Signal Transduction Cascades
J. Am. Chem. Soc., Vol. 120, No. 49, 1998 12735
Scheme 1. Schematic Depiction of the Signaling Systema
Spectral analysis revealed that four HABA molecules were
bound covalently to each molecule of avidinsone per subunit.
The HABA moiety appeared to be coupled only to the biotin-
binding site of avidin, as indicated by the red color. To verify
this contention, biotin was added to the red avidin and an
immediate shift to 356 nm was observed (see spectra in Scheme
1), indicating that biotin, due to its higher affinity for avidin,
was capable of displacing the HABA moiety from the binding
site. Upon dialysis or gel filtration on Sephadex G-25, the
orange color continued to be associated with the protein. To
further prove that HABA is bound covalently to avidin,
deglycosylated avidin3 was modified in the same way and
analyzed by mass spectroscopy. MALDI analysis of the
unlabeled protein showed a m/z value of 14 287-14 290,
corresponding to that for a single subunit of deglycosylated
avidin. The reaction product showed a single peak (m/z
14 727-14 734); the difference in mass (about 440 units) is
consistent with the molecular weight of the HABA derivative
(Figure 3). The HABA must be bound covalently since biotin-
blocked avidin showed a m/z of 14 290 similar to that of the
native avidin.
a Affinity-labeled red avidin (red symbols), λmax ) 504 nm (see
spectrum), does not interact with the anti-HABA antibody (gray
Y-shaped symbols). The cascade is triggered upon addition of biotin
(blue B) or biotin-containing molecules, which expel the covalently
attached HABA moiety (yellow H) from the binding site. The spectrum
shows a shift to orange, λmax ) 356 nm. The HABA group is now
available for subsequent interaction with anti-HABA antibody or avidin.
Reaction with avidin restores the red color.
To determine which residue was modified, the protein was
subjected to trypsin digestion. The hydrolysate was separated
by HPLC (Figure 4), and the sequence of the orange peptide
was analyzed. The analysis showed that Lys-111 was modified
(Table 1). Interestingly, a neighboring residue, Trp-110, is an
important part of the binding pocket,8 which again indicates
that the HABA moiety is located in the binding site (Figure 5).
To show that the HABA moiety is indeed buried in the
binding site and not available for further interaction before its
displacement by biotin, we used a second HABA-binding
molecule for its detection. For this purpose, polyclonal antibod-
ies against HABA were elicited in rabbits. After affinity
purification, the antibodies were examined for their interaction
with the HABAylated avidin. As seen in Figure 6, in the
absence of biotin, the anti-HABA antibody fails to recognize
the HABA, buried in the binding site. Upon addition of biotin,
however, the HABA moiety was expelled and strong binding
of the antibody was detected.
To demonstrate that this system meets the requirements of a
signal transduction cascade and the assemblage of protein
multilayers,9 biotin-saturated HABAylated avidin was incubated
with biotinylated anti-HABA antibodies, followed by additional
cycles of HABAylated avidin and biotinylated antibodies. A
stepwise increase in absorbance could be detected after the
formation of each layer (Figure 7). The cascade could be
initiated using HABAylated avidin and biotin or biotinylated
macromolecules. In either case, the addition of biotin was
crucial to expel the HABA from the binding site, thus enabling
subsequent interaction with the anti-HABA antibodies.
In summary, we have demonstrated, by chemical means the
interplay of molecular recognition and oriented protein assembly
which serves as a prerequisite for the principle of signal
transduction. Using the avidin-biotin system, biotin served as
the effector or trigger of the cascade. Since avidin has four
binding sites, if limited amounts of biotin are added to displace
only one or two HABA molecules, the cascade can be triggered
vectorially in different dimensions. Therefore, this system can
be considered a chemical mimic of signal transduction, which,
group (e.g., an acetyl moiety), to accommodate the azo dye in
its tautomeric form at the binding site.6
For the present work, an HABA-containing affinity label was
designed, such that the dye would remain covalently attached
to the binding site of avidin. The ortho position of the HABA
phenol was thus modified with a reactive functional group,
which, due to steric constraints, forms an intramolecular cyclic
carbamate (Figure 1). This cyclic HABA can be hydrolyzed
by avidin, which enabled us to exploit the principle of forced
catalytic hydrolysis6 to attach the HABA moiety to an appropri-
ate residue in or near the binding site of avidin. The cyclization
of the reagent was necessary, since the activated linear N-
hydroxysuccinimidocarbamate (B) would react with any amino
group, whereas the cyclic compound (C) reacts in situ, i.e.,
selectively, after occupying the binding site of avidin.
Upon addition of the HABA reagent to avidin, an immediate
shift in the spectrum to 504 nm was observed only with the
linear, nonactivated HABA analogue.7 A similar change was
also observed with the cyclic HABA derivative, but in this case,
the shift developed gradually. The latter reaction was allowed
to proceed until no further increase at λmax was observed. No
shift was observed if the cyclic reagent was added together with
biotin (Figure 2), indicating that the vitamin occupied the
binding site and prevented the reaction with HABA. To verify
this assumption, the reaction mixtures were passed through a
Sephadex G-25 column or dialyzed. In mixtures that contained
either the linear precursor or the cyclic reagent in the presence
of biotin, it was possible to dissociate the orange-colored HABA
from the colorless avidin. On the other hand, after reaction
with avidin, the cyclic HABA reagent could no longer be
separated from the protein; the red color remained associated
with the protein, indicating covalent attachment to avidin.
(8) Trp-110 is a part of the binding site but comes from the neighboring
subunit: Livnah, O.; Bayer, E. A.; Wilchek, M.; Sussman, J. L. Proc. Natl.
Acad. Sci. U.S.A. 1993, 90, 5076-5080.
(9) Mu¨ller, W.; Ringsdorf, H.; Rump, E.; Wildburg, G.; Zhang, X.;
Angermaier, A.; Knoll, W.; Liley, M.; Spinke, J. Science 1993, 262, 1706-
1708.
(6) Vetter, S.; Bayer, E. A.; Wilchek, M. J. Am. Chem. Soc. 1994, 116,
6, 9369-9370.
(7) The same effect was observed for deglycosylated avidin, Neutralite
avidin, and streptavidin.