Properties of cyanobacterial UV-absorbing pigments suggest their evolution was driven by optimizing photon dissipation rather than photoprotection
PProperties of cyanobacterial UV-absorbing pigmentssuggest their evolution was driven by optimizingphoton dissipation rather than photoprotection
Aleksandar Simeonov and Karo Michaelian Institute of Biology, Faculty of Natural Sciences and Mathematics,University “Ss. Cyril and Methodius”, 1000 Skopje, Macedonia. Instituto de F´ısica, UNAM, Circuito Interior de la Investigaci´on Cient´ıfica,Cuidad Universitaria, M´exico D.F., Mexico, C.P. 04510.
Abstract
An ancient repertoire of ultraviolet (UV)-absorbing pigments which survive today inthe phylogenetically oldest extant photosynthetic organisms, the cyanobacteria, pointto a direction in evolutionary adaptation of the pigments and their associated biota;from largely UV-C absorbing pigments in the Archean to pigments covering ever moreof the longer wavelength UV and visible in the Phanerozoic. Such a scenario impliesselection of photon dissipation rather than photoprotection over the evolutionary his-tory of life. This is consistent with the thermodynamic dissipation theory of the originand evolution of life which suggests that the most important hallmark of biologicalevolution has been the covering of Earth’s surface with organic pigment moleculesand water to absorb and dissipate ever more completely the prevailing surface solarspectrum. In this article we compare a set of photophysical, photochemical, biosyn-thetic and other germane properties of the two dominant classes of cyanobacterialUV-absorbing pigments, the mycosporine-like amino acids (MAAs) and scytonemins.Pigment wavelengths of maximum absorption correspond with the time dependenceof the prevailing Earth surface solar spectrum, and we proffer this as evidence for theselection of photon dissipation rather than photoprotection over the history of life onEarth. a r X i v : . [ phy s i c s . b i o - ph ] F e b Introduction
Once the subject of mystical and metaphysical interpretations, the explanation of life onEarth has slowly gained a physical-chemical grounding in biochemistry and non-equilibriumthermodynamics. Founded on Boltzmann’s nineteenth century insights into thermodynam-ics, then further elaborated by twentieth century scientists, notably by Ilya Prigogine, non-equilibrium thermodynamics attempts to explain the phenomenon of life as “dissipativestructuring”; an out of equilibrium organization of matter in space and time under an im-pressed external potential for the explicit purpose of producing entropy (Prigogine, 1967;Glansdorff and Prigogine, 1971).Using the formalism of Prigogine’s “Classical Irreversible Thermodynamics” in the non-linear regime, Michaelian (2009; 2011; 2012; 2013; 2016) has proposed a theory for life’sorigin and evolution as microscopic self-organized dissipative structuring of organic pigmentmolecules and the dispersal of these over the entire surface of the Earth as a response to theimpressed high-energy (UV-C to visible) solar photon spectrum prevailing at Earth’s surface.All physicochemical structuring associated with the pigments, such as the photosyntheticorganisms primarily, and heterotrophic organisms secondarily, can be regarded as agents forthe synthesis, proliferation and distribution of the pigments. The theory suggests that itis the thermodynamic imperative of increasing the entropy production of Earth in its solarenvironment that is behind the vitality of living matter as seen in its ability to proliferate,diversify, and evolve.The theory explains satisfactorily, for example, why the three major classes of photosyn-thetic pigments (chlorophylls, carotenoids and phycobilins) of phototrophic organisms dissi-pate most of the absorbed photonic energy into heat (a process known as non-photochemicalquenching, NPQ) while funneling only a minute fraction into productive photochemistry(Horton et al., 1996; Ruban et al., 2007; Staleva et al., 2015; Gupta et al., 2015). Moreover,these organisms often contain a vast array of other organic pigments in addition to the photo-synthetic ones, whose absorption spectra extend outside of the photosynthetically active radi-ation (PAR) in the visible, and into the UV-A, UV-B, and UV-C regions, hence allowing fullcoverage of the past and present incident surface solar spectra (Michaelian, 2012; Michaelianand Simeonov, 2015). In the biological literature these phenomena have been explained pri-marily through invoking the conventional wisdom of photoprotection (Demmig-Adams andAdams, 1992; Mulkidjanian and Junge, 1997; Wynn-Williams et al., 2002; Mulkidjanian etal., 2003; Castenholz and Garcia-Pichel, 2012).Photoprotective roles have especially been attributed to UV-absorbing biological pig-ments (e.g., mycosporine-like amino acids and scytonemins in cyanobacteria and algae;flavonoids and anthocyanins in plants, etc.) since they don’t seem to contribute to photosyn-thesis at all (Moisan and Mitchell, 2001). These theories usually trace the photoprotectiverole of UV-pigments back to the beginnings of cellular life in the early Archean when UVradiation was far more important component of the surface solar spectrum than it is today(Sagan, 1973; Mulkidjanian and Junge, 1997; Garcia-Pichel, 1998; Cockell and Knowland,1999; Mulkidjanian et al., 2003). UV-screening ostensibly conferred pigment-containing or-ganisms Darwinian advantages for survival in the harsh Archean environment of intense UV2adiation.However, from a thermodynamic viewpoint, the UV is the region of the solar spectrumwith the greatest possible entropy production potential per unit photon dissipated. There-fore, under the high UV ambient conditions of our primitive planet, non-equilibrium ther-modynamic principles of increasing the entropy production of Earth in its solar environmentwere probably the motive force for the development of these microscopic dissipative struc-tures in the form of efficient UV-dissipating organic compounds (Michaelian, 2011; 2013),rather than metaphysical forces giving rise to a hypothetical “will to survive” of the indi-vidual cell. The evidence for this inextricable link between UV light and nascent life hasbeen reinforced with the verifications for biogenicity of ∼ Properties of cyanobacterial UV-absorbing pigments
MAAs and a related group of organic compounds called mycosporines represent a large familyof colorless, low-molecular-weight ( <
400 u), water-soluble, usually intracellular secondarymetabolites widespread in the biological world (Dunlap and Chalker, 1986; Carreto et al.,1990; Rosic and Dove, 2011). The exact number of compounds within this family is yetto be determined, since they have only relatively recently been discovered (for a historicaloverview, see Schick and Dunlap, 2002 and ˘Rezanka et al., 2004), and novel molecular speciesare constantly being uncovered. Thus far, however, their number is around 40 (Wada et al.,2015). The name “mycosporine” has to do with them being originally isolated and chemicallyidentified from mycelia of sporulating fungi, where it was thought they played a role in light-induced sporulation (Leach, 1965; Trione and Leach, 1969).
Chemically both MAAs and mycosporines are alicyclic compounds (see Fig. 1) sharing acentral 5-hydroxy-5-hydroxymethyl-2-methoxycyclohex-2-ene ring with an amino compoundsubstituted at the third C-atom and either an oxo or an imino functionality at the firstC-atom (Favre-Bonvin et al., 1976; Ito and Hirata, 1977; Arpin et al., 1979; Karentz, 2001).While most authors don’t make a clear chemical distinction between the two groups, severalauthors (for example: Bandaranayake, 1998; Schick and Dunlap, 2002; Carreto et al., 2011;Molin´e et al., 2014; Miyamoto et al., 2014) when using the term mycosporines refer onlyto those molecular species with a central amino-cyclohexenone chromophore (also calledoxo-mycosporines), and when using the term MAAs refer only to molecules with a centralamino-cyclohexenimine chromophore (also called imino-mycosporines). The N-substitutionon C-3 with different amino acids or amino alcohols is what gives the diversity of molecularstructures within both groups (Korbee et al., 2006; Sinha et al., 2007; Carreto and Carignan,2011). Within the MAA group, the most common amino acid on the C-3 position is glycine,whereas they also have a second amino acid, amino alcohol or an enaminone system attachedat the C-1 position (D’Agostino et al., 2016).This unique molecular structuring and bonding begets their unique spectral properties.MAAs are considered to be one of the strongest UV-A/UV-B-absorbing substances in nature(Schmid et al., 2000); with wavelength absorption maxima ( λ max ) in the 310-362 nm intervaland molar attenuation coefficients ( ε ) between 28100 and 50000 M − cm − (Dunlap andSchick, 1998; Carreto et al., 2005; Gao and Garcia-Pichel, 2011). Their absorption spectraare characterized by a single sharp λ max with a bandwidth of approximately 20 nm and onlyabout 2-3 nm apart from the λ max of structurally similar MAAs (see Fig. 2) which makesit very difficult to distinguish these compounds based solely on their absorption spectra(Karentz, 1994; Carroll and Shick, 1996).The values of λ max and ε are dependent on the degree of derivatization of the centralring and the nature of the nitrogenous side groups (in particular the presence of additional4igure 1: Chemical structures of some common mycosporines and MAAs.conjugated double bonds) (Singh et al., 2010; Wada et al., 2015). Smaller, mono-substitutedmycosporines (typically oxo-mycosporines) have their λ max values shifted to shorter wave-lengths in the UV-B and usually have lower ε values; whereas MAAs (imino-mycosporines)are normally bi-substituted, with higher ε values and λ max values shifted to longer wave-lengths in the UV-A (Portwich and Garcia-Pichel, 2003).For example, the direct metabolic precursor of all mycosporines, 4-deoxygadusol (Fig.1), which has the minimal level of derivatization, has λ max = 268 nm at acidic pH, and λ max = 294 nm at basic pH; mycosporine-glycine (Fig. 1), the simplest oxo-mycosporineand direct precursor of all other mycosporines and MAAs has a λ max = 310 nm, whereaspalythine (Fig. 1), a simple mono-substituted MAA (amino-MAA), has λ max = 320 nm and ε = 36200 M − cm − (Carreto et al., 2005; Gao and Garcia-Pichel, 2011). Palythene (Fig.1), a bi-substituted MAA with an additional conjugated double bond, has one of the mostred-shifted bands of all known MAA species, with λ max = 360 nm and ε = 50000 M − cm − (Uemura et al., 1980).The observed red-shift in λ max is a consequence of the degree of resonance delocalizationinside the molecules; the more efficient is the electron delocalization (i.e. the stronger the π -conjugation character) the lower is the energy requirement for an electronic transition andconsequently the higher are the λ max and ε values (Carreto and Carignan, 2011; Wada etal., 2015).From a thermodynamic perspective, the fate of the electronic excitation energy is alsoa very relevant aspect of the absorption event since it is directly linked to the amount ofentropy produced by the dissipative microscopic structure (i.e. the polyatomic molecule).Nonradiative, vibrational relaxation pathways of the excited state lead to more efficientenergy dissipation and higher entropy production when compared to the fluorescent or phos-phorescent radiative decay channels (W¨urfel and Ruppel, 1985; Michaelian, 2011; 2012;2016). In this respect MAAs prove to be very efficient dissipative structures, although allstudies hitherto have discussed these thermodynamically relevant characteristics only fromthe standpoint of photostability and UV-photoprotection.5igure 2: Complementary absorption spectra of scytonemin and MAAs. (Adapted fromRastogi and Madamwar, 2016).Aiming at fully describing their photophysical and photochemical properties and expand-ing the evidence on the assigned UV-photoprotective role, Conde et al. (2000; 2004; 2007)made several in vitro studies on the excited-state properties and photostability of variousMAAs in aqueous solution (see Table 1). The results showed picoseconds excited state life-times, very low fluorescence quantum yields (e.g., φ F (porphyra 334) = 0 . φ T (porphyra 334) < . φ R (porphyra 334) = 2 − × − ) for all of the MAAs studied.These results are consistent with a very fast internal conversion (IC) process as the maindeactivation pathway of the excited states, which was directly quantified by photoacousticcalorimetry experiments showing that ∼
97% of the absorbed photonic energy is promptlydissipated into the surrounding medium as heat (Conde et al., 2004).A computational study by Sampedro (2011) using the CASPT2//CASSCF protocol(Olivucci, 2005) and employing palythine as a model compound, confirmed these findings.The study indicates that the fast IC processes connecting the S / S and S / S states aredue to the presence of two energetically accessible conical intersection points that can bereached by small geometrical changes in the atomic coordinates. It is now well establishedthat conical intersections (a.k.a. molecular funnels or diabolic points) play a very importantrole in fast, non-radiative de-excitation transitions from excited electronic states to groundelectronic state of molecules, particularly in many fundamental biological molecules, such asDNA/RNA, amino acids and peptides (Schermann, 2008). They enable effective coupling ofthe electronic degrees of freedom of the molecule to its phonon degrees of freedom, therebyfacilitating radiationless decay by vibrational cooling to the ground state (in the processconverting the absorbed high frequency UV photon into many low frequency infrared pho-6ons), which could make them examples of microscopic dissipative structuring of material inresponse to the impressed photon potential (Michaelian, 2016). MAAs and mycosporines are cosmopolitan substances in “optical” habitats - planktonic,benthic and terrestrial; with the largest concentrations detected in environments exposed tohigh levels of solar irradiance (Castenholz and Garcia-Pichel, 2012 and references therein).They are now known to be the most common type of UV-absorbing natural substances,especially among aquatic organisms (Rastogi et al., 2010).While mycosporines have been reported only in the kingdom Fungi (mycosporine-glycineand mycosporine-taurine are exceptions), MAAs are more extensively distributed amongtaxonomically diverse organisms (Karsten, 2008; Carreto and Carignan, 2011). These in-clude: cyanobacteria; heterotrophic bacteria; dinoflagellates, diatoms and other protists; redalgae; green algae; various marine animals, especially corals and their associated biota (for adatabase on the distribution of MAAs, see Sinha et al., 2007). They seem to be completelyabsent from terrestrial plants and animals, but are regularly found in terrestrial cyanobac-teria (Garcia-Pichel and Castenholz, 1993) and terrestrial algae (Karsten et al., 2007).An interesting discovery by Ingalls et al. (2010) reveals that MAAs represent a consid-erable portion of the organic matter bound to diatom frustules, accounting for 3-27% of thetotal carbon and 2-18% of total nitrogen content of the frustules. Previously establishedviews held that MAAs have mainly an intracellular location in these organisms.
The cyclohexenone core of MAAs is derived from intermediates of two fundamental anabolicpathways; the shikimate pathway (Favre-Bonvin et al., 1987; Shick et al., 1999; Portwichand Garcia-Pichel, 2003) and the pentose phosphate pathway (Balskus and Walsh, 2010),with the shikimate pathway being the predominant route for UV-induced MAA biosynthesis(Pope et al., 2015).In MAA/mycosporine biosynthesis both pathways converge at a point where their respec-tive 6-membered cyclic intermediates with similar structures are converted to 4-deoxygadusol(Fig. 1), the common precursor of all MAAs and mycosporines, a reaction catalyzed by thekey enzyme O-methyltransferase (Pope et al., 2015; D’Agostino et al., 2016).These basic biochemical pathways lay at the heart of carbon metabolism, shared by allthree domains of life; the shikimate pathway links carbohydrate catabolism to the biosyn-thesis of the aromatic amino acids and other aromatic biomolecules; similarly the pentosephosphate pathway uses glycolysis for the synthesis of pentose sugars, the nucleotide build-ing blocks (Cohen, 2014). Thus, they are considered to have an ancient evolutionary origin,possibly even dating back to prebiotic times (Richards et al., 2006; Keller et al., 2014).As mentioned in the previous section a very interesting trait of MAAs is that they areextremely prevalent natural compounds produced by a variety of taxonomically very distant7rganisms from simple bacteria to algae and animals. A natural question arises: how canevolutionary so distant organisms share the same MAA encoding genes?Several lines of evidence now suggest that the progenitor of the enzymatic machinery forMAA biosynthesis was probably a cyanobacterium or the cyanobacterial ancestor, while en-dosymbiotic events and prokaryote-to-eukaryote lateral gene transfer events during evolutionaccount for their prevalence among all other biological taxa (Rozema et al., 2002; Waller etal., 2006; Starcevic et al., 2008; Singh et al., 2010; Singh et al., 2012).Numerous in vitro experiments with manipulation of ambient UV light have demonstratedthat wavelengths between 280-320 nm (UV-B) are generally the most effective inducers forthe biosynthesis of MAAs, while UV-A and visible wavelengths have a lesser effect (Sinhaet al., 2001; Rastogi and Incharoensakdi, 2014). Portwich and Garcia-Pichel (2000) pro-posed a certain reduced pterin molecule with a distinct absorption peak at 310 nm as thephotoreceptor involved in the induction of cyanobacterial MAA biosynthesis.
Since their discovery in the 1960’s, authors have struggled to confer a single specific physio-logical function to MAAs. Although from the beginning a UV-photoprotective role seemedmost conspicuous, largely because of their unique UV-dissipating traits and the fact thattheir production is stimulated by UV-B. Later, this theory faced serious challenges, forexample, the failure to find a correlation between intracellular MAA accumulation andUV-resistance in certain coral zooxanthellae (Kinzie, 1993), the phytoplankton
Phaeocys-tis antarctica (Karentz and Spero, 1995), the dinoflagellate
Prorocentrum micans (Lesser,1996), certain cyanobacterial strains (Quesada and Vincent, 1997), and the red alga
Palmariapalmata (Karsten et al., 2003), etc.As a response, many researchers in the field came up with their own suggestions forMAA physiological roles, sometimes very different from the sunscreen role, such as; osmoticregulation, antioxidants, nitrogen storage, accessory pigments, protection from desiccation,protection from thermal stress, reproductive functions in fungi and marine invertebrates, etc.;all of which have also been challenged or discredited (for reviews of the different theoriesof MAA functions and the challenges they face, see: Korbee et al., 2006; Oren and Gunde-Cimerman, 2007; Rosic and Dove, 2011).From a traditional biological standpoint this apparent lack of a clear defining physiologicalfunction for these pigments looks extremely perplexing, especially when taking into accountthe extraordinary prevalence of these compounds in nature. Darwinian theory in its strictesttraditional formulation, with evolution through natural selection operating only at the level ofthe individual, categorically dismisses this kind of phenomena; where an organism wastefullyspends free energy and resources for the synthesis of metabolically expensive, nitrogen-containing compounds with no vital physiological function commensurate with their ubiquityand hence no, or little, benefit for its survival and reproduction. According to Darwiniantheory, such a biosynthetic pathway, with little or no direct utility to the organism, shouldhave been suppressed or completely eliminated through natural selection. However, exactlythe opposite has happened in the course of evolution; MAA biosynthetic genes have not8nly survived but have undergone extensive dissemination across numerous taxa throughhorizontal gene transfer.The failure of Darwinian theory to find a niche for MAAs in its classical “struggle forsurvival” paradigm is a result of it not being soundly grounded in thermodynamics and theuniversal physical laws (for a discussion on this topic, see: Michaelian, 2011; 2012; 2016).From the perspective of non-equilibrium thermodynamics, a metaphysical “will to survive”does not exist, making the search for a particular physiological function of MAAs pointless.But MAAs do have a function and it is a thermodynamic function of energy dissipation, or,more generally, entropy production. This thermodynamic function can be readily inferredfrom their physicochemical properties related to photon dissipation described above. MAAscan be regarded as typical examples of microscopic dissipative structuring of matter for thesole purpose of entropy production through highly efficient dissipation of high-frequency UVphotons into heat (Michaelian, 2016). This is the reason for their “coming into being” andtendency to proliferate over the surface of the Earth, as it is the fundamental reason forthe origin and evolution of life on Earth, and, in fact, the reason for the ubiquity of organicpigments in the neighborhood of stars throughout the universe (Michaelian and Simeonov,2017; Michaelian, 2016).This biological irreversible process of photon dissipation that MAAs and other bio-pigments perform, then couples to a secondary abiotic irreversible process of water evap-oration from surfaces through the heat it releases into its aquatic milieu (Michaelian, 2012).Evidence exist that the profusion of life and chromophoric dissolved organic matter (CDOM)in the sea-surface microlayer (SML) causes significant heating of the ocean surface fomentingevaporation (Morel, 1988; Kahru et al., 1993; Jones et al., 2005; Patara et al., 2012) andeven the irreversible process of hurricane formation and steering (Gnanadesikan et al., 2010).CDOM is the fraction of dissolved organic matter in water (DOM) that interacts withsolar radiation (Nelson and Siegel, 2013). Light energy absorption by CDOM at the surfaceof the ocean usually exceeds that of phytoplankton pigments; 54 ±
15% of the total lightabsorption at 440 nm and >
70% of the total light absorption at 300 nm is due to CDOM(Siegel et al., 2002; Babin et al., 2003; Bricaud et al., 2010; Organelli et al., 2014). Itis a complex and extremely variable mixture of organic pigments such as pheopigments(Bricaud et al., 2010), metal-free porphyrins (R¨ottgers and Koch, 2012), humic and fulvicacids (Carlson and Mayer, 1980; Galgani and Engel, 2016), aromatic amino acids (Yamashitaand Tanoue, 2003) and MAAs (Whitehead and Vernet, 2000; Steinberg et al., 2004; Tilstoneet al., 2010). While it was previously believed that CDOM in the open ocean is chiefly abyproduct of heterotrophic organisms recycling phytoplankton cell contents (Nelson et al.,1998), more recent observations (Romera-Castillo et al., 2010) suggest a large contributionfrom active plankton exudation.Active secretion of MAAs into the surrounding water during surface blooms was demon-strated for the colonial cyanobacterium
Trichodesmium spp (Subramaniam et al., 1999;Steinberg et al., 2004), for the dinoflagellate
Lingulodinium polyedrum (Vernet and White-head, 1996; Whitehead and Vernet, 2000) and for the dinoflagellate
Prorocentrum micans (Tilstone et al., 2010). Interestingly, Tilstone et al. (2010) found far greater MAA con-9entration in the sea-surface microlayer samples when compared to the near-surface (0-2m), and subsurface (0-110 m) samples. Whitehead and Vernet (2000) also concluded thatfree-floating MAAs contributed up to 10% of the UV absorption of the total DOM pool at330 nm during the
L. polyedrum bloom. This exudation of pigments by organisms into theirenvironment would also seem to have little Darwinian advantage.All of the evidence presented suffices to conclude, with some certainty, that MAAs join infunction most of the other bio-pigments in nature which act as catalysts for the dissipationof photons into heat at Earth’s surface and the coupling of this heat to other abiotic entropyproducing processes, such as; the water cycle, hurricanes, water and wind currents, etc.
In 1849, Swiss botanist Carl N¨ageli noted yellowish-brown cyanobacterial sheath coloration(N¨ageli, 1849), and in 1877 coined the name “scytonemin” for the color-producing pig-ment (N¨ageli and Schwenderer, 1877). Although occasionally mentioned in scientific papersduring the twentieth century, scytonemin was not isolated until 1991 when Garcia-Picheland Castenholz (1991) first made a more extensive study of the compound. Proteau et al.(1993) elucidated the chemical structure of scytonemin, which proved to be a completelynovel indolic-phenolic dimeric structure unique among all hitherto known natural organicsubstances. The carbon skeleton of this novel eight-ring homodimeric molecule was giventhe trivial name “the scytoneman skeleton” (Proteau et al., 1993). Already in 1994, an-other scytoneman-type molecule was isolated from the cyanobacterium
Nostoc commune , andtermed “nostodione A” (Kobayashi et al., 1994). Thus far, four additional substances with ascytoneman-type molecular structure, or a structure derived from it, have been isolated fromcyanobacteria: dimethoxyscytonemin, tetramethoxyscytonemin, scytonine (Bultel-Ponc´e etal., 2004) and scytonemin-imine (Grant and Louda, 2013); for which, in this review, we usethe colloquial terms “scytonemins” or “scytoneman pigments”.
Scytonemin (Fig. 3), the representative and most common member of this yet poorly-explored family of aromatic indole alkaloids, is a relatively small molecule (544 u) built fromtwo identical condensation products of tryptophanyl- and tyrosyl-derived subunits linkedthrough a carbon-carbon bond (Proteau et al., 1993). Its IUPAC name is (3E,3’E)-3,3’-Bis(4-hydroxybenzylidene)-1,1’-bicyclopenta[b]indole-2,2’(3H,3’H)-dione.Depending on the redox conditions it can exist in two inter-convertible forms: a pre-dominant oxidized yellowish-brown form which is insoluble in water and only fairly solublein organic solvents, such as pyridine and tetrahydrofuran, and a reduced form (Fig. 3)with bright red color that is slightly more soluble in organic solvents (Garcia-Pichel andCastenholz, 1991; Proteau et al., 1993). Dimethoxy- and tetramethoxyscytonemin can beconsidered as derivatives of reduced scytonemin, where one or both of the ethenyl groupsin the molecule have been saturated by two or four methoxy groups, respectively (Bultel-Ponc´e et al., 2004; Varnali and Edwards, 2010). Another moderate degree of modification of10igure 3: Chemical structures of scytonemin and reduced scytonemin.the parent scytoneman skeleton can also be seen in scytonemin-3a-imine (a.k.a. scytonemin-imine), where the C-3a atom of scytonemin has been appended with a 2-imino-propyl radical(Grant and Louda, 2013).Only the structure of scytonine deviates substantially from the dimeric scytoneman skele-ton, due to the loss of one para-substituted phenol group and ring openings of both cyclopen-tenones where successive methoxylation and secondary cyclizations take place (Bultel-Ponc´eet al., 2004).A full in-depth photophysical and photochemical characterization of scytonemins has yetto be attained; thus far only their elemental spectroscopic properties are known. Scytoneminabsorbs very strongly and broadly across the UV-C-UV-B-UV-A-violet-blue spectral region(see Fig. 2 and Fig. 4), with in vivo λ max at 370 nm and in vitro (tetrahydrofuran) λ max at386 and 252 nm, with smaller peaks at 212, 278 and 300 nm (Garcia-Pichel and Castenholz,1991; Garcia-Pichel et al., 1992; Sinha et al., 1999). Its observed long term persistencein cyanobacterial biocrusts or dried mats exposed to intense solar radiation might be anindication of exceptionally high photostability (Garcia-Pichel et al., 1992; Brenowitz andCastenholz, 1997; Fleming and Castenholz, 2007; Fulton et al., 2012; Lepot et al., 2014).Reduced scytonemin has a similar spectroscopic profile, with in vitro (tetrahydrofuran) λ max (nm) and ε ( M − cm − ) values: 246 (30000), 276 (14000), 314 (15000), 378 (22000), 474(14000) and 572 (broad shoulder 7600) (Varnali and Edwards, 2014). A comparable absorp-tion spectrum is also exhibited by scytonemin-imine, the mahogany-colored, polar derivativeof scytonemin, with slightly different λ max values when measured in acetone (237, 366, 437and 564 nm) and in ethanol (248, 305, 364, 440 and 553 nm) (Grant and Louda, 2013). Con-trary to these three scytoneman-type molecules, the methoxylated derivatives and scytoninedo not absorb strongly in the UV-A region but have very high absorbances in the UV-Cregion with in vitro (methanol) λ max (nm) and ε ( M − cm − ) values for dimethoxyscytone-min: 215 (60354), 316 (18143) and 422 (23015); for tetramethoxyscytonemin: 212 (35928)11igure 4: In vitro absorption spectrum of scytonemin. (Adapted from Sinha et al., 1999).and 562 (5944); and for scytonine: 207 (38948), 225 (37054) and 270 (22484) (Bultel-Ponc´eet al., 2004).Concerning the monomeric scytoneman-type molecules nostodione A and prenostodione,isolated from natural cyanobacterial blooms, it remains debatable whether they are genuinecyanobacterial pigments or just intermediates in the biosynthesis of scytonemin (Ploutnoand Carmeli, 2001; Soule et al., 2009a). Unlike MAAs, scytonemins are exclusively cyanobacterial sheath pigments (Pathak et al.,2016). All phylogenetic lines of sheathed cyanobacteria contain scytonemins (Proteau et al.,1993), notably strains of the genera
Nostoc , Calothrix , Scytonema , Rivularia , Chlorogloeopsis , Lyngbya , Hyella , etc. (Sinha and H¨ader, 2008); as well as cyanolichens of the genera
Peltula , Collema and
Gonohymenia (B¨udel et al., 1997).The mucilaginous extracellular sheath (matrix) consists of heteroglycans, peptides, pro-teins, DNA and different secondary metabolites (Pereira et al., 2009), where scytoneminsare usually deposited in the outer layers, giving the sheath its distinctive dark yellow tobrown color (Ehling-Schulz et al., 1997; Ehling-Schulz and Scherer, 1999). Up to 5% of thedry weight of cultured scytonemin-synthesizing cyanobacteria is due to the pigment, but innatural assemblages this value can be even higher (Karsten et al., 1998). Curiously, Abed etal. (2010) reported two to six times higher concentrations of scytonemin than chlorophyll a in cyanobacterial cryptobiotic soil crusts in the Oman Desert.Scytonemin-producing cyanobacteria typically inhabit highly insolated terrestrial, fresh-12ater and coastal environments such as deserts, exposed rocks, cliffs, marine intertidal flats,shallow oligotrophic fresh waters, hot springs, etc. (Castenholz and Garcia-Pichel, 2012 andreferences therein). In microbial mat communities, especially the extremophilic terrestrialand aquatic colonies, these cyanobacteria occupy the uppermost sunlit layers (Balskus etal., 2011). Scytonemin-imine, for example, was isolated from samples of natural Scytonemahoffmani mats growing under high to intense (300-2000 µ mol quanta m − s − ) photon fluxdensity (Grant and Louda, 2013). The methoxyscytonemins and scytonine were isolatedalongside scytonemin from colonies of Scytonema sp. growing on exposed granite at theMitaraka inselberg in French Guyana, a region subjected to intense UVR-insolation (Bultel-Ponc´e et al., 2004).
The biochemistry and genetics of cyanobacterial scytonemin biosynthesis has extensivelybeen investigated by Soule et al. (2007; 2009a; 2009b), Balskus and Walsh (2008; 2009;2011) and Sorrels et al. (2009). They have confirmed the assumption by Proteau et al.(1993), the discoverers of the scytonemin structure, that the scytoneman molecular scaffoldis actually a condensation product of the aromatic amino acids tryptophan and tyrosine.Michaelian (2011) and Michaelian and Simeonov (2015) have hypothesized that thesewere the first amino acids to enter into a photon-dissipation-driven association with nucleicacids in the prebiotic world, a scenario backed by their high conservation inside the DNA-binding sites of photolyase enzymes (Kim et al., 1992; Weber, 2005); a phylogeneticallyancient family of enzymes, common to all three domains of life (Selby and Sancar, 2006) andeven found in viruses (Srinivasan et al., 2001). Not only do these amino acids absorb in theUV themselves (Michaelian and Simeonov, 2015 and references therein), but they also serveas biosynthetic precursors for most known aromatic UV-absorbing bio-pigments, including:anthocyanins, flavonoids and polyphenols in plants, melanins in heterotrophic organisms,scytonemins in cyanobacteria, etc. (Knaggs, 2003).Eight of the genes that make up the 18-gene scytonemin biosynthesis cluster code forshikimate pathway enzymes for the biosynthesis of tryptophan and tyrosine, while the func-tions of the rest remain unresolved but are suspected to be involved in the coupling of thetryptophan- and tyrosine-derived precursors for the formation of the scytoneman skeleton(Ferreira and Garcia-Pichel, 2016). The expression of the whole gene cluster has been shownto be elicited by exposure to UV-A and UV-B light (Sorrels et al., 2009; Rastogi and In-charoensakdi, 2014).Sorrels et al. (2009) proposed an ancient evolutionary origin for the scytonemin biosyn-thetic pathway based on the combination of the fact that this gene cluster is highly conservedamong evolutionary diverse strains of cyanobacteria (Soule et al., 2007; 2009a), and theirown phylogenetic analyses implying that the cluster is under a purifying selection pressure.Intriguingly, Soule et al. (2009a) observed scytonemin biosynthetic genes even in somecyanobacterial strains incapable of producing the pigment (e.g.,
Anabaena and
Nodularia ),and interpreted this as a case of relic genetic information.13 .2.4 Function: traditional view vs. thermodynamic view
Similarly to MAAs, the Darwinian point of view can only describe scytonemin as an effi-cient protective biomolecule designed to filter out supposedly damaging high frequency UVradiation while at the same time allowing the transmittance of wavelengths necessary forphotosynthesis (Ekebergh et al., 2015).Within the framework of this traditional “struggle for survival” viewpoint, the majority ofauthors define scytonemins as an adaptive mechanism of extremophilic cyanobacteria thatcolonize harsh, inhospitable habitats experiencing high doses of UVR-insolation (Ehling-Schulz et al., 1997; Wynn-Williams et al., 1999; Hunsucker et al., 2001; Sinha and H¨ader,2008; Rastogi et al., 2014).Among the evidence for the accredited photoprotective role is the discovery that up to90% of incident UV photons are prevented from entering sheathed, scytonemin-producingcyanobacterial cells, thus accomplishing significant reduction in chlorophyll a photobleach-ing and maintaining photosynthetic efficiency (Garcia-Pichel and Castenholz, 1991; Garcia-Pichel et al., 1992). Other authors, in addition to the sunscreen role, ascribe supplementarydefensive roles to scytonemin such as protection from oxidative, osmotic, heat and desiccationstress (Dillon et al., 2002; Matsui et al., 2012).Furthermore, scytonemin’s superior UV-C-absorbing capabilities in vivo, experimentallyproven by treating cyanobacterial colonies with 0.5-1.0 W m − UV-C radiation added tonatural solar irradiance (Dillon and Castenholz, 1999), has led many authors to considermodern cyanobacterial production of scytonemins as a relic UV-protection mechanism fromthe pre-Great Oxygenation Event period (Garcia-Pichel, 1998; H¨ader et al., 2003). Indeed,Raman spectral biosignatures of scytonemins, carotenoids and porphyrins were unambigu-ously identified in ∼ ∼ λ max = 237, 366, 437, 564 nm in vitro), extending from theultraviolet (UVB & UVA) into the blue and green of the visible, appears to indicatea photoprotective role beyond shielding only UVR. That is, going on the premise thatevolution generates and retains only advantageous secondary metabolites, then whatis the role of the visible bands in this case?”2. Inability to explain the production of the strongly UV-C/UV-B-absorbing methoxyscy-14onemins and scytonine, in spite of the absence of UV-C wavelengths and the low in-tensity of UV-B in today’s surface solar spectrum. The question is raised by Varnaliand Edwards (2010): “The realization that scytonemin is the parent molecule of per-haps a whole family of related molecules is important in that an analytical challengeis generated to detect these family members in admixture and in the presence of eachother naturally, and also the question is raised about the role of these molecules inthe survival strategy processes involving scytonemin; what subtle changes to the radi-ation absorption process require molecular modification of what apparently is alreadya highly successful radiation protectant, especially when the molecular syntheses areaccomplished in energy-poor situations?”3. Inability to explain why many species of cyanobacteria do not synthesize scytoneminsnor MAAs but, nevertheless, successfully cope with UV-induced cellular damage byemploying only metabolic repair mechanisms (Quesada and Vincent, 1997; Castenholzand Garcia-Pichel, 2000).4. Soule et al. (2007) developed scytoneminless mutant of the cyanobacterium Nostocpunctiforme which proved to have indistinguishable growth rate from the wild typeafter both were subjected to UV-A irradiation. The conclusion of the authors was thatother photoprotective mechanisms can fully accommodate the absence of scytoneminin the mutant.In addition, very efficient absorption and dissipation of high-energy photons is not a prereq-uisite for photoprotection, but it is for dynamical dissipative structuring of material underan external generalized chemical potential. Nature has a simpler way of creating photopro-tective molecules, if this was really the intention, by making them either highly reflective ortransparent to UV wavelengths (Michaelian, 2016).These problems and paradoxes, generated when trying to explain scytonemins fromwithin the Darwinian photoprotection paradigm, can be resolved by employing establishednon-equilibrium thermodynamic principles. In this context, we will first address the ques-tions raised by Grant and Louda (2013) and Varnali and Edwards (2010), and then, based onall the evidence presented, we will assign a thermodynamic dissipative role to scytonemins.The seemingly paradoxical absorption spectra of scytonemin-imine, the methoxyscytone-mins and scytonine, which extend outside of the photoprotectively-relevant part of the spec-trum, make sense only when these bio-pigments are understood as microscopic dissipativestructures obeying non-equilibrium thermodynamic directives related to increasing the globalsolar photon dissipation rate (Michaelian, 2013; Michaelian and Simeonov, 2015; Michaelian,2016). Under these directives, one of the several ways to increase the global solar photondissipation rate is by evolving (inventing) novel molecular structures (pigments) that coverever more completely the prevailing surface solar spectrum (see Michaelian and Simeonov,2015). This is precisely what is observed in the absorption spectra of the different scy-toneman pigments. The strong visible absorption peaks of scytonemin-imine at 437 nm(violet) and 564 nm (green/yellow), of tetramethoxyscytonemin at 562 nm (green/yellow),of dimethoxyscytonemin at 422 nm (violet); and the strong near-UV-C/UV-B absorption15eaks of scytonine (270 nm) and dimethoxyscytonemin (316 nm) is exactly where the photo-synthetic pigments do not peak in absorption (see, for example, Rowan, 1989). It is becauseof this rich assortment of diverse pigment molecules with complementary absorption bandsthat cyanobacterial biofilms, mats and soil crusts in nature tend to have high absorptivities,low albedos and appear almost black in color (Ustin et al., 2009).This fact leads us to an important conclusion on the thermodynamic function of the scy-toneman pigments. We believe that it is most reasonable to consider the photon-dissipationrole of scytonemins as the terrestrial analogue of the function that MAAs perform in theopen aquatic environment. This assertion may be justified on their hydrophobic characterand their inextricable connection to the extracellular polymeric substances (EPSs) of thecyanobacterial sheaths. Ekebergh et al. (2015) have shown that scytonemins have the great-est photostability in vivo, where they are embedded in their natural extracellular matrixmilieu. These extracellular polymeric substances may therefore be playing the role of pro-viding the dissipative medium required to disperse the excess vibrational energy after photonexcitation of the pigment, bringing the system rapidly to the ground state.In wet terrestrial regions of the planet, the thermodynamic role of photon dissipationcoupled to the water cycle is performed mainly by the plant cover, but in arid and semi-arid lands, where vegetation is severely restricted, this function is allotted to microscopicassemblages of cyanobacteria, heterotrophic bacteria, algae and fungi known as biologicalsoil crusts or biocrusts (Evans and Johansen, 1999; Belnap and Lange, 2001). It is theorizedthat these types of microbial communities represented life’s pioneering on dry land and werethe dominant ecosystem on the continents up until the advent of land plants and animals inthe Early Devonian (Beraldi-Campesi et al., 2014).Michaelian (2013) postulated: “The most important thermodynamic work performed bylife today is the dissipation of the solar photon flux into heat through organic pigments inwater. From this thermodynamic perspective, biological evolution is thus just the dispersalof organic pigments and water throughout Earth’s surface... On Earth, organic moleculesare found only in association with water. As described above, this is most likely related tothe efficiency of organic pigments dissipating solar photons using the high frequency watervibrational modes to facilitate their de-excitation. Without water they are poor photondissipaters and easily destroyed by photochemical reactions. This is probably the primordialreason for the association of life with water.”In the context of this citation, we emphasize the fact that cyanobacteria isolated fromdry regions display very high capacity to excrete large amounts of EPSs (Huang et al., 1998;Hu et al., 2003; Roeselers et al., 2007; Rossi et al., 2012), which are the main constituentof the biofilm matrix and together with microbial filaments play a key structural role informing the biocrusts (Mager and Thomas, 2010; Karunakaran et al., 2011). The uniquehydrophilic/hydrophobic nature of the EPSs enables highly efficient water capture and waterstorage within the biocrust by allowing the creation of moistened microenvironments wherewater dynamics is intricately regulated (Colica et al., 2014 and references therein). Hence,crust-covered soils are very hygroscopic and always exhibit higher water content comparedto bare neighboring surfaces (Rossi and Phillips, 2015). This phenomenon is exactly what16e have postulated earlier, life’s fundamental role of “dispersing organic pigments and waterover Earth’s entire surface” (Michaelian, 2013).A very conspicuous analogy between these terrestrial macroscopic and microscopic photon-dissipating biological “carpets” can be drawn. In the same manner as ecological successionof plant coverage leads to old climax forests with higher pigment content and lower albedos(Pokorny et al., 2010; Maes et al., 2011), ecological succession in biocrusts leads to increasein biomass of the late-stage scytonemin-producing cyanobacteria, and consequently accumu-lation of scytonemins in the matrix, an effect macroscopically observed as darkening of thebiocrusted soil (i.e. decrease in albedo) (Couradeau et al., 2016). During dry periods indeserts when water availability is very limited, the heat generated from scytonemin’s photondissipation is expected to go predominantly into sensible heat of the biocrusts instead of intothe latent heat of vaporization of water, and this is exactly what Couradeau et al. (2016)found when they measured ∼ ◦ C higher temperature of biocrust-covered, dark soils incomparison to bare soils. 17
Discussion
In a previous work (Michaelian and Simeonov, 2015) we posited five basic tendencies thatorganic pigment evolution on Earth would have followed: (1) increases in the photon absorp-tion cross section with respect to the pigment physical size, (2) decreases in the electronicexcited state lifetimes of the pigments, (3) quenching of the radiative de-excitation chan-nels (e.g., fluorescence), (4) greater coverage of the surface solar spectrum, and (5) pigmentproliferation and dispersion over an ever greater surface area of Earth.To examine whether these five tendencies are satisfied with the evolutionary invention ofMAAs and scytonemins we compare their properties to those of the aromatic amino acids(AAAs) (see Table 1).Our reason for choosing the AAAs is twofold; (1) they are considered to be among theearliest chromophoric organic molecules used by life with a prebiotic origin (a subject dis-cussed earlier in the text, and in Michaelian (2011) and Michaelian and Simeonov (2015) ingreater detail), and (2) since both MAAs and scytonemins are derived from intermediatesof the shikimate pathway for AAA biosynthesis they most likely appeared later in evolutioncompared to the AAAs, probably when the biosynthetic machinery for the synthesis of theAAAs was already robust; an event that most likely long predated 3.4 Ga, considering thatBusch et al. (2016) demonstrated that the ancestral tryptophan synthase of the last univer-sal common ancestor (LUCA) was already a highly sophisticated enzyme at 3.4 Ga. Thisreasoning is also corroborated by the previously mentioned (see Sect. 2.2.4) identificationof Raman spectral biosignatures of scytonemin in ∼ λ max and ε values of gadusol in water are pH-dependent: 268 nm at pH < ≥ ∼ UV-absorbing bio-pigments λ max (nm) ε ( M − cm − ) Electronic excited state lifetime (ns) Fluorescence quantum yield ( φ F )Aromatic amino acids Phenylalanine (a) 257 195 7.5 0.024Tyrosine (a) 274 1405 2.5 0.14Tryptophan (a) 278 5579 3.03 0.13
Mycosporines and MAAs
Gadusol (b) 269 12400 - non-fluorescentMycosporine- γ -aminobutyric acid (c) 310 28900 - -Mycosporine-glutamic acid (c) 311 20900 - -Palythine (b, c) 320 36200 - non-fluorescentShinorine (b) 333 44700 0.35 0.002Porphyra-334 (b) 334 42300 0.4 0.0016Palythene (c) 360 50000 - - Scytonemins
Scytonemin (d) 252,278,300,384 - - non-fluorescentReduced Scytonemin (d) 246,276,314,378,474,572 30000,14000,15000,22000,14000,7600 - -Scytonemin-imine (e) 237,366,437,564 - - -Dimethoxyscytonemin (d) 316,422 18143,23015 - -Tetramethoxyscytonemin (d) 562 5944 - -Scytonine (d) 225,270 37054,22484 - -
Other poorly characterized cyanobacterial UV-absorbing pigments
Gloeocapsin (f) 392 - - -Microcystbiopterins (g) ∼ ∼
350 10000,3500 - - a)Berezin and Achilefu (2010); (b)Losantos et al. (2015a); (c)Wada et al. (2015); (d)Varnali and Edwards (2014); (e)Grant and Louda (2013); (f)Storme et al. (2015); (g)Lifshits et al. (2016). E ne r g y [ W m - µ m - ] Wavelength [nm] -2
475 Wm -2
672 Wm -2
865 Wm -2 P h e ny l a l a n i n e T y r o s i n e T r yp t oph a n G a du s o l M y c o s po r i n e - g a m a M y c o s po r i n e - g l u t a m i c P a l y t h i n e S h i no r i n e P o r phy r a - P a l y t h e n e Precursors S c y t on e m i n S c y t on e m i n S c y t on e m i n S c y t on e m i n R e du ce d S c y t on e m i n R e du ce d S c y t on e m i n R e du ce d S c y t on e m i n R e du ce d S c y t on e m i n R e du ce d S c y t on e m i n S c y t on e m i n - i m i n e S c y t on e m i n - i m i n e S c y t on e m i n - i m i n e D i m e t hoxy s c y t on e m i n D i m e t hoxy s c y t on e m i n S c y t on i n e G l o e o ca p s i n M i c r o c y s t b i op t e r i n s M i c r o c y s t b i op t e r i n s -2 -2 Figure 5: The expected Earth surface solar spectrum at the given dates since present(Michaelian and Simeonov, 2015) and the maximum absorption ( λ max ) of the mycosporineand scytoneman pigments and their aromatic amino acid precursors. The precursors havestrong absorptions only in the UV-C (230-280 nm) while scytonemins absorb strongly acrossthe UV-C, UV-B, UV-A, violet and blue regions, and mycosporines and MAAs usually havestrong absorption in the UV-B and UV-A regions. Neither scytonemins nor MAAs peakstrongly in the UV-B region from ∼ O and CH CHO)produced by UV-C light on common volcanic gases such as H S, H O and CO , were ab-sorbing strongly in this gap. This gap was later covered by O and O absorption after theGreat Oxygenation Event at ∼ ∼
220 to ∼
700 nm, there is a dip in the ∼
275 to ∼
325 nm interval, and two large maxima at ∼
250 nm and ∼
380 nm. This is exactly the kind of shape that would be expected ifthe selective force for the evolution of this pigment was our proposed Archean surface solarspectrum (Michaelian and Simeonov, 2015). Combining this crucial point with the previouslydiscussed facts on scytonemin, it is tempting to speculate that this pigment had a key rolein photon dissipation during the Archean, being capable of dissipating almost the entireArchean surface solar spectrum. The evolutionary invention of scytonemin’s derivatives,as well as the mycosporines, the MAAs and still many other extinct and extant biologicalpigments, most likely resulted from the necessity to complement scytonemin’s absorptionwith pigments that absorbed wavelengths reaching Earth’s surface but were poorly absorbedby scytonemin itself. This kind of complementary spectral relationship between scytoneminand the MAAs has been well documented by several authors (e.g., Ehling-Schulz and Scherer,1999; Ferroni et al., 2010; Castenholz and Garcia-Pichel, 2012) and is illustrated in Fig. 2.21
Conclusions
The available data on the ubiquity of pigments covering the region from the UV-C to theinfrared, many exuded by the organisms that produce them into the environment, make itincreasingly difficult to assign to them a protective or antenna role within the Darwinianparadigm of the optimization of photosynthesis in benefit of the organism. We believe thatsense can only be made of this by shifting the paradigm from one of “photoprotection” ofthe organism to the thermodynamic optimization of photon dissipation.A number of contemporary pigment lines, most notably scytonemins and the mycosporine-like amino acids, appear to harbor relics of ancient biosynthetic production routes based onthe most ancient of the amino acids, the aromatics. The aromatic amino acids have knownaffinities to their RNA anticodons (Majerfeld and Yarus, 2005; Yarus et al., 2009) and wereperhaps the first antenna pigments for photon dissipation in the UV-C at the beginnings oflife (Michaelian, 2011).These pigment lines absorb and dissipate rapidly in the UV-C as well as the UV-B, UV-Aand the visible. Some of these pigments are exuded into the environment which excludes thepossibility of assigning them a role in photoprotection. Their strong absorption and dissi-pation in regions out of the photosynthetically active radiation (PAR) has been perplexingto perspectives within the Darwinian paradigm since these pigments appear to have littleutility to the organisms themselves. In fact, they absorb exactly where the photosyntheticpigments do not (and where water does not) and appear to have complete coverage of theArchean to present day Earth surface solar spectra.It should be emphasized that our current knowledge of the diversity of cyanobacterial,algal and plant pigments and the thermodynamic function they perform is incomplete. Forexample, there are several indications of even richer diversity of UV-absorbing pigmentsin cyanobacteria than those hitherto characterized and classified into the two groups, my-cosporines and scytonemins. The chemical structure and other elemental properties of one ofthese poorly investigated pigments, named gloeocapsin, have yet to be determined, but initialresults suggest that it is chemically unrelated to both MAAs and scytonemins (Storme et al.,2015). Still other chemically distinct UV-absorbing cyanobacterial pigments, with a uniquepterin structure, have been reported elsewhere (Matsunaga et al., 1993; Lifshits et al., 2016).The wavelengths of maximum absorption of these two ill-defined groups of cyanobacterialpigments are listed in Table 1 and are plotted in Fig. 5. As with the mycosporines and thescytonemins, their absorption properties are consistent with the optimization of dissipationof the prevailing photon spectrum at Earth’s surface.Taken as a whole, these data seem to indicate that, rather than photosynthesis beingoptimized under a Darwinian “survival of the fittest” paradigm, that the origin and evolutionof life is driven by photon dissipation with the net effect of covering Earth’s entire surfacewith pigments and water, reducing the albedo and the black-body temperature at whichEarth radiates into space. It is our hope that this article will incite further investigation intothe proposition that photon dissipation efficacy has been the fundamental driver of biologicalevolution on Earth. 22 eferenceseferences