Mouse klk1b26, a prorenin-converting enzyme, is a member of the klk gene family  and is much more abundantly expressed in the SMGs of male mice than in females [6, 7]. This sex difference in klk1b26 protein expression has been thought to be due to stimulated transcription of the klk1b26 gene in the SMG by androgen [9, 10]. However, we found that when the klk1b26 protein and its mRNA levels in the male and female mouse SMGs were compared based on the wet weight of the glands, the mRNA levels in female SMGs were as much as 20% of that in male SMGs, while klk1b26 protein levels in the female glands were less than 6% of that in the male glands. Thus, we conceived the idea that there was an as yet unidentified mechanism, most probably post-transcriptional, down-regulating klk1b26 expression in the female SMG.
We reexamined klk1b26 mRNA levels in male and female SMGs by RT-PCR using various primer pairs targeting various regions of the mRNA, and found that the klk1b26 mRNA level in female total RNA preparations was estimated to be extremely low compared to that of male total RNA preparations only when the RT-PCR was performed with forward primers targeted near the 5'-terminal region of the mRNA (F15 and F21 primers, where the 5'-nucleotides corresponded to the 15th and 21st nucleotides of klk1b26 mRNA (GenBank: NM 010644), respectively) than when other forward primers targeting other regions of the mRNA were used (Figure 2). The inefficiency of the PCR product formation by F15 or F21 forward primers with the total RNA from female SMGs was not due to the modification of klk1b26 mRNA in female SMGs, because klk1b26 mRNAs from both male and female SMGs had the same nucleotide sequence according to the results of 5' RACE analysis (Figure 3A). The PCR products from male and female total RNA preparations with various primer pairs were subjected to DNA sequencing analysis and confirmed to be identical with the corresponding position of the klk1b26 mRNA sequence. Also, transcript variants for klk1b26 have not been reported/registered in the GenBank database. These results, therefore, suggest that female SMGs contained some inhibitory RNA molecules that interfered with the PCR process in a sequence-specific manner. The results of experiments with the 'exo-mRNA fraction' (Figure 4A, B) also supported this assumption.
Next, we examined the effects of miRNA from mouse SMGs on klk1b26 translation using the in vitro reticulocyte lysate system, because miRNAs arguably plays regulatory roles in gene expression and because such miRNA could be expected to interfere with the PCR reaction in a sequence-specific manner. The miRNA preparations from mouse SMGs prepared by using a PureLink miRNA Isolation Kit (Invitrogen Life Technologies) might contain small RNAs of < 200 nucleotides long other than miRNAs. The female miRNA preparation, but not the male one, inhibited the synthesis of klk1b26 protein in the in vitro translation system (Figure 4C, D). GAPDH translation in the same in vitro system was scarcely decreased by either the miRNA preparation from male or female, but the effects were not at significant levels. Differences between the effects of male and female miRNA preparations on GAPDH translation were not statistically significant. Since the klk1b26 translation interfering RNA was specifically observed in female mouse SMGs, the effects of castration and DHT administration to females or castrated mice were tested. The interfering RNA appeared in the SMG miRNA preparations from castrated mice and disappeared in mice who had undergone DHT administration (Figure 6A, B), indicating that expression of the interfering RNA was down-regulated by androgen.
The miRNA preparation from female SMGs also inhibited product formation in RT-PCR when performed with the primer pair (F21/R552) targeting the 5'-terminal region of klk1b26 mRNA, whereas the male miRNA preparation gave only a slight effect (Figure 5A). It should be noted that the inhibitory effect of the female miRNA preparation on the PCR product formation was clear when the F21 forward primer was used, but was not obvious with the F1, F11, F40, F100 (Figure 5C) or F169 forward primers (Figure 5B). These results implied that the length of RNA contained in the female miRNA preparation, which interacted with klk1b26 mRNA and interfered with klk1b26 translation, was not greater than 31 nucleotides long and that the site on the klk1b26 mRNA where the interfering RNA interacted was within 44 nucleotides from the start site of the mRNA (GenBank: NM 010644). We prepared small RNAs of 20 to 23 nucleotides long (chemically synthesized hsa-miR-325, miR-1497a, and asRNA in Figure 7A) which had partial complementarity with klk1b26 mRNA at its 5'-terminal region, and demonstrated that these small RNAs inhibited the in vitro translation of klk1b26 mRNA (Figure 7B). All of these synthetic small RNAs also inhibited the PCR reaction when the primer pair of F21/R552 was used. However, the inhibitory effect of the RNAs on PCR product formation was not detectable when the F40 or F100 forward primers were used in combination with the R552 reverse primer (Figure 7C). Considering the results shown in Figures 5 and 7 together with that shown in Figure 2A where the inhibition of PCR product formation was also observed with the F15 forward primer, it seemed plausible that the interfering RNA in the female miRNA preparation had partial complementarity with klk1b26 mRNA in its 15th to 44th nucleotide position, and its length was not greater than 25 nucleotides long. Therefore, we synthesized 30-nucleotide-long single-strand DNAs, [15th-44th]ssDNA and [169th-198th]ssDNA, of those sequences corresponding with the klk1b26 mRNA sequence at the 15th to 44th and 169th to 198th positions. Preincubation for hybridization of the female miRNA preparation with [15th-44th]ssDNA, but not with [169th-198th]ssDNA used as a negative control, masked/neutralized the activity of the female miRNA preparation in interfering with klk1b26 translation (Figure 8A, B), supporting our assumption.
Translational repression by miRNAs is thought to occur via several mechanisms: RNA-induced silencing complex (RISC)-mediated enhancement of deadenylation and decapping to destabilize the mRNA; sequestering of the mRNA into cytoplasmic bodies (GW-bodies or P-bodies); or disrupting the binding of translation factors to inhibit translation initiation or elongation [18, 21]. When miRNA affects mRNA such as small interfering RNA (siRNA), miRNAs associated with argonaute proteins in the RISC repress protein expression by binding to regions of complementarity with the target mRNAs. The binding sites of miRNAs were typically located within the 3'-UTR of target mRNAs [15, 18–21]. In this case, the seed sequence (second to eighth sequence from the 5'-end) of miRNA was thought to be very important and should be complementally matched with the target mRNA sequence [13, 14]. However, miRNAs were also reported to repress translation by interacting with coding regions of target mRNAs : mouse Nanog, Oct4 and Sox2 genes were demonstrated to have many naturally occurring miRNA targets in their amino acid coding sequence (CDS) and some of those targets were revealed to not contain the miRNA seed. Tay et al.  also reported that the miRNAs interacting at the CDS regions of their target mRNAs had only a limited effect on the mRNA levels but affected the corresponding proteins more substantially. We hypothesized that a kind of miRNA that was specifically expressed in female mouse SMGs interfered klk1b26 translation process contributing to sexual dimorphism of the klk1b26 protein in the mouse SMGs. It is plausible that such a miRNA, which has partial complementarity with the 5'-terminal region of klk1b26 mRNA, inhibits the klk1b26 protein synthesis by disturbing the binding of translation factors needed for translation initiation or elongation, though the effects of other kinds of small non-coding RNA(s) cannot be completely ruled out. In our preliminary experiments on miRNA profiling for male and female mouse SMGs by microarray assay, mmu-miR-325*, which is the mouse counterpart of primate miR-325, was not expressed in high enough levels to estimate whether it was one of the miRNAs expressed in female SMGs rather than in male SMGs. Though the sequence of mmu-miR-325* [29, 30] has two alternative nucleotides compared with that of Homo sapien s (hsa) miR-325, whose sequence is conserved in Macaca mulatta (mml) and Pongo pygmaeus (ppy) [27, 28], mmu-miR-325* is one of the plausible candidates. Additionally, there might be considerable numbers of miRs that have not yet been found and/or registered in miR database. Further analysis should be carried out for the identification of the miRNA(s) involved in the regulation of klk1b26 expression.
SMGs of male mice but not those of female mice produce various proteins with important biological activities [1–3]. Their expression is androgen dependent [1, 6, 7, 9, 10], and some of them are also responsive to thyroid hormone [6–10, 32, 33]. As implied in the present study, this sex difference observed in mouse SMGs is attributed, at least in part, to the miRNA(s) specifically expressed in the female SMG. More than 900 mature miRNAs are on file for human miRNAs in the miRNA database (miRBase release 16, September 2010; http://www.mirbase.org/) and over 60% of human protein-coding genes are thought to be targeted by miRNAs . Since miRNAs play critical roles for the normal functioning of cells, dysregulation of miRNA is thought to be associated with many diseases. Many miRNAs are reported to have links with cancer , heart disease , and dysfunctions in the nervous system . Among them are those for sex-specific diseases such as breast cancer [38, 39], ovarian cancer [40, 41], and prostate cancer . The detailed mechanisms of the miRNA effects on those diseases largely remain to be elucidated. Furthermore, sexual dimorphism is observed not only in the diseases of reproduction-related organs but also in the prevalence and severity of many common diseases such as cardiovascular diseases, autoimmune diseases, asthma, Alzheimer's disease, and Parkinson's disease . The sexual dimorphism observed in disease traits is thought to result from differential gene regulation in males and females, particularly with respect to sex hormone-responsive genes  including those that encode miRNAs.
Among the organs common in males and females, the mouse SMG is one that typically displays sexual dimorphism. It should, therefore, be noted that the mouse SMG is an important model system to investigate further the role and action mechanisms of sex-specific small non-coding RNAs including miRNAs and to uncover the regulatory mechanisms of such sex-specific RNA expression involving androgen and/or estrogen, and occasionally the combination of sex hormones with other hormones.