DAXX interacts with phage PhiC31 integrase and inhibits recombination

Abstract

Phage PhiC31 integrase has potential as a means of inserting therapeutic genes into specific sites in the human genome. However, the possible interactions between PhiC31 integrase and cellular proteins have never been investigated. Using pLexA-PhiC31 integrase as bait, we screened a pB42AD-human fetal brain cDNA library for potential interacting cellular proteins. Among 61 positives isolated from 10(6) independent clones, 51 contained DAXX C-terminal fragments. The strong interaction between DAXX and PhiC31 was further confirmed by co-immunoprecipitation. Deletion analysis revealed that the fas-binding domain of DAXX is also the region for PhiC31 binding. Hybridization between a PhiC31 integrase peptide array and an HEK293 cell extract revealed that a tetramer, 451RFGK454, in the C-terminus of PhiC31 is responsible for the interaction with DAXX. This tetramer is also necessary for PhiC31 integrase activity as removal of this tetramer resulted in a complete loss of integrase activity. Co-expression of DAXX with PhiC31 integrase in a HEK293-derived PhiC31 integrase activity reporter cell line significantly reduced the PhiC31-mediated recombination rate. Knocking down DAXX with a DAXX-specific duplex RNA resulted in increased recombination efficiency. Therefore, endogenous DAXX may interact with PhiC31 causing a mild inhibition in the integration efficiency. This is the first time that PhiC31 was shown to interact with an important cellular protein and the potential effect of this interaction should be further studied.

Figure 1

Interaction between ΦC31 integrase and human DAXX. (A) Yeast two-hybrid analysis was conducted using ΦC31-integrase-derived pLexA fusion plasmids as bait, a human fetal cDNA-derived pB42AD library and two human DAXX-derived pB42AD fusion plasmids as prey. The overlapping open reading frames correspond to human DAXX. (B) Cell lysates of pEGFP-ΦC31-transfected and pEGFP-N3-transfected HEK293 cells were each incubated with a mouse anti-GFP binding protein on A/G agarose beads. Samples were analyzed by western blotting with a rabbit anti-human DAXX antibody. An unknown protein with higher molecular weight was also recognized by the anti-DAXX antibody. However, it is probably non-specific by a cross-reaction of the polyclonal antibody as it is present in all the samples. (C) Three ΦC31 integrase overlapping peptide arrays with different lengths of peptide (12, 15 and 20mers) for localization of the DAXX interacting region on ΦC31. Serial numbers of spots that were positive for DAXX binding were labeled. (D) Amino acid sequences of the peptide spots corresponding to the array shown in C. The DAXX binding tetramer is indicated by underlining.

Figure 2

Subcellular localization of DAXX and ΦC31 integrase. (A) pDAXX-DsRed and pEGFP-ΦC31 fusion protein constructs were co-transfected into HEK293 cells. (B) HEK293 cells transfected with pEGFP-ΦC31 alone. (C) HEK293 cells transfected with a ΦC31 construct with C-terminal fused EGFP (pΦC31-EGFP) alone. (D) Cells transfected with pEGFP alone. Expressed fluorescent GFP and DsRed were observed under a confocal microscope. The nuclei were stained by DAPI (40X).

Figure 3

Effects of DAXX overexpression and siRNA knock down on efficiency of ΦC31-integrase-mediated recombination in the ΦC31 activity reporter cells PB[EGFP]. The efficiency of recombination was first calibrated using the transfection efficiency in each culture and then expressed as percentages of the cultures transfected with pCMVInt only. PB[EGFP] reporter cells were transfected with A, pCMVInt, B, pCMVInt/pDsRed1, C, pCMVInt/pDAXX-DsRed1, D, pCMVInt/non-silencing siRNA, and E, pCMVInt/siDAXX. ^*P < 0.05, ^**P < 0.01.