Interaction of fluorescently substituted metallacarboranes with cyclodextrins and phospholipid bilayers: fluorescence and light scattering study.

Abstract

We prepared two fluorescein-[3-cobalt(III) bis(1,2-dicarbollide)](-) conjugates. They are sparingly soluble in water and form large aggregates in aqueous solutions. An extensive study on their spectral and aggregation behavior was carried out. To prepare their well-defined dispersion in aqueous systems, we studied the interaction of both probes with two biocompatible amphiphilic systems, cyclodextrins, which are frequently used in drug-delivery systems, and phospholipid membranes, which are the major constituents of cell barriers in living organisms. The presence of fluorescein in both conjugates allows us to study their behavior in detail by steady-state and time-resolved fluorometry, fluorescence correlation spectroscopy, and fluorescence lifetime imaging. The self-assembly of these metallacarboranes in aqueous solutions was studied by dynamic light scattering. The study shows that the compounds interact with cyclodextrins that increases their solubility in water, and they solubilize easily in phospholipid bilayers.

Full text

Page 1

6268
DOI: 10.1021/la904047kLangmuir 2010, 26(9), 6268–6275Published on Web 01/19/2010
pubs.acs.org/Langmuir
©2010 American Chemical Society
Interaction of Fluorescently Substituted Metallacarboranes
with Cyclodextrins and Phospholipid Bilayers: Fluorescence and Light
Scattering Study
Mariusz Uchman,†Piotr Jurkiewicz,‡Petr Cı´gler,§Bohumı´r Gr€ uner,)
Karel Proch? azka,†and Pavel Mat? ejı´? cek*,†
Martin Hof,‡
†DepartmentofPhysicalandMacromolecularChemistry,FacultyofScience,CharlesUniversity,Hlavova2030,
128 40 Prague 2, Czech Republic,‡J. Heyrovsk? y Institute of Physical Chemistry, ASCR, v.v.i. Dolej? skova 3,
18223Praha8,CzechRepublic,§GileadSciencesandIOCBResearchCenter,InstituteofOrganicChemistryand
Biochemistry, ASCR, v.v.i., Flemingovo n. 2, 166 10 Prague 6, Czech Republic, and
Chemistry, ASCR, v.v.i., Area of Research Institutes 1001, 25068 Husinec-?Re? z, Czech Republic
)
Institute of Inorganic
Received October 25, 2009. Revised Manuscript Received December 9, 2009
Wepreparedtwofluorescein-[3-cobalt(III)bis(1,2-dicarbollide)]-conjugates.Theyaresparinglysolubleinwaterand
form large aggregates in aqueous solutions. An extensive study on their spectral and aggregation behavior was carried
out. To prepare their well-defined dispersion in aqueous systems, we studied the interaction of both probes with two
biocompatible amphiphilic systems, cyclodextrins, which are frequently used in drug-delivery systems, and phospho-
lipid membranes, which are the major constituents of cell barriers in living organisms. The presence of fluorescein in
both conjugates allows us to study their behavior in detail by steady-state and time-resolved fluorometry, fluorescence
correlation spectroscopy, and fluorescence lifetime imaging. The self-assembly of these metallacarboranes in aqueous
solutions was studied by dynamic light scattering. The study shows that the compounds interact with cyclodextrins that
increases their solubility in water, and they solubilize easily in phospholipid bilayers.
Introduction
Carboranes and metallacarboranes are interesting compounds
from both the fundamental and applied research points of view
and therefore they have been attracting interest of a number of
researchteamsforquitealongtime.1-3Sofartheyhavebeenused
in medicine mainly as boron carriers for boron neutron capture
therapy, BNCT,4-7but recently very promising attempts have
been made in exploring more sophisticated functional boron-
containing systems for therapeutic use. Here, we mention only
selected papers, which are closely related to our work; however,
otherreferencesandpertinentpiecesofinformationcanbefound
inarticlesreviewingthemedicaluseofboroncompounds.8-11We
focused our recent research almost exclusively on the physico-
chemicalbehaviorof[3-cobalt(III) bis(1,2-dicarbollide)]-anion in
aqueoussolution.Thestimulusforourworkwasthediscoveryof
an exceptionally efficient inhibition of HIV activity by cobalt
bis(dicarbollide) derivatives.12-14The aggregation of metallacar-
boranes in aqueous solutions contributes to the inhibition effi-
ciencyand influences the mode of inhibition.14Fromthe point of
view of medical applications, the challenge number one is the
preparation of a thermodynamically stable dispersion of spar-
inglysolubleboron-containingcompoundsinwater,whichwould
retain the biochemical activity.
In this communication, we present results on the spectral and
aggregation behavior of the fluorescein-[3-cobalt(III) bis(1,2-
dicarbollide)] conjugates in aqueous conditions as a further step
of our research on self-assembling of metallacarboranes in
water.15We synthesized two different metallacarborane deriva-
tives, GB176 (one metallacarborane cluster bound to fluorescein
molecule) and GB179 (two metallacarborane clusters bound to
fluorescein molecule; structures shown in Figure 1), which bear a
fluorescent marker that allows an investigation of their physico-
chemical behavior in detail by fluorometry. Interestingly, an
introduction of the fluorescein moiety to the structural patterns
responsible for the inhibitionofHIV proteasedid not deteriorate
the inhibitor activity substantially.16
Bothconjugatesaresparinglysolubleinaqueousmedia.Thatis
why we studied their interaction with two amphiphilic com-
pounds in order to prepare well-defined dispersions in water. In
this study, we chose cyclodextrins and phospholipid membranes
*Towhomcorrespondenceshouldbeaddressed.Tel:þ420221951292.Fax:
þ420224919752. E-mail: matej@vivien.natur.cuni.cz.
(1) Plesek, J. Chem. Rev. 1992, 92, 269.
(2) Teixidor, F.; Vinas, C.; Demonceau, A.; Nunez, R. Pure Appl. Chem. 2003,
75, 1305.
(3) Grimes, R. N. J. Chem. Educ. 2004, 81, 658.
(4) Soloway, A. H.; Tjarks, W.; Barnum, B. A.; Rong, F. G.; Barth, R. F.;
Codogni, I. M.; Wilson, J. G. Chem. Rev. 1998, 98, 1515.
(5) Hawthorne, M. F. Angew. Chem.-Int. Edit. Engl. 1993, 32, 950.
(6) Barth, R. F.; Soloway, A. H.; Goodman, J. H.; Gahbauer, R. A.; Gupta, N.;
Blue, T. E.; Yang, W. L.; Tjarks, W. Neurosurgery 1999, 44, 433.
(7) Vicente, M. G. H. ed. Anti-Cancer Agents in Med. Chem. 2006, 6, 73.
(8) Lesnikowski, Z. J. Collect. Czech. Chem. Commun. 2007, 72, 1646.
(9) Armstrong, A. F.; Valliant, J. F. Dalton Trans. 2007, 38, 4240.
(10) Sivaev, I. B.; Bregadze, V. V. Eur. J. Inorg. Chem. 2009, 2009, 1433.
(11) Valliant,J.F.;Guenther,K.J.;King,A.S.;Morel,P.;Schaffer,P.;Sogbein,
O. O.; Stephenson, K. A. Coord. Chem. Rev. 2002, 232, 173.
(12) Cigler, P.; Kozisek, M.; Rezacova, P.; Brynda, J.; Otwinowski, Z.;
Pokorna, J.; Plesek, J.; Gruner, B.; Doleckova-Maresova, L.; Masa, M.; Sedlacek,
J.; Bodem, J.; Krausslich, H. G.; Kral, V.; Konvalinka, J. Proc. Natl. Acad. Sci.
U.S.A. 2005, 102, 15394.
(13) Kozisek, M.; Cigler, P.; Lepsik, M.; Fanfrlik, J.; Rezacova, P.; Brynda, J.;
Pokorna, J.; Plesek, J.; Gruner, B.; Grantz Saskova, K.; Vaclavikova, J.; Kral, V.;
Konvalinka, J. J. Med. Chem. 2008, 51, 4839.
(14) Rezacova, P.; Pokorna, J.; Brynda, J.; Kozisek, M.; Cigler, P.; Lepsik, M.;
Fanfrlik, J.; Rezac, J.; Grantz Saskova, K.; Sieglova, I.; Plesek, J.; Sicha, V.;
Gruner, M.;Oberwinkler, H.; Sedlacek, J.; Krausslich, H. G.; Hobza,P.; Kral, V.;
Konvalinka, J. J. Med. Chem. 2009, 52, 7132.
(15) Matejicek, P.; Zednik, J.; Uselova, K.; Plestil, J.; Fanfrlik, J.; Nykanen, A.;
Ruokolainen, J.; Hobza, P.; Prochazka, K. Macromolecules 2009, 42, 4829.
(16) Private communication of Prof. Jan Konvalinka (Institute of Organic
Chemistry and Biochemistry, ASCR, Prague), 2009.

Page 2

DOI: 10.1021/la904047k
6269
Langmuir 2010, 26(9), 6268–6275
Uchman et al. Article
because they suitably interact with the studied boron conjugates.
Cyclodextrins, CDs, solubilize hydrophobic molecules in their
inner cavity and therefore have been used as components in drug
delivery systems.17-21For this study, it is important that the
dimensionsoftheinnerCDcavityarecomparablewiththesizeof
the used carborane cluster.18-22The phospholipid membranes
were chosenbecause theyplayanirreplaceable roleinlivingcells.
We believe that a detailed study of the interaction of cobalt
bis(dicarbollides) with phospholipid membranes, which is still
missing in the literature with few exceptions,23is desirable. For
the BNCT purposes, Hawthorne et al. used highly hydrophilic
boron clusters encapsulated within small unilamellar vesicles.24
We describe here a rather different system, the interaction of
the amphiphilic metallacarboranes with membranes formed by
dioleolylphosphatidylcholine, DOPC, in a form of supported
bilayers on glass and freestanding unilamellar vesicles.
The aggregation behavior of boron conjugates in water as well
as in solutions of cyclodextrins was studied by dynamic light
scattering, DLS. Further, we exploited several fluorescence tech-
niques, like the steady-state and time-resolved fluorescence and
anisotropy, fluorescence correlation spectroscopy, FCS, and
fluorescencelifetimeimaging,FLIMinordertogetdeeperinsight
in the studied systems.
Experimental Section
Materials. GB179 was prepared by ring-opening of the 8-
dioxane cobalt bis(dicarbollide) with fluorescein deprotonated
withNaHintoluene-dimethylethersolution.Thesereactionshave
been in recent years widely used to attachcobalt bis(dicarbollide)
cluster on various organic functional groups and platforms.25In
this particular case, the reaction also proceeds with deprotonated
fluoresceinmolecule,buttheproductisolationisrelativelytedious
due to pronounced amphiphilic behavior of the resulting species.
Nevertheless, the product could be obtained in pure form after
several chromatographic steps and characterized bycombination
of mass spectrometry (MS) and1H and11B NMR methods. The
purity of the compound was assayed by IP-RP high performance
liquid chromatography (HPLC) method used in our laboratories
for separation of cobalt bis(dicarbollide) derivatives being better
then 99.0%.
SynthesisofGB179.Toasuspensionoffluorescein(202mg,0.61
mmol,Aldrich),dried12honaSchlencktypevacuum line,30mL
of toluene-dimethylether (DME) (1:1) was added under nitrogen
followed with NaH (solid, high surface area (2.2 m2/g) made at
Institute of Inorganic Chemistry, 95%, 35 mg, 1.38 mmol). The
slurrywasstirredfor4handthensolutionofvacuum-dried(5hat
70 ?C) [(8-O(CH2CH2)2O-1,2-C2B9H10)(1020-C2B9H11)-3,30-Co]
(525 mg, 1.28 mmol, Katchem Ltd. Prague) in 30 mL of toluene-
dimethylether (DME) (1:1) was syringed through a septum. The
reaction mixture was stirred at ambient temperature for 36 h. The
reaction was quenched by careful addition of MeOH (5 mL)
followed with water (5 mL), the organic solvents were evapo-
rated and the semi-solid residue was treated between HCl (3 M,
3? 20mL) anddiethylether(25mL).Theetherlayerwaswashed
with water (2 ? 20 mL) and ether was then removed in vacuum.
The residue was dissolved in MeOH (ca. 30 mL) and the anionic
cobalt bis(dicarbollide) species were precipitated by an excess
of aqueous solution of CsCl. The orange precipitate was filtered,
washed with 30% aqueous methanol, and dried in vacuum.
The dark orange solid material was treated with hot benzene
(2 ? 20 mL, benzene extracts were discarded) and then dissolved
in CH2Cl2upon addition of few drops of MeOH. This solution
waslayeredwithi-octaneandlefttocrystallizeforthreedays.The
crude product was collected by filtration, dried in vacuum,
dissolved in CH2Cl2-CH3CN solvent mixture (3:1) and eluted
by step gradient increasing the acetonitrile content to 1:1. Last
intensive orange band contained the product. Final purification
wascarriedoutbychromatographyonMerck-LobarC18column
(2.5 ? 35 cmi.d.) using gradient ofaqueous MeOH (75 - 90%) as
the mobile phase, flow rate 5 mL/min. The metallacarborane
products were detected by visual detection and in UV at 308 nm,
the main front band contained the product. The effluent from
multiple injections (MeOH solution, 5 mL) was collected and
evaporated giving the pure GB1793Cs2, according HPLC, MS,
and NMR analyses. Finally, the product was converted to Naþ
saltbytreatmentoftheCsþsaltinEt2O(20mL)withdilutedHCl
(3 M, 4 ? 15 mL), brine (3 ? 15 mL), and water (2 ? 15 mL).
Water (5 mL) was added to ether solution and the volatiles were
removed in vacuum.
Yield of GB1793Cs2120 mg, 14%, HPLC k04.35 (Column:
Separon SGX C8, 250 ? 4 mm ID, Tessek Ltd. Prague, Czech
Republic, Mobile phase 3.0 mM hexylamine acetate buffer, pH
6.0 in 65% aqueous acetonitrile; flow rate 1.0 mL/min, detec-
tion UV at 308 nm), HPLC purity assay >99.0%.11B NMR
δB(128.32 MHz, acetone-d6, Et2O3BF3), δB: 23.25 (s, 2B, B8),
4.53 (d, J = 143, 2B, B80), 0.42 (d, J = 143, 2B, B100), -2.46 (d,
J=147,2B,B10),-4.34(d,J=146,4B,B40,70),-7.24(4B),-7.95
(8B)(2d,overlapB4,7,9,12,90,120),-17.20(d,J=159,4B,B50,
110), -20.25 (d, J = 162, 4B, B5, 11), -22.03 (br. d, 2B, B60),
-28.38 (br. d, J(B,H) = ca.140, 2B, B6).1H {11B} NMR (398.98
MHz, Acetone-d6), δH(ppm): 8.47 (d, J = 8.0, 1H, ArH), 7.985
(t,J=7.6,1H,ArH),7.94(t,J=7.6,1H,ArH),7.714(d,J=1.4,
1H, ArH), 7.579 (m, 4H, ArH), 7.369 (br. t, 2H, ArH), 7.261 (d,
J = 9.6, 1H, ArH), 4.486 (m, 2H, CH2O), 4.219 (s, 8H, CHcarb),
4.173 (br. s, 2H, CH2O), 4.105 (m, 4H, CH2O), 3.938 (t, J = 3.2,
4H, CH2O), 3.639 (m, 6H, CH2O), 3.526 (t, J = 4.8, 2H, CH2O),
3.443 (t, J = 4.2, 2H, CH2O), 3.38 (t, J = 4.4, 2H, CH2O).
1H{11Bselective} 398.98MHz,Acetone-d6),δB-H:2.91(H100),2.73
(H10),2.68(H40,70),2.56(H80),[2.89s,2.11s,1.79s](H4,7,9,12,
90,120),1.65(H60),1.61(H50,110),1.55(H5,11),1.45(H6).MS:m/
z (%) (ESI-) 576.52 (100), 578.92 (4) (calcd. 578.84) [M]2-;
1153.679 (15) 1159.67 (1) (calcd. 1159.70) [M þ H]-.
Another fluorescent metallacarborane conjugate, GB176,
bears fluorescein via thiourea-based spacer. It was prepared by
reactionoffluoresceinisothiocyanatewith8-[2-(2-aminoethoxy)-
ethoxy]cobalt(III) bis(dicarbollide).Thisreactivemetallacarborane
Figure 1. Chemical structure of GB176 and GB179.
(17) Uekama, K.; Hirayama, F.; Irie, T. Chem. Rev. 1998, 98, 2045.
(18) Hardie, M. J.; Raston, C. L. Chem. Commun. 1999, 13, 1153.
(19) Frixa, C.; Scobie, M.; Black, S. J.; Thompson, A. S.; Threadgill, M. D.
Chem. Commun. 2002, 23, 2876.
(20) Ohta, K.; Konno, S.; Endo, Y. Tetrahedron Lett. 2008, 49, 6225.
(21) Ohta, K.; Konno, S.; Endo, Y. Chem. Pharm. Bull. 2009, 53, 307.
(22) Harada, A.; Takahashi, S. J. Chem. Soc., Chem. Commun. 1988, 20, 1352.
(23) Atwell, R. J.; Sridharan, R.; De Levie, R. Proc. Indian Acad. Sci. 1986, 97,
431.
(24) Feakes, D. A.; Shelly, K.; Hawthorne, M. F. Proc. Natl. Acad. Sci. U.S.A.
1998, 96, 2531, and the references therein
(25) For examples see the review article: Semioshkin, A. A.; Sivaev, I. B.;
Bregadze, V. I. Dalton Trans. 2008, 8, 977, and the references therein.

Page 3

6270
DOI: 10.1021/la904047kLangmuir 2010, 26(9), 6268–6275
ArticleUchman et al.
intermediate was obtained using ring-opening of the
8-dioxane cobalt bis(dicarbollide) with ammonia.25
SynthesisofGB176(mixtureof5-and6-isomer,approximately
2:1 respective molar ratio). Fluorescein isothiocyanate (100 mg,
0.257 mmol, Sigma-Aldrich, mixed isomers) and 8-[2-(2-amino-
ethoxy)ethoxy]cobalt(III) bis(dicarbollide) (121 mg, 0.283 mmol)
were dissolved in 10 mL of dry acetonitrile under argon. Triethy-
lamine (65 mg, 0.64 mmol) was added and the reaction mixture
was stirred overnight at ambient temperature. The volatiles were
then evaporated in vacuum and the semisolid was chromato-
graphed on silica column using CH2Cl2-CH3CN solvent mixture
(1:3 to 1:1 gradient). The collected fractions containing a triethyl-
ammonium salt of the product were evaporated to dryness. Finally,
the product was converted to Naþsalt by treatment of the triethy-
lammoniumsaltinEt2O(20mL)withdilutedHCl(3M,4?15mL),
brine(3?15mL),andwater(2?15mL).Water(5mL) wasadded
to ether solution and the volatiles were removed in vacuum.
Yield of GB176 133 mg, 62%.11B NMR, δB(128.32 MHz,
acetone-d6, Et2O3BF3), δB: 23.22 (s, B8), 4.46 (d, J = 110, B80),
0.44(d,J=149,B100),-2.46(d,J=144,B10),-4.36(d,J=150,
B40,70),-7.19,-8.10(overlap,B4,7,9,12,90,120),-17.25(d,J=
156, B50, 110), -20.34 (d, J = 159, B5, 11), -21.98 (br. d, overlap,
B60),-28.60(br.d,J=ca.140,B6).1H{11B}NMR(398.98MHz,
Acetone-d6),δH(ppm):9.29(bs,acidicH),9.00(bs,acidicH),8.48
(s, Ar), 7.96 (s, ArH), 7.89 (m, ArH), 7.78 (s, ArH), 7.65 (s, ArH),
7.20 (d, ArH), 6.74-6.72 (m, xanthene), 6.66-6.63 (m, xanthene),
4.23 (s, CHcarb), 4.19 (br. s, CH2-O), 3.80-3.52 (m, CH2-O).
1H{11Bselective} (398.98 MHz, acetone-d6), δB-H: 2.91 (H100), 2.75
(H10),2.67(H40,70),2.58(H80),[2.89s,2.22s,1.81s](H4,7,9,12,
90,120),1.64(H50,60,110),1.56(H5,11),1.46(H6);MS:m/z(ESI-)
407.9 (calcd. 407.8) [M]2-; 816.5 (calcd. 816.3) [M þ H]-.
R-cyclodextrin, purum (HPLC), and β-cyclodextrin, purum
(HPLC), were purchased from Fluka, and γ-cyclodextrin, hy-
drate, was purchased from Aldrich; they were used as obtained.
Fluorescein isothiocyanate (mixture of 5- and 6-isomer) was
obtained from Sigma-Aldrich. Synthetic lipid, 1,2-dioleoyl-sn-
glycero-3-phosphocholine (DOPC) was purchased from Avanti
Polar Lipids (Alabaster, AL) and used without any further
purification.
Sample Preparation. The aqueous stock solutions of GB176
and GB179 were prepared by mixing 0.8 mg of the probes with
20 mL of filtered deionized water. The mixture was sonicated for
half an hour and stirred overnight. While GB176 was completely
dissolved (c = 5 ? 10-5M), the clear GB179 solution had to be
carefully separated from the sediment. The concentration was
estimatedbyUV-visspectroscopy(c,ca.1?10-5M).Thestock
solutions were not filtered and they were always sonicated before
mixing with other compounds.
The GB176 and GB179 solutions with cyclodextrins were
prepared by dissolution of the probes in the filtered solutions of
R-, β- and γ-CD in such a way to get 30 CD molecules per one
probe molecule.
Liposomal suspensions were prepared as follows: GB176 and
GB 179 were dissolved in methanol and mixed with the chloro-
form solution of DOPC in molar ratio 1:50000 for fluorescence
correlationspectroscopy(FCS)or1:100fortheotherfluorescence
measurements. The organic solvents were evaporated and the
lipid film was suspended in water (Milli-Q3 system, Millipore,
Etten-Leur). The obtained multilamellar vesicles were either
sonicated (FCS) or extruded through polycarbonate membranes
(Avestin, Ottawa, Canada) with 100 nm pores.
Supported phospholipid bilayers (SPBs) for measurement of
lipid lateral diffusion were formed on glass from small sonicated
vesiclessuspendedin10mMHepesbuffer,150mMNaCl,10mM
CaCl2, pH 7.4 directly in the closed perfusion FCS2 chamber
(Biotechs, Butler, U.S.A.), as described in ref 26. The glass was
exposed to UV-produced ozone for 5 min before SPB creation.
Giantunilamellarvesicles(GUVs)werepreparedbyamodified
electroformation method originally developed by Angelova.27
Chloroform solution of GB176 or GB179 and DOPC (molar
ratio 1:50000) was gently applied to the platinum electrodes and
keptforatleast3hundervacuum.Theobtaineddrylipidfilmwas
then hydrated with a 150 mOsm sucrose solution, and an AC
electricalfieldwasappliedtotheelectrodes.Themediumwasthen
exchanged for a 148 mOsm glucose solution. The presence of
glucose in the final solution allowed the sedimentation of lipo-
somesanddecreasedthemovementofvesicles.Themeasurements
were performed in a closed FCS2 chamber at 10 ?C to minimize
membrane undulations.
Methods. Dynamic Light Scattering (DLS). The light
scattering setup (ALV, Langen, Germany) consisted of a 633 nm
He-Ne laser, an ALV CGS/8F goniometer, an ALV High QE
APD detector, and an ALV 5000/EPP multibit, multitau auto-
correlator. DLS data analysis was performed by fitting the
measured normalized intensity autocorrelation function g2(t) =
1 þ β|g1(t)|2, where g1(t) is the electric field correlation function,
tisthelag-time,andβisafactoraccountingfordeviationfromthe
ideal correlation. An inverse Laplace transform of g1(t) with
the aid of a constrained regularization algorithm (CONTIN)
provides the distribution of relaxation times, τA(τ). Effective
angle- and concentration-dependent hydrodynamic radii, RH-
(q,c), were obtained from the mean values of relaxation times,
τm(q,c), of individual diffusive modes using the Stokes-Einstein
equation.
Steady-State and Time-Resolved Fluorometry. Steady-state
fluorescence was measured in 1 cm quartz cuvettes with a Teflon
stopper using a SPEX Fluorolog 3-11 fluorometer. Fluorescence
decays were measured by means of the time-correlated single
photoncountingtechniqueonanEdinburghInstrumentsED299
T fluorometer equipped with an IBH NanoLED-03 excitation
source (496 nm peak wavelength, 0.1 ns fwhm of the pulse, and
1 MHz repetition rate). The measured decays were fitted to the
convolution of multiexponential functions with the instrument
responseprofileusingtheMarquardt-Levenbergnonlinearleast-
squares method. Low values of χ2and random distribution of
residuals were used as criteria of fit.
Time-Resolved Fluorescence Anisotropy. A time-correlated
single photon counting technique was used for measurements of
time-resolved anisotropy. The decays were recorded on an IBH
5000Utime-resolvedfluorometer(IBH,Glasgow,U.K.),equipped
with pulsed diode laser (LDH-P-C-470, 470 nm, fwhm = 80 ps,
10 MHz repetition rate, PicoQuant) and a cooled Hamamatsu-
MCPphotomultiplier.
Fluorescence Correlation Spectroscopy (FCS). The FCS
measurements were performed using an upgraded ConfoCor 1,
Carl Zeiss, Germany. The FCS consists an adapted inverted
confocal microscope Axiovert 135 TV with a water immersion
objective C-Apochromat 40?/1,2 W, pulsed diode laser LDH-P-
C-470 (470 nm, 40 MHz, fwhm of the overall IRF 500 ps) with
PDL800-Bdriver(PicoQuantGmbH,Germany),aproperfluore-
scence filter set (HQ457/10, z470rdc, HQ515/50 - Chroma Tech-
nology, U.S.A.), a single photon counting detector SPCM-
AQR-13-FC (Perkin-Elmer). The hydrodynamic radius RHwas
evaluated from the diffusion coefficient using the Stokes-
Einstein formula.
Z-scanFCS.LateraldiffusionofGB176andGB179inDOPC
SPBs was measured using Z-scan technique,28which was shown
to be the only artifact-free single focus measurement of lateral
diffusion coefficients.29A set of FCS curves was measured at
various Z positions of the focal plane with respect to the bilayer
spacedby0.2μm.ParticularFCScurvesweretreatedaccordingto
(26) Benes, M.; Billy, D.; Benda, A.; Speijer, H.; Hof, M.; Hermens, W. T.
Langmuir 2004, 20, 10129.
(27) Angelova, M. I.; Dimitrov, D. S. Faraday Discuss. 1986, 81, 303.
(28) Benda, A.; Benes, M.; Marecek, V.; Lhotsky, A.; Hermens, W. T.; Hof, M.
Langmuir 2003, 19, 4120.
(29) Dertinger, T.; Pacheco, V.; von der Hocht, I.; Hartmann, R.; Gregor, I.;
Enderlein, J. Chem. Phys. Chem. 2007, 8, 433.

Page 4

DOI: 10.1021/la904047k
6271
Langmuir 2010, 26(9), 6268–6275
Uchman et al. Article
ref 28 and the obtained diffusion times were plotted versus
normalized particle number as described in refs 30 and 31 to
obtain the effective diffusion coefficient and the intercept with
diffusiontimeaxischaracterizingthetypeofdiffusion(i.e.,freeor
hindered).30,31
Fluorescence Lifetime Imaging (FLIM). Fluorescence life-
time imaging (FLIM) measurements were carried out on an
inverted epifluorescence confocal microscope MicroTime 200
(Picoquant,Germany).Weusedconfigurationcontainingapulsed
diode laser (LDH-P-C-470, 470 nm, Picoquant) at 40 MHz
repetition rate, dichroic mirror 490DRLP, band-pass filter
HQ515/50 (Omega Optical), water immersion objective (1.2 NA,
60? ) (Olympus), and detector PDM SPAD (MPD, U.S.A.).
UV-vis Spectroscopy. UV-vis absorption spectra were car-
riedout withaHewlett-Packard8452adiode-arrayspectrometer.
Results and Discussion
The Probes in Water. First of all we studied GB176 and
GB179 (the structures are depicted in Figure 1) in water by
UV-vis and fluorescence spectroscopy. Although the probes
(especially GB179) are sparingly soluble in water as compared
with the parent [3-cobalt bis(1,2 dicarbollide](1-) anion and their
dissolution is slow, the obtained concentrations allowed success-
ful measurements by all spectroscopic methods used. The solubi-
lity in water was estimated by means of UV-vis spectroscopy to
be ca. 5 ? 10-4and 1 ? 10-5M for GB176 and GB179, res-
pectively. The absorption spectra of both probes at different pH
are shown inFigure2.One can see typicalbands offluorescein in
the region 400-550 nm and the band of the metallacarborane
cluster around 300 nm. The fluorophore bands are almost
identical with those of the molecularly dissolved fluorescein with
comparable pH-response.32There are no signs of fluorescein
multimerization, even at relatively high concentrations.33,34A
certain difference between two synthesized probes consists in the
factthatGB179cannotlosetwoprotonsfromthefluoresceinpart
but GB176 can do. It affects the structure of absorption bands at
highpHasshowninFigure2andalsocausesaconsiderablylower
fluorescence quantum yield32of GB179.
The steady-state fluorescence spectra (Inset in Figure 2) of GB
176 and GB 179 are very similar to that of pure fluorescein.
Nevertheless, the shape of the GB179 emission band in basic
buffers reminds the spectrum of fluorescein dianion despite the
fact that it cannot be ionized like GB176. The only significant
difference in the spectroscopic behavior of the probes and pure
fluorescein consists in the emission decay kinetics. The time-
resolved decays of studied compounds are always double-expo-
nential, while that of fluorescein is exponential. This fact is not
surprising,becausethesameobservationwasreportedforDNA-
fluorescein conjugates.35The explanations are analogous: the
pendant fluorescein can acquire several orientations with respect
to the bulky metallacarborane group. Individual conformations
influence differently the electronic structure of the fluorophore
and the allowance ofnonradiative processes depleting the excited
state. The short time is ca. 0.5 ns, the long one ca. 3.5 ns and the
fractionintensityoftheslowercomponentslightlyprevails.Inthe
nextpart,wewillshowthat allfluorescencedecaysofGB176and
GB179 incorporated in complexes with other compounds can be
well described by double-exponential curves with two character-
istic times, while the corresponding fractional intensities do
change considerably (see Table 1).
Because we observed a rich aggregation behavior of both
probes in water,36-38we performed a set of light scattering
measurements in aqueous solutions differing in concentration
(dilution up to 1 ? 10-7M), pH, and ionic strength. They re-
vealed the presence of quite large aggregates with radii around
200 nm (Figure 3). The critical aggregation concentration is
below a detection limit of LS, which is consistent with our
previous observations.36,37The distribution of correlation times
is typically monomodal, but in several cases the bimodal dis-
tribution were also detected. It should be mentioned that we
recently observed bimodal distributions also in other systems
Figure 2. Normalized absorption and fluorescence emission (in
Inset, ex. 495 nm) spectra of (a) GB176 and (b) GB179 in phos-
phate/boratebuffersofvariouspH,1.54,3.45,5.49,6.40,8.17,9.22
and 10.13; the lowest and the highest values of pH are indicated in
graph, and arrows indicate increase of pH.
Table 1. Results of Time Resolved Fluorometry (ex. 495 nm, em. 515
nm; Fraction Intensities Fiand Fluorescence Lifetimes τiObtained
fromFittingofExperimentalData)forGB176andGB179atVarious
Conditionsa
probesampleF1(%)
τ1/nsF2(%)
τ2/ns
τmean/nsb
GB176water
methanol
R-CD
β-CD
γ-CD
DOPC
44
70
39
59
69
20
0.50
0.67
0.57
0.68
0.83
0.98
56
30
61
41
31
80
3.58
3.69
3.72
3.67
3.56
3.76
2.24
1.57
2.49
1.90
1.67
3.20
GB179water
methanol
R-CD
β-CD
γ-CD
DOPC
39
91
47
61
70
83
0.11
0.10
0.01
0.27
0.28
0.35
613.48
1.34
3.26
3.60
3.04
1.81
2.17
0.21
1.74
1.56
1.10
0.60
9
53
39
30
17
aAll solutions with cyclodextrins contain 30 mol equiv of CD per
probe.bτmean= F1τ1þ F2τ2.
(30) Humpolickova, J.; Gielen, E.; Benda, A.; Fagulova, V.; Vercammen, J.;
Vandeven, M.; Hof, M.; Ameloot, M.; Engelborghs, Y. Biophys. J. 2006, 91, L23.
(31) Wawrezinieck, L.; Rigneault, H.; Marguet, D.; Lenne, P. F. Biophys. J.
2005, 89, 4029.
(32) Sjoback, R.; Nygren, J.; Kubista, M. Spectrochim. Acta A 1995, 51, L7.
(33) Arbeloa, L. J. Chem. Soc., Faraday Trans. 1981, 77, 1725.
(34) Arbeloa, L. J. Chem. Soc., Faraday Trans. 1981, 77, 1735.
(35) Sjoback, R.; Nygren, J.; Kubista, M. Biopolymers 1998, 46, 445.
(36) Matejicek, P.; Cigler, P.; Prochazka, K.; Kral, V. Langmuir 2006, 22, 575.
(37) Matejicek, P.; Cigler, P.; Olejniczak, A. B.; Andrysiak, A.; Wojtczak, B.;
Prochazka, K.; Lesnikowski, Z. J. Langmuir 2008, 24, 2625.
(38) Kubat, P.; Lang, K.; Cigler, P.; Kozisek, M.; Matejicek, P.; Janda, P.;
Zelinger, Z.; Prochazka, K.; Kral, V. J. Phys. Chem. B 2007, 111, 4539.

Page 5

6272
DOI: 10.1021/la904047kLangmuir 2010, 26(9), 6268–6275
ArticleUchman et al.
containing metallacarboranes, such as ATP-metallacarborane
conjugates, which bear high charge similarly to the studied
compounds.37It is surprising that neither H- nor J-fluorescein
aggregate bands were detected by UV-vis spectroscopy.33,34It
suggests that the main driving force toward aggregation is the
amphiphilic character of boron clusters and the pendant fluore-
scein does neither participate at, nor significantly affect the
aggregation, at least, in alkaline buffers. The size of aggregates
increases with decreasing pH because the hydrophobicity of the
pH-responsivefluorescein moiety
Figure 3).
The behavior of aqueous solutions is complex. At very low
concentrations, only single molecules have been detected by
fluorescence correlation spectroscopy, FCS. On the contrary,
the fluorescence lifetime imaging, FLIM, clearly shows that at
higher concentrations, some large aggregates slowly sediment to
the bottom of the vessel as shown in Figure 4a,b. The stains have
dimensions slightly above 400 nm (limit of the fluorescence
imaging). This observation is in agreement with DLS data.
Interaction of the Probes with Cyclodextrins. It is a well-
established fact that the parent metallacarborane interact with
cyclodextrins, CDs, and form a variety of complexes differing in
stoichiometry and stability.39,40It is also known that fluorescein
interacts with CDs and weakly binds to them, but the interaction
has almost no impact on the fluorescence properties when
omitting a disruption of nonfluorescent aggregates which leads
to an increase in the overall emission intensity.41We studied the
changes in the aggregation behavior of the probes caused by the
interaction with R-, β- and γ-CD in detail. We used two methods
of preparation of studied solutions, (i) direct dissolution of the
solid sample in an excess of CDs aqueous solution, and (ii) step-
by-step titration of the GB176 solution by CDs (both methods
lead to the same results); however, we did not apply the titration
for GB179 because of a very low solubility and a presence of ill-
defined agglomerates. In the further text, we will focus on the
time-resolved fluorescence and anisotropy results (shown in
Tables 1 and 2), because the steady-state fluorescence did not
provide reliable data and their analysis did not reveal sufficient
increases(depictedin
details elucidating the behavior at the molecular level. For a
further discussion, it is important to remind the reader that large
aggregatesweredetectedinGB176andGB179solutionsbyDLS.
The measurement confirmed that the aggregates disappear only
partly after addition of CDs at elevated concentrations. On the
other hand, no large aggregates were detected by FCS in very
dilute solutions.
The solubility of GB179 increases in neutral and alkaline
solutions with an excess of CDs. The increase reflects the size of
CDscavity:solubilityis2,6,and7-timeshigherthanthatinwater
for R-, β- and γ-CD, respectively. Nevertheless, both GB176 and
GB179 complexes with CDs are insoluble in acidic buffers. This
type of behavior indicates specific interaction of the cyclodextrin
cavity with the metallacarborane conjugates. This assumption is
supportedbychangesintheabsorptionspectraofGB176/CDsin
water (not shown). The fluorescein dianion band has a clearly
pronounced shoulder at the blue edge revealing the presence of
fluorescein monoanion as compared to the spectrum of pure
GB176. Further, a careful inspection of emission spectra of
Figure 3. Distribution of correlation times obtained by DLS at
scattering angle 90? of (a) GB176 and (b) GB179 in buffers with
various pH as indicated in graph.
Figure 4. FLIM images of (a) GB176 and (b) GB179 aggregates
deposited on the bottom of a solution; typical dimensions of the
colorstainsareca.400-500nm,imagesdimensionsare80?80μm.
Table 2. Results of Time-Resolved Fluorescence Anisotropy (ex. 495
nm, em. 520 nm; Amplitudes Ai, Rotational Correlation Times τiand
Residual Anisotropy rresobtained from fitting of experimental data)
for GB176 and GB179 at various conditionsa
probesampleA1
τ1/nsA2
τ2/nsrres
GB176water
methanol
R-CD
β-CD
γ-CD
DOPC
0.170.09 0.3
0.37
0.28
0.24
0.26
0.25
0.36
0.33
0.54
0.68
0.62
2.39
0.020
0.016
0.009
0.003
-0.003
0.002
0.23
0.15
0.17
0.11
0.08
0.14
0.08
0.12
GB179water
methanol
R-CD
β-CD
γ-CD
DOPCb
0.24
0.28
0.33
0.39
0.4
5.41
0.45
0.34
1.09
0.75
1.14
72.80
0.022
0.061
0.045
-0.006
-0.005
-5.040.240.02
aAll solutions with cyclodextrins contain 30 mol equiv of CD per
probe.bDue to very low fluorescence lifetime the fit provides mean-
ingless values.
(39) Rak, J.; Tkadlecovka, M.; Cigler, P.; Kral, V. Chem. Listy 2008, 102, 209.
(40) Chetcuti, P. A.; Moser, P.; Rihs, G. Organometallics 1991, 10, 2895.
(41) Flamigni, L. J. Phys. Chem. 1993, 97, 9566.

Page 6

DOI: 10.1021/la904047k
6273
Langmuir 2010, 26(9), 6268–6275
Uchman et al.Article
GB179/CDs (Figure 5b) reveals that the positions of emission
bands slightly shift and their shapes change reminding the
fluorescein monoanion emission band. It was already mentioned
that interaction of pure fluorescein with CDs does not almost
affect its emission spectra. Therefore we assume that the forma-
tion of a complex between boron clusters and CD molecules
influences the microenvironment polarity of fluorescein and the
acidobasic equilibrium of its individual forms. This premise is
backed up by the previously reported chromatographic behavior
of metallacarborane derivatives on cyclodextrin phases.42How-
ever, a contribution due to the conformation change of probes
cannot be neglected.
The changes in emission kinetics are evidenced by the time-
resolved fluorescence decays from samples with an excess of CDs
(see Table 1). The slow lifetime component, τ2, does not almost
change. The fast component, τ1, increases only little with the size
of CD molecule but the corresponding fraction intensity grows
considerably. Even though the fluorescence response is complex,
we can conclude that the binding of fluorescent conjugates to
bulkyCDmoleculesalterstheprobeconformationandinfluences
itsmotioninthesolution,interactionwithpolarsolventmolecules
and consequently also the ionization.
Different impacts of R-, β-, or γ-CD on fluorescence decays
have been observed during the titration experiments. Figure 6
depicts the dependence of the mean fluorescence lifetime on the
ratio of CD molecules per one GB176 molecule in pure water.
It is obvious that the interaction with the small cavity containing
R-CD does not have a strong impact on the excited state of
fluorescein. The mean lifetime gradually increases with the R-CD
content (both τ2and F2increase, while τ1does not change). It
indicates that the conformation corresponding to slow fluore-
scence decay time becomes more stabilized. The fact that the
changes are relatively modest suggests that the binding is weak
and the pendant fluorescein remains exposed to the aqueous
medium similarly to the conjugates in water. The interaction of
GB176 with γ-CD, which contains the largest cavity, has an
opposite impact on the fluorescence characteristics. In this case,
F1 increases considerably. We can see that the pronounced
changes start at low γ-CD amounts and continue until the
saturation at the CD-to-GB176 ratio ca. 2. The fluorescence
decay kinetics reflects the fact that γ-CD interacts strongly with
the boron cluster. The curve for β-CD is interesting. The slow
component, τ2, continuously decreases, but the short component
and its fraction intensity (τ1and F1) change suddenly at the CD-
to-GB176 ratio ca. 2. It means that, similarly to γ-CD, the
conformation with a faster fluorescent decay is more populated
in the excess of CDs. Hence the interaction of both β-CD and
γ-CD with boron conjugates seems to affect the pendant fluore-
scein in a similar way. However, the complex forming tendency
of β-CD is less pronounced and a certain “critical β-CD concen-
tration”isapparentlynecessarytotriggerthecomplexformation.
It is obvious that more studies aiming at understanding the
interaction in full detail are needed, but we can conclude that
the observed effects are related to different affinity of R-, β- and
γ-CD to the studied boron conjugates and to different stoichio-
metry of created complexes (the association constants of CDs
with the probes have not been determined).39
As concerns the time-resolved anisotropy results, they are
consistentwiththeabove-mentionedconclusions.Theanisotropy
decays of pure conjugates in aqueous media are fast with almost
noresidualanisotropyandthecorrespondingrotationcorrelation
times are short indicating the presence of mobile molecules in
water.Experimentalvaluesaresimilartothoseinmethanolwhere
the probes are well soluble (Table 2). The motion of GB179 is
slower than that of GB176, which correlates well with the
molecular structure of probes (Figure 1). After the addition of
CDs,the rotationtimes increaseslightlyandreflect thesizeofthe
CD molecule as the complex becomes bulkier. The steady-state
anisotropies (not shown) increase appreciably after the addition
ofcyclodextrins,butthisisdue,inmajorpart,totheshorteningof
the fluorescence lifetime and, only partially, to a slower depolar-
ization.
Interaction of the Probes with Phospholipid Membranes.
Inthelastpart,westudiedtheincorporationofprobesGB176and
GB179 in phospholipid bilayers, namely the possibility of their
direct dissolution in the membranes in a fluid state at room
temperature. The molar ratios are given in the Experimental
Section. The analysis of polarity-dependent fluorescence charac-
teristics shows that the synthesized probes do incorporate
in the DOPC vesicles (see Figure 5 and Tables 1 and 2). The
absorptionandemissionspectrashowthatthependantfluorescein
is present in the membrane mainly as monoanion and partly as a
nondissociated neutral molecule. Interestingly, GB179 has unu-
sually low value of fluorescence lifetime in nonpolar media.
The rotational motion of probes in membranes is signifi-
cantly hindered as evidenced by the anisotropy data (Table 2).
Figure 5. Fluorescenceemissionspectra(ex.495nm)of(a)GB176
and (b) GB179 in various solutions as indicated in graph.
Figure 6. Mean fluorescence lifetime (ex. 495 nm, em. 515 nm)
of GB176 as a function of molar equivalents of R- β- and γ-cyclo-
dextrins.
(42) Gruner, B.; Plzak, Z. J. Chromatogr., A 1997, 789, 497.

Page 7

6274
DOI: 10.1021/la904047kLangmuir 2010, 26(9), 6268–6275
ArticleUchman et al.
The FCS autocorrelation curves for the probes in water and in
DOPC vesicles are shown in Figure 7. Curve 1 corresponds to
very dilute aqueous solution of GB176, where the probe is
molecularly soluble. The curves 2 and 3 measured for GB176
and GB179, respectively, in DOPC vesicles reflect the dimen-
sions of the vesicle (radius ca. 50 nm). From the analysis of the
autocorrelation curves we can estimate that ca. 95% of GB176
is solubilized in vesicles and the rest corresponds to free
molecules in water. The less water-soluble probe GB179, it is
practically 100%solubilized andtherefore onlylargeliposomes
were detected.
FCS measurements for probes in supported phospholipids
bilayers on glass, SPBs, yield the lateral diffusion coefficients
in the fluid DOPC membranes at room temperature (25 ?C),
4.8 ( 0.2 and 2.0 ( 0.3 μm2/s for GB176 and GB179, res-
pectively. If we compare the results with our previous measure-
ments of fluorescence dyes in SPBs,43we can conclude that the
probesaremolecularlysolubleinthemembrane,whichisinline
with their amphiphilic structure. The observed significant
difference in their lateral diffusion coefficients might reflect
the different molecular sizes of GB176 and GB179. We should
stress that those membranes are apparently serving as a perfect
matrix, while in proteins as well as in aqueous solutions we
observe the formation of multimolecular aggregates of cobalt
bis(dicarbollides).36,44The GUV (giant unilamelar vesicles)
and SPBs were visualized by means of FLIM (see Figure 8a-c).
The obtained intensity scans show confluent DOPC SPBs.
Here we can see that GB179 has lower mean lifetime in SPB
compared to GB176 (depicted by different colors of the corre-
sponding membrane in Figure 8b,c).
Conclusions
Two different fluorescein-metallacarborane conjugates
GB176, which bears one [3-cobalt bis(1,2-dicarbollide)](-1)
anion and one pendant fluorescein, and GB179, in which the
fluorescein moiety connects two metallacarborane anions, were
synthesized. The fluorescein part of GB176 can dissociate
forming both monoanion and dianion depending on pH. The
fluorescein part of GB179 cannot release any proton. Both
probes, especially GB179, are only little soluble in water as
compared with their parent metallacarborane. Nevertheless,
the behavior of their aqueous solutions could have been studied
in sufficient detail by fluorescence and other spectroscopic
techniques.Eventhoughthesteady-statespectraofthependant
fluorescein are similar to those of the free fluorescein, the time-
resolved fluorescence data show that the photophysics of the
fluoresceinisaffectedbyitsattachmenttothemetallacarborane
cluster. Because of their amphiphilic character the probes self-
assemble in aqueous solutions forming unstable and ill-defined
aggregates. To prepare better-defined dispersions, the interac-
tion of both probes with R-, β- and γ-CD and with phospho-
lipids bilayers was studied.
The cyclodextrins interact with the synthesized probes and
increases their solubility. The study showed that γ-CD which
containsrelativelylargehydrophobiccavityhasthemostefficient
complexationeffect.ThecomplexeswithCDsarestableatneutral
and high pH, but their solubility steeply decreases at low pH.
Even though it certainly requires a further tuning of the size and
rim functionalization, this approach outlines a pathway to the
potentialdeliverysystemformetallacarborane-baseddrugsbased
on noncovalent interactions.
All methods used suggest that both probes can be loaded at
ambient temperatures in phospholipids bilayers thanks to the
amphiphilic nature of metallacarboranes. The presence of fluore-
scein allows the visualization of conjugates embedded in mem-
branesbyFLIMandbyfluorescencemicroscopy.Furthermore,it
is possible to study the lateral diffusion of probes within the
DOPC bilayer by FCS. The FCS data suggest that the probes
moveinthemembraneinaform offreeindividual molecules.We
would like to mention that those FLIM and FCS experiments
Figure 7. Normalizedautocorrelation functions obtainedbyFCS
of (1) GB176 in pure water, (2) GB176 in aqueous solution of the
DOPC vesicles, and (3) GB179 in aqueous solution of the DOPC
vesicles.
Figure 8. FLIM images of (a) GB176 in aqueous solution of the
DOPCcontainingGUV (scale bar 5 μm), (b) GB176 inthe DOPC
containing SPB on glass, and (c) GB179 in the DOPC contain-
ing SPB on glass; images dimensions are (a) 33 ? 33 μm, (b,c)
40 ? 20 μm; molar fraction of the probes in DOPC given in the
Experimental Section.
(43) Przybylo, M.; Sykora, J.; Humpolickova, J.; Benda, A.; Zan, A.; Hof, M.
Langmuir 2006, 22, 9096.
(44) Fanfrlik, J.; Brynda, J.; Rezac, J.; Hobza, P.; Lepsik, M. J. Phys. Chem. B
2008, 112, 15094.

Page 8

DOI: 10.1021/la904047k
6275
Langmuir 2010, 26(9), 6268–6275
Uchman et al. Article
were conducted in model membranes, while further experiments
in living cells are in progress.
Bothprobesmimicthebasicboron-containingstructuresofthe
HIV protease inhibitors.12-14Although the incorporation of the
fluorescein moiety could change biological and solution proper-
ties of the metallacarborane conjugates, this structural alteration
did not significantly influence inhibition activity of the probes.16
Moreover, their aggregation behavior in aqueous solutions is
comparable to the other metallacarborane conjugates.36-38Since
a broad variety of fluorescence techniques are commonly used in
biochemistry, GB176 and GB179 represent valuable alternatives
to the most active metallacarborane HIV protease inhibitors.
However, we are aware that the fluorescein-metallacarborane
conjugates are not supposed to be used in the “real” medical
applications.
Acknowledgment. Theauthorswouldliketoacknowledgethe
financialsupportoftheGrantAgencyoftheAcademyofSciences
oftheCzechRepublic(IAAX00320901andIAA400400621).The
studywasalsosupportedbytheResearchPlansoftheMinistryof
Education of the Czech Republic MSM 0021620857 and
AV0Z40320502. P.M. and M.U. thank Vladimir Dordovic for
assistance with LS and fluorescence measurements. The authors
thank Professor Jan Konvalinka (Institute of Organic Chemistry
and Biochemistry, ASCR, Prague) for HIV protease inhibition
data.