Late Archean rise of aerobic microbial

Jennifer L. Eigenbrode*†‡ and Katherine H. Freeman*

*Department of Geosciences and Penn State Research Center, Pennsylvania State University, University Park, PA 16802; and †Geophysical Laboratory, Carnegie Institution of Washington, Washington, DC 20015

Communicated by John M. Hayes, Woods Hole Oceanographic Institution, Woods Hole, MA, August 30, 2006 (received for review January 18, 2006) We report the 13C content of preserved organic for a 150 zone. Thus, we focus this work on relative water-column depth, million-year section of late Archean shallow and deepwater sedi- and we present kerogen ␦13C data for a well preserved, 150 ments of the Hamersley Province in Western Australia. We find a million-year stratigraphic section from the late Archean, in 13C enrichment of Ϸ10‰ in organic carbon of post-2.7-billion-year- which we distinguish isotopic signatures of shallow versus deep- old shallow-water carbonate rocks relative to deepwater sedi- water sedimentary environments. We specifically evaluate this ments. The shallow-water organic-carbon 13C content has a 29‰ isotopic record after the pronounced depletion in organic 13C range in values (؊57 to ؊28‰), and it contrasts with the less signatures at 2.72 Ga, and we compare it with published isotopic variable but strongly 13C-depleted (؊40 to ؊45‰) organic carbon records from other regions to assess global changes in microbial in deepwater sediments. The 13C enrichment likely represents ecosystems during this important time. microbial habitats not as strongly influenced by assimilation of or other 13C-depleted substrates. We propose that con- Results tinued oxidation of shallow settings favored the expansion of Kerogen ␦13C values for 175 samples from a 2.72- to 2.57-Ga aerobic ecosystems and respiring organisms, and, as a result, section in the Hamersley Province, Western Australia, vary isotopic signatures of preserved organic carbon in shallow settings between Ϫ57‰ and Ϫ28‰ (Fig. 2 A–D), spanning the pub- approached that of photosynthetic biomass. Facies analysis of lished range for late Archean organic carbon. Large variations published carbon-isotopic records indicates that the Hamersley (5–16‰) occurred at 1- to 10-m scales, and they are correlated shallow-water signal may be representative of a late Archean with lithology (Fig. 2 A–D). For example, in the WRL1 core, global signature and that it preceded a similar, but delayed, 13C increases and decreases in dolomitic siltstone relative to shale are enrichment of deepwater deposits. The data suggest that a global- accompanied by enrichment and depletion of ␦13C in kerogens, scale expansion of oxygenated habitats accompanied the progres- respectively. Similar ␦13C–lithologic relationships have been sion away from anaerobic ecosystems toward respiring microbial observed for the 2.52- to 2.46-Ga Transvaal rocks on the communities fueled by oxygenic photosynthesis before the oxy- Kaapvaal Craton (9, 10), although they have a narrower range in genation of the atmosphere after 2.45 billion years ago. carbon isotopic compositions for bulk organic matter (␦org) range. organic carbon ͉ isotopes ͉ methanotrophy ͉ oxygen Facies were defined by sedimentological characteristics, and

they were grouped into shallow-water deposits and deeper slope GEOLOGY esolving changes in early ecosystems and their geochemical and distal basin deposits. Briefly, the upper Carawine Dolomite Rcontext is critical to understanding both the means and pace with abundant stromatolites is well recognized as a shelf deposit of events that culminated in atmospheric oxygenation beginning having platform architecture (11). The Tumbiana very shallow possibly as early as 2.45 gigaannum (Ga) before the present, carbonate environment flanked a large, southerly subsiding rift based on the sulfur isotopic record (1) but certainly by 2.3 Ga, basin (12), and it received a high flux of terrestrial detritus. The based on other geochemical observations (2). Oxygenic photo- Tumbiana Formation has been interpreted as a fluvial– synthesis must have evolved before this event, but conjectures for lacustrine deposit (13) (but others suggest that it may also its onset range from pre-3.5 Ga (3, 4) to 2.7 Ga (5, 6) to 2.4 Ga contain shallow marine deposits; see ref. 14 and Supporting (7). Whenever it occurred, its inevitable impact on ecosystems Discussion, which is published as supporting information on the must have been profound. PNAS web site). The Warrie Member overlies a transgressive The 13C content of sedimentary organic matter represents shoreline deposit in northern exposures, and it contains sedi- mixed contributions from many biological materials. Isotopic mentary structures consistent with a shallow marine shelf envi- signatures within modern and Phanerozoic marine sediments ronment (see Supporting Discussion). Carbonates of the Witte- generally fall in a narrow range, and they reflect those of noom Dolomite and lower Carawine Dolomite largely consist of phytoplankton (8) because heterotrophic respiration imparts platform (shelf) detritus within turbidite deposits (11). Finely small, if any, isotopic shifts in the preserved organic matter. In laminated, dolomite-rich black shales (10–60 wt % dolomite), contrast, late Archean organic carbon displays a wide range in interbedded with platform carbonate deposits in the Carawine 13 ␦13 ϭ 3 13 ͞13 C content (expressed as C, ‰ 10 ( Rsample Rstandard Dolomite, reflect lagoonal or other localized, restricted- Ϫ1), and 13R ϭ 13C͞12C; Fig. 1), with values generally below Ϫ35 circulation settings on the platform. Other finely laminated black to Ϫ40‰ in shales (9, 10) and even lower in some shallow-water shales and dolomitic and tuffaceous siltstones characterize deep- deposits at Ϸ2.7 Ga (Fig. 1). As hypothesized by Hayes (5, 6), strong 13C depletion in organic matter, either in bulk or as the insoluble fraction (kerogen; ␦ker), indicates that late Archean Author contributions: J.L.E. and K.H.F. designed research; J.L.E. performed research; K.H.F. ecosystems incorporated 13C-depleted substrates, particularly contributed new reagents͞analytic tools; J.L.E. analyzed data; and J.L.E. and K.H.F. wrote methane. the paper. If kerogen ␦13C patterns in different sedimentary environ- The authors declare no conflict of interest. ␦ ␦ ments reflect differences in the dominant pathways for carbon Abbreviations: carb, carbon isotopic composition of inorganic carbon of carbonates; ker, ␦ assimilation and cycling, then the isotopic record may also reflect carbon isotopic composition of kerogen; org, carbon isotopic composition of bulk organic matter; Ga, gigaannum before present; Rubisco, ribulose-bisphosphate carboxylase͞ secular changes in microbial communities within different en- oxygenase. vironmental settings. Unlike carbonate isotopic records, which ‡To whom correspondence should be addressed at: Geophysical Laboratory, Carnegie are highly sensitive to marine versus lacustrine distinctions, Institution of Washington, 5251 Broad Branch Road NW, Washington, DC 20015. E-mail: organic-carbon isotopic signatures record the influence of bio- [email protected]. logical productivity and recycling within and below the photic © 2006 by The National Academy of Sciences of the USA

www.pnas.org͞cgi͞doi͞10.1073͞pnas.0607540103 PNAS ͉ October 24, 2006 ͉ vol. 103 ͉ no. 43 ͉ 15759–15764 Downloaded by guest on September 30, 2021 13 Fig. 1. Compilation of published kerogen and total organic carbon ␦ C values (␦org) for all sedimentary rock types. Data from this work are included, and the geochronology is updated (see also Supporting References, which are published as supporting information on the PNAS web site). The top curve is the mean inorganic carbon ␦13C composition for marine carbonate rocks published by Shields and Veizer (16). Values are not corrected for postdepositional alteration.

water slope deposits, whereas massive, pyritic black shales were Discussion deposited in basin settings more distal to the shelf. Observational data and computational models suggest that the The isotopic record shown in Fig. 2 A–D reveals both facies- Archean atmosphere was composed of N2,CO2,CH4, and H2 and time-related signals. Over the entire period, the shallow (17, 18), consistent with the suggestion that the carbon cycle was ␦ facies show a wide range of ker values, with dolomitic shales of dominated by autotrophy, fermentation, acetogenesis, and 13 the restricted platform facies 2–14‰ depleted in C relative to methanogenesis (19, 20). Moderately high CO2 levels (up to 10ϫ the platform carbonates immediately above and below. Kerogens the present atmospheric level) (18, 20, 21) provided ample 13 ␦ in deepwater shales are strongly depleted in C, whereas ker substrate for autotrophs, such that isotope fractionation likely values for the slope facies are intermediate to those for basin and approached maximum values. Assuming an 8‰ difference shallow-water carbonate shelf facies (Fig. 2E). Sedimentology between carbonate and dissolved CO2, carbon fixation by suggests that the slope facies may have received organic matter ribulose-bisphosphate carboxylase͞oxygenase (Rubisco) and the transported and deposited from shallower environments, which Calvin–Benson cycle could account for ␦ker values as low as was mixed with autochthonous inputs. Ϫ37‰, as in most modern eukaryotes (form IB; maximum ␦ To assess temporal changes in ker, we merged the records in Fig. fractionation, Ϸ29‰; refs. 22 and 23). Most modern cyanobac- 2 by using correlations based on available U-Pb geochronology, teria have Rubisco form IB, whereas some marine species have impact spherule beds, and the overall framework of stratigraphic form IA (maximum fractionation, Ϸ24‰; ref. 24). Still, cya- sections (Fig. 3 and Fig. 6, which is published as supporting nobacteria biomass typically yields less fractionation (e.g., 17‰ ␦ information on the PNAS web site). The range of ker values for Synechococcus sp.; ref. 25) because of other factors. If broadens significantly in younger strata over the course of 100 modern cyanobacterial biology and physiology are representa- million years. This broadening is largely the result of a Ϸ23‰ tive of their ancestors, then cyanobacteria most likely would have ␦ enrichment in carbonate-hosted ker values, which approach contributed ␦ker values of 32–25‰. Photoautotrophic purple Ϫ28‰ near the top of the section, and the relative stability of bacteria, which have Rubisco form II (maximum fractionation, deepwater-basin ␦ker values from 2.7 Ga to 2.6 Ga. Regardless of its 19.5‰; ref. 26), would yield similar values. Fractionations are environmental setting, the lowermost unit (i.e., the Tumbiana even lower (Յ14‰; ref. 27) for anoxygenic photoautotrophic carbonate unit) is characterized by very shallow deposition, in which bacteria that employ either the 3-hydroxypropionate pathway or restricted conditions may have existed locally. The extreme ␦ker the reductive tricarboxylic acid cycle. Iron abundances (Ͻ2wt%; values recorded in this unit define the first of two prominent refs. 28 and 29) reported for Hamersley carbonates do not 10–12‰ steps in carbonate-hosted kerogen (Fig. 3) at 2.72 and 2.6 suggest a profuse supply of Fe(II) necessary for an Ga. Both enrichments are exceptional, especially compared with dominated by anoxygenic Fe(II)-oxidizing photoautotrophs, like the notable Cretaceous-to-modern rise in the ␦13C of marine that envisioned for banded iron formation deposits (7, 30). organic carbon, which was only 5–7‰ (15). In between, the ␦ker Sulfur in primary minerals is nearly absent in the Wittenoom signatures for the Warrie Member and Carawine Dolomite are Dolomite (most pyrite occurs in isolated roll-up laminae or as largely consistent (Ϫ38 Ϯ 3‰) for Ϸ50 million years, despite a coatings of clasts), suggesting that sulfur was not the dominant major marine transgression and bolide impact. Shallow marine redox partner for carbon in this system. Considering the limited 13 carbonate ␦ C(␦carb) values remain near 0‰ throughout the study availability of electron donors in shallow environments, the ␦ interval, reflecting stability of global average carb at this time (16) highest ␦ker values probably represent inputs from ancestral (Figs. 1 and 3). cyanobacteria, which is consistent with the finding of molecular All samples experienced relatively low thermal stress (metamor- fossils (2␣-methyl hopanes) most likely from cyanobacteria in phism no greater than prehnite–pumpellyite grade; subgreen- late Archean rocks (31, 32), although difficulties interpreting the schist), consistent with nongraphitic kerogen H͞C ratios of 0.15– molecular record have been noted (7). 0.30). Losses of carbon associated with dehydrogenation could have In all of the above scenarios, photoautotrophy under a high 13 resulted in an isotopic enrichment of 2–2.5‰ (10) for all samples. CO2 atmosphere cannot explain C-depleted values below We see no systematic relationship between total organic carbon Ϫ37‰. The Archean anaerobic communities envisioned by abundance and ␦ker values (see Supporting Discussion); thus, there Kasting, Siefert, and coworkers (17–20) included microorgan- is no obvious evidence for thermally related patterns in ␦ker. There isms that degraded and recycled carbon substrates. In modern is no physical or chemical evidence for residues from nonindigenous settings with anaerobic habitats, such processes yield a 13C- oil migration. depleted biomass. Specifically, recycling of fermentation-

15760 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0607540103 Eigenbrode and Freeman Downloaded by guest on September 30, 2021 GEOLOGY

Fig. 2. Depositional facies and ␦ker profiles. (A–D) Shown are shallow-water carbonates (circles) deposited on a marine platform or other shallow setting (yellow) or turbidite (orange), mixed shale͞carbonate lithologies (triangles), including shallow-water shale facies (blue) or deepwater slope facies (green), massive black shales of deepwater basin facies (black, squares), basalts (pink), carbonate breccia (white), and fluvial sandstones (brown). Symbols are larger than analytical reproducibility. Arrows denote impact-event beds. (A) WRL1 core. (B) RHDH2a core. (C) SV1 core. (D) Tumbiana Formation (Fm.) outcrop. (E) Histograms of ␦ker with respect to depositional facies.

13 derived CO2 or acetate can yield a C-depleted biomass, samples. Methane oxidation requires electron acceptors such as 2Ϫ 2Ϫ especially for anaerobes using the acetyl-CoA pathway (33). O2 or another suitable oxidant, such as SO4 (37, 38) or NO3 Nonfermenting acetogens, which compete for H2 with other (39), as observed for anaerobic methane-oxidizing consortia. organisms that possess greater energy-yielding metabolisms The dramatic drop in ␦ker values at Ϸ2.7 Ga (Fig. 1) may (e.g., methanogens), can produce acetate that is extremely represent a sharp increase in the availability of oxidized electron 13 C-depleted relative to CO2 (Ϸ59‰) (34). Incorporation of this acceptors that was directly (6) or indirectly (40) a consequence product into biomass can also yield low ␦org values. of oxygen production. If so, the Ϸ2.7-Ga ␦org event is evidence Similarly, methane oxidation and assimilation, either aerobi- that oxygenic photosynthesis originated by that time. 13 cally or anaerobically, can produce a biomass with extremely low If the extremely C-depleted ␦ker values of late Archean ␦13C values (5, 6, 35, 36), such as those recorded in our oldest shallow-water facies represent the onset of methane assimilation

Eigenbrode and Freeman PNAS ͉ October 24, 2006 ͉ vol. 103 ͉ no. 43 ͉ 15761 Downloaded by guest on September 30, 2021 Fig. 3. Composite graph of ␦ker data versus time. The curve represents the average in carbonate-hosted ␦ker, but it is not quantitative. For symbol and color key, see Fig. 2. U-Pb geochronology and impact-event beds constrain correlations between cores and outcrops (OC). Inorganic-carbon isotope values from 13 carbonates (␦ Ccarb) are near 0‰ for nonevaporitic units of the Tumbiana Formation (0.0 Ϯ 1.2‰, n ϭ 36) and Carawine (0.0 Ϯ 0.8‰, n ϭ 32) and Wittenoom 13 Dolomites (Ϫ0.5 Ϯ 2.0‰, n ϭ 66), and they are consistent with ␦ Ccarb for well preserved marine carbonates from five other Archean provinces (Ϫ0.3 Ϯ 0.7‰, n ϭ 324) (see Supporting Discussion).

triggered by oxygenic processes (6), we anticipate a subsequent provide additional evidence for oxygen oases in the late Ar- trend, at least in shallow settings, from an ecosystem that chean, and they place the development of oxygen oases in the recycled methane toward an ecosystem based on photoautotro- context of specific depositional environments. The presence of 33 34 phic carbon assimilation and heterotrophic carbon cycling em- environmental O2 is consistent with ␦ S and ␦ S records from ploying electron acceptors that are thermodynamically favored the Ϸ2.63-Ga Hamersley strata that have been interpreted as over fermentative pathways. This expectation is consistent with indicating sulfate-limited sulfate respiration in shallow (but not the Ϸ23‰ enrichment that we observe in carbonate-hosted deep) settings under an anoxic atmosphere (41). An ecological kerogens (Fig. 3). shift toward greater importance of heterotrophic respiration The enrichment of 13C in shallow-water carbonate products is associated with oxygen oases is also consistent with rapid best explained by a decrease in the importance of anaerobic diversification among such organisms some time after the onset processes most likely involving methane recycling and an in- of oxygenic photosynthesis, as suggested in RNA-based phylo- crease in heterotrophic respiration (Fig. 4). A decline in the genetic patterns (42). Our ␦13C data from the Hamersley suc- fermentative supply of hydrogen to methanogenic and aceto- cession document profound changes in shallow-water microbial genic communities must have limited carbon recycling by these ecosystems at the end of the Archean, and they suggest the anaerobic organisms in shallow waters. In deeper settings, ␦ker beginnings of global carbon-cycle reorganization instigated by values remained highly 13C-depleted, indicating that these en- the sustained release of molecular oxygen into an otherwise vironments continued to be dominated by anaerobic communi- anaerobic world. ties and were not (yet) influenced by the major changes in carbon Available data for other localities indicate that the facies–␦ker cycling taking place in shallow waters. relationship is an Archean feature and that the shallow-water Atmospheric oxygenation by Ϸ2.3 Ga (1, 2) must have been enrichment documented by our data may not have been unique preceded by oxidation of ocean surface waters, which likely to the Hamersley Province. When ␦org values from Fig. 1 are began with phototrophic havens in isolated shallow-water envi- sorted by environment based on published lithofacies descrip- ronments (6). These so-called ‘‘oxygen oases,’’ characterized by tions (Fig. 5) and data from sections known to be highly altered local production of O2 and oxidized substrates, contrasted with are removed, a consistent pattern emerges. For shallow facies, the reducing atmosphere and deeper ocean (6, 40). Our data the lowest ␦org values (less than Ϫ45‰) are exclusively from the

15762 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0607540103 Eigenbrode and Freeman Downloaded by guest on September 30, 2021 Fig. 4. Schematic illustration of interpreted ecological changes in carbon 13 cycling and their imprint on the late Archean Hamersley ␦ Cker record. An- aerobic microbial processes in strongly reducing conditions (arrows 1–3) led to 13C depletion relative to photosynthate in preserved organic carbon like that observed for deepwater facies throughout the study interval. By 2.72 Ga, incipient oxygenation of very shallow habitats enabled the incorporation of 2Ϫ methane into biomass by organisms using O2,SO4 , or other suitable electron acceptors (arrows 4–5), producing extremely 13C-depleted biomass. Contin- ued oxidation of shallow-water environments after 2.72 Ga favored microbial respiration (arrows 6–7), which limited the supply of acetate and hydrogen (arrow 1) to methane production and acetatogenesis. Because organic-matter degradation by respiration is accompanied by little isotopic discrimination, the isotopic signature of organic carbon preserved in shallow environments approached values of photosynthetic biomass by Ϸ2.60 Ga.

Hamersley (Fig. 5A), likely reflecting preservational bias and possibly limited occurrence of similar facies and ecosystems. Even so, Kaapvaal shallow-water facies ␦org values show a less distinct and more protracted temporal 13C enrichment. Other shallow-water ␦org values come from the Hamersley, the Wabi- Fig. 5. Temporal distribution of ␦org values of the global record separated goon Belt (Superior Province), and the Belingwe Greenstone into shallow (A) and deepwater (B) facies. Data were reported previously in Belt (Zimbabwe Craton), and they have ␦org values similar to the literature for the Hamersley Province (Pilbara Craton; gray), Kaapvaal

those of the mid-Archean. These values account for the upward Craton (black), Zimbabwe (white), and Superior Province (white with X) and GEOLOGY spread in ␦org data of the 2.7- to 2.6-Ga global record (Fig. 1). in this work (Hamersley Province; gray with black outline; restricted platform Although some 13C enrichment in these data can be attributed facies not included in A). Data from known highly metamorphosed units, units to alteration during Greenschist metamorphism, the isotopic of uncertain facies, and oxide-BIF units were excluded. Dashed lines approx- ␦ record suggests that shallow-water ecosystems dominated by imate the expected org composition if organic carbon were derived solely from photoautotrophs. methane assimilation may have been highly sensitive to envi- ronmental conditions and possibly favored in restricted, stro- 13 matolitic settings. Notably, the Hamersley C enrichment in ronments over this time. Our interpretation is supported by shallow-water units over time coincides with the regional ex- extensive iron deposition and diminishing mass-independent pansion of carbonate shelf environments. Thus, the Hamersley isotope signature in pyrite sulfur from deepwater shales (1) record probably reflects localized-to-regional effects driven by beginning at Ϸ2.45 Ga. Perhaps the lag between shallow and regional tectonic evolution that nonetheless also reflect the deepwater isotopic enrichment reflects a shift in cyanobacterial global rise in oxygenation. Additional, facies-specific high- ␦ habitats from sediments in the photic zone of shallow environ- resolution ker records will help to provide a better understand- ments to photic-zone water columns inclusive of deepwater ing of environmental influences on late Archean and early environments. Proterozoic ␦ patterns. ker In summary, our facies-specific dissection of the organic- Temporally resolved ␦ records for deepwater facies are org carbon isotopic record reveals temporal and spatial patterns that mostly limited to the Kaapvaal and Hamersley regions, although document profound late Archean and early Proterozoic biogeo- 2.69-Ga sulfidic shales with ␦org values less than or equal to Ϫ45‰ have been cited for the Keewatin Group (Superior chemical changes preceding atmospheric oxygenation after 2.45 Ga. Specifically, facies-resolved isotopic records indicate that Province) (35, 43). The 2.72- to 2.59-Ga Hamersley ␦org record (see Fig. 3) shows little variability in deepwater facies; however, microbial communities in shallow environments led a global- scale transition away from purely anaerobic communities. The after 2.59 Ga, other Hamersley and Kaapvaal ␦org data indicate a deepwater temporal 13C enrichment of Ϸ13‰ (Fig. 5B). isotopic depletion characteristic of anaerobic ecosystems and Initiation of this shift lags behind the Hamersley shallow-water methane-assimilating communities was gradually replaced by 13 ␦org enrichment by Ͼ100 million years, yet both approached C-enriched signatures, consistent with a growing importance of Ϫ28‰, approximating the expected photosynthate value, be- cyanobacterial inputs accompanied by heterotrophic respiration. tween Ϸ2.45 and 2.3 Ga. These enrichment patterns suggest that Ultimately, this change was fueled by increased availability of by Ϸ2.45 Ga, the deepwater environments experienced changes electron acceptors derived from oxygenic photosynthesis in in cycling of photosynthate carbon in the upper water column photic zones. These data support the conclusion that oxygenic much like the shift that occurred in shallow environments. We photosynthesis, originating some time before 2.72 Ga, eventually suggest that the 13C enrichment reflects decreased dominance of triggered the rise of aerobic ecosystems and fueled their expan- anaerobic processes associated with the expansion of photoau- sion from shallow settings into the photic zones of deepwater totrophic oxygenation into the photic zones of offshore envi- environments between 2.72 and Ϸ2.45 Ga.

Eigenbrode and Freeman PNAS ͉ October 24, 2006 ͉ vol. 103 ͉ no. 43 ͉ 15763 Downloaded by guest on September 30, 2021 Materials and Methods Marilyn Fogel, Roger Summons, David DesMarais, Yanan Shen, George We used conventional methods to isolate kerogen fractions from Cody, Jim Kasting, Jenn Macalady, Shuhei Ono, and others for con- 13 structive comments on early manuscript drafts. We are grateful to Rio rocks, measure ␦ C values of kerogen (␦ker), and determine organic and inorganic carbon and sulfur elemental abundances Tinto Exploration for providing access to core samples, Matt Hurtgen for field assistance, and Mark Barley and Brian Krapez for field guidance. (Supporting Materials and Methods; Figs. 7 and 8 and Table 1, This work was supported by National Science Foundation Grants which are published as supporting information on the PNAS web EAR-00-73831 and EAR-00-80267 (both to K.H.F.), an American site). Stratigraphy and sedimentology were derived from our Association for Petroleum Geologists Grant-in-Aid (to J.L.E.), a Penn- field and core descriptions, and they were supplemented by sylvania Space Grant Consortium fellowship [National Aeronautics and observations published in the geologic literature. Drill core and Space Administration (NASA) Grant NGT5-40090] (to J.L.E.), NASA outcrop rock samples were collected from a 2.72- to 2.57-Ga Astrobiology Institute Cooperative Agreement NNA04CC06A (Penn section in the Hamersley Province, Western Australia. State Astrobiology Research Center) (to K.H.F. and J.L.E.), and NASA Astrobiology Institute Cooperative Agreements NNA04CC09A (Carne- We thank John M. Hayes for sharing his insight, providing advice on gie Institution of Washington) and the Geophysical Laboratory, Carne- revisions, and handling the review of this manuscript. We also thank gie Institution of Washington (to J.L.E.).

1. Farquhar J, Wing B (2003) Earth Planet Sci Lett 213:1–13. 24. Scott K, Schwedock J, Schrag DP, Cavanaugh CM (2004) Environ Microbiol 2. Holland H (1984) The Chemical Evolution of the Atmosphere and Oceans 6:1210–1219. (Princeton Univ Press, Princeton), p 598. 25. Popp B, Laws E, Bidigare R, Dore J, Hanson K, Wakeham S (1998) Geochim 3. Cloud P (1983) in Earth’s Earliest Biosphere: Its Origin and Evolution, ed Schopf Cosmochim Acta 62:66–77. J (Princeton Univ Press, Princeton), pp 14–31. 26. Robinson J, Scott K, Swanson S, O’Leary M, Horken K, Tabita F, Cavanaugh 4. Rosing M, Frei R (2004) Earth Planet Sci Lett 217:237–244. C (2003) Limnol Oceanogr 48:48–54. 5. Hayes J (1994) in Early Life on Earth, Nobel Symposium No. 84, ed Bengston 27. Sirevåg R (1995) in Anoxygenic Photosynthetic Bacteria: Advances in Photosyn- S (Columbia Univ Press, New York), pp 220–236. thesis, eds Blankenship RE, Madigan MT, Bauer CE (Kluwer, Dordrecht, The 6. Hayes J (1983) in Earth’s Earliest Biosphere: Its Origin and Evolution, ed Schopf Netherlands), Vol 2, pp 871–883. J (Princeton Univ Press, Princeton), pp 291–301. 28. Lindsay J, Brasier M (2002) Precambrian Res 114:1–34. 7. Kopp R, Kirshvink J, Hilburn I, Nash C (2005) Proc Natl Acad Sci USA 29. Veizer J, Clayton R, Hinton RW, Brunn V, Mason T, Buck S, Hoefs J (1990) 102:131–136. Geochim Cosmochim Acta 54:2217–2729. 8. Freeman K (2001) in Stable Isotope , eds Valley JW, Cole DR 30. Konhauser K, Hamade T, Raiswell R, Morris R, Ferris F, Southam G, Canfield (Min Soc Am, Washington, DC), Vol 43, pp 579–605. D (2002) Geology 30:1079–1082. 9. Strauss H, Beukes N (1996) Precambrian Res 79:57–71. 31. Summons R, Jahnke L, Hope J, Logan G (1999) Nature 400:554–557. 10. Des Marais D (2001) in Stable , eds Valley JW, Cole DR 32. Brocks J, Logan G, Buick R, Summons R (1999) Science 285:1033–1036. (Min Soc Am, Washington, DC), Vol 43, pp 555–578. 33. House C, Schopf J, Stetter K (2003) Org Geochem 34:345–356. 11. Simonson B, Schubel K, Hassler S (1993) Precambrian Res 60:287–335. 34. Gelwicks J, Risatti J, Hayes J (1989) Org Geochem 14:441–446. 12. Blake T, Buick R, Brown S, Barley M (2004) Precambrian Res 133:143–173. 35. Schoell M, Wellmer F (1981) Nature 290:696–699. 13. Buick R (1992) Science 255:74–77. 36. Hinrichs K-U (2002) Geochem Geophys Geosyst 3:1042. 14. Sakurai R, Ito M, Ueno Y, Kitajima K, Maruyama S (2005) Precambrian Res 37. Orphan V, House C, Hinrichs K-U, McKeegan K, DeLong E (2001) Science 138:255–273. 293:484–487. 15. Arthur M, Dean W, Claypool G (1985) Nature 315:216–218. 38. Hinrichs K-U, Hayes J, Sylva S, Brewer P, DeLong E (1999) Nature 398:802–805. 16. Shields G, Veizer J (2002) Geochem Geophys Geosyst 3:1031. 39. Raghoebarsing A, Pol A, van de Pas-Schoonen K, Smolders A, Ettwig K, 17. Kasting J (2005) Precambrian Res 137:119–129. Rijpstra W, Schouten S, Sinninghe J (2006) Nature 440:918–921. 18. Kharecha P, Kasting J, Siefert J (2005) 3:53–76. 40. Kasting J, Holland H, Kump L (1992) in The Proterozoic Biosphere: A 19. Kasting J, Siefert J (2002) Science 296:1066–1068. Multidisciplinary Study, eds Schopf J, Klein C (Cambridge Univ Press, Cam- 20. Pavlov A, Kasting J, Brown L, Rages K, Freedman R (2000) J Geophys Res bridge), pp 159–164. 105:981–990. 41. Ono S, Eigenbrode J, Pavlov A, Kharecha P, Rumble D, Kasting J, Freeman 21. Nakamura K, Kato Y (2004) Geochim Cosmochim Acta 68:4595–4618. K (2003) Earth Planet Sci Lett 213:15–30. 22. Roeske C, O’Leary M (1984) Biochemistry 23:6275–6284. 42. Sheridan P, Freeman K, Brenchley J (2003) Geomicrobiol J 20:1–14. 23. Tabita F (1999) Photosynth Res 60:1–28. 43. Strauss H (1986) Geochim Cosmochim Acta 50:2653–2662.

15764 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0607540103 Eigenbrode and Freeman Downloaded by guest on September 30, 2021