Antibodies are Y-shaped immunoglobulin proteins (~150 kDa) built from two identical heavy chains and two identical light chains. Disulfide bonds link the chains, maintaining the Y-shaped structure.
They consist of four polypeptide chains that form a symmetric structure with two identical arms and a stem. Each arm ends in a pair of variable domains (VH and VL) that together form an antigen-binding site, enabling a single antibody to engage two identical epitopes simultaneously.
This bivalent arrangement greatly increases binding strength. A hinge between the arms and stem provides flexibility, allowing the arms to orient to targets at different distances and illustrating how structure supports function.
Heavy and Light Chains
An antibody monomer contains two heavy chains and two light chains. The heavy (H) chains(50–55 kDa) of an antibody consist of one variable domain (VH) and multiple constant domains (CH). In a typical IgG, each heavy chain contains three constant domains (CH1, CH2, CH3) plus a flexible hinge region. Heavy chains from each half of the antibody pair via disulfide bonds at the hinge giving the molecule its “Y” shape.
Light chains weigh roughly 25 kDa and appear in kappa (κ) or lambda (λ) forms. Each one carries a variable domain (VL) followed by a constant domain (CL). Each antibody carries two identical light chains, each folding alongside a heavy chain to complete the antigen-binding site, with its constant region pairing with the heavy chain’s CH1 domain to stabilize the Fab arm.
Each V domain (VH or VL) is ~110 amino acids, and each C domain is similar in size.
Functional Domains: Fab and Fc
An antibody’s Y-shaped architecture can be functionally divided into two arms and a stem, corresponding to the Fab and Fc domains. Each arm is a fragment antigen-binding (Fab) – a unit consisting of the N-terminal portion of one heavy chain (VH and CH1) and one complete light chain (VL and CL) together, forming a complete antigen-binding site at the tip.
Each Fab includes the VH–VL pair (collectively called the Fv region) that contacts the antigen, as well as the adjoining constant domains (CH1 and CL) that support this interaction. The two Fab fragments are identical, so a single antibody is bivalent and can bind two identical epitopes at once. Importantly, the Fab’s sole function is to bind the antigen and it has no intrinsic ability to trigger immune elimination on its own.
The stem of the Y is the fragment crystallizable (Fc) region, composed entirely of heavy-chain constant portions. The Fc is formed by pairing of the two heavy chains’ CH2 and CH3 domains, held together by interchain disulfide bonds in the hinge region.These paired constant domains form a rigid, relatively invariant base that supports the two arms.
The hinge region is a flexible linker between Fab and Fc that allows the two arms to pivot and adapt to different distances and angles when binding antigens
Once the Fab arms lock on to a target, the Fc region recruits immune effectors. For instance,the CH2 domain near the hinge binds complement component C1q to start the classical pathway, while residues in the lower hinge and neighbouring CH2 engage Fcγ receptors on phagocytes and natural killer cells. Further down the tail, the CH3 domain interacts mainly with the neonatal Fc receptor (FcRn) for antibody recycling and with staphylococcal Protein A, a feature widely used in antibody purification.
Variable and Constant Regions
The variable regions (VH on the heavy chain, VL on the light chain) lie at the antibody’s arm tips and are responsible for antigen recognition. Each V domain holds three hyper-variable loops called complementarity-determining regions (CDRs). Together, the six CDRs from one VH–VL pair build a single paratope, the surface that recognizes a specific part of the antigen.
When an antibody binds its target, it is the unique shape and chemistry of its VH–VL interface that ensure tight and specific binding.
In contrast, the constant regions of an antibody are more conserved, forming a stable scaffold and linking the antibody to the immune system’s effector machinery. The heavy chain’s constant portion (CH2 and CH3 in IgG) is responsible for immune effector functions. For example, once the antibody has bound its target, these domains recruit complement proteins and engage Fc receptors on immune cells. These interactions trigger processes like complement activation, phagocytosis, and cell-mediated cytotoxicity.
Heavy Chain Isotypes
Antibody isotype is determined by the heavy chain constant-region sequence. Humans have five isotypes (IgG, IgA, IgM, IgD, IgE), corresponding to the γ, α, μ, δ, and ε heavy chains. These heavy chains differ in the number of constant domains and the hinge region:
- IgG (gamma) heavy chain has three constant domains (CH1–CH3) and a hinge region.
- IgA (alpha) heavy chain has three constant domains and a hinge (IgA1 has a long hinge, IgA2 has a shorter protease-resistant hinge, and secretory IgA is usually dimeric).
- IgM (mu) heavy chain has four constant domains (CH1–CH4) and no classical hinge.
- IgD (delta) heavy chain has three constant domains and a hinge, similar to IgG.
- IgE (epsilon) heavy chain has four constant domains (CH1–CH4) and a short hinge-like linker.
The key structural differences are the presence of three versus four constant domains and the length of the hinge region.
Nanobodies and AI-Guided Antibody Design
Class and domain diversity have inspired new antibody formats. Notably, camelids produce heavy-chain only IgG (subclasses IgG2 and IgG3 that lack the CH1 domain and any light chain. Their single-domain variable regions (VHH) , “nanobodies” are ~12–15 kDa proteins with an extended CDR3 loop that can bind hidden or convex epitopes. This minimal architecture imparts high stability and solubility, and nanobodies can be engineered as monovalent or multivalent constructs. Understanding canonical Ig fold geometry has enabled the design of such smaller binders and even human single-domain variants.
Recent deep learning models can predict antibody structures from sequence and even generate novel antibody (or nanobody) sequences optimized for binding and developability. The platforms learn from vast antibody datasets to directly suggest optimized heavy/light chain sequences based on learned structure–function patterns. These AI approaches leverage the same principles of variable versus constant domain constraints, CDR geometry, and flexibility that underlie traditional antibody classes to create next generation therapeutics.