Control of the murine phosphofructokinase-A gene during muscle differentiation

Tissue-dependent loss of phosphofructokinase-M in mice with interrupted activity of the distal promoter: impairment in insulin secretion

Phosphofructokinase is a key enzyme of glycolysis that exists as homo- and heterotetramers of three subunit isoforms: muscle, liver, and C type. Mice with a disrupting tag inserted near the distal promoter of the phosphofructokinase-M gene showed tissue-dependent differences in loss of that isoform: 99% in brain and 95-98% in islets, but only 50-75% in skeletal muscle and little if any loss in heart. This correlated with the continued presence of proximal transcripts specifically in muscle tissues. These data strongly support the proposed two-promoter system of the gene, with ubiquitous use of the distal promoter and additional use of the proximal promoter selectively in muscle. Interestingly, the mice were glucose intolerant and had somewhat elevated fasting and fed blood glucose levels; however, they did not have an abnormal insulin tolerance test, consistent with the less pronounced loss of phosphofructokinase-M in muscle. Isolated perifused islets showed about 50% decreased glucose-stimulated insulin secretion and reduced amplitude and regularity of secretory oscillations. Oscillations in cytoplasmic free Ca(2+) and the rise in the ATP/ADP ratio appeared normal. Secretory oscillations still occurred in the presence of diazoxide and high KCl, indicating an oscillation mechanism not requiring dynamic Ca(2+) changes. The results suggest the importance of phosphofructokinase-M for insulin secretion, although glucokinase is the overall rate-limiting glucose sensor. Whether the Ca(2+) oscillations and residual insulin oscillations in this mouse model are due to the residual 2-5% phosphofructokinase-M or to other phosphofructokinase isoforms present in islets or involve another metabolic oscillator remains to be determined.

Novel testis- and embryo-specific isoforms of the phosphofructokinase-1 muscle type gene

We have identified novel transcriptional isoforms of the human and mouse genes encoding muscle type phosphofructokinase-1 (PFK-M). These isoforms are expressed specifically in the testis and in the mid-gestation embryo, and have been termed TE-PFK-M (testis- and embryo-specific PFK-M). The 5'UTR of TE-PFK-M is composed of three newly identified exons that lie much farther upstream of the PFK-M coding region than the previously characterized 5'UTR. In addition, this upstream region encodes a series of small polyadenylated transcripts, some of which share the same exons found in the 5'UTR of TE-PFK-M, and which may play some role in regulating TE-PFK-M expression. These findings indicate an even more complex level of control of PFK-M expression than previously thought.

Genomic organization, 5'flanking region and tissue-specific expression of mouse phosphofructokinase C gene

Using a combination of mouse bacterial artificial chromosome (BAC) genomic library screening, long-range polymerase chain reaction (PCR) amplification, genomic walking and DNA sequencing, we have characterized the intron/exon boundaries, the sizes of each intron and 5' flanking region of the mouse PFK-C gene. The gene spans approximately 55 kb and comprises 22 exons separated by 21 introns. All intron/exon splice junctions conform to the GT/AG rule. The mouse PFK-C gene organization is similar to that of the human and rabbit PFK-A and human and mouse PFK-B genes. However, PFK-C has much larger intronic sequences throughout the gene. Anchored PCR was performed to amplify about 1.0 kb of genomic DNA upstream of the translational start site. Sequence analysis of the PFK-C 5' flanking region revealed that it is devoid of TATA and CAAT boxes at the usual positions, but it contained several putative binding sites for transcription factors AP1, GATA1, NKX2.5 and STAT. The 5' flanking region was not enriched in GC dinucleotides and lacked CpG islands and putative binding sites for SP1. Four transcription initiation sites have been identified by full-length RNA ligase-mediated rapid amplification of cDNA ends (RLM-RACE) between -61 and -32 bp from the translation initiation codon. Reverse transcription-PCR analysis revealed that PFK-A, PFK-B and PFK-C genes were expressed, in all mouse tissues tested, at varying levels. PFK-A mRNA was more abundantly expressed in all tissues than were the PFK-B and PFK-C genes. Based on the mouse PFK-C signal normalized to 18S rRNA, the PFK-C mRNA was expressed at the highest levels in the brain, heart, thymus and testicles, whereas low levels were observed in the kidney, liver, muscle, and lung.

Molecular mechanisms of doxorubicin-induced cardiomyopathy. Selective suppression of Reiske iron-sulfur protein, ADP/ATP translocase, and phosphofructokinase genes is associated with ATP depletion in rat cardiomyocytes

Doxorubicin, a cardiotoxic antineoplastic, disrupts the cardiac-specific program of gene expression (Kurabayashi, M., Dutta, S., Jeyaseelan, R., and Kedes, L. (1995) Mol. Cell. Biol. 15, 6386-6397). We have now identified neonatal rat cardiomyocyte mRNAs rapidly sensitive to doxorubicin, or its congener daunomycin, including transcripts of nuclear genes encoding enzymes critical in production of energy in cardiomyocytes: ADP/ATP translocase, a heart- and muscle-specific isoform; Reiske iron-sulfur protein (RISP), a ubiquitously expressed electron transport chain component; and a muscle isozyme of phosphofructokinase. Loss of these mRNAs following doxorubicin or daunomycin is evident as early as 2 h and precedes significant reduction of intracellular ATP. ATP levels in control cardiomyocytes (17.9 +/- 2.9 nM/mg of protein) fall only after 14 h and reach residual levels of 10.4 +/- 0.9 nM (doxorubicin; p = <0.006) and 6.7 +/- 1.9 nM (daunomycin; p = <0. 001) by 24 h. Loss of mRNAs generating ATP was highly selective since mRNAs for other energy production enzymes, (cytochrome c, cytochrome b, and malate dehydrogenase), and genes important in glycolysis (pyruvate kinase and glyceraldehyde-3-phosphate dehydrogenase) were unaffected even at 24 and 48 h. The drugs had no effect on levels of ubiquitously expressed RISP mRNA in fibroblasts. These findings could link doxorubicin-induced damage to membranes and signaling pathways with 1) suppression of transcripts encoding myofibrillar proteins and proteins of energy production pathways and 2) depletion of intracellular ATP stores, myofibrillar degeneration, and related cardiotoxic effects.

Regulation of tissue-specific expression of the skeletal muscle ryanodine receptor gene

The ryanodine receptors (RYR) are a family of calcium release channels that are expressed in a variety of tissues. Three genes, i. e. ryr1, ryr2, and ryr3, have been identified coding for a skeletal muscle, cardiac muscle, and brain isoform, respectively. Although, the skeletal muscle isoform (RYR1) was shown to be expressed predominantly in skeletal muscle, expression was also detected in the esophagus and brain. To analyze the transcriptional regulation of the RYR1 gene, we have constructed chimeric genes composed of the upstream region of the RYR1 gene and the bacterial chloramphenicol acetyltransferase (CAT) gene and transiently transfected them into primary cultured porcine myoblasts, myotubes, and fibroblasts. A 443-base pair region upstream from the transcription start site was sufficient to direct CAT activity without tissue specificity. Deletion of a 61-base pair fragment from the 5'-end of the promoter resulted in a marked reduction of CAT activity in all three tissue types. A similar reduction of expression was observed when using a construct with the first intron in antisense orientation upstream from the promoter. In contrast, the first intron in sense orientation enhanced expression only in myotubes, while expression was repressed in fibroblasts and myoblasts. Gel retardation analyses showed DNA binding activity in nuclear extracts for two upstream DNA sequence elements. Our data suggest that (i) RYR1 gene expression is regulated by at least two novel transcription factors (designated RYREF-1 and RYREF-2), and (ii) tissue specificity results from a transcriptional repression in nonmuscle cells mediated by the first intron.

Characterization of the murine mitochondrial glycerol-3-phosphate acyltransferase promoter

Glycerol-3-phosphate acyltransferase (GPAT) catalyzes the acylation of sn-glycerol 3-phosphate to form 1-acyl-sn-glycerol 3-phosphate, a committed step in triacylglycerol and phospholipid biosynthesis. We have previously reported the cDNA cloning and transcriptional regulation of the murine mitochondrial GPAT (mGPAT). We now report the cloning of the 5'-flanking region of the murine mitochondrial GPAT. The transcription start site was identified by primer extension and RNase protection assays. A TATA box-like motif (TTATTAT) was located between -34 and -29 and a reverse CCAAT box (ATTGG) was located between -78 and -74, relative to the transcription start site. To begin studying mechanisms underlying transcriptional regulation of the mGPAT gene, chimeric luciferase (LUC) plasmids containing serial deletions, from -1447 to -38, of the 5'-flanking region of the murine mGPAT gene were prepared and transfected into 3T3-L1 cells. The fusion construct -1447 GPAT.LUC showed high promoter activity and deletions to -1353, -747, -322, and -86 did not markedly change the promoter activity. With all constructs, luciferase activity was 2-fold higher when plasmids were transfected into 3T3-L1 adipocytes. However, deletion of sequences between -86 and -55 resulted in a 9-fold decrease in LUC activity in both preadipocytes and adipocytes. Deletion of sequences between -55 and -38 did not alter promoter activity. DNase I footprint analysis revealed a protected region between -95 and -65 which included the putative CTF/NF1 binding site. Electrophoretic mobility shift assays demonstrated a single protein-DNA complex formation. Oligonucleotides synthesized according to the CTF/NF1 consensus sequence or the adenovirus NF-1 site showed a different and more complex pattern of protein-DNA interaction and were not able to compete away the mGPAT promoter-protein complex, indicating that a distinct protein was bound to -86/-55, a region important for the basal promoter activity in 3T3-L1 cells. Luciferase activity was increased 2.8- and 8-fold when adipocytes stably transfected with -322 GPAT.LUC were treated with 5 and 25 mM glucose, respectively, in the presence of 10 nM insulin. These results indicate that carbohydrate-responsive sequences are located within -322 base pairs of the mGPAT promoter.