Proteins. Protein domains. Protein families

At a glance

1. A protein is a chain of amino acids joined by peptide bonds, from its free amino end (N-terminus) to its free carboxyl end (C-terminus). Twenty amino acids are encoded by the standard genetic code.
2. Its structure has four levels: the sequence (primary), local α-helices and β-sheets held by backbone hydrogen bonds (secondary), the fold of the whole chain (tertiary), and the joining of several chains (quaternary).
3. The sequence holds the information for the fold: water-avoiding side chains end up inside, charged and polar ones outside.
4. A domain is a part of a chain that folds on its own, usually 40 to 350 amino acids long, often with one job. Large proteins are built from several, and the same domain turns up in many proteins.
5. Proteins that descend from one ancestral gene form a family, such as the globins or the serine proteases. Their folds stay alike long after their sequences have drifted apart.

01Amino acids and the peptide bond

Every amino acid has a central carbon (the α-carbon) carrying an amino group, a carboxyl group, a hydrogen atom and a side chain, the R group. The side chain is what makes each of the twenty different, and how it behaves in water decides where it ends up in the folded protein.

Glycine, whose side chain is a single hydrogen, lets the chain bend sharply; proline, whose side chain loops back onto the backbone, makes it stiff. Tables differ over cysteine, tyrosine, tryptophan and histidine, which sit between the groups, and some put phenylalanine, tyrosine and tryptophan in a group of their own, the aromatic amino acids. All but glycine come in two mirror-image forms, and proteins are built from the L-forms only.

A peptide bond joins the carboxyl group of one amino acid to the amino group of the next, giving off a molecule of water. The bond is flat and rigid, so the chain bends only at the α-carbons between the peptide bonds. A chain is read from its N-terminus to its C-terminus, the direction in which the ribosome makes it.

encoded by the standard genetic code

20amino acidsencoded by the standard genetic code

per turn of an α-helix

3.6amino acidsper turn of an α-helix

in most domains

40 to 350amino acidsin most domains

that code for proteins in the human genome, about

19,500genesthat code for proteins in the human genome, about

02Four levels of structure

In an α-helix the backbone coils, and the C=O of each amino acid forms a hydrogen bond with the N–H of the amino acid four places further on; the side chains point outwards. In a β-sheet, stretches of chain lie side by side, running the same way (parallel) or opposite ways (antiparallel), with hydrogen bonds between neighbouring strands. Loops and turns join these pieces. A few helices or strands in a set arrangement, such as the helix–turn–helix that binds DNA, form a motif; unlike a domain, a motif does not fold on its own.

The fold of the whole chain is driven mainly by water: side chains that avoid it gather in the core, where they also pack tightly by van der Waals contacts, while charged and polar side chains stay on the surface. Disulfide bonds, covalent links between two cysteines, add strength, mostly in proteins that work outside the cell, such as antibodies and insulin; the cytosol is too reducing for most of them to last.

Not all of a protein has to fold. Many proteins, most of all in signalling and gene regulation, have stretches with no fixed shape, intrinsically disordered regions; some fold only when they bind a partner, others work unfolded.

Denaturation undoes the fold but leaves the sequence: heat, extremes of pH or chemicals such as urea break the weak bonds, and the protein stops working. Some proteins fold back once the conditions return, many do not, as when egg white sets on cooking. In the cell, chaperones help new chains fold, and misfolded chains can clump into the amyloid of some diseases (Proteins after translation).

Two broad shapes are worth knowing. Fibrous proteins are long and build structures: collagen, three chains wound into a triple helix with glycine at every third position, or the keratins of hair and skin. Globular proteins are compact, soluble and do most of the cell's chemical work: enzymes, haemoglobin, antibodies.

03Protein domains

A large protein is rarely one continuous fold. Its chain is divided into domains, parts that each fold on their own into a compact, stable structure, usually 40 to 350 amino acids long, joined by short, flexible stretches of chain. Each domain often has its own job, so a protein's domains read like a list of what it can do.

The same domains turn up again and again, in different combinations:

• the kinase domain moves a phosphate from ATP onto a protein;
• the SH2 domain binds a phosphorylated tyrosine, the SH3 domain a short stretch rich in proline;
• the immunoglobulin domain, a sandwich of two β-sheets, builds antibodies and many receptors on the cell surface;
• the zinc finger, a small fold held by a zinc ion, binds DNA in many transcription factors.

Src, a kinase that relays signals inside the cell, carries an SH3, an SH2 and a kinase domain in one chain: two to find its partners and one to act on them.

New proteins have often been made by joining domains that already existed. In the genes of animals, a domain is often encoded by its own exon or run of exons, so recombination within introns can move a domain into a new gene, exon shuffling. Tissue plasminogen activator, the enzyme that starts the breakdown of blood clots, is built this way: a finger domain like those of fibronectin, a domain like epidermal growth factor, two kringle domains and a serine protease domain.

04Protein families

When a gene is duplicated, both copies can be kept and change separately, one taking on a new job. The proteins that descend from one ancestral gene form a family. Their sequences drift apart over time, but their folds change much more slowly. Relatives whose sequences still show the link form a family; more distant ones, whose sequences have drifted too far but whose structure and function still point to one ancestor, form a superfamily. A shared fold alone does not prove a shared ancestor, as some folds have arisen more than once.

• The globins. Myoglobin, which stores oxygen in muscle, and the α and β chains of haemoglobin, which carry it in blood, share one fold, α-helices packed round a haem group: eight in myoglobin and the β chain, seven in the α chain, which lacks the short D helix.
• The serine proteases. Trypsin, chymotrypsin and elastase, digestive enzymes of the pancreas, share one fold and one set of three catalytic amino acids, a serine, a histidine and an aspartate. Each cuts proteins after different amino acids because the pocket that holds the substrate's side chain differs: trypsin after lysine and arginine, chymotrypsin after large water-avoiding amino acids, elastase after small ones. The clotting factors thrombin and factor Xa belong to the family too.
• The protein kinases. About 500 human genes encode protein kinases, 518 in one count; about 480 of them share one kinase domain and form a superfamily of many families.
• The G protein-coupled receptors. The largest group of receptors in the human genome, a superfamily of about 800 genes in five families, each receptor a chain crossing the membrane seven times.

Members of a family in one species, such as the α- and β-globins, are paralogues; the same gene in two species, such as human and mouse β-globin, are orthologues.

Check yourselfAnswer, then say how sure you were: a lucky guess still counts as a weak point.

1What holds an α-helix in shape?

Disulfide bonds between cysteinesHydrogen bonds between the backbone C=O of one amino acid and the N–H of the amino acid four places alongIonic bonds between acidic and basic side chainsPeptide bonds between side chains

2Anfinsen unfolded ribonuclease with urea and a reducing agent, then removed them. The enzyme folded back and worked again. What did this show?

Enzymes need chaperones to foldDisulfide bonds alone decide a protein's shapeThe amino acid sequence holds the information for the protein's foldDenaturation breaks peptide bonds

3Which describes a protein domain?

Any stretch of α-helix in a proteinA separate polypeptide chain in a protein of several subunitsThe region of an enzyme that binds its substrateA part of a polypeptide chain that folds on its own into a compact structure, often with a function of its own

4Myoglobin and the α and β chains of haemoglobin have different sequences but the same fold of α-helices round a haem. Why?

They descend from one ancestral gene that was duplicated and then divergedThey are made from the same mRNA by alternative splicingThey are cut from one precursor proteinAny protein that binds haem must take this fold

5What is the change in the haemoglobin of sickle cell disease?

Valine in place of glutamate at position 6 of the β chainA missing β chainAn extra α chain in each moleculeHaem without its iron