I suggest you read the page following the supposedly cryptic notation:
The above Backus-Naur Form (BNF) is a convenient abbreviation of the following equivalent, more traditionally mathematical definition:
In said definition, it is stated that $M$ and $N$ are simply arbitrary lambda terms. (In a nutshell, it is an inductive definition.)
As to why the author prefers this notation, it is stated right after the definition:
Comparing the two equivalent definitions, we see that the Backus-Naur Form is
a convenient notation because: [...] the use of distinct meta-symbols for different syntactic classes ($x$, $y$, $z$
for variables and $M$, $N$ for terms) eliminates the need to explicitly quantify over
the sets [of variables and lambda terms].