Structural Biology: Key Terms and Glossary

An effector that binds once, between the β chains, in the central cavity of deoxyhaemoglobin and thereby stabilises the T state and promotes unloading.

A right-handed duplex with C3′-endo sugars, adopted by duplex RNA and by DNA–RNA hybrids, with a deep and narrow major groove.

An RNA tertiary contact that packs an adenosine into the minor groove of a helix, abundant in the ribosome and absent from a lone duplex.

A neural-network predictor of protein coordinates from a sequence and a multiple alignment. AlphaFold2, described by Jumper and colleagues in 2021, often reaches near-experimental backbone accuracy, but a confident coordinate set is still not a measured map.

A nucleated fibre of cross-β structure, able to seed its own elongation, and distinct both from amorphous aggregate and from polymerisation of a native fold.

The experimenter whose refolding of reduced ribonuclease showed that sequence can recover native disulphides and activity. Some proteins still need a chaperone or a pro-region on a biological timescale, and a few native states are kinetically trapped.

The unique part of the unit cell from which space-group symmetry generates the remaining copies.

A record of surface height under a tip, with lateral detail typically a few nanometres and with the buried interior invisible.

The usual cellular DNA helix, right-handed, with about 10.5 base pairs per turn in solution and a wide major groove.

A parameter refined with the coordinates that mixes atomic motion, static disorder, and error, and is not a trajectory.

Boyer's account that the three catalytic sites of ATP synthase take turns, and that release of tightly bound ATP is the difficult step.

The annotation, in a structure entry, of which contacts are the oligomer rather than the crystal packing, an annotation that is sometimes wrong.

The condition nλ = 2d sinθ for constructive interference from a set of lattice planes, so a finer spacing d appears at a higher glancing angle. The smallest spacing kept is a nominal resolution, not a promise that every atom is known to that precision.

The serine-protease arrangement of serine, histidine, and aspartate, serine acting as the nucleophile and histidine as the base. The same three residues appear by convergence in subtilisin.

A kinetic helper that binds exposed hydrophobic stretches or cages a chain so it does not aggregate, without in general imposing a fold the sequence does not encode.

A resonance frequency relative to a reference, expressed in parts per million so that it does not depend on the field, and sensitive to the local electrons and thus to conformation.

A change in chemical shift when a partner binds, marking a changed environment that may be a direct contact or an allosteric effect at a distance.

A far-ultraviolet difference in absorption of circularly polarised light that reports approximate secondary-structure content, not which residues are helical.

The arrangement fixed by covalent bonds, including connectivity, stereochemistry, and which cysteines are paired. A different configuration is reached only by making or breaking a covalent bond.

The arrangement reached by rotation about single bonds, so one configuration has many conformations and folding is a search among them.

A covalent join between side chains within the reach of a chemical reagent, identified by mass spectrometry as a sparse upper bound of typically a few tens of ångströms, not as a van der Waals contact.

The repeating fibre diffraction of amyloid, with strands roughly perpendicular to the fibre axis and hydrogen bonds running along it.

A single rotation axis that places subunits in a closed ring, all pointing the same way, as in a twofold dimer.

Stripping of annular or structural lipids by detergent, which can abolish a gate or a site even when the resulting map looks intact.

An amphiphile that replaces the lipid belt of a membrane protein with a micelle, and that must not be mistaken for the bilayer the protein evolved in.

Cyclic symmetry plus perpendicular twofold axes, so a second copy of the assembly is related by a flip.

A camera that records cryo-EM frames as a movie, so beam-induced movement can be aligned before the frames are averaged.

A covalent link between cysteines, formed where it is oxidising, that locks a fold packing has already specified by constraining the unfolded chain.

A portion of a chain, often continuous, that can fold on its own and has its own core.

The bundle of models deposited for an NMR structure, each consistent with the restraints and with covalent geometry, so a splayed region may be mobile or merely under-determined.

A picture of many unfolded conformations narrowing, as native contacts form, towards fewer lower-energy states, without requiring one compulsory pathway.

A comparison of spatial frequencies in two independent half-maps, often quoted where the correlation falls to 0.143, and only a global figure.

Opening and closing of the conduction path, in KcsA at the inner-helix crossing below the filter, and a separate question from which ion the filter prefers.

A self-splicing RNA that binds an exogenous guanosine and uses that ribose to attack the 5′ splice site, with the active site in the structures made of RNA.

A small ribozyme in which a 2′-hydroxyl attacks the adjacent phosphorus in line. Structures of a minimal fragment missed that geometry because peripheral tertiary contacts were absent.

A DNA-reading motif that braces one helix in the major groove with a second helix. Which bases are read depends on the side chains that meet the groove.

A measure of binding cooperativity, not a count of sites. Human haemoglobin has four oxygen sites and a coefficient of about 2.8, strong but short of infinite cooperativity.

Construction of coordinates by aligning a sequence to an experimental template, copying the conserved backbone, and rebuilding insertions and side chains. The model has no R-free, because nothing was measured.

An assembly built from more than one kind of restraint, legitimate only while each method's limit stays visible and unrestrained parts are not drawn as if they were atomic.

Replacement of backbone amide hydrogens by solvent deuterium, slow where the amide is buried or hydrogen-bonded, and a report of protection rather than of a fold or of a guaranteed contact.

The main driving force for collapse, in which burying nonpolar surface releases restricted water and, at physiological temperature, the gain is largely solvent entropy.

A stretch that does not adopt one tertiary structure under native conditions, often because it lacks a hydrophobic core, and that can still function as an ensemble.

A through-bond scalar interaction that groups the spins of one residue, walks assignment along the backbone, and can restrain the backbone angle φ.

A coiled-coil dimer, with leucine on a seven-residue repeat, that pairs the chains while a separate basic region reads the major groove.

The detail that actually varies across a cryo-EM map, which can be much poorer at the periphery than the single global number suggests.

Multi-wavelength anomalous dispersion, which uses differences at two or three wavelengths near an absorption edge to locate scatterers and phase the macromolecule.

The wider groove of B-DNA, whose pattern of hydrogen-bond donors, acceptors, and nonpolar patches can distinguish all four base pairs.

The practice of leaving selected methyl groups protonated in an otherwise deuterated protein so that methyl spectra can probe very large assemblies, usually as reporters of binding and motion rather than as a complete model.

The narrower groove of B-DNA, which does not tell A–T from T–A and is often read as shape, including the narrow groove of an A-tract.

Phasing by placing a similar known structure in the cell and calculating phases from it, with the risk that those phases redraw the search model and preserve a wrong loop.

A heavy-metal salt dried around a particle, used as a low-resolution check of a preparation and not as the unstained high-resolution experiment.

The nuclear Overhauser effect, a through-space transfer of magnetisation that falls as the sixth power of distance and is generally useful only to about 5–6 Å.

Backbone NH groups that stabilise the negatively charged oxygen of the tetrahedral intermediate during serine-protease catalysis.

The site on the large ribosomal subunit where the peptide bond forms. It is an RNA active site: the inner shell is ribosomal RNA, not protein.

The loss of each structure-factor phase when the detector records only intensity, without which an electron-density map cannot be computed.

A per-residue confidence, from 0 to 100, reported with an AlphaFold model. It is not an experimental resolution and not a B-factor, and a score below about 50 should not be drawn as a structure.

The enrichment of lysine and arginine in cytoplasmic loops of bacterial inner-membrane proteins, a check on which side faces the cytoplasm rather than a map of a binding site.

AlphaFold's estimate, for every pair of residues, of error in their relative placement, which can show a hopeless hinge between two locally confident domains.

A specimen in which particles adopt only one or a few views, so missing directions cannot be invented and the reconstruction is anisotropic.

An infectious protein conformation, in which a β-rich form templates conversion of the host protein and no separate nucleic-acid genome is required. Ordinary human amyloidoses do not spread between people in this way.

A stretch, later cleaved, that some proteases require if they are to reach the native conformation on a biological timescale.

The archive of experimental macromolecular models. An entry is one model under stated conditions, and its biological assembly, resolution, and missing residues have to be read before a figure is believed.

The globin histidine, His F8, that coordinates the haem iron on the protein side and moves with helix F when oxygen pulls the iron into the porphyrin plane.

The oxy quaternary structure of haemoglobin, higher in oxygen affinity, whose central cavity is too narrow for 2,3-bisphosphoglycerate.

The fractional disagreement between observed and calculated structure-factor amplitudes on the reflections used in refinement. It falls if the model fits noise as well as signal.

The same amplitude disagreement computed on a set of reflections held out of refinement, so it is not the working R-factor. A large gap above the R-factor is evidence of overfitting.

A size taken from the low-angle, Guinier region of a SAXS curve, easily dominated by a trace of aggregate.

A map of backbone φ against ψ. Steric clashes confine most residues to the right-handed α and the β basins; glycine is allowed more of the map, and proline is fixed in φ.

A stretch that enters the membrane and returns to the same side, lining pores such as those of potassium channels, and missed by a simple span-finding plot.

A dipolar coupling that does not average exactly to zero in a weakly aligning medium, and that reports bond orientation in a common frame rather than a short local distance.

In a crystal data set, the smallest interplanar spacing whose intensities are retained, which limits what a map can usually show and does not make every atom equally certain.

The change, around 2013, when direct electron detectors and statistical classification made near-atomic maps of large complexes and membrane proteins a routine cryo-EM outcome. The experiment did not become infallible.

A three-track network of sequence, distances, and coordinates, described by Baek and colleagues in 2021, less accurate than AlphaFold2 on the same blind test and still dependent on the alignment.

Synthesis and hydrolysis of ATP coupled to rotation of the central stalk, so that one turn cycles the three catalytic states seen together in the F1 crystal.

Single-wavelength anomalous dispersion, which locates anomalous scatterers from one wavelength and, with density modification, phases the macromolecule.

Small-angle X-ray scattering from macromolecules in solution, averaged over orientations and over every particle in the beam, and not a concealed atomic model.

The narrow conduction path of a potassium channel, a braced cage of backbone carbonyls that compensates potassium for the water it loses and does not do the same for the smaller sodium ion.

Adult haemoglobin with a glutamate-to-valine substitution on the β chain, which polymerises as a native deoxy tetramer rather than by unfolding or as amyloid.

Reconstruction of a three-dimensional scattering map by estimating the orientations of many noisy projections of particles that need not form a crystal.

The use of a site's shape and chemistry to choose the next molecule, followed by a structure of the complex and an affinity measured in solution. A site is a constraint, not a medicine, and the notes give no synthetic route.

A family of light-microscopy methods that localise a fluorophore more finely than the diffraction limit, placing a label in a cell rather than showing side chains.

The deoxy quaternary structure of haemoglobin, lower in oxygen affinity, and the state stabilised by salt bridges, protons, and 2,3-bisphosphoglycerate.

An RNA motif that docks a GNRA tetraloop onto a distant helix and thereby packs two duplexes into one fold.

The claim that, under physiological conditions, the native conformation is the lowest free-energy state the sequence can reach, so the cell need not supply steric information the sequence lacks.

A membrane-spanning α helix, the usual response of plasma membranes and bacterial inner membranes to a hydrophobic thickness of about twenty to twenty-five residues.

Transverse relaxation-optimised spectroscopy, which keeps a slowly relaxing component of a nitrogen–proton signal and extends backbone spectra of large proteins without by itself assigning the chain.

The signature sequence of the potassium-channel filter, whose side chains mostly brace the carbonyls that coordinate the dehydrated ion.

The range of sequence similarity, Rost's phrase, in which a genuine homologue is hard to separate from chance and an automatic model is unsafe.

The volume that repeats through a crystal lattice.

Amorphous solid water, produced by cooling faster than ice crystals nucleate, so a hydrated molecule can be imaged without crystalline diffraction or stain.

A local left-handed helix with a zigzag backbone, favoured by alternating purine–pyrimidine sequence and by negative supercoiling, and not a second genome-wide conformation.

A small ββα module whose fold is built by zinc held on cysteines and histidines. The metal stabilises the module and does not itself touch the DNA.

A right-handed helix whose backbone hydrogen bonds run from the carbonyl of residue i to the amide NH of residue i+4.

A closed transmembrane β sheet, typical of bacterial and mitochondrial outer membranes, with side chains alternating between lipid and the pore.

An assembly of extended strands, antiparallel with nearly linear hydrogen bonds between neighbours or parallel with angled bonds.

A four-residue reversal of the chain, with a hydrogen bond from residue i to residue i+3.

Rotation about the backbone N–Cα bond, one of the two free angles once the peptide bond is held planar, and one axis of a Ramachandran plot.

Rotation about the backbone Cα–C bond, the other free angle of a planar peptide unit, and the second axis of a Ramachandran plot.

Torsion about the peptide bond, which is partly double and usually trans, near 180°, rather than cis.