Abstract

   Lignin, a major plant cell wall component, has an important role in
   plant-defense mechanisms against pathogens and is a promising renewable
   carbon source to produce bio-based chemicals. However, our
   understanding of microbial metabolism is incomplete regarding certain
   lignin-related compounds like p-coumaryl and sinapyl alcohols. Here, we
   reveal peripheral pathways for the catabolism of the three main lignin
   precursors (p-coumaryl, coniferyl, and sinapyl alcohols) in the plant
   pathogen Xanthomonas citri. Our study demonstrates all the necessary
   enzymatic steps for funneling these monolignols into the tricarboxylic
   acid cycle, concurrently uncovering aryl aldehyde reductases that
   likely protect the pathogen from aldehydes toxicity. It also shows that
   lignin-related aromatic compounds activate transcriptional responses
   related to chemotaxis and flagellar-dependent motility, which might
   play an important role during plant infection. Together our findings
   provide foundational knowledge to support biotechnological advances for
   both plant diseases treatments and conversion of lignin-derived
   compounds into bio-based chemicals.

   Subject terms: Environmental microbiology, Biocatalysis, Metabolic
   pathways
     __________________________________________________________________

   Plants strengthen their cell walls with lignin to prevent pathogen
   colonization. Here, Martim et al show how Xanthomonas citri degrades
   lignin precursors and detoxifies lignin-derived molecules to overcome
   this plant defense.

Introduction

   The bioconversion of aromatic compounds has a central role in carbon
   cycling^[52]1, plant-pathogen interactions^[53]2–[54]4 and
   detoxification of organic pollutants^[55]5. This process typically
   starts in upper pathways (also named peripheral pathways), which
   converge the structural diversity of aromatic compounds to fewer
   intermediate metabolites, which are further funneled to central carbon
   metabolites through a narrower range of lower pathways^[56]6,[57]7.
   During this convergent process, industrially relevant molecules are
   formed. This offers opportunities for engineering microbial chassis to
   produce chemicals from complex mixtures of aromatic compounds derived
   from abundant wastes such as lignin and mixed-plastics^[58]8,[59]9.

   Hundreds of microorganisms across diverse phyla are known to have the
   potential to metabolize lignin-related monomers, according to the
   eLignin database^[60]7. However, the metabolic pathways of only a few
   species have been characterized so far^[61]7,[62]10. Lignin-related
   monomers include the three main lignin precursors (p-coumaryl,
   coniferyl and sinapyl alcohols) and their respective p-hydroxyphenyl
   (H), guaiacyl (G) and sinapyl (S) derivatives. Although pathways for
   the catabolism of several lignin-related monomers have been described
   in the literature, information on upper pathways for some of them, such
   as p-coumaryl alcohol, sinapyl alcohol and sinapaldehyde, is still
   lacking^[63]11–[64]13.

   The molecular mechanisms related to the bioconversion of aromatic
   compounds have been mainly studied in some model organisms, such as
   Pseudomonas putida KT2440^[65]14,[66]15 and Sphingobium sp.
   SYK-6^[67]16,[68]17, besides other species from the Rhodococcus^[69]18
   and Burkholderia^[70]19,[71]20 genera. These models have been isolated
   either from soil or industrial wastewater, so our understanding of how
   microorganisms from other ecological niches metabolize lignin-related
   compounds remains elusive. Moreover, how these molecules impact
   microbial behavior and physiology is still partially
   understood^[72]21,[73]22.

   Some plant pathogens, such as Xanthomonas species, have a vast arsenal
   of enzymes to degrade components of the plant cell wall such as
   xyloglucan^[74]23 and xylan^[75]24, using the released carbohydrates as
   a source of carbon, energy, and stimuli^[76]25. However, little is
   known about their capacity and molecular strategies to metabolize other
   compounds available in the plant cell wall, especially the phenolics
   related to lignin, a major plant cell wall component^[77]26–[78]29. One
   of the plant defense mechanisms against Xanthomonas infection is to
   increase the lignification of the plant cell wall, which implies an
   increased secretion of monolignols in the infection site^[79]30.
   Therefore, monolignol degradation could be an effective way for the
   pathogen to inhibit lignification. However, it is still unknown whether
   Xanthomonas species are biochemically capable of adopting such a
   strategy.

   In this work, by combining RNA sequencing (RNA-seq) analysis,
   biochemical characterization, and gene knockout studies, we
   investigated the metabolism of lignin-related aromatics in the model
   phytopathogen Xanthomonas citri subsp. citri 306 (X. citri 306). Our
   data revealed missing steps and complete pathways for the catabolism of
   the three main lignin precursors, as well as reductive metabolic
   pathways and efflux approaches to cope with aryl aldehyde toxicity.
   This study also showed that lignin-related compounds activate
   transcriptional responses related to chemotaxis and flagellar-dependent
   motility in the phytopathogen, which might have an important role
   during the infection of the plant host. In summary, this work provides
   insights into the molecular mechanisms involved in plant-pathogen
   interactions and adds missing pieces to the known spectrum of molecular
   strategies for the bioconversion of lignin-related compounds.

Results

The model plant pathogen metabolizes a diverse range of lignin-related
aromatics

   To investigate if X. citri 306 can grow using lignin-related aromatic
   compounds as the main carbon source, we performed growth assays in
   minimal medium supplemented with 21 aromatic monomers, representative
   of H, G and S units (Fig. [80]1). Additionally, we analyzed two complex
   samples of lignin-derived compounds (LDC-I and LDC-II) produced from
   sugarcane bagasse (Supplementary Fig. [81]1 and Supplementary
   Table [82]1). The main aromatic monomers detected in the LDC-I sample
   were p-coumarate, ferulate, 4-hydroxybenzaldehyde, vanillin, along with
   organic acids (acetate, formate, lactate) and sugars (arabinose,
   glucose) (Supplementary Table [83]1). In LDC-II, the most abundant
   molecules detected were acetate and the aromatic compounds catechol,
   3-methoxy-catechol, phenol, guaiacol and pyrogallol (Supplementary
   Table [84]1).

Fig. 1. Summary of growth conditions and analyses of aromatics toxicity and
depletion.

   [85]Fig. 1
   [86]Open in a new tab

   Growth assays in minimal medium (XVM2m) and XVM2m plus 5 mmol L^−^1
   glucose (XVM2m(G)) supplemented with 5 mmol L^−^1 of lignin-related
   aromatic compounds or 1 g L^−^1 LDC-I or 0.3 g L^−^1 LDC-II. LDC =
   lignin-derived compounds. LDC-I is a residual liquid stream resulting
   from the acid precipitation of lignin from an alkaline liquor of
   sugarcane bagasse. LDC-II is a bio-oil rich in aromatic monomers,
   obtained by hydrothermal depolymerization of an alkaline lignin from
   sugarcane bagasse (Supplementary Fig. [87]1 and Supplementary
   Table [88]1). The red gradient represents differences in growth
   parameters from low (blank) to high (red). HPLC analyses were performed
   using XVM2m(G) medium plus 2 mmol L^−^1 aryl alcohols, 1 mmol L^−^1
   aldehydes, 5 mmol L^−^1 acids, or 50 µmol L^−^1 of either aryl alcohols
   or acids. The blue color bar represents the percentage of aromatic
   compound depleted by X. citri 306 after around 15 h of growth (as
   detailed in Supplementary Table [89]2). Conc. concentration, N.A. not
   analyzed. Growth data are shown as mean ± SD of n = 3 or n = 4
   biological replicates.

   Under the tested conditions, only 4-hydroxybenzoate (4HBA), LDC-I and
   LDC-II supported the growth of X. citri 306 (Fig. [90]1, Fig. [91]2a
   and Supplementary Fig. [92]2). After 30 h of X. citri 306 growth in the
   LDC-I condition, we observed the total depletion of
   4-hydroxybenzaldehyde, glucose, acetate, and formate (Supplementary
   Fig. [93]3). Lactate, succinate and arabinose were partially depleted
   (40–55%), with negligible depletion of hydroxycinnamic acids
   (Supplementary Fig. [94]3). In the LDC-II condition added by glucose
   (5 mmol L^−1), we observed the total depletion of hydroquinone and
   acetate, and partial depletion of 3-methoxycatechol (59%),
   4-methylcatechol (40%), and catechol (11%) (Supplementary Fig. [95]3).

Fig. 2. X. citri 306 grows using 4-hydroxybenzoate and lignin-derived
compounds as carbon sources and metabolizes the three main monolignols.

   [96]Fig. 2
   [97]Open in a new tab

   a Growth curves (hours) in XVM2m supplemented with 4-hydroxybenzoate
   (4HBA) and two complex mixtures of lignin-derived compounds (LDC-I and
   LDC-II). (G) – glucose-supplemented media. b HPLC chromatograms showing
   the depletion of the three monolignols and production of intermediate
   metabolites at 20 h post inoculation. 4HBZ 4-hydroxybenzaldehyde,
   p-CaLC p-coumaryl alcohol, p-Ca p-coumarate, p-CaLD p-coumaraldehyde,
   COALC coniferyl alcohol, FA ferulate, SiA sinapate, 3OMG
   3-O-methylgallate, SYR syringate, SINALC sinapyl alcohol. Growth data
   are shown as mean ± SD of n = 3 biological replicates. Source data are
   provided as a source data file.

   To evaluate whether the lack of growth in some conditions was due to
   toxicity or to insufficient carbon and energy supply, we repeated the
   assay supplementing the medium with glucose. Glucose supplementation
   allowed the growth of X. citri 306 in the presence of the model
   aromatic compounds, except for those displaying severe toxicity at
   5 mmol L^−1, such as some aryl aldehydes (Fig. [98]1 and Supplementary
   Fig. [99]2). Overall, the presence of 5 mmol L^−1 of the individual
   aromatic compounds decreased the growth rate when compared to the
   medium containing only glucose (XVM2m(G)), indicating toxicity. The
   aryl aldehydes displayed a more toxic effect than the correspondent
   aryl alcohols and aryl acids (Fig. [100]1 and Supplementary
   Fig. [101]2).

   Next, we investigated if the aromatic compounds are depleted from the
   XVM2m(G) medium by X. citri 306, indicating they are either modified
   into other metabolites or funneled to the central carbon metabolism.
   For this purpose, we analyzed by HPLC the medium supernatant before and
   after bacterial growth using two input concentrations (millimolar
   and/or micromolar) of aromatic compounds (Fig. [102]1 and Supplementary
   Table [103]2). In micromolar aryl alcohol cultivations, only the three
   monolignols (p-coumaryl, coniferyl, and sinapyl alcohol) were
   effectively depleted from the medium by X. citri 306, generating
   oxidized metabolites detected in the medium (Fig. [104]1, Fig. [105]2b,
   and Supplementary Table [106]2). For all the tested aldehydes, a
   depletion higher than 70% was observed at the millimolar condition
   (except for 4-hydroxybenzaldehyde, <30%) (Fig. [107]1 and Supplementary
   Table [108]2). Among the aryl acids tested, X. citri 306 substantially
   depleted only 4-hydroxybenzoate. Together, these results indicate that
   X. citri 306 has pathways to metabolize aromatic compounds as complex
   as the monolignols and effectively metabolizes a more diverse range of
   aryl aldehydes compared to aryl alcohols and aryl acids provided in the
   culture medium (Fig. [109]1 and Supplementary Table [110]2).

Lignin-related compounds induce chemotaxis and flagellar-dependent motility

   To investigate how X. citri 306 responds to lignin-related compounds
   and identify the metabolic pathways involved in their metabolism, we
   performed RNA-seq studies. Six model compounds (coniferyl alcohol,
   4-hydroxybenzoate, 4-hydroxybenzaldehyde, vanillin, syringaldehyde, and
   benzaldehyde) were selected for the RNA-seq analyses as representatives
   of both different molecular structures (H, G and S-type units) and
   different entry points in the metabolic pathways. We also included in
   the tested conditions three complex mixtures of aromatic compounds
   (LDC-I, LDC-II, and aldehydes mix) (Supplementary Table [111]3).

   A total of 278 to 1285 differentially expressed genes (DEGs) in the
   aromatic-containing conditions compared to the control XVM2m(G) were
   identified (Supplementary Data [112]1), evidencing the importance of
   these compounds in modulating various physiological processes. X. citri
   306 discerned subtle structural variations among aromatic compounds, as
   shown by the incomplete overlap of upregulated genes in each condition
   (Fig. [113]3). For example, 4-hydroxybenzaldehyde activated the
   expression of around 400 genes that were not activated by
   4-hydroxybenzoate, although they share a subset of about 200 genes
   activated by both (Fig. [114]3a). This discrepancy suggests that,
   although very similar in terms of molecular structure,
   4-hydroxybenzaldehyde and 4-hydroxybenzoate can elicit substantially
   distinct transcriptional responses in X. citri 306, which might be
   related to the high toxicity of the aldehyde. Partially overlapping
   responses were also observed when comparing the same type of phenolic
   compound with distinct degrees of methoxylation (Fig. [115]3b) and
   complex mixtures with different compositions (Fig. [116]3c,
   Supplementary Table [117]1).

Fig. 3. Transcriptional responses triggered by lignin-related aromatic
compounds.

   [118]Fig. 3
   [119]Open in a new tab

   Venn diagrams comparing the distribution of unique and shared
   upregulated genes in conditions containing (a) 4-hydroxybenzaldehyde
   (4HBZ), benzaldehyde (BZD), or 4-hydroxybenzoate (4HBA) (b) 4HBZ,
   vanillin (VAN), or syringaldehyde (SYALD) and (c) Lignin-derived
   compounds - LDC-I and LDC-II conditions. d Gene ontology (GO)
   enrichment analysis of up regulated genes based on one-side Fisher’s
   exact test as implemented in the clusterProfiler R package^[120]77.
   Circles’ size and color represent the counts and adjusted p-values,
   respectively. Gene ratio corresponds to the number of DEGs related to a
   GO term divided by the total number of genes associated with that GO
   term in the X. citri 306 genome. The differential expression in each
   condition was compared to XVM2m(G) following the criteria log[2] Fold
   Change ≥1 and p-adjusted ≤ 0.05. The analysis of other conditions is in
   Supplementary Fig. [121]4. e Gene set enrichment analysis based on
   weighted Kolmogorov–Smirnov statistic and Over Representation Analysis
   (ORA) to define modules function as implemented in the CEMiTool
   package^[122]31. The size and intensity of the circles correspond to
   the normalized enrichment score (NES) for the module in each condition,
   indicating biological functions enriched in each module. Positive NES
   reflects transcriptional activity above the median, whereas negative
   NES corresponds to transcriptional activity below the median in each
   condition. COALC coniferyl alcohol, MIX aldehyde mixture. Source data
   are provided as a source data file.

   Although X. citri 306 recognizes subtle changes in the structure of
   lignin-related aromatic compounds, Gene Ontology (GO) enrichment
   analysis revealed that biological processes such as signal
   transduction, bacterial-type flagellum assembly, bacterial-type
   flagellum-dependent cell motility, and chemotaxis were enriched and
   upregulated in all conditions featuring lignin-related aromatics
   (Fig. [123]3d, Supplementary Fig. [124]4 and Data [125]2), except in
   the LDC-I and LDC-II conditions. This might be due to the low
   individual concentration of aromatic inducers or to the interference of
   non-aromatic molecules present in these mixtures (Supplementary
   Fig. [126]3).

   In the conditions containing individual aromatic compounds, the
   upregulation of chemotaxis and flagellar genes, including cheAZY
   (XAC1930-32), motAB (XAC3693-94), fliC (XAC1975), flgL (XAC1976), flgG
   (XAC1981), and flgE (XAC1983) suggests that the sensing of
   lignin-related aromatics stimulates a motile state in X. citri 306
   (Supplementary Data [127]1). This observation is consistent with the
   results of the co-expression analysis, where a co-expressed gene module
   (M1), mainly composed of genes involved in flagellar assembly and
   chemotaxis, was identified (Fig. [128]3e). Higher activity within this
   module became especially pronounced in the presence of
   4-hydroxybenzaldehyde (4HBZ) and syringaldehyde (SYALD) while showing
   reduced activity under conditions involving complex lignin-derived
   samples (LDC-I and LDC-II), which can be due to the comparatively lower
   concentration of individual aromatic compounds within these samples or
   to signal interference from other molecules present in these mixtures
   (Fig. [129]3e and Supplementary Fig. [130]3). Between the downregulated
   processes, translation was consistently enriched in all the tested
   conditions, except for 4HBA (Supplementary Fig. [131]4 and
   Data [132]2).

The first steps of monolignols catabolism are performed by aryl alcohol and
aryl aldehyde dehydrogenases

   The first step of coniferyl alcohol catabolism can be catalyzed by an
   NAD^+ dependent aryl alcohol dehydrogenase (ADH), generating
   coniferaldehyde, which is then converted to ferulate by a
   NAD^+-dependent aryl aldehyde dehydrogenase (ALDH)^[133]32,[134]33. For
   p-coumaryl and sinapyl alcohols, this information is still missing, but
   considering their chemical similarity to coniferyl alcohol, we
   hypothesized that their catabolism might follow a similar pathway.
   Thus, to uncover the genes responsible for monolignols catabolism, we
   searched for ADH and ALDH genes upregulated in the presence of
   lignin-related compounds in X. citri 306. Based on their higher
   upregulation levels, genomic context, and the presence of common
   catalytic domains reported for dehydrogenases active on aromatics, we
   selected eight ADH genes and three ALDH genes for cloning, heterologous
   expression, and biochemical activity screening (Fig. [135]4).

Fig. 4. Transcriptomic analysis and activity screening reveal novel ADH and
ALDH enzymes active on aromatic compounds.

   [136]Fig. 4
   [137]Open in a new tab

   The heatmap presents transcription levels (log[2]FC = log[2] Fold
   Change), comparing each growth condition to the reference (XVM2M(G))
   from at least n = 3 biological replicates. Genes were classified as
   upregulated according to the following criteria: log[2]FC ≥ 1, with an
   p-adjusted ≤ 0.05. log[2] Fold Change was calculated using
   edgeR^[138]76 based on likelihood ratio test within a negative binomial
   generalized log-linear model framework. The activity screening was
   performed using purified enzymes, as detailed in Supplementary
   Data [139]3 and Supplementary Tables [140]4 and [141]5, or whole cells
   assays (XAC0129 and XAC0882) as detailed in Supplementary Table [142]6.
   N.S. indicates proteins that were insoluble in E. coli. N.D. indicates
   soluble proteins that did not display activity in the tested
   conditions. % ID = amino acid sequence identity with the most similar
   enzyme sequence listed in the eLignin database (GenBank accession
   number in parentheses). 4HBZ 4-hydroxybenzaldehyde,
   4HBA  4-hydroxybenzoate, COALC coniferyl alcohol, VAN vanillin, SYALD
   syringaldehyde, BZD benzaldehyde. The ADH (green box) enzymes were
   subjected to screening for both direct and reverse reactions, using the
   corresponding substrates listed in the legend. The ALDH (pink box)
   enzymes were screened only for aldehyde dehydrogenation. Colored
   circles indicate the substrates and labels indicate the respective
   co-substrate with which the enzymes were active.

   The activity screening revealed several alcohol and aldehyde
   dehydrogenases active on aromatic compounds, with variable preferences