Insight into molecular and functional properties of NMNAT3 reveals new hints of NAD homeostasis within human mitochondria

Abstract

Among the enzymes involved in NAD homeostasis, nicotinamide mononucleotide adenylyltransferases (NMNAT1-3) are central to intracellular NAD formation. Although NMNAT3 is postulated to be a mitochondrial enzyme contributing to NAD-dependent organelle functioning, information on endogenous proteins is lacking. We report that in human cells a single gene nmnat3 localized on chromosome 3 codes for two mRNA splice variants NMNATv1 and FKSG76, whereas the previously reported NMNAT3v2 transcript is not present. However, NMNAT3v1 and FKSG76 proteins are not detectable, consistent with the finding that an upstream ORF in their mRNAs negatively regulates translation. NMNAT3v1 transfection demonstrates that the protein is cytosolic and inactive, whereas FKSG76 is mitochondrial but operates NAD cleavage rather than synthesis. In keeping with the lack of NMNAT3, we show that extracellular NAD, but not its metabolic precursors, sustains mitochondrial NAD pool in an ATP-independent manner. Data of the present study modify the scenario of the origin of mitochondrial NAD by showing that, in human cells, NMNAT3 is absent in mitochondria, and, akin to plants and yeast, cytosolic NAD maintains the mitochondrial NAD pool.

Figure 1. Expression of NMNAT3 mRNA variants in human cells.

(A) Schematic representation of the ORF of FKSG76, NMNAT3v1 and NMNAT3v2. Colors represent different homology domains. MTS, mitochondrial targeting sequence. Annealing position of primers 1 and 2 (Table 1) used to amplify the ORF of FKSG76 and NMNAT3v2 is shown. (B) Semiquantitative PCR showing the presence of the band of 759 bp related to FKSG76 ORF and the absence of that of 491 bp related to amplification of NMNAT3v2 ORF in HEK293 cells and human brain, skeletal muscle and kidney tissues. (C) Schematic representation of a portion of NMNAT3v1 and FKSG76 transcripts containing their 5′UTR. Colors represent homology domains. Annealing position of primers 3, 4 and 5 (Table 1) used to amplify the fragments of FKSG76 and NMNAT3v2 is shown. (D) Semiquantitative PCR showing the presence of the expected bands of 152 and 147 bp related to amplification of the regions of NMNAT3v1 and FKSG76 shown in (C). (E) Schematic reconstruction of the pre-mRNA structure from which FKSG76 and NMNAT3v1 transcripts originate by alternative splicing. Colors represent homology domains. (F) Comparative analysis of FKSG76 and NMNAT3v1 transcript levels in different human tissues. Agarose gels are representative of at least 4 independent experiments. In (F) columns represent the mean ± SEM of 3 RT-PCR experiments with different human samples.

Figure 2. Effect of NMNAT1 or −3 silencing on cellular NMNAT activity.

(A) Semiquantitative PCR showing silencing of NMNAT1 or NMNAT3 by their respective siRNAs. Note that for NMNAT3 siRNAs able to anneal a sequence present in both NMNAT3v1 and FKSG76 have been used. For NMNAT3 PCR, a sequence present in both NMNAT3v1 and FKSG76 has been amplified. (B) Real-time PCR analysis of the effect of NMNAT3v1 or FKSG76 silencing. (C) NMNAT activity in whole HEK cell extracts and different combinations of the substrates. (D) Effect of NMNAT1 or NMNAT3 silencing on cellular NMNAT activity. (E) Effect of NMNAT1 or NMNAT3 silencing on nuclear or mitochondrial NMNAT activity. Nuclear and mitochondrial NMNAT activities were 2430±390 and 37±7 Fmol/mg prot/30′, respectively. (F) Western blotting evaluation of PARP-1 and VDAC in the nuclear and mitochondrial fractions of control and silenced HEK293 cells. (G) NMNAT3 transcript levels and mitochondrial NMNAT activity in HeLa and HEK cells. In (A) and (G) an experiment representative of 5 is shown. Columns/points represent the mean ± SEM of at least 4 experiments. ** p<0.01; *** p<0.001 vs control (Student's t test).

Figure 3. Expression, intracellular localization and structure of transfected FKSG76 and NMNAT3v1.

(A) Fold increase of mRNA for NMNAT3v1 or FKSG76 upon transfection of the respective expression plasmids. (B) Western blotting analysis of NMNAT3v1 or FKSG76 expression in different cell types as well as NMNAT3v1- and FKSG76-transfected HEK cells. Tubulin is shown as loading control. (C) Immunocytochemical visualization of intracellular distribution of transfected FKSG76 or NMNAT3v1 by means of anti-FLAG antibody. (D) Whole cell NMNAT activity in NMNAT3v1 or FKSG76 transfected cells. (E) Structure of FKSG76 and NMNAT3v1. The ATP-binding domain absent in NMNAT3v1 is shown in green in FKSG76. ATP-binding residues are shown in yellow. The orientation of ATP (red) bound into the catalytic site is also shown. Columns represent the mean ± SEM of 3 (A) and 4 (D) experiments. Western blotting and Immunocytochemistry are representative of 4 independent experiments. *** p<0.001 vs control (Student's t test).

Figure 4. Effect of NMNAT3v1 or FKSG76 transfection on cellular or mitochondrial NAD and membrane potential.

(A) Whole cellular NAD content in control, NMNAT3v1- and FKSG76-transfected cells. Basal NAD content was 12.6±2 nmol/mg prot. (B) Western blotting evaluation of poly(ADP-ribose) (PAR) formation in HEK cells under control conditions or after transfection of FKSG76 and/or PARP1-cd. Tubulin is shown as loading control. (C) Densitometric evaluation of PAR formation shown in (B). (D) Immunocytochemical localization of PAR in HEK cells under control conditions or after transfection of FKSG76 and/or mitoPARP1cd. (E) Effect of FKSG76 transfection on mitochondrial membrane potential. (F) Representative experiment of the effect of FKSG76 transfection on oxygen consumption. The arrow indicates the time when cells were added to the respiration buffer. (G) Oxygen consumption rate in control or FKSG76-transfected cells. Columns represent the mean ± SEM of 4 (A), 5 (C) and 3 (G) experiments. Western blotting and immunocytochemistry are representative of 5 and 2 experiments, respectively.* p<0.05; ** p<0.01; *** p<0.001 vs control (Student's t test).

Figure 5. Identification of FKSG76 and NMNAT3v1 expression under different cellular conditions.

(A) Immunoprecipitation of NMNAT3 from mitochondrial extracts of HEK239 cells or FKSG76-transfected whole cell homogenate. VDAC is shown as a mitochondrial marker. (B) Dot-blot evaluation of the sensitivity of anti-NMNAT3 antibody to recombinant FKSG76. (C) Schematic representation of the 5′UTRs of FKSG76 and NMNAT3v1 cloned in the reported vector used in Luciferase reporter assay (See Methods); the uORF sequences are depicted in red. (D) Effect of uORFs or mutated uORFs (Mut-uORFs) on translational efficiency. (E) Time-dependent accumulation of ubiquitinated-proteins in MG132-exposed (10 µM) cells. Tubulin is shown as loading control. (F) Densitometric evaluation of ubiquitin accumulation shown in (E). (G) Western blotting evaluation of FKSG76 and NMNAT3v1 expression in HEK cells after different times of exposure to the proteasome inhibitor MG132. Positive control for FKSG76 or NMNAT3v1 are shown. Tubulin is shown as loading control. (H) Transcript levels of NMNAT1, −2 and −3 at different times after heat shock (44°C/30′). Note that for NMNAT3 analysis primers able to amplify a common region of NMNAT3v1 and FKSG76 have been used. Columns represent the mean ± SEM of 3 experiments. Western blotting or Dot-blot are representative of at least 3 experiments.* p<0.05; ** p<0.01; *** p<0.001 vs control (Student's t test).

Figure 6. Effect of exogenous NAD or its precursors on cellular and mitochondrial NAD.

(A) Quantification of NAD in control or FKSG67-transfected HEK cells. The effect of NAD, nicotinamide (Nam), NMN, nicotinamide riboside (NR) or nicotinic acid (NA) on NAD contents of FKSG-transfected cells is shown (NAD and its precursors have been added to the incubating media at 1 mM for 48 hrs). Basal NAD content was 12.6±2 nmol/mg prot. (B) Western blotting evaluation of the effect of NAD, Nam, NMN, NR or nicotinic acid NA (1 mM/48 hrs) on depletion of mitochondrial PAR content induced by FKSG76 co-transfection in mitoPARP1cd-transfected cells. Tubulin is shown as loading control. (C) Densitometric analysis on the experiment shown in (B). (D) Effect of oligomycin (10 µM/30 min) and/or glucose deprivation (30 min) on cellular ATP contents. (E) Effects of exogenous NAD (1 mM/3 hrs) on mitochondrial PAR contents in mitoPARP1cd-transfected cells under control conditions or exposed to oligomycin (10 µM) in the presence or absence of glucose. (F) Densitometric analysis on the experiment shown in (E). Columns represent the mean ± SEM of 3 experiments. Western blotting is representative of 3 (E) and 4 (C) experiments.* p<0.05; ** p<0.01 vs control (Student's t test).