301
construct with a linker that is elongated by three residues. Currently,
only two D21V variants with distinctly light-repressed activity
could be identified, and a more detailed analysis of the underlying
linker lengths is therefore precluded. Exhaustive sampling of linker
variants in the D21V context by PATCHY might yield additional
light-repressed variants and allow this type of analysis.
For the H22P background, only hybrid variants with lightinduced activity as the original YF1 H22P were found but none
with light-repressed activity (Fig. 5c). Consistent with a coiledcoil conformation, the linkers of these variants predominantly
comprised 7n residues. A recent EPR investigation [19] revealed
that introduction of the H22P exchange leads to a complete rearrangement of the dimer interface; two N-terminal α helices that
in YF1 are tucked in between the LOV photosensor domains (see
Fig. 1) are displaced and probably unfolded in H22P. Given this
drastic change of the dimer interface, it is perplexing that the
H22P variant transduces signals, let alone in inverted manner.
Despite these differences, signal transduction in H22P apparently
employs a highly similar structural mode as YF1, albeit with
inverted sign [19]. In support of this view, we now find that the
linkers of variants with signal response corresponding to that of
YF1 H22P mostly conformed to 7n residues. By that token, one
would also expect to find H22P linker variants with inverted, i.e.,
light- repressed activity. Our inability to do so may stem from
insufficient sampling of linker sequence space or from an inherent
difference in the signal mechanism of H22P. The above finding
that the light response of the D21V variant could be inverted for
certain linker lengths implies that the former is true and that signal inversion by linker-length variations also applies to the H22P
context.
The PATCHY method offers an efficient route toward libraries of
hybrid genes with a single fusion between defined fragments of
two parental genes A and B. In contrast to related approaches for
the construction of hybrid gene libraries [11–15], PATCHY uses a
simpler protocol and obviates incremental nucleolytic digest of the
parental genes which is difficult to precisely adjust. Rather, PCR
amplification with sets of staggered primers exactly delineates
which fragments of A and B are generated and recombined. The
analysis of PATCHY hybrid libraries benefits from efficient activity
assays, in particular cell-based screening and selection approaches.
We demonstrate the utility of PATCHY for the test case of
hybrid blue light photoreceptors. Catalytic activity and response
to light are largely governed by the length of the linker intervening the constituent photosensor and effector modules of the
receptors. A striking seven-residue dependence of receptor function on linker length is explained by the continuous coiled-coil
3.4 Conclusion
Generation of Hybrid Gene Libraries
construct with a linker that is elongated by three residues. Currently,
only two D21V variants with distinctly light-repressed activity
could be identified, and a more detailed analysis of the underlying
linker lengths is therefore precluded. Exhaustive sampling of linker
variants in the D21V context by PATCHY might yield additional
light-repressed variants and allow this type of analysis.
For the H22P background, only hybrid variants with lightinduced activity as the original YF1 H22P were found but none
with light-repressed activity (Fig. 5c). Consistent with a coiledcoil conformation, the linkers of these variants predominantly
comprised 7n residues. A recent EPR investigation [19] revealed
that introduction of the H22P exchange leads to a complete rearrangement of the dimer interface; two N-terminal α helices that
in YF1 are tucked in between the LOV photosensor domains (see
Fig. 1) are displaced and probably unfolded in H22P. Given this
drastic change of the dimer interface, it is perplexing that the
H22P variant transduces signals, let alone in inverted manner.
Despite these differences, signal transduction in H22P apparently
employs a highly similar structural mode as YF1, albeit with
inverted sign [19]. In support of this view, we now find that the
linkers of variants with signal response corresponding to that of
YF1 H22P mostly conformed to 7n residues. By that token, one
would also expect to find H22P linker variants with inverted, i.e.,
light- repressed activity. Our inability to do so may stem from
insufficient sampling of linker sequence space or from an inherent
difference in the signal mechanism of H22P. The above finding
that the light response of the D21V variant could be inverted for
certain linker lengths implies that the former is true and that signal inversion by linker-length variations also applies to the H22P
context.
The PATCHY method offers an efficient route toward libraries of
hybrid genes with a single fusion between defined fragments of
two parental genes A and B. In contrast to related approaches for
the construction of hybrid gene libraries [11–15], PATCHY uses a
simpler protocol and obviates incremental nucleolytic digest of the
parental genes which is difficult to precisely adjust. Rather, PCR
amplification with sets of staggered primers exactly delineates
which fragments of A and B are generated and recombined. The
analysis of PATCHY hybrid libraries benefits from efficient activity
assays, in particular cell-based screening and selection approaches.
We demonstrate the utility of PATCHY for the test case of
hybrid blue light photoreceptors. Catalytic activity and response
to light are largely governed by the length of the linker intervening the constituent photosensor and effector modules of the
receptors. A striking seven-residue dependence of receptor function on linker length is explained by the continuous coiled-coil
3.4 Conclusion
Generation of Hybrid Gene Libraries