Facile synthesis of tetrafluorotyrosine and its application in pH triggered membrane lysis

Article abstract of DOI:10.1021/ol102610q

"Smart" peptides that lyse membranes in response to biological stimuli are of considerable interest. Herein, we report a facile synthesis of l-tetrafluorotyrosine (l-f₄Y) and its utilization as a pH sensing element. The synthetic route affords

Full text of DOI:10.1021/ol102610q

ORGANIC
LETTERS
2011
Vol. 13, No. 2
236-239
Facile Synthesis of Tetrafluorotyrosine
and Its Application in pH Triggered
Membrane Lysis
Fang Wang,† Luoheng Qin,† Patrick Wong, and Jianmin Gao*
Department of Chemistry, Merkert Chemistry Center, Boston College, Chestnut Hill,
Massachusetts 02467-3801, United States
jianmin.gao@bc.edu
Received November 1, 2010
ABSTRACT
“Smart” peptides that lyse membranes in response to biological stimuli are of considerable interest. Herein, we report a facile synthesis of
L-tetrafluorotyrosine (L-f4Y) and its utilization as a pH sensing element. The synthetic route affords gram quantities of L-f4Y within two days.
Incorporation of this unnatural amino acid into magainin 2 gives a peptide that turns on its membrane-lytic activity upon mild acidification.
As the frontline of host-defense mechanisms across the
biological world,1 membrane-lytic peptides have attracted a
great deal of attention, particularly in the search for novel
antibiotics that are less prone to bacterial resistance.2
Furthermore, inducible membrane lysis by a biological trigger
holds great promise for a number of applications in biotech-
nology. For example, membrane-lytic peptides that respond
to low pH have been utilized in liposomal drug delivery
systems,3 where acidification of endosomes triggers mem-
brane lysis to release the encapsulated drug molecules. More
recently, the pH low insertion peptide (pHLIP) was success-
fully applied to the imaging of acidic cancerous tissues.4
The design of pH-sensitive membrane-active peptides has
been largely relying on the protonation of Glu or Asp side
chains, which display pKa values of 4.3 and 3.9 respectively.
Residues with higher pKa’s are desirable considering the pH
values of (patho)physiological relevance are generally above
5.5.4a We envisioned that pKa values of a tyrosine side chain
† These authors contributed equally to this work.
(1) Zasloff, M. Nature 2002, 415, 389–95.
(2) (a) Scott, R. W.; DeGrado, W. F.; Tew, G. N. Curr.Opin. Biotechnol.
2008, 19, 620–7. (b) Porter, E. A.; Weisblum, B.; Gellman, S. H. J. Am.
Chem. Soc. 2002, 124, 7324–30. (c) Fernandez-Lopez, S.; Kim, H. S.; Choi,
E. C.; Delgado, M.; Granja, J. R.; Khasanov, A.; Kraehenbuehl, K.; Long,
G.; Weinberger, D. A.; Wilcoxen, K. M.; Ghadiri, M. R. Nature 2001, 412,
452–5. (d) Chongsiriwatana, N. P.; Patch, J. A.; Czyzewski, A. M.; Dohm,
M. T.; Ivankin, A.; Gidalevitz, D.; Zuckermann, R. N.; Barron, A. E. Proc.
Natl. Acad. Sci. U.S.A. 2008, 105, 2794–9.
(3) (a) Li, W.; Nicol, F.; Szoka, F. C., Jr. AdV. Drug. DeliVery ReV.
2004, 56, 967–85. (b) Turk, M. J.; Reddy, J. A.; Chmielewski, J. A.; Low,
P. S. Biochim. Biophys. Acta 2002, 1559, 56–68.
(4) (a) Thevenin, D.; An, M.; Engelman, D. M. Chem. Biol. 2009, 16,
754–62. (b) Segala, J.; Engelman, D. M.; Reshetnyak, Y. K.; Andreev, O. A.
Int. J. Mol. Sci. 2009, 10, 3478–87.
10.1021/ol102610q  2011 American Chemical Society
Published on Web 12/10/2010

could be easily tuned to provide novel pH sensing elements
in the design of membrane-lytic peptides.
Scheme 1. Synthesis of Fmoc-L-f4Y-OH
Fluorination has become an attractive strategy in biomo-
lecular engineering because of the minimal steric perturbation
introduced by replacing hydrogen with fluorine atoms.5 On
the other hand, the high electronegativity of fluorine may
dramatically alter the electronic properties of a target
molecule.6 For example, fluorination of a tyrosine side chain
preferentially stabilizes the phenolate, thus affording tyrosine
isosteres with increased acidity. By varying the number of
fluorine substitutions, tyrosine analogues can be created to
display a wide range of pKa values (∼10 for the native
tyrosine and 5.6 for the tetrafluorinated analogue f4Y).7 These
fluorinated tyrosine analogues have been utilized as probes
to elucidate the mechanisms of a number of enzymes,
including kinases and ribonucleotide reductases.7,8 Herein
we show that f4Y serves as an effective pH sensor and
activates the membrane-lytic peptide magainin 2 (mag2)
under mildly acidic conditions.
It came as a surprise to us that there was no chemical
synthesis of f4Y that is stereospecific. The only chemical
synthesis of f4Y was reported by Kang and co-workers that
couples 4-methoxy-tetrafluorobenzyl bromide to diethylac-
etamidomalonate to give the racemic mixture of the O-methyl
protected f4Y (Scheme 1a).9 A biosynthetic route for prepar-
ing fluorotyrosine analogues in the enantiomerically pure
form has been developed by the Cole group and the Stubbe
group.7,8 This approach capitalizes on the enzymatic activity
of tyrosine phenol lyase that converts fluorinated phenols to
the corresponding L-fluorotyrosines in the presence of
pyruvate and ammonia (Scheme 1b). While this strategy
works well with mono-, di-, and trifluorine substituted
phenols, the tetrafluorinated analogue appears to be a poor
substrate for the enzyme. Consequently, the preparation of
L-f4Y takes multiple weeks and requires high concentrations
of the enzyme.
We report a short and efficient synthesis of L-f4Y with
the commercially available L-pentafluorophenylalanine
(dubbed L-f5F for brevity) as the starting material (Scheme
1c). The key transformation of our synthesis is the regiose-
lective nucleophilic addition-elimination (SNAr) reaction of
the pentafluorobenzyl moiety.10 Due to the balance between
the favorable inductive effect and the unfavorable resonance
effect, fluorination on the meta position of the reaction center
best accelerates the SNAr reaction (by 106-fold relative to a
hydrogen) followed by the ortho position (57-fold). With
resonance overwhelming the inductive effect, a fluorine
substituent on the para position actually slows down the
reaction by 2-fold. Given these considerations, the fluorine
on Cꢀ of the f5F side chain should be most activated (by
two meta and ortho fluorines) toward SNAr reactions. The
Boc protected L-f5F was chosen as the starting material. Initial
attempts using hydroxide as the nucleophile failed to give
any product in organic or aqueous media. The reaction
proceeded nicely to completion with methoxide as the
nucleophile. Sodium as the counterion afforded a slightly
faster reaction than potassium presumably because sodium
is a stronger Lewis acid (Table 1). Difficulty was encountered
with the removal of the O-methyl protecting group: BBr3
resulted in a messy mixture; even the milder alternative
(5) (a) Meng, H.; Kumar, K. J. Am. Chem. Soc. 2007, 129, 15615–22.
(b) Marsh, E. N.; Buer, B. C.; Ramamoorthy, A. Mol. Biosyst. 2009, 5,
1143–7. (c) Hodges, J. A.; Raines, R. T. J. Am. Chem. Soc. 2005, 127,
15923–32.
Table 1. Summary of the SNAr Reaction Conditions To Make
f4Y
(6) (a) Woll, M. G.; Hadley, E. B.; Mecozzi, S.; Gellman, S. H. J. Am.
Chem. Soc. 2006, 128, 15932–3. (b) Zheng, H.; Comeforo, K.; Gao, J. J. Am.
Chem. Soc. 2009, 131, 18–9. (c) Zheng, H.; Gao, J. Angew. Chem., Int.
Ed. 2010, 49, 8635–9. (d) Meyer, E. A.; Castellano, R. K.; Diederich, F.
Angew. Chem., Int. Ed. 2003, 42, 1210–50. (e) Waters, M. L. Curr. Opin.
Chem. Biol. 2002, 6, 736–41.
entry
nucleophile (solvent)
Yieldb (%)
1
2
3
4
KOH (tBuOH); KOH (DMSO); NaOH (H2O)
KOtBu (MeOH) f KOMe (MeOH)a, 48 h
NaOMe(MeOH), 36 h
0
>90
>90
88
(7) Kim, K.; Cole, P. A. J. Am. Chem. Soc. 1998, 120, 6851–8.
(8) (a) Seyedsayamdost, M. R.; Yee, C. S.; Stubbe, J. Nat. Protoc. 2007,
2, 1225–35. (b) Seyedsayamdost, M. R.; Reece, S. Y.; Nocera, D. G.;
Na (allyl alcohol) f NaOallyl (allyl alcohol)a
a The nucleophile shown on the right-hand side of the arrow was
Stubbe, J. J. Am. Chem. Soc. 2006, 128, 1569–79
.
generated in situ. b 0% yield means no product was identified by crude 19
F
(9) Filler, R.; Ayyangar, N. R.; Gustowsk., W.; Kang, H. H. J. Org.
Chem. 1969, 34, 534–8.
NMR.
(10) Rodionov, P. P.; Furin, G. G. J. Fluor. Chem. 1990, 47, 361–434.
Org. Lett., Vol. 13, No. 2, 2011
237

11
NaSC2H5 caused cleavage of the Boc protecting group,
is expected at such a high concentration. The membrane lysis
caused by the mag2 derivatives can be easily detected by
the fluorescence increase due to calcein release and dilution.
The membrane-lytic activities of the mag2 mutants are
shown in Figure 1a. Under neutral considitons (pH 7.0), no
which made the purification difficult. Finally, we turned to
sodium allyloxide, which reacted readily with Boc-f5F to give
the O-allyl protected L-f4Y. After converting the Boc group
to Fmoc, the O-allyl protecting group was removed by using
Pd(PPh3)4 in the presence of PhSiH3.12 We confirmed the
absence of epimerization by conjugating Fmoc-L-f4Y-OH to
a chiral amine, the product of which elutes as a single peak
during HPLC analysis (Figure S1, Supporting Information
(SI)). The lack of epimerization under such basic conditions
is presumably because the negatively charged carboxylate
prohibits the deprotonation of the R carbon. The synthetic
route has an overall yield of 72% and can easily afford gram
quantities of L-f4Y for peptide synthesis.
A well characterized membrane-lytic peptide magainin 2
(mag2)13 was used as a model system to test the potential
of f4Y in pH sensing. Mag2 is a 23-residue peptide that folds
into an amphiphilic helix upon membrane binding. Three
mag2 variants were synthesized on the Rink-MBHA amide
resin using standard Fmoc/HBTU chemistry (Table 2,
Table 2. Sequences of the Peptides Used in This Studya
WT mag2
GIGKFLHSAKKFGKAFVGEIMNS
1
2
3
dansyl-GIGK(f4Y)LHSAKK(f4Y)GKA(f4Y)VGEIMNS
dansyl-GIGKYLHSAKKYGKAYVGEIMNS
dansyl-GIGKELHSAKKEGKAEVGEIMNS
a The wild type mag2 sequence (WT) was included for comparison;
residues of interest are shown in bold:f4Y (negative at pH7.0, neutral at
pH5.0); Y (neutral in both); E (negative in both).
peptides 1-3). A dansyl group was installed at the N-termini
of all peptides to facilitate binding affinity measurement and
concentration calibration. Because fluorination renders the
f4Y side chain much less nucleophilic than that of Tyr, the
unprotected f4Y is compatible with the standard peptide
synthesis and cleavage conditions.8a However, the unpro-
tected f4Y did react with dansyl chloride to give the side
chain dansylated peptide. To circumvent this problem, we
used the protected f4Y (Fmoc-f4Y(OAllyl)-OH) in peptide
synthesis. The allyl protecting group was removed on resin
after the dansylation step by using Pd(PPh3)4 in the presence
of PhSiH3. Peptide 2 is the corresponding Tyr control for
the designed pH-responsive peptide 1. For comparison, we
also synthesized peptide 3 incorporating three Glu residues,
which have been widely used as pH sensing elements in
previous reports.3
Figure 1. (a) kinetics of peptide induced cargo (calcein) leakage.
The graph shows observed change in calcein fluorescence at 520
nm (λex ) 485 nm) upon addition of peptide (3 µM) into LUVs of
POPC (500 µM lipids, pH 7.0), followed by the addition of 6 N
HCl (12 µL) to reach pH 5.0. (b) pH profiles of peptide 1 induced
calcein leakage (black squares) and the fraction of membrane-bound
peptides (red dots); 3 µM peptide 1 and 500 µM lipids (POPC)
were used in these experiments. Peptide-liposome binding was
monitored by dansyl fluorescence at 510 nm with 340 nm excitation.
The data were fitted into a sigmoidal function to determine the pKa
values: 5.5 ( 0.1 based on the leakage data and 5.3 ( 0.2 based
on the binding curve.
membrane lysis was observed upon mixing the peptides 1
or 3 with liposomes. The tyrosine variant 2 afforded a small
degree (∼18%, Table S1, SI) of membrane damage. Acidi-
fication to pH 5.0 by HCl addition caused little change for
the control peptides 2 and 3. In sharp contrast, the f4Y mutant
1 displayed a dramatic increase in calcein fluorescence,
showing that the f4Y side chains indeed serve as effective
pH sensors to turn on the membrane-lytic acitivity of mag2.
The concentration dependence of the membrane-lytic potency
yields an EC50 value of 3.7 µM at pH 5.0 (Figure S2, SI).
Under neutral conditions (pH 7.0), no membrane damage
was observed even at 10 µM peptide concentration. The
membrane-lytic activity of peptide 1 was further evaluated
The membrane-lytic activities of peptides 1-3 were tested
by using a calcein leakage assay.14 Large unilamellar vesicles
(LUVs, 70 nm in their hydrodynamic radii) of 1-palmitoyl-
2-oleoyl-sn-glycero-3-phosphocholine (POPC) were prepared
to encapsulate calcein at 40 mM. Significant self-quenching
(11) Feutrill, G. I.; Mirrington, R. N. Tetrahedron Lett. 1970, 1327–8.
(12) Dessolin, M.; Guillerez, M. G.; Thieriet, N.; Guibe, F.; Loffet, A.
Tetrahedron Lett. 1995, 36, 5741–4.
(13) Matsuzaki, K. Biochim. Biophys. Acta 1998, 1376, 391–400.
(14) Lee, H. M.; Chmielewski, J. J. Pept. Res. 2005, 65, 355–63.
238
Org. Lett., Vol. 13, No. 2, 2011

under varied pH conditions (Figure 1b, black squares). The
pH dependence profile shows a clear sigmoidal curve with
the middle point of transition at pH 5.5, which agrees nicely
with the pKa values of f4Y.
lytic peptides. The mag2 mutant harboring f4Y residues
responds quickly to the pH drop from 7.0 to 5.0. No pH
activation was observed for the corresponding mutant that
incorporates Glu resdiues as pH sensors. This is largely
because the pKa value of f4Y falls into this physiologically
relevant region, while that of Glu does not. In addition, the
hydrophobic nature of the neutral f4Y allows their incorpora-
tion onto the membrane-binding face of a peptide, whereas
all previously reported pH sensing elements reside on the
hydrophilic face of an amphipathic structure. Given that
cancer tissues are often acidic, experiments are currently
underway to investigate the potential of the pH-triggered
membrane toxicity in killing cancer cells. Furthermore, we
expect the ease in synthesis and unique physical property of
f4Y will make it suitable for a wide array of applications in
protein biochemistry and engineering.
To gain a further mechanistic understanding of the acid
triggered membrane lysis, we evaluated the fraction of
peptide 1 bound to the liposomes in buffers of varied pH
values (Figure 1b, red dots). The peptide-lipid binding was
conveniently monitored by recording the dansyl fluorescence,
which is sensitive to the local environment. Peptide 1 shows
little binding to vesicles under neutral conditions, and
acidification elicits their partition onto liposomes. Interest-
ingly, the binding curve overlaps with that of the membrane-
lytic potency, indicating that the low pH induced peptide-
vesicle binding accounts for the pH-dependent activity of
peptide 1 toward membranes.
In summary, we have developed a short and effective
synthesis that allows one to prepare grams of L-f4Y within
two days. This protocol compares favorably to the existing
biosynthetic method that takes weeks to months. Our
synthesis should be easily extendable to the preparation of
D-f4Y simply by using the D version of f5F as the starting
material. We further demonstrate that f4Y functions as
effective pH sensors in the design of ‘smart” membrane-
Acknowledgment. This work was supported by the Smith
Family Foundation and the startup fund from Boston College.
Supporting Information Available: Experimental pro-
cedures and characterization of all new compounds. This
material is available free of charge via the Internet at
http://pubs.acs.org.
OL102610Q
Org. Lett., Vol. 13, No. 2, 2011
239