SPS1 deficiency-triggered PGRP-LC and Toll expression controls innate immunity in Drosophila S2 cells

Selenium is an essential trace element in the diet of humans and many other life forms. An appropriate amount of selenium offers many health benefits, such as preventing cancer and heart disease, acting as an antiviral agent, scavenging reactive oxygen species (ROS), and augmenting the immune system and male reproduction (Brigelius-Flohé and Sies, 2016; Gladyshev et al., 2016; Na et al., 2018). The many benefits of selenium may be due to its existence in selenoproteins in the form of the amino acid selenocysteine (Sec) (Hatfield and Gladyshev, 2002; Brigelius-Flohe, 2008; Lu and Holmgren, 2009). Sec is the 21st amino acid and enters into a growing peptide in response to the UGA codon during translation (Lee et al., 1989; Leinfelder et al., 1989; Longtin, 2004; Squires and Berry, 2008; Allmang et al., 2009). The active donor of selenium in Sec biosynthesis is monoselenophosphate (Glass et al., 1993). It is synthesized from selenide and ATP by the enzyme selenophosphate synthetase (SPS, also called SelD or patufet) (Ehrenreich et al., 1992). Only one type of SPS exists in prokaryotes, including Archaea. However, there are two isoforms of SPS, SPS1 and 2, in eukaryotes (Guimaraes et al., 1996). In higher animals, such as mammals, SPS is referred to as SEPHS, because sucrose-phosphate synthase is also designated as SPS. However, there is no sucrose–phosphate synthase in Drosophila melanogaster. The amino acid sequences of SPS1 and 2 are highly conserved. One of the main differences between the sequences of SPS1 and 2 is that they have an arginine and Sec, respectively, in a homologous region (Low et al., 1995). Additionally, only SPS2 has selenophosphate synthesis activity (Xu et al., 2007).

In D. melanogaster, SPS1 deficiency leads to aberrant imaginal-disc morphology and embryonic lethality (Alsina et al., 1998). Furthermore, Sps1 knockdown decelerates cell growth, activates the innate immunity by upregulating AMPs, increases ROS levels, and induces megamitochondria formation in Drosophila S2 cells (Shim et al., 2009; Lee et al., 2011). These phenotypes are caused through downregulation of pyridoxal phosphate (PLP), a biologically active form of vitamin B6 (Lee et al., 2011). In mice, systemic knockout of Sephs1 (Sps1) gradually increases the oxidative stress, thereby impairing gastrulation-related signaling pathways and causing embryonic lethality (Tobe et al., 2016; Bang et al., 2021). In 2H11 cells, SEPHS1 deficiency increases the superoxide level, thereby causing DNA damage, which suppresses the cell proliferation and impairs the cell functions (Jung et al., 2021).

Innate immunity is an important defense system against infections in metazoans (Janeway and Medzhitov, 2002; Lemaitre and Hoffmann, 2007). As with all invertebrates, D. melanogaster depends entirely on innate immunity to thwart infections (Lemaitre and Hoffmann, 2007). Antimicrobial peptides (AMPs) are one of the main effector molecules in the innate immune system. AMPs protect the host by destroying the cell wall of invading microorganisms with cationic and amphipathic peptides (Zhang and Gallo, 2016). The induction of AMP production upon infection is regulated via two distinct signaling pathways – the Toll and IMD pathways – in D. melanogaster (Lemaitre and Hoffmann, 2007; Valanne et al., 2011; Myllymaki et al., 2014). The Toll pathway is used for activating the expression of Drosomycin (Drs) and Metchnikowin (Mtk), and these AMPs are required to protect cells from infections by fungi or Gram-positive bacteria. The active form of spätzle (Spz), cleaved by the spätzle-processing enzyme (SPE), activates the Toll signaling (Kanoh et al., 2015) and finally induces the nuclear translocation of the proteins nuclear factor-kappa B (NF-κB), dorsal-related immunity factor (DIF), and dorsal, thereby activating the expression of AMP genes, including Drs and Mtk (Lindsay and Wasserman, 2014). The IMD pathway is activated upon detecting diaminopimelic acid (DAP)-type peptidoglycans, which are derived from Gram-negative bacteria, via the transmembrane receptor PGRP-LC (Kleino and Silverman, 2014). This transmembrane receptor transduces the signal to downstream factors, including the adaptor proteins (IMD) and NF-κB (relish), and eventually AMPs, such as Drosocin (Dro), Diptericin (Dpt), Attacin (Att), and Cecropin (Cec), are upregulated (Paquette et al., 2010).

Although SPS1 is involved in the regulation of the D. melanogaster innate immune system, the mechanisms whereby SPS1 regulates the AMP production are elusive. In this study, we investigated the signaling components through which SPS1 modulates D. melanogaster innate immunity and found that PGRP-LC and Toll, two genes of transmembrane receptors in the IMD and Toll pathways, respectively, are the primary targets in SPS1-deficiency–induced AMP production.

We have previously reported that Sps1 knockdown upregulates AMPs that are responsible for D. melanogaster innate immunity (Lee et al., 2011). To elucidate which immune pathways are regulated upon SPS1 deficiency, Sps1 was knocked down in S2 cells, and the expression levels of AMPs were measured using RT-qPCR after 5 days. DptB, CecB, Dro, Mtk, and Drs, which are AMP genes widely used to assess whether the innate immune system is activated, were selected as AMP markers. Data analysis revealed that AMPs of both the IMD (DptB, CecB, and Dro) and Toll (Mtk and Drs) pathways were upregulated 10–50 fold (P<0.001) in SPS1-deficient cells (Fig. 1A and B). These results indicate that SPS1 deficiency activates both the IMD and Toll pathways.

To identify the target genes through which SPS1 regulates the expression of the AMPs in the IMD and Toll pathways, the upstream genes in each pathway were individually knocked down along with Sps1.

First, to investigate the IMD pathway, PGRP-SD, the most upstream gene in the IMD pathway and Sps1 were co-knocked down. As shown in Fig. 2A, DptB and Mtk were upregulated upon Sps1 knockdown were not recovered upon PGRP-SD/Sps1 co-knockdown. However, intriguingly, the expression levels of DptB and Mtk were significantly reduced to the background levels (GFP control) upon PGRP-LC/Sps1 co-knockdown (P<0.001) (Fig. 2A). Other AMP genes (CecB, Dro, and Drs) showed similar results (Fig. S1A). These results indicate that SPS1 deficiency affects the IMD pathway by regulating PGRP-LC.

Next, the Toll pathway was likewise investigated. No recovery was observed when SPE and Sps1 were co-knocked down (Fig. 2B). However, when Toll and Sps1 were co-knocked down, the expression of DptB and Mtk was decreased significantly (P<0.001, Fig. 2C). Other AMP genes (CecB, Dro, and Drs) also showed similar results (Fig. S1B). The expression data for all the AMP genes analyzed in this study are summarized in Table 1. The knockdown efficiency was >90% for all the genes (Fig. S1C). Interestingly, the readout AMPs of the Toll pathway were downregulated upon co-knocking down PGRP-LC, which participates in the IMD pathway, and vice versa, (the readout AMPs of IMD pathway were also downregulated by the co-knocking down Toll). These data suggest that there is a crosstalk between the IMD and Toll pathways (Fig. 2). Taken together, it can be concluded that SPS1 regulates AMP expression through a crosstalk between the IMD and Toll pathways at the transmembrane receptor level, namely through PGRP-LC and Toll, respectively.

Since PGRP-LC and Toll are the targets of SPS1, and it has been reported that ecdysone regulates the innate immune system by upregulating PGRP-LC (Rus et al., 2013), we hypothesized that SPS1 deficiency upregulates PGRP-LC and Toll. As shown in Fig. 3A, Sps1 knockdown significantly increased the levels of the PGRP-LC and Toll mRNAs (6.4±0.36 and 3.2±0.45-fold, respectively, n=3), indicating that SPS1 regulates both PGRP-LC and Toll presumably at the transcription level. Interestingly, SPS1 deficiency upregulated PGRP-LC by approximately 2-fold (P<0.001, n=3) compared with that of Toll, suggesting that SPS1 deficiency has a stronger effect on PGRP-LC than on Toll.

To decipher whether upregulation of PGRP-LC or Toll upregulates AMPs, we overexpressed PGRP-LCa, PGRP-LCx, or Toll in S2 cells and examined the expression levels of AMPs. The overexpression of each protein upregulated Mtk (96.16±12.71, 218.72±10.08, and 3.04±0.72-fold for PGRP-LCa, PGRP-LCx, and Toll, respectively, n=3) and DptB (198.55±18.95, 376.74±5.33, and 2.28±0.3-fold for PGRP-LCa, PGRP-LCx, and Toll, respectively, n=3) (Fig. 3B,C). The expression levels of AMPs upon PGRP-LCa, PGRP-LCx, or Toll overexpression is summarized in Table 2. Other AMPs (CecB, Dro, and Drs) were also likewise upregulated when PGRP-LCa, PGRP-LCx, or Toll was overexpressed (Fig. S2). Notably, AMP production was induced more in cells overexpressing PGRP-LC than in those overexpressing Toll although the two proteins were overexpressed to a similar extent. In addition, PGRP-LC overexpression activated the expression of Mtk, which is a target of Toll. This result supports the crosstalk between the IMD and Toll pathways. Taken together, our observations indicate that SPS1 participates in the innate immune system by controlling the expression of the genes of two transmembrane receptors, PGRP-LC and Toll, and the amount of PGRP-LC affects the innate immune system more than the amount of Toll.

SPS1 is known to play an essential role in growth of cells, vitamin-B6 synthesis, and innate immunity in D. melanogaster (Shim et al., 2009; Lee et al., 2011). Among these various functions of SPS1, we focused on how SPS1 affects the immune system, especially on the AMP overproduction upon SPS1 deficiency. Via Sps1 knockdown in S2 cells, we found that SPS1 regulates both the IMD and Toll pathways in the innate immune system.

In this study, we identified PGRP-LC as the primary target gene of the IMD pathway. PGRP-LC is known as the most upstream gene in the sub-cellular IMD pathway (Buchon et al., 2014). Recently, PGRP-SD was found to be involved in the IMD pathway outside the cell. It binds to peptidoglycans (PGNs) that are produced by digestion of bacterial cell walls, and the PGRP-SD/PGN complex helps re-localization of PGN to PGRP-LC on the cell surface (Iatsenko et al., 2016). Our results from the co-knockdown experiments clearly revealed that SPS1 targets PGRP-LC, not PGRP-SD, to induce the IMD pathway upon SPS1 deficiency. Toll was also found to be the primary target of SPS1 to induce the Toll pathway. The induction of both the IMD and Toll pathways is triggered by the upregulation of transmembrane receptors–PGRP-LC and Toll for the IMD and Toll pathways, respectively.

Two different lines of evidence support our findings that the induction of PGRP-LC and Toll pathway triggers the activation of AMP expression. First, knocking down SPS1 (SelD) induced Diptericin expression, whereas knocking down PGRP-LC inhibited Diptericin expression (Foley and O'Farrell, 2004). Second, the activation of PGRP-LC by ecdysone treatment upregulated AMPs (Rus et al., 2013). Altogether, our results indicate that upregulation of PGRP-LC is sufficient to induce the AMP signaling pathway. Toll overexpression has been reported to slightly induce the expression of Drosomycin (approximately 5-fold) (Hu et al., 2004). However, it is unclear why the upregulation of Toll had an insignificant effect on the activation of the innate immune pathways.

We found that the activation of AMP expression occurs as a result of a crosstalk between the IMD and Toll pathways. A crosstalk between these two innate immune pathways has been previously reported (Tanji et al., 2007; Nishide et al., 2019). Tanji et al. showed that overexpression of PGRP-LC or Toll activated both the IMD and Toll pathways. In our study, we also showed that overexpression of PGRP-LC or Toll induces both these pathways. In addition, our co-knockdown experiments showed that knockdown of PGRP-LC or Toll downregulates the AMPs in both the IMD and Toll pathways.

It was reported with microarray analysis that SPS1 deficiency upregulates the genes that participate in defense response, including PGRP-SD, PGRP-LF, and pirk (Lee et al., 2011). In this study, we found that PGRP-SD, unlike PGRP-LC, did not affect the activation of the innate-immune signaling induced upon SPS1 deficiency. Since PGRP-SD itself is a target gene of the IMD pathway and is activated upon PGRP-LC upregulation (Iatsenko et al., 2016), it is upregulated upon SPS1 deficiency presumably because the innate immune system is activated. PGRP-LF and Pirk suppress the IMD pathway and act as negative feedback regulators of this pathway (Maillet et al., 2008; Kleino et al., 2008). Therefore, the upregulation of PGRP-LF and pirk might be due to the upregulation of AMP genes upon SPS1 deficiency.

Recently, SPS1 has been suggested to regulate redox homeostasis (Tobe et al., 2016; Bang et al., 2021; Jung et al., 2021). To examine whether SPS1 affects the innate immune system through ROS in D. melanogaster, we knocked down Sps1 in S2 cells and then treated the cells with the antioxidant N-acetyl cysteine (NAC) to reduce ROS. We found that the expression levels of PGRP-LC and Toll were not consequently changed, and this result suggests that the regulation of the innate immune system through SPS1 is independent of the redox system (Fig. S3A).

In the previous study, we showed that the vitamin B6 metabolism was regulated by the intracellular SPS1 levels in Drosophila S2 cells (Lee et al., 2011). We examined the effect of vitamin B6 on innate immunity in scrutiny in this study. When 4-deoxypyridoxine, an inhibitor of PLP biosynthesis, was administered to the cells without knockdown of Sps1, PGRP-LC and Toll were also upregulated, as in the case of Sps1 knockdown (Fig. S3B). Therefore, it seems that SPS1 controls the expression of PGRP-LC and Toll by regulating the synthesis of vitamin B6. Like in D. melanogaster, the effects of vitamin B6 on immunity were also examined in a mouse model. The deficiency of vitamin B6 that was achieved by feeding the mice a low vitamin B6 diet led to increased immunoglobulin E production, presumably by the upregulation of interleukin-4 (Doke et al., 1997). The deficiency of vitamin B6, however, suppressed immunoglobulin G or immunoglobulin M (Kumar and Axelrod, 1968). Notably, Excess vitamin B6 levels affected immunity in the opposite manner (Inubushi et al. 2000). Therefore, it seems that vitamin B6 regulates immune homeostasis in mammalian systems. However, it remains unclear whether vitamin B6 also regulates immune homeostasis in D. melanogaster. An in vivo study may provide further insights into this issue.

A comparison of the mRNA levels of the AMPs induced upon Sps1 knockdown with those induced upon immune stimulation would be interesting. An in vivo study may also facilitate evaluation of the detailed relationship Sps1 knockdown and immune stimulation. Therefore, to elucidate the function of SPS1 in more detail and to confirm the findings of our current in vitro study, an in vivo study using a fruit fly system may be helpful. As shown by Alsina et al., SPS1-deficient fruit flies die in the late larval stage (Alsina et al., 1998). Thus, it would be intriguing to examine the effects of SPS1 deficiency on the expression of PGRP-LC and Toll, upregulation of AMPs, response upon immune stimulation, and the mechanism how vitamin B6 regulates the innate immunity using Sps1-knockout D. melanogaster larva.

Although our study has some pitfalls, we provide an important finding that SPS1 regulates the innate immune system of D. melanogaster by controlling the expression of PGRP-LC and Toll without any other immune stimulation.

Drosophila Schneider cell line 2 (S2) was purchased from Invitrogen. HyQ SFX-Insect medium was purchased from Hyclone, T3 Megascript kit was purchased from Ambion, PowerUp™ SYBR™ Green Master Mix was purchased from Thermo Fisher, TRIzol reagent was purchased from Invitrogen, Moloney murine leukemia virus reverse transcriptase and nPfu forte DNA polymerase were purchased from Enzynomics. Dimethyldioctadecyl ammoniumbromide was purchased from Sigma-Aldrich, and oligonucleotides were purchased from Cosmo Genetech. The sequences of oligos used for RT-PCR and dsRNA are listed in Supplementary data.

pAcPA PGRP-LCa was produced by PCR amplification of BamHI-PGRP-LCa-KpnI from cDNA of S2 cells, cut and ligated into the pAcPA vector (Shim et al., 2009), containing the actin 5C promoter, with BamHI/KpnI. pAcPA PGRP-LCx and pAcPA Toll were prepared in the same way.

To prepare double-stranded RNA (dsRNA) of Sps1, PGRP-SD, PGRP-LC, SPE and Toll, each gene was amplified with a primer pair. The sequences of each primer are provided in Table S2. Each primer was fused with a T3 promoter sequence (5′-AATTAACCCTCACTAAAGGG-3′) at its 5′ end. In vitro transcription was performed using the T3 Megascript kit according to the manufacturer's protocols and then the dsRNAs were produced by annealing each complementary strand set.

S2 cell culture and RNA interference using dsRNAs were carried out as described previously with minor modification (Shim et al., 2009). Briefly, for RNA interference, 2.5×10⁵ cells were plated on a 24-well plate containing 0.5 ml of HyQ SFX-Insect medium. Four micrograms of dsRNAs were added directly to the medium and incubated for 48 h and cells were split into appropriate culture dishes for further incubation and other experiments.

RT-qPCR was carried out as described with minor modification (Shim et al., 2009). Briefly, total RNA was isolated from the cells using the TRIzol reagent. cDNAs were synthesized from total RNAs with Moloney murine leukemia virus reverse transcriptase and oligo (dT) primers according to the manufacturer's protocols. RT-qPCR was carried out using an ABI 7300 real-time PCR system (Applied Biosystems) as follows. cDNAs were amplified using SYBR Green mix and specific primers for 40 cycles [initial incubation at 50°C for 2 min and then at 95°C for 10 min, and 40 cycles (95°C for 15 s, 55°C for 1 min, and 72°C for 1 min)]. Output data was obtained as Ct values using Sequence Detection Software (SDS) version 1.3 (7300 System, Applied Biosystems) and the differential mRNA expressions of each gene between control and knockdown cell were calculated using the comparative Ct method (Schmittgen and Livak, 2008). rp49 mRNA, an internal control, was amplified along with the target genes, and the Ct value of rp49 was used to normalize the expression of target genes.

Vectors were transfected into S2 cells as described previously (Han, 1996) with minor modifications. Briefly, 2 μg of pAcPA (backbone vector), pAcPA PGRP-LCa, pAcPA PGRP-LCx, and pAcPA Toll were mixed with 100 μl of dimethyldioctadecyl ammonium bromide (125 μg/ml) and 200 μl of HyQ-SFX-Insect media. The mixture was incubated for 30 min at room temperature and then added into a well of a six-well plate containing 2×10⁶ cells.

Each experiment was performed in biological triplicate for statistical analysis. Statistical analyses were performed using an unpaired Student's t-test or one-way ANOVA followed by Tukey's multiple comparison test. A value of P<0.05 was considered significant.

We would like to thank the Seoul National University Research Facility Center for allowing us to use the ABI 7300.