Xylan

The feed medium used in SHIME^® was prepared with distilled water, supplemented with 3 g/L starch (Maizena^®, Unilever Brazil, SP, Brazil), 2 g/L pectin (Sigma-Aldrich, Burlington, MA, USA), 4 g/L gastric mucin type II swine (Sigma-Aldrich), 1 g/L xylan (Megazyme, Bray, Ireland), 1 g/L peptone (Kasvi, São José do Pinhais, PR, Brazil), 1 g/L arabinogalactan (Sigma-Aldrich), 0.4 g/L glucose (Synth, São Paulo, SP, Brazil), 3 g/L yeast extract (Kasvi), and 0.5 g/L L-cysteine (Sigma-Aldrich) [26 (link)].
Stomach conditions were simulated in reactor 1 (R1) and the pH value was adjusted by adding HCl, along with the carbohydrate-based medium. In order to reach the duodenum conditions, in the second reactor (R2), an artificial pancreatic juice was added, composed of 12.5 g/L sodium bicarbonate (LS Chemicals, Maharashtra, India), 6 g/L ox-bile (Sigma-Aldrich), and 0.9 g/L pancreatin (Sigma-Aldrich) [26 (link)]. This carbohydrate-based medium plays an important role in the environmental adaptation and inoculum growth with the formation of a stable and representative community [24 (link)].

Xylan from birchwood, beechwood, and Larchwood, were purchased from Sigma Aldrich (Saint Louis, Missouri). Xylan from quinoa (Chenopodium quinoa) stalks was extracted as described in our previous work.^{10 (link)} Analytical grade xylose, xylobiose, xylotriose, xylotetraose, xylopentaose, and xylohexaose were obtained from Megazyme (Wicklow, Ireland).

We used the wildtype (WT) strain Caulobacter crescentus CB15 (mKate2: GD2 and Venus: GD3) and C. crescentus NA1000 (mKate2: AKS295) strain variants that contained chromosomally incorporated phenotypic markers: fluorescent proteins mKate2 or Venus under a constitutive pLac promoter for most experiments [21 ]. Strains were cultured in Peptone Yeast Extract Broth [22 (link)] (PYE-B) and grown for 30 h at 30 °C. Cells from these cultures were used for growth experiments in M2 minimal medium [16 (link)] containing either xylan (Megazyme, Ireland) or xylose (Sigma Aldrich, Switzerland). Carbon sources were prepared using nanopure water and filter sterilized using 0.40 μm Surfactant-Free Cellulose Acetate filters (Corning, USA). Concentrations for batch experiments ranged from 0.01–0.1% (weight/volume) for batch experiments and 0.05% (weight/volume) for microfluidic experiments for both xylan and xylose. Well-mixed batch experiments in xylan or xylose media were performed in 96-well plates and growth dynamics was measured using a micro-well plate reader (Biotek, USA). See Supplementary Methods for detailed information about growth conditions and media recipes.

Induction of C. campestris haustoria was carried out as described (Bawin et al. 2022 (link)) but with a glass-fiber filter paper (Whatman GF/F, 90 mm) between Cuscuta and the Petri dish that was soaked beforehand with 4 ml of a 2 mg/ml suspension of either xylose (Tokyo Chemical Industry, CAS 58-86-6), mannose (Thermo Fischer Scientific, CAS 3458 to 3428-4), glucose (Sigma Aldrich, CAS 50-99-7), galactose (Tokyo Chemical Industry, CAS 59-23-4), xylan (Megazyme, CAS 9014-63-5), β-1,4-mannan (Megazyme, CAS 9036-88-8), glucomannan (Megazyme, CAS 11078-31-2), or galactomannan (Megazyme, CAS 11078-30-1) in water and subsequently dried. Ten visually similar sites from three or more individual stems were excised and pooled before being frozen in liquid nitrogen. Five biological replicates were harvested for each modality and tested by RT-qPCR for the expression of three endo-β-1,4-mannanase genes (Cc017717, Cc044101, Cc044103) following Bawin et al. (2022) (link) with Cc028808/Cc006757 as references. For each gene, expression values were compared between treatments using a generalized linear model and Tukey multiple pairwise comparisons with BH correction (P ≤ 0.05).