Synthesis of the all-l cyclopentapeptides versicoloritides A, B, and C

Article abstract of DOI:10.1055/s-0032-1316994

Herein we describe the first synthesis of the all-l naturally occurring cyclopentapeptides versicoloritides A, B, and C. We found that the extent of oligomerization and epimerization upon head-to-tail cyclization of the linear pentapeptide precursors could be reduced to acceptable levels by modifying the reaction conditions. Georg Thieme Verlag Stuttgart ? New York.

Full text of DOI:10.1055/s-0032-1316994

2284▌
LETTER
lSetyternthesis of the All-L Cyclopentapeptides Versicoloritides A, B, and C
All-L Cyclopentapeptides Versicoloritides A, B, and C
Harveen Kaur,a,b Amanda M. Heapy,b,c Margaret A. Brimble*a,b,c
a
School of Biological Sciences, The University of Auckland, 3A Symonds St, Auckland, 1142, New Zealand
b
Maurice Wilkins Centre for Molecular Biodiscovery, The University of Auckland, Private Bag 92019, Auckland, 1142, New Zealand
c
School of Chemical Sciences, The University of Auckland, 23 Symonds St, Auckland, 1142, New Zealand
E-mail: m.brimble@auckland.ac.nz
Received: 08.06.2012; Accepted: 01.07.2012
lar, the intramolecular cyclization of linear pentapeptides
is associated with oligomerization and C-terminal epimer-
ization.3a,c,4a The extent of oligomerization and epimeriza-
tion during the cyclization of pentapeptides is primarily
Abstract: Herein we describe the first synthesis of the all-L natural-
ly occurring cyclopentapeptides versicoloritides A, B, and C. We
found that the extent of oligomerization and epimerization upon
head-to-tail cyclization of the linear pentapeptide precursors could
be reduced to acceptable levels by modifying the reaction condi- governed by the sequence of the linear precursor and is
tions.
further influenced by the choice of coupling reagents,
base, solvent, the presence of metal ions, racemization
suppressants, and reaction temperature.3a,c,7 Retro-
synthetic disconnection of the versicoloritides, however,
affords linear pentapeptides containing all-L amino acids.
Linear peptides containing all-L or all-D amino acids pre-
fer to adopt extended conformations in order to minimize
allylic strain.4a This preferred extended conformation,
however, is counteractive for head-to-tail cyclization, as
the success of macrolactamization is promoted by the
ability of a linear precursor to adopt an intermediate con-
formation similar to that of the corresponding cyclic pep-
tide.4a,8
Key words: cyclic peptides, versicoloritides, natural products,
cyclization, total synthesis
In an on-going search for bioactive natural products from
surface-associated marine microorganisms, three new
cyclopentapeptides, versicoloritides A (1), B (2), and
C (3) were recently isolated from the coral-associated fun-
gus Aspergillus versicolor (Figure 1).1 The versicoloriti-
des differ from each other by a single residue and were
found to consist of only L-amino acids with the general se-
quence cyclo-(FPFPX), where X is L-alanine, glycine, or
L-serine in versicoloritides A (1), B (2), and C (3), respec-
tively. Cyclopentapeptides often contain D-amino acids,
N-methylated amino acids, and/or unnatural amino acids
in addition to L-amino acids.2 Natural product cyclopenta-
peptides consisting of only L-amino acids are rarely isolat-
ed, and only a few other naturally occurring all-L
cyclopentapeptides have been found to date.3 Our on-
going interest in the synthesis of natural product cyclic
peptides, the rare occurrence of all-L cyclopentapeptides
and the reported challenges associated with the cyclisa-
tion of all-L linear pentapeptides prompted our interest in
the synthesis of the versicoloritides.3a,4 We herein report
the first synthesis of the three known members of the ver-
sicoloritide family, versicoloritides A (1), B (2), and C
(3).
O
versicoloritide A (1): R = Me
versicoloritide B (2): R = H
versicoloritide C (3): R = CH2OH
N
HN
O
O
NH
R
HN
N
O
O
Figure 1 The versicoloritide family of cyclic peptides, versicolorit-
ides A (1), B (2), and C (3)
Fortunately, the pentapeptide precursors to versicoloriti-
des A (1), B (2), and C (3) contain proline residues and a
glycine residue in the case of versicoloritide B. These res-
idues are known to induce and stabilize turn structures
thereby aligning the N- and C-termini in closer spatial
proximity.4a,9 The linear precursors 4, 5, and 6 to versicol-
oritides A (1), B (2), and C (3), respectively, were there-
fore chosen with the turn-inducing proline residues
located in the middle of the linear sequence rather than po-
sitioned at the N- and C-termini (Scheme 1).4a To mini-
mize steric hindrance during cyclization the smaller
variable residues (alanine, glycine, and serine) were cho-
sen as the C-terminal residues of the pentapeptide precur-
sors 4, 5, and 6. Moreover, the presence of a glycine
residue at the C-terminus of 5 also prevents any opportu-
There are many methods available for the synthesis of
head-to-tail cyclic peptides, of which the final macro-
lactamization reaction is often carried out in solution
phase under high dilution conditions or by using on-resin
cyclization techniques such as side-chain anchoring of a
trifunctional amino acid or the use of ‘safety-catch’ link-
ers.4a,5 Irrespective of the cyclization technique employed,
the synthesis of head-to-tail cyclic peptides containing
less than seven residues is known to be challenging due to
the small-to-medium ring sizes involved.4a,b,5,6 In particu-
SYNLETT 2012, 23, 2284–2288
Advanced online publication: 14.08.2012
0
9
3
6
-
5
2
1
4
1
4
3
7
-
2
0
9
6
DOI: 10.1055/s-0032-1316994; Art ID: ST-2012-D0486-L
© Georg Thieme Verlag Stuttgart · New York

LETTER
All-L Cyclopentapeptides Versicoloritides A, B, and C
2285
nity for racemization during the preparation of versicolo- peptides. The TFA cocktail–peptide mixture was there-
ritide B (2).4a
fore diluted with a mixture of H2O and MeCN, lyophi-
lized, and the resultant oil purified by semipreparative
RP-HPLC to afford the pure linear peptides 4, 5, and 6 as
colorless solids.
The linear precursors 4, 5, and 6 were assembled manual-
ly by Fmoc-SPPS (see Supporting Information). Briefly,
the variable C-terminal residues were attached via an
acid-labile Wang-like linker to in-house prepared amino- With linear peptides 4, 5, and 6 in hand, our attention
methyl polystyrene resin.10
turned to the critical macrolactamization step. In order to
initially evaluate the conditions that affect oligomeriza-
tion, we chose to cyclize linear precursor 5 as the achiral
C-terminal glycine residue offers no opportunity for race-
mization. The cyclization was performed by slowly add-
ing a mixture of linear precursor 5 and HATU (3 equiv) in
DMF dropwise at the rate of 0.5 mL/h to a stirring solution
of DIPEA (5 equiv) in DMF with a final concentration of
0.9 mg/mL (Table 1, entry 1). Analysis of the reaction
mixture by LC–MS after complete addition showed that
the cyclization had proceeded quickly. The linear precur-
sor 5 was undetected, and the presence of both the desired
cyclopentapeptide versicoloritide B (2) and the undesired
corresponding cyclodecapeptide were detected in equal
amounts in a 1:1 ratio as determined by RP-HPLC. In an
attempt to reduce the amount of cyclodimerization, the
above reaction conditions were modified to use either
CH2Cl2 as the reaction solvent at the same concentration
of 0.9 mg/mL or DMF at a higher dilution with a final con-
centration of 0.33 mg/mL, neither of which significantly
altered the cyclic monomer/dimer ratio (Table 1, entries 2
and 3).11 Reversing the order of addition of reagents (i.e.,
adding DIPEA in DMF dropwise to a stirred mixture of 5
and HATU in DMF) resulted in an increase in the amount
of the undesired cyclic dimer.12 We next evaluated the
effect of other activating reagents on the cyclic mono-
mer/dimer ratio. In contrast to HATU, the activating re-
agents BOP and HBTU were both found to improve the
1:1 cyclic monomer/dimer ratio to an acceptable 9:1 ratio
(Table 1, entries 5–8).
O
N
HN
O
O
NH
R
HN
N
O
O
1: R = Me
2: R = H
3: R = CH2OH
NH2-L-Phe-L-Pro-L-Phe-L-Pro-X-COOH
4: X = L-Ala
5: X = Gly
6: X = L-Ser
Scheme 1 Disconnection sites for the synthesis of versicoloritides
A (1), B (2), and C (3)
Successful couplings were followed by deprotection of
the Fmoc group with 20% piperidine in DMF and each
subsequent amino acid (4 equiv) was then attached to the
growing resin-bound peptide using a mixture of HBTU
(3.9 equiv) and DIPEA (8 equiv) in DMF for 45 minutes.
The fully assembled linear resin-bound peptides were
then cleaved from the solid support using a mixture of
TFA, TIS, H2O, and 2,2′-(ethylenedioxy)diethanethiol.
The hydrophobic nature of the linear peptides, however,
rendered the crude peptides miscible in cold diethyl ether
that is normally used during workup to precipitate crude
Table 1 Cyclization of Linear Precursor 5 to 2
Entry
Reagents and conditionsa
(2):Dimerb
1
2
3
4
5
6
7
8
HATU, DIPEA, DMF, 0.9 mg/mL
HATU, DIPEA, CH2Cl2, 0.9 mg/mL
HATU, DIPEA, DMF, 0.33 mg/mL
HATU, DIPEA, DMF, 0.45 mg/mL
PyBOP, DIPEA, DMF, 0.33 mg/mL
BOP, DIPEA, DMF, 0.33 mg/mL
1:1
1:1
1:1
1:3c
major:minord
9:1
9:1
9:1
HBTU, DIPEA, DMF, 0.33 mg/mL
HBTU, DIPEA, CH2Cl2–DMF (4:1), 0.33 mg/mL
a Conditions: 3 equiv of coupling reagent and 5 equiv of base were used.
b The yield ratios were estimated by analytical RP-HPLC peak area.
c Reversed reagent addition order.
d Ratio cannot be determined as phosphoramide generated by PyBOP co-eluted with cyclic monomer.
© Georg Thieme Verlag Stuttgart · New York
Synlett 2012, 23, 2284–2288

2286
H. Kaur et al.
LETTER
Table 2 Cyclization of Linear Precursor 4 to 1
Entry
Reagents and conditionsa,b
Epimer:(1):Dimerc
1
2
3
4
5
6
7
PyBOP, DIPEA, DMF
major:minord
1:8:1
BOP, DIPEA, DMF
BOP, collidine, DMF
3:6:1
BOP, NMM, DMF
2:6:2
HBTU, DIPEA, DMF
1:6:3
HBTU, HOBt, DIPEA, DMF
HBTU, HOBt, DIPEA, CH2Cl2–DMF (19:1)
2:7:1
0:7:3
a Conditions: 3 equiv of coupling reagent and 5 equiv of base were used.
b The final concentration of linear peptide was 0.33 mg/mL.
c The yield ratios were estimated by analytical RP-HPLC peak area.
d Ratio cannot be determined as phosphoramide generated by PyBOP co-eluted with cyclic monomer.
Satisfied with the use of the above conditions for the syn- monomer 1 and undesired cyclic dimer, analytical LC–
thesis of versicoloritide B (2),13 the predominant cyclic MS and RP-HPLC spectra revealed the formation of an
monomer peak was isolated by concentration of the reac- additional compound adjacent to versicoloritide A (1)
tion mixture under reduced pressure, dilution of the re- which exhibited an identical mass to the cyclopentapep-
maining residue with 0.1% TFA–H2O (v/v), and tide (Figure 2, trace A).
purification by semipreparative RP-HPLC. The optical
Isolation of both cyclic monomer peaks by RP-HPLC fol-
rotation value and spectroscopic data of synthetic versi-
lowed by comparison of the 1H NMR data for the
coloritide B (2) were in agreement with that of naturally
synthetic cyclopentapeptide mixture with the reported
values for versicoloritide A confirmed that the predomi-
occurring versicoloritide B (2, Table 2 in Supporting In-
formation). In addition, the cis-conformation of both pro-
nant cyclopentapeptide peak was indeed versicoloritide A
line residues of the natural product was retained in
(1, Figure 3).1
synthetic versicoloritide B (2).1,14
With versicoloritide B in hand we next attempted the syn-
thesis of versicoloritide A (1) by applying the above reac-
tion conditions to the cyclization of linear precursor 4. In
a manner consistent with the cyclization of 5, the cycliza-
tion of precursor 4 using the activating reagents BOP and
HBTU also resulted in a larger proportion of the desired
cyclopentapeptide relative to the undesired cyclodeca-
peptide as confirmed by LC–MS and RP-HPLC (Table
2).15 In addition to the formation of the desired cyclic
Figure 3 Methyl region of the 1H NMR spectrum of a mixture of ver-
sicoloritide A (1) and epi-versicoloritide A
To determine whether the additional cyclic monomer
peak was due to either of the proline residues of the cyclo-
pentapeptide adopting a trans-configuration16 or whether
it was an epimer of versicoloritide A resulting from C-ter-
minal epimerization during macrolactamization,17 we de-
cided to synthesize the possible epimer, namely cyclo-(-L-
Figure 2 RP-HPLC analysis of the crude reaction mixture after com-
plete addition of reagents of the cyclization of the linear epimer (trace
B) and the linear precursor 4 (trace A)
Phe-L-Pro-L-Phe-L-Pro-D-Ala). The linear epimer precur-
sor NH2-L-Phe-L-Pro-L-Phe-L-Pro-D-Ala-COOH was
Synlett 2012, 23, 2284–2288
© Georg Thieme Verlag Stuttgart · New York

LETTER
All-L Cyclopentapeptides Versicoloritides A, B, and C
2287
synthesized in a similar manner to the other precursors by in agreement with naturally occurring versicoloritide C
manual Fmoc-SPPS and purified by RP-HPLC. The epi- (3) (Table 3 in Supporting Information).1
mer precursor was then subjected to identical cyclization
conditions as precursor 4, by treatment with a mixture
HBTU (3 equiv) and DIPEA (5 equiv) in DMF. RP-HPLC
analysis of the reaction mixture after addition of the re-
In conclusion, we herein report the first synthesis of the
naturally occurring all-L cyclopentapeptides versicoloriti-
des A (1), B (2), and C (3). We observed for the three sim-
ilar linear precursor sequences FPFPX significant
agents, as shown in Figure 2, established that the retention
differences in the degree of dimerization and epimeriza-
time of the synthetic cyclic epimer (Figure 2, trace B)
tion observed despite using the same reaction conditions.
matched exactly with the additional peak observed upon
Importantly, the spectroscopic data of the synthetic versi-
cyclization of the all-L linear precursor 4 (Figure 2,
coloritides (1), (2), and (3) confirmed the structure of the
trace A), thereby confirming that epimerization of L-Ala
natural products elucidated by Zhuang et al.1 The peptides
to D-Ala took place upon macrolactamization.
synthesized herein are currently undergoing further bio-
Having established that the additional peak was in fact logical evaluation.
due to C-terminal racemization that took place during the
macrocyclization step, we attempted to reduce the unde-
sired base-catalyzed epimerization by substituting DIPEA
Acknowledgment
We thank the Maurice Wilkins Centre for Molecular Discovery for
financial support.
with less basic tertiary amines such as 2,4,6-collidine and
N-methylmorpholine.7a,18 Substitution of DIPEA with
both 2,4,6-collidine and N-methylmorpholine, however,
increased the extent of epimerization (Table 2, entries 3
and 4). Interestingly, similar observations have been re-
ported by others whereby substitution of DIPEA with the
sterically hindered base collidine was associated with
higher degrees of epimerization, as observed during the
synthesis of a highly N-methylated cyclic peptide NMe-
IB-0121212 and the cyclization of other all-L penta-
peptides.4b,12 A cyclization attempt of precursor 4 using
the racemization suppressant HOBt also failed to improve
the epi-versicoloritide A/versicoloritide A ratio (Table 2,
entry 6). In a further attempt to reduce the extent of
epimerization, we chose to modify the reaction solvent
based on studies that showed that the extent of racemiza-
tion could be altered by the choice of solvent.7a,19 Using
95% CH2Cl2 as the reaction solvent instead of DMF was
found to have a significant effect on C-terminal epimer-
ization, in that the epi-versicoloritide A/versicoloritide A
ratio reflected negligible formation of the undesired epi-
mer (Table 2, entry 7).
Supporting Information for this article is available online at
http://www.thieme-connect.com/ejournals/toc/synlett.SnuIpofoiprmntgirStanoIuifpog
r
t
iornat
References and Notes
(1) Zhuang, Y. B.; Teng, X. C.; Wang, Y.; Liu, P. P.; Wang, H.;
Li, J.; Li, G. Q.; Zhu, W. M. Tetrahedron 2011, 67, 7085.
(2) (a) Houston, D. R.; Shiomi, K.; Arai, N.; Omura, S.; Peter,
M. G.; Turberg, A.; Synstad, B.; Eijsink, V. G. H.; van
Aalten, D. M. F. Proc. Natl. Acad. Sci. U.S.A. 2002, 99,
9127. (b) Wu, W.; Dai, H. Q.; Bao, L.; Ren, B. A.; Lu, J. C.;
Luo, Y. M.; Guo, L. D.; Zhang, L. X.; Liu, H. W. J. Nat.
Prod. 2011, 74, 1303. (c) Fremlin, L. J.; Piggott, A. M.;
Lacey, E.; Capon, R. J. J. Nat. Prod. 2009, 72, 666.
(3) (a) Schmidt, U.; Langner, J. J. Pept. Res. 1997, 49, 67.
(b) Li, F.; Zhang, F. M.; Bin Yang, Y.; Yang, X. Q.; Cao, Q.
E.; Ding, Z. T. Chin. Chem. Lett. 2008, 19, 193. (c) Tang, Y.
C.; Xie, H. B.; Tian, G. L.; Ye, Y. H. J. Pept. Res. 2002, 60,
95.
(4) (a) White, C. J.; Yudin, A. K. Nat. Chem. 2011, 3, 509.
(b) Ehrlich, A.; Heyne, H. U.; Winter, R.; Beyermann, M.;
Haber, H.; Carpino, L. A.; Bienert, M. J. Org. Chem. 1996,
61, 8831. (c) Kaur, H.; Heapy, A. M.; Brimble, M. A. Synlett
2012, 23, 275.
(5) Davies, J. S. J. Pept. Sci. 2003, 9, 471.
(6) Malesevic, M.; Strijowski, U.; Bachle, D.; Sewald, N. J.
Biotechnol. 2004, 112, 73.
Satisfied with the optimized reaction conditions devel-
oped herein to reduce undesired epimerization, we isolat-
ed synthetic versicoloritide A (1) by concentrating the
reaction mixture under reduced pressure, diluting the re-
maining residue with 0.1% TFA–H2O (v/v) and purifying
by semipreparative RP-HPLC. The optical rotation value
and spectroscopic data of synthetic versicoloritide A (1)
were in agreement with that of the naturally occurring ver-
sicoloritide A (1, Table 1 in Supporting Information).1
(7) (a) Carpino, L. A.; Elfaham, A.; Albericio, F. Tetrahedron
Lett. 1994, 35, 2279. (b) Ye, Y. H.; Gao, X. M.; Liu, M.;
Tang, Y. C.; Tian, G. L. Lett. Pept. Sci. 2003, 10, 571.
(8) Jiang, S.; Li, Z.; Ding, K.; Roller, P. P. Curr. Org. Chem.
2008, 12, 1502.
(9) Rizo, J.; Gierasch, L. M. Annu. Rev. Biochem. 1992, 61, 387.
(10) Harris, P. W. R.; Brimble, M. A. Synthesis 2009, 3460.
(11) (a) Blanchette, J. P.; Ferland, P.; Voyer, N. Tetrahedron
Lett. 2007, 48, 4929. (b) Li, P.; Roller, P. P.; Xu, J. C. Curr.
Org. Chem. 2002, 6, 411.
(12) Marcucci, E.; Tulla-Puche, J.; Albericio, F. Org. Lett. 2012,
14, 612.
(13) Versicoloritide B (2)
Having successfully synthesized versicoloritide A (1) and
B (2), we next attempted the synthesis of versicoloritide C
(3). Interestingly, in contrast to the cyclization of precur-
sors 4 and 5, the cyclization of precursor 6 proceeded
smoothly to form versicoloritide C (3) as the predominant
product with negligible dimerization and undetectable
epimerization in all the conditions that were evaluated, us-
ing either DMF or 90% CH2Cl2 as the solvent and using
either HBTU or BOP as the coupling reagent.20 Like ver-
sicoloritide A (1) and B (2), the optical rotation value and
spectroscopic data of synthetic versicoloritide C (3) were
To a solution of DIPEA (28 μL, 161 μmol) in CH2Cl2
(36 mL) was added 5 (18 mg, 32 μmol) and HBTU (36 mg,
95 μmol) in CH2Cl2–DMF (4:1, 15 mL) at 0.5 mL/h. The
reaction mixture was concentrated under reduced pressure,
© Georg Thieme Verlag Stuttgart · New York
Synlett 2012, 23, 2284–2288

2288
H. Kaur et al.
LETTER
diluted with 0.1% TFA–H2O (v/v, 18 mL), and purified by
RP-HPLC to afford 2 as a colorless solid (10 mg, 57%).
HRMS (EI): m/z [M + Na]+ calcd for C30H35N5NaO5:
568.2536; found: 568.2526. IR: 3277, 2956, 1626, 1542,
1447, 1344, 1162, 746, 702 cm–1. [α]D20 –46 (c 0.1, MeOH)
[lit. –43.6 (c 0.1, MeOH)].1 See the Supporting Information
for complete procedures and analytical data.
(16) (a) Francart, C.; Wieruszeski, J. M.; Tartar, A.; Lippens, G.
J. Am. Chem. Soc. 1996, 118, 7019. (b) Gesquiere, J. C.;
Diesis, E.; Cung, M. T.; Tartar, A. J. Chromatogr. 1989,
478, 121.
(17) Yoshiya, T.; Kawashima, H.; Hasegawa, Y.; Okamoto, K.;
Kimura, T.; Sohma, Y.; Kiso, Y. J. Pept. Sci. 2010, 16, 437.
(18) (a) Han, S. Y.; Kim, Y. A. Tetrahedron 2004, 60, 2447.
(b) Carpino, L. A.; Elfaham, A. J. Org. Chem. 1994, 59, 695.
(19) Ye, Y. H.; Li, H. T.; Jiang, X. H. Biopolymers 2005, 80, 172.
(20) Versicoloritide C (3)
(14) (a) Wele, A.; Mayer, C.; Derinigny, Q.; Zhang, Y.; Blond,
A.; Bodo, B. Tetrahedron 2008, 64, 154. (b) Schmidt, G.;
Grube, A.; Kock, M. Eur. J. Org. Chem. 2007, 4103.
(15) Versicoloritide A (1)
To a solution of DIPEA (60 μL, 344 μmol) in CH2Cl2 (80
mL) was added 6 (40 mg, 67 μmol) and HBTU (77 mg, 200
μmol) in CH2Cl2–DMF (5:1, 22 mL) at 0.5 mL/h. After
complete addition of the reagents, the reaction mixture was
concentrated under reduced pressure, diluted with 0.1%
TFA–H2O (v/v, 18 mL), and purified by RP-HPLC to afford
3 as a colorless solid (17 mg, 44%). HRMS (EI): m/z [M +
Na]+ calcd for C31H37N5NaO6: 598.2642; found: 598.2645.
IR: 3307, 2956, 1632, 1523, 1443, 1345, 1165, 746, 700 cm–1.
[α]D20 –135.5 (c 0.2, MeOH) [lit. –118 (c 0.2, MeOH)].1 See
Supporting Information for complete procedures and
analytical data.
To a solution of DIPEA (31 μL, 178 μmol) in CH2Cl2 (44
mL) was added 4 (22 mg, 38 μmol), HBTU (44 mg, 116
μmol), and HOBt (15 mg, 111 μmol) in CH2Cl2–DMF
(4.7:1, 17 mL) at 0.5 mL/h. The reaction mixture was
concentrated under reduced pressure, diluted with 0.1%
TFA–H2O (v/v, 10 mL) and purified by RP-HPLC to 1 as a
colorless solid (5.5 mg, 26%). HRMS (EI): m/z [M + Na]+
calcd for C31H37N5NaO5: 582.2692; found: 582.2675. IR:
3280, 2937, 1633, 1519, 1448, 1346, 1319, 1161, 745, 701
cm–1. [α]D20 –122 (c 0.1, MeOH) [lit. –90.7 (c 1.7, MeOH)].1
See the Supporting Information for complete procedures and
analytical data.
Synlett 2012, 23, 2284–2288
© Georg Thieme Verlag Stuttgart · New York

Copyright of Synlett is the property of Georg Thieme Verlag Stuttgart and its content may not be copied or
emailed to multiple sites or posted to a listserv without the copyright holder's express written permission.
However, users may print, download, or email articles for individual use.