Commentary

Making sense of polymer-based biosensors

Peter S. Heeger†‡ and Alan J. Heeger§
†Department of Medicine, Case Western Reserve University, Cleveland OH 44106; and §Institute for Polymers and Organic Solids, University of California,

Santa Barbara, CA 93106

Ksv of 107 (1). By contrast, quenching of

I n this issue of PNAS, Chen et al. (1)
show feasibility for producing a highly
sensitive biosensor, capable of detecting
By functionalizing the conjugated back-
bone with suitable side-chains, these mac-
romolecules can be made soluble in com-
the fluorescent molecule, stilbene, (the
equivalent of 1.5 repeat units of PPV) has
nanomolar quantities of biologically inter- mon organic solvents and in water. Typical a Ksv of 15. Moreover, the largest ampli-
esting molecules in ‘‘real time.’’ If the molecular weights are on the order of 106 fication of fluorescence quenching re-
phenomenon can be generalized to other Da, corresponding to ⬇1,000 monomer ported earlier in the literature corre-
systems, it could have significant implica- repeat units per macromolecule. sponds to Ksv ⫽ 65 (12).
tions for a wide variety of fields, including The fundamental scientific discovery The huge enhancement of fluorescence
toxicology and medical diagnostics. addressed in the paper by Chen et al. is quenching observed for the MV2⫹兾2,5-
The last two decades have witnessed the that the luminescence of a semiconducting methoxy propyloxysulfonate (MPS)–PPV
emergence of conjugated polymers as an polymer in aqueous solution can be system arises from a combination of two
intriguing class of organic macromole- quenched by using extremely low concen- effects. First, in aqueous solution, there is
cules that have the electrical and optical trations of a cationic electron acceptor an equilibrium:
properties of metals and semiconductors (1). The mechanism for the quenching is
and, in addition, have the processing ad- ultrafast photoinduced electron transfer ⫺SO3⫺Li⫹ N ⫺SO3⫺ ⫹ Li⫹,
vantages and mechanical properties of (10); the excited electron transfers to a
such that the luminescent polymer is neg-
polymers (2). The study of conjugated nearby acceptor within a few hundred
atively charged (anionic). As a result, the
(semiconducting) polymers has resulted in femtoseconds (11), more than four orders
positively charged acceptor (MV2⫹) and
fundamental insights into the understand- of magnitude faster than the luminescence

COMMENTARY
the anionic polymer form a weak complex,
ing of the chemistry and physics of this decay time. Because the electron and the
thereby significantly enhancing the local
novel class of materials (3, 4), and it has hole are separated after the electron
concentration. Second, from the concen-
stimulated the development of a number transfer step (the electron is on the mo-
trations used in the experiments, it ap-
of applications (5, 6). Included among lecular acceptor and the hole is left behind
pears that a single MV2⫹ molecule can
these are all-polymer integrated elec- on the polymer chain), the luminescence is
quench the fluorescence derived from an
tronic circuits (7), photodetectors and so- quenched. The photoinduced electron
entire ⬇1,000 repeat unit polymer chain.
lar cells (8), and flat-panel emissive dis- transfer step is exponentially sensitive to
The mechanism of this remarkable ex-
plays fabricated from polymer light emit- the distance separating the electron on the
tended quenching is not yet understood
ting diodes (9). Chen et al. (1) have now polymer chain from the acceptor. If the
and will need to be more carefully ad-
demonstrated that the luminescent prop- acceptor is removed from the vicinity of
dressed. Nonetheless, the combination of
erties of such polymers can be further the polymer chain by ⬇1 nm, the electron this effect with the enhanced local con-
exploited to develop a new form of bio- transfer rate will be so slow that the radi- centration resulting from the ionic charge
sensor. ative recombination channel (lumines- interaction leads to the million-fold am-
Conjugated polymers such as poly(phe- cence) will again dominate. plification factor.
nylene vinylene) (PPV) and its soluble The Stern-Volmer constant, Ksv, pro- Because MV2⫹ quenches quantitatively
derivatives are known to exhibit photolu- vides a quantitative measure of the lumi- at extremely low concentrations, the re-
minescence with high quantum efficiency nescence quenching: sultant detectable fluorescence could po-
(6). This luminescence can be described in ␾ ⬚兾 ␾ ⫽ 1 ⫹ KSV[quencher], tentially be harnessed as a biosensor; spe-
terms of semiconductor band theory. On cifically, removing minute amounts of
photoexcitation, an electron is excited where ␾° is the intensity of fluorescence in MV2⫹ from the polymer could theoreti-
from the highest occupied energy band the absence of the quencher and ␾ is the cally increase the fluorescence in propor-
(the ␲-band) to the lowest unoccupied intensity of fluorescence in the presence tion to the amount of MV2⫹ removed. By
energy band (the ␲*-band). The excited of the quencher. The equation reveals that connecting the quencher to a biologically
electron and the oppositely charged ␾°兾␾ increases in direct proportion to the interesting ligand through a f lexible
‘‘hole’’ (the empty state in the ␲-band) concentration of the quenching moiety, tether, Chen and colleagues were in fact
attract one another. When the excited and the constant Ksv defines the efficiency able to demonstrate that such an approach
electron recombines with the hole, a pho- of quenching. When all other variables are is feasible (Fig. 2) (1). First, they con-
ton is emitted (luminescence or fluores- held constant, the higher the Ksv, the lower structed the quencher兾ligand moiety by
cence). The wavelength of the absorbed the concentration of quencher required to covalently linking the MV2⫹ to biotin, a
light is determined by the ␲-␲* energy gap quench the luminescence.
and can be manipulated by altering the Chen et al. demonstrate that the lumi-
See companion article on page 12287.
molecular structure of the polymer. Con- nescence of the sulfonated (anionic) and
‡To
jugated polymers have been demonstrated water soluble conjugated polymer, MPS- whom reprint requests should be addressed at: Louis
Stokes Cleveland VA Medical Center, 111K(W), 10701 East
with emission colors that span the full PPV (Fig. 1), is quenched by the cationic Boulevard, Cleveland, OH 44106. E-mail: pxh24@po.
range of the visible spectrum (6). molecule methylviologen (MV2⫹) with a cwru.edu.

PNAS 兩 October 26, 1999 兩 vol. 96 兩 no. 22 兩 12219 –12221

 quench the MPS-PPV, albeit not as effi- will dominate. In the MV2⫹–biotin exam-
ciently as the MV2⫹ alone. The investiga- ple, the quencher is presumably pulled
tors then showed that as little as 1 ⫻ 10⫺8 away from the polymer chain by complex-
M avidin added to a mixture of MV2⫹– ing with the high molecular weight pro-
biotin plus MPS-PPV unquenched the flu- tein, avidin. Whether this hypothesis is
orescence. Importantly, avidin did not indeed correct remains to be formally
have any effect on the quenched state of tested.
MPS-PPV plus MV2⫹ alone (the avidin The potential for adapting the biosen-
effect only occurred in the presence of sor described by Chen et al. for use with
MV2⫹–biotin). Moreover, the specificity antibody:antigen pairs or DNA:DNA (or
of the biotin–avidin interaction was dem- DNA:RNA) interactions is particularly in-
Fig. 1. Chemical structure of a single repeat unit onstrated by showing that addition of an- teresting (Fig. 2). Presently used ap-
of the luminescent polymer MPS-PPV. other protein of similar size to avidin, the proaches for detection of such biologically
cholera toxin receptor, did not unquench relevant molecules include the enzyme-
the fluorescence. linked immunosorbent assay (ELISA), an
small ligand molecule (molecular mass of As noted earlier, when the acceptor is antibody-based heterogeneous assay tech-
⬇250 Da) that specifically binds with high removed from the vicinity of the polymer nique (13, 14). Although this now stan-
affinity (binding constant, Kd ⬇ 10⫺15) to chain by ⬇1 nm, electron transfer to the dard technology is both sensitive and spe-
the much larger protein, avidin (molecular acceptor will be so rare that the radiative cific, it requires two antibodies specific for
mass of ⬇64,000 Da). The MV2⫹–biotin recombination of the electron and the different antigenic regions (epitopes) of
(quencher兾ligand) was able to effectively hole on the polymer chain (luminescence) the molecule to be detected and is a

Fig. 2. Schematic representation of a polymer biosensor. (A) Quencher-Biotin is removed by avidin and results in fluorescence as performed by Chen et al. (1).
(B) Quencher linked to an antibody may be removed from the polymer by its specific antigen and result in fluorescence. (C) Quencher linked to an oligonucleotide
may be removed from the polymer by a specific complementary sequence of DNA or RNA and result in fluorescence.

12220 兩 www.pnas.org Heeger and Heeger

multistep process with several washes tifiable using a standard fluorimeter (1). The reported studies have further dem-
needed to separate bound from unbound Finally, and perhaps most importantly, the onstrated detection of only a single inter-
antigen. Moreover, an ELISA requires a reaction occurs essentially instantly. This action between biotin and avidin under
significant amount of time (hours to days) feature would not only save time (and ideal experimental conditions (1). As
to complete. The PCR methodology, ca- therefore money), but could potentially be noted, the avidin– biotin interaction is
pable of detecting even single copies of extremely useful for ‘‘real time’’ detection high affinity (binding constant, Kd, of
specific RNA or DNA sequences, has of life threatening toxic substances in air 10⫺15) whereas typical antigen–antibody
further revolutionized molecular diagnos- or water. interactions are of much lower affinity [Kd
tics (15). Newer modifications using a Although these ideas are extremely pro- of 10⫺5–10⫺10 (17)]. It therefore remains
fluorescent oligonucleotide primer have vocative, there are many questions to be to be determined whether a quencher兾
even been developed such that it is now addressed regarding the practical use of antibody can be effectively removed from
possible to quantify the results in real time such polymer-based biosensors. How does the polymer to induce a detectable signal.
(16). Nonetheless, the procedure requires the size and charge of the ligand affect the Similar issues are raised for the potential
multiple steps and at least several hours to ability of the quencher兾ligand to quench detection of DNA:DNA interactions by
obtain results. polymer fluorescence? If the ligand teth- this method. Moreover, to be truly useful
The polymer-based biosensor proposed ered to the quencher is significantly larger as a biosensor, the assay needs to be tested
by Chen et al. has the potential to be a than biotin (i.e., antibodies or antibody under less ideal conditions: i.e., detection
significant improvement over these stan- subunits have molecular weights of of relevant molecules from unpurified wa-
dard technologies. If it were possible to 26,000–150,000 Da), it might interfere ter or air samples, or from even blood or
link an antibody or a specific oligonucle- with the complexing of the quencher to serum. Finally, although the work of Chen
otide to the quencher (Fig. 2), then it the polymer chain. In addition, the amino et al. demonstrate that their assay can be
might be possible to rapidly and specifi- acid composition of antibodies can vary quantitative, it remains to be determined
cally detect protein, carbohydrate, nucleic significantly, consequently affecting their whether it is sensitive enough to detect
acid, or other antigens under a variety of overall charge. Quenching may be ineffec- relevant molecules at concentrations use-
conditions. One important advantage to tive if the charge or size of the ligand ful for medical diagnostics or toxicology
the approach is that the polymer-based prevents effective interactions with the studies.
biosensor represents a homogeneous as- polymer. Similarly, it will be important to Despite these numerous unanswered
say. The quencher兾ligand maintains the determine the size and charge restrictions questions, the fundamental discovery that
polymer in an ‘‘off’’ state, and detection of antigens that can be detected in this polymer luminescence can be manipu-
(fluorescence) occurs only when the mol- manner. If the antigen to be detected is lated quantitatively is an important one.
ecule of interest binds to and removes the too small, it might be ineffective in pulling The investigators have opened a new di-
quencher兾ligand. The assay therefore the quencher兾ligand away from the lumi-

COMMENTARY
rection of research that, if generalizable,
does not require differentiating bound nescent polymer. This last point is partic- has the potential for producing a wide
from unbound antigen; only the bound ularly relevant to the studies by Chen et al. range of biosensor applications.
antigen removes the quencher from the in that the detected unquenching of fluo-
polymer and induces a signal (fluores- rescence was likely due, in part, to steric We thank Neil Greenspan (Department of Pa-
cence), eliminating the need for washing effects; the large size of the avidin may thology, Case Western Reserve University
steps (i.e., in ELISAs). The induced flu- have prevented avidin-bound quencher兾 School of Medicine) for his helpful discussions
orescent signal seems to be readily quan- biotin from interacting with the polymer. and insightful comments.

1. Chen, L., McBranch, D. W., Wang, H.-L., Helge- 6. Hide, F., Diaz-Garcia, M., Schwartz, B. & Heeger, 12. Swager, T. (1998) Acc. Chem. Res. 31, 201–207.
son, R., Wudl, F. & Whitten, D. G. (1999) Proc. A. (1997) Acc. Chem. Res. 30, 430–436. 13. Margulies, D. (1994) in Current Protocols in Im-
Nat. Acad. Sci. USA 96, 12287–12292. 7. Drury, C., Mutsaers, C., Hart, C., Matters, M. & munology, eds. Coligan, J., Kruisbeek, A., Margu-
2. Bredas, J. L. & Silbey, R., eds. (1991) Conjugated de Leeus, D. (1998) Appl. Phys. Lett. 73, 108–110. lies, D., Shevach, E. & Strober, W. (Wiley, New
Polymers (Kluwer, Dordrecht, The Netherlands). 8. Yu, G., Gao, J., Hummelen, J., Wudl, F. & Hee- York), Vol. 1, pp. 2.1.2–2.1.20.
3. Salaneck, W. R., Lundström, I. & Rånby, B., eds. ger, A. (1995) Science 270, 1789–1791. 14. Matesic, D., Lehmann, P. & Heeger, P. (1998)
Transplantation 65, 906–914.
(1993) Conjugated Polymers and Related Materials: 9. Heeger, A. & Diaz-Garcia, M. (1998) Curr. Opin.
15. Wang, A., Doyle, M. & Mark, D. (1989) Proc. Nat.
The Interconnection of Chemical and Electronic Solid State Phys. Mat. Sci. 3, 16–22. Acad. Sci. USA 86, 9717–9721.
Structure (Oxford Univ. Press, Oxford). 10. Sariciftci, N., Smilowitz, L., Heeger, A. & Wudl, F. 16. Higuchi, R., Dollinger, G., Walsh, P. & Griffith, R.
4. Heeger, A., Kivelson, S., Schrieffer, J. & Su, W. (1992) Science 258, 1474–1476. (1992) Bio兾Technology 10, 413–417.
(1988) Rev. Mod. Phys. 60, 781–850. 11. Kraabel, B., McBranch, D., Sariciftci, N., Moses, 17. Karush, F. (1978) in Immunoglobulins, eds. Lit-
5. Heeger, A. (1998) Solid State Commun. 107, 673– D. & Heeger, A. (1994) Phys. Rev. B Solid State 50, man, G. & Good, R. (Plenum, New York), pp.
679. 18543–18552. 85–116.

Heeger and Heeger PNAS 兩 October 26, 1999 兩 vol. 96 兩 no. 22 兩 12221