Heat denaturation is a time-honored procedure for protein purification, but do not expect enrichment beyond 3-4 fold. You need to carefully determine the maximal temperature and the maximal time that do not harm your enzyme. For the actual treatment, hold the sample into a boiling water bath under constant swirling until you have reached the pre-determined temperature (thermometer directly in the sample). Then transfer into a second waterbath of that temperature for the time required, finally rapidly chill the sample in ice-water mixture, again swirling constantly. Once the sample is cold, leave it in ice for an hour or so, then spin down precipitated material.

The key point is to do this reproducibly. You need a small volume in a large Erlenmeyer- or round flask, so the surface-to-volume ratio is large for efficient heat transfer and you can swirl without spilling. When you scale up, do not increase the volume of sample, but rather work in several aliquots. That will keep the time to reach the required temperature constant.

It sounds too cold to be able to skip columns altogether. We had to get Taq up to 85C for 1 hour to completely clear out all other protein bands. And then you're still stuck with lipids, nucleic acids, colored compounds from medium. Not sure why you'd ever voluntarily go through the trouble. If you have an FPLC method working, that sounds far preferable to me.

A large DOE (up to 96 samples) with one of the older gradient PCR heat blocks can sometimes show you optimal thermal lysis conditions, but this is just a pretreatment before a conventional FPLC purification usually, as John and Englebert mentioned. 65-70C is also just not high enough to cleanly differentiate any mesophilic protein background, good luck!

Engelbert Buxbaum Thanks for the protocol!