A novel protein crosslinking reagent for the determination of moderate resolution protein structures by mass spectrometry (MS3-D)

Article abstract of DOI:10.1016/j.bmcl.2003.10.043

A new approach to the determination of moderate resolution protein structures, termed MS3-D - Mass Spectrometry in 3 Dimensions - has recently been disclosed. The method involves the formation of covalent crosslinks between reactive residues on the protein surface, the determination of the location of those crosslinks in primary sequence space by mass spectrometry, and then the imposition of a distance constraint upon the location of the respective side chains during distance geometry calculations of protein structure. MS3-D is rapid, requires small amounts of protein, and works in native biochemical conditions. Therefore, it offers the potential for determination of the structures of all proteins expressed by an organism in a high throughput manner. However, the methodology is completely dependent upon the production of chemical crosslinks and technical limitations of available crosslinkers have proven problematic in generalization and automation of the method for the determination of the structures of complete proteomes. Presented herein is the design, synthesis, and proofing of a novel modular protein crosslinking reagent designed to enhance hydrophilicity, provide an increased effective signal to noise ratio for MS3-D, and allow the sampling of a wider variety of side chains during the process.

Full text of DOI:10.1016/j.bmcl.2003.10.043

Bioorganic & Medicinal Chemistry Letters 14 (2004) 427–429
A novel protein crosslinking reagent for the determination of
moderate resolution protein structures by mass
spectrometry (MS3-D)§
Naoaki Fujii,a Richard B. Jacobsen,b Nichole L. Wood,b Joseph S. Schoenigerb
and R. Kiplin Guya,*
aDepartments of Pharmaceutical Chemistry and Cellular and Molecular Pharmacology, University of California at San Francisco,
San Francisco, CA 94143, USA
bBiosystems Research Department, Sandia National Laboratories, Livermore, CA 94551-0969, USA
Received 29 August 2003; revised 16 October 2003; accepted 25 October 2003
Abstract—A new approach to the determination of moderate resolution protein structures, termed MS3-D—Mass Spectrometry in
3 Dimensions—has recently been disclosed. The method involves the formation of covalent crosslinks between reactive residues on
the protein surface, the determination of the location of those crosslinks in primary sequence space by mass spectrometry, and then the
imposition of a distance constraint upon the location of the respective side chains during distance geometry calculations of protein
structure. MS3-D is rapid, requires small amounts of protein, and works in native biochemical conditions. Therefore, it offers the
potential for determination of the structures of all proteins expressed by an organism in a high throughput manner. However,
the methodology is completely dependent upon the production of chemical crosslinks and technical limitations of available cross-
linkers have proven problematic in generalization and automation of the method for the determination of the structures of com-
plete proteomes. Presented herein is the design, synthesis, and proofing of a novel modular protein crosslinking reagent designed to
enhance hydrophilicity, provide an increased effective signal to noise ratio for MS3-D, and allow the sampling of a wider variety of
side chains during the process.
# 2003 Elsevier Ltd. All rights reserved.
Two approaches have been reported for the determin-
ation of moderate resolution protein structures by che-
mical crosslinking, mass spectrometry, and computation
(mass spectrometry in 3 dimensions, MS3-D): the bot-
tom up, in which crosslinked proteins are digested by
proteases prior to analysis,2,3,9 and the top down, in
which intact crosslinked protein is used for the analy-
sis.6 Inherent limitations of existing crosslinking
reagents have proven to be a major technical obstacle in
high throughput MS3-D. First, the hydrophobicity of
crosslinkers with long aliphatic arms is problematic due
to hydrophobic collapse. Second, the efficiency of
detection of the required singly crosslinked protein or
peptides is often low. Finally, available commercial
reagents do not allow for formation of crosslinks of the
same length with diverse chemistries. We report herein
the design, synthesis, and evaluation of a modular affi-
nity-tagged crosslinker, 1, optimized for use in MS3-D.
Hydrophilic crosslinkers exist with PEG chains, with
amino acid cores,7 and with combinations of PEG and
cleavable linkers. However, none were available with
bifunctional lysine reactivity, a noncleavable linker arm,
and an affinity tag. Studies with analogous crosslinkers
of various length, lacking the affinity tag, have revealed
that crosslinkers with a sampling range of 5–10 ang-
stroms are optimal for providing distance geometry
information about the placement of lysine backbone for
threading calculations.4 Thus, combination of a single
amide core and a PEG linker led to designed crosslinker
1, with a À1.7 ClogP value,5 and a similar crosslinking
distance to disuccynimidylsuberate (DSS) (Fig. 1). The
reagent also included an affinity tag (biotin) to facilitate
signal enhancement.
Keywords: Protein crosslinking affinity reagent MS3D.
§Supplementary data associated with this article can be found at doi:
10.1016/j.bmcl.2003.10.043
* Corresponding author. Tel.: +1-415-5027051; fax: +1-415-5140689;
e-mail: rguy@cgl.ucsf.edu
0960-894X/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2003.10.043

428
N. Fujii et al. / Bioorg. Med. Chem. Lett. 14 (2004) 427–429
Figure 1. Design of hydrophilic affinity-tagged crosslinking reagent 1.
Figure 3. Western blot analysis of ubiquitin-1 conjugates confirms
biotinylation of ubiquitin and binding to streptavidin. Ubiquitin (1 mg/
lane) was reacted with varied molar equivalents of 1, purified by SDS-
PAGE on an 18% Tris–Glycine gel, electroblotted to PVDF mem-
brane, and incubated with Streptavidin–AP conjugate to detect the
crosslinking. Panel A shows the blot and molar equivalency of 1. Panel
B shows the quantitative measurement of the relative amounts of the
biotin conjugation.
Figure 2. Synthesis of crosslinking reagent 1. (a) Benzylamine
(0.5 equiv), K2CO3 (3.0 equiv), DMF, 30–40 ꢀC, 2 d, 70%; (b) H2, Pd–
C, MeOH, HCl, rt, 1.5 h quant; (c) dPEG4-biotin acid (0.8 equiv),
HBTU (1.3 equiv), DIPEA (3.0 equiv), DMF, rt, 1 h, quant; (d)
MeOH, NaOH (2.5 equiv), H2O, 45 ꢀC, 7 h, then HCl, 60%; (e) HOSu
(2.0 equiv), PL-EDC (3.4 equiv), 4 A molecular sieves, DMF, CHCl3,
rt, 15 h, 86%.
The synthetic route to 1 (Fig. 2) proceeds smoothly and
allows for large-scale preparations. The crosslinking
subunit is assembled convergently, allowing for later
variation of the pendant crosslinkers. In this case, the
bis-succinimidyl ester is formed using solid-supported
carbodiimide1 to avoid aqueous work up and chroma-
tographic purification. Crosslinker 1 is water soluble,
easily handled, and stable in DMSO solution at À80 ꢀC
for more than 2 months.
ESI-QTOF MS. The relative concentrations of doubly
modified and unmodified ubiquitin were strongly
reduced and desired mono- modified ubiquitin
strongly enhanced in the eluted solution (Fig. 4). Thus,
the species needed for generation of constraints for
MS3-D are essentially the exclusive product of the use
of this crosslinking reagent and procedure.
This study clearly demonstrates the utility of a chemical
protein crosslinker designed to overcome the difficulties
with hydrophobicity, lack of selectivity, and lack of
sensitivity observed when commercially available
crosslinkers are used in the MS3-D process. Cross-
linker 1 is stable and water soluble, has good reactivity,
and is both selective (suppresses uncrosslinked and
multiply-labeled species relative to singly labeled) and
sensitive (enhances crosslinked species relative to
unreacted)
Crosslinking reagent 1 efficiently crosslinks the pendant
lysine sidechains of proteins. Ubiquitin was allowed to
react with increasing amounts of 1, purified by SDS-
PAGE, and detected by streptavidin–AP conjugate
blotting. The only observed species corresponded to the
molecular weight of ubiquitin (Fig. 3A). Saturation of
biotin conjugation was observed slightly above 5-molar
equivalents of 1 (Fig. 3B), corresponding to the six
lysines in the protein. Overall, reactivity is comparable
to DSS. Thus, 1 efficiently crosslinks ubiquitin at equi-
molar concentrations, single crosslinks can be titrated
by using lower concentrations of 1, and large excess of 1
does not cause side reaction with other residues.
Compound 1 also provides a chemically flexible plat-
form for the production of a wide range of crosslinkers.
Studies using tetra O18-labeled 1 to apply isotope driven
MSn methodology,8 cysteine reactive and photoreactive
versions of 1 to allow sampling of a wider range of side
chain placements, and multiple other protein systems
are underway and will be reported in due course.
Protein conjugates of 1 can be purified by avidin affinity
chromatography to remove unconjugated protein (due
to lack of affinity for avidin) and multiply labeled pro-
tein (due to high avidity for avidin). Ubiquitin was
allowed to react with 1 and excess crosslinking reagent
removed by centrifugal size exclusion filtration. The
protein sample was then incubated with streptavidin
agarose. The agarose was subsequently washed and the
bound ubiquitin was eluted from the agarose using
acidic acetonitrile. The eluted sample was analyzed by
Supporting information available
Synthetic procedures, SDS-PAGE/Western blotting
protocols, mass spectrometry protocols and data which
can be found with the online version of this paper.

N. Fujii et al. / Bioorg. Med. Chem. Lett. 14 (2004) 427–429
429
Figure 4. Ubiquitin-1 conjugates purified using streptavidin-agarose. (a) The masses of control and modified ubiquitin were measured by QTOF-
MS. Upper panel, raw data from a portion of a control ubiquitin spectrum showing the +9 to +12 charge states. Middle panel, the same m/z range
following reaction of ubiquitin with 5-molar equivalents of 1. Singly- and doubly-modified, as well as unmodified, ubiquitin can be seen, as sum-
marized in (b). Lower panel, ubiquitin reacted with 1 eluted from streptavidin using 70% acetonitrile/5% TFA; (b) De-convolution (MaxEnt1
software, Micromass) of masses in 1-modified spectra before (upper panel) and after (lower panel) streptavidin purification. The internally cross-
linked, singly modified and hydrolyzed, and singly modified and unhydrolyzed species are labeled in the upper panel. Note the absence of unmodi-
fied and doubly modified protein following streptavidin purification in the lower panel; (c) Expansion of the +9 charge region indicated by a bar in
the lower panel of (a).
Nijtmans, L. G.; De Koster, C. G.; Grivell, L. A.; Van
Acknowledgements
Der Spek, H.; Muijsers, A. O. Protein Sci. 2002, 11, 2471.
3. Bennett, K. L.; Matthiesen, T.; Roepstorff, P. Meth. Mol.
Biol. 2000, 146, 113.
4. Collins, C. J.; Schilling, B.; Young, M.; Dollinger, G.;
Guy, R. K. Bioorg. Med. Chem. Lett. 2003, 13, 4023.
5. Daylight Chem Info, I.
National Science Foundation (CHE-0118481; R.K.G.,
N.F.); Department of Energy (DE-AC04-94AL85000;
J.S.S., R.B.J., N.L.W.).
6. Kruppa, G.; Schoeniger, J. S.; Young, M. M. Rapid
Commun. Mass Spectrom. 2003, 17, 155.
7. Trester-Zedlitz, M.; Kamada, K.; Burley, S. K.; Fenyoe,
D.; Chait, B. T.; Muir, T. W. J. Am. Chem. Soc. 2003,
125, 2416.
References and notes
8. Turecek, F. J. Mass Spectrom. 2002, 37, 1.
9. Young, M. M.; Tang, N.; Hempel, J. C.; Oshiro, C. M.;
Taylor, E. W.; Kuntz, I. D.; Gibson, B. W.; Dollinger, G.
Proc. Nat. Acad. of Sci. U.S.A. 2000, 97, 5802.
1. Adamczyk, M.; Fishpaugh, J. R.; Mattingly, P. G. Tetra-
hedron Lett. 1995, 36, 8345.
2. Back, J. W.; Sanz, M. A.; De Jong, L.; De Koning, L. J.;