Transcript proteinstructure

Protein Structure Helps us
Understand Protein Function
If we do know what a protein does,
its structure will tell us how it does it.
 If we don’t know what a protein
does, its structure might give us
what we need to know to figure out
its function.

IIT Biochemistry: 17 Sep 2007
Slide 1 of 56
Plans for Today

Methods of
Determining
Protein Structure

Crystallography
 NMR
 CryoEM
 Specialty
techniques


IIT Biochemistry: 17 Sep 2007



Levels of Protein
Structure
Hydrogen Bonds
Secondary
structure in
globular proteins
Tertiary Structure
Domains
Slide 2 of 56
Warning: Specialty Content!



I determine protein structures (and
develop methods for determining protein
structures) as my own research focus
So it’s hard for me to avoid putting a lot
of emphasis on this material
But today I’m allowed to do that, because
it’s the stated topic of the day.
IIT Biochemistry: 17 Sep 2007
Slide 3 of 56
Structures: Fourier
transforms of diffraction
results





Position of spots tells you how big the unit cell is
Intensity tells you what the contents are
We’re using electromagnetic radiation, which
behaves like a wave, exp(2ik•x)
Therefore intensity Ihkl = C*|Fhkl|2
Fhkl is a complex coefficient in the Fourier
transform of the electron density in the unit cell:
(r) = (1/V) hkl Fhkl exp(-2ih•r)
IIT Biochemistry: 17 Sep 2007
Slide 4 of 56
F
The phase problem



a
Note that we said Ihkl = C*|Fhkl|2
That means we can figure out
|Fhkl| = (1/C)√Ihkl



But we can’t figure out the direction of F:
Fhkl = ahkl + ibhkl = |Fhkl|exp(ihkl)
This direction angle is called a phase angle
Because we can’t get it from Ihkl, we have a
problem: it’s the phase problem!
IIT Biochemistry: 17 Sep 2007
Slide 5 of 56
b
What can we learn





Electron density map + sequence  we can
determine the positions of all the non-H atoms in
the protein—maybe!
Best resolution possible: Dmin =  / 2
Often the crystal doesn’t diffract that well, so
Dmin is larger—1.5Å, 2.5Å, worse
Dmin ~ 2.5Å tells us where backbone and most
side-chain atoms are
Dmin ~ 1.2Å: all protein atoms, most solvent,
some disordered atoms
IIT Biochemistry: 17 Sep 2007
Slide 6 of 56
What does this look like?


Takes some
experience to
interpret
Automated fitting
programs work
pretty well with
Dmin < 2.1Å
ATP binding to a protein of
unknown function: S.H.Kim
IIT Biochemistry: 17 Sep 2007
Slide 7 of 56
How’s the field changing?



1990: all structures done by
professionals
Now: many biochemists and molecular
biologists are launching their own
structure projects as part of broader
functional studies
Fearless prediction: by 2020,
crystallographers will be either
technicians or methods developers
IIT Biochemistry: 17 Sep 2007
Slide 8 of 56
Macromolecular NMR





NMR is a mature field
Depends on resonant interaction between EM
fields and unpaired nucleons (1H, 15N, 31S)
Raw data yield interatomic distances
Conventional spectra of proteins are too muddy
to interpret
Multi-dimensional (2-4D) techniques:
initial resonances coupled with additional ones
IIT Biochemistry: 17 Sep 2007
Slide 9 of 56
Typical protein 2-D spectrum


Challenge:
identify which
H-H distance is
responsible for a
particular peak
Enormous
amount of
hypothesis
testing required
IIT Biochemistry: 17 Sep 2007
Prof. Mark Searle,
University of Nottingham
Slide 10 of 56
Results



Often there’s a family of structures that
satisfy the NMR data equally well
Can be portrayed as a series of threads
tied down at unambiguous assignments
They portray the protein’s structure in
solution
IIT Biochemistry: 17 Sep 2007
Slide 11 of 56
Comparing NMR to X-ray





NMR family of structures often reflects real
conformational heterogeneity
Nonetheless, it’s hard to visualize what’s
happening at the active site at any instant
Hydrogens sometimes well-located;
they’re often the least defined atoms in an Xray structure
The NMR structure is obtained in solution!
Hard to make NMR work if MW > 25 kDa
IIT Biochemistry: 17 Sep 2007
Slide 12 of 56
What does it mean when NMR
and X-ray structures differ?





Lattice forces may have tied down or moved
surface amino acids in X-ray structure
NMR may have errors in it
X-ray may have errors in it (measurable)
X-ray structure often closer to true atomic
resolution
X-ray structure has built-in reliability checks
IIT Biochemistry: 17 Sep 2007
Slide 13 of 56
Cryoelectron
microscopy



Like X-ray crystallography,
EM damages the samples
Samples analyzed < 100K
survive better
2-D arrays of molecules



Spatial averaging to improve
resolution
Discerning details ~ 4Å resolution
Can be used with crystallography
IIT Biochemistry: 17 Sep 2007
Slide 14 of 56
Circular dichroism




Proteins in solution can
rotate polarized light
Amount of rotation varies
with 
Effect depends on
interaction with secondary
structure elements, esp. 
Presence of characteristic
 patterns in presence of
other stuff enables estimate
of helical content
IIT Biochemistry: 17 Sep 2007
Slide 15 of 56
Poll question:
discuss!

Which protein would yield
a more interpretable CD
spectrum?




(a) myoglobin
(b) Fab fragment of
immunoglobulin G
(c) both would be fully
interpretable
(d) CD wouldn’t tell us
anything about either
protein
IIT Biochemistry: 17 Sep 2007
Slide 16 of 56
Ultraviolet spectroscopy





Tyr, trp absorb and fluoresce:
abs ~ 280-274 nm; f = 348 (trp), 303nm
(tyr)
Reliable enough to use for estimating
protein concentration via Beer’s law
UV absorption peaks for cofactors in various
states are well-understood
More relevant for identification of moieties
than for structure determination
Quenching of fluorescence sometimes
provides structural information
IIT Biochemistry: 17 Sep 2007
Slide 17 of 56
Solution scattering






Proteins in solution scatter X-rays
in characteristic, sphericallyaveraged ways
Low-resolution structural
information available
Does not require crystals
Until ~ 2000 you needed high
[protein]
Thanks to BioCAT, SAXS on dilute
proteins is becoming more feasible
Hypothesis-based analysis
IIT Biochemistry: 17 Sep 2007
Slide 18 of 56
Fiber
Diffraction



Some proteins, like many
DNA molecules, possess
approximate fibrous order
(2-D ordering)
Produce characteristic fiber
diffraction patterns
Collagen, muscle proteins,
filamentous viruses
IIT Biochemistry: 17 Sep 2007
Slide 19 of 56
X-ray spectroscopy





All atoms absorb UV or X-rays at
characteristic wavelengths
Higher Z means higher energy,
lower for a particular edge
Perturbation of absorption spectra
at E = Epeak +  yields neighbor
information
Changes just below the peak yield
oxidation-state information
X-ray relevant for metals, Se, I
IIT Biochemistry: 17 Sep 2007
Slide 20 of 56
Levels of Protein Structure

We conventionally describe proteins at
four levels of structure, from most local to
most global:




Primary: linear sequence of peptide units and
covalent disulfide bonds
Secondary: main-chain H-bonds that define
short-range order in structure
Tertiary: three-dimensional fold of a
polypeptide
Quaternary: Folds of multiple polypeptide
chains to form a complete oligomeric unit
IIT Biochemistry: 17 Sep 2007
Slide 21 of 56
What does the primary
structure look like?




-ala-glu-val-thr-asp-pro-gly- …
Can be determined by amino acid sequencing of
the protein
Can also be determined by sequencing the
gene and then using the codon information to
define the protein sequence
Amino acid analysis means percentages; that’s
less informative than the sequence
IIT Biochemistry: 17 Sep 2007
Slide 22 of 56
Components of secondary
structure




, 310,  helices
pleated sheets and
the strands that
comprise them
Beta turns
More specialized
structures like
collagen helices
IIT Biochemistry: 17 Sep 2007
Slide 23 of 56
An accounting for secondary
structure: phospholipase A2
IIT Biochemistry: 17 Sep 2007
Slide 24 of 56
Alpha helix
IIT Biochemistry: 17 Sep 2007
Slide 25 of 56
Characteristics of  helices




Hydrogen bonding from amino nitrogen
to carbonyl oxygen in the residue 4
earlier in the chain
3.6 residues per turn
Amino acid side chains face outward
~ 10 residues long in globular proteins
IIT Biochemistry: 17 Sep 2007
Slide 26 of 56
What would disrupt this?


Not much: the side chains
don’t bump into one another
Proline residue will disrupt it:



Main-chain N can’t H-bond
The ring forces a kink
Glycines sometimes disrupt
because they tend to be
flexible
IIT Biochemistry: 17 Sep 2007
Slide 27 of 56
Other helices


NH to C=O four residues earlier is
not the only pattern found in
proteins
310 helix is NH to C=O three
residues earlier


More kinked; 3 residues per turn
Often one H-bond of this kind at Nterminal end of an otherwise -helix
  helix: even rarer: NH to C=O
five residues earlier
IIT Biochemistry: 17 Sep 2007
Slide 28 of 56
Beta strands



Structures containing roughly extended
polypeptide strands
Extended conformation stabilized by
inter-strand main-chain hydrogen bonds
No defined interval in sequence number
between amino acids involved in H-bond
IIT Biochemistry: 17 Sep 2007
Slide 29 of 56
Sheets: roughly
planar




Folds straighten H-bonds
Side-chains roughly
perpendicular from sheet
plane
Consecutive side chains
up, then down
Minimizes intra-chain
collisions between bulky
side chains
IIT Biochemistry: 17 Sep 2007
Slide 30 of 56
Anti-parallel
beta sheet



Neighboring strands extend in opposite
directions
Complementary C=O…N bonds from
top to bottom and bottom to top strand
Slightly pleated for optimal H-bond
strength
IIT Biochemistry: 17 Sep 2007
Slide 31 of 56
Parallel
Beta Sheet




N-to-C directions are the same for both
strands
You need to get from the C-end of one
strand to the N-end of the other strand
somehow
H-bonds at more of an angle relative to the
approximate strand directions
Therefore: more pleated than anti-parallel
sheet
IIT Biochemistry: 17 Sep 2007
Slide 32 of 56
Beta turns




Abrupt change in direction
, angles are
characteristic of beta
Main-chain H-bonds
maintained almost all the way
through the turn
Jane Richardson and others
have characterized several
types
IIT Biochemistry: 17 Sep 2007
Slide 33 of 56
Collagen triple helix


Three left-handed helical
strands interwoven with a
specific hydrogen-bonding
interaction
Every 3rd residue
approaches other strands
closely: so they’re glycines
IIT Biochemistry: 17 Sep 2007
Slide 34 of 56



Poll question
Remember that there are
about 3.6 residues per turn
in an alpha helix.
Suppose you had a helical
protein that was sitting on,
not in, a phospholipid
bilayer so that the side
chains point inward and
outward along the surface.
Which of the following
sequences would be the
most stable in this
environment?
IIT Biochemistry: 17 Sep 2007
Slide 35 of 56
Options

Assume side chain of
residue 2 points DOWN
into the bilayer:




(a) GADHKYEKLRG
(b) GLDGIVESVGG
(c) AKRTTVWKDKD
(d) YRNNADRRKLG
IIT Biochemistry: 17 Sep 2007
Slide 36 of 56
Tertiary Structure




The overall 3-D arrangement of atoms
in a single polypeptide chain
Made up of secondary-structure
elements & locally unstructured strands
Described in terms of sequence,
topology, overall fold, domains
Stabilized by van der Waals
interactions, hydrogen bonds,
disulfides, . . .
IIT Biochemistry: 17 Sep 2007
Slide 37 of 56
Quaternary
structure


Arrangement of individual polypeptide
chains to form a complete oligomeric,
functional protein
Individual chains can be identical or
different


If they’re the same, they can be coded for
by the same gene
If they’re different, you need more than
one gene
IIT Biochemistry: 17 Sep 2007
Slide 38 of 56
Not all proteins have all four
levels of structure



Monomeric proteins don’t have
quaternary structure
Tertiary structure: subsumed into
2ndry structure for many structural
proteins (keratin, silk fibroin, …)
Some proteins (usually small ones)
have no definite secondary or tertiary
structure; they flop around!
IIT Biochemistry: 17 Sep 2007
Slide 39 of 56
Note about disulfides
H




Cysteine residues brought S
into proximity under
H
oxidizing conditions can C
form a disulfide
Forms a “cystine” residue
Oxygen isn’t always the
oxidizing agent
Can bring sequence-distant
residues close together and
stabilize the protein
IIT Biochemistry: 17 Sep 2007
H
S
H
H
+
(1/2)O 2
H2O
H
H
C
C
S
S
H
Slide 40 of 56
H
H
C
Hydrogen bonds, revisited

Biological settings, H-bonds are almost
always:



Between carbonyl oxygen and hydroxyl:
(C=O ••• H-O-)
between carbonyl oxygen and amine:
(C=O ••• H-N-)
These are stabilizing structures



Any stabilization is (on its own) entropically
disfavored;
Sufficient enthalpic optimization overcomes
that!
In general the optimization is ~ 1- 4 kcal/mol
IIT Biochemistry: 17 Sep 2007
Slide 41 of 56
Secondary structures in
structural proteins





Structural proteins often have uniform
secondary structures
Seeing instances of secondary structure
provides a path toward understanding them in
globular proteins
Examples:
Alpha-keratin (hair, wool, nails, …): -helical
Silk fibroin (guess) is -sheet
IIT Biochemistry: 17 Sep 2007
Slide 42 of 56
Alpha-keratin



Actual -keratins
sometimes contain helical
globular domains
surrounding a fibrous
domain
Fibrous domain: long
segments of regular helical bonding patterns
Side chains stick out from
the axis of the helix
IIT Biochemistry: 17 Sep 2007
Slide 43 of 56
Silk
fibroin


Antiparallel beta
sheets running
parallel to the
silk fiber axis
Multiple repeats
of (Gly-Ser-GlyAla-Gly-Ala)n
IIT Biochemistry: 17 Sep 2007
Slide 44 of 56