The objective of this study was primarily to determine the structure and biochemical characteristics of the NmeGp1Sd in order to gain a deeper insight into the evolution of metazoan Nme proteins and their functions.
Sponge Group I Nme genes are intron-rich and these introns are relatively short. The same has been found for introns in several other sponge genes [42, 43] and recently in A. queenslandica genome where median intron size is 80 bp . The fourth intron (Figure 1) is likely the most ancient because it is also found in a choanoflagellate Group I Nme homolog. We conclude that the ancestral metazoan Group I Nme gene was intron-rich and probably had all four introns that are still present in most extant basal metazoan homologs. The ancestral gene structure is also well preserved in vertebrate homologs with three out of four introns present. D. melanogaster has only one of the ancestral introns and C. elegans lost all ancestral introns and gained two new ones which likely reflect accelerated evolution in these lineages. Analysis of sponge NmeGp1 promoters showed that some of the motifs essential for human promoter activity are also present in sponges. We did not find these motifs in the corresponding choanoflagellate promoter under the same search parameters, which indicates a possible change in Nme1 regulation in the metazoan lineage.
The phylogenetic tree (Additional file 2) is generally not well resolved or supported, as in comparable earlier studies [1, 2, 24]. Nonetheless, it does provide some new evidence on Group I Nme evolution. Group I Nme gene duplicates in N. vectensis had been proposed to indicate that the split into Nme1 and Nme2 might have occurred very early in metazoan evolution . Another recent study  showed that duplications of Group I Nme genes occurred independently more than once in invertebrates, besides well described duplications and diversifications of Group I Nme to Nme1, 2, 3 and 4 in vertebrate lineage, and that duplication in N. vectensis is also an independent event unrelated to the origin of Nme1 and Nme2. Our results confirm these findings and additionally demonstrate independent duplications in basal metazoans, even within the same lineages (cnidarians and calcareous sponges).
The metastasis suppressor genes Nme are highly conserved in opisthokonts. This implies that they have important roles in basic cellular functions. A number of studies revealed several potential biological roles of Nme genes/proteins, but most are not confirmed in vivo. The expression of the D. discoideum Nme homologues gip17 and guk7.2 is modulated during the developmental phases of cell growth or aggregation due to starvation . In Drosophila, Nme/Awd is required for the proper differentiation of many tissues including the brain, eye, and female reproductive system . Nme protein accumulation is coincident with the functional differentiation of multiple epithelial tissues in the developing mouse . These and many other studies demonstrate that Nme proteins have a critical role in differentiation and development - processes that involve cell migration, a prerequisite for metastasis formation .
The sponge protein NmeGp1Sd was compared primarily with human Nme1 (and Nme2) for several reasons: i) The NmeGp1Sd is highly similar to vertebrate Nme1/Nme2 in primary structure, ii) human Nme1/Nme2 proteins are the most studied and well characterized Group I Nme members iii) Nme1 is linked to metastasis suppression, but the suppression mechanism is not well understood. We expected that NmeGp1Sd analysis and the discovery of its biochemical features would shed some more light into the mechanisms connecting various Nme1 functions and its antimetastatic activity. A recent study by Domazet-Lošo and Tautz  shows that, contrary to what might be intuitively expected, only a subset of cancer genes appeared simultaneously with multicellularity in animals and that the new genetic processes and gene functions that emerged at the basis of the metazoan lineage are major innovations which enabled complex interactions between metazoan cells. Demosponges have only one Group I Nme protein (whereas amniotes have four with potentially different functions), which is probably the ancestral condition for metazoans.
Our results show that the sponge NmeGp1Sd possesses some biochemical characteristics typical for human Nme1 and some typical for human Nme2 protein. Recombinant NmeGp1Sd appears to be predominantly in the hexameric form like human Nme1 protein which was confirmed by gel filtration (Figure 2A and 2B). The sponge NmeGp1Sd and human Nme1 enzymes also have similar levels of kinase activity. Human Nme1 did not display an ability to bind sscDNA, however, human Nme2 can bind sscDNA and showed the same DNA band retardation effect as the sponge homolog. Accordingly we decided to check the possibility that the sponge NmeGp1Sd is functionally more similar to Nme2 than to Nme1. Unlike Nme1, Nme2 is able to cleave c-myc NHE sequence . Therefore, we tested whether the sponge NmeGp1Sd also has DNA topoisomerase-like activity and found out that it is not able to cleave c-myc NHE sequence. We hypothesize that this was the ancestral condition of the Nme Group I protein before duplications and functional diversifications within the Group I Nme. Sponge NmeGp1 may have retained the multiple functions, while these functions have been partitioned between different vertebrate Group I Nme proteins. Recent studies showed that sponge genomes are comparable to genomes of "higher" complex animals, including vertebrates, in terms of gene number and functional repertoire [32, 33]. It was suggested that complex animals differ from their early simple relatives mainly in the more complex regulation of similar sets of genes. From that point of view, it could be speculated that the ability of Nme2 to cleave c-myc is a relatively recent addition to the regulatory networks.
Based on this evidence we then asked the following: What is the biological function of NmeGp1Sd in its native environment - the sponge cell? As it can suppress the migration of CAL 27 cells e.g. replaces the human Nme1 it could thus be involved in migration and/or, possibly, adhesion of sponge cells. It has previously been found that sponges possess genes/proteins for many advanced physiological processes usually linked to more complex metazoan phyla [31, 47]. Migration of cells in complex "higher" animals is limited to specific processes such as embryonic development and immune response and is closely controlled. Cells are anchored in the developed tissues they belong to and do not migrate if not stimulated (with the exception of cells undergoing malignant transformation). The anchorage and migration of cells depends on a vast number of molecules such as cadherines, integrins, fibronectin and collagen. Although sponges do not possess true tissues and organs, they do possess simple varieties of the mentioned proteins [48, 49]. Furthermore, they have complex pelagobenthic life cycles which include development to a larval phase and metamorphosis to sedentary adult form . Cell migration is present in sponges in three different processes: regeneration, larval development and the movement of amoeboid cells through the adult mesohyl; the space between the external and internal cell layers which is composed of galectin, collagen, fibronectin-like molecules and dermatoponin . Although the mesohyl is not a homogenous, organized structure, it resembles a primitive extracellular matrix (ECM). Therefore, both the movement of amoeboid cells through the mesohyl in sponges and the movement of cells through ECM in vertebrates may originate from the same ancient precursor process present in the metazoan last common ancestor. From an evolutionary point of view, tumors likely developed together with the evolution of tissues and organs. It is not yet defined which biochemical functions of Nme are crucial for its metastasis suppressor activity. It is, however, to be expected that its main biological function (besides the well defined NDPK function), would not be to suppress metastasis in "higher" Metazoa but to control normal physiological processes (adhesion, migration) which, if left without control, could lead to metastasis formation. As sponges do not possess tissues and organs they are probably also incapable of forming tumors. The demonstration of the fact that the sponge NmeGp1Sd can replace human Nme1 in human tumor cells encourages us to propose that the NmeGp1 (at least Nme1) function responsible for regulating adhesion/migration of cells was established in the metazoan ancestor well before the Cambrian explosion i.e. before the appearance of diverse groups of multicellular Metazoa. Since tumors likely started developing in parallel with the development of true tissues and organs, this ancient function of Nme1 was then adapted to act in a new highly significant role - the suppression of metastasis.