A variant of Runx2 that differs from the bone isoform in its splicing is expressed in spermatogenic cells

Northern blot analysis

Testes were isolated from 1-, 2-, 3-, 4-, and 10-week-old male mice of the C57BL/6J strain and from 10-week-old male mice of the W∕W^v strain. Spermatocyte and spermatid fractions were prepared from the cell suspension of C57BL/6J testes (Mays-Hoopes et al., 1995). Total cytoplasmic RNA was prepared from testes using Isogen (Nippon Gene, Toyama, Japan). Poly(A)⁺ RNAs were selected using Oligo(dT)-Latex (Takara, Otsu, Japan) and 2 µg of sample was electrophoresed through a 1% (w/v) agarose gel containing 2.2 M formaldehyde. RNA was transferred from the gel to a membrane, and the membrane was hybridized with a ³²P-labeled probe as described previously (Chiba et al., 1997). The probes were prepared either from a HindIII-NotI fragment of murine Runx2 cDNA (corresponding to nt 282 through to nt 473 in NM_001146038.2) or murine Runx1 cDNA sequence, murine PEBP2β∕CBFβ cDNA sequence and β-actin (Satake et al., 1995).

cDNA cloning and sequencing

A cDNA library prepared from murine testicular poly(A)⁺ RNA was provided by Y Nishina (Osaka University, Osaka, Japan). A ³²P-labeled HindIII-NotI fragment of murine Runx2 cDNA was used as the probe. The library was screened under a stringent condition according to the standard method. The cDNA inserts from each of the five isolated clones were excised from the pAP3neo vector and subcloned into the pBluescript II vector. The entire length of the insert was sequenced using the dideoxy-dye terminator method.

5′-Rapid Amplification of cDNA Ends (5′-RACE)

To determine the full length of the transcripts, 5′-RACE was performed following the manual supplied by the manufacturer (Life Technologies). A gene-specific primer (5′-TGTAAATACTGCTTGCAGCC-3′) was annealed to poly(A)⁺ RNAs and cDNA was synthesized. After degrading RNA with RNase H, purified cDNA was tailed with dCTP and TdT. The dC-tailed cDNA was amplified with the anchor primer and a nested gene-specific primer (5′-GTGACCTGCAGAGATTAACC-3′). The double-stranded cDNA was subcloned into the pBluescript II vector and sequenced.

Bioinformatic analysis

The subcellular localizations of proteins were predicted using the PSORT program (Nakai & Horton, 1999). The secondary structures of proteins were predicted using the PSIPRED (McGuffin, Bryson & Jones, 2000), SAM-T99-2d (Karplus et al., 1999), and Profsec (Rost & Eyrich, 2001) programs. These programs were downloaded from the BioInfo MetaServer (http://bioinfo.pl/). The PSI-BLAST program (Altschul et al., 1997) was used for homology searches. Three iterations were used before full saturation was reached. The domain architecture of proteins was analyzed using the SMART tool (Letunic et al., 2002).

Immunoblot analysis

Proteins were extracted from testes using RIPA buffer and 10 µg of the sample was subjected to 8% or 10% (w/v) SDS-polyacrylamide gel electrophoresis (PAGE). The ORF of the testicular Runx2, which encodes a polypeptide of 106 aa, was fused in frame to glutathione-S-transferase (GST) using the pGEX vector. The plasmid was transfected into E. coli and transformed bacteria were lysed in sample buffer after induction with isopropyl-β-D-thiogalactoside. Two micrograms of protein was separated by SDS-PAGE. Proteins were transferred from the gel to a membrane and the blotted membranes were blocked with TBS-T buffer, which contained 20 mM Tris–HCl pH 7.4, 150 mM NaCl, and 0.1% (v/v) Tween 20. The primary antibody was anti-Runx2 serum, which was raised in rabbit using the amino-terminal 15 aa residues (MLHSPHKQPQNHKCG) of the murine testicular Runx2 as an antigen. In some cases, the antiserum was preabsorbed with an excess amount of antigen peptide. The secondary antibody was alkaline phosphatase-conjugated goat anti-rabbit IgG (Promega, Madison, WI, USA). The antibodies were diluted appropriately in TBS-T. Immunologically reacted products were detected using the BCIP/NBT Color Development Substrate Kit (Promega, Madison, WI, USA).

Immunohistochemistry of testicular preparations

To prepare frozen sections, testes from adult male C57BL/6J mice were cut into three pieces and fixed in Zamboni solution for 6 h at 4 °C with agitation. The tissues were immersed sequentially in 10% (w/v), 15%, and 20% sucrose solutions prepared in phosphate-buffered saline (PBS) for 2 h each at 4 °C; embedded in OCT compound (Miles Laboratories, Berkeley, CA, USA); and kept frozen at −80°C until use. The tissues were cryostat-sectioned (7-µm thick sections), air-dried, and post-fixed for 20 min in 4% (w/v) paraformaldehyde prepared in PBS.

The sections were then treated with methanol containing 0.3% (v/v) hydrogen peroxide for 15 min followed by PBS containing 3% (w/v) skimmed milk and 10% (v/v) goat serum for 30 min. In immunohistochemistry, the anti-serum was affinity-purified using a peptide (MLHSPHKQPQNHKCG)-linked Sepharose 4B column. The sections were incubated with appropriately diluted, affinity-purified anti-Runx2 antibody at 4 °C overnight, followed by biotinylated goat anti-rabbit IgG for 30 min. The primary antibody was detected using an ABC Kit (Vector Laboratories, Burlingame, CA, USA). The sections were post-stained with methyl green, and coverslips were mounted on glass slides.

The testicular Runx2 is transcribed specifically in germ cells

To gain insight into the significance of the Runx2 transcript in testes, we first used W∕W^v mice and wild-type C57BL/6 mice. W∕W^v mice lack germ cells except spermatogonia because of mutations in the c-Kit gene (Kubota et al., 2009). RNA was extracted from testes and processed for Northern blot analysis (Fig. 1, Fig. S1). The Runx2 transcript was detected as a broad band of ∼1.8 kb length in wild-type testis, but not in W∕W^v testis (Fig. 1A, lanes 1 and 2). By contrast, Runx1 transcript (Fig. S1), PEBP2β∕CBFβ transcript (Fig. 1B, lanes 1 and 2) and β-actin transcript (Fig. S1) used as loading control, were detected in both RNA samples tested. This indicates that the testes-specific Runx2 transcript is expressed in germ cells, not in somatic cells in testes.

Next, spermatocytes at the pachytene stage and spermatids were purified from a cell suspension prepared from wild-type testis, and then RNA analyzed. The Runx2 transcript was detected in both the spermatocyte and spermatid fractions (Fig. 1A, lanes 3 and 4). The differentiation of germ cells proceeds in a synchronous fashion immediately after birth; therefore, RNA was prepared from the testes of newborn mice and analyzed (Fig. 1A, lanes 5–8). A band of 1.8 kb was detected in 4 week-old testis, which largely contains germ cells at the spermatid stage. Thus, the testicular Runx2 was transcribed in germ cells at the spermatocyte and spermatid stages.

Determination of the structure of the testicular Runx2 transcript

Although the study by Ogawa et al., (2000) predicted the ORF of Runx2 to encode a 106-aa protein, the cDNA and aa sequences have not been published or registered in a public database. Only the junction sequence between exons 1 and 3 is available in the literature. Therefore, we decided to independently clone Runx2 cDNAs from a library prepared from murine testis. The nucleotide sequence that was determined from the obtained cDNA clones and 5′RACE is presented in Fig. 2A and has been deposited at NCBI (accession number, DQ458792). The AUG codons of Met¹ and Met⁶⁹ as well as the termination codon are indicated in bold.

We next compared this cDNA sequence with the murine genomic sequence of Runx2 (AB013129) (Xiao et al., 1998). The so-called exon 1 could actually be split into two small exons that are separated by an intronic sequence of 197 nucleotides. We tentatively designated these smaller exons as 1U (U for upstream) and 1D (D for downstream). The predicted coding region of each exon is shown in Fig. 2B. Exon 1U harbors the Met¹ codon, whereas exon 1D includes the Met⁶⁹ codon. Exons 1U and 1D are also transcribed in bone Runx2 (Xiao et al., 1998) and the ORF of bone Runx2 has been predicted ( NM_009820) (Ducy et al., 1997). The amino-terminal 87 aa residues are common to both the testicular and bone ORFs, whereas the carboxy-terminal 19 aa residues derived from exon 3 are unique to the testicular ORF.

Bioinformatic analysis suggests that the testicular Runx2 variant is authentic

The predicted aa sequence of the testicular Runx2 variant is shown in Fig. 3A. We performed bioinformatic analysis as a first step to examine the theoretical likelihood that this is an authentic protein. According to the PSORTII program, the Runx2 variant had a 43% probability of being a nuclear protein. Moreover, the variant appears to be a basic protein; out of 106 aa residues, 22 are basic and only six are acidic. Secondary structure assignment programs suggested that Runx2 has both α-helical and β-sheet structures (indicated by H and E, respectively; Fig. 3A). In particular, a stretch of 55 aa in the amino-terminus are folded into two distinct α-helices. A basic, nuclear protein with a high α-helical content is reminiscent of a histone. On the other hand, three carboxy-terminal Cys residues may adopt a globular structure with a sulfur bridge (indicated by green, Fig. 3A).

The amino-terminal part of the Runx2 variant (from Met¹ to Ser⁴⁶) showed weak but significant similarity to a 48-aa sequence of a murine protein of unknown function that is expressed in the eye, namely, Q7TPL8/33942100 (SWISS and NCBI IDs, respectively; Fig. 3B). Q7TPL8 possesses features that suggest it is a transcription factor (Fig. 3C). At the amino-terminus, it contains a KRAB (Kruppel-associated box) domain that may function as a transcription-repression domain, and at the carboxy-terminus, it harbors eight zinc fingers that may function as nucleic acid-binding structures. The region of Q7TPL8 that shows similarity to Runx2 is located immediately before the stretch of zinc fingers. This feature suggests that the amino-terminal motif of 46 aa residues in Runx2 represents a functional domain. In addition, the testicular Runx2 also retains a stretch of 19 aa residues (from Met⁶⁹ to Trp⁸⁷) that corresponds to an important transcription activation domain in bone Runx2 (Thirunavukkarasu et al., 1998).

Taken together, these observations increase the probability that the testicular Runx2 is an authentic protein.

Met¹ is used as an initiation codon in the testicular Runx2 variant

We next examined whether Met¹ is indeed used as an initiation codon in testis. We generated an antiserum using the amino-terminal 15 aa of Runx2 (Met¹ to Gly¹⁵) as an antigen. A protein extract was prepared from mouse testis and processed for immunoblot analysis. A clear band of approximately 11 kDa was detected (Fig. 4A, lane 1, indicated by an arrow), whereas inclusion of an excess amount of antigen peptide in the immunoreaction abolished this band (Fig. 4A, lane 2). Lysates from bacteria expressing a GST-Runx2 ORF fusion protein were run as controls in lanes 3 and 4, and GST-Runx2 was detected with the same antiserum. It is therefore highly likely that this ORF is expressed in testis as a polypeptide of 106 aa residues.

The testicular Runx2 is a differentiation stage-dependent nuclear protein

To verify that the truncated variant of Runx2 is expressed, we performed immunohistochemistry on frozen sections of testes (Fig. 4B). Panel a shows the staining pattern in mouse testis that was probed with the affinity-purified anti-Runx2 antibody raised against the Met¹-to-Gly¹⁵ peptide. Positively stained cells were detected within seminiferous tubules. This immunostaining was specific for Runx2 because it completely disappeared when the antibody was preabsorbed with an excess amount of antigen peptide (panel b).

Interestingly, the distribution pattern of positive cells appeared to differ from tubule to tubule. Therefore, each seminiferous tubule containing positive cells were classified according to the differentiation stage. This was judged by the morphology of cells and nuclei (see panel c–i). Runx2 staining was detected in the following differentiation stages: late pachytene spermatocytes at stages VIII and X (panels c and d), diplotene spermatocytes at stage XI (panel e), cells of second meiotic phase at stage XII (panel f), and round spermatids of spermiogenesis phase 1, 2/3, and 5 at stage I, II/III, and V, respectively (panels g, h, and i). In short, the Runx2 variant was detected in various stages of differentiation, from late pachytene spermatocytes to round spermatids. It must be noted that detection of Runx2 protein by immunohistochemistry (Fig. 4B) and detection of the Runx2 transcript by Northern blot (Fig. 1) coincides in terms of germ cell specific expression and expression in pachytene spermatocytes to spermatids.

Finally, as seen in the enlarged view in Fig. 4C, positive staining was always restricted to the nuclei, indicating that the testicular Runx2 is a nuclear protein. The protein was detected as multiple foci in the nuclei of germ cells. The original picture of immunohistochemistry observed under differential interference contrast microscope can be seen as Figs. S2 and S3.