In Vivo
No helper plasmid
We designed plasmid inserts where each of our switch designs was correctly placed to control expression of the far-red fluorescent protein mKate2. The switch-mKate2 transcript was expressed from the pTrc promoter which can be regulated by IPTG in the presence of the repressor lacI.
On the same plasmid the trigger sequence for the switch was expressed from the pBAD promoter fused to the RNA aptamer Broccoli. Broccoli can bind to the chemical DFHBI to give green fluorescence. Transcription from the pBAD promoter can be induced by arabinose in the presence of the regulator araC.
These inserts were ordered from Twist bioscience in kanamycin-resistance plasmids with ColE1 origins.
We intended to use our positive control part, that lacks a switch sequence, to find an intermediate level of mKate2 expression by testing a range of IPTG concentrations. We would then use this IPTG concentration with our other switches and measure mKate2 expression whilst varying the concentration of arabinose. Arabinose which should induce trigger expression. If our triggers operate correctly then varying arabinose concentration should affect the structure of the switch and therefore the mKate2 translation rate.
The ratio of mKate:Broccoli fluorescence per cell can be used as a measure of cooperativity to see if our switches are cooperative in their response to trigger concentration.
Trigger RNAs
mKate2 mRNA
Riboswitch
Triggers bind to riboswitch to affect translation
Translation
Transcription
Transcription
DNA
mKate2 gene
Trigger sequence
pBAD promoter
lac promoter
mKate2 protein
Riboswitch
RNA
Protein
We intended to assemble a plasmid with a compatible origin expressing both lacI and araC but lacked the time or components to carry this out. Instead we first tested a selection of our parts using only background expression of lacI and araC from the genomic copies of these genes. We compared our Homo sym 8, single binding site, Homo sym long and AABB repressive switches to untransformed E. coli DH5⍺ cells.
Cells were grown in LB with kanamycin for approximately 5 hours at 37°C with and without 1g/L arabinose, which should be more than sufficient for maximum expression. 100uL of these populations were then tested in a plate reader measuring mKate2 fluorescence (Ex: 588nm; Em: 633nm) and the optical density at 700nm, a measure of cell growth. (OD600, the standard wavelength for monitoring cell growth was not used as it is close to the mKate2 excitation wavelength.) The results are expressed as mKate2 fluorescence divided by OD which should be proportional to fluorescence per cell.
Although the different switches differed in their mKate fluorescence level these is no significant difference between populations exposed to arabinose and populations without arabinose.
This could be due to three reasons:
-
Our triggers are not being expressed
-
Our triggers are not folding correctly or are rapidly degraded
-
Our triggers do not open our switches as we intended
Using the techniques we had immediately available we could not distinguish between the first two hypotheses (this could be done for instance by a Northern blot if we had a labelled probe to our switch sequence).
We tested the cells for Broccoli fluorescence by resuspending them in PBS with 0, 40 and 80uM DFHBI but found no difference between induced and non-induced cells. This may indicate that our cells are not expressing the correctly folded trigger sequence though we did not have time to optimise this protocol fully.
Pelleted E. coli cells grown for six hours in 5ml cultures with and without arabinose
From left to right:
NC induced+noninduced
Homo Sym 8 induced+noninduced
Long induced+noninduced
Single binding site induced+noninduced
AABB induced+noninduced
pBad33 helper plasmid
We then obtained a copy of the expression plasmid pBAD33 from the Oxford iGEM team. This plasmid has a compatible origin to our switch plasmids and expresses araC. We cotransformed E. coli DH5⍺ cells with pBAD33 and each of our test plasmids and tested mKate2 expression with and without arabinose again.
(Mean of four independent colonies, error bars show the population standard deviation)
Unfortunately we still did not see significant difference in mKate2 expression between the induced and uninduced cells for any of switches indicating that our switches still do not operate as we would like.
One hypothesis is that the increase in the concentration of the regulator araC, expressed from the low-copy number pBAD33, is still insufficient to activate trigger expression from the pBAD promoter on our high-copy test plasmid due the difference between the copy numbers (approximately 15 per cell vs. approximately 700 per cell). Alternatively our triggers may still be folding incorrectly or not binding to our switches as we envisaged.
Further testing with different constructs would be required to resolve these issues.
The expression data however appears to agree with the Salis RBS calculator predictions that increasing the length of the stem, i.e. from 8 to 12 to 18 to 21, in the translationally inactive state of our activatory switches should lower the basal translation rate from our designs. Tuning the basal translation rate of a gene, even in the absence of switching in response to trigger, could be an application of our designs.
Introduction
Conclusion and future work
Although no switching could be observed, either with or without a helper plasmid expressing araC, the results from our in vivo testing are promising in that they confirm that mRNA structures deliberately placed in the 5' UTR of a gene around the RBS can affect translation rate.
However more work is required to satisfactorily test our switches in vivo. The most crucial aspect to investigate would be whether our trigger sequence as being transcribed and if they are being expressed, whether they fold correctly. This could potentially be answer by Northern blotting.
If the trigger sequences are not being transcribe sufficiently then time-consuming work may need to be done to clone araC onto the same plasmid as our test constructs to ensure appropriate regulator expression.
If the triggers are being expressed then they may not be seen using Broccoli fluorescence if broccoli does not fold correctly or if Broccoli fluorescence is too dim to be seen with our setup. Screening could be used to find better-folding constructs by changing the length and sequences of the linkers between our trigger, F30(Bro) and terminator. Brighter versions of Broccoli could be used for instance using four Broccoli repeats: F30(2xdBro) or the brighter red-shifted fluorophore-binding RNA aptamer: Mango.
If this is done then flow cytometry could be used to quantify mKate2 expression and trigger concentration on a single cell basis.
