From biosynthesis and beyond—Loss or overexpression of the cytokinin synthesis gene, iptA, alters cytokinesis and mitochondrial and amino acid metabolism in Dictyostelium discoideum

Cytokinins (CKs) are a class of growth‐promoting signaling molecules that affect multiple cellular and developmental processes. These phytohormones are well studied in plants, but their presence continues to be uncovered in organisms spanning all kingdoms, which poses new questions about their roles and functions outside of plant systems. Cytokinin production can be initiated by one of two different biosynthetic enzymes, adenylate isopentenyltransfases (IPTs) or tRNA isopentenyltransferases (tRNA‐IPTs). In this study, the social amoeba, Dictyostelium discoideum, was used to study the role of CKs by generating deletion and overexpression strains of its single adenylate‐IPT gene, iptA. The life cycle of D. discoideum is unique and possesses both single‐ and multicellular stages. Vegetative amoebae grow and divide while food resources are plentiful, and multicellular development is initiated upon starvation, which includes distinct life cycle stages. CKs are produced in D. discoideum throughout its life cycle and their functions have been well studied during the later stages of multicellular development of D. discoideum. To investigate potential expanded roles of CKs, this study focused on vegetative growth and early developmental stages. We found that iptA‐deficiency results in cytokinesis defects, and both iptA‐deficiency and overexpression results in dysregulated tricarboxylic acid (TCA) cycle and amino acid metabolism, as well as increased levels of adenosine monophosphate (AMP). Collectively, these findings extend our understanding of CK function in amoebae, indicating that iptA loss and overexpression alter biological processes during vegetative growth that are distinct from those reported during later development.


| INTRODUCTION
Cytokinins (CKs) are most well known as phytohormones and are essential plant growth regulators. 1 CKs act as signaling molecules coordinating all aspects of growth and development, and they enable plants to sense and respond to environmental stimuli.Therefore, CKs provide the means to communicate at both the cellular and whole-plant levels with widespread roles in cell division, differentiation, vascular and gametophyte development, senescence, nutrient sensing, and responses to both biotic and abiotic stresses. 1Impactful roles of CKs are not exclusive to plants-in fact, several have been documented in CK-producing organisms spanning various kingdoms.For example, CKs control the cell cycle in apicomplexan parasites, fungi, and pathogenic nematodes, [2][3][4] virulence of bacteria and fungi, [5][6][7][8][9][10] spore germination in amoebae and fungi, [11][12][13] and subcellular protein trafficking in fungi. 4wing to the pleiotropic nature of CKs, much remains to be understood regarding the mechanisms by which CKs influence cellular and developmental processes in organisms other than plants.
The production of CKs is initiated by a rate-limiting step involving one of two biosynthetic enzymes, adenylate isopentenyltransferases (IPTs) or tRNA isopentenyltransferases (tRNA-IPTs). 14,15Disruption of IPT genes (including tRNA-IPTs) in both plant-and nonplant organisms results in drastic reductions in CK levels, with a variety of phenotypes, including decreased shoot growth and increased root growth in Arabidopsis thaliana, 16 decreased colony growth in Physcomitrella patens, 17 decreased virulence in Claviceps purpurea, 8 decreased drought tolerance in Triticum aestivum (wheat), 18 and spore viability defects in the amoeba, Dictyostelium discoideum. 11In plants, studying the roles of individual IPT genes can prove difficult owing to gene redundancy from the multiple wholegenome duplications that have occurred throughout evolution. 19For instance, there are 24 putative IPT genes in Triticum aestivum (wheat) (18 IPT, 6 tRNA-IPT), 18 9 in A. thaliana (7 IPT and 2 tRNA-IPTs), 14,15 and 10 in Oryza sativa (rice) (8 IPT and 2 tRNA-IPTs). 20By contrast, the haploid genome of the CK-producing social amoeba, D. discoideum, is not plagued by such complexities.Therefore, we exploited the genetic tractability of D. discoideum to expand our understanding of CKs outside of plant systems by investigating the role of its only adenylate-IPT gene, iptA, through gene knockout and overexpression.
D. discoideum is a motile, soil-dwelling amoeba that can be studied at the unicellular and multicellular levels due to its unique life cycle.Starvation triggers asexual multicellular development, where tens of thousands of cells merge together to form distinct morphological structures over a 24-h life cycle. 21It is thus straightforward to study a variety of different cellular and developmental processes throughout its life cycle, which has contributed to its usefulness as a model organism and has expanded our understanding of various biological processes. 22. discoideum is a member of the Amoebozoa phylum that diverged shortly after the line leading from plants to opisthokonts (animals and fungi) and thus has both plant-and animal-like characteristics. 23Ks have been established as key signaling molecules in D. discoideum necessary for spore production during the later stages of multicellular development. 11,24isruption of iptA reduces CK levels by 90% with visible morphological phenotypes in the final fruiting body life cycle stage. 11Previously, we profiled CKs throughout the life cycle stages of D. discoideum and found six different development is initiated upon starvation, which includes distinct life cycle stages.
CKs are produced in D. discoideum throughout its life cycle and their functions have been well studied during the later stages of multicellular development of D. discoideum.To investigate potential expanded roles of CKs, this study focused on vegetative growth and early developmental stages.We found that iptA-deficiency results in cytokinesis defects, and both iptA-deficiency and overexpression results in dysregulated tricarboxylic acid (TCA) cycle and amino acid metabolism, as well as increased levels of adenosine monophosphate (AMP).
Collectively, these findings extend our understanding of CK function in amoebae, indicating that iptA loss and overexpression alter biological processes during vegetative growth that are distinct from those reported during later development.

K E Y W O R D S
adenylate isopentenyltransferase, cytokinin, Dictyostelium discoideum, IptA, metabolomics, mitochondrial function, social amoeba, tricarboxylic acid cycle CK forms were produced in varying amounts. 13The highest CK levels were detected in the fruiting body stage, but considerable CK levels were also detected during vegetative growth and early development (aggregation).Moreover, there was a profile shift in the dominant CK forms between the vegetative growth and aggregation life cycle stages compared to those found in fruiting bodies. 13hese findings prompted us to investigate what roles, if any, CKs may have in regulating vegetative growth and early development in D. discoideum.Here, we generated an iptA knockout mutant through CRISPR/Cas9 mediated genome editing and investigated the effects of iptAdeficiency on D. discoideum proliferation, pinocytosis, and cytokinesis.We then observed the ultrastructure of the mutant and wild-type (WT) strains through transmission electron microscopy (TEM) and noted aberrant mitochondrial morphology, which was not anticipated and has not previously been established with iptA or CKs to our knowledge.To investigate the potential role of iptA in mitochondrial regulation, we generated a GFP-IptA overexpression strain in addition to the iptA-deficient strain to assess the metabolic consequences from altered CK dynamics using a metabolomics approach focusing on mitochondrial metabolites.Collectively, these results expand our understanding of CKs to include new roles for IptA in regulating D. discoideum vegetative growth through mitochondrial functions that are different from those previously established during the later stages of multicellular development.

| Culture conditions and chemicals
Strains were thawed from frozen stocks onto SM/2 agar (10 g/L glucose, 10 g/L protease peptone, 1 g/L yeast extract, 1 g/L MgSO 4 •7H 2 O, 1.9 g/L KH 2 PO 4 , and 0.6 g/L K 2 HPO 4 ) with Klebsiella aerogenes (KA) and incubated at room temperature (22°C). 25Several different types of growth media were used: HL5 containing glucose, Lo-Flo, FM minimal medium, FM minimal medium without arginine and lysine, and FM minimal medium without amino acids, which were all purchased from Formedium (Hunstanton, Norfolk, United Kingdom) (Figure 1).Cells were grown axenically in growth media (e.g., HL5-containing glucose) supplemented with 100 μg/mL ampicillin and 300 μg/mL streptomycin on an orbital shaker at 150 rpm.G418 was used at a concentration of 20 μg/mL for knockout and overexpression strains, as appropriate.For all experiments, cells were harvested during the mid-log phase of growth (2-5 × 10 6 cells/mL) and washed twice in cold KK2 buffer (2.2 g/L KH 2 PO 4 , 0.7 g/L K 2 HPO 4 ).All primers or single-guide RNAs were ordered from Integrated DNA Technologies (Coralville, Iowa, USA).

| Strains and plasmids
All experiments were conducted with the D. discoideum parental strain, AX3, which is denoted as either AX3 or wild type (WT) throughout the paper.The D. discoideum iptA-mutant strain was previously determined to have a spore viability defect 11 ; therefore, all strains (WT, mutant, and overexpression strains) were thawed fresh prior to the start of each experiment from frozen stocks onto SM/2 agar with KA for consistency and spores were never used maintaining cultures.Once plaques appeared, and before multicellular structures developed, cells were scraped from the agar plates and cultured axenically in HL5 medium.We were unable to obtain the iptA − strain previously used by Anjard and Loomis, 11 so we generated a new knockout strain using CRISPR-/ Cas9-mediated gene editing.This mutant was used to assess the effects of iptA-deficiency in vegetative amoebae considering it is the key enzyme responsible for CK biosynthesis in D. discoideum.
To generate the iptA knockout strain, hereafter referred to as iptA − , we used the pTM1285 all-in-one sgRNA CRISPR/Cas9 expression vector as previously described. 26he pTM1285 vector was provided by the University of Tsukuba through the National BioResource Project (NBRP) of MEXT Japan.Candidate single guide-RNA (sgRNA) sequences for the iptA gene were designed using CRISPOR and ordered containing overhangs specific to BpiI (Table S1A). 27Single sgRNAs were cloned into the pTM1285 vector using BpiI (New England Biolabs, Whitby, ON, Canada). 26Ligation of the sgRNA was confirmed through PCR with the sense oligo sgRNA and the tracrRV primer (Table S1B).Sanger sequencing was conducted prior to transformation into D. discoideum with the tracrRV primer (Table S1B,C).
Validated pTM1285-sgRNA constructs were transformed into AX3 cells via electroporation using D. discoideum protocols. 28,29Cells in the mid-log phase of growth were pelleted at 500 × g for 5 min and washed twice in ice-cold H50 buffer (20 mM HEPES, 50 mM KCl, 10 mM NaCl, 1 mM Mg 2 SO 4 , 5 mM NaHCO 3 , 1 mM Na 2 HPO 4 ).The cells were resuspended in a total volume of 100 μL ice-cold H50 buffer containing 6 μg of validated pTM1285-sgRNA construct.A 0.1 cm cuvette was prechilled prior to use, and the cell suspension was added to the cuvette, incubated on ice for 5 min and then electroporated using a MicroPulser electroporator (Bio-Rad Laboratories Canada, Mississauga, ON, Canada) with the Dic settings according to the Bio-Rad instruction manual (1 kV, two pulses, 1 ms time constant). 28The cuvette was placed immediately back into ice and kept for 15 min after adding a solution of CaCl 2 and MgCl 2 to a final concentration of 1 mM for each respective compound.The cells were then plated in 100 mm × 15 mm Petri dishes with 10 mL HL5-containing ampicillin (100 μg/mL) streptomycin (300 μg/mL) and kept at room temperature for 8 h.G418 was added to each Petri dish at a final concentration of 20 μg/mL, and the plate was incubated for 4 days.Following the incubation, the remaining cells were counted, washed, serially diluted, and plated (65-150 cells) onto SM/2 agar plates with KA.Individual clones (plaques) were isolated from the SM plates and transferred to 12-well plates containing HL5 with only ampicillin and streptomycin (between 4 and 7 days following plating on SM).Once sufficient growth in the 12-well plate was achieved, the cells were transferred to 100 mm × 15 mm Petri dishes and were scaled up to flasks once the Petri dish was over 70% confluent.A 200 μL aliquot of cells from the confluent Petri dish was washed and then lysed for DNA extraction using 48 μL LyseB buffer (10 mM Tris, pH 8.3, 50 mM KCl, 2.5 mM MgCl 2 , 0.45% NP40, and 0.45% Tween) and 2 μL Proteinase K. 30 To detect indels generated from CRISPR/Cas9, a 764 bp region spanning the sgRNA cut site was amplified via PCR and ran on an agarose gel (Table S1D).PCR amplicons from clones with visible indels were used for Sanger sequencing, and we selected a clone with a 32 base pair mutation that resulted in a frameshift and early truncation of the gene product as the iptA − strain for all further experiments (Figure 2 and Table S1E).A custom antibody was used to validate the knockout through western blot; however, we were unable to detect IptA or GFP-IptA with the antibody.Therefore, RT-qPCR and CK production screens, detailed in the following sections, were used to validate the knockout of iptA, as similarly performed by Lindner et al. 17 in Physcomitrella patens.The full length iptA nucleotide sequence and the selected iptA − strain nucleotide sequence F I G U R E 1 Experimental design and medium selection for targeting select life cycle stages of Dictyostelium discoideum to uncover the role of iptA during vegetative growth and early development.The nutrient-rich HL5 medium was used for routine culturing of vegetative amoebae and select experiments.Vegetative amoebae were also cultured in FM minimal medium to ensure any exogenous cytokinins from the organic ingredients in HL5 medium were not masking any effects of iptA deficiency.For assessing early development, we limited amino acids in the culture medium which has been thoroughly documented to initiate autophagy in D. discoideum.In all experiments that assessed early development, we used previously established methods and cultured vegetative amoebae in FM minimal medium without amino acids except for the metabolomics experiment setup.Early development samples did not contain any streaming or aggregating amoebae in contrast to what is shown in the representative photo in this image.The metabolomics samples were cultured in FM minimal medium without arginine or lysine as opposed to without all amino acids as used in the other experiments because the density of cells required for metabolomics analysis was much higher and resulted in streaming or aggregating cells if cultured without amino acids altogether.To keep early development sample harvest similar across all experiments throughout the manuscript, the metabolomics early development samples were cultured in FM minimal medium without arginine and lysine to avoid streaming or aggregating amoebae.The 24-h fruiting body samples were starved for 24 h on 1% KK2 agar to avoid CK contamination from organic ingredients in nutrient-rich agar medium, such as SM medium.with their corresponding amino acid translations can be found in Table S2.
An N-terminal GFP-IptA overexpression construct was generated and transformed into D. discoideum, hereafter referred to as GFP-IptA, using the pTX-GFP extrachromosomal vector as previously described. 31,32A pTX-GFP empty vector was used as a control for the GFP-IptA strain for the ADP/ATP and NAD + /NADH assays.Strains expressing this vector were supplemented with G418 (20 μg/ mL), and the primers used for amplifying the full-length iptA gene can be found in Table S3.CK profiling was used to assess the CK levels produced in the pTX-GFP empty vector and GFP-IptA overexpression strain and to confirm overexpression of IptA (methodology discussed in a later section).

| RT-qPCR
To confirm the knockout of iptA, we assessed its relative mRNA expression through RT-qPCR.WT and iptA − strains were starved for 24 h in flasks, and replicates of 2 × 10 7 cells were lysed and homogenized using Trizol reagent (Thermo Fisher, Mississauga, ON, Canada) according to manufacturer's instructions to be used for RNA extraction.RNA was isolated according to the manufacturer's instructions using the Ambion PureLink RNA Mini Kit, which included an on-column DNaseI digestion (Thermo Fisher, Mississauga, ON, Canada).Both RNA quantity and quality were measured using a Nanodrop 2000 spectrophotometer (Thermo Fisher, Mississauga, ON, Canada), and absorbance measurements A260/A280  S2).(C) Confirmation of gene KO through reverse transcriptase-quantitative PCR (qPCR) analysis of iptA gene mRNA expression using two different primer sets spanning the iptA gene.Wildtype (AX3) and iptA − vegetative cells were harvested from flasks after 24 h in FM minimal medium.Data were normalized to gpdA and rnlA (data are presented as individual values (dots) and mean ± standard error for three independent).(D) Total CK levels are 99% lower in iptA − than in WT.Fruiting bodies were harvested from KK2 agar, and both intra-and extracellular CK concentrations were analyzed and pooled (data are presented as individual values (dots) and mean ± standard error for four independent replicates).For panels C and D, the statistical significance was determined by a t-test (two-tailed; *p < .05,**p < .01,****p < .0001).and A260/A230 with ratios ~2.0 were accepted as pure for RNA.One microgram of RNA was reverse-transcribed to cDNA using the Bio-Rad 5X iScript RT supermix kit (Bio-Rad Laboratories Canada; Mississauga, ON, Canada) for RT-qPCR, after which samples were diluted 1:1 with nuclease-free water.Samples without reverse transcriptase were included to confirm no genomic DNA contamination.Quantitative PCR (qPCR) was performed on cDNA samples using the 2X SsoAdvanced Universal SYBR Green Supermix (Bio-Rad Laboratories Canada; Mississauga, ON, Canada).A Bio-Rad CFX Connect was used to perform the reactions in triplicate, and gene expression data were normalized to the reference genes, rnlA (mitochondrial large subunit) and gpdA (glyceraldehyde-3-phosphate dehydrogenase).The Bio-Rad CFX Maestro Software was used to analyze the data.All primers used were designed using NCBI Primer Blast (https:// www.ncbi.nlm.nih.gov/ tools/ primer-blast/ ) (Table S4), and data were collected and analyzed as per the MIQE guidelines.

| Cytokinin analysis
Total cytokinin (CK) levels were quantified to confirm the direct effects of gene knockout/overexpression and to assess the overall contribution to CK levels of iptA among the other tRNA-CK biosynthesis genes present in D. discoideum (iptB and iptC).A total of 7.5 × 10 7 cells were plated onto four 100 mm × 15 mm 1% KK2 agar Petri dishes and incubated at 21°C for 24 h in a humidity chamber, as previously described. 13Fruiting bodies were harvested from the agar and separated into extracellular (EC; supernatant from KK2 washes used to collect sample off the plate) and intracellular samples (IC; pelleted fruiting bodies).The samples were flash-frozen in liquid nitrogen and transferred immediately to a −80°C freezer.Samples were subjected to CK extraction and purification and analyzed using high-performance liquid chromatography-positive electrospray ionization-high-resolution tandem mass spectrometry (HPLC-(ESI+)-HRMS/MS), as previously described. 13Levels of the four most abundant CK forms detected (N 6 -isopentenyl adenine, N 6 -isopentenyladenosine (iPR), N 6 -isopentenyladenine-9-riboside-5′ phosphate (iPRP), and discadenine) in both EC and IC samples were quantified individually and combined to determine total CK content for each strain.

| Fruiting body morphology and count
Fruiting body morphology was examined after multicellular development on 1% KK2 agar plates, rather than the typical nutrient-rich SM agar plates with K. aerogenes, to avoid CK contamination from organic ingredients used in SM agar and CKs potentially produced by bacteria. 33 × 10 7 cells were pelleted, washed, and resuspended in 1 mL of KK2.From this cell suspension, 25 μL droplets containing 7.5 × 10 5 cells were deposited onto KK2 agar and incubated at room temperature (22°C) for 24 h.Images were taken from above using a Leica EZ4W stereomicroscope equipped with an internal 5MP CMOS camera.The same magnification was used for all strains, and the sori area was quantified using the Fiji/ImageJ measuring tool (RRID:SCR_00307).From these same images, the total number of fruiting bodies formed and the number of fruiting bodies that contained sori reaching the average sori area of WT were counted (>500 pixels).This experiment was independently replicated five times with three technical replicates per experiment.

| Cell proliferation (liquid culture and solid plates)
Cell proliferation was assessed in liquid culture through cell counts, as previously described, and on solid bacterial plates through measurements of plaque diameter (mm). 13For both experiments, cells in the mid-log phase of growth were collected and washed twice in KK2 buffer.For liquid culture measurements, the cells were resuspended to a density of 2 × 10 5 cells/mL in HL5 medium or 5 × 10 5 cells/mL in FM minimal medium.FM minimal medium was tested in addition to HL5 as it is a defined medium that has low background levels of CKs compared with HL5 (Figure 1). 33Cell counts were then taken every 24 h over a 144-h growth period using a hemocytometer.For solid culture measurements, serial dilutions were performed on the washed cells to plate a total of 15 cells in 15 μL on each plate with Klebsiella aerogenes on SM/2 agar plates.Imaging of plaques occurred every 12 h once the plaques were visible to the eye, starting at 48 h after plating and ending at 84 h after plating.Images of the plaques were captured at the respective time points using a Leica EZ4W stereomicroscope equipped with an internal 5MP CMOS camera (Leica Microsystems Incorporated, Concord, Canada) and were quantified using Fiji/ImageJ (RRID:SCR_00307).Both experiments were independently replicated three times with three technical replicates per experiment.

| Pinocytosis
Pinocytosis was assessed as previously described measuring both uptake and release of the fluorescent marker, fluorescein isothiocyanate (FITC)-dextran. 34Axenic D. discoideum cultures were harvested and resuspended in 5 mL of fresh HL5 medium at a density of 5 × 10 6 cells/ mL.Cultures were then incubated at 21°C and shaken at 150 rpm. 100 μL of a 20 mg/mL FITC-dextran stock solution was added to the suspension.Endocytosis was assessed every 15 min, starting from the addition of FITCdextran, over a 90-min time course.Prior to the addition of FITC-dextran, a background measurement of cells was sampled to subtract background fluorescence and control for differences in protein content.For all sampling periods, a 250 μL cell suspension was collected, washed twice with 750 μL ice-cold Sorenson's buffer (2 mM Na 2 HPO 4 , 14.6 mM KH 2 PO 4 , pH 6.0) and lysed with 500 μL lysis buffer (50 mM Na 2 HPO 4 , 0.2% Triton-X, pH 9.3).Triplicates of 100 μL of lysate were added to separate wells of black, clear bottom 96-well plates.Fluorescence measurements were obtained using a BioTek Synergy HTX plate reader using the following filters: 485/20 nm for excitation and 528/20 nm for emission (BioTek Instruments Incorporated, Winooski, VT, USA).To assess exocytosis, cells were immediately washed with fresh HL5 medium following the 90-min time course and were resuspended in the same volume of HL5 remaining after sampling (without FITC-dextran) and covered in aluminum foil. 34 time-zero measurement was taken using the same sampling method described above to be used as a 100% relative fluorescence measurement for the remaining sample, which was collected after 120 min of incubation.The decrease in fluorescence indicates the amount of FITCdextran released following the 90-min FITC-dextran preincubation.The experiment was independently replicated three times with three technical replicates per experiment.

| Cytokinesis
To assess cytokinesis, a total of 3 × 10 5 vegetative cells were harvested, deposited on individual coverslips, and incubated for 36 h at 22°C in Lo-Florescence HL5 medium or FM minimal medium (Formedium, Hunstanton, Norfolk, United Kingdom; Figure 1).Following incubation, the coverslips were fixed in −80°C methanol for 1 h and then mounted on slides using Dako mounting solution with DAPI (Sigma-Aldrich, Oakville, ON, Canada).Fluorescent micrographs were captured with a Prime BSI scientific complementary metal oxide semiconductor (sCMOS) camera (Teledyne Photometrics, Tucson, AZ, USA) on a custom-built Zeiss Cell Observer Microscope (Intelligent Imaging Innovations, Denver, CO, USA) using a 1.4 NA 63X immersion oil objective lens and LED illumination via a Spectra light engine (Lumencor, Beaverton, OR, USA).Images were processed and analyzed using Adobe Photoshop v. 21.2.12 (RRID:SCR_01499).Three independent replicates were performed, and at least 100 cells were analyzed per replicate.

| Transmission electron microscopy (TEM) and mitochondrial size metrics
Transmission electron microscopy (TEM) was used to assess any aberrant ultrastructural phenotypes from iptA knockout.Vegetative WT and iptA − cells were cultured in either FM minimal medium or starved in FM minimal medium without amino acids for 36 h in 100 mm × 15 mm Petri dishes, as previously described (Figure 1). 35Cells were harvested, washed, and fixed in 2.5% glutaraldehyde (pH 7.2) diluted in 0.1 M of sodium cacodylate buffer for 2 h.The fixed cells were rinsed three times (10 min per wash) with 0.1 M sodium cacodylate buffer.The supernatant was removed, and the cells were post-fixed in 1% osmium tetroxide for 2 h.The cells were rinsed again in distilled water and placed in 0.25% uranyl acetate at 4°C overnight.Following this, the cells were dehydrated in a graded aqueous-acetone series, which were then embedded in 100% epon-araldite resin in a 60°C oven for 48 h to allow the samples to harden correctly.To obtain ultrathin sections, a Reichert Jung Ultracut E Ultramicrotome fitted with a diamond knife was used (approximately 100 nm thick).Sections were retrieved on 300 uncoated copper mesh grids.The sections were stained with 2% aqueous uranyl acetate for approximately 10 min, which was followed by a series of 5-min rinses with distilled water.Samples were stained with lead citrate for 4 min and rinsed two times with distilled water for 5 min.Finally, the samples were observed using a JEOL JEM 1230 transmission electron microscope at 80 kV, and images were taken using a Hamamatsu ORCA-HR digital camera.Mitochondria are easily distinguished from other organelles owing to their higher electron density, which makes them appear darker, and they possess visible and distinct internal cristae structures allowing for differentiation from other organelles.From the images, mitochondrial area and circularity were measured using Fiji/ImageJ, whereby a value of 1.0 indicates a perfect circle (RRID:SCR_00307).

| Metabolomic profile of mitochondrial-related metabolites
Samples from WT, iptA − , and the GFP-IptA strains were harvested under the same conditions as used for assessing ultrastructural phenotypes via TEM, with the addition of a 24-h fruiting body sample for each strain.FM minimal medium without arginine and lysine was used in place of FM minimal medium without amino acids to keep the early developmental sample harvest consistent with the samples collected for TEM analyses (Figure 1).Culturing cells at a high density (required for metabolomics analyses) in FM minimal medium without amino acids caused cells to stream and aggregate.Therefore, we switched the medium to FM minimal medium without arginine and lysine to maintain the high cell density required for the analysis and to avoid the streaming and aggregation of amoebae.FM minimal medium without arginine and lysine has been previously shown to initiate autophagy within 15 min in D. discoideum. 36For the vegetative samples, a total of 2 × 10 6 cells were seeded into four 100 mm × 15 mm Petri dishes with 7 mL HL5 and incubated overnight.The medium was disposed of, and the adherent cells were washed twice with cold KK2 buffer followed by the addition of 7 mL fresh medium-either FM minimal medium or FM minimal medium without arginine and lysine.The cells were incubated at room temperature for 36 h.For the 24-h fruiting body samples, 7.5 × 10 7 cells were plated onto four individual 100 mm × 15 mm 1% KK2 agar Petri dishes and collected after a 24-h incubation period as described in the CK analysis methodology section.All sample types were harvested, washed, pelleted, and flash frozen in liquid nitrogen and kept at −80°C until they were processed for metabolite extraction and purification.
Samples were freeze-dried and endogenous metabolites were extracted with ice-cold 50% acetonitrile (ACN) and purified by solid phase extraction (SPE) using HLB cartridges as previously described. 37The mitochondrialrelated metabolites were assessed using a customized quantification method and grouped into the following categories: amino acids and derivatives, nucleotide metabolites, and energy metabolism which encompasses metabolites from the glycolysis/gluconeogenesis pathways, tricarboxylic acid (TCA) cycle pathway, and pentose phosphate pathway, as previously described (Table S5). 38All samples were spiked with 10 μL of a stable-isotope labeled canonical amino acid mix (0.25 μM final concentration of amino acids; Cambridge Isotope Laboratories, Tewksbury, MA, USA; Table S6).Samples were evaporated to dryness at ambient temperature in a speed vacuum concentrator.The sample residues were redissolved in 500 μL of 90% acetonitrile (acetonitrile: water, v/v).All samples were filtered using 0.2 μm PVDF spin filter with 2 mL receiver tubes (InnoSep Spin, Canadian Life Sciences, Peterborough, Canada) and transferred to 2 mL vials containing 350 μL inserts for high-performance liquid chromatography-high resolution accurate mass-full scan mass spectrometry (HPLC-(HRAM)-FS-MS) analysis.
Samples were resolved with a Kinetex C18 column (2.1 × 50 mm, 2.6 μm).A volume of 25 μL of each sample was injected into a Dionex UltiMate 3000 HPLC (ThermoFisher, Mississauga, Canada) coupled to a QExactive Orbitrap mass spectrometer (ThermoFisher, Mississauga, Canada).A flow rate of 0.3 mL/min was used with a mobile phase of 0.08% acetic acid in water (A) and 0.08% acetic acid in acetonitrile (B).The following gradient was used to elute the analytes: mobile phase B was held at 0% for 1.25 min to help retain the compounds on the column and avoid the metabolite elution in the void volume before increasing to 50% over 2.75 min and to 100% over the next 0.5 min.Solvent B was then held at 100% for 2 min before returning to 0% over 0.5 min for 4 min of column re-equilibration.
Orbitrap QExactive was operated with a heated electrospray ionization (HESI) probe in positive and negative mode. 39Each sample was analyzed in full scan mode using a mass range of m/z 70-900, and data were acquired at 70000 resolution, automatic gain control (AGC) target of 1 × 10 6 , and maximum injection time (IT) of 100 ms.Additionally, pooled sample mixtures from each strain (WT/iptA − /GFP-IptA) or treatment type (FM/FM-AL/24-h) composed of 20 μL of each sample extract was used to generate MS/MS for compound identification.For this, the top 10 data-dependent acquisition experiments were performed separately in positive and negative ionization modes for each of the pooled samples to obtain MS/ MS spectra of the most abundant compounds. 40etabolite processing and identification of all full scan and ddMS2 data was performed in Xcalibur 4.1 software, using a customized quantification method for the list of mitochondrial-related metabolites (Table S5).Metabolites were identified by accurate mass (with a 10 ppm mass error) and comparison of retention times to authentic/labeled standards or by accurate mass and comparison of fragmentation patterns to MS/MS databases when available.Authentic standards were a mix of unlabeled, high-purity compounds for HPLC analysis including sugars, organic acids, and amino acids; additionally, labeled amino acids were spiked into each sample as specified above (Level 1 and highest confidence for metabolite identification). 41For metabolites of interest without a standard, METLIN and PubChem databases were used to identify metabolites by accurate mass and comparing their MS/MS fragmentation patterns (Level 2 confidence) 41 or by accurate mass and MS1 m/z database match (Level 3 confidence) (Table S5). 41To determine the relative concentration of mitochondrial metabolites in the three different strains and treatment types, the peak areas were first normalized based on the median recovery of the 20 stable-isotope labeled canonical amino acid mix in each analyzed sample.To allow for direct comparison of metabolite levels across samples, the individual metabolite levels were then normalized relative to the highest recovery of that particular metabolite across all samples (Table S6).The relative metabolite content of the iptA − or GFP-IptA strains was compared to WT to assess any metabolic consequences in mitochondrial-related metabolism from iptA-deficiency or overexpression for each of the three treatment types.
2.11 | ADP/ATP and NAD + /NADH measurement ADP and ATP contents were measured using the ADP/ ATP Ratio Luminescent Assay Kit according to manufacturer's directions (mak135; Sigma-Aldrich, Oakville, ON, Canada). 5 × 10 5 axenic vegetative cells were washed twice in phosphate buffered saline (PBS), resuspended in 10 μL PBS, and plated in white 96-well plates.A total of three independent experiments were performed, each with two replicates for measuring ATP, ADP, and residual ATP background luminescence.The ADP/ATP ratio was calculated by the following equation: (RLU of ADP-RLU of residual ATP)/RLU of ATP, where RLU represents relative light units.NAD + and NADH content were measured using the NAD + /NADH Fluorometric Assay Kit according to manufacturer's instructions (ab176723; Abcam, Toronto, ON, Canada).9 × 10 7 vegetative cells were washed twice in PBS, resuspended in 225 μL warm lysis buffer provided with the kit, and incubated at room temperature for 15 min in the dark.25 μL of lysed cell suspension was plated in individual wells in a black, clear bottom 96-well plate, and the assay continued according to manufacturer's instructions.A total of three independent experiments were performed, each with two replicates for measuring NAD + , NADH, and total NAD + /NADH.A standard curve for NADH provided with the kit was used to calculate concentrations of the compounds for each reaction.The BioTek Synergy HTX plate reader was used to measure luminescence for the ADP/ATP assay, and fluorescence for the NAD + /NADH assay (excitation: 528/20 nm and emission: 590/35 nm) (BioTek Instruments Incorporated, Winooski, VT, USA).

| Statistical analyses and software
All statistical analyses were performed using GraphPad Prism v. 9.5.0 (RRID:SCR_002798, GraphPad Software, La Jolla, CA, USA).For pairwise comparisons, two-tailed t-tests were performed, and for multiple comparisons, either a one-way ANOVA with Dunnett's multiple comparison test or a two-way ANOVA with the Tukey multiple comparison test was performed, unless otherwise specified.MetaboAnalyst v. 5.0 (RRID:SCR_015539) was used to generate heatmaps for the metabolomics analysis to show the clustering of samples with the Euclidean distance measure, Ward clustering method, and the oneway ANOVA statistical test (Pang et al., 2022).Figures were generated using Prism v. 9.5.0,Adobe Photoshop v. 21.2.12 (RRID:SCR_01499), and Adobe Illustrator v. 25.2 (RRID:SCR_010279).

| Confirmation of iptA knockout and aberrant developmental phenotypes
To investigate the role of IptA in vegetative growth and early development of D. discoideum, CRISPR-/Cas9mediated genome editing was used to disrupt the iptA gene and generate a knockout strain.We screened four different knockout clones through PCR and Sanger sequencing with various indels and selected a clone with a 32 base pair deletion of nucleotides 118-149 from the iptA genomic sequence for all further experiments (Figure 2A,B and Table S2).Predicted translation of the selected clone results in a 44 amino acid protein product, where the first 39 amino acids align with the native IptA amino acid sequence before a frameshift in the reading frame occurs and two stop codons are encountered resulting in early truncation of the protein.RT-qPCR confirmed the absence of iptA transcript in the mutant cells (Figure 2C).Furthermore, we observed a 99% reduction in total cytokinin (CK) levels after 24 h of development, which is consistent with previous work conducted with D. discoideum iptA − cells during the later stages of multicellular development (Figure 2D; Anjard and Loomis,  2008).We confirmed that iptA-deficiency results in aberrant morphology of iptA − fruiting bodies, in which we quantified average sori area, number of fruiting bodies formed, and number of fruiting bodies formed with an average sori area similar to WT (Figure 3A-D).In all cases, we observed a significant reduction in these fruiting body metrics for the iptA − strain (t-test, two-tailed; *p < .05,**p < .01).

| iptA − does not affect vegetative growth or pinocytosis
As CKs are known for their roles in promoting cell division, 2,3,[42][43][44] we first assessed the effects of its iptA loss on vegetative growth in axenic liquid culture and then on solid medium with bacteria (Figure 4A-C).Since we previously determined that CKs are abundant in culture medium containing organic ingredients, such as yeast extract, we used two different media in the axenic vegetative growth experiment to ensure exogenous CKs from the media were not masking any effects of iptA deficiency: nutrient-rich medium HL5 (which contains higher CK levels) and FM minimal medium (which contains minimal CKs). 33We assessed axenic vegetative growth by measuring cell densities every 24 h over a 144-h time course.Significant increases in iptA − cell densities compared with WT were only noted in the FM minimal medium at the 72-and 96-h time points (Figure 4A,B).On solid medium with bacteria, there were no differences observed between WT and iptA − growth rates, as measured through plaque expansion (Figure 4C).
We evaluated the effects of iptA − on pinocytosis to determine if loss of iptA affects the uptake or release of nutrients through measurement of the fluorescent marker, FITC-dextran (Figure 4D,E).Knockout of iptA did not have any effect on endocytosis or exocytosis, as the relative % fluorescence of FITC-dextran uptake and release was similar to WT.The combined analyses indicate that loss of iptA had no major effects on vegetative proliferation (in liquid culture or on agar) or pinocytosis.

| iptA − reduces cytokinesis
To assess cytokinesis, we compared the numbers of nuclei per cell between iptA − and WT cells by staining with DAPI under nutrient-rich and minimal nutrient conditions (Figure 5A-C).Despite there being no effects of iptA − on cell proliferation, there were significant increases in the number of nuclei per cell in the iptA − strain compared with WT in both HL5 and FM minimal medium, indicating that cytokinesis was impaired (Figure 5B,C; p < .0001).There was a more pronounced effect in nutrient-rich conditions, where iptA − cultures contained 30% more cells with two or more nuclei and 10% more cells containing three or more nuclei compared with WT (Figure 5A,B).In FM minimal medium, iptA − cultures contained 20% more with two or more nuclei and 10% more cells containing three or more nuclei compared to WT (Figure 5A,C).

| Loss of iptA results in abnormal mitochondrial morphology
We used transmission electron microscopy (TEM) to examine possible ultrastructural phenotypes from iptA deficiency during vegetative growth and early development using culture conditions previously optimized for D. discoideum and early development (36-h incubation in FM minimal medium without amino acids) (Figures 6A and  S1). 35In this study, early development is defined by the initiation of autophagy through amino acid limitation in the culture conditions, which has been shown to be induced within minutes. 36The early development samples did not contain any streaming or aggregating cells.Visually, the vegetative growth iptA − mitochondria had a rounded shape compared to those of WT in vegetative growth samples cultured in FM minimal medium (Figure 6A).From the vegetative growth TEM images, we quantified the number of mitochondria per cell and their circularity (Figure 6B,C).There was a significant increase in the number of mitochondria per cell in the iptA − strain (unpaired, two-tailed t-test, p < .05),as well as a significant increase in circularity (unpaired, two-tailed t-test, p < .0001).For the circularity metric, a value of 1.0 indicates a perfect circle.The average circularity of mitochondria in the iptA-deficient strain was 0.91, indicating a much higher proportion of rounded mitochondria compared with the WT strain, which had an average circularity of 0.83 (Figure 6C).During early development, iptA − samples cultured in FM minimal medium without amino acids displayed impaired turnover of organelles, as visualized through decreased formation of recycling organelles (i.e., less double-membraned organelles) and cytoplasmic teins under autophagy stimulating conditions compared with WT (Figure S1).

| iptA deficiency and overexpression alter mitochondrial metabolism
The results from the TEM analysis prompted us to explore the metabolic consequences of altered CK dynamics.We first generated an N-terminal GFP IptA-overexpression strain, referred to as the GFP-IptA strain, to assess the effects of IptA-overexpression on mitochondrial metabolism using a metabolomics approach.To verify altered CK dynamics in the generated GFP-IptA strain, we quantified CK levels during vegetative growth in the GFP-IptA strain and a pTX-GFP empty vector control.The total amount of CK produced in the GFP-IptA strain was 2.38 × 10 6 pmol/10 6 cells, which was over 85-fold higher than that of the pTX-GFP empty vector control (Figure 7).In both strains, the classic iP-type CKs were detected, and discadenine was only detected in the GFP-IptA strain (Figure 7).The trends and overall levels of individual CKs produced by the pTX-GFP empty vector control (expressed in AX3 cells) were similar to previous experiments using AX3 cells, in addition to the lack of discadenine in vegetative growth samples (data not shown).Thus, for the metabolomics analysis, we compared the levels of metabolites from both the iptA − and GFP-IptA strains directly to WT cells (AX3) considering the CK analysis indicated no alteration of CK production from the Nterminal GFP molecule (Figures 8 and S2).
We employed full scan, high-resolution mass spectrometry to analyze endogenous metabolites and focused on 70 mitochondrial-related metabolites making up the "mitobolome" described by Chen et al., 38 which we grouped into the following categories: amino acids and derivatives, nucleotide metabolites, and energy metabolism.The latter encompasses metabolites from the glycolysis/gluconeogenesis pathways, TCA cycle pathway, and pentose phosphate pathway (Table S5).The metabolomics samples were cultured under the same conditions as those used for the TEM analyses with the addition of a 24-h fruiting body in the sampling methods and downstream (MetaboAnalyst v 5.0, RRID:SCR_015539; Figures S3-S5). 45This analysis was performed only to assess reliability in the sampling methodology across individual replicates for each of the three strains used in the downstream targeted mitochondrial metabolomics analyses.We did not identify any of the top 400 significantly different features in this analysis.Among the top 400 features shown in each of the three sampled life cycle stages, the trends in feature levels of the WT and iptA − strains were more similar to each other than the GFP-IptA strain in the vegetative amoebae stage (Figure S3).In the early development stage, the feature levels of the WT and GFP-IptA strains were more similar to each other than those in the iptA − strain (Figure S4), and in the 24-h development fruiting bodies, the feature levels of the iptA − and GFP-IptA strains were more similar to each other than those in the WT strain (Figure S5).Among the three sampled life cycle stages analyzed for mitochondrial-related metabolic changes, the vegetative amoebae were the most affected.Specifically, the TCA cycle was the most affected pathway by loss or overexpression of iptA (Figures 8 and S2A).There were significant decreasing trends in metabolite abundance in succinate and fumarate in both strains compared with WT (two-way ANOVA with a false discovery rate of p < .05;Figures 8  and S2A).In the glycolysis pathway, a significant decrease in lactate was observed in both the iptA − and GFP-IptA strains compared with WT (Figures 8 and S2A).In only the GFP-IptA strain, there was a significant decrease in the levels of glucose/fructose (isomers that were indiscernible by the LC method used) and malate (p < .01 and p < .0001,respectively) compared to WT.There were no clear trends observed in the early development samples cultured in FM minimal medium without arginine and lysine for any of the assessed energy metabolism metabolites glycolysis pathway, TCA cycle, or the phosphate pathway (Figure S2B).However, in the 24-h development fruiting body samples, a decreasing trend in succinate level was observed in both the iptA − and GFP-IptA strains compared with WT (Figure S2C; p < .05 and p < .001respectively).The combined energy metabolism pathway metabolite analyses align with the abnormal mitochondrial morphology visualized through TEM (Figure 6).
Mitochondrial dysfunction is well studied in D. discoideum and mitochondrial disease phenotypes are conserved in this organism. 46As such, a set of "readouts" consistent in mitochondrial disease phenotypes have been established in D. discoideum to assess dysregulated intracellular signaling pathways.In many cases, mitochondrial disease phenotypes affect energy metabolism, which results in the chronic activation of the AMP-activated protein kinase (AMPK). 46herefore, using commercially available kits, we assessed energy (ATP and ADP content) and redox metabolism (NAD + and NADH content) in iptA − cells relative to WT and in GFP-IptA cells relative to a pTX-GFP empty vector control during vegetative growth (Figure 9).The levels of ATP were relatively similar in iptA − cells compared with WT cells, but the ADP levels were significantly decreased, resulting in a significant decrease in the ADP/ATP ratio in the iptA − strain (Figure 9A).For the GFP-IptA strain, both ATP and ADP levels were significantly and unequally increased relative to the pTX-GFP empty vector strain, in such a way that created a significant increase in the ADP/ATP ratio (Figure 9A).Interestingly, AMP levels, as assessed in the metabolomics analyses, were significantly elevated in both the iptA − and GFP-IptA strains compared to WT (unpaired, two-tailed t-test; **p < .01,***p < .001; Figure 9B).These findings point to iptA deficiency and/or overexpression being associated with mitochondrial function, as AMPK is activated by elevated AMP/ADP and AMP/ATP ratios and oxidative stress.This chronic activation of AMPK is able to restore ATP levels but leads to characteristic aberrant phenotypic outcomes that have been associated with mitochondrial dysfunction in D. discoideum. 46,47Our results point to elevated AMP/ADP or AMP/ATP ratios as the activators of AMPK over oxidative stress, as the measurements of NAD + and NADH were comparable across the tested strains (Figure 9C).

| iptA-deficiency and overexpression downregulates amino acid metabolism
Within the metabolomics analysis, we examined changes in amino acid levels and their derivatives from loss or overexpression of iptA (Figure 10).Interestingly, the amino acid levels were the most affected by iptA deficiency or overexpression in the 24-h development fruiting body stage compared with the vegetative and early development stages (Figure 10A-C).Specifically, alanine/sarcosine (isomers that were indiscernible by the LC method used), aspartate, glutamate, glutamine, serine, tyrosine, and GABA were all significantly downregulated in both the iptA − and GFP-IptA strains compared to WT (two-way ANOVA with Benjamini, Krieger, and Yekutieli correction controlling for multiple comparisons with false discovery rate of <.05; p < .05; Figure 10C).In only the GFP-IptA strain, there was a significant decrease in the levels of threonine/ homoserine (isomers that were indiscernible by the LC method used) (Figure 10C).Similar decreasing trends in the iptA − and GFP-IptA strains for many of these same amino acid and derivative levels were observed in the vegetative growth samples; however, only arginine levels were significantly downregulated in both the iptA − and GFP-IptA strains (Figure 10A).Additionally, leucine/isoleucine, methionine, and tyrosine were significantly downregulated in the GFP-IptA strain compared with WT (Figure 10A).Like the results of the energy metabolism analyses in the early development samples (cultured in FM minimal medium without arginine or lysine), there were no clear trends in amino acid or derivative levels observed between the iptA − and GFP-IptA strains compared with WT (Figure 10B).
F I G U R E 7 Cytokinin (CK) content (pmol/10 6 cells) of the four most abundant CKs in vegetative Dictyostelium discoideum amoebae pTX-GFP empty vector control and GFP-IptA strains detected using high-performance liquid chromatographypositive electrospray ionization-high-resolution tandem mass spectrometry (HPLC-(ESI+)-HRMS/MS).Individual CK analytes were detected, and levels were compared between strains.Data are presented as individual values and mean ± standard error for four independent replicates.The abbreviations for the CKs shown are as follows: N 6 -isopentenyladenine (iP), discadenine (DA), N 6 -isopentenyladenosine (iPR), and N 6 -isopentenyladenine-9riboside-5′ phosphate (iPRP).DA was not detected in the empty vector control.I G U R E iptA deficiency (iptA − ) and overexpression dysregulate the tricarboxylic acid (TCA) cycle in amoebae.High-performance liquid chromatography-high-resolution accurate mass-full-scan mass spectrometry (HPLC-(HRAM)-FS-MS) was used to detect changes in mitochondrial-related metabolites as a result of iptA deficiency and overexpression.Dictyostelium discoideum WT (AX3), iptA − , and GFP-IptA vegetative cells were cultured in FM minimal medium for 36 h before harvesting cells for analysis (n = 4).Levels of succinate, fumarate, and lactate were all significantly decreased in both the iptA − and GFP-IptA strains compared with WT.In the GFP-IptA strain, the levels of glucose/fructose and malate were significantly decreased compared with WT and were similarly decreased in the iptA − strain compared with WT.This analysis revealed that iptA-deficiency and/or overexpression significantly affects key metabolites in the TCA CKs are necessary for the proper development discoideum and act as key signals for terminal differentiation during fruiting body formation. 11,24The induction of sporulation and the maintenance of spore dormancy are two roles governed by precise regulation of CK production through the primary CK biosynthetic enzyme, IptA. 11,48iptA expression increases throughout multicellular development and peaks during fruiting body formation. 11,48Thus, previous research on iptA and CK production in D. discoideum had focused on the later developmental, reproductive life cycle stages. 11,48In this study, we were particularly interested in dissecting the role of CKs during the vegetative and early development stages of D. discoideum, based on our initial research documenting CK levels during both of these stages, in addition to the fruiting body stage of development. 13herefore, we generated a knockout strain of the primary CK-biosynthesis gene, iptA, focusing our attention on aberrant vegetative growth and early developmental phenotypes in D. discoideum.We examined classical roles of CKs by assessing proliferation, cytokinesis, and nutrient acquisition and found that iptA-deficiency leads to abnormal cytokinesis by a mechanism that does not alter proliferation or nutrient uptake.We then used transmission electron microscopy (TEM) to assess the ultrastructure of the cell, which revealed altered mitochondrial morphology.To gain further insight into CK dynamics, we generated a GFP-IptA overexpression strain for the remaining analyses.Metabolomics analyses of both the iptA − and GFP-IptA strains revealed the downregulation of several TCA cycle metabolites and amino acids, in addition to an increase in the energy metabolite, AMP.
CKs were named based on their earliest established role in promoting cell division or cytokinesis. 42,49Considering this classical role, we assessed cytokinesis in the iptAdeficient strain during both nutrient-rich and minimal nutrient growth and found that cytokinesis was decreased in D. discoideum following loss of iptA.In Arabidopsis, the molecular mechanism underlying mitotic CK-activated cell division was recently discovered. 44CKs were shown to spike precisely at the G2/M transition coincident with rapid nuclear accumulation of the MYB3R4 transcription factor that activates mitotic gene expression.iptA expression in D. discoideum is altered during the early development (between 2 and 6 h of starvation) of both mybB − and MybB-GFP overexpression strains, and the final 24-h development iptA expression levels in the mybB − RNAseq data set were 50-fold lower than native iptA transcript levels. 50,51As such, it would be of interest to assess whether CKs act on the MybB transcription factor in D. discoideum F I G U R E 9 iptA deficiency and overexpression affect energy metabolite levels (AMP/ADP/ATP) but do not affect NAD + /NADH levels.(A) ATP and ADP content is reduced in the iptA − strain relative to WT (AX3) during vegetative growth resulting in a significant decrease in the ADP/ATP ratio.In the GFP-IptA strain, ATP and ADP content is significantly increased relative to the pTX-GFP empty vector control during vegetative growth, resulting in a significant increase in the ADP/ATP ratio.(B) AMP levels were significantly increased in both the iptA − and GFP-IptA strains compared with WT. (C) NADH content was elevated in both the iptA − and GFP-IptA strains relative to WT and empty vector strains during vegetative growth, respectively, resulting in a decreased NAD + /NADH ratio, but in all cases, there was no significance.For panels A and C, metabolite content was determined through commercial kits and analyzed with a plate reader to acquire relative readings compared with WT or empty vector strains, and data are presented as individual values (dots) and ± standard error for three independent experiments.initiate a transcriptional controlling mitosis and cytokinesis, perhaps in manner to that observed in Arabidopsis.
discoideum is a well-established model for studying mitochondrial disease. 46,52,53Phenotypes associated with mitochondrial dysfunction include those with or without defects in oxidative phosphorylation (OXPHOS), which are referred to as either classical mitochondrial disease or mitochondrial-associated dysfunction, respectively. 46The combined data from the analyses of iptA deficiency and overexpression strains provide a unique subset of phenotypes that are not typical of those described in classical mitochondrial disease in D. discoideum, but, rather, more in line with those described in cases of mitochondrialassociated dysfunction.Observable mitochondrialassociated dysfunctions that do not affect OXPHOS are rare and have varied phenotypes depending on the gene studied, which is consistent with our data. 53From the TEM data, iptA deficiency results in abnormal mitochondrial morphology and increased numbers of mitochondria in vegetative amoebae.The rounded mitochondrial morphology in the iptA-deficient cells was the first indication that iptA may be involved in aspects of mitochondrial function, which has not been previously established for CKs or their biosynthetic enzymes in plants or any other CK-producing organisms.A similarly rounded mitochondrial phenotype with increased content of mitochondrial DNA was observed in calcineurin-silenced vegetative cells in D. discoideum, which was speculated to be associated with oxidative stress. 54To verify, the authors assessed gene expression in several oxidative stress-related genes (sodA, sodB, sodC, sod2, and catA) and reported an upregulation in all but one of the measured genes.This indicates that increased intracellular oxidative stress was a result of calcineurin silencing.It was concluded that calcineurin is involved in the interplay between mitochondria and reactive oxygen species in vegetative growth in D. discoideum. 54While we did not assess redox metabolites beyond NAD + and NADH, there are numerous reports on the protective effects of CKs against oxidative stress in mammalian cells, 55,56 plants, [57][58][59][60] and bacteria. 61,62Taken together, it would be of interest to assess oxidative stress in the iptA-deficient and overexpression strains to determine whether CK and/or IptA exerts any protective effects in D. discoideum.
To further assess the potential role of IptA in mitochondrial function during vegetative growth and early development, we employed a metabolomics approach with a focus on mitochondrial-associated metabolites in both iptA-deficient and overexpression strains.We saw similar trends in metabolite profiles in both the iptA − and GFP-IptA strains rather than opposite trends, which are more typical of knockout and overexpression strains.While simple hypotheses would predict these strains have opposite effects, our results can be explained by hormesis patterns whereby very and very high levels can result same effect. 63The TCA cycle the most affected pathway we analyzed during vegetative growth.Succinate, fumarate, and malate were downregulated in both the iptA-deficient and overexpression strains.Moreover, there were increases in the levels of AMP in both the iptA − and GFP-IptA strains, despite ATP levels remaining unaffected among all strains assessed.In cases of either classical mitochondrial dysfunction or mitochondrial-associated dysfunction, the chronic signaling of AMPK appears to be a common feature underlying the similar or unique phenotypes observed in D. discoideum. 53,64,65The higher levels of AMP would, in turn, activate AMPK, leading to downstream signaling that shifts metabolism from an anabolic to a catabolic state and, thereby, stimulating glucose consumption. 66Furthermore, AMPK is likewise activated by decreased glucose levels, which we observed in both the iptA − and GFP-IptA strains.To further support the conducted metabolomic analyses, it would be informative to analyze the gene expression of the snfA/ampkα (encodes the alpha subunit of AMPK) and other mitochondrial genes in iptA-deficient and overexpression cells.
As part of the metabolomic analysis, we assessed the levels of all amino acids in the iptA-deficient and overexpression strains.Similar to the observed decreasing trends in metabolites of the TCA cycle, we observed decreasing amino acid levels in the vegetative amoebae and 24-h fruiting body iptA − and GFP-IptA strains.There were many more significantly decreased amino acids in the 24-h fruiting body samples, which were different from those decreased in vegetative amoebae samples.These findings support the idea of temporally separated dual roles of IptA in early growth versus late development, which aligns with the differing CK profiles observed at the same two stages. 13In vegetative cells, there were decreasing trends in the levels of arginine, leucine/isoleucine, methionine, and tyrosine in the iptA − and GFP-IptA strains.On the contrary, there were decreasing trends in alanine/sarcosine, aspartate, glutamine, glutamate, glycine, serine, threonine/homoserine, tyrosine, and GABA in the 24-h development iptA − and GFP-IptA strains.From a mitochondrial perspective, a decrease in various amino acids would be expected as they directly participate in the synthesis of the 13 mitochondrially encoded proteins necessary for proper function of the respiratory chain. 67Amino acids are also directly involved in various aspects of mitochondrial metabolism and are the site of biosynthesis for several of them-glutamine, glutamate, alanine, proline, and aspartate. 68Lussey-Lepoutre et al. 69 showed that the loss of succinate dehydrogenase, which is an integral component of the mitochondrial respiratory chain complex II and is responsible for oxidizing succinate to fumarate, dysregulates nonessential amino acid metabolism.Alanine, aspartate, glutamate, asparagine, and serine were among the most downregulated amino acids in their analyses.Accompanying these changes were also decreased levels of TCA cycle metabolites, including fumarate, malate, and citrate.Another study used a metabolomics approach to assess the metabolic consequences of fumarate-hydratase-deficiency (the enzyme that catalyzes the conversion of fumarate to malate in the TCA cycle) in mammalian cells and found that the cells became auxotrophic for arginine. 70Similar to these findings in mammalian cells, a study in yeast observed that disruption of mitochondrial respiration resulted in decreased arginine levels, as well as other amino acids derived from the TCA cycle. 71These studies, among many others, illustrate that perturbations in the TCA cycle can result in altered metabolic activity affecting amino acid metabolism, and numerous other pathways. 68,72n the mitochondrial-related metabolites assessed in this study, there were no trends found in the early development samples cultured in FM minimal medium without arginine and lysine.This result was unexpected, as our previous study assessing CK production across the D. discoideum life cycle observed increasing levels of total CKs from the vegetative growth to the early development life cycle stage. 13However, we did notice impaired turnover of organelles and cytoplasmic proteins based on the TEM analysis (Figure S1).In this study, early development was simulated through the initiation of autophagy by culturing cells in FM minimal medium without arginine and lysine for 36 h, as previously described. 35This method was different from our earlier CK profiling assessing early development, for which cells were starved for 6 h in KK2, and samples were collected with distinct streams and aggregation centers. 13A recent publication assessing mitochondrial dynamics in vegetative growth and early development in D. discoideum through bioenergetics and proteomics found that the main mitochondrial processes are downregulated during early development and that there was less mitochondrial yield overall in early developmental samples compared to vegetative cells. 73herefore, sampling a larger population of cells in early development may have helped make the findings in mitochondrial content more comparable to what was analyzed in the vegetative samples.Sampling after the streaming and aggregation centers have formed would provide an insightful follow-up to this analysis.Moreover, performing mitochondrial isolation in both sampled stages may be a better method to account for variation in mitochondrial yield between the sampled life cycle stages.This approach has been validated in mammalian cells where it was found that analyzing isolated mitochondria rather than whole cells is better for revealing mitochondrial-related metabolite changes and depicting mitochondrial dynamics. 38s to note we cannot distinguish aberrant phenotypes are direct result of lack or excess of CKs produced through knockout or overexpression of the iptA gene or whether the lack or excess of the IptA protein impacts biological processes beyond its catalytic role.A comparable example of this comes from Frej et al. 74 who highlighted distinct roles of the biosynthetic enzyme, inositol-3-phosphate synthase, in D. discoideum, beyond its catalytic role that were not rescued by inositol supplementation.As such, supplementation or depletion of CKs in the iptA-deficient and overexpression strains, respectively, would be necessary to distinguish which of the observed phenotypes are associated with a lack/excess of CKs versus a lack/excess of IptA.Determining first whether the observed phenotypes are a result of the lack/excess of CKs versus the lack/excess of IptA will be informative in future experiments.Last, it is interesting that CKs do not affect growth in D. discoideum in a manner similar to plants despite possessing an adenylate isopentenyltransferase gene.It appears that there is no functional redundancy of CKs across CK-producing organisms, and rather the roles of CKs are tailored to the organism that produces them.

| CONCLUSION
This is the first D. discoideum study to report on an IptA overexpression strain and to conduct a metabolomics analysis in response to altered CK dynamics.We hypothesized that a shift in CK forms, observed during growth and early development, reflected distinct functions of CKs during the different stages of the D. discoideum life cycle.Supporting this hypothesis, here we generated iptAdeficient and overexpression strains to document separate roles for CKs during vegetative growth from those identified during the later stages of multicellular development.While there are no visible macro-scale phenotypes from loss or overexpression of iptA during vegetative growth, such as those seen in late development, we documented several biochemical changes at the cellular and molecular levels.We conclude that iptA deficiency affects cytokinesis and both iptA deficiency and overexpression result in subcellular phenotypes consistent with mitochondrialassociated dysfunction, including altered mitochondrial morphology, dysregulated TCA cycle, and amino acid metabolism, and increased AMP levels during vegetative growth.
Figure created with BioRender.com.

F I G U R E 2
Knockout of iptA results in a 99% reduction of total cytokinin (CK) levels.(A) iptA schematic indicating the CRISPR guide RNA cut site (gRNA; red) showing the region of the gene that was targeted for knockout (KO) and the genomic primers used (yellow) for verification of diagnostic iptA KO products.Fwd indicates forward primer, and Rv indicates reverse primer.(B) Schematic of the selected iptA − clone used in all experiments.This clone has a 32 base pair deletion from nucleotides 118-149 of the genomic iptA sequence, which results in a shift in the reading frame and early truncation of the gene product.The translation of the truncated iptA − gene product is shown beneath the cartoon iptA − truncated product; the red line underneath the translation indicates where the sequence deviates from wild type (WT) because of the deletion (see nucleotide and amino acid sequence in Table

F I G U R E 3
iptA deficiency results in aberrant fruiting body morphology.(A) Visual of aberrant iptA − fruiting body morphology with decreased sori size and thin stalks.WT fruiting bodies are shown in the left image, and iptA − fruiting bodies are shown in the right image.Red arrows highlight the sori and stalks in the fruiting bodies, which are smaller in the knockout.The scale bar represents 1 mm.(B) Disruption of iptA generates fruiting bodies containing smaller sori, (C) fewer fruiting bodies formed per number of cells plated, and (D) an overall reduction in sori reaching the average area of WT (AX3).Fiji/ImageJ was used to assess sori area.Data are presented as individual values (dots) and mean ± standard error for five independent experiments, and statistical significance was determined by a t-test (two-tailed; *p < .05,**p < .01).

F I G U R E 4
iptA deficiency does not affect vegetative growth or pinocytosis.(A) The loss of iptA has a minimal effect on vegetative growth in axenic culture under nutrient-rich conditions (HL5) and (B) minimal nutrient conditions (FM).Cell density was determined every 24 h using a hemocytometer.(C) iptA − cells grow similar to WT (AX3) on agar plates seeded with bacteria, as measured through plaque expansion (mm) over an 84-h time course.(D) Relative pinocytosis rates are unaffected by the loss of iptA.WT and iptA − cells were incubated in HL5 containing FITC-Dextran over a 90-min time course, and the uptake of FITC-Dextran was assessed every 15 min.The data were corrected for background signal and expressed as the mean fluorescence change (%) relative to the 0-min time point.(E) Relative exocytosis was assessed using the same cell culture over a 90-min time course.The 2-h time point was normalized to the time 0 reading for each strain.For all panels, the data are presented as mean ± standard error for three independent experiments, and statistical significance comparing iptA − to WT (AX3) was determined by an unpaired t-test (two-tailed; **p < .01).
life cycle stage sample (three life cycle stages assessed: FM minimal medium [vegetative amoebae], FM minimal medium without arginine and lysine [early development], and 24-h development [fruiting bodies]).We first performed a heatmap analysis of the top 400 significantly different features for each respective treatment type.This analysis showed clustering of the individual replicates for each strain, indicating consistency of strain effects and thus F I G U R E 5 Loss of iptA impairs cytokinesis.(A) Representative images from WT (AX3) and iptA − cells stained with DAPI nuclear stain to assess cytokinesis.Red arrows indicate multinucleated cells.The scale bars represent 10 μm.(B, C) The number of nuclei was quantified under nutrientrich conditions (HL5; B) and minimal nutrient conditions (FM; C).Data are presented as individual values (dots) and mean ± standard error for three independent replicates, where at least 300 cells were analyzed per condition.Statistical significance was determined by an unpaired t-test with the Bonferroni-Dunn multiple comparisons test (twotailed; *p < .05,**p < .01,***p < .001).

F I G U R E 6
iptA deficiency alters the morphology and numbers of mitochondria during vegetative growth.(A) TEM images of representative WT (AX3) and iptA − vegetative amoebae cultured in FM minimal medium for 36 h showing altered mitochondria morphology in the iptA − strain compared to WT.The inset shows a zoomed-in section of the mitochondria highlighting the rounded mitochondrial morphology present in the iptA − strain.(B) iptA − cells have a greater number of mitochondria per cell and (C) the mitochondria area is significantly more rounded/circular.For panels B and C, data are presented as individual values (dots) and mean ± standard error for two independent experiments, and statistical significance was determined by an unpaired t test (two-tailed; *p < .05,****p < .0001).The larger scale bars on the bottom left of the TEM images represent 1 μm, and the smaller scale bars on the bottom right panels represent 200 nm.Fiji/ ImageJ was used to assess circularity.
cycle.Metabolites with trends of interest in the glycolysis or TCA cycle are highlighted in boxes on the sides of the figure showing relative normalized metabolite content.Data from the individual metabolite plots were relative to labeled amino acid standards added to each sample and are presented as individual values (dots) and mean ± standard error for four independent replicates.Statistical significance comparing iptA − and GFP-IptA strains to WT for all quantified metabolites was determined through a two-way ANOVA with the Benjamini, Krieger, and Yekutieli correction controlling for multiple comparisons with false discovery rate (FDR) of <.05-pvalues are listed above the bar graph with a line spanning between WT and iptA − or between WT and GFP-IptA to indicate a significant difference in metabolite abundance.Solid red boxes around the individual metabolite plots indicate the metabolite was validated by MS 2 data and have no known isomers, and dashed blue boxes indicate the metabolite has an isomer that was indiscernible by the LC method used (isomer names included in the box titles).
Statistical significance was determined through a two-way ANOVA with the Tukey multiple comparisons test (*p < .05,**p < .01,***p < .001,****p < .0001).For panel B, AMP content was determined through the LC-MS/MS metabolomic analysis and was normalized relative to labeled amino acid standards and individual metabolite values.Data are presented as individual values (dots) and mean ± standard error of the mean for four independent replicates, and statistical significance was determined through a one-way ANOVA with Dunnett's multiple comparison tests (**p < .01,***p < .001).

F I G U R E 1 0
Amino acid and derivative levels decrease in iptA-deficient and overexpression strains in vegetative growth and 24-h fruiting body stages.(A) Amino acid and amino acid derivative metabolite levels detected in vegetative amoebae cultured in FM minimal medium, (B) in early development amoebae cultured in FM minimal medium without arginine and lysine, and (C) in 24-h development fruiting bodies cultured on 1% KK2 agar.High-performance liquid chromatography-high resolution accurate mass-full scan-mass spectrometry (HPLC-(HRAM)-FS-MS) was used to detect amino acids and amino acid derivatives, which are separated by a dashed line on each figure.Individual metabolite levels were normalized relative to labeled amino acid standards added to each sample and are presented as individual values (dots) and mean ± standard error for four independent replicates.Statistical significance comparing iptA − and GFP-IptA strains to WT (AX3) for all detected metabolites was determined through a two-way ANOVA with the Benjamini, Krieger, and Yekutieli correction controlling for multiple comparisons with false discovery rate (FDR) of <.05 -p values are listed above the bar graph with a line spanning between WT and iptA − (blue font) or between WT and GFP-IptA (pink font) to indicate a significant difference in metabolite abundance.Isomers that were indiscernible by the LC method used are both listed under the metabolite plots and are separated with a slash.