Measurement of unbound fractions of mycophenolic acid and its metabolites may prove useful in explaining the complicated pharmacokinetic and pharmacodynamic behavior of this drug as well as in therapeutic drug monitoring. We developed a reliable, accurate, and sensitive liquid chromatography-tandem mass spectrometric method for the simultaneous quantification of mycophenolic acid (MPA), MPA glucuronide (MPAG), and MPA acyl-glucuronide (AcMPAG), total or unbound, in plasma, urine, and tissue extract. This method uses a single internal standard, carboxy-butoxy ether of mycophenolic acid (MPAC), and involves a simple sample preparation step. Aliquots of plasma, urine, or dissolved tissue extract (100 microL) or plasma ultrafiltrate for free analytes (50 microL) are treated with acetonitrile/formic acid mixture (99.5/0.5 v/v) followed by centrifugation and dilution with water. The prepared samples are then injected onto an extraction column (Eclipse XDB-C18 12.5 x 4.1 mm; Agilent Technologies, Palo Alto, CA) and washed with mobile phase composed of acetonitrile/water/formic acid (10/89.5/0.5 v/v/v) at a flow rate of 2.8 mL/min. A switching valve is activated 1 minute after sample injection. The analytes are eluted onto the analytical column (Eclipse XDB-C18 150 x 4.1 mm; Agilent Technologies) with a gradient of 0.5% aqueous formic acid, methanol, acetonitrile, and water. We used a tandem mass spectrometer with electrospray ion source, in which the tandem mass spectroscopy transitions were (m/z): 338-->207 for MPA, 438-->303 for MPAC, and 514-->303 for MPAG and AcMPAG. The dynamic ranges (lower limit of quantitation and upper limit of quantitation) were as follows: 0.05 to 30 mg/L for total MPA and 1 to 300 microg/L for free MPA; 0.5 to 300 mg/L of total MPAG and 0.2 to 60 mg/L for free MPAG; and 0.025 to 15 mg/L of total AcMPAG and 1 to 60 microg/L for free AcMPAG. The precision at lower limit of quantitation was in the range of 8.0% to 11.9% for all three total analytes and 13.8 to 18.7% for the free analytes. Accuracy at lower limit of quantitation was in the range of 100% to 105% for total and 97% to 99% for free analytes. Between-day precision of quality control samples was 4.0% to 6.3% for human plasma spiked with total analytes and 4.5% to 14.4% for spiked plasma ultrafiltrate for free analytes. Mean absolute recovery ranged from 98.5% to 101.7% for MPA (both total and free), from 78.1% to 103.4% for MPAG and from 91.5% to 110.4% for AcMPAG. No significant ion suppression was found under these conditions for any of the analytes. Carryover effect was found to be at a maximum level of 0.02%. This method was successfully applied to analyze over 11,000 samples for total analytes, and over 8000 samples for free analytes in plasma, and has been in operation for nearly 3 years without loss of performance.