Combinatorial Phage Antibody Libraries
79
plasma cells. In the resting state, plasma
cells will be found in higher numbers in
the bone marrow, and this is often a prefer-
able starting point for library construction.
These early experiments illustrated the
validity of the combinatorial approach.
Despite the disruption of the original
in
vivo
heavy–light chain combinations, the
frequency of productive Fabs was higher
than what may have been envisioned
a
priori
. This result can be explained by chain
promiscuity – the ability of a particular
light or heavy chain to accept a number of
different heavy or light chains respectively,
while retaining antigen binding.
The lambda phage system, however,
was hampered by its inef±cient screen-
ing procedure. This method is laborious
and restrictive of the number of clones
that may be examined, which is a lim-
iting factor when working with libraries
prepared from sources other than a hyper-
immunized one. Furthermore, screening
is con±ned to antigens that are available
in large quantities and in a puri±ed form.
What was needed was a library system
in which positive clones could be af±nity-
selected by minute quantities of antigen.
To address these problems, several groups
developed systems to display antibody frag-
ments on the surface of the Ff ±lamentous
phage. We will concentrate here on the
pComb3 phagemid vector system. Other
systems are described in detail elsewhere.
2.2
Constructing Antibody Libraries on the
Surface of a Phage
In vivo
, antibody immunity is driven largely
by selection. Particular B-cells may be trig-
gered to replicate and differentiate on the
basis of recognition of a given antigen by
the antibody displayed on their surface.
B-cells displaying an antibody that has no
speci±city for the same antigen will remain
passive. In this manner, the immune sys-
temisabletohomeinonanantigenicde-
terminant. Pioneering studies suggested
a means by which selection from com-
binatorial libraries might approximate to
native immunity. The work demonstrated
ligand-driven selection, or biopanning,
from peptide libraries displayed on the
surface of a nonlytic ±lamentous phage.
It was reasoned that monoclonal antibod-
ies presented in a similar fashion might
also be identi±ed from within a library of
many antibody phages by interaction with
immobilized antigen. The general scheme
for cloning monoclonal antibodies from
combinatorial libraries on the surface of a
phage is outlined in Fig. 1.
Many different fragments of antibodies
have been displayed on phages including
Fabs, single-chain Fv (scFv) fragments,
and single-domain antibody fragments (in-
cluding human and camelid antibodies).
The absence of antibody constant regions
in Fv fragments means that these con-
structs will not ef±ciently pair to form
an antibody-binding pocket. Therefore, the
variable domains must be joined by means
of linker sequences using PCR primers.
Single-chain antibodies have, however,
been shown to reproduce the binding prop-
erties of the corresponding whole antibody
with varying degrees of ef±ciency. Note
that there are two different types of scFvs:
‘‘light chain ±rst’’ and ‘‘heavy chain ±rst,’’
and the order can subtly affect the binding
properties and stability of the scFv. An-
other feature of the scFv is that different
linker lengths can affect the interaction
between the heavy- and light-chain frag-
ments. For example, very short linkers can
restrict the heavy-chain fragment such that
it will prefer to associate with the light
chain of another scFv more readily than
with its own light chain, thus producing a
previous page 1399 Encyclopedia of Molecular Cell Biology and Molecular Medicine read online next page 1401 Encyclopedia of Molecular Cell Biology and Molecular Medicine read online Home Toggle text on/off