Once transformed bacteria were identified by their ampicillin resistance it was necessary to determine which of them carried the recombinant plasmid, which should have been most of them, for the reasons outlined above. Colonies were picked from ampicillin LM agar plates and grown up in 10 ml LB broth containing 100 pg ml'"* ampicillin, induced and harvested after 3 hours. Total cell protein was examined by running PAGE gels on soluble protein from bacteria lysed by boiling in running buffer for 5 mins. Those bacteria that gave a band of the correct mass (37 KDa) were picked, as they were expressing the protease-GST fusion (these were designated JM 101/HPR). The recombinant protein was also identified by comparison to induced bacteria carrying the parent pGEX-2T plasmid, uninduced recombinant bacteria and JM 101. The recombinant bacteria gave a very strong band at 37 kDa that is heavier than the protein (26 kOa) from bacteria carrying only pGEX-2T (JM 101/pGEX) and is not contained in JM 101. This is shown in the gel below.
A
Column Sample
1 M. W. standards
2 Soluble protein from JM 101/HPR
3 Insoluble material from JM 101/HPR 4 JM 101 total cell protein 5 & 6 Total cell protein from JM 101/HPR
7 Total cell protein from JM 101/ pGEX
Comparison of lanes 2, 5 and 6 with 7 shows the expected increase in mass on forming the fusion protein. This band is absent in the parent bacteria (lane 4) so must have come from the transformed bacteria. The masses of the standards are 76 kDa, 66.25 kDa, 42.7 kDa, 30 kDa and 17.2 kDa.
Various checks were carried out to ensure that the plasmid isolated from the bacteria selected contained the protease gene. The simplest approach is to perform a comparative restriction analysis on the recombinant and parent plasmids. In this approach the plasmids are digested with two restriction endonucleases that each have one unique site in the plasmid and the size of the fragments produced are compared by agarose gel electrophoresis. A partial restriction map for the two plasmids is given in Appendix B. It can be seen that digestion of the plasmids with Sma 1 and Eco RV will give two bands from both plasmids. The recombinant plasmid will give one of these bands about 300 bp heavier due to incorporation of the HIV-1 protease gene (297 bp). These two enzymes were chosen as they can be used in the same buffer system without either of them losing much activity. This eliminates the need to perform the digests consecutively.
Sequencing of the insert will ensure that the protease is in the correct reading frame and that it is inserted in the correct place in the MCS. Isolation and purification of the plasmid for sequencing proved to be problematic, with the usual protocol giving low yields of the plasmid, heavily contaminated with chromosomal DNA. Small scale preparations, using the Stratagene Plasmid Quik™ kit, were also performed and also gave heavily contaminated plasmid. A modification of the described purification was tried, using benzoylated napthoylated DEAE cellulose (BND cellulose) to purify the plasmid DNA from the chromosomal DNA, tRNA and residual bacterial protein. This also proved problematic, due to the very low yields of
plasmid.
The reason for the low yield is not clear as the parent plasmid (pGEX-2T) is a high copy number p l a s m i d , s o pGEX-HPR should also have a high copy number (plasmid copy number is determined by a DNA sequence carried on the plasmid itself). In order to overcome this difficulty the amount of plasmid was amplified by chloramphenicol treatment. When this is added to the bacterial culture host protein synthesis stops but plasmid replication continues as it is dependent on proteins that turn over slowly e.g. DNA polymerases I and III. However, chromosomal DNA replication requires newly synthesised proteins, so it is also halted by chloramphenicol treatment. Thus, the amount of plasmid is amplified over chromosomal DNA and host protein by chloramphenicol treatment, which makes the purification much easier. This strategy was found to be effective, increasing the yield of plasmid around 50-fold (to around 300-500 pg per litre of culture).
It was possible that a reason for the low yield of plasmid was that after a few hours culturing the ampicillin in the medium had been degraded by 3-lactamases secreted into the medium by the resistant bacteria. After this point any bacterium that spontaneously lost the plasmid would be able to grow, and grow much more quickly than those bacteria that still retained the plasmid. These non-transformed bacteria would quickly dominate the bacterial population and result in very low yields of plasmid. To check this hypothesis a viable cell count was performed. A fresh culture was plated out on LB plates containing ampicillin (100 pg ml""*), the rest of the culture was incubated overnight and then plated out as before. The number and size of the colonies on the plates after 24 hours growth were assessed and found to be equivalent. This indicates that the plasmid is well maintained over long growth periods. The low yields of plasmid in the absence of chloramphenicol amplification therefore remains unexplained.
When plasmid preparations were performed to isolate further amounts of pGEX-2T for cloning the purity of the plasmid was checked by agarose gel electrophoresis on the intact plasmid and on the plasmid restricted with Sma 1 and Eco RV. These were run against similarly treated samples of the pure plasmid.
Once isolated and purified attempts were made to sequence the plasmid DNA. Unfortunately, sequencing proved extremely difficult. The dideoxy method,^*'^ was used, with ^ ^ 8 as the radiolabel. There seemed to be impurities in the plasmid preparation that interfered with the DNA polymerase used in sequencing and so at best only partial sequences of the pGEX-HPR around the MCS were obtained. An
alternative explanation is that the dénaturation protocol of the double strand plasmid DNA (heating in the presence of sodium hydroxide) used was not working well.
In an effort to circumvent the problems encountered in sequencing double stranded plasmid DNA, single strand sequencing was examined instead as sequencing single strand DNA is usually much easier. Two alternative approaches to obtaining single strand DNA were investigated.
i) Clone the protease gene into the bacteriophage M13. M13 produces single strand DNA when it infects bacteria, such as JM 101, that express the appropriate cell surface receptor.
ii) Make pGEX-HPR into a phagemid by cloning the origin of replication (oh) from the bacteriophage Ml 3 into it. This will make a bacterial cell transfected with the phagemid produce large quantities of single strand DNA. The ori will be amplified by PCR from a suitable Ml 3 vector.
Approach (I) was taken first. Cloning into Ml 3 was carried out in a very similar way to the cloning into pGEX-2T discussed above. The M13mp18 vector^'*^ was used as the double stranded replicative form (RF). This was handled exactly as if it were a plasmid. Like pGEX-2T M13mp18 RF has a multiple cloning site (MCS) containing, amongst others, unique Bam HI and Eco R1 sites. The MCS is located within two genes required for lactose metabolism. Thus, following the same procedure as for the cloning of the protease, the PCR amplified protease gene was ligated into the MCS of M13mp18. The recombinant Ml 3 vector was then transfected into competent JM 101 in the same way as pGEX-HPR was.
Care was taken to ensure that the competent bacteria used in the transfection retained the receptor for M l3, the sex pili, which is encoded by a bacterial plasmid
known as the F’ episome. This episome also contains two genes (pro A and B) involved in proline biosynthesis. JM 101 does not contain these genes, so growing JM 101 on a minimal medium, such as M9, that does not contain proline will ensure that the F' episome is retained. Thus, the bacteria that were made competent for this procedure were grown from colonies picked from an M9 minimal agar plate.
Transformation was carried out exactly as above and the cells plated out in semi solid medium (top agar) on LB agar plates. The screening was carried out by identifying clear plaques on a bacterial lawn from the transformation. Ml 3 slows the growth of infected bacteria, resulting in areas of low bacterial growth that show up as clear plaques against the lawn of quickly growing, uninfected bacteria. Metabolism of
T ^ .. „■
BCIG (X-gal) in the top agar by lactose metabolising bacteria will result in a blue by product, which makes the bacterial plaque blue. JM 101 is unable to metabolise lactose, having had two of the genes required deleted. However, these two genes are carried on the F’ episome. Thus, those plaques that have a blue colour arise from JM 101 infected with M l3 that contains functional lactose metabolising genes. Successful cloning of a gene into the M13 MCS will result in insertional inactivation of the two lactose metabolising genes it carries, which means no metabolism of the BCIG to the blue compound. Thus, clear plaques arise from bacteria infected with Ml 3 that contains the gene to be cloned. Once clear plaques were identified single strand DNA was prepared from them and sequenced. Similar problems were encountered in sequencing this DNA as were found for sequencing double strand plasmid DNA. Only a partial sequence was obtained by this method (see Figure 2.6.4).
The phagemid approach was explored initially but several problems were encountered early on and so the route was set aside in favour of the Ml 3 procedure. The PCR amplified origin of replication (ori) from M l3 was cloned into the Aat II site of pGEX-2T. The recombinant bacteria harbouring the plasmids required were identified after plating transformed bacteria onto agar plates containing ampicillin. Since the recombinant plasmid contains the Ml 3 ori it is slightly larger than the non recombinant plasmid. The difference in the size of the intact plasmids is too small to reliably detect by gel electrophoresis (the M l3 ori is 123 b.p. and pGEX-2T is 4948 b.p.). However, restriction analysis of plasmid isolated from ampicillin resistant bacteria should have allowed those plasmids containing the ori to be easily identified. Several possible restriction digests were considered, but none gave reliable results and so the route was abandoned. This may be due to the method used to isolate the plasmid, ST ET preparations, instead of the much more time- consuming alkaline lysis/cesium chloride banding procedure, which gives much purer plasmid, used for large-scale preparations. The impurities in the plasmid isolated by the ST ET procedure may interfere with the action of the restriction enzymes and result in poor reproducibility of the digests.
Unfortunately, only partial sequences were ever obtained from the above efforts at sequencing. The most extensive sequence obtained gave the 3' 96 nucleotides of the protease gene but this did not show the site of insertion as it was obtained using the 3' PCR primer.
5’...ATA/GAA7rCT/GTG/GAC/ATA/AAG/CTAn‘AG/TAG/GAC/CTA/CAC/CTT/GTC/AA G/ATA/ATT/GGA/AGA/AAT/CTG/TTG/AGTCAG/ATT/GGT/rGC/AGT/TTA/AAT/TTT 3’
Figure 2.6.4: Partial sequence of HIV-1 protease cloned into pGEX-2T