The critical step in classical prenylation assays is the detection of the radioactive anchor with autoradiography/fluorography. Unfortunately, the sensitivity of this approach is weak since the 3H-labeled anchors emit low-energy radiation and the amount of purified protein with attached anchor is typically very low before expression and modification conditions have been optimized individually for the specific target. The necessary exposure time is a priori unknown and varies widely depending on the target even after optimization of experimental conditions. Several experimenters reported long exposure times of up to several weeks (7 – 30 days , 3 – 14 days , at least one week ). To confirm the efficiency of our protocol, we did a comparison of our detection method to a Western membrane treated with Perkin Elmer "En3Hance-spray", which is the membrane equivalent to a gel soaked in "Amplify" scintillation liquid, and found weak signals after one week and distinct signals only after three weeks (Figure 2, compared to Figure 1). TLC scanning is an efficient alternative to autoradiography/fluorography.
Even having a negative outcome after 2 months exposure does not clarify whether the protein investigated cannot be prenylated or whether the amount of protein after purification is simply too low, for example due to unspecific adsorption to test tubes promoted by the prenyl anchor. It should be noted that the issue is not resolved with a [35S]methionine-based control of translation in a parallel experiment. However, this problem is circumvented by another advantage of the present protocol. It has the possibility to evaluate the amount of protein directly from the Western blot that has been used for TLC scanning. If a band is detected with low intensity of anti-GST- or anti-HA-antibody binding, this indicates that the expression of the target protein must be up-scaled.
Filter binding assays (i.e., the separation of proteins transcribed and translated in vitro in the presence of radioactive anchor precursors from free anchors with filters) provide another quick alternative to autoradiography. The information from such a test is limited since there is no size resolution of the protein mixture and no possibility to directly evaluate the amount of target protein. Furthermore, free radioactive anchors will non-specifically adsorb to the filter material or proteins resulting in high background signals.
To our knowledge, the current method is the first one utilizing a TLC-Scanner for the analysis of putative targets of PTases on Western blot membranes. Compared to autoradiography/fluorography, this approach reduces the detection time from several weeks/months to 20 minutes per lane, resulting in an overall time effort for the whole experiment of about three days, given that the cDNA of the GST- (or HA-tag-) fusion protein is already available. Additionally, this assay detects incorporated 3H-label and translation efficiency of the same reaction, rendering control reactions with [35S]methionine redundant and reducing variability of results caused by pipetting inaccuracies. In conclusion, the TLC scanning method is more sensitive and offers a more reliable way of quantification of any covalently linked 3H-labeled posttranslational modifications in much less time compared with autoradiography. Especially, when conditions for in vitro or in vivo protein expression and incubation still need to be set up or optimized, this method dramatically enhances chances of generating successfully generating reproducible results in reasonable time since the experimental cycle is considerably shortened.
The use of a GST-tag (or a HA-tag) provides a way of removing free radioactive label as well as separation from highly abundant proteins from the rabbit reticulocyte (in vitro test) or cell culture lysate (in vivo test), resulting in lower background signal. In addition for proteins with lower translation efficiency than Rap2A, it offers the opportunity to use bigger reaction volumes or larger cell cultures and load the whole yield on a gel without exceeding its capacity. Furthermore, it provides the opportunity to use the same primers and antibodies for all investigated proteins, making the adaptation of the assay to screening with higher throughput only a small step.
We have shown that the results obtained with our TLC scanning method are in line with those from parallel experiments testing electrophoretic mobility shifts (Figure 7) or subcellular localization changes (Figure 8) due to prenyl anchor attachment. It should be emphasized that the latter two methods are indirect and leave room for alternative interpretations whereas our assays based on TLC scanning provide a much stronger argument. It unequivocally shows anchor incorporation into the target protein directly in vitro as well as in vivo.
Unfortunately, we were not able to find any conditions, which reproducibly allowed detection of the prenylated peptide with HPLC-based purification methods after proteolytic digest. Most of the times, there was no visible difference between wildtype and mutant Rap2A in UV-signals as well as radioactivity measurements, except the fact that the total peptide content was significantly (ca. 3 times) higher for the mutant protein. These findings suggest that, in contrast to the non-prenylated protein, a considerable amount of prenylated protein is lost by unspecific adsorption to the walls of the Eppendorf tubes, vials, tubes and microwell plates used for the reaction and following processing steps. This is in agreement with our observation from the Western blots of the TLC scanning method (Figures 1, 2, 3, 4, 5, 6 and 9), which in almost all cases showed much higher protein content for the non-prenylatable mutant proteins. Additionally, we derived much better results when performing reaction and purification steps the same day, storing the samples in SDS-PAGE sample buffer at -20°C over night. Storage of the protein in the reaction mix or in PBS without detergent resulted in decreased radioactive signals. From these observations, we suggest that there would be even higher losses of the prenylated peptide after digestion, since the properties of the shorter polypeptide are much more dominated by the hydrophobic isoprenoid group, leaving only undetectable amounts of labeled prenylated peptide in solution.
These problems show that the chromatographic method might not be applicable for the small amounts of protein yielded by in vitro transcription/translation. There might be the possibility to overcome most of the troubles by simply increasing the amount of target protein. Results obtained with Rap2A expressed in HeLa-cells, purified by immunoprecipitation and digested with trypsin showed significant discrepancy between the UV-signals of wildtype and mutant protein (data not shown). One peak with a retention time close to the one of FPP had a peak area ~10 times larger for wildtype protein, while all other peaks were nearly identical. Based upon these promising preliminary experiments, future work may find a mass spectrometry-coupled HPLC-based approach useful for in vivo prenylation analysis.
We think that the mechanistic role of the prenyl anchor for the biological function of the proteins studied in this work still requires additional research. For the convenience of the reader, we summarize the current state of knowledge with respect to the molecular and cellular functions of the investigated protein targets for prenylation in the following four paragraphs. Apparently, Rap2A, RasD2, K-Ras and RhoA need the prenyl anchor to get translocated into the right signaling context by membrane association. Rap2 has been shown to promote integrin activation  and to directly bind to the actin cytoskeleton of platelets . Rap2A is regulated by the same GEFs and GAPs as Rap1, but with much lower efficiency for the GAPs. This results in a high ratio of GTP-bound protein. Rap2 may be a slow molecular switch with functions similar to Rap1, but while the latter transduces strong, transient signals, Rap2A could determine the basal level. Thus, Rap1 would be required in the initial step of cell adhesion, which is then maintained by Rap2 signaling .
RasD2/Rhes (ras homolog expressed in striatum) is expressed predominately in the striatum  but also in thyroid glands and pancreas β-cells . It is involved in selected stritial functions, mainly locomotor activity and motor coordination . Unlike the Ras proteins, RasD2 does not activate the ERK-pathway, but it binds and activates phosphoinositide 3-kinase (PI3K). Additionally, RasD2 impairs activation of cAMP/PKA pathway by thyroid stimulating hormone (THS), as well as by activated β2-adrenergic receptor, suggesting a regulating function upstream of activation of the respective heterotrimeric G-protein complex. The mechanism of action implies uncoupling of the receptor from its downstream target .
The Ras proteins have been reported to be involved in many signaling pathways, regulation cell differentiation and proliferation as well as cell shape and motility, to mention only the most important. Ras proteins are GTPases that function as molecular switches, being active in GTP-bound state and inactive when GDP-bound. The different Ras proteins show high homology to each other and collaborate in a complex network, making it hard to distinguish whether their functions are provided by all of them or are unique for a certain type of Ras protein. Nevertheless, there is some experimental data indicating specific functions of K-Ras4B in cell-cell and cell-matrix contacts as well as in apoptosis . These presumptions are supported by the fact that K-Ras4B has a different strategy for membrane association than H-Ras, N-Ras and K-Ras4A, with a polylysine stretch in the hypervariable region instead of cysteine residues as palmitoylation sites. This comes along with localization of K-Ras4B to different microenvironments of membranes and also a trafficking pathway different from the other Ras proteins .
In humans, there are three highly homologous isoforms of Rho GTPases, called RhoA, B and C . Similar to the Ras proteins, their activities are highly overlapping, explaining why reported functions are hardly ever assigned to a certain family member. Regulation of the actin cytoskeleton, particularly the formation of stress fibers, was the first reported function of Rho. Further investigations have revealed roles in the regulation of cell polarity, gene transcription, G1 cell cycle progression, microtubule dynamics and vesicular transport pathways . Thus, it seems that Rho proteins play a major role in vital cell functions such as morphogenesis, chemotaxis, axonal guidance and cell cycle progression .