Combinatorial synthesis of peptide arrays with a laser printer

Article abstract of DOI:10.1002/anie.200801616

(Figure Presented) Special delivery: The freezing of activated amino acid derivatives within solid particles enables a laser printer to deliver these postal packages to defined locations on a solid support with high resolution. Subsequent parallel coupling is initiated simply by melting a whole layer of 20 different amino acid particles (see schematic representation; Fmoc=9-fluorenylmethoxycarbonyl).

Full text of DOI:10.1002/anie.200801616

Communications
DOI: 10.1002/anie.200801616
Biotechnology
Combinatorial Synthesis of Peptide Arrays with a Laser Printer**
Volker Stadler,* Thomas Felgenhauer, Mario Beyer, Simon Fernandez, Klaus Leibe,
Stefan Güttler, Martin Gröning, Kai König, Gloria Torralba, Michael Hausmann,
Volker Lindenstruth, Alexander Nesterov, Ines Block, Rüdiger Pipkorn, Annemarie Poustka†,
F. Ralf Bischoff,* and Frank Breitling*
In memory of Annemarie Poustka
More than forty years ago, Merrifield described the consec-
utive coupling of amino acid monomers to a growing peptide
immobilized on a solid support.[1] His approach was later
expanded to establish the field of combinatorial chemistry,
whereby multiple reactions are carried out in parallel to
synthesize many different peptides.[2] The aim of these
methods is to synthesize and analyze as many peptides as
possible, for example, to identify individual peptides that bind
to a target protein. The one-bead–one-compound method
generates many different peptides readily;[3] however, the
decoding of peptide binders is labor intensive. Furthermore,
problematic peptides, for example, hydrophobic peptides that
bind nonspecifically to any protein, are also synthesized
during library preparation by these methods.
Arrays do not have these drawbacks. The position of a
given peptide on an array corresponds directly to its sequence,
and problematic peptides can be omitted in subsequent
arrays. Peptide arrays were first described by Frank, whose
spot synthesis dominates the field because of its reliability and
wide applicability.[4] High peptide densities can be achieved
with the SC2 method, whereby the individual peptide–
cellulose conjugates synthesized in a first array are separated
and spotted in high density on a secondary support, for
example, glass slides. This variant of the spot synthesis is
particularly useful for the production of multiple replicas of
densely spaced peptide arrays.[5] Arrays with thousands of
oligonucleotides per cm2 can be synthesized by lithographic
methods;[6] however, with these methods only one kind of
monomer can be coupled at a time to spatially defined regions
on the solid support. Thus, 20 ” 10 coupling cycles are
required to synthesize an array of peptides of 10 amino
acids in length. In contrast, only 4 ” 10 coupling cycles are
required to generate an analogous array of oligonucleotides.
This peptide-specific drawback makes it difficult to generate
arrays with longer peptides by lithographic methods.[7]
To generate customized peptide arrays at high density,
high speed, and low cost, we used a modified color laser
printer to “print” the 20 amino acids in the form of solid
amino acid toner particles at defined positions on a glass
support. After printing, an entire layer of all the different
amino acid toners is melted at once to release hitherto
immobilized amino acids and initiate the coupling reaction.
Washing and deprotection steps between the printing pro-
cesses complete the cycle, the repetition of which results into
the combinatorial synthesis of a peptide array (Figure 1). An
advantage of this method is that conventional Fmoc (9-
fluorenylmethoxycarbonyl) chemistry[8] can be used. The
method differs from standard solid-phase synthesis only in
the use of a “solid solvent” (at room temperature), which
immobilizes the amino acids within toner particles until the
beginning of the coupling reaction.
[*] Dr. V. Stadler,[+] Dr. T. Felgenhauer,[+] Dr. M. Beyer, Dr. S. Fernandez,
K. Leibe, K. König, Dr. A. Nesterov, I. Block, Prof. A. Poustka,
Dr. F. R. Bischoff,[+] Dr. F. Breitling[+]
Abteilung Chipbasierte Peptidbibliotheken
Deutsches Krebsforschungszentrum, INF 580
69120 Heidelberg (Germany)
Fax: (+49)6221-421-744
E-mail: v.stadler@dkfz.de
r.bischoff@dkfz.de
f.breitling@dkfz.de
Homepage: http://www.dkfz.de/en/b120_cbpl/index.html
Dr. R. Pipkorn
Zentrale Einheit für Peptidsynthese
DKFZ Heidelberg (Germany)
Dr. G. Torralba, Prof. M. Hausmann, Prof. V. Lindenstruth
Kirchhoff-Institut für Physik, Universität Heidelberg (Germany)
Dr. S. Güttler, M. Gröning
Abteilung für Technische Informationsverarbeitung
Fraunhofer-Institut für Produktionstechnik und Automatisierung
Stuttgart (Germany)
[+] These authors contributed equally to this work.
[†] Deceased on May 3 2008
[**] We thank Dorothea Freidank, Daniela Rambow, Thorsten Kühlwein,
and Jürgen Kretschmer for technical assistance, Martina Schnölzer
(DKFZ) for mass spectrometry, Hans Lehrach for helpful discus-
sions and encouragement, Kazuhiro Ohmori (Sekisui) for advice in
particle composition, Alexander Küller, Michael Grunze, and Reiner
Dahint (University of Heidelberg) for surface-analytical methods,
and the BMBF (grant nos. 03N8710 and NGFN-0313375), the
Helmholtz Society (grant no. VH-VI-108), the HFSPO (grant no.
RGP5/2006), the European Union (grant no. 508399), and the
Volkswagenstiftung (grant no. 79 466) for financial support.
Our peptide laser printer is based conceptually on the
color laser printer OKI C7400, but accommodates 20 instead
of four printing units (each of which contains a particular
amino acid toner), as well as a drive and mounting that enable
the repeated exact positioning (Æ 5 mm) of the solid support
(Figure 2a). A row of approximately 10000 light-emitting
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200801616.
7132
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 7132 –7135

Angewandte
Chemie
(20 ” 20) cm2: The OPC drum is charged evenly by a corona,
and then the pixels that are irradiated with light are dis-
charged. This process is facilitated by the properties of the
OPC material. The drum coating, which is insulating in the
dark, becomes conductive upon exposure to light, so that the
illuminated areas of the drum are neutralized rapidly by
grounding.[9] Subsequently, charged toner particles are trans-
ferred only to those areas previously neutralized by irradi-
ation with light. Thus, the electrostatic pattern is transformed
into the corresponding particle pattern. Finally, the particles
delivered by the OPC drum are collected by a strong electric
field (4 kVmmÀ1) on a solid support, on which a printout is
assembled from the 20 different amino acid toners (Fig-
ure 2b).
Figure 1. Particle-based Merrifield synthesis. a) A laser printer delivers
Fmoc amino acid–OPfp esters embedded within toner particles to
specific locations on a solid support, on which b) the particles are
melted after transfer. Melting enables the amino acid derivatives to
diffuse and undergo coupling to the support. A synthetic cycle is
completed when c) excess monomers are washed away, and d) the
Fmoc protecting group is removed. Repeated coupling cycles generate
a peptide array.
The amino acid toner particles are charged triboelectri-
cally within their cartridges by mild friction of particles
against a foam-rubber drum, whereby a negative charge is
generated on the surface of the particles. We adjusted the
physical properties of the particles to those of commercial
toner particles in terms of size distribution (see Figure S1 in
the Supporting Information), toner transfer (see Figure S3 in
the Supporting Information), melting behavior (see Figure S2
in the Supporting Information), their morphology (see
Figure S4 in the Supporting Information), and, finally, print-
ing performance (Figure 2c, and Figure S8 in the Supporting
Information). We embedded the 20 different Fmoc amino
acid–OPfp esters (OPfp = pentafluorophenyl), which were
chosen for their commercial availability and relative stability
as activated amino acid derivatives for peptide synthesis, into
the corresponding amino acid toner particles (see Table S1 in
the Supporting Information). The component designated to
serve as a solvent in our particle-based Merrifield synthesis
was chosen from higher homologues of standard solvents, for
example, N,N-diphenylformamide. Most such solvents that
are solid at room temperature gave the coupling products in
similar yields to those observed with the standard solvent
N,N-dimethylformamide (DMF) for liquid-phase synthesis,
but some resin components of commercial toners also
performed as well. When we melted the printed particles on
amino-derivatized glass slides,[10] we observed a pattern of
coupled amino acids in the same resolution as that of the
original printout (Figure 2c). Thus, the printing performance
of our amino acid toners in terms of spot resolution and the
amount of toner transferred is nearly indistinguishable from
that of commercial color toners (see Figure S3 in the
Supporting Information). The driver software of our first
peptide laser printer limits the printing resolution to 160000
spots in an area of (20 ” 20) cm2 (Figure 2c): many fewer than
the millions of different peptides that such a method could
generate.
Figure 2. Laser printing: a) The peptide laser printer with 20 different
printing units aligned; the mounting for the support is visible at the
front of the printer. b) A light source (LED row, orange) illuminates
and thereby neutralizes selected areas of an OPC drum (blue), which
is first uniformly charged (yellow) by a corona. Triboelectrically charged
toner particles are transferred to these neutralized areas and from
there by a strong electric field to a solid support. c) Amino acid toner
was printed by the peptide laser printer onto a glass slide derivatized
with free amino groups. The esters (in this case: Fmoc-Ile-OPfp)
embedded in the particles were released by heat, the residual material
was washed away with DMF, and the remaining free amino groups
were blocked with 10% acetic anhydride in DMF. Finally, the Fmoc
protecting groups were removed with 20% piperidine in DMF, and the
newly introduced free amino groups were stained with 0.1% bromo-
phenol blue in methanol.
To explore combinatorial peptide synthesis in more detail,
we synthesized the HA (hemagglutinin) and Flag epitopes,
each on a larger area of our glass surface. We determined the
yield for repetitive coupling by our particle-based method to
be about 90% on average (see Figure S6 in the Supporting
Information) and confirmed the integrity of the synthesized
peptides by mass spectrometry (see Figures S9–11 in the
Supporting Information). This low yield for repetitive cou-
pling relative to yields observed for standard Fmoc synthesis
diodes (LEDs) per 20 cm generates a light pattern on the
surface of a uniformly charged organic photoconducting
(OPC) drum, which rotates with approximately 10000 steps
per 20 cm. The resulting two-dimensional light pattern
comprises around 100 million pixels per (20 ” 20) cm2. The
OPC material translates this light pattern into the corre-
sponding electrostatic pattern of around 100 million pixels per
Angew. Chem. Int. Ed. 2008, 47, 7132 –7135ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7133

Communications
on polymer beads can be explained
partly by the fact that our particle-
based method is not compatible
with the use of preswollen PEG-
derivatized glass slides, as DMF
vapor harms the OPC drums of
the laser printer (see the Supporting
Information; PEG = poly(ethylene
glycol)). We ruled out the possibil-
ity that the nonstandard solvent or
elevated coupling temperatures
induce the racemization of the l-
amino acids (see Figure S7 in the
Supporting Information). Our
observations are in agreement with
the results of Schüttler et al., who
detected no racemization of the
closely related Fmoc amino acid–
OPcp esters at coupling tempera-
tures of 908C (Pcp = pentachloro-
phenyl).[11] Although neither mass
spectrometry nor HPLC data indi-
cate that any other heat-induced
side reaction occurs to a significant
extent (see Figures S7 and S9–S11
in the Supporting Information), this
possibility has to be investigated in
more detail in the future. To our
surprise, none of the Fmoc amino
acid–OPfp esters showed a measur-
able decay rate when stored at 258C
inside our toner cartridges, except
for Fmoc-Arg(Pbf)-OPfp, which
decayed at a moderate rate of 5%
per month (see Figure S5 in the
Supporting Information). This find-
ing is remarkable given the notori-
ous instability of carboxy-activated
Figure 3. Combinatorial synthesis of a peptide array. Permutated Flag and Myc epitopes were
detected by the consecutive addition of epitope-specific mouse monoclonal antibodies (enhanced
chemoluminescence luminescence (ECL, Pierce) readout). a) Glass slides with approximately 5500
peptides generated by combinatorial synthesis; the resulting slides were stained with Flag or Myc
antibodies. Regularly interspersed positive controls (encircled spots in (b,c) correspond to wild-type
epitopes) and some reactive peptide variants are clearly visible. Frames delineate the enlargements
shown in (b,c). b) Characterization of Flag and Myc epitopes. Every amino acid position of each
epitope was exchanged for all 20 different amino acids to characterize the binding requirements of
epitope-specific antibodies. The epitope specificity deduced from individual experiments is given in
the sequence written underneath, whereby the size of an amino acid symbol reflects the importance
of that amino acid for antibody binding. The epitope specificities of 9E10 and Flag M2 antibodies
agree with published results.[13] c) Epitopes were permutated at the two positions highlighted in bold
in the peptide sequence. Interspersed wild-type sequences are encircled. The N-terminal aspartic acid
residue of the Flag epitope can be substituted for other amino acids, whereas the adjacent tyrosine
residue is mandatory for the binding of the Flag M2 antibody. The highlighted isoleucine residue of
the Myc epitope can only be substituted for valine, whereas the adjacent leucine residue is less
important for antibody binding. The results of these more complex permutations all agree with the
experiments in (b). d) Correlation of the staining intensity of the Flag epitope variants in (b) with
published relative affinities,[13] which are written next to the spots stained with the Flag M2 antibody.
Fmoc–arginine derivatives in other solvents.[12] Finally, we
synthesized a peptide array with 5500 different permutated
Flag and Myc epitopes on a microscope slide (400 peptides
per cm2; Figure 3). When we incubated the array with the
corresponding antibodies, all of the permutated peptide
epitopes showed the expected staining pattern, which corre-
lated with previously published results (Figure 3 and Fig-
ure S8 in the Supporting Information).[13]
In summary, we have described the truly combinatorial
synthesis of high-density peptide arrays with a custom-made
peptide laser printer that addresses 20 different amino acid
toners within consecutive combinatorial layers. A variant of
this method, in which a microchip is used to attach amino acid
particles to individual pixel electrodes, has so far enabled us
to address only regular patterns of amino acid particles.[14]
Through the temporary “freezing” of the activated amino
acids within solid particles (Figure 1), the method described
herein enables the separate production and rigorous purifi-
cation of the different amino acid toner particles. These
“postal packages” can be stored for months within the laser
printer, addressed spatially on demand, and only then melted
to initiate combinatorial peptide synthesis. Thus, our
approach adds speed (seconds of printing versus hours of
spotting), flexibility, cost effectiveness, and robustness (see
Figure S5 in the Supporting Information) to the combinato-
rial synthesis of peptide arrays, and at the same time avoids
the drawbacks of other methods. Our peptide laser printer
should make it possible to translate entire genomes of
pathogenic viruses and bacteria into sets of overlapping
peptides, which can be screened for antipathogen antibodies.
Received: April 7, 2008
Revised: May 9, 2008
Published online: July 31, 2008
Keywords: amino acids · biotechnology ·
.
combinatorial chemistry · high-throughput screening · peptides
[1] R. B. Merrifield, J. Am. Chem. Soc. 1963, 85, 2149 – 2154.
[2] a) R. A. Houghten, C. Pinilla, S. E. Blondelle, J. R. Appel, C. T.
Dooley, J. H. Cuervo, Nature 1991, 354, 84 – 86; b) U. Reineke,
R. Volkmer-Engert, J. Schneider-Mergener, Curr. Opin. Bio-
7134
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 7132 –7135

Angewandte
Chemie
technol. 2001, 12, 59 – 64; c) H. M. Geysen, R. H. Meloen, S. J.
Barteling, Proc. Natl. Acad. Sci. USA 1984, 81, 3998 – 4002.
[3] K. S. Lam, S. E. Salmon, E. M. Hersh, V. J. Hruby, W. M.
Kazmierski, R. J. Knapp, Nature 1991, 354, 82– 84.
[4] R. Frank, Tetrahedron 1992, 48, 9217 – 9232.
[9] P. M. Borsenberger, D. S. Weiss in Handbook of Imaging
Materials, 2nd ed. (Eds.: A. S. Diamond, D. S. Weiss), Marcel
Dekker, New York, 2002, pp. 369 – 423.
[10] M. Beyer, T. Felgenhauer, F. R. Bischoff, F. Breitling, V. Stadler,
Biomaterials 2006, 27, 3505 – 3514.
[11] A. Schüttler, W. Meltzow, J. Fꢀhles, H. Zahn, Hoppe-Seylerꢀs Z.
Phys. Chem. 1976, 357, 741 – 744.
[5] A. Dikmans, U. Beutling, E. Schmeisser, S. Thiele, R. Frank,
QSAR Comb. Sci. 2006, 25, 1069 – 1080.
[12] M. H. S. Cezari, L. Juliano, Pept. Res. 1996, 9, 88 – 91.
[13] a) K. Hilpert, G. Hansen, H. Wessner, G. Küttner, K. Welfle, M.
Seifert, W. Hꢀhne, Protein Eng. 2001, 14, 803 – 806; b) J. W.
Slootstra, D. Kuperus, A. Plückthun, R. H. Meloen, Mol.
Diversity 1997, 2, 156 – 164.
[14] M. Beyer, A. Nesterov, I. Block, K. Kꢀnig, T. Felgenhauer, S.
Fernandez, K. Leibe, G. Torralba, M. Hausmann, U. Trunk, V.
Lindenstruth, F. R. Bischoff, V. Stadler, F. Breitling, Science
2007, 318, 1888.
[6] R. J. Lipshutz, S. P. Fodor, T. R. Gingeras, D. J. Lockhart, Nat.
Genet. 1999, 21, 20 – 24.
[7] a) J. P. Pellois, X. Zhou, O. Srivannavit, T. Zhou, E. Gulari, X.
Gao, Nat. Biotechnol. 2002, 20, 922 – 926; b) S. P. Fodor, J. L.
Read, M. C. Pirrung, L. Stryer, A. T. Lu, D. Solas, Science 1991,
251, 767 – 773.
[8] J. Jones in Oxford Chemistry Primers (Ed.: J. Jones), Oxford
University Press, Oxford, 2002, pp. 1 – 92.
Angew. Chem. Int. Ed. 2008, 47, 7132 –7135ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7135