Note: this is a search engine friendly version of my lab notebook, please see the pdf version of this document for a more human friendly, printer friendly version.

Chapter 9
Dabbling in synthetic biology

THIS CHAPTER/PROJECT IS NOT ACTIVE
See the first figure for a pictural overview. See the last page for a summary of the results.
Make toggle with components from systems biology predictions.
Please see the pdf version for figures
Figure 9.1: The basic schema for applying network inference algorithms to synthetic biology problems

9.1  testing pIKE 107

can I switch it? need to order tetracycline

9.2  making a pyruvate / lactose regulated switch

Instead of tetracycline the new switch (if it works) will use pyruvate which derepresses PdhR protein (see Figure ).
Please see the pdf version for figures
Figure 9.2: The potential toggle built using network inference information will replace the tetR gene from Tim Gardner's pIKE107 toggle with pdhR protein. One of the new toggles will use the known repressor site of the PdhR protein that resides prior to the pdhR-aceE operon. The other new toggle will use the new predicted target of pdhR, verified by ChIP: ndh.

9.2.1  Cloning in ndh and pdhR promoters

Thu Sep 13, 2006
Ilaria already has working protocols for sticking these promoters into vectors, so for the most part I'm just copying what she's done already but with new genes/promoters. On the pIKE vector, I'm switching out the pLtet promoter and the tetR genes with pdhR/ndh promoters and pdhR gene respectively. The pdhR promoter is a known target of the pdhR repressor; ndh is a new one that I verified in the chip studies. In this section I'm trying to swap the pLtet promoter with a pdhR one and a ndh one (i.e. to make 2 separate vectors). Then I'll add pdhR to each and see if I can make a pyruvate/IPTG switch 35.

PCR the gene from the genome

I amplified each promoter from genomic DNA36. The primer pairs each added a site for SphI at the 5' end and AatII at the 3' end.
I ran 30 ml PCR reactions using the Easy-A master mix [Stratagene]. The MT for the first 5 cycles was 56 and was then ramped up to 62 for the last 25 cycles (once there were enough sequences with the extra 10 bp or so added by my primers). Yields were unimpressive and post PCR cleanup DNA was pretty dirty. But Ilaria's done this 100x and said all of these things were normal:
Sample DNA (ng/ul) 260/280 260/230 total yield
ndh promoter 45.1 1.4 mg
pdhR promoter 43.4 1.3 mg
pdhR gene 62.1 1.9 mg
Please see the pdf version for figures
Figure 9.3: 1% agarose run for 75 minutes at 120 V. I show the gel after the chunk was remove for purification by razor blade because it was hard to see the PCR fragments while it was there.

Digesting the vector and the inserts

Sep 13, 2006
I minipreped pIKE107 from an overnite culture. Yield:
Sample DNA (ng/ul) 260/280 260/230 total yield
pIKE107 139.9 7.0 mg
I digested 15 ml of vector for 1 hour in a 20 ml double digest with 0.5 ml of each cutter (AatII SphI). N ml of the PCR products were also digested in a similar manner, but for 45 minutes.
The PCR'd insert digestions were purified with a Qiagen PCR purification kit. Yields were crappy as was the nanodrop spec reading. Ilaria said this was all normal, not to worry.
Sample DNA (ng/ul) 260/280 260/230 total yield
digested, PCR purified ndh promoter 28.9 867 ng
digested, PCR purified pdhR promoter 7.7 231 ng
The vector was gel purified (to get rid of the insert). Yield and spec readings for this also sucked. Again Ilaria said it was normal.
See Figure 9.3 for the inserts (uncut) and cut vector on a gel. I ran the pdhR gene on the gel too just to test the primers.
Sample DNA (ng/ul) 260/280 260/230 total yield
digested, gel purified pIKE107 16.2 366 ng

Ligation and transformation of ndh and pdhR promoters

I made 3 ligations. ndh prom + vector, pdhR prom + vector, vector alone (- control). For each I used 10 ml of vector. For ndh prom and pdhR prom, I used 2 ml of the cleaned PCR product. The ligation was run at 16 C for 30 minutes. Followed by the standard ligation proceedure. I plated 75 ml of each transformation (too much).
There were MANY colonies for the two samples and very few on the negative control. I picked 3 colonies for each of the two promoters.

Checking for the proper insert size

I ran PCR reactions with a melting temperature of 57 C for each of the 6 miniprepped samples using 0.5 ml plasmid and 1 ml of 5 uM primer per rxn.
Please see the pdf version for figures
Figure 9.4: 1.5% agarose. I'm not sure why there is a smear below the correct sized band.
Brief Conclusions:   Each plasmid type had two out of three colonies with the correct insert length (see Figure 9.4.

Sequencing the inserts

The four plasmids with the insert lengths of the correct size were spec'd and sent out for sequencing.
Sample DNA (ng/ul) 260/280 260/230 total yield Primer Orientation for sequencing Sequence Correct
pIKE A 109.1 5.5 mg F Y
pIKE B 135.4 6.8 mg R
nIKE B 140.5 7.0 mg F Y
nIKE C 167.0 8.4 mg R
All of the sequence data and chromatagraphs from Agencourt can be found here.

9.2.2  Cloning the pdhR gene into the pdhR-prom and ndh-prom vectors

Thur Sep 21, 2006
I digested 15 ml (appx 2.3 mg ) of pIKE-a and nIKE-b with PstI, AgeI, buffer I, and BSA for 1 hr at 37 C followed by 80 C heat inactivation for 20 minutes. In the same manner, I digested 15 ml of pdhR gene PCR product to which I had added the appropriate restriction sites. The digested pdhR gene was cleaned with a Qiagen PCR purification kit and spec'd:
Sample DNA (ng/ul) 260/280 260/230 total yield
pdhR gene 18.6 558 ng
All 20 ml of the two plasmid digestions were run on a prestained Sybr Safe gel, imaged with the versadoc (using UV) for 1.4 seconds. And cut with a razor blade while viewing on the transilluminator (i.e. no UV).37 The gel slices were cleaned using a Qiagen gel cleanup kit (see Figure ).
10 ml 38 of each cleaned gel slice was used in the ligation reaction. For each ligation, 2 ml of digested, purified pdhR gene was used. A negative control was used for pIKE-a and nIKE-b, consisting of the cut plasmid without insert. Ligations were done for 30 minutes at 16 C with no ligase heat deactivation.
Please see the pdf version for figures
Figure 9.5: 1% agarose run for 85 minutes at 90 V.
The transformation was done as normal except I used 350 ml of SOC (normally I use 250 ml ) and I only plated 50 ml .
Brief Conclusions:   There weren't very many colonies. Next time, I think I'll continue adding 350 ml but I'll plate 75 ml instead of 50 ml . There were enough colonies on each plate (5-20 or so) that I'm pretty sure they'll be some good ones. One negative control plate had one colony. The other had zero.

Checking the pdhR inserts in nIKE and pIKE

Sat Sep 23, 2006
Miniprepped 6 colonies (3 nIKE and 3 pIKE).
Sun Sep 24, 2006
I spec'd the 6 minipreps:
Sample DNA (ng/ul) 260/280 260/230 total yield
nIKE-pdhR A 167.9 8.4 mg
nIKE-pdhR B 160.9 8.0 mg
nIKE-pdhR C 182.7 9.1 mg
pIKE-pdhR A 175.9 8.8 mg
pIKE-pdhR B 117.7 5.9 mg
pIKE-pdhR C 103.0 5.2 mg
and digested 8 ml the colonies with AgeI and PstI 39.
Please see the pdf version for figures
Figure 9.6: 1% agarose run for 55 minutes at 110 V.
Brief Conclusions:   All six of the checked inserts were of the correct size (see Figure 9.6). I'm going to send 2 of each for sequencing as soon as the primers arrive. For now I think I'll use use them assuming they're correct to save time (ok, I know it will be a waste of time if I discover the inserts are actually wrong.)
Brief Update Sun Oct 29 17:32:10 EST 2006:   ndh-pdhR-a and pdh-pdhR-b were both fine (sequenced onthe 28th of September, 2006. pdh-pdhR-a was a little questionable raw sequencing data.

9.2.3  Ilaria's testing of the individual promoters in the U. Alon plasmid

Ilaria screened a lot of different conditions. All of this stuff can be found in her notebook. I want only to show the promoter responses for the final conditions used in the working attempt at switching the toggle (shown below).
Ilaria uses the U. Alon plasmid (Figure ) and a simple model to estimate gene expression from GFP time-series GFP measurements.
Please see the pdf version for figures
Figure 9.7: Promoters to be used for the toggle switch were tested individually across a variety of conditions using the Alon plasmid.
The estimates correspond pretty well to actually expression measurements obtained with qPCR (see Fig ).
Please see the pdf version for figures
Figure 9.8: Gene expression of lacZ estimated from GFP measurements corresponds well with directly measured expression using RT-qPCR.
Ilaria tested the promoters across many conditions; the ones relevant to the pdhR toggle are shown in Figure . She has many others tested conditions. All of the conditions without saturating glucose have a double hump where the cells first eat all of the available nutrients (first hump) and then finally eat the leftover pyruvate (second hump) (this is just the current guesstimate to why this happens. I felt the hump was causing the switch to continually go off with time so this high glucose way keeps a constant signal on the promoter (as long as you don't allow the cells to reach late stationary phase. One interesting result (not shown but in Ilaria's notebook). When I designed the pdhR promoter, I used the 200 bp upstream of the operon. When Ilaria made her initial pdhR promoter strain, she used 500 bp upstream. When she compared the 200bp to the 500bp, the 200bp promoter is around 2x stronger (image not shown see Ilaria's notebook or here Oct 2006 lab meeting slides).
Last, notice that the ndh promoter, although it was strongly enriched in the ChIP study only shows a weak increase in expression when pdhR promoter is derepressed. This might have more to do with the strength of the -10 and -35 than with the binding efficiency of pdhR protein to the ndh promoter. It could also be that ndh needs an activator to express fully.
Please see the pdf version for figures
Figure 9.9: Gene expression estimates for ndh and pdhR promoters.

9.2.4  Testing the pdhR version of the toggle

Mon Sep 25 12:43:38 EDT 2006
I now have two new versions of the toggle, one with a ndh promoter and one with a pdhR promoter. Ilaria has already tested that the 200 bp pdhR operon responds in to pyruvate in her gfp system. I'm a little worried because the lac repressor is so strong. Perhaps I should've used tetR instead of lac? We'll see how this goes. I'm pretty sure it's not going to be a problem to get it off. It's keeping it on at a decent amount that I'm unsure of. Ilaria is also building a pLpdhR promoter that should have more gain to it. I really would like a pLndh.

the plan of action for testing the new potential switch

I'm going to toggle the two new pdhR based switches and the old pIKE107 that I've already tested in the plate reader and I know works (see section 9.1). Based off what I learned in the previous test of the toggle here's what I'm going to do to test these guys:
  1. test nIKE-pdhR, pIKE-pdhR, nIKE, pIKE, and pIKE107
  2. toggle all strains off overnite using appropriate chemical (1% pyruvate for the nIKE and pIKE based plasmids and 500 ng/ml aTc for pIKE107; use LB media
  3. measure OD and fluorescence of the stationary phase culture
  4. wash 2x in LB and dilute 1/200 in LB
  5. measure OD and fluorescence until cells approach early stationary phase
  6. wash 2x in LB and dilute 1/1000 in LB 40 with appropriate chemical to turn GFP on (IPTG or lactose, I'd like to try both)
  7. measure OD and fluorescence of the stationary phase culture
  8. wash 2x in LB and dilute 1/200 in LB
  9. measure OD and fluorescence until cells approach early stationary phase

the plan I tried to take

The background fluorescence from LB was really messing things up. I tried to switch to a Davis minimal + casamino acids format. However, they just grew too slow.

the plan I actually took

I grew the cells in LB media with ampicillin. For the pdhR based plasmids, the LB also always had 2% glucose. Samples were taken every 45-120 minutes. For each sample the cells were washed 2x in Davis media with glucose. This removed the LB media and allowed small quantities of GFP to be measured free from the background fluorescent problems associated with LB. 200 mM IPTG was used to turn both switches on. The toggle on and toggle off states were both done overnite using a 1:1000 dilution. Prior to adding the appropriate switching chemical, the cells were washed 2x in LB. Only one sample was taken for these samples on the morning following the switching (i.e. I only get one datapoint and that point is taken at stationary phase).
For the stable on and stable off states, I sampled every 45-120 minutes until the cells reached early stationary phase. Low OD measurements, where GFP/OD was not accurately estimatable, were removed from the analysis. I turned Tim's pIKE107 as described in Tim's toggle paper in Science.
I expected that I would toggle the new toggle off with pyruvate and on with IPTG. However, this did not work. Using this method, the switch would not stably stay on. My guess for an explanation is that pretty much everything you can feed the bacteria (especially in rich media like LB) gets turned into pyruvate eventually by central metabolism. The pyruvate then turns off the switch. I noticed that this turning off appeared to roughly correspond to the peaks Ilaria saw witht the Uri Alon plasmid that yielded two humps in the expression level of ndh and pdhR promoters when grown on minimal media + pyruvate and casamino acids. Also like Tim's plasmid there was a basal level of expression even in the off state.
The method that actually worked was to allow the cells to reach late stationary phase when putting them into the off state with pyruvate + 2% glucose (by growing overnite and being lazy the next morning about getting started). Wash the cells 2x with Davis minamal and take your sweet time about it. The late stationary combined with washing with Davis ensures that gene expression is essentially halted and the memory manifest in the mRNA population degrades and is forgotten. If I then started the cells (diluting 1:500) in LB + glucose, the GFP was essentially off - hardly more than in a pUC19 strain that contained no GFP. To turn them on just wash 2x in LB dilute 1:1000 in LB + glucose and add 200 mM IPTG. To keep them stably on, don't allow the cells to reach late stationary phase, instead at late log phase (OD 0.8 or so) wash 2x in LB + glucose and dilute 1:500 into LB + glucose, the switch should stay on. Note the ON state and the stable ON states are much lower with this LB + glucose that I could obtain with out having the glucose around. However, the off state is much lower with the LB + glucose too. And when using the LB + glucose, the toggle will stay in the two states stably.
These results can be seen in Figure . Not the y-axis is log scaled. Tim's pIKE1107 toggle has muc expression than the new toggles. However both types of toggle have about the same order of magnitude difference between their off and on states.
Please see the pdf version for figures
Figure 9.10: toggling the toggle notice the y-axis are on a log scale
Brief Conclusions:   It seems like the toggle built using network inference results does stably toggle between two states. I presented these results at BMES in Chicago, Oct 2006. As a little part of a talk at the synthetic biology session. I used this to show how systems biology (and network inference in particular) can be used to do some synthetic biology (which currently reuses the same 3-4 promoters over and over and over and over again [cI, lac, tet, ara]. Unfortunately, I only ran this test one time, I'm not completely confident in these results and would be much confident if I could repeat this a time or two. However, I don't really have plans to do so. Toggling this thing is very time-consuming and I'm moving on to a new side-project with Ilaria working towards measuring GFP and gene expression in single-cells combined with sequencing to look at how changes in promoters cause changes in gene expression and protein production.