Isolation and Identification of Myxobacteria from Saline-Alkaline Soils in Xinjiang, China

Fifty-eight terrestrial and salt-tolerant myxobacteria were isolated from the saline-alkaline soils collected from Xinjiang, China. Based on the morphologies and the 16S rRNA gene sequences, these isolates were assigned into 6 genera, Myxococcus, Cystobacter, Corallococcus, Sorangium, Nannocystis and Polyangium. All the strains grew better with 1% NaCl than without NaCl. Some Myxococcus strains were able to grow at 2% NaCl concentration, suggesting that these strains may be particular type of terrestrial myxobacteria.


Introduction
Myxobacteria are Gram-negative bacteria with rod-shaped vegetative cells that exhibit unique group behavior [1]. They can produce more than 100 types of secondary metabolites with unique structures and functions, and over 600 types of new analogues [2]. Thus myxobacteria have been regarded as ''microbe factories'' for active secondary metabolites, because they have great potential as a prolific source of new drugs for drug discovery programs [3].
Myxobacteria are widely distributed in natural ecosystem. However, currently less than 10% of all the natural species have been isolated [4]. Therefore, it would be of great potential economic value to improve our capacity to isolate new myxobacteria, in order to provide more strains for drug screening and development.
Myxobacteria are considered typical soil microbes [1], [5], [6], and no purified myxobacteria had been confirmed to be able to grow under salt concentration of more than 1.0% [1] until 1998. Previous study indicated that two swarm-forming bacterial strains had homologous 16S rDNA sequences (89.3% and 83.2% similarities) with Nannocystis [7], a lineage of myxobacteria [8], [9]. Some myxobacteria have been found to exist in marine environments as halotolerants or partial halophiles [7], [10][11][12][13]. For example, halotolerant and partially halophilic myxobacterial strains were isolated from a coastal area in Japan, resulting in the discovery of two new genera, Plesiocystis and Enhygromyxa [14], [15]. However, there have been no reports of myxobacteria isolated from terrestrial saline-alkaline environments. Here we report the isolation of 58 myxobacterial strains from saline-alkaline soils in Xinjiang, China. Some of these terrestrial myxobacterial strains could tolerate NaCl concentrations of 2%.

Soil Collection
Twenty-five samples of saline-alkaline soils were collected from Xinjiang, China. No specific permits were required for the described field studies. We state that no specific permissions were required for the locations where the soils were sampled and we confirm that the location is not privately-owned or protected in any way and that the field studies did not involve endangered or protected species. The samples were air-dried immediately after collection and stored at 4uC.

Strain Isolation and Purification
The strains were isolated using conventional myxobacterial isolation methods. Living Escherichia coli was smeared in the form of a cross-streak on WCX (CaCl 2 .2H 2 O 0.1%; agar 1.5%; pH 7.2). After autoclaving, 25 mg/ml cycloheximide was added to the medium agar plate. 100 g soil was added to 1 L distilled water, boiled for 20 min, and then centrifuged to prepare the soil extracts. The central portions of the streaks were incubated with a pea-sized aliquot of the soil samples and moistened with sterile water. For the isolation of cellulose-degrading myxobacteria, slices of filter paper were placed on top of the ISCX agar (KNO 3 0.1%, FeCl 3 ?6 H 2 O 0.02%, K 2 HPO 4 0.1%, MnSO4?7 H 2 O 0.01%, CaCl 2 ?2 H 2 O 0.1%, MgSO4?7 H 2 O 0.1%, yeast extract 0.002%, agar 1.5%, pH 7.2). They were then inoculated in the center of the filter paper with a pea-sized aliquot of the soil samples. All plates were then incubated at 30uC. After three days, myxobacterial fruiting bodies began to be formed and were usually recognizable within 2-4 weeks. The fruiting bodies were picked daily and examined under a dissecting microscope. They were transferred to VY/2 medium (Barkers yeast 0.5%, CaCl 2 ?2 H 2 O 0.1%, agar 1.5%, pH 7.2). For purification, the agar pieces from the fringes of the swarms were picked with needles and transferred to VY/2. The procedure was repeated until the swarms were pure. Rabbit dung pellet method was also used as described by Reichenbach [16].

Systematic Identification
Morphological identification. All strains were incubated in VY/2 medium. Growth and morphogenesis were observed with dissecting microscope, optical microscope, and electronic microscope. The swarms were carefully scraped to glass slides for detection. Fruiting bodies were crashed to release myxospores for detection. According to The Prokaryotes [1] and Bergey's Manual of Systematic Bacteriology [17], the taxonomy of the isolates were determined by the morphologies of vegetative cells, fruiting bodies, myxospores, and swarms?

16S
rDNA sequencing and phylogenetic analysis. Extraction of genomic DNA, PCR amplification and sequencing of the 16S rRNA gene were performed as described by Li et al. [18]. The PCR products were purified with UNIQ-5 column PCR products purification kit (Sangon) and sequenced at Yingjun Company. The 16S rRNA gene sequences of 58 strains were compared against a database of cultured species via BLAST analysis (http://blast.ncbi.nlm.nih.gov/Blast.cgi) and EzTaxon(www.eztaxon.org) in order to retrieve most similar sequences of recognized bacteria. All the sequences were deposited in GenBank, and the accession numbers were obtained (HQ623094-HQ623124, and K862589-KC862608). Multiple alignments were performed with the CLUSTAL_X software package [19]. The phylogenetic trees were constructed by the neighbour-joining [20] using the software packages MEGA version 4.0 [21].The topologies of the phylogenetic trees were evaluated by using the bootstrap resampling method of Felsenstein [22] with 1000 replicates. G+C mol% assay. The G+C content of the genomic DNA was determined by using the HPLC method [23].
Fatty acid analyses. Strains were cultured on VY/2 agar for 5d at 28uC, and the fatty acid methyl esters were prepared according to the protocol of the Sherlock Microbial Identification System (MIDI system; http://www.midi-inc.com/) and analyzed by GC (6890; Hewlett Packard) using the microbial identification software package [24]. Analysis of whole cell hydrolysates was performed by GLC (Agilent 7890A) according to Kä mpfer & Kroppenstedt [25] using the classical method of the Sherlock Microbial Identification System version 6.1 (MIDI) and the standard MIS library TSBA6. All the experiments were carried out in triplicate.
Multilocus Sequence Analysis (MLSA). PCR primers for genes coding for lepA (encoding leader peptidase, a GTP binding membrane protein) and gyrB (encoding DNA gyrase, subunit B) were selected [26]. PCR products were sequenced at Yingjun Company. All the sequences were compared against a database of cultured species via BLAST analysis (http://blast.ncbi.nlm.nih. gov/Blast.cgi). The sequences were deposited in GenBank to get the accession numbers (KC914480-KC914531). The topologies of the phylogenetic trees were evaluated by using the bootstrap resampling method as previously described.
Effects of NaCl on growth. The effects of NaCl on growth were investigated on VY/2 agar media with 0-5% NaCl added. Each strain was inoculated onto the center of the agar plate and incubated at 30uC for 7-30 days. The growth of each myxobacterial strain was monitored by visual inspection.

Collection of Soil Samples and Soil Properties
Most of the samples are sandy soils, and few are clays from saline-alkaline areas in Xinjiang, China. The vegetations are Ligustrum lucidum, Phragmites australis, Chinese tamarisk, Suaeda giauca and Ziziphus zizyphus, and half of the samples were collected from bare alkaline ditches around the areas. We measured the pH, organic matter content and salt content of soils. The pH ranged from 7.6 to 8.7, and the organic matter content was low, in the   range of 5.9-32.6 g/kg. Most of the soil samples from bare alkaline ditches contained high salt content, with the highest one reaching to 346.8 g/kg (Table 1).

Strain Isolation
Using rabbit dung pellets method, WCX method, filter paper method and soil extract method, 58 strains of myxobacteria from 25 soil samples were isolated from Xinjiang, China (Table 2). Rabbit dung pellets are natural medium for myxobacteria growth. Fruiting bodies can be induced in a shorter time (within 2-3 days) by sterilized rabbit dung pellets, and most of fruiting bodies are Myxococci that are easy to be purified. However, natural rabbit dung pellets are rich in organic matter, and meanwhile are easily contaminated by molds during the induction process, covering the fruiting bodies of myxobacteria. Among the 58 strains, only two were isolated by rabbit dung pellets method. Thirty-two strains were isolated by WCX method, and twenty strains by filter paper method, indicating these two conventional methods were able to isolate more myxobacteria species. Myxobacteria grew rapidly in the WCX medium, and they could be observed under optical microscope after culture for 3 days. While they grew slowly in the inorganic salt filter paper; they took 7-14 days or longer time to grow up. In order to imitate a soil environment to maintain the relationship among microbial populations, soil extracts were also added to be components of the growth medium, which resulted in an increased diversity of microbes. Nevertheless, this soil extract method brought about difficulties for purification, causing contaminants, and eventually only four strains were isolated by this method.

Morphology Characteristics
All of these strains were inoculated on VY/2 medium. Some myxobacterial fruiting bodies, swarms, and vegetative cells appeared (Fig. 1, 2, and 3, respectively). According to the main characteristics of myxobacteria previously described [1], [17], these strains were assigned to six genera: Myxococcus (33 strains), Corallococcus (14 strains), Cystobacter (6 strains), Sorangium (2 strains), Nannocystis (2 strains), and Polyangium (1 strain). Myxococcus strains   were widely distributed in the soil environment, and tolerant to stresses, hence they could almost be isolated from every soil sample in the study. As described in Table 3, for Myxococcus, the fruiting bodies were always spherical, pale yellow, yellow, or orange red. Vegetative cells were always rod-shaped with sharp ends, 2-10 mm in length. For Corallcoccus, the fruiting bodies were always coralloid branched in shape; brown or orange red. Vegetative cells were rod-shaped with sharp ends, between 3-7 mm in length. In Cystobacter, the fruiting bodies always were ovoid and brown. Vegetative cells were rod-shaped with sharp ends and between 3-15 mm in length. Radial patterns were recognized within the swarm area.

16S r DNA Sequence Studies
The 16S rRNA gene sequences of 58 strains were compared by BLAST analysis (http://blast.ncbi.nlm.nih.gov/Blast.cgi) and EzTaxon(www.eztaxon.org), together with the morphologies of their vegetative cells, fruiting bodies, myxospores, and swarms, in order to determine their genera or specie. Among the 58 strains myxobacteria, 26 strains were identified to the species level, with the similarity of 99%-100%, while 31 strains were assigned to the genera level ( Table 4). The level of similarity of xj4 and Polyangium sp. was 96%; 1063 and x3t8 showed a similarity of 96% with Sorangium cellulosum and Phaselicystis flava.
A phylogenetic analysis of 16S rDNA sequences from the 26 new isolates identified at the species level is shown in Fig. 4. These 26 strains were grouped into 5 branches: Myxococcus, Corallococcus, Cystobacter, Sorangium, and Nannocystis. For the Myxococcus branch, the level of similarity among 11 strains was 96%. For Corallococcus, the closest relatives among the 10 strains were Corallococcus coralloides and Corallococcus exiguous (level of similarity 100%). The closest relative of 9613 and xi9-1 was Cystobacter disciformis (level of similarity 91%). The closest relative of xj42 was Nannocystis exedens (level of similarity 100%). The level of similarity of x3t8, xj103, and Sorangium cellulosum was also 100%.

Taxonomic Characteristics
Physiology and biochemical characteristics of new strains. The physiological features of eleven new isolates are shown in the Table 5. All 11 strains were negative for oxidase and   DSM14675, DSM 2260, Cystobacter violaceus DSM14727) ( Table 6). All Myxococcus sp. and Corallococcus sp. reference strains contained substantial quantities of C15:0iso branched fatty acids. Many also contained C17:0iso branched fatty acids. In Cystobacter, the reference strains contained the highest quantities of C16:1 v5c branched fatty acids and also many C15:0iso branched fatty acids. DNA base composition. Five of the new isolates were selected to analyze G+Cmol% ( Table 7). The G+C content of these isolates ranged from 70.0-73.4%.
MLSA analysis. The two protein-coding genes were selected for MLSA studies by Santos and Ochman [25]. Note that although the primers used are considered universal, they were not successful for all of our new strains. For example, using the lepA primer, lepA could not be amplified from 7t, x82, xj81, 7613, 161, or x3t8. For lepA and gyrB, the topologies of gene trees with neighbor-joining analyses were similar, though not identical (Figs. 5 and 6, respectively).

Effects of NaCl on Growth of Myxobacteria
As shown in Table 8, the growth of the reference terrestrial strains tested was inhibited by high salinity. All tested strains were found to grow at 1% NaCl (w/v). Only some Myxococcus strains and one Cystobacter sp. strain were able to grow at 2% NaCl (w/v). No reference strains grew at NaCl concentrations over 2% (w/v).
The taxonomy of myxobacteria is based mainly on the characteristics of the fruiting bodies. However, fruiting bodies are often lost during purification, so it is difficult to identify species such as Myxococcus sp. and Corallococcus sp. Furthermore, the comparison of 16S rRNA gene sequences has been shown to be only minimally useful for discriminating closely related strains [27]. MLSA, however, is a promising method for bacterial systematics [26]. Here we used both 16S rRNA gene analysis and MLSA to identify myxobacteria. As shown in Fig. 4 and Fig. 6, the 16S rRNA gene sequence analysis was found to facilitate the division of samples into genera. For species, however, MLSA was found to be the superior method.
Usually terrestrial myxobacteria cannot grow at NaCl concentration higher than 1%. Our work showed that 7 Myxococcus sp. were capable of growing at NaCl concentrations of 2%. They were also found to form fruiting bodies more quickly at NaCl concentrations of 1% than in the absence of NaCl. We therefore concluded that these 7 isolates were specifically adapted to salinealkaline environments. This is the first report describing the phylogenetic and physiological characteristics of myxobacteria isolated from salinealkaline ecosystems. These results provide us with new information about the ecology of myxobacteria, and the methodologies provided here will be useful for further studies of the ecology of these interesting and economically important organisms.
It is clear that the diversity of myxobacteria is low in salinealkaline soils in Xinjiang. This can be explained by the following reasons. First, the number of soil microbes in saline-alkaline soil is limited. The extreme physio-chemical property and condition in the saline-alkaline environment strongly affect the diversity of microbes, resulting in an unfavorable growth of microbes. However, it is the special characteristics of the saline-alkaline environment that may support much more new unidentified species. Second, the isolation methods are traditional. The conventional rabbit dung pellets method, WCX method, direct heat method, antibiotic mixture method and filter paper method have their own limitations, and they are specific to some myxobacteria but not to all of them. For instance, rabbit dung pellets method is usually used for isolating Myxococus fulvus, Corallococcus coralloides, Cystobacter fuscus, Cystobacter ferrugineus, Archangium gephyra, Stigmatella erecta, Myxococcus vtrescens, Myxococus xanthus and Cystobacter velatus [28], [29]; whereas filter paper method is commonly specified to Stigmatella aurantiaca, Chondromyce apiculatus, Corallococcus coralloides, Myxococus.fulvus, Chondromyce pedicularus, Haploangium sp., Chondromyces sp. and Stigmatella sp. [30]. The limitation of isolation methods causes the loss of myxobacteria discovery. Third, the salinity is high in the saline-alkaline soil, which is not suitable for the growth of myxobacteria. Myxobacteria are prone to the habitat rich in organic matter, with pH 6-8. It has been documented that terrestrial myxobacteria can not grow above 1% salinity [1]. The pH in the saline-alkaline soils of Xinjiang ranges from 7.6 to 8.5, and the salinity of soil is high, ultimately the species are scarce. Last, myxobacteria are also affected by soil property. Myxobacteria are abundant in soil; brown soil and black soil are rich in nutrient, and myxobacteria. But sandy soil and decomposed rock soil are lacking myxobacteria [31]. The sandy soil in Xinjiang is not suitable for the growth of myxobacteria.
In the evolution of 16S rRNA sequences, strains 161, 15, 561, 5611, and x3t7 clustered together, with the similarity of 66%. While in the phylogenetic relationship of gyrB, a close cluster is observed. Among them, 15 and 561 clustered with a similarity of 98%; 161 and 561 clustered with a similarity of 98%; 7613 and x3t7 clustered with a similarity of 84%. Hence, phylogenetic relationship of gyrB may be more suitable to reveal the phylogenetic relationship of those strains.