*2.1. Cyanobacteria Origin and Growth Conditions*

Two filamentous non-heterocystous cyanobacteria, previously reported to be present in the arid soil cyanobacterial communities [30,37], were selected to evaluate the effect of water deprivation stress on their EPS secretion: the sheathed *Phormidium ambiguum* Gomont NIES-2121, purchased at NIES Collection, Japan, and originally isolated from an African soil [26], and *Leptolyngbya ohadii* provided by the Department of Plant and Environmental Sciences of the Hebrew University of Jerusalem, Israel, originally isolated in the Negev desert [30]. The latter cyanobacterium produces an EPS fraction, not structured similar to a sheath, that remains attached to the filament when grown in liquid culture hereafter referred to as glycocalyx EPS (G-EPS) [10]. Both cyanobacteria, maintained in the laboratory of the Department of Agriculture, Food, Environment, and Forestry (DAGRI) of the University of Florence (Italy), were grown in flasks containing liquid BG-11 medium [38], at 25 ◦C, under continuous illumination of 15 μmol photons m−<sup>2</sup> s−<sup>1</sup> and continuous stirring at 100 rpm until the stationary phase. These cultures were used for biomass inoculation described below.

#### *2.2. Inoculation and Biocrust Sampling*

In these experiments, the cyanobacterial strains were inoculated (nine replicates for each cyanobacterium) in microcosms of Petri dishes, 150 mm (diameter) × 20 mm (depth), containing 300 g of an autoclaved commercial dried silica sand (VAGA s.r.l., Pavia, Italy), with a granulometry of 0.3–0.6 mm. Before inoculation the cultures were centrifuged at different speeds to optimize their sedimentation. *P. ambiguum* was centrifuged at 7000× *g* and *L. ohadii* was centrifuged at 4000× *g*, for 20 min, at room temperature and the supernatant was discarded. To remove any remaining culture medium, the pellet filaments were resuspended in sterile distilled water, centrifuged again at the same speeds, and the supernatant discarded. The pellet was resuspended in sterile distilled water in a volume to provide enough inoculum to disperse the resuspended filaments spirally on the microcosms [10]. Each microcosm was inoculated with 160 mg (dry weight) of cyanobacterial biomass. The chlorophyll *a* amount inoculated in each microcosm corresponded to 2.54 μg/g sand and 2.9 μg/g sand for *P. ambiguum* and *L. ohadii*, respectively. Inoculated microcosms were maintained inside a Plexiglass incubator with controlled temperature at 25 ◦C and continuous light intensity of 30 μmol photons m−<sup>2</sup> s−<sup>1</sup> to stimulate growth. The microcosms were daily watered with 35 mL of sterilized distilled water (enough volume to wet the entire surface of the biocrust).

After 4 weeks, all microcosms formed a visible crust on the sand substrate and three microcosms (*N* = 3) of each cyanobacterium were randomly collected. The remaining six microcosms were kept in incubation for 5 more days under two different conditions: three (*N* = 3) were subjected to water deprivation stress by stopping the watering and leading to the drying of the biocrusts, while three (*N* = 3) the watering continued and worked as controls (Figure S1). After this additional period (totaling 33 days) the biocrusts in all six microcosms were collected. All the formed crust was collected separated from the sand substrate and gently homogenized with a sterilized spatula. After homogenization, the crusts were weighted and used to determine chlorophyll *a* content, EPS amount, monosaccharidic composition, and MW distribution as described below.

#### *2.3. Biofilm Formation in Liquid Culture Medium*

Cultures of each cyanobacterium were inoculated in 1 L Pyrex Erlenmeyer flasks containing 500 mL of BG-11 culture medium [38] (nine replicates for each cyanobacterium). After inoculation, each replicate had an initial dry weight of 0.1 mg mL<sup>−</sup>1. The chlorophyll *a* amount inoculated in each flask corresponded to 9.65 μg mL−<sup>1</sup> and 6.82 μg mL−<sup>1</sup> for *P. ambiguum* and *L. ohadii*, respectively. The flasks were maintained, without agitation, inside a Plexiglass incubator with controlled temperature at 25 ◦C and continuous light intensity of 30 μmol photons m−<sup>2</sup> s−<sup>1</sup> to stimulate growth. After 4 weeks, 3 biofilms trials

(*N* = 3) of each cyanobacterium were randomly collected. The remaining 6 biofilms trials were kept in incubation for 5 more days, under two different conditions: 3 (*N* = 3) were subjected to water deprivation stress by removing the liquid medium and maintaining the biofilm attached to the glass bottom of the Erlenmeyer flasks, while 3 (*N* = 3) were kept with the liquid medium (Figure S1). After this additional period (totaling 33 days) the 6 biofilms were collected. Samples were used to determine chlorophyll *a* content, EPS amount, monosaccharidic composition, and MW distribution as described below.

#### *2.4. Biocrusts and Biofilm Characterization*

#### 2.4.1. Growth Measurements through Chlorophyll *a*

For biocrusts, chlorophyll *a* content was determined according to the method as reported in Castle et al. (2011) [39]. All biofilms, after resuspension of the water deprived samples in saline solution (150 mM NaCl in sterile distilled water), were homogenized with a sterilized glass rod and collected with serological pipette. The extraction was performed according to the method as reported in Ritchie (2006) and Yéprémian et al. 2016 [40,41]. Briefly, 1 g of homogenized crust and 2 mL of biofilm homogenate were collected into screw-cap vials. The biofilms were centrifuged at 3800× *g* for 10 min, and the supernatant discarded. Next, the weighted crust and the pelletized biofilm were treated with 5 mL of ethanol, at 80 ◦C, for 5 min. Samples were incubated in the dark, at 4 ◦C for 30 min before being centrifuged at 3800× *g* for 15 min. The supernatant was recovered and determined by measuring the absorbance (A) at 665 nm. Chlorophyll *a* content was calculated according to previous study [40]:

$$\text{Chlorophyll } a \text{ [}\mu\text{g/g crust]} = \text{(11.9035 } \times \text{A}\_{665} \times \text{Ve)} \times \text{(g } crsst^{-1}\text{)} \times \text{L} \tag{1}$$

$$\text{Chlorophyll } a \text{ [}\mu\text{g/mL]} = \text{(11.9035 } \times \text{A}\_{665} \times \text{Ve)} \times \text{(Vs}^{-1}\text{)} \times \text{L} \tag{2}$$

where Ve is the volume of ethanol (mL), Vs is the volume of sample (mL), and L is the path length (cm).

### 2.4.2. EPS Isolation, Quantification and Characterization

The EPSs were extracted from the biocrusts in two different fractions according to previous works [17,19,42]. The one easily released into the sand substrate and more water soluble, referred to as loosely bound EPS (LB-EPS), and the one more condensed, firmly attached to cells and sand particles, referred to as tightly bound EPS (TB-EPS). LB-EPS were recovered by resuspending the biocrusts in distilled water and incubating them at room temperature for 15 min. Next, samples were centrifuged at 3800× *g* at 8 ◦C for 30 min and the LB-EPS-containing supernatants collected. This extraction was repeated three times for each sample. TB-EPS were recovered by treating the biocrust pellet resulting from the LB-EPS extraction with 0.1 M Na2EDTA for 16 h at room temperature. Next, samples were centrifuged at 3800× *g* at 8 ◦C for 30 min and the TB-EPS-containing supernatants collected. This extraction was repeated three times for each sample, the last two extractions performed for 120 min each.

Regarding the biofilms, EPS were extracted from the glass-rod homogenized biofilms cultures and centrifuged at 3800× *g* for 30 min at room temperature. The RPS were isolated from the supernatant while the pelleted cells were saved for sheath or G-EPS extraction. For that, the supernatants were concentrated by evaporation using an orbital evaporator at 35 ◦C. Next, the RPS were precipitated by mixing with two volumes of cold (4 ◦C) isopropyl alcohol and incubated at 4 ◦C for 8 h. After centrifuging at 4000× *g* for 15 min, the EPS pellets were resuspended in distilled water. For sheath or G-EPS extraction, the pelleted cyanobacterial cells obtained after the centrifugation of the culture biofilms were washed with 5 mL of 1.5% NaCl solution. After removing the 1.5% NaCl solution, the pellets were resuspended in 5 mL of sterile distilled water at 80 ◦C for 1 h [19]. After centrifuging at

4000× *g* for 30 min, the sheath-containing supernatants for *P. ambiguum* and the G-EPS for *L. ohadii* were collected.

The RPS and TB-EPS extracts were confined in dialysis membranes (12–14 kDa MW cut off, Medicell International London) and dialyzed against distilled water for 24 h, with two changes of water. All the EPS extracts were quantified by the phenol-sulphuric acid assay method [43]. For the determination of the monosaccharidic composition, the samples were hydrolyzed in 2 N trifluoroacetic acid (TFA), for 120 min at 120 ◦C. The tubes containing the samples were then dipped in cool water and the samples were evaporated in an orbital evaporator at 35 ◦C. Subsequently, samples were suspended in HPLC-grade water and evaporated again, repeating this step for one more time. Finally, monosaccharide composition was analyzed with a Dionex ICS-2500 ion exchange chromatograph (Dionex, United States) equipped with an ED50 pulsed amperometric detector operating with a gold working electrode (Dionex) and a CarboPac PA1 column of 250 mm length and 4.6 mm internal diameter (Dionex, Sunnyvale, CA, USA). Eluents were HPLC-grade water (A), 0.185 M NaOH (B), and 0.488 M sodium acetate (C), at a flow rate of 1 mL min−1. Single sugars were identified and quantified based on the retention time of reference standards. Results were expressed in molar ratio.

Molecular size distribution of the EPSs was analyzed using a Varian ProStar HPLC chromatograph (Varian, CA, USA) equipped with a refractive index detector and two columns for Size Exclusion Chromatography (Polysep-GFC-P6000 and 4000, Phenomenex, CA, USA) connected in series. Samples were analyzed with runs of 70 min and with HPLC-grade water as eluent, at a flow rate of 0.4 mL min−1. Dextran (Sigma-Aldrich, Burlington, MA, USA) at different MWs (2, 1.1, 0.41, 0.15, and 0.05 M Da) were used as reference standards. To obtain the % of the different MW classes the ratio between each peak area and the total area under the curve was calculated and the resulting % area was assigned to the corresponding size class according to the retention time of the peak output.

#### *2.5. Data Analysis*

Possible differences in chlorophyll *a* and EPSs amount among the two cyanobacterial strains grown in biocrusts and biofilm-cultures were analyzed using one-way analysis of the variance (ANOVA) at 95% of the significance, followed by Tukey post hoc test. Variables were previously checked for normality and homogeneity of variance using the Shapiro–Wilk and Levene's test, respectively. To correlate parameters, linear regression analyses were performed, and *r*<sup>2</sup> and *p* values are reported. For statistical analysis of the monosaccharidic profiles, Student's *t* test was used to compare the relative amount of each monosaccharide in the control and water deprivation replicates. Furthermore, the number of monosaccharides, diversity, dominance, and equitability indices of sugar residues of the different strains and conditions (sheath, G-EPS, and RPS in liquid cultures, and LB-EPS and TB-EPS in biocrusts) were compared. Shannon diversity index was calculated using the percentiles of a bootstrap distribution with 9999 repetitions. All statistical analysis was performed using Past 4.09 software.

#### **3. Results**

#### *3.1. Effect of Water Deprivation on Chlorophyll a and EPS Production in Sandy Microcosms*

The growth of the cyanobacterial strains was evaluated by measuring the chlorophyll *a* content. Chlorophyll *a* content decreased significantly (*p* < 0.05) during the water deprivation period for both cyanobacteria (Figure 1). The content of chlorophyll *a* decreased in a negative correlation with the EPSs contents (Figure 2), presenting an r2 of 0.87, 0.84, 0.89, and 0.77, respectively; *p* < 0.05, for *P. ambiguum* LB-EPS, *P. ambiguum* TB-EPS, *L. ohadii* LB-EPS, and *L. ohadii* TB-EPS, respectively. The significantly higher amount of chlorophyll a obtained for *P. ambiguum* control (i.e., 4 weeks watered plus 5 days watered), represented twice the amount of chlorophyll *a* produced by *L. ohadii* control (Figure 1).

**Figure 1.** Chlorophyll *a* content in cyanobacterial biocrusts (values represent the mean of *N* = 3, error bars represent SD). Different letters represent significant differences (*p* < 0.05).

**Figure 2.** EPSs contents in *P. ambiguum* (**A**) and *L. ohadii* (**B**) biocrusts (values represent the mean of *N* = 3, error bars represent SD). Different letters represent significant differences (*p* < 0.05) in each graph.

The amount of LB-EPS, the loosely bound polymers, that are weakly attached to cells and sediments, and the TB-EPS, the tightly bound polymers, which have stronger bounds to cells and sediments, changed according to the cyanobacterial strain and growth condition (Figures 2 and S2). For *P. ambiguum*, the only significant difference observed was the increase in the TB-EPS fraction in water deprivation compared to the TB-EPS from the control and all the LB-EPS fractions. On the other hand, for *L. ohadii,* in LB-EPS from water deprived condition was significantly higher compared to those of the control. For this strain, the TB-EPS amount was significantly higher than the LB-EPS amount in each condition. The TB-EPS after water deprivation resulted significantly the highest value among all the EPSs extracted for both cyanobacteria (Figure 2).

When comparing the amount of LB-EPS in each condition for both cyanobacteria, those from *P. ambiguum* in control showed significantly higher value than those of *L. ohadii* in the same condition (Figure S3). After the water deprivation period, the LB-EPSs of these strains were not statistically different. On the contrary, while *P. ambiguum* TB-EPS control

was not statistically different from *L. ohadii* TB-EPS control, after the water deprivation period, *P. ambiguum* TB-EPS showed significantly higher value than *L. ohadii* TB-EPS.

#### *3.2. Changes in Monosaccharidic Composition and Molecular Weight Distribution of Microcosms EPSs*

The EPSs were analyzed in terms of monosaccharide composition, showing differences in its relative abundances between the control and water deprived microcosms. The EPS fractions of *P. ambiguum* were mainly composed of glucose (Figures 3 and S2). The LB-EPS after water deprivation had the highest relative amount of this sugar among the EPSs extracted from this cyanobacterium biocrust. On the other hand, the LB-EPS in the control microcosms had a higher diversity and was also composed by significantly higher molar ratios of uronic acids (Figure 3A,C and Table S1). Moreover, fucose and rhamnose were detected in LB-EPS control and water deprived. The TB-EPS of water deprived *P. ambiguum* microcosms showed lower diversity than the control TB-EPS of this cyanobacterium (Table S1). When comparing the LB-EPS and TB-EPS only in water deprivation conditions, the first fraction showed lower diversity than the latter (Table S1). Galactose was found only after water deprivation in LB-EPS and in both conditions in TB-EPS (Figure 3A,C). Additionally, fucose was detected in the control and water deprived TB-EPS.

**Figure 3.** Monosaccharidic composition of the EPS extracted from biocrusts: (**A**) LB-EPS from *P. ambiguum,* (**B**) LB-EPS from *L. ohadii*, (**C**) TB-EPS from *P. ambiguum*, (**D**) TB-EPS from *L. ohadii*. Molar percentages (%) of single sugars are represented (expressed as moles of the single monosaccharide divided by the total amount of moles of monosaccharides in the EPS × 100). Symbol \*, when present, indicates significant differences between the control and watered-deprived period in each monosaccharide. Fuc, fucose; Rha, rhamnose; GalN, galactosamine; Ara, arabinose; GlcN, glucosamine; Gal, galactose; Glc, glucose; Man, mannose; Xyl, xylose; Fru, fructose; Rib, ribose; GalA, galacturonic acid; GlcA, glucuronic acid.

After 4 weeks watered, *L. ohadii* LB-EPS and TB-EPS were mainly composed of glucose (Figure S2). The control LB-EPS and TB-EPS were mainly composed of glucose, galactose,

and uronic acids (Figure 3B,D). The control TB-EPS showed higher diversity than the TB-EPS extracted after water deprivation (Table S1). For this cyanobacterium, the TB-EPS of water-deprived microcosms showed the highest molar percentage of glucose, followed by galactose. This EPS fraction after water deprivation condition had higher dominance than the control TB-EPS, the same profile observed for the LB-EPS. When comparing the LB-EPS and TB-EPS only in water deprivation conditions, the first fraction showed higher diversity than the second fraction. The TB-EPS demonstrated higher dominance compared with LB-EPS, the same profile shown by *P. ambiguum* water-deprived biocrusts. Moreover, fucose and rhamnose were present in both control and water-deprived EPS fractions.

All the EPS extracted were analyzed in terms of molecular weight (MW) distribution. The results revealed the presence of molecules ranging from 50 kDa to 2 MDa. The *P. ambiguum* LB-EPS and TB-EPS were composed predominantly of MW molecules of 410 kDa, except LB-EPS in water deprivation stress that also showed molecules higher than 2 MDa. The *L. ohadii* LB-EPS after the watered periods was predominantly composed by molecular weight molecules of 1.1 MDa–410 kDa and after a water deprivation of 1.1 MDa. The TB-EPSs extracted from this cyanobacterium were mainly constituted of molecules with MW ranging between 2 MDa–1.1 MDa (80% of chromatogram area) and a small fraction of molecules with MW ranging between 410 kDa–150 kDa (15% of chromatogram area) for all the conditions tested (data not shown).
