The marine environment is a complex ecosystem with an enormous diversity of different life forms often existing in close associations. Among these, microorganism-eukaryote associations have gained significant attention in the past decade [21].

The surfaces of all marine eukaryotes are covered with microbes that live attached in diverse communities often embedded in a matrix, forming a biofilm. The microbial consortia living on various eukaryotes differ significantly from each other and from the microorganisms living in the surrounding seawater [9–16,42,43]. For example, a DGGE based comparison of the microbial community composition of the coral Montastraea franksi and the surrounding seawater revealed almost no overlap [42]. Likewise, Longford et al., [10] found only two (out of a total of one hundred) bacterial species that were common to three different marine sessile eukaryotes. Moreover, host specificity has also been illustrated by studies that have shown the presence of unique stable communities living on geographically distant individuals belonging to the same species [14,44].

In contrast to free living planktonic microorganisms, which often encounter fluctuations in environmental conditions, requiring quick short-term adaptive responses, surface associated microorganisms supposedly have developed more specialised and stable adaptations, specific to the microenvironment created by a particular host. The high level of specificity of microbial communities on various marine eukaryotes highlights the existence of close cross-relationships between microbial epibionts and their eukaryotic hosts. In fact, some epibiotic microorganisms have been shown to be essential for the normal life and development of the eukaryote, for example, being involved in the development of host morphology [45–49]. Host specific bacteria have also been shown to be vertically transferred from the parental eukaryotic organism to its offspring, which indicates the importance of these bacteria for the host. Such inheritance of members of the microbial community has been reported to occur in sponges [16,50,51], bivalves [52] and ascidians [53]. While the exact nature of the relationship between the microorganisms and their hosts remains unclear, it has been hypothesized that the microbial partners construct chemical microenvironments with the eukaryotic host and live in syntrophy, participating in cycling of nutrients, as well as preventing predation of the host via the production of bioactive molecules [50].

Marine surface associated bacteria are often metabolically linked with their host. For example, epiphytes belonging to the Roseobacter lineage are known to degrade the algal osmolyte dimethylsulfoniopropionate (DMSP) yielding the climate-relevant gas dimethylsulfide (DMS) and are regarded as major players in sulphur cycling in the ocean [54,55]. In contrast, some marine eukaryotes heavily rely on the metabolites produced by their microbial symbionts to survive. For example, some marine sponges use the carbon produced by their associated photosynthetic cyanobacteria [56] and may even rely on their autotrophic cyanobacterial symbionts to provide more than 50% of their energy requirements, which allows them to grow in low-nutrient environments [57].

Close metabolic associations between microorganisms and their host can make it difficult to reveal which partner organism is responsible for the production of a particular metabolite. As a result, many bioactive products, previously ascribed to the eukaryotes, have later been found to be produced by associated microorganisms [58–66]. For example, the microbial origin of the cytotoxic compound bryostatin was demonstrated by the identification of polyketide synthase genes, involved in its biosynthesis, in the genome of the bryozoan bacterial symbiont “Candidatus Endobugula sertula” [67]. Moreover, it was proposed that microbially derived bryostatin, found on the larvae of bryozoan Bugula neritina, serves to defend the larvae against potential predators [68].

The interactions between the epibiotic microorganisms and their host, in which microorganisms are thought to acquire nutrients from the eukaryote, while the host benefits from the wide range of bioactives produced by its associated microorganism, seems to be widespread in the marine environment [69]. For example, the gamma-proteobacterium Pseudoalteromonas tunicata, known for the production of several bioactive compounds, is proposed to play a role in defending the host against surface colonisation by producing antimicrobial, antilarval and antiprotozoan compounds [70–73]. Likewise, the surfaces of the healthy embryos of the lobster Homarus americanus are covered almost exclusively by a single gram-negative bacterium, that produces an antifungal compound highly effective against the fungus Lagenidium callinectes, a common pathogen of many crustaceans [74].

The production of antimicrobials by epiphytic microorganisms could also give the producers a distinct advantage in competition with other surface-dwelling microbes. This is especially important given the fierce competition that exists on the surfaces of marine living organisms, that are relatively rich in nutrients compared to seawater, and, therefore, are attractive for numerous microorganisms [21].

The fact that many marine microorganism-host associations are based on metabolic or chemical interactions may explain the abundance of bioactive producing bacteria on living surfaces [17–19,75,76]. Thus, a better understanding of the ecological challenges and the underlying mechanisms involved in such interactions should accelerate the search for novel bioactives.

The limited ability to culture the majority of environmental strains represents a major bottleneck in classical culture-based screening programs for microbial derived bioactives, including those from marine surfaces. It is estimated that the majority (98–99%) of microorganisms cannot be cultured by traditional techniques [79–81]. Nevertheless, marine living surfaces may provide an advantage as, in some cases, a higher percentage of eukaryote associated microorganisms can be readily cultured [82].

Being able to grow the organism in vitro provides great advantages, such as better access to its physiology [83,84]. This may allow the manipulation of different growth parameters to achieve the maximum yield of various products and for their large-scale production via fermentation. Among strategies to improve the culturability of microorganisms, those that attempt to grow the organisms under conditions that mimic the physical and chemical parameters of their natural environment, have been the most successful. For example, by using specialised environmental chambers, Kaeberlein et al., [85] could successfully culture up to 40% of of all microbial cells present in a marine environmental sample.

The close associations often present between marine eukaryotic organisms and their microbial epibionts, clearly impose conditions that are difficult to replicate with standard laboratory procedures. It has, therefore, been proposed that the development of suitable culturing techniques for such organisms should involve conditions that reflect the microenvironment created by their host [86]. This approach has proven successful for the isolation of the sponge associated bacterium Oscillatoria spongeliae [87], for which hyperosmotic medium, resembling the osmolarity of the sponge mesophyle, was used to cultivate the organism.

Cultivation conditions, such as temperature, aeration, pH of the media, incubation time and media composition, can affect the production of the desired metabolite, and, therefore, must be taken into account and fine-tuned [88–94]. Usually this requires a producer strain to be grown in the conditions optimal for the production of the active compound. These conditions can differ significantly from the optimal growth conditions of the strain. In some cases, the producer organism is grown under a variety of conditions in parallel and the differences in metabolic spectra are assessed [95–97]. Marine surface associated microorganisms may also require conditions that resemble their native environment in order to produce the maximum amount of bioactives. For example, several studies have shown an increase in the production of antimicrobial compounds when the surface associated bacteria were grown, in vitro, to form surface attached biofilms [41,98,99]. In addition, Okazaki et al., [100] have shown that marine isolate SS-228 was able to produce the antibiotic compound only when the growth medium was supplemented with powdered Laminaria seaweeds, common in the habitat from which strain SS-228 was obtained. It is now becoming clear that knowledge of a microorgansims’ natural habitat, including the specifics of the host organism, can improve production of microbial derived bioactives.

Sequence information obtained via the sequencing of the environmental DNA (“metagenome”) can greatly assist in understanding the metabolic potential of the organisms present in the environment, and thus guide the development of specialised cultivation conditions. For example Tyson et al., [101] successfully cultured Leptospirillum ferrodiazotrophum by developing a selective isolation strategy. The predicted nitrogen fixing capability of this organism, based on the sequence information of an acid mine drainage biofilm community, underlined the development of that strategy.

Over the past decade genomics has emerged as an alternative to directly culturing microorganisms for the isolation of new bioactives. In particular, functional metagenomics was first developed to tackle the biotechnological potential of unculturable microorganisms. In this approach the DNA obtained from the environment (“environmental DNA”) is inserted into a host organism, such as E. coli, and a functional screen of libraries is performed to detect the desired activity in the clones [102–107]. Some of the successes of this approach were the discovery of terragine A [108], bioactive N-acyl-tyrosine derivatives [109] as well as indirubin [110] from the soil metagenomes.

Functional (meta-)genomics can provide an insight into the genes and gene clusters involved in the production of certain metabolites, and, thus, provide information about the possible biosynthetic pathway leading to that metabolite [111]. Such an approach has been used by Burke et al., [112] to propose the biosynthetic pathway of the antifungal compound tambjamine produced by the marine bacterium P. tunicata. Recently the same approach was successful in identifying two positive clones from a P. tunicata genome library, with different modes of action against the nematode Caenorhabditis elegans (Ballestriero et al., unpublished data). In addition, a functional (meta-) genomics approach provides an advantage for further purification of the bioactive compound produced by a clone, since the “extra” metabolite can be relatively easily pinpointed by using a reference non-bioactive-producing clone [113].

However, despite some success, currently the hit rate of using metagenomic functional screening to obtain bioactive producing clones generally remains low, in the order of 1 in 10,000 [109,114], or even as low as 1 in 730,000 [115] clones screened. This low hit rate is mainly a result of the limited ability of host expression strains to express compounds of foreign origin [78,116,117]. Therefore it is possible that with improvements of such strains the hit rate for positive clones in metagenomics functional screens will greatly increase. For example, recent studies have demonstrated that the use of a variety of host expression strains can assist in the expression of the desired metabolite [118–121]. Specifically, these strains are often chosen based on possible similarities with the producer bacteria (if known), such as, for example, the similarities in codon usage, as well as presence of specific machinery necessary for the production of particular metabolites [119,120].

Alternatively, shotgun sequencing of environmental DNA and subsequent data analysis has the potential to identify genes encoding new structures of known compound classes, e.g., polyketide synthases (PKS) and non-ribosomal peptide synthetases (NRPS) usually involved in production of bioactive secondary metabolites [122]. For example, recent analysis of the genome of P. tunicata has revealed the presence of nine NRPS [123]. Two products of these NRPS with predicted biological activity have been recently identified in the laboratory via heterologous expression, and their presence was confirmed in the original strain of P. tunicata by varying its conditions of growth [124].

Moreover, the availability of sequence data from a variety of microorganisms has further highlighted the importance of developing culturing conditions that would be suitable for the production of bioactives. There are now several examples of genes involved in the biosynthesis of bioactives found in non-bioactive producing organisms, suggesting that, given suitable growth conditions, these organisms have the potential to produce bioactive metabolites. For example, the genome of the myxobacterium Stigmatella aurantiaca DW4/3-1 showed the presence of multiple polyketide synthase/non-ribosomal peptide synthetase gene clusters, which had not been previously observed in this organism [125]. In another example, the previously unknown potential of the production of the antitumour compound terrequinone A in the fungus Aspergillus nidulans was revealed [126].

In the early years of antibiotic discovery the selection of antibiotic producer strains was based on morphology rather than genotypes, resulting in redundancy in many natural product extract libraries [127,128]. The initial success in the discovery of many antibiotic compounds from natural sources included thousands of compounds being described within a few decades. However, the lack of a systematic approach often resulted in the frequent re-discovery of known compounds. Therefore, it is important to put considerable effort into the de-replication process for early detection of both the known producer organisms as well as the known bioactive compounds.

Currently, 16S rRNA gene sequencing [129] and phenotypic characterisation can serve for the identification of producer organisms, and, hence, may reveal whether a given microorganism or its close relatives have been previously known to produce certain bioactives, in order to focus the efforts on organisms with yet uncharacterised activity.

Concerning the early detection of known bioactive compounds, advances in the development of new chemical analysis techniques, coupled with a database evaluation, can serve as tools for rapid detection of known compounds requiring only small quantities of sample and/or minimal efforts in sample preparation [130–132]. Hence, they may greatly assist in preventing the waste of resources, which would otherwise be necessary for scale-up and characterisation of the bioactive compounds.

The exploration of relatively unexplored environments can also assist in finding novel microorganisms and chemical structures, and, hence, minimise the re-discovery of known compounds. The oceans have proven to be a habitat for many unique microbes [133], such as, for example, the recently discovered marine genera Salinispora and Marinophilus [133,134]. In addition, some of the members of these groups were found to produce structurally novel bioactive metabolites, for example, salinosporamides, a family of compounds with cytotoxic activity, were successfully isolated from Salinnispora tropica [135]. Likewise, the structurally novel compounds marineosins [136] and largazole [137] have been recently isolated from marine bacteria belonging to the actinomycete and cyanobacterial groups respectively. Recently discovered marine bioactive natural products including compounds with unique structures are reviewed in [138]. This supports the idea that the unique chemical and physical parameters of the marine environment can lead to the evolution of life forms that could also produce metabolites with novel chemical scaffolds [139,140].

Despite their potential, full characterisation of marine microbially derived bioactives, as well as the development of extraction and purification strategies can be a long and laborious process requiring a great deal of manual work, with little room for automation [78,141].

Purification and structure elucidation of mass limited sample material is considered a major bottleneck. This is usually because the compound of interest often represents less than 1% of the crude extract, which, in most cases, is a mixture of hundreds of different compounds. Therefore, every extract has its unique combination of “contaminants” necessitating a specific approach; as a result, development of the purification strategies remains largely experimental [142].

Furthermore, obtaining adequate quantities of bioactive compounds, necessary for structure elucidation and evaluation, usually requires extensive optimization of conditions and scale-up [143].

To facilitate the chemical characterisation, analytical methods are constantly being developed and improved, one of the major lines of improvement being the possibility of using small quantities of sample and easy sample preparation. For example, the recently developed Ultra High Performance Liquid Chromatography (UPLC) coupled to high resolution mass spectrometers (MS), as well as capillary probe nuclear magnetic resonance spectrometers have greatly assisted the process of natural product discovery from mass limited samples [130,144,145]. Likewise, the newly developed Desorption Electrospray Ionisation Mass Spectrometry (DESI-MS) technique allows for the rapid detection of the compounds requiring minimal effort to be spent on sample preparation [131,132]. These techniques may soon eliminate, or, by some estimates, have already eliminated the purification and chemical characterisation step as a major drawback in natural product discovery [146–150].