What Are the Three Stages of Photosynthesis?

Photosynthesis unfolds in three interconnected stages: light harvesting and water splitting, electron transport and energy-carrier production, and carbon fixation through the Calvin cycle. The first two stages depend on sunlight and take place in the thylakoid membranes of chloroplasts, while the third stage runs in the surrounding fluid called the stroma, using the chemical energy the first two stages produced. Together, these stages convert light energy, water, and carbon dioxide into the sugars that sustain nearly all life on Earth.

Stage One: Capturing Light and Splitting Water

Photosynthesis begins when pigment molecules, primarily chlorophylls and carotenoids, absorb photons. These pigments are organized into large antenna complexes embedded in the thylakoid membrane. The biggest antenna in most land plants is called light-harvesting complex II (LHCII), and it funnels absorbed energy toward a reaction center with remarkable speed. Exactly how that energy hops from one pigment to the next is still an active area of research; the system is so large and intricate that a complete map of all the transfer pathways and their timings has been difficult to nail down even with advanced spectroscopy and computer modeling.¹

Once the energy reaches the reaction center of Photosystem II (PSII), it powers one of the most extraordinary chemical feats in biology: splitting water. A cluster of four manganese atoms and one calcium atom, bound within the PSII protein complex, strips electrons from water molecules and releases oxygen as a byproduct.² The reaction requires the accumulation of four oxidizing equivalents, meaning the cluster cycles through several intermediate states before it can pull apart two water molecules to release one molecule of O₂.³ This is where virtually all the oxygen in Earth’s atmosphere comes from. The electrons liberated from water do not stay at PSII; they are passed along to the next stage.

Stage Two: Electron Transport and Energy-Carrier Production

The electrons freed from water travel down a chain of protein complexes and small mobile carriers embedded in the thylakoid membrane. This electron transport chain links Photosystem II to Photosystem I (PSI), which sits in a different part of the thylakoid system. A small copper-containing protein called plastocyanin acts as the long-range shuttle that carries electrons between the two photosystems. Research on mutant plants with different thylakoid geometries has confirmed that plastocyanin is essential for bridging the physical gap between PSII and PSI, and the typical size of thylakoid stacks in higher plants seems tuned so that plastocyanin does not have to diffuse too far.⁴

As electrons move through the chain, the energy they release is used to pump protons (hydrogen ions) from one side of the thylakoid membrane to the other. This creates a steep proton gradient, a kind of electrochemical pressure difference across the membrane. The gradient serves as a temporary energy store. Protons flow back through a molecular turbine called ATP synthase, and that flow drives the synthesis of ATP from ADP and phosphate.⁵ The proton gradient does not just power ATP production; it also has a regulatory role, helping the chloroplast fine-tune the whole photosynthetic apparatus in response to changing light conditions.⁶

Meanwhile, at Photosystem I, the electrons are re-energized by a second hit of light. The freshly boosted electrons ultimately reduce a molecule called NADP⁺ to form NADPH. By the end of stage two, the plant has converted light energy into two chemical currencies: ATP and NADPH. Neither of those is sugar, though. That comes next.

Stage Three: Carbon Fixation in the Calvin Cycle

The Calvin cycle (sometimes called the Calvin-Benson-Bassham cycle) is where carbon dioxide from the air actually gets built into organic molecules. It runs in the stroma, the fluid surrounding the thylakoid membranes, and it consumes the ATP and NADPH that stages one and two produced. The cycle has three phases of its own: carbon fixation, reduction, and regeneration.

Carbon fixation is the entry point. An enzyme called RuBisCO grabs a molecule of CO₂ and attaches it to a five-carbon sugar called ribulose-1,5-bisphosphate (RuBP). The resulting unstable six-carbon intermediate immediately splits into two molecules of a three-carbon compound (3-phosphoglycerate). Computational studies of the RuBisCO active site have shown that CO₂ is added directly to RuBP without the need for a preliminary docking step, and that a modified lysine residue at the heart of the enzyme plays a critical role not just in this initial fixation but in subsequent chemical steps as well.⁷

In the reduction phase, ATP and NADPH are spent to convert those three-carbon molecules into glyceraldehyde-3-phosphate (G3P), a simple sugar. Some G3P exits the cycle to be assembled into glucose and other carbohydrates, but most of it stays behind for the regeneration phase.

Regeneration is the housekeeping step that keeps the cycle turning. The remaining G3P molecules are reshuffled through a series of reactions to rebuild RuBP, the CO₂ acceptor that started the whole process. This phase is catalyzed by the majority of the cycle’s enzymes: eight out of the eleven enzymes in the Calvin cycle are dedicated to regenerating RuBP.⁸ The last three steps in that regeneration sequence, carried out by ribose-5-phosphate isomerase, ribulose-5-phosphate epimerase, and phosphoribulokinase, are the final gatekeepers that determine how quickly fresh RuBP is available for another round of fixation.⁹ If regeneration falls behind, the entire cycle slows down regardless of how much light and CO₂ are available.

Where It Happens Inside the Chloroplast

The three stages are not just chemically distinct; they are spatially separated. In higher plants, the thylakoid membrane is organized into tightly stacked discs called grana and unstacked sheets called stroma lamellae that connect the grana. Photosystem II concentrates in the grana stacks, while Photosystem I lives primarily in the stroma lamellae.¹⁰ This physical separation means plastocyanin has to travel between the stacked and unstacked regions to shuttle electrons, which is part of why grana diameter matters for photosynthetic efficiency.⁴

The Calvin cycle, by contrast, operates in the stroma itself, the gel-like matrix that fills the chloroplast outside the thylakoid system. ATP and NADPH diffuse from the thylakoid membranes into this space where RuBisCO and the other Calvin-cycle enzymes are dissolved. This arrangement keeps the light-driven chemistry and the carbon-fixation chemistry in close but separate compartments, allowing each to be regulated somewhat independently.

When Oxygen Gets in the Way

RuBisCO has a well-known design flaw: it can react with oxygen as well as carbon dioxide. When it grabs O₂ instead of CO₂, it produces a toxic two-carbon compound called phosphoglycolate instead of the useful three-carbon product. The plant then has to run an energy-expensive salvage pathway called photorespiration to recycle phosphoglycolate back into something the Calvin cycle can use.¹¹ This rescue operation consumes ATP and NADPH and releases some of the CO₂ that was previously fixed, effectively undoing part of the Calvin cycle’s work.¹²

Photorespiration is not a minor nuisance. Under warm, dry conditions when stomata close and internal oxygen levels rise relative to CO₂, the oxygenase reaction can consume a substantial fraction of the plant’s photosynthetic output. The process involves not just the chloroplast but also the peroxisome and mitochondria, making it a cross-organelle salvage relay.¹³ Reducing the cost of photorespiration is one of the biggest targets in crop improvement research, because lowering that waste could meaningfully boost yields.

C4 and CAM Plants: Working Around RuBisCO’s Weakness

Some plants have evolved ways to sidestep the photorespiration problem by concentrating CO₂ near RuBisCO so it almost never encounters oxygen. C4 plants, which include corn, sugarcane, and many tropical grasses, fix CO₂ first in an outer layer of leaf cells using an enzyme called PEP carboxylase, then shuttle the resulting four-carbon acid into an inner layer of cells where it releases CO₂ directly to RuBisCO. This spatial separation keeps CO₂ levels high in the compartment where the Calvin cycle runs.

CAM plants, which include cacti and many succulents, use a similar set of enzymes but separate the two fixation steps by time instead of space. They open their stomata at night to take in CO₂, store it as an organic acid, and then release it internally during the day when the stomata are closed to conserve water. Both strategies have evolved repeatedly over the past 30 million years.¹⁴

These two strategies were long assumed to be mutually exclusive: you either separate the fixation steps spatially (C4) or temporally (CAM), not both. But that assumption turned out to be wrong. The plant genus Portulaca includes species that normally operate as C4 plants yet switch on CAM metabolism when they are drought-stressed. Detailed gene expression mapping has revealed that in at least one species, Portulaca oleracea, C4 and CAM operate in the same cells simultaneously, with carbon fixed by CAM at night feeding directly into the C4 cycle during the day.¹⁴ The discovery that these two “incompatible” modes can coexist has rewritten a core assumption in plant physiology.

How the Chloroplast Protects Itself from Too Much Light

The three stages of photosynthesis work smoothly under moderate light, but too much light is dangerous. When light energy exceeds what the electron transport chain can handle, excess energy can generate reactive oxygen species that damage membranes and proteins. Plants have evolved a suite of protective mechanisms, the most well-known being non-photochemical quenching (NPQ), which safely converts surplus absorbed light energy into heat before it can cause harm.

One key form of NPQ depends on the proton gradient across the thylakoid membrane, the same gradient that drives ATP synthesis. When the gradient builds up faster than ATP synthase can relieve it, it triggers a cascade involving a protein called PsbS and a pigment called zeaxanthin that together act as a molecular pressure valve, dumping excess energy as heat. Research in Arabidopsis has shown that disrupting the regulatory enzyme NTRC causes this quenching mechanism to go into overdrive even under moderate light, dramatically cutting the rate of useful electron transport and slowing growth.¹⁵ Mutant plants lacking both NTRC and PsbS grew faster and recovered higher electron transport rates than plants lacking NTRC alone, demonstrating that protective quenching itself can become a productivity bottleneck when it misfires. Getting the balance right between protection and productivity is one of the trickiest regulatory challenges a leaf faces, and it changes minute by minute as clouds pass overhead.

Ancient Roots of Oxygenic Photosynthesis

Chloroplasts are not native to plant cells. They descended from free-living cyanobacteria that were engulfed by an ancestral single-celled organism more than a billion years ago. Genome comparisons between modern cyanobacteria and plant chloroplasts have confirmed this beyond doubt.¹⁶ Over evolutionary time, much of the cyanobacterial DNA migrated into the host cell’s nucleus, but the chloroplast retains its own small genome and still divides semi-independently inside the cell.¹⁷

Oxygenic photosynthesis, the kind that splits water and releases O₂, was not the first type of photosynthesis on Earth. Older forms of photosynthesis, called anoxygenic, still exist in some bacteria. These organisms use molecules other than water as electron donors, so they never produce oxygen. Structural comparisons of reaction centers in oxygenic and anoxygenic organisms reveal deep similarities, pointing to a shared evolutionary ancestor for both reaction center types.¹⁸ The structure of the reaction center in Heliobacteria, one of the simplest photosynthetic organisms alive today, contains a calcium atom at the electron donor site that is strikingly similar in position to the manganese-calcium cluster in Photosystem II. This suggests the structural scaffolding for water splitting was already present in the ancestor before PSII evolved.¹⁹

Some modern cyanobacteria can still switch between oxygenic and anoxygenic modes depending on their environment. When hydrogen sulfide is abundant, certain species use it as their electron donor instead of water, running anoxygenic photosynthesis. They only switch to the water-splitting, oxygen-releasing mode when sulfide concentrations drop below a threshold.²⁰ This metabolic flexibility offers a living window into how photosynthesis may have operated during Earth’s early history, before atmospheric oxygen levels rose and water splitting became the dominant strategy.