~

by Jerome S. Schultz

'

3-year-Old
bed,

man

recovering

lies

in a hospital

after

a routine

minuscule electrodes or optical sensors; the latter, biomolecules that can

recognize a target substance.
,s. The potential applications of these
devices are as varied as the molecules
that they may incorporate. Medicalcare
stands to benefit most clearly and immediately from blosensors,not only In
clinical testing but also In the manufacture of phannaceuticals and the development of such replacementorgans
as an artificial pancreas for diabetics.
Biosensors are also being used to determine food quality and safety and to.
detect enviromnental pollutants.

brane also prevented the entry of enzymes that could degrade the biocatalyst.)The more glucoseenteredthe sensor, the more oxygen was consumed
by the enzyme.Lower oxygenreadings
translated directly to higher glucose

hip-replacementoperation.Suddenly his heart begins beating Irregularly. As physicians and nurses work
to restore the heart's normal rhythm,
levels.
they draw blood samples to analyze the
, Clark's device never found its way
oxygen, potassium and caldum levels,
into routine patient care. Its accuracy
pH, hematocrit and other factors that
dependedcritically on the rate at which
might provide a clue to the source of
oxygen and glucose dUfused into the
the danger.
sensor,and that rate could change ei2, The laboratory results will typically
ther as the patient's blood oxygen levcomeback in an hour or more, perhaps
el
varied or as dots formed on the sentoo late to be useful Moreover, inforsor's surface. Nevertheless,the device
mation about tile patient's blood chemprovided a conceptualbase for subseIstry 10 or 15 minutes after the arrhyth- "
quent work by Clark and others: a bioT
he
evolution
of
the
first
biosensor
mia has started is not all the physicians
began In the mid-1950s, when Le- logical system sensitive to a particuwantto know; how the blood chemistry
land C. Clark, Jr., of the Chil- lar compound. a ph}!Sicaltransducerto
dlanged before the heart beganbeating
Irregularlywould be evenmore helpful. dren's Hospital ResearchFoundationin convert the chemicalChangelnto a meBlosensors-detectors based on se- Cincinnati invented an electrode de- ter reading and membranesto separate
lectivemolecular components of plants signed to measuredissolved oxygenin
~.o
---~
or anImals-are already making analyt- the blood of patients undergoing sur~ ical results available at some patients' gery. He surrounded a standard platbedsides within a few minutes. Sys- Inum electrodeand referenceelecttode
temSnow being developed may be able with a plastic membrane permeable to
to provide not merely a snapshot of a gases.The voltage bias of the platin~"
patient's condition but a continuous electrodewas set so that the rate of cur"';;
rent flow through the drcult I -biochemicalvideotape."
Modern biosensors evolved from the on the rate at which oxygen diffused',
.marriage of two disparate disciplines: through the membrane,wWch in turn,:
Information technology,exemplified by was directly proportional to the exter":
microcircuits and optical fibers, and na1oxygenconcenttatlon.
molecularbiology. The former provides r By 1962 Clark had extended the:"oxygen electrode" to senseblood glu~~ cose levels. He coated the oxygen sen~,
sor with a layer of gel contaIning a bioJEROMES. SCHULTZis director of the
catalyst:'the enzyme glucose oxidase,:
Center for Blotechnoiogy and BIoengifollowed by a semipermeable dialysis"
neering at the University of Pittsburgh,
membrane that allowed glucose to difwhere he conducts research on blosensors and Immunotechnology. Previousfuse into the sensor but prevented the :
ly he was chairn1anof the chemical enenzyme from diffusing out. (The memgineering department at the University
of MIchigan; between 1985 and 1987 he
also served as deputy director for crossdlsdpl1nary researchat the National Sdence Foundation, where he contributed to the development of the EngineerIng Research Center program. Schultz
receivedhis B.S.and M.S.In chemical engineering from Columbia University and
his doctorate In biochemistry from the
University of Wisconsin.

SCIENTIFICAMERICAN August 1991

HYPODERr.fiC NEEDLE holds a

molea:xIesfrom binding sit~s on thein-:;.; ,

terlor wall of the sensor. An opticalft-::,:
excite the freed dextran; it also carries:
the resulting signal to a detector.
..:

~:3

sensorelementsand protect them from ry out a complex sequenceof reactions out of the laboratory and into common
that yield a responseto amino acidsand use is the developmeni:-generally unthe outside world.
~ The next major innovation came in other biologically important molecules. related to biosensorresearch-of tech.1969, when GeorgeG. Guilbault of Lou- Among those used are banana pulp for niques for stabilizing blomolcculesand
; Islana State University at NewOrleans measuring dopamine, corn kernels for tethering them to surfacesso that they
.built a systemto measure urea in body pyruvate, cucumber leaf for cysteine, retain their activity. Chopping up bafluids. His device used urease, the en- sugarbeets for tyrosine,rabbit liver for nanas to measure dopamine at a pa'zyme that converts urea Into carbon guanine and powdered rabbit muscle tient's bedside Is clearly Impractical.
.dioxide and ammonia. An electrode for adenosinemonophosphate.
Many of the same techniques used to
sensedthe resulting changesin ammo- I y Rechnitz has gone evenfurther in co. affix enzymes or antibodies to a surnium-ion concentration.Guilbault's sen- opting parts of biological systems:one face so that they can be used in labosor marked a significant advance be- of his sensors contains an antennule, ratory assays or biochemical manufaccause It relied on potentiometric detec- or small sensing organ, from a Mary- turing, however,canalso be used to attion, a technique that has sincebecome land blue crab, dissected to exposeits tach molecules to a biosensor.
widely used.
nerve fibers to an electrode. The an- IS Although removing proteins from
to WhereasClark's sensormeasured the tennule-basedsystem canmeasurethe their cellular environment can disrupt
current flowing through the electrode, concentrations of a number of drugs their structure and expose them to
a potentiometric sensor measures the and environmental toxins.
chemical attack, the same techniques
voltage bias required to maintain cur- ,'I, The crab sensor and devices like it that affix proteins to a substrate also
rent flow at zero. The electrode does offer the opportunity to unravel their tend to stabilize them. The chemical
not consume any of the reactants and information-transmitting structure and bonds that hold a protein moleculeto a
thus is less susceptibleto errors caused build simpler sensors that incorporate polymer bead, for example,may crossby changes in external conditions. In the same molecules. Suchsensorsalso link parts of the protein so that It Is
addition, potentiometric systems have point out the advantagesof generality: less likely to unfold and becomesmore
a logarithmic response curve and so although enzymes and antibodies pro- resistant to enzymatic degradaUoT\.
As
can track concentrations over a 100- vide admirable power to detect a single a result, biosensors canbe stored in eifold range.
compound, other biomoleClliesmay be ther moist or dry conditions for a week,
more useful in detecting the presence a month or, in some cases,as long as a
I' I n the decades following the de- of any member of a broad class of year without losing sensitivity.
velopment of these electrochemical
chemicals. Richard F. Taylor of Artbur II~ Advances In membrane technology
methods, roughly 100 different en- D. Uttle, Inc., for example,has built a have also given biosensor workers a
zymes have been used In blosensors. sensorincorporating the membrdnere- free ride in Improving their systems.
Workershave realized, however,that In- ceptor for acetylcholine, which trans- Membranescan nowbe tailored to sepdividual enzymes are not the only use- mits messages from nerve Ilbers to arate solutes based on molecular size,
ful biocatalysts.RecentlyGarryA. Rech- muscles.The device can detect several charge or solubility. One widely used
nitz of the University of Hawaii has different nerve gases.
laboratory assay system has as many
.'3hown that tissue preparations can car- Ii One key factor in moving biosensors as a half dozen membrane layers, each

SCIINl1I'IC AMI-:RICANAugust 1991

:;"~tf

Evolution of the Glucose Biosensor

GAS-PERMEABLE
MEMBRANEIsolates electrodes from blood or
other biological fluids but allows'
oxygen to diffuse Into the potassium chloride solution surrounding the electrodes.

DIALYSISMEMBRANE-permeableto
Ions and small molecules-encloses
gel contaIning glucose oxidase enzyme, which converts glucose to
gluconic acid and consumes oxygen.
The rate at which oxygen reaches
the Inner solution depends inversely
on the level of glucose In the blood.

spectrallypure ligbt sources,such as tran complex cannot diffuse out, but an optical fiber does not travel only
light-emittingdiodesand solid state la- glucose can diffuse In. As glucose enters through the core of the fiber; instead
~ers.In 1969,at the NationalInstitutes the sensor, it displaces dextran from part of it travels through a region that
of Health,GeraldG. Yurek and Robert some ConA binding sites; the higher the extends outward roughly 1,000 angBowmandemonstratedone of the first glucose concentration, the more dex- stroms into the mediwn surrounding
fiber-opticsensorsfor clinical analysis, tran is forced into solution. Meanwhile the core.If a fiber's protectivecladding
.a cololimeter.that measuredthe bind- light transmitted through an optical is removed,any material adhering to,
Ing of dyes to kidney tubules. Optical fiber causes the dextran complex in so- the core can absorb this evanescent
fibers can serveas remotespectropho- lution to fluoresce, thereby producing waveand fluoresce.,t
tometersto measurethe reflection or a signal; dextran bound to ConA on
The late TomasB.Hirschfeld of Lawtransmissionspectnunof a fluid; fluo- the walls of the dialysis fiber is not in rence Uvermore National Laboratory
rimetersto sensereemissionof particu- the path of the light and so produces attachedantibodies to the surface of
lar wavelengthsof light; or turbidime- no signal.
a decladded fiber and measuredthe
tersto measurelight scattering.
,~b The dialysis fiber does not merely natural fluorescenceof antigens that
"... There are three broad categories provide an isolated volume in which bound to them. Evanescent-wave
senof fiber-optic biosensors.The first, a the reaction can take place; it also con- sorscanalsomeasurecompetitivebindstraightforwardextensionof electronic serves the reagents. In a typicallaboraing of the kind employedin my glucose
biosensor techniques,simply detects tory assay. the fluorescein-labeled dex- sensor.If one binds antibodiesto the
changesin a target substance'soptical tran (or other marker) forced into so- fiber and then adds a known quantiproperties rather than in its elecnical lution would be lost. For that matter, ty of fluorescentlylabeledantigenand
ones.The other two, evanescent-wave the ConA (or an antibody) could not be an unknown analyte, the amoUnt of
andsurfaceplasmondevices,makeuse reused after direct e"'"posure to body fluorescencecausedby the evanescent
of the particular way that optical fibers fluids. Packaging the entire system in waveindicatesthe ratio of labeledantia biosensor permits continuous, long- gensto the unlabeledonesin the samtransmitlight.
The first kind of optical biosensor term measurements. The one remain- ple beingtested.~
consistsof a cell bounded by a semi- ing Achilles heel of the device is that it c 'I An alternative to the evanescentpermeablemembrane,reagentsInside is invasive-it must be placed inside tis- wavemethod,a so-calledsurfaceplasthe membrane or bound to its inner sue to maIntaIn contact with the blood. mon deviceworks by meansof a film
surface,a fiber to illuminate the cell Eventually Uliection or blflanunatlon of metal, suchas silver, deposited on
anddetectorsto measurethe changeIn will ensue; no one has yet figured out the surfaceof a thin piece of glass that
opticalproperties.In most cases,a sin- how to make the body accept probes of. acts as a light guide. The conductive
gle fiber both carries light Into the cell this kind for long periods.!
film providesa pathway(the plasmon)
and collects transmitted or reflected; r The second type of optical sensor, for light energyandso changesthe critwhich is based on the evanescent-wave ical angleof incidenceat which light is
light for analysis.
~~ My own work asa biochemicalengi- approach, does not have to infer the trapped within the glass.The new critineeron variousartlfidal-organprojects number of labeled biomolecules dis- cal angle dependsvery sensitivelyon
ledmc'to applytheseopticalteclmiques placed from a receptor from the num- the amountof materialadsorbedon the
to build yet anotherglucosesensor,one ber in solution. It senses competitive metal film. The surfaceplasmonsignal
that might eventuallybe suitable for binding directly. Evanescent-wave de- doesnot requirelabeledmoleculesand
an artificial pancreas.The physiciansI tectors rely on the fact that the ener- competitiveassays;if antibodiesor othworkedwith pointedout that therewere gy of a light wave transmitted through er bioreceptorsare attachedto the demany excellentinsulin pumps, but all
~-thedevicesstill had to be programmed
by.hand based on the results of plnBiosensors and Their Applications
prtckblood tests.Theideal sensorfrom
Substance
aclinicalpoint of viewwould makeconBiological Sensor
Physical Sensor
Measured
tinuousreadings;it would consumeneitherglucose(thus yielding a true equiAntibody to benzo(a)pyrene
Optical-fiber fluorimeter
Benzo(a)pvrene
librtum measurement)nor reagents;it
Creatinine Imlnohydrolase
A.mmonla field-effect transistor
Creatinine
would have no elecnic connectionsto
thebody; and,if at all possible,it would
NADH and dehydrogenase
Oxidation-reduction electrode
Ethanol
be noninvasive,so that long-term inflammationor otherreactionscould be
Polarized light
Antibody to gamma globulin
Gamma globulin
avoided.
Oxygen electrode
Antibody to lidocaine and
lidocaine
~~ The device I built fulfills most, but
ferrocene-lidocaine complex
not all, of the clinical requirements.In
--addition, it servesas an archetypefor
Conductivity measurement
Nerve gas
Acetylcholine receptor
opticalbiosensorsthat candetecta virtually unlimited range of molecules.It
Piezoelectric crystal
Antibody to parathion
1'arathlon
is based on the fluorescent immunoassay teclmique used in clinical analyses.

In this case,ConA, which binds both

glucoseand dextran (a glucose polymer).is bound to the interior of a pollowdialysisfiber,a membraneoriginally developed for an artificial kidney.
The fiber Is filled with a dilute solution
of fluorescein-labeled
dextran.The dex-

Penicillin

Beta-lactamase

Thermistor

Testosterone

Bioluminescenceenzymes:
dehydrogenase and luciferase

Optical-fiber fluorimeter

Theophylline

Antibody to theophylline

Surface plasmon resonance

Vitamin

Bacteria (Escherichia

Oxygen electrode

812

coli)

SCIENIlFIC AMERICAN August 1991

)'\.

f~

':J.

vice, it is possible to measure precisely

how much material in a fluid sample
binds to those receptors.
,30 Becausethe surface plasmonmethod
does not require labeled molecules, It
can be used in biosensors for a wide
range of substances.Indeed,one company, Phannacia,has created a system,
intended for researchlaboratories,that
contains generic surface plasmon sensors. Workers depositwhateverblomolecule they want to investigate on these
sensors.
3' Although biosensors are currently
most prevalent in medical uses, other
applications may prove at least as important in the long run. For example,a
sensor developed by lsao Karube and
Shuichi Suzuki of the Tokyo Institute
of Technology measures biochemical
demand for oxygen, an index of the
level of organic materials in polluted
water. The sensor, based on a yeast,
produces readings in 30 minutes as
opposed to the five days required by
conventionalmethods.
31--Other sensorscould be used to monItor such toxic chemicals as polychlorinated biphenyls (PCBs),chlorinated
hydrocarbons or aromatic compounds.
In the event of an accident, the sensors trigger an automatic, immediate
responseand inform emergencyteams
of exactly what compounds they are
about to confront.
~ Unpleasant and potentially dangerous compounds enter the environment
not only as a result of accidents. In the
caseof food, they accumulateas a part
of the normal process of spoilage.Both
Canadianand Japanesecompaniesare
marketing blosensorsthat measurethe
level of xanthine and other compounds
in fish to determine Its freshness. (In

EVANESCENT-WAVE SENSOR

FIBER-ornc BIOSENSOR
(top) contains
receptor D1oleCtUes
fixed to the ~er
wall of a sernipenneabledialysis D1eD1brane and large fiuorescently labeled
D1olecules(dextran)that bind to the receptors. When glucose D1oleCtUes
diffuse through the D1embrane,they displace labeled D1oleCtUes.
The displaced
D1oleCtUes
are trapped in solution with.
in the D1eD1brane.
They absorbfocused
laser light entering the sensorand produce a signal proportional to the nUD1ber of unlabeled D1oleCtUes
that have
displaced theD1. An evanescent-wave
sensor (bottom) exploits the fact that
SOD1e
of the energy passing through an
optical fiber penetratesthe boundaries
of the light-guiding core. Antibodies or
other blomoleCtUesattachedto the surface of a dedadded fiber can bind the
compound to be sensed,which then absorbs part of the evanescentwav~ and
thereby producesa signal.
SCIEN"nFlC AMERICAN August

1991

.G~

Bedside Analysis
DISPLAY

A:

and-held

der

KEYBOARD

~-/ -

MEMORY
MICROPROCESSOR
ANALOG INTERFACE

SAMPLE

CARTRIDGE

SAMPLE INLET

-""

'-

SENSORS

;\;'~i
CALIBRATIONc
SOLUTION

,
japan the freshness index of fish is
typically printed on the package.)Sensorsto measurethe quality of beef and
other foods are also beingdeveloped.

analyzer

evaluation

currently

at

the

un-

Hospital

of

the University of Pennsylvania

demonstrates how blosensors might
find their way Into clinical use. The
I-STATPCA is one of several bedside
biosensor systems now being readied
for what could be a substantial market.
It simultaneously makes six commonly
requested chemical measurements on
a patient's blood-sodium, potassium,
chloride, urea nitrogen, glucose and hematocrit-producing results In less than
two minutes. The bedside tests cost
more than ones performed In a central
laboratory, but their immediacy may
make them more effective.
The PCAachieves accuracy comparable to that of laboratory equipment by
using a disposable cartridge containing
six blosensors and a calibration sample.
A medical worker places 60 microliters
of blood In the cartridge; the analyzer then measures both the calibration
sample and the patient sample. It displays test results and also stores them,
keyed to time and the patient's Identification number, for later analysis. The
cartridge-baseddesign adopted by manufacturers will make it possible to perform a different set of tests once the appropriate sensors have been developed.

inate the distinction between the biomolecule that senses a compound and
the electrode that sensesthe molecule's
response. Adam Heller of the University of Texas at Austin, for example, has
'31-\ I
ndustrial proces~ control is a fi- inttoduced electton "relays" into a pronal areain which biosensorscould tein so that chemical binding would be
make a substantial difference. Al- telegraphed directly to an electrode inthough detectors in automatedchem- stead of being measured indirectly by
ical process plants can already mea- mediators or by changes in quantities
sure pressltre,temperatureand acidi- such as pH or oxygen consumption.
ty in real time,-biosensorswill be able Such an approach would enable bioto determinethe chemicalcomposition sensors to attain much higher sensitivof materialsin the processflowaswell. ities. If researchers can tailor the chem.
Suchmeasurementsare especiallyim- ical properties of such electroactive
portant in the biotechnologyindustry, molecules as well, they could also build
which currently has no way to moni- more selective biosensors.
tor preciselythe culturing of microor- : ~~ Biosensors will always be less senganisms in fermenters that produce sitive and specific than laboratory assuchdrugs or active proteins as inter- says because of the trade-offs required
feron or insulin. Indeed,biosensorde- to make an instrument that works
velopmentcould feed back synergis- rapidly and reliably in the field. The ultically so that improved manufactur- timate rival of the biosensor, howeving techniquescould allow the cheaper er, is nature itself, and in some cases it
productionof a wider rangeof sensing is clear that researchers will soon be
able to build detectors that are more
molecules.
'3~ Evenas the rangeof biosensorappli- sensitive and specific-and faster to
cations expands dramatically,investi- respond-than the organisms from
gatorsare exploringthe possibilityof a which their molecular mechanisms are
generationof sensorsthat would elim- derived.

FURlliER

READING

F~ER-OYnC SENSORS FOR BIOMEDICAL

APPUCAll0NS. 10hn I. Peterson and
Gerald G. Vurek In Science, Vol. 224,
pages 123-127; April 13, 1984.
CHEMICAL SENSINGIN PROCESSANALYSIs.
T. Hirschfeld, 1. B. Callis and B. R. Kowalski In Science, Vol. 226, pages 312318; October 19, 1984.
A MINlA11JRE OYnCAL GLUCOSE SENSOR
BASED ON AFFINITY BINDING. Sohrab
Mansouri and Jerome S. Schultz In Bioi
Technology, Vol. 2, No. 10, pages 885890; October 1984. .
BIOSENSORS:FUNDAMENTALS AND APPUCAllONS. Edited by Anthony P. F. Turner, Isao Karube and George S. Wilson.
Oxford University Press, 1987.
BIOSENSORS.Garry A. Rechnitz II:1 Chemical and Engineering News, Vol. 66, No.
36, pages 24-31; September 5, 1988.
IMMOBIUZED ENzYMESAND CEil.5, Part D.
Edited by Klaus Mosbach. Methods In
Enzymology, Vol. 137. Academic Press,
1988, .
MOLECULAR ELECl"RONlCS: BIOSENSORS
AND BIOCOMPUTERS.Edited by Felix T.
Hong. Plenum Press, 1989.
BIOSENSORS.ElIzabeth A. H. Hall. Open
University Press, 1990.
PRINCIPLES OF CHEMICAL SENSORS. Jitl
Janata. Plenum Press, 1990.
.

SCIENTIFICAMERICAN August 1991