Ambiente de control

- A surprising result!

We wished to examine the activity of the 2A region of FMDV in the absence of other FMDV sequences which might influence 2A-mediated activity. A further advantage of

removing the FMDV 2A region from the context of the FMDV polyprotein itself, was that the FMDV 3CP^° polyprotein processing would have otherwise considerably complicated the interpretation of the data. It was, therefore, chosen to study the 2A

region of FMDV in an artificial reporter gene polyprotein system, within which the sequence encoding the FMDV 2A region was placed between two reporter genes. The

reporter genes used initially were CAT and GUS in the plasmid pMD2.

In vitro coupled transcription and translation of the plasmid pMD2, encoding [CAT2AGUS], produced the polyprotein [CAT2AGUS], and the cleavage products CAT2A and GUS (as identified by immunoprécipitation with anti-CAT and anti-GUS serum) in agreement with previous reports (Ryan and Drew, 1994). There were considerable amounts of internal initiation products from the coupled transcription / translation reactions, which could have been reduced by improving the Kozak consensus sequence of the translation-initiating codon of CAT, although this is a typical problem with these in vitro systems. However, an assignment of the major internal initiation products, and a densitometric analysis of their relative abundance, provided sufficient information to allow the continued use of plasmid pMD2 and other plasmids based upon it. A comparison of the translation products of pCATGUS and pMD2 demonstrated that the truncated polyproteins, [A’75MetQAT2AGUS] and [Al^^^®^CAT2AGUS] produced by internal initiation at Met-75 and Met-163, respectively, within the CAT gene, and the full length [CAT2AGUS] polyprotein were processed equally efficiently by the FMDV 2A

Unexpectedly, densitometric analyses of the cleavage products, taking the internal initiation into account, demonstrated CAT2A and GUS were not present in equal quantities. The ratio of CAT2A : GUS was calculated to be between 2.4 : 1 and 5.6 : 1 in the rabbit reticulocyte system. In the wheat germ extract system the ratio of CAT2A : GUS was even greater, although it was more variable, at between 6.2 : 1 and 14 : 1. Initially these differences in stoichiometry were thought to be due to trivial reasons such

as degradation, or incomplete synthesis of the GUS protein. However degradation experiments in which the transcription / translation reaction of pMD2 was arrested and the reaction mixture then incubated at 30 °C (the temperature at which transcription / translation reactions are carried out) for up to 24 hours demonstrated that although both CAT2A and GUS were degraded to some extent, the GUS protein was degraded more slowly than the CAT2A protein, and not to the extent which would be required to produce such differences in stoichiometry. One would thus expect that if equal quantities of CAT2A and GUS were made then, should degradation have been a factor requiring consideration, there would have been more GUS than CAT2A detected. It has been reported that the synthesis of protein from larger genes in rabbit reticulocyte lysate is less efficient than the synthesis of smaller proteins (Dasso and Jackson, 1989). If the synthesis of the GUS gene was not being completed for this reason then one might have expected to see a "smear" of protein representing a range of C-terminally truncated GUS proteins. However this was seen neither in the polyacrylamide gel analyses of the transcription / translation reactions nor in the immunoprécipitation of their products. One might also have expected that there would be a greater "smear*" for the wheat germ extract reactions in which the ratio of CAT2A to GUS is greatest. This, however, was not the case either. Thus, it was concluded that the synthesis of GUS was not being hampered by its size.

The only explanation for an excess of CAT2A was that there must have been more CAT2A translated than GUS. This was most likely to be controlled at the level of

translation than transcription since translation reactions programmed with RNA from an

in vitro transcription reaction also displayed this effect. The ratios of CAT2A : GUS can be explained by an 'intermption' in translation, which in some cases results in termination producing CAT2A only, in others it releases CAT2A then continues to produce GUS (termed cleavage), and in others the ribosome continues as normal to produce CAT2AGUS (termed non-cleavage). Such a concept is not compatible with the theory of

a 2A-mediated proteolytic cleavage.

4.2.2 The importance of the 2A RNA sequence in 2A activity

If an interruption in translation was occurring, then it was thought that perhaps the 2A RNA sequence or secondary structure was important for FMDV 2A activity. The

nucleotide sequence encoding 2A in pMD2 had been altered from the wild-type sequence in order to introduce new restriction enzyme sites. It was feared that this might have resulted in a change from 2A wild-type activity. However, plasmid pTG395 which encoded CAT-(5 C-terminal residues of 1D)-2A-GUS, in which the 1D/2A sequences were completely wild-type displayed similar cleavage activity and product ratios (Donnelly et al, 1997). It can therefore be concluded that the activity of the 2A region encoded by pMD2 was not affected by the alteration of the nucleotide sequence to

produce new restriction enzyme sites.

The introduction of a +1 frameshift throughout the FMDV 2A region was employed to study any effects mediated by the RNA sequence encoding 2A, rather than the 2A protein itself. As expected the majority of translation products from the construct encoding [CAT(+1)2AGUS] was the uncleaved polyprotein and its truncated forms (due to internal initiation). Unexpectedly, however, this was accompanied by a small quantity of a product of the same size as CAT2A which was immunoprecipitated by anti-CAT

antibodies. This result cannot be explained but may indicate that the RNA sequence of FMDV 2A does play a role in termination of translation after translation of the 2A region. The 2A RNA sequence was examined for possible secondary structure elements using the

programme RNAdraw (Hofacker et al, available at htp.itc.univie.ac.at.), however, no consistent features were found within the predicted 2A RNA secondary structures of the

aphtho- and cardioviruses. An examination of the primary RNA sequence for the C- terminus of 2A revealed no consistency in codon usage between aphtho- and cardioviruses, although within the aphthoviruses themselves there was reasonably high identity.

4.2.3 The influence of the protein context of 2A on its cleavage ability

The 2A-mediated cleavage occurs co-translationally. It is, therefore, reasonable to

assume that the wild-type FMDV sequence following 2A is not necessary for cleavage

and indeed this was shown by Ryan et a l (1991). In the FMDV virus 2A acts as the species PI-2A, so it is not surprising that PI is influential in mediating 2A activity (Ryan

et al, 1991). This influence may be asserted either by residues close in the primary sequence or those which are spatially close in the tertiary structure of PI-2A. The x-ray crystallographic data of the capsid protomer (see Figure 1.5.1; Lea et al, 1994) demonstrated that the C-terminus of ID was not associated with any other part of the

protomer but with the adjacent protomer in the pentamer. Since 2A-mediated cleavage is

a co-translational process, which may be occurring in the ribosomal exit pore from which

residues distant in the primary sequence would already have been extmded, it is unlikely that residues very distant in the primary sequence are involved. Hence, it was assumed most likely that residues from within ID were required to enhance cleavage activity to 100 %. Indeed, the inclusion of C-terminal ID sequences of 180 and 39 amino acids in length did increase the efficiency of cleavage (this thesis. Section 3.3). However, they

Proteins GFP and CAT were both used to precede the 2A region in artificial reporter gene cassettes (see Section 3.7). No correlation with the preceding reporter gene protein could be found for either the efficiency of cleavage or the ratio of cleavage products. Proteins CAT, NEO and GUS were each encoded after 2A regions. Only when the CAT protein was C-terminal to 2A was the ratio of cleavage products found to be 1 : 1. As predicted no correlation was found for the protein following 2A and the extent of cleavage.

4.2.4 Comparing rabbit reticulocyte lysate and wbeat germ extracts

The ratios of CAT2AGUS cleavage products in both rabbit reticulocyte lysate and wheat

germ extract varied from batch to batch, but an excess of CAT2A always existed and was always greater in the wheat germ extract system. The percentage of cleavage based either

on the quantity of uncleaved material or CAT2A was more consistent in rabbit reticulocyte lysate with average values of 91 % in rabbit reticulocyte lysate and 96 % in wheat germ extract. Qualitatively, in both translation systems the activity of 2A (and mutated 2A regions) was similar.

The relatively similar percentages of cleavage reaction occurring between both batches

and translation systems may be read as an indication that the ability to mediate 2A cleavage resides with the 2A region alone. However, the occurrence of large quantitative differences between the two systems for the ratios of CAT2A : GUS (and smaller batch variations), may infer that the translational machinery of these in vitro systems is playing

a role in the determining the ratio of products (i.e. determining the frequency of termination of translation after 2A). If 2A-mediated cleavage was merely a proteolytic event then its dependence on ribosomes would be expected to be less.