Raman spectroscopic analysis of cyanobacterial ... - Semantic Scholar
PAPER
www.rsc.org/analyst | The Analyst
Raman spectroscopic analysis of cyanobacterial gypsum halotrophs and relevance for sulfate deposits on Mars Howell G. M. Edwards,*a Susana E. Jorge Villar,ab John Parnell,c Charles S. Cockelld and Pascal Leee Received 9th March 2005, Accepted 6th April 2005 First published as an Advance Article on the web 25th April 2005 DOI: 10.1039/b503533c The Raman spectra of cyanobacterial species, Gloecapsa and Nostoc, in clear gypsum crystals from the Haughton Crater, Devon Island, Canadian High Arctic, site of a meteorite impact during the Miocene some 23 Mya, have been recorded using several visible and near-infrared excitation wavelengths. The best spectra were obtained using a green wavelength at 514.5 nm and a confocal microscope with an image footprint of about 2 m in diameter and 2 m theoretical depth. Raman biosignatures for beta-carotene and scytonemin were obtained for one type of colony and parietin and beta-carotene for another; chlorophyll was detected in both types of colony. The different combination of these radiation protectant biomolecules suggests that the two cyanobacterial colonies, namely Nostoc and Gloecapsa, are adopting different survival strategies in the system. Confocal spectroscopic probing of the gypsum crystals exhibited sufficient discrimination for the identification of the biomolecules through the gypsum crystal, in simulation of the detection of extant or extinct halotrophs. This supports the viability of Raman spectroscopic techniques for incorporation as part of the instrumentation suite of a robotic lander for planetary surface exploration for the detection of organisms inside sulfate crystals from previous hydrothermal activity on Mars.
Introduction It is recognised that prokaryotic microorganisms exposed to the evaporation of salt waters in the early geological history of the Earth may have developed responses to combat increased ionic strength in the aqueous solutions, which often resulted in final entrapment in crystalline deposits of mineral salts.1 Thus, in response to the applied stresses of UV-radiation, desiccation and saline osmosis these organisms produce a suite of biochemicals including: N Low-wavelength protective pigments, such as parietin, scytonemin and beta-carotene, which can have dualistic support functions in filtering out the harmful UVB and UVC wavelengths whilst still permitting the transmission of photosynthetically active radiation (PAR) wavelengths for photosynthesis and in cell DNA repair mechanisms.2–4 N Compatible solutes5 such as sugars and polyols (e.g., trehalose) which maintain an osmotic equilibrium across the cell membranes, stabilize proteins and protect against desiccation and freeze–thaw cyclical stresses;6,7 the accumulation of trehalose in cyanobacterial colonies in hot desert environments has been directly attributed to an anti-desiccative role.8 Halophilic and halotrophic microorganisms are found terrestrially in thalassohaline and athalassohaline environments; the former derive from sea water evaporation and sodium chloride is the major component, whereas the latter contain high proportions of Group II cations, such as magnesium and calcium. For example, the Dead Sea has Mg2+as the main cation instead of Na+. Increased concentrations of carbonate ion leads to the formation of soda lakes; these are low in Mg2+ and Ca2+ content where the pH is .10, such as that found in the Wadi Natrun in Egypt which was This journal is ß The Royal Society of Chemistry 2005
used in antiquity as a source of natron (a natural mixture of sodium bicarbonate and carbonate) for the mummification of human remains. The salt concentrations in aqueous lake systems1 can range between 0.8 and 21%, but there are also crystalline salt deposits which can retain viable halophilic or halotrophic organisms. This scenario is very important for current studies of possible exobiology in the solar system. It has been estimated9 that 1.3 M km3 of salt minerals were deposited in this way during the late Permian and early Triassic periods alone, ca. 245–280 Mya. Some of these salt deposits up to 1200 m in thickness are to be found in Siberia, Canada and Northern Europe. Plant spores from extinct species have been isolated from halite (rock salt) deposits10 dating from the Permian period and a similar study of anhydrite and gypsum sulfate deposits of marine origin yielded a Triassic age for these isolates.11 Specimens of gypsum collected at Eilat, Israel, by Oren12 demonstrate bands of active halotrophs with green and reddish-pink pigmentation; in another project, we are studying the key Raman spectral biomolecular markers for the protective chemicals produced in this system which will form the basis of a future report.13 Current interest in the remote exploration of Martian sites of possible exobiological significance is convolved with the realisation and significance that Mars and Earth had similar geological histories and that in early epochs Mars was a warmer and wetter planet than it is now.14–17 If evolutionary processes on Mars followed a similar route to that on Earth, it is reasonable to propose that halophilic or halotrophic species appeared early in Martian geological history18 and there is much recent high-resolution photographic evidence from the Analyst, 2005, 130, 917–923 | 917
NASA Mars Orbiter cameras of features on the Martian surface that support the conjecture for lacustrine water on early Mars. Very recently, the ESA Mars Express orbiter has identified the presence of water–ice in the near subsurface of Mars. Indeed, it is now considered a strong possibility that liquid water still exists in subsurface Mars.19 Also, traces of halite have been detected in the SNC Martian meteorites20,21 and in the Monahans meteorite, whose age is estimated at 4.7 Gya; the liquid inclusions in the halite crystals of the latter contain NaCl, KCl and water. It is acknowledged that the Martian surface is sulfate-dominated rather than chloridedominated as on Earth, but surface-dwelling organisms would nevertheless still require a significant tolerance of high ionic strength. Recent studies22 of microbial colonisation in impactgenerated hydrothermal crystalline gypsum deposits (Fig. 1) in the Haughton Crater, Devon Island, Canadian High Arctic
have demonstrated the presence of cyanobacteria in endolithic habitats up to 50 mm from the crystal margins. The crystalline gypsum was found to exist in the clear selenite form, which facilitated the penetration of PAR radiation to the cyanobacterial colonies and assisted their visual observation within the crystals. The organisms were identified as Gloecapsa alpine (Nageli) Brand and Nostoc commune Vaucher; the dark pigmentation of the cyanobacteria implied that the surrounding crystal matrix did not remove sufficient UVB and UVC radiation for complete radiation protection and that the synthesis of photoprotective pigments was therefore still necessary. The gypsum colonisation in the Haughton Crater has a particular astrobiological relevance with the recent discoveries by the NASA Mars rovers Spirit and Opportunity of sulfate minerals on Mars;23 the presence of sulfate phases in the Martian SNC meteorites also suggests the presence of evaporites at the Martian surface.24 It is interesting to speculate that the colonisation of gypsum deposits on Mars could be a geological niche of microbial activity from periods when there was significant moisture at the Martian surface22 and could therefore provide suitable targets for the search for biomarkers using remote planetary analytical instrumentation. Miniaturised Raman spectrometers are now being considered for future robotic exploration missions to Mars as part of the analytical instrumentation suite on planetary landers and there is much focussed activity in the compilation of Raman spectral databases for terrestrial analogues of possible Martian habitats for extinct or extant exobiological niches.25–27 It is in this context that the present study has been undertaken; we wish to address and verify the experimental realisation of the Raman spectroscopic detection of halotrophic biological colonies in a crystalline gypsum matrix.
Experimental Samples Specimens of colonised selenite, a clear crystalline gypsum, from the Haughton Crater containing cyanobacterial colonies were presented for Raman spectroscopic analysis. These consisted of clear gypsum plates of thickness about 5–7 mm which contained several dark green/black colonies of microorganisms, which have been identified as Gloecapsa and Nostoc. Raman spectroscopy
Fig. 1 (a) Field exposure of a coalesced mass of clear gypsum crystals within melt breccia deposits. Reproduced with permission from Parnell et al., ref. 22, CUP, Cambridge. (b) Map of the Haughton Impact Structure, Devon Island, Canadian High Arctic, showing the location of gypsum samples exhibiting microbial communities—the insert shows the position of the site on Devon Island.
918 | Analyst, 2005, 130, 917–923
Raman spectra were obtained in the near infrared at 1064 nm using a Bruker IFS 66/FRA 106 spectrometer and a Nd3+/ YAG laser with a Raman microscope, 406 lens objective giving a specimen ‘‘footprint’’ of about 20 m and a laser power of 20 mW. Spectra were recorded at 4 cm21 spectral resolution with 4000 scans accumulated over a wavenumber range of 50– 3500 cm21 to improve the signal-to–noise ratio. This microscope did not operate in a confocal mode. A Renishaw InVia confocal Raman microscope with a multiwavelength facility operating at 514.5, 632.8 and 785 nm was also used to assess the effect of excitation wavelength on the recording of spectra from the cyanobacteria/gypsum substrate This journal is ß The Royal Society of Chemistry 2005
in the biogeological system. Using a 506 microscope objective, a specimen ‘‘footprint’’ of about 2 m diameter could be achieved with a theoretical confocal depth of about 2 m. Sixty spectral accumulations were undertaken, each of which requiring about 10 s exposure time, to provide the requisite signal-to-noise ratio for improved spectral quality. The use of the two types of Raman spectrometer in this study enabled the importance of the confocal sampling arrangement in the extraction of spectral data from a complex system to be assessed and the spectral data acquisition to be critically examined over the range of visible to near infrared excitation wavelengths. This could be of fundamental interest to the designers of miniaturised Raman spectrometers for space mission applications where spectral quality versus speed of data acquisition and robustness need to be considered.
Results and discussion Several spectra were obtained with the FT-Raman spectrometer of the regions colonised by the cyanobacterial microorganisms. A band at 1008 cm21, the symmetric sulfate stretching mode from gypsum and a broad, weak band centred at 1330 cm21 assigned to chlorophyll appear in all the spectra. No other features were observed from the microorganisms in the Raman spectra of the specimens using 1064 nm excitation. In the spectra obtained with 785 nm excitation from the regions in the specimen colonised by the cyanobacteria, bands at 1142, 1008, 618, 493, 414, 211 and 181 cm21 appear, all
assignable to the gypsum matrix. A broad band at 1330 cm21 also appears, which is absent from the pure gypsum itself, and this can be attributed to the chlorophyll in the microorganism. Similar results were achieved with the red 633 nm excitation wavelength; no other organic features were identified using these laser wavelengths. The best results from the specimens were obtained with the green 514.5 nm wavelength of excitation; the gypsum spectrum was free from background interference and the characteristic bands at 1139, 1008, 619, 492, 413, 209, 179 and 131 cm21, all assignable to the gypsum matrix, were evident. Several spectra were obtained (three replicates for each colony studied) from different regions of the colonised gypsum matrix; in each case no bands from the host gypsum were identified in the spectra, which confirmed that only the biological components in these specific areas were being sampled spectroscopically. Clearly observable spectral biomarker bands for beta-carotene were located at 1517, 1157 and 1006 cm21; the importance in the spectral discrimination between the organic biomolecules and the inorganic host matrix for this specimen is particularly relevant here since two of the three characteristic biomarker bands for the beta-carotene would be obscured by strong gypsum bands due to sulfate stretching modes at 1140 and 1008 cm21. The spectra (Fig. 2) from these cyanobacterial colonies also exhibit bands at 1630, 1598, 1552, 1454, 1379 and 1281 cm21 which are assignable to scytonemin, an important UVradiation screening protective pigment characteristic of
Fig. 2 Raman spectrum of green cyanobacterial colony (Type I) on gypsum surface; 514.5 nm excitation, wavenumber range 100–2000 cm21, 506 objective lens, 60 spectral scans accumulated, each of 10 s duration. The major spectral features here are biosignatures of scytonemin, a cyanobacterial radiation protection compound, and carotenoids.
This journal is ß The Royal Society of Chemistry 2005
Analyst, 2005, 130, 917–923 | 919
cyanobacteria. This was first identified spectroscopically in cyanobacterial sheaths of Nostoc commune from exposed Antarctic beaches near Mars Oasis. A comparison of the Raman spectrum of scytonemin extracted from these Antarctic specimens28 with that obtained in the current investigation reveals that there are some differences in the relative intensities of some of the Raman bands which can be attributed to the different laser wavelength of excitation used here, namely 514.5 nm compared with 1064 nm used in the earlier study. However, for both excitation wavelengths the band wavenumbers are identical for the beta-carotene and scytonemin biomolecules, and this is highly significant for the accurate assignment of the bands and for their attribution to biomolecular species, i.e. the key biomarkers for scytonemin in our database are all found to be present in several of the colonies being studied here. In the examination of the Raman bands obtained from the colonies, it was clear that in fact two different biological organisms could be identified; some spectra (Fig. 3) gave Raman bands at 1671, 1575 (with a shoulder at 1595 cm21), 1197, 914 and 464 cm21, which are characteristic biosignatures of parietin, an accessory radiation-protective biological pigment. In several species of epilithic Antarctic lichens we have noted the presence of parietin along with beta-carotene, the former is believed to be acting as a UV-radiation protectant and the latter is an oxygen free-radical quencher. The band at 464 cm21 is also a characteristic signature of alpha-quartz , present in many geological specimens and as a
constituent of windborne dust on rock surfaces; however, this band only appears in the present analysis in regions associated with parietin only and not elsewhere. Hence, we can attribute this band to the organic protectant biomolecule in our system. A weak feature is found at 1086 cm21 in only one spectrum from the colonies; this is a characteristic feature of calcium carbonate and must be attributed to a deposit of this compound on the gypsum plate. An interesting observation relates to the appearance of a broad and strong band centred about 1340 cm21, with shoulders at 1292 and 1379 cm21 associated with the presence of parietin, assignable to chlorophyll which is clearly present in a significantly larger amount associated with the parietin than it is with scytonemin, where it occurs with a weaker intensity but associated with the carotene. The Raman spectra indicate the presence of not one but two different organisms in the gypsum specimen; one has significant scytonemin and carotene components and a minor amount of chlorophyll, whereas the other organism has a significant quantity of parietin and chlorophyll but only very little carotene. Clearly, the two organisms are using different survival strategies in their colonisation and adaptation of the sulfate host. Spectroscopic study of the bioorganisms in depth profile The objective of this part of the study was the examination of the feasibility of recording the Raman spectra of a halotrophic community by non-destructive analysis through
Fig. 3 Raman spectrum of green cyanobacterial colony (Type II) on gypsum surface; conditions as for Fig. 2. The major spectral features here are biosignatures for parietin, beta-carotene and chlorophyll and the absence of scytonemin should be noted (cf. the spectrum in Fig. 3).
920 | Analyst, 2005, 130, 917–923
This journal is ß The Royal Society of Chemistry 2005
the specimen; hence, this experiment simulates the conditions required for the observation of biosignatures from extremophiles situated inside UV-radiation transparent crystals such as those which may be found on planetary surfaces or in the subsurface. Again, several options of laser excitation wavelength were considered and preliminary trials indicated that the best results were achieved using the green 514.5 nm laser line. The first experiment was carried out through an inverted gypsum crystal, effectively probing the biological spectral signatures as a function of depth. The results are presented in the form of a spectral stackplot in Fig. 4; here, the spectra have been recorded at the surface and at several depths below the surface. For this procedure the confocal arrangement is essential, otherwise the depth of field in conventional microscopy is not sufficiently small to ensure the collection of the spectroscopic data is accomplished from a spectroscopic footprint of theoretically only a few cubic microns. Fig. 4a demonstrates that at the surface only the spectrum of gypsum is seen, similarly at sampling positions into the crystal, one of which is exemplified by Fig. 4b. In Fig. 4c, however, we now observe the signatures of the organisms and the gypsum host together, probed through some 3 mm of crystal. A comparison of the spectrum in Fig. 4d of the organism ‘‘inside’’ the gypsum crystal, where the gypsum spectrum has been subtracted, with that obtained directly from the organism at the crystal surface
shows the same band wavenumbers and confirms that the same biological colony is being sampled. A second experiment was carried out to assess the effect of the laser excitation wavelength on the Raman spectral data obtained from cyanobacterial colonies inside the gypsum crystal. 785 nm excitation gave the Raman stackplot in Fig. 5; the spectrum in Fig. 5a shows the bands from gypsum only at the crystal surface, whereas that in Fig. 5b shows the spectrum characteristic of the biological colony—the wavenumbers of the latter are coincident with those obtained in the spectrum shown in Fig. 4 with 514.5 nm excitation. A weak gypsum signal is seen in Fig. 5b which reflects a poorer confocal sampling arrangement and a lower discrimination against the inorganic matrix. Finally, in Fig. 6a is shown the Raman spectrum of a cyanobacterial colonised gypsum crystal recorded with 514.5 nm excitation; this spectrum gives only a weak biosignal response from the organisms and very strong gypsum host bands from the region just outside the colony, which is the same as that shown in Fig. 3. The biomolecular Raman signals from the organisms in Fig. 6b are different to those observed in Fig. 5; this procedure demonstrates the viability of the analysis for the extremophile/host matrix system and could find application in the robotic measurements of specimens on planetary surfaces using remote sensing Raman instrumentation.
Fig. 4 Confocal Raman spectroscopic depth study of cyanobacterial colonies, halotrophic extremophiles in a gypsum host matrix; a, spectrum at the gypsum surface; b, spectrum at a position in the crystal about 1 mm below the surface; c, spectrum near the interface between the biological colony and the gypsum host; d, spectrum of the cyanobacterial colony about 3 mm below the surface of the gypsum host crystal. Excitation wavelength 514.5 nm, wavenumber range 100–2000 cm21.
This journal is ß The Royal Society of Chemistry 2005
Analyst, 2005, 130, 917–923 | 921
Fig. 5 Raman spectrum of gypsum crystal and cyanobacterial halotrophic colonies; 785 nm excitation, wavenumber range 100–2000 cm21; a, spectrum obtained at the gypsum surface; b, spectrum obtained at the cyanobacterial stratum, showing weaker gypsum peaks.
It is appreciated that the results obtained in these experiments are from living colonies and it would be highly relevant to extend the investigation into the feasibility of detecting the surviving biomarkers from extinct halophilic or halotrophic colonies and from fossilised cyanobacteria. This would also be of relevance for proposed exobiological analytical experiments and could target suitable terrestrial analogues in the geological record.
Conclusions The viability of Raman spectroscopy for the detection of molecular biosignatures from living halotrophic extremophiles in a gypsum host matrix has been demonstrated. Two types of cyanobacteria in the gypsum matrix have been identified from their Raman spectra, namely Nostoc and Gloecapsa, which confirm also that there are different survival strategies being adopted in terms of the biomolecules being synthesised for radiation protection; one type of cyanobacteria uses scytonemin and carotene for this purpose whereas the other uses mainly parietin and chlorophyll. Chlorophyll was detected in both, which indicates that the PAR is being utilised for photosynthesis. From the Raman spectroscopic experiments undertaken in this current study we can conclude that: N The most favourable wavelength of excitation for the observation of the Raman spectra of biosignatures from halotrophic extremophiles is 514.5 nm. This is quite a surprising result since previous spectroscopic studies of Antarctic epilithic and endolithic extremophiles have generally shown a 922 | Analyst, 2005, 130, 917–923
distinct and variable response to visible excitation for Raman spectroscopy, and the generation of fluorescence emission in this region of the electromagnetic spectrum has swamped the weaker Raman signals. Also, epifluorescence spectroscopy is a standard technique for the microbiological identification and location of cyanobacteria in geological environments and this relies on the fluorescence emission from the cyanobacteria at 532 nm, which is very close to the favoured Raman excitation wavelength found here at 514.5 nm. N The detection of the biomarkers in the cyanobacteria inside a gypsum crystalline host matrix closely matched the spectral quality obtained from the biological molecules in the surface colonies that have been studied previously. In this respect, the requirement for a confocal microscope is paramount and this has been demonstrated in the spectral discrimination between the relatively weaker biomolecular signals and the strong bands from the host matrix. N It was possible to identify the key biomolecular signatures of the individual radiation protectants from the Raman spectra of the cyanobacteria. Even in the case where the spectra obtained from the cyanobacteria were significantly less intense than that from the gypsum it was still possible to use standard spectral subtraction techniques to enhance the bioorganic signals, without compromising the spectral quality for subsequent interpretation purposes. This will have relevance to the in-field data acquisition and processing of Raman spectra from remote-sensing instrumentation initially in terrestrial extreme habitats and eventually from planetary surface landers. This journal is ß The Royal Society of Chemistry 2005
Fig. 6 Raman spectrum of cyanobacterial colony in gypsum crystal; 514.5 nm excitation, wavenumber range 100–2000 cm21; a, specimen sampled near the interface between the cyanobacterial colony and the gypsum host some 3 mm inside the crystal, showing the major bands of gypsum and only very weak signals from the organism, this is easily matched to that of the Type II colony shown in Fig. 4. Although the bioorganic spectrum, b, has increased noise levels compared with the directly recorded spectrum, this would nevertheless be identifiable with its biomolecular composition and demonstrates the viability and necessity for the application of data-processing spectral subtraction techniques for the detection of biomolecular signatures from such a system as that studied here, especially when using in-field remote sensing robotic instrumentation. Howell G. M. Edwards,*a Susana E. Jorge Villar,ab John Parnell,c Charles S. Cockelld and Pascal Leee a Chemical and Forensic Sciences, School of Pharmacy, University of Bradford, Bradford, UK BD7 1DP b Area Geodinamica Interna, Facultad de Humanidades y Educacio´n, University of Burgos, Calle Villadiego s/n, 09001 Burgos, Spain c Geofluids Research Group, Department of Geology and Petroleum Geology, University of Aberdeen, Meston Building, King’s College, Aberdeen, UK AB24 3UE d Planetary and Space Sciences Research Institute, Open University, Milton Keynes, UK MK7 6AA e Mars Institute, SETI Institute and NASA Ames Research Center, Moffett Field, CA 94035-1000, California, USA. E-mail: [email protected]; Fax: 00-44-1274-235350; Tel: 00-44-1274-233787
References 1 H. J. Kunte, H. G. Truper and H. Stan-Lotter, Astrobiology: the Search for Life in the Solar System, ed. G. Horneck and C. Baumstarck-Khan, Springer-Verlag, Berlin, 2002, ch. 12, p. 189. 2 D. D. Wynn-Williams, J. M. Holder and H. G. M. Edwards, Bibl. Lichenol., 2000, 75, 275. 3 C. S. Cockell and J. R. Knowland, Biol. Rev., 1999, 74, p. 311. 4 D. D. Wynn-Williams, in Ultraviolet Radiation in Antarctica: Measurements and Biological Effects, ed. C. S. Weiler and P. A. Penhale, American Geophysical Union, Washington, DC, 1994, p. 243. 5 A. D. Brown, Bacteriol. Revs., 1976, 40, 803. 6 K. Lippert and E. A. Galinski, Appl. Microbiol. Biotech., 1992, 37, 61. 7 G. Malin and A. Lapidot, J. Bacteriol., 1996, 178, 385. 8 N. Hershkovitz, A. Oren and Y. Cohen, Appl. Environ. Microbiol., 1991, 57, 645.
This journal is ß The Royal Society of Chemistry 2005
9 M. A. Zharkov and A. L. I. Anshin, History of Palaeozoic Salt Accumulation, Springer Verlag, Berlin, 1981. 10 W. Klaus, Z. Deutsch. Geol. Ges., 1955, 105, 756. 11 W. T. Holser and I. R. Kaplan, Chem. Geol., 1966, 1, 93. 12 A. Oren, Geomicrobiol. J., 1997, 14, 231. 13 H. G. M. Edwards, E. M. Newton, A. Oren and D. D. WynnWilliams, to be published. 14 M. Carr, Water on Mars, Oxford University Press, Oxford, 1996. 15 S. M. Clifford, J. Geophys. Res.-Planets, 1993, 98, 10973. 16 C. P. McKay, Origins Life Evol. Biosphere, 1997, 27, 263. 17 C. P. McKay, E. I. Friedmann, R. A. Wharton and W. L. Davies, Adv. Space Res., 1992, 12, 231. 18 J. K. Frederickson, D. P. Chandler and T. C. Onstott, SPIE Proc., 1997, 3111, 318. 19 M. C. Malin and K. S. Edgett, Science, 2000, 288, 2330. 20 J. L. Gooding, Icarus, 1992, 99, 28. 21 D. J. Sawyer, M. D. McGehee, J. Canepa and C. B. Moore, Meteoritics Planetary Sci., 2000, 35, 743. 22 J. Parnell, P. Lee, C. S. Cockell and G. R. Osinski, Int. J. Astrobiol., 2004, 3, 247. 23 C. D. Cooper and J. F. Mustard, Icarus, 2002, 158, 42. 24 J. C. Bridges and M. M. Grady, Earth Planetary Sci. Lett., 2000, 176, 267. 25 H. G. M. Edwards and E. M. Newton, in The Search for Life on Mars, ed. J. A. Hiscox, British Interplanetary Society, London, 1999, pp. 78–85. 26 H. G. M. Edwards, E. M. Newton, D. D. Wynn-Williams, D. Dickensheets, C. Schoen and C. Crowder, Int. J. Astrobiol., 2003, 1, 333. 27 A. Wang, L. A. Haskin and E. Cortez, Appl. Spectrosc., 1998, 52, 477. 28 D. D. Wynn-Williams, H. G. M. Edwards and F. Garcia-Pichel, Eur. J. Phycol., 1999, 34, 381.
Analyst, 2005, 130, 917–923 | 923
www.rsc.org/analyst | The Analyst
Raman spectroscopic analysis of cyanobacterial gypsum halotrophs and relevance for sulfate deposits on Mars Howell G. M. Edwards,*a Susana E. Jorge Villar,ab John Parnell,c Charles S. Cockelld and Pascal Leee Received 9th March 2005, Accepted 6th April 2005 First published as an Advance Article on the web 25th April 2005 DOI: 10.1039/b503533c The Raman spectra of cyanobacterial species, Gloecapsa and Nostoc, in clear gypsum crystals from the Haughton Crater, Devon Island, Canadian High Arctic, site of a meteorite impact during the Miocene some 23 Mya, have been recorded using several visible and near-infrared excitation wavelengths. The best spectra were obtained using a green wavelength at 514.5 nm and a confocal microscope with an image footprint of about 2 m in diameter and 2 m theoretical depth. Raman biosignatures for beta-carotene and scytonemin were obtained for one type of colony and parietin and beta-carotene for another; chlorophyll was detected in both types of colony. The different combination of these radiation protectant biomolecules suggests that the two cyanobacterial colonies, namely Nostoc and Gloecapsa, are adopting different survival strategies in the system. Confocal spectroscopic probing of the gypsum crystals exhibited sufficient discrimination for the identification of the biomolecules through the gypsum crystal, in simulation of the detection of extant or extinct halotrophs. This supports the viability of Raman spectroscopic techniques for incorporation as part of the instrumentation suite of a robotic lander for planetary surface exploration for the detection of organisms inside sulfate crystals from previous hydrothermal activity on Mars.
Introduction It is recognised that prokaryotic microorganisms exposed to the evaporation of salt waters in the early geological history of the Earth may have developed responses to combat increased ionic strength in the aqueous solutions, which often resulted in final entrapment in crystalline deposits of mineral salts.1 Thus, in response to the applied stresses of UV-radiation, desiccation and saline osmosis these organisms produce a suite of biochemicals including: N Low-wavelength protective pigments, such as parietin, scytonemin and beta-carotene, which can have dualistic support functions in filtering out the harmful UVB and UVC wavelengths whilst still permitting the transmission of photosynthetically active radiation (PAR) wavelengths for photosynthesis and in cell DNA repair mechanisms.2–4 N Compatible solutes5 such as sugars and polyols (e.g., trehalose) which maintain an osmotic equilibrium across the cell membranes, stabilize proteins and protect against desiccation and freeze–thaw cyclical stresses;6,7 the accumulation of trehalose in cyanobacterial colonies in hot desert environments has been directly attributed to an anti-desiccative role.8 Halophilic and halotrophic microorganisms are found terrestrially in thalassohaline and athalassohaline environments; the former derive from sea water evaporation and sodium chloride is the major component, whereas the latter contain high proportions of Group II cations, such as magnesium and calcium. For example, the Dead Sea has Mg2+as the main cation instead of Na+. Increased concentrations of carbonate ion leads to the formation of soda lakes; these are low in Mg2+ and Ca2+ content where the pH is .10, such as that found in the Wadi Natrun in Egypt which was This journal is ß The Royal Society of Chemistry 2005
used in antiquity as a source of natron (a natural mixture of sodium bicarbonate and carbonate) for the mummification of human remains. The salt concentrations in aqueous lake systems1 can range between 0.8 and 21%, but there are also crystalline salt deposits which can retain viable halophilic or halotrophic organisms. This scenario is very important for current studies of possible exobiology in the solar system. It has been estimated9 that 1.3 M km3 of salt minerals were deposited in this way during the late Permian and early Triassic periods alone, ca. 245–280 Mya. Some of these salt deposits up to 1200 m in thickness are to be found in Siberia, Canada and Northern Europe. Plant spores from extinct species have been isolated from halite (rock salt) deposits10 dating from the Permian period and a similar study of anhydrite and gypsum sulfate deposits of marine origin yielded a Triassic age for these isolates.11 Specimens of gypsum collected at Eilat, Israel, by Oren12 demonstrate bands of active halotrophs with green and reddish-pink pigmentation; in another project, we are studying the key Raman spectral biomolecular markers for the protective chemicals produced in this system which will form the basis of a future report.13 Current interest in the remote exploration of Martian sites of possible exobiological significance is convolved with the realisation and significance that Mars and Earth had similar geological histories and that in early epochs Mars was a warmer and wetter planet than it is now.14–17 If evolutionary processes on Mars followed a similar route to that on Earth, it is reasonable to propose that halophilic or halotrophic species appeared early in Martian geological history18 and there is much recent high-resolution photographic evidence from the Analyst, 2005, 130, 917–923 | 917
NASA Mars Orbiter cameras of features on the Martian surface that support the conjecture for lacustrine water on early Mars. Very recently, the ESA Mars Express orbiter has identified the presence of water–ice in the near subsurface of Mars. Indeed, it is now considered a strong possibility that liquid water still exists in subsurface Mars.19 Also, traces of halite have been detected in the SNC Martian meteorites20,21 and in the Monahans meteorite, whose age is estimated at 4.7 Gya; the liquid inclusions in the halite crystals of the latter contain NaCl, KCl and water. It is acknowledged that the Martian surface is sulfate-dominated rather than chloridedominated as on Earth, but surface-dwelling organisms would nevertheless still require a significant tolerance of high ionic strength. Recent studies22 of microbial colonisation in impactgenerated hydrothermal crystalline gypsum deposits (Fig. 1) in the Haughton Crater, Devon Island, Canadian High Arctic
have demonstrated the presence of cyanobacteria in endolithic habitats up to 50 mm from the crystal margins. The crystalline gypsum was found to exist in the clear selenite form, which facilitated the penetration of PAR radiation to the cyanobacterial colonies and assisted their visual observation within the crystals. The organisms were identified as Gloecapsa alpine (Nageli) Brand and Nostoc commune Vaucher; the dark pigmentation of the cyanobacteria implied that the surrounding crystal matrix did not remove sufficient UVB and UVC radiation for complete radiation protection and that the synthesis of photoprotective pigments was therefore still necessary. The gypsum colonisation in the Haughton Crater has a particular astrobiological relevance with the recent discoveries by the NASA Mars rovers Spirit and Opportunity of sulfate minerals on Mars;23 the presence of sulfate phases in the Martian SNC meteorites also suggests the presence of evaporites at the Martian surface.24 It is interesting to speculate that the colonisation of gypsum deposits on Mars could be a geological niche of microbial activity from periods when there was significant moisture at the Martian surface22 and could therefore provide suitable targets for the search for biomarkers using remote planetary analytical instrumentation. Miniaturised Raman spectrometers are now being considered for future robotic exploration missions to Mars as part of the analytical instrumentation suite on planetary landers and there is much focussed activity in the compilation of Raman spectral databases for terrestrial analogues of possible Martian habitats for extinct or extant exobiological niches.25–27 It is in this context that the present study has been undertaken; we wish to address and verify the experimental realisation of the Raman spectroscopic detection of halotrophic biological colonies in a crystalline gypsum matrix.
Experimental Samples Specimens of colonised selenite, a clear crystalline gypsum, from the Haughton Crater containing cyanobacterial colonies were presented for Raman spectroscopic analysis. These consisted of clear gypsum plates of thickness about 5–7 mm which contained several dark green/black colonies of microorganisms, which have been identified as Gloecapsa and Nostoc. Raman spectroscopy
Fig. 1 (a) Field exposure of a coalesced mass of clear gypsum crystals within melt breccia deposits. Reproduced with permission from Parnell et al., ref. 22, CUP, Cambridge. (b) Map of the Haughton Impact Structure, Devon Island, Canadian High Arctic, showing the location of gypsum samples exhibiting microbial communities—the insert shows the position of the site on Devon Island.
918 | Analyst, 2005, 130, 917–923
Raman spectra were obtained in the near infrared at 1064 nm using a Bruker IFS 66/FRA 106 spectrometer and a Nd3+/ YAG laser with a Raman microscope, 406 lens objective giving a specimen ‘‘footprint’’ of about 20 m and a laser power of 20 mW. Spectra were recorded at 4 cm21 spectral resolution with 4000 scans accumulated over a wavenumber range of 50– 3500 cm21 to improve the signal-to–noise ratio. This microscope did not operate in a confocal mode. A Renishaw InVia confocal Raman microscope with a multiwavelength facility operating at 514.5, 632.8 and 785 nm was also used to assess the effect of excitation wavelength on the recording of spectra from the cyanobacteria/gypsum substrate This journal is ß The Royal Society of Chemistry 2005
in the biogeological system. Using a 506 microscope objective, a specimen ‘‘footprint’’ of about 2 m diameter could be achieved with a theoretical confocal depth of about 2 m. Sixty spectral accumulations were undertaken, each of which requiring about 10 s exposure time, to provide the requisite signal-to-noise ratio for improved spectral quality. The use of the two types of Raman spectrometer in this study enabled the importance of the confocal sampling arrangement in the extraction of spectral data from a complex system to be assessed and the spectral data acquisition to be critically examined over the range of visible to near infrared excitation wavelengths. This could be of fundamental interest to the designers of miniaturised Raman spectrometers for space mission applications where spectral quality versus speed of data acquisition and robustness need to be considered.
Results and discussion Several spectra were obtained with the FT-Raman spectrometer of the regions colonised by the cyanobacterial microorganisms. A band at 1008 cm21, the symmetric sulfate stretching mode from gypsum and a broad, weak band centred at 1330 cm21 assigned to chlorophyll appear in all the spectra. No other features were observed from the microorganisms in the Raman spectra of the specimens using 1064 nm excitation. In the spectra obtained with 785 nm excitation from the regions in the specimen colonised by the cyanobacteria, bands at 1142, 1008, 618, 493, 414, 211 and 181 cm21 appear, all
assignable to the gypsum matrix. A broad band at 1330 cm21 also appears, which is absent from the pure gypsum itself, and this can be attributed to the chlorophyll in the microorganism. Similar results were achieved with the red 633 nm excitation wavelength; no other organic features were identified using these laser wavelengths. The best results from the specimens were obtained with the green 514.5 nm wavelength of excitation; the gypsum spectrum was free from background interference and the characteristic bands at 1139, 1008, 619, 492, 413, 209, 179 and 131 cm21, all assignable to the gypsum matrix, were evident. Several spectra were obtained (three replicates for each colony studied) from different regions of the colonised gypsum matrix; in each case no bands from the host gypsum were identified in the spectra, which confirmed that only the biological components in these specific areas were being sampled spectroscopically. Clearly observable spectral biomarker bands for beta-carotene were located at 1517, 1157 and 1006 cm21; the importance in the spectral discrimination between the organic biomolecules and the inorganic host matrix for this specimen is particularly relevant here since two of the three characteristic biomarker bands for the beta-carotene would be obscured by strong gypsum bands due to sulfate stretching modes at 1140 and 1008 cm21. The spectra (Fig. 2) from these cyanobacterial colonies also exhibit bands at 1630, 1598, 1552, 1454, 1379 and 1281 cm21 which are assignable to scytonemin, an important UVradiation screening protective pigment characteristic of
Fig. 2 Raman spectrum of green cyanobacterial colony (Type I) on gypsum surface; 514.5 nm excitation, wavenumber range 100–2000 cm21, 506 objective lens, 60 spectral scans accumulated, each of 10 s duration. The major spectral features here are biosignatures of scytonemin, a cyanobacterial radiation protection compound, and carotenoids.
This journal is ß The Royal Society of Chemistry 2005
Analyst, 2005, 130, 917–923 | 919
cyanobacteria. This was first identified spectroscopically in cyanobacterial sheaths of Nostoc commune from exposed Antarctic beaches near Mars Oasis. A comparison of the Raman spectrum of scytonemin extracted from these Antarctic specimens28 with that obtained in the current investigation reveals that there are some differences in the relative intensities of some of the Raman bands which can be attributed to the different laser wavelength of excitation used here, namely 514.5 nm compared with 1064 nm used in the earlier study. However, for both excitation wavelengths the band wavenumbers are identical for the beta-carotene and scytonemin biomolecules, and this is highly significant for the accurate assignment of the bands and for their attribution to biomolecular species, i.e. the key biomarkers for scytonemin in our database are all found to be present in several of the colonies being studied here. In the examination of the Raman bands obtained from the colonies, it was clear that in fact two different biological organisms could be identified; some spectra (Fig. 3) gave Raman bands at 1671, 1575 (with a shoulder at 1595 cm21), 1197, 914 and 464 cm21, which are characteristic biosignatures of parietin, an accessory radiation-protective biological pigment. In several species of epilithic Antarctic lichens we have noted the presence of parietin along with beta-carotene, the former is believed to be acting as a UV-radiation protectant and the latter is an oxygen free-radical quencher. The band at 464 cm21 is also a characteristic signature of alpha-quartz , present in many geological specimens and as a
constituent of windborne dust on rock surfaces; however, this band only appears in the present analysis in regions associated with parietin only and not elsewhere. Hence, we can attribute this band to the organic protectant biomolecule in our system. A weak feature is found at 1086 cm21 in only one spectrum from the colonies; this is a characteristic feature of calcium carbonate and must be attributed to a deposit of this compound on the gypsum plate. An interesting observation relates to the appearance of a broad and strong band centred about 1340 cm21, with shoulders at 1292 and 1379 cm21 associated with the presence of parietin, assignable to chlorophyll which is clearly present in a significantly larger amount associated with the parietin than it is with scytonemin, where it occurs with a weaker intensity but associated with the carotene. The Raman spectra indicate the presence of not one but two different organisms in the gypsum specimen; one has significant scytonemin and carotene components and a minor amount of chlorophyll, whereas the other organism has a significant quantity of parietin and chlorophyll but only very little carotene. Clearly, the two organisms are using different survival strategies in their colonisation and adaptation of the sulfate host. Spectroscopic study of the bioorganisms in depth profile The objective of this part of the study was the examination of the feasibility of recording the Raman spectra of a halotrophic community by non-destructive analysis through
Fig. 3 Raman spectrum of green cyanobacterial colony (Type II) on gypsum surface; conditions as for Fig. 2. The major spectral features here are biosignatures for parietin, beta-carotene and chlorophyll and the absence of scytonemin should be noted (cf. the spectrum in Fig. 3).
920 | Analyst, 2005, 130, 917–923
This journal is ß The Royal Society of Chemistry 2005
the specimen; hence, this experiment simulates the conditions required for the observation of biosignatures from extremophiles situated inside UV-radiation transparent crystals such as those which may be found on planetary surfaces or in the subsurface. Again, several options of laser excitation wavelength were considered and preliminary trials indicated that the best results were achieved using the green 514.5 nm laser line. The first experiment was carried out through an inverted gypsum crystal, effectively probing the biological spectral signatures as a function of depth. The results are presented in the form of a spectral stackplot in Fig. 4; here, the spectra have been recorded at the surface and at several depths below the surface. For this procedure the confocal arrangement is essential, otherwise the depth of field in conventional microscopy is not sufficiently small to ensure the collection of the spectroscopic data is accomplished from a spectroscopic footprint of theoretically only a few cubic microns. Fig. 4a demonstrates that at the surface only the spectrum of gypsum is seen, similarly at sampling positions into the crystal, one of which is exemplified by Fig. 4b. In Fig. 4c, however, we now observe the signatures of the organisms and the gypsum host together, probed through some 3 mm of crystal. A comparison of the spectrum in Fig. 4d of the organism ‘‘inside’’ the gypsum crystal, where the gypsum spectrum has been subtracted, with that obtained directly from the organism at the crystal surface
shows the same band wavenumbers and confirms that the same biological colony is being sampled. A second experiment was carried out to assess the effect of the laser excitation wavelength on the Raman spectral data obtained from cyanobacterial colonies inside the gypsum crystal. 785 nm excitation gave the Raman stackplot in Fig. 5; the spectrum in Fig. 5a shows the bands from gypsum only at the crystal surface, whereas that in Fig. 5b shows the spectrum characteristic of the biological colony—the wavenumbers of the latter are coincident with those obtained in the spectrum shown in Fig. 4 with 514.5 nm excitation. A weak gypsum signal is seen in Fig. 5b which reflects a poorer confocal sampling arrangement and a lower discrimination against the inorganic matrix. Finally, in Fig. 6a is shown the Raman spectrum of a cyanobacterial colonised gypsum crystal recorded with 514.5 nm excitation; this spectrum gives only a weak biosignal response from the organisms and very strong gypsum host bands from the region just outside the colony, which is the same as that shown in Fig. 3. The biomolecular Raman signals from the organisms in Fig. 6b are different to those observed in Fig. 5; this procedure demonstrates the viability of the analysis for the extremophile/host matrix system and could find application in the robotic measurements of specimens on planetary surfaces using remote sensing Raman instrumentation.
Fig. 4 Confocal Raman spectroscopic depth study of cyanobacterial colonies, halotrophic extremophiles in a gypsum host matrix; a, spectrum at the gypsum surface; b, spectrum at a position in the crystal about 1 mm below the surface; c, spectrum near the interface between the biological colony and the gypsum host; d, spectrum of the cyanobacterial colony about 3 mm below the surface of the gypsum host crystal. Excitation wavelength 514.5 nm, wavenumber range 100–2000 cm21.
This journal is ß The Royal Society of Chemistry 2005
Analyst, 2005, 130, 917–923 | 921
Fig. 5 Raman spectrum of gypsum crystal and cyanobacterial halotrophic colonies; 785 nm excitation, wavenumber range 100–2000 cm21; a, spectrum obtained at the gypsum surface; b, spectrum obtained at the cyanobacterial stratum, showing weaker gypsum peaks.
It is appreciated that the results obtained in these experiments are from living colonies and it would be highly relevant to extend the investigation into the feasibility of detecting the surviving biomarkers from extinct halophilic or halotrophic colonies and from fossilised cyanobacteria. This would also be of relevance for proposed exobiological analytical experiments and could target suitable terrestrial analogues in the geological record.
Conclusions The viability of Raman spectroscopy for the detection of molecular biosignatures from living halotrophic extremophiles in a gypsum host matrix has been demonstrated. Two types of cyanobacteria in the gypsum matrix have been identified from their Raman spectra, namely Nostoc and Gloecapsa, which confirm also that there are different survival strategies being adopted in terms of the biomolecules being synthesised for radiation protection; one type of cyanobacteria uses scytonemin and carotene for this purpose whereas the other uses mainly parietin and chlorophyll. Chlorophyll was detected in both, which indicates that the PAR is being utilised for photosynthesis. From the Raman spectroscopic experiments undertaken in this current study we can conclude that: N The most favourable wavelength of excitation for the observation of the Raman spectra of biosignatures from halotrophic extremophiles is 514.5 nm. This is quite a surprising result since previous spectroscopic studies of Antarctic epilithic and endolithic extremophiles have generally shown a 922 | Analyst, 2005, 130, 917–923
distinct and variable response to visible excitation for Raman spectroscopy, and the generation of fluorescence emission in this region of the electromagnetic spectrum has swamped the weaker Raman signals. Also, epifluorescence spectroscopy is a standard technique for the microbiological identification and location of cyanobacteria in geological environments and this relies on the fluorescence emission from the cyanobacteria at 532 nm, which is very close to the favoured Raman excitation wavelength found here at 514.5 nm. N The detection of the biomarkers in the cyanobacteria inside a gypsum crystalline host matrix closely matched the spectral quality obtained from the biological molecules in the surface colonies that have been studied previously. In this respect, the requirement for a confocal microscope is paramount and this has been demonstrated in the spectral discrimination between the relatively weaker biomolecular signals and the strong bands from the host matrix. N It was possible to identify the key biomolecular signatures of the individual radiation protectants from the Raman spectra of the cyanobacteria. Even in the case where the spectra obtained from the cyanobacteria were significantly less intense than that from the gypsum it was still possible to use standard spectral subtraction techniques to enhance the bioorganic signals, without compromising the spectral quality for subsequent interpretation purposes. This will have relevance to the in-field data acquisition and processing of Raman spectra from remote-sensing instrumentation initially in terrestrial extreme habitats and eventually from planetary surface landers. This journal is ß The Royal Society of Chemistry 2005
Fig. 6 Raman spectrum of cyanobacterial colony in gypsum crystal; 514.5 nm excitation, wavenumber range 100–2000 cm21; a, specimen sampled near the interface between the cyanobacterial colony and the gypsum host some 3 mm inside the crystal, showing the major bands of gypsum and only very weak signals from the organism, this is easily matched to that of the Type II colony shown in Fig. 4. Although the bioorganic spectrum, b, has increased noise levels compared with the directly recorded spectrum, this would nevertheless be identifiable with its biomolecular composition and demonstrates the viability and necessity for the application of data-processing spectral subtraction techniques for the detection of biomolecular signatures from such a system as that studied here, especially when using in-field remote sensing robotic instrumentation. Howell G. M. Edwards,*a Susana E. Jorge Villar,ab John Parnell,c Charles S. Cockelld and Pascal Leee a Chemical and Forensic Sciences, School of Pharmacy, University of Bradford, Bradford, UK BD7 1DP b Area Geodinamica Interna, Facultad de Humanidades y Educacio´n, University of Burgos, Calle Villadiego s/n, 09001 Burgos, Spain c Geofluids Research Group, Department of Geology and Petroleum Geology, University of Aberdeen, Meston Building, King’s College, Aberdeen, UK AB24 3UE d Planetary and Space Sciences Research Institute, Open University, Milton Keynes, UK MK7 6AA e Mars Institute, SETI Institute and NASA Ames Research Center, Moffett Field, CA 94035-1000, California, USA. E-mail: [email protected]; Fax: 00-44-1274-235350; Tel: 00-44-1274-233787
References 1 H. J. Kunte, H. G. Truper and H. Stan-Lotter, Astrobiology: the Search for Life in the Solar System, ed. G. Horneck and C. Baumstarck-Khan, Springer-Verlag, Berlin, 2002, ch. 12, p. 189. 2 D. D. Wynn-Williams, J. M. Holder and H. G. M. Edwards, Bibl. Lichenol., 2000, 75, 275. 3 C. S. Cockell and J. R. Knowland, Biol. Rev., 1999, 74, p. 311. 4 D. D. Wynn-Williams, in Ultraviolet Radiation in Antarctica: Measurements and Biological Effects, ed. C. S. Weiler and P. A. Penhale, American Geophysical Union, Washington, DC, 1994, p. 243. 5 A. D. Brown, Bacteriol. Revs., 1976, 40, 803. 6 K. Lippert and E. A. Galinski, Appl. Microbiol. Biotech., 1992, 37, 61. 7 G. Malin and A. Lapidot, J. Bacteriol., 1996, 178, 385. 8 N. Hershkovitz, A. Oren and Y. Cohen, Appl. Environ. Microbiol., 1991, 57, 645.
This journal is ß The Royal Society of Chemistry 2005
9 M. A. Zharkov and A. L. I. Anshin, History of Palaeozoic Salt Accumulation, Springer Verlag, Berlin, 1981. 10 W. Klaus, Z. Deutsch. Geol. Ges., 1955, 105, 756. 11 W. T. Holser and I. R. Kaplan, Chem. Geol., 1966, 1, 93. 12 A. Oren, Geomicrobiol. J., 1997, 14, 231. 13 H. G. M. Edwards, E. M. Newton, A. Oren and D. D. WynnWilliams, to be published. 14 M. Carr, Water on Mars, Oxford University Press, Oxford, 1996. 15 S. M. Clifford, J. Geophys. Res.-Planets, 1993, 98, 10973. 16 C. P. McKay, Origins Life Evol. Biosphere, 1997, 27, 263. 17 C. P. McKay, E. I. Friedmann, R. A. Wharton and W. L. Davies, Adv. Space Res., 1992, 12, 231. 18 J. K. Frederickson, D. P. Chandler and T. C. Onstott, SPIE Proc., 1997, 3111, 318. 19 M. C. Malin and K. S. Edgett, Science, 2000, 288, 2330. 20 J. L. Gooding, Icarus, 1992, 99, 28. 21 D. J. Sawyer, M. D. McGehee, J. Canepa and C. B. Moore, Meteoritics Planetary Sci., 2000, 35, 743. 22 J. Parnell, P. Lee, C. S. Cockell and G. R. Osinski, Int. J. Astrobiol., 2004, 3, 247. 23 C. D. Cooper and J. F. Mustard, Icarus, 2002, 158, 42. 24 J. C. Bridges and M. M. Grady, Earth Planetary Sci. Lett., 2000, 176, 267. 25 H. G. M. Edwards and E. M. Newton, in The Search for Life on Mars, ed. J. A. Hiscox, British Interplanetary Society, London, 1999, pp. 78–85. 26 H. G. M. Edwards, E. M. Newton, D. D. Wynn-Williams, D. Dickensheets, C. Schoen and C. Crowder, Int. J. Astrobiol., 2003, 1, 333. 27 A. Wang, L. A. Haskin and E. Cortez, Appl. Spectrosc., 1998, 52, 477. 28 D. D. Wynn-Williams, H. G. M. Edwards and F. Garcia-Pichel, Eur. J. Phycol., 1999, 34, 381.
Analyst, 2005, 130, 917–923 | 923
