Optical sensing of pH and pCO2 using phase-modulation fluorimetry and resonance energy transfer

179

Anaiytrca Chmrea Acta, 272 (1993) 179-186 ElseMer Science Pubhshers B V , Amsterdam

Optical sensing of pH and pC0, using phase-modulation fluorimetry and resonance energy transfer Joseph R Lakowlcz, Hemyk Szmacmslu and Mutlu Karakelle Center for Fluorescence Spectroscopy, Department of Bdogtcal Chetnrstry, Unrvers@ of Maryland, School of Medmtw, 108 North Greene Street, Baitmore, MD 21201 (USA) (Recensed 7th May 1992, revised manuscript received 2nd October 1992)

Abstract We have developed opt& sensors for pH and/or pC0, based on phase shift and modulation measurements of time-resolved fluorescence energy transfer pH-msenatrve donors and pH-sensltlve acceptors were encased m C&-permeable silicon membranes The extent of energy transfer depends on the pH-dependent absorption spectra of the acceptor The measurements were made msensdlve to inner-filter effects of the total intensity by using the phase-modulation method to determme the donor decay times The general apphcaluhty of this method was demonstrated using three donor-acceptor pairs, all of which can be excited urlth an mexpenswe green 543-nm He-Ne laser Keywords Fluonmetry, Donor-acceptor complexes, Energy transfer, pCO,, pH

Opt& measurements of pH and/or pC0, are of wide interest m analytical and chmcal chemtstry [l-3] At present, most optlcal sensmg of pH/pC02 IS based on measurements of the steady-state fluorescence mtennty, as determined by the absorptive and enusswe properties of the sample [4-71 Steady-state intensity measurements are sensltlve to light losses, lamp dnft, probe bleaching, and the optical properties of trssues, and such measurements require frequent recahbratlon To compensate for these dtilculties, wavelength-ratlometnc methods have been proposed, particularly usmg the pH-senatlve fluorophore pyranme [8,9] However, the use of wavelength-ratlometrlc compensation requires that the optical artifacts affect both wavelengths Correspondence to JR Lakowuz, Center for Fluorescence Spectroscopy, Department of B~oloacal Chenustry, Umverslty of Maryland, School of Medicine, 108 North Greene Street, Baltimore, MD 21201 (USA) 0003~2670/93/$06 00 Q 1993 - Elsevler hence

m a smular manner Furthermore, there are relatively few wavelength-ratrometrlc probes avadable, particularly Hrlth long-wavelength absorption and emlssron maxuna Pyranme cannot be excited with Inexpensive laser sources, and its use requires two excitation wavelengths which are not conveniently available usmg laser sources The dlfficultles described above can be clrcumvented using tnne-resolved fluorescence and sensors contammg a fluorescence donor and a pH-sensltwe acceptor (Scheme 1) pH sensors based on mtenslty-based measurements and energy transfer have been described previously by Walt and coworkers [lo-131 The use of lifetime or decay tune measurements offers many advantages for sensmg apphcatlons [14] because decay tunes are generally not dependent on the macroscop~c optical properties of the sample, and are not sensitive to the total Intensity Fluorescence hfetnne measurements are generally thought to require sophlstlcated mstrumentatlon Advances

Publishers B V All nghts reserved

180

JR Lakowzcz et d /Anal Chm Acta 272 (I 993) 179-186

m lasers, optics and electromcs make it easily conceivable to develop Inexpensive and robust Instruments based on the phase shift and/or demodulation methods 1151 However, construction of practical mstrumentatlon requires the use of snnple and mexpenslve laser sources [15-171 Decay tune sensmg of pH(pC0,) 1s hmlted by the lack of pH probes which can be excited Hnth sunple laser sources such as He-Ne or diode laser The fluorescence pH(pC0,) sensor must display good absorption of the laser light, a high quantum yield, and sensltlvlty m the desired pH range It 1s difficult to obtam all these requlrements m a smgle chromophore However, these characterlstlcs are more easily obtamed by usmg a two-part sensor (Scheme 1) In this case the fluorescent donor can be selected for its absorption, ermsslon and decay tune charactenstlcs, without concern for its sensmvlty to pH(pC0,) The acceptor need not be fluorescent, and need only display a change m absorption m response to pH(pC0,) m the wavelength range of the donor emlsslon The donor and acceptor can be covalently lmked (see Dlscusslon) or simply mixed together as described m this report The mechanism of mducmg a pH(pCO,)-dependent change m the donor decay tune 1s fluorescence resonance energy transfer (FRET) The phenomenon of FRET 1s non-radiative energy transfer from the fluorescent donor to the acceptor, without ermsslon and reabsorption of a photon This latter process of ermsslon and reabsorption 1s an mner-filter effect which depends on the size and macroscopic optlcal properties of the sample In contrast, FRET 1s a through-space interaction which can be reliably predicted based on the spectral properties of the donor and acceptor Importantly, FRET decreases the mean decay tnne of the donor, whereas radiative energy

F I

transfer (1 e , reabsorption and emlsslon) does not yreld easily predictable changes m the donor decay time For unlinked donor-acceptor pairs the phenomenon of FRET requires an acceptor concentration m the range of l-10 mM 1181 Such acceptor concentrations result m high optical densltles at the excitation and emlsslon wavelengths, makmg mtenslty measurement d&cult to use m a quantitative manner [12] At these high acceptor concentrations there are, m addition to FRET, inner-filter effects which depend on the excltatlon and observation wavelength and on the detaded macroscopic propertles of the sample and detection optics For mstance, an Increase m absorption can decrease the observed mtenslty of the donor due to preferential absorption of the mcrdent light by the acceptor However, the mcreased absorption wdl also change the penetration depth of the madent light mto the sample, which also alters the observed mtenslty The non-ideal optical propertles of real-world samples result m addmonal dlfflcultles High absorbance or turbldrty results m low signal levels and dlfficultles m measurmg small changes m the signal Intensity These dlfficultles can be avolded by measurement of fluorescence decay times, which are independent of the total mtenslty so long as the donor emlsslon 1s detectable The lrfetnne can be determmed by the rate of change followmg pulsed excltatlon However, these nanosecond timescale processes can be more easily measured by the phase-modulation method [19-221 The decay time mformatlon 1s contained m the phase shift (0) and modulation cm> of the ermsslon relative to the mcldent light In the present report we describe pH(pC0,) sensing based on three different donors and two acceptors The donors were selected to be excitable

Energy transfer

%A kr

Donor chosen for use with inexpensive laser sources

Acceptor displays changes in its absorption spectrum in response to pH or pC0,

Scheme 1 Schematic of pH/pCO, sensors based on energy transfer

JR Lakowzcz et al /Anal

181

Chun Acta 272 (1993) 179-186

wrth a 543-nm He-Ne laser Since the phenomenon of FRET 1s predictable, this method can be extended to use with modulated laser diode sources by the selection of alternative donor-acceptor pairs

MATERIALS AND METHODS

Fabncatlon of the pH(pC0,) sensors Poly(Zhydroxyethy1 methacrylate) (polyHEMA) was synthesized as follows 2-hydroxyethyl methacrylate (Aldrich) monomer was mixed wrth 1% (w/w) 2,2’-azobls1sobutyromtrde (Polysclences) mltlator and degassed The monomer-mltlator mtiure was placed m a template well between two glass plates and cured at 75°C for 60 mm The polyHEMA film was subjected to repeated water extractions for removmg unreacted monomer and catalyst residues Slhcone rubber membrane was prepared usmg a two-part platinum catalyst curing formulation (FGB 001, Admiral) Parts A and B were mixed according to the supplier du-ectlons and degassed The mvrture was cast on a PTFE-coated glass plate using Doctor’s knife and cured at 150°C for 30 mm The donors and acceptors were dissolved at concentrations listed m Table 1 m a 40 mM NaHCO, (Aldrich) solution PolyHEMA hydrogel disks were soaked m the donor-acceptor-blcarbonate solution for 48 h for reaching an equlhbnum intake The equlhbrated hydrogel disks were blot-dried and sandwiched between two clrcular slhcone rubber membranes The circular slhcone membranes were glued together using a moister cure sdlcone adhesive (DC 3145, Dow

Coming) The resultant sensors were completely sealed and retamed then indicator solutions The sensors were stored at 100% relative hurmdlty conditions for retammg then eqmhbrmm water uptake The acceptor concentrations are those expected for sign&ant energy transfer usmg these donor-acceptor pairs Acceptor concentratxons m this range result m a sign&ant fraction of the donor population being Hrlthm the characteristic Forster distances for energy transfer, which 1s near 50 w The donor concentrations were chosen to be as low as possible consistent Hrlth an abUy to easily observe the ermsslon Low donor concentrations were also used to avoid FRET between the donors While such transfer could result m increased transfer to the acceptor [23,24], we wished to avoid donor-to-donor transfer so that the experiments revealed the usefulness of donorto-acceptor transfer for pH(pC0,) sensing Frequency-domaan measurements Multi-frequency phase and modulation data were collected on the frequency-domain mstruments described previously [21,22] For the multi-frequency measurements, we used 568~nm excitation from a cavity-dumped dye laser [22] For testing the pH(pC0,) response of the sensors we used a 543-nm He-Ne laser, which was mtenslty modulated with an acousto-optic modulator This light source was chosen because it IS practlcal for analytical or clmical sensmg applications In phase-modulation fluornnetry, the sample 1s excited with an mtensrty-modulated light source The expenmental observables are the phase shift of the emlsslon (0) and its modulation (m), both relative to the phase and modulation of the source

TABLE 1 Concentrations and wavelengths used m the pH(pC0,) Donor a

sensors

Acceptor

Identlty

Concentration (M)

Identity

Concentration (M)

Eosm R6G TRH

5 x 10-4 1 x 10-4 3 x 10-5

PR PR BTB

4 x 10-S 4 x 10-3 2 x 10-3

a Control (donor-alone) sensors were Identical but did not contain acceptor

Excitation (nm)

Ermsslon filter (mu)

543/563 543/568 543/568

580 600 600

182

.IR Lakowacz et al /Anal Chm Acta 272 (1993) 179-186

[191 These quantlhes are related to apparent phase (~~1 and modulation (7,) hfetnnes by tan I3= ~7~

(1)

m=(1+W%;)-1’2

(2)

where w 1s the modulation frequency m rad s-l The qualifier “apparent” 1s used because the hfetlmes (7p and T,,,,)are true decay times only for smgle-exponential intensity decays It 1sknown that energy transfer between randomly datrlbuted donors and acceptors results m more complex (multi-exponential) intensity decays

0

LW61

RESULTS

Absorption and enusslon spectra of the three donors [l&m, Rhodamme 6G (R6G) and Texas Red Hydrazlde (TRH)], are shown m Fig 1 The donors each show slgnlficant absorption at 543 nm The absorption and emlsslon spectra of the donors are not slgmficantly sensitive to the pH from65to92 Absorption spectra of the acceptors are shown m Fig 2 The absorption of both Phenol Red (PR) and Bromothymol Blue (BTB) increases for decreasing partial pressures of CO, The mcreased absorption IS the result of the increase m pH which occurs upon removal of CO, (Fig 2, inset) Importantly, these acceptors display senwtlvlty to pH m the physlolog&ly relevant range of 6-8, and to pC0, m the physlologtcally relevant range of O-40 Torr The sensitive range can be optmuzed as needed by adjustmg the blcarbonate and/or acceptor concentration The sensors could be made sensitive to pH rather than pC0, by usmg water-permeable membranes If needed, the donor and acceptors could be covalently linked to a polymeric matrw to prevent ddutlon of the chromophores At high pH the absorbance increases m the region of the donor enusslon, so that increased energy transfer IS expected at high pH It 1s also possible to identify donor-acceptor paus m which the extent of energy transfer and/or spectral overlap decreases at high pH

WAVELENGTH

( nm

1

Fu 1 Absorption and emwuon spectra of the donors Eosm, Rhodamme 6G and Texas Red Hydrazlde, ( -_)pH92, (----_)pH65

Frequency response curves of the donor alone controls and the donor plus acceptor-contmmng sensors are shown m Fig 3 The frequency responses shift to higher frequencies m the presence of acceptor, as seen for R6G-PR (top) and TRH-BTB (bottom) A less dramatic acceptordependent shift was seen for Eosm-PR (not shown) The short hfetune of Eosm prevented us from observmg its entire frequency response usmg an upper frequency llrmt of 200 MHz While we are capable of measuring at higher frequencies [22,27], we did not do so because we felt 200 MHz was a reasonable upper hmlt for hfetunesensmg mstrumentatlon, which 1s not hkely to use the hrgher-speed mrcrochannel plate photomultrplier tubes A functional sensor for pH or pC0, does not require measurement of a complete frequency response Measurement at a smgle hght-modula-

JR Lak0nw.z et al /Ad

Chm Acta 272 (1993) 179-186

tlon frequency 1s adequate Such measurements are shown m Figs 4 and 5 for phase and modulation, respectively In these measurements we used the 543~nm He-Ne laser with an acousto-optic modulator The phase angles increased monotonically with increased pa&al pressures of CO, (Fig 41, and the modulation decreased (Fig 51, indicating longer donor decay times as the CO, pressure increases It 1s nnportant to consider the accuracy posable from the phase-modulation data There 1s no single number which characterizes the accuracy because many sensor conflguratlons are possible, and one can obtain various types of data For instance, a single-phase measurement can be performed at one modulation frequency, one could use both the phase and modulation data at the same frequency, or one could measure phase and modulation at several selected frequencies The latter would be possible with an acousto-optlc-

PHENOL

RED

0

PC02

183

0 2

5

10

20

50

100

200

FREQUENCY (MHZ) - Fig 3 Frequency response of the donors m the absence and presence of acceptors Top (0) Rhodamme alone, ? = 6 25 ns, and (0) wth 0004 M Phenol Red, ? = 2 11 ns, energy transfer = 0 83 Bottom (01 Texas Red Hydrazlde alone, ? = 4 30 ns, and (0) wth 0 002 M Bromothymol Blue, ? = 182 ns, energy transfer = 0 78 A, = 568 nm, A, = 600 nm ? IS the mean llfetune from a multi-exponential analysis [22]

1x1

A

modulated laser source Based on conslderable expenence Hrlth phase-modulation fluonmetry, we know that phase angles are easrly accurate to 0 5” and 0 10” seems possible, particularly Hrlth dedlcated mstrumentatlon and numerical methods Assuming an accuracy of 0 5” or l%, the smglefrequency phase and modulation data can provlde pC0, values accurate to 0 25 Torr Hence, this sensing scheme appears to have adequate accuracy for chmcal use m blood gas determmations

PC02[Xl

DISCUSSION 500

550

600 WAVELENGTH

650

700

( nm)

Fig 2 pH(pCO&dependent absorption spectra of the accep tor Phenol Red and Bromothymol Blue The mset shows the relatronshlp between pH and pC02 for our expenmental condrtrons

The energy transfer pH(pC0,) sensors descrrbed m this report contamed hgh acceptor concentrations m order to place the acceptors wrthm the Forster distances for energy transfer There are two disadvantages to this approach

184

JR Lakowux et al /Anal

The first 1s attenuation of the observed mtenslty of the donor due to mner-filter effects which attenuate the donor emlsslon and partial absorption of the Incident light by the acceptor The fraction of the incident light absorbed by the donor 1s approximately gwen by the fractional absorption of the donor at the excitation wavelength [28] Hence, the high acceptor concentration results m decreased donor enusslon due to non-productive absorption by the acceptor A second disadvantage of our unlmked donors and acceptors 1s the dependence of the extent of energy transfer on the acceptor concentration [Cl This dependence 1s given by

-.,

1(~)=I”exp[-&-2~(&)lil]

(3)

(4)

0

5

10

15

20

25

30

35

pCO*( %I FQ 4 Dependence of the donor phase angle on pCOz The donor-acceptor paus were m polyHEMA hydrogel, 25°C (0) Texas Red Hydrazlde+Bromothymol Blue (0002 MI, Aex= 543 mu, A, = 600 run, 133 MHz, (A) Eosm + Phenol Red (0 004 M), A, = 543 nm, A, = 580 nm, 155 MHz, (m) Rhodamme 6G + Phenol Red (0 004 MI, A, = 543 nm, A,, = 600 nm, 133 MHz

Chun Acta 272 (1993) 179-186

I

0

5

Fig 5 Dependence Fig 4 for details

10

15

20

pcoz

(%I

25

30

15

of the donor modulation on pC0,

See

where 7. 1s the decay time of the donor m the absence of acceptors, [Cl, 1s the critical concentration of acceptor, y = [Cl/[Cl,, and R, IS the Forster distance [29,30] It 1s quite possible that the acceptor concentrations can change m a physlologlcal environment due to water evaporation or absorption, or due to osmotic effects The changes m acceptor concentration will require recahbratlon of the pH(pC0,) sensor The two disadvantages discussed above can be avoided by the use of covalently hnked donoracceptor pairs Because the acceptor IS covalently linked to the donor, both are present m a one-toone ratio Hence, there will be mrmmal non-productive absorption due to the acceptor, and the linked probe can be more dilute to avoid mnerfilter effects Importantly, the lmked probes will not be sensitive to dllutlon effects, so that the sensors will require httle If any cahbratlon This advantage of lmked donor-acceptor pairs has been recogmzed by others [lo,111 Several additional donor-acceptor pH probes are shown m Fig 6 A variety of “Gedanken” probes can be imagined with a range of spectral properties The R6G-PR probe (Fig 6, upper left) would display propertles slmllar to that described for the R6G-PR hydrogel sensor (above) The use of FRET as the pH-to-hfetlme transduction mechamsm provides a straightforward path for using other hght or laser sources If desired,

18.5

JR Lokowxz et al /Anal. Chm. Acta 272 (1993) 179-186

Laser diodes are desirable sources for clinical mstrumentatlon because of then slmphaty, rehabrhty and low cost Also, laser diodes can be electronically modulated which elmunates the need for an external modulator Designer probes such as those shown m Fig 6 could have wide apphcatlons m analytical chemistry, clinical chemistry and blomedlcal research

the R6G donor could be replaced wth a ruthemum (Ru) or lanthamde complex These could be excited with electrolummescent lamp [lS] or flash lamp sources because of their long decay tunes The long Ru or lanthamde decay tunes could be measured after decay of the prompt autofluorescence, as 1s now done m the so-called “tlme-resolved nnmunoassays” [31] The Ru complex would be sensltlve to oxygen [32,33], but the lanthamdes are not quenched by oxygen Finally, the use of FRET allows selection of the donor for use with red 633-nm He-Ne or laser diode sources, as shown for the Indocyanine-Thymol Blue probe m Fig 6 (lower left)

H.

,N CZHO 1

This work was supported by a grant (DIR8710401) from the National Science Foundation and grants RR07510 and RR08119 from the National Institutes of Health

H

+, N.

0

T

CA

/A

coon

(4-H f’s y-t4

IfH,l. y-H

:=o fH’ SO,

cn2

o=c OH

H\N,

3Qo 3

;

/

,[CHzb NH

1 on

’ ,st$-

N-NwN=N

0

NH2 To=5 3ns

A,,=365nm, Eu+’

RHODAMINE

6 G -PHENOL

RED

XEX=650nm.

AE,,=670nm,

r.-1

INDOCYANINE

- THYMOL

9ATHOCUPROlN

IEx=485nm, ns

BLUE

Fig 6 Gedanken pH sensors

NH,

AEU--615

s.*3 TRIS

25j1s

nm,

T,-500@

-DISULFONYL

CONGO

RED

AE”=590nm I N,) , ~~-436

I PHENANTHROLINE

ns law I

in water

I RUTHENIUM-PHENOL

RED

186 REFERENCES 1 0 S Wolfbels, F&x-Optx

2 3

4 5 6 7 8 9 10 11 12 13 14 15 16 17

Chermcal Sensors and B~osenSOTS,Vol I, CRC Press, Boca Raton, FL, 1991 0 S Wolfbeis, F&xx-Optic Chermcal Sensors and BlosenSOTS,Vol II, CRC Press, Boca Raton, FL, 1991 R W Murray, R E Dessy, W R Hememan, .I Janata and W R Se& (Eds 1, Chermcal Sensors and Mxromstrumentation (ACS Symposmm Series, Vol 403) American Chemical Society, Washmgton, DC, 1989 LA Saarl and W R Seltz, Anal Chem ,54 (1982) 821 H Offenbacher, 0 S Wolfbels and E Furhnger, Sensors Actuators, 9 (1986) 73 0 S Wolfbels and J H Baustert, J Heterocycl Chem, 22 (1985) 1215 M-R S Fuh, L W Burgess, T Huschfeld, G D Chnstlan and F Wang, Analyst, 112 (1987) 1159 N R Clement and J M Gould, Blochemlstry, 20 (1981) 1534 Z ZhujUn and W R Seitz, Anal Chum Acta, 160 (1984) 47 D M Jordan and D R Walt, Anal Chem , 59 (1987) 437 P Yuan and D R Walt, Macromolecules, 23 (1990) 4611 P Yuan and D R Walt, Anal Chem ,59 (1987) 2391 G Gabor and D R Walt, Anal Chem , 63 (1991) 793 JR Lakowicz, H Szmacmslu and K.W Berndt, Proc SPIE Vol , 1648 (1992) KW Bemdt and JR Lakowlcz, Anal Blochem, 201 (1992) 319 K W Berndt, I Gryczynslu and J R Lakowlcz, Rev SCI Instrum, 61 (1990) 1816 R B Thompson, J K. Frlsoh and J R Lakowlcz, Anal Chem , 64 (1992) 2075

JR Lakowcz et al /Anal Chm Acta 272 (1993) 179-186 18 C Bo~arsk~ and K Slemclu, Photochemistry and Photophysux, Vol 1, CRC Press, Boca Raton, FL, 1989, pp l-57 19 JR Lakow~cz, Prmciples of Fluorescence Spectroscopy, Plenum Press, New York, 1983, pp 81-86 20 E Gratton and M Lunkeman, Blophys J , 44 (1983) 665 21 JR Lakowicz and B P Mahwal, B~ophys Chem ,21(1985) 61 22 JR Lakowicz, G Laczko and I Gryczynslu, Rev Scl Instrum, 57 (1986) 2499 23 C BoJarsb and J Domsta, Acta Phys Hung, 30 (1971) 145 24 A Kawslu and H Szmacmslu, Naturforsch Ted A, 37 (1982) 64 25 I Gryczynslu, W Wlczk, M L Johnson and J R Lakow~cz, Chem Phys Lett , 145 (1988) 439 26 J R Lakowlcz, H Szmacmslu, I Gryczynslu and W Wlczk, J Phys Chem ,94 (1990) 8413 27 G Laczko, JR Lakowcz, I Gryczynslu, Z Gryczynskl and H Malak, Rev Scl Instrum, 61(1990) 2331 28 J Elsmger and J Flores, Anal Blochem, 94 (1979) 15 29 R G Bennett, J Chem Phys , 41 (1964) 3037 30 K.B Elsenthal and S Siegel, J Chem Phys, 41 (1964) 652 31 T Lovgren and K Peterson, m K Van Dyke and R Van Dyke (Eds ), Lummescence Immunoassay and Molecular Applications, CRC Press, Boca Raton, FL, 1990, pp 233253 32 J R Bacon and J N Demas, Anal Chem , 59 (1987) 2780 33 E R Carraway, J N Demas, B A DeGaff and J R Bacon, Anal Chem , 63 (1991) 337