Protein-protein interactions Flashcards
Mitotic cell division cycle (why study, model organism, )
G1: gap between phases and contain major control points; when certain molecules become de/active and control progress of cycle.
S: DNA synthesis
G2: gap between phases and contain major control points; when certain molecules become de/active and control progress of cycle. Check cell size and DNA replication
Mitosis (M): chromosome separation
Paul Nurse in 1970s did experiments on model organism S. Pombe fission yeast in zoology department studying cell cycle and later discovered the process was conserved in all eukaryotes through identifying homologous genes in humans.
Can use yeast cell cycle as one approach to characterise protein-protein interactions.
Cell cycle regulation is important (separation of DNA can’t begin before DNA replication is complete) and cells need to coordinate this complex process.
Mechanism by which G2 to M transition is controlled, model organism, why it’s used,
In fission yeast it’s simplified and easier to manipulate and identify genes, before moving on to more complex organism like human cells to study.
1) Genetics are easy to manipulate and generate mutant strains
2) Molecular biology techniques like DNA cloning, protein expression, etc are easy to apply.
3) Complete genome sequenced in 2002 so well-characterised so easier to study gene function, protein interactions and genetic networks.
4) Grow easily - two hour division time making high-throughput experiments suitable
4) No ethical issues
6) Evolutionary conservation in more complex eukaryotes. Important in some human disease (ex. cancer)
Fission yeast is used as there are multiple reasons it’s easy to work with:
Can see it’s morphology under microscope to know if it’s in mitosis (rod shaped cell with septum at centre) or meiosis (group of four spores in elongated shape). Can also easily identify if mutants affect M phase entry via cell size.
Temperature sensitive mutants allows lethal mutations to be used.
Cyclin dependent kinase (CDKs) are master regulators of S and M phase discovered by Paul Nurse. To activate CDK1 (cdc2), it is dephosphorylation (on Tyr15) which now allows cyclin binding. Now both dephosphorylated and bound to cyclin B (cdc13), CDK1 is active.
Wee1 (kinase) phosphorylates CDK1. Cdc25 (cell division cycle, phosphates) dephosphorylates CDK1
CDK phosphorylates proteins to turn on at different stages of cell division.
What experiments were done to initially identify genes important in cell cycle
Genetic screen:
Mutate yeast with chemical to have one gene with loss of function mutation. Phenotype and identify those with a different morphology to WT.
Note in temperature mutant (conditional phenotype) fission yeast (S. pombe), when grown at 25°C show WT phenotype but at 36°C exhibit the mutation.
Identified 30 genes that influence process (further research showed they’re fundamental) and years later biochemical research showed they all interact in a network.
WT at 25°C and 36°C: grows and divides, undergoing multiple cell cycles
cdc mutants at 36°C: elongated cell
Indicates cycle arrest but growth continues, and cell then dies
Growth and division are separate processes.
Identifying gene
Cloning via complementation: DNA library of thousands of plasmids with each WT gene in the cdc mutant S. pombe. Transform yeast cell to take up one plasmid each at 25°C.
At 36°C, most cells will still exhibit mutation.
The cell with the plasmid matching the mutated gene (complementation) will have WT phenotype rescued.
Sequence plasmid to identify the WT gene responsible for the phenotype.
Now know which gene was responsible for mutant phenotype.
How to identify potential binding proteins/protein-protein interactions in cell cycle and examples
Through yeast genetics studies can identify potential binding partners. Further experiments are use of biochemical methods to confirm. Can also identify if there are homologous genes in other organisms.
1) High copy suppression
suc22+ (overexpression of suc22) rescues cdc22 mutant (arrests in S phase), and vice versa.
Indicates the two proteins interact.
Further experiment/analysis found cdc22 and suc22 were the large and small subunit of ribonucleotide reductase.
cdc2 overexpression saved cdc13, which are CDK1 and cyclin B respectively that interact.
2) Synthetic lethality
cdc2 mutant and wee1 mutant both result in small cells (cell cycle advance) phenotype at 36°C.
cdc2 wee1 double mutant is dead at 36°C
Strong indication that the two proteins work together and suggests they interact
With robotics can do global approach in which the entire genome is sequenced and all combinations of double mutants produced to figure out network
3) Epistasis
cdc25 mutant has elongated phenotype and wee1 mutant has small phenotype at 36°C.
cdc25 wee1 double mutant has the WT phenotype at 36°C. Indicates the are extragenic suppressors for one another (a gene at suppresses the phenotype of the mutation).
This suggests the two proteins interact
How to show binding proteins/protein-protein interactions in cell cycle and examples
Through biochemical methods can show protein-protein interactions and confirm those suggested by yeast genetic studies (less demanding)
1) Yeast two-hybrid (budding yeast)
The GAL1 upstream activating sequence (UAS) is an enhancer in budding yeast upstream from the GAL1 gene open reading frame (ORF). The GAL4 transcription factor (comprised of DNA binding domain; BD and activation domain; AD) binds the GAL1 UAS. RNA polymerase II binds the GAL4 transcription factor and encodes GAL1 (galactose).
Since GAL1 is only transcribed when the GAL4 TF domains (AD and BD) interact or are in close proximity (in same protein complex), can use as basis to identify if two proteins interact.
Make two fusion protein each on a plasmid of AD (without it’s STOP codon) with protein X (prey vector) and BD with protein Y (bait vector).
Use yeast with a GAL1 UAS regulated lacZ + His genes to verify expression. On X-gal media, lacZ produces a blue product (positive control). On His- media, yeast won’t grow if His isn’t expressed (negative control). Two controls are used since there are lots of false positives typically. Also use empty vector negative control. Also use known protein-protein interaction as positive control.
Transform yeast with AD-protein X and BD-protein Y plasmids
If proteins X and Y interact, blue colonies will be shown.
Can use gene from any organism and detect transient/weak interactions, however not good for membrane proteins or transcription factors. Also false positives.
Can fish for new interaction protein. Bait is protein of interest BD fusion and make a cDNA library of prey vectors with one gene fused to each AD. Collect blue colonies and identify which protein the prey is and so the protein-protein interactions the protein of interest is involved in.
2) TAP-tag purification
Highly purify protein while also co-purifying proteins bound to it
Add TAP (tandem affinity purification) tag to C-terminal of a gene from fission yeast (ex. cdc2). Fused to gene in chromosome so it’s expressed at normal levels (less false positives).
TAP tag is calmodulin binding protein (CBP), bridge (with TEV protease cleavage site), then protein A.
A very effective 2 step purification is performed:
Affinity chromatography with IgG beads that protein A binds to (very specific).
Cleave off protein A with TEV protease the perform second affinity column with calmodulin beads that CBP binds to with Ca2+ buffer. Native elution (important for protein complex activity, stability and structure) with EGTA buffer to obtain CBP tagged protein of interest with bound proteins
Directly identify bound proteins with peptide mass fingerprinting in which protein is cleaved into smaller peptides (with trypsin), mass accurately measured with mass spectrometer then compared to database with known or theoretical protein sequences to identify.
Many false positive arise from the high sensitivity of mass spectrometry as well as all antibodies having off-targets as there are some proteins that are sticky and are often caught. CRAPome has database of sticky proteins.
3) Engineering protein kinases
Understand function and identify substrates of protein kinases with analogue sensitive protein kinase mutants (Shokat mutants)
Protein kinases (ex. CDK1 and wee1) are proteins that when ATP bound to a specific cleft, add phosphate groups to other proteins (OH group on Tyr, Ser or Thr residues) to modify their activity. The interaction of the protein kinase with the substrate is temporary and transient binding, so it’s difficult to identify substrates.
Shokat mutant allele of protein kinase is a kinase with a mutation in the ATP binding cleft that binds Shokat mutant chemicals (ATP analogues). The protein kinase with the ATP analogue inhibitor in the cleft can bind its substrate target protein but cannot phosphorylate it (inactive).
This Kinase trap is reversible (temporal control of protein activation at different points in cell cycle) and allows the identification of substrate through TAP-tag.
Doesn’t require temperature shift in method so there is no global stress on the cell or non-specific effects due to change in temperature and not the target protein
Method to interrogate all the proteins in the vicinity of the protein of interest and how method has been developed
Proximity labelling: Obtain information on spatial organisation and dynamic protein interaction networks.
Gene fusion expressed in living cell of protein of interest and a labelling enzyme that biotinylates (covalently conjugates) nearby molecules (promiscuous; non-specifically labels all those in proximity).
Lyse and solubilise cell, then denature proteins with SDS. Perform pulldown assay for biotin (with streptavidin; unlike Ab, is highly specific) and identify proteins bound (and their concentrations) through (quantitative) mass spectrometry
Can perform western blot to see smear of bands of biotinylated proteins. Blot with streptavidin-HRP (HRP is an enzyme with a light product; chemiluminescence western blot) to visualise. Streptavidin is a bacterial protein from Streptomyces avidinii that is a tetramer with 4 high affinity (H bonds, Van der Waals and polypeptide loops) biotin binding sites, one of the strongest non-covalent bonds in nature (Kd = ~10^-15).
Biotin (vitamin B7) is a non-toxic, naturally occurring, easily supplemented to cells / animals co-factor for certain carboxylase enzymes and essential for all prokaryotic and eukaryotic life forms
bioAMP is a biotin free radical that when released from the enzyme, reacts with the first primary amine (lysine) it encounters.
The labelling enzyme can be:
BirA: BioID is a bacterial biotin ligase that dimerises when biotin is bound.
R118 is in disordered loop that closes the active site to retain biotin. R118G (BirA) has 100-fold reduced biotin affinity and 400-fold reduced bio-5’-AMP affinity (premature release of reactive intermediate)
BioID with ATP converts biotin to BioID
On western blot of myc tagged (loading control) WT BirA and BirA* expressed mammalian cells, the WT has few bands (biotinylation of endogenous proteins) but BirA* has large dark smear (massive incorporation of biotin into cell proteome)
Ex. BirA* fused to nuclear envelope protein Lamin A identified interaction network, most of which unsurprisingly nuclear proteins (biotiylation of multiple proteins detectable within 1hr starting labelling and peaks at 24hrs).
BioID2:
smaller fusion protein, higher affinity for biotin
TurboID:
Ascorbate peroxidase (APEX):
Oxidises phenol derivatives to phenoxyl radicals (short lived, <1ms half life) that react with Tyr, Trp, His and Cys
Oxidation catalysed by APEX requires hydrogen peroxide and biotin-phenol (so both need to be supplemented)
APEX2
Pros and cons of different labelling enzymes in biotin proximity labelling
BirA*:
Advantages
-It works equally well for transient & stable interactions
-Biotin is non-toxic, easily supplemented to cells and organisms (~50µM required)
-Labelling radius of 7-15nm reflects the reactivity of the bio-AMP intermediate
-N terminal or C terminal fusion possible
Disadvantages
-Labelling is relatively slow (typically 18h)
-BirA* adds 321 amino acids (can influence protein function)
-Fusion can block protein interaction sites at N or C terminus
-Proximity not direct interaction
APEX:
Advantages
-Labelling is exceptionally rapid
-Can define whole organelle proteomes
-Can measure proximity changes during rapid signalling events (snapshot of dynamic changes) since it labels within 1 min (western blot of cell after 1min compares to and sometimes has more biotinylated substrates than pBirA after 24hrs)
-Labelling radius of ~20nm since the phenoxyl radical is highly reactive with a short half life (<1ms) so can’t diffuse far to label proteins
-N terminal or C terminal fusion possible
Disadvantages
-Requires 1mM hydrogen peroxide (toxic to cells)
-Fusion can block protein interaction sites at N or C terminus
-Proximity not direct interaction
Example of experiments using biotin proximity labelling
GPCR signalling
Fuse APEX to a GPCR (angiotensin-II receptor then β2-adrenoreceptor)
Treat cells with a GPCR agonist, add APEX labelling reagents at defined timepoints after receptor activation
Purify biotinylated proteins
Quantify and compare changes of proteins and in vicinity of receptor and protein concentrations at different timepoints and compared to untreated cells (lots of mass spectrometry)
Measured every 10s for 3mins.
Results:
Activation increases interaction of arrestin and clathrin, evidence of recruitment to early and late endosomes and loss of interaction with effector G proteins
Obtained proximity kinetics for ~1000 proteins (proximity map)
Activation of the β2-adrenoceptor receptor leads to:
Increased interaction of arrestin peaking at the same time as the angiotensin receptor
Delayed recruitment of clathrin compared to angiotensin receptor
Later recruitment to endosomes and lysosomes
Obtained proximity kinetics for ~3600 proteins
Given there are ~20,000 proteins in humans, it may be bystanders and not the β2-adrenoceptor interacting with a fifth of all proteins.
Proteins not in the proximity we know definitively do not interact with the POI under these conditions
Benefit of proximity labelling over affinity chromatography
In both study if proteins interact with each other and identify the members of a protein complex
Affinity chromatography:
Negatives
-Require high-quality antibodies (no antibody has no off targets)
-May miss transient or low-affinity interactions (ex. kinases, phosphatases, modifying enzymes)
-Capturing local protein interaction dynamics is often impossible
-Many false positives: sticky proteins (CRAPome)
Uses:
-Stable interactions
Proximity labelling:
Negatives
-Overexpression of protein of interest can generate spurious results (ex. lots of ribosome detected not because the protein is involved in translation)
Solution: Genomic expression (transgenic animal model expressing fusion protein) instead of transient transfection in cell culture
-Potential to block protein interactions at N or C terminus
Solution: Do parallel experiments fusing at N or C termini; interactors that are unique to one fusion may require that terminus to be ‘free’ to be able to interact
-How to distinguish interacting proteins from bystanders
Solution: Target APEX-GFP to organelle the POI resides in and perform mass spectrometry to know which proteins to exclude in results with POI (like CRAPome).
Compartment specific probes example: 2xFYVE for early endosome targeting motif, Lyn kinase for PM, free GFP for cytoplasm
Uses:
-Transient and stable interactions
-Provides information on interaction dynamics
-Obtain proximity map of all proteins POI encounters in it’s lifetime
Affinity chromatography purifies intact protein complexes under mild conditions hoping important interactions persist and non-specific interactions are not favoured.
Proximity labelling works under strongly denaturing conditions so background/non-specific binding is reduced and only specific is captured.
How does the flight-or-fight response activate CaV1.2 and experiment performed to understand
During excitement, exercise and the fight-or-flight response the force of contraction in the heart increases due to β-adrenergic agonists activating the β-adrenergic receptor which activates PKA, increasing Ca2+ influx through L-type Cav1.2 channels in cardiomyocytes
Suggested that exact the mechanism of how Ca2+ influx increases is through PKA phosphorylating Cav1.2 α1C and/or β2B subunits
Test this model by identifying if Cav1.2 can be stimulated by agonists when the channel can’t be phosphorylated
Mutate all Ser and Thr in Cav1.2 α1C and/or β2B that look like a PKA phosphorylation site (into Ala).
Make 4 transgenic mouse lines expressing L-type calcium channels:
Wild type
All 35 sites mutated in alpha
All 28 sites mutated in beta
All sites in alpha AND beta mutated
Through proximity labelling biotinylation (with biotin-phenol) and mass spectrometry, identified ~3800 proteins in the vicinity of each subunit during lifetime, with over 3000 proteins in common (excellent agreement), although there was no control for bystanders (new mouse needed)
Can build interaction map for the channel
Repeat with WT isolated cardiomyocytes with and without isoproterenol. Quench the reaction, purify the biotinylated proteins and compare different in composition of locality with and without agonist
-Monitored ECG: With agonist has higher bpm
-Volcano plot of fold change of abundance of each protein showed PKA is recruited. A series of APEX experiments in different transgenic mice showed only one protein: Rad consistently changes abundance in the vicitinity of the cardiac L-type calcium channel
Concluded Rad phosphorylation by PKA causes release of Rad from proximity of channel and activates Cav1.2
