Part:BBa_J23103
constitutive promoter family member
Variant RFP (au) J23112 1 J23103 17 J23113 21 J23109 106 J23117 162 J23114 256 J23115 387 J23116 396 J23105 623 J23110 844 J23107 908 J23106 1185 J23108 1303 J23118 1429 J23111 1487 J23101 1791 J23104 1831 J23102 2179 J23100 2547 |
Constitutive promoter family
Parts J23100 through J23119 are a family of constitutive promoter parts isolated from a small combinatorial library. J23119 is the "consensus" promoter sequence and the strongest member of the family. All parts except J23119 are present in plasmid J61002. Part J23119 is present in pSB1A2. This places the RFP downstream of the promoter. Reported activities of the promoters are given as the relative fluorescence of these plasmids in strain TG1 grown in LB media to saturation. See part BBa_J61002 for details on their use.
These promoter parts can be used to tune the expression level of constitutively expressed parts. The NheI and AvrII restriction sites present within these promoter parts make them a scaffold for further modification. JCAraw
Sequence and Features
- 10COMPATIBLE WITH RFC[10]
- 12INCOMPATIBLE WITH RFC[12]Illegal NheI site found at 7
Illegal NheI site found at 30 - 21COMPATIBLE WITH RFC[21]
- 23COMPATIBLE WITH RFC[23]
- 25COMPATIBLE WITH RFC[25]
- 1000COMPATIBLE WITH RFC[1000]
USTC_2009's MEASUREMENT
Note
Parts BBa_J23103 and BBa_J23112 are the same (Rahmi Lale).
Baltimore Biocrew 2019 Characterization
Goal
We, the Baltimore Biocrew, decided to characterize some of the Anderson promoters. These promoters are highly used by iGEM but the relative expression of these promoters have been routinely determined by measuring the fluorescence of a reporter protein. However, the function of a promoter is to start transcription of a gene so it may be more informative to measure the amount of RNA (instead of protein) produced by a reporter gene. Therefore, we decided to further characterize a selection of the Anderson promoters (J23100, J23101, J23103, J23105, J23118) by measuring RNA using reverse-transcription quantitative Polymerase Chain Reaction (qPCR).
Results
We set up samples for qPCR with three or four technical replicates per promoter and no template control samples to measure DNA contamination. We did data analysis using the Livak Method (a standard, comparative method) to determine the relative strength of the promoters from the qPCR data using rrSD as our reference gene, RFP as our target gene, and J23100 as our calibrator sample.
Example:
ΔCT(J23101) = CT(RFP, J23101) – CT(rrSD, J23101)
ΔΔCT(J23101) = ΔCT(J23101) – ΔCT(J23100)
2^(–ΔΔCT) = relative expression ratio
In our first trial of qPCR (8/03/19), we were able to measure the relative strengths for J23100, J23101, J23103, and J23105 which were 1.00, 0.00, 0.81, and 1.93, respectively. Since these strengths did not match the relative expression levels reported by iGEM2006_Berkeley from protein level measurements, we repeated the qPCR (8/10/19) with the same cDNA. The strengths from this second trial were 1.00, 0.00, 0.37, and 0.20. We repeated it again and the relative strengths that we got on 10/12/19 for J23100, J23101, J23103, and J23103 were 1, 0, 2.91, and .32. Next, we made new cDNA by growing new liquid cultures, extracting RNA again, and repeating reverse transcription. From the new cDNA, we repeated the qPCR procedure two more times. The relative strengths for that we got on 9/28/19 for J23100, J23101, J23103, J23105, and J23118 were 1, 24.63, .36, 1.76, and .25. The relative strengths that we got on 10/12/19 for J23100, J23101, J23103, and J23105 were 1, 45.97, 3.20, and 1.26. In addition we measured promoter J23118 twice and got the strengths 1.13 and 1.32.
Here is the relative promoter strengths that we got from the qPCR. Baltimore BioCrew measurements in orange compared to the 2006 Berkeley iGEM, who determined relative strengths from measurements at the protein level, in blue.
To support our RNA measurements we also measured fluorescence of the liquid cultures we used to extract RNA. The cultures were grown overnight so we expected the bacteria to be at the stationary phase, but we measured OD to normalize any differences in growth.
Promoter | OD | fluorescence | fluorescence divided by OD | corrected relative expression | reported relative expression |
---|---|---|---|---|---|
BBa_J23100 | 0.876 | 250 | 285.38 | 1 | 1 |
BBa_J23101 | 0.674 | 255 | 378.33 | 1.33 | 0.7 |
BBa_J23103 | 1.1 | 230 | 209.09 | 0.73 | 0.01 |
BBa_J23105 | 1.08 | 215.74 | 209.09 | 0.76 | 0.24 |
BBa_J23118 | 1.04 | 238 | 228.84 | 0.80 | 0.56 |
After redoing our protocol many times and trying to compare our results to 2006 Berkeley iGEM team we concluded that our data doesn’t quite match theirs. The different results in strengths could be caused by many different factors while doing our protocols. However, we have succeeded in characterizing the different strength Andersons promoters (J23100, J23101, J23103, J23105, J23118) by measuring RNA using Quantitative Polymerase Chain Reaction (qPCR). In conclusion, we successfully reached our goal in bringing new data to the characterizations of 5 different Andersons Promoters. In the future, it would be good for other iGEM teams to try to measure RNA as well so there will be a standard qPCR protocol for iGEM.
To see more detailed methods and our protocols, visit our wiki
Added by KEYSTONE_A 2020 Team
J23103 can be used as constitutive promoter in bacterial cellulose producing strain K. rhaeticus iGEM, which is of very low expression level, as it performs in E. coli.
Figure above: Constitutive promoter average strengths in K. rhaeticus iGEM and E. coli, normalized against J23104. Although all promoters are functional, their relative strengths differ between K. rhaeticus and E. coli. For K. rhaeticus, data is shown as grey bars, with standard deviation of N=3 biological replicates, characterized in liquid HS-medium containing cellulase, measured 3 h post-inoculation. Relative promoter strengths in E. coli are superimposed as black stripes.
References:
Florea, M., Hagemann, H., Santosa, G., Abbott, J., Micklem, C. N., Spencer-Milnes, X., ... & Chughtai, H. (2016). Engineering control of bacterial cellulose production using a genetic toolkit and a new cellulose-producing strain. Proceedings of the National Academy of Sciences, 113(24), E3431-E3440.
•••••
University of Texas at Austin iGEM 2019 |
UT Austin iGEM 2019: Characterization of metabolic burden of the Anderson SeriesDescriptionThe 2019 UT Austin iGEM team transformed the Anderson Series promoters into our 'burden monitor' DH10B strain of E. coli, which contains a constitutive GFP cassette in the genome of the cell. GFP expression fluctuates depending on the number of ribosomes available. Using this strain, we characterized the relative burden (percent reduction in growth rate) of each Anderson Series part. Our results showed a range of growth rate reductions for each of these parts due to ribosomal reallocation from the genome of the host cell, towards the expression of RFP. Anderson Series parts with strong promoters are depicted with darker red colors and Anderson Series parts with weak promoters are depicted with lighter pink colors to show relative RFP expression. We saw a positive correlation between relative promoter strength and metabolic burden; parts with stronger promoters expressed less GFP and had a lower growth rate than parts with weaker promoters. The regression line for the graph below was constructed by measuring the burden of 5 parts that were created by the 2019 UT Austin iGEM team that each contained an Anderson Series promoter (BBa_J23104 or BBa_J23110), an RBS of varying strength, and a BFP reporter. For more information on characterization of these parts through the burden monitor, visit our team’s wiki page: [1]
Importance of Characterizing BurdenAlthough often we cannot avoid using a specific burdensome part, knowing in advance that it is burdensome, and that it has a high chance of mutating into a non-functional genetic device, can help with troubleshooting and coming up with alternatives. In the specific case of fluorescent protein-expressing devices, Fluorescence-activated cell sorting (FACS) can be used to filter out individual cells that meet a certain fluorescence threshold. This way, the cells expressing lower levels of the fluorescent protein are weeded out of the population. |
//direction/forward
//promoter/anderson
//regulation/constitutive
//rnap/prokaryote/ecoli/sigma70
negative_regulators | |
positive_regulators |