Total synthesis, characterization, and conformational analysis of the naturally occurring hexadecapeptide integramide A and a diastereomer

Article abstract of DOI:10.1002/chem.200900945

Integramide A is a 16-amino acid peptide inhibitor of the enzyme HIV-1 integrase. We have recently reported that the absolute stereochemistries of the dipeptide sequence near the C terminus are L-Iva¹⁴-D-Iva ¹⁵. Herein, we describe the syntheses of the natural compound and its D-Iva¹⁴-L-Iva¹⁵ diastereomer, and the results of their chromatographic/mass spectrometry analyses. We present the conformational analysis of the two compounds and some of their synthetic in-termediates of different main-chain length in the crystal state (by X-ray diffraction) and in solvents of different polarities (using circular dichroism, FTIR absorption, and 2D NMR techniques), These data shed light on the mechanism of inhibition of HIV-1 integrase, which is an important target for anti-HIV therapy.

Full text of DOI:10.1002/chem.200900945

DOI: 10.1002/chem.200900945
Total Synthesis, Characterization, and Conformational Analysis of the
Naturally Occurring Hexadecapeptide Integramide A and a Diastereomer
Marta De Zotti,[a] Francesca Damato,[a] Fernando Formaggio,[a] Marco Crisma,[a]
Elisabetta Schievano,[a] Stefano Mammi,[a] Bernard Kaptein,[b] Quirinus B. Broxterman,[b]
Peter J. Felock,[c] Daria J. Hazuda,[c] Sheo B. Singh,[c] Jochen Kirschbaum,[d]
Hans Brꢀckner,[d] and Claudio Toniolo*[a]
Dedicated to the Centenary of the Italian Chemical Society and to the memory of Professor Ernesto Scoffone, founder of pep-
tide chemistry in Italy and a great mentor
Abstract: Integramide A is a 16-amino
acid peptide inhibitor of the enzyme
HIV-1 integrase. We have recently re-
ported that the absolute stereochemis-
tries of the dipeptide sequence near
their chromatographic/mass spectro-
metric analyses. We present the confor-
mational analysis of the two com-
pounds and some of their synthetic in-
termediates of different main-chain
length in the crystal state (by X-ray dif-
fraction) and in solvents of different
polarities (using circular dichroism,
FTIR absorption, and 2D NMR tech-
niques). These data shed light on the
mechanism of inhibition of HIV-1 inte-
grase, which is an important target for
anti-HIV therapy.
the
C
terminus are l-Iva14-d-Iva15.
Keywords: circular dichroism · IR
spectroscopy · NMR spectroscopy ·
peptides · X-ray diffraction
Herein, we describe the syntheses of
the natural compound and its d-Iva14-l-
Iva15 diastereomer, and the results of
Introduction
viral genome into the host cell chromosomes as part of its
infection cycle. Several steps in this integration process are
catalyzed by the enzyme HIV-1 integrase. The integration of
HIV-1 DNA into the host chromosome is achieved by the
integrase protein performing a series of DNA cutting and
joining reactions.
Like all retroviruses, human immunodeficiency virus type 1
(HIV-1) depends upon the integration of a DNA copy of its
[a] Dr. M. De Zotti, F. Damato, Prof. F. Formaggio, Dr. M. Crisma,
Dr. E. Schievano, Prof. S. Mammi, Prof. C. Toniolo
Institute of Biomolecular Chemistry, CNR, Padova Unit
Department of Chemistry University of Padova
via Marzolo 1, 35131 Padova (Italy)
Integrase is an attractive target for anti-HIV therapy be-
cause it is essential for virus replication and, unlike protease
and reverse transcriptase, there are no known counterparts
in the host cell. Furthermore, integrase takes advantage of a
single active site to accommodate two different conforma-
tions of DNA substrates, which may constrain the ability of
HIV to develop drug resistance to integrase inhibitors.[1]
Several classes of HIV-integrase inhibitors are known,[2] in-
cluding some non-a-aminoisobutyric acid (Aib)/isovaline
(Iva) peptides.[3] However, a recent publication from three
of the authors of this paper has dealt with two naturally oc-
curring, nonribosomal peptide molecules (integramides A
and B) that exhibit significant inhibitory activities.[4]
Fax : (+39)049-827-5239
E-mail: claudio.toniolo@unipd.it
[b] Dr. B. Kaptein, Dr. Q. B. Broxterman
DSM Pharmaceutical Products
Innovative Synthesis and Catalysis
P.O. Box 18, 6160 MD Geleen (The Netherlands)
[c] Dr. P. J. Felock, Dr. D. J. Hazuda, Dr. S. B. Singh
Merck Research Laboratories, Rahway
NJ 07065 and West Point, PA 19486 (USA)
[d] Dr. J. Kirschbaum, Prof. H. Brꢀckner
Department of Food Sciences
Integramides are linear peptides, 16 amino acids long and
rich in Aib and Iva residues. There are four Aib and five Iva
residues in the sequence of integramide A, and three Aib
and six Iva residues in integramide B; integramide A has an
Aib residue at position 8, whereas its B analogue has an l-
Interdisciplinary Research Centre for Biosystems
Land Use and Nutrition, University of Giessen
35392 Giessen (Germany)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200900945.
316
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 316 – 327

FULL PAPER
tion of the racemic a-amino amides, followed by optical res-
olution using a broadly specific a-amino amidase.
The segment condensation strategy adopted for the syn-
thesis of the two diastereomeric (l-Iva14-d-Iva15 and d-Iva14-
l-Iva15) hexadecapeptides in solution is illustrated in
Scheme 1. The three planned segments B, C, and D are of
similar length. The synthesis avoids AibACTHNUTRGNEUNG(Iva)-Hyp dipeptide
sequences (Hyp=(2S,4R)-4-hydroxyproline), which are acid
labile and prone to undergo intramolecular cyclization to
2,5-dioxopiperazine when located at the N terminus. In ad-
dition, epimerization in the coupling reaction is prevented
due to the absence of chiral Ca-trisubstituted amino acids in
the last and penultimate positions of the chain.[7] Indeed,
any C activation of amino acid derivatives and peptide seg-
ments, such as A, B, and C, with a C-terminal Ca-tetrasubsti-
tuted amino acid (Iva or Aib), is known to produce an oth-
erwise easily epimerizable 5(4H)-oxazolone intermediate to
a considerable extent.
For the formation of the difficult peptide bonds, in partic-
ular those connecting either two sterically demanding Ca-
tetrasubstituted amino acids (Aib/Iva) or one of these
amino acids followed by an N-alkylated Hyp residue,[7] the
carboxyl group of the Na-protected amino acid or peptide
was activated with the N-ethyl-N’-(3-dimethylaminopropyl)-
Iva residue in this position. In both integramides A and B,
the C terminus is free and the N terminus is acetylated
(Ac). In the initial work,[4] the stereochemical configuration
of the two consecutive C-terminal Iva14-Iva15 residues could
not be determined, although it was known that one is l and
the other is d. More recently, using HPLC and NMR tech-
niques and the two l-Iva14-d-Iva15 and d-Iva14-l-Iva15 syn-
thetic diastereomeric hexadecapeptides, we were able to elu-
cidate this originally unresolved stereochemical problem,
providing unambiguous evidence that the natural inhibitors
contain the l-Iva14-d-Iva15 chiral pair.[5] We also showed that
the stereochemical inversion in the Iva14- Iva15 natural se-
quence is in general not detrimental and might be even
slightly beneficial for activity against the strand-transfer re-
action.
Herein, we describe in detail the strategy we exploited in
solution for the challenging syntheses of the two diastereo-
meric hexadecapeptides, and their extensive characterization
using HPLC, chiral chromatography, and mass spectrometry.
In addition, the results of an in-depth conformational inves-
tigation of the two hexadecapeptides and some of their
short-sequence synthetic intermediates using X-ray diffrac-
tion, CD, FTIR absorption, and NMR techniques are pre-
sented. In our previous communication on these com-
pounds,[5] only the results of the configurational study were
reported; neither the synthetic approach and the related an-
alytical data nor the conformational findings were discussed.
A preliminary NMR spectroscopic conformational investiga-
tion of integramide A was already published,[4] but, as op-
posed to the helix-supporting solvent used in this work (deu-
terated 2,2,2-trifluoroethanol, [D2]TFE), the previous work
was performed in deuterated pyridine, which is not entirely
appropriate for such an analysis because of its known ten-
dency to interact with the hydrogen-bonding donor (pep-
tide) NH groups.
carbodiimide
(HOAt) method.[8] The poorly
reactive secondary alcohol
(EDC)/7-aza-1-hydroxy-1,2,3-benzotriazole
functionalities in the side chains
of the Hyp residues were used
without protection. The C-ter-
minal OtBu carboxylic ester
protection was removed by the
classical trifluoroacetic acid
method in segments B and C,
which lack any AibACTHUNRGTNE(NUG Iva)–Hyp
bond, whereas a mildly acidic
fluoroalcohol (1,1,1,3,3,3-hexa-
fluoroacetone hydrate) was
used to afford the final A–B–
C–D hexadecapeptide.[9]
The HPLC profiles shown in
Figure 1 demonstrate the purity
of the two Ac/OH hexadeca-
peptide diastereomers. As re-
ported in ref. [5], the chromato-
graphic behavior of natural in-
tegramide A is identical to that
Figure 1. HPLC profiles for
the two synthetic hexadeca-
peptide diastereomers l-Iva14-
d-Iva15 (l- d) and d-Iva14-l-
of the l-Iva14-d-Iva15 diastereo- Iva15 (d-l). Conditions: analyt-
ical Phenomenex Kromasil C18
mer, but differs significantly
column (particle size: 5 mm;
from that of the d-Iva14-l-Iva15
pore size: 100 ꢃ); gradient
Results and Discussion
diastereomer.
system: 65% to 80%
B in
A total acid hydrolyzate of
each of the two synthetic hexa-
decapeptide diastereomers was
derivatized with Marfeyꢂs re-
agent according to the litera-
30 min; eluant A, 9:1 0.05%
aqueous trifluoroacetic acid
(TFA)/CH3CN; eluant B: 9:1
CH3CN/ 0.05% aqueous TFA;
Synthesis and characterization: For the large-scale produc-
tion of enantiomerically pure l-Iva and d-Iva, an economi-
cally attractive and generally applicable chemoenzymatic
synthesis developed by DSM was used.[6] The method in-
volves a combination of organic synthesis for the prepara-
flow rate: 1 mLminꢀ1
; room
temperature; absorbance de-
ture procedure.[10] The l-Iva tector at 226 nm.
Chem. Eur. J. 2010, 16, 316 – 327
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
317

C. Toniolo et al.
Scheme 1. Strategy adopted for the synthesis of the two hexadecapeptide diastereomers Ac/OH A–B–C–D[Hyp2,9,13], (-d-Iva14-l-Iva15-) and (-l-Iva14-d-
Iva15-). i) H2/Pd in distilled CH2Cl2; ii) H2/Pd in MeOH; iii) EDC and HOAt (1.3 equiv each) in distilled CH2Cl2, keeping the pH at 8 by adding N-meth-
ylmorpholine (NMM); iv) trifluoroacetic acid (10 equiv) in CH2Cl2; v) EDC, HOAt, and Ac-d-Iva-OH (15 equiv each) in N,N-dimethylformamide
(DMF), keeping the pH at 8 by adding NMM; vi) 1,1,1,3,3,3-hexafluoroacetone hydrate. (EDC, N-ethyl-N’-(3-dimethylaminopropyl)carbodiimide,
HOAt, 7-aza-1-hydroxy-1,2,3-benzotriazole).
(eluted at 27.9 min) and d-Iva (eluted at 30.3 min) enantio-
mers were baseline resolved by HPLC. The average l-Iva/d-
Iva ratio obtained (two analyses for each sample) was 1.48
(ꢁ0.03) for the l-Iva14-d-Iva15 hexadecapeptide and 1.43 (ꢁ
0.03) for the d-Iva14-l-Iva15 hexadecapeptide.[4] These two
experimental values should be compared with the calculated
ratio (3 l-Iva and 2 d-Iva) of 1.50. A synthetic, racemic d,l-
Iva amino acid, used as a standard, gave the expected l/d
ratio (1.00).
Total acid hydrolyzates of natural integramide A and the
synthetic l-Iva14-d-Iva15 hexadecapeptide were separately
converted into Na-trifluoroacetylamino 2-propyl esters. The
resulting derivatives were resolved on a Chirasil-l-Val capil-
lary column and analyzed by selected-ion monitoring (SIM)
mass spectrometry. The presence of Aib, Gly, l-Leu, l-Ile,
and l-Hyp was detected. Moreover, the derivatives of l-Iva
and d-Iva, satisfactorily resolved, provided l-Iva/d-Iva ratios
of 1.48 (natural compound) and 1.42 (synthetic compound).
Trace amounts of d-Leu in the natural compound (1.8%)
and in the synthetic compound (3.8%) are attributed to
minor sequence-dependent racemization during acid hydrol-
ysis and epimerization during the coupling steps in the pep-
tide synthesis.
[Ac-Iva-Hyp-Ile3-Iva-Leu-Aib-Aib-Aib8]+. Ions b11 and b14
(m/z 1160.3 and 1457.3, respectively) are of very low abun-
dance, whereas ions b12 and b15 (m/z 1245.3 and 1556.3, re-
spectively) are remarkably intense. The high intensity of the
b8 and b12 ions is related to the preferred cleavage of the
Aib8–Hyp9 and Aib12–Hyp13 peptide bonds.[10] In the CID-off
mode, the fragment ions b9, b10, and b13 are not generated
(but they are formed when CID 50% is applied, not
shown). The a2 fragment ion (m/z 226.9) results from loss of
CO from the b2 ion. An ion denoted x4 at m/z 411.0 repre-
sents the protonated internal tetrapeptide fragment ion
Hyp9-Leu-Iva-Aib12. Loss of the C-terminal Aib is therefore
assumed to generate the ion at m/z 326.0. The ESI-CID
mass spectrum of the synthetic l-Iva14-d-Iva15 hexadecapep-
tide (Figure 2) and that of natural integramide A show
almost perfect agreement.
Crystal-state conformational analysis: Despite a number of
attempts, we were unable to grow a single crystal from
either of the two hexadecapeptide diastereomers. However,
we succeeded in preparing crystals suitable for X-ray dif-
fraction from four terminally protected segments of integra-
mide A or its d-Iva-l-Iva diastereomer: 1) Z-d-Iva-l-Iva-
Gly-OtBu (Z=benzyloxycarbonyl; OtBu=tert-butoxy), the
C-terminal tripeptide of segment D, already described in
ref. [11]; 2) Z-l-Hyp-l-Iva-d-Iva-Gly-OtBu, a tetrapeptide
that spans the full sequence of segment D; 3) Z-l-Iva-l-
Leu-(Aib)3-OtBu; and 4) Z-l-Ile-l-Iva-l-Leu-(Aib)3-OtBu,
the C-terminal penta- and hexapeptides of segment B. The
The electrospray ionization collision-induced dissociation
(ESI-CID) mass spectrum of the synthetic l-Iva14-d-Iva15
hexadecapeptide (Figure 2) shows the intense peak of the
sodium adduct of the molecular ion (m/z 1653.9), as well as
a regular series of b3–b8 acylium ions (m/z 368.0, 467.0,
580.0, 665.2, 750.2, and 835.3), resulting from the sequence
318
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 316 – 327

Synthesis of Integramide A
FULL PAPER
ration, adopt f,y torsion angles
falling in the left-handed helical
region of the Ramachandran
map. Two intermolecular hy-
drogen bonds involve the pep-
ꢀ
tide N2 H and side-chain Hyp
hydroxyl groups as donors, and
the peptide C2=O2 and C1=O1
as acceptors, respectively. The
-d-Iva-l-Iva- sequence in the
tripeptide
Z-d-Iva-l-Iva-Gly-
OtBu, shorter by one residue at
the N terminus, is also heli-
cal.[11] The structure of the
strictly related tetrapeptide Z-
l-Pro-l-Iva-d-Iva-Gly-OtBu
also shows a nonhelical confor-
mation, with two consecutive b-
turns of type II, III’.[11] These
findings are in line with the
well-known structural bias of
the l-Pro/l-Hyp residues for a
semi-extended (or poly-(Pro)n
Figure 2. Positive ion ESI-MS for the synthetic l-Iva14-d-Iva15 hexadecapeptide.
Table 1. Backbone torsion angles [8] for the peptides studied in this work
molecular structures of the last three peptides are illustrated
in Figures 3, 4, and 5. Their backbone torsion angles are
listed in Table 1.
Torsion Z-l-Hyp-l-Iva-d-
Z-l-Iva-l-Leu-
(Aib)3-OtBu
Z-l-Ile-l-Iva-l-Leu-
(Aib)3-OtBu
angle
Iva-Gly-OtBu
w0
f1
y1
w1
f2
y2
w2
f3
y3
w3
f4
y4
w4
f5
y5
w5
f6
y6
w6
ꢀ170.2(4)
ꢀ56.4(5)
132.5(4)
173.0(3)
55.3(5)
39.2(5)
169.4(4)
72.8(6)
28.2(7)
171.2(5)
ꢀ79.3(8)
171.6(5)[a]
178.0(5)[b]
–
ꢀ177.6(8)
ꢀ58.4(11)
ꢀ23.3(11)
178.0(8)
ꢀ64.6(10)
ꢀ10.9(11)
173.2(8)
ꢀ54.9(10)
ꢀ33.8(10)
ꢀ176.1(7)
ꢀ73.5(10)
ꢀ7.4(10)
ꢀ167.7(8)
50.9(11)
43.3(12)[c]
174.8(10)[d]
–
ꢀ165.6(8)
ꢀ62.3(11)
ꢀ48.0(11)
ꢀ173.1(7)
ꢀ56.2(11)
ꢀ30.8(12)
ꢀ175.3(8)
ꢀ58.8(11)
ꢀ28.1(12)
179.3(8)
ꢀ53.4(11)
ꢀ33.5(10)
ꢀ173.8(7)
ꢀ59.1(11)
ꢀ37.8(12)
ꢀ164.1(9)
46.4(13)
–
–
–
–
–
–
48.5(11)[e]
171.1(10)[f]
–
[a] N4-C4A-C4-OT. [b] C4A-C4-OT-CT1. [c] N5-C5A-C5-OT. [d] C5A-
C5-OT-CT1. [e] N6-C6A-C6-OT. [f] C6A-C6-OT-CT1.
Figure 3. X-ray diffraction structure of Z-l-Hyp-l-Iva-d-Iva-Gly-OtBu
with atom numbering. Hydrogen atoms were omitted for clarity. Dashed
ꢀ
lines represent intramolecular C=O···H N hydrogen bonds.
II) conformation[14] and of the Iva helical screw-sense indif-
ference, due to the small variation in length between its two
side chains.[15]
The pentapeptide molecule (Figure 4), particularly rich in
helicogenic Aib residues,[16] is highly folded. The l-Iva-l-
Leu-Aib-Aib segment is stabilized by three consecutive C=
In the structure of the tetrapeptide (Figure 3), two C=
ꢀ
O···H N intramolecular hydrogen bonds are seen between
ꢀ
the peptide N3 H and urethane C0=O0 groups and between
ꢀ
the peptide N4 H and peptide C1=O1 groups. Both hydro-
gen bonds are weak[12] (Table S1 in the Supporting Informa-
tion). The N-terminal-l-Hyp-l-Iva b-turn structure[13] is non-
helical type II, whereas the next one (-l-Iva-d-Iva-) is heli-
cal type III’. Both Iva residues, independent of their configu-
ꢀ
ꢀ
O···H N intramolecular hydrogen bonds (C0=O0···H N3,
ꢀ
ꢀ
C1=O1···H N4, C2=O2···H N5), generating a full turn of a
310-helix.[17] However, this right-handed structure is not regu-
Chem. Eur. J. 2010, 16, 316 – 327
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
319

C. Toniolo et al.
ꢀ
ꢀ
ꢀ
bond. The hydrogen bonds involve the N3 H, N4 H, N5
ꢀ
H, and N6 H groups as donors and the C0=O0, C1=O1,
C2=O2, and C3=O3 groups as acceptors, respectively. Only
the N-terminal hydrogen bond is rather weak. Also in this
compound, we observe a screw-sense inversion for the C-ter-
minal, helical Aib residue. Two intermolecular hydrogen
ꢀ
bonds (C4=O4···H N1 and C5=O5···H-N2), the latter quite
weak, stabilize the head-to-tail packing mode of the helical
molecules in the crystal.
From this crystal-state conformational analysis, it seems
reasonable to conclude that most of the N-terminal hepta-
peptide segment B of integramide A possesses a strong pro-
pensity for a regular, right-handed, 310-helical structure. On
the other hand, the C-terminal segment D is not helical as
such, most probably because of the presence of a Janus-
headed Hyp residue located at position 1. However, it
would not be surprising to see this segment as well in a reg-
ular folded conformation when incorporated at the C termi-
nus of a pre-existing helical stretch.
Figure 4. X-ray diffraction structure of Z-l-Iva-l-Leu-(Aib)3-OtBu with
atom numbering. Only the major occupancy site of the l-Iva Cg atom is
shown. Hydrogen atoms were omitted for clarity. Dashed lines represent
ꢀ
intramolecular C=O···H N hydrogen bonds.
Solution conformational analysis: We investigated the con-
formational preferences in solution for the two hexadeca-
peptide diastereomers (l-Iva14-d-Iva15 and d-Iva14-l-Iva15)
and for natural integramide A under different solvent and
temperature conditions by means of CD, FTIR absorption
and 2D-NMR spectroscopic techniques.
The CD spectra of the two synthetic compounds in
MeOH, 2,2,2,-trifluoroethanol (TFE), and a 100 mm aque-
ous solution of sodium dodecylsulfate (SDS) are shown in
Figure 6. In Figure 7, the two curves in TFE are compared
with that of natural integramide A in the same fluoroalco-
hol. By changing the solvent, no dramatic alterations are
seen in the CD patterns, which highlights the remarkable
secondary structure stability of the two compounds. This
conclusion is corroborated by the absence of any variation
in the spectra by heating from 5 to 558C in MeOH (not
shown). The shapes of the CD patterns clearly indicate that
both compounds fold in a right-
lar, in particular at the level of the l-Leu2 and Aib4 residues,
in which the corresponding y torsion angles are well below
the typical ꢀ308 value, generating type I instead of type III
b-turns. Nonetheless, the intramolecular hydrogen bonds ex-
ꢀ
hibit normal N···O distances and N H···O angles. The typi-
cal screw-sense inversion with respect to the preceding resi-
dues in the backbone, usually observed at the C terminus of
a
310-helical peptide for a
Ca-tetrasubstituted a-amino
ester,[16] is also noted in this compound. The pentapeptide
molecules are held together head-to-tail through a C4=
ꢀ
O4···H N1 intermolecular hydrogen bond.
At partial variance with its pentapeptide synthetic precur-
sor, the hexapeptide (Figure 5) adopts a regular, right-
handed, 310-helical structure in the crystalline state
(Table 1). Four consecutive, type III b-turns are observed,
ꢀ
each stabilized by a C=O···H N intramolecular hydrogen
handed helical structure of the
mixed 310-/a-type (in MeOH,
the 310-helix seems to prevail,
whereas in SDS the a helix
tends to be more populated).
This information was extracted
from the occurrence of two
negative maxima near 205 and
225 nm and one positive maxi-
mum at about 195 nm, and
225
from the ellipticity ratio [V]R
/
[V]R205, which is known to be
less than 0.50 for a high popula-
tion of 310-helix and more than
0.60 for a high population of a
helix.[18] In general, the highest
ellipticity and, as
quence, the most significant
percentage of a helix is seen in
a conse-
Figure 5. X-ray diffraction structure of Z-l-Ile-l-Iva-l-Leu-(Aib)3-OtBu acetonitrile hemisolvate with atom
ꢀ
numbering. Hydrogen atoms were omitted for clarity. Dashed lines represent intramolecular C=O···H N hy-
drogen bonds.
320
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 316 – 327

Synthesis of Integramide A
FULL PAPER
Figure 7. CD spectra for the two synthetic hexadecapeptide diastereo-
mers l-Iva14-d-Iva15 (l-d) and d-Iva14-l-Iva15 (d-l), and natural integrami-
de A (N) in TFE. Peptide concentration: 1 mm.
from concentrations of 1.0 to 0.1 mm (except for segment C,
but only below 3320 cmꢀ1). This finding suggests that almost
ꢀ
all observed C=O···H N hydrogen bonding is intramolecu-
lar. The ratio between the areas of the band at 3350 cmꢀ1
(associated with hydrogen-bonded NH groups[20]) and the
band(s) above 3400 cmꢀ1 (associated with free, solvated NH
groups) is much lower for segment C, indicating that the
latter peptide exhibits a remarkably reduced tendency to
fold, probably related not only to its short main-chain
length, but to its specific sequence as well, because it lacks
any helicogenic Ca-methylated a-amino acid in the N-termi-
nal doublet. This structural information agrees with that
found from the X-ray diffraction analysis described above.
Consistent with this result, in the two longer segments C–D
and B–C–D (spectra not shown), and in the two Ac/OtBu-
protected hexadecapeptide diastereomers (Figure 8), all con-
taining the C segment, we found that the band at 3350 cmꢀ1
is significantly broadened, which implies some conforma-
tional heterogeneity in this part of the molecule. Because
the FTIR spectra of the two hexadecapeptide diastereomers
are very similar, this spectroscopic technique does not allow
any specific assignment. In these two FTIR spectra, the ex-
tremely weak band above 3400 cmꢀ1 is indicative of the oc-
currence of highly folded species. At concentrations above
1.0 mm, the curves of all Z-protected synthetic intermediates
do not suggest the onset of any significant amount of self-
aggregated species, in contrast with those of the two Ac-
blocked hexadecapeptide diastereomers, which are strongly
indicative of self-aggregation (spectra not shown). Unfortu-
nately, we could not record the spectra of the two final, syn-
thetic Ac/OH hexadecapeptides due to their very low solu-
bilities in CDCl3.
Figure 6. CD spectra for the two synthetic hexadecapeptide diastereo-
mers l-Iva14-d-Iva15 (A) and d-Iva14-l-Iva15 (B) in MeOH (M), TFE (T),
and a 100 mm aqueous solution of SDS. Peptide concentration: 1 mm.
the structure-supporting solvent TFE[19] for both compounds.
The CD patterns of the two hexadecapeptide diastereomers
are similar to each other and to that of natural integrami-
de A, suggesting that the chirality inversion in the Iva14-Iva15
dipeptide segment does not induce any significant change in
the overall secondary structure of the peptide. However,
only the CD curve of the l-Iva14-d-Iva15 peptide matches
almost perfectly that of natural integramide A. Finally, the
CD curves of the Z/OtBu protected pentadecapeptides and
of the Ac/OtBu-protected hexadecapeptide synthetic precur-
sors (not shown) nearly overlap with those of the two final
diastereomeric compounds, showing that the observed 310-/
a-helical conformation is already attained at the pentadeca-
peptide level and also before removal of the OtBu C-termi-
nal protecting group.
ꢀ
Figure 8 reports the conformationally informative N H
stretching region of the FTIR absorption spectra of the
short, Z/OtBu-protected segments B, C, and D (the latter
for the l-Iva14-d-Iva15 diastereomer) in CDCl3 at two con-
centrations. The curves change only slightly on changing
Chem. Eur. J. 2010, 16, 316 – 327
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
321

C. Toniolo et al.
bonds was detected. The 1D
1H NMR spectrum does not
show minor peaks and no se-
quential CaH–CaH NOESY
cross-peaks were found. All
ꢀ
NH NH sequential connectivi-
ties are present, as well as some
ꢀ
NH(i) NH
N
These results are consistent
with a helical structure, which is
not interrupted by the Hyp resi-
dues, as also confirmed by the
presence in the NOESY spec-
trum of all of the connectivities
between the Hyp CdH protons
and the NH proton of both the
preceding and the following
a
ꢀ
residues. A number of C H(i)
a
ꢀ
NH
G
C H(i) NH
N
a
ꢀ
and C H(i) NH
(i+4) cross
peaks were observed. Together,
these data support the presence
of a mixed 310-/a-helical confor-
mation.
ꢀ
Figure 8. FTIR absorption spectra (N H stretching region) of the Z/OtBu protected segments B (A), C (B),
and the l-Iva-d-Iva diastereomer of D (C) at the concentrations 1 mm (1.0) and 0.1 mm (0.1). D) The corre-
sponding spectra for the two synthetic, Ac/OtBu protected hexadecapeptide diastereomers l-Iva14-d-Iva15 (l-
d) and d-Iva14-l-Iva15 (d-l) at 1 mm concentration. Solvent: CDCl3.
The two methyl groups in the
Aib residues belonging to chiral
A more detailed conforma-
tional characterization of inte-
gramide A was carried out by
using 2D NMR spectroscopy.
This study was performed in
[D2]TFE solution, in which the
peptide exhibits a high propen-
sity to fold into a helical struc-
ture as indicated by CD. The
spin systems of the Gly, Leu,
Hyp, and Ile residues were
identified by using DQF-COSY
and TOCSY spectra, whereas
HMQC and HMBC experi-
ments were used for the Aib
and Iva residues. The sequential
assignment
was
performed
using NOESY spectra. Rele-
vant regions of the NOESY
spectrum at tm =250 ms are
shown in Figures 9 and 10. The
proton chemical shift assign-
ment is reported in Table S2 in
the Supporting Information. A
summary of the NOE connec-
tivities is shown in Figure 11.
Despite the presence of three
Hyp residues, no trace of cis
Figure 9. Amide region of the NOESY spectrum (600 MHz, tm =250 ms) of integramide A (1.45 mm in
configuration at the Xxx–Hyp [D2]TFE; 300 K). Medium-range (i!i+2) interactions are marked.
322
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 316 – 327

Synthesis of Integramide A
FULL PAPER
peptides are diastereotopic.
Consequently, the carbon atoms
of the two methyl groups (la-
beled as b and b’) are expected
to resonate at two different
chemical shifts, above and
below d=25 ppm (Figure S1 in
the Supporting Information).
Jung et al.[21] have shown that
the presence of a chiral center
adjacent to two gem-methyl
groups (as in an Aib residue)
induces a chemical shift differ-
ence between these prochiral
methyl carbons not higher than
0.5 ppm. However, if the ob-
served
difference
chemical
in
the
shifts
13C NMR
(termed chemical nonequiva-
lence, CNE) of the two prochi-
ral methyl carbons is 2 ppm or
higher, this indicates the pres-
ence of a stable helical confor-
mation. In Table S3 in the Sup-
porting Information, the CNE
values for the four Aib residues
of integramide A are reported.
All values are higher than
2 ppm, confirming the occur-
rence of a helical structure in
TFE. The stereospecific assign-
ment of the two diastereotopic
Figure 10. Fingerprint region of the NOESY spectrum (600 MHz, tm =250 ms) of integramide A (1.45 mm in
[D2]TFE; 300 K). Medium-range (i!i+2, i!i+3, and i!i+4) interactions are marked.
Figure 11. Summary of the NOESY connectivities for integramide A (1.45 mm in [D2]TFE; 300 K). Peaks are grouped into three classes based upon their
integrated volumes.
Chem. Eur. J. 2010, 16, 316 – 327
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
323

C. Toniolo et al.
Table 3. Average values [8] of torsion angles fm and ym and the relative
standard deviations resulting from the 33 calculated structures (energy
<300 kcalmolꢀ1) of integramide A.
methyl groups was achieved through the Cb-selective
HMQC spectrum, using the method reported by Bellanda
et al.[22]
The conformational properties of integramide A were fur-
ther investigated by distance geometry and restrained mo-
lecular dynamics (MD) calculations. A total of 99 interpro-
ton distance restraints were derived from the NOESY spec-
trum (Table 2) and used in the simulated annealing protocol.
Residue
fm
Df
ym
Dy
d-Iva1
l-Hyp2
l-Ile3
–
–
ꢀ42.6
3.3
ꢁ0.5
ꢁ1.6
ꢁ1.7
ꢁ0.6
ꢁ0.8
ꢁ0.7
ꢁ1.4
ꢁ1.1
ꢁ3
ꢀ84.0
ꢁ1.1
ꢀ80
ꢁ2
ꢀ21.4
ꢀ27.1
ꢀ25.4
ꢀ43.0
ꢀ23.4
ꢀ61.5
ꢀ27
l-Iva4
l-Leu5
Aib6
ꢀ73.3
ꢀ81.5
ꢀ88.3
ꢀ73.5
ꢀ68.3
ꢀ70.6
ꢀ72.7
ꢀ82
ꢁ0.9
ꢁ1.4
ꢁ0.9
ꢁ0.6
ꢁ1.7
ꢁ0.6
ꢁ1.9
ꢁ3
Aib7
Aib8
Table 2. NOE constraints, deviations from idealized geometry, and mean
energies for the NMR-based structure of integramide A.
l-Hyp9
l-Leu10
l-Iva11
Aib12
ꢀ29.6
ꢀ40.4
ꢀ53.1
ꢀ44.3
ꢀ22.1
ꢀ78.7
–
ꢁ1.5
ꢁ1.0
ꢁ0.7
ꢁ1.7
ꢁ0.6
ꢁ0.7
–
number of NOEs
total
99
32
46
21
ꢀ34.2
ꢀ72.8
ꢀ50.4
ꢀ84.1
ꢀ29.6
ꢁ1.3
ꢁ0.6
ꢁ0.8
ꢁ2.8
ꢁ1.5
l-Hyp13
l-Iva14
d-Iva15
Gly16
intraresidue
sequential
i,i+n, n=2, 3, 4
mean rmsd[a] from ideality of accepted structures
bonds [ꢃ]
angles [8]
0.0097
1.13
improper [8]
NOEs [ꢃ]
37.40
0.19
mean energies [kcalmolꢀ1] of accepted structures
Eoverall
Ebond
Eangle
ENOE
299.1
23.7
90.9
165.7
[a] Root-mean-square deviation.
Out of the 150 structures that were generated, 100 had vio-
lations of the NOE restraints lower than 0.5 ꢃ. The 33 struc-
tures with a total energy less than 300 kcalmolꢀ1 were se-
lected, and their superposition is shown in Figure 12. All of
Figure 13. Representation of the 3D structure with the lowest energy ob-
tained for integramide A. The three l-Hyp residues are labeled.
Conclusion
Our interest in the study of integramides originated not only
from their intrinsic activity as effective inhibitors of HIV-1
integrase,[4] but also from the observation that their primary
structures are closely related to those of peptaibols[23]/pep-
taibiotics,[7c,24] a class of naturally occurring peptides that are
characterized by a high occurrence of the strongly 310/a-heli-
cogenic, Ca-methylated a-amino acids Aib[16] and Iva.[15]
Although synthetic methods for peptides based on steri-
cally demanding Ca-methylated a-amino acids have been al-
ready described,[7] the total syntheses of integramide A, a
peptide comprising 16 amino acids, and its d-Iva14-l-Iva15
diastereomer, are notable achievements. Indeed, these com-
pounds present some remarkable challenges due to the steri-
cally hindered nature of multiple amino acids, the acid sensi-
tivity of three tertiary amide bonds, and the propensity of a
few N-terminal dipeptides to cyclize, with the related poten-
tial loss of two residues from the sequence.
Figure 12. Backbone representation of the 33 structures with energy
<300 kcalmolꢀ1 resulting from the MD calculations of integramide A
with the backbone atoms superimposed.
these structures converge to a well-defined, mixed 310-/a-he-
ACHTUNGTRENNUNGlical conformation throughout the sequence, with a back-
bone average pairwise root-mean-square deviation of
(0.26ꢁ0.10) ꢃ (Table 3). The lowest-energy 3D structure,
shown in Figure 13, exhibits a clear amphipathic character
with the three l-Hyp residues at positions 2, 9, and 13 form-
ing the hydrophilic face.
Our very detailed, 3D structural analysis on integrami-
de A and selected short sequences clearly supports the view
that the peptide inhibitor has a predominantly 310-/a-helical
structure with amphipathic features under the variety of ex-
324
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 316 – 327

Synthesis of Integramide A
FULL PAPER
drogen atoms were refined anisotropically. Hydrogen atoms were calcu-
lated at idealized positions and refined using a riding model.
perimental conditions used. Notably, in the C-terminal half
of the molecule, the AibACTHNUTRGNEU(GN Iva) backbone constraints out-
weigh the Hyp conformational preference for the semi-ex-
tended type II poly-(l-Pro)n conformation. This conclusion
agrees well with literature reports on model peptides and on
Z-l-Hyp-l-Iva-d-Iva-Gly-OtBu. Formula: C29H44N4O8; formula weight:
576.7; orthorhombic, space group P212121; unit cell parameters: a=
9.440(2), b=11.768(2), c=29.962(4) ꢃ; V=3328.5(10) ꢃ3; Z=4; 1calcd
=
1.151 mgmꢀ3; crystal size: 0.40ꢄ0.15ꢄ0.07 mm3; data/parameters: 3369/
359; R1 =0.059 [on Fꢂ4s(F)]; wR2 =0.176 (on F2, all data); goodness of
fit on F2: 0.951; largest peak and hole in the final difference Fourier map:
naturally occurring peptaibiotics with repeating Aib
ACHTUNGTRENN(UNG Iva)-l-
Pro
ACHTUNGTRENNUNG
0.264 and ꢀ0.188 eꢃꢀ3
.
acterized by a b-turn ribbon structure in which the regulari-
ty of the 310-helical backbone f,y torsion angles is preserved
Z-l-Iva-l-Leu-(Aib)3-OtBu. Formula: C35H57N5O8; formula weight:
675.9; monoclinic, space group P21; unit cell parameters: a=10.096(2),
b=17.798(3), c=11.554(2) ꢃ; b=105.28(5)8; V=2002.7(6) ꢃ3; Z=2;
1calcd =1.121 mgmꢀ3; crystal size: 0.45ꢄ0.40ꢄ0.10 mm3; data/restraints/pa-
rameters: 3279/42/430; R1 =0.141 [on Fꢂ4s(F)]; wR2 =0.334 (on F2, all
data); goodness of fit on F2: 1.696; largest peak and hole in the final dif-
ference Fourier map: 0.530 and ꢀ0.630 eꢃꢀ3. The Cg atom of the l-Iva
residue is disordered and was refined on two sites with population pa-
rameters 0.55 and 0.45, respectively.
despite a substantial presence of Pro
more than compensated for by the extremely strong folding
propensity of the preceding Aib(Iva) residues. Remarkably,
ACHTUNGERTN(NUNG Hyp) residues, which is
AHCTUNGTRENNUNG
in a b-turn ribbon segment, the orientation of the peptide
carbonyls with respect to the helix axis is not significantly
modified compared to that of these same groups in the 310-
helix, but some of these carbonyl groups cannot form hydro-
Z-l-Ile-l-Iva-l-Leu-(Aib)3-OtBu acetonitrile hemisolvate. Formula:
1
gen bonds with the N-alkylated ProACTHNUTRGNEUNG(Hyp) residues. As a
C41H68N6O9
ꢄ
=
CH3CN; formula weight: 809.5; orthorhombic, space
2
result, these C=O groups are available to interact with hy-
drogen-bonding donor solvents or with other surrounding
peptide (protein) molecules.
group C2221; unit cell parameters: a=13.580(2), b=19.486(3), c=
38.283(4) ꢃ; V=10130(2) ꢃ3; Z=8; 1calcd =1.062 mgmꢀ3; crystal size:
0.30ꢄ0.20ꢄ0.07 mm3; data/restraints/parameters: 4130/44/519; R1 =0.082
[on Fꢂ4s(F)]; wR2 =0.233 (on F2, all data); goodness of fit on F2: 0.926;
largest peak and hole in the final difference Fourier map: 0.332 and
Several approaches have been used to identify peptides
that inhibit HIV-1 integrase.[3] Of particular significance are
the conclusions of Roques and co-workers,[3c] who isolated a
dodecapeptide inhibitor (EBR28) that binds tightly to the
enzyme. An NMR spectroscopic structure analysis showed
that under favorable experimental conditions this peptide
adopts an a-helical conformation with amphipathic proper-
ties. Further evidence suggested that the peptide binds the
integrase catalytic core, thus impairing the enzyme dimeriza-
tion motif that is essential for its activity. Moreover, the hy-
drophobic face of the a helix seems implicated in this inter-
action. We believe that the hexadecapeptide main-chain
length, the stable helical properties, and the strongly amphi-
pathic nature of integramides might mimic the correspond-
ing characteristics of EBR28 and its proposed mechanism of
integrase inhibition. It is also our view that X-ray diffraction
analyses of peptide inhibitor–integrase complexes can be ex-
tremely useful to elucidate the mechanism of action and op-
timize drug candidates that target integration of HIV. The
low conformational flexibility shown by integramide A, typi-
cal of the highly crystalline Aib/Iva-rich peptides,[15,16] will
facilitate such analyses.
ꢀ0.264 eꢃꢀ3. The cocrystallized acetonitrile molecule occupies a (1= , y,
2
1= ) special position and shows orientational disorder.
4
CCDC-726112, 726113, and 726114 contain the supplementary crystallo-
graphic data for this paper. These data can be obtained free of charge
from The Cambridge Crystallographic Data Centre via www.ccdc.cam.a-
c.uk/data_request/cif.
Circular dichroism: The CD spectra were obtained on a Jasco (Tokyo,
Japan) J-715 spectropolarimeter. Cylindrical fused quartz cells (Hellma)
of 0.1 mm path length were used. The values are expressed in terms of
[V]R, residue molar ellipticity (degcm2 dmolꢀ1). Spectrograde MeOH and
TFE (Acros, Geel, Belgium) were used as solvents.
Infrared absorption: The FTIR absorption spectra were recorded by
using a Perkin–Elmer 1720 X FTIR spectrophotometer, nitrogen flushed,
equipped with a sample-shuttle device, at 2 cmꢀ1 nominal resolution,
averaging 100 scans. Cells with path lengths of 0.1, 1.0, and 10 mm (with
CaF2 windows) were used. Spectrograde deuterochloroform (99.8% D)
was purchased from Aldrich (St. Louis, MO). Solvent (baseline) spectra
were recorded under the same conditions.
Chiral liquid chromatography: total hydrolysis:[10] Synthetic l-Iva14-d-
Iva15 hexadecapeptide or its d-Iva14-l-Iva15 diastereomer (0.48 mg) was
hydrolyzed in a 6m aqueous solution of HCl (0.5 mL) for 18 h at 1208C,
and evaporated to dryness. Derivatization: 1m aqueous solution of
NaHCO3 (100 mL) and Marfeyꢂs reagent (Na-(2,4-dinitro-5-fluorophenyl)-
l-alanine amide, FDNP-l-Ala-NH2, Sigma, St. Louis, MO) (1% in ace-
tone, 100 mL) were added. The mixture was sonicated and incubated at
408C for 1 h in a closed vial. A 1m aqueous solution of HCl (100 mL) and
dimethylsulfoxide (200 mL) were added, and aliquots of 5 mL injected
into the HPLC column. Instruments: an HP model 1100 apparatus with a
quaternary pump (Hewlett–Packard, Waldbronn, Germany) and a photo-
diode array detector set at 340 nm; LiChroCART Superspher 60 RP-
Select B column (250 mm ꢄ 4 mm ID; 4 mm particle size; Merck, Darm-
stadt, Germany). Gradient elution: eluant A: 50 mm triethylammonium
phosphate buffer, pH 3.0; eluant B: acetonitrile (100%); gradient: 15%
Experimental Section
Synthesis and characterization of peptides: The strategy of synthesis of
the peptides discussed in this work is illustrated in Scheme 1. Characteri-
zation details are reported in the Supporting Information.
B to 60% B in 75 min; flow rate: 1.2 mLminꢀ1
.
X-ray diffraction: Single crystals of Z-l-Hyp-l-Iva-d-Iva-Gly-OtBu and
Z-l-Iva-l-Leu-(Aib)3-OtBu were grown from ethyl acetate/petroleum
ether solution, whereas crystals of Z-l-Ile-l-Iva-l-Leu-(Aib)3-OtBu were
grown by slow evaporation from acetonitrile solution. Diffraction data
were collected at T=293(2) K with CuKa radiation (l=1.54178 ꢃ) using
a Philips PW 1100 diffractometer in the q–2q scan mode up to q=608.
The structures were solved by direct methods with the SIR 2002 progra-
m.[26a] Refinements were carried out by least-squares procedures on F2,
using all data, by application of the SHELXL97 program.[26b] All non-hy-
Chiral gas chromatography: total hydrolysis: Natural integramide A[4] or
the synthetic l-Iva14-d-Iva15 hexadecapeptide (0.1 mg) was hydrolyzed in
a 6m aqueous solution of HCl (0.5 mL) for 18 h at 1208C and evaporated
to dryness. Then, HCl in 2-propanol (0.5 mL; generated from a mixture
of AcCl/2-propanol 2:8 v/v) was added and the esterification was per-
formed for 1 h at 1008C in a closed vial. Solvents were removed in a
stream of nitrogen, CH2Cl2 (250 mL) and trifluoroacetic anhydride
(50 mL) were added, and the mixture heated at 1008C for 20 min. Sol-
Chem. Eur. J. 2010, 16, 316 – 327
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
325

C. Toniolo et al.
vents were removed in a stream of nitrogen, CH2Cl2 (200 mL) was added,
and 5 mL aliquots were injected into the gas chromatographic (GC)
column. Instruments: a GC–MS apparatus, with an A17 GC and a Shi-
madzu 5000 MS, was used. For chiral analysis of amino acid derivatives, a
Chirasil-l-Val capillary column (25 m ꢄ 0.35 mm ID; Varian, Darmstadt,
Germany) and helium as carrier gas were employed. Solvents and re-
agents were from Merck.
[1] Y. Goldgur, R. Craigie, G. H. Cohen, T. Fujiwara, T. Yoshinaga, T.
Fujishita, H. Sugimoto, T. Endo, H. Murai, D. R. Davies, Proc. Natl.
Acad. Sci. USA 1999, 96, 13040–13043.
[2] a) R. Craigie, J. Biol. Chem. 2001, 276, 23213–23216; b) S. P. Singh,
F. Pelaez, D. J. Hazuda, R. B. Lingham, Drugs Future 2005, 30, 277–
299; c) R. Dayam, R. Gundla, L. Q. Al-Mawsawi, N. Neamati, Med.
Res. Rev. 2008, 28, 118–154; d) T. A. Prikazchikova, A. M. Sychova,
Yu. Agapkina, D. A. Alexandrov, M. B. Gottikh, Russian Chem.
Rev. 2008, 77, 421–434.
[3] a) R. A. Puras Lutzke, N. A. Eppens, P. A. Weber, R. A. Houghten,
R. H. A. Plasterk, Proc. Natl. Acad. Sci. USA 1995, 92, 11456–
11460; b) R. G. Maroun, D. Krebs, M. Rashani, H. Porumb, C. Au-
clair, F. Troalen, S. Fermandjian, Eur. J. Biochem. 1999, 260, 145–
155; c) V. R. de Soultrait, A. Caumont, V. Parissi, N. Morellet, M.
Ventura, C. Lenoir, S. Litvak, M. Fournier, B. Roques, J. Mol. Biol.
2002, 318, 45–48; d) L. Zhao, M. K. OꢂReilly, M. D. Shultz, J.
Chmielewski, Bioorg. Med. Chem. Lett. 2003, 13, 1175–1177; e) K.
Krajewski, C. Marchand, Y.-Q. Long, Y. Pommier, P. P. Roller,
Bioorg. Med. Chem. Lett. 2004, 14, 5595–5598; f) I. Oz Gleenberg,
O. Avidan, Y. Goldgur, A. Herschhorn, A. Hizi, J. Biol. Chem. 2005,
280, 21987–21996; g) H.-Y. Li, Z. Zawahir, L.-D. Song, Y.-Q. Long,
N. Neamati, J. Med. Chem. 2006, 49, 4477–4486; h) P. Cherepanov,
A. L. B. Ambrosio, S. Rahman, T. Ellenberger, A. Engelman, Proc.
Natl. Acad. Sci. USA 2005, 102, 17308–17313; i) A. Armon-Omer,
A. Levin, Z. Hayouka, K. Butz, F. Hoppe-Seyler, S. Loya, A. Hizi,
A. Friedler, A. Loyter, J. Mol. Biol. 2008, 376, 971–982.
Mass spectrometry:[10b] For ESI-CID-MS measurements,
a calibrated
LCQ-MSTM (Thermo Quest, Finnigan MAT, San Josꢅ, CA) was used. N2
served as sheath gas and auxiliary gas. He (purity>99.9990%, Messer-
Griesheim, Krefeld, Germany) was used as collision gas. The tempera-
ture of the heated capillary was 2508C. Sheath gas and auxiliary gas were
set at 40 and 10 relative units, respectively. Ion source CID–MS was per-
formed at 0 and 45% relative collision energies. Sequence analyses were
carried out in the positive ionization mode. The m/z values were record-
ed in the centroid mode and have an accuracy of ꢁ0.5 Da.
NMR spectroscopy: NMR experiments were carried out on a Bruker
AVANCE DMX-600 spectrometer. The peptide concentration was
1.45 mm in deuterated [D2]TFE. The alcohol OH signal was suppressed
by presaturation during the relaxation delay. All homonuclear spectra
were acquired by collecting 512 experiments, each one consisting of 64
scans and 2K data points. The spin systems of protein amino acid resi-
dues were identified by using standard DQF-COSY[27] and CLEAN-
TOCSY[28] spectra. In the latter case, the spin-lock pulse sequence was
70 ms long. The assignment of methyl groups belonging to the same Aib
1
residue was obtained by means of 2D H–13C correlation spectra. To opti-
[4] S. B. Singh, K. Herath, Z. Guan, D. L. Zink, A. W. Dombrowski,
J. D. Polishook, K. C. Silverman, R. B. Lingham, P. J. Felock, D. J.
Hazuda, Org. Lett. 2002, 4, 1431–1434.
[5] M. De Zotti, F. Formaggio, B. Kaptein, Q. B. Broxterman, P. J.
Felock, D. J. Hazuda, S. B. Singh, H. Brꢀckner, C. Toniolo, Chem-
BioChem 2009, 10, 87–90.
[6] T. Sonke, B. Kaptein, W. H. J. Boesten, Q. B. Broxterman, H. E.
Schoemaker, J. Kamphuis, F. Formaggio, C. Toniolo, F. P. J. T. Rutjes
in Stereoselective Biocatalysis (Ed.: R. N. Patel), Marcel Dekker,
New York, 1999, pp. 23–58.
mize the digital resolution in the carbon dimension, HMQC and HMBC
spectra were acquired using selective excitation by means of Gaussian-
shaped pulses with 1% truncation.[29]
The Cb-selective HMQC spectra with gradient coherence selection[30]
were recorded with 224 t1 increments of 270 scans and 2K points each.[31]
A spectral width of 16 ppm centered at 22 ppm in F1 was used, yielding a
digital resolution of 2.36 Hz/pt prior to zero filling. HMBC experi-
ments[32] with selective excitation in the carbonyl region were performed
by using a long-range coupling constant of 7.5 Hz, a spectral width in F1
of 15 ppm centered at 176 ppm, 250 t1 experiments of 800 scans, and 4K
points in F2. The digital resolution in F1, prior to zero filling, was 2.2 Hz/
pt. NOESY experiments were used for sequence-specific assignment. To
avoid the problem of spin diffusion, the buildup curve of the volumes of
NOE cross-peaks as a function of mixing time (50 to 500 ms) was deter-
mined first (data not shown). The mixing time of the NOESY experi-
ments used for interproton distance determination ranged from 200 to
250 ms, that is, in the linear part of the NOE buildup curve. Interproton
distances were obtained by integration of the NOESY spectra using the
XEASY[33] package. The calibration was based on the average of the in-
tegration values of the cross peaks due to the interactions between the
two b-geminal protons of the Leu side-chain residues and the two d-
geminal protons of the Hyp side-chain residues, set to a distance of
1.8 ꢃ. When peaks could not be integrated because of partial overlap, a
distance corresponding to the maximum limit of detection of the experi-
ment (4.0 ꢃ) was assigned to the corresponding proton pair.
[7] a) F. Formaggio, Q. B. Broxterman, C. Toniolo in Houben-Weyl:
Methods of Organic Chemistry, Synthesis of Peptides and Peptidomi-
metics, Vol. E22c (Eds.: M. Goodman, A. Felix, L. Moroder, C. To-
niolo), Thieme, Stuttgart, 2003, pp. 292–310; b) H. Brꢀckner, M.
Currle in Second Forum Peptides (Eds.: A. Aubry, M. Marraud, B.
Vitoux), Libbey, London, 1989, pp. 251–255; c) Peptaibiotics:
Fungal Peptides Containing a-Dialkyl a-Amino Acids (Eds.: C. To-
niolo, H. Brꢀckner), Verlag Helvetica Chimica Acta, Zꢀrich, 2009.
[8] L. A. Carpino, J. Am. Chem. Soc. 1993, 115, 4397–4398.
[9] A. Moretto, M. Crisma, F. Formaggio, B. Kaptein, Q. B. Broxter-
man, T. A. Keiderling, C. Toniolo, Biopolymers 2007, 88, 233–238.
[10] a) R. Bhushan, H. Brꢀckner, Amino Acids 2004, 27, 231–247; b) C.
Theis, T. Degenkolb, H. Brꢀckner, Chem. Biodiversity 2008, 5,
2337–2355.
[11] M. Crisma, A. Moretto, M. De Zotti, F. Formaggio, B. Kaptein,
Q. B. Broxterman, C. Toniolo, Biopolymers 2005, 80, 279–293.
[12] a) C. H. Gçrbitz, Acta Crystallogr. Sect. B 1989, 45, 390–395; b) C.
Ramakrishnan, N. Prasad, Int. J. Protein Res. 1971, 3, 209–231.
[13] a) C. M. Venkatachalam, Biopolymers 1968, 6, 1425–1436; b) G. D.
Rose, L. M. Gierasch, J. A. Smith, Adv. Protein Chem. 1985, 37, 1–
109.
[14] a) E. Benedetti, A. Bavoso, B. Di Blasio, V. Pavone, C. Pedone, C.
Toniolo, G. M. Bonora, Biopolymers 1983, 22, 305–317; b) K. A.
Williams, C. M. Deber, Biochemistry 1991, 30, 8919–8923; c) A.
Yaron, F. Naider, Crit. Rev. Biochem. Mol. Biol. 1993, 28, 31–81;
d) S. J. Eyles, L. M. Gierasch, J. Mol. Biol. 2000, 301, 737–747.
[15] a) F. Formaggio, M. Crisma, G. M. Bonora, M. Pantano, G. Valle, C.
Toniolo, A. Aubry, D. Bayeul, J. Kamphuis, Pept. Res. 1995, 8, 6–15;
b) B. Jaun, M. Tanaka, P. Seiler, F. N. M. Kꢀhnle, C. Braun, D. See-
bach, Liebigs Ann. 1997, 1697–1710; c) G. Valle, M. Crisma, C. To-
niolo, R. Beisswenger, A. Rieker, G. Jung, J. Am. Chem. Soc. 1989,
111, 6828–6833; d) R. Bosch, H. Brꢀckner, G. Jung, W. Winter, Tet-
Distance geometry and MD calculations were carried out using the simu-
lated annealing (SA) protocol of the XPLOR-NIH 2.9.6 program.[34] For
distances involving equivalent or nonstereoassigned protons, rꢀ6 averag-
ing was used. The MD calculations involved a minimization stage of 100
cycles, followed by SA and refinement stages. The SA consisted of 30 ps
of dynamics at 1500 K (10000 cycles, in 3 fs steps) and of 30 ps of cooling
from 1500 to 100 K in 50 K decrements (15000 cycles, in 2 fs steps). The
SA procedure, in which the weights of NOE and nonbonded terms were
gradually increased, was followed by 200 cycles of energy minimization.
In the SA refinement stage, the system was cooled from 1000 to 100 K in
50 K decrements (20000 cycles, in 1 fs steps). Finally, the calculations
were completed with 200 cycles of energy minimization using a NOE
force constant of 50 kcalmolꢀ1. The generated structures were visualized
using the MOLMOL[35] (version 2K.2) program.
326
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 316 – 327

Synthesis of Integramide A
FULL PAPER
rahedron 1982, 38, 3579–3583; e) U. Slomczynska, D. D. Beusen, J.
Zabrocki, K. Kociolek, A. Redlinski, F. Reusser, W. C. Hutton,
M. T. Leplawy, J. Am. Chem. Soc. 1992, 114, 4095–4106.
Biodiversity 2007, 4, 1027–1051; c) T. Degenkolb, J. Kirschbaum, H.
Brꢀckner, Chem. Biodiversity 2007, 4, 1052–1067.
[25] a) I. L. Karle, J. Flippen-Anderson, M. Sukumar, P. Balaram, Proc.
Natl. Acad. Sci. USA 1987, 84, 5087–5091; b) B. Di Blasio, V.
Pavone, M. Saviano, A. Lombardi, F. Nastri, C. Pedone, E. Benedet-
ti, M. Crisma, M. Anzolin, C. Toniolo, J. Am. Chem. Soc. 1992, 114,
6273–6278.
[26] a) M. C. Burla, M. Camalli, B. Carrozzini, G. L. Cascarano, C. Gia-
covazzo, G. Polidori, R. Spagna, J. Appl. Crystallogr. 2003, 36, 1103;
b) G. M. Sheldrick, Acta Crystallogr. Sect. A 2008, 64, 112–122.
[27] M. Rance, O. W. Sørensen, G. Bodenhausen, G. Wagner, R. R.
Ernst, K. Wꢀthrich, Biochem. Biophys. Res. Commun. 1983, 117,
479–485.
[28] a) A. Bax, D. G. Davis, J. Magn. Reson. 1985, 65, 355–360; b) C.
Griesinger, G. Otting, K. Wꢀthrich, R. R. Ernst, J. Am. Chem. Soc.
1988, 110, 7870–7872.
[29] a) C. Bauer, R. Freeman, T. Frenkiel, J. Keeler, A. Shaka, J. Magn.
Reson. 1984, 58, 442–457; b) L. Emsley, G. Bodenhausen, J. Magn.
Reson. 1989, 82, 211–221.
[16] a) G. R. Marshall in Intra-Science Chemistry Reports (Ed.: N. Khar-
asch), Gordon and Breach, New York, 1971, pp. 305–326; b) I. L.
Karle, P. Balaram, Biochemistry 1990, 29, 6747–6756; c) C. Toniolo,
M. Crisma, F. Formaggio, C. Peggion, Biopolymers 2001, 60, 396–
419; d) C. Toniolo, Biopolymers 1989, 28, 247–257; e) C. Toniolo,
M. Crisma, G. M. Bonora, E. Benedetti, B. Di Blasio, V. Pavone, C.
Pedone, A. Santini, Biopolymers 1991, 31, 129–138; f) V. Pavone, B.
Di Blasio, A. Santini, E. Benedetti, C. Pedone, C. Toniolo, M.
Crisma, J. Mol. Biol. 1990, 214, 633–635.
[17] a) C. Toniolo, E. Benedetti, Trends Biochem. Sci. 1991, 16, 350–353;
b) K. A. Bolin, G. L. Millhauser, Acc. Chem. Res. 1999, 32, 1027–
1033.
[18] a) C. Toniolo, A. Polese, F. Formaggio, M. Crisma, J. Kamphuis, J.
Am. Chem. Soc. 1996, 118, 2744–2745; b) F. Formaggio, M. Crisma,
P. Rossi, P. Scrimin, B. Kaptein, Q. B. Broxterman, J. Kamphuis, C.
Toniolo, Chem. Eur. J. 2000, 6, 4498–4504.
[19] M. Goodman, A. S. Verdini, C. Toniolo, W. D. Phillips, F. A. Bovey,
Proc. Natl. Acad. Sci. USA 1969, 64, 444–450.
[30] A. Bax, R. H. Griffey, B. L. Hawkins, J. Magn. Reson. 1983, 55,
301–315.
[20] a) E. S. Pysh, C. Toniolo, J. Am. Chem. Soc. 1977, 99, 6211–6219;
b) S. C. Yasui, T. A. Keiderling, F. Formaggio, G. M. Bonora, C. To-
niolo, J. Am. Chem. Soc. 1986, 108, 4988–4993.
[21] G. Jung, H. Brꢀckner, R. Bosch, W. Winter, H. Schaal, J. Strꢆhle,
Liebigs Ann. Chem. 1983, 1096–1106.
[31] A. Bax, S. Subramanian, J. Magn. Reson. 1986, 67, 565–569.
[32] A. Bax, M. F. Summers, J. Am. Chem. Soc. 1986, 108, 2093–2094.
[33] XEASY-ETH Automated Spectroscopy 1.4 (Version 60801000 of
the XWindow System), see: C. Bartels, T. Xia, M. Billeter, P. Gun-
tert, K. Wꢀthrich, J. Biomol. NMR 1995, 6, 1–10.
[22] M. Bellanda, E. Peggion, R. Bꢀrgi, W. van Gunsteren, S. Mammi, J.
Pept. Res. 2001, 57, 97–106.
[34] C. D. Schwieters, J. J. Kuszewski, N. Tjandra, G. M. Clore, J. Magn.
Reson. 2003, 160, 65–74. Based on X-PLOR 3.851 by A. T. Brunger.
[35] R. Koradi, M. Billeter, K. Wꢀthrich, J. Mol. Graphics 1996, 14, 51–
55.
[23] a) E. Benedetti, A. Bavoso, B. Di Blasio, V. Pavone, C. Pedone, C.
Toniolo, G. M. Bonora, Proc. Natl. Acad. Sci. USA 1982, 79, 7951–
7954; b) H. Brꢀckner, H. Graf, Experientia 1983, 39, 528–530.
[24] a) H. Duclohier, Chem. Biodiversity 2007, 4, 1023–1026; b) B. Leit-
geb, A. Szekeres, L. Manczinger, C. Vꢇgvçlgyi, L. Kredics, Chem.
Received: April 9, 2009
Revised: October 15, 2009
Published online: November 20, 2009
Chem. Eur. J. 2010, 16, 316 – 327
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
327