Horowitz's proposal goes as follows. Suppose an early organism needs organic substance A as a precursor for a vital substance. Substance A eventually comes into short supply because many similar organisms are competing for it. Suppose, further, that A is being produced spontaneously from the following reaction, occurring outside the cell, B + C → A. Any cell that can take up B and C from the environment and produce A now has a selective advantage over all the other cells that depend on an outside source of A.
The uptake of B and C may require, first, a changed membrane protein. This can occur through mutation. Proteins already in the cell might catalyze the reaction B + C → A. Further, mutations could yield a more efficient catalyst. The cell with such new capabilities would now predominate in those parts of the early seas in which survival was possible. And then, in time, B or C or both would come into short supply. What then? A situation like that regarding the formation of A would occur. The cell or cells able to form B from its precursors and C from its precursors would be at a selective advantage. And this situation could repeat itself many times throughout the early history of the earth, which is now reliably estimated as being 4.6 × 109 years old.
The overall result of this early evolution would be the emergence of biosynthetic pathways. Such pathways would be like those known today, in which reactions are catalyzed by highly specific enzymes formed under the control of genes. Also, of necessity, there would be a general direction in this evolution. The biosynthetic pathways would belengthening in the direction of simpler and simpler precursors. Eventually, organisms would probably become so biosynthetically sophisticated as to require only such inorganic precursors as carbon dioxide, water, ammonia, sulfur, and phosphorus or their compounds or both. They might, also, by this time have evolved means for capturing energy from sunlight for biosynthesis. In short, nutritionally, they would have evolved from being chemoheterotrophs to being photoautotrophs.