5.04: A. The Immune System - Antibodies, B- cells, T-cell receptors and T-cells

Search Fundamentals of Biochemistry

Learning Goals 

(Learning goals written by Claude, Sonnet 4.6, Anthropic)

Antibody Structure, Diversity, and Antigen Recognition

• Describe the domain architecture of IgG — two heavy and two light chains linked by disulfide bonds, each chain organized into variable (VH, VL) and constant (CH1–CH3, CL) immunoglobulin domains of ~110 amino acids each — identify the complementarity-determining regions (CDRs/hypervariable loops) within the variable domains as the structural basis of antigen specificity, and explain how the epitope-paratope interaction at the Fab tips of the Y-shaped antibody excludes water and buries predominantly apolar surface area to achieve high affinity.
• Explain how the combinatorial recombination of V, D, and J gene segments at the DNA level, imprecise joining at recombination junctions (adding P- and N-nucleotides), and somatic hypermutation collectively generate the extraordinary diversity of antibody variable regions — enabling recognition of virtually any antigen — and explain why higher-affinity antibodies produced later in the immune response bury more apolar surface area rather than more hydrogen bonds or salt bridges.
• Describe the immunoglobulin fold (IGF) as a structurally conserved β-sandwich domain found not only in antibodies and T-cell receptors but in ~2% of all proteins across all domains of life — including immune, muscle, nervous system, and cell-adhesion proteins — identifying its two antiparallel β-sheets connected by a conserved disulfide bond as the defining structural feature.

T Cells, MHC Proteins, and Antigen Presentation

• Explain why T cells require antigen presentation by MHC proteins rather than recognizing free antigens directly — distinguishing MHC Class I (present on virtually all body cells, display peptides from intracellular proteins including viral and tumor antigens, recognized by CD8⁺ cytotoxic T cells) from MHC Class II (present on antigen-presenting cells like macrophages and dendritic cells, display peptides from engulfed extracellular antigens, recognized by CD4⁺ T helper cells) — and use the "hot dog bun" analogy to explain how the T-cell receptor simultaneously contacts both the peptide and the MHC groove.
• Explain the two-signal requirement for T-cell activation — specific signal (T-cell receptor binding to peptide:MHC complex) plus nonspecific costimulatory signal (CD28 binding B7 on the antigen-presenting cell) plus cytokine signal — as a safety mechanism that prevents inadvertent autoimmune activation, and describe the physiological consequences of CD4⁺ (T helper, protection of extracellular/phagosomal spaces) vs. CD8⁺ (cytotoxic, elimination of infected or tumor cells) T-cell responses.

CAR-T Therapy and De Novo Antibody Design

• Describe CAR-T cell therapy — in which a patient's T cells are engineered to express a chimeric antigen receptor combining the single-chain variable fragment (scFv) of an antigen-specific antibody with T-cell receptor and co-stimulatory intracellular signaling domains — explain how the flexible glycine-serine linker enables scFv formation from VH and VL domains, and distinguish CAR-T from conventional adoptive T-cell transfer and traditional vaccine approaches in terms of mechanism, target, and clinical applicability.
• Explain how recent advances in computational protein design using fine-tuned RFdiffusion models enable the de novo design of antibody variable domains that bind user-specified epitopes with atomic-level precision — and critically note that experimental validation (yeast display screening, structural confirmation) remains essential — connecting this advance to the broader theme of AI-assisted protein design discussed in earlier chapters.

Introduction to the Immune System

Now, let's consider the daunting task faced by the immune system: to recognize all possible "foreign" molecules and respond to them, either by targeting them for elimination or, paradoxically, by recognizing them without responding (a process called tolerance). The same can be said of "self-molecules, and the immune system must recognize them but not respond to them. Otherwise, autoimmune disease might arise in which the body's powerful immune system targets itself.

It is impossible to describe the immune system in depth in a short section. Our goal is to illustrate how the immune system recognizes such a vast number of molecules. We will briefly cover the innate and adaptive immune systems, their differences, and how some cells (macrophages, in particular) in the innate immune system and cells (B and T cells) in the adaptive immune response recognize and respond to target molecules and cells. Finally, we'll discuss how the immune system can respond to similar molecules by recognizing common molecular patterns. Emphasis will be given to recognition. Ways to simplify the complexities of the immune system are presented in a fantastic book by Lauren Sompayrac, How the Immune System Works. (2003, Blackwell Publishing. ISBN: 0-632-04702-X) and adopted here.

We realize we have not yet reached the chapters on carbohydrates, membrane proteins, and nucleic acids. Nevertheless, we present the material in this section to organize it in one location. Users can revisit this page after studying subsequent chapters.

Before we start, think of the variety of chemical species that the immune system should recognize as foreign:

• a bacterial glycan or glycolipid on the outside of the cell
• a viral surface protein, such as the spike protein of the SARS-CoV-2
• bacterial dsDNA (and not host dsDNA)
• viral dsRNA (which is not common in host systems
• a self-protein that has been modified in a tumor cell
• a crystal of urea
• extracellular ATP (a place where it is not usually found)
• a silica particle found in particles like asbestos.

How would you design an immune system to bind each of the "enemy" targets above? That is what we will explore in this section - the binding interactions. What happens after the binding is beyond the scope of this section and falls in the field of signal transduction - how binding events at the cell surface are transferred into intracellular responses.

Three lines of defense protect us from the "enemies," foreign substances (bacteria, viruses, and their associated proteins, carbohydrates, and lipids) collectively called antigens.

• Physical barriers of cells that line our outside surface and our respiratory, GI tract, and reproductive systems.
• The innate immune system (IS) that all animals have. It is composed of scavenger cells like macrophages (MΦ), neutrophils, dendritic cells, and natural killer cells (NK) that can move around the body through the blood and lymph systems and burrow into tissues to meet the enemy where they can engulf and destroy bacteria and "cellular debris." Macrophages start as immature circulating monocytes, which enter tissues by slipping through blood vessel walls. They differentiate into macrophages. There they lie in waiting, ready for the enemy.
• The adaptive immune system, which, as its name implies, can change and adapt to new molecular threats. This branch is better at dealing with viruses, which cause damage within host cells. The adaptive IS consists of B cells that make and secrete protein antibodies that recognize specific foreign molecules and T cells.

In a world that has experienced the most deadly pandemic (COVID) of the last 100 years, and with more to come, immune recognition must be an important part of any biochemistry text. This chapter section could be a whole chapter, but we'll leave it as a very long section. Let's start with the adaptive immune system, which we can co-opt to make vaccines for our major threats.

B Cells and Antibodies

B cells and their differentiated forms (B memory and plasma cells) make antibodies. Antibodies bind to foreign molecules (proteins, glycans, lipids, etc), which might neutralize their effects. For example, an antibody can bind to the hemagglutinin molecule of the influenza virus and prevent its entry into cells. We are all familiar with the utility of vaccines that create antibodies to recognize the spike protein of the acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Antibodies also bind to foreign cells, such as bacteria, which signals other host immune proteins and cells to come in for the kill. Antibodies are secreted by B cells, which also have membrane-bound antibodies on their surface. This antibody acts as a receptor that binds antigens and, through signal transduction, helps activate the B cell. Mature B cells (those that have previously seen antigens) can secrete lots of antibodies quickly. Surface and secreted antibodies can recognize and bind to almost any molecule.

There are many forms of antibodies, also called immunoglobulins (Ig). These include IgA, IgG, IgM, and IgD. We will concentrate on the structure of IgG. It consists of 4 chains (a tetramer), two light chains, and two heavy chains. The light chains form disulfide links with the heavy chains, and disulfide bonds also link the heavy chains. Effectively, it's one big protein molecule (about 160 K). Figure \(\PageIndex{1}\) shows a spacefill, secondary structure, and geometric cartoon rendering of a mouse IgG protein (pdb ID 1IGT).

Figure \(\PageIndex{1}\): Renderings of an IgG antibody

The antibody is shaped as a Y. Foreign molecules (antigens) bind at the end of the top tips of the Y, with both chains contributing to antigen binding. Antiparallel beta sheets dominate both structures.

Each chain consists of a single N-terminal variable domain (V_L or V_H), which participates in antigen recognition. The light chains have an additional constant domain (C_L), while the heavy chains have three constant domains (C_H1-C_H3). The constant domains are not involved in antigen recognition. Still, they are involved in effector functions (such as binding other immune molecules, such as complement proteins) to the antigen-bound antibody heavy chain constant regions. Each domain is about 100 amino acids long. Figure \(\PageIndex{2}\) shows a cartoon showing the domain structures.

Figure \(\PageIndex{2}\): IgG antibody and fragments

Two other features are shown in the figure above. In each variable region of both light and heavy changes, there are hypervariable regions that contribute to the unique binding features of a given antibody. The regions are also called complementarity-determining regions (CDRs). Membrane-bound antibodies that serve as "receptor" proteins have additional domains, as shown in the figure above. The binding site on the antigen recognized by the antibody is called the epitope. The corresponding binding site on the Y-shaped ends of the antibody that recognizes the antigen is called the paratope.

---

Recently Updated:  11/3/24

In Figure \(\PageIndex{2}\), you can see that the intact, full IgG molecule has 12 variable V and constant C domains, each an example of an immunoglobulin domain (IGD) or immunoglobulin fold (IGF). Each has about 110 amino acids, two layers of β-sheets that face each other, with each sheet containing 3-5 antiparallel β-strands, and a disulfide bond connecting the two layers.  Beta-hairpins connect strands in a given sheet while "beta-arches" connect the two sheets. 

The immunoglobulin fold (IGF) and domain (IGD) are ubiquitously found in nature in all domains of life, and proteins with them are part of the Ig Superfamily (IgSF). It's estimated that about 2% of all proteins contain the domain.  They are found in the cytoplasm, membranes, and secreted proteins. The domain is often found on cell membrane proteins and the proteins that bind to them, making them part of a complex IGF "interactome". They are especially abundant in eukaryotes in systems (immune, muscle, and nervous, for example), for which intracellular and extracellular communications are critical. There are over 200 IGD in one of the largest human proteins, titin.  Figure \(\PageIndex{3}\) shows the topology of immunoglobulin-like domains.

Figure \(\PageIndex{3}\):  Topology of immunoglobulin-like domains.  Chidyausiku, T.M., Mendes, S.R., Klima, J.C. et al. De novo design of immunoglobulin-like domains. Nat Commun 13, 5661 (2022). https://doi.org/10.1038/s41467-022-33004-6.  Creative Commons Attribution 4.0 International License.  http://creativecommons.org/licenses/by/4.0/.

Panel a: Three-dimensional cartoon representation of an Ig structure formed by seven β-strands (left) and backbone hydrogen bond patterns (annotated thin lines) between paired β-strands along the sequence (right). Cross-β interactions have higher sequence separation (and higher contact order) than β-hairpins, slowing folding. 

Panel b: β-arches of the cross-β motif belong to two contiguous and distinct Greek key motifs: with two β-strands in each β-sheet (left) and with three β-strands in one β-sheet and one β-strand in the other (right). From the folding and design perspective, the main limiting factor for correctly assembling the Ig structure is the formation of the cross-β motif since the three β-hairpins can form independently.

IGDs in eukaryotes are divided into four main variants: IgI (I set), IgV (V set), IgC1 (C1 set), and IgC2 (C2 set), based on their immunoglobulin structure (e.g., V stands for variable).  They differ in topology and secondary structure but generally resemble two stacked β-sheets or a β-sheet sandwich barrel (with the sheets being the "bread").  A new structural and numbering scheme for the strands shows the Ig domain can contain nine strands, A, B, C, C’, C’’, D, E, F, and G (Caesar Tawfeeq et al.)

Figure \(\PageIndex{4}\) shows  interactive iCn3D models of the IgG2a monoclonal antibody (1IGT) shown in Figure 1 above. The left panel shows the Ig domains of just one variable chain (which contains two domains), while the right panel shows all the IgDs of the full antibody.  The Ig strands (A, B, C, C’, C’’, D, E, F, and G) are colored as shown below.

1IGT Light Chain. Click the image for a popup or use this external link: https://structure.ncbi.nlm.nih.gov/i...xpZS1JFPX99yeA

1IGT Full antibody.  Click the image for a popup or use this external link: https://structure.ncbi.nlm.nih.gov/i...zTPnVZhjmnAQa6

Figure \(\PageIndex{4}\): Ig domain structure of the IgG2a monoclonal antibody (1IGT) light chain (left) and complete antibody (right). (Copyright; author via source). 

---

When we discussed domain structure, we noted that proteins with multiple binding domains can often be selectively cleaved by proteases, with the resulting fragments often retaining binding and other functional properties. The same is true with antibodies. Cleavage by proteases such as pepsin or papain yields fragments with binding activity, as illustrated in the figure above. Selected protease digestion was used to clarify structure/function relationships in antibody recognition.

When antibodies targeting different antigens were sequenced, it was clear that substantial variability was observed in the variable domains of both the light and heavy chains. In those domains, there were also hypervariable regions. The variability and hypervariability originate primarily from the extremely large number of gene segments (also called exons) in the gene encoding the variable domains. The exons can be spliced together at the DNA and RNA levels to produce many different DNA/RNA sequences. These are decoded into the variable and hypervariable regions of the light and heavy chains of the proteins. Somatic mutations are also enhanced in this region.

In-depth: Generation of Antibody Diversity

As B cells mature into antibody-secreting cells (plasma cells), DNA and RNA splicing occur. Splicing for primary RNA transcripts should be no surprise for those who have studied the Central Dogma of Biology. Surprisingly, the DNA genome of B cells changes on their maturation due to the splicing of multiple exons within the variable chain genes to produce unique coding sequences for each B cell clone. There are sets of exons (V, D, and J) or segments within the genes for the variable chain. As the immune cells terminally differentiate, a unique combination of a VDJ segment forms in the DNA genome, so each terminally differentiated B cell is different. When needed (i.e., when their unique antigen binds to the membrane-bound form of the antibody), the cell secretes a monoclonal antibody.

Figure \(\PageIndex{5}\) shows how the different segments become linked in the DNA and can be uniquely spliced in the RNA to produce a unique, monoclonal antibody.

Figure \(\PageIndex{5}\) Generation of Antibody Diversity. Oliver Backhaus DOI: 10.5772/intechopen.72818. Creative Commons Attribution 3.0 License,

The first antibodies produced by the immune system are often of low affinity. Over time, high-affinity (low K_D) antibodies are produced. What differentiates high and low affinity binding at the molecular level? Do high-affinity interactions have many intramolecular H-bonds and salt bridges (ion-ion interactions), or are hydrophobic interactions most important? Crystal structures of many antibody-protein complexes were determined to study the basis of affinity maturation of antibody molecules. Clones of antibody-producing cells with higher affinity are selected by binding and clonal expansion. Investigators studied the crystal structures of four antibodies that bound to the same site (epitope) on the protein antigen lysozyme. Increased affinity was correlated with increased buried apolar surface area and not with increased numbers of H bonds or salt bridges, as described in Table \(\PageIndex{1}\) below.

Antibody	H26-HEL	H63-HEL	H10-HEL	H8-HEL
Kd (nM)	7.14	3.60	0.313	0.200
Noncovalent Interactions
H bonds	24	25	20	23
VDW contacts	159	144	134	153
salt bridges	1	1	1	1
Buried Surface Area
ΔASURF (A2)	1,812	1,825	1,824	1,872
ΔASURF-polar (A2)	1,149	1,101	1,075	1,052
ΔASURF-apolar (A2)	663	724	749	820

Table \(\PageIndex{1}\): Characteristics of Antibody:Hen Egg Lysozyme Complexes (HEL). Data from Y. et al. Nature: Structural Biology. 6, pg 484 (2003)

Many crystal structures of antibody:antigen complexes have been determined. Especially interesting are those in which the antigen is a protein. It is important to understand antibody:protein antigen interactions to develop vaccines against key epitopes in proteins such as the spike protein of SARS-CoV-2. Let's look at the antibody that binds to hen egg white lysozyme (HEWL). The crystal structures of many different IgG antibodies that bind HEWL are known. One recognizes a discontinuous epitope on lysosome consisting of the following amino acids: H15, G16, Y20, R21, T89, N93, K96, K97, I98, S100, D101, G102, W63, R73, and L75. Most of these amino acids are polar, and five are charged.

Figure \(\PageIndex{6}\) shows the interaction of part of the Fab fragment of an antibody that binds to the HEWL epitope just mentioned (3hfm). The light chain is shown in magenta, the heavy chain in dark blue, and the antigen lysozyme in gray. The amino acids' side chains in the HEWL epitope are shown in sticks. Note the complete complementarity of HEWL and Fab surfaces. Water is excluded from the interface.

Figure \(\PageIndex{6}\): Surface interactions between a hen egg white lysozyme epitope and an IgG Fab antibody fragment

Figure \(\PageIndex{7}\) shows an interactive iCn3D model of the same HEWL:Fab complex (3hfm). Lysozyme is shown in black.

Figure \(\PageIndex{7}\): Hen egg white lysozyme:Fab complex. (3hfm) (Copyright; author via source). Click the image for a popup or use this external link: https://structure.ncbi.nlm.nih.gov/i...agQkuAKhgfs177

Here is an external link to an interactive iCn3D model showing a detailed view of the multiple interactions (salt bridges, hydrogen bonds, pi-cation)

Abstract:  Despite the central role of antibodies in modern medicine, no method currently exists to design novel, epitope-specific antibodies entirely in silico. Instead, antibody discovery currently relies on immunization, random library screening, or the isolation of antibodies directly from patients^1. Here we demonstrate that combining computational protein design using a fine-tuned RFdiffusion^2 network with yeast display screening enables the de novo generation of antibody variable heavy chains (VHHs), single-chain variable fragments (scFvs), and full antibodies that bind to user-specified epitopes with atomic-level precision. Bennett, N.R., Watson, J.L., Ragotte, R.J. et al. Atomically accurate de novo design of antibodies with RFdiffusion. Nature 649, 183–193 (2026). https://doi.org/10.1038/s41586-025-09721-5.

T Cells

What happens if a virus makes it into a cell? Antibodies can not bind to them anymore to prevent their entry. Something must be able to recognize a virally infected cell and eliminate it. What about a cancer cell? Wouldn't it be nice if something could recognize a tumor cell as foreign and eliminate it before it divides too much and metastasizes? Those "somethings" are T cells. There are many T cells in a person and many different kinds, including T helper cells (Th), cytotoxic lymphocytes (CTL), and even suppressor T cells. They express distinct protein subsets that differentiate them and determine their functions.

T cells also recognize antigens, but unlike B cells, they recognize only protein fragments. The membrane proteins that recognize protein fragments are called T-cell receptors. In addition, they don't recognize protein antigens in isolation. They must be bound to a protein on the surface of an "antigen" presenting cell (such as a macrophage or dendritic cell). The T cell receptor recognizes and binds simultaneously to the foreign protein fragment and the self "antigen-presenting" protein on the surface of the antigen-presenting cells. The self-protein that binds and presents the foreign protein fragments (peptides) is called a Major Histocompatibility Complex (MHC) protein.

Antigen-presenting cells, such as macrophages and dendritic cells, express MHC Class II molecules on their surfaces. These bind protein fragments from engulfed bacteria (for example) and present them on the surface. T cell receptors bind to the peptide:MHC II complex. All cells in the body have MHC Class I proteins on their surface. If a cell is infected with a virus, viral protein fragments bind to the MHC Class I protein on the cell surface. Now, a T cell can bind through its T cell receptor to the peptide:MHC Class I complex. By displaying a viral protein fragment on its surface, the immune cell can recognize a virally infected cell without entering the cell where the virus is. Sompayrac describes MHC molecules as looking like a hot dog bun. In the groove of the bun lies the peptide fragment, like the hot dog. The T cell receptor recognizes both the bun and the hot dog!

Figure \(\PageIndex{8}\) shows an interactive iCn3D model of an MHC Class I heavy chain complexed with a peptide fragment (i.e., the antigen) of the vesicular stomatitis virus nucleoprotein (2VAA). 

Figure \(\PageIndex{8}\): MHC Class I heavy chain - vesicular stomatitis virus nucleoprotein peptide (antigen) complex (2VAA). (Copyright; author via source). Click the image for a popup or use this external link https://structure.ncbi.nlm.nih.gov/i...NPS9QfLeghCmo8

 

In-depth: Generation of T-Cell Receptor Diversity

We described above how an undifferentiated B cell can produce a wide diversity of antibodies from a single genetic sequence. This occurs through both DNA and RNA splicing. The same processes occur with the alpha and beta chains of T-cell receptors. This is illustrated in Figure \(\PageIndex{9}\). Note that the alpha chains have no D (diversity) coding sequences.

Figure \(\PageIndex{9}\): The diversity of T-cell receptor (TCR)αβ is a result of genetic recombination and diversification mechanisms occurring at the α and β TCR chain loci. Diversity is first created in the germline via recombination of variable V, diversity D (for β chain), and joining J segments. Further diversification occurs through imprecise junctions of these gene segments (addition of P- and N-nucleotides adjacent to the D segment), and the combination of α and β chains

Molecular T-Cell Repertoire Analysis as Source of Prognostic and Predictive Biomarkers for Checkpoint Blockade Immunotherapy. International Journal of Molecular Sciences 21(7):2378 (2020). DOI: 10.3390/ijms21072378. License CC BY

Figure \(\PageIndex{10}\) shows an interactive iCn3D model of the T-cell receptor alpha and beta chains binding to MHC Class 1 protein with a bound peptide (6rp9). The MHC protein complex consists of the histocompatibility antigen A-2 alpha chain and β-2-microglobulin, an 11K subunit of MHC Class I proteins, but not Class II MHC proteins. Bound to it is the nine-amino acid cancer/testis antigen 1 (shown in spacefill). The peptide is sandwiched between the MHC protein complex and the T-cell receptor α and β chains.

Figure \(\PageIndex{10}\): T-cell receptor alpha and beta chains binding to MHC Class 1 protein with a bound peptide (6rp9) (Copyright; author via source). Click the image for a popup or use this external link: https://structure.ncbi.nlm.nih.gov/i...txUWWfbNhNXwQ6

The T-cell receptor consists of two transmembrane protein chains, alpha and beta, each containing a single variable and constant Ig domain, followed by a transmembrane domain. Hence, they are less complicated than an antibody chain. They bind through their extracellular variable domains to a peptide fragment bound to an MHC Class I or Class II membrane protein on the target cell.  The alpha and beta chains of the T-cell receptor, the HLA class I chain, and the associated beta-2 microglobulin are comprised of Ig domains, just like antibodies.  Figure \(\PageIndex{11}\) below shows the Ig domains color-coded to show their A-G beta strands. The MHC Class 1 protein domain that actually binds the peptide antigen (shown in spacefill) is shown in cyan.

Figure \(\PageIndex{11}\): T-cell receptor alpha and beta chains binding to MHC Class 1 protein with a bound peptide (6rp9)

The actual in vivo functional structure is more complicated. The T-cell receptor is found within the much larger T-cell receptor complex (TRC), which contains two copies of the CD3 complex, which itself consists of γ, δ, ε, and ζ chains, as shown in Figure \(\PageIndex{12}\) (A). Part A shows the variable and C domains of the α and β chains of the T-cell receptor in green and dark. The rest of the T-cell receptor complex includes two copies each of the CD3 complex, which consists of one copy of εδζ chains and one copy of εγζ chains.

Figure \(\PageIndex{12}\): T cell receptor structure. Kumaresan Pappanaicken R., da Silva Thiago Aparecido, Kontoyiannis Dimitrios P. Methods of Controlling Invasive Fungal Infections Using CD8+ T Cells. Frontiers in Immunology, 8, 1939 (2018). https://www.frontiersin.org/article/...mmu.2017.01939. DOI=10.3389/fimmu.2017.01939. Creative Commons Attribution License (CC BY).

As mentioned earlier, one function of MHC Class I molecules is to present peptides derived from tumor antigens to T cells. This leads to the activation of other immune cells and, hopefully, the destruction of tumor cells expressing the tumor antigen. Much work has gone into studying immune surveillance and the ultimate destruction of tumor cells to improve our immune response to cancer cells. In early work, T-cells that had infiltrated tumors were isolated from a patient, amplified in the lab by adding a cytokine (e.g., interleukin 2, IL2), a protein growth factor released by activated immune cells, and then reinfused the tumor-specific T cells along with IL2 back into the patient. This adoptive cell transfer (ACT) therapy led to remissions in some patients, but the therapy also could be lethal.

One promising type of immune therapy is chimeric antigen receptor (CAR) T cell therapy (CAR T), in which patients are treated with modified versions of their T-cells. T cells are removed from a cancer patient's blood. A gene is constructed to mimic the V and C domains of the alpha and beta chains of the T-cell receptor and inserted into the patient's T-cell using a viral vector. The gene construct contains, as its tumor antigen-binding motif, the V and C domains of an antibody gene designed to recognize the tumor antigen. The receptor is, hence, a chimeric (formed from parts of different proteins) antigen receptor (CAR) that combines antibody and T cell components. The genetically modified cells are amplified and reinfused back into the patient. Once the collected T cells have been engineered to express the antigen-specific CAR, they are "expanded" in vitro into the hundreds of millions.  This approach is very expensive, so recent successes in using a virus to deliver the appropriate gene construct without removing cells from patients are noteworthy. 

Compare the structure of the chimeric antigen receptor (CAR) in Figure XX-C with the normal T-cell receptors shown in A. The CAR contains two single-chain variable fragments (scFv) derived from the variable domains of the light (VL) and heavy (VH) chains of an antibody that recognizes the tumor cell. These domains are connected by a linker peptide (10-25 amino acids) enriched in glycines, which confers flexibility, and by serines/threonines for hydrogen-bonding interactions. This is attached to an F_C fragment and other intracellular effector domains to create the receptor. Figure \(\PageIndex{13}\) shows the scFv structure. We'll discuss adding the cytoplasmic CD28 domain in a bit.

Figure \(\PageIndex{13}\): Single-chain variable fragment structure used in CART

You can imagine this whole T-cell receptor complex involved in the binding of a tumor peptide antigen presented by a MHC I transmembrane protein on a tumor cell, as illustrated in Figure \(\PageIndex{14}\) (in different colors). The entire interacting structure is called the T cell immunological synapse.

Figure \(\PageIndex{14}\): T cell immunological synapse of T cell with a cancer cell. Zhao Lijun, Cao Yu J. Engineered T Cell Therapy for Cancer in the Clinic. Frontiers in Immunology, 10, 2250 (2019) https://www.frontiersin.org/article/...mmu.2019.02250. Creative Commons Attribution License (CC BY).

Protection Against Autoimmune Recognition - Coreceptors

How can the immune system recognize and bind to any foreign molecule but not self-molecules? The subject of immune tolerance is too specialized to include here, but we will discuss a few features.

The MHC Class I proteins present "self" peptides in their binding pockets. Self-proteins are also cleaved into peptides by proteasomes within the cell. However, the T cell receptor does not recognize and bind to the self-peptide fragment bound to the MHC Class 1 protein. Hence, T cells do not recognize self and turn against their own cells. Once in a while, they do, however, and autoimmune diseases like MS, rheumatoid arthritis, and lupus result.

B cells and T cells must be activated before they can respond. It is important to regulate the "on" switch. If the cells were activated without need, they might turn against the self. In addition to T cell receptor complex binding to foreign peptides, MHC complexes for immune cell activation must bind another protein on the antigen-presenting cell.

In the case of T helper cells, the T cell protein CD28 must also bind the B7 protein on an antigen-presenting cell, like a macrophage expressing an MHC II protein:foreign peptide complex. Hence, there is one specific signal (the peptide:MHC complex binding to the T cell receptor complex) and a nonspecific signal (B7 binding CD28). Why are two signals needed for activation? Again, Sompayrac has a great analogy. A safety deposit box at a bank takes two keys: a specific key (which you have) and a "nonspecific" key (which the bank uses for all boxes) to open the box. Think of it as double security. You don't want to activate immune cells to kill unless you need to.

CD4 and CD8 T cells

Yet other proteins are involved to ensure correct T-cell activation. We'll consider T cells expressing either CD4 or CD8. T-cells expressing these expand after antigen stimulation (infection or immunization). It depends on the T-cell subtype. Let's consider two here:

T cells expressing the protein CD4: After initial stimulation, they differentiate and proliferate into the T helper cells named T_H1, if they produce the cytokine interferon (IFN)-γ, and T_H2, if they produce the cytokine IL4. CD4 is an integral membrane protein that acts as a co-receptor for the MHC Class II:peptide complex on cells such as macrophages. These cells present foreign antigens and antigens derived from microbes taken up into phagosomes by antigen-presenting cells, such as macrophages.  Hence, they protect phagosomal spaces, similar to the role of antibodies, which bind antigens in extracellular spaces and protect against them.  Traditional vaccines targeting foreign proteins, such as the spike protein of the SARS-CoV-2 virus, mainly elicit antibodies that bind antigens in the blood and other extracellular spaces.  Such vaccines are ineffective against antigens such as the malaria parasite in the blood, since the parasite circulates for only an hour before it is internalized by liver cells.

T-cells expressing CD8: These cells produce cytokines (IFN-γ and tumor necrosis factor (TNF)-α) or secrete proteins, which form pore-forming complexes on foreign cells, leading to the lysis of cells such as pathogens or tumor cells. The CD8 protein has an alpha and a beta subunit. They serve as co-receptors for MHC Class I:peptide complex found on tumor cells, for example. MHC Class I proteins are found on most body cells (see below).  Cytotoxic T cells (a type of T cell) express CD8.  Vaccines that elicit a robust CD8 response might be more effective in the defense against intracellular pathogens that cause malaria, tuberculosis, and acquired immune deficiency syndrome (AIDS).    Memory CD8 cells protect against intracellular infections, so efforts are underway to develop CD8 T-cell vaccines.

Figure \(\PageIndex{15}\) shows the multiple co-signals that are required to activate the CD4 T-cell (blue sphere), which has the T-cell receptor complex, the co-receptor CD4, and the CD28 protein. It also displays a cytokine receptor that binds cytokines released by antigen-presenting cells (macrophages shown in pink). This leads to the proliferation and differentiation of activated T cells.

Figure \(\PageIndex{15}\): Co-signals that are required to activate the CD4 T-cell (blue sphere) Salmonella as a Model for Non-Cognate Th1 Cell Stimulation. Frontiers in Immunology 5(621):621 (2014) DOI: 10.3389/fimmu.2014.00621. CC BY 4.0

Sompayrac asks another interesting question. Why is antigen presentation by MHC proteins necessary at all? B cells don't need presentation, since they can bind antigens via membrane-bound antibody molecules. Why do T cells need it? He gives different reasons for Class I and Class II presentations:

Class I MHC (found on most body cells): T cells need to be able to "see" what is going on inside the cell. When virally infected cells bind foreign peptide fragments and present them on the surface, they can be "seen" by the appropriate T cell. It's a way to get a part of the virus, for example, to the surface. They can't hide out in the cell. T cells don't need to recognize extracellular threats since antibodies from B cells can do that. Presentation is also important, since viral protein fragments found outside the cell might bind to the outer surface of a noninfected cell, targeting it for killing by the immune system. That wouldn't be good. It also helps that peptide fragments are presented on the surface. This allows parts of the protein that are buried and not exposed on the surface, which would be hidden from interaction with outside antibodies, to be used in signaling cell infection by a virus.

MHC Class II (found on antigen-presenting cells like macrophages): Two different cells (the presenting cell and the T helper cell) must interact for a signal for immune system activation to be delivered to the body. Again, it is a safety mechanism that prevents nonspecific activation of immune cells. Also, as in the case above, since fragments are presented, more of the foreign "protein" can contribute to the signal that activates the immune system.

Summary

(Summary written by Claude, Sonnet 4.6, Anthropic)

This chapter applies the principles of protein structure, noncovalent binding, and molecular recognition established throughout the course to the immune system — arguably the most sophisticated molecular recognition system in biology, capable of specifically detecting and responding to virtually any foreign molecule while tolerating self.

The adaptive immune system is organized around two cell types — B cells and T cells — each with a distinct strategy for antigen recognition. B cells produce antibodies (immunoglobulins), soluble or membrane-bound proteins capable of binding almost any antigen directly in the extracellular space. The canonical IgG antibody is a Y-shaped disulfide-linked tetramer of two heavy chains (each with VH, CH1, CH2, and CH3 domains) and two light chains (each with VL and CL domains). Antigen recognition occurs at the tips of the Y, where the VH and VL domains from each heavy-light pair cooperate to form the paratope that contacts the antigen's epitope. Within these variable domains, three hypervariable loops per chain — the complementarity-determining regions (CDRs) — make the majority of contacts with antigen. Structural analysis of antibody-lysozyme complexes at increasing affinity reveals that affinity maturation — the progressive increase in antibody affinity during an immune response — correlates predominantly with increased buried apolar (hydrophobic) surface area rather than with more hydrogen bonds or salt bridges, confirming that the hydrophobic effect is the dominant thermodynamic contributor to high-affinity protein-protein recognition. The generation of antibody diversity sufficient to recognize any antigen is achieved through three mechanisms: combinatorial recombination of V, D, and J gene segments at the DNA level during B cell development (producing a unique antibody-encoding sequence in each B cell clone), imprecise junction formation at recombination boundaries (adding or deleting nucleotides to create additional sequence diversity), and somatic hypermutation of variable region genes in antigen-stimulated B cells (further diversifying affinity). The immunoglobulin domain (IGD) — a β-sandwich of two antiparallel β-sheets (3–5 strands each) connected by a conserved disulfide bond and organized into four structural subtypes (IgI, IgV, IgC1, IgC2) — is the structural unit of antibody chains and is one of the most abundant protein domain families in eukaryotic proteomes, found in immune, muscle, nervous system, and cell-surface recognition proteins across all domains of life.

T cells recognize antigens differently from B cells: they can only respond to short peptide fragments (8–25 amino acids) displayed on the surface of other cells by Major Histocompatibility Complex (MHC) proteins. MHC Class I proteins, found on virtually all nucleated body cells, present peptides derived from intracellular proteins — including viral proteins and tumor antigens — in a groove on the cell surface; these complexes are recognized by CD8⁺ cytotoxic T cells through their T-cell receptor (TCR). MHC Class II proteins, found only on dedicated antigen-presenting cells (macrophages, dendritic cells, B cells), present peptides from proteins engulfed from the extracellular environment; these complexes activate CD4⁺ T helper cells. In both cases, the TCR — a heterodimer of α and β transmembrane chains, each with variable and constant Ig domains — simultaneously contacts the peptide and the MHC protein surface (the "hot dog" and the "bun"). TCR diversity is generated by the same VDJ recombination and junctional diversification mechanisms used for antibodies. The complete signaling unit is the T-cell receptor complex, which includes the TCR α/β heterodimer plus multiple CD3 signaling subunits (γδε and εγζ). Critically, T-cell activation requires two signals: the antigen-specific signal (TCR binding to peptide:MHC) and a costimulatory signal (CD28 on the T cell binding B7 on the antigen-presenting cell), plus cytokine signals — a multi-key safety mechanism that prevents inadvertent autoimmune activation. CD4⁺ T helper cells (TH1 and TH2 subtypes) orchestrate immune responses against extracellular and phagosomal threats; CD8⁺ cytotoxic T cells directly kill virally infected and tumor cells by recognizing peptide:MHC Class I complexes. Developing CD8⁺ memory responses is an active vaccine design priority for intracellular pathogens including malaria, tuberculosis, and HIV.

CAR-T cell therapy represents a powerful application of these recognition principles to cancer immunotherapy. A patient's T cells are engineered to express a chimeric antigen receptor — a synthetic fusion protein combining the single-chain variable fragment (scFv) of a tumor antigen-specific antibody (VH and VL domains joined by a flexible Gly-Ser linker) with transmembrane and CD3 signaling domains and costimulatory domains (such as CD28) that eliminate the need for a second costimulatory signal. This allows the engineered T cell to recognize and kill tumor cells expressing the antigen, bypassing the normal two-signal requirement. Recent advances extend antibody engineering further: de novo computational design using fine-tuned RFdiffusion models now enables the generation of antibody variable domains that bind user-specified epitopes with atomic-level precision, followed by experimental validation through yeast display screening — connecting the computational protein design revolution described in earlier chapters directly to the most medically impactful class of therapeutic proteins.