In the genome of Rigosertib molecular weight T. castaneum, two

Dif and one Relish orthologs have been identified [39]. Therefore, in T. castaneum, heterodimers of NF-κB, such as Dif1-Relish or Dif2-Relish, may form in vivo, providing the possible crossing points for the Toll and IMD signaling pathways. Assuming the possibility of crosstalk as described above, we can somewhat explain how the promiscuous activation and usage of signaling pathways that were suggested in this study occur. Most of AMP genes tested in this study were induced by Ec, Ml, Sc, Ecl and Bs. This phenomenon could be explained by any of the three crosstalk hypotheses. PGRP-SA and PGRP-LC may sense both gram-positive and gram-negative bacteria. Sc might be sensed by both GNBP3 and PGRP-LC Bcl-2 apoptosis pathway and signals flowing through the IMD pathway may branch to the Toll pathway via FADD. More promiscuous and frequent heterodimerization among Relish proteins and Dif/Dorsal proteins may result in more complex induction profiles of AMP genes than in Drosophila. For example, when we assume that Tribolium PGRP-SA can recognize both Ec and Ml as mentioned above, the MyD88 knockdown would lead to repressed levels of Cec2 induction

by both Ec and Ml, as shown in this study. The induction of group I genes Att1, Col1 and Def2 by Ec or Ml was suppressed by IMD knockdown. Similarly, this may be explained by hypothesizing that Tribolium PGRP-LC can recognize both Ec and Ml. A phenomenon we observed and should note is that induction levels of some AMP genes by Ml were even elevated after the knockdown of the Toll pathway component MyD88, typically seen in the cases of Def3 and Col1 at 24 h post Ml challenge. This may also be attributed to crosstalk, especially at the levels of transcription factors/response elements. The induction of these AMP genes

seems to be more dependent on the IMD pathways, suggesting the NF-κB-binding motifs regulating the transcription mafosfamide of these genes may have higher affinity to Relish than to Dif/Dorsal. In addition, we hypothesize the signals elicited by Ml is transduced more preferentially by the Toll pathway, but the IMD pathway is also involved. We also hypothesize that these genes are more potently activated by Relish than by Dif/Dorsal. MyD88 knockdown can reasonably reduce the amounts of activated Dif/Dorsal proteins while additional signal-flow via the IMD pathway allows the accumulation of activated form of Relish proteins with time. Under these artificial conditions, accumulating activated Relish can compete for binding to the NF-κB motifs with reduced numbers of activated Dif/Dorsal and eventually overcome Dif/Dorsal to occupy the binding sites. This may lead to elevated transcription of these AMP genes than in the controls, because we postulate Relish is more potent than Dif/Dorsal in terms of transactivation of these genes. Heterodimerization of these transcription factors may also be involved.