Design and Construction of “Synthetic Species”

Synthetic biology is an area of biological research that combines science and engineering. Here, I merge the principles of synthetic biology and regulatory evolution to create a new species with a minimal set of known elements. Using preexisting transgenes and recessive mutations of Drosophila melanogaster, a transgenic population arises with small eyes and a different venation pattern that fulfils the criteria of a new species according to Mayr’s Biological Species Concept. The population described here is the first transgenic organism that cannot hybridize with the original wild type population but remains fertile when crossed with other identical transgenic animals. I therefore propose the term “synthetic species” to distinguish it from “natural species”, not only because it has been created by genetic manipulation, but also because it may never be able to survive outside the laboratory environment. The use of genetic engineering to design artificial species barriers could help us understand natural speciation and may have practical applications. For instance, the transition from transgenic organisms towards synthetic species could constitute a safety mechanism to avoid the hybridization of genetically modified animals with wild type populations, preserving biodiversity.


Introduction
It has been argued that reconstructing a system is the ultimate way of understanding it [1,2]. In order to further comprehend the origin of new species, and to explore possible applications in modern biotechnology, I engineered reproductive isolation between populations of Drosophila melanogaster by generating a synthetic species boundary.
The use of Drosophila is justified because this model organism is leading the fields of regulatory evolution and speciation [3,11]. For example, previous work in several species of Drosophila produced fundamental contributions regarding the genetics of speciation [7][8][9][10]. In addition, key studies of evolution in Drosophila have shown that novelty arises more readily from the recruitment of existing elements into new regulatory networks than from the development of completely new components [3][4][5][6].
However, speciation has been given no attention as a tool for biotechnology and in the context of genetically engineered organisms. Previous artificial speciation experiments produced ''incipient species'' that were not fully isolated or whose speciation genes were unknown [12][13][14][15][16]. Reproductive isolation has been brought about in plants for many decades through polyploidization, which creates individuals that in crosses to parents give rise to sterile progeny [13], and, more recently, in yeast [14]. For the animal kingdom, a parthenogenetic species of lizard was also generated [15], but, again, its speciation genetics are not understood, does not involve transgenesis nor synthetic design and its components can therefore not be reliably and predictably manipulated.
Unlike previous artificial speciation events, the synthetic species boundary described here is a genetic circuit based on the combination of 5 well known preexisting elements leading to reproductive isolation. Building components so well understood that can be reliably and predictably manipulated is one of the goals of synthetic biology. In this case it allows opening and closing speciation gates when desired.

Results
Regulatory evolution acts by using available preexisting genetic elements to generate novelty. Likewise, the synthetic genotype created here is achieved by implementing known elements, however their selection and specific arrangement establishes a previously unknown synthetic species barrier (Fig. 1).
The first element consist of null mutations in the glass (gl) gene [19,20]. The glass product is a transcription factor of 604 amino acids with five zinc-fingers. Mutations in gl specifically abolish photoreceptor cells resulting in blind, but viable flies [19,20]. The gl 60J allele is a spontaneous mutant caused by the insertion of 30 kb of unknown DNA into the gl locus and it is believed to be a null allele [19,20]. Other alleles (gl 3 and gl BS1 ) have also been used for this study.
The second element is formed by the Glass Multimer Reporter (GMR) [20,21], a heterologous promoter construct containing five tandem copies of a 27-bp glass-binding site normally present in the regulatory region of ninaE, the major rhodopsin gene in Drosophila. The GMR promoter can therefore drive glass-dependent expression in the photoreceptor cells of Drosophila eyes.
The yeast protein GAL4 as a third building block can activate transcription in Drosophila from promoters that bear GAL4 binding sites [22,23]. In addition, the GMR sequence has been previously subcloned in front of gal4, thus driving Gal4 expression under the control of Glass (GMR-gal4) [21].
Fourth, a tandem array of five GAL4 binding sites (56UAS, for Upstream Activation Sequence) is employed where GAL4 binds with high affinity to induce the transcription of a downstream located gene.
The fifth element is a ras v12 allele, a mutant form of the Drosophila ras gene [24,25]. Conversion of the glycine residue at position 12 to valine constitutively activates the Ras protein. ras v12 has been previously subcloned behind GAL4 binding sites (UASras v12 ), which permits activation only within cells where GAL4 is expressed [24].
The inherent logic of the design relies on a ''killing module'' (Fig. 1b) and a regulator to switch it ON and OFF (Fig. 1c): The killing module is formed by GMR-gal4 and UAS-ras v12 whose activation is controlled by the presence or absence of the gene glass. When glass gene function is unperturbed, transcription of UAS-ras v12 driven by GMR-gal4 consistently kills 100% of the flies at any temperature from 17uC until 29uC (606/606 lethality at 17uC, 558/558 lethality at 23uC, 330/330 at 25uC, 110/110 at 29uC, Table 1). Pupae arrest at mid pupation and due to abnormal tissue lysis the pupal case ends up almost empty.
More than 20 other UAS transgenes were tested, including UAScaudal [26], UAS-flower LoseA [27] or UAS-eiger [27], but UAS-ras v12 was the only one that resulted in 100% lethality when driven by the presence of glass and the GMR-gal4 transgene at all temperatures (from 17uC to 29uC) ( Table 1).
In the synthetic genotype GMR-gal4/GMR-gal4; UAS-ras v12 , gl 60J /UAS-ras v12 , gl 60J (a glass 60J mutant background, where no Glass protein is present) Ras v12 cannot be produced (OFF state, Fig. 1d). Surprisingly however, in addition to the small eye phenotype (Fig. 2a,b), those flies showed a different wing morphology, with lateral extra veins (Fig 2d, compare with the wt wing pattern shown in 2c). Other alleles (gl 3 and gl BS1 ) were also tested and yielded the same phenotype. Most likely the heat shock promoter (hsp70) of the GMR-gal4 construct is leaky, leading to very low activation of UAS-ras v12 and consequently to the phenotype [28].
When hybrids between Drosophila melanogaster and the synthetic genotype are produced, the ''killing module'' GMR-gal4; UAS-ras v12 is triggered by the presence of the glass gene (Fig. 1d, Table 1, Fig. 3a). This genetic network, while still allowing normal reproduction among flies with the synthetic genotype, completely isolates GMR-gal4/GMR-gal4; UAS-ras v12 , gl 60J /UAS-ras v12 , gl 60J flies from normal D. melanogaster due to hybrid early pupal lethality (Fig. 3a, Table 1). Unlike with the other known and naturally occurring speciation mutations [8], the sex of the parents did not influence the lethality of the hybrids in this case (Table 1, Fig. 3a). Experiments were performed at 17uC because flies of the synthetic genotype grew better and because due to the temperature sensitiveness of the Gal4 it is likely to be the temperature at which the killing module may be less effective. Despite this, the killing module was 100% effective even at 17uC (Table 1, Fig 3a). In the initial population mutations in yellow (y 1 ), which results in mild pigmentation, existed as a polymorphism in some individuals.
It is often difficult to delineate ''species boundaries'' since they may carry identical mutations and are related to one another through common ancestors. However, most biologists agree on a very stringent definition for species, the Ernst Mayr's Biological Species Concept, according to which species consist of populations of organisms that can reproduce with one another, but are reproductively isolated from other such groups [7] (Fig. 3b). This definition leads to a focus on the barriers to reproduction between species [8][9][10][11][12][13][14]. Such barriers represented one of the main problems for Darwin who wrote: ''How can we account for species, when crossed, being sterile (…), whereas, when varieties are crossed, their fertility is unimpaired?'' [11]. Because the postzygotically isolated population generated here conforms to the most stringent definition of species [8][9][10][11][12][13][14], it will subsequently be called Drosophila synthetica.
To further prove that the synthetic genetic network allowed zero gene flow with D. melanogaster, co-cultures of both populations were performed for 13 generations (using D.melanogaster white (w) mutants with white eyes) and not a single hybrid was recovered (Fig. 3c, Table 1). Hybrids would have been easily recognizable by normally sized red eyes, because they would carry a normal copy of gl and two w+ copies from the transgenes ( Table 1), but D. synthetica behaved like a stable species, did not interbreed and maintained its characteristic eyes (Fig. 3c). Identical results were obtained when crossing D. synthetica with other melanogaster strains, including y,w flies and y,w,f flies (Table 1). In all cases synthetica and melanogaster did not interbreed (Table 1).
Assembling synthetic species boundaries can have practical applications. For example, the use of recombinant DNA technology to alter organisms for a specific purpose has raised controversy [18] and is a growing problem due to the increasing number of transgenic organisms approved by regulatory agencies [16][17][18]. A new framework where safety mechanisms are genetically designed along with desired modification could help to gain public support for a technology with the potential to satisfy future medical and nutritional needs [16][17][18]. D. synthetica is the first transgenic organism that cannot reproduce with the original wildtype population. I therefore propose that synthetic species barriers may serve to compartmentalize dangers and protect natural species from interbreeding with emergent transgenic forms, therefore preserving natural biodiversity (Table 1). Moreover, once a genetic network is identified, as is the case for the ''ras-glass'' synthetic boundary described here, opening or closing of the barrier can be controlled at will. In case the interbreeding of populations appears beneficial, targeted strategies can be implemented to reverse hybridization barriers. To test this experimentally, the GAL4-inhibitor GAL80 was expressed from a tubulin promoter (tub-gal80) [29] in D. melanogaster in order to remove hybrid lethality and traverse the species barrier. Males of D. synthetica hybridized successfully with tub-gal80 D. melanogaster females and produced viable hybrids (Table 1), as predicted because Gal80 can block the ''killing module''.

Discussion
The postzygotically isolated population generated here in the genus Drosophila conforms to the most stringent definition of species [8][9][10][11][12][13][14], as well as to the principles of synthetic biology [1,2], and it has been consequently named Drosophila synthetica. I propose the term ''synthetic species'' to distinguish it from ''natural species'', not only because it has been created in the laboratory, but also because it may never be able to survive in the wild, unlike ''natural species''. For example, because the flies created are blind and only survive at lower temperatures, they have potential fitness deficits and it could be argued that the changes could not be arrived at in concert because the ''fitness valley'' will not be traversed in the wild. However, blindness is a common adaptation in caves suggesting that fitness deficits are difficult to predict and depend on environmental conditions [7,11,27].
Interestingly, the generation of Drosophila synthetica matches the Dobzhansky-Muller theoretical model for postzygotic incompatibilities during naturally occurring speciation [7,8], according to which an ancestral population splits into two independent populations that then accumulate mutations (Fig. 3b). Subsequent genetic interactions between those mutations cause hybrid incompatibilities. In particular, it conforms to a derived-ancestral incompatibility (Fig. 3b), in which all substitutions occur in the derived population.
I therefore propose that modifications in transcription factors and appearance of cryptic enhancers upstream of potentially lethal gene products can constitute a normal Dobzhansky-Muller mechanism for speciation (Fig. 3d). The appearance of those cryptic enhancers could be driven by the accumulation of point mutations in regulatory regions (Fig. 3d), in a manner similar to what has been described recently [30], but those enhancers will only be recognised by the ancestral transcription factor which is now missing (or modified) in the derived population (Fig. 3d). When hybridization between the derived and ancestral populations occurs, the genes with cryptic enhancers will be activated by the ancestral transcription factor, causing hybrid lethality and reproductive isolation (Figs. 1,2, Table 1). This could constitute a general mechanism through which regulatory evolution creates species boundaries (Fig. 3d) and may help to define concrete target genes mediating speciation.
One of the predictions of classical evolutionary theory is that organisms that connect two species must exist as part of the gradual divergence process [11]. Because we fully know the mutations forming a reproduction barrier between melanogaster and synthetica, it is feasible to dissect the process and move backwards, showing how populations of intermediate mutants can indeed interbreed with populations at either side of the evolutionary path towards postzygotic isolation (i.e., glass mutants can hybridize with both species (Table 1) and hence connect melanogaster with synthetica, as if they were a ''missing link'') (Fig. 3).
The results shown here provide proof of principle for the transition from ''transgenic animals'' to ''synthetic species'', as defined above, and should spur the debate for its use as a failsafe mechanism in biotechnology. Modifying the binding affinities of one transcription factor and the enhancers it recognises, could be used to engineer reproductive isolation in other living animals, not only in Drosophila. Moreover, the ability to open and close speciation gates when desired reflects one of the goals of synthetic biology -to build components that can be reliably and predictably manipulated-, and preserves flexibility while gaining control over the spread of genetically modified organisms.
One potential caveat could be that this barrier is not irreversible since it can be overturned quite simply if a spontaneous mutation was to arise in any of the components. However, if we think in terms of engineering (or synthetic biology), having fail safe mechanisms in a machine makes it safer; despite they may stop working. The solution is to add more fail safe devices. Identically, adding more synthetic speciation barriers will increase safety. Other transcription factors and enhancers could be easily used to create those extra barriers, because the concept goes beyond any particular element. Importantly, modification of the binding properties, instead of complete elimination of the transcription factor, could also be implemented, reducing the constraint of not finding enough non-essential transcription factors to build several barriers.
In summary, the synthetic species boundary described here can be reliably and predictably manipulated, allowing opening and closing speciation gates when desired, and isolates for the first time a transgenic animal from the original wild-type population.

Methods
High definition and depth of field photographs were obtained with a Keyence VHX-600 microscope. Flies were frozen at 220uC overnight before imaging. For SEM, adults were fixed in 2.5% glutaraldehyde in PBS overnight at 4uC, post-fixed in 1% osmium for 2 h at 4uC, washed, dehydrated in ethanol and with Hexamethyldisilazane until evaporation of the solvent. Samples were coated with 30 nm of gold and observed with a 440 Leica microscope under 20 kV tension.

Nomenclatural Acts
The electronic version of this document does not represent a published work according to the International Code of Zoological Nomenclature (ICZN), and hence the nomenclatural acts contained in the electronic version are not available under that Code from the electronic edition. Therefore, a separate edition of this document was produced by a method that assures numerous identical and durable copies, and those copies were simultaneously obtainable (from the publication date noted on the first page of this article) for the purpose of providing a public and permanent scientific record, in accordance with Article 8.1 of the Code. The separate print-only edition is available on request from PLoS by sending a request to PLoS ONE, Public Library of Science, 1160 Battery Street, Suite 100, San Francisco, CA 94111, USA along with a check for $10 (to cover printing and postage) payable to ''Public Library of Science''. The online version of this work is archived and available from the following digital repositories: PubMed Central, LOCKSS.