In his 2002 paper ‘Plants compared to animals: the broadest comparative study of development’(Meyerowitz, 2002; pp. 1482–1485), Elliot Meyerowitz concluded that as regards the molecular basis of pattern formation and cell–cell signaling, the two fundamental systems underlying development, there is little homology between plants and animals. Similar processes in the two lineages are usually controlled by nonhomologous genes. For example, the master regulatory genes responsible for establishing segmental identity during embryonic development of animals, the Hox homeobox genes, have no evolutionary relationship to master regulators acting in plant development/patterning, the MADS box genes. Similarly, the receptor tyrosine kinases, like Sevenless or Gurken, or the Ras proteins, that are critical components of cell signaling in animals, have no homologous counterparts in plant signaling pathways. This strongly suggests independent evolution of the molecular systems responsible for the assembly of phenotypic forms in plants and animals. However, it appears that beneath this diversification, there is a common and universal molecular tool-kit that orchestrates the translation of both pattern-control and signaling-pathway inputs into selective gene-expression outputs. This is chromatin, which Meyerowitz (2002) acknowledged is practically identical in plants and animals. This extremely high level of conservation is attested by the striking similarity between pea and cow core histones, the employment of DNA and histone modifications and the same organization of SWI/SNF-type chromatin remodeling machinery, and identical chromatin-level regulatory strategies, like the use of Polycomb-class proteins to maintain the pattern of homeotic gene expression. This raises the question, is chromatin an old and finished functional module that can serve with equal efficiency the needs of subsequent evolutionary innovations? The answer is yes and no. Chromatin is certainly ancient and fundamental to gene regulation involved in basic processes of eukaryotic life, but it is by no means a finished, unchangeable module, like, say, the H3–H4 histone tetramer in the nucleosome. On the contrary, it is highly plastic and subject to ongoing evolution. The consequence of its position at the ultimate site of action of gene regulatory systems is that changes in chromatin components are very often lethal or highly deleterious. However, if fixed by natural selection, these components have the potential to reshape or modulate processes underlying basic life strategies. In this issue of New Phytologist, Pedersen et al. (pp. 577–589) provide an example, describing the properties and occurrence throughout the cell cycle of a newly identified class of plant-specific chromatin proteins termed 3 × HMG (high mobility group)-box proteins.
‘Is such a competition involving H1 and 3 × HMG-box proteins taking place during mitosis and meiosis? If so, what purpose could it serve?’
The 3 × HMG-box proteins belong to the HMG-type proteins, ubiquitous eukaryotic nuclear nonhistone proteins with a broad range of functions, all of which are linked to DNA-related activities (Bianchi & Agresti, 2005). The HMG superfamily is subdivided into several groups, each with a characteristic functional sequence motif (Fig. 1). One of them, HMGB (previously known as HMG-1/-2), is named after the ‘HMG-box’ motif, a characteristic L-shaped domain composed of c. 80 amino acids, with strong affinity for the DNA minor groove (Stros, 2010). HMG boxes bind DNA in a sequence-independent manner to induce bending, and this property is the basis of HMGB architectural functions in chromatin. The occurrence of the HMG-box motif is by no means restricted to canonical HMG proteins. This motif is found in a large number of so-called HMG-box proteins, which mostly function as transcription factors. In fact, the canonical HMGB group can be considered a specialized subfamily of HMG-box proteins. The occurrence of HMGB in all lineages of eukaryotes is an indication of a very ancient evolutionary history. It is therefore not surprising that both their structure and function have diverged considerably between animals and plants. The challenge, as already mentioned, is to correlate this diversification with major differences in form and life strategies between these two groups of multicellular eukaryotes.
Within the plant kingdom, the HMG-box family is additionally diversified (Grasser et al., 2007). It consists of four structurally distinct subgroups: canonical HMGB proteins; structure-specific recognition protein 1 (SSRP1); proteins that possess an AT-rich interaction domain (ARID) in addition to the HMG-box; and proteins containing three HMG-box domains (3 × HMG-box proteins) (Stros et al., 2007). While the first three subgroups have been functionally characterized, albeit to different extents, the recently identified fourth subgroup has remained a mystery until now. The results reported by Pedersen et al. provide some insight as to where we should expect these proteins to function.
Pedersen et al. present evidence that the 3 × HMG-box proteins are bona fide architectural factors. While individual HMG-boxes had relatively low affinity for DNA, when arranged as a three-member array in a single molecule, they formed a ligand that bound specific DNA structures with high affinity. Employing an assay used to demonstrate the increased efficiency of ligation of short DNA fragments in the presence of proteins with the potential to facilitate ring closure, the authors showed that the same was true in the case of 3 × HMG-box proteins nonsequence-specific DNA bending activity. These properties of the 3 × HMG-box proteins suggest that they are a variation on a theme already known from studies of canonical HMGB proteins. HMGB proteins were found to be involved in many important DNA transactions, such as the maintenance of genome integrity, DNA replication, transposition, repair and recombination. This notwithstanding, their direct interaction with nucleosomes leads to substantial weakening of DNA binding to core histones and increases the accessibility of DNA to various remodeling and transcription factors.
Did some of these functions undergo modification or refinement with the evolution of 3 × HMG-box proteins, and if so, why was this important for plant cells? A partial answer to these questions may be inferred from the intriguing results of Pedersen et al. concerning the intracellular localization of 3 × HMG-box proteins. Surprisingly, they found that, unlike the nuclear localized Arabidopsis HMGB1, the 3 × HMG-proteins occurred exclusively in the cytosol. This was true in both tobacco protoplasts used for transient transformation and in Arabidopsis plants stably transformed with plasmids expressing GFP-fusion proteins. It was in this second experimental system that the authors noticed selected cells in which the analyzed proteins displayed chromatin localization. More detailed microscopic analyses revealed that the 3 × HMG-box proteins were specifically associated with mitotic and meiotic chromosomes. It thus seems that the cytosol served as a reservoir of 3 × HMG-box proteins until the breakdown of the nuclear membrane, after which they became associated with condensed chromosomes. The possibility of a mitotic/meiotic function for the 3 × HMG-box proteins was strengthened by the unusual pattern of their expression. Not only were they expressed exclusively in proliferating tissues, but their expression during the cell cycle was restricted to the mitotic phase. In order to speculate on the meaning of the mitotic chromosome localization and strange pattern of expression, one must take into account the behavior of chromatin proteins. The widespread use of photobleaching techniques, like FRAP (fluorescence recovery after photobleaching), to study the intracellular mobility of proteins in living cells, has challenged the earlier concept of these proteins being stably bound in chromatin. Such experiments have demonstrated constant exchange between chromatin-bound and free pools of proteins that differ considerably in their mean chromatin residence time (Phair et al., 2004). It now appears that chromatin has a highly dynamic, fluid-like structure and that its functional properties are critically dependent on factors that determine the extent of its fluidity, such as specific variants of histone and nonhistone proteins, and various post-translational modifications (Catez et al., 2004). HMG proteins, particularly of the HMGB subfamily, were shown by FRAP to be in a dynamic competition with histone H1 (another abundant and conserved chromatin architectural element) for binding to the same chromatin sites (Catez & Hock, 2009). In living cells the HMGB proteins compete with H1 and weaken its binding to chromatin (Fig. 2). Even though H1 is 10 times more abundant and binds DNA more strongly than HMGs, competition between these proteins can effectively modulate the local structure of chromatin. Is such a competition involving H1 and 3 × HMG-box proteins taking place during mitosis and meiosis? If so, what purpose could it serve? To look for possible answers, we must once again take into account the variability and peculiarities of seemingly conserved and universal biological phenomena. The land plants differ from most other eukaryotes in that they do not form and use microtubule organizing centers (MTOCs) during chromosome separation in mitosis – their spindle poles lack centrosomes (Murata et al., 2007). Surprisingly, it was found that in plant cells, histone H1 mediates microtubule organizing activity and can induce microtubule nucleation and continuous plus end association (Hotta et al., 2007). Could the competition with 3 × HMG-box proteins serve to relax binding of H1 on the surface of mitotic and meiotic chromosomes, to enhance its function in microtubule nucleation? This example is just one illustration of the possible functional links that could be postulated for this unusual class of plant HMGB-type proteins characterized by Pedersen et al.