• Figure 3.
    Download figureOpen in new tabFigure 3. Schematic of HGT events.

    Bayesian phylogeny based on the 16S rRNA gene from selected taxa is shown. Colored lines indicate putative horizontal gene transfer events, although other possible HGT patterns cannot be definitively excluded. Posterior probabilities are noted at each node.

    DOI: http://dx.doi.org/10.7554/eLife.04266.011

    Figure 4.
    Download figureOpen in new tabFigure 4. Comparison of GH25 muramidase and rRNA divergence.

    (A) Unrooted Bayesian phylogeny of the GH25 muramidase from A. boonei and selected relatives, based on an alignment of 185aa without indels. Taxon of origin for each nucleic acid sequence is indicated by color. Posterior probability is indicated at all nodes with values above 50. Branch lengths represent number of substitutions per site as indicated by scale bar. (B) Unrooted Bayesian phylogeny of the 16S rRNA gene for the same taxa as in (A), based on an alignment of 1,156 bp without indels.

    DOI: http://dx.doi.org/10.7554/eLife.04266.012

    Figure 5.
    Download figureOpen in new tabFigure 5. Conservation of A. boonei GH25 muramidase domain.

    (A) Consensus alignment of 86 GH25 muramidases with insertions and deletions removed. Conservation is indicated by amino acid symbol size and bar graphs below the consensus sequence. Active site residues and highly conserved amino acids modeled below are indicated with red and orange asterisks, respectively. (B) Space-filling model of the active site face of the predicted structure of A. boonei GH25 muramidase domain and (C) S. moellendorffii GH25 muramidase domain. Active site residues are indicated in red and the eight additional residues most highly conserved across all 86 proteins are orange.

    DOI: http://dx.doi.org/10.7554/eLife.04266.013

    Figure 7.
    Download figureOpen in new tabFigure 7. E. coli death following full-length A. boonei lysozyme expression.

    (A) Live/dead stain of BL21 (DE3) E. coli transformed with expression constructs for the full-length lysozyme from A. boonei or a control lysozyme WORiA, a bacteriophage infecting Wolbachia pipientis strain wRi, after overnight growth without induction. PAGE gels of crude E. coli lysates from E. coli expressing the indicated lysozyme after 6 hr of induction are also shown with the expected sizes of lysozymes indicated with arrows. (B) Structure of original full-length A. boonei lysozyme expression plasmid and two spontaneous knockout mutants caused by insertion of 774 bp (mutant 1) and 768 bp (mutant 2) of IS1 transposase sequences. These insertions also resulted in a number of stop codons in the reading frame of the lysozyme. Knockout mutants grew to normal colony size, while all wild type colonies had intact expression plasmids, grew poorly, and died over time in liquid culture.

    DOI: http://dx.doi.org/10.7554/eLife.04266.017

    Figure 8.
    Download figureOpen in new tabFigure 8. Lysozyme expression and relative fitness during A. boonei and M. lauensis coculture.

    (A) Expression of A. boonei GH25 muramidase relative to the control gene elongation factor 1α, after the indicated time of coculture with M. lauensis (M.l) at the specified ratio relative to A. boonei. *p < 0.05, **p < 0.01, by Mann–Whitney U pairwise comparisons. N = 6 for all samples. Primers are listed in Supplementary file 1. (B) Relative fitness of A. boonei vs M. lauensis in monoculture (N = 5) and coculture (N = 4). (C) Growth of A. boonei (red) and M. lauensis (blue) monocultures over time. Significant differences in cell abundance occur at 24, 52, and 64 hr (p < 0.05), and 56 and 60 hr (p < 0.01) based on pairwise Wilcoxon tests. (D) Growth of A. boonei and M. lauensis in coculture over time. Significant differences in cell abundance occur at 48, 52, and 64 hr (p < 0.05) based on pairwise Wilcoxon tests. Error bars are ±SEM for all panels.

    DOI: http://dx.doi.org/10.7554/eLife.04266.018