Degeneration of a homing endonuclease and its target sequence in a wild yeast strain

Abstract

Mobile introns and inteins self-propagate by 'homing', a gene conversion process initiated by site-specific homing endonucleases. The VMA intein, which encodes the PI-SceI endonuclease in Saccharomyces cerevisiae, is present in several different yeast strains. Surprisingly, a wild wine yeast (DH1-1A) contains not only the intein(+) allele, but also an inteinless allele that has not undergone gene conversion. To elucidate how these two alleles co-exist, we characterized the endonuclease encoded by the DH1-1A intein(+) allele and the target site in the intein(-) allele. Sequence analysis reveals seven mutations in the 31 bp recognition sequence, none of which occurs at positions that are individually critical for activity. However, binding and cleavage of the sequence by PI-SceI is reduced 10-fold compared to the S.cerevisiae target. The PI-SceI analog encoded by the DH1-1A intein(+) allele contains 11 mutations at residues in the endonuclease and protein splicing domains. None affects protein splicing, but one, a R417Q substitution, accounts for most of the decrease in DNA cleavage and DNA binding activity of the DH1-1A protein. Loss of activity in the DH1-1A endonuclease and target site provides one explanation for co-existence of the intein(+) and intein(-) alleles.

Figure 1

Alignment of the actual and predicted DNA recognition sequences for VMA inteins from S.cerevisiae, DH1-1A and C.tropicalis. At the top of the figure is shown the protein sequence of the VMA1-encoded H⁺-ATPase in the region overlapping the intein insertion site. The position of the intein insertion site is indicated by an arrow. Base pairs that are shown on a red background in the S.cerevisiae intein^– sequence are critical for DNA cleavage by PI-SceI (20). The staggered line denotes the site of cleavage by PI-SceI. In the DH1-1A and C.tropicalis intein^– sequences, the base pairs shown on a green background are those that differ from the S.cerevisiae intein^– sequence.

Figure 2

Binding of the predicted endonuclease recognition sequences from S.cerevisiae, the DH1-1A intein^– allele and C.tropicalis to purified PI-SceI protein determined by EMSA. Radiolabeled 219 bp substrates containing one of the different recognition sequences was incubated with increasing concentrations of PI-SceI corresponding to 3 × 10^–10, 7 × 10^–10, 1 × 10^–9 and 3 × 10^–9 M in each set of four lanes. The samples were separated by electrophoresis in a 7% polyacrylamide gel. The migration positions of unbound 219 bp fragment DNA and the lower and upper PI-SceI–DNA complexes are indicated by UB, LC and UC, respectively.

Figure 3

Positions of DH1-1A amino acid substitutions within the PI-SceI crystal structure. (A) A ribbon diagram of the PI-SceI structure (colored purple) was generated using SPOCK (52) from coordinates deposited in the RCSB Protein Data Bank (entry 1DFA). The side chains of PI-SceI residues that are substituted in the DH1-1A protein are colored red and are labeled. The endonucleolytic active site is located at the base of the two conserved LAGLIDADG α-helices, which are colored green. (B) The same diagram as in (A) but rotated 180° about the y-axis. Not shown are the side chains for Ile132 and Ile276, which are situated in disordered regions of the crystal structure.

Figure 4

EMSA of binding of PI-SceI, the DH1-1A endonuclease and PI-SceI variants to a S.cerevisiae recognition sequence. The DNA binding activities of PI-SceI, of the DH1-1A protein and of 11 PI-SceI derivatives, each containing one of the DH1-1A substitutions (amino acid substitutions are indicated above each lane) were determined. The endonuclease protein concentration in each lane is 0.7 nM, which corresponds to the K_d for wild-type PI-SceI, and the samples were separated by electrophoresis in a 7% polyacrylamide gel. The migration positions of unbound 219 bp fragment DNA and the lower and upper PI-SceI–DNA complexes are indicated by UB, LC and UC, respectively. Similar relative binding affinities were observed using total protein concentrations of 4 × 10^–9 and 2 × 10^–9 M (data not shown).