Cysteine and methionine are key factors in HPD-induced peroxisome elevation

To explore the impacts of high protein intake on peroxisomes across various Drosophila tissues, we generated a transgenic peroxisomal reporter line, Ubi-GFP-PTS1, that ubiquitously labels peroxisomes in all fly tissues (Fig. 1a). Ubi-GFP-PTS1 adult flies and 3rd instar larvae were provided with either a normal diet (ND) or a high-protein diet (HPD) formulated by Thomas J Baranski et al. [17]. 36 h later, major fly organs were dissected followed by the measurement of peroxisome levels. Compared to the control diet, HPD led to an increased intensity of green fluorescent protein (GFP) signal within the adipose tissues of Ubi-GFP-PTS1 larvae and adult flies (Fig. S1a, b, i), indicating an increase in peroxisome levels under HPD feeding. No significant differences in peroxisome levels were observed in the adult muscles, adult brains, adult egg chambers, adult testes, and larval guts of Ubi-GFP-PTS1 line fed with HPD compared to the control diet (Fig. S1c–f, h, i). However, a slight decrease in peroxisome levels was observed in the adult guts of Ubi-GFP-PTS1 strain when subjected to HPD (Fig. S1g, i).

Fig. 1

Cysteine and methionine are key factors in HPD-induced peroxisome elevation. a Scheme of the construction of the Ubi-GFP-PTS1 line and the experimental setup. b Peroxisome levels in the adipose tissues of Ubi-GFP-PTS1 3rd instar larvae and female flies fed with Suc (5% sucrose) or Suc + YE (5% sucrose + 30% yeast extract) for 36 h. DAPI (blue) labeled nuclei. Scale bar: 100 μm. c Relative fluorescence intensities of peroxisomes in b and Fig. S2. From left to right: 19, 21, 19, 20, 11, 11, 7, 8, 19, 19, 11, 10, 14, 12, 16, and 14 views. d Peroxisome levels in the adipose tissues of Ubi-GFP-PTS1 female flies 36 h following the designated treatments. The boxed areas were enlarged to the lower panel. DAPI (blue) labeled nuclei. Suc: 5% sucrose; Suc + Cys: 5% sucrose + 25 mM cysteine; Suc + Met: 5% sucrose + 25 mM methionine. Scale bar: 100 μm. e Relative fluorescence intensities of peroxisomes in Fig. S3. From left to right: 30, 33, 29, 24, 26, 22, 25, 21, 30, 32, 27, 28, 31, 32, 32, 32, 32, 27, 26, 25, and 31 views. Two-tailed Student’ s t test (c) or one-way ANOVA (e) were performed. ** p < 0.01; **** p < 0.0001; ns, not significant

Yeast extract (YE) serves as the major protein source for Drosophila in natural and experimental environments [20]. To further confirm the effects of HPD on peroxisomes, we adopted a minimal food medium containing 5% sucrose and adjusted the YE content to 30% (based on the calculations derived from ND and HPD; see methods). Subsequently, Ubi-GFP-PTS1 adult flies and 3rd instar larvae were fed with or without 30% YE for 36 h before examination. Consistent with our previous findings, YE feeding promoted peroxisome elevation in the adipose tissues of Ubi-GFP-PTS1 larvae and adult flies compared to the 5% sucrose control diet (Fig. 1b, c). Conversely, no significant changes in peroxisome levels were observed in adult brains, adult egg chambers, adult testes, adult guts, and larval guts in this YE-fed Ubi-GFP-PTS1 line compared to the 5% sucrose control diet, whereas adult muscles showed a slight increase in peroxisome levels (Fig. S2; Fig. 1c). Collectively, HPD specifically increases peroxisome levels in the adipose tissues of Drosophila.

Theoretically, the upregulation of peroxisome levels induced by HPD could be attributed to the overall increase in amino acid concentrations. However, there is a possibility that certain amino acids play a key role in promoting peroxisome elevation. To test this hypothesis, we fed Ubi-GFP-PTS1 female flies with 20 individual amino acids for a duration of 36 h and measured peroxisome levels in their adipose tissues. Among the 20 naturally occurring L-amino acids tested, only cysteine and methionine significantly elevated peroxisome levels in the adipose tissues, whereas threonine slightly lowered peroxisome levels within the adipose tissues (Fig. 1d, e; Fig. S3), indicating that cysteine and methionine are key factors in HPD-induced peroxisome elevation.

HPD elevates peroxisome levels by triggering CG33474 expression

The results above suggested that methionine and cysteine might increase peroxisome levels by activating specific Pex genes. To identify potential peroxisomal genes responsible for increasing peroxisome levels under HPD feeding, we selected methionine, the more extensively studied amino acid [21], and conducted RNA-Seq analyses on w1118 female flies that were fed with or without methionine for a period of 36 h. Compared to the methionine-absent group, over 700 genes were significantly changed in the methionine-fed group. Analyses of the differentially expressed genes revealed that the transcript level of CG33474, a peroxisomal fission gene, was significantly increased in flies fed a methionine-rich diet (Fig. 2a). We further verified the RNA-Seq results by RT-qPCR. Consistent with the RNA-Seq data, RT-qPCR results demonstrated that cysteine and methionine, but not other amino acids, effectively induced the expression of CG33474 (Fig. 2b).

Fig. 2

HPD elevates peroxisome levels by triggering CG33474 expression. a Volcano plot depicting the differentially expressed genes in w1118 female flies treated with Ctrl (1% agarose) or Met (1% agarose + 25 mM methionine) for 36 h. The red and blue dots represented significantly upregulated and downregulated genes under methionine treatment, respectively. b Relative CG33474 mRNA levels in w1118 female flies fed with Ctrl (1% agarose) or various amino acids (1% agarose + 25 mM individual amino acids) for 36 h. n = 3. c Peroxisome levels in the adipose tissue of the designated genotypes fed with Suc (5% sucrose) or Suc + YE (5% sucrose + 30% yeast extract) for 36 h. DAPI (blue) labeled nuclei. Scale bar: 100 μm. d Relative fluorescence intensities of peroxisomes in c. From left to right: 25, 23, 24, and 25 views. One-way ANOVA (b) or two-tailed Student’s t test (d) were performed. *** p < 0.001; **** p < 0.0001; ns, not significant

Next, we quantified the transcript levels of four distinct categories of peroxisome-associated genes in w1118 female flies following a-36-h cysteine or methionine diet: peroxisomal fission genes (Fis, Pex11ab, and Pex11c), peroxisomal organization-related genes (ABCD, Pmp70, CG31454, CG31259, and CG11737), genes involved in peroxisomal membrane assembly (Pex3, Pex16, and Pex19), and genes responsible for peroxisomal protein import (Pex1, Pex2, Pex5, Pex6, Pex7, Pex10, Pex12, Pex13, and Pex14), respectively [22]. None of these genes were concurrently induced by both methionine and cysteine, unlike CG33474 (Fig. S4). Together, it was reasonable to infer that CG33474 might be a crucial peroxisomal gene in HPD-induced peroxisome elevation.

To ascertain whether HPD-induced peroxisome elevation depends on CG33474, we knocked down Luciferase or CG33474 specifically in the adipose tissues and subsequently fed female flies of these genotypes with or without 30% YE for 36 h. Compared to the control diet, knocking down Luciferase in the fat body (lpp-Gal4>UAS-Luciferase-RNAi, Ubi-GFP-PTS1) led to a significant increase in peroxisome levels under HPD feeding as expected (Fig. 2c, d). On the contrary, knockdown of CG33474 in the fat body (lpp-Gal4>UAS-CG33474-RNAi, Ubi-GFP-PTS1) failed to elevate peroxisome levels under HPD feeding (Fig. 2c, d), suggesting that HPD-induced peroxisome elevation requires CG33474.

Cysteine and methionine are key factors in HPD-induced CG33474 expression

To better visualize the expression of CG33474 in vivo, we generated a CG33474-Gal4 knock-in (KI) line using the CRISPR-Cas9 system [23] (Fig. 3a). Notably, the CG33474-Gal4 line also functions as a CG33474 mutant line because the coding region of CG33474 was destroyed (Fig. 3a). We verified the mutation effect of this line through RT-qPCR (Fig. S5a). Additionally, the CG33474-Gal4 homozygous flies are viable and fertile, indicating that CG33474 may not play essential roles under normal developmental conditions. CG33474-Gal4 flies were crossed with UAS-RFP flies, and the offspring were used as a CG33474 fluorescent reporter line. Compared to ND feeding, CG33474-Gal4>UAS-RFP female flies under HPD revealed a significant increase in the percentage of RFP+ flies (flies displaying red fluorescence) (Fig. S5b, c), validating the efficacy of this CG33474 fluorescent reporter system. Next, we fed CG33474-Gal4>UAS-RFP female flies with various diets for 36 h. A significant increase in the proportion of RFP+ flies was observed under the YE condition (Fig. 3b, c), verifying that HPD triggers the expression of CG33474. However, CG33474 exhibited no activation under high-sugar or high-fat diets, as shown by no increase in the percentage of RFP+ flies (Fig. 3b, c), indicating that the trigger of CG33474 is HPD-specific. Next, we tested the roles of individual amino acids in inducing CG33474 expression using CG33474-Gal4>UAS-RFP female flies. Consistent with previous results, dietary cysteine and methionine both significantly potentiated CG33474 expression to a distinctly greater extent than any other amino acids (Fig. 3d), indicating that cysteine and methionine are key factors in HPD-induced CG33474 expression.

Fig. 3

Cysteine and methionine are key factors in HPD-induced CG33474 expression. a A diagram represents the CG33474-Gal4 knock-in approach. Represent images (b) and quantification (c, from left to right: 85, 84, 82, and 81 flies) of CG33474-Gal4>UAS-RFP female flies expressing RFP 36 h following the indicated treatments. 5% Suc: 5% sucrose; 35% Suc: 35% sucrose; 5% Suc + CO: 5% sucrose + 15% coconut oil; 5% Suc + YE: 5% sucrose + 30% yeast extract. d Percentage of CG33474-Gal4>UAS-RFP female flies expressing RFP 36 h following Suc (5% sucrose) or Suc + amino acids (5% sucrose + 25 mM individual amino acids) diets. From left to right: 68, 54, 68, 56, 52, 60, 51, 51, 53, 61, 51, 55, 54, 43, 51, 56, 51, 85, 77, 60, and 52 flies. Fisher’s Exact Test was performed. **** p < 0.0001; ns, not significant

Given that cysteine functions as a scavenger of reactive oxygen species (ROS) [24] and methionine serves both as a metabolic precursor of cysteine and a direct target of ROS [25], we investigated whether alterations in ROS levels could trigger CG33474 expression. To this end, we fed CG33474-Gal4>UAS-RFP female flies with different reductive or oxidative compounds for 36 h. Neither the inhibition of ROS through the administration of Trolox or Vitamin C [26] nor the induction of ROS via Paraquat or H2O2 [27] was able to promote CG33474 expression (Fig. S5d). Altogether, these data suggested that CG33474 expression is probably triggered by certain metabolic products of cysteine and methionine, rather than changes in ROS levels.

Next, we explored whether methionine- and cysteine-induced CG33474 expression is gender-specific. For this purpose, we fed CG33474-Gal4>UAS-RFP female and male flies with or without methionine or cysteine for 36 h. The results showed that in both sexes, the proportion of RFP+ flies significantly rose upon exposure to methionine or cysteine (Fig. S5e), indicating that cysteine- and methionine-induced CG33474 expression is sexual-independent.

CG33474/PEX11G proteins lead to increased peroxisome size

To examine whether CG33474 regulates the size of peroxisomes, we co-transfected CG33474-Flag and GFP-PTS1 (GFP fused to the PTS1 labeled peroxisomes) encoding plasmids into Drosophila S2 cells. The results demonstrated that CG33474 proteins localized to peroxisomes within S2 cells (Fig. 4a). Furthermore, CG33474 overexpression resulted in increased peroxisome size in S2 cells (Fig. 4b), aligning with the previously documented juxtaposed elongated peroxisomes (JEPs) induced by PEX11G in mammalian cells [28].

Fig. 4

CG33474/PEX11G proteins lead to increased peroxisome size. a Fluorescence microscopy images of Drosophila S2 cells co-transfected with CG33474-Flag (red) and GFP-PTS1 encoding plasmids. DAPI (blue) labeled nuclei. GFP-PTS1 (green) indicated peroxisomes. Arrows indicated a significant difference between control and CG33474-expressing cells. Scale bar: 10 μm. Relative peroxisome size (b, n = 10 cells.) and number (c, n = 10 cells.) between control (Ctrl) and CG33474 expressing cells (CG33474) in a. d Fluorescence microscopy images of fat bodies from 3rd instar larvae of the designated genotype that expressed CG33474-Flag (magenta). The boxed areas were enlarged to the lower panel. DAPI (blue) labeled nuclei. Clones were labeled by RFP (red). GFP-PTS1 (green) indicated peroxisomes. Scale bar: 20 μm. Relative peroxisome size (e, n = 10 cells.) and number (f, n = 10 cells.) between control (Ctrl) and CG33474-expressing cells (CG33474) in d. g Fluorescence microscopy images of HEK293T cells transfected with PEX11G-GFP encoding plasmid. The boxed areas were enlarged to the lower panel. DAPI (blue) labeled nuclei. PMP70 (red) indicated peroxisomes. Scale bar: 20 μm. Relative peroxisome size (h, n = 10 cells.) and number (i, n = 10 cells.) between control (Ctrl) and PEX11G expressing cells (PEX11G) in g. Two-tailed Student’s t test was performed. **** p < 0.0001; ns, not significant

To investigate whether CG33474 modulates peroxisome size in vivo, we employed the FLP-out system to co-express CG33474-Flag and GFP-PTS1. Subsequently, we analyzed the impact of CG33474 on peroxisomes using 3rd instar larvae of this specific genotype. The endogenous CG33474-Flag fusion proteins colocalized with GFP-PTS1, a marker of peroxisomes (Fig. 4d). Furthermore, adipocytes expressing CG33474 also exhibited enlarged peroxisome size (Fig. 4e).

To further investigate whether the role of CG33474/PEX11G proteins in modifying peroxisome size is conserved across species, we transfected mammalian HEK293T cells with plasmids encoding CG33474-Flag and PEX11G-GFP, respectively. The results demonstrated that both Drosophila CG33474 (Fig. S6a) and mammalian PEX11G proteins (Fig. 4g) co-localized with the peroxisomal marker PMP70. Moreover, overexpression of either CG33474 (Fig. S6b) or PEX11G proteins (Fig. 4h) in HEK293T cells increased peroxisome size. Notably, overexpression of CG33474 (Fig. 4c, f) or PEX11G (Fig. 4i) did not alter the number of peroxisomes in the respective model organism. However, a slight decrease in the number of peroxisomes was observed upon CG33474 overexpression in HEK293T cells (Fig. S6c). Taken together, our findings proved that the cellular function of CG33474 and PEX11G proteins in altering peroxisome size is evolutionarily conserved between Drosophila and mammals.

Cysteine- and methionine-induced CG33474 expression is primarily in the adipose tissues

As HPD primarily enhances peroxisome levels in the adipose tissues, we examined whether cysteine- and methionine-induced CG33474 expression exhibits adipose specificity. Given that cysteine, a semi-essential amino acid derived from the essential amino acid methionine via the transsulfuration pathway [29], exhibits similar effects to methionine in promoting CG33474 expression, we initially focused on cysteine for preliminary testing. The intensity of red fluorescent protein (RFP) in the adipose tissues of CG33474-Gal4>UAS-RFP female flies increased significantly after a 36-h cysteine feeding, indicative of the fact that cysteine-induced CG33474 expression is predominantly localized to adipose tissues (Fig. S7a–c). This finding was paralleled in CG33474-Gal4>UAS-RFP larvae, where cysteine-induced CG33474 expression was predominantly enriched in the larval fat body (Fig. S7d, e). To solidify this observation, we examined the colocalization of CG33474 with adipose tissues using CG33474-Gal4>UAS-RFP female flies fed with cysteine or methionine for 36 h. Tissue section results revealed that CG33474 proteins co-localized with LipidTOX, a maker of neutral lipids (Fig. 5a; Fig. S7f). Altogether, cysteine- and methionine-induced CG33474 expression is primarily in the adipose tissues.

Fig. 5

CG33474 is required for cysteine- and methionine-induced fat loss. a Images depicting longitudinal sections of CG33474-Gal4>UAS-RFP female flies 36 h following the designated treatments. Suc: 5% sucrose; Suc + Cys: 5% sucrose + 25 mM cysteine; Suc + Met: 5% sucrose + 25 mM methionine. DAPI (blue) labeled nuclei. LipidTOX (green) indicated neutral lipids. Scale bar: 500 μm. b TAG levels in female flies 36 h following the designated treatments. Suc: 5% sucrose; Suc + Cys: 5% sucrose + 25 mM cysteine; Suc + Met: 5% sucrose + 25 mM methionine. w1118 and CG33474 homozygous mutants were used as WT (wild-type) and Mut (mutant), respectively. Data were normalized to protein concentrations. n = 6. c Lipid staining on female flies 36 h following the designated treatments. Suc: 5% sucrose; Suc + Cys: 5% sucrose + 25 mM cysteine; Suc + Met: 5% sucrose + 25 mM methionine. w1118 and CG33474 homozygous mutants were used as WT and Mut, respectively. DAPI (blue) labeled nuclei. LipidTOX (red) indicated neutral lipids. Scale bar: 20 μm. d Relative size of LDs in c. From left to right: 18, 18, 15, 17, 15, and 17 views. e Lipid staining on female flies of the designated genotypes. DAPI (blue) labeled nuclei. LipidTOX (red) indicated neutral lipids. Scale bar: 20 μm. f Relative size of LDs in e. Ctrl: lpp-Gal4>UAS-lacZ; OE: lpp-Gal4>UAS-CG33474-Flag. From left to right: 19 and 17 views. g TAG levels in female flies of the designated genotypes. Ctrl: lpp-Gal4>UAS-lacZ; OE: lpp-Gal4>UAS-CG33474-Flag. Data were normalized to protein concentrations. n = 5. h Lipid staining of fat bodies from 3rd instar larvae of the designated genotypes using Nile Red. DAPI (blue) labeled nuclei. Clones were labeled by GFP (green). Nile Red (magenta) indicated neutral lipids. Arrows indicated fat cells expressing clones. Larvae in the 3rd and 4th columns were treated with 25 mM cysteine (Cys) and 25 mM methionine (Met), respectively, for 36 h. Scale bar: 50 μm. i Relative fluorescence intensities of Nile Red in h. From left to right: 17, 17, 18, 17, 20, 23, 22, 20, 20, 17, 19, 19, 19, and 16 cells. j Representative 3D structural images of Nile Red-stained fat bodies from 3rd instar larvae of the indicated genotypes. Nile Red (magenta) indicated neutral lipids. Scale bar: 25 μm. Two-tailed Student’s t test was performed. * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001; ns, not significant

CG33474 is required for cysteine- and methionine-induced fat loss

In the course of investigating CG33474 expression patterns using LipidTOX, we noticed a significant decline in lipid contents in flies fed with cysteine or methionine (Fig. 5a). This observation prompted us to explore whether cysteine and methionine regulate lipid metabolism in a CG33474-dependent manner. We first confirmed that w1118 female flies fed with a 36-h cysteine or methionine diet exhibited lower lipid contents, as evidenced by measuring their whole-body triglyceride (TAG) levels, an indicator of lipid storage (Fig. S8a, b). Remarkably, CG33474 mutant female flies were protected from cysteine- and methionine-induced decrease in TAG levels (Fig. 5b; Fig. S8c) and lipid droplets (LDs) size (Fig. 5c, d), suggesting that cysteine- and methionine-triggered fat loss requires CG33474. Similarly, in the larval phase, CG33474 mutants also attenuated cysteine- and methionine-induced TAG reduction (Fig. S8d, e). Moreover, fat body-specific overexpression of CG33474 (ppl-Gal4>UAS-CG33474-Flag) of female flies led to a reduced LD size (Fig. 5e, f) and a diminished TAG level (Fig. 5g; Fig. S8f) compared to driver control flies (ppl-Gal4>UAS-lacZ).

We further utilized the FLP-out technique to clonally decrease or increase CG33474 expression in 3rd instar larval fat bodies and visualized LDs through Nile Red staining. Knockdown of CG33474 led to slightly elevated levels of neutral lipids within adipocytes compared to control clones (Fig. 5h, i). Additionally, when CG33474 was knocked down and simultaneously treated with cysteine or methionine for a duration of 36 h also increased neutral lipid levels (Fig. 5h, i), corroborating that cysteine- and methionine-induced lipids decrease requires CG33474. Meanwhile, CG33474-expressing clones significantly reduced neutral lipid levels within adipocytes compared to control clones, suggesting that CG33474 stimulates the breakdown of neutral lipids in a cell-autonomous manner (Fig. 5h, i).

Next, we constructed a Drosophila line (FLP>UAS-GFP-PTS1, UAS-CG33474-Flag) that concurrently overexpressed GFP-PTS1 and CG33474. The results demonstrated that overexpressing CG33474 not only resulted in an elevation in peroxisome levels and a reduction in lipid contents but also facilitated a tight association between LDs and peroxisomes (Fig. 5j; Videos 1, 2), indicating that CG33474-induced fat loss might depend on the biogenesis of peroxisomes. To test this hypothesis, we generated a Drosophila strain (FLP>UAS-CG33474-Flag, UAS-Pex19-RNAi) that simultaneously overexpressed CG33474 and downregulated Pex19, a crucial peroxisome biogenesis factor [30]. We successfully gained a Drosophila Pmp70 antibody, which performed effectively as determined by the co-localization with the green fluorescence emanating from the Ubi-GFP-PTS1 line (Fig. S8g). Then, we employed this Pmp70 antibody to confirm that the knockdown of Pex19 significantly inhibited peroxisome biogenesis (Fig. S8h). Additionally, Nile Red staining revealed that knockdown of Pex19 did not impact neutral lipid levels (Fig. 5h, i). Subsequently, we visualized LDs within the fat bodies of 3rd instar larval from the FLP>UAS-CG33474-Flag, UAS-Pex19-RNAi strain using Nile Red staining. The results demonstrated that suppressing peroxisome biogenesis inhibited CG33474-induced fat loss (Fig. 5h, i), indicating that CG33474-induced fat loss is dependent on peroxisome biogenesis.

Next, we explored the possible effects of other signaling pathways on CG33474-induced fat loss. The Nile Red staining results revealed that the knockdown of bmm (a triglyceride lipase), or ROS scavengers such as Sod1 (located in the cytoplasm) and Sod2 (located in the matrix), failed to abrogate the effect of CG33474 in reducing neutral lipid levels (Fig. S8i, j). This finding suggested that CG33474-induced fat loss is independent of bmm-mediated lipid metabolism process or the Sod-regulated ROS pathway.

Roles of TOR signaling in cysteine- and methionine-induced CG33474 expression

Since a recent study identified that cystine, the oxidized derivative of cysteine, activates TOR signaling in the fat body of fasted Drosophila [31], we conjectured that TOR signaling might be indispensable for cysteine- and methionine-induced CG33474 expression. To probe this hypothesis, we fed w1118 female flies with different compounds for 36 h and assessed the mRNA levels of CG33474 by RT-qPCR. Treatment with the classic TOR allosteric inhibitor rapamycin [32] effectively attenuated the elevation of CG33474 induced by cysteine and methionine (Fig. 6a). Furthermore, exposure to the cell-permeable TOR activator, MHY1485 [33], resulted in a slight increase in CG33474 expression (Fig. 6a).

Fig. 6

Roles of TOR signaling in cysteine- and methionine-induced CG33474 expression. a Relative CG33474 mRNA levels in w1118 female flies 36 h following the designated treatments. Ctrl: H2O; MHY: 100 μΜ MHY1485; Cys: 25 mM cysteine; Cys + Rapa: 25 mM cysteine + 200 μΜ rapamycin; Met: 25 mM methionine; Met + Rapa: 25 mM methionine + 200 μΜ rapamycin. n = 3. b Fluorescence microscopy images of CG33474-Gal4>UAS-GFP 3rd instar larval fat bodies cultured under the indicated conditions. NM: normal Schneider medium; NM + MHY: normal Schneider medium containing 10 μΜ MHY1485; NM + Rapa: normal Schneider medium containing 10 μΜ rapamycin. DAPI (blue) labeled nuclei. Scale bar: 100 μm. T: time. c Relative fluorescence intensities of GFP in b. From left to right: 21, 36, 37, and 33 views. d Relative CG33474 mRNA levels in female flies of the designated genotypes after 36 h of feeding with the indicated treatments. Suc: 5% sucrose; Suc + Cys: 5% sucrose + 25 mM cysteine; Suc + Met: 5% sucrose + 25 mM methionine. n = 3. e Western blotting showing the levels of p-4EBP and np-4EBP in w.1118 female flies under the indicated treatments for 36 h. Suc: 5% sucrose; Suc + YE: 5% sucrose + 30% yeast extract; Suc + Cys: 5% sucrose + 25 mM cysteine; Suc + Met: 5% sucrose + 25 mM methionine. Two-tailed Student’s t test (a, d) or one-way ANOVA (c) were performed. * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001

To further elucidate the roles of TOR signaling in CG33474 expression, we ex vivo cultured fat bodies isolated from CG33474-Gal4>UAS-GFP 3rd instar larvae for a duration of 36 h with different chemicals. The results showed that incubating the fat bodies in Schneider medium significantly induced CG33474 expression as determined by the heightened green fluorescence intensity, and the addition of MHY1485 further potentiated this upregulation (Fig. 6b, c). On the other hand, rapamycin substantially curtailed the upregulation of CG33474 compared to Schneider medium vehicle exposure (Fig. 6b, c).

Ultimately, we employed a genetic approach to substantiate the preceding findings. We collected flies that had been subjected to various diets and analyzed the mRNA levels of CG33474 using RT-qPCR. The results demonstrated that the inhibition of the TOR pathway by knocking down Rheb in the fat bodies (lpp-Gal4>UAS-Rheb-RNAi) effectively diminished cysteine- and methionine-induced CG33474 expression (Fig. 6d). Conversely, overexpression of Rheb in the fat bodies (lpp-Gal4>UAS-Rheb) to activate the TOR pathway was adequate to trigger CG33474 expression even in the absence of cysteine and methionine (Fig. 6d). Together, these data suggested that TOR signaling is required for cysteine- and methionine-induced CG33474 expression.

In addition, we observed that while TOR signaling activity was significantly increased in flies fed with HPD, as determined by elevated eIF4E-binding protein (4EBP) phosphorylation level, it remained unchanged in cysteine- and methionine-treated flies (Fig. 6e). These data showed that while TOR activity promotes CG33474 expression, and a basal level of TOR activity is required for the induction of CG33474. However, an increase in TOR signaling is not required for the upregulation of CG33474. Taken together, our data suggested that TOR serves as a regulator, yet some unidentified amino acid sensing pathway might be directly responsible for the induction of CG33474.

The functions of PEX11G are evolutionarily conserved

Finally, we investigated whether the induction of CG33474/PEX11G is evolutionarily conserved. Mammalian HEK293T cells were cultured in the normal medium supplemented with varying concentrations of cysteine or methionine (with 25 mM cysteine excluded due to its severe cytotoxic effect). Following a 36-h incubation, we quantified PEX11G mRNA levels using RT-qPCR. The results revealed that supplementation with 15 mM cysteine or methionine moderately promoted PEX11G expression in HEK293T cells (Fig. S9a). Analogous to the findings in Drosophila, MHY1485 also increased PEX11G expression, while rapamycin attenuated cysteine- and methionine-induced PEX11G expression in HEK293T cells (Fig. S9b). However, it is noteworthy that the induction of PEX11G by cysteine and methionine in HEK293T cells was less prominent than that in flies. It is possible that stronger effects might be triggered in certain adipocyte cell lines as cysteine- and methionine-induced CG33474 expression is primarily in the adipose tissues in flies.

To assess whether the function of PEX11G in lipid metabolism is evolutionarily conserved, we transfected HEK293T cells with a plasmid encoding PEX11G-GFP and subsequently evaluated the LD index. PEX11G-expressing cells exhibited a decrease in both the size and number of LDs (Fig. 7a–c), indicating an accelerated rate of lipid breakdown.

Fig. 7

The functions of PEX11G are evolutionarily conserved. a Fluorescence microscopy images of HEK293T cells transfected with PEX11G-GFP (green) encoding plasmid and treated with 100 μΜ OA. DAPI (blue) labeled nuclei. LipidTOX (red) indicated neutral lipids. Dotted lines marked the outlines of PEX11G-expressing cells. Scale bar: 20 μm. Relative size (b) and number (c) of LDs between control (Ctrl) and PEX11G-expressing cells (PEX11G) in a. n = 11. d Fluorescence microscopy images of peroxisome-LD contacts in U2OS cells transfected with PEX11G-GFP (green) encoding plasmid or not and treated with 100 μΜ OA. DAPI (blue) labeled nuclei. LipidTOX (magenta) indicated neutral lipids. Scale bar: 10 μm. e Fluorescence microscopy images of peroxisome-LD contacts in U2OS cells transfected with PEX11G-GFP (green) encoding plasmid and treated with 100 μΜ OA. The boxed areas were enlarged to the lower panel. LipidTOX (blue) indicated neutral lipids. PMP70 (red) indicated peroxisomes. Scale bar: 2 μm. f A live cell image depicting U2OS cells transfected with PEX11G-GFP (green) encoding plasmid. The boxed area was enlarged to the right. LipidTOX (red) indicated neutral lipids. Scale bar: 10 μm. g Fluorescence profiles derived from the indicated dotted line scan in f for 60 min. Two-tailed Student’s t test was performed. ** p < 0.01; *** p < 0.001

Next, we investigated the interplay between PEX11G and LDs, as previous studies reported that peroxisomes and LDs are in physical contact within cells [34]. We used the U2OS cell line, which is frequently employed for live imaging of organelle dynamics due to its flat structure. We observed that, in un-transfected U2OS cells stained with PMP70 antibody, peroxisomes exhibited little association with LDs (Fig. 7d). In contrast, peroxisomes in PEX11G-expressing U2OS cells showed a strong association with LDs (Fig. 7d). Furthermore, this interaction is not caused by mis-localization of PEX11G-GFP onto the surface of LDs, since the identical PEX11G-GFP proteins colocalized with the peroxisome marker PMP70 (Fig. 7e), suggesting that PEX11G facilitates the association between peroxisomes and LDs. A similar association between peroxisomes and LDs triggered by PEX11G was also observed in HEK293T cells, albeit to a lesser extent (Fig. S9c; Videos 3, 4). This finding is consistent with our previous discovery that peroxisomes exhibit increased attachment to LDs in larval fat cells (Fig. 5j). The peroxisomes-LDs interaction is further supported by the live-imaging experiment, as PEX11G-tagged peroxisomes tightly associated with LDs for more than 1 h (Fig. 7f, g; Video 5). Altogether, these data strongly suggested that PEX11G promotes inter-organelle contacts between peroxisomes and LDs, potentially accounting for PEX11G-induced fat loss.

Then, we investigate whether PEX11G regulates peroxisomal biosynthesis, import, and assembly. To this end, we overexpressed PEX11G-GFP fusion proteins in HEK293T cells and conducted western blotting to assess the levels of key peroxisomal proteins. The results revealed that PEX11G did not affect proteins that govern peroxisomal morphology, including Dynamin-related protein 1 (Drp1) and Fission factor 1 (Fis1) [30] (Fig. S10a). Additionally, the protein levels of PMP70, PEX5, and PEX19 remained unaltered in PEX11G-expressing cells (Fig. S10b). These results proved that PEX11G does not modulate the biosynthesis, import, or assembly of peroxisomes. Furthermore, the protein levels of Sod1, Catalase, and ACOX1 (a rate-limiting enzyme for peroxisomal β-oxidation [35]) remained unchanged in PEX11G-expressing cells (Fig. S10b). Additionally, PEX11G expression did not colocalize with PEX5, PEX19, or Catalase in HEK293T cells (Fig. S10c–e).

Given the crucial role of peroxisomes in the detoxification of ROS [