Regulation of Toll signaling and inflammation by β-arrestin and the SUMO protease Ulp1

Abstract

The Toll signaling pathway has a highly conserved function in innate immunity and is regulated by multiple factors that fine tune its activity. One such factor is β-arrestin Kurtz (Krz), which we previously implicated in the inhibition of developmental Toll signaling in the Drosophila melanogaster embryo. Another level of controlling Toll activity and immune system homeostasis is by protein sumoylation. In this study, we have uncovered a link between these two modes of regulation and show that Krz affects sumoylation via a conserved protein interaction with a SUMO protease, Ulp1. Loss of function of krz or Ulp1 in Drosophila larvae results in a similar inflammatory phenotype, which is manifested as increased lamellocyte production; melanotic mass formation; nuclear accumulation of Toll pathway transcriptional effectors, Dorsal and Dif; and expression of immunity genes, such as Drosomycin. Moreover, mutations in krz and Ulp1 show dosage-sensitive synergistic genetic interactions, suggesting that these two proteins are involved in the same pathway. Using Dorsal sumoylation as a readout, we found that altering Krz levels can affect the efficiency of SUMO deconjugation mediated by Ulp1. Our results demonstrate that β-arrestin controls Toll signaling and systemic inflammation at the level of sumoylation.

Keywords: Toll; Ulp1; arrestin; inflammation; sumoylation.

Figure 1

Loss of krz leads to an increased production of lamellocytes and activation of Toll signaling. (A–C) Melanotic mass formation in control and krz loss-of-function larvae (arrows). (D and E) Quantification of circulating lamellocytes in the hemolymph of third instar larvae as percentage of all hemocytes. (D) Proportion of lamellocytes increased approximately ninefold in krz^c01503 homozygous larvae and threefold in RNAi knockdown animals, relative to yw controls. (E) Proportion of lamellocytes was significantly lower in Dif¹; krz^c01503 animals, compared to krz^c01503 homozygotes. (F) Live fluorescence image of third instar larvae carrying Drosomycin-GFP reporter as part of the DD1 chromosome. Drs-GFP was highly expressed in krz^c01503 homozygotes. Arrows indicate melanotic masses. (G) Western blot analysis of expression from Drs-GFP and Dpt-lacZ (beta-gal) reporters in whole third instar larvae. Drs-GFP expression was not detected in the DD1 line but was highly elevated in DD1; krz^c01503 animals, whereas Dpt-LacZ levels were not affected. (H) Quantitative RT–PCR of endogenous Drosomycin gene expression in whole second instar larvae. Drs levels were higher in krz^c01503 homozygous larvae but were reduced when krz^c01503 was combined with either loss of Dif or genomic krz rescue (krz5.7). **P < 0.01; *P < 0.05. Error bars represent standard deviation. a.u., arbitrary units.

Figure 2

Loss of krz results in nuclear accumulation of Dorsal and Dif. (A–D′′) Antibody staining of Dorsal and Dif proteins (red) in the fat bodies of third instar larvae. DAPI (blue) marks the nuclei. Dl (A–A′′) and Dif (C–C′′) staining of the fat bodies from control yw larvae showed diffuse subcellular distribution. In the fat bodies of larvae coexpressing krz dsRNA and Dcr-2 with the Cg-GAL4 driver, both Dl (B–B′′) and Dif (D–D′′) were preferentially localized in the nuclei.

Figure 3

Ulp1 is a Krz-interacting protein, and its knockdown results in Toll pathway activation. (A and B) Co-immunoprecipitation of Krz and Ulp1 from Drosophila S2 cells. HA-Krz and Ulp1-V5 were transfected in the indicated combinations, immunoprecipitated with V5 or HA affinity resin, and immunoblotted with the corresponding antibodies. Binding between HA-Krz and Ulp1-V5 was observed in both directions of co-IP. (C) Co-immunoprecipitation of mammalian β-arrestins and SENP1 from Drosophila S2 cells. HA-ARRB1, HA-ARRB2, and Flag-SENP1 were transfected in the indicated combinations, immunoprecipitated with HA affinity resin, and immunoblotted. Flag-SENP1 formed a complex with β-arrestin 2 but not β-arrestin 1. (D) Melanotic mass formation in a third instar larva in which Ulp1 dsRNA and Dcr-2 were expressed using the da-GAL4 driver. Melanotic masses are visible throughout the body. (E) Quantification of circulating lamellocytes in the hemolymph of third instar larvae as percentage of all hemocytes. Proportion of lamellocytes increased ∼15-fold in Ulp1 knockdown animals. (F) Quantitative RT–PCR of endogenous Drosomycin gene expression in whole third instar larvae. Drs levels were 60-fold higher in Ulp1 knockdown animals, compared to controls. (G–H′′) Immunostaining of Dorsal protein (red) in the fat bodies of third instar larvae. DAPI (blue) marks the nuclei. Compared to the Hist-GFP controls (G–G′′), Dl nuclear levels were higher in the Ulp1 RNAi larvae (H–H′′). **P < 0.01. Error bars represent standard deviation. a.u., arbitrary units; IB, immunoblot; IP, immunoprecipitation.

Figure 4

Desumoylating activity of Ulp1. (A) Western blot analysis of overall sumoylation levels in whole third instar larvae. Knockdown of Ulp1 significantly increased global sumoylation in vivo, whereas knockdown of krz did not. (B) Sumoylation of Dorsal in 529SU cells. Cells were treated with control (Bla, beta-lactamase), krz, or Ulp1 dsRNA and transfected with the indicated Dl constructs together with Flag-SUMO^GG. Total cell extracts were immunoblotted with the indicated antibodies. Knockdown of Ulp1 resulted in an increased sumoylation of Dl. Transfection of high levels of Ulp1-SBP completely eliminated Dl sumoylation. A mutated form of Dl (K382R) was not sumoylated under any condition. Arrow indicates the nonsumoylated form of Dl; arrowhead marks the position of the sumoylated form. (C) Quantification of Dl sumoylation in three independent experiments. A ratio of sumoylated to a nonsumoylated form is shown. *P < 0.05. Error bars represent standard deviation.

Figure 5

Krz and Ulp1 synergistically control sumoylation and Toll pathway activity. (A–C) Formation of melanotic masses in third instar larvae. When expressed without Dcr-2 using the Cg-GAL4 driver, knockdown of neither krz nor Ulp1 alone resulted in melanotic mass formation (A and B). However, extensive melanotic masses were observed when krz and Ulp1 dsRNAs were coexpressed (C, arrows). (D) Quantification of circulating lamellocytes in the hemolymph of third instar larvae as percentage of all hemocytes. Proportion of lamellocytes increased ∼10-fold in the double knockdown animals, compared to yw controls. (E) Quantitative RT–PCR of endogenous Drosomycin gene expression in whole third instar larvae grown at 29°, using the ppl-GAL4 driver. Drs levels were 3.5-fold higher in the double knockdown animals, compared to controls. (F and G) Synergistic effects of Krz and Ulp1 on Dl sumoylation in 529SU cells. (F) Cells were treated with control (Bla, beta-lactamase), krz, and/or Ulp1 dsRNA and transfected with the Dl-V5 construct together with Flag-SUMO^GG and increasing (but small) amounts of Ulp1-SBP. Total cell extracts were immunoblotted with the indicated antibodies. Arrow indicates the nonsumoylated form of Dl; arrowhead marks the position of the sumoylated form. (G) Quantification of Dl sumoylation in three independent experiments. Knockdown of both krz and Ulp1 resulted in a significantly higher level of Dl sumoylation, compared to a knockdown of Ulp1 alone. (H) Model summarizing the activity of Krz and Ulp1 in controlling sumoylation and Toll pathway activity. Krz promotes desumoylating activity of Ulp1 toward Dorsal and other targets. A proper balance of sumoylation is important for limiting Toll pathway activity and preventing an inappropriate inflammatory response in the absence of pathogens. **P < 0.01; *P < 0.05. Error bars represent standard deviation.