Although it had been suggested from in silico analysis that the nuclear transport system of trypanosomes had little in common with that of opisthokonts,29
a combination of proteomics, structural studies, and comparative genomics has revealed that the NPC and associated karyopherin transport factors are an ancient feature, and that the overall configuration of this organelle and the transport mechanism and diversity of pathways were established in the last eukaryotic common ancestor (termed LECA). By contrast the lamina and associated macromolecules appear to be more divergent.
The basic NPC architecture, in terms of division into a structural scaffold, consisting predominantly of protocoatomer-related proteins, together with a gating system comprising about ten disordered FG repeat-containing nucleoporins, is conserved across the eukaryotes.30,31
Further, the karyopherin family is likewise ancient and, perhaps more remarkably, based on an ability to reconstruct the majority of karyopherin clades as being widely represented across the eukaryotes, the level of diversity and specificity in nucleocytoplasmic transport is both well conserved across the eukaryotic lineage and a feature of the LECA.32
What these studies imply is that the plural mechanisms of nucleocytoplasmic transport, which are intimately wrapped up with mRNA processing and translocation into the cytoplasm and hence gene expression, are extremely ancient. Minor differences in the compositions between mammalian, yeast, plant and protozoan NPCs may signal some diversification/adaptation of mechanism that reflects the differing transcriptional systems in place in these various lineages, but at a level that is quite detailed and one that is not immediately obvious or accessible from a study of the proteins comprising the NPC and in silico approaches alone. In part, this is due to our rather limited grasp of the precise functions of many individual nucleoporins themselves, and it may well be that some functionality is redundant, such that the presence of a specific subunit from one organism has little or no clear impact on functionality; this is certainly an area where more investigation is warranted. Remarkably, the conservation within NPC functionality even extends to moonlighting nucleoporins in trypanosomes. We have evidence for a nucleoporin, TbNUP92, that appears functionally similar to the mitotic spindle-associated TPR nuclear basket proteins, despite an absence of obvious sequence-based homology and also an FG-repeat nucleoporin, TbNUP53b, potentially involved in transcriptional control and hence analogous to the NUP98 proto-oncogene from Drosophila, albeit again with no obvious primary structural similarity (Holden, J, MPR and MCF, unpublished data, and refs. 9, 33
). It is unclear if these functional similarities in the absence of evidence for sequence relatedness are a result of convergent evolution or a failure in search algorithms to detect such similarity.
In sharp contrast to the NPC/karyopherin system, there is good evidence for significant divergence in the nuclear lamina. While a nuclear lamina has been clearly detectable by ultrastructural analysis in various eukaryotes, and might be expected, in the sense of a basic requirement for structural support of the nucleus (though yeast does manage without apparent support), a knowledge of the molecular basis was restricted to metazoa, where the lamin proteins have been well characterized.35,36
Despite extensive searches of the genomes of yeasts and many other organisms, no evidence for lamin orthologs had been forthcoming, leading to the suggestion that lamins represent a metazoan innovation.
Two studies published in 2012 have, however, overturned this paradigm. First, through a serendipitous route, a clear lamin ortholog was identified for Dictyostelium discoideum
, which even includes a C-terminal prenylation signal, heptad repeat signatures and a potential CDK-1 phosphorylation site consensus;37,38
orthologs are also present in additional Amoebozoa genomes. The significance of this discovery is far reaching. With lamins restricted to Metazoa, lineage-specific innovation is the obvious explanation, but the new data suggest that in fact lamins arose several hundred million years earlier, at least as early as the origin of the unikont lineage, which includes the Opisthokonta and Amoebozoa. Clearly then, the absence from fungi can now be interpreted as a secondary loss, and has profound implications for our views on gene expression mechanisms and lifestyles of rather early eukaryotes close to the radiation of the modern supergroups. This study also provides a salient lesson as well—identification of divergent proteins and detecting an ortholog is often challenging; the tools available have especial difficulty with coiled coil proteins. This topic is discussed further in a recent Extra View article in these pages to which the interested reader is referred.37
In the remaining supergroups, the bikonts, there is so far only one example of a lamina-associated protein for which convincing functional evidence is available. This is the trypanosome NUP-1 protein, which actually has structural similarities with and fulfills many of the reported functions of lamins. NUP-1 forms a fenestrated, apparently fibrillar, network on the inner face of the nucleus, and is implicated in functions including maintaining the structural integrity of the nucleus, positioning of NPCs and telomeres, and mediating telomeric silencing. Interactome analysis also indicates a physical association between the NPC and NUP-1 (Obado S, MCF, MPR and Chait BT, unpublished). In trypanosomes, the role in telomere positioning is particularly critical as monoallelic VSG expression is dependent on heterochomatinization of subtelomeric VSG expression sites, and indeed there is good evidence that this process is disrupted by knockdown of NUP-1.39
While the parallels between NUP-1 and unikont lamins are striking, there are some critical differences. First, there is no obvious sequence relatedness and the amino acid repeats in T. brucei are 144 residues, significantly larger than the heptad repeats of the lamins. Second, NUP-1 is a huge protein of 407 kDa, compared with ~60 kDa for lamins, which means that NUP-1 can reach across a substantial proportion of the trypanosome nuclear volume, and hence may indicate a rather distinct structural organization to the lamins (). However, this may in part be specific to trypanosomes as the orthologs in Leishmania are considerably smaller, albeit still predicted to contain over 2,000 residues. Hence, while there is no evidence for any relationship or common ancestry between NUP-1 and the lamins, we remain agnostic on this issue.
Figure 2. Relative sizes of metazoan and trypanosomatid nuclei and lamina monomers. Blue ovals indicate the relative sizes of a fibroblast and trypanosome nucleus, at 10 μm and 1.5 μm in diameter. The molecular weight of the unikont (more ...)
Also, as mitosis is closed in trypanosomes, the NUP-1 network does not break down as does the mammalian lamin network, but rather it is maintained throughout mitosis, where it may participate in chromosomal segregation (Holden J and MCF, unpublished data). The absence of a massive re-assortment of the nuclear envelope during mitosis is quite significant in terms of maintaining epigenetic marks and the mechanisms by which specific DNA sequences are incorporated into heterochromatin, as in metazoa several lamin-dependent mechanisms that deliver loci to heterochromatin occur as a component of mitosis and reassembly of the nuclear envelope at late anaphase, which clearly cannot be the case for trypanosomes.40
Hence it is unclear just how many of the proteins that control entry, exit and maintenance of heterochromatin are shared across the eukaryotes. However, it seems to us a key point that peripheral organization of developmentally-regulated chromatin (kept inactivated as heterochromatin until needed) at a nuclear lamina composed of coiled-coil proteins is a shared feature of representative unikont and bikont organisms.
A final aspect here are the LINC complex proteins.41
LINC complexes are comprised of trimers of SUN and KASH domain proteins; these trimers then interact within the lumenal space of the nuclear envelope.42,43
These trans-membrane domain proteins are responsible for providing a physical tether for the lamina to the nuclear envelope, as well as a connection between the nuclear lamina and the cytoskeleton. Significantly, SUN and/or KASH domain proteins with the same architecture as those present at the nuclear envelope of metazoa are found in all eukaryotic lineages (including Excavata lineages such as the diplomonads and heterolobosids), with the exception of the trypanosomatids (). By contrast, SUN-like proteins, have different architecture to classical SUN proteins () and their function is likely distinct from the classical SUN proteins of the LINC complex,44
and are found in trypanosomatids as well as elsewhere. This is perhaps consistent with the restriction of the NUP-1/NUP-2 lamina to Trypanosoma and perhaps argues against convergent evolution from a common lamin/SUN/KASH system. If this is the result of loss from a LECA possessing a lamin-based system, or reflects differential evolution of the lamina in the trypanosomatid and unikont lineages remains to be determined. It is also possible that, as trypanosomes likely branched early from the eukaryotic lineage,11
that evolution of the lamin/SUN/KASH system postdates their speciation event. However, the discovery of the D. discoideum
lamin is a further salient lesson in relying too heavily on in silico data alone, if any more are really needed at this juncture, and resolution of this issue will require deeper experimental probing. In summary, the nuclear periphery of the trypanosomes appears to be a mixture of highly conserved elements, together with other elements that may be lineage specific, or are at the very least, extremely divergent ().
Figure 3. Maximum likelihood phylogenetic tree of the SUN domains. The values supporting the separation into two main subfamilies are: approximate likelihood ratio test calculated with PhyML 3.0/non-parametric bootstrap calculated in RAxML 7.2.6/posterior (more ...)