Generation of conditional knockout alleles for PRL-3

Phosphatase of regenerating liver-3 (PRL-3) is a member of the protein tyrosine phosphatase (PTP) superfamily and is highly expressed in cancer metastases. For better understanding of the role of PRL-3 in tumor metastasis, we applied a rapid and efficient method for generating PRL-3 floxed mice and investigated its phenotypes. A BAC retrieval strategy was applied to construct the PRL-3 conditional gene-targeting vector. Exon 4 was selected for deletion to generate a nonfunctional prematurely terminated short peptide as it will cause a frame-shift mutation. Conditional knockout PRL-3 mice were generated by using the Cre-loxP system and were validated by Southern blot and RT-PCR analysis. Further analysis revealed the phenotype characteristics of PRL-3 knockout mice and wildtype mice. In this study, we successfully constructed the PRL-3 conditional knockout mice, which will be helpful to clarify the roles of PRL-3 and the mechanisms in tumor metastasis.


INTRODUCTION
The majority of cancer related deaths are induced by tumor metastasis and recent oncologic researches have focused on the mechanisms of metastasis. Protein tyrosine phosphatases (PTPs) play a fundamental role in regulating diverse cellular processes, including cell proliferation, adhesion and migration [1] . Phosphatase of regenerating liver-3 (PRL-3) is a member of the protein tyrosine phosphatase (PTP) superfamily. Mouse PRL-2 and PRL-3 exhibit 87% and 76% homology to mouse PRL-1 in their amino acid sequences, and three mouse PRL proteins contain a Cterminal consensus sequence for prenylation [2] . PRL-3 encodes a small, -20 kDa tyrosine phosphatase that is located at the cytoplasmic membrane when prenylated at its C-terminus; however, the unprenylated PRL-3 is shifted into the nucleus [3] . The mouse PRL-3 gene has 6 exons, encompassing 30.36 kb of genomic DNA.
In human diseases, PRL-3 is highly expressed in cancer metastases. Moreover, overexpression of PRL-3 also facilitates the cells to form metastasis-like tumors in mouse models. High expression of PRL-3 has been found to be associated with human colorectal tumors [4][5][6] . We have reported that PRL-3 promoted motility and metastasis of mouse melanoma cells [7] . Recently, it has also been reported that silencing of the PRL-3 gene by siRNA reduced cell migration in the multiple myeloma cell line INA-6 [8] . PRL-3 shows promise as a biomarker and prognostic indicator in colorectal, breast, lung and gastric cancers [9][10][11][12] . However, the substrates and molecular mechanisms of action of the PRLs have remained elusive. Recent findings indicate that PRL-3 may down-regulate PTEN expression and signal through PI3K to promote epithelial-mesenchymal transition [13] . In conclusion, PRL-3 is specifically expressed in malignant plasma cells and may have a role in regulating migration of these cells. The progression of other tumor types was positively correlated with PRL-3 expression [14] , suggesting that PRL-3 is a molecular marker for metastatic tumor cells and could be an important target for anticancer therapeutics [5,15,16] . However, the substrates and molecular mechanism by which PRL-3 promotes tumor metastasis are far from clear. To further reveal the physiological significance of PRL-3, we created the PRL-3 conditional knockout mice through the Cre-loxP system.

Construction of PRL-3 conditional gene-target vector
The PRL-3 conditional gene-targeting vector was made based on a BAC-retrieval strategy as described previously [17] . Briefly, the region of the BAC containing the loxP sites and positive and negative selection markers is then excised from the BAC and transformed into ES cells (BAC from Sanger, 129S7/ AB2 clone). After validation by DNA sequencing, the BAC was electroporated into EL350 [18] . Retrieval primers with Amp and TK using pL253 (pBluescriptbased vector) as template and in which the restriction sites for NotI and AvaI were removed. Restriction sites for these enzymes were included in the amplification primers to permit directional cloning of the PCR products into PL253. The 5' homolog (5'-CTACCTCCTACACCTGAGAACAGCTCAG-CACTCCACTCCAAGTCAGTGCCATGGCT-CACAGCTAGTCCCTGCCACCATTC-gcggcc-GCACGAATGGGTTACATCGAACTGGATC-3') and 3' homolog (5'-CTGCCCAGGTCTTGATGT-GGTGGTGACGGCTGTGTAGGCATTTGAT- were designed for retrieving. The next step was introduction of a loxP site into the subcloned DNA. PCR amplified 80-bp homologous arms loxP-NEO-loxP cassette from PL452. This is accomplished by introducing a floxed neomycin resistance (Neo) cassette (PL452) via homologous recombination into the subcloned plasmid DNA, and by removing the DNA fragment between two loxp sites via Cre recombinase induced by 0.1% arabinose. Primers designed for inserting the first loxP between exon 4 and exon 5 were 1Lox-F (5'-GCAGGGAGGGCTTCCTGCTCT-GATCCAGTCACTGCACCTCATGCAGAAAAA-GAGCTGCCGCAAAGACATCCTGGTCATC-CAGCAGCCCAATTCCGATCATATTCAATAACC -3') and 1Lox-R (5'-CCTCCTGCAATAGGGGT-GTCGGCAGGCGCTCTCCTGCCTGGCTCACT-TCCATCTAGTGTGAGCAGCTGTGACTCAAG- . Then, Prl-3 first loxP popped-out plasmid was transformed into EL250 with the final step the introduction of a second loxP site into the subcloned DNA. PCR amplified 80bp homologous arm FRT-NEO-FRT-loxP cassette from PL451 [17] . Primers 2Lox-F (5'-AGCCTCT-CATCGCCTGCCATGTGGGAGCCTGTCTCCT-GGCAGCATTGCATCCCTAGAGACCCTGT-GCAGAGGTCGTGGGTAG_CGACGGTATCGA-TAAGCTTGATATCGAATTCC-3') and 2Lox-R (5'-TCTGCTGAGCGAGAAGCCTGGCTTATC-CAGATTCTGCACTGGGAAGAGAGGCCTTGT-TCCAGGAAACTAACAGGCCAGGCTAGAGGT_ GCTCTAGAACTAGTGGATCCACCTAATAACT-TC-3') were designed. Second PRL-3 loxP plasmids were transformed into DH-5α to get the purified targeting vector.
Chimeric mice were bred with CAGGS-Flpe (The Jackson Laboratory) transgenic mice to delete the FRT flanked Neo cassette to produce the PRL-3 flox allele, which was then bred with EIIa-Cre line to delete the floxed exon 4 to generate the PRL-3⊿ exon4 mice. All these alleles were maintained in the c57BL/6J back-ground.

General characteristics of PRL-3 knockout mice
We measured the body weights of PRL-3 homozygous (PRL-3 -/-) mice and wildtype (PRL-3 +/+ ) mice continuously for 9 weeks and established the growth curve of body weights of PRL-3 knockout mice. Each group had 9 mice and their body weights were measured in g at 1, 2, 3, 4, 5, 6, 7, 8, and 9 weeks. For further phenotype analysis, blood chemistry indicators were measured of the plasma of PRL-3 -/mice and PRL-3 +/+ mice. Each group has 8 mice, which had been fasted for 16 h before sacrifice. Blood collected from mice were put into Eppendorf tubes for aqueous bath at 37°C for 30 min. Then, blood samples were centrifugated at 3,000 rpm for 10 min and the plasma was separated and collected. We detected the plasma levels of cholesterol, triglycereide, high density lipoproteins, low density lipoproteins (LDL) and glucose in PRL-3 -/and PRL-3 +/+ mice. Intraperitoneal glucose tolerance test (IPGTT) was then performed in six PRL-3 -/mice and six PRL-3 +/+ mice were selected at 12 weeks of age. Glucose tolerance measurement was conducted at the second morning after fasting 16 h overnight. Mice were intraperitoneally injected with D-α glucose (0.15 g/mL, dilute in MillQ) at a dose of 1.5 mg/kg. Finally, the values of blood glucose were measured at 0, 15, 30, 45 and 60 min by using Blood Glucose Test Strips.

Statistical analysis
Data were analyzed with Student's t-test and presented as mean±SD. P < 0.05 was considered to have significant difference.

RESULTS
To establish the conditional knockout alleles through Cre-mediated excision, we decided to target exon 4 of the PRL-3 gene by flanking it with two loxP sites. Deletion of exon 4 is predicted to generate a nonfunctional prematurely terminated short peptide as it will cause a frame-shift mutation. We applied a BAC retrieval strategy as described previously [17] to construct the PRL-3 conditional gene-targeting vector, which contained 5'-and 3'-homology arms, FRT flanked Neo cassette, floxed exon 4 and the HSV-TK negative selection marker (Fig. 1).
For gene targeting, 20 μg of NotI linearized targeting vector [19]   ES cells. Positive clones were obtained after G418 selection and screened by long-range PCR. The PCR primers are depicted in Fig. 1. Two clones had undergone correct homologous recombination on the 3' site and 5' site and that the two loxP sites were properly inserted. Targeted clones were further confirmed by Southern blot with 5'-and 3'-probes (Fig. 2). Introduction of an additional EcoRI restriction enzyme site was used to differentiate wildtype (22 kb) versus PRL-3-flox mutated allele (12.7 kb, 8.5 kb) ( Fig. 2A, 2B). After confirmation by Southern blots, positive clones were then used to produce chimera by c57BL/6J blastocyst injection. Twenty-one chimeras (10 females and 11 males) were obtained. After crossing with c57BL/6J mice, 4 out of 10 female chimeric mice and 5 out of 11 male chimeric mice were identified to be capable of germline transmission. The pups of chimeric mice were also genotyped by Southern blot to check whether they contained the loxP mutation. To test Cre-mediated recombination of the PRL-3floxed allele, we crossed PRL-3 floxed mice with EIIA-Cre transgenic mice (FVB/N-TgN(EIIa-Cre)C5379Lmgd, the Jackson Laboratory). The pups (PRL-3 +/mice) that had under-gone Cre-mediated recombination were identified by PCR (Fig. 3A). Then, PRL-3 +/mice were back-crossed to c57BL/6J background. PRL-3 -/mice were obtained by inter-crossing of heterozygous mice (Fig. 3B).
RT-PCR analysis of PRL-3 expression in the heart, muscle and brain tissues of wildtype mice and prl-3 -/mice were performed simultaneously. The PCR products from the tissues of PRL-3 -/mice were 294 bp while the PCR products from the wildtype mice were 425 bp (Fig. 3C). Northern blot analysis of prl-3 expression in the heart, muscle and brain tissues of wildtype mice and prl-3 +/mice were performed simultaneously (Fig. 4).
the EIIa-Cre transgenic mice which express Cre recombinase in the germline. The offspring PRL-3 +/mice were identified by PCR products. Moreover, to further determine the accuracy of PRL-3 recombination in mice, we generated homozygous PRL-3 null mice. To demonstrate correct PRL-3 knockout mice, we detected PRL-3 mRNA expression in the heart, muscle and brain tissues isolated from wildtype mice and PRL-3 -/mice simultaneously. Additionally, the sequencing results and Northern blot analysis demonstrated successful knockout in mRNA level (data not shown). However, Western blotting examination of protein expression in the tissues extracted from wildtype and PRL-3 -/mice separately could not be obtained. Only in the metastasis tissues or tumors ex-DISCUSSION PRL-3 has been shown to play an important role in promoting cell migration and tumor metastasis. It has been reported that PRL-3 exerts its functions by modulating the Rho family GTPase, activating Src, and regulating the PI3K-Akt pathway in a contextdependent manner [20] . In this study, we applied a rapid and efficient method for constructing conditional knockout alleles for PRL-3 that is based on E. coli recombineering rather than restriction enzymes and DNA ligases.
To test whether the neo gene and exon 4 in the PRL-3 floxed allele had been deleted, the heterozygous(PRL-3 floxed) mice with Cre recombinase were crossed with tracted from the wildtype mice, high expression level of PRL-3 protein could have been proved. Previous reports suggested that PRL-3 is the only gene consistently overexpressed in 100% of 18 colorectal cancer liver metastases, which indicates that PRL-3 plays an important role in tumor metastasis [4,21] . Further tests were performed to screen the phenotype of the PRL-3 knockout mice. We established growth curve for PRL-3 -/mice and PRL-3 +/+ mice and found no significant difference in their body weights. Blood chemistry indicators of PRL-3 -/mice and PRL-3 +/+ mice were also measured. Finally, we compared the   Fig. 4 Northern blot analysis. Total RNA of the heart, brain and muscle were extracted from wildtype and PRL-3 heterozygous mice, respectively.

Wildtype Homozygous
Wildtype Homozygous levels of glucose and indicators of lipid metabolism related factors between the homozygous and wildtype mice, and found no significant difference. However, PRL-3 -/mice exhibited abnormal response to IPGTT, which indicated that the mice might be used for further glucose metabolism related investigations. It has been previously reported that PRL-3 has some homology with PTEN in the phosphatase active site and flanking regions, and may act as a lipid phosphatase at the cytoplasmic face of the plasma membrane [21,22] .
In conclusion, we successfully generated the conditional knockout alleles for PRL-3, allowing Cremediated inactivation of PRL-3 during mouse development and in specific adult tissues. Moreover, our conditional PRL-3 knockout mice will be a valuable tool for uncovering the role of PRL-3 and its substrates in tumor metastasis that may be new potential therapeutic targets for cancer treatment.