Functional requirements for a Samd14-capping protein complex in stress erythropoiesis

Acute anemia induces rapid expansion of erythroid precursors and accelerated differentiation to replenish erythrocytes. Paracrine signals—involving cooperation between stem cell factor (SCF)/Kit signaling and other signaling inputs—are required for the increased erythroid precursor activity in anemia. Our prior work revealed that the sterile alpha motif (SAM) domain 14 (Samd14) gene increases the regenerative capacity of the erythroid system in a mouse genetic model and promotes stress-dependent Kit signaling. However, the mechanism underlying Samd14’s role in stress erythropoiesis is unknown. We identified a protein-protein interaction between Samd14 and the α- and β-heterodimers of the F-actin capping protein (CP) complex. Knockdown of the CP β subunit increased erythroid maturation in murine ex vivo cultures and decreased colony forming potential of stress erythroid precursors. In a genetic complementation assay for Samd14 activity, our results revealed that the Samd14-CP interaction is a determinant of erythroid precursor cell levels and function. Samd14-CP promotes SCF/Kit signaling in CD71med spleen erythroid precursors. Given the roles of Kit signaling in hematopoiesis and Samd14 in Kit pathway activation, this mechanism may have pathological implications in acute/chronic anemia.


Introduction
The expansion and differentiation potential of stem/progenitor cells in the hematopoietic system fluctuates in response to changing environmental conditions. Exposure to toxins, pathogens, blood loss, nutrient deficiency, oxygen deficits, etc. activates transcriptional programs and signal transduction pathways in progenitors and precursors to re-establish homeostasis (Bresnick et al., 2018;Socolovsky, 2007). In acute anemia, vital stress response mechanisms are communicated to erythroid progenitor and precursor cells by secreted factors, including stem cell factor (SCF), erythropoietin (Epo), bone morphogenetic protein (BMP)-4, glucocorticoids, and Hedgehog (Bauer et al., 1999; eLife digest Anemia is a condition in which the body has a shortage of healthy red blood cells to carry enough oxygen to support its organs. A range of factors are known to cause anemia, including traumatic blood loss, toxins or nutritional deficiency. An estimated one-third of all women of reproductive age are anemic, which can cause tiredness, weakness and shortness of breath. Severe anemia drives the release of hormones and growth factors, leading to a rapid regeneration of precursor red blood cells to replenish the supply in the blood.
To understand how red blood cell regeneration is controlled, Ray et al. studied proteins involved in regenerating blood using mice in which anemia had been induced with chemicals. Previous research had shown that the protein Samd14 is produced at higher quantities in individuals with anemia, and is involved with the recovery of lost red blood cells. However, it is not known how the Samd14 protein plays a role in regenerating blood cells, or whether Samd14 interacts with other proteins required for red blood cell production.
To shed light on these questions, mouse cells exposed to anemia conditions were used to see what proteins Samd14 binds to. Purifying Samd14 revealed that it interacts with the actin capping protein. This interaction relies on a specific region of Samd14 that is similar to regions in other proteins that bind capping proteins. Ray et al. found that the interaction between Samd14 and the actin capping protein increased the signals needed for the development and survival of new red blood cells.
These results identify a signaling mechanism that, if disrupted, could cause anemia to develop. They lead to a better understanding of how our bodies recover from anemia, and potential avenues to treat this condition.
Samd14 ΔEnh/ΔEnh precursors, an evolutionarily conserved SAM domain in SAMD14 functions to increase Kit signal transduction and erythrocyte regeneration (Ray et al., 2020). Despite structural similarities, the molecular features of the SAMD14 SAM domain are distinct from SAM domains in other proteins. SAM domains have well-recognized roles in signal transduction, including cell surface receptor activation, receptor endocytosis and MAP kinase activities (Nagamachi et al., 2013;Vind et al., 2020;Wang et al., 2018). While Samd14 SAM domain contributes to Samd14 activities, the Samd14 protein lacking a SAM domain retains some activities, suggesting additional functional domains. Here, we discovered that SAMD14 interacts with the barbed end actin capping protein (CP) complex via a non-canonical CP-interaction motif to promote erythroid precursor activity and differentiation during stress erythropoiesis. We also provide evidence that, while Samd14 promotes both SCF-and Epodependent cell signaling in stress erythroid precursors, the Samd14-CP complex is required for SCFdependent Kit signaling but not Epo signaling.

Results
The stress-activated Samd14 protein interacts with capping protein complex Samd14-enhancer (Samd14-Enh) deletion in mice reduces the regenerative capacity of the erythroid system (Hewitt et al., 2017). In a murine Gata1-null-erythroid (G1E) cell line resembling normal proerythroblasts, we deleted the Samd14-Enh with Transcription Activator-Like Effector Nucleases (TALENs) ( Figure 1A; Weiss et al., 1997). Samd14 protein was not detectable in G1E-ΔEnh cells ( Figure 1B). To test whether Samd14 protein interacts with other proteins in proerythroblasts, we infected G1E-ΔEnh cells with GFP-tagged retroviruses expressing empty vector (EV) or hemagglutinin (HA)-Samd14. To normalize levels of exogenous protein that best mimics endogenous levels, infected cells were purified based on a GFP low fluorescence-activated cell sorting (FACS) gating strategy (Figure 1-figure supplement 1A). GFP was used as a selection marker for this and all subsequent experiments involving retroviral infection. Samd14 interacting proteins were immunoprecipitated (IPed) using anti-HA antibody-conjugated beads and enrichment scores were quantitated by mass spectrometry. We detected 21 proteins which were >twofold enriched in HA-Samd14 pulldown samples vs EV (p<0.05) (Supplementary file 1). Gene ontology analysis revealed enrichment of proteins involved in cell-cell adhesion (6 proteins) and barbed end filamentous (F)-actin filament capping (4 proteins). The most enriched Samd14-interacting proteins (>10-fold) (Capzβ, Capzα1, and Capzα2) belonged to the same family of F-actin CPs, which form a CP complex to regulate actin filament assembly and disassembly (Isenberg et al., 1980;Schafer and Cooper, 1995). To validate the interaction, we IPed endogenous Samd14 from wild type (WT) G1E cells using an anti-Samd14 antibody. Capzβ co-IPed with Samd14 ( Figure 1C). To test whether the Samd14 interaction with Capzβ was SAM domain-dependent, we performed co-IPs in G1E-ΔEnh expressing either EV, HA-Samd14, or HA-Samd14 ΔSAM. No Samd14 or CP complex components were detected in IPs from EV-infected cells ( Figure 1D). However, Capzβ, Capzα1, and Capzα2 were pulled down in both HA-Samd14-and HA-Samd14 ΔSAM-expressing cells ( Figure 1D). These results established that Samd14 interacts with CP complex components Capzβ, Capzα1, and Capzα2 in proerythroblasts independent of the SAM domain.
The online version of this article includes the following source data and figure supplement(s) for figure 1: Source data 1. Source Western blot images for Figure 1. (ADD1) is a known to promote spectrin-actin assembly, while CP does not appear to interact with actin filaments (Gardner and Bennett, 1987;Kuhlman and Fowler, 1997).
The online version of this article includes the following source data and figure supplement(s) for figure 3: Source data 1. Source Western blot images for Figure 3.  Capzβ protein compared to control infections ( Figure 3C) with no change in Samd14 expression ( Figure 3D). Capzb knockdown in spleen erythroid precursors decreased the percentage of early R2/ R3 by 2.1-fold and 2.44-fold and increased the percentage of R4/R5 cells by 1.65-fold and 2.35-fold compared to controls ( Figure 3E). In Samd14 ΔEnh/ΔEnh cells, we observed similarly decreased percentages of R2/R3 and increased R4/5 cells (Figure 3-figure supplement 1C). Wright-Giemsa staining to assess cell morphology indicated that Capzb knockdown cultures contained more mature erythroblasts and reticulocytes than controls ( Figure 3F). Capzb knockdown cells were also smaller overall, as measured by the decreased forward scatter-area of the cells in R2-R5 populations, consistent with increased numbers of mature cells (Figure 3-figure supplement 1D). These data demonstrate that Capzb expression opposes erythroid maturation in developing fetal liver progenitors and in splenic stress erythroid precursors. Capzb knockdown data are comparable to previously published data in which Samd14 knockdown decreased R2/R3 percentages and increased R4/5 percentages in erythroid cultures (Hewitt et al., 2015), suggesting that Samd14 and Capzβ may have similar roles in erythropoiesis.
To determine whether CP complex regulates erythroid precursor activity, we performed colony forming unit (CFU) assays in cells infected with control or shCapzb2. Capzb knockdown cells were plated in methylcellulose semisolid media containing SCF and Epo to promote BFU-E and CFU-E activity. Compared to controls, Capzb knockdown cells formed 1.5-fold fewer BFU-E (p=0.016) and threefold fewer CFU-E ( Figure 3G and Figure 3-figure supplement 1E) colonies. Prior work indicated that Capzb depletion increased cell proliferation (Aragona et al., 2013). We quantitated proliferating erythroid precursor cells using Ki67 in CD71 low Kit + . The percentage of Ki67 + cells were 1.6-fold higher in Capzb-depleted cells compared to controls ( Figure 3-figure supplement 1D). To test whether Capzb promotes cellular viability and survival, we analyzed cultured spleen progenitors by flow cytometry for percentages of live (AnnexinV − Draq7 − ), early apoptotic (EA; AnnexinV + Draq7 − ), and late apoptotic cells (LA; AnnexinV + Draq7 + ) cells. In Capzb knockdown cells vs controls, the percentages of dead cells (Draq7 + ) in spleen and bone marrow cultures were higher throughout immunophenotypically defined stages of erythroid maturation (Figure 3-figure supplement 1E and F). Correspondingly, EA (AnnexinV + Draq7 -) and LA (AnnexinV + Draq7 + ) cells increased in Capzb knockdown cells vs control knockdown ( Figure 3H). Capzb knockdown increased the percentage of late apoptotic cells 2.8-fold in GFP + Kit + and 3.68-fold in GFP + CD71 + Ter119 + compared to control ( Figure 3I). The CP complex therefore promotes erythroid precursor activity and survival following acute anemia.
In lineage-depleted spleen cells 3 days after PHZ injection, we confirmed by co-IP that the Samd14-CP interaction occurs in endogenously expressed proteins ( Figure 4C). To determine if the Samd14-CP complex was dynamically regulated in anemia, we compared co-IP in control spleen to spleen isolated from PHZ-treated mice. In PHZ-treated spleen, more Capzb protein was pulled down compared to control spleen (Figure 4-figure supplement 1A), indicating that the frequency and/or stability of the Samd14-CP interaction was increased in acute anemia. No differences were seen in the amount of Capzb pulled down in CD71 + Ter119vs CD71 + Ter119 + cells, or in response to acute SCF stimulation (Figure 4-figure supplement 1B and C). Therefore, Samd14-CP complex formation was not stage-or signal-dependent.
We previously described a SAM domain requirement for Samd14-mediated erythroid progenitor activity in PHZ-induced stress erythroid progenitors. To determine if the Samd14 CPB domain promotes unique Samd14 functions, or whether SAM and CPB domains act synergistically, we utilized an ex vivo genetic complementation assay in which Samd14 is exogenously expressed in Samd14 ΔEnh/    ΔEnh PHZ-treated spleen erythroid progenitor cells. Primary cells infected with either full length Samd14, Samd14 lacking the SAM domain (S14-ΔSAM), CPB domain (S14-ΔCPB), or both (S14-ΔSAMΔCPB) were sorted based on GFP expression and plated in methylcellulose semisolid media for colony assays. Whereas Samd14 promoted BFU-E activity vs empty vector, S14-ΔSAM and S14-ΔCPB had fewer BFU-E colonies compared to Samd14 (1.5-fold and 1.6-fold, respectively) ( Figure 4D). In S14-ΔSAMΔCPB-infected cells, BFU-E activity was 2.2-fold lower than Samd14 (p<0.0001). BFU-E in S14-ΔSAMΔCPB-infected cells were decreased by 1.4-fold (p=0.009) vs S14-ΔSAM, suggesting a synergistic role of both domains to promote Samd14 activity in stress BFU-E ( Figure 4D). Consistent with prior results demonstrating that the SAM domain is not needed for CFU-E colony formation, we observed similar decreases in numbers of CFU-E in S14-ΔCPB (1.5-fold, p=0.0048) and S14-ΔSAMΔCPB (1.7-fold) expressing cells vs full length Samd14. Given the role of Samd14 in stress-activated SCF/Kit signaling (Hewitt et al., 2017;Ray et al., 2020), we tested whether the Samd14 CPB mediates SCF-dependent activation of cell signaling in stress erythroid cultures. 48 hr after infection with either control (EV) or HA-Samd14 retrovirus, lineage-depleted splenocytes from PHZ-treated Samd14 ΔEnh/ΔEnh mice were serum-starved, stimulated with SCF, and analyzed by flow cytometry for AKT and ERK activation. The cell surface marker CD71 (transferrin receptor) can be used to distinguish among stages of erythroid differentiation (Flygare et al., 2011). As each cell stage has distinct cell signaling requirements, we sub-divided GFP + cells into CD71 low , CD71 med , and CD71 high fractions, and then gated for Kit expressing cells by flow cytometry ( Figure 5A). CD71 low cells contained BFU-Es and very few cells expressing the mature erythroid marker Ter119 + , which increased in frequency in CD71 med and CD71 high cells ( Figure 5-figure  supplement 1A). As prior studies indicated, there is a continuum of cell phenotypes from CD71 low Kit + , CD71 med Kit + to CD71 high Kit + characterized by a transition from BFU-E to CFU-E, distinct proliferative indices, and transcriptional states (Li et al., 2019). CD71 med Kit + cells contained more cells with CFU-E potential compared to CD71 high Kit + ( Figure 5-figure supplement 1B and C). Samd14 expression increased the percentage of CD71 low Kit + (1.35-fold) and CD71 med Kit + (1.5-fold) cells ( Figure 5A). In EV-infected Samd14 ΔEnh/ΔEnh , the median fluorescence intensity (MFI) of CD71 low Kit + cells stained with phospho-ERK (pERK) or phospho-AKT (pAKT) antibodies increased by 62.7-fold and 65.3-fold, respectively, after SCF stimulation. Samd14 expression did not alter pAKT or pERK MFI in CD71 low Kit + cells ( Figure 5B and C). Consistent with this, pERK levels were similar between EV-infected WT and Samd14 ΔEnh/ΔEnh cells, indicating that Kit signaling in the CD71 low population is insensitive to Samd14 expression ( Figure 5C). In CD71 med cells, pERK levels were 4.3-fold (p<0.0001) lower in EV-infected Samd14 ΔEnh/ΔEnh cells compared to WT EV-infected cells, indicating that CD71 med cells are sensitive to Samd14 expression levels ( Figure 5D). pERK MFI in Samd14 ΔEnh/ΔEnh CD71 med cells was rescued by Samd14 ( Figure 5B and D). Samd14 expression increased the levels of pERK (4.4-fold) and pAKT (15.3-fold) in SCF-treated CD71 med cells compared to EV controls ( Figure 5D). In CD71 high cells, pAKT and pERK levels in Kit + cells do not change after SCF stimulation. However, pAKT and pERK levels in CD71 high Kit + cells are sensitive to Samd14 expression ( Figure 5E). These results indicated that distinct Kit + populations gated on CD71 show varying signaling response (as measured by pAKT and pERK) to SCF and Samd14, but do not distinguish whether Kit + cells in the CD71 + gates are completely unresponsive to SCF stimulation or whether Kit signaling is occurring via other pathways. Finally, we tested whether the SAM or CPB domains of Samd14 could alter Samd14 sensitivity in the SCFresponsive CD71 med Kit + cells. Deleting either the SAM or the CPB domain of Samd14 reduced levels of pAKT (fourfold) and pERK (twofold) compared to Samd14 in the CD71 med population ( Figure 5F). As an alternative assessment of Kit signal-promoting roles of the Samd14 CPB domain, we examined the activation of Kit transcription in response to SCF stimulation ( Figure 5-figure supplement 1D). Whereas Kit transcript levels were significantly increased in Samd14-expressing cells vs controls, we did not detect significant increases in Kit transcript levels in S14-ΔCBP expressing cells ( Figure 5figure supplement 1D). Both the SAM and CPB domains of Samd14 are required for maximal promotion of Kit signaling in stress erythroid precursors.

Samd14
-LAKGRRHRPSRSRLRDS-Neurabin-1 -SVKNRRQRPSRTRLYDS- The online version of this article includes the following source data for figure 6: Source data 1. Source Western blot images for Figure 6.
central arginine residues are required to mediate CP complex interaction (Johnston et al., 2018). To determine if the aligned arginine residues in the Samd14 CPB domain mediate CP complex binding we mutated the R43/R45 to alanine (Samd14-R43/45 A) in an expression construct ( Figure 6C). To test whether the Samd14-CPB domain function could be replaced by known CPI domains in other proteins, we substituted 17-amino acids in the Samd14-CPB domain with another known CPI motif in CKIP-1 (23.5% protein sequence identity, maintaining central arginine residues). We performed an IP Western blot of HA-tagged Samd14-R43/45 A and the Samd14-Ckip1 CPI proteins expressed in G1E-ΔEnh cells. GFP high and GFP low FACS sorted cells were prepared and analyzed separately to assess whether high/low expression levels of exogenous protein may alter binding affinity ( Figure 6D). Surprisingly, the Samd14-Ckip1 CPI chimeric protein failed to pull down any CP complex subunits (Capzβ, Capzα1, and Capzα2) ( Figure 6E). Whereas Samd14-R43/45 A still interacts with the CP complex, the mutation reduced the amount of CP complex components pulled down ( Figure 6F). Thus, Samd14 R43/45 residues mediate CPB affinity but are not required. Next, we tested whether expression of Samd14-Ckip1 CPI or Samd14-R43/45 A (which interfere with Samd14 binding to CP) impairs Samd14 function. PHZ-treated Samd14 ΔEnh/ΔEnh spleen progenitors expressing Samd14-Ckip1 CPI formed 2.23-fold and Flow cytometry gating to analyze infected (GFP + ) phenylhydrazine (PHZ)-treated cells expressing shControl or shCapzb2 based on CD71 low/med and CD71 high ; representative histograms depict pERK1/2 fluorescence in GFP + CD71 low/med Kit + cells after stem cell factor (SCF) stimulation (10 ng/ml, 5 min) and GFP + CD71 high cells after Epo stimulation (10 min, 5 U/ml). (C) Quantitation of pERK1/2 MFI after SCF (5 min, 10 ng/ml) or Epo (10 min, 5 U/ml) stimulation in GFP + CD71 + Kit + and GFP + CD71 high spleen cells expressing shControl or shCapzb2 (N=3). (D) CD71 low cells are SCF-responsive but Samd14-insensitive, CD71 med are SCF-responsive and Samd14sensitive, and CD71 high cells are unresponsive to SCF. CD71 low/med cells responding to SCF do not respond to Epo stimulation, whereas the CD71 high cells are Epo dependent and Samd14-sensitive. (E) Samd14-CP interaction promotes burst forming unit-erythroid (BFU-E) colony formation, SCF/Kit and Epo/EpoR signaling by phosphorylating ERK1/2 and AKT in PHZ-treated spleen cells, in the absence of the Samd14-CP interaction PHZ-treated spleen cells fails to promote SCF/Kit signaling, but Epo signaling remains unaffected. Error bars represent SD. *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001 (two-tailed unpaired Student's t test).
The online version of this article includes the following figure supplement(s) for figure 7: 1.5-fold fewer BFU-E and CFU-E than Samd14, respectively ( Figure 6G). However, cells expressing Samd14-Ckip1 CPI and Samd14-ΔCPB formed similar numbers of colonies ( Figure 6G). The Samd14-R43/45A-expressing cells (with reduced but not absent CPB) formed fewer BFU-E and CFU-E colonies compared to Samd14 but more than Samd14-ΔCPB ( Figure 6G). These results demonstrate the importance of the Samd14-CP interaction for Samd14-mediated stress erythroid precursor activity. Compared to other functionally validated CPB domains, the Samd14 CPB domain has additional sequence attributes that enable the Samd14-CP protein interaction.
Given that Samd14 promoted SCF signaling within a narrow window of CD71 expression, it is possible that Samd14 regulates other signaling pathways at distinct stages of erythropoiesis. Erythropoietin signaling regulates erythroid maturation and cell survival during stress as well as normal erythropoiesis (Liu et al., 2006;Xiang et al., 2015). To test whether Samd14 regulates Epo signaling, we serum starved and Epo stimulated cultures expressing either EV, Samd14, or Samd14-ΔCPB and assessed AKT activation. pAKT increased 2.28-fold (p=0.0002) in Samd14-expressing cells vs EV control ( Figure 7A). Similarly, pERK increased 5.85-fold in Samd14 expressing cells vs EV controls (Figure 7figure supplement 1). Surprisingly the Samd14-ΔCPB also increased the pAKT levels by 2.3-fold (p=0.0002). This contrasts with our results in SCF stimulated cells, suggesting that the Samd14-CP interaction drives Samd14-dependent Kit signaling but is dispensable for Epo signaling ( Figure 7A). To test the involvement of Capzb in cell signaling, we analyzed ERK1/2 activation levels upon SCF/Epo stimulation after Capzb knockdown. SCF responsive cells (CD71 low , CD71 med , Figure 5B) were analyzed together as GFP + CD71 low/med Kit + cells ( Figure 7B). Capzb knockdown decreased SCF-dependent pERK1/2 activation by 19.3-fold (p=0.001) relative to shControl ( Figure 7C) and reduced the overall magnitude (MFI) of SCF-mediated pERK1/2 activation (Figure 7-figure supplement 1B). Following Epo stimulation, the CD71 high fraction (which is unresponsive to SCF) increased ERK1/2 phosphorylation ( Figure 7C). Epo-stimulated pERK1/2 was unaffected by Capzb knockdown in CD71 high cells. A similar trend was seen using pAKT as a readout for Epo signaling (Figure 7-figure supplement  1C). These data define cell stage-specific requirements for Samd14 in both SCF/Kit and Epo/EpoR signaling pathways ( Figure 7D). Whereas Capzb expression and the CPB domain of Samd14 promote Kit signaling, the interaction between Samd14 and CP is not an important mediator of Epo signaling ( Figure 7E).

Discussion
To reveal mechanisms of anemia stress-activated protein Samd14, and its function in cell signaling, we analyzed Samd14-interacting proteins. We report that Samd14 interacted with a family of F-actin CPs (Capzβ, Capzα1, and Capzα2) in a SAM domain-independent manner. These subunits form the heterodimeric actin CP complex that binds the barbed end of actin during filament assembly/disassembly and plays a role in cell morphology, migration, membrane trafficking, and signaling (Canton et al., 2006;Terry et al., 2018). Mature erythrocytes contain an α 1 β 2 conformation of CP that is present in the cytosol and not associated with the short actin filaments in the erythrocyte membrane skeleton (Kuhlman and Fowler, 1997). While pointed end actin CPs like Tropomodulin (Tmod3) coordinate critical steps in late-stage erythropoiesis (Sui et al., 2014), the role of barbed end capping molecule CP in erythropoiesis was not previously investigated. We show here that Capzb expression opposed erythroid differentiation in fetal liver and stress erythroid precursors, regulated cell size, and promoted cell survival and precursor colony forming activity. The Samd14-CP interaction occurs through a CPI-like domain to mediate Samd14-dependent functions in signaling and progenitor activities. While our studies implicate CP as a mediator of erythropoiesis, the exact mechanism is unclear. Among the potential functions of CP, it is a core component of the dynactin complex which controls early endosome dynamics (Valetti et al., 1999). This represents one potential mechanism whereby CP may function in erythroid precursor cell signaling.
The rate of erythropoiesis can increase by 15-20-fold under stress conditions. Numerous proteins and factors are involved in mechanisms which accelerate, maintain, and ultimately resolve stress erythropoiesis in anemia (Bresnick et al., 2018;Paulson et al., 2020). Stress erythroid progenitors and precursors in spleen and liver have unique cellular behaviors and respond to distinct paracrine signals relative to bone marrow-derived erythroid progenitors (Harandi et al., 2010). Illustrating the requirements for Kit signaling in anemia, SCF deletion in spleen endothelial cells attenuates the anemia stress response and reduces the red blood cell counts without affecting bone marrow hematopoiesis (Inra et al., 2015). Additional stress-specific signaling molecules BMP-4, hedgehog glucocorticoid and GDF15 promote stress erythroid progenitor expansion and homing (Bauer et al., 1999;Hao et al., 2019;Lenox et al., 2005;Perry et al., 2009). Transcriptional mechanisms are also intimately involved in erythroid progenitor/precursor stress responses. The transcription factor GATA1 promotes stress erythropoiesis in addition to roles in developmental and physiological erythropoiesis (Gutiérrez et al., 2008). GATA1 and related transcription factor GATA2 directly activate Samd14 transcription through an intronic enhancer which is present in a network of other genes predicted to control vital aspects of stress erythropoiesis (Hewitt et al., 2017;Hewitt et al., 2015). GATA2 and Samd14, as well as anti-apoptotic genes such as Bcl2l1, are transcriptionally elevated in stress erythropoiesis via similar E-box-GATA composite element sequences (Hewitt et al., 2017). As one node of this network, our results in a stress progenitor/precursor genetic rescue system have provided rigorous evidence for a mechanism in which Samd14 upregulation drives stress progenitor activity through its CPI to coordinate stress-dependent signaling mechanisms.
As a facilitator of cell signaling that is upregulated in response to anemia, Samd14 represents a new constituent driving anemia-specific Kit activities. These conclusions are typified in the Samd14 ΔEnh/ ΔEnh mouse, in which the E-box-GATA composite element is selectively required for stress progenitor responses/activities and anemia-dependent Kit signaling. Similar decreases in stress progenitor activities were observed in other genetic models which attenuate Kit activation (e.g. the Kit Y567F mouse) (Agosti et al., 2009). To establish whether Samd14 mechanisms integrate in other pathways, we also describe a role for Samd14 in Epo signaling. Whereas Capzb knockdown lowered Epo signaling, there were no differences in the ability of full length Samd14 and Samd14-ΔCPB to activate Epo signaling. Thus, both Samd14 and CP appear to have complex-independent functions in erythroid precursor activity and cell signaling. Together, our findings suggest new models for regulation of cell signal transduction in stress erythropoiesis. We envision two important functions, not necessarily exclusive, that result from the interaction of Samd14 with CP through its CPB domain. First, CP may target Samd14, and therefore, specific signaling activities, at specific locations in the cell important for function, such as membrane-associated receptor complexes. Second, the Samd14-CP interaction may influence activities of the larger CP-containing dynactin complex in endocytic trafficking, impacting the rate of ligand-activated receptor turnover in erythroid cell membranes. It is attractive to propose that the temporary upregulation of Samd14 in acute anemia may be tied to a burst of erythropoiesis by controlling the rate of receptor turnover. The model system used in these studies examines stressactivated erythroid precursors after they have expanded. While this permitted the ex vivo analysis of stress BFU-E in culture systems, fewer BFU-E and CFU-E suggests that Samd14-CP mechanisms may play a role in earlier immature progenitor cells to oppose erythroid maturation. Importantly, while we used PHZ injection to induce acute anemia, Samd14 expression is also elevated in clinically relevant anemia models induced by severe bleeding and erythroid radioprotection, suggesting that Samd14 mechanisms are not specific to this model. These experiments have taken important steps toward elucidating Samd14 function in acute anemia, the mechanism whereby the Samd14-CP interaction controls erythroid precursor cell signaling, and how cell signaling networks are optimally coordinated in stress erythropoiesis to accelerate and resolve acute anemia.
Cell lysates were incubated with 5 µg of either anti-Samd14 antibody (Hewitt et al., 2017) or antirabbit IgG isotype (Invitrogen, #10500 C) control overnight. Protein A beads (ThermoFisher #15918-014) were blocked for 1 hr in 1%BSA/PBS prior to IP. Cell lysates were incubated with 60 µl of the preblocked Protein A beads for 2 hr at 4°C. Beads were washed with 1% NP-40 Lysis buffer, boiled in SDS lysis buffer and analyzed by SDS-PAGE.

Mass spectrometry
In-gel samples were reduced with 10 mM DTT and then alkylated with 10 mM iodoacetamide, washed and then digested overnight using trypsin. Tryptic digests were run by LC-MS/MS using a 2 hr gradient on a 0.075 mm × 250 mm Waters CSH C18 column on a U3000 nanoRSLC (Dionex) coupled to a Q-Exactive HF (ThermoScientific) mass spectrometer. Peptides were identified using Mascot (Matrix Science; v.2.6.1), searched with a fragment ion mass tolerance of 0.060 Da and a parent ion tolerance of 10.0 PPM. Deamidated asparagine and glutamine, oxidation of methionine and carbamidomethyl of cysteine, phospho serine, threonine and tyrosine were specified in Mascot as variable modifications. Results were then validated and summarized into Scaffold (Proteome Software; v. 4.8.9) using a 99.0% protein probability with a minimum of two unique peptides with at least 80% peptide probability. Statistically enriched proteins were determined using a 2-way ANOVA.

Quantitative PCR
Total RNA was purified from 0.5 to 2 × 10 6 cells in 1 ml TRIzol (ThermoFisher). For cDNA synthesis 1 µg of RNA was incubated with 200 ng oligo(dT) and 50 ng random hexamer primers at 68°C for 10 min. RNA was incubated with M-MLV reverse transcriptase (NEB) with 10 mM dithiothreitol (DTT), RNAsin (Promega), and 0.5 mM deoxynucleoside triphosphates at 42°C for 1 hr, and then heat inactivated at 98°C for 5 min. Real time PCR was performed with 2 µl cDNA, Power SYBR Green (ThermoFisher) and 200 nM of the appropriate primers on the Quant Studio 3 real time PCR system (Applied Biosystems). Relative expression level for each gene was quantified based on a standard curve of serially diluted control cDNA. Negative control reactions were performed on samples in the absence of reverse transcriptase.

Phospho-flow cytometry
After 2-day culture, primary spleen cells were serum-starved for 2 hr in 1% BSA/IMDM at 37°C and treated with 10 ng/ml SCF or vehicle for the indicated time. For Epo stimulation (5 U/ml), cells were positively selected by Biotin-Ter119 and MojoSort streptavidin-conjugated magnetic beads. Cells were immediately fixed in 2% paraformaldehyde for 10 min at 37°C and permeabilized in 95% methanol overnight at -20°C. Cells were stained with rabbit antibodies against phospho (S473)-AKT (p-AKT) and phospho (Thr202/Tyr204) p44/42 ERK1/2 (p-ERK) (4060, 9101; Cell Signaling) for 30 min, then incubated in APC-conjugated goat anti-rabbit (1:200), PE-Cy7-conjugated Kit (1:200) and PE-conjugated CD71 (1:200) for 30 min at room temperature. Samples were analyzed using a BD LSR II flow cytometer and gates were selected based on fluorescence minus one control. Values for pAKT and pERK levels were calculated by MFI using FlowJo v10.6.2 (BD Life Sciences) and normalized to the maximum overall value within each experiment (Relative MFI) or as fold change of activation over vehicle treated controls.

Ethics
This study was performed in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. All of the animals were handled according to approved institutional animal care and use committee (IACUC) protocols (#18-099-08 FC) of the University of Nebraska Medical Center.

Additional files
Supplementary files • Transparent reporting form • Supplementary file 1. Samd14-interacting proteins. Mass spectrometry analysis of IP protein in EV and Samd14 conditions (N=3). Spectral count of each protein, as determined by Scaffold. Statistical significance between EV and HA-Samd14 conditions determined by Scaffold.

Data availability
The mass spectrometry proteomics data was deposited to the ProteomeXchange Consortium via the PRIDE (Perez-Riverol et al, 2019) partner repository with the dataset identifier PXD030467 and 10.6019/PXD030467. All other data generated or analysed during this study are included in the manuscript and supporting file.
The following dataset was generated: