In the phage universe, there are many examples of so-called satellite phages (or satellite phage-like elements) whose propagation depends on helper phages. Outside the bacterial cell these satellites look like phages, but nucleic acids packaged inside their capsids do not have all the genes that would be required for a successful lytic infection. What do those genomes encode instead? Many possibilities, sometimes a self-replicating plasmid, sometimes genes encoding virulence factors as Merry described in this post from 2013.
Perhaps the most studied and best understood of the satellite phages is P4 from E. coli whose propagation inside a phage capsid depends on phage P2 (or other similar phages). The life cycle of the P4 genetic element is fascinating as it displays multiple and very clever means of propagation (Fig. 1). Outside the cell, P4 looks and acts by all criteria as a phage. It has an icosahedral capsid and a tail, and it can inject its DNA into its E. coli host. However, its capsid and tail proteins are all encoded by the helper phages of the P2 family. The P4 genome itself completely lacks the genes encoding these structural proteins. When P4 infects an E. coli P2 lysogen, it may enter either the lysogenic or the lytic pathway. For the lytic pathway P4 absolutely depends on the capsid, tail, and lysis functions of P2. Interestingly, since the P4 genome is smaller than that of P2, it makes a smaller capsid. To accomplish this, the P4 genome contains a gene (sid) whose product directs the capsid proteins to assemble into a smaller icosahedron. When P4 infects an E. coli that is not a P2 lysogen, the P4 genome is either integrated into the chromosome or establishes itself as an autonomously replicating multicopy plasmid. Glancing at the life cycle diagram, one might guess that this type of helper exploitation would require that the helper phage be able to establish a lysogen state. It is easy to imagine that if the helper phage were virulent, then the helper and satellite would have to infect the host cell pretty much at the same time. The chances of that happening out in nature should be vanishingly low, right? Well... evolution seems to have found a way to overcome the odds.
A recent study discovered and characterized a new satellite-helper system, named "Flayer," that infects diverse Streptomyces species. MindFlayer is the helper and MiniFlayer the satellite. When plated they yield clear plaques, already indicative that they might be exclusively lytic phages. Moreover, their genomes do not appear to have any of the genes characteristic of temperate phages (i.e. able to lysogenize). MiniFlayer's genome does contain genes coding for capsid and tail proteins; what makes it a satellite seems to be the lack genes coding for enzymes involved in its own DNA replication. If both MindFlayer and MiniFlayer can only propagate through a lytic life cycle, how does MiniFlayer manage to infect cells that are being infected at precisely the same time? The electron micrograph shows the characteristic appearance of these phages (Fig. 2). Amazingly, the majority (80%) of the MindFlayer particles have a MiniFlayer attached to their tail. MiniFlayer has a very short tail that likely includes a fiber protein that specifically binds to MindFlayer's tail. There also appears to be some sort of affinity for binding between the capsid proteins. These protein-protein interactions lead to the formation of "bonded pairs" of helper and satellite, thus greatly increasing the chances of simultaneous adsorption to the host cell. And since a picture tells one thousand words, I'll let you ponder on the further physiological, ecological and, yes, even evolutionary implications of this remarkable image.