Biomaterials 26 (2005) 1895–1903

ARTICLE IN PRESS

*Correspondin

Boston Universi

02215, USA. Tel

Department of O

Eye Research I

W. Taylor St.,

fax:+1-312-996-

E-mail addres

davipepp@uic.ed

0142-9612/$ - see

doi:10.1016/j.bio

Activation of membrane receptors by a neurotransmitter conjugatedesigned for surface attachment

Tania Q. Vua,*, Sarwat Chowdhuryb, Niraj J. Munia,c, Haohua Qiana, RobertF. Standaertb, David R. Pepperberga,c,*

aDepartment of Ophthalmology and Visual Sciences, Lions of Illinois Eye Research Institute, University of Illinois at Chicago,

1855 W. Taylor St., Chicago, IL 60612, USAbDepartment of Chemistry, University of Illinois at Chicago, 845 W. Taylor St., Chicago, IL 60607, USA

cDepartment of Bioengineering, University of Illinois at Chicago, 851 S. Morgan St., Chicago, IL 60607, USA

Received 3 April 2004; accepted 1 June 2004

Available online 28 August 2004

Abstract

The derivatization of surfaces with bioactive molecules is a research area of growing importance for cell and tissue engineering.

Tetherable molecules used in such applications must contain an anchoring moiety as well as the biofunctional group, typically along

with a spacer to prevent steric clashes between the target molecule and the tethering surface. Post-synaptic membrane receptors at

chemical synapses in neural tissue mediate signaling to the post-synaptic neuron and are activated by the binding of diffusible

neurotransmitter molecules released by the pre-synaptic neuron. However, little attention has been directed at developing

neurotransmitter analogs that might retain functionality when tethered to a surface that could be interfaced with post-synaptic

receptor proteins. Muscimol (5-aminomethyl-3-hydroxyisoxazole), an analog of GABA (g-aminobutryic acid), is a known potent

agonist of GABAA and GABAC post-synaptic receptors found in retina and other central nervous system tissue. The present paper

reports experiments testing the electrophysiological activity of ‘‘muscimol-biotin’’ on cloned GABA receptors expressed in Xenopus

oocytes. This compound, which is potentially suitable for tethering at avidin-coated surfaces, consists of muscimol conjugated

through an N-acyl linkage to a 6-aminohexanoyl chain that is distally terminated by biotin. We find that muscimol-biotin, as well as

a structurally similar compound (muscimol-BODIPYs) containing a bulky fluorophore at the distal end of the aminohexanoyl

chain, exhibits substantial agonist activity at GABAA and GABAC receptors. Muscimol-biotin and other similarly biotinylated

neurotransmitter analogs, in combination with surface functionalization using avidin-biotin technology, may be useful in

applications involving the controlled activation of neuronal post-synaptic receptors by surface-attached molecules.

r 2004 Elsevier Ltd. All rights reserved.

Keywords: Neurotransmitter; GABA receptor; Muscimol; Biotin; BODIPYs; Nerve tissue engineering; Electrophysiology

1. Introduction

Recent advances in the fields of biomaterials andmicro-/nanotechnology have raised interest in develop-ing neuroprosthetic devices that can interact with nerve

g authors. Biomedical Engineering Department,

ty, 44 Cummington St Room 403, Boston, MA

.: +1-617-358-2832; fax: +1-617-358-2835 (T.Q. Vu);

phthalmology and Visual Sciences, Lions of Illinois

nstitute, University of Illinois at Chicago, 1855

Chicago, IL 60612, USA. Tel.:+1-312-996-4262;

7773.

ses: taniavu@bu.edu (T.Q. Vu),

u (D.R. Pepperberg).

front matter r 2004 Elsevier Ltd. All rights reserved.

materials.2004.06.007

cells under disease conditions to achieve controlledphysiological stimulation of these cells. Approachescurrently under investigation include, for example, theelectrical stimulation of retinal neurons by devicesimplanted within the eye [1–3]. Another approach,distinct from that of direct electrical stimulation, buildson the fact that signal transmission at chemical synapsesin neural tissue involves the activation of membranereceptor proteins on the post-synaptic neuron bychemical neurotransmitter released by the pre-synapticcell. This latter approach envisions nerve cell stimula-tion through the controlled presentation of a receptor-active molecule (neurotransmitter or analog) to post-synaptic receptors and a resulting controlled stimulation

ARTICLE IN PRESS

Fig. 1. Chemical structures of muscimol, muscimol-BODIPY and

muscimol-biotin.

T.Q. Vu et al. / Biomaterials 26 (2005) 1895–19031896

of the post-synaptic neuron. Studies aimed at interfacingneurotransmitter-releasing devices with nerve tissue alsohave been reported [4,5], but this ‘‘chemical’’ approachto nerve cell stimulation remains largely unexplored.Within this latter approach, the availability of neuro-transmitter analogs capable of being tethered to asupporting surface could be useful in the development ofmicro-/nanostructures as neuroprosthetics [e.g., 6] aswell as in advancing fundamental understanding ofintercellular signaling in neural tissue.

The long-term goal of the research reported in thispaper is to develop molecular platforms containingsurface-tethered neurotransmitter analogs that caninteract in controlled and specific fashion with nativepost-synaptic receptors. As the focus of this work wehave chosen types A and C receptors for the neuro-transmitter g-aminobutyric acid (GABA). GABA is themajor inhibitory neurotransmitter in CNS tissue, andGABAA and GABAC receptors, which gate transmem-brane chloride channels, occur in high abundance inneurons such as bipolar cells of the retina [7–10]. Theability to modulate GABAA or GABAC receptoractivity in retinal bipolar cells thus could be useful, forexample, in the development of neuroprosthetic struc-tures for application as a therapy in retinal degenerativedisease.

A key initial step toward developing surfaces with thedesired receptor-modulating property is the identifica-tion of neurotransmitter analogs that both exhibitphysiological activity and are potentially suitable forsurface attachment. A logical approach to achievingtetherability of a known receptor-active compound is toderivatize this compound with a spacer chain (to preventsteric interference between the target receptor and theultimate tethering surface) that is distally terminated by afunctional group capable of mediating surface attach-ment. However, unlike bioactive molecules of relativelyhigh molecular weight (e.g., peptides), neurotransmittersin retina and other CNS tissue are small amino acids(e.g., glutamate and GABA) that possess few sites forpotential conjugation with a tethering moiety. Further-more, even simple alteration of a small-molecule receptorligand (e.g., derivatization of a carboxylic or methylgroup) can dramatically reduce the ligand’s affinity forits target receptor [e.g., 11,12]. Thus, chain-derivatizationof amino acid neurotransmitters and analogs is unlikelyas a general rule to preserve substantial biofunctionalityof the receptor-active parent compound.

Muscimol (5-aminomethyl-3-hydroxyisoxazole) (Fig. 1)is a well-studied agonist of GABAA and GABAC

receptors and exhibits a potency comparable with thatof GABA [11–13]. Recently, two studies have presentedevidence that N-acyl derivatives of muscimol can bind toGABA receptors. First, Wang et al. [14] have shownthat muscimol linked through a 6-aminohexanoyl(6-aminocaproyl) chain to a BODIPYs fluorophore

(here termed ‘‘muscimol-BODIPY’’; Fig. 1) exhibitslocalization to GABAA receptors in retinal neurons.Second, Meissner and H.aberlein [15] have found thatmuscimol conjugated directly (i.e., without an alkylchain) to the fluorophore Alexa Fluor 532 shows highaffinity for GABAA receptors of rat hippocampalneurons. Overall, however, only a few N-acylatedmuscimol derivatives have been described in theliterature [14–16], and none of these studies haveinvestigated the electrophysiological activities of thesederivatives.

In light of the above-summarized results obtainedwith N-acyl muscimol derivatives, and of the wellestablished suitability of avidin-biotin technology formicropatterning and surface functionalization [e.g., 6,17–19], we reasoned that an N-acyl derivative ofmuscimol containing an aminohexanoyl chain termi-nated by a biotin group (‘‘muscimol-biotin’’; Fig. 1)would be of interest to investigate as a candidate agonistat GABAA and GABAC receptors. Here we report thesynthesis of muscimol-biotin and the testing of itselectrophysiological activity at cloned GABAA andGABAC receptors expressed in Xenopus oocytes. TheXenopus oocyte provides a useful model system for thestudy of GABA receptor activity, as the oocytemembrane lacks endogenous GABA receptors [20]. Wefind that muscimol-biotin, as well as muscimol-BODI-PY, exhibits substantial agonist activity at both GABAA

and GABAC receptors. These findings represent, to ourknowledge, the first demonstration of electrophysiolo-gical activity by N-acyl derivatives of muscimol.Preliminary results of this study have been reported[21,22].

2. Materials and methods

2.1. Materials

Muscimol-BODIPY (muscimol, BODIPYs TMR-Xconjugate, product M23400) was purchased from

ARTICLE IN PRESST.Q. Vu et al. / Biomaterials 26 (2005) 1895–1903 1897

Molecular Probes (Eugene, OR) and stored at �20�C.Stock solutions of muscimol-BODIPY (typically, 2.8–10mm) in dimethyl sulfoxide (DMSO) were stored at3�C. Muscimol, biotinamidocaproic acid N-hydroxy-succinimide ester (BAC-NHS), diisopropylethylamine(DIPEA), (1,2,5,6-tetrahydropyridin-4-yl)methylpho-sphinic acid (TPMPA), N-(2-hydroxyethyl)piperazine-N0-(2-ethanesulfonic acid) (HEPES) and bicucullinewere obtained from Sigma (St. Louis, MO); trifluor-oacetic acid (TFA) from Acros (Pittsburgh, PA);acetonitrile (Omnisolvs Cat. No. EM-AX0142-1) fromVWR (West Chester, PA); and N-methylpyrrolidinone(NMP) from Advanced Chemtech (Louisville, KY). Allother chemicals and solvents were reagent grade andwere used without further purification.

2.2. Synthesis of muscimol-biotin

BAC-NHS (19.9mg, 44 mmol) and DIPEA (7.6 ml,44 mmol, 1.0 equiv.) were added to a solution ofmuscimol (5.0mg, 44 mmol) in 1.25ml of NMP. Thereaction mixture was stirred at room temperature for18 h and quenched with TFA (3.4 ml, 44 mmol). Theproduct was purified by reversed-phase HPLC on aWaters Delta-Pak C18 column (25� 100mm) at a flowrate of 10ml/min using a 25-min linear gradient of 0–40% acetonitrile (0.08% TFA) in water (0.1% TFA).Peaks were detected by absorbance at 210 nm; theproduct-containing peak was collected and lyophilizedto afford 16.2mg (81% yield) of muscimol-biotin.Spectral data (1H NMR, 13C NMR and mass spectro-metry) for the HPLC-purified product are as follows: 1HNMR (400MHz, DMSO-d6) d 8.35 (d, 1H, J ¼ 5:6Hz),7.72 (s, 1H), 6.41 (s, 1H), 5.78 (s, 1H), 4.60 (br s, 3H),4.30 (dd, 1H, J ¼ 3:1; 4.5Hz), 4.23 (d, 2H, J ¼ 5:6Hz),4.12 (dd, 1H, J ¼ 3:4; 4.3Hz), 3.09 (m, 1H), 3.00 (dd,2H, J ¼ 5:8; 6.4Hz), 2.82 (dd, 1H, J ¼ 4:9; 7.5Hz), 2.57(d, 1H, J ¼ 12Hz), 2.09 (t, 2H, J ¼ 7:5Hz), 2.04 (t, 2H,J ¼ 7:3Hz), 1.62–1.56 (m, 1H), 1.51–1.48 (m, 5H), 1.42–1.33 (m, 2H), 1.33–1.27 (m, 1H), 1.27–1.17 (m, 2H); 13CNMR (100MHz, DMSO-d6) d 172.73, 172.27, 171.18,170.68, 163.16, 93.70, 61.50, 59.65, 55.86, 38.73, 35.66,35.50, 35.32, 29.39, 28.65, 28.47, 26.53, 25.76, 25.32;HRMS (ESI) m/z calcd for C20H32N5O5S 454.2124,found for (M+H)+ 454.2139; (3.3 ppm error). Stocksolutions of muscimol-biotin (up to B50mm in DMSO)were stored at 3 or �20�C.

2.3. Oocyte preparation and receptor expression

All animal procedures adhered to institutional poli-cies and to the Statement for the Use of Animals inOphthalmic and Vision Research adopted by theAssociation for Research in Vision and Ophthalmology(ARVO). Gravid adult female Xenopus laevis toads(Xenopus One, Ann Arbor, MI) were anaesthetized with

MS-222 (1 g/l), their ovarian lobes excised, and theoocytes removed. Stages V–VI oocytes were selected andstored in frog Ringer solution (i.e., physiological saline)that consisted of 100mm NaCl, 2mm KCl, 2mm CaCl2,1mm MgCl2, 10mm glucose and 5mm HEPES, at pH7.4. To remove the follicular layer, oocytes wereimmersed for 30min at room temperature in Ca2+-freeRinger solution containing 2mg/ml collagenase. Fol-lowing mechanical trituration, each oocyte was injected(Drummond Nanoject II; Drummond Scientific Co.,Broomall, PA) with 50 nl cRNA (0.5mg/ml) and storedin Ringer solution containing 0.1mg/ml gentamycin at19�C. Heteropentameric GABAA receptors (a1b2g2) andhomopentameric GABAC receptors (human r1)were cloned in the pSP64T expression vector andexpressed in the oocytes using procedures similar tothose described previously [23,24]. cRNAs for GABAreceptor subunits were transcribed in vitro fromlinearized cDNAs using the mMessage mMachine(Ambion Inc., Austin, TX).

2.4. Electrical recording

Electrophysiological experiments were conducted onGABAA- and GABAC-expressing oocytes 1–3 days afteroocyte injection. Recordings were obtained at roomtemperature using a two-microelectrode voltage clampamplifier (GeneClamp 500B, Axon Instruments, Inc,Foster City, CA) and low-pass filtered at 10Hz. Cellswere voltage-clamped at �70mV. The pipette-to-bathresistance was typically 2–6MO. Data were sampled ateither 100 or 200Hz by a DIGIDATA 1322A A/Dboard (Axon Instruments) using pCLAMP v.8 software(Axon Instruments, Union City, CA). Test solutions ofmuscimol-biotin and of muscimol-BODIPY (i.e., thesolutions presented to the receptor-expressing oocytes)were prepared from their respective DMSO stocksolutions by dilution into frog Ringer. Unless otherwiseindicated, the amount of carrier DMSO contained in thetest solutions was 1% (v/v) or less. Test solutions wereintroduced into the recording chamber through agravity flow system at a rate of approximately 1ml/min, as controlled by a solenoid manifold operatedunder computer or manual command. The perfusatewas removed from the recording chamber by vacuumsuction through a large bore glass pipette placed on theother side of the cell. Individual oocytes exhibiteddifferent levels of GABA receptor expression, andthis resulted in variations in the amplitude of theresponse elicited by a given concentration of GABAor other test compound. However, functional proper-ties of the receptor such as the relative sensitivitiesand kinetics of responses to differing test com-pounds were largely unaffected by the level of receptorexpression.

ARTICLE IN PRESST.Q. Vu et al. / Biomaterials 26 (2005) 1895–19031898

2.5. Data collection and analysis

Origin 6.0 (OriginLab Corp., Northampton, MA)software was used for data analysis and graphing. Withthe use of a minimum chi-square algorithm, dose–response data were analyzed to obtain best-fit para-meters for the Hill equation:

R=Rmax ¼ cn=ðcn þ ECn50Þ: ð1Þ

Here, R is the peak amplitude of the response elicited bythe test compound at concentration c; Rmax is themaximum peak amplitude; n is the Hill coefficient; andEC50 is the concentration of test compound that elicits ahalf-maximal response.

3. Results

3.1. Muscimol-BODIPY

The present study was motivated by the initial findingthat commercially obtained muscimol-BODIPY, acompound in which an aminohexanoyl spacer linksmuscimol to a BODIPYs fluorophore (Fig. 1), exhibitselectrophysiological activity at GABA receptors. Anexample of this activity is illustrated by Fig. 2A, whichshows responses recorded from a single oocyte expres-sing GABAA receptors. The three illustrated responseswere obtained on treatment of the oocyte with 20 mmmuscimol-BODIPY, 5 mm muscimol and 10 mm GABA,respectively. The peak amplitude of the response to20 mm muscimol-BODIPY was comparable with those ofresponses elicited by 5 mm muscimol or 10 mm GABA.

Fig. 2. Activity of muscimol-BODIPY on GABAA and GABAC receptors

GABA (10mm), muscimol (5mm), and muscimol-BODIPY (20mm). (B) Respon

the co-application of 10mm muscimol-BODIPY and 100mm bicuculline. (C)

10mm GABA, 10mm muscimol and 100mm muscimol-BODIPY. Note diffe

GABAC-expressing oocyte to the application of 10 mm muscimol-BODIPY al

bicuculline. (E) Response of a GABAC-expressing oocyte to 50 mm muscimol-

200mm TPMPA.

Evidence that the response to muscimol-BODIPYrecorded from GABAA-expressing oocytes wasmediated specifically by GABAA receptors came fromresults obtained with bicuculline, a competitive antago-nist selective for GABAA receptors [11,12]. As shown inFig. 2B, the co-application of 100 mm bicuculline and10 mm muscimol-BODIPY blocked almost all of theresponse elicited from a GABAA-expressing oocyte by10 mm muscimol-BODIPY alone.

Muscimol-BODIPY also elicited large responses fromoocytes expressing GABAC receptors. For example, asshown in Fig. 2C, the response elicited by 100 mmmuscimol-BODIPY exhibited a peak amplitude of ordersimilar to the near-saturating peak amplitudes ofresponses to 10 mm GABA and 10 mm muscimol (cf. 25;and see below). The relatively slow response kineticsexhibited by GABAC receptors facilitates analysis of therecovery phase of the GABAC-mediated response. Asillustrated in Fig. 2C, the responses to both muscimoland muscimol-BODIPY (as well as the response toGABA) display qualitatively similar (i.e., approximatelyexponential) recovery kinetics, but the decay of themuscimol-BODIPY-elicited response requires a periodconsiderably longer than that for the decay of theresponse to muscimol. Determinations of the exponen-tial decay constant t characterizing the recovery ofresponses to muscimol and muscimol-BODIPY re-corded from GABAC-expressing oocytes (Table 1, rows1–2) showed that t for muscimol-BODIPY-elicitedresponses exceeded that for muscimol-elicited responsesby more than 3-fold on average. The GABAC receptor isinsensitive to bicuculline [11–13] and, as expected,the co-application of 100 mm bicuculline with 10 mm

. (A) Responses elicited from a single GABAA-expressing oocyte by

se of a GABAA-expressing oocyte to 10mm muscimol-BODIPY, and to

Continuous recording of responses of a GABAC-expressing oocyte to

rences in the recovery kinetics of the responses. (D) Responses of a

one, and to the co-application of 10 mm muscimol-BODIPY and 100mmBODIPY alone and to co-application of 50mm muscimol-BODIPY and

ARTICLE IN PRESS

Table 1

Exponential time constants (t) describing the recovery kinetics of

responses to muscimol, muscimol-BODIPY and muscimol-biotin

recorded from GABAC-expressing oocytes

Test compound Number of trialsa t (s)b

Muscimol (10mm) 22 4.3470.31

Muscimol-BODIPY (100 mm) 13 14.9470.43

Muscimol-biotin (200 or 500mm) 14 14.5670.82

aThe analyzed responses to muscimol, muscimol-BODIPY and

muscimol-biotin were obtained, respectively, from 7, 6 and 4 oocytes.bValues of t obtained by fitting an exponential function [expð�t=tÞ]

to the recovery phase of the responses. Each quoted value is the

mean7SEM for results obtained from the indicated number of trials.

Fig. 3. Preparative, reversed-phase HPLC chromatogram illustrating

the isolation of muscimol-biotin from the reaction mixture (see Section

2). Peak 1 contains N-hydroxysuccinimide, unreacted muscimol, and

hydrolyzed BAC-NHS; peak 2 contains N-methylpyrrolidinone; and

peak 3 contains muscimol-biotin.

T.Q. Vu et al. / Biomaterials 26 (2005) 1895–1903 1899

muscimol-BODIPY to a GABAC-expressing oocyte hadlittle effect on the response elicited by 10 mm muscimol-BODIPY alone (Fig. 2D). However, the response tomuscimol-BODIPY was suppressible by TPMPA, aspecific GABAC receptor antagonist [26] (Fig. 2E).Together, the Fig. 2 results indicate that muscimol-BODIPY is capable of binding to and activating bothGABAA and GABAC receptors.

3.2. Muscimol-biotin

The results obtained with muscimol-BODIPY raisedthe possibility that an analogous conjugate of muscimolterminated with a biotin moiety might also exhibitactivity at GABA receptors. To test this possibility, weprepared ‘‘muscimol-biotin’’ (Fig. 1), a compoundcontaining the same N-acyl attachment and aminohex-anoyl linker as those employed in muscimol-BODIPY,but with the distal end of the aminohexanoyl chainlinked to biotin. Fig. 3 presents a representative,preparative-scale, reversed-phase HPLC chromatogramof the isolation of the muscimol-biotin product, showingclean separation of the product (peak 3) from muscimoland other relatively polar components of the reactionmixture (peaks 1 and 2). The HPLC-purified productwas judged to be 97% pure by 1H NMR spectroscopy.Free muscimol has characteristic resonances in the 1HNMR spectrum at d 3.65 for the CH2 group and d 5.83for the CH group (d 4.14 and d 6.14, respectively, in thepresence of TFA); these peaks were not detected in theproduct (limit of detection: ca. 1%).

Fig. 4 shows results obtained with muscimol-biotin inGABAA-expressing oocytes. Panel A of the figure showsa representative response obtained with the applicationof 100 mm muscimol-biotin. The peak amplitude of thisresponse was comparable with that elicited from thesame oocyte by 10 mm GABA. Furthermore, as illu-strated in Fig. 4B, the activity of 2.5 mm muscimol-biotinat GABAA receptors was suppressed by the co-applica-tion of 100 mm bicuculline with 2.5 mm muscimol-biotin.Fig. 4C shows a family of responses recorded from a

single oocyte upon the presentation of varying concen-trations of muscimol-biotin. Increasing the muscimol-biotin concentration from 50 mm to 1mm increased thepeak amplitude and rate of rise of the response, andlengthened the period of response recovery. Fig. 4Dshows normalized peak amplitudes of responses re-corded from a group of GABAA-expressing oocytes,upon the presentation of varying concentrations ofmuscimol (open circles) and muscimol-biotin (filledcircles). Fitting the Hill equation to each set of datayielded the curves shown in the Fig. 4D. The dose–response function determined for muscimol-biotinexhibits a coefficient n ¼ 1:4; a value higher than thatdetermined for muscimol (n ¼ 0:74). In addition, theEC50 value determined for muscimol-biotin (385 mm)exceeded that determined for muscimol (4.8 mm) byB80-fold.

Fig. 5 shows results obtained with muscimol-biotin inGABAC-expressing oocytes. Panel A shows representa-tive responses obtained with applications of 10 mmmuscimol and 500 mm muscimol-biotin. These tworesponses exhibit similar peak amplitudes, but differconsiderably with respect to their recovery kinetics. Asillustrated by Table 1, the exponential time constant ofrecovery determined for muscimol-biotin-elicited re-sponses, like that determined for muscimol-BODIPY-elicited responses, greatly exceeded that determined forresponses to muscimol. Moreover, as observed formuscimol-BODIPY (Fig. 2E), the response of GA-BAC-expressing oocytes to muscimol-biotin was sup-pressible by TPMPA (Fig. 5B). Fig. 5C showsrepresentative responses of a single GABAC-expressingoocyte to varying concentrations of muscimol andmuscimol-biotin, and Fig. 5D shows normalized peakamplitudes of responses obtained with similar testing ofa group of oocytes. Fitting the Hill equation (Eq. (1)) tothe Fig. 5D data for muscimol (open circles) and

ARTICLE IN PRESS

Fig. 4. Activity of muscimol-biotin on GABAA receptors. (A) Responses of a single GABAA-expressing oocyte to 100mm muscimol-biotin and 10mmGABA. (B) Responses of another oocyte to 2.5mm muscimol-biotin alone, and to the co-application of 2.5mm muscimol-biotin and 100mmbicuculline. (C) Family of responses to varying concentrations of muscimol-biotin and to a single, saturating concentration of muscimol (200 mm,thick trace) recorded from a single GABAA-expressing oocyte. Here and in panel D, the concentrations of carrier DMSO in the applied test solutions

of muscimol-biotin ranged up to 10% (v/v). Labels indicate the concentration of muscimol-biotin applied. (D) Normalized peak amplitudes

(mean7SEM) of responses recorded from GABAA-expressing oocytes (n ¼ 9) upon the application of muscimol (open circles) and muscimol-biotin

(filled circles). Peak amplitudes of all responses obtained from a given oocyte are normalized to the peak amplitude of the saturating response to

muscimol. Curves plot the Hill equation with n ¼ 0:74 and EC50=4.8mm for muscimol; and n ¼ 1:4 and EC50=385mm for muscimol-biotin.

T.Q. Vu et al. / Biomaterials 26 (2005) 1895–19031900

muscimol-biotin (filled circles) yielded n ¼ 1:2 andEC50=2.0 mm for muscimol, and n ¼ 4:4 andEC50=20 mm for muscimol-biotin.

4. Discussion

In this study we have examined the electrophysiolo-gical activity of muscimol-biotin, a chain-derivatizedform of muscimol containing a terminating biotingroup, at GABAA and GABAC receptors expressed inXenopus oocytes. The data of Figs. 4 and 5 show that forboth receptor types, the peak amplitude of themembrane current response elicited by muscimol-biotinincreases with the concentration of this compound, andthe maximal peak amplitude of the response is compar-able with those of responses elicited by muscimol.Furthermore, the response to muscimol-biotin is sup-pressed by known blockers of GABAA and GABAC

receptors. From these findings we conclude thatmuscimol-biotin exhibits agonist activity at GABAA

and GABAC receptors, albeit with a sensitivity con-siderably less than (i.e., an EC50 value considerablygreater than) that of muscimol. To our knowledge, this

is the first demonstration of the activation of aneurotransmitter membrane receptor by a conjugate ofa neurotransmitter analog designed for tethering. Anearlier study of a biotinylated derivative of GABAindicated that the conjugation of GABA with biotinthrough a 5-amino-3-oxapentyl chain preserves theability of GABA to interact with anti-GABA antibody[6]; however this biotinylated GABA lacked significantelectrophysiological activity at GABAC receptors at1mm concentration (unpublished observations).

The evident activity of muscimol-biotin at GABAC

receptors is noteworthy in two respects. First, the dose–response curve for muscimol-biotin at GABAC recep-tors exhibits a Hill coefficient of 4.4, a value consider-ably exceeding the value of 2 or less observed formuscimol and GABA [24]. Homomeric GABAC recep-tors are believed to exist as pentamers with five GABAbinding sites (one at the interface of each pair ofsubunits) and to require the simultaneous binding of atleast three GABA molecules for receptor activation [25].The high Hill coefficient observed for the homomericGABAC receptors in the experiments with muscimol-biotin suggests the possibility that the stabilization offull channel opening by this biotinylated compound

ARTICLE IN PRESS

Fig. 5. Activity of muscimol-biotin on GABAC receptors: (A) responses to 10 mm muscimol and 500mm muscimol-biotin recorded from a single

GABAC-expressing oocyte; (B) response of a single GABAC-expressing oocyte to 50mm muscimol-biotin and to the co-application of 50 mmmuscimol-biotin and 200mm TPMPA; (C) responses recorded from a single GABAC-expressing oocyte upon the presentation of varying

concentrations of muscimol (upper) and muscimol-biotin (lower). Here and in panel D, the concentrations of carrier DMSO in the applied test

solutions of muscimol-biotin ranged up to 10% (v/v). (D) Normalized peak amplitudes (mean7SEM) of responses to muscimol and muscimol-biotin

obtained from GABAC-expressing oocytes (n ¼ 5 for muscimol; n ¼ 6 for muscimol-biotin). Peak amplitudes of all responses obtained from a given

oocyte are normalized to the peak amplitude of the saturating response to muscimol. The fitted curves plot the Hill equation with EC50=2.0mm and

n ¼ 1:2 for muscimol; and EC50=20mm and n ¼ 4:4 for muscimol-biotin.

T.Q. Vu et al. / Biomaterials 26 (2005) 1895–1903 1901

requires the binding of as many as 4 or 5 muscimol-biotin molecules to the receptor [25,27]. The secondconsideration relates to the ionic state of the molecule. Itis conventionally assumed that GABA receptor ligandsbind in zwitterionic form, i.e., a form in which the aminegroup is protonated (positively charged) and a hydroxylgroup is deprotonated (negatively charged) [28].However, muscimol-biotin is not expected to bezwitterionic due to the reduced basicity of the acylatednitrogen and its expected resulting neutrality at physio-logical pH. Muscimol-biotin thus joins a small group offunctional GABA analogs that are non-zwitterionic[28]. Our results demonstrating the activities ofmuscimol-biotin and of muscimol-BODIPY (also ex-pected to be non-zwitterionic at physiological pH)are therefore significant and suggest that it is possibleto compensate for loss of the corresponding positivecharge in muscimol. It is interesting to note in thisregard that the simple analog N-acetylmuscimol wasfound to be inactive as an inhibitor of GABA uptake inrat brain slices, where muscimol was a potent inhibitor[16]. Thus, the nature of the acyl substituent is clearly

important, and the complex acyl substituents inmuscimol-BODIPY and muscimol-biotin may provide,for example, fortuitous hydrophobic or hydrogenbonding interactions with (as yet unknown) parts ofthe receptor that a simple acetyl group cannot. GABAand other amino acid neurotransmitters are smallmolecules that contain at most a few functional groupsamenable to modification and thus present constraintson potential routes for bioconjugation. Our resultssuggest that it is possible to compensate for themodification of critical groups (such as the amine inmuscimol) and, if this finding extends to other neuro-transmitters, it could provide a useful strategy forpreparing conjugates of other amino acid neurotrans-mitters.

The present investigation specifically of a biotinylatedtest compound was motivated by the demonstratedworkability, in other systems, of the surface tethering ofbiomolecules using the biotin-avidin interaction [e.g.,6,17–19,29,30]. The present findings with muscimol-biotin represent what we believe is a significant step inefforts to develop surface-tetherable compounds that

ARTICLE IN PRESST.Q. Vu et al. / Biomaterials 26 (2005) 1895–19031902

retain bioactivity. However, it remains to be determinedwhether the surface attachment of muscimol-biotin willpreserve its physiological activity at GABA receptors.Aminohexanoyl linkers are commonly included in biotinconjugates to separate the conjugated molecule frombiotin, as previous studies have shown that biotin’sbinding to avidin buries the entire biotin moleculeincluding its pentanoic acid side chain [29,30]. Theeffective ‘‘depth’’ at which muscimol is buried within theGABA receptor during its binding at the receptor’sligand-binding site(s) is not known precisely (however,see [31–33]), and this depth, if exceeding the length ofthe aminohexanoyl linker in the present muscimol-biotin, would presumably preclude GABA receptoractivation by avidin-tethered muscimol-biotin. Anotherpotentially complicating issue concerns the orientationof the muscimol moiety. That is, the evidence that theactivation of pentameric GABAA and GABAC receptorsrequires the simultaneous binding of two or moreligands [25] suggests that receptor activation by musci-mol-biotin may involve specific orientational require-ments for the presumably multiple molecules ofmuscimol-biotin that achieve this activation.

5. Conclusion

As an approach to developing surfaces functionalizedwith neurotransmitter analogs, we synthesized musci-mol-biotin, an N-acyl, chain-derivatized form of theGABA analog muscimol, and tested this biotinylatedcompound for electrophysiological activity at GABAA

and GABAC receptors expressed in Xenopus oocytes.The results obtained with muscimol-biotin, and with asimilarly N-acylated, commercially obtained muscimolderivative (muscimol-BODIPY) indicate that bothcompounds exhibit agonist activity, and motivatefurther work to test the activity of a surface-tetheredform of the biotinylated compound. We hypothesizethat it may be possible to engineer such functionalizedsurfaces to achieve controllable stimulation of nativepost-synaptic receptors on neural cells.

Acknowledgements

We thank Ms. Wen Wang for her expert technicalassistance with oocyte injection and voltage clamprecording. This research was supported by NIH grantsEY13693 and EY01792, by a grant from the Universityof Illinois Intercampus Research Initiative in Biotech-nology (IRIB) program, and by an unrestricted awardfrom Research to Prevent Blindness (New York, NY).DRP is a Senior Scientific Investigator of Research toPrevent Blindness.

References

[1] Rizzo III JF, Wyatt J, Humayun M, de Juan E, Liu W, Chow A,

Eckmiller R, Zrenner E, Yagi T, Abrams G. Retinal prosthesis: an

encouraging first decade with major challenges ahead. Ophthal-

mology 2001;108:13–4.

[2] Zrenner E. Will retinal implants restore vision? Science

2002;295:1022–5.

[3] Humayun MS, Weiland JD, Fujii GY, Greenberg R, Williamson

R, Little J, Mech B, Cimmarusti V, Van Boemel G, Dagnelie G,

de Juan E. Visual perception in a blind subject with a chronic

microelectronic retinal prosthesis. Vis Res 2003;43:2573–81.

[4] Peterman MC, Bloom DM, Lee C, Bent SF, Marmor MF,

Blumenkranz MS, Fishman HA. Localized neurotransmitter

release for use in a prototype retinal interface. Invest Ophthalmol

Vis Sci 2003;44:3144–9.

[5] Safadi MR, Washko F, Lagman A, Jaboro C, Auner GW, Iezzi

R, McAllister JP, Abrams G. Development of a microfluidic drug

delivery stimulating device for vision. 2003 Annual Meeting of the

Association for Research in Vision and Ophthalmology, abstract

5082.

[6] Saifuddin U, Vu TQ, Rezac M, Qian H, Pepperberg DR, Desai

TA. Assembly and characterization of biofunctional neurotrans-

mitter-immobilized surfaces for interaction with postsynaptic

membrane receptors. J Biomed Mater Res 2003;66A:184–91.

[7] Yazulla S. GABAergic mechanisms in the retina. Progr Retinal

Res 1986;5:1–52.

[8] Wu SM. Functional organization of GABAergic circuitry in

ectotherm retinas. Progr Brain Res 1992;90:93–106.

[9] W.assle H, Koulen P, Brandst.atter JH, Fletcher EL, Becker CM.

Glycine and GABA receptors in the mammalian retina. Vis Res

1998;38:1411–30.

[10] Lukasiewicz PD, Shields CR. A diversity of GABA receptors in

the retina. Sem Cell Dev Biol 1998;9:293–9.

[11] Chebib M, Johnston GAR. GABA-activated ligand gated ion

channels: medicinal chemistry and molecular biology. J Med

Chem 2000;43:1427–47.

[12] Krogsgaard-Larsen P, Frolund B, Ebert B. GABAA receptor

agonists, partial agonists, and antagonists. In: Enna SJ, Bowery

NG, editors. The GABA receptors. Totowa, NJ: Humana Press;

1997. p. 31–81.

[13] Johnston GAR. GABAC receptors: relatively simple transmitter-

gated ion channels? Trends Pharmacol Sci 1996;17:319–23.

[14] Wang H, Standifer KM, Sherry DM. GABA(A) receptor binding

and localization in the tiger salamander retina. Visual Neurosci

2000;17:11–21.

[15] Meissner O, H.aberlein H. Lateral mobility and specific binding to

GABAA receptors on hippocampal neurons monitored by

fluorescence correlation spectroscopy. Biochemistry 2003;42:

1667–72.

[16] Krogsgaard-Larsen P, Johnston GAR. Inhibition of GABA

uptake in rat brain slices by nipecotic acid, isoxazoles, and

related compounds. J Neurochem 1975;25:797–802.

[17] Bashir R, Gomez R, Sarikaya A, Ladisch MR, Sturgis J, Robinson

JP. Adsorption of avidin on microfabricated surfaces for protein

biochip applications. Biotechnol Bioeng 2001;73:324–8.

[18] Orth RN, Kameoka J, Zipfel WR, Ilic B, Webb WW, Clark TG,

Craighead HG. Creating biological membranes on the micron

scale: forming patterned lipid bilayers using a polymer lift-off

technique. Biophys J 2003;85:3066–73.

[19] Gref R, Couvreur P, Barratt G, Mysiakine E. Surface-engineered

nanoparticles for multiple ligand coupling. Biomaterials 2003;

24:4529–37.

[20] Polenzani L, Woodward RM, Miledi R. Expression of mamma-

lian g-aminobutyric acid receptors with distinct pharmacology in

Xenopus oocytes. Proc Natl Acad Sci USA 1991;88:4318–22.

ARTICLE IN PRESST.Q. Vu et al. / Biomaterials 26 (2005) 1895–1903 1903

[21] Vu TQ, Qian H, Saifuddin U, Rezac M, Desai TA, Pepperberg

DR. Toward development of neurotransmitter-derivatized sur-

faces for interaction with post-synaptic membrane receptors. 2003

Annual Meeting of the Association for Research in Vision and

Ophthalmology, abstract 1064.

[22] Pepperberg DR, Vu TQ, Chowdhury S, Standaert RF, Qian H.

GABA receptor activation by a tetherable analog of muscimol for

application in a neuromodulating molecular device. 2004 Annual

Meeting of the Association for Research in Vision and

Ophthalmology, abstract 4196.

[23] Qian H, Hyatt G, Schanzer A, Hazra R, Hackam AS, Cutting

GR, Dowling JE. A comparison of GABAC and r subunit

receptors from the white perch retina. Visual Neurosci 1997;14:

843–51.

[24] Qian H, Dowling JE, Ripps H. Molecular and pharmacological

properties of GABA-r subunits from white perch retina. J

Neurobiol 1998;37:305–20.

[25] Amin J, Weiss DS. Insights into the activation mechanism of r1GABA receptors obtained by coexpression of wild type and

activation-impaired subunits. Proc Roy Soc London B

1996;263:273–82.

[26] Ragozzino D, Woodward RM, Murata Y, Eusebi F, Overman

LE, Miledi R. Design and in vitro pharmacology of a selective

g-aminobutyric acid C receptor antagonist. Mol Pharmacol

1996;50:1024–30.

[27] Colquhoun D. Binding, gating, affinity and efficacy: the interpreta-

tion of structure-activity relationships for agonists and of the effects

of mutating receptors. Brit J Pharmacol 1998;125:924–47.

[28] Carlier PR, Chow ES, Barlow RL, Bloomquist JR. Discovery of

non-zwitterionic GABAA receptor full agonists and a super-

agonist. Bioorg Med Chem Lett 2002;12:1985–8.

[29] Green NM, Konieczny L, Toms EJ, Valentine RC. The use of

bifunctional biotinyl compounds to determine the arrangement of

subunits in avidin. Biochem J 1971;125:781–91.

[30] Pazy Y, Kulik T, Bayer EA, Wilchek M, Livnah O. Ligand

exchange between proteins. Exchange of biotin and biotin

derivatives between avidin and streptavidin. J Biol Chem

2002;277:30892–900.

[31] Ernst M, Brauchart D, Boresch S, Sieghart W. Comparative

modeling of GABAA receptors: limits, insights, future develop-

ments. Neuroscience 2003;119:933–43.

[32] Miyazawa A, Fujiyoshi Y, Stowell M, Unwin N. Nicotinic

acetylcholine receptor at 4.6 (A resolution: transverse tunnels in

the channel wall. J Mol Biol 1999;288:765–86.

[33] Karlin A. Emerging structure of the nicotinic acetylcholine

receptors. Nature Rev Neurosci 2002;3:102–14.