In mammals, four genes of the B" regulatory subunit family of PP2A have been described, giving rise to six isoforms: PR72 6, PR130 6, PR70 7, PPP2R3L product 8, G5PR 9 and mPR59, an isoform sofar only found in mice 10.

Upon searching the NCBI database for murine B" family members via the BLAST algorithm, we retrieved the complete protein sequence of two murine PR72 isoforms: the human PR72 orthologue [GenBank:BAE21013] (referred to as mPR72/B"α2) and a shorter variant missing the last 41 residues [GenBank:BAC28935] (referred to as mPR72/B"α4). In addition, one mG5PR isoform [GenBank:NP_067504] (referred to as mG5PR/B"γ) and four mPR59 isoforms were retrieved: mPR59 as described by Voorhoeve et al. [GenBank:AAC98973] (referred to as mPR59/B"δ2) 10, two mPR59 variants containing different N-termini [GenBank:AAH59852 and BAE25309] (referred to as mPR59/B"δ1 and mPR59/B"δ3 respectively) and a mPR59/B"δ3 variant harbouring an alternative C-terminus [GenBank:AAH96544] (referred to as mPR59/B"δ4). Additionally, two partial protein sequences of murine PR130 homologues were retrieved, one containing the first 615 amino acids of the PR130 specific N-terminus [GenBank:BAC31413] and the other containing the last 203 amino acids of the specific N-terminus followed by 12 additional residues [GenBank:BAC37349], resembling Xenopus XN73 (referred to as mPR130/B"α3). Of the full-length mPR130 (referred to as mPR130/B"α1), only a predicted sequence [GenBank:XP_135153] could be found.

To confirm these in silico data and to get an idea about the abundancy of these clones, we scanned the NCBI database for murine EST-clones containing the B"-specific N- or C-termini. Only very few EST-clones were found carrying the sequence of the mPR72 C-terminus lacking 41 residues [GenBank:BE852268 and CA984544], the XN73-like truncated mPR130 C-terminus [GenBank:AV356684] and the alternative mPR59 C-terminus [GenBank:BB357167], making it rather unlikely that murine B" family members containing these termini (mPR130/B"α3, mPR72/B"α4 and mPR59/B"δ4) do abundantly exist. In contrast, five EST-clones of the specific N-terminus of mPR59/B"δ3 and multiple (> 15) EST-clones of the other B"-specific N- and C-termini were found, suggesting that mPR72/B"α2, mPR130/B"α1, mG5PR/B"γ, mPR59/B"δ1, PR59/B"δ2, and PR59/B"δ3 are much more abundant, and probably represent the main murine B" isoforms.

Since mPR59/B"δ1 and mPR59/B"δ3 are putatively novel, and to further confirm their existence, we performed a reverse transcription on NIH 3T3 and murine heart RNA using primer E₃ annealing in the mPR59 3' UTR. With the resulting cDNAs as templates and the use of five different primers, annealing with the specific N-termini of PR59/B"δ1–4 (B_δ1, B_δ2, B_δ3/4) and/or the specific C-termini of PR59/B"δ1–3 (E₂) and PR59/B"δ4 (E₁), we were able to generate several DNA fragments, which after sequencing, were shown to correspond to PR59/B"δ1, PR59/B"δ2 and PR59/B"δ3. In vitro transcription-translation reactions of these cDNAs revealed proteins of about 60 kDa (PR59/B"δ1), 55 kDa (PR59/B"δ2) and 40 kDa (PR59/B"δ3) (Figure 1A), implying that not only the messengers of PR59/B"δ1,δ2,δ3 are present in murine cells, but that they can also be translated into proteins. A cDNA for mPR59/B"δ4 could not be amplified using this approach, confirming its low occurrence in the EST database.

A similar approach was undertaken to confirm the existence of the four mPR130/PR72 isoforms. In this case RT-PCR of NIH 3T3 or murine heart RNA led to the amplification of mPR130/B"α1 and mPR72/B"α2 cDNA, but not of mPR130/B"α3 or mPR72/"α4. To confirm this at the protein level, we prepared protein extracts of murine embryonic fibroblast NIH 3T3 cells and several murine adult tissues: heart, in which both human PR130 as PR72 are highly expressed 6; adrenal gland, the tissue from which the mPR130/B"α3 EST-clone [GenBank:AV356684] was isolated, and brain. Three antibodies were used to perform immunoprecipitations of the different mPR130/PR72 isoforms: a PR72 AB, raised against full-length recombinant human PR72 11; a PR130_N AB, raised against the recombinant specific N-terminus of human PR130 (AA 1–664) and a peptide PR72 AB, raised against the first 19 amino acids of the human PR72-specific N-terminus. The results demonstrate immunoreactivity at 140 kDa and 70 kDa, corresponding to mPR130/B"α1 and mPR72/B"α2 (Figure 1B). No other specific bands could be observed, even upon prolonged exposure, confirming our in silico and RT-PCR analysis. Consequently, it can be concluded that should the mPR130/B"α3 and/or mPR72/B"α4 isoforms exist, their expression would be very low or highly restricted in time and/or space.

Together, six main B" family members are present in mice: mPR130/B"α1, mPR72/B"α2, mG5PR/B"γ, mPR59/B"δ1, mPR59/B"δ2 and mPR59/B"δ3, with mPR59/B"δ1 and δ3 being novel ones. A protein alignment of these murine B" isoforms (Figure 2) reveals that they all share a conserved region, harbouring both ASBD and both EF-hand motifs, with the exception of mPR59/B"δ3 which misses 31 amino acids of the first ASBD. They mainly differ within their N- and/or C-termini, which may be responsible for isoform-specific functions. Surprisingly, no evidence of murine PR70 and PPP2R3L product orthologues could be found. The murine proteins most closely resembling these isoforms are PR59/B"δ1 (58% identity, 66% similarity with hPR70) and PR59/B"δ2 (57% identity; 66% similarity with the PPP2R3L product).

An alignment of the nucleotide sequences of the murine PR72/B" isoforms suggests that all novel isoforms are the result of alternative splicing of the genes encoding previously described members. By BLASTN analysis of the murine genome using the murine cDNA sequences, we could solve the genomic organisation of these three murine B" genes (Figure 3). The intron-exon boundaries were deduced by comparing the sequences obtained from the genomic clones and the respective cDNAs. Almost all boundaries were found to follow Chambon's rule (GT-AG) for splice donor and acceptor sites 26 (Table 1).

The genomic organisation of the gene giving rise to the murine PR130 and PR72 isoforms (3222402P14Rik, MGI: 2442104) was deduced from clone RP24-308L17 [GenBank:AC120390], containing the entire gene. To retain the logic in the PR72/B" gene nomenclature, we propose to rename this gene into Ppp2r3a, analogous to the human PPP2R3A gene. The mPR130/PR72 gene is located on chromosome 9F1 (101 Mb). This region corresponds to human chromosome 3q22, where PPP2R3A is located. The mPR130/PR72 gene (Figure 3) spans approximately 106 kb and consists of 15 exons. In addition to the exons giving rise to mPR130/B"α1 and mPR72/B"α2, we could also locate both the exon containing the sequence of the 12 specific C-terminal amino acids of mPR130/B"α3 (exon 2), and the exon responsible for the premature ending of PR72/B"α4 (exon 14B) at the correct sites within the gene. Since PR72/B"α2 is the result of an intra-exonic splicing event within exon 14, PR72/B"α4 might well be the result of a splicing error at this position, possibly explaining its low abundance. In the human PR130/PR72 gene (PPP2R3A), the exon responsible for the putative generation of a human PR72/B"α4 isoform is also present. The NCBI human EST database contains exactly one EST [GenBank:CA427549] containing this exon. In contrast, no sign of the specific C-terminal tail of either mPR130/B"α3 or XN73 could be found in PPP2R3A. In addition, no human EST-clones were retrieved containing the PR130-specific N-terminus followed by nucleotides other than those of the common part of PR130 and PR72.

To unravel the genomic organisation of the murine PR59 gene (MGI: 1335093), we analysed NCBI contig NW_001030923. Currently, this gene's symbol is Ppp2r3a, but as one could easily mistake it for the murine orthologue of PPP2R3A, we propose to change its name into Ppp2r3d (see also further). This gene spans approximately 9 kb and consists of 13 exons. The exons encoding the complete sequences of all mPR59 isoforms were found at the appropriate locations (Figure 3). To find the human Ppp2r3d homologue, we scanned the human genome using the BLASTN program with the different mPR59 isoforms as input. The closest match was the human PR70/PPP2R3L product gene (PPP2R3B/L), followed by the human PR130/PR72 gene (PPP2R3A), confirming the similarity between PR70/PPP2R3L and mPR59 at the genomic level. Therefore, the murine Ppp2r3d and the human PPP2R3B/L gene may have evolved from a common ancestor gene. This hypothesis looks promising since to our knowledge no organism contains both a PR70/PPP2R3L product and a PR59-encoding gene. A phylogenetic tree, based on the protein sequences of various PR72/B" family members in man and mouse, further supports this hypothesis (Figure 4).

Analysis of the mG5PR genomic structure is based on NCBI contig NW_001030500, which is located on chromosome 12C2 (56 Mb, ENSEMBLE ContigView) and contains the entire gene (Ppp2r3c, MGI: 193009). The mG5PR gene (Figure 3) contains 13 exons and closely mimics the genomic organisation of the human gene 9.

With the exception of XN73 and mPR130/B"α3, all B" family members contain a conserved region and specific N- and C-termini. The conserved region, which contains two ASBDs and two EF-hand motifs, is necessary for PR65/A binding 3. Since the first ASBD is not intact in mPR59/B"δ3 (Figure 2), we wondered whether this isoform is still able to bind PP2A. To this end, we made GST-fusion proteins of the main B" family members and overexpressed these in COS-7 cells. After a GST-pull down assay, we evaluated the binding of both the PR65/A and catalytic subunit of PP2A via Western blotting. As expected, all main B" isoforms (hPR130, hPR72, hPR70, mG5PR/B"γ, mPR59/B"δ1 and mPR59/B"δ2) bind PP2A (Figure 5), and are therefore genuine B" subunits. In contrast, mPR59/B"δ3 fails to bind PP2A, suggesting it is not a regulatory subunit of PP2A (Figure 5). Like XN73 and mPR130/B"α3, it might be involved in regulation of PP2A by competing with the other mPR59 isoforms for binding to other binding partners.

We determined the mRNA expression of the genuine murine B" regulatory subunits via Northern blot analysis (Figure 6). Using an antisense probe which hybridises to the first 330 nucleotides of mPR130, transcripts of 4.3 kb and 7.0 kb were detected. These transcripts are probably generated by the use of different polyadenylation sites. However, we can not exclude the possibility that the 4.3 kb transcript represents mPR130/B"α3. Both mPR130 transcripts have a similar expression pattern, being high in kidney and heart, intermediate in skeletal muscle, brain, liver and testis and low in lung and spleen. These findings are consistent with the previously published ubiquitous expression of human PR130 6.

A probe specific for the N-terminal sequence of the mPR72 isoforms, detected a single band at 4.4 kb, likely corresponding to mPR72/B"α2. Similar to the expression pattern of PR72 in humans 6, mPR72/B"α2 is abundant in heart and skeletal muscle and barely detectable in other tissues (Figure 6).

As for mPR130/B"α1 and mPR72/B"α2, the transcripts of mPR59/B"δ1 and mPR59/B"δ2 were visualised using probes hybridising with their specific N-termini. The expression profiles obtained reveal a wide tissue distribution with some minor differences between the two splice variants: mPR59/B"δ1 expression is high in heart and liver, intermediate in kidney, brain and skeletal muscle and rather low in testis, lung and spleen, while mPR59/B"δ2 is highly abundant in heart, liver, brain, kidney and testis and less abundant in lung, skeletal muscle and spleen.

These results, together with the earlier observed broad mG5PR expression pattern 9 indicate a general high expression of all murine PP2A B" subunits in heart.

In order to generate viable knockout mice, it can be important to establish whether a function of the protein of interest can be expected during embryonic development, as a general knockout of such a gene might lead to an embryonic lethal phenotype. Expression of the mPR130/B"α1, mPR72/B"α2, mPR59/B"δ1, mPR59/B"δ2 and mG5PR/B"γ subunits was examined via Northern blotting at embryonic days 7, 11, 15 and 17 (Figure 7). Hybridisation was performed using the same specific probes as for the tissue distribution determination. For mG5PR/B"γ, a probe spanning the first 450 nucleotides of mG5PR/B"γ was used. No additional splice variants, specific for embryonic stages, were observed for any of the main murine B" subunits. Expression of both mPR130/B"α1 and mPR72/B"α2 transcripts increases as embryonic development proceeds. This might suggest a role for these proteins in fetal growth and development. In previous studies, both hPR130 as well as hPR72 have been reported to influence Wnt signalling 151617. In contrast, the expression of mPR59/B"δ1, mPR59/B"δ2 and mG5PR/B"γ remains constant during all stages of embryonic development (Figure 7).

All B" isoforms are highly expressed in heart. Moreover, mPR72/B"α2 expression is nearly restricted to muscle tissue. mPR130/B"α1, which has a broader expression pattern, is also substantially present in the different muscle tissues represented on the Northern blot. Since we have specific antibodies for both isoforms (Figure 1B), we examined the distribution of mPR130/B"α1 and mPR72/B"α2 in the different types of muscle tissue (heart, skeletal and smooth muscle) in more detail via immunohistochemistry.

In longitudinal sections of skeletal muscle, mPR130/B"α1 and mPR72/B"α2 both intensively stain the muscle fibers in a striated pattern (Figure 8A, e–f). To evaluate whether these bands represent the I- or A-bands, we counterstained the sections with iron hematoxylin. This dye gives the A-band a dark blue colour and colocalises with mPR130/B"α1 and mPR72/B"α2 (Figure 8B, f–g). In one of the muscle fibers stained with mPR130/B"α1 AB, we could even observe a darker stained band positioned in the centre of the A-band (Figure 8B, d). This band probably represents the H-zone. In sections of murine myocardium, a similar striated staining pattern could be observed in the longitudinally positioned muscle fibers. Cross-sectional muscle fibers did not show staining of any specific structures (Figure 8A, h–i). In bladder, neither mPR130/B"α1 AB nor mPR72/B"α2 AB stained the smooth muscle fibers of the muscular layer. In contrast, mPR130/B"α1 staining could be seen in the transitional epithelial layer, lining the inner cavity of the bladder (Figure 8A, b–c). In none of the cases, specific staining of any structures was observed using pre-immune serum (Figure 8A, a,d,g).

In conclusion, PR130/B"α1 and PR72/B"α2 staining seems to be a specific feature of the A-band striations of heart and skeletal muscle, whereas staining is very low or absent in smooth muscle.

Using GFP-fusion proteins, we determined the subcellular localisation of the hPR72, mPR59/B"δ1, mPR59/B"δ2, mG5PR/B"γ and hPR70 subunits in COS7 cells. Since the PR130 AB is suitable for direct immunofluorescence, we also evaluated the subcellular localisation of endogenous PR130 (Figure 9). As previously described 11, hPR72 is mainly present in the nucleus (N >> CP). PR130 is present in both the cytoplasm and the nucleus (N > CP). mPR59/B"δ1 (CP = N) and mPR59/B"δ2 (CP > N) appear both in cytoplasm and nucleus but are clearly less abundant in the nucleus than hPR72 and PR130. mG5PR/B"γ is more abundant in the nucleus than the cytoplasm (N > CP), corresponding to earlier published data 9, and additionally is highly abundant in the cell periphery. Surprisingly, unlike PR48 20, hPR70 has a more pronounced cytoplasmic expression (CP >> N). This diversity in observed subcellular localisation patterns demonstrates that each B" subunit may direct the phosphatase to distinct sites and substrates within the cell.

Several B" family members have been suggested to influence cell cycle progression 91011131420. Therefore we analysed the cell cycle profile of HeLa cells upon overexpression of several EGFP-B" fusion-proteins via FACS analysis. Compared to control cells (EGFP), overexpression of all EGFP-B" fusion proteins tested (hPR72, hPR130, hPR70, mG5PR/B"γ, mPR59/B"δ1 and mPR59/B"δ2) caused an increase in the G1 population of transfected cells (Figure 10A). This was even more pronounced in the presence of nocodazole, which arrests cycling cells in G2/M (Figure 10B). These results confirm and extend previously published effects of hPR72 11 and mPR59/B"δ2 10 overexpression on cell cycle progression.

[³⁵S]-Methionine (2.5 mCi/ml), [³²P]UTP, Protein G Sepharose and Glutathione Sepharose beads were obtained from Amersham Pharmacia Biotech. Restriction enzymes and DNA-modifying enzymes were purchased from Fermentas. Pwo proofreading polymerase (used in all PCR reactions) was from Roche Molecular Biochemicals. DNA oligonucleotides were purchased from Sigma-Genosys.

To search for the murine B" homologues in public databases we used the Nucleotide BLAST programs at NCBI http://blast.ncbi.nlm.nih.gov/Blast.cgi. For the alignment of proteins and generation of a phylogenetic tree, we used the multalin program http://bioinfo.genopole-toulouse.prd.fr/multalin/multalin.html. Percentages of similarity and identity between the amino acid sequence of two proteins were calculated using EMBOSS Pairwise alignment algorithm http://www.ebi.ac.uk/emboss/align. Data regarding the localisation of genes on the murine chromosomes were obtained using the ENSEMBLE Contig View http://www.ensembl.org/index.html, while the corresponding regions in human chromosomes were found using the ENSEMBLE Synteny View. Putative nuclear localisation signals have been predicted using the PSORT program http://psort.ims.u-tokyo.ac.jp, whereas putative nuclear export signals were predicted using the NetNES 1.1 Server http://www.cbs.dtu.dk/services/NetNES.

Anti-GST monoclonal antibody was purchased from Sigma. Anti PP2A_C and anti-PR65 monoclonal antibodies were generously supplied by Dr. S. Dilworth (Imperial College, London, UK). Anti-PR72_N pept. antibody was a generous gift of Dr. B. Hemmings (FMI, Basel, CH). Anti-PR72_rec. antibody has been previously described 11. For the generation of the anti- PR130 _{N rec.} antibody, BL21-pLys(RP) E. coli cells were transformed with PR130 AA1-664 in pET15b and induced with 0.2 mM IPTG overnight at 16°C. The inclusion bodies were purified 11, solubilised in 7 M guanidinium hydrochloride, and dialysed against 200 mM Tris. HCl pH 8.2 containing 0.5 M NaCl. The resulting protein solution was used to immunise rabbits (Unité d'Hormonologie, Marloie, Belgium). The antiserum was taken after 5 boosts with the antigen and used at 1/2000 for Western blotting and 1/100 for immunofluorescence. Before immunisation a sample of the preimmune serum was taken as a negative control.

Total RNA from NIH 3T3 cells and murine heart tissue was isolated using GenElute™ Mammalian Total RNA Kit (Sigma) and reverse transcribed using primer E₃ (5'-CTG TGC GCC CAG GAA CTG CGC-3') for the mPR59 isoforms and primers E₅ (5'-CCT TGC AGT CCT CTT CCA CCA GC-3'), E₇ (5'-GAT TCT GTC AGG GCT CTA GCA GG-3') and E₉ (5'-GGA TGC TGT GTA GAG ATG CTC TAG-3') for the mPR72/130 isoforms. The following primers were used to amplify the different mPR59 isoforms: B"δ1 (5'-CTC TGC CAT CAG TCT CTG CCC-3'), Bδ2 (5'-CAG GCC ACA CCC ACG GAT TCG-3'), B"δ3/4 (5'-CCA CAG GGT CTT CAG CCA CGC-3'), E₁ (5'-GCC CTG CCC ACT CTG ACT ATG-3') and E₂ (5'-GAC GAC CTG GAG CCT CTG TGA-3'). The primers used to amplify the different mPR72/130 isoforms are B₁₃₀ (5'-ATG GCA GCA ACT TAC AGA CTT GTG-3'), B₇₂ (5'-ATG ATC AAG GAA ACG TCC TTG CGA AG-3'), E₄ (5'-CTG TTG CTT GGG CAG CCT GAC C-3'), E₆ (5'-AGG TTC AGC TCC AAG AAG GGT GA-3') and E₈ (5'-CTT CAG TCA GTG GAT GAA GAG TAG-3'). The approximate location of the various primers within the genes is indicated in Figure 3.

[³⁵S]-Methionine-labelled proteins were obtained from pBluescript vectors containing the coding regions of the proteins, using the TNT-coupled rabbit reticulocyte lysate system (Promega) with the appropriate RNA polymerase (T3/T7).

Wild-type mice were anaesthetised with an intraperitoneal injection of pentobarbital (Nembutal, CEVA), transcardially perfused with ice-cold saline (NaCl 0.9%, B Braun) and dissected tissues were freeze clamped in liquid nitrogen. Tissues were weighed and homogenised on ice within a threefold volume excess of homogenisation buffer (25 mM Tris pH 7.6, 150 mM NaCl, 1 mM EDTA and 1 mM EGTA) containing protease inhibitors (Complete Protease Inhibitor Cocktail, Roche) using a douncer with 20–30 strokes. The homogenates were centrifuged for 10' at 13,000 g and the supernatant was collected. After preclearing with protein G-Sepharose (Amersham Biosciences), tissue lysates were incubated overnight at 4°C with anti-PR72_{N rec.}, anti-PR130_{N rec.} or anti-PR72_{N pept.}. After addition of 40 μl of protein G-Sepharose for 1 h, the immune complexes were washed 4 times with 1 ml TBS + 0.1% Nonidet P-40 before they were dissolved in Laemmli sample buffer. Bound proteins were separated by SDS/PAGE and transferred to nitrocellulose membranes. Individual proteins were detected with the specified antibodies, revealed by horseradish peroxidase-linked secondary antibodies (DAKO) and developed using the ECL kit (Amersham Biosciences).

Monkey COS7 cells were transfected (FuGENE 6, Roche) with pGMEX-T1, hPR72/B"α2-pGMEX, hPR130/B"α1-pGMEX, hPR70/B"β1-pGMEX, mPR59/B"δ1-pGMEX, mPR59/B"δ2-pGMEX, mPR59/B"δ3-pGMEX and mG5PR/B"γ-pGMEX in a 10-cm dish. 48 h after transfection, cells were rinsed with phosphate-buffered saline (PBS), lysed in 200 μl NET buffer (50 mM Tris pH 7.4, 150 mM NaCl, 15 mM EDTA and 1% NP-40) and centrifugated for 10' at 13,000 g. Cell lysates were incubated at 4°C for 1 h with NENT₂₀₀ buffer (20 mM Tris-HCl pH 7.4, 1 mM EDTA, 0.1% Nonidet P-40, 25% glycerol, 200 mM NaCl) containing 1 mg/ml bovine serum albumin and 25 μl glutathione-Sepharose beads (Amersham Biosciences) on a rotating wheel. The beads were washed 2 times with 1 ml of NENT₃₀₀ (NENT with 300 mM NaCl) containing 1 mg/ml bovine serum albumin and 2 times with 1 ml of NENT₃₀₀. Bound proteins were eluted by addition of 20 μl of Laemmli sample buffer and boiling. The eluted proteins were analysed by SDS-PAGE and Western blotting.

To generate isoform-specific PCR products, we used following primers: 5'-ATG ATC AAG GAA ACG TCC TTG CGA AG-3' and 5' GAA GAG GCT GAA GTC ATT TCA G-3', generating mPR72/B"α2 nt 1–120; 5'-ATG GCA GCA ACT TAC AGA CTT GTG-3' and 5'-GAA TAC CTG TAA CTT AAA GGA C-3', resulting in mPR130/B"α1 nt 1–330; 5'-atg gac tgg a aa gac gtg ctt c-3' and 5'-gcc aaa ctc ctt cat aca gat-3', producing mG5PR/B"γnt 1–450; 5'-ATG GCG CCG CTG ACG CCG CGG-3' and 5'-GAA CGA CGG GAC CCA GGA CG-3', producing mPR59/B"δ1 nt 1–243; and 5'-AGG CCA CAC CCA CGG ATT CG-3' and GTT TCC GTC GCG CTT CCA GAA G-3', resulting in mPR59/B"δ2 54 nt 5'UTR + nt 1–120 of coding region. We subcloned these PCR products into the pBluescript II SK-vector (Stratagene). The constructs were used to generate RNA probes. The RNA probes were transcribed and labeled with [³²P]UTP using Ambion's Strippable RNA probe synthesis and removal kit.

A Mouse Tissue MTN^® Blot containing poly(A⁺) RNA from eight murine tissues and a Mouse Embryo MTN^® Blot containing poly(A⁺) RNA from murine embryos at days 7, 11, 15 and 17 were obtained from Clontech Laboratories. The membranes were hybridised at 64°C using Ambion's Ultrahyb Ultrasensitive hybridisation buffer and washed with Northern Max Low and High Stringency Buffers from Ambion. They were exposed and analysed using ImageQuant (Molecular Dynamics). Probe removal was done using Ambion's Strippable RNA probe synthesis and removal kit. For quantification we used the ImageQuant program from Molecular Dynamics. Absolute values of the different tissues and embryonic stages were first divided by the absolute β-actin values of the corresponding tissue/embryonic stages. These values where then compared with the value of the most intense band on the tissue blot, which was given a value of 100%.

PCR fragments of hPR72/B"α2, hPR70/B"β1, mG5PR/B"γ, mPR59/B"δ1 and mPR59/B"δ2 were cloned into the SmaI restriction site of pEGFP-C1 (Clontech). 1 μg of each plasmid was transfected into COS7 cells grown on a glass coverslip in a 12-well dish. 24 h after transfection, cells were washed in PBS and fixed in PBS containing 4% paraformaldehyde for 10 min. After 3 washes with PBS, cells were mounted in Prolong gold antifade reagent with DAPI (Invitrogen). For immunofluorescence staining of endogenous PR130, COS7 cells grown on a glass coverslip were rinsed with PBS and fixed in PBS with 4% paraformaldehyde for 10 min. Subsequently, cells were incubated with PBS containing 1% bovine serum albumin (BSA, Sigma) for 30 min and then for 45 min with the PR130_{N rec.} primary antibodies, followed by incubation with Alexa Fluor 488 donkey anti-rabbit secondary antibody (Invitrogen) in PBS with 1% BSA. Finally, the cells were washed 3 times in PBS, once in water and mounted in Prolong gold antifade reagent with DAPI (Invitrogen). All slides were examined with a LSM-510 laser scanning confocal microscope (Zeiss, Jena, Germany), using a Plan-Neofluar 40×/1.6 Oil DIC objective.

Asynchronously growing HeLa cells were transfected with pEGFP-C1 or the GFP fusion plasmids of the different B" subunits in two 10-cm dishes per plasmid. If nocodazole (1 μg/ml) was used, it was added at this point for another 16 h before FACS analysis. 48 h after transfection, cells were trypsinised, washed in PBS and fixed for 5 min in 4% paraformaldehyde at room temperature. After washing, the cell pellet was incubated in 0.5 ml of PBS containing 100 μg/ml propidium iodide (Fluka) and 0.1% RNase for at least 1 h at room temperature. The samples were analysed with a Beckman Instruments Coulter Epics XL flow cytometer (Analis) on FL1 (for EGFP) and FL3 (for propidium iodide) using standard procedures, and the System II™ software (Analis) and WinMDI Version 2.8 software for quantification.

Wild-type mice were anaesthetised with pentobarbital (Nembutal, CEVA) and transcardially perfused with PBS containing 4% paraformaldehyde. Organs were removed, postfixed overnight at 4°C and paraffin embedded. Sections (7 μM) were mounted on silan coated glass slides (Menzel gläser). For antigen retrieval, dewaxed and rehydrated tissue sections were boiled in 10 mM citric acid (pH 6.0) for 10 min. After cooling the slides, endogenous peroxidase was removed with 0.03% H₂O₂ in methanol for 20 min and transferred to PBS solution. Subsequently, sections were encircled with a water-repellent PAP-pen (Zymed) and rinsed with PBS. After blocking with PBS containing 1% BSA for 45 min, sections were immunoreacted for 1 h at RT with primary antibodies, washed 3 times with PBS containing 0.05% Tween (PBS-T), incubated with biotinylated goat anti-Rabbit secondary antibody (Vector Labs) at RT for 30 min, and washed three times with PBS-T. After incubation for 30 min with the VECTASTAIN ABC Systems (Vector Labs), sections were washed 3 times in PBS-T, incubated with diaminobenzidine tetrahydrochloride (1 tablet dissolved in 10 ml 50 mM Tris pH 7.6, MP Biomedicals), washed three times in water and incubated with hematoxylin (Sigma) or iron hematoxylin (Sigma) for 2 min. Next, sections were washed in tap water for 5', dehydrated and mounted using Depex. Tissue sections were examined using a light microscope equipped with a digital camera (DC200, Leica Microsystems).