Hydroxymethanesulfonate from Volcanic Sulfur Dioxide: A "Mineral" Reservoir for Formaldehyde and Other Simple Carbohydrates in Prebiotic Chemistry

Abstract

While formaldehyde (HCHO) was likely generated in Earth's prebiotic atmosphere by ultraviolet light, electrical discharge, and/or volcano-created lightning, HCHO could not have accumulated in substantial amounts in prebiotic environments, including those needed for prebiotic processes that generate nucleosidic carbohydrates. HCHO at high concentrations in alkaline solutions self-reacts in the Cannizzaro reaction to give methanol and formate, neither having prebiotic value. Here, we explore the possibility that volcanic sulfur dioxide (SO₂) might have generated a reservoir for Hadean HCHO by a reversible reaction with HCHO to give hydroxymethanesulfonate (HMS). We show that salts of HMS are stable as solids at 90°C and do not react with themselves in solution, even at high (>8 M) concentrations. This makes them effective stores of HCHO, since the reverse reaction slowly delivers HCHO back into an environment where it can participate in prebiotically useful reactions. Specifically, we show that in alkaline borate solutions, HCHO derived from HMS allows formation of borate-stabilized carbohydrates as effectively as free HCHO, without losing material to Cannizzaro products. Further, we show that SO₂ can perform similar roles for glycolaldehyde and glyceraldehyde, two intrinsically unstable carbohydrates that are needed by various models as precursors for RNA building blocks. Zircons from the Hadean show that the Hadean mantle likely provided volcanic SO₂ at rates at least as great as the rates of atmospheric HCHO generation, making the formation of Hadean HMS essentially unavoidable. Thus, hydroxymethylsulfonate adducts of formaldehyde, glycolaldehyde, and glyceraldehyde, including the less soluble barium, strontium, and calcium salts, are likely candidates for prebiotically useful organic minerals on early Earth.

Keywords: Borate; Cannizzaro reaction; Formaldehyde; Hydroxymethanesulfonate; Organic minerals; Origins of life; Prebiotic chemistry.; Sulfur dioxide.

FIG. 1.

Formaldehyde (HCHO) can capture enediol(ate)s before they react to give complex mixtures of unproductive species. However, at high concentrations, HCHO disproportionates in a bimolecular Cannizzaro reaction to yield unproductive methanol and formate. Both problems can be mitigated if an organic mineral can be found that slowly bleeds HCHO into a prebiotic reaction mixture.

FIG. 2.

Possible prebiotic reactions involving volcanic SO₂, HCHO, and other lower carbohydrates. Note how the reaction with ribose involves competition with ring closed forms, here the furanose. Therefore, the equilibrium concentration favors the bisulfite addition product with ribose considerably less than with HCHO, glycolaldehyde, or glyceraldehyde. The same is the case with ketoses. The stereochemistry shown is arbitrary.

FIG. 3.

Comparison by ¹³C NMR spectroscopy of the reaction of DHA with H¹³CHO “bled” into the reaction mixture from ¹³C-labeled HMS (top, signal at ∼74.5) or H¹³CHO added directly (bottom). ¹³C-HMS resonates at 74.2 ppm. Reactions were run in NaCO₃ (1.1 M all at 65°C) H₃BO₃ (0.278 M) buffer at pH 10.4. Note that the manifold of peaks arising from ¹³CH₂OH units in the branched pentoses at 60–65 ppm are essentially the same. Also note the presence of substantially more HCOOH (171 ppm, ionized to give formate) from the Cannizzaro reaction when free H¹³CHO is added directly. MeOH (49.5) was added as an internal standard. The pH-dependent signal from carbonate buffer is at ∼165 ppm.

FIG. 4.

¹H NMR spectra of 1:1 mixture of glycolaldehyde and sodium bisulfite, obtained in D₂O at 25°C. Glycolaldehyde has proton resonances at 4.9 ppm (triplet, C1 of the hydrate) and 3.3 ppm (doublet, C2 of the hydrate, data not shown). Both are missing in the spectrum, which has new resonances at 4.3 and 3.85 ppm (C1, adduct, split by C2 proton, two diastereomers) and 3.5 ppm (C2). Notice also the coupling due to the new stereogenenic center. Absence of detectable amounts of glycolaldehyde suggests that adduct had consumed more than 99% of the substrate.

FIG. 5.

¹³C NMR spectra of a 1:1 mixture of 1-¹³C-labeled D,L-glyceraldehyde and sodium bisulfite, obtained in D₂O at a pH of 4.76 at 25°C. The two resonances (84.3 and 82.4) are assigned to be the two diastereomeric adducts in a ratio of ∼47:53. Glyceraldehyde itself had a single resonance at 89.95 ppm (data not shown), which is assigned to the hydrated species. Again, undetectable of a resonance at 89.95 ppm suggests that adduct had consumed more than 99% of the substrate.

FIG. 6.

¹H NMR spectra of DHA plus sodium bisulfite (in a 1:1 mixture, 10 mM each), obtained in D₂O at 25°C. DHA alone has a proton resonance of 4.38 (with water set at 4.8 ppm; the peak at 3.7 is the hydrate of DHA). Here, the presence of free and adducted DHA allows an equilibrium constant to be calculated, at ≈1.6 × 10^-3M. The integral shows that the (2.57 = integral for total adduct protons, versus 1.3 = integral for total DHA proton, or 2:1 adduct:free DNA) yields an equilibrium constant [DHA][bisulfite]/[adduct] = 1.6 × 10⁻³ M.

FIG. 7.

¹H NMR spectra of ribose (top) and ribose plus sodium bisulfite (a 1:1 mixture, 10 mM each), obtained in D₂O (set at 4.8 ppm) at 25°C. Attention is drawn to the peaks at 4 4.0–4.92, 5.25, and 5.38, which are the alpha and beta isomers of the furanose and pyranose forms. The new resonances at 4.6 (doublet) and 4.67 (broad) are the bisulfite addition products (C1). Note that these represent perhaps only 20–30%% of the total. The estimated equilibrium constant [ribose][bisulfite]/[adduct] ≈ 2 × 10⁻² M.