PolyP (inorganic polyphosphate) is a linear polymer of tens to hundreds of orthophosphate residues linked by high-energy phosphoanhydride bonds. This polymer is present in all living organisms from bacteria to mammals. Until recently, most of the studies on polyP have focused on its function in prokaryotes. In prokaryotes, polyP has been implicated in many unrelated processes ranging from basic metabolism to structural functions. However, polyP analysis and function in higher eukaryotes has been gaining momentum recently. In the present review, we mainly aim to discuss the proposed intracellular functions of polyP in higher eukaryotes and its detection methods.
PolyP (inorganic polyphosphate) is a linear polymer of orthophosphate residues linked by phosphoanhydride bonds as in ATP (Figure 1). PolyP structure is flexible with no inherent tertiary structure. This polymer is present in all living organisms from bacteria to mammals and possesses distinctive properties. At physiological pH, each phosphate carries a monovalent negative charge, making polyP an exceptionally anionic polymer. As a multivalent anion, it has the capacity to chelate different cations and change their availability in cells. Furthermore, in the cellular environment, polyP is a structurally flexible and chemically stable molecule possessing scaffolding properties . PolyP is sometimes considered a molecular fossil owing to its synthesis in the prebiotic Earth and its proposed role as an energy-transfer molecule instead of ATP in the early phases of the origin of life . However, increasing evidence is linking this polymer to numerous and multifaceted cellular functions. These functions depend on the size and concentration of the molecule, its subcellular localization, and the type of organism. We refer the reader to two recent excellent reviews on polyP, one focusing on the emergent role played by polyP in platelets , and a second addressing the functions of polyP in parasitic protozoa . In the present article, we review the currently known intracellular functions of polyP in higher eukaryotes and propose some possible new functions.
Schematic representation of polyP structure
PolyP is particularly abundant in bacteria and yeast, where it is present in the millimolar range and comprises long polymers that can reach hundreds of phosphates in length. By contrast, in mammalian organisms, polyP chains are short and exist in the micromolar range. They have been described in various tissues and subcellular localizations [5,6].
The concentration of polyP in the cell results from the concerted and dynamic balance between synthesis and catalysis. Several enzymes contribute to the synthesis of polyP. PPK (polyphosphate kinase) catalyses the polymerization of the terminal phosphate of ATP into a polyP chain in a fully reversible and highly processive reaction . PPK is found in bacteria and Dictyosteliida amoebae that feed on bacteria and thus may have acquired the bacterial PPK gene by horizontal gene transfer . A second class of polyP-synthesizing enzymes has been identified in amoebae with the sequence and properties of an actin-related protein. This enzyme has been shown to polymerize into an actin-like filament concurrent with its synthesis of a polyP chain [DdPPK2 (Dictyostelium discoideum PPK2)]; however, no genetic evidence has been provided in support of its in vivo polyP-synthesizing activity .
The first, and so far only, eukaryotic enzyme involved in polyP synthesis that has been identified and characterized genetically and biochemically is from the budding yeast Saccharomyces cerevisiae and corresponds to Vtc4 (vacuolar transporter chaperone 4) . Yeast Vtc4 is part of the VTC complex, a heteromeric membrane protein anchored in the vacuole, cell membrane and acidocalcisomes that has been shown to possess polyP polymerase activity. However, owing to the transporter nature of the VTC complex, it has been hypothesized that it could also participate in polyP transport across membranes . The enzyme responsible for polyP synthesis in mammalian cells has not yet been identified, but it is clear that mammalian cells possess neither PPK nor VTC4. They do, however, contain proteins with sequence similarity to DdPPK2. Alternatively, polyP synthesis in mammalian cells might follow a different metabolic pathway. Identification of this synthetic pathway has been hampered by the fact that cell lysates do not show polyP-synthesizing activity. This could be due to the fact that, in cells, this activity might be an energy-dependent membrane-associated process that is disrupted upon protein extraction [5,6]. Alternatively, the very dynamic levels and high rate of turnover of polyP in eukaryotic cells  suggests the presence of multiple polyP phosphatases (see below) that could limit the ability to detect polyP synthesis from crude cell extract. Moreover, it has also been suggested that PPK activity could be driven by a protonmotive force, bypassing the need for Pi and ATP  or even as a by-product of multiple enzyme complexes . What is clear, however, is that the definitive identification of the mammalian polyphosphate kinase will be instrumental for the characterization of these molecules in higher eukaryotes.
In prokaryotes and lower eukaryotes, polyP can be hydrolysed by exo- (PPXs) and endo- (PPNs) polyphosphatases. In yeast, PPX1 hydrolyses the terminal phosphate of polyP to form Pi . The homologous H-Prune protein, from the DDH (dihydrodiol dehydrogenase) superfamily, was the first mammalian PPX to be identified . In yeast, PPN1 cleaves polyP internally to generate shorter polymer chains. No mammalian PPN homologue has been characterized to date. An additional PPN activity has been identified in both yeast and mammalian cells stemming from the Nudix (nucleoside diphosphate linked to X) superfamily member DIPP (diphosphoinositol-polyphosphate phosphatase) . This enzyme represents a multiple substrate enzyme that not only hydrolyses inositol pyrophosphates and nucleotide dimmers , but also works as a PPN.
Several polyP detection techniques are available, each with different advantages. Enzyme-based techniques have been widely used because they are quantitative and easy to perform . However, some of the enzymes used as reagents are chain-length-specific, and thus the product of their activity may not be a true representation of the polyP amount. PAGE combined with the metachromatic stain Toluidine Blue or DAPI staining can resolve mixtures of polyP of approximately 2–450 phosphates in length (Figure 2). The detection limit of this technique is sufficient to detect polyP from bacteria and yeast cells where polyP is very abundant; however, it is not sensitive enough in mammalian cells . Other qualitative or semi-quantitative physical methods include paper chromatography, 31P-NMR, ESI–MS and the analysis of terminal phosphate groups with phosphoglucokinase .
PolyP analysis in S. cerevisiae by PAGE
At the ultrastructural level, polyP can be detected indirectly by immunocytochemistry with the polyphosphate-binding domain of PPX1 . In vivo, polyP can be localized with DAPI staining. Fluorescence of the DAPI–polyP complex shows a shift of the broad emission spectrum with a peak at ~525 nm that can be distinguished from fluorescence of free DAPI or DAPI–DNA complexes, which have emission maxima at ~460 nm . DAPI fluorescence coupled with flow cytometry analysis has also been used to analyse intracellular polyP in mammalian cell populations . However, despite, the use of this technique in mammalian cells, its use in yeast has not been straightforward. Yeast offers the cleanest model system in which to assay the technique's specificity due to the existence of genetic mutations that abolish polyP synthesis. In this system, very few reports exist in which DAPI is used as a way to detect polyP and we find clean results difficult to obtain with this technique (A. Saiardi, unpublished work). Therefore, despite the methodologies available, it is still essential to further develop more sensitive, accurate and easily accessible detection techniques in order for polyP research to become more widespread.
In bacteria, the existence of non-proteic membrane channels is well documented. These channels have a distinctive chemical structure; they are formed of polyP–Ca2+ and the polyester PHB [poly-(R)-3-hydroxybutryrate] and are able to control cation membrane transport [21,22]. Although mammalian cells do not possess pure non-proteic membrane channels, polyP has been found tightly associated with many membrane channel/transporters. It is therefore likely that the role of polyP in controlling cation transport evolved very early and was then integrated into the complexity of the protein world in eukaryotic cells.
In mammalian cells, polyP has been shown to be associated with mPTP (mitochondria permeability transition pore) , TRPA1 (transient receptor potential ankyrin 1)  and TRPM8 (transient receptor potential melastatin 8) channels . Mitochondria possess many different ion-transporting systems whose molecular mechanism is still not fully understood. The mPTP is a supramacromolecular multiprotein non-selective pore complex localized in the inner mitochondrial membrane, providing a permeation pathway to ions and small molecules. The molecular nature and mechanism of mPTP opening is of crucial importance because of its involvement in cell death in several diseases; however, it is poorly understood. The mPTP complexity, together with the fact that polyP is involved in several cellular functions, makes it difficult to assert what the role of polyP is . PolyP could buffer Ca2+; however, its low concentration in mitochondria makes this possibility unlikely unless there is a localized polyP distribution. It has also been suggested that polyP–Ca2+ complexes form aggregates that could be incorporated into the membrane, leading to an overall non-specific increase in membrane permeability. This is seen in mammalian cell membranes during calcium phosphate–DNA transfection . However, the most commonly accepted idea is that polyP could physically participate in the formation of the ion-conducting part of mPTP .
Intracellular polyP is known to activate TRPA1 through electrostatic interactions, probably by maintaining a proper active conformation for channel gating . Charge alone is not sufficient for this function, which suggests that polyP structure also plays a crucial role . In the case of TRPM8, the nature of the interaction with polyP seems to be different from that of TRPA1 since polyP is co-purified with TRPM8 and also associated with PHB . In these complexes, the binding to PHB is covalent, whereas binding to polyP is not; rather, polyP function is usually associated with ion selectivity and transport .
In bacteria, polyP can be used as an energy source. Bacterial genomes encode several polyP-dependent kinases that utilize it as phosphoryl donor. Recently, in mammalian cells, the newly identified human mitochondrial NAD kinase was also shown to use not only ATP, but also polyP . Further evidence of the functional role of polyP in energetic metabolism is the fact that polyP significantly affects mitochondrial respiration under normal physiological conditions . Moreover, polyP levels were shown to be highly dynamic and directly dependent on the metabolic state of mitochondria . This suggests the existence of a feedback mechanism of regulation between the two . The mechanism by which polyP affects mitochondria respiration is not yet understood. However, as for NAD kinase, polyP might serve as an independent source of ATP affecting the activity of ATP synthase and respiratory chain . Other possibilities are that polyP affects Ca2+-buffering properties, as mentioned above, and consequently the regulation of mitochondrial enzymes. Alternatively, polyP might bind and regulate these enzymes directly .
The existence of a direct link between the cellular levels of polyP and inositol pyrophosphates [14,29] highlight further roles for polyP in regulating energetic metabolism. Inositol pyrophosphates are a eukaryote-ubiquitous class of molecules which are central regulators of energetic metabolism (for reviews, see [30–32]). The absence of inositol pyrophosphates from yeast leads to enhanced glycolysis, poor mitochondrial functionality and an overall 3-fold increase in ATP levels . The main enzyme responsible for inositol pyrophosphate synthesis in yeast is Kcs1 [homologous with mammalian IP6K (inositol hexakisphosphate kinase)]. The strain kcs1Δ, besides not possessing inositol pyrophosphates, possesses much reduced levels of polyP  (Figure 2). The molecular mechanism that could explain this relationship is unknown, but pyrophosphates (PPi) have been shown to accelerate the VTC4-dependent polyP polymerase reaction . Therefore it is possible that the pyrophosphate moiety of inositol pyrophosphates could act as the catalyst for this reaction . Alternatively, inositol pyrophosphates could have a protective ‘capping’ effect against the hydrolysis of polyP by PPXs through their covalent attachment to the extremities of polyP . Therefore inositol pyrophosphates might regulate polyP metabolism and, as such, control the availability of cellular free phosphate that is ultimately required for mitochondrial ATP synthesis.
PolyP is able to control gene expression in micro-organisms . In myeloma cells, which are malignant transformation of PCs (plasma cells), intracellular polyP levels are highly increased compared with untransformed PCs. This accumulation is primarily in the nucleolus and specifically in the nucleolar dense fibrillar component and in the fibrillar centre areas . In these areas, transcription levels are low and pre-rRNA transcripts are spliced and modified. Upon transcription inhibition, total and nucleolar polyP decreases with a parallel appearance of polyP-rich structures in the cytoplasm. Moreover, exogenously applied polyP decreases RNA polymerase I activity . Once again, the mechanism by which polyP exerts this function is unknown, but it is possible that polyP regulates RNA pol I directly or regulates factors that are required during transcription.
The polymeric nature of polyP and its capability of co-ordinating cations and water molecules predict that, in cellular regions where this polymer concentration is high, polyP might acquire a hydrogel-like structure. It is possible to envisage this occurring in yeast vacuoles where polyP accumulates; across membranes, where water is already limited; and even in nuclear pores, where the volume is tightly defined.
The ultrastructure of the yeast vacuole is highly dependent on the amount of polyP present. S. cerevisiae wild-type vacuoles are comparatively more electron-dense than those from kcs1Δ yeast, where polyP is reduced or vtc4Δ, from where polyP is absent (C. Azevedo and A. Saiardi, unpublished work) (Figure 3). This could reflect the difference in density of a hydrogel-like structure (Figure 3).
Ultrastructure of S. cerevisiae cells
It is likely that not only the polyP concentration but also polyP's polymeric length will determine its local hydrogel-like status. Therefore polyP local metabolism might control cytosolic molecular fluxes. We speculate that this regulation can be of particular importance in the context of nucleocytoplasmic transport, where the fluxes through the limited space of the nuclear pore can be controlled by polyP hydrogel phase transition.
In higher eukaryotes, polyP has been implicated in many vital aspects of cellular physiology, but, to date, the majority of the data generated has been based on circumstantial evidence for which the molecular mechanisms are unknown. One essential piece of the puzzle that is still missing is the identification of the higher eukaryotic enzyme(s) responsible for polyP synthesis. Once identified, it will be possible to understand in greater detail how the synthesis of these molecules is regulated, which will probably lead to many other important polyP functions being uncovered. Moreover, the development of more specific and easy to perform techniques that will allow a straightforward and reliable polyP detection and quantification will facilitate and increase the expansion of this exciting research field.
Signalling 2013: from Structure to Function: A Biochemical Society Focused Meeting held at the University of York, U.K., 17–19 June 2013. Organized and Edited by Nicholas Brindle (University of Leicester, U.K.), Sandip Patel (University College London, U.K.) and Stephen Yarwood (University of Glasgow, U.K.)
Dictyostelium discoideum PPK2
inositol hexakisphosphate kinase
mitochondria permeability transition pore
transient receptor potential ankyrin 1
transient receptor potential melastatin 8
vacuolar transporter chaperone 4
We thank the members of the Saiardi laboratory and, in particular, Tom Livermore for reading the paper and thoughtful comments before submission, and Jemima Burden for the technical assistance with the EM.
This work was supported by Medical Research Council (MRC) funding to the Cell Biology Unit.