The biological microprocessor, or how to build a computer with biological parts

These are the figures form the following publication:

Moe-Behrens GHG (2013) The biological microprocessor, or how to build a computer with biological parts. Computational and Structural Biotechnology Journal. 7 (8): e201304003. doi: http://dx.doi.org/10.5936/csbj.201304003

(The paper is open access and available for free under the link above)

Abstract

Systemics, a revolutionary paradigm shift in scientific thinking, with applications in systems biology, and synthetic biology, have led to the idea of using silicon computers and their engineering principles as a blueprint for the engineering of a similar machine made from biological parts. Here we describe these building blocks and how they can be assembled to a general purpose computer system, a biological microprocessor. Such a system consists of biological parts building an input / output device, an arithmetic logic unit, a control unit, memory, and wires (busses) to interconnect these components. A biocomputer can be used to monitor and control a biological system.

Figure 1. Natural computing: A platonic idea is an archetype, a blueprint, the essence of various phenomena of the same thing. Systemics and systems biology are such ideas, describing data processing systems in nature in terms of mathematics and formal logic. Systemic ideas have been used as a blueprint for silicon computing. Ideas derived from the observation of nature have also inspired computer models of nature. Engineering ideas behind silicon computer (such as standardized parts, switches, logic gates, input /output device, arithmetic logic unit, control unit, memory, and busses) have been used by synthetic biologists to build computers with biological parts, with the ultimate goal to control data processing in nature.

Figure 2. Four units of a general purpose computer: Input and output device (I/O; I = input signal; O = output signal), the arithmetic logic unit, control unit, memory. Busses (groups of wires) connect these units.

Figure 3. Input/Output (I/O) device: A) In a “digital” biological I/O device input molecules induce due to a set of non-steady state chemical reactions (engineered coherent with a logic scheme) an output molecule. All molecules have a defined concentration translated into Boolean logic; alternative on (1) or off (0). B) In order to do so, normalized molecule concentrations (conz.), which change over time, are defined as off (0), if they are under a certain threshold (tr), and, if they are above, as on (1) C) A switch, which produce an on (induced) or off (not induced) state: The figure gives an example of a switch in a synthetic gene network (adapted from [121]). Off (no detectable EGFP expression): LAcl repressor proteins, which are constitutively expressed, bind to two introns with lac operator (lacO) sites, inducing transcriptional repression of EGFP and TetR respectively. Repression of TetR allows transcription of shRNA, which can subsequently bind to its target sequence, and repress it’s shRNA target. On (EGFP expression induced): isopropyl-b-thiogalactopyrano (IPTG) binds to Lacl proteins. As a consequence, the repressor proteins are inactive, as they change their conformation. Thus, TetR, which represses shRNA, and EGFP get transcribed.

Figure 4. Arithmetic logic unit: Shown are four basic Boolean logic gates (AND, NOT, NOR, and XOR), their symbols and respective truth tables. 1 means that the input (a, b) is sensed or the output (out) is released, whereas 0 means not. In the examples system output = 0 is highlighted as pink, output = 1 as green. A) An AND gate can be based on the transcriptor (T), an asymmetric transcription terminator, which can block RNA polymerase flows one directional. If both terminators are flipped, induced by their respective input signal (a and b), RNA polymerase flows unhindered (full length RNA output). B) Deoxyribozyme based NOT gate: The deoxyribozyme (DNA based catalyst) is in an active form, if no input (in) is present (in = 0). Cleavage activity results in this case in a fluorescent oligonucleotid (F) as output. An oligonucleotide input (in) (in = 1) leads to hybridization of the input strand (green) with the closed loop strand, which is marked purple. This results in an inactive, open loop and the absences of a fluorescent product. C) An RNA aptamer based NOR gate: NOR is an OR gate followed by a NOT gate. Two subsequent RNA devices consist, is this case, each of three functional components: a sensor, made of an RNA aptamer (brown), an actuator component, made of a hammerhead ribozyme (purple), and a cobbling sequence between these parts, the transmitter (blue). Translation of the gene of interest (here GFP), encoded upstream of the device, is only possible in the case of the absence of both inputs (a and b). D) An inter cellular network based XOR gate: The system is built from four Escherichia coli colonies, whereas each colony consists of a strain engineered to contain a single gate. Three cell colonies (cell 1, 2, 3) containing NOR gates and a fourth (cell 4) a BUFFER gate (two subsequent NOT gates; if in = 0, so out = 0; or if in = 1, so out = 1). The cell colonies communicate through quorum sensing, which represent the “wires” of the system. If both inputs (a, b) are present, or if a and b are absent, the system has no output. If either a or b is present, yellow fluorescent protein (YFP) is expressed.

Figure 5. Control unit; central processing unit: A) A final state machine, as shown here, is a theoretical model which can help to understand what is going on in the central processing unit. Simplified: Symbols a and b are written on a tape, which is read by the machine letter by letter from left to right. In this example the tape ends with the final letter b. Each letter provides the instruction to the machine into which state (S1, or S0) it should move; here a means move to state state 0 (S0) and b codes the instruction move to state 1 (S1). The final state of the machine in this example is thus S1. B) Molecular implementation of a final state machine. The upper part of the figure contains the definitions for this machine: The symbols a, b, and t (terminator) are implemented as a sequence of six specific nucleotides. The state (S1 or S0) of the machine is defined by a 5’ overhang (generated during the computing process, see below) consisting of 4 specific nucleotides (inside the frames). The terminator defines the final symbol read. The machine consists of an input molecule, a transition molecule, an output detector and two enzymes Fokl and ligase. The input molecule consist of a Fokl recognition site (F), a spacer x (a certain defined number of nucleotides), a nucleotide sequence defining a and b, a sequence with the remaining symbols (rem = n numbers of a and b in a defined order) and the terminator sequence (t). Fokl is a restriction endonuclease which can bind to F. It cleaves the DNA (without further sequence specificity) on the sense strand 9 nucleotides downstream and the anti-sense strand 13 nucleotides upstream of the nearest nucleotide of the recognition site. Thus the space x defines where Fokl is cutting. The cleavage of the input molecule results in the first intermediate state (S0), an 5’ overhang, reading a. Ligase ligate this product with the transition molecule. This transition molecule determinates the transition between the states, in this example: if a is read, move from S0 to S0. Other transition molecules can be generated defining all the other possible transition rules. These molecules are designed such that the 4 bases long 3’ overhang reads the symbol, the spacer x defines the cutting point of Fokl 1 and the state the machine will transit to (here S0). The input molecule and the transition molecule get ligated. A new digestion with Fokl leaves an 5’ overhang representing S0 and a reading b. This cycle continues until all remaining symbols (rem) are read and state transitions are executed. The last digestion leaves a 5’ overhang with a terminator sequence defining the final state, in general either S0 or S1 (in this example S0). The molecule in its final state, gets ligated to an output detector, engineered to recognize either state 0 or 1. This forms an output-reporting molecule, which can be detected by gel electrophoresis.

Figure 6. Memory: A) Shown is a simplified diagram of a modular memory device, which is a transcriptionally controlled network composed of two transcription factor encoding genes, a sensor gene and a positive (+) auto feedback gene (P-GAL = GAL 1/10 promoter, P-CYC = CYC 1 promoter, DNA BD = sequence encoding a DNA binding domain of the respective transcription factor). The network can be in three states, off, on and memory. The system is in of state, if it has never been exposed to a signal (here galactose). It is on, if galactose is present. In this case the signal induces the synthesis of a transcription factor, the sensor. This triggers the expression of another transcription factor able to bind to its own promoter. The system is in memory state, if the signal is removed. The auto feedback activator is able to initiate its own expression even if the inducing signal is lacking, which means that the system has stored information. B) A rewritable recombinase addressable data module, able to store data within a DNA sequence (simplified adaption from [42]): Serine integrase and excisionase are used to invert and restore specific DNA sequences. The system has two potential inputs; a set and a reset transcription signal. This set signal drives expression of integrase which inverts a DNA element, functioning as a genetic data register. Flipping the register converts the flanking sites (triangle). The system is now in state 1 (S1). Alternatively a reset signal drives integrase and excisionase expression and restores register orientation and the flanking sites. The system is in its other state (S0). The register comprises a promoter, which is driving state dependent, strand-specific transcription of either red or green fluorescent protein, the two possible outputs of the system.