CHAPTER 4. A PROTEOMIC APPROACH FOR SELECTION OF FERMENTATION
5.7 Analysis of the effect of recombinant protein expression on the proteome of
Investigation of the proteomes of the native host strain E.coli DH5a, and the recombinant strain E.coli DH5a pGEX/GST, allowed the opportunity to analyse the effect of expressing a recombinant protein within an E.coli host. Gel images from
E.coli DH5a and E.coli DH5a pGEX/GST were compared and protein spots matched.
Differential analysis of the protein spots was undertaken, and a contour map was generated using Matlab (Figure 5.7). There were considerable differences between the 2 proteomes with the expression of the recombinant protein raising some host protein levels and reducing others.
recombinant versus native
E . c o lipi
Figure 5.7a. Levels of proteins from E.coli DH5a expressing GST compared to proteins from
E.coli DH5a determined by densitometric scanning analysis of the 2-DE gels. Cells were grown at 30°C in glycerol media. Levels of protein are shown in increments of 2.5% per contour, with the mid-point (white) representing zero. Decreases in protein are shown in blue, and increases in red. The box represents the area in which GST has been shown to migrate.
To assess if the protein expression variation between the native and recombinant strains of E.coli that was seen by the differential analysis had an effect on the downstream processing of proteins from these cultures, protein was isolated from the cultures using the anion exchange elution protocol for GST. Soluble protein from E.coli DH5a and E.coli DH5a pGEX/GST cells were separated by centrifugation and loaded on to a HiTrap column (1mL) containing the adsorbent DEAE Sepharose FF (Figures 5.7b, 5.7c). Following loading, bound components were eluted using a linear gradient of 0.01 M NaCI to 0.25M NaCI (10 column volumes), then 0.25M NaCI to 0.35M NaCI (10 column volumes), and 0.35M NaCI to 1M NaCI (10 column volumes) (as described in chapter 2, section 2.8.3.3). As would be expected the protein elution profiles show variation between the native and recombinant strains. This is consistent with the predictions from the proteome analysis and contour maps (Figure 5.7a), which indicate large variation in protein levels throughout the proteome.
native
0.3n E c o 00 C\J mS/cm c 2 Q_ 280nm 0.0 0 10 20 30 40 50 60 30 E 25 O CO 20^
>> 15 5 ? O O Time (min)Figure 5.7b. Isolation of proteins produced from the cell culture of E.coli DH5a by anion- exchange chromatography (DEAE Sepharose FF). Cells were grown in glycerol-rich media and at 30°C. The figure shows total protein as measured at 280nm, and conductivity in mS/cm.
recombinant
G S T E § 02 CN mS/cm c CD e o .r CL 280nm o.a 0 10 20 30 40 50 60 Time (min) 30 20 E œ E H 10 -5 c oo
Figure 5.7c. Isolation of proteins produced from the cell culture of E.coli DH5a pGEX expressing GST by anion-exchange chromatography (DEAE Sepharose FF). Cells were grown in glycerol-rich media and at 30°C. The figure shows total protein as measured at 280nm, and conductivity in mS/cm.
5.8 Discussion
The investigation of the proteome of E.coli cells expressing GST clearly indicate that fermentation temperature affects the proteome of E.coli cells. The proteomes of
E.coli DH5a pGEX/GST grown at each of the 3 temperatures 22°C, 30°C, and 37°C,
exhibit increases and decreases in expression of host proteins.
Cell cultures grown within shake flasks in the varying media showed that GST was produced to a substantial level (-20% total cellular protein). This was confirmed by running protein extracts on SDS-PAGE gels, and also by specific assay for GST. Soluble GST was produced in large quantities and was shown to constitute around 80% of the total GST. The other 20% was in the pelleted fraction of the lysed cells. This could be in the form of inclusion bodies, or GST that has become associated with insoluble proteins. It was hoped that because of the high expression levels of soluble GST within the E.coli cells, the enzyme would provide an ideal system to investigate the production of a recombinant protein using proteomic analysis. Whole cell extracts from selected shake flask cultures were subjected to lEF (pH 3-10) for 40kVhr, and second dimension SDS-PAGE. Analysis of gels at the point of induction and post-induction allowed identification of protein spots that were being over expressed (Figures 5.2.3a, 5.2.3b). It was found that 3 isoforms of GST were being expressed at 3 different isoelectric points (6.5, 6.6, and 6.7 pH units), but all at the same Mw (~26kD). These 3 spots were subsequently identified as GST by MALDI- TOF mass spectrometry (chapter 6, section 6.6). The investigation of the proteins from the pre-induction sample (Figure 5.2.3a) also indicated differences when compared to the host strain DH5a (Figure 4.4a). This indicates that the presence of the plasmid may also be impacting on the cellular proteome. Following the investigation of the E.coli DH5a pGEX/GST system within various media by 2-DE, SDS-PAGE, and specific assay, the system was selected to investigate the effect of temperature on the proteome of E.coli, and therefore the downstream processing of a recombinant protein (GST). Cells were then grown at a larger scale to investigate this parameter.
Cultures grown in the larger scale bioreactors (7L), operated at the 3 temperatures grew to the same optical density and gave comparable yields of GST. One noticeable variable of the cell cultures was the fermentation length. The time to induction for the fermentations grown at 37°C, 30°C, and 22°C, were 14hr, 16hr, and
30hr respectively. Cell cultures grown at 37°C and 30"C reached the point of induction at approximately the same point, but the 22°C fermentation length was approximately 2 fold that of 37°C and 30°C. The productivity of the fermentation, as well as the product titre, are important factors that must be considered in conjunction with any proteomic data. Comparison of the 3 proteomes of E.coli DH5a pGEX/GST and the generation of contour maps comparing proteins from cultures grown at 37°C and 22°C versus 30°C, show that within the region of interest (GST pl=6.7, Mw=26) there are shifts in the protein expression patterns. Within the contour map for the fermentation temperature of 37°C there is an overall increase in expression of E.coli
proteins versus 30°C within the area of interest. The contour map for the fermentation temperature of 22°C versus 30° indicates that there is no significant increase or decrease of E.coli proteins within the area of the GST spots. The proteomic data suggests that the fermentation temperatures of 30°C and 22°C represent the best temperatures to operate the fermentation at in order to reduce host protein contamination within the region of interest, around the GST spot. The operating temperature of 30°C therefore will produce less host protein contamination, but will also give an increased productivity when compared to 22°C, and it is therefore easy to select the fermentation temperature of 30°C over 22°C. The 30°C culture produces less direct host protein contamination versus that of 37°C, but its productivity will be lower, and this must be weighed against ease of purification. In this instance it was decided that that difference in productivity between 37°C and 30°C was negligible, and therefore using the contour maps representing the proteomic data, 30°C was selected as the temperature to operate the fermentations
of E.coli DH5a pGEX expressing the recombinant protein GST.
Isolation of GST from E.coli grown at 22°C and 37°C was also investigated alongside 30°C to assess the effect of fermentation temperature variation on the elution profile. It can be seen that the elution profiles change when material from different fermentation temperatures is loaded. Calculation of purification factor (activity per mL of active fractions divided by activity per mL of load) indicates that the 30°C culture gave a value 14% above that for 22°C, and 23% above that at 37°C. This result agrees well with the predictions from the fermentation temperature selection experiments.
Another interesting feature of these experiments was that they enabled investigation of the effect of production of a recombinant protein within an E.coli host.
Proteins from whole cell lysates of E.coli DH5a, and E.coli DH5a pGEX/GST, grown in glycerol-rich media and at 30°C were investigated by 2-DE. There were significant differences between the 2 proteomes. Areas within the contour map showed large increases or decreases of host proteins. Within the area of interest the GST spot is present and therefore there is a strong increase in protein expression. Variations within the E.coli proteome suggest that the cell is responding to overproduction of the recombinant protein. This metabolic response may be due to the toxic effect that GST has on the cellular environment. This observation is consistent with previous observations where a DNA microarray detected metabolic responses to protein overproduction in E.coli (Oh & Liao, 2000; Jurgen et ai., 2000). Variations in the
E.coii proteome were also observed when material from both cell cultures were
loaded on to an anion exchange chromatography column (DEAE Sepharose FF). GST was isolated from the recombinant cell culture, and an identical elution was performed on the native culture proteins. Again, there were differences between the 2 elution profiles (as well as the expected GST peak). The differences observed agree well with the predictions from the differential analysis and the contour map, where areas throughout the proteome both increase and decrease. The elution profiles indicate that different proteins, or levels of proteins, were binding to the matrix. The host proteins from the sample containing GST (Figure 5.7c) will also be competitively binding to the matix with GST. This may also affect retention of host proteins, although the similar shape of the peaks between 30-45minutes in Figures 5.7b and 5.7c indicate that this may not be the case in these isolations.
This research has shown that proteomic technology presents an opportunity to analyse and optimise fermentation conditions at an early stage of process development. Optimisation using these techniques can influence the way in which fermentation temperature is selected in addition to consideration of final biomass and product titre. This type of analysis at an early stage of process development can allow reduction of potential contamination downstream and therefore avoid complicated and expensive isolation steps.
CHAPTER 6. SELECTION OF FERMENTATION CONDITIONS FOR PRODUCING
AND ISOLATING GST EXPRESSED IN ESCHERICHIA COLI
6.1 Introduction
The work undertaken so far within this thesis has indicated that the proteome of
E.coli exhibits differences when exposed to different fermentation conditions, and
that these variations can be exploited when isolating a target protein from host proteins. Fermentation temperature and media have been investigated using 2-D electrophoresis and have been shown to exhibited significant differences. Isolation of GST by anion exchange chromatography using biomass from the different cell cultures has also been investigated. The next question to ask was whether this proteomic information of the whole cell can aid in isolating the target protein. Within this chapter, the isolation of GST using the selected fermentation conditions was investigated using 2-D electrophoresis in order to visualise the process.
This chapter addresses the isolation of GST. By using information gained from the proteome [and hence the relative position of glutathione S-transferase (GST)] and the protein contaminants, the fermentation conditions of temperature and media were selected, and the recombinant strain was grown in batch culture. Using this material, GST was isolated, and investigated by 2-D electrophoresis, in tandem with mass spectrometry of selected protein spots.
Previous data on biomass production, product titre, and total cell proteome, indicated that the fermentation temperature of 30°C, and glycerol-rich media, present the ideal conditions for producing GST. These parameters were selected to produce material for studies of GST isolation. GST was isolated using anionic exchange chromatography (SP Sepharose fast flow), and affinity chromatography.
Proteomic analysis of purification schemes presents an opportunity to identify proteins that elute under specific conditions within the column. Investigation of various isolation schemes for recombinant proteins expressed within E.coli could lead to a more systematic route being taken to purify recombinant proteins. Identification of proteins that co-elute under specific conditions of, for example, matrix, pH, and conductivity could present a more systematic procedure in which to purify target proteins. This procedure, in conjunction with the E.coli protein database,
could present a realistic opportunity for systematic production of purification protocols for recombinant proteins expressed within E.coli. It would allow researchers and process engineers to purify target proteins more rapidly, and with less optimisation required during scale up. With this in mind, proteins from the elution profile for GST under specific conditions were investigated by mass spectrometry.
6.2 Analysis of the growth of E.coli DH5a pGEX / GST within glycerol-rich