pmf. In intact leaves, illumination typically drives the stroma toward alkaline values near pH 8 while acidifying the lumen. Under strong light the lumen approaches roughly pH 5.5, whereas moderate light yields values nearer pH 6.0, and both compartments relax toward ~7.4 during light-to-dark transitions. Converting these pairs with (2.303 RT/F) ΔpH gives a chemical term of about 148 mV for high light (ΔpH=2.5), 118 mV for low–moderate light (ΔpH=2.0), and essentially zero in darkness (Werdan et al., 1975; Herdean et al., 2016; Hagino et al., 2022). These lumen and stromal pH dynamics and provide a quantitative backdrop for the pmf bars in Figure 4, which stack the computed ΔpH contribution above an electrical base representing Δψ. In vivo, Δψ tends to be smaller in bright light, somewhat larger at moderate light, and negligible after darkening, because ion movements dissipate the field as conditions change (Cruz et al., 2001; Wilson et al., 2021) (Figure 4).

Under high light conditions, rapid water oxidation at PSII and a full Q-cycle at cytochrome b₆f deposit protons into the lumen so that pmf is carried predominantly as ΔpH around the pH 5.5 set point. Anion influx through the thylakoid bestrophin-like channel VCCN1 relieves Δψ and thereby biases the partition toward ΔpH; plants lacking VCCN1 accumulate a larger electrical term and are more vulnerable during rapid light increases, consistent with the logic of the operating-regime map (Sacksteder et al., 2000; Li et al., 2002; Cramer et al., 2011). Under these conditions, ΔpH protonates PsbS within seconds and activates VDE near its acidic optimum, while ΔpH-gated “photosynthetic control” at b₆f restricts plastoquinol oxidation until carbon sinks are fully engaged, protecting PSI during the ramp (Degen and Johnson, 2024).

At low and moderate light, lumen pH stabilizes nearer pH 6.0, the chemical term is smaller, and Δψ contributes a larger fraction of the pmf unless it is actively dissipated. KEA3 becomes important for shaping this regime: by exchanging stromal K⁺ for luminal H⁺, KEA3 hastens ΔpH relaxation after down-steps, limits needless carry-over of qE, and improves PSII quantum efficiency in fluctuating light; genetic and overexpression analyses show that tuning KEA3 levels or regulation speeds recovery from high to low light without sacrificing the capacity to protect during the next up-step (Armbruster et al., 2016; Wang and Shikanai, 2019). Because stromal pH remains near pH 8 in this regime, Calvin-cycle enzymes stay in a light-activated state, but the weaker acidification means that effective Δψ relief is essential to prevent PSI over-reduction when light suddenly increases (Michelet et al., 2013).

During light-to-dark transitions, both compartments relax toward pH 7.4, the ΔpH term collapses, and Δψ decays, extinguishing the driving force for ATP synthesis and allowing NPQ to dissipate. Inner-envelope K⁺/H⁺ antiporters KEA1 and KEA2 are now fully engaged to neutralize the stroma, complementing thylakoid-level H⁺ efflux and explaining the rapid loss of the electrochromic signal at darkening; functional analyses show that KEA1/KEA2 activity is dispensable for maintaining alkaline stroma in the light but is critical for fast down-regulation of stromal pH in the dark, in line with the behavior depicted in the schematic (Correa Galvis et al., 2020; Dukic et al., 2022; Juhaszova et al., 2022).

Partitioning matters because the two components of pmf have distinct physiological consequences. A larger ΔpH fraction promptly engages qE and b₆f control and is therefore inherently protective during up-steps, whereas an oversized or persistent Δψ-dominant pmf is problematic because it fails to generate the ΔpH signal needed for protective feedbacks (Tikhonov, 2018; Wilson et al., 2021). Experimental manipulations that elevate the electrical term at the expense of ΔpH increase PSII damage and depress productivity under fluctuating light, underscoring why leaves are wired to counter-ion flux limits Δψ build-up, thereby allowing proton pumping to manifest primarily as ΔpH during illumination and to release ΔpH quickly when light falls (Cruz et al., 2001).

Electrochromic-shift spectroscopy remains the standard for partitioning pmf and estimating ATP-synthase proton conductance in vivo, but quantitative use requires spectral controls and reporting conventions so that Δψ estimates are not confounded by overlapping pigment signals; coupling ECS with fluorescence-based NPQ kinetics and P700 redox traces allows the ΔpH-dependent switches and PSI safety to be operated alongside the partition itself (Johnson, 2011; Brestic et al., 2015; Tikkanen et al., 2015; Kono et al., 2017). Finally, structural work on the chloroplast ATP synthase provides a mechanistic basis for how proton efflux is gated during transients and explains why adjustments in gH+ markedly reshape ΔpH formation and decay across the three regimes (Johnson et al., 2014; Davis et al., 2017; Wilson et al., 2021).

Most ODE formulations relevant to this review share a common “state-flux” structure. The internal state can be represented by a small set of time-dependent variables, including lumen proton activity (often expressed as lumen pH), trans-thylakoid electric potential (Δψ), and total pmf (Δp = Δψ + (2.303RT/F)ΔpH), together with optional redox states (e.g., PQ pool, PSI acceptor side) depending on model scope. These states evolve as the net result of fluxes that correspond directly to the circuit elements discussed in Sections 2–7:

In a minimal dynamical representation, lumen acidification can be expressed conceptually as:

(Equation 2):

and Δψ can be treated as an electrical state set by a membrane capacitance and net ionic currents:

(Equation 3):

where JH+pump aggregates LEF/CEF-linked proton deposition, JH+ATPase reflects ATP synthase proton efflux (linked to g(H⁺) and substrate availability), Inet captures the balance of proton-generated charge separation and counter-ion currents (e.g., Cl^- influx through VCCN1/CLCe-like terms and cation fluxes), βlumen is an effective lumen buffering capacity, and Cm is an effective thylakoid membrane capacitance. Although different models implement these terms at different levels of detail, the central advantage is that the framework forces explicit accounting of how fluxes and constraints jointly determine ΔpH:Δψ partitioning over time (Lyu and Lazár, 2017, 2022).

A primary contribution of mechanistic ODE models is that they separate three concepts that are easily conflated in narrative descriptions: (i) total pmf magnitude, (ii) pmf parsing into ΔpH versus Δψ, and (iii) the kinetics with which parsing changes during transients. By explicitly representing the competing timescales of proton pumping, ATP synthase discharge, and counter-ion flux, models can reproduce situations where pmf rises quickly but lumen acidification is delayed (Δψ-heavy transient) or where ΔpH persists because ATP synthase throughput is limited by kinetic regulation and/or substrate supply (Lyu and Lazár, 2017, 2023).

A second contribution is that several recent ODE formulations incorporate electrostatic features that are difficult to isolate experimentally, including membrane-surface ion binding and Donnan/diffusion components of electric potential. These features provide a mechanistic route for Δψ behavior that is not simply a “short-lived spike,” and they offer quantitative hypotheses for when electrical contributions could persist or become functionally consequential despite substantial ΔpH expression (Lyu and Lazár, 2022).

Finally, recent work has emphasized sensitivity analysis as a practical bridge between modeling and experiment: by ranking which parameters most strongly influence predicted Δψ/ΔpH dynamics or observable signatures (e.g., ECS or fluorescence kinetics), models can guide which measurements are most informative and which engineering interventions are most likely to shift leaf performance in targeted regimes (Lyu and Lazár, 2024a, b).

To keep this section focused, Table 4 highlights representative ODE-based studies that explicitly simulate Δψ, ΔpH, and pmf dynamics while exploring the influences of ion fluxes, ATP synthase activity, and membrane electrostatics. These studies are not intended as an exhaustive modeling review; rather, they exemplify how formal dynamical models can be used to (i) test the internal consistency of mechanistic interpretations of ECS/fluorescence/P700 data, (ii) quantify how changing conductances shifts ΔpH “timing,” and (iii) identify parameter regimes where ΔpH-gated control might decouple from expected photoprotective outcomes.