System to simulate the activity of living organism - construction of proteome

Highlights

• The presented model - in contrast to others available in Internet - allows simulate the processes of adaptation to external signals.

• The systems available in Internet represent the static form of inter protein relations without possibility to simulate the activity of the sysytem in changed conditions.

• Large number of experiments in silico can be performed using the proposed model due to to possible option allowing changing the parameters.

• Model based on negative feedback loop satisfy all conditions of living organism - stability despite of permanent openness to signals from environment.

Abstract

The living organism maintains homeostasis despite being a thermodynamically open system in constant interaction with its environment. Only a system which implements automatic control mechanisms can meet these criteria. The negative feedback loop provides a way to implement such mechanisms.

While the organism is thermodynamically open, feedback loops active on the level of cells and the organism as a whole maintain specific substrate concentrations and functions which we refer to as “normal” or “physiological”. Each automatically controlled process constitutes a distinct unit associated with a specific metabilism. This unit encompasses all components which enables a specific function – it may therefore be referred to as a “functional unit” or a “functional-structural unit”. Individual units which control specific processes may be further linked with one another, forming a complex network of dependencies.

Receptors play a crucial role in feedback systems. By detecting deviations from the predefined norm – within the scope of a given process – they can trigger counterbalancing actions. They are also involved in inter-process communications. Such communications can be either cooperative – where the product of one process becomes the substrate of another – or coordinative – where the product of one process modulates the behavior of another by affecting the operation of its control function. The latter type introduces a hierarchy of processes.

Proteins are the basic building blocks of biological systems. Their mutual interactions are represented in the form of proteomes, in which they represent structural components. It is, however, difficult to determine what the role a given protein plays in a metabolic process purely by analyzing its structure. Treating proteins as elements of feedback loops provides far greater insight into their functional characteristics. This is why the presented study focuses on functional units rather than on individual proteins.

Introduction

The concentrations of metabolic substrates and the activity levels of biological processes are maintained in a steady state under physiological conditions. This is possible thanks to automatic regulation based on negative feedback loops. Each regulated process constitutes a distinct functional unit which accepts certain substrates and produces certain products. Metabolism itself can be defined as an interplay of such units, which must remain coordinated – and therefore linked together – for the organism to function.

In order to explain the activity which takes place in cells, we must first provide a generalized view of metabolic processes. The function of a cell is dependent on the availability and concentration of various proteins. These proteins, along with their interactions, constitute the so-called proteome. We have determined the structure of most proteins found in cells; however their interactions and metabolic importance often remain a mystery. Notably, many proteins represent intermediate stages in synthesis or degradation of biological constructs and their structural analysis does not – by themselves – contribute to understanding the changes which occur in living cells. Instead, a useful proteome should focus primarily on interactions between essential proteins.

When constructing a proteome we must also determine our basic structural unit – this can be either an individual protein, a group of proteins or a group of proteins subject to regulation (Fig. 1).

The concept of a proteome arrived as a basic consequence of decoding the human genome [1,2]. Comprehensive visualization of the whole molecular system acting in the organism was shifted to the post-translational space – the action of proteins. The proteome was originally treated as network of links, where a link between two or more proteins represents possible interaction between these units. The first proteome developed according to these rules was that of yeast Saccharomyces cerevisiae [3]. This was followed by a similarly constructed Drosophila melanogaster proteome [4].

A significant progress in proteom construction is currently connected with using modern mass-spectrometry analysis [5,6].

Unfortunately, such proteomes are inherently static and represent a large-scale information repository where the biological function of each protein is coded in different forms. Despite the enormous amount of information which goes into assembling this kind of proteome, its availability does not bring us much closer to understanding the reaction of cellular metabolism to changes in environment. When discussing proteomes, the fundamental issue is to define its basic units and the properties of links between such units [[7], [8], [9], [10], [11]]. There is no direct correspondence between individual proteins and specific biological functions, because in most cases many proteins must cooperate to produce a given function (treated as their shared biological “goal”). It is hence often unclear to what degree a specific protein contributes to a certain process. In most cases, the first enzyme in the set of enzymes engaged in a given process is directly involved in its control, and therefore plays a key role. When focusing on controlled processes, it becomes easier to pinpoint the control enzyme in each case.

Acknowledging negative feedback in proteome construction also enables us to determine the outcome of each process as only one compound (whose synthesis is controlled by the receptor) may be regarded as the final product. Regulatory mechanisms dictate, which compounds may be temporarily stored in the cell and which ones must be immediately consumed to ensure biological activity. Thus, a negative feedback loop (controlling a specific process) seems to be better suited for constructing proteomes than an individual protein.

Of course, in order to construct a proteome from structural/functional units, we need to first explain the nature of “feedback”. The relevant mechanisms are discussed below.

For the reasons stated above, we have attempted to construct a proteome which enables us to simulate changes and their outcomes. The design of a proteome built from negative feedback loops was initiated in the framework of the Sano project [12].

All processes occurring in cells are subject to automatic regulation. The task of regulation is to maintain homeostasis, i.e. a state where the basic processes which take place inside cells are maintained at a steady level in spite of spontaneity, which results from a certain imbalance (a permanent deviation from chemical equilibrium), making these processes depend only on the availability of substrates and removal of products (Fig. 2).

The negative feedback loop is a typical regulatory mechanism encountered in automatic systems. In this kind of loop, the receptor (detector) is capable of producing signals which inactivate the associated effector when the concentration of a given product exceeds a certain threshold. The affinity between the receptor and the measured product determines the designed effect – in most cases, this corresponds to the concentration of the product. The negative feedback loop can effectively control a process in the cell, but only if product concentration can be measured directly. This facility can only occur in restricted area, what explains why cells are almost universally very small, regardless of whether they belong to an elephant or a mouse. Additionally, communication between cellular receptors and effectors is often simplified by physically combining both subsystems. In such structures, conformational changes affecting the receptor are automatically recognized by the effector – this change in structure can therefore be regarded as the direct structural communication. Protein complexes which conform to this model are referred to as regulatory enzymes.

A different situation occurs on the level of the organism, where regulation also bases on negative feedback loops, but signals must traverse extremely long distances (compared to intracellular signals). Here, direct contact between the receptor (which registers changes in serum concentrations of certain compounds), and the effector (often found only in certain tissues), cannot occur. Signals transmitted through blood or nerves are typically very weak –which should come as no surprise given that must be capable of rapid distribution and be easily degraded once they have served their purpose. This, however, implies that the signal must undergo significant amplification in target cells. Hormones serve as the most common signal carriers on the level of the organism. Unlike intracellular signals, a hormonal signal is not the direct signal itself: instead, it represents encoded information, since it must differ from other compounds found in the environment, and be protected against ambient “noise”. This process can be likened to voice communication, which is only possible in small spaces where noise levels remain low – however, in order to send voice over long distances, it must be encoded – for example as radio waves – and then decoded and amplified in the receiver.

In regulatory mechanisms the receptor role falls to allosteric proteins which can adopt varying conformations depending on complexation with the product [[13], [14], [15], [16]].

The effector often consists on multiple enzymes. Its role is to generate an effect – for example by synthesizing a product whose concentration can be measured by the receptor, which, in turn, modulates the operation of the effector depending on the achieved effect (eg. product concentration). The receptor and the effector must remain in constant communication, which calls for an information carrier.

In this type of system, both protein components – the receptor and the effector – can be treated as parts of a single structural-functional unit. The notion of a “structural-functional unit” refers to an automatic system which contains all components necessary to ensure autonomy – including the detector and the effector – regardless of its structural complexity or the complexity of the associated signaling mechanisms and the effector’s end product.

The effector is typically composed of multiple proteins (mainly enzymes), although in certain special cases the number of proteins making up the effector may remain small. One example of a single-enzyme effector is hexokinase (which also provides an example of an entire feedback loop embodied by a single protein).

The number of proteins belonging to a structure-function loop usually depends on the structure of its effector. This is because effectors often involve many different enzymes and auxiliary proteins (although complex receptors are not unheard of either). An example is provided by the immune system where many different antibodies and other proteins cooperate to form a comprehensive detection subsystem.

A properly functioning negative feedback loop is characterized by sinusoidal variations in the concentration of its product.

Simulating modifications in the loop’s behavior enables us to determine how the system may react to various input parameters. Such modifications may involve the sensitivity of the receptor (its affinity for the product), resulting in a different set point for the product’s concentration. Since effectors are usually enzymes, the changes may also include e.g. variations in their turnover number, resulting in a different frequency of product synthesis. These phenomena may be studied in silico by visualizing their effect on the shape of the sinusoid which expresses cyclical changes in the product’s concentration depending on the abovementioned parameters.

A software toolkit capable of performing such test simulations has been developed in limited form [17]. Its author – an IT specialists – has identified the need for two additional temporal factors: the time required for information dispatched by the receptor to arrive at the effector, and the timeframe for transfer of information encoded as variations in product concentration.

Taken together, the four parameters presented above enable us to simulate the activity of biological feedback loops, and the effect of any changes. Fig. 3 provides examples of variations in the concentrations of sample products depending on the presets of a given feedback loop.

The simulation must be able to differentiate between control on the level of cells and the organism as a whole. The operation of automatic regulatory systems – cellular and that operating in the organism – differs with respect to its strategic properties. The cell is relatively small, and it is easy for its receptors to directly pick up any change in the concentration of a controlled product. What is more, intracellular receptor proteins are often directly linked to their corresponding effectors (as in the case of regulatory enzymes), which obviates the need for signal amplification. Finally, there is no need to worry about ambient noise which might disrupt signaling, and therefore no need to encode the signal – unlike in long-distance communication.

On the other hand, when dealing with the organism as a whole, signals must be encoded and then greatly amplified at their destination. This is why regulatory systems operating in the organism require different mechanisms.

The cell is a standalone entity, which can function outside of the organism. It may therefore be simulated autonomously. Nearly all metabolic processes are intracellular in scope. Only a handful – such as blood coagulation, immune response or acute-phase protein function – are coordinated on the level of the organism. The reason metabolism occurs mostly inside cells is because this greatly simplifies signaling.

Each process controlled by a negative feedback loop may be treated as a distinct structural-functional unit. If, however, the cell is to be regarded as a component of a living entity, capable of preserving homeostasis, it must communicate and co-act with other similar units. Such co-action gives rise to a network of dependencies. The two basic mechanisms of co-action are cooperation and coordination. The former is a situation where distinct structural-functional units become linked because the product of one serves as the substrate for another (or for many other units). In this case, the affinity of the receptor for the product remains unchanged. On the other hand, coordination implies a hierarchical structure of feedback loops. In this case, the signal produced by one loop modulates the affinity of the receptor in another loop for its designated product.

In the former process, the receptor does not undergo changes modifying affinity, whereas in the latter it does. The system operating initially without changing affinity or the receptor may be changed additionally by the product of other feedback loop playing the role of coordinating signal altering this time the affinity of the receptor (Fig. 4.). Some important processes may be controlled in this way by many coordinating signals as for example synthesis of glutamine. This process typically occurs in organism-cell signaling, where the receptor is additionally covalently modified through phosphorylation. Familiarity with the mechanisms of co-action between various intracellular and those occurring in organism enables us to construct a proteome consisting of negative feedback loops.

When dealing with cellular metabolism, it becomes useful to distinguish processes with regard to which organelles they occur in. This is needed to avoid conflicts in the case of counterbalancing processes (e.g. synthesis and degradation of a product). Even a simplified analysis should acknowledge this specificity – although not all processes need to be simulated upfront; some (such as those related to processing of genetic material) may be set aside in a simplified model. Such simplifications appear necessary in order to come up with efficient digital simulations. In the presented work, the “core” of the proteome comprises processes related to energy management – i.e. glycolysis and the Krebs cycle. These have been accorded particular attention because they underpin a vast array of follow-on synthesis and degradation processes by supplying them with energy (Fig. 5).

Focusing on regulation requires us to present metabolic processes in a different way than in most biochemistry textbooks. In particular, it is possible to link them to other processes whose specifics are well understood. The simulation, however, requires us to be aware of the concentrations of substances which vary depending on the intensity and direction of the given process. Thus, any simulation must refer to specific conditions, which complicates understanding of the problem. Adapting the proposed structure of the proteome to the specific characteristics of cellular metabolism calls for a broad range of parameters, including physiological concentrations of metabolic products, receptor affinities and activity levels of effector subsystems. For making however the simulation functional in practice it is also necessary to introduce certain simplifications choosing selected key processes only.

Simulations of automatic regulatory loops may, however, also focus on isolated systems, as long as their function is sufficiently distinct. One such system is the circadian clock, for which a simulation is currently being prepared independently [18].