The variation of promoter strength in different gene contexts

Background Promoter engineering has been employed as a strategy to enhance and optimize the production of bio-products. There have been many effortless studies searching the best promoter for biological application. However, whether promoter strengths stay unchanged in different gene contexts remains unknown. Results Six consecutive promoters at different strength levels were used to construct six different versions of plasmid backbone pTH1227, followed by inserted genes encoding two polymer-producing enzymes. Some of promoter strengths in the presence of inserted sequences did not correspond to the reported strengths in a previous study. When removing the inserted sequences, the strengths of these promoters returned to their reported strengths. These changes were further confirmed to occur at transcriptional levels. Polymer production using our newly constructed plasmids showed polymer accumulation levels relatively corresponding to the promoter strengths reported in our study. Conclusion Our study revealed the essence of re-assessing promoter strength in a specific gene context. Different gene contexts could result in the variation of promoter strengths, hence this might lead to different outcomes in downstream applications.


Background
Promoter engineering has been considered as one of the critical strategies in the metabolic engineering field. A number of studies have been carried out to predict and characterize the promoter strength using in silico and/or wet-lab experiments [1][2][3][4]. Strong promoters are normally desired to maximize the end products, which could be proteins, amino acids, polymers or biofuel, and reduce toxicity during the growth phase. However, promoter strength and optimal yield of end products are not always correlated proportionally; therefore, promoters that are finetunable and tightly controlled are more beneficial and versatile.
The broad-host range expression vector pTH1227, a pFUS derivative vector, contains an RK2 origin, inducible pTac-lacI q promoter and a reporter gusA gene [5][6][7][8][9]. This vector was used to study the role of the minCDE genes in both S. meliloti and E. coli [5]. The related vector pMP220 has been used in Azorhizobium caulinodans [6]. It was also used for insertional mutagenesis, transcriptional signal localization and gene regulation studies in root nodule bacteria such as S. meliloti, Bradyrhizobium sp. and R. leguminosarum bv. Viciae [8].
A library of engineered promoters of different strengths has been constructed through either mutagenesis of constitutive promoters or chemical synthesis of promoter variants [10][11][12]. By changing the length and the sequence of the spacer between the -35 and -10 regions, a broad range of promoters which differ in strength have been studied in Lactococcus lactis.
Randomizing the spacers resulted in a remarkable change in activities up to 400-fold [11]. It was also emphasized that the context in which the consensus sequences are embedded significantly influences the promoter strength. In a subsequent study, this strategy has been employed to obtain the expression of the pyrG gene encoding CTP synthase at different levels and its effect on the growth rate and nucleotide pool size [10]. Another library of nearly 200 promoters was also obtained using error-prone PCR to examine gene expression in E. coli [12]. It was suggested that the optimization of gene expression could also depend on the genetic background of the strain.
To study gene expression or the effect of a specific promoter on gene expression, a reporter gene located downstream of the promoter is normally used. There are various reporter genes which have been employed to detect gene expression used in different host backgrounds such as gfp gene encoding green florescence protein or gusA gene from E. coli encoding β -glucuronidase [1,2,[12][13][14]. The β -glucuronidase (Gus) product can be easily detected by observation of colour change on media containing chromogenic substrates such as X-GlcA (5-bromo-4-chloro-3indolyl β -D-glucuronide, cyclohexyl-ammonium salt), other 3-indoxyl derivatives, and naphtholβ -D-glucuronide. In addition, it also can be quantitatively measured by spectrophotometry (pnitrophenyl β -D-glucuronide and phenolphthalein-β-D-glucuronide), fluorimetry (4methylumbelliferyl-β-D-glucuronide and 5-dodecanoyl-aminofluorescein-di-β-D-glucuronide) or chemiluminescence (1,2-dioxetane-β-D-glucuronide). Gus has been widely used to study gene expression and other applications because the enzyme is highly stable, resistant and easily detected. The gfp gene was used to examine gene expression in E. coli of nearly 200 promoters created by error-prone PCR [12]. These promoter strengths were also assessed by testing the effect of the expression of the ppc gene encoding phosphoenol pyruvate (PEP) carboxylase on growth yield and encoding deoxy-xylulose-P synthase on lycopene production.
In our study, we employed a broad range of constitutive promoters of different strengths and reassessed their strengths in a specific context. Subsequently we investigated their effect on polymer production.

Strains and media used
All bacteria and plasmids were listed in Table 1. S. meliloti and E. coli strains were cultivated in Tryptone Yeast extract (TY) and Luria-Bertani (LB) media, repectively. Streptomycin (200 µg/ml) and tetracycline (10 µg/ml) were added if necessary.

GusA activity assay
This assay was carried out following the procedure as previously described [15]. Cells were fully grown in TY media, and then the culture was added into the assay buffer at the ratio of 1:4, incubating at room temperature until the mixture turned into a yellow color, at which point sodium carbonate was added to terminate the reaction. Reaction time and the absorbance of the mixture at 420 nm were recorded. Cell density of the culture was measured at 600nm for normalization.

RNA isolation
RNases are present everywhere and very active, hence extra care must be taken to avoid sample degradation by using gloves and RNase-free materials. All solutions are prepared in RNase-free water or DEPC-treated water. Cells were sub-cultured in TY and grown to OD = 0.5 -0.8. Stop solution which is 5% phenol in ethanol was added into the culture and mixed well to stop the cell growth. Next, culture was harvested by centrifuging at 5,000 rpm, 4°C for 10 min.
Cells can be stored at -70°C after flash freezing in liquid nitrogen.
RNA extraction was performed following the hot phenol extraction protocol with some modification [16]. All the following steps should be carried out at 4°C, unless otherwise stated.
The cell pellet from a 50 ml culture was thawed on ice and re-suspended in 960 µl of RNase-free water by vortexing. Next, 480 µl of hot phenol solution was added to an equal volume of cell resuspension, vortexed vigorously and incubated in a water bath at 95°C for 1 min. The sample was spun down at 13000 rpm, 4°C for 10 min. Supernatant was added to 600 µl phenol/chloroform, and vortexed vigorously. Then, it was spun again for 5 mins, and aqueous phase was extracted twice with chloroform. To precipitate nucleic acid, the aqueous phase was added to 1/10 volume of sodium acetate and 2 volume of isopropyl alcohol, inverted to mix and incubated on ice for 30 min. Next, the sample was spun for 10min, washed with -20ºC 70% ethanol and air-dried until the pellet turned translucent. Precipitate was then resuspended in 85 µl water and digested with 5 µl DNase I in 10 µl DNase I buffer at 37°C for 30 min to remove DNA from the sample. The DNA-free sample was confirmed by running on an agarose gel. If DNA is still present in the sample, water was added into the sample up to the volume of 400 µl, and nucleic acid precipitation was repeated as described above, followed by another round of DNA digestion. The RNA sample was further purified using Qiagen RNeasy mini kit, and finally resuspended in 100 µl water. The quantification of RNA was determined using the Nanodrop, and the ratio of A 260 /A 280 should be around 2.0. The quality of RNA was checked by imaging the agarose-formaldehyde gel to observe rRNA bands.

Dot blot
Samples with equal amount of RNA were applied onto a positively charged membrane using the Bio-Dot Microfiltration Apparatus (Bio-Rad). The membrane was placed on the gasket and wetted with 2x SSC. Saran wrap was used to cover unused wells. RNA samples were blotted to each well, followed by washing with 200 µl TE. Vacuum was used to help liquid pass through the membrane. Membrane fixation was performed using the UV-crosslinker instrument, using C-L program. Prehybridization was performed in DIG Easy Hyb buffer (Roche) at 40ºC for 30-45 min, followed by hybridization overnight at the same temperature. Washing and detection steps were carried out as outlined in the manufacturer's protocol for the DIG High Prime DNA Labeling and Detection Starter Kit II (Roche).

Polyhydroxybutyrate (PHB) analysis
Polymer production was evaluated by gas chromatography following a protocol which Engineering. The oven program was set as following: initial temperature was set at 80ºC for 5 min, then ramped to 230ºC at 7.5ºC/min, and continued to ramp to 260ºC at a faster rate 10ºC/min followed by maintaining that temperature for 5 min.

Construction of plasmids with promoters of different strengths
A range of promoters of different strengths (6 out of 12 native promoters) based on a previous study [2] was chosen for further investigation. These promoter sequences (Table 2) were synthesized in the form of extended primers that were annealed to create double-strand DNAs. The backbone plasmid, pTH1227a, is a derivative of pTH1227 that had the XbaI site removed by being digested, blunted and self-ligated as shown in Fig. 1. Promoter P 0 was designed to contain XbaI site next to HindIII site and cloned into pTH1227a at the HindIII and XhoI sites, creating pTH1227b. Next, the synthesized fragment which contains two engineered genes were inserted into the pTH1227b at the XhoI and PstI sites. The rest of the promoter set containing BglII for post-cloning confirmation was cloned into XbaI and XhoI, replacing  (Fig. 3). The rpoD promoter still maintained the low activity which was observed with and without the inserted genes.

Analysis of transcript abundance of gusA and pct genes using dot blot
From the above results, the presence of synthesized genes that changed genetic context has shown the effect on the activity level of some promoters which were indirectly assessed by measuring gusA activity of the reporter gene. As shown in Supplementary 1 and Fig. 4, the abundance of gusA transcripts corresponded with the GusA activity, suggesting that the presence of synthesized genes caused the decrease and increase of gusA transcripts for pTAM2 and pTAM7 plasmids, respectively. The abundance of pct transcripts was also investigated and compared to that of gusA transcripts on the plasmids carrying the synthesized genes (Supplementary 2). As our expectation, the abundance of pct transcripts is similar to that of gusA transcripts since they are transcriptionally fused.

Polymer production in strains that have synthesized genes expressed under these constitutive promoters
To evaluate the different expression profiles on polymer production, these strains were cultivated in define media supplemented with mannitol as a substrate and grown under polymer accumulating conditions. As expected, strains carrying pTAM6 and pTAM7, which had the weakest and strongest promoter strengths produced the least and most amount of PHB, respectively (Fig. 5). The strains carrying pTAM7 which showed promoter strength similar to that of the strain carrying the tac promoter under induction conditions also accumulated a similar amount of PHB. Other strains showing a range of intermediate promoter strengths produced similarly low amount of polymer.

Discussion
The transcriptional levels of some promoters in our study were completely contradictory to previously reported values. For example, we found that the pTAM7 plasmid carrying the ropB1 promoter exhibited the highest levels of expression, while MacLellan et al [2] had previously demonstrated that this same promoter had the lowest level of expression among the same set of promoters. Another promoter that strongly disagreed with MacLellan et al in terms of gene expression level was the Smc1378 gene promoter in the pTAM2 plasmid. In the previous study, expression was the highest of all of the promoters, but we found that the expression was relatively low compared to other promoters in the same set. This finding raised concerns about potential context effects on gene expression. It could be due to plasmid backbone, inserted gene sequence, the reporter gene employed or the cultivation condition. In a previous study, the relative promoter activity has been proved to remain unchanged across different cultivation conditions [17]. The activities of roughly 900 S. cerevisiae and 1800 E. coli promoters are taken into consideration, and their gene expressions changed among different cultivation conditions by a constant factor. In other words, they mostly behaved alike and maintained their relative activities levels across different cultivation conditions. Therefore, the cultivation condition could be ruled out as a possible cause. Among the remaining potential causes, we postulated that inserted gene sequence most probably caused the irrelative change in gene expression of some promoters. Therefore, we removed this inserted sequence from the plasmid construct and compared the GusA activity of the construct with and without these inserted genes.
As our postulation, removing the insert bought back the promoter strength reported in that previous study. This suggested that the downstream sequence probably influenced the transcription process. Other work has shown that the downstream region can have a strong influence on the efficiency of the escape process of RNA polymerase because release of the σ subunit only occurs after the polymerase has transcribed 8-11 nucleotides [18,19]. This region was earlier reported by Bujard and co-workers [20]. They found that the down-stream sequence could change the promoter strength in vivo more than 10-fold. The reason is that RNA polymerase covers a larger region than 35 bp (up to 70 bp); hence, its activity also depends on the flanking regions which could be up-stream or down-stream regions. It was known that this region was involved in promoter escape; however, the mechanism of this process was not understood. By further study of the anti initial transcription sequence (ITS) which was discovered earlier to have an effect on promoter strength, it was found that the function of the anti ITS did not depend on either the stability of RNA:DNA bond or the interaction with core RNA polymerase. It appeared that the function of anti ITS was related to the σ subunit. As a result, the interactions between them could influence the promoter escape. In other studies, it was also observed that characterized promoters often showed variable activities depending on the genetic locus or gene transcribed [11,12,[21][22][23]. A library of variable-strength, constitutive promoters was designed and constructed in bacteria [4]. The length of these promoters is 160 bp including extended sequences at both the 5'-and 3'-ends, so-called insulation sequences. Their promoter strengths were shown to maintain constant relative levels and be independent of genetic contexts. From the above evidence, we reason to suspect that the initial sequence of roughly 20 bp in the synthesized gene fragment may have had some effect on promoter strengths.
Since the promoter activities were indirectly assessed through the expression of gusA reporter gene, there is no clear evidence to prove that these observations happened at the transcriptional level, not translational level. Therefore, we continued to investigate the abundance of RNA transcripts of pct and gusA genes in the pTAM2 and pTAM7 plasmids to verify these observations occurring at the transcriptional level.
The abundance of gusA and pct transcripts were directly proportional to GusA activity. These results support the hypothesis that the downstream sequence influenced on the transcription process which led to the change in the abundance of transcripts and enzyme activity as a consequence. However, which sequence of the downstream region and how far is it from TSS are still not identified. As mentioned earlier, the first 20 amino acids following the TSS might play an important role in either facilitating or impeding the transcriptional process. Therefore, it would be interesting to further investigate these sequences in the future. There have been a number of studies focusing on the effect of the -10 and -35 regions on promoter strength.
However, only the upstream region before the transcription start site (TSS) has been seriously taken into consideration. In our study, it has been shown experimentally that this adverse effect exists. In the presence of engineered genes, relative promoter strength is different from what was reported. Removal of the insert fragment restored their relative strengths consistently across different reporter genes. We also demonstrated that the change occurred at the transcriptional level of both insert fragment and reporter gene.

Conclusions
We selected a range of promoters of different strength that were well-characterized in the literature. However, our study did not agree with previous study, in which the relative promoter strengths have varied not correspondingly. We found that the inserted gene located close to promoters had a significant effect on promoter strengths. We postulated that this effect response may be due to effects on the occurrence during the releasing process of RNA polymerase from the σ subunit.