Characterization of the −148C>T promoter polymorphism in PKLR

We identified the −148C>T mutation in the erythroid-specific promoter of PKLR in 3 unrelated patients with low PK activity. In vitro transfection studies showed that this promoter substitution did not affect promoter activity. We conclude that the −148C>T promoter polymorphism does not cause PK deficiency and, therefore, should be considered a benign polymorphism.

Pyruvate kinase (PK) deficiency is an inherited enzyme abnormality of glycolysis caused by mutations in the gene encoding liver and red blood cell PK (PKLR), and an important cause of hereditary non-spherocytic hemolytic anemia. The clinical phenotype is variable, ranging from severe hemolytic anemia to mild asymptomatic cases1. Until now, more than 190 mutations in the PKLR gene have been reported (http://PKLRmutationdatabase.com), of which only two are located in the erythroid-specific promoter.1,2

A third sequence variation in the promoter of PKLR concerns a C to T substitution at nt −148 (−148C>T). This change occurs at polymorphic frequencies in the North American (allelic frequency 0.007 − 0.013, Ensemble SNP 8177960), Portuguese (0.017)3 and Dutch populations (0.031, this study). The functional relevance of the −148C>T change is unknown but of particular interest since DNA sequence analysis2 of PKLR in 2 unrelated patients with low PK activity (Table 1, individuals 1 and 2) displayed no mutations other than −148C>T. In addition, a third carrier of the −148C>T substitution (individual 3) demonstrated PK activity levels lower than expected while being only heterozygous for the c.1529G>A mutation. We, therefore, hypothesized an association between the −148C>T sequence variation and PK deficiency, as recently also proposed by others.4

Analysis of the proximal part of the erythroid-specific promoter of PKLR (TFSEARCH database)5 revealed a putative binding site for the transcription factor c-myb (nts −151 to −142). This binding site was disrupted by the −148C>T mutation. Subsequent transient transfection assays2 using wild-type and mutant promoter reporter vector constructs in K562 cells showed that the −148C>T substitution did not affect PKLR promoter activity (Figure 1, panel A). In addition, Electrophoretic Mobility Shift Assay (EMSA) with K562 nuclear extract did not show a specific protein-DNA complex with the wild-type nor with the mutant PKLR oligonucleotide probe (data not shown). From these results we conclude that the −148C>T substitution in PKLR does not affect erythroid-specific gene transcription in vitro.

Table 1.Patient characteristics.

Figure 1.The −148C>T sequence variation in the human PKLR promoter does not affect promoter function in K562 cells and is not associated with lower transcript levels in pro-erythroblasts. (A) A PKLR pGl3 promoter reporter gene construct (spanning nts −469 to −1) without (pGL3-PKWT) and with the −148C>T mutation (pGL3-PK148T) was transiently transfected in human erythroleukemic K562 cells. Luciferase activities were calculated relative to the pGL3-SV40 positive control vector. pGL3-Basic was included as negative (promoterless) control. Results are the average of four independent experiments with each sample assayed in duplicate. (B) A cDNA fragment encompassing exon 11 was amplified from total RNA obtained by in vitro production of (pro)erythroblasts from patient 3 (Table 1) and a control subject (C).10 A 1:10 diluted sample of this RT-PCR product was amplified in a second round of PCR (three cycles) and subsequently digested with StyI. As the c.1529G>A change creates a unique restriction site for this enzyme, RT-PCR products amplified from transcripts of the patient’s c.1529A allele are cut into fragments of 298 and 242 bp whereas RT-PCR products from the patient’s c.1529G allele remain uncut (540 bp). The combined intensities of the 242 and 298 fragments are approximately equal to the intensity of the 540 bp fragment, indicating no difference in expression of both alleles. Equal allelic amounts of exon 11, as amplified from the patient’s genomic DNA, served as a control. DNA-PCR product of the patient’s c.1529A allele are cut into 190 and 140 bp fragments whereas the wild-type allele remains uncut (330 bp). M, molecular mass marker; C, control subject.

The role of polymorphic DNA sequence variations in determining susceptibility to disease traits is the subject of much research effort, but it often remains unclear whether they are themselves functionally relevant or just linked to another causative mutation.6,7 We have demonstrated that the −148C>T change is unlikely to affect PKLR expression. Furthermore, the possibility of acquired PK deficiency8 is considered to be highly unlikely in our patients. However, reduced PK activity in our patients could be due to another still unidentified mutation linked to the −148C>T change. Such a mutation could, for instance, reside in an upstream erythroid-specific enhancer element such as the one that has been shown to be essential for high-level PK expression of the rat Pklr gene.9 Therefore, we used RNA from cultured (pro)erythroblasts of this individual to estimate relative transcript levels of both alleles using the c.1529G>A in exon 11 as a marker.10 These analyses suggest no difference in expression of both alleles (Figure 1, panel B). Hence, PKLR promoter activity appears to be unaffected by the −148C>T mutation, nor by any other putative mutation linked to this change. The −148C>T mutation in the erythroid-specific promoter of PKLR should, therefore, be considered a benign polymorphism. The basis for the lowered PK activity in the 3 individuals presented here thus remains to be established.