The Four Levels of Protein Structure, and Which Bonds Hold Each One

What this guide covers

A protein starts out as a line and ends up as a shape. Everything between those two states is what biochemists mean by the levels of protein structure.

In plain terms: primary structure is the order of the amino acids, the way letters make a word. Secondary structure is the local coils and zigzags the chain settles into. Tertiary structure is how the whole chain packs itself into a blob with a definite shape. Quaternary structure is what you get when several of those blobs lock together to work as one unit. Each level is held by a different kind of bond, and each level can be read in the laboratory by a different instrument.

This guide walks through the four levels with real proteins as examples, lists the bonds behind each one, explains why a short research peptide usually never gets past the first level, and matches each level to the method that measures it. For the geometry of the chain itself, the guide to the polypeptide backbone is the companion page.

What is primary structure?

Primary structure is the number and sequence of amino acids in a protein's polypeptide chain or chains, held together by peptide bonds. It is the only level that is covalent from end to end. A general chemistry text gives insulin, with 51 amino acids, as its example, and the gene encoding a protein ultimately sets that sequence.

Hemoglobin makes the point with numbers. UniProt's entry for the human alpha chain (P69905) runs 142 residues including an initiator methionine that is marked as removed, leaving 141 in the mature chain; the beta chain (P68871) runs 147 including a removed initiator methionine, leaving 146. Those subtractions are our arithmetic from the UniProt features. A full hemoglobin molecule has two of each chain, so 2 × 141 + 2 × 146 = 574 residues, which is the figure a biology textbook rounds to about 600.

One change in that sequence can change everything above it. In sickle cell hemoglobin a glutamate at position six of the beta chain is replaced by valine. The textbook describes it as one amino acid out of roughly 600, caused by a single base change in the gene, and the consequence is a protein that assembles differently. That is why primary structure is the level every other level depends on.

Primary structure is also the level a mass spectrometer reads. The molecular mass follows from the sequence, and tandem mass spectrometry breaks a peptide along its backbone so the order of residues can be read from the fragment masses. Steen and Mann describe peptide sequencing by mass spectrometry as the key technology of proteomics. The guide to the peptide sequence covers how a sequence is written and how mass confirms it.

What is secondary structure?

Secondary structure is the local, regular arrangement of the polypeptide backbone, held by hydrogen bonds between backbone groups rather than between side chains. The two common forms are the alpha helix and the beta pleated sheet. Pauling, Corey and Branson proposed two hydrogen-bonded helical configurations of the polypeptide chain in a 1951 paper, and the alpha helix is the one found throughout proteins.

In an alpha helix the chain winds into a right-handed spiral. Each turn holds 3.6 amino acid residues, and the hydrogen bonds run between the carbonyl oxygen of one residue and the amide hydrogen of the residue four positions along the chain, with the side chains pointing outward. Alpha-keratin, the protein of hair and wool, is almost entirely alpha-helical.

In a beta pleated sheet, two or more stretches of chain lie side by side, either parallel or antiparallel, and hydrogen bonds between the backbones of neighboring strands hold the sheet together. Silk fibroin is the classic example, and the chemistry text also lists carboxypeptidase A and lysozyme as proteins with beta sheet regions. Some proteins have little helix at all: gamma globulin, chymotrypsin and cytochrome c are named as examples, while hemoglobin and myoglobin are helical in some regions only.

Secondary structure is the level circular dichroism reads. Greenfield's Nature Protocols method scans the far-ultraviolet region, down to 185 nm or lower, and reports that alpha-helical proteins show negative bands at 222 and 208 nm with a positive band at 193 nm, while well-defined antiparallel beta sheets show a negative band near 218 nm and a positive band near 195 nm. Disordered chains show very low signal above 210 nm and a negative band near 195 nm. CD gives the overall share of each secondary structure in a sample, from 20 micrograms or less, and the paper is explicit that it does not give the residue-by-residue detail that X-ray crystallography or NMR can.

What is tertiary structure?

Tertiary structure is the overall three-dimensional shape of one folded chain, produced by the backbone bending and folding back on itself so that side chains far apart in the sequence come into contact. Where secondary structure is a backbone matter, tertiary structure is mostly a side-chain matter.

The chemistry text lists four interactions that hold it. Ionic bonds form between oppositely charged side chains, such as an aspartate carboxylate next to a lysine ammonium. Hydrogen bonds form between polar side chains. Disulfide linkages form when two cysteine side chains are oxidized to cystine, and the biology text notes this is the only covalent bond that forms during folding. Dispersion forces, the weak attractions between temporary dipoles, matter most where the other three are scarce. The hydrophobic effect sits underneath all of it: water's own hydrogen bonding pushes nonpolar side chains into the interior and leaves polar ones on the surface.

Myoglobin is the standard single-chain example. UniProt's human entry (P02144) lists 154 residues with the initiator methionine removed, so 153 in the mature protein, folded into one globular unit with a heme group. Insulin shows the disulfide contribution. UniProt's entry (P01308) places the B chain at residues 25 to 54 and the A chain at 90 to 110 of the precursor, which is 30 and 21 residues, and records two interchain disulfides (31 to 96 and 43 to 109) plus one within the A chain (95 to 100). Those three bonds are what hold the two chains of insulin in their folded arrangement.

Tertiary structure is the level that needs an atomic-resolution method. The RCSB's guide to structure determination says most structures in the Protein Data Bank were solved by X-ray crystallography, which requires a purified protein grown into a crystal and tends to leave flexible regions invisible. NMR spectroscopy works on the protein in solution and gives an ensemble of models, but is limited to small and medium proteins because the spectra of large ones overlap. Cryo-electron microscopy images many thousands of single particles in a thin layer of ice, and its annual depositions passed NMR's in 2016.

What is quaternary structure?

Quaternary structure is the arrangement of two or more folded polypeptide chains, called subunits, into one working assembly. The chemistry text says it is produced and stabilized by the same kinds of interactions as tertiary structure, which means ionic bonds, hydrogen bonds, disulfides and the hydrophobic effect acting between chains rather than within one. Not every protein has a quaternary level. A single-chain protein such as myoglobin stops at tertiary.

Hemoglobin is the example every textbook reaches for, and for good reason. The RCSB's Molecule of the Month entry describes it as four protein chains, two alpha and two beta, each carrying a ring-shaped heme group with an iron atom where oxygen binds, and each chain structurally similar to myoglobin. The deoxygenated human form is PDB entry 2hhb and the oxygenated form is 1hho; Max Perutz's horse hemoglobin is 2dhb.

What quaternary structure adds is communication between subunits. The RCSB page explains that once the first heme binds oxygen it introduces small changes in that chain, and those changes nudge the neighboring chains into a shape that binds oxygen more easily. The motion spreads through the assembly as what the page calls a large rocking motion of the subunits. A single chain cannot do that. The behavior exists only at the fourth level.

The sickle cell substitution shows the levels interacting in the other direction. One residue changes at the primary level, in the beta chain, and the result shows up at the quaternary level and beyond, as hemoglobin molecules that assemble with each other in a way the normal protein does not.

Which bonds hold each level?

The levels are easiest to keep apart by asking what holds each one together.

One pattern stands out. Only the primary level is covalent all the way through. The disulfide bond is the single covalent contribution above that, and everything else is a weak interaction that gains strength from numbers. That is why heat, pH and solvents can take a protein apart while leaving its sequence untouched, which is the subject of the guide to protein denaturation. The biology text makes the same point: denaturation is often reversible because the primary structure survives it.

Do short peptides have tertiary structure?

Usually not. A chain of a few residues has too few backbone groups to build the hydrogen-bond network of a helix and too few side chains to bury a hydrophobic core. A 2014 review of the unfolded state puts it directly: short peptides are generally unable to form the stable hydrogen bonding and long-range interactions needed to fold into well-defined secondary structures.

That does not mean a short peptide is shapeless. The same review reports that short alanine peptides spend most of their time in a restricted region of backbone angles called polyproline II, with measured fractions of about 0.74 for the alanine dipeptide and 0.84 for trialanine, and that other residues divide between polyproline II and the extended beta-strand region. These are transient preferences in solution, not a fold. The review's own term for the older picture, in which each residue samples every allowed angle at random, is the local random coil model, and it argues the data no longer support that picture either.

For a research peptide this has a practical meaning. Most of the compounds on a research-peptide catalog are chains of a few to a few dozen residues, and what defines them is the primary level: the sequence, the mass that follows from it, and any modifications at the ends. The guide to polypeptide versus protein draws the line between a chain that folds and one that does not, and the guide to amino acids versus protein covers where the scale changes.

Which laboratory methods read each level?

Each level has an instrument that sees it best, and knowing which one answers which question keeps a reader from asking a method for something it cannot give.

1. Primary structure: mass spectrometry. The intact mass checks the composition against the sequence, and tandem mass spectrometry reads the residue order from backbone fragments. This is the identity test on a peptide certificate.
2. Secondary structure: circular dichroism in the far ultraviolet. It reports the overall share of helix, sheet and disordered chain from a few micrograms, but not which residues are in which.
3. Tertiary structure: X-ray crystallography for a crystal, NMR for a small or medium protein in solution, or cryo-electron microscopy for particles frozen in ice. These give atomic coordinates.
4. Quaternary structure: the same three methods, which capture every chain in the assembly at once, as the hemoglobin entries 2hhb and 1hho do.

Every certificate LabFirst Peptides publishes reports identity by LC-MS, purity by RP-HPLC on a C18 column and net content as a separate line, each by lot number, and those three results all describe the primary level: the mass the sequence predicts, how much of the sample is that one compound, and how much of the vial's weight is peptide. The guide to HPLC purity versus peptide identity explains why mass and purity are separate questions, and the certificate index lists the lots with a posted certificate.

Bottom line

Primary structure is the sequence, covalent and set by the gene. Secondary structure is the helices and sheets the backbone's own hydrogen bonds make. Tertiary structure is the fold of one chain, held by side-chain interactions and the hydrophobic effect, with disulfides as the one covalent addition. Quaternary structure is several folded chains working as one, and hemoglobin's four chains are the standard example. Mass spectrometry reads the first level, circular dichroism the second, and crystallography, NMR or cryo-EM the third and fourth. A short peptide lives at the first level, which is the level its certificate describes.

FOR LABORATORY AND IN-VITRO RESEARCH USE ONLY. NOT FOR HUMAN OR ANIMAL CONSUMPTION.

Common questions

More science guides

• N-Acetyl Semax vs Semax: structure, mass and what the certificate shows
• What Is an Oligopeptide? The IUPAC Length Terms Explained
• Parts of an Amino Acid: The Five Parts of the General Structure, Labeled
• What Is a Peptide Bond? Definition, How It Forms and Why It Stays Flat
• Hydrolysis of a peptide bond: the reaction, why it is slow on its own, and how a lab forces it

Sources

• Ball, Hill and Scott, The Basics of General, Organic, and Biological Chemistry, 18.4 Proteins (LibreTexts). Definitions of the four levels, the alpha helix (3.6 residues per turn, carbonyl to amide hydrogen four residues along), beta sheets, the four interactions that stabilize tertiary structure, hemoglobin's four chains, insulin's 51 residues and the named example proteins.
• OpenStax Biology 2e, 3.4 Proteins. Sickle cell hemoglobin's glutamate-to-valine substitution at position six of the beta chain, the roughly 600 residues of hemoglobin, disulfide bonds as the only covalent bond formed in folding, and denaturation leaving primary structure intact.
• RCSB PDB-101, Molecule of the Month: Hemoglobin. Two alpha and two beta chains, each with a heme group and iron atom; each chain similar to myoglobin; the cooperative shape change on oxygen binding; PDB entries 2hhb, 1hho and 2dhb.
• RCSB PDB-101, Methods for Determining Structure. X-ray crystallography (most PDB structures; requires a crystal; flexible regions often invisible), NMR (solution, ensemble of models, limited to small and medium proteins) and cryo-electron microscopy (thousands of single particles in ice; depositions passed NMR in 2016).
• UniProt P69905, Hemoglobin subunit alpha (human). Sequence length 142 with the initiator methionine at position 1 annotated as removed.
• UniProt P68871, Hemoglobin subunit beta (human). Sequence length 147 with the initiator methionine at position 1 annotated as removed.
• UniProt P02144, Myoglobin (human). Sequence length 154 with the initiator methionine annotated as removed.
• UniProt P01308, Insulin (human). B chain at residues 25 to 54 and A chain at 90 to 110 of the 110-residue precursor; interchain disulfides 31 to 96 and 43 to 109 and the intrachain disulfide 95 to 100.
• Greenfield, Using circular dichroism spectra to estimate protein secondary structure, Nature Protocols 2006 2006 doi:10.1038/nprot.2006.202. Far-UV scan range, the characteristic bands for alpha helix (222, 208, 193 nm), beta sheet (218, 195 nm) and disordered chains, sample amounts of 20 micrograms or less, and the statement that CD gives no residue-specific information unlike X-ray crystallography or NMR.
• Toal and Schweitzer-Stenner, Local order in the unfolded state: conformational biases and nearest neighbor interactions, Biomolecules 2014 2014 doi:10.3390/biom4030725. Short peptides are generally unable to form the stable hydrogen bonding and long-range interactions needed to fold into well-defined secondary structures; polyproline II fractions of about 0.74 for the alanine dipeptide and 0.84 for trialanine; the local random coil model.
• Steen and Mann, The ABC's (and XYZ's) of peptide sequencing, Nature Reviews Molecular Cell Biology 2004 (PubMed abstract) 2004 doi:10.1038/nrm1468. Peptide sequencing by tandem mass spectrometry as the key technology of mass-spectrometry-based proteomics.
• Pauling, Corey and Branson, The Structure of Proteins: Two Hydrogen-Bonded Helical Configurations of the Polypeptide Chain, PNAS 1951 1951 doi:10.1073/pnas.37.4.205. The 1951 paper proposing the hydrogen-bonded helical configurations of the polypeptide chain, including the alpha helix.