Parallel synthesis and anti-inflammatory activity of cyclic peptides cyclosquamosin D and Met-cherimolacyclopeptide B and their analogs

Article abstract of DOI:10.1016/j.bmcl.2010.08.033

We report the parallel synthesis of two natural cyclopeptides, isolated from the seeds of Annona squamosa, cyclosquamosin D (A1), and Met-cherimolacyclopeptide B (B) and their analogs. All of the compounds were screened for anti-inflammatory activity by e

Full text of DOI:10.1016/j.bmcl.2010.08.033

Bioorganic & Medicinal Chemistry Letters 20 (2010) 5653–5657
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Parallel synthesis and anti-inflammatory activity of cyclic peptides
cyclosquamosin D and Met-cherimolacyclopeptide B and their analogs
Afef Dellai a, Igor Maricic b, Vipin Kumar b, Sergey Arutyunyan c, Abderrahman Bouraoui a, Adel Nefzi c,
*
a Faculté de Pharmacie, Laboratoire de Pharmacologie Marine, 5000 Monastir, Tunisia
b Laboratory of Autoimmunity, Torrey Pines Institute for Molecular Studies, San Diego, CA 92121, USA
c Chemistry Department, Torrey Pines Institute for Molecular Studies, Port Saint Lucie, FL 34987, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
We report the parallel synthesis of two natural cyclopeptides, isolated from the seeds of Annona squa-
mosa, cyclosquamosin D (A1), and Met-cherimolacyclopeptide B (B) and their analogs. All of the com-
pounds were screened for anti-inflammatory activity by evaluating their inhibitory effects on the
production of pro-inflammatory cytokines using the lipopolysaccharide stimulated macrophage
J774A.1 cell line. Compounds having significant anti-inflammatory activity in suppressing the secretion
Received 6 May 2010
Revised 4 August 2010
Accepted 6 August 2010
Available online 11 August 2010
of IL-6 and TNF-
the natural products.
a have been identified, some of which exhibit activity superior to that observed with
Keywords:
Parallel synthesis
Solid phase synthesis
Natural cyclopeptides
Anti-inflammatory activity
Ó 2010 Elsevier Ltd. All rights reserved.
Cyclic peptides constitute an important group of natural prod-
uct compounds.1 Drugs such as vancomycin, cyclosporin, dapto-
mycin, gramicidin S, and bleomycin belong to this group and
their successes on the market demonstrate that these types of mol-
ecules can be antibiotics and antifungal compounds, as well as
immunosuppressive and even anti-cancer drugs.2 Naturally occur-
ring cyclic peptides continue to hold the attention of synthetic
chemists and biologists alike for their intriguing chemical struc-
tures and potent biological activities.3 Cyclic peptides obtained
from latex, seeds, and roots of plants possess a wide array of bioac-
tivities including cytotoxic activity,4,5 immunosuppressive activ-
ity,6 antimalarial activity,7 vasorelaxant activity,8 cyclooxygenase,
ACE and tyrosinase inhibitory activity,9,10 and anti-inflammatory
activity.11 Recently, cyclic peptides cyclosquamosin D (A1) and
cherimolacyclopeptide B have been isolated from seeds of Annona
squamosa (Fig. 1). The cyclic peptide cyclosquamosin D (A1) was
reported to have an inhibitory effect on the production of pro-
inflammatory cytokines within lipopolysaccharide and Pam3Cys-
stimulated J774A.1 macrophages.11 However, the cyclic peptide
cherimolacyclopeptide B was not active in these assays. In this Let-
ter, we report our results on the parallel synthesis of the cyclic pep-
tides cyclosquamosin D (A1) and Met-cherimolacyclopeptide B (B)
and their analogs. All of the synthetically prepared peptides were
tested for anti-inflammatory activity.
Using the T-bag technology,15 the parallel solid phase peptide
synthesis12 of cyclosquamosin D (A1) and its 14 analogs (A2–
A15) was performed starting from Wang-resin bound valine,13
using Fmoc strategy.14 As outlined in Scheme 1, all syntheses were
performed using standard coupling deprotection cycles in the pres-
ence of DIC, HOBt as activating agents and 20% piperidine in DMF
for Fmoc deprotection. Following deprotection of the N-terminal
protected amino acid, the solid support and all protecting groups
were cleaved using HF.
Following extraction, and lyophilization, the identity of the lin-
ear peptides were confirmed by LC–MS. The crude peptides were
treated overnight in DMF with PyBOP and HOBt in the presence
O
NH2
HO
H
N
N
HO
HN
NH
NH
O
O
O
O
OH
H
N
HN
OH
O
O
O
NH
N
O
O
O
O
N
O
HN
O
NH
N
O
H
O
HN
O
NH
Cyclosquamosin D (A1)
Figure 1. Natural cyclic peptides: cyclosquamosin D and cherimolacyclopeptide B.
S
N
O
H
Cherimolacyclopeptide B
* Corresponding author. Tel.: +1 772 345 4739; fax: +1 772 345 3649.
E-mail address: adeln@tpims.org (A. Nefzi).
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.08.033

5654
A. Dellai et al. / Bioorg. Med. Chem. Lett. 20 (2010) 5653–5657
R4
O
R2
O
O
R6
O
H
N
H
N
H
N
SPPS
O
O
H2N
H2N
N
H
N
H
N
H
N
H
O
R3
O
R1
Fmoc strategy
R7
O
O
O
OH
HF
O
O
R4
R2
R6
O
O
H
H
H
OH
N
N
H2N
N
N
N
N
N
H
H
H
H
R7
O
O
O
R3
O
R1
OH
PyBOP, HOBt, DIEA
R2
H
R3
HN
O
R1
N
HN
O
R4
HN
O
NH
O
O
O
HO
O
O
NH
N
H
NH
R7
R6
Scheme 1. Parallel synthesis of cyclosquamosin D (A1) and its analogs.
of DIEA to generate the corresponding cyclic peptides in good
yields.19 All of the desired compounds were purified using re-
versed-phase high-performance liquid chromatography (RP HPLC).
The purity of these compounds is higher than 90% for all the com-
pounds, based upon analytical traces at k = 214 nm and 254 nm.
In addition to the natural peptide cyclosquamosin D (A1), 14
analogs were synthesized. We performed an alanine screen to
investigate the effect and contribution of the amino acids Xaa1,
Xaa2, Xaa4, Xaa6, and Xaa7 on the anti-inflammatory activity (Ta-
ble 1). Thus, five cyclic peptide analogs (A9, A11–A13, and A15)
were synthesized. We also performed the synthesis of cyclic pep-
tide analogs where we modified or permuted the order of some
amino acids. In analog A4, the tyrosine at position R4 and the serine
at position R6 were reversed, while in analog A8, the glycine at R1
was replaced with tyrosine and the order of amino acids at R2 and
R3 was reversed. In analog A14, the proline and glycine at positions
R2 and R3 were reversed. We also addressed the amino acids bear-
ing functional groups on their side chains. Thus, in analog A2, the
serine at position 6 was replaced by the threonine and in analog
A3, the tyrosine at position R4 and the serine at R6 were both
substituted by the threonine. In analogs A5 and A10, we addressed
the importance of tyrosine at position R4 by replacing it with serine
and O-methyl tyrosine. We also addressed the importance of hav-
ing glycine at position R1 and R2 by replacing them with tyrosine in
analogs A6 and A7, respectively.
Similarly, we performed the parallel solid phase synthesis15 of
13 analogs of cherimolacyclopeptide B (B). In comparison to the
natural cyclic peptide we used the methionine instead of the sulf-
oxide methionine (B1–B13). As outlined in Scheme 2, the parallel
synthesis was performed using Fmoc-chemistry.
Since the natural cyclic peptide cherimolacyclopeptide B (B)
contains glutamine, the methionine containing analogs of cheri-
molacyclopeptide B were synthesized using amino acid side chain
attachment strategy.16 Thus, starting from p-methylbenzhydryl-
amine hydrochloride (MBHAÁHCl) resin, and following neutraliza-
tion of the resin, Fmoc-Glu-O-tBu was attached to the solid
support in the presence of DIC and HOBt. Following Fmoc depro-
tection, and subsequent coupling of the amino acids using standard
Table 1
Amino acids sequences of the cyclic cyclosquamosin D (A1) and its analogs
Xaa1
Xaa2
Xaa3
Xaa4
Xaa5
Xaa6
Xaa7
Cyclosquamosin D (A1)
A2
A3
A4
A5
A6
A7
A8
Val
Val
Val
Val
Val
Val
Val
Val
Val
Val
Val
Val
Val
Val
Val
Gly
Gly
Gly
Gly
Gly
Tyr
Gly
Tyr
Gly
Gly
Ala
Gly
Gly
Gly
Gly
Gly
Gly
Gly
Gly
Gly
Gly
Tyr
Pro
Gly
Gly
Gly
Ala
Gly
Pro
Gly
Pro
Pro
Pro
Pro
Pro
Pro
Pro
Gly
Pro
Pro
Pro
Pro
Pro
Gly
Pro
Tyr
Tyr
Thr
Ser
Ser
Tyr
Tyr
Tyr
Tyr
Tyr(OMe)
Tyr
Tyr
Ala
Tyr
Tyr
Tyr
Tyr
Tyr
Tyr
Tyr
Tyr
Tyr
Tyr
Tyr
Tyr
Tyr
Tyr
Tyr
Tyr
Ser
Thr
Thr
Tyr
Ser
Ser
Ser
Ser
Ser
Ser
Ser
Ser
Ser
Ser
Ala
Leu
Leu
Leu
Leu
Leu
Leu
Leu
Leu
Ala
Leu
Leu
Leu
Leu
Leu
Leu
A9
A10
A11
A12
A13
A14
A14
Tyr

A. Dellai et al. / Bioorg. Med. Chem. Lett. 20 (2010) 5653–5657
5655
O
O
H
N
Fmoc-Glu-OtBu
FmocHN
H2N
O
SPPS
(Fmoc strategy)
O
O
R6
O
R4
O
O
R2
O
H
N
H
N
H
N
H
N
H
N
Fmoc
N
H
N
H
N
H
N
H
R7
O
R5
O
R3
O
R1
O
1) Fmoc deprotection
2) tBu deprotection
HO
O
R2
O
R4
R6
O
O
O
H
H
N
H
H
N
N
N
H2N
N
H
N
H
N
N
H
R1
O
R7
O
R3
H
O
O
R5
HBTU
DIEA, DMF
R2
HN
H
N
R3
HN
R1
O
O
O
R4
NH
O
O
O
HN
O
NH2
O
O
NH
R7
N
NH
H
S
R6
Scheme 2. Parallel synthesis of Met-cherimolacyclopeptide B (B) and its analogs.
Table 2
stepwise Fmoc chemistry, the N-terminal amine was deprotected
and the tert-butyl group on the C of the glutamic acid was re-
a
Amino acids sequences of Met-cherimolacyclopeptide B (B1) and its analogs
moved with a solution of TFA in DCM. The resin-bound linear pep-
tide was treated with HBTU to generate the corresponding cyclic
compounds. The desired cyclic peptides were obtained following
cleavage of the solid support with HF/anisole. All the compounds
were purified using RP HPLC. The purity of these compounds is
higher than 90% for all the compounds, based upon the analytical
traces at k = 214 nm and 254 nm.
We performed the synthesis of the methionine analog B1 of the
cyclic natural peptide cherimolacyclopeptide B. We used an
alanine substitution screen to study the effect and contribution
of each amino acid to the activity (B2–B8). In addition, we synthe-
sized five cherimolacyclopeptide B analogs, where we substituted
the threonine at position R7 by different amino acids bearing the
Xaa1 Xaa2 Xaa3 Xaa4 Xaa5 Xaa6 Xaa7
B1
B2
B3
B4
B5
B6
B7
B8
B9
Fmoc-Glu-
O-tBu
Pro
Ala
Pro
Pro
Pro
Pro
Pro
Pro
Pro
Pro
Pro
Pro
Pro
Ile
Ile
Ala
Ile
Ile
Ile
Ile
Ile
Ile
Ile
Ile
Ile
Ile
Pro
Pro
Pro
Ala
Pro
Pro
Pro
Pro
Pro
Pro
Pro
Pro
Pro
Leu
Leu
Leu
Leu
Ala
Leu
Leu
Leu
Leu
Leu
Leu
Leu
Leu
Met
Met
Met
Met
Met
Ala
Gly
Gly
Gly
Gly
Gly
Gly
Ala
Gly
Gly
Gly
Gly
Gly
Gly
Thr
Thr
Thr
Thr
Thr
Thr
Thr
Ala
Ser
Tyr
Fmoc-Glu-
O-tBu
Fmoc-Glu-
O-tBu
Fmoc-Glu-
O-tBu
Fmoc-Glu-
O-tBu
Fmoc-Glu-
O-tBu
Fmoc-Glu-
O-tBu
Met
Met
Met
Met
Met
Met
Met
Fmoc-Glu-
hydroxyl group on the side chain (Ser, Tyr,
-Tyr) (B9–B13) (Table 2).
Cell culture: Murine macrophage J774A.1 cells were purchased
D-Thr, D-Ser, and
O-tBu
D
Fmoc-Glu-
O-tBu
from the ATCC and cultured in the RPMI 1640 medium supple-
mented with 10% heat inactivated fetal bovine serum (Hyclone),
B10 Fmoc-Glu-
O-tBu
B11 Fmoc-Glu-
O-tBu
B12 Fmoc-Glu-
D
D
D
-Tyr
L-glutamine, penicillin–streptomycin, sodium pyruvate, non-
essential amino acids, HEPES, and 2-mercaptoethanol. For stimula-
tion assays, J774A.1 cells were incubated at 37 °C with the cyclic
-Thr
-Ser
O-tBu
peptides for 1 h, followed by lipopolysaccharide (LPS) (0.1
treatment for additional 4 or 6 h. Cyclic peptides were dissolved
in dimethyl sulfoxide, DMSO at concentration of 20–40 g/mL.
Enzyme-linked immunosorbent assay (ELISA): Anti-mouse IL-6
and TNF- antibodies as well as recombinant mouse IL-6 and
TNF- for ELISA standards were purchased from eBioscience (San
lg/mL)
B13 Fmoc-Glu-
O-tBu
l
a
Diego). Avidin-HRP peroxidase was purchased from Vector
Laboratories and ABTS peroxidase substrate system from KPL.
a

5656
A. Dellai et al. / Bioorg. Med. Chem. Lett. 20 (2010) 5653–5657
Tween solution (0.05%) and then blocked with 200 lL of blocking
buffer (10% FBS in PBS solution) and incubated at room tempera-
ture for 2 h. Following washing four to five times with PBS/Tween,
IL-6, or TNF-
a standards and samples were added (50 lL/well).
Plates were sealed overnight at 4 °C and washed next day five
times with PBS/Tween solution. The biotinylated anti-IL-6 or
anti-TNF-a detection antibodies were diluted to 1 mg/mL in PBS
and added at 100 mL/well. Plates were incubated for 1 h at room
temperature and then washed six times with PBS/Tween. Avidin-
HRP peroxidase conjugate was diluted to 2.5 mg/mL in PBS and
added 100 mL/well and plates were incubated at room tempera-
ture for 30 min. After washing six times with PBS/Tween, ABTS
substrate solution, and peroxidase substrate solutions were mixed
at 1:1 ratio and added at 100 mL/well and plates covered with alu-
minum foil. After 10–20 min the optical density was read at
405 nm using a microplate reader.
Infections with pathogenic agents, including viruses and bacte-
ria can result in overproduction of pro-inflammatory cytokines,
including IL-6 and TNF-a. We have determined whether treatment
with synthetic cyclic peptides has any influence on the secretion of
these cytokines by a murine macrophage cell line, J 774A.1 cells.
These cells were pre-treated with various synthetic cyclic peptides
for 1 h at 37 °C. Following treatment with peptides, cells were
stimulated for 6 h with LPS, a gram-negative bacterial cell wall
component. LPS has been shown to induce signaling for the pro-
duction of cytokines via toll-like receptor 4.
Screening results: All the synthetic cyclic peptides were exam-
ined for their ability to block IL-6 or TNF-a using the J774A.1 cell
line. A wide array of biological activity was observed among the
analogs with some exhibiting no activity with others exhibiting
activity superior to the natural products. Some analogs of both A
and B peptides that were found to show suppression of inflamma-
tory responses in bulk assays were further examined along with
some that showed no activity which were included as negative
controls. As shown in Figure 2, while A10, A11, and A12 signifi-
cantly reduced IL-6 secretion, only A12 was able to block TNF-
a
secretion as well. It is interesting that, although A10 was most
effective (up to ꢀ25%) in suppressing IL-6, it had no effect on the
secretion of TNF-a.
Figure 2. Inhibitory effect of cyclic peptides on the production of TNF-a and IL-6 by
Similar to what was reported for the natural cyclic peptide
cherimolacyclopeptide B, the Met-cherimolacyclopeptide B (B1)
did not show activity. In contrast, both analogs of Met-cherimola-
cyclopeptide B (B1), cyclic peptides B6 and B7 were able to effec-
LPS-stimulated J774A.1 cells. IL-6 secretion after 4 h of LPS treatment (upper panel).
*
Values are mean SD, : p value for A10, A11, A12, B6, B7 is <0.0001. TNF-
a
*
secretion after 4 h of LPS treatment (lower panel). Values are mean SD. : p values
for A12, B6, and B7 are 0.0137, 0.0005, and 0.0268, respectively. These data are
representative of three independent experiments.
tively suppress TNF-
a and IL-6 secretion. Whereas, peptide
analog B10 showed no activity in reducing LPS-induced cytokine
secretion by macrophages and served as a negative control.
Structurally, these peptide analogs are very similar and yet they
generate specific anti-inflammatory activity. The methylation of
the hydroxyl group of the tyrosine (R4) in analog A10 significantly
reduced IL-6 secretion. Similarly, the substitution of the glycine
The cytokine Elisa was performed essentially as described ear-
lier.17,18 Anti-IL-6 or anti-TNF-
a
capture antibodies were diluted
to 1
lg/mL in PBS. Fifty microliters of diluted antibody was added
to the wells of Nunc Maxisorp plate. Plates were then incubated
overnight at 4 °C. The next day they were washed 2Â with PBS/
HO
HO
HO
HO
HN
HO
HO
HN
H
H
N
OH
OH
N
H
HN
O
N
O
NH
N
O
NH
N
O
O
O
O
O
O
HN
O
NH
N
HN
O
O
O
O
HN
O
O
O
O
O
O
O
HN
O
HN
O
O
HN
O
NH
NH
N
N
NH
H
N
H
H
A10
A11
A12
Figure 3. Structures of the individual analogs of cyclosquamosin D showing an inhibitory effect on the production of pro-inflammatory cytokines within lipopolysaccharide
stimulated J774A.1 macrophages.

A. Dellai et al. / Bioorg. Med. Chem. Lett. 20 (2010) 5653–5657
5657
O
Houghten), the State of Florida Funding and the Tunisian Ministry
of Higher Education.
O
NH2
N
H
N
HN
N
NH2
NH
N
O
O
O
O
NH
NH
O
References and notes
OH
O
O
O
HN
NH
O
O
O
O
1. (a) Tamai, S.; Kaneda, M.; Nakamura, S. J. Antibiot. 1982, 35, 1130; (b) Walsh, C.
T. Science 1993, 261, 308; (c) Kannan, R.; Harris, C. M.; Harris, T. M.; Waltho, G.
P.; Skelton, N. J.; Williams, D. H. J. Am. Chem. Soc. 1988, 110, 2946.
2. (a) Wipf, P. Chem. Rev. 1995, 95, 2115; (b) Hamada, Y.; Shioiri, T. Chem. Rev.
2005, 105, 4441; (c) Pomilio, A. B.; Battista, M. E.; Vitale, A. A. Curr. Org. Chem.
2006, 10, 2075.
OH
O
O
O
O
N
NH
N
H
NH
N
NH
H
S
3. Blout, E. R. Biopolymers 1981, 20, 1901.
B6
B7
4. Mongkolvisut, W.; Sutthivaiyakit, S.; Leutbecher, H.; Mika, S.; Klaiber, I.; Moller,
W.; Rosner, H.; Beifuss, U.; Conrad, J. J. Nat. Prod. 2006, 69, 14355.
5. Dahiya, R.; Kaur, K. Arch. Pharm. Res. 2007, 30, 1380.
6. Morita, H.; Shishido, A.; Matsumoto, T.; Itokawa, H.; Takeya, K. Tetrahedron
1999, 55, 967.
Figure 4. Structures of the individual analogs of cherimolacyclopeptide B showing
an inhibitory effect on the production of pro-inflammatory cytokines within
lipopolysaccharide stimulated J774A.1 macrophages.
7. Baraguey, C.; Blond, A.; Correia, I.; Pousset, J.-L.; Bodo, B.; Auvin-Guette, J.-L.;
Mahafacyclin, A. Tetrahedron Lett. 2000, 41, 325.
8. Morita, H.; Lizuka, T.; Choo, C. Y.; Chan, K. L.; Itokawa, H.; Takeya, K. J. Nat. Prod.
2005, 68, 1686.
9. Morita, H.; Shishido, A.; Kayashita, T.; Takeya, K.; Itokawa, H. J. Nat. Prod. 1997,
60, 404.
10. Yahara, S.; Shigeyama, C.; Nohara, T.; Okuda, H.; Wakamatsu, K.; Yasuhara, T.
Tetrahedron Lett. 1989, 30, 6041.
11. Yang, Y. L.; Hua, K. F.; Chuang, P. H.; Wu, S. H.; Wu, K. Y.; Chang, F. R.; Wu, Y. C.
J. Agric. Food Chem. 2008, 56, 386.
12. (a) Merrifield, R. B. J. Am. Chem. Soc. 1963, 85, 2149; (b) Merrifield, R. B. Science
1986, 232, 341.
(R1) and the glycine (R2) by alanine for analogs A11 and A12,
respectively, resulted in reduction of IL-6 secretion (Fig. 3).
Interesting results were obtained for cyclic peptides Met-cheri-
molacyclopeptide B analogs. The substitution of methionine (R5)
and Gly (R6) by alanine for analogs B6 and B7 (Fig. 4), respectively,
resulted in suppression of TNF-a and IL-6 secretion.
In summary, we performed the parallel synthesis of two natural
cyclopeptides and their analogs. We tested all the compounds for
anti-inflammatory activity for their inhibitory effect on the pro-
duction of pro-inflammatory cytokines using lipopolysaccharide
stimulated J774A.1 macrophage line. Modified analog compounds
of cyclosquamosin D (A1) showed significant anti-inflammatory
13. Mitchell, A. R. Biopolymers 2008, 90, 3175.
14. Fields, G. B.; Noble, R. L. Int. J. Pept. Protein Res. 1990, 35, 161.
15. Houghton, R. A. Proc. Natl. Acad. Sci. U.S.A. 1985, 82, 5131.
16. (a) Cabrele, C.; Langer, M.; Beck-Sickinger, A. G. J. Org. Chem. 1999, 64, 4353; (b)
Ficht, S.; Payne, R. J.; Guy, R. T.; Wong, C.-H. Chem.-A Eur. J. 2008, 14, 3620; (c)
Dixon, M. J.; Nathubhai, A.; Andersen, O. A.; van Aalten, D. M. F.; Eggleston, I. M.
Org. Biomol. Chem. 2009, 7, 259.
17. Tang, X.; Maricic, I.; Purohit, N.; Bakamjian, B.; Reed-Loisel, L. M.; Beeston, T.;
Jensen, P.; Kumar, V. J. Immunol. 2006, 177, 7645.
18. Smith, T. R.; Tang, X.; Maricic, I.; Garcia, Z.; Fanchiang, S.; Kumar, V. J. Immunol.
2009, 182, 6959.
activity in suppressing the secretion of IL-6 and TNF-a, in some
cases superior to the activity observed with the natural product.
Similarly, we were able to identify active compounds analogs of
the non-active natural cyclic peptide cherimolacyclopeptide B.
19. 10 mg of linear crude peptide was dissolved in 600 lL of DMF anhydrous.
Acknowledgment
PyBOP (3 equiv) and HOBt (3 equiv) were added to the reaction vessel in the
presence of DIEA (3 equiv). The clear solution was stirred overnight at room
temperature. The completion of the cyclization was monitored by LC–MS. The
DMF was evaporated and the crude product was purified using reversed-phase
high-performance liquid chromatography (RP HPLC). All the compounds were
obtained in good yield ranging from 50 to 60%.
The presented work was supported by Grants from the NIH,
JDRF, and MSNRC (V.K.). NIH (1R03DA025850-01A1, Nefzi), NIH
(5P41GM081261-03, Houghten), and NIH (3P41GM079590-03S1,