OUP user menu

Enterococcal-type glycopeptide resistance genes in non-enterococcal organisms

Robin Patel
DOI: http://dx.doi.org/10.1111/j.1574-6968.2000.tb09032.x 1-7 First published online: 1 April 2000


Although the emergence of vancomycin-resistant enterococci can be attributed, in part, to the increasing use of vancomycin in clinical practice, and glycopeptide use in animal husbandry, the origins of the enterococcal vancomycin resistance genes are not clear. The vancomycin resistance-associated genes in Enterococcus gallinarum, Enterococcus casseliflavus/flavescens, Lactobacillus spp., Leuconostoc spp., Pediococcus spp., and Erysipelothrix rhusiopathiae, are not the source of the high-level vancomycin resistance-associated genes in enterococci. There are, however, environmental organisms which have been found to have gene clusters homologous to the enterococcal vanA, vanB and vanC gene clusters; these include the biopesticide Paenibacillus popilliae, and, to a lesser extent, the glycopeptide-producing organisms Amycolatopsis orientalis and Streptomyces toyocaensis. Still, the exact sources of the enterococcal vancomycin resistance genes remain a mystery.

  • Enterococcus
  • Glycopeptide resistance
  • Vancomycin
  • Paenibacillus popilliae

Among the most dramatic and clinically worrisome examples of resistance to antimicrobial agents in recent years has been the emergence and spread of vancomycin resistance in enterococci. This increase poses important problems, including the death of available antimicrobial agents active against these organisms, in many cases, and the possibility that vancomycin resistance genes can be transferred to other Gram-positive bacteria, especially Staphylococcus aureus, coagulase-negative staphylococci (CNS), Streptococcus pneumoniae, Bacillus spp. and Corynebacterium spp. The transfer of vancomycin resistance to methicillin-resistant S. aureus (MRSA), methicillin-resistant CNS (MR-CNS), high-level penicillin-resistant S. pneumoniae and/or penicillin-resistant Corynebacterium spp. (e.g. Corynebacterium jeikeium) might compromise successful therapy for infections caused by these organisms.

Enterococci with VanA- or VanB-type high-level vancomycin resistance have acquired genetic material which leads to the manufacture of an abnormal ligase (in addition to the normal ligase described below) responsible for the synthesis of the depsipeptide D-alanyl-D-lactate (D-Ala-D-Lac) which is incorporated into a pentapeptide peptidoglycan cell wall precursor (UDP-MurNAc-L-Ala-D-Glu-L-Lys-D-Ala-D-Lac) to which vancomycin binds poorly. In contrast, in vancomycin-susceptible cells, vancomycin complexes with the D-Ala-D-Ala termini of normal pentapeptide peptidoglycan cell wall precursor (UDP-MurNAc-L-Ala-D-Glu-L-Lys-D-Ala-D-Ala) thereby inhibiting cell wall synthesis. Although this normal pentapeptide peptidoglycan cell wall precursor pathway is preserved in enterococci with VanA- or VanB-type vancomycin resistance, there is hydrolysis of some of its components.

VanA-type glycopeptide resistance is characterized by acquired inducible resistance to both vancomycin and teicoplanin. It is mediated by Tn1546 or closely related elements [1]. Tn1546 encodes nine polypeptides that can be assigned to different functional groups: Transposition functions, regulatory proteins (VanR and VanS), synthesis of depsipeptide D-Ala-D-Lac (VanH and VanA) and hydrolysis of precursors of normal peptidoglycan, at least in vitro (VanX and VanY); the function of VanZ is unknown [1]. The vanR, vanS, vanH, vanA and vanX genes are necessary and sufficient for the inducible expression of resistance to glycopeptides [1]. VanY and VanZ are accessory peptides and are not required for resistance. Genetic heterogeneity has been described in the vanA gene clusters of vancomycin-resistant enterococci (VRE) [2], and the vanA gene cluster has been found on the chromosome as well as on plasmids [3].

VanB-type glycopeptide resistance is characterized by acquired inducible resistance to various concentrations of vancomycin, but not to teicoplanin. The vanB gene cluster has homology to the vanA gene cluster, but has been less well studied. It consists of genes encoding polypeptides assigned to regulation of vancomycin resistance genes (vanRB and vanSB), synthesis of the depsipeptide D-Ala-D-Lac (vanHB and vanB), and hydrolysis of precursors of normal peptidoglycan, at least in vitro (vanXB and vanYB) [4]. An open reading frame vanW is also present and encodes a polypeptide of unknown function [4]. Our group has identified sequence variability in the vanB gene of different isolates of enterococci with VanB resistance [5]. The sequence variability observed in these isolates is consistent with the hypothesis that the spread of vancomycin resistance amongst Enterococcus spp. results not just from dissemination of a single clone, but also as the result of horizontal transfer of resistance genes from as yet undefined organisms [6].

High-level enterococcal vancomycin resistance is readily transferred from one enterococcus to another in the laboratory. Additionally, laboratory experiments have achieved the transfer of high-level vancomycin resistance from enterococci to S. aureus[7]. Vancomycin resistance has also been transferred in vitro by conjugation or transformation from enterococci to Streptococcus sanguis, Lactococcus lactis, Streptococcus pyogenes, and Listeria monocytogenes[8,9].

The origins of the vancomycin resistance-associated genes in enterococci are unknown, but the selective pressures favoring the emergence of VRE are more clear. Vancomycin was first licensed in the United States in 1958; however, it took until 1986 to document high-level vancomycin resistance in enterococci. Given the complexity of the mechanism of vancomycin resistance in enterococci, it is possible that it took time to gather together multiple individual genes required for the expression of resistance. This hypothesis is not supported by the multicentric nature of the emergence of VRE in the United States and Europe. On the contrary, the emergence of VRE in the late 1980s suggests that preassembled operons were transferred to enterococci, before or at that time, and selected for under the pressure of dramatically increased vancomycin use in clinical practice and glycopeptide use (e.g. avoparcin and orienticin) in animal husbandry. Prior to the late 1970s, there was little clinical use of vancomycin in the United States; thereafter, however, with the identification of Clostridium difficile as a cause of diarrhea and with the emergence of MRSA and MR-CNS, oral and intravenous vancomycin use, respectively, increased dramatically, likely contributing to the emergence of VRE, especially in the United States. In the United States, the acquisition of vancomycin resistance by multiply drug-resistant (e.g. ampicillin- and aminoglycoside-resistant) enterococci has resulted in VRE which are ‘very resistant enterococci’, which have (unfortunately) spread nosocomially, and have, in some instances, resulted in untreatable infections. In Europe, on the other hand, the use of avoparcin in animal husbandry has created a hospitable, but not necessarily a hospital, environment for the emergence of VRE. In contrast to VRE in the United States, European VRE are less likely to be multiply drug resistant and are less frequently cultured from hospitalized patients, presumably relating to their distinct origin, and perhaps to the absence of as yet uncharacterized human intestinal colonization-promoting factors present in VRE in the United States.

There are several reports of the enterococcal vanA and vanB genes being found in non-enterococcal bacteria isolated from humans. The vanA gene has been found in vancomycin-resistant clinical isolates of Cellulomonas turbata and Arcanobacterium haemolyticum (typically these organisms are vancomycin susceptible) isolated from the stools of two patients during an outbreak of VRE infection in London, England [9]. The vanB gene has been found in a vancomycin-resistant isolate of Streptococcus bovis isolated from a stool swab collected on admission from a patient as surveillance for VRE [10]. The vanA gene cluster has been identified in a Bacillus circulans isolate in a case of catheter-related infection [11]. In the aforementioned cases, the enterococcal vancomycin resistance genes appear to have been transferred into these isolates, and these organisms themselves are unlikely to have been the original source for the vancomycin resistance-associated genes in enterococci.

Two species of enterococci are intrinsically resistant to vancomycin: Enterococcus gallinarum and Enterococcus casseliflavus/flavescens. The vanC-1 gene in E. gallinarum and the vanC-2/3 genes in E. casseliflavus/flavescens encode a D-Ala:D-Ser ligase, which is associated with low-level vancomycin resistance (a D-Ala:D-Ala ligase gene is also present) [12]. In E. gallinarum, three enzymes are sufficient for vancomycin resistance: the VanC-1 ligase for production of D-Ala-D-Ser, the VanXYC protein for hydrolysis of D-Ala-D-Ala and for removal of D-Ala from UDP-MurNAc-pentapeptide ending in D-Ala, and VanT for D-serine synthesis [13]. There is only 37–39% deduced (partial) amino acid sequence identity between the VanC enzymes and VanA/B [14]; therefore the VanC enzymes are not the source of high-level, vanA- and vanB-associated vancomycin resistance in enterococci. As an aside, the origin of and the rationale behind the development of D-Ala:D-Ser ligases are not clear as this pathway does not appear to be metabolically advantageous for the bacteria and seems too ancient to be explained by selective pressure due to the use of glycopeptides in clinical or agricultural practice [15].

Two novel types of vancomycin resistance have recently been described in enterococci. VanD-type vancomycin resistance is distinct from VanA- and VanB-type vancomycin resistance and although associated with synthesis of the depsipeptide D-Ala-D-Lac, does not appear to be transferable [16,17]. VanE-type vancomycin resistance is low-level vancomycin resistance recently described in E. faecalis and associated with a probable D-Ala:D-Ser ligase similar to that present in VanC-type enterococci [18]. Both of these new types of vancomycin resistance are not frequently seen amongst current clinical isolates of VRE and are not the source of glycopeptide resistance in the more common VanA- and VanB-type VRE isolates.

Several bacteria have been known to be resistant to vancomycin since before the identification of vancomycin resistance in enterococci. These include Lactobacillus spp., Leuconostoc spp., Pediococcus spp., and Erysipelothrix rhusiopathiae. Vancomycin resistance in Lactobacillus casei, Pediococcus pentosaceus, Lactobacillus plantarum and Leuconostoc mesenteroides is associated with the presence of a pentapeptide peptidoglycan precursor terminating in D-Ala-D-Lac [12,19]. The peptidoglycan precursor structure in L. mesenteroides (MurNAc-L-Ala-D-Glu-L-Lys-[L-Ala]-D-Ala-D-Lac) [19], and that in L. plantarum (MurNAc-L-Ala-D-Glu-m-Dpm-D-Ala-D-Lac) [20], differ from that in enterococci, but still terminate in D-Ala-D-Lac. It is not clear whether one or two ligases are present in such organisms. A L. plantarum strain deficient for D- and L-lactate dehydrogenase activities, and which therefore produced only trace amounts of D- and L-lactate, has been constructed [20]. These changes drastically affected the peptidoglycan synthesis pathway [20], and a new precursor, terminating in D-alanine was identifiable, in addition to that terminating in D-lactate [20]. In addition, the organism had highly enhanced susceptibility to vancomycin [20]. Whether the origin of the D-alanine-ending precursor in the mutant was a result of a second ligase or mixed substrate specificity of a single ligase which preferentially uses D-lactate is not completely clear. However, since a comparison of the ligase genes from constitutively vancomycin-resistant lactobacilli (including L. plantarum) and L. mesenteroides, has shown that their deduced amino acid sequences are more closely related to each other than to that of a vancomycin-susceptible strain of Lactobacillus leichmannii, the latter is probably the case [21].

Vancomycin resistance in Lactobacillus confusus, Leuconostoc spp., and Pediococcus spp. is not transferable to enterococci using filter mating [22]. Although Leuconostoc spp. generally harbor a number of plasmids, plasmid-free strains of Leuconostoc spp. obtained by curing with novobiocin and ethidium bromide have been shown by some investigators to retain resistance to glycopeptides [19]. There has been one report of a clinical isolate of Lactobacillus viridescens in which glycopeptide resistance was associated with the presence of plasmid DNA and variations in cell wall protein patterns were noted following exposure to vancomycin [23], and another report that ethidium bromide curing of multiple plasmids was associated with concomitant loss of vancomycin resistance in Lactobacillus acidophilus[24]. These observations would suggest that two ligases are present in these organisms and that vancomycin resistance from lactobacilli could be transferable. This issue is controversial.

Although resistance to glycopeptides in Leuconostoc spp., Lactobacillus spp., and Pediococcus spp. is associated with the presence of peptidoglycan precursors terminating in D-Ala-D-Lac, as is the situation for enterococci with VanA- and VanB-mediated vancomycin resistance [12,19], the D-Ala:D-Lac ligase and D-lactate dehydrogenase enzymes involved in the synthesis of the altered precursors in Leuconostoc spp. and Lactobacillus spp. are not closely related to VanA or VanB and VanH or VanHB, respectively [21,2628]. In one study, no polymerase chain reaction (PCR) product was amplified in L. rhamnosus GG with any of three sets of vanA primers used, and enterococcal vanA, vanB, vanH, vanX, vanZ, vanY, vanS and vanR genetic material did not hybridize with DNA of L. rhamnosus GG [29]. More detailed studies have revealed that the levels of identity between the D-Ala:D-Lac ligase in Leuconostoc spp. and Lactobacillus spp. and VanA/B are 26–35%[21]. The D-Ala:D-Lac ligases of the vancomycin-resistant lactic acid bacteria differ in their primary sequence from VanA, VanB and VanD D-Ala:D-Lac ligases, VanC D-Ala:D-Ser ligases, and D-Ala:D-Ala ligases, notably in the ω-loop region, which is critical to catalysis [15]. This suggests that the evolution toward D-Ala-D-Lac formation occurred independently in ligases from lactic acid bacteria and VanA, VanB and VanD enterococci [15]. In lactic acid bacteria, lactate is abundant and the use of D-lactate instead of D-alanine could simply reflect metabolic parsimony [15]. Selective pressure due to the presence of glycopeptide-producing organisms in the environment could also have played a role in the development of D-Ala:D-Lac ligases in lactic acid bacteria [15].

Interestingly, it has recently been demonstrated that insertional inactivation of the rrp-31 gene, which encodes a response regulator of Lactobacillus sakei, increases the vancomycin MIC of this isolate, which is otherwise vancomycin susceptible [30]. Whether this relates to loss of autolysis, related to an altered two-component sensor-regulator system, as has recently been described in S. pneumoniae[31], or altered peptidoglycan precursor synthesis, or another mechanism, remains to be determined [30].

E. rhusiopathiae is resistant to vancomycin with a MIC90 of ≥64 μg ml−1 (range, 4–≥64 mg ml−1) [32], however, to the best of my knowledge, the mechanism of vancomycin resistance in E. rhusiopathiae remains to be described. PCR assays designed to detect the vanA and vanB genes in enterococci have not detected these genes in E. rhusiopathiae (Patel, unpublished results).

For the aforementioned reasons, Lactobacillus spp., Leuconostoc spp., Pediococcus spp., E. rhusiopathiae, and the intrinsically vancomycin-resistant enterococci are not the source of the glycopeptide resistance genes in VRE. There are, however, environmental organisms which have been found to have gene clusters resembling the vanA and vanB gene clusters; these include Paenibacillus popilliae, and, to a lesser extent, Amycolatopsis orientalis and Streptomyces toyocaensis.

Biopesticidal powders containing spores of P. popilliae have been introduced into turf in the Eastern US for the suppression of Japanese beetle populations since the late 1930s. The type strain of P. popilliae (American Type Culture Collection 14706) has a vancomycin MIC of 800 μg ml−1[33]. The putative ligase gene in P. popilliae has 77% nucleotide identity to the sequence of the vanA gene, and was originally designated vanE[33], however since another vanE gene was described [18], the putative ligase gene in P. popilliae was renamed vanF[34]. Putative genes, designated ‘vanYF’, ‘vanZF’, ‘vanHF’, and ‘vanXF’, which resemble the enterococcal vancomycin resistance genes, vanY, vanZ, vanH, and vanX, respectively, have been identified in P. popilliae[34]. The predicted amino acid sequences of the putative proteins are similar to those found in VanA VRE; 61% identity for VanYF, 21% identity for VanZF, 74% identity for VanHF, 77% identity for VanF, and 79% identity for VanXF (Fig. 1). The putative ligase vanF gene has been amplified from a P. popilliae isolate held in dried Japanese beetle hemolymph since 1945 and from a collection of 32 P. popilliae isolates [33]. This provides compelling evidence that the ligase gene present in P. popilliae was not transferred to this organism from an enterococcus.

Figure 1

Alignment of glycopeptide resistance gene clusters of VanA, VanB, and VanD enterococci, P. popilliae, and the glycopeptide-producing organisms A. orientalis and S. toyocaensis[4,17,27,4951]. The percents amino acid identity to vanY, vanZ, vanH, vanA, and vanX, are shown below the respective genes (below the arrows). The percent G and C content of each gene is shown above the genes (above the arrows). Reproduced with permission from Patel et al. [34].

P. popilliae has recently been reported to be a human pathogen [35]; in addition, an organism resembling Paenibacillus alvei, which is intermediately susceptible to vancomycin (vancomycin MIC, 8 μg ml−1), has been reported as a cause of pneumonia and empyema in a patient [36]. The mechanism of resistance to vancomycin in this organism is unknown, however no vanA- or vanB-type gene was present in this organism using PCR (Patel, unpublished results).

Recently, Marshall et al. cloned three genes encoding homologues of VanH, VanA, and VanX from the vancomycin producer A. orientalis C329.2 and the A47934 producer S. toyocaensis NRRL 15009 [27]. The predicted amino acid sequences of these proteins are similar to those found in VanA VRE: 54 to 61% identity for VanH, 59 to 63% identity for VanA, and 61 to 64% identity for VanX (Fig. 1) [27]. Phylogenetic analysis demonstrates that the gene products from S. toyocaensis, A. orientalis, P. popilliae, and VRE form distinct subfamilies of D-Ala:D-Lac ligases and D-lactate dehydrogenases [27,33]. The orientations of vanH, vanA, and vanX, in P. popilliae, S. toyocaensis, and A. orientalis, are identical to the orientations found in VRE, and the signature overlap of the 5′ end of the vanA and homologous genes with the 3′ end of the vanH genes is conserved [27,33]. Southern analysis of total DNA from other glycopeptide-producing organisms, including, A. orientalis 18098 (chloro-eremomycin producer), A. orientalis subsp. lurida (ristocetin producer), and Amycolatopsis coloradensis subsp. labeda (teicoplanin and avoparcin producer), with a probe derived from the vanH, vanA, and vanX cluster from A. orientalis C329.2 has demonstrated cross-hybridizing DNA in all strains [27]. In addition, the vanH, vanA, vanX cluster can be amplified from these other glycopeptide-producing organisms by PCR with degenerate primers complementary to conserved regions of VanH and VanX [27]. In S. toyocaensis there are two D-:D-ligases: a D-Ala:D-Ala ligase and a D-Ala:D-Lac ligase [37].

Marshall et al. hypothesize that the origin of clinically relevant vancomycin resistance lies within the glycopeptide-producing organisms [27]. However, the amino acid identities identified by these authors between the glycopeptide-producing organisms S. toyocaensis and A. orientalis and VanA and VanB VRE are substantially less than the identities reported between these genes in the vanA gene cluster and in P. popilliae (Fig. 1). Furthermore, the G+C content of the P. popilliae vanHF, vanF and vanXF genes is virtually identical to that of the homologous genes in VRE, and significantly different from the homologous genes in S. toyocaensis and A. orientalis (Fig. 1). The vancomycin resistance gene cluster in P. popilliae is more similar to that in VRE than are the gene clusters in S. toyocaensis and A. orientalis. Vancomycin resistance is not transferable from A. orientalis (or the teicoplanin producer A. teichomyceticus) to enterococci [22]. It is possible that the enterococcal vancomycin resistance-associated genes originated in the glycopeptide-producing organisms (where they would presumably be required to prevent such organisms from committing suicide), were then transferred into organisms of similar G+C content to enterococci (like P. popilliae), and thereafter into enterococci. Our group has not, however, been able to transfer vancomycin resistance from P. popilliae to enterococci (Patel, unpublished results). The incomplete identity of the P. popilliae vancomycin resistance genes and the enterococcal vancomycin resistance genes suggests that there exist other organisms which may be the direct sources of vancomycin resistance genes in enterococci.

Given that the G+C contents of the VRE vanH, vanA and vanX genes are 5–10% higher than those of the adjacent vanR, vanS, vanY, and vanZ genes (Fig. 1), it is plausible that the vanH, vanA, and vanX genes have been mobilized as a unit from another source and positioned into contact with appropriate control elements. Transfer of vancomycin resistance from an organism like P. popilliae to an enterococcus may first require transfer of a conjugative transposon from an enterococcus (or another Gram-positive bacterium) to this organism, followed by insertion of vancomycin resistance determinants into the conjugative transposon, and thereafter transfer of the transposon and vancomycin resistance back into an enterococcus [38]. This may account for our difficulty in trying to transfer vancomycin resistance from P. popilliae into an enterococcus. Interestingly, although all North American isolates of P. popilliae we have examined are vancomycin resistant, most Latin American P. popilliae isolates are vancomycin susceptible suggesting that vancomycin resistance has been acquired by P. popilliae (Yousten, unpublished results). Given that we have been able to detect the vanF gene in a P. popilliae isolate from 1945, transfer of vancomycin resistance into P. popilliae would have occurred over five decades ago [33].

The vanH/vanA/vanX gene system may have evolved by mutations of genes encoding proteins involved in normal cell wall biosynthesis. For example, in the Gram-negative bacterium Escherichia coli, which is never challenged by glycopeptide antibiotics (because they cannot penetrate the outer membrane permeability barrier), ddpX (a vanX homologue) is part of a dipeptide transport and degradation system which presumably functions to recapture D-Ala-D-Ala [39]. The catalytic efficiency of D-Ala-D-Ala hydrolysis for DdpX is 25-fold lower than for enterococcal VanX, suggesting that DdpX functions in cell wall turnover, rather than protection [40]. The vanH/vanA/vanX gene system may have been built using a more efficient VanX homologue and mutating the gene that synthesizes D-Ala-D-Ala to synthesize D-Ala-D-Lac. For example, it has been shown that single-residue changes in the active-site regions of D-,D-ligases can cause substantial changes in recognition and activation of hydroxy or amino acids [41].

Other bacteria with known elevated vancomycin MICs include Clostridium innocuum (vancomycin MIC90 16 μg ml−1, range 8–16 μg ml−1), Nocardia spp. (vancomycin MIC90 128 μg ml−1, range 16–128 μg ml−1), and single isolates of Gemella haemolysans (vancomycin MIC>32 μg ml−1) and Corynebacterium spp. (vancomycin MIC 32 μg ml−1) [25,4244]. The mechanism of resistance to vancomycin in these isolates in unknown, although no enterococcal vanA or vanB genes were identified in C. innocuum using a polymerase chain reaction assay (Patel, unpublished results).

Vancomycin intermediately susceptible S. aureus isolates have been recently described in Japan and in the United States [45]. The mechanism of vancomycin resistance in vancomycin intermediately susceptible S. aureus is postulated to be a result of thickened cell wall and resultant increased vancomycin-binding capacity effectively preventing vancomycin molecules from reaching the cytoplasmic membrane by trapping them within the existing cell wall [46]. The vanA and vanB genes have not been detected in these isolates.

Recently, emergence of vancomycin tolerance in S. pneumoniae as a result of loss of function of the VncS histidine kinase of a two-component sensor-regulator system thought to regulate a basic pathway triggering autolysis, has been reported [31]. Although there is 38% identity between the deduced amino acid sequences of VncS and the response regulator, VcnR, and the VanSB-VanRB two-component regulatory system in enterococci, scanning of the pneumococcal genome reveals no homologues of vanH, vanA, vanB or vanX[31]. Although similar two-component systems affect susceptibility to vancomycin in pneumococci and enterococci, their underlying roles in resistance and tolerance appear to be distinct, presumably reflecting different gene products controlled by the response regulator [31].

In conclusion, although the emergence of VRE can be attributed, in part, to the increasing use of oral and parenteral vancomycin in humans since the late 1970s for the treatment of C. difficile and MRSA (and MR-CNS) infections, respectively, and glycopeptide use in animal husbandry [47,48], the exact source of the enterococcal vancomycin resistance genes remains a mystery. Glycopeptide-producing organisms are potential candidates in this regard, especially as the original source of vancomycin-resistant genes. Given the findings in P. popilliae, antimicrobial resistance in glycopeptide-producing organisms may have been transferred to other organisms with a G+C content more similar to enterococci, with vancomycin resistance then being transferred from these organisms into enterococci.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
  30. [30].
  31. [31].
  32. [32].
  33. [33].
  34. [34].
  35. [35].
  36. [36].
  37. [37].
  38. [38].
  39. [39].
  40. [40].
  41. [41].
  42. [42].
  43. [43].
  44. [44].
  45. [45].
  46. [46].
  47. [47].
  48. [48].
  49. [49].
  50. [50].
  51. [51].
View Abstract