Fig 1.
Architectural differences between His-Asp phosphotransfer systems.
(A) Two component phosphotransfer involved in osmoregulation in Escherichia coli [1]. (B) Four component system in Saccharomyces cerevisiae involved in responding to osmotic stress, with the first two components on a single protein, Sln1. A similar architecture in Dictyostelium discoideum (in parentheses) responds to multiple histidine kinases and a range of initiating signals [2, 3].
Fig 2.
Representation of amino acid conservation flanking the reactive histidine of diverse HPt domains.
In each logo, the height of individual columns reflects overall amino acid conservation (hence the information content) at that position, while the height of letters within a given column reflects the relative representation of the corresponding amino acid at that same position [14]. Amino acids are colored according to the scheme of [15]: hydrophobic residues are black; hydrophilic, green; acidic, red; and basic, blue. In (A), the motif was derived from the 241 seed sequences, mostly prokaryotic, used to generate PF01627; in (B), all 344 eukaryotic members of the complete family were analyzed to generate the motif. Note that the search algorithms employed identify regions with the highest overall conservation within the two different sequence sets, and therefore output motifs that situate the phosphorylated histidine (residue “0”) at different relative positions. Furthermore, the phosphorylated histidine is not absolutely conserved in the eukaryotic sequences presented. This is due to certain proteins (such as AHP6; see text) that lack a histidine at the conventional position but are included in PF01627 nevertheless; their biological function may be related to but distinct from phosphotransfer per se.
Fig 3.
φ score distribution in the A. thaliana proteome.
Upper: Histogram representation of the logged distribution of proteins in the A. thaliana proteome as a function of their likelihood of occurrence in the eukaryotic HPt motif profile matrix. Note the several proteins with likelihood scores orders of magnitude greater than those of the bulk distribution. Lower: The same log likelihood scores of the A. thaliana proteome replotted as a box plot. The Q1-Q3 interquartile range (IQR) is boxed with the median (Q2) shown in red; whiskers indicate 3IQR limits. The five most prominent high-scoring outliers, as well as two other proteins significantly closer to the bulk distribution, are again indicated (“+”) in green and red, respectively. Additional information about each of the seven outlying proteins (including duplicates of three entries) appears in Table 1 and Supporting Information S1 Table.
Table 1.
Examination of presumptive HPt domain-containing proteins in A. thaliana.
Fig 4.
Testing of false positive identifications.
The HPt motif likelihood distribution was examined for the human and mouse proteomes, a pseudo-random selection of 20 amino acid-long sequences (residues 2–21 for all proteins in the A. thaliana proteome), and a search of the A. thaliana proteome with a column-scrambled motif. Ideally, no HPt domains would be indicated in any of these searches. The Q1-Q3 IQR and whiskers are indicated as in the previous figure. The few false positives detected, evident in the first and fourth box plots, exhibit very low likelihood scores barely above the 3IQR cutoff.
Fig 5.
Logged motif likelihood score distributions of select eukaryotic proteomes.
Four plant species are followed by four fungi, a protist, and an arthropod. Full proteomes were downloaded from UniProt and their motif likelihood distributions calculated. IQRs and whiskers are as before. The blue horizontal line corresponds to an alternative approach to identifying outliers (“5% cutoff” as explained in Methods) which is independent of the proteome examined. Outliers are enumerated in Table 2, with full details of each candidate given in S1 Table.
Table 2.
Comparison of the number of HPt domains revealed by probabilistic motif searching with their representation in several widely used protein databases.