Chloroplast Structure
A single spinach leaf cell holds 50–100 chloroplasts. Each converts sunlight into chemical bonds. They work fast enough to fix billions of tons of carbon dioxide worldwide every year. Understanding how this happens requires opening the organelle up. We must ask a deceptively simple question. Why does the chloroplast have not one membrane, but three separate membrane systems? Each system does a different job. The answer connects directly to how the organelle is built. It also hints at where it came from. Chloroplasts belong to the plastid family of organelles. Like Mitochondria: Structure and Function, they are double-membrane-bound. They contain their own DNA. They generate ATP using a membrane-embedded electron transport chain. But chloroplasts add a third membrane system found nowhere in mitochondria. This is an internal network of flattened sacs called thylakoids. They are stacked into structures called grana. Stroma lamellae connect these structures. This third compartment is where the light-dependent reactions of photosynthesis occur. Its existence is the single structural feature that most distinguishes a chloroplast from every other organelle in the cell. Why does this matter beyond botany? Chloroplast architecture is a textbook case of structure dictating function. The geometry of the thylakoid membrane maximizes light capture. It also creates the proton gradient that powers ATP synthesis. Meanwhile, the stroma houses the enzymes of carbon fixation in a separate aqueous compartment. Understanding this layout explains several phenomena. It explains why leaves are green. It explains how herbicides that block electron transport kill plants. It also explains why C4 plants evolved specialized chloroplast dimorphism between mesophyll and bundle-sheath cells. This article maps the chloroplast's three membranes and two internal compartments. It works through the classic experiments that revealed where photosynthesis's two halves physically occur. It also previews the deeper story—The Endosymbiotic Theory. That story explains how an organelle came to have its own genome. It also explains how the organelle got its own double membrane in the first place.
- Chloroplasts have three membrane systems: outer envelope, inner envelope, and internal thylakoid membranes—more than mitochondria's two.
- Thylakoids stack into grana (like coins) connected by unstacked stroma lamellae, together forming one continuous internal compartment enclosing the thylakoid lumen.
- The stroma (between inner envelope and thylakoids) contains chloroplast DNA, ribosomes, and the Calvin cycle enzymes, including RuBisCO.
- Light reactions occur in the thylakoid membrane; they pump protons into the lumen, creating a gradient that drives ATP synthase.
- Arnon's isolated-chloroplast experiments (1954) proved the light and dark reactions are physically separable, mapping them onto distinct compartments.
- Chloroplast DNA and 70S ribosomes reflect bacterial ancestry, foreshadowing the endosymbiotic origin of the organelle.
The chloroplast's three-membrane, two-compartment architecture spatially separates light capture (thylakoid membrane and lumen) from carbon fixation (stroma), and this separation is what makes efficient photosynthesis possible.
The core idea
A chloroplast is a lens-shaped organelle, typically 5–10 μm long and 2–4 μm wide, bounded by two concentric membranes that together form the chloroplast envelope. The outer envelope membrane is permeable to small molecules and ions through porin-like channels, while the inner envelope membrane is the true permeability barrier, studded with specific transporters that control traffic of triose phosphates, phosphate, ATP, and metabolites between the cytosol and the organelle interior. Between these two membranes lies a narrow intermembrane space, structurally similar to the mitochondrial intermembrane space but far less biochemically active. Inside the inner envelope lies the stroma, a concentrated protein-rich matrix analogous to the mitochondrial matrix. The stroma contains the chloroplast's own circular DNA genome (typically 120–160 kilobase pairs, present in many copies per organelle), 70S ribosomes, and the full soluble enzyme set of the Calvin-Benson cycle—most abundantly RuBisCO (ribulose-1,5-bisphosphate carboxylase/oxygenase), widely considered the most abundant protein on Earth. This is where atmospheric CO2 is fixed into three-carbon sugars, the so-called "dark reactions," which do not require light directly but depend on ATP and NADPH supplied by the light reactions. Suspended within the stroma is the third membrane system: the thylakoids, flattened membrane sacs enclosing a single continuous internal aqueous space called the thylakoid lumen. Thylakoids stack into cylindrical piles called grana (singular granum), resembling stacks of coins, typically 10–20 disks per granum, connected to one another and to unstacked regions by tubular bridges called stroma lamellae (or intergranal thylakoids). This gives the chloroplast, uniquely among organelles, five distinct aqueous/membrane compartments to track: cytosol, intermembrane space, stroma, thylakoid membrane, and thylakoid lumen. Each houses a different stage of photosynthesis, and the vocabulary matters precisely because mixing up "stroma" and "lumen," or "granum" and "granal stack," leads to mechanistic errors about where light reactions versus carbon fixation occur.
Mechanism
Step 1 — Envelope traffic control. Light energy absorbed by chlorophyll ultimately needs raw materials (CO2, phosphate, water) delivered to the stroma and products (triose phosphates, O2) exported. The outer envelope membrane contains outer envelope proteins (OEPs) forming large channels permeable to metabolites up to roughly 10 kDa, similar in concept to mitochondrial porins (VDAC), though structurally and evolutionarily distinct. The inner envelope membrane is far more selective: the triose phosphate/phosphate translocator (TPT) exchanges newly fixed 3-carbon sugars for inorganic phosphate from the cytosol on a strict 1:1 antiport basis, linking chloroplast carbon export to cytosolic phosphate recycling. Other inner-envelope transporters shuttle ATP/ADP, dicarboxylates, and metabolites needed for amino acid and lipid synthesis, which also occur partly in the stroma.
Step 2 — Light capture in the thylakoid membrane. Embedded in the thylakoid membrane are four major protein complexes arranged to drive linear electron flow: Photosystem II (PSII), the cytochrome b6f complex, Photosystem I (PSI), and ATP synthase. Chlorophyll a, chlorophyll b, and carotenoid pigments organized in light-harvesting complexes (LHCII, LHCI) absorb photons predominantly in the blue (~430 nm) and red (~660 nm) wavelengths, reflecting green light—which is why chloroplasts, and the leaves that pack them, appear green. Absorbed energy funnels to reaction-center chlorophylls, where it drives charge separation.
Step 3 — Water splitting and the proton gradient. PSII oxidizes water at a manganese-calcium oxygen-evolving complex on the lumenal face of the membrane: 2 H₂O → O₂ + 4 H⁺ + 4 e⁻ This releases O2 (the source of essentially all atmospheric oxygen from photosynthesis) and deposits protons directly into the thylakoid lumen. Electrons pass from PSII through plastoquinone to cytochrome b6f, which pumps additional protons from stroma into lumen as it shuttles electrons toward PSI—directly analogous to Complexes III/IV of the mitochondrial chain moving protons across the inner membrane, but here the direction is stroma-to-lumen rather than matrix-to-intermembrane-space. The result is a lumen with a lower pH (roughly pH 5) than the stroma (roughly pH 8) under illumination—a gradient of about 3 pH units, among the steepest ΔpH values found in biological membranes.
Step 4 — ATP synthesis by chemiosmosis. Protons flow back from the lumen through CF0-CF1 ATP synthase embedded in the thylakoid membrane, down their electrochemical gradient, driving rotational catalysis that phosphorylates ADP into ATP in the stroma. This is chemiosmotic coupling, the same principle Peter Mitchell proposed for mitochondria, but oriented so that ATP is released into the stroma rather than a matrix, positioning it exactly where the Calvin cycle needs it. PSI, meanwhile, re-energizes the electron with a second photon and passes it via ferredoxin to ferredoxin-NADP+ reductase, generating NADPH in the stroma.
Step 5 — Carbon fixation in the stroma. ATP and NADPH generated at the thylakoid membrane are consumed a few nanometers away in the stroma, where RuBisCO fixes CO2 onto ribulose-1,5-bisphosphate, initiating the Calvin cycle's carboxylation, reduction, and regeneration phases to build glyceraldehyde-3-phosphate. Some product stays in the stroma to regenerate the cycle and build starch grains (visible as dense stromal inclusions); the rest exits via the TPT for cytosolic sucrose synthesis.
Step 6 — Compartment-specific concentration effects. Because the lumen is a small, enclosed volume (thylakoid lumen volume is a tiny fraction of stroma volume), even modest absolute proton flux produces a large concentration change and thus a strong driving force for ATP synthase—an example of how compartment geometry itself is a functional variable, not just packaging. This is directly analogous to how the small intermembrane space in mitochondria allows a steep proton gradient to build up quickly.
The evidence
The chloroplast's compartmental division of labor was not obvious from microscopy alone; it had to be established biochemically. The decisive work came from Daniel I. Arnon and colleagues at the University of California, Berkeley, in the early 1950s.
The problem. By the 1930s, botanists knew photosynthesis had a light-requiring phase and a phase that seemed not to need light directly (the "dark reactions," later shown by Melvin Calvin's tracer work to be CO2 fixation). But it was unclear whether both phases happened in the same structure, whether isolated chloroplasts could do either reaction at all outside the cell, and—critically for this article—whether the two phases were physically separable within the organelle.
Arnon's isolated chloroplast experiment (1954). Arnon's group ground spinach leaves and isolated intact chloroplasts by differential centrifugation, carefully preserving envelope integrity. They then showed that isolated chloroplasts, given only light, water, ADP, phosphate, and CO2 (no intact cell required), could: 1. Evolve O2 from water splitting when illuminated (Hill reaction, building on Robert Hill's 1937 discovery that isolated chloroplasts reduce an artificial electron acceptor and release O2 even without CO2 present). 2. Synthesize ATP from ADP and phosphate in a light-dependent manner—photophosphorylation—demonstrated using radioactively labeled phosphate (³²P) incorporated into ATP. 3. Fix CO2 into organic acids and sugars, but only when supplied with ATP and reduced pyridine nucleotide (NADPH), and this fixation could proceed in the dark provided ATP and NADPH were supplied artificially. This third result was the crucial proof: carbon fixation does not require light directly—it requires the ATP and NADPH that light reactions produce. Arnon showed the "light reactions" and "dark reactions" could be uncoupled experimentally, with isolated thylakoid membrane fragments carrying out photophosphorylation and electron transport independent of stroma, while stroma extracts (with RuBisCO and Calvin cycle enzymes) fixed CO2 given only ATP and NADPH supplied in a test tube, without any light at all.
Follow-up structural evidence. Electron microscopy in the 1950s–60s (notably by Kenneth Thimann's and Arnon's contemporaries, refined further by freeze-fracture studies in the 1970s) confirmed that the thylakoid membrane is physically continuous and separable from the stroma by gentle mechanical rupture, and that isolated thylakoid vesicles alone retain full electron transport and photophosphorylation capacity while lacking Calvin cycle activity—matching Arnon's biochemical fractionation exactly. Later work using thylakoid-permeant pH probes (9-aminoacridine and related fluorescent dyes) directly measured the light-induced ΔpH of roughly 3 units across the thylakoid membrane, confirming the chemiosmotic model that Peter Mitchell proposed in 1961 (for which he received the 1978 Nobel Prize in Chemistry) applies to chloroplasts as well as mitochondria.
What this collectively proved: photosynthesis's two stages are not merely conceptually distinct but map onto distinct physical compartments—the thylakoid membrane/lumen for light-driven electron transport and proton pumping, and the stroma for ATP/NADPH-driven carbon fixation—exactly the architecture described above.
Worked examples
Example 1 — Calculating thylakoid surface area and its functional payoff. A typical mesophyll chloroplast is roughly a flattened ellipsoid, 5 μm long, 3 μm wide, 2 μm thick, giving an envelope surface area on the order of 40–50 μm². Yet a single chloroplast may contain 40–60 grana, each granum stacking 10–20 thylakoid disks roughly 0.3–0.6 μm in diameter. Estimate one granum's disk area: a disk of diameter 0.4 μm has area π(0.2)² ≈ 0.126 μm² per face. With two membrane faces per thylakoid disk (top and bottom, though they enclose one lumen), and 15 disks per granum: 0.126 μm² × 2 × 15 = 3.77 μm² of thylakoid membrane per granum (rough approximation, ignoring lumenal thickness). With 50 grana per chloroplast: 3.77 × 50 ≈ 188 μm² of thylakoid membrane—roughly four times the chloroplast's own envelope surface area, packed inside a volume of only tens of cubic micrometers. This dramatic amplification of membrane area is the structural reason chloroplasts can house enough photosystem complexes to capture the large photon flux of full sunlight (up to ~2000 μmol photons·m⁻²·s⁻¹ at midday), and explains why grana stacking evolved: it is a space-efficient way to maximize light-harvesting membrane per unit organelle volume, analogous to why mitochondrial cristae fold the inner membrane rather than leaving it smooth.
Example 2 — A herbicide mechanism reasoned from compartment logic. The herbicide DCMU (diuron) is known to block electron flow specifically between PSII and plastoquinone, on the stromal-facing side of PSII. Predict what happens to (a) O2 evolution, (b) the transthylakoid pH gradient, and (c) CO2 fixation, in an illuminated leaf treated with DCMU. Reasoning through the compartments: PSII still absorbs light and attempts charge separation, but because the electron cannot be passed onward to plastoquinone, the oxygen-evolving complex's electron sink is blocked, so O2 evolution stops almost immediately (the manganese cluster cannot be re-reduced to continue turnover cleanly). Because cytochrome b6f no longer receives electrons from plastoquinone, its proton pumping into the lumen collapses, and PSII's direct lumenal proton release also stalls once turnover halts—so the ΔpH gradient dissipates, shutting down ATP synthase output. Without a supply of ATP and NADPH from the light reactions, the stroma's Calvin cycle rapidly exhausts its existing pools and CO2 fixation ceases within minutes, even though RuBisCO itself is not directly inhibited. This chain of reasoning—from a single molecular block at one membrane complex to whole-organelle metabolic arrest—is exactly why DCMU is used experimentally as a specific PSII inhibitor to dissect electron transport steps, and why herbicides targeting this site are agriculturally effective at killing weeds by starving them of ATP and NADPH.
Common misconceptions & traps
| Misconception | Correction |
|---|---|
| "The dark reactions happen at night." | They occur whenever ATP/NADPH are available (normally during daytime alongside light reactions); "dark" only means they don't directly require photons. |
| "Grana and thylakoids are different membrane systems." | Grana are simply stacks of thylakoid disks; stroma lamellae are unstacked thylakoid regions—all one continuous membrane and lumen. |
| "Chloroplasts have two membranes like mitochondria, and that's it." | Chloroplasts have two envelope membranes plus a third internal system, the thylakoid membrane—mitochondria lack this third membrane. |
| "Photosynthesis and respiration are simply reverse reactions happening in the same place." | They occur in different organelles (chloroplast vs. mitochondrion) with different, though structurally analogous, chemiosmotic machinery. |
| "RuBisCO fixes carbon using light energy directly." | RuBisCO uses ATP and NADPH generated by light reactions; the carboxylation step itself does not absorb photons. |
| "More chlorophyll always means faster photosynthesis." | Rate is often limited by CO2 supply, RuBisCO kinetics, or stomatal conductance, not by pigment quantity alone. |
The scientist behind it
Daniel Israel Arnon spent his career at UC Berkeley demonstrating, against real skepticism, that isolated chloroplasts retained the full biochemical machinery of photosynthesis outside a living cell—a claim requiring extraordinarily careful isolation technique, since damaged chloroplasts lose activity within minutes. His laboratory's insistence on rigorous biochemical fractionation, paired with radioisotope tracing of phosphate into ATP, turned photosynthesis from a physiological black box into a mapped set of molecular reactions assignable to specific organelle compartments.
Origins · 1 minDaniel I. Arnon and how this was discovered
Arnon's 1954 papers with Allen and Whatley established photophosphorylation and demonstrated that light and dark reactions of photosynthesis are experimentally separable, laying the biochemical foundation for compartment-specific models of chloroplast function.
Where this matters today
Understanding thylakoid and stroma compartmentalization underlies modern crop science, where engineering chloroplast transformation (chloroplast genetic engineering, exploiting the organelle's own genome) is used to introduce herbicide resistance, pest resistance (e.g., Bt toxin expression), and improved photosynthetic efficiency traits directly into the plastid genome, which is maternally inherited and reduces transgene spread via pollen—an important biocontainment feature in genetically modified crop regulation. In medicine and agriculture together, herbicides such as DCMU-class PSII inhibitors and paraquat (which intercepts electrons at PSI, generating damaging reactive oxygen species) are designed around precise knowledge of thylakoid electron transport chain architecture; resistance-conferring mutations in the psbA gene (encoding the D1 protein of PSII) are tracked in weed populations worldwide as a model of rapid evolution under strong selective pressure. In synthetic biology and bioenergy research, engineers are attempting to re-route or re-engineer thylakoid electron transport to boost biofuel and biohydrogen production, and to improve RuBisCO's notoriously poor carboxylation-to-oxygenation specificity (its wasteful oxygenase side-reaction, photorespiration, costs some C3 crops an estimated 20–30% of potential yield). Understanding the physical separation of light and carbon reactions is also foundational to explaining C4 photosynthesis, where mesophyll and bundle-sheath chloroplasts are structurally and biochemically dimorphic, concentrating CO2 around RuBisCO to suppress photorespiration—a specialization built entirely on the basic chloroplast architecture described here. Finally, this same three-membrane, endosymbiont-derived architecture is the structural evidence at the heart of The Endosymbiotic Theory: chloroplast DNA, 70S ribosomes, and double-envelope membranes closely parallel free-living cyanobacteria, a comparison this article's compartmental map sets up directly.
Ask the tutor: How does grana stacking create lateral heterogeneity between PSII and PSI within the same thylakoid membrane?
Why do chloroplasts have three membranes while mitochondria have only two?
Mitochondria have an outer and inner membrane, with cristae as infoldings of the same inner membrane—still fundamentally two membrane systems. Chloroplasts have that same outer/inner envelope pair, but additionally possess a third, internal membrane system: the thylakoids, forming their own enclosed lumen distinct from the stroma. This extra membrane exists because chloroplasts must physically separate light-driven electron transport and proton pumping (needing a dedicated membrane and an enclosed lumen to build a proton gradient) from the soluble carbon-fixing enzymes of the Calvin cycle. Mitochondria don't need this separation because oxidative phosphorylation's electron transport chain and ATP synthase are the only major energy pathway housed there; chloroplasts run two coupled but spatially distinct processes.
Where exactly does the Calvin cycle occur, and why not in the thylakoid lumen?
The Calvin cycle occurs in the stroma, the fluid matrix surrounding the thylakoids inside the inner envelope membrane. RuBisCO and all other Calvin cycle enzymes are soluble stromal proteins. The thylakoid lumen is unsuitable because it is acidified specifically to drive ATP synthase, and it lacks the enzymes and substrate access needed for carbon fixation; it also has a much smaller volume, poorly suited to housing a full enzyme cycle. The stroma, in contrast, directly receives the ATP and NADPH exported from the thylakoid membrane's stromal face, making it the logical (and evolutionarily selected) compartment for RuBisCO and its supporting cycle.
What is the difference between a granum and a thylakoid?
A thylakoid is a single flattened membrane sac; a granum is a stack of many thylakoid disks piled like coins, typically 10–20 per stack. All thylakoids—whether in grana or in the connecting stroma lamellae—share one continuous internal lumen and one continuous membrane, so a granum is a structural arrangement, not a separate compartment. Grana stacking increases the density of light-harvesting complexes and photosystem II per unit chloroplast volume, since PSII is preferentially localized to appressed (stacked) granal membranes, while PSI and ATP synthase are enriched in unstacked stroma lamellae, creating lateral heterogeneity within the single thylakoid membrane system.
How was it proven that chloroplasts, not some other cell structure, are the site of photosynthesis?
Classic cell fractionation combined with the Hill reaction was decisive: when leaf cells were gently disrupted and centrifuged, the chloroplast-containing fraction—identifiable by green color and confirmed by microscopy—retained the ability to evolve oxygen and, in Arnon's later work, to fix CO2 and synthesize ATP when illuminated, while other cellular fractions (nuclei, mitochondria, cytosol) did not. This localized both light and dark reactions to the isolated organelle itself, independent of any need for an intact cell, definitively assigning photosynthesis to the chloroplast rather than to some diffuse cytosolic process.
Does the thylakoid membrane's proton gradient work the same way as the mitochondrial one?
The underlying chemiosmotic principle is identical—protons are pumped across a membrane by an electron transport chain, and their controlled flow back through ATP synthase drives ATP synthesis, as Peter Mitchell proposed for both organelles. The key difference is geometry and direction: mitochondria pump protons from the matrix into the intermembrane space, and protons flow back into the matrix through ATP synthase, making ATP there. Chloroplasts pump protons from the stroma into the thylakoid lumen, and protons flow back into the stroma through ATP synthase, making ATP in the stroma. The chloroplast's ΔpH (about 3 units) is also typically steeper than the primarily membrane-potential-driven mitochondrial gradient.
Why are chloroplasts green, and does that relate to their internal structure?
Chlorophyll a and b, embedded in light-harvesting complexes across the thylakoid membrane, absorb strongly in blue (~430 nm) and red (~660 nm) wavelengths but reflect and transmit green light (~500–560 nm), which is why chloroplasts and the plant tissues containing them appear green to our eyes. This is directly tied to structure: pigments are anchored in specific protein complexes (LHCII around PSII, LHCI around PSI) positioned within the thylakoid membrane, and the extensive folding of that membrane into grana packs enormous pigment density into a small organelle volume, maximizing photon capture across the visible spectrum except in the green range.
How does chloroplast structure connect to the idea that they were once free-living bacteria?
Several structural features only make sense in light of an endosymbiotic origin: the double envelope membrane resembles a bacterium engulfed within a host vacuole membrane (later lost) plus its own original plasma membrane; the presence of circular DNA and 70S ribosomes (rather than the 80S ribosomes of the eukaryotic cytosol) mirrors bacterial, specifically cyanobacterial, genetics; and the thylakoid membrane itself closely resembles the photosynthetic membranes found in modern cyanobacteria. These structural parallels are the starting evidence base explored more fully under The Endosymbiotic Theory and connect chloroplast architecture to a specific evolutionary history rather than an arbitrary organelle design.