Research Article: Inhibition of a NF-κB/Diap1 Pathway by PGRP-LF Is Required for Proper Apoptosis during Drosophila Development

Date Published: January 13, 2017

Publisher: Public Library of Science

Author(s): Raphael Tavignot, Delphine Chaduli, Fatoumata Djitte, Bernard Charroux, Julien Royet, Norbert Perrimon

Abstract: NF-κB pathways are key signaling cascades of the Drosophila innate immune response. One of them, the Immune Deficiency (IMD) pathway, is under a very tight negative control. Although molecular brakes exist at each step of this signaling module from ligand availability to transcriptional regulation, it remains unknown whether repressors act in the same cells or tissues and if not, what is rationale behind this spatial specificity. We show here that the negative regulator of IMD pathway PGRP-LF is epressed in ectodermal derivatives. We provide evidence that, in the absence of any immune elicitor, PGRP-LF loss-of-function mutants, display a constitutive NF-κB/IMD activation specifically in ectodermal tissues leading to genitalia and tergite malformations. In agreement with previous data showing that proper development of these structures requires induction of apoptosis, we show that ectopic activation of NF-κB/IMD signaling leads to apoptosis inhibition in both genitalia and tergite primordia. We demonstrate that NF-κB/IMD signaling antagonizes apoptosis by up-regulating expression of the anti-apoptotic protein Diap1. Altogether these results show that, in the complete absence of infection, the negative regulation of NF-κB/IMD pathway by PGRP-LF is crucial to ensure proper induction of apoptosis and consequently normal fly development. These results highlight that IMD pathway regulation is controlled independently in different tissues, probably reflecting the different roles of this signaling cascade in both developmental and immune processes.

Partial Text: In Drosophila, bacteria infection triggers NF-κB cascades (called Toll and IMD (Immune Deficiency)) leading to the production of immune effectors and regulators [1–3] [4]. This activation relies on the previous recognition of bacteria derived peptidoglycan (PGN) by host Peptidoglycan Recognition Protein (PGRP) family members. Recognition of Gram-positive PGN by circulating PGRP-SA triggers the maturation of the pro-Spätzle protein into an active ligand for the Toll membrane receptor [5]. The IMD pathway is triggered upon recognition of PGN by either transmembrane associated PGRP-LC or cytoplasmic PGRP-LE [5–10]. Receptor activation leads to the recruitment of the death-domain containing adapters IMD, FADD, and the caspase DREDD. Activated DREDD cleaves IMD, thus allowing its poly-ubiquitination that allows recruitment of the TAK1/TAB2 and the IRD5/Kenny kinase to the receptor complex [11]. These interactions ultimately lead to the nuclear translocation of the transcription factor Relish. In contrast to Toll signaling, IMD pathway activation after bacterial infection is transient and buffered by many repressors [12–14]. This tight control might reflect the essential role played by the IMD cascade in controlling antibacterial response in fly epithelia [4, 15]. Indeed, the constant contact between bacteria and epithelia requires the presence of immune tolerance mechanisms through which the epithelium copes with the continuous input from microbiota derived immune-activating signals [16–19]. The homeobox transcription factor Caudal was one of the first proteins identified as an IMD pathway antagonist [20]. Through its occupation of some IMD target promoters, Caudal blocks IMD-dependent transcription. Negative regulation is also mediated through protein turnover by ubiquitous factors that regulate protein stability of identified IMD pathway components (dUSP36, CYLD, DNR-1, Caspar) [21–23] [24–26]. Some of IMD pathway regulation is also taking place at the level of the PGRP-LC receptor itself and of its PGN ligand. PGRP-LC transcription is under the control of the steroid hormone ecdysone [27]. The number of PGRP-LC molecules at the membrane, depends on intracellular (Pirk) and membrane associated (nonaspanins TM9SF2 and TM9SF4) proteins that by sequestering PGRP-LC in the cytoplasm prevent its localisation at the membrane [16, 28–31]. Another member of the PGRP family, PGRP-LF, antagonizes IMD pathway activation [32]. This transmembrane protein which has no intracytoplasmic tail but has two occluded PGRP domains is unable to bind PGN [32, 33]. Plasma resonance data show that by interacting with PGRP-LC ectodomain, PGRP-LF prevents constant activation of the IMD pathway even in the absence of bacteria [34]. IMD pathway tuning is also mediated through the modulation of ligand availability, via a family of extracellular enzymes called amidases, which degrade PGN into non-stimulatory fragments [17, 18, 35–40]. While the inhibition provided by these regulators appears to be constitutive, the negative regulation brought by amidase and by PIRK is the result of a negative feedback loop. As a result, these factors additively regulate the amplitude of the IMD response. While the detrimental effects of runaway inflammation in mammals are well established, the situation is less clear with regards to Drosophila. The absence of ubiquitous negative regulators leads to a reduced lifespan, which however cannot be ascribed specifically to the constitutive activation of the IMD pathway, as these regulators act upon multiple targets. Modulation of amidase levels causes deregulation of NF-κB activity in the gut, resulting in commensal dysbiosis, stem cell hyper-proliferation, epithelial dysplasia and eventually reduced life span [18, 40]. Importantly, these phenotypes can be partially rescued in germ-free conditions or by inactivating IMD pathway components demonstrating that they are direct consequences of IMD pathway stimulation by bacteria.

We showed here that one of essential role of PGRP-LF is to prevent a bacteria independent constitutive activation of the NF-κB pathway, which otherwise perturbs tergite and genitalia formation during pupariation. We also confirmed previous data showing that a lack of IMD pathway repression by PGRP-LF is leading to AMP production in ectodermal derivatives [33]. Both PGRP-LF expression pattern and loss-of-function phenotype analyses showed that PGRP-LF is mainly acting in ectodermal cells. This is mostly evident in the intestinal tract that is formed during embryogenesis by associating domains of both mesodermal and ectodermal origins. Although IMD pathway is essential in regulating antibacterial response in the mesodermal derived midgut, PGRP-LF is only playing a minor role as an IMD regulator in this tissue. This contrasts with its importance in both neighboring ectodermal derivatives that are the fore and the hindgut. Loss of PGRP-LF function triggers in these tissues a massive AMP production. Interestingly also, is the fact that the effects of inactivating PGRP-LF, and hence of IMD pathway permanent activation, are not the same in all ectodermal derivatives. Whereas in trachea, epidermis or hind/foregut, it only leads to AMP constitutive production, it has profound and deleterious effects on tergites and genitalia. Of note, these are the two known structures whose proper morphogenesis has been shown via P35 overexpression to depend on caspase activity [41]. Removing PGRP-LF, blocks apoptosis and in turn interferes with developmental processes of these adult structures. It could be that apoptosis is also prevented in other ectodermal derivatives but that this has no impact on their development. Since mutations in caspases cause pleiotropic defects during development, it is obvious that PGRP-LF is only antagonizing a limited fraction of them, consistently with its restricted spatial pattern [47].