Materials & Methods

Miscellaneous enzymes and chemicals
All restriction endonucleases were purchased from New England Biolabs (Genesearch Pty Ltd, Arundel, Queensland), Klenow fragment from GeneWorks Pty Ltd (Thebarton, South Australia), and both T4 ligase and T4 DNA polymerase from Roche Diagnostics Australia (Castle Hill, New South Wales). 4',6-diamidine-2-phenylindole' dihydrochloride (DAPI) and 3-Amino Triazole (3AT) were obtained from Sigma-Aldrich (Castle Hill, New South Wales).

The yeast two-hybrid screen
The ProQuest™ yeast two-hybrid system (Invitrogen, Mulgrave, NSW) was employed to screen for potential interacting partners of MID1. In order to generate the "bait" construct, the full-length human MID1 cDNA was subcloned from pBSMID1 [5] using SalI and NcoI into the similarly restricted pDBLeu vector. Subsequent digestion with SalI, end-filling with Klenow fragment and religation generated the full-length MID1 cDNA in-frame with the GAL4 DNA binding domain (Gal4DBD). The construct, pDBLeu-MID1, was verified by sequencing.
The selection of an appropriate cDNA library was deemed critical to maximise the chance of detecting bona fide interacting factors. The expression of the murine Mid1 gene during embryological development is consistent with the clinical presentation of OS [27]. Given the very high level of primary sequence identity between all vertebrate MID1 proteins (unpublished data), a murine 10.5 dpc whole embryo cDNA library (Invitrogen) directionally cloned in the GAL4 activation domain (GAL4-AD) plasmid, pPC86, was selected for use as the "prey" in the two-hybrid screen.
The pDBLeu-MID1 construct was transformed into MaV203 strain (Genotype: MATα, leu2–3, 112, trp1–901, his3Δ200, ade2–101, gal4Δ, gal80Δ, SPAL10::URA3, GAL1::lacZ, HIS3UAS GAL1::HIS3@LYS2, can1R, cyh2R) along with the parental pPC86 plasmid as per the manufacturers instructions (Invitrogen). The HIS3 reporter was used to determine the level of self-activation of the MID1-GAL4DBD fusion based on the level of 3-amino triazole resistance (3ATR) of the fusion protein. A concentration of 50 mM 3AT was found to be an adequate level for the assay, although subsequent analyses were performed on 75 mM 3AT plates to further reduce background transactivation of the reporter genes. To screen for potential interacting proteins, the pDBLeu-MID1 fusion construct was co-transformed with the 10.5 dpc mouse embryo cDNA-pPC86 library according to standard protocols (ProQuest™ Yeast Two-Hybrid Manual, Invitrogen) and plated on a synthetic complete medium (SC-Leu-Trp-His) containing either 50 mM or 75 mM 3AT.
The full-length MID2 (FXY2) cDNA was cloned from pEGFP-FXY2.ORF [15] into pDBLeu using the same strategy as used for MID1. As Mid2 is generally expressed in many of the same tissues during embryological development as Mid1 albeit at considerably lower levels [19], the resultant clone, pDBLeu-MID2, was consequently used as "bait" in a similar screen of the 10.5 dpc embryo cDNA library as well as directly against the identified MID1 interacting clones.
For confirmation of putative interacting clones, constructs and library clone isolates (in both parental vectors and swapped vectors) were re-transformed into the MaV203 yeast strain using the LiAc method [37]. Transformed yeast cultures were incubated for 24 hours at 30°C in selective media. Cultures (10 μl – 0.1 OD600) were then spotted onto selective plates and incubated at 30°C for 48 hours. Replica cleaning of plates was performed as required.

Generation of GFP- and myc-full-length cDNA fusion constructs for immunofluorescence
The generation of full-length MID1-GFP and MID2-GFP fusions in pEGFP have previously been reported [15,5]. A vector for the production of myc-tagged fusion proteins was generated by modification of the pEGFP-N2 vector (Clontech, Palo Alto, CA). Briefly, the GFP coding region was excised from pEGFP-N2 with NotI and BamHI, the 5' overhangs filled using T4 DNA polymerase and the vector religated to give pCMV-N2. Six copies of the myc epitope containing a start codon was amplified from pGEM-6mycT (gift from M. Whitelaw, University of Adelaide), and cloned into the HindIII site of pCMV-N2. To facilitate in-frame insertion of cDNAs directly from the pPC86 library vector, the plasmid was linearised with EcoRI, end-filled and re-ligated to create pCMV-6myc-ΔE. In order to clone Alpha 4 into pCMV-6myc-ΔE, pPC86-Alpha 4 was digested with SalI/AatII and the full-length cDNA insert ligated into SalI/SmaI restricted pCMV-6Myc-ΔE vector. The reading frame of the construct was confirmed by automated sequencing.

Generation of MID1 and MID2 domain-specific deletion constructs for immunofluorescence and yeast two-hybrid analyses
The FNIII domain-specific deletions in both MID1 and MID2 were generated by precise deletion of the domains by a two-step PCR strategy. The other domain-specific deletion constructs were generated using QuickChange™ site-directed mutagenesis (Stratagene, La Jolla, CA) to introduce unique restriction sites, as required, at the start and end of each domain within MID1 and MID2. Exceptions to this were the RING and B-box deletions of MID1 where a native XbaI site located at nucleotides +207–212 was used as the 3' and 5' excision point in the respective constructs, and the CTD deletions in both MID1 and MID2 where a native BamHI site located at nucleotides +1464–1469 was used as the 5' excision point for these constructs. Restriction sites were chosen so that, where possible, the encoded amino acid sequence remained unaltered or only resulted in conservative substitutions. Furthermore, the sites were positioned such that excision of individual domains did not alter the reading frame of the encoded protein (Table 1). In any one construct, a maximum of two restriction sites were introduced. Details of the strategies and primers (for both PCR and site-directed mutagenesis) used in the generation of these deletion constructs will be forwarded upon request to the corresponding author. Each introduced restriction site was confirmed by digestion and sequencing and then independently tested for its effect on the microtubule binding capacity of the respective proteins. In each case, this was determined by direct visualisation of fluorescence of the created GFP fusion protein. No introduced change had any appreciable effect on the microtubular distribution of the proteins.
To further test the function of some of the regions of the MID1 protein, selected separate motifs were generated using the appropriate existing, and/or inserted, restriction sites and ligated in-frame and C-terminal to either EGFP (in pEGFP) or Gal4DBD (in pDBLeu). The following constructs were generated: the MID1 B-box fusion (pDBLeu-M1BB) containing residues 71–213, the MID1 coiled-coil fusion (pEGFP-M1CC) containing residues 214–349, and the MID1 B-boxes plus coiled-coil fusion (pEGFP-M1BBCC) containing residues 71–349.

Transfection and immunofluorescence analysis of GFP-MID1 constructs
Preparations of the various GFP- and myc-tagged expression constructs were made using the Qiagen Midi kit (Qiagen, Clifton Hill, Victoria). Two picomoles (approximately 1 microgram) of each construct were transfected into cultured cell lines (Cos1, HeLa, NIH3T3) using FuGene transfection reagent (Roche Diagnostics Australia). Transfected cells were grown on coverslips in DMEM plus 10% FBS and fixed 24 hours post-transfection as previously described [5].
In test transfections, where only a single GFP expression construct was introduced into cells, control microtubule staining was performed post-fixation using an anti-α tubulin antibody plus an anti-mouse Texas Red-conjugated secondary antibody (Jackson Laboratories, Bar Harbor, Maine). In cells transfected with myc-tagged expression constructs (either alone or in combination with a GFP-tagged expression construct), anti-α tubulin staining was not performed. Instead the Texas Red-conjugated secondary antibody was used in combination with an anti-myc monoclonal antibody (9E10) to detect the expression of the myc-tagged protein. In all cases, nuclei were stained using the DNA-specific stain, DAPI. GFP and Texas Red fluorescence were visualised under appropriate wavelength light on an Olympus AX70 microscope. Images were captured using a Photometrics CE200A Camera Electronics Unit and processed using Photoshop 6.01 software (Adobe Systems Incorporated, San Jose, California).

Immunoprecipitation and western analysis
Preparations of the various GFP- and myc-tagged expression constructs were made using a DNA plasmid Midi kit (Qiagen). Six picomoles (approximately 3 micrograms) of each construct were transfected into approximately 107 Cos1 cells using FuGene transfection reagent (Roche Diagnostics Australia). After 24 hours incubation, cells were scraped from the culture dish and lysed on ice for 30 minutes in 1 ml lysis buffer (50 mM Tris-Hcl pH 7.4, 300 mM NaCl, 5 mM EDTA, 1.0 % Triton X-100). Cell lysates was cleared by centrifugation at 4°C (15 minutes, 16 ×g), and protein extract recovered as supernatant. After pre-clearing 200 μl of protein extract with 10 μl of 50% protein-A sepharose bead slurry, extracts were incubated with 1 μg of antibody for 2 hours at 4°C and then for another 2 hours with 20 μl of fresh 50% protein-A sepharose bead slurry. The beads were washed four times with wash buffer (50 mM Tris-Hcl pH 7.4, 300 mM NaCl, 5 mM EDTA, 0.1% Triton X-100) and protein eluted from the beads by boiling in 2 × SDS load buffer. Proteins were separated by 8% SDS PAGE and blotted onto Hybond-C membranes (Amersham Pharmacia) using a semi-dry transfer apparatus (BioRad). Membranes were blocked, incubated with the appropriate primary antibody, washed, incubated with the appropriate HRP-conjugated secondary antibody and washed again according to established method described in Current Protocols in Cell Biology. Detection was carried out using an enhanced chemiluminescence reagents (ECL) kit (Amersham Pharmacia) as per the manufacturer's instructions. Antibodies used in immunoprecipitation and western blot analysis included; rabbit polyclonal anti-GFP antibody (gift from Pam Silver, Dana-Farber Cancer Institute, Boston), mouse anti-c-myc monoclonal antibody (gift from Stephen Dalton, University of Adelaide), rabbit polyclonal anti-phosphoserine and anti-phosphothreonine antibodies (Zymed) and HRP conjugated anti-rabbit and anti-mouse secondary antibodies (Amersham Pharmacia).

Computer-assisted detection of serine/threonine phosphorylaton sites
Analysis of putative consensus serine/threonine phosphorylation sites in MID1 was performed using NetPhos version 2.0  software [38]. Examination of a multiple sequence alignment of all available MID1 and MID2 sequences was performed in order to determine if any putative phosphorylation site was fully conserved. For this analysis, orthologous MID1 sequences were assessed from a variety of species, including human, mouse, rat, chick, tammar wallaby, zebrafish (partial sequence only) and fugu, and MID2 orthologous sequences from human, mouse and rat (partial sequence only).

Note Added In Proof
During the review of this manuscript, a paper by Trockenbacher et al [Trockenbacher A, Suckow V, Foerster J, Winter J, Krauβ S, Roper H-H, Schneider R. and Schweiger S: MID1, mutated in Opitz syndrome, encodes an ubiquitin ligase that targets phosphatase 2A for degradation. Nature Genetics 2001, 29:287–294] independently reported the interaction of Alpha 4 and MID1. These investigators also showed that MID1, possibly through Alpha 4, regulates the turnover of the microtubule-associated fraction of PP2AC and hence may represent a possible pathological mechanism for the Opitz syndrome phenotype.