Inorganic Pyrophosphate A Product Of Various Biosynthetic Reactions Biology Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

Inorganic pyrophosphate (PPi) is a product of various biosynthetic reactions, which frequently involve the consumption of ATP (Cooperman, et al., 1992, Heikinheimo, et al., 1996, Terkeltaub, 2001, Yang, et al., 2009). These reactions include, but are not limited to the synthesis of DNA (Fig. 1A), carbohydrates, proteins, and fatty acids. They are frequently near equilibrium reactions - the equilibrium constant is approximately one, and the free energy release is near to zero. Because of this, PPi generated has the ability to drive the reactions in the reverse direction; indeed, adding PPi in vitro does exactly that. PPi must therefore be removed to allow the reactions to continue (Kornberg, 1962, Young, et al., 1998).

Soluble Inorganic pyrophosphatases (EC, also known as PPases, which form part of a larger group of phosphoryl transfer enzymes, are ubiquitous enzymes that catalyse the hydrolysis of PPi into inorganic phosphate ions (Pi) (Fig 1B), 1010 times faster than would occur spontaneously in solution (Cooperman, et al., 1992). This hydrolysis provides a 'thermodynamic pull', not only does it make biosynthetic reactions more favourable, but it drives them forward by removing virtually all of the PPi in the cell, reducing it to almost immeasurably low concentrations (Kornberg, 1962). As biosynthetic reactions are vital to life, it logically follows that PPase activity that drives them is too. The importance of PPase activity to growth and mitochondrial activity has been demonstrated in vivo in Escherichia coli and Saccharomyces cerevisiae respectively (Chen, et al., 1990, Lundin, et al., 1991). It has also been calculated that, without PPases, PPi concentration would rise to around 3 M within one hour, a level that would surely be toxic (Heikinheimo, et al., 1996).

There are two families of soluble Inorganic pyrophosphatases, which have convergently evolved (Merckel, et al., 2001). Family I is found across all kingdoms with its most studied examples being PPases from S. cerevisiae (Y-PPase) and E. coli (E-PPase) (Springs, et al., 1981, Heikinheimo, et al., 1996, Pohjanjoki, et al., 1998, Samygina, et al., 2007, Yang, et al., 2009). Family II PPases are much less common, they have been observed in a small number of prokaryotes - that of Bacillus subtilis was the first observed, although several others have since been studied (Shintani, et al., 1998, Young, et al., 1998, Ahn, et al., 2001). Although the two families have very different structures, they represent a very good example of convergent evolution as their active sites and mechanisms of action are very similar (Merckel, et al., 2001). The active sites of the proteins contain several residues that coordinate three or four divalent metal ions. An array of interactions between active site residues, the metal ions, the fissile PPi, and a number of water molecules result in the precise spatial arrangement and pKas that are required for catalysis to occur. Although catalysis involves several slight conformational changes, the hydrolysis itself involves a single phosphoryl transfer step, without the formation of a phosphorylated enzyme intermediate (Heikinheimo, et al., 1996, Ahn, et al., 2001, Merckel, et al., 2001, Yang, et al., 2009).

Family I

As the name suggests, Family I PPases were identified before Family II PPases, they have been extensively studied for at least half a century. Y-PPase and E-PPase are the most studied and crystallised examples of Family I PPases although PPases from a wide range of organisms from all kingdoms have been studied to various extents. These organisms include, but are far from limited to Homo sapiens (Fairchild & Patejunas, 1999), Bovidae (Yang & Wensel, 1992), Hordeum vulgare (Visser, et al., 1998), Thermoplasma acidophilum (Richter & Schafer, 1992) and Pyrococcus horikoshii (Jeon & Ishikawa, 2005). The current mechanism of family I PPases has been developed over the last two decades, but was largely set out in 1996 by Heikinheimo et al.

Family I PPases are sometimes separated into sub-families, but there is not a great deal of consensus on how they should be divided and how many sub-families there should be. Previously, Prokaryotic Family I PPases were known as type-A, while Eukaryotic PPases were known as type-B (Young, et al., 1998). Sivula et al. noted that, while considerably less well characterised than the other two, plant PPases bear more similarity to prokaryotic PPases than to animal and fungal PPases, so divided them into three sub-families: Ia - prokaryotes, Ib - plants, and Ic - animals and fungi (Sivula, et al., 1999).

While their secondary and tertiary structure is well conserved, the primary structure sequence identity of Family I PPases can be relatively low: E-PPase and Y-PPase have a sequence identity of 27% (Heikinheimo, et al., 1996). The tertiary structure of Family I PPases has been likened to both a cup and a hat (Teplyakov, et al., 1994, Fabrichniy, et al., 2006). This basic shape is a five-stranded β barrel formed by strands β1 and β4-β7, surrounding the β barrel there are two large α helices (αA, αB) smaller strands (β3, β4, β9) and one small helix (αC) (Fig 2). The remaining β-strands and loops adopt a less conserved structure. Eukaryotic PPases have N- and C-terminal extensions not present in prokaryotic homologues, these extensions can be seen as forming the 'brim' of the upturned hat of Ic, removing them produces the more cup-like structure of 1a (Teplyakov, et al., 1994, Heikinheimo, et al., 1996).

Y-PPase forms a homodimer, with one subunit rotated approximately 180° respective to the other around a central point between the two. This is held together by hydrogen bonds between Arg51, Trp52 and Trp279', between His87 and His87' and between Trp279, Arg51' and Trp52'. These residues are largely conserved across all eukaryotic PPases, so it is likely that they all from homodimers in the same way. None of the residues required for dimer formation are part of the active site, this has been shown through mutations of active site residues, which do not affect the quaternary structure (Heikinheimo, et al., 1996).

Prokaryotic PPases form homohexamers, as has been demonstrated with E-PPase and T-PPase (Thermus thermophilus). The quaternary structure can be seen as dimer of trimmers (Sivula, et al., 1999). The trimer is formed by three monomers arranged in a circle rotated roughly 120° from each other. This structure is then flipped, placed underneath the previous trimer, and rotated roughly 30°. The active sites are fairly close to the dimer interface, so mutating Glu20, Tyr55, or Lys141 (E-PPase) not only disrupts catalysis, but also prevents hexamer formation. Other than these residues, there are no conserved residues involved in quaternary structure formation (Teplyakov, et al., 1994). Despite interactions between the subunits, the monomer is still active with very little change in Michaelis constants (Borschik, et al., 1985).

The active site of Family I pyrophosphatases is found near the centre of the β barrel, towards the top of the 'cup' (Fig 2). PPases from different species have been noted as containing between 13 and 17 conserved residues, nearly all of which are polar or charged (Fig 3). The active site binds three or four divalent metal ions, depending on the organism, pH, and metal ion concentration. Halonen et al. stated that four is likely to be more common at physiological concentrations, while Samygina et al. argued that three would be more likely as a fourth metal ion could prevent dissociation of the product. The roles of metal ions within the active site are four fold: they stabilise the large negative charge of the substrate, they orientate various side chains within the active site, they activate the hydroxide ion central to catalysis, and they stabilise the transition state during catalysis (Samygina, et al., 2001).

Magnesium provides the greatest level of activity, but several other divalent metal ions can also function as cofactors (Heikinheimo, et al., 1996). Manganese has been used for a number of crystal structures while manganese and cobalt are useful for kinetic experiments as their rates of reaction are roughly 10 times slower than with magnesium (Halonen, et al., 2002).

Among the conserved residues within the active site, all Family I PPases contain a 115DXDXXDX motif (fig. 3)(Y-PPase numbering) (de Graaf, et al., 2006). This motif is essential for correct metal ion and substrate binding and for pKa modifications that will allow catalysis to occur. Several of the active site residues (Glu58, Tyr93, Asp115, Asp120, Asp147 Asp152, Lys154, Tyr192) and several water molecules co-ordinate the magnesium ions within the active site, while some (Glu48, Lys56, Glu58, Asp71, Arg78, Gly94, Asp117, Lys193) are involved in various non-covalent interactions, mainly hydrogen bonding (Teplyakov, et al., 1994, Heikinheimo, et al., 1996, Yang, et al., 2009). Despite many not being directly involved in the mechanism, each of the active site residues is required for catalysis, even conservative mutations in residues that may appear unimportant can result the pKa of a hydroxide ion involved in catalysis designated 'Wat1' or 'Onu' increasing by up to 3 pH units (Salminen, et al., 1995).

Before the substrate binds to the enzyme, the active site contains two metal ions, the pyrophosphate group (with phosphorus atoms arbitrarily named 'P1' and 'P2') enters the active site already datively bound to the second two metal ions. Binding causes slight conformational changes within the active site to accommodate the size and charges of the PPi and metal ions (Heikinheimo, 2001). The sum of the covalent and non-covalent interactions results in the precise positions of the various moieties, several precise pKas, and a decrease in electron density across the fissile phosphoanhydride bond (Heikinheimo, et al., 1996). The aforementioned hydroxide ion, Wat1, is positioned between M1, M2, Asp120, Asp117, and P2, and is a strong nucleophile (Fig 3). This nucleophilic character is strengthened by formation of a low barrier hydrogen bond to Asp117, something that has been shown to significantly increase the rate of enzymatic catalysis (Cleland & Kreevoy, 1994, Heikinheimo, 2001). Wat1, which can almost be seen as an O2- ion then nucleophilically attacks Pδ+ in P2. This causes the cleavage of the P2-O half of the phosphoanhydride bond, producing two separate phosphate groups, which are then able to diffuse out of the active site. The P1 site has a higher affinity for Pi than the P2 site, meaning that P1 is more tightly bound to the enzyme so P2 leaves first, as the two phosphate groups leave the active site they are replaced by water molecules. The enzyme may then rebind PPi, allowing catalysis to occur once more (Heikinheimo, 2001).

Family I PPases can be inhibited by fluoride and calcium ions. Y-PPase has a Kcat of 214±5 s-1 at pH7.2, or 108±4 s-1 at pH8.5, and a Km of 4.7µM at pH7.0 (Springs, et al., 1981, Pohjanjoki, 2000). Fluoride displaces Wat1 from the active site, meaning that the nucleophilic attack cannot occur. 10mM fluoride reduces the Kcat to 2.5±0.3 s-1 at pH7.2 and 0.86±0.03 s-1 at pH8.5 (Baykov, et al., 2000, Pohjanjoki, 2000). Calcium displaces Magnesium from the active site and inhibits catalysis because it has significantly larger ionic radius than magnesium (0.99Å compared to 0.65Å), this causes several conformational changes within the active site, including causing the DXDXXDX loop to move away from the rest of the active site. This disrupts the network of covalent and non-covalent interactions, resulting in a significant decrease in catalytic activity (Samygina, et al., 2001).

Family II

An inorganic pyrophosphatase in from Bacillus subtilis unlike previously observed PPases was first studied in 1967, although it was not properly investigated until 1998 when its gene was isolated, cloned, and expressed (Shintani, et al., 1998, Young, et al., 1998). The protein was shown to exhibit pyrophosphorolytic activity but bore no significant sequence identity to Family I PPases, it was however, shown to have sequence similarity to proteins of unknown function from several other prokaryotic organisms - including Streptococcus mutans, Methanococcus jannaschii, Archaeoglobus fulgidus, and Streptococcus gordonii. A common observation in some of these organisms was that their genomes did not contain a gene for a Family I PPase, it was therefore decided that they were a second family of convergently evolved PPases (Shintani, et al., 1998, Merckel, et al., 2001). Family II PPases belong to the DHH family of enzymes due to several conserved tertiary structure features and a DHH motif within the active site (Rantanen, et al., 2007).

Curiously, although the first Family II PPase was seen in B. subtilis, the closely related B. stearothermophilus has no homologue but does have a Family I PPase (Shintani, et al., 1998). Family II PPases are present in a wide variety of organisms, but are much less common than Family I PPases. Homology searches have shown that portions of the protein have a somewhat significant level of sequence identity to other proteins but homology across the full structure is very rare. It is therefore unclear how the original gene that the proteins in the family have diverged from was transferred between the species and why it has presumably been selected for in favour of the almost ubiquitous Family I (Merckel, et al., 2001).

As the two families of inorganic pyrophosphatases have convergently evolved, the basic tertiary structure of Family II bears no similarity to that of Family I. It consists of, a 20-22kDa N-terminal domain and a 12-13kDa C-terminal domain; the two domains are joined by a 'hinge', allowing the protein to take on open and closed conformations. The N-terminal domain contains a five or six-stranded parallel β-sheet and seven α-helices, somewhat reminiscent of a Rossmann fold. The hinge is formed by part of helix G and a short loop. The C terminal domain largely consists of a five-stranded mixed β-sheet, but also contains one long α-helix and three shorter ones (Ahn, et al., 2001, Merckel, et al., 2001).

Like eukaryotic Family I PPases, Family II PPases appear to function as dimers, with one monomer rotated roughly 180° around a central point between the two. The interface between the two monomers is formed by the loop between strands four and five and strand five itself, it contains three conserved residues: Thr105, Pro108, and Pro115 (Merckel, et al., 2001). At present, there does not appear to have been any studies into the role of each of these residues in dimer formation.

The active site of Family II inorganic pyrophosphatases, which is located at the interface between the two domains, has evolved a very similar structure and mechanism to that of Family I (Ahn, et al., 2001, Merckel, et al., 2001). Like Family I PPases, the active site contains a large number of conserved charged and polar residues and three or four divalent metal ions; Manganese appears to function as a better cofactor than magnesium, although, again, several different metals are able to facilitate catalysis. Three reasons have been given for the apparent preference for manganese ions over magnesium ions. Firstly, the active site contains two histidine residues, histidine, which is not present in the Family I active site, co-ordinates Mn2+ better than Mg2+, as Mg2+ usually only accepts Oxygen atoms as bond donors. Secondly, the ionic radius of Mn2+ is slightly larger than that of Mg2+, allowing better bidentate co-ordination by carboxylic acids. Thirdly, Manganese acts as a stronger Lewis acid, allowing the pKa of the attacking nucleophile to be lowered further (Merckel, et al., 2001).

Pyrophosphate enters the protein when it is in an open conformation, the protein changes to a closed conformation, possibly caused by the binding of PPi, with the C-terminal domain rotating roughly 90-100° on the hinge. The C-terminal partially enters a concave area on the face of the N-terminal domain, with the interface between the two forming the active site around the substrate (Ahn, et al., 2001). Like Family I, a complex network of covalent and non-covalent interactions between various residues, metal ions, PPi and water, causes precise geometries and pKas within the active site, which allow catalysis to occur (Ahn, et al., 2001, Merckel, et al., 2001).

Again, PPi is nucleophilically attacked by a hydroxide ion, causing its cleavage in the same way as in Family I. The two negatively charged Pi groups repel each other, forcing the protein to return to its open conformation, releasing the two molecules of inorganic phosphate. Family II PPases are not inhibited by fluoride or calcium ions, although the reasons why are unclear, it has be speculated that calcium does not inhibit hydrolysis because the more open active site is not affected by larger ionic radius as the active site of Family I is (Gomez-Garcia, et al., 2004).

Self Incompatibility

Self-incompatibility (SI) is a name given to any mechanism by which plants prevent fertilisation by their own pollen (self-pollen), preventing inbreeding and promoting genetic diversity. There are currently three known mechanisms of self-incompatibility in plants: the sporophytically controlled (SSI) mechanism of Brassicaseae, and the gametophytically controlled (GSI) mechanisms of Solonaceae and Papaveraceae, at least one other is known to exist but is yet to be explored at the molecular level (Franklin-Tong & Franklin, 2003). GSI mechanisms are believed to be more widely employed than SSI mechanisms, being present in around 60-90 families (Takayama & Isogai, 2005).

Each of the currently elucidated SI mechanisms is controlled by a region of the genome known as the S-locus, which produces highly polymorphic male and female S-determinants, proteins that induce the SI response. Should a pollen grain expressing a particular male S-determinant land on a stigma expressing the equivalent female S-determinant, indicating that they may come from the same plant, an SI response is triggered. If the S-determinants are not equivalent, ie. the pollen is compatible non-self pollen from a different plant, then no response is triggered and fertilisation is allowed to occur (Takayama & Isogai, 2005). In all three systems, this response results in the cessation of pollen tube growth, although the processes that bring this about differ greatly (Sobotka, et al., 2000, Roalson & McCubbin, 2003, Bosch & Franklin-Tong, 2008).

In Brassicaseae, the S-locus produces a male S-determinant known as S-locus protein 11 (SP11) or S-locus cysteine rich (SCR), and a female S-determinant known as S-locus receptor kinase (SRK). Both the male and female S-determinants have hypervariable regions, which have been shown to be essential to the induction of the SI response. Should SRK detect an incompatible SP11/SCR, it phosphorylates a protein call ARC1. ARC1 is then believed to promote the ubiquitination and proteasomal degradation of pistil proteins that are required for pollen tube growth (Sobotka, et al., 2000, Takayama & Isogai, 2003, Takayama & Isogai, 2005).

In Solonaceae, the male S-determinant is called S-locus F-box (SLF) or S-haplotype-specific F-box (SFB), while the female S-determinant is a ribonuclease known as S-RNase. It is believed that S-RNase enters the pollen tube and degrades mRNA of incompatible pollen, but it is not currently understood how male and female S-determinants interact, allowing the pollen to be recognised as incompatible. Various mechanisms have been suggested including the ubiquitination and proteasomal degradation or the inhibition of S-RNase, but currently there is not enough evidence for the universal acceptance of any one mechanism (Roalson & McCubbin, 2003, Takayama & Isogai, 2005).


Self-incompatibility in Papaver rhoeas (known the field poppy) is triggered by a female S-determinant known as PrsS (P. rhoeas stigma S-determinant) and a male determinant known as PrpS (P. rhoeas pollen S-determinant). PrsS is a 15kDa protein secreted by the pistil, while structural predictions have indicated that PrpS is a transmembrane protein that acts as a receptor for the equivalent PrsS. Only specific PrsS haplotypes are able to bind to their equivalent PrpS causing an SI response. PrpS genes are highly divergent with between 50 and 60% sequence identity. It has been estimated that there are 66 Papaver S-haplotypes (Wheeler, et al., 2009, Poulter, et al., 2010).

The binding of a PrsS to its equivalent PrpS causes a calcium-dependant signalling cascade. Within a few seconds of binding, Ca2+ concentration within the pollen tube increases and continues to increase for several minutes (Bosch & Franklin-Tong, 2008). Pollen tube growth is halted within 5 minutes, while programmed cell death (PCD) can be detected within a few hours. The mechanism by which PCD occurs is currently unclear.

The increase in calcium concentration within nascent pollen tube cells has been shown to have four effects: the F-actin cytoskeleton begins depolymerising (Poulter, et al., 2010), a Mitogen-activated protein kinase (MAPK), known as p56, is activated (Li, et al., 2007) caspases-like activity occurs (Thomas & Franklin-Tong, 2004), and p26, an inorganic pyrophosphatase, is phosphorylated (de Graaf, et al., 2006). Each of these effects has been shown to be a direct result of the increase in calcium concentration but it is not known how they stop pollen tube growth and eventually lead to the PCD of the pollen tube.

Actin depolymerisation can be detected within one minute of PrsS binding and reaches 74% within an hour. Secretory vesicles required by an actively growing pollen tube are transported to the tip by the actin cytoskeleton naturally if this is being broken down vesicles cannot reach the growing tip. Actin depolymerisation can therefore be directly linked to the cessation of pollen tube growth. However, the level of depolymerisation observed is several orders of magnitude higher than is required for growth inhibition. Actin depolymerisation alone has also recently been shown to be capable of triggering programmed cell death (Thomas, et al., 2006). It is currently unknown how the increase in calcium concentration triggers actin depolymerisation but the actin-binding proteins profilin, gelsolin, CAP, ADF, and PrABP80 may be involved in the process (Takayama & Isogai, 2005, Poulter, et al., 2010).

p56 is a 56kDa Mitogen-activated protein kinase (MAPK) that is activated as a result of the increase in calcium concentration. MAPKs are involved in a wide variety of signalling cascades, many of which can result in cell death, indeed MAPKs have been shown to cause cell death in plants as a response to attack from a pathogen, p56 is therefore a good candidate for an effector of PCD. p56 activity can be detected around five minutes after the induction of the SI response, it is therefore unlikely that p56 is involved in growth arrest, which occurs before p56 is activated. MAPKs have been shown to affect actin dynamics though so it is possible that p56 could promote further actin depolymerisation, making growth arrest irreversible (Takayama & Isogai, 2005, Li, et al., 2007).

In mammals, apoptosis, a form of PCD, involves the degradation of cellular structures and the eventual sequestration of the entire contents of the cell into apoptotic bodies, which are then phagocytosed by nearby cells. Apoptosis is triggered by the auto activation of caspases (cysteine aspartate proteases), which cleave a wide variety of targets within the cell including Poly (ADP-ribose) polymerase (PARP). Once apoptosis is underway, the leakage of cytochrome c from the mitochondria can be detected (Jin & El-Deiry, 2005). Each of these can be used as a sign of apoptosis and has also been observed in the dying pollen tube. Caspase inhibitors such as DEVD, a caspase-3 inhibitor, are able to prevent the death of the pollen tube but the Papaver rhoeas genome does not contain any caspase homologues, the proteins responsible for this caspase-like activity are therefore unknown Pollen tube growth arrest still occurs when caspase inhibitors are present, suggesting that the caspase-like activity is only involved with PCD. The caspase-like activity is required to make growth arrest irreversible though - pollen tube growth will resume 15-45 minutes after stopping if caspase inhibitors are present (Thomas & Franklin-Tong, 2004, Jin & El-Deiry, 2005).


Although p26 was first implicated in the SI response of P. rhoeas by Rudd et al. in 1996, much of was is currently known was found in the same lab by de Graaf et al. in 2006. p26 is a heterodimer of two Family I inorganic pyrophosphatases: p26.1a and p26.1b. p26.1a is a 24.4kDa protein with a pI of 6.11, while p26.1b is a 26.5kDa protein with a pI of 6.03. The two proteins have a sequence identity of 78%. p26 is phosphorylated by a Calcium dependant protein kinase (CDPK) within 90 seconds of the induction of the SI response, this is followed by a further increase in the level of phosphorylation within 400s. The phosphorylation of p26 causes a decrease in hydrolytic activity of 50-70% (Rudd, et al., 1996, de Graaf, et al., 2006).

The rapid growth of pollen tubes requires high levels of synthesis of biopolymers, which, as previously discussed, result in the production of inorganic pyrophosphatase, the concentration of which is kept low by p26, allowing tip growth to continue. Knocking down p26 expression resulted in significantly shortened pollen tubes (one fifth of the control length), while the SI response resulted in a doubling of the inorganic pyrophosphate concentration (de Graaf, et al., 2006). This demonstrates a direct link between the pyrophosphorolytic activity of p26, pollen tube growth, and the SI response. It is therefore a logical conclusion that, as the thermodynamic pull of pyrophosphorolysis towards biosynthesis is significantly reduced, pollen tube growth will be significantly slowed or stopped altogether by the phosphorylation of p26 triggered by the SI response (de Graaf, et al., 2006). As PPase activity is required for many metabolic processes, it is possible that the phosphorylation of p26 may play a role in the induction of programmed cell death of pollen tubes, but the two are yet to be mechanistically linked.

The regulation of PPases by phosphorylation has not, at present, been widely observed or studied. Phosphorylation of PPases from rat liver and E. coli has been observed in vitro (Sklyankina & Avaeva, 1990, Vener, et al., 1990), and an inorganic pyrophosphatase is phosphorylated in Streptococcus during purine synthesis (Rajagopal, et al., 2005, Novakova, et al., 2010), but these examples are largely unexplored. The phosphorylation of p26 in the P. rhoeas SI response therefore presents a major new area of scientific research. Understanding how it reduces pyrophosphorolytic activity through the production of crystal structures of the two proteins in various states could not only shed light on the Self-Incompatible response in P. rhoeas but also on signalling in plants as a whole and the regulation of inorganic pyrophosphatases.



Fig 1. A Inorganic Pyrophosphate is a bi-product of many biosynthetic reactions including DNA chain extension, as each nucleotide is added, PPi is released. This is reaction does not yield a high level of free energy, so the PPi produced must be hydrolysed to allow synthesis to continue. B The Hydrolysis of inorganic pyrophosphate as catalysed by inorganic pyrophosphatases. The phosphoanhydride bond is cleaved through the addition of water, producing two molecules of inorganic phosphate. This reaction keeps the cellular concentration of PPi at a minimum, driving biosynthetic reactions that produce PPi as a bi-product forward (Kornberg 1962).







Fig 2. Stereo views of a Superimposition of E-PPase (Blue) and Y-PPase (Red). The two ribbon diagrams largely have the same structure, with slight differences in loops between secondary structure motifs. Inorganic pyrophosphate can be seen bound to the active site at the top of the protein. The almost exact superimposition of the two PPi molecules shows that the active sites have a highly conserved spatial arrangement for optimal catalysis. The β-barrel is visible just to the left of the active site, flanked by α-helices. The C-terminal extension of animal and fungal PPases not present in plant or prokaryotic PPases can be seen at the top of the image. Crystal structures used in the image are 1E6A (Heikinheimo, 2001) and 2AUU (Samygina, et al., 2007).

Fig 3. Stereo views of the active site of Y-PPase with inorganic pyrophosphatase bound. The four manganese ions are show in orange while a fluoride ion in the position that would be occupied by Wat1 is shown in purple, hydrogen bonds are shown in green. Note the large amount of basic and polar residues surrounding the substrate forming the network of covalent and non-covalent interactions, decreasing the electron density across the PPi, stabilising the transition state and forming the catalytic hydroxide ion. The Wat1 binding site is in close proximity to P2 for ease of attack and rapid catalysis. The three aspartate residues of the 115DXDXXDX motif are also visible. Crystal structure 1E6A (Heikinheimo, 2001).



















Fig 4. PPases have been crystallised at various stages throughout the catalytic cycle, slight structural changes occur between each state but catalysis does not occur until the third step. The holoenzyme sits with two magnesium ions in the active site, magnesium pyrophosphate enters, causing the displacement of a water molecule. The network of covalent and non-covalent interactions is formed and the nucleophilic water molecule attacks PPi, resulting in the formation of two magnesium phosphate molecules. A water molecule enters the active site, causing P2 to be released. A second water molecule enter sconcomitantly with this release, forcing P1 out of the active site, and returning the enzyme to its original state, ready for catalysis to occur again (Heikinheimo, 2001)