The authors have declared that no competing interests exist.
Conceived and designed the experiments: GA MVC. Performed the experiments: GA MCM VC. Analyzed the data: GA MCM VC MVC. Contributed reagents/materials/analysis tools: GA MCM MVC. Wrote the paper: MVC.
The most frequent disorder of glycosylation is due to mutations in the gene encoding phosphomannomutase2 (PMM2-CDG). For this disease, which is autosomal and recessive, there is no cure at present. Most patients are composite heterozygous and carry one allele encoding an inactive mutant, R141H, and one encoding a hypomorphic mutant. Phosphomannomutase2 is a dimer. We reproduced composite heterozygosity
The most frequent disorder affecting the transfer of N-linked oligosaccharides to proteins is caused by a deficiency of Phosphomannomutase2 (UniProt: PMM2_HUMAN): accordingly the recommended name for this congenital disorder of glycosylation is PMM2-CDG (MIM#212065), but the same pathology is also known as CDG Ia or Jaeken syndrome.
The major activity of PMM2
PMM2-CDG is transmitted through autosomal recessive inheritance and its clinical spectrum is wide with neurological symptoms and a variable involvement of other organs. More than 85 missense mutations have been described for the gene encoding PMM2 [
The second most frequent mutation, F119L (NM_000303.2:c.357C>A; NP_000294.1: p.Phe119Leu) is common in Scandinavia as well as in other Northern European countries and its incidence decreases moving from the North to the South of the continent. The mutation of F119 into leucine weakens the quaternary structure, which is required for activity, and causes a reduction of the enzymatic activity. Both R141H and F119L are described in DBSNP [
F119L, as other specific mutants have been produced in
Preparations containing a single mutant do not reproduce the real proteins found
Phosphoglucose isomerase from rabbit muscle (commercially available from Sigma Aldrich code P9544-1KU), phosphomannose isomerase from
wt-PMM2, F119L and R141H were inserted into Pet22b+ generating wt-PMM2-Pet22b+, F119L-Pet22b+, R141H-Pet22b+ and expressed in
PMM2 has three enzymatic activities:
Phosphomannomutase activity:
Phosphoglucomutase activity:
Phosphatase activity:
Phosphoglucomutase activity of PMM2 was assayed spectrophotometrically at 340 nm and 32°C by following the reduction of NADP+ to NADPH in 0.3 ml reaction mixture containing Hepes 20 mM, pH 7.5, MgCl2 1 mM, NaCl 150 mM, NADP+ 0.25 mM, BSA 0.1 mg/ml, in the presence of 40 μM glucose 1-phosphate (Glc-1-P), 0.01 mg/ml yeast glucose 6-phosphate dehydrogenase, and 5 μM glucose 1,6-bisphosphate (Glc-1,6-P2). When necessary, a different concentration of Glc-1-P or Glc-1,6-P2 was used. In particular: i) when the effect of Glc-1,6-P2 concentration on the phosphoglucomutase activity was evaluated, the phosphoglucomutase activity was measured in the presence of 40 μM Glc-1P, in addition a fixed protein concentration was used (3.7 μg/ml for F119L, 2 μg/ml for wt and 6.8 μg/ml for the mixture of F119L and R141H 1: 3.5); ii) when the
Phosphomannomutase activity in fibroblasts was also measured similarly and as essentially described by Van Schaftingen et al [
Wild type PMM2, F119L and R141H mutants were individually diluted at 2 μM in Hepes 20 mM, NaCl 150 mM, MgCl2 1 mM (pH 7.5) at 37°C for 30 min. Half of each mixture was incubated with Glc-1,6-P2 at 100 μM for 60 min. All samples were then analyzed by RP-HPLC-ESI-MS carried out on a Q-ToF-Premiere (Waters, Co.) equipped with Alliance binary pump using a Jupiter C4 column (5 μm, 300A°, 50 mm, Phenomenex). The chromatographic runs were carried out using H2O 0.1% Trifluoroacetic acid (buffer A) and acetonitrile 0.1% Trifluoroacetic acid (buffer B), from 10% to 80% of buffer B in 30 min. All mass spectra were acquired from 500 to 2500 m/z values. Horse heart myoglobin was used for tuning Q-ToF instrument and for mass calibration. All data, including ESI-MS on F119L/R141H heterodimers, were acquired and analyzed by MassLynk 4.0.
Analytical size exclusion chromatography was performed using a BioSep-SEC-S3000 column (Phenomenex) equilibrated in Hepes 20 mM pH 7.5, NaCl 150 mM, MgCl2 5 mM. Wild type PMM2 (10 μg), F119L (7 μg) or a mixture of them (5 μg of wt-PMM2 and 3.5 μg of F119L) were analyzed. The chromatography was run at room temperature at 0.5 ml/min on a HPLC system by Shimadzu.
Larger-scale size exclusion chromatography was performed on a Superdex-75 column (1.5x90cm) equilibrated in the same buffer. The chromatography was run at room temperature at 0.7 ml/min on an Akta-prime system. Protein samples (0.5-1mg) (wt-PMM2, F119L, wt plus R141H (1:4.5), F119L plus R141H (1:1 or 1:3.5)) were pre-treated for 30 minutes on ice with EDTA (4 mM final concentration) and then fractionated. Fractions (2 ml each) were analyzed for protein content and enzyme activity.
Alexa Fluor 488 carboxylic acid, succinimidyl ester (AF488) and Alexa Fluor 555 carboxylic acid, succinimidyl ester (AF555), were from Molecular Probes. We performed the conjugation of AF488 with F119L and that of AF555 with R141H following the manufacturer’s instructions. Protein samples, 1 mg of each protein (2 mg/ml in PBS pH 7.2), were treated with 50 μl sodium bicarbonate 1M and added to the reactive dye. The reaction was conducted for 1.5 h at room temperature under gentle stirring. Eventually the unreacted dye was removed by gel filtration.
AF488-F119L and AF555-R141H (3.9 and 6.75 μg respectively) were mixed (in 400 μl Hepes 20 mM pH 7.5, NaCl 150 mM, MgCl2 1 mM). An equal sample was treated for 24h at 4°C with thermolysin (1.5 μg/ml in the presence of CaCl2 1.25 mM). The experiment was repeated three times under identical conditions.
We conducted FRET measurements following the experimental procedure described by Chakraborty
The FRET signal was calculated using the equation Em FRET = Em total−
From measurements of different concentrations of pure AF488-F119L or AF555-R141H, we determined
Fluorescence was measured at room temperature by using a Cary Eclipse Spectrofluorimeter.
Heat-induced melting profiles of wt-PMM2, F119L and R141H were recorded by thermal shift assay or by circular dichroism. In both cases the proteins (0.2 mg/ml) were equilibrated in Hepes 20 mM pH 7.5, MgCl2 1 mM, NaCl 150 mM, DTT 1mM, and the temperature was increased at 0.5°C/min.
Thermal shift assay was performed by using the iCycler iQ Real Time PCR Detection System (Bio-Rad Laboratories, Hercules, CA-USA). The proteins were heated from 25 to 80° in the presence of Sypro Orange 2.4x (Invitrogen Molecular Probes). The normalized melting profiles were obtained using the equation fu(T) = f(T)-fn/fd-fn, where fn represents the minimum value of the fluorescence before the transition and fd represents the maximum value after the transition [
When the melting profile was obtained by circular dichroism (CD) (Jasco J-815 Circular Dichroism Spectrometer), the signal of the protein samples at 220 nm was recorded in the range 20–70°C.
The unfolded fraction was calculated as fu(T) = f(T)—fn(T)/fd(T)—fn(T) where f is the CD ellipticity at 220 nm at temperature T, fn(T), and fd(T) are the values of ellipticity extrapolated at temperature T from the native and unfolded regions of the melting profile.
For limited proteolysis the proteins (wt-PMM2, F119L, R141H, 0.38 mg/ml in Hepes 20 mM pH 7.5, MgCl2 1 mM, NaCl 150 mM) were treated with trypsin (1:50 protease:enzyme ratio) at 37°C [
Urea-induced unfolding of wt-PMM2, F119L and R141H (0.275 mg/ml, Hepes 20 mM pH 7.5, MgCl2 1 mM, NaCl 150 mM) were performed by incubating the pure proteins in the presence of increasing concentration of urea in the presence of Sypro Orange 4x. Fluorescence was recorded at 580 nm (excitation at 485 nm) after 2 h incubation at room temperature. Data were plotted as fraction of protein unfolded. Urea-induced unfolding of PMM2s (wt, F119L, wt/R141H 1:6, and F119L/R141H 1:5 recovered after gel-filtration fractionation) were determined by recording the residual enzymatic activity. The proteins (0.2 mg/ml, Hepes 20 mM pH 7.5, MgCl2 1 mM, NaCl 150 mM) were incubated for 2 hours at 10°C after which the residual phosphoglucomutase activity was measured under standard condition. The reaction mixture (containing appropriate amount of protein) was incubated for known interval times (20–40min), then the reaction was stopped by addition of sodium carbonate 1M and the fluorescence of the NADPH produced was recorded (excitation 340nm, emission 445nm). Data were plotted as normalized residual activity, i.e. the ratio between the activity measured at a given concentration of denaturant divided by the activity in the absence of denaturant. Fluorescence was measured by using a Cary Eclipse Spectrofluorimeter equipped with a high throughput microplate reader.
For western blot analysis polyclonal antibody (10666-1-AP Proteintech) for the detection of PMM2 were used. Adequate amounts of proteins (10 μg) were subjected to SDS-PAGE (15%) and were transferred to PVDF membrane. The detection was performed by using the Immun-Star WesternC chemiluminescence detection kit (Bio-Rad).
Fibroblasts were a gift from Prof. Flemming Skovby (rigshospitalet.dk) who had obtained with informed consent from patients in accordance with The Code of Ethics of the World Medical Association (Declaration of Helsinki) [
Phosphomannomutase and phosphoglucomutase activity of wild type and F119L have already been compared [
Panel A shows the deconvoluted mass spectrum of wt-PMM2 (2 μM) upon RP-HPLC-ESI-MS analysis of the incubation mixture in absence (back peaks) and in presence of Glc-1,6-P2 (250 μM, front peaks), showing an increment of 80 Da attributed to a single phosphorylation event. Panels B and C show the same experiment carried out on F119L and R141H, respectively.
More in detail, wt-PMM2 showed a MW of 27950±1 Da and of 28030±3 Da; F119L showed a MW of 27915±1 Da and of 27995±2 Da, R141H showed a MW of 27931±4 Da and of 28012±2 Da, each pair without and with Glc-1,6-P2, respectively. In a previous paper by us, it was further proved that Glc-1,6-P2 induces a conformational transition in wild type as well as on mutant PMM2 proteins. Therefore, although abortive, the mechanism of phosphorylation and domain closure, which occurs in the wild type enzyme, takes place also in the completely inactive mutant R141H [
We analyzed the quaternary structure of wt-PMM2 (
wt-PMM2 (10 μg), F119L (7μg) and a mixture of them (5 μg of wt-PMM2 and 3.5 μg of F119L) were analyzed by gel filtration on BioSep-SEC-S3000 column at 0.5 ml/min in Hepes 20 mM pH 7.5, NaCl 150 mM, MgCl2 5 mM.
This result is not unexpected and is consistent with our previous study where we had determined the molecular mass of the wt-PMM2 and F119L under native conditions, 52000±200 and 34000±700, by Light-scattering [
One possibility could be that a heterodimer forms. The mutation F119L affects dimerization when present on both subunits of a homodimer F119L/F119L, but has a less damaging effect when present on a single subunit of a complex wt/F119L. This observation is, in our opinion,
We produced R141H and F119L separately in
wt-PMM2 and F119L, alone and in the presence of R141H were fractionated on a Superdex75 column equilibrated in Hepes 20 mM pH 7.5, NaCl 150 mM, MgCl2 1 mM. The fractions were analyzed for the protein content and for the phosphoglucomutase activity (in the presence of 40 μM Glc-1P and 5 μM of Glc-1,6-P2). Before loading, each protein sample was pretreated with EDTA on ice. Panel A and B: F119L and F119L+R141H (1:1 and 1:3.5). Panel C and D: wt-PMM2 and wt-PMM2+R141H (1:4).
The fractions eluted from the column were assayed to test the effect of the inactive subunit on the activity of F119L (
Panel A) The samples (wt-PMM2, F119L, wt-PMM2 plus R141H (1:1 or 1:4), F119L plus R141H (1:1 or 1:3.5) were fractionated on a Superdex75 column. The specific protein ratio of the sample loaded was taken into account in order to calculate the corrected specific activity shown. The protein content of F119L plus R141H (1:1) fr.43,44 and 45 were analyzed by ESI-MS and the deconvoluted mass spectra are shown in panel C, D and E respectively. Panel B) F119L, F119L plus R141H (1:0.5; 1:1; 1:2; 1:6.5) were analyzed in batch and the activity expressed as fold increase relative to pure F119L. The phosphoglucomutase activity was measured in the presence of 40 μM Glc-1P and 5 μM of Glc-1,6-P2.
We carried out a control experiment in which wt-PMM2 was analyzed either alone or in mixture, 1:1 or 1:4, with R141H. The total amount of proteins loaded on the column, was the same. The eluted fractions were assayed and the chromatograms and the enzymatic activity (U/ml) are reported in
In
In order to confirm the fact that R141H enhances the specific activity of F119L, we carried out an experiment in batch where we can control the relative concentration of the two species. We prepared 5 samples containing F119L 0.18 mg (100%), F119L 0.12 mg/R141H 0.06 mg (1:0.5; 66%), F119L 0.09 mg/R141H 0.09 mg (1:1; 50%), F119L 0.06 mg/R141H 0.12 mg (1:2; 33%), F119L 0.024 mg/R141H 0.156 mg (1:6.5; 13%) as previously described and let them equilibrate for 30 min on ice then dialyzed, before assaying the phosphoglucomutase activity. Specific activity was calculated dividing enzymatic units by the amount of F119L. The results are reported in
F119L and R141H were covalently labelled with Alexa Fluor 488 (λex = 495/λem = 519) and Alexa Fluor 555 (λex = 555/λem = 565) respectively. AF448-F119L served as the donor and AF555-R141H as the acceptor. Pure proteins were analyzed by UV-visible spectroscopy (
The dye-protein conjugates (AF555-R141H 0.15 mg/ml, AF488-F119L 0.26 mg/ml) were analyzed by UV-spectroscopy (Panel A). Normalized fluorescence emission spectra of a mixture of AF488-F119L (9.75 μg/ml) and AF555-R141H1 (16.9 μg/ml) recorded upon excitation of 470nm is shown (Panel B, emply circles). The emission spectra obtained when the protein mixture was pre-treated with thermolysin is showed for comparison (Panel B, filled circles).
AF488-F119L and AF555-R141H (3.9 and 6.75 μg respectively) were then mixed in Hepes 20 mM pH 7.5, NaCl 150 mM, MgCl2 1 mM. As a negative control, an equal sample was treated for 24h at 4°C with thermolysin (1.5 μg/ml in the presence of CaCl2 1.25 mM). The spectrum recorded under this condition represents the emission of AF488 and AF555 when no energy transfer due to protein-protein interaction is possible. The protein mixtures were excited at 470 nm and their emission spectrum was recorded (
We measured the
The activity of PMM2 is greatly influenced by the bis-phosphate activators and the relative activity of F119L (open circles in
Protein samples from the Superdex75 fractionations were used. The phosphoglucomutase activity was measured in the presence of 40 μM Glc-1P. The specific protein ratio of the sample F119L+R141H was taken into account in order to calculate the corrected specific activity.
We carried out experiments using variable concentrations of the activator Glc-1,6-P2 at fixed protein concentration, F119L, wt-PMM2 and the mixture of F119L and R141H (1:3.5) and at fixed substrate concentration, 40 μM Glc-1-P. Corrected specific activity was calculated per microgram of active subunit. This means that the specific activity of the mixture (filled circles line in
We tested the stability of F119L and R141H by limited proteolysis, urea induced or thermal induced denaturation and compared it to that of the wild type enzyme (
Panel A: wt-PMM2, F119L, and R141H were treated with trypsin (1:50 protease:enzyme ratio) and the time-course of the reaction was monitored by SDS-PAGE. Panel B: wt-PMM2, F119L, and R141H (0.275 mg/ml) were incubated in the presence of increasing concentration of urea. The incubation was conducted at room temperature in the presence of Sypro Orange 4x, the emitted fluorescence was recorded and data were normalized. Panel C: temperature melting profiles of wt-PMM2, F119L and R141H were recorded by thermal shift assay. The proteins (0.2 mg/ml) were heated from 25 to 80° at 0.5°C/min in the presence of Sypro Orange 2.4x. Panel D: temperature melting profiles of wt-PMM2, F119L and R141H recorded by circular dichroism. The proteins (0.2 mg/ml) were heated from 20 to 70° at 0.5°C/min and the signal at 220nm was recorded. All the experiments were conducted in Hepes 20 mM pH 7.5, NaCl 150 mM, MgCl2 1 mM.
In order to test the stability of the mixed proteins we determined the melting curves in the presence of a denaturant by following the enzymatic activity (
The proteins (wt-PMM2, F119L, wt/R141H 1:6, and F119L/R141H 1:5, 0.2 mg/ml in Hepes 20 mM pH 7.5, MgCl2 1 mM, NaCl 150 mM) were equilibrated with urea (from 0 to 6 M) for 2 hours at 10°C after which the residual phosphoglucomutase activity was measured under standard condition. Data were expressed as residual enzymatic activity.
Mutants that are unstable
Cell extracts (two healthy control fibroblasts and three primary fibroblasts of patients, 10 μg of each) were analyzed and protein were visualized by incubation with polyclonal anti-PMM2 antibody (A).Cell extracts of Cos7 transiently transfected with wt-PMM2-IRES2-EGFP vector or with F119L-IRES2-EGF (10 μg of each), were analyzed and protein were visualized by incubation with polyclonal anti-GFP antibody (B) and polyclonal anti-PMM2 antibody (C) Beta-actin (Panel A and B) or Alpha-tubulin were used as loading controls and purified wt-PMM2 as a standard (Panels B and C).
As seen in
Phosphomannomutase activity was also measured in fibroblast extracts. The activity found in healthy controls was 1.70±0.45 mU/mg, whereas in patients the activity was 0.13 and 0.16 mU/mg as far as F119L/R141H is concerned (lane 3 and 4), 0.19 mU/mg in the genotype F119L/F119L (lane 5).
In order to confirm that F119L is less stable than wt-PMM2
In a previous paper by us we demonstrated that PMM2 works as a dimer and estimated the dissociation constant of wild type and F119L homodimers [
The two chains are represented as cartoons in light or dark grey. F119 on one chain and L119 on the other chain, are shown as spheres in green or red respectively. R141on one chain and H141 on the other chain, are shown as sticks in green or red respectively. The side chains of E93 and R116 forming a salt bridge at the interface, are shown as sticks in orange or blue respectively. The side chains of K115 and N101 forming a H-bond at the interface, are shown as sticks cyan or magenta respectively.
Disease missense mutations reduce the activity of a protein in the cell by different mechanisms: they can lower the specific activity, they can prevent folding or they can affect the stability of the folded protein. In the last case reduced activity is only a secondary effect because unstable mutant proteins have a lower lifetime, and hence a lower concentration in the cell. It is mandatory to distinguish the mechanisms by which mutations exert their deleterious effect. Only for the third case in fact, small molecule drugs can reverse the effect of the mutation by stabilizing the protein, increasing its concentration and hence restoring the activity. Could small molecule drugs, pharmacological chaperones or drugs that affect proteostasis, be used to treat PMM2-CDG as it occurs for other genetic diseases [
It was demonstrated that F119L has an activity that varies from half to one third of that of the wild type, depending on the condition of the assay. As shown in
Then why is F119L/R141H a severe genotype compared with wt/R141H? Let us over-simplify the model and assume that in the cells of the asymptomatic carriers, who carry one wild type allele and one R141H allele, 50% of the proteins are in the form of a heterodimer wt/R141H, 25% in the form R141H/R141H and 25% in the form wt/wt whereas in the cells of the patients, who carry one F119L allele and one R141H allele, 50% of the proteins are in the form of the heterodimer F119L/R141H, 25% in the form R141H/R141H and 25% in the form of the monomers F119L. Since we have shown that the specific activity of F119L/R141H and that of wt/R141H are comparable, the severity of the phenotype should depend on the fraction that is represented by fully active wt/wt in the healthy carriers and by the monomeric inactive F119L in the patients. However in addition to this, we found that fraction represented by F119L/R141H, although being active, is unstable. For this reason in the cells of the patients the reduced phosphomannomutase2 activity should be primarily due to a lower protein concentration and the genotype F119L/R141H should be responsive to pharmacological chaperones and/or to drugs that act on proteostasis. As a proof of concept we have analyzed fibroblasts derived from two patients carrying the genotype F119L/R141H, those deriving from a patient carrying the genotype F119L/F119L and those derived from two healthy controls matched by age, and we have found that the amount of the protein revealed by western blot is much less in the cells of the patients. We are aware of the fact that this type of analysis should be carried out on a higher number of cases and of controls and more importantly on fresh derived lymphoblasts. It is known that the activity of PMM2 is variable when fibroblasts are analyzed and tend to increase upon passages [
To support our finding that the inactive mutant R141H ameliorates the activity of the hypomorphic mutant F119L, we analyzed the literature. We found in the paper by Pirard
In conclusion our paper demonstrates that the activity of F119L/R141H is higher than expected due to the promoting protein dimerization by the inactive mutant. The mutual assistance between the mutants is well described borrowing the epigram about the blind man and the lame from Plato the Younger: " A blind man carried a lame man on his back lending him his feet and borrowing from him his eyes". This effect might not be restricted to PMM2, but might occur with other proteins that are active as dimers.
The reduction of total activity found in the patients carrying F119L/R141H can be attributed to instability and a lower amount of the protein in the cells, better than to an intrinsic reduction of the specific activity. In turn, this opens the possibility that these patients, who represent a large share of the total in Northern Europe, can benefit from a therapy with pharmacological chaperones or drugs that affect protein proteostasis.
We gratefully acknowledge Prof. Flemming Skovby for providing fibroblast lines and Mr Emilio Castelluccio for technical assistance. This work is dedicated to our friend and colleague Dr Maria Malanga.
Congenital disorder of glycosylation due to mutations of phosphomannomutase2
phosphomannomutase2
glucose1,6-bisphosphate
glucose 1-phosphate
Median Effective Concentration
reverse phase high performance liquid chromatography electrospray ionization mass spectrometry
wild type
Förster/Fluorescence resonance energy transfer