Distinct Spatiotemporal Expression of Serine Proteases Prss23 and Prss35 in Periimplantation Mouse Uterus and Dispensable Function of Prss35 in Fertility

PRSS23 and PRSS35 are homologous proteases originally identified in mouse ovaries. In the periimplantation mouse uterus, Prss23 was highly expressed in the preimplantation gestation day 3.5 (D3.5) uterine luminal epithelium (LE). It disappeared from the postimplantation LE and reappeared in the stromal compartment next to the myometrium on D6.5. It was undetectable in the embryo from D4.5 to D6.5 but highly expressed in the embryo on D7.5. Prss35 became detectable in the uterine stromal compartment surrounding the embryo on D4.5 and shifted towards the mesometrial side of the stromal compartment next to the embryo from D5.5 to D7.5. In the ovariectomized uterus, Prss23 was moderately and Prss35 was dramatically downregulated by progesterone and 17β-estradiol. Based on the expression of Prss35 in granulosa cells and corpus luteum of the ovary and the early pregnant uterus, we hypothesized that PRSS35 might play a role in female reproduction, especially in oocyte development, ovulation, implantation, and decidualization. This hypothesis was tested in Prss35(−/−) mice, which proved otherwise. Between wild type (WT) and Prss35(−/−) mice, superovulation of immature females produced comparable numbers of cumulus-oocyte complexes; there were comparable numbers of implantation sites detected on D4.5 and D7.5; there were no obvious differences in the expression of implantation and decidualization marker genes in D4.5 or D7.5 uteri. Comparable mRNA expression levels of a few known protease-related genes in the WT and Prss35(−/−) D4.5 uteri indicated no compensatory upregulation. Comparable litter sizes from WT × WT and Prss35 (−/−)× Prss35 (−/−) crosses suggested that Prss35 gene was unessential for fertility and embryo development. Prss35 gene has been linked to cleft lip/palate in humans. However, no obvious such defects were observed in Prss35(−/−) mice. This study demonstrates the distinct expression of Prss23 and Prss35 in the periimplantation uterus and the dispensable role of Prss35 in fertility and embryo development.


Introduction
Proteases (.600) are categorized into five main groups: metallo, serine, cysteine, aspartic/glutamate, and threonine based on the nature of their active-site catalytic residue. Serine proteases (PRSS) are grouped into 13 classes and 40 families characterized by the presence of serine (Ser) as the nucleophilic amino acid at the enzyme's active site [1][2][3]. PRSS23 and PRSS35 belong to the trypsin class of serine proteases [4].
PRSS35 was originally identified as a novel mouse ovaryselective gene using suppression subtractive hybridization and PRSS23 was subsequently identified as a homologous protease of PRSS35 via BLAST search [4]. Both genes have two exons with the coding region in exon 2. Although their proteolytic activities have not yet been characterized, both possess general features of serine proteases. However, the canonical Ser is replaced by a threonine (Thr) in PRSS35 [4].
Despite their high sequence homology, Prss35 and Prss23 have their unique expression and regulation patterns in the mouse ovary [4,5]. Prss35 mRNA is localized in the theca layers of developing follicles, granulosa cells of preovulatory and ovulatory follicles, and the forming and regressing corpus luteum. Prss23 mRNA is highly detected in the granulosa cells of the secondary/early antral follicles, it is also expressed in the ovarian stroma and theca tissues just before ovulation. Prss35 mRNA is upregulated around the time of ovulation and remains elevated in the developing corpus luteum; while Prss23 mRNA is transiently downregulated after ovulation induction and again in the postovulatory period. Prss35 expression is progesterone-dependent prior to follicle rupture and upregulated by gonadotropins; while Prss23 expression is independent of progesterone and downregulated by gonadotropins.
Dot blots of adult mouse tissues indicate that Prss35 is detectable in the ovary only, and Prss23 is detectable in a wide range of tissues, including a high level in the uterus [4]. A recent study shows that Prss23 is localized in heifer's uterine luminal epithelium (LE) and is upregulated in the heifer uterus from gestation day 7 (D7) to D13 [6], suggesting its uterine regulation during early pregnancy in heifers.
We have been studying the molecular mechanisms for the establishment of uterine receptivity using a mouse model deficient of the third lysophosphatidic acid (LPA) receptor (Lpar3 (2/2) ).
Lpar3 is mainly detected in the preimplantation D3.5 LE in wild type (WT) mice and Lpar3 (2/2) females have delayed uterine receptivity for embryo implantation [7]. Microarray analysis indicated that Prss23 was the most differentially expressed Prss gene between D4.5 WT and Lpar3 (2/2) LE cells. We analyzed the expression of both Prss23 and its homologous protease gene Prss35 in the periimplantation mouse uterus and found their distinct spatiotemporal expression patterns in the LE and the stromal compartment, respectively. Proteases play an important role in proteolysis that is essential for tissue remodeling and functions of the ovary and the uterus, which go through extensive tissue remodeling during estrous cycle and pregnancy [8][9][10]. The spatiotemporal expression of Prss23 and Prss35 in the ovary and the uterus led us to hypothesize that PRSS23 and PRSS35 may be involved in ovarian and uterine functions [4,5]. In this study, we tested the hypothesis on PRSS35 in Prss35 (2/2) mice.

Animals and Genotyping
WT and Lpar3 (2/2) mice (C57BL6/129svj mixed background) were from a colony at the University of Georgia, which was originally derived from a colony at The Scripps Research Institute [7]. They were genotyped as previously described [7]. Prss35 (2/2) mice were derived from the mouse strain B6/129S5-Prss35 tm1Lex / Mmucd (identification number 032535-UCD) purchased from the Mutant Mouse Regional Resource Center (MMRRC) at UC Davis, a NCRR-NIH funded strain repository. Prss35 (2/2) mice were genotyped using tail genomic DNA and four primers: PRSS35 DNA419.18, PRSS35 DNA419.33, PRSS35 DNA419.34 and PRSS35 GT-IRES (Table 1) in PCR reactions. The PCR cycles were set as follows: 10 cycles of 94uC for 15 s, 65uC for 30 s (decreasing 1uC/cycle), and 72uC for 40 s; and 30 cycles of 94uC for 15 s, 55uC for 30 s and 72uC for 40 s. The expected PCR product sizes for WT (PRSS35 DNA419.18 and PRSS35 DNA419.33) and targeted alleles (PRSS35 DNA419.34 and PRSS35 GT-IRES) were 355 bp and 574 bp, respectively. All mice were housed in polypropylene cages with free access to regular food and water from water sip tubes in a reverse osmosis system. The animal facility is on a 12-hour light/dark cycle (7:00 AM to 7:00 PM) at 2361uC with 30-50% relative humidity. All methods used in this study were approved by the University of Georgia Institutional Animal Care and Use Committee (IACUC) and conform to National Institutes of Health guidelines and public law.

Hormonal Treatment
Hormonal treatment on ovariectomized WT mice was done as previously described [13,14]. Briefly, in the vehicle-treated group and the progesterone (P4)-treated group, the ovariectomized virgin C57BL6 mice (recovered for 2 weeks after surgery) were injected with 0.1 ml vehicle (oil) or P4 (2 mg in 0.1 ml oil) three times on 0 h, 24 h, and 48 h, respectively. In the 17b-estradiol (E2)-treated group, the ovariectomized mice were injected with 0.1 ml oil on 0 h and 24 h, then 100 ng E2 (in 0.1 ml oil) on 48 h. In P4+ E2-treated group, the mice were treated the same as the P4-treated group except an additional injection of 100 ng E2 on 48 h. All the mice were dissected 6 hours after the last injection. The total treatment time for P4 was 54 hours and that for E2 was 6 hours. The uteri were dissected and snap-frozen for realtime PCR.
Vaginal Opening, Superovulation, Embryo Implantation, Gestation Period, and Litter Size The vaginas of WT, Prss35 (+/2) , and Prss35 (2/2) females were checked daily from postnatal day 22 until vaginal opening was detected, which was recorded as the age of vaginal opening. Superovulation was done on immature 21 days old WT and Prss35 (2/2) females. They were injected intraperitoneally with 5 IU eCG (equine chorionic gonadotropin, Sigma-Aldrich) and 48 hours later with 5 IU hCG (human chorionic gonadotropin, Sigma-Aldrich). The cumulus-oocyte complexes were collected from the oviduct 16 hours after hCG injection and counted. Embryo implantation was detected on D4.5 and D7.5 using blue dye reaction as previously described [7]. Gestation period and litter size were recorded as previously described [7].

Statistical Analysis
Statistical analyses were done using two-tail, unequal variance Student's t test. The significant level was set at p,0.05.

Expression of Prss23 and Prss35 in Periimplantation Mouse Uterus
In situ hybridization showed expression of Prss23 in the uterine luminal epithelium (LE) but no detectable expression in other uterine compartments of the D3.5 WT uterus (Fig. 1A). The expression level of Prss23 was greatly decreased from LE in the D4.5 WT uterus and it was undetectable in other uterine compartments (Fig. 1C). It was detected in the LE of the D3.5 and the D4.5 Lpar3 (2/2) uteri (Figs. 1B, 1D), which had delayed uterine receptivity for embryo implantation [7]. These data demonstrated that Prss23 was downregulated in the LE upon embryo implantation, which normally initiates around D4.0 in WT mice [11,15]. No significant levels of Prss23 were detectable in the D5.5 WT uterus (Fig. 1E). It reappeared at a relatively low level in the stromal compartment next to the myometrium in the D6.5 WT uterus (Fig. 1F). It remained there in the D7.5 WT uterus and interestingly, it was highly expressed in the D7.5    (Figs. 1G,1I). Positive control indicated Prss23 expression in the granulosa cells of the D4.5 WT ovary (Fig. 1J) [4,5].
In the original study that identified PRSS35 [4], Prss35 was only detectable in the mouse ovary but not other adult mouse tissues, including the uterus examined using dot blot. Our study indicated that Prss35 expression was only detected in the stromal compartment upon implantation during pregnancy, which may explain why dot blot couldn't detect Prss35 in the non-pregnant adult mouse uterus [4]. RT-PCR data from MMRRC website (http:// mmrrc.mousebiology.org/phenotype/Genentech/PRT215N1/ Expression/WT_Panel/Level_I/popups/PRT215N1-Expression-WT_Panel-imageViewer-3349.html) indicate expression of Prss35 in most tissues (no uterus) examined, including several tissues that had undetectable levels of Prss35 by dot blot [4].
Despite their high sequence homology [4], Prss23 and Prss35 had no obvious overlapping spatiotemporal expression in the periimplantation mouse uterus (Figs. 1, 2), suggesting that they have potentially different roles in uterine remodeling during early pregnancy. Since Prss23 is expressed in the LE before embryo implantation, it is possible that it could be involved in the LE preparation for the initial implantation processes, such as embryo attachment. On the other hand, Prss35 is only detected in the stromal compartment (Fig. 2). Interestingly, the main localization of Prss35 is on the mesometrial side of the embryo (Figs. 2E, 2G).

Hormonal Regulation of Prss23 and Prss35 in Ovariectomized Mouse Uterus
In the ovariectomized WT uterus, Prss23 was moderately but significantly downregulated (,1-fold) upon P4 or E2 treatments. There was no significant difference in Prss23 expression levels between vehicle and P4+ E2-treat groups (Fig. 3). Prss35 was dramatically downregulated (.5-fold) upon P4 or E2 treatments. It was also significantly downregulated by P4+ E2 treatment (Fig. 3). The house-keeping gene HPRT1 was not regulated by these hormonal treatments (Fig. 3).
Since the magnitude of uterine Prss23 upregulation during early pregnancy is decreased in heifers with low P4 [6], it suggests that P4 might upregulate Prss23 in the early pregnant heifer uterus. Prss23 mRNA can be induced by E2-activated ERa in MCF-7 breast cancer cells [16]. However, Prss23 is moderately downregulated by P4 and E2 in the ovariectomized mouse uterus under the treatment regimen (Fig. 3). These observations suggest that Prss23 could potentially be differentially regulated by P4 in the uterus in different species and/or under different experimental settings, such as natural pregnancy [6] and ovariectomy (Fig. 3), as well as that Prss23 could be differentially regulated by E2 under different experimental settings, such as breast cancer cells [16] and the ovariectomized mouse uterus (Fig. 3).
Prss35 is regulated by P4 in the mouse ovary [4]. Treatment of eCG-primed mice with steroid synthesis inhibitor trilostane (TRL) reduces ovarian P4 and Prss35 levels and the synthetic progestin R5020 could reverse the inhibitory effect of TRL on ovarian Prss35 expression [4], indirectly indicating that P4 could upregulate Prss35 in the mouse ovary. Prss35 is downregulated by P4 in the ovariectomized mouse uterus (Fig. 3). These observations suggest tissue-specific regulation of Prss35 by P4 in mice.

Non-essential Role of PRSS35 in Embryo Development
When Prss35 (2/2) males were mated with Prss35 (+/2) or Prss35 (2/2) females, the average litter sizes were comparable to that from WT 6 WT crosses (Fig. 7), indicating that PRSS35 did not play a critical role in embryo development.
No Compensatory Upregulation of the mRNA Levels of a Few Other Protease-related Genes in the D4.5 Prss35 (2/ 2) Uterus Compensatory upregulation of other proteases has been reported in the uterus upon deletion of one protease, such as in the case of upregulation of matrix metalloproteinase-3 (MMP-3) and MMP-10 in the MMP-7-deficient uterus [26]. To determine if deletion of Prss35 could lead to compensatory upregulation of other proteases in the uterus, several known uterine expressed protease-related genes were examined in D4.5 WT and Prss35 (2/ 2) uteri. These genes included Prss8 [27], Prss12 [28], Prss23 [4], ISP1 [29], ISP2 [30], Spink3 [31], SerpinA3N [32], and SerpinG1 [33]. Since Prss35 was upregulated in the D4.5 WT uterus upon implantation (Fig. 2C), the expression levels of these genes were examined in the D4.5 WT and Prss35 (2/2) uteri. Comparable mRNA expression levels of these protease-related genes between D4.5 WT and Prss35 (2/2) uteri indicated no compensatory upregulation of these genes in the Prss35 (2/2) uterus (Fig. 8). However, it could not rule out the possible compensatory upregulation of other protease-related genes in the uterus or the possible compensatory upregulation of protease-related genes in other tissues.
In summary, this study demonstrates the distinct spatiotemporal expression patterns of Prss23 and Prss35 in the periimplantation mouse uterus as well as the non-essential role of Prss35 in ovarian function, uterine function, and embryo development using the Prss35 (2/2) mouse model.