Expressional artifact caused by a co-injection marker rol-6 in C. elegans

In transgenic research, selection markers have greatly facilitated the generation of transgenic animals. A prerequisite for a suitable selection marker to be used along with a test gene of interest is that the marker should not affect the phenotype of interest in transformed animals. One of the most common selection markers used in C. elegans transgenic approaches is the rol-6 co-injection marker, which induces a behavioral roller phenotype due to a cuticle defect but is not known to have other side effects. However, we found that the rol-6 co-injection marker can cause expression of GFP in the test sequence in a male-specific interneuron called CP09. We found that the rol-6 gene sequence included in the marker plasmid is responsible for this unwanted expression. Accordingly, the use of the rol-6 co-injection marker is not recommended when researchers intend to examine precise expression or perform functional studies especially targeting male C. elegans neurons. The rol-6 sequence region we identified can be used to drive a specific expression in CP09 neuron for future research.


Introduction
Efficient transgenic techniques are used in various model systems to detect gene expression and assess genetic function. In the nematode Caenorhabditis elegans, for example, gene expression can be monitored using transgenic worms generated by a simple, gonadal microinjection of a plasmid that drives GFP expression under the control of a promoter for a gene of interest [1]. During the course of the DNA transformation procedure, one easy way to select transformed animals is by using easily-detected co-injection markers. In C. elegans, several co-injection markers are commonly used, which include visible fluorescent markers (e.g. ttx-3p::GFP, myo-2p::mCherry) [2,3] and rescuing markers that restore lethal or non-lethal phenotypes (e.g. pha-1, unc-119, dpy-5) [4][5][6]. One type of dominant selectable marker, rol-6(su1006), is widely used, because it shows a dominant roller phenotype that is easily observed and can be used in a wild type background [7,8].
A prerequisite for the use of co-injection markers is that the phenotype induced by the coinjection marker must not interfere with expression or scoring of the gene being tested. In this a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 study, we report that the widely-used rol-6 marker unexpectedly activates the test gene in a male interneuron called CP09 in C. elegans. This unwanted expression could potentially result in misidentification of cell types in a gene expression study as well as affect the results of functional studies that utilize rol-6 as a co-injection marker.

Results and discussion
During the course of experiments to determine the tissue-specific expression pattern of 10 putative synaptic genes in C. elegans, we generated transgenic worms with promoter::GFP test genes using rol-6(su1006) as a co-injection marker (for information on the genes and the promoter fragments assayed, see S1 Table). We noticed that in eight out of 10 transgenic lines, GFP was expressed in the CP09 neuron among other diverse neurons [9]. CP09 is a male-specific interneuron located in the pre-anal ganglion of the male tail that forms chemical and electrical synapses with many other male-specific or sex-shared neurons (Fig 1). Male C. elegans have 10 CP ventral cord neurons (CP00~CP09) [10]. The CP neurons are believed to have similar properties due to a similar developmental origin, in which all CPs are generated from Pn. aapp cells, but some CP neurons are reported to use different neurotransmitters [10][11][12][13]. Interestingly, out of the eight transgenic lines showing CP09 expression, four of the transgenes were expressed in most or many neurons (i.e. many CP neurons), but the other four were expressed exclusively in CP09 among the 10 CP neurons. Therefore, we suspected that the CP09 expression may be an expression artifact.
It is widely known that GFP reporters driven by diverse promoters often show artificial fluorescence in posterior gut cells, in several muscle cells, and even in one neuron called PVT [15,16]. One potential cause of these artifacts was suggested to be an effect of the unc-54 3 0 UTR, which is attached to the GFP coding sequence in most C. elegans vectors [16]. To test whether the unc-54 3 0 UTR can also cause expression in CP09, we replaced the unc-54 3 0 UTR  with the let-858 3 0 UTR in a promoter::GFP fusion for one of the test genes that showed exclusive expression in CP09 among the 10 CP neurons (T19A6.4 gene). When transgenic animals were generated by microinjection of the T19A6.4p::GFP::let-858 3' UTR fusion along with the rol-6 co-injection marker, they still showed CP09 expression, suggesting that at least for this gene factors other than the unc-54 3 0 UTR are likely involved in the generation of the CP09 signal.
The second possibility was that the rol-6 co-injection marker used in the microinjection procedure caused the expression in CP09. To test this idea, we injected an empty GFP vector (pPD95.75), which contains no promoter for the GFP coding sequence, together with the rol-6 co-injection marker (pRF4). The resulting transgenic animal showed a robust GFP expression in CP09 (Fig 2A). However, when the empty GFP vector was injected with another co-injection marker ttx-3p::GFP (expressed in AIY neuron in the head), the CP09 signal was not observed (Fig 2B). Thus, the rol-6 co-injection marker itself can promote transcription in the CP09 neuron.
Homologous recombination between co-injected DNA molecules contributes to the formation of stable extrachromosomal arrays [8]. Most C. elegans vectors have a backbone based on the pUC19 plasmid, and thus have high sequence similarities that are potentially utilized for homologous recombination during extrachromosomal arrays formation. For example, the pRF4 and pPD95.75 plasmids share~2.5 kb sequences that include the E.coli ampicilin resistance gene and origin of replication. We also found a 40 bp homology shared between the pRF4 plasmid and GFP constructs generated by promoter::GFP fusion PCR, which constitutes the multiple cloning site of the vectors (minimum homology with GFP constructs) (Fig 3A).

Co-injection marker-induced artifactual expression
The minimum homology sequence is located upstream of the GFP coding region of pPD95.75 and included in most promoter::GFP fusion constructs, as this inclusion ensures that an artificial intron is placed in front of the GFP sequence for efficient reporter gene expression [16]. The full sequence information of pRF4 is available in S1 Text. (Although pRF4 has been used widely in the C. elegans research community, the accurate pRF4 sequence, to our knowledge, is not yet available in public.) To find a region responsible for CP09 expression, we divided the pRF4 plasmid except for the minimum homology region into two fragments, namely "rol-6" and "vector" fragments, and cloned these into the empty GFP vector pPD95.75 (Fig 3A). When the rol-6 fragment::GFP was injected, the resulting transgenic animal showed a robust GFP expression in CP09 (Fig 3B). However, we could not observe any GFP expression in CP09 when the vector fragment::GFP was used for injection ( Fig 3B). Therefore, it is likely that when a GFP construct is injected together with the rol-6 co-injection marker pRF4, the rol-6 fragment fused to GFP by homologous recombination generates unwanted transcription and GFP expression in CP09 (Fig 3C).
Our results raise an obvious problem in using the rol-6 co-injection marker for gene expression or functional studies especially on the C. elegans male, as this marker can induce unwanted expression in the male CP09 neuron. The rol-6 fragment of pRF4 likely contains a driver sequence that triggers CP09 expression. For expression studies using the rol-6 co-injection marker, any CP09 expression needs to be double-checked by using another type of coinjection marker. For functional studies, it should be determined whether the use of the rol-6 marker affected interpretation of the results. For example, techniques called GRASP (GFP reconstitution across synaptic partners) and iBLINC (in vivo biotin labelling of intercellular contacts) have been developed to visualize synapses formed between specific pairs of neurons that are defined by cell-specific drivers [17,18]. Since CP09 has many synaptic connections with other male-specific and sex-shared neurons (see Fig 1), artifactual expression in CP09 can potentially generate additional synapse signal when using GRASP or iBLINC. We recommend not using the rol-6 marker if studies are designed to examine gene expression or function in the male tail of C. elegans.
To avoid the unwanted CP09 expression, homologous recombination between the rol-6 coinjection marker and any expression constructs may be minimized by reducing their sequence homologies. For example, fusion PCR-based promoter::GFP constructs share a minimum 40 bp homology with the pRF4 plasmid. If a promoter::GFP construct is designed to omit the homologous sequences, it may be possible to suppress CP09 expression caused by the rol-6 coinjection marker. However, it will be difficult to test this idea using plasmid-based GFP constructs, as most C. elegans plasmids share a backbone and usually have a high sequence homology [16].
Cell-specific promotors or drivers are invaluable tools for transgenic research, because they allow us to confine gene expression to subsets of cells or even to a specific cell. Several such drivers have been identified and used extensively in the C. elegans community [19]. In this study, we identified that the rol-6 fragment of pRF4 drives expression in the male-specific CP09 neuron. This sequence can be used as a CP09-specific driver for future research.

C. elegans maintenance
CB4088 him-5(e1490) worms were used as the wild-type reference strain to generate worm populations containing large numbers of males. Worms were grown at 20˚C on standard nematode growth media (NGM) plates with OP50 E. coli as a food source and maintained according to standard methods [20].

Microscopy
Worms were prepared and imaged as described previously [22]. Briefly, 1-day-old males were mounted on 5% agar pads on glass slides using 10~50 mM sodium azide. Worms were observed with fluorescence microscopy (Zeiss Axio Imager.Z2) or confocal microscopy (Nikon Eclipse Ti). Images were processed using AxioVision (Zeiss) or NIS-Elements (Nikon). Figures were prepared using ImageJ software.