Functional characteristics of a highly specific integrase encoded by an LTR-retrotransposon

Abstract

Background: The retroviral Integrase protein catalyzes the insertion of linear viral DNA into host cell DNA. Although different retroviruses have been shown to target distinctive chromosomal regions, few of them display a site-specific integration. ZAM, a retroelement from Drosophila melanogaster very similar in structure and replication cycle to mammalian retroviruses is highly site-specific. Indeed, ZAM copies target the genomic 5'-CGCGCg-3' consensus-sequences. To enlighten the determinants of this high integration specificity, we investigated the functional properties of its integrase protein denoted ZAM-IN.

Principal findings: Here we show that ZAM-IN displays the property to nick DNA molecules in vitro. This endonuclease activity targets specific sequences that are present in a 388 bp fragment taken from the white locus and known to be a genomic ZAM integration site in vivo. Furthermore, ZAM-IN displays the unusual property to directly bind specific genomic DNA sequences. Two specific and independent sites are recognized within the 388 bp fragment of the white locus: the CGCGCg sequence and a closely apposed site different in sequence.

Conclusion: This study strongly argues that the intrinsic properties of ZAM-IN, ie its binding properties and its endonuclease activity, play an important part in ZAM integration specificity. Its ability to select two binding sites and to nick the DNA molecule reminds the strategy used by some site-specific recombination enzymes and forms the basis for site-specific integration strategies potentially useful in a broad range of genetic engineering applications.

Figure 1. The endonuclease activity of ZAM integrase correlates with the presence of a 388 bp fragment from the white locus.

A) Schematic representation of the ZAM integrase “IN” and a carboxy-terminal deleted integrase “ΔIN” used in the in vitro DNA binding assay. The three main domains: the zing finger “HHCC”, the catalytic domain “DDE” and a predictive DNA binding domain “BD” are represented. Nucleotide numbers according to ZAM sequence are indicated below. The hatched box indicated the region deleted to generate the ΔIN protein. The full length and the truncated integrases were expressed in bacteria as GST fusion proteins and fixed on agarose beads. IN and ΔIN purified proteins were analysed on SDS-PAGE electrophoresis followed by coomassie staining (right panel). The molecular masses of proteins are indicated in kilodalton. B) Circular representation of the 2686 bp pUC18 plasmid. Palindromic sequences CGCGCG present in pUC18 are indicated. The Drosophila genomic locus known to be the target of ZAM integration and located 3 kb upstream of the white gene is presented above. The white fragment (from positions 4278 to 4666 according to the drosophila sequence) was cloned in the pUC18 plasmid and is represented by the grey box. The black dot at position 4314 indicates the CGCGCG integration site of ZAM. C) In vitro endonuclease activity of ZAM integrase: pUC, pUC/white and pUC/white1mut plasmids were incubated without (−IN) or with (+IN) purified ZAM-IN. Positions of the supercoiled, nicked (circle) and linear (bar) DNAs are indicated.

Figure 2. LTR binding property of ZAM integrase.

A) ZAM LTR fragments. The grey box represents the full length LTR of ZAM and the solid bars represent the KpnI “K” and HindIII “H” restriction sites at the position 21 and 82, respectively. A full length LTR and truncated PCR product deleted of the first 40 bp of ZAM LTR called “ΔLTR” are presented below. A double stranded oligonucleotide spanning from position 28 to 40 and called “BS” was also used in these experiments. B) In vitro DNA binding assays with ZAM-IN on the LTR fragments. Left panel: The full length “LTR” and the two truncated LTR fragments digested by KpnI “K” or HindIII “H” were tested as indicated above each lane. Middle panel: the full length LTR and a truncated PCR product “ΔLTR” were used in these assays. Right panel: In vitro binding assays with ΔIN on ZAM LTR fragments: the full length “LTR” and the two truncated LTR fragments digested by KpnI “K” or HindIII “H” were tested as indicated above each lane. C) In vitro DNA binding assays performed with a double stranded oligonucleotide from base 28 to 40 according to ZAM sequence. The 13 bp fragment is retained by ZAM-IN. DNA fragments sizes are indicated as “L”. DNA fragments sizes are indicated for each panel. In B and C, bound and unbound fractions are presented in upper and lower panels respectively. The percentage of bound and unbound fractions is presented below each panel in B. In C, 100% of the 274 bp fragment was recovered in the unbound fraction whereas 100% of the 66 and 48 bp fragments were recovered in the bound fraction.

Figure 3. ZAM Integrase interacts with specific genomic DNA sequences.

A) Diagram of the white DNA fragment from nucleotide positions 4278 to 4666. Two PCRs products used in this experiment and called “w4278” and “w4392” are represented underneath. The two AluI restriction sites and the resulting DNA fragment sizes are presented above. The palindromic cleavage site CGCGCG is indicated by a white box. B) In vitro binding assays with ZAM-IN protein performed on the white PCR products “w4278” and “w4392”. C) Assays performed with the white PCR product w4278 digested by AluI. The percentage of bound (upper panels) and unbound (lower panels) fractions is presented below each panel. DNA fragments sizes are indicated for each panel.

Figure 4. ZAM integrase binds two specific genomic DNA sites.

A) Sequence of the Drosophila white locus from base 4278 to 4326. The oligonucleotides w0, w1, w1mut, w2, w3, Δw3 and w3mut used in the experiments are represented under the sequence. The integration site CGCGCG, the sequence of the mutated integration site of the w1mut oligonucleotide and the nucleotides mutated to generate the w3mut oligonucleotide are indicated by boxes. B) Left panel: In vitro binding assays were performed with ZAM integrase “IN” and the double stranded oligonucleotides w0, w1, w1mut, w2, w3, Δw3 and w3mut. w1 and w3 are the only two oligonucleotides retained by ZAM integrase. Right panel: In vitro endonuclease activity of ZAM integrase: pUC/white and pUC/white3mut plasmids were incubated without (−IN) or with (+IN) purified ZAM-IN. Positions of the supercoiled, nicked (circle) and linear (bar) DNAs are indicated. C) Alignment of a conserved motif detected in the ZAM LTR and w3. The first 60 nucleotides of the LTR sequence are presented as the upper sequence. The binding site of w3 is presented below.