Purines were measured by ultra high-performance liquid chromatography UPLC with photodiode array (PDA) ultraviolet detection. The system consisted of a Waters ACQUITY H-Class plus (CH-A) core, quaternary solvent manager, and PDA eLambda detector. Analytes were separated using reverse-phase chromatography on a Waters ACQUITY Premier HSS C18 VanGuard FIT Column, 1.8 µm, 2.1 mm × 100 mm. Elution was conducted at 0.5 mL/min with a stepped gradient of buffer A (10 mM ammonium acetate and 1.5 mM tetrabutylammonium phosphate, pH 5.0) and buffer B (10 mM ammonium phosphate, 1.5 mM TBAP, 25% acetonitrile, pH 7.0 before adding acetonitrile). The gradient consistent of the following sequence: 100% buffer A for 11 min; A linear gradient to 75% buffer B over 1 min, 3 min at 75% buffer B, a linear gradient to 100% buffer B over 4 min, 100% buffer B for 4 min, and a linear gradient to 0% buffer A over 30 s. The column was then re-equilibrated with 100% buffer A for 9.5 min prior to next run.

Reference standards for purines were obtained from Sigma Chemical Co. and prepared in distilled water. Purines in biological samples were identified by comparing their retention times and spectral profiles to known standards, and quantified at a detection wavelength of 254 nm. The injection volume was 5 μL. The column temperature was 40°C and autosampler was 4°C.

To assess performance in biological samples, samples from cultured human cells were prepared as previously described (Sutcliffe et al., 2021). The cells included 3 publicly available induced pluripotent stem cells (iPSCs) in the undifferentiated state, each with and without a c.508C>T stop codon mutation in the HPRT1 gene that causes complete loss of the enzyme (hypoxathine-guanine phosphoribosyltransferase) that is associated with the purine disorder Lesch-Nyhan disease (Fu et al., 2014). The iPSCs included KOLF2.0, NCRM1, and PGP-1. In brief, cultures were grown in 6 well plates to 60%–80% confluency for approximately 1 million cells. The culture medium was freshly replaced 24 h before sample collection.

To measure purines released into the culture medium, a 450 µL aliquot of the medium was mixed with 50 µL of 1 mol/L perchloric acid to produce a final concentration of 0.1 mol/L perchloric acid. The sample was then stored at −80°C. When all samples were ready, they were thawed on ice and a 200 µL aliquot centrifuged for 15 min at 4°C at 13,000 g. Then 12 µL of 3.5 mol/L K₂CO₃ was added and the sample was stored on ice to precipitate potassium perchlorate. The sample was then centrifuged again at 13,000 g for 15 min, and the supernatant further clarified with PVDF microcentrifuge filter tubes (5,000 g for 5 min). Injection volume was 5 µL.

To measure intracellular purines, all culture medium was removed, cells were released from the plate by trypsinization, pelleted at 500 g × 5 min, and the pellet was mixed with 200 µL 0.2 mol/L perchloric acid. At this point, samples could be stored at −80°C. When all samples were ready, they were thawed on ice. Cell pellets were sonicated at 50% amplitude and pulse 2 s on and 2 s off for 1 min using a Q500A cup horn sonicator with the chiller at 4°C (QSONICA, Newtown, CT). The sample was then centrifuged at 13,000 g for 15 min at 4°C. The supernatant containing purines was removed to a fresh tube, the pH neutralized with 11 µL of 3.5 mol/L K₂CO₃, and the sample left on ice for 1 h to precipitate potassium perchlorate. The sample was then centrifuged again at 13,000 g for 5 min, and the supernatant further clarified with PVDF microcentrifuge filter tubes (8,000 g for 3 min). Injection volume was 5 µL. Protein pellets were dissolved in 2% SDS, and the total protein concentration was determined with the Pierce BCA kit (Thermo Fisher Scientific, Rockford IL).

Analytes were identified using a combination of retention time and UV spectrum, and each was quantified by comparison to known standards. For intracellular purines, results were expressed as nmoles per mg total cellular protein, to adjust for any differences in cell numbers at harvest. For extracellular purines, results were corrected according to the volume of tissue culture medium and culture duration (24 h) and expressed as nmoles per mg protein per 24 h.

The energy state of the cells was estimated using the adenylate energy charge (AEC) calculated with the formula ([ATP] + 0.5 × [ADP])/([ATP] + [ADP] + [AMP]). The AEC varies between 0 and 1, with normal healthy cells usually having a value >0.80 (Oakhill et al., 2012). Results for each measure were expressed as average values ± SEM from 3 independent replicates. Comparisons between groups involved ANOVA with post hoc Tukey t-tests.

With a run time of 33 min per sample, the method was capable of resolving and quantifying the most abundant biologically relevant purine metabolites including ATP, ADP, AMP, adenosine, adenine, GTP, GDP, GMP, IMP, ZMP, guanosine, hypoxanthine, inosine, xanthine and uric acid (Figure 2).

Assessing each metabolite at 9 different concentrations confirmed that detection was linear in a range from 0.1 to 100 µM (Figure 3). With a 5 µL volume, all metabolites could be detected down to a level of approximately 100 nM.

To assess reproducibility the same sample was evaluated 5 times within a batch run, as well as in 3 separate batches on different days. The within-run coefficient of variance averaged 0.8% within a single batch, and the between-run variance averaged 1% for most analytes, confirming excellent reproducibility over time.

In keeping with prior studies of other mammalian cells (Shirley et al., 2007; Goettle et al., 2013; Lopez et al., 2020; Sutcliffe et al., 2021), intracellular adenine derivatives in the iPSCs were quantitatively more abundant than guanine derivatives, with ATP being found in greatest amounts (Table 1; Figure 4, top). The AEC fell between 0.81 and 0.88, confirming good preservation of nucleotide energy state. The variance (SEM) between biological samples depended on the total amount measured but averaged 10.7% of the measured value across all measurements.

Also in keeping with prior studies, control lines and their corresponding HPRT1 mutants showed small or no changes for most intracellular purines. Conditioned medium contained no detectable purines, except for hypoxanthine (Table 1; Figure 4 bottom). Although intracellular hypoxanthine was very low and not significantly different between control and mutant cell lines, extracellular hypoxanthine was elevated by an average of 30-fold in the HPRT1 mutants (range 12–225-fold depending on the cell line). This finding is expected since the mutations result in complete loss of the enzyme that recycles hypoxanthine, so cells cannot use it and instead dump it into the extracellular space.

Results for cellular purines were subject to 2-way MANOVA to account for potential relationships among the metabolites, followed by ANOVA for individual purine metabolites separately with a Benjamini–Hochberg correction for multiple comparisons (Table 1). The MANOVA showed significant main effects for both mutation status (controls vs. mutants; p < 0.01) and cell type (KOLF2.0 vs. NCRM1 vs. PGP1; p < 0.001), as well as a significant interaction between these variables (p < 0.05). For individual purine metabolites there were significant effects of mutation status for hypoxanthine in the medium (FDR = 1.3 × 10⁻⁷), as well as intracellular ADP (FDR = 0.0078) and AMP (FDR = 0.0049). There were also significant differences between cell types for ATP (FDR = 0.0016), AMP (FDR = 0.0021), GTP (FDR = 0.0011), GMP (FDR = 0.0011) and HX in medium (FDR = 0.0042).