NAD Biosynthesis: De Novo, Preiss-Handler, Salvage

Most of the NAD+ inside a mammalian cell was never assembled from raw materials. It was rebuilt from a fragment of itself, often within hours of the last enzyme that tore it apart. That single fact reorganises how the whole subject should be taught.

How is NAD+ made in cells?

NAD+ is made through three converging routes: de novo synthesis from the amino acid L-tryptophan via the kynurenine pathway, the Preiss-Handler route from nicotinic acid, and the salvage route that recycles nicotinamide through the enzyme NAMPT. Salvage carries most of the flux in mammalian tissue, which is why NAMPT is described as the rate-limiting step of NAD biosynthesis.

The common oversimplification is that NAD biosynthesis is a linear assembly line fed by dietary vitamin B3. It isn't. All three routes funnel into one of two mononucleotide intermediates, NaMN or NMN, and every route passes through the same family of adenylyltransferases (NMNAT1, NMNAT2, NMNAT3) before a finished dinucleotide exists. This page covers each route enzyme by enzyme, the consumer enzymes that create the demand, how the balance shifts by tissue, what changes with age, and how flux is measured in the laboratory.

What does NAD+ actually do inside a cell?

NAD+ (nicotinamide adenine dinucleotide) has two jobs that are chemically unrelated, and the second one is the reason NAD biosynthesis has to run continuously. The first job is redox carrier: NAD+ accepts a hydride ion to become NADH during glycolysis at glyceraldehyde-3-phosphate dehydrogenase (GAPDH), during the tricarboxylic acid cycle, and during fatty acid oxidation, then surrenders those electrons to Complex I of the electron transport chain. In that role NAD+ is catalytic. It cycles between oxidised and reduced states without being destroyed, and the total pool size never changes.

The second job destroys it. A family of enzymes cleaves the glycosidic bond between nicotinamide and the ribose of NAD+, transferring the ADP-ribose portion somewhere else and releasing free nicotinamide as a by-product. Here NAD+ is a consumed substrate, not a catalyst. Every one of those reactions permanently removes a molecule from the pool and dumps a nicotinamide fragment into the cytosol.

That is the demand side of the equation, and it is what NAD research is ultimately organised around. A separate enzyme, NAD kinase (NADK), phosphorylates NAD+ into NADP+, which serves reductive biosynthesis and antioxidant defence through NADPH. The NADP pool is accounted for separately and is not interchangeable with the NAD+ pool in the way casual summaries imply.

How does the de novo route build NAD+ from tryptophan?

The de novo arm of NAD biosynthesis starts with L-tryptophan and runs through the kynurenine pathway, which takes eight enzymatic steps before anything resembling a nucleotide appears. Tryptophan is opened by indoleamine 2,3-dioxygenase (IDO1) or tryptophan 2,3-dioxygenase (TDO2) to N-formylkynurenine, hydrolysed to kynurenine by arylformamidase, hydroxylated by kynurenine 3-monooxygenase (KMO), and cleaved by kynureninase to 3-hydroxyanthranilic acid.

The branch point sits one step later. 3-hydroxyanthranilate 3,4-dioxygenase (HAAO) produces an unstable intermediate, ACMS, which spontaneously cyclises into quinolinic acid. Competing with that spontaneous reaction is the enzyme ACMSD (2-amino-3-carboxymuconate-6-semialdehyde decarboxylase), which diverts ACMS away from quinolinate and toward picolinic acid and complete oxidation. Research reported by Katsyuba and colleagues showed that inhibiting ACMSD increases de novo flux in liver and kidney tissue, which identifies that branch point as the real gatekeeper of the pathway.

Quinolinate phosphoribosyltransferase (QPRT) then converts quinolinic acid plus PRPP (5-phosphoribosyl-1-pyrophosphate) into nicotinic acid mononucleotide, NaMN, and the de novo route merges with the Preiss-Handler route from that point forward. The classic nutritional figure, roughly 60 mg of dietary tryptophan yielding about 1 mg of niacin equivalent, reflects how inefficient this arm is relative to direct precursor uptake.

What makes the Preiss-Handler route different?

The Preiss-Handler route is the shortest path in NAD biosynthesis, converting nicotinic acid into NAD+ in three enzymatic steps. Jack Preiss and Philip Handler described it in 1958 in the Journal of Biological Chemistry, and the sequence has held up essentially unchanged since.

Step one is NAPRT (nicotinate phosphoribosyltransferase), which condenses nicotinic acid with PRPP to give NaMN. Step two is one of the NMNAT isoforms adenylylating NaMN into NaAD (nicotinic acid adenine dinucleotide). Step three is NADSYN1, a glutamine-dependent NAD synthetase that amidates NaAD into finished NAD+.

The distinction that matters most for experimental design is substrate specificity. Nicotinic acid and nicotinamide are not interchangeable inputs. NAPRT accepts nicotinic acid only; NAMPT accepts nicotinamide only. A cell line that has silenced NAPRT expression cannot use nicotinic acid at all, no matter how much is present in the medium, and becomes entirely dependent on salvage. That dependency is the mechanistic basis for a long line of cancer biology work using NAMPT inhibitors such as FK866, where NAPRT-null tumour lines are selectively vulnerable and NAPRT-positive lines are rescued by nicotinic acid. NAPRT and NADSYN1 expression is concentrated in liver, kidney, and intestine rather than distributed evenly across tissues.

Why does the salvage route carry most NAD biosynthesis?

Salvage dominates because the consumer enzymes keep handing it free substrate. Every sirtuin, PARP, and CD38 reaction releases nicotinamide, and NAMPT (nicotinamide phosphoribosyltransferase, originally described as PBEF or visfatin) immediately recaptures it, condensing nicotinamide with PRPP to form NMN. NMNAT then adenylylates NMN into NAD+. Two steps, closed loop, no new carbon skeleton required.

Work from Revollo, Grimm and Imai in 2004 established NAMPT as the rate-limiting enzyme of mammalian salvage and linked its activity to sirtuin function, and the enzyme has been treated as the control point of NAD biosynthesis ever since. NAMPT transcription is also circadian: the CLOCK:BMAL1 complex drives Nampt expression, SIRT1 feeds back on that complex, and NAD+ concentrations in several tissues oscillate across a 24-hour cycle. Experiments that sample tissue at inconsistent clock times are measuring a moving target.

Compartmentalisation matters just as much as enzyme identity. NMNAT1 is nuclear, NMNAT2 is cytosolic and Golgi-associated, NMNAT3 is mitochondrial. Three physically separate NAD+ pools exist within one cell, they turn over at different rates, and a whole-tissue measurement averages them into a single number that can conceal large compartment-level shifts. Anyone comparing results across papers should check which NAD half-life and which compartment the authors were actually reporting.

Where do nicotinamide riboside and NMN fit into the network?

Nicotinamide riboside and NMN enter the salvage arm rather than forming a fourth independent route. Bieganowski and Brenner described the nicotinamide riboside kinase pathway in 2004, identifying NRK1 and NRK2 as the enzymes that phosphorylate NR directly into NMN, bypassing NAMPT entirely. That bypass is why NR is used experimentally in systems where NAMPT is inhibited or genetically disrupted.

NMN itself is phosphorylated and poorly membrane-permeable, which created one of the more contested questions in the field. The widely reported explanation is that extracellular NMN is dephosphorylated to NR by the ecto-5'-nucleotidase CD73 at the cell surface, the NR crosses the membrane through equivalent nucleoside transporters, and NRK1 re-phosphorylates it inside. A 2019 report in Nature Metabolism from Grozio and colleagues proposed Slc12a8 as a direct NMN transporter in small intestine. Independent groups have disputed that assignment, and the question has not been settled. A page claiming otherwise is overstating the literature.

NR and NMN are also chemically fragile in solution compared with the dinucleotide, which is why peptide degradation pathways and the handling principles in a storage guide are not side issues in precursor work. Hydrolysis of the glycosidic bond proceeds in aqueous conditions and accelerates with temperature.

Which route dominates in which tissue?

The three arms of NAD biosynthesis are not equally available everywhere, and enzyme expression is the limiting variable rather than precursor supply. The table below compares entry points, enzymes, and where each route is most relevant in published work.

What consumes NAD+ fast enough to require constant resynthesis?

Four enzyme families account for the bulk of NAD+ consumption, and all four release nicotinamide back into the salvage loop. Sirtuins (SIRT1 through SIRT7) are NAD+-dependent deacylases that strip acetyl and other acyl groups from histones and metabolic enzymes, consuming one NAD+ molecule per catalytic event and producing 2'-O-acetyl-ADP-ribose plus nicotinamide. Their activity is therefore directly coupled to NAD+ availability rather than merely correlated with it.

PARPs are the heaviest consumers under stress. PARP1 detects DNA strand breaks and polymerises long chains of ADP-ribose onto acceptor proteins, and extensive genotoxic damage can draw down cellular NAD+ sharply, which is the mechanism behind the long-standing link between NAD and PARP-mediated repair.

CD38 and its relative CD157 are ectoenzymes that hydrolyse NAD+ and generate cyclic ADP-ribose, a calcium-mobilising second messenger. Because CD38 faces the extracellular space, it also degrades circulating NMN and NR before they reach their targets.

SARM1 is the outlier. A 2017 report in Neuron from Essuman and colleagues showed that the SARM1 TIR domain possesses intrinsic NAD+ cleavage activity, which reframed axon degeneration as a controlled NAD+ destruction event rather than a passive failure. Collectively, this consumption is the reason NAD biosynthesis is better modelled as a flux than as a stored quantity.

Why do NAD+ pools appear to decline with age?

Reported age-related declines in tissue NAD+ are generally attributed to increased consumption rather than a collapse in synthetic capacity. Camacho-Pereira and colleagues reported in Cell Metabolism in 2016 that CD38 expression rises with age in several mouse tissues and that CD38-null animals are protected from the associated NAD+ decline, which placed ecto-NADase activity at the centre of the hypothesis.

Two other contributors are regularly proposed: chronic low-grade PARP activation driven by accumulated DNA damage, and reduced NAMPT expression in certain tissues, which would throttle the salvage arm of NAD biosynthesis exactly where it carries the most traffic. Research suggests these mechanisms operate together rather than competing.

Honesty about the evidence requires one caveat that most summaries skip. Reported magnitudes of decline vary widely between studies, and a meaningful share of that variation is methodological. NAD+ degrades quickly in unquenched tissue, extraction chemistry differs between laboratories, and older colorimetric assays do not discriminate cleanly between NAD+ and NADH. Comparisons across papers using different analytical platforms should be read as directional, not quantitative. Interested readers can follow how the picture is shifting in NAD research coverage.

What if one route in the network is blocked?

What if NAMPT is pharmacologically inhibited?

The cell loses its ability to recycle nicotinamide and must survive on whatever de novo, Preiss-Handler, or NR-derived input is available. FK866 (APO866) is the reference tool compound for this in published work, and the resulting NAD+ depletion is severe in cell types lacking NAPRT. Supplying nicotinic acid rescues NAPRT-positive cells and does nothing for NAPRT-null cells, which is the cleanest demonstration that the arms of NAD biosynthesis are genuinely separate until they converge.

What if a cell line simply doesn't express NAPRT?

Treat nicotinic acid as an inert addition and plan around the salvage arm instead. NAPRT silencing by promoter methylation has been documented across multiple cancer cell lines, and it is the single most common reason a precursor experiment produces no measurable change. Checking NAPRT and NRK1 expression before selecting a precursor avoids weeks of uninterpretable data.

What if measured NAD+ barely moves after precursor addition?

Consider compartment distribution and consumption rate before concluding the precursor failed. Flux can rise substantially while the steady-state pool stays flat, because increased supply is matched by increased sirtuin, PARP or CD38 turnover. This is exactly why stable isotope tracing has displaced simple pool measurement in careful NAD biosynthesis work.

What does the evidence actually support about raising NAD+?

Here's the honest answer: the biochemistry of NAD biosynthesis is well mapped, and the translation of that biochemistry into predictable outcomes is not. Precursor entry points, enzyme kinetics and compartmentalisation are established. What remains genuinely unsettled is how reliably raising a precursor concentration translates into a sustained change in any specific tissue compartment, how much of an administered precursor is degraded by CD38 before it arrives, and whether direct NMN transport exists at all. Papers that present the first half as settled and the second half as equally settled are overselling. Comparative discussions such as MOTS-c versus NAD are useful only when they keep that distinction visible.

How is NAD biosynthesis measured in the laboratory?

LC-MS/MS is the current reference method, because it resolves NAD+, NADH, NMN, NaMN, NAD, NR and nicotinamide as distinct analytes in a single run rather than reporting a lumped value. Enzymatic cycling assays remain common for total pool quantification and are sensitive, but they do not separate individual intermediates. HPLC with UV detection sits between the two and is still widely used for intermediate profiling.

Sample handling dominates data quality. NAD+ and NADH differ in acid and base stability, which is why protocols split samples into acidic and alkaline extractions, and why quenching speed at collection determines whether a result is real or an artefact of degradation during processing.

The method that actually answers pathway questions is stable isotope tracing. Feeding labelled nicotinamide, labelled nicotinic acid or labelled tryptophan and tracking label incorporation into downstream intermediates shows which arm of NAD biosynthesis is carrying flux, which a static concentration measurement can never reveal. Genetically encoded fluorescent biosensors add compartment-specific, live-cell readouts that bulk extraction destroys.

Analytical rigour on the input side matters just as much. The same separation science used for flux work underpins compound verification, which is why HPLC purity analysis and mass spectrometry verification are standard prerequisites before any NAD biosynthesis experiment begins.

How we researched this page

This page was assembled from the published literature only, with no first-party testing or internal data behind any statement. Primary sources were peer-reviewed papers indexed in PubMed and MEDLINE covering the kynurenine pathway, the 1958 Preiss and Handler characterisation, NAMPT and salvage regulation, the nicotinamide riboside kinase pathway, NAD-consuming enzymes, and age-related NAD+ decline. Enzyme identities, isoform compartmentalisation and reaction stoichiometry were cross-checked against curated enzyme and pathway databases including UniProt, KEGG and BRENDA.

Where the literature is divided, as with the proposed Slc12a8 NMN transporter, the disagreement is stated rather than resolved. Where a figure could not be verified against a primary source, it was left out rather than estimated. No claim on this page describes an outcome in a person.

All compounds discussed on this page are sold for laboratory research use only and are not for human consumption. Real Peptides publishes lot certificates of analysis at /coas/, and the broader meaning of that designation is covered under research use only. Anyone with questions about animal models or companion-animal health should talk to their veterinarian rather than relying on a web page.

Further reading across the NAD library and the analytical verification material: NAD FAQ, NAD history, is NAD legal, NAD benefits research, NAD for sale, NAD shipping, NAD vs SS-31, NAD and addiction research, NAD news 2026, wholesale NAD, is Tylenol NAD, NAD oral taste, is NAD worth it, NAD and men over 40, how to read a COA, HPLC vs mass spec, research quality standards, storage best practices, lyophilized vs liquid stability, and the laboratory locations page.

The most useful mental correction anyone can make about NAD biosynthesis is to stop picturing a reservoir and start picturing a river. The quantity of NAD+ sitting in a cell at any instant says very little on its own, because that quantity is the momentary balance between NAMPT running flat out and a crowd of sirtuins, PARPs and NADases pulling the other way. Two cells can show identical NAD+ concentrations while one is turning its pool over several times faster than the other. Flux is the variable that carries the biology. Concentration is just the water level.

References

Peer-reviewed sources on NAD+ biosynthesis and metabolism indexed in PubMed, listed for research context.

1. Enzymology of extracellular NAD metabolism. Cell Mol Life Sci, 2021. PubMed PMID 33755743.
2. Metabolic Alterations in NADSYN1-Deficient Cells. Metabolites, 2023. PubMed PMID 38132878.
3. Kynurenine pathway metabolites and enzymes involved in redox reactions. Neuropharmacology, 2017. PubMed PMID 26970015.
4. Visfatin--a review. Eur Rev Med Pharmacol Sci, 2011. PubMed PMID 21381495.