The glutathione cycle: Glutathione metabolism beyond the γ-glutamyl cycle
Abstract
Glutathione was discovered in 1888, over 125 years ago. Since then, our understanding of various functions and metabolism of this important molecule has grown over these years. But it is only now, in the last decade, that a somewhat complete picture of its metabolism has emerged. Glutathione metabolism has till now been largely depicted and understood by the γ-glutamyl cycle that was proposed in 1970. However, new findings and knowledge particularly on the transport and degradation of glutathione have revealed that many aspects of the γ-glutamyl cycle are incorrect. Despite this, an integrated critical analysis of the cycle has never been undertaken and this has led to the cycle and its errors perpetuating in the literature. This review takes a careful look at the γ-glutamyl cycle and its shortcomings and presents a “glutathione cycle” that captures the current understanding of glutathione metabolism. © 2018 IUBMB Life, 70(7):585–592, 2018
INTRODUCTION
Glutathione was discovered in 1888 by J. de Rey-Pailhade. The composition of glutathione was established much later, in 1929, through the efforts of Gowland Hopkins as well as Hunter and Eagles who could show that glutathione was a tripeptide of glutamate, cysteine and glycine. The final structure of glutathione as γ-glutamylcysteinyl glycine was determined by Pirie and Pinhey and confirmed through chemical synthesis by Harington and Mead in 1935 1. Since then a lot of work has been carried out in identifying and characterizing the enzymes involved in the metabolism of glutathione. By the 1970s, a fair understanding of the activities of several enzymes that were involved in the biosynthesis and metabolism of glutathione became available. Integrating the different enzymatic reactions, Meister, who had contributed immensely to the field of glutathione biochemistry, came up with a seemingly compelling proposal that appeared to nicely integrate the biosynthesis and degradation of glutathione 2. This sequential reaction of six enzymes constituted the γ-glutamyl cycle whose main purpose was proposed to be in the transport of amino acids across the membrane 3. Despite the elegance of the cycle, the proposed function of the cycle in amino acid transport has subsequently proved incorrect. Furthermore, new understanding on the transport of amino acids and glutathione, identification and characterization of the genes for the enzymes that were previously unidentified in the cycle, identification of γ-glutamyl acting enzymes unlinked to glutathione metabolism, as well as discoveries of new enzymes for the degradation of glutathione, has revealed that the steps relating to the degradation of glutathione, and downstream of the initiation of degradation are now quite different from what was originally proposed 4. Surprisingly, the sequence of reactions and even the pictorial representation of the cycle continue to persist in many textbooks of biochemistry as well as metabolic pathway databases such as Metacyc. Therefore, there is a necessity that the cycle is replaced with a new cycle that captures the new discoveries, while eliminating the earlier errors. This review briefly delineates the γ-glutamyl cycle as it was presented by Meister, discusses the shortcomings of the cycle, and finally presents a “glutathione cycle” that represents the current understanding of glutathione metabolism. Although glutathione is involved in many more reactions such as glutathionylation of proteins, neutralization of superoxides, and detoxification of metabolites through conjugation, they have been discussed in several exhaustive reviews 5, 6, and these aspects of glutathione are not discussed. Only those aspects that are central to the metabolism of glutathione and that were previously part of the γ-glutamyl cycle have been the focus of this article.
THE γ-GLUTAMYL CYCLE AS PROPOSED BY MEISTER
The γ-glutamyl cycle as proposed by Meister (Fig. 1) begins with the biosynthesis of glutathione 2. The biosynthesis involves the cytosolic, non-ribosomal synthesis of this unusual tripeptide by the action of two ATP-dependant enzymes, glutamate–cysteine ligase (GCL) and glutathione synthase (GS).The first enzyme, GCL, which catalyzes the formation of γ-glutamyl-cysteine from glutamate and cysteine in the presence of ATP is the rate limiting step in the biosynthesis. The second enzyme, GS, involves in the ligation of γ-glutamyl-cysteine to glycine in another ATP-dependant reaction to yield γ-glutamylcysteinyl glycine (or glutathione). Following biosynthesis is the initiation of degradation glutathione. In this cycle, the degradation is carried out by γ-glutamyl transpeptidase (γ-glutamyl transferase, GGT), the only enzyme known at that time to degrade glutathione. GGT has a transpeptidation activity that involves in the transfer of the γ-glutamyl group to the amino acid to form a γ-glutamyl amino acid. In the γ-glutamyl cycle, the primary GGT activity that was invoked was the transpeptidase activity, and therefore, the role of the enzyme was to primarily carry out transpeptidation. The transpeptidation reaction involving the amino acids leads to the eventual transport of the amino acid across the membrane in the form of γ-glutamyl amino acids. Furthermore, the model also implied that the plasma membrane bound GGT, which had it active site facing outward facilitated the translocation of the γ-glutamyl amino acid across the membrane into the cytosol. In the fourth enzymatic step, the γ-glutamyl amino acid that is released into the cytosol is acted on by the enzyme, γ-glutamyl cyclotransferase (γ-GCT) to yield 5-oxoproline (pyroglutamic acid) while releasing the amino acid. The next enzymatic reaction involved the ATP-dependant 5-oxoprolinase enzyme that acted on 5-oxoproline to yield glutamic acid. In the sixth step, the cysteinylglycine would be cleaved by cellular peptidases to yield cysteine and glycine. A key player in the entire cycle was the enzyme GGT. The enzyme is predominantly found on kidney tubules, but is also found in the choroid plexus and in the ciliary body and the lens. The kidney is where a lot of amino acid uptake takes place and was one of the reasons that led Meister to propose that amino acid transport was the key function of the cycle 3.
The γ-glutamyl cycle as proposed by Orlowski and Meister in 1970. The red box indicates the part of the cycle that has now been shown to be incorrect.
THE γ-GLUTAMYL CYCLE ALMOST 50 YEARS LATER: A CRITICAL ANALYSIS
The γ-glutamyl cycle was proposed in 1970, 48 years ago. Since then, we have a much better understanding of the functions of glutathione. Glutathione is present in all eukaryotic organisms, from yeast to man. It is only absent in the amitochondrial protozoans 7. It is also an essential metabolite, since in the absence of glutathione biosynthesis, embryonic lethality has been observed in mouse and plants 8-10. In yeasts, also, in the absence of glutathione, growth ceases, unless glutathione is provided from the outside 11, 12. The essential function of glutathione lies in its requirement in iron–sulfur cluster biogenesis 13, 14. The other roles relate to its role in redox properties, detoxification, oxidative stress response, and function in storage of essential cysteine. In addition to these functional aspects, the last decade has seen important advances in our understanding of glutathione metabolism. Some of the missing enzymes of glutathione metabolism have now been identified, that includes the γ-GCT enzyme 15 and the cys–gly dipeptidases 16-18 that were referred to, but not identified at the time the cycle was proposed. In addition, new enzymes involved in degradation have been uncovered, which include the ChaC family of enzymes 19, and the Dug2/Dug3 protein in yeast 20, 21. One also now has a better understanding of amino acid and glutathione transport in different organisms 22. This has led to the realization that while the biosynthesis of glutathione as described in the γ-glutamyl cycle is correctly represented, the latter half of the cycle that relates to the degradation of glutathione and the role of GGT has dramatically altered. These aspects are discussed below before the new cycle based on the current understanding is presented.
GGT Does Not Have a Role in Amino Acid Transport
A key postulate of the γ-glutamyl cycle was that GGT was involved in the transport of amino acids across the plasma membrane. However, work from several groups, even during Meister's lifetime could not confirm the role of GGT in the transport of amino acids 23. Although the GGT enzyme was first discovered as an enzyme that hydrolyzes glutathione to glutamate and cysteinyl-glycine, “glutathionase” 24, it was later shown to have a transpeptidation reaction 25 where it could transfer the γ-glutamyl moiety to the amino acid, and in lieu of the hydrolysis reaction, the enzyme would use the transpeptidase activity leading to the formation of a γ-glutamyl amino acid (Fig. 2). In the cycle, the proposed role of GGT was that it functions in the latter role, with the subsequent translocation of the γ-glutamyl amino acid dipeptide across the membrane into the cell. Eventually, the amino acid is released and glutathione also degraded into its constituent amino acids. Glutathione can then be resynthesized for the transport to commence again. The process required a significant expenditure of energy with a requirement of three ATP molecules and was one of the concerns that were raised about the function of this cycle, especially when less energetically intensive modes of amino acid transport existed. A second major concern was that the proposed cycle could not address how the cycle took into account the specificity of amino acid transport 26, 27. These reservations were eventually proven to be true, and with time, evidence accumulated to indicate that GGT did not play a role in amino acid transport.
The hydrolytic and transpeptidase reactions of γ-glutamyl transpeptidases.
One of the evidences that went against the cycle hypothesis for amino acid transport and which emerged from several laboratories was that the transport of amino acid did not correlate with the levels or activities of GGT 28, 29. Inhibition of the GGT enzyme by neither specific inhibitors nor antibodies interfered with the transport of amino acids. This was also confirmed in patient-derived fibroblast cell lines deficient in GGT activity. Even the transport of those amino acids that were supposed to be good acceptors for the transpeptidation reaction was not affected by inhibiting GGT. Thus, for example, cystine is a very good acceptor for GGT; however, it was observed that at the blood brain barrier microvessels, where GGT was more abundant, inhibiting the activity did not affect the transport of cystine.
The other important piece of data came from the kinetic properties of the enzymes. The two enzymatic activities of GGT were hydrolysis and transpeptidation. The Km (The Michaelis constant) for glutathione for the hydrolysis reaction was in the micromolar range, while the Km toward methionine as the acceptor in the transpeptidation reaction was in the millimolar range 30, 31. However, the concentrations of glutathione and amino acids in the plasma are in the micromolar range. The concentrations of the amino acids was thus 100-fold lower than the Km. Thus, the kinetic parameters were clearly compatible only with the hydrolysis reaction since the concentrations of amino acids were too low to allow the transpeptidation reaction to occur in vivo.
A third line of evidence was based on the observations of patients with GGT deficiency as well as in GGT deficient and knockout mice. These patients and the mice do not exhibit aminoacidurias, but rather displayed “glutathionuria,” which in the case of mouse, showed 2,500-fold higher levels of glutathione in the urine 32-34.
Even with the convincing nature of these findings, a rejection of a role for the γ-glutamyl cycle in amino acid transport never made it to the mainstream literature. This is despite the fact that in a review in 1980 23, some of these issues were discussed. This was due partly to the still incomplete understanding of glutathione and amino acid transport, but was also a result of the strong domination in the field of those in favor of the cycle that resulted in the limited reach of these opposing ideas 35. The counter-argument that was used by those in favor of the cycle was that the γ-glutamyl cycle was only one of the methods that transported amino acids and did not necessarily indicate that other transport systems did not exist, or that the cycle transported all amino acids 36.
The Translocation of γ-Glutamyl Amino Acids (or Glutathione) by GGT Does Not Exist
One of the key aspects of the proposed γ-glutamyl cycle was that the γ-glutamyl amino acids should be transported into the cell. However, a transporter for importing either γ-glutamyl amino acid or, the intact tripeptide glutathione, γ-glutamylcysteinyl glycine has not been identified in mammalian cells. Early studies on glutathione transport had initially suggested that a plasma membrane glutathione transporter exists that can lead to inward movement of glutathione from the extracellular medium to the cytoplasm. However, despite exhaustive efforts, these have not been identified in mammalian cells. The only reports on the identification of such transporters in rat turned out to be artifactual, since the identified protein turned out subsequently to be an Escherichia coli protein 37. The transporters of glutathione that are located on the plasma membrane are those that can efflux glutathione and glutathione conjugates outside. These are the multidrug resistance-associated protein (MRP) and related family of ATP binding cassette (ABC) transporter 38. However, the principal substrate of these pumps is not reduced glutathione but oxidized glutathione and glutathione conjugates 22.
The GGT enzyme is a typical ectoenzyme that contains an uncleaved N-terminal signal sequence. This segment forms a hydrophobic a-helix that spans the membrane once and anchors the catalytic domain of GGT to external surface of the membrane. It is now clear that the principal function of GGT is in salvaging effluxed glutathione conjugates, oxidized glutathione, or even reduced glutathione 39. The breakdown products are glutamate and either cys-gly or bis(cys-gly). These can be either transported as peptides by the peptide transporters 40, or further broken down into the constituent amino acids cysteine and glycine or cystine and glycine, which can be taken up by the cells through specific amino acid transporters.
Thus, glutathione is transported across the plasma membrane in mammalian cells by degradation outside at the surface, transport of the amino acids or dipeptides into the cytoplasm, followed by resynthesis inside the cytosol.
An alternate possibility that would be compatible with the proposed cycle is that the GGT enzyme carries out the transpeptidation reaction at the surface, and that the resulting γ-glutamyl amino acid that is formed, is then transported into the cell. However, this is also not supported by evidence. Although γ-glutamyl-cysteine ethyl esters are able to permeate into cells 41, γ-glutamyl-cysteine is itself not transported. The observation that γ-glutamyl-cysteine feeding to humans can increase glutathione levels 42 is likely to be a consequence of the degradation of the γ-glutamyl-cysteine by GGT, since it is known that GGT can act on the γ-glutamyl peptides 43. There is no evidence for the direct transport of γ-glutamyl amino acid dipeptides. The two peptide transporters, PepT1 and PepT2, are not known to transport γ-glutamyl peptides 44. Mice knockouts deleted for the high affinity PepT2 transporter revealed higher levels only of cysteinyl-glycine in the plasma 40.
In microbial systems (yeast and bacteria) a specific glutathione uptake process exists for the uptake of intact glutathione. The function in microbial systems is also to salvage glutathione from the lysing cells. However, here too, although yeasts have a specific plasma membrane glutathione transporter, and a peptide transporter, neither can transport γ-glutamyl-cysteine 45-47.
The Enzyme γ-GCT That Acts on γ-Glutamyl Amino Acids Is Not Central to Glutathione Metabolism
The γ-GCT enzyme acting on γ-glutamyl amino acids was laid out as one of the essential steps in the cycle. The enzyme activity was detected in extracts, but the enzyme was not identified for a long time. Only recently has the gene been identified as C7ORF24, and the encoded protein (γ-GCT) characterized 15. Once the gene was identified, it became clear from sequence comparisons that the enzyme was indeed absent in yeasts, plants, bacteria as had been suggested by biochemical experiments. Thus, γ-GCT is present only in mammals and does not fit into a universal γ-glutamyl cycle 18. The role of this enzyme was to act on the γ-glutamyl amino acids that were purportedly generated by the transpeptidation reaction of GGT. However, as the transpeptidation reaction does not occur to any significant extent in vivo, there must be some other source of γ-glutamyl amino acids that would justify the presence of such an enzyme exclusively in mammalian cells. It is possible that still unidentified pathways unlinked to glutathione metabolism could be a source of γ-glutamyl amino acids. From glutathione metabolism itself, one possible source of γ-glutamyl amino acids appears to be through the secondary reaction of the biosynthetic enzyme, GCL. The mammalian enzyme, unlike the bacterial enzyme, has a broad specificity, with the ability to accept other amino acids instead of cysteine 48. Thus, in case cysteine is limiting, the GCL enzyme could generate other γ-glutamyl amino acids. These γ-glutamyl amino acids that are formed could then be recycled by the GCT enzyme.
A second proposed reason for the existence of the γ-GCT enzyme is the activity toward γ-glutamyl-cysteine, which enables it to act as a regulator of glutathione synthesis. γ- Glu-cys is a substrate of γ-GCT as well as GS. The activity of GS toward γ-glu-cys is with much greater catalytic efficiency and also with higher affinity than γ-GCT. Thus, under normal conditions, any γ-glu-cys formed will be directed toward glutathione synthesis, but with increased cysteine and glutamate, any excess γ-glu-cys formed could be acted on by γ-GCT to yield cysteine and 5-oxoproline 49.
The Source of 5-Oxoproline in the Cycle Is Glutathione (and Not the Proposed Intermediate γ-Glutamyl Amino Acids)
In the classical γ-glutamyl cycle, 5-oxoproline was generated by the action of γ-GCT on γ-glutamyl amino acids. Thus, the role of 5-oxoprolinase nicely fit into the cycle. However, as mentioned above, most organisms lack a γ-GCT, and yet, they contain a 5-oxoprolinase. This suggests that there must be alternate routes for the generation of cytosolic 5-oxoproline in cells.
The recent discovery of a new class of glutathione degrading enzymes, that is, the ChaC family of γ-GCTs that act specifically on glutathione fills this critical gap in our understanding. This family of glutathione degrading enzymes acts on glutathione to directly yield 5-oxoproline and cys-gly. Members of the ChaC family are found from bacteria to human 19. Humans and mice contain two such enzymes, ChaC1 and ChaC2 50. The ChaC1 enzymes have higher catalytic efficiency, are induced to high levels during development and under stress conditions, while the ChaC2 enzymes have lower efficiency and in mammals are constitutive. Both ChaC1 and ChaC2 act on reduced glutathione. Although the reaction products of both enzymes are the same, the kinetics is different and the enzymes serve two distinct purposes. However, irrespective of the differences in physiological roles, the enzymes yield both cys-gly and 5-oxoproline. This continuous production of 5-oxoproline explains the presence of 5-oxoprolinase in organisms that lack γ-GCT. There are also other sources of 5-oxoproline in the cell unlinked to glutathione metabolism 51. One source linked to glutathione metabolism is by a secondary reaction of the first biosynthetic enzyme, GCL. In the absence of cysteine, GCL with glutamate and ATP forms γ-glutamyl phosphate that can cyclize to 5-oxoproline 52, a reaction that could be prominent in some pathological situations 53.
THE PROPOSED “GLUTATHIONE CYCLE”
Based on what has been described above, it is clear that the γ-glutamyl cycle is incorrect not only in the functional role as an amino acid transporter but also in describing the biochemical fates of glutathione. This is especially so in the light of the recent data that has emerged on glutathione degradation. A new cycle becomes apparent with these new discoveries and we refer to this as the “Glutathione cycle” 4 (Fig. 3). A change in the terminology also ensures the focus on glutathione rather than the γ-glutamyl group. This is important because there are other enzymes acting on γ-glutamyl groups that have no relationship with glutathione metabolism. The γ-GCT as discussed earlier is only found in mammals and is only peripherally involved with the cycle. In addition, there is another γ-GCT, γ-glutamylamine cyclotransferase, that acts on γ-glutamyl amines converting them into 5-oxoproline and the amine 54 (Fig. 3). These γ-glutamyl amines are formed by the metabolism of proteins crosslinked by transglutaminase and are reactions unlinked to glutathione metabolism 55.
The glutathione cycle.
In the Glutathione cycle, the first two steps are the biosynthesis of glutathione. The biosynthesis occurs through the two sequential enzymes, γ-glutamyl-cysteine synthase (or GCL) that makes γ-glutamyl-cysteine from glutamate and cysteine in an ATP-dependant reaction and GS that makes γ-glutamylcysteinyl glycine (or glutathione) in a second ATP-dependant reaction involving γ-glutamyl-cysteine and glycine. The glutathione formed can then be degraded by the cytoplasmic ChaC-family of γ-GCTs that are glutathione-specific degradation enzymes and yield 5-oxoproline and cysteinyl glycine upon hydrolysis of glutathione. The 5-oxoproline formed is cleaved by a 5-oxoprolinase to yield glutamate, while cysteinyl glycine is cleaved by specific cys-gly peptidases to yield cysteine and glycine. The glutamate, cysteine and glycine that are thus released, if needed, can be ploughed back into the synthesis of glutathione. The glutathione formed through the biosynthetic reaction is the reduced form that can form oxidized glutathione with which it equilibrates and determines the redox environment. The oxidized glutathione (as well as glutathione conjugates) can also be effluxed out of the cytosol by specific transporters that are specific for the oxidized glutathione or conjugates during oxidative stress or xenobiotic stress and are also a means of regulating the cellular redox potential 38, 56. The effluxed glutathione and conjugates can be salvaged by the membrane bound GGT [with the active site facing outside (away from the cytosol)]. The GGT enzyme yields glutamate and cys-gly or bis-cys-gly. The glutamate can be transported back inside as can the cys-gly. Alternately, the cys-gly or bis-cys-gly can be broken down by membrane bound peptidases to yield cysteine (or cystine) and glycine. All three of these, cysteine, cystine, and glycine can then be transported inside the cell by specific transporters.
The above constitutes the universal glutathione cycle that would be found in all organisms. However, there will also be organism-specific variations based on the locations of the enzymes and the presence of additional or variant enzyme systems, some examples of which are briefly discussed below.
Yeasts and Fungi
Unlike mammalian cells, in the yeast Saccharomyces cerevisiae, the GGT enzyme is located on the vacuolar membrane. The active site faces the vacuolar lumen enabling it to act on the vacuolar glutathione pools, and like the mammalian GGT, it does not act on the cytoplasmic pools of glutathione. In yeasts, additionally, the glutathione conjugate efflux pumps are also located on the vacuolar membrane pumping the glutathione disulphides and the glutathione conjugates into the vacuole. Furthermore, yeasts and fungi in addition to containing a cytoplasmic ChaC member (GCG1) 19, also contain a second yeast glutathione degrading enzyme, Dug2p/Dug3p in the cytosol that belongs to the N-terminal nucleophilic hydrolases (to which GGT belongs) and cleaves glutathione to give glutamate (rather than 5-oxoproline) and cys-gly 20, 21. The cytoplasmic cys-gly peptidase, Dug1p that belongs to the M20A family (and is a CNDP2 ortholog) is the only major cys-gly peptidase in the cytoplasm of yeast 17.
Mammals
Glutathione biosynthesis in mammals is catalyzed by a GCL that has two subunits, a catalytic subunit, GCLC, and regulatory or modulatory subunit, GCLM. In terms of glutathione degradation, in humans and mouse, there are two ChaC-like proteins in the cytosol, ChaC1 and ChaC2 19, 50. ChaC1 is upregulated during development and during stress conditions while ChaC2 is constitutive and likely to be responsible for the continuous slow degradation of glutathione, where glutathione serves as a reservoir for the rate limiting amino acid, cysteine. Mammalian systems also contain multiple enzymes for cys-gly peptide degradation. These are the cytoplasmic M20A metallopeptidase, CNDP2, the cytoplasmic M19 metallopeptidases that are the leucine aminopeptidases 16, 17 and the plasma membrane peptidases belonging to the M1 family of metallopeptidases aminopeptidase M 57 and the M19 family of metallopeptidases (DPEP1) 58. Although CNDP2 and Aminopeptidase M are specific for cys-gly, DPEP1 can cleave both cys-gly and oxidized cys-gly 59, 60. The mammalian system also has the presence of the γ-GCT that specifically acts on γ-glutamyl amino acids, and may have only a peripheral association with the glutathione cycle.
Plants
In the plant Arabidopsis thaliana, unlike other organisms, the glutathione synthesis partly takes place in the chloroplast 6. Plants have multiple GGTs, and while some of these are located in the vacuole, like in yeast, there are GGT members on the plasma membrane as well 61, 62. MRPs or glutathione conjugate pumps are also vacuole localized 63. Like yeast, although plants lack γ-GCTs that act on γ-glutamyl amino acids, some plants are known to contain variations of glutathione, such as γ-glutamylcysteinyl serine and γ-glutamylcysteinyl-β-alanine (homoglutathione) 6. The significance of these variant structures is still not understood. A. thaliana was also suggested to have a GGT-independent pathway of glutathione degradation 64. These were identified as the cytosolic ChaC family of glutathione degrading enzymes 65, 18, and although Cys-gly peptidases of the M20A family seen in yeasts and humans are lacking in plants, they have been shown to contain a leucine aminopeptidase belonging to the M17 family of metallopeptidases that can cleave cys-gly peptidases 18.
Bacteria
Not all bacteria contain glutathione. However, E. coli and several other gram-negative bacteria contain glutathione 66. In bacteria, the GGT enzyme is located in the periplasmic space, thus again acting on non-cytosolic pools 67. Similar to the other organisms, there is also a member of the ChaC family that acts on cytosolic pools. Bacteria also lack the M20 family of cys-gly peptidases, but contain members of the M17 family seen in plants that can act on cys-gly peptides 67.
The above are some of the examples of organisms that have been relatively well studied in terms of glutathione metabolism. While the core glutathione cycle remains the same in all these organisms, there are variations in each of these cases. As more organisms are studied, it is likely that we will see more variants while the central cycle as proposed (Fig. 3) remains the same.
In concluding, the glutathione cycle that we have presented here captures the current understanding of glutathione metabolism. Being the most abundant non-protein thiol compound with numerous functions within the cell, glutathione plays vital roles in cellular physiology, and therefore its metabolism becomes very important. We hope that a better appreciation of glutathione metabolism through this cycle will further facilitate our understanding of this important metabolite.
ACKNOWLEDGEMENTS
This work was supported by grant-in-aid funding from the Department of Science and Technology (DST), India (project no: SB/SO/BB/017/2014) (to A.K.B.) and an INSPIRE research fellowship (no. IF150103) from DST (to S.Y.). The authors thank Dr. Norman Curthoys and Dr. P. R. Krishnaswamy for critical comments on the manuscript.
