Source
Sørensen, Stiller, Kröninger & Nowack (2025), The FEBS Journal 292(12): 2992–3013, LIT-1560, read in full in NOTE-1241 from the PubMed Central copy. The rhizobia section’s last paragraph, and the fourth paragraph of “Conclusions and future directions”.
What was actually shown
Nothing was tested. The scenario is assembled from systems at different stages:
- Effectors without replacement. The cereal weevil’s peptide ColA enters its 20-million-year-old Sodalis symbiont and regulates its growth: “relatively recently captured endosymbionts can already be targeted by host effector proteins”. Legume NCR peptides control bacteroid division and differentiation.
- Single replacements. RlpA4 in Buchnera, MurF in Moranella, two transaminases in Nardonella complete a function the symbiont lost.
- Few targeted proteins. Six in the diatom’s spheroid bodies and seven in Angomonas’s symbiont, some at the symbiont’s division site.
- Hundreds. The chromatophore and the nitroplast, each with a conserved targeting signal.
In the authors’ words: “a eukaryotic host cell starts relatively early on to target effector proteins into a bacterial endosymbiont … Once a nucleus‐encoded protein with functional redundancy to the endosymbiont attains import capacity, deleterious mutations in the endosymbiont copy can get fixed and—if the resulting process cannot be replaced by metabolite import—the endosymbiont becomes dependent on the imported host protein. … As soon as a general import mechanism with a specific targeting signal evolves, it likely spreads quickly”.
On rhizobia: “we would not consider rhizobia as genetically integrated organelles”. They are not vertically transmitted, keep a soil life, and have no function replaced. Effector manipulation is “an effective alternative to the genetic integration that is always linked to a permanent host/endosymbiont association with vertical transmission”.
What could have failed: the systems could have shown no gradient, for instance hundreds of imported proteins in a horizontally acquired symbiont, or replaced functions with no effector stage anywhere.
What this does not say
- Not that any lineage went through these steps. The stages are different organisms, ordered by the authors. A static gradient across systems fits several histories.
- Not that effectors cause the later steps. Effectors and replacing proteins are different proteins. Whether effector targeting builds the route later used for replacement is not argued.
- Not that vertical transmission is sufficient. Buchnera is vertically transmitted for over 100 million years and shows one imported protein.
- Not the reason dependence is selected. The scenario has mutations in the symbiont copy “get fixed” once the host copy covers them. That is consistent with drift under relaxed selection. The review names no selective advantage for the host or the symbiont.
Connections
- QUESTION-277, the order of the dimensions. Translated into the question’s letters (the translation is mine): host control (C, as effectors) comes first; dependence (D) follows when a redundant imported protein lets the symbiont’s gene go; the distribution of functions across genomes (G) shifts as import grows. Reproductive cohesion (R, vertical transmission) is a condition for the sequence to end in an organelle. Rhizobia are high C, no R and no genetic integration. This gives the question a definite order to test, and it is the reverse of THEORY-1049’s insect picture, where the symbiont’s losses come first and force host supply.
- THEORY-1050 and THEORY-1051. Athreya et al. model D and R with no host control. This scenario makes host control the first step. The two are about different stages and do not conflict.
- LIT-1605. Frank’s model of host control of transmission, unread, is the formal counterpart of R as a precondition.
- CASE-167. The nitroplast’s host can lose UCYN-A when given nitrogen, so the host’s dependence stays conditional even at the hundreds stage.
Note of 2026-10-30: promoted, and the wider thesis filed beside it
The twenty-second status review promoted this reading under ADR-065: it reports the scenario Sørensen and colleagues propose, and NOTE-1241 read the paper in full. Whether endosymbionts become organelles by these steps is THEORY-1211’s question, Proposed, which keeps this entry’s old promote_when.