Perturbation of the Hematopoietic System during Embryonic Liver Development Due to Disruption of Polyubiquitin Gene Ubc in Mice

Disruption of the polyubiquitin gene Ubc leads to a defect in fetal liver development, which can be partially rescued by increasing the amount of ubiquitin. However, it is still not known why Ubc is required for fetal liver development and the nature of the defective cell types responsible for embryonic lethality have not been characterized. In this study, we assessed the cause of embryonic lethality with respect to the fetal liver hematopoietic system. We found that Ubc was highly expressed in the embryonic liver, and the proliferation capacity of fetal liver cells was reduced in Ubc−/− embryos. Specifically, Ubc was most highly expressed in hematopoietic cells, and the proliferation capacity of hematopoietic cells was significantly impaired in Ubc−/− embryos. While hematopoietic cell and hematopoietic stem cell (HSC) frequency was maintained in Ubc−/− embryos, the absolute number of these cells was diminished because of reduced total liver cell number in Ubc−/− embryos. Transplantations of fetal liver cells into lethally irradiated recipient mice by non-competitive and competitive reconstitution methods indicated that disruption of Ubc does not significantly impair the intrinsic function of fetal liver HSCs. These findings suggest that disruption of Ubc reduces the absolute number of HSCs in embryonic livers, but has no significant effect on the autonomous function of HSCs. Thus, the lethality of Ubc−/− embryos is not the result of intrinsic HSC failure.


Introduction
Ubiquitin (Ub) is a small, highly conserved eukaryotic protein that plays a crucial role in diverse cellular signaling pathways, including targeting proteins for proteasomal degradation [1,2,3]. Inside cells, Ub exists in a dynamic equilibrium between free Ub and monomeric/polymeric Ub-substrate conjugate pools [4,5]. The Ub conjugation reaction is mediated by a series of enzymes E1-E3, and the deconjugation reaction is mediated by isopeptidases or deubiquitylating enzymes, during which most Ub is recycled back to the free Ub pool [6,7]. It is believed that maintaining cellular steady-state Ub levels is important for their function and survival [8,9,10]. Although Ub is a unique protein, it is encoded by two different classes of ubiquitin genes; constitutively expressed monomeric Ub ribosomal fusion genes and stressregulated polyubiquitin genes [11,12]. Under stress or even normal conditions, the contribution of polyubiquitin genes towards total cellular Ub levels is very significant [13,14].
In mammals, there are two polyubiquitin genes, Ubb and Ubc, and disruptions of these genes in mice have been shown to exhibit various phenotypes [13,14,15,16,17]. Disruption of Ubb reduced Ub levels in the gonads and hypothalamus, which resulted in infertility, hypothalamic neurodegeneration, metabolic abnormalities, and impaired energy and sleep homeostasis [14,15,16,17]. On the other hand, disruption of Ubc resulted in embryonic lethality with defective fetal liver development [13]. Although ectopic expression of Ub partially rescued the Ubc 2/2 phenotypes by delaying the onset of lethality, the precise mechanism underlying the cause of Ubc 2/2 embryonic lethality is still unknown.
In a study that used mouse embryonic fibroblasts (MEFs) isolated from Ubc 2/2 embryos, the phenotypes of Ubc 2/2 MEFs including reduced proliferation and delayed cell-cycle progression were found to be completely rescued by increasing cellular Ub levels [13]. However, under stress conditions, ectopic expression of Ub was not sufficient to increase the cellular Ub levels observed in wild type cells under stress, resulting in the failure of rescuing the phenotypes of Ubc 2/2 MEFs. Therefore, the phenotypes of Ubc 2/2 MEFs are a direct consequence of reduced cellular Ub levels. In addition, we previously showed that the contribution of Ubc toward total Ub levels is highest in liver among all other tissues investigated in adult mice [13]; therefore, it is highly likely that the defective fetal liver development is closely related to the reduced cellular Ub levels in Ubc 2/2 embryonic liver.
Here, as was observed in Ubc 2/2 MEFs, we found that fetal liver cells exhibit reduced proliferation, presumably due to the reduced cellular Ub levels in fetal liver. Rapid enlargement of the fetal liver during the midgestation period is important because the fetal liver becomes the primary site of hematopoiesis at embryonic days (E) 11-11.5, during which hematopoietic stem cells (HSCs) seed the liver from the aorta-gonad mesenepheros (AGM) region [18]. Through hematopoiesis, all the blood cell types are generated including myeloid and lymphoid lineages [19], therefore it is an essential process for survival. It has also been known that the cycling status of fetal liver HSCs is higher than that in adult bone marrow (BM) stem cells, which are largely quiescent [20]. However, here we demonstrate by non-competitive and competitive reconstitution methods that cell autonomous hematopoietic function in the Ubc 2/2 embryonic liver is essentially intact. These findings show that Ubc 2/2 embryonic lethality is not directly caused by the impaired function of HSCs themselves, but rather by autonomous insufficiency in Ubc 2/2 liver function to support hematopoiesis.

Reduced proliferation and increased apoptosis in Ubc 2/2 fetal liver cells
We have previously demonstrated that targeted disruption of polyubiquitin gene Ubc leads to embryonic lethality between embryonic days (E) 12.5 and 14.5, which is most likely due to the defect in fetal liver development [13]. Although we were able to partially rescue the lethal phenotype of Ubc 2/2 embryos by ectopic expression of Ub, the exact cause of embryonic lethality is still unknown. At stages E11-11.5, the size of the liver rapidly increases because it becomes the principal source of hematopoietic activity. This suggests that fetal liver cells need to undergo rapid proliferation during the midgestation period to reach a dense and homogenous architecture at E13.5. In fact, we have shown that the size of the fetal liver is smaller in Ubc 2/2 embryos when compared to wild type (Ubc +/+ ) at E13.5 [13]. To determine whether the reduced size of Ubc 2/2 fetal liver was due to the reduced proliferation of fetal liver cells, sagittal liver sections from E12.5 and E13.5 embryos were generated and stained with a proliferation marker, Ki-67 or phospho-histone 3 (PH3) ( Figure 1A). As expected, we observed a marked reduction in the proliferation of Ubc 2/2 fetal liver cells, which was demonstrated using two different proliferation markers (Ki-67 and PH3). In addition, not only was reduced cellularity apparent, but also an altered liver architecture with dissociated parenchymal cells were observed in Ubc 2/2 fetal liver ( Figure 1A, upper panel). Therefore, to determine whether apoptosis increased in Ubc 2/2 fetal liver, liver sections from E12.5 and E13.5 embryos were analyzed using the TUNEL assay ( Figure 1B). Although TUNEL-positive cells were rare in wild-type or Ubc +/2 liver sections, positive cells were observed focally in Ubc 2/2 liver sections, especially where the liver structure was altered.
To identify the cell types that were affected in proliferation and apoptosis due to deletion of Ubc, we then processed fetal liver cells from E13.5 embryos to prepare cytospin slides attached with equal number of fetal liver cells. We stained fetal liver cells with CD45 (pan-hematopoietic marker), E-cadherin (epithelial cell marker), AFP (hepatocyte marker), and Ter119 (erythroid cell marker) in combination with a proliferation marker ( Figure 1C) or with an apoptotic cell detection using the TUNEL assay ( Figure 1D). We found that the number of CD45, E-cadherin, AFP, and Ter119positive cells were not largely distinguishable between Ubc +/2 and Ubc 2/2 fetal liver cells. Based on our data, all cell types that we investigated exhibited reduced proliferation in Ubc 2/2 embryonic livers ( Figure 1C). Intriguingly, we found that Ter119-positive Ubc 2/2 fetal liver cells were more apoptotic than any other cell types that we investigated ( Figure 1D). Taken together, it seems that, due to deletion of Ubc, all cell types were affected in proliferation and Ter119-positive cells were affected most in apoptosis. Therefore, Ubc may be necessary for fetal liver cell proliferation and may protect cells from undergoing apoptosis. These results may suggest that Ubc expression is required in an attempt to increase cellular Ub levels in these apoptotic cells. It is also possible that Ubc may be upregulated as a stress response in apoptotic cells, although we cannot exclude the possibility that Ub deficiency in a Ubc 2/2 background may lead to cellular apoptosis.

Ub protein and Ubc expression levels are high in fetal liver
To investigate how cellular Ub levels in the fetal liver are affected during embryonic development, we determined total Ub levels in whole embryos and embryonic livers by indirect competitive ELISA, in which all forms of Ub was converted to monomeric Ub using a Ub-specific protease (Usp2-cc) and the monomeric Ub was quantified ( Figure 2A, left panel) [21]. We found that total Ub levels in both whole embryos and embryonic livers were well maintained throughout the midgestation period from E11.5 to E13.5, with significantly higher Ub levels in embryonic livers. However, a loss of Ubc reduced total Ub levels by approximately 40% in E12.5 embryonic liver (Figure 2A, right panel). In accordance with cellular Ub levels during embryonic development, Ubc mRNA levels, which were determined by quantitative real-time RT-PCR, did not change significantly during the midgestation period (data not shown). Although it was expected that the other polyubiquitin gene Ubb would be upregulated to compensate for the loss of Ubc in Ubc 2/2 fetal liver, because of the lower levels of Ubb than Ubc in fetal liver as well as the fact that the Ub-coding potential of Ubb was about half that of Ubc (9 Ubs per Ubc transcript vs. 4 Ubs per Ubb transcript), it seems that its contribution to increase the total Ub levels was quite minimal ( Figure 2B). In accordance with high Ub levels in embryonic livers, the spatial distribution of Ubc expression in Ubc +/ 2 embryos, which was determined by direct visualization of GFP fluorescence from a GFP-puro r fusion protein knocked in to the Ubc locus, clearly indicated that Ubc was highly expressed in the embryonic liver ( Figure 2C, indicated by an arrow). In contrast, Ubb was not expressed at high levels in the Ubb +/2 embryonic liver, and was comparable to other tissues ( Figure 2C, indicated by an arrow), although both Ubb and Ubc were highly expressed in the embryonic heart ( Figure 2C, indicated by arrowheads). Therefore, given the fact that the contribution of Ubc to the total Ub levels in liver was relatively higher than in other tissues [13], high Ubc expression in the embryonic liver seems to be responsible for the high Ub levels in the embryonic liver.
Ubc expression levels are high in hematopoietic cells and their proliferation capacity is impaired in Ubc 2/2 fetal liver To identify the cell types that exhibit high Ubc expression and the cell types that show altered proliferation capacity upon deletion of Ubc, we processed fetal liver cells, stained with CD45 and Ter119, and analyzed by flow cytometry. We were able to identify CD45-positive cells in both Ubc +/2 and Ubc 2/2 fetal liver cells, which were present at similar frequencies of about 7% ( Figure 3C). Similarly, we found that fetal liver-derived progenitor cell frequency was about 2% in both Ubc +/2 and Ubc 2/2 fetal liver Paraffin-embedded liver sections were prepared from wild-type (Ubc +/+ ) and Ubc 2/2 embryos at E13.5 and stained with the proliferation marker, Ki-67. (Lower panel) Frozen liver sections were prepared from Ubc +/2 and Ubc 2/2 embryos at E12.5 and stained with the proliferation/mitotic marker, PH3 (red), and DNA was visualized using TO-PRO-3 iodide (blue). Ubc +/2 embryos also served as controls because they had no phenotypic difference relative to wild-type embryos. (B) (Upper panel) TUNEL assay of paraffin-embedded E13.5 embryonic liver sections. TUNEL-positive cells were rare in Ubc +/+ embryonic liver sections, but were observed focally in Ubc 2/2 embryonic liver sections. (Lower panel) TUNEL assay of frozen E12.5 embryonic liver sections (red), with DNA visualization using TO-PRO-3 iodide (blue). (C). Cytospin slides were prepared from Ubc +/2 and Ubc 2/2 embryos at E13.5 and stained with CD45 (panhematopoietic marker), E-cadherin (epithelial cell marker), AFP (hepatocyte marker), and Ter119 (erythroid cell marker) in combination with a proliferation marker (PH3 or Ki-67) and DNA was visualized with DAPI. Examples of proliferating cells that were also positive for cell-type specific markers were indicated by arrowheads. (D) Fetal liver cells on cytospin slides were subjected to the double-labeling fluorescence TUNEL assay with appropriate markers. Examples of apoptotic cells that were also positive for cell-type specific markers were indicated by arrowheads. All data are representative images from three different embryos per genotype. Scale bars, 50 mm (upper panels in (A/B) and all panels in (C/D)); 200 mm (lower panels in (A/B)). doi:10.1371/journal.pone.0032956.g001 cells (data not shown). Because total liver cellularity was reduced by about 60% in the Ubc 2/2 fetal liver relative to Ubc +/2 fetal liver ( Figure 3A), these results indicate the concomitant reduction in the absolute number of hematopoietic cells and HSCs/progenitor cells in the Ubc 2/2 fetal liver. Ter119-positive cells were most abundant in both Ubc +/2 and Ubc 2/2 fetal liver cells, although its frequency was reduced by about 10% in Ubc 2/2 fetal liver cells ( Figure 3C). We then monitored GFP fluorescence of these cell populations in Ubc +/2 fetal liver cells, and found that CD45-positive cells exhibited the highest GFP fluorescence, suggesting that Ubc expression levels in hematopoietic cells were much higher than any other cell types ( Figures 3D and 3F). We also demonstrated that CD45-positive cells were highly proliferating and the percentages of Ki-67-positive proliferating cells were significantly reduced in Ubc 2/2 fetal liver cells regardless of the cell types ( Figures 3E and  3G), which is consistent with the immunostaining results (see Figure 1C). Therefore, our data suggest that CD45-positive hematopoietic cells have the highest Ubc transcriptional activity and proliferation capacity in E13.5 embryonic liver, and there is a correlation between Ubc transcriptional activity and proliferation capacity among different cell types.
The capacity of hematopoiesis in Ubc 2/2 fetal liver cells was only slightly affected under competitive microenvironment At the midgestation period, HSCs migrate from the AGM into the liver, and the liver becomes the major site of hematopoiesis [18]. During this period, fetal liver cells proliferate extensively. We hypothesized that the reduced proliferation capacity in Ubc 2/2 fetal liver may hamper the function of HSC to support hematopoiesis, which can be detrimental to embryonic development. Therefore, we decided to examine whether impaired HSC function could be a cause for Ubc 2/2 embryonic lethality. In order to test whether the abnormality in the Ubc 2/2 fetal liver results in cell autonomous defects in HSC function, we assayed the progenitor/HSC activity using the non-competitive and competitive reconstitution method. For non-competitive reconstitution, 1610 6 whole fetal liver cells obtained from E13.5 embryos were transplanted into lethally irradiated mice ( Figure 4A). For competitive reconstitution, 1.5610 5 fetal liver cells were transplanted together with 2610 5 congenic recipient-type bone marrow cells ( Figure 4B). At 4, 8, 12, and 16 week after transplantation,  peripheral blood cells from recipient mice were analyzed to examine the contribution of donor cells to lymphoid (T cells and B cells) and myeloid lineages (data not shown for 8 and 12 weeks). In both competitive and non-competitive settings, loss of Ubc had only a modest impact on HSC function both in terms of repopulation kinetics and total reconstitution of B cell, T cell, and myeloid lineages (Figures 4A and 4B). However, the long-term contribution to myeloid lineages was significantly diminished in the Ubc 2/2 competitive transplants at 16 weeks after transplantation.
To explore this further, we monitored donor-marked granulocyte reconstitution, which, because of their short life span (approximately 4-6 days) in contrast to T cells and B cells, has been used as a measure of an on-going HSC activity [22,23]. Therefore, continued stem cell activity is required to keep generating granulocytes. Both granulocyte frequencies ( Figure 4C, left panel) and granulocyte chimerism ( Figure 4C, right panel) were diminished in the absence of Ubc, suggesting subtle autonomous defects in Ubc 2/2 HSCs. However, because the overall reconstitution of Ubc 2/2 HSCs was generally comparable to control cells, we can conclude that the Ubc 2/2 embryonic lethality may not be directly caused by cell autonomous failure in HSCs, although we cannot exclude the possibility that the reduced number of HSCs may play a role in embryonic lethality.

Discussion
Disruption of the polyubiquitin gene Ubc leads to a defect in fetal liver development, which can be partially rescued by increasing the amount of ubiquitin (Ub) [13]. However, it is still not known why Ubc is required for fetal liver development and the nature of defective cell types that are responsible for the lethality in Ubc 2/2 embryos have not yet been characterized. At embryonic days (E) 11-11.5, right before the onset of Ubc 2/2 embryonic lethality, hematopoietic stem cells (HSCs) migrate into the liver, which rapidly enlarges and becomes the principal site of hematopoietic activity [18]. Therefore, it is possible that the cause of Ubc 2/2 embryonic lethality may be due to the failure in the fetal liver hematopoietic system.
In this study, we tried to identify the cell-type specific role of Ubc in fetal liver development, which cannot be compensated for by other ubiquitin genes. To this end, we first monitored the proliferation capacity in Ubc 2/2 fetal liver cells and found that it was significantly reduced and associated with increased apoptosis. Specifically, we found that Ubc was most highly expressed in hematopoietic cells, and the proliferation capacity of hematopoietic cells was significantly impaired in Ubc 2/2 embryos. Based on flow cytometric analysis, Ubc transcriptional activity seemed to correlate well with the proliferation status of the cells (see Figures 3F and 3G). These results suggest that upregulation of Ubc to maintain enough Ub pools may be required for cell proliferation. Therefore, cells with high Ubc transcriptional activity are highly proliferative with enough Ub pools. However, it is also possible that cells with high Ubc transcriptional activity are simply under stress regardless of their proliferation status or cell types.
All of the polyubiquitin gene knockout phenotypes seem to be related to cell-type specific reduction of Ub levels. For example, since Ubc was highly expressed in MEFs, Ub levels were reduced by ,40% in Ubc 2/2 MEFs [13]. We demonstrated that reduced proliferation, premature senescence, and abnormal cell cycle progression in Ubc 2/2 MEFs were completely rescued by increasing the Ub levels up to wild-type levels [13]. Since Ubb is highly expressed in germ cells, Ub levels were dramatically reduced by ,70% in adult Ubb 2/2 testes, which is mostly comprised of germ cells [14]. Interestingly, although Ub levels were not significantly reduced in Ubb 2/2 ovaries, they were reduced by ,70% in isolated oocytes from Ubb 2/2 mice, suggesting that the effect on total Ub levels due to disruption of polyubiquitin gene depends highly on the contribution of the polyubiquitin gene to the total Ub pools, which varies based on cell types [14]. Hypothalamic phenotypes in Ubb 2/2 mice also seemed to be closely related to the reduced Ub levels in the hypothalamus by ,30%, but not in the whole brain [15]. Therefore, reduced proliferation of Ubc 2/2 fetal liver cells is likely caused by the Ub deficiency in the Ubc 2/2 embryonic liver, in which Ub levels were reduced by ,40%. Currently, it is not clear which cell types in the embryonic liver are affected most due to deletion of Ubc, although we speculate that hematopoietic cells may be affected most, simply based on their high Ubc expression levels. This type of analysis would require the isolation of specific cell types and determination of Ub levels therein.
Interestingly, whereas the frequency of primitive HSCs and hematopoietic cells were maintained (about 2% and 7% of total liver cells, respectively), the reduced cellularity of the Ubc 2/2 fetal livers lead to a reduction in the absolute number of HSCs and hematopoietic cells. We next assessed the cause of Ubc 2/2 embryonic lethality with respect to the failure in the fetal liver hematopoietic system. To investigate whether the hematopoietic defect is caused by the loss of Ubc during embryonic development, we transplanted whole fetal liver cells from wild-type and Ubc 2/2 E13.5 embryos into lethally irradiated recipient mice by noncompetitive and competitive transplantation, and performed peripheral blood analysis to measure donor contribution to hematopoietic lineages at 4, 8, 12, and 16 weeks after transplantation. Under non-competitive conditions, the donor contribution to lymphoid and myeloid lineages was not significantly different between the two genotypes, suggesting that the hematopoietic function of Ubc 2/2 fetal liver remained intact. However, under competitive conditions, Ubc 2/2 fetal liver exhibited a marginal yet significantly reduced capacity to reconstitute hematopoietic function. Therefore, these combined results suggest that although the hematopoietic system in Ubc 2/2 embryos is slightly impaired, the embryonic lethality of Ubc 2/2 mice is not likely caused by an autonomous failure of the hematopoietic system.
Although Ubc 2/2 embryonic lethality may be caused by the reduced number of HSCs and hematopoietic cells with impaired proliferation capacity, our data also suggest that defects in erythroid cell development may also contribute to the Ubc 2/2 embryonic lethality. Although Ubc expression levels in Ter119positive cells were not as high as CD45-positive cells (see Figure 3F), many Ter119-positive cells were apoptotic and the proliferation capacity of Ter119-positive cells was also significantly impaired in Ubc 2/2 fetal liver (see Figure 3G). These were reflected in the reduced frequency of Ter119-positive cells in Ubc 2/2 fetal liver cells (see Figure 3C). In addition, as shown in our previous report [13], the color of Ubc 2/2 fetal liver was very pale and the close examination of fetal liver histology revealed the reduced number of mature enucleated red blood cells, which were stained intensely red with H&E, in Ubc 2/2 fetal liver. Taken together, it is also possible that Ubc 2/2 embryonic lethality could be caused, at least in part, by defects in erythroid cell development, possibly resulting in anemia. Further detailed analysis will prove the cell autonomous and non-cell autonomous role of Ubc in erythroid cell development.
In conclusion, Ubc 2/2 embryonic lethality does not seem to be due to intrinsic failure of the hematopoietic system, but rather may For non-competitive reconstitution, lethally irradiated recipient mice were transplanted with whole fetal liver cells from E13.5 Ubc +/+ and Ubc 2/2 embryos (n = 3 each). For each genotype, whole fetal liver cells from one embryo were used for transplantation into 3 recipient mice. (B) For competitive reconstitution, lethally irradiated recipient mice were transplanted with whole fetal liver cells from E13.5 Ubc +/+ (n = 10) and Ubc 2/2 embryos (n = 7) together with recipient-type whole BM cells. Whole fetal liver cells from three different Ubc +/+ embryos were used for transplantation into 13 recipient mice (3-5 recipient mice/embryo), but 3 out of 13 were not graphed because they were not multilineage engrafted. Similarly, whole fetal liver cells from three different Ubc 2/2 embryos were used for transplantation into 14 recipient mice (4-5 recipient mice/embryo), but 7 out of 14 were not graphed because they were not multilineage engrafted. In both reconstitution experiments, contributions of donor-marked progenitor/HSCs to lymphoid (B cells, T cells) and myeloid lineages were analyzed using peripheral blood drawn at 4 and 16 weeks after transplantation. (C) (Left panel) Donor-marked granulocyte frequency was expressed as % of white blood cells (%WBC) at 4-week intervals after transplantation of whole fetal liver cells from E13.5 Ubc +/+ (n = 11) and Ubc 2/2 embryos (n = 7) into recipient mice in a competitive manner. (Right panel) Donor granulocyte chimerism be due to the reduced number of HSCs. It is also possible that Ubc 2/2 embryonic lethality could, at least in part, be due to defects in the proliferation and/or differentiation capacity of fetal liver epithelial cells such as hepatocytes and ductal cells. Further investigation using these cell types is required to directly prove the role of Ub in fetal liver development and eventually identify the underlying molecular mechanisms for the embryonic lethality of Ubc 2/2 fetal liver.

Mouse studies
All mice were maintained in plastic cages with ad libitum access to food and water. All animal procedures followed National Institute of Health guidelines with the approval of Stanford University Administrative Panel on Laboratory Animal Care (APLAC; #13745) and the University of Seoul Institutional Animal Care and Use Committee (UOS IACUC; UOS-091201-1).
The TUNEL (terminal deoxynucleotidyl transferase dUTP nick end labeling) assay was performed using an ApoptagH peroxidase in situ apoptosis detection kit (Chemicon) for paraffin-embedded sections, and an ApoptagH red in situ apoptosis detection kit (Chemicon) for thaw-mounted frozen sections according to the manufacturer's protocols.

Transplantation and HSC/progenitor reconstitution analysis
Transplantation. Recipient mice (CD45.1) were lethally irradiated with a dose of 800 rad and whole fetal liver cells from donor (CD45.2) were obtained from E13.5 embryos. For noncompetitive transplantation, 1610 6 cells were transplanted by tail vein within 6 hours of irradiation. For competitive transplantation, 1.5610 5 cells were transplanted by tail vein with 2610 5 recipienttype whole bone marrow (BM) cells.

Reconstitution analysis.
Reconstituted mice were periodically bled via tail vein to monitor contribution by donormarked HSC/progenitors in the peripheral blood of B, T, and myeloid lineages at 4, 8, 12, and 16 weeks post-transplant. Using a heat lamp, the tail was heated and a sharp incision was made to collect about 4-6 drops of blood in a tube containing 10 mM EDTA/PBS. An equal volume of 2% dextran/PBS was added to generate a density gradient and incubated at 37uC for 30 min to precipitate red blood cells. The supernatant was collected and the remaining red blood cells were lysed in buffer containing ammonium chloride and potassium bicarbonate. Cells were then stained with anti-CD45.1-PE antibody (to detect recipient cells), was determined as described in Materials and Methods. All data in (A) and (B) are expressed as the means 6 SEM from the indicated number of samples. In (C), open squares and circles represent data from an individual sample and closed squares and circles represent the means 6 SEM from the indicated number of samples. # P,0.1; * P,0.05; ** P,0.01 vs. wild-type (Ubc +/+ ) embryos. doi:10.1371/journal.pone.0032956.g004 anti-CD45.2-FITC antibody (to detect donor cells), anti-Ter119-Cy5PE antibody (erythroid cell marker), anti-B220-Cy7APC antibody (B lymphocyte marker), anti-CD3-APC antibody (T lymphocyte marker), and anti-Mac1-Cy7PE antibody (myeloid cell marker). Using LSR II (BD Biosciences), 1610 6 events were collected and Ter119 2 cells (also PI 2 for live cells) were gated for donor cells (CD45.2 + ) and analyzed for B cells (B220 + ) vs. myeloid cells (Mac1 + , including Mac1 high and Mac1 low ) in donor cell populations. B220 2 and Mac1 2 populations were further analyzed for T cells (CD3 + ). As reported previously, T cells were hardly present in donor cell populations at 4 weeks post-transplant [24].
Donor-marked granulocyte frequency was expressed as % of white blood cells (WBC) at 4-week intervals after transplantation into recipient mice. Donor granulocyte chimerism was determined by analyzing the percentage of Ter119 2 CD3 2 B220 2 Mac1 high side scatter high cells that were also donor + .
Indirect competitive ELISA and quantitative real-time RT-PCR Indirect competitive ELISA and quantitative real-time RT-PCR were carried out as described previously [13,15].

Statistical analysis
Two-tailed unpaired Student's t-tests were used to compare the data between two groups. Although P,0.05 was considered to be statistically significant in most cases, P,0.1 was also noted in certain cases.