Modern Proteomics Methods

Why is proximity labeling used in proteomics?

• traditional MS struggles w/ weak/transient protein interactions and membrane proteins

• proximity labeling solves this by tagging proteins near a protein of interest inside living cells

  • captures native, endogenous interactions that can be identified by MS

• we attach an enzyme to a protein of interest (via ligand/antibody) → enzyme generates a highly reactive species (often free radical) which labels nearby proteins within a few nm

• labeled proteins are then purified and identified by MS

How is the GLP-1 receptor targeted for proximity labeling in the GLP1 paper

Ligand-based targeting:

• GLP-1 (ligand) is conjugated to an enzyme → GLP1 binds GLP-1R → enzyme sits near receptor → initiates labeling reaction

Recombinant tagging approach (transgene):

• GLP-1 or receptor is genetically fused with labelling enzyme

• Notes: GLP1 attached to APEX → GLP1 binds to receptor → biotin phenol is deposited onto protein of interest (biotin phenol is what is deposited and activated)

  • APEC uses H₂O₂ to oxidize biotin phenol to produce radicals

  • can later be isolated w/ streptavidin cuz it binds biotin with high affinity

• ensures controlled, specific labeling

Results of proximity labeling

• measuring cAMP production (downstream effect of GLP1 binding)

• right = transgenically fused approach

Why is observing membrane labeling important in the GLP-1 proximity labeling experiment

• GLP-1R is a cell membrane protein

• if the system is working correctly, labeling should occur at the membrane (ie ligand is binding correctly and enzyme is positioned at the receptor)

• we see labeling when all components (GLP-1-enzyme, biotin-phenol, H₂O₂) are in

  • if labeling occurs only when everything is present, it confirms specific enzymatic labeling

Membrane-localized staining of GLP-1 construct

• membrane localized staining tells us that GLP-1 is successfully binding to GLP-1R and the conjugated enzyme is correctly localized at the membrane

• therefore the GLP-1 construct is functional

• proper localization is essential b/c proximity labelling tags everything nearby

  • if localization is wrong, you label the wrong proteins and get meaningless data

  • correct localization ensures the labeling reflects true biological neighbors

Biotin-Enrichment Approach

• previously, immunofluorescence shows localization and the proteomics workflow identifies which proteins are present

• biotin enrichment allows isolation of all proteins near GLP-1R

  • MS then identifies and quantifies these proteins

• limitation: cannot confirm direct physical binding to GLP-1R b/c labeling is based on distance, not binding

Biotin-Enrichment Approach: Steps

• perform labeling (biotin-phenol + H₂O₂)

• lyse cells

• add streptavidin-coated beads

• biotinylated proteins will bind strongly to streptavidin

• wash away non-biotinylated proteins

• analyze bound proteins by MS

Roepstorff-Fohlmann-Biemann Nomenclature: What determines where the charge goes?

• when a peptide bond breaks, one fragment keeps the charge → detectable in MS

  • charge on N-term → b-ion

  • charge on C-term → y-ion

• if a peptide has more than one charge, we can produce both b and y ions

• when fragmenting peptides, we put it enough energy to break one bond per peptide (no t too much)

Roepstorff-Fohlmann-Biemann Nomenclature: Energy

• when fragmenting peptides, we put it enough energy to break one bond per peptide (not too much)

  • we want clean, interpretable fragment ions

  • if too much energy → multiple bonds break

• each peptide molecule may break at a different position, and across many molecules, we could get a series of fragments covering the sequence

Residue mass vs amino acid mass

Amino acid mass:

• free amino acid (includes full structure)

Residue mass:

• amino acid within a peptide minus H₂O (lost during peptide bond formation)

• MS calculations use residue masses, not free amino acid masses

Read off the sequence: How do you determine a peptide sequence from a MS/MS spectrum?

• use b-ion or y-ion series

• look at mass difference between adjacent peaks, each difference = one amino acid residue mass

• build sequence step-bystep

Why is a y₁₀ ion impossible for a 10 amino acid peptide?

• fragmentation breaks peptide bonds between residues

• therefore the max y-ion is y₉

• for n amino acids → max fragment = n-1

Parent Ion: What does [M+2H]²⁺ represent?

• the intact peptide (original ion, without any fragments missing)

• M= m/z + 2 protons → we have the whole peptide + 2 protons (2+ charge state)

• seeing it means fragmentation was incomplete, but it’s useful because it confirms expected peptide mass

Why are singly charged fragments easier to interpret?

• for z = 1, m/z ≈ actual mass + 1 proton

• easy to calculate: subtract proton mass to get fragment mass

• no need to divide by charge

• the spectra is showing a full series of ions that are all singly charged

How do you use y-ion differences?

• take 2 adjacent y-ions (e.g. y₇ - y₆), calculate mass difference and match the difference to the residue mass of an amino acid

Why are middle fragments more abundant?

• bonds in the middle of the peptide break more easily and the gragments generated from cutting in the middle are more stable than very big or very small fragments

• exception: Proline, remember that the bond before proline is more labile, resulting in strong fragment peaks at that position

Why don’t we always see b- and y-ions

• fragmentation is probabilistic (not every bond breaks equally)

• some fragments are unstable or have low abundance

• you can still determine the sequence if some ions are missing by using nearby ions and known peptide mass

Why does a C-terminal K suggest a tryptic peptide

• trypsin cleaves after Lys or Arg, so peptides often end in K or R

De novo sequencing

• determining peptide sequence directly from MS/MS spectra

  • uses mass difference between fragment ions and doesn’t rely on protein databases

• advantage: can recover full amino acid sequence directly, useful when protein is not in database or detecting mutations or novel peptides

• disadvantages: high error rate, slow, computationally intensive, often the full sequence information is not present

• use as backup option and to find ‘stretches’ of correct sequences (infer sequence directly from fragment ions)

  • primary method is to match spectra to known protein sequences (database search)