Developmental accumulation of inorganic polyphosphate affects germination and energetic metabolism in Dictyostelium discoideum

Significance

The most basic biological polymer is a linear chain of linked phosphate groups, simply called inorganic polyphosphate (polyP). By ablating polyP synthesis in the amoebae, Dictyostelium discoideum, we discovered that polyP regulates basic metabolism, general fitness, and spore germination. We also discovered a massive increase in the level of polyP during developmental progression. Interestingly this accumulation is linked to the accumulation of another family of phosphate-rich molecules, the inositol pyrophosphates. Thus, by using a genetic model, we reveal a link between polyP, ATP production, and inositol polyphosphates.

Abstract

Inorganic polyphosphate (polyP) is composed of linear chains of phosphate groups linked by high-energy phosphoanhydride bonds. However, this simple, ubiquitous molecule remains poorly understood. The use of nonstandardized analytical methods has contributed to this lack of clarity. By using improved polyacrylamide gel electrophoresis we were able to visualize polyP extracted from Dictyostelium discoideum. We established that polyP is undetectable in cells lacking the polyphosphate kinase (DdPpk1). Generation of this ppk1 null strain revealed that polyP is important for the general fitness of the amoebae with the mutant strain displaying a substantial growth defect. We discovered an unprecedented accumulation of polyP during the developmental program, with polyP increasing more than 100-fold. The failure of ppk1 spores to accumulate polyP results in a germination defect. These phenotypes are underpinned by the ability of polyP to regulate basic energetic metabolism, demonstrated by a 2.5-fold decrease in the level of ATP in vegetative ppk1. Finally, the lack of polyP during the development of ppk1 mutant cells is partially offset by an increase of both ATP and inositol pyrophosphates, evidence for a model in which there is a functional interplay between inositol pyrophosphates, ATP, and polyP.

The simplest biological polymer is inorganic polyphosphate (polyP), which consists of a linear chain of phosphates linked by high-energy phosphoanhydride bonds (1). Regardless of its origin, polyP is ubiquitously present in today’s living organisms. The synthesis, enzymology, and role of polyP are best characterized in the bacteria Escherichia coli (2). However, polyP is also present in archaea (3), eukaryotes (4), and human cells (5). Whilst polyP is present in virtually all cellular compartments (6), it accumulates in acidocalcisomes, organelles common to bacteria and mammalian cells (7).

PolyP has been implicated in a range of cellular processes over recent years, but its most basic proposed function is to act as a buffer for the level of free phosphate (8). Phosphate buffering is essential for general metabolism, and the synthesis or degradation of polyP is able to safeguard the level of free phosphate in cells. In addition, the polyanionic nature of polyP makes it a strong chelator of bivalent cations; thus polyP plays a key role in cation homeostasis. The chelation of calcium by polyP might even play a role as a calcium sink, regulating calcium signaling (8). Besides these common roles, many specific functions have been attributed to polyP in eukaryotes (see reviews in refs. 9–11). Recently its ability to function as a protein-folding chaperone (12) and to drive a newly discovered posttranslational protein modification, protein polyphosphorylation (13), have provided further insight into the mechanism of action of this simple polymer.

Little is known about the synthesis or metabolism of polyP in eukaryotes. The yeast Saccharomyces cerevisiae is the only eukaryotic model where polyP physiology has been properly characterized by a genetic approach. In yeast, polyP synthesis is carried out by a subunit of the vacuolar transporter chaperone (VTC), Vtc4 (14). The intact VTC complex is required for transport of polyP into the vacuole lumen (15), leading to the accumulation of polyP in the vacuole. Both exo- and endo polyphosphatases have also been identified in yeast (16, 17). Furthermore, the yeast model has revealed a metabolic link between polyP and the phosphate-rich inositol pyrophosphates. In yeast, the levels of inositol pyrophosphates closely correlate with the level of polyP, and yeast strains lacking inositol pyrophosphates possess almost no polyP (17).

PolyP has also been documented in the social amoeba Dictyostelium discoideum. Its presence in amoebae was first recognized by biochemical extraction in the early 1970s (18). Subsequently, in the late 1980s, the use of ³¹P-NMR analysis revealed that both InsP₆ and polyP are present in Dictyostelium spores (19). Besides yeast, the social amoeba is the only other eukaryote where the enzymology of polyP synthesis has been studied. Kornberg’s laboratory described two enzymes able to synthesize polyP in D. discoideum. The first, DdPPK1 (hereafter refer as Ppk1), was identified by homology to the bacterial polyphosphate kinase (PPK) sequence (20). Dictyostelids represent the only eukaryote class possessing a bacterial-like PPK, suggesting the gene was obtained by a horizontal gene transfer event (20). The second protein complex, named DdPPK2, was identified after biochemical purification of polyphosphate kinase activity and was found to comprise one actin-related protein and two actin-associated proteins (21). Whereas the in vitro biochemical properties of these two enzymes have been studied (20, 22), little is known about D. discoideum polyP metabolism and the contribution of these proteins to in vivo polyP synthesis, or the physiological function of polyP. Therefore, we have used a recently refined biochemical procedure to study polyP metabolism in D. discoideum. Our analysis failed to detect the presence of polyP in cells lacking the bacterial-like Ppk1. Importantly, we observed a dramatic increase of polyP during amoebae development with an accumulation in the spore ultimately affecting germination. These defects are due to the influence of polyP on the energetic status of the cell and inositol pyrophosphate metabolism.

Results

D. discoideum Possesses polyP.

Several polyP detection methods lack specificity (23). Furthermore quantification of polyP using traditional biochemical assays, such as the PPK assay (8) or measurement of free phosphate (Pi) after polyP hydrolysis, are greatly impaired by the inability to extract polyP without copurifying other phosphate rich molecules (see below). Therefore, the only approach to unambiguously establish the presence of polyP in cells is its visual detection once resolved by polyacrylamide gel electrophoresis (PAGE) (24). Furthermore the ability of the phosphoanhydride bonds of polyP to induce photobleaching of DAPI has improved the sensitivity of this method over traditional Toluidine staining (25). We extracted polyP from vegetative growing D. discoideum AX2 cells using phenol/chloroform, which also extracts RNA and inositol pyrophosphates, and resolved 70 μg of RNA by 20% PAGE. DAPI staining revealed the negatively stained smear typical of polyP-DAPI photobleaching in cell extracts (Fig. 1).

Fig. 1.

D. discoideum polyP detected by PAGE analysis. Synthetic polyP standards with an average chain length of 25 (T25) and 65 (T65) were loaded, either untreated or after incubation with the exo- and endopolyphosphatase Ppx1 and Ddp1. Equal amounts of vegetative AX2 D. discoideum phenol chloroform extract, as normalized RNA (70 μg) were treated with DNase, RNase, exo-, and endopolyphosphatase (Ppx1 and Ddp1) or treated in acid at high temperature (acid boiled). Samples were analyzed by 20% PAGE and stained with DAPI. Phenol extract from wild-isolate NC4 cells (Right lane) revealed the same type of polyP. The negative stained smears depend on the polyP property to induce DAPI photobleaching (25). The signal disappears after Ppx1 and Ddp1 incubation or acid treatment, whereas the white fluorescence reveals RNA species degraded by RNase treatment. The results shown are of a representative experiment that was repeated more than three times.

Our extraction procedure copurifies polyP and nucleic acids, mainly RNA; thus to be confident of the nature of the polyP negative smear, we subjected the extract to DNase and RNase treatment. Whereas RNase treatment removed the positive fluorescence visible at the top of the gel (Fig. 1), neither DNase nor RNase affected the negative staining. To test whether polyP was the constituent of this negative staining, we treated the samples with the endopolyphosphatase Ddp1 (17) and the exopolyphosphatase Ppx1 (16). This enzymatic treatment removed the negative smear generated by polyP standards as well as that present in cell extracts (Fig. 1). We further confirmed the smear to be due to polyP by treating with acid at high temperature (acid boiled), because phosphoanhydride bonds are rapidly degraded by these conditions. These analyses demonstrated the presence of polyP in vegetative D. discoideum, not only from the standard laboratory strain AX2, but also in the natural isolate, NC4 (Fig. 1). By comparing the migration of D. discoideum polyP with polyP standards, we estimate that amoebae possess polyP with an average length of ∼50 phosphate residues.

Ppk1 the Only Enzyme Responsible for polyP Synthesis Control General Fitness and ATP Levels.

The characterization of two D. discoideum enzymes Ppk1 (20) and DdPPK2 (22) able to synthesize polyP in vitro suggested that both enzymes might contribute to the synthesis of polyP in vivo. DdPPK2 has been identified as an actin-related protein complex but the specific gene was not cloned. Given that the D. discoideum genome encodes 41 actins and actin-related proteins (26) it was impractical to systematically delete all of these genes. Because only one bacterial-like Ppk1 is present in the Dictyostelium genome, we deleted, by homologous recombination, 69% of the ORF of this gene (ID: DDB_G0293524) as confirmed by Southern blots (Fig. 2 A and B). We referred to the Ppk1 null strain as ppk1. Two independent clones were generated and used in our experiments. The homology between the 688 aa of E. coli PPK1 and the 1,050 aa of D. discoideum Ppk1 is restricted to the C terminus; therefore, our knockout approach deleted >95% of the Ppk1 catalytic domain (Fig. 2A). We first analyzed the general fitness of ppk1 cells by performing a growth assay in rich HL5 medium (Fig. 2C). We observed that ppk1 mutant cells have a growth defect, with a mean generation time of 15 h, compared with 11 h for wild type (WT), demonstrating the importance of Ppk1 for D. discoideum physiology.

Fig. 2.

Ppk1 is the only polyphosphate kinase in D. discoideum affecting growth and basic metabolism. (A) Diagram showing the homologous recombination knockout strategy to generate ppk1 mutant cells. (B) Southern blot of WT and ppk1 genomic DNA. Genomic DNA was digested with BsrGI and probed for the 5′ region, as indicated in A. (C) Cell growth curves in HL5 media. WT, black line; ppk1, dashed line. Average ± SD of three experiments were run in duplicate (*P < 0.05). (D) Phenol chloroform extracts, normalized by loading 90 μg of copurified RNA, from vegetative growing WT and ppk1 cells were analyzed by 20% PAGE and stained with DAPI. The figure shows the result of representative experiments that were repeated more than three times. (E) Quantification of ATP levels in vegetative growing WT and ppk1 amoebae. Average ± SD of three experiments were run in duplicate (**P < 0.01). Phenol chloroform extracts, normalized by loading 65 μg of copurified RNA, from WT (F) and ppk1 (H) cells treated with KCN and BHAM for the indicated times were analyzed by 20% PAGE and stained with DAPI. The results shown are of representative experiments that were repeated more than three times. Parallel ATP level determination, normalized to untreated control (time 0), of KCN- and BHAM-treated WT (G) and ppk1 (I) culture. The data represent the average ± SD of three experiments run in duplicate (*P < 0.05, **P < 0.01).

We next analyzed polyP content by resolving 90 μg of RNA/polyP extract from WT and ppk1 cells by PAGE, revealing the complete loss of the polyP negative stain in ppk1 (Fig. 2D). This analysis demonstrates that Ppk1 is in fact the only enzyme responsible for polyP synthesis in D. discoideum cells. Previous analysis of the level of polyP in ppk1 cells reported that polyP was merely reduced to 20–50% of the WT level (27). There are two explanations for this discrepancy: The first is the use of indirect assays to determine polyP levels. These assays developed using synthetic polyP fail to account for the effect of copurifying interfering molecules. PolyP extraction not only purifies polyP and RNA but also inositol phosphates (Fig. 2D), nucleotides (28), and likely free phosphate. The abundance of these phosphate-containing molecules influences these polyP quantification methods (8). The second explanation is that the previously generated ppk1 line deleted only 114 nt of 3,153 nt, just ∼3% of the Ppk1 ORF (27) and that this mutant retains active Ppk1 enzyme. This hypothesis is supported by the identification of PPK activity in this mutant (27).

Because polyP has been associated with primary metabolism (29) and that any change to Pi buffering seems likely to affect basic metabolism, we decided to investigate whether the level of ATP was affected. In fact, ppk1 cells displayed a 2.5-fold reduction in ATP, offering a simple explanation for the overall fitness defect of these cells (Fig. 2E). We next analyzed the level of polyP after inhibiting both the cytochrome mediated respiration (with KCN) and the alternative oxidase (AOX) pathway with benzohydroxamate (BHAM) (30). Sublethal doses of these drugs induced a ∼40% decrease in the level of ATP in WT amoebae (Fig. 2G), but did not kill the cells during the 2-h time course. During this time course, we also observed an unexpected increase in polyP level (Fig. 2F). Conversely, ppk1 cells did not display any accumulation of polyP nor change in ATP levels (Fig. 2 H and I). Meanwhile, the levels of inositol pyrophosphate were unaffected by these treatments (Fig. 2 F and H).

polyP Levels Increase During D. discoideum Development.

D. discoideum was originally selected as an experimental model to study the transition to multicellularity. Single cells undergo a defined developmental process through a multicellular “slug,” culminating in a spore-containing fruiting body. We decided to investigate polyP metabolism during the substantial physiological changes associated with D. discoideum development. Because the starvation response induces D. discoideum development, we shifted a vegetative growing culture from rich HL5 medium to nonnutrient (phosphate buffer) agar and collected cells at different time points corresponding to the diverse developmental stages. PAGE analysis of phenol extract from WT amoebae revealed an impressive accumulation of polyP during development (Fig. 3A). Densitometry scanning of Toluidine-stained gels (Fig. 3B), which although less sensitive, offers a better dynamic range (31), showed the increase in polyP between vegetative stage and fruiting body was more than 100-fold. The average size of polymers did not substantially change during this accumulation (Fig. 3 A and B). Analysis of ppk1 amoebae confirmed that Ppk1 is the enzyme responsible for this accumulation of polyP (Fig. 3A). Analysis of the wild-isolate NC4 strain confirmed the striking accumulation of polyP during development (Fig. 3C). This accumulation of polyP during the developmental progression is coherent with the increased expression of the Ppk1 gene during the late stages of development as reported by the DictyExpress database (https://dictyexpress.research.bcm.edu/landing/).

Fig. 3.

D. discoideum accumulate polyP during development. (A) WT and ppk1 cells were developed on nutrient-free agar and harvested at five time points. PolyP was extracted and 15 μg of RNA was analyzed by 20% PAGE and stained with DAPI. (B) A total of 70 μg of RNA WT cell extracts from five developmental time points was analyzed by 20% PAGE stained with Toluidine blue and analyzed by densitometry using ImageJ. (C) PolyP was extracted from NC4 cells at three time points during development, analyzed by 20% PAGE, and stained with DAPI. The results shown are of representative experiments that were repeated more than three times.

Ppk1 Is Required for Spore Germination.

We did not observe any developmental phase delay between WT and ppk1 cells. The number of fruiting bodies formed was unaffected in the ppk1 background. The fruiting bodies of the mutant were correctly formed with identifiable basal disks, stalks, and spore heads. However, the overall size of the fruiting bodies was considerably reduced (Fig. 4A). Because fruiting bodies consist of two cell types, we decided to investigate whether polyP accumulates in spores or stalk cells. We isolated pure preparations of spore cells and compared polyP levels extracted from spores to those extracted from whole fruiting bodies. We normalized between these samples by spore number and extracted from 5 × 10⁶ spores in each case. This analysis revealed that a large proportion of polyP accumulates in the spore (Fig. 4B).

Fig. 4.

ppk1 cells show impaired fruiting body formation and spore germination. (A) Representative images of fruiting bodies from WT and ppk1 cells. Images were taken at the same magnification. (B) Analysis by 20% PAGE of polyP extracted from whole fruiting bodies and pure spore preps, stained with DAPI. Time courses of spore germination in HL5 media (C) and HL5 media supplemented with 0.2 M sorbitol (D) over 24 h. WT black line; ppk1 dashed line. The date represent the average ± SD of three experiments run in duplicate (*P < 0.05).

The storage of polyP in the spores suggested an important role for this polymer during spore germination, potentially acting as a store of phosphate, minerals, or energy. In fact, analysis of spore germination in rich HL5 medium revealed a germination delay in the mutant strain (Fig. 4C). The polymeric nature of polyP also offers clear osmotic benefit due to its ability to complex with counterions. Thus, polyP synthesis/degradation can also buffer osmotic pressure by capturing/realizing Pi and cations. We investigated the effect of osmotic stress on germination and found that the defect observed between WT and ppk1 cells was amplified in media supplemented with 0.2 M sorbitol (Fig. 4D).

Altered ATP and Accumulation of Inositol Pyrophosphates in ppk1 Mutant Spores.

In yeast, the level of polyP is metabolically connected to the level of inositol pyrophosphates (17). These molecules (InsP₇ and InsP₈) are notable for their ability to control cellular ATP synthesis (32, 33). Because polyP influences the level of ATP in vegetatively growing cells (Fig. 2E), we wondered if the massive absence of polyP (and therefore phosphoanhydride bonds) during development of ppk1 cells (Fig. 3A) might affect inositol pyrophosphate synthesis or ATP production.

The analysis of inositol phosphates extracted from WT and ppk1 cells revealed that the ratio of InsP₇ and InsP₈ to their precursor InsP₆ is significantly altered (Fig. 5A). The quantification of several experiments revealed that ppk1 fruiting bodies accumulated more than twice the level of inositol pyrophosphates compared with WT (Fig. 5B). Although we were technically unable to quantify ATP in the later stages of D. discoideum development, we observed that during the initial phases of development ATP level in WT cells remained largely unchanged. Meanwhile, the level of ATP in ppk1 cells was substantially increased (Fig. 5C) and reached levels slightly exceeding those of WT amoebae. These results indicated that when polyP cannot be synthesized, as in ppk1, cells accumulated inositol pyrophosphates InsP₇ and InsP₈. This perhaps represents a partial compensation, in which energy is stored in phosphoanhydride bonds of inositol pyrophosphates rather than polyP. It is clear that the additional energy stored in the phosphoanhydride bonds of InsP₇ and InsP₈ in ppk1 will not be equal to that accumulated in WT polyP. However, these results do confirm a metabolic interplay between these two families of phosphate-rich molecules and their ability to regulate or be regulated by cellular energy metabolism.

Fig. 5.

Developmental failure to accumulate polyP affects inositol pyrophosphate and ATP metabolism. (A) Analysis by 35% PAGE of perchloric acid extract from WT and ppk1 cells during development, stained with Toluidine blue. (B) Quantification of the ratio between inositol pyrophosphates (InsP₇ and InsP₈) and their precursor, InsP₆ in fruiting bodies. Quantification by densitometry using ImageJ software, results represent the average ± SD of three independent experiments (***P < 0.0001). (C) Quantification of fold change in ATP between vegetative cells and aggregating cells starved for 9 h. The data represent the average ± SD of six experiments run in duplicate (*P < 0.05).

Discussion

The ability of polyP to covalently modify proteins (13) and to work as a chaperone (12) have recently boosted interest in this polymer. Certainly polyP is no longer the “forgotten polymer” (34) it was a few years ago. The study of polyP has long been inhibited by the lack of reliable techniques, a fact not lost on Kornberg et al. who commented on “the inadequacy of methods to establish the authenticity and size of polyP” (8). Here we have coupled high-quality PAGE analysis with the sensitivity and specificity of DAPI staining to reliably detect polyP in D. discoideum. Furthermore, we were unable to detect any polyP in cells lacking Ppk1. Earlier work described the involvement of Ppk1 in fruiting body development as well as an effect on the predation behavior of the amoebae (27). In this study, we generated a ppk1 mutant amoeba by deleting ∼70% of the Ppk1 ORF, revealing many previously unrecognized phenotypes.

A modest developmental increase of polyP was previously noted (27); however, our analysis has revealed an unprecedented >100-fold increase in the level of polyP during D. discoideum development. The astonishing polyP increase and accumulation in the spore is reminiscent of InsP₆ (phytic acid) accumulation in plant seeds (35). These phosphate-rich molecules share similar biophysical characteristics, including the ability to chelate divalent cations. Therefore, as InsP₆ is important to supply phosphate and cations during plant seed germination (35), polyP might play similar roles during spore germination as demonstrated by the reduction in germination efficiency in ppk1 cells.

Because polyP can act as a Pi donor and buffer, it follows that polyP might regulate ATP metabolism. In yeast, a genome-wide screen revealed the interdependence of polyP with primary metabolism (29). In mammalian cells, alteration of mitochondrial metabolism affects polyP production (36). It has also been suggested that polyP is an activator or even a constituent of the mitochondrial permeability transition pore (37). Using our clean genetic system, we demonstrated the interdependence between polyP and cellular energetic metabolism, through the substantial decrease in cellular ATP in vegetative ppk1 cells. The unexpected increase in the level of polyP after treatment with mitochondrial poisons KCN and BHAM, revealed a further linkage between polyP and energetic metabolism. Nevertheless, these results do appear counterintuitive as these drugs combine to decrease ATP levels. It is possible that our treatment induced a starvation-like response and thus an increase in polyP. However, we did not observe any significant change in ATP level during starvation, and mitochondrial functionality actually increases during development (38). An alternative explanation could be that an increase of polyP in cells with inhibited mitochondria might represent some compensatory mechanism. In bacteria, polyP complexes with β-hydroxybutyrate (PHB) to form calcium selective ion channel (39). This channel induces conductance states similar to those recorded for the mitochondrial permeability transition pore (37). Thus, it has been proposed that polyP plays a key role in defining the mitochondrial permeability transition pore (37). These counterintuitive results invite further work with D. discoideum to define the precise role of polyP in mitochondrial function and the mitochondrial permeability transition pore.

The compensatory increase of inositol pyrophosphates and ATP observed during ppk1 development is supportive of a model in which there is a functional interplay between inositol pyrophosphates, ATP, and polyP. This nascent hypothesis requires further work to be fully validated. However, our characterization of D. discoideum polyP metabolism and the development of ppk1 cells, offer the unique platform to investigate the link between polyP and ATP production. It is noteworthy to remember that inositol pyrophosphates were discovered in D. discoideum (33); thus it is the ideal experimental model to investigate the metabolic connection between these molecules. The yeast vtc4Δ strain and the ppk1 amoebae are the only two eukaryotic models in which a genetic approach has eliminated polyP. Whereas both experimental models offer distinctive advantages, the larger size of amoebae, absence of a cell wall, and a polyP level more similar to mammalian cells, make D. discoideum an excellent model to dissect polyP functions and the ideal model to develop more specific probes for localizing polyP in the cell.

Materials and Methods

Detailed methods are provided in SI Materials and Methods.

Inorganic polyphosphate was extracted from cell pellets at indicated time points. Cells were resuspended in one volume of LETS buffer (100 mM LiCl, 10 mM EDTA, 10 mM Tris⋅HCl, pH 8.0, 0.5% SDS), one volume of acidic phenol (pH 4.0) was added, and samples were vortexed at 4 °C for 5 min. Samples were spun at 17,000 × g for 5 min at 4 °C and aqueous phase was recovered. Two volumes of chloroform were added and samples were again vortexed at 4 °C for 5 min before spinning at 5,000 rpm for 5 min. The aqueous phase was collected and precipitated by adding 2.5 volumes of EtOH and incubating at −80 °C for 1–2 h. After spinning at 17,000 × g for 10 min, pellets were resuspended in 10 mM Tris⋅HCl, pH 7.4, 1 mM EDTA, and 0.1% SDS. RNA concentration was measured by nanodrop and used to normalize samples before loading on PAGE. Inositol pyrophosphates were extracted as described before (31).

SI Materials and Methods

Culture.

D. discoideum lines used were axenic strain AX2, ppk1, and nonaxenic NC4. Axenic strains were cultured in HL5 media in the presence of penicillin and streptomycin (Gibco). Other antibiotics used were, blasticidin (Calbiochem) at a concentration of 10 μg/mL. Axenic cells were grown in HL5 at 22 °C, shaking at 120 rpm, cells were diluted every 2–3 d to ensure they did not surpass 5–6 × 10⁶ cells per milliliter. Growth curves were generated by inoculating cells at 1 × 10⁵ cells per milliliter in HL5 media, supplemented with 0.2 M sorbitol. During development assays, cells were plated on potassium phosphate agar plates [20 mM potassium phosphate, pH 6.2, “KK2” and 2% (vol/vol) agar]. Cells were allowed to develop for between 1 h and 2 d as stated. Clonal selection of mutants and growth of nonaxenic NC4 strains was on Klebsiella aerogenes on Sussman Medium plates.

KCN and BHAM Treatment.

WT and ppk1 cells were grown in shaking culture to a density of 1–2 × 10⁶ cells per milliliter. Cells were then transferred to Petri dishes and allowed to adhere, resulting in dishes at a confluence of ∼50%. Once cells were adherent, media were replaced with fresh HL5, supplemented with 100 μM KCN and 20 μM BHAM. Cells were harvested at indicated time points.

Mutant Construction.

Two regions of Ddppk1 and flanking genomic sequence were amplified by PCR from AX2 genomic DNA. Primer 5F, including a NotI site GCAGCGGCCGCGACAAACCCCCCAGACAATGG and primer 5R, including an EcoRI site GCAGAATTCCATTTGCTAATGGTGGTGCATC, generated a 1,042-bp 5′ sequence. Primer 3F, GCGAAGCTTCGTCGTGTCGAAGTAATGGTTCC, included a HindIII site, whereas primer 3R, GCAGGTACCGGACGCCACCTTTAACTCG, contained a KpnI site and were used to amplify a 925-bp 3′ sequence. The knockout plasmid TMO1-DdPPK1-Bsr was generated by inserting these sequences into the plasmid TMO1 (40), with the 5′ sequence inserted between the NotI and EcoRI sites and the 3′ sequence between the HindIII and KpnI sites. This plasmid was then digested using BssHII and used to transform AX2 cell electroporation. Transformed cells were selected for 8 d in 10 μg/mL of blasticidin. Individual clones were selected by growth on K. aerogenes and screened by PCR. Knockouts were confirmed by Southern blot. We named the newly generated ppk1 knockout strain ASH01.

Southern Blot.

Genomic DNA was purified from AX2 and ppk1 cells. DNA was digested with BsrGI. Digested DNA was run on over 3 h on 0.7% agarose gel and transferred onto Hybond N+ membrane (Amersham) using 10× SSC. Radiolabeled probe was generated using the 5′ arm of the deletion construct and labeling, using Roche HiPrime labeling kit and dCTP32 from Perkin-Elmer.

PAGE.

The 20–35% polyacrylamide gels, as indicated, were prepared using 24 × 16 × 0.1 cm glass plates in 1× TBE. Samples were mixed in with 6× Dye (0.01% Orange G or Bromophenol blue (BBF); 30% glycerol; 10 mM Tris⋅HCl, pH 7.4, 1 mM EDTA). Gels were prerun for 30 min at 300 V and run overnight at 4 °C at 450 V and 3 mA for 20% gels or 700 V and 5 mA for 35% gels. Gels were stained with either DAPI or Toluidine blue as previously described (31). Inositol pyrophosphates were quantified by densitometry using ImageJ software as previously described (31).

Enzymatic Reactions and Acid Hydrolysis.

Phenol extracts were incubated in 45-μL reaction volumes with either DNaseI, RNase A and RNase T1, or Ddp1 and Ppx1 recombinant protein purified as previously described (17). Simultaneous treatment with both exo- (Ppx1) and endo- (Ddp1) polyphosphatases was required to fully degrade extracted polyP. The single use of the exopolyphosphatase Ppx1 often resulted in an incomplete polyP digestion, suggestive of the presence of blocked polyP ends as previously suggested (4). Reactions were conducted in the relevant commercially supplied buffer, RNase buffer (300 mM NaC, 10 mM Tris⋅HCl, pH7.4, 5 mM EDTA) or 1× phosphatase buffer (100 mM KCl, 20 mM Hepes, 6 mM MgSO₄, 1 mM DTT) for 1.5 h at 37 °C. Acid degradation was conducted by incubating samples at 100 °C for 30 min in the presence of 1 M perchloric acid. Acid-treated samples were neutralized by K₂CO₃ before loading.

Spore Preparation.

Cells were developed for 48 h on KK2 agar plates as described above. Fruiting bodies were harvested by scraping cells from plates and then incubated in 0.5% Nonidet P-40 for 3 min, before spinning and washing three times in either KK2 or H₂O. If used for spore germination or whole fruiting body extraction, these cells were pelleted and resuspended in KK2, counted, and either inoculated at indicated spore concentration or used in phenol extraction as described. Pure spore preparations were obtained by passing fruiting body preparations through 114 grade Whatman filter paper after washes to remove stalk cells. Spore cells were then collected and resuspended in KK2, counted, and used for phenol extraction.

Spore Germination.

Spore germination was assayed as previously described (41) by inoculating media, either HL5 or HL5 supplemented with 0.2 M sorbitol, with 1.5 × 10⁶ spores per milliliter. Cells were then counted and scored as either spore, swollen spore, or amoeba at indicated time points.

ATP Quantifications.

ATP was extracted from cell pellets by phenol extraction as previously described; however, after the chloroform step, samples were not ethanol precipitated. Instead, the aqueous phase was collected, diluted 1:10 in water, and immediately used in a luciferase assay using the Molecular Probes ATP Determination Kit.

Statistical Analysis.

Statistical analysis was performed using GraphPad Prism software. Comparison between WT and ppk1 cells was conducted by unpaired t tests of experimental values obtained from three or more independent experiments. Analysis of spore germination compared each time point separately by unpaired t test. Significant differences were indicated as *P < 0.05, **P < 0.01, and ***P < 0.0001.