Protein sequencing

Protein sequencing is the determination of the order of amino acids in a protein or peptide. The first protein to be read was bovine insulin, a chain of 21 amino acids joined to one of 30, and Frederick Sanger and his colleagues published its sequence in papers from 1951 to 1955.^[1][2][3] Sanger received the 1958 Nobel Prize in Chemistry for his work on the structure of proteins, and in 1980 a second one, shared, for determining base sequences in nucleic acids, which have four letters where protein has twenty.^[4][5]

DNA sequencing now supplies the sequence of nearly every protein by inference, which has shifted the purpose of the direct method. A gene gives the chain a cell can make. It does not say which pieces are cut off, which side chains carry a phosphate or a sugar, or how much of each protein a sample holds. Those questions are answered by measuring the protein itself, and the measuring is harder than for DNA because a protein cannot be copied before it is read.

In 1945 Sanger showed that 1-fluoro-2,4-dinitrobenzene reacts with the free amino groups of insulin and attaches a yellow label to them.^[6] Broken into fragments by partial hydrolysis with acid or enzymes, the chain gave short peptides, and the labeled residue at the end of each could be identified. The sequence was assembled from fragments that overlapped, as a torn page is rebuilt from its pieces. Insulin also has disulfide bonds between and within its chains, which had to be cut before the pieces could be separated and then located again, and the paper of 1955 placed them.^[3]

In 1950 Pehr Edman described a chemical cycle that removes the first amino acid of a peptide and leaves the rest intact.^[7] Phenyl isothiocyanate couples to the free amino group at one end under mild alkaline conditions; acid then cleaves off the terminal residue as a derivative that can be identified by chromatography, and the shortened peptide returns for another round. In 1967 Edman and Geoffrey Begg described a machine, the protein sequenator, which ran the cycles automatically in a spinning cup.^[8]

The method reads one end of a chain, and it loses ground at every cycle. If 98% of the chains complete each round, 55% are still in step after 30 cycles and the signal at each position is blurred by chains that lag behind. A chain whose first amino acid is chemically blocked cannot be started at all. Edman sequencing nevertheless stayed the reference method for decades, and its chemistry reappears in the single-molecule methods below.

Mass spectrometry weighs molecules, and for most of the twentieth century proteins were too large and fragile to be turned into ions without breaking. Two methods published in 1988 and 1989 changed that: laser desorption of proteins, by Karas and Hillenkamp and independently by Tanaka and colleagues, and electrospray ionization, by Fenn and colleagues.^[9][10][11] John Fenn and Koichi Tanaka shared half of the 2002 Nobel Prize in Chemistry for these methods.^[12]

In the arrangement now standard, called bottom-up or shotgun proteomics, the proteins in a sample are cut with an enzyme such as trypsin into peptides of a few dozen amino acids at most. The peptides are separated by liquid chromatography, weighed, broken in a second stage and weighed again, and the pattern of fragments is matched against a database of sequences predicted from genomes. A peptide breaks mostly at its backbone bonds, and the masses of neighboring fragments differ by the mass of one amino acid, so in principle the spectrum can be read as a sequence without a database. Leucine and isoleucine, two amino acids with the same composition of atoms, have the same mass and cannot be told apart this way.^[13] The method can identify thousands of proteins in one experiment. It measures an ensemble, it reads sequences that are already in a database more easily than new ones, and it has difficulty with a sample in which the amounts differ enormously: in human plasma, more than ten orders of magnitude in concentration separate albumin from the rarest proteins measured clinically.^[14]

A review of 2021 treats single-molecule sequencing and identification, together with advances in mass spectrometry, as the route to profiling the proteomes of single cells, which is not yet established.^[15] The approaches differ in what they read.

Fluorosequencing keeps the Edman chemistry and adds a camera. Swaminathan and colleagues labeled the cysteine and lysine residues of peptides with dyes, fixed the peptides to a glass surface, and watched under a microscope as each molecule's fluorescence dropped when a cycle of Edman degradation removed a labeled residue. The result is a sparse sequence, the positions of two amino acids in each chain, matched to a reference database. They measured efficiencies above 93% for labeling, survival and cleavage, and worked with zeptomole quantities, about 600 molecules.^[16]

Quantum-Si reads the amino terminus of a peptide in real time. In a 2022 paper its scientists described a mixture of dye-labeled recognizers for N-terminal amino acids and aminopeptidases that cleave the terminal residue, on an integrated semiconductor chip. Each amino acid is identified from the intensity and lifetime of the fluorescence and the kinetics of binding, which let recognizers that bind several amino acids still tell them apart.^[17] Nautilus Biotechnology identifies proteins without reading their sequences. Its platform fixes single intact proteins on a nanofabricated array, probes them in repeated cycles with affinity reagents that bind short stretches of sequence, and identifies each protein by its pattern of binding.^[18]

Nanopores read a chain as it passes through, with no label. The difficulties are described in Nanopore protein sequencing: the alphabet is large, the chain carries no uniform charge, and it folds. In 2021 a group in Delft read the same peptide in a pore 117 times, but an instrument that sequences an unknown protein from end to end does not exist.^[19]

Molecular Reality Corporation does not claim to read protein sequences; its guide describes its work as detection. Its place relative to the groups that are trying is set out in Nanopore protein sequencing, and its pores are described in Nanopore.

• Single-molecule biophysics
• Nanopore sequencing
• Protein Data Bank
• AlphaFold