Joint use of SAS and AUC

Transcription

1 Joint use of SAS and AUC Olwyn Byron School of Life Sciences College of Medical, Veterinary and Life Sciences University of Glasgow, Scotland UK

2 Outline Principles underlying AUC How AUC experiments are typically performed The types of systems that can be analysed The data that result Interpretation of data using hydrodynamic modelling Synergy between SAS and AUC A few examples

3 Outline Principles underlying AUC How AUC experiments are typically performed The types of systems that can be analysed The data that result Interpretation of data using hydrodynamic modelling Synergy between SAS and AUC A few examples

4 AUC tutorials Setting up and running AUC experiments Tutorial paper Lebowitz, J., M.S. Lewis, and P. Schuck, Modern analytical ultracentrifugation in protein science: A tutorial review. Protein Science, (9): p AUC user guide from Demeler lab Hardware.pdf

5 The analytical ultracentrifuge (AUC) was invented by Theodor (The) Svedberg Nobel Prize in Chemistry 1926 awarded to The Svedberg "for his work on disperse systems"

6 Svedberg was an interesting man Married 4 times 12 children! Liked to paint Atomics

7 1960 s-80 s AUC = core biochemical/biophysical technology Advice from the Beckman Model E AUC 1964 manual: The Model E, like a woman, performs best when you care. But you needn t pamper it - just give it the understanding it deserves. image from Analy.cal Ultracentrifuge User Guide Volume 1: Hardware, K. L. Planken & V. Schirf, 2008 (hhp://

8 The modern AUC a high speed preparative UC with optics Beckman Coulter ProteomeLab XL-A/XL-I; k

9 Inside the rotor chamber absorbance slit assembly drive spindle condenser lens for interference optics radiometer monochromator mount image from Analytical Ultracentrifuge User Guide Volume 1: Hardware, K. L. Planken & V. Schirf, 2008 (

10 Inside the Beckman Coulter XL-I Rayleigh interference optics vacuum chamber rotor UV-vis optics sample cell (minus casing)

11 Relationship between data and sample: absorbance image from Ralston, 1993 hhps://

12 2 modes of operation - several data types Sedimentation velocity (SV) Sedimentation equilibrium (SE) In solution Non-destructive Self-cleaning Absolute

13 SV versus SE SV: observe movement of sedimentation boundary Change in (sometimes complex) boundary over time is due to Sedimentation Diffusion SE: rotor spun more slowly so diffusion can balance sedimentation - system reaches thermodynamic equilibrium Observe no change in boundary over time Unless sample is degrading or changing in some other way

14 Sedimentation velocity (SV): shape & homogeneity absorbance t=0 radius t=1 h heterogeneity determination sedimentation (s) & diffusion (D) coefficients (shape) association/dissociation constant (K a /K d ) stoichiometry t=3 h

15 Sedimentation equilibrium (SE): mass & self-association absorbance t=0 radius t=1 h t=3 h t 24 h+ M association/dissociation constant (K a /K d ) stoichiometry non-ideality (B)

16 Outline Principles underlying AUC How AUC experiments are typically performed The types of systems that can be analysed The data that result Interpretation of data using hydrodynamic modelling Synergy between SAS and AUC A few examples

17 Double sector sample holder ( cell ) image from Beckman AUC manual

18 Sample requirements Sample volume SV 360 µl (up to 480 µl) in 12 mm pathlength 90 µl (up to 120 µl) in 3 mm pathlength SE 20 µl (8-channel centrepiece - interference optics only) 80 µl (2- or 6-channel centrepiece) Sample concentration Absorbance optics: A λ in 12 mm pathlength cell λ = nm Interference optics: typically mg/ml Sample reference Absorbance optics: can be column eluant or dialysate better Interference optics: must be dialysate Typical multiplexing: 3 or 7 cells/run Up to 28 samples per run

19 SE: Which speed? Chervenka, C. H. A Manual of Methods for the Analytical Ultracentrifuge. Spinco Division, Beckman Instruments, Palo Alto, 1969

20 SE: Which speed? Rotor speed chosen to optimise shape of equilibrium distribution Rule of thumb: at lowest chosen rotor speed, effective molecular weight (σ) = 1 At subsequent speeds, speed factor = 1.5 Ensures that in global fitting of data at different speeds, data are different from each other

21 Outline Principles underlying AUC How AUC experiments are typically performed The types of systems that can be analysed The data that result Interpretation of data using hydrodynamic modelling Synergy between SAS and AUC A few examples

22 The types of systems that can be analysed Proteins Nucleic acids Carbohydrates Polymers Colloids Complexes

23 Questions that can be answered by AUC Is the sample homogeneous or heterogeneous? If heterogeneous, is it in molecular weight, shape, or both? If heterogeneous, does heterogeneity depend on ph, salt, buffer, etc? Is the sample pure enough for X-ray crystallography, SAXS, SANS or NMR? Does the sample: self-associate? aggregate? What is the molecular weight of the sample, or a mixture of samples? Does the sample bind to a ligand? What is the stoichiometry of binding? What is the Kd? Is the association state/conformation of the sample affected by tagging?

24 More questions that can be answered by AUC What is the sedimentation and diffusion coefficient of the sample? Is it globular or unfolded/disordered? Is the conformation dependent on salt, ph, ligand concentration, deuteration, etc? Do mutations affect Kd, conformation, stoichiometry, etc? Is the sample affected by crowding?

25 Outline Principles underlying AUC How AUC experiments are typically performed The types of systems that can be analysed The data that result Interpretation of data using hydrodynamic modelling Synergy between SAS and AUC A few examples

26 SV: radial movement recorded as function of time

27 SE: data recorded until no change

28 All AUC data analysis software is freely available The RASMB website Reversible Associations in Structural and Molecular Biology Access to freely available software Subscription to AUC-related discussion group Schuck lab (SEDFIT, SEDPHAT) Demeler lab (UltraScan III (including SOMO))

29 Many methods & programs for SV data analysis Too many for comprehensive review here Model independent: dc/dt (Stafford, SedAnal) Eliminates time invariant noise. Resultant curves can be fitted with Gaussians to reveal species content and sedimentation coefficients. c(s) (Schuck, Sedfit) Good for first look at data to get an idea of number of species. Not a proper fit to data. van Holde-Weischet (Demeler, UltraScan III) Diffusion corrected s distribution. Good for detection of aggregates and identification of underlying model. Model dependent: Non-interacting discrete species (Schuck, Sedfit) Up to 4 separate species can be fitted. Self-association (Stafford, SedAnal; Demeler, UltraScan III) Determination of K d, k on, k off, stoichiometry

30 s is influenced by solvent density & viscosity and sample density sample partial specific volume solvent density solvent viscosity s 20,w = ( 1! v" ) 20,w # T,b s ( 1! v" ) # T,b 20,w T,b sedimentation coefficient standardised to solvent of 20 C 1.5 for typical aqueous solvent at 4 C experimental sedimentation coefficient determined in e.g. buffer (b) at T C

31 SEDNTERP : Calculation of ρ, η and partial specific volume online hhp://sednterp.unh.edu/

32 SV: species can resolve into separate boundaries

33 SEDFIT c(s) analysis: how many species + s of species 1: Load SV data hhp://

34 2: Specify parameters hhp://

35 3: Set meniscus, cell base and analysis limits hhp://

36 4: Run hhp://

37 5: Subtract time and radial invariant noise hhp://

38 6: Fit (with solutions to the Lamm equation) hhp://

39 7: Integrate to obtain estimate of concentration of species and weight-average values hhp://

40 Integrating c(s) peaks reveals region of boundary that contains species

41 Integrating c(s) peaks reveals region of boundary that contains species

42 Integrating c(s) peaks reveals region of boundary that contains species

43 Integrating c(s) peaks reveals region of boundary that contains species

44 Integrating c(s) peaks reveals region of boundary that contains species

45 Integrating c(s) peaks reveals region of boundary that contains species

46 Self-association: SE data are the sum of exponentials " # = $%&'()" * ++,-.# / "# * / 01 monomer!! +$%&') / ()" * +()23 / +) /,+,-.# / "# * / 01 +$%&') 4 ()" * +()23 4 +) 4,+,-.# / "# * / 01 +$%&') 5 ()" * +()23 5 +) 5,+,-.# / "# * / n 2 1-n 3 1-n 4

47 Self-association: deconvolution into individual components experimental data = sum of species tetramer dimer monomer

48 Self-association: best model revealed by residuals absorbance residuals absorbance residuals radius (cm) radius (cm)

49 Detergent solubilised proteins: density matching SE In SE bouyant molecular weight is determined: In many AUC expts we want to observe self-association Density matching is a good method for self-associating membrane proteins Burgess, N. K., Stanley, A. M. & Fleming, K. G. (2008). Determination of membrane protein molecular weights and association equilibrium constants using sedimentation equilibrium and sedimentation velocity. In Methods in Cell Biology (J. Correia & H. W Detrich, III, eds.), 84, Academic Press

50 Density matching SE: experimental conditions Experimental conditions adjusted such that: solvent ρ = effective ρ of bound detergent So detergent becomes effectively invisible to centrifugal field SE data can be analysed with standard methods 0

51 BUT..this method works only in certain conditions Solvent density must be adjusted with D 2 O or D 2 18 O Alternatives would be e.g. sucrose or other co-solvent Affect chemical potential Lead to preferential binding and/or exclusion of water or additional co-solvent at protein surface But use of D 2 O or D 2 18 O limits detergents that can be used ρ D 2 O = 1.1 g/ml! of the detergent must be between that of water and D 2 O i.e. 0.9! 1.0 ml/g Eliminates: dodecylmaltoside (ρ = 1.21 g/ml,! = 0.83 ml/g) β-octylglucoside (ρ = 1.15 g/ml,! = 0.87 ml/g) Suitable: C8E5 (! = ml/g) C14SB (density matched by 13% D 2 O in 20 mm Tris-HCl, 200 mm KCl) Dodecylphosphocholine (DPC, density matched by 52.5% D 2 O in 50 mm Tris- HCl, 100 mm NaCl) Burgess, Stanley & Fleming (2008). In Methods in Cell Biology (J. Correia & H. W Detrich, III, eds.), 84, Academic Press

52 Determination of density-matching point for C14SB Determine % of D 2 O required to density match C14SB micelles in background of other buffer components 1mM C14SB in 50mM Tris, 200mM NaCl made in 0, 7, 14, 21, 28, 35, 42% D 2 O Reference solvent the same minus detergent SE observed with interference optics Rotor speed 40k rpm, T = 15 C fringes ρ micelle > ρ solvent ρ micelle = ρ solvent ρ micelle < ρ solvent radius slope of distribution % D 2 O Burgess, Stanley & Fleming (2008). In Methods in Cell Biology (J. Correia & H. W Detrich, III, eds.), 84, Academic Press

53 Outline Principles underlying AUC How AUC experiments are typically performed The types of systems that can be analysed The data that result Interpretation of data using hydrodynamic modelling Synergy between SAS and AUC A few examples

54 s = deviation from sphericity + hydrodynamic hydration #$%" &!#' " =!! ( ) * M, f 0 M, f>f 0 M, f>f 0 M, f>>f 0

55 Sedimentation coefficient is a constraint for SAS modelling s =? S For one sphere f 0 = 6!"R 0 For an assembly of N spheres an approximate solution is where

56 Several freely available programs for HBM A more exact expression for f t together with expressions for other hydrodynamic and related parameters are encoded in HBM software: José García de la Torre et al. (Universidad Murcia, Spain) HYDRO Computes hydrodynamic & other parameters for any bead model HYDROPRO Computes hydrodynamic & other parameters for models constructed from pdb files And many other programs. Mattia Rocco, Emre Brookes Generates HBMs from pdb files, computes hydrodynamic & other parameters with realistic hydration Byron (2008) Methods in Cell Biology Byron (2015) Introduction: Calculation of hydrodynamic parameters in Analytical Ultracentrifugation: Instrumentation, Software and Application (S. Uchiyama, F. Arisaka, W. F. Stafford and T. Laue, editors)

57 SOMO - construction of intelligently hydrated bead models from atomic coordinates B A B D C C E Popped beads AtoB Trans exposed acidic exposed basic exposed polar exposed nonpolar/hydrophobic mainchain buried Water of hydration included in each bead Bead overlaps removed heirarchically Reducing radii + translating bead centres outwards Beads overlapping by > preset threshold are fused ( popped ) Buried beads excluded from hydrodynamic calculations Reduces cpu time Rai, Nöllmann, Spotorno, Tassara, Byron & Rocco, Structure (2005) 13,

58 SOMO is a subprogram of UltraScan III Mattia Rocco/ Borries Demeler/ Emre Brooks Rai et al. (2005) Structure ; Brookes et al. (2010) Eur. Biophys. J; Brookes et al., (2010) Macromol. Biosci.

59 Outline Principles underlying AUC How AUC experiments are typically performed The types of systems that can be analysed The data that result Interpretation of data using hydrodynamic modelling Synergy between SAS and AUC A few examples

60 Outline Principles underlying AUC How AUC experiments are typically performed The types of systems that can be analysed The data that result Interpretation of data using hydrodynamic modelling Synergy between SAS and AUC A few examples

61 Example 1 FLEXIBLE PROTEIN

62 Mystery protein!

63 Elongated model

64 Globular model

65 HPLC-SAXS

66 HPLC-SAXS at least 2 species

67 Rg changes with HPLC elution time

68 Change in Rg is not discrete

69 Example 2 HETERO-ASSOCIATION

70 What is the stoichiometry & structure of human PDC E3BP:E3 subcomplex? E3 forms tight dimer E3 binds to E3BP 2:1 or 1:1?

71 SV titration: stoichiometry is 2:1 Expt 1: SV of E3 alone; SV of E3BP-DD alone Determine their s Expt 2: SV of E3BP-DD+E3 At what ratio does E3BP-DD peak vanish? This reveals stoichiometry: 2:1 Note 2 complex peaks Different conformations s 6 S peak less compact s 8 S peak more compact E3BP E3 E3BP:E3 = 4:1 3:1 2:1 1:1 1:2 1:3 Mischa Smolle Smolle et al., JBC (2006)

72 SE titration: stoichiometry is 2:1 Whole-cell weight-average M (M w,app ) determined e.g. using species analysis in SEDPHAT with 1 species only No model assumed When E3BP-DD is in excess M w,app < M complex until complex is formed When E3 is in excess M w,app < M complex because excess E3 lowers M w,app??? Why M w,app M complex at 2:1??? (E3BP-DD) 2 E3 (E3BP-DD) 1 E3 E3 E3BP-DD:E3 molar ratio Mischa Smolle Smolle et al., JBC (2006)

73 SAXS data characteristic of asymmetric elongated E3BP:E3 complex p(r) vs r ab initio rigid body modelling Mischa Smolle Smolle et al., JBC (2006)

74 Only 1 E3BP-DD lipoyl domain is docked into E3 Brautigam et al. Structure (2006) 14, Mischa Smolle Smolle et al., JBC (2006)

75 Example 3 HETERO-ASSOCIATION: FLEXIBLE PROTEIN + MEMBRANE PROTEIN

76 Pectocins M1& M2: bacteriocins from Pectobacterium spp. have no IUTD - how do they cross the OM? M- class cytotoxic domain plant- like ferredoxin domain Kesha Iosts & Rhys Grinter Grinter, Josts et al., Molecular Microbiology, 2014, 93,

77 PM2 solution conformation inconsistent with P2 1 crystal structure (χ 2 = 1.362) Kesha Iosts & Rhys Grinter Grinter, Josts et al., Molecular Microbiology, 2014, 93,

78 US-SOMO DMD + GAJOE: bimodal ensemble of models that better fits SAXS data (χ 2 = 0.827) US-SOMO DMD: 5,000 models created GAJOE: best fit ensemble selected Kesha Iosts & Rhys Grinter Grinter, Josts et al., Molecular Microbiology, 2014, 93,

79 PM2 re-crystallisation: space group P Kesha Iosts & Rhys Grinter Grinter, Josts et al., Molecular Microbiology, 2014, 93,

80 P conformation permits PM2 to fit through TonB Kesha Iosts & Rhys Grinter Grinter, Josts et al., Molecular Microbiology, 2014, 93,

81 Determination of density-matching point for C14SB: 17.4 %

82 PM2+FupA one species in SV

83 HPLC-SAXS of PM2+FupA PM2+FupA FupA