Articles Cite This: ACS Chem. Biol. XXXX, XXX, XXX−XXX
The Chemoattractant Glorin Is Inactivated by Ester Cleavage during Early Multicellular Development of Polysphondylium pallidum Daniel Heinrich,† Robert Barnett,‡ Luke Tweedy,§ Robert Insall,§ Pierre Stallforth,‡ and Thomas Winckler*,† †
Pharmaceutical Biology, Institute of Pharmacy, Friedrich Schiller University, Jena, Germany Junior Research Group ‘Chemistry of Microbial Communication’, Leibniz Institute for Natural Product Research and Infection Biology − Hans Knöll Institute, Jena, Germany § Cancer Research UK Beatson Institute, Glasgow, United Kingdom ‡
ABSTRACT: Among the amoebozoan species capable of forming fruiting bodies, the dictyostelid social amoebae stand out since they form true multicellular organisms by means of single cell aggregation. Upon food depletion, cells migrate across gradients of extracellular signals initiated by cells in aggregation centers. The model species that is widely used to study multicellular development of social amoebae, Dictyostelium discoideum, uses cyclic adenosine monophosphate (cAMP) as a chemoattractant to coordinate aggregation. Molecular phylogeny studies suggested that social amoebae evolved in four major groups, of which groups 1 and 2 are paraphyletic to groups 3 and 4. During early development, intercellular communication with cAMP appears to be restricted to group 4 species. Cells of group 1 and 2 taxa do not respond chemotactically to extracellular cAMP and likely use a dipeptide chemoattractant known as glorin (N-propionyl-γ-L-glutamyl-L-ornithin-δ-lactam-ethylester) to regulate aggregation. Directional migration of glorinresponsive cells requires the periodic breakdown of the chemoattractant. Here, we identified an extracellular enzymatic activity (glorinase) in the glorin-responsive group 2 taxon Polysphondylium pallidum leading to the inactivation of glorin. We determined the inactivation mechanism to proceed via hydrolytic ethyl ester cleavage of the γ-glutamyl moiety of glorin. Synthetic glorinamide, in which the ethyl ester group was substituted by an ethyl amide group, had glorin-like biological activity but was resistant to degradation by glorinase. Our observations pave the way for future investigations toward an ancient eukaryotic chemotaxis system.
T
positive genetic and biochemical feedback loops.5 Another hallmark of D. discoideum aggregation is the oscillatory secretion and degradation of cAMP in ∼6 min intervals. When full aggregation competence is achieved, D. discoideum cells respond to nanomolar concentrations of cAMP, and they can relay the signal in the field of starving amoebae.9 The enzymes that break down cAMP, in particular the secreted phosphodiesterase DdPDE1, are important for both induction of developmental genes and chemotaxis.10 In their absence cAMP builds up to excessive levels, preventing the pulsatile nature of the signals that control gene expression being perceived and causing chemotactic stimuli to become dulled as receptors saturate at the rear of the cell as well as the front. Breakdown of chemotactic stimuli has assumed an even greater importance with the discovery of self-generated gradients.11 While these are most often described in cases where the attractant is initially homogeneous, recent data have shown that chemoattractant breakdown is important for both sharpening gradients and bringing the attractant levels down
he development of multicellularity is considered to be a major evolutionary transition.1 Multicellularity has originated several times during the evolution of eukaryotes and can occur either by clonal or aggregative development.2 The former starts from a single cell, and daughter cells remain together after cell division by means of specific cell−cell adhesion mechanisms.3 In aggregative development, cells disperse upon cell division and assemble sporadically into aggregates. The latter may be triggered, for instance, in response to stress, since multicellular fruiting bodies containing spores allow survival of unfavorable conditions.4 Aggregation of dictyostelid amoebae is a highly regulated process, in which secreted diffusible signals known as “acrasins” act as chemoattractants and coordinate directional cell movement.5,6 In the best-studied taxon, D. discoideum, aggregation is mediated by cAMP, which acts both extracellularly as a chemoattractant and morphogen and intracellularly as a second messenger.4,5,7,8 A salient feature of the cAMPbased signaling system is that most of its components, including cAMP receptors, adenylyl cyclases, and intracellular and extracellular phosphodiesterases that break down cAMP, are expressed at very low levels in growing cells. Yet upon starvation, these components are strongly induced by means of © XXXX American Chemical Society
Received: January 15, 2018 Accepted: May 17, 2018
A
DOI: 10.1021/acschembio.8b00046 ACS Chem. Biol. XXXX, XXX, XXX−XXX
Articles
ACS Chemical Biology
First, P. pallidum cells were washed free of food bacteria and cells were starved in a phosphate buffer for 1−4 h. Then, aliquots of cell suspensions or cell-free supernatants were spiked with glorin and incubated for various time periods to allow enzymatic glorin degradation. Reactions were performed in a phosphate buffer (pH = 6.2) and terminated by boiling the samples at 95 °C for 10 min. Glorin eluted from a C18 reversephase HPLC column at a retention time (tR) of 7.8 min (Figure 2). In the absence of P. pallidum cells or supernatant, glorin was stable under the assay conditions and not degraded by boiling (data not shown).
below the receptors’ Kd so they can be precisely perceived. Even in common assays (for example, Zigmond chamber assays12) where cells are directly exposed to gradients, chemoattractant breakdown makes chemotaxis markedly more efficient.13 The dictyostelid social amoebae form a monophyletic clade that can be divided into two paraphyletic branches and four major groups (groups 1 and 2 and groups 3 and 4, respectively;14−18 Figure 1). These phylogenetic studies
Figure 1. Phylogeny of dictyostelids. Consensus phylogenetic tree calculated from alignments of 47 concatenated proteins by Singh et al.18 Probabilities were 1.0 for all nodes. Adapted with permission from Singh et al.18 Note that dictyostelids evolved in two paraphyletic branches with groups 1 and 2 as sister clades to groups 3 and 4.
Figure 2. Glorin degradation by starving P. pallidum cells. Cells were starved in phosphate buffer for 1 h. Aliquots of cell-free supernatants were then spiked with glorin and incubated for the indicated time periods.
suggested that all species known to use cAMP as chemoattractants during aggregation belong to group 4. However, knowledge of other, allegedly more ancestral communication systems that regulate aggregation in dictyostelids is scarce. About 40 years ago, it was reported that Polysphondylium violaceum, which is a group-intermediate taxon and outgroup to group 4,17,18 secretes the dipeptide chemoattractant glorin.19,20 Early work in P. violaceum suggested that the glorin-based signaling system might be quite different from the well-known cAMP-based communication system, since parts of the glorinbased communication system were already present in growing cells and were not notably induced upon starvation.21 We have previously shown that several group 1 and 2 organisms, including Polysphondylium pallidum, are able to migrate chemotactically in self-generated glorin gradients, suggesting that they secrete a “glorinase” to break down glorin and sharpen the chemotactic signal.22 In this study, we identified the mechanism of glorinase-induced glorin inactivation.
To determine metabolites generated by glorinase activity, suspensions of starving P. pallidum cells were incubated with glorin in a time frame of minutes to several hours. Under our assay conditions, glorin was quantitatively converted to a metabolite 1 (M1, tR = 5.5 min) within 60−120 min (Figure 2). If samples were incubated for a further several hours, M1 was slowly degraded to a metabolite M2 (tR = 3.5 min; data not shown). When P. pallidum cells were starved for 3 h in a shaken suspension and then separated into cell-free supernatant and washed cells, ∼95% of total glorinase activity was secreted in soluble form, and only ∼5% was cell-bound. Freshly washed cells secreted detectable glorinase activity into the buffer supernatant after 15−30 min, reaching a plateau approximately 4 h after the onset of starvation (Figure 3). The observed steady-state of glorinase accumulation after ∼4 h may suggest that the cells ceased glorinase secretion after a few hours of starvation and that the enzyme in the extracellular fluid remained stable. Alternatively, a steady-state between the rate of glorinase secretion and its extracellular degradation may have been observed. We could not reliably determine if growing P. pallidum cells also secrete glorinase, as it proved unfeasible to reproducibly enrich enzymes from cells growing in the presence of bacteria. This was because it was unclear whether cells growing in phosphate buffer in association with bacteria were in fact in the vegetative state or whether some cells sensed starvation and therefore started to secrete glorinase at a very low level, which impaired the interpretation of results.
■
RESULTS Glorinase Activity in Starving P. pallidum Cells. Although the secretion of glorin has been shown so far only in P. violaceum, it is reasonable to assume that P. pallidum cells use glorin during aggregation since they respond chemotactically to the signal and degrade it; hence, they must secrete a glorinase.22 We developed a glorinase assay to evaluate the metabolism of extracellular glorin during early development of P. pallidum cells. We chose P. pallidum as a model species in our experiments due to the availability of the annotated reference genomea prerequisite for subsequent proteomic approaches aimed at identifying components of the glorinbased communication system. B
DOI: 10.1021/acschembio.8b00046 ACS Chem. Biol. XXXX, XXX, XXX−XXX
Articles
ACS Chemical Biology
glorinase furthermore showed that conversion to M1 was virtually quantitative. Nonenzymatic production of M1 under our assay conditions could be ruled out, since glorinase assays performed with boiled cell supernatants did not reveal any M1 formation, whereas glorin was stable under these assay conditions. Three lines of evidence support our conclusion that enzymatic ester cleavage at the α-carboxyl of the γ-glutamyl moiety is the main route of glorin degradation. First, we compared the isolated enzymatically produced M1 with synthetic M1 produced by saponification of synthetic glorin under alkaline conditions. The two M1 samples showed identical elution in HPLC (Figure 5), and HRMS of synthetic M1 (m/z = 300.1555 [M + H]+) was identical to M1 isolated after enzymatic conversion of glorin.
Figure 3. Developmental regulation of glorinase activity. Cells were starved in phosphate buffer for the indicated time periods. Aliquots of cells were centrifuged, and glorinase activity was measured in the cellfree buffer supernatant of the cells (filled dots; solid line, undiluted samples; dashed line, samples diluted 1:10). Cells were washed once in phosphate buffer and resuspended to the original volume and immediately used for glorinase assays. Open squares represent glorinase activity bound to the cells. The experiment was performed twice with similar results.
Nevertheless, a very low level of cell-bound glorinase was detectable in freshly harvested amoebae/bacteria suspensions (i.e., without starvation time due to washing the cells), suggesting that some cell-bound glorinase activity is present in growing cells even if it is clearly induced in response to starvation. The cell-bound glorinase activity increased by 3-fold during the first hour of starvation and then declined but still remained detectable thereafter (Figure 3). Glorinase Hydrolyzes the Ethyl Ester Group of Glorin. We applied high-resolution mass spectrometry (HRMS) to identify metabolite M1. The mass of M1 (m/z = 300.1553 [M + H]+) revealed a mass difference of Δm/z = 28 to glorin (m/z = 328.1863 [M + H]+), suggesting the loss of the ethyl group of the γ-glutamyl moiety of glorin to yield M1 (C13H21N3O5; MW 299.3229) as suggested in Figure 4. Metabolite M2 matched the propionyl amide-cleaved derivative of M1 (m/z = 244.1291 [M + H]+ calcd for C10H17N3O4; MW 243.2603). The chromatographic profile of glorin upon treatment with
Figure 5. HPLC analysis of purified M1. The metabolite M1 was enzymatically produced from glorin using a supernatant of starving P. pallidum cells as a source for glorinase. In parallel, M1 was generated chemically by saponification of glorin. Both M1 products were purified by preparative HPLC. The retention times of both M1 preparations were then compared on an analytical RP-18 HPLC column.
Second, metabolites of glorin resulting from opening of the ornithine δ-lactam, M3 (MW 345.4) and M5 (MW 317.3), could only be detected in our assays in trace amounts (
