Temporal Microbiome Changes in Axolotl Limb Regeneration: Stage-Specific Restructuring of Bacterial and Fungal Communities with aFlavobacteriumBloom During Blastema Proliferation

23 The intricate relationship between regeneration and microbiota has recently gained attention, 24 spanning diverse model organisms. Axolotl ( Ambystoma mexicanum) is a critically endangered 25 salamander species and a model organism for regenerative and developmental biology. Despite 26 its significance, a noticeable gap exists in understanding the interplay between axolotl 27 regeneration and its microbiome. Here, we analyze in depth bacterial 16S rRNA amplicon 28 dataset that we reported before as data resource and profile fungal community by sequencing ITS 29 amplicons at the critical stages of limb regeneration (0-1-4-7-30-60 days post amputation, 30 “dpa”). Results reveal a decline in richness and evenness in the course of limb regeneration, with 31 bacterial community richness recovering beyond 30 dpa unlike fungi community. Beta diversity 32 analysis reveals precise restructuring of the bacterial community along the three phases of limb 33 regeneration, contrasting with less congruent changes in the fungal community. Temporal 34 dynamics of the bacterial community highlight prevalent anaerobic bacteria in initiation phase 35 and Flavobacterium bloom in the early phase correlating with limb blastema proliferation. 36 Predicted functional analysis mirrors these shifts, emphasizing a transition from amino acid 37 metabolism to lipid metabolism control. Fungal communities shift from Blastomycota to 38 Ascomycota dominance in the late regeneration stage. Our findings provide ecologically relevant 39 insights into stage specific role of microbiome contributions to axolotl limb regeneration. 40 41 42 43 44 45 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint 3 1 | INTRODUCTION 46 The axolotl ( Ambystoma mexicanum ), a species facing critical endangerment in its natural 47 habitat, is a member of the Ambystoma clade within the mole salamander family 48 (Ambystomatidae). Distinguishing salamanders from other amphibians is their remarkable 49 capacity for complete regeneration, encompassing not only limbs and appendages but also 50 certain internal organs. The intricate cellular and molecular mechanisms involved in the 51 extensively researched limb regeneration of axolotls make it a topic of significant scientific 52 interest 1, 2. 53 To regenerate its limbs, the axolotl employs a fascinating process involving the generation of 54 blastema cell mass, resembling the embryonic limb bud. This intricate procedure is significantly 55 influenced by innervation, where the blastema acts as the driving force for redevelopment 56 through a combination of growth and redifferentiation. Despite its complexity, limb regeneration 57 in axolotls unfolds systematically in a stepwise manner. Indeed, it is well established that the 58 regeneration progresses over distinct interdependent stages according to morphogenesis 3. These 59 sequential phases encompass wound healing, dedifferentiation (highlighted by blastema 60 establishment) and re-development stages of regeneration, respectively 4-6. The cell lineage- and 61 regeneration stage-specific expression patterns of various morphogenesis genes are critical for 62 successful limb regeneration and the early steps are highly critical for determining the extent of 63 regenerative response after limb amputation. More specifically, after limb amputation, a 64 specialized wound epidermis forms within ~24 hours through cell migration. Wound epidermis 65 is structurally and molecularly distinct from fully differentiated, intact epidermis. Innervated 66 wound epidermal cells then thicken the tissue, dermal cells migrate beneath it, and proliferates. 67 In the days following re-epithelialization, progenitor cells enter into the cell cycle as well as the 68 accumulate at the tip of the stump, beneath the wound epidermis. These progenitor cells may 69 originate from stem cells or by dedifferentiation. Epigenetic regulation controls both wound 70 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint 4 epidermis formation without scarring and epithelial-to-mesenchymal transition 1, 7, 8 . Together, 71 the activated cells accumulated at the tip of the stump give rise to highly proliferative blastema 72 cells 1, 7, 9. 73 Timing of these events may slightly vary in different salamander species but in axolotl the 74 movement of cells from the dermis into the early limb blastema begin to migrate beneath the 75 wound epithelia for about 5 days post amputation (dpa0-5: “initiation phase”). Next, the wound 76 epithelium thickens and forms an apical epithelial cap (AEC), which promotes the generation of 77 a population of undifferentiated mesenchymal cells called a ‘blastema’, which is formed through 78 a ‘dedifferentiation’ process from differentiated tissues (dpa6-20: “early phase”). In the re-79 development stage, the blastema continues to grow distally via cellular proliferation until the 80 completion of regeneration (dpa21-30: “late phase”) 4-6, 10, 11. 81 Animal microbiomes play key roles in an animal’s development, health, behavior, and protection 82 from pathogens 12-15. Recently, the impact of the host microbiome on regenerative ability of its 83 tissues, organs or even whole-body has gained significant attention even though this is still 84 relatively a new area of research in the field of microbiomes. Indeed, several species, renowned 85 for their regenerative capabilities, such as planarians, sea cucumber, and zebrafish, have been 86 employed in research to investigate whether the microbiome plays a role in regulating the 87 regenerative potential of these hosts or is directly engaged in the regeneration process. These 88 studies revealed that some bacterial species were shown to be causally linked with the 89 regenerative ability of the model animals. For instance, Pseudomonas and Aquitalea species on 90 Planaria, Erwinia carotovara on fruit fly, Aeromonas veronii on Zebrafish, Akkermansia 91 muciniphila and some probiotic Lactobacillus species in intestinal cell proliferation 16. The 92 microbiome, therefore, influence tissue regeneration by modulating immune responses and 93 inflammation, and potentially contributing to the activation of regenerative pathways. 94 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint 5 The axolotl is a permanently aquatic animal and its amputated limb regenerates against the 95 backdrop of microbial communities of the aquaria. This raises the question as to whether and 96 how axolotl’s gut and skin microbiota and water microbiota in the aquarium tank impacts the 97 limb regeneration. Remarkably, however, reports on the impact of microbiome on this highly 98 regenerative animal, the axolotl, and salamanders at large, is very limited, leaving significant 99 knowledge gap . We previously generated high quality 16S rRNA dataset from a longitudinal 100 sampling of axolotl limb tissues over the course of regeneration stages and exceptionally 101 presented the dataset to the interest of the axolotl scientific community 17. 102 In this study, we comprehensively analyze the diversity and temporal dynamics of both bacterial 103 and fungal communities within the same experimental setup. We hypothesized that regenerating 104 limb microbial communities are distinct from the environment; host factors create unfavorable 105 environment for the colonization of pathogens and induce the microbial communities to 106 progressively adapt to evolving micro niches in the course of axolotl limb regeneration. Our 107

demonstrate that both bacteria and fungi communities dramatically restructure over the 108 course of axolotl limb regeneration stages and show lack of resilience at the fully regenerated 109 stage. Our findings also provide evidence that the host limb tissues select microbial taxa that can 110 colonize in the regenerated tissues. 111 112 2 | MATERIAL AND METHOD 113 2 .1 | Ethical Statement and Experimental Design 114 The ethical statement and experimental design were previously described in our previous 115 report17. Briefly, this study was conducted at the Regenerative and Restorative Medicine 116 Research Center (REMER) of the Istanbul Medipol University. Experimental protocols and 117 animal care conditions were approved by the local ethics committee of the Istanbul Medipol 118 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint 6 University (the authorization number 38828770-604.01.01-E.10834). The founders of the initial 119 axolotls colony were procured from the University of Kentucky, USA and reared in optimized 120 conditions18 in aquaria in animal care facility of the university. 121 A total of 54 axolotls were selected at random from a group of siblings. The randomly selected 122 animals were then individually housed in aquariums filled with Holtfreter's solution at a constant 123 temperature of 18 ± 2 °C. They were fed with standard food (JBL Novo LotlM) once daily. The 124 experimental design, shown in Figure 1, included nine animals per group, divided into three 125 biological replicates (R1, R2, R3) to assess reproducibility. During the study, each animal was 126 kept in a separate aquarium. Limb amputation involved anesthetizing the axolotls with 0.1% 127 Tricaine methane sulfonate (Cat. No. E10521 or MS-222, Sigma-Aldrich, St. Louis, MO, USA) 128 and amputating the right forelimb at the mid-zeugopod level. 129 2.2. | Sample Collection 130 After amputation, animals were randomly assigned to six groups, simulating axolotl limb 131 regeneration phases: initiation (dpa 0 and dpa 1), early (dpa 4 and dpa 7), and late (dpa 30 and 132 dpa 60). Tissue samples from three animals were pooled to reduce variation, cryopreserved in 133 liquid nitrogen, and stored at −80 °C until genomic DNA isolation. Sampling specifics included 134 1-mm tissue around the cut site for dpa 0 and dpa 1, removal of newly formed blastema and 0.5 135 mm posterior tissue for dpa 4 and dpa 7, and collection around the original cut site for dpa 30 136 and dpa 60 to analyze microbiota in regenerated tissues. 137 To explore microbiota colonization in axolotl limbs originated from Holtfreter's solution, we 138 obtained 100 ml water samples from 9 separate aquaria throughout the experiment. We 139 combined samples from the beginning (day0-R1), middle (day30-R2), and end (day60-R3) of the 140 experimental timeline, forming the "aqua" control group with three replicates (R1, R2, R3; see 141 Fig. 1a). 142 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint 7 2.3. | DNA Extraction and Amplicon Sequencing 143 Genomic DNA was isolated from skin tissue samples using the DNeasy Blood and Tissue Kit 144 (Qiagen, Catalog No. 69504), while the Metagenomic DNA Isolation Kit (Epicentre, catalog no: 145 MGD08420) was employed for Aquarium water samples. The concentration of the isolated 146 genomic DNA was measured using the Qubit 2.0 Fluorometer (Thermo Fisher Scientific, MA, 147 USA). The internal transcribed spacer 1 (ITS1) region primers (ITS1-30F: 148 GTCCCTGCCCTTTGTACACA and ITS1-217R: TTTCGCTGCGTTCTTCATCG) were 149 selected from a previously published report 19 and V3-V4 region of 16S rRNA gene was 150 amplified in the previous study as described before17. To ensure DNA integrity, an aliquot of the 151 samples was run on a 1.0% agarose gel. For each sample, PCR was performed in a total volume 152 of 25 μ l with two replicates, including approximately 12.5 ng of purified DNA template and 2x 153 KAPA HiFi HotStart Ready Mix. The PCR conditions consisted of an initial denaturation at 95 154 °C for 3 minutes, followed by 25 cycles of denaturation at 95 °C for 30 seconds, annealing at 55 155 °C for 30 seconds, and extension at 72 °C for 30 seconds. The final extension was at 72 °C for 5 156 minutes. Negative controls in genomic DNA isolation and PCR steps were included to control 157 potential contamination and sequenced along with the other samples. 158 Amplicons were purified using the Agencourt AMPure XP purification system (Beckman-159 Coulter Cat. No. A63881, USA). A second PCR, with 8 cycles, was conducted using sample-160 specific barcodes, and the obtained amplicons were purified again. Equimolar concentrations 161 from each library were pooled and sequenced on an Illumina MiSeq sequencer using the MiSeq 162 Reagent Kit v2 (500 cycles). Raw data from this sequencing effort can be found in the NCBI 163 Sequence Read Archive (Accession number: PRJNA1073213). 164 165 166 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint 8 2.4. | Sequence Processing Pipeline and Taxonomic Classification 167 Pair-end reads of the amplicons were first merged using cutadapt 20 after removing primers and 168 low quality sequences. The merged clean sequence reads were then uploaded to the Nephele 169 platform (v.2.30, http://nephele.niaid.nih.gov) to run DADA-ITS pipeline and 16S rRNA 170 pipeline on default parameters and their associated denoising and removing chimeric reads. For 171 taxonomic assignment of the reads the SILVA small-subunit rRNA sequence database (v.132) 21 172 for bacteria and the database for Fungi 22 were used. To run the Phylogenetic investigation of 173 communities by reconstruction of unobserved states (PICRUSt2) 23 analysis we locally reran 174 QIIME2 on Greengenes taxonomy. The resulting amplicon sequence variants (ASVs) table was 175 filtered to exclude chloroplasts, mitochondria, archaea, eukaryotes, or unknown reads were 176 eliminated. As the number of sequences per sample were quite imbalanced, we rarefied ASV 177 tables before starting the downstream analyses and applied center log ratio (CLR) transformation 178 . Downstream analysis of ASVs was carried out using the package Phyloseq24 in R version 4.1.0. 179 2.5. | Community Structures and Diversity 180 The alpha diversity at each sampling time point was calculated using Chao1, Shannon, and 181 Simpson metrics. The Kruskal-Wallis (KW) test and posthoc Dunn’s test were used to identify 182 alpha diversity differences between samples. To test whether bacterial and fungal communities 183 differ in community composition and structure we employed Canonical analysis of principal 184 coordinates (CAP) based on Bray-Curtis distance matrix. Statistically significant differences 185 among life stages were determined with permutational ANOVA (PERMANOVA) performed 186 with adonis function. To test pairwise group differences were tested using pairwise.adonis 187 wrapper function for multiple testing 999 in vegan R package and p values were corrected with 188 False Discovery Rate (FDR) for multiple testing. Indicator species analysis was performed for 189 each in axolotl limb regeneration samples using the indicspecies package in R. Spearman rank-190 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint 9 based correlation coefficients were calculated using from the Hmisc R package. Moreover, 191 pheatmap, lattice, UpSetR and ggplot2 R packages were used for visualizations. Longitudinal 192 changes of taxa of interest was plotted using LOESS() function that fits local polynomial 193 regression within stats (v3.6.2) and ggplot2 R packages. 194 195 3 | RESULTS 196 In our previous work, we provided a comprehensive overview of the experimental design 197 employed in this study, as well as the 16S rRNA dataset generated and shared with the scientific 198 community for further analysis or meta-analysis17. Expanding on this groundwork, we broadened 199 our inquiry by analyzing ITS1 amplicons from identical archival gDNA samples to characterize 200 the temporal changes of the mycobiome during the various stages of limb regeneration. These 201 stages were categorized as dpa0-dpa1 for the initiation phase, dpa4-dpa7 for the early phase, and 202 dpa30-dpa60 for the late phase. In this study, we provide a comprehensive analysis of both 203 bacterial and fungal communities simultaneously. 204 Briefly, we meticulously monitored the amputated axolotls for 60 days and longitudinally 205 collected a total of n=21 samples (comprising 18 limb tissue samples from n=54 adult axolotls 206 and 9 individual aquaria samples with replicates) in total (Figure 1). Subsequent to preprocessing 207 ITS amplicon sequences, conducting quality checks, and removing chimeric sequences using the 208 DADA2 pipeline we obtained 238 amplicon sequence variants (ASVs) with sufficient 209 sequencing depth (28978 reads/sample in the rarefied (ASVs) table). The bacteria ASV table 210 included 116220 reads/sample in the rarefied ASV table. 211 212 213 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint 10 214 3.1. | Dynamic restructuring of bacterial and fungal diversity across regeneration stages 215 We compared species richness and diversity using a variety of alpha-diversity indices including 216 Chao1, Shannon, and Simpson and observed consistent decreasing trend in all metrics of both 217 bacterial diversity (Figure 2A) and fungal diversity (Figure 2B). Unlike fungal community 218 bacterial richness did not reach significance level among samples despite the apparently 219 decreasing trend as measured by Observed species and Chao1 indices (KW, p>0.05, Table S1.1). 220 In contrast, Shannon and Simpson indices steeply increased at dpa30 and dpa60 (late phase or re-221 development of the limbs, KW, P=0.01375 for Shannon and P=0.01423 for Simpson), 222 suggesting bacteria community colonizing the re-growing limb skin was more evenly 223 restructured than fungal community despite the decrease in the number of species. Indeed, 224 Dunn’s posthoc pairwise comparisons for each alpha diversity index for bacterial community 225 showed significant differences between all sample types except between the samples of the 226 initiation phase and late phase while this pattern was absent for the fungal community samples 227 (Table S2.1-S2.2). 228 Next, we analyzed temporal dynamics of both bacteria and fungi communities using Bray Curtis 229 distance and Canonical Analysis of Principal Coordinates (CAP) (Figure 3A), a constrained 230 ordination that maximizes the differences among a priori groups. We reproduced our prior 231 findings17, indicating distinct separation among bacterial communities at each sampling time 232 point (PERMANOVA, P=0.001, pseudo-F=7.191), a result further bolstered by statistically 233 significant hierarchical clustering (see Figure 3B) and confirmed by the SIMPROF significance 234 test (significance level 5%, 999 permutations). These results were mirrored by the CAP and 235 cluster analysis predicted metagenome functional pathway abundances (PICRUSt2) (Figure 3C-236 D, PERMANOVA, pseudo-F=10.461, P=0.001 and SIMPROF (5%, 999 permutations, 237 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint 11 respectively). Interestingly, despite the significant differences in distances between centroids of 238 fungal community samples (Pseudo-F=2.715, P=0.004, see Figure 4A), both hierarchical 239 clustering (Figure 4B) and pairwise PERMANOVA analysis (using Monte Carlo permutation) 240 revealed that fungal communities did not exhibit significant separation at any pairs of dpa timing 241 points (p>0.05), unlike the pairwise contrasts observed in bacterial communities (p<0.05) (Table 242 S3). This analysis suggests that the spatial and temporal dynamics of the fungal community in 243 response to regeneration are distinct from those of the bacterial community. 244 3.2. | Regeneration stages prompt restructuring of bacterial communities, triggering 245 Flavobacterium bloom 246 The heatmap of Top20 most abundant bacteria taxa at genus level showed conspicuous 247 clustering of co-abundant taxa by the limb regeneration stages of axolotl (Figure 5). In the 248 initiation stage (dpa0-1) largely anaerobic gut bacteria ( Bacteroides, Akkermansia, 249 Paenarthrobacter, Odoribacter, Parabacteroides, and unclassified species within the family of 250 Ruminococcoaceae and Lachnospiraceae ) were clustered together while in the early phase 251 (dpa4-7) a bloom of Flavobacterium was observed. Finally, in the late phase (dpa30-60) the 252 bacterial profile continued to evolve, giving rise to the potentially opportunistic bacteria (e.g. 253 Staphylococcus, Pseudomonas) or taxonomically closely related species that are often isolated 254 from soil freshwater creeks, lakes (e.g. Elizabethkingia and Chryseobacterium in the order of 255 Flavobacteriales) that colonize on the skin of the re-growing limbs. 256 Furthermore, we employed LOESS regression, which uses weighted local polynomials to model 257 species abundance data across time, to demonstrate longitudinal shifts of Top20 taxa. Clearly, in 258 the early phase (dpa4-7) a bloom of Flavobacterium persisted during the critical early stage that 259 allows for formation and growth of blastema (Figure 6A). A significant increase in the 260 abundance of Chryseobacterium over dpa60 was also remarkable. Additionally, we performed 261 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint 12 indicator species analysis to unravel indicator taxa at each sampling time point (Figure 6B), 262 which further supported that Flavobacterium is indeed an indicator taxon of the early phase, a 263 critical window for blastema to start proliferating. Notably, Aquabacterium and Acinetobacter 264 were also found to be an indicator taxa within the early phase although their abundances were 265 low. To answer the question whether functional pathway abundances of bacteria communities 266 follow similar trend as taxa in the course of limb regeneration stages we plotted the prediction of 267 metagenome functional pathway abundances (PICRUSt2) in a heatmap (Figure 7). Remarkably, 268 we observed clustering of pathways by limb regeneration stages, mirroring the taxa abundance 269 trend. The pathways in the initiation phase were found to be largely associated with amino acid 270 and nucleic acid biosynthesis while lipid biosynthetic pathways dominated the following early 271 phase. The late phase did not show clustered pathways. Correlation analysis between 272 abundances of bacteria taxa and pathways highlighted Chryseobacterium and Flavobacterium 273 highly significantly associated with the amino acid and lipid pathways, respectively (r=0.40, 274 p<0.05, Figure S1). Together, the diversity and co-abundance of bacteria communities and their 275 clustered pathways point toward stage specific restructuring of bacterial communities across the 276 limb regeneration while restructuring of the fungal communities does not closely follow the 277 regeneration stages. 278 3.3. | Unexpected complexity of fungal communities during the course of limb regeneration 279 The lack of obvious restructuring trend within fungal community was also manifested in the 280 taxonomic diversity of fungal communities. In spite of significant shift in community 281 composition from Basidiomycota dominance to Ascomycota prevalence (Figure 6A) genus level 282 fungal taxa did not cluster by the stages of limb regeneration (Heatmap, Figure 6B). These 283 findings prompted us to investigate degree of temporal relationship between bacterial and fungal 284 communities. We therefore performed Procrustes analysis and Mantel correlation between the 285 Bray-Curtis distance matrices for both communities. Interestingly, both Procrustes and Mantel 286 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint 13 analyses identified a moderately significant relationship between the two datasets (Procrustes 287 correlation=0.73, p=0.001 and Mantel test r=0.69, p=0.001), suggesting both communities 288 influence each other during regeneration. In addition, we found significant correlations (r=±0.40, 289 P<0.05) between bacteria and fungi taxa (Figure 7). In addition to Flavobacterium, Unclassified 290 Neisseriaceae, Chryseobacterium, Elizabethkingia, Unclassified Burkholderiaceae , and 291 Akkermansia exhibited significant correlations with fungal taxa. Notably, Flavobacterium and 292 Unclassified Burkholderiaceae showed predominantly negative correlations. Among fungal taxa, 293 Talaromyces, Aspergillus, and Stachybotrys displayed significant correlations with bacterial 294 taxa. 295 Finally, since Aqua samples were placed farther away from the limb tissue samples in CAP 296 figures we compared the number of unique and shared taxa between limb tissue and aqua 297 samples (Figure 8A-B). The upset figures shows both bacterial and fungal communities are 298 substantially unique at each sampling points (dpa0-dpa60) and few taxa persisted as shared 299 species between these time points, pointing towards highly dynamic temporal shifts of the 300 community compositions over the course of the limb regeneration. As shown in Figure 8A the 301 highest number of unique bacterial species assembled at dpa0 (562 ASVs) and dpa30 (535 302 ASVs) and 36 ASVs are common among the sampling points. As expected, only 14 ASVs are 303 present both in Aqua water and all limb tissues, suggesting the temporal dynamics of the 304 microbiota colonizing regenerating tissues is due largely to the host selection although microbe-305 microbe interactions could also contribute to unique community profiles. 306 307 4 | DISCUSSION 308 The axolotl, with its remarkable ability to regenerate limbs and organs due to sustained neoteny, 309 has strongly attracted the attention of scientific community for translational medicine research 3. 310 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint 14 Despite its recently sequenced expansive genome 25, studies on the axolotl microbiome are very 311 limited. In this pioneering study, we unveil, for the first time, the substantial restructuring of both 312 fungal and bacterial communities throughout the intricate process of limb regeneration. Our 313 findings illuminate the emergence of temporally unique community compositions at sampling 314 days that correspond to the widely accepted three phases of limb regeneration in axolotls. 315 Remarkably, as the axolotl's limbs undergo full regeneration, the microbial communities 316 inhabiting the limb skin undergo significant turnover. This suggests that the axolotl limb skin 317 microbiome lacks resilience in response to the regenerative process. Such insights into the 318 dynamic interplay between microbial communities and the regenerative capacity of axolotls shed 319 new light on the intricate mechanisms underlying limb regeneration. Moreover, they offer 320 valuable implications for translational research aimed at harnessing the regenerative potential of 321 axolotls for therapeutic purposes in humans. 322 One of the most striking observations on limb microbiota is regeneration-induced changes in 323 richness and evenness of limb microbiota as evidenced by significantly decreased alpha diversity 324 metrics for both bacteria and fungi communities. Notably, Shannon and Simpson indices 325 calculated for bacteria communities reversed the downward trend at dpa30 unlike fungal 326 communities. As the axolotl regrows its regenerated limb during this stage, it becomes apparent 327 that the fungal community struggles to recover its original diversity post-treatment. Notably, 328 however, 329 Fungi and bacteria often have different life cycles and growth rates, and regeneration stages 330 might favor bacterial growth or alters resource availability in a way that benefits bacteria over 331 fungi, contributing to the disparity in alpha diversity metrics. Moreover, although fungi 332 community significantly shifts during the course of limb regeneration as manifested by the 333 PERMANOVA main test the specific pairwise differences between individual groups were not 334 statistically significant. Even though this finding can be attributed to the small sample size (3 335 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint 15 replicates) for each pairwise comparison leading to limited power to detect true effects the 336 hierarchical clustering of the samples was not congruent with the stage specific separation. 337 Axolotl limb fungi may therefore have yet unknown subtle ecological complexity in responding 338 to traumatic amputation and regeneration processes. 339 Flavobacterium, among other microbiota, warrants special attention during the course of limb 340 regeneration as our results indicated this bacterium blooms across dpa4-dpa7 where the wound 341 epithelium thickens and forms an apical epithelial cap, which promotes the generation of a 342 population of undifferentiated mesenchymal cells called a ‘blastema’. As a result, signals from 343 the host and environmental cues can potentially promote or hinder differentiation and/or 344 proliferation of the blastema cells in this critical stage. Flavobacterium belongs to the family 345 Flavobacteriaceae in the phylum Bacteroidota. It is a Gram-negative genus comprising 270 346 species, ubiquities in the environment, including soil, seawater, freshwater, plants, and glaciers 347 26, 27. Many species of the genus Flavobacterium are capable of plant growth-promoting ability, 348 hydrolyzing organic polymers, such as complex polysaccharides, cold-adapted species produce 349 polyunsaturated fatty acids 26, although some species (such as F. columnare and F. 350 psychrophilum) reportedly cause infectious diseases in freshwater fish 28, 29 . Nonetheless, 351 considering axolotl was able to complete limb regeneration in this experiment in spite of the 352 bloom of this bacterium the unknown species of Flavobacterium in this context may be 353 positively contributing to the regeneration. While its putative contribution to regenerative 354 mechanisms and epigenetic and antigenic interactions with blastema cells cannot be ruled out 355 evidence in the published literature support that beneficial cutaneous bacteria on amphibians 356 protect against the lethal disease chytridiomycosis caused by Batrachochytrium dendrobatidis 357 (Bd) and salamanders 30-34 . In this study our analysis of fungal community did not detect Bd 358 with the caveat that the majority of the ITS reads were not assigned to Unclassified Fungi. We 359 therefore reasoned that either ITS primers did not amplify Bd, if present, or Flavobacterium 360 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint 16 might compete against unknown fungal species that could potentially impede regeneration 361 mechanisms. Moreover, the prevalence of Flavobacterium exhibited a robust correlation with 362 pathways related to lipid metabolism (see Figure S1), indicating a potential role for this 363 bacterium in influencing regeneration through lipid metabolic processes. Interestingly, a recent 364 study analyzed metabolomic profile of axolotl blastema at dpa11 35. Lipid metabolites such as 365 sphingomyelin and sphingosine, along with polyamine and pyrimidine metabolism, were found 366 to be enriched in blastema tissue. A recent cutaneous microbiota study on Ozark hellbender 367 salamander subspecies ( Cryptobranchus alleganiensis bishopi ), suffering from chronic wounds 368 within their aquatic habitat has uncovered interesting findings 36. The study showed that 369 Acidovorax sp., Flavobacterium sp., Aquabacterium sp., Chryseobacterium sp., and 370 Acinetobacter sp. were found to have higher relative abundance within these wounds compared 371 to intact skin. These results are consistent with our own research, as all four species were 372 identified as indicator species during axolotl limb regeneration studies. Nonetheless, further 373 investigation is needed to determine whether these bacterial species, particularly 374 Flavobacterium, are associated with regenerative capacity of axolotl limb. 375 The stage specific restructuring of bacteria and fungi communities across the limb regenerating 376 phases is fascinatingly chaotic and unstable as we found largely unique ASVs at almost every 377 dpa timepoint (Figure 8A-B). Given that wound epidermis covers the wound within a few hours 378 of amputation and Apical Epithelial Cap (AEC) cells quickly develops in the initiation phase the 379 newly colonizing microbial communities at each dpa point would first have a chance to briefly 380 interact with the wound epidermis but subsequently AEC cells during the critical stage of 381 blastema proliferation. Even though the chaotic stage specific restructuring of microbial 382 communities can be due partly to microbial resource competition the host factors is likely in 383 control of the community composition, provides the microbes less than favorable environment to 384 form continuous biofilms at least until the re-developmental stage is completed. The host factors 385 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint 17 include the proinflammatory immune response to amputation and regeneration. Indeed, 386 macrophages infiltrate amputation site about the same time wound epidermis covers the 387 amputation site and depletion of macrophages impedes limb regeneration 37. Theoretically, 388 genetic and epigenetic signaling between macrophages, AEC cells, nerve cells, blastema, 389 oxidative-stress, and ultimately redecoration of limb skin with protective mucus layer all would 390 have impact restructuring of microbial community attempting to re-colonize regenerating tissues. 391

392 Our findings offer compelling evidence for the first time that both bacterial and fungal 393 communities undergo restructuring across the widely accepted three major phases in axolotl limb 394 regeneration. However, it is noteworthy that the restructuring of the fungal community is less 395 precisely congruent with the timing of these major phases. The notable bloom of Flavobacterium 396 during the critical stage of blastema proliferation, along with the compositional shift of fungi 397 from Basidiomycota dominance in the early phase to Ascomycota dominance in the late phase, 398 raises intriguing questions about their potential impact on regenerative mechanisms in the limb 399 tissues through unknown mechanisms. Future research, therefore, should investigate the 400 contribution of microbial communities to regenerating tissues, specifically the cross talk between 401 Flavobacterium and macrophages, and AEC cells. By unraveling these intricate interactions, we 402 can gain deeper insights into the role of microbiota in the complex process of tissue regeneration, 403 paving the way for potential therapeutic interventions that harness the theraputic potential of 404 microbial communities in enhancing tissue repair and regeneration. 405 AUTHOR CONTRIBUTIONS 406 Conceptualization: SY and TD. Data curation: HA and SY. Formal analysis: SY, BD, and HA. 407 Methodology and resources: TD, BY, and SY. Writing – original draft: SY, Writing – review & 408 editing: SY, TD, HA, and BY. All authors read and approved the final manuscript. 409 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint 18 410 411 ACKNOWLEDGMENTS 412 The financial support for this study was provided by the Medipol University Research Fund. We 413 extend our gratitude to Ms. Ayça Önal and Ms. Damla Yay for their invaluable assistance in 414 capturing axolotl limb photographs. 415 CONFLICT OF INTEREST 416 The authors have no conflicts of interest to declare. 417 418 DATA AVAILABILITY STATEMENT 419 The data that support the findings of this study are openly available in NCBI Bio project at 420 https://www.ncbi.nlm.nih.gov/bioproject/....., reference number ….. 421 422 ORCID: Süleyman YILDIRIM: https://orcid.org/0000-0002-2752-1223 423 Uncategorized References 424 1 . T ajer B, S avag e A M, White d JL. The salam a nder bl a stema w ithin th e b r oa d e r c ontext of me ta z oan 425 reg ene rati on. F ron t Ce ll D ev Biol. 20 23;1 1: 12 06157. 426 2 . Eche verri K, F ei J , Ta naka EM. The Axol ot l 's jo ur ne y to t h e mode rn molec ul ar era . Curr Top Dev Biol . 427 2 022;147: 631-658. 428 3 . Boluk A, Y avuz M, De mircan T . Axolotl : A re sou r c e ful v e r teb ra te mo d el for re genera tion a nd beyond . 429 Dev Dyn. 20 22;251: 1914 -1933. 430 4 . Endo T, Br y ant S V, Gardin er D M. A ste pw is e mode l sy ste m fo r l imb regen e r a ti on . De v Biol. 2004;270 : 431 1 35-145. 432 5 . M c C u sk e r C , B r y a n t S V , G a r d i n e r D M . T h e a x o l o t l l i m b b l a s t e m a : c e l l u l a r a n d m o l e c u l a r m e c h a n i sm s 433 driv ing bl ast ema f ormat ion and limb reg e nera tion in t e trapod s . Reg en era tion ( Oxf) . 20 15;2: 54-71 . 434 6 . Tank P W, Ca rlson BM , C onnelly T G . A staging s y s tem for f oreli mb rege n e r a t i o n in the axolo t l , 435 Amby st oma mexic anum. J Mo r pho l. 197 6; 150: 117 - 1 28. 436 7 . Min S, Whit ed J L . L imb bla st ema f orm ation: How much do we k now a t a gene ti c and epige ne tic le vel? 437 J Bi ol Ch em. 2023;29 9: 10285 8. 438 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint 19 8 . Wan g M H, H su CL , W u CH, et a l. Timing Doe s Ma tt e r : Ne rve-M edia t e d HD AC 1 Pac e s the Tempor al 439 E xpress ion of Morp hogenic Gen e s Durin g Axol ot l Limb R egene r a t i o n. Fron t Cell D e v Bi ol . 2021 ;9 : 440 6 41987. 441 9 . Haa s BJ, W hi te d JL . Adv anc e s in Dec od ing Axolotl Li mb Regener atio n. T r en d s G ene t. 2017;33 : 55 3 -442 5 65. 443 1 0. G a rdin e r DM, M un eoka K, B ryant S V. Th e mig r a t i on o f d ermal c ell s d ur i n g bla st e m a fo r ma ti on i n 444 a xolotl s. D ev Biol. 19 86;118: 4 88 -493. 445 1 1. Yoko yama H. In itia tion o f limb r e gen erati on: th e c ri tical step s fo r reg ene rativ e capa ci ty. D e v Gr ow t h 446 Dif fe r. 20 08;50: 13 - 2 2. 447 1 2. Dia z He ijtz R, W ang S , Anua r F, e t a l. No rmal gut microbio ta modula t e s b r a in d evelo pm ent an d 448 be havi or . Pr oc N atl Aca d Sc i U S A. 2 011; 108: 3047-3052 . 449 1 3. Mo tta EVS, Mor an NA. The ho neyb e e m icrobi ota and i t s impa c t on hea l t h and di se a se. N a t R ev 450 Mic robiol. 2 023 . 451 1 4. Pa n war RB , Seq uei ra R P, Cl arke T B. Micr obio ta - m e diat ed p rot ec t i on ag a inst an tibiotic -re sis t a n t 452 pa thogen s. G ene s Immun. 2021;22 : 255- 267. 453 1 5. Mur e A , Sug iur a Y, Mae d a R , e t a l. I den tific a tion of k ey y ea st s pec i e s and mi crob e -microbe 454 i nt e ra ction s impac ti ng la r v al g r ow th o f D r os ophil a in th e wild. Elif e. 2 023;12 . 455 1 6. D i az- Diaz LM, Rod rig uez-Vi l laf ane A , Garcia - A rrar a s JE . The R ole o f th e Mic r o bi ot a in R egene ra t i o n-456 As s o cia ted Pr o c e sse s. F ront Ce ll D ev Bi ol . 2021 ;9: 7687 83. 457 1 7. Demirc an T, Ilha n AE, Ove zmyradov G , O z t u r k G , Yildirim S . Lon gi tudinal 16S rRNA d ata d eriv ed fro m 458 l imb regen era t i ve ti s su e sa mple s o f ax olotl Ambys toma me xica num. S ci D ata . 2019;6: 7 0. 459 1 8. Kh att a k S, Mura wala P, And rea s H, et a l. O p timized ax olotl (Amby stoma mex ican um) husba ndry , 460 bree ding, m etamo rpho si s , tr a n s g e ne si s and ta moxif en -med ia ted r ec ombina tio n. Na t P rotoc . 2014;9 : 461 5 29-540. 462 1 9. U syk M, Zolnik CP, Pa tel H, L evi MH, Burk RD . N o vel ITS 1 Fung al Prim er s for Cha ra cte riz a t i on o f t h e 463 My cobiom e. mSphe r e . 2017;2 . 464 2 0. Kech in A, Boyar skikh U, Kel A, F ilip e nk o M. cutP rimer s : A New T ool fo r Ac cu rate Cut ting of Prime rs 465 from Re ad s of Targ et ed N ex t G e ne ra tion Seq uenci ng. J Compu t Bi ol. 2017 ;24 : 1138 -1143. 466 2 1. Q ua st C, Pr ue s se E , Yilmaz P , e t al . T he SIL VA r ib o s oma l R N A ge ne da tab as e projec t: improved d ata 467 proc e ssing and w eb -ba sed tool s. Nuclei c Ac ids R e s . 2013;41: D 59 0-596 . 468 2 2 . N i l sso n R H , L a r sso n K H , T a y l o r A F S , e t a l . T h e U N I T E d a t a b a s e f o r m o l e c u l a r i d e n t i f i c a t i o n o f f u n g i : 469 ha ndling dark tax a and pa rall el taxonom i c c lass ifi ca tion s. Nucle ic Ac id s Re s . 2019; 47: D 2 59- D 2 64. 470 2 3. Doug la s GM, Maf f ei VJ, Zanev eld JR , et a l . P IC RU S t2 f or p redic t i o n o f m eta g enome functio n s. Na t 471 Bi ot e chn ol. 2020;38 : 685 -688. 472 24 . M c M u r d i e PJ , H o l mes S. p h y loseq : an R p a ck a ge f o r r e p r od u c i b le i nt e ra c t i ve a n al y s is an d g ra p h i c s o f 473 mi crobiome cen su s dat a. PL oS One . 2 013;8: e 61217. 474 2 5 . N o w o s h i l o w S , S c h l o i ss n i g S , F e i J F , e t a l . T h e a x o l o t l g e n o m e a n d t h e e v o l u t i o n o f k e y t i ss u e 475 forma tion re gulato r s. Na t ur e. 2018;554 : 50-55. 476 2 6. A. W a ś k iewic z L I. Fla vobac te r i um sp p. – Char acte ris tic s , O c curr ence , and To x icity. I n: Batt C , e dito r . 477 E ncyc lopedi a of F oo d Microbio l ogy (S eco nd Edition) . NY, USA : Ac ad emi c Pre ss , 20 14. 478 2 7. Yang L L, Liu HC, Xin YH, Liu Q. Flav obacte rium frigo ritol eran s s p . nov . a nd Flav obacte rium shii s p . 479 no v., is olat ed f r om glac ier s o n th e Tibe ta n Plat ea u. Int J Sy st Ev ol Mic r obi ol. 2023 ; 73. 480 2 8. Loc h TP, Fais al M. Emerging flavo bact erial in fec t i on s in fi sh: A r e view . J Adv Re s . 20 15;6 : 283 -300. 481 2 9. D e Sil va L, He o GJ. Bio film f or m ati on o f pat hog enic bac ter i a i sola t e d fr o m aquatic an imal s . Arch 482 Mic robiol. 2 022;205 : 36. 483 3 0. Bat aille A, L ee - Cruz L , Trip athi B, Ki m H, W a ldman B. Mic robi ome Var iatio n Acro ss Amphibi a n Skin 484 Re gion s: Impli ca tion s f or Chy tridiomy co s is Mit i g ation E f fo rt s. Microb Ec ol. 20 16;7 1: 2 21-232. 485 31 . B a tes K A , C l a re FC, O ' H an l o n S , et a l. A m p h i bi a n ch y t r id io m y co si s o ut b re a k d yn a mi c s ar e l i n ke d w it h 486 ho st skin b ac ter ial communi t y struc ture . Nat Commu n. 2018;9 : 693 . 487 3 2. Bat es KA , Som mer U, H opkins K P, e t a l. Micro biome func tion p redict s amp hibia n chytridiomy co si s 488 di se a se dynami cs . Microbi ome . 2022;1 0 : 44. 489 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint 20 3 3. Mule tz Wolz CR, Ya rwood S A , Camp bell G r ant EH, Fl e i s c her RC, Lip s KR. E ff e c t s o f ho st speci e s a n d 490 e nvironmen t on t h e sk in mic robiom e of P let hodon tid s alama nd er s. J A n im Ec ol. 2 018 ;87: 341-3 53. 491 3 4 . S h a w S D , B e r g e r L , B e l l S , e t a l . B a se l i n e c u t a n e o u s b a c t e r i a o f f r e e - l i v i n g N e w Z e a l a n d n a t i v e f r o g s 492 (L eiopel ma a rchey i a nd Leiop elma hoch st et teri ) a nd i mplica t i on s for thei r r ole in defen s e again s t t h e 493 a mphibia n chy t rid (Ba trac hoch yt r iu m de ndroba tidi s) . J Wildl D i s. 20 14;50: 7 23 -73 2. 494 3 5. Li u Z, L i W, G eng L , et al. C r o s s - spec i e s me tabol omic ana ly si s id enti fie s u ridine a s a p ote n t 495 reg ene rati on promo ting fac tor . Cel l Di sc ov. 20 22;8: 6. 496 3 6. H il l AJ, Gri s ni k M, Wal ke r DM. B ac te r i al Ski n A sse mblage s of Symp a tr ic Sal a mander s Are Prima rily 497 S haped by Ho st G enu s. Mic r o b Eco l. 202 3;8 6: 1364-1 373. 498 3 7. G od win JW, Pint o AR, Ro sen tha l N A. M ac r op h age s ar e req uired fo r a dult sal am ander lim b 499 reg ene rati on. Proc Na tl Acad Sci U S A . 201 3;110: 94 15-9420. 500 501 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 7, 2024. ; https://doi.org/10.1101/2024.03.07.583834doi: bioRxiv preprint