N-Methyl and peptoid scans of an autoinducing peptide reveal new structural features required for inhibition and activation of AgrC quorum sensing receptors in Staphylococcus aureus

Article abstract of DOI:10.1039/c4cc00117f

We report the first N-methyl and peptoid residue scans of a full-length autoinducing peptide (AIP), AIP-III, used by Staphylococcus aureus for quorum sensing (QS). Biological evaluation of these AIP-III analogues uncovered new features of the AIP-III scaffold that can be tuned to develop chemical probes of QS in all four groups of S. aureus (I-IV). The Royal Society of Chemistry.

Full text of DOI:10.1039/c4cc00117f

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
N-Methyl and peptoid scans of an autoinducing
peptide reveal new structural features required for
inhibition and activation of AgrC quorum sensing
receptors in Staphylococcus aureus†
Cite this: Chem. Commun., 2014,
50, 3000
Received 6th January 2014,
Accepted 29th January 2014
Yftah Tal-Gan, Danielle M. Stacy and Helen E. Blackwell*
DOI: 10.1039/c4cc00117f
www.rsc.org/chemcomm
We report the first N-methyl and peptoid residue scans of a full-length
autoinducing peptide (AIP), AIP-III, used by Staphylococcus aureus for
quorum sensing (QS). Biological evaluation of these AIP-III analogues
uncovered new features of the AIP-III scaffold that can be tuned to
develop chemical probes of QS in all four groups of S. aureus (I–IV).
Staphylococcus aureus is a notorious human pathogen that uses
quorum sensing (QS) to assess its local population density.1 Once
a threshold cell density has amassed, the bacteria can activate
group behaviours, including numerous virulence phenotypes,
necessary for host infections.2–5 S. aureus uses the accessory gene
regulator (agr) system for QS, which is regulated by an excreted
autoinducing peptide (AIP) signal and a transmembrane receptor
histidine kinase AgrC. At a sufficiently high cell density – and thus
extracellular AIP concentration – productive binding of the AIP
Fig. 1 Structure of AIP-III. Residues numbered for clarity.
to the extracellular sensor domain of AgrC activates QS signal presents an excellent peptide scaffold for the development of
transduction and alters the expression of group beneficial genes. QS modulators in S. aureus. We now seek to further explore the
Four unique AIP:AgrC pairs have been identified, leading to the AIP-III structure and delineate features that can be further tuned
classification of four S. aureus agr groups (I–IV).6,7
to engender either enhanced agonistic activity in AgrC-III or
As AIP:AgrC binding is essential for S. aureus QS, inhibition increased inhibitory activities against AgrCs-I–IV. We are also
of AIP:AgrC interactions has attracted considerable interest as an interested in developing non-native AIP analogues capable of
approach to attenuate QS, and thereby virulence, in S. aureus. selectively inhibiting one (or two) AgrC receptors, to examine the
Both non-native AIP analogues that displace AIPs and antibodies role of QS in mixed populations of S. aureus groups. As a single
that sequester AIPs away from their cognate AgrC receptor have S. aureus group is often prevalent in certain infection types,4,12 the
been pursued.8–10 Most past work has focused on AIP-I and facility of individual S. aureus groups to use QS as an interference
AIP-II. We recently began examining the AIP-III signal (Fig. 1) mechanism has been postulated. The ability of native AIPs to cross-
and determined key residues on its heptapeptide scaffold essential inhibit non-cognate AgrC receptors supports this hypothesis.4,13
for AgrC-III receptor interactions.11 In addition, we identified a Toward addressing these broad questions, we report herein our
set of AIP-III analogues that were extremely potent inhibitors of synthesis and biological evaluation of a set of new AIP-III analogues
all four AgrC receptors (I–IV) and were capable of completely generated through N-methyl amino acid and N-substituted glycine
blocking QS phenotypes in wild-type strains. These initial struc- (peptoid) scans of full-length AIP-III.
ture–activity relationship (SAR) studies suggested that AIP-III
Our prior SAR11 and structural studies14 of AIP-III and analogues
allowed us to develop a model by which AIP-III binds and activates
the AgrC-III receptor. First, the three hydrophobic residues within
the macrocycle form a triangular knob that is required for initial
AgrC receptor recognition (Phe5, Leu6, and Leu7; Fig. 1). Second,
the Asp4 side chain, as well as additional hydrophilic contacts,
serve to properly orient the exocyclic tail relative to the macrocycle.
Department of Chemistry, University of Wisconsin-Madison, 1101 University
Avenue, Madison, Wisconsin 53706-1322, USA. E-mail: blackwell@chem.wisc.edu;
Fax: +1 608 265-4534; Tel: +1 608 262-1503
† Electronic supplementary information (ESI) available: Full details of peptide
synthesis and characterization, biological assay protocols, dose response curves
for AIP-III analogues. See DOI: 10.1039/c4cc00117f
3000 | Chem. Commun., 2014, 50, 3000--3003
This journal is ©The Royal Society of Chemistry 2014

View Article Online
Communication
ChemComm
Third, the side chain of Ile1 on the tail, when properly oriented,
serves as a fourth hydrophobic anchor that is critical for AgrC-III
activation. Modification of any one of these three facets yielded
analogues that were incapable of AgrC-III activation, and
many exhibited competitive antagonistic activity instead (against
AgrCs-I–IV). Most notably, AIP-III with an Asp4 - Ala4 modifica-
tion (AIP-III D4A) is the most potent pan-group AgrC inhibitor
reported to date.11 Truncated analogues (lacking the tail) that
retained the hydrophobic knob were all AgrC inhibitors.
This past work demonstrated that the majority of the side chain
functionalities in AIP-III are essential for AgrC recognition. We
sought to explore if other structural elements could be altered in
AIP-III to modulate its activity (i.e., backbone amide NHs and side
chain positioning), and performed N-methyl (N-Me) and peptoid
(n) scans of its primary sequence. In each peptoid analogue, the
side chain of a specific residue is shifted from the C-a to the
amide. We previously showed that a peptoid–peptide hybrid of
truncated AIP-I was capable of modest QS inhibition in S. aureus;15
peptoid AIP mimics are yet to be investigated further. Likewise,
N-Me amide modifications remain largely unexamined in the
context of AIPs. In the one reported study to date, George et al.
performed an N-Me scan of the macrocycle of truncated AIP-II
Fig. 2 (A) Structure of AIP-III nN2 DKP. (B) Key SAR of AIP-III.
(tAIP-II) and discovered H-bonds essential for cognate and non- AIP-III nN2 DKP (Fig. 2A), which was also forwarded to bio-
cognate AgrC inhibition,16 underscoring the potential utility of logical analysis.
such studies for investigating AIP SAR.
We examined the ability of the AIP-III analogues to modulate
We used a straightforward Boc/Fmoc hybrid solid-phase the AgrC receptors (I–IV) in cell-based assays using two sets of
peptide synthesis protocol to construct the seven N-Me AIP-III S. aureus reporter strains. To measure AgrC antagonism, we
analogues and six peptoid AIP-III analogues (each with one used S. aureus strains (groups-I–IV) harbouring P3-gfp reporter
residue modifications, listed in Table 1; AIP-III nC3 not examined plasmids.17 In these strains, the AIP:AgrC complex activates
due to synthesis constraints). Peptide macrocyclization was AgrA (the intracellular response regulator), which then binds the
performed post-cleavage using our reported method,11 after P3 promoter and induces gfp transcription. Thus, competitive
which the peptides were purified to homogeneity by RP-HPLC AgrC inhibition by an exogenous peptide was quantified by
(see ESI† for full details). Interestingly, replacement of Asn2 measuring GFP fluorescence. For the AgrC agonism assay, we
with its peptoid counterpart resulted in almost quantitative used a set of S. aureus agr-null strains each carrying a P3-blaZ
formation of an amide bond between the amino terminus and reporter plasmid and agrCA from groups-I, -II, -III, or -IV.13,18 AgrC
the nAsn2 side chain to yield a diketopiperazine (DKP). This agonism was then quantified by measuring b-lactamase activity.
cyclic segment did not hinder the macrocyclization reaction,
The activities of the N-Me and peptoid AIP-III analogues in
resulting in clean formation of a bicyclic AIP-III analogue, AgrCs-I–IV are summarized in Table 1. We first scrutinized their
Table 1 Structures of the AIP-III analogues in this study and their IC50 values against AgrCs I–IV and EC50 values in AgrC-IIIa
b
b
b
b
b
Peptide/AIP-III
modification
AgrC-I IC50
(nM)
AgrC-II IC50
(nM)
AgrC-III IC50
(nM)
AgrC-IV IC50
(nM)
AgrC-III EC50
(nM)
Sequence
AIP-III
N-Me-I1
N-Me-N2
N-Me-C3
N-Me-D4
N-Me-F5
N-Me-L6
N-Me-L7
nI1
nN2 DKP
nD4
nF5
nL6
nL7
I-N-(C-D-F-L-L)
5.05
60.8
8.04
137
172
4.49
41000c
—
44.0
126
162
13.8
—
5.63
12.8
3.19
134
7.98
1.95
—
8.53
49.2
9.54
106
36.9
6.59
81.6
—
22.2
84.8
53.2
0.839
41000c
—
406
—
75.8
44000
—
198
—
—
—
—
—
d
NMeI-N-(C-D-F-L-L)
I-NMeN-(C-D-F-L-L)
I-N-(NMeC-D-F-L-L)
I-N-(C-NMeD-F-L-L)
I-N-(C-D-NMeF-L-L)
I-N-(C-D-F-NMeL-L)
I-N-(C-D-F-L-NMeL)
nI-N-(C-D-F-L-L)
(I-nN)-(C-D-F-L-L)
I-N-(C-nD-F-L-L)
I-N-(C-D-nF-L-L)
I-N-(C-D-F-nL-L)
I-N-(C-D-F-L-nL)
41000c
—
—
90.2
e
—
162
—
41000c
41000c
25.1
4.25
4.28
28.3
75.6
—
206
41000c
41000c
—
947
—
—
—
—
a
b
c
See ESI for details of reporter strains and assay procedures. See ESI for 95% confidence ranges. 100% inhibition was not observed over the
d
e
conc. tested. No activation observed over the conc. tested. No inhibition observed over the conc. tested.
This journal is ©The Royal Society of Chemistry 2014
Chem. Commun., 2014, 50, 3000--3003 | 3001

View Article Online
ChemComm
Communication
antagonism activity trends against non-cognate AgrC receptors primary amine at the AIP-III N-terminus is important for not
relative to native AIP-III. Modifications to Ile1 yielded analogues only non-cognate, but also cognate AgrC interactions. However,
that were B2–10-fold less active than AIP-III (with only one the ability of AIP-III nN2 DKP to effectively inhibit AgrC-III,
exception), revealing the importance of both the Ile side chain despite the incorporation of the amino terminus within the
position and an N-terminal primary amine in most non-cognate DKP scaffold, weakens this hypothesis.19 Indeed, AIP-III nN2
receptor interactions. However, the N-Me Asn2 analogue displayed DKP was the strongest AgrC-III inhibitor identified in this study
analogous inhibitory activity as AIP-III, suggesting this amide NH (IC50 = 25.1 nM). It is possible that a more basic, secondary
was not important. We were unable to directly evaluate AIP-III amine at the N-terminal Ile1 has a detrimental effect on receptor
nN2 (see above), but note that while AIP-III nN2 DKP significantly binding. Within the macrocycle, modification of Asp4 yielded
lost the ability to inhibit AgrC-I and -IV (420-fold and 10-fold moderate AgrC-III inhibitors, as expected in view of our prior
changes, respectively, relative to AIP-III), it displayed analogous studies.11 Modifications to either Leu6 or Leu7 resulted in
activity as AIP-III against AgrC-II.
largely inactive analogues, further emphasizing the importance
Turning to the AIP-III macrocycle, we found that N-Me of these two hydrophobic residues for AgrC interactions in
modifications to both Cys3 and Asp4, and peptoid modification general (Fig. 2B).
to Asp4, typically caused modest to large reductions in inhibitory
To streamline the AgrC agonism assays, we only evaluated
activity relative to AIP-III (B4- to 30-fold changes, Table 1), AIP-III analogues that displayed o100% antagonism in the GFP
indicating the significance of these two amide NHs and the reporter strains. None of these analogues were capable of
Asp4 side chain position for certain non-cognate receptor inter- activating non-cognate AgrC receptors (see ESI†); however, four
actions. These effects were the most apparent for AgrC-I, and the analogues were capable of AgrC-III activation (Table 1), and we
least apparent for AgrC-II (with AIP-III N-Me-D4 displaying almost analyse these data here. Most notably, the N-Me Asn2 analogue
comparable activity as AIP-III against this receptor). The results was B4-fold more active than AIP-III (EC50 = 75.8 nM). This result,
for Phe5 were perhaps more interesting. We previously observed combined with the antagonistic activity observed for AIP-III nN2 DKP
that converting Phe5 to either D-Phe or Ala in AIP-III yielded a against AgrC-III, suggests that N-alkylation of Asn2 is highly bene-
weakly active, non-cognate AgrC inhibitor;11 however, in the ficial for AgrC-III binding. In turn, N-methylation of Cys3 resulted in
current study we found that Phe5 could be replaced with either a 410-fold reduction in agonistic activity relative to AIP-III, high-
its N-Me or peptoid counterpart with virtually no change in lighting again the significance of this amide NH for general AgrC
antagonistic activity. In fact, AIP-III nF5 was the most potent receptor recognition. The N-Me and peptoid Phe5 analogues again
inhibitor identified in this study (IC50 = 839 pM), with a 10-fold yielded interesting results – both were able to activate AgrC-III as well
increased potency against AgrC-IV relative to AIP-III. These as native AIP-III. This outcome strengthens our hypothesis about the
results are notable, as they suggest that the nPhe side chain minimal structural requirement for AIP-III:AgrC receptor binding,
can mimic that of L-Phe in the native AIP-III.
and extends it to now include AgrC-III (Fig. 2B).
In contrast to Phe5, the amide NHs and side chain position-
Lastly, we examined the selectivities of the AgrC inhibitors
ing of Leu6 and Leu7 in AIP-III were critical for non-cognate identified herein. Few receptor selective AgrC inhibitors are
AgrC inhibition, as N-Me or peptoid modification of these known; prior work has uncovered four AgrC-III selective inhibitors
C-terminal residues virtually abolished activity. George et al. and one AgrC-II selective inhibitor (selectivity defined as 415-fold
observed analogous results for tAIP-II when either of its two stronger activity in one group relative to others).11,13 We
C-terminal residues (Leu4 and Phe5) were N-methylated.16 discovered the first AgrC-IV selective inhibitor in this study,
These data also mesh with our earlier study, which showed AIP-III nF5 (Table 1). As the sensor domains of AgrC-I and AgrC-IV
that conversion of either Leu6 or Leu7 to Ala obliterated the have nearly 90% sequence homology,5 the selectivity observed for
resulting AIP-III analogue’s inhibitory activity.11 Interestingly, AIP-III nF5 is significant, and reinforces the view that very subtle
while the AIP-III D-Leu6 analogue was also inactive as a cross- changes in the AIP-III structure can strongly tune its interactions
receptor antagonist, the D-Leu7 analogue displayed analogous with AgrC receptors.
antagonistic activity as the native AIP-III.11 Together, these
In summary, we report the first systematic N-Me and peptoid
results indicate that the side chain of Phe5 (in the proper scans of a full-length AIP from S. aureus, AIP-III. Evaluation of
orientation), but not its amide NH, is a crucial component for these new analogues in AgrC antagonism and agonism assays
non-cognate AgrC inhibition, whereas both the amide and the revealed key H-bond and side chain interactions that are critical for
side chain are the important contributions at Leu6 and Leu7 AgrC-I–IV receptor inhibition and AgrC-III activation. These SAR
(with the orientation of the Leu7 side chain less critical relative data allowed us to define a minimal structural requirement for
to Leu6). We propose, based on these results, that the AIP-III AIP-III:AgrC interactions. In addition, we identified the first
segment spanning from the Phe5 a-carbon to the Leu7 a-carbon group-IV selective AgrC inhibitor, by replacing the AIP-III Phe5
(orientation controlled by the macrocycle) serves as a minimal residue with its peptoid counterpart. The analogues reported herein
requirement for non-cognate AgrC binding (Fig. 2B).
provide further insights into the mechanisms of AgrC activation
We next analysed the AgrC-III antagonism assay data to and inhibition by AIPs, and constitute new – and potentially more
deduce any new structural features critical for AIP-III to bind biostable – tools to study QS in S. aureus.
(but not activate) its cognate receptor (Table 1). Modifications of
This work was supported by the Office of Naval Research
Ile1 resulted in relatively inactive analogues, suggesting that a (N00014-07-1-0255) and the Burroughs Welcome Fund. We thank
3002 | Chem. Commun., 2014, 50, 3000--3003
This journal is ©The Royal Society of Chemistry 2014

View Article Online
Communication
ChemComm
10 N. Amara, B. P. Krom, G. F. Kaufmann and M. M. Meijler, Chem.
Rev., 2011, 111, 195–208.
11 Y. Tal-Gan, D. M. Stacy, M. K. Foegen, D. W. Koenig and
H. E. Blackwell, J. Am. Chem. Soc., 2013, 135, 7869–7882.
12 V. Fleming, E. Feil, A. K. Sewell, N. Day, A. Buckling and
R. C. Massey, J. Bacteriol., 2006, 188, 7686–7688.
Profs. Alexander Horswill and Richard Novick for S. aureus
strains, and Dr. Noam Freeman for valuable discussions.
Notes and references
1 H. F. Chambers and F. R. DeLeo, Nat. Rev. Microbiol., 2009, 7, 629–641.
2 E. A. George and T. W. Muir, ChemBioChem, 2007, 8, 847–855.
13 G. J. Lyon, J. S. Wright, T. W. Muir and R. P. Novick, Biochemistry,
2002, 41, 10095–10104.
3 J. R. Li, W. L. Wang, S. X. Xu, N. A. Magarvey and J. K. McCormick, 14 Y. Tal-Gan, M. Ivancic, G. Cornilescu, C. C. Cornilescu and
Proc. Natl. Acad. Sci. U. S. A., 2011, 108, 3360–3365.
4 G. Ji, R. Beavis and R. P. Novick, Science, 1997, 276, 2027–2030.
5 R. P. Novick and E. Geisinger, Annu. Rev. Genet., 2008, 42, 541–564.
H. E. Blackwell, J. Am. Chem. Soc., 2013, 135, 18436–18444.
15 S. A. Fowler, D. M. Stacy and H. E. Blackwell, Org. Lett., 2008, 10,
2329–2332.
6 P. McDowell, Z. Affas, C. Reynolds, M. T. Holden, S. J. Wood, 16 E. A. George, R. P. Novick and T. W. Muir, J. Am. Chem. Soc., 2008,
S. Saint, A. Cockayne, P. J. Hill, C. E. Dodd, B. W. Bycroft,
W. C. Chan and P. Williams, Mol. Microbiol., 2001, 41, 503–512.
7 S. Jarraud, G. J. Lyon, A. M. Figueiredo, L. Gerard, F. Vandenesch,
J. Etienne, T. W. Muir and R. P. Novick, J. Bacteriol., 2000, 182, 6517–6522.
130, 4914–4924.
17 R. N. Kirchdoerfer, A. L. Garner, C. E. Flack, J. M. Mee,
A. R. Horswill, K. D. Janda, G. F. Kaufmann and I. A. Wilson,
J. Biol. Chem., 2011, 286, 17351–17358.
8 M. Thoendel, J. S. Kavanaugh, C. E. Flack and A. R. Horswill, Chem. 18 G. J. Lyon, P. Mayville, T. W. Muir and R. P. Novick, Proc. Natl. Acad.
Rev., 2011, 111, 117–151. Sci. U. S. A., 2000, 97, 13330–13335.
9 C. P. Gordon, P. Williams and W. C. Chan, J. Med. Chem., 2013, 56, 19 We note that DKPs have previously been shown to weakly inhibit the
1389–1404.
agr system in group-III S. aureus. See ref. 3 and 11.
This journal is ©The Royal Society of Chemistry 2014
Chem. Commun., 2014, 50, 3000--3003 | 3003