Same Box, Different Contents: Viruses and Cells May Have Swapped Shells

A team at the University of Miami has identified small viruses whose protein shells resemble the chemical compartments inside bacteria and archaea, and argues the two have traded roles repeatedly over evolutionary time, in both directions.

The study, led by Abelardo Aguilar with senior author Antoni Luque, was published in Nature Communications on October 14, 2026 ^[1]. It offers a model built from sequences and structures, not a fossil caught in the act ^[1].

One shape, two jobs

The story starts with a protein shape called the HK97 fold. Cells and viruses both use it to build hollow shells ^[1].

In a virus, the shell is a capsid, a protein coat that protects the genetic material inside. In bacteria and archaea, the same fold builds encapsulins, compartments that enclose cargo proteins and carry out chemistry. The jobs include iron storage, sulfur metabolism and responses to oxidative stress ^[1][2].

Aguilar said the shells have different roles in cells and in viruses. They can enclose biochemical reactions, or they can store and protect a viral genome ^[2]. Same box, different contents.

An earlier proposal held that encapsulins descended from viruses of this type. But there was a puzzle. Known viral capsids are structurally more complex than encapsulins, which is awkward if the simpler thing came from the more elaborate one ^[1]. The authors went looking for smaller, simpler viruses that might fill the gap.

Hunting in environmental DNA

Many viruses cannot be grown in a laboratory. Luque said that sequencing gives researchers information about viruses whose hosts are unknown and which cannot be cultured ^[2].

So the team worked from data. They searched environmental metagenomes, which are pooled DNA sequences taken from soil, water and other natural samples, in public databases including IMG/M and MGnify. They looked for small viral genomes that encode HK97-fold proteins ^[1].

To find and characterize them, they used hidden Markov models, statistical profiles that recognize members of a protein family even when their sequences have drifted apart. They also used protein language models, software trained on large collections of protein sequences ^[1].

The search turned up viral shell proteins with molecular similarities to encapsulins. The genomes lacked portal and tail genes ^[1]. Those are the parts many viruses use to deliver their DNA into a host. Without them, these viruses are stripped-down things, closer in spirit to a bare shell.

What the analyses show

The team then ran phylogenetic analyses, which sort proteins into family trees by relatedness, along with structural comparisons. In the trees, viral and encapsulin proteins did not sort into two clean branches. They came out mixed together ^[1].

The structures added a second clue. Encapsulins resemble procapsids, the immature form of a viral shell before it is finished ^[1].

The authors read these observations as support for bidirectional evolution. In their account, HK97-fold viruses and encapsulins have exchanged roles more than once, and the procapsid stage may have made the switch possible ^[1]. A simpler, unfinished shell is an easier thing to repurpose.

The paper also proposes that HK97-fold shells began life as lower-symmetry biochemical compartments. It predicts that some viral capsids may still retain biochemical functions ^[1].

What this does and does not show

Nobody watched a virus become a compartment, or the reverse. The researchers worked from sequence and structural information in databases. They did not isolate the new viruses or observe the proposed transitions in a laboratory ^[1][2].

The model is an interpretation of patterns. The mixed groupings in the trees fit repeated exchange in both directions, but the paper does not claim to have found a single ancestor or a proven chain of descent ^[1].

Nor does it settle how viruses, or life, began. The hosts of the new viruses are not known, and no one has yet shown what the viruses do there.

That is a limit, not a flaw. Much of what is known about viral diversity comes from sequence data, because most viruses refuse to grow on a dish.

Why it matters

Origin-of-life researchers care about this corner of biology for a plain reason. Life needs two things at once. It needs a way to keep information safe, and it needs a place to run chemistry. A compartment can concentrate and protect molecules, and heredity depends on preserving information ^[1].

The HK97 fold appears to have served both purposes in different lineages, and possibly swapped between them. That gives researchers a concrete protein family in which to study how a container for genes and a container for chemistry might share a history.

There is also a practical side. Aguilar said the work raises the possibility that viral capsids could facilitate biochemical reactions ^[2]. If some do, viral-like shells might be used or engineered to contain biochemical activity, a research question the paper points toward for biotechnology ^[1]. That is a possibility the model predicts. It has not been demonstrated.

The researchers said non-infective virus-like particles, shells that carry no working infection, deserve more attention and further study ^[2]. For decades such particles were easy to dismiss as debris or dead ends. This work suggests they may be doing, or may once have done, something more interesting.

The study was supported by the Gordon and Betty Moore Foundation and the U.S. National Science Foundation, according to the University of Miami ^[2].

What comes next

The immediate next step, according to the researchers, is to investigate what virus-like particles do and whether capsids can facilitate biochemical reactions ^[2]. Isolating one of the new viruses, finding its host, or building and testing a shell in the laboratory would turn a model into a test. Until someone does, the idea of two-way traffic between viruses and cellular compartments rests on patterns in sequences, and the experiments that could confirm it or sink it have yet to be run.