POMC Neuron BBSome Regulation of Body Weight is Independent of its Ciliary Function

Abstract The BBSome, a complex of several Bardet-Biedl syndrome (BBS) proteins including BBS1, has emerged as a critical regulator of energy homeostasis. Although the BBSome is best known for its involvement in cilia trafficking, through a process that involve BBS3, it also regulates the localization of cell membrane receptors underlying metabolic regulation. Here, we show that inducible Bbs1 gene deletion selectively in proopiomelanocortin (POMC) neurons cause a gradual increase in body weight, which was associated with higher fat mass. In contrast, inducible deletion of Bbs3 gene in POMC neurons failed to affect body weight and adiposity. Interestingly, loss of BBS1 in POMC neurons led to glucose intolerance and insulin insensitivity, whereas BBS3 deficiency in these neurons is associated with slight impairment in glucose handling, but normal insulin sensitivity. BBS1 deficiency altered the plasma membrane localization of serotonin 5-HT2C receptor (5-HT2CR) and ciliary trafficking of neuropeptide Y2 receptor (NPY2R).In contrast, BBS3 deficiency, which disrupted the ciliary localization of the BBSome, did not interfere with plasma membrane expression of 5-HT2CR, but reduced the trafficking of NPY2R to cilia. We also show that deficiency in BBS1, but not BBS3, alters mitochondria dynamics and decreased total and phosphorylated levels of dynamin-like protein 1 (DRP1) protein. Importantly, rescuing DRP1 activity restored mitochondria dynamics and localization of 5-HT2CR and NPY2R in BBS1-deficient cells. The contrasting effects on energy and glucose homeostasis evoked by POMC neuron deletion of BBS1 versus BBS3 indicate that BBSome regulation of metabolism is not related to its ciliary function in these neurons.


Introduction
2][3] Strong evidence points to the central nervous system in obesity susceptibility, 4 , 5 which is consistent with the critical role of the brain, particularl y the hypothalam us, in the r e gulation of body w eight by inte gr ating the external and internal signals and enacting appropriate and consequential metabolic and behavioral responses to maintain energy balance. 6 , 7Proopiomelanocortin (POMC) neur ons ar e essential for energy homeostasis due to their importance in mediating the effects of various signals. 8r oopiomelanocortin pr oducts such as α-melanocyte stim ulating hormone , whic h acti v ates the melanocortin 4 r ece ptor (MC4R), have potent and long-lasting anorectic and weightreducing effects.While several molecules in POMC neurons have been identified as playing a key role in the control of energy homeostasis, our understanding of the molecular basis of metabolic regulation by POMC neurons remain incomplete.
Bardet-Biedl syndrome (BBS) proteins have emerged as key components in metabolic control. 9The requirement of BBS pr oteins for meta bolic homeostasis is supported by the high pr ev alence of obesity and other metabolic disturbances including type 2 diabetes in individuals that lack functional BBS proteins. 10 , 11A total of 8 conserved BBS proteins (BBS1, BBS2, BBS4, BBS5, BBS7, BBS8, BBS9, and BBS18) interact together to form a complex, termed BBSome. 12BBSome assemb l y r elies on another protein complex, referred to as BBS chaperonin complex, which contains 3 BBS proteins (BBS6, BBS10, and BBS12). 13][16] For instance, the BBSome was found to mediate the ciliary localization of r ece ptors implicated in metabolic regulation such as the neur ope ptide Y2 r ece ptor (NPY 2 R). 17 Such findings have led to the notion that BBS-associated obesity is caused by loss of ciliary localization of these receptors. 18 , 19However, the BBSome has been implicated in many cellular processes not related to cilia including the delivery to the plasma membrane of r ece ptors inv olv ed in the r egulation of energy homeostasis such as the serotonin 5-HT2C receptor (5-HT 2C R). 16 , 20 , 21 More r ecentl y, w e demonstr ated the importance of the BBSome for the regulation of mitochondria function by modulating the activity of dynamin-like protein 1 (DRP1). 224][25] Thus, loss of BBS3 interrupt the ciliary localization of the BBSome.Interestingly, Bbs3 gene null mice do not phenocopy many of the features displayed by BBSome-deficient mice. 26In particular, Bbs3 null mice exhibited a modest increase in body weight and fat mass whereas mice lacking elements of the BBSome display overt obesity and associated diseases including type 2 diabetes. 11 , 21 , 27 , 28These results suggest that BBSome regulation of body weight and energy homeostasis is not related to its function in cilia.In the current study, we used in vivo and in vitro systems to assess the effects of disrupting ciliary localization of the BBSome through deletion of BBS3 compared to disruption of the complex itself through BBS1 deletion.For the in vivo studies, we elected to target POMC neurons since the obesity phenotype was largely recapitulated by disruption of the BBSome, though constituti v e Bbs1 gene a b lation, in these neurons. 16The use of inducible POMC Cre mouse model allowed us to determine whether the obesity phenotype evoked by congenital BBSome deficiency is recapitulated by the disruption of this complex in adult stage.

Cell Culture and Constructs
We used a CRISPR-Cas9 system to produce cell lines lacking BBS1 or BBS3 gene.Specifically, pSpCas9-BBS1g2 and pSpCas9-BBS3g2 were used to generate the retinal pigment epithelial RPE1-BBS1 −/ − and RPE1-BBS3 −/ − cells, respecti v el y.The guide sequence used for BBS1 is CCTTTGA GCA C-CTTCAGGCG and for BBS3 is GTGTTTGAC ATGTC AGGTC A. Sta b le clones were selected with blasticidine and indels were identified by Sanger sequencing.It should be noted that wild-type RPE1 cells deri v ed fr om the same clones as the n ull cells wer e used as contr ols.Sta b le clones wer e also v erified by Western b lot.Human NPY 2 R and 5HT 2C R were tagged with a green fluorescent protein (GFP) at C-terminus as descripted previously. 16

Animals
All animal testing was performed based on guidelines set forth by the National Institutes of Health and appr ov ed by the Uni v ersity of Iowa Animal Care and Use Committee.Mice were housed in groups of 3-5 per cage and maintained on 12-h light-dark cycle with lights on at 6 am .Room temperatur e w as maintained at 22 • C. Food and water were available ad libitum except when the mice were fasted as indicated below.
Bbs1 fl/fl mice that carry floxed alleles around exon 3 of the Bbs1 gene were generated pr eviousl y. 29 Bbs3 fl/fl mice were generated by inserting LoxP sites around exon 5 of the Bbs3 gene using an EuMMCR targeting vector (PG00267 Y 7 A07).Cr e-r ecombinase leads to the excision of exon 5. Female Bbs1 fl/fl and Bbs3 fl/fl mice wer e cr ossed with POMC CreERT2 mice 30 (provided by Dr. Joel Elmquist, UT Southwestern) to generate POMC CreERT2 /Bbs1 fl/fl and POMC CreERT2 /Bbs3 fl/fl mice, respecti v el y.Littermate Bbs1 fl/fl and Bbs3 fl/fl mice were used as contr ols.To la bel POMC neur ons, POMC CreERT2 /Bbs1 fl/fl and POMC CreERT2 /Bbs3 fl/fl mice were further crossed with the ROSA (Stop fl/fl -tdTomato) r e porter transgenic mice (J ackson La borator y, strain n umbers 007914).Cr e r ecombination r emov es the stop site, leading to the expression of the fluorescent tdTomato protein in POMC neurons.
The genotype of the mice was determined by polymerase chain r eaction (PCR) anal ysis of tail DNA using the following conditions: 95 • C for 30 s, 60 • C for 30 s, and 72 • C for 30 s for a total of 34 cycles.Primers used for genotyping are provided in the Ta b le S1 .To induce Cr e r ecombinase expr ession, 6-wkold POMC CreERT2 /Bbs1 fl/fl and POMC CreERT2 /Bbs3 fl/fl mice and their littermate controls (Bbs1 fl/fl and Bbs3 fl/fl mice) were administrated with 75 mg/kg of tamoxifen for 5 d via intraperitoneal (IP) injection.In the body weight study, a small subset of male POMC CreERT2 /Bbs1 fl/fl mice ( n = 4) treated with vehicle (corn oil) was used as an additional control.

Validation of Deletion of Bbs1 and Bbs3 Gene in POMC Neurons
Single cell RNA PCR was used to verify selective Bbs1 and Bbs3 gene deletion in POMC neurons of the hypothalamic arcuate nucleus.A 3-mo-old male POMC CreERT2 /Bbs1 fl/fl /ROSA and POMC CreERT2 /Bbs3 fl/fl /R OSA mice wer e intraperitoneall y injected with tamoxifen (75 mg/kg) for 5 consecuti v e da ys.Ten da ys after the first tamoxifen injection, mice were sacrificed and perfused with a cold cutting solution containing (in m m ) 92 NMDG, 2.5 KCl, 1.2 NaH 2 PO 4 , 30 NaHCO 3 , 20 HEPES, 25 glucose, 5 sodium ascorbate, 2 thiourea, 3 sodium pyruvate, 10 MgSO 4 , and 0.5 CaCl 2 .Retrieved brains were submerged in a chilled and aerated (95% O 2 , 5% CO 2 ) cutting solution before sections (280 μm) containing the hypothalamic arcuate nucleus were obtained using a vibr atome .Br ain slices w ere tr ansferred to an aerated and heated (33

Analysis of Body Weight, Adiposity, and Food Intake
Body weight analysis was performed by weighing mice once a week for 20 wk, from weaning.Nuclear magnetic resonance (NMR, LF50, Bruker minispec) was used to measure body composition (fat mass and lean mass) in mice.At the end of experiments, mice were sacrificed and various fat pads (interscapular br own, perir enal, gonadal, and inguinal fat pads), li v er, and kidneys were dissected and weighed.To measure food intake, mice were housed in individual cages.After 3 d of acclimation to individual housing, daily and cum ulati v e food intake wer e measur ed over a 4-d period at 20 wk of age.

Glucose and Insulin Tolerance Tests
Glucose tolerance test was performed in overnight fasted mice.Blood samples were taken from the tail to measur e b lood glucose at baseline before mice were injected with glucose (2 mg/kg body w eight, IP, Sigma-Aldric h).To test for the glucose-reducing effect of insulin, mice were fasted for 5 h.After determining baseline glucose levels, mice were injected with insulin (0.5 unit/kg body weight, IP, No vo Nor disk).Blood glucose was measured at 15, 30, 60, and 120 min after injection of glucose or insulin.Blood glucose levels were determined using a glucometer (OneTouch Ultra 1, LifeScan Inc.).

Cell Transfection, Infection, and Immunofluorescence
RPE1, RPE1-BBS1 −/ − , and RPE1-BBS3 −/ − cells were gown in DMEM/F12 medium.The cells were seeded in 24 wells with glass cover slide the day before transfection.Density of the cells was 95% confluent in order to promote cilia formation at transfection.Lipofectamine 3000 transfection r ea gent w as used according to man ufactur e's instructions.Forty hours post transfection, the cells were serum starv ed ov ernight.RPE1-BBS1 −/ − cells infected with an adeno-associated virus 2/5 expressing the mouse Bbs1 gene (AA V -Bbs1 , ∼100 particles/cell) for 48 h were also subject to overnight serum starvation.The cells were then fixed with 4% paraformaldehyde (PFA) for 20 min in room temper ature , bloc ked with bloc king solution containing 5% normal goat serum and 0.1% Triton X-100 in phosphate-buffered saline (PBS) before incubation with the primary antibody (1:250), at 4 • C overnight.The cells were washed with PBS for 10 min, 3 times followed by incubation with secondary antibody, diluted at 1:2000, in room temperature for 1 h.The cells were washed again with PBS for 10 min, 3 times befor e ther e wer e mounted with VECTASHIELD mounting solution with DAPI (4 ,6-diamidino-2phenylindole) that stain cell nuclei.
To test whether rescue of mitochondria dynamic r estor e localization of 5HT 2C R to the plasma membrane and NPY 2 R to the cilia, RPE1 and RPE1-BBS1 −/ − cells were transiently transfected with 5HT 2C R-YFP or NPY 2 R-GFP.After 6 h transfection, the cells were infected with FIV-DRP1S637A virus ( ∼100 particles/cell) for 48 h.It is worth noting that the FIV-DRP1S637A expresses an shRNA that suppresses endogenous DRP1 expression.Cells were then processed for cilia staining with the ARL13B antibody as a bov e.
All images were captured using confocal microscopy (Zeiss LSM880).We counted the number of cilia that are NPY 2 R-positive or -negati v e in each ima ge befor e av era ging the data for each r e plicate.

Brain Slice Immunostaining
Mice dee pl y anesthetized with ketamine and xylazine were perfused with PBS (3 mL/min; 15 mL) followed by 4% PFA/HistoChoice Tissue Fixati v e (Amr esco) in PBS (3 mL/min; 45 mL) using Harvard PHD 22/2000 Syringe Pump.Entire brain was extracted and incubated in the same fixative overnight at 4 • C. F ixed br ains w er e w ashed 3 times with PBS and incubated in 30% sucr ose/PBS ov ernight with 1 change of solution after 4-6 h of initial incubation.Br ains w ere vibr atome-sectioned with 30 μm thickness.Immunostaining was performed on brain sections to detect POMC or NPY 2 R as described pr eviousl y 20 by using a 1:250 dilution of a rabbit polyclonal anti-POMC or -NPY 2 R antibodies.Processed brain sections were mounted using VECTASHIELD mounting medium with DAPI.For each animal, NPY 2 R co-colocalization with cilia was quantified in m ultiple confocal ima ges (Zeiss LSM880) befor e calculating the av era ge for each mouse.

Mitochondrial-targeted GFP
Mitochondrial morphology was analyzed using mitochondrialtargeted GFP (mito-GFP) as described before. 22Briefly, cells grown on coverslips precoated with 0.1% gelatin were transduced with adenovirus expressing mito-GFP (MOI 50) for 48 h, fixed in 4% paraformaldehyde and then mounted in VECTASHIELD mounting medium with 4,6-diamidino-2phenylindole (DAPI).Images were acquired using a Zeiss LAM 880 confocal microscope.National Institutes of Health ImageJ was used for morphometry analysis of length and form factor (length/width) as pr eviousl y r e ported. 22The parameters for form factor are set with minimum value of 1 for perfectly circular mitochondria.

Mitochondrial Morphometry by Transmission Electron Microscopy
Cells were fixed with 2.5% glutaraldehyde (electron microscopy grade) and 2% paraformaldehyde in PBS and then washed with ice-cold 0.1 m sodium cacodylate 3 × 10 min on ice followed by postfixation in 1% osmium tetroxide, 0.8% potassium ferrocyanide in 0.1 m sodium cacodylate for 3 h on ice.After 3 washes in ice-cold ddH 2 O for 10 min each, cells were stained in 2% uranyl acetate for 2 h before dehydration in an ethanol series of ice-cold 20%, 50%, 70%, and 90%, followed by 3 washes in 100% ethanol at room temperature for 10 min each.Samples w ere infiltr ated in 67% ethanol/33% Dur cupan ACM (Fluka; Sigma-Aldrich) for 3 h at room temperature with agitation, then 33% ethanol/67% Durcupan ACM for 3 h at room temperature followed by 3 changes of 100% Durcupan for 8 h each at 22 • C with agitation.The Dur cupan-infiltr ated cells were then flatmounted between 2 mold-release glass slides and polymerized at 60 • C for 2 d.Semithick sections were cut using a Leica EM UC7 ultramicrotome and placed on 50-mesh uncoated copper clamshell grids.
The specimens were irradiated for ∼30 min before initiating a tilt series to limit anisotropic specimen thinning during image collection.During data collection, the illumination was held to near parallel beam conditions.In each sample, several electron micr oscopy ima ges wer e captur ed using Hitac hi HT7800 r andomly, all at the same magnification.The mitochondrial length and form factor wer e measur ed using the ImageJ Fiji area and perimeter tools.To avoid bias, all the mitochondria in an image wer e measur ed.A total of at least 3 experiments were performed eac h time .

Western Blot Assays
Pr oteins wer e extracted by homogenizing the cells or tissue in tissue lysate buffer (50 m m HEPES, pH 7.5, 150 m m NaCl, 1 m m MgCl2, 1 m m CaCl2, 10 m m NaF, 5 m m EDTA, 1% Triton, 2 m m sodium orthovanadate, and Roche cocktail protease inhibitor ta b let).Pr otein samples (20 μg) wer e subjected to SDS PAGE, electr o-transferr ed on a pol yvinylidene fluoride membr ane , then probed with primary antibodies (1:1000) targeting BBS3, DRP1, or phospho-DRP1 followed by a secondary antirabbit antibody (1:10 000).The protein membrane was striped to probe for β-actin, which was used to normalize loading.Protein expr ession w as visualized with ECL detection kit (GE healthcar e) and imaged with Sapphire Biomolecular Imager from Azure Biosystems.

Da ta Anal ysis
The data are expressed as means ± SEM.Data were analyzed using t-test or 2-way analysis of variance (ANOVA) with repeated measures.When ANOVA reached significance, a post-hoc comparison was made using Fisher's test or Tukey's test.GraphPad PRISM 9.1.0was used for statistical analysis.A P < .05v alue w as consider ed statisticall y significant.

Inducible Loss of the Bbs1 Gene in POMC Neurons Causes Obesity
To investigate the metabolic consequences of adult-onset disruption of POMC neuron BBSome , w e gener ated mice that ena b le selecti v e inducib le deletion of the Bbs1 gene in POMC neur ons by br eeding mice expr essing tamoxifen-inducib le Cr e in POMC neurons (POMC CreERT2 ) with Bbs1 fl/fl mice.To validate Cre r ecombinase in POMC neur ons, POMC CreERT2 mice wer e cr ossed with a r e porter mouse model expr essing fluor escent tdTomato (ROSA), in a Cre-dependent manner.First, we confirmed the Cremediated recombination by assessing tdTomato fluor escent pr otein expression 2 wk after POMC CreERT2 /ROSA mice were treated with tamoxifen ( Figure S1A and B ).It should be noted that in the forebrain, tdTomato fluorescent protein was expressed almost exclusi v el y in the arcuate n ucleus of the hypothalamus, which is consistent with the well-known fact that the vast majority of POMC neurons are located in this nucleus. 31ext, we verified that tdTomato fluorescent protein mostly colocalized with POMC ( Figure S1C ) confirming the specificity of POMC CreERT2 mice.We also used single cell RNA PCR to confirm POMC neuron-specific loss of Bbs1 gene expression in POMC CreERT2 /Bbs1 fl/fl /R OSA mice ( Figur e S1D ) v alidating our str ate gy and mouse model.
To determine the consequence of adult-onset Bbs1 gene deletion in POMC neurons, we assessed the effect of tamoxifen treatment on body weight.There was no difference in body w eight betw een POMC CreERT2 /Bbs1 fl/fl mice and littermate contr ols befor e and during administration of tamoxifen, at 6 wk of age ( Figure 1 A and B).However, 4 wk after tamoxifen (10 wk of age), body weight of male mice began to di v erge with POMC CreERT2 /Bbs1 fl/fl mice gaining more weight than the controls, with POMC CreERT2 /Bbs1 fl/fl mice weighing about 4.4 g more than the controls at 20 wk of age ( Figure 1 A and C).Body weight of female POMC CreERT2 /Bbs1 fl/fl mice began to di v erge 6 wk after tamoxifen treatment (12 wk of age), but ther e w as no statistical differ ence compar ed to contr ols ( Figur e 1 B).Nonetheless, female POMC CreERT2 /Bbs1 fl/fl mice gained significantly more weight than the controls at 20 wk of age ( Figure 1 D).
It should be noted that no significant difference in body weight was observed in male POMC CreERT2 /Bbs1 fl/fl mice treated with corn oil r elati v e to contr ol animals tr eated with tamoxifen ( Figure S2A ) indicating that the excess weight gain in tamoxifen treated POMC CreERT2 /Bbs1 fl/fl mice is not due to the presence of the POMC CreERT2 tr ansgene .
The increase in body weight in POMC CreERT2 /Bbs1 fl/fl mice was due to an increase in adiposity as indicated by the higher fat mass ( Figure 1 E and F) and weight of individual fat pads including brown adipose tissue and inguinal, gonadal, peri-renal white adipose tissues ( Figure S2B and C ). Liver mass was also significantl y elev ated ( P < .05) in POMC CreERT2 /Bbs1 fl/fl mice r elati v e to controls (females: 1.16 ± 0.04 versus 1.09 ± 0.03 g, males: 1.98 ± 0.12 versus 1.62 ± 0.07 g).However, there was no difference in lean mass between POMC CreERT2 /Bbs1 fl/fl mice and contr ols ( Figur e 1 G and H), which was further confirmed by the lack of difference in the weight of the kidneys ( Figure S2B and C ).Consistent with the obesity phenotype, both male and female POMC CreERT2 /Bbs1 fl/fl mice exhibited elevated food intake compared to their controls ( Figure 1 I and J).

Inducible Loss of the Bbs3 Gene in POMC Neurons Does Not Affect Body Weight
To determine whether inducib le POMC neur on Bbs3 gene deletion recapitulate the body weight phenotype induced by Bbs1 gene deficiency, w e gener ated a new mouse model that ena b les Cre-mediated specific deletion of the Bbs3 gene (Bbs3 fl/fl ) by inserting 2 lox P sequences to flank exon 5 of the Bbs3 gene ( Figure S1E ).Next, we crossed the Bbs3 fl/fl mice with the POMC CreERT2 mice to generate POMC CreERT2 /Bbs3 fl/fl mice to allow inducible Bbs3 gene specifically in POMC neurons ( Figure S1D ).
We assessed how inducible deletion of the Bbs3 gene in POMC neurons affect body weight and adiposity.However, tamoxifen treatment as above (for 5 d at 6 wk of age) caused no significant change in body weight in male and female POMC CreERT2 /Bbs3 fl/fl mice r elati v e to littermate contr ols ( Figur e 2 A-D).Consistent with the unaltered body weight, fat mass and lean mass were compara b le between POMC CreERT2 /Bbs3 fl/fl mice and controls ( Figure 2 E-H).The weight of various fat pads and kidneys were also not different between the 2 groups of mice ( Figure S3A and B ).

Effects of POMC Neuron Bbs1 Versus Bbs3 Gene Deletion on Glucose Handling and Insulin Sensitivity
Gi v en the importance of POMC neurons for the control of glucose metabolism and peripheral insulin sensitivity, 32 , 33 we asked whether inducible Bbs gene deletion affects glucose handling and insulin action.However, no significant change in fasting blood glucose was noted in POMC CreERT2 /Bbs1 fl/fl mice (male: 89.1 ± 3.8 mg/dL, female: 76.1 ± 2.7 mg/dL) r elati v e to controls (male: 82.9 ± 4.5 mg/dL, female: 66.7 ± 3.1 mg/dL).Next, we performed a glucose tolerance test, which r ev ealed that both male and female POMC CreERT2 /Bbs1 fl/fl mice have glucose intoler ance .In control mice, following acute glucose challenge, blood glucose levels started to decline after peaking at 15-30 min and returned to near normal values after 2 h.However, in POMC CreERT2 /Bbs1 fl/fl mice, after glucose administration, blood glucose levels peaked at 30 min and r emained elev ated at 60 min indicating reduced glucose excursion from blood ( Figure 3 A and  B).Calculating the area under the curve confirmed the glucose intolerance in POMC CreERT2 /Bbs1 fl/fl mice ( Figure 3 C and D).
To assess insulin's ability to stimulate glucose mobilization, we performed an insulin tolerance test.In control mice, insulin treatment caused a robust decrease in blood glucose ( Figure 3 E-H).The blood glucose lowering effect of insulin was significantly attenuated in POMC CreERT2 /Bbs1 fl/fl mice demonstrating the contribution of Bbs1 gene in POMC neurons to insulin sensitivity.
Fasting blood glucose was not different in POMC CreERT2 /Bbs3 fl/fl mice r elati v e to controls (male: P = .3,female: P = .1).Inter estingl y and despite the lack of body w eight phenotype , glucose toler ance test sho ws a trend to war d impairment in both male and female POMC CreERT2 /Bbs3 fl/fl mice when compared to littermate controls ( Figure 4 A and B).The glucose intolerance in POMC CreERT2 /Bbs3 fl/fl mice was more evident when calculating the area under the curve ( Figure 4 C and D).However, insulin tolerance test r ev ealed no differ ence in insulin sensitivity between POMC CreERT2 /Bbs3 fl/fl mice and contr ols ( Figur e 4 E-H).Thus, glucose handling and insulin sensiti vity ar e differ entiall y affected by POMC neur on deletion of the Bbs1 gene versus Bbs3 gene.

Validation of BBS1 −/ − and BBS3 −/ − RPE1 Cells
To understand the molecular mechanisms underlying differential metabolic effects induced by loss of BBS1 gene versus BBS3 gene in POMC neurons, we used CRISPR-Cas9 technology to generate RPE1 cells that lack these genes indi viduall y.We confirmed the absence of BBS3 expression in BBS3 −/ − cells by Western blot ( F igure S4 ).How ever, the lac k of specific BBS1 antibodies precluded us from confirming the loss of this protein in BBS1 −/ − cells.Nonetheless, we found that the endogenous BBS2 protein lost its ciliary localization in both BBS1 −/ − and BBS3 −/ − cells ( Figures S5A and S6A-B ) indicating that the tr affic king of the BBSome to cilia is disrupted.Importantl y, ciliar y localization of the endogenous BBS2 w as r estor ed in BBS1 −/ − cells infected with AA V -BBS1 ( Figure S6C ) confirming that loss of the endogenous BBS2 protein in cilia of BBS1 −/ − cells is due to BBS1 gene deficiency.We further observed that BBS3 protein, which is absent in BBS3 −/ − cells locate throughout cilia in control cells, whereas it appears as punctate in BBS1 −/ − cells, localizing at the base of the cilium ( Figure S5B-C ).Co-staining with γ -tubulin, a basal body marker, r ev ealed that in BBS1 −/ − cells, BBS3 did not colocalize with γ -tubulin, but was adjacent to it, inside the cilium ( Figure S5D ), likely in the transition zone.Thus, BBS1 deficiency disrupts ciliary localization of BBS3.

Loss of BBS1, But Not BBS3, Alters Trafficking of Receptors T ha t Regula te Metabolic Function
Evidence implicating BBS proteins in the tr affic king of receptors inv olv ed in contr ol of energy homeostasis, such as 5-HT 2C R and NPY 2 R, led us to assess how deletion of BBS1 and BBS3 genes affect localization of these r ece ptors.Consistent with our previous report, 16 the 5-HT 2C R locates pr edominantl y in the plasma membrane of control cells with no localization in cilia ( Figure 5 A).Notably, the plasma membrane localization of 5HT 2C R w as alter ed in BBS1 −/ − , but not BBS3 −/ − cells.In BBS1 −/ − cells, the 5HT 2C R appears stuck in the cytoplasm, likely in the late endosome. 16Next, we analyzed the effect of BBS1 versus BBS3 gene deletion on NPY 2 R tr affic king, whic h pr edominantl y localizes to the ciliary membr ane . 16 , 17In control cells, significant amount of NPY 2 R (62% ± 4%) localized in cilia ( Figure 5 B  and C).Nota b l y, this co-localization was substantially reduced in BBS3 −/ − cells and absent in BBS1 −/ − cells.
We also investigated the NPY 2 R localization in POMC neurons of the arcuate nucleus of the hypothalamus in control, POMC CreERT2 /Bbs1 fl/fl mice and POMC CreERT2 /Bbs3 fl/fl mice.In control mice, 33% ± 4% of cilia of POMC neur ons ar e equipped with NPY 2 R ( Figure 6 A and B).Inter estingl y, this NPY 2 R/cilia co-localization was significantly reduced in POMC neurons of POMC CreERT2 /Bbs3 fl/fl mice and absent in POMC neurons of POMC CreERT2 /Bbs1 fl/fl mice.It should be noted that ciliary localization of NPY 2 R in non-POMC cells was not different between control, POMC CreERT2 /Bbs1 fl/fl , and POMC CreERT2 /Bbs3 fl/fl mice.In contrast, ciliary localization of NPY 2 R was disrupted in POMC neurons of POMC CreERT2 /Bbs1 fl/fl mice.Together, these data indicate that BBS1, but not BBS3, is r equir ed for the tr affic king of key meta bolic r ece ptors to plasma membrane and cilia.

Absence of BBS1, But Not BBS3, Affects Mitochondria Dynamics
Next, we assessed whether the control of mitochondria by the BBSome depends on its ciliary function.Mitochondrial-targeted GFP-mediated analysis of mitochondrial morphology showed that BBS1 gene deficiency caused mitochondria hyperfusion as indicated by the increased mitochondrial form factor and length in BBS1 −/ − cells ( Figure 7 A-C), which is consistent with our previous findings. 22Strikingly, these changes in mitochondria were not recapitulated in BBS3 −/ − cells.The role of DRP1 in mediating mitochondrial alterations in BBS led us to measur e DRP1 phosphor ylation at Ser616 [pDRP1(Ser616)].A significant decrease in pDRP1(Ser616) was detected in Bbs1 −/ − , but not Bbs3 −/ − cells ( Figure 7 D-F).Analysis of mitochondrial morphology with transmission electron microscopy (TEM) confirmed that BBS1 −/ − , but not BBS3 −/ − cells display elongated mitoc hondria ( F igur e 7 G).Mor eov er, infecting BBS1 −/ − cells with a DRP1S637A mutant, a mimetic of the dephosphorylated state 34 , 35 that rescues the decrease in pDRP1(Ser616) phosphor ylation 22 r ev ersed the c hanges in mitoc hondrial morphology in BBS1 −/ − cells without affecting mitochondria in BBS3 −/ − cells.These findings demonstrate that BBSome regulation of mitochondria is independent of its ciliary function.
F inally, w e asked whether the mitoc hondria defects may explain the r ece ptor mistr affic king e voked b y BBS1 deficiency.To test this, we examined the effect of restoring mitochondrial dynamic with DRP1S637A on 5-HT 2C R and NPY 2 R localization in BBS1 −/ − cells.Inter estingl y, 5-HT 2C R localization in the plasma membrane was restored in BBS1 −/ − cells infected with DRP1S637A ( Figur e 8 A).Similarl y, ciliar y localization of NPY 2 R w as substantiall y r escued by DRP1S637A in BBS1 −/ − cells ( Figur e 8 B and C).These results point to mitochondrial defects as the underlying cause of receptor mislocalization associated with BBSome deficiency.

Discussion
Pr oopiomelanocortin neur ons ar e key r egulators of energy homeostasis by controlling downstream neurocircuits that promote satiety, increase energy expenditure, and contribute to weight loss. 8Here, we demonstrate contrasting metabolic effects e voked b y POMC neuron deletion of a ke y component of the BBSome, BBS1, versus BBS3, which mediate the BBSome trafficking to cilia.Loss of the Bbs1 gene in POMC neurons triggered a significant increase in body weight and adiposity.In contrast, POMC neuron Bbs3 gene deletion failed to affect body weight and fat mass.Mor eov er, POMC neur on Bbs1 gene deletion caused glucose intolerance and insulin resistance, whereas loss of the Bbs3 gene in these same neurons is associated with normal insulin sensitivity and mild glucose intoler ance .Interestingly, disruption of the BBSome hindered the tr affic king and localization of 5-HT 2C R and NPY 2 R, 2 important r ece ptors for energy balance.On the other hand, BBS3 deficiency did not interfere with plasma membrane localization of 5-HT 2C R, but reduced the tr affic king of NPY 2 R to cilia.Mor eov er, deficiency in the BBSome, but not BBS3, impaired mitochondria dynamics through reduction in DRP1 activity.Rescuing the diminished DRP1 activity restored the proper localization of 5-HT 2C R and NPY 2 R in plasma membrane and cilia, r especti v el y, in BBSome-deficient cells.These data highlight the importance of the BBSome in POMC neurons for control of energy and glucose balance while BBS3 appears minimall y inv olv ed in glucose handling.Together, our findings indicate that POMC neuron BBSome regulation of energy homeostasis is not linked to its ciliary function.
The increase in body weight and adiposity by inducible Bbs1 gene a b lation in POMC neur ons is in line with the effects of constituti v e deletion of this gene in these neurons. 16This argues against the idea that BBS-associated obesity is caused by neur odev elopmental defects that arises fr om pr enatal loss of BBS genes. 36It should be noted, however, that the weight gain and adiposity evoked by inducible Bbs1 gene ablation in POMC neur ons w as less pr onounced r elati v e to those ev oked by its constituti v e deletion.This difference in weight gain and adiposity may reflect the more severe impact of BBSome disruption during embryonic stage versus adult-onset phase.Alternatively, this may be due to a broader loss of the Bbs1 gene in mice bearing the constituti v e POMC Cre .Indeed, constituti v e POMC Cre mice displayed Cr e expr ession not onl y in POMC neur ons, but also in other neuronal populations, 37 which may have contributed to the pronounced obesity phenotype in POMC Cre mice r elati v e to POMC CreET2 mice.Irr especti v e of this issue, our data demonstrate that the BBSome in POMC neurons is r equir ed for energy balance.The pr ev alence and sev erity of obesity in BBS3 patients is v aria b le within and among families.For instance, Young et al. r e ported that most BBS3 patients of northern European descent displayed no or mild obesity phenotype, 38 whereas most BBS3 patients of Arab-Bedouin and Iranian families were found to exhibit marked obesity. 39 , 40We pr eviousl y r e ported that the phenotype de veloped b y global Bbs3 null mice do not entir el y mimic those observed in other BBS knockout mice. 26In particular, Bbs3 null mice displayed minimal increase in body weight, adiposity, and plasma leptin.Here , w e extend these findings by demonstrating that in contrast to Bbs1 gene a b lation, Bbs3 gene deletion in POMC neurons failed to affect body weight and adiposity.These results indicate that increased body weight in Bbs3 null mice is not due to loss of BBS3 function in POMC neurons.These findings also show that the contribution of the POMC neuron BBSome to energy homeostasis is independent of its interaction with BBS3 and by extension from its ciliary function.The notion that BBSome control of energy balance is not linked to cilia is further supported by the minimal or no effect on body weight and adiposity of loss of cilia ( Ift88 gene) in the leptin sensiti v e neur ons, wher eas Bbs1 gene deletion in these same neurons causes frank obesity. 20 , 41Rather, the BBSome appears to influence body weight through the regulation of the plasma localization of key r ece ptors inv olv ed in energy homeostasis such as the 5-HT 2C R.
Type 2 diabetes is a common observation in BBS patients. 42 , 43sing BBS mice , w e demonstr ated that this phenotype is due to mis-tr affic king of the insulin r ece ptor. 28Our curr ent inv estigation r ev ealed the importance of Bbs1 gene in POMC neur ons for the regulation of glucose metabolism and insulin sensitivity as indicated by the glucose intolerance and insulin resistance observed in POMC CreERT2 /Bbs1 fl/fl mice.Remarkably, inducible deletion of the Bbs3 gene in POMC neurons caused slight impairment in glucose tolerance despite normal body weight and adiposity.This indicates that the glucose phenotype in BBS mice is independent of obesity.This is in line with our previous findings of insulin resistant in global BBS mice even when kept lean using caloric restriction. 28Our results add to the growing body of evidence pointing to the importance of POMC neurons in the regulation of glucose metabolism and insulin sensitivity.Previous studies have implicated various proteins and signaling pathw ays in POMC neur ons in the control of peripheral glucose metabolism and insulin action. 32 , 33We extend these findings by implicating BBS proteins in POMC neurons in glucose tolerance and insulin sensitivity.However, additional work is needed to understand how loss of BBS proteins in POMC neurons alters these parameters.
The differential body weight and metabolic phenotypes r esulting fr om loss of BBS1 v ersus BBS3 seem to hav e its explanation in the distinct inv olv ement of these 2 proteins in the handling of metabolic receptors.Consistent with previous findings, 16 , 17 we showed that BBS1 is r equir ed for pr oper localization of 5-HT 2C R and NPY 2 R to plasma membrane and cilia, respecti v el y, in cultured cells and POMC neurons.Importantly, we show that BBSome regulation of the trafficking of these r ece ptors relate to mitochondria function as indicated by the restored 5-HT 2C R and NPY 2 R localization following rescue of mitochondrial defects in BBS1 −/ − cells.On the other hand, loss of BBS3 did not affect the tr affic king of 5-HT 2C R nor the mitochondria dynamic, but reduced the ciliary localization of NPY 2 R. The dispensability of BBS3 for 5-HT 2C R tr affic king is consistent with the notion that BBSome transport of r ece ptors to plasma membrane is independent of its ciliary function. 20However, the fact that ∼34% of NPY 2 R still traffic to cilia in absence of BBS3 is puzzling because BBS3 is inv olv ed in the recruitment and tr affic king of the BBSome to cilia, 26 which we confirmed by showing that loss of BBS3 alter the ciliary localization of the BBSome component BBS2.Of note, cr ystal structur e anal ysis has demonstrated that BBS3 promote the BBSome entry into cilia through its binding to BBS1. 24This interaction is abolished by single point mutations in the BBS3-BBS1 interface pr ev enting the import of BBSome into cilia.Mutations in BBS1 (e.g.M390R) also disrupted the interaction of the BBSome with BBS3. 24Thus, it is not clear how the BBSome mediates tr affic king of NPY 2 R to cilia in absence of BBS3.One potential explanation is that in absence of BBS3, other proteins may mediate tr affic king of the BBSome cargos such as the NPY 2 R to the ciliary membr ane .For instance , tubby family protein TULP3 has been implicated in the ciliary localization of v arious G pr otein-coupled r ece ptors. 44 , 45Howev er, futur e studies are warranted to test whether compensatory mechanisms overcome the loss of BBS3 and mediate the tr affic king of NPY 2 R to cilia.Additional studies are also needed to determine whether loss of ciliary localization of NPY 2 R contribute to the metabolic defects evoked by BBSome deficiency in POMC neurons.
Our findings reinforce the notion that the BBSome is a key modulator of mitochondria dynamic and function by regulating the activity of DRP1, a core component of the canonical mitochondrial fission machinery in mammals.However, the exact mechanisms underlying BBSome regulation of DRP1 lev els and acti vity ar e not full y understood.We pr eviousl y demonstrated that DRP1 pr otein w as significantl y decr eased in mitochondria isolated from BBS1-deficient cells, 22 indicating that absence of the BBSome disrupts mitochondrial localiza-tion of DRP1, which may lead to its de gr adation decreasing its total and phosphorylated levels.Alternatively, dysregulation in DRP1 phosphorylation when the BBSome is absent 22 may promote its de gr adation.Mor e studies ar e necessar y to test these possibilities.
Our findings provide additional evidence that point to a neuronal origin of the obesity associated with BBS.Specifically, our results show that defects in the first-order neurons (e.g.POMC neurons) as the underlying cause of obesity in BBS.Importantly, such defects can be b ypassed b y tar geting do wnstream neurocircuits.Indeed, we showed that stimulation of MC4R cause a significant decrease in food intake and body weight in BBS mice. 21This observation has led to several clinical trials that demonstrated the efficacy of an MC4R agonist, Setmelanotide, as a treatment option for obesity in individuals with BBS. 46 , 47etmelanotide is now appr ov ed for the management of obesity in BBS patients, improving their health-related quality of life. 48][51]

Figure 1 .
Figure 1.Pr oopiomelanocortin neur on-specific inducib le Bbs1 gene deletion increases body weight.( A-B ) Effect of tamoxifen treatment (at 6 wk of age) on body weight of male ( A ) and female ( B ) POMC CreERT2 /Bbs1 fl/fl mice and littermate controls (CTL).( C-D ) Change in body weight (at 20 wk r elati v e to 6 wk of age) in males ( C ) and females ( D ). ( E-F ) Fat mass in males ( E ) and females ( F ) at 20 wk of age.( G -H ) Lean mass in males ( G ) and females ( H ) at 20 wk of age.( I-J ) Cumulative 5 d food intake of males ( I , n = 10-12 animals/group) and females ( J , n = 12-15 animals/group) at 21 wk of age.* P < .05 and * * P < .01versus CTL.

Figure 2 .
Figure 2. Proopiomelanocortin (POMC) neuron-specific inducible Bbs3 gene deletion does not alter body weight.( A-B ) Effect of tamoxifen treatment (at 6 wk of age) on body weight of male ( A ) and female ( B ) POMC CreERT2 /Bbs3 fl/fl mice and littermate controls (CTL).( C-D ) Change in body weight (at 20 wk r elati v e to 6 wk of age) in males ( C ) and females ( D ). ( E-F ) Fat mass in males ( E ) and females ( F ) at 20 wk of age.( G -H ) Lean mass in males ( G ) and females ( H ) at 20 wk of age.
NaH 2 PO 4 , 24 NaHCO 3 , 5 HEPES, 12.5 glucose, 2 MgSO 4 , and 2 CaCl 2 .G seals on arcuate tdT omato + or tdT omato − cells were made with an electrophysiological recording pipette and r emov ed fr om the brain slice.Pipette tips containing a single neur on wer e br oken off into RN A extr action buffer and further processed for single cell PCR for Bbs1 , Bbs3 , Pomc , and S18 genes according to the man ufactur e's pr otocol (PicoPur e RNA Isolation Kit, Thermo Fisher).