Transcript Solid state NMR assignment of a whole virus particle

SSNMR studies of intact
bacteriophage virus
Assignment of the coat protein of pf1
filamentous phage
A. Goldbourt1, B.J. Gross1, L. Day2, A.E. McDermott1
1Chemistry
2The
Department, Columbia University, New York
Public Health Recent Institute, Newark, NJ
Filamentous bacteriophage
2000nm
6nm
Pf1 cartoon model
G. Stubbs,
Rep. Prog. Phys. (2001) 64, 1389
Pf1 belongs to the filamentous Bacteriophage (Inovirus)
family of organisms known to attack bacteria.
Members of the family include Pf1, Pf3 and Xf (Class-II) as
well as M13, fd, f1, If1, and IKe (Class-I).
The host bacteria for Pf1 (and Pf3) are Pseudomonas
aeruginosa of different strains (strain K for Pf1).
Most virions consist of a long single stranded circular DNA
encapsulated by multiple copies of protein subunits. The
DNA loop is stretched from one end of the virus to the
other in an unknown conformation.
On the surface of the coat protein, several additional
functional proteins are docked. These proteins are crucial
for the bacterial infection and for the reassembly process
of the virion.
Pf1 is the longest known virion
CTXF filamentous bacteriophage codes for
Cholera Toxin
Waldor and Mekalanos
Science (1996) 272, 1910.
It has been discovered that ctxAB, which codes for Cholera
Toxin, is found in the genome of the filamentous phage CTXF.
Only upon infection of Vibrio Cholera by CTXF does it become
toxic.
– The first gene product, cep, closely resembles the genes for the capsid
proteins of M13 and Pf1.
The alignment of the three phages is shown below
Alignment of Pf1, M13 and cep
M13:--AEGDDPAKAAFNSLQASATEYIGYAWAMVVVIVGATIGIKLFKKFTSKAS
PF1:-GVIDTSAVESAITDGQGDMKAIGGYIVGALVILAVAGLIYSMLRKA----cep: DAGLVTEVTKTLGTSKDTVIALGPLIMGVVGAIVLIVTVIGLIRKAK----
N-terminus
Surface exposed
Hydrophobic
region
Positively
Charged,
DNA contact region?
EM of partially purified CTXF
Uses of filamentous phages
Pf1 readily aligns in a magnetic field and is
regularly used for protein partial alignment in
order to obtain RDC constrains.
M13 and T7 are used as sequencing primers.
Virions like fd, f1 and M13 are used for peptide
phage display.
pf1 virion
GVIDTSAVESAITDGQGDMKAIGGYIVGALVILAVAGLIYSMLRKA
Longest known phage, 2000nm.
Coat protein with 46 amino acids, all-helical
Nucleotide/subunit ratio of 1:1 – smallest
known
Undergoes a phase transition at 10oC
Side view of
1PJF
Showing the
repeating
subunits
Top view of
1PFI,
DNA in the
center
Known models for Pf1
Models for the structure of Pf1 were published by 3 groups:
– X-ray powder diffraction, pdb codes (1-4)IFM and 1QL1 for the
low temperature form (Marvin et al.)
– 2IFN and 1QL2 for the high-temperature form (Marvin).
– 1PFI, a new calculation of existing data incorporates the DNA and
predicts a different symmetry then Marvin (Day et al.)
– 1PJF, SSNMR structure of the coat protein backbone for an
aligned sample, model of the phage was built with the symmetry
from 1QL1 (Opella et al.)
Structure of Pf1
The helix is assumed to have a kink in the center, or to be gently curved
the N-terminus can be a loop (1PJF,1QL1,4IFM) or helical (1PFI, 1IFM) and
probably depends on the solvent (surface exposed)
The locations of the sidechains are unresolved in the x-ray data and are
partially obtained from alternative spectroscopic experiments.
The helical bundle symmetry around the DNA is on debate – 27 units in 5
turns (1QL1/2) or 71 subunits in 13 turns (1PFI) or alternatively, 71 dimers
in 26 turns (Day, unpublished). And perhaps, a non-rational ratio.
SSNMR dipolar waves (opella, 1PJF)
1QL1
gently curved helix
15N-H
experimental
dipolar coupling
1QL1
Advantages of MAS NMR
What is the coat protein’s exact structure?
– Refinement of X-ray diffraction data requires a fitting of the protein
structure + the whole helical bundle.
– Different coat protein models can give different results for the helical
bundle arrangement.
– A more accurate structure of the protein will directly contribute to the
solution of the whole virus structure by reducing the number of fit
parameters for the X-ray powder diffraction data.
Is it an inside out DNA?
– Model 1PFI suggests that the structure is ‘inside out’: The bases
point outside towards the coat protein and the phosphates point
towards the inside.
– Hopefully, 31P/13C NMR experiments …
Site specific information
What is the nature of the phase transition?
– The pf1 virus undergoes a reversible phase transition at 10oC. It is
known that the overall length of the virus increases slightly (~100nm ;
15 turns). A model assumes reorganization of the helix bundle.
– Site-specific information will be obtained with ssNMR experiments.
Site specific information
What is the nature of the phase transition?
– The pf1 virus undergoes a reversible phase transition at 10oC. It is
known that the overall length of the virus increases slightly (~100nm ;
15 turns). A model assumes reorganization of the helix bundle.
– Site-specific information will be obtained with ssNMR experiments.
What kind of contacts exist with the DNA?
– Mainly qualitative data exist: Quenching of Tyr40 fluorescence signal,
Lys45 & Arg44 compensate for phosphate charges etc.
– With site specific information contacts to the DNA can be obtained
Where is pf1 coming from?
The virus was prepared in the laboratory of
Loren day
at the ‘Public Health Recent Institute’ in Newark, NJ
The Pseudomonas aeruginosa host was grown on a
13C-Glucose/15N-ammonium chloride M9 media and pf1
was purified and isolated.
Precipitation was done in the McDermott lab (a protocol
developed by L. Day) using PEG, 5mM MgCl2 and
Ethylene Glycol as a cryo-protectant.
Assignment summary for pf1
N
REMARK:
All assignments
shown are site
specific.
Data for ‘residue
type’ assignments are
not included
Next:
sidechains
G1
V2
I3
D4
T5
S6
A7
V8
Q9
S10
A11
I12
T13
D14
G15
E16
G17
D18
M19
K20
A21
I22
G23
CO
CA
CB
CG
-
-
-
-
-
-
-
-
CD
Site specifically assigned
Multiple conformations
Not-assigned
not in sidechain
unresolved
N
G24
Y25
I26
V27
G28
A29
L30
V31
I32
L33
A34
V35
A36
G37
L38
I39
Y40
S41
M42
L43
R44
K45
A46
CO
CA
CB
CG
-
-
-
-
-
-
Backbone: assigned
42/46 (91%)
well resolved: 38/46 (83%)
Sidechains: assigned: 35/40 (85%)
CD
13C-13C
correlation
Cb
Cb/g
Sidechain of Lys20
Cg-Cb
Cg-Cd
O
a
b
g
Cd-Cg
d
Cb-Cg
Cb-CO
Ca-Cg
• Red color on the
peaks serve as a
guide to the eye.
• Negative peaks
were beyond the
contour level
threshold
Ca-CO
Ca-Cb
Ca
Ca
CO
Cb
13C-13C
correlation
Glu9
O
Cb-Cg
Cb-Cd
a
b
g
d
Cg-Cd
Cg-Cb
Ca-CO
• One bond contacts are underlined
Ca-Cb
13C-13C
correlation
Cg2/1-Cb
Cg1/2-Cb
Val2
O
(Cb-CO)
a
b
g1
g2
Ca-Cb
Ca-CO
STRIP PLOTS FROM 750MHz 3D EXPERIMENTS
RED: NCACX
BLUE: NCOCX
29
31
Strip plots from 3D
33
35
37
39
41
43
45
G17 G17
47
49
51
i-1
D18
i
55
55
57
59
Ca (ppm)
61
G15
Q16
Q16
i
i-1
Similar 15N shifts
121.9
121.8
121.7
Similar 13CO/13CA shifts, sequential 15N
176.9
177.0
Assignment of Alanine residues
The advantage of two-bond contacts
DCP: NCA (750)
RAD, 6ms (750)
A7
A46
Alanine
Ca
A36
Ca
O
A34
A21
A11
Ca
Cb
N
a
b
36
21
V35Ca-A34CO
V8Ca-A7CO
CO
Ca
I12Ca-A11CO
CO
RAD, 6ms (600)
Cb
Ca
Initial structural information
The TALOS derived secondary structure
TALOS
‘Torsion Angle Likelihood Obtained
from Shifts and sequence similarity’
Database of ~80 Proteins with
Cornilescu, Delgalio, Bax,
J. Bio. NMR (1999) 13, 289.
– Known X-ray structure < 2Å resolution
– Known NMR chemical shifts
For every 3-amino acid sequence (e.g. GQG), search
matching chemical shifts in the database, give a score
for (i ) N,HA,CA,CB,CO secondary shift difference and
(ii) amino acid identity.
The best 10 matches are used to derive the dihedral
angle for the center a.a. (Q).
Remark : (i) TALOS predicts 3% in error !!
(ii) The database is derived for proteins in solution
and we look at a well organized helical bundle of proteins!
Weights : HA>CO>CA~CB>N
Example – Ser6
10 best
predictions
Dihedral
angle in pdb
file 1PFI
f
Protein matches
from database
y
Best matching
triplets
Prediction is good of 9/10 results agree
TALOS score
Example – Ser6
TALOS results in a text file
y
Df
-40.000
8.000
7.000 50.130
10
Good
5 T -67.000 -42.000
9.000
8.000 57.810
10
Good
6 S -62.000 -40.000
5.000
6.000 51.160
10
Good
7 A -65.000 -38.000
4.000
8.000 41.410
10
Good
8 V -67.000 -44.000
9.000
7.000 33.250
10
Good
# code
4
f
10 best
predictions
D -63.000
Dihedral9 E
angle in pdb
file 1PFI
-64.000 -36.000
Dy
score matches prediction
10
Good
result
6.000 10.000 32.960
Protein matches
from database
f
y
Best matching
triplets
Prediction is good of 9/10 results agree
TALOS average
score
TALOS results
Comparison to 1PFI model (Day)
0
Y
-20
-40
-60
-80
F
-100
-120
G V I D T S A V E S A I T D G Q G D M K A I G G Y I V G A L V I L A V A G L I Y S M L R K A
Y, NMR/TALOS
f, NMR/TALOS
Y, 1PFI model
F, 1PFI model
TALOS results
Comparison to 1PJF model (Opella)
0
Y
-20
-40
-60
-80
-100
-120
F
G V
I
D T
S
A
V
E
S A
I
T
D G Q G
Y, NMR/TALOS
f, NMR/TALOS
D M K A
I
G G Y
I
V
G A
L
V
I
L
A
V
A
G L
Y, 1PJF model
F, 1PJF model
I
Y
S M L
R K A
TALOS results
190
Comparison to 1QL1 model (Marvin)
150
110
0
Y
-20
-40
-60
-80
-100
F
-120
G V I D T S A V E S A I T D G Q G D MK A I G G Y I V G A L V I L A V A G L I Y S M L R K A
-140
-150
-160
Y, NMR/TALOS
f, NMR/TALOS
Y, 1QL1 model
F, 1QL1 model
Summary
Almost complete assignment of pf1, a filamentous bacteriophage
virus, 35MD molecular weight, has been obtained.
Weak DNA sugar-peaks have been observed.
TALOS results suggest a completely helical structure. No significant
loop/kink region has been predicted.
Current work in progress:
– Structural distance constrains will be obtained with emphasis on intermolecular contacts.
– Site specific information will enable to probe the phase transition.
– DNA-protein contacts will be obtained by low-temperature experiments.
Thanks to
$$$ Columbia University $$$
Loren Day Initiated the project, prepares constant flow
of samples and enlightens us with all we know about
pf1!!
Ben Gross – from him I inherited this project, thanks
and good luck!! (of course he also work hard – 600
3D’s, 400 experiments, assignments!!!)
NYSBC (and Boris Itin!!)
2D NCA/NCO experiments – J. Lorieau
Ann McDermott and the wonderful group at Havemeyer
3rd floor.
DNA signals at the 2D 13C-13C spectra
C5
C3
C1/4
C3
Additional expected regions
For C5 (65ppm) and C2 (41ppm)
C1/4
RAD mixing, 6ms
Obtained with 60Hz line
broadening in both dimensions
Information from XPD
R
Powder diffraction layer
lines encompass the
symmetry of the whole
helical virion and fit to a
simple expression –
I(line l)=tn+um
integer
# of turns
# of subunits
Bessel function Jn(R)
A top view of the helical bundle
• DNA is in the center
and was not included in
the models, except for
1PFI
• Does the helix have a
‘kink’?, what is the
structure of the Nterminus? Where do the
subunits make contacts?
1PJF
What do we have so far?
Almost complete assignment has been achieved (for the hightemperature form) with 2D and 3D experiment on 600 and 750
MHz spectrometers.
DNA signals from the sugars have been detected but they are
weak. Low temperature experiment are underway in order to try
and obtain the DNA-coat protein specific contacts.
31P-31P DQ and 31P/13C REDOR experiments failed to produce
any correlations but 1D spectra suggest strong DNA dynamics.
These also will require low temperatures
The low temperature form was precipitated and spectra will be
used to probe the site-specific changes of the phase transition