Materials and Methods for Selecting Phage Displaying
Ligand Binders
Field of the Invention The present invention relates to materials and methods for selecting ligand binders displayed on the surface of phage particles that bind to a target ligand, and more particularly to a method for selecting phages displaying ligand binders using the interaction between the ligand binder and an infection mediating complex comprising a target ligand and a component capable of making the phage infective towards bacteria having a certain type of infection mediating structure.
Background of the Invention
The display of peptides and polypeptides on the surface of bacteriophage (phage) fused to one of the coat proteins is a powerful technique for the selection of specific ligands. The display of small peptides as fusions with phage coat proteins was first described by Smith in 1985 and has since been used in a range of different applications.
The principles behind phage display technology are as follows. Nucleic acid encoding the peptides or polypeptides for display is cloned into a phage or a phagemid vector (i.e. a vector comprising origins of replication derived from a phage and a plasmid) that can be packaged as single stranded nucleic acid in a bacteriophage coat. When phagemid vectors are employed, a "helper phage" is used to supply the functions of replication and packaging of the phagemid nucleic acid. Since the cloned nucleic acid is expressed fused to the coat-anchoring part of one of the phage coat proteins (typically the p3 or p8 coat proteins in the case of filamentous phage) , it is displayed on the surface of the phage, thus linking phenotype to genotype. The resulting phage will express both wild type coat protein (encoded by the helper phage) and modified coat protein (encoded
by the phagemid) when a phagemid vector is used, and only modified coat protein when a phage vector is used. Selection based on phenotype will give the genotype that can be sequenced, multiplied and transferred to expression systems.
Several randomized combinatorial peptide libraries have been constructed to select for peptides that bind to different targets, e.g. cell surface receptors or DNA (reviewed in Kay, 1995) . Proteins and multimeric proteins have been successfully phage-displayed as functional molecules (EP 0349578A, EP 0527839A, EP 0589877A, Chiswell and McCafferty, 1992) . Functional antibody fragments (Fab, single-chain Fv (scFv) ) have been expressed (McCafferty et al . , 1990; Barbas et al . , 1991; Clackson et al . , 1991), and some of the shortcomings of human monoclonal antibody technology have been superseded since human high affinity antibody fragments have been isolated (Marks et al . , 1991; Hoogenboom and Winter, 1992) .
A major problem with phage display work aiming at isolating binding proteins with improved affinity or specificity for a target ligand resides in the selection of phage expressing the protein fragments of the desired quality. A widely used method for selection is "panning" , in which phage stocks displaying ligands are exposed to solid-phase coupled target molecules. However, panning often results in high backgrounds due to the binding of non-specific phage.
Methods known as SAP (Selection and Amplification of Phages; WO 95/16027) and SIP (Selectively-Infective Phage; EP 0614989A) employ selection based on the amplification of phages in which the displayed ligand specifically binds to a ligand binder. In one embodiment of the SAP method, this is achieved by using non-
infectious phage and connecting the ligand binder of interest to the N-terminal part of p3. Thus, if the ligand binder specifically binds to the displayed ligand the otherwise non-infective ligand-expressing phage is provided with the parts of p3 needed for infection. The selection can thus be based on kinetic parameters (Duenas et al . , 1996) . This is because as the ligand-binder interaction is reversible, the degree of complex formation will be concentration dependent. Phage carrying high affinity ligands, i.e. with high equilibrium constants for association (which in turn depends on the kinetic rate constants) can thus be selected for by varying the concentration of ligand.
Summary of the Invention
In order to allow a rational use of the phage display technology, there is a continuing need for efficient screening methodologies to achieve rapid and accurate selection of ligand binders displayed by phage.
The present invention provides materials and methods for selecting phage displaying specific ligand binders, e.g. antibodies or antibody fragments. Large antibody libraries constructed using conventional technology with phage or phagemid vectors can be used in these methods.
In contrast to the SAP and SIP methods described above, this method does not rely upon creation of phage particles that have been engineered to be non-infectious, allowing the use of ordinary phages and helper phages. Instead, phage retaining their wild-type ability to infect bacteria via one type of infection mediating structure (e.g. F-pili) are engineered to display ligand binders so that binding with an infection mediating complex comprising a ligand which specifically binds to the ligand binder causes the phage to become infectious towards bacteria having a different infection mediating
structure (e.g. N-pili) to which the phage were initially substantially non-infectious .
Thus, the infectivity of phages for bacteria is controlled by selective use of host cells and ligands combined with infection mediating proteins or polypeptides that make phage infective towards a given type of bacteria on binding of the infection mediating complex to the displayed ligand binder. The present invention can therefore be used in combination with conventional libraries.
Accordingly, in one aspect, the present invention provides a method for selecting phages displaying ligand binders that specifically bind to a target ligand, the method comprising:
(a) constructing a library of ligand binders displayed on the surface of phage particles, wherein the phage are substantially incapable of infecting bacteria having a first type of infection mediating structure but are capable of infecting bacteria having a second type of infection mediating structure;
(b) constructing an infection mediating complex comprising a target ligand and a component capable of mediating infection via the first type of infection mediating structure;
(c) exposing the infection mediating complex to the phage particles displaying the ligand binders so that the specific binding between a ligand binder displayed on the phage and the target ligand causes the phage to become infective towards bacteria having the first type of infection mediating structure; and,
(d) adding bacteria having the first type of infection mediating structure so that the infective phage particles produced in step (c) can infect the bacteria with nucleic acid encoding the ligand binder.
In the above method, the "infection mediating structure" is a structure of the bacteria which phage can use to infect the bacteria via interaction with phage coat proteins. Examples of infection mediating structures in filamentous phage include F-pili which interact with Ff p3 or minor coat protein, and N-pili or I-pili which interact with IKe p3 protein. Embodiments of the invention employing different types of pili are described in more detail below. However, other types of phage can also be used in the invention, such as λ phage, in which case the lamB gene product of the bacteria could be used as the infection mediating structure (see Bradley et al . , 1980) .
By way of example, a library of ligand binders contained in, e.g. M13 filamentous phages, will not significantly infect bacteria expressing N-pili, but which are devoid of F-pili. Phage particles displaying ligand binders, such as antibodies specific for a certain antigen, can be given the ability to infect N-pili carrying bacteria if they bind to an infection mediating complex comprising the N-terminal part of p3 from N-pili restricted phage like IKe (Khatoon et al . , 1972) and the antigen. The antigen may be linked to the N-terminal part preferably comprising the glycine rich linker region of IKe p3
(Marzari et al . , 1997) as a fusion protein or chemically linked using any of several known methods. Conversely, phage normally restricted to infect N-pili carrying bacteria can be exposed to antigen linked to p3 from M13 of Ff phages to allow this phage to infect F-pili expressing bacteria. It may also be possible to manipulate the infectivity of phage, e.g. F-pili restricted phage, to become N-pili restricted by substituting the N-terminal region (s) of p3 from such phages to an N-terminal region from F-pili restricted phage .
In embodiments employing phagemids to generate the phages displaying the ligand binders, plasmids containing the genetic information for the ligand binder can be prepared from the selectively infected bacteria, introduced into F-pili (or N-pili) carrying bacteria and packaged into phage particles by use of a helper phage. As an example, a M13 helper phage could be used to infect F-pili carrying bacteria, leading to the production of phage particles capable of infecting F-pili bacteria, but substantially incapable of infecting N-pili carrying bacteria.
Alternatively, if phage vectors were used, the plasmids introduced into the F-pili (or N-pili) carrying bacteria would carry all the genetic information for formation of phage particles and there would be no need for a helper phage. The phage particles thus produced can then be subjected to an additional round of selection using the infection mediating complex comprising the antigenic epitope and the N-terminal part of N-pili restricted p3 and N-pili carrying bacteria.
In an alternative manner, phage particles restricted on one type of pili might be changed to be restricted on another type of pili. It has been demonstrated that M13 phage with the ability to infect bacteria expressing F- pili can be given the additional ability to infect bacteria expressing N-pili. This was achieved through transplantation of the N-terminal part of p3 from an N- pili restricted phage (IKe) to p3 of M13 (Marzari et al . , 1997) . The resulting chimeric p3 contained domains that bound both types of pili. Alternatively, the removal of the F-pili binding domain of, e.g. M13 p3 , and transplantation of the N-terminal domain of IKe p3 would result in a phage with changed pili restriction now being able to infect N-pili expressing bacteria, but not F-pili expressing bacteria. The construction of such a phage
could be achieved through genetic manipulation of the phage genome using conventional techniques.
In a preferred embodiment, the present invention can be started using either N or F-pili bacteria, in conjunction with their respective phages and fusion proteins. In embodiments employing phage vectors, steps (a) to (d) can be used, and optionally the phage particles displaying ligand binders that are formed as a result of the method can be subjected to a second round of selection, starting the process again at step (b) .
It will be apparent that the above method can be repeated one or more times employing the general principles set out above to reduce the number of ligand binders selected and/or to obtain ligand binders having improved properties, e.g. in the case of antibody fragments, improved binding specificity or affinity for an antigen.
The method can additionally comprise one or more of the following steps:
(e) amplifying the nucleic acid transferred to the bacteria to multiply and/or selecting for bacteria infected by the phage; and/or, (f) recovering the nucleic acid from the bacteria; and/or,
(g) using the nucleic acid recovered in step (f) to repeat steps (a) to (d) .
In some embodiments, the present invention includes the step of constructing the library of phage particles displaying the ligand binders, the method comprising: (a') infecting bacteria having a second type of infection mediating structure with phage comprising nucleic acid encoding a library of ligand binders, or a phagemid vector comprising nucleic acid encoding a library of ligand binders and a helper phage, wherein the
ω ω t to in o in o in in
μ- μ- TJ CD J 3 3 Ω μ- OJ tr TJ Pi Ω rt 3 μ- £ o μ- μ- m Hi TJ tr rr 3 μ- J
3 3 tr TJ tr Φ Φ 0 3 ri ØJ tr μ- O rr φ 3 3 tr Hi 3 3 rr μ- tr μ- ø) Φ 3 tr
Hi Hi ØJ φ OJ Pi Pi M Hi Φ < OJ Q 3 0J rt Hi μ- Hi Hi 3 ii 0J 3 < Pi Hi 0J
Φ Φ LQ Ω LQ μ- μ- TJ φ μ- LQ TJ TJ rt rr ø) Φ Ω rt Φ φ CD m LQ Pi μ- μ- Φ LQ
Ω Ω φ μ- Φ OJ øj r Ω Ω 3 φ n 0 Ω rr tr Ω Ω - rt Φ Φ 3 0) Ω Φ
.— *. ft rr Hi ^ rr rt μ- . — . rt OJ LQ 0J ^-^ μ- CD Pi Hi ft Φ ft rt li ≤ LQ rt rt α μ- μ- ØJ μ- TJ Ω μ- μ- ω tr μ- TJ 0) ^ ØJ CQ TJ 3 μ- Ω μ- ft rt tr CQ tr μ- μ- 0
0 <! 3 Ω OJ — 3 3 μ- — 0 0J øJ r φ — μ- Φ Hi ii J ØJ 0 tr tr a Φ Φ 0J 3 3 li
3 Φ Pi n LQ in 3 3 tr Φ Pi 3 Ω 0 rt φ n μ- 3 0J Φ μ- P. i LQ LQ tr rt LQ Hi LQ μ- li tr li LQ Ω Φ 3 μ- Φ CD r
ØJ 3 rt rt μ- μ- Φ CD μ- Ω 3 Φ μ- 0- 0 Ω .. Hi Φ rt μ- 0J 3 rt TJ LQ CQ μ- Φ CD tr Φ
Φ 0 tr 3 Ω £ rt 3 ØJ 0 Φ r 3 3 O μ- CD n O φ 3 Φ Φ tr 0 TJ 3 Ω rt 0J
Pi Pi S φ TJ ri Hi 3 Pi 0 CD tr 3 Ω φ CQ Pi Pi li 0J Hi μ- O ii Ω J μ- μ- OJ μ- φ 0 3 Φ rt C0 μ- Hi rt CD rt m 01 0J CD μ- μ- LQ 3 01 rt 3 3 rt φ
3 øj r rt 3 CD CD Ω Ω ØJ rt ø rt tr rt φ CD 3 rt tr OJ øj Φ μ- Ω * tr Pi Ω Φ ii
LQ rt Pi OJ LQ μ- rt rt r ri rt μ- rt OJ Φ n Ω TJ Ω tr μ- rt 3 0) Φ Φ rt ri μ- CQ ι-i Pi 3 3 μ- LQ 3 μ- 3 a 3 3 *< rt Φ tr φ 3 μ- O J Hi TJ rt 3 μ- TJ tr 3 LQ tr μ- LQ H 0 φ Ω 3 Hi rt CQ Ω μ- Ω 3 Hi ØJ Φ 0J 3 a ri øj tr
ØJ LQ tr Φ Φ CD Φ 3 rt rt LQ Φ Φ μ- 3 rt tr 3 rt tr Ω Φ iQ li Ω tr O φ Φ ØJ
Ω 0J rt rt J rt -• μ- Ω 0J i-i μ- μ- LQ - 0J 0 li CD rt rt 3 rt Φ rr LQ rt CQ Ω ≤ H tr < 3 CD ft 0 H" Hi 3 3 Ω 3 CQ rt μ- μ- Φ tr tr 0J Φ φ rt rt Φ OJ Φ μ- μ- LQ rt μ- Hi 0J LQ Pi TJ rt rr TJ rt rt φ Ω O rt 0 0 3 <!
H H Φ μ- Φ ^ OJ LQ ri 3 •^ Ω tr tr φ 0 O H TJ r-1 3 0 tr Pi Hi rt μ- ØJ μ- 3 ri LQ 3 μ- μ- ø) ØJ 3 LQ μ- Φ ØJ rt OJ Φ li 3 3 φ Hi Φ 3 li
0J Ω μ- 0J 3 3 rt 3 3 Ω 3 μ- O LQ μ- Φ rt Ω ,— . CQ 3 TJ μ- OJ LQ Φ rt øj 3 OJ LQ Hi rr P- rt tr Hi 3 O Φ TJ øj 3 tr rt 0J Φ μ- CQ ri 3 r rr 3 Φ Φ μ- 3 OJ Φ Ω Hi μ- 0J CD r • rt Φ 3 - μ- 3 3 0 Hi Φ rt CD
0J H tr rt Ω OJ 3 ri Ω Ω 0) σ Φ ϋ — 3 μ- Hi n <! Φ rr 3
<! Φ ØJ Ω μ- tr rt Hi 3 Hi Φ rt rt TJ TJ r rt Pi CD s: μ- Φ ØJ Φ Hi μ- Ω Ω φ
-• tr C μ- ~. < OJ LQ φ μ- μ- Pi Φ Φ μ- OJ tr ØJ OJ μ- Φ tr 3 <! rt rt Ω 01 Pi rt OJ CD
3 μ- 3 0J 0 rl Ω n 0 tr ØJ H l to 3 μ- Hi μ- μ- rr μ- rt Ω Φ μ- TJ Hi rt
LQ 3 CD 3 3 CD 0J rt μ- 3 LQ "< LQ TJ rt Ω Φ 3 øj Φ 3 μ- Φ O 0J μ- 0J
LQ Φ P. μ- rt μ- OJ φ Φ Φ rr Ω LQ 3 3 tr n 3 rt CD LQ 3 Ω 0 3 0 ft 0J μ- rt rt rt LQ O ØJ CD rt tr rt tr øJ Φ rt 0 3 tr Φ O TJ Hi ^ 3 li μ- tr tr μ- CD Hi 3 Φ rt μ-
Φ rr rt μ- 3 Pi τ 3 0J Pi Hi ØJ μ- <J Φ O Φ Φ tr rt tr Φ øj φ rr 3 Pi μ- TJ 3 < μ- H. μ- 3 φ 3 3 ri r 0J TJ μ- Pi 0 rt
Hi Φ Pi ø Φ 0 Φ μ- ØJ μ- rt μ- LQ LQ 3 p. ØJ Hi 01 3 Ω tr tr μ- Hi μ- Hi Φ tr rt 3 Pi 3 rt 3 μ- LQ 0J rt Φ 3 tr μ- 11 Ω rt ØJ n 0J a "< ι-i μ- J 11 μ- μ- 0 Φ μ- LQ μ- Hi Ω 0J 3 μ- ri Φ CQ H ^ rt Φ LQ ØJ rt μ- Φ
CD tr 3 3 Hi 3 ø 3 Φ 3 Pi μ- O CQ α Φ CD 3 ri φ li μ- 3 μ- rt CQ ø) Pi Pi LQ rr rt ØJ LQ Ω Φ Pi LQ 3 μ- 3 rt 01 ri μ- < 3 Hi 0 3 rt LQ μ- Φ μ- μ- rt CQ 0J rt øj o li Φ øj TJ LQ Φ Hi Ω rt Φ CD H Ω 3 Ω 3 co CD μ- tr rt 3 TJ tr rt Φ rt Φ . 01 0 Ω 0J a rt TJ CD 0 Hi 0) LQ Φ rt 3 μ- tr Pi ri Φ μ-
Ω n ifl ≤ 3 Φ 0 3 0 a rr li Hi CD ft μ- TJ rt 3 Φ TJ 3 0J rt rt μ- 3 OJ
Φ τ o 0) CD TJ Ω øJ Ω 0 3 - Pi σ <. TJ LQ Hi Φ 3 μ- H 3 Hi tr
Φ ^ 0 rt tr O 3 Ω tr Φ φ 3 μ- μ- tr 0J μ- Ω μ- 3 LQ φ
0 tr φ Φ μ- h-1 M rt ØJ ri ri φ 3 Pi 0J Ω tr 0 tr 3 H LQ Ω Ω φ
Hi 0 Φ Pi ft X 0 φ TJ 3 Ω Φ CD rt Pi Φ LQ O μ- Hi l-1 LQ Φ 0) ft tr rt
Hi Ω rr 3 rt 11 rr μ- tr φ CD Φ 3 3 φ CQ 3 3 0) μ- 0
0 0 0J rt 0 Φ Φ Φ 3 0 ii J Pi ø) - i li Ω 0 Hi
3 3 rt 0 Hi X T3 ι-i P. CQ ØJ μ- rt Φ rt 3 φ Φ tr μ- rt Φ 3 O rt Φ rt rt 3 ØJ - X LQ rr 11 tr rr 0 rt Φ Φ μ- Φ Φ Hi øj
infection mediating structure so that the infective phage particles produced in step (c) can infect the bacteria with nucleic acid encoding the ligand binder;
(e) using a helper phage to infect the bacteria, wherein the helper phage is restricted so that the phage particles produced by the bacteria are substantially incapable of infecting bacteria having a second type of infection mediating structure but are capable of infecting bacteria having the first or a further type of infection medaiting complex;
(f) constructing an infection mediating complex comprising a target ligand and a component capable of mediating infection via the second type of infection mediating structure; (g) exposing the infection mediating complex to the phage particles displaying the ligand binders so that specific binding between ligand binder displayed on the phage and the target ligand causes the phage to become infective towards bacteria having the second type of infection mediating structure;
(h) adding bacteria having the second type of infection mediating structure so that the infective phage particles produced in step (c) can infect the bacteria with nucleic acid encoding the ligand binder;
Of course, the above steps can be carried out simultaneously, sequentially and in a different order to that set out above. By way of example, the order of steps (c) , (d) and (e) can be varied.
In some embodiments, the method includes the steps of:
(i) optionally allowing the nucleic acid transferred to the bacteria to multiply and/or optionally selecting for bacteria infected by the phage; and, (j) optionally recovering the nucleic acid from the bacteria.
Thus, in embodiments employing helper phage, steps (a) to (h) can be repeated as necessary to obtain one or more ligand binders having the desired specificity.
In a further aspect, the present invention provides phage and/or phage bound to an infection mediating complex for use in the above method.
Embodiments of the present invention will now be further described by way of example and not limitation with reference to the accompanying drawings .
Brief Description of the Drawings
Figures 1 and 2 show schematically embodiments of the invention employing a phagemid vector system and using N- and F-pili expressing bacteria and M13K07 and IKe as helper phages (a chimeric variant of M13K07 with the same pili restriction as IKe phage) . An alternating use of ligand binder fused to an F-pili binding phage protein and ligand binder fused to an N-pili binding phage protein, to achieve efficient selection of phage- expressed ligands.
Figure 3 shows schematically an embodiment of the invention employing a phage vector system to select antibodies capable of binding to an antigen.
Figure 4 shows an inhibition assay using approximately 1 x 103 IKe phage to infect N-pili bearing cells.
Detailed Description
A number of different types of phage are known having different specificities for infecting target bacteria. In embodiments of the invention using filamentous phage, the phage generally utilize surface structures on bacteria known as pili, which are used by the bacteria, e.g. for specific binding of macromolecules and for
uptake of foreign DNA, to enter the target bacterium in order to complete their life cycles. Different types of pili have been recognized on gram-negative bacteria (Bradley, 1980) . Long, flexible pili of the F type are utilized by the commonly used filamentous phage of the Ff type (e.g. M13, fd, fl) for target cell binding and subsequent infection. The Ff phage coat protein responsible for this recognition are named p3. Ff p3 has been characterized in detail (Beck and Zink, 1981) and is composed of a polypeptide of 424 amino acid residues synthesized as a preprotein with an 18 residues leader sequence. The p3 protein is structurally organised in three domains, Nl , N2 and CT (carboxy terminal), where the Nl domain has been shown to be involved in penetration of the bacterial membrane and the N2 domain contains the binding surface essential for specific F- pili recognition (Stengele et al . , 1990).
A related filamentous phage, named IKe, has a variant of the p3 coat protein that recognizes a shorter, rigid type of pili, called N-pili. This N-pili binding domain is located upstream of the penetration domain, contrary to the N2 domain of fd and completely different (Endemann et al . , 1992). The IKe penetration domain is homologous to the Nl domain of fd phages, as well as the C-terminal regions (coat-anchoring) . Attempts to exchange the p3 molecule of IKe with the p3 of Ff phages has failed, although deletion mutants of IKe p3 are assembled into Ff phages that can attach to N-pili, but are non-infectious (Bross et al . , 1988; Endemann et al . , 1993). A chimeric fd phage, with the N-pili binding domain of phage IKe p3 grafted in front of the Nl domain of wild type fd p3 , was able to attach and infect both N- and F-pili carrying bacteria (Marzari et al . , 1997), thus showing that the fd penetration domain is functional in the infection of bacteria carrying either N or F pili. The restriction of filamentous phage for certain target bacteria seems thus
![Figure imgf000014_0001](https://patentimages.storage.***apis.com/97/bb/c9/cf1acd16af0800/imgf000014_0001.png)
> Hi ri Hi Ω 0J ØJ Φ Ω TJ s
Pi 3 rt OJ ø) 0J OJ H rr rt to øJ Ω μ- > Hi ?v $, Ω rt φ Φ CQ 3 Φ ii 0 3 3 x 0 tr μ- Φ μj ri 3 3 3 ri 3 01 tr 0J CD 01 3 CD O 3 tr 0 O
CQ CD H rt OJ 3 rt rt øι 3 ØJ CQ rt μ- ØJ Pi rt 0 Φ <! 0 Ω CD ri O μ- ri μ- Ω 3 M 01 LQ TJ μ- μ- 3 CO LQ TJ tr LQ 3 μ- rt rt μ- 0. rt rt Φ rt 3 s; Ω H tr ii ri φ μ- 3 tr tr TJ rt φ 0 ØJ CQ CD tr - 3 tr 3 φ φ tr tr Φ TJ tr Φ Φ
Φ μ- 3 3 φ Φ 0 0 (-■ ri 0J α 3 Ω 3 0 Φ 0 Φ LQ ri Φ ft Φ 3 tr Φ Ω tr rt Φ H- 3 X Pi P- Φ 3 Ω • ; CO H tr Pi ri tr μ- tr rt 0J μ- rt φ μ- 3 3 rt • μ- μ- - Ω 0 μ- CD μ- • Φ TJ rt 0J 01 μ- Φ 3 μ- LQ LQ CD 3 tr P. 0 LQ LQ CO Φ φ rt OJ 0 0 tr TJ rt Φ Ω H tr <! 3 φ O Φ Φ r Φ μ- 3 rt CQ CD rt μ- rt Hi Hi μ- r ri 01 3 Φ Φ μ- Hi μ-1 rt 3 3 01
3 tr Φ tr CD 0 s tr 0 3 μ- 0J a CQ 3 tr Φ 0J tr Φ O φ ri rt
Φ 0 Hi f μ- 0 Φ 3 TJ Φ rt P. 0 ft OJ X Φ φ iQ LQ 0 o 3 0 P. μ- Hi IQ O μ- li 3 μ- 3 ri li μ- tr Φ 3 Φ 3 0J CD 3 Φ CQ rt ro α CD μ- Φ
3 Φ OJ μ- Φ rt LQ TJ 0 0 rt Φ ri CQ α 3 rt 3 rt rt μ- ø) TJ rt ø rt OJ
LQ μ- tr LQ * OJ rt 0J ri Hi rt rr % Hi " TJ Ω Φ tr CQ O LQ tr tr rt
3 0 0J 0 Φ 3 tr 3 Φ Φ Φ TJ 0J OJ 01 μ- rt Φ 3 φ o 0J Φ Φ Ω ri
CQ < 3 r CQ Pi Φ rt CQ ft μ- ri i ØJ Ω Φ 3 Φ TJ 3 μ- 3 3 > <! ^ Pi Φ Φ
TJ μ- φ Pi Φ % μ- Φ tr 3 Φ ri rt 3 rt CD 01 < Ω LQ CQ 3 Φ 3 φ TJ CQ
Φ CD 0J Ω LQ 3 Φ - tr CQ Φ 0 N μ- tr Φ φ Φ Φ ^ ft tr tr j rt
Ω ØJ tr 3 01 0 Φ rt μ- φ ri •^ LQ 0 3 μ- 3 Hi Pi TJ φ Φ ØJ ^ ri μ- μ- μ- 0 3 3 3 μ- 3 3 3 μ- CQ 3 Φ Hi φ rt 3 Φ 3 μ- ri X O Ω LQ D μ-
Hi 3 TJ 3 CO CD μ- 3 CQ Pi rt 3 Φ 3 μ- Hi rt t-y 0J O OJ rt rr Φ rt 3 Ω μ- 1 Pi 0 Φ rt 3 Hi μ- μ- rt 0) CD m 0 0 0 μ- rt Pi 3 tr 3 1 tr n rt
Ω rt φ Φ Hi Ω s ri 0 <J Φ 3 3 μ- Φ 3 ^ μ- Hi 3 li Ω μ- μ- 3 TJ Φ μ- TJ Φ Hi μ- μ- tr Hi li 3 M 3 3 Φ Ω LQ LQ 3 ri Pi 01 iQ 3 3 3 Ω -y li Λ μ- 01 O rt Φ φ rt M μ- Ω 3 ft < Ω 3 l-1 OJ CQ OJ μ- LQ rt φ 3 J Ω 3
< r 0 tr Φ LQ rt TJ rt μ- rt TJ Φ tr 3 Pi μ- 3 J s rt 3 rr rt Φ μ- tr Φ
CD li li Φ CD 0) μ- tr μ- 0 tr 01 3 ØJ 3 LQ Pi φ μ- Hi tr ω O O a CD ØJ μ-
01 Φ Φ - 3 0 01 0 3 Φ ii rt 3 Ω Φ OJ ω Ω μ- 0 0 Φ rt Hi Hi . μ- LQ CD 3 i-1
3 Pi tr 3 IP 3 rt μ- LQ l-1 3 3 μ- LQ 3 ri ii Φ 3 Φ rt t
Pi Φ μ- μ- Φ φ 3 0 3 0 φ Φ N Pi 01 Hi OJ 3 TJ 3 rt rt CD rt H μ-
Ω 3 iQ 0 CQ tr 0 ** Ω Φ rt Φ 3 OJ μ- 1 CD 3 μ- 3 Hi 0J li Ω tr tr Φ TJ 3 3
0 rt CD OJ TJ μ- Hi 0J P. tr ii tr Ω Pi Ω P. o rt φ rt Φ μ- O ϋ J Ω μ-
H μ- Φ 3 0 Φ 3 3 3 μ- Φ 0J φ ØJ ØJ CQ ri μ- 0_ 3 CD Hi TJ rt rt
0 Pi LQ Ω Pi rt CD ØJ 11 01 ri ^ 01 3 5 O Φ ri tr TJ 3 μ-
ØJ 3 0 • μ- μ- Φ tr μ- tr rt rj_ Φ Ω μ- Pi μ- 2. 3 3 Φ 0J μ- TJ ØJ li ri øj
HI Hi Ω 01 Φ 3 φ μ- TJ 3 μ- 3 LQ ^ OJ 1 rt . Ω CD tr ^ O Φ
HI 0 OJ 5 LQ 3 0J ø) OJ CD Pi tr ri 0J 3 TJ Hi rt 01 rt CD μ- H ft μ- 3 iQ 11 TJ φ ω μ- Φ 3 tr i μ- O μ- φ >* LQ μ- Φ li
3 rr K . ØJ 3 rt CQ rt Hi TJ Pi • tr Ω μ- li 3 ri Φ CQ μ- O Φ μ- OJ Φ OJ Hi tr φ Ω 3 3 μ- Φ 3 s μ- <; μ- J 3 3 Ω rt 3 Ω Hi CQ Φ Φ Pi 0 Φ CD μ- rr rt TJ σ a . ^ Φ øj tr Ω ØJ O
^ rt 3 rt li 0 Ω H μ- Ω Φ i 0 rt μ- μ- μ- 1 3 ø) 01 rt LQ μ- Φ μ- ØJ rt 01 rt TJ 0 ØJ li tr1 ri 0 3 3 LQ J rt Ω LQ 3 TJ tr 3
Hi tr rt <! IP μ- μ- 3 0 μ- 3 tr φ μ- CD ri Pi LQ 0J μ- μ- O Φ H, CΛ φ μ-
0 o tr μ- 3 3 0 rt Φ 3 μ- LQ ^ CD Φ 3 O 3 01 Φ 3 μ- rt n Pi 0 rt Φ Ω 3 μ- X s- φ 3 01 - li rt Pi μ- 3 ri Ω 11 μ- μ- *< 3 μj LQ μ- • rt μ- » 3 0J CD 0 i μ- CD Φ tr O
ØJ Φ CD rt 3 3 φ CQ tr rt rt Pi 3 Φ tr øj μ- 3 0 Λ J 3
CD 0 CD Φ 3 TJ φ tr 0 OJ 5 Pi 3 0J 0 μ- 3 3 tr 3 μ- rt Hi Φ CD -y ^ 3 N i 3 3 Ω Φ £. tr μ- O
0) tr CQ μ- ØJ 0 rt rt 01 ^ Φ Φ Pi h-1 tr Φ CD μ- Hi r OJ ft ØJ 0 ^ ri tr tr 3 3 Φ rt 3 3 μ- μ- ~
LQ < rr rt ri Φ Φ rt Pi φ ii tr CD Ω 3 rt ft
Φ μ- Φ μ- rr CD CQ 0 Φ Φ tr CD Φ rr rt 3 3 φ 5 α Φ φ iQ LQ φ
antigen. Techniques for genetically engineering antibodies are well known in the art and typically involve manipulating nucleic acid encoding the immunoglobulin variable region, or the complementarity determining regions (CDRs) of an antibody.
As used herein, the term "antibody" should be construed as covering antibody fragments, derivatives, functional equivalents and homologues of antibodies, including any polypeptide comprising an immunoglobulin binding domain, whether natural or synthetic. Chimeric molecules comprising an immunoglobulin binding domain, or equivalent, fused to another polypeptide are also included. The cloning and expression of chimeric antibodies are described in EP 0120694A and EP 0125023A.
It has been shown that fragments of a whole antibody can perform the function of binding antigens. Examples of binding fragments are (i) the Fab fragment consisting of VL, VH, CL and CHI domains; (ii) the Fd fragment consisting of the VH and CHI domains; (iii) the Fv fragment consisting of the VL and VH domains of a single antibody; (iv) the dAb fragment (Ward et al . , Nature 341:544-546, 1989) which consists of a VH domain; (v) isolated CDR regions; (vi) F(ab')2 fragments, a bivalent fragment comprising two linked Fab fragments (vii) single chain Fv molecules (scFv) , wherein a VH domain and a VL domain are linked by a peptide linker which allows the two domains to associate to form an antigen binding site (Bird et al . , Science, 242:423-426, 1988; Huston et al . , PNAS USA, 85:5879-5883, 1988); (viii) bispecific single chain Fv dimers (PCT/US92/09965) and (ix) "diabodies" , multivalent or multispecific fragments constructed by gene fusion (WO 94/13804; Holliger et al . , PNAS Sci . USA, 90:6444-6448, 1993) .
In step (a) of the above methods, the library of
potential ligand binders is constructed using techniques well known in the art employing phage or phagemid/helper phage. In the latter case, a phagemid vector can be used to transform bacteria having a second type of infection mediating structure different from the first with a library of candidate ligand binders, e.g. by ligating antibody fragments obtained by PCR amplification of cDNA derived from antibody producing cells into the phagemid vector. A helper phage can then be used which is restricted to infecting bacteria having the second type of infection mediating structure to infect the bacteria transformed with the candidate ligand binders. The helper phage may be so restricted as it is specific for that type of bacteria (e.g. the IKe helper phage is specific for N-pili carrying bacteria) , or is chimeric and has had its infectivity altered (e.g. the M13K07 ch helper phage, engineered to be specific for N-pili carrying bacteria) . Alternatively, the bacteria can be transformed with nucleic acid encoding the library of ligand binders using a phage. In either case, this will result in the production of a library of phage particles displaying the ligand binder fragments as fusions with the coat protein, the phage particles being capable of infecting via the second type of infection mediating structure, but restricted so as to be substantially incapable of infecting via the first type of infection mediating structure. By way of example, where a F-pili bacteria is used later in the method, a N-pili restricted phage is used in this step such as an IKe phage or a p3 chimeric helper phage based on M13K07. Thus, in this example, the phage particles displaying the ligand binders will be unable to infect bacteria displaying F- pili. Of course, the methods described above can be readily adapted by the skilled person to employ other types of infection mediating structure, pili and/or bacteria.
In step (b) , the phage particles displaying the ligand binders are exposed to infection mediating complexes comprising a target ligand and a component capable of mediating infection via the first type of infection mediating structure. In this step, more that one type of infection mediating complex can be simultaneously or sequentially exposed to the library of ligand binders displayed on the phage particles. The component is preferably all or a part of a phage coat protein capable of specifically mediating infection of bacteria via the first type of infection mediating structure, e.g. F-pili where the initial phage used to construct the library of phage particles was N-restricted phage, or N-pili where the initial phage used to construct the library of phage particles was F-pili restricted phage.
In embodiments in which the target ligand is a peptide or protein, the infection mediating complex can be produced recombinantly, e.g. by being expressed as a fusion. Alternatively, methods for synthetically producing the component capable of mediating infection and/or the target ligand can be used. Techniques for coupling synthetically produced target ligands and components for mediating infection are well known in the art. The synthetic approach allows the construction of infection mediating complexes which are non-peptidyl, e.g. nucleic acid or carbohydrate ligands which can be used to screen phage displayed transcription factors. The synthesis of peptidyl and non-peptidyl ligands is well known in the art, see for example J.M. Stewart and J.D. Young, Solid Phase Peptide Synthesis, 2nd edition, Pierce Chemical Company, Rockford, Illinois (1984) , in M. Bodanzsky and A. Bodanzsky, The Practice of Peptide Synthesis, Springer Verlag, New York (1984) ; and Applied Biosystems 430A Users Manual, ABI Inc., Foster City, California).
Recombinant expression of peptides is also well known, and protocols for this are contained in standard texts
J OJ to to in o in o in in
0)
CD
Φ
Φ
Ω rt μ-
0
3
3 ø) ii
?r
Φ ii
^__.
01
3 rt μ- tr μ-
0 rt μ-
Ω ri
Φ
CD μ-
CD rt
ØJ
3
Ω
Φ
Φ rt
Ω
^ —
.
H tr
Φ
nucleic acid can optionally also be multiplied in the bacteria to further improve the amplification relative against background. It is also possible to recover or salvage the nucleic acid from the bacteria, e.g. by lysis or use of a helper phage . Techniques and vectors for carrying out these steps are provided in Molecular Cloning: a Laboratory Manual: 2nd edition, Sambrook et al . , 1989, Cold Spring Harbor Laboratory Press. Many known techniques and protocols for manipulation of nucleic acid, for example in preparation of nucleic acid constructs, mutagenesis, sequencing, introduction of nucleic acid into cells and gene expression, and analysis of proteins, are described in detail in Current Protocols in Molecular Biology, Ausubel et al . eds . , John Wiley & Sons, 1992.
It will be apparent that some or all of the steps set out above can be repeated to further enrich phage carrying ligand binders having increasingly high specificity for the target ligand. Thus, by way of example, phage stock can be prepared from the bacteria having the first type of infection mediating structure (e.g. a first type of pili) and containing the nucleic acid encoding the ligand binders selected in the first round of selection by infecting the bacteria with phage or helper phage restricted to that first type of bacteria. This in turn leads to the production of phage particles displaying the ligand binder which are restricted so that they can infect the first type of bacteria. A further infection mediating complex comprising a target ligand and component capable of mediating infection other than by the first type of infection mediating structure (e.g. the second type of pili or via another structure) can be used, so that the specific interaction of the ligand binder and ligand causes the phage to become infective towards first or a further type of bacteria. It will be clear that these steps can be repeated as required,
taking phage particles restricted so that they are incapable of infecting bacteria, and using an infection mediating complex to make those phage in which the displayed ligand binder binds to ligand of the complex infective towards the bacteria.
In a preferred embodiment, the bacteria used to provide the selectivity of infection by the complex formed between the phage particles and the infection mediated complex are N- or F-pili bacteria. For example, a library of ligand binders, e.g. contained in M13 filamentous phages, cannot significantly infect bacteria that express N-pili and which are devoid of F-pili. Thus, phage particles displaying ligand binders, e.g. antibodies specific for a certain antigen, can be given the ability to infect such N-pili carrying bacteria if they bind to an infection mediating complex comprising a ligand capable of binding the ligand binder and a component capable of mediating infection via N-pili, such as a part of a coat protein that can interact with N- pili, e.g. the N-terminal part of p3 from N-pili restricted phage IKe.
The method of the invention will now be described with reference to figures 1 to 4. In figures 1 and 2, embodiments of the invention employing a phagemid vector system is shown involving the following steps.
A. An N-pili restricted phage is used, for example an IKe phage or a p3 chimeric helper phage based on M13K07. Such a chimeric helper phage can be constructed by replacing the F-pili binding domain of M13 p3 with the N- pili binding domain from phage IKe p3.
B. N-pili bacteria are transformed with a library of antibody fragments as ligands, obtained by ligation of PCR amplified cDNA derived from antibody producing cells
or another convenient source into a phagemid vector.
C. The N-pili bacteria are infected with the M13 based p3 chimeric helper phage stock.
D. An F-pili binding p3 -ligand fusion protein (infection mediating complex) is constructed, composed of
M13, Nl and N2 domains linked to a polypeptide or glycoprotein antigen.
E. The fusion protein is mixed with phage originating from the transformed N-pili bacteria, and hence displaying the antibody library, and with F-pili bacteria.
F. F-pili bacteria that have been infected with chimeric phage are selected for by a selection marker (e.g. ampicillin resistance) on the plasmid used.
G. A phage stock is prepared from the selected F-pili bacteria, using a F-pili dependent helper phage (e.g. M13K07) . This phage stock contains phage with the genetic information for the selected antibodies and displays the antibodies at their surfaces.
H. An N-pili binding p3 -ligand fusion protein is constructed, composed of the N-terminal domain of IKe phage p3 linked to the polypeptide or glycoprotein antigen.
I . The N-pili binding fusion protein is mixed with phage originating from the infected F-pili bacteria, and with N-pili bacteria.
J. N-pili bacteria that have been infected with chimeric phage are selected for by a selection marker on the plasmid used.
K. A phage stock is prepared from the selected N-pili bacteria, using a N-pili dependent helper phage (e.g. M13K07 ch) . This phage stock contains phage with the genetic information for the selected antibodies and displays the antibodies at their surfaces.
Steps D-K are optionally repeated in order to further enrich phage carrying the ligand binders, i.e. antibodies, of interest. The plasmids harbouring the genetic information encoding these antibodies are salvaged. The coding sequence can be then be sequenced to determine the primary structure of the selected antibodies, and/or used for expression of the selected antibodies .
Thus, figure 1 shows an example using a phagemid system for selecting antibodies that bind to an antigen. In this example, N-pili carrying bacteria are transformed with phagemid vectors comprising nucleic acid encoding a library of ligand binders and a chimeric helper phage
(M13K07) . This leads to the production of a library of phages displaying the antibodies as M13 fusions, the phage being capable of infecting N-pili carrying bacteria via the chimeric IKe coat protein, but restricted not to infect F-pili carrying bacteria.
In the first round of selection, the phages are mixed with an infection mediating complex comprising one or more antigens (target ligands) fused to p3 of M13. Thus, the binding of antibody to antigen causes the phage to become infective towards F-pili carrying bacteria.
This process is repeated in the second round of selection using a different M13 helper phage, leading to the production of phage displaying the antibodies selected in the first round of selection and that are restricted to infect N-pili carrying bacteria. These two rounds of
selection can then be repeated as necessary
The embodiment of the invention shown in figure 2 differs from that in figure 1 in that in the first round of infection an IKe helper phage is used instead of the M13K07 chimeric helper phage.
Figure 3 shows an embodiment employing a phage vector to produce the library of phage particles displaying the antibodies, as opposed to the phagemid vector and helper phage systems shown in figures 1 and 2. This system differs from the helper phage embodiments shown in figures 1 and 2 in that each round of selection employs the bacteria displaying the same type of infection mediating structure (N-pili) .
Example: Production of a infection mediating complex and selection of antigen specific F-pili restricted phage using IKe p3-fusion protein The ability of an IKe p3 -fusion protein to mediate infection of F-pili restricted R408 phage displaying antibody fragments into N-pili expressing E. coli was investigated.
The fusion protein, LIKebAD2, containing IKe p3 receptor and penetration domain and the AD2 peptide (from cytomegalovirus) was constructed. For the AD2 epitope two complementary oligonucleotides were used containing a 5'KpnI site and a 3 'EcoRI site with the sequence 5'- CCCCGGTACCGCCAACGAGACTATCTACAACACTACCCTCAAGTATGGAGATTGAAT TCCCCC-3'. The two oligonucleotides were allowed to hybridize, digested with Kpnl and EcoRI, gelpurified and subsequently ligated into Kpnl- and EcoRI-digested pUC19 expression vector, resulting in the intermediate vector pAD2.
The LIKep3b fragment comprising the genetic information
for the receptor and penetration domain of IKe p3 was isolated from PCR-amplified dsIKe phage DNA using the primers :
LIKe5' (Hindlll) : 5' -CCAAGCTTCAGCTAAGGCGTAATTATGAAAAGAAAAATAATAGCA-3' and LIKeb3' ( Kpnl ) : 5' -ATCTTCCTTTGTCAGTGAGGTACCAGTTGATC-3' .
The fragment was digested with Kpnl and Hindlll, gelpurified and ligated into Kpnl- and Hindlll -digested pAD2 vector, resulting in the expression vector pLIKebAD2.
The pLIKebAD2 vector was transformed into E. coli cells and the transformed cells expressed the fusion protein after induction with IPTG. The expressed protein was verified on Western blot using anti-AD2 antibodies. The fusion protein was purified by ion exchange chromatography and the concentration determined by competitive ELISA to be lxlO"5 M. The fusion protein was also shown to inhibit IKe phage infection of N-pili expressing cells in a dose-dependent manner (figure 4) . The fusion protein did not inhibit infection of M13 phage of F-pili expressing bacteria, and a corresponding fusion protein comprising the AD2 epitope and the Nl and N2 domains of M13 p3 did not inhibit infection of IKe phage into N-pili expressing bacteria.
Using purified LIKebAD2 fusion protein antigen specific selection was performed from a human scFv R408 phage display library (2xl09 members) having ampicillin as a selection marker. 10 μl LIKebAD2 fusion protein diluted to 2xl0"7 M was mixed with an equal volume of phage stock (1013 cfμ/ml) and incubated overnight at +4°C. A negative control was made using PBS instead of fusion protein.
100 μl N-pili expressing E. coli cells in log phase were added to each mixture and the samples were incubated 1 h
at 37°C, 75 rpm in 5 ml round bottom tubes. Each sample was diluted with LB medium to a final volume of 800 μl, plated onto two LA plates (diameter 15 cm) containing 100 μg/mL ampicillin and 1% glucose, and incubated overnight at 37°C. The resulting colonies were transferred to 20 ml LB medium containing ampicillin and glucose as above, and incubated at 37°C, 250 rpm for 3 h and then transferred to a 1 1 flask containing 200 ml of the same medium and incubated for an additional 3 h. The cells were used for preparation of phagemide DNA, which was then electroporated into electrocompetent ToplOF' E. coli cells. The transformed cells were grown to log phase in selective medium and infected with R408 helper phages, induced with IPTG and grown for 16 h at 25°C. The phage containing supernatant was PEG-precipitated, redissolved in 100 μl of PBS, and the titer determined. The phage stock from the first selection with LIKebAD2 was used for a second round of selection, using the same parameters as for the first selection. Phage stocks were made from selected positive and negative bacterial clones, and analysed on antigen-specific ELISA. The results demonstrated that presence of the fusion protein in the reaction mixture allowed rescue of antigen specific phage from the library (Table 1) . Already after the first round of selection was a phage stock obtained that bound more efficiently to the antigen coated wells than to control antigen (BSA) . After a further selection round clones were prepared and such clonal phage stocks bound to antigen with an appreciable OD reading of 0.374, whereas they only bound to the control antigen BSA at low background levels as examplified in Table 1. In contrast, unspecific phage originating from the library bound to the antigen at lower levels. The OD value obtained for the specific clone was similar to that generated by a control phage (pI3) displaying antibodies with a known specificity for the AD2 epitope of cytomegalovirus .
Table 1. AD2 specificity of selected phage as determined by ELISA
References :
The references described herein are expressly incorporated by reference.
1. Barbas, C. et al . (1991), Proc . Natl . Acad. Sci . U.S.A., 88, 7978-7982.
2. Beck, E. and Zink, B. (1981), Gene, 16, 35-38.
3. Bradley, D.E. (1980), Plasmid, 4, 155-169.
4. Bradley, D.E. et al . , (1980), J. Bacteriol . , 143, 1466-1470. 5. Bross, P. et al . (1988), J. Gen. Microbiol . , 134, 461- 471.
6. Chiswell, D.J. and McCafferty, J. (1992), Trends Biotechnol., 10, 80-84.
7. Clackson, T. et al . (1991), Nature (London), 352, 624- 628.
8. Duenas, M. and Borrebaeck, C.A.K. (1994), Bio/Technology, 12, 999-1002.
9. Duenas, M. et al . (1996), Mol . Immunol., 33, 279-285.
10. Endemann, H. et al . (1992), Mol. Microbiol., 6, 471- 478.
11. Endemann, H. et al . (1993), J. Virol., 67, 3332-3337.
12. Hoogenboom, H.R. and Winter, G. (1992), J. Mol. Biol . , 227, 381-388.
13. Kay, B.K. (1995), Perspect . Drug Discovery Des . , 2, 251-268.
14. Khatoon, H. et al . , (1972), Virology, 48, 145-155.
15. Krebber, C. et al . (1995), FEBS Letters, 377, 227-231.
16. Marks, J.D. et al . (1991), J. Mol. Biol., 222, 581- 597. 17. Marzari, R. et al . (1997), Gene, 185, 27-33.
18. McCafferty, J. et al . (1990), Nature (London), 348, 552-554.
19. Smith, G.P. (1985), Science, 228, 1315-1317.
20. Stengele, I. et al . (1990), J. Mol. Biol., 212, 143- 149.