The synthesis and screening of 1,4,5,8-naphthalenetetracarboxylic diimide-peptide conjugates with antibacterial activity

Article abstract of DOI:10.1016/S0968-0896(01)00108-0

We have employed an initial combinatorial approach followed by systematic lead optimization to investigate a series of novel molecules that exhibit antimicrobial activity against Gram-negative and Gram-positive bacteria. The new molecules contain various sequences of amino acids, generally L-lysine and glycine, attached to the 1,4,5,8-naphthalenetetracarboxylic diimide aromatic unit. Systematic structure-activity studies found that increasing positive charge enhanced activity and molecules containing one naphthalenetetracarboxylic diimide unit as well as at least seven lysine residues were optimum for antimicrobial activity. The naphthalenetetracarboxylic diimide derivatives were found to be inactive against mammalian cell lines, making them excellent antimicrobial candidates. Our results indicate that combining positive charge with aromatic and/or hydrophobic elements may be an interesting new approach to antimicrobial agents and adds an important new dimension to the field of cationic peptides.

Full text of DOI:10.1016/S0968-0896(01)00108-0

Bioorganic & Medicinal Chemistry 9 (2001) 2015–2024
The Synthesis and Screening of
,4,5,8-Naphthalenetetracarboxylic Diimide–Peptide
Conjugates with Antibacterial Activity
1
a
a
b
a,
Chandra T. Miller, Ramal Weragoda, Elzbieta Izbicka and Brent L. Iverson *
aDepartment of Chemistry and Biochemistry, The University of Texas at Austin, Austin, TX 78712, USA
bInstitute for Drug Development, San Antonio, TX 78245, USA
Received 21December 2000; accepted 9 March 2001
Abstract—We have employed an initial combinatorial approach followed by systematic lead optimization to investigate a series of
novel molecules that exhibit antimicrobial activity against Gram-negative and Gram-positive bacteria. The new molecules contain
various sequences of amino acids, generally l-lysine and glycine, attached to the 1,4,5,8-naphthalenetetracarboxylic diimide aro-
matic unit. Systematic structure–activity studies found that increasing positive charge enhanced activity and molecules containing
one naphthalenetetracarboxylic diimide unit as well as at least seven lysine residues were optimum for antimicrobial activity. The
naphthalenetetracarboxylic diimide derivatives were found to be inactive against mammalian cell lines, making them excellent
antimicrobial candidates. Our results indicate that combining positive charge with aromatic and/or hydrophobic elements may be
an interesting new approach to antimicrobial agents and adds an important new dimension to the field of cationic peptides. # 2001
Elsevier Science Ltd. All rights reserved.
Introduction
The homopolymer poly-l-lysine also displays anti-
bacterial activity against both Gram-positive and
Gram-negative bacteria. The naturally occurring
homopolymer contains anywhere between 25 and 30
9
The emerging crisis in antibiotic resistance has renewed
interest in the investigation of novel chemical structures
1
,2
with antimicrobial activity. Two major approaches
for combating resistant microbial pathogens include (1)
targeting the ‘resistance factors’ and (2) the generation
of novel antimicrobial agents, ideally with new modes of
lysine residues and has been isolated from Streptomycin
9
,10
albulus.
A minimum of 10 lysine residues have been
found necessary to inhibit bacterial growth using syn-
thetic peptides, and increasing the number of lysine
11
residues enhanced bacterial growth inhibition.
2
action.
Among the different classes of antimicrobial agents
currently being investigated are the cationic peptides
that exhibit strong activity against both Gram-negative
and Gram-positive bacteria. This class of antibiotics
currently consists of a rapidly expanding list of natu-
rally occurring cationic peptides that have been isolated
from bacteria, mammals, amphibians, plants, fungi,
The biological activity of N-substituted naphthalimides
1
2
has been studied since the early 1940’s. Biological
potencies of simple analogues include anti-tumor, anti-
neoplastic and anesthetic activity. To the best of our
knowledge, however, there has never been a systematic
study examining the antimicrobial properties of these
species. More recently, naphthalimide and 1,4,5,8-
naphthalenetetracarboxylic diimides (NDI) derivatives
have also attracted attention because of their ability to
3,4
insects, crustaceans and humans. The cationic peptides
as a class possess two elements in common, the presence of
both basic (i.e., arginine and lysine) and hydrophobic
amino acids. These cationic peptides are thus amphipathic
1
3,14
intercalate into double-stranded DNA.
1,3ꢀ5
molecules.
The antibacterial mode of action has been
Herein is reported the synthesis and preliminary eva-
luation of the antimicrobial properties of a series of
novel cationic NDI–peptide hybrid molecules. Initial
combinatorial studies identified lead structures that
were optimized using systematic structure–activity
approaches. Several of the more positively-charged
molecules exhibited activity against Gram-negative and
generally attributed to the formation of ion-channels or
other types of membrane disruption/disorganization,
2ꢀ8
although research remains active in this area.
*
8
Corresponding author. Tel.: +1-512-471-5053; fax: +1-512-471-
696; e-mail: biverson@utxvms.cc.utexas.edu
0
968-0896/01/$ - see front matter # 2001Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(01)00108-0

2016
C. T. Miller et al. / Bioorg. Med. Chem. 9 (2001) 2015–2024
especially Gram-positive bacteria, with no detectable
activity against mammalian cell lines in preliminary
studies. Taken together, our results indicate the cationic
peptide motif of antimicrobial activity is not restricted
to molecules composed of amino acids only, perhaps
paving the way for new generations of novel cationic
molecules with potent antimicrobial activity.
properties. The library was based on a so-called dimer,
in which two NDI units are separated by four amino
acids (Fig. 1), and glycine residues on each terminus.
Nineteen amino acids were used as building blocks for
the polypeptides in the library. Cysteine was not inclu-
ded in the study due to potential complications arising
from disulfide bond formation. Fmoc-NovaSyn TGA
resin was evenly divided into 19 reaction vessels,
wherein each vessel represented an amino acid in the
tetrapeptide. The NDI was attached to glycine prior to
solid-phase synthesis. Consequently, only three amino
acid positions were varied in the initial library. The
compounds were not mixed after the addition of the
third amino acid. In this way, a total library of
approximately 6900 dimers were divided among nine-
teen vessels with 361compounds per vessel. The mate-
rial from each vessel was cleaved from the resin then
placed on filter paper at known concentrations in
preparation for testing.
Results
To investigate the antimicrobial activity of compounds
based on peptide hybrids of the NDI moiety, a series of
protected NDI derivatives (2–5) were synthesized in
preparation for solid-phase synthesis (Schemes 1–3).
Split synthesis was employed to create a library of
compounds varying in hydrophobic and electrostatic
The Kirby–Bauer disk assay was used to identify active
5ꢀ17
1
mixtures.
In these assays, E. coli and Bacillus subtilis
were used as representative Gram-negative and Gram-
positive bacterial strains, respectively. The antibiotic
streptomycin was used as a calibration in all tests.
Screening of the 19 different libraries indicated that at
300 nmol on the filter, only the mixture containing a
lysine residue in the X3 position of the tetrapeptide
exhibited slight antibacterial activity in B. subtilis. No
activity was observed in E. coli. The 19 libraries were
not deconvoluted further, but rather, a series of studies
were initiated to identify parameters important for
antimicrobial activity.
A number of molecular parameters were investigated,
including the influence of (1) linker amino acid
sequence; (2) increasing positive charge; (3) arginine
versus lysine; (4) l versus d amino acids; (5) altering the
number of NDI units; (6) N-alkylamino substituents
versus peptide containing substituents; (7) variation of
the aromatic moiety. The qualitative and quantitative
Scheme 1. Synthesis of diimide amino acid precursor. Reagents: (i) b-
alanine benzyl ester p-toluene sulfonate salt, BocN(CH
NEt; (ii) H , Pd/C; (iii) TFA/CH Cl ; (iv) Fmoc-Gly-OC
,6-lutidine. For compounds with lysine and arginine residues, Fmoc-
2 2 2 2
) NH , i-Pr -
a
2
2
2
6 6
F , HOBT,
a
2
Lys(Boc)-OC
6
F
6
6 6
and Fmoc-Arg(PMC)-OC F were substituted.
0
0
Scheme 2. Synthesis of N,N-bis(2,2 -ethylamine) 1,4,5,8-naphthalene-
Cl
Scheme 3. Synthesis of N,N-bis(2,2 -dimethylmino)-1,4,5,8-naphtha-
lenetetracarboxylic diimide. (i) 2-(dimethylamino)ethylamine, THF.
tetracarboxylic diimide. (i) TFA/CH
2
2
.
n
Figure 1. Variation of three of the four amino acids in the polypeptide linker where X represents one of the 19 amino acids.

C. T. Miller et al. / Bioorg. Med. Chem. 9 (2001) 2015–2024
2017
structure–activity relationship (SAR) of each derivative
was explored first using the Kirby–Bauer assay. The
amounts used for each test ranged from 33–300 nmol
and were used to assay the biological activity against
two bacterial reference strains, E. coli and B. subtilis.
Each study was repeated five times to assess reproduci-
bility. On average, the studies resulted in zones of
inhibition with a difference of ꢁ2 mm.
nmol disks, and activity against E. coli on the 100 nmol
disks. The bottom line is that within this series of
molecules, a minimum of two positive charges are nee-
ded to show activity against B. subtilis and at least five
positive charges are needed to show activity against E.
coli in this assay system.
Compounds 35 and 36 contain only d-lysine, but are
otherwise identical to the l-lysine containing com-
pounds 27 and 33, respectively. In both cases, the d-
lysine containing analogues displayed antimicrobial
activities that were indistinguishable from the molecules
containing l-lysine, indicating that the stereochemistry
of the lysine residue is not relevant to overall activity. In
addition, l-arginine was used in place of l-lysine in
compounds 37 and 38. Here, the arginine-containing
compounds displayed similar but measurably lower
activity than the reference compounds 27 and 33.
Finally, as important controls, an NDI unit alone (39)
and lysine containing peptides of differing lengths were
analyzed (40–43), but failed to show any measurable
antimicrobial activity. These latter results are consistent
with literature precedent in which at least 10 lysine resi-
dues are required for significant antimicrobial activity in
Fifty NDI–peptide conjugates, 8–39 and 44–61, ranging
in selected properties were prepared using solid-phase
chemistry and analyzed in an effort to probe potential
relationships between chemical structure and anti-
microbial activity. RP HPLC was used to purify each
compound and molecular weight was characterized by
mass spectrometry.
Six compounds were initially synthesized (8–13), vary-
ing in the location of a single l-lysine residue in the
context of five glycine residues and two NDI aromatic
units. The resulting compounds were inactive in both E.
coli and B. subtilis at 300 nmol per disk (Table 1). Fif-
teen dimers (14–28) were then synthesized containing
the various possible sequences of two l-lysine residues
in the context of four glycine residues and two NDI
aromatic units. At 300 nmol, all but one, compound 28,
displayed similar activities against B. subtilis, but were
inactive against E. coli. Compound 28 contained lysine
residues on both the N- and C-termini and was com-
pletely inactive. It is unclear why having the two lysine
residues in the terminal positions precluded activity with
9
peptides.
A series of derivatives with a single NDI unit attached
to various numbers of glycine, l-lysine, d-lysine and l-
arginine residues were prepared (44–60) and analyzed
for antimicrobial activity using the same Kirby–Bauer
disk diffusion assay (Table 2). The results with these so-
called ‘NDI monomers’ were entirely consistent with the
results obtained with the NDI dimers listed in Table 1.
In particular, with the NDI monomers, increasing anti-
microbial activity was correlated with increasing posi-
tive charge (44–54). Again, a minimum of two positive
charges was needed for activity against B. subtilis and at
least five positive charges were needed for activity
against E. coli. Increasing the number of l-lysine resi-
dues to greater than six had only a minimal effect on
activity in this assay. Substitution of l-lysine residues
with d-lysine (55) or l-arginine (56) also had little effect.
In addition, the NDI unit position within the peptide
conjugate did not matter (57–60) and having three NDI
units in the context of a longer peptide showed no
advantage (61) in terms of antimicrobial activity.
28, while the activities of the all other compounds in the
series, 14–27, were not noticeably dependent on the
location of the lysine residues.
The results using compounds 14–27 indicated that
greater positive charge (+2 vs +1) produced generally
greater activity, so the influence of positive charge was
investigated in more detail using compounds 29–34.
Compounds 29–34 have total charges ranging from ꢀ2
to +6. A clear trend toward greater activity with
increasing positive charge was observed in the series.
Particularly noteworthy are compounds 33 and 34, with
total charges of +5 and +6, respectively, which showed
substantial activity against B. subtilis even with the 33
Table 1. Observed and calculated HR-MS (FAB) data for selected
compounds
As a first attempt to see if other naphthalene derivatives
would function in a manner similar to the NDI unit,
another aromatic moiety, 1,8-dialkoxynaphthalene, was
used in place of the NDI unit in compounds 62 and 63.
No activity was observed with either compound.
Finally, two NDI derivatives were investigated in which
aminoethyl (6) or (dimethylamino)ethyl (7) groups were
attached in place of the peptide. These molecules dis-
played the highest activity of any of the derivatives in
the Kirby–Bauer disk diffusion assay, against both E.
coli and B. subtilis.
Compound
Molecular weight
Molecular formula
Observed
Calculated
14
27
28
33
34
46
47
49
53
54
61
62
63
1230.4741
1229.4623
1229.4674
1442.6892
1514.7654
638.29337
766.38857
1022.5802
1335.7904
1663.0634
1963.7926
944.6322
1230.4730
1229.4652
1229.4652
1442.6857
1514.7670
638.29384
766.38880
1022.5787
1335.7901
1663.0642
1963.7869
944.6296
C
C
C
C
58
58
58
70
H
H
H
H
66
65
65
92
N
N
N
N
14
14
14
17
O
O
O
O
17
17
17
17
C
74
H
102
N
18
O
17
C
C
C
31
H
40
N
N
7
O
O
O
8
37
H
52
9
9
49
H
N
76 13
11
C
C
C
63
79
93
H
103
H
145
H
109
N
18
O
23
O
23
O
11
O
14
O
14
16
26
9
N
N
N
N
In order to probe further the antimicrobial activities of
the NDI–peptide conjugates, minimum inhibitory con-
C
47
H
82
15ꢀ17
centrations (MIC’s) were determined for selected
compounds varying in their number of NDI units and
1357.8600
1357.86115
C
70
H
113
13

2018
C. T. Miller et al. / Bioorg. Med. Chem. 9 (2001) 2015–2024
cationic properties (Table 3). The same general trends
were seen in the MIC’s as with the Kirby–Bauer studies.
In general, activity increased with increasing positive
charge, and was rather insensitive to the number of
NDI units (as long as at least one was present), or the
stereochemistry of the lysine residues. However, as
opposed to the Kirby–Bauer assays, there was a dra-
matic difference between derivatives with six versus seven
l-lysine residues. The molecules with seven l-lysine
residues, 52 and 53, showed significantly higher activity
than the corresponding molecule with only six l-lysine
residues (51). Increasing to 10 l-lysine residues, 54, did
not increase activity further. Interestingly, in B. subtilis,
compounds 52, 53 and 54 containing 7 and 10 lysine
residues, respectively, displayed the same MIC, 3.1 mM,
as the antibiotic streptomycin.
NDI–peptide conjugates 52, 53, and 54. The dis-
crepancy in results between measured MICs and Kirby–
Bauer assays can be attributed to the differences in the
molecular weights of the compounds. The Kirby–Bauer
method is highly dependent on diffusion and thus the
molecular weight and solubility of a compound,
whereas molecular weight is less important when deter-
mining the MIC by broth dilution since diffusion is no
longer a significant consideration. This rationale may
also explain why the zone of inhibition of the most
active NDI–peptide conjugates, 52, 53, and 54, were less
than streptomycin 66 (data not shown) whereas the
MIC’s were identical in B. subtilis.
Poly-l-lysines ranging in molecular weights (2900,
111,000, and 134,000 g/mol) were also tested to deter-
mine the correlation between enhanced antimicrobial
activity and increasing molecular weight/positive
charge. Poly-l-lysine (2900 g/mol) 43 displayed an MIC
of 50 mM whereas the larger poly-l-lysine (111,000 and
In contrast to the Kirby–Bauer assay results, the MICs
in Table 4 indicate that the amino alkyl NDI derivatives
6
and 7 are significantly less active in culture than the
Table 2. Peptide and Dimer results; variation of charge, location of chromophore, increasing positive charge, d versus l amino acids and arginine
versus lysinea
E. coli
B. subtilis
3
00
100
33
300
100
33
Charge
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
GAGGGGAK
GAGGGKAG
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
1
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
1
8
1
8
7
9
1
7
1
1
1
11
11
1
NI
NI
NI
1
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
+1
+1
+1
+1
+1
+1
0 GAGGKGAG
1GAGKGGAG
2GAKGGGAG
3 KAGGGGAG
4 GAGKGKAG
5 GAKKGGAG
6 GAGKKGAG
7 KAKGGGAG
8 KAGKGGAG
9 GAKGGKAG
0 GAGGKKAG
1 KAGGGKAG
2 KAGGKGAG
3 GAGGGKAK
4 GAGGKGAK
5 GAGKGGAK
6 GAKGGGAK
7 GAKGKGAG
8 KAGGGGAK
9 GAEGEGAG
0 GAGGGGAG
1 GAKKKGAG
2 GAKKKKAG
3 KAKKKKAG
4 KAKKKKAK
5 GAKGKGAG (D)
6 KAKKKKAG (d)
7 GARGRGAG
8 RARRRRAG
9 H-A-OH
0
0
NI
NI
NI
NI
NI
NI
+2
NI
NI
NI
NI
NI
NI
+2
+2
+2
0
NI
NI
NI
NI
+2
0
0
0
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
+2
+2
0
NI
1
NI
NI
NI
NI
NI
NI
NI
+2
ꢀ2
0
1
2
8
0
1
1
8
11
NI
9
1
12
NI
12
NI
1
NI
NI
NI
NI
NI
8
NI
9
NI
NI
NI
NI
14
1
12
8
+6
NI
+5
+2
1
0
NI
10
NI
8
NI
NI
0
NI
NI
1
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
0
0 Lys–Lys–Lys
+3
+4
+5
+n
1 Lys–Lys–Lys–Lys
2 Lys–Lys–Lys–Lys–Lys
3 Poly-l-lysine (2900 g/mol)
aDiameters of zones of inhibition (DZI), mm 300, 100 and 33 nmol.

C. T. Miller et al. / Bioorg. Med. Chem. 9 (2001) 2015–2024
2019
1
34,000 g/mol) derivatives (64 and 65) exhibited an
dilutions starting at 100 mM were used with the MTT
18
MIC of 0.78 mM. Since the shorter poly-l-lysine (2900
g/mol) 43, containing 15 l-lysine residues, displayed
relatively poor MIC activity, these data confirm the
importance of the NDI moiety for enhancing activity in
small oligo-l-lysine peptide conjugates.
assay. None of the NDI–peptide conjugates showed
activity even at 100 mM. On the other hand, the ami-
noalkyl derivatives 6 and 7 displayed IC50 values of 1
mM and 390 nM in both cell lines. The mammalian cell
activity observed with 6 is consistent with literature
1
9,20
reports of this compound.
To probe the bactericidal activity of the most active
compounds, minimal lethal concentrations (MLC’s)
were measured for the most active NDI–peptide con-
jugates (52–54), the aminoalkyl NDIs (6, 7) and strep-
tomycin (66). Briefly, in these experiments, broth from
each B. subtilis culture of the MIC determination
experiments was plated to find the minimum con-
centration of compound in which all bacteria had been
killed, as opposed to just growth arrested, following
incubation with compound in solution. Interestingly,
the NDI–peptide conjugates displayed bactericidal
activity similar to that of streptomycin, but the ami-
noalkyl NDI derivatives 6 and 7 were not bactericidal even
at the highest concentrations examined, 100 mM. Thus,
these latter two compounds appear to be bacteriostatic, as
opposed to bactericidal.
Discussion
Positively-charged NDI–peptide conjugates display
antimicrobial activity against the representative Gram-
negative E. coli and especially the Gram-positive B.
subtilis, with increasing activity being correlated to
increasing numbers of positively-charged lysine residues
in the conjugate. Maximal activity is seen with seven or
greater lysines. The activity is dependent on the pre-
sence of at least one NDI residue, but the position of
this unit within the molecule does not appear to be cri-
tical and additional NDI units provide no discernable
benefit. Importantly, the NDI–peptide conjugates dis-
play bactericidal activity and have no activity against
the mammalian cell lines tested.
The activities of selected NDI derivatives in mammalian
cell lines were investigated using normal (NCM-460)
and cancer (HT-29) cell lines (Table 5). Two-fold serial
Taken together, the data for all of the NDI–peptide
conjugates investigated provide circumstantial support
Table 3. Monomer results; variation of charge, location of chromophore, increasing positive charge, d versus l amino acids and arginine versus
lysinea
E. coli
B. subtilis
3
00
100
33
300
100
33
Charge
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
6
6
6
6
7
4 GAG
5 GAK
6 KAK
7 KKAK
8 KKKAK
9 KKKKAK
0 KKKKKAG
1 KKKKKAK
2 KKKKKKAK
3 KKKKKKKAG
4 KKKKKKKKKAK
5 KKKKKAG (d)
6 RRRRRAG
7 KKKKAKK
8 KKKAKKK
9 KKAKKKK
0 KAKKKKK
1 GAKGKGAKGKGAG
2 KKKKDK
NI
NI
NI
NI
NI
8
1
1
1
1
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
1
1
1
14
NI
NI
NI
NI
0
0
+1
7
NI
1
2
2
1
2
1
10
8
NI
NI
+5
3
1
2
1
NI
NI
NI
NI
1
0
1
1
1
7
NI
NI
NI
1
2
1
1
1
1
2
1
0
1
1
3
2
1
1
1
6
NI
1
11
11
11
12
NI
NI
NI
20
1
9
8
9
NI
NI
NI
NI
NI
NI
NI
14
13
14
13
1
12
12
12
11
8
12
10
10
+6
+6
+6
+6
9
NI
NI
NI
1
1
9
1
NI
NI
NI
NI
NI
NI
NI
+5
+6
3 KDKKKKDK
2+
[H
3
NA1NH
3
]
NA1N(CH
6
1
0
5
NI
20
25
9
[CH
3
)
2
3
)
2
]2+
6
1
20
1
aDiameters of zones of inhibition (DZI), mm 1300, 100 and 33 nmol.

2020
C. T. Miller et al. / Bioorg. Med. Chem. 9 (2001) 2015–2024
for a mode of action involving the bacterial membranes
or components thereof, as it is unlikely that highly
positively-charged molecules can penetrate into the
interior of bacterial cells. In addition, the observed lack
of amino acid sequence dependence and equivalent
activity of d-lysine containing derivatives argue against
a specific receptor mediated process. Rather, the activity
seems more correlated with the overall physical proper-
ties of the molecules, namely the positive charges of the
lysine residues coupled to the aromatic surfaces of the
NDI unit. Since no activity with the NDI–peptide con-
jugates was seen in mammalian cells, similar membrane-
based interactions presumably do not occur.
derivatives 6 and 7. Only 6 and 7 displayed relatively
strong activity in mammalian cells. Sami et al. have
shown that the mammalian cell activities of naphthali-
mide analogues are related to DNA intercalation and
2
1ꢀ23
enhanced DNA interactions parallel cytotoxicity.
Assuming a DNA intercalation mechanism of action,
the 3-fold difference in IC50 values observed in our
study is likely due to greater membrane permeability of
the tertiary amino groups of 6 compared to the primary
amino groups of 7.
To the best of our knowledge, the antimicrobial prop-
erties of the NDI species 6 and 7 have not been repor-
ted. Although 6 and 7 both show strong activity in the
Kirby–Bauer assay, they appeared to be bacteriostatic,
not bactericidal. Such distinct behavior underscores the
The activity of the NDI–peptide conjugates can be
contrasted with the activities of the aminoalkyl NDI
Table 4. Minimum inhibitory and minimum lethal concentrations of selected compounds reported in mM and mg/mL
E. coli
B. subtilis
MIC (mM)
MIC (mg/mL)
MIC (mM)
MIC (mg/mL)
MLC (mM)
MLC (mg/mL)
3
4
4
4
5
5
5
5
5
3
3
5
6
7
4
4
4
4
6
6
6
4 KAKKKKAK
50
>100
>100
76
>77
>90
>100
5125
5125
58
64
67
72
72
42
18
20
>41
>54
>66
145
87
25
>100
>100
>103
26
38
>77
>90
50
7 KKAK
8 KKKAK
9 KKKKAK
51
0 KKKKKAG
5 KKKKKAG (d)
1 KKKKKAK
2 KKKKKKAK
3 KKKKKKKAG
3 KAKKKKAG
6 KAKKKKAG (d)
4 KKKKKKKKKAK
50
50
50
25
50
50
50
25
26
25
29
4.2
1
72
72
3.1
3.14.2 2.5
50
50
3.15.2 2.5
25
25
12.5
19
20
25
1
2+
[H
[CH
0 Lys–Lys–Lys
1 Lys–Lys–Lys–Lys
3
NA1NH
3
]
50
50
8.8
10
>100
>100
>35
>41
)
2
NA1N(CH
3
)
2
]2+
3
>100
>100
>100
50
>100
>100
>100
50
>41
>54
>66
145
87
2 Lys–Lys–Lys–Lys–Lys
3 Poly-l-lysine (2900 g/mol)
4 Poly-l-lysine (111,000 g/mol)
5 Poly-l-lysine (134,000 g/mol)
6 Streptomycin
0.78
0.78
0.39
0.78
0.78
3.1
104
1.1
104
4.6
25
36
Table 5. Mammalian cell results obtained from the MTT assay of various compounds tested against human normal and colon carcinoma cell lines
NCM-460
IC50
HT-29
IC50
6
5
5
5
2
6
7
KAK
>100 mM
>100 mM
>100 mM
>100 mM
>100 mM
1 mM
>100 mM
>100 mM
>100 mM
>100 mM
>100 mM
1 mM
0 KKKKKAG (l)
5 KKKKKAG (d)
4 KKKKKKKKKAK
7 GAKGKGAG
2
+
[H
3
NA1NH
3
]
([CH
3
)
2
NA1N(CH
3
)
2
]2+
390 nM
390 nM

C. T. Miller et al. / Bioorg. Med. Chem. 9 (2001) 2015–2024
2021
likely different mode of action of 6 and 7 compared to
the NDI–peptide conjugates such as 52–54.
round-bottom flask equipped with gas inlet stopper and
dropping funnel. Di-tert-butyl dicarbonate (34 mL, 147
mmol) was suspended in 100 mL of dichloromethane
and slowly added to the flask via dropping funnel. The
reaction was stirred 18 h and concentrated in vacuo.
Conclusion
H O (100 mL) was added to the flask and the filtrate
2
The NDI–peptide conjugates such as 52–54 should be
considered as starting points for further antibacterial
drug development due to relatively strong antimicrobial
activity coupled to the complete lack of observed activ-
ity in mammalian cell lines. At this point, it is tempting
to ascribe the antimicrobial activity to membrane inter-
actions/disruption in analogy to previous studies with
was partitioned between CH Cl and 10% Na CO . The
2
2
2
3
organic layer was washed with Na CO (4ꢂ100 mL),
2
3
water (4ꢂ100 mL), brine (150 mL), dried over anhy-
drous sodium Na SO , filtered and concentrated to
2
4
1
produce a viscous yellow oil (20.3 g, 86%). H NMR
(250 MHz, CDCl ) d 5.65 (d br, 1H), 3. 18 (s, 1H), 2.82
3
(q, 2 H, J=6 Hz), 2.78 (t, 2 H, J=6 Hz), 1.46 (s, 9
13
3
,4
poly-l-lysine and cationic peptides.
Nevertheless,
H); C NMR (63 MHz, CDCl ) d 156.1, 42.9, 41.1, 30.8,
3
detailed mechanistic studies with the most active NDI–
peptide conjugates are currently being planned that will
address in detail their mode of action.
28.0; HR-MS (FAB) m/z 161.1284 (161.1290 calcd for
+
C H N O , M +H).
2
7
17
2
0
N-(2-tert-Butoxycarbonylaminoethyl)-N -(2-carboxyethyl)-
1,4,5,8-naphthalenetetracarboxylic diimide (1). 1,4,5,8-
naphthalenetetracarboxylic dianhydride (34.06 g, 127
The studies described herein have demonstrated that the
relatively hydrophobic and electron deficient aromatic
NDI unit can be used to enhance dramatically the anti-
microbial activities of positively-charged peptides, and
this activity enhancement is likely independent of any
preformed secondary structure. Such findings add a new
dimension to the field of antimicrobial cationic peptides.
Taken one step further, our results indicate that a
systematic investigation of novel hydrophobic and posi-
tively-charged units, including non-peptide systems, might
uncover a fertile new source of potent antimicrobial
agents.
mmol),
mono-tert-butoxycarbonylaminoethylamine
(20.31g, 12 7 mmol), b-alanine benzyl ester p-toluene-
sulfonate salt (44.63 g, 127 mmol) and N,N-diisopro-
plyethylamine (24 mL, 140 mmol) were suspended in 1 L
of i-PrOH. The mixture was heated at reflux for 24 h
then concentrated in vacuo to give a pink solid. The
solid was dissolved in CH Cl (550 mL) and the organic
2
2
layer was washed with 0.2 M sodium citrate buffer
(4ꢂ250 mL, pH 4.5), saturated NaHCO (5ꢂ250 mL),
3
brine (250 mL), and then dried over anhydrous Na SO .
4
2
The combined organic layers were concentrated to
afford a statistical mixture of products, including 1.
Experimental
General considerations
The resulting tan solid was suspended in absolute EtOH
(2 L) and purged with argon for 3 h. To the mixture was
added 10% Pd/C (15 g) and then the solution was
Proton and carbon NMR were obtained on a Bruker
AC 250 spectrometer or a Bruker AMX-500 spectro-
meter. Chemical shifts are reported in ppm relative to
tetramethylsilane (d units). Chemical ionization mass
spectra and fast atom bombardment (FAB) mass spec-
tra were recorded on a Finnigan TSQ-700 and Micro-
mass ZAB spectrometer at the Analytical Facility of
The University of Texas at Austin. Combinatorial
chemistry and parallel synthesis were performed manu-
ally on an apparatus designed in our laboratory. To
insure proper mixing of the reactants the apparatus was
attached to a teflon block, argon outlet, vacuum mani-
fold and connected to a shaker. HPLC data were
obtained on a Hewlett-Packard 1090 UV–Vis Spectro-
photometer. Unless indicated, all solvents and materials
were used without further purification. Amino acids and
coupling reagents were purchased from Novabiochem,
Advanced Chemtech and BACHEM. Anhydrous
charged with H at atmospheric pressure for 72 h. The
2
solvents were evaporated and the resulting solid was
dissolved in 200 mL of 20% TEA/CH Cl . The solution
2
2
was filtered through Celite and the filtrate was con-
centrated under vacuum onto 100 g of silica. The
resulting solid was purified by column chromatography
using a gradient from 0 to 10% MeOH in 90:10 CH Cl
2
2
/TEA. The product was taken from the second band,
concentrated and dissolved in 10% MeOH/CH Cl (200
2
2
mL) and HOAc (30 mL). Pentane (300 mL) was added
to the solution to facilitate precipitation of the product.
The mixture was sonicated, filtered, washed with more
pentane (30 mL) and dried in vacuo to yield 1 as a tan
1
powder (18.4 g, 30%). H NMR (250 MHz, DMSO-d )
6
d 8.61(s, 4H), 6.91(t, 1H , J=6 Hz), 4.25 (t, 2H, J=7.4
Hz), 4.12 (t, 2H, J=6 Hz), 3.27 (br, 2H), 2.61(t, 2H,
1
3
J=7.5 Hz), 1.21 (s, 9H); C NMR (63 MHz, DMSO-
d6) d 172.5, 162.2, 162.0, 155.8, 130.2, 130.1, 125.8,
125.6, 77.6, 37.6, 36.1, 32.1, 28.0; HR-MS (FAB) m/z
(
99.8%) DMF was purchased from Aldrich and stored
over molecular sieves. All reactions were performed
under argon atmosphere and glassware used for SPPS
were pre-treated with SIGMACOTE to prevent the
resin from sticking to the glass.
+
481.1467 (481.1486 calcd for C H N O , M +H).
3
2
4
23
8
N-2-(N-ꢀ-9-Fluorenylmethoxycarbonylglycyl)aminoethyl-
N -(2–carboethyl)-1,4,5,8-naphthalenetetracarboxylic dii-
0
mide (2). Compound 1 (2 g, 4.15 mmol) was suspended
Mono-tert-butoxycarbonylaminoethylamine
in CH Cl (12 mL) and was added slowly TFA (12 mL).
2
2
Ethylenediamine (57 mL, 852 mmol) was dissolved in
CH Cl (100 mL) and added to a 500-mL three-neck
The solution stirred for 30 min, concentrated and
residual TFA was removed by azeotropic evaporation
2
2

2022
C. T. Miller et al. / Bioorg. Med. Chem. 9 (2001) 2015–2024
from heptane (5ꢂ30 mL). The mixture was triturated
with ether (5ꢂ30 mL), filtered and dried in vacuo. The
resulting solid was dissolved in DMF (15 mL) and
Fmoc-Gly-OC F (1.92 g, 4.15 mmol) was added fol-
32.2, 27.4, 25.4, 20.8, 19.2, 18.1, 17.1, 16.1; d HR-MS
(FAB) m/z 1026.3705 (1026.3707 calcd for
+
C H N O S , M +H).
7
5
4
56
12 1
6
5
0
0
lowed by the addition of 1-hydroxybenzotriazole
HOBT, 0.561 mg, 4.15 mmol) and 2,6-lutidine (1.32 g,
2.5 mmol). The mixture was stirred for 18 h then
N N-bis(2,2 -tert-Butoxycarbonylaminoethyl)-1,4,5,8-naph-
thalenetetracarboxylic diimide. The bis-Boc compound
was formed as a by-product during the synthesis of 1.
Chromatographic separation of the statistical mixture
using a gradient from 0 to 10% MeOH in 90:10
CH Cl /TEA and isolation of the first band afforded a
(
1
poured into citric acid buffer (100 mL, 1 M, pH 4.5) and
filtered. The yellow solid was thoroughly rinsed with
H O (3ꢂ30 mL), ether (3ꢂ30 mL) and dried in the pre-
2
2
2
1
sence of P O in a vacuum dessicator for 7 h. The solid
2
tan solid in 15% yield. H NMR (250 MHz, CDCl ) d
5
3
was triturated with heptane (4ꢂ50 mL) (to remove the
8.75 (s, 4H), 4.87 (br, 2H), 4.38 (t, 4H, J=5.7 Hz), 3.57
3
1
excess pentafluorophenol and HOBT), filtered and dried
overnight to give 2 (2.5 g, 92%) as an orange powder.
(m, 4H), 1.22 (s, 18H); C NMR (63 MHz, CDCl ) d
3
163.14, 156.18, 130.94, 128.43, 128.25, 40.62, 39.19,
28.14, 21.67; HR-MS (FAB) 552.2208 (552.2220 calcd
1
H NMR (250 MHz, DMSO-d ) d 8.61(s, 4H), 8.02 (t,
6
+
1
H, J=6 Hz), 7.86 (d, 2H, J=7.3 Hz), 7.62 (d, 2H,
for C H N O , M +H).
2
8
32
4
8
J=7.3 Hz), 7.43 (m, 1H, J=6.4 Hz ), 7.39 (t, 2H, J=7.3
Hz), 7.27 (t, 2H, J=7.6 Hz), 4.24 (t, 2H, J=6.9 Hz ),
0
0
N,N -bis(2,2 -Ethylamine)-1,4,5,8-naphthalenetetracarb-
oxylic diimide (6). The above compound (1g, .18 1
mmol) was suspended in CH Cl (12 mL) and was
4
3
.15 (br, 2H), 4.06 (d, 2H, J=6.4 Hz ), 3.60 (br, 1H),
.50 (d, 2H, J=6.4 Hz ), 3.47 (m, 2H), 2.61(t, 2H,
2
2
1
3
J=7.6 Hz); C NMR (63 MHz, DMSO-d ) d 172.4,
added dropwise TFA (12 mL). The solution stirred for
30 min, was concentrated in vacuo and the remaining
TFA was removed by azeotropic evaporation (5ꢂ30
mL) from heptane. The solid was triturated with ether
(4ꢂ25 mL), filtered and dried in vacuo to give rise to 6
6
1
1
3
69.57, 162.5, 162.1, 156.1, 143.7, 143.3, 140.5, 130.3,
27.6, 127.0, 126.0, 125.7, 125.2, 120.0, 63.1, 46.4, 43.5,
6.5, 36.4, 32.0; HR-MS (FAB) m/z 661.1941 (661.1934
+
calcd for C H N O , M +H).
3
6
29
4
9
1
as a yellow powder (630 mg, 97%). H NMR (250 MHz,
Preparation of L and D N-2-(N-ꢀꢀ-Fluorenylmethoxycar-
0
DMSO-d ) d 8.72 (s, 4H), 7.93 (br, 4H), 4.35 (t, 4H,
6
13
bonyl-N-"-tert-butoxycarbonyl-lysyl)aminoethyl-N -(2-
J=5.9 Hz), 3.19 (t, 4H, J=5.6 Hz); C NMR (63 MHz,
DMSO-d ) d 162.7, 129.8, 126.0, 125.7, 37.5, 36.9; HR-
carboethyl)-1,4,5,8-naphthalenetetracarboxylic
3) and (4). Compounds 3 and 4 were prepared accord-
ing to the procedure for the preparation of 2. However,
and
diimide
6
(
MS (FAB) m/z 353.1245 (353.1249 calcd for
+
C H N O , M +H).
1
8
17
4
4
Fmoc-l-Lys(Boc)-OC6F5
Fmoc-d-Lys(Boc)-
0
0
OC6F5 (2.63 g, 4.15 mmol) was used instead of Fmoc-
1
N,N -bis-(2,2 -Dimethylamino)ethylamine-1,4,5,8-naph-
thalenetetracarboxylic diimide (7). The synthesis of 7 is
a slight modification of the synthesis reported by
Gly-OC F . H NMR (250 MHz, DMSO-d ) d 8.55 (s,
6
5
6
4H), 8.07 (br, 1H), 7.83 (d, 2H,O J=7.3 Hz ), 7.61(t,
2H, J=7.3 Hz ), 7.30 (m, 1H), 7.38 (d, 2H, J=6 Hz ),
7.28 (br, 2H), 7.07 (br t, 1H) 4.24 (t, 2H, J=6.4 Hz ),
4.11 (t, 2H, J=7.3 Hz), 4.04 (d, 2H, J=6.8 Hz), 3.78
2
4
Tanious. 1,4,5,8-naphthalenetetracarboxylic dianhy-
dride (1g, 3.72 mmol) and N, N-Dimethylethylenedia-
mine (8.1 g, 91.8 mmol) were suspended in THF (15
mL) and heated at reflux for 8 h. The excess volatiles
were removed in vacuo, then the mixture was taken up
in CH Cl (100 mL), washed with a saturated sodium
(
(
br, 1H), 3.57 (br, 2H), 3.39 (m, 2H), 2.85 (br, 2H), 2.63
t, 2H, J=7.3 Hz), 1.49 (m, 2H), 1.38 (s, 9H) 1.29 (br m,
1
3
4
1
1
4
8
H); C NMR (63 MHz, DMSO-d ) d 172.5, 63.8,
6
2
2
62.4, 162.1, 155.8, 155.6, 143.7, 140.5, 131.6, 128.9,
28.3, 126.5, 125.7, 123.9, 121.8, 77.3, 55.9, 53.7, 47.5,
5.4, 34.7, 29.9, 29.7, 28.2, 22.1; HR-MS (FAB) m/z
bicarbonate solution (2ꢂ20 mL) followed by a water (20
mL) wash and concentrated. The resulting orange solid
was suspended in water (5 mL) and acidified to pH 6
with cold HCl. Trituration of the mixture resulted in the
formation of a precipitate that was filtered and dried.
+
32.3218 (832.3194 calcd for C H N O , M +H).
4
5
46
5
11
G
N-2-(N-ꢀꢀ-9-Fluorenylmethoxycarbonyl-N -2,2,5,7,8-
pentamethylchroman-6-sulfonyl-L-argyl)aminoethyl)-N -
Recrystallization of the solid from water afforded pure 7
1
0
2-carboethyl)-1,4,5,8-naphthalenetetracarboxylic diimide
5). Compound 5 was prepared according to the proce-
(1.4 g, 92%). H NMR (250 MHz, CDCl ) d 8.74 (s,
3
(
(
4H), 4.35 (t, 4H, J=6.7 Hz ), 2.68 (t, J=6.8 Hz, 4H),
3
1
2.35 (s, 12H); C NMR (63 MHz, CDCl ) d 162.8,
3
dure for the preparation of 2. However, Fmoc-
Arg(PMC)-OC F (3.43 g, 4.15 mmol) was used instead
130.9, 126.05, 126.4, 56.9, 45.7, 380.6; HR-MS (FAB)
m/z 409.1874 (409.1875 calcd for C H N O ,
M +H).
6
5
22 25
4
4
1
+
of Fmoc-Gly-OC F . H NMR (250 MHz, DMSO-d ) d
6
5
6
8
(
1
(
.71(s, 4H), 8.02 (br, 1H), 7.98 (d, 2H, J=7.3 Hz), 7.61
t, 2H, J=7.5 Hz), 7.37 (d, 2H, J=7.4 Hz), 7.30 (m,
H), 7.27 (br, 2H), 7.07 (br t, 1H), 6.89 (br, 1H) 6.56
br, 1H), 4.18 (br m, 7H), 3.38 (br, 1H), 3.39 (m, 2H),
Combinatorial library
1.5 g of Fmoc-Gly-functionalized NovaSyn TGA resin
(0.17 mmol/g loading) was deprotected with 2ꢂ20 mL
of 50% piperidine/DMF (2ꢂ15 min), washed with
DMF (5ꢂ10 mL), i-PrOH (5ꢂ10 mL), DMF (5ꢂ10
mL), i-PrOH (5ꢂ10 mL) and DMF (5ꢂ10 mL). 674 mg
of 2 (1.2 mmol, 4 equiv), 1.95 mL of PyBOP/DMF
(0.4 M, 4 equiv) and 1.95 mL of NMM/DMF (0.8 M, 8
2
4
.98 (br, 2H), 2.00 (s, 6H), 1.90 (br t, 1H), 1.72 (br t,
1
3
H), 1.53 (m, 1H), 1.35 (br, 4H), 1.21 (s, 9H); C NMR
(
63 MHz, DMSO-d ) d 170.7, 172.2, 162.5, 162.1, 156.1,
6
1
1
1
55.8, 152.3, 143.6, 140.5, 134.5, 134.0, 131.6, 128.8,
28.2, 126.5, 126.0, 125.8, 123.9, 122.7, 121.3, 121.1,
18.7, 117.8, 73.4, 65.7, 55.6, 53.3, 47.6, 45.4, 36.2, 34.1,

C. T. Miller et al. / Bioorg. Med. Chem. 9 (2001) 2015–2024
2023
equiv) were added to the reaction and shaken for 45
min. The resin was washed with i-PrOH (3ꢂ10 mL),
DMSO (3ꢂ10 mL) followed by the wash procedure
described above. The resin was suspended in 50 mL
DCM with stirring and the mixture was divided into 20
reaction vessels using disposable pipettes. Each vessel
represented one of the standard amino acids, whereas
one corresponded to the control. The DCM was filtered,
the resin was washed with DMF and deprotected as
previously described. The appropriate Fmoc-amino
acids (4 equiv); Ala, Arg(Pmc), Asn(Trt), Asp(OBu),
Glu(OBu), Gln(Trt), Gly, His(Trt), Ile, Leu, Lys(Boc),
Met, Phe, Pro, Ser(t-Bu), Thr(t-Bu), Trp(Boc), Tyr(t-
Bu), Val, were coupled to the resin using PyBOP/NMM
Monomer concentrations were determined by UV
measurements at 386 nm (e386=23,300, 1cm) and dimer
concentrations were determined at 386 nm
(e386=27,400, 1cm). Stock concentrations ( 15 , 5, and
1.66 mM) of the compounds to be tested were prepared
using sterile dd H O. The desired compound (20 mL)
2
was added to a sterile 4 inch concentration disk (Difco)
and allowed to dry overnight to afford compounds of
300, 100, and 33 nmol, respectively.
Minimum inhibitory concentrations (MICS)
In addition to the Kirby–Bauer test, the MIC of various
compounds were determined in microtiter plates by
making 2-fold serial dilutions of each compound in
Mueller–Hinton broth, adding the test organism and
(
see above). The resin was washed with DMF, i-PrOH,
DMF, i-PrOH, DMF and the resin from the vessels was
suspended in DCM (50 mL) with stirring and was
evenly divided into the 19 reaction vessels. This process
was repeated until all of the units had been coupled,
however, the resin was not mixed and pooled after the
addition of the last amino acid. Instead, 674 mg of 2
was dissolved in 1.95 mL of DMF, 260 mL of the solu-
tion was added to each vessel and 260 mmL of PyBOP/
DMF (0.4 M, 4 equiv) and 260 mL m of NMM/DMF
1
6,17
incubating the plate for 16–20 h.
E. coli JM109 and
B. subtilis 1012M15 were grown in Muller–Hinton
broth for 4 h and diluted to an OD600 of 0.1. Microtiter
plates were covered with a sterile lid and incubated at
ꢃ
37 C for 18 to 20 h and the MICs were determined. The
wells were examined for bacterial growth at OD600 and
the MIC was defined as the lowest concentration of the
agent that prevented growth of the test organism. The
activity in mammalian cells of selected compounds were
determined through 2-fold serial dilutions starting at
(
0.8 M, 8 equiv) were added and shaken for 45 min. The
resin was washed with i-PrOH (3ꢂ10 mL), DMSO
1
8
(
3ꢂ10 mL), followed by the wash procedure described
100 mM concentrations using the MTT assay.
above. The resin in each vessel was then suspended in
DCM (5 mL) and transferred to twenty 10 mL glass
vials with a screw cap. The resin was dried in vacuo and
the acid labile side chains and mixtures were cleaved
References and Notes
from the resin with 5 mL TFA/H O/TIS (95:2.5:2.5, 18
2
1. Oren, Z.; Hong, J.; Shai, Y. J. Biol. Chem. 1997, 272, 1463.
2. Setti, E. L.; Micetich, R. G. Curr. Med. Chem. 1998, 5, 101.
h). The resin was filtered off and washed several times
with TFA (0.5 mL) and DCM (0.5 mL). Each solution
was transferred to a 50 mL plastic tube and triturated
with cold ether (20 mL). The products precipitated and
the solutions were centrifuged for 15 min, the ether was
decanted and cold ether (20 mL) was added. The pro-
cess was repeated five times to remove residual TFA
and the resulting solids were dried in vacuo for 1h,
3
2
. Hancock, R. E.; Falla, J. T. In Biotechnology of Antibiotics,
nd ed.; Strohl, W. R., Ed.; Marcel Dekker: New York, 1978;
pp 471–494.
4. Tossi, A.; Sandri, L.; Giangasporo, A. Biopolymers 2000,
55, 4. See also the other comprehensive reviews in this issue.
5. Hancock, L.; Lehrer, R. Trends in Biotech. 1998, 16, 82.
6. Vaara, M.; Vaara, T. Anti. Agents Chemother. 1983, 24, 107.
7. He, K.; Ludtke, J. S.; Huang, H. W. Biochemistry 1995, 34,
dissolved in 10 mL of H O and lyophilized. All mixtures
2
15614.
were analyzed on RP HPLC (C ) monitored by UV
1
8. Nikaido, H.; Vaara, M. Micro. Rev. 1985, 49, 1 .
9. Shima, S.; Matsuoka, H.; Iwamoto, T.; Sakai, H. J. Anti-
biot. 1984, 37, 1149.
8
(l=386 nm) over a 20 min gradient. The gradient con-
sisted of 90% TFA/H O (0.07:99.93) to 100% TFA/
CH CN (0.07:99.93).
2
1
1
1
0. Coste, J.; Le-Nguyen, D.; Castro, B. Tetrahedron Lett.
990, 31, 205.
1. Lee, S.; Mihara, H.; Aoyagi, H.; Kato, T.; Izumiya, N.;
3
Parallel synthesis
Yamasaki, N. Biochim. Biophys. Acta 1986, 862, 211.
2. Cheng, C. C. In Cancer Chemotherapuetic Agents; Foye,
W. O., Ed.; American Chemical Society: Washington, DC,
995; pp 239–260.
13. Yeng, S.; Gabbay, E. J.; Wilson, D. W. Biochemistry 1982,
21, 2070.
1
Compounds 8–39 and 44–63 were synthesized using the
procedure described above. However, the compounds
were not mixed after couplings and in some cases,
Fmoc-Lys(Boc) or Fmoc-Arg(Pmc) resin was used
instead of Fmoc-Gly resin.
1
14. Lokey, R. S.; Kwok, Y.; Guelev, V.; Pursell, C. J.; Hurley,
L. H.; Iverson, B. L. J. Am. Chem. Soc. 1997, 119, 7202.
1
5. Alper, P. B.; Hendrix, M.; Sears, P.; Wong, C. J. Am.
Chem. Soc. 1998, 120, 1965.
6. Phillips, I.; Williams, J. D.; Wise, R. In Laboratory Meth-
Biological screening: Kirby–Bauer assay
1
The following studies were conducted in the absence of
light due to the sensitivity of the NDI units. The com-
pounds synthesized via combinatorial and parallel
synthesis were tested for biological activity against B.
subtilis and E. coli using the Kirby–Bauer disk diffusion
ods in Antimicrobial Chemotherapy; Garrod, L., Ed.; Churchill
Livingston: Edinburgh, 1978; pp 3–30.
1
7. Waterworth, P. M. In Laboratory Methods in Anti-
microbial Chemotherapy; Garrod, L., Ed.; Churchill Living-
ston: Endinburgh, 1978; pp 31–40.
18. Sami, S. M.; Dorr, R. T.; Aniko, M. S.; Alberts, D. S.;
Bhashyam, S. I.; Remers, W. A. J. Med. Chem. 1995, 39, 1609.
1
5ꢀ17
The tests were performed using the refer-
method.
ence strains E. coli JM109 and B. subtilis 1012M15.

2024
C. T. Miller et al. / Bioorg. Med. Chem. 9 (2001) 2015–2024
19. Schutz, V. S.; Kurz, J.; Plumpe, H.; Bock, M. O.; Otten,
H. Arzn-Forsch. 1971, 21, 739.
22. Sami, S. M.; Dorr, R. T.; Aniko, M. S.; Alberts, D. S.;
Remers, W. A. J. Med. Chem. 1995, 38, 983.
23. Sami, S. M.; Dorr, R. T.; Aniko, M. S.; Alberts, D. S.;
Remers, W. A. J. Med. Chem. 1996, 39, 4978.
24. Tanious, F. A.; Yen, S.; Wilson, W. D. Biochemistry 1991,
30, 1813.
2
1
2
0. Paull, K. D.; Nasr, M.; Narayanan, V. L. Arzn-Forsch
984, 34, 1243.
1. Sami, S. M.; Dorr, R. T.; Alberts, D. S.; Remers, W. A. J.
Med. Chem. 1993, 36, 765.