Sulfonyl 3-alkynyl pantetheinamides as mechanism-based cross-linkers of acyl carrier protein dehydratase

Article abstract of DOI:10.1021/ja4042059

Acyl carrier proteins (ACPs) play a central role in acetate biosynthetic pathways, serving as tethers for substrates and growing intermediates. Activity and structural studies have highlighted the complexities of this role, and the protein-protein interactions of ACPs have recently come under scrutiny as a regulator of catalysis. As existing methods to interrogate these interactions have fallen short, we have sought to develop new tools to aid their study. Here we describe the design, synthesis, and application of pantetheinamides that can cross-link ACPs with catalytic β-hydroxy-ACP dehydratase (DH) domains by means of a 3-alkynyl sulfone warhead. We demonstrate this process by application to the Escherichia coli fatty acid synthase and apply it to probe protein-protein interactions with noncognate carrier proteins. Finally, we use solution-phase protein NMR spectroscopy to demonstrate that sulfonyl 3-alkynyl pantetheinamide is fully sequestered by the ACP, indicating that the crypto-ACP closely mimics the natural DH substrate. This cross-linking technology offers immediate potential to lock these biosynthetic enzymes in their native binding states by providing access to mechanistically cross-linked enzyme complexes, presenting a solution to ongoing structural challenges.

Full text of DOI:10.1021/ja4042059

Communication
pubs.acs.org/JACS
Sulfonyl 3‑Alkynyl Pantetheinamides as Mechanism-Based Cross-
Linkers of Acyl Carrier Protein Dehydratase
Fumihiro Ishikawa, Robert W. Haushalter, D. John Lee, Kara Finzel, and Michael D. Burkart*
Department of Chemistry and Biochemistry, University of California, San Diego, 9500 Gilman Drive, La Jolla, California 92093-0358,
United States
S
* Supporting Information
participation in these systems remains largely uncharacterized,
ABSTRACT: Acyl carrier proteins (ACPs) play a central
role in acetate biosynthetic pathways, serving as tethers for
substrates and growing intermediates. Activity and
structural studies have highlighted the complexities of
this role, and the protein−protein interactions of ACPs
have recently come under scrutiny as a regulator of
catalysis. As existing methods to interrogate these
interactions have fallen short, we have sought to develop
new tools to aid their study. Here we describe the design,
synthesis, and application of pantetheinamides that can
cross-link ACPs with catalytic β-hydroxy-ACP dehydratase
(DH) domains by means of a 3-alkynyl sulfone warhead.
We demonstrate this process by application to the
Escherichia coli fatty acid synthase and apply it to probe
protein−protein interactions with noncognate carrier
proteins. Finally, we use solution-phase protein NMR
spectroscopy to demonstrate that sulfonyl 3-alkynyl
pantetheinamide is fully sequestered by the ACP,
indicating that the crypto-ACP closely mimics the natural
DH substrate. This cross-linking technology offers
immediate potential to lock these biosynthetic enzymes
in their native binding states by providing access to
mechanistically cross-linked enzyme complexes, presenting
a solution to ongoing structural challenges.
and those systems that have been studied structurally show
disordered ACPs.6 To overcome the kinetic disadvantage of
diffusion control, type-II enzymes are thought to associate in
dynamic complexes, the compositions of which can change
during the course of a particular biosynthetic pathway.7 To serve
as a covalent tether, an ACP must be post-translationally
modified with a 4′-phosphopantetheine (PPant) prosthetic arm.
It is in this covalent, protein-bound form that an ACP reacts with
the active site of a catalytic partner protein.8 The ACP must
therefore functionally interact with all enzyme domains
responsible for loading carbon units, condensing these units,
modifying the condensation product, and releasing the final
product from the synthase. Each step relies on specific protein−
protein interactions to recognize and channel the biosynthetic
intermediate to the proper partner enzyme.
The binding affinities of an ACP for its enzymatic partners are
relatively weak (Kd ∼ 1−10 μM), allowing it to interact reversibly
with each enzyme in the biosynthetic pathway and move on to
the next. Therefore, attempts to visualize ACPs in complex with
partner enzymes for structural analysis have been met with only
modest success. To overcome this obstacle, we have devised a
general methodology for covalent cross-linking of ACPs to
partner domains by taking advantage of the unique post-
translational modification of ACPs (Scheme 1).8 In this
approach, we synthesize analogues of pantetheine that contain
mechanism-based inhibitors of FAS enzymes and append them
to ACPs. When an inhibitor-loaded crypto-ACP interacts with its
enzymatic partner through protein−protein interactions
(Scheme 1a), the synthetic PPant side arm is allowed to enter
the pocket of the catalytic protein (Scheme 1b), where it can
react with an active-site residue and form a covalent bond that
traps both the ACP and the enzyme in a bound conformation
(Scheme 1c). The formation of the cross-linked species, which
was initially applied to the study of ketosynthase domains,8b is
highly dependent on ACP-mediated protein−protein interac-
tions and has been applied as a tool in elucidating ACP−partner
protein interactions. Similar strategies have also been successfully
applied to probe peptidyl carrier protein interactions within
nonribosomal peptide synthetase enzymes.9
cyl carrier proteins (ACPs) are highly conserved, small (∼9
AkDa), acidic proteins found in fatty acid synthase (FAS) and
polyketide synthase (PKS) biosynthetic pathways.1 The ACP
serves as a central tether for all starting materials and
intermediates in these pathways, mediating each chemical
transformation in the elongation and elaboration of the
metabolites. As the structures and activities of FAS and PKS
enzymes continue to be refined, the role of the ACP has been
subjected to further scrutiny.2 Whether incorporated into a large
megasynthase (type-I systems) or as discrete enzymes
functioning independently (type-II systems), ACPs are highly
flexible species that must interact with multiple catalytic enzyme
partners through highly specific protein−protein interactions.3
Complicating such binding events is the ability of type-II ACPs to
sequester their elongating intermediates, shielding them from
reactivity until favorable protein interactions are achieved.4
The need for new technologies to capture ACP dynamics and
protein−protein interactions has become apparent in recent
years.5 Because of the inherent technical issues posed by the
cloning and expression of large, multidomain synthases, ACP
In the present study, we extended the repertoire of
chemoenzymatic ACP cross-linking tools to study the
interactions of ACPs and dehydratase (DH) domains in FAS.
β-Hydroxy-ACP DH enzymes are found in all FAS pathways and
Received: April 27, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja4042059 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
was attributed to rapid release of the reactive functionality from
the crypto-ACP by nucleophilic cleavage of the thioester bond in
2 before the formation of the ACP−DH complex or to the
decomposition of 2 itself. We therefore began investigating
potential replacements of this warhead for increased stability.
To this end, we recently reported the design, synthesis, and
utility of DH-specific probe 3 containing a sulfonyl 3-alkyne
reactive warhead (Figure 1).13 Replacement of the reactive
thioester moiety of 1 with the nonhydrolyzable sulfone in 3 not
only eliminated the instability of nucleophilic cleavage but also
provided an α-proton with low enough acidity to avoid
nonenzymatic allene formation while still facilitating DH-specific
labeling. In addition, our previous studies suggested that
mechanism-based inhibition of FabA by 3-decynoic acid
analogues is dependent upon the pKa of the α-proton of the
suicide substrate scaffold (thioester > oxoester ≫ acid/amide),13
which led us to focus on the use of the sulfonyl 3-alkyne scaffold
to obtain second-generation ACP−DH cross-linking reagents.
To this end, pantetheine analogues 4a and 4b were synthesized
from the convergent assembly of commercially available 2-
nonyn-1-ol, cysteamine for 4a or 3-chloropropylamine for 4b,
and a p-methoxybenzyl-protected pantothenic acid (synthetic
methods are provided in the Supporting Information).
Our biochemical studies began with an evaluation of the ability
of 4a and 4b to modify the E. coli FAS ACP (AcpP). Using the
CoA biosynthetic enzymes CoaA, CoaD, and CoaE along with
4′-phosphopantetheinyl transferase (Sfp), we were able to
confirm loading of the analogues 4a and 4b onto the apo-ACP
by means of conformationally sensitive urea polyacrylamide gel
electrophoresis (urea-PAGE) analysis.14 A characteristic increase
in electrophoretic mobility was observed upon AcpP loading of
fatty acid pantetheines 4a and 4b to give the crypto-ACPs (Figure
2a). The conversion of the apo-ACP to the crypto-ACPs14b was
accomplished in quantitative yield at 37 °C for 1 h. Upon
addition of FabA to the corresponding crypto-ACPs obtained
using 4a and 4b, clear bands indicating irreversible covalent
cross-linking between the ACPs and the DH domain were
observed after 12 and 24 h at 37 °C. These products resulted in
observed gel shifts of the FabA band from ∼20 kDa for FabA to
∼45 kDa for the cross-linked AcpP−FabA complex upon analysis
by sodium dodecyl sulfate PAGE (SDS-PAGE). The observation
of FabA consumption at 12 and 24 h (Figure 2b) clearly showed
that 4a efficiently facilitated cross-linking between the crypto-
AcpP and FabA, suggesting that the longer alkyl linker more
closely mimicks the native substrate (R)-3-hydroxydecanolyl-
ACP and correctly aligns a His residue with the α-proton of the
sulfonyl 3-alkyne in the FabA active site. A slight gel shift due to
nonspecific probe labeling during SDS-PAGE sample prepara-
tion was seen in all bands within probe-containing lanes.
Subsequent screening allowed us to investigate the pH
dependence of the cross-linking process (Figure 2c). While the
maximum labeling of probe 3 was observed at pH 8.0 or 8.5,13 the
cross-linking reaction to form 7 (Scheme 2) was unaffected by
pH. As an ultimate test of the cross-linking efficiency, we
conducted time-course studies using 4a as the ACP−DH cross-
linking reagent at 37 °C and pH 7.0 (Figure 2d). This study
indicated that the improved stability of 4a allowed successful
cross-linking between the crypto-AcpP and FabA. Using
pantetheinamide 4a, we observed greatly improved cross-linking
efficiency in yields greater than 90%, as estimated by the band
intensity after 36 h at pH 7.0.
Scheme 1. General Representation of the Cross-Linking
Strategy for a FAS Dehydratase (DH) and Its ACP Loaded
with a Cross-Linking Pantetheine Probe: (a) Protein−Protein
Interaction; (b) Switchblade Mechanism; (c) Mechanism-
Based Inactivation and Cross-Linking
in any PKS whose product contains olefinic or saturated
methylene units. The enzyme FabA catalyzes two kinds of
chemical reactions in the Escherichia coli fatty acid biosynthetic
pathway: dehydration of (R)-3-hydroxydecanoyl-ACP to give
(E)-2-decenoyl-ACP and isomerization of (E)-2-decenoyl-ACP
to give (Z)-3-decenoyl-ACP.10 3-Decynoyl-N-acetylcysteamine
(1) (Figure 1), which was first reported by Bloch to inactivate
Figure 1. Structures of DH probes and cross-linking agents.
FabA rapidly and irreversibly,11 is the first instance of a
“mechanism-based inhibitor”. Here the enzyme target catalyzes
the formation of a covalent bond between the inhibitor and an
active-site residue, thus permanently inactivating the enzyme. In
this work, we expanded the use of pantetheinamides by installing
a sulfonyl 3-alkyne reactive functionality to facilitate specific
cross-linking of the ACP and DH domains.
Our studies originally began with the preparation of cross-
linker 2 to expand the use of this historical inhibitor scaffold and
guide specific cross-linking of ACP and DH domains.12
However, when this pantetheine analogue was investigated,
only low levels of cross-linking were observed. This observation
On the basis of the precise stereochemical, mechanistic, and
structural studies of FabA inactivation by 1,10 we propose that 4a
B
dx.doi.org/10.1021/ja4042059 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
Scheme 2. Schematic Representation of the Cross-Linking of
ACP and DH Domains Using Probe 4a
Figure 2. Evaluation of ACP−DH cross-linking reagents 4a and 4b. (a)
Pantetheinamides 4a and 4b were appended enzymatically to apo-AcpP
using a one-pot chemoenzymatic method. The conversion of E. coli apo-
AcpP to the corresponding crypto-ACP when treated with 500 μM 4a or
4b was visualized by urea-PAGE (a = apo, h = holo, c = crypto). (b) SDS-
PAGE analysis of the cross-linking of AcpP and FabA upon treatment
with 500 μM 4a or 4b at 37 °C for 12 or 24 h. (c) SDS-PAGE gel
showing the pH dependence of the cross-linking reaction for 12 h at 37
°C. (d) Time-course studies depicting the cross-linking of AcpP to FabA
with 500 μM 4a. The cross-linking reaction of 4a loaded onto the ACP
was accomplished in high yield (>90%) after 36 h at 37 °C. (e, f) Effect
of carrier protein (CP) identity on FabA cross-linking. Pantetheinamide
4a was used to modify (e) the FAS and PKS CPs AcpP (FAS), Fren
(PKS), Act (PKS), and Otc (PKS) and (f) the NRPS CPs VibB and
EntB. The results indicate the clear preferential interaction of FabA with
AcpP from FAS over the other ACPs from PKS and the peptidyl CPs
from NRPS.
and 4b loaded onto ACP act via the mechanism shown Scheme 2.
When the DH and crypto-ACP interact (5), the reactive 3-
decynoyl unit is positioned precisely in the FabA active site via
protein−protein interactions, allowing the active-site residue
His70 of FabA to deprotonate the acidic α-sulfone proton,
generating electrophilic allene 6. This in turn reacts with the
His70 residue to form AcpP−FabA cross-linked products 7 and 8.
Next, we tested the ability of the ACP−DH cross-linking
reaction to visualize protein−protein interactions between native
and non-native ACP·DH pairs.8 In addition to the FAS carrier
protein AcpP, the type-II PKS carrier proteins FrenACP (from
Streptomyces roseofulvus), ActACP (from Streptomyces colicolor),
and OtcACP (from Streptomyces rimosus) as well as the NRPS
carrier proteins VibB (from Vibrio cholerae) and EntB (from E.
coli) were loaded with 4a and incubated with FabA.8 Cross-
linking with FabA was seen only with FrenACP, ActACP, and
OtcACP, while VibB and EntB remained un-cross-linked (Figure
2e,f). A clear preference for E.coli AcpP could be seen by
comparison of the intensities of the cross-linked bands. This
result reflects the sequence and activity-based homology between
type-II FAS ACPs and type-II PKS ACPs.15 Indeed, DH-specific
fluorescent probes based on this sulfonyl alkyne scaffold labeled
DH enzymes from recombinant type-I and type-II FAS and PKS
systems.13 The lack of cross-linking between FabA and the NRPS
ACPs indicates a distinct specificity in protein−protein
interactions that are not satisfied with these pairs. These
experiments indicate that, as seen with ACP−KS cross-linking,7
the formation of ACP·DH complexes is driven by the native
protein−protein interactivity within the ACP·DH pairing.
Finally, we wished to understand the relevance of studying
pantetheinamide probe 4a as a true surrogate of ACP−DH
interactions. The suitability of this probe to mimic sequestration
of the natural substrate within the hydrophobic cavity of E. coli
AcpP was determined by solution-phase protein NMR spectros-
copy. We linked 4a to 15N-labeled AcpP, purified the crypto-
AcpP, and collected the heteronuclear single-quantum coherence
(HSQC) spectrum. A comparison with the holo-AcpP spectrum
under identical conditions is shown in the HSQC overlay and
corresponding chemical shift perturbation plot in Figure 3. In
particular, major perturbations of residues Asp35, Thr39, and
Glu47 within helix II and Ala59 and Glu60 in helix III are indicative
of substrate sequestration, as previously demonstrated for AcpP
and other ACPs within type-II FAS and PKS pathways.2d,4,16 Of
particular interest are the large perturbations of Thr63 and Tyr71
in helix IV, which display interactions with the sulfonyl 3-alkynyl
moiety. These results demonstrate that association with FabA
allows the release of the probe from the ACP binding pocket into
the active site of the enzyme and indicate that comparison of the
ACP in the free and cross-linked states could shed light on the
molecular details of this “switchblade” mechanism.
In summary, we have demonstrated new tools for the study of
protein−protein interactions between ACPs and DHs in FAS
and PKS based on the development of a pantetheinamide with a
C
dx.doi.org/10.1021/ja4042059 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
ACKNOWLEDGMENTS
■
This work was funded by NIH R01GM094924 and
R01GM095970. We thank Drs. X. Huang and A. Mrse for
NMR, Dr. Y. Su for MS, and Dr. J. J. La Clair for figure assistance.
REFERENCES
■
(1) Mercer, A. C.; Burkart, M. D. Nat. Prod. Rep. 2007, 24, 750.
(2) (a) Smith, S.; Tsai, S. C. Nat. Prod. Rep. 2007, 24, 1041. (b) Khosla,
C.; Tang, Y.; Chen, A. Y.; Schnarr, N. A.; Cane, D. E. Annu. Rev. Biochem.
2007, 76, 195. (c) Leibundgut, M.; Maier, T.; Jenni, S.; Ban, N. Curr.
Opin. Struct. Biol. 2008, 18, 714. (d) Haushalter, R. W.; Filipp, F. V.; Ko,
K. S.; Yu, R.; Opella, S. J.; Burkart, M. D. ACS Chem. Biol. 2011, 6, 413.
(3) (a) Parris, K. D.; Lin, L.; Tam, A.; Mathew, R.; Hixon, J.; Stahl, M.;
Fritz, C. C.; Seehra, J.; Somers, W. S. Structure 2000, 8, 883. (b) Qiu, X.;
Janson, C. A. Acta Crystallogr. 2004, D60, 1545.
́
(4) Płoskon, E.; Arthur, C. J.; Kanari, A. L.; Wattana-amorn, P.;
Williams, C.; Crosby, J.; Simpson, T. J.; Willis, C. L.; Crump, M. P. Chem
Biol. 2010, 17, 776.
(5) (a) Chan, D. I.; Vogel, H. J. Biochem. J. 2010, 430, 1. (b) Crosby, J.;
Crump, M. P. Nat. Prod. Rep. 2012, 29, 1111.
(6) (a) Jenni, S.; Leibundgut, M.; Maier, T.; Ban, N. Science 2006, 311,
1263. (b) Maier, T.; Jenni, S.; Ban, N. Science 2006, 311, 1258.
(7) Veyron-Churlet, R.; Bigot, S.; Guerrini, O.; Verdoux, S.; Malaga,
W.; Daffe’, M.; Zerbib, D. J. Mol. Biol. 2005, 353, 847.
(8) (a) Worthington, A. S.; Burkart, M. D. Org. Biomol. Chem. 2006, 4,
44. (b) Worthington, A. S.; Rivera, H.; Torpey, J. W.; Alexander, M. D.;
Burkart, M. D. ACS Chem. Biol. 2006, 1, 687. (c) Kapur, S.;
Worthington, A. S.; Tang, Y.; Cane, D. E.; Burkart, M. D.; Khosla, C.
Bioorg. Med. Chem. Lett. 2008, 18, 3034. (d) Worthington, A. S.; Hur, G.
H.; Meier, J. L.; Cheng, Q.; Moore, B. S.; Burkart, M. D. ChemBioChem
2008, 9, 2096.
(9) (a) Qiao, C.; Wilson, D. J.; Bennett, E. M.; Aldrich, C. C. J. Am.
Chem. Soc. 2007, 129, 6350. (b) Liu, Y.; Zheng, T.; Bruner, S. D. Chem.
Biol. 2011, 18, 1482. (c) Sundlov, J. A.; Shi, C.; Wilson, D. J.; Aldrich, C.
C.; Gulick, A. M. Chem. Biol. 2012, 19, 188. (d) Mitchell, C. A.; Shi, C.;
Aldrich, C. C.; Gulick, A. M. Biochemistry 2012, 51, 3252.
(10) (a) Leesong, M.; Henderson, B. S.; Gilling, J. R.; Schwab, J. M.;
Smith, J. L. Structure 1996, 4, 253. (b) Schwab, J. M.; Li, W. B.; Ho, C. K.;
Townsend, C. A.; Salituro, G. M. J. Am. Chem. Soc. 1984, 106, 7293.
(11) (a) Brock, D. J.; Kass, L. R.; Bloch, K. J. J. J. Biol. Chem. 1967, 242,
4432. (b) Helmkamp, G. M., Jr.; Brock, D. J.; Bloch, K. J. J. J. Biol. Chem.
1968, 243, 3229. (c) Kass, L. R. J. Biol. Chem. 1968, 243, 3223. (d) Endo,
K.; Helmkamp, D. M.; Bloch, K. J. Biol. Chem. 1970, 245, 4293.
(12) Meier, J. L.; Haushalter, R. W.; Burkart, M. D. Bioorg. Med. Chem.
Lett. 2010, 20, 4936.
(13) Ishikawa, F.; Haushalter, R. W.; Burkart, M. D. J. Am. Chem. Soc.
2012, 134, 769.
(14) (a) Rock, C. O.; Jackowski, S. J. Biol. Chem. 1982, 257, 10759.
(b) La Clair, J. J.; Foley, T. L.; Schegg, T. R.; Regan, C. M.; Burkart, M.
D. Chem. Biol. 2004, 11, 195.
(15) (a) Revill, W. P.; Bibb, M. J.; Hopwood, D. A. J. Bacteriol. 1996,
178, 5660. (b) Crosby, J.; Byrom, K. J.; Hitchman, T. S.; Cox, R. J.;
Crump, M. P.; Findlow, I. S.; Bibb, M. J.; Simpson, T. J. FEBS Lett. 1998,
433, 132. (c) Arthur, C.; Cox, R. J.; Crosby, J.; Rahman, M. M.; Simpson,
T. J.; Soulas, F.; Spogli, R.; Azafranska, A. E.; Westcott, J.; Winfield, C. J.
ChemBioChem 2002, 3, 253.
Figure 3. Solution-phase NMR spectra of crypto-ACP demonstrating
probe sequestration. (a) 1H,15N-HSQC spectra of purified crypto-ACP
containing 15N-enriched E. coli apo-AcpP16b modified chemoenzymati-
cally with 4a (blue) and holo-AcpP (green). Significant shifts are
annotated by residue number. (b) Structure of AcpP highlighting the
residues that showed the largest chemical shift perturbations. (c)
Chemical shift perturbation plot of crypto-AcpP vs holo-AcpP.
sulfonyl alkyne scaffold. The synthesis of this probe is
straightforward, offering easy access to synthetic modulation.
The probe itself is stable toward one-pot chemoenzymatic
methodology, is sequestered within the hydrophobic cleft of
crypto-AcpP, and forms the designed AcpP−FabA cross-linked
product in high yield. This scaffold could offer optimized
pantetheine analogues for use in investigating protein−protein
interactions for type-I and type-II PKS biosynthetic pathways.
This sulfonyl 3-alkynyl pantetheinamide provides a general
approach for elucidating interactions of ACPs with other DH-like
enzymes, such as the product template (PT) domain in fungal
nonreducing PKSs17 and the thioester hydrolase (TH) domain
in fungal type-I iterative PKSs.18
(16) (a) Kosa, N. M.; Haushalter, R. W.; Smith, A. R.; Burkart, M. D.
Nat. Methods 2012, 9, 981. (b) Zornetzer, G. A.; Fox, B. G.; Markley, J. L.
Biochemistry. 2006, 45, 5217.
(17) Crawford, J. M.; Korman, T. P.; Labonte, J. W.; Vagstad, A. L.;
Hill, E. A.; Kamari-Bidkorpeh, O.; Tsai, S. C.; Townsend, C. A. Nature
2009, 461, 1139.
ASSOCIATED CONTENT
* Supporting Information
Experimental details and additional data. This material is
available free of charge via the Internet at http://pubs.acs.org.
■
S
(18) (a) Moriguchi, T.; Kezuka, Y.; Nonaka, T.; Ebizuka, Y.; Fujii, I. J.
Biol. Chem. 2010, 285, 15637. (b) Scott, A. I.; Beadling, L. C.;
Georgopapadakou, N. H.; Subbarayan, C. R. Bioorg. Chem. 1974, 3, 238.
AUTHOR INFORMATION
Corresponding Author
mburkart@ucsd.edu
■
Notes
The authors declare no competing financial interest.
D
dx.doi.org/10.1021/ja4042059 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX