A molecular keypad lock: A photochemical device capable of authorizing password entries

Article abstract of DOI:10.1021/ja065317z

This paper describes a new concept in the way information can be protected at the molecular scale. By harnessing the principles of molecular Boolean logic, we have designed a molecular device that mimics the operation of an electronic keypad lock, e.g., a

Full text of DOI:10.1021/ja065317z

Published on Web 12/19/2006
A Molecular Keypad Lock: A Photochemical Device Capable
of Authorizing Password Entries
David Margulies, Clifford E. Felder, Galina Melman, and Abraham Shanzer*
Contribution from the Department of Organic Chemistry, The Weizmann Institute of Science,
RehoVot 76100, Israel
Received July 25, 2006; E-mail: abraham.shanzer@weizmann.ac.il
Abstract: This paper describes a new concept in the way information can be protected at the molecular
scale. By harnessing the principles of molecular Boolean logic, we have designed a molecular device that
mimics the operation of an electronic keypad lock, e.g., a common security circuit used for numerous
applications, in which access to an object or data is to be restricted to a limited number of persons. What
distinguishes this lock from a simple molecular logic gate is the fact that its output signals are dependent
not only on the proper combination of the inputs but also on the correct order by which these inputs are
introduced. In other words, one needs to know the exact passwords that open this lock. The different
password entries are coded by a combination of two chemical and one optical input signals, which can
activate, separately, blue or green fluorescence output channels from pyrene or fluorescein fluorophores.
The information in each channel is a single-bit light output signal that can be used to authorize a user, to
verify authentication of a product, or to initiate a higher process. This development not only opens the way
for a new class of molecular decision-making devices but also adds a new dimension of protection to
existing defense technologies, such as cryptography and steganography, previously achieved with molecules.
Introduction
computers, electronic logic circuits control the operation of a
variety of devices around us,32 from calculators and store
The increasing role electronic devices play in our daily lives,
as well as our constant need to pursue superior technologies,
have raised a wide interest in the development of molecular
systems mimicking the operation of electronic logic gates and
circuits.1-31 Besides their integration in the heart of digital
automation to video games and music equipment. In the past
few years, the feasibility of molecular logic gates to be integrated
into practical decision-making devices has also been demon-
strated. Although these devices so far include merely an
interactive game,33 a few moleculators34 (see also refs 35-50),
and a comparator,50 such progress raises an unavoidable thought
(1) de Silva, A. P.; McClenaghan, N. D. Chem.-Eur. J. 2004, 10, 574-586.
(2) de Silva, A. P.; Gunaratne, H. Q. N.; McCoy, C. P. Nature 1993, 364,
42-44.
(3) de Silva, A. P.; McClenaghan, N. D. Chem.-Eur. J. 2002, 21, 4935-
(22) Ashkenasy, G.; Ghadiri, M. R. J. Am. Chem. Soc. 2004, 126, 11140-11141.
(23) Deonarine, A. S.; Clark, S. M.; Konermann, L. Future Gener. Comput.
Syst. 2003, 19, 87-97.
4945.
(4) Uchiyama, S.; McClean, G. D.; Iwai, K.; de Silva, A. P. J. Am. Chem.
Soc. 2005, 127, 8920-8921.
(24) Montenegro, J.-M.; Perez-Inestrosa, E.; Collado, D.; Vida, Y.; Suau, R.
Org. Lett. 2004, 6, 2353-2355.
(5) Callan, J. F.; de Silva, A. P.; Magri, D. C. Tetrahedron 2005, 61, 8551-
8588.
(25) Szacitowski, K. Chem.-Eur. J. 2004, 10, 2520-2528.
(26) Petitjean, A.; Kyritsakas, N.; Lehn, J.-M. Chem.-Eur. J. 2005, 11, 6818
- 6828.
(6) Magri, D. C.; Brown, G. J.; McClean, G. D.; de Silva, A. P. J. Am. Chem.
Soc. 2006, 128, 4950-4951.
(7) Balzani, V.; Credi, A.; Venturi, M. ChemPhysChem 2003, 3, 49-59.
(8) Credi, A.; Balzani, V.; Langford, S. J.; Stoddart, J. F. J. Am. Chem. Soc.
1997, 119, 2679-2681.
(27) Kompa, K. L.; Levine, R. D. Proc. Natl. Acad. Sci. 2001, 98, 410-414.
(28) Shiraishi, Y.; Tokitoh, Y.; Hirai, T. Chem. Commun. 2005, 42, 5316-
5318.
(9) Collier, C. P.; Wong, E. W.; Belohradsky, M.; Raymo, F. M.; Stoddart, J.
F.; Kuekes, P. J.; Williams, R. S.; Heath, J. R. Science 1999, 285, 391-
394.
(29) Straight, S. D.; asson, J. A.; Kodis, G.; Bandyopadhyay, S.; Mitchell, R.
H.; Moore, T. A.; Moore, A. L.; Gust, D. J. Am. Chem. Soc. 2005, 127,
9403-9409.
(30) Sousa, M. d.; Castro, B. d.; Abadb, S.; Mirandab, M. A.; Pischel, U. Chem.
Commun. 2006, 19, 2051-2053.
(10) Pina, F.; Roque, A.; Melo, M. J.; Maestri, M.; Livia Belladelli; Balzani,
V. Chem.-Eur. J. 1998, 1184-1191.
(11) Pina, F.; Maestri, M.; Balzani, V. Chem. Commun. 1999, 107-114.
(12) Gunnlaugsson, T.; MacDo´nail, D. A.; Parker, D. Chem. Commun. 2000,
93-94.
(31) Tian, H.; Wang, Q.-C. Chem. Soc. ReV. 2006, 35, 361-374.
(32) http://www.educypedia.be/electronics/circuits.htm
(33) Stojanovic, M. N.; Stefanovic, D. Nat. Biotechnol. 2003, 21, 1069-1074.
(34) Margulies, D.; Melman, G.; Shanzer, A. J. Am. Chem. Soc. 2006, 128,
4865-4871.
(13) Raymo, F. M. AdV. Mater. 2002, 14, 401-414.
(14) Raymo, F. M.; Giordani, S. Proc. Natl. Acad. Sci. 2002, 99, 4941-4944.
(15) Guo, X.; Zhang, D.; Wangab, T.; Zhu, D. Chem. Commun. 2003, 914-
915.
(35) de Silva, A. P.; McClenaghan, N. D. J. Am. Chem. Soc. 2000, 122, 3965-
3966.
(16) Guo, X.; Zhang, D.; Zhu, D. AdV. Mater. 2004, 16, 125-127.
(17) Stojanovic, M. N.; Semova, S.; Kolpashchikov, D.; Macdonald, J.; Morgan,
C.; Stefanovic, D. J. Am. Chem. Soc. 2005, 127, 6914-6915.
(18) Stojanovic, M. N.; Mitchell, T. E.; Stefanovic, D. J. Am. Chem. Soc. 2002,
124, 3555-3561.
(19) Baytekin, H. T.; Akkaya, E. U. Org. Lett. 2000, 2, 1725-1727.
(20) Turfan, B.; Akkaya, E. U. Org. Lett. 2002, 4, 2857-2859.
(21) Saghatelian, A.; Volcker, N. H.; Guckian, K. M.; Lin, V. S.-Y.; Ghadiri,
M. R. J. Am. Chem. Soc. 2003, 125, 346-347.
(36) Remacle, F.; Speiser, S.; Levine, R. D. J. Phys. Chem. B 2001, 105, 5589-
5591.
(37) Stojanovic, M. N.; Stefanovic, D. J. Am. Chem. Soc. 2003, 125, 6673-
6676.
(38) Langford, S. J.; Yann, T. J. Am. Chem. Soc. 2003, 125, 11198-11199.
(39) Margulies, D.; Melman, G.; Felder, C. E.; Arad-Yellin, R.; Shanzer, A. J.
Am. Chem. Soc. 2004, 126, 15400-15401.
(40) Guo, X. F.; Zhang, D. Q.; Zhang, G. X.; Zhu, D. B. J. Phys. Chem. B
2004, 108, 11942-11945.
9
10.1021/ja065317z CCC: $37.00 © 2007 American Chemical Society
J. AM. CHEM. SOC. 2007, 129, 347-354
347

A R T I C L E S
Margulies et al.
regarding our ability to create a molecular analogue for every
electronic product.
An important electronic logic device, which has not yet been
mimicked at the molecular level, is a keypad lock.32 This device
is used for numerous applications in which access to an object
or data is to be restricted to a limited number of persons. What
complicates an electronic keypad circuit over a simple logic
gate is the fact that its output signals are dependent not only on
the proper combination of the inputs but also on the correct
order by which these inputs are introduced. In other words, one
needs to know the exact password that opens this lock.
The development of a molecular-scale keypad lock is a
particularly attractive goal as it represents a new approach for
protecting information at the molecular scale. Molecular security
is currently achieved by steganography or cryptography meth-
ods,51 used to hide or to encrypt a message, respectively.
Invisible inks are a classic example of chemical steganography.
These materials provide hidden images that cannot be photo-
copied, thus securing the authentication of a variety of products.
Enhanced security can be achieved using erasable materials52
and their compression into high-density storage devices.53
Molecular cryptography has also been accomplished through
various methods for encrypting messages on DNA strands.54-56
In this paper, we propose a password entry, considered as
the front line of defense against intruders, as a third approach
for securing information at the molecular scale. A password
entry is used neither to hide nor to encrypt information but rather
to prevent access from an unauthorized user.51 We prove the
feasibility of this approach by demonstrating a reconfigurable
molecular keypad lock whose fluorescence is revealed only in
response to correct sequences of three input signals. We first
describe the design and function of a programmable molecular
device, capable of executing a range of Boolean functions. Then,
we depict the way these functions could be integrated into a
more complex circuitry at the molecular scale. Specifically, we
have developed a kinetically controlled, priority-AND molecular
logic gate, capable of authorizing different photoionic pass-
words.
Figure 1. (Left) Three-input molecular AND logic gate and (right) three-
input electronic keypad circuit. What complicates an electronic keypad
circuit over a simple logic gate is the fact that its output signal is dependent
not only on the proper combination of the inputs but also on the correct
order by which these inputs are introduced. The lock, obtained from the
integration of several AND logic gates, generates an electronic output only
when the relevant password keys (ABC) are pushed in this exact order.
logic gates, that can control an opening of a door or a safe, an
alarm system, and more. Figure 1 describes a 3-input AND logic
gate (left) and a particular example of a simple electronic lock
(right), which also processes three input signals. Both devices
generate a high output in response to three high inputs, namely,
A and B and C; however, the electronic keypad lock is more
selective, as it produces an electrical signal to only one (ABC)
out of six possible combinations of ABC, ACB, BAC, BCA,
CAB, and CBA. Specifically, in the electronic lock, which
integrates three 2-input AND logic gates, LED D1 will switch
ON only when buttons S1, S2, S3 are pushed in the right order.
Pin 1 is held HIGH by R1. This enables gate U1A, and when
button S1 is pressed, the output at pin 3 will go HIGH. It locks
itself ON through R2 and enables gate U1B, by taking pin 4
HIGH. Now, if S2 is pressed, the output of gate U1B will lock
itself ON through R3, and by taking pin 9 HIGH, enable gate
U1C. Pressing S3 will cause gate U1C to do the same thing,
only this time its output at pin 8 through R5 turns LED D1
ON. Each of the other buttons in the keypad can be connected
to a RESET button SR, so if pressed the circuit will reset and
the password entry will fail. During this moment, pin 1 goes
LOW, hence the output at pin 3 will go LOW. It locks itself
OFF through R2 and disables gate U1B, by taking pin 4 LOW.
In a similar way, the output at pin 6 will go LOW and the output
at pin 8 will go LOW. This will switch LED D1 OFF.
Results and Discussion
An Electronic Keypad Lock. An electronic keypad lock is
a common security device, comprised of several interconnected
(41) Andreasson, J.; Kodis, G.; Terazono, Y.; Liddell, P. A.; Bandyopadhyay,
S.; Mitchell, R. H.; Moore, T. A.; Moore, A. L.; Gust, D. J. Am. Chem.
Soc. 2004, 126, 15926-15927.
(42) Okamoto, A.; Tanaka, K.; Saito, I. J. Am. Chem. Soc. 2004, 126, 9458-
9463.
(43) Margulies, D.; Melman, G.; Shanzer, A. Nat. Mater. 2005, 4, 768-771.
(44) Coskun, A.; Deniz, E.; Akkaya, E. U. Org. Lett 2005, 7, 5187-5189.
(45) Li, F.; Shi, M.; Huang, C.; Jin, L. J. Mater. Chem. 2005, 15, 3015-3020.
(46) Qu, D.-H.; Wang, Q.-C.; Tian, H. Angew. Chem., Int. Ed. 2005, 44, 5296
-5299.
Design. To realize a molecular logic device capable of
distinguishing between different input sequences, we exploit our
experience in the design and synthesis of fluorescent biosensors
mimicking the function of siderophores;57-71 these are the
(47) Remacle, F.; Weinkauf, R.; Levine, R. D. J. Phys. Chem. A 2006, 110,
177-184.
(48) Baron, R.; Lioubashevski, O.; Katz, E.; Niazov, T.; Willner, I. J. Phys.
Chem. A 2006, 110, 8548-8553.
(57) Shanzer, A.; Libman, J.; Lazar, R.; Tor, Y. Pure Appl. Chem. 1989, 61,
1529-1534.
(49) Lederman, H.; Macdonald, J.; Stefanovic, D.; Stojanovic, M. N. Biochem-
istry 2006, 45, 1194-1199.
(58) Berner, I.; Yakirevitch, P.; Libman, J.; Shanzer, A.; Winkelmann, G. Biol.
Met. 1991, 4, 186-191.
(50) Liu, Y.; Jiang, W.; Zhang, H.-Y.; Li, C.-J. J. Phys. Chem. B. 2006, 110,
14231-14235.
(59) Lytton, S. D.; Mester, B.; Libman, J.; Shanzer, A.; Cabantchik, Z. I. Anal.
Biochem. 1992, 205, 326-333.
(51) Schneier, B. Secrets and Lies: Digital Security in a Networked World;
John Wiley& Sons, Inc.: New York, 2000.
(60) Yakirevitch, P.; Rachel, N.; Albrecht-Gary, A.-M.; Libman, J.; Shanzer,
A. Inorg. Chem. 1993, 32, 1779-1787.
(52) Kishimura, A.; Yamashita, T.; Yamaguchi, K.; Aida, T. Nat. Mater. 2005,
4, 546-549.
(61) Jurkevitch, E.; Hadar, Y.; Chen, Y.; Yakirevitch, P.; Libman, J.; Shanzer,
A. Microbiology 1994, 140, 1697-703.
(53) Braeckmans, K.; Smedt, S. C. d.; Roelant, C.; Leblans, M.; Pauwels, R.;
Demeester, J. Nat. Mater. 2003, 2, 169-173.
(62) Weizman, H.; Ardon, O.; Mester, B.; Libman, J.; Dwir, O.; Hadar, Y.;
Chen, Y.; Shanzer, A. J. Am. Chem. Soc. 1996, 118, 12368-12375.
(63) Zanninelli, G.; Glickstein, H.; Breuer, W.; Milgram, P.; Brissot, P.; Hider,
R. C.; Konijn, A. M.; Libman, J.; Shanzer, A.; Cabantchik, Z. I. Mol.
Pharmacol. 1997, 51, 842-852.
(54) Clelland, C. T.; Risca, V.; Bancroft, C. Nature 1999, 399, 533-534.
(55) Mao, C.; LaBean, T. H.; Reif, J. H.; Seeman, N. C. Nature 2000, 407,
493-496.
(56) Tanaka, K.; Okamotoa, A.; Saito, I. BioSystems 2005, 81, 25-29.
9
348 J. AM. CHEM. SOC. VOL. 129, NO. 2, 2007

Molecular Keypad Lock
A R T I C L E S
natural microbial iron chelators secreted by bacteria to bind
extra-cellular iron and mobilize it into their cells. Harnessing
this ingenious bacterial machinery to control over iron-ligand
binding state, we were able to demonstrate a range of devices,
including molecular carriers,63-71 a molecular switch based on
iron translocation,72 and more recently a reconfigurable molec-
ular logic gate endowed with calculating skills.39 The latter
device has been further developed to a complete Boolean
network of seven logic gates. Remarkably, this array has shown
a preferential response to particular sequences of input signals,
which can encode photoionic passwords.
A molecule assembled as Fluorescein-Linker-Pyrene (FLIP)
is a reconfigurable logic gate that can operate as a molecular
keypad lock (Figure 2a). The binding site of FLIP resembles
the natural coprogen siderophore having three hydroxamic acids
capable of selectively binding iron in great efficiency and a
proline group, which facilitates its folding around the metal ion.
Two fluorophores, pyrene (donor) and fluorescein (acceptor),
are attached at both terminus of the receptor chain, enabling
the molecule to fluoresce at two distinct regions, blue and green,
respectively. To force the fluorophores to approach in close
proximity to the ferric ion, two of the hydroxamate binding sites
were directly attached to the fluorophores in a way that dictates
an oriented position upon binding. Empirical force field (EFF)
calculations70 have shown that in the presence of iron(III), the
hydroxamate binding groups wrap around the metal ion in a
roughly octahedral arrangement that constrains the molecular
backbone into a highly ordered helical twist. In the most stable
conformation, the two fluorescent groups are situated parallel
to each other, with their centroids 7 Å apart and with a projection
angle close to 30° (Figure 2b).
Although a reversible binding of iron can induce merely two
states on the molecule, i.e., bound or nonbound (Figure 2a,b),
the effect of acid and base input signals on FLIP is somewhat
more complex, as fluorescein can equilibrate between several
ionization states, consisting of no less than seven protolytic
forms.73-80 As each of the states possesses a characteristic
absorption spectrum, it is possible to follow the changes that
occur upon protonation.
Figure 2c depicts a few examples. In ethanol, two peaks are
observed with maximums at 344 and 500 nm, corresponding
to the pyrene and fluorescein units, respectively. Under these
Figure 2. Reconfigurable molecular logic gate capable of performing as a
molecular keypad lock. (a) Processing molecule is composed as Fluorescein-
Linker-Pyrene (FLIP). The functional linker mimics the structure of a
natural bacterial iron(III) carrier (siderophore), whereas the two fluorophores,
pyrene and protonable fluorescein, function as blue and green emission
channels, respectively. (b) Calculated structure of the iron complex
performed by EFF force field program.70 (c) Changes that occur on the
absorption spectra of FLIP in response to acid and base input signals.
Changes are observed at 500 nm, where the protonable fluorescein absorbs,
whereas the spectrum of the chemically inert pyrene unit (344 nm) remains
unchanged. In ethanol solution, the monoanion fluorescein F(-1) is
prevalent. Addition of sodium acetate results in a characteristic peak of the
highly fluorescent dianionic form F(-2), whereas insertion of acetic acid
provides the colorless non emissive lactonic species, F(0). Conditions: 8.7
µM FLIP in ethanol, inputs: 2% aqueous solutions of AcOH or AcO-Na+,
0.01N.
(64) Ardon, O.; Nudelman, R.; Caris, C.; Libman, J.; Shanzer, A.; Chen, Y.;
Hadar, Y. J. Bacteriol. 1998, 180, 2021-2026.
(65) Nudelman, R.; Ardon, O.; Hadar, Y.; Chen, Y.; Libman, J.; Shanzer, A. J.
Med. Chem. 1998, 41, 1671-1678.
(66) Shanzer, A.; Libman, J. Met. Ions Biol. Syst. 1998, 35, 329-354.
(67) Palanche´, T.; Marmolle, F.; Abdallah, M. A.; Shanzer, A.; Albrecht-Gary,
A.-M. JBIC 1999, 4, 188-198.
(68) Weizman, H.; Shanzer, A. Chem. Commun. 2000, 2013-2014.
(69) Meijler, M. M.; Arad-Yellin, R.; Cabantchik, I.; Shanzer, A. J. Am. Chem.
Soc. 2002, 124, 12666-12667.
(70) Felder, C.; Shanzer, A. Biopolymers 2003, 68, 407-421.
(71) Kornreich-Leshem, H.; Ziv, C.; Gumienna-Kontecka, E.; Arad-Yellin, R.;
Chen, Y.; Elhabiri, M.; Albrecht-Gary, A.-M.; Hadar, Y.; Shanzer, A. J.
Am. Chem. Soc. 2005, 127, 1137-1145.
(72) Zelikovich, L.; Libman, J.; Shanzer, A. Nature 1995, 374, 790-792.
(73) Diehl, H.; Markuszewski, R. Talanta 1989, 36, 416-418.
(74) Klonis, N.; Sawyer, W. H. J. Fluoresc. 1996, 6, 147-157.
(75) Klonis, N.; Sawyer, W. H. Photochem. Photobiol. 2000, 72, 179-185.
(76) Kubista, M.; Sjoback, R.; Nygren, J. Anal. Chim. Acta 1995, 302, 121-
125.
conditions, fluorescein seems to exist mainly in its monoanion
form F(-1). Although, a protolytic equilibrium between the
mono- and dianion species might provide a more accurate
description for FLIP in ethanol, for simplicity we will refer to
this state as monoanion F(-1) predominant form. This inter-
pretation is in good agreement with previous studies, which
regard the monoanion as the predominant species in ethanol.78
(77) Martin, M. M.; Lindqvist, L. J. Lumin. 1975, 10, 381-390.
(78) Martin, E.; Pardo, A.; Guijarro, M. S.; Fernandez-Alonso, J. I. J. Mol. Struct.
1986, 142, 197-200.
(79) Sjoback, R.; Nygren, J.; Kubista, M. Spectrochim. Acta A 1995, 51, L7-
L21.
(80) Zanker, V.; Peter, W. Chem. Ber. 1958, 91, 572-580.
9
J. AM. CHEM. SOC. VOL. 129, NO. 2, 2007 349

A R T I C L E S
Margulies et al.
Scheme 1. Fluorescence Outputs of FLIP Can Interchange in
Such a Way That a Boolean Network Is Establisheda
Figure 3. (a) Fluorescence spectra for states A-E excited at 344 nm,
corresponding to Scheme 1: (A) 1 mL FLIP 0.5 µM in ethanol, (B) A +
a A green fluorescence emission is achieved in ethanol when the ferric
ion is not bound and the fluorescein (Flu) is in its monoanion F(-1)
fluorescent forms (A). Addition of iron(III) to the ethanol solution of FLIP
results in a fluorescence quenching by the metal ion (B) and almost a total
elimination of fluorescein emission. When a base is added to B, the
monoanion fluorescein F(-1) deprotonates to its more fluorescent dianion
F(-2) form, resulting in a faint green luminescence from the iron complex
(C). Extraction of iron from state C or, alternatively, addition of a base to
state A generates state D, in which the metal quencher is not bound and
fluorescein is in its highly emissive dianion F(-2) state (D). A blue
fluorescence (E) is generated in acidic media when the fluorescein is not
emitting (Off) and iron(III) is not bound. Consequently, the FRET process
does not take place and the green emission of the fluorescein unit is replaced
with a blue pyrene luminescence.
1 eq FeCl3, (C) B + 10 µL AcO-Na+(aq) 2 M, (D) A + 10 µL AcO-Na+
(aq)
2 M, (E) A + 10 µL HCl(aq) 1 M or 20 µL AcOH(aq) 1 M. (b) Repeated
scans under continuous excitation at 344 nm revealed light induced
fluorescence enhancement of the blue channel in state E. Whereas in ethanol
(A) a minor enhancement of pyrene emission intensity was observed,81 in
acidic environment (E) this transformation is far more significant. Each
scan corresponds to 18 s of irradiation of E in AcOH.
Upon addition of acetic acid to FLIP in ethanol, a total
elimination of fluorescein absorption is observed, indicating that
the neutral fluorescein F(0) is obtained in its colorless, non
conjugated, lactonic form. The fact that the acidification of the
solution takes place in an organic solvent rather than in a polar
aqueous solution is the reason for a preferential formation of
the lactonic species over the quinoid or zwitterion neutral
forms.75 Alternatively, when a base is added to FLIP in ethanol,
an enhancement of the absorption spectrum of fluorescein at
500 nm is observed, and a typical absorption peak of the dianion
F(-2) is formed. In opposed to pH-dependent transformations
of fluorescein absorption, the absorbance of the pyrene aromatic
unit (344 nm), having no protonable functional groups, remained
unchanged. The relevant emissions bands, corresponding to each
of the states in Figure 2c, are presented in Figure 3.
Five Fluorescent States. Through a reversible binding of
iron and/or protonation of the fluorescein unit, several chemical
states can be obtained from FLIP. Scheme 1 describes the
conditions by which the fluorescence outputs of FLIP can be
controlled, in a way that a Boolean network is established. A
key principle for understanding this scheme is the fact that
mainly the dianion F(-2) and monoanion F(-1) forms con-
tribute to the fluorescein emission at 525 nm. In addition, the
first is known to possess higher quantum yield.73 The corre-
sponding fluorescence spectra for states A-E are shown in
Figure 3a.
fluorescent form (Figure 2c). In this case, fluorescence resonance
energy transfer (FRET) takes place between the donor (pyrene)
and acceptor (fluorescein) leading a predominant emission from
the fluorescein unit. Addition of iron(III) to the ethanol solution
of FLIP results in a fluorescence quenching by the metal ion
(B) and almost a total elimination of fluorescein emission is
obtained from FLIP-Fe(III), as expected from a fluorescent
siderophore analogue upon iron binding.63-71
State C describes the change that occurs when a base is added
to the ferric complex in ethanol. As depicted in Figure 2c, under
such conditions the monoanion fluorescein unit F(-1) depro-
tonates to its more fluorescent dianion F(-2) form,73 resulting
in an increased emission from the complex, in such a way that
a faint green luminescence is observed even from the bound
fluorescein. Extraction of iron(III) from state C or, alternatively,
addition of a base to state A generates state D, in which the
metal quencher is not bound and fluorescein is in its highly
emissive state F(-2). This results in the strongest emission at
525 nm (D). A blue fluorescence (E) is generated in acidic media
when the fluorescein is not emitting (Off) and iron(III) is not
bound. Acidification of the solution results in protonation of
the fluorescein unit to its transparent, non-emissive lactonic form
(Figure 2c). Consequently, the FRET process does not take
place, and the green emission of the fluorescein unit is replaced
with a blue pyrene luminescence. The same transformation
occurs in higher acidity (e.g., HCl as an input) when a
fluorescein cation F(+1) starts to appear;34 therefore, for the
Five different fluorescence states can be obtained from FLIP,
namely, a green luminescence at 525 nm (A), no emission (B),
faint green emission (C), intense green fluorescence (D), and
blue emission at 390 nm (E). State A describes a green
fluorescence emission achieved in ethanol when the ferric ion
is not bound and the fluorescein is in its monoanion F(-1)
9
350 J. AM. CHEM. SOC. VOL. 129, NO. 2, 2007

Molecular Keypad Lock
A R T I C L E S
Figure 4. (a) Molecular NOT logic gate. (b) Simultaneous YES/NOT molecular logic gates. Conditions: as in Figure 3.
general case of acidic environment (E), the green channel is
described as “Off” luminescence.
ing of fluorescence induced by unprotonated hydroxamates or
by the chemical input themselves, iron interfering in the FRET
process, as well as change in the polarity of the solvent85 might
also play a role. That is to say that the chemical inputs might
have some additional effects on FLIP. For example, it should
be expected that the emission intensities of state D, obtained
by adding a base to A (route AfD), would not be identical to
those obtained by extraction of iron from C (route CfD), due
to the different chemical reagents and starting conditions
involved. Nevertheless, in this experimental setup, these dif-
ferences are minor; hence, the overall picture depicted in Scheme
1 is perfectly suitable for the design of various Boolean functions
at the molecular scale.
Interestingly, the blue emission intensity in state E was
enhanced with increasing irradiation time (Figure 3b). This
indicates that in addition to FRET exclusion, a rapid photo-
chemical process further contributes to the blue-channel en-
hancement in acidic environment. In an in-depth article, Sun et
al. have suggested that unstable pyrene photoproducts might
be responsible for the enhanced emission.81 Figure 3b shows
the changes in the blue-channel intensity obtained under repeated
scans in ethanol (A) and in acidic environment (E). Both
processes lead to an enhancement in the blue channel, which
clearly supports the observation of Sun et al.;81 however, in
acidic environment, this transformation is far more significant.
The tendency of pyrene to form a radical in an acidic
environment82 and the capacity of a siderophore cavity to
scavenge it83,84 might be the reason that this phenomenon is so
profoundly observed in state E. As far as Boolean logic is
concerned, this photoreaction leads to an improved sensitivity
of the blue channel, which does not change the overall outcome
of the processor. Irreversibility might be a concern, but this is
already an inherent problem of many chemical-based logic
systems, which has been addressed elsewhere.43
Quenching of pyrene85 or fluorescein65 in siderophore ana-
logues or other hydroxamate-based ligands has been addressed
before and is considered to occur by energy transfer85 or electron
transfer65 processes. Heavy-atom-assisted intersystem crossing
is usually ruled out because the emission is generally not
quenched by gallium, being heavier than iron. For FLIP, the
situation is somewhat more complex, as not only two distinct
fluorophores are involved in the emission process, but also
FRET comes as an additional parameter whose influence on
the quenching mechanism deserves further investigation. In this
respect, it is noteworthy that Scheme 1 depicts a simplified
model corresponding to the processes we consider most crucial
for the generation of the different optical states. Other processes,5
such as spatial separation between donor and acceptor, quench-
Elementary Logic Operations. Fundamental logic gates are
the basic building blocks for the construction of superior logic
devices. This is true for automaton33 and moleculators34-50 as
well as for a keypad lock (Figure 1). Switching between states
A-E (Scheme 1), a Boolean network of various fundamental
logic gates can be established (Figures 4 and 5). The underline
principle we use is based on chemical input multiplicity,39 that
is, a combinatorial recognition approach that enables FLIP to
process a variety of the chemical inputs with a minimum amount
of receptors. Although molecular logic gates can be comprised
of several receptors, each recognizes a specific chemical input,3,6
in our design: (i) a single chemical input can target a specific
receptor, (ii) distinct inputs can be recognized by the same
receptor, and finally (iii) a single chemical species can target
several receptors simultaneously. Similar principles enable the
olfactory system to recognize about 10 000 different volatile
chemicals with only 1000 types of receptors.86
Targeting the siderophoric receptor with iron(III) as input,
route AfB results in green emission quenching and a NOT
logic gates in the green channel (Figure 4a). Alternatively, using
an acidic input that target the fluorescein unit, route AfE
provides, parallel YES/NOT outputs in the blue and green
channels, respectively (Figure 4b). Combination of the inputs
above, route BrAfE corresponds to a NOR gate at 525 nm
(Figure 5a). In this case, each of the inputs targets a different
receptor, in a way that either iron (III), acid, or both eliminate
of the emission at 525 nm.
(81) Sun, Y.-P.; Ma, B.; Lawson, G. E.; Bunker, C. E.; Rollins, H. W. Anal.
Chim. Acta 1996, 319, 379-386.
(82) Kubat, P.; Civi, S.; Muck, A.; Jim, B.; Zima, J. J. Photochem. Photobiol.
A 2000, 132, 33-36.
(83) Denicola, A.; Souza, J. M.; Gatfi, R. M.; Augusto, O.; Radi, R. Free Radical
Biol. Med. 1995, 19, 11-19.
It is also possible to realize logic function with different inputs
that target the same receptor. Starting from the metal complex
(84) Yavin, E.; Weiner, L.; Arad-Yellin, R.; Shanzer, A. J. Phys. Chem. A 2001,
105, 8018-8024.
(85) Bodenant, B.; Fages, F.; Delville, M.-H. J. Am. Chem. Soc. 1998, 120,
7511-7519.
(86) Malnic, B.; Hirono, J.; Sato, T.; Buck, L. B. Cell 1999, 96, 713-723.
9
J. AM. CHEM. SOC. VOL. 129, NO. 2, 2007 351

A R T I C L E S
Margulies et al.
Figure 5. (a) Molecular NOR logic gate. Conditions: as in Figure 3. (b) Molecular OR logic gate. Inputs: 20 µL saturated EDTA(aq), 20 µL DFO(aq) 5 mM.
in basic environment (C), the ferric ion can be selectively
removed by two different chelators, i.e., by ethylenediamine
tetraacetic acid (EDTA) or by the natural desferrioxamine (DFO)
siderophore (CfD). A basic solution (C) is required to maintain
the ionization state of fluorescein upon addition of EDTA, which
is also an acid. Because the ferric ion is removed by each of
the chelators or both, an OR logic gate at 525 nm is obtained
(Figure 5b).
state, both will ultimately result in a high output at 525 nm.
However, as the extraction of a ferric ion from a siderophore
by EDTA is inhibited in basic environment,87 a vast difference
in the reaction rate between the two paths is observed.
When base is added before EDTA (3 min interval time), only
a faint green emission is detected from the bound fluorescein
unit. Monitoring the emission intensity over time reveals that
only after several hours a full recovery of fluorescence is
achieved (Figure 6a). On the other hand, when EDTA is first
added to the solution, followed by insertion of base, an intense
emission at 525 nm from the nonbound fluorescein is im-
mediately observed (Figure 6b). These observations are con-
sistent with previous studies, revealing a mechanistically
interpreted kinetic scheme that involves two parallel pathways
for siderophores-EDTA iron exchange.87 One involves a direct
attack of the EDTA on the ferric ion, whereas in the other the
hydroxamate binding groups are first protonated87,88 followed
by a rapid attack of the competing ligand. Therefore, when
EDTA is first added, the solution becomes acidic, leading to
partial protonation of the hydroxamate residues in FLIP-Fe-
(III) and a rapid iron extraction. On the contrary, when base is
initially added, the ferric ion becomes tightly bound, which
causes an inhibition of the metal exchange process. Conse-
quently, by limiting the gate to operate solely for a few minutes,
this gate becomes a 2-input priority-AND molecular logic gates
that process EDTA (E) and base (B) as inputs and generates a
high output only in response to sequence EB.
When the ferric complex is present in ethanol (B), an acidic
chemical input can simultaneously affect the fluorescein unit and
the iron-binding domain in a way that XOR/INHIBIT (CrBfE)
and AND/INHIBIT (BfEfD) operations are obtained.39 The
latter routes (DrCrBfEfD), previously described as part
of a molecular arithmetic system,39 not only complete the
operation of FLIP to seven fundamental logic operations of YES,
NOT, NOR, OR, AND, XOR, and INHIBIT but also provide
the basis for realizing a keypad lock at the molecular scale.
A Molecular Keypad Lock. To acquire a molecular keypad
lock, the next step must be the development of a mechanism
by which a molecule could distinguish between different input
sequences. Electronic locks are built by connecting several
elementary logic gates (Figure 1); however, it has already been
demonstrated that molecular logic systems should not necessarily
follow circuitry, but can rather make use of a variety of
approaches for molecular logic reconfiguration.34 For route
BfD (Scheme 1), this could be achieved by controlling the
kinetics of iron extraction, in a way that reconfigures a simple
AND gate into a 3-input priority-AND molecular logic device.
Starting from the iron complex in ethanol (state B in Scheme
1) and using an acidic iron chelator (EDTA) and a base
(AcO-Na+) as inputs, an AND39 logic gate is obtained at 525
nm. EDTA alone extracts the iron, but also protonates the
fluorescein to in its inactive form (state E). Insertion of acetate
to B provides a basic environment, yet, the iron is still bound
(state C). Only when both inputs are present, the solution is
basic and the iron is removed, which leads to an intense
fluorescence at 525 nm (D).
To realize a molecular keypad lock, it is essential to increase
the number of inputs, as two bits (E and B) provide only limited
numbers of password combinations, namely, EB or BE. Steps
toward molecular AND gates endowed with multiple bits
processing have recently been reported.6,16 For FLIP this has
been achieved simply by introducing the excitation beam (344
nm) as an additional input signal.10,36 In opposed to chemical
inputs, a light input signal has the advantage that it does not
accumulate in solution, hence its operation is restricted solely
to the period of activation.
Now it is possible to operate this AND gate through two
possible pathways (Figure 6). In the first route (BfCfD), a
base is first added, followed by addition of the EDTA, whereas
in the second route (BfEfD), the order of insertion is
switched. Because both pathways lead to an identical chemical
Figure 7a describes a molecular keypad lock that “opens”
by inserting the right sequences of three keys: EDTA (E), base
(87) Santos, M. A.; Bento, C.; Esteves, M. A.; Farinha, J. P. S.; Martinho, J.
M. G. lnorg. Chim. Acta 1997, 258, 39-46.
(88) Monzyk, B.; Crumbliss, A. L. J. Am. Chem. Soc. 1982, 104, 4921-4929.
9
352 J. AM. CHEM. SOC. VOL. 129, NO. 2, 2007

Molecular Keypad Lock
A R T I C L E S
Figure 6. Two-input priority-AND molecular logic gate that processes EDTA (E) and base (B) as inputs and generates a high output only in response to
sequence EB. The exchange reaction between the siderophore analogue and EDTA is inhibited in basic environment,87 therefore a vast difference in the
reaction rate between the two paths is observed. (a) When base is added before EDTA (route BfCfD), only a faint green emission is observed due to
ionization of the bound fluorescein unit. Monitoring the emission intensity over time revealed that only after several hours a full recovery of fluorescence
is achieved. (b) When EDTA is first added to the solution (route BfEfD), followed by insertion of base, an intense emission at 525 nm from the nonbound
fluorescein is immediately observed. Conditions: 1 mL of 0.5 µM of FLIP-Fe(III) in ethanol. Inputs: 10 µL of aqueous NaOAC (2 M), 20 µL of saturated
EDTA.
Figure 7. Molecular keypad lock. (a) Keys B, E, and U hold the relevant inputs (Base, EDTA, and UV, respectively) applicable for generating a strong
green emission at 525 nm, only when a correct password (EBU) is entered. In addition, only sequence EU will generate a high output at 390 nm. The other
keys can hold a high concentration of fluorescence quenchers, so if pressed the password entry will fail. (b) Outputs in the blue and green fluorescence
channels, corresponding to the six possible input sequences, inserted with an interval time of 3 min.
(B), and UV light (U). Only keys E, B, and U hold the correct
inputs that can initiate a strong fluorescence at 525 nm, whereas
keys E and U are enough for generating high emission at 390
nm. The other keys can hold a high concentration of fluores-
cence quenchers, so if pressed, the password entry will fail. What
makes this device possible is the fact that its operation time is
restricted to 9 min, during which, each input operates for 3 min
before insertion of the next. In this way, there is a sufficient
time for EDTA to extract the iron in acidic environment, but
not enough for removing the iron under basic conditions. At
any point during this period, if a high output is detected in the
blue or green channel, it authorizes a valid password.
fluorescence, as only then is the iron quencher rapidly removed
and the fluorescein in its dianion F(-2) fluorescent state; hence,
the excitation beam will find an emitting species at 525 nm
(state D). In addition, only sequence EU can generate a high
output at 390 nm (state E). Consequently, this molecular device
is now applicable for authorizing two different users, in a manner
similar to a simple ATM banking machine (Figure 7a). One
holds the exact combination keys for activating the green
channel, whereas the other can activate the blue channel using
a different password entry.
Conclusions
This paper describes a new concept in the way information
can be protected at the molecular scale. By harnessing the
principles of molecular Boolean logic,1 we have designed a
reconfigurable molecular lock that has the power to authorize
two different code entries, EBU that activates the green channel,
and EU that triggers the blue channel.
The information in each channel is a single-bit light output
signal that can be used to authorize a user, to verify authentica-
tion of a product, or to initiate a higher process. Additional
security comes from the fact that this device operates in the
Figure 7b shows the outputs in the blue and green fluores-
cence channels, obtained for the six possible input combinations,
i.e., UEB, UBE, EUB, EBU, BUE, and BEU. When light input
signal is applied first (UEB or UBE), then the fluorescence
output is generated by the iron complex in ethanol (state B in
Scheme 1) having low output in both green and blue emission
channels. In codes BUE and BEU, the complex in basic solution
is excited (state C), which results in a faint green luminescence
output below the threshold line. Only sequence EBU provides
the right password entry for the generation of an intense green
9
J. AM. CHEM. SOC. VOL. 129, NO. 2, 2007 353

A R T I C L E S
Margulies et al.
fluorescence mode; hence, it has the potential be concealed down
to the level of a single molecule.89 In addition, the accumulation
of chemical inputs that prevent this lock, as well as many other
molecular logic systems from resetting,43 might be advantageous
in some cases, as it limits the code entry to be inserted only
once.
To conclude, we hope to have drawn attention to the potential
use of molecular logic gates for security purposes, by demon-
strating a new class of decision-making device at the molecular
scale. Although the molecular lock presented here can so far
process a limited number of password entries comprised merely
of two and three input bits, it inspires the imagination toward
the next generation of confidential identification technologies
based on molecules. In fact, common security inks already
function as simple YES logic gates that produce a fluorescence
output, only in response to black light (UV) as an input. Future
programmable molecular locks might have the capacity to
process multiple password entries that provide theoretically
unbreakable protection against forgery. Although this goal would
probably require more powerful processors that might run on
solid supports, it can be reasonably envisioned that, similar to
the rapid evolution of molecular logic gates, improved security
logic systems, based on molecules, would also be developed.
To the best of our knowledge, this is the first example of a
molecular system capable of processing password entries; hence,
it can add a new dimension of protection to other defense tech-
nologies, such as cryptography and steganography, previously
achieved with molecules. A few years ago, Bancroft et al. have
shown the way a message could be encrypted on a DNA strand
(cryptography) and be hidden inside a tiny microdot (stegan-
ography).54 The concept proposed in this paper refers to the
possibility of preventing access to such information, in a manner
similar to a password entry requested for accessing electronic
computers.51 In this way, even when both the location of the
information and the decryption keys are known, it is kept secure.
There is an additional merit to this work. The keypad lock
presented here might just be a particular example for the general
case of molecular systems, capable of distinguishing between
different times of events. It is generally believed that the power
of Boolean computing would find application in molecular
agents capable of diagnosing biological disorders and produce
the right signaling or even therapeutic outputs.1,90 Considering
that, the level of chemical agents in the blood stream is
constantly changing, in a way that a disease can be pronounced
in a particular order of different biochemical signals, bio-
molecular devices capable of distinguishing between different
chemical sequences, might have considerable advantages over
simple molecular logic gates.
Acknowledgment. We thank Mrs. Rachel Lazar, for her
skillful assistance and Mr. Vladimir Sergeyev for designing the
electronic lock circuit. E-mail: Vladimir.S@scent-tech.com.
Support from the G. M. J. Schmidt Minerva Center on
Supramolecular Architecture, the Helen and Martin Kimmel
Center for Molecular Design, and Mordechai Glikson Fund are
greatly acknowledged. A.S. holds the Siegfried and Irma
Ulmann professorial chair.
Supporting Information Available: Synthesis of FLIP,
chemical structures of EDTA and DFO, and fluorescence spectra
of the molecular keypad lock. This material is available free of
charge via the Internet at http://pubs.acs.org.
(89) Weiss, S. Science 1999, 283, 1676-1683.
(90) Margolin, A. A.; Stojanovic, M. N. Nat. Biotechnol. 2005, 23, 1374-1376.
JA065317Z
9
354 J. AM. CHEM. SOC. VOL. 129, NO. 2, 2007