Sortase A-catalyzed transpeptidation of glycosylphosphatidylinositol derivatives for chemoenzymatic synthesis of GPI-anchored proteins

Article abstract of DOI:10.1021/ja906611x

Several peptides/small proteins and glycosylphosphatidylinositol (GPI) derivatives were synthesized and compared as substrates of sortase A (SrtA), a bacterial transpeptidase, for enzymatic coupling. It was observed that peptides containing the LPKTGGS and LPKTGGRS sequences as sorting signals at the peptide C-terminus were effectively coupled to GPI derivatives having one or two glycine residues attached to the phosphoethanolamine group at the nonreducing end. This reaction was employed to prepare several analogues of the human CD52 and CD24 antigens, which are naturally GPI-anchored glycopeptides/ glycoproteins. It was further observed that the trisaccharide GPI analogues 5 and 6 were better SrtA substrates than monosaccharide GPI analogue 4, suggesting that steric hindrance of the GPI analogues does not affect their peptidation reaction mediated by SrtA. Therefore, this synthetic strategy may be useful for the preparation of more complex GPI-anchored peptides, glycopeptides, proteins, and glycoproteins.

Full text of DOI:10.1021/ja906611x

Published on Web 01/15/2010
Sortase A-Catalyzed Transpeptidation of
Glycosylphosphatidylinositol Derivatives for Chemoenzymatic
Synthesis of GPI-Anchored Proteins
Zhimeng Wu, Xueqing Guo, Qianli Wang, Benjamin M. Swarts, and Zhongwu Guo*
Department of Chemistry, Wayne State UniVersity, 5101 Cass AVenue, Detroit, Michigan 48202
Received August 11, 2009; E-mail: zwguo@chem.wayne.edu
Abstract: Several peptides/small proteins and glycosylphosphatidylinositol (GPI) derivatives were synthe-
sized and compared as substrates of sortase A (SrtA), a bacterial transpeptidase, for enzymatic coupling.
It was observed that peptides containing the LPKTGGS and LPKTGGRS sequences as sorting signals at
the peptide C-terminus were effectively coupled to GPI derivatives having one or two glycine residues
attached to the phosphoethanolamine group at the nonreducing end. This reaction was employed to prepare
several analogues of the human CD52 and CD24 antigens, which are naturally GPI-anchored glycopeptides/
glycoproteins. It was further observed that the trisaccharide GPI analogues 5 and 6 were better SrtA
substrates than monosaccharide GPI analogue 4, suggesting that steric hindrance of the GPI analogues
does not affect their peptidation reaction mediated by SrtA. Therefore, this synthetic strategy may be useful
for the preparation of more complex GPI-anchored peptides, glycopeptides, proteins, and glycoproteins.
Introduction
forms and sufficient quantities, which is currently difficult to
achieve by either biological or chemical means. Therefore, a
Many surface proteins and glycoproteins are associated with
the cell membrane through glycosylphosphatidylinositol (GPI)
anchoring.1-3 GPI-anchored proteins and glycoproteins play a
pivotal role in numerous biological events.4-8 For example, the
human CD52 and CD24 antigens (Figure 1) are representative
GPI-anchored glycopeptide/glycoprotein antigens.9-11 The CD52
antigen is ubiquitously expressed by human lymphocyte and
sperm cells9,10 and has been demonstrated to play an important
role in the human immune system and the human reproductive
process. The CD24 antigen is expressed by T cells12 and
neurons13,14 and overexpressed on a number of carcinoma cell
lines.15,16 To study the functions of GPI-anchored proteins and
glycoproteins at the molecular level, it is necessary to have
access to these molecules and their derivatives in homogeneous
feasible method for synthesizing GPI-linked peptides/proteins
and glycopeptides/glycoproteins is highly desirable.17-21
Every naturally GPI-anchored protein or glycoprotein has its
polypeptide C-terminus linked to the phosphoethanolamine
group at the nonreducing end of the GPI core glycan (Figure
1). Accordingly, a general method for the preparation of various
natively linked GPI-protein/glycoprotein conjugates would be
feasible once a reaction for site-specific attachment of peptides/
proteins to the phosphoethanolamine group of GPIs is estab-
lished. For this purpose, an enzyme-catalyzed ligation method
is particularly appealing, not only because enzymatic reactions
are usually highly specific and efficient but also because
enzymes are especially suitable for complex structures such as
large proteins and intricate GPIs. For this purpose, we became
interested in sortases, a class of transpeptidases found in Gram-
positive bacteria.22
(1) Ferguson, M. A. J.; Williams, A. F. Annu. ReV. Biochem. 1988, 57,
285–320.
(2) Englund, P. T. Annu. ReV. Biochem. 1993, 62, 121–138.
(3) Hwa, K. AdV. Exp. Med. Biol. 2001, 491, 207–214.
(4) Robinson, P. J. Cell Biol. Int. Rep. 1991, 15, 761–767.
(5) Ferguson, M. A. J. Parasitol. Today 1994, 10, 48–52.
(6) Gaulton, G. N.; Pratt, J. C. Semin. Immunol. 1994, 6, 97–104.
(7) Kasahara, K.; Sanai, Y. Glycoconjugate J. 2000, 17, 153–162.
(8) Loertscher, R.; Lavery, P. Transplant Immunol. 2002, 9, 93–96.
(9) Treumann, A.; Lifely, M. R.; Schneider, P.; Ferguson, M. A. J. J. Biol.
Chem. 1995, 270, 6088–6099.
Results and Discussion
Sortases catalyze transpeptidation reactions on the bacterial
cell wall, known as the “cell-wall sorting reactions”, which
anchor surface proteins onto bacterial cell walls.22 The reaction
mechanisms of sortase-catalyzed transpeptidation have been
established.23 For instance, Staphylococcus aureus sortase A
(10) Schro¨ter, S.; Derr, P.; Conradt, H. S.; Nimtz, M.; Hale, G.; Kirchhoff,
C. J. Biol. Chem. 1999, 274, 29862–29873.
(11) Kay, R.; Rosten., P. M.; Humphries, R. K. J. Immunol. 1991, 147,
1412–1416.
(17) Tanaka, Y.; Nakahara, Y.; Hojo, H.; Nakahara, Y. Tetrahedron 2003,
59, 4059–4067.
(12) Williams, L. A.; Hock, B. D.; Hart, D. N. J. Blood 1998, 88, 3048–
3055.
(18) Xue, J.; Shao, N.; Guo, Z. J. Org. Chem. 2003, 68, 4020–4029.
(19) Shao, N.; Xue, J.; Guo, Z. Angew. Chem., Int. Ed. 2004, 43, 1569–
1573.
(13) Kadmon, G.; Kowitz, A.; Altevogt, P.; Schachner, M. J. Cell Biol.
1992, 110, 209–218.
(20) Paulick, M. G.; Wise, A. R.; Forstner, M. B.; Groves, J. T.; Bertozzi,
C. R. J. Am. Chem. Soc. 2007, 129, 11543–11550.
(21) Becker, C. F. W.; Liu, X.; Olschewski, D.; Castelli, R.; Seidel, R.;
Seeberger, P. H. Angew. Chem., Int. Ed. 2008, 47, 8215–8219.
(22) Marraffini, L. A.; DeDent, A. C.; Schneewind, O. Microbiol. Mol.
Biol. ReV. 2006, 70, 192–221.
(14) Rougon, G.; Alterman, L. A.; Dennis, K.; Guo, X. J.; Kinnon, C. Eur.
J. Immunol. 1991, 21, 1397–1402.
(15) Jackson, D.; Waibel, R.; Weber, E.; Bell, J.; Stahel, R. A. Cancer
Res. 1992, 52, 5264–5270.
(16) Lim, S.; Oh, S. Pathol. Res. Pract. 2005, 201, 479–486.
9
10.1021/ja906611x  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 1567–1571 1567

A R T I C L E S
Wu et al.
Figure 1. Generic structure of the human CD52 and CD24 antigens.
Figure 2. Anchorage of surface proteins to cell-wall proteoglycans by SrtA in S. aureus. SrtA recognizes LPXTG near the C-terminus of a target protein.
The Cys184 of SrtA performs a nucleophilic attack at the peptide bond between T and G in LPXTG, resulting in a thioester intermediate with the carboxyl
group of the C-terminal T linked to Cys184. This reactive intermediate reacts with the cross-bridge N-terminus of a cell-wall proteoglycan to anchor the target
protein to the cell-wall peptidoglycan.
Scheme 1. Synthesis of GPI-Anchored Proteins via SrtA-Catalyzed
(SrtA) was found to recognize a pentapeptide LPXTG (known
Transpeptidation of GPIs
as the sorting signal) near the substrate protein C-terminus, break
the peptide bond between T and G in the sorting signal to form
a reactive SrtA-substrate protein complex, and then link the
substrate protein to cell-wall peptidoglycans (Figure 2).22
Specifically, the thiol group of Cys184 at the active center of
SrtA reacts with LPXTG to form a thioester,24 which then
transfers the acyl group to the N-terminus of the cross-bridge
of the cell-wall proteoglycan.23
SrtA has been shown to accept various peptide/protein
substrates, so long as they bear the sorting signal LPXTG, and
a range of amino nucleophiles. SrtA has thus been utilized as a
powerful tool25,26 for ligating peptides and proteins;27-30 linking
peptides/proteins to nucleic acids,31 lipids,32 and carbohy-
drates;33 and labeling proteins on living cell surfaces.34-36 On
the basis of these findings, we envisioned that SrtA might be
used to couple proteins and glycoproteins to GPIs for the
synthesis of GPI conjugates, as outlined in Scheme 1, provided
that the substrate protein 1 contains a sorting signal at its
C-terminus and that GPI 2 has a Gly residue linked to the
(23) Huang, X.; Aulabaugh, A.; Ding, W.; Kapoor, B.; Alksne, L.; Tabei,
K.; Ellestad, G. Biochemistry 2003, 42, 11307–11315.
(24) Ton-That, H.; Mazmanian, S. K.; Alksne, L.; Schneewind, O. J. Biol.
Chem. 2002, 277, 7447–7452.
(25) Budisa, N. ChemBioChem 2004, 5, 1176–1179.
(26) Tsukiji, S.; Nagamune, T. ChemBioChem 2009, 10, 787–798.
(27) Mao, H.; Hart, S. A.; Schink, A.; Pollok, B. A. J. Am. Chem. Soc.
2004, 126, 2670–2671.
(28) Parthasarathy, R.; Subramanian, S.; Boder, E. T. Bioconjugate Chem.
2007, 18, 469–476.
(32) Antos, J. M.; Miller, G. M.; Grotenbreg, G. M.; Ploegh, H. L. J. Am.
Chem. Soc. 2008, 130, 16338–16343.
(29) Pritz, S.; Kraetke, O.; Klose, A.; Klose, J.; Rothemund, S.; Fechner,
K.; Bienert, M.; Beyermann, M. Angew. Chem., Int. Ed. 2008, 47,
3642–3645.
(33) Samantaray, S.; Marathe, U.; Dasgupta, S.; Nandicoori, V. K.; Roy,
R. P. J. Am. Chem. Soc. 2008, 130, 2132–2133.
(34) Popp, M. W.; Antos, J. M.; Grotenbreg, G. M.; Spooner, E.; Ploegh,
H. L. Nat. Chem. Biol. 2007, 3, 707–708.
(30) Kobashigawa, Y.; Kumeta, H.; Ogura, K.; Inagaki, F. J. Biomol. NMR
2009, 43, 145–150.
(35) Tanaka, T.; Yamamoto, T.; Tsukiji, S.; Nagamune, T. ChemBioChem
2008, 9, 802–807.
(31) Pritz, S.; Wolf, Y.; Kraetke, O.; Klose, J.; Bienert, M.; Beyermann,
M. J. Org. Chem. 2007, 72, 3909–3912.
(36) Yamamoto, T.; Nagamune, T. Chem. Commun. 2009, 1022–1023.
9
1568 J. AM. CHEM. SOC. VOL. 132, NO. 5, 2010

Sortase A-Catalyzed Peptide-GPI Ligation
A R T I C L E S
Scheme 2. SrtA-Mediated Chemoenzymatic Synthesis of CD52
Analogues
Scheme 3. Synthesis of GPI Analogues 5 and 6
Nevertheless, 4 was used as a positive control to evaluate SrtA-
catalyzed ligation of more complex structures. GPI analogues
5 and 6 containing the GPI core trisaccharide are excellent
mimics of natural GPI anchors. Moreover, because 5 and 6 have
one and two Gly residues, respectively, linked to the phospho-
ethanolamine group at the carbohydrate nonreducing end,
comparison of these analogues would be able to demonstrate
whether the number of Gly residues in the GPI substrate has a
major impact on its reactivity as a peptide acceptor. Peptide 7
differs from peptides 8 and 9 mainly in regard to the position
of the sorting signal: the sorting signal in 7 is exposed at the
C-terminus, while it is only close to the C-terminus in 8 and 9.
According to Pritz,31 the C-terminal sequence may affect the
reaction rate, efficiency, and equilibrium. The structural differ-
ences between 8 and 9 are that 8 has an acetylated N-terminus
and that 9 has an additional amino acid at the C-terminus.
Analysis of the reactions of peptides 7-9 would show how the
different peptide structures might affect the SrtA-catalyzed
coupling reactions between peptides and GPIs.
phosphoethanolamine moiety. It is worth noting that the resulting
GPI-anchored protein/glycoprotein 3 would differ slightly from
the natural products as a result of the insertion of a pentapeptide
sequence between the GPI and the native protein. How this
minor structural change may affect the biological functions of
GPI-anchored proteins and glycoproteins is an interesting
question to study. We also expected this synthetic method to
be applicable as well to the coupling of GPIs to proteins and
glycoproteins that are not GPI-anchored in nature. These non-
natural products would be useful for various biological studies
and the investigation of GPI-directed targeting of molecules to
specific cell or tissue types.37,38
The above synthetic strategy was explored in the preparation
of analogues of the human CD52 and CD24 antigens. Initially,
our investigation was focused on the coupling reaction of GPI
analogues 4, 5, and 6 with CD52 peptides 7, 8, and 9 (Scheme
2). Our previous studies revealed that SrtA accepts 4 as a
nucleophilic peptide acceptor,39 though only a monosaccharide
GPI analogue and a very short peptide were involved therein.
The peptide substrates 7, 8, and 9, which contain both the
CD52 peptide sequence and the sorting signal LPKTG, were
prepared with an automatic peptide synthesizer employing the
conventional Fmoc-amino acid chemistry. GPI analogue 4 was
previously prepared in our laboratory,39 while GPI analogues 5
and 6 were synthesized according to the procedures shown in
Scheme 3. First, trisaccharide 1618 was converted into Schmidt
glycosyl donor 18 in two steps, including oxidative hydrolysis
of the anomeric center at the reducing end of 16 and trichlo-
roacetimidation of hemiacetal 17 by trichloroacetonitrile in the
presence of DBU. The reaction of 18 with 2-phenylethanol
promoted by TMSOTf afforded an inseparable mixture of R
and ꢀ anomers (3:1) that were separable by column chroma-
tography after deacetylation with NaOMe/MeOH, eventually
yielding 20. The stereochemistry of 20 was confirmed by the
1JC,H coupling constants of the anomeric carbons. Phosphory-
lation of 20 was achieved by the H-phosphonate method using
21 and 22 as the phosphorylating reagents. The condensation
reactions of 20 with 21 and 22 went smoothly in the presence
of pivaloyl chloride and were followed by in situ oxidation of
the resultant H-phosphonates using iodine. Finally, the benzyl
and Cbz groups were simultaneously removed with 10% Pd/C
(37) Ilangumaran, S.; He, H.-T.; Hoessli, D. C. Immunol. Today 2000, 21,
2–7.
(38) Trotter, J.; Klein, C.; Kramer, E. Neuroscientist 2000, 6, 271–284.
(39) Guo, X.; Wang, Q.; Swarts, B. M.; Guo, Z. J. Am. Chem. Soc. 2009,
131, 9878–9879.
9
J. AM. CHEM. SOC. VOL. 132, NO. 5, 2010 1569

A R T I C L E S
Wu et al.
Figure 3. HPLC diagrams of SrtA-mediated coupling reactions of 8 with (a) 4, (b) 5, and (c) 6. HPLC conditions: C18 column; gradient eluent, 10 to 60%
CH3CN/H2O; monitoring at 220 nm.
under a hydrogen atmosphere to give 5 and 6, respectively,
which were ready for enzymatic ligation with peptides.
The enzymatic reactions outlined in Scheme 2 were performed
in 0.3 M Tris-HCl buffer (pH ∼7.5) containing 0.15 M NaCl,
5 mM CaCl2, and 0.2 mM mercaptoethanol. The peptide, GPI,
and SrtA concentrations were set at 0.5 mM, 2.5 mM, and 12.6
µM, respectively. The reaction mixtures were incubated at 37
°C, and the reaction progress was monitored by HPLC and
MALDI-TOF mass spectrometry (MS). After 24 h of incuba-
tion, the reactions were quenched with 0.1% trifluoroacetic acid
(TFA), and the reaction yields were eventually determined using
HPLC. It was observed that the reactions of 7 with 4 and 5
were very slow, and after 24 h of incubation, the corresponding
desired products 10 and 12 were formed in only ∼10% yield
while the majority of 7 was unaffected. These results suggested
that 7 was probably not a good SrtA substrate. On the contrary,
the reactions of 8 with 4 and 5 were very fast, offering 10 and
12 in 61 and 78% yield, respectively (Figure 3). Under the same
conditions, the reaction of 8 and 6, which contained two Gly
residues, was also fast and gave conjugate product 14 in an
82% yield. Furthermore, the reactions of 9 with 4, 5, and 6
were similar to the reactions of 8, affording 11, 13, and 15 in
∼65, 81, and 78% yield, respectively. Large-scale reactions of
8 with 4 and 5 were also performed. In this case, the ligation
products 10 and 12 were purified by HPLC to obtain the samples
in milligram quantities that were sufficient to perform NMR
characterization in addition to MS analysis.
monosaccharide GPI analogue 4. The results suggest that SrtA
may accept more complex GPI analogues. In addition, 5 and 6,
which contain one and two Gly residues, respectively, were
similar as peptide acceptors of the SrtA-catalyzed transpepti-
dation reactions. These studies further demonstrated that at-
taching the sorting signal directly to the peptidic C-terminus,
such as in the case of 7, does not give the best SrtA substrate.
For efficient recognition by and reaction with SrtA, the sorting
signal should be close to the C-terminus but not exposed. The
observation that peptide substrates with either a dipeptide (such
as KS39 or GS) or a tripeptide (GRS) at the C-terminus gave
almost identical results suggests that beyond the sorting signal,
SrtA may require only a dipeptide to be present at the C-terminus
and that the structure of this dipeptide can be variable.
The reactions of 8 with 4 and 5 were analyzed in detail to
further understand the reactivity of SrtA. In these studies, the
two reactions were carried out under the same conditions, and
reaction aliquots were obtained at various time points, quenched
with 0.1% TFA, and then analyzed by HPLC. As shown in
Figure 4, both reactions were fast and reached equilibrium within
6 h. Evidently, the reaction of 5 was faster than that of 4. After
30 min of incubation, 5 gave 60% conversion, as opposed to
only 30% for 4. To further examine the kinetics of these
reactions, we derived the Km and Vmax values of 4 and 5 by
fitting their initial reaction rates obtained at various concentra-
tions (0.625, 1.25, 2.5, 5.0, 10.0 mM) [with fixed concentrations
of 8 (0.5 mM) and SrtA (12.6 µM)] to the Michaelis-Menten
equation using Lineweaver-Burk plots. The Vmax values of 4
and 5 were 8.4612 and 12.9761 µM/sec, respectively, and the
corresponding Km values were 2.0928 and 0.8784 µM. These
results clearly indicate that the reaction of 5 is faster and more
These studies revealed that SrtA accepts both monosaccharide
and trisaccharide GPI analogues as the peptide acceptor in
transpeptidation reactions, and more interestingly, trisaccharide
analogues 5 and 6 seemed to be better substrates than the
9
1570 J. AM. CHEM. SOC. VOL. 132, NO. 5, 2010

Sortase A-Catalyzed Peptide-GPI Ligation
A R T I C L E S
Figure 4. Kinetics analysis of SrtA-catalyzed reactions of 8 with 4 and 5. (a) Results of the reaction between 4 (2.5 mM) and 8 (0.5 mM). (b) Results of
reaction between 5 (2.5 mM) and 8 (0.5 mM). (c) Experiments to determine Km and Vmax for 4 at a fixed concentration of 8 (0.5 mM). (d) Experiments to
determine Km and Vmax for 5 at a fixed concentration of 8 (0.5 mM).
Scheme 4. SrtA-Mediated Chemoenzymatic Synthesis of CD24
Analogue 26
MS and HPLC. We found that the reaction afforded the desired
GPI-protein conjugate 26 [MALDI-TOF MS (m/z): calcd for [M
+ H]+, 4210.2; found, 4210.6] in 72.9% yield according to HPLC
assays. This result demonstrates the potential of SrtA to be used
for the synthesis of more complex GPI-anchored peptides/proteins.
Conclusion
In summary, the synthesis of several analogues of the human
CD52 and CD24 antigens through SrtA-catalyzed ligation
between peptides and GPI analogues has been described. It was
observed that peptides with different lengths and sequences had
very similar reactivities. On the other hand, the more sterically
demanding trisaccharide GPI analogues 5 and 6 were better SrtA
substrates than the monosaccharide GPI analogue 4. On the basis
of the present results, we are optimistic that the strategy may
be applicable to more complex peptides and GPIs and that it
may be developed into a general and practical method for the
synthesis of GPI-anchored peptides, glycopeptides, proteins, and
glycoproteins. The ligation of intact GPIs and full-size proteins/
glycoproteins is now under investigation.
efficient than that of 4. Therefore, the increased steric hindrance
of trisaccharide 5 in comparison to 4 did not seem to have any
impact on its reactivity with the acyl-enzyme complex. On the
contrary, 5 and 6 may fit the catalytic site of SrtA better than
4. To prove this postulation and address its rationale, additional
structure-reactivity relationship studies are necessary. Never-
theless, these observations highlight the potential of accomplish-
ing site-specific ligation between peptides and more complex
GPI analogues via SrtA-catalyzed transpeptidation.
To further investigate whether the chemoenzymatic synthesis
strategy is applicable to more complex peptides and proteins,
we then used solid-phase peptide synthesis to prepare a small
protein 25 containing the full sequence of the CD24 antigen
and the sorting signal and evaluated its conjugation with GPI
analogue 5 in the presence of SrtA (Scheme 4). The coupling
reaction was carried out under the same conditions as described
above, while the reaction progress was monitored by MALDI-TOF
Acknowledgment. This work was supported by Wayne State
University and in part by the NSF (CHE-0715275).
Supporting Information Available: Experimental details for
the preparation of SrtA and compounds 5-9 and 25; conditions
and protocols for the enzymatic reactions of 4-6 with 7-9 and
of 5 with 25 to prepare 10-15 and 26, respectively; HPLC
results for the enzymatic reactions; and MS and NMR spectra
of all intermediates and products. This material is available free
of charge via the Internet at http://pubs.acs.org.
JA906611X
9
J. AM. CHEM. SOC. VOL. 132, NO. 5, 2010 1571