HCBs occur around the world and are responsible for most aquatic environment pollution [9,10]. Researches of HCBs have been concentrated on the physical, chemical and bio-ecological methods for the control of cyanobacteria and the removal of nitrogen and phosphorous [5,9]. Little is known about the microbial community of cyanobacteria with heterotrophs and the interactions between them [2,3]. Previous studies demonstrated that the oxic cyanobacterial layer of eutrophic water was mainly composed by cyanobacteria and aerobic heterotrophic microorganisms, and the relationships between them were complicated [31,32]. Therefore, it is necessary to obtain the axenic M. aeruginosa from the complex microbial community and further research the interactions between cyanobacteria and heterotrophs.
Traditionally, cyanobacterial purification methods including antibiotic treatment and lysozyme treatment had been applied for eliminating heterotrophs from cyanobacteria and algae [20,21,25,33], and the purification effects were depended on the concentrations and types of antibiotic or lysozyme [22-25]. With a series of antibiotic and lysozyme procedures, the axenic cyanobacteria such as Anabaena flos-aquae, Aphanothece nidulans, Arthrospira platensis and Arthrospira spp. were obtained [22,23]. While the sensitivities of xenic cyanobacterium Microcystis 905 to five antibiotics employed in the present study are quite different, in particular, four of the tested antibiotics have the inhibition effects on cyanobacterium growth. Furthermore, the lysozyme can inhibit both the cyanobacterium and heterotroph simultaneously (Supporting Information of Table S1 and Fig. S1), it is quite difficult to eliminate the heterotrophs from xenic Microcystis 905 culture by antibiotics treatment or lysozyme treatment methods. Researches indicate the bloom forming cyanobacteria in freshwater or seawater are more often occurred in nutrient-rich environments, and the cyanobacteria are surrounded by diverse communities of heterotrophic bacteria [31,32,34]. The difficulty in obtaining the axenic Microcystis 905 is probably due to the lack knowledge of heterotrophs in xenic culture.
Heterotrophs can colonize within the enclosed region or directly adhere to the surface of a cyanobacterium colony . By transferring and culturing xenic culture of Arthrospira platensis in fresh sterile medium, the axenic A. platensis is obtained by the technique of single-trichome manipulation performed with a microtrowel . Considering the xenic Microcystis 905 can easily form single cyanobacterial colony on BG11 agar plate and the growth rates of Microcystis and heterotrophs are significantly different, heterotrophs are removed by solid-liquid alternate cultivation method and micropipette technique, which by picking and transferring the single cyanobacterial colony to BG11 liquid medium under the microscope. This method not only guarantees the minimum initial growth density of cyanobacterial cells, but also ensures the purity of cyanobacterial cells, thus results in the successful separation of the axenic Microcystis 905A. It is also successfully applied to purify other strain such as axenic Microcystis 907A. In spite of the traditional standard plate method based on solid-liquid alternate cultivation for obtaining axenic culture is time-consuming, the protocol that we have developed for purifying axenic Microcystis 905A culture maybe suitable for separating axenic strains from a commensal, and potentially syntrophic, symbiosis. These results indicate that this technique is at least applicable to unicellular cyanobacteria.
Molecular biological techniques such as denaturing gradient gel electrophoresis (DGGE) and fluorescence in situ hybridization have been used to investigate the purity of cyanobacterial culture [17,30]. DGGE results suggest that a number of bacteria including α-proteobacteria, β-proteobacteria, γ-proteobacteria, Bacteroidetes and Actinobacteria have been detected in the cyanobacterial cultures, and the Sphingomonadales are the prevalent group among the Microcystis-associated bacteria ; in another study, the heterotrophs, for instance, Aeromicrobium alkaliterrae, Halomonas desiderata and Staphylococcus saprophyticus are also identified from the Arthrospira platensis culture . The heterotrophic bacteria, such as α-proteobacteria and bacteria from the Bacteroidetes-group, are reported to associate with Diatoms in nature as well as in stock cultures . We observe that the heterotrophs strain B905-1 and B905-2 are closely related to Pannonibacter sp. and Chryseobacterium sp., respectively. Besides the identification of heterotrophs, it seems that more attention should be paid to the interactions between heterotrophs and the cyanobacterium M. aeruginosa. It is suggested that the interaction between heterotrophs and cyanobacterium is symbiosis or parasitic [3,36], and the heterotrophs are difficult to isolate from cyanobacterium during the formation of cyanobacterial or algal colony [1,37].
Heterotrophs can enhance or suppress the growth of cyanobacteria, or even kill them [31,34]. To better understand the general interaction between heterotroph and cyanobacterium, the effect of the strain B905-1 on the cyanobacterium M. aeruginosa FACHB-905A is studied. It is showed that the growth rate of the xenic Microcystis 905 is much faster than that of the axenic xenic Microcystis 905A under both static cultivation and shaking cultivation conditions. The results indicate that the heterotroph B905-1 has a promoting effect on the growth of axenic Microcystis 905A. In consideration of the initial cell number of Microcystis 905 is (2.2 ± 0.2) × 106 cell mL-1 and heterotroph B905-1 is (0.64 ± 0.07) × 106 cell mL-1, it is not surprising that the growth-promoting effect of the 1:10 treatment is much better than the 1:100 treatment. Interestingly, the Microcystis 905A is unable to form colonies in the 1:1 treatment group on BG11 agar medium. Although M. aeruginosa is a kind of photosynthetic bacterium (or autotrophic bacteria) and it grows well under the light with inorganic nutrients, which are supplied by BG11 liquid medium, it is not surprising that axenic Microcystis 905A could not divide at the heterotroph-cyanobacterium ratio of 1:1, as the heterotroph B905-1 can effectively compete nutrients with axenic Microcystis 905A.
The growth-promoting effect of heterotrophs on algae has recently been observed in other studies, for example, the growth of toxic dinoflagellate Alexandrium fundyense is promoted substantially by Alteromonas sp. , and the attached bacteria provide co-existing for diatom Thalassiosira weissflogii to form transparent exopolymer particles . Interpretation of such phenomenon might be explained by the symbiotic interaction that the bacteria deliver vitamins for algae , or the addition of bacteria changes the available nutrient concentration such as extracellular organic carbon or dissolved organic matter [2,4,14,17,31]. In a previous study, the growth rate and metabolic products of Shewanella putrfaciens, Brochothrix thermosphacta and Pseudomonas sp. show a remarkable increase no matter cultured individually or in all possible combinations compared to the control cultures . Difference from the above-mentioned microorganisms, axenic diatoms are unable to form bioﬁlm when purified from bacteria . Although the axenic Microcystis 905A grows well under the liquid culture condition, it could not form cyanobacterial colonies on the BG11 agar plate without the addition of strain B905-1, indicating the presence of heterotroph B905-1 is indispensable for the growth of axenic Microcystis 905A on BG11 agar plate. The different growth phenomenon of Microcystis 905A in solid and liquid BG11 medium is mainly attributed to the phosphate. It is reported that reactive oxygen species (ROS) were produced when phosphate was autoclaved together with agar, and total colony counts of Gemmatimonas aurantiaca in liquid medium (without agar) were remarkably higher than those grown on solid medium (with agar) . In the same way, there may be some ROS produced in BG11 solid medium and the ROS is likely a contributing factor to the growth inhibition of Microcystis 905A. It is speculated that the heterotrophic bacterium B905-1 closely associated with cyanobacterium likely consume nutrients that released by Microcystis 905, and may also produce vitamins and other beneﬁcial metabolites useful for cyanobacterial growth [32,34]. Nevertheless, the presence of strain B905-1 for the cyanobacterial colony formation mechanism needs to be further studied.
Previous study also indicates that the enhancement growth of axenic Microcoleus chthonoplastes PCC 7420 is upon the addition of a filtrate obtained from the closely related xenic culture of Microcoleus sp. M2C3, and the stimulated effect could be due to the release of certain growth factors and vitamins by associated aerobic heterotrophic microorganisms . Most of the strains are able to secrete active substance to inhibit or enhance the growth of cyanobacteria . Possible mechanisms may include various types of interactions from nutrient cycling to the production of growth-inhibiting and cell-lysing compounds . Our results demonstrate that strain B905-1 has the potential to promote Microcystis 905A growth, whereas Microcystis 905A provides organic matter for associated bacterial proliferation. In a comparable study it is pointed out that bacteria have the potential to control diatom growth, and their interactions are regulated by multiple signals involving common biomolecules such as proteins, polysaccharides and respective monomers . In accordance with previous observations, we also find the associated bacterium has promoting effect on the growth of cyanobacterium M. aeruginosa. Increasing knowledge on molecular mechanisms of microbial interactions are crucial to better understand or predict nutrient and organic matter cycling in aquatic environment, and also to better understand the role of such associated bacterium for the formation mechanism of HCBs and eutrophication control.
Up to now, most studies on the interaction between heterotrophs and cyanobacteria are performed in pure cultures [32,34,41], and the growth of the axenic cyanobacteria is almost promoted by the heterotrophs [8,32,34]. However, the interaction can be profoundly different in nature, as most microbes are not axenic but grow together in communities. The complex communities or microbial networks often result in surprisingly coordinated multicellular behaviour, e.g. dinoflagellates can feed on associated bacteria and heterotrophs also attack and lysis the cyanobacteria . Furthermore, the heterotrophs are considered as playing a significant role in carbon cycling and cyanobacterial photosynthesis . All these studies suggest that the relationship between heterotrophs and cyanobacteria in nature is complex and manifold, further analysis is needed to have a full understanding of the microbial communities surrounding cyanobacteria.