Design and synthesis of bis-biotin-containing reagents for applications utilizing monoclonal antibody-based pretargeting systems with streptavidin mutants

Article abstract of DOI:10.1021/bc100030q

Previous studies have shown that pretargeting protocols, using cancer-targeting fusion proteins, composed of 4 anti-CD20 single chain Fv (scFv) fragments and streptavidin (scFv₄-SAv), followed by a biotinylated dendrimeric N-acetyl-galactosamine blood clearing agent (CA), 1, then a radiolabeled DOTA-biotin derivative (a monobiotin), 3a, can provide effective therapy for lymphoma xenografts in mouse models. A shortcoming in this pretargeting system is that endogenous biotin may affect its efficacy in patients. To circumvent this potential problem, we investigated a pretargeting system that employs anti-CD20 scFv₄-SAv mutant fusion proteins with radioiodinated bis-biotin derivatives. With that combination of reagents, good localization of the radiolabel to lymphoma tumor xenografts was obtained in the presence of endogenous biotin. However, the blood clearance reagents employed in the studies were ineffective, resulting in abnormally high levels of radioactivity in other tissues. Thus, in the present investigation a bis-biotin-trigalactose blood clearance reagent, 2, was designed, synthesized, and evaluated in vivo. Additionally, another DOTA-biotin derivative (a bis-biotin), 4a, was designed and synthesized, such that radiometals (e.g., ¹¹¹In, ⁹⁰Y, ¹⁷⁷Lu) could be used in the pretargeting protocols employing scFv₄-SAv mutant fusion proteins. Studies in mice demonstrated that the CA 2 was more effective than CA 1 at removing [¹²⁵I]scFv₄-SAv-S45A mutant fusion proteins from blood. Another in vivo study compared tumor targeting and normal tissue concentrations of the new reagents (2 and [¹¹¹In]4b) with standard reagents (1 and [¹¹¹In]3b) used in pretargeting protocols. The study showed that lymphoma xenografts could be targeted in the presence of endogenous biotin when anti-CD20 fusion proteins containing SAv mutants (scFv ₄-SAv-S45A or scFv₄-SAv-Y43A) were employed in combination with CA 2 and [¹¹¹In]4b. Importantly, normal tissue concentrations of [¹¹¹In]4b were similar to those obtained using the standard reagents (1 and [¹¹¹In]3b), except that the blood and liver concentrations were slightly higher with the new reagents. While the reasons for the higher blood and liver concentrations are unknown, the differences in the galactose structures of the clearance agents 1 and 2 may play a role.

Full text of DOI:10.1021/bc100030q

Bioconjugate Chem. 2010, 21, 1225–1238
1225
Design and Synthesis of Bis-Biotin-Containing Reagents for Applications
Utilizing Monoclonal Antibody-Based Pretargeting Systems with Streptavidin
Mutants
D. Scott Wilbur,*,†,§ Steven I. Park,†,‡,# Ming-Kuan Chyan,§ Feng Wan,‡,§ Donald K. Hamlin,§ Jaideep Shenoi,#
Yukang Lin,# Shani M. Wilbur,# Franz Buchegger,‡,# Anastasia Pantelias,‡,# John M. Pagel,|,# and
Oliver W. Press|,⊥,#
Departments of Radiation Oncology, Medicine, and Biological Structure, University of Washington, Seattle, Washington, and
Clinical Research Division, Fred Hutchinson Cancer Research Center, Seattle, Washington. Received January 15, 2010;
Revised Manuscript Received May 15, 2010
Previous studies have shown that pretargeting protocols, using cancer-targeting fusion proteins, composed of 4 anti-
CD20 single chain Fv (scFv) fragments and streptavidin (scFv4-SAv), followed by a biotinylated dendrimeric N-acetyl-
galactosamine blood clearing agent (CA), 1, then a radiolabeled DOTA-biotin derivative (a monobiotin), 3a, can provide
effective therapy for lymphoma xenografts in mouse models. A shortcoming in this pretargeting system is that endogenous
biotin may affect its efficacy in patients. To circumvent this potential problem, we investigated a pretargeting system
that employs anti-CD20 scFv4-SAv mutant fusion proteins with radioiodinated bis-biotin derivatives. With that
combination of reagents, good localization of the radiolabel to lymphoma tumor xenografts was obtained in the presence
of endogenous biotin. However, the blood clearance reagents employed in the studies were ineffective, resulting in
abnormally high levels of radioactivity in other tissues. Thus, in the present investigation a bis-biotin-trigalactose blood
clearance reagent, 2, was designed, synthesized, and evaluated in vivo. Additionally, another DOTA-biotin derivative
(a bis-biotin), 4a, was designed and synthesized, such that radiometals (e.g., 111In, 90Y, 177Lu) could be used in the
pretargeting protocols employing scFv4-SAv mutant fusion proteins. Studies in mice demonstrated that the CA 2 was
more effective than CA 1 at removing [125I]scFv4-SAv-S45A mutant fusion proteins from blood. Another in vivo study
compared tumor targeting and normal tissue concentrations of the new reagents (2 and [111In]4b) with standard reagents
(1 and [111In]3b) used in pretargeting protocols. The study showed that lymphoma xenografts could be targeted in the
presence of endogenous biotin when anti-CD20 fusion proteins containing SAv mutants (scFv4-SAv-S45A or scFv4-
SAv-Y43A) were employed in combination with CA 2 and [111In]4b. Importantly, normal tissue concentrations of
[
111In]4b were similar to those obtained using the standard reagents (1 and [111In]3b), except that the blood and liver
concentrations were slightly higher with the new reagents. While the reasons for the higher blood and liver concentrations
are unknown, the differences in the galactose structures of the clearance agents 1 and 2 may play a role.
as a treatment for patients with B-cell lymphomas (1). While it
seems likely that changes being evaluated in the current RIT
treatment regimens will increase efficacy and/or decrease
toxicity (2-4), newer approaches to RIT, such as cancer
“pretargeting” (5-7), may provide more significant improve-
ments in the treatment of lymphoma. One of the pretargeting
approachesbeinginvestigatedusesmonoclonalantibody-streptavidin
(mAb-SAv) conjugates (8-10), or related scFv4-SAv fusion
proteins, in protocols where they are administered in an initial
step, followed in subsequent steps by administration of a blood
clearance agent, then a biotin derivative labeled with a
therapeutic radionuclide. In preclinical investigations, it has been
conclusively shown that pretargeting protocols in mouse models
using mAb-SAv conjugates or fusion proteins, reactive with
CD20 (11), or other lymphoma cell surface antigens (12, 13),
either alone or in combinations (14), provides a more efficacious
therapy than obtained with conventional RIT.
INTRODUCTION
Radioimmunotherapy (RIT1) employing radiolabeled anti-
CD20 monoclonal antibodies (mAb) has proven to be efficacious
* Address correspondence to D. Scott Wilbur, Ph.D., Department
of Radiation Oncology, University of Washington, Box 355016, 616
N.E. Northlake Place, Seattle, WA 98105, Phone: 206-616-9246,
E-mail: dswilbur@u.washington.edu.
† These investigators contributed equally to this manuscript.
‡ Current Addresses: Steven I. Park, M.D., Division of Hematology/
Oncology, University of North Carolina, Chapel Hill, NC 27599-7305;
Feng (Connie) Wan, Ph.D., Intellectual Ventures, 11235 SE 6th St.,
Suite A 200, Bellevue, WA 98004; Franz Buchegger, M.D., University
Hospital of Lausanne, CH-1011 Lausanne, Switzerland; Anastasia
Pantelias, Bill and Melinda Gates Foundation, Seattle, WA.
§ Department of Radiation Oncology, University of Washington.
| Department of Medicine, University of Washington.
⊥ Department of Biological Structure, University of Washington.
# Fred Hutchinson Cancer Research Center.
Although the mAb-SAv/radiolabeled biotin pretargeting ap-
proach benefits from the very high affinity of the biotin-SAv
binding pair, a potential limitation in the approach is the fact
that patient serum contains significant amounts of biotin (15).
On the basis of animal studies (16), it is possible that
endogenous biotin in patient blood can bind with administered
mAb-SAv conjugates or fusion proteins to block the binding
of subsequently administered blood clearance and radiolabeled
biotin derivatives. To circumvent this problem, new mAb-SAv/
1Abbreviations: ASOR, asialoorosomucoid protein; CA, blood clearance
agent; CDI, carbonyldiimidazole; DCC, N,N′-dicyclohexylcarbodiimide; DOTA,
1,4,7,10-tetraazacyclododecane tetraacetic acid; EDC, 1-ethyl-3-(3-dimethy-
laminopropyl)carbodiimide; FP, fusion protein; HOBT, 1-hydroxybenztriazole;
mAb, monoclonal antibody; PEG, poly(ethylene glycol); Ra/Ni, Raney-Nickel;
RIP, relative inhibitory power; RIT, radioimmunotherapy; RB, radiolabeled
biotin; SAv, streptavidin; scFv, single chain fusion of light and heavy chains
in variable region of mAb; TCDI, thiocarbonyldiimidazole; TFA, trifluoroacetic
acid; TFP-OTFA, 2,3,5,6-tetrafluorophenyl-O-trifluoroacetate; WT, wild-type
(native) protein sequence
10.1021/bc100030q  2010 American Chemical Society
Published on Web 07/02/2010

1226 Bioconjugate Chem., Vol. 21, No. 7, 2010
Wilbur et al.
Figure 1. Chemical structures for monobiotin- and bis-biotin-blood clearance agents (1, 2) and nonmetalated DOTA-biotin derivatives (3a, 4a).
biotin binding pairs were designed and prepared. Site-directed
mutations of the biotin-binding pocket in streptavidin (SAv)
were performed to prepare SAv mutants that have decreased
binding with biotin (17). Subsequently, those SAv mutants were
used to prepare scFv4-SAv mutant fusion proteins (18). It was
hypothesized that the decrease in biotin binding in the SAv
mutants could be offset by using bis-biotin derivatives, which
would have increased binding due to dual binding with a single
SAv molecule (19) and an avidity effect (20). Therefore, a series
of bis-biotin derivatives were synthesized and their binding with
SAv mutants was assessed. Importantly, it was demonstrated
that the bis-biotin derivatives could bind with the SAv mutants
in the presence of biotin (21). Later pretargeting studies
demonstrated that tumor xenografts could be effectively targeted
using scFv4-SAv mutant fusion proteins, a blood clearance agent,
and a radioiodinated bis-biotin derivative, even when the mice
were placed on a normal (biotin-containing) diet (22). While
the studies successfully demonstrated that a mAb-SAv/biotin-
based pretargeting approach targeted tumors in the presence of
endogenous biotin, the concentrations of radioactivity in normal
tissues were much higher than those obtained in other pretar-
geting studies. The higher normal tissue concentrations appeared
to be caused by a lack of efficient blood clearance when using
either of the two blood clearing agents studied. Those agents
included [1] a monobiotin-hexadecyl-N-acetylgalactose deriva-
tive, 1, (Figure 1) and [2] a galactosylated and bis-biotinylated
human serum albumin derivative (22).
the antibody-SAv mutant fusion protein. Additionally, a bis-
biotin derivative containing the metal chelation moiety DOTA,
4a, was designed and synthesized due to our interest in using
radiometals, such as indium-111 (111In) and yttrium-90 (90Y) in
pretargeting protocols. The bis-biotin-DOTA derivative, 4a, was
designed to be similar to the bis-biotin-iodobenzoyl derivative
previously shown to bind selectively with SAv mutants in the
presence of biotin (21). In addition to syntheses of the new
reagents, studies were conducted in mice to demonstrate that
the new reagents functioned as designed. A study was conducted
in mice to compare the blood clearance of a scFv4-SAv mutant
fusion in the presence of endogenous biotin when the clearing
agents 1 and 2 were employed. Another study was conducted
to evaluate whether 111In-labeled 4a, [111In]4b, could successfully
target tumors in athymic mice bearing Ramos lymphoma. In
that study, an anti-CD20 scFv4-SAv mutant fusion protein was
administered, followed by blood clearance with 2, then admin-
istration of an 111In-labeled biotin derivative. For comparison,
the biodistribution of [111In]4b alone was evaluated in mice.
The results obtained in the synthesis of 2 and 4a, preparation
of radiolabeled biotin derivatives, [111In]3a/[111In]4a, and use
of those reagents in animal studies are described herein.
EXPERIMENTAL PROCEDURES
General. Chemicals purchased from commercial sources were
analytical grade or better, and were used without further
purification. Di-tert-butyl-dicarbonate, 2,2′-(ethylenedioxy)bis-
(ethylamine), nitromethanetrispropionic acid, acetobromo-R-D-
galactose (2,3,4,6-tetra-O-acetyl-R-D-galactopyranosyl bromide),
3-iodopropionic acid, 1,1′-carbonyldiimidazole (CDI), 1,1′-
thiocarbonyldiimidazole (TCDI), Amberlite IR120 (hydrogen
form), and most other chemicals were obtained from Sigma-
Aldrich (Milwaukee, WI). L-Aspartic acid R-t-butyl ester, 10,
In the current studies, two new reagents, 2 and 4a (Figure
1), were synthesized for use in pretargeting protocols with
scFv4-SAv mutant fusion proteins. A new biotin-containing
blood clearance agent, 2, was designed and synthesized to
evaluate the hypothesis that the unacceptably high concentrations
of a radioiodinated bis-biotin derivative in normal tissues
previously obtained was due to ineffective blood clearance of

Bis-Biotin Reagents for Pretargeting
Bioconjugate Chem., Vol. 21, No. 7, 2010 1227
was obtained from Novabiochem (AMD Biosciences, Inc., San
Diego, CA). D-Biotin, 9, N-Boc-trioxatridecanediamine, 16, and
N-(phthalimido)amino-dPEG4-acid, 19, were gifts from Quanta
BioDesign, Powell, OH; aminobenzyl-DOTA (S-2-(4-Ami-
nobenzyl)-1,4,7,10-tetraazacyclododecane tetraacetic acid), 31,
was obtained from Macrocyclics, Inc. (Dallas, TX). Biotin-
sarcosine-aminobenzyl-DOTA, 3a, was prepared as previously
described (23) (currently available from Macrocyclics, Inc.,
Dallas, TX). 2,3,5,6-Tetrafluorophenyl trifluoroacetate, TFP-
OTFA, was prepared as previously described (24). Sephadex
LH-20 gel filtration resin was obtained from Fluka (Sigma-
Aldrich, Milwaukee, WI). Melting points were obtained in open
capillary tubes on a Mel-Temp II apparatus with a Fluke 51
K/J electronic thermometer and are uncorrected.
HPLC effluent. The HPLC equipment used consisted of a
Hewlett-Packard quaternary 1050 gradient pump, Waters 601
UV detector, and a Beckman model 170 radioisotope detector.
Analysis of the HPLC data was conducted on Hewlett-Packard
HPLC ChemStation software. The separations used a C-18
column (Altima C-18, 5 mm, 250 × 4.5 mm; Alltech, Deerfield,
IL) with a gradient eluting at a flow rate of 1 mL/min. The
gradient was composed of MeOH as solvent A and 0.1% HOAc
in water as solvent B. The gradient started at 60% B, which
was held for 2 min, then the percent of MeOH was increased
linearly to 100% over the next 10 min. After reaching 100%
MeOH, it was held there for 8 min.
5-N-(tert-Butyloxycarbonyl)aminoisophthalate-1,3-bis-O-
(2,3,5,6-tetrafluorophenyl) Ester, 6. This compound was
prepared from aminoisophthalic acid, 5, by reaction with di-
tert-butyl dicarbonate in basic DMF/H2O solution, followed by
esterification with tetrafluorophenol and EDC in DMF as
previously reported (21).
N-(tert-Butyloxycarbonyl)-3,6-dioxa-1,8-octanediamine,
7. To a solution of 2,2′-(ethylenedioxy)bis(ethylamine) (403 mL,
2760 mmol) in 200 mL of dichloromethane, di-tert-butyl
dicarbonate (20 g, 91.6 mmol) with 200 mL of dichloromethane
was added dropwise over 4 h. After the addition was completed,
the reaction solution was allowed to stir at room temperature
for 1 h. Dichloromethane was evaporated by rotoevaporator
under vacuum, and then the residue was dissolved in water (200
mL) and washed with hexanes (2 × 200 mL). The aqueous
solution was then extracted with dichloromethane (3 × 100 mL).
The fractions of dichloromethane were combined and washed
with water (3 × 100 mL), dried over anhydrous Na2SO4, and
then evaporated under vacuum on a rotoevaporator to yield
14.11 g (62%) of 7 as a colorless oil (used without further
purification). 1H NMR (CDCl3, 500 MHz): δ 1.45 (s, 9H), 1.57
(s, 2H), 2.88 (t, J ) 5.4 Hz, 2H), 3.32 (q, J ) 5.2 Hz, 2H),
3.54 (q, J ) 5.5 Hz, 4H), 3.62 (s, 4H), 5.18 (br s, 1H). HPLC:
tR ) 4.1 min. (Alternate syntheses have been previously
reported 26, 27).
Radioactive Materials. All radioactive materials were handled
according to approved protocols at the University of Washington
and Fred Hutchinson Cancer Research Center. Standard methods
for safely using radioactive materials were employed (25).
Sodium [125I]iodide and [111In]indium chloride were purchased
from PerkinElmer Life Sciences (Billerica, MA). Measurement
of 125I and 111In was conducted on a Capintec CRC-15R
Radioisotope Calibrator using calibration numbers provided by
the manufacturer. Samples containing 125I or 111In were counted
in a Packard Cobra II gamma counter (PerkinElmer Life and
Analytical Services, Waltham, MA). Radioactivity counts were
imported into an Excel Spreadsheet (Microsoft Corp., Redmond,
WA) where calculations of percent injected dose (%ID) and
percent injected dose per gram (%ID/g) were made.
1
Spectral Analyses. H NMR spectra were obtained on a
Bruker AV 500 (500 MHz) instrument. Proton chemical shifts
are expressed as ppm using tetramethylsilane as an internal
standard (δ ) 0.0 ppm). Low-resolution mass spectral (LRMS)
and high-resolution mass spectral (HRMS) data were obtained
on a Bruker APEX III 47e Fourier transform mass spectrometer
using electrospray ionization. For analysis, the samples were
dissolved in 50/50 MeOH/H2O and were introduced by an
integral syringe infusion pump (Cole Parmer series 74900).
N,N′-Bis(N-(tert-butyloxycarbonyl)-N-(8-amino-3,6-diox-
aoctanylamino)-5-(N-tert-butyloxycarbonyl)aminoisophtha-
lamide, 8. A 1.25 g (5.09 mmol) quantity of 7 was added to a
solution containing 6 (1.50 g, 2.42 mmol), Et3N (1.01 mL, 7.27
mmol), and anhydrous CH3CN (150 mL). The resultant solution
was stirred at room temperature for 30 min, and then evaporated
to dryness under vacuum on a rotoevaporator. The crude product
was purified by silica gel column (2.5 cm ×25 cm) eluting with
10% MeOH/EtOAc to yield 1.57 g (81%) of 8 as a colorless,
tacky solid. 1H NMR (CDCl3, 500 MHz): δ 1.41 (s, 18H), 1.53
(s, 9H), 3.28-3.31 (m, 4H), 3.55 (t, J ) 5.3 Hz, 4H), 3.64 (s,
8H), 3.67 (s, 8H), 5.23 (s, 2H), 7.10 (s, 2H), 7.87 (s, 1H), 7.99
(s, 1H), 8.17 (s, 2H). HRMS (ES+) calcd for C35H60N5O12
(M+H)+: 764.4052. Found: 764.4030. HPLC: tR ) 13.7 min.
2-(N-Biotinoyl)-L-aspartate-1-O-tert-butyl Ester, 11. This
compound was prepared from biotin, 9, by reaction with 2,3,5,6-
tetrafluorophenyl trifluoroacetate (TFP-OTFA) in an initial step,
followed by reaction of the intermediate biotin tetrafluorophenyl
ester with L-aspartic acid R-tert-butyl ester, 10, as previously
reported (28).
4-(O-Tetrafluorophenyl)-2-(N-Biotinoyl)-L-aspartate-1-O-
tert-butyl Ester, 12. This compound was prepared by reaction
of 11 with TFP-OTFA following a previously reported proce-
dure (28).
N,N′-Bis(N-(biotinyl)-ꢀ-N-(tert-butyloxycarbonyl)-N-(8-
amino-3,6-dioxaoctanyl)aspartamidyl)-5-aminoisophthala-
mide, 13. A 1.50 g (2.02 mmol) quantity of 8 was stirred with
trifluoroacetic acid (10 mL) at room temperature for 10 min.
Excess trifluoroacetic acid was evaporated under a stream of
argon, and then the residue was washed with ethyl acetate (2
Chromatography. All synthetic reactions were monitored
by HPLC. Samples were assessed on a system that contained a
Hewlett-Packard quaternary 1050 gradient pump, a variable
wavelength UV detector (254 nm), and a Varex ELSD MKIII
or an Alltech ELSD 2000 evaporative light-scattering detector.
Analyses of HPLC data were conducted using Hewlett-Packard
HPLC ChemStation software. Reversed-phase HPLC chroma-
tography was carried out on an Alltech Altima C-18 column (5
µm, 250 × 4.5 mm) using a gradient solvent system at a flow
rate of 1 mL/min. A gradient of MeOH and H2O/0.1% HOAc
was used. Starting with 40% MeOH, the initial solvent mixture
was held for 2 min, then the gradient was increased to 100%
MeOH over the next 10 min, then held at 100% MeOH for 8
min. Retention times (tR) under these conditions for compounds
are provided with the compound experimental.
Some synthesized compounds were purified from crude
reaction mixtures using a Biotage SP Flash Purification System
(Charlottesville, VA) on a reversed-phase C18 FLASH 25+M
column. The purification used a gradient mixture composed of
MeOH and 0.1% aqueous HOAc solution (pH 3.25). The
gradient started with 25% MeOH/75% HOAc/H2O, and that
mixture was held for 2 min. Following that, the % MeOH was
increased linearly to 100% over the next 10 min, then held at
100% for 8 min. Fractions were collected based on UV detection
at 215 or 254 nm. Fractions containing pure products were
determined by analytical HPLC and were combined, solvent
evaporated, and isolated to provide the yields listed.
The 111In-labeling reactions of biotin derivatives 3a and 4a,
to prepare [111In]3b and [111In]4b, were followed by radio-
HPLC, and the products were purified by isolation from the

1228 Bioconjugate Chem., Vol. 21, No. 7, 2010
Wilbur et al.
× 40 mL) and dried under vacuum for 3 h. The deprotected
aniline derivative was dissolved in anhydrous DMF (20 mL),
and then Et3N (1.13 mL, 8.09 mmol) and 12 (2.34 g, 4.14 mmol)
were added, respectively. The resultant solution was stirred at
room temperature for 1 h. After solvents were evaporated under
vacuum on a rotoevaporator, the crude product was purified by
silica gel column (2.5 cm ×25 cm) eluting with 50% MeOH/
EtOAc to yield 1.84 g (74%) of 13 as a colorless tacky solid.
1H NMR (CD3OD, 500 MHz): δ 1.38-1.43 (m, 4H), 1.44 (s,
18H), 1.51-1.75 (m, 8H), 2.19-2.28 (m, 4H), 2.62-2.72 (m,
6H), 2.89-2.92 (m, 2H), 3.15-3.19 (m, 2H), 3.33-3.35 (m,
4H), 3.53 (t, J ) 5.4 Hz, 4H), 3.56 (t, J ) 5.4 Hz, 4H),
3.61-3.67 (m, 12H), 4.29 (dd, J ) 4.8, 8.0 Hz, 2H), 4.47 (dd,
J ) 4.8, 8.0 Hz, 2H), 4.61 (t, J ) 6.6 Hz, 2H), 7.24 (s, 2H),
7.49 (s, 1H). LRMS (ES+) calcd for C56H90N11O16S2 (M+H)+:
1236.6. Found: 1236.8. HPLC: tR ) 11.0 min.
15-N-(Phthalimido)-(3,6,9,12-tetraoxa)pentadecanoic acid,
19. This compound was obtained from Quanta BioDesign
(Powell, OH) and was used without further purification. H
NMR (CDCl3, 500 MHz): δ 2.62 (t, J ) 6.4 Hz, 2H), 3.57-3.63
(m, 10H), 3.65-3.67 (m, 2H), 3.74-3.77 (m, 4H), 3.91 (t, J )
5.8 Hz, 2H), 7.72 (dd, J ) 3.2, 5.4 Hz, 2H), 7.85 (dd, J ) 3.2,
5.4 Hz, 2H), 9.38 (br s, 1H).
1
1-(15-N-(Phthalimido)-(3,6,9,12-tetraoxa)pentadecanoy-
lamino)methane-tris(N-(4,7,10-trioxa-13-amino)-N-(tert-bu-
tyloxycarbonyl)-3-propionylamine), 20. A 0.30 g (0.77 mmol)
quantity of 19 was dissolved in 50 mL of anhydrous DMF. To
that solution was added 0.89 g (0.77 mmol) of 18, 0.32 g (1.54
mmol) of DCC and catalytic amount of HOBT. This solution
was stirred at room temperature over a weekend. Et3N was added
to neutralize the residual TFA in the acid (from removal of tBu
ester). When the acid and DCC were mixed, a white precipitate
was noted immediately. To purify, the solvent was removed
under vacuum and the residue loaded onto a silica gel column
and eluted with a gradient of 100% EtOAc to 50% MeOH/
Nitromethane-tris(3-propionate-O-2,3,5,6-tetrafluorophe-
nyl ester), 15. A 5.0 g (18.1 mmol) quantity of nitromethane-
trispropionic acid, 14, was dissolved in 10 mL of anhydrous
DMF under argon atmosphere. To that solution was added 10.4
mL (7.29 g, 72.2 mmol) of Et3N, followed by addition of 12.5
mL (18.9 g, 72.2 mmol) of TFP-OTFA at 0 °C. The reaction
was allowed to come to room temperature and was stirred for
an additional 30 min. A portion of the DMF was removed under
vacuum, and the resultant light brown liquid was poured into
200 mL cold water (ice bath) with stirring. The colorless solid
was collected and dried to give 12.28 g (96%) of 15. mp
1
EtOAc to yield 0.85 g (72%) of 20 as a pale yellow oil. H
NMR (CDCl3, 500 MHz): δ 1.43 (s, 27H), 1.73-1.78 (m, 12H),
1.99 (t, J ) 7.8 Hz, 6H), 2.19 (t, J ) 7.8 Hz, 6H), 2.38 (t, J )
5.9 Hz, 2H), 2.43 (br s, 1H), 3.21 (q, J ) 5.9 Hz, 6H), 3.29 (q,
J ) 6.0 Hz, 6H), 3.52-3.55 (m, 12H), 3.57-3.60 (m, 22H),
3.63-3.66 (m, 14H), 3.72 (t, J ) 5.8 Hz, 2H), 3.73 (t, J ) 5.8
Hz, 2H), 3.90 (t, J ) 5.9 Hz, 2H), 5.13 (br s, 3H), 6.53 (br s,
3H), 7.73 (dd, J ) 3.2, 5.4 Hz, 2H), 7.85 (dd, J ) 3.2, 5.4 Hz,
2H). HRMS (ES+) calcd for C74H130N8NaO25 (M+Na)+:
1553.9045. Found: 1553.8990. HPLC: tR ) 14.2 min.
1
133.4-134.4 °C. H NMR (CDCl3, 500 MHz): δ 2.52 (t, J )
7.9 Hz, 6H), 2.82 (t, J ) 7.9 Hz, 6H), 7.00-7.06 (m, 3H).
HRMS (ES+) calcd for C28H15F12NNaO8 (M+Na)+: 744.0504.
Found: 744.0518. HPLC: tR ) 15.2 min.
N-(15-Amino-(3,6,9,12-tetraoxa)pentadecanoylami-
no)methane-tris(N-(4,7,10-trioxa-13-amino)-N-(tert-butyloxy-
carbonyl)-3-propionylamine), 21. Hydrazine (106 mg, 3.3
mmol) was added to a solution of 20 (245 mg, 0.16 mmol) in
absolute EtOH (20 mL) and stirred at room temperature
overnight. The solvent was removed under vacuum, and the
residue was dissolved in MeOH. The solution was filtered to
remove precipitates and then loaded onto a Sephadex LH-20
gel filtration column for purification. Combined fractions yielded
Nitromethane-tris(N-(4,7,10-trioxa-13-tridecaneamino)-N-
(tert-butyloxycarbonyl)-3-propionylamine), 17. A 8.40 g (26.2
mmol) quantity of N-(tert-butyloxycarbonyl)-4,7,10-trioxa-1,13-
tridecanediamine, 16, was dissolved in 70 mL anhydrous THF.
The solution was cooled to -78 °C using a dry ice acetone
bath. Then, 5.64 g (8.0 mmol) of 15 was added. The reaction
mixture changed from yellow to clear while it was allowed to
come to room temperature with stirring overnight. The THF
was removed on a rotary evaporator and the residue was
dissolved in 100 mL EtOAc, washed with 1 N HCl (2 × 50
mL), 1 N Na2CO3 (3 × 50 mL), and brine (50 mL). The EtOAc
was evaporated by rotary evaporator, and the residue was dried
1
0.18 g (78%) of 21 as pale yellow oil. H NMR (CDCl3, 500
MHz): δ 1.43 (s, 27H), 1.73-1.79 (m, 12H), 1.99 (t, J ) 7.6
Hz, 6H), 2.21 (t, J ) 7.6 Hz, 6H), 2.45 (t, J ) 5.9 Hz, 2H),
3.05 (br s, 2H), 3.22 (q, J ) 6.0 Hz, 6H), 3.28 (q, J ) 6.1 Hz,
6H), 3.53 (q, J ) 6.3 Hz, 12H), 3.58-3.61 (m, 18H), 3.62-3.68
(m, 22H), 3.72 (t, J ) 5.8 Hz, 2H), 5.12 (br s, 3H), 7.07 (br s,
3H), 7.51 (s, 1H). HRMS (ES+) calcd for C66H129N8O23
(M+H)+: 1401.9165. Found: 1401.9122. HPLC: tR ) 12.0 min.
2,3,4,6-Tetra-O-acetyl-galactopyranosyl-1-ꢀ-pseudothio-
urea Hydrobromide, 23. This compound was prepared by
reaction of 2,3,4,6-tetra-O-acetyl-R-D-galactopyranosyl bromide,
22, to form thiourea intermediate, 23, in a procedure similar to
that previously described (29). To a solution of 2.02 g (0.026
mol) of thiourea in 24 mL acetone was added 10 g (0.024 mol)
of 2,3,4,6-tetra-O-acetyl-R-D-galactopyranosyl bromide, 22. The
mixture was refluxed for 20 min, then cooled on ice. The solid
was collected and rinsed with 10 mL of ice-cold acetone. The
collected solid was suspended in 20 mL ice-cold acetone, then
refluxed for 15 min, cooled on ice, and filtered. The solid was
rinsed with 10 mL of ice-cold acetone, air-dried, and then dried
under vacuum to yield 4.15 g (69%) of 23 as a colorless solid.
This compound was used in the next step without further
purification.
1
under vacuum to give 8.65 g (92%) of 17 as a yellow oil. H
NMR (CDCl3, 500 MHz): δ 1.43 (s, 27H), 1.73-1.79 (m, 12H),
2.13 (t, J ) 7.9 Hz, 6H), 2.25 (t, J ) 7.9 Hz, 6H), 3.21 (q, J )
5.8 Hz, 6H), 3.33 (q, J ) 5.9 Hz, 6H), 3.52-3.57 (m, 12H),
3.58-3.61 (m, 12H), 3.63-3.66 (m, 12H), 5.08 (br s, 3H), 6.65
(br s, 3H). HRMS (ES+) calcd for C55H105N7NaO20 (M+Na)+:
1206.7312. Found: 1206.7245. HPLC: tR ) 13.2 min.
Aminomethane-tris(N-(4,7,10-trioxa-13-amino)-N-(tert-bu-
tyloxycarbonyl)-3-propionylamine), 18. A 0.95 g (0.83 mmol)
quantity of 17 was dissolved in 30 mL of absolute EtOH. To
that solution was added 3 g of a 50% RaNi/H2O slurry with
stirring. The mixture was put under hydrogen (45 psi) and was
stirred overnight. Hydrogen pressure dropped from 45 to 40
psi. Hydrogen was removed with a H2O aspirator vacuum, and
argon was introduced to the mixture; this procedure was repeated
3×, and then the EtOH solution was filtered through a Celite
column. The solvent was removed from the filtrate to yield
0.95 g (99%) of 18 as colorless oil. 1H NMR (CDCl3, 500 MHz):
δ 1.44 (s, 27H), 1.66 (t, J ) 7.9 Hz, 6H), 1.73-1.79 (m, 12H),
2.05 (br s, 2H), 2.21 (t, J ) 7.8 Hz, 6H), 3.22 (q, J ) 5.8 Hz,
6H), 3.32 (q, J ) 6.0 Hz, 6H), 3.54 (q, J ) 6.0 Hz, 12H),
3.58-3.61 (m, 12H), 3.63-3.66 (m, 12H), 5.14 (br s, 3H), 6.59
(br s, 3H). HRMS (ES+) calcd for C55H108N7O18 (M+H)+:
1154.7751. Found: 1154.7754. HPLC: tR ) 11.6 min.
2,3,4,6-Tetra-O-acetyl-galactopyranosyl-1-ꢀ-(3′-thiopropi-
onic acid), 24. This compound was prepared from the pseudot-
hiourea intermediate, 23, as previously described (30). Briefly,
a mixture of 0.9 g (1.85 mmol) 23, 0.37 g (1.99 mmol)
3-iodopriopionic acid, 0.32 g (167 mmol) Na2S2O5, and 0.30 g
(2.15 mmol) potassium carbonate in 1.5 mL acetone/1.5 mL

Bis-Biotin Reagents for Pretargeting
Bioconjugate Chem., Vol. 21, No. 7, 2010 1229
H2O was stirred at room temperature for 45 min. To the mixture
was added 7 mL of HCl (5%) and 7 mL of CHCl3. The CHCl3
layer was separated, washed with H2O, dried with MgSO4, and
evaporated to give 0.77 g (96%) of 24 as an oil. This compound
was used in the next step without further purification.
anhydrous DMF (2 mL), was added dropwise to a solution of
25 (54.9 mg, 0.094 mmol) in anhydrous DMF (2 mL). The
resultant solution was stirred at room temperature overnight.
Volatile materials were removed by rotary evaporator under
vacuum, and then the crude product was dissolved in 50%
MeOH/water and purified by flash chromatography (Biotage).
After purification, a 60 mg (0.017 mmol; 91%) quantity of 29
was obtained as a colorless oil. 1H NMR (CD3OD, 500 MHz):
δ 1.38-1.46 (m, 4H), 1.52-1.65 (m, 8H), 1.72-1.78 (m, 12H),
1.95 (s, 9H), 1.96-2.00 (m, 6H), 2.03 (s, 9H), 2.04 (s, 9H),
2.15 (s, 9H), 2.16-2.24 (m, 10H), 2.51-2.55 (m, 6H),
2.66-2.71 (m, 4H), 2.86-3.01 (m, 15H), 3.16-3.29 (m, 15H),
3.41 (t, J ) 5.1 Hz, 2H), 3.50-3.54 (m, 12H), 3.57-3.72 (m,
63H), 4.13 (s, 6H), 4.28-4.31 (m, 2H), 4.44-4.50 (m, 4H),
4.76 (d, J ) 9.6 Hz, 3H), 5.10-5.17 (m, 6H), 5.43 (d, J ) 2.9
Hz, 3H), 7.38 (s, 1H), 7.86 (s, 1H), 7.87 (s, 1H). LRMS (ES+)
2,3,4,6-Tetra-O-acetyl-galactopyranosyl-1-r-(3′-thiopropi-
onate-O-2,3,5,6-tetrafluorophenyl ester), 25. To a solution of
24 (772 mg, 1.77 mmol) in anhydrous DMF (15 mL) at 0 °C,
was added dropwise TFP-OTFA (0.448 mL, 2.6 mmol),
followed by addition of Et3N (0.362 mL, 2.6 mmol). The
mixture was stirred in an ice bath for 1 h. Then, half of the
solvent was removed under vacuum, and the remaining solution
was extracted with EtOAc (3 × 100 mL). The EtOAc solution
was dried with MgSO4 and evaporated by rotary evaporator
under vacuum to afford 1.038 g (100%) of 25 as a pale yellow
1
oil. H NMR (CDCl3, 500 MHz): δ 2.00 (s, 3H), 2.05 (s, 3H),
C
151H242KN19NaO64S5 (M+H+Na+K)+ calc 3567.4. Found
2.08 (s, 3H), 2.17 (s, 3H), 2.98-3.04 (m, 1H), 3.09-3.12 (m,
2H), 3.13-3.19 (m, 1H), 3.99 (t, J ) 6.5 Hz, 1H), 4.14-4.19
(m, 2H), 4.60 (d, J ) 10.0 Hz, 1H), 5.08 (dd, J ) 3.4, 10.0 Hz,
1H), 5.28 (t, J ) 9.9 Hz, 1H), 5.46 (d, J ) 3.5 Hz, 1H),
7.01-7.08 (m, 1H). HRMS (ES+) C23H24F4NaO11S (M+Na)+
calcd: 607.0873. Found: 607.0865. HPLC: tR ) 16.3 min.
3567.5. HPLC: tR ) 11.2 min.
N-(15-(N,N′-Bis(N-(biotinyl)-ꢀ-N-(tert-butyloxycarbonyl)-
N-(8-amino-3,6-dioxaoctanyl)aspartamidyl)-5-aminoisoph-
thalamido)-(3,6,9,12-tetraoxa)pentadecanoylamino)methane-
tris(N-(4,7,10-trioxa-13-amino)-N-(galactose-1-r-(3′-
thiopropionylamine), 2. A 83 mg (0.024 mmol) quantity of
29 was added to a clear solution of Na metal (0.55 mg; 0.024
mmol) in anhydrous MeOH (1 mL). The resultant solution was
stirred for 2 h at room temperature. The mixture was loaded
onto an Amberlite IR120 ion-exchange resin (H+ form) and
eluted slowly. The fractions containing 2 were combined, and
the solvent was removed under vacuum to yield 67.5 mg (0.022
mmol; 94%) as a pale yellow oil. 1H NMR (CD3OD, 500 MHz):
δ 1.38-1.46 (m, 4H), 1.54-1.66 (m, 8H), 1.73-1.79 (m, 12H),
1.96-2.00 (m, 6H), 2.17-2.30 (m, 10H), 2.42 (t, J ) 5.7 Hz,
2H), 2.55 (t, J ) 6.9 Hz, 6H), 2.68-2.79 (m, 4H), 2.88-2.93
(m, 4H), 2.97-3.03 (m, 4H), 3.16-3.29 (m, 15H), 3.42 (t, J )
5.1 Hz, 2H), 3.46-3.56 (m, 25H), 3.57-3.72 (m, 63H), 3.77
(dd, J ) 7.4 Hz, 11.4 Hz, 3H), 3.87 (d, J ) 2.0 Hz, 3H),
4.28-4.31 (m, 2H), 4.34 (d, J ) 7.3 Hz, 2H), 4.47-4.50 (m,
2H), 4.72-4.74 (m, 2H), 7.37 (s, 1H), 7.88 (s, 1H), 7.93 (s,
1H). HRMS (ES+) C127H218N19O52S5 (M+H)+ calc: 3001.3602.
Found: 3001.3223. HPLC: tR ) 8.4 min.
N,N′-Bis(N-(biotinyl)-ꢀ-N-(butyloxycarbonyl)-N-(8-amino-
3,6-dioxaoctanyl)aspartamidyl-5-N-(DOTA-benzylamido)am-
inoisophthalamide, 32. A solution of 13 (170 mg, 0.137 mmol),
thiocarbonyldiimidazole (40.8 mg, 0.206 mmol), and anhydrous
DMF (4 mL) was stirred at room temperature for 1 h. After the
solution was washed with 10% EtOAc/hexanes (5 × 15 mL),
the volatile materials were removed under vacuum on a
rotoevaporator, and the isothiocyanate 30 was dried under
vacuum for 2 h. The crude 30 was dissolved in anhydrous DMF
(4 mL), and then aminobenzylDOTA, HCl salt, 31 (100 mg,
0.137 mmol), and pyridine (111 µL, 1.374 mmol) were added,
respectively. The resultant solution was stirred at room tem-
perature for 16 h. Volatile materials were evaporated under
vacuum on a rotoevaporator. The residue was dissolved in DMF/
MeOH/water (5 mL, 1/2/2) and purified by flash chromatogra-
phy (Biotage). Fractions containing product were combined and
solvent removed to yield 185 mg (75%) of 32 as a colorless
solid. 1H NMR (CD3OD, 500 MHz): δ 1.37-1.42 (m, 4H), 1.44
(s, 18H), 1.51-1.75 (m, 8H), 2.26 (t, J ) 7.2 Hz, 4H), 2.69 (t,
J ) 7.1 Hz, 4H), 2.74 (d, J ) 12.9 Hz, 2H), 2.95 (dd, J ) 4.4,
12.9 Hz, 2H), 3.00-3.19 (m, 6H), 3.22-3.26 (m, 3H),
3.31-3.42 (m, 11H), 3.58-3.76 (m, 26H), 3.85-3.98 (m, 5H),
4.36 (dd, J ) 4.5, 7.5 Hz, 2H), 4.56 (dd, J ) 4.5, 7.5 Hz, 2H),
4.59 (t, J ) 6.5 Hz, 2H), 7.36 (d, J ) 8.3 Hz, 2H), 7.42 (d, J
) 8.3 Hz, 2H), 8.04 (s, 2H), 8.10 (s, 1H). HRMS (ES+) calcd
for C80H123N16O24S3 (M+H)+: 1787.8053. Found: 1787.8137.
HPLC: tR ) 10.4 min.
N-(15-(N,N′-Bis(N-(biotinyl)-ꢀ-N-(tert-butyloxycarbonyl)-
N-(8-amino-3,6-dioxaoctanyl)aspartamidyl)-5-aminoisoph-
thalamido)-(3,6,9,12-tetraoxa)pentadecanoylamino)methane-
tris(N-(4,7,10-trioxa-13-amino)-N-(tert-butyloxycarbonyl)-3-
propionylamine), 27. A solution of 13 (135 mg, 0.109 mmol),
1,1′-carbonyldiimidazole (35.4 mg, 0.218 mmol) and anhydrous
DMF (8 mL) was stirred at room temperature for 1 h. The
solution was triturated with 10% EtOAc/hexanes (50 mL). The
solution was decanted, and the residue, containing 26, was
washed with EtOAc (2 × 10 mL), then dried under vacuum
for 2 h. A solution containing 21 (104 mg, 0.74 mmol), Et3N
(20.6 µL, 0.148 mmol), and anhydrous DMF (8 mL) was added
to the flask containing 26, and the resultant solution was stirred
at room temperature overnight. The volatile materials were
removed by rotary evaporation under vacuum, and the crude
residue was purified by Sephadex LH-20 gel filtration column
to yield 173 mg (88%) of 27. 1H NMR (CD3OD, 500 MHz): δ
1.42 (s, 27H), 1.43 (s, 9H), 1.44 (s, 9H), 1.52-1.77 (m, 24H),
1.97 (t, J ) 7.6 Hz, 6H), 2.15-2.27 (m, 10H), 2.41 (t, J ) 6.0
Hz, 2H), 2.62-2.72 (m, 6H), 2.88-2.93 (m, 2H), 3.11 (t, J )
6.8 Hz, 6H), 3.14-3.19 (m, 2H), 3.24 (t, J ) 6.9 Hz, 6H),
3.33-3.35 (m, 4H), 3.42 (t, J ) 5.2 Hz, 2H), 3.48-3.52 (m,
12H), 3.53-3.72 (m, 60H), 4.30 (dd, J ) 4.9, 7.9 Hz, 2H),
4.48 (dd, J ) 4.9, 7.9 Hz, 2H), 4.61 (t, J ) 6.3 Hz, 2H), 7.86
(s, 1H), 7.99 (s, 1H), 8.12 (s, 1H). HRMS (ES+)
C123H215N19Na2O40S2 (M+2Na)2+ calcd: 1354.2300. Found:
1354.2239. HPLC: tR ) 13.7 min.
N-(15-(N,N′-Bis(N-(biotinyl)-ꢀ-N-(tert-butyloxycarbonyl)-
N-(8-amino-3,6-dioxaoctanyl)aspartamidyl)-5-aminoisoph-
thalamido)-(3,6,9,12-tetraoxa)pentadecanoylamino)methane-
tris(N-(4,7,10-trioxa-13-amino)-3-propionylamine), 28. A 139
mg (0.052 mmol) quantity of 27 was stirred in 2 mL of TFA
for 45 min. The excess TFA was removed under a stream of
air. The residue was washed with diethyl ether and was placed
under high vacuum overnight. The crude residue was dissolved
in H2O and was purified over a Sephadex LH-20 gel filtration
column to yield 92.8 mg (0.036 mmol; 69%) of 28. This
compound was used in the next step without further character-
ization.
N-(15-(N,N′-Bis(N-(biotinyl)-ꢀ-N-(tert-butyloxycarbonyl)-
N-(8-amino-3,6-dioxaoctanyl)aspartamidyl)-5-aminoisoph-
thalamido)-(3,6,9,12-tetraoxa)pentadecanoylamino)methane-
tris(N-(4,7,10-trioxa-13-amino)-N-(2,3,4,6-tetra-O-acetyl-
galactose-1-r-(3′-thiopropionylamine), 29. A solution, containing
28 (42.3 mg, 0.019 mmol), Et3N (21 mL, 0.15 mmol), and

1230 Bioconjugate Chem., Vol. 21, No. 7, 2010
Wilbur et al.
N,N′-Bis(N-biotinyl-N-(8-amino-3,6-dioxaoctanyl)asparta-
midyl-5-N-(DOTA-benzylamido)aminoisophthalamide, 4a. A
solution containing 32 (185 mg, 0.103 mmol) and trifluoroacetic
acid (4 mL) was stirred at room temperature for 20 min. Excess
trifluoroacetic acid was evaporated under a stream of argon,
and then the residue was dissolved in 50% MeOH/water and
purified by flash chromatography (Biotage). Fractions containing
product were combined and solvent removed to yield 165 mg
to the avidin-agarose beads as a wash. The filter was vortexed
to mix, then was placed in a microcentrifuge and spun at 1000
g for 1 min. The filtrate was combined with the previous filtrate.
The microcentrifuge filter containing avidin-agarose beads was
placed in a tube, and then that tube and the tube containing the
combined filtrates were counted in a gamma counter. The
quantity of 111In-labeled biotin derivative remaining in the filter
(% bound) was calculated as [radioactivity in filter]/[radioactivity
in filter] + [radioactivity in filtrates] × 100.
1
(95%) of 4a as a colorless solid. mp 211-213 °C. H NMR
(CD3OD, 500 MHz): δ 1.33-1.41 (m, 4H), 1.47-1.71 (m, 8H),
2.23 (t, J ) 7.2 Hz, 4H), 2.66-2.78 (m, 6 H), 2.92 (dd, J )
4.9, 13.0 Hz, 2H), 2.97-3.17 (m, 6H), 3.19-3.23 (m, 3H), 3.33
(t, J ) 5.3 Hz, 4H), 3.56 (t, J ) 5.1 Hz, 4H), 3.62-3.74 (m,
28H), 3.90-4.21 (m, 6H), 4.33 (dd, J ) 4.8, 7.5 Hz, 2H), 4.53
(dd, J ) 4.9, 7.4 Hz, 2H), 4.69 (t, J ) 6.4 Hz, 2H), 7.32-7.43
(m, 4H), 8.00 (s, 2H), 8.07 (s, 1H). LRMS (ES+) calcd for
C72H107N16O24S3 (M+H)+: 1675.6. Found: 1675.5. HPLC: tR
) 6.3 min.
Lymphoma Cell Line. The human Ramos B lymphoma cell
line (American Type Culture Collection, Bethesda, MD) was
maintained in log-phase growth in RPMI-1640 medium supple-
mented with 10% heat-inactivated fetal calf serum in a 5% CO2
incubator.
Synthesis and Purification of anti-CD20 1F5 scFv4-SAv
Fusion Proteins. The production and purification of 1F5 scFv4-
SAv-WT fusion protein and the related site-directed SAv mutant
fusion proteins, 1F5 scFv4-SAv-S45A and 1F5 scFv4-SAv-
Y43A, have been previously described (17, 18). Briefly, scFv-
SAv fusion genes were produced by molecularly fusing the
single-chain variable regions (scFv) of the murine anti-CD20
1F5 antibody to either the full-length SAv wild-type gene or
SAv mutant genes. The resultant fusion proteins were expressed
in the periplasmic space of E. coli and spontaneously formed
stable soluble tetramers of 1F5 scFv4-SAv with a molecular
weight of 174 kDa. The 1F5-SAv fusion proteins were
formulated at a concentration of 3 mg/mL in phosphate buffered
saline (PBS) containing 5% sorbitol.
Preparation of 1F5 [125I]scFv4-SAv-S45A Fusion Protein.
Fusion proteins were radioiodinated with Na[125I]I (PerkinElmer,
Waltham, MA) by the chloramine-T method in a manner similar
to that previously published (32). Briefly, 50 µL of a 1 mg/mL
solution of chloramine-T (Sigma-Aldrich, St. Louis, MO) was
added to 500 µL fusion protein. Next, the Na[125I]I was added
and the solution vortexed. Following a 1 min incubation at room
temperature, 5 µL of a 1 mg/mL solution of sodium metabisulfite
was added, and the solution was vortexed. The contents were
pipetted onto a prewashed PD10 column and eluted with PBS,
collecting 600 µL fractions. The amount of radioactivity in each
fraction was measured in a Capintec CRC-7 Radioisotope
Calibrator, and early fractions with the highest activity were
pooled to obtain the labeled fusion protein.
Blood Clearance of 1F5 [125I]scFv4-SAv Fusion Protein.
Female FoxN1NU athymic nude mice, aged 6-8 weeks, were
obtained from Harlan Sprague-Dawley (Indianapolis, IN) and
housed under protocols approved by the Fred Hutchinson Cancer
Research Institute Institutional Animal Care and Use Committee.
Mice were fed a standard diet (containing biotin) and separated
into three groups of 4 mice each. Mice were injected intrave-
nously with 500 µg (2.8 nmol) of 1F5 [125I]scFv4-SAv-S45A
fusion protein. After 24 h, mice were injected intravenously
with equimolar doses of 1 (5.8 nmol; 50 µg) or 2 (5.8 nmol;
17.4 µg). Blood samples were drawn at serial time points before
and after injection of 1 or 2, and the blood samples were counted
in a gamma counter. The percent of injected dose per gram of
blood (%ID/g) was calculated, and the mean values were plotted
using KaleidaGraph software (Synergy Software, Reading, PA).
A tabulation of the blood clearance data is provided as Table
S1 in the Supporting Information.
Biodistributions of [111In]3b and [111In]4b in Tumor-
Bearing Mice Pretargeted with 1F5 scFv4-SAv Fusion
Proteins. Female FoxN1NU athymic nude mice were separated
into four groups and were fed either a standard diet (group 1)
or biotin-free diet (Harlan Teklad, Madison, WI) (groups 2-4).
Ramos cells (10 × 106) were injected subcutaneously in the
right flank of each mouse. Mice were monitored until palpable
tumor nodules appeared (7-10 days) and mice with tumors of
Preparation of [111In]indium-labeled 3a, [111In]3b. The
radiolabeling reactions were conducted with metal free reagents
and labware. Methods for demetalation have been previously
described (31). To a 7-17 µL aliquot of a 3 mg/mL (20-50
µg) solution of 3a in H2O was added 80-100 µL 500 mM
NH4OAc solution, pH 5.3 (demetalated), followed by 2-4 µL
(0.63-1.22 mCi) of [111In]indium chloride. This mixture was
placed in a heat block at 85 °C for 30-60 min, then purified
by radio-HPLC to give 72-95% isolated yield from a peak at
4.5 min. Following isolation from the HPLC, the eluant solvents
were removed by evaporation under vacuum (90-100%), and
the predetermined amount of PBS was added to make the animal
injectate. Radiochemical purity was determined by an avidin
bead assay. In that assay, the preparations had 95-99% of the
[
111In]3b bound with avidin.
Preparation of [111In]Indium-labeled 4a, [111In]4b. This
reaction was conducted in a manner similar to preparation of
[
111In]3b. To a 27-33 µL aliquot of 3 mg/mL (80-100 µg) 4a
in H2O was added 80-100 µL 500 mM NH4OAc solution, pH
5.3 (demetalated) followed by 6-10 µL (1.42-2.81 mCi) of
[
111In]indium chloride. This mixture was placed in a heat block
at 85 °C for 30 min, then purified by HPLC to give 45-94%
isolated yields from a peak at 7.5 min. Following isolation from
the HPLC, the eluant solvents were removed by evaporation
under vacuum (90-100%) and the predetermined amount of
PBS was added to make the animal injectate. Analyses by avidin
bead assay showed that 85-95% of the [111In]4b bound with
avidin.
Avidin Bead Assay. The avidin beads were prepared for use
by the following procedure. A solution of avidin-agarose from
the manufacturer (Sigma-Aldrich #A9207-5 mL or Immu-
noPure Immobilized Avidin Gel, Pierce Biotechnology, Inc.
Rockford, IL), was vortexed, and then a 300 µL aliquot of the
suspended avidin-agarose beads was added to a microcentrifuge
filter (Fisher #07-200-387; 0.45 mm cellulose acetate spin
column with centrifuge tube filter). The filter was placed in a
microcentrifuge and spun at 1000 g for 1 min. The filtrate was
discarded.
The binding assay was conducted as follows. A 300 µL
aliquot of PBS was added to filter (with avidin-agarose) and
vortexed to thoroughly mix. The beads were spun at 1000 g
for 1 min, and then the PBS filtrate was discarded. The PBS
washing step was repeated, and then an aliquot (up to 2 µL) of
111In-labeled biotin ([111In]3b or [111In]4b) was added to the
suspended beads, and this suspension was vortexed to mix
thoroughly. The 111In-labeled biotin was incubated with the
beads for 10 min at room temperature. The mixture was vortexed
to mix, then placed in microcentrifuge and spun at 1000 g for
1 min. The filtrate was set aside, and 300 µL of PBS was added

Bis-Biotin Reagents for Pretargeting
Bioconjugate Chem., Vol. 21, No. 7, 2010 1231
Scheme 1. Synthetic Steps to Prepare the Bis-Biotin Derivative 13, Used for Syntheses of Bis-Biotin Clearing Agent, 2, and
Bis-Biotin-DOTA Derivative, 4a
a (a) DMF, H2O, NaOH, di-tBu-dicarbonate, 0 °C to rt, 30 h; (b) DMF, tetrafluorophenol, EDC, rt, 16 h; (c) Et3N, CH3CN, 7, rt, 30 min, 81%;
(d) TFP-OTFA, Et3N, rt, 10 min; (e) H2O, acetone, NaHCO3, 10, rt, 6 h; (f) TFP-OTFA, DMF, Et3N, 0 °C, 30 min; (g) TFA, rt, 10 min, evap,
DMF, Et3N, rt, 1 h, 74%.
similar sizes (5-10 mm3) were selected for the experiment.
Groups of 5 mice were injected intravenously with 500 µg (2.9
nmol) of 1F5 scFv4-Sav-WT (groups 1 and 2) or one of the
SAv mutant fusion proteins, 1F5 scFv4-SAv-S45A (group 3)
or 1F5 scFv4-SAv-Y43A (group 4). After 20 h, the mice were
injected with either 50 µg (5.8 nmol) of 1 (groups 1 and 2) or
17.5 µg (5.8 nmol) of 2 (groups 3 and 4). Four hours after
administration of clearing agent, 20 µCi on 1.2 nmol (∼1 µg
3a or ∼2 µg of 4a) of the 111In-labeled biotin derivatives,
yield. It should be noted that the synthesis was conducted in
this manner so that the aspartate carboxylates remained as tert-
butyl esters in bis-biotin derivative 13, since that is required to
prepare the isocyanate functionality in a later step of the
synthesis.
A key part of the design of the blood-clearing agent 2 was
the inclusion of three galactosyl groups at the termini of
poly(ethylene glycol) (PEG) spacer arms. A PEG-containing
tetrafunctional molecule, 21, was utilized as a core for the
branched galactosyl portion of 2. The synthesis to prepare 21
is shown in Scheme 2. In the synthesis, nitromethane-tris-
propionic acid, 14, was converted to the tris-tetrafluorophenyl
ester 15 in 96% yield by reaction with tetrafluorophenyl-O-
trifluoroacetate (TFP-OTFA) and triethylamine in DMF. Reac-
tion of 15 with an excess of N-Boc-trioxatridecanediamine, 16,
provided the adduct 17 in 92% yield. Reduction of the nitro
group in 17 by hydrogenation over Raney-Nickel gave nearly
a quantitative yield of the amino derivative 18. Reaction of the
hindered amine in 18 with the phthalimido-protected amino-
dPEG4-acid, 19, and DCC/HOBT in DMF at room temperature
for 72 h provided a 72% yield of adduct 20. Removal of the
phthalimide protecting group using hydrazine in ethanol for 16 h
provided a 78% yield of the tetra-functional intermediate, 21.
The tetra-O-acetyl-galactose derivative 25 used in the syn-
thesis of 2 was prepared as shown in Scheme 3. Nucleophilic
displacement of the bromine in acetylbromo-R-D-galactose, 22,
by reaction of thiourea in refluxing acetone produced the
ꢀ-pseudothiourea derivative 23 in 69% yield. Reaction of 23
with 3-iodopropionic acid in a basic acetone/water mixture gave
a 96% yield of the galactose-thiopropionic acid derivative, 24.
Preparation of the tetrafluorophenyl ester derivative, 25, was
achieved in nearly quantitative yield by reaction of 24 with TFP-
OTFA.
The final sequence of reactions, which combined the three
components to produce 2, is shown in Scheme 4. The reaction
sequence began with conversion of the bis-biotin derivative 13
to the phenylisocyanate, 26, by reaction with carbonyldiimida-
zole (CDI). Reaction of 26 with 21 provided the adduct 27 in
88% yield. Deprotection of the N-Boc and tert-butyl esters in
27 was accomplished in one step by reaction in neat trifluoro-
acetic acid to provide a 69% yield of intermediate 28. Reaction
of 28 with the tetra-O-acetyl-galactose derivative 25 provided
29 in 91% yield. In the final step, the acetyl protecting groups
were removed from 29 using an acid form of an ion exchange
column to provide a 94% yield of the targeted blood clearance
agent 2.
[
111In]3b (groups 1 and 2) or [111In]4b (groups 3 and 4), were
administered. Mice were euthanized, and tumors and normal
organs were harvested 24 h after injection of radioactivity.
Selected tissues and tumors were weighed and counted in a
gamma counter. Counts were corrected for decay using standard
aliquots of the injectate. The percent injected dose per gram
(%ID/g) values for tissues and tumor xenografts were calculated.
The mean values and standard deviations were plotted for each
tissue. A tabulation of the pretargeting data is provided as Table
S2 in the Supporting Information.
Biodistributions of [111In]4b. Female FoxN1NU athymic nude
mice,aged6-8weeks,wereobtainedfromHarlanSprague-Dawley
(Indianapolis, IN) and housed under protocols approved by the
Fred Hutchinson Cancer Research Institute Institutional Animal
Care and Use Committee. Mice were fed a standard diet
(containing biotin) and separated into three groups of 5 mice
each. Mice were injected intravenously with ∼30 µCi on 2 µg
of [111In]4b. Groups of mice were euthanized at 1, 4, and 24 h
post injection of [111In]4b, and preselected tissues were harvested
and counted in a gamma counter. The percent of injected dose
per gram (%ID/g) of blood or tissue was calculated. A tabulation
of the biodistribution data is provided as Table S3 in the
Supporting Information.
RESULTS
Synthesis of the Blood Clearance Agent, 2. The synthesis
of 2 utilized a convergent approach, combining a bis-biotin
component, a tetrafunctional poly(ethylene glycol) (PEG)
component, and a galactosyl component. The syntheses of the
individual components are shown in Schemes 1-3. The
synthetic steps to prepare the bis-biotin derivative 13 are shown
in Scheme 1. In the synthesis, aminoisophthalic acid, 5, was
modified with di-tert-butyl-dicarbonate, and then tetrafluorophe-
nol and EDC were used to prepare N-Boc-aminoisophthalate
ditetrafluorophenyl ester, 6. Reaction of 6 with mono-Boc
protected diamino-dioxaoctane, 7, and triethylamine in anhy-
drous DMF provided the tri-N-Boc protected aminoisophthalic
acid derivative, 8. After removal of the N-Boc groups on 8 with
neat TFA, the resulting alkyl amines were reacted with the
tetrafluorophenyl ester of the biotin-aspartate adduct 12, which
was prepared from biotin, 9, as previously described (28). That
reaction provided the anilino-bis-biotin derivative 13 in 74%
Syntheses of the Bis-Biotin-DOTA Derivative, 4a. Partial
synthesis of the radiometal-binding bis-biotin derivative 4a is
shown in Scheme 5. The initial steps of the synthesis involved
preparation of bis-biotin derivative 13, as described previously
(and shown in Scheme 1). Conversion of the aniline in 13 to

1232 Bioconjugate Chem., Vol. 21, No. 7, 2010
Wilbur et al.
Scheme 2. Synthetic Steps to Prepare the Branched Polyethylene Glycol (PEG) Core, 21, Used in Synthesis of Bis-Biotin Clearing
Agent, 2a
a (a) TFP-OTFA, DMF, Et3N, 0 °C, 30 min, 96%; (b) 16, THF, -78 °C - rt, 92%; (c) RaNi, H2, EtOH, H2O, 16 h, 99%; (d) 19, DMF, DCC,
HOBT, rt, 72 h, 72%; (e) NH2NH2, EtOH, rt, 16 h, 78%.
Scheme 3. Synthetic Steps to Prepare the Galactosyl Derivative, 25, Used in Synthesis of Bis-Biotin Clearing Agent, 2a
a (a) Thiourea, acetone, ∆, 20 min, 69%; (b) iodopropionic acid, K2CO3, acetone, H2O, rt, 45 min, HCl, 96%; (c) TFP-OTFA, Et3N, DMF, 0 °C,
1 h, 100%.
the phenylisothiocyanate functionality in 30 was accomplished
by reaction with thiocarbonyldiimidazole (TCDI). The isothio-
cyanate derivative, 30, was prepared and used in this coupling
reaction, rather than the phenylisocyanate, 26, due to the low
reactivity of the aniline. Reaction of 30 with aminobenzylDOTA,
31, using pyridine as the base, provided the adduct 32 in 75%
yield. Removal of the tert-butyl esters in 32 was accomplished
in neat TFA to provide the bis-DOTA-biotin compound 4a in
95% yield.
purified). In subsequent labeling optimization studies, it was
found that consistently high 111In-labeling yields (90+%) could
be obtained for the reaction of 4a by conducting the reactions
in 0.5 M ammonium acetate buffer at pH 5.3 and lowering the
reaction temperature to 45 °C (from 85 °C).
Comparison of Blood Clearance with 1 and 2. An experi-
ment was conducted to determine the effectiveness of 1 and 2
for decreasing 1F5 scFv4-SAv-S45A mutant fusion protein
concentration in blood. In the experiment, the blood concentra-
tions were assessed in three groups of four athymic FoxN1Nu
mice after administration of 1F5 [125I]scFv4-SAv-S45A. Serial
blood samples were obtained and counted to assess the blood
concentrations. A logarithmic plot of the blood concentrations
of 1F5 [125I]scFv4SA-S45A in the three groups of mice, as a
function of time, is provided in Figure 2 (data are provided as
Table S1 in Supporting Information). One group of mice did
not receive a blood clearance agent (group 1, red line). Mice in
the second and third groups received a single dose of either 1
or 2 given intravenously 24 h after 1F5 [125I]scFv4-SAv-S45A
administration. Administration of the bis-biotin-trigalactose
reagent, 2, removed 70% of 1F5 [125I]scFv4-SAv-S45A from
the bloodstream within 2 h of injection (group 3, green line).
However, only a 29% reduction in blood concentrations of 1F5
Radiolabeling. The 1F5 scFv4-SAv-S45A mutant fusion
protein was radioiodinated in a standard protocol using chloram-
ine-T as the oxidant. Several 111In-labeling reactions of the biotin
derivative 3a conducted for these studies provided isolated
radiochemical yields (72-95%) of [111In]3b, with radiochemical
purities of 95-99% being obtained as measured in an avidin
binding assay. It should be noted that the labeling yields reported
are for the isolated product, and in some examples, not all of
the [111In]3b was isolated from the HPLC eluant making the
radiolabeling yields appear lower. In labeling optimization
studies, radiolabeling yields of 97+% were obtained for
[
111In]3b under a variety of reaction conditions studied.
Radiolabeling reactions of 4a with 111In were more variable,
providing ranges of isolated radiochemical yields from 45% to
94% under the reaction conditions used. The radiolabeled
product was purified by radio-HPLC. Analysis of the radio-
chemical purity of [111In]4b, as determined by the avidin bead
assay, was 85-95% for the preparations (some were HPLC
[
125I]scFv4-SAv-S45A was observed at the same time point when
1 was used (group 2, blue line). Similar blood clearance data
were obtained in mice administered the fusion protein, 1F5
[
125I]scFv4-SAv-Y43A (data not shown).

Bis-Biotin Reagents for Pretargeting
Bioconjugate Chem., Vol. 21, No. 7, 2010 1233
Scheme 4. Synthetic Steps Combining Bis-Biotin Derivative, 13, with PEG Core, 21, and Galactosyl Derivative, 25, in the Preparation
of Bis-Biotin Clearing Agent, 2a
a (a) CDI, DMF, rt, 1 h; (b) Et3N, DMF, rt, 16-20 h, 88% [from 13]; (c) TFA, 45 min, 69%; (d) Et3N, DMF, rt, 16-20 h, 91%; (e) MeOH,
Amberlite IR-120 [H+ form], 94%.
Biodistributions of [111In]3b or [111In]4b in Tumor-Bear-
ing Mice Pretargeted with 1F5 scFv4-SAv Fusion Proteins.
A pretargeting experiment was conducted to determine if the
new clearing agent 2 was effective in decreasing normal tissue
concentrations of [111In]4b and to determine if [111In]4b could
be used in a pretargeting protocol to effectively target tumor
xenografts in the presence of endogenous biotin. The biodis-
tribution study did not include groups of mice on biotin-deficient
diets when scFv4-SAv mutant fusion proteins were administered,
as no difference in tumor concentration was seen previously in
pretargeting experiments using the same fusion proteins in mice
placed on biotin-deficient diets or standard diets (22).
In the pretargeting experiment, four groups of female
FoxN1NU athymic nude mice (5 mice/group) bearing Ramos
lymphoma xenografts were administered the fusion proteins
(FP), blood clearance agents (CA), and radiolabeled biotin
derivatives (RB) as depicted in Scheme 6. Group 1 was fed a
biotin-deficient diet and groups 2 through 4 were fed standard
diets. Twenty hours after anti-CD20 fusion protein administra-
tion, the mice received a blood clearing agent, either 1 or 2,
followed 4 h later by either [111In]3b or [111In]4b. Mice were
sacrificed 24 h after administration of [111In]3b or [111In]4b, and
selected tissues were harvested. The tissues were counted in a
gamma counter, and the concentration of 111In in each tissue
was calculated as %ID/g. A bar graph showing the concentra-
tions in tumor xenografts and selected tissues is provided as
Figure 3 (data is provided as Table S2 in Supporting Informa-
tion). An additional biodistribution study was conducted in
athymic mice injecting [111In]4b alone. In that biodistribution,
the only tissue that had more than 1% ID/g was kidney (1.57
( 0.10, 1.54 ( 0.11, and 1.03 ( 0.20 at 1, 4, and 24 h pi,
respectively), and most tissue concentrations were less than
0.2%ID/g (Table S3 in Supporting Information).
Figure 3 shows the dramatic effect that endogenous biotin
can have on tumor concentrations of radiolabeled biotin in
pretargeting protocols. The tumor and tissue concentrations after
administration of wild-type SAv fusion protein (1F5 scFv4-SAv-
WT), the blood clearance agent 1, and 111In-labeled DOTA-
biotin, [111In]3b, are represented as groups 1 (red bars) and 2
(blue bars). Significant tumor concentrations (8.6 ( 1.2% ID/
g) of [111In]3b were obtained when the mice were fed biotin-
free diets (group 1), whereas when mice were on normal diets
(group 2), the concentrations of [111In]3b in tumor (0.25 (
0.1%ID/g) were very low, indicating that little or no tumor
targeting was attained. Furthermore, the substitution of 2 for 1,
or [111In]4b for [111In]3b, did not significantly change the
concentration of radioactivity in tumor xenografts when pre-
targeted with 1F5 scFv4SA-WT (data not shown). In contrast,
pretargeting using the SAv mutant fusion proteins, scFv4-SAv-
S45A or scFv4-Y43A, in mice fed a standard diet exhibited high
concentrations of radioactivity (8.52 ( 4.4% ID/g and 6.71 (
1.3% ID/g, respectively) in tumor xenografts. These results
demonstrate that the blocking of radiolabeled biotin by endog-
enous biotin observed with the 1F5-SAv-WT can be overcome
by using the lower-affinity SAv mutants. The tissue concentra-
tions of [111In]3b and [111In]4b were of a similar magnitude for
the three fusion proteins, except in the blood and liver. The
blood concentrations in the groups administered SAv mutant
fusion proteins, scFv4-SAv-S45A and scFv4-SAv-Y43A, were
much higher (0.18 ( 0.06%ID/g and 0.58 ( 0.13%ID/g) than
in the groups administered scFv4-SAv-WT (0.008 ( 0.002%ID/g
and 0.004 ( 0.001%ID/g), indicating that CA 2 was less
efficient at clearing the SAv mutant fusion proteins than CA 1

1234 Bioconjugate Chem., Vol. 21, No. 7, 2010
Wilbur et al.
Scheme 5. Synthetic Steps to Prepare Bis-Biotin-DOTA Derivative, 4a, and Radiolabeling to Obtain [111In]4ba
a (a) TCDI, DMF, rt, 1 h; (b) 31, pyridine, rt, 16 h, 75%; (c) TFA, rt, 20 min, 95%; (d) [111In]InCl3, NH4OAc, H2O, pH 5.3, 85 °C, 30 min,
45-94%.
Scheme 6. Schematic Depicting the Time Course of Reagent
Injections (Panel A) and the Reagents Used in Each Study
Group (Panel B) in the Pretargeting Experimenta
Figure 2. Blood clearance of 1F5 [ 125I]scFv4-SAv-S45A fusion protein
after administration of 1 or 2. Graph (log scale) of 1F5 [125I]scFv4-
SAv-S45A fusion protein concentration (%ID/g) when no blood
clearance agent was used (group 1; red line), or after administration of
one of the biotinyl-galactosyl-blood clearance agents, 1 (group 2, blue
line) or 2 (group 3, green line). In the experiment, three groups of
athymic FoxN1Nu mice (n ) 4 per group) were injected intravenously
at t ) 0 with 2.8 nmol of 1F5 [ 125I]scFv4-SAv-S45A fusion protein,
followed 24 h later with an intravenous injection of 5.8 nmol of either
a In the experiment, the reagents outlined in the table (panel B) were
administered to 4 groups of mice (n ) 5/group) following the timeline
shown (panel A). The fusion protein (FP) was initially injected (t )
0 h), followed 20 h later by a clearing agent (CA, 1 or 2). Four hours
later, the radiolabeled biotin (RB) derivative, [111In]3b or [111In]4b,
was administered. Mice were sacrificed at 48 h post the fusion protein.
Tumor xenografts and selected tissues were harvested and counted to
determine concentrations (%ID/g) of 111In-labeled biotin derivative in
the tissues.
1 or 2. Serial blood samples were obtained from the tail vein, and 125
I
was counted in a gamma counter. The %ID/g was calculated based on
separately counted injection standards. Average %ID/g values are
plotted with error bars representing ( standard deviation. The plotted
data are provided as Table S1 in the Supporting Information.
(1.77 ( 0.4% ID/g and 3.99 ( 0.8% ID/g, respectively) than
were obtained in groups administered the scFv4-SAv-WT fusion
protein (0.43 ( 0.1% ID/g and 0.36 ( 0.18%ID/g).
DISCUSSION
was with the SAv WT fusion protein. Similarly, liver concentra-
tions in the groups administered SAv mutant fusion proteins,
scFv4-SAv-S45A and scFv4-SAv-Y43A, were somewhat higher
The ultimate goal of our studies is to develop a pretargeting
protocol based on the streptavidin/biotin binding pair that

Bis-Biotin Reagents for Pretargeting
Bioconjugate Chem., Vol. 21, No. 7, 2010 1235
lactosamine-biotin reagent, 1, was designed and synthesized at
NeoRx Corporation, Seattle, WA, and was provided as a gift),
to affect clearance of the wild-type SAv conjugates of, or fusion
proteins with, antibodies. The asialoglycoprotein receptors, also
referred to as Ashwell receptors, are hepatic plasma membrane
receptors responsible for regulating the concentrations of
circulating glycoproteins (34, 35). This regulation is controlled
by removal of sialic acid termini on glycoproteins to expose
terminal galactose residues, allowing binding with the asia-
loglycoprotein receptor. It has been shown that asialogycopro-
teins are bound to liver membranes by a saturable, calcium-
dependent process. Bis-biotin blood-clearance reagent 2 was
designed to bind with SAv in antibody conjugates or fusion
proteins in the blood, then bind with the asialoglycoprotein
receptors in the liver to clear them from blood. The structure
of 2 was designed based on information provided in several
literature reports. Studies by Lee et al. reported that conjugation
of different monosacharides to proteins resulted in varying
binding affinities with isolated rabbit liver membranes (36). In
those studies, it was demonstrated that proteins conjugated with
thio-D-galactose moieties were effective at inhibiting binding
of a standard, [125I]iodo-asialoorosomucoid ([125I]ASOR). In-
deed, depending on the number of thio-D-galactosyl moieties
conjugated, measured relative inhibitory power (RIP) values
were as high as 10 000× the [125I]ASOR. The studies also
demonstrated that a nine-atom linker between the sugar and
protein provided a higher RIP then when there was a two- or
five-atom linker. Those early studies set a basis for the use of
galactosyl conjugates for removal of proteins from blood using
the asialoglycoprotein receptors and began to define some of
the structural requirements of the conjugates for efficient
binding.
The difficulty in synthesizing a bis-biotin counterpart to the
large clearing agent 1 led us to consider whether the 16
galactosyl moieties were required for effective blood clearance.
There are literature reports suggesting that as few as 3 galactosyl
moieties may be effective. In early studies, Lee recognized that
a number of biological systems require clustering of glycosides
for biological activity, and that such clustering was often
obtained from branched structures, so he prepared a number of
triglucose and trigalactose derivatives for study (37). Evaluation
of mono-, bis-, and tris-galactosyl derivatized bovine serum
albumin binding with rabbit liver membrane demonstrated that
the tris-galactosyl conjugates were better at inhibiting binding
of [125I]ASOR than the bis-galactose conjugate, which was better
than the monogalactose conjugate (38). In another study, Lee
et al. demonstrated that the number of branched galactose
moieties provided dramatic differences in inhibitory potency,
with tetra > trisfbisfmono (39). The 50% inhibitory value
was ∼1 mM for the monogalactose conjugate, whereas it was
∼1 nM for the trigalactose conjugate. This large difference in
binding inhibition based on the number of galactose residue in
branched chains was also noted by Bezouska et al. (40). It was
noted by Baenziger and Maynard that the galactose residues
may bind two sites simultaneously and speculated that the sites
might be 25-30 Å apart (41). In a later study, Lee et al.
synthesized “cluster ligands” with 1, 2, 3, 4, or 6 terminal
galactose moieties (42). They found that the highest binding
was obtained with the highest number of galactose moieties (i.e.,
6 moieties), but that the binding affinity increased only a modest
amount for the compounds containing 4 or 6 galactose moieties
relative to the high binding of the compound that had 3 galactose
moieties in a “flexible” structure. On the basis of their results,
it was hypothesized that there are three galactose binding sites
in the asialoglycoprotein receptor and the binding occurs in a
triangle configuration with distances between binding sites of
15, 22, and 25 Å. Some years later, Biessen et al. added
Figure 3. Biodistribution of 111In-labeled biotin derivatives in tumor-
bearing mice pretargeted with anti-CD20 scFv4-SAv fusion proteins.
Bar graph depicting tissue concentrations (%ID/g) of 111In-labeled biotin
derivatives [111In]3b or [111In]4b in mice bearing Ramos lymphoma
xenografts after administration of the following pretargeting reagents
and diets. Group 1 (red bars): 1F5 scFv4-SAv-WT, CA 1, [111In]3b,
biotin-deficient diet. Group 2 (black bars): 1F5 scFv4-SAv-WT, CA 1,
[
111In]3b, standard diet. Group 3 (green bars): 1F5 scFv4-SAv-S45A,
CA 2, [111In]4b, standard diet. Group 4 (blue bars): 1F5 scFv4-SAv-
Y43A, CA 2, [ 111In]4b, standard diet. In the pretargeting protocol, a
2.8 nmol quantity of the fusion protein was injected; then after 24 h,
5.8 nmol of either blood clearance agent, 1 or 2, was injected, followed
4 h later by injection of 1.2 nmol of the 111In-labeled biotin derivative,
[
111In]3b or [111In]4b. Mice were euthanized 24 h after 111In-labeled
biotin injection; and blood, tumor, and selected normal tissues were
harvested, weighed, and counted to determine %ID/g. Average %ID/g
values are shown ( standard deviation. The plotted data are provided
as Table S2 in the Supporting Information.
provides good tumor targeting of therapeutic radionuclides in
the presence of endogenous biotin, while having minimal
localization of the radionuclides in normal tissues. Although
endogenousbiotinhasnotprecludedtheuseofantibody-streptavidin
conjugates or fusion proteins in clinical studies (33), it has the
potential to significantly diminish the blood clearance and tumor
targeting of the radiolabeled biotin derivatives in patients. In
the pretargeting approach being investigated, a recombinant
tetrameric fusion protein, combining the scFv portion of the
anti-CD20 antibody 1F5 with streptavidin mutants that have
decreased binding for biotin, is used in combination with a bis-
biotin derivative that is radiolabeled with a diagnostic or
therapeutic radionuclide. Due to its bivalency, bis-biotin deriva-
tives have higher binding avidities than biotin, allowing their
preferential binding with SAv mutants in the presence of
endogenous biotin. The SAv mutants used in this investigation
have a serine or tyrosine in the biotin-binding pocket replaced
with an alanine residue (S45A and Y43A, respectively (17)).
In initial proof-of-principle studies, the SAv-Y43A mutant was
found to have a 67-fold lower affinity for biotin compared to
the wild-type (WT) SAv, and the off-rate for the bis-biotin was
found to be 640 times slower than that of biotin when binding
SAv-Y43A (21). This combination of reagents was previously
found to target radionuclides to lymphoma xenografts in the
presence of endogenous biotin, but normal tissues had significant
concentrations of the radionuclides (22). Thus, the primary goal
of this investigation was to design and synthesize a galactose-
containing bis-biotin reagent that would bind in vivo with a
scFv4-SAv fusion protein, such that more efficient blood
clearance of the fusion proteins could be attained. In addition
to preparation of an effective clearing agent, a bis-biotin
derivative that could be radiolabeled with 111In and 90Y was
prepared such that those radiometals could be used in the
pretargeting protocols with mutant SAv fusion proteins.
The majority of our pretargeting investigations have used the
biotin-hexadecyl-galactosamine reagent, 1 (also referred to as
N-acetylgalactosamine-biotin or NAGB; note: the N-acetylga-

1236 Bioconjugate Chem., Vol. 21, No. 7, 2010
Wilbur et al.
hydrophilic poly(ethylene glycol) spacers of varying lengths (4,
9, 10, 13, and 20 Å) between galactose moieties and the branch
point in trigalactose derivatives (or 38 Å between anomeric
oxygen atoms on galactose residues) and evaluated their
competitive binding vs [125I]ASOR to isolated parenchymal liver
cells (43). They found that the binding affinity of the trigalactose
cluster compounds was strongly correlated with the length of
the spacer between the galactosyl residues and the branching
point of the galactoside, with a factor of 2000× higher affinity
with the 20 Å spacer vs the 4 Å spacer.
The information from the literature provided the basis for
the structure of galactose-containing portion of 2. Our previous
studies demonstrating successful in vivo targeting of tumor
xenografts with a radioiodinated bis-biotin derivative led us to
retain the main structural features of that portion of the
compound. Those structural features included (1) appropriate
functional groups to provide a distance between biotin car-
boxyate carbonyl functionalities of ∼40 Å, (2) carboxylate
groups on the carbon alpha to the biotinamide functionalities,
and (3) diamino-dioxaoctane spacer arms to increase aqueous
solubility of compound. The distance between biotin moieties
was chosen as the distance that was adequate to readily bind
the two biotin-binding pockets on one face of the streptavidin
mutants (19). The addition of the carboxylates alpha to the
biotinamide provided protection from in vivo cleavage of the
biotinamide bonds by the enzyme biotinidase (28). The structural
properties that were previously shown to bind asialoglycoprotein
receptors in liver were combined with the bis-biotin structural
properties desired for binding with streptavidin mutants to design
blood clearance agent 2.
loglycoprotein receptor is directed at galactose derivatives, it
has been reported that the avidity of agents with terminal
N-acetylgalactosamine moieties is greater than those terminated
with galactose moieties (41). Indeed, studies have shown that
tris-N-acetyl-D-galactosamine derivatives can have subnanomo-
lar binding with the asialoglycoprotein receptor (45). Another
possible contributing factor is higher denticity (i.e., more
galactosamine residues), which could increase the rate of cell
internalization (20). Yet another possible contributing factor is
the binding avidity of the bis-biotin with the mutant fusion
protein. It seems likely that the difference in the binding of 1
with the mutant proteins contributes to the higher liver
concentrations, as the liver concentration of [111In]4b at the time
of sacrifice was about 2× higher when the fusion protein 1F5
scFv4-SAv-Y43A was used than when 1F5 scFv4-SAv-S45A
was used. It is not possible to determine the reason for the
differences in liver concentrations with the data obtained,
particularly with the complexity of the internalization process
(46) and receptor/ligand recycling that has been observed (47, 48).
Further studies need to be conducted to understand which
parameters affect the observed higher liver concentrations.
In conclusion, the pretargeting approach using a combination
of SAv mutant fusion proteins, bis-biotin-trigalactose blood
clearance agent 2, and 111In-labeled DOTA-bis-biotin, [111In]4b,
was found to target lymphoma tumor xenografts in high
concentrations, resulting in far superior biodistributions com-
pared with the 1F5 scFv4-SAv-WT conjugate system in the
presence of endogenous biotin. While the new blood clearance
agent, 2, has been shown to greatly decrease the radioactivity
concentrations in most normal tissues relative to CA 1, the fact
that higher blood and liver concentrations were obtained
indicates that the structure of 2 is not optimal. Optimization of
the structure of 2 might include use of N-acetylgalactosamine
residues in the place of galactose residues and incorporation of
a higher number of galactosamine residues (i.e., 4-6). While
CA 2 is not optimized, we believe the results obtained are
adequate to support investigation of 2 and 4a in pretargeting
therapy studies using scFv4-SAv mutant fusion proteins in
tumor-bearing mice fed standard diets.
An investigation of pretargeting using bis-biotin derivatives
2 and [111In]4b was conducted in mice. The quantities of bis-
biotin clearing agent 2 (5.8 nmol) and radiolabeled bis-biotin
[
111In]4b (1.2 nmol) used in this investigation were the same
as in the previous studies employing fusion proteins constructed
with wild-type (WT) SAv, the NAGB clearing agent, 1, and
radiolabeled DOTA-biotin, [111In]3b. The quantity of those
reagents and the times for their administration were optimized
in a series of experiments (18, 44). Those optimization studies
were not repeated in this investigation; instead, a limited series
of experiments varying the doses of mutant SAv fusion protein
and 2 within a more narrow range were conducted. Those studies
verified that the quantities previously optimized also worked
optimally with the new reagents.
ACKNOWLEDGMENT
We thank Quanta BioDesign (Powell, Ohio) for providing
D-biotin, 9, N-Boc-trioxatridecanediamine, 16, and N-(phthal-
imido)amino-dPEG4-acid, 19, used in the studies. We thank Dr.
Patrick Stayton (Bioengineering, University of Washington,
Seattle, WA) for assistance in the preparation and purification
of the fusion proteins containing mutant streptavidins S45A and
Y43A. We thank NIH (P01 CA44991, R01 CA76287, R01
109663, R01 136639, R01 CA113431, and K08 CA095448)
and Lymphoma Research Foundation (Postdoctoral Fellowship
Grant # 82360) for funding the studies.
A blood clearance study demonstrated that the fusion protein
1F5 [125I]scFv4-S45A could be more efficiently cleared from
blood in mice after administration of 2 than when 1 was
administered. This is shown in Figure 2. Importantly, use of 2
in a pretargeting study, where both 1F5 scFv4-SAv-S45A and
1F5 scFv4-SAv-Y43A fusion proteins (in separate groups) were
evaluated with bis-biotin [111In]4b, resulted in normal tissue
concentrations that were much lower than obtained in our
previous pretargeting studies employing those fusion proteins
with CA 1 and a radioiodinated bis-biotin derivative. The tissue
concentration data obtained in the pretargeting study are shown
as a bar graph in Figure 3. The most striking differences between
Supporting Information Available: Tables S1 and S2
provide the data plotted to prepare Figures 2 and 3, and Table
S3 provides biodistribution data for [111In]4b. This material is
available free of charge via the Internet at http://pubs.acs.org.
[
111In]4b and [111In]3b concentrations in tissues were found in
LITERATURE CITED
the blood and liver. The higher blood concentrations are most
likely brought about by less efficient blood clearance using 2;
however, it is unclear why the liver concentrations of [111In]4b
are somewhat higher. There may be several possible contributing
factors to the observed differences in liver concentrations. For
example, a contributing factor may be the difference in the
nature of the galactose derivative in the clearing agents, 1 and
2. In 1, the galactose derivative is an N-acetyl-galactosamine
moiety, whereas in 2, it is a galactose moiety. Although most
literature relating to developing ligands for binding to asia-
(1) Press, O. W. (2008) Evidence mounts for the efficacy of
radioimmunotherapy for B-cell lymphomas. J. Clin. Oncol. 26,
5147–50.
(2) Press, O. W. (2003) Radioimmunotherapy for non-Hodgkin’s
lymphomas: a historical perspective. Semin. Oncol. 30, 10–21.
(3) Sharkey, R. M., Press, O. W., and Goldenberg, D. M. (2009)
A re-examination of radioimmunotherapy in the treatment of non-
Hodgkin lymphoma: prospects for dual-targeted antibody/radio-
antibody therapy. Blood 113, 3891–5.

Bis-Biotin Reagents for Pretargeting
Bioconjugate Chem., Vol. 21, No. 7, 2010 1237
(4) Buchegger, F., Press, O. W., Delaloye, A. B., and Ketterer, N.
(2008) Radiolabeled and native antibodies and the prospect of
cure of follicular lymphoma. Oncologist 13, 657–67.
(5) Green, D. J., Pagel, J. M., Pantelias, A., Hedin, N., Lin, Y.,
Wilbur, D. S., Gopal, A., Hamlin, D. K., and Press, O. W. (2007)
Pretargeted radioimmunotherapy for B-cell lymphomas. Clin.
Cancer Res. 13, 5598s–5603s.
(6) Sharkey, R. M., Karacay, H., Cardillo, T. M., Chang, C. H.,
McBride, W. J., Rossi, E. A., Horak, I. D., and Goldenberg,
D. M. (2005) Improving the delivery of radionuclides for imaging
and therapy of cancer using pretargeting methods. Clin. Cancer
Res. 11, 7109s–7121s.
design and use of multivalent ligands and inhibitors. Angew.
Chem., Int. Ed. 37, 2755–2794.
(21) Hamblett, K. J., Kegley, B. B., Hamlin, D. K., Chyan, M. K.,
Hyre, D. E., Press, O. W., Wilbur, D. S., and Stayton, P. S. (2002)
A streptavidin-biotin binding system that minimizes blocking by
endogenous biotin. Bioconjugate Chem. 13, 588–98.
(22) Hamblett, K. J., Press, O. W., Meyer, D. L., Hamlin, D. K.,
Axworthy, D., Wilbur, D. S., and Stayton, P. S. (2005) Role of
biotin-binding affinity in streptavidin-based pretargeted radio-
immunotherapy of lymphoma. Bioconjugate Chem. 16, 131–8.
(23) Yau, E. K., Theodore, L. J., and Gustavson, L. M. (1996)
Pretargeting Methods and Compounds, U. S. Patent 5,541,287.
(24) Gamper, H. B., Reed, M. W., Cox, T., Virosco, J. S., Adams,
A. D., Gall, A. A., Scholler, J. K., and Meyer, R. B., Jr. (1993)
Facile preparation of nuclease resistant 3′ modified oligodeoxy-
nucleotides. Nucleic Acids Res. 21, 145–50.
(25) Wilbur, D. S., Hamlin, D. K., Vessella, R. L., Stray, J. E.,
Buhler, K. R., Stayton, P. S., Klumb, L. A., Pathare, P. M., and
Weerawarna, S. A. (1996) Antibody fragments in tumor pretar-
geting. Evaluation of biotinylated Fab’ colocalization with
recombinant streptavidin and avidin. Bioconjugate Chem. 7, 689–
702.
(26) Sigal, G. B., Mammen, M., Dahmann, G., and Whitesides,
G. M. (1996) Polyacrylamides bearing pendant asialoside groups
strongly inhibit agglutination of erythrocytes by influenza virus:
the strong inhibition reflects enhanced binding through coopera-
tive polyvalent interactions. J. Am. Chem. Soc. 118, 3789–3800.
(27) Braun, M., Camps, X., Vostrowsky, O., Hirsch, A., Endress,
E., Bayerl, T. M., Birkert, O., and Gauglitz, G. (2000) Synthesis
of a biotinylated lipofullerene as a new type of transmembrane
anchor. Eur. J. Org. Chem. 1173–1182.
(28) Wilbur, D. S., Hamlin, D. K., Chyan, M. K., Kegley, B. B.,
and Pathare, P. M. (2001) Biotin reagents for antibody pretar-
geting. 5. Additional studies of biotin conjugate design to provide
biotinidase stability. Bioconjugate Chem. 12, 616–23.
(29) Theodore, L. J., and Axworthy, D. B. (2001) Cluster clearing
agents, U.S. Patent 6,172,045.
(30) Ponpipom, M. M., Bugianesi, R. L., Robbins, J. C., Doebber,
T. W., and Shen, T. Y. (1981) Cell-specific ligands for selective
drug delivery to tissues and organs. J. Med. Chem. 24, 1388–95.
(31) Sandmaier, B. M., Bethge, W. A., Wilbur, D. S., Hamlin,
D. K., Santos, E. B., Brechbiel, M. W., Fisher, D. R., and Storb,
R. (2002) Bismuth 213-labeled anti-CD45 radioimmunoconjugate
to condition dogs for nonmyeloablative allogeneic marrow grafts.
Blood 100, 318–26.
(32) Press, O. W., Eary, J. F., Badger, C. C., Martin, P. J.,
Appelbaum, F. R., Levy, R., Miller, R., Brown, S., Nelp, W. B.,
Krohn, K. A., Fisher, D., DeSantes, K., Porter, B., Kidd, P.,
Thomas, E. D., and Berstein, I. D. (1989) Treatment of refractory
non-Hodgkin’s lymphoma with radiolabeled MB-1 (Anti-CD37)
antibody. J. Clin. Oncol. 7, 1027–2038.
(7) Goodwin, D. A., and Meares, C. F. (1997) Pretargeting: General
Principles. Cancer (Suppl.) 80, 2675–2680.
(8) Axworthy, D. B., Fritzberg, A. R., Hylarides, M. D., Mallett,
R. W., Theodore, L. J., Gustavson, L. M., Su, F.-M., Beaumier,
P. L., and Reno, J. M. (1994) Preclinical evaluation of an anti-
tumor monoclonal antibody/streptavidin conjugate for pretargeted
90Y radioimmunotherapy in a mouse xenograft model. J. Im-
munother. 16, 158.
(9) Axworthy, D. B., Reno, J. M., Hylarides, M. D., Mallett, R. W.,
Theodore, L. J., Gustavson, L. M., Su, F., Hobson, L. J.,
Beaumier, P. L., and Fritzberg, A. R. (2000) Cure of human
carcinoma xenografts by a single dose of pretargeted yttrium-
90 with negligible toxicity. Proc. Natl. Acad. Sci. U.S.A. 97,
1802–7.
(10) Theodore, L. J., Fritzberg, A. R., Schultz, J. E., and Axworthy,
D. B. (2000) Evolution of a pretarget radioimmunotherapeutic
regimen, in Radioimmunotherapy of Cancer (Abrams, P. G., and
Fritzberg, A. R., Eds.) pp 195-221, Marcel Dekker, New York.
(11) Press, O. W., Corcoran, M., Subbiah, K., Hamlin, D. K.,
Wilbur, D. S., Johnson, T., Theodore, L., Yau, E., Mallett, R.,
Meyer, D. L., and Axworthy, D. (2001) A comparative evaluation
of conventional and pretargeted radioimmunotherapy of CD20-
expressing lymphoma xenografts. Blood 98, 2535–2543.
(12) Pagel, J. M., Pantelias, A., Hedin, N., Wilbur, S., Saganic,
L., Lin, Y., Axworthy, D., Hamlin, D. K., Wilbur, D. S., Gopal,
A. K., and Press, O. W. (2007) Evaluation of CD20, CD22, and
HLA-DR targeting for radioimmunotherapy of B-cell lympho-
mas. Cancer Res. 67, 5921–8.
(13) Pagel, J. M., Hedin, N., Subbiah, K., Meyer, D., Mallet, R.,
Axworthy, D., Theodore, L. J., Wilbur, D. S., Matthews, D. C.,
and Press, O. W. (2003) Comparison of anti-CD20 and anti-
CD45 antibodies for conventional and pretargeted radioimmu-
notherapy of B-cell lymphomas. Blood 101, 2340–8.
(14) Pagel, J. M., Orgun, N., Hamlin, D. K., Wilbur, D. S., Gooley,
T. A., Gopal, A. K., Park, S. I., Green, D. J., Lin, Y., and Press,
O. W. (2009) A comparative analysis of conventional and
pretargeted radioimmunotherapy of B-cell lymphomas by target-
ing CD20, CD22, and HLA-DR singly and in combinations.
Blood 113, 4903–13.
(15) Baker, H. (1985) Assessment of biotin status: clinical implica-
tions. Ann. N.Y. Acad. Sci. 447, 129–132.
(16) Rusckowski, M., Fogarasi, M., Virzi, F., and Hnatowich, D. J.
(1995) Influence of endogenous biotin on the biodistribution of
labelled biotin derivatives in mice. Nucl. Med. Commun. 16, 38–
46.
(17) Klumb, L. A., Chu, V., and Stayton, P. S. (1998) Energetic
roles of hydrogen bonds at the ureido oxygen binding pocket in
the streptavidin-biotin complex. Biochemistry 37, 7657–7663.
(18) Lin, Y., Pagel, J. M., Axworthy, D., Pantelias, A., Hedin, N.,
and Press, O. W. (2006) A genetically engineered anti-CD45
single-chain antibody-streptavidin fusion protein for pretargeted
radioimmunotherapy of hematologic malignancies. Cancer Res.
66, 3884–92.
(19) Wilbur, D. S., Pathare, P. M., Hamlin, D. K., and Weerawarna,
S. A. (1997) Biotin reagents for antibody pretargeting. 2.
synthesis and in vitro evaluation of biotin dimers and trimers
for crosslinking of streptavidin. Bioconjugate Chem. 8, 819–832.
(20) Mammen, M., Chio, S.-K., and Whitesides, G. M. (1998)
Polyvalent interactions in biological systems: implications for
(33) Knox, S. J., Goris, M. L., Tempero, M., Weiden, P. L.,
Gentner, L., Breitz, H., Adams, G. P., Axworthy, D., Gaffigan,
S., Bryan, K., Fisher, D. R., Colcher, D., Horak, I. D., and
Weiner, L. M. (2000) Phase II trial of yttrium-90-DOTA-biotin
pretargeted by NR-LU-10 antibody/streptavidin in patients with
metastatic colon cancer. Clin. Cancer Res. 6, 406–414.
(34) Ashwell, G., and Morell, A. G. (1974) The role of surface
carbohydrates in the hepatic recognition and transport of circulat-
ing glycoproteins. AdV. Enzymol. 41, 99–128.
(35) Ashwell, G., and Harford, J. (1982) Carbohydrate-specific
receptors of the liver. Annu. ReV. Biochem. 51, 531–54.
(36) Lee, R. T., and Lee, Y. C. (1980) Preparation and some
biochemical properties of neoglycoproteins produced by reductive
amination of thioglycosides containing an omega-aldehydoag-
lycon. Biochemistry 19, 156–63.
(37) Lee, Y. C. (1978) Synthesis of some cluster glycosides suitable
for attachment to proteins or solid matrices. Carbohydr. Res.
67, 509–514.
(38) Kawaguchi, K., Kuhlenschmidt, M., Roseman, S., and Lee,
Y. C. (1980) Synthesis of some cluster galactosides and their

1238 Bioconjugate Chem., Vol. 21, No. 7, 2010
Wilbur et al.
effect on the hepatic galactose-binding system. Arch. Biochem.
Biophys. 205, 388–95.
(39) Lee, Y. C., Townsend, R. R., Hardy, M. R., Lonngren, J.,
Arnarp, J., Haraldsson, M., and Lonn, H. (1983) Binding of
synthetic oligosaccharides to the hepatic Gal/GalNAc lectin.
Dependence on fine structural features. J. Biol. Chem. 258, 199–
202.
fusion protein and an antibody-streptavidin chemical conjugate
for pretargeted anti-CD20 radioimmunotherapy of B-cell lym-
phomas. Blood 108, 328–36.
(45) Lee, R. T., and Lee, Y. C. (1987) Preparation of cluster
glycosides of N-acetylgalactosamine that have subnanomolar
binding constants towards the mammalian hepatic Gal/GalNAc-
specific receptor. Glycoconjugate J. 4, 317–328.
(40) Bezouska, K., Ta´borsky, O., Kubrycht, J., Pospisil, M., and
Kocourek, J. (1985) Carbohydrate-structure dependent recogni-
tion of desialylated serum glycoproteins in the liver and
leucocyes. Biochem. J. 227, 345–354.
(46) Wall, D. A., Wilson, G., and Hubbard, A. L. (1980) The
galactose-specific recognition system of mammalian liver: the
route of ligand internalization in rat hepatocytes. Cell (Cam-
bridge, Mass.) 21, 79–93.
(41) Baenziger, J. U., and Maynard, Y. (1980) Human hepatic
lectin. Physiochemical properties and specificity. J. Biol. Chem.
255, 4607–13.
(42) Lee, R. T., Lin, P., and Lee, Y. C. (1984) New synthetic cluster
ligands for galactose/N-acetylgalactosamine-specific lectin of
mammalian liver. Biochemistry 23, 4255–61.
(43) Biessen, E. A. L., Beuting, D. M., Roelen, H. C. P. F., van de
Marel, G. A., Van Boom, J. H., and Van Berkel, T. J. C. (1995)
Synthesis of cluster galactosides with high affinity for the hepatic
asialoglycoprotein receptor. J. Med. Chem. 38, 1538–46.
(44) Pagel, J. M., Lin, Y., Hedin, N., Pantelias, A., Axworthy, D.,
Stone, D., Hamlin, D. K., Wilbur, D. S., and Press, O. W. (2006)
Comparison of a tetravalent single-chain antibody-streptavidin
(47) Oka, J. A., and Weigel, P. H. (1983) Recycling of the
asialoglycoprotein receptor in isolated rat hepatocytes. Dissocia-
tion of internalized ligand from receptor occurs in two kinetically
and thermally distinguishable compartments. J. Biol. Chem. 258,
10253–62.
(48) Townsend, R. R., Wall, D. A., Hubbard, A. L., and Lee, Y. C.
(1984) Rapid release of galactose-terminated ligands after
endocytosis by hepatic parenchymal cells: evidence for a role
of carbohydrate structure in the release of internalized ligand
from receptor. Proc. Natl. Acad. Sci. U. S. A. 81, 466–70.
BC100030Q