Phosphorus, like other essential chemical elements – carbon, nitrogen, oxygen, and sulfur, is central to life. Inside cells, phosphorus is found in the form of salts and esters of phosphoric acid (orthophosphates, or simply Pi), and is widely present in a range of macromolecules (Hunter, 2012; Kamerlin et al., 2013; Westheimer, 1977). Phosphates and orthophosphates are critical for metabolic reactions in cells, and adding or removing phosphate groups to organic molecules (sugars, nucleotides, and phospholipids) is central to metabolism (Westheimer, 1987). However, a systems-level understanding of how phosphate balance is maintained in cells, in conjunction with the many metabolic roles of phosphate, remains a poorly addressed but fundamental biological question. Building an overarching organizational logic of how phosphates are integral to and mediate metabolic information flow has therefore not been systematically attempted.

Most cells (particularly free-living microbes or sessile organisms) experience fluctuations in available nutrients. Therefore, cells have evolved multiple regulatory mechanisms to sense and adapt to nutrients. In the contexts of phosphate availability, such regulatory responses have been well studied in plants, bacteria, or model eukaryotic cells such as the budding yeast Saccharomyces cerevisiae. For example, in yeast, the cellular transcriptional response to extended phosphate limitation is mediated by the PHO pathway (Lenburg and O'Shea, 1996; Ljungdahl and Daignan-Fornier, 2012; Mouillon and Persson, 2006; Secco et al., 2012). This system functions through the transcription factors, Pho4 and Pho2, which control the expression of phosphate metabolism related genes called the PHO regulon. This regulon includes genes that encode high- and low-affinity phosphate transporters, acid and alkaline phosphatases, vacuolar polyphosphate (polyP) synthesis proteins and vacuolar polyphosphatase (Bun-Ya et al., 1991; Lemire et al., 1985; Martinez and Persson, 1998; Ogawa et al., 2000; To-E et al., 1973), which collectively help in the acquisition, assimilation and release of phosphate. Several such examples of elaborate transcriptional-regulatory networks that respond to phosphate starvation are also known in plants (Fabiańska et al., 2019; Stigter and Plaxton, 2015; Zhang et al., 2014), bacteria (Lamarche et al., 2008) and other organisms (Bergwitz and Jüppner, 2011).

In contrast to our extensive knowledge about regulated gene expression during phosphate starvation, how distinct nodes in cellular metabolism are themselves coupled to phosphate balance is surprisingly not as clear. Effectively, our collective understanding of phosphate demand and utilization, and the role of phosphates in regulating the metabolic state of the cell, remains poor. In this perspective, we first identify metabolic processes that are irrevocably coupled to phosphate balance, and therefore overall cellular metabolic homeostasis. We illustrate how cellular metabolism depends on intracellular phosphate amounts, and can maintain Pi levels, by effectively ‘squeezing’ internal phosphate availability through specific metabolic outputs. We conceptualize phosphates in central metabolism into cycles, sources, and sinks. By defining and distinguishing these roles, it now becomes possible to more systematically contextualize how phosphates will control biochemical information flow in cells, and thereby determine cellular metabolic states.

Phosphates are integral components of lipids, nucleic acids, proteins, and sugars. They are present in many energetically costly metabolites, including reducing equivalents in the form of pyridine nucleotides (NADH and NADPH), cellular energy currencies (represented by adenosine triphosphate (ATP), phosphoribosyl phosphate (PRPP), pyrophosphates (PPi), pyridoxal phosphate (PLP)), and signaling molecules like inositol phosphates. Each of these molecules has substantial metabolic roles. Indeed, phosphates are ubiquitous, because of their unique chemistry. They easily ionize (and are retained within cells), yet phosphate bonds are stable, and so phosphates are ideally suited to link nucleotides, function as intermediary metabolites, and serve as energy sources (Westheimer, 1987). Under physiological conditions, many phosphate-containing metabolites participate in coupled biochemical reactions that effectively cycle the phosphate moieties (Westheimer, 1987). This type of role of phosphate-containing molecules as part of continuous cycles is relatively well-understood biochemically, and a few illustrative reactions are shown in Figure 1. Two classic examples would be ATP + AMP <->2 ADP, and the continuous cycling of ATP <-> ADP (with the concurrent release or use of a Pi molecule, and a water molecule) (Figure 1A). Other examples include reactions where hydrolysis of a phosphoanhydride bond drives product formation with either the incorporation of phosphate(s) into (Figure 1B), or the release of phosphate(s) from the product (Figure 1C).

Figure 1

Download asset Open asset

Importantly, such instances can be viewed as contained biochemical systems. If there is a fixed amount of phosphates, no new phosphate molecules are required for such a system to continue indefinitely, as long as the total amounts of these metabolites are unchanged. These enclosed ‘phosphate cycles’ use phosphates integrally in their reaction mechanisms. Therefore, although these reactions require phosphates, they can be considered distinct from the metabolic reactions that can immediately affect the total phosphate pools available in the cell, or reactions that must most rapidly change when phosphates become limiting (these are discussed later). Of course, the absolute amounts of these metabolites do change, depending on the metabolic needs of the cell. By first separating out these closed phosphate cycles, we can now conceptualize how other, distinct phosphate utilizing/dependent reactions (which are not part of these phosphate cycles) can be critical for phosphate homeostasis, and consequently regulate overall carbon metabolic flux.

Phosphate limitation can be viewed as two types – (1) extreme phosphate starvation, due to a severe or continuing depletion of phosphate in the extracellular environment, or (2) transient limitations in phosphate availability and/or in intracellular Pi levels. Our understanding of transcriptional responses to extreme phosphate starvation is reasonably advanced, at least in model systems like S. cerevisiae and plants (Fabiańska et al., 2019; Ljungdahl and Daignan-Fornier, 2012; Secco et al., 2012; Stigter and Plaxton, 2015; Zhang et al., 2014). Here, upon phosphate starvation, a combination of increased Pi acquisition (via extracellular Pi uptake), and increased Pi mobilization (via intracellular Pi stores) enables cells to restore intracellular Pi levels. This is illustrated in Figure 2A. In budding yeast for example, both Pi acquisition and mobilization from various reserves are mediated by the activation of the PHO transcriptional response, which regulates gene expression of multiple phosphate transporters and acid phosphatases. However, such elaborate responses will require time scales of min–hours, involving multiple, regulated transcription and translation steps.

Figure 2

Download asset Open asset

In contrast to severe phosphate starvation, there can also be transient limitations in internal phosphate levels. In such contexts, even before the phosphate-responsive gene-regulatory programs fully come into effect, changes in metabolism can immediately restore intracellular phosphate levels. We illustrate how this can happen, using a hypothetical reaction network (Figure 2B). In such a reaction network, some reactions will release phosphate, while others will incorporate/consume phosphate. Purely based on standard rules of mass action, when phosphate availability is reduced, the flux through reactions releasing phosphate will increase, while those that consume phosphates will decrease. Under such conditions, cells can therefore immediately mobilize Pi from intracellular metabolic reserves by biochemical processes driven by metabolite abundances and the law of mass action (Hackett et al., 2016). Thus, while regulated gene expression changes in response to severe phosphate starvation can carry out longer term adaptive responses, metabolic rewiring can rapidly maintain or restore intracellular Pi levels in response to transient fluctuations. This is the essence of dynamic metabolic regulation, where cells modulate their metabolic fluxes over short- and long-time scales to match their requirements (Metallo and Vander Heiden, 2013). Summarizing, while the long-term transcriptional, adaptive responses are well studied, we will substantiate the idea of how immediate metabolic rewiring can restore phosphates, through reactions that release or absorb Pi, stating specific examples.

A notable point emerges if such mobilization of metabolic Pi stores by mass-action based processes occur: as we will illustrate subsequently, there will be an accompanying rewiring of overall carbon flow, where reactions that release Pi are favored. This is associated with the accumulation of other central carbon (glucose) derived metabolites. This therefore highlights an underappreciated point of how tightly central carbon metabolism is coupled to phosphate homeostasis. By clearly identifying nodes of metabolism that will change with phosphate availability, we can predictably address how intracellular Pi changes will re-route carbon flux through very specific nodes of carbon metabolism to alter cellular metabolic states.

Extending beyond closed biochemical reactions where phosphates continuously cycle (discussed in the earlier section), we can now organize phosphate-dependent metabolic reactions into classes that would affect the internal phosphate amounts distinctly. These metabolites can be classified into two groups – phosphate sources and sinks (Figure 3A). Here, sources of phosphate can now be defined as metabolites which themselves may not contain phosphate/phosphorus, but importantly their synthesis results in phosphate release. These metabolites will therefore restore phosphate pools and balance Pi during transient Pi fluctuations. However, in order to do so, they should satisfy the following criteria; their synthesis should increase in response to phosphate limitation (and be accompanied by Pi release), the metabolic flux towards these molecules should be (conditionally) high, and they need not always accumulate in the cell. This definition therefore alters the idea of a source from a phosphate-containing molecule, to include molecules that themselves do not contain phosphates, but are made concurrent with Pi release. The metabolic precursors of these sources will be phosphate-containing molecules. As we will exemplify shortly, many metabolic pathway intermediates or ‘end products’ that do not themselves contain phosphates will act as very effective phosphate sources, since Pi will be released during the synthesis of these molecules.

Figure 3

Download asset Open asset

Contrarily, sinks of phosphate will be phosphate-containing metabolites, where the synthesis of these metabolites results in the storage of excess Pi. Sinks therefore should satisfy the following criteria; the synthesis of these molecules should decrease in response to phosphate limitation, they should have a slow turnover rate, and should accumulate in abundance.

By using such a definition, it becomes clear that a source and a sink need not necessarily be the same molecules, and in fact may often not even be in the same biochemical pathway. However, their respective synthesis must be anti-correlated in the cell, in response to phosphate limitation. This central idea is illustrated in Figure 3B. This source-sink balance will therefore be substantially governed by the allocation and amount of intracellular Pi. In such a context, the cellular response to transient phosphate depletion can now be understood as follows: whenever there is a transient Pi depletion, increased synthesis towards a Pi source will immediately occur, and this will rapidly release phosphate (Figure 3B). If phosphate pools are not immediately restored, the phosphate sinks (which are molecules that are slowly turned over) can be further mobilized to release inorganic phosphates. In summary, simple metabolic rewiring towards the synthesis of Pi sources can initially restore Pi, and the subsequent degradation/mobilization of sinks can continue this process of Pi release. Such a definition of sources and sinks contrasts with simply phosphate-rich metabolites involved in biochemical cycles (as described earlier). Making this classification now allows a systematic investigation of their distinct roles during changes in intracellular phosphate, and this will provide a predictable metabolic state signature for the cell.

What therefore might be the cellular sources of phosphates, and can they now be precisely identified? We systematically investigated the central carbon derived metabolic pathways (as obtained from KEGG [Ogata et al., 1999]) to identify possible cellular phosphate sources and sinks, using an example of yeast metabolic networks. From this comprehensive analysis, trehalose, serine, chorismate, inositol, and glycerol synthesis, all of which are accompanied by the release of Pi, represent the likely examples of phosphate sources by satisfying all required criteria as defined earlier (Figure 3C). Note here that none of these metabolites – trehalose, serine, chorismate, inositol or glycerol – contain phosphates. Instead, all of these metabolites are made via intermediate reactions where Pi is released, illustrating this idea of what a Pi source might be (and that this need not be a phosphate-containing metabolite). Trehalose is a major ‘storage’ carbohydrate that accumulates in cells at very high concentrations (Chester, 1963; François and Parrou, 2001; Stewart et al., 1950), and a recent report provides strong evidence of it being a phosphate-restoration source (Gupta et al., 2019; Gupta and Laxman, 2020). Trehalose is made from the sugar phosphates, glucose-6-phosphate (G6P) and UDP-glucose in two steps, where the last step (catalyzed by Tps2) results in trehalose formation and the release of one molecule of Pi (Figure 3C). In glucose replete conditions, these two steps are involved in futile cycling of trehalose during glycolysis, and trehalose synthesis allows proper glycolytic flux by maintaining phosphate balance (van Heerden et al., 2014b; van Heerden et al., 2014a). Pertinently, dampening the PHO response (which mimics phosphate limitation), results in carbon (glucose) flux rerouted towards trehalose synthesis, and away from the pentose phosphate pathway (PPP) and glycolysis (Gupta et al., 2019; Gupta and Laxman, 2020). This is largely governed by mass action, with no changes in glycolytic/PPP enzyme amounts required. Multiple lines of evidence indicate that rerouting metabolic flux towards trehalose restores the Pi balance (Gupta et al., 2019). Further, if the last step in trehalose synthesis (which releases phosphate) is removed (tps2Δ cells), this decreases the cellular phosphate pools (Gupta et al., 2019). Taken together, all these data satisfy the criteria for trehalose to be a phosphate source. A related, predictive case can be made for glycerol, which is synthesized during the lower arm of glycolysis and is accompanied by the release of Pi. Glycerol biosynthesis from dihydroxyacetone phosphate (DHAP) occurs in two steps, where the first step results in the synthesis of glycerol-3-phosphate (catalyzed by Gpd1 and Gpd2), and the second step results in the synthesis of glycerol and the release of one Pi molecule by the action of glycerol-3-phosphatases (Hor2 and Rhr2) (Austin and Mayer, 2020; Figure 3C). In line with the role of glycerol as a phosphate source, glycerol levels are elevated in response to phosphate limitation (Kazemi Seresht et al., 2011). Moreover, consistent with a role of trehalose and glycerol syntheses in phosphate restoration, the metabolic phosphatases involved in these biochemical conversions, trehalose-6-phosphate synthase/phosphatase and glycerol-3-phosphatases, respectively, are also transcriptionally upregulated during phosphate limitation (Ogawa et al., 2000). This suggests an amplified response (through the transcriptional induction of these enzymes) as a way to mobilize metabolic reserves of Pi. Thus, when phosphates are transiently limited, glucose-6-phosphate and dihydroxyacetone phosphate will reroute away from glycolysis and the PPP, and toward trehalose and glycerol respectively. While this can restore phosphate balance, it will also result in a reduction of the overall glycolytic and PPP flux.

Three other major biosynthetic processes derived from glucose metabolism result in abundant products and also release phosphate. Serine biosynthesis from 3-phosphoglycerate (3PG) occurs in three steps, and the last step releases one Pi molecule. Chorismate synthesis (a precursor for aromatic amino acids and the vitamins, p-aminobenzoate and p-hydroxybenzoate) requires one molecule of erythrose-4-P (E4P) (from the pentose phosphate pathway) and two molecules of phosphoenolpyruvate (PEP), and occurs in seven steps. Here, three molecules of Pi are released (Figure 3C). The prediction therefore is that flux towards these molecules will increase when phosphates are limiting. Indeed, consistent with this prediction, independent studies with phosphate starvation have all observed that amino acid amounts strongly increase (Boer et al., 2010; Gupta et al., 2019), along with reduced metabolites of the PPP and glycolysis. These data therefore are consistent with these metabolites acting as phosphate sources. Finally, a predictive case can be made for myo-inositol (inositol) biosynthesis as a source of phosphate. Inositol is perhaps the most abundantly present stereoisomer in yeast and mammalian cells (Desfougères et al., 2019). Inositol biosynthesis from glucose-6-phosphate (G6P) occurs in two steps. In the first step, Ino1 converts glucose-6-phosphate to inositol-1-phosphate. The second step is catalyzed by inositol monophosphatases (Inm1 and Inm2), and inositol-1-phosphate is converted to inositol and one molecule of Pi is released (Figure 3C). Although the inositol biosynthetic pathway is relatively poorly studied in yeast, and there is insufficient metabolic data in this (Pi relevant) context, inositol monophosphatase, Inm1, is upregulated under low phosphate conditions (Kazemi Seresht et al., 2011). Therefore, we can predict that this inositol synthesis, accompanied by Pi release, can help restore internal Pi levels. To summarize, phosphate sources can modulate the overall carbon flux distribution in a metabolic network depending on the internal phosphate levels, and will thereby determine the metabolic state of the cell.

What then are the sinks of phosphates? Nucleoside-, di-, tri-phosphates, NAD(P)⁺, inositol pyrophosphates, and polyphosphates (PolyP) are all putative phosphate sinks (and not sources). Is there evidence that is consistent with the plausibility of these metabolites as phosphate sinks? Going by the earlier described criteria of a sink, these metabolites are abundant in cells (Belenky et al., 2007; Ermakova et al., 1981; Gakière et al., 2018; Kukko and Saarento, 1983; Ljungdahl and Daignan-Fornier, 2012; Pestov et al., 2004; Rao et al., 1998). A second criterion is that upon phosphate limitation, their levels will reduce due to both decreased synthesis and increased degradation (mobilization), and this will release phosphate. The case for nucleotides as phosphate sinks: as sinks, the synthesis of these molecules should decrease in response to phosphate limitation, they should have a slow turnover rate, and should accumulate in abundance. These criteria appear to be well-satisfied for nucleotides (mono-, di-, tri phosphates). Phosphate limitation results in a rapid decrease in both their synthesis and steady-state levels (Ashihara et al., 1988; Boer et al., 2010; Klosinska et al., 2011). In a recent study, the reduction in (but not absence of) PHO gene expression (phosphate transporters, phosphatases etc.), which mimics phosphate limitation, also revealed reduced de novo nucleotide synthesis. This was not due to reduced amounts of nucleotide biosynthesis enzymes, but instead was governed by mass action (Gupta et al., 2019). Further, phosphate limitation also increases the amounts of 5’-, 3’-nucleotidases and nucleotide pyrophosphatase enzymes, which thereby increases nucleotide degradation (and will allow phosphate release) (Alipanah et al., 2018; Hammond et al., 2003; Kennedy et al., 2005; Misson et al., 2005; Rittmann et al., 2005; Shimano and Ashihara, 2006; Uhde-Stone et al., 2003; Wasaki et al., 2006; Yin et al., 2007). Phm8, a nucleotidase, is transcriptionally induced (a longer term response) upon carbon, nitrogen, and Pi starvation (Xu et al., 2013). The relavance of these nucleotidases (in the context of Pi sinks) will become clearer in the following paragraph.

Using very similar analyses, the phosphate moiety of NAD(P)⁺ and its intermediate metabolites also become potential phosphate sinks. Although direct estimates of these molecules during phosphate limitation have not been made, we can predict that there will be reduced synthesis of these metabolites, as they utilize nucleotide triphosphates as their phosphate donor, and nucleotide synthesis itself decreases under these conditions. Indeed, consistent with this prediction, during low Pi conditions, pyridine nucleotides, NAD⁺ and NAD⁺ intermediates (NMN and NaMN) are broken down by specific nucleotidases, and this will also release phosphate (Bogan and Brenner, 2010). Indeed, somewhat overlooked reports suggest connections between phosphate signaling and NAD(P)⁺ metabolism (Bogan and Brenner, 2010; Kato and Lin, 2014).

In this context of nucleotides, as well as NAD(P)⁺ molecules as Pi sinks, two relevant points emerge. In an insightful review, Bogan and Brenner summarize that ribonucleoside monophosphates (which are potential substrates of 5’-nucleotidases) come from both salvage and de novo synthesis. The 5’-nucleotidases themselves allow the reverse reactions of nucleoside kinases (and therefore oppose salvage pathways, and nucleotide generation) (Bogan and Brenner, 2010). These opposing substrate cycles, and the relative activity of phosphorylation/dephosphorylation, determine how much nucleoside monophosphates (and eventually triphosphates) are present, or if they will be converted to nucleosides. A more recent review points to the fact that upon extreme Pi starvation, as a ‘last resort’ cells induce enzymes that release Pi from nucleotides, since it would otherwise be illogical for cells to deplete its pools of nucleotides, which are critical for future cell division (Austin and Mayer, 2020). These points are all entirely consistent with nucleotides (and NAD related molecules) being Pi sinks, which will serve to eventually release Pi if phosphate levels within cells are not restored.

Finally, inositol pyrophosphates and polyphosphates are likely phosphate sink candidates. Inositol pyrophosphates have the highest proportion of phosphate groups for any metabolite, containing more phosphates than carbon atoms (Saiardi, 2012), and they regulate the expression of PHO related transcripts in yeast (Auesukaree et al., 2005; Lee et al., 2008; Lee et al., 2007), suggesting cross-talk between inositol pyrophosphate metabolism and phosphate signaling. Polyphosphates (polyP) are linear polymers of many phosphate residues, ranging from few to hundreds, and are linked by the same high-energy phosphoanhydride bonds that are found in ATP. PolyPs are mainly stored inside the vacuole, with small pools also found in cytosol, nucleus and mitochondria (Gerasimaitė and Mayer, 2016; Saiardi, 2012). Evidence from some organisms (bacteria, yeasts) suggest that polyphosphates are effective phosphate reservoirs, and upon phosphate starvation, these stores are degraded by the action of exo- and endo-polyphosphatases to restore phosphate (Brown and Kornberg, 2008; Kulaev et al., 1999; Kulaev and Kulakovskaya, 2000; Shirahama et al., 1996). We would therefore predict these metabolites to collectively be intracellular phosphate sinks, with inositol phosphates and polyphosphates likely to be the major sinks. All these are now directly testable predictions.

Two broad points emerge when we define phosphate sources and sinks in this way. First, we reiterate a close coupling of glucose metabolism with internal phosphate homeostasis. Altering flux toward different arms of glucose metabolism – glycolysis, the pentose phosphate pathway, and trehalose/glycogen synthesis – will therefore result in substantial changes in phosphate release/utilization (as shown earlier in Figure 3C). Second, we can now with some confidence predict key metabolic state changes (based on metabolic flux through these defined sources and sinks) due to a transient phosphate limitation (or ‘squeeze’). Primarily, trehalose, aromatic amino acids/chorismate, serine, glycerol, and inositol (phosphate sources) will all increase, while nucleotides, polyphosphates and (some) inositol pyrophosphates (phosphate sinks) will decrease. Flux through glycolysis and the pentose phosphate pathway will correspondingly decrease. Amounts of these hallmark metabolites therefore represent a concise metabolic signature of a phosphate-limited cell. In summary, this conceptualization of the phosphate sources and sinks together with their biochemical reactions illustrate the importance of these metabolites in maintaining the overall phosphate levels in the cell. Second, mass-action driven metabolic processes play a fundamental role in phosphate homeostasis by regulating flux through the key phosphate-related metabolic nodes. Collective evidence therefore indicates that there is an intimate relationship and interdependence between phosphate balance and carbon metabolism.

While bacteria do not have internal pools of phosphates, in eukaryotic cells, phosphates reside in multiple cellular compartments, including the cytoplasm, nucleus, vacuole, endoplasmic reticulum, mitochondria and so on (Beauvoit et al., 1989; Kato and Lin, 2014). Phosphate supply inside the cell therefore depends on the transport of phosphate across the plasma membrane and membranes of various organelles mediated by transporter proteins. The content and functions of phosphates vary in different organelles. Inorganic phosphate is required for the production of ATP in both the cytosol and mitochondria. In the cytosol, Pi stimulates the activity of hexokinase and phosphofructokinase, which catalyze irreversible steps of glycolysis. Moreover, Pi is a necessary substrate for glyceraldehyde-3-phosphate dehydrogenase, which catalyzes the first reaction of the payoff phase in glycolytic pathway (and therefore is a major consumer of phosphate). In mitochondria, inorganic phosphate is required for oxidative phosphorylation, as the ATP synthase complex uses it for producing ATP (Ferreira and Pedersen, 1993). Vacuoles have high amounts of inorganic phosphates in the cell. Estimates of the Pi concentration in vacuoles is ~100 mM, which is ~100 times higher than cytosolic concentrations (Okorokov et al., 1980). The mitochondrial polyPs, which are 14 residues long, constitute ~10% of the total cellular polyP content (Beauvoit et al., 1989). In yeast, cell envelopes also contain polyP, with ≥20% of the total cellular polyP content (Kulaev and Kulakovskaya, 2000). Vacuoles are the major polyP storage compartments in yeast, with the total polyP concentration of ~200 mM (Kornberg, 1999; Urech et al., 1978). By increasing the mobilization and transport of phosphate into the cytosol, vacuolar polyP helps maintain cellular phosphate homeostasis during limiting-phosphate conditions. A genetic screen in yeast revealed the distribution of proteins important for polyP content in various compartments: cytoplasm (59 proteins), nucleus (56 proteins), mitochondria (42 proteins), vacuole (20 proteins), and endoplasmic reticulum (13 proteins), collectively indicating that multiple groups of proteins help manage polyP metabolism in distinct cellular compartments (Freimoser et al., 2006). In line with this, two different vacuolar transport sub-complexes: Vtc4/Vtc3/Vtc1 and Vtc4/Vtc2/Vtc1 are present in the cell, the first on the vacuolar membrane and the second in endoplasmic reticulum and nucleus (Gerasimaitė and Mayer, 2016). Together, these studies suggest that cellular phosphate pools have distinct subcellular localizations. This adds additional (and currently entirely unaddressed) complexity in how phosphate homeostasis is maintained in the cell. We know relatively little about how intracellular pools of phosphates control metabolic information flow. Comprehensively addressing the role of phosphates in mediating metabolic information flow will therefore require such knowledge of how phosphates in these distinct pools function in unison.

Surprisingly, an integrative understanding of phosphates as central regulators of metabolic information transfer in cells has been missing. Because phosphates are everywhere, this appears to be an intractable problem. However, by going beyond investigating molecular responses to phosphate limitation, and by accounting for and systematically reorganizing both short- and long-term responses (mediated by metabolic/mass action and molecular control respectively), we can obtain insights into coordinated process that maintain phosphate levels. In this article, we have conceptualized and categorized phosphates into metabolic cycles, and then sinks and source. This is a first step toward building an integrative understanding of how phosphates control metabolic information flow. Since the metabolic control of phosphate homeostasis depends on central carbon metabolism and vice versa, identifying control-hubs that facilitate cross-talk between these two systems should therefore be an exciting area of future inquiry.

1. 1. Alipanah L
   2. Winge P
   3. Rohloff J
   4. Najafi J
   5. Brembu T
   6. Bones AM

   (2018) Molecular adaptations to phosphorus deprivation and comparison with nitrogen deprivation responses in the diatom Phaeodactylum tricornutum

   PLOS ONE 13:e0193335.

   https://doi.org/10.1371/journal.pone.0193335

   • PubMed
   • Google Scholar
2. 1. Ashihara H
   2. Li X-NI
   3. Ukaji T

   (1988) Effect of inorganic phosphate on the biosynthesis of purine and pyrimidine nucleotides in Suspension-Cultured cells of Catharanthus roseus*

   Annals of Botany 61:225–232.

   https://doi.org/10.1093/oxfordjournals.aob.a087547

   • Google Scholar
3. 1. Auesukaree C
   2. Tochio H
   3. Shirakawa M
   4. Kaneko Y
   5. Harashima S

   (2005) Plc1p, Arg82p, and Kcs1p, enzymes involved in inositol pyrophosphate synthesis, are essential for phosphate regulation and polyphosphate accumulation in Saccharomyces cerevisiae

   Journal of Biological Chemistry 280:25127–25133.

   https://doi.org/10.1074/jbc.M414579200

   • Google Scholar
4. 1. Austin S
   2. Mayer A

   (2020) Phosphate homeostasis - A vital metabolic equilibrium maintained through the INPHORS signaling pathway

   Frontiers in Microbiology 11:1367.

   https://doi.org/10.3389/fmicb.2020.01367

   • PubMed
   • Google Scholar
5. 1. Beauvoit B
   2. Rigoulet M
   3. Guerin B
   4. Canioni P

   (1989) Polyphosphates as a source of high energy phosphates in yeast mitochondria: A ³¹ P NMR study

   FEBS Letters 252:17–21.

   https://doi.org/10.1016/0014-5793(89)80882-8

   • Google Scholar
6. 1. Belenky P
   2. Racette FG
   3. Bogan KL
   4. McClure JM
   5. Smith JS
   6. Brenner C

   (2007) Nicotinamide riboside promotes Sir2 silencing and extends lifespan via nrk and Urh1/Pnp1/Meu1 pathways to NAD+

   Cell 129:473–484.

   https://doi.org/10.1016/j.cell.2007.03.024

   • PubMed
   • Google Scholar
7. 1. Bergwitz C
   2. Jüppner H

   (2011) Phosphate sensing

   Advances in Chronic Kidney Disease 18:132–144.

   https://doi.org/10.1053/j.ackd.2011.01.004

   • PubMed
   • Google Scholar
8. 1. Boer VM
   2. Crutchfield CA
   3. Bradley PH
   4. Botstein D
   5. Rabinowitz JD

   (2010) Growth-limiting intracellular metabolites in yeast growing under diverse nutrient limitations

   Molecular Biology of the Cell 21:198–211.

   https://doi.org/10.1091/mbc.e09-07-0597

   • PubMed
   • Google Scholar
9. 1. Bogan KL
   2. Brenner C

   (2010) 5′-Nucleotidases and their new roles in NAD+ and phosphate metabolism

   New Journal of Chemistry 34:845–853.

   https://doi.org/10.1039/b9nj00758j

   • Google Scholar
10. 1. Brown MR
    2. Kornberg A

    (2008) The long and short of it - polyphosphate, PPK and bacterial survival

    Trends in Biochemical Sciences 33:284–290.

    https://doi.org/10.1016/j.tibs.2008.04.005

    • PubMed
    • Google Scholar
11. 1. Bun-Ya M
    2. Nishimura M
    3. Harashima S
    4. Oshima Y

    (1991) The PHO84 gene of Saccharomyces cerevisiae encodes an inorganic phosphate transporter

    Molecular and Cellular Biology 11:3229–3238.

    https://doi.org/10.1128/MCB.11.6.3229

    • PubMed
    • Google Scholar
12. 1. Chester VE

    (1963) The dissimilation of the carbohydrate reserves of a strain of Saccharomyces cerevisiae

    Biochemical Journal 86:153–160.

    https://doi.org/10.1042/bj0860153

    • PubMed
    • Google Scholar
13. 1. Desfougères Y
    2. Wilson MSC
    3. Laha D
    4. Miller GJ
    5. Saiardi A

    (2019) ITPK1 mediates the lipid-independent synthesis of inositol phosphates controlled by metabolism

    PNAS 116:24551–24561.

    https://doi.org/10.1073/pnas.1911431116

    • PubMed
    • Google Scholar
14. 1. Ermakova SA
    2. Mansurova SE
    3. Kalebina TS
    4. Lobakova ES
    5. Selyach IO
    6. Kulaev IS

    (1981) Accumulation of pyrophosphate and other energy-rich phosphorous compounds under various conditions of yeast growth

    Archives of Microbiology 128:394–397.

    https://doi.org/10.1007/BF00405919

    • PubMed
    • Google Scholar
15. 1. Fabiańska I
    2. Bucher M
    3. Häusler RE

    (2019) Intracellular phosphate homeostasis - A short way from metabolism to signaling

    Plant Science 286:57–67.

    https://doi.org/10.1016/j.plantsci.2019.05.018

    • PubMed
    • Google Scholar
16. 1. Ferreira GC
    2. Pedersen PL

    (1993) Phosphate transport in mitochondria: past accomplishments, present problems, and future challenges

    Journal of Bioenergetics and Biomembranes 25:483–492.

    https://doi.org/10.1007/BF01108405

    • PubMed
    • Google Scholar
17. 1. François J
    2. Parrou JL

    (2001) Reserve carbohydrates metabolism in the yeast Saccharomyces cerevisiae

    FEMS Microbiology Reviews 25:125–145.

    https://doi.org/10.1111/j.1574-6976.2001.tb00574.x

    • PubMed
    • Google Scholar
18. 1. Freimoser FM
    2. Hürlimann HC
    3. Jakob CA
    4. Werner TP
    5. Amrhein N

    (2006) Systematic screening of polyphosphate (poly P) levels in yeast mutant cells reveals strong interdependence with primary metabolism

    Genome Biology 7:R109.

    https://doi.org/10.1186/gb-2006-7-11-r109

    • PubMed
    • Google Scholar
19. 1. Gakière B
    2. Hao J
    3. de Bont L
    4. Pétriacq P
    5. Nunes-Nesi A
    6. Fernie AR

    (2018) NAD + biosynthesis and signaling in plants

    Critical Reviews in Plant Sciences 1:259–307.

    https://doi.org/10.1080/07352689.2018.1505591

    • Google Scholar
20. 1. Gerasimaitė R
    2. Mayer A

    (2016) Enzymes of yeast polyphosphate metabolism: structure, enzymology and biological roles

    Biochemical Society Transactions 44:234–239.

    https://doi.org/10.1042/BST20150213

    • PubMed
    • Google Scholar
21. 1. Gupta R
    2. Walvekar AS
    3. Liang S
    4. Rashida Z
    5. Shah P
    6. Laxman S

    (2019) A tRNA modification balances carbon and nitrogen metabolism by regulating phosphate homeostasis

    eLife 8:e44795.

    https://doi.org/10.7554/eLife.44795

    • PubMed
    • Google Scholar
22. 1. Gupta R
    2. Laxman S

    (2020) tRNA wobble-uridine modifications as amino acid sensors and regulators of cellular metabolic state

    Current Genetics 66:475–480.

    https://doi.org/10.1007/s00294-019-01045-y

    • Google Scholar
23. 1. Hackett SR
    2. Zanotelli VR
    3. Xu W
    4. Goya J
    5. Park JO
    6. Perlman DH
    7. Gibney PA
    8. Botstein D
    9. Storey JD
    10. Rabinowitz JD

    (2016) Systems-level analysis of mechanisms regulating yeast metabolic flux

    Science 354:aaf2786.

    https://doi.org/10.1126/science.aaf2786

    • PubMed
    • Google Scholar
24. 1. Hammond JP
    2. Bennett MJ
    3. Bowen HC
    4. Broadley MR
    5. Eastwood DC
    6. May ST
    7. Rahn C
    8. Swarup R
    9. Woolaway KE
    10. White PJ

    (2003) Changes in gene expression in Arabidopsis shoots during phosphate starvation and the potential for developing smart plants

    Plant Physiology 132:578–596.

    https://doi.org/10.1104/pp.103.020941

    • PubMed
    • Google Scholar
25. 1. Hunter T

    (2012) Why nature chose phosphate to modify proteins

    Philosophical Transactions of the Royal Society B: Biological Sciences 367:2513–2516.

    https://doi.org/10.1098/rstb.2012.0013

    • Google Scholar
26. 1. Kamerlin SC
    2. Sharma PK
    3. Prasad RB
    4. Warshel A

    (2013) Why nature really chose phosphate

    Quarterly Reviews of Biophysics 46:1–132.

    https://doi.org/10.1017/S0033583512000157

    • PubMed
    • Google Scholar
27. 1. Kato M
    2. Lin S-J

    (2014) Regulation of NAD+ metabolism, signaling and compartmentalization in the yeast Saccharomyces cerevisiae

    DNA Repair 23:49–58.

    https://doi.org/10.1016/j.dnarep.2014.07.009

    • Google Scholar
28. 1. Kazemi Seresht A
    2. Palmqvist EA
    3. Olsson L

    (2011) The impact of phosphate scarcity on pharmaceutical protein production in S. cerevisiae: linking transcriptomic insights to phenotypic responses

    Microbial Cell Factories 10:104.

    https://doi.org/10.1186/1475-2859-10-104

    • PubMed
    • Google Scholar
29. 1. Kennedy EJ
    2. Pillus L
    3. Ghosh G

    (2005) Pho5p and newly identified nucleotide pyrophosphatases/ phosphodiesterases regulate extracellular nucleotide phosphate metabolism in Saccharomyces cerevisiae

    Eukaryotic Cell 4:1892–1901.

    https://doi.org/10.1128/EC.4.11.1892-1901.2005

    • PubMed
    • Google Scholar
30. 1. Klosinska MM
    2. Crutchfield CA
    3. Bradley PH
    4. Rabinowitz JD
    5. Broach JR

    (2011) Yeast cells can access distinct quiescent states

    Genes & Development 25:336–349.

    https://doi.org/10.1101/gad.2011311

    • PubMed
    • Google Scholar
31. 1. Kornberg A

    (1999) Inorganic polyphosphate: a molecule of many functions

    Progress in Molecular and Subcellular Biology 23:1–18.

    https://doi.org/10.1007/978-3-642-58444-2_1

    • PubMed
    • Google Scholar
32. 1. Kukko E
    2. Saarento H

    (1983) Accumulation of pyrophosphate in Escherichia coli. relationship to growth and nucleotide synthesis

    Archives of Microbiology 136:209–211.

    https://doi.org/10.1007/BF00409846

    • PubMed
    • Google Scholar
33. 1. Kulaev I
    2. Vagabov V
    3. Kulakovskaya T

    (1999) New aspects of inorganic polyphosphate metabolism and function

    Journal of Bioscience and Bioengineering 88:111–129.

    https://doi.org/10.1016/S1389-1723(99)80189-3

    • PubMed
    • Google Scholar
34. 1. Kulaev I
    2. Kulakovskaya T

    (2000) Polyphosphate and phosphate pump

    Annual Review of Microbiology 54:709–734.

    https://doi.org/10.1146/annurev.micro.54.1.709

    • PubMed
    • Google Scholar
35. 1. Lamarche MG
    2. Wanner BL
    3. Crépin S
    4. Harel J

    (2008) The phosphate regulon and bacterial virulence: a regulatory network connecting phosphate homeostasis and pathogenesis

    FEMS Microbiology Reviews 32:461–473.

    https://doi.org/10.1111/j.1574-6976.2008.00101.x

    • PubMed
    • Google Scholar
36. 1. Lee Y-S
    2. Mulugu S
    3. York JD
    4. O'Shea EK
    5. O’Shea EK

    (2007) Regulation of a Cyclin-CDK-CDK Inhibitor Complex by Inositol Pyrophosphates

    Science 316:109–112.

    https://doi.org/10.1126/science.1139080

    • Google Scholar
37. 1. Lee YS
    2. Huang K
    3. Quiocho FA
    4. O'Shea EK

    (2008) Molecular basis of cyclin-CDK-CKI regulation by reversible binding of an inositol pyrophosphate

    Nature Chemical Biology 4:25–32.

    https://doi.org/10.1038/nchembio.2007.52

    • PubMed
    • Google Scholar
38. 1. Lemire JM
    2. Willcocks T
    3. Halvorson HO
    4. Bostian KA

    (1985) Regulation of repressible acid phosphatase gene transcription in Saccharomyces cerevisiae

    Molecular and Cellular Biology 5:2131–2141.

    https://doi.org/10.1128/MCB.5.8.2131

    • PubMed
    • Google Scholar
39. 1. Lenburg ME
    2. O'Shea EK

    (1996) Signaling phosphate starvation

    Trends in Biochemical Sciences 21:383–387.

    https://doi.org/10.1016/0968-0004(96)10048-7

    • PubMed
    • Google Scholar
40. 1. Ljungdahl PO
    2. Daignan-Fornier B

    (2012) Regulation of amino acid, Nucleotide, and phosphate metabolism in Saccharomyces cerevisiae

    Genetics 190:885–929.

    https://doi.org/10.1534/genetics.111.133306

    • PubMed
    • Google Scholar
41. 1. Martinez P
    2. Persson BL

    (1998) Identification, cloning and characterization of a derepressible na+-coupled phosphate transporter in Saccharomyces cerevisiae

    Molecular and General Genetics MGG 258:628–638.

    https://doi.org/10.1007/s004380050776

    • Google Scholar
42. 1. Metallo CM
    2. Vander Heiden MG

    (2013) Understanding metabolic regulation and its influence on cell physiology

    Molecular Cell 49:388–398.

    https://doi.org/10.1016/j.molcel.2013.01.018

    • PubMed
    • Google Scholar
43. 1. Misson J
    2. Raghothama KG
    3. Jain A
    4. Jouhet J
    5. Block MA
    6. Bligny R
    7. Ortet P
    8. Creff A
    9. Somerville S
    10. Rolland N
    11. Doumas P
    12. Nacry P
    13. Herrerra-Estrella L
    14. Nussaume L
    15. Thibaud MC

    (2005) A genome-wide transcriptional analysis using Arabidopsis thaliana affymetrix gene chips determined plant responses to phosphate deprivation

    PNAS 102:11934–11939.

    https://doi.org/10.1073/pnas.0505266102

    • PubMed
    • Google Scholar
44. 1. Mouillon JM
    2. Persson BL

    (2006) New aspects on phosphate sensing and signalling in Saccharomyces cerevisiae

    FEMS Yeast Research 6:171–176.

    https://doi.org/10.1111/j.1567-1364.2006.00036.x

    • PubMed
    • Google Scholar
45. 1. Ogata H
    2. Goto S
    3. Sato K
    4. Fujibuchi W
    5. Bono H
    6. Kanehisa M

    (1999) KEGG: kyoto encyclopedia of genes and genomes

    Nucleic Acids Research 27:29–34.

    https://doi.org/10.1093/nar/27.1.29

    • PubMed
    • Google Scholar
46. 1. Ogawa N
    2. DeRisi J
    3. Brown PO

    (2000) New components of a system for phosphate accumulation and polyphosphate metabolism in Saccharomyces cerevisiae revealed by genomic expression analysis

    Molecular Biology of the Cell 11:4309–4321.

    https://doi.org/10.1091/mbc.11.12.4309

    • PubMed
    • Google Scholar
47. 1. Okorokov LA
    2. Lichko LP
    3. Kulaev IS

    (1980) Vacuoles: main compartments of potassium, magnesium, and phosphate ions in Saccharomyces carlsbergenis cells

    Journal of Bacteriology 144:661–665.

    https://doi.org/10.1128/JB.144.2.661-665.1980

    • PubMed
    • Google Scholar
48. 1. Pestov NA
    2. Kulakovskaya TV
    3. Kulaev IS

    (2004) Inorganic polyphosphate in mitochondria of Saccharomyces cerevisiae at phosphate limitation and phosphate excess

    FEMS Yeast Research 4:643–648.

    https://doi.org/10.1016/j.femsyr.2003.12.008

    • PubMed
    • Google Scholar
49. 1. Rao NN
    2. Liu S
    3. Kornberg A

    (1998) Inorganic polyphosphate in Escherichia coli: the phosphate regulon and the stringent response

    Journal of Bacteriology 180:2186–2193.

    https://doi.org/10.1128/JB.180.8.2186-2193.1998

    • PubMed
    • Google Scholar
50. 1. Rittmann D
    2. Sorger-Herrmann U
    3. Wendisch VF

    (2005) Phosphate starvation-inducible gene ushA encodes a 5' nucleotidase required for growth of Corynebacterium glutamicum on media with nucleotides as the phosphorus source

    Applied and Environmental Microbiology 71:4339–4344.

    https://doi.org/10.1128/AEM.71.8.4339-4344.2005

    • PubMed
    • Google Scholar
51. 1. Saiardi A

    (2012) How inositol pyrophosphates control cellular phosphate homeostasis?

    Advances in Biological Regulation 52:351–359.

    https://doi.org/10.1016/j.jbior.2012.03.002

    • PubMed
    • Google Scholar
52. 1. Secco D
    2. Wang C
    3. Shou H
    4. Whelan J

    (2012) Phosphate homeostasis in the yeast Saccharomyces cerevisiae, the key role of the SPX domain-containing proteins

    FEBS Letters 586:289–295.

    https://doi.org/10.1016/j.febslet.2012.01.036

    • PubMed
    • Google Scholar
53. 1. Shimano F
    2. Ashihara H

    (2006) Effect of long-term phosphate starvation on the levels and metabolism of purine nucleotides in suspension-cultured Catharanthus roseus cells

    Phytochemistry 67:132–141.

    https://doi.org/10.1016/j.phytochem.2005.10.013

    • PubMed
    • Google Scholar
54. 1. Shirahama K
    2. Yazaki Y
    3. Sakano K
    4. Wada Y
    5. Ohsumi Y

    (1996) Vacuolar function in the phosphate homeostasis of the yeast Saccharomyces cerevisiae

    Plant and Cell Physiology 37:1090–1093.

    https://doi.org/10.1093/oxfordjournals.pcp.a029058

    • PubMed
    • Google Scholar
55. 1. Stewart LC
    2. Richtmyer NK
    3. Hudson CS

    (1950) The preparation of trehalose from yeast¹

    Journal of the American Chemical Society 72:2059–2061.

    https://doi.org/10.1021/ja01161a050

    • Google Scholar
56. 1. Stigter KA
    2. Plaxton WC

    (2015) Molecular mechanisms of phosphorus metabolism and transport during leaf senescence

    Plants 4:773–798.

    https://doi.org/10.3390/plants4040773

    • PubMed
    • Google Scholar
57. 1. To-E A
    2. Ueda Y
    3. Kakimoto SI
    4. Oshima Y

    (1973) Isolation and characterization of acid phosphatase mutants in Saccharomyces cerevisiae

    Journal of Bacteriology 113:727–738.

    https://doi.org/10.1128/JB.113.2.727-738.1973

    • PubMed
    • Google Scholar
58. 1. Uhde-Stone C
    2. Zinn KE
    3. Ramirez-Yáñez M
    4. Li A
    5. Vance CP
    6. Allan DL

    (2003) Nylon filter arrays reveal differential gene expression in proteoid roots of white lupin in response to phosphorus deficiency

    Plant Physiology 131:1064–1079.

    https://doi.org/10.1104/pp.102.016881

    • PubMed
    • Google Scholar
59. 1. Urech K
    2. Dürr M
    3. Boller T
    4. Wiemken A
    5. Schwencke J

    (1978) Localization of polyphosphate in vacuoles of Saccharomyces cerevisiae

     Arch Microbiol 116:275–278.

    https://doi.org/10.1007/BF00417851

    • Google Scholar
60. 1. van Heerden JH
    2. Wortel MT
    3. Bruggeman FJ
    4. Heijnen JJ
    5. Bollen YJ
    6. Planqué R
    7. Hulshof J
    8. O'Toole TG
    9. Wahl SA
    10. Teusink B

    (2014a) Fatal attraction in Glycolysis: how Saccharomyces cerevisiae manages sudden transitions to high glucose

    Microbial Cell 1:103–106.

    https://doi.org/10.15698/mic2014.01.133

    • PubMed
    • Google Scholar
61. 1. van Heerden JH
    2. Wortel MT
    3. Bruggeman FJ
    4. Heijnen JJ
    5. Bollen YJM
    6. Planque R
    7. Hulshof J
    8. O'Toole TG
    9. Wahl SA
    10. Teusink B

    (2014b) Lost in transition: start-up of glycolysis yields subpopulations of nongrowing cells

    Science 343:1245114.

    https://doi.org/10.1126/science.1245114

    • Google Scholar
62. 1. Wasaki J
    2. Shinano T
    3. Onishi K
    4. Yonetani R
    5. Yazaki J
    6. Fujii F
    7. Shimbo K
    8. Ishikawa M
    9. Shimatani Z
    10. Nagata Y
    11. Hashimoto A
    12. Ohta T
    13. Sato Y
    14. Miyamoto C
    15. Honda S
    16. Kojima K
    17. Sasaki T
    18. Kishimoto N
    19. Kikuchi S
    20. Osaki M

    (2006) Transcriptomic analysis indicates putative metabolic changes caused by manipulation of phosphorus availability in rice leaves

    Journal of Experimental Botany 57:2049–2059.

    https://doi.org/10.1093/jxb/erj158

    • PubMed
    • Google Scholar
63. 1. Westheimer FH

    (1977) The hydrolysis of phosphate esters

    Pure and Applied Chemistry 49:1059–1067.

    https://doi.org/10.1351/pac197749071059

    • Google Scholar
64. 1. Westheimer F

    (1987) Why nature chose phosphates

    Science 235:1173–1178.

    https://doi.org/10.1126/science.2434996

    • Google Scholar
65. 1. Xu YF
    2. Létisse F
    3. Absalan F
    4. Lu W
    5. Kuznetsova E
    6. Brown G
    7. Caudy AA
    8. Yakunin AF
    9. Broach JR
    10. Rabinowitz JD

    (2013) Nucleotide degradation and ribose salvage in yeast

    Molecular Systems Biology 9:665.

    https://doi.org/10.1038/msb.2013.21

    • PubMed
    • Google Scholar
66. 1. Yin Y
    2. Shimano F
    3. Ashihara H

    (2007) Involvement of rapid nucleotide synthesis in recovery from phosphate starvation of Catharanthus roseus cells

    Journal of Experimental Botany 58:1025–1033.

    https://doi.org/10.1093/jxb/erl264

    • Google Scholar
67. 1. Zhang Z
    2. Liao H
    3. Lucas WJ

    (2014) Molecular mechanisms underlying phosphate sensing, signaling, and adaptation in plants

    Journal of Integrative Plant Biology 56:192–220.

    https://doi.org/10.1111/jipb.12163

    • PubMed
    • Google Scholar

Author details

1. Ritu Gupta

   Institute for Stem Cell Science and Regenerative Medicine (inStem), Bangalore, India

   Present address

   National Institutes of Health, Bethesda, United States

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-0563-6599

2. Sunil Laxman

   Institute for Stem Cell Science and Regenerative Medicine (inStem), Bangalore, India

   For correspondence

   sunil@instem.res.in

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-0861-5080

• 2,425

  views

• 274

  downloads

• 13

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.