OUP user menu

Cryptococcus neoformans: A sugar-coated killer with designer genes

John R. Perfect
DOI: http://dx.doi.org/10.1016/j.femsim.2005.06.005 395-404 First published online: 1 September 2005

Abstract

Cryptococcus neoformans has become a common central nervous system pathogen as the immunocompromised populations enlarge world-wide. This encapsulated yeast has significant advantages for the study of fungal pathogenesis and these include: (1) a clinically important human pathogen; (2) a tractable genetic system; (3) advanced molecular biology foundation; (4) understanding of several virulence phenotypes; (5) well-studied pathophysiology; and (6) robust animal models. With the use of a sequenced genome and site-directed mutagenesis to produce specific null mutants, the virulence composite of C. neoformans has begun to be identified one gene at a time. Studies into capsule production, melanin synthesis, high temperature growth, metabolic pathways and a variety of signaling pathways have led to understandings of what makes this yeast a pathogen at the molecular level. Multiple principles of molecular pathogenesis have been demonstrated in virulence studies with C. neoformans. These include evolutionary differences between the varieties of C. neoformans in their genes for virulence, quantitative impact of genes on the virulence composite, species and site-specific importance of a virulence gene, gene expression correlation with its functional importance or phenotype and the impact of a pathogenesis gene on the host immune response. C. neoformans has now become a primary model to study molecular fungal pathogenesis with the goal of identifying drug targets or vaccine strategies.

Keywords
  • Cryptococcus
  • Molecular biology
  • Pathogenesis
  • Virulence
  • Genes

1 Introduction

Cryptococcus neoformans is an encapsulated, heterobasidiomycetous fungus that has dramatically risen from a rare cause of human infections with less than 300 reported cases before 1955 to a common world-wide opportunistic pathogen in the rising immunocompromised populations over the last two decades. Cryptococcus is one of the leading pathogens to be grown from the cerebrospinal fluid in the world today. The yeast is a particularly fascinating fungal pathogen because it crosses the entire spectrum of host immunity. For instance, it produces infections in apparently immunocompetent patients without any known underlying disease but on the other hand, commonly the yeast invades a severely immunosuppressed host as a result of HIV infection, organ transplantation, malignancy or receiving high doses of corticosteroids [1]. Therefore, this encapsulated yeast has developed a variety of tools to successfully attack mammalian hosts. Furthermore, this yeast has the ability to simply colonize the host's respiratory tract without production of disease, but it also has a unique propensity to invade the central nervous system of the mammalian host. In fact, it can disseminate from a localized infection to produce disease in any organ of the human body.

A variety of recent research and clinical developments have allowed this yeast to be elevated to the center stage for study of fungal pathogenesis. However, it should be acknowledged that there are both disadvantages and advantages for the study of C. neoformans as a model fungal pathogen. The disadvantages center around: (1) it is a basidiomycete and not the more common ascomycete infection in human disease; (2) the yeast contains a unique capsule; (3) the field lacks a critical number of investigators for its study; (4) primarily, molecular studies focus only on virulence factors; and (5) cryptococcosis is not the most common deep-seated human pathogen. It could be stated that C. neoformans is not a model fungal pathogen for any fungus except Cryptococcus. However, this is a very narrow viewpoint, which this investigator rejects and conversely, believes that this basidiomycete is actually an ideal yeast in which to understand the fundamentals of molecular fungal pathogenesis. With this yeast it is now possible to dissect what it means to be a human fungal pathogen one gene at a time and integrate these findings into an understanding of the specific genotypes that produce the phenotypes to control the production of disease. The insights gained from molecular studies of this yeast pathogen will allow us to identify drug targets, study antifungal compounds, define drug resistance mechanisms and/or prepare mutants or fungal products for protective vaccines.

In the following discussion, six major factors will be discussed that represent specific areas of advantages that make C. neoformans an ideal fungus to study.

2 Important clinical pathogen

There is presently no established clinical test to detect latent cryptococcosis, and this reduces our ability to assess the magnitude of infection with this yeast. However, it has been found that the majority of adults possess antibody to C. neoformans, and in New York City most children acquire antibodies to cryptococcal antigens before the age of 10 [2]. These observations suggest asymptomatic cryptococcal infections are frequent, and occur at a young age, but the immune response is generally very effective in preventing disease. It is hypothesized that the yeasts from the initial infection can remain latent in the host for long periods of time, and reactivation of infection can occur as the host's immune system becomes compromised. Best estimates for rates of invasive cryptococcosis in the United States in the pre-AIDS era predicted an overall incidence of 0.8 cases per million persons per year [3]. In 1992, during the peak of the AIDS epidemic within the United States, the rate soared to approximately five cases of cryptococcosis per 100,000 persons in several urban areas [4]. During the widespread use of fluconazole and the advent of HAART in the late 1990s, rates of symptomatic infection reduced to one case per 100,000 persons per year [5]. In the United States, C. neoformans now frequently represents an infection that identifies a disadvantaged patient who is either undiagnosed or untreated for HIV infection [6] or a patient receiving corticosteroids. Less-developed countries ravaged by HIV infection, such as these in sub-Saharan Africa, report prevalence rates of 15–45% of those with advanced HIV infection succumbing to cryptococcosis [7,8]. In these areas, cryptococcal meningitis is the most common cause of central nervous system infections, and without treatment, mortality is 100% within the first two weeks of hospitalization [9]. Despite several antifungal drugs (amphotericin B, flucytosine, fluconazole) available for the treatment of cryptococcal meningitis, it is clear that clinical drug resistance frequently occurs. In fact, the acute mortality of cryptococcal meningitis despite present therapies still runs between 10% and 25% in the most medically advanced countries [1]. Cryptococcosis is a common life-threatening infection, which produces significant morbidity and its management needs to be improved.

3 Genetic system

The life-cycle of C. neoformans involves two distinct forms: asexual and sexual [10]. Human infection is primarily characterized by the asexual budding yeast forms in tissue and fluids. However, it is hypothesized that a sexual form, the basidiospore, with its small size, might be the infectious propagule. The mating type loci have been extensively studied and shown to be involved in the virulence of C. neoformans var. neoformans (serotype D) [11]. However, recent work with C. neoformans var. grubii (serotype A) has suggested that the mating type loci of this variety have little or no contribution to virulence [12]. It will be helpful to have further strains studied to determine if this is a varietal or strain difference in virulence between strains with different mating type loci. Recent studies have also made substantial progress in understanding the molecular signaling networks that control the sexual cycle and haploid fruiting, and these studies may have specific implications in understanding both morphogenesis and virulence [10]. A significant technical advance in molecular pathogenesis studies has resulted from the availability of congenic pairs of both C. neoformans var. neoformans and C. neoformans var. grubii to allow for genetic crosses [11,12]. Matings can now be used to confirm specific mutants and to reduce background mutations when null mutants are created. Furthermore, studies can continue to elucidate the impact of the mating loci on pathogenesis and morphogenesis, and attempt to understand why there is a major mating-type bias (alpha) in specimens from both nature and human infections [13]. The mating loci have been sequenced, annotated and several genes studied for their impact on morphogenesis and virulence [1416], but there is still much to be learned about the linkage of sex and pathogenesis.

4 Molecular biology foundation

There have been many studies performed over the last decade that have advanced the molecular biological aspects of C. neoformans. A series of molecular-based typing studies have identified individual strains [1719]. This precise genotyping has been able to distinguish relapsed infection from re-infection with a novel strain [20] and led to grouping strains into several genotypes [21]. With the use of molecular evolutionary studies, it has been predicted that C. neoformans var. neoformans (serotype D) and C. neoformans var. grubii (serotype A) have been separated from each other genetically for over 18 million years [22] and the genetic separation is even longer for C. neoformans var. gattii from the other two varieties. This genetic separation of the varieties can have relevance in the study of molecular pathogenesis. For instance, it might help explain the differences observed between varieties of Cryptococcus regarding the impact of certain virulence genes. For example, the STE12 gene is important for the virulence composite of a serotype D variety neoformans strain but the same gene is not for a serotype A (variety grubii) strain [23,24]. Conversely, the STE20 gene is important for a serotype A strain in its virulence composite but not a serotype D strain [25]. Furthermore, the signaling pathway through cyclic AMP has a varietal difference for virulence in that PKA1 is important for virulence through its control of capsule and melanin production in a serotype A strain but this is not the case in a serotype D strain [26]. As previously noted, the mating loci appear to play a different role in the virulence composite between the two varieties [11,12].

Importantly, several transformation systems with DNA delivery by electroporation, biolistics, or Agrobacterium and using both auxotrophic complementation and dominant selection markers have been developed [2733]. Homologous DNA recombination can average between 2% and 10% of transformants with biolistic DNA delivery for creation of site-directed mutants and the use of epistasis experiments are now routinely performed. There have been multiple strategies used to identify essential genes in C. neoformans[34,35] and both antisense [36] and RNA interference [37] techniques have been successfully employed in this yeast.

The fulfillment of molecular postulates for virulence [38] is now expected for all of the molecular pathogenesis studies performed in C. neoformans. Currently, approximately 40–50 genes have been validated for their virulence potential by this standard molecular strategy (Table 1). The use of random signature-tagged mutant libraries have begun to be analyzed [33,39]. The completion of the sequencing, alignment and annotation of three cryptococcal genomes by Stanford University, TIGR, British Columbia Genome Center and the Duke/Whitehead Institute consortium is imminent (http://www.neo.genetics.duke.edu; http://www.tigr.org.tdby/e2kl/cnal; http://www.genome.stanford.edu; http://www.genome.ou.edu.cneo.html; http://www.bcgsc.bc.ca). The power of this informational resource has substantially elevated the genomic potential of this yeast for study. For example, one recent off-shoot of the genome sequencing projects has been the development of microarray transcriptional profiling in C. neoformans[40]. Despite effective use of cDNA library subtraction protocols [41], differential display RT-PCR [42] and SAGE analyses [43] for global analysis of transcription in C. neoformans, the microarray allows for rapid generation of more immediate identifiable data to make new hypotheses or confirm/refute old ones. As powerful as the global profiles for studying gene regulations are, studies in C. neoformans have demonstrated that gene regulation does not always correlate with gene function and that specific null mutants must be evaluated for virulence phenotype when a gene's importance is implicated by its expression profile. For instance, the C. neoformans isocitrate lyase (ICll) gene is highly up-regulated in its expression at the site of a central nervous system infection but the null mutant for ICll does not have an attenuated phenotype in the same animal model [44]. Furthermore, it is apparent that “virulence” genes in C. neoformans under standard in vivo conditions will have a variable or quantitative impact on the virulence composite. For instance, some null mutants will be completely eliminated from the host with 100% host survival while others will have prolonged survival but animals will still die of infection (Table 2). It has also been shown that the classical phenotypes of virulence (capsule, melanin, temperature and intracellular growth) in C. neoformans are controlled by many genes and several of these genes have now been identified and characterized (Table 3).

View this table:
Table 1

Site-specific gene disruptants in Cryptococcus neoformans and their impact on virulence in animal models

Attenuated virulenceHypervirulentNo impact
ADE2[113]AOX 1[41]PKR1[42]FKB1
URA 5[32]CPA1[76]CAS1[52]CBP1[114]
MET 3[62]PLB1[79]APP1[111]STE11[14]
MET 6[115]URE1[78]STE7[14]
ILV2[73]MAN1[116]CPA2[76]
TPS1[75]CAS2[117]CPK1[14]
CNA1[66]STE12D[24]RAS2[76]
CNB1[118]STE20A[25]ICL1[44]
RAS1[70]MFα 1,2,3[119]CCP1[120]
PKA1[121]CAP 10,[47,48]SXI1[122]
59,60,[49,50]
64[51]
STE20D[25]
LAC1[58]SKN 7[99]GPB1[16]
SOD1[81,123]VPH1[84]STE12A[23]
SOD2[82]MPK 1[72]
CAC1[124]TSA1[83]
CCN1[69]SPE3/LYS9[74]
CLC1[60]CTS1[125]
CPRa[126]FHB1[127]
PAK1[25]APP1[111]
GPA1[46]
  • Genes are given names and numbers consistent with Saccharomyces homologs in many cases or designated letters taken from the manuscript. (References are in brackets.)

  • Impact on virulence dependent on host.

View this table:
Table 2

Examples of gene disruptants relative impact on virulence in mice

SevereModerateMild
MAN1 ILV2PAK1URE1
CNB1 MET3PKA1PLB1
CNA1CAC1STE 12/20
RAS1GPA1SOD1
TPS1SPE3/LYS9AOX1
SOD2MET6FHB1
  • Complete elimination of yeast from host and full survival of animals.

  • Reduction in yeast counts compared to wild type and majority of animals will survive.

  • Statistically reduced yeast counts in tissue and prolonged survival compared to wild type but all animals eventually succumb to infection.

View this table:
Table 3

Genes linked to the major virulence phenotypes

Capsule formationMelanin productionTemperature (high temperature growth 37–39 °C)Intra-cellular growth defect
GPA1CAS1GPA1MET3CNA1RAS1PLB1
CAC1CAS 2CAC1STE12CNB1SOD2SOD1
PKA1CAP 60PKA1CLC1CPA1TSA1SKN7
PKR1CAP 10PKR1CCN1ILV2IPC1
MAN1CAP 64VPH1TPS1SPE3/LYS9AOX1
CAP 59VPH1LAC1TPS2MPK1
STE 20STE12IPC1MGA2STE 20

Finally, improvements in vector designs such as stable plasmids, in vivo regulated promoter systems and more complete microarrays are still needed tools and they are being developed. However, the majority of molecular tools are currently available to perform cutting-edge, molecular biology experiments with this pathogenic yeast.

5 Virulence phenotype

C. neoformans has several well-characterized virulence phenotypes. The three classical phenotypes under genetic control have been: (1) capsule production; (2) melanin formation; and (3) the ability of a yeast strain to grow well at 37 °C [45].

The most distinctive feature of C. neoformans is the polysaccharide capsule containing an unbranched chain of α1,3-linked mannose units substituted with xylosyl and β-glucuronyl groups. The possible mechanisms for capsular protective effects for the yeast against the host include antiphagocytosis, complement depletion, antibody unresponsiveness, inhibition of leukocyte migration, dysregulation of cytokine secretion, brain enhancement of HIV infection, l-selectin and tumor necrosis factor receptor loss, and negative charge [45]. Several C. neoformans genes in capsular synthesis and formation have been identified, and site-directed gene mutants created which produce hypocapsular or acapsular strains. All capsule-deficient mutants have been found to be less virulent than the parental strains [4651]. On the other hand, genetic manipulation of the capsular structure by either a change in its structure or increased formation of capsule can make strains hypervirulent [52,53]. Furthermore, it has been shown that the capsular structure is dynamically enlarged during infection and even at different sites of infection [54,55]. The molecular biology of capsule production will need to interface with further understanding of its biochemistry to make further progress in this area, but recently, it has been shown that the capsular polysaccharide is attached to the cell wall of the yeast by glucan bridges [56].

The production of melanin is observed in many pathogenic fungi [57] and C. neoformans possesses particularly prominent laccase activity that converts diphenolic compounds to melanin. There are several laccase-like genes in C. neoformans and disruption of one of these genes (LAC1) has led to “an albino” mutant on specific agar plates that is attenuated for virulence in animal models [58]. Further studies have now assigned several genes to melanin production and regulation in C. neoformans[53,5863]. It has been suggested that a primary mechanism for melanin's importance is its capacity to act as antioxidant but there are other possible mechanisms which the yeast might use melanin for protection from the host including cell wall integrity and charge, interference with antifungal susceptibility, abrogating antibody-mediated phagocytosis, and protection from extreme temperatures [45].

The ability to grow at 37 °C is a simple and intuitively obvious phenotype for any invasive mammalian pathogen. In the fungal kingdom there are over 20,000 different fungal species and yet less than two dozen consistently cause human disease. These pathogenic fungi have developed the molecular tools to survive at host temperatures while the nonpathogenic fungi rarely possess this innate ability. This phenotype of high temperature growth is vividly illustrated among the Cryptococcus species, since some of these species possess capsules and/or others produce melanin. However, besides C. neoformans, none of these species can grow well at 37 °C, and all of the non-neoformans species are generally considered non-pathogenic in humans. There also appears to be some evolutionary drift in high temperature growth in that serotype B and C (variety gatti) and serotype D (variety neoformans) isolates appear generally to be more sensitive to growth inhibition and killing at high environmental temperatures 37–39 °C compared to serotype A isolates (variety grubii) [64].

The ability to grow at 37 °C in C. neoformans was first shown to be under direct genetic control and linked to virulence by the elegant studies of Rhodes and colleagues [65]. From these initial experiments, substantial progress in the genetic understanding of high temperature growth in C. neoformans has recently been made. The first molecular studies on high temperature growth demonstrated that calcineurin A mutants grew normally at 30 °C but were not viable at 37 °C, and were found not to be pathogenic in several mammalian models of infection [66]. This phenotype for the signaling molecule, calcineurin, in C. neoformans would not have been predicted from studies with the model yeast, Saccharomyces cerevisiae, since the mutation of the homologous gene in this yeast does not result in temperature-sensitivity for growth. This is a prime illustration of the importance of studying a true pathogen rather than a surrogate model.

From these molecular studies, there has been further significant progress in identifying the many genes associated with high temperature growth in C. neoformans. First, screens to examine for transcriptional regulation of genes at high temperature have led to the identification of genes such as COX1 (cytochrome oxidase c subunit 1) and AOX1 (alternative oxidase), which are up-regulated in their expression to a shift from low to high temperatures [41,67]. With COX1, it was found that there are differences in temperature-regulated transcription between variety grubii and neoformans strains, but this difference in COX1 transcription did not impact on the differences between the varieties in terms of growth or virulence [63,67]. Furthermore, although an aox1 mutant did possess a measurable defect in virulence, it did not possess a temperature-sensitive growth defect [41]. Conversely, the calcineurin gene, while necessary for growth at 37 °C, is not regulated by shifts in temperatures. Therefore, cryptococcal genes regulated by temperature are not always involved in high temperature growth, and genes involved in high temperature growth are not always regulated by temperature.

There are many other C. neoformans genes that have been identified as important for high-temperature growth. VPH1 (vacuolar APT-ase) [68] and CCN1 (homolog of CLF1 used in RNA/DNA replication and splicing) [69] have been found to be required for high temperature growth, and genes in several signaling pathways such as RAS1, CNA1, CNB1, MPK1 and CTS1 have been implicated in high temperature growth [66,7072]. Also, it has been shown that basic amino acid metabolism genes such as ILV2 gene and a SPE3/LYS9 chimeric gene are required for high temperature growth [73,74]. In the study of the stress protectant sugar, trehalose, it was shown that mutants in the TPS1 (trehalose phosphate synthase) and TPS2 (trehalose-6 phosphate phosphatase) genes exhibited a phenotype in which they can grow normally at 30 °C but are inhibited or killed at 37 °C [75]. Some genes, such as STE20α and CPA1 (cyclophilin), are required for very high temperature growth at 39–40 °C, but are dispensable for growth at 37 °C [25,76]. Finally, with microarray transcriptional profiling of C. neoformans regulated genes at 37 °C compared to 25 °C, a transcription factor MGA2 was found to be upregulated and the null mutant had reduced growth at 37 °C. Similarly, MGA2 is a homolog of a S. cerevisiae gene, which is involved in regulating transcription in response to cold and hypoxia and specifically activates expression of a fatty acid desaturase gene. The null mutant of C. neoformans MGA2 also displayed another phenotype which was hypersensitivity to cell membrane stresses such as fluconazole exposure. This finding vividly illustrates that these networks for high temperature growth are linked to other stress response pathways.

There are several points of significance to our present understanding of temperature-related genes in C. neoformans. First all null mutants which demonstrate a temperature-sensitive phenotype at 37 °C are completely avirulent in mammalian models, and would likely make excellent antifungal targets. Second, genes that are regulated by temperature are not necessarily required for high-temperature growth. Third, temperature-related genes cannot be predicted from other model yeasts, and some even vary in importance between varieties of C. neoformans. Fourth, temperature-related genes can have a variety of functions. These genes may participate in certain stress resistant pathways, signaling networks, and in basic metabolism. Fifth, studies in S. cerevisiae have identified over 70 genes for which the null mutants cannot grow at 37 °C; therefore, the approximately dozen identified temperature-sensitive genes in C. neoformans are probably just a fraction of the genes which represent potential genetic weak spots for inhibition of C. neoformans growth in the host.

There are several other known phenotypes that have been associated with attenuated virulence in C. neoformans by gene analysis, and these include urease production [78], phospholipases secretion [79], mannitol production [80], oxidative stress reduction [41,8183], and vacuolar acidification [84].

6 Immunology

The host responses to C. neoformans are relatively well-understood. There is general agreement supported by many studies that a strong cellular immune response producing granulomatous inflammation is essential for containment of this infection [85,86]. This inflammation is primarily a Th-1 polarized response with requirement of cytokines such as tumor necrosis factor, interferon gamma and IL-2 and chemokines MCP-1 and MIP-1α for recruitment of inflammatory cells [8790]. It is also clear that effector cells against C. neoformans include CD4+ lymphocytes, CD8+ lymphocytes, and activated macrophages [91,92]. For example, the risk for disseminated cryptococcosis rises precipitously for patients with HIV infection when CD4 counts fall below 100 cells µl−1[93]. Other studies have identified several innate factors (serum, saliva, and complement) that discourage infection. Creative experiments over the last decade have shown that humoral immunity can play an important role in the elimination of cryptococci from the body [94,95]. With the use of molecular biology, several genes and encoded proteins have now been identified which may help elicit a protective immune response [96,9698]. In fact, with the use of molecular tools, it has now been possible to create C. neoformans strains which secrete host cytokines [99] which might aid in their own destruction. Finally, it is clear that different C. neoformans strains, those with switching colony morphology, or even yeast strains with a specific null mutation can produce a variable immune response during infection [100103].

7 Animal models

One of the primary advantages of using C. neoformans in fungal virulence studies is the robustness of the animal models. Routine mammalian models include rabbits, rats, and mice. There is tremendous flexibility in the models in terms of the types of strains that can be used, and the use of various drug treatments for infection. Both survival and burden of infection have been used as virulence composite endpoints, and over 40–50 site-directed null mutants have been evaluated using these models. It is clear that these models are reproducible, and can distinguish various degrees of virulence potential associated with specific C. neoformans genes. Review of many null mutant outcomes shows that the animal models reveal genes, which produce hypovirulence, hypervirulence or no impact on the virulence composite (Table 1). It has also been noted that most specific site-directed gene (null) mutants have concordant impact on virulence in several animal models. However, there are genes which are important for the virulence composite in one model but not in another [78]. We await further studies on infection-site specific genes but early studies do suggest that specific genes and their encoding proteins may be important to infection at a specific body site such as the central nervous system [104].

For expansion of host-fungal studies, several models of cryptococcosis in less complex organisms have been recently developed. Amoebae [105], Dictyostelium[106], nematodes [107] and Drosophila[107] have been studied with C. neoformans as potential surrogates for mammalian infection models. In fact, the nematode model (C. elegans) has shown similar virulence assessment for some mutants compared to the mammalian models [108].

There are several important issues that need to be discussed as we examine the genetic basis for virulence in C. neoformans. We should use some caution because the vast majority of the molecular work is performed in only several prominent cryptococcal strains. Paradigm-setting studies will need to be re-confirmed in other wild-type strains. Furthermore, in these strains, the virulence composite is dynamic with measurable microevolution occurring in the laboratory [55,102,109,110]. For example, a strain can change its relative virulence through multiple, nonselective in vitro passages by producing an attenuation of virulence and conversely, passing the strain through an animal will amplify its relative virulence composite. This rapid loss or acquisition of virulence needs to be further understood and carefully controlled for in all animal experiments. On the other side of the infection encounter, the change of the host genetic background can have profound impact on outcome of infection. For instance, the app1 mutant of C. neoformans is avirulent in an immucompetent host, but when inoculated into a severely immunosuppressed host it becomes highly virulent [111]. There will continue to be many examples of host genetic loci controlling the virulence manifestation of a yeast strain by modulating the net immunosuppression of the host.

C. neoformans is a unique, sugar-coated, killer yeast. It particularly attacks the immunologically weakened host. Although clinical management of infections is possible, there remains substantial failure rates. The domestication of this fungus in the research laboratory has now made it a facile fungal pathogen for the study of molecular pathogenesis. There is now a great potential to further understand this yeast as a beast for producing disease through the genes which make it so [112]. These insights can lead to new designer drugs and vaccine strategies for management of cryptococcosis.

Acknowledgement

Dr. Perfect is supported by Public Health Service Grant AI28388.

References

  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].
  52. [52].
  53. [53].
  54. [54].
  55. [55].
  56. [56].
  57. [57].
  58. [58].
  59. [59].
  60. [60].
  61. [61].
  62. [62].
  63. [63].
  64. [64].
  65. [65].
  66. [66].
  67. [67].
  68. [68].
  69. [69].
  70. [70].
  71. [71].
  72. [72].
  73. [73].
  74. [74].
  75. [75].
  76. [76].
  77. [78].
  78. [79].
  79. [80].
  80. [81].
  81. [82].
  82. [83].
  83. [84].
  84. [85].
  85. [86].
  86. [87].
  87. [88].
  88. [89].
  89. [90].
  90. [91].
  91. [92].
  92. [93].
  93. [94].
  94. [95].
  95. [96].
  96. [97].
  97. [98].
  98. [99].
  99. [100].
  100. [101].
  101. [102].
  102. [103].
  103. [104].
  104. [105].
  105. [106].
  106. [107].
  107. [108].
  108. [109].
  109. [110].
  110. [111].
  111. [112].
  112. [113].
  113. [114].
  114. [115].
  115. [116].
  116. [117].
  117. [118].
  118. [119].
  119. [120].
  120. [121].
  121. [122].
  122. [123].
  123. [124].
  124. [125].
  125. [126].
  126. [127].
View Abstract