Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2011 Aug 15.
Published in final edited form as: Anal Chem. 2010 Aug 15;82(16):6838–6846. doi: 10.1021/ac1008018

Analysis of Inorganic Polyphosphates by Capillary Gel Electrophoresis

Andrew Lee 1, George M Whitesides 1,*
PMCID: PMC2926810  NIHMSID: NIHMS225465  PMID: 20704373

Abstract

This paper describes the development of a method that uses Capillary Gel Electrophoresis (CGE) to analyze mixtures of inorganic polyphosphate ((Pi)n). Resolution of (Pi)n on the basis of n, the number of residues of dehydrated phosphate, is accomplished by CGE using capillaries filled with solutions of poly(N,N-dimethylacrylamide) (PDMA) and indirect detection by the UV-absorbance of a chromophore, terephthalate, added to the running buffer. The method is capable of resolving peaks representing (Pi)n with n up to ~70; preparation and use of authentic standards enables the identification of peaks for (Pi)n with n = 1 - 10. The main advantages of this method over previously reported methods for analyzing mixtures of (Pi)n (e.g., gel electrophoresis, CGE using polyacrylamide-filled capillaries) are its resolution, convenience, and reproducibility; gel-filled capillaries are easily regenerated by pumping in fresh, low-viscosity solutions of PDMA. The resolution is comparable to that of ion-exchange chromatography and detection of (Pi)n by suppressed conductivity. The method is useful for analyzing (Pi)n generated by the dehydration of Pi at low temperature (125 - 140 °C) with urea, in a reaction that may have been important in prebiotic chemistry. The method should also be useful for characterizing mixtures of other anionic, oligomeric or polymeric species without an intrinsic chromophore (e.g., sulfated polysaccharides, oligomeric phospho-diesters).

INTRODUCTION

This paper describes a technique for analyzing mixtures of inorganic oligo- and poly-phosphates, (Pi)n, by capillary gel electrophoresis. Samples of condensed inorganic phosphate are typically mixtures of (Pi)n with different chain length, n. Mixtures of (Pi)n can have a wide range in n; (Pi)n with estimated values of n as high as 1000 have been reported.1 The analytical method developed here characterizes samples of (Pi)n at high resolution by separating and detecting each species in a mixture. Our primary motivation to develop a new method of analysis was to explore the synthesis and reactivity of (Pi)n relevant to the chemical origins of life (i.e., the prebiotic chemistry leading to self-replicating systems in a “pre-RNA” or “RNA world”).2-7

Species of (Pi)n in aqueous solution are anionic, and differ from each other in the number of residues of condensed phosphate and in net negative electrostatic charge. Capillary electrophoresis (CE), in its most straightforward mode of operation – capillary zone electrophoresis (CZE): that is, electrophoresis of analytes through free solution, combined with optical detection of chromophoric analytes – cannot resolve and detect (Pi)n. We developed a method that addresses the two problems of i) separating (Pi)n in mixtures and ii) detecting and quantifying each (Pi)n. Capillary gel electrophoresis (CGE), using capillaries filled with aqueous solutions of poly(N,N-dimethylacrylamide) (PDMA), resolved cyclic and linear (Pi)n in order of their electrophoretic mobility. Addition of the chromophoric anion, terephthalate (1,4-(CO2-)2C6H4, abbreviated as TP2-), to the running buffer enabled the detection of separated (Pi)n by indirect UV absorbance.

We demonstrated the resolution of mixtures of (Pi)n (with n up to ~70), and the identification of components in mixtures with the use of authentic standards. The areas of peaks in electropherograms, determined by indirect detection, allowed us to quantify the relative concentration of each species in a mixture of (Pi)n. In addition to analyzing commercially available samples of (Pi)n, prepared by the thermal dehydration of Pi (> 220 °C), we analyzed (Pi)n generated by dehydration reactions that might have occurred on the prebiotic earth, and thus might have been involved in the chemical origins of life.5-6,8

Motivation

(Pi)n and adenosine triphosphate (ATP) have an essential functional group in common: residues of dehydrated phosphate connected by phosphoanhydride bonds. That (Pi)n is simpler in composition than ATP, but in principle may provide the same chemical function (e.g., activation of –OH groups), has led to the suggestion that (Pi)n is a molecular fossil – a species important in the origin of life – and a precursor to ATP.9-13 We wished to explore the dehydration of Pi in detail, and needed a method that was more convenient and reproducible than those reported so far for the resolution of mixtures of (Pi)n varying in chain length. The method we have developed should also be useful for investigating the biochemistry of (Pi)n,9-11, and for developing (Pi)n as a reagent in chemical synthesis. In addition, the method may be useful in applications for quality control: (Pi)n is a component of many commercial materials (e.g., fertilizers, food products, detergent formulations, building materials).14

Previously Reported Methods for Separating Mixtures of (Pi)n

Polyacrylamide Gel Electrophoresis (PAGE) can resolve mixtures of (Pi)n, with n ~ 2 - 450. Analysis by PAGE involves gel electrophoresis (typical runs require ≥ 3 hours) and the detection of (Pi)n by staining gels with the cationic dye Toluidine Blue O (TBO)1,15-16; autoradiography can detect species of (Pi)n synthesized from 32P-ATP.17 The disadvantages to PAGE are the difficulty of the experiments and the time required for analysis. We found it difficult to cast 20% gels that are homogenous and provide reproducible resolution; to run a gel requires several hours (for separation, staining and destaining).

Anion-exchange chromatography is useful both for analyzing mixtures of (Pi)n and for preparing samples of purified (Pi)n. The best results for the chromatographic resolution of (Pi)n are chromatograms showing up to ~50 peaks, in runs of less than 30 minutes;18-19 this method requires HPLC instrumentation with a suppressed conductivity detector and an on-line KOH gradient generator (for minimizing the amount of CO2/CO32- adsorbed from atmosphere). Distinguishing samples containing (Pi)n with n > 45 is difficult by this method, and the resolution of species with n ~ 100 has not been demonstrated.

Other qualitative or semi-quantitative methods used to analyze mixtures of (Pi)n include paper chromatography20-21, 31P NMR22-27, ESI-MS28, and the analysis of terminal phosphate groups with phosphoglucokinase.20

Capillary Electrophoresis of (Pi)n and DNA

Capillary electrophoresis separates and resolves analytes based on differences in electrophoretic mobility. Since both the amount of negative charge and hydrodynamic drag of (Pi)n increase with n, in a way similar to that of single-stranded DNA, we expected poor resolution in the analysis of (Pi)n in free-solution capillary electrophoresis (i.e., CZE).29 Capillaries filled with a sieving matrix however are capable of resolving DNA in order of chain length.30-32 The use of replaceable solutions of entangled polymer (such as linear acrylamide or PDMA) in CGE enabled the automated and massively parallel analysis of samples of DNA and was essential to the completion of the Human Genome Project.33-34

Previous reports of methods for analyzing (Pi)n by CGE used capillaries coated and filled with linear polyacrylamide (capillaries were prepared by filling capillaries with aqueous acrylamide and polymerizing in situ).35-36 We, and others, found that capillaries prepared this way had short lifetimes;37-40 these capillaries could not be re-used, required time-consuming preparation of new capillaries, and limited the reproducibility of the method. Stover36,41 and Wang29,35 detected (Pi)n by indirect UV absorbance in CGE experiments (chromate or pyromelltic acid were the chromophores added to the running buffer). These demonstrations did not however identify peaks for (Pi)n beyond P3, nor did they quantify resolved (Pi)n.

EXPERIMENTAL DESIGN

CZE – Separation of (Pi)n in Free Solution

We used the results of CZE experiments to guide the development of a CGE technique. Although the resolution of mixtures of (Pi)n in free solution is poor, CZE experiments are easier to run than CGE experiments, and allowed us to i) test different chromophores for indirect detection, and identify terephthalate (TP2-) as an optimal choice; ii) measure μ for analytical standards of (Pi)n in free solution, and determine the influence of pH and net electrostatic charge on the mobility of (Pi)n (pH in the ranges of 6.8 - 7.1 and 8.4 - 8.7); iii) optimize the composition of the running buffer. Table 1 gives literature values of pKa for (Pi)n. Analysis by CZE required the migration of (Pi)n to the anode and the use of capillaries with suppressed electro-osmotic flow. Capillaries covalently modified by reaction of the fused silica surface with a copolymer of N,N-dimethylacrylamide and 3-methacryloxy-propyltrimethoxysilane42 had electro-osmotic flow < 0.1 cm2 kV-1 min-1.

Table 1.

Values of pKa of Inorganic Polyphosphates.

Abbreviation Name pKa
P1 orthophosphatea 2.15, 7.20, 12.35
P2 pyrophosphatea 0.8, 2.2, 6.7, 9.4
P3 tri(poly)phosphateb 0.5, 1.0, 2.4, 6.5, 9.4
cyclo-P3 trimetaphosphatea 2.05c
(Pi)n polyphosphated ~1-2, 7.2-8.2d
Open in a new tab
a

Values at infinite dilution and 25 °C, taken from Ref57.

b

Values at infinite dilution and 25 °C, taken from Ref58.

c

Results of titration of cyclo-P3 cannot distinguish the pKa of the three ionizable groups of cyclo-P3; each of the three groups has a value of pKa of approximately 2.05.

d

Values taken from Ref14; ranges are inferred from titration curves for long-chain polyphosphates, and describe the strongly acidic hydrogen at each residue of phosphate (pKa of 1-2) and two weakly acidic hydrogen at the ends of the chain (pKa of 7.2-8.2).

CGE – Resolution of (Pi)n in Order of n

To achieve a combination of high resolution, speed, and convenience in the analysis of (Pi)n, we used CGE with solutions of PDMA (average molecular weight 57 kDa, 9.1% w/v) as the sieving medium. A key advantage of using a solution of PDMA39,43-47 is its low viscosity (< 100 cP), compared to that of other water-soluble polymers applied in CGE, such as polyacrylamide48-49, polyvinylpyrrolidone50, hydroxypropyl cellulose51, polyethylene glycol52-53. We prepared gel-filled capillaries by pumping solutions of PDMA into capillaries for 10 minutes at 30 psi. Regenerating gel-filled capillaries, by pumping in fresh solutions of PDMA (after 5 runs of 20 minute duration each), ensured the reproducibility of the separation media.

Detection of (Pi)n by Indirect UV Absorbance

(Pi)n cannot be detected by absorbance at wavelengths above 200 nm.54 Efforts to derivatize (Pi)n chemically with chromophores have not been successful.55-56 Electrophoresis however provides the opportunity to detect anions without chromophores, by indirect UV-absorbance. We used the anionic chromophore terephthalate (TP2-) for indirect detection, by adding terephthalic acid to solutions of running buffer (pH > 6.8; composition of the buffer discussed below). TP2- carries current in the capillary and migrates to the anode during electrophoresis. Migration of TP2- by the detector produces a steady-state signal in UV absorbance, generating the baseline of electropherograms. Sample zones containing (Pi)n also migrate towards the anode. Within a sample zone, the current is carried in part by (Pi)n, rather than TP2-. Zones containing (Pi)n are detected by a decrease in UV absorbance, resulting from a decrease in the concentration of TP2-.

The sensitivity of detection of (Pi)n depends primarily on the extinction coefficient of TP2-, and the mobilities of TP2- and (Pi)n. Terephthalate, TP2-, absorbs at λ = 254 nm (ε = 8.2·103 M-1 cm-1) and has mobility (μ ~ 28 cm2 kV-1 min-1) close to that of (Pi)n (26 - 34 cm2 kV-1 min-1 in free solution, pH = 7.0). The similarity in μ for TP2- and (Pi)n is important for limiting dispersion in electromigration and de-stacking, which result in the asymmetry and broadening of peaks.59 The estimated limit of detection of residues of Pi is ~ 0.2 μM.

Ions in the Running Buffer

The running buffer, by design, contains only one type of anion: TP2-. This characteristic avoids complications in the indirect detection of (Pi)n by the absorbance of TP2-. Anions in addition to TP2- would lead to system zones (that show up as negative peaks unrelated to (Pi)n), artifacts in peak shape, and complications in the analysis of peak areas.59-63 We therefore used buffers prepared by adding terephthalic acid to solutions of 18.0 - 24.0 mM bis-tris (pKa = 6.46) or tris (pKa = 8.06). The resulting buffers contained TP2- (3.0 mM), protonated amine (6.0 mM bis-tris-H+ or tris-H+), and free amine (12.0 - 18.0 mM of bis-tris or tris).

pH

Running buffers made with bis-tris or tris allowed us to compare the resolution of (Pi)n at two ranges of pH: 6.8 - 7.2 (bis-tris) and 8.4 - 8.7 (tris). These values of pH are well beyond the pKa of the first ionizable group of phosphate residues (~2) but are near the pKa of the second ionizable groups of the terminal residues of phosphate of (Pi)n (~6.3 - 7.2) (Table 1). Values of mobility for (Pi)n are sensitive to the pH of the running buffer, particularly for n < 5.

Addition of Polyethylene Glycol (PEG) to Reservoirs of Running Buffer

In CGE, capillaries are filled a solution of PDMA (density ~1.06 g/mL); the open ends of the capillary are immersed into reservoirs of running buffer that do not contain PDMA (~1.01 g/mL). As a result, solutions of PDMA leak out of the capillary into the buffer reservoirs, under gravity. The reproducibility of this experiment was therefore poor; the current decreased by > 10% within two hours, and retention times increased by > 5% in each subsequent run.

The solution to this problem was to add polyethylene glycol (PEG) to the reservoirs of running buffer, generating solutions isodense with the solution inside the capillary. PEG is water-soluble and uncharged; it does not migrate during electrophoresis. CGE experiments using reservoirs of running buffer with 9.0% PEG (w/v) showed stable currents and improved run-to-run reproducibility. This procedure was essential for maintaining a constant medium for separation; we routinely collected data for 120 minutes of electrophoresis for each preparation of a filled capillary (enough for 5-6 typical analyses of (Pi)n)64

Loading Samples by Electrokinetic Injection

During electrokinetic injection, an applied field forces anions in the sample to enter the capillary and migrate towards the anode. Rather than transferring plugs of solution into the capillary, electrokinetic injection only transfers anions. This procedure avoids the creation of discontinuities in the capillary and improves the reproducibility of separation medium.

The total number of ions injected is determined by the electrical field, conductivity of the running buffer, duration of the injection, and to a smaller extent, the conductivity of the sample (further discussion in the Supporting Information). Typical injections (175 V cm-1 for 2.0 s) correspond to the loading of ~ 0.1 nmol of charge. The amount of each ion injected depends on both the concentration and mobility of the ion, as well as the concentration and mobility of all other ions in the sample. Our quantitative treatment of the data accounts for the bias in sampling caused by electrokinetic injection (discussed in the Results section and Supporting Information).65-66

Sample Preparation

The procedure described above results in the injection and detection of all anions in a sample, not just (Pi)n.67 Analysis of (Pi)n by indirect detection works best on samples that are free of salts besides (Pi)n. The following steps were useful for preparing samples free of additional anions, originating from either synthesis or preparative separation: i) adsorption to anion-exchange resin; ii) elution of anions other than (Pi)n (e.g., Cl-) with 0.1 M Na2CO3; iii) elution of (Pi)n with concentrated 2.0 M NH4HCO3; iv) removal of NH4HCO3 under vacuum.

Analytes: Authentic standards of (Pi)n (n = 1 – 10)

Commercially available oligophosphates of a single chain length are limited to P1, P2, P3, and cyclo-P3. Preparative-scale separation of a mixture of (Pi)n (117% polyphosphoric acid69), by anion-exchange chromatography, generated samples of purified (Pi)n, n = 4 - 10, that served as analytical standards. By analyzing standards added to mixtures of (Pi)n, we identified peaks representing (Pi)n with n = 1 - 10.

Analytes: Mixtures of (Pi)n

We demonstrated the resolution of the method by analyzing commercially available samples of higher (Pi)n, covering a range of average length (n̄): 117% polyphosphoric acid, n̄ = 17, n̄ = 21, n̄ = 48, n̄ = 65. In addition, we prepared samples of (Pi)n generated by the dehydration of NH4H2PO4 in mixtures with urea, using conditions reported by Orgel et.al., which were presumed to be plausible in prebiotic chemistry.8,21

Internal standard(s)

We added an internal standard, K+CH3SO3-, to each sample prior to analysis by CE. The peak observed for the internal standard allowed us to: i) monitor the reproducibility of the method; ii) define the mobilities of (Pi)n relative to that of a standard (CH3SO3-) (μP,rel in Eq. 1); iii) determine the relative and absolute concentrations of resolved (Pi)n, by comparing the areas of peaks for CH3SO3- and (Pi)n. In Eq. 1, tCH3SO3- and tP are the retention times for CH3SO3- and (Pi)n; V is the voltage, L1 is the length of the capillary, and L2 is the distance between the inlet and detector.

μP,rel=μP−μCH3SO3−=L1L2V(1tCH3SO3−−1tP) (1)

The electrophoretic mobility of CH3SO3- in free solution was near that of (Pi)n, but peaks for CH3SO3- and (Pi)n did not overlap (μCH3SO3- = 26.7 cm2 kV-1 min-1). For experiments requiring additional internal standards, we also used Cl-, CF3CO2-, and CH3-C6H4-SO3-.

RESULTS & DISCUSSION

Resolution and Detection of Lower Oligophosphates by CZE

We analyzed mixtures of P1, P2, P3, and cyclo-P3 by CZE, using coated capillaries and running buffer composed of 3.0 mM TP2- and 18.0 mM bis-tris (pH = 6.9) or 24.0 mM tris (pH = 8.4);70 data from these experiments are shown in the Supporting Information. The results demonstrated i) the indirect detection of anions, enabled by the steady-state absorbance signal of TP2- at 254 nm, and ii) mobilities in the order cyclo-P3 > P3 > P2 > P1 (at pH = 6.9) that correspond to the number of ionizable groups, values of pKa, and structures of these (Pi)n (i.e., the radius of cyclo-P3 is constrained in a way that P3 is not).

Resolution of Mixtures of (Pi)n by CGE

The resolution of mixtures of (Pi)n with n > 5 required the use of a sieving gel. Best results were obtained with capillaries filled with solutions of 9.1% PDMA (w/v; average molecular weight 58.9 kDa) in running buffer (24.0 mM tris, 3.0 mM terephthalic acid, pH = 8.4). We filled capillaries (100 μm internal diameter and 57 cm in length) by pumping in solutions of PDMA with positive pressure (30 psi of ultra-high purity N2 applied to the inlet) for 10 minutes. During electrophoresis, the ends of the capillary were immersed into reservoirs of running buffer containing 9% PEG (w/v; average molecular weight of 1.5 kDa).

The upper trace in Figure 1A shows the separation of (Pi)n in a commercially available mixture of sodium polyphosphate (reported chain length of n̄ ~17). The sample had a concentration of 19.0 mM (in phosphate residues); P3 (2.0 mM) and CH3SO3- (2.0 mM) were added to serve as internal standards. For comparison, the lower trace in Fig. 1A shows the analysis of the same sample by CZE.

Figure 1. Resolution of (Pi)n by CGE.

Figure 1

Figure 1

Open in a new tab

A) Analysis of sodium polyphosphate (average chain length 17) in capillaries filled with 9.1% PDMA (w/v) gel (upper trace) or running buffer alone (24.0 mM tris, 3.0 mM terephthalic acid, pH = 8.4, 25 °C) (lower trace). Electrophoresis was performed by applying 14.6 kV across capillaries having a length of 57 cm (50 cm from the inlet to detector). B),C) Expanded views of the upper and lower traces in 1A), respectively.

We identified peaks for P1, P2, and P3 (marked with a dotted line in Fig. 1B) by comparing the traces in Fig. 1A to those collected for samples containing standards added to the mixture. In CGE experiments, mobilities of (Pi)n with n > 3 are lower than that of CH3SO3-. In CZE, the mobilities of all (Pi)n are greater than that of CH3SO3-. The contrast in the order of mobility indicates size-sieving during CGE, and the separation of (Pi)n in order of n. The series of peaks detected by CGE (shown in Fig.1B) suggests species as large as ~P35 in the sample.

Preparation of (Pi)n, n = 4 - 10

Identification of peaks observed for broad distributions of (Pi)n required oligophosphate standards that are not commercially available. We separated a mixture of (Pi)n (n ~ 1 - 10, derived from 117% polyphosphoric acid69) on an anion-exchange chromatography column (Cl- form) by elution with a gradient of KCl(aq). A second application of anion-exchange chromatography removed KCl from fractions containing (Pi)n (HCO3- form; elution withNH4HCO3). Removal of NH4HCO3 by vacuum generated oligophosphate standards as the ammonium salt.

CZE traces analyzing purified (Pi)n, n = 4 - 10, show one major peak in each sample (Fig. 2). Small peaks next to the major peak and a small peak for P1 (top trace in Fig. 2 for reference) suggest small amounts of (Pi)(n + 1) or (Pi)(n - 1), and P1. The relatively clean traces suggest that the conditions used in the preparation of samples do not cause extensive hydrolysis of (Pi)n or equilibration in chain length. Characterization of oligophosphate standards by 31P NMR is available in the Supporting Information.

Figure 2. Standards of (Pi)n, n = 1 - 10.

Figure 2

Open in a new tab

CZE traces analyzing a mixture of commercially available P1, P2, P3, cyclo-P3, and CH3SO3- (top trace) and analytical standards P4 - P10 purified by anion-exchange chromatography (no internal standard added). CZE was performed by applying 14.6 kV across coated capillaries (57 cm in length, 50 cm from the inlet to detector) filled with running buffer composed of 18.0 mM bis-tris and 3.0 mM terephthalic acid (pH = 6.9, 25 °C).

Identification of Peaks for n = 3 - 10 in Mixtures of (Pi)n

Figure 3 shows CGE data analyzing mixtures of (Pi)n (19.0 mM in phosphate residues; n̄ ~ 17) and added P3, P4, P6, P8, or P10 (2.0 mM). The x-axis of traces in Fig. 3 are in units of mobility relative to CH3SO3- μp,rel (Eq. 1). The traces show the alignment of peaks from run to run, and allowed us to assign peaks with increased area to specific (Pi)n (dotted lines in Fig. 3). The dotted line for n = 20 in Fig. 3B is based on the reasonable assumption that the peaks continue in order of n.71

Figure 3. Identification of P3 – P10 in mixtures of (Pi)i by CGE.

Figure 3

Open in a new tab

Traces are for the analysis of mixtures of (Pi)n (19 mM in total phosphate, pH ~ 8) with the addition of analytical standards for (Pi)n (2 mM). Mixtures were resolved by CGE at 14.6 kV using capillaries filled with 9.1 % PDMA (w/v) and running buffer (24.0 mM tris, 3.0 mM TP2-, pH = 8.4). Dotted lines mark peaks for species identified by added standards. Traces for samples with added P7 or P9 are available in the Supporting Information.

Mixtures of (Pi)n with n up to ~70

The CGE data in Fig. 4 characterizes commercially available mixtures of (Pi)n with different distributions in chain length. Samples analyzed in traces A-D are in order of increasing average chain length. The results show that peaks with n up to ~70 can be resolved in a single run; longer (Pi)n may potentially be resolved using lower concentrations of PDMA.

Figure 4. Resolution of commercially available mixtures of (Pi)n.

Figure 4

Open in a new tab

Samples consisting of commercially available mixtures of polyphosphate glass (20 mM in total phosphate, Na+ salt) and CH3SO3- (2 mM) were analyzed by CGE. After electrokinetic injection (4.0 s at 10 kV), (Pi)n were separated by electrophoresis at 14.6 kV in coated capillaries (57 cm in length, 50 cm between inlet and detector) filled with solutions of 9.1% PDMA (w/v) and running buffer (24.0 mM tris, 3.0 mM TP2-, pH = 8.4). Peaks in trace C, for n ~ 10 - 40, appear less sharp than the peaks in other runs. The reason for this difference is unclear; one possibility is that the peaks are affected by irregularities in pH within capillaries during separation, caused by a small mismatch in pH between samples (~ 8) and the running buffer (pH = 8.4).72

Shapes of Peaks

The asymmetric peaks observed in both CZE and CGE experiments are typical for electropherograms collected by indirect UV absorbance. (Pi)n with mobility greater than that of TP2- show up as peaks broadened to the left, while (Pi)n with mobility less than that of TP2- show up as peaks broadened to the right. These shapes are the results of dispersion by electromigration, originating from i) differences in mobility between analytes and TP2-, and ii) non-uniform electric fields inside sample zones. Discussion of the origin of the shapes of peaks is available in the Supporting Information, as well as in Ref. 59.

Quantitative Analysis of (Pi)n: Areas of Peaks

Three contributions determine the area of a peak: i) the response of [TP2-] to (Pi)n; ii) the amount of (Pi)n transferred from the sample to the capillary by electrokinetic injection; iii) residence time of the analyte passing the detector. Analysis of the areas of peaks for (Pi)n and CH3SO3- enables the quantification of resolved (Pi)n. To account for the effect of iii), areas of peaks are adjusted by the factor (1 / ti), where t is the retention time of analyte i.73 In Eq. 2, the ratio of adjusted areas for analyte i and CH3SO3- (Ai and ACH3SO3-) is related to the concentrations [i]S and [CH3SO3-] of the sample; zi and zCH3SO3- are the electrostatic charge of i and CH3SO3-; μCH3SO3- and μi are the mobilities of i and CH3SO3-; μC+ is the mobility of the cation in the sample zone (bis-tris-H+ or tris-H+).

AiACH3SO3−=[i]S[CH3SO3−]⋅(zizCH3SO3−)⋅(μC++μiμC++μCH3SO3−) (2)

The Supporting Information contains our derivation of Eq. 2 and a discussion of the quantitative aspects of electrokinetic injection and indirect UV absorbance.

Equation 2 reveals the advantage of analyzing electropherograms by comparing peaks for (Pi)n and CH3SO3-. The ratio (Ai / ACH3SO3-) depends on the ratio ([i]S / [CH3SO3-]) but does not depend on the concentration of other ions in the sample, or the voltage and duration of electrokinetic injection.

Calibration of peak areas to concentrations of P1, P2, P3 and cyclo-P3

Analysis of samples of containing P1, P2, P3, or cyclo-P3, and added CH3SO3- by CZE demonstrated the relationship between peak areas and concentrations. Fig. 5 shows values of (Acyclo–P3 / ACH3SO3-) determined from the analysis of samples containing cyclo-P3 and CH3SO3-. Samples had ratios of cyclo-P3 to CH3SO3- that spanned a factor of 104. The results are consistent with a linear dependence of (Ai / ACH3SO3- on ([i]S / [CH3SO3-]). The dotted line with slope m was obtained by fitting to the points with adjusted weighting.74 A value of 0.98 for the slope m is consistent with the complete ionization of all three acidic groups of cyclo-P3 at pH = 6.8.

Figure 5. Quantitative Calibration of cyclo-P3.

Figure 5

Open in a new tab

The plot shows values for (Acyclo–P3 / ACH3SO3-) determined in the analysis of cyclo-P3 CH3SO3- by CZE, using coated capillaries (57 cm in length, 50 cm between inlet and detector) filled with running buffer (18.0 mM bis-tris, 3.0 mM TP2-, pH = 6.9). Points for ([cyclo-P3] / [CH3SO3-]) = 0.1, 1.0, or 10.0 are the average taken from eight experiments. Standard deviations for (Acyclo–P3 / ACH3SO3-) are < 5% of the measured values; error bars are not visible in plots with a logarithmic scale. Points at other values of ([cyclo-P3] / [CH3SO3-]) are from one or two trials. The slope m is for the line determined by fitting the data with adjusted weighting.74

The calibration of ratios of peak areas to ratios of concentrations (of cyclo-P3 and CH3SO3-) provides a way to measure the amount of cyclo-P3 in mixtures. Data collected for the quantitative analysis of P1, P2, and P3 are available in the Supporting Information. Calibration of peak areas for (Pi)n of greater chain length should also be straightforward in CGE and enable the quantification of resolved (Pi)n in mixtures.

Analysis of (Pi)n Generated by Dehydrating Pi

The method developed in this paper should be useful for exploring reactions that may have generated (Pi)n on the prebiotic earth. CGE allowed us to characterize samples prepared by dehydrating Pi in mixtures with urea, using conditions originally reported by Orgel.8,21,75 The mechanism of dehydration by urea has not been determined. Urea is however essential; heating NH4H2PO4 in the absence of urea only generates a small amount of P2.

We synthesized (Pi)n by evaporating solutions of NH4H2PO4 (0.2 mmol) and urea (20.0 mmol), and heating to temperatures ≥ 125 °C. After 48 hours of heating mixtures to either 125 °C or 140 °C, we cooled the samples to 25 °C, added water (2.0 mL), and removed insoluble material by centrifugation. Analysis of the samples by CGE (data in Fig. 6) showed the formation of (Pi)n with n > 40, and distinguished the composition of mixtures generated at 125 °C and 140 °C. Cyclo-P3 is the most abundant species in mixtures prepared at 140 °C. In contrast, the mixture prepared at 125 °C does not contain cyclo-P3, despite containing linear (Pi)n with n > 40. The reason for the preferential formation of cyclo-P3 over linear (Pi)n at 140 °C is not clear; the synthesis of cyclo-P3 is however potentially important for prebiotic chemistry. Reactions of (Pi)n with -OH groups, leading to polyphosphorylated compounds, likely depend on whether (Pi)n are cyclic or linear. Reactions of linear (Pi)n potentially transfer phosphate residues from either terminal or middle positions of the chain, while reactions of cyclo-P3 are ring-opening and can generate a triphosphate group similar to that of ATP. The resolution and convenience of CGE should enable a broad survey of conditions for the synthesis of (Pi)n and for reactions of (Pi)n with -OH groups. The results should be helpful in refining hypotheses for the importance of dehydrated phosphate in the chemical origins of life.

Figure 6. Analysis of (Pi)n Generated under Possible Prebiotic Conditions.

Figure 6

Open in a new tab

A) Samples of (Pi)n, generated by heating mixtures of NH4H2PO4 and urea, were 20.0 mM in total phosphate with pH ~ 7. Analysis by CGE used coated capillaries filled with solutions of PDMA (9.1% w/v) and running buffer (3.0 mM TP2- and 24.0 mM tris, pH = 8.4); the protocol for CGE was the same used to collect the data in Fig. 3. B) Close-up views of peaks observed in the range μ = 3 - 9 cm2 kV-1 min-1. Peaks marked with a * have not been identified. The reason for the irregular shapes of peaks for n ~ 10 - 15 in the analysis of (Pi)n synthesized at 125 °C is unclear, but is probably related to the preparation of the specific sample (peaks for n ~ 5 - 20 appear sharp in the analysis of other mixtures using the same procedure (Figs. 1 and 3)). One possible reason for peak broadening in some parts of the trace, but not others, is a non-uniform pH in the capillary during separation, resulting in electrodispersion. Irregularities in pH may originate from either i) a mismatch in pH between running buffer and sample, or ii) bias towards the injection of the fully deprotonated form of (Pi)n, resulting in a pH for the sample zone that is different from the pH of the sample.

CONCLUSION

Previously reported methods using electrophoresis (slab gels or CGE) to analyze (Pi)n successfully resolved species on the basis of their size (n). These methods provided a way to qualitatively characterize mixtures of (Pi)n at high resolution, but involved difficult and time-consuming experiments with limited reproducibility. The method wehave demonstrated in this paper exploits the sieving properties of low-viscosity solutions of PDMA, and is capable of high resolution and quantitative analysis. The advantages of CGE using solutions of PDMA as the separation medium are convenient preparation, rapid analysis, and reproducibility (enabled by refilling capillaries with fresh solutions of gel). Our identification of P1-P10 and resolution up to P70 validates a method that will be useful in studies of (Pi)n relevant to prebioitic chemistry and biochemistry.

In addition to characterizing the composition of samples of (Pi)n, the method we have developed is potentially useful for analyzing other anionic, oligomeric species without a sensitive chromophore. Examples relevant to prebiotic chemistry are oligomers of phosphate condensed with organic compounds (e.g., structures with formula (PO3--RO)n or (P2O62--RO)n)). Examples of biological polymers include teichoic acid, hyaluronic acid, and sulfated polysaccharides such as heparin and chondroitin sulfate.

Supplementary Material

1_si_001

ACKNOWLEDGEMENTS

We acknowledge the Harvard University Origins of Life Initiative for research support. We thank Douglas B. Weibel, Katherine L. Gudiksen, Paul J. Bracher, and Dosil Pereira de Jesus for technical assistance and helpful discussion. A. Lee acknowledges salary support from NIH GM051559.

Footnotes

SUPPORTING INFORMATION AVAILABLE.

Additional information as noted in the text, as well as information about experimental procedures and sources of chemicals, is available free of charge at http://pubs.acs.org.

REFERENCES

  • 1.Clark JE, Wood HG. Anal. Biochem. 1987;161:280–290. doi: 10.1016/0003-2697(87)90452-0. [DOI] [PubMed] [Google Scholar]
  • 2.Joyce GF. Nature. 1989;338:217–224. doi: 10.1038/338217a0. [DOI] [PubMed] [Google Scholar]
  • 3.Joyce GF. Nature. 2002;418:214–221. doi: 10.1038/418214a. [DOI] [PubMed] [Google Scholar]
  • 4.Gilbert W. Nature. 1986;319:618–618. [Google Scholar]
  • 5.Keefe AD, Miller SL. Origins Life Evol. Biosphere. 1996;26:15–25. doi: 10.1007/BF01808157. [DOI] [PubMed] [Google Scholar]
  • 6.Keefe AD, Miller SL. J. Mol. Evol. 1995;41:693–702. doi: 10.1007/BF00173147. [DOI] [PubMed] [Google Scholar]
  • 7.Pasek MA, Kee TP, Bryant DE, Pavlov AA, Lunine JI. Angew. Chem. Int. Ed. Engl. 2008;47:7918–7920. doi: 10.1002/anie.200802145. [DOI] [PubMed] [Google Scholar]
  • 8.Osterberg R, Orgel LE, Lohrmann R. J. Mol. Evol. 1973;2:231–234. doi: 10.1007/BF01654004. [DOI] [PubMed] [Google Scholar]
  • 9.Rao NN, Gomez-Garcia MR, Kornberg A. Annu. Rev. Biochem. 2009;78:605–647. doi: 10.1146/annurev.biochem.77.083007.093039. [DOI] [PubMed] [Google Scholar]
  • 10.Kornberg A. J. Bacteriol. 1995;177:491–496. doi: 10.1128/jb.177.3.491-496.1995. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Kornberg A, Rao NN, Ault-Riche D. Annu. Rev. Biochem. 1999;68:89–125. doi: 10.1146/annurev.biochem.68.1.89. [DOI] [PubMed] [Google Scholar]
  • 12.Orgel LE, Lohrmann R. Acc. Chem. Res. 1974;7:368–377. [Google Scholar]
  • 13.Lohrmann R, Orgel LE. Nature. 1973;244:418–420. doi: 10.1038/244418a0. [DOI] [PubMed] [Google Scholar]
  • 14.Phosphoric Acid and Phosphates . Kirk-Othmer Encyclopedia of Chemical Technology. 4th ed. Vol. 18. John Wiley & Sons; New York: 1991. pp. 669–718. [Google Scholar]
  • 15.Ogawa N, DeRisi J, Brown PO. Mol. Biol. Cell. 2000;11:4309–4321. doi: 10.1091/mbc.11.12.4309. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Robinson NA, Wood HG. J. Biol. Chem. 1986;261:4481–4485. [PubMed] [Google Scholar]
  • 17.Gomez-Garcia MR, Kornberg A. Proc. Natl. Acad. Sci. USA. 2004;101:15876–15880. doi: 10.1073/pnas.0406923101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Baluyot ES, Hartford CG. J. Chromatogr. A. 1996;739:217–222. [Google Scholar]
  • 19.Sekiguchi Y, Matsunaga A, Yamamoto A, Inoue Y. J. Chromatogr. A. 2000;881:639–644. doi: 10.1016/s0021-9673(99)01278-9. [DOI] [PubMed] [Google Scholar]
  • 20.Greenfield S, Clift M. Analytical Chemistry of the Condensed Phosphates. 1 ed. Pergamon Press; New York: 1975. [Google Scholar]
  • 21.Osterberg R, Orgel LE. J. Mol. Evol. 1972;1:241–250. doi: 10.1007/BF01660243. [DOI] [PubMed] [Google Scholar]
  • 22.Rao NN, Roberts MF, Torriani A. J. Bacteriol. 1985;162:242–247. doi: 10.1128/jb.162.1.242-247.1985. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Gard DR, Burquin JC, Gard JK. Anal. Chem. 1992;64:557–561. [Google Scholar]
  • 24.Crutchfield MM, Callis CF, Irani RR, Roth GC. Inorg. Chem. 1962;1:813–817. [Google Scholar]
  • 25.Teleman A, Richard P, Toivari M, Penttilla M. Anal. Biochem. 1999;272:71–79. doi: 10.1006/abio.1999.4165. [DOI] [PubMed] [Google Scholar]
  • 26.Moreno B, Urbina JA, Oldfield E, Bailey BN, Rodrigues CO, Docampo R. J. Biol. Chem. 2000;275:28356–28362. doi: 10.1074/jbc.M003893200. [DOI] [PubMed] [Google Scholar]
  • 27.Glonek T, Lunde M, Mudgett M, Myers TC. Arch. Biochem. Biophys. 1971;142:508–513. doi: 10.1016/0003-9861(71)90514-5. [DOI] [PubMed] [Google Scholar]
  • 28.Choi BK, Hercules DM, Houalla M. Anal. Chem. 2000;72:5087–5091. doi: 10.1021/ac000044q. [DOI] [PubMed] [Google Scholar]
  • 29.Wang TL, Li SFY. J. Chromatogr. A. 1998;802:159–165. [Google Scholar]
  • 30.Heller C. Electrophoresis. 1999;20:1962–1977. doi: 10.1002/(SICI)1522-2683(19990701)20:10<1962::AID-ELPS1962>3.0.CO;2-H. [DOI] [PubMed] [Google Scholar]
  • 31.Heller C. Electrophoresis. 2001;22:629–643. doi: 10.1002/1522-2683(200102)22:4<629::AID-ELPS629>3.0.CO;2-S. [DOI] [PubMed] [Google Scholar]
  • 32.Kopecka K, Drouin G, Slater GW. Electrophoresis. 2004;25:2177–2185. doi: 10.1002/elps.200305951. [DOI] [PubMed] [Google Scholar]
  • 33.Karger BL, Guttman A. Electrophoresis. 2009;30:S196–S202. doi: 10.1002/elps.200900218. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Dovichi NJ, Zhang JZ. Angew. Chem.-Int. Edit. 2000;39:4463–4468. doi: 10.1002/1521-3773(20001215)39:24<4463::aid-anie4463>3.0.co;2-8. [DOI] [PubMed] [Google Scholar]
  • 35.Wang TL, Li FY. J. Chromatogr. A. 1999;834:233–241. [Google Scholar]
  • 36.Stover FS. J. Chromatogr. A. 1999;834:243–256. [Google Scholar]
  • 37.Polyacrylamide-filled capillaries resulted in sharply decreasing current and increasing retention times within 40 min., and permitted only ~ 3 runs per capillary.
  • 38.Ruiz-martinez MC, Berka J, Belenkii A, Foret F, Miller AW, Karger BL. Anal. Chem. 1993;65:2851–2858. doi: 10.1021/ac00068a023. [DOI] [PubMed] [Google Scholar]
  • 39.Madabhushi RS. Electrophoresis. 1998;19:224–230. doi: 10.1002/elps.1150190215. [DOI] [PubMed] [Google Scholar]
  • 40.Vegvari A, Hjerten S. J. Chromatogr. A. 2002;960:221–227. doi: 10.1016/s0021-9673(01)01436-4. [DOI] [PubMed] [Google Scholar]
  • 41.Stover FS. J. Chromatogr. A. 1997;769:349–351. [Google Scholar]
  • 42.Cretich M, Chiari M, Pirri G, Crippa A. Electrophoresis. 2005;26:1913–1919. doi: 10.1002/elps.200410368. [DOI] [PubMed] [Google Scholar]
  • 43.Heller C. Electrophoresis. 1999;20:1978–1986. doi: 10.1002/(SICI)1522-2683(19990701)20:10<1978::AID-ELPS1978>3.0.CO;2-G. [DOI] [PubMed] [Google Scholar]
  • 44.Wang YM, Liang DH, Ying OC, Chu B. Electrophoresis. 2005;26:126–136. doi: 10.1002/elps.200406162. [DOI] [PubMed] [Google Scholar]
  • 45.Rosenblum BB, Oaks F, Menchen S, Johnson B. Nuc. Acids Res. 1997;25:3925–3929. doi: 10.1093/nar/25.19.3925. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.He H, Buchholz BA, Kotler L, Miller AW, Barron AE, Karger BL. Electrophoresis. 2002;23:1421–1428. doi: 10.1002/1522-2683(200205)23:10<1421::AID-ELPS1421>3.0.CO;2-L. [DOI] [PubMed] [Google Scholar]
  • 47.Barbier V, Viovy JL. Curr. Opin. Biotechnol. 2003;14:51–57. doi: 10.1016/s0958-1669(02)00007-1. [DOI] [PubMed] [Google Scholar]
  • 48.Zhou HH, Miller AW, Sosic Z, Buchholz B, Barron AE, Kotler L, Karger BL. Anal. Chem. 2000;72:1045–1052. doi: 10.1021/ac991117c. [DOI] [PubMed] [Google Scholar]
  • 49.Kotler L, He H, Miller AW, Karger BL. Electrophoresis. 2002;23:3062–3070. doi: 10.1002/1522-2683(200209)23:17<3062::AID-ELPS3062>3.0.CO;2-G. [DOI] [PubMed] [Google Scholar]
  • 50.Gao QF, Yeung ES. Anal. Chem. 1998;70:1382–1388. doi: 10.1021/ac970999h. [DOI] [PubMed] [Google Scholar]
  • 51.Giovannoli C, Anfossi L, Tozzi C, Giraudi G, Vanni A. J. Sep. Sci. 2004;27:1551–1556. doi: 10.1002/jssc.200401797. [DOI] [PubMed] [Google Scholar]
  • 52.Fung EN, Pang HM, Yeung ES. J. Chromatogr. A. 1998;806:157–164. doi: 10.1016/s0021-9673(97)00812-1. [DOI] [PubMed] [Google Scholar]
  • 53.Huang MF, Huang CC, Chang HT. Electrophoresis. 2003;24:2896–2902. doi: 10.1002/elps.200305432. [DOI] [PubMed] [Google Scholar]
  • 54.Shamsi SA, Danielson ND. Anal. Chem. 1995;67:1845–1852. [Google Scholar]
  • 55.Wang PG, Giesel RW. Anal. Biochem. 1995;230:329–332. doi: 10.1006/abio.1995.1482. [DOI] [PubMed] [Google Scholar]
  • 56.Wang PG, Giese RW. Anal. Chem. 1993;65:3518–3520. [Google Scholar]
  • 57.Smith RM, Martell AE. Critical Stability Constants. Plenum Press; New York: 1989. [Google Scholar]
  • 58.Smith RM, Martel AE. NIST Critically Selected Stability Constants of Metal Complexes. National Institute of Standards and Technology; Gaithersburg, MD: 1997. [Google Scholar]
  • 59.Poppe H, Xu X. In: High Performance Capillary Electrophoresis: Theory, Techniques, and Applications. 1 ed. Khaledi M, editor. Vol. 146. Wiley & Sons; New York: 1998. [Google Scholar]
  • 60.Gas B, Kenndler E. Electrophoresis. 2004;25:3901–3912. doi: 10.1002/elps.200406159. [DOI] [PubMed] [Google Scholar]
  • 61.Beckers JL, Bocek P. Electrophoresis. 2003;24:518–535. doi: 10.1002/elps.200390060. [DOI] [PubMed] [Google Scholar]
  • 62.Bruin GJM, Vanasten AC, Xu XM, Poppe H. J. Chromatogr. 1992;608:97–107. [Google Scholar]
  • 63.Macka M, Haddad PR, Gebauer P, Bocek P. Electrophoresis. 1997;18:1998–2007. doi: 10.1002/elps.1150181120. [DOI] [PubMed] [Google Scholar]
  • 64.We used PEG instead of PDMA in buffer reservoirs because commercial PDMA (~$100 per g) is much more expensive than PEG (~ $0.1 per g); each set of experiments requires ~ 1 g of polymer; PDMA can be obtained inexpensively however, by preparing it from N,N-dimethylacrylamide.
  • 65.Huang XH, Gordon MJ, Zare RN. Anal. Chem. 1988;60:375–377. [Google Scholar]
  • 66.Rose DJ, Jorgenson JW. Anal. Chem. 1988;60:642–648. [Google Scholar]
  • 67.Satow T, Machida A, Funakushi K, Palmieri R. J. High. Resolut. Chromatogr. 1991;14:276–279. [Google Scholar]
  • 68.Cohn WE, Bollum FJ. Biochim. Biophys. Acta. 1961;48:588–590. doi: 10.1016/0006-3002(61)90058-0. [DOI] [PubMed] [Google Scholar]
  • 69.The nomenclature (i.e., 117%) is based on comparing the ratio of phosphorus to oxygen in samples of dehydrated phosphate to the ratio of a standard solution of 85% phosphoric acid.
  • 70.Capillaries modified by the method of Cretich (Ref. 42) were used over the course of several months in more than 100 runs, without observable change in the retention times for analytical standards.
  • 71.The alignment is not perfect; small differences from run to run are possibly due to dispersion in electromigration, caused by differences between samples in pH or ionic strength.
  • 72.Another possibility is that electrokinetic injections may result in concentrations of (Pi)n at the inlet that can lead to precipitation.
  • 73.Hilser VJ, Freire E. Anal. Biochem. 1995;224:465–485. doi: 10.1006/abio.1995.1075. [DOI] [PubMed] [Google Scholar]
  • 74.Values of the points span a factor of 104. Errors are correlated with the independent variable (ratio of concentrations in the sample). A line fit to these points is determined mostly by the value and error for the highest ratio tested, 102. We obtained a better fit by linear regression of log (Ai / ACH3SO3- ) to log ([i]S / [CH3SO3-]) with a slope constrained to a value of one. Fitting the equation log y = log m + log x to the data provides a more accurate result for points that are separated by orders of magnitude than fitting to the equation y = mx (the form of Eq. 2); y is (Ai / ACH3SO3- ), x is ([i]S / [CH3SO3-]), and m is a proportionality constant obtained by fitting to the data.
  • 75.Handschuh G, Lohrmann R, Orgel LE. J. Mol. Evol. 1973;2:251–262. doi: 10.1007/BF01654094. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1_si_001
NIHMS225465-supplement-1_si_001.pdf (1.1MB, pdf)

RESOURCES