Introduction

The extracellular polysaccharidic (EPS) matrix of biological soil crusts (BSCs) plays an important role in the adaptation and survival of the microbial communities living in these complex assemblages under environmental conditions that, in many cases, are very harsh (Rossi et al. 2012a; Colica et al. 2014; Chen et al. 2014). Indeed, EPS matrix of BSCs provides soil stabilization against erosion, enhances moisture content, and allows positive interactions (e.g., nutrient availability) with the surrounding environment in geographical areas characterized by drought, high solar irradiation, and high or very low temperatures (Pointing and Belnap 2012). In oligotrophic soils, such as the sandy soil of the Mongolian desert, the capacity of the colonizing microorganisms to synthesize EPSs is key to foster the aggregation of soil particles and to promote the development of BSCs. The onset of these complex microbial aggregates, typically cyanobacteria-dominated, increases biological activity and enhances the turnover of substances, primarily carbon and nitrogen, in and out of the soil (Pointing and Belnap 2012).

In a well-developed BSC, the EPS-based matrix is usually synthesized by pioneer organisms, such as filamentous cyanobacteria, or by later colonizing microalgae and microfungi (Belnap 2001), while it is contemporarily decomposed and consumed by the microcosm that feeds on it (Yu et al. 2012; Chen et al. 2014). EPSs represent a notable C-source, which is readily available to the heterotrophic members of the community (Mager and Thomas 2011). As the result of complex biotic interactions within the crust communities, EPS chemical and macromolecular characteristics vary according to the respective combined activities of producers and consumers and according to the contributing species and their relative abundances.

The characteristics of the EPS-based matrix depend also on a wide array of abiotic factors including soil type, environmental conditions, and nutrient levels, which reportedly influence EPS synthesis by cyanobacteria (Hu et al. 2002; Nicolaus et al. 1999). In general, microbial interactions are heavily influenced by environmental stress gradients (“stress gradient hypothesis” (SGH)), which determine the relative frequency of facilitation and competition between the species. According to SGH, environmental constraints produce changes in the community, which are instrumental to fulfill the ecological functions of the BSCs (Li et al. 2013).

The excretion of EPSs and the chemical characteristics of these macromolecules can deeply influence the adaptation capability of the microbial community residing in the BSCs to environmental stresses. EPSs have varying degrees of complexity, including fractions with a hydrophobic character and fractions with a hydrophilic character. The fractions with a hydrophobic character, which owes mostly to the presence of deoxysugars, boost the capacity to adhere to solid substrates, while fractions with a hydrophilic character promote ionic interactions. The immobilization of minerals and nutrients from the surrounding environment due to the anionic nature of EPSs enhances the availability of these compounds for the microflora residing in the BSCs (Rossi et al. 2012a). In addition, the hygroscopic nature of EPSs is very important to confer tolerance to water stress. Indeed, in soil systems, the secretion of EPSs was demonstrated to enhance water availability at the topsoil (Colica et al. 2014). The EPS-based matrix of BSCs plays also an important structural role in enhancing soil stability against erosion, creating the conditions for an optimal water diffusion through the crust profile (Rossi et al. 2012b).

The present study was aimed at investigating the monosaccharidic composition and molecular weight (MW) distribution of EPSs of BSCs collected in the hyper-arid plateau of Hobq Desert (Inner Mongolia, China) and the possible correlations between EPS characteristics and the age of BSCs. The experimental area was chosen as it displays induced BSCs (henceforth IBSCs) as the result of inoculation-based procedures carried out by employing cyanobacteria (Wang et al. 2009). Since the inoculation was carried out in different sites in a period spanning from 2002 to 2007, IBSCs in the sampling sites are of different ages. The inoculation sites, even if constituted by the same kind of sandy soils, were characterized by different gradients of radiation, slope of sand dunes, moisture, and wind erosion conditions, which affected the growth speed of the crustal communities in the years (Colica et al. 2014).

Material and methods

Experimental area and inoculation procedure

The characteristics of the experimental area have been reported by Colica et al. (2014) and Chen et al. (2014). Hobq Desert is a hyper-arid plateau at an altitude of 1040 m above sea level, located in Dalatequi County, Inner Mongolia, China, characterized by moving sandy dunes. The climate belongs to a typical continental monsoon pattern, with minimal average temperature of −34.5 °C and maximum temperature of 40.2 °C. Mean annual evaporation (2448 mm) far exceeds mean annual precipitation (126 mm). From 2002 to 2007, large volumes of a mixed cyanobacterial culture constituted by Microcoleus vaginatus (Gom.) and Scytonema javanicum (Born et Flah) were spray-inoculated with a 10:1 w/w ratio onto the dune surfaces in four sites, one site per year (2002, 2004, 2005, and 2007), in order to trigger the formation of IBSCs. At the time of inoculation, the soil in the four experimental areas consisted of unconsolidated sand, almost devoid of any plant coverage.

The employed technology has been reported in details by Chen et al. (2006) and Wang et al. (2009). The two strains had been previously isolated from natural BSCs growing in the experimental area and then mass cultivated in raceway ponds in a greenhouse, up to volumes of 10,000 L, before being mixed and spray-inoculated onto the sites (~35 μg Chl a cm−2 of soil). After the inoculation, the sand was intermittently irrigated for 15–18 days with 20 mm of groundwater per day by automatic micro-irrigation sprinkler facilities, until cyanobacterial crusts appeared at the topsoil.

Sampling site description and sampling procedure

The inoculated sites were approximately rectangular, with a road flanking one of the two long sides of each rectangle, from which cyanobacterial mixed cultures were sprayed onto the sites. The first three sites (henceforth named 02, 04, 05), which are located at the northern edge of the desert in a fenced remediation field, measured 300 × 100 m, while the fourth site (henceforth named 07) measured 2000 × 100 m. While sites 02 (40.3662° N, 109.8449° E, inoculated in 2002) and 04 (40.3671° N, 109.8448° E, inoculated in 2004) are close to each other, site 05 (40.3664° N, 109.8501° E, inoculated in 2005) is 400 m apart from the two. Site 07 (40.3923° N, 109.8154° E, inoculated in 2007) is located about 6 km away from sites 02, 04, and 05, deeper into the Hobq desert.

Sample collection was performed in summer of 2010, according to the method described by Colica et al. (2014). It was considered that, due to the inoculation procedure and due to the fact that the height of the dunes increased moving perpendicularly from the road, the side of the sites close to the road received higher amounts of inocula and water. Each site was thus sampled in three different points along a hypothetical transect moving perpendicular from the longest sides of each sampling sites to the opposite longest side, starting from the road. The first sampling point was always at a distance of 10 m from the road (indicated with suffix -1 near the number indicating the sample site), the second at a distance of 50 m (indicated with suffix -2), and the last one at a distance of 90 m (indicated with suffix -3). At least three sub-samples were collected in each sampling point to perform all the experiments (experimental replicates, n). Control samples were collected in non-inoculated sites outside the fenced remediation field where sites 02, 04, and 05 are located and outside site 07. Control sites were characterized by unconsolidated sandy soil with the presence of sparse sub-shrub vegetation. Occasionally, natural BSCs were present. Both natural BSCs and unconsolidated sand were used as controls. Samples were picked up with a thickness ranging from 5 to 10 mm, depending on the structure of the crusts, stored in sterile Petri dishes, transported to the lab, and dried out before analysis.

Pigment extraction and quantification

IBSC samples were fragmented, homogenized, weighed, and extracted at 4 °C overnight with absolute acetone. Extracts were clarified by centrifugation at 3600×g for 20 min and the supernatants analyzed by a Cary 50 Spectrophotometer (Varian, USA). Absorbance contributed by chlorophyll a was measured at 663 nm. Absorbance at 750 nm was subtracted from all the measured absorbances to account for residual scattering. Corrected absorbance was calculated using the trichromatic equations of Garcia-Pichel and Castenholz (1991), and an extinction coefficient of 92.6 L g−1 cm−1 for chlorophyll a in 100 % acetone was used (Vernon 1960).

EPS extraction and quantification

In order to enhance the recovery of both tightly bound and loosely bound carbohydrates, IBSCs were treated with Na2EDTA solution, which chelates metal cations (especially Ca and Mg) that otherwise favor the cohesion between the polymeric strands of EPSs. IBSCs were disassembled, homogenized in a mortar, and weighed. For each extraction, about 100 mg of crust was treated with 20 mL Na2EDTA 0.1 M for 15 min at 25 °C, according to the method of Underwood et al. (1995) for the extraction and quantification of extracellular carbohydrates from soil, before being centrifuged for 20 min at 3500×g to recover the extracts devoid of any soil particle in suspension. EPS-containing supernatants were stored at 4 °C before analysis. The above-described extraction procedure was repeated four times on each sample in order to make the treatment more effective. Total extracellular carbohydrate content, a valid index to gauge EPS content (Underwood et al. 1995), was determined in the extracts using phenol-sulfuric acid assay, according to Dubois et al. (1956).

Determination of the monosaccharide composition

In order to remove the excess of Na2EDTA, EPS-containing supernatants were confined in nitrocellulose tubular membranes (14.000 Da MW cutoff, Medicell International, UK) and dialyzed against Milli-Q-grade water for 2 days, with three changes of water, before being dried. Dried EPSs were hydrolyzed with 2 N trifluoroacetic acid (TFA) for 45 min at 120 °C. Afterward, they were ultra-filtered using centrifugal filters (Amicon Ultra-4) at 8200×g for 20 min, dried on a rotary evaporator, resuspended in deionized water, and analyzed using a Dionex ICS-2500 ion exchange chromatograph (Dionex, USA) 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). Eluents used were Milli-Q-grade water (A), 0.185 M sodium hydroxide (B), and 0.488 M sodium acetate (C). In the first stage of the analysis (from injection time to 7 min), the eluent consisted of 84 % A, 15 % B, and 1 % C; in the second stage (from 7 to 15 min), the eluent consisted of 50 % B and 50 % C; in the final stage (from 15 to 30 min), the eluent was that of the first stage. The flow rate was kept at 1 mL min−1. Peaks for each sugar were identified on the basis of the retention time of known standards. The utilized analytical method is in accordance with Xu et al. (2013).

Size fraction analysis

Dried extracts were dissolved in deionized water, transferred to Eppendorf tubes, and clarified by ultracentrifugation at 13,000×g in order to remove the coarse particulate. The extracts were subsequently analyzed using a Varian ProStar HPLC chromatograph (Varian, USA), equipped with a refractive index (RI) detector and two columns for size exclusion chromatography (SEC), Polysep-GFC-P 6000 and 4000 (Phenomenex, USA) connected in series. The columns (700-mm length and 7.8-mm internal diameter) have separation ranges of 100 kDa to 15 MDa and 0.3 to 400 kDa, respectively. Samples were analyzed with runs of 65 min using Milli-Q-grade water as eluent at a flow rate of 0.4 mL min−1. Dextran (Sigma-Aldrich, USA) at different MWs (2000, 485, and 76 kDa) and saccharose (MW = 0.34 kDa) were used as standards.

Statistical analysis

Every measurement was performed in triplicate (experimental replicates, n). Data were processed with one-way analysis of the variance (ANOVA) at the 95 % significance level and linear regressions to validate the significance and correlations of the data. In all cases, R 2 and P values are reported in the text and/or in the figures. Statistical analyses were performed using GraphPad Prism version 5.00 for Windows (GraphPad Software, USA).

The data deriving from monosaccharide composition and MW distribution, organized in matrices, were processed using principal component analysis (PCA), which is a multivariate analytical tool to reduce the dimensionality of data, without the loss of information, to a set of linear combinations (principal components (PCs)) (Rencher 1995). The coefficients of linear combinations are called loadings, and the PCs account for the variance of the original data. PC1 is the first PC and accounts for the greatest amount of the total variance, while PC2 and PC3 account, respectively, for the second and the third largest variances. Data were plotted to observe possible differences between the variables. For PCA analysis, software Primer-6 (Primer-E Ltd, UK) was used.

Results

Chlorophyll a and total carbohydrate content of IBSCs

The overall distribution of phototrophic abundance in the IBSCs sampled in the investigated sites was determined as chlorophyll a content and was patchy and uneven even within the same sampling sites (Table 1). The highest chlorophyll a abundances were observed for samples 02-3 and 04-1 (8.35 and 9.13 μg g−1 soil, respectively), with no statistically significant difference between the two (P = 0.17, P > 0.05). Considering the whole set of data, no statistically significant correlation between the phototrophic abundance and the age of the crusts was found (P = 0.40, P > 0.05).

Table 1 Phototrophic abundance (expressed as chlorophyll a content in the soil, μg g−1) and total extracellular carbohydrate content (expressed as carbohydrate content in the soil, μg g−1) of the IBSC samples
Full size table

In general, phototrophic abundance was at least two times higher in IBSCs, compared to control samples. Controls were made up of natural BSCs and bare sand sampled in an area devoid of BSCs. Controls made up of natural BSCs showed chlorophyll a contents ranging between 0.16 and 1.65 μg g−1 soil, while in bare sand, average chlorophyll a contents were extremely low, ranging from a detected minimum of 0.01 μg g−1 soil to a maximum of 0.11 μg g−1 soil, while in one control sample (04-Cs), the level was too low to be detected.

In the case of the total extracellular carbohydrate content (i.e., high and low MW carbohydrate content) of IBSCs, values ranging from 0.03 to 1.99 mg carbohydrate g−1 of soil (sample 04-3) were found (Table 1). Samples collected in sites 05 and 07 showed lower amounts of carbohydrates, ranging from 0.03 to 0.33 mg g−1 of soil. Bare sand showed total carbohydrate contents ranging from 0.03 to 0.15 mg g−1 of soil while natural BSCs showed contents ranging between 0.11 and 0.81 mg g−1 of soil.

Only considering sites 02 and 04, which are characterized by the presence of 8- and 6-year-old IBSCs, chlorophyll a content correlated well with the total extracellular carbohydrate content, showing a linear distribution (R 2 = 0.83, P = 0.04; data not shown). However, considering the whole set of data, no statistically significant correlation was found.

Monosaccharide composition of EPSs extracted from IBSCs and soil samples

The EPSs extracted from the IBSCs under study showed a high complexity in terms of number of different types of monosaccharides. Up to 13 different types of monosaccharides were identified and quantified (Table 2). The hexoses galactose and glucose were the overall dominant sugars. A hydrophilic fraction (represented by galacturonic and glucuronic acids) and a hydrophobic fraction (represented by the deoxysugars, rhamnose and fucose) were always detected. Occasionally, the amino sugars galactosamine and glucosamine were also present, although the former sugar type was detected only in traces. Fructose and ribose were present in traces in some of the samples, although not in the samples collected in site 07.

Table 2 Relative abundances of the single monosaccharides (calculated as moles of the single monosaccharide divided by the total amount of moles of monosaccharides in the EPS × 100) in the EPSs extracted from IBSCs in the four sampling sites
Full size table

The relative abundances of glucose, galactose, and mannose were cumulatively higher in IBSCs collected in site 07 (R 2 = 0.78, P < 0.01), compared to the other three sites.

The overall highest contents in uronic acids (i.e., glucuronic acid + galacturonic acid) were found in IBSCs collected in sites 02 and 04. Statistically, the relative abundance of uronic acids increased with the age of the crusts (R 2 = 0.75, P < 0.05), assuming a linear distribution (Fig. 1). In sites 02 and 04, the relative abundances of uronic acids ranged from 13.28 mol% (sample 04-3) to 20.22 mol% (sample 02-1). In site 05, uronic acid contents ranged from 14.68 to 16.17 mol%, while in site 07, the contents were significantly lower, ranging from 2.97 to 5.95 mol%.

Fig. 1
figure 1

Uronic acids (moles of glucuronic + galacturonic acid per total moles of monosaccharides) in the EPS matrix of the IBSC samples collected in the four sampling sites. 02, samples collected in 2002 site (8-year-old IBSCs); 04, samples collected in 2004 site (6-year-old IBSCs); 05, samples collected in 2005 site (5-year-old IBSCs); 07, samples collected in 2007 site (3-year-old IBSCs). Values are means of three different measurements and SD bars

Full size image

Control samples constituted by bare sand showed a carbohydrate content which was too low to be analyzed by IEC; thus, only control samples made up of natural BSCs were analyzed for the monosaccharide composition of their EPS matrices. EPSs extracted from natural BSCs showed an uronic acid content ranging from 9 to 11 mol%, while galactose and glucose were the first and the second most abundant monosaccharides (ranging between 25.04 and 31.06 mol% and between 25.9 and 29.7 mol%, respectively). No statistically significant differences were found in the composition of the EPSs extracted from different control BSCs.

PCA applied to the set of data of the monosaccharidic composition of the EPSs showed three PCs that cumulatively explained 91.9 % of the variance of the data (Table 3). Galactose, glucose, mannose, and uronic acids showed high loadings on all the components. The first PC (explaining 41.7 % of the variance) showed galactose and glucose, respectively, with positive and negative effects. Mannose, galacturonic, and glucuronic acids were with negative effects. In the second PC (explaining 34.9 % of the variance), galacturonic and glucuronic acids had positive effects, while galactose and glucose showed negative effects. The third PC (explaining 15.3 % of the variance) showed galactose, mannose, and galacturonic acid with negative effects, while glucuronic acid had positive effects.

Table 3 Coefficients of linear combination (loadings) of the first three linear components, according to the monosaccharide composition profiles
Full size table

Scatter plots for PC1 × PC2 and PC1 × PC3 are shown in Figs. 2 and 3. In the first plot, a clustering of the EPSs extracted from samples collected in site 07 is visible (Fig. 2). This cluster separated from the other samples. A dense cluster in the center of the plot collected EPSs from samples taken in sites 04 and 05 and from natural BSCs (C, in the figure). EPSs extracted from samples collected in site 02 did not show any clustering. EPSs from samples 02-2 and 02-3 located close on the plot and were characterized by high amounts of galacturonic acid.

Fig. 2
figure 2

Score plot of PC1 vs PC2 showing the distribution of the samples according to the monosaccharidic composition of the EPS extracted from the IBSCs. 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.. 02, samples collected in 2002 site (8-year-old IBSCs); 04, samples collected in 2004 site (6-year-old IBSCs); 05, samples collected in 2005 site (5-year-old IBSCs); 07, samples collected in 2007 site (3-year-old IBSCs)

Full size image
Fig. 3
figure 3

Score plot of PC1 vs PC3 showing the distribution of the samples according to the monosaccharidic composition of the EPS extracted from the IBSCs. 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. 02, samples collected in 2002 site (8-year-old IBSCs); 04, samples collected in 2004 site (6-year-old IBSCs); 05, samples collected in 2005 site (5-year-old IBSCs); 07, samples collected in 2007 site (3-year-old IBSCs)

Full size image

The scatter plot PC1 × PC2 showed a marked separation between EPSs from samples 02-2 and 02-3, and from sample 02-1, the latter characterized by high amounts of glucuronic acid, primarily, and galactose, secondarily.

The cluster constituted by EPSs extracted from samples collected in site 07 separated significantly from EPSs of samples collected in site 02.

The PC1 × PC3 plot showed that, while sample 02-1 separated from samples 02-2 and 02-3 on the first PC, samples 02-2 and 02-3 separated along the third PC (Fig. 3). Sample 02-1 was characterized by higher relative contents in glucuronic acid and galactose, while sample 02-2 was characterized by lower relative amount of mannose, glucose, and galacturonic acid.

MW distribution analysis

The EPS matrices extracted from the IBSC samples were analyzed for their MW distribution by means of SEC, and the fractions obtained were classified in five groups, according to the MWs of the standards used (see “Material and methods” section). The groups were classified into the following: class A, having MW ≥2000 kDa; class B, having MW in the range 2000–485 kDa; class C, having MW in the range 485–72.6 kDa; class D, having MW between 72.6 and 0.34 kDa; and class E, having MW ≤0.34 kDa. The EPSs obtained from the matrices of the IBSCs showed a very high variability in terms of size distribution of their molecular fractions, even considering EPSs extracted from IBSCs of the same age (Fig. 4). For instance, the EPS fractions belonging to class B showed the widest relative variations within the 12 EPS samples analyzed, ranging from 14.2 % in the EPS of sample 05-3 to 95 % in the EPS of sample 07-2. On the contrary, size groups A (MW ≥2 106 Da) and E (MW ≤0.34 Da) showed variations in a narrower range.

Fig. 4
figure 4

Amount of the extracted EPSs belonging to the five reference size fractions. A, MW ≥2000 kDa; B, 2000 ≤ MW ≥ 485 kDa; C, 485 ≤ MW ≥ 72.6 kDa; D, 72.6 ≥ MW ≤ 0.34 kDa; E, MW ≤0.34 kDa. 02, samples collected in 2002 site (8-year-old IBSCs); 04, samples collected in 2004 site (6-year-old IBSCs); 05, samples collected in 2005 site (5-year-old IBSCs); 07, samples collected in 2007 site (3-year-old IBSCs). Values are mean of three different measurements (n = 3), and SD bars are reported in the figure

Full size image

Some differences were observed by comparing samples collected in the earlier-inoculated sites (i.e., sites 02 and 04) and those of the younger IBSCs (collected in sites 05 and 07). The majority of EPSs extracted from samples collected in sites 02 and 04 showed a much higher percentage of molecules in the fraction B (MW in the range 2000–485 kDa) in comparison with the EPSs extracted from IBSCs of sites 05 and 07, which, on the other side, showed a higher percentage of molecules in the fraction D (MW between 72.6 and 0.34 kDa). A striking exception is represented by sample 07-2, in which EPSs belonged almost entirely (~95 %) to class B. However, taking into account the whole set of data, any statistically significant correlation was found between the age of the crusts and the relative amount of the single fractions.

PCA applied to the set of data of the MW distribution showed that the first two PCs explained cumulatively 96.3 % of the variance (Table 4). Variables B and D showed high loadings on the first component, which explains 79.6 % of the variance, while on the second component, high loadings were shown by variables C and D.

Table 4 Coefficients of linear combination (loadings) of the first three linear components, according to the MW distribution profiles
Full size table

Discussion

The results reported in the present study showed that earlier-inoculated IBSCs (sites 02 and 04) have higher extracellular carbohydrate contents and higher phototrophic abundances compared to later-inoculated IBSCs (sites 05 and 07). However, no statistically significant correlation between the carbohydrate or the chlorophyll a contents of the IBSCs and their age was found. In any case, it was evident that the inoculation of the dunes induced the formation of stable and solid crusts, whose presence caused a significant increase in the amount of carbohydrates in the crusted soil in comparison with the non-inoculated soils. On the other hand, the lack of any sound correlation between the amount of carbohydrates and the age of the crusts points out the complexity of the process leading to the presence of the EPS in the IBSCs. Indeed, the presence of the polysaccharide matrix of IBSCs is the consequence of two contrasting processes: the synthesis of the polymer, at the beginning—mostly due to the biosynthetic activity of the phototrophs sprayed on the sand dunes—and the degradation of the polymeric carbohydrates and their utilization as C-source by the chemoheterotrophic microflora growing over time in the IBSCs. The complexity of this phenomenon and the influence of the microenvironmental conditions are also pointed out by the patchy distribution of phototrophic abundance, even in IBSCs sampled at a very close distance in the same sampling sites.

All the EPSs extracted from the IBSCs had a high complexity in terms of number of different types of monosaccharides. It is worth mentioning that the presence of a large number of qualitatively different monomers represents an advantage for the microflora residing in the IBSCs as it confers a high plasticity against environmental constraints. Indeed, as it was previously suggested by Mazor et al. (1996), EPSs represent an energy and C-source for the microflora embedded in the polymeric matrix of crusts and the variety of sugars in the exudates increases the array of possible recruitable microbial species. For example, it was shown that under acidic conditions, EPSs are hydrolyzed to galactose, glucose, and mannose, thus stimulating the activity of microorganisms capable of utilizing these carbohydrates (Mazor et al. 1996). It was also previously shown that carboxylic and amino acids are two preferentially utilized C-sources in BSCs (Yu et al. 2012), inducing the presence of specific consumers. In the present study, it was found that the relative abundances of glucose, mannose, and galactose vary even considering IBSCs of the same age, suggesting differences in utilization rates between microbial communities living in crusts at a very close distance one to another. According to PCA, the relative abundances of galactose, glucose, mannose, and uronic acids explain much of the variance of the data, thus confirming the relevant role of these sugars in the differentiation of the IBSCs. The scatter plot shows that EPSs extracted from IBSCs collected in site 07 clustered apart from EPSs extracted from all the other samples (especially from site 02) and from natural BSCs. PCA analysis also pointed out that, while EPSs extracted from samples collected in site 07 clustered together, EPSs extracted from samples collected in site 02 showed significant difference between one another and located apart on the plot. These observations are a further confirmation of the complexity of the process leading to the presence of EPSs in IBSCs and show that the processes leading to the development of IBSCs and to the secretion of the exopolysaccharide matrix are not simply related to the age of the crusts. It is more probable that the characteristics of EPSs depend on the characteristics of the inoculated sites and on eventual differences in stress gradients.

Considering the correlation between the relative abundances of the monosaccharides and the age of the investigated IBSCs, a positive linear correlation was only found for uronic acids (R 2 = 0.746, P < 0.05), while no significant correlation (P > 0.05) was found for the other sugars. It is worth stressing that the increased presence of uronic acids may play a very important role in establishing favorable living conditions to the microflora residing in the crusts. Indeed, uronic acids are hydrophilic molecules capable of conferring water-absorptive characteristics to EPSs and it was previously shown that the desiccation tolerance of Nostoc commune is related to the presence of these acidic sugars in the EPS synthesized by this cyanobacterium (Tamaru et al. 2005). Indeed, hydration is the most vital factor controlling photosynthetic production in BSCs (Lange 2001). In our study, the significantly higher relative amount of uronic acids in EPSs extracted from older IBSCs seems to be supportive of their progressive good adaptation to water stress. In addition, it points out once again the contribution of the phototrophic microbial fraction in the synthesis of the EPS matrix of IBSCs, being these acidic sugars present in most of the EPSs synthesized by cyanobacteria and microalgae (Rossi and De Philippis 2015). The increase in the uronic acids content of the EPS matrix may also contribute to establish the acidic conditions needed for the chemical hydrolysis of the polymeric sugars (Yu et al. 2012), in this way favoring the flourishing of the chemoheterotrophs residing in the crusts.

Considering the MW of the EPSs extracted from IBSCs, the polymers obtained from earlier-inoculated crusts (i.e., sites 02 and 04) were shown to be mostly composed of fractions belonging to the higher MW classes B and C, while later-inoculated IBSCs showed a significant amount of the fraction belonging to class D. According to PCA, size class B represents the variable with the highest loading on the first component (explaining much of the variance), while size class D showed high loadings on the first and the second components. It can therefore be inferred that the difference in MW distribution of EPSs of the samples investigated is mostly explained by higher or lower relative abundances of molecules in the two classes B and D, effectively differentiating the IBSCs according to their age.

It is well known that cyanobacteria and microalgae generally produce EPSs with high MW. Indeed, about 80 % of the so far analyzed cyanobacterial EPSs showed apparent MWs bigger than, or equal to, 1000 kDa (i.e., within class B in this paper), the remaining ones showing MWs ranging from 100 to 460 kDa (Pereira et al. 2009). However, since phototrophic abundance did not show any statistically significant correlation with the age of IBSCs, the higher relative abundance of high MW EPSs in oldest crusts cannot be immediately explained by an increased synthetic activity of cyanobacteria and microalgae. More likely, the differences in the MW of the EPS fractions are explainable by hypothesizing an increase in the consumption of simple sugars, in the range between 72 and 0.34 kDa, in the older IBSCs, which has the consequence of increasing the relative abundance of the other fractions. This hypothesis is supported by a recent study carried out in the same area (Chen et al. 2014), which showed that the activity of sugar-degrading enzymes (sucrase and dehydrogenase) is higher in older IBSCs, suggesting a consumption of low-MW carbohydrates by the microbial community. Moreover, the increased presence of microfungi with IBSC age (Colica et al. 2014) probably further contributed to increase the complexity of the phenomenon observed. Indeed, it is known that microfungi are key components of BSCs (Belnap 2001), and their onset, besides enhancing the stability of the crust structure due to the formation of the mycelium, could have contributed to EPS synthesis during the developmental process of the crusts. Uronic acids are constituents of the cell walls of fungi, being present as high as 25 % in sporangiophore walls and 12 % in the hyphal walls (Martens and Frankenberger 1990). Fungal-secreted EPSs have also been shown to include uronic acids between the components (Martinez et al. 1986). Consequently, the contribution of microfungi to the final characteristics of the exocellular matrix in older IBSCs cannot be excluded.

In conclusion, our results show that the characteristics of the EPS of the matrix of the investigated IBSCs cannot be explained only in relation with the age of the crusts and the activity of phototrophic microorganisms, but most properly, the biotic interactions ongoing between EPS producers (cyanobacteria, green microalgae, and microfungi as major) and consumers (e.g., microfungi, heterotrophic bacteria) have to be taken into account. Moreover, it has to be stressed that the presence of environmental stressors (i.e., soil moisture, crust coverage, and species diversity) seems to be capable of differentiating the developmental level of IBSCs more than the biotic factors (Li et al. 2013), with a consequent differentiation in the characteristics of the EPS matrix of the crusts hardly correlating with the age.