Resonance Assignment for Proteins

1
Resonance Assignment for Proteins

• Classical homonuclear (1H-1H) assignment methods
• 1. Spin system assignments
• 2. Sequence-specific assignments
• 3. Sequential vs. Main-chain Directed Assignment
• Modern methods Use of heteronuclear shift
  correlation, triple resonance experiments, etc.

2
Resonance assignments

• in order to be able to actually solve the
  structure of a protein, we first have to assign
  the spectrum
• each peak corresponds to some proton within some
  amino acid residue. Is the sharp peak at -0.8
  ppm a valine, leucine or isoleucine methyl?
• even if we knew it was a valine methyl, which
  valine does it belong to?
• even if we knew it was Val30, which of the two
  methyls is it?

sequence of lysozyme KVFGRCELAAAMKRHGLDNYRGYSLGNW
VCAAKFESNFNTQATNRNTDGSTDYGILQINSRWWCNDGRTPGSRNLCNI
PCSALLSSDITASVNCAKKIVSDGNGMNAWVAWRNRCKGTDVQAWIRGCR
L
3
Levels of resonance assignment

• spin system assignment is it Val, Ile or Leu?
• sequence-specific assignment is it Val 30 or Val
  87?
• stereospecific assignment is it the pro-R or
  pro-S methyl of Val 87?

4
Classical protein NMR the basic plan
In classical protein NMR, assignments are made
by using 2-dimensional experiments to establish
correlations between different 1H resonances.
Recognition of characteristic patterns and
networks of correlations then allows assignments
to be made. Resonances are correlated either
through-bond, mediated by the scalar coupling,
or through-space, mediated by the spin dipolar
coupling (nuclear Overhauser effect).
through-bond (J-coupling)
through-space (nOe)
H
H
5
Basic features of 2D spectra
HA
HB
2 Å
HA
HB
diagonal peak correlation of a resonance with
itself
1H chemical shift (ppm)
crosspeak correlation of two different
resonances by short interatomic distance or
through-bond connection
1H chemical shift (ppm)
6
Spin systems and scalar coupling networks

• a spin system is a set of 1H resonances connected
  (either directly or indirectly) by 1H-1H scalar
  couplings
• generally this means networks of 1H in which each
  1H is connected to another member of the network
  by three or fewer covalent bonds--longer-range
  couplings are generally small, so experiments
  based on resonance correlation via scalar
  coupling will generally not detect four- and
  five-bond couplings

Hc
indirect connection
H
H
H
H
Ha
Hb
geminal coupling (two-bond) J -12 to -15 Hz
vicinal coupling (three-bond) J 2-14 Hz
example of a spin system
7
2D COSY/TOCSY--gtspin systems

• COSY and TOCSY give crosspeaks when resonances
  are linked through scalar coupling
• COSY gives crosspeaks when 2- and 3-bond
  couplings are present
• in TOCSY, longer range correlations are seen due
  to relays of 3-bond couplings
• these two techniques can be used to assign spin
  systems through recognition of coupling patterns
• recognition of the patterns at right also takes
  into account qualitative chemical shift
  information--the beta methyl of alanine, for
  instance, might be anywhere from 0.9-1.7 but is
  never 3 or 4.

o crosspeaks visible in COSY , crosspeaks
visible in TOCSY
8
Example of lysine spin system
CO
Hd
HN
Ha
Hb
Hg
He
NH3
e
a
b
g
d
Hd
Hg
Hb
He
Ha
9
Sequence-specific assignments

• suppose we have the sequence of our protein from
  some independent measurement
• suppose weve assigned an isoleucine spin system,
  and theres only one isoleucine in the sequence
  (unique), at position 48. Then we know our
  isoleucine is Ile48.
• there wont be very many unique amino acid
  residues in a protein, however.
• but there will be many unique dipeptide sequences
• but in order to use this fact, we need to be able
  to connect adjacent residues.

unique residues (arrows) and unique
dipeptide sequences in lac repressor
10
Linking spin systems using nOes

• because the nOe depends upon interatomic
  distance and not upon J coupling, it can be used
  to connect spin systems which are adjacent in
  space but not part of the same spin system, for
  instance two residues adjacent in the sequence
• general nomenclature for
• interatomic distance between
• atoms A and B in residues i and j
• dAB(i,j)

• nOe correlations are denoted using the distance
  nomenclature, e.g. dbN(i,i1) nOe or
  dbN (i,i1) correlation
• daN(i,i1), dNN(i,i1), and sometimes
  dbN(i,i1) are used to connect adjacent residues

11
2D NOESY linking spin systems
diagonal no magnetization transferred
5.HN/6.HN
4.HN/5.HN
crosspeaks intersection of chemical shifts of
atoms which are close in space, i.e. lt 5 Å
1H
portion of 2D NOESY of P22 cro
showing dNN(i,i1) correlations-- can walk
along the chain from one residue to the
next. Residues 3-7 shown.
6.HN/7.HN
3.HN/4.HN
1H
12
Sequential assignment

• the technique of making the spin-system
  assignments, followed by sequence-specific
  assignment using unique fragments of sequence, is
  known as sequential assignment (Wuthrich)
• there are alternatives to this protocol one is
  known as main-chain directed assignment
  (Englander). This technique does not focus on
  assigning all the spin systems first. Rather, it
  focuses on the backbone and links sizable
  stretches of backbone residues via sequential
  (i,i1) nOes and other nOes that are
  characteristic of secondary structures. This
  technique is particularly useful when there is
  some knowledge of secondary structure beforehand.

13
Close interatomic distances in secondary
structures
parallel beta-sheet
antiparallel beta-sheet
alpha-helix
type I turn
type II turn
14
Close interatomic distances in 2ndary structures
15
youll often see nOes associated with secondary
structure charted in this way
residue

• a thick bar means a strong nOe (short distance),
  a thin bar means a weak nOe (long but still
  visible distance)
• these sorts of charts allow one to make
  secondary structure assignments more or less
  concurrently with sequential assignments. As we
  will see, coupling constants and chemical shifts
  also aid in secondary structure assignment

16
...you can see that it would be easiest to link
adjacent residues in helices with sequential
amide-amide nOes, whereas in beta sheets
(strand) sequential alpha-amide nOes are stronger
d2.8 Å
d2.2 Å
17
Summary of main-chain directed approach
1. assign a few unique spin systems and use as
entries onto the backbone
Arg Tyr Ser Ala Ala Asn Trp
2. walk down the backbone using sequential and
other backbone nOes
3. fill in missing spin system assignments
backbone refers to alpha and amide protons
18
Summary of sequential approach
1. assign most or all spin systems
Arg Tyr Ser Ala Ala Asn Trp
2. connect adjacent spin systems using backbone
nOes to identify unique dipeptides
3. assemble larger sections of sequence-specific
assignments from dipeptide fragments, until the
whole protein has been assigned
backbone refers to alpha and amide protons
19
Assignment methods that use heteronuclear shift
correlation

• for larger proteins (gt10-15 kD), assignment
  methods based on 2D homonuclear 1H-1H correlation
  methods (COSY/TOCSY/NOESY) dont work very well
  because of overlapping resonances and broad
  linewidths.
• an alternative (which is now used even for small
  proteins in most cases) is to use heteronuclear
  shift correlation experiments on 13C, 15N
  labelled samples.
• in these experiments, magnetization is
  transferred from 1H to 13C and/or 15N through
  large one-bond scalar couplings.
• Some relevant scalar coupling constants

20
15N-1H HSQC based techniques

• as we have seen, one of the simplest types of
  heteronuclear shift correlation is the HSQC
  experiment, which correlates 1H chemical shift to
  the chemical shift of a 15N or 13C connected by a
  single bond
• heteronuclear shift correlation can be combined
  with homonuclear experiments such as 1H-1H NOESY
  or TOCSY to yield 3-dimensional spectra

21
3D HSQC-NOESY and HSQC-TOCSY
these planes can be thought of as a 15N-1H HSQC
these planes can be thought of as a 1H-1H NOESY
the 15N shift dimension can resolve peaks that
would overlap in a 2D NOESY
view of a 3D NOESY experiment
22
Triple-resonance experiments

• there is a whole raft of experiments that use
  both 13C and 15N correlations to 1H nuclei
• the beauty of these experiments is that they can
  connect adjacent residues without requiring any
  nOe information--its all through-bond scalar
  coupling interactions. Makes sequence-specific
  assignment more reliable.
• they also use mostly one-bond couplings, which
  arent very sensitive to the protein conformation
  (unlike, say, three-bond couplings, which vary
  significantly with conformation, as we will see)
• limiting factors 13C is expensive and these
  expts can be tricky