Diversity of cultivated aerobic poly-hydrolytic bacteria in saline alkaline soils

Sample characteristics

Surface soil samples (0–5 cm depth) were collected into sterile plastic Petri dishes at five locations in Central Asia, Egypt and California. Each individual sample comprised a composite of 4 subsamples randomly collected in a 3–5 m² area. Samples from Kenya and Tanzania were collected in sterile plastic bags (Whirl-Pak®; Nasco, Fort Atkinson, WI, USA) and vials using disposable sterile tongue depressors as described previously (Duckworth et al., 1996). The samples were kept at 4 °C before analysis. At most locations, the top soil layer was desiccated at the sampling time with a 20% maximum content of moisture. The selection of the samples was based on an immediate measurement of pH of a 1:5 water extract using a field pH-conductivity meter (model WTW 340i; WTW, Weilheim, Germany). Only those soils showing the pH of the water extract above 9.5 were selected for sampling. In total, more than 70 saline alkaline soil samples were obtained. Some of their characteristics are presented in Table 1. The content of total soluble salts was estimated in the laboratory by gravimetry after extraction of 2 g dry soil homogenized with 5 ml water followed by filtration through 0.2 µm filter and drying at 105 °C. Carbonate alkalinity in the soluble fraction was determined by acid titration monitored by a pH meter, using 5 g dry soil extracted with 20 ml water and after centrifugation at 10,000 × g for 10 min a 10 ml aliquot was titrated to pH 4.5 with 0.1 M HCl providing the value of total soluble carbonate alkalinity (NaHCO₃ + Na₂CO₃).

Enrichment, isolation and cultivation of pure cultures of haloalkaliphilic aerobic hydrolytic bacteria

The general methods for the cultivation of aerobic alkaliphiles have been described elsewhere (Grant, 2006). The basic sodium carbonate mineral medium for cultivation of moderately salt-tolerant alkaliphiles contained 0.6 M total Na⁺ and 1 g l⁻¹ K₂HPO₄ and was strongly buffered at pH 10. After sterilization, the medium was supplemented with 1 mM MgSO₄7H₂O and trace metal solution (Pfennig & Lippert, 1966). The enrichments were performed in 20 ml medium contained in 100 ml bottles closed with rubber septa (to prevent evaporation during prolonged incubation) inoculated with 1 g soil. Incubation was performed on a rotary shaker at 100 rpm and 28 °C. After achieving growth and polymer degradation, the cultures were plated on solid media of the same composition. Five different polymers were used as substrates at concentration 1 g l⁻¹: CMC, soluble starch, casein, powdered alpha-keratin and emulsified olive oil prepared according to Sorokin & Jones (2009). Testing of pure cultures also included 3 additional polymers: beech xylan, amorphous cellulose and chitin prepared as described by Sorokin et al. (2015). In the case of CMC, xylan and olive oil, the solid medium was supplemented with 0.2 g l⁻¹ and in the case of chitin and starch—with 20 mg l⁻¹ yeast extract. Growth of the xylanase-positive cultures on xylan was also tested in liquid culture containing 20 mg l⁻¹ yeast extract. The pure cultures were isolated from individual colonies and checked for purity by repeated re-inoculation on to solid media. The culture purity and endospore formation was also checked by phase contrast microscopy (Zeiss Axioplan Imaging 2; Zeiss, Göttingen, Germany) and, finally, by nucleotide sequencing. The pH profiling of growth and hydrolytic activities was performed on solid media containing 0.6 M total Na⁺ in the form of either NaCl (for pH 5–8) or NaHCO₃–Na₂CO₃ (for the pH range 8–11). The media at pH range 5–8 were buffered with a mixture of potassium phosphates (50 mM) and HEPES (50 mM).

Detection of hydrolytic activities

All activities were detected using plate assays. Beta-1,4-endoglucanase and endoxylanase activities were visualized by using sequential flooding of the plates with 0.1% (w/v) Congo Red and 1 M NaCl each with 30 min incubation (Teather & Wood, 1982). The hydrolysis of keratin, emulsified olive oil, and amorphous chitin and cellulose was directly observed by formation of clarification halos around the colonies (Sorokin & Jones, 2009; Sorokin et al., 2015). The hydrolysis of starch was visualized after flooding the plates with 0.05 N J₂ solution, containing 1% KJ. The hydrolysis of casein was visualized by flooding the plates with 10% (w/v) trichloroacetic acid. For several strains the pH profile and thermotolerance of endoglucanase activity were measured in culture supernatant by agar diffusion approach and measurements of reducing sugar release with DNS (Miller, 1959).

16S rRNA gene sequence and analysis

Genomic DNA was extracted from colony biomass using alkaline SDS cell lysis at 65 °C for 30 min followed by pH neutralization and DNA purification using the Wizard MaxiPreps Purification resin (Promega, Madison, WI, USA). For this, the following steps were taken: (1) cell material taken from solid medium was resuspended in 100 µl of buffer I; (2) 125 µl of lyzing buffer II was added and the resulted mixture was vortexed and (3) incubated at 65 °C for 30 min; (4) 125 µl neutralizing buffer III was added, the resulted mixture was vortexed, centrifugated at 10,000 g for 10 min; (5) 200 μl of the Wizard MaxiPreps resin (Promega) was added to the supernatant and next purification steps were made according to the Wizard DNA Extraction System manufacturer‘s instructions. The final DNA concentration was generally > 10 m kg ml⁻¹, D₂₆₀:D₂₈₀ > 1.8, RNA contamination was less than 1%. Buffer I: 50 mM Tris–HCl, pH 8.0, 10 m M EDTA, 50 µg/ml pancreatic RNAse. Lyzing buffer II: 1% SDS in 0.2 M NaOH. Neutralizing buffer III: 2.5 M CH₃COOK, pH 4.5. The 16S rRNA gene was amplified with bacterial forward primer 11f and the reverse universal primer 1495r. Sequencing was performed commercially using standard Sanger sequencing techniques. The obtained sequences were analyzed using SILVAngs web interface (Quast et al., 2013) on 07.03.2017. The Project summary and settings are shown in Table S1. The 16S rRNA gene sequences of 13 isolates, possibly representing novel taxa, together with the most identical sequences from the Ganbank, verified by BLASTn, were aligned in MAFFT 7 (Katoh et al., 2002). The Maximum Likelihood phylogenetic analysis with General Time Reversible model (G + I, 4 categories, Nei & Kumar, 2000) was performed in MEGA 6 (Tamura et al., 2013).

Isolation and identification of pure cultures of aerobic hydrolytics from saline alkaline soils

A total of 179 strains with one of five polymer degrading activities have been isolated. From the general colony morphology and microscopy, the isolates were obviously dominated by two large groups—actinomycetes (formation of aerial or substrate mycelium) and endospore-forming bacilli. Furthermore, isolates obtained with proteins as substrate also included Gram-negative bacteria. The identification by 16S rRNA gene sequencing generally confirmed this conclusion. The two largest groups of isolates from the saline soda soils are typical hydrolytics belonging to the phyla Actinobacteria and Firmicutes (Fig. 1, Table 2) which may reflect a combination of the specific habitat (Table S2), sampling methods and culture conditions (Duckworth et al., 1996).

The general phylogenetic distribution of the isolates is shown on a Krona diagram, obtained in the course of SILVAngs analysis (Fig. 1) and in the sample-dependent taxa clustering (Table S1). The Actinobacteria were mostly represented by two genera—Nocardiopsis and Streptomyces, and they were closely related to halotolerant alkaliphilic strains and species of these two genera found previously in haloalkaline habitats, in particular in Kenyan and Chinese soda lakes and saline alkaline soils (Grant & Jones, 2016). The relatively low diversity within the otherwise extremely diverse genera of these Actinobacteria indicates that haloalkaline conditions are rather selective for a few highly adapted species. Only two isolates from this group were distantly related to known species. One strain might represent a new genus in the Micromonosporacea with a closest relative from the genus Salinispora, while the second isolate is a distant member in the family Glycomycetaceae (Figs. S1A and S1B, respectively).

Same low genetic diversity was also observed in the second largest group represented by the genus Bacillus. Most of the isolates were closely related to the known alkaliphilic (B. pseudofirmus, B. horokoshii and B. akibai), or haloalkaliphilic (B. halodurans, B. daliensis, and B. alkalisediminis) species. The only exception was a single isolate only distantly related (95% sequence similarity) to B. mannanilyticus—a low salt-tolerant alkaliphilic species producing beta-mannanase (Akino, Nakamura & Horikoshi, 1987; Nogi, Takami & Horikoshi, 2005) (Fig. S1C).

A relatively minor group of isolates enriched with proteins belonged to the proteobacterial class Gammaproteobacteria. A subgroup of three isolates was closely related (99% sequence similarity) to species of the genus Alkalimonas, a known amylolytic haloalkaliphile (Ma et al., 2004). Four isolates were closely related to a haloalkaliphilic member of the genus Aliidiomarina, A. soli, isolated from a soda soil in Inner Mongolia (Xu & Wu, 2017). The third gammaproteobacterial subgroup is represented by 4 proteolytic strains distantly related to organisms in the genus Lysobacter in the Xanthomonadaceae (95–96% sequence similarity). Three out of four strains of this subgroup clustered with an undescribed haloalkaliphilic isolate from Mono Lake (ML-122, 99% similarity), while the fourth strain was distant (96% similarity to ML-122). Therefore, this subgroup probably consists of two novel species and together with the Mono Lake strain ML-122 might represent a new genus in the family Xanthomonadacea (Fig. S1D).

Finally, a significant group of actinobacteria with strong polyhydrolytic potential belonged to the Cellulomonas/Isoptericola clad within the family Promicromonosporaceae (Fig. S1E). The Cellulomonas species are known for their cellulolytic activity and include a haloalkaliphilic isolate from a Kenyan soda lake (Jones et al., 2005), while the genus Isoptericola mostly include halotolarant representatives, although the described neutrophic species apparently have only a limited hydrolytic activity (Schumann & Stackebrandt, 2014).

Hydrolytic spectra of the soda soil isolates

Most of the actinobacteria and bacilli isolates enriched with CMC or starch, were polyhydrolytic, being able to degrade all tested polymers, except for the insoluble native cellulose and chitin (Table 3). Only three actinobacterial isolates showed the ability to hydrolyse amorphous cellulose on the plate assay and only one of the three (DS33), a relative of Salinispora, was actually capable of growth with cellulose as substrate. Six isolates showed a potential to grow with amorphous chitin (Table 3). On the other hand, most of the endo-glucanase and endoxylanase positive actinobacteria and bacilli isolates utilized beech xylan as the growth substrate, which indicates that they are rather specialized in the mineralization of soluble hemicelluloses.

The isolates enriched with proteins belonged to the Gammaproteobacteria and Firmicutes. All of them, as expected, showed highest hydrolytic potential against casein, and many of them did not have endoglucanase, endoxylanase or lipase activities (Table 3). So, they can be considered as dedicated proteolytics. Indeed, proteolytics are the most well-studied group of alkaliphilic hydrolytics.

For the pH profiling, four strains from actinomycetes and from bacilli were selected for test on solid medium containing 0.6 M total Na⁺ with CMC + yeast extract as substrate. The solid medium is not optimal for the profiling but it was chosen for two reasons: (1) the mycelium-forming actinomycetes do not grow homogenously in liquid media and their growth is often estimated by radial colony increase; (2) test on solid medium permitted simultaneous estimation of both growth and endoglucanase activity. The results (Table 4) demonstrated that the tested actinomycetes are facultative moderate alkaliphiles, while the bacilli isolates are obligate alkaliphiles. The endoglucanase activity of both groups had a very broad pH range from six to 11 with an optimum for actinomycetes from eight to 10 and for the bacilli from nine to 10.5.

Overall, the results of this study demonstrated that saline alkaline soils represent a potentially valuable resource of aerobic haloalkaliphilic bacteria capable of producing multiple alkalistable hydrolytic enzymes. Most of the haloalkaliphilic polyhydrolytic isolates belong to Actinobacteria (genera Streptomyces and Nocardiopsis) and the genus Bacillus. We consider the actual capability of a large proportion of the soda soil aerobic haloalkaliphilic isolates to utilize xylan and starch as growth substrates as one of the principal findings of this extended screening. Such organisms definitely represent an interesting object for further investigation of their haloalkalistable hydrolases, particularly with a potential for application in laundry detergent production.