Modeled microgravity alters apoptotic gene expression and caspase activity in the squid-vibrio symbiosis

Identification of the extrinsic and intrinsic apoptosis network within the host Euprymna scolopes

To elucidate the apoptosis machinery in the bobtail squid, a detailed map of the putative pathways for apoptosis in the squid host was generated by data-mining the reference transcriptome and genome of E. scolopes (Fig. 2). This search identified 293 transcripts, representing 137 unique genes, involved in the regulation of apoptosis in the host squid (Table S1; Fig. S1). The resultant network revealed an elaborate web of parallel and hierarchical interactions and pro-and anti-apoptotic effectors that appear to govern extrinsic, receptor-mediated, and intrinsic, stress-induced cell death via transcriptional, translational, and post-translational means (Fig. 2).

Fig. 2figure 2

Putative pathways for apoptosis in Euprymna scolopes. The candidates identified from the genome and reference transcriptome included effectors of both extrinsic/receptor-mediated, and intrinsic/stress-associated, apoptosis. Representative interactions were mapped by cross referencing the multi-species KEGG pathway for apoptosis (ko04215) with peer-reviewed literature

For the extrinsic apoptosis network, two tumor necrosis factor receptor superfamily members (Tnfrsf11, Tnfrsf21) and five Tnf-associated factors (Traf 2, 3, 4, 6, and 7) were discovered in E. scolopes (Fig. 2, Table S1). The closely related initiator caspases-8 and -10 were also identified, as well as death domain-associated adaptors Fadd, Daxx, and the stress-associated kinase Ripk1. Interestingly, neither Fas ligand nor its corresponding receptor was found in the reference transcriptome despite the presence of components known to operate downstream (e.g., Fadd). Nevertheless, E. scolopes was determined to harbor all the effectors required to assemble a complete death-inducing signaling complex (DISC) for extrinsic, receptor-mediated apoptosis, which serves as the activation assembly for extrinsic initiator caspases-10 and -8 [27]. Additionally, transcripts associated with three LPS-binding proteins (Lbps) and the LPS-induced transcription factor, Litaf, which have been previously described in E. scolopes [28] were also observed in the putative extrinsic, receptor-mediated network (Fig. 2).

With regards to the intrinsic apoptosis pathway, transcripts encoding both pro-apoptotic (Bnip3, Bcl10, Bcl2L13, Bax, Bak) and anti-apoptotic (Bcl2L1, Mcl1, Ar1) members of the Bcl2 protein family were found in E. scolopes (Fig. 2; Table S1). A variety of mitochondrial proteins (i.e., Aifm1, Aifm3, Diablo, Cytc, Htra2) were identified as well as Apaf1 and the intrinsic initiator caspase-9. Thus, all the requisite parts of the apoptosome, which is the activation complex of the stress-induced pathway for cell death were found in the bobtail squid (Fig. 2). However, the Bh3-only protein Bid, which typically represents a critical junction for crosstalk between the intrinsic and extrinsic pathways for apoptosis in animals, was conspicuously absent. Likewise, the presence of atypical initiator caspase-2 was confirmed along with Pidd1, both of which are constituents of the pro-death, p53-inducible, PIDDosome complex that forms in response to DNA damage [29]; however, no evidence of the third member of this assembly, Raidd/Cradd, was found in the bobtail squid reference transcriptome or genome.

In addition to Bcl2-type repressors, transcripts encoding numerous inhibitors of apoptosis proteins (Iaps) were discovered in the host squid. Included among these were Xiap, Diap2, three Baculoviral Iap-repeat containing proteins (Birc7b, Birc3, and Birc6), three Bcl2-associated athanogene regulators (Bag 1, 4, and 6), cell death regulator Aven, the peptidyl isomerase Pin1, as well as heat shock proteins Hsp90 and Hsp7c (Table S1). Further examination revealed the presence of numerous cathepsin (B, K, L, L1, L2, Z) and calpain (1–3, 5, 7–9, 11, B, D) proteases, several of which have been previously reported in E. scolopes [30]. The cathepsins recovered from the reference transcriptome were primarily cysteine-dependent, except for cathepsin B, which is a serine protease. Intriguingly, neither aspartic acid-type cathepsins (D and E) were found in the host squid.

Analysis of the E. scolopes reference transcriptome revealed a variety of nucleus-localizing proteins affiliated with apoptosis (Table S1). Broadly, this included transcription factors (i.e., Aatf, Atf4, Jun, E2f, Hox, Irf, Litaf, NF-κB, p73), p53-type transactivators (i.e., Aspp1, Ccar1), and apoptotic mediators of chromatin condensation and DNA fragmentation (i.e., Acin1, Dffb/Cad, EndoG) (Table S1). Transcripts encoding numerous regulators of apoptotic translation were likewise present in the squid, most notably the eukaryotic initiation factors Eif2, Eif3f, Eif3j, Eif4b, and Eif4e (Table S1). Interestingly, however, no transactivating factors of internal ribosomal entry site (IRES)-mediated translation were present in the transcriptome of E. scolopes, and the nuclear lamin protease and executioner caspase, caspase-6, was not observed.

Extrinsic and intrinsic apoptosis gene expression under LSMMG conditions

To ascertain the impact of modeled microgravity on the light organ cell death event in the host animal, 41 apoptosis-related genes identified in the reference transcriptome were selected for gene expression analysis using the NanoString nCounter assay (Table 1). These target genes were representative of key components of the apoptosis machinery depicted in Fig. 2. The expression of these target transcripts was quantified over time in hatchling paralarvae incubated within the HARVs under LSMMG and gravity control conditions in the presence and absence of the symbiotic bacterium V. fischeri ES114. Overall, the NanoString results indicated that there was an increase in the transcription of genes associated with both receptor-mediated and stress-induced apoptosis under LSMMG conditions compared to gravity controls (Fig. 3, 4).

Table 1 NanoString CodeSet for apoptosis genes in Euprymna scolopesFig. 3figure 3

Extrinsic and intrinsic gene expression in symbiotic hatchlings in gravity controls relative to LSMMG. Expression of extrinsic, receptor-mediated apoptosis genes a Lbp1, b Lbp3, c Litaf, and d Fadd. Expression of intrinsic, stress-induced apoptosis genes e Bnip3, f Bak, g Diablo, and h Aifm3. Expression is conveyed as log2 fold-change (log2FC). Positive log2FC values denote higher expression in the gravity control group (solid bars). Negative log2FC values indicate up expression in LSMMG (hatched bars). Error bars are the standard error of the mean. Asterisks denote significant differences between the conditions (* = p ≤ 0.10, ** = p ≤ 0.05)

Fig. 4figure 4

Apoptotic caspase expression in symbiotic hatchlings under gravity and low shear modeled microgravity (LSMMG) conditions. Heatmaps representing the transcriptional expression of pro-death caspases -2, -3, -8, -9, and -10 genes in gravity (left) and LSMMG (right) conditions. Per the color scale, red indicates a negative Z-score and lower-than-average expression, whereas green signifies a positive Z-score and higher-than-average expression

With regards to effectors of extrinsic, receptor-mediated, apoptosis, the LPS binding proteins 1 and 3, along with the endotoxin-induced transcription factor Litaf were significantly up-expressed in modeled microgravity between 6 and 10 h post-hatching (Fig. 3). Likewise, transcripts encoding the adaptor protein Fadd were more abundant during the early hours of development in LSMMG; however, these results were not statistically significant from the gravity controls (Fig. 3).

A similar trend was observed in the NanoString results for genes affiliated with intrinsic, stress-induced apoptosis. The Bh3-only protein Bnip3, membrane permeabilizer Bak, Diablo, and caspase-dependent Aifm3 were significantly up expressed in LSMMG between 8 and 12 h compared to gravity controls (Fig. 3). Bnip3, which overcomes Bcl-2 type suppression of pro-death effectors Bax/Bak, increased overtime under LSMMG peaking at 16 and 18 h in LSMMG compared to the unit gravity controls (Fig. 3). Bnip3 is known to be induced under hypoxic conditions, however, measurements of the dissolved oxygen content were consistent across all treatments at 24 h revealing no difference in oxygen availability in the HARVs (Fig. S2).

Expression of pro-death caspases in host animal under LSMMG conditions

The identification of caspase-2, -3, -7, -8, -9, and -10 transcripts in the reference transcriptome suggested the presence of canonical pathways for extrinsic and intrinsic apoptosis in E. scolopes (Fig. 4). Additionally, caspase-driven apoptosis has a known role in the bacteria-induced development of the light organ, given that treatment with the pan-caspase inhibitor Z-VAD-FMK has been shown to reduce the number of dying cells observed at 24 h post-inoculation [30]. These results suggested that caspase regulation, either at the level of mRNA and/or protein, was a contributing factor to the accelerated developmental timeline of apoptosis in the host light organ under LSMMG conditions (Fig. 1d).

NanoString analysis was used to examine the differential gene expression of the putative initiator (i.e., casp2, casp8, casp9, and casp10) and executioner (i.e., casp3) caspase genes in symbiotic and aposymbiotic squid under microgravity and gravity conditions. Results revealed that under LSMMG conditions there was a pronounced increase in the pro-death caspase transcripts between 6 and 10 h post-colonization by V. fischeri (Fig. 4), which aligns with early-onset apoptosis previously observed in paralarvae squid exposed to modeled microgravity [15]. Aposymbiotic control animals that were not exposed to V. fischeri showed no differential expression under LSMMG or gravity conditions at these same time points (Fig. S3).

The architecture of caspase domain structure in Euprymna scolopes

To evaluate the extent of similarity between the E. scolopes caspase (EsCasp) enzymes and homologs in other animals, the domain architecture of EsCasps was mapped and compared to sequences derived from Homo sapiens, Rattus norvegicus, Danio rerio, and Xenopus laevis (Fig. 5). Additionally, the Mediterranean mussel, Mytilus galloprovincialis, and the East Asian octopus, Octopus sinensis, were included in the analysis to distinguish lineage-specific modifications within the Mollusca phylum and/or among cephalopods (Fig. 5, Additional File 3). Domain analysis via SMART, InterProScan, and the NCBI Conserved Domain Database, as described in the Methods, revealed that the architecture of caspases-2, -3, -7, -8, and -9 in E. scolopes is largely consistent with that of other animals, including both vertebrates and invertebrates alike (Fig. 5). Typically, the EsCasps ranged from 327 to 634 residues in length and weighed between 41.4 and 72.1 kDa (Table 2). The main exception to this was EsCasp10_2X, which was only 297 amino acids long and had a molecular weight of 34.4 kDa (Table 2).

Fig. 5figure 5

Protein sequence-based analysis of modular domain architecture for the initiator and executioner caspases of E. scolopes, H. sapiens, R. norvegicus, D. rerio, X. laevis, M. galloprovincialis, and O. sinensis. The X-axis indicates residue position in the primary peptide sequence. Red boxes represent death effector domains (DED) whereas green boxes are caspase activation and recruitment domains (CARD). The 45 kDa CASc precursor is illustrated by a yellow box, with the resulting p20 and p10 subunits shaded medium and dark blue, respectively. Variants of squid caspase enzymes are indicated by the notation nX where n is the isoform number

Table 2 Initiator and executioner caspases in Euprymna scolopes

Several EsCasps exhibited key structural elements characteristic of caspases in other animals (Fig. 5). For example, the extrinsic initiator caspase-8 of E. scolopes (EsCasp8) exhibited dual twin death effector domains (DED) at the N-terminus, which is characteristic of both vertebrates and invertebrate caspase-8 proteins; however, this dual DED domain was missing in the initiator EsCasp10 isoforms when compared to vertebrates H. sapiens, D. rerio, and X. laevis (Fig. 5). EsCasp10 did harbor a singular caspase activation and recruitment domain (CARD), suggesting a potential misannotation.

The CARD domain was also observed in EsCasp2 and -9 initiator caspases, which is typical of most vertebrates and invertebrates (Fig. 5). Interestingly, caspase-9 in M. galloprovincialis (MgCasp9) exhibited dual DED in place of the traditional CARD, suggesting a potential misannotation of this protein in M. galloprovincialis. In all sequences examined, a carboxyl-terminal catalytic (CASc) domain was present (Fig. 5, yellow box). The CASc is a 45 kDa precursor that is cleaved during activation to yield the p20 and p10 subunits. Intriguingly, despite the presence of a complete CASc, analysis of EsCasp2_2x and EsCasp9 revealed the smaller p10 subunit was not detected in these two isoforms.

Executioner caspases-3 and -7 in animals typically lack pro-domains and are therefore shorter as a result (Fig. 5). The executioner caspases of E. scolopes (EsCasp3 and -7) varied from 213 to 370 residues in length and had a molecular weight ranging from 23.9 to 41.5 kDa (Table 2; Table S2), thus were comparable to both the vertebrate and invertebrate species targeted in this analysis. Most EsCasp3 and -7 isoforms exhibited a conserved QACRG pentapeptide in the catalytic region or equivalent (Table 2).

Phylogenetic analysis of host squid caspases

Maximum likelihood analysis of the initiator caspases revealed two distinctive clusters primarily consisting of caspases-8, -9 and -10, and caspase-2, respectively (Fig. 6a), that were reflective of the function of whether the initiator caspases were activated by extrinsic or intrinsic triggers. EsCasp8 clustered with all other caspase-8 proteases sourced from NCBI but was closest to proteases from O. sinensis. Similarly, EsCasp2 also branched closely with both invertebrate and vertebrate caspase-2 initiators.

Fig. 6figure 6

Maximum likelihood phylogenetic analysis of the caspase enzymes in E. scolopes, H. sapiens, R. norvegicus, D. rerio, X. laevis, M. galloprovincialis, and O. sinensis. a Unrooted phylogenetic tree for initiator caspase-2, -8, -9, and -10. b Unrooted phylogenetic tree for executioner caspase-3 and -7. Both trees were generated assuming the WAG model of amino acid substitution, with 1000 bootstrap iterations, in MEGA X. Branch support values are expressed as percentages. Sequences are labeled with the first letter of the corresponding genus and species (e.g., Euprymna scolopes, Es)

Unexpectedly, EsCasp9 clustered with the two sequences that were likely misannotated as isoforms of caspase-10. Although the monophyly of the group is indeterminate due to the low outer branch support value (42%), these sequences share several distinguishing features that suggest they may represent a group of caspase-9-like initiators that is specific to Euprymna (Fig. 6a). First, the domain organization of EsCasp9, EsCasp10_1X, and EsCasp10_2X emulates that of canonical caspases -2 and -9, with an N-terminal CARD domain and a C-terminal CASc region (Fig. 5). Second, although the sequences of this group exhibit 94 to 95% identity with one another, their percent identity scores with the other initiator caspases are very low, between 15 and 29%, which indicates they may have distinct structural features (Fig. S4). Third, EsCasp9, EsCasp10_1X, and EsCasp10_2X all exhibit a conserved “QACQP” catalytic pentapeptide (Table 2). Caspases, uniformly, possess a QACXG motif in the active site, with the middle position specifying the catalytic cysteine residue [31]. Given its crucial importance to the protease activity of caspases, variations within this pentapeptide are limited and may denote similar functions.

The maximum likelihood estimation of executioner caspases formed two several distinctive groups primarily separated into vertebrate and invertebrate clusters (Fig. 6b). For example, the vertebrate caspases-7 and -3 from H. sapiens, R. norvegicus, D. rerio, and X. laevis formed distinctive groupings, that excluded most of the molluscan isoforms. Interestingly, two caspase-7 isoforms from E. scolopes (EsCasp7_2X and EsCasp7_4X) had a higher association with the vertebrate executioner caspases than the other molluscan orthologs for caspase-3 and -7 (Fig. 6b). This phylogeny may allude to a greater degree of differentiation, or perhaps expansion, among the executioner caspases of the squid, which has previously been documented in bivalves [32, 33]. However, the percent identity among the executioners was 47% which makes it difficult to discern whether these sequences represent isoforms of the same gene(s) or an expansion of caspases-3 and -7 in E. scolopes (Fig. S4).

Overall, there appeared to be a relatively high degree of similarity among the executioner caspases of E. scolopes. However, the close affiliation of EsCasp7_2X and EsCasp7_4X with vertebral orthologs suggests the presence of at least two distinct executioner-type caspases in the squid. Unsurprisingly, both the initiator and executioner caspases of E. scolopes were most closely related to O. sinensis, and to a lesser extent, M. galloprovincialis, which is consistent with their categorization as both soft-bodied cephalopods and, more broadly, mollusks.

Caspase activity and mitigation during the onset of symbiosis and light organ morphogenesis under LSMMG and gravity conditions

The activities of executioner caspases-3/7, extrinsic initiator caspase-8, and intrinsic initiator caspase-9 were quantified in host animals in the presence and absence of the symbiont V. fischeri using Caspase Glo activity kits. Note, due to the similar mechanisms of activity, the kits could not resolve between caspase-3 and caspase-7 activities. All caspases tested were significantly more active in symbiotic animals at 16 h post-colonization, the peak of cell death [19], compared to 16 h aposymbiotic controls and newly hatched paralarvae (Fig. 7). In all three treatments, there was a 26-fold increase in executioner caspase activity (Fig. 7a) compared to the extrinsic and intrinsic initiators (Fig. 7b, c).

Fig. 7figure 7

Caspase activity and protease inhibition during bacteria-induced apoptosis in the normal squid-vibrio symbiosis. The activity of a executioner caspase-3/-7, b extrinsic initiator caspase-8, and c intrinsic initiator caspase-9 was measured in the light organs of hatchling, 16 h aposymbiotic (apo), and 16 h symbiotic (sym) squid. d Apoptosis was quantified at 16 h after protease inhibitor treatment including pan-caspase inhibitor z-VAD-FMK (ZVAD), the caspase 8 inhibitor Ac-IETD-CHO (C8i), caspase 9 inhibitor Ac-LEHD-CMK (C9i), Pefabloc (Pefa) and dimethyl sulfoxide (DMSO) controls. Error bars are the standard error of the mean. Asterisks denote significant differences between the datasets test (* = p ≤ 0.05, ** = p ≤ 0.01). Comparisons that were not significant are labeled “ns”

To test whether the caspases could be inhibited in the host animals, pre-treatment with inhibitors targeting different caspases was used. Treatment with the caspase-8 inhibitor Ac-IETD-CHO, the pan-caspase inhibitor Z-VAD-FMK, and the serine protease inhibitor Pefabloc for two hours prior to the start of the symbiosis significantly reduced apoptotic cell levels at 16 h in symbiotic hatchlings compared to untreated symbiotic controls but did not completely inhibit apoptosis (Fig. 7d). Interestingly, although the Caspase-Glo assays indicated caspase-9 was active during the apex of cell death (Fig. 7c), treatment with the caspase-9 inhibitor Ac-LEHD-CMK did not reduce the level of pycnotic nuclei observed at 16 h relative to the untreated or DMSO-treated controls (Fig. d). Additionally, the coupling of different inhibitors showed no additive effects (Fig. 7d). The colonization of the host squid did not appear to be affected as symbiotic animals exhibited normal luminescence levels throughout the experimental treatments.

To elucidate the effects of LSMMG stress on caspase activity, 10 h symbiotic animals were examined with Caspase Glo activity kits (Fig. 8). This time point was chosen as it coincided with the higher expression of caspase transcripts in LSMMG (Fig. 4). Results indicated that LSMMG-treated animals exhibited higher levels of caspase-3/7 and caspase-8 activity compared to gravity conditions in symbiotic animals (Fig. 8a, b). Interestingly, no differences were observed in caspase-9 activity in symbiotic animals incubated under LSMMG or gravity conditions, (Fig. 8c). Executioner activity was observed to be several orders of magnitude higher than both initiator caspases, exhibiting up to a 32-fold increase compared to the initiator caspases in both aposymbiotic and symbiotic animals (Fig. 8a-c). Treatment with the caspase-8 inhibitor Ac-IETD-CHO, the pan-caspase inhibitor Z-VAD-FMK, and the serine protease inhibitor Pefabloc abrogated the significant increase in light organ apoptosis observed at 10 h in LSMMG in symbiotic animals (Fig. 8d).

Fig. 8figure 8

Caspase activity and protease inhibition during light organ apoptosis in gravity and LSMMG conditions at 10 h post-inoculation. The activity of a executioner caspases 3/7, b extrinsic initiator caspase-8, and c intrinsic initiator caspase-9 was measured in the light organ of symbiotic (sym) hatchlings at 10 h in gravity (blue) and modeled microgravity (red). d Apoptosis was quantified in symbiotic hatchlings at 10 h in both conditions following protease inhibitor treatment including pan-caspase inhibitor z-VAD-FMK (ZVAD), the caspase 8 inhibitor Ac-IETD-CHO (C8i), caspase 9 inhibitor Ac-LEHD-CMK (C9i), Pefabloc (Pefa) and dimethyl sulfoxide (DMSO) controls. Error bars are the standard error of the mean. Asterisks denote significant differences between datasets (* = p ≤ 0.05, ** = p ≤ 0.01). Comparisons that were not significant are labeled “ns”

留言 (0)

沒有登入
gif