Transforming TpoR TMD mutations are few in number and are highly restricted by position and substituting amino acid type. (A) Sequence-function map showing the results of saturation mutagenesis of human TpoR TMD (residues 488-516) in Ba/F3 cells. WT sequence and amino acid number in the mature polypeptide are shown to the right and left of the heat map, respectively. Each square reflects the combined data from all synonymous variants at that position. Color of each square indicates the log ratio enrichment score for variant frequencies in Ba/F3 cells cultured for 48 hours without cytokines compared with the unselected library, according to the scale below the map. Scores are relative to WT (white squares with dots; score of 0). Red shades indicate enrichment. Blue shades indicate depletion. Diagonal lines are proportional to standard deviations of scores across 6 replicate selections. TMD helix model shows the predicted locations of sites where mutations caused constitutive activity, colored by approximate enrichment score. (B) Selected variants from (A) were transduced individually into Ba/F3 cells and tested for growth in IL-3, Tpo (100 ng/ml), or no cytokine (RPMI-1640). Data are plotted as fold-change over starting cell density (dotted line) after 48 hours. Individual data points represent the mean of technical triplicates from each of 3 independent experiments. Error bars indicate standard deviation. No statistically significant differences were observed in IL-3 or Tpo; significance values are indicated for all variants compared with WT in no cytokine (1-way analysis of variance [ANOVA]: ****P < .0001). (C) Western blot analysis of STAT5 and ERK1/2 activation status. Ba/F3 cells expressing each variant were serum/cytokine starved for 7 hours prior to treatment with 50 ng/mL recombinant human Tpo (hTpo; WT only) or without cytokines for 30 minutes. Lysates were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotted with antibodies against the indicated targets. *Nonspecific background band in phosphorylated ERK 1/2 (pERK1/2) immunoblots; #glycosylation variants of hemagglutinin (HA)-tagged human TpoR (hTpoR) protein. (D) Densitometric analysis of phosphorylated STAT5 (pSTAT5) and pERK1/2 signal from 3 independent experiments performed as shown in (C). All immunoblots are shown in supplemental Figure 2. Signal intensities are reported as fold-increase over WT with no cytokine treatment (indicated by dotted line at 1). Significance values are indicated for all variants compared with the inactive V501G variant for pSTAT5 or pERK1/2 (1-way ANOVA: ****P < .0001; ***P < .001; *P < .05). (E) Plot shows correlation between log ratio enrichment scores in the DMS screen (A) and cell counts from independent growth experiments (B). Error bars indicate the standard deviations of enrichment scores calculated in Enrich2 (y-axis) and standard deviation of final cell counts in the experiment from panel (B) (x-axis). ns, not significant.