A short protein segment helps direct SOSEKI to plant cell edges

Finding the targeting signal

A protein’s location can matter as much as its activity. Cell polarity—the uneven distribution of cellular components—helps organisms coordinate cell division, differentiation and development. Yet the mechanisms that position polar proteins in plant cells remain incompletely understood.

In a bioRxiv preprint posted on November 2, 2026, Andriy Volkov and colleagues investigate SOSEKI1, or SOK1, in Arabidopsis thaliana. The study has not been peer reviewed. It combines engineered proteins, fluorescence microscopy, computational predictions and evolutionary analysis to separate the features that direct SOK1 to the membrane from those that concentrate it at a particular edge.

SOSEKI proteins often accumulate where two cell faces meet. In Arabidopsis root cortex cells, SOK1 occupies the inner apical edge, facing the root’s center. Other members of the family show different patterns, making them useful for investigating how related proteins acquire distinct destinations.

Direction and concentration are separable

SOK1 contains a DIX domain that allows protein molecules to assemble into polymers. Removing this domain, or disrupting its assembly function, left SOK1 unevenly distributed at the membrane but spread over a broader area rather than tightly focused at an edge.

Replacing the native DIX domain with unrelated polymerization domains restored focused targeting. The authors interpret this as evidence that polymerization concentrates the protein, rather than supplying the instructions for which membrane region to occupy.

Simply linking several targeting domains together did not reproduce that effect. Constructs containing two, four or six tandem copies remained diffusely polar. The result leaves two possibilities: focused localization requires assemblies larger than six units, or polymerization contributes something beyond increasing the number of membrane-binding segments.

To narrow the targeting instructions, the researchers systematically replaced sections of SOK1’s previously identified polarity domain and tested progressively shorter fragments. These fragments were attached to a human DIX polymerization domain and a fluorescent marker, allowing their locations to be observed in Arabidopsis roots.

The authors report that a 25-amino-acid segment, corresponding to SOK1 residues 213–237, was sufficient for consistent polar targeting in this engineered setting. Crucially, the finding concerns a fragment paired with a polymerization domain—not an isolated peptide independently recreating the full localization pattern.

A proposed two-sided membrane interaction

Computational tools predicted an amphipathic helix within the targeting region: a spiral with water-attracting, positively charged residues concentrated on one face and hydrophobic residues on the other. Such an arrangement could help a protein associate with a membrane.

Mutations supported the importance of these chemical properties. Replacing two hydrophobic leucines abolished membrane association. Several changes affecting the charged face also disrupted targeting, although one lysine substitution produced an intermediate pattern, with protein both at the polar membrane and in the cytosol.

The helix remains a structural prediction rather than an experimentally determined structure. The authors propose that it first brings SOK1 near the membrane, after which protein acyltransferase enzymes attach a lipid chain to a conserved cysteine. Polymerization would then concentrate the anchored proteins at an edge. The lipid-attachment mechanism draws on earlier work described in the preprint.

That sequence is a model, not a fully established pathway. The authors explicitly leave open whether the helix contacts membrane lipids directly or instead interacts with an acyltransferase, and whether removal of lipid modifications also helps maintain polarity.

Conserved destinations, evolving signals

The evolutionary analysis covered 1,086 SOSEKI sequences from 199 plant species. It supports an origin from a single ancestral land plant SOSEKI gene, followed by duplications and diversification.

When targeting domains from different species were tested in Arabidopsis roots, most directed their engineered fusion proteins toward the inner membrane. Exceptions included mostly cytosolic constructs, a maize domain with membrane-wide distribution and edge enrichment, and a Marchantia domain targeting the opposite side from SOK1.

Reciprocal expression in Marchantia also showed opposite orientations: its native SOSEKI pointed toward the growing notch, whereas Arabidopsis SOK1 pointed away. This supports a genuine difference in how the two proteins respond to polarity cues across these experimental contexts.

The authors infer that inner-directed targeting was broadly ancestral, with additional patterns emerging after gene duplication. However, these are reconstructions from modern sequences and localization experiments. What an equivalent “inner” direction meant in the earliest land plants remains unresolved, as do the biological functions of the diversified targeting patterns.