Site-selective azide incorporation into endogenous RNase A via a "chemistry" approach

Article abstract of DOI:10.1039/c2ob26561c

Site-selective labeling of endogenous proteins represents a major challenge in chemical biology, mainly due to the absence of unique reactive groups that can be addressed selectively. Recently, we have shown that surface-exposed lysine residues of two end

Full text of DOI:10.1039/c2ob26561c

Organic &
Biomolecular
Chemistry
View Article Online
View Journal | View Issue
PAPER
Site-selective azide incorporation into endogenous
RNase A via a “chemistry” approach†
Cite this: Org. Biomol. Chem., 2013, 11,
353
Xi Chen,a,b Lars Henschke,a Qianzhen Wu,a Kasturi Muthoosamy,b Boris Neumannc
and Tanja Weil*a,b
Site-selective labeling of endogenous proteins represents a major challenge in chemical biology, mainly
due to the absence of unique reactive groups that can be addressed selectively. Recently, we have shown
that surface-exposed lysine residues of two endogenous proteins and a peptide exhibit subtle changes in
their individual reactivities. This feature allows the modification of a single residue in a highly site-selec-
tive fashion if kinetically controlled labeling conditions are applied. In order to broaden the scope of the
“kinetically-controlled protein labeling” (KPL) approach and highlight additional applications, the water-
soluble bioorthogonal reagent, biotin–TEO–azido–NHS (11), is developed which enables the site-selective
introduction of an azido group onto endogenous proteins/peptides. This bioconjugation reagent fea-
tures a biotin tag for affinity purification, an azido group for bioorthogonal labeling, a TEO (tetra-
ethylene oxide) linker acting as a spacer and to impart water solubility and an N-hydroxysuccinimidyl
(NHS) ester group for reacting with the exposed lysine residue. As a proof of concept, the native protein
ribonuclease A (RNase A) bearing ten available lysine residues at the surface is furnished with a single
azido group at Lys 1 in a highly site-selective fashion yielding azido–(K1)RNase A. The K1 site-selectivity is
demonstrated by the combined application and interpretation of high resolution MALDI-ToF mass
spectroscopy, tandem mass spectroscopy and extracted ion chromatography (XIC). Finally, the water
soluble azide-reactive phosphine probe, rho–TEO–phosphine (21) (rho: rhodamine), has been designed
and applied to attach a chromophore to azido–(K1)RNase A via Staudinger ligation at physiological pH
indicating that the introduced azido group is accessible and could be addressed by other established
azide-reactive bioorthogonal reaction schemes.
Received 7th August 2012,
Accepted 30th October 2012
DOI: 10.1039/c2ob26561c
www.rsc.org/obc
interactions.6 In vitro, the site-specific chemical modification
of proteins represents a key tool for numerous biochemical
Introduction
Over the past few years, the site-selective modification of pro- studies7,8 and has earned also high commercial interest
teins has received increasing interest in chemical biology.1–4 for the development of next generation biotherapeutics
Many important in vivo and ex vivo events are triggered by with improved pharmacokinetic properties for disease
modification processes that alter the structure and function of treatment.9,10
proteins.5 These processes are commonly denoted as post-
Even though PTMs play a crucial role in many areas and
translational modifications (PTMs) and they play a crucial role they can be readily accomplished in living organisms,11 the
in many cellular events such as the in vivo phosphorylation, site-selective chemical modification of endogenous proteins
which regulates endocytosis and downstream protein–protein still represents a major challenge. Since many copies of each
amino acid are present in a typical protein, it is often con-
sidered impossible to produce a homogeneously functiona-
aInstitute of Organic Chemistry III/Macromolecular Chemistry and Biological
lized variant.12 In contrast, recombinant proteins can be
Chemistry, University of Ulm, Albert-Einstein 11, 89069 Ulm, Germany.
furnished with reactive functionalities via point mutations
E-mail: tanja.weil@uni-ulm.de; Fax: (+)49-73150-22883; Tel: (+)49-73150-22880
such as cysteine residues that allow the site-specific attach-
ment of further functionalities.13 In addition, non-canonical
amino acids (NCAA) bearing reactive functionalities could be
introduced by more elaborate biological techniques such as
residue-specific NCAA incorporation14,15 or nonsense codon
suppression16,17 that facilitate bioorthogonal post-modifi-
cations such as 1,3 Huisgen cycloaddition reactions (“click
bDepartment of Chemistry, National University of Singapore, 3 Science Drive 3,
117543, Singapore
cProteome Factory AG, Magnusstr. 11, Haus 2, 1. OG, 12489 Berlin, Germany
†Electronic supplementary information (ESI) available: The auto-syringe pump
setup in KPL, RNase A functional assay, HRMS (ESI) spectra of K1 modified
KETAAAKF (1–8), interpretation of the false positive signals in Fig. 4a, the argon
column chromatography setup, details of organic synthesis as well as the full
NMR spectra of all novel compounds. See DOI: 10.1039/c2ob26561c
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem., 2013, 11, 353–361 | 353

View Article Online
Paper
Organic & Biomolecular Chemistry
reactions”)18–20 or Sonogashira cross-coupling reactions21,22 on high labeling yield. The water soluble azido-reactive phosphine
proteins. Other strategies involve enzymatic posttranslational probe, rho–TEO–phosphine (21), is prepared demonstrating
modifications of proteins.23,24 These biological approaches are the potential to site-selectively introduce additional functional
nowadays widely used but they can be technically more groups into azido–(K1)RNase A via Staudinger ligation at physio-
demanding and they often require several optimization cycles logical pH in PBS buffer. In addition, we have shown pre-
until the protein of choice is expressed and purified in accep- viously that KPL allows modifying lysozyme and somatostatin
table quantities.
as well again stressing the potential of our approach for the
Labeling of native proteins via chemical approaches usually introduction of azido groups into proteins other than RNase A.
lacks site-selectivity since addressing a single functional group The approach reported herein is fast, convenient and mini-
is challenging and heterogeneous protein mixtures are often mizes the risk for protein misfolding. It complements the
obtained. A few attempts have been proven successful such as portfolio of recombinant techniques or chemoenzymatic
targeting accessible disulfide bonds25 or the N-terminus of a approaches and opens up various opportunities for precisely
protein.26–29 However, possible limitations of these approaches functionalizing endogenous proteins and peptides with azido
include the necessity to apply non-physiological reaction con- groups.
ditions, lack of accessible disulfides for intercalation or
N-termini not suitable for further chemical modifications. In
addition, heterogeneous mixtures containing both native and
Experimental section
modified proteins are often obtained, and the purification of
Methods and materials
the modified protein can be challenging. Recently, affinity-
Ribonuclease A was obtained from MP Biomedicals, LLC (Cat.
No. 101076). Pierce Monomeric Avidin Agarose for affinity
chromatography was from Thermo Scientific (Cat. No. 20267).
MALDI-ToF-MS spectra of protein samples were recorded on a
Bruker Daltonics mass spectrometer using sinapinic acid as
the matrix. The analysis of the protein modification site by
nanoLC-MS2(ESI) was accomplished at Proteome Factory AG
(Germany). NuPAGE® Novex 4–12% Bis-Tris gel and its corres-
ponding NuPAGE MES SDS running buffer for gel electropho-
resis were purchased from Invitrogen. Phosphate buffered
saline, ultra-pure grade, was purchased from 1st BASE (Cat.
No. 2040). A Pierce BCA assay kit for protein quantifications
was obtained from Novagen (Cat. No. 71285-3). Baker’s yeast
ribonucleic acid (S. cerensiae) was purchased from Sigma-
Aldrich (Cat. No. R6750). SYBR® Safe DNA gel stain (10 000×
concentrated) in DMSO was obtained from Invitrogen Molecular
Probes (USA). Ultra-pure grade Tris (Cat. No. 1400) and ultra-
pure grade EDTA (Cat. No. 1050) were purchased from 1st
BASE. 384-Well black microtiter plates for RNase A functional
assay and sterile 96-well plates were received from Greiner Bio-
One Ltd (UK) and Nunclon Surface (Denmark), respectively.
guided concepts based on the specificity of ligand–receptor
interactions have received considerable attention for the site-
specific modification of endogenous proteins.30–33
We have recently reported the concept that surface-exposed
lysine residues of native proteins and peptides can reveal
subtle differences of their individual reactivity, which allows
introducing reactive groups in a highly site-selective fashion.
This “kinetically-controlled protein labeling” (KPL) approach has
been applied successfully to furnish the model proteins RNase
A, lysozyme C and the peptide hormone somatostatin-14 with
a single biotin ligand in a highly site-selective fashion.34 In
addition, the protein RNase A was also functionalized with an
ethynyl group.34 In order to further extend this bioconjugation
scheme to the attachment of even more attractive and versatile
groups, we have developed herein the novel azido-containing
biotin-tagged bioorthogonal reagent biotin–TEO–azido–NHS
(11) for the site-selective introduction of a bioorthogonal azido
group onto proteins and peptides. Biotin-tagged affinity
conjugation agents have received considerable attention in
chemical biology, particularly when an affinity separation is
required.35–37 Besides, the azido substituent reacts with a high
chemoselectivity in the presence of proteinogenic groups via
e.g. copper-free click labeling38 and Staudinger ligation label-
ing;39,40 both approaches have even been performed in living
Protein modifications
(A) “KINETICALLY-CONTROLLED PROTEIN LABELING” FOR THE PREPARATION
organisms. Therefore, the azide-mediated “tag-and-modify”41 OF AZIDO–(K1)RNASE A (13). A solution of the in situ prepared
concept that this functionality offers allows chemical modifi- biotin–TEO–azido–NHS (11) (0.1
M in DMF) (58.4 μL,
cations under rather mild conditions and is particularly suited 5.84 μmol, 2 equiv.) was diluted in 2 mL of cold DMF (anhy-
for sensitive proteins. In addition, the azido group also re- drous). The resultant DMF solution was injected dropwise
presents an interesting IR probe for elucidating protein into the stirring RNase A (12) solution (40 mg, 2.92 μmol,
folding, protein dynamics and electrostatics42 which further 1.0 equiv.) in PBS buffer (0.1 M, pH 7.2) (40 mL) via an auto-
stresses the significance of decorating proteins with azides.
syringe pump during 200 min at +4 °C in a cold room
In this contribution, the protein RNase A is furnished with (Fig. S1†). After the injection was completed, the resultant
an azido group site-selectively at Lys 1 and further modified protein solution was further stirred at +4 °C overnight (step I:
via Staudinger ligation following an entirely chemistry-based kinetic conjugation). Ethanolamine solution (7.84 mg,
scheme for RNase functionalization. The novel azide-contain- 128 μmol, 43.8 equiv.) was added and the solution was stirred
ing reagent biotin–TEO–azido–NHS (11) has been developed at RT for an additional 9 h in order to quench any unreacted
that reacts with native RNase A under KPL conditions with a NHS ester (12) (step II: quenching step). The protein solution
354 | Org. Biomol. Chem., 2013, 11, 353–361
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
was freeze dried, re-constituted in 4 mL of MQ-H2O and dia- of azido–(K1)RNase A (13) only display a non-fluorescent gel
lyzed against MQ-H2O (MWCO 3.5 kDa) overnight in order to band at ∼14 kDa.
remove biotin-containing small molecules that could perturb
the subsequent affinity purification (step III: 1st biotin-removal
step). The obtained protein solution (∼8 mL) exclusively con-
tained azido–(K1)RNase A (13) and excess of unreacted native
RNase A (12). Due to the loading capacity limit, 0.5 mL of this
solution was loaded into a Pierce Monomeric Avidin Resin
Results and discussion
Design of the key building blocks for KPL of endogenous
RNase A
Column (2 mL of resin supplied, Thermo Scientific, Prod.
#20267) and the column was eluted by PBS buffer (pH 7.2),
yielding the recovered RNase A solution and quantified by BCA
assay (2.28 mg) (step IV: recovery step). Afterwards, the column
was eluted by a 2 mM (+)-biotin solution in PBS buffer
(pH 7.2), giving mono-modified RNase A (13) (step V: affinity
elution). The resultant azido–(K1)RNase A (13) solution con-
tained undesired (+)-biotin molecules and needed to be
removed via a second dialysis (MWCO 3.5 kDa) (step VI:
2nd biotin-removal step). The obtained protein solution was
azido–(K1)RNase A (13) which was quantified by BCA assay
(195 μg) and then lyophilized to yield the targeted azido–(K1)
RNase A (13). On account of recovered RNase A, the labeling
yield was calculated to be 84%. MALDI-ToF-MS: m/z 14 336
[M + H]+.
RNase A is selected as a model protein since it is stable, well-
characterized and of therapeutic relevance as one of its close
variants, onconase, is currently under clinical investigations as
a promising cancer therapeutic.43,44 Endogenous RNase A con-
tains as many as ten lysine residues on its surface, and accord-
ing to solvent accessibility45 calculations, Lys 1 displays the
highest solvent accessibility.34 However, except for Lys 41, all
other lysine residues of RNase A are located on the protein
surface and should still be readily accessible (Fig. 1a). KPL
could be applied if one Lys residue has a higher reactivity com-
pared to other Lys residues based on, e.g., a better solvent
accessibility or activating or deactivating interactions of neigh-
boring amino acids.34 In this case, by adding a low concen-
tration of the bioconjugation reagent to a large excess of native
protein over an extended period of time, the most reactive
amino acid can be labeled selectively. Herein, the novel multi-
functional reagent (11) has been designed that offers the site-
selective incorporation of an azido group via KPL (Fig. 1b). In
addition, the water-soluble azide-reactive probe, rho–TEO–
phosphine (21), facilitates efficient post-modifications of
azido-containing proteins in aqueous solution at physiological
pH via Staudinger ligation to prove the accessibility of the
newly introduced azido group (Fig. 1c). The two steps allow an
endogenous protein to be converted to its site-selectively modi-
fied variants (Fig. 1d) where all related bioconjugation and
(B) STAUDINGER LIGATION LABELING OF AZIDO–(K1)RNASE A (13)
USING
RHO–TEO–PHOSPHINE (11). Rho–TEO–phosphine (21)
(1.2 mg, 992 nmol) wetted by DMSO was added into azido–
(K1)RNase A (13) (100 μg, 3.5 nmol) solution in PBS buffer
(200 μL, 1×, pH 7.2) under Ar in a small glass vial. The resul-
tant violet solution was shaken under Ar at 37 °C for 24 h. Gel-
electrophoresis reveals a highly fluorescent band at ∼15 kDa
suggesting the rhodamine was covalently attached onto azido–
(K1)RNase A via Staudinger ligation. Pure azido–(K1)RNase A
(13) and the control reaction using native RNase A (12) instead
Fig. 1 The crystal structure of RNase A (PDB file: 1RCA.pdb) with nine out of ten lysine residues highlighted as a yellow ball-and-stick model (above); calculated
solvent accessibility of each lysine residue where lysine 1 (K1) (violet sphere) gives the highest value (below); (b) the chemical structure of reagent 11—biotin–TEO–
azido–NHS; (c) the chemical structure of the phosphine probe—rho–TEO–phosphine (21); (d) the “kinetic protein labeling” approach for site-selectively introducing
an azido handle onto endogenous RNase A (12) using reagent 11 and further Staudinger ligation labeling of azido-functionalized RNase A by phosphine probe 21.
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem., 2013, 11, 353–361 | 355

View Article Online
Paper
Organic & Biomolecular Chemistry
Fig. 2 (a) Design of biotin–TEO–azido–NHS (11) featuring a biotin tag for affinity purification, a TEO-spacer for imparting water solubility, an azido group for
bioorthogonal labeling and an amine-reactive NHS ester for targeting lysine. (b) The total synthetic scheme of biotin–TEO–azido–NHS (11): (i) NaN3, H2O, 75 °C,
20 h, then NaOH (aq), 60%. (ii) BrCH2COOtBu, K2CO3, DMF, RT, 48 h, 68%. (iii) Succinic anhydride, DMF, RT, overnight, 81%. (iv) (Boc)2O, dioxane, RT, overnight, 34%.
(v) (+)-Biotin, EDC·HCl, DMF, RT, 73%. (vi) TFA, DCM, RT, 30 min then NH3 (aq), 99%. (vii) 4, EDC·HCl, DMF, RT, 48 h, 69%. (viii) TFA, DCM, RT, 6 h, 93%. (ix) NHS,
EDC·HCl, DMF, RT, 24 h, 99%.
purification steps proceed under very mild conditions (pH 7.2) been characterized in detail by HRMS, LC-MS and NMR spec-
and without the necessity to apply recombinant techniques.
troscopy. Since biotin–TEO–azido–COOH features a tertiary
amide bond, both of its 1H-NMR and carbon NMR spectra
show two sets of signals due to the presence of rotamers.
Synthesis and characterization of biotin–TEO–azido–NHS (11)
1
Therefore, 2D-NMR including H–1H COSY and HMQC as well
The multifunctional, bioorthogonal linker biotin–TEO–azido–
NHS (11) (TEO: tetraethylene oxide; NHS: N-hydroxysuccinim-
idyl ester) has been designed to meet several unique character-
istics such as a biotin tag for affinity purification of proteins, a
TEO spacer for imparting water solubility and an amine-reac-
tive NHS ester group for targeting lysine residues as well as the
azido group as a reactive handle (Fig. 2a). The total synthesis
as DEPT 135 spectra have been recorded to fully assign all the
NMR signals (see ESI†).
Synthesis of the water soluble azide-reactive probe—rho–TEO–
phosphine (21)
scheme is depicted in Fig. 2b. Briefly, 2-azidoethylamine (2)46 Subsequently, the water soluble phosphine probe—rho–TEO–
is prepared via nucleophilic attack of 2-bromoethylamine (1) phosphine (21) has been designed featuring an azido-reactive
with sodium azide under heating. Then, 1 equiv. of tert-butyl- phosphine group, a water soluble TEO spacer and a rhodamine
2-bromoacetate is treated with 2 equiv. of 2-azidoethylamine B chromophore (Scheme 1). Previously, similar rhodamine–
(2) under basic conditions to afford tert-butyl 2-azidoethyl- phosphine conjugates have been reported and applied, e.g. as
aminoacetate (3) which is subsequently acylated by succinic bioorthogonal chemical reporters.47 The rho–TEO–phosphine
anhydride to yield 4. In parallel, 4,7,10-trioxa-1,13-tridecane- presented herein features a highly hydrophilic tetraethylene
diamine (5) is protected to yield N-Boc-4,7,10-trioxa-1,13-tri- oxide (TEO) linker for water solubility. Synthesis of 21 is
decanediamine (6) and then coupled with (+)-biotin via accomplished in six steps. First, rhodamine B piperazine
EDC (1-ethyl-3-(3-dimethylaminopropyl)carbodiimide) conden- amide48 is prepared from rhodamine B (14) using a new pro-
sation to obtain biotin–TEO–NH(Boc) (7). Afterwards, 7 is cedure via a rhodamine B acyl chloride intermediate (15).
deprotected by trifluoroacetic acid (TFA) to give biotin–TEO– Then, 1-methyl-2-iodoterephthalate (17) is coupled to diphe-
NH2 (8), which is then coupled with 4 in the presence of nylphosphine under the catalysis of palladium acetate to yield
EDC to yield 9. Subsequently, 9 is deprotected by TFA to 2-diphenylphosphino-1-methylterephthalate (18).49 Thereafter,
afford biotin–TEO–azido–COOH (10). Finally, 10 is coupled this phosphine intermediate is connected to the water soluble
with N-hydroxysuccinimide to give the desired bioorthogonal linker, tert-butyl 12-amino-4,7,10-trioxadodecanate, using
reagent 11 in its activated NHS ester form. Biotin–TEO–azido– HBTU (O-benzotriazol-1-yl-N,N,N′,N′-tetramethyl-uronium hexa-
NHS (11) is synthesized in nine steps with an overall fluorophosphate) as the coupling reagent. After deprotection
yield of 30% starting from 2-bromoethylamine hydrobromide of the tert-butyl ester group, the resultant free acid (20) is
salt (1).
attached to rhodamine B piperazine amide (16) via HBTU
Since the final bioconjugation reagent biotin–TEO–azido– coupling yielding the rho–TEO–phosphine (21) azide probe in
NHS (11) represents an active species and displays only limited an overall yield of 21% starting from 2-iodo-1-methyltere-
stability, the free acid form, biotin–TEO–azido–COOH (10), has phthalate (17).
356 | Org. Biomol. Chem., 2013, 11, 353–361
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
(Fig. 3b, middle). It is noteworthy that no multi-modified side-
products have been detected. Ethanolamine is added to the
reaction mixture to quench any remaining unreacted NHS
ester and the solution is purified on the monomeric avidin
resin. Azido–(K1)RNase A (13) which carries a biotin motif is
reversibly bound into the resin while excess of unreacted
RNase A is recovered by eluting the resin column with PBS
buffer (pH 7.2). Thereafter, azido–(K1)RNase A (13) is eluted
from the resin with PBS buffer (pH 7.2) containing 2 mM of
(+)-biotin. Finally, the free biotin is removed by gel filtration,
yielding azido–(K1)RNase A (13) which is quantified by the
BCA assay. 13 exhibits a single peak with m/z 14 336 [M + H]+
matching the calcd M.W. of 14 335 g mol−1 (Fig. 3b, right).
Considering the amount of recovered RNase A, the labeling
yield is about 84% and minor product loss might be due to
the purification steps including dialysis.
It is noteworthy that the total amount of the bioconjugation
reagent relative to the amount of protein required for this re-
action step does not represent a fixed number but merely
depends on the properties of the reagent itself (e.g. its water-
solubility, reactivity, steric hindrance of the reactive group) as
well as the protein of choice. Therefore, it is advisable that a
kinetic labeling study is conducted in order to identify the
optimal reagent/protein ratio. This kinetic labeling study of a
new reagent/protein pair could be easily investigated via
MALDI-ToF-MS spectrometry. In the case that a less water
soluble or less reactive labeling reagent is used, the required
amount of the bioconjugation reagent could also be less than
e.g. one equiv. However, if the reagent is very water-soluble but
of limited stability in water such as (11), which is deactivated
during the labeling process, the required quantity to achieve
an optimal labeling result exceeds one equiv. In this study, the
application of 0.5 equiv. of biotin–TEO–azido–NHS (11) results
in a very low labeling ratio and very little amounts of azido–
(K1)RNase A (13) are formed (data not shown). Therefore,
2 equiv. of 11 have been used thus yielding an improved label-
ing efficiency (Fig. 3b, middle).
RNASE A FUNCTIONAL ASSAY. The enzymatic activities of both
azido–(K1)RNase A and recovered RNase A relative to native
RNase A are determined by an RNase A functional assay34 in
order to understand the effects of the azide-modification as
well as the optimized KPL procedures on the activity of this
enzyme. According to the experimental results, the retained
enzymatic activity of the K1 functionalized RNase A variant,
azido–(K1)RNase A (13), is about 29% of native RNase A
(Fig. S3†) which is lower than the retained activity of biotin–
LC–(K1)RNase A (88%)34 which carries a smaller biotin substi-
tuent. This activity loss often occurs after chemical modifi-
cation and in the case of RNase A, it might be due to the fact
that the modified Lys 1 residue is located in close vicinity to
the active site of RNase A and a more bulky group (e.g. biotin–
TEO–azido moiety) might further reduce the enzymatic
activity. However, the introduction of functional groups close
to the active center could be very attractive for achieving
responsive proteins whose activity could be switched upon an
external stimulus e.g. by attaching responsive polymers.50 It is
Scheme 1 Synthesis of the water soluble rho–TEO–phosphine probe (21):
(i) SOCl2, RT, overnight. (ii) Piperazine, DCM, RT, overnight, 31% from 14.
(iii) HPPh2, Pd(OAc)2, K2CO3, MeCN, 85 °C, 24 h, 86%. (iv) tert-Butyl 12-amino-
4,7,10-trioxadodecanate, HBTU, DIEA, DMF, RT, 48 h, 60%. (v) TFA, DCM, RT,
overnight, 99%. (vi) HBTU, rhodamine B piperazine amide (16), DIEA, DMF, RT,
48 h, 42%.
Preparation of azido–(K1)RNase A via an optimized
“kinetic protein labeling” (KPL) procedure
PREPARATION OF AZIDO–(K1)RNASE A VIA AN OPTIMIZED KPL SCHEME.
The general procedures of the KPL (Scheme S1) are given in
the ESI,† featuring six essential steps. In comparison to our
previous KPL study,34 an optimized labeling procedure has
been applied herein for the preparation of site-selectively
mono-functionalized azido–(K1)RNase A. First, an increased
kinetic bioconjugation time (from 100 min to 200 min) and an
overnight incubation after labeling (instead of 1 h of incu-
bation) were applied in order to increase the labeling
efficiency of 11. As a result, only 2 equiv. of the bioconjugation
reagent compared to 8 equiv. used in the previous protein
ethynylation were applied that allowed saving the precious
reagent 11. Additionally, an auto-syringe pump set-up
(Fig. S1†) as opposed to tedious manual addition used in a pre-
vious biotinylation protocol34 has been applied which guaran-
teed higher reproducibility. Dialysis was chosen instead of gel
filtration which better suits a larger scale reaction. Further-
more, in step I, the bioconjugation reaction is performed at a
reduced temperature (+4 °C) in a cold room and not at RT to
further increase the site-selectivity.
In short, 2 equiv. of biotin–TEO–azido–NHS (11) in DMF
solution (1/20 v/v relative protein solution volume) are injected
dropwise into the RNase A solution at +4 °C in PBS buffer (pH
7.2) during 200 min via an auto-syringe pump yielding mono-
modified RNase A co-existing with native RNase A which was
used in excess (Fig. 3a). Herein, minor amounts of the organic
solvent DMF are applied in order to reduce the autohydrolysis
of 11 in aqueous solution. Even though 2 equiv. of 11 are
applied, it is still kept in deficiency since 11 reveals auto-
hydrolysis in water and is partially deactivated before reacting
with RNase
A
(12). According to the MALDI-ToF-MS
spectrum, only mono-modified RNase A (13) is formed in the
presence of native RNase A, which has been used in excess
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem., 2013, 11, 353–361 | 357

View Article Online
Paper
Organic & Biomolecular Chemistry
Fig. 3 (a) “Kinetic protein labeling” scheme of the site-selective introduction of an azido group at K1 of RNase A (12) applying biotin–TEO–azido–NHS (11);
(b) MALDI-ToF-MS spectra of native RNase A (12) (left), KPL reaction mixture (middle) and purified mono-modified azido–(K1)RNase A (13); (c) full interpretation of
the MS2 spectrum of modified KETAAAKF (1–8), suggesting that the modification site occurred at K1 rather than K7.
Fig. 4 (a) The XIC chromatographs of modified KETAAAKF (1–8) (top, 2+ and 3+) as well as unmodified KETAAAKF (1–8) (bottom, 1+ and 2+) proving the full
modification at K1; (b) the MALDI-ToF-MS spectrum of azido–(K1)RNase A revealing mono-modification of RNase A. Both data demonstrate the site-selective modifi-
cation at K1.
noteworthy that the recovered RNase A displays the same enzy- recorded in order to differentiate the two sites (Fig. 3c). All the
matic activity of native RNase A suggesting that the optimized b and y ions of this peptide fragment have been elucidated
KPL procedure and the purification protocol have no negative and are marked in the MS2 spectrum, proving that the modifi-
impact on the activity of RNase A (Fig. S3†).
cation occurred at K1 rather than K7.
IDENTIFICATION OF THE LOCATION OF THE AZIDO-GROUP AT THE SURFACE
LABELING SITE-SELECTIVITY AT K1. Some of the peptide fragments
OF AZIDO–(K1)RNASE A (13). Identification of the modification containing free thiol groups cannot be determined since they
site is achieved by using nanoLCMS2.34 Briefly, this protein are generally lost after Michael addition with the polyacryl-
sample is subjected to gel-electrophoresis, in-gel digested by amide gel during in-gel digestion. In order to circumvent this
chymotrypsin and the resultant peptide fragments are ana- limitation, the extracted ion chromatographs (XIC) of both
lyzed by nanoLC-MS2. The peptide fragment, KETAAAKF (1–8), modified KETAAAKF (1–8) (Fig. 4a top, [M + 2H]2+ and
has been identified to contain the biotin–TEO–azido moiety [M + 3H]3+) and unmodified KETAAAKF (1–8) (Fig. 4a bottom,
(calcd M.W. of modified KETAAAKF: 1518.790 g mol−1) due to [M + H]+ and [M + 2H]2+) fragments have been recorded. While
the presence of the parent ions that match m/z 760.403 the modified KETAAAKF (1–8) shows intensive peaks in its XIC
[M + 2H]2+ and m/z 507.271 [M + 3H]3+ (Fig. S4†). Since this chromatograph at tR 16.22 min, unmodified KETAAAKF (1–8)
peptide fragment contains two lysine residues, i.e. Lys 1 and peaks have not been identified. The peak at tR 13.05 min in
Lys 7, the MS2 spectrum of this peptide fragment has been Fig. 4a, and the peaks at tR 28.28 min and tR 29.73 min in
358 | Org. Biomol. Chem., 2013, 11, 353–361
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
Fig. 5 (a) Staudinger ligation of azido–(K1)RNase A (13) with the rho–TEO–phosphine probe (21); (b) gel-electrophoresis image of the Staudinger ligation labeling
reaction (left), pure azido–(K1)RNase A (middle) and the control reactions using native RNase A (12) instead of azido–(K1)RNase A (13) to react with rho–TEO–phos-
phine (21) (right) (above: fluorescence; below: Coomassie blue stain; Pre-Stained Protein Marker II (AppliChem) was used as the gel ladder).
Fig. 4a (bottom) represent false positive signals due to their complements recombinant techniques. Compared to other
unmatched charges (see ESI, Fig. S4†). This information col- endogenous protein labeling methods, KPL reveals both
lected from XIC proved the full modification of the K1 residue advantages and limitations. On the one hand, the KPL labeling
of RNase A. As RNase A is mono-modified (Fig. 4b), the com- scheme is straightforward and a modified protein variant can
bined information demonstrates the K1 site-selectivity of the be readily achieved directly from the endogenous protein. As a
“kinetic protein labeling” approach yielding azido–(K1)RNase A result, high quantities of modified proteins might become
(13). However, this study does not exclude the possibility of accessible in particular if removal of the native protein is not
other lysine residues than K1 that might have been modified crucial for the downstream reactions (e.g. in the preparation of
to a very low extent which is below the detection limit of protein biohybrids and high performance protein chips etc.
MALDI-ToF-MS spectrometry.
where only modified protein can react). Besides, compared to
other techniques, e.g. N-terminal transamination which
employs highly reactive pyridoxal-5-phosphate (PLP),27 disul-
fide intercalation25 which necessitates a reducing reagent to
reduce the disulfide bond(s) or others, both the labeling and
purification steps of the KPL can be performed under mild,
physiological pH conditions.
On the other hand, the synthesis of the bioconjugation
reagent can be tedious and often requires multistep reactions.
In addition, two to four days are required to accomplish all
essential synthesis and purification steps in KPL and upscal-
ing in a batch process has not yet been demonstrated.
Additionally, the KPL labeling only occurs at the most reactive
lysine residues which might not be the desired site. However,
this challenge might be addressed when the origin of the site-
selectivity is better understood and more reactive lysine resi-
dues can be introduced at the desired position by point
mutations.
Staudinger ligation of azido–(K1)RNase A (13) by rho–TEO–
phosphine (21)
In the last step, the accessibility of the azido group on azido–
(K1)RNase A (13) for the Staudinger ligation modification as
well as the azide-reactivity of the rho–TEO–phosphine (21) is
investigated. For this purpose, the azido–(K1)-RNase A (13) is
labelled with the probe (21) under an Ar atmosphere at 37 °C
for 24 hours (Fig. 5a). The resultant protein solution is analysed
via gel electrophoresis under UV irradiation to test if the rhod-
amine chromophore has been covalently attached to azido–(K1)
RNase A (Fig. 5b). According to the gel image, the protein
product displays a gel band with intense emission (Fig. 5b, lane
1). On the other hand, azido–(K1)RNase A (13) (Fig. 5b, lane 2)
and native RNase A (12) as the control which carries no azido
group (Fig. 5b, lane 3) only reveal non-fluorescent gel bands.
This study demonstrates the availability of the introduced azido
group for subsequent Staudinger ligation in aqueous media at
physiological pH as well as the reactivity of the phosphine
probe, rho–TEO–phosphine (21), toward azides.
Conclusions
In summary, the synthesis of the novel bioorthogonal linker,
biotin–TEO–azido–NHS (11), has been developed allowing the
site-selective introduction of a single azido group on proteins
Discussion of the advantages and limitations of the KPL
approach
KPL represents a straightforward and efficient approach to via the KPL approach. The scheme of the KPL approach con-
site-selectively modify endogenous proteins and therefore it tains in total six key steps after which azido–(K1)RNase A (13)
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem., 2013, 11, 353–361 | 359

View Article Online
Paper
Organic & Biomolecular Chemistry
is achieved from native RNase A (12). Azido–(K1)RNase A (13)
contains a single azido group as shown by the MALDI-ToF-MS
Acknowledgements
spectrum and the modification site is located at K1 as demon- We acknowledge funding from the German Research Foun-
strated by nanoLC-MS2 analysis which is consistent with pre- dation (DFG) via the SFB 625 as well as grant P3246029 and
vious biotinylation results.34 K1 site-selectivity has been the Volkswagenstifung (Aktenzeichen 86 366) for their finan-
further verified from the XIC chromatographs of both modi- cial support of this work.
fied and unmodified KETAAAKF (1–8) peptide fragments
suggesting that all detected K1 fragments have been modified.
In order to assess the reactivity of the introduced azido group,
the water soluble azide-reactive chromophore, rho–TEO–phos-
phine (21), has been prepared facilitating the successful label-
Notes and references
ing of azido–(K1)RNase A (13) yielding a fluorescent rho–(K1)
RNase A (22) variant.
1 T. L. Foley and M. D. Burkart, Curr. Opin. Chem. Biol., 2007,
11, 12–19.
The work described herein represents an extension of our
previous attempt to introduce a bioorthogonal ethynyl group
for click labeling via the KPL approach.34 The present study
aims at introducing the more attractive bioorthogonal azido
group to further broaden the general applicability of KPL. The
azido group has proven to be an ideal bioorthogonal anchor
since aside from the copper(I)-catalyzed click reaction that
both azido and ethynyl can participate in,20 azido groups also
allow other attractive posttranslational modifications such as
copper-free click reactions38 or Staudinger ligations that
proceed under mild physiological conditions which can be
performed even in living animals. The introduction of the
azido group by this new bioconjugation reagent, biotin–TEO–
azido–NHS (11), has been accomplished following an opti-
mized procedure which allowed increasing the labeling
efficacy.
2 M. B. Francis and I. S. Carrico, Curr. Opin. Chem. Biol.,
2010, 14, 771–773.
3 L. Yi, H. Sun, A. Itzen, G. Triola, H. Waldmann, R. S. Goody
and Y. W. Wu, Angew. Chem., Int. Ed. Engl., 2011, 50, 8287.
4 M. Hejjaoui, S. Butterfield, B. Fauvet, F. Vercruysse, J. Cui,
I. Dikiy, M. Prudent, D. Olschewski, Y. Zhang, D. Eliezer
and H. A. Lashuel, J. Am. Chem. Soc., 2012, 134, 5192.
5 E. Basle, N. Joubert and M. Pucheault, Chem. Biol., 2010,
17, 213–227.
6 H. B. Moeller, J. Praetorius, M. R. Rutzler and R. A. Fenton,
Proc. Natl. Acad. Sci. U. S. A., 2010, 107, 424–429.
7 X. Michalet, S. Weiss and M. Jager, Chem. Rev., 2006, 106,
1785–1813.
8 J. Y. Yang and W. Y. Yang, J. Am. Chem. Soc., 2009, 131,
11644–11645.
9 J. M. Harris and R. B. Chess, Nat. Rev. Drug Discovery, 2003,
2, 214–221.
Modified azido-proteins represent attractive macromolecu-
lar building blocks that offer the great opportunity to attach 10 S. J. Bell, C. M. Fam, E. A. Chlipala, S. J. Carlson, J. I. Lee,
a variety of functional groups e.g. via Staudinger ligation.
This opens up exciting prospects for achieving protein
M. S. Rosendahl, D. H. Doherty and G. N. Cox, Bioconjugate
Chem., 2008, 19, 299–305.
libraries in a straightforward fashion to allow a deeper inves- 11 G. Walsh, Drug Discovery Today, 2010, 15, 773–780.
tigation of the impact of unnatural substituents such as 12 S. D. Tilley and M. B. Francis, J. Am. Chem. Soc., 2006, 128,
polymer chains, lipids or peptides on cell uptake, cell
1080–1081.
trafficking and cytotoxicity. Such investigations could have a 13 Y. Kim, S. O. Ho, N. R. Gassman, Y. Korlann, E. V. Landorf,
high impact on the rational design of improved protein or
protein–polymer therapeutics. In this context, it would be
F. R. Collart and S. Weiss, Bioconjugate Chem., 2008, 19,
786–791.
attractive to further improve the therapeutic potential of 14 A. J. Link, M. L. Mock and D. A. Tirrell, Curr. Opin. Biotech-
clinically relevant RNase A variants by introducing the appro- nol., 2003, 14, 603–609.
priate post-modifications by high content protein synthesis. 15 K. L. Kiick, E. Saxon, D. A. Tirrell and C. R. Bertozzi, Proc.
The novel bioorthogonal linker presented herein could be Natl. Acad. Sci. U. S. A., 2002, 99, 19–24.
applied in principle for the site-selective introduction of an 16 L. Wang, A. Brock, B. Herberich and P. G. Schultz, Science,
azido group onto other proteins and peptides as well and we 2001, 292, 498–500.
have already demonstrated the modification of the protein 17 H. S. Park, M. J. Hohn, T. Umehara, L. T. Guo,
lysozyme and the peptide hormone somatostatin following
this strategy. These findings underline that the KPL
E. M. Osborne, J. Benner, C. J. Noren, J. Rinehart and
D. Soll, Science, 2011, 333, 1151–1154.
approach is not only applicable if the protein of interest 18 A. Dondoni, Angew. Chem. Int., Ed. Engl., 2008, 47,
carries a lysine residue at the N-terminus since the most 8995–8997.
accessible lysine residue will be modified. In addition, the 19 J. E. Moses and A. D. Moorhouse, Chem. Soc. Rev., 2007, 36,
KPL approach could probably be extended to recombinant 1249–1262.
proteins where an accessible lysine residue is introduced via 20 H. C. Kolb, M. G. Finn and K. B. Sharpless, Angew. Chem.,
a point mutation. In this way, it might be also suitable for Int. Ed. Engl., 2001, 40, 2004–2021.
dual labeling strategies in combination with cysteine point 21 H. Pfeiffer, A. Rojas, J. Niesel and U. Schatzschneider,
mutations.
Dalton Trans., 2009, 4292–4298.
360 | Org. Biomol. Chem., 2013, 11, 353–361
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
22 N. Li, R. K. Lim, S. Edwardraja and Q. Lin, J. Am. Chem. 36 A. A. Poloukhtine, N. E. Mbua, M. A. Wolfert, G. J. Boons
Soc., 2011, 133, 15316–15319. and V. V. Popik, J. Am. Chem. Soc., 2009, 131, 15769–15776.
23 J. Z. Yao, C. Uttamapinant, A. Poloukhtine, J. M. Baskin, 37 L. Chaisemartin, P. Chinestra, G. Favre, C. Blonski and
J. A. Codelli, E. M. Sletten, C. R. Bertozzi, V. V. Popik and
A. Y. Ting, J. Am. Chem. Soc., 2012, 134, 3720–3728.
24 D. Rabuka, Curr. Opin. Chem. Biol., 2010, 14, 790–796.
25 S. Shaunak, A. Godwin, J. W. Choi, S. Balan, E. Pedone,
J. C. Faye, Bioconjugate Chem., 2009, 20, 847–855.
38 P. V. Chang, J. A. Prescher, E. M. Sletten, J. M. Baskin,
I. A. Miller, N. J. Agard, A. Lo and C. R. Bertozzi, Proc. Natl.
Acad. Sci. U. S. A., 2010, 107, 1821–1826.
D. Vijayarangam, S. Heidelberger, I. Teo, M. Zloh and 39 S. S. van Berkel, M. B. van Eldijk and J. C. van Hest, Angew.
S. Brocchini, Nat. Chem. Biol., 2006, 2, 312–313. Chem. Int., Ed. Engl., 2011, 50, 8806–8827.
26 J. Alam, T. H. Keller and T. P. Loh, J. Am. Chem. Soc., 2010, 40 C. I. Schilling, N. Jung, M. Biskup, U. Schepers and
132, 9546–9548. S. Brase, Chem. Soc. Rev., 2011, 40, 4840–4871.
27 J. M. Gilmore, R. A. Scheck, A. P. Esser-Kahn, N. S. Joshi 41 J. M. Chalker, G. J. Bernardes and B. G. Davis, Acc. Chem.
and M. B. Francis, Angew. Chem., Int. Ed. Engl., 2006, 45,
5307–5311.
28 G. B. Postnikova, Biokhimiya, 1994, 59, 1445–1457.
29 K. F. Geoghegan and J. G. Stroh, Bioconjugate Chem., 1992,
3, 138–146.
30 S. Tsukiji, M. Miyagawa, Y. Takaoka, T. Tamura and
I. Hamachi, Nat. Chem. Biol., 2009, 5, 341–343.
31 M. Kunishima, S. Nakanishi, J. Nishida, H. Tanaka,
Res., 2011, 44, 730–741.
42 H. Taskent-Sezgin, J. Chung, P. S. Banerjee, S. Nagarajan,
R. B. Dyer, I. Carrico and D. P. Raleigh, Angew. Chem., Int.
Ed. Engl., 2010, 49, 7473–7475.
43 W. Ardelt, K. Shogen and Z. Darzynkiewicz, Curr. Pharm.
Biotechnol., 2008, 9, 215–225.
44 M. Lillich, X. Chen, T. Weil, H. Barth and J. Fahrer, Biocon-
jugate Chem., 2012, 23, 1426–1436.
D. Morisaki, K. Hioki and H. Nomoto, Chem. Commun. 45 M. J. Thompson and R. A. Goldstein, Proteins, 1996, 25,
(Camb), 2009, 5597–5599. 38–47.
32 G. Chen, A. Heim, D. Riether, D. Yee, Y. Milgrom, 46 M. K. Sinha, O. Reany, M. Yefet, M. Botoshansky and
M. A. Gawinowicz and D. Sames, J. Am. Chem. Soc., 2003,
125, 8130–8133.
33 Z. Chen, B. V. Popp, C. L. Bovet and Z. T. Ball, ACS Chem.
Biol, 2011, 6, 920–925.
E. Keinan, Chemistry, 2012, 18, 5589–5605.
47 P. V. Chang, J. A. Prescher, M. J. Hangauer and
C. R. Bertozzi, J. Am. Chem. Soc., 2007, 129, 8400–8401.
48 T. Nguyen and M. B. Francis, Org. Lett., 2003, 5, 3245–3248.
34 X. Chen, K. Muthoosamy, A. Pfisterer, B. Neumann and 49 S. T. Laughlin, N. J. Agard, J. M. Baskin, I. S. Carrico,
T. Weil, Bioconjugate Chem., 2012, 23, 9.
P. V. Chang, A. S. Ganguli, M. J. Hangauer, A. Lo,
J. A. Prescher and C. R. Bertozzi, Methods Enzymol., 2006,
415, 230–250.
35 J. Szychowski, A. Mahdavi, J. J. Hodas, J. D. Bagert,
J. T. Ngo, P. Landgraf, D. C. Dieterich, E. M. Schuman and
D. A. Tirrell, J. Am. Chem. Soc., 2010, 132, 18351–18360.
50 P. M. Mendes, Chem. Soc. Rev., 2008, 37, 2512–2529.
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem., 2013, 11, 353–361 | 361