Introduction

Carbonaceous chondrites are fragments of ancient asteroids that provide a record of the chemistry of the early solar system1,2. They contain a variety of organic compounds and their delivery to early Earth may have played an essential role in the chemistry that led to the origin of life3,4,5,6,7,8,9,10,11. Hydrogen cyanide (HCN) has been detected in water extracts of the Murchison meteorite upon acidification12, which was a surprising result because cyanide is a highly reactive compound and thought to have been completely consumed by reactions in the parent asteroid. Isotope ratio mass spectrometry indicated that this cyanide was extraterrestrial in origin; however, low 13C enrichment and low 15N enrichment of released cyanide in the Murchison meteorite suggested that this source of cyanide was separate from the cyanide responsible for the synthesis of some extraterrestrial amino acids and other organic compounds13. Furthermore, the compounds responsible for released cyanide in meteorites are still unknown.

Here, we report releasable cyanide abundance (often referred to as total cyanide) in acid-digested distillates of various meteorites by chemical derivatization (Supplementary Fig. 1) and liquid chromatography with fluorescence detection and time-of-flight mass spectrometry (LC-FD/ToF-MS). We identify two iron cyanocarbonyl complexes, [FeII(CN)5(CO)]3− and [FeII(CN)4(CO)2]2−, in the Lewis Cliff 85311 meteorite by liquid chromatography-high resolution orbitrap mass spectrometry. These extraterrestrial organometallic compounds are a source of free cyanide (HCN/CN−) and also bear a striking similarity to portions of the active sites of [NiFe]- and [FeFe]-hydrogeneses, which suggests that these compounds may have played an important role during the origin and early evolution of life on Earth.

Results

Abundance and species of cyanide in meteorites

We analyzed Allan Hills (ALH) 83100, Allan Hills 84001, Graves Nunataks (GRA) 06100, Lewis Cliff (LEW) 85311, Lewis Cliff 90500, Lonewolf Nunataks (LON) 94102, Murchison, and Roberts Massif (RBT) 04133 for acid-releasable cyanide. All of these meteorites are carbonaceous chondrites with the exception of ALH 84001, which is a martian meteorite (see Table 1 for classifications and additional information14,15,16). Figure 1 shows the unambiguous identification of cyanide in LEW 90500, a CM2 chondrite. There is a single peak at ~5.6 min. in the fluorescence chromatogram in addition to a single peak at ~5.7 min. in the extracted ion chromatogram (slight delay in retention time due to the mass spectrometer coming after the fluorescence detector), which is an exact match to our cyanide standard (in the form of an NDA-cyanide derivative). We also measured the same fluorescence and mass peaks corresponding to cyanide for the other CM chondrites analyzed. All five of the CM chondrites we analyzed contain cyanide with LEW 85311 containing the highest concentration at 2472 ± 38 nmol CN·g−1 meteorite (Table 1). The cyanide abundance for Murchison in our study (95 ± 1 nmol CN·g−1 meteorite) was noticeably lower than the abundance (~400 nmol CN·g−1 meteorite) previously reported by Pizzarello, although our extraction and analysis methods differed12. Method blanks were almost completely absent of cyanide resulting in a clean baseline (see Fig. 1 and Supplementary Note 1). We did not detect any (<0.1 nmol·g−1) releasable cyanide in the CR2 (heated), CV3, or martian meteorite (see Supplementary Note 2). Cyanide can be found in nature, but most terrestrial sources are anthropogenic and are derived from industrial processes17. The observation that cyanide was group-specific (CMs only) among the Antarctic meteorites analyzed suggests that its source was indigenous to the meteorite rather than a pervasive terrestrial contaminant. Furthermore, Antarctic meteorites tend to be less contaminated compared to meteorites collected elsewhere18,19,20 and cyanide abundances did not track with the meteorite’s weathering grade, i.e., meteorites that have experienced greater terrestrial weathering effects did not have greater cyanide abundances.

Table 1 Summary of cyanide abundances in meteorites
Full size table
Fig. 1
figure 1

Identification of cyanide released from CM chondrites. a Fluorescence chromatograms (λex 252 nm, λem 483 nm) of naphthalene-2,3-dicarboxaldehyde-cyanide derivative from the LEW 90500 meteorite, KCN standard, and method blank. b Extracted ion chromatograms (m/z 251.08 with a ±0.03 window; [M + H]+) of naphthalene-2,3-dicarboxaldehyde-cyanide derivative from the LEW 90500 meteorite, KCN standard, and method blank

Full size image

All carbonaceous chondrites in the CM group have experienced some degree of pre-terrestrial aqueous alteration21, which has been attributed to melting of ice inside asteroids from short-lived radionuclides (26Al), electromagnetic induction, and/or impact heating22. During this aqueous alteration stage, fluids in the asteroid were likely low to moderate in temperature (0–150 °C) and alkaline in pH based on mineralogy and oxygen isotope data23,24,25,26. The prior detection of releasable cyanide in the water-soluble portion of the Murchison meteorite12,13 suggested that this cyanide was not trapped in a mineral phase and would have been accessible to aqueous fluids in the meteorite parent body. Table 1 also shows CM chondrites associated with a numerical scale that estimates the degree of aqueous alteration (lower numbers mean more aqueously altered) from Alexander et al.16. There is a noticeable decrease in cyanide abundance for more aqueously altered CM chondrites. This trend suggests that the protracted aqueous alteration stage in the meteorite parent body may have altered or destroyed the compounds responsible for this cyanide, similar to previous studies involving the abundance of meteoritic amino acids and N-heterocycles7,20,27.

Meteoritic insoluble organic matter (IOM) is a potential source of cyanide in CM chondrites because it represents the majority of organic matter in meteorites and is known to contain nitriles28. However, we measured a similar concentration of released cyanide in the Murchison meteorite to those measured in previous studies of hot water extracts12,13, the latter result indicates that the cyanide species would be water-soluble, which would rule out IOM as a source. If there was a high concentration of cyanide in meteorite parent bodies, cyanide may be present in polymeric form (we do not consider HCN polymers to be part of the IOM because they can be extracted in water yielding a yellow–brown solution29). However, it is unlikely that the cyanide detected in this study originated mainly from HCN polymer because the total cyanide liberated from the HCN polymer using our methods is only ~0.02% (wt/wt). The meteorite sample would need to be almost entirely HCN polymer in order to release the levels of cyanide we detect in LEW 85311 (see Supplementary Note 3). Organonitrile compounds do not appear to be a significant source of releasable cyanide in CM chondrites because they release little to no detectable cyanide when analyzed using our method (see Supplementary Note 4).

It is widely believed that many of the soluble organic compounds in carbonaceous chondrites were synthesized during the aqueous alteration stage(s) that occurred in the parent body, which also resulted in changes to the mineral composition21. Reaction-path calculations simulating aqueous alteration in CM parent bodies suggest that metal-organic compounds (along with carboxylic acids and amides, which agrees with experimental measurements of CM chondrites3) should dominate the composition of organic compounds30. Recently, dihydroxymagnesium carboxylates [(OH)2MgO2C-R]− were reported in the soluble organic fraction of various meteorites31.

We analyzed both water and base extracts of LEW 85311 by inductively coupled plasma mass spectrometry in order to help elucidate potential speciation of cyanide in these meteorites. Both types of extractions had relatively similar element contents with high concentrations of Mg, Na, Ca, K, and Al. Appreciable amounts of Fe, Ni, Ag, and Pt were also detected (Supplementary Table 1). Many of these elements are known to form simple cyanide salts such as NaCN and KCN; however, simple cyanides are only a minor contributor to the overall acid-releasable cyanide based on the analysis of LEW 90500 (see Supplementary Note 5). Iron forms particularly stable complexes with cyanide ligands due to cyanide’s strong σ-donor properties. Using negative-ion electrospray ionization (ESI) mass spectrometry, we observed changes in iron oxidation state, loss of cyanide ligands, and aggregation of counterions to reduce charge in ferricyanide [FeIII(CN)6]3− and ferrocyanide [FeII(CN)6]4− reference standards, which agrees with previous measurements of these complexes under ESI conditions32. A major identifying fragment for iron cyanide complexes is [FeII(CN)3]− at m/z 133.9447.

We identified this [FeII(CN)3]− anion in high resolution mass spectra of base extracts of LEW 85311 meteorite based on accurate mass measurements and its distinct isotope pattern including a triplet peak with 13C, 15N, and 57Fe isotopologues (Fig. 2). This mass fragment was not associated with either ferrocyanide or ferricyanide, rather it resulted from the fragmentation of iron cyanocarbonyl complexes [FeII(CN)5(CO)]3− and [FeII(CN)4(CO)2]2− (Supplementary Fig. 2). [FeII(CN)3]− is also a known fragment ion for both [FeII(CN)5(CO)]3− and [FeII(CN)4(CO)2]2− in the negative-ion ESI mass spectrum33. We identified the parent ions [H2FeII(CN)5(CO)]− and [HFeII(CN)4(CO)2]− using accurate mass measurements, their distinctive isotope patterns, and the observation that a single peak for each complex was present in their respective extracted ion chromatogram (Fig. 2 and Supplementary Fig. 2). We estimate that ~70% of the released cyanide was derived from these two organometallic compounds assuming similar responses of m/z 133.9447 to ferrocyanide (see Supplementary Note 6). It is also noteworthy to point out that cyanide ligands stabilize CO binding to iron, which makes these iron cyanocarbonyl complexes very stable.

Fig. 2
figure 2

High resolution ESI mass spectra. a Mass spectrum of LEW 85311 meteorite extract along with simulated isotope pattern of [FeII(CN)3]−. b Mass spectrum of LEW 85311 meteorite extract along with simulated isotope pattern of [H2FeII(CN)5(CO)]−. c Mass spectrum of LEW 85311 meteorite extract along with simulated isotope pattern of [HFeII(CN)4(CO)2]−. Accurate mass measurements and isotope patterns support the identification of two iron cyanocarbonyl complexes and a shared fragment ion. H+ serves as counterions which reduce the overall charge of these species to −1. Simulated isotope patterns (in red) were generated using a Gaussian profile and a mass resolution of 65,000 resolution (full-width at half-maximum) in the Thermo Scientific XCalibur software

Full size image

[FeII(CN)5(CO)]3− and [FeII(CN)4(CO)2]2− (both cis and trans forms) have been synthesized in the laboratory from solutions containing ferrous salts (FeCl2), CN−, and CO33, and similar synthetic routes were likely available during the aqueous alteration stage(s) on asteroids30. Additionally, the discovery of iron cyanocarbonyl complexes in aqueously altered meteorites suggests that these organometallic compounds could have also been produced by geochemical processes on early Earth34,35,36,37. Thus, we emphasize that the study of meteorites may help identify unusual or overlooked geochemical reactions (particularly those that are ancient or rarely seen today), as well as reveal new prebiotic geochemical compounds that were not actively being considered in scenarios relevant to the origin of life.

Discussion

Both exogeneous and endogenous iron cyanocarbonyl complexes may have been important for prebiotic chemistry on early Earth. In the case of meteorites, aqueous leaching or dissolution of CM chondrites followed by UV irradiation of iron cyanocarbonyl complexes would have led to the release of cyanide from their metal centers38. Shallow ponds, streams, or puddles may have been favorable environments for cyanide release because they would have allowed greater UV penetration depth with less dilution effects. Some of this cyanide may have escaped into the atmosphere, particularly if the solution was acidic. Thus, exogenous cyanide delivery via meteorites may have provided a boost to the cyanide abundance produced on early Earth34; see Supplementary Discussion on estimated cyanide abundance on early Earth. Reactions involving cyanide were likely critical for the synthesis of numerous biologically relevant compounds and were driven, in part, by photochemistry of organometallic compounds. For example, photochemical reduction of cyanohydrins by UV irradiation of iron cyanocarbonyl complexes in the presence of sulfite would have likely produced simple sugars and amino acid precursors since this process was previously demonstrated with ferrocyanide39,40 (and we would expect similar photochemical behavior).

On early Earth, molecular hydrogen (H2) was one of the earliest energy sources available and the ubiquity of hydrogen metabolism and deeply rooted lineages in both archaea and bacteria has led to the belief that hydrogen metabolism is ancient in origin41. Hydrogenases catalyze the reversible oxidation of molecular hydrogen (H2 ⇋ 2 H+ + 2e−), which takes place at their organometallic active sites. In [NiFe]-hydrogenases as well as the phylogenetically unrelated [FeFe]-hydrogenases, the active sites have iron atoms coordinated to CO and CN ligands, which are assumed to be essential for enzyme activity because they are always present42. These CO and CN ligands are also considered unusual43 since they are not found in other metalloenzymes. Figure 3 shows that the two iron cyanocarbonyl complexes found in LEW 85311 share structural similarities with the active sites of [NiFe]-hydrogenase from Desulfovibrio gigas43,44 and [FeFe]-hydrogenase from Clostridium pasteurianum45, specifically the FeII(CN)(CO) moieties with similar geometries33,46. It is conceivable that iron cyanocarbonyl complexes may have served as the building blocks of these bimetallic active sites through dimerization (for [FeFe]-hydrogenases) or association with a nickel-containing species (for [NiFe]-hydrogenases) in conjunction with the loss of CN ligands (possibly through photodissociation).

Fig. 3
figure 3

Meteoritic organometallic compounds compared to active sites. a [FeII(CN)5(CO)]3− (boxed) and active-site structure of [NiFe]-hydrogenase from Desulfovibrio gigas (1FRV). b [FeII(CN)4(CO)2]2− (boxed, cis form shown) and active-site structure of [FeFe]-hydrogenase from Clostridium pasteurianum (3C8Y). Regions that are shaded blue indicate structural similarity. A bridging ligand between metals is not shown for clarity

Full size image

A large number of model complexes mimicking the active site of [FeFe]-hydrogenases have been synthesized in the laboratory, although this number is much less for model complexes of [NiFe]-hydrogenase active sites due to the difficulty of synthesizing the heterobimetallic site47. Nevertheless, there have been reports of synthesizing model complexes of [NiFe]-hydrogenase active sites using Fe(CN)(CO)-containing precursors. For example, Perotto et al. demonstrated that stable heterobimetallic analogs of the active sites of [NiFe]-hydrogenases can be readily synthesized from reactions of fac-[Fe(CO)3(CN)2I] and Ni-centered compounds containing polydentate ligands that were stirred for 1–12 h in acetonitrile48. Biomimetic complexes that are catalytically active have been synthesized in the laboratory47, which suggests the possibility that primitive organometallic compounds could play a role in catalysis and hydrogen metabolism without a protein scaffold on early Earth.

Methods

Meteorite sample preparation

All glassware and ceramics were rinsed with ultrapure water, wrapped in aluminum foil, and heated to 500 °C overnight in order to remove organic compounds. All meteorites were pulverized to a fine powder using a porcelain mortar and pestle under a positive pressure ISO 5 High-Efficiency Particulate Air (HEPA) filtered laminar flow hood (Labconco).

We acid-digested and distilled the following amounts of each meteorite: 204.3 mg ALH 83100 (specific 272, parent 33), 203.6 mg ALH 84100 (specific 425, parent 15), 206.7 mg GRA 06100 (specific 49, parent 47), 202.8 mg LEW 85311 (specific 78, parent 3), 209.3 mg LEW 90500 (specific 79, parent 2), 201.1 mg LON 94102 (specific 29, parent 15), 204.4 mg Murchison (USNM 5451.1), and 207.7 mg RBT 04133 (specific 21, parent 0). Each powdered meteorite sample was placed into a MicroDIST sample tube of the distillation apparatus (Lachat Instruments) along with 5.8 mL ultrapure water. 750 μL of 9 M H2SO4 was added to the sample tube as a releasing agent for cyanide and 1.5 mL of 0.08 M NaOH was added to the top of the tube as a trapping solution for cyanide. The sample tube was separated from the NaOH trapping solution by a gas-permeable membrane within the tube. The distillation tube was sealed with a press (Lachat Instruments) and placed into a hot oil bath with a temperature range of 134–142 °C for 30 min. The final distillate volume in the top tube after distillation was ~2–3 mL.

Control samples were run in parallel with the meteorites, which included procedural (reagent) blanks, a serpentine blank (a hydrated magnesium silicate), and KCN standards.

Derivatization

Supplementary Figure 1 shows the derivatization scheme for cyanide using naphthalene-2,3-dicarboxaldehyde (NDA) and glycine. The meteorite and control sample distillates were processed using the following optimized conditions: derivatization for 15 min at room temperature using a mixture of 50 μL 0.1 M glycine, 50 μL 10 mM sodium borate buffer (pH 9.1), 50 μL meteorite extract or control sample, and 50 μL 1 mM NDA (modified from prior studies49,50). Glycine and sodium borate buffer solutions were dissolved in ultrapure water, KCN standards were dissolved in 0.08 M NaOH, and NDA was dissolved in methanol. Glycine, KCN, and NDA standard solutions were made fresh daily from the powder. We determined that the NDA-cyanide derivative showed <1% degradation after 2.25 h and <15% degradation after 17.5 h at room temperature. Due to the stability of the derivative, we derivatized and immediately analyzed one sample at a time to result in optimal response.

Total cyanide analysis

We used a Waters Acquity ultra performance liquid chromatograph coupled to a Waters Acquity fluorescence detector and Waters LCT Premier time-of-flight mass spectrometer for total cyanide analysis. Chromatographic separation was achieved using a prefilter and guard column followed by an Acquity BEH C18 2.1 × 50 mm column (1.7 μm particle size) and an Acquity BEH Phenyl 2.1 × 150 mm column (1.7 μm particle size) with a 25 μL injection volume. Samples were eluted using an isocratic flow of 35% 50 mM ammonium formate buffer (pH 8) with 8% methanol and 65% methanol at a flow rate of 150 μL min−1 and column temperature of 30 °C. The fluorescence excitation wavelength for the NDA-cyanide derivative was 252 nm and the emission was monitored at 483 nm. MS settings were as follows: positive ion mode, ESI capillary voltage +3.5 kV, cone voltage 30 V, desolvation temperature 350 °C, and source 120 °C. The N2 cone and desolvation gases were flowed at 50 L hr−1 and 650 L hr−1, respectively. Mass spectra were acquired in V-mode (mass resolution ~5000) with a range of m/z 50–500. Cyanide identification in the meteorite was made by comparison of the fluorescence and single ion mass peak (m/z 251.08 ± 0.03, which corresponds to the protonated molecular ion) retention times to standards. Cyanide concentration was calculated by comparing the fluorescence and mass peak areas of the sample to our standard concentration curve.

Measurement of cyanide in standard solutions

We performed a seven-point linearity study using non-distilled solutions of KCN with a concentration range of 0.01–2 μM (or 60 fmol–12 pmol on column). The curve was highly linear for fluorescence (R2 = 0.9997) and for single ion mass detection at the theoretical protonated mass of 251.08 ± 0.03 Da (R2 = 0.9989) (Supplementary Fig. 3). The limit of detection was 0.1 nmol CN·g−1 meteorite (60 fmol on column) with a signal-to-noise-ratio of 3:1. The 0.5 μM KCN standard was processed through the entire method from acid-digestion and distillation to derivatization and analyzed using LC-FD/ToF-MS in order to determine recovery. The average recovery for cyanide was 99% based on eight measurements with mass and fluorescence detection.

Inductively coupled plasma mass spectrometry

ICP-MS analysis was performed in medium resolution (m/Δm = 4000 at 1% peak height) using a Thermo Finnigan Element 2 ICP-MS in the Department of Geology Plasma Laboratory at the University of Maryland. All samples, blanks, and standards were aspirated via cyclonic nebulizer into the plasma source in 2% HNO3 using Ar and N2 carrier gases with a flow rate of 100 μL min−1. The source was flushed extensively with ultrapure water between samples. The isotopes measured were: 23Na, 25Mg, 27Al, 39K, 43Ca, 47Ti, 49Ti, 56Fe, 57Fe, 59Co, 60Ni, 62Ni, 63Cu, 65Cu, 66Zn, 67Zn, 68Zn, 107Ag, 109Ag, 111Cd, 112Cd, 114Cd, 115In, 194Pt, 195Pt, 196Pt, and 197Au. Multiple mass stations were monitored for elements most susceptible to isobaric interferences (even in medium resolution), and 115In spiked into each sample was used to track instrumental drift. Supplementary Table 1 shows the summary of elements measured in hot water and hot base extracts of LEW 85311.

High resolution mass spectrometry

Two samples of LEW 85311 (309.1 and 337.4 mg) were extracted under the following procedure: 500 μL of 0.08 M NaOH was added to each powdered LEW 85311 sample in 2.5-mL glass ampoules, flame-sealed, and put into an oven at 80 °C for ~20 h. A 0.08 M NaOH method blank was run in parallel with meteorite extracts.

Sample extracts were analyzed using a Thermo Scientific Accela High Speed LC coupled to a Thermo Scientific LTQ Orbitrap XL hybrid mass spectrometer. Separation was accomplished by injecting 20 μL sample solution onto a SIELC Primesep B2, 2.1 × 150 mm column (5 μm particle size). Mobile phase (A) was 100 mM ammonium acetate, pH = 4, mobile phase (B) was 100 mM ammonium acetate, pH = 9, mobile phase (C) was ultrapure water, and mobile phase (D) was acetonitrile. The elution method was as follows: initial mobile phase composition was set to 10% A, 70% C, and 20% D and held for 5 min, linear gradient to 50% B, 30% C, and 20% D over 15 min, mobile phase composition was returned to the initial conditions at 15.1 min, and the column was re-equilibrated for 15 min to finish the run. For optimal separation, the flow rate was set at 200 μL min−1.

The Thermo Scientific LTQ Orbitrap XL hybrid mass spectrometer was equipped with an electrospray ionization (ESI) source and operated in negative-ion mode. Parameters for ESI were as follows: the nitrogen gas for desolvation of the electrospray was set to 40 for the sheath gas, 5 for the auxiliary gas, and 1 for the sweep gas (all in arbitrary units). The ion transfer capillary voltage and temperature were −21 V and 275 °C, respectively. Full scan spectra in negative-ion mode were acquired over a mass range of m/z 120–320 and automated gain control (AGC) was set to 5 × 105 ions with a maximum injection time of 400 millisecond. The mass resolution was set to 30,000 (which translates to a mass resolution closer to 60,000 at m/z 200) to obtain good chromatographic peaks (~14 mass spectra were collected over a chromatographic peak width of 10 s). External calibration was performed using a mixture of sodium dodecyl sulfate, sodium taurocholate, and Ultramark 1621 in an acetonitrile-methanol-water solution containing 1% acetic acid.

Elemental compositions were calculated from the negative ion using C, H, O, N, Na, K, and Fe. Accurate mass measurements (usually <1 ppm mass error) enabled assignment of molecular formulas, which were consistent with predicted isotope patterns. While the orbitrap mass analyzer is known for its high mass accuracy and resolving power, its spectral accuracy51 (i.e., accuracy in measuring the abundances of isotopic peaks) has been known to underrepresent heavier isotopologues52, which may be the case for a couple of low intensity isotopic peaks belonging to iron cyanocarbonyl complexes. Nevertheless, accurate mass and distinct isotope patterns with good spectral accuracy enabled confident molecular assignments in our study.