Table 1.
Summary of the human PGM1 homologs.
Fig 1.
A multiple sequence alignment of human PGM1 and its paralogs.
Strictly conserved residues are in white font on a red background; highly conserved residues are in red font. Secondary structure elements from the human PGM1 crystal structure (PDB-ID 5EPC) are above the alignment: α-helices are shown as coils, β-strands as arrows, and turns by TT. Residues with known disease-associated missense variants are highlighted with triangles (gold for PGM1, cyan for PGM3). Residues with variants from the ExAC database corresponding to disease-related mutants in PGM1 or PGM3 are indicated with black star. For other symbols, see key on figure. The amino acid sequence of PGM3 has been altered to account for its circular permutation relative to the other proteins; borders of translocated segments (~60 aa) are indicated by green/pink arrows (see also S1 Fig).
Fig 2.
Crystal structures of human PGM1 and a PGM3 homolog.
(A) Human PGM1 (PDB ID 5EPC) colored by structural domain: domain 1 (magenta, residues 1–191); domain 2 (green, residues 192–304); domain 3 (orange, residues 305–421), and domain 4 (cyan, residues 422–562). Bound metal ion in the active site is shown as gray sphere. (B) A homolog of PGM3 (N-acetylphosphoglucosamine mutase from C. albicans; PDB ID 2DKD) colored as in (A). The domains are: domain 1 (residues 1–191); domain 2 (residues 192–311); domain 3 (residues 312–456); and domain 4 (residues 457–544). (C) A structural superposition of PGM1 (colored by domain) and the PGM3 homolog (white). (D) A schematic of the domain arrangement of PGM1 and PGM3, indicating the residues transposed by the circular permutation in domain 1 of PGM3. Key active site regions are indicated.
Fig 3.
Sequence preferences for key functional regions of the PGM1 paralogs.
Selected regions of paralog-specific multiple sequence alignments are shown for the phosphoserine (P-Ser), metal-binding, sugar-binding, and PO4-binding loops. Identical residues are highlighted with red background; similar residues are shown in red font. Asterisk (*) indicates the catalytic serine residue.
Fig 4.
Close-up view of key active site regions of the PGM1 paralogs.
(A) The catalytic serine and metal-binding loop (regions i- ii) in a superposition of PGM1 (cyan), a PGM3 ortholog (green), and a related bacterial enzyme (PDB ID 2Z0F) (pink). The catalytic serine (shown in its dephospho-state) and the three conserved aspartates (Asp-A,B,C) that coordinate bound metal are highlighted. The bacterial enzyme has a metal-binding loop sequence equivalent to that of PGM2/2L1, and is included to show its structural similarity despite the proline between Asp-A and Asp-B. (B) The sugar-binding and PO4-binding loops (regions iii-iv) of PGM1 and the PGM3 ortholog. Colors as in (A). The bound substrate (N-acetylglucosamine 1-phosphate) from the enzyme-ligand complex in 2DKD is shown in yellow; bound sulfate ion in 5EPC is shown in light blue. In the sugar-binding loop, the conserved glutamate found in all paralogs, and the nearby Ser/Asn (PGM1 vs. PGM3) are labeled. In the PO4-binding loop, the two conserved arginines (Arg-A,B) found in PGM1/3/5 are highlighted, along with the conserved glutamate found in PGM2/2L1/3.
Fig 5.
Phylogenetic and sequence relationships among human PGM1 and its paralogs.
(A) A phylogenetic tree was constructed using MEGA7 [27] to show the evolutionary relationships among the five proteins (see also Methods). Human sequences are highlighted with blue triangles. Bacterial sequences/sub-groups are indicated by black stars. Red circles mark branch points for the emergence of the different paralogs (e.g., between PGM2 and PGM2L1). Selected sequences from two other enzyme sub-groups (PMM & PNGM) within the large α-D-phosphohexomutase superfamily are shown as a distinct cluster (gray); these enzymes are found exclusively in bacteria or archaea. (B) Amino acid sequence identity matrix for the PGM1 paralogs (sequence references on Table 1). Top-right section of the matrix shows amino acid identities; lower-left shows pairwise sequence distances as calculated by seqinR [28]. Sequence identities with PGM3 were calculated using the altered version of its sequence to account for the circular permutation, as in Fig 1.
Fig 6.
Tissue-specific protein expression profiles of the PGM1 paralogs.
Data are from ProteomicsDB [35] for 30 human tissues. (A) Protein expression levels are shown on a log scale using their iBAQ scores. Brackets at left indicate clustering of tissues into groups with similar protein expression patterns. (B) RNA expression levels. Data are normalized across each row by Z-score, emphasizing the relative RNA expression for the paralogs in each tissue. (Z-score is the standard deviation from the mean per row). Gray cells indicate that the corresponding tissues are missing in the respective database.
Fig 7.
Residues affected by disease-related mutants mapped onto the structures of (A) PGM1 (PDB ID 5EPC) and (B) a homolog of PGM3 (PDB ID 2DKD).
Residues affected by PGM1 missense variants are highlighted in yellow; those in PGM3 in green.
Table 2.
Known and predicted disease-associated missense variants in the PGM1 paralogs.