The search and discovery of novel microbes that produce new secondary metabolites can be expected to remain significant in the race against new and emerging diseases and antibiotic resistant pathogens (Berdy, 2012; Manivasagan et al., 2013). Actinomycetes are widely distributed in various marine habitats, ranging from sea sand (Hong et al., 2008), mangrove sediments (Hong et al., 2009; Hong, 2013; Azman et al., 2015), sea water (Zhang L. et al., 2012), coastal sediments (Yu et al., 2015), and deep sea sediments (Zhang et al., 2015; Chen et al., 2016). The increasing number of literature on marine actinomycetes strongly supported the view that marine environments including deep sea are significant sources for search and discovery of both diverse actinomycetes resources and secondary metabolites (Skropeta and Wei, 2014; Xu et al., 2014).

Deep sea habitats show extreme variations in available nutrients, light, oxygen concentration, pressure, salinity, and temperature. Deep-sea organisms have developed unique biochemical metabolic and physiological capabilities, which not only ensure their survival in this habitat but also provide potential for the production of novel metabolites absent in terrestrial microorganisms (Fenical, 1993; Bull et al., 2000; Skropeta and Wei, 2014). Through molecular ecology studies, actinobacterial operational taxonomic units (OTUs) have been identified from deep sea sediments. Most of those foreseeably exhibit novel species, genera and families (Stach and Bull, 2005; Chen et al., 2016). Diverse species of actinomycetes cultured from the deep seafloor surface, including the deepest sea sediment samples from the Mariana Trench, have shown great biosynthetic capacities and thus a potent source of novel natural products (Pathom-aree et al., 2006d; Abdel-Mageed et al., 2010). With the breakthrough of technological barriers associated with deep sea actinobacteria isolation strategies, such as sample collection and cultivation under standard laboratory conditions, more and more deep sea actinobacteria and their natural products have been identified. Here we review the recent progress on deep sea actinomycetes and their metabolites from literature during year 2006–2016.

The vast oceans cover 70% of the world's surface, with 95% greater than 1,000 m deep. Deep sea environments are divided into the bathyal zone (depths between 200 and 2,000 m), the abyssal (depths between 2,000 and 6,000 m) and the hadal zone (depths below 6,000 m) (Harino et al., 2009). Below sea level pressure is increased by depth, thereby in the deepest part of the trenches, the pressure varying from 10 atm at the shelf- slope interface to >1,000 atm. At bathyal depths temperatures taper off rapidly with increasing depth to 2°C. Deep-sea species must adjust their biochemical processes to survive in low temperatures, because the cold reduces chemical reaction rates. Oxygen concentration drops along with the depth, oxygen-minimum layer in mid-water, usually between 300 and 1,000 m depth. Light intensity decreases exponentially with depth in the water column. No photosynthetically useful light reaches the sea floor below about 250 m (Thistle, 2003).

Start at about 200 m depth, the deep sea is characterized by high pressure, low temperature, lack of light and variable salinity and oxygen concentration (Das et al., 2006), at the shelf break, where a clear change of fauna from shallow to deep water is observed (Thistle, 2003). According to Haefner (2003), in cold deep sea mud the diversity of life can be remarkably high with species richness rivaling that of tropical rain forest. Studying the species level of microbial diversity, finding a large number of rare species which more than half of them considered as new species and more than 95% is unidentified, furthermore the expanding of biodiversity reach to the 5,000 m in depth to abyssal which the peak amount of species at the depths of 3,000 m and beyond (Skropeta, 2008). On earth abyssal hills are the most abundant of biomass, but on wider abyss the ecological impact of the habitat heterogeneity is largely unexplored (Durden et al., 2015).

However, so far only a few actinomycetes have been isolated from deep sea. It is because of technological barriers associated with isolation strategies. Therefore, we are in the pace to develop efficient cultivation methods to recover the actinobacteria population from extreme deep sea habitats. To achieve the task, firstly collection of samples from deep sea plays a pivotal role. In recent years several advancements have been developed in the context of sample collection from deep sea such as modified sediment grab and designer-built bounce corer (Fenical and Jensen, 2006), remote-operated submarine vehicle (Pathom-aree et al., 2006d), neuston sampling devices (Hakvåg et al., 2008), multi-core sampler (Xu et al., 2009), gravity or piston cores (D'Hondt et al., 2009), and untethered coring device (Prieto-Davó et al., 2013).

It is crucial to cultivate deep sea actinomyetes under standard laboratory conditions. There are several factors that influence the isolation, such as pre-treatment of dry heat (Shin et al., 2008), media composition (Luo et al., 2011; Pan et al., 2013; Song et al., 2015), dilution factor (Pathom-aree et al., 2006a), seawater requirement (Song et al., 2015), artificial seawater (Pan et al., 2013; Pesic et al., 2013) and incubation time (Song et al., 2015). It has also shown the addition of different antibiotics on selective media can inhibit the growth of fungal and bacterial contamination in order to enhance the actinomycetes growth similar to those used in isolation of actinomycetes from terrestrial sample. Long term freeze storage of deep sea sediment samples at −80°C has shown to prevent the growth of fast-growing bacteria which in results enhance the actinomycetes population (Ulanova and Goo, 2015). For the initial isolation of Streptomyces, cultivation temperatures have also influenced the recovery from deep sea sediment samples. Optimal growth temperature generally ranging from 25 to 30°C for successful cultivation of deep sea actinomycetes (Jeong et al., 2006; Luo et al., 2011; Pesic et al., 2013).

Heat pre-treatment procedures have been used effectively for the selective isolation of members of several actinomycete taxa and also inhibited growth of bacterial and fungal colonies. Moreover, actinomycete spores and hyphae are more sensitive to wet than dry heat hence relatively low temperature regimes are used to pretreat water and soil suspensions. Although heat pretreatment procedures decrease the ratio of bacteria to actinomycetes on isolation plates, the numbers of actinomycetes may also be reduced (Williams et al., 1972; Pathom-aree et al., 2006a,b,c,d). Pathom-aree et al., isolated actinomycetes from Norwegian fjord sediments support that the numbers of actinomycetes were reduced when used heat pretreatment for isolation; fewer actinomycetes were isolated on selective media inoculated with suspensions treated at 55°C as opposed to 50°C. Similarly, higher counts were generally recorded on isolation plates seeded with non-heat pretreated suspensions (Pathom-aree et al., 2006d).

For the other method, Jensen et al., 2005 used dry and stamp method for isolation actinomycetes from tropical Pacific Ocean and found that using this method for isolation of actinomycetes showed good recovery of 44%. In addition, Ulanova and Goo (2015) found that the majority of actinomycete-like colonies were also isolated using dry stamping technique from subseafloor sediments at the Nankai and Okinawa Troughs.

Novel actinomycete species isolated from deep sea environment between 2006 and 2016, have yielded an impressive array of novel species with the highest number found at depths of abyssal zone and deeper. Different media has been used by researchers (Table 1). It is worth to be noticed that long time culturing and low temperature were employed for some of the novel isolates (Table 1). Only one novel Microbacterium marinum was obtained by pretreatment at 55°C, 6 min, others were from none heat pretreated samples (Table 1). The novel deep sea actinomycete species including 21 species under 13 genera with the maximum number from Microbacterium (n = 4), followed by Dermacoccus (n = 3), Streptomyces (n = 3) and Verrucosispora (n = 2), and one novel species for each of the other 9 genera (Table 1).

The numbers of novel microbial metabolites from deep sea sediment samples have been increasing, especially from deep sea streptomycetes. Eight genera of actinomycetes were reported to produce secondary metabolites, among which Streptomyces is the richest producer (Table 2). Earlier culture dependent studies strongly suggested that Streptomyces species are present in considerable number in deep sea sediment samples (Jensen et al., 2005; Pathom-aree et al., 2006d). In addition several novel species of deep sea derived Streptomyces strains with distinct metabolites have been reported which indicates deep sea Streptomyces are really worth in the context of novel natural products discovery (Pan et al., 2015; Song et al., 2015).

The deepest sea sediment samples from the Mariana Trench have been shown to possess great biosynthetic capacities. Seven dermacozines A–G were reported from the actinobacteria Dermacoccus abyssi sp. nov., strains MT1.1 and MT1.2 isolated from Mariana Trench sediment collected at a depth of 10 898 m. Dermacozines F and G displayed moderate cytotoxic activity against the leukemia cell line K562 with IC₅₀ values of 9 and 7 mM, respectively, whereas dermacozine C also exhibited high radical scavenger activity with an IC₅₀ value of 8.4 mM (Abdel-Mageed et al., 2010).

In recent years, South China Sea has been emerging as a potentially abundant source of novel species/genera of marine actinomycetes. Some bioactive compounds, such as pseudonocardians A-C, grincamycins B-F, and abyssomicins J-L were reported. Natural products derived from deep sea actinomycetes discovery have displayed a wide range of bioactivities, such as antitumor, antimicrobial, antifouling, and anti-fibrotic activities (Table 2).

Lobophorins H and I together with three known analogs, O-β-kijanosyl-(1 → 17)-kijanolide, lobophorins B and F were yielded by Streptomyces sp. 12A35, isolated from a deep sea sediment sample collected at a depth of 2,134 m in South China Sea (Pan et al., 2013). While, lobophorins E and F, along with two known analogs lobophorins A and B were discovered from the products of the deep sea Streptomyces sp. SCSIO 01127, was isolated from sample collected at a depth of 1,350 m in the South China Sea (Niu et al., 2011). The gene cluster involved in biosynthesis of lobophorin was the first type I PKS gene cluster identified from the deep sea derived Streptomyces. Three glycosyltransferases (GTs) LobG1-LobG3 genes-inactivation mutants yielded five different glycosylated metabolites, and the result suggested that LobG3 as an iterative GT to attach two L-digitoxoses (Li et al., 2013). Desotamides B, C and D together with a known desotamide A were obtained from deep sea derived Streptomyces scopuliridis SCSIO ZJ46, recovered from sediment sample collected at a depth of 3,536 m in the South China Sea (Song et al., 2014). A 39 kb gene cluster governing the biosynthesis of the anti-infective desotamides has been isolated from the strain. Desotamides A and B and a new desotamide G have been obtained by heterologous expression of desotamide gene cluster in Streptomyces coelicolor M1152 (Li et al., 2015).

Heronamides D, E, and Fare discovered from the products of Streptomyces sp. SCSIO 03032, which was isolated from deep sea sediment sample collected at a depth of 3,412 m in the Bay of Bengal, Indian Ocean (Zhang W. et al., 2014). The gene cluster governing the biosynthesis of heronamide has been isolated from strain SCSIO 03032. The gene inactivation study confirmed that P450 enzyme encode HerO as an 8-hydroxylase for tailoring heronamide biosynthesis. Feeding experiments with labeled small carboxylic acid molecules confirmed the migrated double bonds in the conjugated diene-containing side chain of heronamides (Zhu et al., 2015).

Marformycins A-F were obtained from fermentation broth of deep sea sediment-derived Streptomyces drozdowiczii SCSIO 1014, which was isolated from sample collected at a depth of 1,396 m in South China Sea. All compounds exerted selective anti-microbial activity against Micrococcus luteus, Propionibacterium acnes, and P. granulosum. Marformycins A-E displayed inhibitory activity against M. luteus with MICs of 0.25, 4.0, 0.25, 0.063, and 4.0 μg/mL, respectively, while they did not displayed any cytotoxicity (Liu D. et al., 2015). It is suggested that these compounds may be used as promising candidatures for anti-infective drug leads. The gene cluster that responsible for the biosynthesis of marformycin is about 45 kb in size and has been identified from strain SCSIO 10141. The gene inactivation studies indicated that three NRPS proteins MfnC, MfnD, MfnE, a free adenylation (A) enzyme MfnK, and a free peptidyl carrier protein (PCP) MfnL were essential for the generation of the marformycin core scaffold. Further, MfnN was found to use an intact cyclodepsipeptide intermediate as its substrate (Liu J. et al., 2015).

The discovery of novel actinomycete taxa with unique metabolic activity from deep sea samples, and novel compounds with the greatest biogenic, metabolic diversity and biological activities clearly illustrate that indigenous deep sea actinomycetes indeed exist in the oceans and are an important source of novel secondary metabolites. Other function of deep sea actinobacteria is also interesting such as oil degradation and biosurfactant production (Wang et al., 2014). It is worth to be noticed that no heat pretreatment, dry and stamp method and low temperature incubation were more productive for actinomycetes isolation from some deep sea samples. With the development of culture independent techniques, more productive strategy of strain isolation guided by the deep sea actinomycetes distribution or direct cloning and heterologous express the functional genes could be approached.

MK contribute the introduction, deep sea environment and biodiversity, actinomycete cultivation, novel taxa, and Table 1. PS contribute sample collection, Table 2 and biosynthesis of secondary metabolites from deep sea streptomycetes. KH and ZD conceived the idea and revised the whole manuscript.