Stereospecific pseudoproline ligation of N-terminal serine, threonine, or cysteine-containing unprotected peptides

Article abstract of DOI:10.1021/ja991153t

This paper describes an orthogonal and stereospecific method for ligating free peptide segments to form a monosubstituted pseudoproline bond with a hydroxymethyl moiety at the C2 carbon. The pseudoproline ligation, comprising both the oxaproline and thiap

Full text of DOI:10.1021/ja991153t

J. Am. Chem. Soc. 1999, 121, 9013-9022
9013
Stereospecific Pseudoproline Ligation of N-Terminal Serine,
Threonine, or Cysteine-Containing Unprotected Peptides
James P. Tam* and Zhenwei Miao
Contribution from the Department of Microbiology and Immunology, Vanderbilt UniVersity, MCN A5119,
NashVille, Tennessee 37232
ReceiVed April 12, 1999
Abstract: This paper describes an orthogonal and stereospecific method for ligating free peptide segments to
form a monosubstituted pseudoproline bond with a hydroxymethyl moiety at the C2 carbon. The pseudoproline
ligation, comprising both the oxaproline and thiaproline ligations, initially involves an imine capture of a
peptidyl glycoaldehyde ester with an N-terminal cysteine, serine, or threonine peptide segment and then two
spontaneous cyclization reactions. The thiazolidine or oxazolidine ester formed in the first cyclization undergoes
an O,N-acyl transfer to form an pseudoproline bond. The thiaproline ligation can be carried out exclusively
with unprotected peptide segments in both aqueous and nonaqueous pyridine-acetic acid conditions. However,
the oxaproline ligation is best performed in a nonaqueous pyridine-acetic acid mixture with unprotected peptide
segments except for those containing N-terminal nucleophilic amino acids such as Cys, His, and Trp.
1
Pseudoproline ligation is not only regioselective but also stereospecific. 2D H NMR studies of dipeptide
models, Z-Xaa-ψPro-OMe, indicate that the newly created C2 stereocenter of the pseudoproline ring affords
only an R-epimer and the C2-hydroxymethyl-substituted pseudoproline exhibits high preference for cis
conformation. Three of the model peptides have more than 50% cis isomers. Finally, this novel method has
been used successfully in ligating two segments of 24 and 35 amino acids under mild conditions to synthesize
three analogues of bactenecin 7, an antimicrobial peptide containing 59 amino acid residues.
Introduction
aminoacid)inpeptidesandproteinshasbeenwelldocumented,9-12
and proline has a relatively high ability to impart the cis-isomer.
Orengo et al.13 have shown the occurrence of a cis Pro bond in
70 out of 205 Pro-containing proteins in a nonredundant
database of 214 protein structures. In bovine pancreatic ribo-
nuclease A, two of the four Xaa-Pro bonds are found to be
cisoid.14 Furthermore, the cis-trans isomerization of Xaa-Pro
bonds in proteins is a slow conformational interconversion
process that has been shown to be the rate-limiting step in their
refolding pathways.15,16
The combined effect of the structural peculiarity of Pro and
the cis-trans isomerization of an Xaa-Pro bond has led the
development of various mimetics and analogues that are
intended to produce Pro-like reverse turns and gain better control
of the cis-trans ratio.17-24 Toward this end, pseudoprolines
(ψPro) derived from Cys, Ser, and Thr as thiazolidine-4-
Proline, an imino acid, plays a unique role in the structures
and folding pathways of peptides and proteins.1,2 Structurally,
it often serves as a hinge amino acid because of its frequent
occurrence in the locale where peptide chains turn and reverse
directions. Indeed, Pro has been found with a strong preference
in turns occupying the i, i + 1 (types I and II), i + 2 (type IV),
or i + 3 (type VIII) position.3-5 Pro-rich sequences also exist
as extended helices in tandemly repetitive occurrences as found
in collagens and antimicrobial peptides.6,7 In peptides and
peptidomimetics, the effect of Pro is believed to confer bioactive
conformation or recognition or to increase bioavailability
because of the increased lipophilicity and protease-resistance
of the imide bond over the secondary amide bond.8
The cis-trans isomerization of Xaa-Pro bonds (Xaa ) any
* To whom correspondence should be addressed: James P. Tam,
Department of Microbiology and Immunology, Vanderbilt University, MCN
A5119, Nashville, TN 37232. Telephone: (615)-343-1465. Fax: (615)-
343-1467. E-mail: Tamjp@ctrvax.vanderbilt.edu.
(1) MacArthur, M. W.; Thornton, J. M. J. Mol. Biol. 1991, 218, 397-
412.
(2) Schmid, F. X. Annu. ReV. Biophys. Biomol. Struct. 1993, 22, 123-
143.
(3) (a) Wilmot, C. M.; Thornton, J. M. J. Mol. Biol. 1988, 203, 221-
232. (b) Hutchinson, E. G.; Thornton, J. M. Protein Sci. 1994, 3, 2207-
2216.
(9) Rose, G. D.; Gierasch, L. M.; Smith, J. A. AdV. Prot. Chem. 1985,
37, 1-109.
(10) Reimer, U.; Scherer, G.; Drewello, M.; Kruber, S.; Schutkowski,
M.; Fischer, G. J. Mol. Biol. 1998, 279, 449-460.
(11) Bisang, C.; Weber, C.; Inglis, J.; Schiffer, C. A.; van Gunsteren,
W. F.; Jelesarov, I.; Bosshard, H. R.; Robinson, J. A. J. Am. Chem. Soc.
1995, 37, 1-109.
(12) Schultz, G, D.; Schirmer, R H. Principles of Protein Structure;
Springer-Verlag: New York, 1978; p 25.
(13) Orengo, C. A. Flores, T. P.; Taylor, W. R.; Thornton, J. M. Protein
Eng. 1993, 6, 485-500.
(4) Richardson, J. S. AdV. Protein Chem. 1981, 34, 168-339.
(5) (a) Yao, J.; Feher, V. A.; Espejo, B. F.; Reymond, M. T.; Wright, P.
E.; Dyson, H. J. J. Mol. Biol. 1994, 243, 736-753. (b) Yao, J.; Dyson, H.
J.; Wright, P. E. J. Mol. Biol. 1994, 243, 754-766.
(6) Engel, J.; Prockop, D. J. Annu. ReV. Biophys. Biophys. Chem. 1991,
20, 137-152.
(7) Nicolas, P.; Mor, A. Annu. ReV. Microbiol. 1995, 49, 305-333.
(8) (a) Yaron, A. Biopolymers 1987, 26 Suppl. 215-222. (b) Yaron, A.;
Naider, F. Crit. ReV. Biochem. Mol. Biol. 1993, 28, 31-81.
(14) Lin, L.-N.; Brandts, J. F. Biochemistry 1983, 22, 559-563.
(15) (a) Mayr, L. M.; Landt, O.; Hahn, U.; Schmid, F. X. J. Mol. Biol.
1993, 231, 897-912. (b) Mayr, L. M.; Schmid, F. X. J. Mol. Biol. 1993,
231, 913-926.
(16) Reinstadler, D.; Fabian, H.; Naumann, D. Proteins 1999, 34, 303-
316.
(17) (a) Chalmers, D. K.; Marshall, G. R. J. Am. Chem. Soc. 1995, 117,
5927-5937. (b) Takeuchi, Y.; Marshall, G. R. J. Am. Chem. Soc. 1998,
120, 5363-5372.
10.1021/ja991153t CCC: $18.00 © 1999 American Chemical Society
Published on Web 09/16/1999

9014 J. Am. Chem. Soc., Vol. 121, No. 39, 1999
Tam and Miao
Figure 1. Structures of pseudoprolines (ψPro), 4-Thz and 4-Oxa,
derivatized respectively from Cys and Ser/Thr, exhibiting cis-trans
isomerization of the Xaa-ψPro bond. In this study, monosubstituted
ψPro with a hydroxymethyl moiety at C2 are denoted as: a, thiaproline
(SPro), X ) S, R3 ) H; b, oxaproline (OPro), X ) O, R3 ) H; c,
5-methyloxaproline (OProMe), X ) O, R3 ) CH3.
Figure 2. General scheme of ψPro ligation of an acyl segment 1 with
an amine segment 2a-c to form a thiaproline bond 5a and an oxaproline
bond 5b or 5c.
carboxylic acid (4-Thz) and oxazolidine-4-carboxylic acid (4-
Oxa) have been developed as Pro surrogates (Figure 1). These
surrogates can be accessed by various carbonyl derivatives to
afford the corresponding mono- and disubstituted derivatives
at the C2 carbon to modulate the cis-trans isomers of the imidic
Xaa-Pro bond.22-24 Model dipeptides based on Xaa-ψPro
showed that monosubstitution at C2 generally stabilizes cis
conformation while substitution at C4 decreases cis isomers.
Disubstitution at C2, such as dimethyl-4-Thz, produces >95%
cis isomer.22,23
Thus far, incorporation of an unsubstituted or monosubstituted
ψPro has been used in small peptides. The 4-Thz derivatives
have been found in the synthesis of small bioactive peptides
such as somatostatin,25 angiotensin II,26 oxytocin,27 and various
protease inhibitors.28,29 In contrast, 4-Oxa and its substituted
analogues are rarely used30 because of synthetic difficulties.
Although several methods have been developed based on either
the acid or metal-catalyzed condensation of aldehydes or ketones
to form ψPro,31-33 these methods are conducted with small
protected peptides and generally under conditions which are not
suitable for the synthesis of large ψPro-containing peptides.
Over the past six years, we have explored orthogonal ligations
of free peptides to form amide bonds in aqueous solutions, which
afford products without further need for a deprotection step.34-40
Thiaproline ligation is perhaps the first example that validates
the principle of orthogonal ligation to produce spontaneously
an amide bond by ligating two unprotected peptide segments
(Figure 2).34 It utilizes an acyl peptide segment bearing a
glycoaldehyde ester (peptideyl-OCH2CHO) 1 to capture an
N-terminal nucleophile (Ntn)-Cys segment 2a, through an imine
3a which rapidly tautomerizes to a thiazolidine ester 4a. The
O-ester is then rearranged through an O,N-acyl shift to a
monosubstituted thiaproline (SPro) bond 5a containing a hy-
droxymethyl moiety at C2 of the thiaproline ring. However,
the stereochemistry of the newly created chiral center at C2
has not been determined. The SPro bond formed at the ligation
site is a proline mimetic that retains the amide backbone
structure.34,40 We envision that similar chemistry may be feasible
for obtaining the analogous oxaproline (OPro) with the Ntn-
Ser 2b or Ntn-Thr 2c segment, thus providing another pathway
to pseudoprolines based on imine capture, ring-chain tautomer-
ization, and O,N-acyl migration 3-5 (Figure 2). However, under
similar conditions, oxaproline ligation using Ntn-Ser or Thr
segments 2b,c has not been successful in aqueous buffers at
pH 4 to 6.
(18) Genin, M. J.; Johson, R. L. J. Am. Chem. Soc. 1992, 114, 8778-
8783.
(19) Gramberg, D.; Weber, C.; Beeli, R.; Inglis, J.; Bruns, C.; Robinson,
J. A. HelV. Chim. Acta 1995, 78, 1588-1606.
(20) Delaney, N. G.; Madison, V. J. Am. Chem. Soc. 1982, 104, 6635-
6641.
(21) Magaard, V. W.; Sanchez, R. M.; Bean, J. W.; Moore, M. L.
Tetrahedron Lett. 1993, 34, 381-384.
(22) Kern, D.; Schutkowski, M.; Drakenberg, T. J. Am. Chem. Soc. 1997,
119, 8403-8408.
(23) Dumy, P.; Keller, M.; Ryan, D. E.; Rohwedder, B.; Wo¨hr, T.;
Mutter, M. J. Am. Chem. Soc. 1997, 119, 918-925.
(24) Keller, M.; Sager, C.; Dumy, P.; Schutkowsky, M.; Fischer, G. S.;
Mutter, M. J. Am. Chem. Soc. 1998, 120, 2714-2720.
(25) (a) Pattaroni, C.; Lucietto, P.; Goodman, M.; Yamamoto, G.; Vale,
W.; Moroder, L.; Gazero, L.; Gohring, W.; Schmied, B.; Wunsch, E. Int.
J. Pept. Protein Res. 1990, 36, 401-417. (b) Mieke, D. F.; Pattaroni, C.;
Delaet, N.; Toy, A.; Goodman, M.; Tancredi, T.; Motta, A.; Temussi, P.
A.; Moroder, L.; Bovermann, C.; Wunsch, E. Int. J. Pept. Protein Res.
1990, 36, 418-432.
(26) Samanen, J.; Cash, T.; Narindray, D.; Brandeis, E.; Adams, W.;
Weidemann, H.; Yellin, T. J. Med. Chem. 1991, 34, 3036-3043.
(27) Rosamond. J. D.; Ferger, M. F. J. Med. Chem. 1976, 19, 873-876.
(28) Tsuru, D.; Yosimoto, T.; Koroyama, N.; Furukawa, S. J. Biochem.
1997, 104, 580-586.
The chemistry to obtain oxazolidine and thiazolidine through
a 1,2-dinucleophile of an amino alcohol or amino thiol and a
carbonyl is one of the best known reactions41-44 and ring-chain
tautomerization of both thiazolidine and oxazolidine has been
extensively studied.45-48 The capture step in the oxaproline
(34) (a) Liu, C. F.; Tam, J. P. J. Am. Chem. Soc. 1994, 116, 4149-
4153. (b) Liu, C. F.; Tam, J. P. Proc. Natl. Acad. Sci. U.S.A. 1994, 91,
6584-6588.
(35) Tam, J. P.; Lu, Y.-A.; Liu, C. F.; Shao, J. Proc. Natl. Acad. Sci. U.
S.A. 1995, 92, 12485-12489.
(36) Tam, J. P.; Lu, Y.-A. Tetrahedron Lett. 1997, 38, 5599-5602.
(37) Zhang, L.; Tam, J. P. Tetrahedron Lett. 1997, 38, 3-6.
(38) Tam, J. P.; Yu, Q. Biopolymers 1998, 46, 319-327.
(39) Liu, C. F.; Rao, C.; Tam, J. P. Tetrahedron Lett. 1996, 37, 933-
936.
(40) Liu, C. F.; Chang, R.; Tam, J. P. J. Am. Chem. Soc. 1996, 118,
307-312.
(29) Kiso, Y.; Yamaguchi, S.; Matsumoto, H.; Mimoto, T.; Kato, R.;
Nojima, S.; Takaku, H.; Fukazawa, T.; Kimura, T.; Akaji, K. In Peptides:
Frontiers of Peptide Science; Proceedings of the 15th American Peptide
Symposium; Tam, J. P., Kaumaya, P. T. P., Eds; Kluwer/Escom: Boston,
1999; pp 667-669.
(41) (a) Kallen, R. G. J. Am. Chem. Soc. 1971, 93, 6227-6235. (b)
Kallen, R. G. J. Am. Chem. Soc. 1971, 93, 6236-6248.
(42) Wolfe, S.; Militello, G.; Ferrari, C.; Hasan, S. K.; Lee, S. L.
Tetrahedron Lett. 1979, 41, 3913-3916.
(30) Rahfeld, J.; Schutkowski, M.; Faust, J.; Neubert, K.; Barth, A.;
Heins, J. Biol. Chem. Hoppe-Seyler 1991, 372, 313-318.
(31) Mutter, M.; Nefzi, A.; Sato. T.; Sun, X.; Wahl, F.; Wohr, T. Peptide
Res. 1995, 8, 145-153.
(32) Wo¨hr, T.; Wahl, F.; Nefzi, A.; Rohwedder, B.; Sato. T.; Sun, X.;
Mutter, M. J. Am. Chem. Soc. 1996, 118, 9218-9227.
(33) Seebach, D.; Sommerfeld, T.; Jiang, Q.; Venanzi, L. M. HelV. Chim.
Acta 1994, 77, 1313-1330.
(43) Garner, P.; Park, J. M. J. Org. Chem. 1986, 52, 2361-2364.
(44) Schmolka, I. R.; Spoerri, P. E. J. Org. Chem. 1957, 22, 943-946.
(45) Lambert, J. B.; Majchrzak, M. J. Am. Chem. Soc. 1980, 102, 3588-
3591.
(46) (a) Fu¨lo¨p, F.; Berna´th, G.; Mattinen, J.; Pihlaja, K. Tetrahedron
1989, 45, 4317-4324. (b) Fu¨lo¨p, F.; Pihlaja, K. Tetrahedron 1993, 49,
6701-6706. (c) Szakonyi, Z.; Fu¨lo¨p, F.; Berna´th, G.; Evanics, B.; Riddell,
F. Tetrahedron 1998, 54, 1013-1020.

Stereospecific Pseudoproline Ligation
J. Am. Chem. Soc., Vol. 121, No. 39, 1999 9015
ligation is also the formation of an imine. As opposed to the
thiaproline ligation, the step from 3b,c to 4b,c is not favored in
aqueous conditions, which is in accord with precedents in the
literature.45-48 Oxazolidine ring-chain tautomerization generally
prefers the open imine form, and the oxazolidine has been found
to be 104 times less stable than the analogous thiazolidine.48
Furthermore, the ring closure of the imine to oxazolidine is a
disfavored process in accordance to the Baldwin rule 1,5-endo-
trigonal addition.49 These factors account for the difficulty in
achieving oxazolidine ligation in aqueous conditions. However,
this difficulty can be viewed as an opportunity for site-specific
orthogonal ligation between an Ntn-Cys and an Ntn-Ser or Ntn-
Thr peptide to yield a thiaproline or oxaproline bond of multiple
segments in a specific order without protecting groups.
The goals of this paper focus on the development of an
orthogonal ligation method to afford a pseudoproline, particu-
larly an oxaproline, at the ligation site, the regiospecificity of
pseudoproline ligation, the stereochemistry of the newly created
C2 carbon with the hydroxymethyl substitution, as well as the
application of this newly developed chemistry to two-segment
ligation to yield large peptides.
Figure 3. Synthetic scheme of peptide glycoaldehyde esters 8a-d
from an acetal resin support 6 (left) and amine segments 10a-l from
an MBHA resin 9 (right).
Results
conditions of 90% DMSO and 10% aqueous buffers but at a
very slow rate.51 Since unprotected peptide building blocks were
used for the oxaproline ligation, more desirable and compatible
polar solvents such as pyridine, acetic acid, and low-molecular-
weight alcohol were studied. Similar to the analogous thiaproline
ligation, the initial step of oxaproline ligation is known to be
imine formation, which is both acid- and base-catalyzed, with
the optimal pH in the range of 4 to 6.34 Pyridine-acetic acid
mixtures would provide the desired pH range and are known
to be excellent solvent combinations for dissolving unprotected
peptide segments. Furthermore, both solvents are volatile and
can be removed by lyophilization. Based on this rationale, we
explored the rates of oxaproline ligation between Leu-Ile-Leu-
Asn-Gly-OCH2CHO 8a and Ser-Phe-Lys-Ile-amide 10a in seven
solutions ranging from 0 to 100% pyridine in acetic acid (data
not shown). A mixture of pyridine-acetic acid at a 1:1 molar
ratio was found to be suitable to mediate oxaproline ligation in
78% yield after 45 h at 20 °C. Under similar conditions, other
mixtures with a 2:3, 2:1, 1:2 and 4:1 molar ratio of pyridine-
acetic acid also resulted in acceptable yields of 74, 71, 69 and
62%, respectively. However, pyridine alone gave only 49%
yield, while no observable product was found with acetic acid
alone. The ligation product 11a containing an oxaproline with
a hydroxymethyl substitution at C2 carbon was identified by
both chemical and spectroscopic methods and will be discussed
later in detail.
We then studied this pyridine-acetic acid mixture (1:1, mol/
mol) under various conditions including temperature and
cosolvent mixture to accelerate the ligation between 8a and 10a
(Table 1). Elevating the temperature from 20 °C to 40 °C
accelerated the reaction about 2-3-fold. Polar cosolvents such
as TFE accelerated the reaction 2-3-fold and > 70% of ligation
was achieved at 6.5 h in 50% TFE and 50% pyridine-acetic
acid (1:1, mol/mol). Aprotic polar solvents such as DMF slightly
retarded the ligation rate as compared to pyridine-acetic acid
(1:1, mol/mol) alone, while the aqueous environment of 50%
H2O was strongly inhibitory.
Synthesis of Acyl and Amine Peptide Segments. Unpro-
tected peptide building blocks consisting of an acyl segment
bearing a glycoaldehyde at its C terminus and an amine segment
carrying the Ntn-amino acid such as Ser, Thr, or Cys were
prepared by stepwise solid-phase method.
Several methods for preparing the acyl segment bearing a
C-terminal glycoaldehyde have been developed, including an
indirect method using an n + 1 strategy that either chemically
or enzymatically couples a single amino acid derivative as an
acetal with an unprotected peptide ester or thioester.34,40 A more
convenient method for obtaining the peptidyl glycoaldehyde 8
from a peptide glyceryl ester precursor 7 has also been
developed using an acetal resin 6 (Figure 3).50 This method
requires Fmoc/tBu chemistry because the acetal is unstable to
a strong acid such as anhydrous TFA which is used for
deprotection in the tBoc/Bzl strategy for peptide synthesis. After
the TFA-mediated cleavage of the peptidyl ester from the acetal
resin 6, the unprotected glyceryl ester segment 7 carrying a 1,2-
diol moiety was then transformed into aldehyde 8 by periodate
oxidation under aqueous conditions at pH 2 to 7. This two-step
conversion scheme has the advantage of avoiding exposure of
the sensitive aldehyde moiety to anhydrous TFA and the
subsequent purification steps. All acyl segments 8a-d bearing
a glycoaldehyde were prepared by the two-step method.
The amine segments 10a-l, including those containing Ser
10a, Thr 10b, and Cys 10c at the N termini, were achieved by
tBoc/Bzl chemistry to afford completely free peptides after
cleavage by HF from the p-methylbenzhydrylamine resin 9
(Figure 3). All unprotected building blocks, the acyl segments
8a-d and amine segments 10a-l, were purified by HPLC and
confirmed by TOF-MALDI mass spectrometry and amino acid
analysis.
Conditions for Oxaproline Ligation. Previously, we found
that oxaproline ligation could be achieved in nearly anhydrous
(47) Valters, R. E.; Fu¨lo¨p, F.; Korbonits, D. AdVances In Heterocyclic
Chemistry; Academic Press: New York, 1995; Vol. 64, pp 251-321.
(48) Fu¨lo¨p, F.; Marttinen, J.; Pihlaja, K. Tetrahedron 1990, 46, 6545-
6552.
(49) (a) Baldwin, J. E. J. Chem. Soc., Chem. Commun. 1976, 734-736.
(b) Baldwin, J. E. Tetrahedron 1982, 38, 2939-2947.
(50) Botti, P.; Pallin, D. P.; Tam, J. P. J. Am. Chem. Soc. 1996, 118,
10018-10024.
We also investigated the effects of other amine bases in acetic
acid on the ligation between Asp-Ser-Phe-Gly-OCH2CHO 8b
(51) Tam, J. P.; Rao, C.; Liu, C. F.; Shao, J. Int. J. Pept. Protein Res.
1995, 45, 209-216.

9016 J. Am. Chem. Soc., Vol. 121, No. 39, 1999
Tam and Miao
Table 1. Solvent Effect on the Oxaproline Ligation between 8a
and 10aa
Met 10j, and Tyr 10k as well as with those Ntn-amine segments
bearing acidic side chains such as Asp 10l. Ntn-Lys 10g and
Ntn-Arg 10h were also found to be excluded from the ligation
reaction, probably because of the low pKa of the pyridine-
acetic acid condition and the protonated state of the strongly
basic side chains of lysine and arginine. However, ligation
reactions were observed with those Ntn-segments 10a-f. Rate
analysis showed that ligation occurred almost spontaneously
with Ntn-Cys segment 10c, rapidly with those Ntn-amino acid
segments such as Ser 10a, Thr 10b, Trp 10d, and His 10e, but
very slowly with Asn 10f.
yield (%)
solvent
2.5 h
6.5 h
20 h
45 h
pyridine/HOAc
TFE
DMF
42.1
59.8
10.5
8.7
50.5
76.6
28.4
7.6
69.2
83.1
49.6
3.1
75.6
83.7
55.9
2.1
H2O
a 50% solvent in 50% pyridine-HOAc (1:1, mol/mol). TFE,
trifluoroethanol; DMF, N,N-dimethylformide; HOAc, acetic acid.
Table 2. Effect of Base on the Oxaproline Ligation between
Asp-Ser-Phe-Gly-OCH2CHO, 8b, and Ser-Phe-Lys-Ile-amide, 10a,
in 50% TFE and 50% Base/Acetic Acid (1:1, mol/mol)a
The Ntn-Ser 10a, Ntn-Thr 10b, or Ntn-Asn 10f segment
provided a predominant single ligation product. However, Ntn-
Trp 10d segment gave two, while Ntn-His 10e yielded three
major isomeric products. The stereochemistries of these isomeric
products from 10d,e have not been characterized at this time.
However, these results also show that imine ligation is fairly
general and Ntn amino acids containing other weak base
nucleophiles could participate in ligation similar to Ntn-Cys,
-Ser, and -Thr.
yield (%)
base
pKa
6 h
10.5 h
24 h
pyridine
DMAP
imidazole
NMIm
DIEA
piperidine
5.25
63.8
66.8
79.4
80.4
38.1
8.9
70.6
67.4
82.2
85.2
49.5
15.4
71.8
69.8
83.6
88.4
66.1
16.8
6.95
6.95
10.60
11.32
It should be noted that there are two R-amines and an ꢀ-amine
in the reactants of 8a and 10a-l. Thus, a number of products
could be obtained including oligomers, cyclized products, or
cyclodimerizations through intramolecular ligation of 8a, and
side-chain acylation of ꢀ-amine of lysine. On the basis of the
combined use of HPLC and MS analysis, no detectable side
products stemming from oligmerization or cyclization of 8a were
found. End group analysis was used to determine the regiospeci-
ficity between R- and ꢀ-amines in 11a-c. The results revealed
that the ligation occurred exclusively on the N-terminal R-amines
of 10a-c. These results suggest that the reaction of pseudopro-
line ligation is highly regiospecific with an Ntn-amino acid.
However, minor side products, usually <10%, were detected.
These include hydrolysis of the glycolaldehyde 8a to the
corresponding to free carboxylic acid and transesterification to
the TFE ester when TFE was used as a cosolvent.
a DMAP, 4-(dimethylamino)pyridine; NMIm, N-methyl imidazole;
DIEA, N,N-diiospropylethylamine.
Table 3. Regiospecificity of N-Terminal Amino Acid in the
Ligation between the Acyl Segment L-I-L-N-G-OCH2CHO 8a and
Different Amine Segments X-F-K-I-amide 10a-k in
Pyridine-Acetic Acid (1:1, mol/mol)
[ψP] in producta
segment X )
yield (%) 36 h
relative rate
[OP] 11a
[OPMe] 11b
[SP] 11c
11d
11e
11f
S 10a
T 10b
C 10c
W 10d
H 10e
N 10f
K 10g
R 10h
A 10i
M 10j
D 10k
Y 10l
79.6
74.2
95.3
77.9
73.3
44.4
<1
1
0.95
>1000b
1.2
1.1
0.3
Competitive experiments were employed to establish the
orthogonality between thiaproline ligation and oxaproline liga-
tion. Under aqueous conditions at pH 4 to 6, thiaproline ligation
product Z-Ala-SPro-OMe 17f was exclusively obtained between
Z-Ala-OCH2CHO 15a and Cys-OMe 16c in the presence of Ser-
OMe 16a and Thr-OMe 16b. Similarly, when an acyl segment
15a and three amine segments 16a-c were mixed in an
equimolar ratio in pyridine-acetic acid (1:1, mol/mol), only
the thiaproline ligation between 15a and 16c to afford 17f was
observed. However, in the absence of 17c, the oxaproline
ligation between 15a and 16a,b went smoothly in pyridine-
acetic acid. Under the condition of aqueous buffers or pyridine-
acetic acid mixture (1:1, mol/mol), competitive experiments
using unprotected peptide segments Leu-Ile-Leu-Asn-Gly-
OCH2CHO 8a, Ser-Phe-Lys-Ile-NH2 10a, Thr-Phe-Lys-Ile-NH2
10b, and Cys-Phe-Lys-Ile-NH2 10c gave similar results. These
results are consistent with the literature48,51 that the formation
of thiazolidine 4a is >1000-fold faster than that of oxazolidine
4b,c. Thus, thiaproline and oxazoproline ligations are semi-
orthogonal, i.e., thiaproline ligation can be performed in the
presence of Ntn-Ser/Thr or other N-terminal amino acid-
containing peptide segments under either aqueous conditions
or pyridine-acetic acid mixture (1:1, mol/mol). However,
regioselective oxaproline ligation can be carried out in the
pyridine-acetic acid conditions in the absence of Ntn-Cys, -His,
and -Trp peptide segments.
<1
<1
<1
<1
<1
a [OP], 2-hydroxymethyl-oxaproline; [OPMe], 2-hydroxymethyl,
5-methyl-oxaproline; [SP], 2-hydroxymethyl-thiaproline; two isomeric
products for 11d and three isomeric products for 11e and one of 11f
were obtained; for their possible structures, see ref 51. b Rate for
thiazolidine ester formation (tR ) 31.9 min), but the rate for amide
product (tR ) 29.9 min) through O,N-acyl transfer is 0.9 relative to
11a.
and 10a (Table 2). Weak bases such as imidazole, N-meth-
ylimidazole (NMIm) or 4-(dimethylamino)pyridine (DMAP) in
a 1:1 molar ratio with acetic acid gave results similar to pyridine,
while strong bases such as the tertiary and secondary alkylamine
were not effective. These results are consistent with the
importance of using a proper solvent mixture to achieve imine
capture for oxaproline ligation.
Regiospecificity, Orthogonality, and Steric Factors. To
determine the regiospecificity of oxaproline ligation in the
pyridine-acetic acid mixture, the acyl segment 8a bearing an
glycoaldehyde ester was chosen to react with various amine
segments 10a-l bearing either Ntn- or non-Ntn-amino acids to
form 11a-l (Table 3). In the amine segments, an internal lysine
residue is present to serve as a side-chain amine to test the
regiospecifity of the ψPro ligation. No ligation products were
observed with non Ntn-amine segments 10i-l such as Ala 10i,
Finally, the steric influence of the C-terminal amino acids of
the acyl segments on oxaproline ligation was also studied (Table

Stereospecific Pseudoproline Ligation
J. Am. Chem. Soc., Vol. 121, No. 39, 1999 9017
Table 4. Effect of C-Terminal Amino Acids of Different Acyl
Segments 8a-d on the Oxaproline Ligation with Amine Segments
10a and 10b in 50% TFE and 50% Pyridine-Acetic Acid (1:1,
mol/mol) for 30 h
Table 5. Stereochemistry of Model Dipeptides Z-Xaa-ψPro-OMe
17a-e a
yield relative
producta
ligating segments (%)b
rate
LILNG-[OP]-FKI
LILNG-[OPMe]-FKI
DSFG-[OP]-FKI
DSFG-[OPMe]-FKI
LILA-[OP]-FKI
11a
11b
12a
12b
13a
13b
14a
8a + 10a
8a + 10b
8b + 10a
8b + 10b
8c + 10a
8c + 10b
8d + 10a
8d + 10b
80.4
76.2
83.7
81.1
66.9
65.3
54.8
51.9
1
0.95
1
0.95
0.8
LILA-[OPMe]-FKI
0.75
0.55
0.50
compd
structure
X
R1
R2
C2
cis/trans
IAYGGFL-[OP]-FKI
IAYGGFL-[OPMe]-FKI 14b
17a
17b
17c
17d
17e
Z-A-[OP]-OMe
Z-V-[OP]-OMe
Z-A-[OPMe]-OMe
Z-V-[OPMe]-OMe
Z-V-[SP]-OMe
O
O
O
O
S
Me
iPr
Me
iPr
iPr
H
H
Me
Me
H
R
R
R
R
R
68:32
56:44
54:46
43:57
40:60
a C-terminal amino acids of the acyl segments 8a-d are bold and
underlined; [OP] and [OPMe] see Table 3. b Yield was calculated from
HPLC.
a 15a, R1 ) Me; 15b, R1 ) iPr; 16a, X ) O, R2 ) H; 16b, X ) O,
R2 ) Me; 16c, X ) S, R2 ) H; [OP], [OPMe] and [SP] see Table 3.
4). We have found previously significant steric effect of the
C-terminal (CT) amino acids in thiaproline ligation that
undergoes a 5,5-bicyclic intermediate during O,N-acyl migra-
tion.34 The acyl segments 8a and 8b with nonhindered CT-Gly
gave higher yields than 8c with CT-Ala and 8d with CT-Leu.
Furthermore, kinetic study showed that the ligation rate of Ser-
OMe 16a with Z-Val-OCH2CHO 15b was found to be 6-fold
slower than that with Z-Ala-OCH2CHO 15a, and 10-fold slower
than that with Z-Gly-OCH2CHO 15c.
products from oxaproline ligation were obtained. Furthermore,
only the R epimer at C2 was found.
Racemization was a concern because of the prolonged period
for pseudoproline ligation. NR-(2,4-Dinitro-5-fluorophenyl)-L-
valinamide (Marfey’s reagent)52 was used for evaluating the
racemization because it can distinguish the D,L-diastereomers
in the selected pseudoproline ligation products 17a,c. HPLC
analysis showed that 17a gave 0.3% D-Ala and 3.9% D-Ser,
and 17c gave 0.4% D-Ala and 5.4% D-Thr. The control
experiments in which L-amino acids were subjected to acid-
hydrolytic conditions and derivatiztion by Marfey’s reagent
similar to 17a,c showed that Ala, Ser, and Thr gave 0.4, 5.3,
and 5.8% D-diastereomers, respectively. These results suggest
that oxaproline ligation in the pyridine-acetic acid mixture (1:
1, mol/mol) does not cause significant racemization.
Stereochemistry at C2 Carbon of ψPro Ring and cis-
trans Conformational Ratio of Xaa-ψPro Bond. To determine
the stereochemistry of the newly created C2-stereocenter of
oxaproline and the cis-trans conformational preference of the
oxaproline containing peptides, five dipeptides 17a-e were
examined by the 2D 1H NMR studies at 20 °C. For each sample,
two sets of independent NMR data were obtained based on the
assignment of the 2D (DQF) COSY, indicating only two cis-
trans isomers with the R epimer at the C2 carbon was found in
each compound. Table 5 summaries the C2 stereochemistry of
the oxaproline ring and the cis-trans ratios of the Xaa-ψPro
bonds in compounds 17a-e.
Product Characterization and the Absence of Ester
Intermediates in Oxaproline Ligation. A striking difference
was observed between the oxaproline and thiaproline ligations.
In thiaproline ligation, two stable R,S epimers of the thiazolidine
esters 4a were observed. The two esters of R,S epimers at the
C2 stereocenter then underwent O,N-acyl migration to form a
single amide product 5a (detected by HPLC). In oxaproline
ligation, no oxazolidine esters 4b,c as stable intermediates were
detected by HPLC in various pyridine-acetic acid ratios (1:0
to 0:1, mol/mol, data not shown). Chemical and spectroscopic
analyses were employed to determine the ligation products, since
the isomeric ester and amide were unable to be distinguished
by MS analysis.
Esters are generally susceptible to aminolysis by concentrated
hydroxyamine or basic hydrolysis, while amides are usually
stable under the same conditions. Treatment of oxaproline
ligation products 11a,b and 12a,b and thiaproline ligation
product 11c with 1 M hydroxylamine at pH 9 or 0.1 M LiOH
for 20 h failed to give the aminolysis or hydrolysis products.
In contrast, thiazolidine ester (4a in Figure 2) intermediate of
11c gave the expected aminolysis and hydrolysis products. More
importantly, no ester bond peak in the range 1720-1800 cm-1
was observed from the FT-IR spectra of the ψPro-containing
peptides 11a-c and 12a,b. In addition, treatment of Z-Ala-ψPro-
OMe 17a-e (see Table 5), which were ligation products of
15a,b and 16a-c, under different aqueous alkaline buffers at
pH 9.1 to 10.8 for 72 h, afforded only the hydrolyzed product
of the methyl ester. Finally, 17a-e were confirmed to be amide
The C2 stereochemistry of the oxaproline ring was determined
by the NOE cross-peak between the C2-H and C4-H of the
oxaproline ring.22,23,53-56 In 17a-e, a comparatively strong NOE
cross-peak between the C2-H and C4-H was observed in both
cis and trans isomers. In contrast, no such a cross-peak between
(52) (a) Marfey, P.; Ottesen, M. Carlsberg Res. Commun. 1984, 49, 585-
590. (b) Szo´ka´n, G.; Mezo¨, G.; Hudecz, F. J. Chromatogr. 1988, 444, 115-
122.
(53) Seebach, D.; Lamatsch, B.; Amstutz, R.; Beck, A. K.; Dobloer, M.;
Egli, M.; Fitzi, R.; Gautschi, M.; Herrado´n, B.; Hidber, P.; Irwin, J. J.;
Locher, R.; Maestro, M.; Maetzke, T.; Mourio`o, A.; Pfammatter, E.; Plattner,
D. A.; Schickli, C.; Schweizer, W. B.; Seiler, P.; Stucky, G. HelV. Chim.
Acta 1992, 75, 913-934.
(54) Szila´gyi, L.; Gyo¨rgydea´k, Z. J. Am. Chem. Soc. 1979, 101, 427-
432.
(55) Parthasarathy, R.; Korrytnyk, W. J. Am. Chem. Soc. 1976, 98, 6634-
6643.
(56) Wu¨rthrich, K. NMR of Proteins and Nucleic Acids; John Wiley and
Sons: New York, 1986; pp 117-125.
1
rather than ester products by H NMR studies. For example,
1
the 1D H NMR of the thiazolidine ester intermediate of 17e
showed a peak at 8.12 ppm indicating the azomethyl proton of
the open imine form resulting from the tautomerization of the
thiazolidine.46,48 None of the oxaproline ligation products 17a-d
or the amide bond product 17e of thiaproline ligation showed
1
any peak at the region 8.0-9.0 ppm in their 1D H NMR
spectra. Thus, these results confirmed that only the amide

9018 J. Am. Chem. Soc., Vol. 121, No. 39, 1999
Tam and Miao
Figure 4. Expanded regions of the 2D 1H NMR NOESY spectrum of 17d showing the assignment of C2 stereocenter and the cis and trans
isomers. The NOE cross-peaks are labeled with R, cis, and trans for the C2 R-epimer, cis and trans conformers, respectively. a and b in Figure 4A
indicate the NOE cross-peaks between C2 and C2-hydroxymethyl protons.
C4-H and C2-hydroxymethyl protons was found. These results
indicate that the C2 stereocenter is an R rather than an S epimer.
Other observable NOE cross-peaks also support that C2 centers
in 17a-e are R epimers. These include the NOE cross-peaks
between the C2-hydroxymethyl protons and the axial C5-H
but no NOE peaks between the C5-H and C2-H.
The cis-trans isomer ratios of the Xaa-ψPro bond was
measured by conventional NOE experiments.54,55 A typical
pattern56 of the RHi-C2-Hi+1 (i ) Xaa, i + 1 ) ψPro) NOE
cross-peak was observed in 17a-e (Figure 4). This pattern, i.e.,
the RHi-δHi+1 in a proline containing peptide, is characteristic
of the trans Xaa-ψPro bond.22-24,56 The cis isomer in each
compound is characterized by the NOE cross-peak between the
âHi and C4-Hi+1 (Figure 4).23,56 In 17b, 17d, and 17e, NOE
cross-peaks between C2-Hi+1 and âH-Val or CH3-Val were
also observed to confirm the assignment of cis isomers (Figure
Figure 5. Orthogonal ligation of Bac 7 analogues 20a, 20b, and 20c
at Gly24-Pro25 indicated by an arrow between acyl segment 18 and amine
segments 19a, 19b, and 19c.
4). In addition, the NOE cross-peaks between the RHi and C2i+1
-
in Figure 3. Three 35-amino acid-containing-segments with Ntn-
Ser 19a, -Thr 19b, and -Cys 19c were synthesized by tBoc/Bzl
chemistry. Peptide segments 18, 19a, 19b, and 19c were purified
by HPLC and identified by TOF-MALDI MS and amino acid
analysis. The purified peptide segments were readily soluble in
the pyridine-acetic acid mixture suggesting that this condition
is a suitable solvent. The oxaproline ligations were carried out
in pyridine-acetic acid mixtures (1:1, mol/mol), and in each
case, the amine segment 19a or 19b was used in slight excess.
The ligation reaction was relatively slow and required >35 h
for completion. However, HPLC monitoring showed that the
reaction proceeded cleanly and predominantly gave a single
product together with the unreacted starting materials and
hydrolyzed acyl segment 18 as observable byproducts (Figure
6A, B). The thiaproline ligations between 18 and 19c were
performed in both pyridine-acetic acid (1:1, mol/mol) and
aqueous conditions with the acyl segment 18 slight excess. In
contrast to oxaproline ligation, the HPLC patterns of thiaproline
ligations (Figure 6C, D) were complicated by the presence of
the R,S-ester intermediates 20d,e which converted to a single
amide product 20c after 20 h in aqueous buffer at pH 6.2 or
after 40 h in more acidic condition of pyridine-acetic acid (1:
1, mol/mol).
hydroxymethyl protons were observed in cis isomers. However,
the cis-trans ratios vary depending on the side chains around
the pseudoproline bonds.22-24 For example, Val at the i position
has an unfavorable effect on cis conformation because of the
steric hindrance between its large side chain and C4-Hi+1
.
Synthesis of Bactenecin 7 Analogues by Two-Segment
Ligation. Bac 7 is a highly cationic antimicrobial peptide
isolated from the large granules of bovine neutrophils.57 It
contains 59 amino acid residues with molecular weight about 7
kD, characterized by a high content of proline (47%) and
arginine (29%) and three tandem repeats of a tetradecamer.58
This exceptional Arg-rich peptide can also provide a test for
oxaproline ligation in pyridine-acetic acid because the guani-
dine side chain is susceptible to side reaction with aldehydes.
Furthermore, it is known that proline-rich sequences such as
Bac 7 are difficult to obtain by stepwise synthesis due to the
propensity of the proline residue to cleave off as a cyclic
dipeptide. To obtain synthetic analogues for studying the
mechanism of antimicrobial activity of Bac 7 which is not
known yet, we undertook the synthesis of Bac 7 analogues 20a,
20b, and 20c by a two-segment ligation strategy.
The ligation site was at Pro25 where the Gly-Pro sequence
offers the least steric hindrance (Figure 5). A 24-amino acid-
containing acyl segment 18 with a glycoaldehyde ester at the
C terminus was synthesized on an acetal resin 6 as described
The ligation sites of 20a-c were shown to be amide bond
due to the absence of the ester peaks at 1710-1780 cm-1 in
their FT-IR spectroscopies. They were further confirmed by the
treatment with 1.0 M H2NOH at pH 9.1 and 0.1 M LiOH for
10 h. Under these conditions, no hydrolysis or aminolysis
product was detected by HPLC. Furthermore, no evidence of
(57) Gennaro, P.; Skerlavaj, B.; Romeo, D. Infect. Immun. 1989, 57,
3142-3146.
(58) Frank, R. W.; Gennaro, R.; Schneider, K.; Przybylski, M.; Romeo,
D. J. Biol. Chem. 1990, 265 18871-18874.

Stereospecific Pseudoproline Ligation
J. Am. Chem. Soc., Vol. 121, No. 39, 1999 9019
Table 7. Comparison of Antimicrobial Activity of Bac 7 and
Three ψPro Analogues 20a-c from Pseudoproline Ligation
MIC (µM)a
peptide
E. coli
S. aureus
Bac 7
20a
20b
20c
0.24
0.31
0.23
0.27
0.23
0.29
0.22
0.28
a MIC, minimal inhibition concentration, obtained by radial diffusion
assay with underlay gel containing 1% agarose in 10 mM phosphate
buffer.
react selectively with aldehydes under very mild conditions to
afford oxazolidines or thiazolidines, a criterion important for
the orthogonal ligation of free peptide segments.34 This paper
confirms that the Ntn-Cys and Ntn-Ser/-Thr can be exploited
successfully for regiospecific oxaproline and thiaproline ligation
to afford monosubstituted ψPro at the ligation site using
unprotected peptides.
Although the thiaproline ligation proceeds smoothly in
aqueous solutions at pH 4 to 6, no reaction is observed for
oxazolidine ligation of Ntn-Ser or Ntn-Thr peptides under the
same conditions. In contrast, oxaproline ligation requires the
use of nearly anhydrous conditions.51 We have found that
pyridine-acetic acid mixtures for oxaproline ligation can
achieve reasonable rates, minimize side reactions and solubilize
free peptide segments.
Thiaproline and oxaproline ligations are semi-orthogonal, i.e.,
the former ligation of an Ntn-Cys peptide can be performed in
the presence of an Ntn-Ser or Ntn-Thr peptide. Such orthogo-
nality may be useful for assembling multiple free peptide
segments based on the regioselectivity of different Ntn-amino
acids without the need for a protection scheme.
Figure 6. HPLC profiles of oxaproline and thiaproline ligations. A
and B, oxaproline ligations of 18 with 19a and 19b in pyridine-acetic
acid (1:1, mol/mol) to form 20a and 20b. C, thiaproline ligation between
18 and 19c in pyridine-acetic acid (1:1, mol/mol). First, two intermedi-
ates as thiazolidine esters 20d,e were formed, then 20d,e transferred
to the amide bond product 20c under the same condition. D, thiaproline
ligation between 18 and 19c in aqueous buffers, first at pH 5.2 for 10
h to form thiazolidine esters 20d,e, and then the O,N-acyl transfer was
carried out at pH 6.6 for 20 h to form the amide bond product 20c. *,
impurities from pyridine. h, hydrolysis product of 18. i, possible imine
intermediate (3c in Figure 2).
Thiaproline ligation in pyridine-acetic acid conditions or
aqueous buffers proceeds in a two-step-reaction scheme with
the isolated R,S epimers of the thiazolidine esters as stable
intermediates (Figures 2 and 5). In the first step, imine capture
and ring-chain tautomerization are fast. The subsequent step of
O,N-acyl transfer of the thiazolidine esters to form the amide
SPro bonds is the slow step which can proceed smoothly under
acidic conditions because of the low pKa (6.24) of thiazolidine.41
This step is also pH dependent and is significantly accelerated
in more basic conditions as shown in the thiaproline ligation of
18 and 19c. However, in oxaproline ligation no oxazolidine ester
intermediate was detected by HPLC. A plausible explanation
is that the oxazolidine ester is unstable, >104 fold less stable
than its analogous thiazolidine ester,48 and transfers into the
stable amide OPro bond or equilibrates back to the starting
materials.
In pseudoproline ligation, the stereochemistry of the newly
created stereocenter at C2 of the OPro ring is a concern. Using
Z-Xaa-ψPro(R)-OMe 17a-e as models, only one epimer
(>95%) was detected by HPLC and NMR. In each compound,
the single C2 epimer of the oxaproline was assigned an R
configuration by 2D NOESY (Table 5). Thus, the oxaproline
ligation under our described conditions is not only regiospecific
but also stereospecific and does not produce diastereomers
during synthesis.
Table 6. Summary of the Synthesis of Bac 7 Analogues 20a, 20b,
and 20c
product
segments
ligation site [ψP]
[OP]25
yield (%)
20a
20b
20c
19a + 18
19b + 18
19c + 18
82.4
83.0
84.1
25
[OPMe
[SP]25
]
Arg modification by the glycoaldehyde as a branched peptide
was found. Both the oxaproline and thiaproline ligations gave
excellent yields ranging from 82 to 85% after 40 h (Table 6).
Furthermore, to examine the Pro surrogates derived from
pseudoproline ligation on the antimicrobial activity of Bac 7,
Bac 7 and three analogues 20a-c were assayed against a Gram
positive (Staphylococcus aureus) and a Gram negative (Es-
cherichia coli) species using the radial diffusion method.59 The
results (Table 7) showed very close activities between Bac 7
and its three analogues 20a-c. The detailed results of structure-
activity studies will be published elsewhere.
Discussion
Two plausible mechanisms can be advanced to account for
the stereospecific nature of the oxazolidine ligation. First, the
oxazolidine formation is stereospecific, which results in only
the C2-R epimer. This single epimer undergoes the O,N-acyl
migration to give diastereomeric selectivity to the ligation
reaction. The second mechanism is that the oxazolidine forma-
tion is not stereospecific but the O,N-acyl migration is ste-
The 1,2-amino alcohol of an N-terminal serine and threonine
as well as the 1,2-amino thiol of an N-terminal cysteine represent
unique combinations of dinucleophiles which are not found
elsewhere in a free peptide. Furthermore, they are known to
(59) Turner, J.; Cho, Y.; Dinh, N.; Waring, A. J.; Lehrer, R. I. Antimicrob.
Agents Chemother. 1998, 42 2206-2214.

9020 J. Am. Chem. Soc., Vol. 121, No. 39, 1999
Tam and Miao
reospecific because the C2-R epimer proceeds at a faster rate
than the C2-S epimer. The unreacted C2-S epimer undergoes
an equilibration to a mixture of R,S epimers through oxazolidine
ring-chain tautomerization. Our experiments as well as literature
precedents generally support the second proposed mechanism.
Experimental Section
General. 1D 1H NMR was obtained on Bruker AC 400 instrument.
Analytical HPLC was run on a Shimadzu 10A system using Vydac
C18 column (4.6 × 250 mm, 5 µm) with a flow rate of 1.0 mL/min,
monitored at 225 nm. Preparative HPLC was performed on a Waters
600 equipment with a Vydac C18 column (22 × 250 mm). All HPLC
was carried out with reversed phase linear gradient of buffer A, 0.05%
TFA in H2O, and buffer B, 60% CH3CN in H2O with 0.04% TFA.
Low-resolution mass spectra (LRMS) were obtained by MALDI-TOF
method on a PerSeptive Biosystems Voyager Elite 2 instrument and
high-resolution mass spectrometries (HRMS) were determined by ESI
Fourier transform ion cyclotron resonance method (ESI-FT-ICR)61 on
a PerSeptive Biosystems Mariner instrument. FT-IR spectra were
recorded on an ATI Mattson Genesis Series FTIR spectrometer with
the sample filmed on CaF2 surface.
Condensation reactions between a nonchiral aldehyde com-
ponent and chiral 1,2-amino alcohols or 1,2-amino thiols such
as Ser, Thr, or Cys generally produce a mixture of epimers at
the C2 atom in N-unsubstituted oxazolidine or thiazolidine
derivatives. However, upon acylation or methoxylcarbonylation,
a single C2 diastereomer cis to the C4-configuration of Ser,
Thr, or Cys has been found to be the only observed prod-
uct.54,55,60 It has been reported that Cys derivatives react with
benzaldehydes and p-tolualdehyde to produce a mixture of C2
thiazolidine epimers. Upon acetylation by acetic anhydride in
pyridine at 25 °C, however, only a single diastereomer is
obtained.55 To account for this observation, it is proposed that
the epimerization of the C2 atom occurs at the Schiff base
intermediate and the acetylation of the R-(C2)-epimer is fast
compared with the S-(C2) epimer. A similar mechanism for
stereospecific control can be invoked for both oxaproline and
thiaproline ligations, i.e., the first step of ester formation
produces a mixture of epimers at C2 and the O,N-acyl migration
of the R-epimer is fast compared with the S-epimer.
2D NMR Spectroscopy. Spectra were recorded at 500.13 MHz using
Bruker DRX 500 instrument. Chemical shifts were calibrated by the
residual CHCl3 signal as reference at 7.26 ppm. Phase sensitive NOESY
spectra used for resonance assignments were recorded using the TPPI
procedure with mixing time of 400 ms. In the d1 dimension, 1024 real
data points were collected using 16 acquisitions per FID with a 1.5 s
relaxation delay; 2048 real data points were obtained in the d2
dimension. Scalar connectivities were recorded from 2D double
quantum filtration (DQF) COSY experiments. Complete proton reso-
nance assignments were fulfilled with the (DQF) COSY experiments,
in some cases, with the aid of H-13C HSQC experiments. All NMR
data were transferred to Indigo II workstation and processed using
FELIX software.
1
The pseudoprolines obtained in our syntheses are monosub-
stituted with a hydroxymethyl side chain at the C2 carbon. The
C2-substitution enhances the cis isomer content of the Xaa-
ψPro bonds. Monosubstitution at C2 with a methyl group
increases the cis isomer to 38% compared with 18% obtained
from the unsubstituted ψPro surrogates derived from Cys or
Ser.23,24 However, substitution at the C5 decreases the cis
isomer,22 as shown in our study, from 68% in 17a to 60% in
17c and 56% in 17b to 43% in 17d. The trend showing that the
monosubstituted ψPro in our series ranges from 40% to 68%
is consistent with the literature.22-24 The slight increase in our
observed result could be due to the nature of the C2 substitution
which may form a hydrogen bond in the cis Xaa-ψPro isomer
in CDCl3 used in the NMR experiments.
Solid-Phase Peptide Synthesis. The amine peptide segments 10a-
l, 19a, 19b, and 19c were synthesized on resin using Boc/Bzl and
HBTU/HOBt strategy.62 The peptides were cleaved from the resin by
anhydrous HF-anisole (95:5, v/v), and then purified by HPLC. The
amino acid analysis and MS gave the desired results. 10a, Ser-Phe-
Lys-Ile-NH2, tR ) 11.2 min (20-60% B), HRMS m/e 493.3147 (M +
H+, C24H41N6O5 requires 493.3138); 10b, Thr-Phe-Lys-Ile-NH2, tR
)
13.4 min, HRMS m/z 507.3291 (M + H+, C25H43N6O5 requires
507.3295); 10c, Cys-Phe-Lys-Ile-NH2, tR ) 13.1 min, HRMS m/z
509.2923 (M + H+, C24H41N6O4S requires 509.2910); 10d, Trp-Phe-
Lys-Ile-NH2, tR ) 21.2 min, HRMS m/z 592.3628 (M + H+,
C32H46N7O4 requires 592.3611); 10e, His-Phe-Lys-Ile-NH2, tR ) 11.5
min, HRMS m/z 543.3429 (M + H+, C27H43N8O4 requires 543.3407);
10f, Asn-Phe-Lys-Ile-NH2, tR ) 11.3 min, HRMS m/z 520.3260 (M +
H+, C25H42N7O5 requires 520.3247); 10g, Lys-Phe-Lys-Ile-NH2, tR
)
In general, we have found that replacing Pro with SPro and
OPro based on ligation scheme does not result in significant
changes in biological activity, as demonstrated in the synthesis
of three ψPro analogues 20a-c which exhibit identical anti-
bacterial activities as the native peptide Bac7. Similar results
have been observed in the synthesis of HIV-1 protease SPro
analogues.40
11.1 min, HRMS m/z 534.3768 (M + H+, C27H48N7O4 requires
534.3774); 10h, Arg-Phe-Lys-Ile-NH2, tR ) 11.9 min, HRMS m/z
562.3822 (M + H+, C27H48N9O4 requires 562.3829); 10i, Ala-Phe-Lys-
Ile-NH2, tR ) 14.4 min, HRMS m/z 477.3191 (M + H+, C24H41N6O4
requires 477.3189); 10j, Met-Phe-Lys-Ile-NH2, tR ) 17.2 min, HRMS
m/z 537.3236 (M + H+, C26H45N6O4S requires 537.3223); 10k, Tyr-
Phe-Lys-Ile-NH2, tR ) 15.6 min, HRMS m/z 569.3463 (M + H+,
C30H45N6O5 requires 569.3451); 10l, Asp-Phe-Lys-Ile-NH2, tR ) 12.5
min, HRMS m/z 521.3070 (M + H+, C25H41N6O6 requires 521.3088);
19a, Ser-Arg-Pro-Ile-Pro-Arg-Pro-Leu-Pro-Phe-Pro-Arg-Pro-Gly-Pro-
Arg-Pro-Ile-Pro-Arg-Pro-Leu-Pro-Phe-Pro-Arg-Pro-Gly-Pro-Arg-Pro-
Ile-Pro-Arg-Pro-OH, tR ) 22.4 min, LRMS m/z 3980.71 (M + H+,
3980.83 calcd for C188H302N59O37); 19b, Thr-Arg-Pro-Ile-Pro-Arg-Pro-
Leu-Pro-Phe-Pro-Arg-Pro-Gly-Pro-Arg-Pro-Ile-Pro-Arg-Pro-Leu-Pro-
Phe-Pro-Arg-Pro-Gly-Pro-Arg-Pro-Ile-Pro-Arg-Pro-OH, tR ) 22.5 min,
LRMS m/z 3994.49 (M + H+, 3994.86 calcd for C189H304N59O37); 19c,
Cys-Arg-Pro-Ile-Pro-Arg-Pro-Leu-Pro-Phe-Pro-Arg-Pro-Gly-Pro-Arg-
Pro-Ile-Pro-Arg-Pro-Leu-Pro-Phe-Pro-Arg-Pro-Gly-Pro-Arg-Pro-Ile-
Pro-Arg-Pro-OH, tR ) 22.9 min, LRMS m/z 3995.27 (M + H+, 3996.90
calcd for C188H302N59O36S). Peptide glycoaldehydes 8a-d were syn-
thesized on the acetal resin 6 using Fmoc/tBu strategy (Figure 3).
Sequential coupling was accomplished by the HBTU/HOBt method62
with 2.5 equiv of amino acids, and the Fmoc group was deprotected
by 20% piperidine in DMF. Final cleavage of peptides from the resin
The C2-hydroxymethyl OPro and SPro derivatives are also
amino acid chimeras of proline and γ-hydroxyl amino acid.
These types of amino acids have recently been explored by
Marshall et al. in the synthesis of peptide hormone mimetics of
the “bioactive form” of angiotensin II.17 Furthermore, the use
of different acyl segments bearing other aldehydes and ketones
may yield other substituted ψPro at the ligation site to control
the cis-trans isomers. Although we have focused our attention
to peptide synthesis, various applications through non-peptides
or peptide mimetics suitable for combinatorial library syntheses
can also be envisioned. In conclusion, pseudoproline ligation
provides a convenient approach to the synthesis of a ψPro-
containing peptide using a 1,2-dinucleophile with a glycoalde-
hyde ester under mild conditions and holds promise for other
applications for the synthesis of peptide mimetics using the
orthogonal approach.
(61) Marsh, A. J.; Henrickson, C. L.; Jackson, G. S. Mass Spectrom.
ReV. 1998, 17, 1-35.
(62) Dourtoglou, V.; Gross, B. Synthesis 1984, 572-574.
(60) Seebach, D.; Aebi, J. D. Tetrahdron Lett. 1984, 25, 2545-2548.

Stereospecific Pseudoproline Ligation
J. Am. Chem. Soc., Vol. 121, No. 39, 1999 9021
was performed with TFA-glycerol-anisole-thioanisole (90:4:3:3, 40
mL/g resin) for 3 h. The residue was dissolved in buffer B and
lyophilized to dryness after removal of the TFA and resin. The produced
peptidyl glyceryl esters 7a-d were oxidized by sodium periodate (4-6
equiv) in aqueous buffers at pH 2 to 6 for 10 min. Preparative HPLC
gave the purified peptide glycoaldehydes 8a-d (50-70% yield) which
were characterized by amino acid analysis and MS. 8a, Leu-Ile-Leu-
Asn-Gly-OCH2CHO, tR ) 12.8 min (20-60% B), HRMS m/z 571.3466
(M + H+, C26H47N6O8 requires 571.3455); 8b, Asp-Ser-Phe-Gly-OCH2-
CHO, tR ) 12.6 min, HRMS m/z 467.1788 (M + H+, C20H27N4O9
requires 467.1778); 8c, Leu-Ile-Leu-Ala-OCH2CHO, tR ) 15.7 min,
HRMS m/z 471.3169 (M + H+, C23H43N4O6 requires 471.3182); 8d,
Ile-Ala-Tyr-Gly-Gly-Phe-Leu-OCH2CHO, tR ) 25.5 min, HRMS m/z
782.4091 (M + H+, C39H56N7O10 requires 782.4089). Similarly, acyl
segment 18, Arg-Arg-Ile-Arg-Pro-Arg-Pro-Pro-Arg-Leu-Pro-Arg-Pro-
Arg-Pro-Arg-Pro-Leu-Pro-Phe-Pro-Arg-Pro-Gly-OCH2CHO, was pre-
pared in 74.4% yields, tR ) 17.4 min, LRMS m/z 2980.18 (M + H+,
2980.60 calcd for C135H227N51O26).
Optimization of Ligation Condition. a. Pyridine-Acetic Acid
Ratios. A mixture of peptide glycoaldehyde 8a (7.2 mg, 12.7 µmol)
and peptide 10a (6.3 mg, 12.8 µmol) was divided into seven equal
portions. Pyridine-acetic acid (300 µL) was added to each part to bring
the final concentration to 6.0 × 10-3 M with pyridine-acetic acid molar
ratios of 1:0, 4:1, 2:1, 1:1, 2:3, 1:2, and 0:1, respectively. The reactions
were followed by HPLC.
b. Solvents. Peptide glycoaldehyde 8a (3.8 mg, 6 µmol) and peptide
10a (2.96 mg, 6.0 µmol) were dissolved in 600 µL of pyridine-acetic
acid (1:1, mol/mol), divided into four equal portions, and 150 µL of
TFE, DMF, H2O, and pyridine-acetic acid (1:1,mol/mol) were added
to each. The reaction was followed by HPLC, and the results were
shown in Table 1.
c. Bases. Peptide glycoaldehyde 8b (5.6 mg, 12µmol) and peptide
10a (5.9 mg, 12 µL) were dissolved in 600 µL of TFE and 120 µL of
acetic acid. After being divided into six equal portions, TFE and a
different base were added to bring the final concentration to 6 × 10-3
M with an acetic acid-base molar ratios of 1:1. The reaction was
followed by HPLC and summarized in Table 2.
Regiospecific Ligation. a. Competitive Reaction. Z-Alanine gly-
coaldehyde 15a (100 mg, 0.40 mmol) was dissolved in 3 mL of TFE
and 3 mL of pyridine-acetic acid (1:1, mol/mol), and then H-Ser-
OMe HCl 16a (65 mg, 0.42 mmol), H-Thr-OMe HCl 17b (70 mg,
0.41 mmol), and H-Cys-OMe HCl 16c (70 mg, 0.41 mmol) were added.
The reaction was monitored by HPLC. First, the ester bond product
Z-Ala-OCH2-Thz-COOMe (tR ) 27.7 min, 20-70% B, 30 min) formed
between 15a and 16c was observed. Then, the ester bond product
rearranged into the amide product Z-Ala-SPro-OMe 17f (tR ) 20.5 min,
20-70% B, 30 min) with half time t1/2 ) 17 h.
b. Regiospecificy of N-Terminal Amino Acid. Peptide glycoalde-
hyde 8a (8.6 mg, 15 µmol) was dissolved in 600 µL of acetic acid and
divided into 10 equal portions. To each portion, an N-terminal segment
(10a-l, 1.5 µmol) in 240 µL of pyridine was added. The mixture was
shaken for 45 h at room temperature, and the reaction was followed
by HPLC. The ligation products were confirmed by chemical analysis
and MS, and the yields were shown Table 3. 11a, tR ) 27.6 min (20-
50% B in 30 min, then 50-90% B in 15 min) LRMS m/z 1045.0 (M
+ H+, 1045.3 calcd for C50H84N12O12); 11b, tR ) 28.2 min, LRMS
m/z 1058.7 (M + H+, 1059.3 calcd for C51H86N12O12); 11c, tR ) 29.9
min, LRMS m/z 1060.8 (M + H+, 1061.4 calcd for C50H84N12O11S);
11d, tR ) 32.6, 33.6 min, LRMS m/z 1143.9 (M + H+, 1144.3 calcd
for C58H89N13O11); 11e, tR ) 23.3, 24,7, 33.6 min, LRMS m/z 1095.9
(M + H+, 1096.4 calcd for C53H87N14O11);. 11f, tR ) 27.5 min, LRMS
m/z 1071.8 (M + H+, 1072.3 calcd for C51H86N13O12);
Steric Effect of C-Terminal Amino Acids at the Acyl Segments.
Peptide glycoaldehyde 8a (0.95 mg, 1.5 µmol) and Ntn-serine peptide
10a (0.74 mg, 1.5 µmol) were dissolved in a solution of 150 µL TFE,
90µL of pyridine and 60µL of acetic acid. The ligation reaction was
carried out at room temperature for 45 h, and the process was monitored
by HPLC. Ligations between 8b-d and 10a,b were performed under
the same conditions. The results are shown in Table 4. The ligation
products were confirmed by chemical analysis and MS, and the yields
are shown Table 4. 11a, tR ) 27.6 min (20-50% B in 30 min, then
50-90% B in 15 min) LRMS m/z 1045.0 (M + H+, 1045.3 calcd for
C50H84N12O12); 11b, tR ) 28.2 min, LRMS m/z 1058.7 (M + H+, 1059.3
calcd for C51H86N12O12); 12a, tR ) 23.0 min, HRMS m/z 941.4758 (M
+ H+, C44H65N10O13 requires 941.4732); 12b, tR ) 23.2 min, HRMS
m/z 955.4878 (M + H+, C45H67N10O13 requires 955.4889); 13a, tR
)
25.9 min, HRMS m/z 945.6127 (M + H+, C47H81N10O10 requires
945.6136); 13b, tR ) 26.4 min, HRMS m/z 945.6127 (M + H+,
C47H81N10O10 requires 945.6136); 14a, tR ) 34.6 min, LRMS m/z 1256.0
(M + H+, 1256.5 calcd for C63H93N13O14); 14b, tR ) 35.8 min, LRMS
m/z 1270.0 (M + H+, 1270.5 calcd for C63H93N13O14).
Model Dipeptide Synthesis for Stereochemistry Study. Z-Alanine
glycoaldehyde (15a, 53 mg, 0.20 mmol) synthesized according to an
earlier described procedure51 was dissolved in 2 mL of TFE and 2 mL
of pyridine-acetic acid (1:1,mol/mol), and then H-Ser-OMe (16a, 65
mg, 0.42 mmol) was added. The mixture was stirred at room
temperature for 36 h. The product Z-Ala-OPro-OMe (17a) was
1
characterized by HPLC, MS, and H NMR. Similarly, Z-Val-OPro-
OMe (17b) from Z-Val-OCH2CHO 15b and 16a, Z-Ala-OProMe-OMe
(17c) from 15a and Thr-OMe 16b, Z-Val-OProMe-OMe (17d) from 15b
and 16b and Z-Val-SPro-OMe 17e from 10b and Cys-OMe 16c were
obtained.
17a: tR ) 17.4 min (20-70% B, 30 min), yield 89.3% based on
15a. HRMS m/z 367.1512 (M + H+, C17H23N2O7 requires 367.1505).
1H NMR (500 MHz, 298 K, CDCl3, 10 mg/mL):
C2, R; cis (68%, by NH-Ala): 7.31-7.36 (m, 5H, C6H5-Z), 5.60
(d, 1H, J ) 7.5, NH-Ala), 5.49 (m, 1H, H-C2), 5.02-5.14 (m, 2H,
CH2-Z), 4.52 (t, 1H, J ) 7, H-C4), 4.43 (t, 1H, J ) 7, axial H-C5),
4.33 (t, 1H, J ) 7, equatorial H-C5), 4.31 (m, 1H, RH-Ala), 3.86
(dd, 1H, J ) 3.5, J ) 9, CH2-C2), 3.82 (dd, 1H, J ) 3.5, J ) 9,
CH2-C2), 3.76 (s, 3H, OMe), 1.37 (d, 3H, J ) 7, CH3-Ala). NOE:
H(C4)-H(C2), R; H(C4)-×-CH2(C2), R; CH3(Ala)-H(C4), R-cis;
RH(Ala)-CH2(C2), R-cis; CH3(Ala)-H(C2), R-cis; axial H(C5)-
H(C2) and equatorial H(C5)-×- H(C2), one epimer.
C2, R; trans (32%): 7.31-7.36 (m, 5H, C6H5-Z), 5.49 (m, 1H,
H-C2), 5.39 (d, 1H, J ) 7.5, NH-Ala), 5.02-5.14 (m, 2H, CH2-Z),
4.68 (d, 1H, J ) 5.5, H-C4), 4.50 (m, 1H, RH-Ala), 4.38 (dd, 1H, J
) 7, J ) 9.5, axial H-C5), 4.15 (dd, 1H, J ) 7.5, J ) 9.5, equatorial
H-C5), 3.97 (dd, 1H, J ) 4, J ) 13, CH2-C2), 3.89 (dd, 1H, J )
4.5, J ) 13, CH2-C2), 3.75 (s, 3H, OMe), 1.41 (d, 3H, J ) 7, CH3-
Ala). NOE: H(C4)-H(C2), R; RH(Ala)-CH(C2), R-trans; axial
H(C5)-H(C2) and equatorial H(C5)-×-H(C2), one epimer.
17b: tR ) 17.7 min (35-80% B, 30 min), yield 77.3% based on
15b. HRMS m/z 395.1827 (M + H+, C19H27N2O7 requires 395.1818).
1H NMR (500 MHz, 298 K, CDCl3, 10 mg/mL):
C2, R, cis (56%, by CH3-Val): 7.36-7.29 (m, 5H, C6H5-Z), 5.55-
5.50 (m, 1H, NH-Val), 5.55-5.50 (m, 1H,H-C2), 5.01-5.12 (m, 2H,
CH2-Z), 4.43 (t, 1H, J ) 7, H-C4), 4.41 (d, 2H, J ) 7.5, H-C5),
4.28 (d, 2H, J ) 8.5, H-C5), 4.09 (dd, 1H, J ) 7, J ) 9, RH-Val),
4.01 (dd, 2H, J ) 3.5, J ) 13, CH2-C2), 3.90 (dd, 2H, J ) 3.5, J )
13, CH2-C2), 3.68 (s, 3H, OMe), 2.17-2.03 (m, 1H, âH-Val), 1.12
(d, 6H, J ) 7, 2CH3-Val), 1.02 (d, 6H, J ) 7, 2CH3-Val). NOE:
H(C4)-H(C2), R; âH(Val)-H(C4), R-cis; RH(Val)-CH2(C2), R-cis;
axial H(C5)-H(C2) and equatorial H(C5)-×-H(C2), one epimer;
C2, R; trans (44%): 7.36-7.29 (m, 5H, C6H5-Z), 5.55-5.50 (m,
1H, NH-Val), 5.55-5.50 (m, 1H, H-C2), 5.01-5.12 (m, 2H, CH2-
Z), 4.64 (t, 1H, J ) 9, H-C4), 4.62 (d, 2H, J ) 7.5, H-C5), 4.17 (t,
1H, J ) 7.5, RH-Val), 4.12 (dd, 2H, J ) 1.5, J ) 9, H-C5), 3.82
(dd, 2H, J ) 4, J ) 18, CH2-C2), 3.75 (s, 3H, OMe), 1.99-1.92 (m,
1H, âH-Val), 0.97 (d, 6H, J ) 6.5, 2CH3-Val), 0.93 (d, 6H, J ) 6.5,
2CH3-Val). NOE: H(C4)-H(C2), R; RH(Val)-H(C2), R-trans; CH3-
(Val)-axial H(C5), R-trans; axial H(C5)-H(C2) and equatorial
H(C5)-×-H(C2), one epimer.
17c: tR ) 21.9 min (20-70% B, 30 min), yield 78.4% based on
15a. HRMS m/z 381.1662 (M + H+, C18H25N2O7 requires 381.1652).
1H NMR (500 MHz, 298 K, CDCl3, 10 mg/mL):
C2, R, cis (60%, by NH-Ala): 7.33-7.31 (m, 5H, C6H5-Z), 5.57
(d, 1H, J ) 7 HZ, NH-Ala), 5.42 (s, 1H, H-C2), 4.53 (d, 1H, J ) 7
HZ, RH-Ala), 4.24 (d, 1H, J ) 11, HC4), 4.18-4.15 (m, 1H, H-C5),
4.11 (d, 2H, J ) 14, CH2-C2), 3.95 (d, 2H, J ) 13, CH2 -C2), 3.75
(d, 3H, J ) 7, OMe), 1.47-1.43 (m, 3H, CH3-C5), 1.37-1.34 (m,

9022 J. Am. Chem. Soc., Vol. 121, No. 39, 1999
Tam and Miao
3H, CH3-Ala). NOE: H(C4)-H(C2), R; H(C5)-×-H(C2), R; RH-
(Ala)-CH2(C2), R-cis.
(m, 1H, âH-Val), 0.95 (d, 6H, J ) 6.5, 2CH3-Val), 0.91 (d, 6H, J )
6.5, 2CH3-Val). NOE: H(C4)-H(C2), R; RH(Val)-H(C2), R-trans.
Evaluation of Racemization. The peptide samples 17a,c as well
as reference amino aicids, L-Ser, L-Thr, and L-Ala, were treated by the
following procedure: (1) hydrolysis in 6 M HCl aqueous solution at
110 °C for 24 h, (2) reaction of about 5 µmol of the sample with 200
µL of 1% NR-(2,4-dinitro-5-fluorophenyl)-L-valinamide in acetone and
40 µL of 1.0 M sodium bicarbonate in H2O at 40 °C for 1 h, (3) HPLC
analysis on a Vydac C18 column (4.6 × 250 mm) using a linear gradient
(35-65% B, in 40 min, at 340 nm), (4) comparison with the HPLC of
standard single D- or L-amino acid derivative to distinguish the D-,L-
diastereomers. The HPLC analysis of reference L-Ser gave 94.7%
L-diastereomer (tR ) 16.0 min) and 5.3% D-diastereomer (tR ) 16.8
min, racemization from hydrolysis). Similarly, the results of HPLC
showed that reference L-Thr: 94.2% L-diastereomer (tR ) 17.4 min)
and 5.8% D-diastereomer (tR ) 17.8 min, racemization from hydrolysis);
The reference L-Ala: 99.6% L-diastereomer (tR ) 22.5 min) and 0.4%
D- diastereomer (tR ) 24.9 min, racemization from hydrolysis); Z-Ala-
OPro-OMe 17a: 96.1% L- and 3.9% D-Ser diastereomers as well as
99.7% L- and 0.3% D-Ala diastereomers; Z-Ala-OProMe-OMe 17c:
94.6% L- and 5.4% D-Thr diastereomers as well as 99.6% L- and 0.4%
D-Ala diastereomers.
Synthesis of Bac 7 Analogues by Two-Segment Ligation. The Bac
7 analogues 20a, 20b, and 20c were synthesized by a two-segment
ligation strategy between acyl segment 18 and amine segments 19a-
c. 18, 18.0 mg in 3 mL pyridine-acetic acid mixture (1:1, mol/mol)
was divided into three equal portions, and then 19a (8.0 mg), 19b (8.1
mg), and 19c (8.2 mg) were added to each portion, respectively. The
well dissolved solutions stood at room temperature for 35 h, and the
reactions were monitored by HPLC (Figure 6, Table 6). The ligation
products were confirmed by chemical analysis and MS. 20a, tR ) 25.7
min (20-55% B in 30 min), LRMS m/z 6940.9 (M + H+, 6942.4 calcd
for C323H526N110O62); 20b, tR ) 25.8 min, LRMS m/z 6955.9 (M +
H+, 6956.4 calcd for C324H528N110O62); 20c, tR ) 26.3 min, LRMS m/z
6958.1 (M + H+, 6958.5 calcd for C323H526N110O62S).
C2, R; trans (40%): 7.33-7.31 (m, 5H, C6H5-Z), 5.64 (d, 1H, J )
8, NH-Ala), 5.28 (d, 1H, J ) 3, H-C2), 5.13-4.99 (m, 2H, CH2-Z),
4.50 (d, 1H, J ) 7, H-C4), 4.35-4.31 (m, 1H, H-C5), 4.23 (d, 1H,
J ) 13, RH-Ala), 4.05 (d, 2H, J ) 13, CH2-C2), 3.86 (d, 2H, J )
13, CH2-C2), 3.77 (d, 3H, J ) 5.5, OMe), 1.47-1.43 (m, 3H, CH3-
C5), 1.37-1.34 (m, 3H, CH3-Ala). NOE: H(C4)-H(C2), R; H(C5)-
×-H(C2), R; CH3(Ala)-CH2(C2) R-trans; RH(Ala)-H(C2), R-trans.
17d: tR ) 21.9 min (35-80% B, 30 min), yield 73.3% based on
15b. HRMS m/z 409.1959 (M + H+, C20H29N2O7 requires 409.1975).
1H NMR (500 MHz, 298 K, CDCl3, 10 mg/mL):
C2, R; cis (43%, by CH3-Thr): 7.34-7.29 (m, 5H, C6H5-Z), 5.52
(d, 1H, J ) 9, NH-Val), 5.46 (s, 1H, H-C2), 5.12-5.00 (m, 2H,
CH2-Z), 4.25 (d, 1H, J ) 8.5, H-C4), 4.24 (d, 1H, J ) 7, RH-Val),
4.22 (m, 1H, H-C5), 4.08 (dd, 2H, J ) 2.5, J ) 15, CH2-C2), 3.87
(dd, 2H, J ) 2.5, J ) 15, CH2 -C2), 3.75 (s, 3H, OMe), 2.05-1.98
(m, 1H, âH-Val), 1.42 (d, 3H, J ) 6, CH3-C5), 1.05 (d, 6H, J ) 7,
2, CH3-Val), 0.99 (d, 6H, J ) 6.5, 2CH3-Val). NOE: H(C4)-H(C2),
R; H(C5)-CH2(C2), R; RH(Val)-H(C2), R-trans.
C2, R, trans (57%): 7.34-7.29 (m, 5H, C6H5-Z), 5.52 (d, 1H, J )
9, NH-Val), 5.29 (d, 1H, J ) 5, H-C2), 5.12-5.00 (m, 2H, CH2-Z),
4.16 (d, 1H, J ) 9, H-C4), 4.17-4.14 (m, 1H, H-C5), 3.98 (d, 1H,
J ) 8, RH-Val), 3.95 (d, 2H, J ) 13, CH2-C2), 3.77 (dd, 2H, J ) 5,
J ) 17, CH2 -C2), 3.71 (s, 3H, OMe), 1.98-1.92 (m, 1H, âH-Val),
1.46 (d, 3H, J ) 6, CH3-C5), 0.98 (d, 6H, J ) 7, 2CH3-Val), 0.92
(d, 6H, J ) 7, 2CH3-Val). NOE: H(C4)-H(C2), R; H(C5)-CH2(C2),
R; âH(Val)-H(C4), R-cis; CH3(Val)-H(C4), R-cis.
17e: tR ) 23.4 min (35-80% B, 30 min), yield 83.6% based on
15b. HRMS m/z 411.1599 (M + H+, C19H27N2O6S requires 411.1590).
1H NMR (500 MHz, 298 K, CDCl3, 10 mg/mL):
C2, R, cis (60%, by CH3-Val): 7.35-7.30 (m, 5H, C6H5-Z), 5.55
(d, 1H, J ) 8.5, NH-Val), 5.42 (t, 1H, J ) 9.5, HO-CH2), 5.19 (t,
1H, J ) 12, H-C2), 5.12-5.01 (m, 2H, CH2-Z), 4.59 (dd, 1H, J )
6.5, J ) 15, RH-Val), 3.90 (dd, 2H, J ) 6, J ) 18, CH2-C2), 3.73
(s, 3H, OMe), 3.49 (m, 1H, H-C4), 3.08 (d, 2H, J ) 13, H-C5),
2.14-2.07 (m, 1H, âH-Val), 1.27 (d, 6H, J ) 7, 2CH3-Val), 0.96
(d, 6H, J ) 7, 2CH3-Val). NOE: âH(Val)-H(C2), R-cis; CH3(Val)-
H(C2), R-cis CH3(Val)-H(C4), R-cis.
C2, R; trans (40%): 7.35-7.30 (m, 5H, C6H5-Z), 5.62 (d, 1H, J )
9, NH-Val), 5.42 (t, 1H, J ) 9.5, HO-CH2), 5.12-5.01 (m, 2H, CH2-
Z), 4.92 (dd, 1H, J ) 6, J ) 21, H-C2), 4.16 (dd, 1H, J ) 6, J )
15.5, RH-Val), 3.77 (dd, 2H, J ) 6, J ) 17.5, CH2 -C2), 3.65 (s,
3H, OCH3), 3.49 (m, 1H, H-C4), 3.34 (d, 2H, H-C5), 1.95-1.89
Acknowledgment. We thank Dr. Marcus Voehler and Mr.
Zhijun Li for help with 2D NMR experiments, Dr. Leonardo
Di Donna for help with HRMS determination, Dr. Jin-Long
Yang for assistance in antibacterial assays. This work was in
part supported by US Public Health Service NIH Grant
CA36544 and GM57145.
JA991153T