Phosphofructokinase-1 (PFK1), the enzyme that converts fructose 6-phosphate into fructose 1,6-bisphosphate, is the textbook answer to this question. It catalyzes the first committed step of glycolysis and is heavily regulated by the cell’s energy status, making it the classic “gatekeeper” of the pathway. But the honest picture is more interesting than a single enzyme with a single title. Modern research has shown that control over glycolytic flux is shared among several enzymes and shifts depending on the tissue, the cell’s oxygen supply, and whether the cell is healthy or cancerous.
Why PFK1 Gets the Title
Glycolysis has ten steps, but only three are essentially irreversible under normal cellular conditions: the reactions catalyzed by hexokinase, PFK1, and pyruvate kinase. Because those steps can’t easily run backward, they act as natural control points. Of the three, PFK1 has traditionally received the most attention as the rate-limiting enzyme because it sits at the gateway to the pathway’s committed phase. Once PFK1 does its job, the glucose molecule is locked into being broken down for energy; there’s no turning back toward storage or other fates.
PFK1 responds to an impressive array of signals. When the cell is low on energy, molecules like AMP and ADP activate PFK1, ramping up glycolysis to produce more ATP. When energy is abundant, ATP itself and citrate (an intermediate from the next major energy pathway, the citric acid cycle) inhibit PFK1, slowing things down. This push-and-pull responsiveness is what makes PFK1 so effective as a gatekeeper: it constantly adjusts the pace of glycolysis to match what the cell actually needs.1Nature Communications. Structural basis for allosteric regulation of human phosphofructokinase-1
The Fructose 2,6-Bisphosphate Amplifier
If PFK1 is the gatekeeper, fructose 2,6-bisphosphate (Fru-2,6-P2) is the molecule that tells the gatekeeper how wide to open the door. Fru-2,6-P2 is one of the most potent activators of PFK1 known. It’s not a glycolytic intermediate itself but a signaling molecule produced and degraded by a single bifunctional enzyme called PFK-2/FBPase-2. This enzyme can either make Fru-2,6-P2 (activating glycolysis) or break it down (slowing glycolysis), depending on which of its two catalytic activities is switched on.
The switch depends on hormonal signals and varies by tissue. In the liver, the hormone glucagon (which rises during fasting) triggers a chemical modification that inactivates the Fru-2,6-P2-making side and activates the Fru-2,6-P2-destroying side. The result: PFK1 loses its most powerful stimulator, glycolysis slows, and the liver shifts toward releasing glucose into the blood instead. In heart muscle, the opposite pattern plays out: several signaling pathways activate the Fru-2,6-P2-making side, keeping glycolysis humming to meet the heart’s constant demand for energy.2PubMed Central. 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase: head-to-head with a bifunctional enzyme that controls glycolysis This tissue-specific tuning is one reason the “rate-limiting step” question doesn’t have a perfectly universal answer. The relative importance of PFK1 depends partly on how much Fru-2,6-P2 is floating around, and that varies from organ to organ.
The Other Two Irreversible Steps
Hexokinase, the very first enzyme in glycolysis, phosphorylates glucose as it enters the cell. In many tissues, hexokinase is inhibited by its own product, glucose 6-phosphate. When downstream glycolysis slows and intermediates pile up, hexokinase gets the message and backs off. In some contexts, particularly in rapidly dividing cells, hexokinase exerts substantial control over overall glycolytic flux. One study of cancer cell glycolysis explicitly identified hexokinase and PFK1 together as the rate-limiting enzymes that cap the overall glycolytic rate.3Journal of Biological Chemistry. Evidence That Does Not Support Pyruvate Kinase M2 (PKM2)-catalyzed Reaction as a Rate-limiting Step in Cancer Cell Glycolysis
Pyruvate kinase, the final enzyme in the pathway, catalyzes the irreversible conversion of phosphoenolpyruvate and ADP into pyruvate and ATP.4PubMed Central. An overview of structure, function, and regulation of pyruvate kinases It, too, is allosterically regulated. In yeast, fructose 1,6-bisphosphate (the product of PFK1) powerfully activates pyruvate kinase, creating a feed-forward loop: when PFK1 is active and producing fructose 1,6-bisphosphate, pyruvate kinase is stimulated to keep pace so intermediates don’t pile up.5PubMed. The allosteric regulation of pyruvate kinase by fructose-1,6-bisphosphate In mammalian tissues, different pyruvate kinase isoforms exist in different organs, each with distinct regulatory properties that fine-tune the endpoint of glycolysis to local metabolic needs.
Why “Rate-Limiting Step” Is an Oversimplification
The traditional idea of a single rate-limiting step implies that one enzyme acts as an absolute bottleneck while every other enzyme in the pathway has negligible influence on the overall rate. Metabolic Control Analysis (MCA), a quantitative framework developed in the 1970s and 1980s, replaced this intuitive notion with something more precise: the concept of a flux control coefficient. Each enzyme in a pathway gets a number between zero and one describing how much influence it has over the pathway’s total flux. Importantly, all those numbers add up together, meaning control is always shared among the enzymes in the pathway, never perfectly concentrated in just one.6PubMed Central. Metabolic control analysis: a tool for designing strategies to manipulate metabolic pathways
When researchers applied MCA to real biological systems, the results often challenged textbook expectations. In one analysis of glycolysis in potato tuber tissue under aerobic conditions, PFK1 had a surprisingly low control coefficient. Far more control over flux lay in the steps around phosphoenolpyruvate and beyond, well past PFK1 in the pathway.7PubMed Central. Metabolic Control Analysis of glycolysis in tuber tissue of potato (Solanum tuberosum): explanation for the low control coefficient of phosphofructokinase over respiratory flux A study of glycolysis in human heart muscle during hibernation (a state of reduced blood flow) found a pattern of regulation that used coordinated modulation across multiple sites, differing substantially from the traditional single-bottleneck model.8PubMed. Metabolic control analysis of anaerobic glycolysis in human hibernating myocardium replaces traditional concepts of flux control
None of this means PFK1 isn’t important. In many mammalian tissues under standard conditions, PFK1 probably does hold the largest single share of flux control. But calling it “the” rate-limiting step suggests a binary on/off bottleneck that doesn’t reflect how the pathway actually behaves. The control is distributed, and the distribution shifts with conditions.
How Cancer Cells Rewire the Bottleneck
Cancer cells are famous for their voracious glucose consumption, a phenomenon often called the Warburg effect. Even when oxygen is plentiful, many tumor cells prefer to run glycolysis at full speed and ferment the resulting pyruvate into lactate rather than sending it into mitochondrial respiration. This metabolic shift involves changes at several glycolytic control points.
One player that has attracted enormous research attention is pyruvate kinase M2 (PKM2), an isoform of the final glycolytic enzyme that is preferentially expressed in many tumors. PKM2 can flip between a highly active tetrameric form and a less active dimeric form. In its less active state, it slows down the last step of glycolysis, which may seem counterproductive but actually allows glycolytic intermediates to pile up and be diverted into biosynthetic pathways the tumor needs for growth. PKM2 also has a moonlighting role: it can translocate to the cell nucleus and influence gene expression, further promoting aerobic glycolysis.9PubMed Central. PKM2-driven metabolic reprogramming in digestive system tumors: mechanisms, therapeutic advances, and clinical challenges
For a while, some researchers proposed PKM2 as a rate-limiting step in tumor glycolysis. But experimental evidence pushed back. Work published in the Journal of Biological Chemistry argued that hexokinase and PFK1 remain the true rate-limiting enzymes in cancer cells, while PKM2 and lactate dehydrogenase serve a different function: rapidly clearing upstream intermediates to prevent traffic jams in the pathway.3Journal of Biological Chemistry. Evidence That Does Not Support Pyruvate Kinase M2 (PKM2)-catalyzed Reaction as a Rate-limiting Step in Cancer Cell Glycolysis In other words, PKM2’s unusual behavior reshapes how intermediates are distributed but doesn’t dictate the pathway’s overall throughput. The upstream enzymes still hold that role.
When Oxygen Drops, the Rules Change
Cells that find themselves short on oxygen need to ramp up glycolysis quickly, because their mitochondria can no longer produce ATP efficiently. The master regulator behind this shift is a transcription factor called HIF-1α (hypoxia-inducible factor-1α). Under low-oxygen conditions, HIF-1α accumulates and switches on the genes for multiple glycolytic enzymes and glucose transporters, broadly increasing the pathway’s capacity rather than just tweaking a single step.10PubMed. Regulation of glycolysis by the hypoxia-inducible factor (HIF): implications for cellular physiology This transcriptional boost affects hexokinase, PFK1, pyruvate kinase, and the transporters that bring glucose into the cell in the first place.
But transcription takes time, and cells sometimes need to respond faster than gene expression allows. Research on intestinal epithelial cells and on human platelets (which lack nuclei entirely and can’t perform conventional gene activation) revealed that cells can still boost glycolysis under low oxygen through non-transcriptional mechanisms. Investigators found that glycolytic enzymes and glucose transporters physically assemble into complexes under hypoxic conditions, a response that occurs even when transcription or translation is blocked. This assembly appears to be independent of HIF-1α-driven gene expression.11PubMed Central. Hypoxia induces a glycolytic complex in intestinal epithelial cells independent of HIF-1-driven glycolytic gene expression It’s a striking finding because it means the cell has at least two independent mechanisms for ramping up glycolysis under stress: one slow (new enzyme production) and one fast (physically reorganizing the enzymes it already has).
The Glycolytic Metabolon
The idea that glycolytic enzymes don’t just float freely in the cytoplasm but physically cluster together has been around for decades. Early work described glycolytic enzymes binding to structural proteins in skeletal muscle and to the inner surface of red blood cell membranes, forming what was called a multienzyme complex that compartmentalizes the glycolytic process.12PubMed. Supramolecular organization of glycolytic enzymes
Recent research has given this concept fresh molecular detail. A 2024 study in Nature Metabolism showed that hexokinase 1 (the first glycolytic enzyme) can be chemically modified in a way that promotes its attachment to the outer mitochondrial membrane, where it then helps organize a full glycolytic metabolon. This clustering appears to coordinate glycolytic ATP production with mitochondrial ATP production, creating a kind of assembly line on the mitochondrial surface.13PubMed Central. Organization of a functional glycolytic metabolon on mitochondria for metabolic efficiency
Why does physical clustering matter for the rate-limiting step question? If glycolytic enzymes are organized into a complex, intermediates can be handed directly from one enzyme to the next rather than diffusing randomly through the cytoplasm. This “substrate channeling” could reduce the effective bottleneck at any single step because the intermediates never accumulate in the bulk solution where they’d be diluted. It challenges the assumption behind classical rate-limiting analysis, which envisions each enzyme encountering its substrate only through random collision in a well-mixed pool.
Glycolytic Oscillations in Yeast
One of the more unusual windows into glycolytic control comes from yeast, where glycolysis can oscillate. Under specific conditions, concentrations of glycolytic intermediates, NADH, and ATP rhythmically rise and fall in a synchronized pattern. Populations of yeast cells can even synchronize their oscillations with each other through the exchange of acetaldehyde between cells.14Biophysical Journal. Transduction of Intracellular and Intercellular Dynamics in Yeast Glycolytic Oscillations
Temperature experiments on yeast extracts revealed that PFK1 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH, the sixth step) are the most temperature-sensitive steps of glycolysis, meaning temperature changes most readily disrupt the oscillation pattern through those enzymes.15PubMed Central. Control of glycolytic oscillations by temperature Intracellular potassium concentration also oscillates in sync with NADH and ATP, suggesting that ion balance is tied into the rhythm as well.16PubMed. On the coupling of intracellular K+ to glycolytic oscillations in yeast These oscillation studies underscore a point that applies broadly: the dynamic behavior of glycolysis depends on feedback loops and energy-sensing molecules operating across multiple enzymes, not just at one bottleneck.
Red Blood Cells as a Special Case
Red blood cells offer a naturally simplified system for studying glycolytic control because they lack mitochondria and rely entirely on glycolysis for ATP. Without the citric acid cycle or oxidative phosphorylation to worry about, every ATP molecule a red blood cell uses comes from breaking glucose down to lactate.
Research on human red blood cell glycolysis examined the influence of magnesium on various glycolytic enzymes and found that hexokinase was the most sensitive to magnesium levels, while steps further downstream like phosphoglycerate kinase showed no rate-limiting behavior and no magnesium dependence.17PubMed. The regulatory role for magnesium in glycolytic flux of the human erythrocyte This is consistent with the general finding that control concentrates in the early irreversible steps, but the specific distribution can shift depending on conditions like ion concentrations and substrate availability. In a cell whose entire energy supply depends on glycolysis, the consequences of shifting that control can be medically relevant: inherited deficiencies in pyruvate kinase, for instance, are one of the most common enzymatic causes of chronic hemolytic anemia.
Post-Translational Modifications Add Another Layer
Beyond allosteric regulation and gene expression changes, glycolytic enzymes can be chemically modified after they’re made. Phosphorylation, acetylation, and other post-translational modifications can alter an enzyme’s activity, its tendency to form complexes, or its location within the cell. Plant cells provide a particularly clear example: multiple glycolytic and citric acid cycle enzymes have been identified as targets of post-translational modifications that allow rapid and reversible changes in metabolic flux, helping the plant adjust to shifting environmental demands without waiting for new proteins to be built.18PubMed Central. Regulation of plant glycolysis and the tricarboxylic acid cycle by posttranslational modifications
In mammalian cells, the O-GlcNAcylation of hexokinase 1 described earlier in the metabolon context is one such modification, directly linking nutrient sensing to enzyme organization on the mitochondrial surface.13PubMed Central. Organization of a functional glycolytic metabolon on mitochondria for metabolic efficiency Phosphorylation of the bifunctional enzyme PFK-2/FBPase-2, as described by hormone-driven signaling in liver and heart, is another.2PubMed Central. 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase: head-to-head with a bifunctional enzyme that controls glycolysis These modifications mean that the rate-limiting character of any given enzyme can be toggled on a timescale of seconds to minutes, much faster than changes in gene expression. They’re another reason the identity of the “bottleneck” isn’t fixed: it can be reprogrammed in real time.
When Glycolysis and Gluconeogenesis Run Simultaneously
Glycolysis breaks glucose down; gluconeogenesis builds it back up. In tissues like the liver, both pathways share several reversible steps and are usually kept from running at full speed simultaneously, because doing so would create a futile cycle that wastes ATP without accomplishing anything. The irreversible steps catalyzed by hexokinase, PFK1, and pyruvate kinase in glycolysis are bypassed by different enzymes in gluconeogenesis precisely to allow independent control of each direction.
But the separation isn’t perfect. Modeling work using Markov chain analysis estimated that the energetic cost of futile cycling through the early phosphorylation steps can be substantial, with a lower bound of roughly 15 ATP molecules hydrolyzed per glucose molecule that cycles through these reactions.19PubMed Central. Futile cycles revisited: a markov chain model of simultaneous glycolysis and gluconeogenesis This is far more than the net yield of glycolysis itself, which produces only two ATP per glucose under anaerobic conditions. It highlights why tight regulation at PFK1 (and its gluconeogenic counterpart, fructose-1,6-bisphosphatase) matters so much: even modest simultaneous activity in both directions can be energetically expensive. The rate-limiting step, in this context, is as much about preventing wasteful backward flow as it is about controlling forward throughput.