Synthesis of rhamnosylated arginine glycopeptides and determination of the glycosidic linkage in bacterial elongation factor P

Article abstract of DOI:10.1039/c6sc03847f

A new class of N-linked protein glycosylation-arginine rhamnosylation-has recently been discovered as a critical modification for the function of bacterial elongation factor P (EF-P). Herein, we describe the synthesis of suitably protected α- and β-rhamnosylated arginine amino acid “cassettes” that can be directly installed into rhamnosylated peptides. Preparation of a proteolytic fragment of Pseudomonas aeruginosa EF-P bearing both α- and β-rhamnosylated arginine enabled the unequivocal determination of the native glycosidic linkage to be α through 2D NMR and nano-UHPLC-tandem mass spectrometry studies.

Full text of DOI:10.1039/c6sc03847f

View Article Online
View Journal
Chemical
Science
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: R. J. Payne, S.
Wang, L. Corcilius, P. P. Sharp, A. Rajkovic, M. Ibba and B. Parker, Chem. Sci., 2016, DOI:
1
0.1039/C6SC03847F.
Volume
7
Number
1
January 2016 Pages 1–812
This is an Accepted Manuscript, which has been through the
Royal Society of Chemistry peer review process and has been
accepted for publication.
Chemical
Science
Accepted Manuscripts are published online shortly after
www.rsc.org/chemicalscience
acceptance, before technical editing, formatting and proof reading.
Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.
You can find more information about Accepted Manuscripts in the
author guidelines.
Please note that technical editing may introduce minor changes
to the text and/or graphics, which may alter content. The journal’s
standard Terms & Conditions and the ethical guidelines, outlined
in our author and reviewer resource centre, still apply. In no
event shall the Royal Society of Chemistry be held responsible
for any errors or omissions in this Accepted Manuscript or any
consequences arising from the use of any information it contains.
ISSN 2041-6539
EDGE ARTICLE
Francesco Ricci et al.
Electronic control of DNA-based nanoswitches and nanodevices
rsc.li/chemical-science

Page 1 of 7
Pleas Ce hd eo mni oc ta al dS j cu iset n mc ea rgins
View Article Online
DOI: 10.1039/C6SC03847F
Chemical Science
EDGE ARTICLE
Synthesis of Rhamnosylated Arginine Glycopeptides and
Determination of the Glycosidic Linkage in Bacterial Elongation
Factor P
Received 00th January 20xx,
Accepted 00th January 20xx
a
a
b
c
c
d
Siyao Wang, Leo Corcilius, Patrick P. Sharp, Andrei Rajkovic, Michael Ibba, Benjamin L. Parker,
DOI: 10.1039/x0xx00000x
a
Richard J. Payne*
www.rsc.org/
A new class of N‐linked protein glycosylation – arginine exclusive to higher eukaryotic organisms, it is now established that
rhamnosylation – has recently been discovered as a critical widespread O‐ and N‐linked glycosylation of bacterial proteins also
1
1,
modification for the function of bacterial elongation factor P (EF‐ occurs with similar linkage types to the eukaryotic counterparts.
P). Herein, we describe the synthesis of suitably protected ‐ and
1
2
However, since the field of bacterial protein glycosylation is still in

‐rhamnosylated arginine amino acid “cassettes” that can be its infancy, the effect of glycosylation on structure and function is
directly installed into rhamnosylated peptides. Preparation of a still not well understood. Very recently, a number of new
proteolytic fragment of Pseudomonas aeruginosa EF‐P bearing glycosylation motifs have been discovered in bacteria, including S‐
1
3‐15
both ‐ and ‐rhamnosylated arginine enabled the unequivocal glycosylation
of cysteine and N‐glycosylation of the guanidine
1
6‐19
determination of the native glycosidic linkage to be through 2D side chain of arginine (Arg) residues.
This latter modification
NMR and nano‐UHPLC‐tandem mass spectrometry studies.
was first observed as a ‐N‐acetyl‐D‐glucosamine (‐GlcNAc)
modification on specific Arg residues within human proteins bearing
the death domain motif, including TRADD, FADD, TNFR1 and the
Protein glycosylation is the most ubiquitous post‐translational
1
6, 17
modification in nature, estimated to occur on more than 50% of
kinase RIPK1.
Intriguingly, this modification was shown to be
1
human proteins. In eukaryotes, glycosylation is a key modulator of
performed by a type III secretion system effector protein called
NleB produced by enteropathogenic strains of Escherichia coli.
Functionally, ‐GlcNAcylation of specific Arg residues within death
domains by NleB prevents cell death pathways by blocking
downstream signalling that would normally lead to apoptosis or
biological recognition events, including cell adhesion,
2
differentiation and growth. In addition, the presence of glycans can
affect the structure and activity of a given protein and aberrant
glycan structures have been implicated in a number of serious
3
4
1
6, 17
illnesses including cancer and autoimmune diseases. Eukaryotic
necroptosis of E. coli‐infected cells.
protein glycosylation can be divided into two main classes: O‐
glycosides, whereby a glycan is covalently linked to the hydroxyl In more recent work, Lassak and Jung and co‐workers and Rajkovic
moiety of serine, threonine, tyrosine, hydroxylysine or et al. discovered L‐rhamnosylation of Arg residues within bacterial
5
, 6
hydroxyproline with ‐ or ‐anomeric configuration,
or N‐ translation Elongation factor P (EF‐P), a protein required for
1
8,
glycosides, in which N‐acetylglucosamine is ‐linked to the amide preventing ribosome stalling when translating polyproline motifs.
side chain of an asparagine residue (within an Asn‐Xaa‐Ser/Thr
Importantly, the number of eukaryotic proteobacterium Shewanella oneidensis, the pathogenic bacteria
proteins known to be glycosylated has dramatically increased in Pseudomonas aeruginosa and Neisseria meningitidis. In all three
recent years owing, in major part, to improvements in bacteria rhamnosylation of Arg‐32 was shown to be critical for the
1
9
To date, the modification has been discovered on EF‐P of the ‐
6
,
7
18
consensus sequence).
1
9
20
8‐10
glycoproteomic mass spectrometry methodologies.
function of EF‐P in assisting translation and for the pathogenicity of
1
8
P. aeruginosa in vitro. The glycosylation was also shown to be
performed by EarP, a rhamnosyltransferase that employs dTDP‐L‐
rhamnose as the substrate. An important piece of information that
was not revealed in these studies was the stereochemistry at the
anomeric center of L‐rhamnose appended to the side chain of Arg‐
32. Establishing the stereochemical configuration of this centre is
critical to determine the conformational effect of the carbohydrate
on the underlying peptide backbone and to ascertain whether EarP
is an inverting or retaining glycosyltransferase. In this work we were
interested in developing an efficient synthetic route toward suitably
protected rhamnosylated arginine amino acids with both ‐ and ‐
While glycosylation was initially deemed to be a modification
a.
School of Chemistry, The University of Sydney, Sydney, NSW 2006, AUSTRALIA.
ACRF Chemical Biology Division, Walter and Eliza Hall Institute of Medical
b.
Research, 1G Royal Parade, VIC3052, Australia.
Department of Microbiology and Center for RNA Biology, Ohio State University,
c.
Columbus, Ohio, USA
Charles Perkins Centre, The University of Sydney, NSW 2006, Australia.
d.
Electronic Supplementary Information (ESI) available: materials and methods,
compound characterization, 1D and 2D NMR and mass spectrometry data of
glycopeptides. See DOI: 10.1039/x0xx00000x
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1‐3 | 1
Please do not adjust margins

Pleas Ce hd eo mni oc ta al dS j cu iset n mc ea rgins
Page 2 of 7
EDGE ARTICLE
Chemical Science
configuration. In doing so we would be able to synthesize purification, was treated with thiophosgene and CaCO to afford
View Article Online
3
DOI: 10.1039/C6SC03847F
glycopeptides bearing both rhamnosyl configurations for the glycosyl isothiocyanate 4 as predominantly the β‐isomer (α:β
2
8
unequivocal determination of the anomeric configuration of the 1:9). In this case β‐selectivity was presumed to result from base‐
rhamnose moiety on EF‐P. While this work was in progress, Li et al. catalysed isomerization of the α‐glycosylamine to the
1
8
27
reported an NMR study on recombinant S. oneidensis EF‐P that on thermodynamically favoured equatorial (β) isomer. Importantly,
the basis of a JC1,H1 coupling constant in an undecoupled HSQC 2D NOESY NMR experiments of the two glycosyl isothiocyanates 3
experiment (J = 167 Hz) and the absence of a H1‐H5 NOE (Nuclear and 4 confirmed the respective anomeric configurations (See
Overhauser Effect) would suggest that the rhamnose is ‐ Supplementary Information).
2
1
configured at the anomeric centre. Herein, we report a
comprehensive analysis of the stereochemistry of the
rhamnosylation of P. aeruginosa EF‐P through 2D NMR
spectroscopy, HPLC, nanoLC and tandem mass spectrometric
comparison of synthetic rhamnosylated peptides with the
proteolytic fragment of isolated P. aeruginosa EF‐P and
recombinant P. aeruginosa EF‐P both bearing the modification. Our
studies provide strong evidence that independently supports the
notion that rhamnose is ‐linked to arginine in bacterial EF‐P
proteins. We also reveal that the ‐anomer of rhamnosylated
arginine is configurationally labile and undergoes base‐promoted
endocyclic ring opening anomerization to the ‐anomer which, to
our knowledge has not been previously observed in native N‐linked
glycosides.
We began by targeting proteolytic fragments of P. aeruginosa EF‐P
bearing both ‐ and ‐configured rhamnosylated Arg. Hu and co‐
workers have reported the synthesis of peptides containing N‐
2
2
21
GlcNAcylated and N‐rhamnosylated Arg through a solid‐phase
guanidinylation strategy whereby the requisite glycosyl isothiourea
was reacted with a resin‐bound ornithine‐containing peptide. While
the direct glycosylation of resin‐bound peptides is attractive from a
synthetic standpoint, we were concerned about the configurational
stability of the rhamnose linkage under the acidic and basic
conditions employed for cleavage and deprotection of the
glycopeptide following solid‐phase peptide synthesis (SPPS).
Therefore we opted for a building block strategy, in which pre‐
fabricated α‐ and β‐rhamnosylated arginine building blocks 1 and 2
could be robustly interrogated by 2D NMR techniques and the
stability of the individual glycosylamino acids examined under acidic
and basic conditions before direct incorporation into glycopeptides
as ‘cassettes’ through Fmoc‐strategy SPPS (Scheme 1).
Scheme 1. A) Stereoselective synthesis of
In order to mask the nucleophilicity of the guanidinium side chain of rhamnosyl isothiocyanates and B) Synthesis of suitably protected
Arg during SPPS, glycosylamino acid building blocks 1 and 2 were rhamnosylated arginine building blocks and with ‐ and
‐ and ‐configured
3
4;
1
2

‐
designed with an acid‐labile 2,2,4,6,7‐pentamethyldihydro‐ anomeric configuration, respectively
.
benzofuran‐5‐sulfonyl (Pbf) side chain protecting group, commonly
employed for the protection of the parent amino acid in Fmoc‐ With the two rhamnosylated isothiocyanates 3 and 4 in hand, we
2
3
SPPS. Synthesis of the two rhamnosylated arginine building blocks next sought to introduce the Pbf‐protected guanidine moiety
and 2 began with preparation of the α‐ and β‐configured glycosyl (Scheme 1B). Direct treatment of isothiocyanates 3 and 4 with
1
2
1
isothiocyanates 3 and 4, both of which were accessed from deprotonated Pbf‐NH proved to be the optimal means to generate‐
2
2
4
29
rhamnosyl acetate 5 as a common precursor (Scheme 1A). To Pbf protected thioureas from 3 and 4 (Scheme 1B). These were
2
5
access α‐isothiocyanate 3, rhamnosyl chloride 6 was treated with not isolated, but rather treated directly with EtI and K CO in MeCN
2 3
KSCN in the presence of catalytic TBAI in refluxing acetonitrile. to afford ethyl isothioureas 8 and 9 in good yields. Subsequent
Under these conditions complete 1,2‐trans (α) stereoselectivity was Hg(II)‐promoted guanidinylation of each of the isothioureas with
afforded due to anchimeric assistance from the neighbouring Fmoc‐Orn‐OAll then gave the corresponding glycosylamino acid
2
6
acyloxy substituent. Alternatively, treatment of the glycosyl derivatives 10 and 11 in excellent yields. Finally, deprotection of the
2
7
acetate 5 with SnCl and TMSN provided glycosyl azide 7. This allyl ester moiety with tetrakis(triphenylphosphine)palladium(0)
4
3
azide was reduced to the glycosylamine using H ,Pd/C and, without and phenylsilane afforded the suitably protected rhamnosylated
2
2
| J. Name., 2012, 00, 1‐3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Page 3 of 7
Pleas Ce hd eo mni oc ta al dS j cu iset n mc ea rgins
Journal Name
COMMUNICATION
arginine building blocks 1 and 2 ready for direct incorporation into
glycopeptides by Fmoc‐SPPS.
View Article Online
DOI: 10.1039/C6SC03847F
At this stage, we set out to fully characterize building blocks 1 and 2
via NMR spectroscopy. Unfortunately, despite screening a number
of solvents and temperatures in NMR experiments, proton signals
from the pyranosyl ring were extremely broad, thus making it
difficult to assign the configuration of the anomeric centre through
2D NMR experiments. This presumed rotameric (or tautomeric)
effect was alleviated by deprotection of both the acetyl and Pbf
moieties on building blocks 1 and 2 (via acidolytic deprotection of
the Pbf followed by deacetylation by treatment with sodium
methoxide in MeOH). These conditions were selected to mirror the
intended deprotection sequence of the rhamnosylated peptide
following assembly by Fmoc‐SPPS. Deprotection of building block 2
Scheme 2. Proposed base‐catalyzed anomerization of ‐rhamosylated Arg
to‐rhamnosylated Arg.
(
1
with ‐anomeric configuration) proceeded smoothly to afford triol
2 in 36% yield following HPLC purification. The anomeric
configuration of 12 was unequivocally determined through 1D and
D NMR analysis (see Supplementary Information). Unexpectedly,
2
significant anomerization was observed when the same
deprotection sequence was applied to glycosylamino acid 1 bearing

‐configuration at the anomeric centre. Specifically, upon
treatment of 1 with TFA followed by methoxide, a 4:5 :mixture
of anomers was generated (see Supplementary Information). Upon
further examination, it was determined that the anomerization
resulted from the deacetylation step, suggesting a base‐catalyzed
mutarotation event. Thankfully, reversal of the deprotection
sequence (i.e. treatment with NaOMe in MeOH prior to treatment
with TFA) led to almost complete retention of the anomeric
stereochemistry (α:β 9:1), suggesting that the presence of the
electron‐withdrawing Pbf protecting group was necessary to
suppress base‐catalysed anomerization. The configuration of the
anomeric linkage was able to be confirmed by 2D NOESY analysis of
the deprotected glycosylamino acids 12 and 13 (see Supplementary
Information for details). Importantly, ‐configured rhamnosylated
Arg 12 exhibited a distinct NOE between H1 and H5 as well as
between H1 and H3, whereas these NOEs were absent in the
corresponding ‐anomer 13 (see Supplementary Information).
1
HSQC experiments without H decoupling were also performed to
1
13
determine the H‐ C coupling constant at C‐1 (J
= 167 Hz for 13
C1,H1
Scheme 3. Synthesis of Lys‐C proteolytic fragments of P. aeruginosa EF‐P
bearing ‐rhamose (14) and ‐rhamnose (15) anomeric configurations.
and 154 Hz for 12, see Supplementary Information for details).
Deprotected glycosylamino acids 12 and 13 were also found to be
stable in 0.1% TFA in MeCN/H O, which is a common eluent system
2
With the two target rhamnosyl arginine building blocks 1 and 2 in
hand, we began assembly of target peptide fragments 14 and 15
bearing ‐ and ‐configured rhamnosyl arginine, respectively via
Fmoc‐SPPS (Scheme 3). These glycopeptides correspond to the Lys‐
C proteolysed fragment of P. aeruginosa EF‐P that has been
characterized via MS(MS ). Synthesis of these glycopeptides
began by pre‐loading of Chemmatrix® Trtyl‐OH resin with Fmoc‐
Lys(Boc)‐OH, followed by standard Fmoc‐SPPS protocols to
generate resin‐bound hexapeptide 16. Building blocks 1 or 2 (2.2
equiv.) were then coupled onto the resin‐bound peptide using 1‐
used to purify peptides and glycopeptides. The above finding that
the deprotected rhamnosylarginine isomerizes under basic
conditions highlights a unique behavior of the Arg‐N‐linked glycan
motif as the more common Asn‐N‐linked and O‐linked glycans are
not susceptible to anomerization under these conditions.
Mechanistically, we propose that anomerization of ‐
rhamnosylated arginine proceeds through endocyclic ring opening
that is facilitated by a lone pair on one of the guanidine ‐nitrogen
atoms (see Scheme 2). This mechanistic proposal is supported by
previous observations that mannosyl thiourea derivatives are
susceptible to base‐catalysed endocyclic ring opening
3
19
[
3
bis(dimethylamino)methylene]‐1H‐1,2,3‐triazolo[4,5‐b]pyridinium
‐oxid hexafluorophosphate (HATU, 2.0 equiv.), 1‐Hydroxy‐7‐
3
0
anomerization.
31
azabenzotriazole (HOAt, 3.0 equiv) and sym‐collidine (2.4 equiv.)
as the base. These conditions facilitated complete incorporation of
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1‐3 | 3
Please do not adjust margins

Pleas Ce hd eo mni oc ta al dS j cu iset n mc ea rgins
Page 4 of 7
EDGE ARTICLE
Chemical Science
the glycosylamino acids (as judged by HPLC‐MS analysis) to afford glycopeptide. The optimal deprotection sequence oVfierweAsritnic‐leboOunlninde
resin‐bound glycopeptide 17 and 18, which were further elongated 19 therefore involved release of the DsOidIe: 10c.1h0a3i9n/C6pSrCo0te38ct4e7dF
to the desired resin bound nonapeptides 19 and 20 following glycopeptide from the resin using HFIP, careful deacetylation with
standard Fmoc‐SPPS protocols. Deprotection of the β‐ methoxide (pH 8‐9) followed by global deprotection with an
rhamnosylated peptide 20 was carried out through acidolytic side acidolytic cocktail comprising TFA/i‐Pr SiH/H O to afford
3
2
chain deprotection and cleavage from the resin followed by predominantly α‐rhamnosylated peptide 14 (α:β 9:1) in 17% yield
Zemplén deacetylation at pH 8‐9 to afford peptide 15 in 30% based on the original resin loading. It should be noted that during
isolated yield (based on the original resin loading). In contrast, the the above experiments we were also able to confirm the identity of
configurationally unstable α‐rhamnosylated peptide 19 was cleaved the proposed imine intermediate that forms via the base‐catalysed
from the resin using
a
weakly acidic solution of endocyclic ring opening anomerization (Scheme 2). Specifically,
hexafluoroisopropanol (HFIP) in DCM (30% v/v) to afford the fully following co‐treatment of 14 with methoxide and sodium
side‐chain protected peptide, which was then carefully borohydride, the ring opened imine could be trapped as the
deacetylated under Zémplen conditions before global side chain corresponding amine (see Supplementary Information).
deprotection with TFA:i‐Pr SiH/H O to afford predominantly the ‐
3
2
We next extensively characterized glycopeptides 14 and 15 with a
view to performing comparative studies with native P. aeruginosa
EF‐P for determination of the anomeric configuration of the
rhamnose moiety in the native protein. 2D NMR experiments were
initially used to unequivocally confirm that the rhamnopyranosyl
moiety of glycopeptides 14 and 15 was α‐ and β‐configured,
glycopeptide 14 (9:1 :). This Zémplen deacetylation reaction
required careful monitoring by UPLC‐MS as either excessively high
pH (> 9) or extended reaction time still led to significant
anomerization. Indeed, as with amino acid 13, treatment of ‐
rhamnosylated EF‐P glycopeptide 14 with dilute methoxide (pH 9.0)
for 2 h led to significant anomerization to the corresponding ‐
configured glycopeptide 15 (1:2 ratio 14:15, see Scheme 3 and
Supplementary Information). Unfortunately, an attempt to avoid
anomerization completely through on‐resin deacetylation of 19
with 10vol.% hydrazine monohydrate in DMF, followed by cleavage
from the resin with an acidic cocktail, led to significant
anomerization [1:1 14():15(), see Supplementary Information].
Thus, careful off‐resin treatment with sodium methoxide in MeOH
was deemed optimal for deacetylation of the ‐configured
1
respectively. We first performed H coupled HSQC experiments that
1
13
provided a H‐ C coupling constant at the anomeric centre of 167
Hz for the ‐linkage of 14 and 156 Hz for the ‐linkage of 15, very
similar to the values obtained for the parent rhamnosylamino acids
1
and 2 (vide supra) (Figure 1A). It should be noted that the
coupling constant for 14 was also in agreement with that
1
determined for recombinant S. oneidensis EF‐P homologue ( J
=
C‐H
2
1
1
67 Hz) reported by Li et al. Selective ROESY irradiation at H1 was
also performed which
showed a NOE with H5 of
5 (with a diaxial H1‐H5
A)
B)
1
relationship, see Figure 1B)
but not for 14 (see Figure
1C). Finally, we developed
analytical HPLC‐MS
D)
conditions to separate the
two anomeric glycopeptides
1
4 and 15 with a retention
time difference of ~1 min
see Figure 1D for individual
(
and overlayed HPLC traces).
C)
Having prepared the two
possible anomers of the
Lys‐C
glycopeptides
proteolysed
of P.
aeruginosa EF‐P (14 and 15)
and confirmed their
anomeric stereochemistry,
we next performed
comparative analysis to the
Lys‐C glycopeptide
produced from isolated P.
aeruginosa EF‐P prepared
as reported previously via
1
13
1
Figure 1. A) zoom in of the H‐ C HSQC without proton decoupling for: H‐SGR(‐Rha)NAAVVK‐OH (14), JCH = 167 Hz (left)
19
1
and H‐SGR(‐Rha)NAAVVK‐OH (15), JCH = 156 Hz (right); B) selective ROESY experiments for H‐SGR(‐Rha)NAAVVK‐OH (15).
Black = zoom in of H NMR spectrum, Blue = Selected ROESY by irradiating H‐1; Red: Selected ROESY by irradiating H5. The
nano‐ultra high pressure
1
through‐space NOE relationship between diaxial H5 and H1 can be observed; C) selective ROESY experiments for H‐SGR(‐
1
Rha)NAAVVK‐OH (14). Black = zoom in of H NMR spectrum, Blue = Selected ROESY by irradiating H‐5; Red: Selected ROESY
by irradiating H1. No NOE was observed; D) analytical HPLC traces of H‐SGR(‐Rha)NAAVVK‐OH 14 (top), H‐SGR(‐
4
| J. Name., 2012, 00, 1‐3
This journal is © The Royal Society of Chemistry 20xx
Rha)NAAVVK‐OH 15 (middle), overlay of HPLC chromatograms of 14 and 15 (bottom); gradient: 100% A for 2 min then 0 to
0% B over 30 min, Waters Atlantis® T3 C18 Column at 0.2 mL/min
2
Please do not adjust margins

Page 5 of 7
Pleas Ce hd eo mni oc ta al dS j cu iset n mc ea rgins
Journal Name
COMMUNICATION
liquid chromatography (nanoUHPLC) coupled to tandem mass fragment arising from a loss of the rhamnose sugar fVrioewmArtthicele tOrniplinlye
spectrometry (see Supplementary Information). Separation of the charged precursor‐ion under collision‐induDceOdI: 1d0i.s1s0o3c9i/aCt6ioSnC0(3C8I4D7)F.
two anomeric rhamnosylated glycopeptides 14 and 15 was The difference in stability of the ‐ verses β‐linkage prompted us to
achieved with nanoUHPLC using isocratic elution on a 50 cm column investigate relative fragment ion distributions under CID MS/MS as
(Figure 2A). Analysis of the Lys‐C proteolysed glycopeptide fragment a further tool to differentiate between the two anomers. The
of purified P. aeruginosa EF‐P clearly showed a retention time synthetic ‐ and β‐rhamnosylated glycopeptides were analysed by
alignment with the ‐rhamnosylated glycopeptide 14 providing CID and higher collisional dissociation (HCD) MS/MS with increasing
convincing evidence that the rhamnose moiety appended to Arg‐32 normalised collision energies (NCEs), specifically 10 – 40 NCE with
of P. aeruginosa EF‐P is ‐linked. We further confirmed this steps of 5 NCE. As expected, no fragment ion masses were unique
retention time‐based evidence by spiking either the synthetic ‐ to the ‐ or β‐linkage with either fragmentation approach.
rhamnosylated (14) or β‐rhamnosylated glycopeptide (15) into a However, we observed subtle differences in the intensity of the
sample of Lys‐C proteolysed P. aeruginosa EF‐P, followed by fragment‐ion distributions. The intensity of the doubly charged
analysis by nanoUHPLC‐MS/MS (see Figure 81, Supplementary fragment arising from a loss of the rhamnose sugar from the triply
Information). Addition of the ‐rhamnosylated (14) glycopeptide charged precursor‐ion was the major ion significantly different for
increased the intensity of a single chromatographic peak while the two anomers with either CID or HCD. We hypothesise that this
spiking in the β‐rhamnosylated glycopeptide (15) produced two diagnostic fragment ion may arise from differences in the C‐N bond
peaks of similar intensity, thus further supporting the ‐anomeric strength at the anomeric centre. Specifically, due to back donation
configuration of rhamnose in P. aeruginosa EF‐P. Interestingly, we of electron density from the endocyclic oxygen into the * of the
also observed <5% of the β‐rhamnosylated peptide in the C1‐N bond in the ‐anomer (i.e. the endo anomeric effect) the C‐N
proteolysed sample. This observation is consistent with the bond of this anomer would be expected to be longer and weaker
endocyclic ring opening anomerization observed during preparation than the corresponding ‐anomer, thus leading to more extensive
of the synthetic glycopeptides. It is important to note that this loss of rhamnose. The optimal fragmentation energy to maximize
anomerization occurred despite only neutral buffer conditions the intensity differences and distinguish between ‐ and β‐anomers
being used and suggests that rhamnosylated peptides and proteins was 10 – 15% NCE and 30 – 40% NCE for HCD and CID, respectively
(whether expressed, isolated or synthesised) must be handled (see Figure 83, Supplementary Information). However, the
carefully during isolation, purification and other manipulations to fragmentation efficiency of HCD at 15 NCE was only approximately
prevent extensive anomerization. In order to further strengthen our 20% with the un‐fragmented precursor ion being the major ion in
observations on the anomeric configuration of rhamnose in EF‐P, MS/MS (see Figure 84, Supplementary Information). Therefore, all
we produced recombinant P. aeruginosa EF‐P in E. coli and further work was performed with CID at 35 NCE which produces the
performed in vitro rhamnosylation with EarP using dTDP‐L‐ maximal difference in the diagnostic fragment ion intensity
1
8, 19
rhamnose as a substrate (see Supplementary Information).
between ‐ and β‐anomers, and produces extensive peptide
Analysis of the Lys‐C proteolysed glycopeptide fragment also backbone fragmentation efficiency.
suggested ‐configuration of the rhamnose moiety based on
CID MS/MS analysis of the Lys‐C glycopeptide fragment from
isolated P. aeruginosa EF‐P displayed very similar fragmentation
retention time analysis with synthetic 14 and 15 by nanoUHPLC.
Finally, we performed fragmentation of the rhamnosylated
glycopeptide from P. aeruginosa with ETD MS/MS which localised patterns to the ‐rhamnosylated glycopeptide 14 (Figure 3D).
1
9
the glycosylation site to Arg32 (see Figure 82, Supplementary Quantification of this fragment‐ion intensity (normalized to the
Information).
precursor‐ion intensity) highlighted a significant difference between
the ‐ and β‐configured glycopeptides (Figure 3E). The intensity of
this fragment‐ion from glycopeptides produced from P. aeruginosa
To further confirm the ‐configurational assignment of EF‐P
rhamnosylation, MS/MS analysis was performed on glycopeptides EF‐P and the in vitro rhamnosylated recombinant EF‐P are clearly
4 and 15 bearing ‐ and ‐configuration, respectively, and the consistent with the ‐configuration. Taken together, the NMR,
data compared to glycopeptides produced from Lys‐C digestion of UHPLC and tandem mass spectrometry
1
data provide strong
isolated P. aeruginosa EF‐P and the recombinant protein with in evidence that rhamnosylation of Arg‐32 in P. aeruginosa EF‐P (and
vitro EarP‐transferred rhamnose. Subtle differences were observed
in the fragmentation patterns between the two anomeric
rhamnosylated glycopeptides 14 and 15 (Figure 2B and 2C). This
included a difference in the intensity of the doubly charged
by extension the conserved Arg residue of other bacterial EF‐P
proteins ) is ‐configured.
2
1
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1‐3 | 5
Please do not adjust margins

Pleas Ce hd eo mni oc ta al dS j cu iset n mc ea rgins
Page 6 of 7
EDGE ARTICLE
Chemical Science
2
1
homologue. Finally, we demonstrate that the ‐rhamnosylated
View Article Online
DOI: 10.1039/C6SC03847F
arginine moiety can undergo unusual endocyclic ring opening
anomerization to the ‐anomer. As such, synthetic or recombinant
glycopeptides and glycoproteins bearing ‐rhamnosyl arginine must
be carefully handled to prevent unwanted loss of configurational
purity. Importantly, the work described here lays the foundation
for future synthetic and biochemical studies on bacterial
glycopeptides and glycoproteins, including those bearing the
rhamnosylated arginine motif, with a view to understanding the
importance of glycosylation for bacterial protein structure and
function.
Acknowledgements
This research was supported by an Australian Research Council Future
Fellowship (FT130100150). We would also like to acknowledge the support
of Associate Professor Chris Burns (Walter and Eliza Hall Medical Research
Institute), Dr Ian Luck (The University of Sydney) and Rodney Tollerson (Ohio
State University).
Notes and references
1.
2.
3.
4.
R. Apweiler, H. Hermjakob and N. Sharon, Biochim.
Biophys. Acta, 1999, 1473, 4.
K. W. Moremen, M. Tiemeyer and A. V. Nairn, Nat. Rev.
Mol. Cell. Biol., 2012, 13, 448.
Figure 2. Characterization of rhamnosylated glycopeptides by nanoUHPLC‐
MS/MS. A) Extracted‐ion chromatographic elution profiles of synthetic
glycopeptides 14 and 15, an equimolar mixture of 14 and 15, and Lys‐C
digested EF‐P purified from P. aeruginosa or recombinant EF‐P (from E. coli)
with in vitro rhamnosylation by EarP. The extracted ion mass window was
S. Hakomori, Proc. Natl. Acad. Sci. U. S. A., 2002, 99,
10231.
P. M. Rudd, T. Elliott, P. Cresswell, I. A. Wilson and R. A.
Dwek, Science, 2001, 291, 2370.
349.86‐349.87 m/z corresponding to the triply charged glycopeptide. B) Ion‐
trap CID MS/MS analysis of the ‐linked synthetic glycopeptide. C) Ion‐trap 5.
CID MS/MS analysis of the ‐linked synthetic glycopeptide. D) Ion‐trap CID
MS/MS analysis of the glycopeptide identified from P. aeruginosa purified 6.
P. V. D. Steen, P. M. Rudd, R. A. Dwek and G. Opdenakker,
Crit. Rev. Biochem. Mol. Biol., 1998, 33, 151.
R. J. Payne and C.‐H. Wong, Chem. Commun., 2010, 46,
21.
2+
EF‐P. E) Distribution of the diagnostic fragment ion [M+2H‐rhamnose]
normalized to the precursor ion intensity. **P<0.01 (students t‐test). All b‐ 7.
and y‐type ions are annotated within 0.6 Da and represent the fragment 8.
containing the rhamnose glycan still attached via glycosidic linkage to the
B. G. Davis, Chem. Rev., 2002, 102, 579.
K. K. Palaniappan and C. R. Bertozzi, Chem. Rev. 2016,
DOI: 10.1021/acs.chemrev.6b00023
fragment ion.
9.
S. Gaunitz, G. Nagy, N. L. Pohl and M. V. Novotny, Anal.
Chem., 2016, DOI: 10.1021/acs.analchem.6b04343
M. Thaysen‐Andersen, N. H. Packer and B. L. Schulz, Mol.
Cell. Proteomics, 2016, 15, 1773‐1790.
10.
Conclusions
11.
H. Nothaft and C. M. Szymanski, Nat. Rev. Microbiol.
2
010, 8, 765.
H. L. Tytgat and S. Lebeer, Microbiol. Mol. Biol. Rev., 2014,
8, 372.
N‐linked rhamnosylation has recently been uncovered as an 12.
unusual modification of a conserved Arg residue in bacterial EF‐P
7
13.
T. J. Oman, J. M. Boettcher, H. Wang, X. N. Okalibe and W.
A. van der Donk, Nat. Chem. Biol., 2011, 7, 78.
J. Stepper, S. Shastri, T. S. Loo, J. C. Preston, P. Novak, P.
Man, C. H. Moore, V. Havlíček, M. L. Patchett and G. E.
Norris, FEBS Lett., 2011, 585, 645.
H. Wang, T. J. Oman, R. Zhang, C. V. Garcia De Gonzalo, Q.
Zhang and W. A. van der Donk, J. Am. Chem. Soc., 2013,
136, 84.
proteins that is critical for its role in preventing ribosome stalling
during polyproline translation. In this work we have described an
efficient synthetic route to suitably protected ‐ and ‐
rhamnosylated arginine ‘cassettes’ that can be directly incorporated
into Fmoc‐SPPS for the preparation of rhamnosylated peptides. We
also describe the synthesis of Lys‐C proteolytic fragments of P.
aeruginosa EF‐P bearing both ‐ and ‐rhamnosylated arginine.
14.
15.
Through a combination of 2D NMR and nano‐UHPLC coupled to 16.
tandem mass spectrometry and using isolated and recombinant P.
aeruginosa EF‐P as standards we provide evidence that strongly
suggests that the native protein bears an ‐configured rhamnose
residue on Arg‐32. These findings complement a recent anomeric
coupling constant analysis on the full length S. oneidensis
J. S. Pearson, C. Giogha, S. Y. Ong, C. L. Kennedy, M. Kelly,
K. S. Robinson, T. W. F. Lung, A. Mansell, P. Riedmaier and
C. V. Oates et al., Nature, 2013, 501, 247.
S. Li, L. Zhang, Q. Yao, L. Li, N. Dong, J. Rong, W. Gao, X.
Ding, L. Sun, X. Chen, S. Chen, F. Shao, Nature, 2013, 501,
17.
242.
6
| J. Name., 2012, 00, 1‐3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Page 7 of 7
Pleas Ce hd eo mni oc ta al dS j cu iset n mc ea rgins
Journal Name
COMMUNICATION
18.
J. Lassak, E. C. Keilhauer, M. Fürst, K. Wuichet, J. Gödeke,
A. L. Starosta, J.‐M. Chen, L. Søgaard‐Andersen, J. Rohr, D.
N. Wilson, S. Häussler, M. Mann, K. Jung, Nat. Chem. Biol.,
View Article Online
DOI: 10.1039/C6SC03847F
2015, 11, 266.
19.
A. Rajkovic, S. Erickson, A. Witzky, O. E. Branson, J. Seo, P.
R. Gafken, M. A. Frietas, J. P. Whitelegge, K. F. Faull, W.
Navarre, A. J. Darwin, M. Ibba, MBio, 2015, 6, e00823.
T. Yanagisawa, H. Takahashi, T. Suzuki, A. Masuda, N.
Dohmae and S. Yokoyama, PloS One, 2016, 11, e0147907.
X. Li, R. Krafczyk, J. Macosek, Y. Li, Y. Zou, B. Simon, X.
Pan, Q. Wu, F. Yan, J. Hennig, K. Jung, J. Lassak and H.‐
G. Hu, Chem. Sci., 2016, 7, 6995.
M. Pan, S. Li, X. Li, F. Shao, L. Liu and H.‐G. Hu, Angew.
Chem. Int. Ed., 2014, 53, 14517.
L. A. Carpino, H. Shroff, S. A. Triolo, E.‐S. M. E. Mansour,
H. Wenschuh and F. Albericio, Tetrahedron Lett., 1993, 34,
2
0.
1.
2
2
2.
3.
2
7
829.
24.
25.
26.
27.
W. Chen, L. Gu, W. Zhang, E. Motari, L. Cai, T. J. Styslinger
and P. G. Wang, ACS Chem. Biol., 2010, 6, 185.
Q. Wang, J. Fu and J. Zhang, Carbohyr. Res., 2008, 343,
2
989.
G. Wulff and G. Röhle, Angew. Chem. Int. Ed., 1974, 13,
57.
1
M. T. Blázquez‐Sánchez, F. Marcelo, M. d. C. Fernández‐
Alonso, R. del Villar‐Guerra, A. Samadi, F. J. Cañada, J.
Jiménez‐Barbero and C. Vicent, Eur. J. Org. Chem., 2015,
2
015, 6180.
28.
29.
30.
J. L. Jiménez Blanco, B. Sylla, C. Ortiz Mellet and J. M.
García Fernández, J. Org. Chem., 2007, 72, 4547.
J. Li, G. Zhang, Z. Zhang and E. Fan, J. Org. Chem., 2003,
6
8, 1611.
J. M. Benito, C. O. Mellet, K. Sadalapure, T. K. Lindhorst, J.
Defaye and J. M. G. Fernández, Carbohydr. Res., 1999,
3
20, 37.
31.
Y. Zhang, S. M. Muthana, D. Farnsworth, O. Ludek, K.
Adams, J. J. Barchi and J. C. Gildersleeve, J. Am. Chem.
Soc., 2012, 134, 6316.
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1‐3 | 7
Please do not adjust margins