The complestatins as HIV-1 integrase inhibitors. Efficient isolation, structure elucidation, and inhibitory activities of isocomplestatin, chloropeptin I, new complestatins, A and B, and acid-hydrolysis products of chloropeptin I

Article abstract of DOI:10.1021/np000632z

From the screening of a microbial extract library, isocomplestatin (1), a new axial-chiral isomer of complestatin (2) which is a known rigid bicyclic hexapeptide, was identified as a potent natural product inhibitor of HIV-1 integrase, a unique enzyme responsible for viral replication. Isocomplestatin showed inhibitory activities (IC₅₀) in coupled 3′-end processing/strand transfer (200 nM), strand transfer (4 μM), and HIV-1 replication (200 nM) in virus-infected cells. Attempted large-scale isolation of 1 by the literature method, used for the isolation of complestatin, led to lower yield and limited availability. We have developed several new, two-step, high-yielding absorption/elution methods of isolation based on reverse-phase chromatography at pH 8 that are applicable to scales from one gram to potential industrial quantities. We have also discovered and determined the structure of two new congeners of 1, namely, complestatins A (4) and B (5), with almost equal HIV-1 integrase activity. They differ from 1 at C2′ and C3′ of the tryptophan moiety (residue F). Selective acid hydrolysis of chloropeptin I (3), itself a known acid-catalyzed rearranged isomer of 1 and 2 (8′- vs 7′-substitution in tryptophan residue F, respectively), an isomer of complestatin, and isocomplestatin resulted in a number of fragments (6-10) with retention of most of the HIV-1 integrase activity. The structure - activity relationship as revealed by these compounds could possibly lead to the design of better inhibitors or understanding of the HIV-1 integrase target.

Full text of DOI:10.1021/np000632z

874
J . Nat. Prod. 2001, 64, 874-882
Th e Com p lesta tin s a s HIV-1 In tegr a se In h ibitor s. Efficien t Isola tion , Str u ctu r e
Elu cid a tion , a n d In h ibitor y Activities of Isocom p lesta tin , Ch lor op ep tin I, New
Com p lesta tin s, A a n d B, a n d Acid -Hyd r olysis P r od u cts of Ch lor op ep tin I
Sheo B. Singh,*,† Hiranthi J ayasuriya,† Gino M. Salituro,† Deborah L. Zink,† Ali Shafiee,† Brian Heimbuch,†
Keith C. Silverman,† Russell B. Lingham,† Olga Genilloud,‡ Ana Teran,‡ Dolores Vilella,‡ Peter Felock,§ and
Daria Hazuda§
Merck Research Laboratories, P.O. Box 2000, Rahway, New J ersey 07065, Centro de Investigacio´n Ba´sica,
Laboratorios Merck Sharp & Dohme de Espan˜a, S. A. J osefa Valca´rcel 38, 28027 Madrid, Spain,
and Merck Research Laboratories, West Point, Pennsylvania 19486
Received December 29, 2000
From the screening of a microbial extract library, isocomplestatin (1), a new axial-chiral isomer of
complestatin (2) which is a known rigid bicyclic hexapeptide, was identified as a potent natural product
inhibitor of HIV-1 integrase, a unique enzyme responsible for viral replication. Isocomplestatin showed
inhibitory activities (IC50) in coupled 3′-end processing/strand transfer (200 nM), strand transfer (4 µM),
and HIV-1 replication (200 nM) in virus-infected cells. Attempted large-scale isolation of 1 by the literature
method, used for the isolation of complestatin, led to lower yield and limited availability. We have
developed several new, two-step, high-yielding absorption/elution methods of isolation based on reverse-
phase chromatography at pH 8 that are applicable to scales from one gram to potential industrial
quantities. We have also discovered and determined the structure of two new congeners of 1, namely,
complestatins A (4) and B (5), with almost equal HIV-1 integrase activity. They differ from 1 at C2′ and
C3′ of the tryptophan moiety (residue F). Selective acid hydrolysis of chloropeptin I (3), itself a known
acid-catalyzed rearranged isomer of 1 and 2 (8′- vs 7′-substitution in tryptophan residue F, respectively),
an isomer of complestatin, and isocomplestatin resulted in a number of fragments (6-10) with retention
of most of the HIV-1 integrase activity. The structure-activity relationship as revealed by these compounds
could possibly lead to the design of better inhibitors or understanding of the HIV-1 integrase target.
AIDS is caused by the human immunodeficiency virus
(HIV) and remains one of the leading causes of loss of
human lives worldwide. HIV-1 encodes three enzymes, (i)
reverse transcriptase, (ii) integrase, and (iii) protease, that
are collectively responsible for viral replication and infec-
tivity. The discovery of many approved therapeutic agents
that inhibit reverse transcriptase and protease has been
helpful in prolonging and in some cases saving lives of
AIDS-infected patients. Integrase is responsible for the
multistep process of integration into the host cell genome
that includes cleavage of two nucleotide bases from the 3′-
end of each strand of viral DNA (3′-end processing) and
transfer of the processed 3′-ends into the host cell (human)
target DNA (strand transfer).1 Significantly, integrase is
not present in host cells and is specific to retroviruses.
Integration is essential for retroviral replication; therefore
inhibition of integrase should provide a means for inhibit-
ing HIV-1. The demonstration that compounds1e that
inhibit the integrase-catalyzed strand transfer reaction can
effectively block viral replication validates this approach
for the discovery and development of a new class of
antiretroviral agents targeting HIV-1 integrase.
integrase that include equisetin, phomasetin from Phoma
sp.,2 and integric acid from Xylaria sp.3 Continued screen-
ing with the recombinant integrase4 enzyme using an in
vitro assay led to the discovery and identification of
isocomplestatin (1) from Streptomyces sp. as a potent
inhibitor of HIV-1 integrase. Isocomplestatin inhibited the
coupled 3′-processing/strand transfer assay with an IC50
of 0.2 µM. Isocomplestatin also inhibited the strand trans-
fer activity of recombinant integrase in staged assays and
integration using isolated HIV-1 preintegration complexes
with comparable potency, IC50 values of 3-4 µM. Inhibition
of HIV-1 replication was observed in cell culture with an
IC50 of 0.2 µM.
Isocomplestatin (1), complestatin (2), and chloropeptin
I (3) are bicyclic hexapeptides that differ at the position of
the phenyl-indole ring junction. Complestatin (2) was
originally isolated in 1980 as an inhibitor of an alterna-
tive pathway of complement (IC50 ) 2 and 0.5 µM),5a and
a planar structure was published nine years later.5b-d
Chloropeptin I (3) was reported in 1994 along with com-
plestatin (a.k.a. chloropeptin II) as inhibitors of gp120-CD4
binding (IC50 ) 2-1.3 µM).6 The absolute stereochemistry
of chloropeptin I was determined in 1996 by a combina-
tion of hydrolysis, high-temperature molecular dynamics
and Monte Carlo conformational searching, and NMR
spectroscopy.6d More recently, chloropeptin I, which was
prepared from HCl-catalyzed rearrangement of complesta-
tin, was also shown to enhance the plasminogen binding
and fibrinolysis.7 Because of the structural complexities
and various biological effects displayed by these com-
pounds, they have attracted numerous efforts toward total
synthesis,8 but none have been successfully completed to
date.
Screening of natural product extracts provides an op-
portunity for the discovery of structurally diverse enzyme
inhibitors and receptor ligands that are usually not found
by screening synthetic chemical collections or combinatorial
libraries. We have recently reported the discovery of a
number of novel natural product inhibitors of HIV-1
* To whom all correspondence should be addressed. Tel: 1(732)594-3222.
Fax: 1(732)594-6880. E-mail: sheo_singh@merck.com.
† Merck Research Laboratories, Rahway, NJ .
‡ Laboratorios Merck Sharp & Dohme de Espan˜a.
§ Merck Research Laboratories, West Point, PA.
10.1021/np000632z CCC: $20.00
© 2001 American Chemical Society and American Society of Pharmacognosy
Published on Web 06/15/2001

The Complestatins as HIV-1 Integrase Inhibitor
J ournal of Natural Products, 2001, Vol. 64, No. 7 875
F igu r e 1. Structures of isocomplestatin (1), complestatin (2), chloropeptin I (3), and complestatins A (4) and B (5).
During the isolation of isocomplestatin (1) we observed
that in the presence of mineral acid it was converted, at
least in part, to chloropeptin I (3). We investigated this
phenomenon and reported9 an efficient TFA-catalyzed
rearrangement of 1 to chloropeptin I (3). We also discovered
two new congeners of isocomplestatin from cultured broth
of Streptomyces sp., namely, complestatin A (4) and com-
plestatin B (5), which possess oxidized tryptophan instead
of tryptophan. Despite the structural similarities among
these four compounds, their solubility behavior is highly
dissimilar. The virtual insolubility of isocomplestatin in
many common organic solvents presented a tremendous
challenge for the large-scale isolation of this compound
from fermentation broth. The inefficiencies and problems
associated with existing methods for the isolation of
complestatin and isocomplestatin prompted us to develop
an efficient new method for its isolation to provide the
needed quantities for chemical and biological studies. In
this paper several methods for the isolation of isocom-
plestatin, stereochemical elucidation, the isolation and
structure elucidation of the new complestatins A and B,
the experimental details of the conversion of isocomplesta-
tin to chloropeptin I, the detailed description of the
hydrolysis10 of chloropeptin I, and the anti HIV-1 activities
of the natural products and derivatives are described.
F igu r e 2. Isolation scheme of isocomplestatin.
Resu lts a n d Discu ssion
isocomplestatin that is free from chloropeptin I and developed
Isola t ion of Isocom p lest a t in . (a ) Bioa ssa y-Gu id ed
Sm a ll-Sca le Isola tion of Isocom p lesta tin . Size exclusion
chromatography (Sephadex LH-20) of methyl ethyl ketone
extracts of the fermentation broth of Streptomyces sp. (MA7234)
followed by reverse-phase HPLC gave isocomplestatin (50 mg/
L).
(b) La r ger Sca le Isola tion of Isocom p lesta tin . Attempts
to scale-up the isolation of isocomplestatin by using the above
method, the method described by Tanaka et al.6b for isolation
of complestatin, and various other methods at lower pH (see
Experimental Section) led to the conversion of isocomplestatin
or complestatin6b,9 to chloropeptin I due to low pH and resulted
in an extremely poor recovery due to its poor solubility. We
exhaustively examined various methods for the isolation of
a reverse-phase Amberchrom resin based absorption/elution
method that is schematically described in Figure 2. In this
method we took advantage of the anionic groups of isocom-
plestatin, which allowed the formation of salts with good
solubility in aqueous methanol. Thus the fermentation broth
filtrates and 70% aqueous acetone extracts of mycelia were
separately chromatographed on Amberchrom resin, and the
salts of isocomplestatin were eluted with 50% aqueous metha-
nol. Treatment of the salts with aqueous citric acid in methanol
followed by filtration and washing with solvents produced
isocomplestatin in greater than 90% yield. This process readily
led to the isolation of gram quantities of isocomplestatin and
is potentially amenable for industrial scale isolation. Other
resins such as HP20 and SP207 could be easily substituted

876 J ournal of Natural Products, 2001, Vol. 64, No. 7
Singh et al.
F igu r e 3. ChemDraw 3D model of isocomplestatin (1) showing NOESY correlations (mix ) 300 ms).
for the isolation of isocomplestatin, but Amberchrom was found
to be superior. Use of either mineral or the strong organic acids
such as acetic or trifluoroacetic acids, during regeneration of
free isocomplestatin from salts, caused significant formation
of chloropeptin I.
Similar chromatography of the aqueous acetone extract or
broth filtrate on Amberchrom resin at pH 2.5 and elution with
50-100% aqueous acetone at pH 2.5 gave excellent recovery
of isocomplestatin. However, a significant portion (∼15-40%)
of isocomplestatin was converted to chloropeptin I. Therefore,
these experiments further confirmed our previous observation
that chloropeptin I is an isolation artifact. Since Tanaka et
al.6a also used hydrochloric acid during their original isolation
procedure of chloropeptin I, it is reasonable to assume that
their chloropeptin I isolate is also an isolation artifact.
However, Hegde et al.11 recently reported the occurrence of
chloropeptin I as a natural product.
molecular models.6 Like complestatin, the acid hydrolysis of
isocomplestatin produced (R)-4-hydroxyphenylglycine and (R)-
3,5-dichloro-4-hydroxyphenylglycine, indicating that compound
1 and complestatin (2) are not enantiomers. In addition, we
have observed that the slight structural changes in these
compounds (cf. chloropeptin I, complestatins A and B) led to
significant changes in the 1H NMR shifts of a number of
protons located in residues D (H-4 and H-8) and F (H-1R, H-3).
This precludes the possibility of epimerization of any chiral
centers and leaves the possibility of opposite axial chirality of
residues D and F. The different axial chirality of tryptophan
between the two compounds may cause a significant difference
in the specific rotation without causing much difference in the
other properties, including 1H NMR shifts. This could explain
the existence of the small difference of the 1H NMR shifts of
the H-3 of residue F (vide infra). The NOESY correlations and
ChemDraw 3D modeling led to the elucidation of complete
stereochemistry including axial chirality of compound 1, as
illustrated in Figure 3. Compound 1 revealed the NOESY
correlations between Trp H-2 and Trp-H-2′, and Trp-H-3â and
H-5′, as well as H-6′ and H-8 (residue D), indicating that indole
nitrogen (tryptophan N-1′) is pointing down, similar to chlo-
ropeptin I. The H-3R appeared as a doublet of doublets at δ
2.87 (J ) 12, 2.4 Hz) and the H-3â appeared as an apparent
triplet at δ 3.44 (J ) 12 Hz) in the 1H NMR spectrum of
compound 1, whereas both of these protons appeared at δ 2.89
in complestatin (2).5d The J values of the protons at C-3 of 1
are consistent with the dihedral angles (ΦH-2R,H-3â ) -176°,
and ΦH-2R,H-3R ) 60°) of the isomer shown in Figure 3 and
Isola tion of Com p lesta tin A a n d B. Complestatins A and
B together with a mixture of isocomplestatin and chloropeptin
I were isolated from a 6 L fermentation of Streptomyces sp.
MA7234. The fermentation broth was filtered, and the filtrate
was acidified and extracted with ethyl acetate and chromato-
graphed on Sephadex LH-20. Elution of the column with
methanol followed by 1:1 methanol-acetone gave a mixture
of four compounds. Purification of this mixture by reverse-
phase HPLC gave the minor constituents complestatin B (5),
complestatin A (4), and a mixture of isocomplestatin (1) and
chloropeptin I (3). As shown earlier, there was a significant
loss of isocomplestatin, during gel filtration on LH-20 and
preparative HPLC, due to its insolubility, which helped in the
relative enrichment and eventual isolation of minor congeners
complestatins A and B. Like chloropeptin I, both complestatins
A and B exhibited better solubility in methanol and other polar
organic solvents. HPLC analysis of broth filtrates and the
solvent washings of precipitated isocomplestatin from an
Amberchrom column showed the presence of complestatins A
and B, indicating that they are true natural products.
Str u ctu r e Elu cid a tion : Isocom p lesta tin (1) a n d Ch lo-
r op ep tin I (2). The physical, chemical, and spectral properties
would be inconsistent with the dihedral angles (ΦH-2R,H-3â
)
-80°, and ΦH-2R,H-3R ) 36°) of the opposite axial isomer.
Therefore, on the basis of all the data structure 1 is assigned
to isocomplestatin.14
Our observation of acid-catalyzed rearrangement9 of the
complestatin ring system to the chloropeptin ring system has
been confirmed with model systems.8c The rearrangement
outcome is unaffected by the stereochemistry of the R-carbon
of tryptophan, but corresponding acyclic systems do not
undergo rearrangement.8c
Com p lesta tin A (4) a n d Com p lesta tin B (5). ESI mass
spectral analysis of complestatin A (4) and complestatin B
(5) showed pseudomolecular ions at m/z 1041 (M + H)+
and m/z 1057 (M + H)+, respectively. High-resolution ESI-
FTMS analyses of the molecular ion of both compounds
gave exact masses of 1041.1137 and 1057.1081, analyzing for
C61H46Cl6N7O16+H and C61H46Cl6N7O17+H, respectively. These
formulas consist of one and two extra oxygen atoms in
complestatin A and B, respectively, compared to complestatin,
isocomplestatin, and chloropeptin I. UV spectra of isocom-
plestatin, complestatin A, and complestatin B were qualita-
tively identical. The 1H and 13C NMR spectra (Table 1) of
isocomplestatin (1), complestatin A (4), and complestatin B
1
including H and 13C NMR spectra of 1 are identical to those
of complestatin (2)5,12 except for their specific rotations and
the 1H NMR chemical shifts of H-3 of tryptophan residue F,
indicating that the two compounds were not completely
identical. Compound 1 showed specific rotation13 of -11.2°,
which is opposite in sign of the reported specific rotation
(+16.3°) of complestatin.5 Complestatin has been previously
converted to chloropeptin I ([R]D -18.7°) by treatment with
hydrochloric acid.6a,7,12 Similar acid treatment of compound 1
with HCl or TFA produced chloropeptin I ([R]D -28°)13 with
identical physical, chemical, and 1H and 13C NMR spectral
properties including axial chirality and conformation as de-
termined by the comparison of NOESY correlations and

The Complestatins as HIV-1 Integrase Inhibitor
J ournal of Natural Products, 2001, Vol. 64, No. 7 877
Ta ble 1. 1H (600 and 500 MHz) and 13C NMR (125 MHz) Assignment of Isocomplestatin (1), Complestatin A (4), and Complestatin B
a
(5) in DMSO-d6
isocomplestatin (1)
δH, J , Hz
complestatin A (3)
δH, J , Hz
complestatin B (4)
δH, J , Hz
position
δC
δC
δC
Hydroxyphenylglycine (A)
171.6
1
2
3
4,8
5,7
6
171.4
55.8
127.8
128.2
115.3
157.2
171.4
55.8
128.1
128.2
115.3
157.2
5.05, d, 6.2
56.0
127.9
128.4
115.5
157.4
5.06, d, 6.0
5.06, d, 6.0
7.10, d, 8.5
6.76, d, 8.5
7.10, d, 9.0
6.74, d, 9.0
7.09, d, 8.4
6.75, d, 8.4
NH
8.52, d, 6.5
8.43, d, 6.6
8.42, d, 6.6
Tyrosine (B)
168.7
61.5
1
2
3
168.5
61.2
35.1
168.5
61.3
35.0
5.06, m
3.05, m
5.06, m
3.0, m
5.07, m
3.04, m
35.1
4
5
6
7
8
9
134.4
130.6
121.7
155.2
123.2
131.5
31.3
134.3
130.7
121.5
156.1
122.1
131.3
31.9
134.2
130.5
121.5
155.8
122.9
131.6
31.2
7.18, dd, 8.5, 1
7.06, dd, 8.5, 2.3
7.18, dd, 8.5, 1.5
7.15, dd, 8.4, 2.4
7.18, dd, 8.4, 1.8
7.14, dd, 8.4, 2.4
6.84, brd, 8.5
7.81, dd, 8.5, 1.5
2.97, s
6.76, dd, 8.4, 2.4
7.78, dd, 9.0, 2.0
3.0, s
6.78, dd, 8.4, 2.4
7.80, dd, 8.4, 2.4
3.0, s
NCH3
3,5-Dichloro-4-hydroxyphenylglycine (C)
169.5
1
2
3
4,8
5,7
6
169.2
51.4
131.1
126.9
122.0
148.7
169.3
51.6
131.2
127.0
122.0
148.4
5.19, d, 6.2
7.33, s
51.9
131.1
127.2
122.0
148.9
5.18, d, 6.6
5.19, d, 6.0
7.37, s
7.36, s
NH
8.73, d, 6.0
8.90, d, 6.0
8.91, d, 6.0
3,4-Dihydroxyphenylglycine (D)
167.4
1
2
3
4
5
6
7
8
167.6
55.0
167.6
54.9
5.56, d, 8.3
5.46, d, 2.2
55.3
127.2
113.0
149.7
140.2
130.1
126.1
5.69, d, 7.8
5.71, d, 9.0
5.76, d, 2.4
126.3
110.5
149.6
139.4
131.1
129.5
127.7
112.9
149.5
140.6
130.5
125.8
5.76, d, 2.4
5.06, d, 2.0
8.28, d, 9.4
9.40, s
5.69, d, 2.4
8.61, d, 9.6
9.45, s
5.73, d, 2.4
8.53, d, 9.0
9.46, s
NH
OH
3,5-Dichloro-4-hydroxyphenylglycine (E)
170.9
1
2
3
4,8
5,7
6
169.9
51.1
131.9
126.7
121.7
148.2
170.6
54.2
130.5
126.1
121.9
148.7
5.56, d, 8.6
7.28, s
54.5
131.9
126.4
122.2
148.7
5.54, d, 9.0
5.54, d, 8.4
7.02, s
7.02, s
NH
7.82, d, 8.6
7.14, d, 9.0
7.18, d, 8.4
Tryptophan (F)
1
2
3
170.5
57.0
28.3
167.8
48.2
29.1
166.7
47.3
35.1
4.15, ddd, 11, 9, 3
R: 2.87, dd, 12, 2.4
â: 3.40, t, 12
3.45, t, 8.0
3.00, m
3.46, m
3.00, m
1′
2′
3′
4′
5′
6′
7′
8′
9′
NH
NH
10.87, s
NH
178.9
43.7
126.2
121.5
115.5
127.8
110.9
142.2
10.6, s
NH
178.4
75.5
128.2
122.5
114.6
127.0
110.9
140.9
10.5, s
123.6
111.8
126.3
118.4
123.6
134.3
114.4
136.2
7.25, d, 2.1
3.65, t, 3.9
7.42, d, 8.4
6.82, d, 8.4
7.08, d, 9.0
6.74, dd, 7.8, 1.2
7.12, d, 7.8
6.75, dd, 7.8, 1.2
7.23, s
6.76, d, 1.2
6.71, d, 1.2
9.35, brs
8.55, brs
9.31, brs
3,5-Dichloro-4-hydroxyphenyl Oxo Acid (G)
1
2
3
4,8
5,7
6
164.5
185.7
127.2
130.5
122.5
156.8
164.1
183.6
127.2
130.7
122.7
156.5
163.8
184.0
127.2
130.5
122.0
156.5
7.75, s
7.82, s
7.83, s
a
Unassigned OH groups each broad singlets, 1: δ 9.47, 9.85, 10.07; 4: δ 6.51, 9.47, 9.97, 10.08; 5: δ 6.19, 9.49, 9.98, 10.09.

878 J ournal of Natural Products, 2001, Vol. 64, No. 7
Singh et al.
studies failed to define the stereochemistry at the new asym-
metric carbon of the tryptophan moiety. The conformation of
complestatins A and B appeared to be similar to isocomplesta-
tin except for the oxidized tryptophan due to relatively free
rotation caused by the sp3 carbon.
Hyd r olysis of Ch lor op ep tin I. Brief accounts of the
selective hydrolysis of chloropeptin I (3) to fragments 6-9
(Figure 6) have already been described.10 The experimental
details of the hydrolysis, characterization of a new hydrolytic
product, and the biological activities of the fragments are
described here. The terminal hydroxyphenyl glycine of 3 was
selectively hydrolyzed by heating with trifluoroacetic acid at
60 °C to quantitatively give compound 6. Heating of chloropep-
tin I (3) in a vacuum hydrolysis tube under inert atmosphere
with a mixture of acetic acid-HCl-thioglycolic acid gave
compound 7. Prolonged heating of the reaction mixture led to
the formation of compounds 8-10 (Figure 6).10 The progress
of the reaction was monitored by analytical HPLC. The
hydrolytic reaction led to the formation of compound 6 in less
than 1 h, which was subsequently converted to compound 7.
The formation of the latter compound was maximized at about
2 h and upon further heating was converted to compounds
8-10. The time course of the reaction has been described
previously.10 Compound 9 was isolated from the reaction
mixture and is characterized here for the first time. Although
chloropeptin I has good solubility in methanol and many other
polar organic solvents, its solubility in an aqueous environment
was poor even at high temperature. Acetic acid as a cosolvent
was helpful in the dissolution of chloropeptin I. Thioglycolic
acid was useful in preventing decomposition of tryptophan.
The cleavage of the terminal 4-hydroxymethyl glycine led
to the cis-trans isomerization of the tertiary N-methyl amide
in compounds 6 and 7. These compounds each showed two
HPLC peaks that could be easily separated by preparative
HPLC. However, the HPLC analysis of the fractions, suppos-
edly containing a single compound, revealed that the fractions
consisted of the same two compounds present in the starting
F igu r e 4. COSY and HMBC correlations of fragments of complest-
atins A (A) and B (B).
(5) were essentially identical except for the signals originating
from the tryptophan residue (F) and downfield shift of aromatic
protons of residue D. This indicated that the extra oxygen(s)
were located in the tryptophan residue (F) in both of these
compounds. The 13C NMR spectrum of complestatin A (4)
revealed a signal for a carbonyl group at δ 178.9 and a methine
carbon at δ 43.7. Complestatin B (5) exhibited a signal for a
carbonyl group at δ 178.9 and an oxygenated quaternary
carbon at δ 75.5 instead of olefinic carbons at δ 123.6 and 111.8
for C-2′ and C-3′ of tryptophan, respectively. These assign-
ments were confirmed by HMBC correlations of NH-1′ to C-3′
in complestatin A (Figure 4A) and NH-1′ to C-2′ and C-3′, and
H-2 to C-3′, in complestatin B (Figure 4B).
Ster eoch em istr y of Com p lesta tin s A a n d B. The NOE-
SY (mix ) 300 ms) spectrum of complestatin A (4) exhibited
essentially identical correlations to that observed for isocom-
plestatin and chloropeptin I6c except for the residue F and thus
indicated identical backbone stereochemistry of amino acids.
Within the residue F, H-3′ showed NOESY correlation to H-5′,
and NH-1′ showed correlation to H-8′. Similar NOESY cor-
relations were also observed for complestatin B. Extensive
NOESY and molecular modeling (see Figure 5, which shows
the models of 3′R and 3′S stereoisomers of complestatin B)
F igu r e 5. ChemDraw 3D generated models of complestatin B (5) with 3′R and 3′S stereochemistry.
F igu r e 6. Structures of hydrolytic fragments of chloropeptin I.

The Complestatins as HIV-1 Integrase Inhibitor
J ournal of Natural Products, 2001, Vol. 64, No. 7 879
(6 and 7) were evaluated in the HIV-1 viral spread assay.2e
These compounds exhibited IC95 values of 5, 5, 5, and 7.5 µM,
respectively (Table 2). Isocomplestatin (1) in the same assay
showed an IC50 value of 0.2 µM.
Ta ble 2. Biochemical and Antiviral Activities of
Isocomplestatin and Related Compounds (1-10)
HIV-1 integrase (IC50 µM)
HIV-1 viral spread
compound
coupled
strand transfer
(IC95 µM)
HIV-1 In tegr a se Activity. Isocomplestatin (1) inhibited
in vitro HIV-1 integrase coupled and strand transfer activities
with IC50 values of 0.2 and 4.0 µM, respectively. Chloropeptin
I (3), complestatin A (4), and complestatin B (5) were less
active than isocomplestatin and exhibited IC50 values of 0.4-
1.7 and 5-12.5 µM in the coupled and strand transfer assays,
respectively. The comparable activities of these compounds are
not surprising due to their structural similarity. This was
further confirmed from the molecular model comparison of
these compounds. The molecular models (ChemDraw 3D) of
all four compounds were highly similar, except for a slight and
expected deviation of the tryptophan-containing 16-membered
ring of chloropeptin I. The hydrolytic fragments (6 and 7) with
intact macrocyclic rings retained almost all of the activities
and exhibited IC50 values of 0.6, 0.3 µM and 8.0, 6.0 µM in
coupled and strand transfer assays, respectively. However,
opening of one of the macrocyclic rings caused a significant
decrease in the integrase activity (see compounds 8-10 in
Table 2). Isocomplestatin inhibits preintegration complexes
from HIV-1-infected cells with an IC50 comparable to that
observed for inhibition of strand transfer using complexes
assembled on the viral donor DNA with recombinantly ex-
pressed integrase. This is similar to what has been observed
previously for the diketoacid inhibitors of integrase.2e The
observation that isocomplestatin is active both in strand
transfer and against HIV-1 preintegration complexes suggests
this compound may be an alternative platform on which to
design novel strand transfer inhibitors.
It should be noted that in contrast to the diketoacid
inhibitors, which are much weaker inhibitors of 3′ processing
than strand transfer, isocomplestatin and the derivatives
reported herein inhibit 3′ processing with a characteristically
10-fold better IC50 than the IC50 measured in strand transfer
(Table 2). These data suggest that isocomplestatin may
preferentially bind to the uncomplexed enzyme, unlike the
diketoacid, which requires binding the viral DNA end.2e Also,
unlike the diketoacid, isocomplestatin inhibits the HIV-1
disintegration activity of both the intact protein and the
catalytic core domain of integrase (amino acids 50-212) with
comparable potency (IC50 0.5 µM); thus, isocomplestatin binds
to the core domain of integrase. Isocomplestatin inhibits the
integrase protein from several related retroviruses. Isocom-
plestatin inhibited recombinant feline immunodeficiency virus
with an IC50 value of 0.5 µM and was a more potent inhibitor
of recombinant simian immunodeficiency virus, with an IC50
value of 0.1 µM.
1
3
4
5
6
7
8
9
10
0.2
0.4
0.8
1.7
0.6
0.5
3.8
7.8
28
4
5
12.5
12.5
8
6
45
86
5
5
NDa
ND
5
7.5
ND
ND
ND
>100
a
ND (not determined)
Ta ble 3. Other Activities of Isocomplestatin
assay
IC50 (µM)
HIV disintegration
3′-end processing
integration
rFIV
0.5
0.2
3.0
0.5
0.10
0.25
3.0
0.2
2.0
rSIV
rAMV
rHIV (PIC)
rHIV
gp120-CD45
mixture due to isomeric equilibration. The rate of the isomeric
equilibration was different for each isomer and was solvent
dependent. For example, in aqueous CH3CN + 0.1% TFA
equilibration of the faster eluting isomer was more rapid than
that of the slower eluting isomer. In less than 20 min, the
fraction containing the faster eluting isomer was equilibrated
to a ∼50:50 mixture consisting of both isomers, but it took
more than 1 h for the same equilibration of the slower eluting
isomer. The stability and the rate of the equilibration were
reversed in methanol. For example in the case of compound
7, the slower eluting isomer was completely converted (>90%)
to the faster eluting isomer after 24 h in MeOH, as evidenced
1
by HPLC. The H NMR spectra of both compounds 6 and 7 in
DMSO-d6 showed a 2:1 isomeric ratio even at higher temper-
atures (50-100 °C). However the preponderance of only one
of the two isomers was observed when the 1H NMR spectra
were recorded in THF-d8 in the presence of saturating amounts
of LiCl. This phenomenon of LiCl-induced change in isomeric
population is consistent with the observation originally made
for N-methylated compounds, i.e., cyclosporin.15 The exchange-
able hydroxy protons were broadened and were present
throughout the spectrum recorded in THF-d8. The 1H NMR
spectrum of compound 8 also showed the presence of two major
isomers due to cis-trans isomerization of the tertiary N-
methyl amide group; however the isomers were not separable
by HPLC. It is interesting to note that in the presence of
terminal 4-hydroxyphenylglycine this cis-trans isomerization
of N-methyl amide is not observed (cf. isocomplestatin, com-
plestatins, and chloropeptin I).
Biologica l Activity. Isocomplestatin was isolated following
bioassay-guided fractionation using an in vitro HIV-1 integrase
assay that recapitulates the staged process of viral integra-
tion.4 Subsequently, isocomplestatin and all of the other
compounds were evaluated in a number of related biological
and biochemical assays including the HIV-1 multiple cycle
assay for viral replication (viral spread), and the data are
shown in Tables 2 and 3. Complestatin (2) and chloropeptin I
(3) were reported6a,b to inhibit the binding of gp120 to CD4
with IC50 values of 2 and 1.3 µM, respectively. Like complesta-
tin, isocomplestatin was shown to affect the formation of
syncitia at µM levels in cell culture. Although this is higher
than the concentration required to inhibit HIV-1 replication
(vide infra), we cannot exclude the possibility that the antiviral
activity of isocomplestatin is all or in part due to the effect on
the HIV-1 envelope GP120/CD4 receptor interaction. Isocom-
plestatin (1), chloropeptin I (3), and the hydrolytic fragments
In conclusion we report absolute stereochemistry and ef-
ficient processes for isolation of isocomplestatin, which can be
implemented from a gram to an industrial scale. We have
comprehensively evaluated these compounds against HIV-1
integrase and various other retroviral targets. In addition, we
have reported two new oxidized congeners of isocomplestatin,
which has doubled the number of the members of this small
family of bicyclic hexapeptides. Selective hydrolysis of chlo-
ropeptin I and the retention of most of the biological activity
by the bicyclic fragments shed light on the structure-activity
relationship of this class of compounds and may lead to better
inhibitor design and understanding of HIV-1 integrase target.
Exp er im en ta l Section
Gen er al Exper im en tal P r ocedu r es. Amberchrom CG161,
a styrenic-based resin, was obtained from TosoHaas Corpora-
tion. HP20 and SP207 were obtained from Mitsubishi Chemi-
cal Company. Optical rotations were measured on a Perkin-
Elmer model 241 polarimeter. All NMR spectra were recorded
on Varian Inova 400, 500, or 600 MHz instruments operating
at 400, 500, and 600 MHz for 1H and 100, 125, and 150 MHz
for 13C nuclei. An HP1100 was used for analytical HPLC.
LCMS was performed on a Thermo Quest LCQ instrument
using electrospray ionization (ESI) or atmospheric pressure

880 J ournal of Natural Products, 2001, Vol. 64, No. 7
Singh et al.
chemical ionization (APCI). High-resolution mass spectral
analysis was performed on a Thermo Quest FTMS using
electrospray ionization.
brown powder consisting of a mixture of isocomplestatin and
chloropeptin I.
Isola tion of Isocom p lesta tin (1) by Am ber ch r om Ch r o-
m a togr a p h y a t p H 2.5. A 10 L fermentation was filtered,
and the cells were extracted twice with 4 L and once with 2 L
of 70% aqueous acetone by stirring at room temperature
overnight. The pH of the extract was ∼7.0. The quantitative
HPLC analysis of the acetone extract (9 L) revealed the
presence of 2.5 g of isocomplestatin. A small portion (100 mL)
of extract was concentrated to a small volume and acidified
with 4 N HCl to pH 2.5. The acidification resulted in the
formation of a precipitate that was collected by filtration to
give isocomplestatin (9 mg). The filtrate was charged onto 10
mL of Amberchrom preequilibrated at pH 2.5 with 30%
aqueous acetone. The column was washed with 30% and 50%
aqueous acetone at pH 2.5 and eluted with 70% aqueous
acetone to give 12.5 mg of a 2:1 mixture of isocomplestatin
and chloropeptin I.
Isola tion of Isocom p lesta tin F r ee of Ch lor op ep tin I.
A 13 L fermentation of the culture was filtered through Celite
to give 9 L of filtrate (pH ) 8.4). Mycelia were separated and
extracted five times by shaking with 2 L each of 70% aqueous
acetone to give a combined extract of 9 L. A quantitative HPLC
analysis (Zorbax RX C-8, 4.6 × 250 mm, 50% aqueous CH3-
CN + 0.1% TFA, 1 mL/min) indicated the presence of ∼1.3
and 1.1 g (>90%) of isocomplestatin (tR ) 8.8 min) in extracts
A and B, respectively, while chloropeptin I (tR ) 9.6 min) was
not detectable.
(1) By Am ber ch r om Ch r om a togr a p h y a t p H >8. Eight
liters of filtrate (pH ) 8.4) was charged to a 1 L, thoroughly
washed Amberchrom column at a flow rate of 15 mL/min. The
column was washed with 15 L of water followed by 4 L each
of 10% and 20% aqueous methanol. Isocomplestatin was eluted
with 66% aqueous acetone (0.6 L) and finally with acetone (1
L). All eluates containing isocomplestatin were combined and
concentrated under reduced pressure to yield a mostly aqueous
solution, which was lyophilized to give salts of isocomplestatin
as a brown powder (1.1 g). The fractions eluting with 50-66%
acetone contained complestatins A (4) and B (5).
F er m en ta tion of MA7234. Streptomyces sp. (MA7234) was
grown on an ISP Medium2 plate (DIFCO, Detroit, MI) for 7
days at 28 °C. Growth was scraped from the plate and
inoculated into a 250 mL baffled flask containing 50 mL of
ATCC-2 media [soluble starch (20 g); dextrose (10 g); NZ amine
E (5 g); Difco beef extract (3 g); Difco bacto peptone (5 g); Fidco
yeast extract (5 g); CaCO3 (1 g); in 1 L distilled water; pH was
7.0 prior to the addition of calcium carbonate]. Flasks were
shaken at 220 rpm for 4 days at 28 °C, and 1.5 mL of the ATCC
culture broth was transferred to 250 mL nonbaffled production
flasks containing 45 mL of KHC media containing dextrin (20
g); â-cyclodextrin (10 g); primary yeast (10 g); tomato paste
(20 g); CoCl2-6H20 (0.005 g); in 1 L distilled water at pH 7.2.
The inoculated flasks containing production media were
shaken at 220 rpm for 8 days at 28 °C. The culture broths
were pooled together and extracted as described below.
Bioa ssa y-Gu id ed Isola tion of Isocom p lesta tin . Fer-
mentation broth (0.5 L) was extracted with methyl ethyl
ketone (0.5 L), and the extract was filtered and evaporated to
produce 300 mg of residue. The residue was dissolved in 100
mL of methanol and was charged onto a 1.5 L Sephadex LH-
20 column in methanol. Elution of the column with methanol
afforded the HIV-1 integrase active fractions in 1.2 column
volumes. The active fractions were concentrated under reduced
pressure to produce 46 mg of a brown gum. A 20 mg portion
of the gum was purified on a reverse-phase Zorbax RX C-8
HPLC column (22.4 × 250 mm) eluting with aqueous CH3CN
containing 0.1% TFA at a flow rate of 10 mL/min. The column
was initially eluted with a 50 min gradient of 10-40% CH3-
CN followed by a 10 min gradient to 70% CH3CN. Activity was
concentrated in a broad peak eluting at 76 min. Lyophilization
of the combined active fractions gave isocomplestatin (1, 12.0
mg, 50 mg/L) as a buff-colored powder. The UV, 13C NMR, and
mass spectral data were identical to complestatin.5
In itia l La r ge-Sca le Isola tion of Isocom p lesta tin a n d
Isola tion of Com p lesta tin s A a n d B Usin g Acid ic Con d i-
tion s. A 6 L fermentation grown in liquid (pH ) 7.6) media
was filtered in order to separate the cells from the extract.
The filtrate (4 L) was acidified to pH 2.0 by addition of 4 N
HCl and extracted with ethyl acetate (2 × 4 L). The ethyl
acetate extract was concentrated to dryness to give 1.5 g of
residue, which was purified on a 2 L Sephadex LH-20 column
in methanol. The column was eluted with three column
volumes of methanol followed by two column volumes of a 1:1
mixture of methanol-acetone to afford 180 mg of a mixture
of isocomplestatin and related compounds as a buff-colored
powder. Analysis of the powder by HPLC (Zorbax RX C-8, 4.6
× 250 mm, 50% aqueous CH3CN containing 0.1% TFA, 1 mL/
min) indicated it to be an approximately 10:1:1:1 mixture of
isocomplestatin (1, tR ) 8.8 min), chloropeptin I (3, tR ) 9.6
min), complestatin A (4, tR ) 5.7 min), and complestatin B (5,
tR ) 4.3 min). A 53 mg portion of the residue was chromato-
graphed on a reverse-phase Zorbax SB CN HPLC column (22.4
× 250 mm) using a gradient of aqueous CH3CN + 0.1% TFA
at a flow rate of 10 mL/min. The column was eluted with 20%
of CH3CN for 12 min, then a linear gradient to 40% in 10 min,
held for 10 min, followed by a linear gradient to 60% of CH3-
CN in 10 min to give 1 (5 mg containing ∼10% of chloropeptin
I, 3), 4 (3.5 mg), and 5 (4 mg) as buff powders.
The aqueous acetone extract was concentrated to a total
volume of ∼5 L and diluted with 4 L of water. The pH of this
solution was adjusted to 8.5 by addition of 10% aqueous sodium
bicarbonate and chromatographed on a similar Amberchrom
column to give 1.4 g of isocomplestatin salt that was contami-
nated with large amounts of fatty compounds (1H NMR).
(2) By SP 207 Ch r om a togr a p h y a t p H >8. One hundred
milliliters (containing 14.4 mg of isocomplestatin) of filtrate
(pH ) 8.4) was charged on a 60 mL SP207 column. The column
was washed with water (400 mL) and 100 mL each of 10%
and 20% methanol, and isocomplestatin salt was eluted with
50% aqueous methanol (80 mL) followed by 160 mL each of
50% aqueous acetone and acetone. The combined solution was
concentrated under reduced pressure to yield mostly aqueous
solution that was lyophilized to give 21 mg (containing 13 mg,
90% yield) of isocomplestatin salts.
(3) By HP 20 Ch r om a togr a p h y a t p H >8. One hundred
milliliters of the filtrate was charged onto a 60 mL HP20
column, and the chromatography was repeated in an identical
manner as described for SP207 to give 19 mg (containing 13
mg, 90% yield) of isocomplestatin salts.
Regen er a tion of Isocom p lesta tin fr om Isocom p lesta -
tin Sa lts. Isocomplestatin salts (1.1 g) prepared from the
filtrate by Amberchrom chromatography were dissolved in 50
mL of methanol and acidified by addition of 30 mL of 10%
aqueous citric acid. This led to the formation of a precipitate
of salt-free isocomplestatin. The precipitate was collected on
a scintered glass funnel and washed with 10% aqueous citric
acid (50 mL), water (50 mL), methanol (100 mL), ethyl acetate
(100 mL), and finally hexane (100 mL). The resulting powder
was dried in a vacuum desiccator overnight to afford 1.03 g
(two step yield ∼90%) of highly pure (HPLC, LCMS, 1H NMR)
isocomplestatin (1) as a brown solid.
The cells were extracted with 70% aqueous acetone (2 ×
1.5 L), and the extract was concentrated under reduced
pressure to a volume of 1.2 L, cooled, and acidified to pH 2.0
by addition of 4 N HCl. The acidified solution was extracted
with EtOAc (2 × 1.2 L), and the EtOAc extract was concen-
trated under reduced pressure to a volume of 150 mL and
extracted twice with 150 mL of 5% aqueous NaHCO3 solu-
tion. The sodium bicarbonate extracts were pooled, cooled,
acidified to pH ) 2.0, and re-extracted with EtOAc (2 × 250
mL). The EtOAc extract was concentrated to dryness, shaken
in methanol, and then centrifuged to give 55 mg of a dark
A similar regeneration process produced 0.5 g of isocom-
plestatin from 1.4 g of salt isolated from the aqueous acetone
extract and 12 mg each from SP207 and HP20 chromatogra-

The Complestatins as HIV-1 Integrase Inhibitor
J ournal of Natural Products, 2001, Vol. 64, No. 7 881
phy. The fatty impurities present with isocomplestatin salts
were easily removed during the organic solvent washes of the
regeneration process and led to highly pure isocomplestatin.
Small amounts of complestatins A and B that were present
with isocomplestatin were also washed out due to their better
solubility in methanol.
(2/3H, d, J ) 1.5 Hz), 6.06 (1/3H, brs), 6.78 (1/3H, dd, J ) 8,
2.5 Hz), 6.92, 6.93 (1H, t, J ) 7. 5 Hz), 7.03 (1/3H, dd, J ) 8,
2 Hz), 7.05 (1/3H, d, J ) 7 Hz), 7.10 (1/3H, d, J ) 8 Hz), 7.15
(2/3H, dd, J ) 8, 2 Hz), 7.18 (2/3H, dd, J ) 8, 2 Hz), 7.23 (2/
3H, d, J ) 9 Hz), 7.25 (2/3H, d, J ) 9 Hz), 7.26 (1H, s), 7.39
(4H, s), 7.59 (1H, dd, J ) 9, 2 Hz), 7.68 (1H, dd, J ) 8.5, 2
Hz), 7.87, 7.88 (2H, s), 7.93 (1H, d, J ) 9 Hz), 8.06 (1H, d, J
) 8.5 Hz), 8.53 (1/3H, d, J ) 6.5 Hz), 8.63 (2/3H, d, J ) 8.5
Isocom p lesta tin (1): tR 8.8 min (Zorbax RX C-8, 4.6 × 250
22
mm, 50% aqueous CH3CN + 0.1% TFA, 1 mL/min); [R]D
-11.2° (c, 1.6, DMSO); [R]D -60° (c, 0.25, CH3OH-0.01 N
Hz), 9.98 (1/3H, brs), 10.23 (2/3H, brs); 1H NMR (THF-d8
+
22
NaOH, 2:1) {lit. [R]D26 +24.5° (c, 0.13, CH3OH-0.01 N NaOH,
saturating LiCl, δ) 3.06 (3H, s, N-CH3), 3.13 (3H, m), 3.69 (1H,
m), 4.65 (1H, m), 5.27 (1H, d, J ) 6.0 Hz), 5.69 (1H, m), 5.79
(1H, d, J ) 9 Hz), 5.81, 5.97 (1H, d, J ) 2.5 Hz), 6.09, 6.20
(1H, d, J ) 2.5 Hz), 6.35 (1H, d, J ) 9.5 Hz), 6.81 (1H, dd, J
) 8.5, 2.0 Hz), 6.85 (1H, t, J ) 7.5 Hz), 7.14 (1H, d, J ) 7.0
Hz), 7.21 (1H, d, J ) 8.5, 2.5 Hz), 7.31 (1H, d, J ) 8.5 Hz),
7.32 (1H, brd, J ) 8 Hz), 7.42 (1H, brs), 7.51 (2H, s), 7.78 (2H,
s), 7.88 (2H, s), 7.99 (1H, brd, J ) 8 Hz), 8.19 (2H, d, J ) 6.5
Hz), 10.4 (1H, brs); HR-ESI-FTMS (m/z) 11770731 [(M + H)+,
calcd for C53H39Cl6N6O13 1177.0706].
2:1)};5d [R]D26 +16.3° (c, 1.6, DMSO);6a UV λmax (CH3CN-H O
2
+ 0.1% TFA): 205, 240, 290 nm; HR-ESI-FTMS (m/z) 1326.1155
[(M + H)+, calcd for C61H46Cl6N7O15 1326.1183]; 1H and 13C
NMR (see Table 1).
Ch lor op ep tin I (3): tR 9.6 min (Zorbax RX C-8, 4.6 × 250
22
mm, 50% aqueous CH3CN + 0.1% TFA, 1 mL/min); [R]D
-
26
84° (c, 0.5, CH3OH); [R]D -28° (c, 0.5, DMSO) {lit.6c [R]D
22
-18.7° (c, 0.16, DMSO)}; UV λmax (CH3CN-H2O + 0.1%
TFA): 205, 240, 290 nm; HR-ESI-FTMS (m/z) 1326.1161 [(M
+ H)+, calcd for C61H46Cl6N7O15 1326.1183]; 1H and 13C NMR
(see ref 7b).
Min er a l Acid Hyd r olysis of Ch lor op ep tin I. A solution
of chloropeptin I (100 mg) in acetic acid (2 mL), 8 N HCl (0.5
mL), and thioglycolic acid (0.1 mL) was added to a vacuum
hydrolysis tube, degassed by pulling vacuum, and purged with
nitrogen. The solution was heated at 110 °C for 24 h and was
allowed to cool to room temperature. Solvents were removed
under reduced pressure to give a powder, which was dissolved
in methanol and chromatographed on a Primesphere 5µ C-8
(22 × 250 mm) column and eluted with an aqueous CH3CN +
0.1% TFA gradient (10% aqueous CH3CN for 10 min, 10 to
70% in 60 min) at a flow rate of 8 mL/min. The fractions were
analyzed by analytical HPLC, and similar fractions were
mixed and lyophilized to give the following four major frac-
tions: A (5-10 min, 21.4 mg), compound 10 (34-40 min, 2.2
mg), B (41-49 min, 28.4 mg), C (50-59 min, 30.1 mg). Fraction
A was rechromatographed on the same column and eluted with
20% aqueous CH3CN + 0.1% TFA for the first 20 min followed
by a 20 min gradient to 45% at a flow rate of 8 mL/min.
Lyophilization of the fractions gave TFA salts of N-methyl-4-
hydroxyphenylglycine (7-11 min, 10.1 mg), 4-hydroxyphe-
nylglycine (12-14 min, 5.3 mg [HR-ESI-FTMS (m/z) 168.0664
(M + H)+, calcd for C8H10NO3+H 168.0661]), 3,5-dichloro-4-
hydroxyphenylglycine (25-32 min, 3 mg, [R]25D -10° (c, 0.1, 4
Com p lesta tin A (4): tR 5.7 min (Zorbax RX C-8, 4.6 × 250
mm, 50% aqueous CH3CN + 0.1% TFA, 1 mL/min); [R]D22 +10°
(c, 0.1, CH3OH); UV λmax (CH3CN-H O + 0.1% TFA) 205, 240,
2
290 nm; HR-ESI-FTMS (m/z) 1342.1137 [(M + H)+, calcd for
1
C
61H46Cl6N7O16 1342.1132]; H and 13C NMR (see Table 1).
Com p lesta tin B (5): tR 4.3 min (Zorbax RX C-8, 4.6 × 250
mm, 50% aqueous CH3CN + 0.1% TFA, 1 mL/min); [R]D22 +20°
(c, 0.1, CH3OH); UV λmax (CH3CN-H O + 0.1% TFA) 205, 240,
2
290 nm; HR-ESI-FTMS (m/z) 1358.1120 [(M + H)+, calcd for
1
C
61H46Cl6N7O17 1358.1081]; H and 13C NMR (see Table 1).
Rea r r a n gem en t of Isocom p lesta tin to Ch lor op ep tin I.
Isocomplestatin (112 mg) was heated in 1 mL of anhydrous
TFA at 60 °C for 15 min. The progress of the reaction was
analyzed by a reverse-phase HPLC (Zorbax RC C-8, 4.6 × 250
mm, 50% aqueous CH3CN + 0.1% TFA, flow rate 1 mL/min).
The conversion of isocomplestatin (tR ) 8.8 min) to chloropeptin
I (tR ) 9.6 min) was complete in less than 15 min. TFA was
evaporated under reduced pressure to give homogeneous (by
HPLC, 1H NMR, and 13C NMR) chloropeptin I (110 mg). An
analytical sample was prepared by reverse-phase preparative
HPLC using one of the gradients described earlier followed
by lyophilization, giving buff-colored chloropeptin I (3). The
1H and 13C NMR spectral data of chloropeptin I were identical
to that reported in the literature.6b
N HCl), reported [R]25 for R-isomer -81.6°),5d 4-hydroxyphe-
D
nylglycine methyl ester (33-38 min, 1.5 mg), and compound
10 (63-65 min, 0.3 mg) all as amorphous powders. Fraction
B was similarly chromatographed eluting with 20% aqueous
CH3CN + 0.1% TFA for the first 10 min followed by a 60 min
gradient to 50% CH3CN but held at 35% at 30th min onward
at a flow rate of 8 mL/min. The like fractions were combined
and lyophilized to give compound 10‚TFA (31-32 min, 3.3 mg),
3,5-dichloro-4-hydroxyphenyl oxo-acid (33-35 min, 6 mg),
compound 9‚TFA (40 min, 0.8 mg), and compound 8‚TFA (41-
42 min, 1.0 mg) as amorphous powders. Fraction C was
accordingly chromatographed eluting with 20% aqueous CH3-
CN + 0.1% TFA for the first 10 min followed by a 70 min
gradient to 45% CH3CN at a flow rate of 8 mL/min. The
fractions eluting from 67 to 71 and 76 to 78 min were combined
to give the two interconvertible isomers of compound 7‚TFA
(5.1 mg) as a buff powder. 7: tR 7.0 and 9.5 min (Zorbax RX
C-8, 4.6 × 250 mm, 40% aqueous CH3CN + 0.1% TFA, 1 mL/
min); 1H NMR (DMSO-d6, 100 °C, δ) 2.86 (3H, s, N-CH3), 3.00
(3H, m), 3.28 (1H, m), 4.39 (1H, m), 4.79, (1H, dd, J ) 11.5, 3
Hz), 5.33 (2/3H, d, J ) 8.5 Hz), 5.36 (1H, d, J ) 9 Hz), 5.48
(1/3H, d, J ) 9 Hz), 5.51 (1/3H, m), 5.52 (1/3H, d, J ) 9 Hz),
5.56 (2/3H, d, J ) 9 Hz), 5.74 (2/3H, d, J ) 1.5 Hz), 5.77 (1/
3H, d, J ) 1.5 Hz), 5.89 (2/3H, d, J ) 1.5 Hz), 6.05 (1/3H, brs),
6.77 (1H, dd, J ) 8, 2.5 Hz), 6.95, 6.96 (1H, t, J ) 7. 5 Hz),
7.03 (1/3H, dd, J ) 8, 2 Hz), 7.09 (2/3H, d, J ) 7 Hz), 7.12
(1/3H, d, J ) 8 Hz), 7.15 (2/3H, d, J ) 7 Hz), 7.18 (2/3H, dd,
J ) 8, 2 Hz), 7.23 (1/3H, s), 7.24 (2/3H, s), 7.25 (1H, brs), 7.28,
7.29 (2H, s), 7.49 (1H, dd, J ) 9, 2 Hz), 7.68 (1H, dd, J ) 8.5,
2 Hz), 7.97 (2/3H, d, J ) 8.5 Hz), 8.02 (1/3H, d, J ) 9 Hz),
8.19 (2/3H, d, J ) 8.5 Hz), 8.22 (1/3H, d, J ) 9 Hz), 8.40 (2/
3H, d, J ) 6 Hz), 8.47 (1/3H, d, J ) 8.5 Hz), 9.96(1/3H, brs),
10.23 (2/3H, brs); 1H NMR (THF-d8 + saturating LiCl, δ) 3.03
H yd r olysis of Ch lor op ep t in I w it h Tr iflu or oa cet ic
Acid . Isocomplestatin (20 mg) was heated in anhydrous TFA
(0.2 mL) at 60 °C for 15 min to give chloropeptin I. The reaction
mixture was further heated, and the progress of the reaction
was examined by HPLC (Zorbax RC C-8, 4.6 × 250 mm, 10-
30% aqueous CH3CN + 0.1% TFA in 20 min then 30-80% CH3-
CN in another 20 min, flow rate 1 mL/min, gradient A). The
reaction was mostly complete after 24 h but heated for an
additional 3 days to afford exclusively compound 6 (tR ) 38.4
min) and 4-hydroxyphenylglycine (tR ) 5.2 min). Chloropeptin
I eluted at 37.5 min in gradient A. TFA was evaporated under
reduced pressure, and the residue was chromatographed using
a reverse-phase HPLC (Zorbax RX C-8, 22 × 250 mm) and
eluted with a 40 min gradient of 40-60% aqueous CH3CN +
0.1% TFA at a flow rate of 10 mL/min. Two peaks were
collected and lyophilized to give 1.5 mg of 4-hydroxyphenylg-
lycine and 10 mg of compound 6 as amorphous powders.
Reexamination of the compound 5 by HPLC using gradient A
showed two peaks, tR ) 38.4 and 37.2 min, in a ∼1:1 ratio due
to cis-trans isomerization of N-methyl tertiary amide. The
ratio of the two compounds varied depending on the way the
sample was stored and how it was analyzed by HPLC.
25
4-Hydroxyphenylglycine: [R]D -51.4° (c, 0.18, 2 N HCl) {lit.
[R]D25 for R-isomer -156° (c, 1, 1 N HCl)16 and -125° 5d}. 6: tR
8.55 and 9.7 min (Zorbax RX C-8, 4.6 × 250 mm, 55% aqueous
1
CH3CN + 0.1% TFA, 1 mL/min); H NMR (DMSO-d6, 70 °C,
δ) 2.84 (3H, s, N-CH3), 3.00 (3H, m), 3.28 (1H, m), 4.88 (1H,
dd, J ) 11.5, 3 Hz), 4.96 (1H, d, J ) 5.5 Hz), 5.04 (1H, m),
5.36 (1H, d, J ) 8.5 Hz), 5.52 (1H, brd, J ) 9 Hz), 5.57 (1H, d,
J ) 8.5 Hz), 5.72 (2/3H, d, J ) 2 Hz), 5.82 (1/3H, brs), 5.92

882 J ournal of Natural Products, 2001, Vol. 64, No. 7
Singh et al.
(3H, m), 3.07 (3H, s, N-CH3), 3.57 (1H, m), 5.12 (1H, brm),
5.37 (1H, brm), 5.88 (1H, m), 5.90 (1H, d, 9.5 Hz), 5.95 (1H, d,
J ) 9 Hz), 5.97 (1H, d, J ) 2.0 Hz), 6.09 (1H, d, J ) 2.0 Hz),
6.32 (1H, d, J ) 9.0 Hz), 6.82 (1H, dd, J ) 8.0, 2.0 Hz), 6.92
(1H, t, J ) 8 Hz), 7.09 (1H, d, J ) 7.0 Hz), 7.15 (2H, dd, J )
8, 2 Hz), 7.20 (1H, d, J ) 7 Hz), 7.22 (1H, dd, J ) 8, 2 Hz),
7.34 (1H, d, J ) 7.5 Hz), 7.38 (1H, brs), 7.60 (2H, s), 7.88 1H,
brs), 7.90 (2H, s), 9.96 (1H, brs); HR-ESI-FTMS (m/z) 961.1330
[(M + H)+, calcd for C45H37Cl4N6O10 961.1325]. 8: tR 8.6 min
(Zorbax RX C-8, 4.6 × 250 mm, 30% aqueous CH3CN + 0.1%
TFA, 1 mL/min); 1H NMR (DMSO-d6, δ) Tyr (B), 2.50 (3H, brs,
N-CH3), 3.04 (2H, m, H2-3), 3.82 (1H, m, H-2), 6.92 (2H, d, J
) 8 Hz, H-6,8), 7.21 (2H, d, J ) 8.4 Hz, H-5,9); dichlorohy-
droxyphenyl glycine (C), 5.06 (1H, d, J ) 8 Hz, H - 2), 7.25
(2H, s, H-4,8), 9.07 (1H, d, J ) 7.6 Hz, NH); dihydroxyphenyl
glycine (D), 5.55 (1H, d, J ) 8.8 Hz, H-2), 5.95 (1H, brs, H-4),
6.96 (1H, brs, H-8), 8.60 (1H, d, J ) 9.2 Hz, NH); dichlorohy-
droxyphenyl glycine (E), 5.43 (1H, d, J ) 8.4 Hz, H-2), 7.30
(2H, s, H-4,8), 8.35 (1H, d, J ) 7.6 Hz, NH); Trp (F), 2.99 (1H,
m, H-3), 3.23 (1H, m, H-3), 4.39 (1H, m, H-2), 6.94 (1H, t, J )
8 Hz, H-6), 7.01 (1H, d, J ) 7.2 Hz, H-7), 7.26 (1H, d, J ) 9.2
Hz, H-5), 7.48 (1H, brs, H-2), 10.52 (1H, brs, H-1); HR-ESI-
FTMS (m/z) 979.1442 [(M + H)+, calcd for C45H39Cl4N6O11
979.1431). 9: tR 7.3 min (Zorbax RX C-8, 4.6 × 250 mm, 30%
Refer en ces a n d Notes
(1) For reviews see: (a) Brown, P. O.Curr. Top. Microbiol. Immunol.
1990, 157, 19. (b) Goff, S. P. Annu. Rev. Genet. 1992, 26, 527. (c) Katz,
R. A.; Skalka, A. M. Annu. Rev. Biochem. 1994, 64, 133. (d) Farnet,
C. M.; Wang, B.; Lipford, J . R.; Bushman, F. D. Proc. Natl. Acad.
Sci. U.S.A. 1996, 90, 2399. (e) Hazuda, D. J .; Felock, P.; Witmer, M.;
Wolfe, A.; Stillmock, K.; Grobler, J . A.; Espeseth, A.; Gabryelski, L.;
Schleif, W.; Blau, C.; Miller, M. D. Science 2000, 287, 646.
(2) (a) Singh, S. B.; Zink, D. L.; Goetz, M. A.; Dombrowski, A. W.;
Polishook, J . D.; Hazuda, D. L. Tetrahedron Lett. 1998, 39, 2243. (b)
Hazuda, D. J .; Uncapher, C. U.; Felock, P.; Hastings, J .; Pramanik,
B.; Wolfe, A.; Bushman, F.; Farnet, C.; Goetz, M.; Williams, M.;
Silverman, K.; Lingham, R.; Singh, S. Antiviral Chem. Chemother.
1999, 10, 63.
(3) (a) Singh, S. B.; Zink, D.; Polishook, J .; Valentino, D.; Shafiee, A.;
Silverman, K.; Felock, P.; Teran, A.; Vilella, D.; Hazuda, D. J .;
Lingham, R. B. Tetrahedron Lett. 1999, 40, 8775. (b) Singh, S. B.;
Felock, P.; Hazuda, D. J . Bioorg. Med. Chem. Lett. 2000, 10, 235.
(4) Hazuda, D. J .; Hashings, J . C.; Wolfe, A. L.; Emini, E. A. Nucleic
Acids Res. 1994, 22, 1121.
(5) (a) Kaneko, I.; Fearon, D. T.; Austen, K. F. J . Immunol. 1980, 124,
1194. (b) Kaneko, I.; Kamoshida, K.; Takahashi, S. J . Antibiot. 1989,
42, 236. (c) Seto, H. Pure Appl. Chem. 1989, 61, 365. (d) Seto, H.;
Fujioka, T.; Furihata, K.; Kaneko, I.; Takahashi, S. Tetrahedron Lett.
1989, 30, 4987.
(6) (a) Tanaka, H.; Matsuzaki, K.; Nakashima, H.; Ogino, T.; Matsumoto,
A.; Ikeda, H.; Woodruff, H. B.; mura, S. J . Antibiot. 1997, 50, 58. (b)
Matsuzaki, K.; Ogino, T.; Sunazuka, T.; Tanaka, H.; Oh mura, S. J .
Antibiot. 1997, 50, 66. (c) Matsuzaki, K.; Ikeda, H.; Ogino, T.;
Matsumoto, A.; Woodruff, H. B.; Tanaka, H.; Oh mura, S. J . Antibiot.
1994, 47, 1173. (c) Gouda, H.; Matsuzaki, K.; Tanaka, H.; Hirono, S.;
Oh mura, S.; McCauley, J . A.; Sprengeler, P. A.; Furst, G. T.; Smith,
A. B., III. J . Am. Chem. Soc. 1996, 118, 13087.
(7) Tachikawa, K.; Hasumi, K.; Endo, A. Thrombosis Res. 1997, 87, 571.
(8) (a) Carbonnelle, A.; Zamora, E. G.; Beugelmans, R.; Roussi, G.
Tetrahedron Lett. 1998, 39, 4471. (b) Beugelmans, R.; Roussi, G.;
Zamora, E. G.; Carbonnelle, A. Tetrahedron 1999, 55, 5089. (c) Elder,
A. M.; Rich, D. H. Org. Lett. 1999, 1, 1443.
(9) J ayasuriya, H. J .; Salituro, G. M.; Smith, S. K.; Heck, J . V.; Gould,
S. J .; Singh, S. B.; Homnick, C. F.; Holloway, M. K.; Pitzenberger, S.
M.; Patane, M. A. Tetrahedron Lett. 1998, 39, 2247. The sample used
for the rearrangement studies in this paper was isocomplestatin
instead of complestatin.
(10) Singh, S. B.; J ayasuriya, H.; Hazuda, D. J .; Felock, P.; Homnick, C.
F.; Sardana, M.; Patane, M. A. Tetrahedron Lett. 1998, 39, 8769. The
sample used for the hydrolytic studies in this paper was isocom-
plestatin instead of complestatin.
(11) Hegde, V. R.; Dai, P.; Patel, M.; Gullo, V. P. Tetrahedron Lett. 1998,
39, 5683.
1
aqueous CH3CN + 0.1% TFA, 1 mL/min); H NMR (DMSO-
d6, δ) chemical shifts are reported as an average of the two
cis-trans tertiary N-methyl amide isomers, Tyr (B), 2.48 (3H,
brs, N-CH3), 3.01 (2H, m, H2-3), 3.55 (1H, m, H-2), 6.85 (2H,
d, J ) 8 Hz, H-6,8), 7.17 (2H, d, J ) 8.4 Hz, H-5,9);
dichlorohydroxyphenyl glycine (C), 5.08 (1H, brs, H-2), 7.22
(2H, s, H-4,8), 8.83 (1H, brs, NH); dihydroxyphenyl glycine (D),
5.46 (1H, d, J ) 9.2 Hz, H-2), 6.05 (1H, brs, H-4), 6.62 (1H,
brs, H-8), 8.69 (1H, d, J ) 8.0 Hz, NH); dichlorohydroxyphenyl
glycine (E), 5.35 (1H, d, J ) 8.0 Hz, H-2), 7.30 (2H, s, H-4,8),
8.56 (1H, d, J ) 8.0 Hz, NH); Trp (F), 2.99 (1H, m, H-3), 3.23
(1H, m, H-3), 4.39 (1H, m, H-2), 6.94 (1H, t, J ) 8 Hz, H-6′),
7.01 (1H, d, J ) 7.2 Hz, H-7′), 7.27 (1H, d, J ) 8.4 Hz, H-5),
7.47 (1H, brs, H-2′), 10.61 (1H, brs, H-1′); HR-ESI-FTMS (m/
z) 979.1444 [(M + H)+, calcd for C45H39Cl4N6O11 979.1431).
10: tR 5.6 min (Zorbax RX C-8, 4.6 × 250 mm, 40% aqueous
1
CH3CN + 0.1% TFA, 1 mL/min); H NMR (DMSO-d6, δ) Tyr
(B), 2.48 (3H, brs, N-CH3), 3.06 (2H, m, H2-3), 3.77 (1H, m,
H-2), 6.96 (2H, d, J ) 8 Hz, H-6,8), 7.24 (2H, d, J ) 8.4 Hz,
H-5,9); dihydroxyphenyl glycine (D), 5.48 (1H, d, J ) 8.8 Hz,
H-2), 6.07 (1H, d, J ) 2 Hz, H-4), 6.93 (1H, d, J ) 2 Hz, H-8),
8.68 (1H, d, J ) 8.4 Hz, NH); dichlorohydroxyphenyl glycine
(E), 5.25 (1H, d, J ) 8.4 Hz, H-2), 7.35 (2H, s, H-4,8), 8.53
(1H, d, J ) 8.8 Hz, NH); Trp (F), 3.02 (1H, t, J ) 12.8 Hz,
H-3), 3.26 (1H, dd, J ) 15.2, 4.8 Hz, H-3), 4.44 (1H, dd, J )
11.6, 6 Hz, H-2), 6.94 (1H, t, J ) 8 Hz, H-6′), 7.03 (1H, d, J )
6.8 Hz, H-7′), 7.27 (1H, d, J ) 7.6 Hz, H-5), 7.47 (1H, brs, H-2′),
10.59 (1H, brs, H-1′); HR-ESI-FTMS (m/z) 762.1757 [(M + H)+,
calcd for C37H34Cl2N5O9 762.1734].
(12) The authentic samples of complestatin and chloropeptin I were not
available to us.
(13) We encountered some problems in measurement of optical rotation
of the buff-colored DMSO solutions of isocomplestatin and chloropep-
tin I. These solutions in polarimeter cells with 1 dm path length
caused low-energy (less than 60%) deflection and fluctuation presum-
ably due to high viscosity and color of the sample. This problem was
circumvented by substitution of cells with 1 cm path length. This
phenomenon could lead to some inconsistencies in the observed
rotations of these compounds.
(14) An attempt to convert isocomplestatin to complestatin by heating at
200 °C in DMSO was not successful due to decomposition.
(15) Kock, M.; Kessler, H.; Seebach, D.; Thaler, A. J . Am. Chem. Soc. 1992,
114, 2676.
(16) Aldrich Handbook of Fine Chemicals; 1998-1999.
Ack n ow led gm en t. The authors wish to thank Dr. Z. Guan
for high-resolution mass spectral data.
NP000632Z