Tải bản đầy đủ - 0 (trang)
1 Introduction: Nature as a Complex System

1 Introduction: Nature as a Complex System

Tải bản đầy đủ - 0trang



(chains of automobile movements), stock markets (chains of mutually beneficial

financial transactions), and family trees (chains of procreation events), to name

but a few. The common feature in all of these systems is that they change their

state based on the accumulated action–reaction interactions over time, between

the individual elements in the system. Each individual action–reaction event is

typically relatively simple and easily understood, but the accumulation of them is

extraordinarily complicated.

In a sense, complex systems can be envisioned to be like intersecting lines of

dominos. When each individual domino in any one line falls, we can know with

near certainty that the next one will be knocked over and will also fall. We can

further know how the following one will fall and when it will fall. But, when two

lines of falling dominos intersect each other, it becomes more difficult to know what

will happen at the intersection points. One line may prematurely set off the other

line. Alternatively, the two lines may mutually halt each other. A further possibility

is that the two lines will pass through each other without affecting each other at

all. Several other options may arise. Now multiply that uncertainty by hundreds, or

thousands, or millions of intersecting lines of dominos, and it becomes essentially

impossible to predict the overall state of the system with any degree of certainty

even over short intervals into the future.

It was, perhaps, weather forecasting that gave the greatest impetus to the foundation of complex systems science. By the 1970s, it was starting to dawn on scientists

that no amount of mathematical modeling could accurately predict weather for more

than a period of a few days. This was because weather events display an extreme

sensitivity to their initial conditions, an effect commonly known as the butterfly

effect, after a paper presented by Edward Lorenz to the AAAS in Washington

D.C. on December 29, 1979, entitled “Predictability: Does the Flap of a Butterfly’s

Wings in Brazil Set off a Tornado in Texas?”

The problem is that weather is comprised, effectively, of innumerable, interacting, highly localized chains of action–reaction weather events. While each

individual event and the chain it sets off (e.g., warming of a particular location

by the sun) can be modeled in not inconsiderable detail, it is more difficult to

model and predict how two local chains interact with and influence each other. As

more such local chains are added to the model, it very quickly and nonlinearly

becomes an increasingly impossible task.

One of the reasons is that complex systems display a property known as amplification. When there are many chains present and interacting with each other, a small

event in even one of the remote chains can quickly be transmitted and amplified

through the other chains to create a powerful overall effect somewhere entirely

different. Thus, the butterfly flapping its wings in Brazil could potentially set off a

tornado in Texas. We will discuss the origin and character of amplification in later


Biological systems and life itself have a very similar character. They comprise

of innumerable biochemical entities interacting with each other in action–reaction

chains to thereby create something that is more than the simple sum of their parts.

Most of the individual interactions are, in themself, not overly complicated. The



individual chain that is created or changed by the interaction is similarly amenable

to understanding. However, when two or more chains interact and influence each

other, then things quickly get complicated. When many, many chains are present

and interacting, then the system rapidly becomes an impenetrable maze. For this

reason, complex systems have also been termed chaotic (i.e., described by chaos

theory). In a sense they can also be considered to be cascade processes; that is,

the system cascades along a seemingly inevitable path driven by the timing and

character of the individual action–reaction events between the system elements.

Biology has recently started examining these webs of biochemical interactions

that comprise life, in a subdiscipline known as systems biology. The field that draws

inspiration from such biological systems has therefore come to be called systems

chemistry. It is around this concept that efforts devoted to biomimicry and bioinspiration involving the concept of life itself have coalesced. A new journal, Journal

of Systems Chemistry, was established recently to cater to studies of this type.1



Complex systems typically display a range of common features that arise out

of their unique character, involving, as it does, interacting, localized, chains of

action–reaction. We do not discuss all of these common features here, but will

mention a few of the most pertinent. These include:






Autonomy/autonomous agents

Nonequilibrium processes

Time and path dependence

A key feature of complex systems is that they often systematically expand to fill

the space allotted to them. For example, what we call biology has expanded from

a prebiotic primordial state to, effectively, cover the Earth. A threatening virus that

arises in some very particular location will generally, over time, tend to expand its

reach to more and more locations.

This property is very different from what is common for human-made creations.

Our buildings, machines, and the like inevitably get old, degrade, become impossible to maintain, and ultimately disappear. In a sense, extinction is inherent in

human-made devices. But Nature’s way of making things is different. Nature makes

things, which make new things, and so on. This “machines-making-machines” property of biology is called self-replication, and it involves the incorporation within



the system of a capacity for renewal. Such a capacity is absent in human-made


Thus, the DNA molecule of biology is unique in that it replicates itself. It does

so over and over, creating the outwardly moving cascade that we call biology.

Self-replication is likely a form of the property of amplification noted earlier. Somehow, when the individual elements in a complex system engage in

action–reaction events, they can end up creating something that is more than the

simple sum of their parts. How they do that is a key topic in systems biology.

Studying, understanding, and creating amplication and self-replication is similarly

a central theme of systems chemistry.1

In other complex systems, amplification often seems to arise out of individual

events occurring in a synchronous fashion. That is, the timing of the individual

events tends to make the events mutually reinforce one another, to thereby achieve

a larger overall effect than might otherwise have been the case. For example,

extreme weather tends to derive from synchronous confluences of a collection

of more moderate weather events that end up combining to creating a “perfect

storm.” There seems little doubt that amplification in biological systems and life

itself has similar origins. In later sections in this chapter, we discuss work aimed

at elucidating amplification effects in chemistry.

We should note here also that, when we speak of amplification, we refer not

only to unexpected, nonlinear enlargement effects, but also to nonlinear contraction

effects. Nature can shrink back just as rapidly and vigorously as it can expand.

Emergence is the phenomenon where amplification leads to new patterns of

behavior by the system. These patterns, which have typically not existed before,

emerge spontaneously from the background. Emergence commonly derives from

another property of complex systems: evolution.

Because they are so dynamic, complex systems tend to undergo continuous

change in response to changes in the external conditions. In effect, they adapt to

the changing environment. This process of adaption is known as evolution and one

of its physical manifestations is emergence. Understanding emergence is a key topic

in systems chemistry and biology research.1 For example, how did Nature come to

adopt a left-handed chirality in all of its amino acids? Can we, through chemistry,

come to better understand that process or maybe even simulate it? Understanding,

or better still, creating and harnessing systems that evolve is a prime target of

systems chemistry.

Feedback is the process where an outcome in one part of the system dramatically

influences—amplifies or diminishes—a process that takes place in another part of

the system. Thus, one typically speaks of a feedback loop, where a system generates something that dramatically influences the system that has created it. Examples

include the allosteric responses of biology discussed in an earlier chapter. Another

famous example is that of the bacterial growth curve of an isolated population

of microbes on a new, but static food source.2a Initially, the microbial population grows exponentially as there is an apparently unlimited source of food (the

“exponential phase”). But then, once the bacterial population has exploded, the

population first stabilizes (the “stationary phase”) and then declines in an extreme,



nonlinear way as the decreasing food availability and the toxic products generated by the microbes during their consumption of the food create a vigorously

destructive feedback loop (the “death phase”).

Because of the amplification involved, feedback effects are also of interest in

systems chemistry. How do they work? Can we create such effects?

Many complex systems are known to contain within them mini, self-contained

systems that display elements of complexity. These “systems-within-systems” may

act totally autonomously from the rest of the system, giving them the property of

autonomy. Some researchers also speak of autonomous agents.

An example of an autonomous agent is you. Your body is a complex system

that exists autonomously within Nature, which is also a complex system, albeit

much larger. Indeed, Nature contains many autonomous agents, including all of

the microbes, organisms, and other known biological entities.

But the concept of autonomy does not extend only to living things, it goes

much further than that. For example, while you are an autonomous agent, there

are subsystems within your body that are themselves autonomous. These systems

go all the way down to the molecular level. For example, it could be argued that

proteins are autonomous agents in the way that they fold. Protein folding has

been and is the subject of a large volume of research. The problem that arises

is that protein folding is a relatively quick process, occurring on a time scale of

seconds. However, studies show that if a protein were to search for its global energy

minimum by sampling every conformation available to it, it would typically take

innumerable years to fold.2b How, then, do proteins fold so quickly? This question

is known as the Levinthal paradox 2b and the answer may lie in systems chemistry.

Proteins likely fold by a cascade process in which each local conformational change

sets off another local conformational change and so on, until there are no more

conformational changes available. At that stage, the protein is fully folded. In other

words, the individual residues in the protein act autonomously and interact with the

other residues in a series of action–reaction events to set up a one-way sequence

of conformational changes that cause the protein to fold.

This explanation, if it is generally correct, illustrates another more microscopic

feature of complex systems; namely, an apparent prevalence of nonequilibrium (or

“one-way”) processes at the most basic, fundamental level of the system. Many

(although certainly not all) complex systems seem to display a “one-way” character

at their most elemental level. There is often no searching for, nor achieving, an

equilibrium. For example, the complex system of a family tree propagates exclusively in a one-way manner; you cannot unmake the birth of a child or obtain a

new child from someone already dead. This is not to say that all action–reaction

steps in complex systems involve nonequilibrium processes. Some may occur as

equilibrium processes under thermodynamic control. In a later section we describe

examples of complex systems driven by equilibrium and nonequilibrium molecular


The apparent prevalence of nonequilibrium processes has much to do with the

fact that in many complex systems the individual action–reaction events are pathway and time dependent. As noted above, timing and synchronicity appear to play a



large part in the phenomenon of amplification. But timing can only ever be important if pathway is also important. In other words, an element can only ever be at

the right place at the right time. It cannot be only at the right time.

Understanding the fundamental character of nonequilibrium processes, as well

as path and time dependence in chemical reactions and processes, is another, highly

fundamental avenue of research in systems chemistry. The field of chemistry has a

well-developed and rich foundation in chemical reactions and processes driven by

their thermodynamics under equilibrium conditions. Our understanding of chemical

processes under nonequilibrium conditions is far less developed.3

We should note here too that nonequilibrium processes are sometimes termed

kinetic processes; that is, processes driven by the kinetics of reaction, not by its

thermodynamics. In the case of chemical reactions, an alternative terminology

is mechanical processes; that is, processes driven by the mechanics of reactant

collisions.3 Other terms may also be used. In the sections that follow, we use the

original terms employed by the researchers involved. Where applicable, we have

added the expression nonequilibrium for improved clarity.



In the following sections we discuss several examples of research in systems chemistry. This research is mainly drawn from recent scientific publications in chemistry

journals and should be considered to offer only a representative slice of this new



Self-Replication, Amplification, and Feedback

Many areas of interest in systems chemistry are developed around the concepts of

self-replication and amplification. A possible method of self-replication in chemistry and biology is autocatalysis, or “making molecules that make themselves.”4

The two primary methods employed to foster autocatalysis are:5

1. Using the reaction product as a templating molecule that goes on to facilitate

the synthesis of further reactants (Figure 15.1a).

2. Cross-catalytic systems in which products of multiple reactions catalyze the

counterpart reactions (Figure 15.1b).

In cross-catalyzed systems, two reaction trajectories produce separate products

A and B, with product A then catalyzing the formation of B and vice versa.

Figure 15.2 depicts a reaction scheme exhibiting both autocatalytic and crosscatalytic mechanisms.5

Autocatalytic reactions are of fundamental interest. They potentially also have

implications for our understanding of the origin of life and its molecular building

blocks. We discuss these topics in a later section.















= αcP












Figure 15.1 Schematic illustrations by von Kiedrowski of two common mechanisms

for self-replication: (a) representation of template directed self-replication, and (b) representation of cross-catalytic self-replication. (Reprinted as an open-source image from

ARKIVOC: Ref. 5.) Self-Replication While many self-replicating systems employ autocatalytic reactions using one of the above models, there are exceptions including

self-replicating systems promoted by product-induced solvation of reactants.6 One

example of such a system involves lipophilic amine and hydrophilic aldehydes

reversibly generating amphiphilic imines that aggregate and form micelles. The

micelles increase the concentration of the dissolved lipophilic reactant allowing for

a product-induced increase in reactants. In its use of a feedback loop, this system

was said to offer a chemical model of homeostasis.7

In further work, multiple variations of lipophilic amines were introduced into the

system allowing for self-replicating chemical “selection” of the “fittest” products.6

One example in this respect involved the use of a mix of both templated autocatalysis and cross-catalysis for the condensation of 3-aminobenzamidine and phenoxy

acids.5 By varying aromatic substituent groups of the reactants, the mechanism of

autocatalysis could be selected (Figure 15.2). With NO2 substituted phenoxy acids,

the product exhibited the classic square-root dependence of autocatalytic systems

in which the templates interact to form a stable complex inhibiting the reaction.





10 R1 = t-Bu

13 R1 = Me




















10 R1 = t-Bu

13 R1 = Me

H2 N


























12 R3 = t-Bu

14 R3 = Me



12 R2 = NO2

15 R2 = H

Figure 15.2 Reaction scheme exhibiting both autocatalytic and cross-catalytic mechanism. (Reprinted with permission. Copyright Wiley-VCH: Ref. 4.)

When the NO2 was replaced with H, the resulting template displayed first order

dependence, indicating a cross-catalytic system in which the template formed a less

stable complex with itself. Amplification: Asymmetric Autocatalysis Another important

concept in autocatalytic systems chemistry is the potential for enantioselectivity

in the self-replication of chiral products, or asymmetric autocatalysis. One early

system focused on the addition of dialkylzinc to pyridine-3-carbaldehyde to which

a 20 mol % enantiomeric excess (ee) of chiral product was added. Seeding with

chiral product had the effect of autocatalytically generating enantioselectivity,

which matched the seeded product in an otherwise racemic reaction.8

This work was later extended to produce consecutive autocatalytic reactions

capable of producing almost enantiomerically pure product upon addition of minimal amounts (<0.0001% ee) of enantio-enriched product. Through the use of

consecutive reactions, the enantioselectivity imparted by a slight excess was, effectively, amplified to produce enantiomerically pure, chiral pyrimidyl alkanol.

In a first example of organocatalytic asymmetric autocatalysis, autoamplification of enantiomeric excess was observed for the Mannich reaction of acetone

and an iminoethyl gloxylate when a small amount of chiral product was added

(Figure 15.3).9 Interestingly, both chiral products can be selected for, depending

only on the choice of initial product chirality. Comparable unseeded reactions

produced racemic mixtures.

Further work studying the origin of homochirality revealed that stochastic fluctuations in the distribution of enantiomers may account for significant enantio

enrichment of products, generating, at random, an excess of (S) or (R) form.10

While many efforts in asymmetric autocatalysis have focused on autoamplification

of existing chiralty, this work provided insight into the origin of chirality and its















(30 mol%, 99% ee (R))


4 days, RT


40% yield[a], 96% ee (R)[29]











4 days, RT




(15 mol%, 98% ee (S))



48% yield , 85% ee (S)[29]


Yields after subtraction of the initially added product catalyst










without product catalyst

4 days, RT


31% yield, 9.4% ee (S)[30]

Figure 15.3 The first asymmetric organoautocatalytic system: Mannich reaction of acetone with N -PMP-protected α-imino ethyl glyoxylate. (Reprinted as an open-source image

from J. Syst. Chem.: Ref. 9.)

ability to generate products with significant enantiomeric excess based on natural

stochastic variations in initial conditions.

The biological context of symmetry-breaking reactions has been studied by

Eschenmoser and colleagues,11 who demonstrated that copolymerization of heterochiral pyranosyl tetramers into pyranosyl-RNA resulted in a ligation process that

was highly stereoselective. Shortly thereafter, Feringa and van Delden highlighted

the origin of chirality in relation to several stereoselective systems.12

While broad in scope, the field of autocatalysis and self-replication has historically focused on the implications for our understanding of the origin of life;

in particular, the generation and replication of the information encoded RNA and

DNA molecules necessary for the propagation of life.

Autocatalytic replication of nucleic acids was first reported using complementary

trimers linking to form a templating hexamer.13 A simpler system was later discovered in which single complementary adenosine base pairs were autocatalyzed by the

resulting product.14, 15 In elaboration of this work, an experiment was devised in



which a bimolecular reaction (A + B) was used to generate a product (A + B → T),

where the product (T) then templated and catalyzed further synthesis.16 Selfreplication of larger peptide chains from fragments was also demonstrated using a

similar bimolecular templating approach.17 Using cross-catalytic systems, 10–23

deoxyribozymes were activated from inactive precursors using complementary pairs

of cyclized deoxyribozymes to produce their autocatalytic linear versions.18


Emergence, Evolution, and the Origin of Life

While not necessarily autocatalytic, efforts in systems chemistry have been

made to generate fundamental biological molecules from prebiotically plausible

conditions. Synthesis of pyrimidine ribonucleotides has been shown from simple

precursers—cyanamide, cyanoacetylene, glycolaldehyde, glyceraldehyde, and

inorganic phosphate.19

As mentioned, studies of autocatalysis are often directed at elucidating the origin

of life and its molecular building blocks. Enzymes are nature’s agents for replication

of complex and simple biomolecules. How, though, were the first organisms able

to replicate the molecules (nucleotides) with which enzymes are encoded?

Both Pross20 and Eschenmoser21 have proposed that replication and metabolism

developed concurrently not separately. Regardless of the order in which they came

about, it remains true that while racemization is the most thermodynamically stable

state of a heterochiral mixture, there are many systems, including enzymes, that

perform processes with often remarkably high stereoselectivity.

The level to which Darwin’s ideas on evolution influenced subsequent studies

is rarely seen in any field of experimental science. But while we know how complex life is, the origin of life—that is, the transfer of inanimate building blocks

into animate organisms—remains unexplained. Many systems chemists are seeking to explain this transition using models. These models attempt to study complex

molecular networks and their phenomena, self-reproducing, self-organizing, and

autocatalytic systems with a basis of the origin of life, and applications to combinatorial and symmetry-breaking systems.1 We discuss some examples here.

One model in this respect, by Kuhn,22 considers probability estimates that a

homochiral oligomer of adequate length will form from a heterochiral pool and

successfully replicate, given that the physical interactions rely on chance. Citing

information theory, this probability was interpreted with respect to a “knowledge”

term, which indicated how much information is required to form the replicating

oligomer. Knowledge increases with time due to growth in chain lengths or if

environmental conditions either kill the oligomer or force it to adapt otherwise,

via stabilization effects. The model, which was applied to the hairpin structures of

RNA and DNA, suggested that chemical entities that successfully replicated and

adapted were essential to the formation of life.

In another model,20 Pross refers to the work performed by Joyce and colleagues,23 which showed that two RNA enzymes could not coexist over time in

the presence of a substrate that was essential to both. Pross20 argued that while

there is value in being able to monitor “survival of the fittest” experimentally,



the appearance of competitive behavior should not be considered the important

observation. While it could be interpreted that the more adaptive enzyme drove

the lesser into extinction, the observation is not so much that “molecules exhibit

Darwinian behavior” but rather that “biology can be explained by chemistry.” That

is, the actions and evolutionary paths of organisms should be attributed to actions

on the atomic scale, and not vice versa. The behavior seen in the RNA enzymes

is therefore not analogous to evolution but rather a potential cause of evolution.

Pross proposed, in effect, that evolution is based on increasing “dynamic kinetic

stability.” By this he means that the system as a whole moves to a state of kinetic

(nonequilibrium) stability. It does so independent of thermodynamic barriers experienced by its individual components.

An example of such a system in Nature is photosystem II, the universal protein

that contains the active sites of photosynthesis, which turns over and rebuilds

itself approximately every 30 minutes. The analogy to cell division and death,

regardless of the complexity of the organism, follows similarly. Thus, knowledge of

the conditions that maintain this form of autocatalysis can, if repeated in laboratory

settings, yield elegant chemistry. By understanding how complex systems work,

coupled with a background in evolutionary study, we may therefore not only learn

from, but also be able to ultimately mimic evolution, at least in principle.

15.3.3 Autonomy and Autonomous Agents: Examples of Equilibrium

and Nonequilibrium Systems Equilibrium Systems: Dynamic Combinatorial Libraries (DCLs)

Dynamic combinatorial chemistry involves the creation of a dynamic combinatorial library (DCL) under thermodynamic control and at equilibrium.25 DCLs

constitute an important basis of systems chemistry in the field of screening and

sensing.16 Where standard screening techniques involve the interaction of one sensor compound with its corresponding molecule, systems chemistry involves the

interaction of three or more molecules to form a unique complex identifiable by

simple spectroscopy. Creating understandable results from complex mixtures is

one of the hallmarks of systems chemistry and lends itself to screening large numbers of molecules simultaneously in, for example, mass screenings and sensor

applications.16, 24, 26

To illustrate the concept of a DCL, consider the following example. In a system

where reactants A, B, and C are present, along with a template T, which favors

binding trimers, complexity quickly becomes evident. The binding constant of T to

each of the monomer and dimer combinations must be considered, as well as the

trimer combinations. The trimer combinations alone are affected by concentration,

which dictates that for A3 T (Figure 15.4) to accumulate, it must be a considerably

better binder than ABCT if A, B, and C are in equal concentrations.27 The possibility of larger oligomers must also be factored even if they do not bind to the

template, in order for the proper concentrations of oligomers to be modeled. All

these factors and more must be considered when examining such DCL systems

chemistry. It can lead to nonintuitive results.












Dynamic mixture











Model 1













Figure 15.4 A simple interaction model for a DCL containing only one building block,

A, and a template,T. (Reprinted with permission. Copyright Wiley-VCH: Ref. 27.)

In this field, the relation between biology and systems chemistry is clear. Classical chemistry would dictate that each oligomer be isolated and tested against T to

find its unique binding constant, slowly building up a web of interactions. Isolation

and step-by-step examination is impossible in biology, so methods were developed

to examine systems as a whole.

A proof-of-principle study showing how DCLs are capable of sensing different molecules in a complex mixture was reported by Severin and co-workers

(Figure 15.5).27, 28 A solution, containing two metal ions and three coordinating dye molecules, was created as an initial DCL. The UV–visible spectrum of

this solution served as the baseline of the system. Addition of a mixture of dipeptides to the baseline DCL resulted in a perturbation of the metal–dye interactions,

resulting in unique UV–Vis shifts. This complex system, containing three dye

molecules and two metals, allowed for the distinction of closely related dipeptides

(Figure 15.6). In a similar study, a sensor was developed to distinguish Gly-Gly-His

from His-Gly-Gly and Gly-His-Gly.29 Nonequilibrium Systems Moving beyond thermodynamic control,

similar principles can be used to create kinetically controlled, or nonequilibrium

networks. Such systems are of interest because, in both chemistry and biology, what

reacts fastest is often just as important as what binds strongest. Two examples of

such nonequilibrium DCL-like systems are discussed briefly here.

In the thermodynamically controlled examples above, the selectivity allowed by

the system was rarely higher than that provided by the affinity between the template and the binding molecule. In nonequilibrium DCL-like systems, it becomes

possible to enhance the selectivity as demonstrated by Gleason, Kazlauskas, and

Tài liệu bạn tìm kiếm đã sẵn sàng tải về

1 Introduction: Nature as a Complex System

Tải bản đầy đủ ngay(0 tr)