| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Department of Pharmacology, University of Wisconsin-Madison, Madison, Wisconsin 53706-1532, USA;
Department of Molecular Biology and Biotechnology, University of Sheffield, Western Bank, Sheffield S10 2TN, United Kingdom;
Department of Biochemistry and Molecular Biology, Mount Sinai School of Medicine, New York, New York 10029-6574, USA
1Correspondence: Department of Pharmacology, University of Wisconsin, 1300 University Ave., Madison, WI 53706-1532, USA. E-mail: bkkay{at}facstaff.wisc.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT SH3 DOMAINSWW DOMAINSOTHER MODULES/PROTEINS THAT BIND...BIOPHYSICAL REASONS WHY PROLINE...THE REGULATION OF PROLINE...POLYPROLINE PEPTIDOMIMETICSFUTURE PROSPECTSREFERENCES |
|---|
. Other examples include
protein-interaction modules, such as Src homology (SH) 2 and 3
domains, phosphotyrosine binding domains (PTB), postsynaptic
density/disc-large/ZO1 (PDZ) domains, WW domains, Eps15 homology (EH)
domains, and 14–3-3 proteins that typically recognize linear regions
of 3–9 amino acids. Each of these domains has been the subject of
recent reviews published elsewhere (2
3
4
5
6
7)
. Among the
primary structures of many ligands for protein–protein interactions,
the amino acid proline is critical. In particular, SH3, WW, and several
new protein-interaction domains prefer ligand sequences that are
proline-rich. In addition, even though ligands for EH domains and
14–3-3 domains are not proline-rich, they do include a single proline
residue. This review highlights the analysis of those protein–protein
interactions that involve proline residues, the biochemistry of
proline, and current drug discovery efforts based on proline
peptidomimetics.—Kay, B. K., Williamson, M. P., Sudol, M. The
importance of being proline: the interaction of proline-rich motifs in
signaling proteins with their cognate domains. |
SH3 DOMAINS |
|---|
|
TOPABSTRACTSH3 DOMAINSWW DOMAINSOTHER MODULES/PROTEINS THAT BIND...BIOPHYSICAL REASONS WHY PROLINE...THE REGULATION OF PROLINE...POLYPROLINE PEPTIDOMIMETICSFUTURE PROSPECTSREFERENCES |
|---|
, ras GTPase activator, spectrin, and tight junction protein
ZO-1. Computer-aided analysis of protein sequences has even suggested
that SH3 domains may also exist in bacteria (7a)
. For a
long time, the function of the SH3 domain in eukaryotes was enigmatic;
however, since the domain was present in so many cytoskeletal proteins,
it was assumed to play a role in mediating protein–protein
interactions (8)
or directing cell compartmentalization
(9)
. The SH3 domain clearly had an important function for
some proteins, as mutations in the SH3 domains of cellular Src
(10)
and Abl (11
, 12)
are activating.
The ligand specificity of SH3 domains was first revealed when a
-cDNA expression library was screened with a glutathione
S-transferase (GST)-Abl SH3 fusion protein (13)
. Two
different cDNA clones were isolated and the regions responsible for
their binding to the Abl SH3 domain were shown to be proline-rich
segments (14)
. Examination of other SH3 ligands identified
in the same manner (15
16
17)
has suggested that SH3 domains
recognize proline-rich sequences containing the core PxxP, where x
denotes any amino acid (18
, 19)
.
With the advent of combinatorial peptide libraries, ligand specificity
has been determined rapidly for a large number of different SH3
domains. The SH3 domains of Src and the phosphatidyl inositol 3-kinase
(PI3K) regulatory subunit select proline-rich sequences from peptides
displayed on beads (20)
and phage (21
22
23)
.
Phage libraries have also been used to identify the specificity of SH3
domains from Abl, amphiphysin I, cortactin, Crk, Fyn, Grb2, Lyn, Nck,
p53BP2, PI3K, PLC, Tsk, and Yes (22
, 24
25
26
27)
. The
optimal ligand preference for each domain varies around the PxxP core
(Table 1
), demonstrating that a great deal of protein–protein interaction
specificity can be encoded by proline-rich sequences.
|
Investigations into the structure of peptide-SH3 complexes have shown
that peptide ligands can bind in two orientations with respect to the
SH3 domain (28
, 29)
. The peptides, which bind in either an
N to C or C to N terminal orientation relative to the SH3 domain, have
been classified as either class I or class II ligands, respectively.
The orientation of the peptide is dictated by the location of a
positively charged residue, relative to the PxxP core, which forms a
salt bridge with an acidic residue in the SH3 domain (Fig. 1A
). Thus, peptides with the motif +xxPxxP and PxxPx+ (where +
refers to a positively charged amino acid) correspond to class I and
class II motifs, respectively. Some SH3 domains, like that present in
Src, have the capacity to bind peptides of either class (i.e.,
RPLPPLP and PPVPPR; where
the scaffolding prolines are underlined), although the biological
implications of this are unclear. (One frequently raised possibility is
that ligand binding in two orientations could access different partners
in multicomponent protein complexes.) In both orientations, not only do
proline and leucine side-chains fit into the same hydrophobic pockets
on the SH3 domain (Fig. 1A
), but also very similar hydrogen
bonds can be made from peptide carbonyls in the proline-rich ligand
(30)
. Typically, the Kd values of
synthetic peptides binding SH3 domains are 1–100 µM, although amino
acid analoging experiments have led to the development of higher
affinity peptide ligands (31)
. The
Kdvalues of full-length proteins that interact with
SH3 domains are significantly lower than for peptides; for example, the
HIV Nef protein binds to the Hck SH3 domain with a
Kd of ~250 nM
(32)
.
|
There are three shallow pockets within the SH3 domain where the
peptide ligands bind. Two of the pockets are 25 Å long and 10 Å wide,
large enough to accommodate each of the prolines in the PxxP motif,
accompanied by a hydrophobic residue (i.e., A, I, L, V, and P). These
two pockets are parallel to each other and have been termed the ‘LP
dipeptide pockets’ (33)
. A third pocket, termed the
‘specificity pocket’ is bound by two loops (i.e., RT, nSrc) of the
SH3 domain (34)
. Within this pocket sit residues either
NH2- or carboxyl-terminal of the positively
charged residue in class I or class II ligands, respectively.
Besides the two orientations described above, a third SH3 ligand
corresponding arrangement has been described. In this arrangement,
residues spaced apart in a protein sequence come together, because of
the tertiary structure of the protein, to interact with the
peptide-binding groove of the SH3 domain. One example of this
arrangement occurs within the p53 binding protein, p53BP2. The SH3
domain of p53BP2 interacts with two segments of the L3 loop of p53, in
a manner analogous to the contacts made in canonical SH3-PxxP
complexes, with the positions and orientations of the interacting
residues determined by the overall 3-dimensional structure of p53,
rather than as part of a PxxP core (35)
. It should be
mentioned that p53 does contain two PxxP motifs that have been shown to
be involved in growth suppression through interactions with other SH3
domain-containing cellular proteins (36)
.
Recently, a novel ligand sequence has been observed for the SH3 domain
of the epidermal growth factor receptor substrate, Eps8, which has been
discovered to form intertwined dimers (37)
. Because the
site of dimerization includes the interface of the SH3 domain that
typically interacts with PxxP motifs, the specificity of the
intertwined dimers is altered. Analysis of the ligand preferences of
this dimer has revealed it to recognize not PxxP, but PxxDY instead
(37a)
. It will be interesting to learn how common this
unusual arrangement is among other SH3 domains.
Analysis of several SH3 domain interactions has suggested that
some form stable complexes and others may be regulated in the cell. For
example, the interaction between Sos and Grb2, which is constitutively
complexed in cells, can be inhibited by phosphorylation of
serine/threonine residues in the proline-rich carboxyl-terminal segment
of Sos, after receptor activation (38)
. Phosphorylation of
a PxxP site in the Wiskott-Aldrich syndrome protein (WASP) can inhibit
its interaction with the SH3 domain of the cytoskeletal-associated
protein, PSTPIP (39)
. Although the sites have not been
mapped, phosphorylation of the Drosophila-enabled (Ena)
protein by Abl can inhibit the interaction of Ena with certain SH3
domains (40)
. Finally, binding of phosphopeptide ligands
to the Crk SH2 domain alters its conformation such that a proline-rich
insert in the Crk SH2 domain becomes an SH3 domain-binding site
(41)
. Very likely, the interaction between other SH3
domains and ligand-containing proteins is regulated by allosteric
changes (42)
. A recent publication capitalizes on this
conclusion by generating temperature-sensitive SH3 domains for the
purpose of dissecting SH3 domain function in vivo
(43)
. Such an approach may be applicable to other protein
interaction modules.
Although most SH3 domain-mediated interactions are
intermolecular, several intramolecular interactions have been
described. In Src (44
, 45)
, Hck (46
, 47)
, and
Abl (48)
, the SH3 domains interact with a linker region
between the SH2 and catalytic kinase domains of these proteins, thereby
adopting a catalytically inactive conformation (49)
. In
some cases, this conformation is stabilized by the interaction of the
SH2 domain with phosphorylated tyrosine residues present near the
carboxyl termini of such proteins. The elucidation of these various
intramolecular interactions explains the observations that mutations in
the SH3 domains of Src (10)
and Abl (50)
are
activating, because the mutant forms fail to adopt an inactive
conformation. Another example occurs in the cellular oncogene, Itk,
where a proline-rich sequence, KPLPPTP (residues 155–160) binds to an
SH3 domain (residues 170–232) within the same protein
(51)
. The intramolecular interaction is quite stable, even
without the involvement of an SH2 domain, although the interaction of
the Itk SH3 domain with a soluble form of the KPLPPTP peptide sequence
is undetectable.
|
WW DOMAINS |
|---|
|
TOPABSTRACTSH3 DOMAINSWW DOMAINSOTHER MODULES/PROTEINS THAT BIND...BIOPHYSICAL REASONS WHY PROLINE...THE REGULATION OF PROLINE...POLYPROLINE PEPTIDOMIMETICSFUTURE PROSPECTSREFERENCES |
|---|
. The WW domain
was initially identified by computer-aided analysis of imperfectly
repeated sequences in the murine form of Yes-associated protein (YAP;
http://www.bork.embl-heidelberg.de/Modules.ww-gif.html).
Although the WW domain resembles the SH3 domain functionally by
displaying affinity toward proline-rich ligands, their structures are
distinct (53
54
55)
. WW domains have an antiparallel
three-stranded ß-sheet structure that forms a shallow binding pocket
for ligands containing PPxY or PPLP core motifs, usually flanked
by additional prolines (53
, 56
, 57)
. Recently another
motif, with a preliminary consensus PxxGMxPP (Table 1)
, was proposed
for ligands interacting with a subset of WW domains present in proteins
participating in the pre-mRNA splicing machinery (58)
.
These WW domains have a distinguishing feature of three consecutive
tyrosine residues located centrally within them.
The first structure of a WW domain was that of human YAP in
complex with its cognate ligand, solved by Nuclear Magnetic Resonance
spectroscopy (54)
. Shortly thereafter, the structure of
Pin1 WW domain was solved by X-ray (59)
. These two
structures are almost completely superimposable. The WW domain is the
most compact globular structure known to occur naturally, and it is the
smallest ß-sheet module that folds as a monomer in solution without
disulfide bridges or cofactors (60)
. The hallmarks of the
binding pocket of the WW domain of human YAP include three hydrophobic
amino acids, leucine, tyrosine, the second conserved tryptophan and
histidine, as confirmed by structural and mutational studies
(54)
. Two prolines of the ligand (PPxY) form
van der Waals contacts with the second tryptophan, whereas the terminal
tyrosine of the ligand fits into a hydrophobic pocket and is
coordinated by a hydrogen bond from the conserved histidine residue
(Fig. 1B
). The Kd of interaction for
WW-ligand complex formation is in the high nM to
low µM values for proline-rich ligands, and in the low µM values
for phosphoserine or phosphothreonine containing ligands, depending
upon the domain-ligand pair, buffer conditions, and the binding assay
(57
, 61
, 62)
. Phosphorylation of the terminal tyrosine in
the ligand (PPxY) abolishes the binding in vitro,
suggesting that this modification could represent a negative
regulation mechanism for a subset of WW domains in
vivo (63)
.
Immediately after its discovery, the WW domain attracted attention
because the signaling complexes it mediates have been implicated
directly or indirectly in several human diseases including Liddle’s
syndrome of hypertension, muscular dystrophy, and Alzheimer’s and
Huntington’s diseases (64
65
66
67
68
69)
. Liddle’s syndrome
results from genetic lesions that affect ß and subunits of the
amiloride-sensitive epithelial sodium channel (70)
. Most
of the mutations that have been reported in patients with Liddle’s
syndrome represent deletions that encompass a minimum of 12 residues,
which includes the PPxY motif. Three independent Liddle’s syndrome
patients were recently characterized who have single-point mutations in
one of the three crucial positions of the PPxY motif in the ß subunit
of the sodium channel (70
71
72)
. These mutations (i.e.,
P615S, P616L, and Y618H) substantiate biologically the conclusions
first generated by in vitro alanine scanning of a short
region within YAP WW domain interacting proteins (53
, 63)
.
Liddle’s mutations abolish binding of sodium channel subunits to the
Nedd-4 protein, which in addition to containing three WW domains
includes an ubiquitin-ligase catalytic domain (66
, 73
, 74)
. Nedd-4 normally participates in targeting the epithelial
sodium channel subunits for ubiquitin-directed proteolysis, but in
mutant epithelial cells the channel has a long half-life, leading to a
sodium imbalance and subsequently high blood pressure. Given that
epithelial sodium channels are phosphorylated on serine, threonine, and
possibly tyrosine residues, it will be interesting in the future to
learn whether or not these modifications modulate the interaction
between the sodium channel and the WW domains of Nedd-4 and thereby
change the rate of channel degradation (75
, 76)
.
Recently, the PPxY motif has been observed in a number of transcription
factors (i.e., c-Jun, AP2, NF-E2, C/EBP
, PEBP2/CBF) where
it may play a role in transcriptional activation. For example, the
hematopoietic transcription factor, NF-E2, contains two PPxY motifs
that can be recognized by the WW domains contained within certain
ubiquitin ligases (61
, 77)
. Interaction of WW
domain-containing proteins with this motif is likely to be important
for transcriptional activation, as deletion of one of the two PPxY
motifs in NF- E2 (61)
or the single copy in PEBP2
(78)
inhibits their ability to transactivate genes. Thus,
the presence of the PPxY motif within transcription factors may
function to recruit WW domain-containing proteins such as YAP
(78)
and Npw38 (79)
, which have been
discovered to act as transcriptional coactivators. Conversely, it is
possible in some cases that WW domain-containing proteins serve to
negatively regulate transcription. Recently, PQBP-1, a novel
polyglutamine tract binding protein with a WW domain, has been shown to
inhibit transcription activation by Brn-2, although the role of its WW
domain is yet to be defined (80)
.
|
OTHER MODULES/PROTEINS THAT BIND PROLINE-RICH LIGANDS |
|---|
|
TOPABSTRACTSH3 DOMAINSWW DOMAINSOTHER MODULES/PROTEINS THAT BIND...BIOPHYSICAL REASONS WHY PROLINE...THE REGULATION OF PROLINE...POLYPROLINE PEPTIDOMIMETICSFUTURE PROSPECTSREFERENCES |
|---|
. The EVH1 (Enabled, VASP,
Homology 1) domain is a protein interaction
module present in Ena, vasodilator-stimulated phosphoprotein (VASP),
and the WASP family of proteins that regulate the dynamics of the actin
cytoskeleton (82
83
84
85)
. EVH1 domains bind the proline-rich
consensus sequence (D/E)FPPPP, which is present in ActA, a protein on
the surface of the bacterial pathogen Listeria monocytogenes
as well as in vinculin and zyxin, two protein components of focal
adhesions (84)
. The EVH1 domain has also been detected in
Homer, a neuronal protein enriched in excitatory synapses
(86)
, and which binds to proline-rich motifs (i.e., PPxxF)
within glutamate and inositol trisphosphate receptors
(86
87
88)
and Shank, a postsynaptic density protein
(89)
. Recently, the 3-dimensional structure of the EVH1
domain has been solved for the mouse Ena protein (90)
.
Much like SH3 and WW domains, aromatic residues within the EVH1 domain
create an interaction surface for proline-rich peptide ligands that
adopt a polyproline II (PP II) conformation (see below), even though
the binding site is V-shaped instead of flat.
Two other proteins that bind proline-rich peptides should be mentioned
as well. A cytosilic protein, CD2 binding protein (CD2BP2), was
recently described in a pathway that regulates CD2-triggered T
lymphocyte (91)
. The CD2BP2 protein contains a domain that
was shown to interact with a tandemly repeated PPPGHR sequence (Table 1)
and its structure was recently solved (91a)
. Profilin
was originally described as a regulator of actin polymerization
(92)
. Within its structure, profilin, like SH3, WW, and
EVH1 domains, exposes stacked aromatic and hydrophobic residues to form
an interface that binds proline-rich ligands (82
, 93
, 94)
.
The binding of polyproline has been shown to be essential for profilin
function (95)
; profilin require a minimum of 6–8
consecutive prolines for high affinity (92
, 94
, 96)
. Some
of these cores are flanked by leucines, providing a basis for
speculation about degeneracy of certain proline-rich ligands that might
interact with both profilin and SH3 and WW domains with similar
affinities (81
, 94
, 97)
. Indeed, it has been speculated
that profilin may provide the link between signaling pathways and
remodeling of the actin cytoskeleton (98)
, a good example
of the versatile biological role made possible by the promiscuity
exhibited by proline-rich regions and their ligands.
|
BIOPHYSICAL REASONS WHY PROLINE IS A COMMON BINDING MOTIF |
|---|
|
TOPABSTRACTSH3 DOMAINSWW DOMAINSOTHER MODULES/PROTEINS THAT BIND...BIOPHYSICAL REASONS WHY PROLINE...THE REGULATION OF PROLINE...POLYPROLINE PEPTIDOMIMETICSFUTURE PROSPECTSREFERENCES |
|---|
angles
of ~-65°. It also restricts the conformation of the residue
preceding the proline because of the bulk of the
N-substituent and results in a strong preference for a
ß-sheet conformation (99)
. As a consequence, polyproline
sequences tend to adopt the PP II helix, which is an extended structure
with three residues per turn. This implies that the two prolines in the
SH3 domain ligand core, PxxP, are on the same face of the helix and are
thus well placed to interact with the protein. The PP II helix is an
unusual structure: the prolines form a continuous hydrophobic strip
round the surface of the helix, while the backbone carbonyls present
ideal hydrogen bonding sites, being both conformationally restricted
(and therefore poorly hydrated) and electron-rich (Fig. 1C
).
Therefore, PP II helices present an easily accessible hydrophobic
surface, as well as a good hydrogen-bonding site. The accessibility of
PP II helices is greatly enhanced by the fact that they are frequently
found either at the NH2- or carboxyl termini of
proteins where they form extended structures that have been described
as ‘sticky arms’ (100)
.
PP II helices are common in globular proteins with solved 3-dimensional
structures (101)
, where they are generally solvent exposed
and amphipathic (102)
. They are probably even more common
in proteins that have not been characterized structurally, because they
tend to occur in extended regions that are hard to characterize using
X-ray diffraction or NMR spectroscopy. It does not require an unbroken
sequence of prolines to make a PP II helix; in fact, this type of
left-handed helix can form in globular proteins without protein in the
helix (101)
. By contrast, a single nonproline residue may
be enough to interrupt a PP II helix in a completely exposed chain
(103)
, and it has commonly been observed that the flanking
prolines around or within the core PxxP SH3 ligand sequence function to
maintain the required PP II structure (104)
.
The relative rigidity of polyproline stretches means that they lose
little conformational entropy on binding and thus bind more favorably
than other exposed (i.e., nonglobular) peptide sequences. Of course,
proline-rich sequences cannot bind as tightly as globular domains can;
however, weaker binding can be of great advantage, as it allows the
binding of proline-rich regions to be modulated rapidly. It also
permits large changes to be made in Kd by small
changes in the sequence of the proline-rich sequence or of its binding
domain, either by sequence changes or by covalent modification such as
phosphorylation. The entropic stabilization has been estimated to be
~1 kcal mol-1 per amino acid, both
experimentally (105)
and theoretically (100)
.
The interaction is largely hydrophobic and, therefore, does not require
highly complementary surfaces. This accounts for the large sequence
variability seen in polyproline binding and for the remarkable ability,
described above, of SH3 domains to bind polyproline sequences in both
orientations, two features that give proline-rich ligands great
versatility in signaling pathways (30)
.
Because proline-rich regions are exposed, their on-rates and off-rates
for binding can be very fast. However, the price to be paid for these
fast rates is that the complexes are not structurally very well defined
on a nanometer scale. Therefore, proline-rich sequences are commonly
found in situations requiring the rapid recruitment or interchange of
several proteins, such as during initiation of transcription, signaling
cascades, and cytoskeletal rearrangements. Here, the role of
proline-rich regions is not to provide a structurally defined complex
but rather to bring proteins together in such a way that subsequent
interactions are more probable. For example, the proline-rich region on
Sos1 binds to the two SH3 domains of Grb2, thereby bringing Sos1 to the
cell membrane following receptor activation, where Sos1 in turn then
activates the Ras pathway (100)
. It is significant that
the proline-rich protein is part of an ‘adaptor’ system bringing
together other proteins. Proline-rich regions also seem to participate
in other such adaptor systems, for example, in synaptic vesicle
endocytosis (106)
.
Some proteins contain tandem proline-rich repeats. These generally seem to have a different function, in that they are involved in the formation of networks of interactions that lead to precipitation or rigid meshes, as, for example, in salivary protein/polyphenol interactions or insect eggshells (100). An interesting role of multiple repeats appears to be in the actin-based movement of Listeria monocytogenes, where each repeat, within the EVH1 domain of ActA, contributes to the rate of actin-based movement (107). This latter provides yet another example of the ‘adaptor’ role of proline-rich regions. The binding of ActA (as well as zyxin and vinculin) to VASP and of VASP to profilin are both mediated by proline-rich regions, thus using proline-rich ligands indirectly to bring together ActA and actin (82).
|
THE REGULATION OF PROLINE-DEPENDENT INTERACTIONS |
|---|
|
TOPABSTRACTSH3 DOMAINSWW DOMAINSOTHER MODULES/PROTEINS THAT BIND...BIOPHYSICAL REASONS WHY PROLINE...THE REGULATION OF PROLINE...POLYPROLINE PEPTIDOMIMETICSFUTURE PROSPECTSREFERENCES |
|---|
. Because an intramolecular
interaction is generally more energetically favorable than the
equivalent intermolecular interaction, the manipulation of
intramolecular interactions, by conformational change or
phosphorylation, is likely to be common.
Phosphorylation is a very common mechanism for regulation of protein
function and is used extensively in the context of proline-rich regions
(108)
. Proline-rich regions often contain serine or
threonine, which frequently occur as the residue immediately preceding
the proline. Specific enzymes control phosphorylation and
dephosphorylation: for example, the microtubule-associated protein 2
(MAP2) is developmentally regulated by phosphorylation, which controls
its binding to tubulin via a proline-rich region
(109)
. Both MAP2 kinase and p34cdc2
(108)
have a preference for the sequence Px(S/T)P, in
which the two prolines are separated by two residues, suggesting that
the proline-dependent kinase may well be recognizing a PP II
conformation in the substrate. As noted above, some WW domain ligands
represent target sites for proline-directed serine-threonine kinases
and have a phosphorylatable residue frequently in the vicinity of the
prolines. When decorated by phosphorylation, the serines and threonines
that flank selected PPxY cores of WW domain ligands have the ability to
positively or negatively regulate the binding (Korosi, T., Chang, A.,
and Sudol, M., unpublished results). Very recently, WW domains of two
proteins, Ess1/Pin1 and Nedd-4, were shown to recognize phosphoserine
or phosphothreonine containing ligands (Table 1)
in a
phosphorylation-dependent manner (62)
. It is likely that,
in addition to the PPxY binding pocket, these WW domains may have
another crevice that accommodates the phosphoserine or phosphothreonine
residues. It remains to be determined how phosphorylation affects the
ligand binding of the 160 WW domains known so far.
It has also been shown that phosphorylation changes the
cis/trans interconversion rate of the peptide bonds between
serine/threonine and proline. A candidate enzyme is the peptidyl-prolyl
cis/trans isomerase, Ess1/Pin1, which catalyzes the
isomerization of phosphorylated Ser/Thr-Pro (110)
. This
activity is essential for progression through the cell cycle
(111
, 112)
and restoration of the microtubule-binding
activity of 
, a protein component of neurofibrillar tangles of
Alzheimer’s patients (113)
. Peptide bond isomerization
could alter binding either directly, by changing the shape of the
ligand, or indirectly, by affecting enzyme-catalyzed hydrolysis of the
bond. Therefore, it will be of great interest when the 3-dimensional
structures of the Ess1/Pin1 WW domain complexed with its phosphorylated
ligands (59
, 62)
are determined.
Proline can have several other specific functions in proteins
unrelated to the theme of binding elaborated here. Proline can induce
ß-turns, particularly if preceded by tyrosine (114)
or
followed by phenylalanine or tryptophan residues (115)
.
Proline can also introduce bends in transmembrane helices, and because
of its rigidity can act as a conformational ‘switch’, allowing parts
of proteins to adopt alternative conformations such as domain-swapped
dimers (116)
.
|
POLYPROLINE PEPTIDOMIMETICS |
|---|
|
TOPABSTRACTSH3 DOMAINSWW DOMAINSOTHER MODULES/PROTEINS THAT BIND...BIOPHYSICAL REASONS WHY PROLINE...THE REGULATION OF PROLINE...POLYPROLINE PEPTIDOMIMETICSFUTURE PROSPECTSREFERENCES |
|---|
or
electroporated into mast cells (117)
and led to a
biological response (i.e., acceleration of progesterone stimulated
oocyte maturation, inhibition of mast cell activation). Although most
peptides fail to cross the membrane on their own, in the future it
should be possible to link them to peptide segments of
Antennapedia (118
119
120)
, Kaposi fibroblast
growth factor (121)
, or HIV Tat (122)
, which
do have the capacity to cross the plasma membrane. Recently, such an
approach has been used to introduce a dimeric peptide segment of Sos
into fibroblasts where it was observed to block the endogenous Ras
signaling pathway (123)
. Work from the laboratory of
Stephan Feller (University of Wurtzburg) has shown that a high-affinity
peptide ligand for the NH2-terminal SH3 domain of
Crk, when attached to a cell permeable peptide sequence, has biological
activity when added to tissue culture cells (Kardinal, C., and Feller,
S., unpublished observation).
A number of proline-rich peptides, which are antimicrobial and adopt a
PP II conformation, such as indolicidin (124)
and
bactenecin (125)
, likely act by crossing cell membranes on
their own. When neutrophils are incubated with the PR-39 peptide, a
proline- and arginine-rich antimicrobial peptide, NADPH oxidase
production of superoxide anion O2- is blocked.
The peptide is presumed to act by binding to the SH3 domains of the
NADPH oxidase subunits, thereby blocking assembly of a functional
enzyme (126)
. Furthermore, when NIH 3T3 cells are
incubated with the 15 residue NH2-terminal
segment of PR-39, the peptide crosses the membrane and binds SH3
domain-containing proteins, such as p130Cas
(127)
.
There has been great interest in designing peptidomimetic antagonists
of protein-protein interactions. Such antagonists should have value in
disrupting specific protein–protein interactions in the cell for the
purposes of evaluating the functional consequences of the interaction
and drug discovery. Stuart Schreiber and colleagues (Harvard
University) have attempted to design inhibitors of the Src SH3 domain
by initially synthesizing a library of compounds that contained
nonpeptidic elements fused to the NH2 terminus of
the core sequence PLPPLP. A combinatorial library of compounds, which
was synthesized on beads, was incubated with a biotinylated Src SH3
domain, and positive beads were revealed when incubated with
streptavidin-linked alkaline phosphatase and a chromogenic substrate.
Only those beads with nonpeptidic elements that fit into the
specificity pocket of the Src SH3 domain yielded positive beads,
because the Kd of the PLPPLP peptide by itself is 1
mmol/l (128
, 129)
. One ligand was selected for further
development. Through screening a second-generation ‘tuning library’,
ligands with nonpeptidic elements were identified that fit the
specificity (130)
and LP pockets (33)
.
Other research groups have attempted to create peptidomimetics of the
scaffolding prolines of SH3 ligands. This could have major consequences
for drug design, because it allows a search for novel functionalities
to replace the proline that can interact with new regions of the SH3
and hence enhance binding. Novel oligomers containing proline analogs
have been observed to adopt a PP II conformation (131
, 132)
. It also appears to be possible to replace proline by other
N-substituted amino acids. Recently, a paper from the laboratory of
Wendell Lim (University of California, San Francisco) demonstrated that
the two prolines in the PxxP motif could tolerate substitution by
N-substituted glycine, or ‘peptoid’, residues (133)
.
From a survey of six SH3 domain-peptide ligand pairs, N-substituted
ligands were discovered to bind better than the original peptides. In
particular, one ligand was identified that selectively bound the
NH2-terminal SH3 domain of Grb2 with 100-fold
greater affinity than the original 12-mer peptide sequence. Thus, in
conjunction with cell permeable peptides, it may be possible in the
future to generate peptidomimetics that are potent antagonists of
specific SH3 and WW domain based protein–protein interactions in
cultured cells and whole animals.
|
FUTURE PROSPECTS |
|---|
|
TOPABSTRACTSH3 DOMAINSWW DOMAINSOTHER MODULES/PROTEINS THAT BIND...BIOPHYSICAL REASONS WHY PROLINE...THE REGULATION OF PROLINE...POLYPROLINE PEPTIDOMIMETICSFUTURE PROSPECTSREFERENCES |
|---|
are more defined and the
structural determinants of the domains are illuminated, one should be
able to make such predictions with the aid of computer
modeling/simulation packages. Such analyses could provide useful
paradigms for analogous future approaches directed toward
characterizing the human proteome.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTSH3 DOMAINSWW DOMAINSOTHER MODULES/PROTEINS THAT BIND...BIOPHYSICAL REASONS WHY PROLINE...THE REGULATION OF PROLINE...POLYPROLINE PEPTIDOMIMETICSFUTURE PROSPECTSREFERENCES |
|---|
, Crk, and Grb2. Proc. Natl. Acad. Sci. USA 93,1540-1544
