Impact of temperature on the time required for the establishment of primordial biochemistry, and for the evolution of enzymes

Abstract

All reactions are accelerated by an increase in temperature, but the magnitude of that effect on very slow reactions does not seem to have been fully appreciated. The hydrolysis of polysaccharides, for example, is accelerated 190,000-fold when the temperature is raised from 25 to 100 °C, while the rate of hydrolysis of phosphate monoester dianions increases 10,300,000-fold. Moreover, the slowest reactions tend to be the most heat-sensitive. These tendencies collapse, by as many as five orders of magnitude, the time that would have been required for early chemical evolution in a warm environment. We propose, further, that if the catalytic effect of a “proto-enzyme”—like that of modern enzymes—were mainly enthalpic, then the resulting rate enhancement would have increased automatically as the environment became cooler. Several powerful nonenzymatic catalysts of very slow biological reactions, notably pyridoxal phosphate and the ceric ion, are shown to meet that criterion. Taken together, these findings greatly reduce the time that would have been required for early chemical evolution, countering the view that not enough time has passed for life to have evolved to its present level of complexity.

Whereas enzyme reactions ordinarily occur in a matter of milliseconds, the same reactions proceed with half-lives of hundreds, thousands, or millions of years in the absence of a catalyst (Fig. 1) (1). Yet life is believed to have taken hold within the first 25% of Earth’s history (2). How could cellular chemistry, and the enzymes that make life possible, have arisen so quickly? Here, we show that because of an extraordinarily sensitive relationship between temperature and the rates of very slow reactions, the time required for early evolution on a warm earth was very much shorter than it might appear. That sensitivity also suggests some likely properties of an evolvable catalyst, and a testable mechanism by which its ability to enhance rates might have been expected to increase as the environment cooled.

Fig. 1.

Half-lives (t_1/2) and first order rate constants (k) of some biological reactions proceeding spontaneously in the absence of a catalyst in neutral solution at 25 °C and 100 °C. For references, see (3).

Rapid substrate turnover is necessary to support the metabolism of an organism at the enzyme concentrations found in cells, but the same reactions, in the absence of enzymes, proceed vastly more slowly (Fig. 1). For example, the decarboxylation of orotidine 5′-phosphate (OMP), the final step in the biosynthesis of pyrimidines—and thus nucleic acids—proceeds with a half-life of 0.017 s at the active site of OMP decarboxylase. In neutral solution in the absence of the enzyme, the same reaction proceeds with a half-life of 78 million years (1). It is natural to ask how enzymes arose to meet so formidable a challenge.

The Time Required for Primordial Chemistry to Become Established

The rates of simple reactions, even if they are immeasurably slow at ordinary temperatures, can often be estimated by first determining their rates at elevated temperatures. Plots of the logarithm of the observed rate constants against the reciprocal of temperature (Arrhenius plots), which have been shown in some cases to be linear over many orders of magnitude (3), can then be used to estimate rate constants at ordinary temperatures by extrapolation. A reaction’s heat of activation (ΔH^‡, usually expressed in kcal/mol) is directly related to the negative slope of the Arrhenius plot, furnishing a direct measure of the sensitivity of the reaction rate to changing temperature (Eq. 1):

[1]

where k is the rate constant, R is the gas constant and T is the absolute temperature.

A familiar generalization—included in many textbooks and traceable to Harcourt’s pioneering experiments on the “iodine clock” reaction (4)—holds that the rates of chemical reactions in water usually double when the temperature is raised from 20° to 30 °C (ΔH^‡ = 12 kcal/mol), or 70-fold when temperature is raised from 25 to 100 °C.

Except for a few rapid reactions, that generalization is seriously misleading (Table 1). With the recent accumulation of observations on spontaneous reaction rates at elevated temperatures, it has become apparent that the sensitivity of most uncatalyzed reactions to temperature is actually much more pronounced than that of the iodine clock. For example, the rate of uncatalyzed amide, peptide and urea hydrolysis (t_1/2 ∼ 500 y at 25 °C, ΔH^‡ = 23 kcal/mol), increases ∼3,000-fold when the temperature is raised from 25 to 100 °C (5). The rate of hydrolysis of O-glycoside bonds [t_1/2 ∼ 18 × 10⁶ y at 25 °C, ΔH^‡ = 35 kcal/mol (6)] increases ∼190,000-fold when the temperature is raised from 25 to 100 °C; while the hydrolysis of aliphatic phosphate monoester dianions [t_1/2 ∼ 1.1 × 10¹² y at 25 °C, ΔH^‡ = 47 kcal/mol (7)] is accelerated ∼10,000,000-fold when the temperature is raised from 25 to 100 °C.

Table 1.

Temperature effects on the rates of biological reactions in the absence of a catalyst, compared with temperature effects on k_cat for a typical enzyme reaction (in italics, from Table 4)

	t1/2 (25 °C)	ΔH‡	k100°/k25°	k0°/k25°
Phosphomonoester hydrolysis (7)	1.1 × 1012 yr	47.0	10,300,000	0.0006
Uroporphyrinogen decarboxylation (21)	2.3 × 109 yr	41.2	1,440,000	0.0016
Amino acid decarboxylation (14)	1.1 × 109 yr	38.4	560,000	0.0024
Anhydrocellobiitol hydrolysis (6)	1.1 × 108 yr	35.2	190,000	0.0040
Phosphodiester hydrolysis (22)	3.1 × 107 yr	29.5	27,000	0.0096
Fumarate hydration (23)	7 × 105 yr	28.9	22,000	0.0105
Chloroacrylate dehalogenation (24)	1.0 × 103 yr	26.7	10,500	0.015
Peptide hydrolysis(internal) (25)	600 yr	22.9	2,900	0.027
Mg-ATP + glucose, 1 M (19)	6 yr	25.6	2,600	0.019
Bicarbonate dehydration (26)	27 sec	16.2	220	0.075
Typical enzyme reaction (Table 4)		12.2	70	0.14

In addition to the startling magnitudes of these rate accelerations, Table 1 shows that the slowest reactions are most sensitive to temperature. These combined tendencies result is a marked “leveling” of rates, so that even very slow reactions become measurable at temperatures not far removed from the boiling point of water (vertical arrows, Fig. 1). It is therefore of special interest that the earliest branches of the tree of life—as indicated by analysis of molecular sequences and structures—are largely populated by thermophilic organisms (8). Even now, temperatures near 100° are present near submarine vents and in hot springs that support organisms of widely differing phylotypes (9).

If life originated under these conditions, an idea that has been widely [(10) and references cited therein], but not universally (11), accepted, then the time required for prebiotic chemistry to become established would have been greatly abbreviated (by as many as five orders of magnitude at 100 °C) compared with the time required at 25 °C. Conversely, primordial chemistry would have been far more difficult at 0° than at 25 °C (Table 1, final column), aggravating the problem of understanding how chemical evolution could have occurred within a reasonable period of time.

These observations lend credence to the view that life originated on a warm earth, and they counter the view that insufficient time has passed for life to have evolved to its present level of complexity.

Evolution of Proto-Enzymes: a Hypothesis

In a warm environment in which these chemical transformations were proceeding without great difficulty, even a modest catalyst would have produced useful rates of reaction. In a cooler environment, however, many reactions would have become so sluggish that they would have required the action of much more powerful catalysts. For a reaction such as the decarboxylation of OMP, with a t_1/2 of 10⁸ years at 25 °C (1), a proto-enzyme that reduced that half-life by a factor of 10³, 10⁶, or even 10⁹, would have conferred no significant competitive advantage on the host organism. How, then, did cells develop the ability to produce useful rates of reaction at ordinary temperatures?

Suppose that a small molecule (perhaps a metal ion or an organic acid or base) was once present in a warm aqueous environment where it produced a small but significant acceleration of a particular reaction. In principle, that increase in rate might have been achieved either by lowering ΔH^‡ or by raising the value of TΔS^‡ (the entropy of activation) (Fig. 2) (for thermodynamic definitions, see SI Appendix). If that primitive catalyst enhanced the rate of reaction entirely by raising the value of TΔS^‡, then the rate enhancement it produced would have been unaffected by changing temperature (Fig. 2A). But if that catalyst enhanced the reaction rate by lowering ΔH^‡, then the rate enhancement that it produced would have increased automatically as the surroundings cooled (Fig. 2B), quite apart from any selective pressures that might also have been at work.

Fig. 2.

Expected variation with temperature of the rate enhancement (k_cat/K_m) produced by two catalysts, acting (A) by enhancing a reaction’s entropy of activation, or (B) by reducing a reaction’s enthalpy of activation.

The plausibility of this latter scenario for enzyme evolution depends on whether primitive nonbiological catalysts tended to lower activation barriers by reducing the value of ΔH^‡, or by raising the value of TΔS^‡. To address that question, we turned to the behavior of nonenzymatic catalysts acting in aqueous solution.

Do Primitive Catalysts Act by Lowering ΔH^‡? A Test

Substantial rate accelerations have been reported for a variety of nonenzymatic catalysts acting in water (including metal ions, simple organic molecules, antibodies, and other proteins) (Table 2), but there appears to be relatively little information about the temperature dependence of these catalytic effects. We decided to examine the action of two molecules catalyzing particularly difficult reactions, phosphate ester hydrolysis (the Ce^IV ion) and amino acid decarboxylation (pyridoxal phosphate, or PLP). By comparing the temperature dependence of k_cat for the catalytic complex with that of k_non for the substrate alone, we sought to determine whether these nonenzymatic catalysts act by reducing ΔH^‡ or by enhancing TΔS^‡. The results, reported below, were unequivocal.

Table 2.

Estimated rate enhancements produced by nonenzymatic catalysts acting in water

catalyst (or solvent)	reaction	kcat/knon
Transfer to vapor phase (27)	OH- + CH3Br	1.3 × 1018
CeIV (28)	H2O + CH3OPOO-2	1.0 × 1014
Transfer to cyclohexane (solvent effect) (29)	H2O + CH3OPOO-2	2.5 × 1012
Pyridoxal phosphate (PLP) (28)	2-aminoisobutyric acid (AIB) decarboxylation	2.3 × 109
HMPA (solvent effect) (30)	Kemp elimination	9 × 107
Metal complexation (cis-) (25, 31)	OH- + gly-gly	9 × 107
Computer-aided enzyme design (32)	Kemp elimination	1.2 × 106
Catalytic antibody (33)	aldol cleavage	3.6 × 105

The hydrolysis of phosphate esters is strongly catalyzed by soluble complexes of dinuclear Ce^IV (12, 13)—the most abundant lanthanide in the earth’s crust and the only lanthanide that is stable in the +4 oxidation state. The hydrolysis of unactivated phosphate monoesters—a common event in metabolism and cell signaling—is intrinsically one of the slowest biochemical reactions known (estimated t_1/2 = 1.1 × 10¹² y at 25 °C) (7). When we examined the temperature dependence of the effect of Ce^IV on the hydrolysis of two phosphate monoesters (SI Appendix), the results were unequivocal. The rate enhancement by Ce^IV, estimated by comparing the rate constant for the catalyzed reaction (k_cat) with the rate constant for the uncatalyzed reaction (k_non), was achieved entirely by reducing the enthalpy of activation, while the effect of Ce^IV on TΔS^‡ was actually unfavorable (Table 3).

Table 3.

Effects of nonenzymatic catalysts on thermodynamics of activation (kcal/mol, where ΔG^‡ = ΔH^‡ - TΔS^‡)

	k (s-1) (25 °C)	ΔG‡ (25 °C)	ΔH‡	TΔS‡
Phosphate ester hydrolysis(1)
4-npP-2	3.4 × 10-9	29.0	27.6	−1.4
4-npP∶CeIV	5.2 × 10-5	23.2	12.1	−11.1
Rate enhancement	1.5 × 104	−5.8	−15.5	−9.7
(2)
MeP-2	2 × 10-20	44.3	47.0	2.7
MeP-2∶CeIV	1.9 × 10-6	25.2	6.3	−18.9
Rate enhancement	9.6 × 1013	−19.1	-40.7	-21.6
Amino acid decarboxylation
(3)
AIB	1.1 × 10-18	41.8	39.9	−1.9
AIB∶PLP	2.6 × 10-9	29.1	21.3	−7.8
Rate enhancement	4.4 x 1011	−12.7	-18.6	-5.9

In each case, the effect of the catalyst on ΔH^‡ is favorable, but its effect on TΔS^‡ is unfavorable. For experimental details, see SI Appendix.

The decarboxylation of amino acids, another reaction that is intrinsically very slow [t_1/2 ∼ 10⁹ y in neutral solution at 25 °C (14)] is catalyzed by enzymes that ordinarily use PLP as a cofactor. PLP by itself is known to be an effective catalyst of this reaction, accounting for a major fraction of the rate enhancement produced by the enzyme (15). Examining the temperature dependence of the PLP rate enhancement of the decarboxylation of 2-aminoisobutyrate (SI Appendix), we observed once again that catalysis arises entirely from a reduction in the enthalpy of activation for the catalyst-substrate complex (Table 2).

These findings are not unprecedented. Enthalpic effects have also been shown to predominate in systems in which catalysis is much less pronounced, including general base catalysis of the bromination of acetoacetate by glycolate (16), in the covalent catalysis of the hydrolysis of 4-nitrophenyl acetate by imidazole (17), in general base catalysis of the aminolysis of carboxylic esters by alkylamines (18), and in the Mg^II-catalyzed methanolysis of ATP (19).

It is interesting to consider the properties of present-day enzymes in the light of these findings. The final entry at the bottom of Table 1 indicates the average behavior of k_cat for 12 different enzymes shown in Table 4. Despite differences in mechanism, the heats of activation are remarkably similar to each other, with ΔH^‡ ∼ 12 kcal/mol. If ΔH^‡ values for k_cat of these highly evolved enzyme reactions are compared with ΔH^‡ values for the uncatalyzed reactions in Table 1, it is evident that all these enzymes act by lowering the value of ΔH^‡.* Thus, the corresponding increase in affinity, as the enzyme-substrate complex in such reactions passes from the ground state to the transition state, is mainly enthalpic in origin, consistent with chemical mechanisms that typically involve the formation of new electrostatic and hydrogen bonds in the transition state (20).

Table 4.

Values of ΔH^‡ for some biological reactions in the presence (k_cat) and absence (k_non) of an enzyme at neutral pH (literature references in parentheses)

Bond cleaved	Enzyme	ΔH‡ (kcat)	ΔH‡ (knon)
C─C	OMP decarboxylase	11.0 (34)	37.1 (1)
C─N	Dihydroorotase	12.3 (35)	24.7 (35)
	Cytidine deaminase	14.9 (36)	22.1 (36)
	Urease	9.9 (37)	22.9 (5)
	Trypsin + casein	12.0 (38)	22.9 (25)
C─O	Invertase	11.5 (39)	27.3 (40)
P─O	Alkaline phosphatase	12.7 (41)	47.0 (7)
	Staphylococcal nuclease	10.8 (42)	29.5 (22)
	Hexokinase	11.6 (19)	12.8 (19)
S─O	Alkyl sulfatase	12.4 (43)	24.6 (44)
C─H	Mandelate racemase	15.4 (45)	31.9 (46)
	Fumarase	13.9 (47)	28.9 (48)
C─Cl	Chloroacrylate dehalogenase	9.4 (24)	26.7 (24)
	average	12.2	28.0

From an evolutionary standpoint, it is unlikely that the common enthalpy-lowering effect of present-day enzymes is fortuitous. As the environment cooled, a primitive catalyst that reduced ΔH^‡ would have offered a selective advantage over a catalyst that raised TΔS^‡ by an equivalent amount (Fig. 2). An entropic catalyst, in contrast, would have been useful only for reactions with intrinsically low enthalpic barriers in which a large rate enhancement is unnecessary and substrate specificity is of special importance.^† We propose that enthalpy-lowering mechanisms became common because they are so temperature-dependent; and because there is almost no limit—at least in principle—to the benefit that might arise from the action of a purely “enthalpic” catalyst. Natural selection has presumably resulted in the evolution of enzymes toward greater catalytic power and specificity, but at the most fundamental level—the thermodynamics of substrate activation—PLP and the Ce^IV ion resemble modern-day enzymes and furnish plausible models for their evolution.