Correction: Bottom-Up Engineering of Biological Systems through Standard Bricks: A Modularity Study on Basic Parts and Devices

Background: Modularity is a crucial issue in the engineering world, as it enables engineers to achieve predictable outcomes when different components are interconnected. Synthetic Biology aims to apply key concepts of engineering to design and construct new biological systems that exhibit a predictable behaviour. Even if physical and measurement standards have been recently proposed to facilitate the assembly and characterization of biological components, real modularity is still a major research issue. The success of the bottom-up approach strictly depends on the clear definition of the limits in which biological functions can be predictable. Results: The modularity of transcription-based biological components has been investigated in several conditions. First, the activity of a set of promoters was quantified in Escherichia coli via different measurement systems (i.e., different plasmids, reporter genes, ribosome binding sites) relative to an in vivo reference promoter. Second, promoter activity variation was measured when two independent gene expression cassettes were assembled in the same system. Third, the interchangeability of input modules (a set of constitutive promoters and two regulated promoters) connected to a fixed output device (a logic inverter) expressing GFP was evaluated. The three input modules provide tunable transcriptional signals that drive the output device. If modularity persists, identical transcriptional signals trigger identical GFP outputs. To verify this, all the input devices were individually characterized and then the input-output characteristic of the logic inverter was derived in the different configurations. Conclusions: Promoters activities (referred to a standard promoter) can vary when they are measured via different reporter devices (up to 22%), when they are used within a two-expression-cassette system (up to 35%) and when they drive another device in a functionally interconnected circuit (up to 44%). This paper provides a significant contribution to the study of modularity limitations in building biological systems by providing useful data on context-dependent variability of biological components. Citation: Pasotti L, Politi N, Zucca S, Cusella De Angelis MG, Magni P (2012) Bottom-Up Engineering of Biological Systems through Standard Bricks: A Modularity Study on Basic Parts and Devices. PLoS ONE 7(7): e39407. doi:10.1371/journal.pone.0039407 Editor: Mark Isalan, Center for Genomic Regulation, Spain Received February 28, 2012; Accepted May 24, 2012; Published July 20, 2012 Copyright: 2012 Pasotti et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: This project was partially funded by the Italian Ministero dell’Universita’ e della Ricerca through the FIRB ITALBIONET project. No additional external funding was received for this study. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: paolo.magni@unipv.it

S cell,J23118,GF P 32 S cell,P lacIQ,RF P 34 (red fluorescence) and the 100· S cell,P LlacO1,RF P 34 S cell,J23118,GF P 32 (green fluorescence) ratios, estimating the maximum percent contribution that GFP and RFP could give to red and green fluorescence measurements respectively, were computed. No detectable RFP contribution could be observed in the green fluorescence acquisitions, while the resulting ratio was 1.4% when red fluorescence of GFP-expressing cells was acquired. This can be considered as reasonably low crosstalk value.

Preliminary design of the interconnected system
The first design of the interconnected system included a logic inverter with a strong RBS (BioBrick BBa B0034) upstream of the tetR gene. Such system, however, always remained in the OFF state even when the uninduced P lux promoter, whose basic activity was very low, was assembled upstream. Only the promoterless logic inverter gave a high output when tested (data not shown). This result is consistent with previous findings in which such logic inverter was tested in similar conditions [15]. In order to construct a logic inverter that could switch from the ON state to the OFF state in a range of RPUs between 0.05 and 2, which is exhibited by a number of easy-to-retrieve promoters from the Registry of Standard Biological Parts [14], the RBS upstream of tetR was changed. A much weaker candidate was chosen (BioBrick BBa B0031). It gave the expected effect (see main text), as the switch point occurred when the input RPUs were ∼0.14.
The low copy vector condition was chosen to characterize the system because such condition can give more reliable results than in high copy vectors [17,27]. Moreover, attempts in cloning the interconnected circuit in a high copy vector (pSB1A2, which has a pUC19-derived replication origin) gave no successful transformants, suggesting that one (or more) of the modules causes a high metabolic burden when present in >100 DNA copies.
A set of four synthetic constitutive promoters of different strengths was chosen as the INPUT1. All of them are 35-bp long and share a common structure [S1]. Inducible lacI-and luxR-regulated promoters were chosen as INPUT2 and INPUT3, respectively. Single-cell analysis reported in [17, 28, S2] showed that these two systems produce a homogeneous response in an induced cell population in presence of sub-saturating concentrations of IPTG or HSL. KRX E. coli strain was used in this study because it overexpresses the LacI repressor through a lacI expression cassette with the lacIq mutation, carried in the F plasmid. This allows the tight transcriptional control of lacI-regulated promoters without including a lacI gene in the circuit.

Characterization of individual promoters in a high copy vector
It is common knowledge that the activity of genetic parts in high copy number vectors can be nonlinearly affected by the overloading of cell machinery due to the high copy number of the DNA-encoded functions [17]. This was confirmed by comparing the activity of promoters characterized via the same reporter device in low copy and high copy vectors (see Figure S1): the RPUs of the two strongest promoters, P LlacO1 and P R , were respectively 4.4-and 2.3-fold lower in high copy when compared to low copy. This means that, given a reporter device, the ratio between the activity of the promoter of interest and the reference is lower in high copy when compared to low copy. The other promoters did not show such a large difference (<1.3-fold). The observed large-entity variations could be due to saturation effects in transcription/translation processes that occur for the strongest promoters in high copy condition, while such effects were absent for the other (weaker) promoters.