Isotope and Affinity Tags in Photoreactive Substance P Analogues to Identify the Covalent Linkage within the NK-1 Receptor by MALDI-TOF Analysis

Article abstract of DOI:10.1021/ac034512i

Photoreactive analogues of substance P (biotin sulfone-spacer (amino pentanoic or Gly ₃)-Arg-Pro-Lys-Pro-(pBzl)-Phe-Gln-Phe-Phe-Gly-Leu-Met(O ₂)NH₂) with or without isotope (deuterium) labeling have been synthesized. Deuteriums were present on (d)-biotin or epibiotin sulfone (D₃), on the Gly₃ spacer linker (D₆), or on the Gly in position 9 of SP (D₂). Therefore, peptide analogues could be either unlabeled or tri-, penta-, or hexadeuterated. Results obtained with the use of these peptide analogues show that (d)-biotin sulfone and epibiotin sulfone are not recognized with the same affinity by streptavidin, with (d)-biotin sulfone displaying better affinity for the protein. Photolabeling of the human NK-1 receptor with a 1:1 molar ratio of nondeuterated and deuterated photoreactive substance P (SP) analogues in position 5, followed by combined digestions, purification, and MALDI-TOF mass spectrometry analysis, made the identification of the domain of the receptor covalently linked by the photoreactive SP analogue easier. Indeed, doublets in mass spectra were specific for the covalent complex whereas single peaks could be attributed to contaminating species. This method is particularly suitable when minute amounts of complex have to be analyzed, as in the case of highly hydrophobic G-protein coupled receptors.

Full text of DOI:10.1021/ac034512i

Anal. Chem. 2003, 75, 6536-6543
Isotope and Affinity Tags in Photoreactive
Substance P Analogues To Identify the Covalent
Linkage within the NK-1 Receptor by MALDI-TOF
Analysis
Emmanuelle Sachon, Olivier Tasseau, Solange Lavielle, Sandrine Sagan, and Ge´ rard Bolbach*
UMR 7613 CNRSsUniversite´ Pierre et Marie Curie, Structure et Fonction de Mole´cules Bioactives, case courrier 182,
4 Place Jussieu, 75252 Paris Cedex 05, France
years ago, we had developed a procedure involving biotin/
streptavidin purification, which does not require the use of a
radioactive photoactivable ligand.3,4 After enzymatic digestions,
chemical cleavage, or both, the biotin present at the N-terminus
of the photoactivable ligand is used to fish with streptavidin-coated
magnetic beads the ligand-receptor complex out of the cellular
magma.4-6 Characterization of the digested complex is then
achieved by mass spectrometry, mainly MALDI-TOF because of
its high sensitivity (subpicomolar range), its high resolution, and
great tolerance to biological media. Recent results using these
successive steps have shown the interest and the capabilities of
this procedure to define the interaction domain between SP and
the human NK-1 (hNK-1) receptor.4-6
Despite its apparent versatility and ability, this strategy must
however be optimized. Indeed, because of the presence of peaks
hard to interpret in the mass spectra, a lot of rather tedious blanks
and control experiments are needed before getting final results.
The problem is to determine whether these peaks are related to
the photoreactive ligand-receptor complex or to contaminating
peptides still present after the purification procedure. The fact that
the enzyme used can sometimes cleave the receptor and the
ligand, and that chymotrypsin-like activity always derives from
trypsin during long-time incubations,7 renders interpretation of
the mass spectra even more complicated. Consequently, various
enzymatic or chemical digestions and different digestion times
are needed to interpret unambiguously the data obtained from
mass spectrometry.
P hotoreactive analogues of substance P (biotin sulfone-
spacer (amino pentanoic or Gly3)-Arg-P ro-Lys-P ro-(pBzl)-
P he-Gln-P he-P he-Gly-Leu-Met(O2 )NH2 ) with or without
isotope (deuterium) labeling have been synthesized. Deu-
teriums were present on (d)-biotin or epibiotin sulfone
(D3 ), on the Gly3 spacer linker (D6 ), or on the Gly in
position 9 of SP (D2 ). Therefore, peptide analogues could
be either unlabeled or tri-, penta-, or hexadeuterated.
Results obtained with the use of these peptide analogues
show that (d)-biotin sulfone and epibiotin sulfone are not
recognized with the same affinity by streptavidin, with (d)-
biotin sulfone displaying better affinity for the protein.
P hotolabeling of the human NK-1 receptor with a 1 :1
molar ratio of nondeuterated and deuterated photoreac-
tive substance P (SP ) analogues in position 5 , followed
by combined digestions, purification, and MALDI-TOF
mass spectrometry analysis, made the identification of the
domain of the receptor covalently linked by the photore-
active SP analogue easier. Indeed, doublets in mass
spectra were specific for the covalent complex whereas
single peaks could be attributed to contaminating species.
This method is particularly suitable when minute amounts
of complex have to be analyzed, as in the case of highly
hydrophobic G-protein coupled receptors.
Among the different strategies used to elucidate the interaction
domain of a ligand within a receptor protein, photolabeling1,2 is
particularly suitable and complementary to mutagenesis studies.
After photolabeling, the ligand-receptor complex is subject to
one or more enzymatic digestions, chemical cleavage, or both,
subsequent purification step(s) being then required to recover
the fragments of interest. Besides the design of the photoactivable
probes, this purification step is often the cornerstone of such
studies. Indeed, HPLC separation requires nanomole amounts of
radioactive materials to overcome the adsorption losses. This
quantity is a real hindrance to the study of hydrophobic proteins
such as recombinant GTP binding-protein coupled receptors that
are not easily purified in high amounts from mammal cells. A few
We have now developed a strategy complementary to the
procedure we used so far. This method, derived from the isotope-
coded affinity tag method in proteomic studies,8,9 is based on the
isotopic (deuterium) labeling of the photoactivable ligand, to get
(3) Girault, S.; Chassaing, G.; Blais, J.-C.; Brunot, A.; Bolbach, G. Anal. Chem.
1 9 9 6 , 68, 2122-2126.
(4) Girault, S.; Sagan, S.; Bolbach, G.; Lavielle, S.; Chassaing, G. Eur. J. Biochem.
1 9 9 6 , 240, 215-222.
(5) Lequin, O.; Bolbach, G.; Frank, F.; Convert, O.; Girault-Lagrange, S.;
Chassaing, G.; Lavielle, S.; Sagan, S. J. Biol. Chem. 2 0 0 2 , 277, 22386-22394.
(6) Sachon, E.; Bolbach, G.; Chassaing, G.; Lavielle, S.; Sagan, S. J. Biol. Chem.
2 0 0 2 , 277, 50409-50414.
(7) Keil-Dlouha, V.; Zylber, N.; Tong, N.; Keil, B. FEBS Lett. 1 9 7 1 , 16, 287-
290.
(8) Gygi, S. P.; Rist, B.; Gerber, S. A.; Turecek, F.; Gelb, M. H.; Aebersold, R.
Nat. Biotechnol. 1 9 9 9 , 17, 994-999.
(9) Goshe, M. B.; Smith, R. D. Curr. Opin. Biotechnol. 2 0 0 3 , 14, 101-109.
* To whom correspondence should be addressed. E-mail: bolbach@
ccr.jussieu.fr. Telephone: (33) 1 44 27 31 80. Fax: (33) 1 44 27 38 43.
(1) Dorman, G.; Prestwich, G. D. Trends Biotechnol. 2 0 0 0 , 18, 64-77.
(2) Hatanaka, Y.; Sadakane, Y. Curr. Top. Med. Chem. 2 0 0 2 , 2, 271-288.
6536 Analytical Chemistry, Vol. 75, No. 23, December 1, 2003
10.1021/ac034512i CCC: $25.00 © 2003 American Chemical Society
Published on Web 10/24/2003

Figure 1. Schematic representation of the chemical modifications incorporated into the sequence of SP as described in the text.
a direct unambiguous reading of the relevant peaks related to the
ligand-receptor covalent complex. The photolabeling experiments
are carried out with a 1:1 ratio of deuterated and nondeuterated
photoactivable nonradioactive ligand. After photolabeling, diges-
tion, and affinity (biotin/ streptavidin) purification, MALDI-TOF
analysis provides single peaks (with a characteristic isotopic
pattern) for species nonrelated to the ligand-receptor covalent
complex and double peaks separated by the difference in mass
between the deuterated and nondeuterated species (singly charged
ion) that are specific of the complex. This easy and reliable reading
out of the relevant species requires neither tedious controls nor
high numbers of repetitive experiments. In addition, the proteolytic
stability of the ligand can also be followed, depending on the
localization of the deuterium atoms in the peptide sequence
relative to the position of the photoactivable amino acid and the
N-terminal biotin sulfone. Nevertheless, the number of deuteriums
incorporated in the peptide sequence has to be high enough to
prevent too much interference between the isotopic patterns of
deuterated and nondeuterated species. Typically, a labeling using
five or six deuterium atoms is well adapted for peptides with mass
of <3000 u.
for chemical modifications). [D3]-Bapa[(pBzl)Phe5, Met(O2)11]SP
corresponds to the trideuterated peptide (D3), [D3]-Bapa[(pBzl)-
Phe5, [D2]Gly9, Met(O2)11]SP to the pentadeuterated peptide (D5),
and Biot(O2)-([D2]Gly)3-[(pBzl)Phe5, Met(O2)11]SP to the hexa-
deuterated analogue (D6). The pentadeuterated analogue [D3]-
Bapa[(pBzl)Phe5, [D2]Gly9, Met(O2)11]SP has two deuterated
positions located on each side of the photoactivable amino acid.
EXPERIMENTAL SECTION
Syntheses. (a) Oxidation of (d)-Biotin Sulfone. (d)-Biotin
(2.44 g, 10 mmol) in suspension in acetic acid (30 mL) was stirred
at room temperature for 65 h in the presence of H2O2 (10 mL,
35% in water). Dissolution was first observed within a few hours,
and then (d)-biotin sulfone slowly precipitated. This precipitate
was filtered off, washed twice with diethyl ether, and dried in
vacuo. (d)-Biotin sulfone was obtained as a white solid: 2.34 g
(85% yield), mp >260°, Rf (TLC media, CHCl3/ CH3OH/ CH3COOH
9/ 1/ 0.5, p-dimethylaminocinnamaldehyde detection10) 0.2; 1H
2
NMR (500 MHz, H6-DMSO) δ 1.4-1.5 (m, 2H, CH2), 1.5-1.6
(m, 2H, CH2), 1.6-1.7 (m, 2H, CH2), 2.25 (t, J ) 7.6 Hz, 2H, CH2-
CO), 3.0 (d, 1H, J ) 14.5 Hz, H9), 3.2 (q, 1H, H6), 3.3 (m, 3H, H9
+ H2O), 4.4 (m, 2H, H7, H8, J H7,8 ) 9.8 Hz), 6.6 (s, 1H, NH), 6.7
(s, 1H, NH); ESI(+) m/ z 277 [M + H+], ESI(-), m/ z: 275 [M -
H-].
(b) Deuteration of (d)-Biotin Sulfone at Carbons 6 and
9 . A 1 N NaOD solution was prepared by dropwise additions of
D2O (19 mL) to Na (437 mg, 19 mmol). (d)-Biotin sulfone (1.05
g, 3.8 mmol) was added to the NaOD solution, and the resultant
The potential of this strategy combining affinity and isotope
tags was investigated by photolabeling the hNK-1 receptor
expressed in Chinese hamster ovary (CHO) cells. For this
purpose, photoactivable analogues of SP were synthesized, i.e.,
[D3]-Bapa[(pBzl)Phe5, Met(O2)11]SP, [D3]-Bapa[(pBzl)Phe5, [D2]-
Gly9, Met(O2)11]SP, Biot(O2)-([D2]Gly)3-[(pBzl)Phe5, Met(O2)11]-
SP and the nondeuterated analogues Bapa[(pBzl)Phe5, Met(O2)11]-
SP and Biot(O2)-(Gly)3-[(pBzl)Phe5, Met(O2)11]SP, all incorporating
p-benzoyl-L-phenylalanine ((pBzl)Phe) at position 5 (see Figure 1
(10) McCormick, D. B.; Roth, J. A. Anal. Biochem. 1 9 7 0 , 34, 226-231.
Analytical Chemistry, Vol. 75, No. 23, December 1, 2003 6537

Table 1. Summary of Peptide Syntheses and Analytical Data
mass spectrometrya
retention time (min)b
(d)-biotin
purity % (yield %)b
MH+
calc
MH+
meas
(d)-biotin
peptides
sulfone
epimerc
sulfone
epimerc
Bapa[(pBzl)Phe5, Met(O2)11]SP
1859.90
1862.92
1864.93
1937.93
1931.89
1859.94 ( 0.06
1862.96 ( 0.06
1864.93 ( 0.06
1937.89 ( 0.06
1931.89 ( 0.06
8.2
97.2 (46)
98.0 (12)f
97.1 (20)f
97.3 (70)
98.0 (53)
[D3]-Bapa[(pBzl)Phe5, Met(O2)11]SP
8.5d
8.23d
7.1
9.01e
8.88e
96.8 (12)f
96.1 (18)f
[D3]Bapa[(pBzl)Phe5, [D2]Gly9, Met(O2)11]SP
Biot(O2)-([D2]Gly)3-[(pBzl)Phe5, Met(O2)11]SP
Biot(O2)-(Gly)3-[(pBzl)Phe5, Met(O2)11]SP
6.9
a
Calculated and measured (MALDI-TOF) masses are both reported. b The retention times of the purified peptides were determined with isocratic
conditions with 60% B using solvents A (H2O/ 0.1% TFA) and B (CH3CN 60%, H2O 40%, TFA 0.1%). c The epimer of (d)-biotin sulfone has a (R)
configuration at C6, the carbon R to the sulfur atom. d Peak 1 corresponds to trideuterated (d)-biotin sulfone (NMR analysis, data not shown).ePeak
2 corresponds to trideuterated biotin sulfone epimer at C6, R to the sulfur atom (NMR analysis, data not shown). f A fraction containing both
diastereomers was also collected, 6 and 10%, respectively.
mixture was stirred 30 min at room temperature. The mixture
was neutralized by slow additions of 6 N HCl solution. After
filtration, the resulting solution was first desalted on a C18 reversed-
phase column (0.1% TFA aqueous solution), and then deuterated
biotin sulfone was eluted with a solution of acetonitrile in 0.1%
TFA (4:6). This elution allowed partial separation of the two
deuterated diastereomers of biotin sulfone ((d)-biotin sulfone and
its epimer at C6). [D3]-(d)-Biotin sulfone eluted first and can be
isolated. The following fractions containing both trideuterated (d)-
biotin sulfone and its epimer epibiotin sulfone were purified again.
Altogether, [D3]-(d)-biotin sulfone was obtained pure (447 mg),
whereas the trideuterated epibiotin sulfone was always obtained
as a mixture of both diastereomers (458 mg): Rf (TLC media,
CHCl3/ CH3OH/ CH3COOH 9/ 1/ 0.5, detection with p-dimethy-
laminocinnamaldehyde); [D3]-(d)-biotin sulfone 0.17, [D3]-epibiotin
sulfone 0.40; H NMR (250 MHz, H6-DMSO) for trideuterated-
(d)-biotin sulfone δ 1.3-1.8 (m, 6H, 3CH2), 2.2 (t, 2H, CH2CO),
4.3 (m, 2H, H7, H8), 6.6 (s, 1H, NH), 6.7 (s, 1H, NH); and for a
1/ 1 ratio of trideuterated-(d)-biotin sulfone and its epimer δ 1.3-
1.8 (m, 6H, 3CH2), 2.2 (t, 2H, CH2CO), 3.93 (d, 1H, JAB ) 9.8 Hz,
for H7 of trideuterated epibiotin sulfone), 4.2 (d, 1H, JAB ) 9.8 Hz,
H8 of trideuterated epibiotin sulfone), 4.35 (q, 2H, H7, H8 of
trideuterated-(d)-biotin sulfone), 6.7 (s, 1H, NH), 7.0 (s, 1H, NH);
ESI(-) m/ z 278 [M - H-].
of the Boc-protecting group, (d)-biotin sulfone or the mixture of
trideuterated biotin sulfone diastereomers was coupled manually
overnight (5-fold excess) after dissolution in N-methylpyrrolidone
and activation by dicyclohexylcarbodiimide and 1-hydroxylben-
zotriazole. Coupling efficiency was monitored with the Kaiser test.
After completion, the peptidyl-resin was transferred into the Teflon
vessel of an HF apparatus and the peptide was cleaved by
treatment with 1.5 mL of anisole, 0.25 mL of dimethyl sulfide, and
10 mL of anhydrous HF per gram of peptide-resin for 1 h at 0 °C.
After evaporation of HF in vacuo, the resin was first washed (3×)
with diethyl ether and then eluted (3×) with acetic acid/ H2O (1:
1). Lyophilization of the extract led to crude peptides, which were
purified by reversed-phase HPLC with a C8 (7 µm) Symmetry
Prepcolumn (7.8 × 300 mm); separation was accomplished using
various acetonitrile gradients in aqueous 0.1% TFA at a flow rate
of 6 mL/ min with UV detection fixed at 220 nm. Before pooling,
the purity of the collected fractions was ascertained by analytical
HPLC, with a C8 (10 µm) Lichrospher 100 RP column (4 × 250
mm) in isocratic mode at a flow rate of 1.5 mL/ min with UV
detection at 220 nm. Separation of the diastereomers of the two
following peptides [D3]-Bapa[(pBzl)Phe5, Met(O2)11]SP and [D3]-
Bapa[(pBzl)Phe5, [D2]Gly9, Met(O2)11]SP was achieved by pre-
parative HPLC. Further data on synthesized peptides are sum-
marized in Table 1.
1
2
To firmly establish the stereochemistry of the trideuterated
epibiotin sulfone, the exchange reaction was also performed on
biotine sulfone in 1 N NaOH to measure all coupling constants of
the different protons and to assign the stereochemistry at C6 from
the Karplus equation. Complete assignments of the NMR spec-
trum gave the following resonances and coupling constants for
epibiotin sulfone: 1H NMR (500 MHz, 2H6-DMSO) δ 1.4-1.5 (m,
2H, CH2), 1.5-1.6 (m, 2H, CH2), 1.6-1.7 (m, 2H, CH2), 2.20 (t, J
) 7.6 Hz, 2H, CH2CO), 3.17 (m, 2H, H6 and H9), 3.53 (m, 1H, J
H8,9′ ) 7.4 Hz, H9′), 3.94 (m, 1H, J H6,7 ) 6.6 Hz and J H7,8 ) 9.8
Hz, H7), 4.30 (m, 1H, J H8,9 ) 5.8 Hz, J H9,9′ ) 14.2 Hz, H8), 6.7 (s,
1H, NH), 7.0 (s, 1H, NH).
(c) P eptide Syntheses. Peptides syntheses were carried out
on an ABI model 431A peptide synthesizer using t-Boc strategy
and starting from a R-p-methylbenzhydrylamine resin (substitution
0.9 mmol/ g of resin). All NR-Boc amino acids (10-fold excess)
were assembled using dicyclohexylcarbodiimide and 1-hydroxy-
benzotriazole as coupling reagents; only a 5-fold excess of Boc-
Met(O2), Boc-[D2]Gly, and Boc-(pBzl)Phe were used. After
coupling Boc-aminopentanoic acid (10-fold excess) and removal
Biotin-Streptavidin Affinity. 2 -(4 ′-Hydroxyazobenzene)-
benzoic Acid (HABA) Titration Curves.1 1 The titrations were
performed at 500 nm by successive additions of 4 µL of biotin or
biotinylated peptide solutions (100 µM) to 300 µL of streptavidin
(3.03 µM) and HABA (206 µM) solution in 0.2 M NH4OAc, 0.2%
NaN3 (w/ v), pH 7.0, as described.11 The absorbencies, corrected
for dilution, were plotted versus the amount of biotin.
Streptavidin-Coated Magnetic Beads. A solution consisting
of a 1:1 ratio mixture (10 pmol/ µL) of deuterated and nondeu-
terated peptides was prepared in 1 mL of 50 mM Tris-Cl (pH 7.4)
containing 0.1% bovin serum albumin. Samples were gently stirred
for 1 h with 100 µg of streptavidin-coated magnetic beads before
washes and MALDI-TOF analysis as described previously.3
Cell Culture and Membrane P reparation. CHO cells
expressing the human NK-1 receptor (6 pmol/ mg of proteins)
were cultured in DMEM supplemented with 100 IU/ mL penicillin,
100 IU/ mL streptomycin, and 10% fetal calf serum at 37 °C in a
humidified atmosphere of 5% CO2. Membranes were prepared as
(11) Lavielle, S.; Chassaing, G.; Marquet, A. Biochim. Biophys. Acta 1 9 8 3 , 759,
270-277.
6538 Analytical Chemistry, Vol. 75, No. 23, December 1, 2003

described4 and stored in 50 mM Tris-Cl (pH 7.4), 1 mM EDTA, 1
mM MgCl2, 1 mM MnCl2, 1 mM phenylmethanesulfonyl fluoride
(PMSF), 5 µg/ mL soybean trypsin inhibitor (STI), and 10%
glycerol at -80 °C.
patterns. Comparing the ion intensities of the ligand from
controlled experiments with known matrix/ analyte ratios and from
photolabeling experiments evidenced that, in these last experi-
ments, peptides in the 1-10-fmol range are expected on the target.
The mass spectra were externally calibrated based on the
protonated molecules of both the matrix and a known peptide
(neurotensin or [Tyr8]SP). Alternatively, an internal calibration
can also be done with the peptide ligand.
A custom-made program was developed (Visual Basic V6.0)
to deduce the isotopic pattern of the doublet taking into account
the unlabeled/ labeled molar ratio for comparison with experi-
mental data. Finally, the Protein Analysis WorkSheet software
(PAWS, freeware edition, Proteomics, http:/ / www.proteomic-
s.com) was used to research matching masses between the mass
spectrometry data and the hNK-1 receptor sequence.
P hotolabeling Experiments.4 -6 Membranes (1-2 mg of
proteins) from CHO cells expressing the human NK-1 receptor
were incubated for 10 min at room temperature with the peptide
mixture (used at a concentration of 10-fold their affinity sKis for
the NK-1 receptor). The membrane preparation was then irradi-
ated on ice at 365 nm at a distance of 6-10 cm for 40 min. After
irradiation, the sample was centrifuged for 2 min at 13 000 rpm,
washed with Tris-Cl buffer, and centrifuged again. Finally, pho-
tolabeled membranes were incubated for 6 h at room temperature
in a denaturation buffer consisting of 17 mM dithiothreitol and
3% SDS in 30 µL of 50 mM Tris-Cl (pH 8.0).
Trypsin Digestion. After denaturation, photolabeled mem-
branes were digested for 15 h at 22 °C with 100 µg of trypsin in
1 mL of 50 mM NH4HCO3 (pH 8.0). Digestion was stopped by
adding 10 µL of PMSF (0.1 M) and 10 µL of STI (5 µg/ mL). The
digested sample was then incubated with 100 µg of streptavidin-
coated magnetic beads as described previously4 for 1 h at 4 °C.
The sample was then either submitted to further digestion with
endo-GluC or analyzed by MALDI-TOF MS.
RESULTS
[D3]-(d)-Biotin sulfone, obtained by H2O2 oxidation in acetic
acid, was prepared by hydrogen-deuterium (H-D) exchange in
1 N NaOD solution. According to this procedure, the three protons
R to the sulfone were readily and totally H-D exchanged, since
no resonance R to SO2 was observed in the NMR spectrum of
the crude material. After desalting on a C18 reversed-phase column,
elution allowed us to recover only [D3]-(d)-biotin sulfone as a pure
enantiomer after two purification steps, its structure being
determined by comparison with the biotin sulfone NMR spectrum.
The structure of the second eluted molecule, which could not be
isolated as a pure isomer but coeluted with trideuterated (d)-biotin
sulfone, has also been established from NMR experiment and
corresponds to [D3]-epibiotin sulfone. The assignments of all
resonances of trideuterated biotin sulfone diastereoisomers can
Endo-GluC Digestion on Streptavidin-Coated Magnetic
Beads. After purification and washing,4 the beads were incubated
at 37 °C for 15 h with 20 µg of endo-GluC in 20 µL of 100 mM
Tris-Cl (pH 7.8). Digestion was stopped by the addition of 10 µL
of PMSF (0.1 M) and 10 µL of STI. A 100-µg aliquot of streptavidin-
coated magnetic beads were again added for 1 h at 4 °C before
the washing steps.4
CNBr Cleavage. After denaturation of membranes, 100 µL of
70% formic acid was added with a few crystals of CNBr for 15 h
at room temperature. After endo-GluC digestion, 100 µL of 70%
formic acid was added to the beads. The supernatantscontaining
the ligand-receptor complexswas transferred to a new tube and
incubated with a few crystals of CNBr for 15 h at room temper-
ature. Tubes were kept in the dark under argon atmosphere.
MALDI-TOF Mass Spectrometry. Mass spectra were ac-
quired on a MALDI-TOF Voyager Elite (Applied Biosystems,
Framingam, MA) in the reflector mode, and the delayed extraction
conditions were optimized for the working range (m/ z < 4500),
and typical resolution m/ ∆m ∼ 5000 was achieved. Saturated
solution of R-cyano-4-hydroxycinnamic acid in acetonitrile/ water
(0.1% TFA) (4:1, v/ v) was freshly prepared, and 3 µL was added
to the streptavidin-coated magnetic beads for 10 min. Two different
sample preparations were studied. The first one corresponds to
the elution of the peptide solution by blocking the beads with a
magnet, as previously described,3 and 1 µL is deposited on the
target holder. The second one was the direct deposit of 1 µL of
matrix containing beads. Both preparations generally gave similar
results in terms of sensitivity and resolution, although the second
one provided more durable and intense signals during acquisition
of the mass spectra.
1
be attributed by H NMR analysis in D6-DMSO of the diastere-
oisomeric mixture, since the chemical shifts and J3 coupling
constant patterns of the junction protons of the two diastereoiso-
mers are completely different. To unambiguously attribute the
absolute configuration at C6 of the second eluted molecule,
epibiotine sulfone was prepared by treating biotin sulfone in 1 N
NaOH.
Diastereomeric mixtures of trideuterated biotinylated peptides
were prepared with the diastereomeric mixture of trideuterated
(d)-biotin and epibiotin sulfones, due to the rather low recovery
yield (30%) of pure [D3]-(d)-biotin sulfone starting from (d)-biotin
sulfone. Both peptide diastereomers of [D3]-Bapa[(pBzl)Phe5, Met-
(O2)11]SP and [D3]-Bapa[(pBzl)Phe5, [D2]Gly9, Met(O2)11]SP were
successfully separated by preparative HPLC. The absolute con-
figuration at C6 of the biotin sulfone moiety of each isomeric
peptide was also unambiguously attributed by 500-MHz 1H NMR
analysis of both diastereomers (H7, H8 chemical shifts and
coupling patterns, data not shown). In both cases, the trideuterated
(d)-biotin sulfone-substituted peptides eluted first, before the
epimeric trideuterated biotin sulfone-substituted isomer. Thus, in
the following sections, peak 1 of [D3]-Bapa[(pBzl)Phe5, Met(O2)11]-
SP and [D3]-Bapa[(pBzl)Phe5, [D2]Gly9, Met(O2)11]SP will cor-
respond to the (d)-biotin sulfone isomer, while peak 2 of these
peptides will correspond to epibiotin sulfone.
Due to the very small amounts of peptides, the signal-to-noise
ratio was not sufficiently high if the TOF mass spectrum was
averaged on a limited number of the laser shots. Thus, averaging
10-20 mass spectra, each of 256 laser shots, was needed to get
a reliable mass spectrum with a good statistic on the isotope
Streptavidin quenching and subsequent elution of mixtures of
these diastereomeric peptides, [D3]-Bapa[(pBzl)Phe5, Met(O2)11]-
SP or [D3]-Bapa[(pBzl)Phe5, [D2]Gly9, Met(O2)11]SP, by strepta-
vidin-coated magnetic beads were analyzed by MALDI-TOF mass
Analytical Chemistry, Vol. 75, No. 23, December 1, 2003 6539

(O2)11]SP in a 1:1 ratio shows a preferential recovery (1:2 ratio)
of the nondeuterated peptides, that is, the peptide bearing (d)-
biotin sulfone versus the peptide with the epibiotin sulfone at the
N-terminus. The epibiotin sulfone-substituted peptides could never
be totally quenched even with a large excess of streptavidin-coated
magnetic beads. Similar results were obtained with incubation
performed with the diastereomers (peak 1 and peak 2) of [D3]-
Bapa[(pBzl)Phe5, [D2]Gly9, Met(O2)11]SP and Bapa[(pBzl)Phe5,
Met(O2)11]SP. The lower affinity for streptavidin of the epibiotin
sulfone moiety was also ascertained by titration of streptavidin in
solution with HABA (Figure 2). Moreover, as previously ob-
served,11,12 although avidin and streptavidin are supposed to
present four independent/ equivalent binding sites, the “third and
fourth” binding sites of the tetrameric avidin did not bind the
biotinylated peptides, without spacer13 between (d)-biotin sulfone
and the first amino acid, with the same avidity as the two other
ones. Even in the presence of an aminopentanoic spacer between
biotin sulfone and the peptide sequence, the lower affinity for
streptavidin is still observed with peak 1 of [D3]-Bapa[(pBzl)Phe5,
Met(O2)11]SP for approximatively one-fourth of the streptavidin
binding sites. With peak 2 (epibiotin sulfone), this difference in
affinity occurred already at midpoint addition, that is for half of
the binding sites of the tetrameric streptavidin. Consequently,
diastereomeric mixtures of trideuterated biotinylated peptides
cannot be used without prior separation for photolabeling experi-
ments, since a 1:1 ratio recovery of nondeuterated and deuterated
peptides is mandatory for nonambiguous attribution of peaks
specific for the peptide-receptor covalent complex by MALDI-
TOF mass spectrometry. With the peptides modified by a
Figure 2. Displacement curves of HABA from streptavidin by
successive additions of biotin, [D3]-Bapa[(pBzl)Phe5, Met(O2)11]SP
peak 1 (biotin sulfone) or [D3]-Bapa[(pBzl)Phe5, Met(O2)11]SP peak
2 (epibiotin sulfone). Absorbance is corrected for dilution as AoxVo -
AixVi (with Ao corresponding to absorbance of the complex HABA/
streptavidin alone with a volume Vo and Ai to absorbance of the
complex in the presence of a volume Vi of biotin or biotin sulfone-
containing peptide) and plotted versus the added biotin.
spectrometry, as previously described.2-5 Incubation of [D3]-Bapa-
[(pBzl)Phe5, Met(O2)11]SP (peak 1) and Bapa[(pBzl)Phe5, Met-
(O2)11]SP in a 1:1 ratio allows a 1:1 recovery of both nondeuterated
and trideuterated peptides. The behavior was different with the
trideuterated epibiotin sulfone-substituted peptides. Incubation
with streptavidin-coated beads and then elution of [D3]-Bapa-
[(pBzl)Phe5, Met(O2)11]SP (peak 2) and Bapa[(pBzl)Phe5, Met-
Figure 3. (a) Partial positive ion MALDI-TOF mass spectrum of the peptides after photolabeling, combined digestion (trypsin and endoprotease
V8), and affinity extraction using the isotope-unlabeled Bapa[(pBzl)Phe5, Met(O2)11]SP and D5-labeled ligand (molecular ratio 1:0.6); (b) region
around the unbound ligand; (c) calculated isotope pattern of the ligand; (d) region of the ligand covalently attached to the fragment of the
receptor [173-177]; and (e) its expected isotope pattern. All ions are protonated molecules, and the m/z values refer to those of the first peak
of the isotope pattern. Relevant peaks are marked with asterisks for those corresponding to (/) the intact ligands (doublets with a m/z shift of
5) and to (//) the partially digested ligand (doublets with a m/z shift of 3). Intensity is given in arbitrary units (au).
6540 Analytical Chemistry, Vol. 75, No. 23, December 1, 2003

Figure 4. (a) Partial positive ion MALDI-TOF mass spectrum of the peptides after photolabeling, combined digestion (trypsin and endoprotease
V8), and affinity purification using the isotope-unlabeled Biot(O2)-(Gly)3-[(pBzl)Phe5, Met(O2)11]SP and D6-labeled ligand (molecular ratio 1:1);
(b) region around the unbound ligand; (c) calculated isotope pattern of the ligand; (d) region of the ligand covalently attached to the fragment
of the receptor [173-177]; and (e) the corresponding expected isotope pattern. All ions are protonated molecules, and the m/z values refer to
those of the first peak of the isotope pattern. Relevant peaks are marked with one full circle for those involving the intact ligand (doublets with
a m/z shift of 6) and two for the partially digested ligand (doublets with a m/z shift of 6). Intensity is given in arbitrary units (au).
p-benzoylphenylalanine in position 5 of SP, the two biotinylated
diastereomers could be separated by preparative HPLC, but total
recovery of both diastereomers was not quantitatively achieved.
Thus, it turned out that incorporation of bideuterated glycine for
labeling the peptide would be easier to use than deuterated biotin
alone. In the case of SP, [D2]-glycine can be introduced in the
sequence of SP (position 9); otherwise the aminopentanoic spacer
can be replaced by three deuterated glycines. Thus, we synthe-
sized and purified both diastereomers of [D3]-Bapa[(pBzl)Phe5,
[D2]Gly9, Met(O2)11]SP (D5) and also Biot(O2)-([D2]Gly)3-[(pBzl)-
Phe5, Met(O2)11]SP (D6). Indeed, introduction of a deuterated
reporter in the sequence of the peptide can give an index of the
proteolytic stability of the photoactivable peptide during the
different incubation steps, before mass spectrometry analysis,
while the hexadeuterated spacer may be used with any photore-
active peptide.
Figure 3 shows the MALDI-TOF mass spectrum using a
mixture of the unlabeled peptide Bapa[(pBzl)Phe5, Met(O2)11]SP
and labeled peptide (D5) in a molar ratio close to unity. The isotope
pattern of the unbound ligand (Figure 3b) allows the deduction
of the real unlabeled/ labeled molar ratio, i.e., 1.6. The recalculated
isotope pattern for this experimental 1.6 ratio is given in Figure
3c. From these data, the isotope pattern of the covalently linked
peptides can be calculated and compared to the experimental
pattern. Peaks showing a doublet with a m/ z shift of 5 are specific
for the covalent complex and could be unambiguously assigned
(PAWS) to the covalent attachment of the intact ligand to the
fragments 169STTETMPSR177 and 173TMPSR177, confirming previ-
ous results.6 A good agreement in the isotope pattern is found
for the ligand attached to the fragment 173TMPSR177as depicted
in Figure 3d (experimental) and Figure 3e (calculated). Double
peaks with a m/ z shift of 3 also confirm the degradation of the
bound and unbound ligand, since fragment (1-8) contains isotope
labeling only through the biotin group (D3).
The same experiments were done with unlabeled Biot(O2)-
(Gly)3-[(pBzl)Phe5, Met(O2)11]SP and the hexadeuterated (D6)
analogue used in a molar ratio of 1:1. A typical MALDI-TOF mass
spectrum is given in Figure 4. All the relevant peaks appearing
as a doublet, with a m/ z shift of 6, at m/ z 2940.25, 2522.14, 1931.89,
2189.99, and 1599.66 are easily found. They correspond respec-
tively to a protonated molecule consisting of the intact photore-
active ligand linked to the receptor domain [169-177] (mass
measured 1008.36, mass expected 1008.46), [173-177] (mass
measured 590.25, mass expected 590.29), or alone (m/ z at 1931.89)
and to the partially digested ligand (1-8) attached or not to the
receptor fragment [173-177]. In this experiment, isotope labeling
is only present on the N-terminal region of the ligand. Thus, both
the intact and the digested ligands have the same isotope labeling.
It must be mentioned that, for these ions, the intensity ratio
between the labeled and unlabeled species is close to unity as
expected from the initial ligands mixture. Thus, ions shifted by a
m/ z of 6 but with very different relative intensities must be
considered as irrelevant and due to nonspecific adsorption on the
magnetic beads during the purification step. Typically, ions at m/ z
2262.04 and 2268.09 in Figure 4 are assigned to be irrelevant.
Indeed, they generally appeared in mass spectra of both photo-
labeled and control samples (not shown).
The same experiment was also done with combined digestions
(trypsin, endoprotease V8, and CNBr) to precisely the site of
photoinsertion of the photoreactive ligands within the domain
Analytical Chemistry, Vol. 75, No. 23, December 1, 2003 6541

Figure 5. Partial positive ion MALDI-TOF mass spectrum of the peptides after photolabeling, combined digestion (trypsin, endoprotease V8
and CNBr), and affinity purification using (a) the isotope-unlabeled and D5-labeled ligand mixture (1:0.6) and (b) the isotope unlabeled and D6
labeled ligand mixture (1:1). Insets show details of the mass spectra involving the intact ligands bound to the receptor fragment [173-177]. All
ions are protonated molecules, and the m/z values refer to those of the first peak of the isotope pattern. Intensity is given in arbitrary units (au).
[173-177] of the hNK-1 receptor (Figure 5). Single peaks at m/ z
1944.98, 1976.09, and 1979.97, generally present in all mass
spectra, were assigned to be irrelevant. The MALDI-TOF mass
spectra for the unlabeled and D5 labeled ligands showed new
features at m/ z 2386.23 and 2402.22 (Figure 5a), indicating a mass
shift of -64 and -48 u with respect to the ligand attached to the
[173-177] fragment of the receptor, respectively. Similar shifts
were also observed on the new features appearing when a 1:1
mixture of unlabeled and D6 labeled ligands was used (Figure
5b), indicating that similar products derived from CNBr cleavage
were produced. It should be underlined that new species with
similar shifts are also detected to a smaller extent. From the
doublets profile, these species were attributed to involve the
partially digested ligand (1-8) as described above (data not
shown). These fragmentations correspond to the covalent linkage
of the photoreactive peptide to the CγH2 of methionine-174 side
chain that leads, after CNBr cleavage, to the epoxide/ ketone and
ethylenic derivatives of the photoprobe.6,14
recognized with the same affinity by streptavidin. These experi-
ments established that the (S) absolute configuration at C6, is
mandatory for high-affinity binding to streptavidin.
Structural analyses of the binding pocket of (d)-biotin in
streptavidin or avidin have previously been reported.16,17 These
studies show that ureido oxygen of the biotin molecule forms three
hydrogen bonds with the side chain of N12, S16, and Y33 (avidin)
or N23, S27, and Y43 (streptavidin). The tetrahedral oxyanion formed
is one of the predominant stabilizing interactions in biotin-avidin
or biotin-streptavidin complexes. On the other hand, the presence
of the sulfur atom is not a prerequisite for high-affinity binding
(10-15 M) of biotin to avidin since dethiobiotin and biotin sulfone
still bind the protein with dissociation constants of, respectively,
5 × 10-13 and 10-13 M.12 However, the carboxylate of the caproyl
side chain of (d)-biotin forms H-bond interactions with S73 and
S75 (avidin) or S88 and N49 (streptavidin).18 With the amide group
of biotinylated peptide, this stabilizing interaction should not exist.
However, the orientation of the valeric side chain, pointing to the
opposite direction in epibiotin sulfone relative to (d)-biotin sulfone,
DISCUSSION
It is clear from this study that the two C6 diastereoisomers in
position R to the sulfur atom of biotin (Figure 1) are not
(15) Livnah, O.; Bayer, E. A.; Wilchek, M.; Sussman, J. L. FEBS Lett. 1 9 9 3 ,
328, 165-168.
(16) Weber, P. C.; Wendoloski, J. J.; Pantoliano, M. W.; Salemme, F. R. J. Am.
Chem. Soc. 1 9 9 2 , 114, 3197-3200.
(12) Green, N. M. Methods Enzymol. 1 9 9 0 , 184, 51-67.
(17) Livnah, O.; Bayer, E. A.; Wilchek, M.; Sussman, J. L. Proc. Natl. Acad. Sci.
U.S.A. 1 9 9 3 , 90, 5076-5080.
(18) Pazy, Y.; Kulik, T.; Bayer, E. A.; Wilchek, M.; Livnah, O. J. Biol. Chem. 2002,
(13) Finn, F. M.; Titus, G.; Hofmann, K. Biochemistry 1 9 8 4 , 23, 2554-2258.
(14) Sachon, E.; Milcent, T.; Sagan, S.; Convert, O.; Chassaing, G.; Lavielle, S.
Tetrahedron Lett. 2 0 0 2 , 43, 7485-7489.
277, 30892-30900.
6542 Analytical Chemistry, Vol. 75, No. 23, December 1, 2003

may result in the loss of stabilizing interactions leading to the
decrease observed herein in the binding affinity of epibiotin
sulfone-containing peptides for streptavidin.
Proteomics strategies actually combine the use of the so-called
light biotin-tag (nondeuterated) with heavy biotin-tag (deuterated)
to quantify proteins in control and “abnormal” or experimental
situations.8,9 However, such epimerization of biotin sulfone at the
C6 asymmetric center may only occur at very basic pH (g11) and
fortunately cannot arise in experimental conditions used in
proteomics studies.
ously to branched peptides containing the photoreactive (in
position 5) analogue cleaved between Phe8 and (D2)Gly9, meaning
that two deuteriums were lost in the purified branched peptide.
In contrast, all doublets shifted by a m/ z of 5 can unequivocally
match a receptor fragment covalently linked to the entire photo-
reactive peptide. Therefore, previous tedious analyses of mass
spectra because of the possible digestion of the photoprobe ligand
itself become more easily analyzed in the presence of such
“deuterium reporter”. This method can be applied to any bioactive
peptide containing Gly, Ala, Leu, Phe, Tyr, or Val or amino acids
that can be replaced with one of these latter amino acids, without
a concomitant loss in peptide affinity. Indeed at least these six
amino acids mentioned above are commercially available under
deuterated and N-t-Boc-protected forms and may be used in solid-
phase peptide synthesis. In addition, a deuterated Gly may be
introduced as a spacer arm between biotin and the first amino
acid of the peptide sequence. Besides, this methodology can be
used for any enzymatic digestion, specific (as with endoprotease
V8, trypsin, and Arg-C) or nonspecific (as with pepsin), since only
doublets are relevant to the complex and should be analyzed to
identify the receptor domain covalently linked.
Finally, such a strategy (use of 1:1 ratio between deuterated
and nondeuterated peptides) should be generalized in photola-
beling studies using mass spectrometry analysis. However, if this
strategy is very useful to identify by MALDI-TOF relevant peaks,
it cannot, in any case, identify the amino acid in the receptor
covalently linked to the photoreactive ligand. For that purpose,
combined digestions as described herein or tandem mass spec-
trometry is needed. PSD-MALDI-TOF has already been applied
to the sequencing of a branched peptide to deduce the residue
involved in the covalent linkage.19 However, the minute amount
of branched peptides in our experiments (a few femtomoles) put
an obstacle in this way and all attempts to get relevant information
in PSD-MALDI were unsuccessful. ESI-MS/ MS would be a
reliable alternative to get the sequence of the branched peptide.
Its sensitivity, as shown in recent proteomic studies when coupled
with liquid chromatography,20,21 lies in the low-femtomole range.
Furthermore, this ionization mode is compatible with the affinity
purification methods.22,23 The interpretation of tandem mass
spectra would be easier by comparing the fragmentation pattern
of unlabeled and labeled relevant branched peptide ions, especially
if the deuterium labeling is carried out on both sides of the ligand
relative to the photoreactive residue. Finally, LC-MS and LC-
MS/ MS20,21 could be interesting alternative techniques to analyze
the complex mixture of peptides after photolabeling and digestion
without the purification step.
The second main result of this study shows the suitable use
of an equimolar mixture of deuterated and nondeuterated photo-
reactive peptides to identify specific peaks after photolabeling,
purification, and MALDI-TOF analysis of minute amount of
covalent complex between the peptide and the protein receptor.
The methodology has previously been used without deuterated
peptides and was shown to be rather tedious because numerous
experiments were required to conclude the identification of the
receptor domain.4-6 Here, doublets in MALDI-TOF mass spectra
could in almost all cases be attributed specifically to the covalent
complex between the receptor and the photoreactive peptide.
Indeed, we have done again the photolabeling of the hNK-1
receptor with the analogue of SP photoreactive in position 5
recently reported,6 but this time using a 1:1 ratio mixture of penta-
or hexadeuterated peptide and the corresponding nonlabeled
photoreactive peptides. From one single experiment using the
mixture of deuterated and nondeuterated photoreactive peptides,
we could conclude the receptor domain was covalently linked
while several experiments were required before this could be
discerned when only the unlabeled peptide was used. The results
obtained in this study confirm that, after CNBr cleavage, the
species with a mass shift of -64 and -48 u, with respect to the
unlabeled ligand attached to the [173-177] fragment of the
receptor previously characterized, is really related to the covalent
complex and corresponds to the ethylenic and epoxide/ ketone
derivatives of this branched peptide.6,14 Moreover, the use of
deuterated peptides confirms that the photoreactive SP analogues
are sensitive to trypsin or trypsin-derived chymotrypsin-like activity
during long incubation times.4-7 For example, the pentadeuterated
peptide contains three deuteriums on the N-terminal biotin sulfone
and two others on Gly9 in the photoreactive analogue peptide
sequence. After trypsin incubation, MALDI-TOF analysis identified
doublets shifted by a m/ z of 3. These ions correspond unambigu-
(19) Fournier, I.; Chaurand, P.; Bolbach, G.; Lutzenkirchen, F.; Spengler, B.;
Tabet, J. C. J. Mass Spectrom. 2 0 0 0 , 35, 1425-1433.
(20) Washburn, M.P.; Wolters, D., Yates, J. R., III. Nat. Biotechnol. 2 0 0 1 , 19,
242-247.
(21) Romijn, E. P.; Krijgsveld, J.; Heck, A. J. J. Chromatogr., A 2 0 0 3 , 1000, 589-
608.
(22) Gerber, S. A.; Scott, C. R.; Turecek, F.; Gelb, M. H. J. Am. Chem. Soc. 1 9 9 9 ,
Received for review May 14, 2003. Accepted September
16, 2003.
121, 1102-1103.
(23) Li, Y.; Ogata, Y.; Freeze, H. F.; Scott, C. R.; Turecek, F.; Gelb, M. H. Anal.
Chem. 2 0 0 3 , 75, 42-48.
AC034512I
Analytical Chemistry, Vol. 75, No. 23, December 1, 2003 6543