N-acetylaspartate from fat cells regulates postprandial body temperature

Abstract

N-acetylaspartate (NAA), the brain's second most abundant metabolite, provides essential substrates for myelination through its hydrolysis¹. However, the physiological roles of NAA in other tissues remain unknown. Here, we show that aspartoacylase (ASPA) expression in white adipose tissue (WAT) governs blood NAA levels for postprandial body temperature regulation. Genetic ablation of Aspa in mice resulted in systemically elevated NAA levels, and the ensuing accumulation in WAT stimulated pyrimidine production. Stable isotope tracing confirmed higher incorporation of glucose-derived carbon into pyrimidine metabolites in Aspa knockout cells. Additionally, serum NAA levels positively correlated with the abundance of the pyrimidine intermediate orotidine 5'-monophosphate, and this relationship predicted lower body mass index in humans. Using whole-body and tissue-specific knockout mouse models, we observed that fat cells provided plasma NAA and suppressed postprandial body temperature elevation. Moreover, unopposed NAA from adipocytes greatly enhanced whole-body glucose disposal exclusively in WAT. Exogenous NAA also increased plasma pyrimidines and lowered body temperature. These data place WAT-derived NAA as an endocrine regulator of postprandial body temperature and define broader roles for metabolic homeostasis.

Fig. 1. Constitutive Aspa deletion significantly affects body weight and substrate use in mice fed a normal chow diet.

Aspa^WT and Aspa^KO mice were fed a normal chow diet for 12 weeks and then subjected to metabolic phenotyping. a, Visual representation of Aspa^WT and Aspa^KO mice at 8 weeks of age. b, Body weight measurements of Aspa^WT and Aspa^KO mice over 12 weeks starting at 6 weeks of age (n = 6 WT, 10 KO mice per group). Data represent mean ± s.e.m. **P < 0.01 by an ordinary two-way ANOVA. c, Body composition by Echo magnetic resonance imaging, shown as a percentage of body weight (n = 11 WT, 9 KO mice per group). Data represent mean ± s.e.m., and statistical analysis was performed using an unpaired two-tailed Student’s t-test. d,e, Mice were individually housed and monitored in CLAMS cages during a 96-h period with measurements of food intake (d) and RER (e). Data represent mean ± s.e.m. Statistical analysis of d and e was performed within CalR and using ANCOVA with lean body mass as a covariate (n = 7 WT, 6 KO mice per group). f, ΔRER, defined as the difference between the average of the lowest 10% (light phase) and the highest 10% (dark phase) of RER values (n = 7 WT, 6 KO mice per group). *P < 0.05 by an unpaired two-tailed Student’s t-test. g, Tissue weights from WAT (scWAT and vWAT) depots, shown as a percentage of body weight (% BW) (n = 11 WT, 9 KO mice per group). h, Plasma leptin levels from ad libitum-fed mice (n = 8 WT, 9 KO mice per group). i,j, Representative histological analysis by H&E staining and mean adipocyte size (μm²) of scWAT (i) and vWAT (j) across three to five fields of view (n = 4 mice per group); scale bars, 100 μm. k–m, Relative abundance of NAA, glycolytic metabolites and pyrimidine metabolites in the scWAT (k), vWAT (l) and plasma (m) of Aspa^WT and Aspa^KO mice, measured by IC–MS targeted analysis. G6P, glucose-6-phosphate; F6P, fructose-6-phosphate. In c, d and f–j, data are represented as box-and-whisker plots using the Min-to-Max method in GraphPad Prism: box limits, 25th to 75th percentiles; centre line, median; whiskers, minimum and maximum values. In k–m, data are shown as the z score of log₂-transformed values and relative changes denoted in the colour bar (n = 4 mice per group). In f–m, *P < 0.05, ****P < 0.0001 by an unpaired two-tailed Student’s t-test. Source data

Fig. 2. ARIC reveals that NAA and the pyrimidine metabolite OMP predict body composition.

a, Schematic for the analysis of the ARIC study dataset using an age-, sex-, race- and BMI-adjusted linear regression model. BMI (kg m⁻²) groups: normal, 18.5 < BMI ≤ 25; overweight, 25 < BMI ≤ 30; obesity, BMI > 30. b, Relative (Rel) NAA abundance among BMI groups and male and female participants (male + female: n = 1,308 normal, n = 2,088 overweight (Over), n = 1,790 obesity; male: n = 500 normal, n = 1,012 overweight, n = 740 obesity; female: n = 808 normal, n = 1,076 overweight, n = 1,050 obesity). Shown are mean ± 95% confidence interval. P < 0.05 (^anormal versus obesity, ^bnormal versus overweight) by a two-way ANOVA with a Sidak multiple comparisons test. c, NAT8L and ASPA expression in the scWAT of persons with a metabolically normal lean (MNL) status and those with metabolically unhealthy obesity (MUO) (n = 5 per group). Data are represented as box-and-whisker plots using the Min-to-Max method in GraphPad Prism: box limits, 25th to 75th percentiles; centre line, median; whiskers, minimum and maximum values. *P < 0.05, **P < 0.01 by an unpaired two-tailed Student’s t-test. d, Estimated effect of OMP on NAA for each BMI category. The β coefficient is shown with the respective 95% confidence intervals and statistical significance from a linear regression model (male + female: n = 1,308 normal, n = 2,088 overweight, n = 1,790 obesity; male: n = 500 normal, n = 1,012 overweight, n = 740 obesity; female: n = 808 normal, n = 1,076 overweight, n = 1,050 obesity). e, Estimated effect of OMP on NAA for each BMI category and sex. The β coefficients for each BMI category are shown. Error bars in e represent s.d. The figure shows statistical assessments of the regression analysis, and P values (circles) summarize the strengths of the associations within each group. f,g, Relative abundance of NAA and pyrimidine metabolites in the medium of scWAT explants incubated for 24 h (f) and in scWAT explants (g) (n = 5 replicates per group). Relative abundance was normalized to an internal standard, followed by log₁₀ transformation. Data are represented as box-and-whisker plots using the Min-to-Max method in GraphPad Prism: box limits, 25th to 75th percentiles; centre line, median; whiskers, minimum and maximum values. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 by an unpaired two-tailed Student’s t-test. Source data

Fig. 3. NAA from adipose tissue regulates postprandial body temperature.

a,b, Body temperature (temp) of male (a) and female (b) Aspa^fl/fl and Aspa^fKO mice following fasting (n = 23 Aspa^fl/fl, 29 Aspa^fKO males; n = 8 Aspa^fl/fl, 9 Aspa^fKO females) and refeeding (n = 23 Aspa^fl/fl, 29 Aspa^fKO males; n = 8 Aspa^fl/fl, 9 Aspa^fKO females). For the change in body temperature (ΔTemp), in paired males (n = 23 Aspa^fl/fl, 29 Aspa^fKO) and females (n = 8 Aspa^fl/fl, 9 Aspa^fKO), data are mean ± s.e.m., and *P < 0.05 by an unpaired two-tailed Student’s t-test. c, Plasma leptin levels in male mice following fasting (n = 22 Aspa^fl/fl, 23 Aspa^fKO) and refeeding (n = 18 Aspa^fl/fl, 22 Aspa^fKO). d,e, Relative abundance of NAA and pyrimidine metabolites in the scWAT (d) and plasma (e) of Aspa^fl/fl and Aspa^fKO mice in fasted and refed states measured by IC–MS analysis (n = 4 mice per group per condition) and shown as log₁₀-transformed values normalized to Aspa^fl/fl fasted levels. f, Hyperinsulinaemic–euglycaemic clamp (n = 5 per group). Glucose infusion rate (GIR), glucose production rate (GPR) during basal (empty bars) and clamp (striped bars) conditions, glucose disposal rate (GDR), and glucose uptake in vWAT and gastrocnemius muscle. Data are mean ± s.e.m. In a–e, data are represented as box-and-whisker plots using the Min-to-Max method in GraphPad Prism: box limits, 25th to 75th percentiles; centre line, median; whiskers, minimum and maximum values. For a–e, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 by an ordinary two-way ANOVA followed by Fisher least significant difference (LSD) tests. For f, *P < 0.05 by an unpaired Student’s t-test (glucose infusion rate, glucose disposal rate and glucose uptake). Glucose production was not different after an ordinary two-way ANOVA. Source data

Fig. 4. Acute NAA supplementation enhances pyrimidine synthesis and reduces body temperature.

a, Schematic of the gavage experiment. Body temperature measurements and plasma collection in WT mice were performed following a 16-h fast, 15–30 min following gavage and after 1 h of refeeding. b, Body temperature measurements in fasted (n = 13 water, 13 NAA), post-gavage (n = 13 water, 13 NAA) and refed mice (n = 8 water, 8 NAA). *P < 0.05 by an ordinary two-way ANOVA followed by Fisher LSD tests. c, Plasma NAA levels of fasted (n = 4 water, 4 NAA), post-gavage (n = 9 water, 9 NAA) and refed mice (n = 4 water, 4 NAA). d, Tissue NAA levels in the liver, scWAT and vWAT after gavage (n = 4 mice per group per condition). e, Relative abundance of CarbAsp across the liver, scWAT and vWAT following gavage (n = 4 mice per group per condition). f, Relative abundance of plasma CarbAsp (n = 9 mice per group per condition), OMP (n = 9 mice per group per condition) and UMP (n = 4 mice per group per condition) following gavage. g, Immunoblot of scWAT from Aspa^WT mice following water or NAA gavage, probed for ASPA, phospho-CAD (pCAD) S1859, total CAD (tCAD), phospho-AKT S473, total AKT (tAKT), phospho-TSC2 T1462, total TSC2 (tTSC2), phospho-4EBP1 T37/46 and total 4EBP1 (t4EBP1). HSP90 served as the loading control (n = 3 mice per group per condition). h, Model for NAA’s mechanism of action in regulating de novo pyrimidine synthesis. NAA binds within the ATCase domain of CAD, promoting increased CAD activity. Carbamoyl-P, carbamoyl phosphate; N-carbamoyl-asp, N-carbamoyl aspartate; DHO, dihydroorotase; DHODH, DHOA dehydrogenase; PPi, pyrophosphate; PRPP, phosphoribosyl pyrophosphate; UMPs, UMP synthase; UDP, uridine 5′-diphosphate; UTP, uridine 5′-triphosphate; CTP, cytidine 5′-triphosphate; CDP, cytidine 5′-diphosphate; CMP, cytidine 5′-monophosphate; UCK, uridine cytidine kinase. In b–f, data are represented as box-and-whisker plots using the Min-to-Max method in GraphPad Prism: box limits, 25th to 75th percentiles; centre line, median; whiskers, minimum and maximum values. In c–f, data are shown as log₁₀-transformed values. *P < 0.05, **P < 0.01, ****P < 0.0001 by an unpaired two-tailed Student’s t-test. NS, no significance. Panels a and h created with BioRender.com. Source data

Extended Data Fig. 1. Aspa is highly expressed in differentiated adipocytes and across adipose tissue depots.

a, Aspa gene expression across mouse tissues (n = 3 mice/tissue). Data represent mean ± s.e.m. *P < 0.05, **P < 0.01, ****P < 0.0001 by ordinary one-way ANOVA followed by Dunnett’s multiple comparisons test of brain vs different tissues. Expression of Aspa and adipogenic genes (Adipoq, Fabp4, and Pparg2) in pre-adipocytes (Pread) and mature adipocytes (Adipo) in mice: (b) SVF-derived adipocytes, (c) 3T3-L1 adipocytes (n = 3 biological replicates/group). Data represent mean ± s.e.m. d, Expression of ASPA and adipogenic genes (ADIPOQ, FABP4, and PPARG2) in human primary pre-adipocytes (Pread) and mature adipocytes (Adipo) (n = 3 biological replicates/group). Data represent mean ± s.e.m. e, Relative abundance of NAA and TCA metabolites in 3T3-L1 pre-adipocytes (Pread) and mature adipocytes (Adipo) (n = 5 biological replicates/group). Data represented as box-and-whisker plots using the Min-to-Max method in GraphPad Prism: box limits, 25^th to 75^th percentiles; center line, median; whiskers, minimum and maximum values. f, Immunoblot of 3T3-L1 cells and human primary cells probed for ASPA, ADIPOQ, and PPARγ before (-d) and after adipocyte differentiation (+d). HSP90 served as the loading control. g, Immunofluorescent staining of ASPA (red) and co-labeling of lipid droplets (green, LipidTox), and nuclei (blue, DAPI) of primary human adipocytes before and after adipocyte differentiation. h, Correlation between expression of ADIPOQ and ASPA in WAT of people before and after weight loss (WL) (n = 15 samples/group). P-value for Pearson’s r (r) <0.001 by one-sided t-test performed on the linear regression. b-e, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 by unpaired two-tailed Student’s t-test. f,g representative of 3 independent experiments. Source data

Extended Data Fig. 2. Aspa^KO mice display decreased glucose tolerance and plasma leptin levels while maintaining insulin sensitivity.

Aspa^WT and Aspa^KO mice were fed NCD for 12 weeks. a, Body weight gain of Aspa^WT and Aspa^KO mice over 12 weeks starting at 6 weeks of age (n = 6 WT,10 KO mice/group), shown as a percentage of initial body weight. Data represent mean ± s.e.m. ***P < 0.001 by ordinary two-way ANOVA. b, Glucose tolerance test (GTT) (n = 11 WT,9 KO mice/group) and (c) insulin tolerance test (ITT) (n = 10 WT,9 KO mice/group) with corresponding area-under-curve (AUC) measurements. Data represent mean ± s.e.m. *P < 0.05, **P < 0.01 by mixed-effects model followed by Sidak’s multiple comparisons test. AUC: **P < 0.01 by unpaired two-tailed Student’s t-test. d, Plasma insulin levels following 4 h fast (n = 10 WT,9 KO mice/group). Statistical analysis by unpaired two-tailed Student’s t-test. e, Representative histological analysis by H&E of ad-libitum fed liver across 3-5 fields of view (n = 4 mice/group); scale bars, 100μm. f, Ad-libitum fed liver triglyceride levels (n = 5,4 mice/group). Statistical analysis by unpaired two-tailed Student’s t-test. g, Mice were individually housed and monitored in CLAMS cages during a 96 h period with measurements of energy expenditure (EE). Data represent mean ± s.e.m. Statistical analysis was performed within CalR by ANCOVA with lean body mass as a covariate (n = 7 WT,6 KO mice/group). h, Representative histological analysis by H&E of ad-libitum fed BAT across 3-5 fields of view (n = 4 mice/group); scale bars, 100μm. i, BAT depot tissue weight, shown as a percentage of body weight (n = 11 WT,9 KO mice/group). b-d,f,i Data represented as box-and-whisker plots using the Min-to-Max method in GraphPad Prism: box limits, 25^th to 75^th percentiles; center line, median; whiskers, minimum and maximum values. Source data

Extended Data Fig. 3. Broad metabolic pathway alterations occur in WAT from Aspa^KO mice.

Aspa^WT and Aspa^KO mice were fed NCD for 12 weeks and tissues were used for reverse-phase protein array (RPPA) analysis of (a) scWAT and (b) vWAT from Aspa^WT and Aspa^KO mice. Relative abundance of (c) AMP and (d) GMP in scWAT and vWAT of Aspa^WT and Aspa^KO mice, measured by IC/MS targeted analysis (n = 4 mice/group). Data are mean ± s.e.m. e, Ratio of NAD/NADH across scWAT (n = 8 WT,6 KO mice/group) and vWAT (n = 8 WT,9 KO mice/group). Data are mean ± s.e.m. a-b Data shown as z-score of log₂ transformed values and relative changes denoted in the colour bar (n = 4 mice/group). a-e ^#P < 0.1, *P < 0.05 by unpaired two-tailed Student’s t-test. Source data

Extended Data Fig. 4. Aspa^−/− adipocytes shunt glucose-derived carbon into pyrimidines.

a, Schematic of metabolite profiling of SVF-derived differentiated adipocytes generated from Aspa^WT and Aspa^KO mice. Glucose incorporation into metabolite pools was measured by incubating cells with U-¹³C glucose. Isotopolog distribution for indicated metabolites was measured after 24 h using IC-MS targeted profiling. b, Relative abundance of intracellular NAA (n = 3 biological replicates/group). Data represent mean ± s.e.m. ****P < 0.0001 by unpaired two-tailed Student’s t-test. c, Schematic showing contribution of ¹³C₆-labeled glucose tracing into early pyrimidines via aspartate (m + 2 and m + 3 isotopologs), and R5P and downstream pyrimidines (m + 5 isotopologs). G6P, glucose-6-phosphate; 6-PG, 6-phosphogluconate; R5P, ribose-5-phosphate; F6P, fructose-6-phosphate; 3PG, 3-phosphoglycerate. d, Fractional labeling of aspartate, early and later pyrimidines, and the pentose phosphate pathway metabolite, R5P. DHOA, dihydroorotate; OMP, orotidine 5’-monophosphate; UMP, uridine monophosphate. Isotopolog data are corrected for ¹³C natural abundance (n = 3 replicates/group). Data are mean ± s.e.m. *P < 0.05, **P < 0.01, ***P < 0.001 by unpaired two-tailed Student’s t-test. e, Immunoblot of Aspa^WT and Aspa^KO primary SVF cells ± differentiation, probed for ASPA, phospho-CAD (S1859), total CAD, phospho-S6K (T389), total S6K, and ADIPOQ. HSP90 served as the loading control. Representative of 3 independent experiments. f, Representative Oil-Red O (ORO) staining of lipid accumulation in differentiated Aspa^WT and Aspa^KO SVF adipocytes. Scale bar is 0.783 cm. Panels a and c created with BioRender.com. Source data

Extended Data Fig. 5. Aspa^KO adipocytes show altered mitochondrial metabolism profiles.

Profiling of mitochondria from differentiated Aspa^WT and Aspa^KO primary SVF cells. a, Immunoblot of Aspa^WT and Aspa^KO primary SVF cells ± differentiation, probed for markers of mitochondrial complexes I-V (CI-CV). HSP90 served as the loading control. Representative of 2 independent experiments. b, Representative images from high-throughput microscopy (HTM) following immunofluorescent labeling of lipid droplets (green, BODIPY), mitochondria (red, MitoTracker), and nuclei (blue, DAPI) of primary SVF-derived adipocytes (n = 8 replicates/group). Imaging was performed with a 60x objective (2K x 2K pixels²). c, High-content analysis of differentiated adipocytes from HTM: (top) nuclei count, mitochondrial area, mitochondrial fluorescence intensity, (bottom) number of big lipids (denoted as lipid count), big lipid area, and mean fluorescence intensity (n = 8 replicates/group). Data represented as box-and-whisker plots using the Min-to-Max method in GraphPad Prism: box limits, 25^th to 75^th percentiles; center line, median; whiskers, minimum and maximum values. *P < 0.05, **P < 0.01, by unpaired two-tailed Student’s t-test. d, Oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) with corresponding area-under-curve (AUC) measurements. (n = 12 replicates/group/experiment, 2 independent experiments). Data for kinetic assessments of OCR and ECAR are mean ± s.e.m. AUC measurements are represented as box-and-whisker plots using the Min-to-Max method in GraphPad Prism: box limits, 25^th to 75^th percentiles; center line, median. Source data

Extended Data Fig. 6. ASPA expression governs postprandial body temperature.

a, Schematic of fast/re-feeding experiment. Body temperature measurements and plasma of Aspa^WT and Aspa^KO male mice were taken following 16 h fast and after 6 h of refeeding. Necropsy data from fasted and refed mice: (b) Body weight, (c) vWAT and (d) scWAT depot weights, shown as % of body weight. Fasted (n = 5 WT,5 KO mice/group), refed (n = 6 WT,9 KO mice/group). e, Body temperature measured by rectal probe. Fasted (n = 10 WT,12 KO mice/group) and refed (n = 10 WT,12 KO mice/group). For ΔT (n = 10 WT,12 KO mice/group), data are mean ± s.e.m.,*P < 0.05 by unpaired two-tailed Student’s t-test. f, Plasma leptin levels. Fasted (n = 9 WT,7 KO mice/group), refed (n = 8 WT,6 KO mice/group). g, Immunoblot of scWAT from Aspa^WT and Aspa^KO mice following fasting and refeeding, probed for phospho-CAD (S1859), total CAD (tCAD), DHODH, phospho-S6K (T389), total S6K (tS6K), and ADIPOQ. HSP90 served as the loading control (n = 3 mice/group/condition). b-f Data represented as box-and-whisker plots using the Min-to-Max method in GraphPad Prism: box limits, 25^th to 75^th percentiles; center line, median; whiskers, minimum and maximum values. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 by ordinary two-way ANOVA followed by Fisher LSD tests. Panel a created with BioRender.com. Source data

Extended Data Fig. 7. Generation of the mouse model for conditional Aspa gene deletion.

a, Aspa^fl/fl mice were generated using CRISPR/Cas9 gene editing. Exon 2 of the Aspa gene was targeted by sgRNAs designed complementary to intronic sequences flanking the exon, followed by insertion of loxP sequences by DNA donor oligonucleotides. b, Validation of floxed allele by genotyping. PCR analyses of floxed alleles at the targeted loci in genomic DNA extracted from ear clips of wild-type (+/+), fl/+, and fl/fl mice. PCR products were run on agarose gels with expected band sizes: wild-type (+) 217 bp and loxP allele (fl) 251 bp. Representative of 5 independent experiments. c, Immunoblot validation of adipocyte-specific Aspa deletion in scWAT vs liver, probed for ASPA, PPARγ, and HSP90 (n = 3 mice/group). d, Body weight measurements of Aspa^fl/fl and Aspa^fKO male mice fed NCD over 12-weeks starting at 6 weeks of age; shown as percent weight gain (n = 7 Aspa^fl/fl,10 Aspa^fKO mice/group). Data represent mean ± s.e.m. No significant difference across genotypes, determined by mixed-effects model. e, vWAT and (f) scWAT depot tissue weights, shown as a percentage of body weight (n = 7 Aspa^fl/fl,6 Aspa^fKO mice/group). Statistical analysis by unpaired two-tailed Student’s t-test. Mice were individually housed and monitored in CLAMS cages during a 96 h period with measurements of: (g) Energy expenditure (EE); (h) Respiratory exchange ratio (RER); (i) Food intake n = 5 Aspa^fl/fl,5 Aspa^fKO mice/group. Statistical analyses of (g, h, i) were performed by ANCOVA with lean body mass as a covariate. j, Representative histological analysis by H&E and mean adipocyte size (μm²) of vWAT and scWAT across 3-5 fields of view (n = 4 mice/group); scale bars, 50 μm. ****P < 0.0001 by unpaired two-tailed Student’s t-test. Necropsy data from fasted and refed Aspa^fl/fl and Aspa^fKO male mice: (k) vWAT and (l) scWAT, shown as % of body weight. Fasted (n = 13 Aspa^fl/fl,11 Aspa^fKO mice/group and refed (n = 10 Aspa^fl/fl,10 Aspa^fKO mice/group. m, Plasma insulin levels in fasted and refed Aspa^fl/fl and Aspa^fKO male mice (n = 6 mice/group). k-m Statistical analyses by ordinary two-way ANOVA followed by Fisher LSD tests. No significant difference across genotypes or after fasting or re-feeding. e,f,i-m Data represented as box-and-whisker plots using the Min-to-Max method in GraphPad Prism: box limits, 25^th to 75^th percentiles; center line, median; whiskers, minimum and maximum values. Source data

Extended Data Fig. 8. Adipose tissue Aspa deletion does not affect liver pyrimidines.

a, Relative abundance of NAA and pyrimidine metabolites in the liver of Aspa^fl/fl and Aspa^fKO mice in fasted and refed states. Metabolites were measured by IC-MS analysis (n = 4 mice/group/condition) and shown as log₁₀ transformed values normalized to Aspa^fl/fl fasted levels. b, Representative histological analysis by H&E of livers of Aspa^fl/fl and Aspa^fKO mice in fasted and refed states across 3-5 fields of view (n = 4 mice/group/conditon); scale bars, 100μm. c, Fasted and refed liver triglyceride levels (n = 4 mice/group/condition). Data represented as box-and-whisker plots using the Min-to-Max method in GraphPad Prism: box limits, 25^th to 75^th percentiles; center line, median; whiskers, minimum and maximum values. *P < 0.05, **P < 0.01, ****P < 0.0001 by ordinary two-way ANOVA followed by Fisher’s LSD tests. Source data

Extended Data Fig. 9. Alphafold modeling docks NAA into the ATCase domain of CAD.

a, AlphaFold prediction of entire CAD protein structure based on theoretical modeling using information from UniProt ID PYR1_HUMAN. b, Aspartate transcarbamylase (ATCase) domain of CAD (green, 1918-2225 region) and NAA (deprotonated form) in predicted binding site. c, Binding energy of NAA and aspartate calculated using high-stringency docking studies. Source data