There were a number of good comments. I respond to a few here. One pointed out that the interstate highway system was largely built in the 1950s and 1960s, rather than continuously. As a result, it presents a "bloc obsolescence" problem necessitating heavy expenditures on maintenance and rebuilding--costs exacerbated by the unanticipated wear and tear resulting from the vast increase in usage of the system. This makes the problem of financing the necessary expenditures an urgent one.
Another comment points out perceptively that imposing tolls on users of the interstate highways could create a negative externality by deflecting users to non-toll roads, thus increasing congestion and wear and tear on those roads. That is an argument for tolling all roads--to which two objections are raised in comments. One is that tolls will be prohibitive on roads that are lightly traveled, assuming that their light traffic doesn't reduce maintenance costs to trivial levels (as it would not). What is true is that tolls will be kept down in order to avoid deflecting users of these roads to more congested roads, adding to the congestion on them. The second objection is that it is infeasible to impose tolls on busy commuter roads such as the Minneapolis bridge that collapsed, because it would slow down traffic too much. But this is a short-term objection. Technology is rapidly coming on line, and at rapidly falling cost, to enable tolls to be charged (and varied with time of day to minimize congestion) without need for toll booths, but instead through a system of sensors in the pavement and cameras overhead.