I am in New York City today and tomorrow, having a bit of a vacation with my wife. Well, it’s more like she needs some much needed rest and I am keeping her company. I do stay a bit too long at the hotel doing some grant-related work, much to her chagrin. Earlier today we went to the David Letterman show and I think if some of you want to find me in the audience (perhaps my students want to make fun of my giggling face), it might be too late as the show is probably on right now. It was a lot of fun. What fascinated me the most was the attention to detail by those running the telecast. It is quite an astounding display of efficiency: everyone knows exactly what she/he needs to do at a given time.
It is also fun to think about amides, especially about some of the more heretical considerations, if you will. My lab has been tackling the physical organic fundamentals of amide bond formation and extracted some much-needed new value out of this linchpin of synthesis (you can see my post of February 27). It is remarkable how many tricks we can teach this old dog. I will talk at length about some additional underappreciated aspects we currently find rather exciting, but this will happen after we publish our next piece of this saga. As I was watching David Letterman, I kept comparing his well-greased show business machine to some of the mechanisms of chemistry. I refer to the ones where everything happens for a good reason and under “spatiotemporal” control (in other words, at the right place and at the right time). It occurred to me today that I might not have a good answer to the collapse of the tetrahedral intermediate, which is central to understanding amide bond formation. Take a look at the graphic below. Let’s not worry too much about pH and where our protons are supposed to be. These are finer details. For the time being, I am just going to illustrate the collapse using anionic oxygen center that is part of the well-known tetrahedral intermediate (it can also be OH, you know). Here comes the $64,000 question… When we discuss amide bond formation, we teach it in terms of the “blue arrows” you see here. For the life of me, I cannot think of a study (either theoretical or experimental) that has seriously considered an alternative shown in red. This really bugs me and those of you who are of the opinion that this “red” idea is pure heresy, please direct me to the primary literature. Dear undergraduate students: please do not get all flustered and confused, I am certainly hopeful that there is a solid study illuminating the overall process so that we can sleep well tonight. I just can’t think of it.
amphoteros 
Andrei, you got me so excited that we had all missed something fundamental in 1st year organic! But thinking about it some more, isn’t this just a semantic difference? In my mind, it would also be acceptable to simply displace the leaving group, with that being the *only* arrow push. Formally, this gives you a carbocation, but it’s just the third resonance structure of the product you already drew. So the “accepted,” the “heresy,” and this third mechansism all lead to the same species, just different representations of it.
The way we push arrows in the classical mechanism implies a contribution of the oxygen lone pairs into the nacent pi-system, which is true. But it’s also true that the nitrogen’s lone pairs are involved as well. And both these features are represented by the resonance structures. After all, isn’t the whole point of resonance structures to supplement our simple model of bonding when the Lewis dot formalism breaks down? Every year we hammer into the undergraduates that point about “resonance structures are not different molecules” but even in our own work, it can be surprisingly easy to overlook the ramifications of that statement!
Well, that’s just it. This was a tricky question that exposed the inadequacy of Lewis structures in explaining reactivity. But that’s what we have in our disposal when we draw structures, don’t we? The correct way is to always have a “holistic” view of a given molecule and consider the HOMO/LUMO interaction. Thus, what I wrote is not heretical when we consider the molecule as a whole. Oh those cute lines we use to represent bonds… They mean SO LITTLE!