It has been suggested that temperature and pH are strong explanatory causes of protein evolutionary constraint. However, a general understanding of how changes in temperature and pH may universally impact protein evolution is lacking. Moreover, while pH is locally regulated in the different microbial compartments (e.g. periplasm, cytoplasm), temperature is not. A priori, these variables should act at different rates over evolutionary timescales with respect to genome evolution, since there are genes encoding for proteins allocated in different compartments. To test whether this is true, we can use a proxy for the number of evolutionary events (e.g. mutations, duplications, etc.). One simple yet unregarded way is to calculate the diversity of unique sequences assigned to a biochemical function (KEGG) across species. Given similar diversity of species in particular environments, the expectation for any KEGG is to be represented by a similar number of sequences. We can give answer to the question comparing this metric between metastable communities with differential values of temperature and pH. In this work, we examine the sequence diversity among 17 metagenomes from El Tatio geothermal field (Chile), spanning temperatures of 45-62{degrees}C, and pH values of 7.2-9.3. We found that, when controlling for abundance, higher temperature (>56{degrees}C) increases, while basic pH (>8.1) decreases the expected by-function diversity. The inclusion of these variables improved the prediction of sequence diversity from metagenomic abundance. Between-compartment diversity was more affected by pH than by temperature. Protein length distributions and individual, well-known taxa from hydrothermal springs showed compatible trends. Overall, our results suggest that sequence diversity is universally governed by the signature buffering of the stable compartments to some of the outside-cell factors, besides the taxonomic trends or the particularities of any function. Altogether, we provide evidence against the temperature-induced gene loss hypotheses, and in favor of the confined evolution hypothesis.