The case
Almost every aphid species has 60 to 80 bacteriocytes holding Buchnera, a close relative of E. coli, transmitted maternally to eggs and embryos. The abstract of the genome paper (LIT-1596) says the mutualism “is so obligate that neither can reproduce independently”. The 640,681-bp chromosome is a seventh of E. coli’s. It keeps genes for the amino acids the host cannot make and lacks those for the ones it can, as well as genes for cell-surface components, regulators and defence. The authors read Buchnera as viable only in the bacteriocyte.
Its genes evolve faster than free-living relatives’, with unusually many replacement substitutions and A+T-rich changes. Moran (LIT-1573) explains this by Muller’s ratchet in small asexual populations, with mutational bias, so much of the reduction is not selected.
In the record, NOTE-114 reports Godfrey-Smith’s placement of aphid–Buchnera as both an organism and a Darwinian individual. The entry it reads, Wilson and Barker, misplaces Buchnera in a human gut.
What it can show
- Dependence through degradation. The symbiont’s dependence was produced largely by drift, by the abstracts’ account. That separates the route to dependence from any selection for the collective (QUESTION-277).
- Constituent autonomy lost while the whole thrives. Buchnera cannot live outside the host. The partnership is one of the most successful in insects. It illustrates CLAIM-060: the whole’s success does not preserve a constituent’s viability on its own.
- High cohesion by transmission. Vertical transmission couples the partners’ reproduction. The abstracts report the two genomes as complementary, and report no gene transfer to the host. Whether any occurred is for a reading to check. Compare the mealybug (CASE-171) and Paulinella (CASE-174).
The case is documented in works the record has not read. What it shows rests on their abstracts until they are read.
Note of 2026-10-27 (later): a host protein inside Buchnera, and what the model says of the host
- Host-encoded protein in the symbiont. McCutcheon et al. (LIT-1571, read in NOTE-1213) report that host-encoded proteins have been identified by immunofluorescence “in the cytoplasm of aphid and mealybug endosymbionts”. The study for the aphid looked at a single protein, and the mechanism of import is unknown. The paper does not say that the protein comes from a gene transferred from Buchnera: it speaks of horizontally transferred bacterial genes on host genomes in general. So the case’s statement, that its sources report no gene transfer to the host, still holds of those sources. But the host already reaches into the symbiont, which bears on control (C) in QUESTION-277. Buchnera APS’s genome (641 kb) is well above the 400 kb and 200 kb thresholds of THEORY-1049.
- The host’s side, on Athreya et al. LIT-1570 (read in NOTE-1211) lists aphid–Buchnera as mutually obligate and vertically transmitted. Its model predicts that the faster-evolving partner, the symbiont, becomes obligate first, and that the host’s dependence follows only if partners die inside the collective more than they are born there (THEORY-1051). That turns the case into a measurement: the demography of Buchnera within its host. Nothing in the sources reports it.
Note of 2026-10-27 (batch 2): the genome paper read in full
LIT-1596 has been read in full (NOTE-1236, THEORY-1103). So has the review LIT-1560 (NOTE-1241). The case’s first paragraph mostly stands. Six corrections:
- “Neither can reproduce independently” is cited, not shown. It is the paper’s opening sentence, cited to earlier work. The genome shows the symbiont’s side only.
- Almost all, not all. The non-essential amino-acid genes are “almost completely” missing, and cysteine is still made, from reduced sulphate. The host’s glutamate and aspartate are precursors of Buchnera’s essential amino acids, so the pathways interlock as well as complement.
- The drift attribution has no source here. The genome paper does not mention drift, Muller’s ratchet or population size. “Dependence … produced largely by drift” rests on LIT-1573 alone, which is Unfetched.
- On gene transfer to the host, the genome paper says nothing. It does not study the host genome. The host-encoded protein reported in Buchnera (the note above) is RlpA4. Its gene came to the aphid from another bacterium, not from Buchnera. Proteomics found no host protein reliably enriched in Buchnera (LIT-1560). The paper itself proposes that the host supplies phospholipid, or the enzymes to make it, “like mitochondria do”. That is untested.
- Loss of self-regulation, not host control. Buchnera lacks nearly all its own transcriptional regulators and keeps the genes they once regulated. That bears on C as lost self-regulation, not as host control.
- Envelope loss above the threshold. At 641 kb, Buchnera APS has already lost phospholipid and LPS synthesis. That is above THEORY-1049’s 400 kb boundary for complete envelope loss, and consistent with its remark that erosion begins in larger genomes.
Campbell et al. (LIT-1590, Table 1) cite 886 to 8,223 Buchnera cells per aphid egg across three aphid species, from earlier work.