Generation and characterization of CD19-iCre mice as a tool for efficient and specific conditional gene targeting in B cells

The Cre/loxP system is a powerful tool for generating conditional gene knockout (KO) mice and elucidate gene function in vivo. CD19-Cre and Mb1-iCre transgenic mice are commonly used for generating B cell-specific KO mice and investigate the development, as well as the physiological and pathophysiological roles of B cells. However, the CD19-Cre line low efficiency and the Mb1-iCre line occasional ectopic recombination represent challenges for their use. Thus, we developed a CD19-codon-improved Cre (CD19-iCre) knock-in mouse with the T2A-iCre sequence inserted into the Cd19 locus, just before the stop codon. The CD19-iCre mice were compared with existing models, crossed with the Rosa26-EYFP reporter mice, and their recombination activity in B cells carrying different Cre alleles was assessed. CD19-iCre mice showed more effective Cre recombination in the early B cell developmental stages compared with the CD19-Cre mice. The efficiencies of the CD19-iCre and Mb1-iCre lines were similar; however, the B lineage-specific recombination was more stringent in the CD19-iCre line. Furthermore, the utility value of the CD19-iCre model was superior than that of the CD19-Cre mice regarding deletion efficiency in IL10-floxed mice. Thus, the CD19-iCre line is a valuable tool for highly efficient gene targeting specific to the B cell compartment.

www.nature.com/scientificreports/ cell stage in the bone marrow during B cell development 14 . In contrast, the other B cell-specific Cre driver line Mb1-iCre, which was created by replacing the coding exons of the Cd79a (Mb1) locus with iCre recombinaseencoding cDNA, showed efficient Cre deletion in bone marrow B cells, including in pro-B cells, when crossed with R26 EYFP mice 11,13 . Thus, researchers often use the Mb1-iCre line for the deletion or manipulation of a gene to analyze early B cell development in the bone marrow. With respect to peripheral tissues such as the spleen and lymph nodes, both Cd19 Cre/+ R26 EYFP and Mb1-iCre/ Rosa26-EYFP (Mb1 iCre/+ R26 EYFP ) lines targeted B cells with high frequency (approximately 80-93% and ~ 99% of B cells, respectively) 11,13 . Therefore, these data indicate that Mb1-iCre mice are more efficient than CD19-Cre mice in the recombination of loxP sites during B cell development. However, it should be noted that after crossing with R26 EYFP reporter mice, there can be a small fraction of EYFP + T cells detected in the Mb1-iCre line, suggesting minor aberrant loxP recombination 11 . In addition, several studies reported the occurrence of Mb1-iCre-mediated recombination in the germline in some cases [15][16][17][18] . Although deletion efficiency using the same Cre driver can also vary depending on the floxed alleles, such unexpected Cre expression in the germline can lead to gene deletions being passed to the subsequent generation. Therefore, breeding strategies should be carefully adjusted for the maintenance of these lines 19 .
To overcome these limitations of the currently available B cell specific Cre lines, we developed a new B-lineage Cre mouse model, which is a Cd19-T2A-iCre knock-in line (CD19-iCre), in which the T2A-iCre sequence is inserted in-frame to the 3′ end of the Cd19 coding sequence (immediately upstream of the stop codon). By crossing with the R26 EYFP line, we show that the CD19-iCre driver line enables highly efficient and specific ablation of floxed genes in the B cell lineage from the pro-B cell stage without infidelity of Cre-mediated recombination in T cells. When crossed with IL10-floxed mice, the CD19-iCre line was found to be superior to CD19-Cre mice with regard to the disruption of IL-10 production in B cells. Thus, the CD19-iCre line provides a new option for generating B cell-specific deficient mice with high specificity and efficiency, which further facilitates investigating the functional roles of a gene throughout B lineage cells.

Results
Generation of the CD19-iCre knock-in mice. To generate a mouse line that expresses the iCre recombinase 20 in CD19-expressing B lineage cells without disturbing its endogenous Cd19 expression, we generated a CD19-T2A-iCre knock-in line, in which the T2A-iCre sequence was inserted in-frame to the 3′-end of the Cd19 coding sequence (immediately upstream of the stop codon) (Fig. 1a). Following the induction of double-strand DNA breaks by the Cas9 enzyme mediated by a guide RNA (gRNA), the homology arms guided the T2A-iCre to be inserted in-frame with the Cd19 open reading frame by homologous directed repair. This targeting strategy using self-cleaving 2A sequences 21 was designed to mediate bicistronic translation, which is potentially more efficient than the well-described internal ribosome entry site sequence. Upon translation of the chimeric CD19-T2A-iCre mRNA, the T2A sequence led to ribosome skipping 22 , resulting in the co-expression of CD19 and iCre as discrete proteins in CD19-expressing cells. Consequently, iCre expression was regulated by the Cd19 promoter in tandem with the endogenous CD19 expression. In the present study, the targeting construct was correctly introduced in embryonic stem (ES) cells, which was confirmed by polymerase chain reaction (PCR) analysis. This was followed by deletion of a neomycin resistance gene (Neo R ) cassette (Neo cassette) by Flp cDNA transfection into targeted ES cells to avoid unpredictable Cre activity 23 . After confirming the Neo cassette deletion by PCR, the targeted ES cells were transferred into blastocysts and the mouse germline (Fig. 1b). Heterozygous and homozygous CD19-iCre mice (referred to as Cd19 iCre/+ and Cd19 iCre/iCre , respectively) were found to be viable, fertile, and born at the expected Mendelian frequencies (data not shown). The sequences for the T2A-iCre have been inserted between the last amino acid and the stop codon in exon 14 of Cd19 locus. Neomycin resistance gene cassette (Neo) is flanked by Frt sites and inserted downstream of the untranslated region. The KI allele is obtained after Flippase (Flp) site-directed recombination of the selection marker. (b) PCR for detection of iCre transgene KI allele with genomic DNA from wild-type (WT), heterozygous and homozygous CD19-iCre mice. Amplicons of 484 and 247 bp are identified by primer pairs (p1-p2 and p1-p3) specific for the WT and iCre alleles, respectively. (c) Flow cytometry of the surface expression of CD19 on B220 hi B cells of peripheral blood from WT, heterozygous and homozygous CD19-Cre and CD19-iCre mice. Right, mean fluorescence intensity (MFI) of the staining of CD19. (d) Quantitative RT-PCR of mRNA encoding CD19 in WT and CD19 iCre/iCre B cells, normalized to the expression of β-actin. (e) Western blotting analysis of whole-cell lysates of splenic B cells from WT, CD19 iCre/iCre , and CD19 Cre/Cre mice with antibodies specific for CD19, Cre, and β-actin. The arrowheads and arrow indicate fusion proteins (CD19-T2A-iCre) and Cre/iCre proteins, respectively. (f) Absolute number or frequency of each B cell subset from bone marrow, spleen, mesenteric lymph node (mLN) and peritoneal cavity (PEC) and from WT and CD19 iCre/+ mice on the basis of total cell count and flow cytometry analysis. B cell subsets were as follow: pre-pro-B (IgM − IgD − B220 + CD19 − CD43 + ), pro-B (IgM − IgD − B220 + CD19 + CD43 + ), pre-B (IgM − IgD − B220 + CD19 + CD43 − ), immature B (IgM + IgD -B220 low CD19 + ), recirculating B (Rec: IgM + IgD + B220 hi CD19 + ) and plasma cells (PC; CD138 + TACI + ); follicular (FO; CD19 + CD93 -CD21 low CD23 hi ), marginal zone (MZ; CD19 + CD93 − CD21 hi CD23 low ); B1a (IgM + CD5 + CD43 + B220 low CD19 + ), B1b (IgM + CD5 − CD43 + B220 low CD19 + ); germinal center B cells (GC; CD19 + Fas + CD38 − ). Data are representative of three independent experiments (b, d, e), or pooled from three (c) or two (f) independent experiments. Data are presented as mean ± SD. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ns, not significant. The P values were obtained by one-way ANOVA with Tukey's post hoc test (c) or two-tailed unpaired t test (d, f). www.nature.com/scientificreports/ To understand the consequences of the CD19-iCre targeted insertion strategy on CD19 expression, we analyzed peripheral B cells in mice heterozygous or homozygous for the knock-in allele by flow cytometry and compared the collected data with that of a previously published CD19-Cre line 10 , in which the Cre was inserted in exon 2 and the CD19 coding sequence was disrupted, leading to a CD19 deficiency in the homozygous situation. CD19-Cre homozygous mice showed complete loss of CD19 expression, whereas the CD19-iCre homozygous mice retained CD19 expression, albeit there was a significant reduction in its levels compared with that in B cells from wild-type (WT) mice. The mean fluorescence intensity (MFI) of CD19-stained B cells in CD19-iCre mice decreased by about fourfold compared with WT mice (Fig. 1c). Moreover, Cd19 mRNA and protein levels were also reduced in splenic B cells from Cd19 iCre/iCre mice (Fig. 1d, e), as assessed by quantitative RT-PCR and western blot analysis, respectively. It should be noted that high levels of released iCre and CD19 along with subtle levels of the uncleaved CD19-T2A-iCre fusion protein were detected, indicating efficient T2A self-processing in B cells (Fig. 1e). These results suggest that the targeted insertion of T2A-iCre might interfere with Cd19 expression or mRNA stability. Heterozygous mice for the Cre allele are commonly used as Cre deleter lines. Mice heterozygous for the Cre insertion in CD19-Cre mice, which retained one functional CD19 allele, showed reduced surface CD19 levels (the MFI was approximately half of that of WT mice) (Fig. 1c). Cd19 iCre/+ mice also showed reduced CD19 expression, but significantly higher compared with CD19-Cre mice (Fig. 1c). However, B cells from the Cd19 iCre/+ mice did not show any further apparent alterations and exhibited normal development in the bone marrow, spleen, lymph nodes, and peritoneal cavity (Fig. 1f).

Efficient recombination at different B cell developmental stages with the CD19-iCre transgene.
To determine the efficiency of Cre expression, CD19-iCre mice were crossed with the R26 EYFP line to generate Cd19 iCre/+ R26 EYFP mice. Almost all CD19 + B cells expressed EYFP in the spleen and mesenteric lymph nodes (mLN), but EYFP was not detected in CD19 − non-B cells (Fig. 2a), suggesting high B cell-specific Cre expression in the CD19-iCre lines. Next, we examined the efficiency of Cre-mediated deletion in B cell subsets of CD19-iCre mice and compared them side-by-side with previously published B cell specific Cre-driver lines: the CD19-Cre and the Mb1-iCre mice that were also crossed with the R26 EYFP mice. A previous study demonstrated that Mb1-iCre mice have an earlier and/or more efficient loxP site recombination in developing B cells than CD19-Cre mice 11 . Our results also confirmed these findings, since almost all of the pro-B cells expressed EYFP in Mb1 iCre/+ R26 EYFP mice, whereas only ∼50% of the same subset of cells were EYFP-positive in Cd19 Cre/+ R26 EYFP mice (Fig. 2b).
A comparison of our newly developed CD19-iCre line showed highly efficient Cre-mediated activation of EYFP expression in pro-B cells, similar to that in the Mb1-iCre line. In the case of the very early B lineage stage     (Fig. 2b). Other B cell subsets in the spleen (marginal zone B cells) and bone marrow (plasma cells and recirculating B cells) were comparable among the three Cre lines (Fig. 2b). These results demonstrate the highly efficient Cre-mediated recombination in B lineage cells from the pro-B cell stage in CD19-iCre mice.
Highly specific recombination in B cells with the CD19-iCre transgene. A previous study indicated aberrant loxP recombination in T cells from Mb1 iCre/+ R26 EYFP mice, albeit with a relatively low frequency 11 . To detect ectopic Cre-mediated activation of EYFP expression in Cd19 iCre/+ R26 EYFP and Mb1 iCre/+ R26 EYFP mice, flow cytometry analysis was conducted. In agreement with a previous report 11 , a small number of splenic CD4 + and CD8 + TCRβ + T cells from Mb1 iCre/+ R26 EYFP mice were found to be EYFP + (Fig. 3). In contrast, EYFP expression was virtually absent in T cells from Cd19 iCre/+ R26 EYFP mice (Fig. 3). Together, these data indicate the highly targeted expression of iCre in CD19-iCre mice, thereby demonstrating that this transgenic line is a very valuable resource to study B cell-specific gene functions.    www.nature.com/scientificreports/ CD19-iCre mice are a valuable line for studying B cell function. We have previously shown that lipopolysaccharide (LPS)-activated B cells produce IL-10 after stimulation with B cell antigen receptor (BCR) 24,25 . To test the utility of the CD19-iCre driver line for general knockout studies, this line was crossed with IL10 floxed (Il10 f/f ) mice 26 to generate Cd19 iCre/+ Il10 f/f animals. We compared these Cd19 iCre/+ Il10 f/f mice with CD19-Cre/IL10 f/f (Cd19 Cre/+ Il10 f/f ) mice regarding the recombination efficiency of floxed alleles in splenic B cells. The deleted Il10 floxed alleles in B cells, but not T cells, sorted from spleens with high purity (> 98%) were confirmed by genomic PCR in Cd19 Cre/+ Il10 f/f and Cd19 iCre/+ Il10 f/f mice (Fig. 4a). However, it is noteworthy that the recombination efficiency in B cells from Cd19 iCre/+ Il10 f/f mice was superior to that in Cd19 Cre/+ Il10 f/f mice. Enzyme-linked immunosorbent assay (ELISA) showed that Cd19 Cre/+ Il10 f/f B cells had compromised IL-10 secretion compared with control Cd19 Cre/+ B cells (Fig. 4b), in agreement with a previous study 27 . Notably, IL-10 production was significantly reduced in Cd19 iCre/+ Il10 f/f B cells than in Cd19 Cre/+ Il10 f/f B cells (Fig. 4b). Collectively, these data suggest that CD19-iCre mice are valuable Cre drivers to manipulate genes in B cells.

Discussion
In this study, we describe a new B cell-specific Cre line, named CD19-iCre, which is the third-generation line after CD19-Cre and Mb1-iCre. Our findings indicate that CD19-iCre can recombine loxP sites with high efficiency in pro-B cells of Cd19 iCre/+ R26 EYFP mice and show efficient Cre-mediated deletion of loxP sites during B cell development in several tissues, which is comparable to that of the Mb1-iCre line but greater than that of the CD19-Cre mice. Indeed, the Il10 floxed allele was more efficiently deleted in splenic B cells from Cd19 iCre/+ Il10 f/f mice compared with those from Cd19 Cre/+ Il10 f/f mice, and, consequently, Cd19 iCre/+ Il10 f/f B cells expressed less IL-10 than Cd19 Cre/+ Il10 f/f B cells.
Despite the previously reported studies on Mb1-iCre mice attributing the efficiency of the Cre deletion to promoter-specific differences between the Mb1-iCre and the CD19-Cre mouse lines, our present study shows that this may not be the case as the results exhibited highly efficient Cre-mediated recombination in pro-B cells of Cd19 iCre/+ R26 EYFP mice, which was comparable to that of Mb1 iCre/+ R26 EYFP mice. Given the endogenous expression of CD19 in the pro-B cell population, these results suggest that CD19-iCre mice can excise the floxed sequence more effectively than CD19-Cre mice. The enhanced efficiency of the CD19-iCre mice may be due to the use of the improved iCre, which has been reported to be more efficiently expressed in mouse cells 20 . Another possible underlying reason can be the difference in the design strategy of the Cre-transgenic lines. In the CD19-Cre line, in addition to the regulatory elements of the endogenous CD19 locus, a β-globin polyadenylation site was introduced to permit polyadenylation of the Cre mRNA. In the CD19-iCre line, we introduced the T2A-iCre sequence before the 3′-stop codon, with all other transcription processes of the Cre mRNA being retained, which might affect Cre expression in early B cell stage. Since the Cd19-Cre line used in this study possesses the Neo cassette, the recombination efficiency of the CD19-Cre mouse may be low; however, this is somewhat unlikely since it has been previously reported that excision of the Frt-flanked Neo cassette from the CD19-Cre allele leads to reduced Cre expression by an unknown mechanism 28 .
Conditional gene targeting based on the Cre/loxP system is a powerful technology for the analysis of gene function, but it is key that the expression of Cre transgenes occurs in a targeted cell-specific manner. Despite the differential sensitivity of individual loxP-flanked target gene alleles to Cre-mediated recombination, at least in the R26 EYFP reporter system, Mb1-iCre mice can exhibit undesired recombination in T cells. By contrast, our CD19-iCre mouse line did not show such ectopic recombination. Nonetheless, the proportion of detectable EYFP + T cells was minor in Mb1 iCre/+ R26 EYFP mice, in cases in which the target genes that are to be deleted and analyzed in B cells are critical for T cell development and function. Thus, it is important to carefully assess the recombination in T cells and interpret the phenotype. Until now we did not observe Cre-mediated recombination in the germline of Cd19 iCre/+ R26 EYFP mice; however, we cannot exclude the possibility that the CD19-iCre line induce germline deletions.
The T2A-based approach in CD19-iCre mice may produce the equimolar amounts of iCre relative to CD19 protein from a single multicistronic transcript. Generally, a 2A peptide, when placed between genes can induce ribosomal skipping during protein translation, thereby producing discrete proteins, whereas unsuccessful skipping and continued translation may result in a fusion protein. We could detect very little CD19-T2A-iCre fusion protein in Cd19 iCre/iCre B cells, which might explain the observed reduction in surface CD19 levels in CD19-iCre mice. However, given the reduced Cd19 expression in Cd19 iCre/iCre B cells, it seems most likely that T2A insertion may affect the expression or mRNA stability of Cd19. Although minor reduction in CD19 expression did not induced apparent alterations of B cell development in Cd19 iCre/+ mice, we advocate using Cd19 iCre/+ mice as controls for conditional BKO mice generated with this strain, as is the case with CD19-Cre or Mb1-iCre lines.
In summary, our newly developed CD19-iCre mice showed more effective Cre recombination at the early B cell stages, as well as throughout B cell development, when compared with that of the commonly used CD19-Cre line. Although the efficiency of the CD19-iCre line was comparable to that of the Mb1-iCre line, more stringent B lineage-specific recombination was observed in the novel model. Thus, its high efficiency and specificity make the CD19-iCre line the most reliable and robust model available for gene targeting and manipulation of B cells in vivo.
Genomic PCR analysis. Genomic DNA isolated from sorted cells or tail biopsies was used for PCR analysis.
Statistical analysis. GraphPad Prism 6 (GraphPad Software) was used for all statistical analyses. Statistical significance was determined by two-tailed paired or unpaired Student's t test for two groups or one-way ANOVA with Tukey's post hoc test for multiple groups. Comparisons of two nonparametric data sets were done by the Mann-Whitney U test. A p value of less than 0.05 was considered statistically significant.