E. coli BL21(DE3) seems to be the dominant choice for the production of membrane proteins for the purpose of solving their structures and the characterization of their functions . Membrane proteins are in some cases produced by this strain, but the product usually appears as inclusion bodies , and this is generally undesirable due to downstream problems . It is also clear that relatively few membrane protein structures have today been solved and furthermore that until 2005, more than 70% of the solved structures were of non-eukaryotic origin . The lack of good protein production methods is thought to be one of the dominant factors behind the low success .
Miroux and Walker have shown that the production of eukaryotic membrane proteins with the BL21(DE3)-system is generally very toxic to the cells and that most of them did not even survive induction . They thus developed the BL21(DE3) mutant strains, C41(DE3) and C43(DE3), the "Walker strains", which seems to be much better suited for production. Through subsequent genomic and proteomic studies , it was shown that the original P
promoter of these strains, which is generally stronger than the original P
variant, had reverted to the original wild type sequence. It was clearly shown that the lower strength of the lac promoter led to a lower transcription rate of the T7 RNA polymerase, and that this in turn, had a positive effect on the cellular viability leading to higher total production levels. The mechanism by which the viability is restored has not yet been revealed. Wagner and coworkers suggested that it was due to the lower rate of transcription of the membrane protein, due to a decreased T7RNAP synthesis, and used this to modulate these levels using a tunable T7Lys expression system . This resulted in increased cell viability and in several cases also to an increase in the total production. In some cases, however, the use of C41(DE3) and C43(DE3) gave still better results. The tuning of T7RNAP levels was already suggested by Miroux and Walker , but with the comment that it might not be altogether successful for the difficult eukaryotic membrane proteins that still remain to be successfully produced.
In this paper, we have described a different approach to membrane protein production by reducing the substrate uptake rate, and thus indirectly the protein synthesis rate, using the PTS-mutants. The production is introduced on low copy number plasmids with tunable promoter properties. The advantage, besides a lower synthesis rate, is that it avoids the ATP-costly transcription of specific messengers, which we believe to be an additional reason for the reduction in cell viability upon induction. In many cases in the literature it has been indicated that the cell after induction is lacking energy and will seek alternative sources for the supply of both carbon and energy, which can lead to e.g. ribosome degradation since these are by far the largest cellular resources available [5, 14]. Energy-costly processes should thus be avoided as far as possible to save the cell from internal degradation.
Another large drawback of production systems involving elements of periplasmic or outer membrane transport is the possible saturation of the inner membrane secretory system (Sec), and this seemed to be another consequence of high and rapid membrane protein production in the BL21(DE3), pET-vector based system . The accumulation of precursors of periplasmic and outer membrane proteins in the cytosol is of course also a waste of building blocks and energy. In the same study it was also shown that overexpression leads to the induction of the acetate-phosphotransacetylate pathway probably because of the need for additional energy which, under these particular circumstances, might not be possible to obtain by increased oxidative phosphorylation.
The glucose permease mutants and the BL21(DE3) strain, all induced by arabinose, are described in this paper as being equally effective in protein expression, as we were able to express three out of five proteins with all of them. However, the yield of protein per cell for the PPA652 mutant was twice as high as for the other producing systems. The glucose permease mutants grow more slowly than the BL21(DE3) cells and therefore need approximately twice the time (approx 7 h) to reach the same cell density. The mutant cells can however grow to a higher cell density since they produce no acetic acid, and this may lead to an even further increase in the total production level. This ability has earlier been shown when mutants were grown in a fermenter to an OD600 of approximately 100 . The limit is set by the oxygen-transfer capacity, which in the fermenter is characterized by a KLa of approx 800 h-1 compared to microtitre plates with a KLa of approximately 130 - 188 h-1 [15, 16]. Theoretically, this comparison should allow an OD600 of approximately 20 on the best plates using the mutants when no acetic acid is formed, which is about 5 times higher than the density reached today.
The question of whether it is the low growth rate or the lack of acetic acid production that is the reason for the better production has not been totally clarified. The data suggest the latter, since BL21(DE3), in spite of its high growth rate, shows similar production results to our glucose permease mutants, and the reason could then be the low acetic acid production of BL21(DE3) at high glucose concentrations in minimal medium. The reason why this B strain is not subject to overflow metabolism in the same way as the K-12 strains has, to our knowledge, not been explained in detail. It has however been suggested that B-strains have a more active glyoxylate shunt than K-12 strains , which results in the direction of pyruvate into other metabolic functions and in the reduction of acetic acid formation.
The comparative data presented in Figure 4 and Table 3 also suggest that the lower the acetic acid production the higher is the protein production. This is not however the only factor that is important, since it seems that a too low glucose uptake, maybe only leading to maintenance levels of glucose in the double mutant, also reduces production. This again suggests the importance of a high-energy status of the cell.