Partitioning of the Amino Acids Phenylalanine and Tryptophanylphenylalanine

The pH-metric technique has a practical low log P limit of detection that is approximately -2. For hydrophilic molecules with log P of -2, the octanol-containing and the aqueous titration curves differ in the buffer region by about 0.01 pH units. This is about the level of reproducibility of a good research-grade pH electrode. One would not expect the technique to work with amino acids; most amino acids are very hydrophilic and are charged in the interval pH 2-12. Even when the molecule is uncharged in the overall sense, it still exists in the zwitterionic form, with a plus-charged center and a minus-charged center. Since we were able to determine ion-pair partitioning in many other substances, with log P often in the range between 0 and -2 and were able to demonstrate the dependence of the phenomenon on background salt concentration, we thought that it would be productive to try measuring log P of some amino acids. What we immediately saw was met with some scepticism: although we were sometimes unable to characterize the zwitterion log P, we often saw convincing indication of the monocation partitioning at low pH and the monoanion partitioning at high pH, producing an "inverted" parabola log D curve, with the minimum at the isoelectric point. This was different from what was observed with the shake-flask method, employing pH buffers. Our solutions contained 0.15 m NaCl or KC1, and it appeared reasonable to us that ion pair partitioning was taking place.

Fig. 16a shows the lipophilicity profile of phenylalanine in 0.027 and 0.161m NaCl. Also shown in the plot are measured values of log D using radiolabeled phenylalanine (data kindly provided by Dr Keith Chamberlain, 1994, Rothamsted Experimental Station, UK). The agreement is superb over much of the pH range. The measured log P of the zwitterion was -1.4. We were also able to see evidence for the partitioning of the amino acid cation in low-pH solution. Fig. 16a shows slight salt dependence, most prominent at pH 2.

Fig. 16b shows the lipophilicity profile of tryptophanylphenylalanine (Trp-Phe), obtained from pH-metric data in 0.15 M KC1. Lipophilicity profiles of amino acids determined by the shake-flask method, using phosphate buffers, often show a parabolic shape with a maximum log D at the isoelectric point (pH 5.24). The plot in Fig. 16b is controversial because it reveals as smaller logD at the isoelectric point than at lower pH, where the cationic species predominates. It appears that the ion-pair partitioning is more prevalent than zwitterion partitioning.

Precedence for the "inverted" behavior exists in the literature, though it appears not to have received the attention we think it deserves. Fig. 16c displays the lipophilicity

~1 ' 3 ' 5 ' 7 ' 9 pH

Trp — Phe Lipophilicity Profile

log D

Trp — Phe Lipophilicity Profile

11 13

Figure 16. Lipophilicity profiles of amino acids, (a) Phenylalanine at two different ionic strengths and with radiolabelled shake-flask measured log D values (squares). See text, (b) Tryptophanylphenylalanine. See text (c) Tryptophanylphenylalanine using data of Akamatsu et al. [35]. See text.

plot ofTrp-Phe constructed from the data of Akamatsu et al. [35]. Using conventional buffers they saw a maximum log D at the isoelectric point. However, when they attempted to measure log D in low pH solutions in the presence of chloride and other anions more lipophilic than phosphate, they observed log D values higher than that found at the isoelectric point. Fig. 16c shows such points at pH 1 and 13. It would seem that salt has a key influence on the lipophilicity of charged species such as amino acids. Our measured lipophilicity profiles of hydrophilic amino acids (e. g., glycine, aspartic acid, tyrosine) in the 0.15 m NaCl or KC1 medium often look like cases (j), (k), or (1) in Fig. 12.

0 0

Post a comment