In any discussion of biosignatures, one of the first and most pressing questions is how we can differentiate between biosignatures and ‘abiosignatures’. In this usage, the term ‘abiosig-natures’ refers to a signature derived from an abiotic (i.e. chemical or physical) process. Only by demonstrating that a given signature can only be produced by biological activity can we conclude that it represents evidence of life. Therefore, in order to interpret a mineral, organic or isotopic signature as a biosignature, we must also be certain that it cannot be produced by abiotic processes, i.e. that it does not mimic a biosignature. It is worth noting that such interpretations should always be made with the awareness that as our understanding increases, what was once assumed to be a biosig-nature may cease to be considered so in light of new information.

The stability of minerals over long periods of time and a wide range of conditions means that they have the potential to preserve biosignatures. Chains of single-domain magnetite crystals produced by magnetotactic bacteria represent one example of a mineral biosignature that has received attention recently (Sukumaran 2005). These bacteria form magnetite crystals as part of a magnetosome (Fig. 1), which allows them to orient their movement and stay at the geochemical interface between oxic and anoxic conditions in the sediments where their optimal growth conditions occur (Kopp and Kirshvink 2008). These magnetofossils are present in the geologic record back to the Cretaceous, and potentially to the Archean (Chang and Kirshvink 1989; Kopp and Kirshvink 2008). However, recognition of these crystals as biosignatures requires careful assessment in order to differentiate them from abiotically produced crystals (Kopp and Kirshvink 2008). The observation of magnetite crystals has been proposed as evidence for life in the meteorite ALH84001 (Thomas-Keprta et al. 2002), but this argument has yet to be accepted as definitive. Ongoing study seeks to further define the biogenic characteristics of magnetofossils (Kopp and Kirshvink 2008) and to add other lines of evidence to support their biogenicity, such as isotopic ratios (Mandernack et al. 1999).

Because life on Earth is based on carbon chemistry, organic compounds have a great deal of potential as biosig-natures. In fact, it was organic geo-chemists who originally used the term “biological marker compounds” or “biomarkers” to describe organic compounds that could survive over the long periods and potential diagenetic alterations associated with geological processing, and still retain an ability to be related to a specific biological origin (Eglinton et al. 1964). Determination of effective biomarker compounds requires that their production by modern organisms be characterized. For instance, one class of biomarker compounds produced by bacteria, the hopanoids, is found throughout the geologic record (Summons et al. 1999; Brocks et al. 2003), and one hopanoid class in particular, the 2-methyl hopanoids, have been used to illuminate the role of cyanobacteria in the rise of photosynthesis (Summons et al. 1999). However, as with all biosignatures, understanding how these signatures are synthesized by modern organisms is crucial to their interpretation. Recent research in measuring the production of bacteriohopanpolyols, particularly 2-methyl bacteriohopanpolyol in modern cyanobacteria, has found that, although they are produced by the majority of species, the most prolific marine picocyanobacteria do not produce them, whereas two prolific marine nitrogen-fixing species do (Talbot et al. 2008). This new information does not call into question the utility of these compounds as a biomarker of microbial life in general, but it does raise important questions regarding the extent to which they can be used as a marker of specific species.

Because of the complexity of organic chemistry and its fundamental role in life as we know it, organic bio-markers are of great interest as a focus of the search for life on other astronomic bodies. The state of knowledge concerning organic biosignatures and their application was recently summarized by Summons et al. (2008), who describe a series of general patterns of measurable molecular biosignatures that may be used to identify life. These patterns include i) molecular distributions that display a preference for one specific spatial arrangement of atoms (known as a stereoisomer) over another; ii) molecules constructed of repeating constitutional sub-units (e.g. proteins, which are constructed of amino acids); and iii) uneven distributions of structurally related compounds (biological systems will often produce very large amounts of a small sub-set of potential compounds rather than a continuum of compounds built from smaller sub-units). These general patterns have been developed from our observations of life on Earth, but by focusing on general properties of biological compounds we may be able to develop a framework that will also enable us to recognize signatures of life that is based on different chemistries.

Isotopic biosignatures are changes in stable isotope ratios that can be directly related to biological activity. Specific isotopic distribution patterns are also included in the Summons et al. (2008) description of organic biosignatures. However, isotopic biosignatures can exist for nearly all elements because, with a few exceptions, all elements comprise an array of stable isotopes. The term isotope comes from the Greek meaning ‘same place’ because all isotopes of a given element occupy the same place in the periodic table. This means that isotopes of an element all have the same basic chemical properties, the primary difference being their atomic mass, which varies according to the number of neutrons in the nucleus. In radioactive isotopes the nucleus is unstable and decays via one of several mechanisms. The extent of radioactive decay is the basis for much of our dating of geological materials (Faure 1986) and has therefore contributed a great deal to our understanding of the timing of biosignatures observed over the history of life on Earth. However, because the abundance of radioactive isotopes is constantly changing, they cannot preserve a direct isotopic biosignature over long periods of time. In contrast, stable isotopes, isotopes in which the mass difference does not cause the atom to be unstable, do not change over time and can therefore preserve a biosignature (Faure 1986). For example, carbon has two stable isotopes, carbon 12 and carbon 13, which compose approximately 98.9% and 1.1%, respectively, of any mass of carbon on Earth. For any sample of carbon-bearing molecules, we can determine the precise ratio of the isotopes of carbon, and express it in delta notation as follows:

δ13C = (Rsample – Rstandard)/Rstandard * 1000 in units of per mil (%) (1),

where R is the ratio ¹³C/¹²C. This expression for isotopic ratios can be generalized to any atom by replacing ¹³C and ¹²C with the isotopes of interest. Note that by convention in geo-chemistry, the heavier (and less abundant) isotope is divided by the lighter (more abundant) isotope. In our example, ¹³C has the same chemical properties as ¹²C, but is one atomic mass unit heavier, hence when it is involved in processes in which mass plays a factor, it behaves in different ways. These differences in behaviour result in isotopic ‘fractionation’ of the ratio of ¹³C to ¹²C in a sample. This isotopic fractionation can be described by a fractionation factor, which is characteristic for the process:

α = 1000 + δ¹³C_A/1000 + δ¹³C_B (2),

where A and B are two components or phases of a chemical process, such as the reactant and product, or gas phase and aqueous phase. Both the stable isotopic ratios and the fractionation factors that change them can be characteristic for different processes and have been used as biosignatures of life on Earth.

Biosignature research uses all of the foregoing types of signatures, ideally in combination, to generate an interpretation of the history of life. Developing this interpretation is contingent on studies focused in two areas. The original focus of biosignature research was the acquisition of samples from the entire geologic record that could be used to provide a history of life on Earth. Canada is a logical location for this type of research because our rocks record a broad swath of geologic history, including some of the oldest rocks known. The second focus of the research requires study of modern analogues to aid in the interpretation of the geological samples. In other words, to interpret the signatures observed in the rock record, we must have some grasp of the processes that created those signatures. By using examples of modern systems that are analogous to past systems, we can explain the mechanisms by which biosignatures form, we can differentiate between biosignatures and abiosignatures, and we can determine what these signatures tell us about life and/or its environment at the time. The remainder of this paper highlights some of the current research on both of these fronts, and how it is contributing to our understanding of the origins and evolution of life.

One of the most fundamental unanswered questions concerning life on Earth is the timing of its origin. Because Canada hosts some of the oldest rocks on Earth, e.g. the Nuvvuagittuq greenstone belt in Northern Quebec (Cates and Mojzsis 2007; O’Neil et al. 2008), this country is well positioned to contribute samples that can provide unique insights into this question. Interpreting signatures in these ancient samples as evidence of early life on Earth hinges on the issue of biosignatures versus abiosignatures. In general, the geologic record shows that organic carbon is isotopically depleted in ¹³C compared to inorganic carbon. Schidlowski (2001) summarized the data on organic and inorganic isotopes (Fig. 2), and stated that isotopically depleted organic carbon (δ¹³C <-16%) is a bona fide biological signature. Furthermore, it has been widely stated that isotopic depletion in ¹³C is a result of biological activity. However, this conclusion is now being reconsidered in the light of recent research demonstrating that abiogenic hydrocarbon formation via chemical reactions between water, dissolved inorganic carbon and minerals is not only capable of producing reduced organic carbon, but that this organic carbon is strongly isotopically depleted with respect to its inorganic source. Methane with isotopic compositions ranging from -19 to -53% relative to initial isotopic compositions of -4.1% for the carbon source, was described by Horita and Berndt (1999), and δ¹³C values ranging from -40 to -50% for hydrocarbons produced via the Fischer-Tropsch synthesis, corresponding to fractionations of 38%, were reported by Taran et al. (2007). McCollom and Seewald (2006) also documented the production of strongly depleted hydrocarbons (δ¹³C of -44 to -50%), corresponding to similar fractionations of 31-36%, during abiogenic reactions under hydrothermal conditions (250ºC, 325 bar) (Fig. 2). Comparison of these ranges with those observed in the geologic record demonstrates that isotopic depletion of ¹³C in organic matter cannot be considered an unambiguous biosignature. This point is emphasized by recent evidence of abiogenic hydrocarbons in natural systems, based on unique ¹³C and ²H signatures in hydrocarbons from Kidd Creek mine in Northern Ontario (Sherwood Lollar et al. 2002). Subsequently, evidence of abiogenic hydrocarbons has also been found in the Lost City hydrothermal system near the Mid Atlantic Ridge (Proskurowski et al. 2008).

The observation of abiogenic production of isotopically depleted hydrocarbons, both in laboratory studies and in the field, raises significant questions regarding the conclusion of Schidlowski (2001) and others that isotopically depleted organic carbon observed in ancient rocks is a biosignature. Since the presence of hydrocarbons is a necessary precondition for life, and since hydrocarbons formed by abiogenic processes may not be isotopically distinguishable from biogenic hydrocarbons, it is extremely difficult to assess whether the isotopic depletions observed in ancient rocks are indeed a biosignature that records the origins of life on Earth. It follows that observation of isotopically depleted organic matter on other planets or solar system bodies can not be unambiguously interpreted as a biosignature. To overcome this limitation, we need to find multiple converging lines of evidence that demonstrate the presence and activities of life.

Another example of how our understanding is changing with the acquisition of new information has to do with oxygenation of the atmosphere. The rise of atmospheric oxygen, often referred to as the Great Oxidation Event (GOE; Canfield 1998), is one of the most significant events in the history of life on Earth; this event represents one of the greatest impacts that life has had on the global environment, and on all subsequent life. In this instance, microbial life drastically affected its environment and changed its properties, making the Earth’s atmosphere an ‘extreme’ environment to the anaerobic organisms then living under the anoxic conditions that prevailed. However, it also created the dis-equilibrium between reduced carbon and free oxygen that provided the thermodynamic energy basis for aerobic life on Earth, and is arguably a requirement for the development of terrestrial macrobiota. This disequilibrium is accepted as a biosignature of photo-synthetic life on Earth; hence, spectroscopic evidence of oxygen in the atmospheres of new planets being discovered around distant stars is a potential biosignature. However, the extent to which atmospheric oxygen can be considered a biosignature is defined by our ability to be certain that it could not also be created by abiotic processes. Ongoing research into the oxygenation of Earth’s atmosphere is one approach that contributes to our ability to answer this question.

Because of the crucial role it has played in the development of life on Earth, the occurrence and timing of the increase in concentration of oxygen in the atmosphere is the subject of multiple lines of current research, including new approaches in isotopic analysis. A prime example is the use of Mass-independent Isotopic Fractionation (MIF) of sulphur as a tracer of the sulfur cycle, and, through geo-chemical connection, as a measure of the abundance of oxygen in the atmosphere (Farquhar et al. 2000). MIF occurs by significantly different processes than the mass dependent fractionation described in the previous section. As the name suggests, it is not related to the mass difference between isotopes, but to other atomic properties. The specific mechanisms of MIF are the subject of ongoing research (Theimens 2006). However, using the MIF of sulphur 33 (Δ³³S), Farquhar et al. (2000) have demonstrated significant changes in the Earth’s atmospheric sulphur cycle (Fig. 3), which they have used to argue that a massive increase in atmospheric oxygen occurred at 2450 Ma.

Rouxel et al. (2005) have applied new approaches in iron isotope analysis to the same question and have shown that the range in iron isotopic compositions was drastically reduced within the same timeframe as that proposed by the sulphur MIF signal for the GOE (Fig. 3). The decrease in negative δ⁵⁶Fe signatures observed across the GOE is interpreted by Rouxel et al. (2005) as the result of the oxidation and precipitation of large amounts of iron. The continued presence of positive δ⁵⁶Fe during the GOE is argued to be evidence of ongoing redox stratification in the ocean at this time; i.e. the deep ocean was not fully oxic. The timing of the changes in δ⁵⁶Fe coincides very closely with the changes in Δ33S, indicating that these two independent isotope systems are recording evidence of the same event.

The timeframe for these isotopic excursions is concurrent with the formation of large deposits known as banded iron formations. Banded iron formations are alternating layers of iron oxide-rich and silicate-rich minerals that range in thickness from meter to sub-millimeter scales and range in age from ca. 2.7 to 1.9 Ga. They have long been related to the oxygenation of the Earth’s atmosphere, but recent research is investigating new roles that microbial activity likely played at several steps in the formation of these structures. Konhauser et al. (2005) developed a model indicating that heterotrophic (i.e. requiring organic compounds of C and N for nourishment) microbial metabolism was an important component in their genesis, whereas Johnson et al. (2008) applied iron isotope systematics to suggest a specific role for microbial iron reduction. Although our knowledge of the complex processes that led to the production of banded iron formations is by no means complete, by bringing together multiple lines of evidence from multiple biosignatures, we are getting much closer to a conceptual model of what occurred at this crucial time in Earth history.

Microbial impacts on the environment can also occur at much smaller scales than the oxidation of the atmosphere, and can be far more difficult to recognize. However, these smaller-scale effects have the potential to engender new biosignatures that may be preserved over long geologic timescales. For example, Furnes et al. (2004) reported the occurrence of (microbial) microtubules in volcanic glasses from the Barberton Greenstone Belt in South Africa and in samples from modern oceanic crust, that have been interpreted as a new biosignature (Fig. 4). Although modern examples of surface etching by microorganisms can be observed in the laboratory, the timescales required to form such extended tubules are too great to allow experimental testing. Without the ability to directly demonstrate microbial creation of these tubules, researchers are bringing together multiple lines of evidence, including element mapping, carbon isotopes, scanning transmission x-ray microscopy (STXM), transmission electron microscopy (TEM), molecular genetic techniques, and dating methods (Banerjee et al. 2006, 2007; Benzerara et al. 2007). These techniques are being applied to samples of pillow lava from South Africa, Australia and the Abitibi Greenstone Belt in Canada, to determine their origin and whether the tubules constitute a biosignature.

To provide the foundation needed to interpret biosignatures in the geologic record, research characterizing modern biosignatures is required. This research must address a wide range of questions/conditions, since much of life’s diversity arises from the ability of microbial life to evolve metabolic capabilities that allow it to inhabit almost every known environment on Earth. Microbes have been found to thrive in extreme ranges of temperature, salinity, pH, radiation, pressure, water activity and oxygen availability. Furthermore, they are extremely metabolically diverse and can obtain energy from a wide array of sources, both autotrophically and heterotrophically. If research succeeds in characterizing the potential biosignatures that may be associated with this wide range of adaptability and metabolic flexibility, the information can then be used as a modern template by which either ancient terrestrial or extraterrestrial biosignatures are interpreted.

The range of natural environments within Canada provides important opportunities to investigate the formation of biosignatures, the mechanisms by which life creates them, and fundamental aspects of the limits of life. This environmental diversity is used to advantage by the Canadian Space Agency’s ‘Canadian Analogue Research Network’ (CARN), which supports a wide range of research at sites across Canada, including the three main sites at the McGill Arctic Research Station on Axel Heiberg Island, the Houghton Mars Project on Devon Island, and the Pavilion Lake Research Project in British Columbia. Research conducted at these sites focuses on how life survives in extreme environments and the signatures it leaves behind. Documentation of microbial biosignatures generated at these sites will increase our knowledge of the capabilities of life and allow us to interpret evidence of life not only from Earth’s geologic record, but also in samples from Mars and elsewhere.