Stable silencing of β-lactoglobulin (BLG) gene by lentivirus-mediated RNAi in goat fetal fibroblasts

β-lactoglobulin (BLG), a dominant allergen in goat milk, is difficult to remove by traditional biochemical methods. Its elimination from goat milk by genetic modification therefore poses a major challenge for modern goat breeders. A shRNA targeting BLG mRNA with high interference efficiency was identified, with which lentiviral vectors were used for mediating stable shRNA interference in goat-fetal fibroblast cells. Apart from high efficiency in the knockdown of BLG expression in these cells, lentivector-mediated RNAi manifested stable integration into the goat genome itself. Consequently, an in vitro model for goat BLG-content control was compiled, and a goat-cell line for accompanying transgenetic goat production created.


Introduction
Although nutritious for humans, especially infants, goat milk and its by-products are considered important allergens in food allergy, whence the restrictions against their wide use in the food industry. b-lactoglobulin (BLG), a major protein in milk (56%-60% of total cow-whey proteins), is considered a dominant allergen (Sharma et al., 2001;Fritsche, 2003).
Several biochemical approaches have been applied, in the attempt to reduce the allergenic potential of milk proteins, especially BLG. Heat treatment, enzymatic hydrolysis, fermentation and glycation (Fritsche, 2003;Hattori et al., 2004;Ehn et al., 2005), as well as the Maillard reaction (Guanhao Bu et al., 2010), are examples. Notwithstanding, these always entail increased costs and the accumulation of unexpected by-products. Hence, how to decrease BLG content in goat milk, and thus deaden milk allergy by genetic modification, remains a major challenge among modern goat-breeders.
RNA interference (RNAi) is now a widely-used technique for the sequence-specific knockdown of target mRNA (Dorsett and Tuschl, 2004;Mittal, 2004). The stable induction of RNAi usually involves short hairpin RNA (shRNA) expressed from plasmid or viral vectors. shRNA has a stem-loop structure that mimicks endogenous microRNA (miRNA), besides sharing the same cleavage mechanism and transport pathways (McCaffrey et al., 2002;Miyagishi et al., 2004). In animal cells, mature 21-25 nt miRNAs are produced in cytoplasm, with incorporation of the antisense strand into the RNA-induced silencing complex (RISC), thereby mediating gene silencing, this including mRNA cleavage and translational repression. Lentiviral vectors are potent and efficient in enhancing the capacity for stable RNAi mediating-gene silencing in animal cells (Stewart et al., 2003;Xie et al., 2008). This capacity is critical for further application, especially gene knockdown in transgenic animals.
The aim was to obtain goat transgenic fetal fibroblasts with stable RNAi BLG-gene suppression. An in vitro cell model to mimic in vivo goat mammary epithelial cells was built by inducing BLG overexpression in goat fetal fibroblasts. The model was then used to examine the efficiency of shRNA induced BLG expression silencing, as well as the lentiviral vectors for mediating stable shRNA interference in goat fetal fibroblast cells.

Construction of OE-BLG overexpression plasmids
Total RNA was isolated from goat mammary gland tissues, whereupon reverse transcription was carried out to prepare cDNA. A full-length wild-type goat BLG open reading frame (758 bp) was obtained by PCR, using cDNA as templates, as well as forward (5'-GCACTCGAGAGA AGCACGA-3') and reverse (5'-GGCGGATCCCTTTA TTGCTGAAGGAGGA-3') primers. Xho and BamH restriction sites (underlined) were respectively flanked with initiation and stop codons by PCR. The resultant PCR products were digested with Xhoand BamH, purified and cloned as overexpressional fragments into N1 vectors (Invitrogen). The final construct was termed OE-BLG.

Cell culture, transient transfection
Goat embryo fibroblast (GEF) cells (donated by Prof. Feng Wang, Nanjing Agricultural University) were grown in Dulbecco's modified Eagle's minimal essential medium (DMEM)-F12 (Gibco) supplemented with 15% fetal bovine serum (FBS) (Gibco) at 37°C, and 5% carbon dioxide (CO 2 ). 24 h before transfection, 3 x 10 5 cells were seeded into each well of a 6-well plate (Costar), and cultured in the growth medium without antibiotics, to so achieve 90%-95% confluence at the appropriate time. For transfection, 4 mg of each shRNA expression construct was used per Zhang et al. 681 Table 1 -Oligonucelotides used in producing shRNA against goat BLG mRNA.
Oligos Sequence well. Lipofectamine 2000 Reagent (Invitrogen) was used as the transfection reagent, according to manufacturer's instructions, i.e., mingling 4 mg of DNA with 6 mL of lipofectamine per well. For co-transfection, each shRNA construct was co-transfected with N1-BLG (2:1) into GEF cell cultures using lipofectamine. After 6 h past transfection, the lipofectamine & DNA mixture was removed from the wells and replaced by a fresh medium. After 48 h past transfection, cells were assayed by quantitative RT-PCR, with non-transfected and non-shRNA vector transfected cells as controls. 293FT virus producer cells (Invitrogen) were cultured in a complete D-MEM medium containing 10% FBS supplemented with 0.1 mM MEM Non-Essential Amino Acids, 1 mM sodium pyruvate and 2 mM Lglutamine. 293FT cells were maintained in the complete medium containing 500 mg/mL Geneticin (Invitrogen).

Real-time RT-PCR analysis
Total RNA was isolated 48 h after transfection using Trizol (Invitrogen) according to manufacturer's instructions. All RNA samples were reverse transcribed in a 25 mL reaction mixture at 37°C for 15 min with M-MLV reverse transcriptase (Promega), using Oligo-dT. The cDNA samples were stored at -20°C, prior to quantification and analysis. Primers were designed for BLG (Z19569) which span intron(s) to prevent amplification of genomic DNA, using the forward primer (5'-GGC GAG TGT GCT CAG AAG A-3') and the reverse primer(5'-GGG CTC AGC ACT GTT TTC C-3'). Real-Time RT-PCR (ABI 7300) was performed using FastStart Universal SYBR Green Master (Roche), according to manufacturer's protocol. PCR thermal cycle reactions were denaturation at 95°C for 10 min followed by 40 cycles of denaturation and annealing/extension at 95°C for 15 s and 60°C for 1 min. Amplification specificity during real-time RT-PCR was monitored by evaluating the melting curve and examining the products of real-time RT-PCR on agarose gel for the absence of nonspecific bands. The comparative Ct (threshold cycle) method (Livak and Schmittgen, 2001) was used to determine the relative amount of RNA transcripts of BLG gene present in the cells. The expression levels of each group of cells were normalized to b-actin levels, using the forward primer (5'-TGA ACC CCA AAG CCA ACC-3') and the reverse primer (5'-AGA GGG GT A CAG GGA CAG CA-3').

Production of lentivirus in 293FT cells
For the generation of lentivirus, 293FT cells were transfected with 9 mg of ViraPower Packaging mix (Invitrogen) along with 5 mg of pLenti-U6-shRNA2 plasmid by lipofectamine. 24 h post transfection, the transfection mix was replaced by a fresh culture medium (without antibiotics). The virus-containing supernatant was harvested 72 h post transfection, cleared by centrifugation (3000 rpm/min, 15 min, and 4°C), and then filtered through a 0.45 mm filter (Millipore). Lentivirus aliquots were subsequently prepared and stored at -80°C, until use in transduction. Virus generation and infection of target cells were carried out in class II biological safety cabinets in class II containment laboratories, with access restricted to authorized staff.

GEF-cell transduction with recombinant lentivirus and selected by Blasticidin
The day before transduction, goat embryo fibroblast cells were seeded into 6-well cell-culture plates at a density of 5 x 10 4 cells per well. On the day, they were transduced with 1 mL of the already described virus-containing supernatant. At 24 h post transduction, the mix was replaced with fresh cell-culture medium. 48 h post transduction, the cells were collected for RNA isolation and qRT-PCR analysis.
In order to produce stably transduced cells 48 h post transduction, 5 mg/mL of Blasticidin (Invitrogen) were added to the medium (DMEM/F12 + 15% FBS), as a means of selecting clones containing the insert. The blasticidincontaining medium was replenished every 3 days. The cells remained in the selective medium for 2 weeks, after which they were left to grow to 70% confluence before RNA isolation and qRT-PCR analysis. 2-fold serial dilutions of Blasticidin were prepared in a preliminary test, 5 mg/mL of Blasticidin being considered as the final selective concentration. After 12 days, Blasticidin-resistant GEF cell pools were established.

Statistical analysis
Real-time PCR results were analyzed by using the 2 -DDCT method (Livak and Schmittgen, 2001). All the experiments were performed in triplicate. Data were analyzed with the SPSS 13.0 Student t-test (SPSS Inc., Chicago, IL) to evaluate the significance of differences in gene expression. The results were expressed as mean ± STD. The level chosen to define whether treatments were significantly different from mock controls was p < 0.05.

Results shRNAs induce transient silencing of BLG in GEF
Individual shRNA efficiency in inducing RNAi in GEF cells was tested in transiently transfected cells. Three shRNA constructs against BLG were examined ( Figure 1). As EGF cells have no de novo BLG gene expression under the in vitro culture system, an in vitro model was established to investigate BLG RNAi, by subcloning the fulllength cDNA fragment of goat BLG gene into an overexpression plasmid (OE-BLG), and co-transfecting with shRNA expression plasmids. Synchronous with OE-BLG plasmid co-transfection, all the anti-BLG shRNA constructs resulted in a significant reduction (p < 0.05) in the level of BLG mRNA, compared to cells that were only transfected with OE-BLG plasmids, considered as positive 682 Lentivirus-mediated stable RNAi in GEF control ( Figure 2). As expected, as negative control, a luciferase-specific shRNA (shRNA-NC) failed to reduce BLG mRNA levels, thereby showing no significant difference from cells that were only transfected with OE-BLG. All told, the interference efficiency that shRNAs induced in GEF cells was acknowledged, and the specialty of shRNA targeting confirmed. Out of the three shRNA constructs, shRNA2 presented the highest efficiency in silencing BLG expression in GEF cells, with a reduction of approximately 90%, when compared with the positive control, whereas in the other two, although significantly high, this was less than 75% ( Figure 2). Thus, shRNA2 was chosen for sequential establishment of BLG-knockdown goat-cell lines.

Lentivirus-mediated shRNAs integrate into goat genome and provide stable BLG knockdown in GEF
The shRNA2 expression cassette was removed from pGP-U6-shRNA plasmid by restriction digestion, for posterior subcloning into a lentivector system, newly denominated "plenti-U6-shRNA2". On using these plasmids together with a virus packaging mix, shRNA2-recombined lentiviruses were produced for infecting GEF cells. After remaining 12-days in the final Blasticidin-containing selective medium, surviving GEF colonies were defined, whereas control cells without lentivirus infection had expired, thereby demonstrating the establishment of a BLGknockdown GEF cell-line.
To confirm the integration of lentivector-mediated shRNA into cell-line genome, genome DNA was isolated from BLG-knockdown GEF cells for detection of shRNA expression cassettes. Through analysis with agarose gel electrophoresis, these transgenetic fragments were detected in PCR products, but not in non-virus-infected control GEF cells (Figure 3). Further confirmation by gel extraction and gene sequencing provided additional evidence of, shRNA-expression-cassette stable integration into the goat genome.
In order to test BLG RNAi in established BLGknockdown GEF, OE-BLG was transfected into this cell line, while using normal GEFs as control. After 48 h of transfection, the mRNA was isolated and relative levels of BLG in each group of GEFs analyzed. Significantly, the high efficiency of RNAi against BLG in BLG-knockdown GEFs was observed, whereas in control GEFs, BLG expression remained at a relatively higher level (p < 0.01) (Figure 4). All told, these data confirmed the establishment of BLG-knockdown GEFs with the ability to express de novo shRNA against BLG.
shRNA interference efficiency correlates with the secondary structure of target mRNA Among all the shRNA constructs, shRNA2 presented the highest interference efficiency (Figure 2). This difference in efficiency can be explained by the interaction between target mRNA and complement RNA produced by shRNA. To investigate whether the BLG mRNA secondary structure could affect RNAi interference efficiency, this structure was calculated at 37°C. The calculated image showed shRNA1 and shRNA3 target regions to be selfparing in a large part and folded in stem structure (Figure 5). On the contrary, the shRNA2 target region was shown to unfold and stretch in a loop structure ( Figure 5). Such diversity in mRNA secondary structure indicated variable Zhang et al. 683 Figure 2 -Detection of shRNA efficiency against BLG. BLG mRNA levels were relatively lower in GEF cells co-transfected with both shRNA and BLG overexpression plasmids (2:1). Relatively higher BLG mRNA levels indicate those co-transfected with shRNA-NC and BLG overexpression plasmids (2:1), and those only transfected with BLG overexpression plasmids. shRNA-2 manifested the highest efficiency in BLG expression silencing. The amount of BLG mRNA overexpression was determined using Real-Time RT-PCR and normalized to b-actin mRNA. All the experiments were done in triplicate. The different letters indicate a significant difference of relative levels in each group (p < 0.05) and error bars the s. d.. Control, non-transfected GEF cells.

Figure 3
-Lentivector-mediated shRNA were stably integrated into the goat genome. In order to construct BLG-knockdown GEF cells, GEF cells were transfected with lentivector-mediated shRNA. After 15-days of selection with Blasticidin, GEF genomic DNA was isolated for plenti-Lacz-shRNA2 PCR identification. PCR products were analyzed by 1% agarose gel electrophoresis. The electrophoresis image indicates lentivector-mediated shRNA as being stably integrated into the goat genome. Control, non-infected GEF cells. molecular behavior while complement-RNA targeting, this resulting in differences in interference efficiency.

Discussion
In the present study, BLG expression silencing was with the lentivirus system mediating a stable RNAi in goat embryo fibroblasts. Lentivirus-mediated shRNA expression leads to long-term ability in silencing BLG expression. In the established BLG-knockdown GEFs, induced BLG expression was suppressed to 10% lower than in positive control cells (Figure 4). Such high efficiency had also been found in former shRNA transient transfection experiments (Figure 2). All together, the data indicate the capacity of the lentivirus system to effectively integrate sufficient DNA copies into goat genome to thus greatly deaden BLG expression, even in the disappearance of the interference caused by transient transfection.
Due to growing popularity, RNA-interference has become widely used in innumerous cases of gene silencing. Through induction by plasmids or virus vectors, RNAi has broadened its range of application (Bartel, 2004). DNA vectors facilitate the delivery of shRNA expression constructs into animal cells, to so achieve continuous longterm expression from either polor pol promoters, such as U6 (Brummelkamp et al., 2002). Besides infecting a wide range of cells in vitro and in vivo (Naldini et al., 1996), the lentivirus system is notably effective in shRNA delivery (Abbas-Terki et al., 2002). This system can also provide higher integration efficiency, whereas with conventional protocols depending on homologous recombination, integration efficiency was much lower, i.e., about 10 -6 . Even compared to the classical DNA microinjection (DNA-MI) technique, the system results in a four to eight-fold higher rate of transgenic animals per embryo treated (Pfeifer, 2004). In all these applications, nearly all the F 0 generation animals expressed the transgene (Lois et al., 2002;Pfeifer et al., 2002). Although not with goats, lentiviral transgenesis has been applied with high efficiency in producing trangenic large farm animals, such as pigs and cattle (Hofmann et al., 2003(Hofmann et al., , 2004Whitelaw et al., 2004). Hence, by using established BLG-knockdown GEF, our research work on transgenic goat production can be further developed.
shRNA2 interference efficiency is the highest in transient transfection expression. The assumption that BLG mRNA secondary structure may affect interference efficiency was amply sustained through subsequent computer calculations. As shRNA1 and shRNA3 target regions appeared over a large part of self-paring and folds in stem 684 Lentivirus-mediated stable RNAi in GEF Figure 5 -shRNA target regions in BLG mRNA. BLG mRNA secondary structure at 37°C was calculated to analyze mRNA target structural effects on RNAi efficiency. shRNA1 and shRNA3 target region structures were self-paring and stem-like, whereas the shRNA2 target region was relatively unfolded and looped.
structures, these could hardly unfold for complementary RNA guiding RISC complex binding and function. On the contrary, as the shRNA2 target region showed unfolding and stretching in loop structures, this secondary structure is apparently more feasible for targeting proceedures (Shao et al., 2007), thereby displaying higher interference efficiency in silencing BLG expression, whereby the assumption that secondary structure analysis of target mRNA is a credible method for predicting shRNA interference efficiency.
In conclusion, we are the first to use the RNAi mechanism via the lentivirus system to knockdown goat BLG expression. Our shRNAs successfully target BLG mRNA and reduce expression to a low level. Compared to the conventional protocol depending on homologous recombination, the letivirus system can provide higher integration efficiency, thus demonstrating its potencial as an efficient tool for transgenetic application. In addition, the establishment of BLG-knockdown GEF can be of use for further genetic engineering to produce transgenic goats.