Auxiliary Solvents and Buffer Effects in Octanol-Buffer Partitioning: Standardization of Data Using the ADW Workstation Technical Note ADW-01 ANALIZA, Inc. March 28, 2000

Summary This Technical Note discusses the effects of introducing a compound dissolved in auxiliary solvents into an octanol-buffer partitioning system, and the effects of buffer properties on the distribution coefficient. For drug profiling during discovery it is important to maintain data standardization to account for such effects in a consistent manner. The approach undertaken with the ADW workstation is described. Auxiliary Solvent Effects Screening a wide range of compounds with varying physicochemical properties poses unique challenges to implementation of an automated workstation such as ADW. All must be analyzed automatically by the same workstation in a uniform manner. These difficulties are typically handled on a case by case basis during manual shake-flask operations. Certain compounds may be readily soluble in the aqueous phase, and some dissolve well in the octanol phase. When using the ADW system in a screening mode it is clearly desirable not to make a priori decisions regarding which phase to use to dissolve a given compound. Standardization of sample preparation provides a logical solution, but not without a consequence. Any auxiliary solvent that is introduced into a partitioning system will, by definition, change the properties of that system. This effect must be quantified and taken into account when measuring the true distribution coefficient. Since most combinatorial libraries are prepared and stored in DMSO, this solvent is an attractive choice. Over 50 organic compounds dissolved in DMSO were partitioned in an octanol-buffer system at pH of 7.4, and the logD values observed were plotted versus the logD values for the same compounds partitioned in the same octanol-buffer system without addition of DMSO. The volume of the DMSO solution was 20% of total volume of the octanol-buffer system. The data obtained are shown in Figure 1. The observed correlation could be described as: LogDOctanol-DMSO-Buffer, pH 7.4 = a + b* LogDOctanol-Buffer, pH 7.4 N = 53; r2 = 0.9753

 All rights reserved

(1)

1

2

LogD

Octanol-DMSO-Buffer

1

0

-1

-2

-3

-4 -4

-3

-2

-1

0

1

2

3

LogDOctanol-Buffer

where a = -0.347 ± 0.023, b = 0.877 ± 0.020, N is the number of compounds tested, and r2 is the correlation coefficient. These values are valid only for 20% v/v DMSO in the octanol-buffer system using buffer composed of 0.15 M NaCl in 0.01 M sodium phosphate buffer, pH 7.4. They are likely to change with the DMSO quantity in a system as well as with changes in the buffer composition and/or pH. A similar correlation was observed when compounds dissolved in methanol were introduced as 20% v/v of an octanol-buffer system with the regression coefficients being a = 0.096± 0.030, and b = 0.734± 0.029 (N = 43; r2 = 0.9413). The strong correlation shown above enables transformation of the data obtained with DMSO or methanol as auxiliary solvents (under specific buffer and quantity conditions) into the corresponding values which would have been obtained without auxiliary solvents. The price paid in terms of the potential deviations from these relationships for any individual compound should be considered together with the value of achieving standardized and automated sample preparation without the need for individual attention to each compound, and with the ability to directly screen combinatorial libraries already available in DMSO. The ADW software has an internal user-selected capability (via a simple checkbox) to automatically consider samples that are dissolved either in the system phases, DMSO, or in methanol. This choice implies standardization of the buffer components and DMSO or methanol quantities that are introduced into the system. The ADW workstation automatically places the proper amount of sample after the user selects the auxiliary solvent during the assay setup dialog. The standardization of buffers by ANALIZA ensures accurate transformation of the data to the true value of LogD. Buffer Effects LogP data, representing perhaps the most studied structural parameters to date, have been measured using a variety of techniques and using many buffers. It is not surprising that an agreement among the data obtained using different methods for the  All rights reserved

2

same compound can not be guaranteed. Since the data obtained using the ADW workstation represent direct partitioning (similar to shake-flask), a method-related variability does not exist for data obtained using this workstation. However, buffer properties are subtler and require further studies before standardization could be accomplished. Buffer effects (aside from pH) include all aspects of the aqueous solvent that may affect the partitioning behavior of the solute. Of particular importance is the ionic strength of the buffer. To study buffer effects, partitioning experiments with the ADW system were performed, and the results were compared with those obtained in manual shake-flask experiments and with those reported in the literature. While the agreement between the data obtained with ADW system and those obtained with manual shake-flask partitioning was excellent (usually within 1-2%), results for some of the compounds tested did not agree with literature data. Analysis of the literature data compiled by Hansch et al [1] has shown that the logD values for many compounds indeed vary dramatically. For example, logD values for atenolol at pH 7.4 vary from –2.00 to –0.11 [1, p.127], and for chlorpromazine at the same pH from 1.90 up to 3.50 [1, p.149]. Typical experimental data obtained with ADW system in parallel with manual shake-flask partitioning for propranolol in octanol-buffer systems with varied buffer composition are shown in the following Table. An extensive number of examples are available from ANALIZA. Buffer Effect on Propranolol Partitioning in Octanol-Buffer Systems Buffer Composition

0.01 M Universal Buffer1 0.01 M K-Phosphate Buffer 0.01 M Na-Phosphate Buffer + 0.15 M NaCl 0.01 M Universal Buffer1 + 0.15 M NaCl 0.10 M K-Phosphate Buffer 0.01 M Universal Buffer1 0.10 M K-Phosphate Buffer 0.07 M Na-Bicarbonate Buffer

pH

LogD

7.4 7.4 7.4 7.4 7.4 11.0 11.0 11.0

0.015 ± 0.009 0.010 ± 0.007 0.632 ± 0.009 0.697 ± 0.011 0.981 ± 0.008 1.147 ± 0.004 1.530 ± 0.002 2.440 ± 0.010

1

– Universal buffer is composed of a mixture of acetic, phosphoric, and boric acids with NaOH. The ionic composition of octanol-buffer systems used for partitioning of organic compounds clearly has a significant impact on the data obtained for the same compound at the same pH. Unfortunately only pH-value, but not the buffer composition, is commonly reported in the literature. Recent studies at ANALIZA have indicated that there are certain general trends regarding the effects of the ionic composition of the aqueous phase on the partitioning of organic compounds in octanol-buffer systems. A physicochemical model for these effects is currently under development and will be published in the future. The recognition of buffer effects shifts the discussion from directly comparing with literature data to analyzing the best practice to follow for constructing a self-consistent

 All rights reserved

3

internal compound database. ANALIZA has chosen to standardize the buffers used in its pre-filled partitioning plates as a solution to data self-consistency. Thus, for drug profiling and screening using ANALIZA’s standardized partitioning system plates (supplied at 6 pH level), comparing data obtained from different compounds could be performed in a self consistent manner that is free from buffer property variations. ANALIZA’s choice of a universal buffer allows the use of the same ionic composition over the entire range of pH from 2.0 up to 10.0 (and upward, if required). To compensate for the variation in the ionic strength of the buffer with pH, 0.01 M universal buffer is used with an addition of 0.15 M NaCl. The 1-octanol/0.01 M universal buffer0.15M NaCl systems at pH values of 2.0, 4.5, 6.6, 7.4, 8.5, and 10.0 are provided by ANALIZA in sealed pre-filled 96-deep well plates under strict quality control program.

Reference: 1. Hansch, C. and Leo, A., Exploring QSAR: Fundamentals and Applications in Chemistry and Biology, American Chemical Society, Washington, DC, 1995

 All rights reserved

4