The Half Decent Pharmaceutical Chemistry Blog

Nowadays, organic chemists normally cross the boundaries of traditional synthesis and play with stuff biologists thought to be their own, personal territory. This phenomenon has had a terrific impact: for instance, the conidia Curvularia Lunata plays a key role in the industrial preparation of cortisol acetate. It has made easier to yield the drug, skipping many steps, so, as a result, cortisol acetate has become cheaper, as its synthesis has been shortened.

During our half-decent lab course we, too, had a go at mixing synthesis and enzymes. Our aim was to prove how easily you can perform an enantioselective reduction of a beta-keto-ester (methyl acetoacetate) using normal brewer's yeast.
The most important advantage of the "biological" way lies in its simplicity: to yield this very product (with an asymmetric synthesis) would be incredibly more complicated, whereas the use of a yeast dramatically reduces the problem.

We weren't interested in the final yield very much: it was the specific rotation what we wanted to determine, above all.

In a 250 mL conical flask, with a ground-glass neck, we prepared the following solution: 64 g of sucrose, 0.4 of Na₂HPO₄ and 100 mL of water.
Because we will introduce a living being in this piece of glassware, we must create an environment where our cells could live and enzymes normally work. So, we set the temperature to approximately 35°C.

The flask was placed on the stirrer and we waited until all sucrose was dissolved. Then, we added 14 g of commercial brewer's yeast: I mean, it came from the supermarket just in front of the department, not from, say, a Chinese supplier.

We stirred the mixture again, but this time, following the instructions we had been given, we stirred rather vigorously for at least 15 minutes.  Actually, it took a little bit more than that to get a decent, homogeneous suspension.

Finally, we added the last, missing ingredient, the reagent, as 2 mL of methyl acetoacetate were poured in the flask.

Before leaving the lab, we assembled the reaction apparatus: the conical flask and a reflux valve (thanks to their ground glass junctions)  were joined together. At the other end of the valve, we put a rather short silicon tube, which provided the link to a Pasteur pipette.
This was partly immersed in a test tube full of glycerin. Obviously, we fixed this part of the apparatus so that it was in oblique position.

This system wasn't for show: it was a brilliant solution to exhaust the carbon dioxide produced by the yeast. Interestingly, when we arrived in the lab, on the next day, we could appreciate everything had worked from the small bubbles in the test tube.

So, once the mixture had been stirred overnight, at room temperature, we filtered the solution with the help of Celite, which is useful whenever cells are involved in such procedure, unless you want to have your filter rapidly blocked and useless (or so obstructed that the filtration becomes an endless agony).

We saturated the resulting solution with NaCl and, then placed it in a separatory funnel, where we extracted it thrice, with 30 mL of dichloromethane each time.
The organic phases were collected and traces of water removed with anhydrous sodium sulphate.

In this solution, however, there were still either methyl acetoacetate and our product. To check if that was really how things were, we ran a TLC (not shown) where the mobile phase consisted of a solution dichloromethane/methanol (98:2). The products were detected only once the layer had been treated with KMnO₄ and NaOH (yellow stains).

At this stage, the two molecules had to be separated chromatographically in a column packed with silica gel. Now, unfortunately, none of us actually prepared anything: the columns were prepared either by our professor or by PhD students.

Still, at least I can describe what they did: half of the volume of the column was filled with silica gel, which was then placed in a beaker. There we added 100 mL of the aforementioned mobile phase; meanwhile, a little bit of cotton was put right over the bottom valve of the column, together with 5 mL of mobile phase. Then, the suspension of silica gel was poured in, some mobile phase added, the bottom valve was opened, for a short amount of time, to let the stationary phase settle.

With a pipette, they loaded our sample (letting it running down the inner surface and poured some more MP). Before that, a bed of sand was laid in order to get an accurate loading and, as a result, an excellent separation, since sand provides a higher density, so your analytes gather right at the beginning of the stationary phase.

We discarded the first 10 mL and I began to (manually) collect everything coming out of the column in test tubes, while making sure the stationary phase never got dry.

While collecting the samples, I performed several TLCs (same conditions of the previous, preparatory one) to determine the content of any test tube that had already been filled, using a standard solution of methyl acetoacetate as reference.

Predictably, the first analyte to be eluted was the reagent (more hydrophobic than the product), while, at the sixth TLC, one stain confirmed all the product was in one test tube only. Nevertheless, we put the content of those collected immediately before and after, too, in a round-bottom flask.

 

Once we reached a constant weight (0.1118 g), the oil was placed in a 10 mL volumetric flask and chloroform was added to reach the final volume.

Finally, we measured the optical rotation (+ 0.40°): so, we really had yielded the S-(+)-methly-3-hydroxybutanoate we were looking for and its specific rotation (+ 35.78°) was pretty close to the one found in literature (+ 38°).