Correlating enzyme annotations with a large set of microbial growth temperatures reveals metabolic adaptations to growth at diverse temperatures

Abstract

BackgroundThe ambient temperature of all habitats is a key physical property that shapes the biology of microbes inhabiting them. The optimal growth temperature (OGT) of a microbe, is therefore a key piece of data needed to understand evolutionary adaptations manifested in their genome sequence. Unfortunately there is no growth temperature database or easily downloadable dataset encompassing the majority of cultured microorganisms. We are thus limited in interpreting genomic data to identify temperature adaptations in microbes.ResultsIn this work I significantly contribute to closing this gap by mining data from major culture collection centres to obtain growth temperature data for a nonredundant set of 21,498 microbes. The dataset (10.5281/zenodo.1175608) contains mainly bacteria and archaea and spans psychrophiles, mesophiles, thermophiles and hyperthermophiles. Using this data a full 43% of all protein entries in the UniProt database can be annotated with the growth temperature of the species from which they originate. I validate the dataset by showing a Pearson correlation of up to 0.89 between growth temperature and mean enzyme optima, a physiological property directly influenced by the growth temperature. Using the temperature dataset I correlate the genomic occurance of enzyme functional annotations with growth temperature. I identify 319 enzyme functions that either increase or decrease in occurrence with temperature. Eight metabolic pathways were statistically enriched for these enzyme functions. Furthermore, I establish a correlation between 33 domains of unknown function (DUFs) with growth temperature in microbes, four of which (DUF438, DUF1524, DUF1957 and DUF3458_C) were significant in both archaea and bacteria.ConclusionsThe growth temperature dataset enables large-scale correlation analysis with enzyme function- and domain-level annotations. Growth-temperature dependent changes in their occurrence highlight potential evolutionary adaptations. A few of the identified changes are previously known, such as the preference for menaquinone biosynthesis through the futalosine pathway in bacteria growing at high temperatures. Others represent important starting points for future studies, such as DUFs where their occurrence change with temperature. The growth temperature dataset should become a valuable community resource and will find additional, important, uses in correlating genomic, transcriptomic, proteomic, metabolomic, phenotypic or taxonomic properties with temperature in future studies.Electronic supplementary materialThe online version of this article (10.1186/s12866-018-1320-7) contains supplementary material, which is available to authorized users.

Machine Learning Applied to Predicting Microorganism Growth Temperatures and Enzyme Catalytic Optima

“…To build a machine learning model that can predict OGT from the amino acid composition of proteins encoded by an organism's genome we first established a training dataset. To this end, we downloaded an OGT dataset ( https://doi.org/10.5281/zenodo.1175608 ), which contains data for 21,498 microorganisms, including bacteria, archaea and eukarya 14 . Using this dataset, all proteins from 5,761 organisms from RefSeq (https://www.ncbi.nlm.nih.gov/refseq/) and Ensembl genomes (http://ensemblgenomes.org/) could be associated with an OGT value (we refer this as the annotated dataset), while proteins from an additional 1,803 organisms could not be associated with an OGT value (we refer this as the unannotated dataset) ( Fig.…”

“…2c). Second, we seized on the fact that the average temperature optimum of catalysis ( T opt ) of at least five enzymes from an organism shows a Pearson correlation above 0.75 with growth temperature 14 . Essentially, the catalytic optimum of an enzyme tends to be close to the organism growth temperature.…”

“…Here we use a recently published dataset which contains OGTs for 21,498 microorganisms 14 , and combine this with genomic resources to build a machine learning model. The model predicts OGT from a file containing all proteins encoded by an organism's genome.…”

Biodegradation of endocrine disruptor Bisphenol A by Pseudomonas putida strain YC-AE1 isolated from polluted soil, Guangdong, China

“…The growth rate of microorganisms becomes slow below or above the optimum growth temperature because of a reduced rate of cellular production [35]. Enzyme thermal stability and activity is correlated to an organism's growth temperature and also, degradation is an enzyme-controlled activity hence as the temperature increases, the cellular growth and physiological functions increase to an optimum value [36]. Fouda [37] reported a relatively high optimum temperature (35-40°C) for BPA biodegradation by Klebsiella pneumoniae J2 and Enterobacter asburiae L4.…”

Biodegradation of Endocrine Disruptor Bisphenol A by Pseudomonas putida Strain YC-AE1 Isolated from Polluted Soil, Guangdong, China

“…2(d). Enzyme thermal stability and activity is correlated to an organism's growth temperature and also, degradation is an enzyme-controlled activity hence as the temperature increases, the cellular growth and physiological functions increase to an optimum value [43]. Fouda [44] reported a relatively high optimum temperature (35-40 o C) for BPA biodegradation by Klebsiella pneumoniae J2 and Enterobacter asburiae L4.…”