Optimal SELECT(s) in stored procedures

[Marek Erban]

Look at two following cases.
What use of SELECT is more optimal and faster (and why)?
(all is in stored procedure, all conditions use right indexes, SELECT from A always returns one row, SELECT from B or C can return more rows - not in this "INTO" case but in case of using CURSORs)

Case 1:
SELECT F1 INTO L1 FROM A WHERE K1 = 123;
SELECT F2 INTO L2 FROM B WHERE K2 = L1;
or
SELECT F2 INTO L2 FROM A, B WHERE (A.K1 = 123) AND (B.K2 = A.F1);
?

Case 2:
SELECT F1 INTO L1 FROM A WHERE K1 = 123;
SELECT F2 INTO L2 FROM B WHERE K2 = L1;
SELECT F3 INTO L3 FROM C WHERE K3 = L1;
or
SELECT F2 INTO L2 FROM A, B WHERE (A.K1 = 123) AND (B.K2 = A.F1);
SELECT F3 INTO L3 FROM A, C WHERE (A.K1 = 123) AND (B.K3 = A.F1);
?