Predicting Electrophoretic Mobility of Protein-Ligand Complexes for Ligands from DNA-Encoded Libraries of Small Molecules.

Selection of target-binding ligands from DNA-encoded libraries of small molecules (DELSMs) is a rapidly developing approach in drug-lead discovery. Methods of kinetic capillary electrophoresis (KCE) may facilitate highly efficient homogeneous selection of ligands from DELSMs. However, KCE methods require accurate prediction of electrophoretic mobilities of protein-ligand complexes. Such prediction, in turn, requires a theory that would be applicable to DNA tags of different structures used in different DELSMs. Here we present such a theory. It utilizes a model of a globular protein connected, through a single point (small molecule), to a linear DNA tag containing a combination of alternating double-stranded and single-stranded DNA (dsDNA and ssDNA) regions of varying lengths. The theory links the unknown electrophoretic mobility of protein-DNA complex with experimentally determined electrophoretic mobilities of the protein and DNA. Mobility prediction was initially tested by using a protein interacting with 18 ligands of various combinations of dsDNA and ssDNA regions, which mimicked different DELSMs. For all studied ligands, deviation of the predicted mobility from the experimentally determined value was within 11%. Finally, the prediction was tested for two proteins and two ligands with a DNA tag identical to those of DELSM manufactured by GlaxoSmithKline. Deviation between the predicted and experimentally determined mobilities did not exceed 5%. These results confirm the accuracy and robustness of our model, which makes KCE methods one step closer to their practical use in selection of drug leads, and diagnostic probes from DELSMs.

Slow-Equilibration Approximation in Kinetic Size Exclusion Chromatography.

Kinetic size exclusion chromatography with mass spectrometry detection (KSEC-MS) is a solution-based label-free approach for studying kinetics of reversible binding of a small molecule to a protein. Extraction of kinetic data from KSEC-MS chromatograms is greatly complicated by the lack of separation between the protein and protein-small molecule complex. As a result, a sophisticated time-consuming numerical approach was used for the determination of rate constants in the proof-of-principle works on KSEC-MS. Here, we suggest the first non-numerical (analytical) approach for finding rate constants of protein-small molecule interaction from KSEC-MS data. The approach is based on the slow-equilibration approximation, which is applicable to KSEC-MS chromatograms that reveal two peaks. The analysis of errors shows that the slow-equilibration approximation guarantees that the errors in the rate constants are below 20% if the ratio between the characteristic separation and equilibration times does not exceed 0.1. The latter condition can typically be satisfied for specific interactions such as receptor-ligand or protein-drug. The suggested analytical solution equips analytical scientists with a simple and fast tool for processing KSEC-MS data. Moreover, a similar approach can be potentially developed for kinetic analysis of protein-small molecule binding by other kinetic-separation methods such as nonequilibrium capillary electrophoresis of equilibrium mixtures (NECEEM).

Slow-equilibration approximation in studying kinetics of protein adsorption on capillary walls.

Adsorption of proteins on inner capillary walls affects the quality of capillary electrophoresis (CE) analyses. Coating the capillary surface with an anti-adhesive layer is a method typically used to suppress protein adsorption. The successful development of methods for prevention of protein adsorption requires quantitative characterization of the surface ability to adsorb and desorb a protein. It can be done by determining kinetic rate constants of adsorption, kad, and desorption, kdes. We have recently developed a pattern-based method for determination of kad and kdes for protein interaction with a capillary wall. The protein is moved through the capillary in a CE instrument by pressure and a temporal pattern of protein propagation through the detector is recorded. The experimental pattern is fitted with a numerical solution of the protein mass transfer to find kad and kdes. The fitting procedure is not "transparent" and can be complicated. In the present work, we obtained approximate analytical solutions of the protein mass transfer equations in the case of slow-equilibration during adsorption and desorption of the protein. These analytical solutions allow us to introduce a fitting-free parameter based method for determination of kad and kdes. It uses simple explicit expressions for kad and kdes in terms of experimental characteristics easily measured in capillaries. We tested the accuracy of the method by applying it to signals simulated with numerical solutions of protein mass-transfer equations. For the slow equilibration approximation the accuracy of kad and kdes was better than 12%.

Using nonequilibrium capillary electrophoresis of equilibrium mixtures (NECEEM) for simultaneous determination of concentration and equilibrium constant.

Nonequilibrium capillary electrophoresis of equilibrium mixtures (NECEEM) is a versatile tool for studying affinity binding. Here we describe a NECEEM-based approach for simultaneous determination of both the equilibrium constant, K(d), and the unknown concentration of a binder that we call a target, T. In essence, NECEEM is used to measure the unbound equilibrium fraction, R, for the binder with a known concentration that we call a ligand, L. The first set of experiments is performed at varying concentrations of T, prepared by serial dilution of the stock solution, but at a constant concentration of L, which is as low as its reliable quantitation allows. The value of R is plotted as a function of the dilution coefficient, and dilution corresponding to R = 0.5 is determined. This dilution of T is used in the second set of experiments in which the concentration of T is fixed but the concentration of L is varied. The experimental dependence of R on the concentration of L is fitted with a function describing their theoretical dependence. Both K(d) and the concentration of T are used as fitting parameters, and their sought values are determined as the ones that generate the best fit. We have fully validated this approach in silico by using computer-simulated NECEEM electropherograms and then applied it to experimental determination of the unknown concentration of MutS protein and K(d) of its interactions with a DNA aptamer. The general approach described here is applicable not only to NECEEM but also to any other method that can determine a fraction of unbound molecules at equilibrium.

Prediction of protein-DNA complex mobility in gel-free capillary electrophoresis.

Selection of protein binders from highly diverse combinatorial libraries of DNA-encoded small molecules is a highly promising approach for discovery of small-molecule drug leads. Methods of kinetic capillary electrophoresis provide the high efficiency of partitioning required for such selection but require the knowledge of electrophoretic mobility of the protein-ligand complex. Here we present a theoretical approach for an accurate estimate of the electrophoretic mobility of such complexes. The model is based on a theory of the thin double layer and corresponding expressions used for the mobilities of a rod-like short oligonucleotide and a sphere-like globular protein. The model uses empirical values of mobilities of free protein, free ligand, and electroosmotic flow. The model was tested with a streptavidin-dsDNA complex linked through biotin (small molecule). The deviation of the prediction from the experimental mobility did not exceed 4%, thus confirming that not only is the model adequate but it is also accurate. This model will facilitate reliable use of KCE methods for selection of drug leads from libraries of DNA-encoded small molecules.

Pre-equilibration kinetic size-exclusion chromatography with mass spectrometry detection (peKSEC-MS) for label-free solution-based kinetic analysis of protein-small molecule interactions.

Here we introduce pre-equilibration kinetic size-exclusion chromatography with mass-spectrometry detection (peKSEC-MS), which is a label-free solution-based kinetic approach for characterizing non-covalent protein-small molecule interactions. In this method, a protein and a small molecule are mixed outside the column and incubated to approach equilibrium. The equilibrium mixture is then introduced into the SEC column to initiate the dissociation process by separating small molecules from the complex inside the column. A numerical model of a 1-dimensional separation was constructed to simulate mass chromatograms of the small molecule for varying rate constants of binding.

One-dimensional approach to study kinetics of reversible binding of protein on capillary walls.

We introduce a method for kinetic characterization of reversible binding of protein onto the inner capillary wall. In essence, a short plug of the protein solution is propagated through the capillary by pressure, and the protein is detected at the distal capillary end. The signal versus time profile is fitted with a numerical model which uses the rate constants of adsorption, kad, and desorption, kde, as fitting parameters. The values of kad and kde which result in the best fit are considered to be the sought ones. We first used COMSOL multiphysics software to develop a numerical model with two-dimensional (2D) equations of mass transfer. Although 2D models in general can describe experiments more accurately than one-dimensional (1D) models, computing 2D models takes much more time (many hours to find two parameters: kad and kde). We used the fact that the capillary is narrow and long to develop a simplified model with 1D equations of mass transfer. Our comparison of the 1D and 2D models showed that the errors of the 1D approximation were less than 5%, whereas the computation of the 1D model was 100 times faster. We finally used the 1D approach to study kinetics of reversible binding of conalbumin to the uncoated fused-silica capillary walls. We determined kad, kde, and a diffusion coefficient, D. The obtained value of D is in excellent agreement with literature data which suggests that the values of kad and kde (for which there are no literature data) are also calculated correctly. Our approach for finding kad and kde will facilitate quantitative characterization of protein adsorption on capillary walls as well as properties of passivating materials used for capillary coating.

Kinetic size-exclusion chromatography with mass spectrometry detection: an approach for solution-based label-free kinetic analysis of protein-small molecule interactions.

Studying the kinetics of reversible protein-small molecule binding is a major challenge. The available approaches require that either the small molecule or the protein be modified by labeling or immobilization on a surface. Not only can such modifications be difficult to do but also they can drastically affect the kinetic parameters of the interaction. To solve this problem, we present kinetic size-exclusion chromatography with mass spectrometry detection (KSEC-MS), a solution-based label-free approach. KSEC-MS utilizes the ability of size-exclusion chromatography (SEC) to separate any small molecule from any protein-small molecule complex without immobilization and the ability of mass spectrometry (MS) to detect a small molecule without a label. The rate constants of complex formation and dissociation are deconvoluted from the temporal pattern of small molecule elution measured with MS at the exit from the SEC column. This work describes the concept of KSEC-MS and proves it in principle by measuring the rate constants of interaction between carbonic anhydrase and acetazolamide.

Reducing pH gradients in free-flow electrophoresis.

Small-volume continuous-flow synthesis (small-volume CFS) offers a number of benefits for use in small-scale chemical production and exploratory chemistry. Typically, small-volume CFS is followed by discontinuous purification; however, a fully continuous synthesis-purification combination is more attractive. Milli free-flow electrophoresis (mFFE) is a promising continuous-flow purification technique that is well suited for integration with small-volume CFS. The purification stability of mFFE, however, needs to be significantly improved before it can be feasible for this combination. One of the major sources of instability of mFFE is attributed to the ions produced as a result of electrolysis. These ions can form pH and conductivity gradients in mFFE, which are detrimental to separation quality. The severity of these gradients has not been thoroughly studied in mFFE. In this paper, we have experimentally demonstrated that detrimental pH gradients occur at flow rates of 8 mL/min and less, and electric field strengths of 25 V/cm and greater. To decrease the pH gradients, it is necessary to evacuate H(+) and OH(-) as soon as they are generated; this can be done by increasing local hydrodynamic flow rates. We calculated the necessary flow rate, to be applied at the electrode, which can effectively wash away both ions before they can cause a detrimental pH gradient. These optimized flow rates can be attained by designing a device that incorporates deep channels. We have confirmed the effectiveness of these channels using a prototyped device. The new design allows mFFE users to work over a wider range of flow-rate and electric-field conditions without experiencing significant changes in pH.

Extracting kinetics from affinity capillary electrophoresis (ACE) data: a new blade for the old tool.

We describe a mathematical approach that enables extraction of kinetic rate constants from thousands of studies conducted over the past two decades with affinity capillary electrophoresis (ACE). Previously, ACE has been used almost exclusively for obtaining equilibrium constants of intermolecular interactions. In this article, we prove that there exists an analytical solution of partial differential equations describing mass transfer in ACE. By using an in silico study, we demonstrate that the solution is applicable to experimental conditions that are typically used in ACE and found in most historical ACE experiments. The solution was validated by extracting rate constants from previously published ACE data and closely matching independently obtained results. Lastly, it was used to obtain previously unknown rate constants from historical ACE data. The new mathematical approach expands the applicability of ACE to a wider range of biomolecular interactions and enables both prospective and retrospective data analysis. The obtained kinetic information will be of significant practical value to the fields of pharmacology and molecular biology.

Universal drag tag for direct quantitative analysis of multiple microRNAs.

Using microRNA (miRNA) as molecular markers of diseases requires a method for accurate measurement of multiple miRNAs in biological samples. Direct quantitative analysis of multiple miRNAs (DQAMmiR) has been recently developed based on a classical hybridization approach. In DQAMmiR, miRNAs are hybridized with excess fluorescently labeled complementary DNA probes. Capillary electrophoresis (CE) is used to separate the unreacted probes from the hybrids and the hybrids from each other. The challenging separation is achieved by using two types of mobility modifiers. Single-strand DNA binding protein (SSB) is added to the running buffer to bind and shift the single-stranded unreacted probes from the double-stranded hybrids. Different drag tags are built into the probes to introduce significant differential mobility between their respective hybrids. For the method to be practical it requires a universal extendable drag tag. Polymers are a logical choice for making extendable drag tags. Our recent theoretical work suggested that short peptides could provide a sufficient mobility shift to facilitate DQAMmiR. Here, we experimentally confirm this prediction in the analysis of five miRNAs: mir10b, mir21, mir125b, mir145, and mir155. We conjugated four fluorescently labeled DNA molecules with peptides of 5, 10, 15, or 20 neutral amino acids in length; the fifth probe was peptide-free. The peptide tags showed no interference with SSB binding to the probes and facilitated separation of the five hybrids. The mobilities of the five hybrids were used to refine the previously suggested theory. The analysis was performed in both a pure buffer and in cell lysate. Our analysis of the experimental data suggests that using DNA-peptide probes can readily facilitate simultaneous analysis of more than 10 miRNAs.

Theoretical modeling of masking DNA application in aptamer-facilitated biomarker discovery.

In aptamer-facilitated biomarker discovery (AptaBiD), aptamers are selected from a library of random DNA (or RNA) sequences for their ability to specifically bind cell-surface biomarkers. The library is incubated with intact cells, and cell-bound DNA molecules are separated from those unbound and amplified by the polymerase chain reaction (PCR). The partitioning/amplification cycle is repeated multiple times while alternating target cells and control cells. Efficient aptamer selection in AptaBiD relies on the inclusion of masking DNA within the cell and library mixture. Masking DNA lacks primer regions for PCR amplification and is typically taken in excess to the library. The role of masking DNA within the selection mixture is to outcompete any nonspecific binding sequences within the initial library, thus allowing specific DNA sequences (i.e., aptamers) to be selected more efficiently. Efficient AptaBiD requires an optimum ratio of masking DNA to library DNA, at which aptamers still bind specific binding sites but nonaptamers within the library do not bind nonspecific binding sites. Here, we have developed a mathematical model that describes the binding processes taking place within the equilibrium mixture of masking DNA, library DNA, and target cells. An obtained mathematical solution allows one to estimate the concentration of masking DNA that is required to outcompete the library DNA at a desirable ratio of bound masking DNA to bound library DNA. The required concentration depends on concentrations of the library and cells as well as on unknown cell characteristics. These characteristics include the concentration of total binding sites on the cell surface, N, and equilibrium dissociation constants, K(nsL) and K(nsM), for nonspecific binding of the library DNA and masking DNA, respectively. We developed a theory that allows the determination of N, K(nsL), and K(nsM) based on measurements of EC50 values for cells mixed separately with the library and masking DNA (EC50 is the concentration of fluorescently labeled DNA at which half of the maximum fluorescence signal from DNA-bound cells is reached). We also obtained expressions for signals from bound DNA (measured by flow cytometry) in terms of N, K(nsL), and K(nsM). These expressions can be used for the verification of N, K(nsL), and K(nsM) values found from EC50 measurements. The developed procedure was applied to MCF-7 breast cancer cells, and corresponding values of N, K(nsL), and K(nsM) were established for the first time. The concentration of masking DNA required for AptaBiD with MCF-7 breast cancer cells was also estimated.

Theoretical estimation of drag tag lengths for direct quantitative analysis of multiple miRNAs (DQAMmiR).

To better understand the regulatory roles of miRNA in biological functions and to use miRNA as molecular markers of diseases, we need to accurately measure amounts of multiple miRNAs in biological samples. Direct quantitative analysis of multiple miRNAs (DQAMmiR) has been recently developed by using a classical hybridization approach where miRNAs are hybridized with fluorescently labeled complementary DNA probes taken in excess, and the amounts of the hybrids and the unreacted probes are measured to calculate the amount of miRNAs. Capillary electrophoresis was used as an instrumental platform for analysis. The problem of separating the unreacted probes from the hybrids was solved by adding SSB to the run buffer. A more difficult problem of separating hybrids from each other was solved by attaching different drag tags to the probes. Biotin and a hairpin-forming extension on the probe were used as two drag tags in the proof-of-principle work. Making DQAMmiR a generic approach requires a generic solution for drag tags. Peptides have been suggested as drag tags for long oligonucleotides in DNA sequencing by electrophoresis. Here we theoretically consider short peptides of different lengths as drag tags for DQAMmiR. We find analytical equations that allow us to estimate mobilities of RNA-DNA hybrids with peptide drag tags of different lengths. Our calculations suggest that the mobility shifts required for DQAMmiR can be achieved with the length of peptide chains in the ranges of 5-20 residues for five miRNAs and 2-47 residues for nine miRNAs. Peptides of these lengths can be feasibly synthesized with good yield and purity. The results of this theoretical study will guide the design and production of hybridization probes for DQAMmiR.

Sequential-dissociation kinetics of non-covalent complexes of DNA with multiple proteins in separation-based approach: general theory and its application.

Binding of multiple proteins to DNA is crucial in many regulatory cellular processes. The kinetics of assembly and disassembly of DNA-multiple protein complexes is very difficult to study in detail due to the lack of suitable experimental approaches. A separation-based approach has been recently proposed to resolve disassembly kinetics of such complexes. While conceptually simple, the separation-based approach generates experimental data with very complex patterns. The analysis of these patterns is a challenging problem on its own. Here we report on a mathematical approach that can extract a solution for the experimental data obtained in separation-based analysis of sequential dissociation of a DNA complex with multiple proteins. This case describes the dissociation of proteins one-by-one from the complex. Generally speaking, a mathematical solution of such problems requires calculations of multiple integrals. Our approach reduces this procedure to taking double integrals and constructing their superposition. We tested this approach with the experimental data obtained for three-step sequential dissociation of complexes of DNA with two protein copies.

Simplified universal method for determining electrolyte temperatures in a capillary electrophoresis instrument with forced-air cooling.

Temperature increase due to resistive electrical heating is an inherent limitation of capillary electrophoresis (CE). Active cooling systems are used to decrease the temperature of the capillary, but their capacity is limited; and in addition, they leave "hot spots" at the detection interface and at the capillary ends. Until recently, the matter was complicated by the lack of a fast and generic method for temperature determination in efficiently and inefficiently cooled regions of the capillary. Our group recently introduced such a method, termed "Universal Method for determining Electrolyte Temperatures" (UMET). UMET is a probe-less approach that requires only measuring current versus voltage for different voltages and processing the data using an iterative algorithm. Here, we apply UMET to develop a Simplified Universal Method of Temperature Determination (SUMET) for a CE instrument with a forced-air cooling system using an Agilent 7100 CE instrument (Agilent Technologies, Saint Laurent, Quebec, Canada) as an example. We collected a wide set of empirical voltage-current data for a variety of buffers and capillary diameters. We further constructed empirical equations for temperature calculation in efficiently and inefficiently cooled parts of the capillary that require only the data from a single 1-min voltage-current measurement. The equations are specific for the Agilent 7100 CE instrument (Agilent Technologies) but can be applied to all kinds of capillaries and buffers. Similar SUMET approaches can be developed for other CE instruments with forced-air cooling using our approach.

Two-peak approximation in kinetic capillary electrophoresis.

Kinetic capillary electrophoresis (KCE) constitutes a toolset of homogeneous kinetic affinity methods for measuring rate constants of formation (k(+)) and dissociation (k(-)) of non-covalent biomolecular complexes, C, formed from two binding partners, A and B. A parameter-based approach of extracting k(+) and k(-) from KCE electropherograms relies on a small number of experimental parameters found from the electropherograms and used in explicit expressions for k(+) and k(-) derived from approximate solutions to mass transfer equations. Deriving the explicit expressions for k(+) and k(-) is challenging but it is justified as the parameter-based approach is the simplest way of finding k(+) and k(-) from KCE electropherograms. Here, we introduce a unique approximate analytical solution of mass transfer equations in KCE termed a "two-peak approximation" and a corresponding parameter-based method for finding k(+) and k(-). The two-peak approximation is applicable to any KCE method in which: (i) A* binds B to form C* (the asterisk denotes a detectable label on A), (ii) two peaks can be identified in a KCE electropherogram and (iii) the concentration of B remains constant. The last condition holds if B is present in access to A* and C* throughout the capillary. In the two-peak approximation, the labeling of A serves only for detection of A and C and, therefore, is not required if A (and thus C) can be observed with a label-free detection technique. We studied the proposed two-peak approximation, in particular, its accuracy, by using the simulated propagation patterns built with the earlier-developed exact solution of the mass-transfer equations for A* and C*. Our results prove that the obtained approximate solution of mass transfer equations is correct. They also show that the two-peak approximation facilitates finding k(+) and k(-) with a relative error of less than 10% if two peaks can be identified on a KCE electropherogram. Importantly, the condition of constant concentration of B is always satisfied in macroscopic approach to studying kinetics at equilibrium (MASKE) whether or not B is in excess to A* and C*, and, thus, the two-peak approximation is applicable to MASKE. It completes a toolset of fitting-free methods for processing MASKE data and makes MASKE a simple practical method for finding k(+) and k(-) of "fast", "slow", and "intermediate-rate" non-covalent interactions.

Peak-shape correction to symmetry for pressure-driven sample injection in capillary electrophoresis.

Pressure-driven sample injection in capillary electrophoresis results in asymmetric peaks due to difference in shapes between the front and the back boundaries of the sample plug. Uneven velocity profile of fluid flow across the capillary gives the front boundary a parabolic shape. The back side, on the other hand, has a flat interface with the electrophoresis run buffer. Here, we propose a simple means of correcting this asymmetry by pressure-driven "propagation" of the injected plug, with the parabolic sample-buffer interface established at the back. We prove experimentally that such a propagation procedure corrects peak asymmetry to the level comparable to injection through electroosmosis. Importantly, the propagation-based correction procedure also solves a problem of transferring the sample into the efficiently cooled zone of the capillary for capillary electrophoresis (CE) instruments with active cooling. The suggested peak correction procedure will find applications in all CE methods that rely on peak shape analysis, e.g., nonequilibrium capillary electrophoresis of equilibrium mixtures.

Method for determination of peak areas in nonequilibrium capillary electrophoresis of equilibrium mixtures.

Nonequilibrium capillary electrophoresis of equilibrium mixtures (NECEEM) facilitates determination of both the kinetic constants (k(off)) and the equilibrium constants (K(d)) of complex dissociation from a single experiment. A typical NECEEM electropherogram consists of two peaks and an "exponential bridge" between them, smoothly merging into the peaks. The values of k(off) and K(d) are usually calculated with simple algebraic formulas, by utilizing the areas of the peaks and the bridge. Accurate determination of the two constants requires accurate positioning of the two boundaries separating the bridge from the peaks. Here, we propose a more systematic method for the determination of boundaries between the peaks and the bridge. The method involves a simple geometrical analysis of a NECEEM electropherogram based on an assumption of symmetry in ordinary electrophoretic peaks. To test the method, we (i) constructed a series of computer-simulated NECEEM electropherograms, (ii) determined the two boundaries with our method, and (iii) calculated the values of k(off) and K(d). We found that the deviation of the calculated values from those used to simulate the electropherograms did not exceed 15% for k(off) and 25% for K(d), as long as the peaks and the bridge were visually identifiable. We finally applied the method to the determination of K(d) and k(off) values for the interaction between AlkB protein and its DNA aptamer. The developed method for rational boundary determination in NECEEM will facilitate accurate data analysis in a simple and efficient manner.

The structures of Thermoplasma volcanium phosphoribosyl pyrophosphate synthetase bound to ribose-5-phosphate and ATP analogs.

Phosphoribosyl pyrophosphate (PRPP) synthetase catalyzes the transfer of the pyrophosphate group from ATP to ribose-5-phosphate (R5P) yielding PRPP and AMP. PRPP is an essential metabolite that plays a central role in cellular metabolism. The enzyme from a thermophilic archaeon Thermoplasma volcanium (Tv) was expressed in Escherichia coli, crystallized, and its X-ray molecular structure was determined in a complex with its substrate R5P and with substrate analogs ß,γ-methylene ATP and ADP in two monoclinic crystal forms, P2(1). The ß,γ-methylene ATP- and the ADP-bound binary structures were determined from crystals grown from ammonium sulfate solutions; these crystals diffracted to 1.8 Å and 1.5 Å resolutions, respectively. Crystals of the ternary complex with ADP-Mg(2+) and R5P were grown from a polyethylene glycol solution in the absence of sulfate ions, and they diffracted to 1.8 Å resolution; the unit cell is approximately double the size of the unit cell of the crystals grown in the presence of sulfate. The Tv PRPP synthetase adopts two conformations, open and closed, at different stages in the catalytic cycle. The binding of substrates, R5P and ATP, occurs with PRPP synthetase in the open conformation, whereas catalysis presumably takes place with PRPP synthetase in the closed conformation. The Tv PRPP synthetase forms a biological dimer in contrast to the tetrameric or hexameric quaternary structures of the Methanocaldococcus jannaschii and Bacillus subtilis PRPP synthetases, respectively.