86552513

Nitrosation of CD36 regulates endothelial function and serum lipids

Authors

Melissa A. Luse

1,2

, Wyatt J. Schug

1,2

, Luke S. Dunaway

, Shruthi Nyshadham

, Skylar A. Loeb

1,2

Alicia Carvalho

Rachel Tessema

Caitlin Pavelic

1,3

T.C. Stevenson Keller IV

1,2

Xiaohong Shu

Claire A. Ruddiman

1,3

Anna Kosmach

Timothy M. Sveeggen

Ray Mitchell

Pooneh Bagher

Richard D. Minshall

Norbert Leitnger

1,3

Linda Columbus

Kandice R. Levental

Ilya Levental

Miriam Cortese-Krott

Brant E. Isakson

1,2

* To whom correspondence should be addressed

Affiliations

Robert M. Berne Cardiovascular Research Center, University of Virginia School of Medicine

Department of Molecular Physiology and Biophysics, University of Virginia School of Medicine

Department of Pharmacology, University of Virginia School of Medicine, Charlottesville, VA,

USA

College of Pharmacy, Dalian Medical University, Dalian China

Department of Cellular and Integrative Physiology, University of Nebraska Medical Center,

Omaha, Nebraska, USA

Department of Pharmacology, University of Illinois-Chicago

Department of Chemistry, University of Virginia

Heinrich Heine University, Dusseldorf, Germany

E: brant@virginia.edu

P: +1 434.466.3394

• Running title: Nitrosation of CD36

Abstract

During obesity, endothelial cells (ECs) become lipid laden leading to endothelial dysfunction. We

demonstrate endothelium downregulates caveolin-1 (Cav1) in mouse and human in response to

obesity. Using an EC-specific Cav1 knockout mouse, we find mice are hyperlipidemic regardless

of diet, but retain endothelial cell function. Whereas initially this was thought to be due to Cav1

mediate endocytosis, we find instead the mice have significantly increased nitric oxide (NO) in

response to the lack of Cav1. The presence or absence of NO toggled inversely EC lipid content

and plasma lipid in mice. We found the fatty acid translocase CD36 was directly nitrosated by

endogenous NO at the same cysteines that are palmitoylated on CD36. The nitrosation of CD36

prevented it’s trafficking to the plasma membrane and decreased lipid uptake. The physiological

effect of this mechanism was a reliance on NO for endothelial function. This work suggests that

CD36 nitrosation occurs as a protective mechanism to prevent EC lipotoxicity and preserve

function.

Teaser

Nitric oxide regulates serum lipids and endothelial cell lipid content through nitrosation of CD36.

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

Introduction

Endothelial cells (ECs), strategically positioned as gatekeepers of peripheral tissues, play a crucial

role in the uptake and transportation of materials from the bloodstream to tissues

1-3

. In the context

of metabolic syndrome, characterized by a state of nutritional excess, the finely tuned transport

system of ECs faces a challenge. Excessive fatty acid cargo transported into the endothelium can

disrupt the equilibrium of fatty acid storage and handling, leading to endothelial dysfunction

4, 5

. As

a consequence, lipids begin to accumulate within the endothelium, instigating a state of lipotoxicity

that further contributes to and exacerbates endothelial dysfunction

6-8

. In metabolic syndrome, the

intricate process of how ECs normally take in lipids becomes dysregulated, emphasizing the

importance of understanding these mechanisms to inhibit the onset and progression of endothelial

dysfunction

. ECs possess inherent protective mechanisms that aim to combat lipotoxicity and

dysfunction; however, these compensatory mechanisms fail as the lipid burden increases. This

lipid-dependent dysregulation not only compromises the physiological functions of ECs but also

sets the stage for the development of various cardiovascular diseases associated with obesity

10, 11

Caveolae, specialized membrane invaginations have long been implicated in cardiovascular and

metabolic syndrome, with historical relevance attributed to their role in low-density lipoprotein

(LDL) transcytosis, promoting atherogenesis

12-16

. Caveolin-1 (Cav1) emerges as a central figure in

the intricate landscape of EC function, particularly through its association with caveolae. Recent

investigations have unveiled additional layers of complexity in Cav1's regulatory functions,

demonstrating its role in both lipid uptake and lipolysis

4, 17-19

. Notably, Cav1 extends its influence

beyond lipid metabolism and transport within ECs to include the deliberate regulation of endothelial

nitric oxide synthase (eNOS)

20, 21

. Cav1 has been recognized as an integral modulator of eNOS

activity and the subsequent production of nitric oxide (NO), a critical mediator of vascular

homeostasis

22, 23

. The absence of caveolae disrupts the normal inhibitory relationship between Cav1

and eNOS, leading to an over production of NO. This negative regulation can be modified to aptly

tune NO concentrations in ECs. Here, we draw a link between the lipid handling and eNOS

regulatory roles of Cav1 establishing, for the first time, a role for NO in lipid handling. This new

role for NO emphasizes the coordinated interplay between Cav1, NO, lipid uptake, and EC

dysfunction.

Lipid uptake into non-fenestrated endothelium requires the transfer of fatty acids from the

circulation into ECs

and regulation by key proteins, including lipoprotein lipase, scavenger

receptor-B1, and notably, CD36

. CD36 is abundantly expressed in the endothelium and serves as

a crucial fatty acid transporter, allowing for the uptake of lipids into ECs

24, 26, 27

. In humans,

polymorphisms in the CD36 gene are associated with hyperlipidemia, metabolic syndrome, and

type 2 diabetes, highlighting the clinical relevance of CD36 in metabolic homeostasis

. However,

the exact regulatory mechanisms behind its function within the endothelium remain not fully

understood. One proposed mechanism for CD36 regulation is through the posttranslational

modification palmitoylation

29-31

. Palmitoylation is a reversible modification which plays an

important role in shuttling modified proteins to and from the plasma membrane. The reversible

cycles of palmitoylation and depalmitoylation allow for dynamic re-localization of proteins within

the cell

32-35

. Specifically, work by Hao et al. show APT1, acyl-protein thioesterase, to cleave

palmitoyl groups from CD36 driving the internalization of CD36 from the plasma membrane

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

Proposed here, is novel post translational modification of CD36, nitrosation. Nitrosation is an

important post translational modification which can alter protein localization and function

36-38

. Like

palmitoylation, nitrosation is readily reversible

allowing for rapid protein modifications to occur

altering cellular function. Here, we present nitrosation of CD36 via endogenous NO by loss of Cav1

inhibition in obesity. We find this post translational modification works in tandem with

palmitoylation as a competitive regulatory mechanism for controlling lipid transport in the adipose

endothelium. Our evidence suggests the cellular changes in CD36 post translational modifications

between palmitoylation and nitrosation in endothelium regulate blood lipids.

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

Results

100

Caveolin 1 expression decreases in endothelial cells during metabolic syndrome in mouse and

101

human. We initially sought to identify EC transcripts that may be involved with lipid transport into

102

endothelium during obese conditions. Using our scRNAseq of normal chow (NC) and high fat diet

103

(HFD) adipose

, we subsetted the isolated ECs based on expression of specific marker genes

104

known to be present in endothelium (Supplementary figure 1A-F). ECs were then further

105

classified by their vascular origin (i.e. arterial, venous, capillary, lymphatic; Supplementary figure

106

1B). Differential gene expression analysis was performed and a reduction in Cav1 expression in

107

HFD mice (Figure 1A) was observed in ECs from all vascular clusters (Figures 1B-C). We next

108

used a published scRNAseq data set of human adipose tissue from both healthy individuals and

109

patients with metabolic syndrome

and also observed CAV1 expression was decreased in adipose

110

ECs from obese patients (Figure 1D-F). The diet dependent manner of this gene expression data

111

suggested Cav1 may be an important component to lipid handling during obesity.

112

113

Endothelial cell Caveolin 1 can regulates serum and cellular lipids. To investigate how the

114

down regulation of Cav1 contributes to metabolic syndrome progression, we selectively deleted

115

Cav1 from endothelium (Figure 2A). Loss of Cav1 in ECs was extensively validated (Figure 2B-

116

D; Supplementary figures 2A-B). Importantly, Cav1 protein levels in adipocytes were unaltered

117

(Supplementary figure 2C). Deletion of Cav1 from endothelium did not change mouse weight,

118

epididymal fat pad mass or food consumption (Supplementary figures 2D-G). Despite the lack of

119

weight gain phenotype, EC Cre

Cav1

fl/fl

mice when compared to EC Cre

Cav1

fl/fl

mice, had

120

significantly increased serum triglycerides, cholesterol, and LDL regardless of diet (Figures 2E-

121

G). Importantly, EC Cre

Cav1

fl/fl

mice fed a NC diet became hyperlipidemic, even surpassing

122

triglyceride levels of HFD controls. HDL levels were unchanged and non-esterified fatty acids

123

(NEFA) levels were unaffected by loss of Cav1 from ECs (Figures 2H-I). Next, adipose arteries

124

were stained with Nile red to visualize en face intracellular lipid accumulation in endothelium.

125

Arteries from EC Cre

Cav1

fl/fl

mice had significantly decreased levels of lipid droplets relative to

126

EC Cre

Cav1

fl/fl

controls (Figures 2J). Thus, it appears the lack of Cav1 in endothelium prevents

127

lipid accumulation in endothelial cells and promotes the accumulation of lipids in serum (Figure

128

2K). Despite the hyperlipidemic phenotype of the mice, their ability to process glucose was

129

improved regardless of diet (Figure 2L, Supplementary figure 2H). The enhanced glucose

130

sensitivity provided initial evidence the endothelium remained more metabolically healthy

131

compared to their EC Cre

Cav1

fl/fl

littermates, possibly due to decreased intracellular lipid

132

accumulation from lack of Cav1.

133

134

We next examined the mechanism underlying the observed reduction in endothelial lipid uptake

135

with human adipose microvascular endothelial cells (HAMECs). We recapitulated decreased lipid

136

uptake in HAMECs with Cav1 deficiency (Figure 3A-C). Because caveolin organizes cholesterol

137

in the plasma membrane into distinct lipid domains

, we hypothesized that the disruption of the

138

cholesterol would also alter lipid uptake. To test this, we treated HAMECs with dimethyl- β-

139

cyclodextrin (Cyclo) and observed similar results to the depletion of Cav1 (Figure 3D-F). Because

140

lack of Cav1 and disruption of lipid domains inhibited lipid uptake, we predicted that endocytosis

141

inhibitors would also inhibit lipid uptake. However, using two different endocytosis inhibitors

142

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

(dynosore and genistein), we observed no change in lipid uptake after treatment with lipids (Figure

143

3G-H).

144

145

A potential role for endothelial nitric oxide in regulating lipids. It is well described there is an

146

increase in nitric oxide (NO) in the absence of Cav1 due to its role in eNOS inhibition

20, 21, 43-45

. An

147

increase in NO is also observed with the depletion of cholesterol in the plasma membrane which

148

disrupts Cav1-eNOS interactions.

46-48

At 20-weeks of age (12 weeks on NC diet) we measured

149

nitrite and nitrate levels (nitrite and nitrate are stable oxidation products of NO) from Cav1

fl/fl

mice

150

and found EC Cre

Cav1

fl/fl

mice had significantly increased serum nitrite and nitrate (Figure 3I).

151

Furthermore, mice on a HFD that we demonstrated in Figure 1 had a loss of Cav1 in adipose EC,

152

also had significantly increased nitrite and nitrate (Figure 3J). These results suggested NO derived

153

from loss of Cav1 could regulate endothelial lipid uptake.

154

155

To test the hypothesis the Cav1 loss is causing a compensatory increase in NO and disrupting lipid

156

uptake, we used HAMECs genetically devoid of Cav1, a condition that previously inhibited lipid

157

uptake (Figure 3). We treated these cells with L-NAME, a well-known inhibitor of nitric oxide

158

synthase

49, 50

, to examine the role of NO in endothelial lipid uptake. After lipid treatment, cells

159

devoid of Cav1 were indistinguishable from control treated cells (Figure 4A). We repeated this

160

experiment in vivo by adding L-NAME in the drinking water of EC Cav1

fl/fl

mice for 2-weeks.

161

Adipose arteries from EC Cre

Cav1

fl/fl

mice had lipid droplets present in the endothelium (Figure

162

4B), a result strikingly different than the EC Cre

Cav1

fl/fl

that did not receive L-NAME water

163

(Figure 2). These data indicate decreased lipid uptake from the loss of EC Cav1 may be rescued

164

with the inhibition of nitric oxide production.

165

166

Since the EC Cre

Cav1

fl/fl

mice given L-NAME were able to accumulate lipids within the

167

endothelium, we hypothesized that their blood lipids would be decreased after drinking L-NAME

168

water relative to before the onset of treatment. We took blood before giving the mice L-NAME

169

water and then took blood after their 2-week treatment and saw a significant decrease in both serum

170

triglycerides and cholesterol after L-NAME consumption (Figure 4C). Similarly, HFD mice were

171

also given L-NAME drinking water for 2-weeks and compared to HFD mice given regular water

172

(Figure 4D). The L-NAME treated group had significantly lower serum triglycerides and

173

cholesterol compared to the control water group (Figure 4D). To determine if the source of NO,

174

we used eNOS

fl/fl

mice either on a red blood cell (RBC) or EC Cre

. Loss of eNOS from RBCs had

175

no effect on serum lipids (Figure 4E). However, deletion of eNOS from endothelium significantly

176

decreased serum triglycerides and lowered cholesterol (Figure 4F). Thus, endogenous NO from

177

endothelium is capable of regulating serum and intracellular lipids.

178

179

NO modification of CD36 regulates lipid uptake. Next, we sought to identify the target of NO

180

within the endothelium that may be responsible for regulating lipid uptake. Because there was an

181

association between changes in serum and intracellular lipid levels and changes with alterations in

182

Cav1, we performed lipid domain fractionation and blotted for CD36, a fatty acid translocase

183

present on endothelium

24, 27, 51

, which has been shown to colocalize with Cav1

52-54

. We also

184

observed CD36 in caveolar enriched lipid fractions (Figure 5A). However, when visualized within

185

adipose endothelial arteries en face, we find Cav1 and CD36 signals did not overlap (Figure 5B).

186

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

This data shows they are in lipid domains which are in close proximity to one another, but not likely

187

within caveola (e.g.,

55, 56

). To determine if the CD36 is necessary for lipid uptake in adipose

188

endothelium, we used CD36 siRNA mediated knockdown of CD36 (siCD36) in HAMECs. Cells

189

treated with siCD36 had significantly decreased lipid uptake when comparted to controls (Figure

190

5C). Specificity of the CD36 antibody used are shown in Supplementary figure 3A-B. Cells

191

treated with siCD36 did not have altered expression of Cav1 (Supplementary figure 3C). To

192

assess whether CD36 is sufficient for lipid uptake, we used HEK293T (HEK) cells which do not

193

endogenously express CD36 or eNOS (Supplementary figures 3D-F). Lipid uptake in HEK cells

194

was only observed with CD36 transfection (Figure 5D). Because CD36 was sufficient for lipid

195

uptake, and NO regulated serum and intracellular lipid accumulation independent of Cav1, we

196

tested whether NO targeted CD36. Using an NO donor and an S-nitrosothiol (DETA/NO and

197

SNOG), we were able to block lipid uptake in HEK cells transfected with CD36 (Figure 5D). A

198

common method by which NO regulates protein function is by nitrosation (first described by

199

Stamler et al

37, 57

). Using a biotin-switch assay

, we found CD36 is nitrosated in the presence of

200

NO (Figure 5E). To further solidify the association between NO and CD36, we compared lipid

201

uptake in mouse adipose arteries expressing or genetically lacking eNOS. The lack of NO from

202

eNOS-deficient mice caused an accumulation of lipids in the endothelium (Figure 5F), that was

203

blocked by the CD36 inhibitor SSO (Figure 5F). The sum of the data indicates increased NO,

204

generated from lack of Cav1, may regulate CD36 lipid transport by nitrosation.

205

206

To identify which cysteines may be nitrosated on CD36, we turned to GPS-SNO

, a software

207

designed to predict sites of nitrosation. We found three specific cysteines (3, 313, 466) on CD36

208

likely to undergo nitrosation (Figure 6A). These three cysteines lie in critical regions for the

209

function and localization of the CD36 protein (Figure 6B). Cysteines 3 and 466 are palmitoylated

29-

210

and cysteine 313 is involved in the formation of disulfide bridges that are present in the fatty acid

211

binding pocket of the protein

60, 61

. To determine which cysteines on CD36 may be nitrosated, we

212

used mutated versions (C→A) of CD36 at C3, C313, and C466. First, we used a version of the

213

plasmid where all three cysteines of interest were mutated (C466A/313A/3A; Figure 6C). Minimal

214

lipid uptake was seen in cells containing the C466A/313A/3A mutant CD36. To dissect which

215

individual or combination of cysteines was responsible for the loss of lipid uptake, individual and

216

double mutants were made. DETA/NO treatment inhibited lipid uptake in each single cysteine

217

mutation. We next tried a double mutation of C313A/3A and saw a decrease, however not

218

statistically significant, in lipid uptake with DETA/NO administration. A biotin switch assay was

219

performed, confirming that CD36 was still nitrosated in the C313A/3A CD36 mutants

220

(Supplementary figure 3G). Due to the persistent nitrosation of CD36 in the C313A/3A mutants,

221

an additional double mutation was created to probe the combinatorial role of cysteine 466 and 3

222

nitrosation. Mutation of C466A/3A inhibited lipid uptake mimicking the phenotype seen in the

223

C466A/313A/3A CD36 mutant (Figure 6C). With the mutation of C466A/3A, nitrosation of CD36

224

was no longer detected via biotin switch (Figure 6D). The cumulative lipid uptake data

225

(Supplementary figure 3H) and detection of nitrosation indicate cysteines 3 and 466 to be crucial

226

for CD36 mediated lipid uptake.

227

228

Next we hypothesize that nitrosation of cysteines 3 and 466 on CD36 inhibits palmitoylation,

229

inhibiting proper localization of CD36 to the plasma membrane and subsequently decreasing lipid

230

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

uptake (Figure 7A). To assess CD36 localization with loss of Cav1, an environment with increased

231

NO, we used caveolin lipid domain fractionation and siRNA for Cav1. With siCav1, CD36 is no

232

longer enriched in the caveolin/plasma membrane fraction (Figure 5A: fraction 3) and is was

233

shifted towards fraction 4 overlapping with the ER marker calnexin (Figure 7B). Total CD36

234

expression at the RNA or protein level does not change with loss of Cav1 (Supplementary Figure

235

4). If loss of Cav1 and therefore an increase in NO is disrupting the localization of CD36 this effect

236

should be reversed with L-NAME. To test this, HAMECs were treated with either control (Cntrl)

237

or Cav1 siRNA (siCav1) then treated with L-NAME or a vehicle control. The surface of the cells

238

was biotinylated to capture only plasma membrane proteins. Cells lacking Cav1 had decreased

239

plasma membrane CD36; however, cells treated with L-NAME, restored CD36 at the plasma

240

membrane (Figure 7C). Because palmitoylation of CD36 is necessary for its plasma membrane

241

localization

, we tested the effect of NO on palmitoylation levels with the addition of DETA/NO

242

to HEK293T cells. Using an IP-ABE assay, we were able to detect decreased palmitoylation of

243

CD36 in the presence of NO (Figure 7D). Since cysteines 3 and 466 on CD36 are important for

244

protein localization

, we transfected HAMECs with either a wild type or C466A/3A mutant version

245

of the CD36 plasmid to assess localization of the protein. HAMECs receiving the C466A/3A

246

mutant CD36 had protein sequestration inside the cells as opposed to the plasma membrane staining

247

seen in the wild type protein (Figure 7E). Additionally, cells transfected with wild type CD36 and

248

treated with DETA/NO for 30 minutes had decreased CD36 present at the plasma membrane

249

(Figure 7E). Last, we wanted to modulate CD36 localization in vivo using both EC Cre

and EC

250

Cre

Cav1

fl/fl

mice and L-NAME drinking water. EC Cre

mice demonstrate a diffuse and punctate

251

staining pattern for CD36 (Figure 7F). Secondary only staining controls for CD36 are show in

252

Supplemental figure 4B. EC Cre

mice lose the punctate staining pattern and CD36 becomes

253

compartmentally sequestered in ECs similar to a pattern of endoplasmic reticulum (ER). CD36

254

staining in EC Cre

mice administered L-NAME have a more similar staining pattern to their EC

255

Cre

controls (Figure 7F) and is quantified using Pearson’s correlation coefficient. The data

256

therefor suggests nitrosation may prevent CD36 transport to the plasma membrane and in so doing,

257

decreases intracellular endothelial lipids, and increases plasma lipids.

258

259

NOS inhibition worsens endothelial dysfunction. With an increase in EC lipid accumulation, we

260

would expect to see endothelial dysfunction. To assess the impact of lipids on the EC transcriptome

261

we performed bulk RNA-sequencing on HAMECs treated with lipids, L-NAME, and DETA/NO.

262

HAMECs receiving both lipid and L-NAME treatment had a marked decrease in mitochondrial

263

oxidative phosphorylation genes and a significant increase in lipid storage genes (Figure 8A,

264

Supplementary figures 5A-D). To investigate the functional effects of decreased mitochondrial

265

gene expression we used a Seahorse bioenergetics assay to perform a mitochondrial stress test and

266

found HAMECs treated with lipids had a decrease in maximal oxygen consumption (Figure 8B),

267

but the addition of both lipids and L-NAME decreased maximal oxygen consumption even further.

268

Supplementing with DETA/NO, we reversed the decreased oxygen consumption and restored

269

mitochondrial function to control levels even in the presence of lipids. This functional data

270

supported the bulk RNA-sequencing analysis. Both loss of Cav1 and L-NAME alone did not affect

271

mitochondrial bioenergetics (Supplementary figures 5E-F). To assess how this dysfunction

272

translates to arterial function we used pressure myography to measure the vasodilatory capacity of

273

third order mesenteric arteries. Using C57Bl/6 mice, the presence of L-NAME does not alter

274

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

acetylcholine (ACh) induced vasodilation (Figure 8C), an effect commonly observed for small

275

arteries

62-66

. However, if the same arteries were exposed to lipids with L-NAME, preventing

276

nitrosation of CD36 and allowing an influx of lipids into endothelium, only then was the ACh

277

dilation significantly decreased (Figure 8C). We interpret this data as evidence that NO’s effect on

278

vascular function of small arteries may be preferentially to regulate lipid uptake via CD36.

279

280

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

Discussion

281

ECs are woven throughout every major organ system and facilitate the transport of critical cellular

282

nutrients and supplies. Cellular energy needs can fluctuate, requiring ECs to constantly adapt their

283

transport function. However, during times of excess lipid burden, such as in metabolic syndrome

284

and obesity, this crucial energy mediator can become dysfunctional, causing other organs systems

285

to fail. ECs must have a way to control lipid build-up or else they risk organs becoming insulin-

286

resistant. Here, we demonstrate one mechanism in which the endothelium protects itself from

287

excess lipid accumulation and the subsequent lipotoxicity that follows. Our data suggests a novel

288

mechanism whereby NO modifies the fatty acid transporter CD36, preventing its localization to the

289

plasma membrane and decreasing EC lipid accumulation (Supplementary figure 6). This work is

290

a culmination of three exciting findings. First, at the cellular level, altering either palmitoylation or

291

nitrosation of CD36 at the same cysteines regulates endothelial lipid uptake and cellular function.

292

Second, at the tissue level, we provide evidence for the physiological importance of nitric oxide as

293

a mediator of lipid accumulation.

294

295

Our initial data demonstrated in mouse and humans a loss of Cav1 in adipose endothelium in

296

response to obesity. To better understand how the loss of Cav1 impacts endothelial function in

297

obesity, we generated EC Cav1 deficient mice. Originally, we expected the phenotypes of an EC

298

specific Cav1 deletion to be similar to the global Cav1 knockout mouse

67, 68

. Cav1 null mice are

299

resistant to HFD-induced weight gain, insulin resistant

, glucose insensitive

, and

300

hyperlipidemic

. Our EC specific Cav1 deletion had no weight gain differences compared to

301

controls and surprisingly enhanced glucose sensitivity. Deletion of EC Cav1 and Cav1 null mice

302

had a matching phenotype with elevated serum lipids. Many of the phenotypes associated with the

303

Cav1 null mice have been attributed to adipocyte loss of Cav1. The data we generated helps to tease

304

out the contributions of EC Cav1 compare to adipocyte Cav1 to whole animal metabolism.

305

306

Interestingly, the phenotypes seen with loss of EC Cav1 mimic that off an EC-specific loss of CD36.

307

EC CD36 knockout mice have increased postprandial serum triglyceride levels, decreased

308

intracellular lipid droplet formation, and enhanced glucose sensitivity

. Conversely,

309

overexpression of CD36 in muscle reduces serum triglycerides and increases glucose

310

insensitivity

. Together, this drives the idea that these two proteins exist in a regulatory pathway

311

related to lipid entry into endothelium.

312

313

We considered two hypotheses for how Cav1 regulates lipid uptake. The first, is that Cav1 does so

314

through endocytosis. However, we found that inhibition of endocytosis with dynosore and genistein

315

showed no effect on lipid uptake. This finding was surprising as the literature surrounding Cav1

316

and lipid uptake suggests its mechanism for doing so is through endocytosis

. The second

317

possibility we considered is that increased NO associated with loss of Cav1 inhibits lipid uptake.

318

Nitrosation from NO has been shown to regulate protein function and localization

38, 71

. Therefore,

319

we proposed that Cav1 regulates lipid uptake by modulating endothelial NO levels.

320

321

According to our results, the addition of NO to cysteines C3 and C466 on CD36 inhibits CD36

322

from being trafficked to the plasma membrane and instead retains the translocase within the ER.

323

Therefore, less lipids can enter endothelium with nitrosation of CD36. Converse to nitrosation,

324

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

previous work has identified palmitoyl groups as key regulators of CD36 localization

29-31, 72

325

Specifically, the N and C-terminal cysteines (C3, C7, C464, C466) have been demonstrated as sites

326

of palmitoylation for proper CD36 localization to the plasma membrane. The palmitoylated CD36

327

at the plasma membrane functions to take in lipids

29-31, 72

. This mechanism has been identified in

328

HEK293T, 3T3L1, adipocytes, and macrophages

29-31, 72

. We found identical cysteines, C3 and

329

C466, can be either nitrosated or palmitoylated on CD36. We propose ECs can use the post-

330

translational modifications of nitrosation or palmitoylation on CD36 for localization at the plasma

331

membrane and modulate lipid uptake.

332

333

Once lipids enter endothelium, their fate is not fully understood. Literature points to four known

334

outcomes for lipids, re-esterification for storage in lipid droplets, direct use for ATP production

335

through oxidative phosphorylation, remain in the cell inducing various cellular stress pathways, or

336

transported to other tissues

. Here, we show robust staining of lipid droplets in endothelium both

337

upon lipid stimulation as well as at homeostasis. We hypothesize that the excess lipids which are

338

taken into the endothelium with NOS inhibition are both stored within lipid droplets and reside

339

within the cytoplasm inducing cellular stress via lipotoxicity. Our mitochondrial respiration data

340

show when lipids are added to HAMECs the cells don’t necessarily utilize the extra fatty acids for

341

fuel. In fact, this addition of lipid decreased the cells mitochondrial function. It has been

342

demonstrated that the alteration of Cav1 dynamics in ECs can alter mitochondrial fusion and fission

343

dynamics directly causing mitochondrial dysfunction

. We hypothesize this is likely due to

344

lipotoxicity induced by lipids that were not stored, which can ultimately affect cellular respiration.

345

Our data highlights two destinations for lipids upon entry into endothelium.

346

347

Moving from the cell to the tissue, how does the excess lipid accumulation in EC causing

348

mitochondrial dysfunction alter arterial dilation, the essential physiological mechanism whereby

349

blood flow and blood pressure are regulated? Current literature focuses on NO modulating

350

endothelial-induced vasodilation. This effect of NO has been demonstrated a plethora of times in

351

large arteries, but in small arteries, it has become clear other mechanisms strongly predominate

62-

352

. In agreement with these observations, our data shows no change in ACh induced vasodilation in

353

the presence of L-NAME. However, it is the combined presence of lipids and L-NAME that

354

ultimately decreased vasodilatory capacity. Our findings are in alignment with results from ECs

355

treated with lipids and L-NAME, showing a dramatic decrease in mitochondrial function. Based on

356

our data, L-NAME moves CD36 away from the plasma membrane to the ER in control and lipid

357

treated arteries. However, it is the presence of the lipids that alter EC function by being taken up

358

into the cells when NO, or mechanisms to make NO, is deplete. In this way, our data points to

359

nitrosation, a crucial regulator of various cellular processes by protein post-translational

360

modification, as a major function for NO in small arteries.

21, 23, 74

. This may also explain the

361

difference between utilization of NO in small versus large arteries.

362

363

Currently, it is thought that elevated serum triglycerides can lead to the development of type 2

364

diabetes and metabolic syndrome

75, 76

. How then, do the EC Cav1

fl/fl

mice with elevated

365

triglycerides have an enhanced ability to process glucose? We hypothesize that due to the decrease

366

in intracellular lipids, ECs remain glucose sensitive and can maintain glycemic control. This would

367

present EC lipid accumulation and lipotoxicity as an emerging predictor for the development of

368

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

metabolic syndrome. This idea is well-supported by other studies which describe endothelial

369

dysfunction contributing to the etiology of metabolic syndrome

. Here, we demonstrate it is the

370

intracellular lipid accumulation in the microvasculature, rather than elevated serum triglycerides,

371

that plays a larger role in the onset of endothelial dysfunction.

372

373

Our work is not without alternative explanations. For example, the increased NO metabolites

374

measured in our study could be a byproduct of peroxynitrite activity

. Vascular peroxynitrite has

375

been shown to increase with obesity

79, 80

. Additionally, loss of Cav1 under pathological conditions

376

has been linked to eNOS uncoupling

. It is conceivable cysteine oxidation or nitrosation by

377

peroxynitrite in HFD could prevent palmitoylation. Because the main NO donor used for our

378

experiments (DETA/NO) provides NO rather than peroxynitrite, our results provide strong

379

evidence for NO delivering the nitrosation modification. In addition, most of the data presented

380

here is in non-obese mice where eNOS uncoupling is not prevalent. It is also possible other parts

381

of the pathway may also be nitrosated. For example, the biphasic effects of cysteine oxidation and

382

nitrosation has recently been shown to regulate key endothelial proteins like vascular endothelial

383

growth factor (VEGF)

and Cav1

21, 23

. It is even possible eNOS itself may be both nitrosated and

384

palmitoylated which could help move CD36 to and from the plasma membrane. Further work is

385

warranted in these areas.

386

Last, elevated serum lipid levels are clinically characterized as hyperlipidemia which continues to

387

be a crucial indicator of metabolic syndrome and cardiovascular disease risk. For patients with

388

elevated cholesterol, statins are a first line prescription to manage the progression of hyperlipidemia

389

by targeting the LDL receptor and the synthesis of cholesterol in the liver. However, up to 30% of

390

patients on statins are intolerant and never reach healthy cholesterol levels

. This represents nearly

391

200 million people worldwide that have uncontrolled serum lipid levels. In addition, statin use does

392

not come without risks, including increased risk of diabetes and damaged muscle. Thus, there is a

393

need for a new lipid lowering modalities. CD36 is well-documented to regulate lipid uptake into

394

cells but, crucial parts of this fatty acid translocases’ biology are not fully understood. Exploiting

395

the mechanisms by which CD36 imports lipids could be a novel way to manage hyperlipidemia.

396

Our work adds important mechanistic insight into how the properties of CD36 could be targeted for

397

drug development by focusing on balancing the post-translational modification of palmitoylation

398

and nitrosation

399

400

401

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

Materials and Methods

402

Animals

403

Only male mice were used, 20-23 weeks of age, on a C57Bl/6 genetic background, and were cared

404

for under the provisions of the University of Virginia Animal Care and Use Committee and

405

followed the National Institutes of Health guidelines for the care and use of laboratory animals.

406

Animals were subject to a 12hr light dark cycle. Mice were housed on standard corn cobb bedding

407

expect for metabolic studies where mice were fasted overnight on wood chip bedding. A C57BL/6n

408

mice were purchased from Taconic. The inducible, EC-specific Cav1 knockout mice

409

(Cdh5ER

T2+

/Cav1

fl/fl

) were generated by crossing Cdh5ER

T2+

/Cav1

WT/WT

mice (a kind gift from Dr

410

Ralf Adams, Max Plank Institute, Germany) with Cdh5ER

T2-

/Cav1

fl/fl

mice

79, 84, 85

. To conditionally

411

induce Cav1 deletion in the vascular endothelium, Cdh5ER

T2-

/Cav1

fl/fl

(EC Cre

Cav1

fl/f

) and

412

Cdh5ER

T2+

/Cav1

fl/fl

(EC Cre

Cav1

fl/fl

) littermates received intraperitoneal (I.P.) injections of

413

Tamoxifen (1 mg in 0.1 ml peanut oil) at six weeks of age for 10 consecutive days. For diet studies,

414

littermates were randomly assigned to normal chow (NC; 5% Kcal from fat, house chow) or high

415

fat diet (HFD; 60% Kcal from fat Bio-Serv-The Foster Corp F3282) starting at 8 weeks of age for

416

12 consecutive weeks. Additionally, C57BL6J (strain #: 000664) and eNOS transgenic mice

417

(strain#: 002684) were purchased from Jackson Laboratories and used between the ages of 20-22

418

weeks. The eNOS

fl/fl

mice on both the endothelial and red blood cell (RBC) cre were used as

419

described in

. Age of all mice at the completion of NC or HFD are 20 weeks. For L-NAME

420

administration to NC mice, L-NAME (98%, Fisher Scientific: AAH6366606) was provided in the

421

drinking water at 60mg/dL for 14 days. For L-NAME administration to HFD mice, L-NAME was

422

provided in the drinking water at 100mg/dL for 14 days. All experiments were performed on a

423

minimum of four mice. For all assessments of blood, blood was collected via terminal cardiac

424

puncture using a syringe fitted with 25G needle, coated with EGTA to prevent clotting and

425

deposited in gold cap blood collection tubes. Blood lipids (Cholesterol, LDL, HDL, and

426

triglycerides) were processed by UVA clinical laboratory. Serum nitrite and nitrate levels were

427

measured as described in Leo et al. 2021

. Non-esterified fatty acids (NEFA) were measured using

428

Fujifilm NEFA absorbance kit (Fujifilm Medical systems: 633-52001). Diet and experimental

429

procedures for ROSA26-eYFP

+/+

Cdh5-CreER

T2+

mice used for scRNAseq are described in depth

430

in Dunaway and Luse et al. 2023.

431

432

Genomic excision

433

DNA was extracted from lung tissue and digested using proteinase K (Bioline: BIO-37084) for

434

genomic excision gels. A set of three primers were used F: TTCTGTGTGCAAGCCTTTCC, R1:

435

GTGTGCGCGTCATACACTTG, R2: GGGGAGGAGTAGAAGGTGGC. Non-excised Cav1

436

mice have a band at ~7Kb and Cav1 excision is seen as an absence of a band.

437

438

Tissue culture

439

Human adipose microvascular endothelial cells (HAMECs) were acquired from ScienCell (#7200)

440

and cultured according to supplier guidelines. All HAMECs used were under passage 10. mRNA

441

knockdown was performed using siRNA; Cav1 (30nM, Origene: SR319567) and CD36 (10nM,

442

Thermo Fisher: s534752). 10mM Dimethyl-b-cyclodextrin (Thermo: 235530050) was used to

443

deplete HAMECs of cholesterol. Knockdown of Cav1 was achieved via nucleofection using a

444

LONZA nucleofector

, while lipofectamine (Thermo Fisher: L3000008) was used for knockdown

445

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

of CD36. Briefly, cells were washed with PBS and incubated in serum free media containing 10mM

446

dimethyl-b-cyclodextrin at 37°C for 1hr. Cholesterol depletion was validated using a

447

Cholesterol/Cholesterol ester-glo assay (Promega: J3190). Endocytosis was blocked using two

448

drugs Dynosore (selleckchem: S8047) and Genistein (Fisher: G02721) at 80µM and 200µM

449

respectively for 30min at 37°C. For lipid treatment experiments, cells were treated with 50µM lipids

450

(12:5µM linoleic acid, 25µM oleic acid, 12.5µM palmitic acid). Briefly, lipid stocks were made in

451

ethanol and for each experiment the appropriate volume of lipids were dried down using nitrogen

452

gas and then resuspended in cell culture media. The media was then sonicated and vortexed three

453

times each and allowed to warm to 37°C. Lipid media was passed through a 0.45µM filter before

454

being added to cells for 30 minutes. All cells to be imaged were washed once with 1X PBS and

455

then fixed with 4% paraformaldehyde for 10 minutes at room temperature. Cellular lipid staining

456

was accomplished through the use of BODIPY493/503 (Thermo Fisher: D3922) dissolved first in

457

DMSO at a concentration of 1mg/mL then further diluted in PBS to fixed cells at 1µM. Cells were

458

treated with L-NAME at concentration of 1mMol for 18hours either 72hrs after a control or siCav1

459

transfection. HAMECs were transfected with mCherry-CD36-C-10 (Addgene: Plasmid #55011),

460

8µg of plasmid was used per 1 million cells using nucleofection.

461

462

HEK293T were cultured in DMEM, high glucose (Gibco, 11965-092) supplemented with

463

1% Pen-Strep (Gibco, 15140122), 10% FBS (Avantor: 97068-085). Cells were transfected with

464

mCherry-CD36-C-10 (Addgene: Plasmid #55011), 5µg per 10cm dish or 0.5µg per well of a 6-well

465

dish, using Lipofectamine 3000 (Thermo Fisher: L3000001). Point mutations to the mCherry-

466

CD36-C-10 plasmids were generated with the help of Genscript. HEK293T cells were treated with

467

NO donors either DETA/NO-NONoate (DETA/NO: 50µM 30 minutes, Fisher: AC328650250) or

468

S-Nitrosoglutathione (SNOG: 100µM 30 minutes, Fisher: 06-031-0) for nitrosation experiments.

469

HEK293T cells treated with lipids followed the same procedure state above for HAMECs. Cells

470

were fixed for 10min at room temperature with 4% paraformaldehyde and then stained with bodipy

471

in PBS at 1µM.

472

473

Lipid raft fractionation

474

HAMECs were homogenized in 25mM MES and 150mM NaCl then centrifuged at 1000g for

475

10min at 4°C. Supernatant was collected and mixed with various percentages of Optiprep (Sigma:

476

D1556) to create a discontinuous density gradient. Samples were centrifuged at 380 x 1000 RPM

477

for 2 hours at 4°C in a Beckman ultracentrifuge. Post centrifugation fractions were removed from

478

top to bottom to isolate lipid rafts based on density. Lipid raft fractionation was performed 72hrs

479

after SiCav1 or control treatment.

480

481

Biotin Switch

482

Biotin switch assay was performed using the S-Nitrosylated Protein Detection Kit (Cayman

483

chemical: 10006518) and HEK293T cells. Briefly, cells were washed with supplied wash buffer

484

and free thiols were blocked. Samples were split into two providing internal controls where one

485

sample did not receive the reducing and detection agents (biotin). After biotin was added to thiols

486

containing nitrations, streptavidin beads (Thermo Fisher: 65601) were used to isolate

487

biotinylated/nitrated proteins. Protein bound to the beads was eluted using 5x laemmli buffer at

488

room temperature for 15 minutes and then heated at 95°C for 1 minutes. Sample was loaded

489

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

according to western blotting procedures (below) and immunoblotted for CD36 (Thermo fisher

490

PA1-16913) or mcherry (abcam: ab213511) tag on CD36.

491

492

Real-time quantitative PCR

493

Total RNA was extracted from mouse tissues using the Aurum Total RNA Fatty and Fibrous Tissue

494

Extraction kit (Biorad: #732-6870). RNA from cells was extracted using Zymo Research R1055

495

Quick-RNA MiniPrep Kit, Zymo Research Kit (Genesee: 11-328). RNA concentration was

496

measured using the Nanodrop1000 spectrophotometer (Thermo Fisher). RNA was stored at -80°C

497

before reverse transcription with SuperScript III First-Strand Synthesis system (Thermo Fisher:

498

18080051) using random hexamer primers on 1μg of template RNA. Real-time quantitative PCR

499

was performed using Taqman Gene Expression Master Mix (Thermo Fisher: 4369016) and Taqman

500

Real-Time PCR assays in MGB-FAM for Cav1 (Hs00971716_m1; Mm00432403_m1), CD36

501

(Hs00354519_m1; Mm00432403_m1), and were normalized to β-2-microglobin/B2M in VIC-PL

502

(Hs00364808_m1; Mm00437762_m1). Reactions were run in a CFX Real-Time Detection System

503

(Applied BioSystems) and threshold cycle number (CT) was used as part of the 2-DDCT method

504

to calculate fold change from control.

505

506

Western Blotting

507

Cell and tissue lysates were generated in RIPA (50mmol/L Tris-HCL, 150mmol/L NaCl, 5mmol/L

508

EDTA,1% deoxycholate, 1% Triton-X100) in PBS and pH adjusted to 7.4) supplemented with

509

protease inhibitor cocktail (Sigma: P8340). Lysates were rocked at 4°C for 30-60 min to solubilize

510

proteins, sonicated briefly, and centrifuged for 15 min at 12,000 rpm to pellet cell debris. Protein

511

concentration was determined using the Pierce BCA method (Thermo Fisher: 23227). 20μg of total

512

protein was loaded into each sample well. Samples were subjected to SDS gel electrophoresis using

513

8% or 4-12% Bis-Tris gels (Invitrogen) and transferred nitrocellulose membranes for

514

immunoblotting. Membranes were blocked for 1 hour at room temperature in a solution containing

515

3% BSA in Tris buffered saline, then incubated overnight at 4°C with primary antibodies against

516

Cav1 (abcam: ab32577, 1:2000), CD36 (Human tissue Thermofisher: PA1-16813 1:500, mouse

517

tissue abcam: ab124515 1:500), Calnexin (abcam: ab22595 1:1000), and eNOS (BD transduction:

518

610297 1:1000). Membranes were washed and incubated in LiCOR IR Dye secondary antibodies

519

(1:10,000) for 1 hour and viewed/quantified using the LiCOR Odyssey CLx with Image Studio

520

software. Licor Total Protein stain was used for loading normalization. Representative western blot

521

images have been cropped for presentation.

522

523

Single-Cell RNA sequencing (scRNAseq)

524

A previously published scRNA-seq dataset from human adipose tissue was used to create these

525

data

. Endothelial cells were subsetted, using subset() function, from previously published data

526

using clusters defined by the original authors and confirmed by our arterial, venous, capillary, and

527

lymphatic endothelial markers. Endothelial cells from visceral adipose tissue were used for analysis

528

for more appropriate comparison with epididymal adipose ECs.

529

530

scRNAseq from mouse adipose endothelium has been previously described in

. Briefly, the

531

generation of single cell indexed libraries was performed by the School of Medicine Genome

532

Analysis and Technology Core, RRID:SCR_018883, using the 10X Genomics chromium controller

533

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

platform and the Chromium Single Cell 3′ Library & Gel Bead Kit v3.1 reagent. Around 5,000

534

cells were targeted per sample and loaded onto each well of a Chromium Single Cell G Chip to

535

generate single cell emulsions primed for reverse transcription. After breaking the emulsion, the

536

single cell specific barcoded DNAs were subjected to cDNA amplification and QC on the Agilent

537

4200 TapeStation Instrument, using the Agilent D5000 kit. A QC run was performed on the

538

Illumina Miseq using the nano 300Cycle kit (1.4 Million reads/run), to estimate the number of

539

targeted cells per sample using the Cellranger 3.0.2 function. After run completion, the Binary base

540

call (bcl) files were converted to fastq format using the Illumina bcl2fastq2 software raw reads in

541

fastq files were mapped to the mm10 reference murine genome and assigned to individual cells by

542

CellRanger 5.0. Data from two separate experiments representing cells from 12 mice were analyzed

543

in RStudio (2022.07.1) with the Seurat package (4.3.0). Sequencing yielded 27,944 cells with

544

54,100 features. To ensure high quality data, cells were excluded if they contained less than 200

545

genes, more than 5000 genes, if their transcriptome was more than 5% mitochondrial encoded, and

546

more than 5% hemoglobin beta. This resulted in a final data set of 17,164 cells. Data was combined

547

using SCTransform, normalized, and 3000 variable features were chosen. UMAPs were generated

548

using 20 principal components. Clusters were generated using a resolution of 1. Non-endothelial

549

cells were excluded based on low expression of Pecam1 and Cdh5 as well as high expression of

550

non-endothelial markers (Col1a1, Acta2, Cd3g, Ptprc, Ccr5, Adipoq).

551

552

En Face Imaging

553

Epididymal fat pad arteries (100-200µm in diameter) were collected and stripped of adipose and

554

connective tissue. Arteries were subsequently fixed in 4% paraformaldehyde on ice for 10 minutes.

555

For en face preparation, arteries were cut longitudinally with microdissection scissors and pinned

556

open on polymerized Sylgard 184 (Electron Microscopy Sciences) using tungsten wire (0.0005”,

557

ElectronTubeStore). For immunostaining, vessels were then permeabilized with 0.5% TritonX100

558

in PBS for 30 minutes at room temperature and blocked in 1% bovine serum albumin, Fraction V

559

(BSA Sigma) in 0.5% TritonX100/PBS. Primary antibody staining was performed overnight at 4°C

560

in 0.1% BSA in 0.5% TritonX100/PBS. Cav1 (Abcam: ab211503, 1:200), CD36 (abcam:

561

ab124515, 1:100), Calnexin (abcam: ab219644, 1:100) primary antibodies were used. Samples

562

were incubated with secondary antibodies at 1:500 for 1-2 hours at room temperature. For Nile red

563

staining, samples were immediately stained after pinning. Nile red (Thermo: N1142) was used at

564

5µM in PBS for 30min at room temperature. Nuclei were stained using DAPI (Invitrogen: D1306,

565

final concentration 0.1mg/mL) in addition to mounting with Prolong Gold Antifade Mountant

566

(Invitrogen: P36930). Images were collected using an Olympus FV3000 with a 60X oil emersion

567

lens and post processing was completed using FIJI.

568

569

Mitochondrial Stress Test

570

HAMECs were plated (80,000 cells per well) in complete medium in XFe96 cell culture microplate

571

and allowed to settle for one hour at room temperature before overnight culture. Sixteen hours prior

572

to assessment HAMECs were treated with L-NAME (see above) and subsequently four hours prior

573

to assessment HAMECs were treated with 50 μM lipids as well as DETA/NO (see above).

574

Immediately prior to the assay the media was changed to mitochondrial stress test medium

575

(Corning: 50-003-PB) including L-NAME or DETA/NO as indicated. Mitochondrial activity was

576

assessed by measurement of O

consumption rate on a Seahorse XFe96 instrument (Agilent

577

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

Technologies). The rate of O

change was measured every 13 minutes for a 3-minute interval before

578

sequential challenge with 1) 1 μM oligomycin (Sigma-Aldrich: 75351), 2) 2 μM BAM15

579

(Cayman Chemical Company: 17811) and 3) 10 μM antimycin A (Sigma-Aldrich: A8674) and 10

580

μM rotenone (Sigma-Aldrich: R88751G). Maximal OCR was measured as post-BAM15 OCR

581

minus post-antimycin A/rotenone OCR. Data represented here have been normalized (pmol O

/μg

582

total protein) to total protein in each well of assay plate using BCA assay (Thermo: 23227).

583

584

RNA sequencing

585

ECs were treated as described above with the same methods used for “Mitochondrial stress test”.

586

RNA was isolated using Quick-RNA MiniPrep Kit (Genesee Scientific: 11-328). When necessary,

587

samples were further concentrated and purified with the RNeasy MinElute Cleanup Kit (Qiagen:

588

74804). PE150 reads were generated with the Illumina NovoSeq platform. The paired-end reads

589

were mapped to the hg38 reference genome using STAR (v 2.7.9a). Gene counts were generated

590

with FeatureCounts in the Rsubread (v. 2.8.8) package. DESeq2 (v. 1.34.0) was used to calculate

591

log fold change and adjusted p-values. The adjusted p-values were then adjusted for multiple

592

comparisons between groups by the Benjamini-Hochberg procedure using p.adjust(). Ensemble IDs

593

were converted to gene symbols using the EnsDb.Hsapiens.v75 and org.Hs.eg.db packages. For

594

gene set enrichment analysis, genes were ranked by log2foldchange and pathway analysis was

595

conducted with clusterprofiler (v 4.2.2). Raw data and normalized counts have been deposited in

596

NCBI’s Gene Expression Omnibus and are accessible through the GEO accession (GSE260897).

597

598

Pressure Myography

599

Third order mesenteric arteries (diameter between 100-200µm) were removed from 14-week old

600

C57BL/6J mice. Arteries were cannulated on two glass pipettes tips and equilibrated to 80mmHg

601

as previously described

. For vasoreactivity measurements vessels were either incubated with

602

50µM L-NAME or water for 20 minutes. Subsequently for vessels receiving lipid treatments 10µM

603

lipids (10µM oleic acid solubilized in ethanol) were added to the circulating buffer for an additional

604

20 minutes. Vessels were pre-constricted with 1µM phenylephrine (PE) and then exposed to serial

605

doses of Acetylcholine (ACh) to assess vasodilation. Three male mice were used for the

606

experiments using lipids with 1-2 vessels per mouse averaged for each sample value. Five mice

607

were used for experiments without lipids with 1-2 vessels averaged per each sample value.

608

609

Acyl–biotinyl exchange and Western blotting to analyze palmitoylation

610

Palmitoylation analysis was done using acyl–biotinyl exchange (ABE) as previously described

611

One 10 cm plate with HEK-293 cells at 80-90% confluence was lysed in 2% SDS-containing

612

buffer and free cysteines were blocked by 10 mM NEM. Then, palmitoylated cysteines were

613

liberated by 0.4 M hydroxylamine and labeled with biotin-HPDP (Pierce). Biotinylated proteins

614

were pulled down using streptavidin-magnetic beads (Dynabeads, Thermo Scientific) and eluted

615

with 1% 2-mercaptoethanol. Three chloroform/methanol precipitations were performed between

616

each step to remove chemicals. After elution, a Western blot of the eluate (palmitoylated fraction)

617

and input (total protein) was performed for the different Rab proteins (endogenous) or TMD

618

probes (anti-RFP). For quantification, densitometry analysis was performed in BioRAD

619

ChemiDoc MP.

620

621

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

Statistical Analysis

622

Statistical analysis was performed using Prism GraphPad 9 and R studio for scRNAseq and bulk

623

RNA sequencing analysis. Specific statistical test type and N values are listed in each

624

corresponding figure legend. Data represented are the mean with SEM (+ and -) shown as error

625

bars. All N values represent different experiments (in vitro) and individual mice (in vivo). For

626

image quantification, three fields of view were used to average each N value.

627

628

629

630

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

References

631

1. Mehrotra D, Wu J, Papangeli I and Chun HJ. Endothelium as a gatekeeper of fatty acid

632

transport. Trends in Endocrinology & Metabolism. 2014;25:99-106.

633

2. Yazdani S, Jaldin‐Fincati JR, Pereira RV and Klip A. Endothelial cell barriers: Transport

634

of molecules between blood and tissues. Traffic. 2019;20:390-403.

635

3. Wang H, Liu Z, Li G and Barrett EJ. The vascular endothelial cell mediates insulin

636

transport into skeletal muscle. American Journal of Physiology-Endocrinology and Metabolism.

637

2006;291:E323-E332.

638

4. Kuo A, Lee MY and Sessa WC. Lipid droplet biogenesis and function in the endothelium.

639

Circulation research. 2017;120:1289-1297.

640

5. Stoletov K, Fang L, Choi S-H, Hartvigsen K, Hansen LF, Hall C, Pattison J, Juliano J,

641

Miller ER and Almazan F. Vascular lipid accumulation, lipoprotein oxidation, and macrophage

642

lipid uptake in hypercholesterolemic zebrafish. Circulation research. 2009;104:952-960.

643

6. Jang C, Oh SF, Wada S, Rowe GC, Liu L, Chan MC, Rhee J, Hoshino A, Kim B and

644

Ibrahim A. A branched-chain amino acid metabolite drives vascular fatty acid transport and

645

causes insulin resistance. Nature medicine. 2016;22:421-426.

646

7. Listenberger LL, Han X, Lewis SE, Cases S, Farese Jr RV, Ory DS and Schaffer JE.

647

Triglyceride accumulation protects against fatty acid-induced lipotoxicity. Proceedings of the

648

National Academy of Sciences. 2003;100:3077-3082.

649

8. Unger RH and Scherer PE. Gluttony, sloth and the metabolic syndrome: a roadmap to

650

lipotoxicity. Trends in Endocrinology & Metabolism. 2010;21:345-352.

651

9. Boutagy NE, Gamez-Mendez A, Fowler JW, Zhang H, Chaube BK, Esplugues E, Lee S,

652

Horikami D, Zhang J and Citrin KM. Dynamic metabolism of endothelial triglycerides protects

653

against atherosclerosis in mice. The Journal of Clinical Investigation. 2024.

654

10. Boutagy NE, Singh AK and Sessa WC. Targeting the vasculature in cardiometabolic

655

disease. The Journal of Clinical Investigation. 2022;132.

656

11. Simionescu M. Implications of early structural-functional changes in the endothelium for

657

vascular disease. Arteriosclerosis, thrombosis, and vascular biology. 2007;27:266-274.

658

12. Fernández-Hernando C, Yu J, Suárez Y, Rahner C, Dávalos A, Lasunción MA and Sessa

659

WC. Genetic evidence supporting a critical role of endothelial caveolin-1 during the progression

660

of atherosclerosis. Cell metabolism. 2009;10:48-54.

661

13. Fernández-Hernando C, Yu J, Dávalos A, Prendergast J and Sessa WC. Endothelial-

662

specific overexpression of caveolin-1 accelerates atherosclerosis in apolipoprotein E-deficient

663

mice. The American journal of pathology. 2010;177:998-1003.

664

14. Wang H, Wang AX and Barrett EJ. Caveolin-1 is required for vascular endothelial insulin

665

uptake. American Journal of Physiology-Endocrinology and Metabolism. 2011;300:E134-E144.

666

15. Kakava S, von Eckardstein A and Robert J. Regulation of low-density lipoprotein

667

transport through endothelial cells by caveolae. Atherosclerosis. 2023;375:84-86.

668

16. Le Master E, Paul A, Lazarko D, Aguilar V, Ahn SJ, Lee JC, Minshall RD and Levitan I.

669

Caveolin-1 is a primary determinant of endothelial stiffening associated with dyslipidemia,

670

disturbed flow, and ageing. Scientific Reports. 2022;12:17822.

671

17. Shamsaldeen YA, Ugur R, Benham CD and Lione LA. Diabetic dyslipidaemia is

672

associated with alterations in eNOS, caveolin‐1, and endothelial dysfunction in streptozotocin

673

treated rats. Diabetes/Metabolism Research and Reviews. 2018;34:e2995.

674

18. Li X, Sun X and Carmeliet P. Hallmarks of endothelial cell metabolism in health and

675

disease. Cell metabolism. 2019;30:414-433.

676

19. Matthaeus C, Lahmann I, Kunz S, Jonas W, Melo AA, Lehmann M, Larsson E, Lundmark

677

R, Kern M and Blüher M. EHD2-mediated restriction of caveolar dynamics regulates cellular

678

fatty acid uptake. Proceedings of the National Academy of Sciences. 2020;117:7471-7481.

679

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

20. Bernatchez P, Sharma A, Bauer PM, Marin E and Sessa WC. A noninhibitory mutant of

680

the caveolin-1 scaffolding domain enhances eNOS-derived NO synthesis and vasodilation in

681

mice. The Journal of clinical investigation. 2011;121.

682

21. Chen Z, DS Oliveira S, Zimnicka AM, Jiang Y, Sharma T, Chen S, Lazarov O, Bonini

683

MG, Haus JM and Minshall RD. Reciprocal regulation of eNOS and caveolin-1 functions in

684

endothelial cells. Molecular Biology of the Cell. 2018;29:1190-1202.

685

22. Leo F, Suvorava T, Heuser SK, Li J, LoBue A, Barbarino F, Piragine E, Schneckmann R,

686

Hutzler B and Good ME. Red blood cell and endothelial eNOS independently regulate circulating

687

nitric oxide metabolites and blood pressure. Circulation. 2021;144:870-889.

688

23. Bakhshi FR, Mao M, Shajahan AN, Piegeler T, Chen Z, Chernaya O, Sharma T, Elliott

689

WM, Szulcek R and Bogaard HJ. Nitrosation-dependent caveolin 1 phosphorylation,

690

ubiquitination, and degradation and its association with idiopathic pulmonary arterial

691

hypertension. Pulmonary circulation. 2013;3:816-830.

692

24. Son N-H, Basu D, Samovski D, Pietka TA, Peche VS, Willecke F, Fang X, Yu S-Q,

693

Scerbo D and Chang HR. Endothelial cell CD36 optimizes tissue fatty acid uptake. The Journal of

694

clinical investigation. 2018;128:4329-4342.

695

25. Abumrad NA, Cabodevilla AG, Samovski D, Pietka T, Basu D and Goldberg IJ.

696

Endothelial cell receptors in tissue lipid uptake and metabolism. Circulation research.

697

2021;128:433-450.

698

26. Peche V, Pietka T, Jacome-Sosa M, Samovski D, Palacios H, Chatterjee-Basu G, Dudley

699

A, Beatty W, Meyer G and Goldberg I. Endothelial cell CD36 regulates membrane ceramide

700

formation, exosome fatty acid transfer and circulating fatty acid levels. Nature communications.

701

2023;14:4029.

702

27. Dawson DW, Pearce SFA, Zhong R, Silverstein RL, Frazier WA and Bouck NP. CD36

703

mediates the in vitro inhibitory effects of thrombospondin-1 on endothelial cells. The Journal of

704

cell biology. 1997;138:707-717.

705

28. Love-Gregory L, Sherva R, Sun L, Wasson J, Schappe T, Doria A, Rao DC, Hunt SC,

706

Klein S and Neuman RJ. Variants in the CD36 gene associate with the metabolic syndrome and

707

high-density lipoprotein cholesterol. Human molecular genetics. 2008;17:1695-1704.

708

29. Hao J-W, Wang J, Guo H, Zhao Y-Y, Sun H-H, Li Y-F, Lai X-Y, Zhao N, Wang X and

709

Xie C. CD36 facilitates fatty acid uptake by dynamic palmitoylation-regulated endocytosis.

710

Nature communications. 2020;11:4765.

711

30. Wang J, Hao J-W, Wang X, Guo H, Sun H-H, Lai X-Y, Liu L-Y, Zhu M, Wang H-Y and

712

Li Y-F. DHHC4 and DHHC5 facilitate fatty acid uptake by palmitoylating and targeting CD36 to

713

the plasma membrane. Cell reports. 2019;26:209-221. e5.

714

31. Tao N, Wagner SJ and Lublin DM. CD36 is palmitoylated on both N-and C-terminal

715

cytoplasmic tails. Journal of Biological Chemistry. 1996;271:22315-22320.

716

32. Guan X and Fierke CA. Understanding protein palmitoylation: Biological significance and

717

enzymology. Science China Chemistry. 2011;54:1888-1897.

718

33. Greaves J, Carmichael JA and Chamberlain LH. The palmitoyl transferase DHHC2 targets

719

a dynamic membrane cycling pathway: regulation by a C-terminal domain. Molecular biology of

720

the cell. 2011;22:1887-1895.

721

34. Zhang MM, Tsou LK, Charron G, Raghavan AS and Hang HC. Tandem fluorescence

722

imaging of dynamic S-acylation and protein turnover. Proceedings of the National Academy of

723

Sciences. 2010;107:8627-8632.

724

35. Salaun C, Greaves J and Chamberlain LH. The intracellular dynamic of protein

725

palmitoylation. Journal of Cell Biology. 2010;191:1229-1238.

726

36. Kharitonov VG, Sundquist AR and Sharma VS. Kinetics of nitrosation of thiols by nitric

727

oxide in the presence of oxygen. Journal of Biological Chemistry. 1995;270:28158-28164.

728

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

37. Stamler JS, Simon DI, Osborne JA, Mullins ME, Jaraki O, Michel T, Singel DJ and

729

Loscalzo J. S-nitrosylation of proteins with nitric oxide: synthesis and characterization of

730

biologically active compounds. Proceedings of the National Academy of Sciences. 1992;89:444-

731

448.

732

38. Lohman AW, Weaver JL, Billaud M, Sandilos JK, Griffiths R, Straub AC, Penuela S,

733

Leitinger N, Laird DW and Bayliss DA. S-nitrosylation inhibits pannexin 1 channel function.

734

Journal of Biological Chemistry. 2012;287:39602-39612.

735

39. Gould N, Doulias P-T, Tenopoulou M, Raju K and Ischiropoulos H. Regulation of protein

736

function and signaling by reversible cysteine S-nitrosylation. Journal of Biological Chemistry.

737

2013;288:26473-26479.

738

40. Dunaway LS, Luse MA, Nyshadham S, Bulut G, Alencar GF, Chavkin NW, Cortese-Krott

739

MM, Hirschi KK and Isakson BE. Obesogenic diet disrupts tissue specific mitochondrial gene

740

signatures in the artery and capillary endothelium. Physiological Genomics. 2023.

741

41. Emont MP, Jacobs C, Essene AL, Pant D, Tenen D, Colleluori G, Di Vincenzo A,

742

Jørgensen AM, Dashti H and Stefek A. A single-cell atlas of human and mouse white adipose

743

tissue. Nature. 2022;603:926-933.

744

42. Porta JC, Han B, Gulsevin A, Chung JM, Peskova Y, Connolly S, Mchaourab HS, Meiler

745

J, Karakas E and Kenworthy AK. Molecular architecture of the human caveolin-1 complex.

746

Science Advances. 2022;8:eabn7232.

747

43. Shu X, Keller TS, Begandt D, Butcher JT, Biwer L, Keller AS, Columbus L and Isakson

748

BE. Endothelial nitric oxide synthase in the microcirculation. Cellular and molecular life

749

sciences. 2015;72:4561-4575.

750

44. Ju H, Zou R, Venema VJ and Venema RC. Direct interaction of endothelial nitric-oxide

751

synthase and caveolin-1 inhibits synthase activity. Journal of Biological Chemistry.

752

1997;272:18522-18525.

753

45. Murata T, Lin MI, Huang Y, Yu J, Bauer PM, Giordano FJ and Sessa WC. Reexpression

754

of caveolin-1 in endothelium rescues the vascular, cardiac, and pulmonary defects in global

755

caveolin-1 knockout mice. The Journal of experimental medicine. 2007;204:2373-2382.

756

46. Mahammad S and Parmryd I. Cholesterol depletion using methyl-β-cyclodextrin. Methods

757

in membrane lipids. 2015:91-102.

758

47. Martens JR, Sakamoto N, Sullivan SA, Grobaski TD and Tamkun MM. Isoform-specific

759

localization of voltage-gated K+ channels to distinct lipid raft populations: targeting of Kv1. 5 to

760

caveolae. Journal of Biological Chemistry. 2001;276:8409-8414.

761

48. Seto S, Krishna S, Yu H, Liu D and Khosla S. Impaired Acetylcholine-Induced

762

Endothelium-Dependent Aortic Relaxation by Caveolin-1 in. 2013.

763

49. Rees D, Palmer R, Schulz R, Hodson H and Moncada S. Characterization of three

764

inhibitors of endothelial nitric oxide synthase in vitro and in vivo. British journal of

765

pharmacology. 1990;101:746.

766

50. Silva FCd, Araújo BJd, Cordeiro CS, Arruda VM, Faria BQ, Guerra JFDC, Araújo TGD

767

and Fürstenau CR. Endothelial dysfunction due to the inhibition of the synthesis of nitric oxide:

768

Proposal and characterization of an in vitro cellular model. Frontiers in Physiology.

769

2022;13:978378.

770

51. Febbraio M and Silverstein RL. CD36: implications in cardiovascular disease. The

771

international journal of biochemistry & cell biology. 2007;39:2012-2030.

772

52. Pohl J, Ring A, Korkmaz U, Ehehalt R and Stremmel W. FAT/CD36-mediated long-chain

773

fatty acid uptake in adipocytes requires plasma membrane rafts. Molecular biology of the cell.

774

2005;16:24-31.

775

53. Ring A, Le Lay S, Pohl J, Verkade P and Stremmel W. Caveolin-1 is required for fatty

776

acid translocase (FAT/CD36) localization and function at the plasma membrane of mouse

777

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

embryonic fibroblasts. Biochimica et Biophysica Acta (BBA)-Molecular and Cell Biology of

778

Lipids. 2006;1761:416-423.

779

54. Gerbod-Giannone M-C, Dallet L, Naudin G, Sahin A, Decossas M, Poussard S and

780

Lambert O. Involvement of caveolin-1 and CD36 in native LDL endocytosis by endothelial cells.

781

Biochimica et Biophysica Acta (BBA)-General Subjects. 2019;1863:830-838.

782

55. Shvets E, Bitsikas V, Howard G, Hansen CG and Nichols BJ. Dynamic caveolae exclude

783

bulk membrane proteins and are required for sorting of excess glycosphingolipids. Nature

784

communications. 2015;6:6867.

785

56. Lim JE, Bernatchez P and Nabi IR. Scaffolds and the scaffolding domain: an alternative

786

paradigm for caveolin-1 signaling. Biochemical Society Transactions. 2024:BST20231570.

787

57. Stamler JS. Redox signaling: nitrosylation and related target interactions of nitric oxide.

788

Cell. 1994;78:931-936.

789

58. Forrester MT, Foster MW, Benhar M and Stamler JS. Detection of protein S-nitrosylation

790

with the biotin-switch technique. Free Radical Biology and Medicine. 2009;46:119-126.

791

59. Xue Y, Liu Z, Gao X, Jin C, Wen L, Yao X and Ren J. GPS-SNO: computational

792

prediction of protein S-nitrosylation sites with a modified GPS algorithm. PloS one.

793

2010;5:e11290.

794

60. Pepino MY, Kuda O, Samovski D and Abumrad NA. Structure-function of CD36 and

795

importance of fatty acid signal transduction in fat metabolism. Annual review of nutrition.

796

2014;34:281-303.

797

61. May SC and Sahoo D. Now in 3D! Novel insights into CD36 structure and function. Ann

798

Blood. 2021;6:1-12.

799

62. Ottolini M, Daneva Z, Chen YL, Cope EL, Kasetti RB, Zode GS and Sonkusare SK.

800

Mechanisms underlying selective coupling of endothelial Ca2+ signals with eNOS vs. IK/SK

801

channels in systemic and pulmonary arteries. The Journal of physiology. 2020;598:3577-3596.

802

63. de Wit C, Jahrbeck B, Schfer C, Bolz S-S and Pohl U. Nitric oxide opposes myogenic

803

pressure responses predominantly in large arterioles in vivo. Hypertension. 1998;31:787-794.

804

64. Bolz SS, De Wit C and Pohl U. Endothelium‐derived hyperpolarizing factor but not NO

805

reduces smooth muscle Ca2+ during acetylcholine‐induced dilation of microvessels. British

806

journal of pharmacology. 1999;128:124-134.

807

65. Garland J and McPherson GA. Evidence that nitric oxide does not mediate the

808

hyperpolarization and relaxation to acetylcholine in the rat small mesenteric artery. British

809

journal of pharmacology. 1992;105:429.

810

66. Earley S, Pauyo T, Drapp R, Tavares MJ, Liedtke W and Brayden JE. TRPV4-dependent

811

dilation of peripheral resistance arteries influences arterial pressure. American Journal of

812

Physiology-Heart and Circulatory Physiology. 2009;297:H1096-H1102.

813

67. Razani B, Engelman JA, Wang XB, Schubert W, Zhang XL, Marks CB, Macaluso F,

814

Russell RG, Li M and Pestell RG. Caveolin-1 null mice are viable but show evidence of

815

hyperproliferative and vascular abnormalities. Journal of Biological Chemistry. 2001;276:38121-

816

38138.

817

68. Cohen AW, Razani B, Wang XB, Combs TP, Williams TM, Scherer PE and Lisanti MP.

818

Caveolin-1-deficient mice show insulin resistance and defective insulin receptor protein

819

expression in adipose tissue. American Journal of Physiology-Cell Physiology. 2003;285:C222-

820

C235.

821

69. Pojoga LH, Underwood PC, Goodarzi MO, Williams JS, Adler GK, Jeunemaitre X,

822

Hopkins PN, Raby BA, Lasky-Su J and Sun B. Variants of the caveolin-1 gene: a translational

823

investigation linking insulin resistance and hypertension. The Journal of Clinical Endocrinology

824

& Metabolism. 2011;96:E1288-E1292.

825

70. Ibrahimi A, Bonen A, Blinn WD, Hajri T, Li X, Zhong K, Cameron R and Abumrad NA.

826

Muscle-specific overexpression of FAT/CD36 enhances fatty acid oxidation by contracting

827

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

muscle, reduces plasma triglycerides and fatty acids, and increases plasma glucose and insulin.

828

Journal of Biological Chemistry. 1999;274:26761-26766.

829

71. Lohman AW, Straub AC and Johnstone SR. Identification of connexin43 phosphorylation

830

and S-nitrosylation in cultured primary vascular cells. Gap Junction Protocols. 2016:97-111.

831

72. Thorne RF, Ralston KJ, de Bock CE, Mhaidat NM, Zhang XD, Boyd AW and Burns GF.

832

Palmitoylation of CD36/FAT regulates the rate of its post-transcriptional processing in the

833

endoplasmic reticulum. Biochimica et Biophysica Acta (BBA)-Molecular Cell Research.

834

2010;1803:1298-1307.

835

73. Jiang Y, Krantz S, Qin X, Li S, Gunasekara H, Kim Y-M, Zimnicka A, Bae M, Ma K and

836

Toth PT. Caveolin-1 controls mitochondrial damage and ROS production by regulating fission-

837

fusion dynamics and mitophagy. Redox Biology. 2022;52:102304.

838

74. Schiattarella GG, Altamirano F, Tong D, French KM, Villalobos E, Kim SY, Luo X, Jiang

839

N, May HI and Wang ZV. Nitrosative stress drives heart failure with preserved ejection fraction.

840

Nature. 2019;568:351-356.

841

75. Opie LH. Metabolic syndrome. Circulation. 2007;115:e32-e35.

842

76. Denisenko YK, Kytikova OY, Novgorodtseva TP, Antonyuk MV, Gvozdenko TA and

843

Kantur TA. Lipid-induced mechanisms of metabolic syndrome. Journal of Obesity. 2020;2020.

844

77. Hink U, Li H, Mollnau H, Oelze M, Matheis E, Hartmann M, Skatchkov M, Thaiss F,

845

Stahl RA and Warnholtz A. Mechanisms underlying endothelial dysfunction in diabetes mellitus.

846

Circulation research. 2001;88:e14-e22.

847

78. Romero N, Radi R, Linares E, Augusto O, Detweiler CD, Mason RP and Denicola A.

848

Reaction of human hemoglobin with peroxynitrite: isomerization to nitrate and secondary

849

formation of protein radicals. Journal of Biological Chemistry. 2003;278:44049-44057.

850

79. Ottolini M, Hong K, Cope EL, Daneva Z, DeLalio LJ, Sokolowski JD, Marziano C,

851

Nguyen NY, Altschmied J and Haendeler J. Local peroxynitrite impairs endothelial transient

852

receptor potential vanilloid 4 channels and elevates blood pressure in obesity. Circulation.

853

2020;141:1318-1333.

854

80. Daneva Z, Marziano C, Ottolini M, Chen Y-L, Baker TM, Kuppusamy M, Zhang A, Ta

855

HQ, Reagan CE and Mihalek AD. Caveolar peroxynitrite formation impairs endothelial TRPV4

856

channels and elevates pulmonary arterial pressure in pulmonary hypertension. Proceedings of the

857

National Academy of Sciences. 2021;118:e2023130118.

858

81. Potje SR, Grando MD, Chignalia AZ, Antoniali C and Bendhack LM. Reduced caveolae

859

density in arteries of SHR contributes to endothelial dysfunction and ROS production. Scientific

860

reports. 2019;9:6696.

861

82. Kang DH, Kim Y, Min S, Lee SY, Chung KY, Baek I-J, Kwon K, Jo H and Kang SW.

862

Blood flow patterns switch VEGFR2 activity through differential S-nitrosylation and S-oxidation.

863

Cell Reports. 2023;42.

864

83. Toth PP, Patti AM, Giglio RV, Nikolic D, Castellino G, Rizzo M and Banach M.

865

Management of statin intolerance in 2018: still more questions than answers. American Journal of

866

Cardiovascular Drugs. 2018;18:157-173.

867

84. Cao G, Yang G, Timme TL, Saika T, Truong LD, Satoh T, Goltsov A, Park SH, Men T

868

and Kusaka N. Disruption of the caveolin-1 gene impairs renal calcium reabsorption and leads to

869

hypercalciuria and urolithiasis. The American journal of pathology. 2003;162:1241-1248.

870

85. DeLalio LJ, Keller AS, Chen J, Boyce AK, Artamonov MV, Askew-Page HR, Keller IV

871

TS, Johnstone SR, Weaver RB and Good ME. Interaction between pannexin 1 and caveolin-1 in

872

smooth muscle can regulate blood pressure. Arteriosclerosis, thrombosis, and vascular biology.

873

2018;38:2065-2078.

874

86. Kenwood BM, Weaver JL, Bajwa A, Poon IK, Byrne FL, Murrow BA, Calderone JA,

875

Huang L, Divakaruni AS and Tomsig JL. Identification of a novel mitochondrial uncoupler that

876

does not depolarize the plasma membrane. Molecular metabolism. 2014;3:114-123.

877

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

87. Billaud M, Lohman AW, Straub AC, Parpaite T, Johnstone SR and Isakson BE.

878

Characterization of the thoracodorsal artery: morphology and reactivity. Microcirculation.

879

2012;19:360-372.

880

88. Wan J, Roth AF, Bailey AO and Davis NG. Palmitoylated proteins: purification and

881

identification. Nature protocols. 2007;2:1573-1584.

882

883

884

885

886

887

888

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

Acknowledgments

889

We thank Anne Kenworthy (University of Virginia) for contributing her expertise on caveolin-1

890

and its role in caveolae and lipid nanodomain organization. We thank the University of Virginia

891

School of Medicine Research Histology Core and Blood lab.

892

893

Funding:

894

Work was supported by NIH grants HIL120840 (N.L.), HL007284 (C.M.P), HL120840 (B.E.I),

895

HL137112 (B.E.I), University of Virginia Launchpad (B.E.I), Lipedema Foundation (B.E.I)

896

HL007284 (M.A.L and L.S.D), AHA915176 (M.A.L), R35GM131829 (L.C.)

897

898

Author contributions

899

Conceptualization: BEI, MAL, MK, NL

900

Methodology: PB, AK, TS, IL, RM

901

Investigation: MAL, WJS, LSD, SN, SKL, AC, RT, CP, TSK, XS, CAR, LC, KRL, MK,

902

903

Visualization: LC

904

Supervision: BEI

905

Writing—original draft: MAL, BEI

906

Writing—review & editing:

907

908

Competing interests: Authors declare that they have no competing interests

909

910

Data and materials availability: All data are available in the main text or the supplementary

911

materials

912

913

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

Figures

914

FIGURE 1:

915

916

917

918

Figure 1: Adipose endothelial cell caveolin 1 decreases with metabolic disease. (A) Volcano

919

plot showing differential gene expression in adipose ECs from HFD-fed mice relative to NC-fed

920

mice fed over 12 weeks; genes are displayed as up- (red) or down-regulated (blue). (B) UMAP of

921

murine EC Cav1 expression across vascular clusters. (C) Dot plot showing Cav1 expression among

922

ECs of different vascular origins and diet-fed mice. (D) Volcano plot showing differential gene

923

expression in human adipose ECs from obese individuals relative to lean individuals; genes are

924

displayed as up- (red) or down-regulated (blue). (E) UMAP of human EC CAV1 expression across

925

vascular clusters. (F) Dot plot showing human CAV1 expression among ECs of different vascular

926

origins from both lean and obese humans.

927

928

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

FIGURE 2:

929

930

Figure 2: Genetic deletion of caveolin 1 in endothelial cells increases serum lipids and

931

decreases intracellular lipids. (A) Schematic outlining the timeline and diet schedule for Cav1

fl/fl

932

mice. (B) gDNA excision gel of the Cav1 second exon from lung. (C) qPCR of Cav1 RNA isolated

933

from lung. (D) en face immunostaining of adipose arteries. N= 3 (each averaged from 3 fields of

934

view), Scale bar = 20µm, (right) quantification in arbitrary fluorescence units (AFU). (E)

935

triglycerides, (F) cholesterol, (G) LDL, (H) HDL, (I) fasting non-esterified fatty acids (NEFA)

936

from serum of EC Cre

Cav1

fl/fl

and EC Cre

Cav1

fl/fl

fed a normal chow (NC), N= 8-10 mice per

937

group. (J) Lipid staining (Nile red) in an en face adipose artery quantified (below) using lipid

938

droplet area per nuclei. N=5 mice (average from 3 fields of view each), Scale bar = 10 µm. (K)

939

Diagram showing loss of Cav1 inhibits lipid uptake. (L) Glucose tolerance test (GTT) area under

940

the curve (AUC) quantification. N = 8-10 mice. P-values listed are from unpaired t-tests.

941

942

943

944

945

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

FIGURE 3:

946

947

948

Figure 3: Potential role for nitric oxide, but not endocytosis, in the absence of endothelial cell

949

caveolin 1. (A) Knockdown of Cav1 RNA shown via qPCR. (B) Lipid (Bodipy) staining in human

950

adipose microvascular endothelial cells (HAMECs), quantified in (C), N= 4 (with each as an

951

average of 3 fields of view). (D) Cholesterol concentration after cyclodextrin treatment (+cylco,

952

10mM) or control treatment (-cyclo). (E) Lipid staining (Bodipy) of HAMECs after cyclodextrin

953

and lipid treatments, quantified in (F). (G) Lipid (Bodipy) staining and quantification (H) of cells

954

treated with vehicle or endocytosis inhibitors (dynosore (80µM) and genistein (200µM)) before

955

lipid treatment. (I). Nitrite (left) and nitrate (right) levels in EC Cre

Cav1

fl/fl

and EC Cre

Cav1

fl/fl

956

mice, N= 5 mice. (J) Nitrite and nitrate levels in NC- and HFD-fed mice at 12 weeks of diet. N = 3

957

mice per group. 50µM lipids were used for treatments in all cells for 15 minutes. Scale bars = 25µm

958

throughout. P-values listed are from unpaired t-tests. 50µM lipids (12:5µM linoleic acid, 25µM

959

oleic acid, 12.5µM palmitic acid)

960

961

962

963

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

FIGURE 4:

964

965

966

Figure 4: L-NAME rescues loss of lipid uptake seen with deletion of endothelial caveolin 1.

967

(A) HAMECs treated with control (siCntrl) or Cav1 (siCav1) siRNA followed by L-NAME and

968

stained for lipid accumulation (Bodipy) after treatment with 50µM lipids (50µM lipids (12:5µM

969

linoleic acid, 25µM oleic acid, 12.5µM palmitic acid). N=4 (each from an average of 3 fields of

970

view). (B) Lipid staining (Nile red) in en face adipose arteries of EC Cre

Cav1

fl/fl

and EC Cre

971

Cav1

fl/fl

mice given L-NAME (60mg/dL) in their drinking water. N=7 (each from an average of 3

972

fields of view). (C) Cholesterol and triglyceride concentrations from blood taken before and after

973

EC Cre

Cav1

fl/fl

mice were administered L-NAME in their drinking water. N=5-6. P-values listed

974

are from paired t-tests. (D) Cholesterol and triglyceride concentrations from mice fed a high fat diet

975

(HFD) and administered either regular drinking water or L-NAME water. (E) Cholesterol and

976

triglyceride concentrations from RBC Cre

eNOS

fl/fl

and RBC Cre

eNOS

fl/fl

mice. N=6-7. (F)

977

Cholesterol and triglyceride concentrations from EC Cre

eNOS

fl/fl

and EC Cre

eNOS

fl/fl

mice. N=6.

978

P-values listen for (D-F) are from unpaired t-tests.

979

980

981

982

983

984

985

986

987

988

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

FIGURE 5:

989

990

991

992

Figure 5: CD36 and nitric oxide may both be necessary for lipid uptake in adipose endothelial

993

cells. (A) Caveolin lipid raft fractionation in HAMECs with the cell surface marked with biotin, the

994

Cav1-rich lipid raft marked as fraction 3, and calnexin marking ER fractions. (B) Immunostaining

995

of an en face adipose artery showing Cav1 (cyan) and CD36 (magenta) in various planes of view

996

(XY, XZ, YZ). Arrows highlight locations in which Cav1 and CD36 are next to one another. Yellow

997

bars show the location in which the XZ and YZ planes were taken from. Scale bar = 10µm (C)

998

Control (Cntrl) vs. siCD36-treated HAMECs treated with lipids (50µM lipids; (12:5µM linoleic

999

acid, 25µM oleic acid, 12.5µM palmitic acid)) (stained with Bodipy) and quantified. Scale bar =

1000

25µm. (D) HEK cells transfected with CD36 plasmid or empty vector (control) and subsequently

1001

treated with lipids and one of two NO donors: DETA/NO or SNOG. Quantification of lipid uptake

1002

(below). Scale bar = 25µm, N=5 distinct experiments. (E) Biotin switch assay, with subsequent

1003

streptavidin bead pull down. All samples treated with DETA/NO. EV = empty vector, CD36 =

1004

CD36 plasmid, -B or +B = without or with biotin added, respectively. (F) Adipose artery en face,

1005

lipid accumulation quantified N = 4, three images were averaged for each value, SSO is Sulfo-N-

1006

Succinimidyl oleate a CD36 inhibitor. Scale bar = 20µm. All P-Values denoted are unpaired t-

1007

tests.

1008

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

FIGURE 6:

1009

1010

1011

Figure 6: Cysteines 3 and 313 on CD36 are nitrosated inhibiting lipid uptake. (A) Table

1012

showing three cysteine positions on CD36 protein predicted by GPS-SNO to be potential points of

1013

nitrosation. (B) GPS-SNO-predicted cysteines (3, 313, 466) on CD36 protein with labeled domains.

1014

Figure created using GPS-SNO software. Yellow region denotes fatty acid binding pocket of CD36;

1015

the disulfide bridges supporting the fatty acid pocket are marked from residues 243-334. Grey

1016

denotes the intracellular domains, dark blue denotes the transmembrane regions, and teal denotes

1017

the ectodomain of CD36. (C) HEK cells transfected with mutated versions (C→ A) of the CD36

1018

plasmid. All three cysteine sites (3, 313, 466) on the top, then from top to bottom sites 466, 313, 3,

1019

both 313 and 3, and lastly both 466 and 3 on the bottom. Each condition was treated with 50µM

1020

DETA/NO and 50µM lipids. Quantification for each condition to the right, N = 4-5 with 3 images

1021

averaged for each N value. (D) Biotin switch assay, with subsequent streptavidin bead pull down.

1022

All samples treated with DETA/NO. EV = empty vector, WT = CD36 plasmid, C466/3A and

1023

C466/313/3A are mutant plasmids with mutations at specified sites, -B or +B = without or with

1024

biotin added respectively. CD36 plasmid is mCherry tagged.

1025

1026

1027

1028

1029

1030

1031

1032

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

FIGURE 7:

1033

1034

Figure 7: Nitrosation of CD36 inhibits its plasma membrane localization. (A) Structure of

1035

CD36 protein with transmembrane and ectodomains marked. Cysteines present in CD36 are marked

1036

in red; (left) cysteines in the ectodomain participate in disulfide bonds, (middle) CD36 with four

1037

cytoplasmic cysteines palmitoylated, and (right) CD36 with proposed cysteines nitrosated. (B)

1038

Caveolin lipid raft fractionation from HAMECs treated with Cav1 siRNA. Cav1-rich lipid rafts

1039

appear in fraction three when Cav1 is present (Figure 5A). Black box marks a shift in fractions

1040

where CD36 is enriched. The cell surface is marked with biotin and Calnexin marks ER fractions.

1041

1042

were treated with control siRNA (Cntrl) or siCav1 and then half were given L-NAME. (D) IP-ABE

1043

assay on HEK293T cells transfected with CD36-mcherry. Detection of palmitoylated proteins via

1044

western blot. (E) HAMECs treated with no plasmid control, WT CD36, or C466/3A mutant

1045

plasmid. DAPI denotes nuclei, and magenta marks CD36. Scale bar = 10µm. (F) En face adipose

1046

arteries from EC Cre

Cav1

fl/fl

(top), EC Cre

Cav1

fl/fl

(middle), and EC Cre

Cav1

fl/fl

+ L-NAME

1047

(bottom) mice stained for CD36 (magenta) and ER with calnexin (green). Quantification of CD36-

1048

Calnexin overlap using Pearson’s correlation coefficient on the right, N= 4 mice, statistics represent

1049

one-way ANOVA with multiple comparisons.

1050

1051

1052

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

1053

1054

FIGURE 8

1055

1056

Figure 8: NOS inhibition drives endothelial dysfunction. (A) Heatmap showing enrichment of

1057

mitochondrial genes (left) in control treated cells and a decrease in expression with administration

1058

of lipids + L-NAME; statistics in supplemental figure 5A. Right, heatmap showing enrichment of

1059

lipid storage genes in L-NAME treated cells and a decrease in expression with non-lipid treated

1060

cells (control). Statistics and GSEA plot in supplemental figure 5B-E. (B) Maximum oxygen

1061

consumption normalized to total cellular protein from mitochondrial stress test performed via

1062

seahorse assay, N = 3-4, P values from one-way ANOVA with a multiple comparisons test. (C)

1063

Vasodilation of 3

order mesenteric arteries when exposed to L-NAME and/or lipids. A pre-

1064

treatment of 10µMoleic acid was used. Oleic acid was also present during treatments. N= 3-5

1065

1066

1067

1068

1069

1070

1071

1072

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

Supplementary Materials for

Dynamic nitrosation – palmitoylation of endothelial CD36 regulates serum

lipids

Melissa Luse et al.

*Corresponding author. Email: brant@virginia.edu

This PDF file includes:

Supplementary Text

Figs. S1 to S6

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

Fig. S1.

Supplementary Figure 1: Filtering and subsetting scRNAseq clusters for EC enrichment.

UMAP of all cells (lean and obese patients) in Emont data set showing expression pattern of CDH5

(A) and PECAM1 (B). (C) Filtered human adipose ECs clustered by vascular origin. UMAP of all

adipose ECs (from HFD and NC mice) before filtration with expression patterns of Cdh5 (D) and

Pecam1 (E). (F) Filtered murine adipose ECs clustered by vascular origin.

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

Fig. S2.

Supplementary Figure 2: Caveolin 1 deletion from endothelial cells does not alter body

mass. (A) Western blot from lung, showing loss of Cav1 protein when EC Cre is present,

quantification on left is normalized to total protein loading control. Unpaired t-test was used, N =

4-5. (B) Adipose artery en face with secondary only staining. (C) Adipose tissue cross section

stained with Cav1, Scale bar = 50µm. (D) Weight of mice (in grams; g) over the course of the

12-week NC or HFD diets of Cav1

fl/fl

mice (E) Percent weight gain calculated using ((final mass-

initial mass/Initial mass)*100 of high fat diet (HFD) and normal chow (NC) mice. (F)

Epididymal fat mass and food intake (G) of EC Cre

(N.=4) and EC Cre+ (N= 9) Cav1

fl/fl

mice.

(H) Glucose tolerance test, blood glucose over time after glucose injection. N= 8-10 mice for all

experiments for E,F, and H.

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

Fig. S3.

Supplementary Figure 3: Protein composition of HAMEC and HEK cells. (A) Western blot

showing loss of CD36 band with siCD36 in HAMECs, below is total protein (B) HEK cells with

no transfection (top) and HEK cells transfected with CD36 plasmid (bottom). CD36 plasmid is

tagged with mCherry (C) Cav1 RNA expression with CD36 knockdown in HAMECs measure via

q-PCR. (D) HEK cells transfected with CD36 plasmid with mcherry reported. (E) HEK cell

transfections with CD36 plasmid with subsequent CD36 immunoblot. IP = immunoprecipitation.

(F) Cav1 and eNOS immunoblots in HAMEC vs. HEK cells. (G) Biotin switch assay to detect

nitrosation of CD36 using C313/3A and C466/313/3A mutant versions of the CD36 plasmid. -B

an +B represent without or with biotin respectively. All samples were treated with DETA/NO. (H)

Cumulative HEK293T lipid uptake data.

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

Fig. S4.

Supplementary Figure 4: CD36 protein and RNA levels are unaltered with loss of Caveolin

1. (A) CD36 expression in control and siCav1 treated cells measured via q-PCR. (B) Adipose

artery en face stained with Alexa Four 568 only. (C) CD36 immunoblot from EC Cre

and EC Cre

Cav1

fl/fl

mice and quantified in (D) by normalizing to total protein. N = 4 mice.

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

Fig. S5.

Supplementary Figure 5: Oxygen consumption and mitochondrial gene expression is

decreased with lipids. (A) P values of mitochondrial genes, using DESeq2 and the Benjamini-

Hochberg correction as described in the methods. (B and C)Gene set enrichment analysis (GSEA)

for regulation of lipid storage genes. GSEA plot for Lipid + L-NAME treated HAMECs. P values

for lipid storage genes. (D) HAMEC FABP4 expression were calculated using DESeq2 and the

Benjamini-Hochberg correction as described in the methods. N= 3 for all bulk RNA sequencing

data. (E) Oxygen consumption trace for mitochondrial stress test measured via seahorse assay.

siCntrl (black) or siCav1 (red) HAMECs were used. Maximum oxygen consumption (right) and

Oxygen consumption over time normalized to total cellular protein. (F) Mitochondrial stress test

for HAMECs treated with L-NAME (grey), Lipids (orange), Lipids + L-NAME (red), Lipids +

DETA/NO (blue) or control (black) media. Maximum oxygen consumption for control and L-

NAME treated cells (right).

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint

Fig. S6.

Supplementary Figure 6: Summary figure showing proposed mechanism for nitrosation-

palmitoylation of CD36 which regulates endothelial lipid uptake.

The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.09.588733doi: bioRxiv preprint