Introduction

During early embryonic development, organizers play an important role in patterning and directing the differentiation of surrounding cells into specific tissues and organs. The embryonic organizer establishes the developmental polarity in vertebrates and similarly, the mound/slug tip in Dictyostelium act as organizers (Rubin and Robertson, 1975) playing a pivotal role in guiding collective cell migration and patterning. Despite numerous investigations on Dictyostelium tip formation, the processes by which the tip establishes and maintains the primary developmental axis remains elusive. Gaining insights into the mechanisms regulating tip organizer function will offer a valuable understanding of the orchestrated cell movements and intricate processes underlying development.

Study of the social amoeba Dictyostelium discoideum has helped us to better understand many fundamental problems about development (Annesley and Fisher, 2009; Müller-Taubenberger et al., 2013; Dunn et al., 2018; Martín-González et al., 2021; Huber et al., 2022). Dictyostelium has demonstrated its value in a range of applications, including the investigation of human diseases ( Eichinger et al., 2005; Müller-Taubenberger et al., 2013; Storey et al., 2022). Dictyostelium are free living, amoeboid, soil protists and alternate their life cycle between unicellular and multi- cellular stages. The amoebae prey on bacteria and upon starvation, the cells aggregate gradually forming a tipped multicellular mound. The mound further develops to form a slug which undergoes differentiation and morphogenesis to form a fruiting body consisting of a dead stalk supporting a ball of spores (Raper, 1940; Kessin, 2001).

Diffusible molecules such as cAMP, adenosine, ammonia, and a chlorinated hexaphenone ‘differentiation-inducing factor’ (DIF) regulate multicellular development in Dictyostelium (Bloom and Kay, 1988; Williams, 1988; Gross, 1994; Mahadeo and Parent, 2006; Shu et al., 2006). The gradient of these morphogens along the slug axis determines the cell fate, either as prestalk (pst) or as prespore (psp) cells. cAMP regulates PKA activity, adenosine signaling, and morphogenetic cell movements to control tip development in Dictyostelium (Mann and Firtel, 1993; Siegert and Weijer, 1995). Adenosine, a by-product of cAMP hydrolysis, acts as an inhibitory morphogen suppressing additional tip formation (Schaap and Wang, 1986), however the mechanism involved in tip formation is not completely understood.

Adenosine deaminase (ADA) or its isoforms regulate adenosine homeostasis by catalysing the breakdown of adenosine to generate inosine and ammonia, and are conserved among bacteria, invertebrates, vertebrates including mammals (Cristalli et al., 2001). Two isoforms of ADA are known in humans including ADA1 and ADA2 and ADGF shares a structural similarity to ADA2. ADA2 is associated with the activation of lymphocytes that play a role in the immune response to infections (Marone et al., 1980). Absence of ADA2 activity results in the build-up of deoxyadenosine and its phosphorylated metabolite, dATP that is toxic to lymphoid precursors (Notarangelo, 2016). Consequently, complete ADA2 deficiency is characterized by lymphopenia and severe combined immuno-deficiency disorder (SCID) (Gaspar, 2010). The loss-of-function mutations in adgf/ada2 in humans also manifest in a variety of vascular and inflammatory symptoms, including early-onset of stroke and systemic vasculopathy (Zhou et al., 2014).

ada2 disruption in mice leads to liver dysfunction and perinatal mortality (Wakamiya et al., 1995) and transgenic overexpression of ADA2 in mice results in aberrant heart and kidney development (Riazi et al., 2005). Absence of adgf/ada2 in frogs manifests in reduced body size with altered polarity (Iijima et al., 2008). In Drosophila, six ADGFs have been identified and are reported to promote cell proliferation by depleting extracellular adenosine (Zurovec et al., 2002). Loss of function of adgfA in Drosophila promotes melanotic tumour formation and larval death (Dolezal et al., 2005). In Arabidopsis thaliana, tRNA adenosine deaminase 3 (tad3) is necessary for telomere length homeostasis (Bose et al., 2020). Adenosine deaminases are known to interact with dipeptidyl peptidase IV (DPP), CD26 (cluster of differentiation 26) expressed on T-cells and the adenosine receptor, A2AR (expressed on dendritic cells) to facilitate cell-cell signaling (Moreno et al., 2018). Thus, adgf plays a pivotal role in the regulation of cell proliferation and development in several organisms.

Four isoforms of ADA are annotated in the Dictyostelium discoideum genome (Eichinger et al., 2005) including adenosine deaminase (ada; DDB_G0287371), adenosine deaminase acting on tRNA-1 (DDB_G0278943), adenosine deaminase tRNA-specific (DDB_G0288099) and adenosine deaminase-related growth factor (adgf; DDB_G0275179), and their role in growth and development is not known. In this study, we examined the role of adgf during multicellular development in D. discoideum and demonstrate that ammonia derived from adenosine deamination is directly involved in tip organizer development.

Results

Differential regulation of adgf expression during growth and development

To verify the blasticidin (bsr) insertion in adgf ^-, a diagnostic PCR was carried out and the integration was validated (Figures 1A and 1B), and qRT-PCR analysis confirmed the absence of adgf expression (Figure 1C). In vertebrates, adenosine deaminases are expressed in a tissue specific manner to control growth and development. To know if adgf expression is differentially regulated during development in Dictyostelium, qRT-PCR was performed using RNA obtained from different stages of development. The time point of analysis includes 0 h, 8 h, 12 h, 16 h, 20 h and 24 h. adgf expression peaks at 16 h (Figure 1D) implying an important role for adgf later in development. At this time point, the expression of the other three isoforms of ADA were not significantly different from WT, suggesting that they do not compensate for the loss of adgf (Figure 1E).

Figure 1.

Dictyostelium ADGF shares a strong structural similarity with human ADA2

The D. discoideum adgf gene is predicted to encode a protein of 543 long amino acids and belongs to the metallo-transferase superfamily (Figure 2A), containing a ADA domain and an N- terminal domain similar to human ADA2 (Figure 2B). Using the ADGF sequence from the protein family database (https://www.ebi.ac.uk/interpro/), an analysis was carried out using the online tools, SMART (Simple Modular Architecture Research Tool: http://smart.embl-heidelberg.de), to know the presence of structurally similar domains. This study suggests that the protein domains of Dictyostelium ADGF are conserved and show high degree of similarity with human ada2.

Figure 2.

D. discoideum ADGF (DdADGF) shares 37.47% identity with human ADA2. ADGF is present in other social amoeba, sharing a sequence similarity of 59.5% with Dictyostelium fasciculatum, 61.2% with Dictyostelium pupureum, 53.6% with Dictyostelium lacteum and 62.1% with Polysphondylium pallidum (www.uniprot.org). Multiple sequence alignment reveals the presence of an N-terminal signal sequence characteristic of extracellular proteins (Figure 2C) and conserved histidine and glutamine residues in the active site of both D. discoideum ADGF and human ADA2 (Figure 2D).

The phylogenetic relation of ADGF to the classic ADA subfamily has been reported previously (Maier et al., 2005). To determine the evolutionary relationship of D. discoideum ADGF with that of other organisms, a phylogenetic analysis was carried out and the alignment was created using MEGAX software (Kumar et al., 2018). The evolutionary history of ADGF was inferred using a Maximum Likelihood approach with Bootstrap analysis (100 iterations) as described by Felsenstein (1985). The resulting phylogenetic tree indicated that DdADGF is closely related to the ADGF proteins of other Dictyostelid members, including D. purpureum, Heterostelium pallidum, and Cavenderia fasciculata. D. discoideum ADGF forms a distinct clade, likely representing a distant relative of its vertebrate homolog (Figure 2E). Additionally, the structure of D. discoideum ADGF was predicted by homology modeling with AlphaFold (https://alphafold.ebi.ac.uk/), using the crystal structure of human ADA2, CECR1 (Cat eye syndrome critical region protein 1) as a template (PDB-3LGD) (Figures 2F-H). Alignment of Dictyostelium ADGF with human ADA2 yielded an RMSD value of 1.003 Å, indicating strong structural similarity (Eidhammer et al., 2000; Koehl, 2001).

adgf controls aggregate size in Dictyostelium

To understand the role of adgf during D. discoideum development, mutant lines with lesions in the gene adgf were plated at a density of 5x10⁵ cells/cm² on non-nutrient phosphate buffered agar plates (pH 6.4) and monitored thereafter. In comparison to WT (37.6 ± 2.7 mm²), the aggregates of adgf⁻ were larger in size (62.6 ± 5.4 mm²; p value = 0.0003), and thus the number of aggregates formed by the mutants were fewer (39.67 ± 3.703) than the WT (67.00 ± 2.543) (Figure 3A). To determine the pathways that are affected impairing the tissue size in the mutant, RNA expression of countin (ctn) and small aggregates (smlA) were examined and their levels were found to be reduced significantly compared to controls (Figure 3B). Countin factor (CF) regulates group size by reducing the expression of cell-cell adhesion proteins cadherin (cadA) and contact site A (csaA) (Siu et al., 1985). The expression of cadA, csaA and cell-cell adhesion were also quantified and the mutants displayed enhanced cadA and csaA expression and higher cell-to-cell contacts than the WT (Figures 3C and 3D) suggesting that adgf regulates the overall size of the aggregates.

Figure 3.

cAMP chemotaxis significantly influences cell-cell adhesion (Konijn et al., 1967) and to know if chemotaxis is impaired in the mutant, an under-agarose chemotaxis assay was carried out. The chemotactic activity was not significantly different between the two cell types (Figure 3E), implying that the increased mound size in the mutant may not be due to altered chemotactic activity.

adgf mutants form large, tipless mounds

Subsequent to plating on KK2 agar plates, WT cells formed mounds by 8-9 h, and culminate to form fruiting bodies by the end of 24 h. In contrast, the adgf⁻ lines were blocked as rotating mounds with no tips till 30 h (Figure 3F; Supplementary videos S1 and S2). The mound arrest phenotype could be mimicked by adding the ADA specific inhibitor deoxycoformycin (DCF) to WT cells (Figure 3G). After 36 h, a fraction of adgf⁻ mounds (34.77 ± 4%) formed fruiting bodies with bulkier spore sac and a stalk (Figures 3H and 3I). This late recovery from the mound arrest may be due to the expression of other ADA isoforms after 36 h.

Reduced ADA activity leads to high adenosine levels in the adgf mutant

If the function of adgf is compromised, ADA enzyme activity is expected to be low. To verify this, total ADA activity in WT and adgf⁻ mounds was measured using a commercial kit. This technique relies on the conversion of inosine (by ADA) to uric acid and was measured at A293 nm. Although ADA activity was reduced significantly in the mutant (Figure 4A), it was not completely abolished and possibly, other isoforms of ADA may have a basal activity at 16 h.

Figure 4.

ADGF quenches extracellular adenosine and if this process is impaired as expected in adgf⁻, the mutants will have increased extracellular adenosine. Using a commercial kit, total adenosine levels were measured and the adenosine levels were elevated both at 12 h and 16 h in the mutant, (Figure 4B) and this difference was highly significant at 16 h (377.6 ± 18.86 nM in the WT; 490.8 ± 7.241 nM in the mutant, p = 0.0002). adgf expression, at this time point was also high in WT cells. The elevated adenosine levels at 16 h may be due to reduced ADA activity and enhanced expression of genes such as 5’ nucleotidase (5’nt), phosphodiesterases (pdsA and regA) that are involved in adenosine formation. Hence, their expression levels were examined. Although the expression of both 5’nt and regA were enhanced at 8 h, pdsA levels remain unaltered. Interestingly, the expression of all three genes trended lower at 12 h but were significantly upregulated at 16 h (Figure 4C) suggesting an important role of these genes at a specific time in development.

Addition of ADA or overexpression of adgf cDNA restored tip development in adgf⁻

adgf mutants carry significantly high levels of adenosine. By administering ADA enzyme on top of the adgf⁻ mounds (Figure 5A), excess adenosine can be quenched resulting in ammonia formation, possibly rescuing the phenotype. Indeed, addition of 10 U ADA onto mutant mounds restored tip development. Besides, overexpression of WT adgf cDNA (driven by actin15 promoter) rescues the developmental defects of the adgf mutants (Figures 5B and 5C), confirming that the developmental block is the result of adgf gene disruption alone. However, adgf cDNA overexpression in WT cells does not result in any observable defects (Figure 5D). Taken together, these findings support an important function of adgf in tip development.

Figure 5.

WT or its conditioned media (CM) restored tip formation in adgf⁻ mounds

In order to determine whether adgf is necessary for tip formation in a cell-autonomous manner, WT and the mutant cells were reconstituted in different proportions and plated. In a mix of 50% WT: 50% mutant, WT cells were able to rescue the defects of adgf⁻ cells but with 20% WT and 80% adgf⁻, the rescue was partial (31 ± 4 %) (Figure 5E) suggesting that the mound arrest phenotype of the mutant is due to the absence of some secreted factor(s).

Further, when developed in the presence of WT CM, adgf⁻ cells formed tipped mounds and eventually fruiting bodies (Figure 5F). Conversely, in the presence of CM from adgf⁻ cells, WT developed as large mounds with no tips (Figure 5G) and the size was comparable to the mounds formed by adgf⁻. These findings imply that the developmental phenotype of adgf⁻ is not due to a cell autonomous defect but due to faulty secreted factor signaling.

Volatile ammonia rescued the mound arrest phenotype of adgf⁻

If ADGF function is compromised, ammonia levels are expected to be low. Hence, the concentration of ammonia from WT and adgf⁻ mounds were measured using a commercial kit. At the mound stage, total ammonia levels were significantly reduced in adgf⁻ lines (Figure 6A). By mixing sodium hydroxide and ammonium chloride (Thadani et al., 1977), ammonia could be generated, and in such conditions, tip formation was restored in adgf⁻ mounds (Figure 6B).

Figure 6.

Physically separated WT restored tip development in the mutant

WT mounds are expected to release lot of volatiles including ammonia, possibly rescuing the mound arrest phenotype of the mutants nearby. To verify if this is correct, adgf⁻ cells were developed in one half of a compartmentalized Petri dish and WT cells on the other side. Interestingly, the mound arrest of adgf⁻ was completely rescued by 3.5 hours (Figure 6C) in the presence of WT although they were physically separated from each other. This suggests that volatiles, likely to be ammonia released from WT is sufficient enough to rescue the mutant phenotype.

The CM from adgf⁻ is expected to have high adenosine, possibly generating low or no ammonia. Addition of ADA to the CM of the mutant in one compartment of the partitioned dish led to a partial rescue (57 % ± 2) of adgf⁻ kept on the other side of the dish after 3.5 h (Figure 6D) implying that the ammonia generated in such conditions may not be enough for a full rescue.

Adenosine deamination alone drives the rescue of adgf⁻

To know if adenosine deamination is exclusively responsible for the rescue of the mutant, 10 ml cold KK2 buffer mixed with different concentrations of adenosine (10 µM, 0.1 mM, 1 mM) was added on one side of the compartmentalized dish and the other side of the dish had the mutant on phosphate buffered agar arrested at the mound stage. After 3.5 h of the addition of ADA enzyme (10 U) to the buffered solution containing adenosine, a full rescue was observed (Figure 6E), strongly indicating that volatile ammonia generated from adenosine deamination alone is rescuing the defect. Ammonia is known to be produced during protein/RNA catabolism in Dictyostelium (White and Sussman, 1961; Hames and Ashworth, 1974; Schindler and Sussman, 1977; Walsh and Wright, 1978) but the total protein and RNA levels were not significantly different between WT and adgf mutants (Figure S1 A-B) suggesting that ammonia released from adenosine deamination plays a crucial role in tip formation.

Caffeine, not ammonia rescued the mound size of adgf ^-

Given that adenosine levels were elevated in the mutants, we attempted to rescue the large mound size observed in the adgf mutants by treating them with the adenosine antagonist, caffeine. Caffeine rescued this early developmental defect in a dose-dependent manner (Figure 6F). This suggests that adgf may be one of the regulators of group size in Dictyostelium. Since the ammonia levels were lower in the mutant, we tested whether exposure to ammonia could rescue the large aggregate size of the adgf mutant. Exposing the mutants to 0.01 M ammonia soon after plating or six hours after plating, had no effect on the aggregate size of the mutants (Figure 6G), suggesting that while some early effects may be mediated through adenosine receptors, the later effects appear to be independent and likely influenced by ammonia.

Faulty expression of cAMP relay genes in adgf⁻

cAMP signaling is crucial for tip development and determining cell fate in Dictyostelium (Schaap and Wang, 1986; Saxe et al., 1993; Firtel, 1996; Singer et al., 2019). Hence, the expression of genes involved in cAMP relay were measured in WT and adgf⁻ cells by qRT-PCR. Total cAMP levels and acaA gene expression were both low in the adgf⁻lines (Figures 7A and 7B). Treating the cells with the pkA activator 8-Br-cAMP or cyclic-di-GMP, the activator of adenyl cyclase (Wang and Schaap, 1985; Chen and Schaap, 2012), reversed the adgf⁻ mound arrest phenotype (Figures 7C and D). Increasing the cyclic-di-GMP dose from 0.5 mM to 1 mM resulted in the formation of multiple tips (Figure 7E). These observations further reinforce that the mound arrest phenotype of the adgf⁻ lines is due to faulty cAMP signaling. Blocking the phosphodiesterase (pde4) activity could also lead to higher cAMP levels transiently. When treated with 3-Isobutyl-1-methylxanthine (IBMX), a known PDE4 inhibitor (Bader et al., 2007; Siegert and Weijer, 1989), there was no effect on tip formation in adgf⁻mounds (Figure 7F), although caffeine restored tip formation (Figure 7G). Thus, in spite of impaired expression of cAMP relay genes, the chemotaxis activity is not altered in the mutant.

Figure 7.

Circular instead of spiral cAMP waves in adgf⁻ mounds

Low acaA expression, reduced cAMP levels and enhanced cell-cell contacts in adgf⁻ is likely to impair cAMP wave propagation and to ascertain if this is true, the cAMP signal propagation in WT and adgf⁻ mounds were compared using dark field optics. In contrast to the WT, which displayed a spiral wave propagating center, the adgf mutant exhibited a circular wave propagating center throughout, suggesting that the cAMP relay is impaired (Figure 7H; Supplementary videos S3 and S4).

Ammonia restores tip formation by regulating acaA/pde4 expression

adgf expression peaks at 16 h and to find if adgf expression is regulated by the substrate or the product, WT cells treated with different concentrations of adenosine (100 nM, 1 µM, 0.5 mM), were plated on KK2 agar plates and harvested at 16 h for RNA isolation. Independently, WT mounds on KK2 agar plates were exposed to different concentrations of volatile ammonia (0.1 mM, 1 mM, 10 mM) for 3 h and thereafter, the mRNA expression levels of adgf, acaA and pde4 were examined. All the rescue experiments involving ammonia is for 3.5 h time period. With 100 nM adenosine, the expression of both adgf and acaA decreased when compared to controls but at higher adenosine concentrations (1 µM, 0.5 mM), the expression levels of these two genes were comparable to controls (Figure 8A). Interestingly, pde4 expression decreased gradually with increasing adenosine concentrations (100 nM to 0.5 mM), but increased steadily with increasing ammonia levels.

Fig 8.

In mounds exposed to 0.1 mM ammonia, the expression of adgf decreased but exposure to 1 mM, and 10 mM ammonia, respectively caused a 3 and 2.3-fold upregulation of acaA expression. This observation suggests that ammonia may be rescuing the mound arrest phenotype by enhancing acaA expression and thus cAMP levels. In similar conditions, a considerable increase in pde4 expression was observed (Figure 8B), which may be necessary for controlling cAMP production in response to NH3 treatment. In conclusion, the expression of adgf is influenced both by the substrate adenosine as well as the product ammonia in a dose dependent manner.

adgf ^- cells sort to prestalk in a chimera with WT

To determine if adgf favours the differentiation of one cell type over the other, the expression of prestalk (pst) and prespore (psp) markers were examined in the mutant by realtime PCR. While the expression of pst markers, ecmA and ecmB were significantly upregulated, the psp-specific pspA gene showed reduced expression in the mutant (Figure 8C), suggesting that WT adgf favours psp expression.

To know the cell type preference of the adgf mutants in a chimera with WT, a fraction (20%) of any one cell type was labelled with a cell tracker stain DIL and the development was tracked thereafter. In the slugs that formed after reconstituting labelled WT or adgf ^OE cells with 80% unlabelled adgf⁻ cells, the fluorescence was largely confined in the psp region. Conversely, when adgf⁻ cells stained with DIL were reconstituted with unlabelled WT or adgf ^OE, the fluorescence was restricted to the pst part of the slug. In the reconstituted slugs, consistently, labelled WT or adgf ^OE cells occupy the psp, whereas the adgf⁻ cells end up in the pst region (Figures 9A, 9B and 9C). Thus, the distribution of cell types in the mound/slug is significantly influenced by adgf, as the mutant cells sort out in a mixture with WT cells and differentiate to pst cells.

Figure 9.

Further, when stained with the prestalk marker, neutral red (Yamamoto and Takeuchi, 1983), the staining was intense in the pst and ALCs of adgf⁻ slugs, and such an intense coloration was not apparent in WT especially in the ALCs (Figure S2A). To further investigate if reduced ADA activity affects cell type specific marker expression or their pattern in slugs, pst (ecmA-GFP, ecmO-GFP) or psp (pspA-RFP) lines were treated with the ADA inhibitor DCF. While there was no change in pst/psp patterns in the mounds and slugs that formed after a long delay, the ecmA- GFP was intense in the tip when compared to WT (Figures S2B). Similar to NR staining pattern in adgf mutant slugs, a prominent expression of ecmA-GFP was observed in the ALC region also. Visually, there was no change in ecmO or psp marker expression in DCF treated slugs (Figures S2C and S2D) suggesting that impaired ADGF activity affects tip specific expression significantly.

adgf acts downstream of the histidine kinase, dhkD

In an effort to identify the pathway by which adgf acts during tip development, we selected a number of mutant lines with similar mound arrest phenotype such as cAMP receptor B (carB ^-), LIM domain containing protein (limB ^-), mound mutants (mndA ^-) and histidine kinase (dhkD ^-) (Saxe et al., 1993; Carrin et al., 1996; Chien et al. 2000; Singleton and Xiong, 2013) and treated the mounds with ADA enzyme. Treatment with 10 U ADA or exposure to ammonia restored tip formation in dhkD⁻ mounds and not others tested (Figures S3A and S3C). An increase in ADA concentration (20 U), resulted in the development of multiple tips in dhkD^– (Figures 9D, 9E and 9F). These findings imply that adgf acts downstream of dhkD in controlling tip development. Similarly, we tested few mutants with multiple tips phenotype, such as tipped mutant (tipA ^-), culinB (culB ^-), autophagy mutants (atg7 ^-, atg8^-, and atg9^-) (Stege et al., 1999; Wang and Kuspa, 2002; Otto et al., 2003; Otto et al., 2004, Tung et al., 2010) by adding the ADA inhibitor, DCF (1 mM) to the cell suspension /agar plates and if rescued, those mutants are likely to be in the same pathway controlling tip development. However, DCF treatment had no impact on mutants with multiple tips (Figure S3B).

Impaired expression of other deaminases also results in aberrant tip formation

Several pathways control ammonia levels during development and knockouts in other deaminases (2-aminomuconate deaminase: DDB_G0275081, adenosine monophosphate deaminase: DDB_G0292266, dCTP deaminase: DDB_G0293580, threonine deaminase: DDB_G0277245, N-acetyl glucosamine deaminase: DDB_G0286195, glucosamine-6-phosphate deaminase: DDB_G0278873) show a partial mound arrest phenotype (Figures S4A-F) although the expression of some of these candidates is stronger during development suggesting a prominent and unique role of adenosine deamination in tip development.

High adenosine levels and other related purines do not block tip development

ADA catalyses the conversion of adenosine/deoxy adenosine respectively to inosine/ deoxy inosine. As adgf mutants have high adenosine levels, we investigated if the mound arrest could be mimicked in WT by treating with adenosine. Surprisingly, addition of adenosine does not lead to mound arrest in WT (Figure S5A). It is possible that ADA/ADGF converted adenosine analogues to inosine analogues, which in some manner affected development. Hence, WT cells were treated with an adenosine analogue (2’-deoxyadenosine), or guanosine (10 µM each) and they do not cause a mound arrest phenotype either. Similarly, treating adgf⁻ mounds with inosine does not restore tip formation (Figures S5B and S5C) suggesting that the blocked tip development is due to faulty adenosine deamination alone.

Cross kingdom rescue of the adgf ^- mound arrest phenotype

Just like WT Dictyostelum rescuing the mound arrest phenotype of the mutant, we examined if bacteria physically separated from the mutants would rescue the phenotype. To check this, Klebsiella aerogenes and adgf mutants were incubated adjacent to each other within a compartmentalized KK2 agar plate. After 12 h, tip formation was restored in the mutants while in the same time frame, the mounds in controls failed to form tips. Possibly, Klebsiella on KK2 with no nutrients would die releasing ammonia restoring tip development in Dictyostelium (Figure S6).

Discussion

Dictyostelium ADGF is likely to be secreted growth factor

Multiple sequence alignments, experiments with conditioned media and cell type reconstitution with WT suggests that ADGF is a secreted molecule. AMP deaminase, previously characterized in the mollusk Helix pomatia, has been identified as a member of ADGF family (Tzertzinis et al., 2023). It is important to note that ADGF is also known to act on 5′-adenosine monophosphate (AMP) and deoxyadenosine. Studies on the characterization and expression of this growth factor has shown that it is a secreted protein critical for embryonic and larval development in Pacific abalone (Hanif et al., 2022), highlighting the broader role of ADGF in development across species. The human ADA2 enzyme belongs to the novel family of ADGF and crystal structure of ADA2 reveal the presence of a catalytic- and two unique ADA2-specific domains with novel folds, responsible for protein dimerization and interaction with cell surface receptors. Furthermore, the presence of a number of N-glycosylation sites, conserved disulfide bond, and a signal peptide in ADA2 indicate that ADA2 is specifically adapted to function in the extracellular environment (Zavialov et al., 2010).

Possible reasons for increased mound size in the adgf mutant

Enhanced cell adhesion influences cohesion and can impact the mound size (Roisin-Bouffay et al., 2000). The cell-cell adhesion genes cadA and csaA (Coates and Harwood, 2001) were upregulated in the mutant, leading to enhanced adhesion and larger mound size. While over-secretion of ctn leads to stream breaking resulting in smaller aggregate formation (Brock and Gomer, 1999), disruption of the ctn gene prevents this process, inducing large aggregate formation. ctn in turn regulates smlA (Brock et al., 2003), impacting the overall aggregate size. Thus, adgf mutants have reduced ctn and smlA expression manifesting large mound formation. Indeed, cells treated with adenosine are known to form large aggregates (Schaap and Wang, 1986) and the first genetic evidence supporting the previous work is from adgf mutants, which carry excess extracellular adenosine and form large aggregation streams.

Pathways generating ammonia in Dictyostelium

Proteolysis during starvation is believed to be the main source of volatile ammonia in Dictyostelium (Hames and Ashworth, 1974; Schindler and Sussman, 1977; White and Sussman, 1961), while RNA degradation is also attributed to yield ammonia during starvation (Walsh and Wright, 1978). During development, amino acids, total protein and RNA levels are reported to reduce with a significant increase in ammonia levels, thus equivocating the source of volatile ammonia (Hames and Ashworth, 1974). Furthermore, several deaminases or enzymatic reactions in Dictyostelium may also potentially generate ammonia (Supplementary Table S1). The annotated D. discoideum genome contains five evolutionarily conserved family of ammonium transporter/methylammonium permease/rhesus protein (Amt/Mep/Rh) encoding genes (Eichinger et al., 2005), amtA, amtB, amtC, rhgA, and rhgB, which help in regulating ammonia levels during growth and development. amtA and amtC antagonistically control developmental processes and are involved in ammonium sensing or transport (Follstaedt et al., 2003; Kirsten et al., 2005; Singleton et al., 2006).

Role of ammonia during Dictyostelium development

During early development in Dictyostelium, ammonia is known to inhibit aggregation (Schindler and Sussman, 1979; Williams et al., 1984), affect aggregate territory size (Thadani et al., 1977), aggregate density (Feit, 1988), cell fate (Gross et al., 1988), culmination (Davies et al., 1993), and fruiting body size (Lonski, 1976). Ammonia is essential for differentiation in submerged clumps (Sternfeld and David, 1979) and favours prolonged slug migration called ‘slugging’ over culmination and influences psp over pst differentiation (Newell et al., 1969). Ammonia plays a crucial role in orienting cell masses and accelerating the movement of aggregating cells (Bonner et al., 1986). Enzymatic removal of ammonia causes the quick transition from slug to fruiting thus controlling the morphogenetic pathways (Schindler and Sussman, 1977). By suppressing DIF biosynthesis (Neave et al., 1983), ammonia favors psp over pst development in Dictyostelium (Davies et al., 1993; Riley and Barclay, 1990). Further, ammonia prevents the developmental transition of slugs to fruiting bodies (Bradbury and Gross, 1989; Wang and Schaap, 1989). It is interesting to note that ammonia induces tip development (as in adgf mutants), but subsequently, tips move away from a source of ammonia. Migration of slugs/ fruiting bodies away from ammonia, reflects a process by which fruiting bodies position themselves from other structures to increase the possibility of spore dispersal (Kosugi and Inouye, 1989). Ammonia by enhancing the collective movement of cells is believed to exert negative chemotaxis (Bonner et al., 1986). Ammonia is known to inhibit tip formation in slugger mutants, but in strains such as NC4 (Gee et al., 1994), ammonia exerts no effect on tip formation as in AX4 (data not shown). The highly acidic, autophagic vesicles in pst cells (Gross, 2009) are believed to catalyze the breakdown of proteins and RNA, also generating ammonia.

ADGF and its substrate, adenosine could serve as a localized source of ammonia. Indeed, cells carrying high ATP, its derivatives cAMP (Bargoda et al., 2009; Singer et al., 2019) and adenosine (Schaap and Wang, 1986), end up in the tip (Hiraoka et al., 2022). Further, 5’- nucleotidase promoter activity is high in pstAB cells (Ubeidat et al., 2002) that significantly covers the tip region suggesting high adenosine levels in the tip. However, extracellular 5’-AMP can be derived from multiple sources such as hydrolysis of cAMP, ATP and RNA (Carpousis et al., 1999).

As a gas, ammonia can diffuse generating a gradient; high in the slug front compared to the posterior. Our results (Fig. 6A) also show that the amount of ammonia released from adenosine is in the same order of magnitude as that from other sources (Yoshino et al., 2007). In adgf mutants, ammonia levels may not be sufficient enough to neutralize the acidic vesicles and hence, NR staining is intense in the pst region and the ALCs (Bonner, 1952). Thus, excess acidification if not neutralized may lead to mound arrest and if ammonia acts on the acidic vesicles, then, collective migration of cells leading to tip formation may be favoured.

Ammonia is crucial in regulating differentiation, metabolism, and gene expression (Liu et al., 2022, Stein et al., 2013). Ammonium has been shown to induce aberrant blastocyst differentiation, disrupt normal metabolic processes, alter pH regulation, and subsequently affect fetal development in mice (Lane and Gardner, 2003). Newt Triturus exposed to ammonia show dorsalization of the ventral marginal zone highlighting the capacity of ammonia to alter embryonic axis formation (Yamada, 1950). Furthermore, research on avian embryos has revealed significant ammonia content in developing eggs, suggesting a potential role for ammonia in the energy metabolism during ontogenesis (Needham, 1926).

Ammonia’s effect on cAMP signaling in Dictyostelium

Ammonia is known to increase intracellular cAMP levels in D. discoideum (Riley and Barclay, 1990; Feit et al., 2001) but Schindler and Sussman, (1977) and Williams et al., (1984) reported that high ammonia levels inhibit the synthesis and release of cAMP. However, some of these experiments use ammonium carbonate (Schindler and Sussman, 1977), or ammonium chloride (Williams et al., 1984), as a source of ammonia and the possibility of carbonate and chloride ions interfering with development cannot be ruled out. In slugs, ammonia is shown to exert very little effect on cAMP levels (Schaap et al., 1995). In Dictyostelium mucoroides, high ammonia levels are known to block the production of extracellular cAMP. The effect of volatile ammonia on aggregation seems to be species specific notably in P. violaceum, P. pallidum, and D. mucoroides resulting in wider aggregation territories, while it was not observed in other species such as D. discoideum and D. purpureum (Bonner and Hoffman, 1963).

Adenylate cyclase (AC) activity is regulated by various factors (Steer, 1975). At physiological levels, ammonium ions increase AC activity in the rat brain by 40% (Yeung et al., 1989), but lowered its activity in the liver by about 30% (Wiechetek et al., 1979). Ammonia has been shown to increase the activity of AC-G in Dictyostelium sori (Cotter et al., 1999). High ammonia levels by altering the pH, can affect the activity of numerous enzymes whose activity is pH dependent and thus the activity of AC can be impacted by pH variations. adgf mutants with low ammonia levels have reduced cAMP levels and a gradual increase in ammonia, cause a steady but significant increase in acaA expression. In other systems (Rat brain), ammonia is known to interact with manganese (Rivera-Mancía et al., 2012; Lu et al., 2020) and this divalent cation is known to increase acaA expression in Dictyostelium (Loomis et al., 1979; Khachatrian et al., 1987). Possibly, ammonia interacts with metal ions like manganese to form ‘ammine complexes’ (Lipkowski and Galus, 1973), particularly in aqueous environments or under specific conditions where the appropriate ligands are available and enhances cAMP levels thus rescuing the mound defects of the mutant.

The decision between the formation of the tip (pst cells) and the ALCs are controlled by the tip’s production of ammonia, which prevents the migration of ALCs towards the tip (Sternfeld and David, 1982; Feit et al., 1990). Consistent with this observation, our results show that adgf mutant lines that have low ammonia levels exhibit enhanced ecmA and ecmB marker expression. Mutants impaired in DIF-1 synthesis fail to develop pst-O cells in the slug and subsequently also fail to form the basal disc in the fruiting body (Thompson and Kay, 2000). Similar to our observations of cells with high adenosine occupying the tip of the slug, cells with high ATP (that degrades to adenosine) levels, also end up in the centre of the tip (Hiraoka et al., 2022).

Tip organizer development in Dictyostelium depends on the differentiation and appropriate sorting of pst and psp cells (Kay et al., 1978; Williams et al., 1989; Saxe et al., 1993; Williams, 2006), a process that relies on signaling of different morphogens including cAMP, adenosine, DIF and ammonia (Bloom and Kay, Williams, 1988; 1988; Riley and Barclay, 1990; Gross et al., 1994; 2006). These morphogens modify one another’s effects, and determine the choice of the differentiation pathway as well as the spatial arrangement of cells (Bloom and Kay, 1988). Thus, the rescue of the adgf mutant upon exposure to ammonia is likely due to cAMP signaling and cell-cell contact.

Possible reasons for reduced cAMP levels in the adgf mutant

The increased expression of PDE’s (intra and extracellular) and reduced ACA expression could well explain the lower cAMP levels in the mutant. ADA significantly regulates adenosine levels, thus reducing the activation of adenosine-mediated receptors (Van Haastert, 1983). Two classes of adenosine receptors including adenosine alpha- and beta-receptors are known to be expressed in Dictyostelium (Theibert and Devreotes, 1984). When adenosine concentrations are high, beta- receptors bound with adenosine inhibit the binding of cAMP to its receptors, thereby inhibiting cAMP signaling (Newell, 1982; Van Haastert, 1983; Theibert and Devreotes, 1984). High adenosine levels in the adgf mutant may also reduce cAMP levels via a similar mechanism.

adgf activity is likely to be different between pst and psp cells as reconstitution with different cell types shows that cells with low adenosine and high ammonia levels end up in psp while cells with excess adenosine and low ammonia attain pst cell fate. Cell death also results in adenosine/ammonia formation and thus adgf activity is likely to be higher in pst than psp to maintain low extracellular adenosine. The mound/slug tip is believed to carry high adenosine levels restricting additional tip formation favouring lateral inhibition (Wang and Schaap, 1985) and our results also show that adgf⁻ cells with high adenosine goes to pst than the psp region.

Distorted cAMP waves in mutant

Collective cell movement within mounds is essential for cell sorting and the progression from mounds to finger structures (Kellerman and McNally, 1999). In WT mounds, cAMP waves propagate in a spiral pattern, while in mutants, waves propagate in a concentric circular pattern. Initially, cAMP waves appear as concentric circles that, upon symmetry breaking, form spiral waves (Siegert and Weijer, 1995). With successive wave propagation, the circular ring distorts, and one end curls toward the pulsatile center, creating a spiral wave around the organizing center. In adgf mutants, however, defective cAMP relay prevents this transition, likely due to lower cAMP levels. A surge in phosphodiesterase inhibition is thought to regulate this shift from concentric to spiral wave propagation (Palsson and Cox, 1996).

The mound/slug tip of Dictyostelium generates cAMP pulses (Traynor et al., 1992), and is known to suppress additional tip formation (Farnsworth, 1973). Adenosine, that inhibits tip formation represses pde4 expression whereas ammonia, that promotes tip elongation, increases pde4 expression. Thus, the restoration of cAMP signaling and spiral wave propagation possibly leads to the rescue of the adgf ^- mound arrest phenotype. Surprisingly adenosine exerts little effect on tip development in WT cells and studies of Inouye (1989) also reinforce our observations, showing that adenosine does not significantly influence the conversion of pst to psp cells in shaking culture conditions.

Adenosine deamination drives tip organizer development

A crucial evidence supporting that adenosine deamination is singularly responsible for the rescue of mound arrest comes from experiments where partitioned dish with buffered solution containing adenosine on one side and the mutants on the other half of the dish. Addition of ADA enzyme (10 U) to the buffered solution with adenosine fully restored the tipped mound formation in the other half of the dish. Although ammonia’s role in Dictyostelium development is well established, this report is the first to show a novel role of ammonia in tip development. If natural variants of Dictyostelium with no tips exist in soil, there is a likelihood that they will form fruiting bodies when WT Dictyostelids are nearby. It is important to note that development of an organism in its natural habitat is strongly influenced by volatiles nearby (Schulz-Bohm et al., 2017).

ADA in organizer development and gastrulation in vertebrates

ada/adgf expression is found to be significantly high during gastrulation stages (Figure S7) in several vertebrates (Pijuan-Sala et al., 2019; Tyser et al., 2021) and the process of gastrulation in higher organisms and collective cell movement within the Dictyostelium mound are remarkably similar (Weijer, 2009), suggesting an overlooked role of ammonia in organizer development. It is likely that ADA plays a conserved, fundamental role in regulating morphogenesis in Dictyostelium and other organisms including vertebrates. The human homologue of adgf, CECR1, is a potential gene for the genetic disorder, Cat-eye syndrome and thus, Dictyostelium may also serve as a model to study this condition.

dhkD functions upstream of adgf

The histidine kinase dhkD, a member of the two-component family of histidine kinases, appears to function upstream of adgf, adding to our understanding of the signaling cascade in Dictyostelium. Histidine kinases play essential roles in regulating developmental transitions, with dhkC, for example, initiating the late developmental program that ultimately leads to fruiting body formation (Singleton et al., 1998). This function of dhkC is regulated in part by ammonia levels, as low ammonia concentrations have been shown to inhibit the dhkC phospho-relay, thereby influencing developmental outcomes (Kirsten et al., 2005). In Dictyostelium, several histidine kinases, including dhkA, dhkB, dhkC, and dokA, coordinate cAMP signaling to modulate the activity of PKA, a cAMP-dependent protein kinase that is essential for proper development (Anjard and Loomis, 2003). These findings suggest the interdependencies within the signaling network in regulating multicellular development in Dictyostelium.

Model of ADGF action

High extracellular cAMP at the slug front (Singer et al., 2019) is expected to generate high adenosine, which however could arise from other sources and thus ammonia levels are expected to be higher in the slug front compared to the back. Possibly, ammonia generated extracellularly is quenched rapidly by the intracellular acidic compartments of pst cells, increasing its pH and favouring collective cell movement driving mound/ slug tip development. It is important to note that ammonia is known to promote psp than pst differentiation. Although ammonia is likely to be formed from several sources in Dictyostelium, a critical threshold concentration of ammonia may be necessary for tip development. The failure of even one pathway may result in a drop-in ammonia levels, exerting an effect on development (Figure 10). Overexpression of ecmA and downregulation of pspA (as measured by realtime PCR) in the mutant suggests that there is a change in cell type differentiation as suggested by the sorting experiments. This could suggest that the differentiation of the cells is blocked at an early prestalk cell fate. Thus, ammonia is likely to be important for the maturation of tip cells.

Fig 10.

Our study shows that adgf acts downstream of dhkD and in silico studies have shown that dhkD interacts with adgf (Figure S8A). Therefore, it is likely that in the presence of ammonia, dhkD activates the phosphorelay, in turn increasing intracellular cAMP levels, leading to tip formation. Histidine kinase dhkC is known to phosphorylate regA in Dictyostelium (Thomason et al.,1999). However, based on docking we found no direct interaction of regA with ADGF (data not shown). Thus, the integration of cytokinin and histidine kinase signaling ensures coordinated multicellular organization in Dictyostelium (Aoki et al., 2020). Previous work in other systems have found ADA to be interacting with CD 26 (dpp IV) (Tanaka, 1993) and docking DdADGF with AprA (dpp8) (Figure S8B) also support the observations. However, adding ADA inhibitor to aprA mutant (dpp8), show no effect on the phenotype (data not shown).

Organisms from different kingdoms often interact in complex ways such as symbiosis, predation, competition etc., which can have significant effects on ecosystem dynamics (Helgason et al., 1998; Harrison, 2005). Such interactions are often mediated through volatiles affecting the physiology of each other just like the bacterial volatiles affecting the development of Dictyostelium. However, it needs to be ascertained if this particular volatile is ammonia.

Supplementary Results

Genes involved in autophagy and tipped mound formation do not show altered expression in the mutant

Autophagy helps maintain cellular homeostasis by promoting the breakdown of damaged proteins and organelles (Mizushima et al., 2008), and ammonia is recognised to be a diffusible regulator of autophagy in human cells (Eng et al., 2010). To investigate if autophagy is impaired in the mutant, the expression of autophagy markers atg8 and atg18 were examined and there was no discernible change in the expression levels when compared to WT (Figure S9A). The tipless to tipped mound transition as well as late developmental gene expression is regulated by gbfA, a G-box binding factor (GBF). However, in the adgf mutant the gbfA expression levels were comparable with no significant difference between the WT at 16 h (Figure S9B). Without the assay for autophagy as well as assay for GBF activity, these expression studies alone are not conclusive to rule out the role of these factors in tip formation.

ADA does not deaminate cytokinin

ADA is known to interact with 5’AMP (Hanif et al., 2022) and in Dictyostelium, 5′-AMP serves as a direct precursor of cytokinin, playing a significant role in cellular signaling and development (Taya et al., 1978). Although docking studies suggest that binding of ADGF with cytokinins (zeatin, dihydrozeatin, isopentenyl adenine) is moderate, (data not shown), we carried a functional assay by mixing cytokinin and ADA with KK2 buffer in one side of the Petridish and found no rescue of the mutant (Figure S10) in such conditions suggesting no functional activity between ADA and cytokinin.

Supplementary discussion

Plausible routes of caffeine action rescuing the mound arrest phenotype

The adgf mound arrest could be rescued by the addition of the adenosine antagonist, caffeine but treatment with the pde4 inhibitor, IBMX failed to rescue the mound arrest. Caffeine is known to reduce adenosine levels in blood plasma (Conlay et al., 1988) and also increase ammonium levels in urine samples of rabbits (Bernheim and Bernheim, 1945). Thus, the mound rescue upon caffeine treatment may be a result of reduced adenosine and increased ammonia levels. With respect to caffeine action on cAMP levels, the reports are contradictory. Caffeine has been reported to increase adenylate cyclase expression thereby increasing cAMP levels (Hagmann, 1986) whereas Alvarez-Curto et al., (2007) found that caffeine reduced intracellular cAMP levels in Dictyostelium. Although, caffeine is moderately potent in inhibiting PDE enzyme activity, the in vivo concentrations are likely to be low to be associated with effective PDE inhibition (Burg and Werner, 1975; Daly, 1993).

Material and Methods Bioinformatic analyses of ADGF

The genomic sequence and the protein sequence of ADGF were obtained either from dictybase (http://dictybase.org) or NCBI database (https://www.ncbi.nlm.nih.gov/). Within the Dictyostelium ADGF, the ADA domain was identified by SMART and BLAST (http://smart.embl-heidelberg.de/) (https://blast.ncbi.nlm.nih.gov/Blast.cgi) (Altschul et al., 1990; Letunic and Bork, 2018) analyses. A phylogenetic tree was generated by aligning multiple amino acid sequences of ADGF from several taxa. Neighbour Joining approach and the MUSCLE alignment tool of the MEGAX programme were used for constructing the tree (Saitou and Nei, 1987). The tertiary structure of ADGF was obtained using the online programme alphafold (https://alphafold.ebi.ac.uk/) (Jumper et al., 2021), and PyMol was used for viewing the images. The expression profiles of ada and ada2 during mouse and human gastrula development were retrieved from the marionilab.cruk.cam.ac.uk/MouseGastrulation2018/ and human-gastrula.net/ databases (Pijuan-Sala et al., 2019; Tyser et al., 2021), respectively. Protein-protein docking was carried out using High Ambiguity Driven Protein-Protein Docking (HADOCK) version 2.4.

Culture and development of Dictyostelium discoideum

D. discoideum (WT-AX4) cells or the mutant adgf derived from AX4, (DBS0237637) were cultured in modified maltose-HL5 medium (Formedium, UK) with 100,000 U/L penicillin and 0.1 g/L streptomycin. Three independent insertional mutants (GWDI_17_D_7, GWDI_47_C_1, GWDI_132_H_3) in the adgf gene (insertion in exon 2 in all three mutants) were obtained from the GWDI bank (https://remi-seq.org), in Dictyostelium stock centre, North Western University, USA (Gruenheit et al., 2021). The mound defects were identical in all three and GWDI_47_C_1 alone was characterised further. The culture was raised as a monolayer in Petri plates or grown in an Erlenmeyer flask in shaking conditions at 150 rpm and 22 °C, to a cell density of 2-4x10⁶ cells/ml. The cells were also grown on SM/5 agar plates supplemented with Klebsiella aerogenes at 22 °C (2 g/L glucose, 2 g/L protease peptone, 0.4 g/L yeast extract, 1 g/L MgSO4.H2O, 0.66 g/L K2HPO4, 2.225 g/L KH2PO4, 1% Bactoagar, pH 6.4). For developmental assays, freshly starved cells were washed twice with ice cold KK2 buffer (2.25 g KH2PO4 and 0.67 g K2HPO4 per litre, pH 6.4), and plated on 1% non-nutrient KK2 agar plates at a density of 5x10⁵ cells/cm² (Nassir et al., 2019). Thereafter, the plates were incubated in dark conditions at 22 °C for development.

Dictyostelium genomic DNA isolation

WT and adgf ^- cells grown axenically were harvested, and the pellet was resuspended in 1 ml of lysis buffer (50 mM Tris-Cl, pH 8; 10 mM EDTA; 0.8% SDS). To this mixture, 200 µl of Nonidet P-40 (NP40), a non-ionic detergent, was added, vortexed and centrifuged at 12,000 g for 15 min at room temperature (RT). The resultant pellet was gently vortexed, resuspended in 500 µl of lysis solution containing 200 µg/ml Proteinase K, and incubated at 65 °C for 30 min. 300 µl of phenol: chloroform was added to this suspension, centrifuged at 18,000 g for 10 min at RT, and the resultant aqueous phase was extracted carefully. An equal amount of chloroform was added and centrifuged at 18,000 g for 10 min at RT. Following this, 750 µl of pure ethanol was added to the suspension, which was then centrifuged at 12,000 g for 15 min at 4 °C. Genomic DNA was precipitated by adding twice the volume of absolute ethanol and 1/10^th the volume of 3 M sodium acetate. After a 10 min centrifugation at 15,000 g, the pellet was cleaned using 70% ethanol and stored at 4 °C. Electrophoresis was conducted in TAE buffer at 50 V using a Medox power pack system (India), and the integrity of DNA was confirmed on a 1% agarose gel.

Validation of the adgf mutant

To validate the blasticidin (bsr) cassette insertion in the adgf mutant, WT and adgf⁻ genomic DNA were isolated, and a diagnostic PCR was performed using the gene and bsr insert specific primers in accordance with the guidelines provided in the GWDI website. A qRT-PCR was carried out to confirm the absence of adgf expression in the mutant cells. The primers used for mutant validation are listed in Supplementary Table S2.

Quantitative real-time PCR (qRT-PCR)

Total RNA was extracted from WT and adgf⁻ cells using Trizol reagent (Favorgen, USA) at specified intervals (every four hours from 0 to 24 h). cDNA was synthesized from the total RNA using a PrimeScript 1st Strand cDNA Synthesis Kit (Takara, Japan). Random primers from the manufacturer were used to generate the cDNA from a template of 1 µg total RNA. qRT-PCR was performed using SYBR Green Master Mix (Thermo Scientific, USA) and 1 µl of cDNA. The expression levels of adgf, acaA, cAMP receptor A (carA), phosphodiesterases (pdsA, regA), 5’ nucleotidase (5’nt), extracellular matrix A (ecmA), extracellular matrix B (ecmB), prespore A (pspA), countin (ctn), and small aggregate (smlA) were quantified with a QuantStudio Flex 7 (Applied Biosystems, USA). The mitochondrial large RNA subunit (rnlA) served as an internal control. qRT-PCR data analysis was conducted according to the method described by Schmittgen and Livak (2008). The primer sequences used for qRT-PCR are provided in Supplementary Table S4.

Generation of adgf over expression construct

The full-length 1.7 kb adgf sequence was PCR-amplified using ExTaq polymerase (Takara, Japan) with WT genomic DNA as the template. The amplified product was then ligated into the pDXA-GFP2 vector at the HindIII and KpnI restriction sites. Both adgf ^- and WT cells were electroporated with this vector, and G418-resistant clones (10 μg/ml) were isolated for further analysis. The expression of adgf was confirmed by semi-quantitative PCR. The corresponding primer sequences are listed in Supplementary Table S3.

Transformation of Dictyostelium discoideum

WT and adgf ^- cells grown axenically were harvested, washed twice with ice-cold EP buffer (10 mM KH2PO4, 10 mM K2HPO4, 50 mM sucrose, pH 6.2), and resuspended in 100 µl of EP++ solution (10 K2HPO4 mM, 10 mM KH2PO4, 50 mM sucrose, 1 mM MgSO4, 1 mM NaHCO3, 1 mM ATP, 1 µM CaCl2) containing 10 µg of plasmid vector (Nassir et al., 2019) in pre-cooled cuvettes (BioRad, USA). The cells were electroporated using a BTX ECM830 electroporator (Harvard Apparatus, USA) at 300 V with 2 ms pulses and five square wave pulses at 5-second intervals. The cells were then transferred to a Petri dish with 10 ml of HL5 medium and incubated at 22 °C. After 24 hours, G418 (10 µg/ml) was added to the medium, and resistant colonies were selected for further analysis.

Preparation of conditioned media (CM)

WT and adgf⁻ cells grown in HL5 medium were collected at the mid-log (ML) phase, resuspended in KK2 buffer at a density of 1x10⁷ cells/ml, and incubated at 22 °C with shaking for 20 hours. The clarified supernatant obtained after centrifugation was used for the experiments.

Assay for ADA activity

The total ADA activity from Dictyostelium was determined as per the protocol of the manufacturer (Abcam, USA; Cat No: ab204695). For sample preparation, 5x10⁵ cells/cm² were seeded onto KK2 agar plates and at the mound stage, ice cold ADA assay buffer was flooded, then vigorously pipetted to disrupt the mound integrity. The cell homogenate was agitated on a rotary shaker at 4 °C for 15 min and then centrifuged at 12,000 rpm for 10 min in a cold microfuge tube. The supernatant was subjected to the ADA assay. This ADA test detects inosine, generated from adenosine break down through a multi-step reaction. The intermediate formed combines with the probe to produce uric acid, which is quantified at OD 293 nm. BCA (Bicinchonic acid) assay kit (Thermoscientific, USA) was used for determining the protein concentration. One unit of ADA activity is defined as the amount of enzyme that hydrolyses adenosine to yield 1 µmol of inosine per min under the assay conditions.

Adenosine quantification

Total adenosine levels from Dictyostelium mounds were measured according to the protocol using the Adenosine Assay Kit (Abcam, USA; Cat No: ab211094). Cells grown in HL5 were collected, washed and plated on KK2 agar plates. The lysis buffer was added to the plates with mounds and mixed thoroughly. 50 μl of the lysate was mixed with 2 U of ADA and was subjected to incubation for 15 min at RT. Adenosine quantification involves the use of ADA and after a series of enzymatic reactions, an intermediate is formed which reacts with the adenosine probe generating a fluorescent product. Using a spectrofluorometer (Perkin Emer, USA; λEx = 544 nm/λEm = 590), the fluorescence intensity can be measured which is proportional to the concentration of adenosine (Perkin Emer, USA; λEx = 544 nm/λEm = 590). The adenosine levels were quantified using the adenosine standard curve.

Quantification of ammonia

Ammonia assay kit (Sigma-Aldrich, USA; Cat No: AA0100) was used for estimating the total ammonia levels. WT and adgf ^- cells developed on KK2 agar plates were sealed with parafilm and incubated at 22 °C. The mounds were collected using lysis solution, and the debris was removed by centrifugation at 10,000 g for 10 min. The supernatant was used for further analysis. For the ammonia assay, 1 ml of assay reagent was mixed thoroughly with either 100 µl of samples or standards, incubated for 5 min at RT, and the absorbance was measured at 340 nm. Then, 10 µl of L-glutamate dehydrogenase (GDH) solution was added to each cuvette, and after a 5-min incubation at 25 °C, the absorbance was measured again at 340 nm using a spectrophotometer (Eppendorf, Germany). In the presence of GDH, ammonia combines with α- ketoglutaric acid and reduced NADPH to produce L-glutamate and oxidised NADP+. The decrease in absorbance at 340 nm is proportional to the ammonia concentration. The ammonia standard curve was used to calculate the ammonia levels.

Volatile ammonia generation

To generate ammonia, 1 ml of 1 N NaOH and 1 ml of NH4Cl (concentrations used 0.1 mM, 1 mM, 10 mM) were mixed thoroughly (Thadani et al., 1977; Feit et al., 1990) and from this mix, 2 ml was aliquoted in the upper half of the Petri dish. The other half of the plate with the adgf⁻ mounds on KK2 agar were inverted, sealed and incubated at 22 °C. To determine whether WT mounds, physically separated from the mutants could rescue the mound arrest, WT (1x10⁶ cells/cm²) and adgf⁻ (5x10⁵ cells/cm²) cells were developed on KK2 agar plates on either side of compartmentalised and sealed Petri plates. adgf⁻ cells developed on either side of the dish served as controls.

Quantification of cAMP

Using the cAMP-XP test kit and following the manufacturer’s instructions, total cAMP levels were determined from both the WT and the mutant (Cell signaling, USA). WT and adgf⁻ mounds were disrupted and collected in 1 ml ice cold KK2 buffer. After centrifuging the pellet, 100 μl of 1X lysis solution was added, and the mixture was allowed to rest on ice for 10 min. Following that, 50 μl of lysate and 50 μl of HRP-conjugated cAMP solution were added to the test plates, which were then shaken horizontally at RT. After a 3-hour incubation, the wells were emptied and washed three times with 200 μl of 1X wash buffer. 100 μl of stop solution was added to halt the reaction, and the absorbance was measured at 450 nm using a spectrophotometer (Biorad, USA). The cAMP standard curve was used to determine the cAMP levels.

Visualisation of cAMP waves using dark field optics

5x10⁵ cells/cm² were developed on KK2 agar plates under moist, dark conditions to monitor the propagation of the cAMP wave. Using a Nikon DS-5MC camera mounted on a Nikon SMZ- 1000 stereo zoom microscope, a time-lapse video of the aggregation was captured in real-time (Nikon, Japan). cAMP optical density waves were displayed by subtracting the image pairs (Siegert and Weijer, 1995) after processing the images with the NIS-Elements D program or ImageJ (NIH, USA).

Under agarose cAMP chemotaxis assay

The under-agarose cAMP chemotaxis assay (Woznica and Knecht, 2006; Singh and Insall, 2022) was carried out with cells obtained from WT and adgf⁻ mounds. Briefly, the cells were starved in KK2 buffer at 1x10⁷ cells/ml density. Three parallel troughs, each measuring 2 mm in width, were set up on a Petri dish containing 1% agarose. In the outer two troughs, 100 µl of WT and adgf⁻ cell suspension was aliquoted respectively and 10 μM cAMP was added to the central trough. Cell migration toward cAMP was recorded every 30 s over a total duration of 20 min using the NIS-Elements D software and an inverted Nikon Eclipse TE2000 microscope (Nikon, Japan). 35 cells were tracked each time and subsequently analysed using ImageJ (NIH, USA).

Reconstitution of WT with mutant cells

WT and adgf⁻ cells cultured in HL5 medium, were harvested at the ML phase by centrifugation, rinsed twice with ice-cold KK2 buffer. The cells mixed in different ratios (1:9, 2:8 and 1:1) were adjusted to a final density of 5x10⁵ cells/cm², plated on 1% KK2 agar plates, and incubated at 22 °C for development.

Tracking cell fate after reconstitution

Dictyostelium cells harvested from fresh HL5 media were resuspended at a density of 1x10⁶ cells/ml in KK2 buffer, followed by incubation at 22°C for 1.5 h in shaking conditions with 0.2 µM DIL, (1,1’-Dioctadecyl-3,3,3’,3’-Tetramethylindocarbocyanine Perchlorate, Invitrogen, USA) a cell tracker dye. DIL has been used extensively as a cell tracker in C. elegans, Xenopus and mice (Schultz and Gumienny, 2012; Xu et al., 2020; Erdogan et al., 2016). WT or adgf⁻ cells labelled with DIL were reconstituted with the unlabelled cells in a ratio of 1:9, 2:8 and 1:1 and plated for development. In the ratios mentioned, the smaller fraction represents the stained cells. DIL is a lipophilic dye that selectively stains the plasma membranes of living cells, allowing visualization and analysis of cell morphology.

Neutral red staining of mounds and slugs

WT and adgf⁻ cells harvested from HL5 cultures were resuspended at a density of 1x10⁷ cells per ml in KK2 buffer and treated with 0.005% neutral red (NR) solution for 15 min at RT in shaking conditions (Bonner, 1952). The stained cells were washed twice with KK2 buffer, plated in buffered agar plates and thereafter, NR stained slugs were observed using an upright microscope.

Cell-cell adhesion assay

After 14 h of development on phosphate buffered agar plates, WT and adgf⁻ mounds were disaggregated by repeated pipetting and vortexing using ice-cold KK2 buffer. Dissociated cells were resuspended in KK2 buffer and incubated at 150 rpm for 45 min. Single and non-adherent cells were counted using a Neubauer chamber (Lam et al., 1981).

Treatment of Dictyostelium cells with different compounds

A highly specific ADA inhibitor, DCF that inhibits both extra and intra-cellular ADAs (Cha et al., 1975) was mixed with the WT cell suspension and plated on 1% non- nutrient KK2 agar plates at a density of 5x10⁵ cells/cm². Of the different DCF concentrations tested (10 nM, 50 nM, 100 nM and 1 µM), 100 nM showed complete tip inhibition.

For the enzymatic rescue assay, 10 U of bovine ADA dissolved in KK2 buffer (Sigma-Aldrich, USA) was added onto mutant mounds. The concentrations of 8-Br-cAMP (2 mM), cAMP (0.1 mM, 0.5 mM), c-di-GMP (0.1 mM, 0.5 mM), adenosine (1 µM, 10 µM, 100 µM, 1 mM), caffeine (10 nM, 100 nM, 1 µM) and IBMX (3-isobutyl-1-methylxanthine 0.5 mM) were based on previous publications (Chen et al., 2017; Chen and Schaap, 2012; Nassir et al., 2019; Siegert and Weijer, 1989) and these compounds were added independently either on top of the mounds or supplemented while plating. All the fine chemicals were from Sigma-Aldrich, USA except when mentioned.

Microscopy

Microscopy was performed using a Nikon SMZ-1000 stereo zoom microscope with epifluorescence optics, Nikon 80i Eclipse upright microscope, or a Nikon Eclipse TE2000 inverted microscope, connected with a digital sight DS-5MC camera (Nikon). Image processing was carried out using NIS-Elements D (Nikon) or Image J.

Statistical tools

Data analyses were carried out using Microsoft Excel (2016). Statistical significance was determined by paired or unpaired, one-tailed Student’s t-test (GraphPad Prism, version 6).

Acknowledgements

We gratefully acknowledge the help of NBRP Nenkin, Japan and Dictyostelium Stock Center (DSC), Northwestern University, USA for providing various strains and plasmids used in this study. We thank Prof. Richard Gomer, Texas A&M, USA for providing aprA mutant strain. We thank Prof. Kei Inouye, Kyoto University, Japan and Prof. Kees Weijer, University of Dundee, UK for their comments on an earlier version of the manuscript.

Additional information

Author contributions

P.H. and R.B. conceptualized the work. P.H. performed all experiments. P.H. and R.B. analysed the data and wrote the manuscript.

Additional files

Supplementary file

Supplementary videos