Monolayer assemblies of a de novo designed 4-α-helix bundle carboprotein and its sulfur anchor fragment on Au(111) surfaces addressed by voltammetry and in situ scanning tunneling microscopy

Article abstract of DOI:10.1021/ja020943r

Mapping and control of proteins and oligonucleotides on metallic and nonmetallic surfaces are important in many respects. Electrochmical techniques based on single-crystal electrodes and scanning probe microscopies directly in aqueous solution (in situ SPM) have recently opened perspectives for such mapping at a resolution that approaches the single-molecule level. De novo design of model proteins has evolved in parallel and holds promise for testing and controlling protein folding and for new tailored protein structural motifs. In this report we combine these two strategies. We present a scheme for the synthesis of a new 4-α-helix bundle carboprotein built on a galactopyranoside derivative with a thiol anchor aglycon suitable for surface immobilization on gold. The carboprotein with thiol anchor in monomeric and dimeric (disulfide) form, the thiol anchor alone, and a sulfur-free 4-α-helix bundle carboprotein without thiol anchor have been prepared and investigated for comparison. Cyclic and differential pulse voltammetry (DPV) of the proteins show desorption peaks around -750 mV (SCE), whereas the thiol anchor desorption peak is at -685 mV. The peaks are by far the highest for thiol monomeric 4-α-helix bundle carboprotein and the thiol anchor. This pattern is supported by capacitance data. The DPV and capacitance data for the thiolated 4-α-helix bundle carboproteins and the thiol anchor hold a strong Faradaic reductive desorption component as supported by X-ray photoelectron spectroscopy. The desorption peak of the sulfur-free 4-α-helix bundle carboprotein, however, also points to a capacitive component. In situ scanning tunneling microscopy (in situ STM) of the thiol anchor discloses an adlayer with small domains and single molecules ordered in pin-striped supramolecular structures. In situ STM of thiolated 4-α-helix bundle carboprotein monomer shows a dense monolayer in a broad potential range on the positive side of the desorption potential. The coverage decreases close to this potential and single-molecule structures become apparent. The in situ STM contrast is also strengthened, indicative of a new redox-based tunneling mechanism. The data overall suggest that single-molecule mapping of natural and synthetic proteins on well-characterized surfaces by electrochemistry and in situ STM is within reach.

Full text of DOI:10.1021/ja020943r

Published on Web 12/10/2002
Monolayer Assemblies of a De Novo Designed 4-r-Helix
Bundle Carboprotein and Its Sulfur Anchor Fragment on
Au(111) Surfaces Addressed by Voltammetry and In Situ
Scanning Tunneling Microscopy
Jesper Brask,† Hainer Wackerbarth,† Knud J. Jensen,*,‡ Jingdong Zhang,†
Ib Chorkendorff,§ and Jens Ulstrup*,†
Contribution from the Department of Chemistry, Buildings 201 and 207, and Department of
Physics and ICAT, Building 309, Technical UniVersity of Denmark, DK-2800 Lyngby, Denmark,
and Department of Chemistry, The Royal Veterinary and Agricultural UniVersity,
ThorValdsensVej 40, DK-1871 Frederiksberg C, Denmark
Received July 9, 2002
Abstract: Mapping and control of proteins and oligonucleotides on metallic and nonmetallic surfaces are
important in many respects. Electrochemical techniques based on single-crystal electrodes and scanning
probe microscopies directly in aqueous solution (in situ SPM) have recently opened perspectives for such
mapping at a resolution that approaches the single-molecule level. De novo design of model proteins has
evolved in parallel and holds promise for testing and controlling protein folding and for new tailored protein
structural motifs. In this report we combine these two strategies. We present a scheme for the synthesis
of a new 4-R-helix bundle carboprotein built on a galactopyranoside derivative with a thiol anchor aglycon
suitable for surface immobilization on gold. The carboprotein with thiol anchor in monomeric and dimeric
(disulfide) form, the thiol anchor alone, and a sulfur-free 4-R-helix bundle carboprotein without thiol anchor
have been prepared and investigated for comparison. Cyclic and differential pulse voltammetry (DPV) of
the proteins show desorption peaks around -750 mV (SCE), whereas the thiol anchor desorption peak is
at -685 mV. The peaks are by far the highest for thiol monomeric 4-R-helix bundle carboprotein and the
thiol anchor. This pattern is supported by capacitance data. The DPV and capacitance data for the thiolated
4-R-helix bundle carboproteins and the thiol anchor hold a strong Faradaic reductive desorption component
as supported by X-ray photoelectron spectroscopy. The desorption peak of the sulfur-free 4-R-helix bundle
carboprotein, however, also points to a capacitive component. In situ scanning tunneling microscopy (in
situ STM) of the thiol anchor discloses an adlayer with small domains and single molecules ordered in
pin-striped supramolecular structures. In situ STM of thiolated 4-R-helix bundle carboprotein monomer shows
a dense monolayer in a broad potential range on the positive side of the desorption potential. The coverage
decreases close to this potential and single-molecule structures become apparent. The in situ STM contrast
is also strengthened, indicative of a new redox-based tunneling mechanism. The data overall suggest that
single-molecule mapping of natural and synthetic proteins on well-characterized surfaces by electrochemistry
and in situ STM is within reach.
1. Introduction
ET mechanisms of immobilized functional redox proteins and
enzymes,5,8-11 and mechanisms of coupled chemical and bio-
catalytic processes.8,9,11 Biotechnological perspectives include
protein-membrane interactions and drug delivery,12,13 biologi-
cal corrosion,14 biocompatibility of metallic implantates,15 and
biofilms.16 To this come areas under headings such as on-line
“bioelectronic function”, “biosensors” etc.,8,9,17-20 covering
Mapping of the structural organization and electron transfer
(ET) function of proteins and other biomolecules, such as
oligonucleotides, adsorbed on solid surfaces is crucial in several
contexts. Fundamental contexts are adsorption patterns,1-4 two-
dimensional molecular interactions including phase transitions,5-7
* Corresponding authors: kjj@kvl.dk and ju@kemi.dtu.dk.
† Department of Chemistry, Technical University of Denmark.
‡ Department of Chemistry, The Royal Veterinary and Agricultural
University.
(5) Zhang, J.; Chi, Q;, Kuznetsov, A. M.; Hansen, A. G.; Wackerbarth, H.;
Christensen, H. E. M.; Andersen, J. E. T.; Ulstrup, J. J. Phys. Chem. B
2002, 106, 1131-1152.
(6) Castner, D. G.; Ratner, B. D. Surf. Sci. 2002, 500, 28-60.
(7) Nielsen, L. K.; Bjørnholm, T.; Mouritsen, O. G. Nature 2000, 404, 353.
(8) Shipway, A. N.; Willner, I. Acc. Chem. Res. 2001, 34, 421-432.
(9) Gilardi, G.; Fantuzzi, A. Trends Biotechnol. 2001, 19, 468-476.
(10) Faraday Discussion Chem. Soc. 2000, 116. Special issue on bioelectro-
chemistry.
§ Department of Physics and ICAT, Technical University of Denmark.
(1) Proteins at Interfaces II. Fundamentals and Applications; Horbett, T. A.,
Brash, J. L., Eds.; ACS Symposium Series 602; American Chemical
Society: Washington, DC, 1995.
(2) Tirrell, M.; Kokkoli, E.; Biesalski, M. Surf. Sci. 2002, 500, 61-82.
(3) Kasermo, B. Surf. Sci. 2002, 500, 656-677.
(11) Jeuken, L. J. C.; Jones, A. K.; Chapman, S. K.; Cecchini, G.; Armstrong,
F. A. J. Am. Chem. Soc. 2002, 124, 5702-5713.
(4) Service, R. P. Science 1998, 282, 396-399.
9
94
J. AM. CHEM. SOC. 2003, 125, 94-104
10.1021/ja020943r CCC: $25.00 © 2003 American Chemical Society

Synthesis of 4-R-Helix Bundle Carboprotein
A R T I C L E S
broadly bioelectrochemistry understood as interfacial electro-
chemical ET between metallic electrodes and adsorbed redox
metalloproteins or other biomolecules. Electronic communica-
tion between the electrodes and the biomolecules can be
controlled by electrode modification, covalent linking, and
electromagnetic radiation where, for example, photoisomeriza-
tion of molecular links to the electrode can be exploited.8
trodes. This enables characterization of the protein and partial
structure monolayers by electrochemical approaches based on
atomically planar electrode surfaces. Molecular monolayer
characterization includes cyclic (CV) and differential pulse
voltammetry (DPV), interfacial capacitance data, X-ray photo-
electron spectroscopy (XPS), and in situ STM directly in
aqueous buffer media. Such combinations have only recently
been introduced in the context of natural or synthetic macromole-
cules.5,32-38 The present investigation shows that the combina-
tion of totally synthetic proteins with state-of-the-art physical
electrochemistry indeed holds promise for addressing biomo-
lecular structure and function at the single-molecule and
supramolecular levels.
Ultimate goals of these efforts in areas broadly denoted as
“bioelectronics” extend to bringing electronic control into true
nanoscale and single-molecule levels. Other goals are to combine
different nanoscale functions into biotechnological devices and
circuits. Efforts along new lines are also needed, however, with
the common notion of “single-molecule electronic technology”.
This implies first that interfacial bioelectrochemical ET should
be combined with state-of-the-art physical electrochemistry, such
as use of single-crystal electrode surfaces and in situ scanning
probe techniques, particularly in situ scanning tunneling mi-
croscopy (STM).21-23 The latter images single-molecule function
directly in the biological aqueous buffer media. A second
approach addresses systematically molecular structure-function
relations, where secondary and tertiary protein-folding and
stability features can be controlled and tailor-made, and the
resulting bioelectrochemical function monitored.
2. Experimental Section
2.a. Synthesis of Carboproteins and Anchor Molecule. 2.a.1.
General Procedures. Tritylmercaptoacetic acid was prepared in 76%
yield from chloroacetic acid.39 4-Aminophenyl R-D-Galp, 1, was
prepared in 92% yield by catalytic hydrogenation of 4-nitrophenyl R-D-
Galp.40 Boc2NOCH2COOH (Boc2-Aoa-OH) was prepared in 77% yield
from Boc-Aoa-OH.25 For synthetic details for peptide aldehyde Ac-
YEELLKKLEELLKKAG-H, 5, and the 4HB carboprotein without
1
anchor, 12, refer to ref 41. H and 13C NMR spectra were recorded at
500 and 125.7 MHz, respectively, on a Varian Unity Inova 500
spectrometer. Electrospray ionization mass spectrometry (ESI MS) was
performed on a Micromass LCT mass spectrometer. For further general
procedures and instrumentation, see refs 25 and 42.
In this report we combine these two approaches, based on
de novo designed 4-R-helix bundle (4HB) carboprotein
structures.24-26 These are intermediate-size structures simplified
compared with natural proteins. Their secondary and tertiary
structural elements can be varied systematically, enabling new
levels of detail in molecular structure-function relations. Several
groups have developed synthetic 4HB proteins by using different
strategies.27-31 Two of the present authors have introduced the
use of monosaccharides as templates for the de novo design of
4HB proteins denoted as carboproteins.24-26 Such a 4HB
carboprotein and its anchor fragment for protein immobilization
on Au-electrode surfaces are target molecules in the present
report. The 4HB carboprotein and its partial molecular structure
are brought to immobilization on single-crystal Au(111) elec-
2.a.2. 4-(Tritylmercaptoacetamide)phenyl r-D-Galp, 2. TrSCH2-
COOH (213 mg, 0.64 mmol) was dissolved in dimethylformamide
(DMF) (10 mL), and PyBOP (331 mg, 0.64 mmol), HOBt (97 mg,
0.64 mmol), and DIPEA (218 µL, 1.27 mmol) were added. After 5
min, compound 1 (157 mg, 0.58 mmmol) was added. The dark solution
was stirred for 18 h and then concentrated to dryness, dissolved in
MeCN (3 mL), and purified by prep. C18 RP-HPLC. Product 2 was
1
obtained as a white solid. Yield 300 mg; 88%. m.p. 115-117 °C. H
NMR (DMSO-d6, D2O-exch), δ: 7.4-7.2 (m, 15H, Tr), 7.25 (d, 8.8
Hz, 2H, aryl), 6.98 (d, 8.8 Hz, 2H, aryl), 5.29 (d, 2.8 Hz, 1H, H-1),
3.8-3.7 (m, 4H, H-2/3/4/5), 3.50 (dd, 6.5 Hz, 11.4 Hz, 1H, H-6a),
3.4-3.3 (m, 1H, H-6b), 2.99 (s, 2H, CH2S). 13C NMR (DMSO-d6), δ:
166.9 (NHCO), [154.5, 134.1, 121.6, 118.4] (aryl, OPhNH), [145.2,
130.3, 129.3, 128.0] (aryl, Tr), 99.8 (C-1), 73.2 (C-5), [70.4, 69.6, 69.1]
(C-2/3/4), 67.3 (CPh3), 61.4 (C-6), 38.1 (CH2S).
2.a.3. 4-(Tritylmercaptoacetamide)phenyl 2,3,4,6-tetra-O-(Boc2-
Aoa)-r-D-Galp, 3. Compound 2 (80 mg, 0.14 mmol) and Boc2-Aoa-
OH (238 mg, 0.82 mmol) were dissolved in pyridine/DCM (1:1, 4 mL).
DMAP (10 mg, 0.082 mmol) and N,N′-diisopropylcarbodiimide
(DIPCDI) (126 µL, 0.81 mmol) were added. After 30 min, TLC
(EtOAc-hexane, 2:3) showed several products and additional DIPCDI
(12) Lasic, D. D. Liposomes: From Physics to Applications, 2nd ed.; Elsevier:
Amsterdam, 1995.
(13) Storm, G.; Crommelin, D. J. A. Pharm. Sci. Technol. Today 1998, 1, 19-
31.
(14) Telegdi, J.; Keresztes, Z.; Pa´linkas, G.; Ka´lma´n, E.; Sand, W. Appl. Phys.
A 1998, 66, S639-S642.
(15) Walavaara, B.; Askendal, A.; Lundstro¨m, I.; Tengwall, P. J. Biomater. Sci.
Forum 1998, 8, 41-48.
(16) Telegdi, J.; Shaban, A.; Beczner, J.; Keresztes, Z.; Ka´lma´n, E. Mater. Sci.
Forum 1998, 289, 77-82.
(17) Eggins, B. Biosensors. An Introduction; Wiley: Chichester, UK, 1997.
(18) Wang, J. Electroanalysis 2001, 13, 983-988.
(19) Piedade, J. A. P.; Fernandes, I. R.; Oliveira-Brett, A. M. Bioelectrochemistry
2002, 56, 81-83, and references there.
(32) Andersen, J. E. T.; Møller, P.; Pedersen, M. V.; Ulstrup, J. Surf. Sci. 1995,
325, 193-205.
(20) Palecek, E.; Fojta, M.; Jelen, F. Bioelectrochemistry 2002, 56, 85-90.
(21) Nanoscale Probes of the Solid/Liquid Interface; Gewirth, A. A., Siegent-
haler, H., Eds.; Kluwer: Dordrecht, 1995.
(33) Zhang, J.; Chi, Q.; Dong, S.; Wang, E. Bioelectrochem. Bioenerg. 1996,
39, 267-274.
(34) Andersen, J. E. T.; Olesen, K. G.; Danilov, A. I.; Foverskov, C. E.; Møller,
P.; Ulstrup, J. Bioelectrochem. Bioenerg. 1997, 44, 57-63.
(35) Friis, E. P.; Andersen, J. E. T.; Madsen, L. L.; Møller, P.; Ulstrup, J.
Electroanal. Chem. 1997, 431, 35-38.
(22) Imaging of Surfaces and Interfaces; Lipkowski, J., Ross, P. N., Eds.; Wiley-
VCH: New York, 1999.
(23) Danilov, A. I. Russ. Chem. ReV. 1995, 64, 767-781.
(24) Jensen, K. J.; Barany, G. J. Pept. Res. 2000, 56, 3-11.
(25) Brask, J.; Jensen, K. J. J. Pept. Sci. 2000, 6, 290-299.
(26) Brask, J.; Jensen, K. J. Bioorg. Med. Chem. Lett. 2001, 11, 697-700.
(27) DeGrado, W. F.; Summa, C. M.; Pavone, V.; Nastri, F.; Lombardi, A. Annu.
ReV. Biochem. 1999, 68, 779-819.
(36) Friis, E. P.; Andersen, J. E. T.; Kharkats, Yu. I.; Kuznetsov, A. M.; Nichols,
R. J.; Zhang, J.; Ulstrup, J. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 1379-
1384.
(37) Chi, Q.; Zhang, J.; Friis, E. P.; Anderen, J. E. T.; Ulstrup, J. Electrochem.
Commun. 1999, 1, 91-96.
(38) Chi, Q.; Zhang, J.; Nielsen, J. U.; Friis, E. P.; Chorkendorff, I.; Canters,
G. W.; Andersen, J. E. T.; Ulstrup, J. J. Am. Chem. Soc. 2000, 122, 4047-
4055.
(39) Bell, A. B.; Hughes, D. W.; Lock, C. J. L.; Valliant, J. F. Can. J. Chem.
1996, 74, 1503-1511.
(40) Westphal, O.; Feier, H. Chem. Ber. 1956, 89, 582-588.
(41) Brask, J.; Dideriksen, J. M.; Nielsen, J.; Jensen, K. J., submitted.
(42) Boas, U.; Brask, J.; Christensen, J. B.; Jensen, K. J. J. Comb. Chem. 2002,
4, 223-228.
(28) Choma, C. T.; Lear, J. D.; Nelson, M. J.; Dutton, P. L.; Robertson, D. E.;
DeGrado, W. F. J. Am. Chem. Soc. 1994, 116, 856-865.
(29) Robertson, D. E.; Farid, R. S.; Moser, C. C.; Urbauer, J. L.; Mulholland,
S. E.; Pidikiti, R.; Lear, J. D.; Wand, A. J.; DeGrado, W. F.; Dutton, P. L.
Nature 1994, 368, 425-432.
(30) Rau, H. K.; DeJonge, N.; Haehnel, W. Angew. Chem., Int. Ed. 2000, 39,
250-253.
(31) Mutter, M.; Vuilleumier, S. Angew. Chem., Int. Ed. Engl. 1989, 28, 535-
554.
9
J. AM. CHEM. SOC. VOL. 125, NO. 1, 2003 95

A R T I C L E S
Brask et al.
(126 µL) was added. After further 30 min, TLC (EtOAc-hexane, 2:3)
showed one major component with Rf 0.65. The suspension was
concentrated to dryness and CH3CN (5 mL) was added. Solid N,N′-
diisopropylurea was filtered off and the mixture was purified by prep.
C18 RP-HPLC. Product 3 was obtained as a viscous oil, which
crystallized when a solution in EtOH was poured over ice water. The
crystals were filtered off and dried in vacuo. Yield 142 mg; 62%. m.p.
5.13 (dd, 10.7 Hz, 3.5 Hz, 2H, H-2), 4.8-4.4 (m, 18H, CH2 Aoa/H-5),
4.23 (dd, 11.5 Hz, 7.0 Hz, 2H, H-6a), 4.15 (dd, 11.4 Hz, 5.3 Hz, 2H,
H-6b), 3.72 (s, 4H, CH2S), [1.46, 1.46, 1.42, 1.40] (singlets, 144H,
Boc). 13C NMR (DMSO), δ: [167.9, 167.9, 167.7, 167.1] (CO Aoa),
[153.1, 135.9, 121.8, 119.5] (aryl), [150.6, 150.5, 150.5, 150.5] (CO
Boc), 96.8 (C-1), [84.9, 84.8, 84.8] (C(CH3)3), 73.0 (CH2 Aoa), [69.9,
68.3, 67.9] (C-2/3/4/5), 44.1 (CH2S), [28.7, 28.7, 28.7] (C(CH3)3).
2.a.8. 4-(Mercaptoacetamide)phenyl 2,3,4,6-tetra-O-(Aoa)-r-D-
Galp disulfide, 9. Compound 8 (75 mg, 26 µmol) was deprotected in
TFA/DCM (1:1, 4 mL) for 30 min. The colorless solution was then
concentrated to dryness, dissolved in H2O/CH3CN (10:0.2, 10 mL),
and lyophilized to give template 9 as a white powder. Yield 56 mg;
1
80-83 °C. H NMR (CDCl3), δ: 7.5-7.2 (m, 15H, Tr), 7.20 (d, 9.0
Hz, 2H, aryl), 6.93 (d, 9.0 Hz, 2H, aryl), 5.76 (d, 3.3 Hz, 1H, H-1),
5.70 (dd, 10.7 Hz, 3.2 Hz, 1H, H-3), 5.59 (d, 3.4 Hz, 1H, H-4), 5.27
(dd, 11.1 Hz, 3.9 Hz, 1H, H-2), 4.7-4.4 (m, 9H, CH2 Aoa/H-5), 4.24
(dd, 11.3 Hz, 6.5 Hz, 1H, H-6a), 4.20 (dd, 11.2 Hz, 6.4 Hz, 1H, H-6b),
3.29 (s, 2H, CH2S), [1.55, 1.54, 1.51, 1.50] (singlets, 72H, Boc). 13C
NMR (CDCl3), δ: [167.5, 167.3, 167.1, 167.0] (CO Aoa), 166.5
(NHCO), [153.3, 133.7, 122.0, 118.0] (aryl, OPhNH), [150.7, 150.7,
150.6, 150.5] (CO Boc), [144.5, 130.1, 129.0, 127.9] (aryl, Tr), 95.9
(C-1), [85.2, 85.2, 85.1, 85.0] (C(CH3)3), [72.8, 72.6, 72.5] (CH2 Aoa/
CPh3), [69.4, 69.0, 68.3, 67.6] (C-2/3/4/5), 62.5 (C-6), 37.2 (CH2S),
[28.8, 28.7, 28.7] (C(CH3)3).
1
98% (incl. 8× TFA). H NMR (DMSO, D2O-exch.), δ: 7.54 (d, 8.9
Hz, 4H, aryl), 7.07 (d, 9.3 Hz, 4H, aryl), 5.75 (d, 3.5 Hz, 2H, H-1),
5.6-5.5 (m, 4H, H-3/4), 5.25 (dd, 10.1 Hz, 3.7 Hz, 2H, H-2), 4.7-4.3
(m, 18H, CH2 Aoa/H-5), 4.23 (dd, 11.5 Hz, 7.1 Hz, 2H, H-6a), 4.16
(dd, 11.0 Hz, 5.1 Hz, 2H, H-6b). ESI MS, Calcd for C44H60N10O30S2:
1272.29 Da. Found: m/z 1295.42 [M + Na]+, 1273.42 [M+H]+.
2.a.9. 4HB-Carboprotein Dimer, 10. Template 9 (11 mg, 5.2 µmol)
was dissolved in DMSO (1 mL). Peptide 5 (8 mg, 3.4 µmol) was
dissolved in 0.1 M acetate buffer, pH 4.76 (4 mL). To initiate reaction,
54 µL (0.28 µmol) of the template solution was transferred to the
peptide solution. After 3 h, the mixture was directly injected to prep.
C4 RP-HPLC. After lyophilization of fractions containing product,
carboprotein 10 was obtained as a white powder. Yield 6 mg; 100%
(incl. 32× TFA). ESI MS, Calcd for C764H1260N170O230S2: 16571 Da
(average isotope). Found: m/z 1656.50 [M + 10H]10+, 1506.76 [M +
11H]11+, 1381.64 [M + 12H]12+, 1275.52 [M + 13H]13+, 1184.66 [M
2.a.4. 4-(Mercaptoacetamide)phenyl 2,3,4,6-tetra-O-(Aoa)-r-D-
Galp, 4. Compound 3 (33 mg, 20 µmol) was deprotected in trifluoro-
acetic acid (TFA)/DCM/Et3SiH (1:1:0.1, 4.2 mL) for 1 h. The colorless
solution was then concentrated to dryness, dissolved in DCM (2 mL),
and extracted with H2O (4 × 2 mL). The combined aqueous phases
were freeze-dried to give template 4 as a white powder. Yield 22 mg;
102% (incl. 4× TFA). 1H NMR (D2O), δ: 7.38 (d, 8.8 Hz, 2H, aryl),
7.12 (d, 8.7 Hz, 2H, aryl), 6.02 (d, 3.5 Hz, 1H, H-1), 5.81 (dd, 10.7
Hz, 3.0 Hz, 1H, H-3), 5.74 (d, 2.9 Hz, 1H, H-4), 5.54 (dd, 10.5 Hz,
3.8 Hz, 1H, H-2), 4.8-4.3 (m, 11H, CH2 Aoa/H-5/6), 3.35 (s, 2H,
CH2S). ESI MS, Calcd for C22H31N5O15S: 637.15 Da. Found: m/z
675.97 [M + K]+, 660.01 [M + Na]+, 638.15 [M+H]+.
+ 14H]14+, 1105.77 [M + 15H]15+, 1036.82 [M + 16H]16+
.
2.a.10. N-Phenyl-mercaptoacetamide disulfide, 11. Tritylmercap-
toacetic acid (200 mg, 0.60 mmol) was dissolved in DCM (10 mL).
Aniline (60 µL, 0.66 mmol) was added, followed by DIPCDI (139
µL, 0.90 mmol). After 18 h, the mixture was evaporated to dryness
and the residue dissolved in TFA/DCM/Et3SiH (1:1:0.1, 10.5 mL). After
30 min, the clear solution was again evaporated to dryness. The residue
was dissolved in DMSO/MeCN (1:1, 4 mL) and neutralized with 1 M
KHCO3 (1 mL). The resulting suspension was stirred in an open flask
for 18 h, before being concentrated to 1 mL, dissolved in MeCN (5
mL), and injected to prep. C18 RP-HPLC. Concentration of fractions
gave the title compound as white crystals. Yield 74 mg; 74%. m.p.
2.a.5. 4HB-carboprotein, 6. Template 4 (16 mg, 14.6 µmol) was
dissolved in H2O (2.1 mL). Peptide 5 (10 mg, 4.2 µmol) was dissolved
in MeCN/1 M acetate buffer, pH 4.76 (1:1, 2 mL). To initiate the
reaction, 100 µL (0.70 µmol) of the template solution was transferred
to the peptide solution. After 3 h, the mixture was directly injected to
prep. C4 RP-HPLC. After lyophilization of fractions containing product,
carboprotein 6 was obtained as a white powder. Yield 8 mg; 113%
(incl. 32× TFA). ESI MS, Calcd for C382H631N85O115S: 8287 Da
(average isotope). Found: m/z 1656.84 [M + 5H]5+, 1381.83 [M +
1
157-159 °C (lit. 161-163 °C43). H NMR (CD3OD), δ: 7.57 (d, 8.4
6H]6+, 1184.86 [M + 7H]7+, 1036.96 [M + 8H]8+, 921.86 [M + 9H]9+
829.76 [M + 10H]10+, 754.38 [M + 11H]11+
,
Hz, 4H, aryl), 7.31 (t, 7.3 Hz, 4H, aryl), 7.11 (t, 7.5 Hz, 2H, aryl),
3.68 (s, 4H, COCH2S). ESI MS, Calcd for C16H16N2O2S2: 332.07 Da.
Found: m/z 355.27 [M + H]+, 333.30 [M + Na]+.
.
2.a.6. 4-(Mercaptoacetamide)phenyl r-D-Galp disulfide, 7. Com-
pound 2 (197 mg, 0.34 mmol) was deprotected with TFA/DCM/Et3-
SiH (1:1:0.1, 21 mL) for 30 min. The colorless solution was then
concentrated to dryness, suspended in Et2O (50 mL), and extracted with
H2O (2 × 25 mL). DMSO (5 mL) was added to the aqueous phase.
The solution was stirred for 18 h in an open flask, and then purified
by prep. C18 RP-HPLC. Product 7 was obtained as white needles. Yield
2.b. Analytical Gel Filtration. Gel filtration was performed with a
Superdex 75 HR 10/30 column (Amersham Biosciences, Sweden)
connected to a Waters HPLC system (600 control unit, 996 PDA
detector, 717 Plus autosampler, Millenium 32 control software). A 100-
µL carboprotein sample was injected from a 2 mg/mL solution in 10
mM phosphate, pH 7.0 buffer. Elution buffer was 50 mM phosphate,
100 mM NaCl, pH 7.0. Samples were eluted with 1 mL/min flow rate
and detected by UV absorption at 215 nm. The column was calibrated
with albumin, ovalbumin, chymotrypsinogen A, ribonuclease A,
aprotinin, and vitamin B12.
2.c. Circular Dichroism Spectroscopy. CD spectra were recorded
on a Jasco J-710 instrument at 20 °C (Jasco, Inc., USA). Carboprotein
solutions were 20 µM (carboprotein 6) and 10 µM (carboprotein 10)
in 10 mM NaH2PO4, pH 7.0. The carboprotein concentration was
determined gravimetrically. The mean residue ellipticity was calculated
as [θ] ) θ/(10 × l × c × n), where θ [mdeg] is the measured ellipticity,
l [cm] the path length, c [mol L-1] the carboprotein concentration, and
n the number of amino acid residues in the carboprotein, that is, 60 for
6 or 120 for 10.
1
86 mg; 75%. m.p. 248-250 °C (decomp.). H NMR (DMSO, D2O-
exch.), δ: 7.46 (dd, 1.8 Hz, 6.5 Hz, 4H, aryl), 7.02 (dd, 2.2 Hz, 6.8
Hz, 4H, aryl), 5.30 (d, 2.3 Hz, 2H, H-1), 3.8-3.3 (m, 16H, H-2/3/4/
5/6/CH2S). 13C NMR (DMSO), δ: 167.5 (NHCO), [154.6, 134.0, 121.8,
118.6] (aryl), 99.8 (C-1), 73.2 (C-5), [70.4, 69.6, 69.1] (C-2/3/ 4), 61.4
(C-6), 44.1 (CH2S).
2.a.7. 4-(Mercaptoacetamide)phenyl 2,3,4,6-tetra-O-(Boc2-Aoa)-
r-D-Galp disulfide, 8. Compound 7 (80 mg, 0.12 mmol) and Boc2-
Aoa-OH (338 mg, 1.16 mmol) were suspended in pyridine/DCM (1:1,
4 mL). DMAP (14 mg, 0.12 mmol) and DIPCDI (180 µL, 1.16 mmol)
were added. Additional DIPCDI (2 × 180 µL) was added in two
portions during the next 4 h,. After a further 16 h, the suspension was
concentrated to dryness and CH3CN (5 mL) was added. Solid N,N′-
diisopropylurea was filtered off and the mixture was purified by prep.
C18 RP-HPLC. Product 8 was obtained as a clear, colorless oil. Yield
86 mg; 26%. 1H NMR (DMSO), δ: 7.56 (d, 8.8 Hz, 4H, aryl), 7.04 (d,
8.8 Hz, 4H, aryl), 5.67 (d, 3.4 Hz, 2H, H-1), 5.6-5.5 (m, 4H, H-3/4),
2.d. Preparation of Samples for Voltammetry, In Situ STM, and
XPS. Single-crystal gold bead electrodes for voltammetry and capaci-
(43) Schimelpfenig, C. W. J. Org. Chem. 1962, 27, 3323-3324.
9
96 J. AM. CHEM. SOC. VOL. 125, NO. 1, 2003

Synthesis of 4-R-Helix Bundle Carboprotein
A R T I C L E S
Figure 1. Chemical synthesis of 4HB-carboprotein monomer 6 and 4HB-carboprotein dimer 10. Ligation of hexadecapeptide aldehyde 5 to aminooxy-
functionalized templates 4 and 9 proceeded quantitatively by oxime formation.
tance measurements were prepared by the method of Clavilier44 and
Hamelin.45 A Au(111) disk (Surface Preparation Laboratory, The
Netherlands, www.surface-prep-lab.com, 12-mm diameter, 1 mm thick)
was used as substrate in STM and XPS. Before use, the electrodes
were electropolished in 0.1 M H2SO4 (10 V, followed by soaking in 1
M HCl) and annealed in a hydrogen flame. The disk substrate was
then cooled to room temperature, while the bead was kept for 1 min
above a Millipore water surface, followed by immersion into the water.
Substrate and bead electrodes were finally transferred to aqueous buffer
solution of the carboprotein or anchor molecule for several hours. The
carboprotein solutions were 0.23 mM protein in 0.1 M K2HPO4/KH2-
PO4 buffer (pH 6.9, 18 h), the anchor fragment 15 mM solution in
methanol (several hours). The samples were thoroughly rinsed with
Millipore water.
2.e. Voltammetry and Capacitance Measurements. Glassware in
electrochemical and in situ STM measurements was cleaned as
reported.37 The hang meniscus method was used in voltammetric and
capacitance measurements.37,38 CV, DPV, and interfacial capacitances
were recorded using an Autolab system (Eco Chemie, The Netherlands)
controlled by the general-purpose software. Parameters in DPV and
capacitance measurements were the same as previously37,38 (10 mV
s-1 scan rate and 4.05 mV step potential in DPV, 100 Hz and 5 mV
modulation amplitude in capacitance measurements). A coiled bright
platinum wire and a saturated calomel electrode (SCE) were used as
counter and reference electrode, respectively. All potentials refer to
the SCE. Solutions were deoxygenated by purified argon (Chrompack,
5N) before use and an argon atmosphere maintained above the solutions
during experimental recordings.
Inc., USA) equipped with a multichannel detector. A Mg-anode operated
at 200 W was used to generate X-rays. The pressure in the chamber
during data acquisition was less than 1 × 10-9 Torr. The pass energy
was either 25.0 or 50.0 eV. All spectra were acquired at room
temperature and are referenced to the Au(4f7/2) line at 84.00 eV.
3. Results
3.a. Synthesis and Structure of Carboprotein and Anchor
Fragment. Commercially available 4-nitrophenyl R-D-galacto-
pyranoside (R-D-Galp) was chosen as starting material for the
template synthesis, as reduction of the nitro group to an amine
allowed functionalization of the aglycon of the monosaccharide.
Taking advantage of the selectivity of the PyBOP coupling
reagent for forming amide bonds, chemoselective coupling of
TrSCH2COOH to 4-aminophenyl R-D-Galp, 1, gave 4-(trityl-
mercaptoacetamide)phenyl R-D-Galp, 2, in 88% yield (Figure
1). Next, the four hydroxyls were acylated with Boc2-Aoa-OH,
using a previously developed protocol,25 to give protected
template 3 in 62% yield. An alternative strategy, in which Boc2-
Aoa-OH acylation of 4-nitrophenyl R-D-Galp was followed by
reduction of the nitro group, proved unsuccessful, because
reduction by catalytic hydrogenation lead to a complex mixture.
Treatment of 3 with TFA leads to facile and quantitative removal
of all protecting groups (eight Boc and one Tr). Triethylsilane
was added to scavenge the liberated trityl cations, and thus to
prevent reattachment of the protecting group onto the thiol upon
concentration of the mixture.46 The lyophilized deprotected
template 4 was dissolved in a mixture of MeCN and aqueous
acetate buffer, pH 4.76, together with 50% excess of the
hexadecapeptide aldehyde Ac-YEELLKKLEELLKKAG-H, 5.
The 4HB-carboprotein monomer, 6, was formed during 3 h
under Ar by chemoselective oxime ligation47 and was obtained
in quantitative yield after prep. HPLC. The peptide aldehyde 5
was synthesized on solid-phase by a BAL strategy.42,48
2.f. In Situ STM. A PicoSPM instrument (Molecular Imaging Co.,
USA) with a bipotentiostat for independent control of substrate and
tip potential, and in-house-built three-electrode KEL-F cells were used,
in the constant-current mode. The substrate was a Au(111) disk,
compare with above. Reference and counter electrodes were platinum
wires; 10 mM phosphate (pH 6.8) was supporting electrolyte. Tungsten
tips were prepared and coated as described.38
2.g. XPS Measurements. The XPS spectra were recorded using a
Perkin-Elmer surface analysis system (Physical Electronic Industries,
(45) Hamelin, A. J. Electroanal. Chem. 1996, 411, 1-11.
(46) Pearson, D. A.; Blanchette, M.; Baker, M. L.; Guindon, C. A. Tetrahedron
Lett. 1989, 30, 2739-2742.
(44) Clavilier, J.; Faure, R.; Guinet, G.; Durand, R. J. Electroanal. Chem. 1980,
107, 205-209.
9
J. AM. CHEM. SOC. VOL. 125, NO. 1, 2003 97

A R T I C L E S
Brask et al.
Figure 2. The two reference structures, N-phenyl-mercaptoacetamide
disulfide, 11, and 4HB carboprotein without thiol anchor, 12.
Because of the difficulties with keeping 6 in the reduced form,
that is, to avoid disulfide formation, the 4HB-carboprotein dimer,
10, was prepared (Figure 1). Hence, 2 was treated with TFA to
deprotect the thiol group. The intermediate was extracted into
water, to which DMSO was added to facilitate oxidation to the
disulfide. 4-(Mercaptoacetamide)phenyl R-D-Galp disulfide, 7,
was formed over 18 h and purified to a white solid in 75%
yield. Boc2-Aoa-OH functionalization of all eight hydroxyl
groups proceeded, using the same protocol as for the monomer,
to give the desired protected template 8 in a modest yield of
26%. Next, 8 was quantitatively deprotected with TFA to give
template 9, which subsequently was ligated to peptide aldehyde
5 to give the 4HB-carboprotein dimer, 10, likewise in quantita-
tive yield. The ligation reaction was performed in aqueous
acetate buffer pH 4.76, containing a trace of DMSO, with 50%
excess of peptide aldehyde 5.
The thiol anchor aglycon of carboproteins 6 and 10 was
prepared as the disulfide 11 to provide a reference compound
in the electrochemical studies of the carboproteins. TrSCH2-
COOH and aniline were coupled with DIPCDI, after which the
trityl group was removed with TFA and Et3SiH scavenging,
and the resulting thiol oxidized to the disulfide with DMSO.
The desired compound, N-phenyl-mercaptoacetamide disulfide,
11, was obtained as a white solid in 74% overall yield (Figure
2).
Figure 3. ESI-MS of 4HB-carboprotein monomer 6 (top) and 4HB-
carboprotein dimer 10 (bottom). The insets show the deconvoluted spectra
(MassLynx transform algorithm). Calculated molecular masses (average
isotope distribution) for the two carboproteins are 8287 and 16 571 Da,
respectively.
Likewise, the 4HB carboprotein without thiol anchor, 12, was
prepared to provide a sulfur-free reference structure (Figure 2).
Preparation of 12 is reported elsewhere.41 Carboproteins 6 and
10 were characterized by ESI MS (Figure 3) and analytical gel
filtration (Figure 4). In both cases MS gave a series of multiply
protonated species, which could be deconvoluted to the putative
molecular ion, shown in the insets of Figure 3. For both
monomer and dimer carboprotein the correct mass was found,
8287 and 16 571 Da, respectively. Gel filtration clearly il-
lustrated the size difference of the two structures. By calibrating
the column with a series of globular proteins of known size,
the size of the 4HB-carboprotein monomer 6 could be estimated
to 10 707 Da. Similar for the 4HB-carboprotein dimer 10, a
size estimate of 22 064 Da was obtained. Hence, the dimer was
found to be slightly more than twice the size of the monomer.
Nonglobular shapes of the carboproteins can potentially explain
the difference between these results and the masses determined
by MS. The elution profiles (inset in Figure 4) showed close to
symmetrical peak shapes. However, the profile of the later
eluting 4HB-carboprotein monomer 6 revealed some disulfide
formation due to air-oxidation of the structure in the pH 7.0
Figure 4. Gel filtration (size exclusion chromatography) of 4HB-carbo-
protein monomer 6 and 4HB-carboprotein dimer 10. The linear correlation
between elution volume (Ve) and log MW was established from calibration
with albumin, ovalbumin, chymotrypsinogen A, ribonuclease A, aprotinin,
and vitamin B12. The inset shows the elution profiles of 6 (dashed line)
and 10 (fully drawn line).
phosphate buffer. Finally, carboproteins 6 and 10 were char-
acterized by CD spectroscopy (Figure 5). The spectra of both
structures showed a high content of R-helix. Based on the mean
residue ellipticity at 222 nm, the helicity was calculated to 57%
for 6 and 61% for 10.49
(47) Rose, K. J. Am. Chem. Soc. 1994, 116, 30-33.
(48) Jensen, K. J.; Alsina, J.; Songster, M. F.; Va´gner, J.; Albericio, F.; Barany,
G. J. Am. Chem. Soc. 1998, 120, 5441-5452.
9
98 J. AM. CHEM. SOC. VOL. 125, NO. 1, 2003

Synthesis of 4-R-Helix Bundle Carboprotein
A R T I C L E S
Figure 5. CD spectra of 4HB-carboprotein monomer 6 and 4HB-
carboprotein dimer 10, 10 mM phosphate buffer, pH 7.0, 20 °C.
Figure 6. Cyclic voltammograms of Au(111) in 100 mM phosphate, pH
6.9. Fully drawn line, bare electrode; dashed line, Au(111) after adsorption
of 4HB-carboprotein 6. Scan rate 10 mV s-1. Scan range: -0.2 to 1.2 V.
The inset shows an enlarged view of the voltammogram in the potential
range of Au(111)-surface reconstruction, -0.2 to 0.3 V.
3.c. Cyclic and Differential Pulse Voltammetry. Voltam-
metric data for reductive desorption and interfacial capacitances
were recorded for the 4HB-carboprotein 6, the 4HB-carboprotein
dimer 10, the 4HB-carboprotein without thiol anchor 12, and
for the anchor fragment 11. The carboproteins are stable at acidic
and neutral pH. Reductive desorption is usually effected under
basic conditions to avoid electrochemical dihydrogen evolution.
This interference was of minor importance in the present
investigation. All voltammetric and capacitance data were
recorded using 100 mM phosphate buffer (pH 6.9) where the
proteins are stable.
The fully drawn lines in Figure 6 show cyclic voltammograms
of Au(111) in the potential ranges of Au-oxide formation and
reduction and lifting of the surface reconstruction (inset). The
dashed lines show voltammograms in the same potential ranges
when the 4HB-carboprotein 6 is adsorbed. Oxidation is strongly
enhanced and the reconstruction peak disappears on carboprotein
adsorption, indicative of significant surface coverage. This is
substantiated by the data in Figures 7-9. Figure 7A, B shows
differential pulse voltammograms in the reductive desorption
region of the anchor fragment 11 (Figure 7A) and the three
different forms of the 4HB carboprotein (Figure 7B) in 0.1 M
phosphate buffer, pH 6.9. Figure 7C shows cyclic voltammo-
grams of the 4HB-carboprotein thiol monomer 6. The DPV of
anchor 11 displays a strong peak at -685 mV, with a shoulder
Figure 7. Differential pulse and linear sweep voltammograms for reductive
desorption of anchor fragment 11 and the three 4HB-carboprotein forms 6,
10, and 12, from Au(111) in 100 mM phosphate buffer, pH 6.9. Scan rate:
10 mV s-1. (A) Anchor 11. (a) First scan, (b) second scan, (c) third scan.
(B) 4HB carboproteins. Dashed line, 4HB carboprotein 6; dotted line, 4HB-
carboprotein dimer 10; fully drawn line, 4HB carboprotein without anchor
12. (C) Three successive linear sweep scans. The peak position in the first
scan is at -728 mV. The peak positions of the subsequent sans are shifted
to more positive potentials and the peaks become smaller with each scan.
on the negative side (Table 1). This accords with a reductive
CV desorption peak at -702 mV reported previously for the
anchor 11.50 The observation of a well defined desorption peak
clearly separate from the onset of dihydrogen evolution is not
common for thiol desorption at neutral pH. The two DPV peaks
accord, interestingly, with almost equally abundant highly
ordered domains and regions of adsorbate structural disorder
visible by in situ STM, cf. below. The DPV peak area, calibrated
toward the CV, corresponds to a coverage of (1.6 ( 0.3) ×
(50) Brask, J.; Wackerbarth, H.; Jensen, K. J.; Zhang, J.; Nielsen, J.U.; Andersen,
(49) Chen, Y.-H.; Yang, J. T.; Chau, K. H. Biochemistry 1974, 13, 3350-3359.
J. E. T.; Ulstrup, J. Bioelectrochemistry 2002, 56, 27-32.
9
J. AM. CHEM. SOC. VOL. 125, NO. 1, 2003 99

A R T I C L E S
Brask et al.
This pattern also resembles reductive desorption of the combined
anchor and scaffold fragments.50 Following conclusions from
reported patterns for reductive desorption of nonylthiolate,51 the
different patterns are likely to reflect different concentrations
and protonation states of liberated thiol anchor or combined
anchor and template fragments. Desorption at high pH liberates
deprotonated, negatively charged thiolate, which is repelled from
the surface at the high negative potentials. The adsorbate is
liberated as electrically neutral protonated thiolate at neutral pH,
where the thiol remains in significant concentrations close to
the electrode surface and is re-adsorbed much more easily than
the thiolate anion.
Figure 7B shows DPV of three different 4HB-carboprotein
forms, the thiol monomer 6, the disulfide dimer 10, and the
sulfur-free carboprotein without anchor 12 (Table 1). All three
forms show a peak around -750 mV but the following
differences are notable:
(1) The peak is by far the highest for the carboprotein
monomer 6. By comparison with the CV (Table 1) the peak
area corresponds to (2.4 ( 0.4) × 10-11 mol cm-2 or 40-80%
monolayer coverage assuming a 4HB radius of 10-12 Å, cf.
below. The peak area for disulfide 10 desorption corresponds
to only 30-50% monolayer coverage.
(2) The ratio between the coverages of anchor 12 and 4HB-
carboprotein 6 is about 5. This corresponds to an effective radius
ratio between the protein and the anchor molecule of 2.2 in the
packing on the Au(111) surface. The diameter of the four
parallel R-helices on the carbohydrate scaffold is 20-24 Å. The
radius of the thiol anchor in extended conformation is between
2.5 and 4 Å, giving a carboprotein/anchor coverage ratio
between 6 and 20. The observed ratio is thus on the lower side
of the ratio expected from sterically unhindered structures. This
could be indicative of lateral attraction in the carboprotein
monolayer on the Au(111) surface.
Figure 8. Differential capacitances on adsorption of the three 4HB
carboproteins on Au(111) from 100 mM phosphate buffer, pH 6.9. (A) Pure
phosphate buffer (4), 4HB carboprotein 6 (0), carboprotein 6 (O) after a
single negative potential excursion beyond the reductive desorption potential.
(B) Comparison of anchor 11 and 4HB carboproteins 6, 10, and 12. Anchor
fragment 11 (O), 4HB-carboprotein dimer 10 (0), 4HB-carboprotein
monomer 6 (4), 4HB carboprotein without anchor 12 (]).
(3) The CV and DPV peak potentials for the 4HB-monomer
6 are -743 and -728 mV, respectively. The peak potentials
for 4HB-carboprotein monomer 6 and for dimer 10 are very
close (i.e., -743 and -754 mV, respectively). This indicates
that disulfide adsorption is accompanied by disulfide bond
breaking. The lower coverage may then reflect sterically
hindered access and less adsorption of the disulfide form via
covalent Au-S bond formation.
(4) Subsequent scans of the adsorbed thiol monomer and
disulfide dimer give significantly smaller surface populations
but the peaks are largely undistorted. This follows the pattern
of the thiol anchor.
(5) The sulfur-free carboprotein 12 also gives a voltammetric
peak close to the reductive desorption peak of the thiol and
disulfide 4HB-carboproteins 6 and 10. The peak potential is
shifted negatively compared with 6 and 10 (i.e., to -759 mV).
The peak intensity is the smallest among the carboproteins.
There can be two origins of this peak. Oximes and oxime ethers,
which are the functionality linking the peptides to the template
in carboproteins, are known to undergo irreversible electro-
chemical four-electron or successive two-electron reductions
with cleavage of the NsO bond,52 in the same potential region
as observed here. The data in Figure 7 differ, however, from
Figure 9. XPS overview spectrum of Au(111) after adsorption of 4HB
carboprotein 6. Pass energy 50 eV. (Inset) XPS spectrum in the S(2p) region.
Pass energy, 25 eV.
10-10 mol cm-2, or close to a monolayer assuming one-electron
transfer reductive desorption. Subsequent scans disclose more
conspicuous double peaks of decreasing peak areas but in the
same potential range. This is different from reductive desorption
in strongly basic solution, pH 11.9, where subsequent scans give
desorption peaks strongly displaced toward positive potentials.50
(51) Yang, D.-F.; Wilde, C. P.; Morin, M. Langmuir 1996, 12, 6570-6577.
(52) Kapetanovic, V.; Aleksic, M.; Zuman, P. J. Electroanal. Chem. 2001, 507,
263-269.
9
100 J. AM. CHEM. SOC. VOL. 125, NO. 1, 2003

Synthesis of 4-R-Helix Bundle Carboprotein
A R T I C L E S
Table 1. Total Charge Determined from DPV and Interfacial Capacitances of the Three 4-R-Helix Bundle Carboproteins and the Thiol
Anchor Molecule in the Reductive Desorption Potential Region (First Two Columns)a
peak,
mV
DPV,
µC‚cm
capacitance,
µC‚cm-2
net charge,
µC‚cm
coverage,
pmol‚cm-2
-2
-2
%
4HB-carboprotein monomer 6
4HB-carboprotein dimer 10
sulfur-free carboprotein 12
anchor fragment 11
-743 ( 19
-754 ( 2
-759 ( 12
-685 ( 18
3.7 ( 1.4
2.4 ( 0.4
1.3 ( 0.3
17.3 ( 3.3
0.8
0.7
0.6
1.6
20
30
50
10
2.9
1.7
0.7
30 ( 15
18 ( 4
15.7
163 ( 34
a “%” indicates the capacitive contribution to the total charge. The fourth column shows the Faradaic contribution and the fifth column the coverage in
picomoles per square centimeter.
Table 2. Atomic Concentration of the Elements C, O, and N in
Surface-Immobilized 4HB Carboprotein 6 on Au(111) Obtained
electrochemical oxime reduction data by involving an adsorbed
monolayer, rather than diffusion of bulk oxime. The observation
that subsequent scans record a successively decreasing peak at
the same potential is also not obviously in keeping with
significant irreversible Faradaic oxime reduction. Alternatively
the peak could be caused by desorption of “unspecifically
adsorbed” sulfur-free carboprotein. This is supported by the
interfacial capacitance peak in the same potential range, cf.
below. Such a feature would be expected to be far more
important for the sulfur-free carboprotein 12 than for the thiol
and disulfide forms 6 and 10 at high coverage, where linking
of the carboprotein via the thiol anchor will block the access of
the helix parts of the carboprotein to the Au(111) surface. Steric
hindrance would also block the access of the potentially
reducible oxime groups to the surface from other carboprotein
fragments.
3.d. Interfacial Capacitance Data. Adsorption of 4HB-
carboproteins 6, 10, and 12, and of anchor fragment 11, as well
as the reductive desorption dynamics is reflected in interfacial
differential capacitance patterns (Figure 8). The strong peak at
0.1 V is caused by adsorption and desorption of mono- and
dihydrogen phosphate ions. The square symbols show the
capacitance after adsorption of 4HB-carboprotein 6. The
capacitance has decreased from the peak value of about 130
µF cm-2 in pure buffer to less than 10 µF cm-2 in the whole
potential range. The figure shows, further the differential
capacitance after a single negative potential excursion beyond
the reductive desorption potential. Partial reappearance of the
phosphate adsorption peak is noted, in keeping with the data
from DPV. Figure 8B shows the differential capacitances of
all the structures, that is, carboproteins 6, 10, and 12, and anchor
fragment 11. Together with the voltammetric data the following
observations are notable:
from XPS
atomic concentration, %
N
O
C
calcd
measd
19.8
18.4
14.6
12.2
65.6
67.2
carboprotein 6 on Au(111). The carboprotein monolayer gives
a marked lowering of the Au(4f) intensity compared with bare
Au(111). The bare surface thus discloses 60% Au-atom
concentration determined by the intensity of the Au(4f) transition
but 30% after carboprotein adsorption, indicative of extensive
carboprotein coverage. The atomic composition of the elements
C, N, O from the carboprotein stoichiometry and the XPS
spectra accord well such as summarized in Table 2. The inset
in Figure 9 shows the XPS spectrum in the sulfur 2p region.
Although significantly broadened, two peaks at 162.2 and 163.3
eV could always be clearly distinguished. These are character-
istic of thiolate headgroup interactions with the gold sur-
face,38,53,54 and indicative of strong carboprotein adsorption via
the thiol anchor.
3.f. In Situ STM of 4HB-Carboprotein Monomer 6 and
Anchor Fragment 11. The voltammetric, interfacial capaci-
tance, and XPS data point coherently toward extensive mono-
layer coverage of the Au(111)-electrode surface either by anchor
fragment 11 or full 4HB carboprotein 6 and 10. This was
supported by in situ STM under full electrochemical potential
control, directly in the aqueous buffer. In situ STM discloses a
monolayer at the molecular and supramolecular levels. Figure
10A, B shows representative in situ STM images of anchor 11
adsorbed on Au(111) in 10 mM aqueous buffer solution, pH
6.8. Domains of supramolecular organization can clearly be
distinguished even though the individual domains do not possess
long-range order. A large number of intermediate-size pits (30-
50 Å) are also apparent. The domains are oriented in three
directions rotated relative to each other and in almost equal
abundance. Molecules in each domain seem to be oriented in
rows or “pin stripes”, illustrated by the height profiles along a
given row and perpendicular to a set of rows (Figure 10A). High
anchor molecule coverage and significant local structural order
thus emerge. The domains are separated by disordered regions
in between. As noted the areas of the ordered domains and the
disordered domain boundary regions correlated roughly (Figure
7). Figure 10B shows a high-resolution in situ STM image of
a monolayer domain of anchor fragment 11. Different bias
(1) Anchor 11 shows a single strong peak with a peak height
of ≈25 µF cm-2 at the reductive desorption potential compared
with a featureless background of ≈5 µF cm-2
.
(2) All the carboproteins also show a peak close to the DPV
peak potentials of reductive desorption, but less intensive than
for the anchor 11. The 4HB-carboprotein monomer 6 shows a
lower capacitance than the 4HB-carboprotein dimer 10 and the
4HB-carboprotein 12 without linker. As for the anchor 11, the
monomer 6 also shows a featureless background but weak
structural features are apparent at ≈0.0 V and ≈0.4 V, that is,
around the potential of zero charge, for the 4HB-carboprotein
dimer 10 and 4HB carboprotein without anchor 12.
(3) Peak features in the phosphate adsorption region reappear
after a negative potential excursion beyond the reductive
desorption potential.
3.e. XPS Data. Figure 9 shows a representative overview
spectrum in the binding energy range 0-1000 eV of 4HB-
(53) Bain, C. D.; Troughton, E. B.; Tao, Y.-T.; Evall, J.; Whitesides, G. M.;
Nusso, R. G. J. Am. Chem. Soc. 1989, 111, 321-335.
(54) Walcsak, M. M.; Alves, C. A.; Lamp, B. B.; Porter, M. D. J. Electroanal.
Chem. 1995, 396, 103-114.
9
J. AM. CHEM. SOC. VOL. 125, NO. 1, 2003 101

A R T I C L E S
Brask et al.
Figure 11. In situ STM images of 100 nm × 100 nm area of 4HB
carboprotein 6 adsorbed on Au(111), 10 mM phosphate buffer, pH 6.8.
Constant current mode. Tunneling current, 0.3 nA. Scan rate, 9.8 s-1. (A)
Sample potential, 0.34 V. Bias voltage, -0.5 V. (B) Sample potential, -0.71
V. Bias voltage, 0.35 V.
potentials, and sample potentials closer to the reductive desorp-
tion, cf. below, were here needed for optimal imaging. Both
single molecules and the supramolecular striped domain struc-
ture can, however, be distinguished.
Figure 11 shows, finally, two in situ STM images of the Au-
(111)-surface covered by 4HB-carboprotein 6 at two different
substrate potentials. Figure 11A is recorded at a potential on
the far positive side of the reductive desorption potential. The
Au(111)-surface appears fully covered, with the same 30-50
Å pits scattered over the surface as for the anchor fragment 11.
Figure 11B is recorded at a potential, close to the reductive
desorption potential. The adsorption pattern here is quite
different from the pattern at high potential. Smaller coverage
is apparent, with individual single-molecule size structures
distinguishable. The controlling in situ STM bias voltage had
to be set differently for optimal image quality in the two images,
but the “molecular” in situ STM contrasts are noted to disclose
Figure 10. In situ STM of the anchor fragment 11 adsorbed on Au(111),
10 mM phosphate, pH 6.8. Constant current mode. (A) Image area, 100
nm × 100 nm. Tunneling current, 0.3 nA. Sample potential, -460 mV.
Bias voltage, 0.2 V. Scan rate, 13 s-1. The height profiles are recorded
perpendicular to a set of rows (a) and along a given row (b) indicated by
the lines “a” and “b” in the image. (B) Image area 10 nm × 10 nm.
Tunneling current, 3 nA. Sample potential, -760 mV. Bias voltage, 0.15
V. Scan rate, 9.8 s-1
.
9
102 J. AM. CHEM. SOC. VOL. 125, NO. 1, 2003

Synthesis of 4-R-Helix Bundle Carboprotein
A R T I C L E S
a different character close to the reductive desorption potential
(Figure 11B) and far from this potential (Figure 11A). A similar
but less pronounced difference is apparent for the two images
of anchor 11 in Figure 10A, B.
Natural and synthetic redox proteins hold exciting perspec-
tives in new biotechnology, approaching control of biotechno-
logical function at the nanoscale and single-molecule levels.
Prerequisites are that the proteins can be immobilized in ordered
functional states on well-defined solid substrates, which also
interface to electrical, mechanical, and other external circuits.
In this report we have shown that 4HB proteins can be brought
to immobilization on single-crystal, atomically planar Au(111)
surfaces. CV, DPV, interfacial capacitance, and XPS data have
mapped a coherent pattern of close to monolayer coverage of
both 4HB carboprotein 6 and thiol anchor 11. The protein
monolayer is functional in the sense that the proteins chemi-
sorbed via covalently attached thiol anchors display the expected
features of well-defined Au-S bond formation. The disulfide
4HB-carboprotein dimer 10 is adsorbed in the same potential
range, but the peak area is much smaller. The carboprotein
without linker 12 shows a peak at the lowest potential and has
the smallest peak area of the three carboproteins. The reductive
desorption peak of the thiol 4HB-carboprotein monomer 6 is
therefore also likely to hold a capacitive component (Table 1).
Discussion
The 4HB carboproteins constitute a new class of synthetic
proteins with structures based on R-helices mounted on a
monosaccharide scaffold.24-26 Other classes of 4HB synthetic
proteins based on peptide templates27,30,31 and disulfide link-
ing28,29 have been reported and subjected to extensive physical-
chemical characterization. Hence, recently developed chemical
methodologies have put model proteins within the reach of
synthetic chemistry. Merits of the synthetic proteins are that
they hold options for systematic sequence and structure varia-
tion, so that the effects of charge, hydrophobic surface patches,
ligation, and broad secondary and tertiary structural features
become available. In these ways their perspectives differ from
those offered by microbiological structure modification (site
directed mutagenesis) and from attempts toward total synthesis
of natural redox and nonredox proteins,55-57which still pose
stricter limitations in these respects.
We have, further, obtained in situ STM images of both the
adsorbed anchor 11 and the full 4HB carboprotein 6. The
resolution discloses both supramolecular organization and single-
molecule features, which has not been reported previously for
4HB proteins. The in situ STM data also define electrochemical
potential ranges of adsorbed monolayer stability. In situ STM
of redox and nonredox proteins under full potentiostatic control
is, further, still novel and has only been reported in a very few
cases.32-38,59 The anchor fragment 11 forms monolayers with
lateral domain order but small lateral extension of the individual
domains. The domains extend in three different directions
oriented 60° relative to each other and separated by almost
equally abundant interdomain regions. This accords with two
peaks in the DPV for reductive desorption of roughly the same
intensity ratio. 4HB carboprotein 6 clearly forms dense mono-
layers but without lateral order such as for anchor 11. Lack of
two-dimensional lateral order is a common observation in the,
still very few, reported cases of in situ STM of biological
macromolecules. The character of the monolayer and imaging
contrast changes drastically as the electrochemical potential for
reductive desorption is approached. A decrease in the coverage
is clearly distinguishable, accompanied by increased apparent
spatial resolution toward the single-molecule level. The in situ
STM contrast also changes toward strongly enhanced tunneling.
With reference to recent theoretical approaches to molecular in
situ STM mechanisms,60-62 it could be suggested that different
in situ STM tunneling channels are operative close to the
reductive desorption potential and far from this potential.
Electron tunneling far from reductive desorption is effected by
superexchange via the lowest unoccupied molecular orbital
(LUMO) of the molecular adsorbate. This level is far off
resonance relative to the Fermi levels of the substrate and tip
in these potential ranges. As the Au(111) substrate potential
approaches the reductive desorption potential, the LUMO is
For the purpose of studying proteins adsorbed on gold
surfaces, it is desirable that a single thiol or disulfide moiety is
present at a specific position in the structure. Except for the
very first reported carboprotein, in which the pyranoside aglycon
was used to link the structure to a solid phase during synthesis
of the four peptide strands,24 the potential of functionalizing
carboproteins at the template aglycon has not previously been
utilized.58 It was envisioned that a thiol anchor at the free
aglycon of a carboprotein structure would give adsorption
without interference with the folding of the structure. The 4HB-
carboprotein monomer, 6, was prepared based on this strategy.
However, as analytical gel filtration result (Figure 4) showed,
some degree of air oxidation to the disulfide form was
unavoidable. To circumvent the presence of both monomers and
dimers, the 4HB-carboprotein dimer, 10, was assembled on
dimeric template 9. In both cases, chemoselective oxime ligation
between the peptide aldehyde and the aminooxy-functionalized
template gave the desired carboprotein in quantitative yield and
high purity, as judged by HPLC and MS (Figure 3). Previous
studies of the 4HB carboprotein without thiol anchor, 12, have
indicated that the four peptide strands on the d-Galp template
do interact to form a bundle structure.41 Hence, it is reasonable
to assume that carboprotein 6, which only differs from 12 in
the aglycon, forms a similar 4HB structure. Likewise with the
dimeric structure 10, although it is uncertain if this carboprotein
forms two 4HB structures or if all peptide strands interact to
form a 8HB structure. The CD spectra of 6 and 10 supported
the hypothesis of bundle structures as both carboproteins were
found to be highly R-helical (Figure 5). Because sample
concentrations were determined gravimetrically on milligram
quantities of material, there are some uncertainties as to the
exact amount of R-helix in the structures, calculated to 57%
and 61% for 6 and 10, respectively. In a previous study, the
helicity of carboprotein 12 was calculated to 64%.41
(59) Andolfi, L.; Cannistrarro, S.; Canters, G. W.; Facci, P.; Ficca, A. G.; Van
Amsterdam, I. M. C.; Verbeet, M. P. Arch. Biochem. Biophys. 2002, 399,
81-88.
(55) Christensen, H. E. M.; Hammerstad-Pedersen, J. M.; Holm, A.; Iversen,
G.; Jensen, M. H.; Ulstrup, J. Eur. J. Biochem. 1994, 224, 97-101.
(56) Hellinga, H. W.; Marvin, J. S. Trends Biotechnol. 1998, 16, 183-189.
(57) Kennedy, M. L.; Gibney, B. R. Curr. Opion. Struct. Biol. 2001, 11, 485-
490.
(60) Kuznetsov, A. M.; Sommer-Larsen, P.; Ulstrup, J. Surf. Sci. 1992, 275,
52-64.
(61) Friis, E. P.; Kharkats, Y. I.; Kuznetsov, A. M.; Ulstrup, J. J. Phys. Chem.
A 1998, 102, 7851-7859.
(58) Jensen, K. J.; Brask, J. Cell. Mol. Life Sci. 2002, 59, 859-869.
(62) Kuznetsov, A. M.; Ulstrup, J. J. Phys. Chem. A 2000, 104, 11531-11540.
9
J. AM. CHEM. SOC. VOL. 125, NO. 1, 2003 103

A R T I C L E S
Brask et al.
brought close to the Fermi level of the Au(111) electrode and
now mediates in situ STM by temporary LUMO population and
depopulation. This enhances tunneling significantly. Unfortu-
nately, the expected reVersible dependence of this effect cannot
be tested due to the significant adsorbate depopulation which
accompanies the redox process.
further, be distinguished in electrochemical potential ranges far
from the reductive desorption potential and close to this
potential.
The primary target molecules of our investigation, the 4HB
carboproteins, do not hold a metal redox center. Histidines and
other residues suitable for heme group attachment can, however,
be inserted into the carboproteins similar to other 4HB
structures.27-31 The hydrophobic heme group could also be
trapped by noncovalent hydrophobic interactions in the space
between the four organized R-helices, which have been designed
with a hydrophobic interior and a hydrophilic surface in contact
with the outer aqueous solution. Such a confinement would
endow the enclosed heme group with considerable internal
mobility. This would detract from options of addressing electron
transfer distances and related notions in electron transfer theory66
but hold other favorable perspectives relating to efficient
electronic transmission in the in situ STM mode and more
broadly, toward working biosensor function. A mobile heme
group could thus efficiently both accept and donate electrons
at short distances from the substrate electrode and tip, if these
two steps were separated by internal translation. Work along
these lines is in progress.
Heme groups and blue copper-center binding to synthetic 4HB
proteins has been extensively reported recently and characterized
spectroscopically and voltammetrically.28,29,63-65 The 4HB car-
boproteins offer similar options, but our approach has differed
somewhat from the previous reports. We have, first, exploited
a comprehensive, state-of-the-art physical electrochemical ap-
proach. A primary basis has been the use of bead and disk Au-
(111) electrodes. The surfaces of these electrodes are atomically
planar over up to several hundred nanometers. The single-crystal
electrodes have been combined with electrochemical (CV, DPV,
and capacitance measurements) and spectroscopic (XPS) char-
acterization, with focus on the Au-S bonding and electrode
coverage of adsorbed carboprotein and anchor structures. The
single-crystal electrochemical approach has, secondly, paved the
way for addressing the adsorbate monolayers by in situ STM
toward the single-molecule level. Single-molecule resolution has
been achieved for both the anchor fragment 11 and the full 4HB
carboprotein 6, but only the former shows supramolecular order
on the Au(111) surface. Different in situ STM contrasts can,
Acknowledgment. Financial support from The Danish
Technical Science Research Council (grant 26-00-0034) is
acknowledged.
(63) Chen, X. X.; Discher, B. M.; Pilloud, D.L.; Gibney, B. R.; Moser, C. C.;
Dutton, P. L. J. Phys. Chem. B 2002, 106, 617-624.
JA020943R
(64) Katz, E.; Heleg-Shabtai, V.; Bardea, A.; Willner, I.; Rau, H.; Haehnel, W.
Biosens. Bioelectron. 1998, 13, 741-756.
(65) Shifman, J. M.; Gibney, B. R.; Sharp, R. E.; Dutton, P. L. Biochemistry
2000, 39, 14813-14821.
(66) Kuznetsov, A. M.; Ulstrup, J. Electron Transfer in Chemistry and Biology.
An Introduction to the Theory; Wiley: Chichester, UK, 1999.
9
104 J. AM. CHEM. SOC. VOL. 125, NO. 1, 2003