I believe it wasn't known to him or he didn't understand it enough to trust it.  PubKey recovery was certainly known prior to 2009.  There are a number of other "nuts and bolts" decisions (oversights?) which make me thing Satoshi's background was not cryptography.  He probably had exposure to and experience working with cryptography but wasn't a cryptographer.

In addition to PubKey recovery here are just a few other "quirks":
* Transaction Malleability (Bitcoin Specific). It is cumbersome to fix to fix today but with a different txn structure it would be impossible.
* Signature Malleability (ECDSA).  ECDSA can have multiple signatures for the same digest the solution is to either not make the signature part of the txn hash or limit Bitcoin to a single form.
* DER encoding serves no purpose.  Even if OpenSSL was used the DER bits could be striped.
* Lack of Compressed Keys.  Optimally the only valid key would be a compressed key and thus not only is key size reduced but there is only one format to consider.
* Unusual choice of double hashing.  Normally done to prevent length extension attacks which don't exist in Bitcoin.

This isn't to say Satoshi wasn't a genius, the magic in Bitcoin isn't the cryptographic primitives used.  It is in the way he elegantly used existing systems (digital signatures & hashing algorithms) to create a timestamp and consensus finding system that is very simple and yet very difficult to attack.