1.1 Time line for H,C,N,O chemistry: formation of elements

The current standard model of cosmology based on the Wilkinson Microwave Anisotropy probe (WMAP) sets the age of the Universe at 13.772±0.059 billion years [4], [5]. The era from then to now is displayed as a chemical time line (Fig. 1), with the latter part of the time showing the chemical systems dependent on nucleotide-code-based synthesis on Earth. Here we are concerned with earlier chemical systems that arose following the cataclysmic production of large quantities of oxygen, nitrogen and carbon in the era of the first stars, prior to 12.7 Gya.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Chemical Time Line.

https://doi.org/10.1371/journal.pone.0103036.g001

In the primordial nucleo-synthesis [6] the isotopes of hydrogen and helium, together with a small amount of lithium were produced. Later in the expansion of the universe (at roughly 400,000 years) the free electrons re-combined with these nuclei to produce neutral atoms of H, D, ³He, ⁴He and ⁷Li [7]. Accretion of first generation stars from essentially H and He was aided by dark matter “haloes” in which baryonic matter became gravitationally concentrated, then collapsed into massive stars, particularly in the range 63 to 130 solar masses [8], [9]. The predicted end-product of nucleo-synthesis in these “population III” stars was a stellar core comprising mostly oxygen, with some carbon and lesser amounts of nitrogen and higher mass elements. The life of these stars was short and within them terminal He burning lead to collapse during a brief period of “pair production” followed by explosive oxygen burning reactions that blew the star apart, or at least caused it to shed much of its mass. The related stellar ultraviolet radiation re-ionized the inter-galactic medium at redshift Z = 11 [10], [11] and by redshift Z = 6 (at 0.9 Gy) large galaxies were prolifically giving birth to stars [12]. The conditions that could lead to polymer amide production therefore first began to exist about 1 Gy into the life of the universe once the necessary elements H, C, N and O were dispersed, the radiation and baryon temperatures had fallen, and the next generation of smaller stars had begun to condense. Locally the density would be raised to at least that of the warm, dense clouds observed in our galaxy, approximately 10⁷ – 10⁸ H₂ molecules cm⁻³ in a cloud interior equilibrium radiation environment of 100 K or more.

1.2 From H,C,N,O to amino acids in space

Our thesis in the present paper is that a polymer of amino acids (HPA) could have formed in gas phase chemistry as the next stage of chemical evolution after formation of the numerous smaller molecules that have been catalogued [13]. If correct, HPA would comprise the first solid density material and the properties of HPA, particularly its richness in hydrogen bond sites, would provide adhesion to start the accretion of other free molecules, also of the metal ions present at low density in early times. Although the simplest amino acid glycine has not been definitively observed in space [13]–[15] its gas phase production is expected. For our proposal to be valid, before interstellar “dust” or “ice” exists to provide the solid density route to glycine [16], there has to be a purely gas phase route from simple molecules to glycine and other amino acids at the relatively low density and temperature of a warm dense cloud. In the laboratory amino acids may be able to form purely in the gas phase when appropriate gas mixtures are subject to electric discharge [17] or ionizing radiation [18]. However, it seems that none of these demonstrations has been performed in conditions where contact with a surface was impossible during the duration of the experiment and therefore an element of surface or volume solid density reaction chemistry could not be completely excluded. In [17], for example, a water phase reaction appeared to account for all of the amino acid synthesis and in [18], a water surface was present. Moreover, the thermal energy and density available in these experiments would allow some reactions to proceed that would not occur at the lower temperatures of interstellar space. A more definitive result obtains when likely individual reactions in the chain are studied in a helium-buffered flow tube [19], in which the reaction rate is high enough to ensure that no contact with a surface occurs. The difficulty with this complementary approach, however, is that very many reactions, possibly numbering in the thousands, have to be studied in order to establish a most probable gas phase reaction sequence to amino acids. For example, a promising gas phase reaction [19] to produce (ionized) glycine, that of ionized hydroxylamine with acetic acid depends upon the synthesis of hydroxylamine which to date is only known to be possible in the solid or liquid state.

In theoretical work toward a gas phase formation route Maeda and Ohno [20] have searched the potential energy surface of glycine to identify a neutral chemistry reaction to glycine via ammonium ylide CH₂NH₃, an energetic isomer of methylamine CH₃NH₂, in reaction with CO₂. As ammonium ylide is in principle available via the dissociative recombination of protonated methylamine [20], and the latter may form via the radiative association of NH₃ and CH₃⁺ [21], [22], the elements of a truly gas phase route to glycine (and other amino acids) are becoming visible. Glycine has not been detected astronomically, but in view of the reactions we describe below, it would dimerize and pass into HPA fairly rapidly along with other amino acids, reducing the chance of its observation as a free molecule.

1.3 Gas phase polymerization of amino acids: background

We address the gas phase reactions of amino acids via ab-initio modeling and determine that peptide bonds can form for many amino acid pairs in warm dense cloud conditions, simply following a gas phase collision of two amino acids. The water molecule that is ejected during peptide bond formation remains bound to the nascent di-peptide, and this small molecular complex can grow by accretion, adding additional amino acids or small polar molecules, particularly water.

Our reference point for the study of early HPA was a study of hydrophobic polymer amide of biological origin [1] in which we observed entrapment and ordering of water in vesicles and tubes ranging from nanometers to microns, with movement of these structures demonstrated from 233–298 K. The biological polymer is typically 25 nm long as a stretched-out beta sheet of 75–81 amino acids. The National Center for Biotechnology Information (NCBI) web site [23] provides extensive data on this ancient protein. It is physically tough, some forms functioning in ocean vents at 383 K [24]. Biologically it acts as a rotor component in a 5 nm lipid membrane for the synthesis of cellular ATP. When reconstituted in hydrated systems with lipid it will form ion-conducting pores, even when the substrate is silicon [25].

For the ab initio calculation of amino acid collisional polymerization we mostly took pairs of amino acids from the ancient sequence rather than random pairs, based on the assumption that a polymer sequence that has existed essentially unchanged on Earth for 3.8 Gyr (Fig. 1) might reasonably be expected to have a prior history, particularly as it is only a factor of three to get to 12.7 Gyr ago. A contrasting polymer is polyethylene glycol, which does not appear to be able to condense in such conditions.

State 1

Separate amino acid equilibrium energies were computed using ab initio total energy/equilibrium geometry density functional calculations (B3LYP/6-31G).

State 2

The chosen pair of amino acids was manually assembled with the C-terminus of the first <0.4 nm from the N-terminus of the second. This manual initial molecular geometry was refined prior to any ab initio calculation by subjecting the pair to a molecular mechanics energy minimization involving the MMFF94 force field. The chosen conformer was then used in an ab initio calculation (B3LYP/6-31G) to obtain more accurately the energy of the second of the five states depicted on the abscissa of Figure 2, the amino acid pair in proximity. The graph was adjusted to set the sum of the separate amino acid energies to zero. Other energies are plotted relative to this.

State 3

A trial transition state in the “trans” configuration was then constructed from the amino acid pair, using Spartan input commands, in a “concerted” configuration, as described by Jensen et al. [30]. We took the amide bond to form via a single “concerted” transition state rather than via a “two-stage” process with two sequential transitions states. In [30] the authors explored these two pathways in detail and found that the concerted and two-stage processes have very similar energy barriers to amide bond formation, with a slightly lower energy to the two-step path. There has not been an experimental determination of the transition state path, to distinguish between these possibilities. Again, starting with the approximate transition state, the B3LYP/6-31G calculation was run to refine the transition state energy and atomic coordinates. These energies are depicted as the third state on the abscissa of Figure 2.

State 4

Under constraints that preserved the new C-N and O-H bonds that had formed, the system was energy-minimized using MMFF94, to obtain the dipeptide configuration with a water molecule that is usually attached via one or more hydrogen bonds to the dipeptide. Once again the energy and coordinates were refined using the B3LYP/6-31G calculation to obtain the values plotted as state 4 on the abscissa of Figure 2, the dipeptide with attached explicit water.

State 5

Lastly, separate calculations with B3LYP/6-31G were performed with an isolated water molecule and the isolated dipeptide, to generate the fifth state of Figure 2.

For the calculations on other polymers and poly amino acids the same general approach was taken.

4.1 Stability of amino acids and HPA

We compare the life of amino acids in space to the rate of their incorporation into polymer. Amino acids trapped in inert gas matrices at low temperature, simulating the free molecule, have been decomposed by short wave (100 – 200 nm) ultraviolet light from a hydrogen lamp [33]. Depending upon the supposed space environment the half-life of an amino acid to ultraviolet decomposition can vary from 300 yr in a diffuse interstellar medium (DISM) to 3×10⁷ yr in a dark cloud (DC) [33]. Above a certain density, amino acids will collide with each other on a much shorter timescale than this and polymerize into peptides.

The resistance of peptides to short wave ultraviolet light would be expected to be greater than individual amino acids because the energy of a photon absorbed into one bond is communicated throughout many more vibrational modes [34], which equilibrate collisionally, or radiatively [35] with the cool black-body spectrum of the cloud before energy has a chance to re-group and cause dissociation. Even with deep ultraviolet light the destruction efficiency is fairly low. Experimental data on Gly-Trp with 145 nm exposure [36] shows a destruction quantum efficiency of 1.3×10⁻². Longer wavelength irradiation of Gly-Gly at 206 nm [37] shows a quantum efficiency of 2.2×10⁻². This data is typically at room temperature, whereas even greater stability against photodecomposition would be expected at lower temperature, especially for longer peptides. Beta sheet peptide structures are very stable in UV light [38].

4.2 Polymerization Kinetics

The first stage in amino acid polymerization is the formation of a hydrogen-bonded collision complex (discussed above). In the collision of very small molecules the excess kinetic energy of approach has to rapidly be removed, either radiatively, or by a third body, for the di-molecular complex to be stabilized before the molecules simply fly apart. Larger molecules such as amino acids behave differently upon collision because the kinetic energy of approach is much less than the internal energy and is absorbed via distribution throughout internal rotational and vibrational modes. Consider the collision of molecules containing respectively M and N atoms in a gas at temperature T. The product complex will have 3(M+N−1) internal energy modes and the energy in these modes will equal the initial energy 3(M+N−1)kT/2 plus an increment corresponding to the average kinetic energy donated in the collision, 3 kT/2. The final internal energy per mode will be (M+N)kT/(2(M+N−1)), a relatively small fractional increase. For example, glycine (M = 10) colliding with alanine (N = 13) leads to an effective internal temperature increase by a factor of 23/22 = 1.045. This small excess in temperature does not cause dissociation after the complex has formed. The hydrogen bonding of amino acids in a “collision complex” therefore proceeds as a two-body process. At 100 K the bi-molecular gas kinetic reaction rate () is of the order of 2×10⁻¹¹ cm³ sec⁻¹ so that, for example, a fractional glycine density of 10⁻⁷, i.e. 1 cm⁻³, in a warm dense cloud could collide to form a hydrogen-bonded di-peptide complex in 1600 years, which is much shorter than the amino acid ultraviolet decomposition lifetime of 3×10⁷ yr in that environment [33]. The amino acids, once formed, therefore have a high probability of forming the hydrogen-bonded precursor to the dimer.

Moving on to the kinetics of polymerization, Figure 2 shows that the typical binding energy of the hydrogen-bonded di-peptide complex (the amino acid pair in proximity) is 60 kJ/mol (0.6 eV), which is shown below to be sufficiently deep to hold the complex together for 1million years at a temperature of 100 K. The excess energy of formation has already been lost on a timescale of 10⁴ seconds by collisions with surrounding thermal He atoms and H₂ molecules (radiative equilibration with the black body radiation field at the cloud temperature takes two orders of magnitude longer at 100 K). In general the dissociation rate, in the rapid energy exchange (thermalized) limit [35] is given bywhere k_D is the unimolecular dissociation rate constant, A is the Arrhenius pre-exponential constant, the activation energy of dissociation, k Boltzmann’s constant and T the temperature. The value of A will range from 10¹² sec⁻¹ for “tight” transition states to 10¹⁵ sec⁻¹ for “loose” transition states, as defined in [35].

Hydrogen-bonded dimer complexes form at the rate where is the collision cross section, the average particle velocity and the amino acid number density. If the system reaches equilibrium with respect to this reaction, the ratio of dimer to free amino acid iswhere is the dimer number density. As an example, at cm⁻³, and T = 100 K, the dimer to monomer ratio can range from 3×10⁵ to 3×10⁸ as the pre-exponential coefficient A varies from 10¹⁵ to 10¹². In these circumstances an amino acid would be overwhelmingly in the di-peptide hydrogen bonded complex.

The final step in polymerization involves passage through the transition state and expulsion of water. With the existence of a stable amino acid pair in proximity the system will progress to form a peptide bond if the transition state energy barrier is less than or of the same order as the typical 60 kJ/mol (0.6 eV) association energy calculated above. Of those studied here, Glu-Ala, Asp-Ile, Ser-Leu and Asn-Pro have transition barriers of less than 0.6 eV (Figure 1), but Lys-Phe, Phe-Gly, Ala-Ala, Gly-Gly and Ala-Gly have considerably higher barriers. In Fig. 2 it is seen that, once formed, the peptide-bonded pairs are stabilized by the continued association of the H₂O molecule released in bond formation. Although generally the water-stabilized system is at slightly higher energy than that of the initial amino acid pair in proximity, the increased entropy associated with the water will tend to stabilize the peptide bond against dissociation back into the hydrogen-bonded pair.

4.3 Formation of tri-peptides and higher polymers

Gas phase condensation of selected amino acids will proceed to tripeptides and higher in typical warm dense clouds. Each water of condensation will stick to the growing molecular system, thereby nucleating the first molecular assemblies of water. At low temperatures such an assembly resembles amorphous ice, with a hydrogen-bonded polypeptide around its surface. To the extent that polar side groups are plentiful, the water will be distributed around the peptide, but more hydrophobic polypeptides will push the water to one side, where it will coalesce with itself (Fig. 3). Other simple molecules such as CO, NH₃, CH₃OH and especially H₂O will collide frequently with the growing complex, finding ready hydrogen bonding to the collective water with the result that interstellar “ice” particles grow in association with the peptide nuclei. The possible evolution in dark interstellar clouds of water-cored peptide-skinned “vesicles” has therefore to be considered. The composition of these first beta sheet peptides will not be random, but will reflect the energetics of the very many pair-wise interactions of interstellar amino acids, many of which will not correspond to those currently present on Earth. Certain combinations will be greatly favored over others on account of the height of the transition state barrier.

4.4 Comparison with polyethylene glycol

Another candidate molecule that can hydrogen bond in a di-molecular complex and in principle polymerize producing water of condensation is ethylene glycol. However, we find that the hydrogen-bonded di-ethylene glycol pair has less binding energy than the corresponding amide pair (−40 kJ/mol vs −60 kJ/mol), and this gives dimer to monomer ratios (at 100 K and n_EG = 1 cm⁻³) ranging between 0.015 and 1.5×10⁻⁵. Moreover, the transition state barrier into the di-ethylene glycol O-C-C-O-C-C-O structure is 290 kJ/mol (Figure 5), ensuring that negligible amounts of polymer will form, in contrast to the amino acid case. We found that although poly-ethylene glycol (PEG) hydrogen bonds to the surface of a water cluster, it does not link transversely to another PEG molecule, and therefore no skin-forming structure, analogous to the peptide beta sheet, can form.

4.5 Outcomes

In summary, beginning approximately 12.7 Gya the Universe, with polymer amide, could have acquired a material that provided a surface and a solid density “platform” to host chemistry in a manner more favorable to many reactions than that of random diffusion in space. The beginnings of polymer amide complexity in this era may have been individual and essentially random, but a trend would be expected to favor polymerization of hydrophobic amino acids on the surface of the water accretion. Even in this era the temperature and available amino acid spectrum would have affected the composition of the polymer formed. Water would have been trapped in bulk form for the first time, in low density conditions where its nucleation would otherwise have been impossible. Such processes would precede poly-aromatic hydrocarbons (PAHs) as possible nuclei for interstellar “dust” [39] because PAH production is only thought to be significant in carbon rich red giant stars, and even there the production rate may not be sufficient to account for the apparent carbon particle accretion rates [40], [41]. Further, polymers involving rings are considered to require a catalyst and a surface to form, and before the first topology arose there was no surface. The composition of the first particles, or “dust” in the early universe is a subject of debate. There is evidence that the spectrum of early ultra-luminous infrared galaxies (ULIRGs) from the period at 1 Gy could be dominated by water emissions, and not mineral dust [42]. This would be consistent with an environment in which gas phase, rather than grain surface chemistry could dominate.

In more recent times accretion around a proto-star could involve HPA of H/C/N/O composition because that type of molecule has a structure to withstand harsh conditions in space and more importantly it is rich in hydrogen bonds for adhesion, the essential actuator of accretion. We are about to start analysis on meteorites that have fallen to Earth having accreted 5 billion years ago when the first elements of our solar system were starting to adhere to one another. HPA if present in such meteorites would not diffuse out based of the fact that rare earth elements in 5 billion years move by less than 700 nm [43]. In ultra-clean slices from the Chelyabinsk and Sutter’s Mill meteorites we will search via focused ion beam and mass spectrometry for polymer using techniques already applied to HPA analysis [1]. Mass spectroscopy has been performed on both meteorites [44], [45] yielding elemental data supportive of H/C/N/O content with amino acids detected in Sutter’s Mill samples. If rugged polymers rich in hydrogen bonds like HPA play an essential role at accretion, we should detect it in these meteorites.

The outer atmosphere of Earth is a “laboratory” environment in which the requisite H,C,N,O reside at low density and temperature. Although amino acids are known to be plentiful in the lower atmosphere [46], there are intriguing questions as to their possible rate of formation and/or polymerization in the upper atmosphere. If existent, this could have implications for water nucleation and precipitation.