Many microtubule-associated proteins have been shown to move along the microtubules. In order to study migration of the ubiquitin ligase MID1 along the microtubules, we transfected HeLa cells with GFP-tagged MID1 (MID1-GFP) and analysed them in a laser-scan microscope for FRAP. As described previously [10], [18], wild-type MID1-GFP showed a defined microtubule-associated pattern. Full recovery of the GFP signal after bleaching was obtained within a few seconds (30–60 sec, Fig. 1a). Similarly fast recovery was observed in cell-bodies of F11 cells, a hybrid cell line of mouse neuroblastoma cell line N18TG-2 and embryonic rat dorsal-root ganglion (DRG) neurons (Fig. 1b), while fluorescence returned significantly more slowly in axons (60–120 sec, Fig. 1b, c). Detailed analysis of the recovery of the green signal in axons demonstrated recovery from the centrosome and from the periphery suggesting a bi-directional mechanism (Fig. 1d).

Microtubule-dependency of the transport of MID1-GFP was further analysed in HeLa cells treated with drugs that interfere with microtubule dynamics. While a few bundles were still left after 3 hours, all microtubules were destroyed after treatment with 100 ng/ml colcemide over a period of 16 hours. As shown in figure 2A, after 16 h, apart from few green dots resulting from protein diffusion in the cell, no fluorescence recovery of the MID1-GFP signal was observed even after 180 seconds. By contrast, treatment of the cells for 5 hours with 5 µg/ml taxol, a drug that stabilizes microtubules (Fig. 2b), significantly accelerated fluorescense recovery indicating that the transport of MID1-GFP relies on intact microtubule-dynamics.

Two different classes of molecules are known to actively transport proteins along the microtubules. While kinesins transport towards the plus ends of microtubules and therefore the cell periphery, dyneins are adjusted to the minus ends, which locates at the organizing centre of microtubules. As suggested from the bi-directional transport seen in axons of F11 cells (see above), inhibition of both molecule classes significantly influences the recovery rate of the MID1-GFP signal. Treatment of cells with 10 mM of Erythro-9-(2-Hydroxy-3-Nonyl)Adenine (EHNA), an inhibitor of dynein activity [19], led to inhibition of retrograde transport and fluorescence recovery only took place from the centrosome (Fig. 3b and d). On the other hand, treatment of cells with aurintricarboxylic acid, an inhibitor of kinesin activity [20], resulted in a slow-down of recovery from both directions as compared to the mock-treated control (Fig. 3a, c and d). Both analyses were performed in axons.

Various mutations in the MID1 proteins have been identified in OS patients [1], [3], [10], [12]–[17]. Most of those are located in the C-terminus of the protein and interfere with its binding to microtubules [10]. Apart from these, missense mutations in the B-box1 domain of the MID1 protein have been identified [1], [14]. Proteins carrying these mutations still bind to microtubules (BA et al., manuscript in preparation), but fail to bind the α4 protein [7] (BA et al., manuscript in preparation). We have used MID1-GFP proteins carrying three different mutations in this domain, C145S , A130T (both patients identified in this study, see Methods section) and ΔVTC [13] for FRAP analysis (Fig. 4a). For comparison, we analysed two mutations in the coiled-coil domain, C266R [13] and L295P [1], which also do not affect microtubule binding of the protein (Fig. 4b). Interestingly, while the last two showed fluorescence recovery comparable to the wild-type (Fig. 4b, c and d), no recovery was seen with either of the B-Box1 mutations (Fig. 4a and d), suggesting that functional B-box1 domain is required for the correct transport of MID1 at the microtubules.

Since MID1 interacts with α4, and thereby with PP2A, through its B-box1 domain, the previous FRAP data pointed towards a functional role for α4 and PP2A in the microtubule-associated transport of the MID1 protein. In order to confirm this, we knocked down α4 with specific RNAi oligonucleotides and analysed these HeLa cells for fluorescence recovery of MID1-GFP. Confirming our hypothesis, no recovery could be observed in cells with α4 knock-down (Fig. 5b and d) even after 180 seconds , while the recovery rate in cells treated with non-silencing siRNAs (control, Fig. 5a and d) was comparable to the wild-type. Efficiency of the knock-down is shown on a Western-blot using an anti-α4 antibody (Fig. 5c). Interestingly, while tubulin staining showed no changes in microtubules organization in cells with α4 knock-down compared to cells treated with nonsilencing siRNAs (data not shown), microtubule association of MID1-GFP seemed to be slightly disturbed after α4 knock-down (Fig. 5b).

Similarly, treatment of cells with the PP2A inhibitors okadaic acid (OA) and fostriecin (FST) led to a complete inhibition of fluorescent recovery (Fig. 6b and c). Full recovery after 30–60 seconds was seen in the mock-treated control (Fig. 6a and c), while no recovery was seen after OA and FST treatment after 180 seconds, once again indicating a central involvement of PP2A activity in the microtubule-associated transport of MID1-GFP.

It has been suggested previously that MAP kinase and PP2A regulate the phosphorylation status of MID1 on serine 96. Therefore, its PP2A dependent dephosphorylation could be necessary for proper microtubule-associated transport along the microtubules. To test this hypothesis, we produced three different point mutations on serine 96. Two of them, S96D and S96E, brought negative charges and thereby simulated continuous phosphorylation (Fig. 7b), whereas S96A completely inhibited phosphorylation (Fig. 7c). As expected, while S96A showed recovery rates comparable to the wild-type (Fig. 7a, c and d), neither S96D nor S96E were transported or showed fluorescent recovery even after 180 seconds (Fig. 7b and d). In contrast to the wild-type, transport of the phosphorylation free S96A mutant along the microtubules was not inhibited by okadaic acid treatment prior to analysis (Fig. 7e). Also, cells overexpressing wild-type MID1-GFP that were pretreated with the MAPK inhibitor UO126 did not react on okadaic acid treatment, confirming a functional role of MAPK dependent phosphorylation in the regulation of the transport of the MID1-GFP protein along the microtubules (Fig. 7f). Surprisingly, contrary to previously suggested data [11], all S96 mutant proteins showed undisturbed microtubule-associated pattern. Also, UO126 treatment did not interfere with microtubule-association of the wild-type protein (Fig. 7f).

Two novel mutations found in OS patients have been used: One patient presented with hypertelorism, broad nasal bridge, strabismus, cleft lip and palate, hypospadias and small ears with a right pre-auricular pit. He was found to have a de novo 388G>A mutation in MID1, predicting an A130T change in the B-box1 domain. The second patient had hypertelorism, down-slanting palpebral fissures, broad nasal bridge, posteriorly rotated ears, cleft lip and palate, and hypospadias. He was found to harbour a de novo 433T>A MID1 mutation, predicting a C145S change in the B-box1 domain.

In-vitro mutagenesis experiments were performed on MID1 in the pEGFP-C1 vector (MID1-GFP) [10] and according to the instruction manual provided with the QuickChange® Site Directed Mutagenesis Kit (Stratagene).

1×10⁵ HeLa per well of a 6-well plate were grown for 24 hours on glass cover slips and transfected with 1 µg plasmid DNA using 5 µl of lipofectamine (Invitrogen) according to the manufacturer's instructions.

2×10⁵ F11 cells per well of a 6-well plate were grown for 24 hours in Ham's F-12 medium supplemented with 10% FCS on glass coverslips and were transfected using 5 µl of lipofectamine and 8 µl of Plus Reagent (Invitrogen) in 2 ml OptiMEM. After 2 h medium was changed to Ham's F-12 containing 18% FBS. Medium was changed again after 3 h to Ham's F-12 supplemented with 10% FBS.

Fluorescence recovery after photobleaching (FRAP) analysis of MID1-GFP was carried out on cultured F11 neurons and HeLa cells. Transfected cells were randomly selected and bleached for between 0.5 and 1.1 s using 100% power from a 25 mW argon ion laser. Recovery was imaged at low laser power, and cells were examined for ∼4 min, with imaging approximately every 2 sec. Percent recovery was determined by subtracting arbitrary average background values outside the cell from average values from the whole cell or selected subregions, then dividing this by the difference in fluorescence intensity of the selected region before bleaching and the average background values (I_spot−I_bkgd (post-bleach)/I_spot−I_bkgd (pre-bleach). Each of the traces was calculated from an average percent recovery of n = 5 cells at each time point, starting at 4 s post bleach.

Series of 50–100 single section images were collected with the help of a time series programme. For imaging, the laser power was attenuated to 2% of the bleach intensity. Images were acquired on an LSM 510 (Carl Zeiss, Jena, Germany) with the Planapochromat 63-/1.4 objective. EGFP fluorescence was detected using the 488 laser line of an argon laser (25% of 25-mW nominal output) in conjunction with a LP 505 filter. All live cell imaging was done at room temperature. Images were analysed with the Zeiss LSM image examiner software and Image J.

20–24 h after transfection, cells were treated for 30 min with 50 nM okadaic acid (Sigma), 10 µM U0126 (Sigma), 200 nM fostricien (Sigma), 10 µM aurintricarboxylic acid (Calbiochem) or 1 mM Erythro-9-(2-Hydroxy-3-Nonyl) Adenine (Calbiochem). In addition, cells were treated with 100 ng/ml colcemid (Biochrom) for 3 and 16 h or 5 µg/ml taxol (Sigma) for 5 h. Cells were washed with PBS and analysed under a LSM 510 microscope as described above.

1×10⁵ HeLa cells were seeded in a 6-well plate 24 hours before transfection. Cell were incubated with a solution containing 1.5 µg DNA, 3 µl of siRNA oligo (α4 oligonucleotide 1: sense CGUUGCUAUGGCAUCUCAAdTdT/antisense UUGAGAUGCCAUAGCAACGdTdT, α4 oligonucleotide 2: sense GUACCUUUUGGUGCCAGCGdTdT/antisense CGUUGGCACCAAAAGGUACdTdT, non-silencing oligonucleotide: sense UUCUCCGAACGUGUCACGUdTdT/antisense ACGUGACACGUUCGGAGAAdTdT, Qiagen), 2 µl dreamfect (Biosiences), and 100 µl of OptiMEM (Sigma) in 2 ml of medium without antibiotics. Cells were incubated at 37°C in a humidified incubator with 5% CO2 v/v for 72 h.

200 µg of proteins were loaded on a 12% SDS-page, blotted on a PVDF membrane (Roche) and incubated with a specific anti-α4 antibody (Trockenbacher et al., 2001) at 4°C overnight. For loading control, the same blot was incubated with an anti-tubulin antibody.