Homologues of 6-methyl-7,8-dihydropterin (6-Me-7,8-PH2) and 6-methyl-5,6,7,8-tetrahydropterin (6-Me-PH4), expanded in the pyrazine ring, were synthesized to determine the effect of increased strain on the chemical and enzymatic properties of the pyrimidodiazepine series. 2-Amino-4-keto-6-methyl-7,8-dihydro-3H,9H-pyrimido[4,5-b] [1,4]diazepine (6-Me-7,8-PDH2) was found to be more unstable in neutral solution than 6-Me-7,8-PH2. Its decomposition appears to proceed by hydrolytic ring opening of the 5,6-imine bond, followed by autooxidation. 6-Me-7,8-PDH2 can be reduced, either chemically or by dihydrofolate reductase (Km = 0.16 mM), to the 5,6,7,8-tetrahydro form (6-Me-PDH4). This can be oxidized with halogen to quinoid dihydropyrimidodiazepine (quinoid 6-Me-PDH2), which is a substrate for dihydropteridine reductase (Km = 33 microM). Whereas quinoid 6-methyldihydropterin was found to tautomerize to 6-Me-7,8-PH2 in 95% yield in 0.1 M tris(hydroxymethyl)aminomethane hydrochloride (Tris-HCl), pH 7.4, quinoid 6-Me-PDH2 gives only 53% 6-Me-7,8-PDH2, the remainder decomposing via an initial opening of the diazepine ring. Additional evidence for the extra strain in the pyrimidodiazepine system is the cyclization of quinoid 6-N-(2'-aminopropyl)divicine to quinoid 6-Me-PH2 in 57% yield in 0.1 M Tris-HCl, pH 7.4. By comparison, no quinoid 6-Me-PDH2 is formed from the homologue quinoid 6-N-(3'-aminobutyl)divicine. A small (2%) yield of 6-Me-PDH4 is found if the unstable C4a-carbinolamine intermediate is trapped by enzymatic dehydration and reduction. Although phenylalanine hydroxylase utilizes 6-Me-PDH4 (Km = 0.15 mM), the maximum velocity of tyrosine production is 20 times slower than that with 6-Me-PH4, indicating that a ring opening reaction is not a rate-limiting step in the hydroxylase pathway. Further, the maximum velocities of 2,5,6-triamino-4(3H)-pyrimidinone, 2,6-diamino-5-(methylamino)-4(3H)-pyrimidinone, and 2,6-diamino-5-(benzylamino)-4(3H)-pyrimidinone span a 35-fold range. These cofactors would theoretically form the same oxide of quinoid divicine if oxygen activation involves a carbonyl oxide intermediate. Thus, the limiting step is also not transfer of oxygen from this hypothetical intermediate to the phenylalanine substrate.