Importance of a Conserved Lys/Arg Residue for Ligand/PDZ Domain Interactions as Examined by Protein Semisynthesis

Article abstract of DOI:10.1002/cbic.201600322

PDZ domains are ubiquitous small protein domains that are mediators of numerous protein–protein interactions, and play a pivotal role in protein trafficking, synaptic transmission, and the assembly of signaling-transduction complexes. In recent years, PDZ domains have emerged as novel and exciting drug targets for diseases (in the brain in particular), so understanding the molecular details of PDZ domain interactions is of fundamental importance. PDZ domains bind to a protein partner at either a C-terminal peptide or internal peptide motifs. Here, we examined the importance of a conserved Lys/Arg residue in the ligand-binding site of the second PDZ domain of PSD-95, by employing a semisynthetic approach. We generated six semisynthetic PDZ domains comprising different proteogenic and nonproteogenic amino acids representing subtle changes of the conserved Lys/Arg residue. These were tested with four peptide interaction partners, representing the two different binding modes. The results highlight the role of a positively charged amino acid in the β1–β2 loop of PDZ domains, and show subtle differences for canonical and noncanonical interaction partners, thus providing additional insight into the mechanism of PDZ/ligand interaction.

Full text of DOI:10.1002/cbic.201600322

Accepted Article
Title: Importance of a conserved Lys/Arg residue for ligand/PDZ domain interactions
as examined by protein semisynthesis
Authors: Søren Wittrup Pedersen; Griffin E. Moran; Vita Sereikaite; Linda M.
¨
Haugaard-Kedstrom; Kristian Strømgaard
This manuscript has been accepted after peer review and the authors have elected
to post their Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by using the Digital
Object Identifier (DOI) given below. The VoR will be published online in Early View as
soon as possible and may be different to this Accepted Article as a result of editing.
Readers should obtain the VoR from the journal website shown below when it is pu-
blished to ensure accuracy of information. The authors are responsible for the content
of this Accepted Article.
To be cited as: ChemBioChem 10.1002/cbic.201600322
Link to VoR: http://dx.doi.org/10.1002/cbic.201600322
A Journal of
www.chembiochem.org

ChemBioChem
10.1002/cbic.201600322
Importance of a conserved Lys/Arg residue for ligand/PDZ domain interactions as examined by
protein semisynthesis
Søren W. Pedersen, Griffin E. Moran, Vita Sereikaitė, Linda M. Haugaard-Kedström and Kristian
Strømgaard
Center for Biopharmaceuticals, Department of Drug Design and Pharmacology, University of
Copenhagen, Universitetsparken 2, DK-2100 Copenhagen, Denmark
Correspondence should be addressed to S.W.P (sorenwp@gmail.com) or K.S.
(kristian.stromgaard@sund.ku.dk)
1

ChemBioChem
10.1002/cbic.201600322
Abstract
PDZ domains are ubiquitous small protein domains that are mediators of numerous protein-protein
interactions, and play a pivotal role in protein trafficking, synaptic transmission and facilitating
assembly of signaling-transduction complexes. In recent years, PDZ domains have emerged as novel
and exciting drug targets for diseases in the brain in particular, and understanding the molecular details
of PDZ domain mediated interactions is of fundamental importance. These interactions are mediated by
PDZ domains binding to the protein partner through either a C-terminal peptide or internal peptide
motifs. Here, we examined the importance of a conserved Lys/Arg residue in the ligand binding site of
the second PDZ domain of PSD-95 by employing a semisynthetic approach. Six semisynthetic PDZ
domains were successfully obtained comprising different proteogenic and non-proteogenic amino acids
representing subtle changes of the conserved Lys/Arg residue, and these were tested towards four
different peptide interaction partners, representing the two different binding modes. The results
highlight the role of a positively charged amino acid in the β1-β2 loop of PDZ domains, and show
subtle differences for canonical and non-canonical interaction partners, thus providing additional
insight into the mechanism of PDZ/ligand interaction.
Keywords
semisynthesis; protein-protein interactions; PDZ domains; non-proteogenic amino
acids; protein engineering
2

ChemBioChem
10.1002/cbic.201600322
Introduction
Protein-protein interactions (PPIs) are essential for almost all cellular and biochemical processes and
hence are a promising, but also challenging, class of potential drug targets.[1-4] Nature has evolved
numerous protein families that govern and facilitate such PPIs,[5, 6] one example being the protein
family of membrane-associated guanylate kinases (MAGUKs), which assist in forming multiprotein
complexes near cellular membranes.[7, 8] The family includes the postsynaptic density protein 95 (PSD-
95), which is critical for synaptic formation and function due to its ability to organize cellular
components within the post synaptic density (PSD) primarily mediated via three PSD-95/discs
large/zonula occludens 1 (ZO-1) (PDZ) domains.[9]
PDZ domains are one of the largest families of protein domains that facilitate PPIs, consisting of 256
human PDZ domains found in 142 different proteins.[10] PDZ domains play a pivotal role in trafficking,
synaptic transmission and facilitating assembly of signaling-transduction complexes by acting as
scaffolding proteins for integral membrane and cytosolic proteins.[9, 10] Notably, PDZ domains have
recently emerged as novel and interesting drug targets, in particular in relation to diseases in the brain
such as neurodegeneration, stroke and pain.[11-13] PDZ domains typically consist of 90-100 amino acids,
which form six β-strands (β1-β6) and two α-helices (α1-α2). The ligand binding site is a shallow pocket
located between the β2 strand and the α2 helix, where the ligand forms an extended antiparallel β
strand (Fig. 1A).[14-16] The canonical peptide binding mode involves a peptide C-terminal carboxylate,
which interacts with a conserved GLGF motif located just before the β2 strand in the PDZ domain (Fig.
1B). The GLGF motif, or carboxylate binding site, forms a cradle of backbone amides, which aligns a
hydrogen bond network to the peptide C-terminal carboxylate and hence stabilizes the PDZ
domain/peptide interaction. Furthermore, an extended backbone-backbone hydrogen bond network
3

ChemBioChem
10.1002/cbic.201600322
between the ligand and the β2 strand,[16, 17] together with side chain interactions, primarily governed by
the four last amino acids, are also fundamental for the canonical binding of C-terminal peptides to PDZ
domains. The four C-terminal amino acids of the peptide ligand contains a motif (P-3-P0), where amino
[18, 19]
acids in position P0 and P-2 are particular important for affinity and selectivity.
So-called class I
PDZ domains, such as the PDZ domains of PSD-95, recognize C-terminal peptides containing a T/S-X-
V/L-COOH (X being any amino acid) motif, where a hydrophobic pocket at P0 accommodates the V/L
residue and a conserved His in the α2 helix form a hydrogen bond to the side chain hydroxyl group of
the T/S residue.[15, 20] It is well established that a subset of PDZ domains recognize internal peptide
motifs, exemplified by the non-canonical interaction between the β-finger in neuronal nitric oxide
synthase (nNOS) and PSD-95, syntrophin and ZO-1.[20-22] The internal binding motifs form either an
antiparallel β-finger or an extended conformation, which aligns the β2 strand in the PDZ domain and
form interactions similar to the canonical binding mode: The P0 site accommodates an aromatic side
chain (Phe111) of the nNOS β-finger and in the P-2 site a hydrogen bond is formed between His residue
in the α2-helix and a hydroxyl group (Thr109) in nNOS.[16, 21, 23-25] In addition, structural analysis of
PDZ domains have identified interactions between the side chain of a highly conserved basic residue in
the flexible β1-β2 loop, Arg/Lys (Lys165 and Arg318 in PSD-95 PDZ2 and PDZ3, respectively),
which together with the GLGF loop stabilizes the carboxylate of C-terminal ligands through a water-
mediated hydrogen bond (Fig. 1A).[15, 21, 26] It is known that Arg318 in PSD-95 PDZ3 can act as a
switch allowing anionic strength to modulate the affinity of ligands to PDZ3[27] and interestingly, a
Lys165Arg mutation in PSD-95 PDZ2, disrupted the binding to nNOS, but not the canonical binding to
the C-terminal of Shaker-type K+ channel Kv1.4.[28]
4

ChemBioChem
10.1002/cbic.201600322
PSD-95 is one of the most abundant proteins in the PSD in neurons and also one of the most studied
PDZ containing proteins, and thus provides an excellent model system for studies of ligand/PDZ
domain interactions. PSD-95 is known to interact via its PDZ domains with the C-terminal of the
GluN2B subunit of the N-methyl-D-aspartate (NMDA) glutamate-type receptor and nNOS. The PSD-
95/nNOS/NMDA receptor complex have been demonstrated to be important in excitotoxicity, which is
a process in the brain believed to be involved in diseases such as ischemic stroke,[9] and notably
peptide-based inhibitors of PSD-95 PDZ domains have shown promising results in both preclinical and
clinical studies as a potential treatment of ischemic stroke.[29, 30] Thus, understanding the molecular
details of PDZ domain mediated interactions of PSD-95 is also of fundamental importance for
development of improved inhibitors of PSD-95.
Here we address the importance of the conserved Arg/Lys in PDZ domains by employing expressed
protein ligation (EPL) to introduce non-proteogenic mutations at position Lys165 in PSD-95 PDZ2
(Fig. 1C) and examine the effect on canonical and non-canonical peptide ligand interactions. Altering
the length of the Lys side chain or replacing with an Arg had no effect on ligand affinity in either
binding mode. Interestingly removing the positive charge of the side chain, as in citrulline and acetyl-
ornithine, did not influence non-canonical interaction, but had a subtle effect upon the canonical
interaction. Thus, this provides mechanistic insight into the importance of the positive charge of the
conserved Lys/Arg side chain in PDZ domains, and in particular underpins putative differences
between canonical and non-canonical interactions, which are fundamental for the understanding of the
largest family of modulators of PPIs.
5

ChemBioChem
10.1002/cbic.201600322
Results
To explore the importance of the conserved Lys/Arg residue for peptide ligand binding to PDZ
domains, we employed the prototypical PSD-95 PDZ2 domain, as well as four different 5/6-
tetramethylrhodamine (TAMRA) labelled peptide ligands, representing both canonical ligands, the
GluN2B C-terminal peptide (TAMRA-YEKLSSIESDV-COOH, GluN2B-COOH) and CRIPT (TAMRA-
DTKNYKQTSV-COOH, CRIPT-COOH), as well as non-canonical ligands, a nNOS β-finger mimetic
31-32]
(TAMRA-cyclo(CTHLETTFTGDGTPKTIRVTQpG), nNOS β-finger)[16,
and a C-terminally
amidated GluN2B peptide (TAMRA-YEKLSSIESDV-CONH2, GluN2B-CONH2). In order to evaluate the
importance of the Lys/Arg residue, we replaced Lys165 in PSD-95 PDZ2 with Arg, as well as four
non-proteogenic amino acids: First two variants of Lys, homolysine (hLys) and ornithine (Orn), where
the number of methylene groups to the primary amine is varied, and then two variants of Arg, citrulline
(Cit) and acetyl-ornithine (AcOrn), where the degree of protonation, as well as hydrogen bonding, is
varied (Fig. 1C).
In order to introduce the non-proteogenic amino acids into PDZ domains, we explored a semisynthetic
methodology for the generation of the PSD-95 PDZ2 domain comprising chemical mutations in
position 165.[16] The strategy required the generation of two fragments: A synthetic 23-mer N-terminal
peptide fragment (ΔNPDZ2) and a recombinantly expressed 72-mer C-terminal protein fragment
(ΔCPDZ2), which would be ligated at a V178C site, a mutation which has been shown not to affect
ligand binding.[16] The N-terminal fragment ΔNPDZ2, comprising the residue of interest, Lys165 and a
C-terminal hydrazide, was prepared by solid phase peptide synthesis (SPPS) (Fig. 2A), while the C-
terminal fragment ΔCPDZ2, containing a N-terminal Cys, was generated by bacterial overexpression
6

ChemBioChem
10.1002/cbic.201600322
(Fig. 2B). The two fragments then underwent chemo-selective ligation and subsequent refolding to
generate the binding competent, full-length PSD-95 PDZ2 domain (Fig. 2C).
Semisynthesis of PSD-95 PDZ2 domains. Six different N-terminal peptide fragments (ΔNPDZ2-Lys,
ΔNPDZ2-hLys, ΔNPDZ2-Orn, ΔNPDZ2-Arg, ΔNPDZ2-Cit and ΔNPDZ2-AcOrn) containing mutations
in position 165 were synthesized by Fmoc-based SPPS using a hydrazine-preloaded 2-chlorotrityl resin
and an automated peptide synthesizer. The fragment ΔNPDZ2-Lys was used to generate wild-type
semisynthetic PSD-95 PDZ2 as a control. Three of the four non-proteogenic amino acids, hLys, Orn
and Cit, were commercially available as appropriately protected building blocks. However, the
acetylated ornithine building block, Fmoc-AcOrn-OH, used to generate ΔNPDZ2 AcOrn, was prepared
from Fmoc-Orn(Boc)-OH (Scheme S1). First, the carboxylic acid was protected as an allyl ester (1),
then the N-Boc side chain protection group was removed by treatment with trifluoroacetic acid (TFA)
and subsequently acetylated with acetyl chloride. Finally, the carboxylic acid was reconstituted by
treating 2 with Pd(PPh3)4, generating Fmoc-AcOrn-OH (3) in good yield (32%) (Fig. S1). With all
Lys/Arg analogues in hand, the six N-terminal peptides comprising C-terminal hydrazide were
prepared by SPPS and obtained in good yields (16-34%) and excellent purities (>90%) (Table S1). The
C-terminal protein fragment (ΔCPDZ2) was obtained using a protocol similar to previously
described.[16] In brief, a modified fusion protein of the full-length PSD-95 PDZ2 containing a N-
terminal polyhistidine tag (MHHHHHHHRS) and a Factor Xa (FXa) cleavage site (IEGR) at position
V178C was expressed in E. coli. After metal ion affinity chromatography, the isolated protein was
subjected to FXa cleavage and the remaining Cys-terminal protein fragment (ΔCPDZ2) was isolated in
good yield and purity (13 mg/L, >95% pure) (Fig. S2, Table S2).
7

ChemBioChem
10.1002/cbic.201600322
To generate full-length PDZ domains, ligations of the two fragments were initiated by oxidation of the
C-terminal hydrazide of synthetic peptide (ΔNPDZ2) with NaNO2 for in situ generation of a C-terminal
azide (Fig. 2A), which was converted to a C-terminal thioester peptide by addition of 2-
mercaptophenyl acetic acid (MPAA).[33] Subsequently, the recombinant C-terminal protein fragment,
ΔCPDZ2, was added to the ligation mixture and ligations were complete after 2 h with ligation
efficiencies between 68-81% evaluated by depletion of reactive C-terminal protein fragment (Fig. S3).
The full-length semisynthetic PSD-95 PDZ2 domains were isolated by gel filtration, which removed
non-ligated fragments and ligation buffer components while simultaneously refolding the protein into
binding buffer. All six semisynthetic proteins were obtained in a purity of >95%, as observed by UPLC
(Fig. 2D and S5, Table S3) and SDS-PAGE (Fig. 2E). The secondary structure of the semisynthetic
PSD-95 PDZ2 domains, as well as the recombinantly expressed PSD-95 PDZ2 domain (PDZ2 WT),
were examined and compared by circular dichroism (CD) spectroscopy (Fig. 2F). Gratifyingly, all
spectra overlapped with a minimum at 212 nm, thus suggesting that all semisynthetic PDZ2 domains
had a secondary structure and fold similar to that of PDZ2 WT.
Binding between semisynthetic PDZ domains and interaction partners. In order to address the
effect of the chemically introduced mutations in the PSD-95 PDZ2 domain, we examined the binding
of four different peptide ligands, GluN2B-COOH, CRIPT-COOH, GluN2B-CONH2 and nNOS β-finger, to
the semisynthetic PDZ domains using a fluorescence polarization (FP) assay.[16, 34] These ligands
represent, as previously described, canonical (GluN2B-COOH and CRIPT-COOH) and non-canonical
(GluN2B-CONH2 and nNOS β-finger) binding modes of the peptide/PDZ domain interaction. The C-
terminally amidated ligand, GluN2B-CONH2, has previously been shown to have very low affinity
towards the PSD-95 PDZ2 domain (KD > 500 μM), mainly due to of the lack of charge in the C-
8

ChemBioChem
10.1002/cbic.201600322
terminal and not being able to engage in the hydrogen bond network to the backbone of the GLGF
motif.[16, 21] The nNOS β-finger ligand is a cyclic peptide analogue derived from nNOS, which has been
shown to mimic the native non-canonical interaction between nNOS and PDZ domains with
comparable binding affinities (KD ~1.0-1.8 uM). [16, 31-32]
First, the KD values for binding of the four peptide ligands to PDZ2 WT and the semisynthetic control
protein (PDZ2-Lys) were determined by saturation binding experiments (Fig. 3). The affinities for the
canonical ligands, GluN2B-COOH and CRIPT-COOH, were found to be in the lower micromolar range
(4.9 μM and 17 μM for PDZ2 WT and 5.4 μM and 17 μM for PDZ2-Lys, respectively) and for the non-
canonical ligands, nNOS β-finger and GluN2B-CONH2, the KD values were ranging from the low to the
high micromolar range (1.2 μM and 314 μM for PDZ2 WT and 1.2 μM and 240 μM for PDZ2-Lys,
respectively). The KD values for these ligands are generally the same for the expressed PDZ2 WT and
the semisynthetic PDZ2-Lys, which strongly indicates that the semisynthetic strategy provides
structurally intact PDZ domains. Next, affinities of the four ligands were determined for the five
mutated semisynthetic PDZ domains (Fig. S6, Table S4). First, the two proteins mutated with Lys
analogues, PDZ2-hLys and PDZ2-Orn, respectively were examined, and here binding was generally
unaffected by either side chain elongation (hLys) or shortening (Orn). Compared to PDZ2 Lys, the
mutations only gave moderate changes in affinity for both canonical and non-canonical ligands,
exemplified by binding of CRIPT-COOH to PDZ2-hLys (KD = 6.7 μM) and PDZ2-Orn (KD = 3.4 μM),
respectively (Fig. 4). Interestingly, replacing Lys with Arg, PDZ2-Arg, as also found in related PDZ
domains, [15] led to a marginal increase in the affinity with KD = 2.7 μM and 156 μM for GluN2B-COOH
and GluN2B-CONH2, respectively, while affinity of the nNOS β-finger decreased by 1.8 fold (KD = 2.1
μM). PDZ domains comprising Arg analogues, PDZ2-Cit and PDZ2-AcOrn, both showed binding
9

ChemBioChem
10.1002/cbic.201600322
properties different from those containing either Lys or Arg in the same position. Interestingly, both
introduction of Cit or AcOrn led to increase in KD values for GluN2B-COOH (18 μM and 49 μM,
respectively) and CRIPT-COOH (17 μM and 52 μM). In contrast, when examining the nNOS β-finger, a
slight increase in affinity to both PDZ2-Cit (KD = 1.1 μM) and PDZ2-AcOrn (KD = 0.89 μM) was
observed. Finally, binding of GluN2B-CONH2 to PDZ2-AcOrn was slightly decreased (KD = 464 μM),
while no change in affinity was observed for binding to PDZ2-Cit (KD = 259 μM).
Discussion
Alteration of Lys in the β1-β2-loop changes binding preferences of canonical PDZ2 binders. The
C-terminal of the GluN2B subunit is a canonical binder to PSD-95 PDZ2 and the importance of
hydrogen-bonding[16, 17] and side chain mutations[34] have previously been studied for this interaction,
but so far no X-ray crystal structure has been solved for this PDZ2 domain in complex with a peptide
ligand. In contrast, the first X-ray crystal structure of a ligand/PDZ complex was solved with the
CRIPT C-terminal peptide bound to PSD-95 PDZ3, which is currently the established model for the
canonical binding mode.[15] The CRIPT peptide has preference for the PDZ3 vs. the PDZ2 domain of
PSD-95, with KD values of 2.3 μM and 16 μM for PSD-95 PDZ3 and PSD-95 PDZ2, respectively,[18, 34]
still it is assumed that that CRIPT binds to both PDZ2 and PDZ3 in the canonical binding mode given
the high sequence similarity of the proteins, including a conserved His (PDZ2 H225, PDZ3 H372) in
the α2 helix and the carboxylate binding site (PDZ2 G169-F172, PDZ3 G322-F325).[35] One notable
difference between the two domains is the conserved basic residue in the loop region between β1-β2
(PDZ2 K165, PDZ3 R318), which facilitates a hydrogen bond to the C-terminal carboxylate of the
peptide ligand through a water molecule (Fig. 1A). Although the β1-β2 loop of PDZ2 contains the
same number of amino acids as PDZ3, the X-ray crystal structure of apo PDZ2 indicate that this loop is
10

ChemBioChem
10.1002/cbic.201600322
bending away from the GLGF site in contrast to PDZ3. This does however not seem to affect the
position of the basic moiety from the conserved Lys/Arg residue as the position of the Lys amino group
in PDZ2 still overlaps with the Arg guanidinium group in PDZ3.
Here, we examined the molecular details of this conserved basic residue in PSD-95 PDZ2, and by
employing a semisynthetic approach we could introduce refined chemical modifications in its side
chain. Three of the five semisynthetic PDZ2 domains, PDZ2-hLys, PDZ2-Orn and PDZ2-Arg, have
positively charged side chains, with variation in both the number of methylene groups separating the
amino group and the nature of the basic moiety (Fig. 1C). Compared to PDZ2-Lys, extending, PDZ2-
hLys, or shortening, PDZ2-Orn, the side chain of Lys165 has only limited effect on binding affinity
towards GluN2B-COOH and CRIPT-COOH, whereas replacing with Arg induces a small affinity
increase. Despite the differences, the side chains of hLys and Arg have a similar distance from the
terminal nitrogen of the amine/guanidinum to the backbone (Fig. 1C) thereby making them similarly
positioned to form a hydrogen bond to the water molecule. In contrast, introduction of the Arg
analogues, PDZ2-Cit and PDZ2-AcOrn, that do not have a net charge, but can still participate in
hydrogen bonding, led to a moderate decrease in affinity of the canonical ligands. This suggests that the
net charge of the basic side chain may play a role in the canonical peptide/PDZ domain interaction. A
similar tendency has previously been observed for the peptide/syntrophin interaction, where Lys86,
corresponding to Lys165 in PSD-95 PDZ2, was mutated to Met leading to a loss in binding affinity
(ΔΔGbinding = 1.5 kcal mol-1).[36]
Another interesting difference between the six side chain modifications is the ability to stabilize the
position of the charge. The side chain of Arg has three nitrogens compared to one in Lys and its
analogues, which allows Arg to form a larger number of interactions.[37] In the X-ray crystal structure
11

ChemBioChem
10.1002/cbic.201600322
of CRIPT/PDZ3,[15] the Arg318 may form hydrogen bonds to the water molecule near the GLGF motif,
as well as to the backbone carbonyls of Leu379 and Gly383 given the close proximity (3.0 Å and 2.7
Å, respectively) to the latter. These additional intramolecular interactions could stabilize the position of
the positive charge in PDZ2-Arg, but not in PDZ2-Lys, which could explain the slightly lower KD
values for the canonical binders.
Non-canonical PDZ2 domain interactions are charge independent. PSD-95 PDZ2 interacts with the
β-finger of nNOS through a non-canonical binding, where the ligand backbone aligns with the β2
strand forming an extended β-sheet.[16, 21] The X-ray crystal structure of the nNOS β-finger and the
syntrophin PDZ domain suggests that this interaction is not mediated by a water molecule near the
GLGF-site, which can be explained by the β-finger not having a free carboxylate moiety.[15] This
suggests that the basic residue (Lys86 in syntrophin) might not be directly involved in non-canonical
interactions. When comparing the sequences of PDZ domains that interact with nNOS (Fig. S7),[22, 38-41]
it is seen that both Lys and Arg are found in the position of the conserved basic residue. Thus, the
specific nature of the basic residue is apparently not important for the non-canonical interaction. In
contrast, it has previously been reported that the interaction between nNOS and PSD-95 PDZ2 was
completely disrupted upon an Arg to Lys mutation in this position.[28] Here, we observed that
regardless of the substitution introduced into position 165 of PSD-95 PDZ2, the affinity towards the
nNOS β-finger was changed less than 2 fold. The same observations were seen for the affinity of
GluN2B-CONH2 having a KD value for PDZ2 WT in the high micromolar range. Thus, it seems that the
removal of the charge in this position does not affect the non-canonical interaction, which is in
agreement with examination of the nNOS/syntrophin interaction, where the Lys86Met mutation
induced only a very modest change in binding (ΔΔGbinding = 0.5 kcal mol-1).[36] In addition, our results
12

ChemBioChem
10.1002/cbic.201600322
are in agreement, with the conclusions drawn from the X-ray crystal structure of the nNOS/syntrophin
complex, that the conserved basic residue is less important for the non-canonical interaction.
Conclusion. Here, we have addressed the importance of a basic residue (Lys or Arg), which is
conserved in many PDZ domains, for the binding of both canonical and non-canonical ligand/PDZ
domain interactions. We took advantage of a semisynthetic strategy for PDZ domains and introduced
non-proteogenic amino acids instead of Lys165 in the PSD-95 PDZ2 domain. Specifically, we
incorporated two Lys analogues by introducing hLys and Orn, and similarly two Arg analogues, Cit
and AcOrn. This provided succinct information on differences in the canonical and non-canonical
interaction in PDZ domains, where the former prefers amino acids bearing charged side chains,
whereas in the latter there was no difference in ligand affinities regardless of the mutation introduced.
Thus employing a semisynthetic approach to generate PDZ domains, allowed us to decipher the
molecular details about a key residue in PDZ domain interactions, and the methodology offers the
opportunity to introduce a wealth of additional amino acids into PDZ domains.
Experimental details
General information. Unless otherwise stated, starting materials were obtained from commercial
suppliers and used without further purification. Fmoc protected amino acids, coupling reagents and
trifluoroacetic acid (TFA) were purchased either from Peptides International (Louisville, KY, USA),
Sigma Aldrich (St. Louis, MO, USA) or Iris Biotech (Marktredwitz, Germany). All reagents and
solvents were used without further purification. The reactions for the Fmoc-AcOrn-OH synthesis were
monitored by thin-layer chromatography (TLC) using silica gel coated aluminum plates 60F-254
13

ChemBioChem
10.1002/cbic.201600322
(Merck). A CombiFlashRfTM system (Teledyne ISCO, Lincoln, NE, USA) was used for flash column
chromatography with samples being loaded onto GraceResolvTM flash cartridges (Grace, Columbia,
MD, USA). Silica gel chromatography was performed using Merck silica gel 60, 230-400 mech. NMR
spectroscopy was performed on Bruker 400 MHz apparatus. 1H NMR chemical shifts were recorded in
ppm, using CDCl3 and TMS as internal standard (0.00 ppm). The following abbreviations are used: s,
13
singlet; d, doublet; t, triplet; q, quartet; m, multiplet; br, broad. C NMR shifts are recorded in ppm
using the residual non deuterated solvent as internal standard (CDCl3 13C 77.0 ppm). NMR spectra
were analyzed using MestReNova version 10.0 by Mestrec Laboratories. Mass spectra were obtained
with an Agilent 6410 Triple Quadrupole Mass Spectrometer instrument using electron spray coupled to
an Agilent 1200 HPLC system (ESI-LC/MS) with a C18 reverse phase column (Zorbax Eclipse XBD-
C18, 4.6 × 50 mm), autosampler and diode-array detector using a linear gradient of the binary solvent
system of H2O/acetonitrile (ACN)/formic acid (A: 95/5/0.1 and B: 5/95/0.086) with a flow rate of 1
mL/min, and gradient elution from 0-100% B in 5 min. Purity was determined by ultra performance
liquid chromatography (UPLC) using an Acquity UPLC BEH C18 1.7µm 2.1 × 50 mm column over a
gradient of 5-95% Buffer B in 7 min.
Synthesis of Fmoc-AcOrn-OH
Allyl 2-((((9H-fluoren-9-yl)methoxy)carbonyl)amino)-5-((tert-butoxycarbonyl)amino) pentanoate
(1). Fmoc-Orn(Boc)-OH (1.36 g, 3 mmol) was dissolved in dimethylformamide (DMF) (10 mL) and
Cs2CO3 (1.96 g, 6 mmol) was added to reaction mixture, followed by slow addition of allylbromide
(0.39 mL, 4.5 mmol). The reaction mixture was stirred for 3 h at room temperature (rt). The mixture
was diluted with H2O (10 mL). The aqueous layer was extracted with ethyl acetate (EtOAc) (3 × 15
mL), the combined organic phases were washed with brine (20 mL), dried (MgSO4) and excess solvent
14

ChemBioChem
10.1002/cbic.201600322
removed in vacuo. The crude product was purified by silica gel chromatography (heptane/EtOAc 3:2)
to afford 1 as colorless liquid (1.21 g, 82%). δH (400 MHz, CDCl3) 7.82–7.71 (2 H, m), 7.65–7.55 (2 H,
m), 7.45–7.36 (2 H, m), 7.36–7.28 (2 H, m), 5.90 (1 H, ddt, J 16.5, 11.0, 5.8), 5.38–5.20 (2 H, m), 4.65
(2 H, d, J 5.8), 4.41 (3 H, d, J 7.1), 4.22 (1 H, t, J 6.9), 3.14 (2 H, s), 1.79 (2 H, m) and 1.44 (9 H, s); δC
(100 MHz CDCl3) 172.1, 156.1 (2 × C), 143.9 (2 × C), 141.5 (2 × C), 131.6 (2 × C), 127.9, 127.2,
125.2 (2 × C), 120.1, 119.2, 118.0, 79.5, 67.1, 66.2, 53.8, 47.4, 40.1, 30.1, 28.6 (3 × C) and 26.2.
Allyl 2-((((9H-fluoren-9-yl)methoxy)carbonyl)amino)-5-acetamidopentanoate (2). 1 (1.2 g, 2.4
mmol) was treated with TFA (5.5 mL, 72 mmol). The reaction mixture was stirred at rt and after 20
min the reaction mixture was diluted with dichloromethane (DCM) (15 mL). The solvent was
subsequently removed in vacuo. The procedure was repeated three times to afford the primary amine.
The crude primary amine was dissolved in DCM (10 mL) and cooled to 0 °C, followed by addition of
triethylamine (1 mL, 7.2 mmol) and acetyl chloride (0.205 mL, 2.88 mmol). The reaction was then
heated to rt and stirred for another 2 h until all starting material was consumed. The solvent was
removed in vacuo and the residue was purified by silica gel column chromatography (DCM/methanol
(MeOH) 5:1) to afford 2 as a white solid (848 mg, 81% over two steps). δH (400 MHz, CDCl3) 7.77 (2
H, dd, J 7.5, 1.0), 7.60 (2 H, dd, J 7.5, 4.3), 7.44–7.36 (2 H, m), 7.31 (2 H, tt, J 7.5, 1.3 Hz), 5.90 (1 H,
ddt, J 16.5, 11.0, 5.8), 5.68 (1 H, s), 5.49 (1 H, d, J 8.2), 5.38–5.22 (2 H, m), 4.65 (2 H, d, J 5.8), 4.40
(3 H, m), 4.22 (1 H, t, J 6.9), 3.34–3.21 (2 H, m), 1.96 (3 H, s), 1.94–1.67 (2 H, m) and 1.58 (2 H, q, J
8.1, 7.4); δC (100 MHz, CDCl3) 172.0, 170.3, 143.8 (2 × C), 141.5 (2 × C), 131.5 (2 × C), 127.9, 127.2,
125.2 (2 × C), 120.2, 120.1, 119.3, 67.2, 66.3, 53.7, 47.3, 39.1, 30.4, 25.6 and 23.4; LCMS obs. [M+H]
437.2, calcd. [M+H] 437.2.
15

ChemBioChem
10.1002/cbic.201600322
2-((((9H-fluoren-9-yl)methoxy)carbonyl)amino)-5-acetamidopentanoic acid (3). 2 (829 mg, 1.9
mmol) and Pd(PPh3)4 (220 mg, 0.19 mmol) were dissolved in dry tetrahydrofuran (THF) (50 mL) and
placed under nitrogen, morpholine (0.33 mL, 3.8 mmol) was added dropwise and the reaction mixture
was stirred at rt for 20 min. The solvent was removed in vacuo and the resulting residue was dissolved
in DCM (10 mL) and washed with 1M HCl (2 × 5 mL), dried (MgSO4) and excess solvent removed in
vacuo. The crude product was purified by silica gel chromatography (DCM/MeOH 10:1) to afford 3 as
a white solid (241 mg, 32%). δH (400 MHz, CDCl3) 7.73 (2 H, d, J 7.5), 7.57 (2 H, m), 7.43–7.32 (2 H,
m), 7.28 (2 H, dd, J 7.5, 1.2), 6.30 (1 H, s), 5.83 (1 H, d, J 7.5), 4.36 (3 H, d, J 7.2), 4.18 (1 H, t, J 7.0),
3.26 (2 H, td, J 13.7, 7.5), 1.95 (3 H, s), 1.91–1.65 (2 H, m) and 1.63–1.49 (2 H, m); δC (100 MHz,
CDCl3) 174.8, 172.0, 156.5, 143.8 (2 × C), 141.4 (2 × C), 127.9 (2 × C), 127.2 (2 × C), 125.2 (2 × C),
120.1 (2 × C), 67.2, 53.6, 47.2, 39.4, 30.0, 25.3 and 23.0; LCMS obs. [M+H] 397.2, calcd. [M+H]
397.21; UPLC >95% purity.
Solid-phase peptide synthesis (SPPS)
Resin loading with hydrazine. 2-Chlorotrityl chloride resin (4.26 mmol, 1.42 mmol g−1 loading, Iris
Biotech) was washed alternatingly with DMF (3 × 5 mL) and DCM (3 × 5 mL) twice. Resin was
swollen in DMF/DCM (1:1) for 30 min and subsequently washed with DMF (3 × 5 mL). 16 mmol
NH2NH2 and 5 mmol triethylamine (TEA) in DMF was added to the resin and agitated for 30 min. The
resin was washed with DMF and the procedure was repeated. After the second coupling, the resin was
washed by alternating DMF and DCM and capped using 5 mL DMF/MeOH/N,N-
diisopropylethlyamine (DIPEA) (8:1.5:0.5) for 10 min, washed with DMF and repeated once.
16

ChemBioChem
10.1002/cbic.201600322
Manual coupling of Fmoc-Gly-OH. The resin was washed twice alternating with DMF (3 × 5 mL)
and DCM (3 × 5 mL) before flow wash with DCM. 2 eq. of Fmoc-Gly-OH, 4 eq. 2-(1H-7-
azabenzotriazole-1-yl)-1,1,3,3-tetramethyl uronium hexafluorophosphate (HATU) and 4 eq. DIPEA in
DMF were added directly to the resin. After 1 h of coupling, the resin was flow washed with DMF.
Subsequently, the resin was capped using 5 mL DMF/MeOH/DIPEA (8:1.5:0.5) for 10 min, washed
with DMF and repeated once. Finally, the resin was washed with DCM and dried overnight under
vacuum.
Microwave assisted Fmoc-SPPS. All peptides were synthesized in 0.1 mmol scale using the prepared
hydrazine linked resin. Automated microwave-assisted SPPS was carried out using a Liberty Blue
microwave peptide synthesizer (CEM, Matthews, NC, USA) equipped with a 30 mL Teflon reaction
vessel. Fmoc-protected amino acids were made up as 0.2 M solutions in DMF. An activator solution
was made up as 0.5 M N,N’-diisopropylcarbodiimide (DIC) in DMF with 0.01% DIPEA and an
activator base solution was made up as 1 M ethyl 2-cyano-2-(hydroxyimino)acetate (Oxyma) in DMF
with 0.01% DIPEA. Peptide synthesis was performed using the following settings: deprotection: 150
W, 15 s, 75 °C followed by 30 W, 60 s, 90 °C, coupling: 190 W, 180 s, 75 °C followed by 36 W, 120 s,
90 °C. The introduction of Fmoc-hLys(Boc)-OH and Fmoc-Orn(Ac)-OH in ΔNPDZ2 hLys and
ΔNPDZ2 AcOrn, respectively, was performed manually using 2 eq. Fmoc-protected amino acid, 2 eq.
HATU and 4 eq. DIPEA and agitated for 1 h at rt. The coupling efficiency was monitored by Kaiser
Test (Sigma-Aldrich, St. Louis, MO, USA). After the coupling was completed, the resin was capped by
treating with 5 mL DMF/MeOH/DIPEA (8:1.5:0.5) for 10 min, washed with DMF and repeated once.
The synthesis continued on the peptide synthesizer.
17

ChemBioChem
10.1002/cbic.201600322
Cleavage of peptides from resin. Peptides were cleaved from resin using a cleavage cocktail
containing TFA/triisopropylsilane (TIPS)/H2O (90:5:5) for 2.5 h at rt. After cleavage, the peptides were
precipitated using ice cold diethyl ether and subsequently spun down. The isolated pellet was dissolved
in a mixture of buffer A (95% H2O, 5% ACN, 0.1% TFA) and buffer B (95% ACN, 5% H2O, 0.1%
TFA) and lyophilized.
Purification of peptides. Lyophilized peptides were purified by preparative reverse phase high
performance liquid chromatography (RP-HPLC) (Waters, Milford, MA, USA) using a ZORBAX
300SB-C18 PrepHT 21.2 × 250 mm 7 µm column (Agilent, Santa Clara, CA, USA) over a gradient of
20-50% Buffer B over 25 min. 10 mL fractions were collected using an automated fraction collector
and pure fractions were pooled and lyophilized. Peptides were characterized by LC-MS and purity
determined by UPLC.
Fluorescently labeled peptides were synthesized and purified as previously described.[16]
Recombinant protein expression
DNA subcloning. Codon optimized DNA gene cassettes were purchased from Invitrogen (Carlsbad,
CA, USA) containing HindIII and KpnI surrounding the PDZ2 domain comprising residues 155-249 of
PSD-95. This insert was cloned into a pRSET expression vector from Invitrogen (Carlsbad, CA, USA)
and verified by DNA sequencing (Eurofins MWG Operon, Germany). Two mutations (V178C and
Y190W) were introduced by standard site-directed mutagenesis protocol (Stratagene) by primers
(forward primer V178C 5’-CTA TCG CGG GTG GTT GTG GTA ATC AGC ATA TTC -3´; reverse
primer V178C 5´-GAA TAT GCT GAT TAC CAC AA CCA CCC GCG ATA-3´; forward primer
Y190W 5’- GCG ATA ACA GCA TTT GGG TGA CCAA AAT TAT-3´ reverse primer Y190W; 5´-
18

ChemBioChem
10.1002/cbic.201600322
CAA TAA TTT TGG TCA CCC AAA TGC TGT TAT CGC-3´) . The factor Xa (FXa) sites were
introduced by PCR amplification with elongated (+ 16 bp) primers sFXA (5´-NNN ATC GAG GGA
AGG TGT NNN-3´) and asFXA thereby inserting four extra amino acids (Ile-Glu-Gly-Arg) next to the
V178C mutation.
Protein overexpression. Expressions of PDZ2 domains were generally conducted as previously
described.[16] In brief, E. coli BL21-DE3 pLys cells were transformed with PDZ2 containing pRSET
cDNA and grown over night (ON) in 6 L ZYM5052 medium. The cells were harvested, resuspended
and lysed into solid guanidinium hydrochloride (GnHCl) and sodium dihydrogen phosphate giving a
final concentration of 6 M and 0.05 M, respectively. pH was adjusted to 7.5 and the protein was
purified using 2 × 5 mL HisTraps (GE Healthcare Life Sciences) according to the protocol. To generate
the C-terminal fragment of PDZ2, the PDZ2 IEGR construct was dialyzed into 50 mM TRIS, 150 mM
NaCl, 1 mM CaCl2, pH 7.5 ON at 4 °C. 1 unit of Factor Xa (HTI technologies, Essex Junction, VT,
USA) were added per mg of protein generating the desired Cys-terminal fragment after ON incubation
at 4 °C. A reverse HisTrap and RP-HPLC purification was performed afterwards according to previous
described protocols.[16]
Ligation and refolding. The lyophilized synthetic peptide (2.0 mmol, 1.5 eq.) was dissolved in 300 μL
acidic buffer (6 M GnHCl, 0.2 M sodium phosphate, pH 3.0) and cooled to -10 °C. 60 uL of a 0.2 M
NaNO3 solution in acidic buffer (6 M GnHCl, 0.2 M sodium phosphate, pH 3.0) was added dropwise to
the peptide solution. After incubation for 20 min, the C-terminal hydrazide was fully converted and the
peptide was mixed with 10 mg (1.3 mmol) of the C-terminal protein fragment dissolved in 300 μL
ligation buffer (6 M GuHCl, 0.2 NaPi, 60 mM MPAA, 40 mM TCEP, pH 7.1). The ligation was
incubated at rt for 2 h and 1 M DTT was added to give a final DTT concentration of 10 mM. ÄKTA
19

ChemBioChem
10.1002/cbic.201600322
Explorer 100 Air (GE Healthcare) equipped with a Superdex HiLoad 16/600 75 pg (GE Healthcare) gel
filtration column equilibrated with 50 mM sodium phosphate, pH 7.4. The ligation mixture was
injected and the full-length PDZ domain collected after approximately 84 mL. The purified protein was
concentrated in Amicon Ultra-15 cutoff 10 kDa (EMD Millipore, Billerica, MA, USA) and
concentration was determined using absorbance at λ280 nm measured on a NanoDrop 1000 spectrometer
(Thermo Fisher Scientific, Waltham, MA, USA). Purity was determined by UPLC and SDS-PAGE gel.
Fluorescence polarization. Dilution series of the PDZ2 domains were prepared in 50 mM sodium
phosphate and ranged from 1-500 μM. TAMRA labeled peptides (400 nM) were mixed and diluted 1:1
with the different concentrations of PDZ2 in a black flat-bottom 396 well-plate (Corning Life Science,
Corning, NY, USA). 11 measurements were made for each binding curve using a Safire plate-reader
(Tecan) at excitation/emission wavelengths of λ545 nm/λ575 nm. Each experiment was performed in
triplicate. The concentration of the fluorescent probe was 200 nM in all the FP experiments. The
instrumental Z-factor was adjusted to obtain maximum signal and the G-factor calibrated from the
blank and background wells to give an initial milli-polarization (mP) of 20. The background signal was
subtracted from the obtained fluorescent polarization signal and fitted to one-sited binding model.
Circular dichroism. Circular dichroism (CD) experiments were performed using an Olis DSM 100
CD Spectrometer (Olis Inc., Bogart, GA, USA) in 1 mm quartz cuvettes (Starna Scientific Ltd., Essex,
England). CD spectra were obtained using protein samples at a concentration of 200 μg/mL and data
was collected in millidegrees of ellipticity (θ), which was converted into mean residue ellipticity
(ϴMRE) by the equation:
[ ]
Molar weight
[
]
[
]
[
]
Number of backbone amides
20

ChemBioChem
10.1002/cbic.201600322
where l is the pathway of the cuvette in mm and the MRW is the mean residue weight. The obtained
spectrum is an average of 3 measurements at each wavelength.
Acknowledgements
This research was financed by the Danish Council for Independent Research – Medical Sciences.
21

ChemBioChem
10.1002/cbic.201600322
Figures
Fig. 1. Structures and sequences of PDZ domains and introduced amino acids. (A) X-ray crystal
structures of PSD-95 PDZ2 (orange), PSD-95 PDZ3 (light blue) and syntrophin-PDZ (grey) domains
display a similar fold. The conserved basic residues coordinate a water molecule and form hydrogen
bonds to the C-terminal carboxylate of the ligand. (B) The three PDZ domains all contain a conserved
Lys/Arg residue (▼, bold) positioned in the β1-β2 loop. (C) Proteogenic and non-proteogenic amino
acids introduced instead of Lys165.
22

ChemBioChem
10.1002/cbic.201600322
Fig. 2. Semisynthesis of PDZ2 domains introducing Lys/Arg analogues. (A). Synthesis of the N-
terminal peptide fragment containing Lys/Arg analogues of PSD-95 PDZ2. The peptide thioester was
prepared using Fmoc-based SPPS. Hydrazine was loaded on a 2-chlorotrityl chloride resin, followed by
chain elongation and cleaved using TFA. The C-terminal hydrazide was converted into a thioester in
situ by oxidation with NaNO2 followed by thiolysis with 4-mercaptophenylacetic acid (MPAA). (B)
The recombinant C-terminal fragment was obtained by expression of a fusion protein containing a
histidine tag (His tag) and a FXa site (IEGRC). After expression in E. Coli, the protein was
enzymatically digested with FXa, yielding the ΔCPDZ2 fragment with a N-terminal Cys. (C) Ligation
of the synthetic peptide and recombinant protein fragments, ΔNPDZ2 and ΔCPDZ2 respectively, was
performed under denaturing conditions in 6 M GnHCl. After ligation, the semisynthetic PDZ2 was
refolded and purified by gel filtration (>95% pure). (D) LC-MS characterization of semisynthetic
PDZ2-Lys; chromatogram, mass trace and deconvoluted mass spectra. (E) 16% SDS-PAGE gel
displaying the purified expressed and semisynthetic PDZ2 domains. Left lane is the protein ladder
(Novex Sharp Pre-stained Protein Standard, Thermo). (F) CD spectra of the six Lys165 PDZ2 mutants
and the expressed PDZ2 WT, demonstrating highly similar secondary structures. The spectra were
recorded three times between 250–190 nm at 20 °C with 1 nm increments. Each curve is an average of
three measurements of the same sample.
23

ChemBioChem
10.1002/cbic.201600322
24

ChemBioChem
10.1002/cbic.201600322
Fig. 3. Ligand binding to recombinantly expressed and semisynthetic PDZ2 domains. Saturation
curves from FP assays for the expressed PDZ2 domain (upper left) and semisynthetic PDZ2 domain
(upper right). Increasing amounts of PDZ2 domains were titrated against TAMRA-labeled peptides; up
to 250 μM for the C-terminal peptide ligands and 25 μM for the cyclic nNOS β-finger. Assays were
performed in triplicates and KD values were determined from 50% of mPBmax
.
25

ChemBioChem
10.1002/cbic.201600322
Fig. 4. Effect on peptide binding of introducing Arg/Lys analogues into PSD-95 PDZ2. The fold
change was found by dividing the KD values for the four peptide ligands and the semisynthetic PDZ
domains by the KD value obtained from PDZ2-Lys. Affinity increase is displayed as left shifted fold
change whereas affinity decrease as right shifted fold change.
26

ChemBioChem
10.1002/cbic.201600322
References
References
[1]
[2]
[3]
[4]
M. R. Arkin, Y. Tang, J. A. Wells, Chem. Biol. 2014, 21, 1102-1114.
S. Gul, K. Hadian, Expert Opin. Drug Discovery 2014, 9, 1393-1404.
L. N. Makley, J. E. Gestwicki, Chem. Biol. Drug Design 2013, 81, 22-32.
H. Jubb, A. P. Higueruelo, A. Winter, T. L. Blundell, Trends Pharmacol. Sci. 2012, 33, 241-
248.
[5]
[6]
B. T. Seet, I. Dikic, M. M. Zhou, T. Pawson, Nat. Rev. Mol. Cell Biol. 2006, 7, 473-483.
S. Yellaboina, A. Tasneem, D. V. Zaykin, B. Raghavachari, R. Jothi, Nucleic Acids Res. 2011,
39, D730-735.
[7]
[8]
[9]
C. Y. Zheng, G. K. Seabold, M. Horak, R. S. Petralia, Neuroscientist 2011, 17, 493-512.
C. Oliva, P. Escobedo, C. Astorga, C. Molina, J. Sierralta, Develop. Neurobiol. 2012, 72, 57-72.
E. Kim, M. Sheng, Nat. Rev. Neurosci. 2004, 5, 771-781.
[10] W. Feng, M. Zhang, Nat. Rev. Neurosci. 2009, 10, 87-99.
[11] K. K. Dev, Nat. Rev. Drug Discov. 2004, 3, 1047-1056.
[12] C. N. Chi, A. Bach, K. Strømgaard, S. Gianni, P. Jemth, BioFactors 2012, 38, 338-348.
[13] Z. Khan, M. Lafon, Curr. Med. Chem. 2014, 21, 2632-2641.
[14] H. J. Lee, J. J. Zheng, Cell Commun. Signal. 2010, 8, 8.
[15] D. A. Doyle, A. Lee, J. Lewis, E. Kim, M. Sheng, R. MacKinnon, Cell 1996, 85, 1067-1076.
[16] S. W. Pedersen, S. B. Pedersen, L. Anker, G. Hultqvist, A. S. Kristensen, P. Jemth, K.
Strømgaard, Nat. Commun. 2014, 5, 3215.
[17] J. N. Eildal, G. Hultqvist, T. Balle, N. Stuhr-Hansen, S. Padrah, S. Gianni, K. Strømgaard, P.
Jemth, J. Am. Chem. Soc. 2013, 135, 12998-13007.
[18] M. A. Stiffler, J. R. Chen, V. P. Grantcharova, Y. Lei, D. Fuchs, J. E. Allen, L. A. Zaslavskaia,
G. MacBeath, Science 2007, 317, 364-369.
[19] R. Tonikian, Y. Zhang, S. L. Sazinsky, B. Currell, J. H. Yeh, B. Reva, H. A. Held, B. A.
Appleton, M. Evangelista, Y. Wu, X. Xin, A. C. Chan, S. Seshagiri, L. A. Lasky, C. Sander, C.
Boone, G. D. Bader, S. S. Sidhu, PLoS Biol. 2008, 6, e239.
[20] B. Z. Harris, B. J. Hillier, W. A. Lim, Biochemistry 2001, 40, 5921-5930.
[21] B. J. Hillier, K. S. Christopherson, K. E. Prehoda, D. S. Bredt, W. A. Lim, Science 1999, 284,
812-815.
[22] B. H. Chang, T. S. Gujral, E. S. Karp, R. BuKhalid, V. P. Grantcharova, G. MacBeath, Chem.
Biol. 2011, 18, 1143-1152.
[23] R. R. Penkert, H. M. DiVittorio, K. E. Prehoda, Nat. Struct. Mol. Biol. 2004, 11, 1122-1127.
[24] H. Tochio, Q. Zhang, P. Mandal, M. Li, M. Zhang, Nat. Struct. Mol. Biol. 1999, 6, 417-421.
[25] H. Tochio, Y. K. Mok, Q. Zhang, H. M. Kan, D. S. Bredt, M. Zhang, J. Mol. Biol. 2000, 303,
359-370.
[26] O. Kocher, G. Birrane, K. Tsukamoto, S. Fenske, A. Yesilaltay, R. Pal, K. Daniels, J. A. Ladias,
M. Krieger, J. Biol. Chem. 2010, 285, 34999-35010.
[27] C. N. Chi, A. Engström, S. Gianni, M. Larsson, P. Jemth, J. Biol. Chem. 2006, 281, 36811-
36818.
[28] S. H. Gee, S. Quenneville, C. R. Lombardo, J. Chabot, Biochemistry 2000, 39, 14638-14646.
27

ChemBioChem
10.1002/cbic.201600322
[29] A. Bach, B. H. Clausen, M. Møller, B. Vestergaard, C. N. Chi, A. Round, P. L. Sørensen, K. B.
Nissen, J. S. Kastrup, M. Gajhede, P. Jemth, A. S. Kristensen, P. Lundström, K. L. Lambertsen,
K. Strømgaard, Proc. Natl. Acad. Sci. U.S.A. 2012, 109, 3317-3322.
[30] M. D. Hill, R. H. Martin, D. Mikulis, J. H. Wong, F. L. Silver, K. G. Terbrugge, G. Milot, W.
M. Clark, R. L. Macdonald, M. E. Kelly, M. Boulton, I. Fleetwood, C. McDougall, T.
Gunnarsson, M. Chow, C. Lum, R. Dodd, J. Poublanc, T. Krings, A. M. Demchuk, M. Goyal,
R. Anderson, J. Bishop, D. Garman, M. Tymianski, ENACT trial investigators, Lancet. Neurol.
2012, 11, 942-950.
[31] S. Seedorff, C. Appelt, M. Beyermann, P. Schmieder, Biochem. Biophys. Res. Commun. 2010,
395, 535-539.
[32] C. Hoppmann, S. Seedorff, A. Richter, H. Fabian, P. Schmieder, K. Rück-Braun, M.
Beyermann, Angew. Chem. Int. Ed. 2009, 48, 6636-6639.
[33] G. M. Fang, Y. M. Li, F. Shen, Y. C. Huang, J. B. Li, Y. Lin, H. K. Cui, L. Liu, Angew. Chem.
Int. Ed. 2011, 50, 7645-7649.
[34] A. Bach, C. N. Chi, T. B. Olsen, S. W. Pedersen, M. U. Roder, G. F. Pang, R. P. Clausen, P.
Jemth, K. Strømgaard, J. Med. Chem. 2008. 51, 6450-6459.
[35] Y. Ivarsson, FEBS Lett. 2012, 586, 2638-2647.
[36] B. Z. Harris, F. W. Lau, N. Fujii, R. K. Guy, W. A. Lim, Biochemistry 2003, 42, 2797-2805.
[37] S. Sokalingam, G. Raghunathan, N. Soundrarajan, S. G. Lee, PLoS One 2012, 7, e40410.
[38] H. Cui, A. Hayashi, H. S. Sun, M. P. Belmares, C. Cobey, T. Phan, J. Schweizer, M. W. Salter,
Y. T. Wang, R. A. Tasker, D. Garman, J. Rabinowitz, P. S. Lu, M. Tymianski, J. Neurosci.
2007, 27, 9901-9915.
[39] K. Schuh, S. Uldrijan, M. Telkamp, N. Rothlein, L. Neyses, J. Cell Biol. 2001, 155, 201-205.
[40] G. M. Riefler, B. L. Firestein, J. Biol. Chem. 2001, 276, 48262-48268.
[41] B. L. Firestein, D. S. Bredt, J. Biol. Chem. 1999, 274, 10545-10550.
28

ChemBioChem
10.1002/cbic.201600322
Table of Contents
PDZ domains are mediators of numerous protein-protein interactions, and understanding the molecular
details these interactions are of fundamental importance. Here, we applied expressed protein ligation to
generate six semisynthetic PDZ domains containing proteogenic and non-proteogenic amino acids.
This allowed us to highlight the importance of a conserved Lys/Arg residue in PSD-95 PDZ2.
29