The effects of parent body processes on amino acids in carbonaceous chondrites

UPLC-FD Measurements and Total Amino Acid Abundances

Typical UPLC-FD chromatograms of the 6 M HCl-hydrolyzed, hot-water extracts from the CM1 chondrites SCO 06043 and MET 01070, the CR1 chondrite GRO 95577, the serpentine blank, and an amino acid standard mixture are shown in Fig. 1. Similar UPLC-FD chromatograms were obtained for the other meteorite samples. The amino acid peak numbers correspond to the designations given in Table 1. Peaks in the meteorite extracts were only identified as amino acids if the retention time of the UV fluorescence peak coincided with the corresponding mass peak of the OPA/NAC derivatized amino acid standard (see 1, 2). Several peaks labeled with an “X” in the fluorescence chromatograms for the meteorite samples and the serpentine blank are unidentified primary amine compounds from either the chemical workup or analytical artifacts that did not interfere with amino acid identification and quantification (Fig. 1). All of the identified amino acids and corresponding abundances are given in Tables 2 and 3. The absolute abundances reported were corrected for the desalting and derivatization recoveries using the d,l-norleucine internal standard; the recoveries ranged from 60–80% for the meteorite samples studied. Many other amino acid peaks were detected in the meteorite extracts (e.g., see Murchison in Fig. 2), but were not identified due to a lack of amino acid standards. All of the Antarctic meteorite samples revealed high intensity peaks (see 1, 2, peak no. 33) corresponding to the C₆ amino acid ε-amino-n-caproic acid (EACA). This finding is consistent with a previous analysis of Antarctic CM2 chondrites and the most likely source of EACA is terrestrial contamination from the Nylon-6 storage bags used during sample collection and curation of Antarctic meteorites (Glavin et al. 2006). The abundance of EACA in the Antarctic meteorites ranged from approximately 400 to 51,000 ppb (Table 2), suggesting that exposure of these meteorites to Nylon-6 contamination is variable. It should be emphasized that the EACA contamination in these meteorites did not interfere with the ability to detect trace levels of other amino acids present in the samples. Since EACA is a potential terrestrial contaminant and its abundance is highly variable in meteorites, we excluded this amino acid from the total amino acid abundances and the ratio of free to total amino acids reported in Table 2.

Much lower abundances ranging from 4 to 295 ppb of the amino acids glycine, α-alanine, BALA, α-, β-, and γ-amino-n-butyric acid (ABA), and α-aminoisobutyric acid (α-AIB) were found in the aqueously altered chondrites MET 01070, SCO 06043, and GRO 95577 (Table 2). The abundances of α-AIB in these type 1 chondrites (30 to 175 ppb) are 100 to 1,000 times lower than the α-AIB abundances found in the CM2 Murchison and the CR2 EET 92042 (Table 2). A total amino acid abundance of approximately 16,000 ppb was measured in LON 94102, which is in the range of amino acid abundances previously measured in the CM2 chondrites Murchison and LEW 90500 (Table 2). Previous bulk oxygen isotope data (Clayton and Mayeda 1999), as well as hydrogen, carbon, and nitrogen isotopic analysis of the IOM of LON 94102 suggest that this meteorite is among the least altered of the CM chondrites (Alexander et al. 2010).

From Table 2, the low total amino acid abundances measured in MET 01070 (approximately 710 ppb) and GRO 95577 (approximately 1,600 ppb) are consistent with previous amino acid analyses of these type 1 chondrites (Botta et al. 2007; Martins et al. 2007). Although aspartic acid, glutamic acid, and serine have all been identified previously in these meteorites, we were unable to confirm their presence above serpentine blank levels (Fig. 1, Table 2). These amino acids are difficult to identify by mass due to the low abundances and poor ionization of these compounds during the UPLC-ToF-MS analysis. The CM1 chondrite SCO 06043 has a very similar amino acid composition and total abundance (660 ppb) compared to MET 01070 (Table 2). The total amino acid abundance of the CI1 Orgueil (6,700 ppb) is higher than the CM1 and CR1 meteorites studied (Table 2). These Orgueil results are consistent with amino acid abundances measured in a previous HPLC-FD analysis of a different fragment of this meteorite (Ehrenfreund et al. 2001). Overall, the total amino acid abundances found in these aqueously altered type 1 chondrites are lower than those found in the less altered CM and CR type 2 and 3 chondrites which ranged from approximately 9,400 ppb for the CM2 LEW 90500 up to 320,000 ppb for the CR2 chondrite EET 92042 (Table 2). High amino acid concentrations ranging from 180,000 to 249,000 ppb have previously been reported for the CR2 meteorites EET 92042 and GRA 95229 (Martins et al. 2007; Pizzarello et al. 2008), but these values did not contain many of the C₅ amino acids that we included in our total amino acid abundance calculations reported in Tables 2 and 3. We found that the primitive CR3 chondrite QUE 99177 had a similar distribution of amino acids compared to EET 92042, but the total amino acid abundance of QUE 99177 (81,000 ppb) was lower than in EET 92042 (Table 2).

UPLC-FD/ToF-MS Measurements of C₂ to C₁₀ Amino Acids

We investigated the masses corresponding to the OPA/NAC derivatives of the acyclic C₂ to C₁₀ amino alkanoic acids in the sample extracts (see Fig. 2). Only trace levels (<10 ppb) of glycine (peak 9) and l-alanine (peak 12) in the serpentine procedural blank were measured by UPLC-FD/ToF-MS (Fig. 2), which indicates that minimal amino acid contamination of the meteorite samples occurred during the processing procedure. A few small unidentified nonfluorescent peaks were detected in both the serpentine blank and meteorite sample extracts and are marked with an “X” (Fig. 2). Based on the UPLC-FD/ToF-MS data, the Murchison meteorite contains a very large diversity of amino acids with evidence for C₂ up to C₁₀ acyclic amino alkanoic acid isomers detected above background (Fig. 2). The masses plotted in Fig. 3 represent a 0.02 Da window (the peak width at half maximum) centered around the corresponding theoretical masses of the mono-protonated positive ions of C₂ to C₁₀ OPA/NAC amino acid derivatives, and not the masses of the individual unlabeled amino acids. Based on the total ToF-MS areas in Fig. 3 (data not shown), the C₁₀ amino acids in Murchison are a factor of 10–20 times less abundant than the area of each of the C₂ through C₆ amino acids.

Amino acids with carbon numbers exceeding C₁₀ may also be present in Murchison, but these were not observed with confidence above the detection limit of the UPLC-ToF-MS instrument under the conditions employed. This may either be a result of meteorite chemistry or biases in the extraction and workup for the more hydrophilic amino acids. Diamino acids have also been identified in Murchison meteorite extracts using GC-MS (Meierhenrich et al. 2004). OPA derivatives of diamino acids have been separated using HPLC (Satyanarayana et al. 2001); however, our OPA/NAC derivatization and UPLC-FD/ToF-MS analytical technique was not optimized for the identification of these compounds. Nevertheless, these results are consistent with previous analyses of the Murchison meteorite that have identified over 80 amino acids, including both cyclic and acyclic primary α-amino alkanoic and alkanedioic acids with carbon numbers ranging from C₂ up to C₈ (Cronin et al. 1981; Cronin and Pizzarello 1983, 1986; Cronin and Chang 1993). To our knowledge, this study is the first report of amino acids in the Murchison meteorite that exceed C₈. Based on our UPLC-FD/ToF-MS data and the numerous possible structural isomers and enantiomers for amino acids with carbon numbers up to C₁₀, the Murchison meteorite likely contains hundreds of individual amino acids. Unfortunately, due to their low abundance, incomplete separation, and a lack of standards, the identity of many specific compounds cannot be confirmed. Future optimization of our LC separation conditions and synthesis of amino acid standards that are not commercially available will be required to determine the identities of the detected C₆ to C₁₀ amino acids in Murchison and other carbonaceous meteorites.

Several unidentified peaks above background also appeared in the m/z 393.15 mass window of the Murchison extract, corresponding to the C₆ amino acid isomers (Fig. 2); however, we only confirmed the identity of two of these peaks: (EACA, peak 33) and the norleucine internal standard (I.S.) used to estimate amino acid desalting recoveries. For EET 92042, the intensity of the EACA peak was so high that the I.S. peak is present, but not easily visible at the scale shown (Fig. 2). A few of the unidentified C₆ amino acid peaks were also present in the type 1 chondrites Orgueil and SCO 06043; however, the total number of C₆ isomers is clearly less in these carbonaceous chondrites compared to Murchison (Fig. 2). In contrast to Murchison, the CR chondrites GRO 95577 and EET 92042 did not contain any evidence for amino acids with more than six carbons above serpentine blank levels (Fig. 2). This finding is particularly surprising for EET 92042, which contains the highest reported amino acid abundance of any meteorite to date (Martins et al. 2007). The Orgueil and SCO 06043 meteorite extracts contain some evidence for the presence of C₇ and C₈ amino acids above serpentine background levels; however, the intensity of these amino acid peaks are reduced compared to Murchison (Fig. 2). These data are consistent with earlier studies showing reduced amino acid abundances in more aqueously altered type 1 carbonaceous chondrites compared to type 2 chondrites (Botta et al. 2007; Martins et al. 2007).

The UPLC-FD/ToF-MS instrument was specifically optimized for separation of the C₅ acyclic amino alkanoic acids and the retention times for all possible C₅ amino acid isomers and enantiomers were identified based on the analysis of standards (Fig. 3). The UPLC-FD/ToF-MS chromatograms centered at m/z 379.1328 ± 0.02 (the peak width at half maximum) correspond to the C₅ amino acids in the standard, the serpentine blank, and the MET 01070, SCO 06043, and GRO 95577 meteorite extracts (Fig. 3). Chromatograms showing the C₅ amino acid distributions in Orgueil, Murchison, LON 94102, EET 92042, and QUE 99177 are not shown here, but have been published previously (Glavin and Dworkin 2009). Although complete separation of all 23 possible C₅ amino acid isomers and enantiomers could not be achieved under the chromatographic conditions employed, all of the C₅ amino acids were accounted for and we observed no interference or coelution of the d- and l-enantiomers of isovaline, valine, norvaline, and 3-aminopentanoic acid (3-apa) with other C₅ amino acids. As with a number of rare C₅ amino acids, the d- and l-enantiomers of 3-apa were clearly separated (Fig. 3), however they could not be identified due to the lack of optically pure 3-apa standards.

A comparison of the individual and total abundances of the C₅ amino acids in the meteorites is shown in Table 3. The only C₅ amino acid that could not be quantified in any of the meteorite extracts due to coelution was 3-amino-3-methylbutanoic acid (3-a-3-mba); therefore, only upper limits are given for 3-a-3-mba. The total abundances of all other C₅ amino acids were determined by comparing the peak areas with standards (Table 3). The CM1 chondrites MET 01070 and SCO 06043 and the CR1 chondrite GRO 95577 had low total C₅ amino acid abundances ranging from 220 to 500 ppb, much lower than found in the CM2 and CR2/3 chondrites (Table 3). The most likely explanation for the low abundances in type 1 chondrites is decomposition of amino acids during aqueous alteration on the meteorite parent body. This is not surprising, since the decomposition of aliphatic moieties in the free and macromolecular organic matter by low temperature chemical oxidation during aqueous alteration on the parent body could result in the low amino acid levels observed in the type 1 chondrites (Sephton et al. 2004; Cody and Alexander 2005). Thermal decomposition by α-decarboxylation of α-amino acids to CO₂ and the corresponding primary amine during aqueous alteration on the parent body is another possibility (Simmonds et al. 1972); however, as noted previously, many of the volatile amines produced from α-decarboxylation (including the butylamines from the C₅ amino acids) were not detected in our UPLC-FD/ToF-MS analyses of the meteorite extracts and may have been lost during chemical workup.

Presence of Free versus Bound Amino Acids

It has been shown previously that acid hydrolysis of the water extracts from carbonaceous meteorites will readily release bound amino acids. These bound amino acids have their amino group chemically protected and are not subject to OPA/NAC derivatization or exist as a precursor to the amino acid and would have a different mass and retention time compared to the OPA/NAC labeled amino acid. The increase in amino acid concentration in the hydrolyzed extracts observed in many carbonaceous chondrites is believed to be a result of derivatives and/or acid-labile precursors that are converted to amino acids after acid hydrolysis. For example, in the Murchison meteorite most of the acid-labile precursors are low molecular weight derivatives including mono- and dicarboxylic acid amides, hydroxy acid amides, lactams, caboxylactams, N-acetylamino acids, and substituted hydantoins (Cronin 1976a,b; Cooper and Cronin 1995).

Analysis of the nonhydrolyzed water extracts of the CM1s SCO 06043 and MET 01070, and the CR1 GRO 95577 showed a similar distribution of free amino acids compared to the total (free + bound) amino acids in the acid-hydrolyzed extracts. From the UPLC-FD/ToF-MS data (not shown), we calculated the total abundance of free amino acids (not including EACA) in the nonhydrolyzed extracts (Table 2). The results are similar to previous analyses of the CM2 chondrites Murchison, LEW 90500, and ALH 83100 (Glavin et al. 2006) and the CM1 chondrites MET 01070 and ALH 88045 (Botta et al. 2007) which reported that free amino acids on average represent approximately 40 to 50% of the total acid-hydrolyzable amino acid abundance. In contrast, Pizzarello and coworkers found that the abundance of amino acids in the relatively unaltered Antarctic CR2 chondrite GRA 95229 hot water extract did not increase substantially after acid hydrolysis, with free amino acids representing over 70% of the total amino acid abundance on average (Pizzarello et al. 2008). We also measured a high relative abundance of free amino acids (free/total = 64%) in another Antarctic CR2 chondrite EET 92042 (Table 2). It is possible that bound amino acids are more resistant to oxidation during aqueous alteration than free amino acids, resulting in a lower free/total amino acid ratio in more aqueously altered chondrites. However, if this were true we would expect a higher ratio of free to total amino acids in the less altered CR3 chondrite QUE 99177 compared to the CR2 EET 92042, and this was not observed (Table 2). Amino acid analyses of the nonhydrolyzed water extracts from additional carbonaceous chondrites that experienced minimal aqueous alteration will be required to further investigate this hypothesis.

d/l Enantiomeric Ratios

Chirality is another important measurement tool to help discriminate between biotic and abiotic origins of amino acids in meteorites, even if present at low abundances. On Earth, biology almost exclusively uses l-amino acids, whereas abiotic syntheses result in equal mixtures of l- and d-amino acids. Furthermore, protein amino acids will readily racemize (approach d/l = 1) on geologically short time scales (Bada 1972). Therefore, if we assume that the protein amino acids found in carbonaceous chondrites were racemic (d/l approximately 1) prior to their fall to Earth, then their current d/l ratios can be used to determine the relative degree of terrestrial l-amino acid contamination they have experienced on Earth. The enantiomeric ratios (d/l) for aspartic and glutamic acids, serine, alanine, β-ABA, valine, norvaline, and isovaline are reported in Table 4 when abundances permitted quantitation.

The enantiomeric ratios measured for the protein amino acids (italicized in Table 4) in the CM chondrites Murchison (USNM 6650), LEW 90500, and LON 94102, and the CR chondrites EET 92042 and QUE 99177 were mostly found to be racemic (d/l approximately 0.9 to 1.0) within analytical errors (±0.05 to 0.3, Table 4), indicating that these amino acids are primarily abiotic in origin and not the result of terrestrial contamination. However, a lower d/l ratio for glutamic acid of 0.57 measured in Orgueil (Table 4) may indicate that some terrestrial l-glutamic acid is present in this meteorite. The relatively low d/l ratios of aspartic and glutamic acids in Murchison (USNM 5453) of 0.67 and 0.61, indicate that this fragment has a higher degree of terrestrial l-protein amino acid contamination than the other Murchison sample used in this study (USNM 6650) which has much higher d/l ratios for aspartic and glutamic acids (d/l = 0.91 and 0.96, respectively) (Table 4). Despite these significant differences in d/l aspartic and glutamic acid ratios in the two samples of Murchison, the nonprotein amino acids β-ABA, norvaline, and isovaline in these samples had identical d/l ratios within errors (Table 4), suggesting that terrestrial contamination is an unlikely source for these nonprotein amino acids. Although aspartic and glutamic acids were not detected in MET 01070, SCO 06043, and GRO 95577, the d/l ratio for alanine measured in these meteorites was racemic (d/l approximately 1) within analytical uncertainties (Table 4), indicating both an extraterrestrial origin for alanine and low levels of terrestrial l-amino acid contamination in these meteorites.

Although the d- and l-enantiomers of valine were separated using the LC conditions employed, there was a relatively large unidentified nonfluorescent coeluting isobaric contaminant at approximately the same retention time as l-valine in some of the analyses including the serpentine blank (Fig. 3), which may have led to erroneously high l-valine abundances, and hence lower d/l valine ratios (approximately 0.2 to 0.5) for Orgueil, Murchison (USNM 5453), SCO 06043, and GRO 95577 (Table 4). Therefore, lower limits were placed on these ratios. The nonprotein amino acids β-ABA and norvaline detected in these meteorites were racemic (d/l approximately 1) within analytical error, indicating an abiotic origin for these amino acids (Table 4). One notable exception was the enantiomeric ratio of the nonprotein amino acid isovaline, which ranged from 0.61 to 0.80 in Orgueil, SCO 06043, GRO 95577, and Murchison (Table 4). Terrestrial isovaline contamination of these meteorites after falling to Earth is highly unlikely since no isovaline was detected in any of the relevant collection site environments or any of the laboratory procedural blanks. Although some fungal peptide contamination of both d- and l-isovaline cannot be completely ruled out, it is important to note that isovaline is most commonly found in the d-configuration in fungal peptides (Keller et al. 1990; Brückner et al. 2009), which would increase (and not decrease) the d/l isovaline ratio. In addition, many common fungal peptides also contain l-alanine (d/l approximately 0.06) at similar abundances to isovaline (Brückner et al. 2009); therefore, if any significant fungal peptide contamination of the meteorites had occurred, the d/l alanine ratios in these meteorites should have been affected as well. However, the d/l alanine ratios in all of the meteorites analyzed were racemic within analytical errors showing no evidence for significant terrestrial l-alanine contamination (Table 4). The detection of nonracemic d/l isovaline ratios corresponding to l-isovaline excesses in Murchison and other carbonaceous chondrites have been reported previously (Cronin and Pizzarello 1997; Pizzarello and Cronin 2000; Pizzarello et al. 2003, 2008; Glavin and Dworkin 2009). Possible mechanisms for the origin of the l-isovaline asymmetry observed in some carbonaceous chondrites are discussed below.

Relative Abundance of β-alanine as an Indicator for Aqueous Alteration

A comparison of the relative molar abundances (glycine = 1.0) of α-alanine (d+l enantiomers), β-alanine, α-AIB, and isovaline (d+l enantiomers) measured in nine different carbonaceous chondrites is shown in Fig. 4. The order of carbonaceous chondrites in Fig. 4 are based on the approximate degree of aqueous alteration inferred from mineralogical and isotopic evidence with the most altered CI1 Orgueil and other type 1 chondrites on the left and the most primitive, least altered CR chondrites on the right (Zolensky and McSween 1988; Kallemeyn et al. 1994; Zolensky and Browning 1994). The relative degree of aqueous alteration among carbonaceous chondrites within the same group and of the same petrologic type (e.g., the difference in degree of alteration between the three CM2 chondrites shown in Fig. 4) is less certain, although recent isotopic analyses of the IOM in LON 94102 suggests that this meteorite is less altered than Murchison (Alexander et al. 2010). Based on the data in Fig. 4, there appears to be a correlation between the relative abundance of β-alanine and degree of aqueous alteration in the meteorites studied. The relative abundance of β-alanine increases from a β-alanine/glycine ratio of approximately 0.1 to 0.2 for the least altered CR chondrites EET 92042 and QUE 99177 up to approximately 2.7 for the most aqueously altered CI1 Orgueil. We also find that within the corresponding CM and CR chondrite groups, the type 1 chondrites have a much higher relative abundance of β-alanine compared to the less altered type 2 chondrites (Fig. 4). Thus, in addition to mineralogy and isotopic measurements, the relative abundance of β-alanine should also be considered when comparing the degree of aqueous alteration between carbonaceous chondrites.

In contrast to β-alanine, there do not appear to be any strong trends in relative α-alanine, α-AIB, and isovaline abundances with aqueous alteration (Fig. 4). However, it is worth noting that the relative abundances of α-AIB and isovaline tend to be higher in carbonaceous chondrites that experienced less aqueous alteration (Fig. 4). The relative abundance of α-alanine ranged from 0.2 for Orgueil up to 1.3 for EET 92042. The high α-alanine abundance in EET 92042 was not observed in QUE 99177 and cannot be explained by a large contribution of terrestrial l-alanine contamination since the d/l alanine ratio in EET 92042 was 0.97 (Table 4). Assuming that Strecker synthesis was the primary mechanism for the formation of α-alanine in EET 92042 and QUE 99177, it is possible that the parent body of EET 92042 had a higher ratio of acetaldehyde to formaldehyde (the Strecker precursors for formation of α-alanine and glycine, respectively) compared to QUE 99177.

The weakly altered CM2 chondrite LON 94102 is clearly different compared to the other more altered CM2 chondrites (or any other meteorite analyzed in this study) and appears to follow the opposite trend in relative α-AIB and isovaline abundances, with lower α-AIB/glycine and isovaline/glycine ratios compared to Murchison and LEW 90500 (Fig. 4). One possibility for the low α-AIB and isovaline abundances in LON 94102 is that this meteorite originated from a chemically distinct parent body from other CM chondrites that was depleted in acetone and 2-butanone, the ketone precursors required for the formation of α-AIB and isovaline, respectively, by Strecker-cyanohydrin synthesis. However, based on a variety of petrologic properties of eleven different CM chondrites that show similar bulk compositions, it has been suggested that textural differences seen among individual CM chondrites may reflect progressive alteration of similar hypothetical CM3 starting materials in different regions of the same parent body (Rubin et al. 2007). Therefore, we cannot rule out the possibility that the relative abundances of amino acids in LON 94102 are typical for weakly altered CM chondrites. Comparing the relative distribution of amino acids in LON 94102 to other weakly altered CM2 chondrites, such as QUE 97990 and Yamato (Y)-791198 (Metzler et al. 1992; Rubin et al. 2007; Chizmadia and Brearley 2008; Maeda et al. 2009), may help confirm this hypothesis. Previous amino acid analyses found more than ten times the relative molar abundance of α-AIB (α-AIB/glycine = 2.7 to 3.5) in Y-791198 (Shimoyama et al. 1985; Shimoyama and Ogasawara 2002) compared to what we observed in LON 94102 in this study (Fig. 4). However, it should be pointed out that we have not yet analyzed Y-791198 using the same UPLC-FD/ToF-MS technique used for LON 94102. To our knowledge, no amino acid measurements have been reported yet for QUE 97990.

The similarity in the distributions of α-alanine, β-alanine, α-AIB, and isovaline among type 1 chondrites across three separate carbonaceous chondrite groups (CI, CM, and CR) that are distinct from the less altered type 2 and type 3 chondrites (Fig. 4), may point toward degree of aqueous alteration, rather than unique parent body chemical compositions, as an explanation for the amino acid variations observed in these chondrites. Although the high relative abundance of β-alanine and the simple distribution of amino acids found in the CI1 chondrites Orgueil and Ivuna has been attributed to a distinct chemical composition on the CI parent body compared to the CM parent body (Ehrenfreund et al. 2001), the amino acid data from this study suggest that this might not be the case, and extended aqueous alteration, rather than differences in the chemical composition of the parent body, may be responsible for the high relative abundance of β-alanine over other α-amino acids in CI1 chondrites.

Possible Formation Pathways for Amino Acids Found in Meteorites

As stated above, β-amino acids cannot be formed by the Strecker mechanism, but can be synthesized by Michael addition of ammonia to α,β-unsaturated nitriles followed by reduction/hydrolysis in water, a mechanism that was proposed for the formation of β-alanine in spark discharge experiments (Miller 1957). For example, the formation of β-alanine by Michael addition of ammonia to cyanoacetylene during aqueous alteration on the parent bodies of CI, CM, and CR chondrites may have occurred. The reaction of HCN, NH₃, and carbonyl compounds (aldehydes and ketones) leads to the formation of α-amino, imino, and α-hydroxy acids in aqueous solution via the Strecker-cyanohydrin pathway and all of these expected Strecker byproducts have been identified in the Murchison meteorite (Peltzer and Bada 1978; Peltzer et al. 1984; Lerner and Cooper 2005). Alternative sources for the amino acids in carbonaceous chondrites such as irradiation of ices formed in the interstellar medium (Bernstein et al. 2002; Elsila et al. 2007; Hudson et al. 2008) and Fischer-Tropsch Type (FTT) catalysis (Hayatsu et al. 1971) that do not require aqueous activity for their formation in the parent body have also been proposed. Several α-amino acids and one β-amino acid (β-alanine) have been synthesized by Fischer-Tropsch, however, no γ- or δ-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. Some meteoritic γ- and δ-amino acids found in Murchison could be produced from the hydrolysis of lactams that have been identified in this meteorite (Cooper and Cronin 1995; Pizzarello et al. 2006). Although FTT products are expected to be predominantly straight-chain hydrocarbons (Hayatsu and Anders 1981), the opposite is found for the carbonaceous chondrites where branched-carbon chain amino acid isomers dominate over straight-carbon chain isomers (e.g., Murchison, LEW 90500, EET 92042, QUE 99177, GRA 95229) or are approximately equal abundance (Orgueil, MET 01070, SCO 06043, GRO 95577, LON 94102) (see Fig. 5b) (Martins et al. 2007). The predominance of branched over straight-chain amino acid isomers has also been observed in Murchison and suggests that ion and radical stability was important during carbon chain formation of the interstellar amino acid precursors (Cronin and Chang 1993; Herbst 1995). Identifying variations in the structure of C₅ amino acids in meteorites has also been used to help constrain parent body processes for the formation of these compounds (Glavin et al. 2010).

Variations in C₅ Amino Acid Structure Between Carbonaceous Chondrite Groups

For this study, we calculated the relative molar abundances of the C₅ amino acid isomers in each meteorite as a function of amine position (α-, β-, γ, or δ-) and valeric acid carbon chain structure (n-, sec-, iso-, or tert-) to determine if any synthetic relationships or chemical patterns exist between the CI, CM, and CR chondrite groups. These results, normalized to the total number of possible structural isomers, are illustrated in Figs. 5a and 5b. The dashed line corresponding to a relative abundance of 1 indicates the expected ratio if there was an equal probability of forming all of the C₅ amino acid isomers. Therefore, for the purposes of this discussion, enhanced and depleted relative abundances refer to values that fall above and below the dashed line, respectively. The high relative abundances of C₅α-amino acids compared to β-, γ-, and δ-amino acids in the CM chondrites Murchison and LEW 90500 and the CR chondrites EET 92042 and QUE 99177 provide strong evidence that the C₅α-amino acids in these meteorites were formed by Strecker-cyanohydrin synthesis (Fig. 5a). However, the opposite trend is observed for the more aqueously altered type 1 chondrites Orgueil, SCO 06043, MET 01070 and GRO 95577, which have a much higher relative abundance of the γ- and δ-amino acids compared to the α- and β-amino acids (Fig. 5b). One possibility is that more extensive aqueous alteration in the type 1 carbonaceous chondrites results in an increase in the rate of hydrolysis of amino acid precursors yielding elevated levels of γ- and δ-amino acids in these carbonaceous chondrites. As mentioned earlier, some meteoritic γ- and δ-amino acids found in Murchison could have been produced directly from the hydrolysis of lactams present in the meteorite (Cooper and Cronin 1995; Pizzarello et al. 2006). It is possible that the high abundances of C₅γ- and δ-amino acids detected in type 1 carbonaceous chondrites could have formed from lactams by a similar process; however, lactams contain secondary (and not primary) amines and we are unable to confirm the presence of these compounds in the nonhydrolyzed meteorite extracts using our analytical method. Another possible explanation for the high relative abundances of C₅γ- and δ- amino acids is that these amino acids are more stable to thermal decomposition than the α- and β-amino acids. For example, it has been shown previously for the C₄ amino acids that α- and β-aminobutyric acid have much higher decarboxylation rates than γ-aminobutyric acid in aqueous solution at elevated temperatures (Li and Brill 2003). Future experiments will be required to determine if the relative abundances of the C₅ amino acids measured in this study reflect a similar amino acid decarboxylation rate trend.

In general, the relative abundances of the C₅β-amino acids in the type 1 chondrites seem to be slightly higher than in the type 2 and type 3 chondrites (Fig. 5a); however, unlike the C₃ amino acid β-alanine, we see no clear trend in C₅β-amino acid abundance with degree of alteration in the meteorites analyzed. LON 94102 has a unique C₅ amino acid distribution compared to the other CM2 chondrites and is only slightly enriched in α-, γ-, and δ-amino acids compared to β-amino acids. One possible explanation for this observation is that LON 94102 experienced more aqueous alteration than Murchison and LEW 90500 leading to an increased relative abundance of γ- andδ-amino acids due to their slower decomposition rates and/or increased production from the hydrolysis of lactams in LON 94102. However, bulk oxygen isotope data (see Fig. 4 of Clayton and Mayeda 1999) and recent isotopic analysis of the IOM in LON 94102 indicates that this meteorite is much less, not more, altered than Murchison and LEW 90500 (Alexander et al. 2010). Therefore, an alternate explanation is that LON 94102 originated on a chemically distinct parent body from these CM2 chondrites.

A comparison of the relative abundances of the C₅ amino acids as a function of carbon-chain (valeric acid) structure yields additional insight into the mechanism of formation of these amino acids. As illustrated in Fig. 5b, we find a predominance of branched-chain (sec- and iso-) over straight-chain (n-) C₅ amino acids in the CMs Murchison and LEW 90500 and the CRs EET 92042 and QUE 99177. However, this is not the case for the aqueously altered type 1 chondrites Orgueil, MET 01070, SCO 06043, and GRO 95577 where branched-chain amino acids are not favored over the straight-chain acids and roughly equal abundances of the n-, sec-, and iso-valeric acid structures are observed with only a slight depletion of tert-valeric acid (Fig. 5b). Once again, we find that LON 94102 has a distinct distribution of C₅ amino acids compared to other CM2 and CR2 chondrites with respect to the valeric acid chain structure with a distribution that is very similar to the type 1 chondrites. Since LON 94102 has been classified as a CM2, the relative distribution of C₅ amino acids in this meteorite is puzzling. With the exception of LON 94102, we conclude that aqueous alteration on the parent bodies of CM and CR chondrites changes the distributions of C₅ amino acids from predominately α-amino acids to predominately γ- and δ-amino acids with relative enrichments in the n- and tert-valeric acid structures. Although at present we have not identified the mechanism(s) responsible for the variations seen in the C₅ amino acid distributions for these carbonaceous chondrites, parent body alteration seems to be the most reasonable explanation.