Studies in History and Philosophy of Science Part C: Studies in History and Philosophy of Biological and Biomedical Sciences
Creating parts that allow for rational design: Synthetic biology and the problem of context-sensitivity
Introduction
Synthetic biology is an emerging and multifaceted discipline. The projects pursued by synthetic biologists range from the creation of minimal genomes, to the generation and study of synthetic cells, to the more general aim of turning biology into a discipline that can use engineering principles to rationally design novel biological systems (Deplazes, 2009, O’Malley et al., 2008). This paper focuses on this latter project of turning biology into a true engineering discipline. To reach this goal, synthetic biologists are aiming to develop a set of pre-characterised and standardised parts that can be reliably used for the construction of novel biological systems. Due to its focus on parts, I will refer to this approach as the ‘parts-based approach’ in synthetic biology.
The parts-based approach is in many ways a continuation of earlier projects in the life sciences.1 Protein engineering and genetic engineering have been pursued since the 1970s and the idea of designing novel entities with altered properties is not novel (see, e.g., Danielli (1972) for an interesting account of biology entering the ‘synthetic phase’). However, the parts-based approach followed in synthetic biology shows features that are new to biology. One of these features is the goal of creating a library of pre-characterised and standardised functional parts.
The idea of having a library of characterised biological parts raises a series of issues that need to be addressed if such an approach is to be successful. One such question is what the characterised parts should actually look like (Bennett, 2010). This question, which is the focus of this paper, becomes important because of the challenge that the context-sensitivity of biological entities poses to the goal of reliable construction in biology. In order to make the rational design of novel functional systems a reality, the parts-based approach needs pre-characterised parts that behave reliably, even if they are used in different contexts.
In Section 2, I will discuss the context-sensitivity of biological entities and the challenge it poses for the construction of novel functional systems. In Sections 3 Independent and orthogonal biological parts, 4 How to achieve functional independence, I will discuss some of the properties that, according to synthetic biologists, should turn pre-characterised parts into entities that can be used independent of their context. I will further discuss in these sections how synthetic biologists think that parts with such properties can be created. In Section 5, I will analyse a recent example of the successful use of pre-characterised parts in synthetic biology, namely the ‘expression operating unit’ (EOU) developed at the BioFAB center at UC Berkeley and Stanford University.2 The EOU is a framework within which some pre-characterised parts can be used in a predictable manner, even if part of their context changes. However, the analysis of this example will show that this success is not due to any artificially modified biological parts. It seems that the strategies and concepts discussed in Sections 3 and 4 are not responsible for the success of the EOU and it remains unclear how biology can avoid trial-and-error approaches and become more predictable. In the second half of the paper, I will claim that the initial question of how parts can be made predictable might be the wrong one to begin with and that the focus might have to be shifted away from parts and their properties towards the notion of capacities. In Sections 6 The use of the notion of a ‘functional’ element in synthetic biology, 7 Lessons learned from Cummins’ account of functional analysis, I will turn my attention to the literature on functional analysis in philosophy of science and use Robert Cummins’ account of functions as capacities to develop the notion of a condition space. The condition space is a dynamic and interconnected network of conditions that are required for the execution of a capacity of a biological element. I will show that an account using the notions ‘capacities’ and ‘condition space’ allows us to make sense of the functional independence that the EOU displays. I will conclude that it is the condition space of a capacity that has to be studied and made independent in order to achieve reliable functional composition in biology.
Section snippets
What does ‘reliable construction’ mean in synthetic biology?
There are two different aspects of construction in biology that have to be distinguished, namely the physical and the functional composition of parts (Canton, Labno, & Endy, 2008). Physical composition refers to the problem of how to establish a physical link between different biological molecules that are combined in a new system. If the parts of interest are, e.g., DNA-based entities, then one way of tackling the problem of physical composition would be to define the physical ends of DNA
Independent and orthogonal biological parts
According to Lucks et al., one of the key properties that a part must possess in order to be reliable is the ‘independence’ property. This property is described as ‘non-interference’: a part is independent if it does not interfere with the host machinery. In general, ‘interference’ means that there is interaction and ‘non-interference’ can therefore be defined as the absence of what could be called positive and negative interactions. By the absence of negative interactions I mean that there are
How to achieve functional independence
It has been claimed that abolishing unwanted interactions of biological parts, without simultaneously altering the basic function of the part, is key to achieving a larger degree of independence (Bennett, 2010, Endy, 2005). Aiming at independence then means getting rid of interactions that are not required for the functioning of the molecule of interest. This approach of increasing independence by simplification or ‘weeding out’ has been compared to refactoring in software development, where
The expression operating unit: the behaviour of parts becomes predictable
The EOU is a piece of DNA that combines different elements, which together serve the function of expressing a single polypeptide. The elements forming the EOU are: two insulator sequences at each end of the unit, a promoter, a 5′-untranslated region (5′-UTR) containing a ribosome binding site (RBS), a gene of interest (GOI) and a terminator (Bennett, 2010, Fig. 2). The insulators at the borders of the EOU play an important role, as they are supposed to insulate the expression unit from the
The use of the notion of a ‘functional’ element in synthetic biology
The term ‘function’ is used in different ways in synthetic biology and this can complicate the discussion about the refinement of functional elements. Looking at the literature on the parts-based approach, it becomes evident that some authors refer to parts (i.e. promoters, DNA-binding domains, etc.) as functional entities (see, e.g., Arkin, 2008, Endy, 2005), whereas other authors claim that only combinations of parts (i.e. devices), can be called functional (see, e.g., Andrianantoandro et
Lessons learned from Cummins’ account of functional analysis
Going back to the issue raised in Section 6, i.e. the different uses of the notion of a function in synthetic biology, we can start to see how Cummins’ account of functional analysis can be applied to the questions that arise in synthetic biology. According to Cummins, it makes perfect sense to talk of a single part as the carrier of a function; whenever a part has the ability to contribute to a higher-level system capacity, that part can be said to have a function.
But does this mean that a
Conclusions: capacities, condition spaces and the rational design of biological systems
The example of the EOU showed that functional biological parts can display independence of their context. At the same time it became clear that the independence observed is not due to any particular modification of the parts involved; the promoters and 5′-UTRs that behave predictably within the EOU were not specially modified to behave that way. The question of whether parts can be made independent remained unanswered and so did the question of how the parts-based approach could avoid the
Note added in proof
Since this paper has been submitted, the complete data on the expression operating unit has been published online (Mutalik et al., 2013a, Mutalik et al., 2013b).
Acknowledgments
I thank Claire Marris and Nikolas Rose for their support and critical input during the writing of a first version of this manuscript. I also thank the members of the former BIOS centre at the London School of Economics, where a large part of the work for this paper was carried out, for stimulating discussions. Jordan Bartol, Shareena Edmonds and Nicolas Wüthrich, as well as two anonymous reviewers, are acknowledged for their very helpful comments on the final version of this paper. This work
References (41)
- et al.
Regulation of gene expression via the core promoter and the basal transcriptional machinery
Developmental Biology
(2010) - et al.
Structural rearrangement of human lymphotactin, a C chemokine, under physiological solution conditions
The Journal of Biology Chemistry
(2002) - et al.
Zinc finger proteins: New insights into structural and functional diversity
Current Opinion in Structural Biology
(2001) - et al.
Toward scalable parts families for predictable design of biological circuits
Current Opinion in Microbiology
(2008) Exploration, iterativity and kludging in synthetic biology
Comptes Rendus Chimie
(2011)- Andrianantoandro, E., Basu, S., Karig, D., & Weiss, R. (2006). Synthetic biology: New engineering rules for an emerging...
Setting the standard in synthetic biology
Nature Biotechnology
(2008)- et al.
Engineering life: Building a fab for biology
Scientific American
(2006) - et al.
Engineering polydactyl zinc-finger transcription factors
Nature Biotechnology
(2002) - et al.
Synthetic biology
Nature Reviews Genetics
(2005)