Introduction
The identification of the survival strategies adopted by extremophiles in environmentally stressed geological scenarios, which represent terrestrial ‘limits of life’ situations (Wynn-Williams & Edwards Reference Wynn-Williams and Edwards2000), is a key factor in the astrobiological exploration of our Solar System, especially in the search for extinct or extant life on Mars through the remote robotic monitoring of the Martian regolith (Wynn-Williams Reference Wynn-Williams and Hiscox1999; Wynn-Williams et al. Reference Wynn-Williams, Edwards and Newton2000; Villar & Edwards Reference Villar and Edwards2006). The characterization of the protective chemicals that have been synthesized by microbial communities in stressed geological environments in response to low-wavelength, high-energy ultraviolet (UV) radiation insolation, low temperatures, extreme desiccation and hypersalinity is central to the establishment of biomarkers in the geological records that are indicative of relict or living organisms (Edwards et al. Reference Edwards, Holder and Wynn-Williams1998; Cockell & Knowland Reference Cockell and Knowland1999; Villar et al. Reference Villar, Edwards and Worland2005).
It is clear that the production of radiation protectant biomolecules by photosynthetic colonies in extreme environments is a prime requirement (Edwards et al. Reference Edwards, Holder and Wynn-Williams1998; Cockell & Knowland Reference Cockell and Knowland1999) for the survival, adaptation and control of their environments in strategies that encompass, first, the avoidance of the low-wavelength Solar radiation using geological matrices and substrates and also their ability to minimize and repair cell damage (Vincent et al. Reference Vincent, Castenholz, Dournes and Howard-Williams1993; Wynn-Williams et al. Reference Wynn-Williams, Edwards and Garcia-Pichel1999). The necessity of limiting the transmission of radiation in the UVC (200–280 nm), UVB (280–320 nm) and UVA (320–380 nm) regions of the electromagnetic spectrum, whilst at the same time permitting the access of photosynthetically active radiation at about 500 nm to activate the chlorophyll in the organisms, is a vital requirement of these survival strategies.
The ability of photosynthetic cyanobacterial colonies (e.g. stromatolitic, mats, and biofilms) to survive high Solar irradiance at or near to the surface of their geological extremophilic niches is associated with their synthesis of the UV-radiation screening pigment, scytonemin (Fig. 1), whose structure was first elucidated by Vincent et al. (Reference Vincent, Castenholz, Dournes and Howard-Williams1993) and Proteau et al. (Reference Proteau, Gerwick, Garcia-Pichel and Castenholz1993) on the basis of 13C and 1H NMR spectra. The pigment was found to absorb strongly in the UVA, UVB and UVC regions of the electromagnetic spectrum with a λmax value of 370 nm and absorptions at 300 and 250 nm.
Scytonemin structure.
The protection afforded by the presence of scytonemin in cyanobacterial colonies in the geological record is indicated by the results of experiments, which demonstrated that 90% of the incident UVA is prevented from entering the cells (Garcia-Pichel & Castenholz Reference Garcia-Pichel and Castenholz1991; Garcia-Pichel et al. Reference Garcia-Pichel, Sherry and Castenholz1992) and that, in addition, the photobleaching of chlorophyll a was suppressed significantly. Increased exposure to UV-radiation flux was found to stimulate an increased production of scytonemin by the cyanobacteria. Hader et al. (Reference Hader, Kumar, Smith and Worrest2003) and Hansucker et al. (Reference Hansucker, Tissue, Potts and Helm2001) have reported a study of the effects of Solar UV radiation on aquatic ecosystems and have noted the supportive roles of mycosporin, such as amino acids and scytonemin, in protective functions for cyanobacteria.
Scytonemin, uniquely synthesized by cyanobacteria in radiation stressed environments, was first characterized vibrationally in 2000 using Raman spectroscopy (Edwards et al. Reference Edwards, Garcia-Pichel, Newton and Wynn-Williams2000) from the non-destructive analysis of material extracted from cyanobacterial colonies and its presence verified in situ from its key spectral biomarkers in intertidal cyanobacterial mats of Lyngbya cf. aestuarii. At that time, the assignment of the vibrational bands in the Raman spectrum was accomplished by comparison with related model structural components, such as tryptophan and para-hydroxybenzaldehyde: several vibrational mode attributions were necessarily tentative and approximate. The subsequent discovery of scytonemin in cyanobacterial colonies from a wide and diverse range of habitats, including hot desert salterns, desert varnishes, cold desert endoliths and chasmoliths (Edwards & Hargreaves Reference Edwards, Hargreaves, Boeyens and Ogilvie2008), in which the UV-radiation protection pigment occurs along with carotenoids and accessory light-harvesting pigments now necessitate a finer investigation of the assignment of the Raman spectral modes of scytonemin to corroborate the earlier studies. This requirement has recently gained importance (Bultel-Ponce et al. Reference Bultel-Ponce, Felix-Theodose, Sarlhou, Ponge and Bodo2004) with the discovery of three new pigments based on scytonemin, namely, dimethoxyscytonemin, tetramethoxyscytonemin and scytonin, in extracts from Scytonema sp.; although no vibrational spectroscopic work has yet been carried out on these derivatives of scytonemin, a definitive assignment of the molecular modes associated with the key structural components of scytonemin is now clearly necessary as the parent of these molecular protectants. Hence, we report here the results of ab initio calculations of scytonemin, from which our knowledge of the molecular structural behaviour of this important radiation protectant biomolecule will be enhanced.
The survivability of scytonemin residues in the terrestrial geological environment after biological activity has ceased is of especial relevance to the forthcoming adoption of miniaturized Raman spectroscopic instrumentation on a Mars robotic lander as part of the Pasteur package in the European Space Agency AURORA/ExoMars programme. An example of the relevance of this approach is provided by our recent Raman spectroscopic analysis of an ancient stromatolitic formation from early Earth evolution in an aqueous environment by the cyanobacterial colonization of rocks. Figure 2 shows a section of a specimen from the North Pole Dome, Trendall, Pilbara, North West Australia, which dates from 3.45 Gya. The characteristic silicified stromatolitic striations are clearly visible in this specimen as dark bands and the presence of scytonemin in a small niche of about 10 μm diameter is unquestionably indicated from the Raman spectrum (Edwards et al. Reference Edwards, Jorge Villar, Pullan, Hofmann, Hargreaves and Westall2007; Pullan et al. Reference Pullan2008). It is, of course, conjectural as to whether the scytonemin has been present in the rock from its earliest deposition or perhaps results from a later ingressive contamination, but the spectroscopic evidence for its presence in the geological record of this specimen is incontrovertible.
A section of a geological specimen of a stromatolite from the North Pole Dome, Trendall, Pilbara, North West Australia, which dates from 3.45 Gya, in which scytonemin has been identified as residue of cyanobacterial life.
Mars, significantly smaller than Earth with an equatorial radius of 3397 and 6371 km, respectively, and of correspondingly smaller mass, 6.42×1023 kg compared with that of the Earth at 5.97×1024 kg, would have therefore cooled much more rapidly following its planetary formation and it is possible that on a wet Mars life could have started earlier in Martian planetary history. The existence of the cyanobacterial stromatolitic colonies on early Earth, with the scytonemin protection afforded against radiation exposure in the early atmosphere and dating from about 3.5 Gya in our geological record, therefore could imply that such an occurrence could have pertained on Mars. These rocks from the Pilbara Craton have undergone geological modification in the form of greenschist facies metamorphism and would not be expected to contain remnants of functionalized syngenetic molecules. In that case, the essential terrestrial chemical radiation protectant would have also been required on Mars and scytonemin has been identified as a prime biomarker in the search for life on Mars, for example in cryptoendolithic geological niches similar to the Mars analogue sites at Pilbara in Australia and Barberton in South Africa. This chert (Fig. 2) is highly fractured and contains good zones of contaminating fluid flow throughout the chert. In addition, endolithic micro-organisms can easily colonize along fracture zones.
Experimental details
Scytonemin, Lyngbya sp. was obtained from Calbiochem (Merck KgaA, Darmstadt, Germany) in 90.7% purity (by high-performance liquid chromatography (HPLC)) and was used without further purification.
Fourier transform Raman spectroscopy
Raman spectra were recorded using a Bruker IFS 66/FRA 106 instrument (Ettlingen, Germany) with an Nd3+/YAG solid state laser operating at 1064 nm, with a power at the sample of 97 mW and a liquid nitrogen cooled InGaAs detector. Spectra were accumulated over 200 scans at a spectral resolution of 4 cm−1 with a spectral footprint of approximately 100 μm. The spectral wavenumber range was typically 4000–80 cm−1. Wavenumbers of sharp Raman bands are correct to better than ±1 cm−1. A daily interferometer function check is required and is achieved by recording the Raman spectrum of sulphur. Spectra were not corrected for instrument response. The spectrometer was controlled by a PC with instrument control software (Bruker Opus NT version 3).
Fourier transform infrared spectroscopy
Infrared (IR) transmission measurements were carried out between 4000 and 650 cm−1 with a Digilab Scimitar S Series spectrometer, equipped with a deuterated triglycine sulphate (DTGS) detector and a KBr beam splitter. The samples were presented in the form of a KBr disc and the data were recorded at a resolution of 4 cm−1 with the accumulation of 64 scans.
Computational details
The strategy was to first define a conformer search by molecular mechanics; the molecular mechanics calculations resulted in potential energy minima at about 60° and 120° for the torsional angle C5—C1—C22—C34 with a low rotational barrier in between these limits and showed potential energy maxima near 0° and at 180°. The numbering refers to the skeletal atomic positions shown in Fig. 3. Similar calculations were carried out making use of the semi-empirical methods Modified Neglect of Differential Overlap (MNDO), Austin Model 1 (AM1) and Parametrized Model number 3 (PM3), which also gave approximately the same potential energy surface for the molecule. Both of the molecular mechanics and semi-empirical calculation methods did not give low-energy structures when the two para-substituted phenol rings were in a planar conformation and the lowest energy; the most stable structure was given when these were not coplanar and were rotated out-of-plane by about 50° with respect to each other.
Scytonemin structure as usually presented in the literature in the planar conformation providing the atomic numbering used in the calculations and text.
The energy minima obtained by the molecular mechanics and semi-empirical methods were subjected to ab initio calculations for which the package program Spartan (Spartan'O4 Wavefunction, Inc. Irvine, CA) was used. The geometry optimizations have been carried out at an HF/6-31G** level of sophistication. The minima have been verified by frequency calculations made at the same level and no imaginary frequencies were created. The characteristic parameters for the minima obtained in these calculations, designated as Conformer 1 and Conformer 2, are reported in Table 1 along with those for Conformers 3–5, which represented other low-energy, possible structures for scytonemin. Geometry optimizations at 20°, 90° and 180° were also undertaken to provide additional quantitative information on the molecular potential energy surface. 1H and 13C NMR chemical shifts, electronic absorption spectra, IR and Raman wavenumbers have all been calculated at the HF/6-31G** level for Conformer 1, but only the electronic and vibrational spectra are reported here. The configuration of Conformer 1 is shown in Fig. 4.
Conformer 1 (132.5° rotation about the C—C dimer bond).
Calculated parameters for the five lowest energy conformers of scytonemin
Configuration Interaction Singles (CISs) calculations were carried out making use of the program package Gaussian (Frisch et al. Reference Frisch1998) with key words CIS=(Nstates=13), CIS=Direct to obtain absorption spectra; the program package Spartan calculates only the first five singlet excitation energies to the first lowest excited state.
The Raman wavenumbers and activities were obtained from calculations using the program package Gaussian for the monomer as Spartan does not provide Raman activities.
Results and discussion
Scytonemin has two identical planar chemical moieties that can rotate around the C—C (C1—C22) bond that links the two halves of the molecule. The potential energy surface (potential energy versus angle of rotation about the axis C5—C1—C22—C34) has a maximum at 0°; at this point the molecule is planar. The two carbonyls face each other and at the other molecular extreme the two aromatic moieties of each molecular half actually touch each other. Steric hindrance cannot therefore allow this molecule to adopt this planar conformation and the potential energy of this structure is high; despite this, scytonemin has always been represented as a planar molecule in the literature, a structure that we believe has been engendered by the mistaken conception that this structure would best serve the electronic orbital delocalization in the dimer molecule.
There exists a second maximum energy conformation at about 180° rotation with a relative energy of 60 kJ mol−1 above that of the global minimum; this is a planar molecular conformation, but now the two carbonyls are at opposite sides of the C1—C22 bond. At this position, each carbonyl oxygen and the H atom from the aromatic moiety on the other molecular half are in very close proximity, for example the O35 to H29 distance=1.360 Å. This would undoubtedly cause significant steric hindrance if the molecule is forced into planarity in this conformation. Upon energy optimization, however, this O—H distance becomes 1.975 Å, which considerably reduces the steric hindrance, but this can only be achieved at the cost of lengthening the C1—C22 bond and requires a tilting of the two halves of the dimer.
There are two energy minima noted on the potential energy surface between 0° and 180° rotation. The global primary minimum occurs at a rotation of 132.5° (Fig. 4) about the C1—C22 bond and the second minimum occurs at a rotation of 58.9°; however, the secondary minimum is 4 kJ mol−1 higher in energy than the global primary minimum energy conformation. There is a third maximum energy noted on the potential energy surface, with a low rotational barrier between these two energy minima at about 90° rotation. Here, the two molecular halves are orthogonal to each other and this conformation is calculated to be 10 kJ mol−1 higher in energy compared with the global primary minimum conformational energy.
It can be concluded from these calculations that scytonemin must be non-planar and flexible to rotation about the C1—C22 bond axis. The structural parameters for the minimum energy conformations are not presented in detail here, but are available on request. It is interesting to note further that the rotor bond (C1—C22) is shortest at the global primary energy minimum (1.4625 Å), longer at the second minimum (1.4635 Å) and lengthens even more to 1.4670 Å in Conformer 5 (the 20° rotation conformer). It further increases to 1.482 Å in Conformer 4 and decreases again to 1.4690 Å in Conformer 3. As the scytonemin molecule becomes more planar overall, the two carbonyl bonds lengthen (1.187 Å at 20°, 1.189 Å at 58.9°, 1.190 Å at 90° and 1.192 Å at 132.5° and 180° rotations).
The highest occupied molecular orbital (HOMO) orbital has an anti-bonding character over the (O)C—C—C—C(O) atom series (C5—C1—C22—C34) and apparently has no effect on the carbonyls. The lowest unoccupied molecular orbital (LUMO) orbital has a bonding character over the (O)C—C—C—C(O) atom series (C5—C1—C22—C34) and an anti-bonding character over the carbonyls. The preference of the E configuration over the Z configuration for the double bonds linking the phenols in scytonemin is 24 kJ mol−1 and the E configuration has a smaller delta EHOMO–LUMO value of 0.03 eV (3 kJ mol−1). The relative HOMO–LUMO energy differences show that the bonding–anti-bonding energy gap decreases as the molecule becomes more planar and is greatest in the orthogonal conformation. Conversely, the introduction of steric hindrance forces the molecule to adopt a significantly non-planar configuration.
The lowest-energy conformer of the two minimal potential energy forms is also the one with a lower dipole moment, but both low-energy conformations have a significant dipole moment. This will be significant for the spectroscopic analyses and interpretation of the IR and Raman spectra, since in the strictly planar high-energy form of scytonemin in the molecular structure that is usually portrayed in the literature and in which both carbonyl groups are diametrically opposed to each other, the molecular symmetry point group is D 2h with a centre of inversion, whereas the non-planar conformational structures are of much lower symmetries. The conclusion is that the former structure would be subject to the strict operation of the rule of mutual exclusion and there would be no coincidences between the observed Raman and IR spectral bands; the non-planar forms of scytonemin would not exhibit this effect and several coincidences should then be observable between the Raman and IR spectral data. Since the IR spectrum of scytonemin has never been reported hitherto for comparison with the Raman spectrum, there is thus an opportunity to test experimentally the conclusions of our ab initio and potential energy calculations and to compare these with the observed vibrational spectra for scytonemin here for the first time.
Comparison of Raman and IR spectra of scytonemin
The Raman and IR spectra of scytonemin are shown in Figs 5 and 6, respectively. Molecules, which possess a centre of symmetry, strictly have no coincidence in the fundamental vibrational wavenumbers between the IR and Raman spectra, as a consequence of the rule of mutual exclusion.
Vibrational spectra of scytonemin; upper spectrum, FTIR spectrum in the wavenumber range 3600–650 wavenumbers, spectral resolution 4 wavenumbers, 64 spectral scans; lower spectrum, FT-Raman spectrum in the wavenumber range 3200–300 wavenumbers, 1064 nm excitation, spectral resolution 4 wavenumbers, 2000 spectral scans.
Expanded wavenumber scale FTIR and FTR spectra from Fig 5, wavenumber range 2000–550 wavenumbers.
The non-planar form, low energy conformation proposed for the scytonemin on the basis of our calculations has lost the centre of symmetry that exists in the published literature structure and would not exhibit mutually exclusive Raman and IR bands. Hence, we should expect to observe some coincidences between Raman and IR spectral data. The observed Raman and IR spectra reported here (Figs 5 and 6) indeed show that there are several significant coincidences (see Table 2), which provide the observed Raman and IR spectral wavenumbers and also the spectral data from the ab initio calculations on which the molecular assignments and mode descriptions have been made. Notably, the spectral bands at 1110 cm−1 (δ(COH) phenolic), 1270 cm−1 ν(CC) ring breathing pyrole, 1440 cm−1 ν(CCH) p-disubstituted aromatic ring, 1590 cm−1 ν(CCH) aromatic ring quadrant stretch and 3020 cm−1 ν(CH) vinylic all have Raman and IR coincidences, which proves that scytonemin cannot be a planar centrosymmetric molecule as depicted in the literature and shown in Fig. 1, so lending weight to the thesis proposed here that the true structure is significantly non-planar.
Observed and calculated Raman and IR spectral bands with assignments based on the ab initio calculations
Electronic spectra
The electronic energy levels calculated for the Conformer 1 of the scytonemin shown in Fig. 3 enable a prediction to be made of the electronic spectrum for this molecule. The observed electronic spectrum reported in the literature for scytonemin and the calculated transitions from the present study are shown in Table 3:
Electronic absorption spectrum, observed and calculated, for scytonemin
It is seen that the calculated and observed electronic absorption spectrum of scytonemin agree very well; of particular significance are the low-wavelength absorption features in the UVC region, which confer upon the scytonemin molecule its ability to protect the organisms against high-energy UV radiation insolation in their stressed geological environments. Our calculations reveal that the monomer from which scytonemin is derived, has a significantly different electronic absorption spectrum to that of its dimer and the two critical absorption bands at 344 and 212 nm are absent. This means that the protective role of the monomeric molecule would not be as effective as its dimeric scytonemin.
Conclusions
The results of the ab initio structural calculations on the cyanobacterial protective molecule, scytonemin, predicting the stability of the non-planar dimer are supported by the observed Raman and IR spectral data. The importance of this structure in the absorption of high-energy, low-wavelength UV radiation in cyanobacterial colonies is apparent in comparison with the monomeric structural analogue. The molecular assignments and confirmation of the vibrational spectroscopic bands in scytonemin will provide a firm basis for the detection of this key biomolecule in complex geological and biogeological matrixes, which are of direct relevance to the survival of terrestrial extremophiles and the search for extinct and extant life on Mars in future space missions that use IR and Raman spectrometers in remote robotic geological and biogeological exploration. Following the recent discovery of derivatives of scytonemin, it is even more critical that the behaviour of the parent molecule in this important biological protective class of compound is better understood and defined so that the changes of molecular structure can be properly assessed for the prediction of the survivability of organisms in geological niches in stressed terrestrial and extraterrestrial environments.
Acknowledgements
Tereza Varnali thanks Bogazici University Research Fund (Project Number 08B508) for research funding to carry out this project; Howell Edwards and Michael Hargreaves thank the EPSRC for funding this research.
