Figure 1. The 0–40 min region of the LC-FD and ToF-MS chromatograms of the procedural blank and Almahata Sitta meteorite acid-hydrolyzed, hot-water extracts from the GSFC analyses. Similar LC-FD and ToF-MS chromatograms were obtained for the nonhydrolyzed extracts and are available upon request. The traces corresponding to the masses of the C2–C6 amino acids were plotted as the sum over a given mass range (peak width at half maximum of ∼200 ppm) for the Almahata Sitta meteorite sample only. Separation was achieved using a Waters BEH C18 column (2.1 × 50 mm, 1.7 μm bead) followed by a second Waters BEH phenyl column (2.1 × 150 mm, 1.7 μm bead). The conditions for separation of the OPA/NAC (15 min derivatization) amino acid derivatives at 30 °C were as follows: flow rate, 150 μL min−1; solvent A (50 mm ammonium formate, 8% methanol, pH 8.0); solvent B (methanol); gradient, time in minutes (%B): 0 (0), 35 (55), 45 (100). The peaks were identified by comparison of the retention time and exact molecular mass to those in the amino acid standard run on the same day. Peaks in the chromatograms that did not correspond to the same UV fluorescence and mass retention times of the standard amino acids tested were not identified. (See Table 1 for the identities of the numbered peaks.)
Download figure to PowerPoint
Figure 2. The 21–41 min region of the LC-ToF-MS single ion chromatograms (m/z = 379.13) in positive electrospray ionization mode from the GSFC analyses. OPA/NAC derivatization (15 min) of the five-carbon (C5) amino acids in the standard mix and the acid-hydrolyzed, hot-water extracts from the procedural blank and the Almahata Sitta meteorite. Similar LC-ToF-MS single ion chromatograms were obtained for the nonhydrolyzed extracts and are available upon request. To separate all possible C5α-, β-, γ-, and δ-amino alkanoic acid isomers the same Waters columns discussed in Fig. 1 caption with the following chromatographic conditions for the mobile phase at 30 °C were used: flow rate, 150 μL min−1; solvent A (50 mm ammonium formate, 8% methanol, pH 8.0); solvent B (methanol); gradient, time in minutes (%B): 0 (15), 25 (20), 25.06 (35), 44.5 (40), 45 (100). The peaks were identified by comparison of the retention time and exact molecular mass to those in the C5 amino acid standard run on the same day. Peak identifications are given in Table 1.
Download figure to PowerPoint
Figure 4. The relative molar abundances of the C5 amino acids in Almahata Sitta ureilite (data from Table 3) compared to several other carbonaceous chondrites as a function of (a) amine position (α-, β-, γ-, and δ-) and (b) valeric acid carbon chain structure (n-, sec-, iso-, and tert-) normalized to the total number of possible structures. The dashed line corresponds to the expected relative abundance if the amino acids were formed by a completely random synthetic process. The data for the CI, CM, and CR chondrites were taken from Glavin and Dworkin (2009). It is apparent from the data that there is structural similarity in C5 amino acid abundances within a carbonaceous chondrite group, and differences between groups. For the purpose of this manuscript, these plots illustrate that Almahata Sitta has a unique distribution of C5 amino acids based on both carbon chain structure and amine position compared to other CI, CM, and CR carbonaceous chondrites.
Download figure to PowerPoint
Amino Acid Analyses of the Almahata Sitta Meteorite
Typical HPLC-FD and LC-ToF-MS chromatograms of the hot-water extracts from the GSFC and SIO analyses of the Almahata meteorite and procedural blank are shown in Figs. 1–3. The peaks labeled with an “X” in the chromatograms are desalting artifacts and non-UV fluorescent mass peaks that did not interfere with the ability to identify the amino acids and amines in the extracts. The most abundant primary amines detected in the Almahata Sitta acid-hydrolyzed extracts were isopropylamine (149 ppb), ethylamine (104 ppb), glycine (69 ppb), and 4-amino-2-methylbutanoic acid (65 ppb). Trace to low levels (∼1–21 ppb) of aspartic and glutamic acids, alanine, β-alanine (BALA), α-, β-, and γ-amino-n-butyric acid (ABA), α-aminoisobutyric acid (α-AIB), ethanolamine, methylamine, and the five-carbon (C5) amino acids valine, norvaline, isovaline, 3-, 4-, and 5-aminopentanoic acid, 3-amino-2,2-dimethyl-propanoic acid, and 4-amino-3-methylbutanoic acid were also identified (Tables 2 and 3). As an interfering mass peak (m/z = 379.13) was detected in both the nonhydrolyzed and acid-hydrolyzed procedural reagent blank at the same retention time as l-valine (Fig. 2), accurate quantification of this amino acid was problematic and the values reported for l-valine in Table 3 should be considered to be upper limits.
Several unidentified peaks in the m/z = 393.15 and m/z = 407.16 mass traces corresponding to the C6 and C7 aliphatic amino acid isomers, respectively, were also detected in the Almahata Sitta extracts (Fig. 1). However, we were only able to confirm the identity of two of the C6 amino acid peaks: ε-amino-n-caproic acid (EACA) and the norleucine internal standard (IS) used to estimate amino acid desalting recoveries. Analyses of Antarctic meteorites have previously shown that EACA in bound form is derived from contamination from the Nylon-6 storage bags used in sample collection and curation (Glavin et al. 2006). However, Almahata Sitta was collected with clean aluminum foil (Jenniskens, personal communication), and as EACA is present entirely in free and not bound form in Almahata Sitta (Table 2), this amino acid cannot be derived from nylon. Additional analyses of amino acid standards will be required to identify the remaining C6 and C7 amino acid isomers in Almahata Sitta. No peaks in the chromatograms corresponding to the α-, β-, γ-, and δ-amino alkanoic acids containing more than seven carbons were detected in this meteorite. Many of the C6 and C7 acyclic primary α-amino alkanoic acids have previously been found in the Murchison meteorite (Cronin et al. 1981; Cronin and Pizzarello 1986), and primary aliphatic amino acids with as many as 10 carbons (C10) have been detected in Murchison meteorite extracts using our LC-FD/ToF-MS technique.
A similar distribution of free amino acids and amines was also detected in the nonhydrolyzed water extracts of Almahata Sitta (Fig. 3). Most of the amino acids detected by HPLC-FD from the SIO analyses are near background levels except for glycine, the most abundant amino acid detected both at SIO and GSFC. Although the HPLC-FD instrument used at SIO does not have the enhanced sensitivity and mass identification capability as the LC-ToF-MS needed to quantify accurately trace levels of amino acids, several volatile amines (eethanolamine, methylamine, ethylamine, and isopropylamine) present at concentrations ranging from ∼10 to 150 ppb were identified and quantified by HPLC-FD at SIO (Fig. 3; Table 2). Mass peaks corresponding to these four volatile amines were confirmed by LC-ToF-MS in the Almahata Sitta extracts analyzed at GSFC. However, these volatile amine peaks were too small to quantify and may have been partially lost during the extraction and processing procedure used at GSFC. Ammonia was also identified in both the Almahata Sitta and procedural reagent blank extracts from SIO (Fig. 3); however, the most likely source of the ammonia is incomplete evaporation of the NH4OH after desalting.
From the data in Tables 2 and 3, we calculate that the ratio of free amino acids to total (free + bound) amino acids identified in Almahata Sitta is 0.23 ± 0.06. This ratio is lower than the ratio of approximately 0.4 measured in the CM carbonaceous chondrites Murchison and Lewis Cliffs 90500 (Glavin et al. 2006). The parent body of Murchison and other CM carbonaceous chondrites experienced much lower temperatures compared to asteroid 2008 TC3 (Clayton and Mayeda 1984). Therefore, one possible explanation for the difference between the ratio of free to total amino acids is that free amino acids in the presence of iron or other metal cations are less stable to thermal decomposition and more readily oxidized at elevated temperatures than amino acids in bound form. In contrast to the CM chondrites, we observed a much higher abundance of amines in Almahata Sitta compared to amino acids with a total amine to total amino acid ratio of 1.1 ± 0.3 (Table 2) compared to a ratio of 0.1–0.5 for the Murchison meteorite (Botta and Bada 2002). This result may indicate a higher rate of thermal decomposition of amino acids to amines at elevated temperatures on the Almahata Sitta parent body. The total amino acid abundance measured in Almahata Sitta (∼275 ppb) is 15–900 times lower than reported in the CI chondrites Orgueil and Ivuna, CMs Murchison and Murray, and Antarctic CRs Elephant Moraine (EET) 92042 and Graves Nunataks 95229 (Ehrenfreund et al. 2001; Glavin et al. 2006; Martins et al. 2007). The low amino acid abundances and relatively high concentration of amines in Almahata Sitta compared to other carbonaceous chondrites is not surprising given that the Almahata Sitta meteorite was subjected to much higher temperatures (as inferred from its mineralogy) than these CI, CM, and CR chondrites, which experienced only the relatively low temperatures of 0–150 °C during aqueous alteration (Zolensky et al. 1993; Clayton and Mayeda 1999).
Unusual Amino Acids and D/L Enantiomeric Ratios
The nonprotein amino acids α-AIB and isovaline detected in Almahata Sitta are not common amino acids on Earth, and are thus characteristic of amino acids of apparent extraterrestrial origin. Biological α-AIB and isovaline have been detected in acid-hydrolyzed extracts of a variety of fungal peptides (Brückner et al. 2009). However, terrestrial contamination as the sole source of α-AIB and isovaline detected in Almahata Sitta is not consistent with the high relative abundance of free α-AIB and isovaline (∼50% in free form; Tables 2 and 3) measured in this meteorite as fungal derived α-AIB and isovaline would be entirely in the peptide bound form (Brückner et al. 2009), and not the free form as observed in Almahata Sitta. Moreover, the presence of several unusual five-carbon β-, γ-, and δ-amino acids that are not found in fungal peptides (Table 3), and racemic d/l ratios of the amino acids alanine, β-amino-n-butyric acid (β-ABA), 2-amino-2-methylbutanoic acid (isovaline), and 2-aminopentanoic acid (norvaline) in the Almahata Sitta meteorite (Table 4), provide additional evidence that these amino acids have an abiotic origin and are indigenous to the meteorite.
Large l-enantiomeric excesses (ee) of isovaline exceeding 15% have been measured in the CM chondrite Murchison (Cronin and Pizzarello 1997; Pizzarello et al. 2003; Glavin and Dworkin 2009) and similar l-isovaline ee were recently reported for the first time in the CI Orgueil (Glavin and Dworkin 2009). The large l-isovaline excesses observed in CI and CM chondrites is inconsistent with UV circularly polarized light as the primary mechanism, as UV CPL has only been shown to produce amino acid excesses of a few percent (Flores et al. 1977; Takano et al. 2007). As l-isovaline ee have only been observed in the more aqueous altered CM and CI chondrites and not in the most pristine, unaltered carbonaceous chondrites (such as the Antarctic CRs Queen Alexandra Range 99177 and EET 92042), amplification of a slight initial l-isovaline imbalance may have occurred during aqueous alteration on the meteorite parent bodies (Glavin and Dworkin 2009). Based on the total abundance data for isovaline in Table 3, we calculated an l-isovaline enantiomeric excess (lee = l% − d%) in Almahata Sitta of 3.7 ± 5.1%. A similar lee was obtained for the free isovaline (Table 3). The lee for isovaline in Almahata Sitta is indistinguishable from zero within measurement error and is consistent with the lack of any mineralogical evidence for aqueous activity on the Almahata Sitta meteorite parent body (Zolensky et al. 2009), but does not rule out UV CPL as a possible source of a small amount of l-isovaline asymmetry in Almahata Sitta that is within our analytical error of a few percent.
Only trace amounts (<1–10 ppb) of l-aspartic acid, l-serine, glycine, β-alanine, l-alanine, and l-valine in the procedural and serpentine blank could be detected by standard RP-HPLC-FD and LC-ToF-MS (Figs. 2 and 3), which indicates that minimal amino acid contamination of the samples occurred during the processing procedure. Nevertheless, given the extremely low abundances of amino acids in Almahata Sitta and low enantiomeric ratios of the protein amino acids aspartic and glutamic acids (d/l ∼ 0.5–0.6) in the hydrolyzed extracts (Table 4), we cannot rule out the possibility that some terrestrial protein amino acid contamination of Almahata Sitta occurred after its fall. A similarly low d/l glutamic acid ratio of approximately 0.6 previously measured in the Antarctic CR chondrite EET 92042 was attributed to terrestrial amino acid contamination from exposure to the Antarctic ice or during sample curation (Martins et al. 2007). Although the d- and l-enantiomers of valine were separated using the LC conditions employed, there was a relatively large unidentified coeluting mass peak in the m/z = 379.13 mass trace at the same retention time as l-valine in the procedural reagent blank (Fig. 2), which may have led to artificially high l-valine abundances, and hence lower d/l valine ratios (∼0.2–0.4) in the extracts (Table 4). Biologically derived l-amino acid contamination of Almahata Sitta, possibly from exposure to soil at the landing site during the approximately 2 month residence time in the Nubian Desert prior to collection, may have lowered the initial d/l ratios of these protein amino acids to the d/l ratios presently observed. Amino acid analyses of the Martian meteorites Allan Hills 84001 and Nakhla have shown that terrestrial amino acid contaminants can be rapidly absorbed by meteorites from the landing site environment (Bada et al. 1998; Glavin et al. 1999). Future analysis of amino acids in soil collected from the meteorite fall site as well as compound-specific stable isotopic analyses (CSIA) would help constrain the origin of these protein amino acids. Isotopic measurements of the amino acids and amines in Almahata Sitta were not possible given their very low abundances and the limited mass of sample available for this study. Based upon a detection limit of 1 nmol for carbon CSIA using our gas chromatography-isotope ratio mass spectrometry instrument at GSFC (Elsila et al. 2009), we would require a minimum of several grams of meteorite for carbon isotope measurements of the most abundant amino acid glycine.
The Origin and Survival of Amino Acids on Asteroid 2008 TC3
The detection of even part-per-billion levels of amino acids and amines in Almahata Sitta is surprising given that mineralogical evidence of the meteorite points to fractional melting and shock heating temperatures of approximately 1100–1300 °C on asteroid 2008 TC3 (Herrin et al. 2009). Amino acids will rapidly decompose when heated to temperatures above 500–600 °C, even for very short durations (Rodante 1992). For example, heating experiments of the Murchison meteorite showed that only 3% of the amino acids inside the meteorite survived heating to a temperature of 550 °C for approximately 30 s, and all of the amino acids were destroyed after the meteorite fragments were melted using a CO2 laser at a temperature of approximately 1200 °C for 10 s (Glavin and Bada 2001). Based on these experimental results, it seems very unlikely that any indigenous amino acids would be present in the Almahata Sitta meteorite if these organic compounds formed (or were incorporated into the asteroid from carbonaceous precursor materials) prior to or during the shock heating and partial melting experienced by asteroid 2008 TC3.
One possibility is that the amino acids and amines detected in Almahata Sitta were synthesized directly from their chemical precursors (e.g., HCN, H2, CO2, NH3, and CH4) after asteroid 2008 TC3 cooled to lower temperatures. The formation of amino acids has been experimentally observed by Fischer-Tropsch type catalytic reactions of CO, H2, and NH3 in the gas phase at 200–500 °C in the presence of nickel-iron (Hayatsu et al. 1971; Yoshino et al. 1971). Aliphatic amines have also been produced in similar fast Fourier transformation experiments (Kölbel and Trapper 1966). Therefore, it is possible that the amino acids and amines detected in the Almahata Sitta meteorite formed by similar catalytic reactions at elevated temperatures on 2008 TC3. In contrast, there is chemical evidence that the complex distribution of α-amino acids observed in Murchison and other CM chondrites formed from the reaction of aldehydes, ketones, NH3, and HCN by Strecker-cyanohydrin synthesis on its asteroidal parent body during a low temperature aqueous alteration phase (Peltzer and Bada 1978; Ehrenfreund et al. 2001; Lerner and Cooper 2005). Alternative sources for the amino acids in carbonaceous chondrites (e.g., from interstellar ices) that do not require aqueous activity have also been proposed (Bernstein et al. 2002). The high relative abundance of the five-carbon α-amino acids over β-, γ-, and δ-amino acids in the CM and CR chondrites studied provides additional evidence that Strecker synthesis was active on the parent bodies of these meteorites (Fig. 4).
In contrast to the CM and CR chondrites, the Almahata Sitta amino acid distribution is unique and is dominated by the five-carbon γ-amino acids (Fig. 4), which cannot be produced by the Strecker route. In addition, the formation of amino acids by the Strecker mechanism requires aqueous activity on the meteorite parent body, and there is no mineralogical evidence for aqueous alteration in the Almahata Sitta meteorite (Zolensky et al. 2009). Several α-amino acids and one β-amino acid (β-alanine) have been synthesized by Fischer-Tropsch; however, no γ-amino acids were reported (Yoshino et al. 1971). Additional studies of the Fischer-Tropsch based synthesis of amino acids are clearly needed using modern analytical methods to ascertain the distribution of amino acids produced by this process. It has been suggested that some meteoritic γ-amino acids could be produced from the hydrolysis of lactams that have been found in Murchison (Cooper and Cronin 1995; Pizzarello et al. 2006). It is possible that the γ-amino acids detected in Almahata Sitta could have formed from lactams by a similar process; however, lactams contain secondary (and not primary) amines and are therefore not detectable using our OPA/NAC derivatization and LC-FD/ToF-MS analytical method.
It has been suggested that polymict ureilites such as Almahata Sitta represent regolith that subsequently formed on the surface of a daughter asteroid from the remnants of the proto-ureilite asteroid and other carbon-rich asteroid impactors that reaccreted in a rubble pile (Downes et al. 2008). This could have introduced meteorites of unrelated origin, which carried amino acids. Subsequent collisions could have spread this material throughout the asteroid. Another possible explanation for the presence of amino acids and amines in Almahata Sitta is that the carbonaceous chondrite precursor material of asteroid 2008 TC3 contained a very high initial abundance of amino acids or their precursors prior to impact of the parent asteroid and the subsequent heating of this material during reaccretion did not reach high enough temperatures to destroy completely all the amino acids and amines. If this formation model for asteroid 2008 TC3 is correct, it seems unlikely that any of the carbonaceous precursor material found in Almahata Sitta was originally CV chondrite-like given that the CV chondrites Allende and Mokoia have been found to be essentially devoid of amino acids with total abundances of <1 nmol g−1 (Cronin and Moore 1971, 1976). Based on amino acid evidence alone, it is possible that the carbonaceous precursor material for asteroid 2008 TC3 originally contained a higher abundance of amino acids, similar to the range of abundances reported for CI, CM, and CR carbonaceous chondrites. However, future analyses of CVs and other ureilites using modern high sensitivity analytical techniques may reveal that these meteorites contain similar levels of indigenous amino acids to Almahata Sitta. Although we are currently unable to rule out either one of these possible explanations for the presence of amino acids and amines in Almahata Sitta, formation of these compounds after asteroid 2008 TC3 cooled to lower temperatures is the more likely explanation given the high fractional melting and shock heating temperatures on 2008 TC3.
Given the higher abundance of amines compared to amino acids in Almahata Sitta (Table 2), it is reasonable to conclude that the amines could have been produced by thermal decomposition of amino acids on the parent body at elevated temperatures (Simmonds et al. 1972). We did not observe any evidence for thermal decomposition of amino acids into amines in a pure amino acid standard carried through the identical extraction procedure as the meteorite samples, therefore the amines observed in Almahata Sitta were not produced by our extraction method. As several simple amines have been detected in soil samples collected from the Atacama Desert in Chile (Skelley et al. 2007), we cannot rule out the Nubian Desert soil in Sudan as a possible terrestrial source of these amines in Almahata Sitta. However, if the amines in Almahata Sitta are extraterrestrial in origin and thermal decomposition by α-decarboxylation of the amino acids glycine, alanine, and α-AIB was the primary source of methylamine, ethylamine, and isopropylamine, respectively, then the relative abundances of these amines should be consistent with the decarboxylation rates measured for the corresponding amino acids.
From the total abundance data reported for the hydrolyzed Almahata Sitta extracts in Table 2, we calculate the following molar ratios from the molecular weights: methylamine/glycine = 0.44, ethylamine/alanine = 9.4, and isopropylamine/α-AIB = 37. Assuming that these three amines are produced entirely by α-decarboxylation of the corresponding amino acids, we would conclude from the molar ratios that α-amino acid stability follows the order of glycine > alanine > α-AIB. However, direct measurement of the decarboxylation rates of these α-amino acids in aqueous solution at 310–330 °C and 275 bar follow the exact opposite order of stability of α-AIB > alanine > glycine (Li and Brill 2003). Although the amine to amino acid ratios observed in Almahata Sitta do not follow the expected trend for thermal decarboxylation, it should be emphasized that the amino acid decarboxylation rates reported by Li and Brill (2003) were measured in solution and the relative rates could be very different in the solid state in the absence of any aqueous activity on asteroid 2008 TC3. In addition, we cannot rule out the possibility that the amine to amino acid molar ratio trend observed in Almahata Sitta is related to evaporative loss of amines during the extraction procedure and the relative volatility of these compounds. In any case, if the amines detected in Almahata Sitta are indigenous to the meteorite and derived directly from interstellar precursors or amino acid decarboxylation as has been previously suggested for the Murchison meteorite (Pizzarello et al. 1994), these amines could not have been exposed to the high temperatures associated with thermal alteration of the parent body.
In contrast to Almahata Sitta, a much more complex distribution and higher abundances of C1–C5 aliphatic amines have been identified in Murchison (Jungclaus et al. 1976; Pizzarello et al. 1994), and 16 of the 20 amines identified could be produced from the corresponding amino acids present in Murchison by α-decarboxylation (Simmonds et al. 1972). Based on published Murchison data from Pizzarello and coworkers (Cronin and Pizzarello 1983; Pizzarello et al. 1994), we calculate the following amine to amino acid molar ratios in Murchison: methylamine/glycine = 0.72, ethylamine/alanine = 0.40, and isopropylamine/α-AIB = 0.07. These ratios are consistent with α-decarboxylation as the primary source of amines based on the amino acid decarboxylation rates discussed above. In addition, nitrogen isotopic measurements of amino acids and volatile bases (including ammonia and amines) in Murchison show that these compounds are enriched in 15N and fall in a narrow range (δ15N = + 94 ± 8‰), consistent with their formation from similar precursors (Pizzarello et al. 1994). Future nitrogen isotopic measurements of both amines and amino acids in Almahata Sitta will be necessary to help constrain their origins.