Reduction in structural disorder and functional complexity in the thermal adaptation of prokaryotes

Abstract

Genomic correlates of evolutionary adaptation to very low or very high optimal growth temperature (OGT) values have been the subject of many studies. Whereas these provided a protein-structural rationale of the activity and stability of globular proteins/enzymes, the point has been neglected that adaptation to extreme temperatures could also have resulted from an increased use of intrinsically disordered proteins (IDPs), which are resistant to these conditions in vitro. Contrary to these expectations, we found a conspicuously low level of structural disorder in bacteria of very high (and very low) OGT values. This paucity of disorder does not reflect phylogenetic relatedness, i.e. it is a result of genuine adaptation to extreme conditions. Because intrinsic disorder correlates with important regulatory functions, we asked how these bacteria could exist without IDPs by studying transcription factors, known to harbor a lot of function-related intrinsic disorder. Hyperthermophiles have much less transcription factors, which have reduced disorder compared to their mesophilic counterparts. On the other hand, we found by systematic categorization of proteins with long disordered regions that there are certain functions, such as translation and ribosome biogenesis that depend on structural disorder even in hyperthermophiles. In all, our observations suggest that adaptation to extreme conditions is achieved by a significant functional simplification, apparent at both the level of the genome and individual genes/proteins.

Figure 1. Structural disorder and optimal growth temperature of prokaryotes.

(A) Average structural disorder of proteins in prokaryotes was predicted by the IUPred algorithm , , averaged over all proteins in the proteome, and is shown as a function of the OGT of the bacteria (borders of OGT classes marked by vertical dashed lines). (B) Because the exact value or range of OGT is not reported for all prokaryotes, which, however, are classified as psychrophiles (OGT 5–17°C), mesophiles (20–42°C), thermophiles (45–75°C) and hyperthermophiles (75–105°C), average disorder within these groups has also been calculated. The horizontal line shows median of disorder, whereas the grey box represent standard error of mean (SEM). Error bars show the highest and lowest value observed within that group. Asterisks mark if difference of average disorder of the given group and mesophiles is significant (one asterisk: significant, p<0.05, three asterisks: highly significant, p<0.0001).

Figure 2. Taxonomic distribution of structural disorder in bacteria.

Average structural disorder in bacteria in different odrers has been calculated and is shown by color coding. Orders are given by name, and genera within are colored by boxes that reflect the respective average level of disorder, such as white (0–16%), yellow (16–20%), ochre (20–24%), orange (24–28%) and red (above 28%). Generally, bacteria that belong to the same genus tend to have similar average disorder, but no general correlation between closely related orders is apparent.

Figure 3. Average disorder of proteins and genome size.

(A) The size of proteome (actually the number of annotated genes in the genome) is shown as a function of average predicted disorder of proteins in prokaryotes with known status of thermal adaptation. Particular groups are also shown highlighted, such as hyperthermophiles (B), Actinobacteria (C) and halophiles (D).

Figure 4. Transcription factors in the thermal adaptation of prokaryotes.

It is assessed how transcription factors, i.e. a group of proteins of an essential function that depends on structural disorder is affected by thermal adaptation. (A) The average length of annotated transcription factors (error bars, SEM) is shown for the four groups psychrophiles, mesophiles, thermophiles and hyperthermophiles. The average length of TFs in mesophiles with the same average proteome size as thermophiles (meso-tehrmo) or hyperthermophiles (meso-hyper), is also shown. (B) The average level of predicted disorder of annotated transcription factors (error bars, SEM) for the four groups. The average disorder of groups meso-tehrmo and meso-hyper, as defined above, is also shown. (C) The ratio of TFs among all annotated genes is shown for the four groups and meso-thermo and meso-hyper, as defined above. In all three panels, asterisks mark if difference of average from that of mesophiles is significant (one asterisk: significant, p<0.05, three asterisks: highly significant, p<0.0001)