The role of variant histone H2AV in Drosophila melanogaster larval hematopoiesis

Within the nucleus the compaction of chromatin is important in not only protecting DNA from damage but also in controlling access of the transcriptional machinery to the genome (Lorch and Kornberg, 2015). Chromatin compaction is carried out with the aid of highly conserved histone proteins, which are assembled into the nucleosome structure. Each nucleosome is composed of 146 bp of DNA wrapped around an octamer of histones. This octamer includes two molecules each of histones H2A, H2B, H3 and H4 (Luger et al., 1997). In order for gene expression to occur, the compaction of chromatin must be relieved, allowing transcriptional machinery access to the DNA. One method whereby this is thought to be regulated is with the replacement of the replication-dependent canonical histones with replication-independent histone variants that can alter nucleosome structure and function (Luger et al., 1997).

Two variants of H2A that have been highly conserved in eukaryotes are H2AZ and H2AX (Malik and Henikoff, 2003). H2AZ can be found at gene regulatory regions and has been implicated in both the transcriptional activation of genes and the maintenance of gene silencing (Billon and Cote, 2013). By contrast, H2AX is important in the DNA damage repair response (Scully and Xie, 2013). In particular, phosphorylation of the C-terminal tail of H2AX is crucial for mediating the assembly of the repair machinery in response to double-strand breaks in chromosomal DNA. Unlike many other eukaryotes, the model organism Drosophila melanogaster has only one H2A variant, H2AV, which is a structural and functional chimera of H2AZ and H2AX (Baldi and Becker, 2013).

Previous studies of ATP-dependent chromatin-remodeling enzymes have shown a role for these complexes in Drosophila hematopoiesis (Badenhorst, 2014). NURF (nucleosome remodeling factor), a chromatin-remodeling complex that catalyzes nucleosome sliding, was found to play a role in hematopoiesis. Mutants of NURF display hemocyte-containing black melanotic masses, as well as increases in hemocyte number (Badenhorst, 2014; Kwon et al., 2008). Domino (Dom), a catalytic subunit of the Drosophila TIP60 complex (dTIP60), has also been implicated in the regulation of hematopoiesis. dTIP60 facilitates the acetylation and exchange of the phosphorylated H2A variant (H2AVph) with non-phosphorylated H2AV (Kusch et al., 2004). dom loss-of-function mutants in Drosophila larvae form black melanotic masses that are composed of necrotic prohemocytes (Braun et al., 1997). These results suggested a role for Dom in prohemocyte survival. This function appears to have been conserved in mice, where a Dom homolog was shown to be important both in primitive hematopoiesis in the embryo and in the maintenance of hematopoietic progenitor cells in the adult bone marrow (Fujii et al., 2010; Ueda et al., 2007).

Given the role that the dTIP60 complex plays in both H2AV exchange and hematopoiesis, we thought it would be informative to study a potential role for H2AV itself in hematopoiesis. Drosophila has three types of hemocytes that are most similar to members of the myeloid lineage of blood cells found in vertebrates (Evans et al., 2003). The most prevalent Drosophila blood cells are the plasmatocytes, which make up 90-95% of the mature hemocytes and are phagocytic in function. The second most prevalent blood cells are the crystal cells, which comprise ∼5% of the mature hemocytes and contain prominent cytoplasmic granules. Crystal cells aid in melanization, which is important for wound healing and the innate immune response. Finally, the rarest of blood cells are the lamellocytes, which are seldom seen under normal conditions. If an immune challenge is detected, the differentiation of these cells can be upregulated drastically. The role of the lamellocytes is to encapsulate and neutralize any objects too large to be phagocytized by plasmatocytes (Evans et al., 2003).

The larval hematopoietic organ, the lymph gland, is a bilaterally symmetric structure that flanks either side of the dorsal vessel or heart tube (Fig. 1). Multiple lobes are located on either side of the heart and are termed primary, secondary and tertiary lobes. The primary lobes contain not only undifferentiated prohemocytes, but also differentiated hemocytes. The secondary and tertiary lobes commonly contain undifferentiated prohemocytes during larval stages. The primary lobes of the lymph gland, which are located at the anteriormost region of the hematopoietic organ, can each be subdivided into three distinct zones. The differentiated hemocytes are located in the outermost region called the cortical zone. The undifferentiated prohemocytes are located in the central region called the medullary zone. Finally, there is a small subset of cells located in the posterior medial portion of each primary lobe called the posterior signaling center (PSC) (Jung et al., 2005). It is the role of the PSC to send signals to the other cells of the lymph gland, thereby orchestrating their differentiation. The PSC sends out multiple signals that not only keep the prohemocytes of the medullary zone in an undifferentiated state, but that also lead to the differentiation of hemocytes within the cortical zone (Krzemien et al., 2007; Mandal et al., 2007). We now show that the absence of H2AV causes a variety of hematopoietic abnormalities in Drosophila larvae, including the formation of melanotic masses.

Unlike most other eukaryotes studied thus far, D. melanogaster has only one histone H2A variant, H2AV. This protein is thought to be a functional chimera of the H2AX and H2AZ proteins present in most other eukaryotes, including humans (Baldi and Becker, 2013). Comparison of the Drosophila H2AV protein and the two most common H2A variants in humans reveals that H2AV is much more closely related to human H2AZ than to human H2AX (Fig. 2A,B). However, the C-terminal tail of Drosophila H2AV does contain an H2AX-like SQXY peptide that can be phosphorylated in response to DNA damage (Madigan et al., 2002). The rest of the human H2AX protein is far more closely related to the canonical H2A proteins of human and Drosophila, than to human H2AZ and Drosophila H2AV. These observations are consistent with a model in which the H2AZ gene family arose via duplication of, and divergence from, an ancestral H2A-like gene early during eukaryotic evolution (Malik and Henikoff, 2003). On the other hand, genes encoding H2AX appear to have arisen multiple times during eukaryotic evolution. This most frequently occurred by a more recent duplication of an H2A-encoding gene followed by the addition of sequences encoding an SQXY-containing C-terminal tail to one duplicated copy. In D. melanogaster, sequences encoding an SQXY-containing tail appear instead to have been added to an unduplicated H2AZ-encoding gene, thereby giving rise to the gene encoding H2AV.

In order to study the role of the His2Av gene that encodes H2AV in D. melanogaster, we examined His2Av⁸¹⁰/His2Av⁸¹⁰ homozygous mutants and observed a previously unreported formation of black melanotic masses during larval stages (Fig. 2C). His2Av⁸¹⁰ is a null allele that contains a 311 bp deletion that removes the region surrounding and including the second exon of the gene (van Daal and Elgin, 1992). His2Av⁸¹⁰/His2Av⁸¹⁰ mutants arrest in development during late third instar or prepupal stages (Kotova et al., 2011). In order to test whether the formation of the melanotic masses was due to the loss of H2AV, rather than some other unknown mutation in the genetic background of the homozygous animals, we crossed His2Av⁸¹⁰/+ heterozygotes with flies heterozygous for Df(3R)BSC524, a large deletion mutant with molecularly defined endpoints in the same region of the genome (Cook et al., 2012). This chromosomal deficiency (Df) lacks 312,855 contiguous base pairs on the right arm (R) of chromosome 3, including the entire His2Av gene. In heterozygous His2Av⁸¹⁰/Df(3R)BSC524 larvae we observed the formation of similar melanotic masses, implying that the absence of H2AV is indeed the cause of this phenotype (Fig. 2D).

To further test whether the absence of H2AV alone is responsible for the observed melanotic masses, we introduced a transgene expressing an H2AV-RFP fusion protein under the control of the His2Av promoter into His2Av⁸¹⁰/His2Av⁸¹⁰ animals. No melanotic masses were observed (data not shown). These results again imply that the absence of H2AV is indeed the cause of this phenotype.

The formation of melanotic masses in Drosophila has been commonly linked to the immune response (Minakhina and Steward, 2006; Rodriguez et al., 1996). Generally, hemocytes undergo melanization as an immune response to a present threat. However, in certain mutants, this response occurs without the presence of any discernable immune challenge (Harrison et al., 1995; Lemaitre et al., 1995; Minakhina and Steward, 2006; Zettervall et al., 2004). In order to determine whether the melanotic masses we observed included hemocytes and, if so, which types of hemocytes, we immunostained the melanotic masses using antibodies against three different hemocyte cell types (Evans et al., 2014). A Nimrod (P1) antibody was used to detect plasmatocytes, a Prophenyloxidase (Propo) antibody to detect crystal cells, and an Atilla (L1) antibody to detect lamellocytes. We identified plasmatocytes and crystal cells within the melanotic masses (Fig. 2E), but did not find any lamellocytes (Fig. 2F). These results imply that the melanotic tumors observed in the absence of H2AV are composed of hemocytes.

In order to understand whether the absence of H2AV was having an effect on the cells of the lymph gland and whether the lymph gland was a possible source of the hemocytes that resulted in the formation of the melanotic masses, we first looked for the expression of H2AV in wild-type lymph gland lobes. We wished to ascertain whether H2AV would be found at varying expression levels within specific zones and cell types of the lymph gland. However, immunostaining revealed ubiquitous and uniform expression of H2AV in all of the different zones within the primary lymph gland lobes (Fig. 3A). We also looked at H2AV phosphorylation in the lymph gland in order to discern whether significant levels of DNA damage were present. We saw only small numbers of cells within the lymph gland that were positive for H2AVph under normal conditions (Fig. 3B,B′). In order to ascertain whether the cells of the lymph gland might be contributing to the hemocytes found within the melanotic masses that formed in the absence of H2AV, we looked at the primary lymph gland lobes of His2Av⁸¹⁰/His2Av⁸¹⁰ and His2Av⁸¹⁰/+ third instar larvae.

We found that the heterozygous mutants had larval lymph glands very similar to those of wild-type (WT) animals. Further investigation using antibodies against specific zones and cell types within the lymph glands of heterozygotes showed a WT pattern of expression of these markers (Fig. 3C,E,G). P1 was used to label plasmatocytes within the cortical zone (Fig. 3C), while Antennapedia (Antp) and Propo were used to label the PSC and the crystal cells, respectively (Fig. 3E,G).

In contrast to the WT-like phenotype seen in heterozygous larvae, examination of the His2Av⁸¹⁰/His2Av⁸¹⁰ larvae revealed a structural loss of most of the primary lymph gland lobes (Fig. 3D,F,H). A majority of the differentiated plasmatocytes and crystal cells had dispersed from the primary lobes in homozygous mutant larvae (Fig. 3D,H), along with a majority of the undifferentiated prohemocytes normally found within the medullary zone of the primary lobes (Fig. 3D,F,H). Although some crystal cell staining could still be seen within the primary and secondary lobes of the lymph gland (Fig. 3H), the staining for Propo appeared hazy, suggesting that the crystal cells might have burst and released the Propo protein into the surrounding area.

Interestingly, the one portion of the lymph gland that appeared to remain intact in His2Av⁸¹⁰/His2Av⁸¹⁰ larvae was the Antp-positive PSC (Fig. 3F,H). Some of the homozygous mutant larvae did contain ectopic clusters of Antp-positive cells distinct from the PSC within the primary lobes (Fig. 3H′), as well as the abnormal presence of Antp-positive cells within the secondary or tertiary lobes of the lymph gland (Fig. 3H). Normally, Antp is restricted to a discrete region within the primary lobes of the lymph gland. Although the area occupied by the PSC appeared to be expanded within some of the homozygous mutant larvae relative to heterozygotes, the number of Antp-positive cells was similar in the primary lobes of the lymph glands in both genotypes (Fig. S1). Rather, the proportion of the primary lobe occupied by Antp-positive PSC cells increased dramatically from a mean of 4% in heterozygotes to 79% in homozygotes (Fig. S1).

Together, these results suggest that H2AV is required within prohemocytes and its absence might alter the response of these cells to signals originating from the PSC. Consistent with this hypothesis, the Hedgehog (Hh) ligand that prevents differentiation of prohemocytes was detected by immunostaining in the PSC cells of both heterozygous and homozygous His2Av⁸¹⁰ mutant lymph glands (Fig. S2A,B) (Mandal et al., 2007). Likewise, levels of Wingless (Wg) protein expression appeared similar in heterozygous and homozygous His2Av⁸¹⁰ mutant lymph gland lobes (Fig. S2C,D). Past studies have shown that Wg plays a dual role within the lymph gland, maintaining undifferentiated prohemocytes and dictating PSC cell number (Sinenko et al., 2009).

In order to determine whether the His2Av⁸¹⁰/His2Av⁸¹⁰ lymph gland phenotype could be rescued, we introduced a transgene expressing an H2AV-RFP fusion protein into the homozygous mutant background. We were able to rescue most of the dispersal phenotype of the lymph glands of these larvae (Fig. S3A-F). However, we did observe an increase in plasmatocyte differentiation in both the primary and secondary lobes (Fig. S3C,D). Crystal cells, present in both primary and secondary lobes, appeared to have burst (Fig. S3E,F, arrowheads). It is possible that the incomplete nature of the rescue was due to the RFP domain that is fused to the H2AV protein. In this regard, others have previously reported that a similar fusion of GFP can disrupt some of the normal functions of H2AV in vivo (Kotova et al., 2011).

In order to determine whether the premature differentiation of prohemocytes in the absence of H2AV was a cell-autonomous property, we drove expression of UAS-His2Av^HMO5177 RNAi with a domeless-GAL4 driver. A UAS-GFP transgene acted as a reporter for the location of domeless-GAL4 and UAS-His2Av RNAi expression. Importantly, domeless-GAL4 is expressed in the prohemocytes but not in the PSC of the primary lobes of the lymph gland (Jung et al., 2005). domeless-GAL4 expression initially occurs in a majority of the prohemocytes of the primary lobe, but continued expression then becomes restricted to the undifferentiated prohemocytes within the medullary zone as the cells within the cortical zone start to differentiate. domeless is also expressed in the prohemocytes of the secondary lobes. In order to test whether we were achieving a significant reduction of H2AV protein levels, we carried out immunostaining with an anti-H2AV antibody (Fig. 4A-B′). We observed a dramatic reduction of H2AV protein levels in a majority of the prohemocytes found within the primary lobes of the lymph gland, whereas H2AV protein was still expressed in the cells of the PSC, as expected (Fig. 4A,A′).

Lymph glands in which levels of H2AV protein were greatly reduced in the prohemocytes, via domeless-GAL4-driven RNAi, displayed a marked increase in the differentiation of plasmatocytes and crystal cells in comparison to controls (Fig. 5A-C,E-G). This increased differentiation of plasmatocytes and crystal cells was seen not only in the medullary zone of the primary lobes, but was also present extensively within the secondary lobes of the lymph gland (Fig. 5C,G). By contrast, control lymph glands did not display differentiation within the secondary lobes at similar stages (Fig. 5A,E). We also looked for differentiation of lamellocytes, but did not see any substantial increase in the number of differentiating lamellocytes in comparison to controls (data not shown). Since differentiation can be regulated by the presence of the PSC, we also looked at the expression of Antp within these lymph glands using an Antp-specific antibody (Fig. 5I-K). However, we did not see any discernable differences in the number or location of Antp-positive cells in the PSC of domeless-GAL4-driven UAS-His2Av RNAi lymph glands. These results are consistent with a cell-autonomous function of H2AV within the prohemocytes of the lymph gland.

In order to test whether H2AV was also required within cells of the PSC for normal prohemocyte differentiation, we drove UAS-His2Av RNAi and UAS-GFP expression with an Antp-GAL4 driver that is expressed in the PSC cells but not in the other cells of the lymph gland (Mandal et al., 2007). Expression of the Antp-Gal4 driver matched Antp protein expression as could be seen by immunostaining (Fig. 5L). We used anti-H2AV immunostaining to verify that the expected reduction in H2AV protein levels occurred only within the cells of the PSC and not in other cells within the lymph gland (Fig. 4B,B′). In lymph glands with greatly reduced levels of H2AV in the PSC, plasmatocyte differentiation appeared normal and the medullary zone was maintained (Fig. 5D). However, a decrease in crystal cells was observed in comparison to controls (Fig. 5H). Given the possible bursting of the crystal cells observed in the His2Av⁸¹⁰/His2Av⁸¹⁰ mutants and the reduction in crystal cell number when an Antp-GAL4 driver was used to greatly reduce levels of H2AV in the PSC, it is possible that proper H2AV expression within the PSC is required for crystal cell differentiation and maintenance.

In Drosophila, the H2AX-like function of H2AV requires a C-terminal SQAY motif that is phosphorylated on the serine when double-strand breaks in DNA are present nearby (Madigan et al., 2002). In order to determine whether loss of the H2AX-like function of H2AV led to the hematopoietic phenotypes seen in His2Av⁸¹⁰/His2Av⁸¹⁰ larvae, we tested for rescue with a previously described H2AV mutant protein that lacks the C-terminal 14 amino acids including the SQAY motif (H2AV^CT) (Clarkson et al., 1999). This mutant protein was also able to rescue most of the lymph gland dispersal phenotype (Fig. 6B,E). The number of Antp-positive PSC cells in the H2AV^CT-rescued His2Av⁸¹⁰/His2Av⁸¹⁰ mutant larvae primary lobes was similar to that in heterozygous larvae with or without the H2AV^CT-encoding transgene (Fig. S1). However, some abnormalities were present in the H2AV^CT-rescued homozygous mutant larvae, including the presence of Antp-positive cells in the secondary lobes (Fig. 6B). Crystal cell and plasmatocyte differentiation appeared to be increased, and some shedding of hemocytes was still observed (Fig. 6E), although not to the extent seen in unrescued His2Av⁸¹⁰/His2Av⁸¹⁰ mutants.

Because the rescue of normal larval hematopoiesis by H2AV^CT was incomplete, we wondered whether this mutant protein might itself have a dominant phenotype. Expression of H2AV^CT protein altered the normally WT lymph gland phenotype of His2Av⁸¹⁰/+larvae (Fig. 6A,D). H2AV^CT caused crystal cell and plasmatocyte differentiation to no longer be restricted to the cortical zone, and some early shedding of differentiated hemocytes was seen (Fig. 6D).

The phenotype seen when H2AV^CT is expressed in a H2AV⁸¹⁰/+ background suggested a possible dominant effect. To further test this hypothesis, we examined larvae containing the H2AV^CT protein in a His2Av wild-type background (Fig. 6C,F). We found that H2AV^CT again caused a phenotype that included differentiation of crystal cells and plasmatocytes within the secondary lobes (Fig. 6F), and shedding of hemocytes from the primary lobes.

We wished to distinguish among different hypotheses that might explain the dominant lymph gland phenotypes caused by the presence of the mutant H2AV^CT protein described above. The observed dominant H2AV^CT phenotype could be due to the absence of a short peptide upstream of the SQAY motif. Alternately, the phenotype might be caused by the dilution of wild-type H2AV proteins capable of responding to DNA damage by mutant H2AV^CT proteins, thereby causing increased levels of cell death within the organism. This in turn might result in increased mobilization of hemocytes from the lymph gland in response to cell damage and/or cell death. Finally, it is possible that an increased dosage of mutant or wild-type H2AV might cause a dominant phenotype in larval hematopoiesis.

Overexpression of wild-type H2AV protein in the lymph gland via either a domeless-GAL4 or an Antp-GAL4 driver resulted in less cohesion of the Antp-positive cells of the PSC (Fig. S4C-E). Overexpression of wild-type H2AV in the PSC via Antp-GAL4 also resulted in lower numbers of differentiated plasmatocytes and crystal cells (Fig. S4D,E). However, a similar alteration in differentiated hemocyte numbers was not observed when the overexpression of wild-type H2AV was driven via a domeless-GAL4 driver (Fig. S4A,B). Similar alterations in hemocyte differentiation or PSC cell cohesion were also observed when we drove expression of either a mutant H2AV protein that mimicked constitutive phosphorylation of the serine 137 residue within the SQAY motif (H2AV^SE; Fig. S4F-J) or a mutant H2AV protein that was non-phosphorylatable due to a serine-to-alanine substitution (H2AV^SA; Fig. S4K-O) with domeless-GAL4 or Antp-GAL4 (Kotova et al., 2011). Taken together, these results show that increasing the dosage of H2AV proteins, whether wild type or mutant, can also cause alterations in larval hematopoiesis.

We discovered that absence of the variant histone protein H2AV results in the formation of larval melanotic masses containing plasmatocytes and crystal cells. Previous studies have proposed that the formation of melanotic masses can be due either to the response of a normal immune system to abnormal tissue formed during development, or to a developmental defect in the hemocytes of the lymph gland (Rizki and Rizki, 1980; Rodriguez et al., 1996; Watson et al., 1991). Our data showing the loss of a majority of the primary lymph gland lobes in the His2Av⁸¹⁰ null mutant, as well as the early differentiation of the medullary zone and secondary lobe prohemocytes when H2AV levels were reduced via RNAi, are consistent with the latter model. Our results demonstrate an important role for H2AV during normal hemocyte differentiation and dispersal. Interestingly, studies using a human histiocytic lymphoma cell line or normal macrophages differentiated with macrophage colony stimulating factor (M-CSF; CSF1) have shown an upregulation of the His2Av-related human H2A.Z (H2AFZ) gene during macrophage differentiation (Baek et al., 2009). These results imply an evolutionarily conserved role for the closely related H2AV and H2AZ histone variants in blood cell differentiation.

The presence of black melanotic masses in Drosophila larvae is not restricted to His2Av mutants. This phenotype has previously been observed in mutants of two different ATP-dependent chromatin-remodeling complexes. Dom, which is a catalytic subunit of the dTip60 complex, plays a role in H2A variant exchange in nucleosomes, as well as in DNA damage repair (Kusch et al., 2004). dom loss-of-function mutants display black melanotic masses that are composed of melanized lymph glands (Braun et al., 1997). Mutants have shown that the vertebrate homolog of dom is required for both embryonic and adult hematopoiesis in the laboratory mouse (Fujii et al., 2010; Ueda et al., 2007). Loss of a subunit of another ATP-dependent chromatin-remodeling complex, NURF, also causes melanotic masses (Badenhorst, 2014). In addition, melanotic masses have been observed in mutations that affect various signaling pathways. For example, constitutive activation of the JAK-STAT pathway via the dominant gain-of-function Hop^TUM mutation results in the formation of melanotic masses (Hanratty and Dearolf, 1993). Constitutive activation of the Toll pathway via the dominant gain-of-function Tl^10b mutation also causes melanotic masses (Lemaitre et al., 1995). These observations raise the question of whether the closely related variant histones H2AV and H2AZ might be required to repress these evolutionarily conserved signaling pathways in hematopoietic cells.

Although the majority of the cells in the primary lymph gland lobes in His2Av mutants are lost, the Antp-positive cells comprising the PSC are spared and can be seen adjacent to the cardioblasts of the dorsal vessel. In addition, these cells express the Hh ligand that normally prevents premature differentiation of hemocyte precursors. The presence of Antp-positive cells can also be observed in posterior lymph gland lobes, where Antp is not normally expressed. In this regard, previous studies have shown that His2Av can function as a Polycomb Group (PcG) gene, and PcG proteins are known to be important for repressing the transcription of homeotic genes such as Antp (Swaminathan et al., 2005). In particular, it has been reported that Antp expression is expanded in the central nervous system of larvae that are mutant for His2Av. Reduction of H2AV levels via RNAi in the prohemocytes of the primary lobes, as well as in the secondary lobes, led to increased differentiation of plasmatocytes and crystal cells. This suggests that H2AV also acts downstream of the signals that originate from the PSC and that maintain the prohemocytes of the medullary zone in an undifferentiated state.

Reduction of H2AV levels via RNAi causes a less severe phenotype than that of His2Av⁸¹⁰ null mutants, in that the primary lobes of the lymph gland are preserved. However, there is a loss of the undifferentiated prohemocytes found within the medullary zone, as these cells differentiate into mature hemocytes. Previous studies in the testis of Drosophila have shown an important role for H2AV in the maintenance of both the germline and cyst stem cells (Morillo Prado et al., 2013). Together, these results suggest a possible role for H2AV in the transcriptional control of genes important for stem cell maintenance in general. In this regard, the closely related H2AZ protein of mammals has been reported to be important for the differentiation of embryonic stem cells in culture (Creyghton et al., 2008).

H2AV might be exerting its effects on the lymph gland through various signaling pathways that have been shown to orchestrate prohemocyte differentiation. Two pathways that might be affected are the Hh and Wg signaling pathways. Hh has been implicated in maintaining prohemocytes in an undifferentiated state (Mandal et al., 2007). However, we observed Hh expression in the PSC of both heterozygous and homozygous mutant His2Av⁸¹⁰ larval lymph glands. Wg has been reported to not only maintain the prohemocyte population in an undifferentiated state, but also to dictate PSC cell number (Sinenko et al., 2009). We did not detect any significant differences in the staining of prohemocytes and PSC cells with anti-Wg antibodies in homozygous versus heterozygous mutant His2Av⁸¹⁰ larval lymph glands. These results suggest that loss of H2AV might alter the intracellular responses to these ligands rather than their expression.

Drosophila H2AV is a chimeric protein that plays the roles of two widely conserved variant histones, H2AX and H2AZ (Baldi and Becker, 2013). H2AX is important for the DNA damage repair response (Scully and Xie, 2013), while H2AZ is important for both transcriptional activation and gene silencing (Billon and Cote, 2013). Previous studies have shown that H2AV^CT, which lacks H2AX function, is able to rescue the lethal phenotype seen in His2Av⁸¹⁰ null mutants, allowing the organisms to progress to pupation and adulthood (Clarkson et al., 1999). We found that H2AV^CT was able to partially rescue the His2Av⁸¹⁰ null larval hematopoietic phenotype, arguing that an H2AZ-like function rather than an H2AX-like function of H2AV is required for hematopoiesis. Nevertheless, differentiation within the lymph gland still appeared disrupted and partial loss of the primary lymph gland lobes could be seen. In addition, the expression of Antp was at times seen to expand into the posterior lobes of the lymph gland. This lack of full rescue could be due to a decreased stability of the H2AV^CT protein. However, the presence of H2AV^CT in an otherwise His2Av wild-type background was sufficient to cause abnormalities of the lymph gland lobes. Furthermore, overexpression of wild-type or of phosphorylation mutants of H2AV also caused hematopoietic abnormalities. Together, these results imply that a precise dosage of H2AV protein is essential for normal hematopoiesis in Drosophila. Similar alterations in differentiation might also occur in other organs and tissues. In this regard, care should be taken when using His2Av-GFP and His2Av-RFP transgenes, which are popular markers in live imaging.

The formation of black melanotic masses in the His2Av⁸¹⁰ null mutant establishes larval hemocytes as a useful tool for further studies of H2AV function. Furthermore, given the role that H2AV plays not only in undifferentiated prohemocytes, but also in the germline and cyst stem cells found in the testis (Morillo Prado et al., 2013), it will be interesting to test whether H2AV also regulates stem cells found in other tissues.

Mutant or transgenic fly stocks used in these studies included: domeless-GAL4, UAS-GFP (Grigorian et al., 2013; Mandal et al., 2007); Antennapedia-GAL4, UAS-GFP (Grigorian et al., 2013; Mandal et al., 2007); His2Av⁸¹⁰/TM3,Sb¹ (Bloomington Drosophila Stock Center); His2Av^CT (Clarkson et al., 1999); UAS-His2Av^WT (Kotova et al., 2011); UAS-His2Av^SE (Kotova et al., 2011); UAS-His2Av^SA (Kotova et al., 2011); UAS-His2Av^HMO5177 RNAi (Bloomington Drosophila Stock Center); and His2Av-RFP (Schuh et al., 2007). Additional stocks with balancer chromosomes expressing dominant larval markers were obtained from the Bloomington Stock Center. All GAL4×UAS-RNAi crosses were carried out at 29°C. All other stocks and crosses were grown at 25°C in standard laboratory conditions.

Primary antibodies were against Antennapedia (mouse; 1:4; Developmental Studies Hybridoma Bank), P1 (mouse; 1:10) (Kurucz et al., 2007), Propo (rabbit; 1:1000) (Muller et al., 1999), L1 (mouse; 1:10) (Kurucz et al., 2003), H2AV (rabbit; 1:1000) (Leach et al., 2000), H2AVph (mouse; 1:25; Developmental Studies Hybridoma Bank) (Lake et al., 2013), Hedgehog (rabbit; 1:200) (Taylor et al., 1993), Wingless (mouse; 1:50; Developmental Studies Hybridoma Bank) and GFP [either rabbit (1:1000; Abcam, ab6556) or mouse (1:250; Abcam, ab1218)]. All fluorescent secondary antibodies were conjugated to Alexa Fluor dyes (Thermo Fisher Scientific; mouse 594, A11032; mouse 488, A11029; rabbit 594, A11037; rabbit 488, A11034) and were used at 1:500. Samples were counterstained with the DNA dye TOTO-3 iodide (Thermo Fisher Scientific) added at 1:250-1:500 to the Vectashield (Vector Labs) in which the samples were mounted.

Lymph glands were dissected from wandering third instar larvae. Lymph glands were dissected by pulling out mouth hooks with lymph glands attached. Tissue was fixed for 10 min in a 1:3 solution of 37% formaldehyde: 1×PBS. Immunostaining was carried out as previously described (Grigorian et al., 2011).

We are grateful to I. Ando, R. Glaser, S. Heidmann, H. Muller, K. Suyama, A. Tulin, the Bloomington Drosophila Stock Center, FlyBase and the Developmental Studies Hybridoma Bank for providing stocks, database access and antibodies.