Fig. 5

Hsf1 mediates the response to misfolded myosin. a, a’ Co-injection of unc45b-mo with the construct −445/−310(unc45b)gata2:gfp (Figure S3f in Additional file 8, construct 25) leads to an increase of GFP expression (a’) compared with embryos injected with the plasmid alone (a). b, b’ Co-injection of unc45b-mo with the construct −445/−310mut(unc45b)gata2:gfp in which the Hsf1 binding site was destroyed by four point mutations (Figure S3f in Additional file 8, construct 26) does not show an increase of GFP (b’) compared with embryos injected with the transgene alone (b). The basal muscle expression was unaffected by the point mutations. Thus, the Hsf1 recognition sequence is important for the transgene’s response in embryos with a myosin folding defect but not for basal expression in the muscle cells. c, c’ The −445/−385(unc45b)gata2:gfp construct (Figure S3i in Additional file 8, construct 35) lacking the Mef2 binding site but containing the Hsf1 recognition sequence does not drive any GFP expression when injected into wild-type embryos (c). However, co-injection of unc45b-mo with this transgene triggers activation of GFP expression in muscle cells (c’). d, d’ − 445/ − 385mut(unc45b)gata2:gfp (Figure S3i in Additional file 8, construct 42) carries point mutations in the Hsf1 binding site in addition to a deletion of the Mef2 binding site. This construct does not show GFP expression when injected alone (d) or in combination with unc45b-mo (d’). Thus, in this construct both the basal expression in muscle cells and the misfolded myosin response are abolished. e–e”’ Knock-down of Hsf1 (hsf1-mo) abolished the misfolded myosin response. Transgenic embryos stably expressing Tg(−505/−310(unc45b)gata2:gfp) were either not injected (e) (basal muscle expression) or injected with unc45b morpholinos (e’) (unc45b-mo, misfolded myosin induced expression) or double injected with morpholinos (e”) directed against unc45b and hsf1 (hsf1-mo) or with hsf1-mo alone (e”’). Co-injection of hsf1-mo and unc45b-mo (e”) blocked the induction of the transgene as observed by injection of unc45b alone (e’). Injection of hsf1-mo alone (e”’) (compare with (e”) or (e)) did not alter basal muscle expression, demonstrating that Hsf1 is only required for the misfolded myosin response and not for basal muscle expression of the transgene. f–f” Injection of a hsf1cont-mo harboring five mismatches does not prevent the ability of −505/−310(unc45b)gata2:gfp to respond to the accumulation of unfolded myosin. Tg(−505/−310 (unc45b)gata2:gfp) was either not injected (f), or injected with the unc45b-mo (f’), or unc45b-mo and hsf1cont-mo together (f”). Expression of GFP reporter in double injected embryos (f”) is as high as in embryos injected with unc45b-mo alone (f’). g–g” Knock-down of Hsf2 (hsf2-mo) does not impair the response of Tg(−505/−310(unc45b)gata2:gfp) to misfolded myosin. Tg(−505/-310(unc45b)gata2:gfp) embryos were either not injected (g) or injected with unc45b-mo (g’), or unc45b-mo and hsf2-mo (g”). h–h”’ Co-injection of the plasmid encoding Hsf1-mOrange fusion protein rescued the misfolded myosin response. Tg(−1.8unc45b:tfp) embryos were either not injected (h), or injected with unc45b-mo (h’) or with unc45b-mo and hsf1-mo (h”) or with unc45b-mo, hsf1-mo and hsf1-morange (h”’). The triple-injected embryos showed TFP reporter expression (h”’) comparable to that of embryos injected with unc45b-mo alone (h’). i–k Knock-down of Hsf1 (hsf1-mo) reduced the expression of unc45b mRNA from the endogenous gene in unc45b mutants. In situ hybridization with unc45b probe on either wild-type embryos (i), unc45b mutants (j), or unc45b mutants injected with hsf1-mo (k). unc45b mutants were unequivocally identified by the lack of well-formed myofibrils and the total lack of motility. All embryos are 72 h old and are shown anterior left and dorsal up