The distribution and effect of expressed Willin constructs on mammalian

cells

4.2.1. Construction of the pWillin-FLAG plasmid

A previously constructed plasmid (pWillin-GFP) expressing Willin fused to the N-terminus of EGFP was shown to express this chimeric protein (Willin-GFP) such that it had an intracellular distribution that was different from the ERMs and Merlin, and that varied according to the cell type within which it was expressed and

even within a cell type it was dependent on growth factor activation, cell-cell contact, and as-yet other undetermined factors (Gunn-Moore et al., 2005). To see if Willin with a smaller tag would behave in the same way, pWillin-FLAG was constructed. The most convenient cloning sites in the pCMV-Tag 4A vector were SacI and BamH I; however, there is a single SacI site within Willin at base pair 1507. Therefore, a two-step cloning strategy was devised. First, pWillin-GFP was digested with the restriction enzyme SacI, releasing the first 1507bp fragment of Willin. Then, using pWillin-GFP as a template, the forward primer 5’ CCACCTCGAGCTCTTCAG 3’ containing a SacI site, and reverse primer 5’

CGGGATCCCACAACAAACTCTGGAAC 3’ containing a BamH I site, a PCR product containing the remaining 338bp of Willin was produced and subsequently digested with the restriction enzymes SacI and BamH I. The pCMV Tag-4A vector was also digested with SacI and BamH I, and the shorter fragment of Willin was ligated with it. The ligation mixture was transformed intoE. coliand prepared as usual, and restriction digest analysis showed the ligated plasmid was present. This plasmid was then digested with SacI, treated with alkaline phosphatase (section 2.1.3) and the longer Willin fragment ligated with it. As two orientations for this insert were possible, several digests were performed to test both the presence and direction of the insert, and the complete plasmid was further confirmed by sequencing. Figure 4.1 summarises the cloning strategy of pWillin-FLAG.

Willin GFP Neo/Kan P_CMV SacI SacI Willin-GFP

Figure 4.1. Two-step cloning strategy for pWillin-FLAG. pWillin-GFP was digested with SacI to produce a 1507bp fragment; pWillin-GFP was also used as the template for PCR of the last 338bp of Willin. The two fragments were then sequentially ligated into pCMV-Tag4A to produce pWillin-FLAG.

4.2.2. Expression of Willin-FLAG

Willin-FLAG was transfected into HEK-293 cells to test for expression by Western blot analysis. A positive control lysate containing a truncated FLAG-Ezrin protein was provided by Dr. Fleur Davey. The Willin-FLAG protein was expected to

Sac I

Sac I Sac I

be about 72kDa, but as is usually seen with Willin, the detected protein had a lower apparent molecular weight (Figure 4.2).

1 2

Figure 4.2. Truncated FLAG-Ezrin positive control (Lane 1) and whole cell extract from Willin-FLAG transfected HEK-293 cells (Lane 2) were separated on a 4-12% Bis-Tris gel and the subsequent nitrocellulose membrane probed with anti-FLAG M2 1:500 and secondary anti-mouse-HRP (Santa Cruz) 1:10,000.

Immunocytochemistry was also performed in HEK-293 cells transfected with pWillin-FLAG (see section 2.2.8) to study the distribution of the Willin-FLAG protein. It was observed that Willin-FLAG has a comparable distribution pattern to Willin-GFP (see Figure 1.17B): localisation is punctate throughout the membrane and cytoplasm in the vast majority of cells (Figure 4.3). In addition, Willin-FLAG- expressing cells showed the same tendency towards high levels of cell death 48 hours post-transfection as those expressing Willin-GFP. This indicated that the GFP tag did not have a significant effect on Willin localisation and was not solely responsible for increased levels of cell death, and could thus continue to be used for other studies.

Co-expression of Willin-GFP and Willin-FLAG was attempted, but due to technical problems with the FLAG antibodies that were not FITC-tagged, images could not be obtained to show co-localisation.

Figure 4.3. HEK-293 cell expressing Willin-FLAG. Image is a maximum projection of 46 Z-sections acquired on Olympus IX70 Deltavision RT microscope with

Coolsnap 2HQ camera (Roper Sci). Images were deconvolved and converted to TIF format with SoftWorx (Applied Precision) and projection assembled using ImageJ software. Z-sections were 0.2m thick.

4.2.3. Construction of the pWillin-DsRed plasmid

Under my supervision, undergraduate project student Jessica Davis

constructed a pWillin-DsRed plasmid that could be used for co-localisation studies with GFP-tagged proteins of interest. PCR was not required; instead, the GFP sequence was digested from the pWillin-GFP plasmid using BamHI and NotI and replaced with DsRed2 that had been digested from pDsRed2-Mito with BamHI and NotI (Figure 4.4).

Figure 4.4. Cloning strategy for pWillin-DsRed2. DsRed2 was digested from the pDsRed2-Mito plasmid with BamHI and NotI, while GFP was digested from the pWillin-GFP plasmid. The DsRed2 fragment was then ligated into the Willin vector.

Once the sequence of the plasmid produced was verified by sequencing to ensure it was in frame, COS-7 cells were transfected with the plasmid. As can be seen in Figure 4.5, the Willin-DsRed2 protein was found to form aggregates in the

cytoplasm, and did not display the characteristic Willin distribution usually found with Willin-GFP. DsRed1 previously showed this tendency, and though DsRed2 is supposed to show drastically reduced aggregation (Clontech manual,

http://www.clontech.com/upload/images/ctq/full/CTQJUL01.pdf), it is likely that the cause of aggregation is DsRed itself. This construct could therefore not be used.

Figure 4.5. A single Z-section snapshot of a COS-7 cell expressing Willin-DsRed2. Single section snapshots were created using LCS Lite V2.61 Build 1538 and cropped with Adobe Photoshop 7.0 (no other post-processing)