Tuning into xanthan: A conserved yet flexible polysaccharide utilization system in Microbacterium

xanthan, Microbacterium, polysaccharide utilization, regulation, microbial 19 ecosystems, evolution 20 21 Word counts: 22

204 words 23 Text: 4,846 words 24 25 26 *Address correspondence to Marion Eisenhut, [email protected] 27 28 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint 2

29 Bacteria encounter structurally complex extracellular polysaccharides in natural environments, 30 yet the regulatory and evolutionary basis of their utilization remains poorly understood. Here, 31 we isolated a soil-derived Microbacterium strain, named Microbacterium xanthanicum UB-LE1, 32 that grows on xanthan as the sole carbon source. We dissected the genetic and regulatory 33 architecture underlying this capability. Genome sequencing combined with transcriptomic and 34 proteomic profiling uncovered a discrete, strongly inducible regulon associated with xanthan 35 utilization, encoding 23 proteins with five secreted proteins and three candidate transcriptional 36 regulators. DNA-affinity purification sequencing confirmed two regulators binding to operons 37 within the xanthan utilization locus. Comparative genomics across the Microbacteriaceae 38 revealed conserved and lineage-specific features of this system and supports recent 39 acquisition and modular integration of the locus, with at least two predominant architectural 40 variants possibly shaped by substrate availability and ecological specialization. Coordinated 41 induction at both the transcript and protein levels, together with two experimentally validated 42 regulators, points to tight regulatory control of complex polysaccharide degradation in 43 Microbacterium xanthanicum UB-LE1. 44 Together, these findings provide mechanistic and evolutionary insight into how bacteria 45 adapt to complex extracellular carbohydrates, expand current knowledge of xanthan turnover 46 in microbial ecosystems, and establish a framework for exploring the emergence and 47 diversification of specialized polysaccharide utilization pathways across bacterial taxa. 48 49 IMPORTANCE 50 Microorganisms are central drivers of carbon turnover in soils and other terrestrial ecosystems, 51 determining the availability of nutrients and shaping microbial community structure. A 52 significant portion of soil carbon is contained in extracellular polysaccharides, yet the pathways 53 by which microorganisms degrade these complex polymers remain poorly understood. 54 Xanthan, a structurally complex and widely produced microbial exopolysaccharide, represents 55 a persistent and largely overlooked carbon pool. By dissecting the genetic, regulatory, and 56 .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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint 3 evolutionary basis of xanthan utilization in Microbacterium xanthanicum UB-LE1, this study 57 advances our understanding of how soil bacteria adapt to complex extracellular carbohydrates 58 and how substrate availability shapes the emergence and diversification of specialized 59 metabolic pathways. Importantly, the identification of additional xanthan-active enzymes and 60 regulatory components in M. xanthanicum UB-LE1 opens opportunities for targeted 61 modification of xanthan structure and properties, paving the way for new biotechnological 62 applications in food, materials, and industrial biotechnology, while linking microbial ecology to 63 functional innovation. 64 65

66 Microorganisms play a central role in shaping carbon turnover in soils and other terrestrial 67 ecosystems through the secretion and degradation of extracellular polymers. Among these 68 polymers, exopolysaccharides (EPS) are secreted into the environment to form biofilms, which 69 provide structural support for microbial communities and act as carbon sources for specialist 70 degraders. Understanding how microorganisms sense, transport, and metabolize structurally 71 complex EPS is essential to uncovering both ecological interactions and the evolution of 72 specialized metabolic pathways. 73 Xanthan is a high-molecular-weight, anionic heteropolysaccharide produced by 74 members of the Xanthomonas genus. It consists of a β-(1,4)-glucan backbone substituted with 75 trisaccharide side chains containing acetylated or pyruvylated mannose and glucuronic acid. 76 In plant-pathogenic Xanthomonas campestris, xanthan contributes to virulence by suppressing 77 plant immune responses, protecting the bacteria from host-derived and environmental 78 stresses, and facilitating adhesion, aggregation, and colonization of the apoplast (1–3). Loss 79 of xanthan biosynthesis attenuates virulence, highlighting its central role in host–pathogen 80 interactions. Beyond its role in plant pathogenesis, xanthan persists in soil for extended 81 periods, with portions reported to resist degradation for several months (4) due to its high 82 molecular weight and complex chemical structure, providing a stable carbon pool for microbial 83 communities. 84 .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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint 4 The structural complexity of xanthan has also driven its widespread industrial use. Its 85 high viscosity and shear-thinning properties make it a ubiquitous additive in food, cosmetics, 86 personal care, and industrial formulations (5, 6). Emerging applications in biomedical 87 engineering (7) and geoengineering (4) highlight the growing relevance of understanding 88 xanthan metabolism and biodegradation. Because it is biodegradable, identifying microbial 89 enzymes that can modify or degrade xanthan has both ecological and biotechnological 90 significance, offering potential avenues for tailored modification of xanthan properties for novel 91 applications. Despite the industrial and ecological relevance of xanthan, the mechanisms by 92 which fungi and bacteria degrade this polysaccharide (8) are only partially understood. 93 Xanthan degradation typically begins with an extracellular xanthan lyase (PL8) that liberates 94 terminal pyruvylated mannose residues, followed by an endoxanthanase (GH9 or GH5 family) 95 that cleaves the backbone into tetrasaccharides (8). Complete turnover requires intracellular 96 enzymes to remove acetyl groups and cleave the resulting oligosaccharides into 97 monosaccharides (8, 9), as well as transport systems to import these products and regulatory 98 proteins to coordinate expression of the catabolic machinery. A variety of xanthan-degrading 99 bacteria have been isolated from diverse environments including soil (10), sludge (11), and 100 mammalian intestines (8). These include Bacillus GL1 (12) and Microbacterium XT11 (13), 101 both of which possess secreted PL8 xanthan lyases, membrane-bound GH9 enzymes, and 102 intracellular GH38 and GH3 enzymes (10), and also include Paenibacillus nannanis (9, 14, 103 15), Bacteroides and Ruminococcus (8), Cohnella sp. 56 (VKM B-36720) (16), and 104 Paenarthrobacter ilicis (17). While many xanthan-degrading enzymes are known, information 105 on their regulatory control is scarce. In other extracellular polysaccharide systems, catabolic 106 genes are frequently organized into co-regulated clusters. For instance, polysaccharide 107 utilization (pul) loci in human gut Bacteroidetes comprise genes for extracellular enzymes, 108 transporters, intracellular enzymes, and regulators (18). Secretion of large or high-molecular-109 weight proteins is energetically costly, and these systems are typically under tight 110 transcriptional control (18). Similarly, gene clusters for pectin degradation in Erwinia 111 chrysanthemi (19) or rhamnogalacturonan degradation in Bacillus subtilis (20) are regulated 112 .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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint 5 by specific transcription factors or two-component systems. However, no such regulatory 113 frameworks have been comprehensively described for xanthan utilization. 114 To investigate how bacteria adapt to structurally complex extracellular carbohydrates, 115 we focused on a newly isolated, soil-derived Microbacterium strain capable of using xanthan 116 as its sole carbon source. This strain, M. xanthanicum UB-LE1, provides a unique system to 117 connect ecological persistence of xanthan with the microbial strategies required for its 118 degradation. By integrating genomic, transcriptomic, and proteomic data, we map the 119 functional components involved in xanthan utilization, including secreted enzymes, 120 intracellular catabolic proteins, and regulatory elements coordinating their expression. 121 Comparative analyses across related Microbacteria indicate that these pathways combine 122 both conserved mechanisms and lineage-specific adaptations, reflecting evolutionary 123 pressures imposed by substrate availability and ecological specialization. Studying M. 124 xanthanicum UB-LE1 thus illuminates how bacteria detect, import, and metabolize complex 125 polysaccharides in natural environments, while also providing a framework for exploiting these 126 systems in biotechnology. 127 128

129 Isolation of Microbacterium xanthanicum UB-LE1 as xanthan-degrading bacterium 130 To isolate a naturally occurring xanthan-degrading bacterium, a topsoil sample from the 131 eastern Teutoburger Forest in Bielefeld, Germany, at approximately 52,01387° N, 8,48334° E 132 was enriched for capable xanthan degraders. After several rounds of single colony restreaking, 133 a single orange colony was isolated and provisionally named UB-LE1. Nanopore sequenced 134 DNA assembled into a full genome as a single contig with a total size of 3,410,522 bp, a GC 135 content of 70.46%, and BUSCO completeness of 99.2% (details of genome assembly and 136 annotation in Laker et al. (Companion paper submitted to MRA). Since UB-LE1 showed the 137 highest average nucleotide identity (ANI) values of 98.22% with Microbacterium sp. Leaf436 138 (Table S2) and TYGS classified UB-LE1 as closest to Microbacterium enclense NIO-1002 139 (details of phylogenetic analysis in Laker et al. (Companion paper submitted to MRA)), the 140 .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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint 6 strain was designated Microbacterium xanthanicum UB-LE1 (in the following M. xanthanicum 141 UB-LE1). 142 M. xanthanicum UB-LE1 is a Gram-positive, rod shaped (1 µm by 3 to 4 µm) bacterium 143 that lacks a flagellum and is non-motile (Fig. S1A). Its growth optimum was detected at 30 °C, 144 pH 7, without addition of NaCl (Fig. S1B-D). Growth experiments in replicates identified M. 145 xanthanicum UB-LE1 as an efficient user of xanthan as the sole carbon source with growth 146 rates on xanthan of µ = 0.0653 + 0.0185 h -1 in comparison to glucose with µ = 0.102 + 0.004 147 h-1 (Fig. 1). 148 Fig. 1. Growth and sampling of M. xanthanicum UB-LE1 for OMICS studies. Growth of M . 149 xanthanicum UB-LE1 in M9 medium with either glucose or xanthan as C-source was performed on a 150 shaker at 30 °C and 180 rpm. Sampling time points for transcriptomic and proteomic analyses are 151 indicated with arrows (blue: glucose sampling; yellow: xanthan sampling). Three biological replicates 152 were grown under each condition and sampled. 153 154 Response to xanthan on transcript and protein level 155 To gain a comprehensive view on M. xanthanicum´s response to xanthan we performed a 156 comparative OMICS approach. Samples for transcriptomic and proteomic (cellular and 157 extracellular samples) analyses were taken after 34 h (+ glucose) and 48 h (+ xanthan), 158 respectively, in the exponential grow phase and at a similar cell density (Fig. 1). 159 .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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint 7 The transcriptome analysis detected 3,085 transcripts with xanthan as the carbon 160 source and 3,093 transcripts with glucose as the carbon source (Table S1: mothertable). 161 Samples were well separated in a principal component analysis (PCA) with 49% variation 162 represented in the first principal component indicating the major source of variation in the 163 transcriptome experiment is indeed the carbon source (Fig. 2A). The proteome analysis of the 164 cellular protein detected 1,285 proteins with xanthan as the carbon source and 1,274 proteins 165 with glucose as the carbon source (Table S1: mothertable). Samples were well separated in a 166 PCA analysis with 51% of variation represented in the first component (Fig. 2B). The genome 167 contained 202 genes with products that were predicted to be secreted by the secretion protein 168 (SP) or twin-arginine translocation (TAT) pathways (Table S1: mothertable). Seventy-nine of 169 these predicted extracellular proteins were detected in the supernatant with xanthan as the 170 carbon source and 56 proteins were detected with glucose as the carbon source. The 171 extracellular proteome samples carried higher variation as only 38% of variation in the first 172 principal component separated both groups (Fig. 2C). 173 Statistical analysis detected 257 transcripts with significantly ( q < 0.01) higher 174 abundance in xanthan-dependent growth compared to glucose and 186 transcripts with 175 significantly lower abundance in xanthan, that is higher abundance in glucose-dependent 176 growth (Fig. 2D). When the 34 transcripts with at least 32-fold induction in xanthan were 177 examined, they included MMX123_00115, MMX123_00472 and -473, MMX123_01228, 178 MMX123_01822 and -1823, genes in a genome region from gene MMX123_02529 to gene 179 MMX123_02549, genes in a genome region from gene MMX123_02601 to MMX123_02608, 180 and genes in a genome region from MMX123_03133 to MMX123_03136 (Fig. 2D). Among the 181 proteins, statistical analysis detected 149 cellular proteins as higher abundant in xanthan 182 compared to glucose and 144 cellular proteins as higher abundant in glucose (Fig. 2E, Table 183 S1: mothertable). When the proteins with fold-changes greater than 32-fold were examined, 184 they included the protein encoded by MMX123_00473, proteins which are encoded by genes 185 positioned between MMX123_02531 and MMX123_02549, MMX123_02679, 186 MMX123_03136, MMX123_03140, and MMX123_03143. Many of these are identical to the 187 .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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint 8 maximally induced transcripts with additions of MMX123_02679 and MMX123_03143. Among 188 the proteins in the supernatant with secretion signals, 17 were detected as significantly higher 189 190 Fig 2. RNA-seq and proteomics analyses. Profiles of M. xanthanicum UB-LE1 grown under two 191 different carbon source conditions. Principal component analysis of (A) transcriptome, (B) cellular 192 proteome, and (C) extracellular proteome profiles from M. xanthanicum UB-LE1 grown on glucose or 193 xanthan as the sole carbon source. Yellow points represent samples grown on glucose medium and 194 TranscriptomeCellular proteomeExtracellular proteome A B C D E F .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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint 9 blue points represent samples grown on xanthan medium. Biological replicates are marked by their 195 labels. Volcano plot showing differential gene expression (D), differential cellular proteome (E) 196 accumulation and differential extracellular proteome (F) accumulation between xanthan- and glucose-197 grown cells. Each point represents a gene or protein. Not significantly regulated events are colored in 198 gray, significantly xanthan up-regulated events are colored in yellow, significantly glucose up-regulated 199 events are colored in blue. 200 201 abundant in xanthan and 4 as significantly higher abundant in glucose (Fig. 2F). The proteins 202 with fold-change greater than 32-fold included four of the five genes annotated as having 203 secreted protein products in the region MMX123_02528 to gene MMX123_02550 and five 204 additional proteins MMX123_00619, MMX123_02209, MMX123_02445, MMX123_02738, 205 and MMX123_02792. 206 The gene products of region MMX123_02528 - MMX123_02550 were consistently 207 detected as the most induced gene products in xanthan-dependent growth with similar 208 inductions in transcripts and proteins (Fig. S2). 209 210 Analysis of CAZymes and putative sugar utilization loci in the genome 211 In accordance with the focus on xanthan utilization we inspected the genomic repertoire of M. 212 xanthanicum UB-LE1 for carbohydrate active enzymes (CAZymes) and observed their 213 response towards the different carbon sources. In total, 106 genes are annotated as 214 CAZymes. The majority with 52 are glycoside hydrolases (GHs) followed by 39 215 glycosyltransferases (GTs), 5 carboesterases (CEs), 5 auxilliary enzymes, 3 polysaccharide 216 lyases, and 1 carbohydrate binding module (CBM) protein (Fig. S3). Of these 106 annotated 217 CAZymes, 14 CAZymes also contain a secretion signal. Seven of these belong to the GH43 218 family, with five significantly transcriptionally more abundant in xanthan, one significantly more 219 abundant in supernatant protein in xanthan. One of the CAZymes with secretion signal is 220 annotated as a putative cellulase (GH5 family) and is significantly more abundant in the 221 extracellular proteome in xanthan. Two of the 14 CAZymes with secretion signal are increased 222 significantly at least 32-fold on transcriptional and secreted protein level, namely a GH9 family 223 .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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint 10 protein and a PL8 family protein. When pul-loci are defined as those, which contain a GH in 224 close proximity to a transport system consisting of periplasmic binding proteins, 15 pul-loci 225 (pul-locus I – XV) were identified in the M. xanthanicum UB-LE1 genome (Table S1: pul loci). 226 Of these, five contained at least three transcripts with significantly higher abundance in 227 xanthan medium while not any locus contained at least three transcripts with significantly 228 higher abundance in glucose medium. Induction in xanthan ranged from about 8-fold to up to 229 225-fold for a transcript in the locus with the highest induction, that is the locus 230 MMX123_02528 – MMX123_02550 (Table S1: pul loci). 231 232 Identification of the xanthan utilization (xut) locus 233 Apparently, OMICS analyses identified a genome region between gene MMX123_02528 to 234 gene MMX123_02550 as consistently upregulated during growth with xanthan as the carbon 235 source on both, the transcript and protein level (Fig. 2, Fig. S2). A genome-wide feature 236 analysis of GC content and transcript abundance patterns identified the same region 237 (MMX123_02528 - MMX123_02550) as having a reduced GC content compared to the 238 surrounding regions and as carrying genes with higher transcript abundance (Fig. 3A). 239 The genome analysis identified four additional regions longer than 10 kb of reduced 240 GC content relative to the surrounding genome (Fig. 3A, Table S1: lowGC content regions) 241 but none of these overlaps with regions of higher transcript abundance (Fig. 3A). Based on its 242 characteristic combination of low GC content and high expression levels during growth on 243 xanthan, genomic region 4 was designated the xanthan utilization (xut) locus (Fig. 3A, Table 244 S1: lowGC content regions). The xut-locus comprises 23 genes (Fig. 3B) with functional 245 annotations provided in Tables S1 (Table S1: xut_interproscan) and naming in Table S3. The 246 gene xanthan utilization regulator 1, xutR1 (MMX123_02528), encoded on the reverse strand, 247 contains a DeoR-type helix-turn-helix domain and a periplasmic binding domain, suggesting a 248 sugar-responsive transcription factor (TF). The adjacent gene xutA (MMX123_02529) encodes 249 a secreted GH9 enzyme likely capable of cleaving the xanthan backbone following removal of 250 .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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint 11 terminal pyruvylated mannose residues (9). The next gene, xutB (MMX123_02530), encodes 251 a putative 2-hydroxyacid dehydrogenase. 252 253 Fig. 3: The xanthan utilization ( xut) locus and its organization in the genomic context in M. 254 xanthanicum UB-LE1. A) Circular genome map of the UB-LE1 genome. Track 1 displays forward 255 coding sequences (CDS; dark gray) and track 2 shows reverse CDS (light gray). In both tracks, rRNAs 256 are shown in green and tRNAs in brown. Track 3 represents GC content with positive GC content in 257 .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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint 12 black and negative in blue-gray. Track 4 depicts GC skew with positive GC skew in bright red and 258 negative in dark red. Track 5 displays the log2 fold change (log2FC) with high log2FC in yellow and low 259 log2FC in blue. The position of the xut-locus is highlighted in light teal over all tracks, with the ±10 kbp 260 surrounding region shown in cyan. B) Organization of the xut-locus with its gene annotation. Genes are 261 shown as directional arrows, representing gene orientation and their colors denote functional gene 262 annotation. Transcript abundance for each gene is shown as two-colored rectangles below each gene. 263 The first rectangle displays the transcripts per million (TPMs) under growth on glucose and the second 264 rectangle represents the TPMs under xanthan growth, with increasing intensity from white to red 265 indicating higher expression levels. 266 267 Genes xutC1 to xutI (MMX123_02531–02539) form a candidate operon on the forward 268 strand. The first three genes encode a likely ABC transporter system, including a lipid-269 anchored binding protein (xutC1) and membrane-associated transporter components (xutC2–270 C3). The intracellular enzymes XutD (GH3) and XutE (GH38) are similar to proteins, which 271 hydrolyze glucose- and mannose-containing linkages. The downstream genes xutF–xutH 272 encode proteins with predicted roles in sugar modification, including oxidoreductase and 273 isomerase activities, while xutI encodes a glycerate kinase. 274 A second operon on the reverse strand includes xutJ (MMX123_02540), encoding a 275 secreted PL8 xanthan lyase that likely cleaves terminal pyruvylated mannose residues, 276 preceded by a second ABC transporter system ( xutK1–K3, MMX123_02541-02543) and the 277 putative esterase XutL ( MMX123_02544). The regulator XutR2 ( MMX123_02545), a TetR-278 type TF, may act as a repressor. A third operon comprises TF xutR3 (MMX123_02550) and 279 genes xutM1–M3 (MMX123_02546-02548), encoding a third transporter system, and XutN 280 (MMX123_02549), a predicted FAD-dependent oxidoreductase. The organization of the locus 281 suggests at least four promoters are required for full expression (Fig. 3B). 282 283 284 285 286 .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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint 13 Identification of transcriptional regulators for the xut-locus 287 The position of three TFs within the xut-locus suggests they may be regulating the expression 288 of genes within and hence bind some or all the putative transcriptional starts (Fig. 3B). DNA 289 Affinity Purification sequencing (DAP-seq) was used to determine genome-wide binding-sites 290 for the three TF candidates XutR1, XutR2, and XutR3. The three candidates were expressed 291 in vitro with halo-tags and detected on Western Blots (Fig. S4). XutR2 protein was incubated 292 B A xut1-operon xut2-operon xut3-operon XutR3 XutR2 xut1-operon xut2-operon xut3-operon .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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint 14 Fig. 4: XutR2 and XutR3 bind genome regions in the xut-locus. DAP-seq binding data are shown 293 for A) XutR2 and B) XutR3 across the xut-locus and one adjacent flanking gene on either side. For each 294 TF, the top track depicts per-base read coverage of DAP-seq reads mapped the genome using Bowtie2. 295 The second track shows the corresponding empty-vector control coverage, scaled to the experimental 296 track to facilitate direct comparison of signal-to-noise ratios. The third track displays binding sites 297 identified by MACS3 and is based on the narrowPeak output. The width of the bars indicates the called 298 peak region, while the color intensity reflects the peak score. The bottom track illustrates schematic of 299 the xut-locus genomic architecture. Arrows indicate coding strand orientation, with gene colors 300 corresponding to the functional categories defined in Fig. 3B. Adjacent flanking genes are displayed in 301 light gray. 302 303 with DNA and sequencing of the bound DNA yielded two peaks with read pileups above 8,000 304 reads with a summit at genome position 2,766,454, which puts it 15 bases upstream of itself 305 and 114 bases upstream of the first gene, xutL, in the pentacistronic transcriptional unit within 306 the xut-locus (Fig. 4A). 307 The second peak is located at 1,117,528 in between MMX123_00995 and 308 MMX123_00996, neither of which is changed in transcript abundance in xanthan- or glucose-309 dependent growth. XutR3 yielded one peak at genomic position 2,772,971, which is 35 bases 310 upstream of itself, the first gene in a second pentacistronic operon (Fig. 4B). Additional peaks 311 were not detected. That is, XutR2 controls transcription of the xut2-operon (xutJ - xutL) and 312 XutR3 regulates expression of the xut3-operon (xutM1 – xutR3). Transcriptional control of the 313 large xut1-operon (possibly xutB, xutC1 - xutI) remains elusive. 314 315 Conservation of xut-loci among Microbacterium sp. 316 To identify additional Microbacterium species carrying a xut-locus, proteins encoded by 317 MMX123_02528–MMX123_02550 were queried against 1,206 available Microbacterium 318 genomes. Twelve genomes contained the core xanthan-degrading enzymes GH9 (XutA), GH3 319 (XutD), GH38 (XutE), and the xanthan lyase PL8 (XutJ) (Fig. 5,(10)). Three of these displayed 320 the type I arrangement, characterized by separation of GH3/GH38 from PL8 and conservation 321 .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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint 15 of intervening genes, as observed in M. xanthanicum UB-LE1, the leaf-derived strains M. sp. 322 Leaf436 and M. testaceum StLB037, and the wastewater-derived M. sp S1037. In contrast, 323 M. testaceum DSM 20166 lacked the locus and was unable to grow on xanthan (Fig. S5). 324 325 Fig. 5: Synteny plot of genes belonging to the xut-locus region in different Microbacterium 326 species. Color coding represents functional annotation according to M. xanthanicum UB-LE1 (see Fig. 327 3B). Chromosome-level genome assemblies are shown in bold. Genomic coordinates are listed in Table 328 S4. Assemblies marked with Type I: Gene organization of the xut-locus in Microbacterium species with 329 the xanthan lyase ( xutJ) is oriented opposite to upstream genes, e.g., GH3 ( xutD), GH38 ( xutE). 330 Assemblies marked with Type II: Gene organization of the xut-locus in Microbacterium species with 331 the xanthan lyase ( xutJ) is co-oriented with upstream genes, e.g., GH3 ( xutD), GH38 ( xutE). The xut-332 locus organization in M. testaceum NS220 could not be classified and is marked with a question mark, 333 as the xanthan lyase (xutJ) is located on a separate contig from the other locus genes. 334 .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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint 16 Eight genomes showed the type II arrangement, in which the genes encoding GH3, 335 GH38, and PL8 are co-located on the same strand. Known xanthan degraders, such as 336 Microbacterium sp. XT11 and fully sequenced strains including M. oleivorans I46 carried this 337 organization, whereas strains lacking the locus, such as M. oleivorans DSM 16091 (Fig. S5), 338 were deficient in xanthan utilization. 339 Across both types, the core gene block from x utR1 to xutE is conserved, whereas 340 downstream genes xutF, xutG, and xutH are variably distributed. The xanthan lyase xutJ is 341 present in all strains but differs in genomic context. The second transport system and its 342 regulator (XutR2) are restricted to type I, while the third transport system and XutR3 are 343 consistently present in type I and variably retained in type II strains. 344 345

346 The identification of a tightly regulated xanthan utilization (xut) locus in M. xanthanicum UB-347 LE1 provides insight into how soil bacteria adapt to structurally complex and spatially 348 heterogeneous carbon sources. From an eco-evolutionary perspective, xanthan represents a 349 particularly challenging substrate: it is chemically recalcitrant (REFs from introduction) and 350 produced in highly localized biological contexts, for example during biofilm formation by 351 Xanthomonas species (3). The strong and selective induction of the xut-locus in response to 352 xanthan (Fig. 2, Fig.3), in contrast to other polysaccharide utilization (pul) loci (Table S1: pul 353 loci), indicates that M. xanthanicum UB-LE1 follows a strategy of conditional investment into 354 extracellular degradation. This strategy is consistent with the high energetic costs associated 355 with secreting large CAZymes, such as PL8 and GH9, which release diffusible oligo- and 356 monosaccharides(21) into the environment. 357 The presence of multiple, co-regulated transport systems within the same locus (Fig. 358 3B, Fig. 4) likely enhances substrate capture and limits loss of released sugars to competing 359 microorganisms. Such coupling of extracellular hydrolysis with high-affinity uptake may 360 stabilize the producer phenotype and reduce exploitation by non-degrading “cheaters” (22). 361 The genomic potential of M. xanthanicum UB-LE1 further suggests a role as an early colonizer 362 .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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint 17 or primary degrader, analogous to pioneer degraders described in marine microbial 363 communities that initiate polymer breakdown and structure subsequent cross-feeding 364 interactions (23). The occurrence of xut-loci in leaf-associated Microbacterium strains (Fig. 5) 365 indicates that xanthan degraders may also influence plant–microbe interactions by 366 destabilizing pathogen biofilms or by redirecting carbon flow through interference competition 367 (24). Such activity could indirectly benefit plant hosts under pathogen pressure, while 368 simultaneously providing the degrader access to released sugars. Co-regulation of a 369 phenylacetic acid degradation operon may further contribute to competitive fitness in biofilm-370 associated niches, where detoxification of antimicrobial compounds can provide a selective 371 advantage (25). Notably, the broad repertoire of additional sugar transporters encoded by M. 372 xanthanicum UB-LE1 suggests that xanthan utilization is embedded within a metabolically 373 flexible, generalist lifestyle rather than representing a narrow specialization, consistent with 374 the enrichment of CAZymes in plant-associated bacteria (26–28). 375 The genomic architecture of the type I xut-locus (Fig. 5) supports the view that xanthan 376 utilization is an adaptive trait shaped by horizontal gene transfer, inducibility, and modular 377 evolution. The reduced GC content (Fig. 3A), tight clustering of functionally linked genes (Fig. 378 3B), and the presence of cognate transcriptional regulators within the locus (Fig. 3B, Fig. 4) 379 are all consistent with acquisition as a pre-assembled functional module. The organization of 380 the xut-locus parallels the general logic of pul-loci described in Bacteroidetes (29), combining 381 secreted and non-secreted CAZymes with tripartite transport systems. Unlike canonical 382 Bacteroidetes pul-loci (30), however, the xut-locus includes LacI-type regulators, here 383 designated XutR1 and XutR3 (Fig. 3B) that likely function as both sensors and transcriptional 384 regulators via fused substrate-binding domains (Table S1: interproscan). At least one such 385 regulator is conserved across both type I and type II xut-locus arrangements in Microbacterium 386 (Fig. 5), suggesting functional constraints on regulatory architecture. The apparent mobility of 387 this system is further supported by recent reports of xanthan utilization loci encoded on 388 plasmids (17). 389 .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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint 18 Comparison of type I xut-loci in Microbacterium with type II arrangements, in which 390 xanthan-associated genes are dispersed across the genome, suggests that complete xanthan 391 utilization requires auxiliary functions beyond the core degradative enzymes. These include 392 enzymes, such as a 2-hydroxyacid reductase (XutB), a FAD-dependent oxidoreductase 393 (XutN), and additional sugar-modifying activities (XutF, XutG, XutI; Table S3, Fig. 5, Fig. 6) 394 that likely enable assimilation of unusual xanthan-derived sugars into central metabolism (Fig. 395 6). 396 397 Fig. 6: Hypothetical model of xanthan degradation and utilization (type I) in M. xanthanicum UB-398 BL1. Extracellular xanthan is initially processed by a secreted xanthan lyase (PL8 family, XutJ), which 399 removes terminal pyruvylated mannose residues, followed by an endoxanthanase (GH9 family, XutA) 400 that cleaves the β-(1,4)-glucan backbone into xanthan-derived oligosaccharides. These 401 oligosaccharides as well as pyruvylated mannose are captured by surface-associated substrate-binding 402 proteins and imported via dedicated transport systems (XutC1/2/3, XutK1/2/3, XutM1/2/3) encoded 403 within the xut-locus. In the cytoplasm, an esterase (likely XutL) removes acetyl groups (Ac), and 404 glycosidases (GH3 family, XutD and GH38 family, XutE) further hydrolyze the oligosaccharides into 405 monosaccharides glucose (G), mannose (M), glucuronic acid (GA), which are subsequently assimilated 406 .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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint 19 into central carbon metabolism. The proteins XutB, XutF, XutG, XutI, and XutN are likely involved in this 407 metabolic conversion. XutH is a candidate enzyme to cleave pyruvylated (Pyr) mannose. The question 408 mark indicates hypothetical or unspecified functions. Expression of the xanthan utilization machinery is 409 tightly controlled by the substrate-responsive transcriptional regulators XutR2 and XutR3 encoded within 410 the xut-locus. The differently colored suns indicate transcriptional control by the respective XutR 411 transcription factor. Protein functions are not functionally tested but suggested according to interproscan 412 predictions. 413 414 Together, our findings position xanthan utilization as a model system for understanding 415 how bacteria adapt to structurally complex, spatially heterogeneous extracellular 416 carbohydrates in natural environments. The xut-locus in M. xanthanicum UB-LE1 illustrates 417 how modular gene clusters, tight regulatory control, and metabolic integration enable 418 conditional investment into energetically costly extracellular degradation strategies. This 419 architecture links ecological opportunity to evolutionary innovation, highlighting how substrate 420 availability can drive the emergence and diversification of specialized catabolic pathways. 421 Beyond advancing fundamental insight into microbial polysaccharide degradation, our work 422 establishes a framework for exploring and engineering xanthan-active systems. The regulatory 423 logic and enzymatic repertoire uncovered here provide entry points for the targeted 424 modification of xanthan structure and properties, connecting microbial ecology with future 425 biotechnological and material applications. 426 427

428 Isolation of M. xanthanicum UB-BL1 and culture conditions 429 Details of sample collection and strain isolation of M. xanthanicum UB-LE1 is described in 430 Laker et al. (Companion paper submitted to MRA). 431 M. xanthanicum UB-LE1 was grown at 30 °C and 180 rpm in defined synthetic M9-432 medium (KH2PO4, 3 g L-1; NaCl, 0.5 g L-1; Na2HPO4, 6.78 g L-1; NH4Cl, 1 g L-1) containing 0.001 433 g L- 1 Biotin and Thiamin, 0.12 g L-1 MgSO4, 0.033 g L-1 CaCl2 and trace elements (EDTA, 0.05 434 g L-1; FeCl3 x 6H2O, 0.0083 g L- 1; ZnCl2, 0.00084 g L-1; CuCl2 x 2H2O, 0.00013 g L-1; CoCl2 x 435 .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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint 20 2H2O, 0.00010 g L-1; H3BO3, 0.00010 g L-1). Experiments were conducted with UB-LE1 cells, 436 starting OD OD600 of 0.1, grown on 5 g L -1 xanthan or 6 g L -1 glucose to assess the influence 437 of the carbon sources. To obtain biological replicates, three or four batches of UB-LE1 cells 438 were grown on each carbon source. 439 440 RNA isolation and RNA-seq analysis 441 Cells for RNA isolation were harvested after 34 h (+ glucose) and 48 h (+ xanthan) by 442 centrifugation at 4 °C, 8,000 x g for 5 min. Pellets were frozen in liquid nitrogen and stored at 443 −80 °C. Total RNA was extracted using the Quick RNA miniprep kit (Zymo Research, USA) 444 following the manufacturer’s instructions. RNA quality and quantity were measured using a 445 Bioanalyzer 2100 (Agilent, USA). rRNA was depleted using RNA Depletion llumina® Ribo-446 Zero Plus (Illumina, USA). Libraries were prepared using the TruSeq RNA Library Prep Kit v2 447 kit (Illumina, USA), pooled, and sequenced 100 bp single-end on an Illumina NextSeq 2000 448 instrument (Illumina, USA). 449 All computational analyses were performed using default parameters unless otherwise 450 specified. Transcripts were quantified with Kallisto v0.44 (31) with parameters ‘--single -l 200 -451 s 20’ and the M. xanthanicum UB-LE1 transcript sequences as index). Downstream analyses 452 were conducted in R v4.5.2 https://www.R-project.org/ (32) and plotted with ggplot2 v3.5.2 453 (33). Genes with zero counts across all samples were excluded prior to analyzes. Principle 454 components were determined with prcomp (scale=T) and differential transcript accumulation 455 was determined with edgeR v4.6.3 (34), classic mode) and p-values were corrected with 456 Benjamini-Hochberg. A circular genome map of UB-LE1 plotted with pyCirclize 457 (github.com/moshi4/pyCirclize) visualized gene content, GC content, and genome-wide log2-458 fold changes 459 460 Protein isolation and proteome analysis 461 Samples were centrifuged at 8,000 x g for 5 min, washed with phosphate-buffered saline (PBS) 462 containing cOmplete Protease Inhibitor (Roche, Germany), and stored at -80 °C until 463 .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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint 21 lyophilized. The supernatant fraction was passed through a sterile 0.2 µm filter (Fisher 464 Scientific) and stored at -80 °C until lyophilized. Six to twelve mg of lyophilized material were 465 resuspended in 1 mL of 100 mM ammonium bicarbonate (AmBic; Sigma-Aldrich), disrupted 466 mechanically using a Precellys 24 homogenizer (Bertin Instruments) with 0.1 mm glass beads 467 (BioSpec Products) at 6,500 rpm for 30 s in three cycles, refrozen in liquid nitrogen, and 468 lyophilized. 469 Proteins were reduced and alkylated by sequential treatment with trifluoroethanol 470 (VWR Chemicals), AmBic, 200 µM dithiothreitol (DTT), and 200 µM iodoacetamide (Sigma-471 Aldrich) for 90 min at RT, and again treated with DTT for 60 min at RT. Tryptic digestion was 472 performed by overnight-incubation at 37 °C with sequencing-grade Trypsin Gold (1 µg µL -1; 473 Promega) in AmBic and H2O (1:1). Peptides were purified with filter-aided sample preparation 474 (Wisniewski et al., 2009; Sep-Pak C18 Vac Cartridges) and reconstituted in 98% LC-MS water, 475 2% acetonitrile, and 0.1% trifluoroacetic acid (VWR Chemicals), diluted to 0.5-1.5 µg µL -1. 476 Peptides were analyzed by LC–ESI–MS/MS on a Q Exactive Plus Orbitrap instrument (Thermo 477 Fisher Scientific), linear gradient from 3.2% to 76% acetonitrile with 0.1% formic acid over 478 80 min, positive ion mode from 5 to 92 min with a resolution of 70,000, AGC target of 3 x 10 6 479 ions, maximum injection time (IT) of 64 ms, and a scan range of 350-2000 m/z. Data-480 dependent acquisition selected the top 10 precursor ions with a maximum IT of 100 ms, AGC 481 target of 2 x 105, isolation window of 1.6 m/z, and MS2 resolution of 17,500. 482 Raw data were processed using MaxQuant version 2.5.2 with default parameters 483 except “Match between runs” enabled. Proteins were considered identified if represented by 484 at least two unique peptides detected in a minimum of two biological replicates. For label-free 485 quantification (LFQ), intensity values were log 2-transformed and statistically significant 486 differential accumulation was determined by Welch’s t-test with missing values omitted 487 followed by Benjamini Hochberg correction (35). Secreted peptides were predicted with 488 DeepLocPro in Gram-positive mode (36), and carbohydrate-active enzymes were annotated 489 with dbCAN3 (37). 490 491 .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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint 22 Synteny analysis 492 The protein sequences of UB-LE1 xut-locus genes were queried against 1,206 downloaded 493 proteomes from NCBI (retrieved 16.10.2025 with taxon filter “Microbacterium”; option 494 “annotated genomes only”) using BLASTP (e-value = 60%). Genes located 10 kbp upstream and downstream of the UB-LE1 xut-496 locus were queried with BLASTP against the proteomes with xut-locus proteins with 497 Microbacterium sp. MM2322 as a negative control (GCF_964186585.1; (38)). The xut-locus 498 regions were plotted with pyGenomeViz (https://github.com/moshi4/pyGenomeViz) with color 499 codes based on BLASTP results. 500 501 DNA-affinity purification sequencing (DAP-seq) and analysis 502 UB-LE1 was cultivated on LB agar medium over night at 30 °C. Genomic DNA was isolated 503 using the Nucleospin Microbial DNA Kit (Macherey-Nagel, Germany, including RNase 504 treatment), assessed on a 0.8 % agarose gel, and quantified using a High Sensitivity (HS) 505 dsDNA assay kit on a Qubit 4 fluorometer (Thermo Fisher Scientific, USA). For DNA library 506 preparation, 5 μg genomic DNA was fragmented, end-repaired, A ‑tailed, adapter ligated and 507 amplified from 15 ng following (39). 508 DAP-seq experiments were conducted as described by Schiller et al. (2025). Briefly, 509 coding sequences of genes MMX123_02528 (xutR1), MMX123_02545 (xutR2), and 510 MMX123_02550 (xutR3) were cloned in frame with an N-terminal Halo-tag encoded in the 511 pFN19A T7 SP6 Flexi vector (Promega, USA; discontinued) using Gibson assembly (40). 512 Primer sequences are listed in Table S5. Halo:XutR1, Halo:XutR2, Halo:XutR3, and a 513 barnase(-) vector control were expressed with the TnT SP6 High-Yield Wheat Germ Protein 514 Expression System (Promega, USA) using 2 µg plasmid DNA per reaction. Proteins were 515 bound to Magnet HaloTag beads (Promega, USA) for 1 h and washed three times with a PBS-516 NP40 solution. Expression and bead-binding of the Halo-Tag fusion proteins were verified by 517 western blot analysis with anti-HaloTag monoclonal antibody (Promega, USA) and Goat Anti-518 Mouse IgG H&L (HRP) (Abcam, UK). For DNA binding, 1 ng amplified DNA library and 1 µg 519 .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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint 23 salmon sperm DNA (Thermo Fisher Scientific, USA) were added to the bead-bound proteins 520 and incubated for 1 h at 19 °C before four washes with PBS-NP40 solution and elution. The 521 optimal number of PCR cycles for final amplification was determined by qPCR in duplicate 522 using 1 µL recovered DNA and Luna Universal qPCR Master Mix (New England Biolabs, USA), 523 Final libraries were amplified from 23 µL of recovered DNA using the Illumina TruSeq Strand 524 A primer (5'-ACACTCTTTCCCTACACGACGCTCTTCCGATCT-3') and the i7 8-baseindex 525 primer (5'-CAAGCAGAAGACGGCATACGAGAT-[i7]-GTGACTGGAGTTCAGACGTGT-3') 526 with 15-19 PCR cycles. For sequencing, 5 µL of amplified DNA for each sample were pooled 527 and purified twice with AMPure XP beads (Beckman Coulter, USA) at a bead-to-DNA volume 528 ratio of 1:1. Library size distribution and DNA concentration were assessed using HS DNA 529 Analysis Kit on a Bioanalyzer (Agilent Technologies, USA). and Qubit dsDNA HS Assay Kit 530 (Thermo Fisher Scientific, USA), respectively before sequencing as single ‑end reads on a 531 NextSeq2000 (Illumina, USA). 532 Sequencing data were processed with the automated DAP-Seq analysis pipeline v1.0.1 533 described in (39) using parameters ‘useDuplicates=True subsample=200000’, with the 534 Brassica napus Darmor-bzh v10 genome ( GCA_905183035.1,(41)) used as the random 535 source file. Briefly, raw reads were trimmed with Trimmomatic v0.39 (42) and aligned to the M. 536 xanthanicum genome (Laker et al., Companion paper submitted to MRA) using Bowtie2 v2.5.4 537 (43) Aligned reads were filtered by mapping quality with samtools v1.20 (44) and sambamba 538 v1.0.1 (45) and then randomly subsampled to 200,000 per sample. Peaks were called with 539 MACS3 v3.0.2 (46), using the corresponding empty-vector control for each TF as input control. 540 Peaks with a fold enrichment < 3 relative to background were discarded. Genes were 541 considered bound if a peak summit overlapped with their promoters, defined as the 150 bp 542 region upstream of the transcription start site. 543 Per-base coverage across the xut-locus was obtained from processed Bowtie2 544 alignments using samtools depth and peak locations and scores were extracted from MACS3 545 narrowPeak output. The xut-locus coordinates and peak widths were displayed as 546 gene/feature models using the R package gggenes https://wilkox.org/gggenes/ (47). 547 .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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint 24 DATA AVAILABILITY 548 The RNA-Seq data of the glucose and xanthan growth conditions are available in 549 GenBank/ENA under BioProject accession number PRJEB110862. The GenBank/ENA 550 accession number for the raw RNA-Seq read data under glucose are ERR16928898, 551 ERR16928899, ERR16928900 and under xanthan are ERR16928901, ERR16928902, 552 ERR16928903. 553 The mass spectrometry proteomics data have been deposited to the 554 ProteomeXchange Consortium via the PRIDE (48) partner repository with the dataset 555 identifier PXD069319. 556 Strain M. xanthanicum UB-LE1 has been deposited in the Deutsche Sammlung von 557 Mikroorganismen und Zellkulturen under the accession number DSM 120304 and in the 558 Laboratorium voor Microbiologie, Universiteit Gent, collection under the accession number 559 LMG 34097. 560 561

562 We thank Ulrike Harke and Sanne Wiersma for the handling of M. xanthanicum UB-LE1 and 563 the class (2023) of the master module ‘Methods and examples of functional genome research’ 564 at Bielefeld University for experimental support. The authors would like to thank Jungbunzlauer 565 Holding AG for their support and for providing a scholarship to fund this project, as well as for 566 supplying “cell free” xanthan. Additionally, we would like to thank Dietmar Althaus for his aid 567 in obtaining sampling permission. We thank the NGS team of the Bielefeld University Omics 568 CF NGS Unit (in development) and CeBiTec as well as the technical staff of the CeBiTec 569 Technology Platform Genomics, particularly Eva Schulte-Bernd, Yvonne Kutter, and Katharina 570 Hanuschka for technical assistance. 571 572 FUNDING 573 This work was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research 574 Foundation) – SFB1535 - Project ID 458090666. We acknowledge Bielefeld University core 575 .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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint 25 funding and support for the publication costs by the Open Access Publication Fund of Bielefeld 576 University and the DFG. This work was supported by the BMBF-funded de.NBI Cloud within 577 the German Network for Bioinformatics Infrastructure (de.NBI). 578 579 AUTHOR CONTRIBUTIONS 580 Bianca Laker, Computational data analysis, Manuscript writing | Michael Thomas, Strain 581 isolation, Proteomics, Manuscript writing | Wiebke Weber, Experiments, Data analysis | Prisca 582 Viehoever, RNA-seq, DAP-seq | Anja Meierhenrich, DAP-seq | Levin J. Klages, Genome 583 sequencing, Computational data analysis | Tobias Busche, Genome sequencing, 584 Computational data analysis | Karsten Niehaus, Experimental design, Supervision | Andrea 585 Bräutigam, Experimental design, Data analysis, Supervision, Manuscript writing | Marion 586 Eisenhut, Experimental design, Data analysis, Supervision, Manuscript writing 587 588 AUTHOR ORCIDs 589 Bianca Laker ORCID: 0000-0002-5792-0102 590 Michael Thomas ORCID: 0000-0003-4646-5324 591 Wiebke Weber ORCID: 0009-0007-0721-3163 592 Prisca Viehoever ORCID: 0000-0003-3286-4121 593 Anja Meierhenrich ORCID: 0000-0001-6721-4220 594 Levin J. Klages ORCID: 0000-0003-1634-3705 595 Tobias Busche ORCID: 0000-0001-9211-8927 596 Karsten Niehaus ORCID: 0000-0003-4078-9870 597 Andrea Bräutigam ORCID: 0000-0002-5309-0527 598 Marion Eisenhut ORCID: 0000-0002-2743-8630 599 600 601 602 603 .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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint 26 ADDITIONAL FILES 604 The following material is available online. 605 Supplemental material 606 Table S1: Additional data with mothertable, secreted CAZymes, CAZyme analysis, pul loci, 607 low GC content regions 608 Table S2: Average nucleotide identity (ANI) values for Microbacterium xanthanicum UB-LE1 609 with closely related species. 610 Table S3: Genes of the xut-locus in M. xanthanicum UB-BL1. 611 Table S4: Genomic coordinates of sequence regions used for synteny analyses of the xut-612 locus. 613 Table S5: Primers used during initial PCR for TF candidate amplification. 614 Figure S1: Features of M. xanthanicum UB-LE1. 615 Figure S2: Scatterplot comparing fold-changes between transcriptome and cellular proteome. 616 Figure S3: Pie chart representing the number of proteins assigned to CAZyme classes in M. 617 xanthanicum UB-LE1. 618 Figure S4: Western blotting to prove successful in vitro expression of XutR1, XutR2, and 619 XutR3. 620 Figure S5: Ability of Microbacterium species to utilize xanthan as carbon source. 621 622

623 1. Dow JM, Daniels MJ. 1994. Pathogenicity determinants and global regulation of 624 pathogenicity of Xanthomonas campestris pv. campestris. Curr Top Microbiol Immunol 625 192:29–41. 626 627 2. Yun MH, Torres PS, El Oirdi M, Rigano LA, Gonzalez-Lamothe R, Marano MR, 628 Castagnaro AP, Dankert MA, Bouarab K, Vojnov AA. 2006. Xanthan induces plant 629 susceptibility by suppressing callose deposition. Plant Physiol 141:178–87. 630 631 3. Torres PS, Malamud F, Rigano LA, Russo DM, Marano MR, Castagnaro AP, 632 Zorreguieta A, Bouarab K, Dow JM, Vojnov AA. 2007. Controlled synthesis of the DSF 633 cell-cell signal is required for biofilm formation and virulence in Xanthomonas 634 campestris. Environ Microbiol 9:2101–9. 635 636 4. Berninger T, Dietz N, González López Ó. 2021. Water ‑soluble polymers in agriculture: 637 xanthan gum as eco‑friendly alternative to synthetics. Microb Biotechnol 14:1881–638 1896. 639 .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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint 27 640 5. Freire CD, Pinto FR, Almeida D, Gil MM. 2024. Linseed and xanthan gum in algae 641 pasta: textural, sensory and antioxidant characteristics. Int J Food Sci Technol 642 59:5477–5489. 643 644 6. Krystyjan M, Dobosz-Kobędza A, Sikora M, Baranowska HM. 2022. Influence of 645 Xanthan Gum Addition on the Short- and Long-Term Retrogradation of Corn Starches 646 of Various Amylose Content. Polymers (Basel) 14:452. 647 648 7. Singhvi G, Hans N, Shiva N, Kumar Dubey S. 2019. Xanthan gum in drug delivery 649 applications, p. 121–144. In Natural Polysaccharides in Drug Delivery and Biomedical 650 Applications. Elsevier. 651 652 8. Berezina O V., Rykov S V., Schwarz WH, Liebl W. 2024. Xanthan: enzymatic 653 degradation and novel perspectives of applications. Appl Microbiol Biotechnol 654 108:227. 655 656 9. Moroz O V., Jensen PF, McDonald SP, McGregor N, Blagova E, Comamala G, 657 Segura DR, Anderson L, Vasu SM, Rao VP, Giger L, Sørensen TH, Monrad RN, 658 Svendsen A, Nielsen JE, Henrissat B, Davies GJ, Brumer H, Rand KD, Wilson KS. 659 2018. Structural Dynamics and Catalytic Properties of a Multimodular Xanthanase. 660 ACS Catal 8:6021–6034. 661 662 10. Sun Z, Liu H, Wang X, Yang F, Li X. 2019. Proteomic Analysis of the Xanthan-663 Degrading Pathway of Microbacterium sp. XT11. ACS Omega 4:19096–19105. 664 665 11. Muchová M, Růžička J, Julinová M, Doležalová M, Houser J, Koutný M, Buňková L. 666 2009. Xanthan and gellan degradation by bacteria of activated sludge. Water Science 667 and Technology 60:965–973. 668 669 12. Nankai H, Hashimoto W, Miki H, Kawai S, Murata K. 1999. Microbial System for 670 Polysaccharide Depolymerization: Enzymatic Route for Xanthan Depolymerization by 671 Bacillus sp. Strain GL1. Appl Environ Microbiol 65:2520–2526. 672 673 13. Qian F, An L, Wang M, Li C, Li X. 2007. Isolation and characterization of a xanthan-674 degrading Microbacterium sp. strain XT11 from garden soil. J Appl Microbiol 675 102:1362–1371. 676 677 14. Khianngam S, Akaracharanya A, Tanasupawat S, Lee KC, Lee J-S. 2009. 678 Paenibacillus thailandensis sp. nov. and Paenibacillus nanensis sp. nov., xylanase-679 producing bacteria isolated from soil. Int J Syst Evol Microbiol 59:564–568. 680 681 15. Jensen PF, Kadziola A, Comamala G, Segura DR, Anderson L, Poulsen J-CN, 682 Rasmussen KK, Agarwal S, Sainathan RK, Monrad RN, Svendsen A, Nielsen JE, Lo 683 Leggio L, Rand KD. 2019. Structure and Dynamics of a Promiscuous Xanthan Lyase 684 from Paenibacillus nanensis and the Design of Variants with Increased Stability and 685 Activity. Cell Chem Biol 26:191-202.e6. 686 687 16. Han R, Baudrexl M, Ludwig C, Berezina O V., Rykov S V., Liebl W. 2024. 688 Identification of a novel xanthan-binding module of a multi-modular Cohnella sp. 689 xanthanase. Front Microbiol 15:1386552. 690 691 17. Thomas M, Schlüter A, Fjodorova J, Rückert C, Busche T, Niehaus K. 2026. Genomic 692 and proteomic characterization of a newly isolated Paenarthrobacter ilicis strain and 693 its plasmid-mediated xanthan degradation. Microbiol Spectr 14:e0007526. 694 695 .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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint 28 18. McKee LS, La Rosa SL, Westereng B, Eijsink VG, Pope PB, Larsbrink J. 2021. 696 Polysaccharide degradation by the Bacteroidetes: mechanisms and nomenclature. 697 Environ Microbiol Rep 13:559–581. 698 699 19. Reverchon S, Nasser W, Robert ‑Baudouy J. 1991. Characterization of kdgR , a gene 700 of Erwinia chrysanthemi that regulates pectin degradation. Mol Microbiol 5:2203–701 2216. 702 703 20. Hirooka K. 2022. RhgKL and CcpA directly regulate the rhiLFGN-rhgR-yesTUV 704 operon involved in the metabolism of rhamnogalacturonan type I in Bacillus subtilis. 705 Biosci Biotechnol Biochem 86:1383–1397. 706 707 21. Cezairliyan B, Ausubel FM. 2017. Investment in secreted enzymes during nutrient-708 limited growth is utility dependent. Proceedings of the National Academy of Sciences 709 114:E7796–E7802. 710 711 22. Allison SD. 2005. Cheaters, diffusion and nutrients constrain decomposition by 712 microbial enzymes in spatially structured environments. Ecol Lett 8:626–635. 713 714 23. Enke TN, Datta MS, Schwartzman J, Cermak N, Schmitz D, Barrere J, Pascual-García 715 A, Cordero OX. 2019. Modular Assembly of Polysaccharide-Degrading Marine 716 Microbial Communities. Current Biology 29:1528-1535.e6. 717 718 24. Rendueles O, Ghigo J-M. 2015. Mechanisms of Competition in Biofilm Communities. 719 Microbiol Spectr 3. 720 721 25. Hooppaw AJ, McGuffey JC, Di Venanzio G, Ortiz-Marquez JC, Weber BS, Lightly TJ, 722 van Opijnen T, Scott NE, Cardona ST, Feldman MF. 2022. The Phenylacetic Acid 723 Catabolic Pathway Regulates Antibiotic and Oxidative Stress Responses in 724 Acinetobacter. mBio 13:e0186321. 725 726 26. Levy A, Salas Gonzalez I, Mittelviefhaus M, Clingenpeel S, Herrera Paredes S, Miao 727 J, Wang K, Devescovi G, Stillman K, Monteiro F, Rangel Alvarez B, Lundberg DS, Lu 728 T-Y, Lebeis S, Jin Z, McDonald M, Klein AP, Feltcher ME, Rio TG, Grant SR, Doty SL, 729 Ley RE, Zhao B, Venturi V, Pelletier DA, Vorholt JA, Tringe SG, Woyke T, Dangl JL. 730 2018. Genomic features of bacterial adaptation to plants. Nat Genet 50:138–150. 731 732 27. Zhalnina K, Louie KB, Hao Z, Mansoori N, da Rocha UN, Shi S, Cho H, Karaoz U, 733 Loqué D, Bowen BP, Firestone MK, Northen TR, Brodie EL. 2018. Dynamic root 734 exudate chemistry and microbial substrate preferences drive patterns in rhizosphere 735 microbial community assembly. Nat Microbiol 3:470–480. 736 737 28. Vorholt JA. 2012. Microbial life in the phyllosphere. Nat Rev Microbiol 10:828–840. 738 739 29. Grondin JM, Tamura K, Déjean G, Abbott DW, Brumer H. 2017. Polysaccharide 740 Utilization Loci: Fueling Microbial Communities. J Bacteriol 199:199(15):e00860–16. 741 742 30. Terrapon N, Lombard V, Gilbert HJ, Henrissat B. 2015. Automatic prediction of 743 polysaccharide utilization loci in Bacteroidetes species. Bioinformatics 31:647–655. 744 745 31. Bray NL, Pimentel H, Melsted P, Pachter L. 2016. Near-optimal probabilistic RNA-seq 746 quantification. Nat Biotechnol 34:525–527. 747 748 32. R Core Team. 2025. R: A Language and Environment for Statistical Computing. 749 Vienna, Austria. 750 751 .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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint 29 33. Wickham H. 2016. ggplot2. Springer International Publishing, Cham. 752 753 34. Chen Y, Chen L, Lun ATL, Baldoni PL, Smyth GK. 2025. edgeR v4: powerful 754 differential analysis of sequencing data with expanded functionality and improved 755 support for small counts and larger datasets. Nucleic Acids Res 53: gkaf018 756 757 35. Diz AP, Carvajal-Rodríguez A, Skibinski DOF. 2011. Multiple Hypothesis Testing in 758 Proteomics: A Strategy for Experimental Work. Molecular & Cellular Proteomics 759 10:M110.004374. 760 761 36. Nielsen H. 2025. Practical Applications of Language Models in Protein Sorting 762 Prediction: SignalP 6.0, DeepLoc 2.1, and DeepLocPro 1.0. Methods Mol Biol 153–763 175. 764 765 37. Zheng J, Ge Q, Yan Y, Zhang X, Huang L, Yin Y. 2023. dbCAN3: automated 766 carbohydrate-active enzyme and substrate annotation. Nucleic Acids Res 51:W115–767 W121. 768 769 38. Eisenhut M, Schulze T, Matten FNJ, Sengewald A, Böhner P, Falk D, Klinge CV, 770 Langschwager R, Lohoff S, Meyer N, Ponomarchuk V, Pünter F, Sherlock J, Pschik D, 771 Klages LJ, Busche T, Wobbe L, Bräutigam A, Laker B. 2024. Complete genome 772 sequences of Chitinophaga sp. strain MM2321 and Microbacterium sp. strain 773 MM2322, isolated from a sand sample in Harsewinkel, Germany. Microbiol Resour 774 Announc 13: e0070024. 775 776 39. Schiller K, Meierhenrich A, Zenker S, Sielmann LM, Laker B, Bräutigam A. 2025. 777 Optimized DNA affinity purification sequencing determines relative binding affinity of 778 transcription factors. Biol Methods Protoc 10:bpaf082. 779 780 40. Gibson DG, Young L, Chuang R-Y, Venter JC, Hutchison CA, Smith HO. 2009. 781 Enzymatic assembly of DNA molecules up to several hundred kilobases. Nat Methods 782 6:343–345. 783 784 41. Rousseau-Gueutin M, Belser C, Da Silva C, Richard G, Istace B, Cruaud C, Falentin 785 C, Boideau F, Boutte J, Delourme R, Deniot G, Engelen S, de Carvalho JF, 786 Lemainque A, Maillet L, Morice J, Wincker P, Denoeud F, Chèvre A-M, Aury J-M. 787 2020. Long-read assembly of the Brassica napus reference genome Darmor-bzh. 788 Gigascience 9:giaa137 789 790 42. Bolger AM, Lohse M, Usadel B. 2014. Trimmomatic: a flexible trimmer for Illumina 791 sequence data. Bioinformatics 30:2114–2120. 792 793 43. Langmead B, Salzberg SL. 2012. Fast gapped-read alignment with Bowtie 2. Nat 794

9:357–359. 795 796 44. Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, Marth G, Abecasis G, 797 Durbin R. 2009. The Sequence Alignment/Map format and SAMtools. Bioinformatics 798 25:2078–2079. 799 800 45. Tarasov A, Vilella AJ, Cuppen E, Nijman IJ, Prins P. 2015. Sambamba: fast 801 processing of NGS alignment formats. Bioinformatics 31:2032–2034. 802 803 46. Zhang Y, Liu T, Meyer CA, Eeckhoute J, Johnson DS, Bernstein BE, Nusbaum C, 804 Myers RM, Brown M, Li W, Liu XS. 2008. Model-based Analysis of ChIP-Seq (MACS). 805 Genome Biol 9:R137. 806 807 .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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint 30 47. Wilkins D. 2023. gggenes: Draw Gene Arrow Maps in “ggplot2.” R package version 808 0.5.0. 809 810 48. Perez-Riverol Y, Bandla C, Kundu DJ, Kamatchinathan S, Bai J, Hewapathirana S, 811 John NS, Prakash A, Walzer M, Wang S, Vizcaíno JA. 2025. The PRIDE database at 812 20 years: 2025 update. Nucleic Acids Res 53:D543–D553. 813 814 .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 April 22, 2026. ; https://doi.org/10.64898/2026.04.20.719765doi: bioRxiv preprint