That’s the title I originally proposed for an opinion article just published in American Journal of Botany.  Although microbes can provide a wide range of benefits to plants, from nitrogen fixation by rhizobia bacteria in root nodules (my main research focus) to defense against pathogens, it’s been bothering me for years that the only known benefit from many so-called Plant-Growth-Promoting Rhizobacteria (PGPR) is that they “provide hormones.”  The metabolic cost to plants of hormone synthesis is negligible (in contrast to the cost of making roots, for example), so production of plant-hormone mimics by microbes would only be expected to benefit plants if the microbe-manipulated plant phenotype is superior to the phenotype the plant would get from its own hormones alone.  That would seem to require :

• The microbial cells have more information than the plant does about how much auxin or whatever the plant needs.  Otherwise, manipulation of the plant by microbes is as likely to hurt the plant as help it.  Hence “poison” in the title.
• Microbes that make a plant-benefiting hormone somehow have greater fitness (survival and reproduction) than their own mutants nearby in the soil that save the cost of making the hormone but retain their parent’s ability to use root exudates, etc.  Otherwise, the no-hormone mutants would quickly displace their parents.

These both seem unlikely to me, in general.

But plant phenotypes shaped by natural selection in past environments could perhaps be nonoptimal in a consistent direction in a new environment (“the conservatory”) in which case hormonal manipulation in the opposite direction could be beneficial.  I may do a more-comprehensive review of the PGPR literature to see whether convincing reports of plants benefiting from manipulation by microbes all come from environments very different from where the plants evolved.

Also, natural selection has often favored individual-plant competitiveness over plant-community efficiency, a major theme of my book.  Could manipulation of crops by microbes be an easier way to make crop plants more “cooperative”, relative to plant breeding or genetically engineering of crops?  For example, could microbial hormones make leaves more vertical, or limit wasteful water use on hot afternoons?

Maybe, but how do we ensure that crops are exposed mainly to the “right” microbes?  We can inoculate seeds with selected strains, but as soon as the roots get out in the soil, they’ll be exposed to a huge diversity of microbial cells, each with its own evolution-shaped agenda.  My article therefore suggested that, to see if there’s any real potential for consistent benefits to crops from manipulative microbes, we might want to start with rhizobia in root nodules, which have the following benefits:

• They’re more abundant than (most?) soil microbes and have access to enough resources that making hormones (which many of them can do) is something they can easily afford to do, metabolically.
• Because of their interactions with crops that farmers control, rhizobia are potentially easier to control (in the sense of being able to switch to strains that make more or less of some plant hormone), relative to microbes out in the soil.

Walid Sadok and I have submitted a proposal to USDA to explore this approach with soybean.  It’s just occurred to me that an intercropped legume might even release enough hormones from its root nodules to manipulate a nonlegume companion crop.