We describe a new NMR method for monitoring the individual hydrogen exchange rates of the hydroxyl groups of tyrosine (Tyr) residues in proteins. The method utilizes (2S,3R)-[beta(2),epsilon(1,2)-(2)H(3);0,alpha,beta,zeta-(13)C(4);(15)N]-Tyr, zeta-SAIL Tyr, to detect and assign the (13)C(zeta) signals of Tyr rings efficiently, either by indirect (1)H-detection through 7-8 Hz (1)H(delta)-(13)C(zeta) spin couplings or by direct (13)C(zeta) observation. A comparison of the (13)C(zeta) chemical shifts of three Tyr residues of an 18.2 kDa protein, EPPIb, dissolved in H(2)O and D(2)O, revealed that all three (13)C(zeta) signals in D(2)O appeared at approximately 0.13 ppm ( approximately 20 Hz at 150.9 MHz) higher than those in H(2)O. In a H(2)O/D(2)O (1:1) mixture, however, one of the three signals for (13)C(zeta) appeared as a single peak at the averaged chemical shifts, and the other two appeared as double peaks at exactly the same chemical shifts in H(2)O and D(2)O, in 50 mM phosphate buffer (pH 6.6) at 40 degrees C. These three peaks were assigned to Tyr-36, Tyr-120, and Tyr-30, from the lower to higher chemical shifts, respectively. The results indicate that the hydroxyl proton of Tyr-120 exchanges faster than a few milliseconds, whereas those of Tyr-30 and Tyr-36 exchange more slowly. The exchange rate of the Tyr-30 hydroxyl proton, k(ex), under these conditions was determined by (13)C NMR exchange spectroscopy (EXSY) to be 9.2 +/- 1.1 s(-1). The Tyr-36 hydroxyl proton, however, exchanges too slowly to be determined by EXSY. These profound differences among the hydroxyl proton exchange rates are closely related to their relative solvent accessibility and the hydrogen bonds associated with the Tyr hydroxyl groups in proteins.