We have examined the reaction of Escherichia coli tryptophan indole-lyase with fluoro, chloro, methyl and hydroxytryptophans using steady-state kinetics, rapid-scanning and single wavelength stopped-flow spectrophotometry, and rapid chemical quench methods. All of the 16 tryptophan derivatives examined are substrates for alpha, beta-elimination catalyzed by tryptophan indole-lyase. The steady-state kinetic parameter, kcat/Km, did not show a consistent trend with the steric bulk of the substituent, but Km increased for larger substituents. Rapid-scanning stopped-flow spectra show that all tryptophan analogues undergo covalent reaction with the pyridoxal-5'-phosphate cofactor to give equilibrating mixtures of external aldimine and quinonoid intermediates, but the relative amounts of each intermediate are strongly dependent on the nature and position of the substituent. The dissociation constants for external aldimine formation, Kd, obtained from single-wavelength stopped-flow experiments decreased for most substituted tryptophans, which suggests that part of the binding energy is derived from hydrophobic interactions between the enzyme and the indole ring of tryptophan. In contrast, the rate constants of quinonoid intermediate formation and reprotonation and of indole elimination were quite variable, depending on the position and the nature of the substituent. Overall, 6-substituted tryptophans have the most consistent reactivity, which indicates that there may be space in the enzyme active site near the 6-position. There is a good linear correlation between log (kcat/Km) and log (kf/Kd) (apparent second order rate constant for quinonoid intermediate formation), with a slope of 0.66. This suggests that quinonoid intermediate formation contributes only about 66% of the activation energy for the reaction, and thus a later step in the reaction must be partially rate-limiting. Rapid chemical quench experiments demonstrate a 'burst' of indole in the reaction of L-tryptophan under single turnover conditions, confirming that a step subsequent to the elimination is partially rate-determining. In contrast, 5-methyl-L-tryptophan does not exhibit a significant 'burst', suggesting that 5-methylindole elimination is nearly completely rate-determining. These results support the proposed mechanism and demonstrate that there are significant effects of aryl substituents on the distribution of covalent intermediates and on the rate-determining step in the alpha, beta-elimination reaction catalyzed by E. coli tryptophan indole-lyase.