Efficient delivery of streptavidin to mammalian cells: Clathrin-mediated endocytosis regulated by a synthetic ligand

Article abstract of DOI:10.1021/ja0258733

The efficient delivery of macromolecules to living cells presents a formidable challenge to the development of effective macromolecular therapeutics and cellular probes. We describe herein a novel synthetic ligand termed "Streptaphage" that enables effici

Full text of DOI:10.1021/ja0258733

Published on Web 05/04/2002
Efficient Delivery of Streptavidin to Mammalian Cells:
Clathrin-Mediated Endocytosis Regulated by a Synthetic
Ligand
Stephen L. Hussey and Blake R. Peterson*
Contribution from the Department of Chemistry, The PennsylVania State UniVersity,
152 DaVey Lab, UniVersity Park, PennsylVania 16802
Received February 8, 2002
Abstract: The efficient delivery of macromolecules to living cells presents a formidable challenge to the
development of effective macromolecular therapeutics and cellular probes. We describe herein a novel
synthetic ligand termed “Streptaphage” that enables efficient cellular uptake of the bacterial protein
streptavidin by promoting noncovalent interactions with cholesterol and sphingolipid-rich lipid raft subdomains
of cellular plasma membranes. The Streptaphage ligand comprises an N-alkyl derivative of 3â-
cholesterylamine linked to the carboxylate of biotin through an 11-atom tether. Molecular recognition between
streptavidin and this membrane-bound ligand promotes clathrin-mediated endocytosis, which renders
streptavidin partially intracellular within 10 min and completely internalized within 4 h of protein addition.
Analysis of protein uptake in Jurkat lymphocytes by epifluorescence microscopy and flow cytometry revealed
intracellular fluorescence enhancements of over 300-fold (10 µM ligand) with >99% efficiency and low
toxicity. Other mammalian cell lines including THP-1 macrophages, MCF-7 breast cancer cells, and CHO
cells were similarly affected. Structurally related ligands bearing a shorter linker or substituting the protonated
steroidal amine with an isosteric amide were ineffective molecular transporters. Confocal fluorescence
microscopy revealed that Streptaphage-induced uptake of streptavidin functionally mimics the initial cellular
penetration steps of Cholera toxin, which undergoes clathrin-mediated endocytosis upon binding to the
lipid raft-associated natural product ganglioside GM1. The synthetic ligand described herein represents a
designed cell surface receptor capable of targeting streptavidin conjugates into diverse mammalian cells
by hijacking the molecular machinery used to organize cellular membranes. This technology has potential
applications in DNA delivery, tumor therapy, and stimulation of immune responses.
Introduction
molecular cargo. Alternatively, cellular membranes can be
chemically altered or permeabilized to facilitate macromolecular
The effectiveness of therapeutics and cellular probes depends
critically on the efficient delivery of molecules into living cells.
Whereas many hydrophobic compounds of low molecular
weight diffuse freely through low-polarity cell membranes,
macromolecules such as proteins and DNA typically require
active transport mechanisms to gain access to intracellular
receptors.1 To enhance the cellular uptake of poorly permeable
molecules, covalent or noncovalent modification by cationic,
hydrophobic, or amphipathic polymers2-9 or hydrophobic
lipids10-12 can favorably modify the chemical properties of
uptake.13-15 However, the molecular mechanisms underlying
many of these methods are not well understood. As a conse-
quence, efficiencies of macromolecular delivery are often limited
by variability, toxicity, and unpredictable cell-type specificity.
Hence, novel methods that enhance uptake of poorly permeable
molecules are needed in basic cell biology, tumor therapy, and
genetic therapy.
Mammalian cells internalize specific small molecules, mac-
romolecules, and particles through endocytosis, an active
transport process dependent on time, temperature, pH, and
energy.16,17 This process targets extracellular materials to specific
membrane-sealed compartments termed endosomes within the
cytoplasm. The most common fate of molecules internalized in
* Corresponding author. E-mail: brpeters@chem.psu.edu. Phone: (814)
865-2969. Fax: (814) 863-8403.
(1) Smith, D. A.; van de Waterbeemd, H. Curr. Opin. Chem. Biol. 1999, 3,
373-378.
(2) Fischer, P. M.; Krausz, E.; Lane, D. P. Bioconj. Chem. 2001, 12, 825-
841.
(3) Garnett, M. C. Crit. ReV. Ther. Drug Carrier Syst. 1999, 16, 147-207.
(4) Hawiger, J. Curr. Opin. Chem. Biol. 1999, 3, 89-94.
(5) Lindgren, M.; Hallbrink, M.; Prochiantz, A.; Langel, U. Trends Pharmacol.
Sci. 2000, 21, 99-103.
(10) Bendas, G. BioDrugs 2001, 15, 215-224.
(11) Rait, A.; Pirollo, K.; Will, D. W.; Peyman, A.; Rait, V.; Uhlmann, E.;
Chang, E. H. Bioconj. Chem. 2000, 11, 153-160.
(12) Zelphati, O.; Wang, Y.; Kitada, S.; Reed, J. C.; Felgner, P. L.; Corbeil, J.
J. Biol. Chem. 2001, 276, 35103-35110.
(6) Schwarze, S. R.; Hruska, K. A.; Dowdy, S. F. Trends Cell Biol. 2000, 10,
290-295.
(13) Lu, Y.; Friedman, R.; Kushner, N.; Doling, A.; Thomas, L.; Touzjian, N.;
Starnbach, M.; Lieberman, J. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 8027-
8032.
(7) Wender, P. A.; Mitchell, D. J.; Pattabiraman, K.; Pelkey, E. T.; Steinman,
L.; Rothbard, J. B. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 13003-13008.
(8) Schwarze, S. R.; Ho, A.; Vocero-Akbani, A.; Dowdy, S. F. Science 1999,
285, 1569-1572.
(14) Walev, I.; Bhakdi, S. C.; Hofmann, F.; Djonder, N.; Valeva, A.; Aktories,
K.; Bhakdi, S. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 3185-3190.
(15) Wojda, U.; Goldsmith, P.; Miller, J. L. Bioconj. Chem. 1999, 10, 1044-
1050.
(9) Umezawa, N.; Gelman, M. A.; Haigis, M. C.; Raines, R. T.; Gellman, S.
H. J. Am. Chem. Soc. 2002, 124, 368-369.
9
10.1021/ja0258733 CCC: $22.00 © 2002 American Chemical Society
J. AM. CHEM. SOC. 2002, 124, 6265-6273
6265

A R T I C L E S
Hussey and Peterson
this fashion is degradation or recycling of material back to the
cell surface, but internalized substances can also be released
into the cytosol.16,17 One of the most important and best-
characterized endocytic mechanisms is receptor-mediated en-
docytosis via clathrin-coated pits. Clathrin is a protein that
assembles at cellular plasma membranes into a basketlike
concave framework of hexagons and pentagons that form pit
structures.18-20 Some endogenous proteins, such as the membrane-
associated epidermal growth factor receptor, become clustered
in coated pits upon binding to soluble ligands. This clustering
results in invagination of the pit to form clathrin-coated vesicles
that internalize the receptor-ligand complex and enable activa-
tion of signal transduction pathways.16 Internalized clathrin-
coated vesicles become intracellular endocytic compartments
that are acidified to between pH 6.5 and 5.0 as a consequence
of activation of proton pumps in endosomal membranes. This
acidification activates proteolytic and other enzymes and
promotes dissociation of ligands from bound receptors. Many
membrane-associated receptors are subsequently recycled back
to the cell surface, whereas ligands are typically degraded. This
and other endocytic mechanisms are exploited by many viruses,
toxins, and symbiotic microorganisms to gain entry into cells.16
Cholera toxin produced by Vibrio cholerae comprises a
protein complex that penetrates cells by co-opting the molecular
machinery controlling clathrin-mediated endocytosis.21,22 This
toxin consists of an A (activating) subunit that activates
intracellular adenylyl cyclase activity and a pentameric B
(binding) subunit that binds with high affinity to the plasma
membrane-associated small molecule ligand, ganglioside GM1.23
The GM1 ligand is a sphingolipid comprising a pentasaccharide
linked to the lipid ceramide (N-acyl sphingosine). Each mono-
mer of the toxin B subunit binds to the pentasaccharide moiety
of GM1, which is localized to specific lipid raft domains in
cellular plasma membranes. These lipid rafts form a liquid-
ordered phase that is enriched in cholesterol and sphingolipids,
and lipid rafts play key functional roles in segregation and
concentration of membrane proteins.24-26 Many proteins linked
to cholesterol, glycosylphosphatidyl inositol (GPI), or saturated
alkyl chains are localized to lipid rafts,27 which are essential
membrane features that control activation of numerous signal
transduction pathways.28 Binding of Cholera toxin to GM1
targets this protein to lipid rafts, promoting clathrin-mediated
endocytosis through endogenous vesicular transport mecha-
nisms.21,22,29 In addition to Cholera toxin, other proteins that
undergo clathrin-mediated endocytosis upon binding to lipid
rafts include the epidermal growth factor receptor30,31 and Shiga
toxin.32
We report here the synthesis of a novel cholesterol-derived
ligand (1) termed here “Streptaphage” (one that eats streptavi-
din). This ligand associates with lipid raft subdomains in plasma
membranes of mammalian cells and efficiently promotes dose-
dependent cellular uptake of the bacterial protein streptavidin
(SA) through the mechanism of clathrin-mediated endocytosis.
This compound is structurally related to a previously reported
fluorescent cholesterylamine derivative that promotes endocytic
cellular uptake of antifluorescein antibodies and associated
Protein A from Staphlococcus aureus.33
Streptaphage (1) was designed to enable uptake of SA protein
by promoting strong noncovalent interactions with lipid rafts
of cellular plasma membranes. This compound comprises an
N-alkyl derivative of 3â-cholesterylamine34 linked to the car-
boxylate of D-biotin (vitamin H) through an 11-atom tether. The
cholesterylamine moiety was chosen as a membrane anchor
because cholesterol is a highly abundant membrane-associated
steroid that is functionally linked to endocytosis35-37 and is
covalently attached to proteins involved in signal transduc-
tion.28,38
Streptaphage (1) differs from cholesterol by substituting a
3â-N-alkylamino headgroup for the 3â-cholesterol hydroxyl
group to afford an amphiphilic molecule that is protonated at
physiological pH (7.4). This amino group was incorporated
because 3â-cholesterylamine binds much more tightly to model
membranes than cholesterol as evidenced by ca. 30-fold slower
off-rates of intervesicle transfer.34 Structurally related cationic
cholesterol derivatives that form nontoxic liposome complexes
with DNA have been used as cellular transfection reagents.39
Other synthetic cholesterylamine derivatives linked to the protein
ligands estradiol40 and mannose-6-phosphate41 have been previ-
ously described, and lipid derivatives of biotin have also been
reported.42,43
The biotin moiety of Streptaphage (1) provides a very high-
affinity ligand (Kd ∼ 100 fM)44 for SA. The SA-biotin system
is one of the strongest receptor-ligand interactions found in
nature, and this complex has been extensively studied by
thermodynamic, kinetic, and mutagenic methods in solution,44
(32) Merritt, E. A.; Hol, W. G. J. Cur. Opin. Struct. Biol. 1995, 5, 165-171.
(33) Hussey, S. L.; He, E.; Peterson, B. R. J. Am. Chem. Soc. 2001, 123, 12712-
12713.
(16) Mukherjee, S.; Ghosh, R. N.; Maxfield, F. R. Physiol. ReV. 1997, 77, 759-
803.
(17) Mellman, I. Annu. ReV. Cell DeV. Biol. 1996, 12, 575-625.
(18) Kirchhausen, T. Annu. ReV. Biochem. 2000, 69, 699-727.
(19) Schmid, S. L. Annu. ReV. Biochem. 1997, 66, 511-548.
(20) Wang, L. H.; Rothberg, K. G.; Anderson, R. G. J. Cell Biol. 1993, 123,
1107-1117.
(34) Kan, C. C.; Yan, J.; Bittman, R. Biochemistry 1992, 31, 1866-1874.
(35) Subtil, A.; Gaidarov, I.; Kobylarz, K.; Lampson, M. A.; Keen, J. H.;
McGraw, T. E. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 6775-6780.
(36) Rodal, S. K.; Skretting, G.; Garred, O.; Vilhardt, F.; van Deurs, B.; Sandvig,
K. Mol. Biol. Cell. 1999, 10, 961-974.
(21) Shogomori, H.; Futerman, A. H. J. Biol. Chem. 2001, 276, 9182-9188.
(22) Lencer, W. I.; Hirst, T. R.; Holmes, R. K. Biochim. Biophys. Acta 1999,
1450, 177-190.
(37) Ikonen, E. Curr. Opin. Cell Biol. 2001, 13, 470-477.
(38) Mann, R. K.; Beachy, P. A. Biochim. Biophys. Acta 2000, 1529, 188-
202.
(23) Spangler, B. D. Microb. ReV. 1992, 56, 622-647.
(24) Hooper, N. M. Mol. Membr. Biol. 1999, 16, 145-156.
(25) Brown, D. A.; London, E. Annu. ReV. Cell DeV. Biol. 1998, 14, 111-136.
(26) Simons, K.; Ikonen, E. Nature 1997, 387, 569-572.
(27) Simons, K.; Ikonen, E. Science 2000, 290, 1721-1726.
(28) Simons, K.; Toomre, D. Nat. ReV. Mol. Cell. Biol. 2000, 1, 31-39.
(29) de Haan, L.; Hirst, T. R. J. Nat. Toxins 2000, 9, 281-297.
(30) Couet, J.; Sargiacomo, M.; Lisanti, M. P. J. Biol. Chem. 1997, 272, 30429-
30438.
(39) Nakanishi, M.; Noguchi, A. AdV. Drug. DeliVery ReV. 2001, 52, 197-
207.
(40) Hussey, S. L.; He, E.; Peterson, B. R. Org. Lett. 2002, 4, 415-418.
(41) Barragan, V.; Menger, F. M.; Caran, K. L.; Vidil, C.; Morere, A.; Montero,
J. L. Chem. Commun. 2001, 1, 85-86.
(42) Swamy, M. J.; Marsh, D. Biochemistry 2001, 40, 14869-14877.
(43) Marsh, D.; Swamy, M. J. Chem. Phys. Lipids 2000, 105, 43-69.
(44) Stayton, P. S.; Freitag, S.; Klumb, L.; Chilkoti, A.; Chu, V.; Penzotti, J.;
To, R.; Hyre, D.; Trong Le, I.; Lybrand, T.; Stenkamp, R. Biomol. Eng.
1999, 16, 39-44.
(31) Waugh, M. G.; Lawson, D.; Hsuan, J. J. Biochem. J. 1999, 337, 591-597.
9
6266 J. AM. CHEM. SOC. VOL. 124, NO. 22, 2002

Cellular Uptake of Streptavidin Regulated by a Synthetic Ligand
A R T I C L E S
Scheme 1
Figure 1. X-ray crystal structure (PDB 1SWR) of tetrameric streptavidin
(ribbon diagrams) bound to biotin (CPK models).
in the solid state,45-47 and at solid-liquid interfaces.48-50
Structural studies have revealed that SA forms a tetramer that
binds four biotin molecules positioned in pairs at opposite faces
of the protein (Figure 1). The ease of functionalization of
biomolecules with biotin or SA has made this system extremely
useful for myriad biotechnological applications such as affinity
separations,51-54 diagnostic assays,53-55 biomolecular imaging,
and the delivery of therapeutics.56-66 In addition, numerous
conjugates of SA with fluorophores, enzymes, probes, and
particles have been reported and many are commercially
available.
Streptavidin is a particularly interesting target for enhanced
cellular uptake because this protein has clinical applications in
anticancer therapies employing monoclonal antibodies that bind
tumor-associated antigens.67-70 Pretargeting anticancer immu-
notherapeutics covalently link either SA or biotin to a mono-
clonal antibody that specifically recognizes tumor cells. Infusion
of this antibody conjugate results in binding to specific tumor-
associated antigens in vivo. Biotin or SA conjugates are
subsequently administered to deliver toxins or imaging agents
specifically to tumor cells. By uncoupling the kinetically slow
step of antibody binding to tumor antigens from the fast kinetics
of streptavidin-biotin recognition, this approach minimizes the
duration of exposure to conjugated anticancer agents.71 However,
efficient endocytosis of these antibody conjugates by tumor cells
is necessary to achieve optimal therapeutic effects, and many
tumor antigens are not rapidly internalized. Hence, the identi-
fication of compounds that improve the delivery of SA-linked
therapeutics into tumor cells may enhance the effectiveness of
these agents.72 We report here the synthesis of a nontoxic ligand
that enables regulated dose-dependent delivery of SA conjugates
into diverse mammalian cell lines through the well-characterized
mechanism of clathrin-mediated endocytosis.
(45) Freitag, S.; Le Trong, I.; Chilkoti, A.; Klumb, L.; Stayton, P. S.; Stenkamp,
R. J. Mol. Biol. 1998, 279, 211-221.
(46) Weber, P. C.; Ohlendorf, D. H.; Wendoloski, J. J.; Salemme, F. R. Science
1989, 243, 85-88.
(47) Hyre, D. E.; Le Trong, I.; Freitag, S.; Stenkamp, R. E.; Stayton, P. S. Protein
Sci. 2000, 9, 878-885.
(48) Luna-Perez, V. H.; O’Brien, M. J.; Opperman, K. A.; Hampton, P. D.;
Lopez, G. P.; Klumb, L.; Stayton, P. S. J. Am. Chem. Soc. 1999, 121,
6469-6478.
(49) Buranda, T.; Jones, G. M.; Nolan, J. P.; Keij, J.; Lopez, G. P.; Sklar, L. A.
J. Phys. Chem. B 1999, 103, 3399-3410.
(50) Wilchek, M.; Bayer, E. A. Biomol. Eng. 1999,16, 1-140.
(51) Hermanson, G. T.; Mallia, A. K.; Smith, P. K. Immobilized Affinity Ligand
Techniques; Academic Press: San Diego, CA, 1992.
(52) Bayer, E. A.; Wilchek, M. J. Chromatogr. 1990, 510, 3-11.
(53) Wilchek, M.; Bayer, E. A. Methods Enzymol. 1990, 184, 5-13, 14-45.
(54) Diamandis, E. P.; Christopoulos, T. K. Clin. Chem. 1991, 37, 625-636.
(55) Schetters, H. Biomol. Eng. 1999, 16, 73-78.
(56) Oehr, P.; Westermann, J.; Biersack, H. J. J. Nucl. Med. 1988, 29, 728-
729.
(57) Hnatowich, D. J.; Virzi, F.; Rusckowski, M. J. Nucl. Med. 1987, 28, 1294-
1302.
(58) Kalofonos, H. P.; Rusckowski, M.; Siebecker, D. A.; Sivolapenko, G. B.;
Snook, D.; Lavender, J. P.; Epenetos, A. A.; Hnatowich, D. J. J. Nucl.
Med. 1990, 31, 1791-1796.
Results and Discussion
Synthesis of Cholesterylamine Derivatives. Streptaphage
(1) was prepared from the known33 2-nitrobenzenesulfonamide-
protected cholesterylamine 2 as shown in Scheme 1. This
approach employed the amine synthesis methodology of Fuku-
yama73 to couple Boc-3-chloropropylamine74 and install the
protected primary amine of 3. Removal of the Boc carbamate
(59) Alvarez-Diez, T. M.; Polihronis, J.; Reilly, R. M. Nucl. Med. Biol. 1996,
23, 459-466.
(60) Rosebrough, S. F.; Hashmi, M. J. Pharmacol. Exp. Ther. 1996, 276, 770-
775.
(61) Ogihara-Umeda, I.; Sasaki, T.; Toyama, H.; Oda, K.; Senda, M.; Nishigori,
H. Cancer Res. 1994, 54, 463-467.
(62) Rusckowski, M.; Paganelli, G.; Hnatowich, D. J.; Magnani, P.; Virzi, F.;
Fogarasi, M.; DiLeo, C.; Sudati, F.; Fazio, F. J. Nucl. Med. 1996, 37, 1655-
1662.
(63) van Osdol, W. W.; Sung, C.; Dedrick, R. L.; Weinstein, J. N. J. Nucl.
Med. 1993, 34, 1552-1564.
(69) McDevitt, T. C.; Nelson, K. E.; Stayton, P. S. Biotechnol. Prog. 1999, 15,
391-396.
(64) Figge, J.; Bakst, G.; Weisheit, D.; Solis, O.; Ross, J. S. Am. J. Pathol.
1991, 139, 1213-1219.
(70) Yu, A.; Choi, J.; Ohno, K.; Levin, B.; Rom, W. N.; Meruelo, D. DNA Cell
Biol. 2000, 19, 383-388.
(65) Emans, N.; Biwersi, J.; Verkman, A. S. Biophys. J. 1995, 69, 716-728.
(66) Bonfils, E.; Mendes, C.; Roche, A.; Monsigny, M.; Midoux, P. Bioconj.
Chem. 1992, 3, 277-284.
(67) Wilbur, D. S.; Pathare, P. M.; Hamlin, D. K.; Stayton, P. S.; To, R.; Klumb,
L.; Buhler, K. R.; Vessella, R. L. Biomol. Eng. 1999, 16, 113-118.
(68) Stayton, P. S.; Hoffman, A. S.; Murthy, N.; Lackey, C.; Cheung, C.; Tan,
P.; Klumb, L.; Chilkoti, A.; Wilbur, D. S.; Press: O. W. J. Controlled
Release 2000, 65, 203-220.
(71) Press, O. W.; Corcoran, M.; Subbiah, K.; Hamlin, D. K.; Wilbur, D. S.;
Johnson, T.; Theodore, L.; Yau, E.; Mallett, R.; Meyer, D. L.; Axworthy,
D. Blood 2001, 98, 2535-2543.
(72) Heitner, T.; Moor, A.; Garrison, J. L.; Marks, C.; Hasan, T.; Marks, J. D.
J. Immunol. Methods 2001, 248, 17-30.
(73) Fukuyama, T.; Jow, C. K.; M., C. Tetrahedron Lett. 1995, 36, 6373-6374.
(74) Wu, S.; Lee, S.; Beak, P. J. Am. Chem. Soc. 1996, 118, 715-721.
9
J. AM. CHEM. SOC. VOL. 124, NO. 22, 2002 6267

A R T I C L E S
Hussey and Peterson
Scheme 2
Scheme 3
Figure 2. Differential interference contrast (DIC) and epifluorescence
micrographs of Jurkat lymphocytes treated with Streptaphage (1, 10 µM)
and green fluorescent SA Alexa Fluor 488 (20 µg/mL). Fluorescence
excitation (Ex.) and emission (Em.) wavelengths (nm) are explicitly shown.
Scheme 4
and acylation with commercially available D-biotinamidoca-
proate N-hydroxysuccinimidyl ester added the biotin side chain
of 4, which was further deprotected to afford 1.
To probe the importance of the protonated cholesterylamine
headgroup of Streptaphage (1), the isosteric amide 6 was
prepared as shown in Scheme 2. In addition, the length of the
linker between biotin and the steroid was investigated by
synthesis of the shorter chain variant 8 as shown in Scheme 3.
To compare the influence of amide and amine steroid head-
groups on affinity for cellular plasma membranes, the related
fluorescent probes 9 and 10 were synthesized as shown in
Scheme 4.
Ligand-Regulated Uptake of SA by Mammalian Cells. To
qualitatively evaluate whether combinations of these compounds
and SA would affect mammalian cells, compound/SA-treated
Jurkat T-lymphocytes were analyzed by epifluorescence
microscopy. As shown in Figure 2 (Panel A), cells treated with
Streptaphage (1) for 1 h followed by addition of SA conjugated
to green fluorescent Alexa Fluor 488 revealed bright fluores-
cence at the cellular periphery of >99% of viable cells within
10 min of addition of the protein conjugate. This peripheral
fluorescence was primarily localized at the cellular plasma
membrane as determined by comparison with cells treated with
previously described fluorescent plasma membrane probes.33,75
9
6268 J. AM. CHEM. SOC. VOL. 124, NO. 22, 2002

Cellular Uptake of Streptavidin Regulated by a Synthetic Ligand
A R T I C L E S
Remarkably, cells examined 4 h after protein addition revealed
bright green fluorescence nearly exclusively in defined compart-
ments in the interior of practically all viable cells, indicating
that the complex between SA and 1 was internalized (Figure 2,
Panel B). This effect was also observed when SA and 1 were
preequilibrated prior to addition of the complex to cells (Figure
2, Panel C), even though partial precipitation of the preformed
amphipathic protein-ligand complex was observed. Surpris-
ingly, SA addition to cells treated with the isosteric amide
analogue 6 or the short-chain analogue 8 did not appreciably
affect cellular fluorescence (micrographs provided in the Sup-
porting Information).
To quantify the compound dose-dependent uptake of SA
observed by microscopy, treated cells were analyzed by flow
cytometry. As shown in Figure 3 (Panel A), pretreatment of
cells with 1 for 1 h followed by addition of green fluorescent
SA for 4 h resulted in remarkable enhancements of cellular
fluorescence with significant effects observed at concentrations
of 1 as low as 100 nM. The intracellular fluorescence of cells
pretreated with 10 µM 1 was increased by over 300-fold
compared to cells treated with SA alone. Competition by
addition of D-biotin (100 µM) completely blocked uptake
promoted by 1 (1 µM), confirming that SA interacts specifically
with 1 at plasma membranes. Preequilibration of SA and 1 also
engendered dose-dependent uptake of SA, but the magnitude
of the effect was diminished by 10-fold at 10 µM 1 (Figure 3,
Panel B). This reduction was a consequence of aggregation and
precipitation of the preformed SA-1 complex as observed by
microscopy. In contrast, compounds 6 and 8 did not substantially
enhance uptake of SA by Jurkat lymphocytes. At compound
concentrations of 10 µM, the uncharged amide 6 conferred only
a 3-fold enhancement of cellular fluorescence and the shorter
chain variant 8 enhanced cellular fluorescence by only 5-fold
(Figure 3, Panels C and D). This result illustrated the importance
of the protonated secondary amine of 1, and revealed the
dramatic influence of linker length on biological activity.
Significant effects of the linker length on the affinity of other
biotin derivatives for avidin and streptavidin proteins have been
previously reported.76,77
Since cholesterol is a critical component of mammalian
plasma membranes, the cholesterylamine moiety of Streptaphage
(1) should enable delivery of SA into numerous mammalian
cell lines. To test this hypothesis, uptake of SA by the suspension
macrophage cell line THP-1, the adherent breast cancer cell line
MCF-7, and adherent chinese hamster ovary (CHO) cells was
analyzed by flow cytometry and epifluorescence microscopy.
To release adherent cell lines from flask surfaces, cells were
treated with the protease trypsin for analysis by flow cytom-
etry.78 As shown in Figure 3 (Panel E), flow cytometry revealed
significant compound-mediated enhancements of cellular fluo-
rescence in all cell lines examined. Compound-mediated cellular
fluorescence was consistently observed in >99% of viable cells,
illustrating the high efficiency of this delivery method. Exami-
nation of these cell lines by epifluorescence microscopy
confirmed that the compound-induced fluorescence was intra-
Figure 3. Analysis of ligand dose-dependent cellular uptake of SA Alexa
Fluor 488 (20 µg/mL) by flow cytometry. Each bar represents the median
fluorescence of 10 000 living cells.
cellular and resulted from protein uptake (micrographs provided
in the Supporting Information). These experiments confirmed
that incorporation of Streptaphage (1) in cellular plasma
membranes provides a highly efficient and general method for
intracellular delivery of SA protein conjugates to mammalian
cells.
(75) Creaser, S. P.; Peterson, B. R. J. Am. Chem. Soc. 2002, 124, 2444-2445.
(76) Hofmann, K.; Titus, G.; Montibeller, J. A.; Finn, F. M. Biochemistry 1982,
21, 978-984.
(77) Leary, J. J.; Brigati, D. J.; Ward, D. C. Proc. Natl. Acad. Sci. U.S.A. 1983,
80, 4045-4049.
Molecular Modeling of SA-Ligand Complexes. The effect
of linker length on the biological activity of Streptaphage (1)
(78) Daugherty, P. S.; Iverson, B. L.; Georgiou, G. J. Immunol. Methods 2000,
243, 211-227.
9
J. AM. CHEM. SOC. VOL. 124, NO. 22, 2002 6269

A R T I C L E S
Hussey and Peterson
Figure 5. Analysis by flow cytometry of Jurkat lymphocytes treated with
fluorescent derivatives of cholesterylamine. Cells were washed 3× with
1% BSA in PBS prior to analysis.
to SA, the greater linker length of 1 enables favorable interac-
tions between the protonated amine of 1 and lipid headgroups.
This protonated amino group is critical for uptake of SA as
evidenced by the dramatically diminished ability of the un-
charged amide 6 to affect the fluorescence of SA-treated cells
(Figure 3, Panel C).
Molecular Recognition of Cellular Plasma Membranes by
Structurally Related Fluorescent Probes. To further investi-
gate the importance of the protonated cholesterylamine head-
group for molecular recognition of cellular plasma membranes,
Jurkat lymphocytes were treated with fluorescent probes 10 and
11. Epifluorescence microscopy revealed that cells treated with
10 or 11 exhibited green fluorescence localized persistently and
nearly exclusively at cellular plasma membranes (micrographs
provided in the Supporting Information). However, compound
10 bearing a free amino group exhibited significantly higher
affinity for cellular plasma membranes compared with the
analogous amide as evidenced by qualitative differences in
fluorescence intensity by microscopy. Furthermore, quantitative
differences of 2-fold (1 µM) to 4-fold (0.1 µM) were measured
by flow cytometry (Figure 5). The enhanced affinity of the
protonated amine of 10 for plasma membranes may derive from
favorable electrostatic interactions with anionic phospholipid
headgroups.80 However, the cholesterol hydroxyl group is
thought to form favorable H-bonding interactions with amide
and alcohol functionality of sphingolipids in cholesterol-rich
lipid rafts,27 and these H-bonding interactions should be
strengthened by the presence of a protonated amino substituent.
Comparison of fluorescent probes 10 and 11 with SA-bound 1
and 6 suggests that the enhanced affinity for plasma membranes
engendered by a free amino group contributes significantly to
the stability of membrane-bound complexes between SA and
1. However, it is remarkable that amide 6 does not appreciably
promote association between SA and cellular plasma membranes
(Figure 3, Panel C). This result may relate to the much longer
linker between the biotin derivatives and cholesterylamine
compared with the fluorescent probes. The short hydrophilic
thiourea linker between fluorescein and cholesterylamine pre-
sumably interacts favorably with the hydrophilic lipid head-
groups of the plasma membrane. In contrast, the long relatively
hydrophobic linker in the biotin derivatives may magnify
differences in membrane affinity between the amide and amine
by positioning the hydrophilic biotin moiety distal to hydrophilic
headgroups of membrane lipids.
Figure 4. Ribbon and electrostatic solvent-exposed surface models of SA
bound to Macromodel-minimized ligands (CPK models) inserted into a
model membrane bilayer (CPK models). Panel A: SA bound to Streptaphage
(1). Panel B: SA bound to the short-chain analogue 8.
was examined by molecular modeling of SA-ligand complexes
interacting with model membranes. Modeling employed sub-
stitution of two bound biotin ligands on one face of the
tetrameric X-ray structure of SA with Macromodel (v. 6.5)
Amber*-minimized79 models of 1 and 8 in extended chain
conformations. As shown in Figure 4, the area proximal to the
ligand-binding site exhibits a concave topology as a consequence
of nearby solvent-exposed loops on the protein exterior. This
revealed that the binding site of SA allows complete penetration
of the cholesterylamine moiety of 1 into a flat model membrane
bilayer constructed from minimized phosphatidylcholine lipids
(Figure 4, Panel A). In contrast, the shorter linker of 8 prohibits
complete insertion of the steroid into the model membrane
(Figure 4, Panel B). These results indicated that when bound
(79) Mohamadi, F.; Richards, N. G. J.; Guida, W. C.; Liskamp, R.; Lipton, M.;
Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C. J. Comput. Chem.
1990, 11, 440-467.
(80) Roy, M. O.; Leventis, R.; Silvius, J. R. Biochemistry 2000, 39, 8298-
8307.
9
6270 J. AM. CHEM. SOC. VOL. 124, NO. 22, 2002

Cellular Uptake of Streptavidin Regulated by a Synthetic Ligand
A R T I C L E S
previously described.35-37,82 Treatment with these compounds
predominantly trapped SA at the cellular plasma membrane,
which indicated that uptake of SA was controlled by a clathrin-
dependent mechanism.
Cholera toxin invades cells by binding to the lipid raft-
associated ganglioside GM1 in cellular plasma membranes.23,29
Hence, fluorescent derivatives of the Cholera toxin B subunit
have been employed as markers of lipid rafts.21 To investigate
whether green fluorescent cholesterylamine 10 or the SA-1
complex might similarly colocalize with lipid rafts, confocal
fluorescence microscopy was employed to compare localization
in cellular plasma membranes with the membrane distribution
of red fluorescent Cholera toxin B subunit. As shown in Figure
7, compound 10 and the SA-1 complex exhibited a heteroge-
neous distribution in plasma membranes of Jurkat lymphocytes.
This distribution substantially colocalized with Cholera toxin,
consistent with the cholesterylamine moiety of 10 and 1
associating with lipid rafts. These results indicate that both 10
and complexes between SA and 1 preferentially bind to the
sphingolipid and cholesterol-rich liquid-ordered phase of lipid
rafts in cellular plasma membranes.
Streptaphage (1) Exhibits Minimal Toxicity to Cells in
Culture. Streptavidin is sufficiently nontoxic for clinical
applications such as pretargeting immunotherapeutics.67-71
Additionally, structurally related cationic cholesterol derivatives
exhibit low toxicity and have been used as transfection
reagents.39 To evaluate the toxicity of Streptaphage (1) alone
and in combination with SA, compound-treated Jurkat lympho-
cytes were analyzed by flow cytometry. As shown in Figure 8,
Streptaphage (1) does not adversely affect the viability of
lymphocytes even at concentrations as high as 100 µM.
Similarly, SA protein alone at 0.2 mg/mL in cell culture was
nontoxic. Under conditions that promote uptake of SA by over
300-fold (10 µM 1, 0.02 mg/mL SA), only a slight reduction
in viability was observed. However, cellular viability was
dramatically affected by coadministration of 1 and SA at 10-
fold higher concentrations, which may result from rupture of
cellular membranes due to overwhelmingly excessive protein
delivery.
Figure 6. DIC and epifluorescence micrographs of Jurkat lymphocytes
treated with Streptophage (1, 10 µM) for 1 h followed by green fluorescent
SA Alexa Fluor 488 (20 µg/mL) for 4 h. Panels A-C: Cells colabeled
with red fluorescent Lysotracker Red dye. D, E: Cells pretreated with
chlorpromazine (100 µM). F, G: Cells pretreated with hypertonic sucrose
(400 mM). H, I: Cells pretreated with methyl-â-cyclodextrin (10 mM).
Streptaphage (1) Enables SA to Associate with Lipid Rafts
and Access a Clathrin-Mediated Endocytosis Pathway.
Streptaphage (1) was found to target SA conjugates to defined
intracellular compartments. These compartments were identified
as acidic endosomes by addition of the acid-sensitive red
fluorescent endosome probe Lysotracker Red (Figure 6, Panels
A-C), indicating that the protein is internalized by an endocytic
mechanism. To examine whether uptake of SA involved
clathrin-mediated endocytosis, specific inhibitors of this pathway
were investigated. Cationic amphipathic drugs such as chlor-
promazine specifically block clathrin-mediated endocytosis by
disrupting the assembly/disassembly of clathrin associated with
coated pits and endosomes.20 In addition, cells treated with
hypertonic sucrose exhibit defects in clathrin-coated lattices that
block clathrin-mediated endocytosis.81 Uptake of SA promoted
by 1 was inhibited by chlorpromazine (Figure 6, Panels D and
E) and hypertonic sucrose (Figure 6, Panes F and G). In addition,
cellular cholesterol was also critical for ligand-regulated uptake
of SA. Cells pretreated with methyl-â-cyclodextrin to deplete
plasma membrane cholesterol levels trapped SA at the plasma
membrane. Similar effects of this cholesterol-binding compound
on lipid rafts and clathrin-mediated endocytosis have been
Summary and Conclusions
We report the synthesis of a novel small-molecule ligand
termed Streptaphage (1) that enables efficient uptake of SA by
mammalian cells by hijacking the molecular machinery used
by cells to organize cellular membranes. This ligand comprises
biotin linked to cholesterylamine, which promotes clathrin-
mediated endocytosis by anchoring SA in lipid raft subdomains
of cellular plasma membranes. Streptaphage (1) associates with
lipid rafts by functionally mimicking cholesterol, which packs
with sphingolipids through hydrophobic and hydrogen bonding
interactions to form the liquid ordered phase that characterizes
lipid rafts.24-27,83
The Cholera toxin B subunit gains access to the cellular
interior by similarly binding a lipid raft-associated small
molecule. This pentameric protein penetrates cellular membranes
(82) Kilsdonk, E. P. C.; Yancey, P. G.; Stoudt, G. W.; Bangerter, F. W.; Johnson,
W. J.; Phillips, M. C.; Rothblat, G. H. J. Biol. Chem. 1995, 270, 17250-
17256.
(83) Veiga, M. P.; Arrondo, J. L. R.; Goni, F. M.; Alonso, A.; Marsh, D.
Biochemistry 2001, 40, 2614-2622.
(81) Heuser, J. E.; Anderson, R. G. W. J. Cell Biol. 1989, 108, 389-400.
9
J. AM. CHEM. SOC. VOL. 124, NO. 22, 2002 6271

A R T I C L E S
Hussey and Peterson
Figure 8. Analysis of toxicity of 1 and/or SA by flow cytometry. Jurkat
lymphocytes were treated with compounds for 24 h and viability quantified
by forward and side-scatter dot plots. Dead cells were counterstained with
propidium iodide.
Figure 7. Confocal fluorescence microscopy of Jurkat lymphocytes treated
with red fluorescent Alexa Fluor 594-conjugated Cholera toxin B subunit
(8 µg/mL) for 5 min. Panels A-D: Cells pretreated with green fluorescent
cholesterylamine 10 (10 µM) for 1 h. Panels E-H: Cells pretreated with
1 (10 µM) for 1 h, followed by addition of green fluoresent SA Alexa Fluor
488 (20 µg/mL) for 5 min. White arrows illustrate points of colocalization
at the plasma membrane. Overlap images (D and H) display colocalization
as yellow pixels.
Figure 9. Model of a plasma membrane lipid raft segment containing
Streptaphage (1) and ganglioside GM1. The cholesterylamine moiety of 1
partitions preferentially into the liquid-ordered phase of the exoplasmic
(outer) leaflet, which primarily comprises cholesterol, sphingomyelin, and
glycosphingolipids. The inner leaflet predominantly contains cholesterol
and saturated glycerolipids such as phosphatidylserine.
by binding the raft-associated natural product ganglioside GM1.
This natural product is a sphingolipid that comprises ceramide
bearing a pentasaccharide headgroup (Figure 9). Binding of
Cholera toxin to GM1 enables this protein to localize in lipid
rafts and undergo clathrin-mediated endocytosis by accessing
endogenous mechanisms of vesicular transport.21,22,29 Because
lipid rafts contain cholesterol packed with sphingolipids,
synthetic analogues of GM1 replacing the ceramide lipid moiety
with a cholesterol ester have yielded non-natural GM1 mimics
more potent than the natural product in mediating actions
of Cholera toxin.84 This previously reported result and the
results described herein illustrate the functional equivalence of
sphingolipid and cholesterol derivatives in mediating protein
uptake through this mechanism.
Analysis by confocal fluorescence microscopy revealed that
the SA-1 complex and Cholera toxin-GM1 colocalize in lipid
rafts. Furthermore, both proteins are internalized by clathrin-
mediated endocytosis. Thus, uptake of the SA-1 complex
functionally mimics the initial steps of cellular penetration
employed by the Cholera toxin-GM1 complex. A proposed
model of Streptaphage (1) associated with a portion of a lipid
raft27 is shown in Figure 9. In this model, 1 is shown in the
(84) Pacuszka, T.; Bradley, R. M.; Fishman, P. H. Biochemistry 1991, 30, 2563-
2570.
9
6272 J. AM. CHEM. SOC. VOL. 124, NO. 22, 2002

Cellular Uptake of Streptavidin Regulated by a Synthetic Ligand
A R T I C L E S
exoplasmic leaflet of the bilayer because 3â-cholesterylamine
is unequally distributed in model liposomal membranes and
favors the outer leaflet.34
of production of lipid raft components.88,89 Lipid raft-rich plasma
membranes of MDR cancers may permit selective cellular
labeling by Streptaphage (1) or related compounds for targeted
delivery of antitumor agents.
Streptaphage (1) was found to exhibit remarkable low toxicity
to cells in culture, and SA protein is sufficiently nontoxic for
clinical applications such as pretargeting immunotherapeutics.
These results suggest that ligand-regulated uptake of SA
conjugates and fusion proteins may have applications in tumor
therapy. Many proteins such as serum albumin tend to ac-
cumulate in solid tumors due to the enhanced microvasculature
of tumors and the lack of functional lymphatic drainage systems
in tumor tissue.85,86 Hence, endosome-activated antitumor agents
conjugated to serum albumin can exhibit some selectivity for
solid tumors and promote tumor regression in nude mice.85,87
SA-linked cytotoxins that become activated upon protolysis or
acidification in endosomes may similarly accumulate in tumor
tissue. Subsequent administration of synthetic ligands that
promote endocytosis may trigger tumor regression. Streptaphage
(1)-regulated cytotoxicity of SA proteins linked to endosome-
activated toxins may be particularly effective because the fate
of internalized SA protein appears to be irreversible entrapment
in endosomal compartments due to the high affinity of bound
cholesterylamine for membranes. This approach of labeling lipid
rafts of cell membranes with small molecules that internalize
SA toxins has the potential to be useful for targeting multidrug
resistant (MDR) cancers, which exhibit distinct plasma mem-
brane compositions as a consequence of dramatic upregulation
The synthesis of Streptaphage (1) analogues bearing acid-
labile or other reactive linkers may facilitate escape of SA
proteins from acidic endosomes into the cytosol or nucleus. This
approach may be useful for stimulation of immune responses
by regulating uptake and processing of SA-linked protein
antigens in antigen presenting cells such as macrophages,
B-cells, and dendridic cells. Streptaphage (1) may also enable
regulated gene delivery since avidin conjugated to the DNA-
binding polymer polyethyleneimine has previously been em-
ployed for cellular transfection.15 This general strategy of
coupling protein ligands to molecules that associate with lipid
rafts has potential for regulated delivery of diverse ligand-
binding proteins.
Acknowledgment. We thank the NIH (CA83831) for finan-
cial support. S.L.H. thanks the DOD for a predoctoral fellow-
ship.
Supporting Information Available: Experimental procedures,
compound characterization data, and additional micrographs
(PDF). This material is available free of charge via the Internet
at http://pubs.acs.org.
JA0258733
(85) Kratz, F.; Roth, T.; Fichiner, I.; Schumacher, P.; Fiebig, H. H.; Unger, C.
J. Drug Target. 2000, 8, 305-318.
(86) Maeda, H.; Wu, J.; Sawa, T.; Matsumura, Y.; Hori, K. J. Controlled Release
2000, 65, 271-284.
(88) Lavie, Y.; Fiucci, G.; Liscovitch, M. AdV. Drug DeliVery ReV. 2001, 49,
317-323.
(89) Lavie, Y.; Liscovitch, M. Glycoconj. J. 2000, 17, 253-259.
(87) Kratz, F.; Beyer, U. Drug DeliVery 1998, 5, 281-299.
9
J. AM. CHEM. SOC. VOL. 124, NO. 22, 2002 6273