Nitrilases

1
Nitrilases
Self-terminating, homo-oliogomeric spirals with
industrial applications
Trevor Sewell University of Cape Town
with lots of help from Mark Berman (Cape
Town) Paul Chang (Cape Town) Dakshina M.
Jandhyala and Michael Benedik (Houston) Paul
Meyers (Cape Town) Ed Egelman (Virginia) Dennis
Burford (Cape Town) Helen Saibil (London)
and the EMU at UCT Mohamed Jaffer Brendon
Price Miranda Waldron James Duncan William
Williams
The Wellcome Trust
2
Establishing the principles underlying the
oligomeric structure of the nitrilases.
3
Insights into the structures of nitrilases and
GroEL from 3D electron microscopy
Trevor Sewell
with lots of help from Mark Berman (Cape
Town) Dakshina M. Jandhyala and Michael Benedik
(Houston) Paul Meyers (Cape Town) Ed Egelman
(Virginia) Dennis Burford (Cape Town) Helen
Saibil (London)
and the EMU at UCT Mohamed Jaffer Brendon
Price Miranda Waldron James Duncan William
Williams
The Wellcome Trust
4
Why nitrilases are interesting

• Cleave non-peptide C-N bonds
• Used in industrial processes e.g. manufacture of
  acrylic acid - efficient and environmentally
  friendly
• Detoxification of cyanide waste - bioremediation
• Role in plants - in synthesis of auxin - is one
  of few biological roles properly documented
• Variety of different reported sizes of apparently
  homogeneous material
• Apparent link between quaternary structure and
  activity in some enzymes

5
What we know

• Cysteine, lysine and glutamic acid at active site
• pH optimum 7.6 - 8.0
• Molecular weight of subunit 37kD
• Close relatives all have large molecular weights
  - reported number of subnits varies in different
  species from monomers and dimers, to tens and
  occasionally hundreds.
• Sequences of over 400 members of the nitrilase
  superfamily
• Atomic structure of two (now four) distant
  members of the superfamily.
• The B. pumilus enzyme complex measures 10nm x
  10nm x 20nm

6
The Structure of Nitrilases
Self-terminating, homo-oliogomeric spirals with
industrial applications
Trevor Sewell, Biotechnology Department UWC and
EMU, University of Cape Town
Ndoriah Thuku (UWC) Margot Scheffer(UCT) Mark
Berman (UCT) Paul Chang (UCT) Dakshina M.
Jandhyala(Houston) Xing Zhang (Houston) Michael
Benedik (Tamu) Paul Meyers (Cape Town) Ed Egelman
(Virginia) Arvind Varsani(Cape Town) Helen Saibil
(London)
and the EMU at UCT Mohamed Jaffer Brandon
Weber Brendon Price Miranda Waldron James
Duncan Sean Karriem
The Wellcome Trust Carnegie Corporation
7
Useful Industrial Enzymes
Nicotinic Acid Mandelic Acid Ibuprophen Detoxifica
tion of cyanide
8
Reactions catalysed
Nitrilase - cyanide dihydratase - B. pumilus,
P.stutzeri
Cyanide hydratase - G. sorghi
9
Nit active site
10
Putative catalytic mechanism
11
(No Transcript)
12
(No Transcript)
13
Topology diagram of the a-b-b-a-a-b-b-a dimer
structure found in both DCase and Nit. Nit
labelling. Pace et al (2000)
To Fhit domain
To Fhit domain
14
Location of the active site
15
(No Transcript)
16
Two questions concerning the quaternary
structure

• Homologous nitrilases have subunit molecular
  weights around 40 kDa but are generally reported
  to occur in complexes with 2 - 18 subunits. Why
  is this?
• Nitrilases from several Rhodococcus species are
  inactive as dimers but form active decamers or
  dodecamers on incubation with substrate. Why is
  this?

17
What we did

• Reconstructed a 3D map from negatively stained
  images to a resolution of 2.5nm using SPIDER
• Located homologues in the PDB and aligned them to
  our sequences with GENthreader.
• Developed a dimer model for our enzymes based on
  the non-spiral forming homologues.
• Located the dimer model within the density with
  CoLoRes in SITUS and O.

18
The Process

• Negative stain (uranyl acetate on carbon film)
• Image using low dose
• Digitize film
• Select images
• Classify images
• Starting model using a common-lines based method
• Match images to projections of model
• Reconstruct new model
• Check resolution of structure

iterate
19
Negatively stained native B. pumilus nitrilase,
pH8
20
Multi-reference alignment
Iterative 3D reconstruction
21
Averages of the 84 image sets used in the
reconstriction of the cyanide dihydratase from P.
stutzeri AK61
22
The refinement of the structure of the nitrilase
from Pseudomonas stutzeri (7008 images)
video made by Paul Chang
23
The refinement of the structure of the nitrilase
from Bacillus pumilus (11661 images)
video made by Paul Chang
24
B. pumilus nitrilase (pH 6)
bulge
ridge
P. stutzeri nitrilase (pH 8)
25
Evidence for the global dyad Reconstruction with
no imposed symmetry
26
Cylindrical projection of P. stutzeri nitrilase
32
1.6 nm vertical displacement between local two
fold axes
z (nm)
0
-180
180
0
f ()
70.5
70.5
76.5
76.5
96.5
96.5
Angular offset between local two-fold axes ()

27
The cylindrical projection shows that successive
local two fold axes are separated by increasing
angular rotations but a constant shift along the
helix axis. The projections of the subunits also
appear increasingly elongated along v, because
they are closer to the helix axis.
28
We know the sequences of the B. pumilus enzyme,
thanks to Michael Benedik and Dakshina Jandhyala
at the University of Houston, and the P.
stutzeri enzyme due to Atsushi Watanabe et
al, (1998) BBA, 1382, 1-4. They have 70
sequence homology. A search for structurally
homologous enzymes in the Protein Data Bank using
GenTHREADER produced two enzymes Nit and
DCase. These have less than 20 sequence
homology to our enzymes.
29
Two family members are tetramers
Nit
DCase
30
In the tetramer there are two interacting
surfaces almost at right angles to one another
Surface A alpha helix
Surface B beta sheet
Nit
DCase
31
Topology diagram of the a-b-b-a-a-b-b-a dimer
structure found in both DCase and Nit. Nit
labelling. Pace et al (2000)
To Fhit domain
To Fhit domain
32
Superposition of the alpha carbons of DCase and
Nit
DCase
Nit
cys 169, lys 127, glu 54 catalytic triad
33
An alignment of the nitrilase sequences with Nit
and DCase by GenTHREADER
34
From the sequence comparisons we conclude that

• The insertions and deletions in our enzymes
  relative to NIT and DCase are in outer loops and
  will not impinge on the tertiary structure that
  is crucial to the fold.
• A major difference between our enzymes and the
  tetramers is the existence two significant
  insertions and the C-terminal extension.

35
Need to fit model into density
The two fold axes must coincide
36
Dimer with A surface associating modeled on
residues 10-291 of Nit
Surface A
Surface B
C - terminal
C - terminal
Surface B
37
Dimer with B surface associating modeled on
residues 10-280 of Nit
Surface A
C - terminal
C - terminal
Surface B
Surface A
38
4 ways to align global dyad to dimer axis
A surface mating
B surface mating
This was repeated for the other handedness
39
What is wrong with the B surface models?
Steric clash between NH5 and NS13 and NH3 in the
neighbouring dimer
Poor fits
Unexplainable gaps in density
40
The final, left-handed, 14-subunit model
41
Termination of the helix

• The C surface is flexible and operates as a hinge
  between the subunits.
• As subunits are added at terminus of the spiral
  new opportunities arise for interactions across
  the groove.
• The addition of a further subunit will occur if
  the energetic considerations favour this in
  preference to interactions across the groove
  which result in steric hindrance which would
  prevent the addition of a further subunit.

42
Contacts a and b result in the terminal dimer
having an inwards tilt of 12 degrees thus
preventing the addition of a further dimer. .
I
B
a
43
Contacts c and d are between helices NH2. The
contact area has a local pseudo-dyad axis.
M
d
d
K
glu 82
D
c
c
B
lys 86
44
N
L
b
J
d
H
c
M
F
a
K
D
b
I
B
d
G
c
a
E
C
A
(a)
Cylindrical projection
32
z (nm)
0
-71
71
147
-147
244
-320
320
0
-244
f ()
(b)

45
Superposition of the P. stutzeri nitrilase dimer
model onto the A surface Nit dimer
Insertions thought to be responsible for the C
surface interactions
Deletion causes steric hindrance and would
prevent C surface interactions
46
A prominent ridge on the outer surface was not
filled by the initial model. A four stranded
segment of sheet from bovine superoxide dismutase
fills the density has the correct number of
residues and mates with the ends in left handed
models only.
47
10x(?)
8x(?)
Crosslinking with glutaraldehyde
6x(?)
4x
the protein from the column was diluted 32 fold
and crosslinked with the glutaraldehyde concentrat
ion indicated for 1.25 hrs.
3x
2x
nitrilase monomer

Incompletely unfolded conformational isomers?
0
.002
.005
.01
.02
.05
.1
.2
48
The flexible C surface
49
The location of the active site and B surface
50
Does the quaternary structure have functional
significance?
Nagasawa et al (2000) have found that isolated
dimers of the related nitrilase from Rhodococcus
rhodochrous J1 are inactive. However in the
presence of certain substrates they assemble to
form an active decamer. ( A decamer is required
to produce one turn of the spiral.) We do not yet
know whether this occurs in our case as we don't
yet know how not to produce the spiral in our
enzymes.
51
The enzyme from B. pumilus forms long fibres at
pH 5.4
52
Unidirectional shadowing shows that the long
helices are left handed.
53
The handedness of the spiral

• Defined length oligomers from B. pumilus form
  long helices at pH 5.4. These are shown by
  shadowing to be left handed.
• Our dimer model fits better into left handed
  spirals than right handed spirals as shown by
  SITUS correlation co-efficients.
• Only in left handed spirals is there empty space
  in the map to accommodate the insertions relative
  to non spiral-forming homologues.

54

• What came out of the study?
• A new, defined size, short, spiral,
  homo-oligomeric quaternary structure
• The handedness of the spiral
• The conserved interface (A surface)
• The residues involved in a previously
  undiscovered interface (C surface)
• A model of this interface which would explain its
  flexibility
• A reason for the termination of the spiral
• A reason for the variety of subunit sizes in the
  enzymes

55
Structural transitions in B. pumilus nitrilase
pH 6
pH 8

• The transitions between pH 6 and pH 5.4 may
  involve the titration of a histidine.
• The drop in pH from 8 to 6 results in reduced
  occupancy of the terminal subunits.

pH 5.4
56
Regular helix having 9.4 residues per turn ( for
dimer model Dv76.7 , Dz1.58 nm )
57
B. pumilus
P. stutzeri
G. sorghi
Potential for two salt bridges in
pumilus Repulsion in stutzeri - no long
fibres One salt bridge in sorghi - always fibres
58
Activity increases when structural transition
occurs. Could this mean that 2 extra sites per
18mer become active?
59
(No Transcript)
60
The Effect of Surface Mutations on Activity
Mutant
Surface
Change and location
Activity
B pumilus



1. Delta 303
A
Vgtg-gtstop
Full activity
2. Delta 293
A
Matg-gtstop
Partial activity
3. Delta 279
A
Ytat-gtstop
Inactive
4. Y201D/A204D
A
Ytat-gtDgac, Agcg-gtDgac
Inactive
5. Delta 219-233
C
MKEMICLTQEQRDYF was deleted. 235 Egaa-gtNaac
Inactive
6. 90
D
EAAKRNE-gtAAARKNK
Full activity




P stutzeri



7. Delta 310
A
Sagt-gtstop
Inactive
8. Delta 302
A
Vgtg-gtstop
Inactive
9. Delta 296
A
Qcag-gtstop
Inactive
10. Delta 285
A
Ytat-gtstop
Inactive
11. Delta 276
A
Kaaa-gtstop
Inactive
12. Y200D/C203D
A
Ytac-gtDgac, Ctgc-gtDgac
Inactive
13. Delta 220-234
C
MKDMLCETQEERDYF deleted.
Inactive




Hybrids



14. Pum Stu
A
Residues 1-286 from B. pumilus, 287-end from P.
stutzeri
Full Activity
15. Stu Pum
A
Residues 1-286 from P. stutzeri, 287-end from B.
pumilus
Inactive
61
The only histidines in pumilus that are not in
stutzeri.
The ATCC pumilus has no histidines in the tail -
its properties are being studied
62
Rhodococcus rhodochrous J1
20nm
Negatively stained fibres of J1 nitrilase
(0.45mg/ml) buffered in 20mM KH2PO4, 50mM NaCl at
pH 7.8. Magnification 50000x
63
(No Transcript)
64
G. sorghi CHT reconstructions
WT1 (film)
WT2 (CCD)
Mutant R87Q(CCD)
65
Gloeocercospora sorghi cyanide hydratase
Surprise! Quaternary helix is right handed
66
What's empty?
C terminal extension
C surface linker as before
67
What interactions stabilize the spiral?
B. pumilus
P. stutzeri
G. sorghi
68
active
mutant
charge
no
E82V
-
no
R91Q

yes
D92Q
-
no
Y217D
a-surface
no
Y217E
a-surface
R91Q
E82V
69
(No Transcript)
70
We think we know where all the bits of the
molecule are at coarse resolution.
We think we know what stabilizes the spiral and
causes its termination.
We think that the spiral is essential for
activity.
71
Biotechnological uses?
Can the knowlege we have gained be used to
enhance Stability Activity Ease of
Purification Ease of Immobilization ????
72
B. pumilus has a complex internal structure which
changes during its life cycle. It is therefore
relevant to ask where the nitrilase is located in
the hope that it may give a clue to its function.
73
(No Transcript)