All animal experiments were performed according to procedures approved by the animal welfare committees of the government of Upper Bavaria, Germany (ROB-55.2-2532.Vet_02-20-32) and performed in compliance with German Animal Welfare Laws. Animals were housed in individually ventilated cages at room temperature (22°C) with a 12-h light–dark cycle. All mouse housing and husbandry occurred on standard chow diet (Ssniff). We used 12 weeks old C57BL/6J (Janvier) wild-type mice, which we housed at 30°C for 7 days for thermoneutral acclimatization. For cold exposure and cold adaptation, mice were housed at 4°C for 24 h and 1 week, respectively.

We used an immortalized brown preadipocyte cell line (imBAT) (11), which we differentiated into mature brown adipocytes in vitro. The cells were induced at confluence for 2 days and differentiated for 2 days followed by 1 day of cultivation in standard medium [DMEM GlutaMax (Gibco), 10% fetal bovine serum (Sigma-Aldrich), 1% penicillin-streptomycin (Sigma-Aldrich)]. For induction the standard medium was supplemented with 850 nM human insulin (Sigma-Aldrich), 1 µM dexamethasone (Sigma-Aldrich, in 100% ethanol), 1 µM T3 (Sigma-Aldrich, in 1 M NaOH), 1 µM rosiglitazone (Cayman Chemicals, in 100% DMSO), 500 nM IBMX (Sigma-Aldrich, in 100% DMSO) and 125 nM indomethacin (Sigma-Aldrich, in 100% ethanol). The differentiation medium contained 1 µM T3 and 1 µM rosiglitazone. All treatments were performed on the 5^th day of differentiation. Proteasome inhibitors bortezomib (Selleckchem), epoxomicin (Millipore) and MG-132 (Calbiochem) were used at 100 nM in 100% DMSO and incubated for 6 h. For β3-adrenergic stimulation, cells were treated with 1 µM CL316,243 (Tocris, in distilled H₂O) for 6 h.

For primary cell experiments, mature adipocytes were differentiated from preadipocytes, isolated from the stromal-vascular fraction of adipose tissue of 4 weeks old C57BL/6J (Janvier) wild-type mice. The mice were sacrificed by cervical dislocation, the respective adipose tissues were harvested and pooled. The tissues were minced, weighted and digested in DMEM/F-12 (Sigma-Aldrich), 1% PenStrep (Sigma-Aldrich), 15 mg/mL fatty acid free BSA (Sigma-Aldrich), 1 mg/mL collagenase type 2 (Worthington) and 0.1 mg/mL DNase 1 (Roche) for 30 – 45 min at 37°C. For BAT, the digestion mix was supplemented with 1.2 U/mL Dispase (Roche). The digestion was stopped by adding DMEM/F-12 (+10% FBS, +1% PenStrep) in a 1:5 ratio. The digest was filtered through a 100 µm strainer and centrifuged at room temperature for 10 min at 500 xg. The mature adipocyte and the supernatant were aspirated, the pellet suspended in DMEM/F-12 (+10% FBS, +1% PenStrep) and filtered through a 70 µm strainer. After another centrifugation step at room temperature for 10 min at 500 xg, the pellet was again suspended in DMEM/F-12 (+10% FBS, +1% PenStrep), filtered through a 30 µm strainer and eventually plated in T75 flasks. The medium was changed the next day, to remove debris from the digestion and then every other day until the cells reached confluency. To differentiate the preadipocytes into mature adipocytes, cells were induced for 2 days and differentiated for 2 days followed by 1 day cultivation in standard medium. The induction medium consisted of DMEM/F-12 (+10% FBS, +1% PenStrep) supplemented with 340 nM human insulin (Sigma-Aldrich), 1 µM dexamethasone (Sigma-Aldrich, in 100% ethanol), 1 µM T3 (Sigma-Aldrich, in 1 M NaOH), 1 µM rosiglitazone (Cayman Chemicals, in 100% DMSO), 500 nM IBMX (Sigma-Aldrich, in 100% DMSO). The differentiation medium for white adipocytes from gonadal white adipose tissue (GWAT) consisted of DMEM/F-12 (+10% FBS, +1% PenStrep) with 10 nM human insulin and 2 µM T3. For beige and brown adipocytes from subcutaneous adipose tissue (SCAT) and BAT, respectively, DMEM/F-12 (+10% FBS, +1% PenStrep) was supplemented with 10 nM human insulin, 2 µM T3 and 1 µM rosiglitazone.

For RNA interference experiments, we reverse transfected imBAT and primary brown adipocytes with SMARTpool siRNA (Dharmacon) on the 3^rd day of differentiation. Lipofectamine™ RNAiMAX transfection reagent (Thermo Fisher Scientific) was used according to manufacturer’s instructions. SMARTpool siRNAs for Nfe2l1, Ubqln1, 2 and 4 were used in a final concentration of 30 nM for single knockdowns. For double and triple knockdown of Ubqln1, 2 and 4 SMARTpool siRNAs were mixed in an equimolar ratio with final concentrations of 60 nM and 90 nM, respectively. After 24 h the transfection mix was replaced with the standard medium and the cells were incubated for another 24 h, after which we treated or directly harvested them.

To extract total RNA from frozen adipose tissue and cells we used the NucleoSpin^® RNA kit (Macherey-Nagel) according to the manufacturer’s instructions. RNA concentrations were measured on a NanoDrop spectrophotometer (Thermo Fisher Scientific). To synthesize complementary DNA (cDNA), we reverse-transcribed 500 ng RNA with the Maxima™ H Master Mix 5x (Thermo Fisher Scientific) in a total volume of 10 μl. cDNA was diluted 1:40 with RNase-free H₂O. Relative gene expression was quantified using quantitative real time-PCR. Each reaction contained 4 µL cDNA, 5 µL PowerUp™ SYBR Green Master Mix (Applied Biosystems) and 1 µL of 5 µM primer stock (full primer list in Supplementary Table 1 ). We used standard run conditions for Applied Biosystems SYBR Green Gene Expression Assays (2:00 min 50°C, 10:00 min 95°C, 40 cycles of 0:15 min 95°C, 1:00 min 60°C). Cycle thresholds (Cts) of genes of interest were normalized to TATA-box binding protein (Tbp) levels by the ΔΔCt-method and displayed as relative copies per Tbp or relative expression normalized to experimental control groups.

To visualize and quantify lipid content in adipocytes, we used Oil Red O staining. Confluent cells were harvested 48 h after transfection, washed once with DPBS and incubated with zinc formalin solution (Merck) for 5 min at room temperature. Afterwards, zinc formalin was replaced with fresh zinc formalin and the cells were fixed for 48 h at 4°C. Adipocytes were washed with 100% isopropanol and dried at 37°C. Cells were stained with 0.5 ml 60% vol/vol Oil Red O solution (Sigma), 40% vol/vol distilled H₂O for 1 h at room temperature and immediately washed 4x with distilled H₂O afterwards. Pictures from each well were taken with a digital camera as well as with a microscope at 200x magnification before ORO stain was eluted with 100% isopropanol and transferred into a clear 96-well plate. To quantify the staining, absorption of the eluates was measured at 500 nm on a Tecan plate reader. Absorption was normalized by subtracting the absorbance of 100% isopropanol and displayed as fold change to Scramble.

For extracellular flux analysis, we reverse transfected imBAT and plated them at 25,000 cells per well in a 24-well Seahorse plate in two independent experiments. To ensure even distribution, cells were seeded in 50 µl of medium and kept for 2 h at room temperature. Then, 175 µl medium were added, and the cells were allowed to fully attach at 37°C overnight. On the day before the experiment, Seahorse cartridges with the sensors were equilibrated in XF calibrant solution (Agilent) at 37°C in a non-CO2 chamber overnight. The experiment was performed on the 5th day of differentiation. Cells were treated with 100 nM bortezomib for 6 h or the corresponding amount of DMSO (Sigma-Aldrich) as a control treatment. XF DMEM pH 7.4 medium (Agilent) was supplemented fresh on the day of the experiment with 10 mM glucose, 1 mM pyruvate and 2 mM L-glutamine (Seahorse medium, all reagents from Agilent). Cells were carefully washed 2x with 1 ml Seahorse medium before incubating them with 500 µl of Seahorse medium for approximately 45 min at 37°C without CO₂. During the measurement norepinephrine (Sigma-Aldrich), oligomycin (Sigma-Aldrich), FCCP (Sigma-Aldrich) and rotenone–antimycin A (Sigma-Aldrich) were injected via the port of the Seahorse cartridge. Final concentrations were 1 µM norepinephrine (in 100% H₂O), 1 µM oligomycin, 4 µM FCCP and 0.5 µM rotenone–antimycin A (all in 100% DMSO). The protocol was 3 min mix, 2 min wait and 3 min measure for each time point. For normalization of respiration to live cell count, the cell plate was retrieved after the run, and we quantified the cell count using the CyQUANT^® Cell Proliferation Assay Kit (Thermo Fisher Scientific) according to manufacturer’s instructions.

We lysed cells in RIPA buffer [150 mM NaCl (Merck), 5 mM EDTA (Merck), 50 mM Tris pH 8 (Merck), 0.1% wt/vol SDS (Carl Roth), 1% wt/vol IGEPAL^® CA-630 (Sigma-Aldrich), 0.5% wt/vol sodium deoxycholate (Sigma-Aldrich)] freshly supplemented with protease inhibitors (Sigma-Aldrich) in a 1:100 ratio. The lysate was centrifuged at 4°C for 30 min at 21,000xg, to clear the lysate from debris and lipids. Protein concentrations were determined using the Pierce BCA Protein Assay (Thermo Fisher Scientific) according to the manufacturer’s instructions. For western blotting, lysates were adjusted to a final concentration of 0.85 µg/µl in 1x Bolt™ LDS Sample buffer (Thermo Fisher Scientific) supplemented with 5% vol/vol 2-mercaptoethanol (Sigma-Aldrich). For SDS-PAGE we used Bolt™ 4–12% Bis-Tris gels (Thermo Fisher Scientific) with Bolt™ MOPS SDS running buffer. After separation, proteins were transferred onto a 0.2 µm PVDF membrane (Bio-Rad) using the Trans-Blot^® Turbo™ system (Bio-Rad) at 12 V, 1.4 A for 16 min. The membrane was briefly stained with Ponceau S (Sigma-Aldrich) to verify successful transfer and blocked in 1x Roti-Block (Carl Roth) for 1 h at room temperature. Primary antibodies (Cell signaling) were used in a 1:1,000 ratio in 1x Roti-Block overnight at 4°C. After washing 4× for 10 min with TBS-T (200 mM Tris (Merck), 1.36 mM NaCl (Merck), 0.1% vol/vol Tween 20 (Sigma)), secondary antibodies (Santa Cruz) were applied in a 1:10,000 ratio in Roti-Block for 1 h at room temperature. Then membranes were washed again 3× for 10 min in TBS-T and developed using SuperSignal West Pico PLUS Chemiluminescent Substrate (Thermo Fisher Scientific) and a Chemidoc imager (Bio-Rad). We analyzed the digital images with the Image Lab software (Bio-Rad). Uncropped images can be found in ( Supplementary Figure 1 ).

For the proteasomal activity assay, we lysed cells with a buffer containing 40 mM TRIS pH 7.2 (Merck), 50 mM NaCl (Merck), 5 mM MgCl₂(hexahydrate) (Merck) and 10% vol/vol glycerol (Sigma-Aldrich) freshly supplemented with 2 mM β-mercaptoethanol (Sigma-Aldrich) and 2 mM ATP (Sigma-Aldrich). To measure chemotrypsin-like, trypsin-like and caspase-like activity, we used the Proteasome Activity Fluorometric Assay (UBP Bio) according to the manufacturer’s instructions. For the normalization of proteasomal activity to total protein content, we used the Bio-Rad Protein Assay Kit II to determine protein concentration in our samples.

Data were expressed as the mean ± standard error of the mean (SEM). For comparing three and more groups we used one-way ANOVA with Bonferroni post-hoc test as indicated in the figure legends. For comparing two variables we used two-way ANOVA with Bonferroni post-hoc test. Analysis was performed using R, Microsoft Excel and/or GraphPad Prism. P <0.05 was considered significant, as indicated by asterisks and letters in the figures legends.

As nothing is known about ubiquilins in BAT, we first investigated their gene expression levels in primary and immortalized adipocytes as well as in BAT ex vivo and second how these are impacted by cold and proteasome inhibition. In primary cells from intrascapular brown, inguinal subcutaneous and gonadal white adipose tissue (WAT), Ubqln1, 2 and 4 expression was found to be expressed whereas Ubqln3 was undetectable ( Figure 1A ). Ubqln2 and Ubqln4 were expressed at similar levels in all three adipocyte types. Interestingly, Ubqln1 was approximately higher expressed compared to Ubqln2 and 4 with highest expression in adipocytes from the subcutaneous WAT. In BAT from 12-week-old mice, which were kept at 4°C for 24 h (cold exposure) and 7 days (cold adaptation), we found that Ubqln1 and Ubqln4 expression was higher compared to mice kept at 30°C (thermoneutrality) ( Figure 1B ). We confirmed that the cold challenge was successful by increased expression of thermogenic markers Ucp1 and Ppargc1a ( Supplementary Figures 2A, B ). Compared to primary brown adipocytes, we found a similar expression pattern in imBAT, which remained unchanged upon stimulation with beta3-agonist CL316,243 ( Figure 1C ). Proteasome inhibition with epoxomicin, MG132 or bortezomib respectively, resulted in higher Ubqln1 expression compared to DMSO treated controls ( Figure 1D ). We were unable to detect Ubqln3 mRNA in any of our samples, which is in line with a previous report that Ubqln3 is specific to testis (9). Thus, Ubqln1, 2 and 4 are expressed in adipose tissue, yet only Ubqln1 is positively regulated by cold and proteasome inhibition.

Next, we determined the significance of ubiquilins in brown adipocyte homeostasis. We used RNA interference to silence the gene expression of Ubqln1, 2 and 4 in immortalized and primary adipocytes to investigate their functional relevance. We explored established surrogate markers of stress responses and adipocyte function linked to BAT proteostasis (4). While our RNAi strategy was efficient, knockdown of any of the Ubqln genes did not result in changes in the mRNA expression of surrogate stress markers, Ddit3 and Ccl2 ( Figures 2A, B ). Knockdown of Ubqln2 resulted in lower Adipoq expression in both immortalized and primary brown adipocytes, which suggests that adipocyte health is affected under these conditions. Since Ubqln1, 2 and 4 share the same domain structure and their sequences are highly similar, we hypothesized that they could compensate for each other (12). This was also supported by upregulation of Ubqln2 after knockdown of Ubqln1 in primary brown adipocytes ( Figure 2B ). To circumvent this effect, we studied all possible combinations of Ubqln1, 2 and 4 siRNAs in single, double and triple Ubqln knockdown experiments in imBAT ( Figure 2C ). This resulted in lower Adipoq expression in conditions with Ubqln2 knockdown compared to the control cells ( Figure 2D ). Ddit3 and Ccl2 expression remained largely unaffected. Hence, loss of Ubqln1, 2 or 4 did not increase ER stress or inflammation marker genes, alone or in combination.

As we observed a reduction in Adipoq mRNA levels in both immortalized and primary brown adipocytes, we next investigated how loss of ubiquilins impacts adipocyte health. To our surprise, the expression levels of other adipocyte marker genes such as Pparg, Plin1, Fabp4, Fasn and Cebpa remained unchanged by Ubqln single or triple knockdown ( Figure 3A . However, compared to the control cells we observed a lower lipid content in imBAT ( Figures 3B, C and Supplementary Figure 3 ). Next, we investigated if loss of ubiquilins impacts NST. We analyzed norepinephrine-induced respiration and mitochondrial stress tests as the gold standard of measuring NST in cells. Neither single nor triple knockdown affected basal cellular respiration, the response to norepinephrine stimulation or maximal respiratory capacity ( Figures 3D, E ). To investigate if ubiquilins participate in the coordinated cellular response to proteasome inhibition, which has been shown to impair mitochondrial function in brown adipocytes (4), we treated imBAT cells with the compound bortezomib. While pretreatment with bortezomib led to lower levels of respiration, this response was independent of single or triple knockdown of ubiquilins ( Figures 3F, G ).

As we found that Ubqln1 expression in brown adipocytes is strongly induced by proteasome inhibition, we hypothesized that Ubiquilins are involved in maintaining proteostasis. Indeed, knockdown of Ubqln1, 2 and 4 resulted in higher Atf3 gene expression, albeit the effect was small but significant ( Figure 4A ). Other stress markers related to accumulation of misfolded proteins or ER stress such as Atf4, Atf6 and sXbp1 were unchanged and Hspa5 was only higher in the triple knockdown condition. Another important player for brown fat proteostasis is the transcription factor Nfe2l1, which activates the transcription of proteasomal subunits in a bounce-back-mechanism in response to proteasomal inhibition and cold (4). It is conceivable that Nfe2l1 counteracts accumulation of misfolded proteins and ER stress as a consequence of diminished Ubqln expression through increased proteasome formation. However, we neither observed differences in Nfe2l1 expression nor in the expression of two of its downstream targets Psma1 and Psmb1. We also tested whether, the induction of Atf3 in response to Ubqln knockdown was mediated by Nfe2l1, but Atf3 expression was unchanged by Ubqln triple knockdown in combination with Nfe2l1 knockdown ( Supplementary Figure 4 ). In line with these findings, single Ubqln knockdown ( Supplementary Figure 5 ) or triple Ubqln knockdown neither affected the levels of ubiquitinated proteins nor Nfe2l1 protein levels, neither under basal conditions nor after proteasome inhibition ( Figure 4B ). Triple knockdown did also not result in lower proteasomal activity ( Figure 4C ). Thus, ubiquilins are dispensable to maintain proteostasis in brown adipocytes under basal and proteasome inhibited conditions and their absence is not linked to Nfe2l1 activation.