3.1. Polyethyleminie transfection assay
In order to establish the selectable system of promoter library, the transfection efficacy is important in this my experiment. DNA transfection technique was shown to deliver genes into various cell lines successfully, (Boussif, Zanta et al. 1996) its transfection efficiency is higher than other cationic polymers (Demeneix, Behr et al. 1998; Read, Singh et al. 2005).
In this experiment, Different doses of PEI/DNA complex were tried to transfect efficiently into various cell lines, including Balb/3T3 (murine fibroblast), 293 (human kidney cell), B16-F10 (murine melamona cell), A549 (human lung carcinoma cell), DBTRG-05MG (human glioblastoma cell) and HepG2 (human hepatocellular carcinoma cell) cell lines.
The efficiency rates of PEI transfection for Balb/3T3 and 293 cells were 64.95% and 60.92%, respectively (Figure 7). For A549 and B16-F10 cells, the efficiency rates were 26.58% and 25.84%, respectively (Figure 7, 8) However, (Figure 8), the efficiency rates only were 11.28% and 4.24%
for the DBTRG-05MG and HepG2 cells, respectively. Therefore, PEI transfections were easy for Balb/3T3, 293, A549 and B16-F10 cell lines.
PEI transfections were difficult for DBTRG-05MG and HepG2 cells.
Because PEI transfection for Balb/3T3, 293, A549 and B16-F10 cells were better transfection efficiency than HepG2 and DBTRG-05M, these
cells were selected as target cells . Furthermore, Balb/3T3 and 293 cells are belonging to immortal, non-tumorous. In contrast, A549 and B16-F10 cells are belonging to tumorous. They can be studied the genes expression between non-tumor and tumor cells.
3.2. Assay of NF-κB expression level in various cell lines
As previous description, NF-κB activities are higher in tumor cells than normal cells. Therefore, the NF-kB activities were monitored in Balb/3T3, 293, A549 and B16F10 cells to determine whether the difference exists in the selected target cell line.
In this experiment, pNF-κB–hrGFP vector was used to detect NF-κB activity in Balb/3T3, 293, A549 and B16F10 cell lines (Figure 9, 10). Two plasmids were co-transfected into target cells: one plasmid containing the TFBS which drove the green fluorescence protein (hrGFP) gene; the other containing CMV promoter which drove the red fluorescence protein (AsRed). Co-transfection with pAsRed-N2 could avoid the difference in transfection efficiency at different experiment groups. Moreover, pARE-hrGFP containing ARE site (prokaryotic TFBSs) was as negative control vector to normalize the degree of transgene expression in different cell lines. The expression folds formula is described as below:
CMV-AsRed X ARE-hrGFP TFBSs-hrGFP
Expression Fold =
In Balb/3T3 and B16F10 cells, pNF-κB-hrGFP expression was 12-fold and 34-fold than pARE-hrGFP, 30-fold and 12-fold than
pAP-1-hrGFP (Figure 9). In 293 and A549 cells, pNF-κB-hrGFP expression was 4-fold and 28-fold than pARE-hrGFP, 1.2-fold and 21-fold than pAP-1-hrGFP (Figure 10). The NF-κB activities were lower in normal-like (293 and Balb/3T3) cells than tumor cells (A549 and B16F10).
Interestingly, all cell lines had higher NF-κB activities than other transcription factors in this experiment (Figure 9, 10).
3.3. Construction and selection of NF-κB based promoter library
Promoters’ activity dependent to transcription factors and NF-κB has higher activity in tumor cells. Therefore, a new NF-κB based promoter library was created to select a new promoter sequence which can be overexpressed in tumor cell but not in normal cell.
3.3.1. Construction of NF-κB based promoter library
NF-κB based promoter library was created by randomly ligating oligonucleotides containing NF-κB, CRE, MEF-2 or Sp1 sites. First,two complementary oligonucleotides were synthesized for each individual binding sites, later they were annealed and phosphorylated to yield short DNA fragments (the section 2.2.1).
After oligonucleotides with different binding sites were ligated, cleaned up (the section 2.2.2) and then observed their sizes by DNA electrophoresis. The sizes of DNA fragments were between 100 bp and 300
bp after ligation (Figure 11). An average size of joint DNA fragments was
“200” bp. These primers’ were about “20” bp sizes before ligation.
Therefore, the promoter library had 1,048,576 (4200/20= 4 10=1,048,576) variety of combinations probably. Adaptors with Hind III restriction enzyme site were added to ligate the end s of promoter fragments and then they were inserted to B16.4 vector (with anti-phOx gene). After ligation, the ligation efficiency was determined by transfomation. The results showed that the negative control (the cutting vector alone treated with T4 ligase) and my sample had about 100 and 2568 colonies respectively (the section 2.2.1.3). However, total ligation products only got 1/50 volume to transfom into E. coli on the plate (Ttotal 15 μl ligation product get 2μl to transform into E .coli by electroporation, and then total 1000 μl transformable volume got 100μl to spread on the plate. Therefore, total ligation products only got 1/50 volume transfom into E. coli on the plate.).
Therefore, there are 123,400 varieties in promoter library actually.
3.3.2. Selection of NF-κB based promoter library
The expression of NF-κB based promoter library in Balb/3T3 and A549 cell lines were selected by MBSCS to high-activity promoters respectively. Balb/3T3 is normal-like cell line, and A549 is lung carcinoma cell line. NF-κB based promoter library transfected into normal and tumor cells to compare difference of gene expression and find tumor specific promoters.
3.3.2.1. Selection of NF-κB based promoter library in Balb/3T3
Original NF-κB based promoter library was transfected into Balb/3T3 and analyzed gene expression by flow cytometry (Figure 12). Fluorescent expression of B2a plasmid (with CMV promoter and anti-phOx-eB7) was 19.48% and NF-κB based promoter library with anti-phOx-eB7 was 7.062%. Supposition, CMV promoter is certain to express. Different cell types, staining and transfection efficacy causes the fluorescence expression not arrived at 100%. When the fluorescent intensity (FLH-1) exceeds 101 values that gene is turned on.
How have many activities promoters in my established promoter library?
The ratio was calculated by the following formula (Figure 12):
In original promoters library, the ratio of active promoters
fluorescent intensity of CMV promoter - negative control = 7.062-3.48/19.48-3.48 ratio =
fluorescent intensity of promoter library – negative control
22.38%
=
In overall promoters, active promoters were 22.38% approximately.In another word, total 123,400 varieties kinds promoters had gene expression only were 27626 varieties.
In order to further analyzed promoter library activities, the observation of gene expression in original promoter library must did first. Under the fluorescent intensity (FLH-1) exceeds 101 values, most cell counts lain on FLH-1=101~3 x 101 region (M2). In FLH-1=3x 101~102 region (M3), cell counts were fewer than M2 region. However, in FLH-1 >102 region (M4),
cell counts were almost zero. According to the distribution (cell counts and fluorescent intensity) of gene expression in original NF-κB based promoter library, fluorescent intensity (FL1-H >101) was divided into three groups (Figure 13):
(1) Low fluorescent intensity, FL1-H =101~3 x 101
This ratio (M2) was 5.63% in overall situation. Under the overall expression ratio (M1), it occupied 79.75% (M2/M1).
(2) Medium fluorescence intensity, FL1-H =3x 101~102
The ratio (M3) was 0.77% in overall situation. Under the overall expression ratio (M1), it occupied 10.91 % (M3/M1).
(3) High fluorescence intensity, FL1-H >102
The ratio (M4) was 0.66%. Under the overall expression ratio (M1), it occupied 9.35% (M4/M1).
In order to distinguish the cells with high, medium or low expressive promoters, transfectants were isolated by the modified magnetic beads separation method. In the separation of processing, the competitor (glycine-phOx) was added to further exclude low expressive cells and remain the high-expressive cells by surface expression of anti-phOx molecule to competing with the competitor. After MBSCS separation, B2a and origin NF-κB promoter library vectors were transfected into Balb/3T3 cell and its gene expressive rates were 35.84% and 3.1.32%, respectively (Figure 14).
In the same way, the ratio of activities promoter in isolated promoter
library was calculate by the following formula:
In isolated promoters library (after elution), the ratio of active promoters
After isolation of promoter library (after elution), its’ overall gene expression that compare to original library are increase or decrease by the following formulato calculate expression times.
fluorescent intensity of B2a - negative control
fluorescent intensity of islated promoter library - negative control
Ratio = = 85.967%
Second transfection, expression level of B2a
First transfection, expression level of B2a
= (After elution) X
Original promoter library Expression
times = X
Isolated promoter library
3.2-fold
=
Because transfection and staining twice had different efficiency, CMV promoter activity divides by each other to remove from this interference.
The isolated promoter library was 3.2-fold than origin. Such the isolated promoter library had higher gene expression than origin. In the preliminary experiment, this strategy seems to be practicable.
The comparison between original and isolated promoter library in the low, medium or high fluorescence intensity ratios were able to analyze
further changes of fluorescence intensity. According to prior way, isolated promoter activity could divide into three groups (Figure 15):
(1) Low fluorescent intensity, FL1-H =101~3 x 101
This ratio (M2) was 25.50% in overall situation. Under the overall expression ratio (M1), it occupied 81.42% (M2/M1).
(2) Medium fluorescence intensity, FL1-H =3x 101~102
The ratio (M3) was 5.27 % in overall situation. Under the overall expression ratio (M1), it occupied 16.83 % (M3/M1).
(3) High fluorescence intensity, FL1-H >102
The ratio (M4) was 0.66%. Under the overall expression ratio (M1), it occupied 0.64% (M4/M1).
The comparison between original (Figure 13) and isolated (Figure 15) promoter library activities, two results were shown and analyzed by histogram (Figure 16). In the low fluorescent intensity region, the ratio of isolated library (M2/M1) was increased 1.59% than original. And in medium fluorescent intensity region, the ratio of isolated library (M3/M1) increased 5.91 % than original.
However, gene expression rate decreased 8.71% in the high fluorescence intensity region. This cause knew unclearly. But total gene expression is enhancement (3.2-fold) after MBSCS. Therefore, MBSCS is the effective method that can exclude no and low expression of promoters.
MBSCS could divide three procedures. The first procedure is washing, the second procedure is competing and the third procedure is eluting (Figure 5). In order to further analyze three procedures of MBSCS (Figure
17). The changes of three different fluorescent intensity regions were analyzed by flow cytometry. In three procedures, washing procedure of gene expression in low fluorescent intensity was higher level (83.51%) than competing (81.42%) or eluting (80.92%). However, in medium or high fluorescent intensity was lower level (16.59 or 0.54%) than competing (18.26 or 18.26 %) or eluting (18.98 or 0.64%), but they were no significant (Figure 18).
After isolation, the ratio of medium/high-active promoters in overall promoters was calculated by following formula.
The ratio of “medium/high-active” promoters in overall promoters Gene expression of medium/high-active
promoter- negetive control
Gene expression of B2a - negetive control = 5.69%
=
If pick 100 colonies and purify their DNA after isolated plasmids transformed into E. coil, there will be 5.69 colonies with medium / high-active promoters among 100 colonies possibly. On the basis of Laboratory manpower and select probability, picked 120 colonies purified their DNA by minipreparation; maybe there will be 6.82 colonies with medium/high-active promoters. These plasmids transfected into Balb/3T3 and their gene expressions were analyzed by flow cytometry (Figure 21).
There were 25 plasmids by selecting and they divided into three groups simply as following.
High expression (100~50% Gate, proportion of gene expression):
B-2 (53.26%), B-8 (51.02%), B-25 (94.73%), B-26 (96.66%), B-27
(83.76%), B-28 (63.92%), B-29 (92.35%), B-31 (62.21%) and B-35 (62.28%)
Medium expression (50%~10% Gate, proportion of gene expression):
B-1 (25.65%), B-15 (28.03%), B-19 (41.6%), B-50 (15.48%), B-59 (22.08%)
Low expression (10%~ Gate, proportion of gene expression):
B-86 (8.95%), B-90 (1.48%), B-96 (2.91%), B-102 (3.16%)
Because these plasmids were purified by mimipreparation, DNA solutions involved chemical compounds or RNA probably. The purity of DNA solutions is important in transfection. Therefore, these select plasmids were purified further by midipreparation to analyze gene expression in different cells and find the tumor specific promoters.
3.3.2.2. Selection of NFκB-based promoter library in A549
In the same select isolated Balb/3T3 promoters’ way, NF-κB based promoter library was transfected into A549 cell and selected by MBSCS (Figure 19). The observation of gene expression in three procedures (Figure 20), washing procedure of gene expression in low fluorescent intensity is higher level (92.83%) than competing (85.44%), but is lower than eluting (96.46%). In medium fluorescent intensity, washing (7.17%) is lower level than competing (14.91 %) but is higher level than eluted (3.11%). Interestingly, in high fluorescent intensity, eluting (0.61%) is higher level than washing (0.32%) or competing (0.22%). These results were no significant (Figure 20).
After MBSCS separation, eluted cell pellets were collected, extracted
plasmids and then transform into E. coil. Picked 120 colonies and purified plasmids by mimipreparation to analyze their gene expression. These plasmids transfected into A549 and their gene expressions were analyzed by flow cytometry (Figure 22). 12 pieces of plasmids were selected by their gene expression and divided into from three groups simply as following:
High expression (100~50% Gate):
A-2 (40.25%), A-10 (61.45%)
Medium expression (50%~10% Gate):
A-5 (25.1%), A-6 (27.98%), A-8 (17.12%), A-9(19.11%), A-11 (28.65%)., A-12 (18.42%)
Low expression (10%~ Gate):
A-23 (4.54%), A-107 (6.48%), A-120 (2.05%), A-141 (2.78)
These select plasmids were purified further by midipreparation to analyze gene expression in different cells and find the tumor specific promoters.
3.4. To compare with gene expressions of differential NF-κB based promoters in differential cell lines
3.4.1. Gene expressions of differential NF-κB based promoters from Balb/3T3 selection
Selectable plasmids from Balb/3T3 were transfected by PEI into B16F10, 293, Balb3T3 and A549 cells. Gene expression levels in four cells
were ordered and compared tumor with normal-like cells in each experiment (Table 2). B-2 promoter in Balb/3T3 and 293 cells are lower activities than other promoters, but gene expression in A549 and B16F10 cells are higher than other promoters. B-31, B-35, B-28 and B-51 are ubiquitous promoters. B-28 is medium expression than other promoters;
B-31, B-35 and B-51 are high expression than other promoters in these cells.
3.4.2. Gene expressions of differential NF-κB-based promoters from A549 selected
The same way of Balb/3T3, gene expression levels in four cells were ordered and compared tumor with normal-like cells in each experiment.
Maybe A-2 and A-111 vector is a ubiquitous promoter, A-29 in A549 high expression than other cells (Table 3).
3.5. DNA sequence of NF-κB-based promoters and TFBSs
Sequence of B-51 has been verified completely by autosequencing,
but the complete information of sequence for B-2, B-6, B-31 and B-28 vectors have not been got yet. Because the length of these sequences is so long (over 1 kb), these sequences have only got their partial sequences (about 200 bp). There are one NF-κB site, two CREB sites, one Sp1 site
and four MEF-2 sites in the sequence of B-51 vector. B-51 promoter has high gene expression in Balb/3T3 and 293 cells than A549 and B16F10 cells (Table 3). The partly sequences of B-2, B-6, B-31, B-28 promoters showed that they have different elements in their sequences (Figure 23-24).
B-2 promoter has two NF-κB sites and CREB sites; B-6 promoter has one CREB site, three Sp1 sites and two MEF-2 sites. Both promoters are high activities of gene expression in tumor cells (A549 and B16F10) but low activities in non-tumorigenic cells (293 and Balb/3T3 cells). B-31 promoter has two Sp1 sites and one MEF-2 site; promoter is ubiquitous and high active for expression in four cell lines. B-28 promoter has one Sp1 sites and one MEF-2 site and is highly active for gene expression in Balb/3T3 cells.
In addition, I also selected and gained high expression promoters from A549 cells transfected with NF-κB based library, the results showed A-2 and A-111 are active in all cell lines. Because their insert sequences are so long (over 1kb) or certain unknown problem, these sequences cannot be obtained. Therefore, these promoters are not discussed.
Chapter 4 Discussion
In this study, the modified magnetic beads separation system (MBSCS) used to isolate lower gene expression and medium / high expression of promoters. Magnetic beads separation system does not separate high-expression and low-expression on cell surface. The low expressive population in cells transfected with promoter library occupies the most proportion of transfectants, it strongly interferes the selection for the promoter with high expression activity. Therefore, I hoped to exclude the promoters with low expression activities when the promoter library is selected with MBSCS. The competitor (glycine-phOx) was designed and used to decrease low-expression of promoters. By competing with glycine-phOx, low expressive transfectants would catch the competitor and be released from magnetic beads separation system. Thus, the low expressive transfectants could be excluded and increase the proportion of high expressive transfectants, later the promoter with high expression activity would be easier to identify.
Without competitor, the low fluorescence intensity ratio in total fluorescence intensity is about 79.83%, medium fluorescence intensity is 10.92% after selection in Balb/3T3 cell, and competitor treatment raises the medium fluorescence intensity to 16.83%. Furthermore, I analyzed the fluorescent intensities of transfectants after washing, competing and eluting procedures, the results showed that the proportions for medium
fluorescent intensity is 16.59% after washing, 18.26% after competing and 18.98% after eluting in BalB/3T3 cell (Figure 18). In other cell, A549, medium fluorescent intensity is 7.17% after washing and 14.19%
after competing. However, after eluting, medium fluorescent intensity is only 3.11%. Interesting, high fluorescent intensity is 0.61 after eluting, and is higher expression than washing (0.32%) or competing (0.22%) (Figure 20).
According to these data, this system seems to be practicable to enhance the proportions for medium fluorescent intensity. However, the decrease of lower expression is not significant. Maybe it is due to the amounts of competitor that is not enough. The results may be improved by adding more competitors or decrease cell numbers.
Ubiquitiously or specifically high expressive promoters both were isolated from the promoter library; such as B-2 promoter is overexpressive in tumor cells (A549 and B16F10 cells). B-6 promoter is A549-specific promoter. How many NF-κB binding site in those sequences is not clear (because their complete sequences are not available) but partial sequence results showed they contain several predicted TFBSs (NF-κB, Sp1, CREB or MEF-2) in their sequences (Figure 24). B-2 promoter sequence had two NF-κB binding sites which is conformable to previous results that NF-κB activities are more active in tumor than normal cells (Figure 9-10). The activities of NF-κB in tumor cells usually are aberrant (Rayet and Gelinas 1999). In addition, NF-κB has higher affinity to DNA than other transcription factors. Many researches indicate that the dimers of NF-κB have high affinity for NF-κB binding sites (from 1013 to 1010 M) in comparison with most transcription factors
(whose affinity for cognate sites is generally close to 109 M) (Urban and Baeuerle 1990; Chen-Park, Huang et al. 2002; Bosisio, Marazzi et al.
2006), and can generate very stable complexes (half-life is about 45 minutes) on DNA (Zabel and Baeuerle 1990).
However, B-6 promoter may not had the NF-κB site and it is high expressive in A549 cells. Because B-6 promoter’s complete sequence is not available yet, therefore the distribution of TFBSs in B-6 promoter is still unclearly. Promoter library were isolated by MBSCS and then individual plasmid was selected by transfecting each single plasmid into target cell to precisely determine its expressive activity. In transfection, one cell could gain more than one plasmid that may be the season why the plasmids with low expressive activities would be isolated after MBSCS. Although MBSCS still has some defects for efficiently selecting high expressive promoter, the B-51 and B-35 promoters are isolated and they are high expression in Balb/3T3 cell. Therefore, MBSCS should be practicable.
As previous section (Chapter 1), 3 different TFBSs can interact with NF-κB to strengthen gene expression. B51 promoter sequence is complete sequenced and the result showed that there are one NF-κB, one Sp1, two CREB and four MEF-2 binding sites in the sequence. The result indicated that these TFBSs can assemble to form a high expressive promoter as our prediction. Recently, NF-κB–DNA complexes are further stabilized by cooperative protein–protein interactions with other sequence-specific transcription factor and architectural proteins, giving rise to multimolecular protein–DNA complexes called enhanceosomes. In vitro the interferon beta
(IFN-β) enhanceosome (Maniatis, Falvo et al. 1998) remains completely stable for over 10 hours (Yie, Merika et al. 1999). In this experiment, the selection of three TFBSs (Sp1, CREB, MEF-2 binding sites) ligated NF-κB binding sites randomly to create novel promoters. These three factors are related to tumor development, survival and metastasis. In addition, these also correlate with NF-κB in regulatory gene expression. The creation of NF-κB based promoter library by ligating randomly NF-κB, Sp1, CREB, or MEF-2 binding sites can isolate high expression or specific promoter successfully in different cell lines (Table 3).
Li et al. synthesized the novel muscle promoters at 1999 (Li, Eastman et al. 1999) by random assembly of E-box, MEF-2, TEF-1, and SRE sites.
They obtained high expression promoter in muscle by pick over hundreds of individual clones, and screening of for transcriptional activity in vitro and in vivo. However, they cannot control individual promoter by picking is high expression or not. Therefore, magnetic beads separation system is used in my experiment to improve this defect. In addition, Li et al. used known, muscle specific and high-expression transcription factor to create new muscle promoters. The probability of created high expression promoters is higher than my promoter library, because the selection of four transcription factors exists not only in tumor but also normal cells. That is to say, the high and tumor specific promoters are selected difficultly in my experiment.
However, tumor specific promoters (B-2, B-6) and ubiquitous promoters (B-31) were isolated and selected respectively by magnetic beads separation system combining with the competitor system (MBSCS)
in my experiment. MBSCS is able to select high or specific promoters easily. This strategy not only let us understand the context of high efficient promoter in tumor cells, but also let us knows which factors in the tumor cells are important. This may take advantage of it to more thorough understanding about tumor cells, and find the better treatment for cancer.
Figure 1 Tumor specific promoters made and obtained procedures