doi:10.1016/j.jmb.2005.02.056 J. Mol. Biol.

(2005) 348, 399408

Structural Basis for APPTPPPLPP Peptide Recognition

by the FBP11WW1 Domain
Jose Ricardo Pires1,2*, Christoph Parthier3, Rodolpho do Aido-Machado1
Urs Wiedemann2, Livia Otte4, Gerald Bohm5, Rainer Rudolph3 and
Hartmut Oschkinat2*
1
Departamento de Bioqumica WW domains are small proteinprotein interaction modules that recognize
Medica, Instituto de Ciencias proline-rich stretches in proteins. The class II tandem WW domains of the
Biomedicas, Universidade formin binding protein 11 (FBP11) recognize specifically proteins contain-
Federal do Rio de Janeiro ing PPLPp motifs as present in the formins that are involved in limb and
Av. Brigadeiro Trompowiski s/n kidney development, and in the methyl-CpG-binding protein 2 (MeCP2),
CCS, Rio de Janeiro, RJ associated with the Rett syndrome. The interaction involves the specific
21941-590, Brazil recognition of a leucine side-chain. Here, we report on the novel structure
2 of the complex formed by the FPB11WW1 domain and the formin fragment
Forschungsinstitut fur
APPTPPPLPP revealing the specificity determinants of class II WW
Molekulare Pharmakologie
domains.
Robert-Rossle-Str. 10, 13125
q 2005 Elsevier Ltd. All rights reserved.
Berlin, Germany
3
Institut fur Biotechnologie
Martin-Luther-Universitat
Halle (Saale), Kurt-Mothes-Str.
3, 06120 Halle (Saale), Germany
4
Institut fur Medizinische
Immunologie, Charite
Universitatsmedizin Berlin
Hessische Str. 3-4, 10115 Berlin
Germany
5
ACGT Progenomics AG
Weinbergweg 22, 06120 Halle
(Saale), Germany
Keywords: WW domain; NMR structure; FBP11; proline-rich peptide;
*Corresponding authors proteinprotein interaction

Introduction these structural investigations and ligand binding

WW domains are proteinprotein interaction
modules forming three-stranded antiparallel
b-sheet structures.1,2 They recognize peptides con-
taining proline-rich motifs (Figure 1).35 A number
of structures revealed the mechanism of proline
recognition and specificity achievement.68 From
J.R.P. and C.P. contributed equally to this work.
Abbreviations used: FBP11, formin binding protein 11;
rmsd, root-mean-square deviation; PDB, Protein Data
Bank; MeCP2, methyl-CpG-binding protein 2; NOE,
nuclear Overhauser enhancement.
E-mail addresses of the corresponding authors:
jrmpires@gmx.net; oschkinat@fmp-berlin.de
studies employing peptide libraries it became
apparent that peptides as short as five residues
may bind with an affinity of 60 mM to WW
domains,9 utilizing hydrophobic interactions and
few hydrogen bonds. As initially observed for SH3
domains, the proline-rich ligands can bind in
different orientations to WW domains.7,8
The WW domain containing formin binding
proteins (FBPs) were originally identified by screen-
ing of mouse limb bud expression libraries for
binders of a conserved proline-rich region present
in formin isoforms. One of the candidates was
formin binding protein 11 (FBP11), which is, as most
FBPs, localized in the cell nucleus and assumed to
be connected to pre-mRNA splicing. It possesses a

0022-2836/$ - see front matter q 2005 Elsevier Ltd. All rights reserved.
400
Figure 1. The WW domain sequences and classes.
Amino acid sequence of selected WW domains grouped
in classes according to ligand specificity.4 Classes/ligand
motifs: class II PPLPp (magenta), class III
(p/f)P(p/g)PPpR or (p/f)PPRgpPp (orange), class IV
(poS/poT)P (green) and class I PPxY (blue). Conserved
amino acid residues are in bold; amino acid residues on
the b-strand face where binding occurs are grey-shaded;
amino acid residues responsible for class specificity are
colored. The asterisk (*) indicates sequences assumed to
be FBP11 orthologs.
modular architecture comprising two WW domains
followed by several FF domains.10,11 FBP11 WW
domains are capable of binding the sequence
APPTPPPLPP present in most formin isoforms,
and further functional screening of cDNA
expression libraries have identified eight putative
ligands for FBP11 WW domains (WW domain
binding proteins, WBP-3 to WBP-10). These pro-
teins contain proline-rich sequences including
leucine residues. Among these proteins, WBP-5
and WBP-8 comprise the longest poly-proline
sequence, PPPPPPPLPP. Analysis of the proline-
rich region within the identified WBPs resulted in
the consensus sequence PPLP, defined as the
binding motif of class II WW domains.12
The network of possible physiologically relevant
interactions involving FBP11 is still emerging and in
the past years the proteins Huntingtin,13,14 neural
WiskottAldrich syndrome protein (N-WASP),15
and MeCP216 (WBP-10 homologous protein) were
shown to be able to interact with FBP11. Huntingtin
and MeCP2 are proteins related to the human
genetic disorders Huntingtons disease and Rett
syndrome, respectively. N-WASP plays a role in
actin cytoskeleton reorganization. It was demon-
strated that the FBP11 WW domains and proline-
rich regions in these proteins are essential for their
interactions. Formins and the protein MeCP2
contain the WW domain class II binding motif
PPLPp, whereas in N-WASP and Huntingtin the
motif PPLPp is absent or replaced by long poly-P
stretches. Substitution analysis of the peptide
APPTPPPLPP has shown that the mutation L to P
PPLPp Peptide Recognition by Class II WW Domains
reduces the strength of the interaction significantly.
In this case, specificity towards peptide ligands
with the motif PPLPp is governed by hydrophobic
interactions involving a residue other than proline.4
This result shows that specificity depends on the
shape of a hydrophobic amino acid, a rare situation
in the field of proteinpeptide interactions.
Of special interest in this context is the complex
of the FBP11WW1 domain and the peptide
APPTPPPLPP. The new complex structure reported
here is the first involving a WW domain of the class
II subfamily and expands our understanding about
the mechanism of recognition of proline-rich motifs
adopted by WW domains.
Results and Discussion
Binding constants and epitope mapping
To substantiate the contribution of WW domain
interactions to the formation of complexes invol-
ving FBP11 we determined dissociation constants
for the complex formed by FBP11WW1 domain and
the peptides GPPPPPPPLPP, GPPPPPPPLP (WBP-
5/WBP-8), APPTPPPLPP (formin) and the L to P
mutants GPPPPPPPPPP and APPTPPPPPP (Table 1
and Figure 2(a)). The micromolar constants
obtained are in agreement with the constants
recently determined by other authors using the
surface plasmon resonance technique for the
FBP11WW1 domain and proline-rich peptides.17
The weak interactions determined here are typical
for proteinprotein interaction domains recogniz-
ing proline-rich motifs and could be related to
the formation of transient complexes or may be
supported by other interactions to form tight
complexes.
The KD values presented in Table 1 indicate two
effects. First, the L to P mutation increases the
dissociation constants by around 100 mM. This effect
was previously described for the class II WW
domain FE65, which binds poly-proline peptides
interrupted by a single leucine residue with higher
affinity compared to the same length poly-proline
peptides.18 Second, longer uninterrupted proline
stretches contribute to the stabilization of the
complex, as seen by the difference in the dis-
sociation constants of FBP11WW1 complexes with
ligands derived from formin and WBP5/WBP8,
respectively.
Additionally, the dependence of the binding
constants on pH was investigated. For the peptide
Table 1. Dissociation constants of complexes formed by
the FBP11WW1 domain and proline-rich peptides
Peptide KD (mM)
GPPPPPPPLPP 145G4
GPPPPPPPPPP 242G23
GPPPPPPPLP 226G7
APPTPPPLPP 371G22
APPTPPPPPP 473G54
PPLPp Peptide Recognition by Class II WW Domains
Figure 2. Ligand binding and epitope length analysis.
(a) Comparison of peptide ligands GPPPPPPPLP (;),
GPPPPPPPLPP (C) and bGPPPPPPPPPP (B) binding to
3.5 mM FBP11WW1 domain in 50 mM TrisHCl buffer
(pH 7.5), 150 mM NaCl, 5 mM EDTA at 25 8C, determined
by fluorescence spectroscopy. (b) pH dependence of
the dissociation constant (Kd ) for the complex
FBP11WW1:GPPPPPPPLP in the range from pH 5.5 to
pH 8.5 at 25 8C, calculated pKaZ6.7. (c) Interaction of the
peroxidase-labeled FBP11WW1 domain with a library of
cellulose-bound truncation analogs of the ligand
APPTPPPLPP. Higher signal intensities (darker spots)
indicate stronger binding of the WW domain to the
respective peptide.
401
GPPPPPPPLP the KD values increased twofold
when the pH was shifted from 8.5 to 5.5 (Figure
2(b)). Fitting the sigmoidal shaped data observed
(Figure 2(b)), allowed the calculation of an apparent
pKa value of 6.7 that was attributed to the ionizable
group His20 (the pKa value of the imidazole group
of histidine found in proteins may vary in the range
between 6 and 7). This is the first evidence of the
involvement of the location of His20 in the
composition of the binding site. The additional
positive charge of the protonated histidine residue
in the otherwise largely hydrophobic binding sur-
face could interfere negatively with the interaction
between ligand and the WW domain, which leads
to the somewhat lower affinity at pH !7.5.
Interestingly, His20 is a partially conserved residue
among WW domains of class II. It is only replaced
by other aromatic or aliphatic residues (Figure 1).
This highlights the importance of the hydrophobic
character of His20 for binding.
Length analysis of the peptide APPTPPPLPP by
application of a peptide library obtained by SPOT
synthesis yielded the five amino acid residues long
sequence PPLPP as the minimal epitope for achiev-
ing higher affinity (Figure 2(c)). The ligand epitope
determined has the same length as compared to
the one determined for the YESkinase-associated
protein (YAP65) WW domain (PPPPY).9 In the case
of FBP11, the C-terminal residue of the binding
epitope is essential and preferably proline,
although it is also exchangeable by other amino
acid residues.4 In the YAP65 case the same is true
for the N-terminal proline residue.9
Structure of the FBP11WW1 domain and of its
complex with peptide APPTPPPLPP
The solution structure of the FBP11WW1 domain
is shown in Figure 3(a) and the 0
structure of0 its
complex with the peptide A1 PPTPPPLPP10 in
Figure 3(b)(d). The free domain structure was
derived based on 851 experimental restraints and is
well defined from residues 1542. For detailed
structure statistics see Table 2. FBP11WW1 shows
the well-known WW fold. A hydrophobic cluster on
one side of the b-sheet is formed by the highly
conserved residues W17, Y29 and P42. On the other
side W39, Y28 and S37 are conserved in most WW
domains, and the variable residues T18, H20, S22,
P23, D24 and Y30, form the binding surface for the
proline-rich peptide.
Loop 1 is composed of residues P23, D24 and
G25, and its structure is less conserved. It shows a
backbone root-mean-square deviation (rmsd) value
of 3.8 A and 4.4 A to the loops in the WW
prototype19 and YAP65,9 respectively. In our struc-
ture, the position of loop 1 is defined by several
nuclear Overhauser enhancements (NOEs)
observed between W39 and loop residues P23 and
D24. Equivalent NOEs were not observed for the
class I WW domain of YAP65.9
By addition of the ligand APPTPPPLPP, pertur-
bation of W39 chemical shifts is observed and many
402 PPLPp Peptide Recognition by Class II WW Domains

Figure 3. Structure of the FBP11WW1 domain and of its complex with peptide APPTPPPLPP. (a) Stereo view of an
ensemble of the 15 best structures selected for the FBP11WW1 domain. b-Strand regions are colored blue and loop
regions are orange. Selected side-chains are displayed and labeled. (b) Stereo view of an ensemble of the 15 best
structures selected for the FBP11WW1:APPTPPPLPP complex. The domain was colored as for (a) and the ligand was
colored green. (c) Top view of a representative structure showing the residues involved in binding. (d) Ligand bound to
the surface of the FBP11 WW1 domain. The surface was colored by hydrophobicity. Blue represents hydrophilic, green
represents intermediate and brown represents hydrophobic regions. H-bonds are displayed in yellow.

of the NOEs observed in the free state between W39
and residues P23 and D24 disappear. These results
suggest that complex formation involves a smooth
accommodation of the domain surface during
ligand binding. This hypothesis was confirmed by
the determination of the structure of the complex,
where loop 1 no longer deviates significantly from
other WW domain structures.
The calculation of the complex structure was
supported by 40 intermolecular NOEs, 15 of them
were manually assigned and a further 25 NOEs
automatically using ARIA.20,21 The ligand was
restrained to a poly-proline type II left-hand helical
conformation (PPII), as observed for other com-
plexes involving proline-rich ligands. However,
similar results could be obtained without restrain-
ing the ligand conformation using only the 40
intermolecular NOEs (data not shown). Seven of the
manually assigned NOEs occurred between L8 0 and
Y28, W39, S22 and P23. These restraints define
precisely the position and orientation of L8 0 with
respect to the binding surface. Eight further NOEs
assigned involved proline residues: between P5 0
and H20; between P6 0 and Y30; between P9 0 and
S37/Y28; and between P10 0 and W39. The complete
structure statistics can be found in Table 2.
Starting the inspection of the complex at the
N-terminal residues of the ligand, we see that the
ring of P6 0 is situated perpendicular to the b-sheet,
equidistant from Y30, T18, Y28 and H20 (Figure
3(c)). These four residues form a hydrophobic
groove on the surface of the domain where the
ring of P6 0 is buried (Figure 3(d)). The hydrophobic
nature of this interaction explains the dependence
of the complex dissociation constant on pH: at low
pH, His20 will be protonated, increasing the polar
character of the recognition site, thus destabilizing
the complex. The P6 0 carbonyl group is also
PPLPp Peptide Recognition by Class II WW Domains 403

Table 2. Structure statistics

WW alone Complex
A. Number of restraints
Intra-residual 344 273
Sequential 147 130
Medium-range 69 48
Long-range 210 150
Intermolecular 40
Hydrogen bonds (two restraints each) 11 14
Dihedral angles (f,j) 28 40
Ambiguous 51 33
Total 851 742
B. rmsd from experimental restraints
NOEs (A ) 0.025G0.001 0.016G0.002
Dihedral angles (deg.) 1.9G0.1 0.66G0.09
C. CNS potential energy (kcal molK1)
Etotal 340G6 328G6
Ebonds 6.1G0.5 3.3G0.2
Eangles 38G3 28G2
Eimpropers 6.2G0.9 5G1
Edihedral 209G3 236G4
EvdW 48G2 45G3
Enoe 27G2 9G2
Ecdih 6.0G0.8 1.9G0.5
D. Ramachandran plot analysis (%)
Residues in most favored regions 78.6 85.1
Residues in additionally allowed regions 18.6 8.0
Residues in generously allowed regions 2.8 3.2
Residues in disallowed regions 0.0 3.7
E. rmsd (A ) between average structure and selected set of structures
WW domain residues 1542
Backbone 0.24G0.06 0.39G0.13
All non-H 0.69G0.07 0.89G0.08
WW domain residues 1542Cligand residues 5 0 10 0
Backbone 0.37G0.11
All non-H 0.82G0.07

involved in H-bonding to the Y28 hydroxyl group.
The ring of P7 0 is positioned parallel with the b-
sheet, contacting the surface of the domain. The P7 0
carbonyl group forms a hydrogen bond involving
the S37 hydroxyl group.
The second recognition site in FBP11WW1 is
formed by Y28 and W39. It is conserved in all WW
classes. As determined from the intermolecular
NOEs, Leu8 0 of the class II binding motif PPLPp is
placed at this binding site. The methyl groups of L8 0
contact S22 and P23 located in loop1 and Y28. One
methyl group is directed to the surface of the
domain, the second is more exposed (Figure 3(c)
and (d)). Consequently, the degenerate chemical
shifts observed for both Leu methyl groups in the
free ligand at dZ0.89 ppm split into two signals at
dZ0.86 ppm and 0.92 ppm upon binding. P9 0 is the
second ligand residue located in the binding site
formed by Y28 and W39. This residue repeats the
orientation of P6 0 starting a second turn of the PPII
helix. As in the case of P6 0 , the ring of P9 0 is
perpendicular to the b-sheet and parallel with the
W39 indole ring. The carbonyl groups of P9 0 and
P10 0 may form a bi-furcated H-bond involving the
W39 indolic NH.
The structure of the complex FBP11WW1:
APPTPPPLPP is in agreement with our binding
constants measurements, epitope mapping and
previous substitution analysis data.4 The residues
that form the minimal ligand epitope are P6 0 , P7 0 ,
L8 0 , P9 0 and P10 0 . From our structure we can see that
all these five residues are involved in interactions
with the WW domain, whereby P10 0 shows only
contacts involving backbone atoms. The ring of P10 0
is directed towards the solution. P10 0 contributes
mainly as an H-bond acceptor via its carbonyl group
and is thus found to be essential for higher affinity
in the length analysis, and in our binding constant
measurements, but unspecific in the substitution
analysis. P5 0 seems to be already outside the
epitope. In the length analysis it does not seem to
improve the affinity significantly, and the substi-
tution analysis does not show specificity at this
position. Accordingly, in the complex structure the
ring of P5 0 is directed towards the solution and does
not seem to contribute directly to the interaction.
Residues P6 0 , P7 0 , L8 0 and P9 0 show high specificity
in the interaction with the WW domain and in all
cases the respective side-chains contact the WW
domain surface. The clearest contributions to
specificity are due to P6 0 and P9 0 . These residues
are in equivalent positions of consecutive turns of
the PPII helix. Their proline rings make extensive
hydrophobic contacts to Y30 and/or Y28 in the case
of P6 0 and to W39 in the P9 0 case.
The specificity for leucine at position 8 0 instead of
Val, Ile or Ala is less obvious, but a possible factor
contributing to this specificity is the shape of the
404 PPLPp Peptide Recognition by Class II WW Domains

Figure 4. Comparison of WW and SH3 domain complexes. (a) Schematic representation of a WW domain showing
important regions for ligand recognition. Residues conserved for all WW domains are displayed in blue. Positions
changing between the WW domain specificity classes are shown in green, next to which the amino acid types and their
corresponding WW domain classes are indicated in parentheses. Superposition of FBP11WW1 (grey ribbon, blue side-
chains, class II) complexed to APPTPPPLPP (cyan) with: (b) dystrophinWW domain (red, class I) in complex with SPPPY
(yellow); (c) Pin1WW domain (red, class IV): in complex with (poS)PT(poS)PS (yellow); and (d) the C-CRK N-terminal
SH3 domain (red) in complex with PPPALPPK (yellow). (e) PRP40WW1 (red, class II) and (f) PRP40WW2 (red, class II)

binding site dictated by the less conserved residues
S22, P23, D24 and G25, which interact with L8 0 .
A proline side-chain at position 8 0 would not fit as
perfectly as a leucine residue to the surface of the
domain without distorting the PPII helix confor-
mation or without changing other interactions,
explaining the lower affinity shown by the L to P
mutants (Table 1). A rationalization of the impor-
tance of proline in position 7 0 is less straight-
forward. The ring of P7 0 is positioned parallel with
the b-sheet, contacting the surface of the domain. Its
d methylene group is close to the aromatic ring of
Y30 making further hydrophobic contacts. This
could be one contribution to the specificity for
proline at this position, besides the stabilization of
the PPII helix.
Comparison of complexes of different WW
families and SH3 domains
Comparisons of the FBP11WW1:APPTPPPLPP
complex with other WW and SH3 complexes are
shown in Figure 4(a)(f). Several conserved
features can be derived from these comparisons,
and even more interesting are the differences
between the complexes, which are responsible for
specificity. The summary of these comparisons is
shown in Figure 4(a). This Figure shows a ribbon
representation of a typical WW domain fold,
with a top view of the binding site. Conserved
residues involved in binding are displayed with
their side-chains in blue, while the position of
key variable residues, responsible for subfamily
PPLPp Peptide Recognition by Class II WW Domains
specificity are in green, and only b-carbon atoms are
shown.
Figure 4(b) compares the structure of the com-
plex FBP11WW1:APPTPPPLPP with the complex
dystrophinWW:SPPPY.22 FBP11WW1 makes use of
two hydrophobic grooves for recognizing its pro-
line-rich ligand. One groove is common to all WW
domains, it is the one formed by W39 and Y28.
These residues are very conserved in WW domains
of all classes. The second hydrophobic groove is
formed by T18, H20, Y30 and Y28, and is spatially
distinct from the specificity recognition site of
dystrophin, which uses residues I30, H32 and Q35
of the second and third WW domain b-strands for
recognizing a tyrosine in the ligand. This Y
recognition site is conserved in all domains that
recognize the PPxY motif but is absent from the
FBP11WW1 domain. Interestingly, the aromatic
residue at position 30 superimposes well with the
tyrosine of class I PPxY ligands, as seen in Figure
4(b). As a consequence, all WW domains possessing
an aromatic residue at position 30 should be unable
to bind to the PPxY motif, and should use instead a
similar second hydrophobic recognition site as in
FBP11WW1 for binding the ligand. This principle
was suggested before as one of the rules that
characterize differences between class I and class II
WW domains.23 However, the binding studies for
the tandem WW domains of the FBP11-homologous
yeast protein PRP40, which also possess an aro-
matic residue at position 30, indicated that PRP40
WW domains can bind both PPLPp and PpxY-
containing peptides.24 Nevertheless, this study
emphasizes that peptide ligands containing the
class II motif seemed to have a higher affinity than
those containing PPxY. Moreover, PRP40WW2
exhibited a more favorable interaction with PpxY-
containing peptides than PRP40WW1. Figure 4(e)
and (f) show the comparison of the FBP11WW1:
APPTPPPLPP complex with the PRP40WW1 and
WW2 domains, respectively. The difference in the
affinity of the two PRP40 WW domains for the
PpxY-containing peptides seems to correlate with
the orientation of the aromatic residue at position
30. For PRP40WW1, the Y recognition site is more
obstructed by the aromatic residue at position 30
compared with PRP40WW2 (Figure 4(e) and (f)).
The overall poor affinity of the PRP40WW domains
for peptides containing the PPxY motif can also be
related to the lack of H32, previously determined to
be important in the recognition of the Y-containing
ligand.25
Figure 4(c) compares the structure of the complex
FBP11WW1:APPTPPPLPP with the complex Pin1
WW:(pS)PT(pS)PS.26 It reveals two major differ-
ences: P23 in loop 1 and H20 in the first b-strand of
FBP11WW1 are replaced by arginine residues in the
Pin1 WW 0 domain. Correspondingly
0
P5 0 and L8 0
of the A1 PPTPPPLPP10 ligand are replaced by
phosphoserine residues in the Pin1 ligand. This
emphasizes the importance of loop 1 for domain
specificity in both cases, Pin1 and FBP11WW1. The
NC terminal orientation of the ligand peptide in
405
the FBP11WW1 complex is the same as in Pin1 and
opposite that in the dystrophin complex. Due to the
pseudo-symmetry of the proline-rich ligand it is
likely that some proteins could bind to FBP11 in the
opposite orientation, depending on flanking inter-
actions to the proline-rich region, as observed in
SH3 domain complexes, where the orientation of
the ligand is dependent on the presence of an
arginine or lysine C or N-terminal to the proline-
rich motif.
Figure 4(d) compares the structure of the complex
FBP11WW1:APPTPPPLPP with the complex
formed by the N-terminal SH3 domain of the
proto-oncogene product c-Crk and peptide
PPPALPPK27 in which the leucine plays a similar
role, and because FBP11WW domains compete with
SH3 domains for binding to similar ligands. W39
and H20 are the edging aromatics in the binding site
of FBP11WW1. They superimpose nicely to the
edging Trp and Phe of the SH3 domain. About
halfway between these two edging aromatic resi-
dues in both cases there is a third aromatic residue,
a tyrosine that divides the binding surface in two
hydrophobic binding grooves. In these two grooves
two turns of a PPII helix of the ligand are fitted.
However, from the comparison in Figure 4(d), it
becomes obvious that there are many differences
between these two classes of domains. The central
aromatic ring is oriented differently and hence it
serves different purposes in both domains. In SH3 it
is perpendicular to the surface often serving as an
H-bond donor and contact area for two proline
residues that surround it forming a conserved
umbrella-like structure,6 whereas in WW domains
it is oriented along the surface providing a
hydrophobic area for contacting only one proline
ring. Hydrogen bonds to this aromatic residue are
not always observed in WW domains.
Also, there is no equivalent for FBP11WW1 Y30 or
loop 1 residues in the SH3 domain and there is no
equivalent for the R/K recognition site of SH3
domains in the FBP11WW1 domain. Thus, although
in some cases both domains can bind to the same
proline-rich ligand, they have different ligand
predilections.
Concluding remarks
The structure of the complex of FBP11WW1 with
peptide APPTPPPLPP complements our previously
derived structurefunction relationships using
screening of peptide libraries and binding constant
measurements.4 The singular feature of this com-
plex is the specificity invoked by a leucine residue.
The new structure shows that this amino acid binds
instead of proline in the conserved hydrophobic
groove (Figure 4(a), blue). The topology of this
groove is modulated by variable amino acid
residues that compose loop 1 and which are
responsible for the specificity for Leu. The rather
flexible side-chain of leucine as compared to proline
contributes to a perfect fit of the PPII helix, allowing
four consecutive residues (P6 0 P9 0 ) to have their
406
side-chains contacting the binding surface of the
domain. We expect that FBP11WW1 interacts with
the proline-rich regions of proteins formin and
MeCP2 containing the PPLPp motif in the way
shown in our structure. For other proteins like
Huntingtin and N-WASP, which were also shown to
bind to FBP11WW1 but do not possess the PPLPp
motif, we could expect that the predicted lower
affinity could be compensated to some extent by the
very long all-proline stretches and the additive
contributions of neighboring WW domains.
Considering the classification of WW domains
according to ligand predilections, two different
tendencies are presented in the recent literature.4,7,17
Kato et al.17 and Wiesner et al.24 have shown
examples that some WW domains are promiscuous,
displaying binding propensities of classes II and III
or I and II, respectively. Based on these results it is
suggested that borderlines between classes are not
well defined. On the other hand, in the work by Otte
et al.,4 different binding propensities were described
for members of the same WW domain class,
consequently class II was divided into domains
preferring either PPLPp or p/jPPPPP while class
III was divided into domains binding preferably
either (p/f)P(p/g)PPpR or (p/f)PPRgpPp.
The structural comparisons of WW domains
presented here show similarities in the binding
modes of WW domains of different classes explain-
ing the fragility of classification. On the other hand
the present detailed structural investigation contri-
buted to understanding the role of variable residues
in recognition mechanisms and thus to understand
differences in binding constants that can be impor-
tant in tuning the complex net of interactions.
Materials and Methods
WW domain preparation
The FBP11WW1 sequence, residues W17 to D44
(Figure 1), was obtained in the expression plasmid
pGEX-2TK (Amersham Pharmacia Biotech) encoding an
N-terminal fusion protein of glutathione-S-transferase
and FBP11WW1 domain (GST-WW). The N terminus of
the FBP11WW1 domain was later extended by four
residues, A13 to M16 from the FBP11WW1 gene sequence,
by QuikChange site-directed mutagenesis (Stratagene).
GST-WW was expressed at 37 8C in Escherichia coli
strain BL21 (DE3) grown in shake flasks. To obtain 13C
and 15N incorporation into the protein, cells were grown
in mineral salt medium containing [13C]glucose as the
only carbon source and [15N]ammonium chloride as
nitrogen source, respectively. Unlabeled GST-WW was
expressed in cells grown in LB medium. Protein induction
was achieved at A600Z0.7 by addition of IPTG to a final
concentration of 1 mM in the culture medium. Recombi-
nant GST-WW was purified from soluble cell lysate using
glutathione-Sepharose 4B affinity resin (Amersham
Pharmacia Biotech) following the manufacturers recom-
mendations. The WW domain was cleaved from GST
using bovine thrombin (Sigma) yielding a 41 residue
FBP11WW1 domain construct including nine N-termi-
nal residues (G4SRRASVGS12) originating from the
PPLPp Peptide Recognition by Class II WW Domains
expression vector. Final purification of the WW domain
was achieved under denaturing conditions using
Source15RPC reversed phase HPLC column (Amersham
Pharmacia Biotech) following standard procedures invol-
ving a solvent system consisting of acetonitrile/water
containing 0.05% (v/v) trifluoroacetic acid. Fractions
containing the purified WW domain were pooled and
lyophilized. Protein identity and homogeneity were
confirmed by electrospray mass spectrometry.
Proline-rich peptide ligands
Peptide ligands (GPPPPPPPLP, bGPPPPPPPLPP,
bGPPPPPPPPPP, CbbAAPPTPPPLPP, CbbAAPPTP
PPPPP, bZbeta-alanine) were obtained from automated
solid phase peptide synthesis on chlortrityl resin (Nova-
biochemtech) using FMOC strategy. The peptide
N-terminal amino group and C-terminal carboxy group
were retained unmodified. After cleavage of the peptides
from the resin using hexa-fluoro-isopropanol/dichlor-
methane the peptides were purified by reversed phase
HPLC using a Licrospher 100 RP column (Merck)
following standard protocols involving a water/aceto-
nitrile solvent system containing 0.05% trifluoroacetic
acid. Pooled fractions of the pure peptides were
lyophilized and analyzed by electrospray mass
spectrometry to verify identity and homogeneity.
Epitope length analysis
The cellulose membrane-bound peptide library was
prepared according to standard SPOT synthesis proto-
cols28 using a Spot synthesizer (Abimed, Langenfeld,
Germany) as described.29 The library was incubated
with the peroxidase-labeled WW domain following the
procedure described in the literature.4
Determination of KD by fluorescence
Stock solutions of the WW domain and peptide ligands
were prepared by dissolving the lyophilized powder in
10 mM potassium phosphate (pH 6.0), 100 mM NaCl,
0.1 mM EDTA, which allowed the WW domain to refold
spontaneously (data not shown). Concentrations of the
polyproline ligand stock solutions were determined
gravimetrically; concentration of the WW domain was
determined by measuring absorption at 280 nm and using
the calculated absorption coefficient 3280Z15,470 MK1
cmK1. For fluorescence measurements the WW domain
was diluted into 2 ml of the corresponding buffer to a final
concentration of 3.5 mM in a stirring quartz cuvette.
Excitation of tryptophan fluorescence was carried out at
295 nm and change of fluorescence emission intensity at
340 nm upon ligand addition was monitored using a
Fluoromax-2 spectrofluorometer (ISA SPEX; excitation
and emission monochromator slits set to 2.5 nm, tem-
perature-controlled by a cryostat). The emission intensity
(Fobs) was plotted as a function of total ligand concen-
tration ([L0]) added to the sample and least-squares fitted
according to equations (1) and (2) (assuming 1:1 complex
formation) from which the dissociation constant KD can
be calculated:
WL

Fobs Ffree Fsat K Ffree (1)
W0

where Ffree and Fsat are the fluorescence intensity without
ligand and with a saturating concentration of ligand,
respectively. [W0] is the total WW domain concentration
PPLPp Peptide Recognition by Class II WW Domains 407

used in the titration, while [WL] corresponds to the characteristic of secondary structure and other unam-
fraction of WW domain bound to ligand, which can be biguous NOEs that allowed the calculation of a prelimi-
obtained according to the law of mass action: nary 3D structure. Further assignments of NOEs were
obtained automatically using ARIA20,21 version 1.2 and
W0
L0
KD
WL
checked manually.
2 Dihedral angle restraints (f,j) based on chemical shifts
s
 were obtained from CSI32 and inter-strand hydrogen-
W0
L0
KD 2 bond restraints based on NOEs characteristic of secon-
K K W0
L0
(2)
2 dary structure pattern were used in the calculations with
ARIA using CNS33 version 1.1.

pH titration of ligand binding Complex

In addition to the procedure described above for the
Determination of dissociation constants (KD) derived
complex, 15 intermolecular NOEs were identified and
from binding of peptide ligand GPPPPPPPLP to
manually assigned by comparisons of 2D and 3D NOESY
FBP11WW1 domain was performed by fluorescence
spectra with and without ligand, allowing the calculation
over a pH range of 5.5 to 8.5 at 25 8C involving different
of a preliminary 3D structure of the complex. A further 25
buffer systems. Titrations at pH !7 were done using
intermolecular NOEs were assigned automatically by
50 mM sodium phosphate, 150 mM NaCl, 5 mM EDTA,
ARIA. Dihedral angles restraints (f,j) were added
measurements in the range of pH 7 to pH 8.5 using 50 mM
restraining the ligand conformation to a PPII helix and a
TrisHCl, 150 mM NaCl, 5 mM EDTA, respectively.
further three intermolecular H-bonds were added based
Observed dissociation constants (Kobs) were recorded as
on preliminary 3D structures.
a function of pH and fitted to the sigmoidal curve
For the FBP11WW1 and its complex 200 structures each
described by equation (3):
were calculated in the last ARIA iteration. The 30
Kxh C Kx 10pHKpKa structures with lowest energy were selected for the
Kobs Z (3) calculation of an average structure and the 15 structures
1 C 10pHKpKa
with the smallest rmsd values to the average structure
where Kxh represents the KD value for the protonated were selected as representative ensemble.
species, Kx represents KD displayed by the unprotonated
species. The least-squares fit yields the pKa value for the
Atomic coordinates and NMR restraints
group, which changes ionization within the pH range
probed.
The atomic coordinates, NMR restraints, and chemical
shifts assignments for the FBP11WW1 and its complex
NMR spectroscopy were added to the RCSB Protein Data Bank, PDB codes
1YWJ and 1YWI, respectively.
FBP11WW1 alone
Multidimensional NMR experiments30,31 for the assign-
ment of backbone and side-chain 1H, 13C and 15N, chemical
shifts were acquired in a 1.8 mM [U-15N,13C]FBP11WW1 Acknowledgements
domain sample dissolved in 10 mM phosphate buffer (pH
6.0), 100 mM NaCl, 0.1 mM DTT, 0.1 mM EDTA, 90% H2O,
We thank Dr Mark Bedford, Harvard Medical
10% 2H2O. For the derivation of distance restraints a 3D 13C
NOE spectroscopy (NOESY) heteronuclear multiple quan- School, USA, for providing the plasmid containing
tum coherence (HMQC) (100 ms mixing time), a 3D 15N the original FBP11WW1 domain sequence, Dirk
NOESY heteronuclear single quantum coherence (HSQC) Wildemann, Max Planck Institute Halle, Germany,
(150 ms) and a 2D NOESY (200 ms) were obtained. for assistance in solid phase peptide synthesis and
Dr Hauke Lilie for fruitful discussions. C.P. was
supported by a grant from Land Sachsen-Anhalt,
Complex Germany. J.R.P. acknowledges FAPERJ and CNPq
For the complex structure determination unlabeled agencies (Brazil) for funding.
ligand was added in a molar ratio of 2:1 (ligand/ domain)
and the same spectra as described above were acquired.
Ligand chemical shift assignments were obtained by
13
C/12C filtered total correlated spectroscopy (TOCSY)
References
and NOESY. All NMR spectra were acquired in a Bruker
1. Bork, P. & Sudol, M. (1994). The WW domain: a
DRX600 spectrometer at 298 K, processed with
signaling site in dystrophin? Trends Biochem. Sci. 19,
XWINNMR (Bruker AG) and analyzed with Sparky
531533.
(T.D. Goddard & D.G. Keneller, University of California,
2. Macias, M. J., Hyvonen, M., Baraldi, E., Schultz, J.,
San Francisco).
Sudol, M., Saraste, M. & Oschkinat, H. (1996).
Structure of the WW domain of a kinase-associated
Structure calculations protein complexed with a proline-rich peptide.
Nature, 382, 646649.
FBP11WW1 alone 3. Sudol, M. & Hunter, T. (2000). New wrinkles for an
old domain. Cell, 103, 10011004.
The 3D NOESY spectra were analyzed manually for 4. Otte, L., Wiedemann, U., Schlegel, B., Pires, J. R.,
unambiguous HNHN, Ha Ha and HNH a NOEs Beyermann, M., Schmieder, P. et al. (2003). WW
408
domain sequence activity relationships identified
using ligand recognition propensities of 42 WW
domains. Protein Sci. 12, 491500.
5. Hu, H., Columbus, J., Zhang, Y., Wu, D., Lian, L.,
Yang, S. et al. (2004). A map of WW domain family
interactions. Proteomics, 4, 643655.
6. Ball, L. J., Kuhne, R., Schneider-Mergener, J. &
Oschkinat, H. (2005) Recognition of proline-rich
motifs (PRMs) by proteinprotein interaction
domains. Angew. Chem. Int. Ed. Engl. In the press.
7. Macias, M. J., Wiesner, S. & Sudol, M. (2002). WW and
SH3 domains, two different scaffolds to recognize
proline-rich ligands. FEBS Letters, 513, 3037.
8. Zarrinpar, A. & Lim, W. A. (2000). Converging on
proline: the mechanism of WW domain peptide
recognition. Nature Struct. Biol. 7, 611613.
9. Pires, J. R., Taha-Nejad, F., Toepert, F., Ast, T.,
Hoffmuller, U., Schneider-Mergener, J. et al. (2001).
Solution structures of the YAP65 WW domain and the
variant L30K in complex with the peptides
GTPPPPYTVG, N-(n-octyl)-GPPPY and PLPPY and
the application of peptide libraries reveal a minimal
binding epitope. J. Mol. Biol. 314, 11471156.
10. Chan, D. C., Bedford, M. T. & Leder, P. (1996).
Forming binding proteins bear WWP/WW domains
that bind proline-rich peptides and functionally
resemble SH3 domains. EMBO J. 15, 10451054.
11. Bedford, M. T. & Leder, P. (1999). The FF domain: a
novel motif that often accompanies WW domains.
Trends Biochem. Sci. 24, 531533.
12. Bedford, M. T., Chan, D. C. & Leder, P. (1997). FBP
WW domains and the Abl SH3 domain bind to a
specific class of proline-rich ligands. EMBO J. 16,
23762383.
13. Faber, P. W., Barnes, G. T., Srinidhi, J., Chen, J.,
Gusella, J. F. & MacDonald, M. E. (1998). Huntingtin
interacts with a family of WW domain proteins. Hum.
Mol. Genet. 7, 14631474.
14. Passani, L. A., Bedford, M. T., Faber, P. W., McGinnis,
K. M., Sharp, A. H., Gusella, J. F. et al. (2000).
Huntingtins WW domain partners in Huntingtons
disease post-mortem brain fulfil genetic criteria for
direct involvement in Huntingtons disease patho-
genesis. Hum. Mol. Genet. 9, 21752182.
15. Mizutani, K., Suetsugu, S. & Takenawa, T. (2004).
FBP11 regulates nuclear localization of N-WASP and
inhibits N-WASP-dependent microspike formation.
Biochem. Biophys. Res. Commun. 313, 468474.
16. Buschdorf, J. P. & Stratling, W. H. (2004). A WW
domain binding region in methyl-CpG-binding pro-
tein MeCP2: impact on Rett sydrome. J. Mol. Med. 82,
135143.
17. Kato, Y., Nagata, K., Takahashi, M., Lian, L., Herrero,
J. J., Sudol, M. & Tanokura, M. (2004). Common
mechanism of ligand recognition by group II/III WW
domains: redefining their functional classification.
J. Biol. Chem. 279, 3183331841.
18. Ermekova, K. S., Zambrano, N., Linn, H., Minopoli,
G., Gertler, F., Russo, T. & Sudol, M. (1997). The WW
domain of neural protein FE65 interacts with proline-
rich motifs in Mena, the mammalian homolog of
Drosophila enabled. J. Biol. Chem. 272, 3286932877.
(Received 9 November 2004; received in revised form 18 February 2005; accepted 22 February 2005)
PPLPp Peptide Recognition by Class II WW Domains
19. Macias, M. J., Gervais, V., Civera, C. & Oschkinat, H.
(2000). Structural analysis of WW domains and design
of a WW prototype. Nature Struct. Biol. 7, 375379.
20. Nilges, M. & ODonoghue, S. I. (1998). Ambiguous
NOEs and automated NOE assignment. Prog. NMR
Spectrosc. 32, 107139.
21. Linge, J. P., ODonoghue, S. I. & Nilges, M. (2001).
Automated assignment of ambiguous nuclear
Overhauser effects with ARIA. Methods Enzymol.
339, 7190.
22. Huang, X., Poy, F., Zhang, R., Joachimiak, A., Sudol,
M. & Eck, M. J. (2000). Structure of a WW domain
containing segment of dystrophin in complex with
b-dystroglycan. Nature Struct. Biol. 7, 634638.
23. Sudol, M. (1998). From Src homology domains to
other signaling modules: proposal of the protein
recognition code. Oncogene, 17, 14691474.
24. Wiesner, S., Stier, G., Sattler, M. & Macias, M. J. (2002).
Solution structure and ligand recognition of the WW
domain pair of the yeast splicing factor Prp40. J. Mol.
Biol. 324, 807822.
25. Toepert, F., Pires, J. R., Landgraf, C., Oschkinat, H. &
Schneider-Mergener, J. (2001). Synthesis of an array
comprising 837 variants of the hYAP WW protein
domain. Angew. Chem. Int. Ed. Engl. 40, 897900.
26. Verdecia, M. A., Bowman, M. E., Lu, K. P., Hunter, T.
& Noel, P. (2000). Structural basis for phosphoserine-
proline recognition by group IV WW domains. Nature
Struct. Biol. 7, 639643.
27. Wu, X., Knudsen, B., Feller, S. M., Zheng, J., Sali, A.,
Cowburn, D. et al. (1995). Structural basis for the
specific interaction of lysine-containing proline-rich
peptides with the N-terminal SH3 domain of c-Crk.
Structure, 3, 215226.
28. Frank, R. (1992). Spot synthesis: an easy technique
for the positionally addressable, parallel chemical
synthesis on a membrane support. Tetrahedron, 48,
92179232.
29. Kramer, A. & Schneider-Mergener, J. (1998). Synthesis
and screening of peptide libraries on continuous
cellulose membrane supports. Methods Mol. Biol. 87,
2539.
30. Sattler, M., Schleucher, J. & Griesinger, C. (1999).
Heteronuclear multidimensional NMR experiments
for the determination of proteins in solution employ-
ing pulsed field gradients. Prog. NMR Spectrosc. 34,
93158.
31. Kay, L. E. (1995). Pulsed field gradient multi-dimen-
sional NMR methods for the study of protein
structure and dynamics in solution. Prog. Biophys.
Mol. Biol. 63, 277299.
32. Wishart, D. S. & Sykes, B. D. (1994). The 13C chemical-
shift index: a simple method for the identification of
protein secondary structure using 13C chemical-shift
data. J. Biomol. NMR, 4, 171180.
33. Brunger, A. T., Adams, P. D., Clore, G. M., DeLano,
W. L., Gros, P., Grosse-Kunstleve, R. W. et al. (1998).
Crystallography & NMR system: a new software suite
for macromolecular structure determination. Acta
Crystallog. sect. D, 54, 905921.
Edited by P. Wright