Efficient synthesis and applications of peptides containing adenylylated tyrosine residues

Article abstract of DOI:10.1002/anie.201103203

The enzyme DrrA of the human pathogen Legionella pneumophila adenylylates specifically a tyrosine of the GTPase Rab1. An efficient synthesis route using Fmoc solid phase peptide synthesis led to Tyr-adenylylated peptides and allowed the generation of mono

Full text of DOI:10.1002/anie.201103203

Communications
DOI: 10.1002/anie.201103203
Post-Translational Modification
Efficient Synthesis and Applications of Peptides containing
Adenylylated Tyrosine Residues
Cornelis Smit, Julia Blꢀmer, Martijn F. Eerland, Michael F. Albers, Matthias P. Mꢀller,
Roger S. Goody, Aymelt Itzen,* and Christian Hedberg*
Post-translational modification (PTM) is a versatile strategy
to influence the biological activity of proteins and enzymes.
One fascinating example of the control of enzymatic activity
is the regulation of bacterial glutamine synthetase. Adenyly-
lation, which consists of the covalent transfer of an adenosine
monophosphate (AMP) group from an adenosine triphos-
phate (ATP) precursor to the hydroxy function of a specific
tyrosine residue of glutamine synthetase, renders the enzyme
less active (Scheme 1).[1a] Recently, it was discovered that
certain pathogenic bacteria are able to influence the signaling
of eukaryotic cells by adenylylating small GTPases at either
tyrosine or threonine residues.[1b,2,3] Small GTPases act as
molecular switches and are active or inactive when bound to
GTP or GDP, respectively.[4] The Vibrio parahaemolyticus
effector protein VopS adenylylates a specific threonine
residue in the Rho subfamily (Thr37 in RhoA, Thr35 in
Cdc42 and Rac1) of small GTPases.[2] The substrates of the
effector protein DrrA/SidM from the human pathogen
Legionella pneumophila are GTPases from the Rab subfam-
ily, with the modified tyrosine residue being in the switch 2
region (Tyr77 in Rab1b).[3] VopS belongs to the Fic family of
adenylyltransferases, which contains more than 2700 mem-
bers of sequence-homologous proteins.[5] Although the sub-
strate of VopS has been identified, the physiological protein
substrates of DrrA and the remaining members of the FIC
family proteins are less clear. We hypothesized that identifi-
cation of physiological substrates of adenylylating proteins
(e.g. FIC domains) could be aided by antibodies that
specifically recognize adenylylated proteins in eukaryotic
cells or cell lysates. The generation of a specific anti-tyrosine-
AMP antibody has been hampered by the lack of efficient
synthesis methods, which has not allowed the preparation of
pure multimilligram necessary quantities of a peptide antigen
bearing an adenylylated tyrosine for immunization.[6,7,8a]
We report herein a readily applicable and general
procedure for the synthesis of peptides carrying adenylylated
tyrosine residues by using a standard 9-fluorenylmethoxycar-
bonyl (Fmoc) solid-phase synthesis. We envisioned a Tyr-
AMP building-block approach that relies fully on the global
removal of protective groups under acid conditions, thereby
allowing its direct and uncomplicated application in the Fmoc
solid-phase synthesis of peptides.
As shown by a recent example in the literature, 2’,3’-
bis(ester)-protected adenosine moieties are prone to depuri-
nation under acidic conditions (Scheme 2A).[8b] We hypothe-
sized that the depurination reaction of the commonly used
2’,3’-bis(ester)-protected adenosine is induced by nucleophilic
participation of the 2’-ester group. To circumvent the decom-
position under acidic conditions, the 2’,3’-bis(ester) protecting
groups were replaced with a 2’,3’-isopropylidene acetal, and
the deactivating N6,N6-bis(Boc) protecting groups were
introduced at the adenosine moiety to further stabilize the
system (Scheme 2B).[9] We employed O-cyanoethyl (CNE) as
the protecting group for the phosphodiester, thus allowing for
instant deprotection during the removal of the first Fmoc
groups. This resulted in stabilization of the phosphodiester
linkage in a monoanionic form and generation of the amino
acid building block 1 (Scheme 2C). The phosphodiester
moiety of 1 was cleaved by either direction D or E, thereby
creating two different sets of phosphoramidite precursors (3
and 5), together with their coupling partners (2 and 4).
Scheme 1. Adenylylation of proteins at tyrosine residues.
[*] C. Smit, M. F. Eerland, M. F. Albers, Dr. C. Hedberg
Max-Planck-Institut fꢀr molekulare Physiologie
Abt. Chemische Biologie
Otto-Hahn-Strasse 11, 44227 Dortmund (Germany)
E-mail: christian.hedberg@mpi-dortmund.mpg.de
J. Blꢀmer, M. P. Mꢀller, Prof. Dr. R. S. Goody, Dr. A. Itzen
Max-Planck-Institut fꢀr molekulare Physiologie
Abt. Physikalische Biochemie
Otto-Hahn-Strasse 11, 44227 Dortmund (Germany)
E-mail: aymelt.itzen@mpi-dortmund.mpg.de
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201103203.
9200
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 9200 –9204

temperature over two hours,
resulted in phosphodiester 6 in
45% yield (Scheme 2, step c;
1 mmol scale) after chromato-
graphic purification. Next, we
investigated
the
coupling
between Fmoc-tyrosine allyl
ester 4 and adenosine-derived
phosphoramidite 3 under identi-
cal conditions and we retrieved
56% of phosphodiester 6 after
chromatographic
purification
(Scheme 2, step d). Further opti-
mization of the protocol accord-
ing to route d, by employing a
mixture of tetrazole/diisopropy-
lammonium tetrazolide as the
activator, resulted in a yield of
76% of 6 on a 10 mmol scale
(protocol 3, see the Supporting
Information). Deallylation of 6
with PhSiH3 and [Pd(PPh3)4][12]
gave the desired amino acid
building block 1 in 95% yield
as a white amorphous solid after
purification and lyophilization
on a C18 Sep-Pak cartridge.
The Rab1 (AMP) antigen
was synthesized on a tentagel
resin[13] functionalized with
a
RAM-anchored Fmoc-(Trt)-cys-
teine amide (7, Scheme 3;
RAM = Rink amide resin, Trt =
triphenylmethyl). Coupling of
the Fmoc-protected amino acids
Scheme 2. A) Participation of the 2’-ester leads to depurination upon peptide cleavage under acidic
conditions. B) A switch in the protecting-group strategy allows for cleavage. C) Transformation into the
final building block. D),E) Alternative disconnection sites for the preparation of the CNE-protected
phosphadiester lead to two different routes. Synthesis of the protected Tyr-AMP building block. Reagents (10 equivalents) was carried out
and conditions: a) (iPr)2NP(Cl)O(CH2)2CN (1.2 equiv), CH2Cl2, DIPEA (3 equiv) 30 min, 08C!RT, quant.
b) (iPr)2NP(Cl)O(CH2)2CN (1.2 equiv), CH2Cl2, DIPEA (3 equiv) 30 min, 08C!RT, quant. c) 1. Tetrazole
(3 equiv), acetonitrile, 08C!RT 3 h; 2. TBHP (2.8m in (CH2Cl)2, 1.3 equiv), 45%. d) 1. tetrazole
(3 equiv), acetonitrile, 08C!RT, 3 h; 2. TBHP (2.8m in (CH2Cl)2, 1.3 equiv), 56%. Optimized protocol:
d) 1. tetrazole (1.2 equiv), diisopropylammonium tetrazolide (2.2 equiv), acetonitrile, 08C!RT, 3 h;
2. TBHP (5m in decane, 1.3 equiv) 76%. e) [Pd(PPh3)4] (5 mol%), THF, PhSiH3 (1.5 equiv), RT, 5 h,
95%. Boc=tert-butoxycarbonyl, Bz=benzoyl, DIPEA=N,N-diisopropylethylamine.
using standard HBTU/HOBt
activation[14] on an automated
peptide synthesizer,[15] except
for the protected adenylylated
Fmoc-protected tyrosine build-
ing block 1 (2.5 equiv). In this
case, 1 was coupled manually by
employing HATU/HOAt as the
activating reagent,[16] followed
The synthesis commenced with an investigation of the
coupling efficiency of the two phosphoramidites, where 5 was
prepared from Fmoc-l-tyrosine allyl ester (4),[10] as well as the
conversion of N6N6-bis(Boc)-2’,3’-isopropylideneadenosine
2[9] into the corresponding phosphoramidite building block
by coupling of residual Fmoc-protected amino acids accord-
ing to the automated protocol. The N-terminal Fmoc group
was replaced with an N-acetyl group and subsequently
detached from the resin, as well as globally deprotected by
applying a mixture of trifluoroacetic acid/triisopropylsilane/
water (TFA/TIPS/H2O 90:5:5).[17] Concentration of the solu-
tion in vacuo at ambient temperature and subsequent tritu-
ration with diethyl ether yielded the crude peptide. At this
stage, only traces (< 3%) of depurinated peptide could be
detected by ESIMS, but was not visible by 31P NMR
spectroscopy. After purification by reverse-phase HPLC
(C18) and lyophilization, 8 was isolated in 61% yield. To
verify the generality of the method, we synthesized two
additional peptide sequences with adenylylated tyrosine
3.[11]
Diisopropylaminocyanoethyl
phosphochloridate
(1.2 equiv) together with excess DIPEA (3 equiv) in anhy-
drous dichloromethane were used for preparation of building
blocks 3 and 5 (Scheme 2, steps a and b). Coupling of
stoichiometric amounts of tyrosine phosphoramidite 3 with
adenosine derivative 2 in acetonitrile by employing an excess
of anhydrous tetrazole (2 equiv) as the promoter, followed by
oxidation with anhydrous tert-butylhydroperoxide (TBHP) in
1,2-dichloroethane (2.8m, 1.3 equiv) at 08C to ambient
Angew. Chem. Int. Ed. 2011, 50, 9200 –9204
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 9201

Communications
Scheme 3. Solid-phase synthesis of Rab1 Y-77 (adenylyl)-peptide 8.
AA=amino acid, HBTU=O-(benzotriazol-1-yl)-N,N,N’,N’-tetramethyl-
uronium hexafluorophosphate, HOBt=1-hydroxybenzotriazole,
HATU=O-(7-azabenzotriazol-1-yl)-N,N,N’,N’-tetramethyluronium hexa-
fluorophosphate, HOAt=7-aza-1-hydroxybenzotriazole, TIS=triisopro-
pylsilane.
residues (Ac-GSGA-Y*(AMP)-AGSGC-NH2 (S4, 63%) and
EVYRGAE-Y*(AMP)-AVDG (S5, 59%), both on
a
0.1 mmol scale; see the Supporting Information).
The Rab1-antigen peptide 8 was conjugated to Keyhole
Limpet Hemocyanin (KLH) carrier protein through MBS
coupling (MBS = N-hydroxysuccinimidyl-3-maleimidoben-
zoate) and injected in two rabbits by employing Freundꢀs
complete adjuvant.[18] Five booster immunizations were
carried out over 8 weeks, and final bleeding, serum isolation,
selection over antigen peptide 8 immobilized on sepharose,
and depletion over the non-adenylylated Rab1 sequence Ac-
TITSSYYRGAHGC-NH2 (S3, see the Supporting Informa-
tion) resulted in approximately 6 mg of monoselective poly-
clonal IgG from each animal. For immunosorbent assay
analysis, Rab1-antigen peptide 8 was conjugated to bovine
serum albumin (BSA) through an MBS linkage, and purified
by ultrafiltration (30 kDa cut-off; see the Supporting Infor-
mation), which resulted in BSA-8. Comparison of antisera
from each animal by immunosorbent assay over BSA-8
showed no major differences in the antibody titer (see the
Supporting Information). In a Western blot analysis, Rab1b-
AMP could be clearly discriminated from wild-type (wt)
Rab1b by the derived antibody (Figure 1A). The level of
binding of the a-Tyr-AMP antibody to Rab1b-AMP is
approximately 20-fold higher than for unmodified Rab1b.
This finding suggests a major contribution of the AMP moiety
to the antibody binding and only a weak recognition of the
Figure 1. Affinity and specificity of the derived a-Tyr-AMP antibody.
A) Western blot analysis with the indicated amounts of highly purified
Rab1b and Rab1b-AMP by using the a-Tyr-AMP-antibody (1:100
dilutions), thus demonstrating the specific recognition of the Tyr-AMP
modification. B) Western-blot analysis of highly purified adenylylated
and unmodified forms of BSA, Rab1b, and Cdc42 (1 or 0.1 mg, protein
sample per lane). The a-Tyr-AMP antibody (1:100 dilution) strongly
binds to all the tested adenylylated proteins, thus indicating additional
binding activity for Thr-AMP (Cdc42). C) Competition of the a-Tyr-AMP
antibody in the presence of GMP or AMP. Samples (0.1 mg each) were
carried out as indicated in (B). The Western blots were incubated with
the a-Tyr-AMP antibody in the presence and absence of 5 mm GMP or
AMP. AMP, but not GMP, competes moderately with antibody binding
to Rab1b-AMP and BSA-AMP. Both AMP and GMP compete with the
antibody for Cdc42-AMP detection. In all Western blots, IRDye800-
conjugated donkey anti-rabbit IgG was used as a secondary antibody.
peptide backbone that was part of the immunization. Under
these experimental conditions, as little as 10 ng Rab1b-AMP
could be detected with the a-Tyr-AMP antibody. The experi-
ments were repeated with different adenylylated proteins to
investigate the specificity of the derived a-Tyr-AMP antibody.
In addition to Rab1b-AMP and BSA-8, the adenylylated
9202 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 9200 –9204

Cdc42 (Cdc42-AMP) that had been preparatively adenyly-
lated using VopS was investigated. VopS has been reported to
adenylylate Cdc42 specifically on Thr35,[2] and can thus serve
as a control as to whether the a-Tyr-AMP antibody can
discriminate between adenylylated threonine and tyrosine
residues (Figure 1B). Apparently, the a-Tyr-AMP antibody
recognizes adenylylated BSA-8 and Rab1b-AMP specifically
over the unmodified proteins. However, the antibody also
recognizes Cdc42-AMP, thus indicating that it can also detect
adenylylated threonine residues. This again suggests a major
contribution of the AMP group to antibody binding, with only
a small influence by the modified amino acid side chain and
the peptide backbone.
To evaluate the binding specificity of the generated a-Tyr-
AMP antibody further, we have performed nucleotide com-
petition experiments. The binding of the a-Tyr-AMP to 8-
BSA (wt-BSA as control), AMP-Rab1b (wt-Rab1b as con-
trol), and AMP-Cdc42 (wt-Cdc42 as control) was investigated
in the presence and absence of either GMP or AMP
(Figure 1C). In the presence of GMP, no impairment of the
binding of the a-Tyr-AMP antibody to BSA-8 or Rab1b-AMP
could be detected. However, the signal corresponding to the
binding of the a-Tyr-AMP antibody decreased significantly
on incubating the samples with AMP, thus demonstrating the
relevance of the adenine base of the AMP moiety for
interaction with the antibody. Intriguingly, both GMP and
AMP competed with the binding of the a-Tyr-AMP antibody
to Cdc42-AMP, although the competition with AMP
appeared to be slightly more effective than with GMP. This
observation could possibly hint at a strong recognition of the
furanoside residue and the phosphate group of GMP/AMP,
thereby leading to a substantial degree of competition
between AMP and GMP in the absence of the Rab1b peptide
sequence.
To further investigate the applicability of the Rab1
antibody we performed pull-down experiments of adenyly-
lated proteins from cell lysates. For this purpose we exoge-
nously added adenylylated Rab1 and Cdc42 to mammalian
(simian) Cos7 cell lysates. We were able to preferably pull
down adenylylated Rab1, thus indicating a specificity of the a-
Tyr-AMP antibody for Rab1-AMP binding even in the
presence of the competitive environment of the cell lysate
(Figure 2).
The generated tyrosine-AMP-specific antibody will help
identify the physiological substrates of proteins containing an
FIC domain, for example, BepA[19] well as DrrA.[3] In
addition, the peptide-derived tools could be employed to
establish whether adenylylation of proteins is a more general,
currently underappreciated, process in eukaryotic cells in the
absence of pathogenic adenylylating proteins.
Received: May 10, 2011
Revised: June 20, 2011
Published online: August 25, 2011
Keywords: adenylylation · antibodies · peptides ·
.
post-translational modifications · Rab proteins
[1] a) E. R. Stadtman, J. Biol. Chem. 2001, 276, 44357 – 44364; b) for
a recent review on adenylylation, see A. Itzen, W. Blankenfeldt,
R. S. Goody, Trends Biochem. Sci. 2011, 36, 221 – 228.
[2] M. L. Yarbrough, Y. Li, L. N. Kinch, N. V. Grishin, H. L. Ball, K.
Orth, Science 2009, 323, 269 – 272.
[3] M. P. Mꢁller, H. Peters, J. Blꢁmer, W. Blankenfeldt, R. S. Goody,
A. Itzen, Science 2010, 329, 946 – 949.
[4] For a recent review covering the topic of small GTPases as
regulators of vesicular transport, see A. Itzen, R. S. Goody,
Semin. Cell Dev. Biol. 2011, 22, 48 – 56.
[5] L. N. Kinch, M. L. Yarbrough, K. Orth, N. V. Grishin, PLoS
ONE 2009, 4(6), e5818.
[6] Y. H. Hao, T. Chuang, H. L. Ball, P. Luong, Y. Li, R. D. Flores-
Saaib, K. Orth, J. Biotechnol. 2011, 151, 251 – 254.
[7] R. A. Al-Eryania, L. Yan, H. L. Ball, Tetrahedron Lett. 2010, 51,
1730 – 1731.
[8] a) For an early example of interassembly synthesis of adenyly-
lated tyrosine, see J. A. Todhunter, D. L. Purich, Biochem.
Biophys. Res. Commun. 1974, 59, 82 – 85; b) G. J. van der He-
den van Noort, H. S. Overkleeft, G. A. van der Marel, D. V.
Filippov, J. Org. Chem. 2010, 75, 5733 – 5736.
[9] H. Ikeuchi, M. E. Meyer, Y. Ding, J. Hiratake, N. G. J. Richards,
Bioorg. Med. Chem. 2009, 17, 6641 – 6650.
Figure 2. Pull down of adenylylated Rab1/Cdc42 from mammalian cell
lysates. Preparative adenylylated Rab1 and Cdc42 (0.1 mg each) were
exogenously added to buffer or 100 mg Cos7 cell lysate. A biotinylated
a-Tyr-Rab1-AMP antibody was immobilized on magnetic streptavidin
beads and used to pull down the Rab1-AMP and Cdc42-AMP samples.
A) Loading control containing sample mixtures before pull down.
B) Pull-down experiment of (A). In all Western blots, IRDye800-
conjugated donkey anti-rabbit IgG was used as a secondary antibody.
[10] S. Ficht, R. J. Payne, R. T. Guy, C.-H. Wong, Chem. Eur. J. 2008,
14, 3620 – 3629; for a previous example of a CNE-phosphoroa-
midite building block of the Fmoc-protected tyrosine methyl
ester, see C. D. Claeboe, R. Gao, S. M. Hecht, Nucl. Acids Res.
2003, 31, 5685 – 5691.
[11] For a recent review on phosphoroamidite chemistry, see C.
Hoebartner, F. Wachowius, Chem. Biol. Nucl. Acids 2010, 1 – 37.
Angew. Chem. Int. Ed. 2011, 50, 9200 –9204
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 9203

Communications
[12] M. Dessolin, M. G. Guillerrez, N. Thieriet, F. Guibe, A. Loffet,
ally deprotected peptide from the resin. After filtering from the
resin, an additional 10% H2O was added to the combined
cleavage mixture, which was aged for 60 min to ensure complete
removal of the isopropylidene acetal, prior to concentration and
precipitation with diethyl ether.
Tetrahedron Lett. 1995, 36, 5741.
[13] Acquired from Rapp Polymere GmbH, Tꢁbingen, Germany.
[14] V. Dourtoglou, B. Gross, V. Lambropoulou, C. Ziodrou, Syn-
thesis 1984, 572.
[15] Solid-phase peptide synthesis was carried out on an Applied
Biosystems peptide synthesizer equipped for the Fmoc protocol.
[16] L. A. Carpino, J. Am. Chem. Soc. 1993, 115, 4397.
[17] Between each cleavage cycle, the resin was washed with
trifluoroethanol (CF3CH2OH, 3 ꢂ 2 mL) to solubilize any parti-
[18] Peptide conjugation, immunization, and antibody purification
were performed by Biogenes GmBH, Berlin, Germany
(www.biogenes.de).
[19] D. V. Palanivelu, A. Goepfert, M. Meury, P. Guye, C. Dehio, T.
Schirmer, Protein Sci. 2011 20, 492 – 499.
9204 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 9200 –9204