Fig 1.
Abundance levels and enzymatic activity of HvPho1 during barley endosperm development.
(A) HvPho1 protein abundance from barley endosperm extracts analyzed via immunoblot at 2 d intervals. Only one band is visible just above 100 kDa in accordance with an expected mass of 105 kDa. (B) Relative quantification of the data from panel A (blue line) and activity from panel C (H2O control; red line). (C) Starch phosphorylase activity probed in 2 day intervals as for panel A but with native gels and Lugol coloring of activity products. Strong synthetic activity appears as a dark stained band and is marked with a black arrow. White bands and smears represent amylolytic activities. All gels include a recombinant HvPho1 control as the right-most band. The different redox treatments are indicated next to each gel. (D) Immunoblot (top) and native gel (bottom) analysis of HvPho1 protein abundance on buffer soluble (S) protein and buffer insoluble (P) protein fractions of barley endosperm between 0 and 8 DAF. Numbers indicate the DAF. Arrows mark the position of the two relevant bands in the immunoblot and the position of the (single) activity band in the zymogram.
Fig 2.
Apparent molecular weights and affinity of HvPho1 constructs.
(A), (B), (C) Total barley endosperm soluble protein loaded onto a 26/60 Superdex S-200 SEC column. (A) Protein fractions as analyzed by SDS-PAGE/immunoblot using anti-HvPho1 polyclonal antibodies and (B) Native gels using glycogen as in gel glycosyl acceptor and G1P as glycosyl donor. The first lane (marked M) on each immunoblot is protein molecular weight ladder. The first lane in each native gel (marked C) is the recombinant HvPho1 control. Arrows on top of the gels indicate molecular weight of the protein fraction according to column calibration. (C) Chemical cross-linking of HvPho1 dimers in solution. HvPho1 was incubated for the indicated times with 0,15% (v/v) glutaraldehyde. Formation of HvPho1 cross-linked dimers is indicated. (D) Protein fractions as analyzed by SDS-PAGE/immunoblot using anti-HvBeIIb polyclonal antibodies. Total barley endosperm protein was either incubated either with (left gel) or without (right gel) 1 mM ATP and 2.5 mM protein phosphatase inhibitor cocktail prior to size separation via SEC. Lane 1: protein molecular weight ladder, lanes marked 2: protein fractions from SEC ranging between 330 kDa and 190 kDa; lane 3: recombinant HvBeIIb purified from E. coli. Brown arrows to the right of both gels indicate HvBeIIb. (E) SEC profile of HvPho1 (blue) and HvPho1ΔL78 (black). The inserts are DLS analyses of the hydrodynamic sizes of the two proteins. The green and brown lines represent the SEC peak fractions of the respective proteins used for SEC analysis. Arrows on top indicate molecular weight according to column calibration. (F) The affinity of recombinant HvPho1, HvPho1F50 and HvPho1ΔL78 for amylopectin and starch assessed by analysis of starch bound and unbound protein in vitro in two ways: Top: By incubation of the proteins with 5, 1 and 0.5 mg∙ml-1 amylopectin and successive analysis of soluble (S) and pellet–amylopectin bound (P) fraction with SDS-PAGE. The concentrations of amylopectin are indicated over each S/P pair. Bottom: analysis via native gel with in gel starch as an interaction partner. Rf values are plotted versus the starch concentration in the gels. The resultant affinity constants for half maximum binding are given.
Fig 3.
Crystal structure and modes of polysaccharide binding.
(A) Thin Layer Chromatography (TLC) of a crystallization drop of HvPho1 containing 10 mM maltoheptaose. A ladder with 1 mM sugar markers is included and two dilutions from the crystallization drop were applied next to it. 1 μl of solution was loaded in each lane. (B) Overall structure of the HvPho1 dimer in the crystals. One monomer is shown as a semitransparent gray surface with the α-carbon trace as a stick model. The second monomer is shown as a cartoon model. Colors are in rainbow from dark blue in the N-terminus through light blue, green, yellow and orange to red at the C-terminus. The protein co-factor pyridoxal phosphate group is show as spheres (pink carbons) and lies at the interface between both, the N-terminal and C-terminal subdomains. The missing loops are indicated with dashed green lines. (C) Structural overlay of the native HvPho1 structure (green) and rabbit glycogen phosphorylase B (PDB-code 1P2B). 1P2B has maltoheptaose (shown as ball and sticks) bound in the glycogen storage site of glycogen phosphorylase B, although only maltopentaose has been explicitly modelled. 2 different angles are presented. Parts of HvPho1 corresponding to the L78 insert are colored pink with the last modeled residues indicated by sequence numbers. (D) Recombinant HvPho1, when stored at 15°C, degrades over time into specific degradation products called F50 and F50s. Left: initial degradation products after 1 week. Middle: a stable F50 band, indicated with a square, was apparent after 4 weeks incubation. It was excised from SDS-PAGE, gel eluted, trypsin digested and the resulting proteolytic fragments analyzed by MALDI-TOF. The sequence to the right shows full-length HvPho1. Labelled in red are MALDI-TOF recorded trypsin fragments (they do not cover the L78 insertion). (E) Superposition of HvPho1 (green carbon backbone) on the maltotriose binding site of AtPHS2 (gray carbon backbone). The maltotriose from the AtPHS2 crystal, not present in the HvPho1 structure, is shown with cyan carbons. For clarity, only side chains are shown except for W405/W363. Hydrogen bonds to the maltotriose in AtPHS2 are highlighted with yellow dashed lines and the corresponding residues are labelled in the figure, first with the Pho1 residue numbering, then with the AtPHS2 numbering.
Fig 4.
Structural details of the active site of HvPho1 and acceptor recognition.
(A) Mode of binding of maltotetraose in the active site of HvPho1 (which is depicted as a semi-transparent ribbon). Maltotetraose is shown as ball and stick with yellow carbons and can also be seen in the same color in panel C. A modeled maltose is in ball and stick with orange carbons. Contacting amino acids are shown as stick models with green carbons, while those involved in stacking interactions with the glucose units are depicted with gray carbons. The pyridoxal phosphate is also depicted, with pink carbons, in the lower left corner. (B) Movement of a loop of HvPho1 in response to maltotetraose binding. Residues 422–427 of the HvPho1 complex with maltotetraose are highlighted with all atom sticks (orange carbons) and the maltotetraose is shown as sticks with cyan carbons. The thicker ribbons represent the α-carbon trace of HvPho1 bound to maltotetraose (orange), HvPho1 in the native structure (green), HvPho1 in complex with acarbose (gray), EcMalP in complex with maltopentaose (yellow) and rabbit muscle glycogen phosphorylase (white). (C) Superposition of maltopentaose in the binding site of MalP (white, from PDB_code 1e4o), maltotetraose in HvPho1 and acarbose in HvPho1; plus PLP groups and selected details from the native HvPho1 structure. For clarity, only the mentioned groups, the pyridoxal-5’-phosphates (PLP) and, in the case of all three structures of HvPho1 reported here, Tyr900 and Tyr905 are depicted. Details from the maltotetraose complex are shown with yellow carbons, details from the acarbose complex with cyan carbons and details from the native structure with pink carbons. The four glucose units in the maltotetraose complex overlap well with the MalP structure for sub-sites +1 to +4.
Fig 5.
De novo production of α-1,4 glucans by HvPho1.
(A) Proton NMR analysis of de novo synthesis of α-1,4 glucans by HvPho1. HvPho1 and HvPho1Asp383Ala (0,1 mg∙ml-1) were incubated with G1P (25 mM) as sole substrate. The figure shows production of 1,4 glycosidic bonds, usage of G1P and generation of reducing ends recorded over time by proton NMR spectroscopy. (B) Plot of the generation of 1,4 glycosidic linkages over the number of reducing ends indicating approximate lengths of glucans produced over time. (C) Recombinant HvPho1 (0.05 mg∙ml-1) and HvBeIIa (0.05 mg∙ml-1) were incubated with G1P (50 mM). Samples were taken at the start of the reaction (0 h) after 2 h and after 4 h. (D) The last sample (4 h) from panel C was debranched using CrIsa1 (0.01 mg∙ml-1) at 25°C overnight. Each sample was boiled, filtered and loaded on a CarboPac PA-100 ion-exchange column and analyzed with pulsed amperometric detection. The gradient used could not distinguish between species with more than 32 glucoses.
Fig 6.
Model describing the effect of the L78 insertion on polysaccharide binding to HvPho1.
(A) HvPho1 forms a homodimer in solution with an enlarged molecular size. The formation of the dimer is brought about by the crystallographic dimer interface. The flexible nature of the L78 insertion could block access of larger glucans to the protein´s surface. (B) The specific degradation products of HvPho1 are the F50s which probably lack L78. Our crystal structure does not contain the L78 insertion and might therefore represent the F50s rather than the full-length enzyme. The F50s provide better access to larger polysaccharides like starch or amylopectin. (C) HvPho1ΔL78 lacks the L78 insertion but it also lacks a break in the protein chain. Affinity of larger polysaccharides is similar to full-length HvPho1 as the main protein backbone is closed and restricts access to this area.
Table 1.
Data collection and refinement statistics.