Proteomics > Which service should I request? > Interaction Proteomics > Affinity Purification - AP MS

Affinity Purification followed by mass spectrometry - AP MS

General description

The general concept of affinity purification mass spectrometry (AP-MS) is well described in:

Richards, A. L., Eckhardt, M. & Krogan, N. J. Mass spectrometry‐based protein–protein interaction networks for the study of human diseases. (Mol Syst Biol 17, e8792 (2021)):
"AP-MS experiments utilise epitope tagging, where short peptide or protein tags (for example, FLAG-, TAP-, Strep-Tag, or c-myc) are fused to the protein of interest - either in the context of an exogenous expression construct or under the gene’s endogenous promoter using gene editing technologies like CRISPR-Cas. The resulting bait protein functions as an affinity capture probe for interacting, or prey proteins, eliminating the need for specific antibodies to proteins of interest, as would be the case in lower throughput immunoprecipitation (IP) experiments. The affinity tag can easily be purified on a matrix recognising the epitope. After washing steps to eliminate non-specific interactors, interacting proteins can be identified via LC-MS."

Remark

The term "affinity purification" is actually misleading, since modern AP-MS workflows typically use single-step affinity enrichment protocols! The resulting material is therefore NEVER pure.

Studies like:
Keilhauer EC, Hein MY, Mann M. "Accurate protein complex retrieval by affinity enrichment mass spectrometry (AE-MS) rather than affinity purification mass spectrometry (AP-MS)" (Mol Cell Proteomics. 2015 Jan;14(1):120-35)
have therefore correctly suggested to better use the term affinity enrichment followed by MS (AE-MS). We still decided to go with the term AP-MS, since AP-MS is unfortunately still more frequently used in the literature.

The FGCZ supports almost any type of AP-MS workflow and typically takes care of the following steps:

• Digestion of the enriched proteins
• Label-free nanoflow LC-MS analysis
• Relative identification and quantification of proteins (relative to a control sample/condition)
• Scoring of protein-protein interactions

Pilot phase experiments typically skip the last stage of the analysis pipeline and rather focus on getting enrichment conditions optimised/make sure that sample preparation is well aligned with our downstream LC-MS.

Default Workflow

Our default workflow for the analysis of AP-MS samples can be summarised by

Protein digestion

A central step of any bottom-up proteomics workflow is protein digestion (typically using sequencing-grade Trypsin). This step can happen in solution, or while proteins are still attached to a solid support. Accordingly, we offer two sample submission forms:

Sample form	Protocol	Comments
On-beads (the enriched proteins are still attached to the affinity matrix)	on-beads digestion	Please note that many affinity reagents (antibodies) are proteins or coupled to the affinity resin using proteins (Protein A/G). These proteins will be co-digested during on-bead digestion and add signals to your sample matrix. Bring the beads after your last wash step.
In-solution (the enriched material has already been detached from the affinity matrix by mechanisms like pH shift, competition or alike. The specifics depend on used affinity reagent.)	in-solution digestion	Whenever possible, use volatile buffering agents for the elution buffer. Limit the usage of buffer additives like detergents, salts, crowding agents to a bare minimum! Check with us if your additives are compatible with in-solution digestion /LC-MS analysis. We might need to add a clean-up/concentration step by TCA precipitation should that not be the case!

NOTES:

• NEVER bring samples with frozen beads!
  • If you bring protein on beads, NEVER bring them frozen and cover them with a small volume of transfer buffer. NEVER let the beads dry completely!
• Choose an affinity matrix that can be precipitated by either low-speed centrifugation or magnetic separation.
  • We do not have special equipment to handle exotic solid supports!__
• Carefully titrate the volume of the affinity matrix to match the binding capacity of your affinity matrix/reagent
  • You need to maximize bait capture while minimizing background, and still allowing efficient bead dispersion during washing and digestion.
  • Consult with us beforehand!

LC-MS analysis

Our default LC-MS workflow for affinity purification samples uses label-free high-resolution accurate mass (HR-AM) data acquisition on a nanoflow LC-MS system following data dependent acquisition (DDA) logic. Other data acquisition schemes (like PRM, DIA, SureQuant, ...) are technically possible, but need to be discussed upfront! The same applies for the usage of metabolic (SILAC) or chemical labeling (isobaric labeling by TMT).

Identification and quantification of proteins

For the identification and quantification of proteins we are applying state-of-art software tools/pipelines optimised for HR-AM LC-MS data. Typically your data is matched against one or several Uniprot reference proteomes:
https://www.uniprot.org/help/reference_proteome

Some general considerations:

• If a protein is not present in a database, it cannot be identified nor quantified!
• Your order summary should include all necessary information that helps us choosing the right reference proteome(s). Please inquire if your organism is not covered by a uniprot reference proteome.
• Many AP-MS experiment involve transgenic animals/cell lines or include so called spike-in proteins that were added to your samples during sample preparation process. These are typically not part of an organism-specific reference proteome. Please make sure to upload a corresponding FASTA-formated protein sequence(s) of these spike-in proteins, so we can add them to searches, e.g.:

>sp|P42212|GFP_AEQVI Green fluorescent protein OS=Aequorea victoria OX=6100 GN=GFP PE=1 SV=1
MSKGEELFTGVVPILVELDGDVNGHKFSVSGEGEGDATYGKLTLKFICTTGKLPVPWPTL
VTTFSYGVQCFSRYPDHMKQHDFFKSAMPEGYVQERTIFFKDDGNYKTRAEVKFEGDTLV
NRIELKGIDFKEDGNILGHKLEYNYNSHNVYIMADKQKNGIKVNFKIRHNIEDGSVQLAD
HYQQNTPIGDGPVLLPDNHYLSTQSALSKDPNEKRDHMVLLEFVTAAGITHGMDELYK

Scoring of protein-protein interactions

The results of protein identification and quantification pipeline (sample-wise protein abundance estimates) are forwarded to statistical tools that compile lists of biologically relevant interactions.

• This process depends heavily on the availability of suitable control samples and replicates.
• All samples used by this statistical assessment need to be measured in the same order. Do NOT place controls in order a and assume that they can be used for scoring interactions in order b. These are apple to orange comparisons and we will not support such attempts in any way.

Experimental Design

As always in science nothing is as important as a good experimental design!
The specifics depend on the biological question you are trying to solve. Assuming your are trying to probe the social context of protein of interest (POI) in an unbiased manner, we suggest your design should cover the following aspects:

Protein extraction

Ensure efficient extraction of your POI in a buffer that maintains its endogenous interactions during the entire sample handling process.

Controls

The most important aspect is controlling the specificity and efficiency of your protein enrichment. We therefore encourage the usage of affinity tags like FLAG, HA, Strep, ... where you can use generic purification procedures and reagents offering confirmed performance. Using affinity reagents of unknown quality is the biggest risk for the success of your AP-MS experiments and requires careful testing and optimisations! Using fancy technology like LC-MS can NEVER rescue poor enrichments!

In any case, we NEED a minimal neg. control that will be used to calculate enrichments. This could be an unspecific affinity reagent, or a cellular system where the affinity tag is fused to an "out-of-context" protein (a protein that has no natural clients in the studied system, for instance GFP in a human cell line). The untagged/not induced/parental cell line/strain is always the last resort, because it does not control for many aspects of your purification. Omitting the affinity reagent instead of replacing it with an unspecific, closely related agent is an idea of similar quality and might reward you with unspecific hits. It is also a good habit that the control should mimic the abundance level and cellular localisation of your POI. For instance, a FLAG-tagged GFP located in the cytosol of a cell will for sure encounter different proteins, compared to the same protein send to the nucleus or attached to the nuclear membrane. The importance of good controls can not be stressed enough!

Replication & blocking

The statistical analysis of your data requires replicates. For AP-MS one typically replicates the affinity purification and applies the principle of blocking, since it is - most likely - the biggest source of variability in your data. Practically that means:

• try to affinity purify all samples in parallel (not on different days, by different people, using different batches of reagents)
• submit all sample in the same order, so they are measured together and the data stays comparable

All other approaches inflate the variance and you will need to scale-up replicate numbers to arrive at some level of statistical significance!

Comparing biological conditions

Typical AP-MS experiments explore the social context of protein in a specific biological condition (cell line, growth condition, cell cycle stage, ...). The possibilities are endless. We can only do this one at a time having minimal controls for each condition. Comparing across conditions is currently not explicitly covered by our analysis pipeline, but well supported by providing linked visualisations.

Testing your design & protocol using standard laboratory equipment

We heavily encourage testing the efficiency of your pulldown using standard procedures such as SDS-PAGE and in-gel protein detection and/or Western blotting by probing input, supernatant and eluate of your sample before preparing samples for LC-MS/MS. This often helps determining the optimal input and affinity reagent amounts.

Expected outcome

Depending on the applied enrichment protocol, starting material and completeness of the reference proteome we typically identify and quantify hundreds of proteins per sample. By contrasting the sample-wise protein abundance matrix with a control sample or sample set, it becomes possible to narrow down with proteins show a sig. co-enrichment. We typically provide access to the raw LC-MS data, combined identification & quantification results, as well as interaction scoring.

Proteome-scale AP-MS examples

The human BioPlex project provides a good example how modern AP-MS workflows are applied at large scale. Their latest publication describing the project is:
https://doi.org/10.1016/j.cell.2021.04.011

In short, the authors applied single-step affinity purification (HA tag-mediated) combined with HA peptide elution. The tagged proteins were expressed using a human ORFeome collection in HEK293T cells. Another impressive example is the recent yeast interactome from the Mann lab described in
https://www.biorxiv.org/content/10.1101/2021.10.24.465633v1

Here, GFP fusion proteins were used as bait proteins and enriched by GFP-specific nanobodies (GFPtrap). Like in their previous paper, single-step affinity purification was favored, and combined with on-bead digestion.

Benchtop protocols

Hesketh, G. G., Youn, J.-Y., Samavarchi-Tehrani, P., Raught, B. & Gingras, A.-C. Proteomics, Methods and Protocols. Methods Mol Biology 1550, 115–136 (2017).
Describes the parallel implementation of both BioID and FLAG AP-MS allowing simultaneous exploration of both spatial and temporal aspects of protein interaction networks.

Native Isolation of 3×HA-Tagged Protein Complexes to Characterize Protein-Protein Interactions

https://currentprotocols.onlinelibrary.wiley.com/doi/10.1002/cpz1.29

FAQs

Questions?
Check our dedicated Interaction Proteomics FAQs page or contact us at proteomics at fgcz.ethz.ch