The avidin-biotin complex as a tool in molecular biology

Full text

METHODS
OF
BIOCHEMICAL ANALYSIS VOLUME
26
The
Use
of
the Avidin-Biotin Complex
as
a
Tool in Molecular Biology
EDWARD
A
.
BAYER
and
MEIR
WILCHEK.
Department
ofBaophysics.
The
Weizmunn
Institute
of
Science.
Rehwot.
Israel
I
.
I1
.
111
.
IV
.
v
.
VI
.
Introduction
...............................................
Principle
...................................................
5
1
.
Amino Reagents
7
Preparation of Reactive Biotinyl Derivatives
...............................
2
.
Carboxyl and Sugar Reagents
.......................................
7
3
.
Thiol Reagents
....................................................
8
4
.
Phenol and Imidazole Reagents
......................................
8
Assays for Avidin and Biotin
............................................
Purification Studies: Affinity Chromatography
..........................
9
1
.
Natural Biotin-Containing Systems
...................................
9
2
.
"Tailor-Made" Biotin-Containing Systems
............................
11
3
.
Experimental Procedures
...........................................
14
A
.
Preparation of Affinity Columns
...............................
14
a
.
Activation of Sepharose
.................................
14
b
.
Biotin Affinity Columns
................................
14
Biocytin
............................................
14
Biocytin Sepharose
...................................
14
Polymeric Biotin-Containing Columns
..................
14
c
.
Avidin Affinity Columns
................................
15
Cyanogen Bromide Induced Immobilization
............
15
Periodate-Induced Immobilization
.....................
15
Other
Coupling Methods
.............................
15
B
.
Isolation Procedures
.........................................
15
Locathation Studies: Affinity Cytochemistry
.............................
16
1
.
General Considerations
.............................................
16
B
.
Amino Acid Residues
....................
............
20
3
.
Localization of Receptors
...........................................
22
4
.
Other Systems
........................
.....................
23
5
.
Restrictions of the Method
..........................................
25
6
.
Experimental Procedures
...........................................
26
...................................................
9
...................................
2
.
Localization of Functional Groups
17
A
.
Sugars
......................................................
17
A
.
Preparation
of
Avidin-Conjugated Markers
.....................
26
a
.
Fenitin-Avidin Conjugates via Reductive Alkylation
.......
26
1
Methods
of
Biochemical Analysis, b701ume
26
Edited by David Glick
Copyright
0
1980
by
John
Wiley
&
Sons, Inc.

2
EDWARD
A.
BAYER
AND
MEIR
WILCHEK
b.
Glutaraldehyde Method
.......
.....................
26
c. Other Coupling Methods
.....
d. Other Markers
..............
e.
Analysis of Conjugates
...............
B.
Direct Biotinylation of Functional Groups
a.
Sialic Acid Residues
.............
b. Galactose and N-Acetylgalactosa
Biotinylation of Biologically Active
Pro
C.
D.
Interaction Between Biotinylated Binding Protein and Receptor
...
29
VII.
Miscellaneous Systems ............................
1.
Phage Inactivation Studies
........
B.
Experimental Procedures
..........
................
29
2.
Lymphocyte Stimulation
.................
......................
31
3.
Hormone-Receptor Interactions .....................................
33
IX.
Antibiotin Antibodies
..............................................
40
X.
Conclusions ...............................................
...
41
VIII.
The
Biotin Transport
S
m:
An Affinity Labeling Study
................
35
Acknowledgments
.......................
.......................
42
References
......
...............................................
42
I.
INTRODUCTION
The high affinity constant between the glycoprotein avidin and the
vitamin
biotin prompted early attention to the nature of this complex.
To
obtain further insight into the properties of the avidin-biotin complex, in
the
early 1950s Fraenkel-Conrat and co-workers
(
1952) purified avidin
and studied the effect of chemical modification on
its
activity.
No
further
interest was taken in the complex until the end of the decade when Wakil
et al. (1958) and Lynen et
al.
(1959) discovered the coenzyme function of
covalently bound biotin. It became clear that avidin could
be
used as a tool
for characterizing biotin-requiring enzymes. In fact, the spatial position
of the avidin-bound biotin-containing subunit
of
transcarboxylase was
ultimately localized by high resolution electron microscopy (Green
et
al.,
1972;
Green, 1972).
Sice
1963
Green has been the leading figure in the efforts to
understand this unique interaction by various biophysical and
biochemical methods (Green, 1975). However the innate reason for the
strong interaction between biotin and avidin is not yet known. Judging
from the structure of biotin (Figure
l),
it
is
difficult
to
understand why
such a simple molecule should
possess
such
an
unprecedented affinity for
a given protein. Even more surprising, only the intact ureido ring is

THE
USE
OF THE
AVIDIN-BIOTIN
COMPLEX
Figure
1.
The structure
of
biotin.
3
required for this strong interaction.
Regarding the other partner of this complex (Table
I),
it
is surprising
that the four tryptophan residues of each avidin subunit vie for the biotin
molecule. There is no perceivable reason for tryptophan, which generally
participates in charge-transfer complexes
or
hydrophobic interactions, to
have affinity for the ureido group instead
of
other more hydrophobic
components
of
the biotin molecule. Nevertheless, even though we do not
yet fully understand this interaction, it provides a powerful tool for study
in the following areas:
(1)
the isolation of biotin-derivatized materials by
affinity chromatography,
(2)
affinity labeling and identification studies,
(3)
affinity cytochemical labeling for localization studies by fluorescence
and electron microscopy,
(4)
the inhibition of bacteriophages, and
(5)
the
study of cell surface molecular interactions.
In
this respect the avidin-biotin complex represents
a
complementary
approach and/or a potential replacement for lectins and antibodies in
biological interactions that exploit the specific binding between a protein
and a ligand. This chapter describes in more detail previous contributions
to the application of the avidin-biotin complex and provides some
suggestions about the direction of its prospective use. Naturally, we will be
unable to cover all possible applications; it seems that the potential of the
TABLE
I
Some Important Characteristics
of
Avidin"
Molecular weight
Subunit molecular weight
KD
(avidin-biotin complex)
&A
1
mg/ml)
Oligosaccharidelsubunit
Mannose/subunit
Glucosamine/subunit
Tryptophan/subunit
67,000
15,600
-
10-15
1.54
96,000
1
4-5
3
4
"Modified from Green
(1965).

4
EDWARD
A.
BAYER
AND
MEIR
WILCHEK
avidin-biotin complex as a tool in molecular biology is unlimited, and that
its successful implementation is directly dependent on the needs and
imagination
of
the user.
11.
PRINCIPLE
The rationale behind our approach is
as
follows: biotin, bound to a
macromolecule,
is
still
available for the high affinity interaction with
avidin (Becker and Wilchek,
1972).
Thus (in addition to biotin-requiring
enzymes) biotin-derivatized hormones, phages, lectins, antibodies, and
other binding proteins can interact with avidin; and if the avidin
is
immobilized
or
covalently bound
to
a
potentially perceptible probe, the
avidin-biotin complex can
be
used for the localization
or
isolation of the
compounds
above
and/or their receptors (Figure
2).
The major
B-X/
\AB
LEGEND
@
-
MEMBRANE RECEPTOR
8-X
-
BlOTlNY
L-
REAGENT
3
-
BIOTINYLATED BINDING PROTEIN
@P
-
AVIOIN
PmBE
@
Im-
IMMOBILIZED
AVIDIN
Figure
2.
complex as a probe
in
molecular biology.
Schematic representation
of
the rationale
behind
the
use
of
the avidin-biotin

THE
USE
OF
THE
AVIDIN-BIOTIN
COMPLEX
5
restriction concerns methods for the introduction (attachment) of biotin
to a given component of the experimental system.
Only in one case-that of biotin-requiring enzymes--has nature
provided us with a native, covalently bound, biotinylated protein. Various
laboratories have demonstrated through the years that the biotin moiety
of these proteins is capable of interacting with avidin (Knappe, 1970;
Moss
and Lane, 197
1).
Accordingly, avidin has been used for the isolation
and structural determination
of
the biotin-containing subunits.
Therefore, in other
cases
it is necessary to devise methods for the artificial
emplacement
or
covalent attachment of the biotin molecule to a specified
component of the experimental system. The latter
is
subsequently
evaluated by an appropriate avidintontaining conjugate.
111. PREPARATION
OF
REACTIVE
BIOTINYL DEWATIVES
Proteins contain a variety
of
functional groups, some
of
which are
important
for
their activity and some not. In any given protein, chemical
modification of an essential functional group may destroy directly or
indirectly, its biological activity and/or specificity. Since we are interested
in preserving these properties of the protein, a selection of groupspecific
reagents must
be
available. Therefore it would
be
advantageous to have
biotinyl derivatives that can
be
bound’
to
different
classes
of functional
groups. If a given biotinyl derivative interferes with the biological activity
or specificity of a modified protein, an alternative derivative
can
be
used
in its place. Accordingly, we have prepared a selection of biotinyl
derivatives that can
be
covalently bound to a variety of functional groups,
including mines, thiols, imidazoles, and phenols,
as
well
as
carboxyls.
Since many of the important cell receptors are glycoproteins, biotin
derivatives that can interact with sugar residues have also been prepared.
Some of the biotin derivatives that we have found useful are listed in the
scheme in Figure
3,
and the methods of preparation of a selected few are
summarized in Sections
111.1
to 111.4. It should
be
noted that these
reagents are not only applicable for direct coupling to a protein, but also
can
be
used after prior enrichment of a given functional group
and
subsequent attachment with an appropriate biotinyl derivative to the
extraneous functional group. For example, thiolylation of a protein with
homocysteine lactone generates free sulfiydryl groups, which sub-
sequently
can
be
reacted with a bromoacetyl
analog
of biotin.
A
second
example
is
the biotinylation
of
a glycoprotein by way of the oligo-
saccharide moiety, since the first step in such a procedure consists
of
periodate-oxidation of vicinal hydroxyls to aldehydes. The latter
can
be

r
On
OI
I
v-v-z
0
0
I
O=U
o=v
r
0
o=
v
0"
r
Q
+5/
z
I
V
r
z
0=2
6-$
rI
0-0
o=v
I
z
z
I
0=0
\
r
0,
6
o=v
I
I
z
I
z
I
o=v
/
6

THE
USE
OF
THE
AVIDIN-BIOTIN COMPLEX
7
used to introduce different functional groups to which a suitable biotin
derivative subsequently can be bound.
1.
AminoReagents
1.
Biotinyl-N-hydroxysuccinimzde
ester
(BNHS).
The method described
by Becker et al. (1971) and Bayer and Wilchek (1974) is as follows. Di-
cyclohexylcarhdiimide
(0.8
g) was added to a solution
of
dimethyl-
formamide (12 ml) containing biotin
(
1 g) and N-hydroxysuccinimide (0.6
g).
The suspension was stirred overnight at room temperature. The
dicyclohexylurea precipitate was filtered and the filtrate evaporated
under reduced pressure. The residue was washed well with ether, and the
product was recrystallized from isopropanol.
An alternative method for preparation of this compound was described
by Jasiewicz
et
al.
(1976) by use of
N,N'-carbonyl-diimidazole
as a
coupling reagent.
2.
Bwtinyl-p-nitrophyrophenyl
ester
(pBNP). Bayer and Wilchek (1977)
synthesized p-BNP
by
a slight modification
of
the earlier procedure
described by Becker et al. (1971). To biotin
(244
mg
1.0
mmole)
suspended in
3
ml of methylene chloride was added p-nitrophenol(l75
mg,
1.3
mmole) and
dicyclohexylcarbodiimide
(DCC) (206 mg, 1 mmole).
After stirring this mixture
for
24
hr at 25"C, it was filtered and the filtrate
taken to dryness under reduced pressure. The yellow gummy residue was
washed several times with absolute ether, and taken up in isopropanol.
Following filtration, the solution was reduced to minimum volume and
allowed to crystallize overnight. The crystals of pBNP were collected by
filtration and washed with anhydrous ether.
Another preparative procedure for this compound was described
recently by Bodanszky and Fagan (1 977). The
o-
and m-nitrophenyl esters
of
biotin can be prepared in a similar manner.
2.
Carboxyl
and
Sugar
Reagents
Biotin
hydrazide
(BHZ) was synthesized by a modification
of
the
procedure used by Heitzmann and Richards (1974). Thionyl chloride
(1
ml) was added slowly to a chilled solution
(10
ml)
of
methanol (in an
ice-saline bath).
To
this solution biotin (1 g) was added, and
it
was left
overnight at room temperature. The solvent was evaporated to dryness.
Methanol (10 ml) was added, and the solvent was again evaporated to
dryness. The residue was redissolved in
5
ml of methanol; hydrazine
hydrate
(1
ml) was added, and the reaction was allowed to proceed
overnight at room temperature. The precipitate (biotin hydrazide) was
filtered and washed with ether.
A
second crop may
be
obtained
by
con-

8
EDWARD
A.
BAYER
AND
MEIR
WILCHEK
centration
of
the filtrate. The samples were recrystallized from dimethyl-
formamide.
Biotin hydrazide can
be
used directly with the aldehyde derivatives of
periodate-oxidized sugars.
For
reaction with carboxyl groups, car-
bodiimides (water soluble
or
otherwise) must
be
added as a coupling
reagent. Other biotin derivatives for carboxyl groups can
be
prepared by
monosubstitution of biotin to diamines and coupling of the latter
derivatives to carboxyls via carbodiimide.
3.
ThiolReagents
Thiol reagents were prepared either by substitution of biotin hydrazide
or
monosubstituted biotinyl-amines with
bromoacetyl-N-hydroxysuc-
cinimide ester,
or
with bromoacetic anhydride (Wilchek and Givol,
1977).
Biotinyl derivatives containing mercury were prepared by coupling
biotin-hydrazide with the N-hydroxysuccinimide ester of phydroxy-
mercuribenzoate.
Bbtznyl-bTomoaceiyl
hyharde.
Biotin hydrazide
(260
mg) was dissolved in
0.5M
sodium bicarbonate
(10
ml) and treated with bromoacetic an-
hydride
(520
mg) in
4
ml of dioxane at
0°C.
After
15
min the precipitate
was
filtered off, dissolved in isopropanol and precipitated with ether.
4.
Phenol
and
ImidPzole Reagents
Biotinyl reagents for the phenol and imidazole reagent functional
groups were prepared by reduction of biotinyl-pnitroanilide with sodium
dithionite, followed by diazotization with sodium nitrite. This reaction
should be performed immediately before use.
1.
Bbtznyl-pnitroal2i.
Biotin
(244
mg) was dissolved in dimethyl-
formamide
(3
ml), and triethylamine
(0.14
ml)
was added. The solution
was cooled, and isobutylchloroformate
(0.16
ml) was added. After
5
min,
pnitroaniline
(1
50
mg) was added. The reaction mixture was left
at
room
temperature overnight. Upon adidtion of ethyl acetate, the product
precipitated and
was
collected. The product was recrystallized from
isopropanol.
2.
Biotinyl-diaroanilh.
Biotinyl-p-nitroanilide was dissolved in
dimethylformamide and water was added until the suspension became
slightly turbid.
To
this suspension, an excess of crystalline dithionite was
added. After
10
min the solution was acidified with
IN
hydrochloric acid
to pH
2.0
and treated with sodium nitrate at
0°C.
After
5
min a sample was
brought
to
pH
8.0
and reacted with phenol
or
imidazole. The appearance
of
a deep yellow
or
reddish color is a sign of reaction.

THE
USE
OF
THE
AVIDIN-BIOTIN
COMPLEX
9
IV. ASSAYS
FOR
AVIDIN AND BIOTIN
A
variety
of
method
for
the assay
of
avidin and/or biotin are presently
available. Biotin may
be
analyzed by a selection of biological procedures
based on the use of appropriate microorganisms (see McCormick and
Wright,
1970).
Biological assays, although by far the most sensitive
(representative lower limit:
10
pg of biotin), are cumbersome to perform
and typically take several days
to
obtain results. Biotin content may also
be
analyzed chemically, using
pdimethylaminocinnamaldehyde
(McCormic
and Roth, 1970).
Avidin may
be
assayed using ['4C]-biotin to a lower limit
of
20 ng
of
avidin (Wei, 1970). Since avidin is highly antigenic (Korenman and
OMalley,
1970),
the avidin content
of
a given solution can
be
assessed by
radioimmunoassay
.
Reciprocal methods
for
the detection of either avidin
or
biotin are also
available, although the sensitivity is generally reduced. One method is
based on the increased absorbance
of
avidin at 233 nm
(heZz3
=
2.4
x
104M-'/mole biotin) upon complex formation with biotin (Green,
1970).
The quantitative displacement by biotin of the avidin-dye (4'-hydroxy-
azobenzene-2-carboxylic acid) complex forms the basis
for
another assay
of
both
biotin and avidin (Green,
1970).
The quenching of tryptophan
fluorescene in avidin upon complex formation with biotin provides yet
another rapid and sensitive assay for both biotin and avidin in solutions
free
of
fluoroescent contaminants (Lin and Kirsch,
1977).
The latter
procedure affords improved sensitivity, and
free
biotin may
be
determined in amounts as small as 20 ng. Biotin
(5-10
ng/ml)
or
avidin
(100-300 ng/ml) also
can
be
assayed by the phage technique (Becker and
Wilchek, 1972).
V. PURIFICATION STUDIES:
AFFINITY
CHROMATOGRAPHY
1.
Natural
Biotin-Containing
Systems
It is interesting
to
note that in the early attempts at specific isolation of
biologically active compounds (affinity chromatography), the avidin-
biotin complex was used as a model system for demonstrating the appli-
cability
of
such an approach. Again, the reason for this is the high affinity
constant, that applies even under the most unfavorable conditions.
Biotin was first coupled to cellulose by way
of
an ester bond, and some
retardation of avidin was obtained on such columns (McCormick,
1965).
In the first study using Sepharose
as
a carrier
for
affinity chromatography
(Cuatrecasas and Wilchek,
1968),
this complex was again used to show the
superiority
of
this matrix over previously used camers. Thus when

0.2
0.
I
d
E
0
W
N
0
2
i
a
m
a
m
a
$
a
01
0.:
A.
6M
Gtmnidine
-HC,
p
n
1.5
nct,
pnzo
Acetic
Acid,
pnJ.0
Acetote
h
,HC
::4y',
.*
I
nct,
pnzo
I
I\
*,
.
I
L
I.
\
8.
6M
Gwtidine-HCI
unsubslirulrd
6M
(iwnidihe
-
HCl
4
EFFLUENT,
ML
Figure
4.
Affinity chromatography
of
commercial avidin on biocytin-Sepharose
(A$)
and
unsubstituted Sepharose
(C)
columns.
The
columns
(0.5
X
5
cm)
were equilibrated with
0.2
M
sodium bicarbonate, pH
8.7,
and
0.75
mg
of avidin (in
0.5
mi of the same buffer) was
attempted by varying the conditions as indicated
(arrows).
The small protein peak that
emerges early in
A
and
B
represents an impurity that does not bind ['%]biotin.
10

THE
USE
OF
THE
AVIDIN-BIOTIN
COMPLEX
11
biocytin was coupled to cyanogen bromide activated Sepharose and a
solution of avidin was passed over the column, the avidin was
so
strongly
adsorbed that no single agent (e.g., salts, acids, bases,
or
concentrated
solutions of biotin) was effective in eluting the avidin from the biotin
column. Only the combination of
6
M
guanidine-hydrochloride at pH 1.5
was capable of eluting the avidin (Figure
4).
Even under these drastic
conditions, however, avidin could
be
purified 4000-fold directly from egg
white without appreciable loss of biological activity.
The reverse approach was also taken to purify biotin-containing
compounds (Bodanszky and Bodanszky, 1970). Biotinyl peptides from
transcarboxylase
were
adsorbed to an avidin-Sepharose affinity column
and could
be
eluted only under the above-mentioned drastic conditions.
In this case the yields were quite low due to the extraordinary interaction
between immobilized avidin and the biotin-containing peptides. In a
more recent study (Rylatt et al., 1977) the biotin-containing tryptic
peptides of pyruvate carboxylases from liver mitocondria in various
mammalian and avian species were isolated by complexing with soluble
avidin. Subsequent separation of the protein-peptide complex from free
peptides was achieved by precipitation of the former with zinc chloride.
The avidin-biotinyl-peptide complex was irreversibly dissociated by
70%
formic acid.
A more extensive investigation in which the avidin-biotin complex was
used, involved the separation of biotin-containing subunits from the
apoenzyme of biotin-requiring enzymes. Two different approaches have
been employed. One study used an avidin column for the adsorption of
transcarboxylase (Berger and Wood, 1974). The enzyme was dissociated
into subunits at pH 9.0, and only the subunits containing the biotin
remained bound to the column. In this manner the nonbiotinylated
subunits were isolated in a nonactive state. Upon addition of purified
biotin-containing subunits, the enzyme was reconstituted in a highly
active form.
A
similar approach was used to isolate apo(acety1-CoA-
carboxylase) completely free of the holoenzyme (Landman and
Dakshinamurti, 1973,1975). In another approach, biotin was coupled to
Sepharose followed by an excess
of
avidin.
To
this immobilized complex,
pyruvate-carboxylase was added. Again, in this way the biotin-containing
subunit could
be
separated from the remaining subunits (Lane et al.,
1970).
2.
“Tailor-Made”
Biotin-Containing
Systems
The avidin-biotin complex can be used
as
a tool for purification, even
for systems in which biotin is not a native component (see Figure
2).
Such
systems usually employ a sandwich-type approach (Wilchek and Gorecki,

12
EDWARD
A.
BAYER
AND
MEIR
WlLCHEK
1973). The standard protocol involves the coupling of biotin
or
a hapten
to one of the interacting species, followed by controlled incubation with
the counterpart. The mixture is then applied
to
an avidin
or
antibody
column. Alternatively, the biotin-containing partner may
be
bound
initially to the immobilized avidin and a solution comprising the second
species is then applied to the column. Theoretically, after extensive
washing the interacting system can be dissociated and the underivatized
member can be isolated. On the other hand, the interacting system can
be
separated theoretically
as
a complex by introduction of huge excesses of
biotin. Because of the inefficiency of the latter procedure, considering the
strong interaction between avidin and biotin, however,
it
is
preferable to
dissociate the interacting complex by other means (e.g., as described
above for the isolation of transcarboxylase subunits).
As
an example, we have tried both approaches on the purification
of
the
receptor for insulin (Ginsberg and Wilchek, unpublished
work).
[1251]Insulin was reacted with
biotinyl-N-hydroxysuccinimide
ester and
mixed with a membrane extract from turkey erythrocytes. The complex
was not adsorbed to the avidin affinity column. On the other hand, when
the biotinylated insulin was bound
to
the avidin column and subsequently
the extract was applied, the receptor was adsorbed to the column. It seems
that prior formation of the biotin-derivatized insulin-receptor complex
shields the biotin moiety from further interaction with the avidin column.
Using the second approach, however, afforded no advantage over direct
coupling
of
insulin to Sepharose because similar dissociation conditions
(3M
guanidine-hydrochloride)
were
required in both cases to remove the
receptor from the column.
Our lack of luck with the insulin receptor, however, seems not to
be
a
general phenomenon among other interacting systems.
For
example, a
method for gene enrichment from
the
total
DNA
of
an organism
has
recently been described (Pellegrini
et
al.,
1977;
Manning
et
al.,
1977).
Purified
RNA,
from the corresponding gene, was covalently attached
to
biotin
by
means
of
a cytochrome
t
bridge. The modified
RNA
was
hybridized
to
the total
DNA
preparation. Only the
DNA,
which
recognized
RNA,
underwent hybridization; thus the specific population
could
be
separated from the other
DNA
through the action
of
the avidin-
biotin interactions.
This
was accomplished either by affinity chromato-
graphy on an avidin column
or
by
gradient separation on avidin-
containing microspheres. The gene was obtained in high yields,
4240%
pure.
The
avidin-biotin complex
was
also
used for the retrieval of thymocytes
artificially labeled with
BNHS
Uasiewicz
et
al.,
1976).
In this study the
cells
were
adsorbed to avidin immobilized on nylon mesh.
No
attempt was

THE
USE
OF
THE
AVIDIN-BIOTIN
COMPLEX
13
made to release the cells from the solid support nor
to
reconstitute native
cells. This study would have been more valid if the biotin had been bound
to the cells in a reversible manner, either through the use of reagents that
can
be
split chemically,
or
by enzymatic cleavage using biotinidase, which
removes biotin from proteins in natural systems (Koivusalo et al. 1963).
We have used this technique to separate successive generations
of
proliferating yeast cells (Bayer, Niedermeyer, and Skutelsky, unpubl-
ished work). Yeasts
were
harvested in mid-log phase and subjected to
biotinylation. The cells-still viable-were allowed to double in number,
and after one generation two different populations
of
cells were observed
in the electron microscope: one exhibiting prominent surface label
following ferritin-avidin treatment, and a second that was completely void
of surface label after the same treatment. In some cases unlabeled buds on
labeled cells were seen. The biotin-derivatized parent generation thus
could
be
separated from the daughter cells by sequestration
of
the former
on an appropriate solid support containing either avidin
or
antibiotin
antibodies (Berger, 1975). In the latter case the cells could
be
released
after immobilization by concentrated solutions of biotin. We have also
found that preferential agglutination on
10%
fetal calf serum, bovine
serum albumin,
or
sucrose provides a suitable method for separating
larger numbers of cells.
Other biotinylated proteins have also been prepared
for
receptor
perturbation and subsequent purification studies. Thus adrenocortico-
tropic hormone (ACTH) (Hofmann and
Kiso,
1976)
and
insulin
(Hofmann et al., 1977; May et
al.,
1978) have been modified with biotin
for the purpose of isolating the respective receptors. However no such
isolation
has
yet been described.
Biotinylated-lectins and antibodies have also been retrieved on avidin
columns for their separation from the underivatized protein (Bayer et al.,
1976b; Skutelsky and Bayer, 1976b). This procedure, however,
was
instituted for analytical purposes rather than
for
purification. The use of
this class
of
biotinylated binding proteins
is
described below Section VI,
on affinity cytochemistry.
The avidin-biotin complex has also been used for purification studies in
a relatively new approach termed “affinity partitioning” (Flanagan and
Barondes, 1975). This technique involves the action of an aqueous,
polymer two-phase system. The addition of a polymer-ligand that
partitions predominantly into one phase has been shown to cause a shift
of the ligand-receptor into the same phase. Thus biotinylated-binding
proteins can
be
complexed with poly(ethylene-oxide) avidin, and
this
complex is capable of partitioning the respective receptor from the
dextran phase.

14 EDWARD A. BAYER
AND
MEIR
WILCHEK
The avidin-biotin complex was used to isolate an affinity-labeled
oligonucleotide fragment from
E.
coli
23-S
ribosomal
RNA.
The avidin
was used to “fish out” biotinylated affinity-labeled nucleotides. The
basicity
of
avidin then permitted the isolation of the complex on a
phosphocellulose column (Eckermann and Symons,
1978).
3.
Experimental
Procedures
A.
PREPARATION
OF
AFFINITY
COLUMNS
a.
Adviation
of
Sepharose
(Axen
et
al.,
1967).
Sepharose 4B (10 g)
was washed
well
with distilled water, resuspended in 10 ml of distilled
water, and the suspension was stirred constantly with a magnetic stirrer.
Cyanogen bromide
(1.25
g)
was added and the pH maintained between
10.0 and 1
1
.O
by
the dropwise addition
of
2N
sodium hydroxide. After
10
min the activated
gel
was filtered and washed extensively with cold 0.1M
sodium bicarbonate. The activation procedure can
be
performed in the
presence
of
water-miscible organic solvents. The typical procedure in our
laboratory includes dissolving cyanogen bromide
(100
g) in dimethyl-
formamide
(50
ml), and the activation is performed in
2M
Na,Co, for
2
min. Dirnethylformamide is more advantageous than acetonitrile,
because a yellow product is often obtained with the latter solvent.
Coupling
of
ligands to activated Sepharose is usually performed in
0.1M
bicarbonate solutions
for
16
hrs at
4°C.
b.
Biotin
Affinity
Columns
SioCytin (Buye~
and
Wilchek
1974).
Biocytin is synthesized as follows:
BNHS (340 mg, 1 mmole) was suspended in
3
ml of dimethylformamide,
hot a-tBoc-lysine (400 mg) dissolved in 4 ml of sodium bicarbonate was
added, and the suspension was adjusted to pH
8.5.
The reaction was
carried out for
4
hr. The solvent was evaporated under pressure and
tBoc-biocytin was precipitated by the addition of
5
ml of 10% citric acid.
The crystals were filtered and washed with water.
A
second crop of
crystals can be isolated from the mother liquor.
The tBoc-group was removed
by
treatment with
4N
hydrochloric acid
in
dioxane (5ml). After
20
min ether was added. The crystals formed are
filtered,
washed
well
with ether, and dried
in
vacua
BiOcytin
Sepharose.
Biocytin-hydrochloride was coupled to activated
Sepharose
as
described above. About
25
mg of biocytin was used to
prepare
20
ml of substituted Sepharose.
PolymeriG
Biotin-Containing
Columns.
The synthesis of poly-L-lysine
or
polyacrylic hydrazide and its subsequent coupling to Sepharose was

THE
USE
OF
THE
AVIDIN-BIOTIN
COMPLEX 15
described earlier in this series (Wilchek and Hexter,
1976).
Poly-L-lysyl-
Sepharose
(3
g,
containing about
60
pmole of lysine) was suspended in
absolute dioxane
(3
ml). An excess of biotin
(65
mg) was added followed
by an equimolar amount of
DCC (50
mg). The reaction was allowed to
proceed overnight. The biotin-conjugated poly-lysyl-Sepharose was
washed several times with large volumes of dioxane, methanol, and
distilled water, respectively, and stored in about
5
ml of distilled water.
Alternatively, the column above can
be
prepared using BNHS at pH
8.5.
Other biotin-containing columns were prepared by coupling biotin to
diamine-
or
dihydrazide-coated Sepharose.
c.
Avidin
Affinity
Columns
Cyamgen
Bromide
induced Immobiliration
(Bodanszky
and
Bodanszky,
1970).
Avidin
(20
mg in
20
ml of
0.1M
sodium bicarbonate) was added
to
10
g
of
activated Sepharose (above), and the suspension was stirred
overnight at
4°C.
The gel was washed (the supernatant and washings
checked by absorption at
282
nm to determine the amount of protein
coupled) and stored suspended in water with the addition of a few crystals
of
sodium azide.
Avidin
(20
mg) was dissolved in
20
ml
of
0.1M
sodium acetate buffer, pH
4.5,
and sodium m-periodate
(22
mg)
was added. The reaction was carried out for
3
hr at
4"C,
after which
ethylene glycol
(2
ml) was added. The solution was dialyzed overnight at
4°C.
The contents
of
the dialysis bag were then allowed to interact for
3
hr
at room temperature with either adipic-hydrazido Sepharose
or
poly
acryl-hydrazido Sepharose. The extent of immobilization
(>95%)
was
checked by measuring the absorbance
(282
nm) of supernatant fractions.
Using this method, the affinity constant for biotin was somewhat reduced.
In addition to the above-mentioned methods,
avidin has been coupled to other columns, notably glass beads via the
N-hydroxysuccinimide active ester (Manning et al.,
1977).
Avidin has also
been immobilized on nylon mesh by partial hydrolysis of the nylon and
subsequent coupling using a water soluble carbodiimide (Jasiewicz et al.,
1976).
Polymethacrylate spheres have
also
been used for the same
purpose (Manning et al.,
1975).
We have also coupled avidin to fixed,
intact, erythrocyte membranes as a mode of immobilization.
Periodate-induced immobilization.
Other
Coupling
Methods.
B.
ISOLATION
PROCEDURES
The lack of chromophore in biotin renders the isolation of biotin-
containing peptides difficult to follow. Two approaches were taken by
Bodanszky and Bodanszky
(1970)
to overcome this obstacle. One

16
EDWARD
A.
BAYER
AND
MEIR
WILCHEK
approach employed radioactive biotin-containing peptides. In another
more elegant method
4-hydroxyazobenzene-2’-carboxylic
acid was
applied
to the avidin column, giving a pink color. The color was displaced
upon addition of biotin-containing peptides. In a typical experiment
using this approach, a l-ml column of avidin
(2
mg/ml) was able to bind
0.06
pmole of biotin. The same amount of biotin-containing peptides was
adsorbed.
However elution was extremely difficult; using 6M guanidine
hydrocholoride, pH 1.5, only
25%
of the biotin-containing peptide was
eluted from the avidin column.
A
different approach to the isolation
of
biotin-containing peptides was
taken by Rylatt
et
al. (1977). Soluble avidin was used to bind the biotin-
containing peptides in solution. It was found that zinc sulfate is capable of
precipitating
the
avidin-biotin complex without concomitant dissociation.
After washing out non-biotin-containing peptides, the avidin-biotin
complex
was
dissociated
by
incubating
with
70% formic acid for 1 hr at
room temperature. Under these conditions avidin was irreversibly
denatured
and
could no longer interact with biotin.
VI. LOCALIZATION STUDIES:
AFFINITY
CYTOCHEMISTRY
1.
General Considerations
One of the most cogent areas of recent interest concerns the specific
localization
of functional groups, biologically active components and
receptors on cell surfaces. The use of affinity methods
for
the localization,
visualization, and subsequent evaluation of specific cellular components
by light, fluorescent,
or
electron microscopy has been termed “affinity
cytochemistry” (Bayer et al., 1976b). In general, the technique is based on
the
preparation of a mixed conjugate, comprising a biologically active
molecule (e.g., antibodies, lectins, hormones) attached chemically to a
potentially demonstrable probe (e.g., fluorescein, ferritin, peroxidases,
hemocyanin), whereby the resultant product retains both detectibility and
biological activity.
For
use in light, fluorescent, and electron microscopical
studies, these probes have been coupled to a wide spectrum of biologically
active molecules, including antibodies (Singer and Schick, 1961
;
Avrameas, 1969; Raff, 1976), lectins (Nicholson and Singer, 1971; Ash
and Singer, 1976), hormones (Jarrett and Smith, 1975). lipoproteins
(Anderson
et
al., 1976), vitamins (Bayer et al., 1978b), sugars (Monsigny
et
al.,
1976), cations (Danon
et
al., 1972), and anions (Bayer and
Skutelsky, unpublished work). Because the defined electron-opacity of
the ferritin iron core affords
superior
resolution qualities, ferritin is the
electron microscopic marker of choice.

THE
USE
OF
THE
AVIDIN-BIOTIN
COMPLEX
17
The preparation of ferritin conjugates represents the major problem
inherent in the above-described method. Procedures currently available
for covalent coupling of ferritin to the biologically active counterpart are,
for the most part, cumbersome and inefficient. The resultant complex
is
of high molecular weight-ften a multimer-thus affecting both the
physical and chemical binding characteristics as well as the biological
activity of the conjugate.
It has been shown, however, that the use of the high affinity avidin-
biotin complex can circumvent some
of
the problems relating to ferritin-
protein conjugation (Bayer et al., 1976a). In addition, this method may be
employed to unify and facilitate certain aspects of affinity cytochemical
techniques (Heitzmann and Richards, 1974; Bayer et al., 1976b, 1978~;
Skutelsky and Bayer, 1976a).
The following steps are involved in this approach.
(a)
Biotin
is
attached
via an appropriate reactive derivative, either directly to cell surface
functional groups (sugars, amino acids, etc.) or to a biologically active
molecule (antibody, lectin, etc).
(b)
In the latter
case,
the biotinylated
conjugate is incubated with an appropriate target (intact cells or tissue,
enzymatically
or
chemically treated cells, membrane preparations, sub-
cellular fractions, defined macro molecules).
(c)
Subsequent incubation
with ferritin-avidin conjugates permits ultrastructural visualization of the
given cell surface receptor.
(d)
Proper controls using nonbiotinylated
preparations and/or unconjugated ferritin should always be
implemented.
An alternative method, which constitutes a permutation of the fore-
going method, has
also
been attempted in our laboratory. Biotinylated
membrane sites are saturated with free avidin. Since the latter is a
tetramer, subsequent interaction with biotin-conjugated ferritin also
results in specific labeling of cell surfaces. This method is somewhat less
tedious than that involving fenitin-avidin conjugates because prepara-
tion and analysis of protein-protein conjugates are precluded. However
the resultant cell surface label is much less uniform, and consequently the
method is less reliable. Figure
5
summarizes
both
approaches.
2.
Localization
of
Functional
Groups
A.
SUGARS
The avidin-biotin complex
was
first
used
as
an affinity cytochemical probe
by
Heitzmann and Richards (1974) for the localization of biotin-tagged
sites
on
membrane preparations of
Acho&losnur
laidlawii
and on
erythrocyte
ghosts.
In our laboratory, in collaboration with Dr. Ehud Skutelsky, ferritin-

ia
EDWARD
A.
BAYER
AND
MEIR
WILCHEK
B
I
..............
=...
U
-x
.....
........
....................................
L
-
...........................................
:::
MEkdRhNE
!:!:!:!:!::;,
...............
APPROACH
i
B
A
...........
I--
-
\
"tf3
BIOTIN
-
LABELED
MEMBRANE
SITE
-
FERRlTlN -AVIDIN
CONJUGATE
key
i
O
B
......................
.:.:.;.:.:
M
EM@R,A.fj'r
::!:!:!:!:
...................................
.......
APPROACH
2
FERRITIN-BIOTIN
CONJUGATE
B
Figure
5.
chemistry
using
the
avidin-biotin system.
Schematic representation
of
two
possible approaches
for
ultrastructural cyto
avidin conjugates (FAv)
were
used
in a variety of studies on intact
erythrocytes and lymphocytes. Thus erythrocytes from various
mammalian species (e.g., mouse, rat, rabbit, and human) were treated
with periodate under conditions causing the cis-hydroxyl groups
of
sialic
acid (SA) to
be
selectively oxidized
to
the corresponding aldehydes
(Skutelsky et
al.,
1977).
The latter were reacted with biotin hydrazide,
resulting in biotinylated erythrocytes. The reaction is schematically
shown in Figure
6.
Treatment of these cells with FAv revealed the SA in
the cell surface (Figure
7).
This method
was
shown to
be
superior to such
other
electron microscopic methods previously used for the localization
of
SA
as
cationized femtin (Danon et
al.,
1972)
and colloidal iron (Gasic et
al.,
1968), since
only
surface SA and not anionic groups
were
detected.
As
Figure
7
indicates, the femtin particles
were
somewhat removed
from
the
membrane surface. Therefore
we
were
able
to
calculate the average
distance of
the
SA
from the erythrocyte surface. Values of
50
to
70
W
were

THE
USE
OF
THE
AVIDIN-BIOTIN
COMPLEX
19
OH
CH, CHN
II
0
HNUNH
P
(
J-(cH,)~C-NH-N=CH
I,
AC-HN
Figure
6.
residues.
Schematic description
of
selective biotinylation
of
membrane-bound sialic acid
obtained for rat and human, whereas in rabbits the SA was juxtaposed to
the lipid bilayer (Bayer et al.,
1977).
We
were
also able to follow the fate of
SA
during the development of the erythrocyte from its precursors
(Skutelsky et al.,
1977).
Very recent studies have shown that the same procedures can
be
used
to
study alterations in the surface in various examples of diseased blood
cells. Thus striking alterations in the topography
of
SA were observed on
erythrocytes derived from thalassemic patients (Kahane et al.,
1978).
These observations were in accordance with the decreased life span
of
these cells in
the
circulation. In another study with lymphatic leukemia
cells from both human and bovine sources, we have observed dramatic
increases in the surface labeling
of
SA, compared with that of normal

20
EDWARD
A.
BAYER
AND
MEIR
WILCHEK
Figure
7.
Labeling of membrane-bound sialic acid sites on
the
human erythrocyte
by
periodate-induced biotinylation followed by treatment with femtin-avidin conjugates. Note
the measurable distance separating the ferritin particles from the plane of the membrane.
lymphocytes (Skutelsky et al.,
1978).
In the bovine leukemic cell, a dense,
multilayered deposition of FAv was obtained, extending to more than
500
A
from the lipid bilayer. Likewise, the density
of
SA on lymphoid cells,
derived from chronic lymphatic leukemia patients, was nearly twice that
of
lymphocytes derived from healthy donors.
The same biotinyl derivative can
also
be
used for the localization
of
galactose residues on the cell surface by prior treatment with galactose
oxidase (Heitzmann and Richards,
1974).
In a similar manner this
procedure, combined with enzymatic techniques,
can
be
used
as
a tool
for
the localization of sialyl-blocked, membrane-bound, galactose-containing
sites (Figure
8).
Treatment
of
aldehydes generated by galactose oxidase
with
an
unrelated hydrazide, followed by reduction with borohydride,
blocks free galactosyl groups. Subsequent enzymatic digestion of
SA
residues with neuraminidase exposes penultimate galactose
(or
N-
acetylgalactosamine), which may now
be
subjected to successive treatment
with galactose
oxidase
and biotin hydrazide.
B.
AMINO
ACID
RESIDUES
The avidin-biotin complex is not only useful for the localization
of
cell
surface sugars, it can also
be
used for the study
of
amino acid functional
groups of cell surface proteins
or
polypeptides. Thus the N-hydroxy-

RCONHN=CH
H
Galanos? oxidas?: Newaminidale
~
RCONHNH,
H
OH
OH
RCONHN=CH
HO
F0\
Galactose oxidare
biotin hydrazide
HO
OH
Y
Y
OH
(a)
(bJ
Figure
8.
Schematic illlustration representing differential biotinylation
of
membrane-bound
galactosyl
(or
N-acetylgalaaosamine) residues in the free
(a)
or
sialyl-blocked
(b)
form.
The
hydrazide
derivatives
of
R
is structurally unrelated
to
biotin and does not form a complex with avidin.

22
EDWARD
A.
BAYER
AND
MEIR
WILCHEK
sucanimide ester of biotin reacts fairly specifically with lysine residues.
Use of this reagent at pH
5.0
limits
its interaction
to
a a-amino groups.
The diazonium salt of biotin
can
be used in the study of tyrosine and
histidine residues, and
is
reversible upon reduction with dithionite.
The bromoacetyl derivative
can
be used for cysteines,
or,
following
reduction with
thiols,
it is also suitable for localization of cell surface
cystine bridges. Depending on pH, the bromoacetyl reagent may also be
used to localize methionines, histidines. and lysines. Thus a variety of
amino
acid groupspecific reagents are available, and comparative studies
of the localization and analytical isolation of labeled surface proteins from
various cell types and species are currently in progress in our laboratory.
3.
Localization
of
Receptors
Of greater significance is the use of the avidin-biotin complex for the
localization and evaluation of receptors on cell surfaces. Cell surfaces are
known to
possess
a variety of receptors for hormones, antibodies, lectins,
drugs, toxins, effectors, and a variety of other biologically active com-
pounds. The amount of these receptors is usually very low, and very
highly radioactive labeled compounds must
be
used
to
quantify the
number of receptors per cell.
The use of electron microscopic techniques obviates some of the
difficulties by permitting specific labeling with non-penetrating markers
for quantitative analysis. The direct use of ferritin-conjugated binding
proteins introduces a factor of uncertainty, since the binding proteins are
usually of much lower molecular weight than the marker. The conjugated
binding protein may exhibit reduced activity and/or altered specificity,
thus restricting the reliability of the method. Coupling a small molecule,
such
as
biotin, to the binding protein, under defined conditions, will only
slightly affect the binding characteristics. Following interaction
of
the
biotinylated binding protein with the cell surface enables more specific
labeling of the receptor. The cell may be fixed either before
or
after the
labeling procedure; consequently, these conditions are amenable to
kinetics studies. Following fwation, the biotin-tagged sites are available
for further interaction with FAv. Figure
9
presents the procedure
for
specific labeling
of
receptor sites.
We have exploited this approach (Bayer et
al.,
1976b) and have coupled
biotin to various lectins, including concanavalin
A
(Con
A)
peanut
agglutinin (PNA), soybean agglutinin, phytohemagglutinin, and wheat
germ agglutinin. The biotinylated Con
A
(B-Con
A)
had properties
similar to those of the native protein-namely, it bound to Sephadex and
agglutinated erythrocytes. When we applied the B-Con
A
to
erythrocytes

THE
USE
OF
THE
AVIDIN-BIOTIN
COMPLEX
23
8
A8
Biotinylated Membrane
--
lectin or
+
receptor
antibooy
Figure
9.
Schematic procedure
for
B
d
B-lectin- Ferritin- Specifically-
labeled
+
avidin labeled
membrane- conjugate membrane-
receptor receptor
specific labeling
of
receptor sites by a biotinylated
-
-
binding protein (antibody, lectin, hormone, effector, etc.), followed by interaction with an
appropriate avidin-conjugated marker (ferritin, hernocyanin, peroxidase, fluorescein, etc.).
followed by FAv, the cell surface was heavily labeled with ferritin.
a-
Methyl-mannoside was effective in preventing interaction between the
surface receptor and B-Con
A
but did not interfere with that between FAv
and B-Con
A.
In a cognate study, when B-PNA, which has no exposed
receptors on human erythrocytes, was incubated
with
the latter cells, no
ferritin label was detected. When sialyl residues were removed with
neuraminidase, however, the newly exposed B-PNA-receptor complex
could
be
localized by the treatment above.
Biotinylated antibodies, elicited against erythrocytes, could be used to
detect membrane-bound antigens in a similar manner. It is interesting to
note that biotinylated whole antiserum can be used, since only the active
cell-binding antibodies complex with the membrane, while other
nonrelated biotinylated proteins are removed during subsequent cen-
trifugation steps.
Because the number of receptor sites to Con A, PNA, and antibodies is
relatively high,
we
are now trying to apply this approach to the localization
of biologically important receptors present in very limited quantities on
the cell surface, that is, hormone receptors (Riesel et
al.,
1977).
Thus
biotinylated insulin and human chorionic gonadotropin (hCG)) are
presently being used
for
the localization
of
these receptors on lympho-
cytes and ovary cells, respectively.
4.
Othersystems
The
planar distribution of surface membrane proteins in
Acholeplasma
hdhwi
has been investigated using the avidin-biotin complex (Wallace et
al.,
1976).
The membranes were labeled with biotin via the N-hydroxy-
succinimide ester. Under the conditions reported, only membrane
protein constituents were derivatized. Biotinylated-membrane prepara-

24
EDWARD
A.
BAYER
AND
MEIR
WILCHEK
tions
were
then treated with ferritin-avidin conjugates in a temperature-
controlled chamber and fixed by drying in an atmosphere of
dry
nitrogen
gas.
It was found that membrane proteins were relatively dispersed either
in the paracrystalline or smectic phases (below or above the phase
transition, respectively). On the other hand, patches were observed at
temperatures in the midphase transition range. These results indicate
that
the
physical state of membrane lipids can influence the relative
location of protein constituents within the plane of the membrane
surface.
We have used the avidin-biotin complex for the direct visualization
of
the interaction between liposomes and the cell surface (Bayer et al.,
1978a). Biotin was covalently attached to the head groups of appropriate
lipids, again via
biotinyl-N-hydroxysuccinimide
ester. Liposomes, con-
sisting of
5%
biotinylated lipids, were interacted with various types of
cells. Following furation at an appropriate time interval, the distribution
of
biotinylated lipids
on
the cell surface was evaluated using ferritin-
avidin conjugates. It was shown that the extent and mode of liposome-cell
interaction was dependent upon both the lipid content of the liposome
and the cell
type.
Ferritin-avidin conjugates were
also
used to determine the relative
position of the
4s
and ribosomal
RNA
genes in HeLa cell mitochondria1
DNA (Angerer
et
al., 1976). These authors found that at least 19
4s
RNA
genes are present in the HeLa mitochondrial-DNA genome.
Macromolecules have also been a subject
for
electron microscopic
investigations using the avidin-biotin complex. In a pioneering work,
Green and co-workers (Green et al., 197
1)
used bivalent biotin-containing
compounds to study the orientation
of
avidin subunits.
A
modification of
this approach (Green et
al.,
1972) permitted determination of the position
of the biotin carboxyl carrier protein
of
native transcarboxylase. In an
application of this technique to macromolecules that are not native biotin-
associated systems,
we
have used the avidin-biotin complex
to
localize the
positions of oligosaccharide residues on collagen (unpublished work).
The use of the avidin-biotin complex in ultrastructural or related
labeling studies is not limited to FAv, and other markers can
be
used.
Although ferritin exhibits superior resolution qualities for the analysis of
labeled material in thin sections, freeze-etching replicas, shadow casting,
or negatively stained sections by transmission electron microscopy; other
markers--namely peroxidases, hemocyanin, and phages-are
appropriate for evaluation by the latter techniques. Horseradish
peroxidase may
be
used for labeling studies in both ultrastructural and
light microscopic analysis. Hemocyanin or phages, conjugated to avidin,
may eventually prove to be excellent markers for the detection of minute

THE
USE
OF THE
AVIDIN-BIOTIN
COMPLEX
25
amounts of receptors or as defined markers for adaptation of the
technique to scanning electron microscopy.
A
fluorescent form
of
avidin has already been applied to fluorescent
microscopy. Heggeness and Ash
(1977)
used biotinylated heavy-
meromyosin and biotinylated antiactin antibodies to visualize the
distribution of nonmuscle contractile proteins in fibroblast cells.
Extension
of
this technique to ultrastructural analysis may enable double-
labeling studies for determination of the interrelationship between
exocellular receptors and intracellular contractile components. The
combination of two ultrastructural labels (e.g., peroxidase
as
the
intracellular marker and femtin conjugates as the extracellular label)
may pave the way for such studies.
The use of the avidin-biotin complex
as
a general probe in affinity
cytochemistry
is
appealing for a variety of reasons:
1.
Only one conjugate (e.g., ferritin-avidin, fluorescent-avidin) need
be
prepared and characterized for
all
affinity systems.
2.
Biotin can be attached to most small ligands and macromolecules
efficiently and under very mild conditions.
3.
In most cases, the size, the physical characteristics, and the
biological activity of biotin-derivatized proteins are only nominally
affected.
4.
Crude preparations of binding proteins (e.g., wble antiserum
instead of antiwies or plant extracts instead
of
lectins)
can
be bio-
tinylated, and, following dialysis, may
be
used for localization studies
without further purification.
5.
The biotin-avidin complex is of exceptionally high affinity and
stability.
6.
The use of this system allows kinetics studies, since
furation
and
subsequent localization via the conjugated marker
can
be performed at
any stage during the probe (binding protein) receptor interaction.
7.
The avidin-biotin complex, in conjunction with standard affinity
cytochemical systems (direct conjugation of binding protein to marker),
may be used
for
double-labeling studies.
8.
Both avidin and biotin are commercially available in large
quantities.
5.
Restrictions
of
the Method
Although the avidin-biotin complex affords a highly versatile method
for
specific ultrastructural labeling studies, the potential user should be
aware of several limiting
or
interfering factors. Initially, when applying
the biotin-avidin interaction to a given experimental system, it must be

26
EDWARD
A. BAYER
AND
MEIR
WILCHEK
determined whether the latter comprises a biotin-containing, biotin-
recognizing,
or
biotin-free system. Avidin also possesses a biological role,
albeit
as
yet undefined, and is produced in the oviducts of various avian
and reptilian species as well
as
by selected strains of bacteria (Green,
1975). Avidin has been found
to
bind selectively to condensed chromatin
(Heggeness, 19’77) in
a
manner apparently unrelated
to
its biotin-binding
properties. Our own observations confirm this phenomenon, since FAv
binds “unspecifically”
to
subcellular fractions that are prepared from
osmotically shocked intact cells and undoubtedly contain large amounts
of adsorbed nucleic acids. It should
be
noted that avidin is a basic glyco-
protein, and either ionexchange properties or its oligosaccharide moiety
might
be
responsible for a variety of side interactions. Consequently,
it
is
emphasized that proper controls should be implemented in all applica-
tions of the avidin-biotin complex to affinity cytochemical studies.
6.
Experimental
Procedures
A.
PREPARATION
OF
AVIDIN-CONJUGATED MARKERS
a.
Femtin-Avidin
Conjugates
Via
Reductive
Alkylation.
(Bayer et
al.,
1976a). Commercial avidin (15 mg) in
5
ml of acetate-buffered
saline,
(ABS)
pH
4.5,
was added to femtin
(100
mg, 1 ml). Sodium
m-periodate (Merck,
0.66
ml, 0.1M solution) was added to a final concen-
tration of 10mM. The mixture was stirred for
3
hr in ice, dialyzed for
6
hr
against
ABS
at 4”C,
and
followed by a second dialysis overnight at 4°C
against borate-buffered saline, pH
8.5.
A
fresh solution of sodium boro-
hydride (10 mg/ml in
0.0
1M sodium hydroxide) was prepared, and
0.5
ml
was added
to
the femtin-avidin conjugates in an ice bath. After 1 hr the
solution was dialyzed against phosphate-buffered saline (PBS), pH 7.0.
The conjugates
were
washed twice by centrifugation
(100,000
x
g,
3
hr)
and resuspension in
PBS,
and finally resuspended to
1
mg of ferritin per
milliter.
This method
of
unidirectional conjugation has been shown to provide
increased yields of active, unit-paired conjugates. Reductive alkylation
is
therefore recommended over the conventional glutaraldehyde
techniques
for
the preparation of ferritin avidin conjugates.
b.
Glutaraldehyde Method.
(Heitzman and Richards, 1974).
Commercial avidin (15 mg in
3
ml of PBS) was added
to
a
solution
of
ferritin
(1
ml, 100 mg) and stirred at room temperature. Glutaraldehyde
(440
pl,
0.5%
solution) was added slowly
to
a final concentration
of
0.05%.
The reaction was allowed
to
proceed for
1
hr at room temperature and

THE
USE
OF
THE
AVIDIN-BIOTIN
COMPLEX 27
then stopped with 0.1M ammonium bicarbonate. The conjugates were
dialyzed overnight against
PBS.
Large aggregates were removed by
centrifugation at 10,000
X
g
for
30
min. The supernatant was sub-
sequently centrifuged at 100,000
X
g
for
3
hr. The pellet, consisting of
free ferritin and ferritin-avidin conjugates, was resuspended in
PBS,
and
the
100,000
X
g
centrifugation step was repeated. (The supernatant
fractions containing free avidin were saved for future preparations
of
ferritin-avidin conjugates.) The washed conjugates were resuspended in
PBS
to a final concentration of
I
mg of fenitin per milliliter
(AMo
=
I.
1).
Alternatively, to achieve unidirectionality, ferritin can be treated with
an excess of glutaraldehyde
(1
5%)
and subsequently interacted with
avidin following gel filtration on Sephadex
G-25
to remove free glutar-
aldehyde (Otto et al., 1973).
c.
Other Coupling Methods.
Another unidirectional approach for
the preparation
of
ferritin-avidin conjugates was described by Angerer et
al. (1976). In this procedure, ferritin was bromoacetylated and reacted
with thiolated avidin.
d.
Other
Markers.
Other markers (e.g., peroxidases, hemocyanin)
can
be
substituted for ferritin, and the respective avidin conjugates can
be
prepared as above, using equimolar amounts
of
the prospective marker.
In
the
case
of
large markers (hemocyanin, phages, etc.), free avidin can
be
removed from the reaction mixutre by centrifugation as above. With
smaller markers (peroxidases, etc.) the conjugates can
be
separated from
the reactants by gel chromatography.
Fluorescein-derivatized avidin was prepared as follows.
To
a solution of
avidin (14 mg) in 0.01M phosphate buffer, pH 7.4
(1
ml), was added
fluorescein-isothiocyanate
(300
ug in
0.1
ml
of
0.5M sodium carbonate
buffer, pH 9.5). The solution was
stirred
overnight in the cold. The
conjugate was separated from the free ligand by passage of the reaction
mixture over Sephadex G-25. The conjugate was visibly present in the
void volume. About 6 fluorescein groups were attached
per
avidin
molecule by this procedure.
A
somewhat different procedure was
described by Heggeness and Ash (1977).
e.
Analysis
of
Conjugates.
The relative size of ferritin-avidin conju-
gates can
be
tested by gel filtration on a Sepharose
6B
column. The extent
of ferritin conjugation was assayed by affinity chromatography on a
biotincontaining affinity column. The difference in absorbance
(AMo
applied effluent) represents the amount of active ferritin-avidin con-
jugate.

28
B.
EDWARD
A.
BAYER
AND
MEIR
WILCHEK
DIRECT BIOTINYLATION
OF
FUNCTIONAL
GROUPS
a.
Sialic Acid Residues
(Skutelsky
et
al.,
1977).
Cells
(lo*
ml) were
washed and resuspended in PBS
(1
ml). Sodium m-periodate was added to
a final concentration
of
Id,
and the reaction was allowed to proceed for
30
min in an ice bath. The cells were then washed twice with PBS and
resuspended in a solution of biotin hydrazide
(2.5
mg/ml). After
1
hr at
room temperature, the cells
were
washed three times in PBS and fixed in
2%
glutaraldehyde
(1
ml in PBS).
Galactose
and
N-Acetylgalactosamine Residues (Heitzmann
and
Richards,
1974).
The primary hydroxyl groups of these sugars were
treated with galactose oxidase, and the resultant aldehydes were inter-
acted with biotin hydrazide. Cells (approximately
lo*)
were washed twice
with
PBS
and mixed with sodium borohydride
(2mM
in PBS) to quench
the effect
of
endogenous, oxidized membrane components. The cells
were
washed twice in buffer, and biotin hydrazide
(2.5
mg/ml PBS) was
added. The suspension was treated with galactose oxidase
(10
units,
Sigma Chemical)
for
3
hr at
37"C,
washed twice with buffer, and fixed in
glutaraldehyde.
b.
C. BIOTINYLATION
OF
BIOLOGICALLY ACTIVE PROTEINS
For
relatively stable, biologically active proteins, such as antibodies,
lectins, and polypeptide hormones, the following biotinylation procedure
has
proved
to
be
effective.
Biotinyl-N-hydroxysuccinimide
ester (BNHS) dissolved in dimethyl
formamide
(DMF),
was added
to
a solution, pH
7.0
or
higher,
of
the
desired protein in a
1-10
to
1-100
v/v
and
5-1
mole/mole ratio.
For
example, an aliquot
(0.1
ml) containing
0.5
pmole of BNHS
(1.7
mg/ml
DMF)
was added to a solution containing goat anti-rabbit
IgG
antibodies
(16
mg of protein in
1
ml
of
PBS). The solution was kept at room
temperature for
4
hr and dialyzed overnight at 4°C against PBS, with one
buffer change. Biotinylated antibodies
or
lectins may be stored at
-20°C.
Whole antiserum
or
unpurified lectins can
be
biotinylated in the manner
just described and used for affinity cytochemical studies in their
unrefined state. The reaction with BNHS can be restricted mostly to
a-amines by performing the reaction between
pH
5.0
and
6.0.
Relatively unstable proteins
(or
those subject to loss of biological activity
upon chemical modification) may require additional
or
alternative
treatment-for example, modification of cysteines
or
tyrosines, or
separation
of
biotinylated proteins from underivatized material after
milder biotinylation procedures.

THE
USE
OF
THE
AVIDIN-BIOTIN COMPLEX
29
D.
INTERACTION BETWEEN BIOTINYLATED BINDING PROTEIN AND
RECEPTOR
Viable cells
(lo*)
or
cells fixed with
2%
glutaraldehyde for
30
min, were
washed and incubated with an appropriate solution of biotinylated lectins
or
antibodies (0.5-1.0 mg of protein
per
milliliter of
PBS
for
30
min at
rcmm temperature. Normal
or
optimal conditions of interaction should
be
used with any other protein type (hormones, toxins, effectors, etc.).
Controls comprise labeling with underivatized protein samples
or
use of
appropriate inhibitors. After incubation, cells are washed, fixed with
glutaraldehyde, and treated with
2%
bovine serum albumin.
E.
L.(XALIZATION OF BIOTINYLATED SITES
Further treatment with ferritin-avidin conjugates (FAv) (1 mg of
protein
per
milliliter) results in specific labeling of biotin-tagged sites.
Cells whose functional groups have been directly biotinylated are labeled
with FAv by similar treatment. Samples prepared in this manner are then
processed for electron microscopy (Luft, 1961). Controls, comprising
pretreatment of FAv with biotin, should
be
employed.
VII. MISCELLANEOUS
SYSTEMS
1.
Phage
Inactivation
Studies
A.
GENJ?.RAL
CONSIDERATIONS
One of the earliest studies that demonstrated the availability of biotin,
artificially complexed with a living system, for subsequent interaction
with avidin, involved phage inactivation (Becker and Wilchek, 1972). In
this work phages, to which biotin was covalently attached by means of the
N-hydroxysuccinimide ester, were found to be inactivated with avidin
(Table
11).
It was known previously that bacteriophages, modified by
chemical attachment of haptens, are still viable and can be inactivated by
specific antibodies against the covalently linked modifier. This technique
was first demonstrated by Makela (1966) and Haimovich and
Sela
(1966)
and
was used for
the
detection and quantification
of
antibodies and
haptens. An analogous study was performed by Becker and Wilchek
(1972) to show that not only antibody-antigen complexes are capable of
phage inactivation, but other interacting systems can be used, provided
one component can be bound to the phage and the counterpart is multi-
valent. The avidin-biotin system was an ideal subject for this study, once
again because
of
the high affinity constant of interaction. It
was
assumed
that the latter property would render this method the most sensitive for

30
EDWARD
A.
BAYER
AND
MEIR
WILCHEK
w-
Figure
Rotin
concn
(ng/rnl)
0.
Inhibition
of
avidin-induced inactivation of biotin-derivatizej- phages by ree
biotin. Biotin
(0.
I
ml) at the concentration given was incubated with
0.1
ml
of
an avidin
solution
(20
pg/ml avidin in
50
mM
sodium phosphate buffer, pH
6.8).
A
suspension
(0.1
mi) of biotin-modified bacteriophage
(see
text)
was
added and the mixture incubated at
37°C
for
1
hr. The surviving phage titer
was
determined, and the percentage inhibition of
avidin-induced phage inactivation was calculated by comparing plaque-forming units
per
milliliter in
the
absence and
in
the presence
of
various concentrations of biotin.
quantitative estimates of biotin. It was found that avidin was indeed
capable of inactivating the biotin-derivatized T,-bacteriophage from
forming plaques, and that free biotin in solution can prevent the observed
inactivation. An interesting observation derived from this study was that
only
2
moles of biotin
per
mole
of
avidin is required for the complete
inhibition of phage inactivation.
Figure
10
indicates that at least
2
biotin
binding sites
per
avidin molecule must
be
unoccupied before avidin can
inactivate the biotin-modified bacteriophage. The extremely narrow
range
in
which inhibition by free biotin occurs may
be
a
consequence
of
the relative homogeneity
of
avidin preparations versus the hetero-
geneous population
of
antibodies. Inhibition by free hapten
of
inactiva-
tion of chemically modified bacteriophages by antibodies usually occurs
over several orders of magnitude of free hapten concentration.
B.
EXPERIMENTAL
PROCEDURES
a.
Preparation
of
Biotin-Modified
Bacteriophage.
The bacterio-

THE
USE
OF
THE
AVIDIN-BIOTIN
COMPLEX
TABLE
I1
Effect of Avidin on
T,
Bacterio hage and
31
T,-BIO Preparation
B
Bacteriophage titer
(plaque-forming
Reaction mixture units/ml)
T,
+
buffer
4.5
x
103
T4
+
avidin
4.5
x
103
T,-BIO
+
buffer
4.4
x
103
T,-BIO
+
avidin
5.0
X
10'
aBacteriophage and buffer
(0.05
M
sodium
phosphate, pH
6.8,
containing
20
pg
of
gelatin
per milliliter)
or
avidin
(0.1
mg/ml)
dissolved in
this buffer were incubated at
37°C
for
1
hr. The
bacteriophage titer was determined
as
describ-
ed previously.
phage
T4
suspension was dialyzed for 24 hr against 0.1M borate buffer
(pH 8.5).
A
solution
(0.1
ml) of
biotinyl-N-hydroxysuccinimide
ester (20
mg/ml in dioxane) was added to the
T4
suspension
(1.0
ml). After stirring
for 10 min at
4"C,
the mixture was diluted 100-fold in
50mM
sodium
phosphate buffer (pH
6.8).
The diluted suspension was dialyzed for
24
hr
against the same buffer. The
derivatized-bacteriophage
preparation was
kept at
4°C
in the buffer above containing 20 pg/ml gelatin.
b.
Inactivation
of
Bacteriophage
by
Avidin.
Aliquots
(0.1
ml) of
avidin solutions in 50
mM
phosphate buffer (pH
6.8),
containing 20
pg/ml gelatin, were added to bacteriophage suspensions (0.1 ml) in the
same buffer and incubated at 37°C for
1
hr. At this time
2.5
ml of soft
agar, containing
3
x
lo9
bacteria
per
milliliter, was added, and the mixture
was poured onto plates of hard agar. Plaque-forming units were counted
after incubation at 37°C for
15
to
18
hr. Inhibition of phage inactivation
was accomplished by preincubation of biotin at various concentrations
with enough avidin (1
.O
pglml) to cause 95% inactivation. The biotin-
derivatized phage suspension was then added, and the mixture incubated
at
37°C
for
1
hr. Assay of bacteriophage was performed according to
Adams
(
1959).
2.
Lymphocyte Stimulation
Lymphocytes are triggered to grow and divide upon interaction with a
variety of agents including the phytomitogens (Ling and Kay, 1975),
anti-immunoglobulins (Moller, 1972), and single chemicals such as

32
EDWARD
A.
BAYER
AND
MEIR
WILCHEK
sodium periodate (Novogrodsky and Katchalski,
1971).
To
further
understand the mechanism of lymphocyte activation, it was important to
ascertain whether different mitogens trigger the cells by affecting
different membrane sites,
or
whether the triggering signal is localized at a
single unique site.
Logically, the first step in lymphocyte stimulation is the binding of the
above-mentioned agents with sites on the cell surface membrane.
To
prove this premise,
we
once again turned to the avidin-biotin complex to
probe
this system (Wynne
et
al.,
1976).
Biotin was bound chemically to a
variety of functional groups on the cell surface to determine whether
avidin would stimulate the modified cell. Two different modes of
chemical modifications
were
used
to
insert a mitogenic site onto mem-
brane constituents’:
in
one the ligand is inserted onto carbohydrate
moieties, and in the second the ligand is attached to protein components
of
the membrane. Biotin-conjugated cells, modified via lysyl, sulfhydryl,
STlMU
LATlO
N
-&
-
BIOTINYLATED-
Fob?
ANTlOlNlTROPHENOL
ANTIBODY FRAGMENTS
@
-
AVlDlN
Figure
11.
dinitrophenylated lymphocytes
with
biotinylated-Fabanti-DNP-antibod
y
fragments.
Schematic representation of avidin-induced stimulation following treatment
of

THE
USE
OF
THE
AVIDIN-BIOTIN
COMPLEX
33
and phenol groups on the membrane surface, were agglutinated but not
stimulated when cultured with avidin. On the other hand, avidin-induced
stimulation of mouse and rat lymphocytes was attained upon biotin con-
jugation
of
sialyl moieties on membrane-bound oligosaccharides,
achieved via successive treatments with periodate, biotin hydrazide, and
borohydride reduction (Table
111).
The conclusions derived from these
studies were as follows:
1.
Triggering sites can
be
grafted onto lymphocyte membranes.
2.
The chemically modified lymphocyte will
be
triggered
to
undergo
blastogenesis by proteins specific to the ligand only if the latter
is
attached
to the proper position at the particular triggering site.
3.
Sugars may comprise the specific triggering site of the lymphocyte.
4.
In a related study we came to a fourth conclusion that multivalency
is a requirement for stimulation (Ravid
et
al., 1978). The crosslinking
criterion was corroborated by the observed enhancement of stimulation
in dinitrophenylated lymphocytes by the combined action
of
biotinylated-
Fab-anti-DNP antibody fragments and avidin (Figure 1
1).
3.
Hormone-Receptor
Interactions
The binding for purification purposes of biotin to hormones such as
adrenocorticotropic hormone (ACTH), insulin, and hCG was discussed
earlier. A more
detailed
study on the effect
of
avidin on hormone action
was performed with biotinyl-insulin (Hofmann et al., 1977). Insulin was
selectively biotinylated with
BNHS
at the phenylalanine amino terminus
of the B-chain to give a biologically active modified hormone. The ability
of the latter to bind to an avidin column was demonstrated. Avidin did not
interfere with the biological activity
of
underivatized insulin but was
found
to
interfere with that of the biotinylated hormone. It
is
interesting
to note that the biotinylated insulin, attached to avidin-Sepharose,
retained
15%
of the ability to stimulate rat epididymal adipocytes.
No
leakage of the immobilized hormone could
be
detected. This study thus
confirms the early observations that insulin action is an exocellular
phenomenon (Cuatrecasas,
1969).
Another biotinyl derivative of insulin was prepared wherein Lys-29 on
the Bchain was modified (May et al., 1978). This derivative had biological
activity indistinguishable from that of native insulin. Upon complexing to
avidin in a
1
:
1
ratio, the maximal biological response was similar to that of
the
active hormone, but the potency of the complex was reduced to
5%.
In our laboratory we have studied the effect
of
biotinylation via
BNHS
on
hCG
(Riesel et al., 1977).
It
was shown that derivatization at high
ligand-to-hormone ratios impaired the ability
of
hCG to interact with
testicular luteinizing hormone (LH) receptors. On the other hand, the

TABLE111
Specific Interaction
of
Proteins with Lymphocytes Modified by Conjugation
of
Chemical
Groups
onto the Cell Surface
Chemical group Specific binding
protein Stimulation
Reagents Functional
group
on cell introduced
-
Group A
Biotin
h
ydrazide
DNP-hydrazine
Aldehyde (NaIO, generated) Biotin Avidin Positive
DNP Anti-DNP
Ig
Group
B
cw
Trinitrobenzene sulfonic acid
A
N-DNP-t-aminocaproyl-N-h
ydroxysuccinimide ester e-NH,
a-N-Bromoacetyl-c-N-DN P-lysine'
I
,3-Difluoro-2,4-dinitrobenzene
2,4-Dinitro-N,N-di(2-chloroethyl)aniline
(N-DNP-nitrogen
TNP Anti-DNP Igb Negative
DNP
mustard)
Diazotized m-nitroaniline
Diazotized parsanilic acid
Biotin N-h ydroxysuccinimide ester
Tyrosine MNP
Arsanilic acid Anti-arsanihc acid
Ig
E-NH, Biotin Avidin
"Also
reacts with -SH groups.
bTNP and MNP also interact with anti-DNP
Ig.

THE
USE
OF
THE
AVIDIN-BIOTIN
COMPLEX
35
controlled attachment
of
BNHS
to hCG, followed by separation of the
native and biotinylated hormone via an antibiotin-antibody affinity
column, yields a conjugate that exhibits a biotin-to-hormone ratio of
1
:
1.
The ability
of
the latter conjugates to activate ovarian adenylate cyclase
was investigated.
A
biphasic interaction was observed (Figure
12),
which
probably indicated the presence of more than one species
of
biotinylated
hCG, that is, derivatization of the hormone at different positions. It is
known that some of
the
lysines of
LH
are important for the hormone-
receptor interaction, but others are not (Liu et al.,
1974).
Thus at high
hormone concentrations, competition between various classes
of
the bio-
tinylated hCG (depending on the essentiality of the modified residue)
would account
for
the biphasic curve in Figure
12.
VIII.
THE
BIOTIN
TRANSPORT
SYSTEM:
AN AFFINITY
LABELING
STUDY
The affinity labeling
of
the biotin transport system in yeast cannot
legitimately
be
included as a prime example in the use of the avidin-biotin
complex in molecular biology.
The
biotin transport system constitutes a
Figure
12.
compared
to
that
of
underivatized
hCG
(circles).
Ability
of
biotinylated
hCG
(triangles) to interact
with
testicular membranes

36
EDWARD
A.
BAYER
AND
MEIR
WILCHEK
biotin-recognizing system; hence the use of avidin as a probe is circum-
stantial. On the other hand this experimental system is of particular
interest because
it
provided the medium for the debut and development
of
the above-described cell-based affinity studies (Bayer and Wilchek,
1978). Thus in addition
to
affinity labeling studies, affinity cytochemistry
and affinity purification were used to investigate this system.
To
date transport systems for this essential vitamin have been described
for
bacteria (Lichstein and Waller, 1961
;
Waller and Lichstein, 1965a,b),
yeast (Rogers and Lichstein, 1969a,b), and intestinal cells derived from
various mammalian species (Spencer and Brody, 1964; Berger et al.,
1972). The biotin transport system in the yeast
Sacchurmyces
cerevisiae
is
the best characterized
of
those cited. The uptake
of
biotin by these cells
was shown to
be
pH and temperature dependent, stimulated by glucose,
and inhibited by biotin analogs-features consistent with an active,
energy-requiring, camer-mediated phenomenon.
With this in mind, we wanted to use the biotin transport system in yeast
as a model experimental system to determine the reaction of affinity
probes on intact cells.
Our
initial approach involved affinity labeling, a
technique that was introduced to selectively inhibit proteins by the specific
modification of an amino acid residue at
or
near the active site (Jakoby
and
Wilchek, 1977). The first examples of affinity labeling studies
described the use of these compounds for the site-directed inactivation of
specific isolated proteins (e.g. enzymes, antibodies) (Singer, 1967). The
expressed
goal of the original proponents of the method, however, con-
centrated on its potential use in drug therapy. This goal has yet to be
achieved. Perhaps the approach nearest
to
the desires of the originators
of
this method is exemplified by the affinity labeling of a defined trans-
port
system, because to perform its function a drug must either affect a
given membrane constituent
or
pass through the cell membrane. The
importance of the cell surface in modem biology is reflected in the
observations that many biological functions are regulated by membrane-
bound constituents and many signals are transferred from the outside.
For
our purpose,
the
biotin transport system provided an excellent
model system for these affinity studies. It was known that the ureido ring
is required for recognition and binding of the vitamin, whereas the valeric
acid side chain is less important. It
was
therefore assumed that modifica-
tion
of
the terminal carboxyl group with reactive functional groups would
furnish an appropriate affinity label for transport components. The
affinity label should thereby serve to inactivate biotin uptake, whereas
other transport systems should remain unaffected. To this end, three
reactive biotinyl derivatives (the bromoacetic hydrazide, the N-hydroxy-
succinimide ester, and the pnitrophenyl ester) were prepared and

THE
USE
OF
THE
AVIDIN-BIOTIN
COMPLEX
37
reacted with yeast cells (Becker et al.,
1971).
It was found that the biotinyl-
bromoacetic hydrazide failed to inhibit biotin transport. The N-hydroxy-
succinimide ester partially inhibited the transport, and biotinyl-p
nitrophenyl ester (BNP) at
low
concentration
(10
@)
inactivated the
biotin transport almost completely (Figure
13).
BNP-inactivation was
found to be very specific, since nonreactive biotinyl derivatives or
unrelated active esters did not affect biotin uptake. The BNP-induced
inhibition of biotin uptake could be partially prevented by free biotin,
indicating the specificity of the affinity labeling reaction.
The evaluation of the number
of
transport components per cell was
attempted using three different approaches. The first (Bayer et al.,
1978b) involved radioactive BNP. The use of the radioactive compound
can generally be considered the most sensitive method for the examina-
tion of intracellular labeling. Unfortunately
it
is difficult to obtain a
purified radioactive derivative of high specific activity. In addition, non-
specific sidd reactions and/or interactions are to be found (and to be
Minutes
Figure
13.
Effects of
various
compounds
on
biotin
transport.

38
EDWARD
A.
BAYER
AND
MEIR
WILCHEK
anticipated) in a complex experimental system. Because
of
these diffi-
culties, a second approach was considered using biotin-modified
T,
bacteriophages (Becker et al., 1972). Specifically, the interference in the
avidin-induced inactivation
of
modified phages by extracts
of
yeast mem-
branes was attempted.
No
such interference was observed. Finally, the
affinity cytochemical approach was taken and proved
to
be
successful.
Ferritin-biotin conjugates were used for the localization, and determina-
tion of the amount
of
receptors
per
yeast spheroplast was enumerated
(Bayer et al., 1978b). An average of
4000
receptor proteins
per
cell was
found, compared to
8000
per
cell
as
estimated from tracing the fate of the
radioactive affinity label through subcellular fractions.
Even though the number
of
transport components is
very
low, attempts
were
made to isolate such components. It was decided to isolate the
transport components via spheroplast preparation, subsequent lysis, and
affinity chromatography of membrane extracts on avidin-Sepharose.
I
I
1
I
I
1
I
i
0
5
10
15
20
25
Time
of
Uptoke
(min)
Figure
14.
Reactivation of biotin transport in spheroplasts and mercaptoethanol-treated
cells. Effect ofpBNPon ['%]biotin uptake: cells (B) and spheroplasts (SB) treated withpBNP
(10
fl;
spheroplasts from pBNP-treated cells;
and
pBNP-treated cells subjected to mer-
captoethanol (Bm); cells
(C)
and spheroplasts
(S)
controls withoutpBNP-treatment.

THE
USE
OF
THE
AVIDIN-BIOTIN
COMPLEX 39
Thus attempts were made to prepare spheroplasts from affinity-labeled
cells.
To
prepare spheroplasts in high yields, intact cells were treated with
a solution containing mercaptoethanol and
EDTA,
followed
by
the
enzymatically induced removal of the cell wall by glusulase. Surprisingly,
it was discovered that the spheroplasts, prepared in the above described
manner from affinity-labeled (hence inactivated cells), regained their
ability to transport biotin (Figure 14) (Viswanatha et al.,) It was thus
interpreted that the biotin label was liberated from the transport
component. Next it was necessary to determine the stage during sphero-
plast production at which the label was released. It was found that simply
treating labeled cells with mercaptoethanol was sufficient to cause total
reversal of BNP-induced inhibition (Bayer et al., 1975). The revived cells
were
again sensitive to BNP inhibition. This property demonstrated that a
transport system in an intact cell could
be
specifically switched on and off
upon successive treatments by a site-specific reagent (BNP) and a group-
30
20
-
$
10
-
:o
x
E
a2
1
0
Z
30
3
E
c
;
20
10
0
10
20
X)
40
50
Time
of
Uptake
(min
1
Figure 15.
cells
(A-H)
treated as indicated in Figure
16.
Switching on and
off
the biotin transport system in
yeast
cells. Biotin uptake by

40
EDWARD
A.
BAYER
AND
MEIR
WILCHEK
specific reagent (mercaptoethanol) (Figures
15
and
16).
This constituted
the first demonstration
of
a specific
inactivation-reactivation
sequence of
a specified cell-based system.
A
second inference derived from this study was that a cysteine residue
was
located
at the binding site in the transport system (Figure 17). From
the conclusions
of
this study,
it
became clear that the steps shown in
Figure
18
should
be
taken for the isolation of biotin transport com-
ponents.
As
this scheme indicates, one
of
the major steps is the specific
adsorption
to
an avidin-affinity column, followed by the elution
of
transport components by the action of mercaptoethanol
or
other thiols.
Considering the correlation between reactivation of uptake
with
the thiol-
induced release
of
the label, it is anticipated that the abovedescribed
procedure will ultimately enable the isolation of viable transport com-
ponents (Bayer, 1976).
IX. ANTIBIOTIN ANTIBODIES
To add an extra dimension of versatility to the avidin-biotin system,
antibiotin antibodies have been elicited to serve
as
additional biotin-
binding proteins (Berger, 1975). Why do we need antibiotin antibodies
when antibodies
to
other haptens can be produced much more easily?
The answer is that in this case we will then have a native interacting system
(i.e., avidin-biotin), plus an artificial interacting system (i.e., antibodies to
CELLS
A
B
c
D
E
F
G
H
Figure
16.
the
biotin transport
system
in
yeast
cells.
Scheme
employed
to
cause
inactivation-reactivation-inactivation
sequence
of

THE
USE
OF
THE
AVIDIN-BIOTIN
COMPLEX
41
H
:
p2
Q
-c
-
NH-CH-L-NH-
Figure
17.
Proposed mechanism
for
thiol reactivation.
biotin). These two systems therefore can
be
used in complementary
fashion.
For
example, the moment avidin is added to biotinylated
lymphocytes, the observed stimulation is an irreversible process. This
differs from lectin-induced stimulation of lymphocytes because lectins,
up to
6
hr after their addition, can
be
removed from the lymphocyte
surface by the competitive sugar, thereby aborting the stimulatory
process. Thus when direct mimicry of lectin-induced stimulation is
desired, antibiotin antibodies, rather than avidin, can be used to effect
stimulation of biotinderivatized cells. Addition
of
free biotin has been
found to reverse the process up to
6
hr after the addition of antibodies.
Antibiotin antibodies are also particularly applicable in other cases
wherein the intrinsic affinity between biotin and avidin is too strong to
be
useful (e.g., in isolation and purification procedures). We mentioned
previously that
we
have used an antibiotin-antibody column for separa-
tion of biotinylated hCG from the native hormone. Presumably antibiotin
antibodies will prove
to
be
expedient in future applications of this sort.
X.
CONCLUSIONS
This chapter has presented the many applications
of
the avidin-biotin
complex in molecular biology. We have also provided some of the
potential advantages of this system as a general probe, taking into account
certain restrictions in its use. We cannot, of course, foresee all its possible
uses, given our own limited imagination and/or restricted knowledge. We
do hope, however, that this review will serve to stimulate the imagination
of the readers and convert some of them to users of the method.

42
EDWARD
A. BAYER
AND
MEIR
WILCHEK
Figure 18.
transport components from yeast.
Steps
involved
in
the
affinity labeling, localization, and isolation
of
biotin
Acknowledgments
We
thank
Dr.
Ehud Skutelsky
for
collaborative studies in developing
the
techniques used
for
ultrastructural studies.
We
are indebted
to
Mrs.
Tsipora Sheer
for
the excellent secretarial
work
involved
in
preparing the
manuscript.
References
Adams,
M.
H.
(1959),Battmophnges, Wiley-Intencience, New York,
p.450.
Anderson,
R.
G.
W.. Goldstein,
J.
L.,
and
Brown,
M.
S.
(1976),
Proc.
Not.
Acad.
Sci.
US.,
73,
Angerer,
L.,
Davidson,
N.,
Murphy, W., Lynch, D., and Attardi,G. (1976),Cell,
9,81-90.
Ash,
J.
F.
and Singer,
S.
J,
(1976).
Proc.
Nd.
Acad.
Sci.
US.,
73,4575-4579.
Avrameas,
S.
(l969),
Irnmunochnn&y,
6,43-52.
Axen,
R.,
Porath,
J.,
and Emback,
S.
(1967),
Nature,
414, 1302-1304.
Bayer,
E.
A. (1976),Isr.
J.
Med.
Sci.,
14,
1361.
Bayer.
E.
A.
and Wilchek.
M.
(1974). MeWEnzymol., 34,265-267.
Bayer,
E.
A. and Wilchek.
M.
(1977), Methods
EnzymoL,
46,613-617.
Bayer,
E.
A.
and Wilchek,
M.
(1978),
Methods
Enzymol.,
in press.
Bayer,
E.
A,, Viswanatha,
T.,
and Wilchek,
M.
(1975),
FEES
Left.,
60,309-312.
2434-2438.

THE
USE OF THE
AVIDIN-BIOTIN
COMPLEX
43
Bayer, E. A., Skutelsky,
E.,
Wynne, D., and Wilchek, M.
(1976a),
J.
Hictochem.
Cytochm.,
24,
Bayer, E. A., Wilchek, M., and Skutelsky,
E.
(1976b), FEBS
Lett.,
68,240-244.
Bayer, E. A., Wilchek,
M.,
and Skutelsky,
E.
(1977),
Isr.
J.
Med. Sci., 13,953.
Bayer, E.
A.,
Rivnay,
B.,
Wilchek, M., and Skuetlsky,
E.
(1978a),
Isr.
J.
Med. Sci.,
in press.
Bayer, E.
A.,
Skutelsky, E., Viswanatha, T., and Wilchek, M.
(1978b), Mol. Cell. Biochcm. 19,
Bayer, E.
A.,
Skutelsky,
E.,
and Wilchek,
M.
(1978c), Methods EnzymoL,
in press.
Becker,
J.
M.
and Wilchek,
M.
(1972),Biochim. Biophys. Actu,
264,
165-170.
Becker,
J.
M.,
Wilchek, M., and Katchalski, E.
(1971),
Proc.
Nat.
Acad.
Sci
U.S.,
68,
2604-
Becker, J. M., Wilchek,
M.,
and Katchalski, E.
(1972), Fed.
Proc.,
31,837.
Berger,
E.,
Long, E., and Semenza,
G.
(1972), Biochim. Biophys.
Actu,
255,873-887.
Berger,
M.
(1975), Biochemisby, 14,2338-2342.
Berger,
M.
and Wood, H.G.
(1974).
J.
Biol.
Chem.,
250,927-933.
Bodanszky,
A.
and Bodanszky,
M.
(1970), ExpeTi.edu, 26,327.
Bodanszky,
M.
and Fagan, D.
T.
(1977),
J.
Am. Chcm.
Soc.,
99,235-239.
Cuatrecasas,
P.
(1969),
Proc.
Nat. Acad. Sci.
US.,
63,450-457.
Cuatrecasas, P. and Wilchek,
M.
(1968),
Biochem.
Biophys.
Res. Commun., 33,235-239.
Danon, D., Goldstein, L., Marikovsky, Y., and Skutelsky, E.
(1972).
J.
Ultrmtruct.
Res.,
38,
Eckermann, D.
J.
and Symons,
R.
H.
(197~
EUT.
J.
Biochem.,
82,225-234.
Flanagan,
S.
D. and Barondes,
S.
H.
(1975),
J.
Biol.
Ch.,
250,
1484-1489.
Fraenkel-Conrat, H., Snell, N.
S.,
and Ducay,
E.
D.
(1952a),ArchBiochem. Biophys., 39.80-96.
Fraenkel-Conrat, H., Snell, N.
S.,
and Ducay, E. D.
(1952b). Arch. Bioch. Biophys., 39,
Gasic, G. J., Berwick, L., and Sorrentino,
M.
(1968),Lab. Invest., 18,63-71.
Green,
N.
M.
(1970), Methods Enzymul.,
MA,
424-427.
Green,
N.
M.
(1972),
in
Protein-Protein Interattimu
(23rd
Coll. Gesell Biol. Chem. Mosbach/
Baden), R. Jaenicke, and
E.
Helmreich, Eds., Springer-Verlag, Berlin, pp.
183-2 11.
Green,
N.
M.
(1
975), Av. Protein
Chem.,
29,85- 133.
Green,
N.
M.,
Konieczny, L., Toms, E.
J.,
and Valentine, R. C.
(197
l),
Biochem.
j.,
125,
Green,
N.
M.,
Valentine, R. C., Wrigley,
N.
G., Ahmad, F., Jacobson, B. E., and Wood, H.
G.
Haimovich,
J.
and Sela,
M.
(1966),
J.
Immunol., 97,338-343.
Heggeness,
M.
H.
(1977), Stain Technol.,
52,
165-169.
Heggeness,
M.
H. and Ash,
J.
F.
(1977)~. Cell
Biol.,
73,783-788.
Heitzmann, H. and Richards, F.
M.
(1974),
Proc.
Nat. Acad. Sci
US.,
71,3537-3541.
Hofmann, K. and Kiso,
Y.
(1976),
Proc.
Nat. Acad. Sci
U.S.,
73,35 16-35
18.
Hofmann, K., Finn, F.
M.,
Friesen,
H.
-J.,
Diaconescu, C., and Zahn, H.
(1977),
Proc.
Nat
Jakoby, W. B. and Wilchek,
M.,
Eds.
(1977), Methods in Enzymology,
Vol.
46: AfiniOLabeling,
Jarrett,
L.
and Smith,
R.
M.
(1975),
Proc.
Nat.
Acad.
Sci
US.,
72,3526-3530.
Jawiewicz, M. L., Schoenberg, D. R., and Mueller,
G.
C.
(1 976),
Exp.
Cell Res.. 100.2 13-2 17.
Kahane,
I.,
Polliack, A., Rachmilewitz,
E.
A.,
Bayer,
E.
A.,
and Skutelsky,
E.
(1978), Nature,
Knappe,
J.
(1970), Ann. Rev.
Biochem.,
39,757-776.
933-939.
23-29.
2607.
500-510.
97-108.
781-791.
(1972),
J.
BWl.
Chem.,
247,6284-6298.
Acad.
Sci
U.S.,
74,2697-2700.
Academic Press, New York.
271,674-675.

44
EDWARD
A.
BAYER
AND
MEIR
WILCHEK
Koivusalo, M., Elorriaga,
C.,
Kaziro, Y.. andOchoa,
S.
(1963)~. Bwl.
Ch.,
938,1038-1042.
Koreman,
S.
G.
and OMalley,
B.
W.
(1970), MehnikEnzytnoL, MA, 427-430.
Landman, A. D. and Dakshinamurti, K.
(1973),Anal.
Biochem.,
56,191-195.
Landman, A. D. and Dakshinamurti, K.
(1975),
Biochem.J.,
145,545-548.
Lane, M. D., Edwards J., Stoll, E., and Moss,
J.
(1970),
Vihm.
HUT.,
48,345-363.
Lichstein,
H.
C.
and Waller,
U.
R.
(196l)J.
Baclnwl.,
81.65-69.
Lin, H.
J.
and Kirsch,
J.
F.
(1977), Ad.
Biochem.,
81.442-446.
Ling, N.
R.
and Kay,
J.
E.
(1975),
Lymphocyte
Shmhiota,
2nd ed., North-Holland,
Amsterdam.
Liu,
W.,
Yang, K., Burleigh,
B.
D.,
and Ward, D.
N.
(1974).
in
Hmne
Biding
and
Target
Cell
Activation
in
Testir,
A.
R. Means, and M. Dufau. Eds., Plenum Press, New York,
pp.89-110.
Luft, J.
H.
(1961),
J.
Bzophys.
Bioch.
Cybl.,
9,409-414.
Lynen,
F.,
Knappe, J., Lorch, E., Jutting,
G.,
and Ringelmann, E.
(1959), Angm.
Chcm.
71,
Makela,
0.
(1966).
Immunology,
10,81-86.
Manning, J. E., Hershey, N. D., Broker,
T.
R., Pellegrini,
M.,
Mitchell, H. K.,
and
Davidson,
Manning,
J..
Pellegrini, M.,
and
Davidson,
N.
(1977),BiocMhy, 16, 1364-1370.
May, J.M., Williams, R. H.,
and
de Haen,
C.
(1978).
J.
Bwl.
Ch.,
253,686-690.
McCormick, D. B.
(1965), Ad.
Bioch.,
13, 194-198.
McCormick, D.
B.
and Wright, L. D., Eds.
(1970),MethodrinEnzymology,
Vol.
18A,
Academic
McCormick, D. B. and Roth.
J.
A.
(1970),
Mew
Enzymol.,
18A, 383-385.
Moller,
G.,
Ed.
(
1972),
Transplunkrtiun
Reuiews,
Vol.
1
1.
Monsigny, M., Kieda,
C.,
Gros,
D.,
and Schrevel,
J.
(1976),
Proc.
Sixth
Eur.
Cong.
Electron
Moss,
J.
and
Lane,
M. D.
(1971),Adv.
Enzyml.,
35,321-442.
Nicholson,
G.
L. and Singer,
S.
J.
(1971),
Proc.
Nat.
Ad. Sci
US.,
68,942-945.
Novogrodsky, A. and Katchalski, E.
(197
I),
FEES
Lcft.,
12,297-300.
Otto, H., Takamiya, H., and
Vogt,
A.
(1973)J.
Immuml.
Methodc,
3,
137-145.
Pelligrini, M., Holmes, D.
S.,
and Manning.
J.
(1977),NwfeicAcidrRes., 4,2961-2973.
Raff, M.
C.
(1976). Sci. Am., 434,30-39.
Ravid, A., Novogrodsky,
A.,
and Wilchek,
M.
(1978),
Eur.
J.
Immunol.,
in
press.
Riesel,
R,,
Bayer,
E.
A.,
Wilchek, M.,
and
Amsterdam, A.
(l977),
Isr.
J.
Med.
Sci.,
13,968.
Rogers,
T.
0.
and Lichstein, H.
C.
(1969a),
J.
Bachl.,
100,557-564.
Rogers,
T.
0.
and Lichstein, H.
C.
(1969b).
J.
Bachl.,
100,565-572.
Rylatt, D.
B.,
Keech, D.
B.,
and Wallace,
J.
C.
(1977),Arch.
Biochem. Bqbhys.,
183, 113-122.
Singer,
S.
J.
(1967). Adv.
ProUin
Ch.,
24, 1-54.
Singer,
S.
J.
and Schick, A.
F.
(1961),
J.Blophys.Biochem.
Cytol.,
9,519-537.
Skutelsky, E. and Bayer
,
E.
A. (1976a),
In.
J.
Med.
Sci.,
12. 1355-1356.
Skutelsky,
E.
and Bayer,
E.
A.
(1976b),
Proc.
Skth
Eur.
Cong.
Electron
Microsc.
Jerusala,
Skutelsky. E., Danon, D., Wilchek, M., and Bayer,
E.
A.
(1977),
J.
Lrlrrcrrtruct.
Res.,
61,
Skutelsky, E., Bayer,
E.
A., Wilchek, M., Matzner, Y., Kahane,
I.,
and Polliack, A.
(1978),
Spencer, R. P. and Brody,
K.
R.
(1964), Am.
J.
Physwl.,
206,653-657.
Viswanatha,
T.,
Bayer,
E.,
and Wilchek, M.
(1975),
Biochim.
Bzophys.
Ach,
401,152-156.
Wakil,
S.J.,
Titchener,
E.
B., and Gibson, D. M.
(1958),
Biochim. Biophys.
Acto, 29,225-236.
481-486.
N.
(1975),
Chromosoma
(Berlin)
53,
107- 1 17.
Press,
New York.
Microsc.
Jerusalem,
3940.
198-200.
325-335.
Twelfth
Int.
Leuk.
Cult.
Conf.,
in press.

THE
USE
OF
THE
AVIDIN-BIOTIN COMPLEX
45
Wallace,
B.
A.,
Richards,
F.
M., and Engelman, D. M.
(1976)J.
Mol.
Bzol.,
107,255-269.
Waller,
J.
R.
and Lichstein, H.
C.
(1965a)J.
Bacterial..
90,843-852.
Waller,
J.
R.
and Lichstein, H.
C.
(1965b)J. Bactnwl. 90,853-856.
Wei,
R.
D.
(1970),
McthodcEnzymoL,
MA,
424-427.
Wilchek, M. and
Givol,
D.
(1977),
MethodtEnrymof.,
46,
153-157.
Wilchek,
M.
and Gorecki, M.
(1973),FEBSLett., 31,149-152.
Wilchek, M. and Hexter,
C.
S.
(1976),
Mcthodr
Biochen
Anal.,
23,347-385.
Wynne,
D.,
Wilchek,
M.,
and
Novogrodsky,
A.
(1976). Bkh. Biophys.
Rcs.
Commun.,
68,
730-739.