What is Dr. Zubair Khalid's research focus?

Dr. Zubair Khalid specializes in molecular virology, mRNA vaccine development, and computational biology, with a focus on avian pathogens like IBDV and Avian Reovirus.

Where is Dr. Zubair Khalid currently working?

Dr. Zubair Khalid is a Postdoctoral Research Associate at the University of Maryland (UMD), specifically within the Department of Animal and Avian Sciences.

Chaperonin Proteins: Structure, Function, and Mechanism

By Dr. Zubair Khalid, DVM, MS, PhD · 2026-08-11

Introduction to Chaperonin Proteins

Protein folding is the process by which a linear polypeptide chain acquires its native three-dimensional structure. While the primary amino acid sequence contains the thermodynamic information required for folding, the crowded intracellular environment—with protein concentrations reaching 300–400 mg/mL—poses a severe challenge. Unfolded or partially folded proteins expose hydrophobic surfaces that are prone to aggregation, a state that is both cytotoxic and energetically costly to reverse. To manage this problem, cells have evolved a diverse arsenal of molecular chaperones: proteins that bind to non-native polypeptides and facilitate their folding without becoming part of the final structure.

Chaperonins are a specific subclass of molecular chaperones defined by their distinctive architecture: a large, barrel-shaped complex composed of two stacked rings, each enclosing a central cavity. Unlike Hsp70 and Hsp90 systems, which bind client proteins on their surfaces and act in an ATP-regulated cycle of binding and release, chaperonins physically enclose their substrates within an isolated chamber. This encapsulation provides a protected environment where a single polypeptide can fold in isolation, free from the risk of aggregation with other cellular components. The term "chaperonin" was introduced in the late 1980s to distinguish these large double-ring complexes from other chaperone families, and the name is reserved for the Hsp60 (heat shock protein 60) family and its eukaryotic counterpart TRiC/CCT.

Chaperonins are found in all domains of life. In bacteria, the paradigmatic chaperonin is GroEL, which works with its co-chaperonin GroES. In eukaryotes, mitochondria and chloroplasts contain GroEL homologs, while the cytosol contains the more complex chaperonin TRiC (TCP-1 ring complex), also called CCT (chaperonin containing TCP-1). Archaea possess thermosome, a Group II chaperonin. The essential nature of these proteins is underscored by the fact that GroEL is essential for E. coli viability, and TRiC is essential for the viability of all eukaryotic cells. Chaperonins are not merely passive folding chambers; they are ATP-driven molecular machines whose conformational changes are tightly coupled to substrate folding.

Structure of Chaperonins

All chaperonins share a conserved core architecture: two rings of subunits stacked back-to-back, forming a cylindrical complex with a central cavity. Each ring typically contains seven to nine subunits, and the two rings are arranged with their cavities facing inward. The overall molecular mass of a chaperonin complex ranges from approximately 800 kDa to nearly 1 MDa. The subunits themselves are approximately 55–60 kDa and share a conserved three-domain architecture: an equatorial domain, an intermediate domain, and an apical domain.

The equatorial domain contains the ATP-binding site and mediates most of the inter-ring and intra-ring contacts. It is the most conserved region of the subunit. The apical domain forms the opening of the central cavity and contains the hydrophobic residues that interact with substrate proteins. The intermediate domain connects the equatorial and apical domains and transmits conformational changes between them. ATP binding and hydrolysis in the equatorial domain induce movements in the intermediate domain that are amplified into large rigid-body rotations of the apical domain, opening or closing the cavity.

The central cavity is the site of protein folding. Its interior surface is lined with hydrophobic residues in the unoccupied state, which is critical for substrate capture. Upon ATP binding and lid association, the cavity surface becomes more hydrophilic, creating an environment that promotes the burial of hydrophobic residues within the folding protein—a condition that mimics the dilute aqueous environment of the test tube but with the added benefit of isolation.

Group I Chaperonins (GroEL/GroES)

Group I chaperonins are found in bacteria, mitochondria, and chloroplasts. The best-characterized member is E. coli GroEL, a tetradecamer of 14 identical subunits arranged in two heptameric rings. Each subunit is 57 kDa, giving the complex a molecular mass of approximately 800 kDa. GroEL functions with a co-chaperonin, GroES, a heptameric ring of 10 kDa subunits that acts as a lid, capping the central cavity.

The GroEL structure can be described as a "double doughnut." Each ring has a central cavity approximately 45 Å in diameter in the absence of nucleotide. The apical domains line the opening of the cavity and present a ring of hydrophobic residues—including Leu, Val, and Phe side chains—that bind to exposed hydrophobic patches on unfolded substrate proteins. The equatorial domains form the floor of the cavity and contain the ATP-binding pockets, which are occupied by ATP in a cooperative manner. The two rings of GroEL communicate allosterically: ATP binding in one ring stimulates substrate release and GroES binding in that ring, while simultaneously promoting the release of GroES and folded substrate from the opposite ring.

GroES is a dome-shaped heptamer that binds to the apical domains of GroEL in the presence of ATP. The inner surface of GroES is hydrophilic and contributes to the character of the folding cavity. Together, the GroEL ring and GroES lid enclose a volume of approximately 85,000 Å³, sufficient to accommodate proteins up to about 60 kDa in size. The GroEL/GroES complex is often referred to as the "Anfinsen cage" in honor of Christian Anfinsen, whose work established that the amino acid sequence determines protein structure.

Group II Chaperonins (TRiC/CCT)

Group II chaperonins are found in the eukaryotic cytosol and in archaea. Unlike Group I chaperonins, they do not require a separate co-chaperonin lid. Instead, the apical domains contain a built-in helical protrusion that acts as a lid, closing the cavity in response to ATP binding. This structural difference means that Group II chaperonins undergo a more subtle conformational change during their functional cycle, with the lid elements moving as a iris-like aperture.

The eukaryotic chaperonin TRiC/CCT is considerably more complex than GroEL. It is composed of eight distinct subunit types (CCT1 through CCT8, encoded by genes TCP1, CCT2, CCT3, CCT4, CCT5, CCT6, CCT7, and CCT8 in humans), each present in one copy per ring. The two rings are arranged in a back-to-back fashion, giving a hetero-oligomeric complex of 16 subunits with a molecular mass of approximately 900 kDa. The subunit diversity is functionally significant: different subunits present different amino acid side chains to the central cavity, allowing TRiC to recognize a broader range of substrates than the homo-oligomeric GroEL.

TRiC is responsible for folding approximately 10% of cytosolic proteins in eukaryotes, including the cytoskeletal proteins actin and tubulin, which are obligate substrates. It also folds proteins involved in cell-cycle regulation, chromatin remodeling, and signal transduction. The requirement for eight different subunits means that TRiC assembly is itself a complex process, and mutations in individual CCT subunits are associated with specific disease phenotypes, including sensory neuropathy and male infertility.

The ATP-Driven Folding Cycle

The chaperonin folding cycle is a precisely orchestrated sequence of conformational changes driven by ATP binding and hydrolysis. The cycle can be divided into four stages: substrate binding, encapsulation, folding, and release. The entire cycle for GroEL takes approximately 15–30 seconds at 37°C, depending on the substrate and ATP concentration.

Substrate Binding

The cycle begins when an unfolded or partially folded protein binds to the hydrophobic apical domains of an open GroEL ring. This binding is mediated by the exposure of hydrophobic side chains on the substrate, which are normally buried in the protein's interior. In the crowded cellular environment, these hydrophobic surfaces would otherwise drive aggregation. GroEL captures these non-native species with high affinity, with dissociation constants in the nanomolar range. The binding is promiscuous with respect to sequence but selective for the conformational state: GroEL binds molten globule-like states and other partially folded intermediates, but not native proteins or fully unfolded polypeptides.

The substrate can bind to either ring of GroEL, but binding to one ring is sufficient to initiate the cycle. In the absence of nucleotide, GroEL exists predominantly in a "tense" state with high affinity for substrate. ATP binding to the equatorial domains of the same ring induces a cooperative conformational change that weakens substrate affinity and prepares the ring for GroES binding.

Encapsulation by the Lid

ATP binding to all seven subunits of a GroEL ring is cooperative: the binding of ATP to one subunit increases the affinity of neighboring subunits for ATP. This cooperativity ensures that the ring acts as a unit, avoiding partial transitions. Once ATP is bound to all seven subunits, the apical domains rotate upward and outward, and GroES binds to the now-exposed surface. GroES binding is accompanied by a large conformational change in which the apical domains move approximately 60 Å, and the central cavity doubles in volume.

The substrate protein, which was bound to the apical domains, is displaced into the central cavity as GroES caps the ring. This process is often described as "forced unfolding" because the conformational change in the apical domains can actively stretch or unfold parts of the substrate that are in stable local structures. The result is that the substrate is released into the cavity in a more unfolded state than when it entered, giving it a fresh start for folding.

The opposite ring, which is not bound to GroES, undergoes a reciprocal conformational change. ATP binding in the substrate ring is communicated through the equatorial domains to the opposite ring, which is induced to release its GroES and any folded substrate it contains. This negative cooperativity between rings ensures that only one ring is in the "folding-active" state at any time.

Folding Inside the Cage

Inside the encapsulated cavity, the substrate protein is free to fold in isolation. The cavity is large enough to accommodate proteins up to approximately 60 kDa in the case of GroEL, and the interior surface is now hydrophilic, which promotes the burial of hydrophobic residues in the folding protein. The folding process inside the cage is fundamentally the same as folding in dilute solution: the protein explores its conformational landscape, seeking the native state. However, the cage provides two critical advantages. First, it prevents aggregation, because the substrate is physically isolated from other proteins. Second, the confined volume can, in some cases, accelerate folding by stabilizing compact intermediates (the "confinement effect").

The ATP molecules bound to the equatorial domains are hydrolyzed to ADP over a period of approximately 10–15 seconds. This hydrolysis is not synchronized across the ring; rather, it occurs stochastically, with each subunit hydrolyzing its ATP independently. The timing of hydrolysis sets the "residence time" of the substrate in the cage. The hydrolysis of ATP to ADP reduces the affinity of GroES for the apical domains, priming the complex for disassembly.

Release of Folded Protein

The cycle is completed when ATP binds to the opposite ring. This binding is communicated through the equatorial domains and triggers the dissociation of GroES and the release of the substrate from the first ring. If the substrate has folded to its native state, it is released into the cytosol, where it can perform its biological function. If the substrate has not yet reached its native state, it is released in a non-native form and can rebind to GroEL for another round of folding.

The release step is driven by the binding of ATP to the trans ring, not by the hydrolysis of ATP in the cis ring. This ensures that the cycle is directional and that the system does not waste energy. The overall cycle consumes 14 ATP molecules per ring per round (7 for binding and 7 for hydrolysis), plus 7 more for the opposite ring when it becomes the active ring. This energy expenditure is the price the cell pays for ensuring that protein folding proceeds without aggregation.

Substrate Recognition and Specificity

Chaperonins recognize their substrates through a combination of structural features rather than specific amino acid sequences. The primary recognition element is the exposure of hydrophobic surface area. In native proteins, hydrophobic side chains are buried in the core, away from water. In unfolded or partially folded states, these residues are exposed and available for interaction with the hydrophobic patches on the chaperonin apical domains.

The apical domains of GroEL present a ring of hydrophobic residues—including Leu2, Val3, and Phe4 in the helix H region—that form a binding surface for substrate proteins. This surface is complementary to the hydrophobic patches exposed on non-native proteins. The binding is of moderate affinity, allowing the substrate to be released into the cavity upon ATP binding. The affinity is tuned so that GroEL binds non-native proteins with micromolar to nanomolar affinity but does not bind native proteins, which lack exposed hydrophobic surfaces.

GroEL is remarkably promiscuous in its substrate range. In E. coli, GroEL interacts with approximately 250–300 different proteins, representing about 10% of the proteome under normal growth conditions. These substrates range in size from 10 to 60 kDa and include proteins with diverse folds and functions. Many GroEL substrates are proteins that fold slowly or have complex topologies that make them prone to aggregation. Under conditions of stress, such as heat shock, the number of GroEL substrates increases dramatically as more proteins become unfolded.

TRiC/CCT, by contrast, has a more restricted substrate range, reflecting its more complex subunit composition. The eight different subunits of TRiC present different amino acid side chains to the cavity, creating a mosaic surface that recognizes specific structural features. TRiC substrates include actin and tubulin, which are obligate substrates, as well as proteins involved in cell-cycle control (such as Cdc20 and Cdh1), chromatin remodeling (such as histone deacetylases), and signaling (such as G-protein β subunits). The substrate specificity of TRiC is determined in part by the electrostatic character of the cavity, which is more hydrophilic than that of GroEL, and by the specific geometry of the apical domain protrusions.

Chaperonins in Health and Disease

Given their essential role in protein folding, it is not surprising that chaperonin dysfunction is associated with a range of human diseases. These conditions, collectively termed chaperonopathies, arise from mutations in chaperonin genes or from alterations in chaperonin expression or function.

Mutations in the HSPD1 gene, which encodes the mitochondrial chaperonin Hsp60, cause an autosomal dominant form of hereditary spastic paraplegia (SPG13) and a recessive form of hypomyelinating leukodystrophy. These disorders are characterized by progressive degeneration of motor neurons and white matter in the central nervous system. The disease-causing mutations, such as V72I and D29G, impair the ATPase activity or the assembly of the chaperonin complex, leading to mitochondrial dysfunction and impaired folding of mitochondrial proteins.

Mutations in HSPE1, encoding the mitochondrial co-chaperonin Hsp10, are rarer but have been reported in patients with mitochondrial encephalopathy. The clinical presentation includes developmental delay, seizures, and lactic acidosis, consistent with impaired mitochondrial protein folding.

In the cytosol, mutations in individual CCT subunits have been linked to specific diseases. Mutations in CCT5 cause hereditary sensory neuropathy, a condition characterized by loss of pain and temperature sensation in the distal limbs. The mechanism involves impaired folding of tubulin and consequent defects in microtubule dynamics in sensory neurons. Mutations in CCT4 are associated with male infertility, likely due to defects in spermatogenesis that require functional TRiC.

Chaperonin overexpression is a feature of many cancers. Cancer cells have high rates of protein synthesis and are under chronic proteotoxic stress due to mutations that produce misfolded proteins. Upregulation of TRiC subunits, particularly CCT2 and CCT3, has been observed in hepatocellular carcinoma, breast cancer, and colorectal cancer. The increased chaperonin capacity supports the survival of cancer cells by preventing the accumulation of toxic protein aggregates. This has made chaperonins an attractive target for cancer therapy; inhibitors of TRiC ATPase activity are being explored as potential anticancer agents.

Chaperonins are also implicated in neurodegenerative diseases, including Alzheimer's, Parkinson's, and Huntington's diseases. These disorders are characterized by the aggregation of misfolded proteins into amyloid fibrils and inclusion bodies. Chaperonins can, in principle, suppress aggregation by binding to aggregation-prone intermediates and promoting their folding. However, the capacity of the chaperonin system can be overwhelmed by the high concentration of aggregation-prone proteins in these diseases. In some cases, chaperonins may even contribute to pathology by stabilizing toxic oligomeric species.

Experimental Methods to Study Chaperonins

The study of chaperonin structure and function has relied on a combination of biophysical, biochemical, and structural techniques. Each method provides complementary information about different aspects of the chaperonin mechanism.

X-ray crystallography has provided the highest-resolution views of chaperonin structure. The first crystal structure of GroEL was determined in 1994 at 2.8 Å resolution, revealing the domain architecture and the arrangement of subunits in the double-ring complex. Subsequent structures of GroEL with bound nucleotide, with GroES, and with substrate proteins have illuminated the conformational changes that occur during the folding cycle. The major limitation of crystallography is the requirement for well-ordered crystals, which has made it difficult to capture the dynamic, heterogeneous states of the chaperonin cycle.

Cryo-electron microscopy (cryo-EM) has emerged as a powerful complementary technique, particularly for large, dynamic complexes. Single-particle cryo-EM has been used to determine structures of GroEL-GroES complexes in different nucleotide states and of TRiC in its open and closed conformations. The advantage of cryo-EM is that it can capture multiple conformational states from a single sample, providing a movie-like view of the conformational landscape. Recent advances in detector technology and image processing have pushed cryo-EM resolutions to near-atomic levels (2–4 Å), rivaling crystallography for many complexes.

ATPase assays are essential for characterizing the enzymatic activity of chaperonins. The ATPase activity of GroEL can be measured using a coupled enzyme assay in which ADP production is linked to the oxidation of NADH, which is monitored spectrophotometrically at 340 nm. Typical assays are performed at 25–37°C in a buffer containing 50 mM Tris-HCl (pH 7.5), 10 mM MgCl₂, 10 mM KCl, and 1 mM ATP. The ATPase activity of GroEL is approximately 0.1–0.2 µmol ATP hydrolyzed per minute per mg of protein, and this activity is stimulated 2- to 4-fold by the addition of GroES and substrate protein.

Single-molecule studies have provided unique insights into the dynamics of the chaperonin cycle. Fluorescence resonance energy transfer (FRET) between labeled GroEL subunits has been used to monitor the conformational changes of individual complexes in real time. These studies have revealed that the GroEL ring can populate multiple conformational states and that the timing of ATP hydrolysis is stochastic. Optical trapping experiments have measured the forces generated by GroEL during substrate encapsulation, demonstrating that the chaperonin can actively unfold stable protein structures.

Mass spectrometry has been used to identify chaperonin substrates and to characterize the stoichiometry of chaperonin complexes. Native mass spectrometry can preserve non-covalent interactions and has been used to determine the mass and subunit composition of intact chaperonin complexes. Hydrogen-deuterium exchange mass spectrometry provides information about the conformational dynamics of chaperonins and their substrates. For a broader discussion of how mass spectrometry is applied to protein analysis, see Mass Spectrometry Work for Proteins and Mass Spec Identify Proteins. Quantitative approaches to measuring chaperonin levels in cells rely on methods described in Quantify Proteins and Quantitative Determination of Proteins.

Common Misconceptions and Pitfalls

Students frequently encounter several conceptual difficulties when studying chaperonins. Clarifying these points is essential for a correct understanding of the system.

Confusing chaperones with chaperonins. All chaperonins are chaperones, but not all chaperones are chaperonins. Chaperones are a broad class of proteins that assist in protein folding, including Hsp70, Hsp90, and the small heat shock proteins. Chaperonins are specifically the large double-ring complexes (GroEL, TRiC, thermosome) that encapsulate their substrates. The distinction matters because the mechanism of action is fundamentally different: chaperones bind and release substrates on their surface, while chaperonins enclose them in a cage.

Thinking chaperonins dictate folding information. Chaperonins do not contain information about the native structure of their substrates. The folding information is entirely contained in the amino acid sequence of the substrate protein. Chaperonins provide an environment in which folding can occur without aggregation, but they do not instruct the protein how to fold. This is demonstrated by the fact that many proteins can fold correctly in dilute solution without any chaperonin, and that chaperonins can fold proteins that are not their natural substrates.

Assuming ATP hydrolysis is not required for folding. ATP hydrolysis is essential for the chaperonin cycle, but it is important to understand what the energy is used for. ATP binding and hydrolysis drive the conformational changes that open and close the cavity, and they provide directionality to the cycle. The energy is not used to actively fold the substrate; rather, it is used to reset the machine for another round of substrate binding and encapsulation. The folding itself is driven by the intrinsic thermodynamic properties of the polypeptide chain.

Believing the cage is a passive container. The chaperonin cavity is not simply a passive box. The interior surface of the cavity changes character during the cycle, from hydrophobic (substrate-binding) to hydrophilic (folding-promoting). This change is functionally important: it promotes the burial of hydrophobic residues in the substrate and prevents the substrate from rebinding to the cavity walls. The confinement effect, in which the limited volume of the cavity stabilizes compact folding intermediates, also contributes to folding efficiency.

Overlooking the allosteric communication between rings. The two rings of a chaperonin do not function independently. They communicate through the equatorial domains, such that the state of one ring influences the state of the other. This inter-ring communication ensures that only one ring is active at a time and provides the directionality of the cycle. Students often think of the two rings as independent, but they are functionally coupled.

Assuming all chaperonins are the same. Group I and Group II chaperonins differ in their subunit composition, their requirement for a co-chaperonin, and their substrate specificity. GroEL is a homo-oligomer that requires GroES, while TRiC is a hetero-oligomer with a built-in lid. These differences have functional consequences: TRiC can fold a different range of substrates and does not require a separate lid protein.

Summary and Key Takeaways

Chaperonins are essential molecular machines that facilitate protein folding by providing an isolated environment in which unfolded proteins can fold without aggregating. Their double-ring architecture, ATP-driven conformational changes, and substrate recognition mechanisms are conserved across all domains of life. The study of chaperonins has provided fundamental insights into the principles of protein folding and the cellular strategies for managing proteostasis.

Frequently Asked Questions

Is chaperonin a protein?

Yes, a chaperonin is a protein. Specifically, it is a large, multi-subunit protein complex composed of multiple copies of chaperonin subunits. GroEL, for example, is a complex of 14 identical protein subunits. Chaperonins are themselves gene products that are synthesized on ribosomes and fold into their functional quaternary structure.

What are chaperonin proteins?

Chaperonin proteins are a class of molecular chaperones characterized by a double-ring, barrel-like architecture. They function by encapsulating unfolded or partially folded substrate proteins within a central cavity, where folding can occur in isolation from the cellular environment. The two major groups are Group I chaperonins (e.g., GroEL in bacteria, Hsp60 in mitochondria) and Group II chaperonins (e.g., TRiC/CCT in eukaryotic cytosol, thermosome in archaea).

How do chaperonins differ from other chaperones?

Chaperonins differ from other chaperones in their mechanism of action. Hsp70 and Hsp90 chaperones bind to hydrophobic patches on unfolded proteins and use ATP to drive cycles of binding and release, but they do not enclose their substrates. Chaperonins, by contrast, physically encapsulate their substrates within a large central cavity, providing a protected folding environment. This encapsulation is the defining feature of chaperonins.

What is the role of ATP in chaperonin function?

ATP plays two roles in the chaperonin cycle. First, ATP binding drives the conformational changes that allow substrate encapsulation: it weakens substrate affinity, promotes lid (GroES or built-in lid) binding, and enlarges the central cavity. Second, ATP hydrolysis provides directionality to the cycle, setting the residence time of the substrate in the cage and priming the complex for disassembly. The energy from ATP hydrolysis is not used to fold the substrate directly; it is used to reset the machine.

What is the structure of a chaperonin?

A chaperonin is a large complex of approximately 800–900 kDa, composed of two stacked rings of subunits. Each ring contains seven to nine subunits, and each subunit has three domains: equatorial (ATP binding and inter-ring contacts), intermediate (signal transduction), and apical (substrate binding and cavity opening). The two rings enclose a central cavity where protein folding occurs. Group I chaperonins require a separate co-chaperonin lid (GroES), while Group II chaperonins have a built-in lid.

Why are chaperonins important in the cell?

Chaperonins are essential for cell viability because they prevent protein aggregation and promote the folding of a significant fraction of the proteome. In E. coli, GroEL is essential for survival, and in eukaryotes, TRiC is essential for all cells. Chaperonins are particularly important under stress conditions, such as heat shock, when protein unfolding is increased. They also play critical roles in the folding of specific proteins, such as actin and tubulin in eukaryotes.

What happens if chaperonins malfunction?

Chaperonin malfunction leads to the accumulation of misfolded and aggregated proteins, which is cytotoxic. In humans, mutations in chaperonin genes cause specific diseases, including hereditary spastic paraplegia (Hsp60 mutations), hypomyelinating leukodystrophy, and hereditary sensory neuropathy (CCT5 mutations). Chaperonin dysfunction is also implicated in cancer, where overexpression supports tumor cell survival, and in neurodegenerative diseases, where chaperonin capacity is overwhelmed by aggregation-prone proteins.

Key Takeaways

Chaperonins are large, double-ring protein complexes (GroEL in bacteria, TRiC in eukaryotes) that encapsulate unfolded proteins in a central cavity to promote folding.
The chaperonin structure consists of stacked rings of subunits, each with equatorial, intermediate, and apical domains; the apical domains line the cavity and bind substrates.
The ATP-driven cycle involves substrate binding, encapsulation by a lid (GroES for Group I, built-in protrusions for Group II), folding in the cavity, and release driven by ATP binding to the opposite ring.
Chaperonins recognize substrates by exposed hydrophobic surfaces, not by specific sequences; GroEL is promiscuous, while TRiC has a more restricted substrate range.
Chaperonin dysfunction causes chaperonopathies, including hereditary spastic paraplegia and sensory neuropathy, and is linked to cancer and neurodegeneration.
Key experimental methods include X-ray crystallography, cryo-EM, ATPase assays, and single-molecule FRET; mass spectrometry is used for substrate identification and complex characterization.
Chaperonins do not encode folding information; they provide a protected environment where the amino acid sequence dictates the native structure.