Antimicrobial activity and protease stability of peptides containing fluorinated amino acids

Article abstract of DOI:10.1021/ja075373f

Selective fluorination of peptides results in increased chemical and thermal stability with simultaneously enhanced hydrophobicity. We demonstrate here that fluorinated derivatives of two host defense antimicrobial peptides, buforin and magainin, display moderately better protease stability while retaining, or exhibiting significantly increased bacteriostatic activity. Four fluorinated analogues in the buforin and two in the magainin series were prepared and analyzed for (1) their ability to resist hydrolytic cleavage by trypsin; (2) their antimicrobial activity against both Gram-positive and Gram-negative bacterial strains; and (3) their hemolytic activity. All but one fluorinated peptide (M2F5) showed retention, or significant enhancement, of antimicrobial activity. The peptides also showed modest increases in protease resistance, relative to the parent peptides. Only one of the six fluorinated peptides (BII1F2) was degraded by trypsin at a slightly faster rate than the parent peptide. Hemolytic activity of peptides in the buforin series was essentially null, while fluorinated magainin analogues displayed an increase in hemolysis compared to the parent peptides. These results suggest that fluorination may be an effective strategy to increase the stability of biologically active peptides where proteolytic degradation limits therapeutic value.

Full text of DOI:10.1021/ja075373f

Published on Web 11/28/2007
Antimicrobial Activity and Protease Stability of Peptides
Containing Fluorinated Amino Acids
He Meng† and Krishna Kumar*,†,‡
Contribution from the Department of Chemistry, Tufts UniVersity, Medford, Massachusetts
02155, and Cancer Center, Tufts-New England Medical Center, Boston, Massachusetts 02110
Received July 18, 2007; E-mail: krishna.kumar@tufts.edu
Abstract: Selective fluorination of peptides results in increased chemical and thermal stability with
simultaneously enhanced hydrophobicity. We demonstrate here that fluorinated derivatives of two host
defense antimicrobial peptides, buforin and magainin, display moderately better protease stability while
retaining, or exhibiting significantly increased bacteriostatic activity. Four fluorinated analogues in the buforin
and two in the magainin series were prepared and analyzed for (1) their ability to resist hydrolytic cleavage
by trypsin; (2) their antimicrobial activity against both Gram-positive and Gram-negative bacterial strains;
and (3) their hemolytic activity. All but one fluorinated peptide (M2F5) showed retention, or significant
enhancement, of antimicrobial activity. The peptides also showed modest increases in protease resistance,
relative to the parent peptides. Only one of the six fluorinated peptides (BII1F2) was degraded by trypsin
at a slightly faster rate than the parent peptide. Hemolytic activity of peptides in the buforin series was
essentially null, while fluorinated magainin analogues displayed an increase in hemolysis compared to the
parent peptides. These results suggest that fluorination may be an effective strategy to increase the stability
of biologically active peptides where proteolytic degradation limits therapeutic value.
of peptides.5,7,8 Because of the broad spectrum activity and
ancient lineage of these peptides, it has been suggested that
Introduction
The emergence of bacterial resistance to common antibiotics
poses a serious threat to human health1 and has rekindled interest
in antimicrobial peptides.2,3 Both plants and animals have an
arsenal of short peptides with diverse structures that are part of
the innate immune system and are deployed against microbial
pathogens. The common distinguishing characteristic among
these peptides is their ability to form facially amphipathic
conformations, segregating cationic and hydrophobic side
chains.2,4 The classification of these peptides has usually been
on the basis of secondary structure. Both R-helical (magainins
and cecropins) and â-sheet (bactenecins and defensins) second-
ary structure elements are well represented among the 500 plus
members that have been hitherto identified.2,4-6 Most eukaryotes
express a combination of such peptides from many different
classes within tissues that provide the first line of defense against
invading microbes. The architectural details reveal the mech-
anism of actionspositive charges help the peptides seek out
negatively charged bacterial membranes, and the interaction of
the hydrophobic side chains with the acyl chain region of lipid
bilayers eventually leads to membrane rupture or translocation
bacterial resistance may be completely thwarted or slowed down
enough to offer a reasonably long therapeutic lifetime for
suitable candidates.2,9,10
Strategies to modulate antimicrobial activity of host defense
peptides have relied mainly on substitution at single (or multiple)
sites by 1 of the other 19 natural amino acids. This approach
has resulted in several improved variants, most notably the
[Ala8,13,18]-Magainin II amide (M2, Figure 1).11,12 On the other
hand, general principles gleaned from the study of natural
peptides have been utilized in the design of antimicrobial
peptides and polymers using unnatural building blocks. Several
of these constructs based on â-peptides,13-17 D,L-R-peptides,18
and arylamide polymers19,20 show impressive bactericidal activ-
ity. We21,22 and others23-25 have recently described peptide
(7) Matsuzaki, K. Biochim. Biophys. Acta 1999, 1462, (1-2), 1-10.
(8) Shai, Y. Biochim. Biophys. Acta 1999, 1462, (1-2), 55-70.
(9) Perron, G. G.; Zasloff, M.; Bell, G. Proc. R. Soc. London, Ser. B 2006,
273, (1583), 251-256.
(10) Yeaman, M. R.; Yount, N. Y. Pharmacol. ReV. 2003, 55, (1), 27-55.
(11) Zasloff, M. Proc. Natl. Acad. Sci. U.S.A. 1987, 84, (15), 5449-5453.
(12) Chen, H. C.; Brown, J. H.; Morell, J. L.; Huang, C. M. FEBS Lett. 1988,
236, (2), 462-466.
(13) Porter, E. A.; Wang, X. F.; Lee, H. S.; Weisblum, B.; Gellman, S. H. Nature
2000, 404, (6778), 565-565.
† Tufts University.
(14) Porter, E. A.; Weisblum, B.; Gellman, S. H. J. Am. Chem. Soc. 2002, 124,
(25), 7324-7330.
‡ Tufts-New England Medical Center.
(1) Coates, A.; Hu, Y. M.; Bax, R.; Page, C. Nat. ReV. Drug DiscoVery 2002,
(15) Schmitt, M. A.; Weisblum, B.; Gellman, S. H. J. Am. Chem. Soc. 2004,
126, (22), 6848-6849.
1, (11), 895-910.
(2) Zasloff, M. Nature 2002, 415, (6870), 389-395.
(3) Hancock, R. E. W.; Sahl, H. G. Nat. Biotechnol. 2006, 24, (12), 1551-
1557.
(16) Raguse, T. L.; Porter, E. A.; Weisblum, B.; Gellman, S. H. J. Am. Chem.
Soc. 2002, 124, (43), 12774-12785.
(17) Liu, D. H.; DeGrado, W. F. J. Am. Chem. Soc. 2001, 123, (31), 7553-
7559.
(4) Tossi, A.; Sandri, L.; Giangaspero, A. Biopolymers 2000, 55, (1), 4-
30.
(18) 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, (6845), 452-455.
(5) Brogden, K. A. Nat. ReV. Microbiol. 2005, 3, (3), 238-250.
(6) Ganz, T. Nat. ReV. Immunol. 2003, 3, (9), 710-720.
9
10.1021/ja075373f CCC: $37.00 © 2007 American Chemical Society
J. AM. CHEM. SOC. 2007, 129, 15615-15622
15615

A R T I C L E S
Meng and Kumar
assemblies that incorporate highly fluorinated residues that have
superior thermal and chemical stability. The increased hydro-
phobicity of fluorinated amino acids fuels the formation of stable
secondary and quaternary protein structures in aqueous solutions
if the nonpolar fluorinated surfaces can be segregated away from
water. Indeed, appropriately designed fluorinated peptides show
a higher affinity for membranes as in the case of cell lytic
melittin26 and can also direct discrete oligomer formation in
biological membranes.27,28 We envisioned that increased mem-
brane affinity and greater structural stability could yield peptide
variants that have increased potency and stability toward
proteolytic enzymes than known antimicrobials.
Experimental Section
Materials. Boc-Lys(2-Cl-Z)-Merrifield resin, 4-methylbenzhydry-
lamine (MBHA) resin, Boc-L-amino acids, and 2-(1H-benzotriazol-1-
yl)-1,1,3,3-tetramethyluronium hexafluorophosphate (HBTU) were
purchased from Novabiochem. Dichloromethane (ACS grade, Fisher),
dimethylformamide (sequencing grade, Fisher), N,N-diisopropylethy-
lamine (biotech grade, Aldrich), trifluoroacetic acid (New Jersey
Halocarbon), anisole (anhydrous, Aldrich), trifluoroethanol (TFE, 99%,
Acros), and acetonitrile (Optima grade, Fisher) were used without
further purification. Hydrogen fluoride was purchased from Matheson
Gas. Melittin (from honey bee venom), trypsin (from bovine pancreas),
and benzoyl arginine ethyl ester (BAEE) were purchased from Sigma.
Fresh human red blood cells (type B) were procured from Research
Blood Components L.L.C., Boston, MA. Chemical synthesis of
hexafluoroleucine was carried out according to a previously disclosed
procedure with slight modifications.29 Solvents for reversed-phase high-
performance liquid chromatography (RP-HPLC) had the following
compositions: Solvent A, H2O/CH3CN/TFA (99/1/0.1); Solvent B, CH3-
CN/H2O/TFA (90/10/0.07).
Peptide Synthesis, Purification, and Characterization. Peptides
were synthesized manually using the in situ neutralization protocol for
t-Boc chemistry30 on a 0.075 mmol scale. Boc-Lys(2-Cl-Z)-Merrifield
and MBHA resins were used for buforin and magainin peptides. The
dinitrophenyl protecting group on histidine was removed using a 20-
fold molar excess of thiophenol. Peptides were cleaved from resin by
treatment with HF/anisole (90:10) at 0 °C for 2 h and then precipitated
with cold Et2O. Crude peptides were purified by RP-HPLC [Vydac
C18, 10 µ, 10 mm × 250 mm]. The purities of peptides were judged to
be greater than 95% by analytical RP-HPLC [Vydac C18, 5 µ, 4 mm ×
250 mm]. The molar masses of peptides were determined by MALDI-
TOF MS. Peptide concentrations were determined by quantitative amino
acid analysis.
Figure 1. Amino acid sequences of antimicrobial peptides. The net charges
at pH 7.40 are given in parentheses. L: 5,5,5,5′,5′,5′-2S-hexafluoroleucine.
(M + H+); BII5 m/z calcd (M) 2002.2, obsd 2003.5 (M + H+); BII5F2
m/z calcd (M) 2218.1, obsd 2218.9 (M + H+); BII6 m/z calcd (M)
1847.2, obsd 1848.5 (M + H+); BII6F2 m/z calcd (M) 2063.1, obsd
2064.2 (M + H+); BII10 m/z calcd (M) 1477.8, obsd 1479.2 (M +
H+); BII10F2 m/z calcd (M) 1693.7, obsd 1695.2 (M + H+); M2 m/z
calcd (M) 2476.4, obsd 2496.1 (M + Na+); M2F2 m/z calcd (M)
2692.3, obsd 2693.6 (M + H+); M2F5 m/z calcd (M) 3114.2, obsd
3115.5 (M + H+).
Antimicrobial Activity. Minimal Inhibitory Concentrations (MICs)
were measured against Gram-negative Escherichia coli (ATCC 23716)
and Gram-positive Bacillus subtilis (SMY) using mid-logarithmic phase
cells. Bacteria from a single colony were grown overnight in Luria
Bertani broth at 37 °C with agitation. An aliquot was taken and diluted
(1:50) in fresh broth and cultured for ∼2 h. The cells (OD590 ∼0.5)
were diluted to a concentration of 5 × 104 colony forming units/mL
(CFU/mL) for the BII series peptides and a concentration of 5 × 105
CFU/mL for the M2 series. Serial dilution (2-fold) of peptide solutions
was performed in a sterile 96-well plate (MICROTEST) in duplicate
to a final volume of 50 µL in each well, followed by addition of 50 µL
of cell suspension. The concentrations of peptides tested were from
0.3 µg/mL to 256 µg/mL. Subsequent to incubation at 37 °C for 6 h,
the absorbance at 590 nm was monitored using a microtiter plate reader
(VERSAmax). The MIC was recorded as the concentration of peptide
required for the complete inhibition of cell growth (no change in
absorbance).
MALDI-TOF MS Characterization: BII1 m/z calcd (M) 2432.4,
obsd 2434.9 (M + H+); BII1F2 m/z calcd (M) 2649.3, obsd 2650.7
(19) Tew, G. N.; Liu, D. H.; Chen, B.; Doerksen, R. J.; Kaplan, J.; Carroll, P.
J.; Klein, M. L.; DeGrado, W. F. Proc. Natl. Acad. Sci. U.S.A. 2002, 99,
(8), 5110-5114.
(20) Liu, D. H.; Choi, S.; Chen, B.; Doerksen, R. J.; Clements, D. J.; Winkler,
J. D.; Klein, M. L.; DeGrado, W. F. Angew. Chem., Int. Ed. 2004, 43, (9),
1158-1162.
(21) Bilgic¸er, B.; Fichera, A.; Kumar, K. J. Am. Chem. Soc. 2001, 123, (19),
4393-4399.
(22) Bilgic¸er, B.; Xing, X.; Kumar, K. J. Am. Chem. Soc. 2001, 123, (47),
11815-11816.
(23) Tang, Y.; Ghirlanda, G.; Vaidehi, N.; Kua, J.; Mainz, D. T.; Goddard, W.
A.; DeGrado, W. F.; Tirrell, D. A. Biochemistry 2001, 40, (9), 2790-
2796.
Hemolysis Assay. Fresh human red blood cells (hRBCs) were
centrifuged at 3500 rpm and washed with PBS buffer until the
supernatant was clear. The hRBCs were then resuspended and diluted
to a final concentration of 1% (v/v) in PBS and stored at 4 °C. Serial
dilution (2-fold) of peptides in PBS in a 96-well plate resulted in a
final volume of 20 µL in each well, to which 80 µL of hRBCs were
added. The plate was incubated at 37 °C for 1 h, followed by
centrifugation at 3500 rpm for 10 min using a SORVALL tabletop
centrifuge. An aliquot (50 µL) of supernatant was transferred to a new
(24) Tang, Y.; Tirrell, D. A. J. Am. Chem. Soc. 2001, 123, (44), 11089-11090.
(25) Lee, K. H.; Lee, H. Y.; Slutsky, M. M.; Anderson, J. T.; Marsh, E. N. G.
Biochemistry 2004, 43, (51), 16277-16284.
(26) Niemz, A.; Tirrell, D. A. J. Am. Chem. Soc. 2001, 123, (30), 7407-7413.
(27) Bilgic¸er, B.; Kumar, K. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, (43),
15324-15329.
(28) Naarmann, N.; Bilgicer, B.; Meng, H.; Kumar, K.; Steinem, C. Angew.
Chem., Int. Ed. 2006, 45, (16), 2588-2591.
(29) Xing, X.; Fichera, A.; Kumar, K. Org. Lett. 2001, 3, (9), 1285-1286.
(30) Schnolzer, M.; Alewood, P.; Jones, A.; Alewood, D.; Kent, S. B. H. Int. J.
Pept. Protein Res. 1992, 40, (3-4), 180-193.
9
15616 J. AM. CHEM. SOC. VOL. 129, NO. 50, 2007

Peptides Containing Fluorinated Amino Acids
A R T I C L E S
Table 1. Antimicrobial and Hemolytic Activities and
Hydrophobicities of the Peptides
MICa
(
µ
g/mL)
b
HC50
(µg/mL)
solvent B (%)
for elutionc
peptide
E. coli
B. subtilis
BII1
BII1F2
BII5
BII5F2
BII6
BII6F2
BII10
BII10F2
M2
20
5
5
10
2.5
5
5
400
400
400
400
400
400
400
400
175
20
31.0
37.0
32.1
38.4
33.0
39.5
29.0
35.7
56.5
62.1
85.8
5
>256
>256
80
20
>256
40
>256
10
2.5
2.5
40
2.5
2.5
10
M2F2
M2F5
11
a MICs represent at least two independent experiments, each in duplicate,
and are reported as the highest determined values. bThe HC50 values were
extracted by extrapolation of the fitted curve to 50% lysis of hRBCs. When
no detectable hemolytic activity was observed at 400 µg/mL, a value of
400 µg/mL was used. cPercentage solvent B (9:1:0.007 CH3CN/H2O/
CF3CO2H) required for elution on RP-HPLC (J.T. Baker C18, 5 µ, 4 mm ×
250 mm).
Figure 2. Helical wheel diagrams and sites of fluorination. (A) Buforin
series: the arrangement of nonpolar residues is based on ref 38, Leu 18
and 19 (blue) were substituted with hexafluoroleucine; (B) magainin
series sites of fluorination: residues Leu 6 and Ile 20 in M2F2 and Leu 6,
Ala 9, Gly 13, Val 17, and Ile 20 in M2F5; (C) model structure of BII10.54
Leu 18 and 19 are shown in space-filling depiction; (D) NMR
structure of magainin 2 in dodecylphosphocholine micelles (PDB code:
2mag).
Figure 3. Hemolytic activities of peptides against type B hRBCs relative
to melittin. Each data point is the average of at least two independent
experiments with two replicates.
96-well plate containing 50 µL of H2O in each well. Release of
hemoglobin was monitored at 415 nm using a microtiter plate reader.
Percentage hemolysis was calculated using eq 1:
The remaining full-length peptide concentration was normalized with
respect to the initial concentration. Pseudo-first-order rate constants
were obtained by data (<initial 20 min) using eq 2:
(A415,peptide - A415,buffer
)
y ) a + b ‚ e-k*t
(2)
Percentage hemolysis ) 100 ‚
(1)
(A415,complete hemolysis - A415,buffer
)
where y is the normalized concentration of peptides; k, the pseudo-
first-order rate constant; t, the reaction time in mins; and b, the initial
concentration of peptides. Each fragment cleaved from the full-length
peptides was identified by ESI-MS so that cleavage patterns could be
established and compared.
Circular Dichroism. Circular dichroism (CD) spectra were re-
corded at 25 °C on a JASCO J-715 spectropolarimeter fitted with
a PTC-423S single position Peltier temperature controller using
a 1 cm path length cuvette. TFE titrations were carried out in
PBS buffer at a fixed peptide concentration (10 µM). Four scans were
acquired per sample and averaged to improve the S/N ratio. A
baseline was recorded and subtracted after each spectrum. Mean
residue ellipticities ([θ], deg‚cm2‚dmol-1) were calculated using
eq 3:
where complete hemolysis is defined as the average absorbance of all
wells containing 400 µg/mL melittin. Negative controls consisted of
hRBCs alone.
Protease Stability of Peptides. The proteolytic stability of peptides
toward trypsin (from bovine pancreas, EC 3.4.21.4) was determined
by an analytical RP-HPLC assay. A standard substrate, N-R-benzoyl-
L-arginine ethyl ester (BAEE), was used to calibrate enzymatic activity
by measuring absorbance at 254 nm. The enzyme concentration (in 1
mM HCl) was determined by absorbance at 280 nm (ꢀ1% ) 13.3 cm-1).
In a typical trypsinization experiment, the buforin peptides (0.2 mM)
and 0.1 µg trypsin or the magainin peptides (0.25 mM) and 0.2 µg
trypsin in 200 µL of PBS buffer (pH 7.4, 10 mM PO43-, 150 mM
NaCl) were used. The amount of enzyme was optimized so that the
kinetics of proteolytic reactions could be assayed by RP-HPLC
(detection at 230 nm). The peptides were incubated with trypsin at
37 °C over a period of 3 h. Aliquots (10 µL) were taken at different
reaction times, diluted with 0.2% TFA (440 µL), and stored at
-80 °C. A C18 analytical column [J.T. Baker C18, 5 µ, 4 mm × 250
mm] was used for separation and quantitation of digested products.
[θ] ) θobs × MRW/10 ‚ l ‚ c
(3)
where θobs is the measured signal (ellipticity) in millidegrees, l, the
optical path length of the cell in cm, c, the concentration of the peptide
in mg/mL, and MRW, the mean residue molecular weight (molecular
9
J. AM. CHEM. SOC. VOL. 129, NO. 50, 2007 15617

A R T I C L E S
Meng and Kumar
Figure 4. Protease stability as probed using analytical RP-HPLC. Aliquots of reaction mixtures were removed, quenched, and analyzed by RP-HPLC and
ESI-MS. (A) Relative rates of proteolytic cleavage of fluorinated peptides compared to hydrocarbon controls; (B) fragments BII5*(5-14) and BII5*(5-17)
appearance and degradation; (C) fragment BII1*(6-21) appearance and degradation; (D) fragment M*(1-14) appearance and degradation. Rate constants
for degradation of fluorinated peptides are relative to the parent hydrocarbon peptide. Each data point is the average of two independent experiments.
weight of the peptide divided by the number of residues). Percentage
helical contents were calculated using eq 4:
according to amino acid composition using the program SEDNTERP,
assuming the PSV of hexafluoroleucine to be equivalent to that of
leucine.
[θ]222 × 100
Results and Discussion
Helical content (%) )
(4)
2.5
n
-40 000 ‚ 1 -
[
]
Model Peptides and Sites of Fluorination. Two antimicro-
bial peptides, buforin II (BII)33 and magainin II amide (M2),12
were chosen as templates for fluorination. While both peptides
are capable of exerting their bactericidal activity at low
micromolar concentrations, their modes of action are quite
distinct. Although both are initially drawn to negatively charged
bacterial membranes by electrostatic interactions, M2 causes
cell lysis by forming torodial pores in lipid bilayers,5,7 whereas
BII penetrates into the cell and kills bacteria presumably by
binding intracellular DNA and RNA.34,35 Furthermore, the
peptides chosen are selective against bacterial cells as is evident
from their low hemolytic activity. With regard to their potency,
while the peptides in the M2 series operate at the lower limit
of MIC concentration typically observed,4 the BII peptides in
principle could benefit from increased efficacy. Both translo-
cation of BII into cells and pore formation by M2 seem to be
partially controlled by hydrophobic interactions.7,36,37 The
transfer free energies from water to n-heptanol for side chains
of hexafluoroleucine (-2.5 kcal/mol) and leucine (-2.1 kcal/
where [θ]222 is the mean residue ellipticity at 222 nm, and n, the number
of residues.31,32
Analytical Ultracentrifugation. Equilibrium sedimentation analysis
was performed for M2, M2F2, and M2F5 at 25 °C on a Beckman
ProteomeLab XL-I analytical ultracentrifuge. Peptides dissolved in PBS
were loaded into equilibrium cells at concentrations of 50 and 100 µM.
Absorbance data at 230 nm were acquired at three different rotor speeds
(35 000, 40 000, and 45 000 rpm) after equilibration for 18 h. Data
obtained were fitted using eq 5 that describes the sedimentation of a
single ideal species using Igor Pro v5.03:
Abs ) A′ exp(H × M[x2 - x0 ]) + B
(5)
2
where Abs ) absorbance at radius x, A′ ) absorbance at reference radius
x0, H ) (1 - Vh F)ω2/2RT, Vh ) partial specific volume (PSV ) 0.7673
mL/g), F ) density of solvent (1.0017 g/mL), ω ) angular velocity in
radians/s, R ) gas constant in g/mol‚K, T ) absolute temperature (298
K), M ) apparent molecular weight (Da), and B ) solvent absorbance
(blank). The partial specific volumes of peptides were estimated
(33) Park, C. B.; Yi, K. S.; Matsuzaki, K.; Kim, M. S.; Kim, S. C. Proc. Natl.
Acad. Sci. U.S.A. 2000, 97, (15), 8245-8250.
(31) Chen, Y. H.; Yang, J. T.; Chau, K. H. Biochemistry 1974, 13, (16), 3350-
(34) Park, C. B.; Kim, H. S.; Kim, S. C. Biochem. Biophys. Res. Commun. 1998,
244, (1), 253-257.
3359.
(32) Javadpour, M. M.; Barkley, M. D. Biochemistry 1997, 36, (31), 9540-
(35) Kobayashi, S.; Takeshima, K.; Park, C. B.; Kim, S. C.; Matsuzaki, K.
Biochemistry 2000, 39, (29), 8648-8654.
9549.
9
15618 J. AM. CHEM. SOC. VOL. 129, NO. 50, 2007

Peptides Containing Fluorinated Amino Acids
A R T I C L E S
Figure 5. Helical content of buforins and fluorinated analogues in TFE/aqueous buffer probed by circular dichroism spectroscopy. (A) BII1 and BII1F2;
(B) BII5 and BII5F2; (C) BII6 and BII6F2; (D) BII10 and BII10F2.
mol)25 point to the superior hydrophobicity of the fluorinated
congener. We envisaged that incorporation of hexafluoroleucine
at selected positions would simultaneously increase membrane
affinity and provide greater protease stability. We also explored
fluorinated derivatives of buforin peptides truncated at the
N-terminus. The sequences of peptides and the fluorinated
analogues are shown in Figure 1. Since these peptides adopt
amphipathic helical conformations, sites of fluorination were
selected on the nonpolar face of helices with the help of helical
wheel diagrams. In the case of the buforin peptides, leucine
residues at positions 18 and 19 on the nonpolar face, which
form part of the putative DNA/RNA binding sequence,34,38 were
replaced with hexafluoroleucine. M2F2 was constructed by
replacing Leu6 and Ile20 with hexafluoroleucine. M2F5 was
designed by replacing five residues on the nonpolar face with
hexafluoroleucine to yield a highly fluorinated surface (Fig-
ure 2).
with improvements in MIC values ranging from 3-fold to greater
than 25-fold, while the fourth derivative (BII5F2) had equivalent
activity as compared to the control peptides. On the other hand,
M2F2 exhibited same MIC values as M2 while M2F5 was 4-
and 16-fold less active against B. subtilis and E. coli. A survey
of the antimicrobial host defense peptides suggests that the most
active members have MICs in the 1-3 µM range.4,39,40 It is
therefore not surprising that BII5F2 and M2F2 show essentially
the same efficacy as BII5 and M2 as these peptides are already
operating at the potency limit. Furthermore, the data obtained
demonstrate that the bioactivity of these peptides is either
retained at the frequently invoked maximal potency or improved
by means of fluorination. Selective, controlled fluorination may
therefore serve as a powerful tool to enhance the activity of
known antimicrobial peptides to optimal levels. The only peptide
to have lost its potency upon fluorination, M2F5, forms a helical
bundle in aqueous solution (Vide infra).
Antimicrobial Activity. The antimicrobial activity was
assessed as a minimal inhibitory concentration (MIC) using
turbidity assays against both Gram-positive (B. subtilis) and
Gram-negative (E. coli) bacteria (Table 1). All of the fluorinated
peptides had better or similar antimicrobial activities relative
to the parent peptides with the exception of M2F5. Three of
the four fluorinated buforin analogues had increased potency,
Hemolytic Activity and Hydrophobicity. The ability of the
peptides to disrupt the membrane integrity of mammalian cells
was interrogated by a hemolysis assay using type B hRBCs.13
The buforin analogues had hemolytic activity essentially the
same as those associated with the control peptides indicating
that passage across the membrane was not compromised by
fluorination (Figure 3). In addition, the data are also indicative
of the fact that these peptides do not operate by membrane lysis
(36) Tachi, T.; Epand, R. F.; Epand, R. M.; Matsuzaki, K. Biochemistry 2002,
41, (34), 10723-10731.
(37) Kobayashi, S.; Chikushi, A.; Tougu, S.; Imura, Y.; Nishida, M.; Yano, Y.;
Matsuzaki, K. Biochemistry 2004, 43, (49), 15610-15616.
(38) Yi, G. S.; Park, C. B.; Kim, S. C.; Cheong, C. FEBS Lett. 1996, 398, (1),
87-90.
(39) Porter, E. A.; Weisblum, B.; Gellman, S. H. J. Am. Chem. Soc. 2005, 127,
(32), 11516-11529.
(40) Hilpert, K.; Volkmer-Engert, R.; Walter, T.; Hancock, R. E. W. Nat.
Biotechnol. 2005, 23, (8), 1008-1012.
9
J. AM. CHEM. SOC. VOL. 129, NO. 50, 2007 15619

A R T I C L E S
Meng and Kumar
Protease Stability. A major limitation of antimicrobial
peptide function is their inactivation by proteases. We examined
the susceptibility of the various peptides against trypsin,14,43
a
commercially available protease. Trypsin specifically catalyzes
the hydrolysis of C-terminal amide bonds of lysine and arginine,
making it an ideal enzyme to examine this parameter as the
substrates used in our study contain multiple cleavage sites in
the form of residues with cationic side chains. Poor protease
stability severely limits the clinical use of many therapeutic
peptides.44,45 Results reported here will therefore inform future
investigations involving fluorination of peptide-based drug
candidates.
All fluorinated peptides were generally more stable toward
trypsin digestion (Figure 4). The BII analogues BII5F2, BII6F2,
and BII10F2 were ∼2.3-fold more resistant to hydrolysis, while
BII1F2 was similar to BII1. The initial P1 site of cleavage,
according to the nomenclature of Schechter and Berger,46 was
different in BII5F2 (R14) than BII5 (R17). This was also the
case when one compares the initial cleavage site in BII1F2 and
BII1 (See Supporting Information). In addition, the initial
cleavage fragment BII1F2(6-21) accumulated and persisted
much longer than BII1(6-21). In both cases, the presence of
hexafluoroleucine at the P1′ and P2′ sites confers solid protection
to the R17 cleavage site. In addition, BII6F2 is a much more
effective antimicrobial than BII6, which may make BII1F2
retain potency even after suffering proteolytic degradation. A
similar trend was observed for the magainin analogues. M2F2
was more stable to proteolysis relative to M2 by 35%, whereas
M2F5 was fiercely resistant to degradation, with >78% of the
peptide remaining in solution after 3 h. In contrast, M2 is
completely hydrolyzed in less than 40 min. The initial fragment
resulting from cleavage of M2F2(1-14) accumulated in higher
amounts than M2(1-14) and only underwent minimal pro-
teolytic degradation over 3 h. The presence of a single
hexafluoroleucine residue at position 6 (P2′ site) in M2F2(1-
14) confers a dramatic advantage in protecting the K4 amide
bond. Collectively, hexafluoroleucine at either the P1′ site and/
or P2′ site is sufficient to provide protection against catalyzed
hydrolytic damage.
The mechanism by which hexafluoroleucine is able to impart
greater protease stability requires further detailed investigation.
To a first approximation, the effect could either be attributed
to the increased steric bulk of the fluorinated amino acid or to
electronic perturbation of the amide bond. The acid dissociation
constants for hexafluoroleucine (pKa values of 7.51 and 2.79)
indicate that the acidity of ammonium group is increased ∼100
fold compared to leucine (pKa values of 9.71 and 2.61).47 The
substitution of a single fluorine for hydrogen is often considered
to be isosteric; however, the side chain volume of hexafluoro-
leucine is calculated to be ∼37 Å3 larger than that of leucine.25
Taken together, the evidence suggests that while a moderate
electronic perturbation may still be operational, it is more likely
that the protease protection is a result of steric occlusion of the
peptide from the active site. In the case of M2F5, the
Figure 6. Circular dichroism spectra and effect of TFE on the helical
content for magainins. (A) CD spectra of M2, M2F2, and M2F5 in PBS
and 50% TFE in PBS; (B) TFE effect on the helical content of M2, M2F2,
and M2F5.
as is the case in the magainin series. On the other hand, M2F2
was moderately more hemolytic than M2 at the MIC value41
(Figure 3).
It has been demonstrated previously that increased hydro-
phobicity correlates with hemolytic activity.4,36,42 Accordingly,
we evaluated the hydrophobicity of the peptides based on their
retention times on reversed phase HPLC.26,36 The introduction
of hexafluoroleucine resulted in increased hydrophobicity of
parent peptides with an average increase of 3.1((0.2)% solvent
B per Leu (or Ile) substituted. The two peptides exert their action
using fundamentally different mechanisms, and it appears that
increased hydrophobicity correlates with greater antimicrobial
activity in the buforin series, with essentially no perturbation
of hemolytic potential. On the other hand, the magainin
derivatives show a significant increase in the ability to disrupt
hRBC membranes. Overall, these results are consistent with
previous studies that showed increased hydrophobicity results
in higher hemolytic tendency, indicating that a fine balance
between hydrophobicity and charge as in the case of magainin-
like peptides is crucial for redesign with fluorinated analogues.
It is likely that a hydrophobicity maximum of the parent peptide
(>75% Solvent B required for elution in RP-HPLC under the
conditions specified in Table 1) exists, beyond which fluorina-
tion may not result in retention of selectivity for bactericidal
activity over mammalian cell permeabilization.
(43) Rozek, A.; Powers, J. P. S.; Friedrich, C. L.; Hancock, R. E. W.
Biochemistry 2003, 42, (48), 14130-14138.
(44) Latham, P. W. Nat. Biotechnol. 1999, 17, (8), 755-757.
(45) Sato, A. K.; Viswanathan, M.; Kent, R. B.; Wood, C. R. Curr. Opin.
Biotechnol. 2006, 17, (6), 638-642.
(41) The peptide M2F5 was significantly more hemolytic than M2. The
hydrophobicity of this peptide was significantly increased upon fluorination
leading to both reduced antimicrobial activity and higher hemolytic activity.
(42) Wieprecht, T.; Dathe, M.; Beyermann, M.; Krause, E.; Maloy, W. L.;
MacDonald, D. L.; Bienert, M. Biochemistry 1997, 36, (20), 6124-6132.
(46) Schechte, I.; Berger, A. Biochem. Biophys. Res. Commun. 1967, 27, (2),
157-162.
(47) Zhang, C.; Ludin, C.; Eberle, M. K.; Stoeckli-Evans, H.; Keese, R. HelV.
Chim. Acta 1998, 81, (1), 174-181.
9
15620 J. AM. CHEM. SOC. VOL. 129, NO. 50, 2007

Peptides Containing Fluorinated Amino Acids
A R T I C L E S
Figure 7. Representative equilibrium analytical ultracentrifugation traces for M2 (A) and M2F5 (B) [25 °C, 35000 rpm at 230 nm]. Fits to a single ideal
species model are shown as a solid line with residuals in the top frame. Conditions: [peptide] ) 50 µM, 10 mM phosphate, pH 7.40, 137 mM NaCl, 2.7
mM KCl. The observed apparent molecular weights were 2413 (M2, calc. 2478 for monomer) and 12 436 (M2F5, calc. 12 460 for tetramer). Plots of ln(A)
vs r2 for M2 (C) indicate a single ideal species, while nonrandom residuals for M2F5 (D) suggest presence of other aggregation states.
conformational stability is greatly increased, thereby denying
the protease access to the labile amide.
Overall, the R-helical content is related to more pronounced
antibacterial action in the case of buforins. In contrast, higher
amounts of secondary structure correlates with increased
hemolysis and formation of aggregates in the magainin series.
Circular Dichroism and Analytical Ultracentrifugation.
Antimicrobial potency has been shown to correlate closely with
R-helical content in several instances. Accordingly, we used CD
spectroscopy to probe the influence of fluorination on secondary
structure formation in TFE/water mixtures, a solvent known to
mimic the nonpolar environment of membranes. All peptides
with the exception of M2F5 were unstructured in aqueous
solutions. However with increasing amounts of TFE, the
peptides adopted an R-helical structure, consistent with their
putative membrane active conformations. The fluorinated pep-
tides exhibited an increased propensity for adopting R-helical
conformations in both the buforin and magainin series peptides
(Figures 5 and 6). In the case of the buforin peptides, higher
R-helical content is coupled closely with antimicrobial potency.
At 50% TFE, M2, M2F2, and M2F5 were ∼60% helical.
However, only M2F5 was significant helical (∼40%) in buffered
aqueous solutions with no exogenous TFE (Figure 6). This
dramatic difference necessitated further examination of the
respective oligomeric states. M2 was monomeric as judged by
analytical ultracentrifugation while both M2F2 and M2F5 had
a tendency to populate multiple oligomeric states (Figure 7 and
Supporting Information). Indeed, M2F5 appears to form stable
helical bundles in aqueous solutions providing an explanation
for decreased antimicrobial activity, increased hemolytic activity,
and greatly enhanced protease stability.32,48,49
Conclusions
Stabilization of helical peptide assemblies by incorporation
of fluorinated amino acids has recently been realized by several
groups.21-24,50,51 However, incorporation of such amino
acids into bioactive peptides has been limited.52,53 We have
demonstrated here that incorporation of hexafluoroleucine at
selected sites in antimicrobial peptides results in simul-
taneous enhancement of potency and increased resistance to
protease degradation. Five of six fluorinated peptides (M2F2,
BII1F2, BII5F2, BII6F2, and BII10F2) used in this
study showed significantly enhanced (or similar) bactericidal
activity. The lone exception, M2F5, which is highly fluorinated,
had a proclivity to form helical bundles in aqueous solutions
and had diminished antimicrobial power and increased hemo-
lytic activity. Furthermore, this peptide was extremely resistant
to protease action. The other five analogues showed in-
creased or similar protease stability as well as altered cleavage
(50) Son, S.; Tanrikulu, I. C.; Tirrell, D. A. ChemBioChem 2006, 7, (8), 1251-
1257.
(51) Lee, H. Y.; Lee, K. H.; Al-Hashimi, H. M.; Marsh, E. N. G. J. Am. Chem.
Soc. 2006, 128, (1), 337-343.
(52) Vine, W. H.; Hsieh, K. H.; Marshall, G. R. J. Med. Chem. 1981, 24, (9),
1043-1047.
(48) Papo, N.; Shai, Y. J. Biol. Chem. 2005, 280, (11), 10378-10387.
(49) Kuroda, K.; DeGrado, W. F. J. Am. Chem. Soc. 2005, 127, (12), 4128-
4129.
(53) Hsieh, K. H.; Needleman, P.; Marshall, G. R. J. Med. Chem. 1987, 30,
(6), 1097-1100.
9
J. AM. CHEM. SOC. VOL. 129, NO. 50, 2007 15621

A R T I C L E S
Meng and Kumar
patterns. The hemolytic properties of these peptides were
comparable to the parent control peptides from which they were
derived. Poor protease stability of short R-peptides severely
limits their therapeutic value. Our study highlights the utility
of fluorinated peptides in reviving promising leads that suffer
such a fate. Further, it is evident that one could perturb peptide-
receptor interactions using fluorinated derivatives.
for helpful discussions and careful reading of the manuscript.
This work was supported in part by the NIH (GM65500), NSF,
and the Massachusetts Technology Transfer Center. The AUC
and ESI-MS facilities at Tufts are supported by the NIH
(1S10RR017948) and NSF (0320783).
Supporting Information Available: HPLC data detailing
protease stability assays and accompanying kinetic analysis. ESI-
MS data for all peptide fragments generated in protease
degradation assays, circular dichroism spectra, and AUC data
for peptides in tabular format. This material is available via the
Internet at http://pubs.acs.org/.
Acknowledgment. We thank Dr. A. Fichera for initial help
with the synthesis of hexafluoroleucine and Prof. M. d’Alarcao
(54) The model structure was generated by automated homology modeling
using SWISS-MODEL v3.5 (AACODE: P55897). Schwede, T.;
Kopp, J.; Guex, N.; Peitsch, M. C. Nucleic Acids Res. 2003, 31, 3381-
3385.
JA075373F
9
15622 J. AM. CHEM. SOC. VOL. 129, NO. 50, 2007