PPi-dependent phosphofructotransferase (phosphofructokinase) activity in the mollicutes (mycoplasma) Acholeplasma laidlawii

A PPi-dependent phosphofructotransferase (PPi-fructose 6-phosphate 1-phosphotransferase, EC 2.7.1.90) which catalyzes the conversion of fructose 6 phosphate (F-6-P) to fructose 1,6-bisphosphate (F-1, 6-P2) was isolated from a cytoplasmic fraction of Acholeplasma laidlawii B-PG9 and partially purified (430-fold). PPi was required as the phosphate donor. ATP, dATP, CTP, dCTP, GTP, dGTP, UTP, dUTP, ITP, TTP, ADP, or Pi could not substitute for PPi. The PPi-dependent reaction (2.0 mM PPi) was not altered in the presence of any of these nucleotides (2.0 mM) or in the presence of smaller (less than or equal to 300 microM) amounts of fructose 2,6-bisphosphate, (NH4)2SO4, AMP, citrate, GDP, or phosphoenolpyruvate. Mg2+ and a pH of 7.4 were required for maximum activity. The partially purified enzyme in sucrose density gradient experiments had an approximate molecular weight of 74,000 and a sedimentation coefficient of 6.7. A second form of the enzyme (molecular weight, 37,000) was detected, although in relatively smaller amounts, by using Blue Sepharose matrix when performing electrophoresis experiments. The back reaction, F-1, 6-P2 to F-6-P, required Pi; arsenate could substitute for Pi, but not PPi or any other nucleotide tested. The computer-derived kinetic constants (+/- standard deviation) for the reaction in the PPi-driven direction of F-1, 6-P2 were as follows: v, 38.9 +/- 0.48 mM min-1; Ka(PPi), 0.11 +/- 0.04 mM; Kb(F-6-P), 0.65 +/- 0.15 mM; and Kia(PPi), 0.39 +/- 0.11 mM. A. laidlawii B-PG9 required PPi not only for the PPi-phosphofructotransferase reaction which we describe but also for purine nucleoside kinase activity. a dependency unknown in any other organism. In A. laidlawii B-PG9, the PPi requirement may be met by reactions in this organism already known to synthesize PPi (e.g., dUTPase and purine nucleobase phosphoribosyltransferases). In almost all other cells, the conversion of F-6-P to F-1,6-P2 is ATP dependent, and the reaction is generally considered to be the rate-limiting step of glycolysis. The ability of A. laidlawii B-PG9 and one other acholeplasma to use PPi instead of ATP as an energy source may offer these cytochrome-deficient organisms some metabolic advantage and may represent a conserved metabolic remnant of an earlier evolutionary process.

Although reasonably assumed to be present in glucosefermenting mollicutes, F-6-P kinase activity has only been reported in one study, in which it was observed to be * Corresponding author. stimulated by ATP in the presence of NADP and triphenyltetrazolium chloride (M. E. Tourtellotte, Ph.D. thesis, University of Connecticut, Storrs, 1960). The reaction was monitored by formazan production, and no mention was made of the possible role of PP,. The product of the F-6-P kinase reaction is F-1,6-P2, a compound little studied in mollicutes, except by Neimark (16), who has shown that it activates NAD-dependent lactic dehydrogenase and can thereby modulate pyruvate metabolism in these organisms.
Although growth and biosynthesis are considered possible only at an adenylate energy charge (ECA) of >0.8 (10), we found that during log-phase growth, some mollicutes have an ECA of <0.8 (4). We hypothesized that if PPi spared ATP at the F-6-P kinase locus, ECA values might be affected. PP, might be involved in this effect for two reasons: When studying cytoplasmic extracts from some mollicutes, we found that PPi is a product of dUTPase activity (33,34); and unlike any other cell, the nucleoside kinase activity in these organisms is PP1 dependent (30,31). We found that in cytoplasmic extracts of A. laidlawii B-PG9 and Acholeplasma florum LlT, the synthesis of F-1,6-P2 from F-6-P is PP1 dependent. We also report on the partial purification of the PPi-PFP enzyme, its characterization, and its possible role in the metabolism of mollicutes.

MATERIALS AND METHODS
Materials. Tetrasodium [32P21PPi (1.44 Ci/mmol) was purchased from Amersham Corp., Arlington Heights, Ill. Blue Sepharose CL-6B, Phenyl Sepharose CL-6B, and DEAE-Sephacel were purchased from Pharmacia Fine Chemicals, Piscataway, N.J. Coomassie blue protein dye reagent was purchased from Bio-Rad Laboratories, Richmond, Calif. Polyethyleneimine thin-layer chromatography plates containing 0.55 meq of polyethyleneimine g of cellulose-' were obtained from Analtech, Inc., Newark, Del. Each enzyme sample was tested for the presence of contaminating (and interfering) membrane-associated NADH oxidase activity (20). To test for NADH oxidase activity, all reaction components except enzyme and F-6-P were combined in a cuvette. The enzyme was added next, and the reaction was monitored for about 1 min. The PPi-PFP reaction was begun by the addition of F-6-P. NADH oxidase activity was only detected at very low levels in crude hypotonic lysates and in fractions in the early stages of enzyme purification. The PPi-PFP reaction rate was calculated from the change in A340 before and after the addition of F-6-P. Each active enzyme sample was tested at two to five concentrations to determine the maximum specific activity. All assays were conducted on a Gilford 260 spectrophotometer and recorder (Gilford Instrument Laboratories, Inc., Oberlin, Ohio) with cells of 1-cm path length. The temperature was between 21 and 22°C. In some experiments with PP1, we also added either fructose 2,6-bisphosphate (F-2,6-P2) ( The content of the dialysis bag was chromatographed on a Blue Sepharose column (0.9 x 10 cm) equilibrated in TBM buffer. The PPi-PFP activity was eluted in a stepwise fashion, by using 5 mM ATP, then 10 mM tetrasodium PPi, and finally 2 mM KCl in TBM buffer. PPi-PFP activity eluted with 2 mM KCl was used for all characterization studies after dialysis against IBM buffer or TBM buffer. One unit of PPi-PFP was defined as the amount of enzyme which converted 1 ,umol of F-6-P to F-1,6-P2 min-' at 21 to 220C.
After incubation for 30 min at 37°C, the reactions were stopped by heating at 90°C for 2 min. After cooling, samples of each reaction mixture were incubated at 370C for 30 min with 0.3 mM NADH and the following (per milliliter of reaction mixture): 2 U of aldolase, 3 U of triose phosphate isomerase, and 0.3 U of a-glycerol phosphate dehydrogenase. The reaction was stopped by heating at 90°C for 2 min. After cooling, 25 ,ul of enzyme-treated or untreated samples were cochromatographed with nonradioactive samples of F-1,6-P2 and PP1 on polyethyleneimine plates using a solvent which contained 0.25 M LiCl2 and 1 N formic acid. In this system, F-1,6-P2 migrates ahead and clear of tetrasodium PP1. Resolved components were visualized by FeCl3sulfosalicylic acid spray (29) and by radioautography. Areas identified as F-1,6-P2 were scraped off the plates and counted for radioactivity (30).
Other methods. (i) Method 1. Protein was assayed by the Coomassie blue method according to the instructions of the manufacturer, by using bovine serum albumin as the standard.
(ii) Method 2. To estimate enzyme purity, samples of our purified PPi-PFP activity were concentrated by dialysis against kappa buffer (20) containing 35% (wt/vol) sucrose at 4°C. A quantity of 200 p.l of the concentrate containing 7.2 p.g of protein was applied to slab gels and electrophoresed. Slab gels (0.75 mm by 24 cm by 24 cm) were made by the method of Davis (8). Electrophoresis was performed with a Protean II slab gel electrophoresis system (Bio-Rad Laboratories, Richmond, Calif.) equipped with a ECPS 3000/150 power supply and a Volthour Integrator VH-1 (Pharmacia, Inc., Piscataway, N.J.). Samples were electrophoresed initially for 30 min at 150 V gel-1, then by 100 V gel-l, until 900 V * h of electrophoresis was obtained (approximately 5 h). Gels were stained by using the KODAVUE Electrophoresis Visualization Kit (Eastman Kodak Co., Rochester, N.Y.) according to the instructions of the manufacturer. By using the high-molecular-weight protein standards kit (Pharmacia), we determined that in our system, the lower limit of protein detection was 10 ng per band. (iii) Method  Approximately 100 jig of protein was layered on each gradient. After centrifugation, each gradient was removed from the top of the tube with an Auto Densi-Flow IIC (Buchler Instruments Div., Nuclear-Chicago Corp., Fort Lee, N.J.). Fractions (0.4 ml each) were collected and were assayed for protein or PPi-PFP activity as described above.
Standards (Sigma Chemical Co., St. Louis, Mo.) were P-amylase (sweet potato, Mr = 200,000), alcohol dehydrogenase (bakers' yeast, Mr = 141,000), creatine phosphokinase (rabbit muscle, Mr = 81,000), bovine serum albumin (Mr = 67,000), and ovalbumin (chicken, Mr = 45,000). An estimate of the sedimentation coefficient was calculated by the method of Martin and Ames (15). We also estimated the moleculat weight by polyacrylamide gel electrophoresis by using the conditions and standards described for estimation of enzyme purity.
(vi) Method 6. To calculate the Km and Vinax of the reaction, we determined the reaction velocity and moles of F-1,6-P2 synthesized minute-1 milligram-' at various concentrations of PP, and F-6-P. In  Kja(pp,), 0.39 ± 0.11 mM. In the same companion series of experiments, also described in the text, the effect of PPi concentrations on the initial velocities of the PP,-PFP reaction at three concentrations of F-6-P was determined (data not shown). The combined data from these experiments were used to obtain the kinetic derived constants. each F-6-P level over the range of 0.005 to 0.30 mM. All reaction conditions were otherwise identical, and each assay contained 0.25 ,ug of the same batch of partially purified (430-fold) PPi-PFP. The Mg2+ concentration was held constant at 6.09 mM, and no allowance was made for bound or free Mg2+ (5). In initial calculations, reciprocal substrate concentrations were plotted versus the reciprocal values of experimentally determined initial velocities; all plots were linear. The same initial velocities were fitted to the intersecting initial velocity equation of Cleland (7) by using Fortran program. The equation for these initial velocity studies in the direction of F-1,6-P2 formation (5) was: v = VmaxABl(KiaKb + KbA + KaB + AB), where v is the observed velocity, A and B are substrate concentrations, Ka and Kb are the apparent Michaelis constants, and Kia is the dissociation constant of the first enzyme-substrate A complex in a sequential mechanism. We plotted the computer-derived best-fit reciprocal values of initial velocities fitted to the above equation versus reciprocal substrate concentrations of F-6-P at three different concentrations of PP, (Fig. 1). The computer-derived kinetic constants are noted in Results. 180d Cells were harvested from 2 liters of growth medium (approximately 2 x 1012 cells). b Data from one of four experiments performed (four separate batches of crude lysate). The average recovery from all four expdriments, relative to DEAE-Sephacel (I), was 3 ± 2.2% (n = 4; range, 1 to 6%) (average recovery of 6% for this experimento). ' Micromoles of F-1,6-P2 synthesized minute-' milligram of protein-' in standard assay. d Purification was 180-fold, and the specific activity was 12 p.mol of F-1,6-P2 synthesized min-' mg-'. The average purification was 302 + 107-fold (n = 4; range, 180 to 430-fold), and the average specific activity was 13.2 ± 0.91 (n = 4; range, 13.7 to 15.5 ,umol of F-1,6-P, synthesized min-' mg-1). the supernatant after ultracentrifugation and overnight dialysis; it apparently contains an inhibitor of PPi-PFP activity which is removed by chromatography on DEAE-Sephacel, since the enzyme recovery is higher after passage through the matrix (Table 1). We therefore calculated two recovery patterns, one with the total activity in the crude lysate representing 100% recovery and the other with the total activity in the eluate of passage 1 through DEAE-Sephacel representing 100% recovery. With this second alternative, we calculated an average recovery of ca. 3% with approximately 300-fold average purification.

Purification of
In most experiments, we found that there were two peaks of PPi-PFP activity elutable from columns of Blue Sepharose matrix. One peak was apparently not retarded by the matrix and was always found in the wash buffer. This nonretarded activity accounted for approximately 5 to 10% of the total activity recovered from Blue Sepharose. These two forms of PP1-PFP activity were never resolved on other matrices. All our studies were conducted on the larger portion of PPi-PFP activity which was. eluted from the Blue Sepharose after application of 2 M KCI. Estimation of enzyme purity. We detected two visible bands after slab gel electrophoresis and staining of 7.2 ,ug of protein from a 430-fold-purified enzyme preparation. The most prominent band had an Rf of 0.63, corresponding to a molecular weight of 81,000, and the other band had an Rf of 0.80, with a molecular weight of 37,000. Since we are able to detect bands containing 10 ng of protein, we calculated that at least 99% of our preparation was composed of only these two protein species. However, in other radiographic studies of this preparation with [32P21PPi (see Table 3), we found spots identified as triose [32P]phosphates derived from 32Plabeled F-1,6-P2 (data not shown). Therefore, our 430-foldpurified preparation was apparently contaminated with aldolase and perhaps with other enzymes, and our preparations are only partially purified.
pH and metAl requirements. The optimum (maximum specific activity) was at a pH of 7.4. The response over the range of pH 3.49 to 10.3 was that of a smooth hyperbolic curve. Mg2+ was required for activity although Co2+ or Mn2+ could substitute for Mg2+. At a PP, concentration of 2.06 mM, the maximum nanomoles of F-1,6-P2 synthesized minute-1 milligram of protein-1 were: 14.4 (for Mg2+ at 6.1 mM), 9.0 (for Mn2+ at 0.10 mM), and 5.7 (for Co2+ at 0.20 mM). Higher concentrations of Mn2+ or Co2+ were inhibitory. Combinations of these metals at their respective concentrations were neither additive nor gave decreased re-sponses. Ca2+ or Zn2+ could not substitute for Mg2+, Co2+, or Mn2 . Incorporation of EDTA (50 ,uM) into reaction mixtures did not significantly change any recovered value (<5% change).
Specificity of the tetrasodium PP1 requirement. The PFP activity of A. laidlawii B-PG9 required tetrasodium PPi and various deoxyriboand ribonucleotides could not be substituted (Table 2). Furthermore, these nucleotides did not inhibit or enhance the PPi-dependent activity since in the presence of equimolar concentrations of any nucleotide and PPi, there was no appreciable change in activity compared with data from reaction mixtures without nucleotides. Also,

PPi-PFP IN A. LAIDLAWII
Pi interfered with the PP1-dependent reaction ( Table 2). In studies designed to test the purity of the CTP and ATP used, we found (data not shown) by using PP1-PFP from Propionibacterium freudenreichii (Sigma) that our CTP and ATP were contaminated with PP1 (approximately 4 and 0.1 ,ug of PP1 per mg of nucleotide, respectively) (data not shown). We assumed that the low level of activity we detected with CTP and perhaps also with ATP was due to this contaminating PP1. By using the 430-fold-purified preparation, we found that citrate, AMP, GDP, phosphoenolpyruvate, or (NH4)2SO4 did not affect the specific activity, but that F-2,6-P2 (200 ,uM) did, in some experiments, stimulate the reaction by about 5%. Increasing the F-2,6-P2 concentration to 2 mM did not alter our findings. We did not test the effect of F-2,6-P2 on PPi-PFP activity from cells in different growth stages, as was done by Wu et al. with peas (39). Identification of 32P-labeled product F-1,6-P2. We found that the synthesis of putative F-1,6-P2 product from F-6-P and tetrasodium [32P2]PPi was diminished when nonradioactive tetrasodium PPi was added to the reaction mixture (Table 3). We also found that after reaction with aldolase, triose phosphate isomerase, and a-glycerophosphate dehydrogenase, the amount of synthesized 32P-radioactive material that had comigrated with the F-1,6-P2 standard was reduced and that there was a concomitant appearance of 32P-radioactive compounds more mobile than F-1,6-P2.
Molecular weight determination. From our sucrose density gradient studies, we determined that the enzyme which eluted from Blue Sepharose in 2 M KCl has an approximate molecular weight of 74,000 (Fig. 2). The sedimentation coefficient of this material is about 6.7. From our slab gel a Putative 32P-labeled F-1,6-P2 was synthesized as described in the text in the presence of various amounts of nonradioactive PP,. Samples of each of these reaction mixtures were heat stopped and were then treated with a mixture of enzymes which would specifically convert any synthesized 32p_ labeled F-1,6-P2 to trioses. The treated and untreated reaction mixtures were mixed with nonradioactive standards and chromatographed. After chromatographic resolution, areas identified as F-1,6-P2 were removed and counted for radioactivity. The PPi-PFP enzyme preparation used in this experiment was also used in the estimation of enzyme purity as described in the text. electrophoresis studies, we determined that there were two proteins in our preparations. The most abundant form had a molecular weight of 81,000, and the other form had a molecular weight of 37,000 (data not shown). Although we did no enzyme studies with these electrophoretically separated proteins, we assumed that the band at 81,000 molecular weight was the 74,000-molecular-weight PP1-PFP active peak observed in our sucrose density gradient studies and that the 37,000-molecular-weight band was the PPi-PFP active material that did not stick to Blue Sepharose. These results suggested the presence of two forms of the enzyme, presumably a monomer and dimer. Probably the 74,000-molecularweight material, which was the value we used as the most probable approximate molecular weight, was the dimer.
Stability. The partially (430-fold) purified enzyme lost approximately 25% of its specific activity when it was kept in kappa buffer with 10% glycerol (vol/vol) at -22°C for about 6 months. The specific activity of this stored preparation was reduced by about 60% after exposure to 70°C for 2 min and was completely lost after 10 min.
PP,-PFP activity in acholeplasmas. We also found PPi-PFP activity in the crude lysates from three different batches of freudenreichii. By using ATP contaminated with PP1 as a VOL. 165, 1986 phosphorus donor, we found low levels of PFP activity. By using PP1 (Fisher) or the ATP contaminated with PP1, we compared PP1-PFP activity in both crude and partially purified preparations. With either enzyme preparation, the specific activity was about 33.5-fold higher with PP, than with ATP samples. With propionibacteria, Wood and Goss (37) have shown that the enzymes of phosphorylation, aside from glucokinase, are specific for only one phosphate donor or acceptor. Others have shown that either ATP or PP, could serve as phosphorus donors in this reaction (24). Because of the relatively slight activity with ATP, we cannot be absolutely sure that PP, is the only permissible phosphorus donor in the PPi-PFP reaction of A. laidlawii B-PG9.
The absence of any effect, positive or negative, on PP1-PFP activity of low levels of ADP, AMP, GDP, phosphoenolpyruvate, citrate, or ammonium ion, and the absence of any appreciable effect of F-2,6-P2, suggest that either the PPi-PFP activity in A. laidlawii B-PG9 cannot be allosterically modified by thenm or that it is fully activated. Other PP1-PFP or PP1-PFKs, but not all, are generally known to be modified by these compounds (27,32,39). F-2,6-P2 has little or no effect on the PP1-PFK from propionibacteria or from various other cells (6,9,37). Furthermore, the presence of 2 mM of any one of a variety of riboor deoxyribonucleotides and ADP did not affect the velocity of the reaction in the presence of 2 mM PPi. This finding supports our view of both the dependence of the reaction on PP1 and its essential unresponsiveness to other riboand deoxribonucleotides.
Pi was inhibitory to the PP,-dependent reaction ( Table 2); this was expected, since Pi is a product. However, the role of Pi is more complicated, because we found in preliminary experiments that the PP1-PFP reaction was reversible (D. DeSantis and J. D. Pollack, unpublished data). The reverse reaction; (F-1,6-P2 to F-6-P) requires Pi, and PP1, ADP, ATP or other riboor deoxyribonucleotides tested could not substitute. Furthermore, the reverse reaction of the partially purified PP,-PFP from A. laidlawii B-PG9, like the reverse reaction of the PP1-PFP in E. histolytica, can use arsenate instead of P1 (23). This reverse reaction is detected by the reduction of NADP in the presence of rabbit muscle phosphoglucose isomerase and yeast glucose 6-phosphate dehydrogenase. Although our partially purified PPi-PFP had no detectable phosphoglucose isomerase activity, and the reaction was conducted with excess aldolase, triose phosphate isomerase, oa-glycerophosphate dehydrogenase, and NADH (thereby favoring the oxidation of NADH), it is still possible that the forward reaction (F-6-P to F-1,6-P2) was impeded. Therefore, the kinetic data we have presented should be viewed with some reservation, since the effects of any competition for F-1,6-P2 by the reverse reaction on the kinetics we present for the forward reaction velocities are undetermined. What we have attempted to establish is that the presumed rate-limiting step of glycolysis in A. Iaidlawii B-PG9 is essentially or totally PP1 dependent.
Kruger and Dennis (13) report that the auxiliary enzymes commonly used in the assay for PFP or PFK may be contaminated with UDP-glucose pyrophosphorylase and that such contamination can account for apparent activity of PP1-PFP. That is, in the presence of UDP-glucose and PPi, the enzyme catalyzes the synthesis of UTP and glucose 1-phosphate. PFK from rabbit muscle can accept UTP, as well as ATP, as a donor in the phosphorylation of F-6-P to F-1,6-P2. In this case, the requirement for PPi falsely appears to be due to a PP,-dependent phosphofructotransferase, when in fact, PP, is required for the synthesis of UTP which drives a nucleotide-dependent phosphofructokinase. In our studies, there is essentially no activity with UTP or dUTP as phosphorus donors with our partially purified PFP (Table 2). Also, we started the reactions with F-6-P, not with PPi, and then only after we had first determined the rate in the absence of F-6-P. Therefore, even if PP1, UDP-glucose, and UDP-glucose-pyrophosphorylase were all present after our purification steps, our assay procedure would have detected their combined activity. Any oxidation of NADH in the complete reaction mixture minus F-6-P was only observed in crude fractions, and then only in traces. We corrected for this oxidation whenever it was detected. This activity was attributed to contamination of crude fractions with membrane-bound NADE oxidase activity. This activity is localized in the membrane fraction of the nonsterol mollicutes. In the sterol-requiring mollicutes, NADH oxidase activity is localized in the cytoplasm (19). Also, in the presumed absence of UDP-glucose, the conversion of ATP-PFK to PP1-PFP reported in spinach leaf cytosol and rabbit muscle was not considered by us to be applicable (2,13,35).
We searched for PP1-PFP activity in other mollicutes to establish its general presence in this group of microorganisms. We found PP1-PFP activity in crude cytoplasmic fractions of A. florum, but none in Spiroplasma floricola 23-6T or Mycoplasma gallisepticum S6T (J. D. Pollack and M. V. Williams, unpublished data). Our negative findings with the latter two organisms were to some degree expected, because as noted, crude cytoplasmic fractions of the sterolrequiring M. gallisepticum contain NADH oxidase activity (19). This is also true for S. floricola 23-6T (J. D. Pollack, unpublished data). In these cases, since our assay monitors PPi-PFP activity by the oxidation of NADH, crude cytoplasmic extracts from these two organisms, which contain NADH oxidase activity, would inhibit or nullify the test. To determine whether these two mollicutes and other sterolrequiring species have any PP1-PFP activity, a different technique that does not rely on NADH oxidation must be used. The earlier report of ATP-PFK activity in M. gallisepticum (M. E. Tourtellotte, Ph.D. thesis) cannot be verified by our procedure.
The presence of a PP1-PFP implies that in A. laidlawii B-PG9, and possibly in other acholeplasma, glycolysis is dependent on PP1. This may be related to the observation that during growth in modified Edwards medium, some mollicutes in log-phase growth (by viable cell count) have a lower than expected ECA (ranging from 0.69 to 0.76) (3). Karl (10) indicates that biosynthesis and growth are possible only at ECA . 0.8. If PP1 spares ATP, this action may permit log-phase growth with an apparently less than optimal ECA status and may result in relatively high yields of cells (3). In A. laidlawii B-PG9, levels of cellular PP1 may be dependent on the PP1-synthesizing activities of dUTPase (33,34), adenine, and the hypoxanthine-guanine phosphoribosyltransferases (30). PP, may also arise from the activities of polymerase (11), phosphoenolpyruvatecarboxytransphosphorylase, or pyruvate diphosphate dikinase (23,(36)(37)(38). These last two enzymes are generally found in organisms with PP1-PFP activity (36,38). We have detected low levels of phosphoenolpyruvate-carboxytrans- The fact that PP1 has an important role in A. laidlawii B-PG9 metabolism is supported by the additional observation that the purine nucleoside kinase activities in this and other species of acholeplasmas tested is PP, dependent, a requirement which is unknown in any other organism (30,31). We did not find inorganic pyrophosphatase (EC 3.6.1.1) in extracts of A. laidlawii B-PG9 by using tetrasodium [32P2]PPi (V. V. Tryon and J. D. Pollack, unpublished data).
Also, O'Brien et al. did not find any inorganic pyrophosphatase activity in A. laidlawii B-PG9 (17). Although all our studies were with cell extracts, the ability to use PP; rather than ATP may offer some advantage to growing cells. Furthermore, the absence of cytochrome pigments in this group of organisms (21) infers some possible shortage of ATP, perhaps compensated by a sparing action of PPi not only at the F-6-P level of glycolysis but also in the phosphorylation of some purine nucleosides by the action of purine nucleoside kinases. The requirement for PP1 at these loci may represent a conserved remnant of earlier developmental processes, as PPi has been proposed in a study of Desulfotomaculum species to be an evolutionary precursor of ATP (14).