TB6: protein interactions

Why study protein itneractions?

• Interactions between molecules defines life - proteins are key players of this

  • The specific, reversible intermolecular interactions are key to understanding living organisms

what are the types of different protein interactions?

Protein protein interactions

protein ligand interactions

protein nucleic acid interactions

What are protein-protein interactions involved in?

• Regulation and signalling

• Localisation

• Assembly of multicomponent cellular machines

  • Eg. complexes such as ATP-synthase, chaperonins and the proteasome would not function without PPIs between different subunits of them

• Generative cooperativity, allowing tuning of responses

• Providing mechanical support to cells

  • Eg. keratin, collagen and microtubules all rely on PPIs

• The immune system

what are protein-ligand interacitions involved in>

• Enzyme interactioins

  • This can be with substrates, or with co-factors

• Receptors and signalling molecules interacting to allow signal transduction

• Allosteric modulators interacting with protein targets

• Membrane lipids

• Drugs

what are protein-nucleic acid interactions involved in?

• Processes such as transcription, translation, replication, gene regulation and processing of transcripts all rely on PNIs

  • These are essential to biology!!

what are the properties of most protein interactions?

In biological systems, protein interactions are specific (and reversible most of the time)

why is reversibility of protein interactions important?

This reversibility allows responses to be made under different conditions, and also allows protein turnover

how is specificity achieved in protein interactions?

• Specificity is achieved through the unique spatial array of chemical groups formed as a globular protein folds 

  • This is achieved through the rich diversity of exposed chemical groups on a protein surface

    • Such as positive vs negative charges, polar vs apolar patches, H-bons donors vs acceptors

  • The range of surface amino acids are less important for folding stability as they ddo not fold into the hydrophobic core - this allows them to be much more varied to form different interactions

  • This variation on the surface produces a pattern of patches that convey 2+3D complementarity with other proteins - this allows different interactions

    • These interactions between proteins are dominated by weak noncovalent interactions - these convey specificity and reversibility of intermolecular binding

2 classes of PPIs?

• reversible PPIs and obligate oligomers

describe reversible PPIs

• Reversible PPIs have reversibility and can be studied to determine affinity and regulation of these interactions

• These are often involved in signalling pathways and regulatory processes

describe obligate oligomers?

• Obligate oligomers have low reversibility. Once the single monomers are produced they are assembled to form an oligomeric complex. They only really dissociate when the complex is turned over

  • The PPI here provides function

describe the physical properties of PPI surfaces

• Wide range of affinities

• Mostly high specificity

• Shape and sizes of the PPI interfaces are large and variable

  • They are often flat and undulating in morphology though - this differs to active sites

• For those that are buried, the amount of polar and non-polar residues is similar to non interface surfaces

  • This is to prevent aggregation

• Interactions are largely driven through specific interactions conferred by the amino acids present:

  • Salt bridges

  • H bonds

  • Arrangements of small hydrophobic patches

• As well as the types of amino acids present, shape complementarity is also important

• Often enriched in aromatic amino acids

describe the physical properties of buried PPI surfaces?

• Amount of non-polar / polar surface area buried is similar to non-interface protein surfaces

  • for something that isnt always oligomerised, needs to remain soluble to prevent aggregation

  • largely driven through specific itneractions

    • salt bridges

    • hydrogen bonds

    • arrangments of small hydrophobic patches (rather than a sole large one)

  • shape complementarity is important for these interactions

• enriched in aromatic amino acids F, Y, W

• generally smaller in area, less hydrophobic, and more charge, than those between subunits in obligate oligomeric proteins

How can we determine the buried surface area to determine the surface areas of PPIs interaction surfaces?

• Rolling a probe sphere along the surface of the protein allows determination of the surface area

  • this can be done for both individual proteins and the complex to get: Buried surface area = Ax + Ay - Axy → so it is the sum of the proteins by themself subtracted from the entire protein complex area

• Can see a wide range of buried surface

What different regions to PPI interfaces have?

• a core

• a rim

describe the core of a PPI interaction surface?

• the core: centre of the interaction

  • water is completely excluded

  • water wont freely exchange

describe the rim of a PPI interaction surface?

accessible to water and water is exchangeable

how can water molecules contribute to the hydrophobic effect at PPI surfaces?

• Water molecules can contribute to the hydrophobic effect

  • release of water from hydrophobic residues can enhance affinity

  • pulling waters off of charged, hydrated amino acids may be needed to form salt bridges

what affects can water have at PPI interfaces?

• improve complementarity by filling gaps between the proteins

• contribute to interfacial H bonds

  • eg. by bridging between proteins

• screens unfavourable electrostatic interactions

what are hotspots in PPI surfaces?

• positions in PPI interfaces that are important for complex stability

  • these residues have more contribution to the interface

how can we determine PPI hotspots?

• can be determined with alanine scans

  • change each of the residues to alanine and measure the affinity

  • ∂∂G is the change in ∂G

    • those greater than 1.5 in this case is considered the cut off for an important residue

what can substitutions at hot spots do?

reduce affinity

what residues are typically hotspots? what interactions may they participate in?

arg, tyr, trp

hydrophobic, aromatic and polar interactions

what interactions with pi electrons are common at interfaces?

• pi electrons can interact with cations

• pi-pi stacking of aromatic rings

why use kd rather than ka

• this puts affinity into physically intuitive units

• this is an indication at the point of which there is half saturation

how can we get interaction dynamics from ka/kd?

when we assume proteins interact as rigid bodies, what are we assuming? what is the equation?

we assume little or no change in structure, other than side chain rearrangements

this is a simple bimolecular association

describe the conformational changes that can occur at large interface PPIs?

• loop/secondary structure rearrangements

• disorder-order transitions of disordered regions

• domain rearrangements in multi-domain protein

what are the different types of conformational change models for different PPIs?

• Induced fit (conformational change)

• Conformational selection

describe conformational selection of protein binding

• conformational selection - sometimes, the protein adopts a different conformation - this can only bind the other protein

  • and so the initial binding is dependent on the formation of the binding conformation

describe induced fit of protein binding

induced fit - the protein binds the other protein in a conformation that is not the most stable, forming an encountered complex. whilst bound, it rearranges, forming the most stable confirmation

equation for induced fit protein binding?

equation for conformational selection protein binding?

What aspects need to be considered when quantifying binding interactions?

• Circumstances of interactions occurring

  • subcellular localisation

  • binary or within a larger complex

  • PTMs necessary?

• Information wanted from the assay

  • complex formation

  • binding affinity - kd

    • is the affinity sufficiently lower than a wt - can use this in mutagenesis

  • binding kinetics - kon and koff

  • binding thermodynamics - H and S

    • can do SPR, ITC to determine different components of the equations

  • mass, stoichiometry

    • may need size sensitive approaches

  • where is the interface? what does the complex look like?

• Possible sample conditions

  • in vivo

    • eg. may want to know the stage of the cell cycle

    • may need to do in cultuted cells

  • crude mixture or cell extracts

    • crude extracts are easier to work with, but ability to quantify the data is limited

  • purified component or reconstitution

What does a yeast 2 hybrid assay exploit?

• exploits the 2 domain architecture of the yeast transcriptional activator Gal4p

• Domains have distinct functions

• can exploit this, splitting the protein in half and attaching the DNA binding domain and activation domain to different proteins

what are the 2 domains in the gal4p activator used in a two hybrid assay?

• DNA binding domain binds UAS

• the activation domain then recruits rnap for gene transcription

how does the yeast 2 hybrid work?

• The activation domain will only be in the right position if X and Y interact

  • can use a reporter gene

    • eg. lacz- blue/white screening

    • eg. fluorescence expression

advantages of yeast 2 hybrid?

• Low cost

• easy to use

disadvantages of yeast 2 hybrid?

• transcription activation is sensitive

  • there is a high false positive rate

• proteins must be nuclear localised

• yeast may not be relevant cellular context

• limited to binary PPIs

  • Y3H exists but needs more effort

• Requires a well defined hypothesis - need to know the identity of both

describe the use of in-vivo chemical cross linking to determine PPIs?

• incorporate photo-activatable chemical group into bait protein (in cells)

  • this is the protein you already have knowledge of

• Introduce UV light

  • this activates the cross linker, inducing non-specific covalent cross linking

    • this cross links interacting proteins

  • (traps complex)

• This is done in a crude mixture. can then isolate the different cross-linked complexes

  • isolate/purify cross-linked complexes

• detect cross-linked complexes

• In gel Digest

  • this maintains the cross links

• identify cross-linked complexes

  • detect via mass spec and then a databbase

describe the use of FRET to determine PPIs?

• X is attached to CFP, Y is attached to YFP

  • if do not interact - low fret efficiency

  • if they do - high fret

    • get energy transfer

• FRET via photoluminescence

• This tests for bimolecular interactions

describe the use of BRET to determine PPIs?

bioluminescence resonance energy transfer

• X is attached to luciferase

  • in the presence of luciferin and oxygen, this produces light via an oxidation reaction, and produces light

• Y is labeled with YFP, should get an energy transfer

describe the use of split protein sensors to determine PPIs?

• X and Y are known

• we cut another protein with a reporter activity in half

  • fragments on their own lack activity

  • if the proteins interact, this brings the C and N termini together, allowing the reporter to show activity

• Examples:

  • GFP + GFP variants

  • those with measurable enzymatic activity

How can protein interactions be detected in crude mixtures?

use of a pull down

describe the use of pull downs to determine PPIs

• matrix anchored linker - this has a crude mixture poured over it

  • bait protein binds the ligand

  • anything binding the protein, which binds the ligand, will not come off the column (it is retained)

pros of pull downs for PPI identification?

• Discover new binders

• test known binders

• detect large complexes, not only binary PPIs

  • this is due to the use of the crude mixture

cons of pull downs for PPI identification?

• Detect large complexes - may be indirect

• Relatively low sensitivity, not suitable for weak binders

  • this is because the washes through the column can induce dissociation

• not in vivo

how can we get quantitative binding data? include example

• Mostly measured on purified proteins in reconstituted complexes

• Anything that can be detected that changes in proportion to complex formation can be used to measure kd

  • eg. in haeomoglobin, oxygen binding can be detected via uv-vis spectrophotometry

    • extinction coefficient is sensitive to whether oxygen is bound or not

    • the two forms oxygen bound or unbound are very different - this produces different intensities, allowing the ozygen bound to be determined

what is the EMSA used to detect?

protein DNA interactions

main principle of EMSA?

DNA bound to protein is larger than free DNA, and so moves much more slowly in the gel

set up of EMSA?

• use an agarose gel

• we are detecting the nucleic acids, not the protein!!

• proteins binding the dna slow it down, producing a shift

• this can be done as a function of protein cocnentration

  • the bands are quantified to determine free dna vs complexed

what are EMSA gels stained with?

Stain DNA before the gel with covalent dyes/radioisotopes. Or after the gel with a base intercalating fluorophore (eg. ethidium bromide or sybr-GOLD)

what does varying protein concentration allow for in EMSA?

• we measure the change in intensity of bands for free DNA and protein bound DNA

• we use this to get Kd

describe the use of competition experiments in EMSA. what do they reveal?

• use target DNA, protein and an unlabelled competitor DNA

• start with a labelled DNA-protein complex and add in an unlabelled competitor DNA

• If there is a transition after the addition of the competitor, this indicates specificity

• we vary competitor concentration, rather than protein concentration

describe the use of a super shift EMSA

• Free DNA, add some protein (eg. crude mix of different TFs)

  • we then add abs against different tfs

  • this is a larger protein complex , this further perturbs movement through the gel, producing another shift
    
    

How can we analyse EMSAs?

• Gel image intesities can be quantified

• Total band intensities are proportional to concentration

  • DNA free and or DNA complex

• Binding parameters can be quantified

Langmuir isotherm equation for single site ligand binding?

what do we plot when measuring single site ligand binding?

• plot ñ on y vs free ligand concentration

  • this produces a rectangular hyperbola or a direct plot

• or ñ on your vs log free ligand concentration

  • this produces a sigmoidal plot, where it is much easier to extract kd

main assumption for single site ligand binding plots?

assume ligand free = ligand total - this assumes ligand is in excess

what do we plot when measuring multiple site, identical and independent ligand binding?

plot ñ/Free ligand on y and ñ on x. this is a scatchard plot. this gives a straight line with gradient of -1/kd

could also do a log scale for this. a positive deflection indicates positive cooperativity

scatchard eq for measuring multiple site, identical and independent ligand binding?

what do we need to consider when quantifying binding data?

• Does the signal change proportionally with complex formation?

• Can we detect the signal sensitively (high signal-to noise ratio) and in the concentration range we need for detection

• Do I need to label something to detect it and what are the risks of introducing artefacts?

  • Unlabelled (‘native’) always preferred but may not be possible.

  • could ;labelling artefacts prevent complex formation?

• How much sample is required for the technique? Can we makenenough of it in the right form?

  • eg. may need specific PTMs

  • eg. is it folded?

• Is it stable, and is it soluble at the required concentrations?

list of methods used to quantify binding data?

• ITC

• SPR

• Thermal shift assays

• fluorescence polarisation

• microscope thermophoresis

• NMR

• stopped flow

describe thermal sift assays

• when you bind something, the protein becomes stabilised

  • this is because there are more contacts present

• can do cdmelt, DSC, DSF - diffferentways to measure melting tempertature

• best done with small ligands

  • for PPIs, both have a specific melt point, and so can become difficult

  • can be done with ID blinding partners

Describe the fluorescence polarisation experiment?

• Consider a cuvette containing fluorophores (eg. fluorophore labelled proteins)

• Light absorbance and emitted fluorescence depends on the direction of the transition dipole moment.

  • so if the molecule rotates, the dipole moment also rotates

• If polarized light is used to excite, this will only excites a subset of fluorophores (photoselection)

  • this is photoselection

  • those with a transition dipole aligned with the polarised light will be excited more than those not fully aligned

• We use this and the emission polarisation is detected

  • before detection, it is passed through polarisers again

  • can use grids parallel or perpendicular to the inital polariser

    • we detect both of these

Why do we observe a change in anisotropy in FP experiments?

• Use a small ligand with a fluorophores tag, this is very fast tumbling

  • so in the time for excitation and detection, molecules tumble through all orientations and so have low polarisation

• large molecules are low tumbling

  • in the time for excitation and detection, the molecule has tumbled slowly, and so the orientation hasnt changed as much

  • this will have a high anisotropy

  • so a ligand binding will have a high anisotropy in comparison to the free ligand

• so when the ligand binds the large protein, it will have high anisotropy - high A

  • this change in anisotropy is detected!!

What is the polarisation anisotropy?

• The polarisation anisotropy, A is defined as the degree of depolarisation divided by the total fluorescence output

  • (intensity of the parallel polarised light - the intensity of the perpendicular polarised light) / the sum of the two

  • this is dimensionless, and between 0-0.5

equation for polarisation anisotropy?

what is the plot for anisotropy of binding? what are the different points corresponding to?

• We measure anisotropy as a function of protein concentration

• At no protein concentration, the ligand has low anisotropy

• At fully bound, it has a high anisotropy

• Smallest A → Largest A = ∂A

• ∂Amax = difference between fully unbound and fully bound

Equations for fitting of anisotropy data?

how to we get A in terms of free protein and Kd from the Langmuir isoform equation?

• ñ can be scaled from the langmuir isoform to get A in terms of free protein and Kd

  • A = ∂A[Pfree]/Kd+[Pf] +Amin

  • but the values of these scaling values is not importnat, we only care that ∂A is large and that A and Amin are detectable

  • the Amin is because the ligand on its own will have some anisotropy

why does microscale thermophoresis work?

the impact of heat on the protein, which is dependent on the size, which is dependent on the complex and the interaction with hydration cell

describe the process of microscale thermophoresis (MST)

• molecules are flown into capillaries

  • these often have different concentrations

• IR laer introduces a temperature gradients

• molecules will diffuse

• we detect fluorescence in a specific region

• fluorescence on y as a function of time

• laser on - fluorophore diffuses away from the point of heat

• turn laser off - diffuse back into the space

• very simple and sensitive!!

• The extent of movement depends on molecular hydration shell, charge and size

  • these change on binding of the ligand/protein partner

• detection via covalently attached fluorophores, fused GFP or intrinsic tryptophan fluorescence

how do we normalise MST data?

• Normalise using Fmin and Fmax - looking at the proportional change in relation to complex formed

• critical to measure a wide range, as this allows us to define the plateau regions

what are some key considerations in experiments to quantify binding data?

• many assume ligand free and ligand total are true

  • this only holds if there is a lot of excess ligand!!!

• mutagenesis is useful

  • but remember there will be 2 interfaces!!

• gold standard is using multiple orthogonal approaches

  • eg. a tm measurement and spr

• need a big change in anisotropy - need to consider which components you label

What are the two classes of techniques when measuring protein-ligand interactions by NMR?

• protein detected approaches

• ligand detected approaches

what do protein detected NMR approaches measure?

measuring free or bound protein

what do ligand detected NMR approaches measure?

• measure free ligand signal

• some physical properties have been imparted to the free ligand

How can 1H 15N HSQCs be used to measure protein-ligand interactions?

• Can look at chemical shift change

  • these are sensitive to changes in environment and conformation and therefore upon binding of a ligand there will be a shift

• we can calculate the chemical shift perturbations for each residue

how do we find the CSPs via HSQCs when studying protein detected NMR for interactions?

• We use an N15 labelled protein at a fixed concentration, and can introduce variable ligand concentrations

  • this can be done on the same sample via a titration

  • can also be done using many different samples with protein at different ligand concentrations to be recorded one at a time

• then do spectral overlay of different spectrums with different concentratons

• can use this to trace and watch the CSPs happening

• not everything moves - only those involved in binding do

  • 
    

what happens in CSPs when a molecule only binds one specific site?

if it binds to one site, only a subset of resides should be impacted by binding

what do many residues undergoing CSP indicate in protein-ligand interactions?

indicates a conformational change or non specific binding

how do we get the CSP from HSQC?

• we can determine the CSP via Euclidian geometry - we have to take into account the difference in chemical shifts for nitrogen

  • we do this by introducing a scaling factor of 5

what values of CSP are significant`? what does this tell us?

• for the ∂delta value, anything more than 0.1 is significant

  • if there is a significant CSP for a residue, we asume that the ligand is binding direcly to the residue or very close by

how is chemical shift data derived from HSQCs used to determine Kd?

• Using the chemical shift data, we know that the environment is changing because of the ligand binding

• We use the same formulas to fit this data as before - we can use the single site binding equation to do this

• Can do a multiple spectra overlay

  • the chemical shift information for each residue can be quantified

    • we do this for all to determine whether these are involved in the same binding event or not

  • all residues have CSPs - these need to be normalised becuase of the details of the interaction

  • if the kd is the same - its the same binding event

describe how CSPs look for weak interactions?

• If there is a weak interaction (high koff), residues will start at the free position and move along the line to the fully bound position

  • we can track the movement

describe how CSPs look for strong interactions?

• however if koff is very low, there is a slow exchange regime

  • within the timescale, the ligand doesn't come off

  • so we observe 2 populations - ligand free and ligand bound population

  • for his tight binding, the slow exchange produces an intensity drop of the free form, and an increase of the bound form - the intensities are directly proportional to the population

  • the slow exchange can be difficult to interpret - we dont see an incremental movement, and so assignment of the fully free and bound is difficult

are fast or slow exchange interactions easier to analyse via CSPs?

• Fast exchange, weak binding events are easier to study - this is becuase peaks move in a line over the titration, making it easy to assign peaks

how can NMR HSQCs elucidate multiple different binding interactions?

• When we plot the different perturbations, we can see multiple different binding events by fitting different trend lines to different residues. this can be used to validate the presence of different binding sites and determine their kd

how can 1H, 13C methyl CSPs be used for large complexes?

• methyl groups only produce peaks, other proton sites are deuterated - this reduces the number of signals. and therefore reduces the number of amino acids being detected

• 
  

how can structure activity relationships be determined using NMR for drug design?

• protein is screened with lots of small molecule fragments

  • can detect some interaction and confirm specificity

  • this can then be independently optimised to increase its affinity

  • whilst bound we can screen another pool of fragments and find things that bind elsewhere, but nearby

    • this is again optimised to increase affinity

  • we then produce a chemical linker to link these together

  • this can them both bind, greatly increasing affinity

What are the advantages of protein detected NMR approaches?

• Exploits exquisite sensitivity of chemical shifts to report on interactions

• Quantitative information on affinity

• 2D HSQCs can be recorded quickly (minutes)

  • depends on the protein concentration

• High resolution information on location of binding site

  • this is because there is one signal for every amino acid

• Very good (best) at detecting weak interactions

• Pooled small molecule screening possible (but requires deconvolution)

  • if one pool you see aa CSP, dont know which molecule binds -need to deconvolute to determine what is binding

What are the disadvantages of protein detected NMR approaches?

• Requires sample that is labelled, either 15N or 13C, which means usually sample must be expressed and purified from in E. coli

  • e.coli may not be the best expression. system for the protein you are studying

• Typical protein concentrations are 10-100’s μM so cannot easily quantify low Kd

  • if there is a low kd, protein concentration needs to be very below the kd

  • and so not good for quantification of tight binding ligands

• Chemical shift assignments needed for locating binding site

• Requires good quality data, well-resolved HSQCs – the size limitation of NMR can be limiting (broad signals and spectral complexity)

What are the different ligand detected approaches used in NMR to detect protein interactions?

• STD-NMR

• WaterLOGSY

• Relaxation editing

• Transferred NOE

• Paramagnetic relaxation enhancement

• diffusion editing

what do ligand detected NMR approaches rely on?

all detect different chemical properties depending on the bound and free ligand

unbound - fast tumbling fast diffusion and slow relaxation

Process for STD-NMR?

• Protein is selectively irradiated

• Upon ligand binding, some magnetisation is transferred to the ligand, and this can be detected

• We can use difference spectra

  • we record 1D spectra with no protein irradiation - this produces no magnetisation

  • we then record a second experiment with selective saturation of the protein - here, the magnetisation is transferred

  • there will be intensity differences between the peaks on the NMR, differences can be calculated to determine which resonances are perturbed

  • these are used to determine which molecules bind

what 2 1D spectra are used in STD-NMR?

• one where the protein is irradiated

• one where the protein is not irradiated

Describe the process of water-logsy

• Instead of saturating the protein, we instead saturate water

• Bulk water is selective irradiated with a radifrequency pulse

• Some of this is transferred to ligands through space via NOEs

• however the NOE transfer depends on molecule tumbling times

  • small molecules that have bound a large protein will have reduced tumbling and so their NOE value will be smaller and flipped

How does relaxation editing work?

• Looks at which molecules behave like large molecules

  • when bound to a protein, will have some properties

• looking for a change in linewidth

• as molecules get bigger, linewidth increases and signal decreases

• based on CPMG, a way to measure t2, which is related to linewidth

what are the cons of some ligand detected NMR experiments?

• techniques rely on the detection of the free ligand

  • transfer of properties from the bound to free satte

• if there is no exchange bewteen the two, can get no results

  • tight binders cannot detect the diferences