FIG 3.

Identification of the T. gondii gene GRA28 as an inducer of CCL22 in human cells. (A) Gene expression correlations across 396 T. gondii expression microarrays between TgMYR1 and 4 additional “bait” genes (top) and 5 candidate CCL22-inducing effectors (bottom). dN/dS ratios are also shown to illustrate the high level of positive selection acting on this class of genes. (B) Effect of deleting 5 candidate genes on CCL22 secretion in THP-1 cells, showing that ΔGRA28 parasites induced significantly less CCL22 than wild-type controls (>100-fold reduction; ***, P < 0.0001). PRUΔToxofilin parasites also induced significantly lower levels of CCL22 in THP-1 cells (1.4-fold reduction; **, P < 0.01). Each blue dot indicates a genetically distinct knockout clone. (C) ΔGRA28 parasite clones also induce significantly less CCL22 from primary human second trimester placental villous explants. (D) MA plot (where the log₂ fold change is on the y axis and the log-transformed average of normalized counts is on the x axis) of RNA-seq analysis performed on THP-1 cells infected with WT or ΔGRA28 T. gondii (RH strain). CCL22 and the chemokines CXL1 and CXL2 were the most highly GRA28-dependent transcripts, while a handful (n = 64) of other genes had significantly higher transcript abundance in WT parasites than in ΔGRA28 parasites (P_adj < 0.001; log₂FC > 2; blue symbols). CCL17 (red arrow, green symbol), a chemokine that is typically coregulated with CCL22, did not show any evidence of being induced by GRA28.