Let A,A′ be two rings, p:A→A′ a homomorphism, S a multiplicative subset of A, S′=ρ(S) and ρ′:S−1A→S′−1A′ the homomorphism corresponding to p (no. 1, Proposition 2).
(i) For every A′-module M′ there exists a unique S−1A-isomorphism
j:S−1ρ∗(M′)→ρ∗′(S′−1M′) such that j(m′/s)=m′/ρ(s) for all m′∈M′,s∈S.
(ii) For every A-module M, there exists a unique isomorphism
j′:(S−1M)⊗S−1A(S′−1A′)→S′−1(M⊗AA′) of S′−1A′*-modules*
such that j′((m/s)⊗(a′/s′))=(m⊗a′)/(ρ(s)s′).
(i) If we consider S′−1M′ as an A-module by means of the composite homomorphism iM′S′∘p, the homotheties induced by the elements of S are bijective, hence there exists a unique homomorphism j with the stated property (no. 2, Proposition 3). As ρ(S)=S′, j is surjective; moreover, if m′∈M′, s∈S, m′/ρ(s)=0, there exists t′∈S′ such that t′m′=0; as there exists t∈S such that ρ(t)=t′,t′.m′=0 in ρ∗(M′), hence m′/s=0 in S−1ρ∗(M′).
(ii) As (S−1M)⊗S−1A(S′−1A′)=(M⊗AS−1A)⊗S−1A(S′−1A′) and
S′−1(M⊗AA′)=(M⊗AA′)⊗A′(S′−1A′), the existence of
j′ follows from the associativity of tensor products.
Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.
Nothing.