The Deep-sea Hypersaline Anoxic Basins (DHABs) of the Mediterranean Sea are one of the most extreme oceanic realms known on Earth. The bottom sediments of these regions are completely anoxic and covered by a thick and dense brine (from tens to hundreds of meters), which hampers oxygen exchange. In particular, in the L’Atalante basin, the anoxic conditions are present since more than 50.000 years . These conditions have been assumed for a long time to be excessively harsh to allow the life of multicellular eukaryotes, at least until the recent discovery of three new species belonging to Loricifera, a group of microscopic invertebrates (Metazoa). These three species were apparently able to live and complete their entire life cycle without access to free oxygen . Using different and independent analyses based on incubations with radioactive tracers and specific fluorogenic probes (e.g. CellTracker Green), quantitative micro X-ray and infrared spectroscopy, and accurate analyses of different components of life cycles, Danovaro et al.  concluded that the loriciferans inhabiting the L’Atalante basin are metabolically active and show specific adaptations to the anoxic conditions. Furthermore, SEM and TEM analyses provided evidence that the cellular tissues were not degenerated.
Bernhard et al. , conducted an investigation in the same deep hypersaline anoxic basin (L’Atalante). Due to technical difficulties, the sampling of the bottom sediments, i.e., beneath the brines, was not possible. However, the authors collected sediment samples from the hypoxic redoxcline. Bernhard and coworkers found 10 specimens of Loricifera in the lower halocline, and only one from normoxic sediments (a specimen that resembles the recently described Spinoloricus cinziae; cf. ). They treated with DAPI (a fluorochrome used for DNA staining) the two Loricifera specimens (one from the halocline and one from the normoxic condition) and found that they were weakly stained. Moreover, the staining of the same individuals with Rose Bengal revealed the presence of a putative oocyte, but no other identifiable internal organs were observed. The presence of cadavers and animal remains (e.g., dead copepods and their exuviae) in the L’Atalante basin was pointed out by both teams [2, 3]. Bernhard et al.  suggested the possibility that benthic storms, namely those reported in the North-Western Mediterranean (which are ca 3000 km apart), could have transported the cadavers of the small crustaceans (and thus also the Loricifera) within the basin. However, invoking major physical processes to explain the presence of dead copepods in the system is not necessary since these very small organisms (size in the order of 150 μm) are able: (i) to swim in the boundary layer and thus they can be transported by deep-sea currents , and (ii) to enter in the system by simple sedimentation. In addition, there are no signs of storm events from the perfectly undisturbed and stratified sediments of the DHABs .
The analysis of SSU rRNA from the sediments of the halocline of the L’Atalante as well as of the normoxic sediments revealed the presence of a very low contribution of reads belonging to multicellular organisms, which were mainly represented by pelagic crustaceans. Nevertheless, Bernhard and collaborators could not find sequences of the nematodes that they reported as dominant taxon in all samples and, according to ultrastructural analysis, showed the presence of healthy tissue in normoxic sediments at the time of sampling. The authors also conducted in situ incubation experiments using CellTracker Green on normoxic sediments and on one sediment sample from the upper halocline of L’Atalante, but did not find Loricifera in their sample and thus could not obtain data from these experiments. However, they found labeled nematodes (indicating esterase activity at the time of incubation) and concluded that this reflected the nematode viability because no parasitic or scavenging prokaryotes were found on the nematode cuticle.
In our opinion, these results corroborate our previous findings  as we used the same analyses utilized by Bernhard et al.  to demonstrate the viability of nematodes in normoxic sediments as a proof of viability of the Loricifera in the permanently anoxic sediments of the L’Atalante basin (see below for a detailed analysis).
Moreover, the observation reported by Bernhard et al.  of living nematodes present at the oxic/anoxic interface (i.e., potentially moving actively in hypoxic-anoxic conditions) is interesting though not novel to science . Organisms living at the normoxic/anoxic interface have been defined as Thiobios almost 50 years ago . Interestingly, when the existence of these organisms was proposed, it was initially rejected: “Thiobenthos does not exist” . Nowadays, the existence of the Thiobios (or Thiobenthos) is universally accepted by the scientific community , and in our opinion the loriciferans inhabiting the L’Atalante basin represent an example of the possibility of metazoan life in anoxic conditions.
The conclusions made by Bernhard et al.  diverge from those proposed by Danovaro et al.  in four main points. Bellow, we discuss the different methodologies and approaches utilized for providing evidence of the presence of metazoans living in anoxic conditions and compare our conclusions with those presented by Bernhard et al. .
Evidence based on cell tissue staining
Danovaro et al.  pointed out that loriciferans found in the sediments of L’Atalante Basin could be perfectly stained with Rose Bengal, while all specimens of nematodes and copepods retrieved from the same sediment samples did not show or showed only a very weak staining. Furthermore, signs of the Rose Bengal staining in loriciferans was found in all their tissues/organs (e.g., brain, muscles, oocytes and epidermis cells).
In contrast, Bernhard et al.  suggested that the loriciferans stained with Rose Bengal were actually dead and the positive staining color was resulting from the presence of anaerobic bacteria, archaea and/or fungi living within the exoskeleton of the decaying loriciferans. However, this hypothesis should be ruled out because the investigation of some of the loriciferans found in the L’Atalante basin, and recently described as Spinoloricus cinziae , showed that specimens had very clean and non-decaying bodies. Indeed, a SEM analysis rendered images showing the absence of a single bacterium on the lorica or on the many spinoscalids of the head of the specimens investigated. This would not be possible if the animal has been dead. Furthermore, if the Rose Bengal staining was due to prokaryotes/fungi colonizing the degraded animal tissues, as suggested by Bernhard and co-workers, then a positive staining should be observed also in the dead nematodes and copepods found in the same samples.
All of the microscopic methodologies previously utilized (confocal laser microscopy, contrasting-phase microscopy, SEM and TEM) [2, 4] did not show any sign of degraded tissues, and the loriciferans were found either fully retracted, or partially retracted or fully extended (figures 7–11 in ). Moreover, since dead loriciferans are usually seen as fully extended and not stained by Rose Bengal, these features provide a good indication that these organisms were active at the time of sampling and responded to changes in the surrounding environment. Thus, although the Rose Bengal per se is not sufficient to prove that loriciferans were alive (as also reported by Danovaro et al. ), the criticism of Bernhard et al.  is not supported by the data presented. Therefore, according to available results, the loriciferans were collected alive.
Evidence based on incorporation of radiolabeled substrates
Another technique utilized to provide evidence of active metabolism of loriciferans extracted from the anoxic sediments is the incorporation of radiolabeled organic substrates . In contrast, Bernhard et al.  suggested that the radioactivity incorporated by loriciferans could be due to bacteria or archaea present within their body. In this regard, it is known that heterotrophic prokaryotes (either bacteria and archaea) can uptake leucine [11, 12]. However, the TEM investigations at the ultrastructural level of loriciferans from the L’Atalante basin by Danovaro et al.  provided evidence of the complete lack of abundant or aggregated prokaryotes, within the body of loriciferans. As reported above, SEM analyses also demonstrated the absence of prokaryotes in the lorica or in the many spinoscalids of the head providing evidence that the incorporation of radiolabeled leucine occurred within the tissues of loriciferans. Moreover, the magnitude of the radiolabeled substrate uptake makes highly improbable alternative explanations, even assuming a potential contribution from the symbiotic bacteria present within the animal tissues.
Evidence based on metabolism (CellTracker Green labelling)
Bernhard et al.  stated that although esterase activity in loriciferans was clearly detected by CellTracker Green labeling performed by Danovaro and co-workers , bacteria also react to this fluorescent dye  and, hence, could produce the fluorescent reactivity observed inside the loriciferans. However, confocal laser microscopy analysis carried out on Spinoloricus cinziae (which was analyzed to the highest detail; ), revealed that the fluorescence was clearly present in different sections and parts of the body (from the head to abdomen and posterior lorica), and not only in specific parts of the body where the potential symbiotic bacteria were found. In addition, the conclusions made by Bernhard et al.  appear contradictory, because they incubated the nematodes collected from the halocline and normoxic samples with CellTracker Green and concluded that these nematodes were alive since they showed positive reactions. If Bernhard et al.  can state that this is a proof for the viability of nematodes, it should be as well a proof for the viability of the Loricifera incubated with CellTracker Green , which also showed the absence of parasitic organisms attached to their body.
Evidence from molecular analyses
Extracting and sequencing RNA from living organisms can provide additional information on their ability to survive in anoxic conditions. Bernhard and co-workers  could not obtain any anoxic sediment sample below the brines of the DHAB (and thus any specimen of anoxic metazoans) to perform their molecular analyses.
Nor they were able to obtain rRNA sequences from any other taxon microscopically identified (mainly nematodes) both in normoxic and halocline sediments. All of the sequences of metazoans they found were affiliated to animal taxa belonging mainly to planktonic crustaceans.
The authors invoked four possible explanations: i) lack of primers’ specificity towards nematodes inhabiting the systems investigated; ii) an insufficient amount of sediment used for RNA extraction; iii) a nematode cuticle which does not protect the rRNA from degradation processes after nematode death and iv) the presence of signatures of pelagic copepods, which can have masked those of nematodes.
However, as far as the primer specificity is concerned, there is no reason to hypothesize that the primer pairs selected are not suitable for nematodes, since the region cover by that primers fall within the region we successfully amplified and sequenced from deep-sea nematodes . Therefore, the primers utilized should have worked at least for the nematodes that Bernhard et al.  found in their normoxic sediments.
The nucleic acid extraction procedure carried out directly on a few grams of sediments that the authors used (i.e., in situ lysis approach) is known to be appropriate for molecular analysis of unicellular eukaryotes, but not for meiofauna, especially when very low abundances are encountered [15, 16]. This is confirmed by the fact that independently of the samples analyzed, including those from normoxic conditions, the largest majority of the sequences found were affiliated to benthic fungi and protists , with only a very small percentage of sequences affiliated to metazoans (0.02-6.5 %).
The conclusion that no nematodes were found due to decomposition processes of their RNA occurring post-mortem contradicts what the authors reported based on viability and ultrastructural analyses. This appears even more evident from the analysis of normoxic samples, where the authors identified living nematodes as well as in all other deep-sea sediments worldwide.
The isolation of nematodes from the sediments and the subsequent nucleic acid extraction are indispensable to avoid the masking effects related to the presence within the total RNA pools of sequences belonging exclusively to non-target organisms (as reported by Bernhard et al. ).
From the evidence provided by Bernhard et al.  we conclude that their analyses were biased in several aspects and thus could not allow any evaluation or proof of the presence/absence of living metazoans in their samples.