Characterization of genetically modified mice for phosphoglycerate mutase, a vitally-essential enzyme in glycolysis

Glycolytic metabolism is closely involved in physiological homeostasis and pathophysiological states. Among glycolytic enzymes, phosphoglycerate mutase (PGAM) has been reported to exert certain physiological role in vitro, whereas its impact on glucose metabolism in vivo remains unclear. Here, we report the characterization of Pgam1 knockout mice. We observed that homozygous knockout mice of Pgam1 were embryonic lethal. Although we previously reported that both PGAM-1 and -2 affect global glycolytic profile of cancers in vitro, in vivo glucose parameters were less affected both in the heterozygous knockout of Pgam1 and in Pgam2 transgenic mice. Thus, the impact of PGAM on in vivo glucose metabolism is rather complex than expected before.


Introduction
Glucose is utilized as energy source, from bacterias to higher eukaryotes. Glycolysis provides energy supply in the form of ATP through oxidation of carbon atoms in glucose. As it also supports the synthesis of macromolecules [1], glycolysis is evolutionally conserved as one of vitally essential metabolisms.
Dysregulation in glycolysis, either the pathological enhancement or impairment, is closely related to human disease states. For example, most of cancer cells display enhanced glycolytic feature, known as the Warburg effect. On the other hand, impaired glycolysis is much involved in dysfunction in various tissues and degenerative disorders [2,3]. In humans and other vertebrates, glucose is delivered to various tissues through circulating blood. Even during prolonged fasting (no calorie feeding), blood sugar levels are minimally maintained, which enables various tissues and cells, e.g. neurons in brain, to utilize glucose from circulating blood. The patients with impaired uptake of glucose into cells display higher glucose levels in blood, a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 known as diabetes or impaired glucose metabolism. Indeed, diabetes are observed both in human glycolytic enzyme deficiencies [4,5] and in knockout (KO) mice for glycolytic enzymes [6][7][8][9].
Besides its role in glycolysis, PGAM would play a key role connecting glycolysis to physiological homeostasis. PGAM is post-translationally regulated by several modification, including the phosphorylation, acetylation and ubiquitination by environmental stimuli or under stress conditions [13][14][15]. Moreover, PGAM activity modulates mitochondrial function [16] and the pentose phosphate pathway [17], implicating its involvement in the defense against oxidative stress [12]. Noteworthy, both PGAM-1 and -2 support cancerous cell growth through its nonenzymatic function. PGAM interacts with Chk1 kinase under oncogenic Ras expression, followed by boosts of glycolytic profiles [18].
However, the global impacts of PGAM in vivo is still unclear. Here we characterized the profiles of glucose metabolisms in mice with ablation of Pgam1 or global overexpression of Pgam2 in vivo. Homozygous knockout of Pgam1 is embryonic lethal, while its global heterozygous knockout mice displayed comparable glucose parameters in vivo, compared to those in wild type. Moreover, global overexpression of Pgam2 unlikely affects the in vivo profile for glucose metabolisms. Thus, the impact of PGAM on in vivo glucose metabolism is rather complex than expected before.

Ethical statements
All procedures for animal experiments were approved and conducted in accordance with the principles and guidelines of the Animal Care and Use Committees of Kyoto University Graduate School of Medicine (Med Kyo 15556 and Med Kyo 15558).

Generation of mouse models
Pgam2-Tg mice were generated previously [13]. Pgam2-Tg is a strain of transgenic C57BL/6 mice that overexpresses the Pgam2 with a 3xFLAG tag under the cytomegalovirus immediateto-early enhancer element and chicken β-actin promoter (CAG) [19].
Generation of the conditional knockout mouse of Pgam1 is previously reported [18]. For the study of whole body Pgam1 knockout mice, Pgam1 +/mice were generated by crossing between Pgam1 flox/+ mice and CAG-Cre Tg mice, which ubiquitously express Cre recombinase under the CAG promoter. Subsequently, Pgam1 +/-; CAG-Cre micewere crossed with wild-type mice and the genotypes of progenitor mice were analyzed by using PCR. Resulting Pgam1 +/mice without CAG-Cre transgene were applied for further experiment.

Mouse experiments
Intraperitoneal glucose tolerance test (IPGTT) was performed as previously described [20]. After 16 h of fasting, 1.5 g/kg body weight glucose was intraperitoneally injected into the mice. The blood glucose level of the mice was measured at 0, 15, 30, 60, and 120 min using a Glutest sensor (Sanwa Kagaku Kenkyusyo Co., Ltd. Aichi, Japan). For the high-fat diet study, 8-weekold male mice were fed diets with 60% of calories from fat (high-fat diet, Research Diets D12492) for 10 weeks. Body weight was measured every week. Finally, IPGTT was performed when the mice 18 weeks old. Blood tests for several biological markers were performed by Unitech Co. (Kashiwa, Japan). Muscle muss and body fat mass in mice were measured using magnetic resonance imaging (Latheta LCT-100, Hitachi).

RNA analysis
Total RNA was extracted with TRIzol (Invitrogen). cDNA pools were prepared by the Rever-Tra Ace qPCR RT kit (Toyobo, Osaka, Japan). Real-time quantitative PCR was carried out using Thunderbird SYBR qPCR mix (Toyobo). The level of expression of target gene was monitored by the Thermal Cycler Dice Real-Time system (Takara Bio., Kusatsu, Japan) and. Gene expression levels were normalized to Rpl13a mRNA and presented as values relative to wildtype mice.
The primer sequences are shown as below.

Statistical analysis
All data are expressed as the mean ± standard error of the mean (SEM) Comparisons between two independent groups were analyzed using an unpaired Student's two-tailed t-test.

Heterozygous Pgam1 knockout mice were viable with reduced PGAM acitvity
Although Pgam1 homozygous KO mice were embryonic lethal, Pgam1 +/− mice are viable (Fig  2A). The mRNA levels of Pgam1 were decreased by approximately 50% in several tissues of Pgam1 +/− mice (Fig 2B, upper panel), whereas the levels of Pgam2 mRNA were not affected in the examined tissues from these mice (Fig 2B, lower panel). PGAM1 and PGAM2 exhibit overlapping but distinct tissue distributions; PGAM1 is dominantly expressed in several tissues including the liver, white adipose tissue, aorta, and brain, whereas PGAM2 is found mainly in the muscle. The other tissues including the lung, heart, skin, and bone express both isoforms https://doi.org/10.1371/journal.pone.0250856.g001 [11,13]. Consistently, the enzymatic activities of PGAM in Pgam1 +/− mice were significantly decreased in several tissues including the cerebrum, kidney, and skin, or modestly decreased in the lung, liver, and white adipose tissue (WAT), but not in the heart, brown adipose tissue (BAT), and muscle, compared to those in control ( Fig 2C). As several posttranslational modification for PGAM is heavily operating [26], unknown compensatory mechanisms by such modification may preserve its enzymatic activity in Pgam1 heterozygous knockout mice.

In vivo parameters for glucose metabolism were less affected by heterozygous Pgam1 knockoout
Recently we reported that PGAM status deeply affects glycolytic profiles in cancer cells with oncogenic Ras expression, but not in standard cells in vitro [18]. We examined the profiles of glycolytic mRNAs in the tissues of Pgam1 +/− mice. We observed no significant difference in the profiles of glycolytic mRNAs between wild-type and heterozygous knockout of Pgam1 mice (Fig 3A).
It was reported that brain-specific knockout of Glut4 in vivo, glucose transporter in glycolysis, displayed impaired glucose tolerance and insulin sensitivity [27]. As glycolysis is essential energy source in all tissues, it is possible that the impaired glycolysis in the limited organs could affect glucose metabolisms in vivo [28]. To examine this possibility, we next examined glucose metabolism of Pgam1 +/− mice in vivo, as mice with one allele of glucokinase displayed diabetic phenotype [8]. In comparison to wild-type mice, Pgam1 +/− mice displayed similar profiles for in vivo glucose metabolism (intraperitoneal glucose tolerance test (IPGTT) and serum glucose level), in addition to body weight, visceral-fat weight, muscle mass, total-cholesterol, and other biochemical blood markers (Fig 3B-3E). Collectively, in vivo parameters for glucose metabolism were less affected by heterozygous Pgam1 knockout.

Global overexpression of Pgam2 does not affect glucose tolerance in vivo
Next, we evaluated the impact of PGAM overexpression in vivo. We previously reported that heart-specific Pgam2-transgenic (Tg) mice displayed almost normal glycolytic features in the heart [16]. However, it remains unclear whether global PGAM overexpression affects glycolytic profiles in vivo. As it has been demonstrated that the overexpression of either PGAM isoform confers similar physiological impact [11][12][13], we evaluated the in vivo parameters of glucose metabolism in Pgam2-Tg mice, in which a Pgam2-FLAG transgene is driven by the CAG promoter. Pgam2-Tg mice exhibited a significant increase of PGAM protein in the whole body [18]. Pgam2-Tg mice also displayed their similar features with respect to body weight and IPGTT both under physiological ( Fig 4A) and high fat-induced obesity conditions (Fig 4B  and 4C), compared with those of wild-type mice. Thus, the parameters for glucose metabolism are less affected by PGAM overexpressing status in vivo under physiological conditions.

Discussion
Here we reported on the profiles of in vivo glucose parameters in genetic models of glycolytic enzyme PGAM. Consistent to the findings in the ablation of other glycolytic enzymes, the embryonic lethality of homozygous knockout of Pgam1 supports the notion that glycolysis is vitally essential metabolism in vivo. Unexpectedly, however, Pgam1 +/− and Pgam2 -Tg mice displayed comparable profiles of in vivo glucose parameters, as far as we tested.
Several genetic mutations of PGAM with its impaired enzymatic activity, have been reported in human PGAM deficient patients [29][30][31]. However, little is reported on the link between human PGAM deficiency and diabetes, except one case report on the adult patient of PGAM deficiency (51 years old) accompanied with type 2 diabetes [32]. Thus, it is still controversial whether human PGAM deficiency causes diabetes or not. Here, we observed that the in vivo parameters for glucose metabolism are less affected in Pgam1 +/− mice. Moreover, Pgam2 -Tg in normal and high-fat diet conditions displayed the normal glucose tolerance. However, as Pgam1 +/mice retained substantial PGAM enzymatic activity in several tissue including liver, muscle, WAT and BAT, it would be worthy to investigate tissue-specific Pgam1 homozygous knockout or Pgam2 knockout model in the future.
Different from our observations in genetic model mice for PGAM, several reports suggest that impaired glycolysis results in diabetes or impaired glucose metabolism both in human patients [4,5] and in knockout (KO) mice for the other glycolytic enzymes [8,9]. The discrepancy between our study and the inactivation of the other glycolytic enzymes could be possibly due to the different impact of individual glycolytic enzymes on glycolytic pathway. Interestingly, the catalytic reaction of PGAM is not rate-limiting step during glycolytic pathway under standard conditions, even when PGAM activity is reduced by approximately half of normal mean value both in human PGAM deficient patient [32] and in Pgam1 heterozygous knockout mice (this study) [18]. As PGAM enzyme activity is most prominent among glycolytic enzymes [30], the residual PGAM activity may be sufficient to maintain glycolysis in human or mice PGAM impairment. Interestingly, overall glycolysis is much affected if majority of PGAM proteins were degraded via ubiquitin-dependent proteolysis of PGAM under senescence-inducing stress [13], rather consistent to the embryonic lethality of Pgam1 −/− mice (in this study). In this setting, PGAM would be rate-limiting in glycolysis. Moreover, we recently reported that the global enhancement of glycolytic profiles in cancerous cells, the Warburg effect, is provoked in vitro by the cooperation of PGAM with Chk1 kinase under oncogenic Ras expression [18]. Besides cancers, it is well known that the active proliferation of T cells after immunological stimuli or stress is accompanied by the enhanced glycolysis [33]. Pgam1 deficient T cells are impaired in maintenance of such glycolytic enhancement, followed by immunological dysfunction [34]. Thus, the impact of PGAM on glycolytic regulation would be cellular-or tissue-context dependent. Indeed, the boosts of glyclolytic mRNAs were observed in Pgam2-Tg mice under chemical carcinogenic protocol [18], while those in Pgam2-Tg mice under standard conditions were comparable to control [18]. It is possible that the modulation of PGAM proteins would be required in vivo for its physiological impact, as several post-translational regulations of PGAM were reported in vitro [26].
Further investigation might provide valuable clue to understand the impact of PGAM on glucose metabolisms in vivo.