Recently, we proposed a natural scenario of grand unified theories with anomalous U (1) A gauge symmetry, in which doublet-triplet splitting is realized in SO(10) unification using the Dimopoulos-Wilczek mechanism, and realistic quark and lepton mass matrices can be obtained in a simple way. The scenario has an additional noteworthy feature that the symmetry breaking scale and the mass spectrum of super heavy particles are determined essentially by anomalous U (1) A charges. Therefore, once all the anomalous U (1) A charges are fixed, the gauge coupling flows can be calculated. We examine several models in which gauge coupling unification is realized. Examining the conditions for the coupling unification, we show that when all the fields except those of the minimal SUSY standard model become super heavy, the unification scale generically becomes just below the usual GUT scale, Λ G ∼ 2 × 10 16 GeV, and the cutoff scale becomes around Λ G . Since the lower GUT scale leads to a shorter lifetime of nucleons, proton decay via the dimension-six operator, p → e + π 0 , should be observed in future experiments. On the other hand, a cutoff scale lower than the Planck scale may imply the existence of an extra dimension in which only gravity modes can propagate. * ) Downloaded from 598 N. Maekawa neutrino, to be unified into a single multiplet, which is important in investigating neutrino masses.Second, one of the most difficult problems is the "doublet-triplet (DT) splitting problem". Generally, a fine-tuning is required to obtain the light SU (2) L doublet Higgs multiplet of the weak scale while keeping the triplet Higgs sufficiently heavy to suppress dangerous proton decay. There have been several attempts to solve this problem. 10), 11) Among them, the Dimopoulos-Wilczek mechanism provides a promising way to realize DT splitting in the SO(10) SUSY GUT. 11) -14)Finally, there is a rather theoretical problem, which has not been emphasized in the literature. If we adopt an ajoint Higgs field A to break the GUT gauge group, the superpotential is generically given by W = ∞ n A n . In vacua, the GUT gauge group is generically broken to U (1) r , where r is the rank of the GUT gauge group. It is unnatural to obtain the standard gauge group SU (3) C ×SU (2) L ×U (1) Y from this superpotential, because the model has an infinite number of vacua other than the standard vacuum. The point of this problem is that the infinite number of terms A n lead to an infinite number of solutions to the F -flatness conditions. In order to avoid this problem, we have to restrict the number of terms. At least for SU (5) unification, we can impose renormalizability to avoid this problem. Then the superpotential becomes W = A 2 + A 3 , which naturally gives the standard gauge group below the GUT scale. However, for SO(10) or E 6 unification, this cannot be done, because A 3 is not allowed under the gauge symmetry. Moreover, in the context of the Wilsonian renormalization group, renormalizability is not a principle to be imposed, but a resulting feat...