OUP user menu

Iron and heme utilization in Porphyromonas gingivalis

Teresa Olczak, Waltena Simpson, Xinyan Liu, Caroline Attardo Genco
DOI: http://dx.doi.org/10.1016/j.femsre.2004.09.001 119-144 First published online: 1 January 2005


Porphyromonas gingivalis is a Gram-negative anaerobic bacterium associated with the initiation and progression of adult periodontal disease. Iron is utilized by this pathogen in the form of heme and has been shown to play an essential role in its growth and virulence. Recently, considerable attention has been given to the characterization of various secreted and surface-associated proteins of P. gingivalis and their contribution to virulence. In particular, the properties of proteins involved in the uptake of iron and heme have been extensively studied. Unlike other Gram-negative bacteria, P. gingivalis does not produce siderophores. Instead it employs specific outer membrane receptors, proteases (particularly gingipains), and lipoproteins to acquire iron/heme. In this review, we will focus on the diverse mechanisms of iron and heme acquisition in P. gingivalis. Specific proteins involved in iron and heme capture will be described. In addition, we will discuss new genes for iron/heme utilization identified by nucleotide sequencing of the P. gingivalis W83 genome. Putative iron- and heme-responsive gene regulation in P. gingivalis will be discussed. We will also examine the significance of heme/hemoglobin acquisition for the virulence of this pathogen.

  • Porphyromonas gingivalis
  • Iron/heme/hemoglobin binding
  • Iron/heme transport
  • Outer membrane receptor
  • Gingipain

1 Introduction

Porphyromonas gingivalis is a black-pigmented, Gram-negative anaerobic bacterium, which has been implicated as a major etiological agent in the development and progression of chronic periodontitis [13]. Periodontal diseases are infectious conditions, which may progress over several years with episodes of exacerbation and remission. Formation of a periodontal pocket, destruction of supporting connective tissues, and loss of alveolar bone take place in the progression of this disease. Initiation and progression of periodontal diseases is the result of a complex interaction between the bacteria colonizing the gingival crevice and the immune and inflammatory responses of the host. P. gingivalis cells are found in healthy sites [4], but significantly elevated levels of the organism are observed in periodontal lesions [5, 6]. Increased systemic and local immune responses to P. gingivalis have been found in subjects with various forms of the disease, and serum IgG antibody responses to this organism have been reported to be elevated in individuals with a history of periodontal diseases as compared to controls [7, 8]. Although the bacterium may be eliminated by successful therapies [9], it causes recurrent infections [5].

In the periodontal pocket, P. gingivalis is found predominantly as a component of complex biofilms containing multiple bacterial species. It has been documented that this pathogen has an ability to aggregate with other oral microorganisms, such as Streptococcus gordonii or Prevotella intermedia, the first oral bacteria colonizing on tooth surfaces, thus providing attachment sites, supplying growth substrates and reducing oxygen tension [10, 11]. P. gingivalis produces a variety of enzymes, particularly the trypsin-like proteases, whose activity has been found to be critical for this pathogen. Of these enzymes, gingipains are considered to be essential for the growth and survival of the bacterium by providing a source of nutrients [12]. Moreover, P. gingivalis adheres to a variety of host cell surfaces [2] and the proteolytic activity of this organism plays a key role in the disruption of host-defense mechanisms, the penetration and destruction of the host connective tissues, as well as development and maintenance of inflammation in periodontal pockets [13, 14]. P. gingivalis has also been implicated as a risk factor for other conditions including cardiovascular disease [15, 16] and pre-term delivery of low birth weight infants [17, 18].

The ability of a pathogen to colonize and proliferate within an environmental niche in the host is essential for the establishment and progression of an infection. Growth of microorganisms depends in part upon the ability of a pathogen to scavenge essential nutrients. It is well established that bacterial growth and subsequent colonization is dependent on the ability to acquire iron. P. gingivalis utilizes iron and protoporphyrin IX (PPIX; Fig. 1(a)) for growth due to its iron requirement and the inability to synthesize the protoporphyrin ring [19, 20]. This anaerobe has attracted much attention during recent years and research over the past decade has significantly increased our knowledge of its mechanisms of iron and heme utilization and their contribution to P. gingivalis virulence. Considerable attention has been given to the characterization of various secreted and surface-associated proteins. Of these, gingipains have been linked directly to disease pathogenesis due to their ability to degrade host structural and defense proteins, as well as their high capability to bind heme and iron- or heme-containing proteins [21].

Figure 1

Structures of porphyrins. The basic structure of porphyrin consists of four pyrrole rings linked by four methene bridges (a). Four methyl, two vinyl, and two propionate side chains are attached. Fe2+ is added to protoporphyrin IX via ferrochelatase, to yield heme (a,b). The term heme refers to reduced, ferrous or Fe(II) iron protoporphyrin IX, whereas the term hemin (b) refers to the oxidized, ferric or Fe(III) form of the molecule. The μ-oxo dimer of heme (b) is a compound composed of two Fe(II) PPIX moieties bridged by an oxygen atom. Structures of selected non-natural metalloporphyrins (c) are also shown.

In this review we discuss the mechanisms of iron and heme uptake utilized by P. gingivalis. The P. gingivalis W83 genome has been sequenced (The Institute for Genome Research – TIGR; http://www.tigr.org) [20] and putative genes that may be involved in iron/heme acquisition have been identified (Table 1). We also discuss putative iron- and heme-responsive gene regulation mechanisms in this bacterium and review the significance of the heme/hemoglobin utilization systems for P. gingivalis virulence.

View this table:
Table 1

Summary of P. gingivalis proteins that are/may be involved directly or indirectly in iron/heme utilization

ProteinDescription and functionStrainAccession/protein ID/locusReference
Hemin-binding protein (OMP26)Hemin-regulated protein; binding of heminW50Not identified[154, 155]
Hemin-binding protein (OMP80)Hemin-regulated protein; binding of heminW50Not identified[154, 155]
Hemin-regulated receptor (HemR)Binding of heminATCC53977U54787/AAC44980[190]
HmuYBinding of heminA7436, W83AF200358/AAF07986, AAQ66587[20, 138, 162]
Hemin utilization receptor (HmuR)Binding of hemin and hemoglobin and transport of heme across the outer membraneA7436, W83U87395/AAB47566, AAQ66588
[20, 138, 162]
IhtA – TonB-dependent receptorUptake and transport of iron/hemeW83AAQ65844[20]
IhtB-hemin-binding protein FetB (cobalt chelatase or ferrochelatase; formerly immunoreactive 33-kDa antigen PG125)W50, W83AF195649/AAF03904, AAQ65845[20, 161, 168, 247]
IhtC – periplasmic binding protein (lipoprotein, putative)W83AAQ65846[20]
IhtD – iron compound ABC transporter, permeaseW83AAQ65847[20]
IhtE – iron compound ABC transporter, ATP-binding proteinW83AAQ65848[20]
TonB-linked hemin receptor Tlr (in W83 contains an authentic frame shift)Transport and binding proteinsW50, W83AF155223/AAD37808[20, 163]
Conserved domain protein (htrA)W83AAQ65824[20]
Iron compound ABC transporter, ATP-binding protein (htrB)W83AAQ65825[20]
Iron compound ABC transporter, permease protein (htrC)W83AAQ65826[20]
Iron compound ABC transporter, periplasmic iron compound-binding protein, putative (htrD)W83AAQ65827[20]
TonB-linked adhesion (Tla)Protein expressed under low level of heminW50Y07618/CAA68897[176]
35-kDa hemin-binding protein and co-aggregation factor (HBP35) or thioredoxin, putativeBinding of hemin381, W83AB059658/BAC53728, AAQ65800[20, 183]
32-kDa hemin-binding proteinBinding of heminW50Not identified[167]
30-kDa (the molecular mass of the heated protein 24 kDa) hemin-binding protein;Hemin- and iron-regulated protein; binding of hemin381Not identified[156]
Fragment 1 of the protein obtained after CNBr digestionAAB46901[156]
Fragment 3 of the protein obtained after CNBr digestionAAB46902[156]
RagA protein (temperature-regulated receptor antigen)TonB-dependent receptor involved in iron uptakeW50, W83AJ130872/CAA10226, AAQ65420[20, 160]
RagB lipoprotein (55-kDa immunodominant antigen)Hemin-binding lipoproteinW50, W83AJ130872/CAA10227, AAQ65421[20, 160]
Hemagglutinin HagACellular processes; hemagglutination of red blood cells381, W83U41807/AAB17128, AAQ66831[20, 108]
Hemagglutinin HagB381, W83Z35494/CAA84627, AAQ66947[20, 99]
Hemagglutinin HagB, degenerate (in W83 contains one or more premature stops or frame shifts)W83PG1674[20]
Hemagglutinin HagC381, W83Z27394/CAA81786, AAQ66949[20, 99]
Hemagglutinin HagD (in W83 contains an authentic frame shift)381, W83U68468/AAB49691, PG1844[20, 108, 118]
Hemagglutinin HagE381, W83AF026946/AAD01810, AAQ66991[20, 118]
Hemagglutinin, putativeW83AAQ65612, AAQ66396[20]
Hemagglutinin-related proteinW83AAQ65642[20]
Hemagglutinin, truncatedW83PG2198[20]
Thiol protease/hemagglutinin PrtT precursor, putativeProtein degradation and hemagglutinationW83AAQ66480[20]
Thiol protease/hemagglutinin PrtT precursor (in W83 contains an authentic frame shift)W83PG1548[20]
Arginine-specific cysteine protease (RgpA)Degradation of proteins, binding of hemin, PPIX, hemoglobin and other iron- and heme-containing proteinsW83AAQ65700[20]
H66U15282/AAA69539[102, 114]
HG66P28784[110, 248]
W50L26341/AAC18876[249, 250]
W 50X82680/CAA57997[104, 120]
Arginine-specific cysteine protease (RgpB)Degradation of proteinsATCC33277D64081/BAA10963[105, 251]
HG66U85038/AAB41892, P95493[119, 122, 123]
Lysine-specific protease (Kgp)Degradation of proteins, binding of hemin, PPIX, hemoglobin and other iron- and heme-containing proteinsW83 HG66AF017059/AAC26523 U54691/AAA99810[252] [112]
HemolysinCellular processes; degradation of red blood cellsW83AAQ66861[20]
Ferrous iron transport protein B, putative (FeoB-1)Transport and binding proteinsW83AAQ66162[20]
Ferrous iron transport protein B (FeoB-2)Transport and binding proteinsW83AAQ66370[20]
FerritinIron storageATCC3327, W83AB016086/BAA31691, AAQ66365[20, 57]
Heme uptake protein A and B or ribosomal protein S1 (90-kDa protein)Proteins similar to B. fragilis hupA and hupB proteins, to hemin-binding protein from Prevotella intermedia, to ribosomal protein S1 from B. fragilis, and to ribosomal 30S protein S1 from B. thetaiotamicronW50, W83AF143945/AAD33925, AAQ66372[B. Ross et al., unpublished, 20]
HemH (porphyrin ferrochelatase)Biosynthesis of cofactors, prosthetic groups and carriersW83AAQ65369[20]
HemD (uroporphyrinogen III synthase)W83AAQ65436[20]
HemN (coproporphyrinogen III oxidase)W83AAQ65670[20]
HemG (protoporphyrinogen oxidase)381, W83AB074530/BAB93002, AAQ67110[20, 71]
PorS proteinA gene responsible for black pigmentation and extracellular protease productionATCC33277D64132/BAA31965, AAQ66244[K. Nakayama and D.B. Ratnayake, unpublished, 20]
PorR pigmentation and extracellular proteinase regulatorRegulatory protein for pigmentation and extracellular protease productionATCC33277D64132/BAA31964, AAQ66245[20, 226]
Ferric uptake regulator (Fur)Regulatory functionA7436, W83AF171224/AAF89746, AAQ65662[T. Karunakaran, W. Simpson, and C.A. Genco, unpublished, 20]
Iron-dependent repressor, putativeRegulatory functionW83AAQ66163[20]

2 Iron and heme availability in the human host

Free iron ions are difficult to obtain by microorganisms since Fe3+ is insoluble in water solutions at physiological pH. Insoluble ferric ion (Fe3+) may be used under acidic conditions by some microorganisms, such as Lactobacillus species [22]; however, a majority of bacteria mediate the transport of ferrous or ferric ion complexed to siderophores or heme. The free Fe3+ concentration in equilibrium with environmental Fe3+–hydroxide polymer is in the order of 10−9 M, which is below the concentration of 10−7 M required to support bacterial growth [23]. In humans, iron is generally not present in the free form, but is sequestered by iron-binding complexes that maintain the concentration of free iron within the host at approximately 10−18 M [24]. The majority of iron is found intracellularly in the form of hemoglobin or ferritin, whereas extracellular iron is bound by transferrin found in serum, and lactoferrin present within mucosal surfaces. A significant amount of iron is coordinately bound to heme (Fig. 1(a) and (b)), which is the prosthetic group of many biologically active proteins (i.e., hemoglobin, myoglobin, cytochromes, catalases, peroxidases) involved in various processes, and is also a regulatory molecule that mediates cellular responses to oxygen, iron and heme itself. The biological utility of heme depends upon the propensity of the iron moiety to alternate between two oxidation states (Fe2+ and Fe3+; Fig. 1 (b)). In an aqueous solution in the absence of proteins or reducing agents, iron protoporphyrin is found in its oxidized form (hemin; Fig. 1 (b)). Due to the toxic (i.e., oxidative) nature of heme, the molecule is not allowed to circulate in its free form within the human host. When heme-containing molecules are degraded, released heme is rapidly bound by molecules such as hemopexin and serum albumin (Kd ∼ 10−12 and ∼10−8 M, respectively) [25, 26]. Due to the action of these scavenging proteins, the concentration of free heme within the human host is maintained at very low levels.

3 Energy requirement for iron/heme transport into Gram-negative bacterial cell

In Gram-negative bacteria, the outer membrane forms a permeability barrier for substrates of >600 Da [27]. Heme is a hydrophobic molecule and has a tendency to form aggregates at physiological pH [28]. The size of these dimers limits the ability of heme to readily transverse the bacterial outer membrane through porin channels. Thus, heme or other iron-containing molecules are not transported across the outer membrane of Gram-negative bacteria by passive diffusion, but are recognized and bound by specific receptors at the cell surface. Heme is further removed from the bacterial receptor and transported into the cell by an energy requiring process.

ATP and GTP, the best-known energy sources, are lacking in the bacterial outer membrane and periplasm. Energy for the transport of iron or heme across the outer membrane into the periplasmic space in most Gram-negative organisms is provided by TonB protein in association with ExbB and ExbD proteins (Fig. 2), which form a complex in a molar ratio of 1:7:2, respectively [2931]. This system energizes outer membrane transport by the proton motive force (the electrochemical potential) of the cytoplasmic membrane. It is well known that TonB–ExbB–ExbD complex mediates the energy input from the cytosolic ATP pool [32], but the precise mechanism by which chemiosmotic potential and ATP catalysis energize the transport remains to be elucidated.

Figure 2

Schematic illustration of the protein complex involved in transduction of energy from the cytoplasmic membrane to the outer membrane of Gram-negative bacteria. Since no energy source exists in the outer membrane, TonB-dependent receptors must utilize energy provided by the proton motive force of the cytoplasmic membrane through TonB–ExbB–ExbD complex. TonB is the energy transducer, while ExbB and ExbD are required for coupling the proton motive force to energization of TonB and for the retrieval of TonB from the outer membrane after the energy transduction process. The TonB box of outer membrane receptors is located in their N-terminal fragments and binds to the region around amino acid 160 of one of the subunit of TonB dimer. The picture provides a static condition of the overall protein complex and does not show dynamic structure alterations that take place upon ligand binding, releasing and subsequent iron/heme transport. OM, outer membrane; CM, cytoplasmic membrane.

The amino acid sequences of TonB proteins identified in various bacteria are homologous to the E. coli TonB protein. The E. coli TonB is a 26-kDa protein, which lies in the periplasmic space with the N-terminal region associated with the cytoplasmic membrane (Fig. 2) [31, 3336]. The N-terminal hydrophobic membrane-spanning region contains a histidine residue (His20), which probably responds to the proton motive force by binding a proton. This process causes a structural change in TonB that converts the protein into an energized form. The resulting form interacts with outer membrane receptors and changes their conformation, which allows for dissociation of heme or Fe3+, Fe2+ complexes from their binding sites and heme or iron diffusion through the channel that is opened at the same time. ExbB and ExbD proteins, forming the TonB–ExbB–ExbD complex, are also bound to the cytoplasmic membrane (Fig. 2). ExbB spans the cytoplasmic membrane three times. Its N-terminal region is present in the periplasmic space, and most of the protein lies in the cytoplasm [34, 35]. ExbD is located in the periplasm, and its N-terminal region is associated with the cytoplasmic membrane [34, 35].

Receptors, which require energy supplied via the TonB system, are termed TonB-dependent or TonB-linked and share amino acid homology in several regions termed TonB boxes. The TonB box represents the domain of the bacterial receptor, which physically interacts with the TonB protein (Fig. 2) [3741]. The C-terminal domain of TonB forms a cylindrical-shaped dimer that interacts with the TonB boxes usually at the region around amino acid residue 160 of the TonB (Fig. 2) [41]. TonB seems to distinguish between the ligand-free and the ligand-loaded states of the outer membrane receptor [42]. Based on an analysis of E. coli outer membrane receptor FhuA it is likely that signaling of ligand-loaded receptor is required for energy transfer and active transport of iron across the outer membrane [37].

The energy-dependence of hemin transport in P. gingivalis [43] suggests the presence of a TonB analog anchored in the cytoplasmic membrane of this bacterium which might function in the transport of heme. These findings have been supported by the sequencing of the P. gingivalis W83 genome [20]. P. gingivalis possesses a gene encoding a putative TonB protein (protein ID: AAQ65946), as well as genes encoding putative ExbB (AAQ65815, AAQ65943) and ExbD (AAQ65816) proteins [20].

Iron/heme acquisition systems in Gram-negative bacteria also require periplasmic transport machinery for the passage of iron or heme through the cytoplasmic membrane [4447]. In contrast to transport across the outer membrane, transport of these compounds across the cytoplasmic membrane is carried out using ATP hydrolysis. All iron sources are transported through the periplasmic space and cytoplasmic membrane by common systems, which belong to the family of ATP-binding cassette (ABC) transporters [45]. These systems utilize a soluble periplasmic substrate binding protein, one or two hydrophobic integral membrane spanning permeases, and one or two hydrophilic proteins with ATPase activity [48]. In P. gingivalis W83, several operons encoding components of ABC transport systems and additional genes for ATP-binding proteins have been identified [20]. Some of these transport systems may be involved in iron/heme transport.

4 Iron storage in P. gingivalis

Iron may be deposited in ferritins (contain iron) present in eukaryota and prokaryota or bacterioferritins (contain heme) found only in prokaryota [4951]. Iron-storage proteins play a dual function in iron metabolism: they serve as an iron reserve and take part in detoxification [52, 53]. E. coli possesses both ferritin and bacterioferritin genes, whose products are highly homologous [54]. E. coli bacterioferritin is composed of a single type of subunit, forming a spherical shell of 24 polypeptides and 12 hemes around the polynuclear mineral iron core. The mechanism for release of iron from the core is unknown. It has been proposed that the iron enters the protein by the reduction of Fe3+ to Fe2+ by the apo-protein and the transport occurs through channels formed in the protein [52, 53]. In contrast, P. aeruginosa has two ferritin genes (bfrA and bfrB), whose products form two types of polypeptide subunits, which are present in variable proportions in the 24-mer [55, 56].

A gene encoding a putative ferritin was found in P. gingivalis W83 strain, and the ftn gene and its product were characterized in P. gingivalis ATCC 33277 strain [57]. The P. gingivalis protein with a molecular mass of 18 kDa formed ferritin-like particles with a spherical structure containing an inner electron-dense core, and non-heme iron was found in this molecule. Analysis of its amino acid sequence showed similarity to Bacteroides fragilis ferritin [58] and ferritins identified in other bacteria. Moreover, conserved amino acid residues (Glu27, Tyr34, Glu61, Glu62, His65, Glu107 and Gln141) critical in iron chelation for eukaryotic ferritin H chains and forming the ferroxidase center [59], were identified in the corresponding sites in the P. gingivalis protein [57]. Ferritin has also been found to be crucial for P. gingivalis to survive in iron-restricted conditions under both hemin and transferrin starvation [57].

Of studied pathogens, P. gingivalis is the-best characterized bacterium which also accumulates iron on the surface of the cell. When grown on blood-containing media, it binds heme as μ-oxo dimer, [Fe(III)PPIX]2O (Fig. 1 (b)) [60, 61]. Some monomeric heme molecules are also present in the surface heme layer, resulting from the equilibrium with μ-oxo dimer of heme [62]. Both compounds undergo aggregation due to hydrophobic effects [6264], which is proposed as a mechanism of pigment accumulation [60]. The ability of P. gingivalis to store hemin at the cell surface appears to provide a nutritional advantage in the iron-limited conditions of a healthy periodontal environment [65, 66].

5 Iron/heme requirements for P. gingivalis growth

Iron has been shown to play a crucial role in the growth and virulence of P. gingivalis [67]. The majority of genes required for the de novo porphyrin biosynthetic pathway are absent in P. gingivalis and the bacterium must acquire PPIX from the environment [6870]. Putative hemD (an uroporphyrinogen III synthase), hemN (a coproporphyrinogen oxidase), hemG (a protoporphyrinogen oxidase) and hemH (a porphyrin ferrochelatase) genes encoding enzymes involved in the PPIX biosynthetic pathway have been identified in the P. gingivalis W83 genome [20], and additionally hemG was identified and characterized experimentally in P. gingivalis 381 strain [71]. However, this does not allow proper heme biosynthesis and the bacterium must employ compensatory mechanisms to overcome the heme limitation problem. Contradictory reports indicate that tetrapyrrole rings can replace hemin to support the growth of P. gingivalis [68] or according to others [72], the PPIX alone does not stimulate the growth of this pathogen. This discrepancy may be explained by the fact that residual endogenous iron levels (about 2.7 × 10−5 M) might be satisfactory for the iron requirement of P. gingivalis in the presence of PPIX [68, 72]. By virtue of the enzyme ferrochelatase, the cells may have the ability to catalyze the insertion of iron into the protoporphyrin ring to form hemin as it was reported for other bacteria [7376]. Ferrochelatase catalyzes the terminal step in heme biosynthesis and inserts ferrous ion into protoporphyrin IX by a mechanism that is still poorly understood. The expression of HemH in P. gingivalis may permit exogenous PPIX, once captured and transported into the cell, to substitute as a growth factor for heme in iron-replete conditions. It is also worth mentioning that HemH might act in the reverse reaction in vivo as a dechelatase by extracting the ferrous ion out of heme, as has been reported for Haemophilus influenzae [74]. The presence of HemN, HemD and HemG in P. gingivalis W83 may impart an ability to generate essential porhyrins from related protoporphyrins by modification of side chains of the tetrapyrrole ring. The capability in vitro to bypass the growth requirement for heme by using other iron sources in combination with exogenous non-iron porphyrins would imply that at least some of the porphyrin capture systems in P. gingivalis can recognize non-iron porphyrins. All these data suggest that as a pathogen living in vivo on a rich medium, P. gingivalis would appear to have lost selective pressure to maintain its ability to make tetrapyrroles de novo. It is likely that transport of heme into the bacterial cell may require less energy then the biosynthesis of the PPIX ring [77], thus, the bacterium has evolved efficient transport systems to import exogenous heme (Fig. 3).

Figure 3

P. gingivalis iron and heme utilization. A schematic overview of major iron and heme acquision systems in P. gingivalis is shown. Only proteins encoded by the genes whose sequences have been deposited in the databases are indicated. The majority of them have also been characterized by mutational and/or biochemical analyses. CM, cytoplasmic membrane; OM, outer membrane; PS, periplasmic space; gingipains Kgp (indicated in black) and HRgpA (indicated in grey) are shown as membrane-associated (larger squares) and excreted, soluble (smaller squares) forms. HmuR, TonB-dependent heme/hemoglobin receptor; IhtA, TonB-dependent iron/heme receptor; Tlr, TonB-dependent heme receptor; RagA, TonB-dependent heme/receptor; RagB, hemin-binding lipoprotein; FeoB-1 and FeoB-2, ferrous iron transport proteins; ABC, ATP-binding cassette transport system, which usually is composed of a soluble periplasmic substrate binding protein, one or two hydrophobic integral membrane spanning permeases, and one or two hydrophilic proteins with ATPase activity.

In vitro, P. gingivalis grows well in a medium containing hemin or hemoglobin as an iron source; however, the concentration of hemoglobin required to support its growth is much lower than that of hemin (1.7 × 10−9 M and 1–5 × 10−6 M, respectively) [78, 79]. The gingival crevicular fluid present in diseased periodontal pockets contains a variety of proteins carrying iron or heme, including transferrin and hemoglobin [80, 81]. In vitro heme can be acquired by P. gingivalis from a range of hemoproteins at low concentrations (<10−5 M), including hemoglobin, hemoglobin bound to haptoglobin and hemin complexed to hemopexin, indicating that this bacterium has a mechanism for removing the heme from the host iron- or heme-binding proteins [72, 82]. Several studies have demonstrated that both [14C]hemin and [59Fe]hemin are accumulated by P. gingivalis. This indicates that the iron and the porphyrin ring (i.e., the entire heme moiety) are taken into the P. gingivalis cell [43].

In addition to heme, transferrin may represent one of the most important sources of iron for P. gingivalis. This protein is found in large amounts in gingival crevicular fluid from patients with destructive periodontitis [80, 81]. Iron-saturated transferrin was found to support the long-term growth of P. gingivalis [78, 83, 84]. Tazaki et al. [84] found that the binding of human transferrin to P. gingivalis cells occurred rapidly, reversibly and specifically (Kd= 1.4 × 10−6 M). This binding was increased when cells were grown in iron-limited conditions. However, specific receptors for transferrin in this organism have not been identified. P. gingivalis mutants deficient in one or several proteases (described below) were used to demonstrate the key role of specific enzymes in degradation of transferrin and its subsequent utilization for growth [82, 85]. The lack of both arginine- and lysine-specific gingipain activities was associated with the absence of degradation of transferrin and the inability of bacteria to grow in the presence of transferrin as iron source. It has also been found that the lysine-specific gingipain activity is more critical than the arginine-specific gingipain activity [82, 85]. It is likely that transferrin is degraded by gingipains and liberated iron is captured by one or more putative iron transport systems recently identified in the P. gingivalis W83 (Table 1) [20].

6 Hemagglutinins and proteases produced by P. gingivalis

6.1 Hemagglutinins and hemolysins

Porphyromonas gingivalis possesses both hemagglutinating and hemolytic activities [8694]. The ability to agglutinate and lyse red blood cells is a feature that distinguishes this organism from other black-pigmented anaerobes [95]. Hemagglutinins/adhesions not only mediate the adsorption and invasion of bacteria into host cells, but also take part in iron/heme utilization [9698]. P. gingivalis possesses five major hemagglutinin genes (hagA, hagB, hagC, hagD and hagE) [99]. The gene sequences for six newly identified putative hemagglutinin-like proteins have also been identified in the P. gingivalis W83 genome (Table 1) [20]. Four of these are duplications in the genome of HagA and HagD adhesion domain-related sequences.

Various hemolysins have previously been identified and characterized from P. gingivalis cells and extracellular vesicles [9092]. The activity of some of these proteins has been shown to be regulated by environmental hemin [90]. A new hemolysin for the release of iron and PPIX has been recently found in P. gingivalis W83 strain (AAQ66861) [20]. This protein shows homology only to the hemolysin previously characterized in Prevotella melaninogenica [100].

6.2 Proteases

In addition to hemolytic activity, several proteases have been identified in the P. gingivalis W83 genome [20]. This bacterium possesses three major proteolytic activities: (1) trypsin-like, (2) collagenolytic, and (3) glycylprolyl peptidase activities. Numerous proteins with thiol-dependent trypsin-like cleavage specificity can be isolated from cells and culture supernatants, but biochemical and genetic analyses indicate that many of these are derived by proteolytic processing of larger, cell-associated, primary gene products [101, 102]. Recent studies indicate that P. gingivalis contains a family of genes, which comprise part of the protease sequences and also possess hemagglutinin/adhesion domains [20, 101108]. The best-characterized of these enzymes are the cysteine proteases known as gingipains (Fig. 4). It is likely that the presence of the hemagglutinin/protease activities plays a significant role in binding and lysis of red blood cells, and that proteolysis may be an important non-specific step in the utilization of iron- and heme-binding proteins by this pathogen.

Figure 4

Protein structure of P. gingivalis gingipains. The gingipains Kgp and HRgpA are produced as polyproteins composed of the non-covalent complexes of the catalytic domain and hemagglutinin/adhesion domains (HA) in contrast to RgpB, which possesses the enzymatic activity only. Homologous regions encompassing three hemagglutinin/adhesion domains are shown by similar patterns. The second hemagglutinin domain, comprising the HbR receptor, is shown in black. The dotted region in RgpB indicates the Ig-like motif. The C-terminal parts involved in secretion and attachment of gingipains to the outer membrane are shown as white circles.

6.2.1 Gingipains

Several P. gingivalis gingipain genes identified in various strains, termed by different names and accession numbers, have been deposited in databases. In addition, heterogenous nomenclature of the genes and their products has been employed in the literature. Comparison of sequences of these protease genes and mutants analysis showed that the genetic structure of the gingipains in P. gingivalis is less complicated than suspected and new names as well as the classification of these enzymes have been recently suggested [109]. Detailed characterization of gingipains has been reviewed by many authors (for references see below) and will not be covered extensively in this review.

Two major types of gingipains were found in P. gingivalis, which are usually encoded by three genes (Table 1; Fig. 4). The arginine-specific gingipains (Arg-gingipains; cleave polypeptide chains after arginine residues) are encoded by two related genes rgpA and rgpB, while the lysine-specific gingipain (Lys-gingipain; cleaves polypeptide chains after lysine residues) is encoded by the single kgp locus [102, 110113]. The P. gingivalis lysine-specific gingipain K (Kgp), and an arginine-specific gingipain R1 (HRgpA), are purified as non-covalent complexes of the catalytic domain associated with four polypeptide chains derived from the hemagglutinin domains (Fig. 4) [102, 110118]. The protease regions of Kgp and HRgpA are divergent, but their hemagglutinin domains are very similar to each other [110, 111]. The latter also share significant amino acid sequence homology to hemagglutinin A (HagA), a protein containing four contiguous direct repeats each of which contains a functional hemagglutinin domain [108]. Arg- and Lys-gingipains are produced as soluble (extracellular) and membrane-associated forms [106, 109, 117, 119, 120]. It has been proposed that the gingipain polyproteins are secreted and then attached to the outer membrane, possibly through the C-terminal fragment, since this segment is not found in the soluble RgpA or Kgp released from the outer membrane [114, 115, 121].

In contrast to Kgp and HRgpA, a second arginine-specific gingipain R2 (RgpB) contains only the catalytic domain (Fig. 4) [122]. The tertiary structure of RgpB indicates that this molecule is composed of two distinct regions [123]. The N-terminal catalytic fragment possesses the fold typical for α-hydrolases with topological similarity to caspases, and the C-terminal fragment contains an Ig-like motif. The catalytic domain of RgpB is almost identical with HRgpA indicating that any differences in enzymatic activity between these two enzymes may be due to the presence of the hemagglutinin/adhesion domains present in HRgpA. However, recent data showed that changes in four amino acids around the active site of HRgpA relative to RgpB may also play a role in their different activities [124].

The Arg-gingipains play an important housekeeping role in the proteolytic processing and maturation of other surface proteins, including Kgp [125, 126]. Gingipains have also been linked directly to disease pathogenesis due to their ability to degrade host structural and defense proteins [21]. Recently, it has been shown that attachment and detachment of P. gingivalis to epithelial cells were mediated by gingipain adhesion and Rgp catalytic domains, respectively [127, 128]. The adhesion/hemagglutinin domains of Kgp and HRgpA, as well as HagA have been shown to be involved in co-aggregation with other bacteria [11], the process facilitating P. gingivalis colonization in periodontal pockets. In addition, hemagglutinin domains of gingipains have been implicated in tissue colonization either directly through adhesion to extracellular matrix proteins [94, 95] or indirectly by processing the fimbrillin subunit of fimbriae [111].

Gingipains are also considered to be important virulence factors due to their ability to bind heme and heme-containing proteins. Kgp and HRgpA have been demonstrated to bind hemin, protoprophyrin IX, and other porphyrins and metalloporphyrins [88, 129, 130]. They are also able to degrade hemoglobin [82, 87, 131, 132], haptoglobin and hemopexin [82], transferrin [82, 85], and lactoferrin [133, 134]. This role of gingipains has been extensively studied through mutational analysis [135139]. Generally, the gingipain mutants are less or non-pigmented, suggesting that the accumulation of iron/heme on the cell surface has been affected. Chen et al. [137] isolated and characterized non-pigmented mutants that had a transposon Tn 4351 DNA [140] within kgp gene. The mutants were unable to bind and degrade hemoglobin. Simpson et al. [138] found that a non-pigmented P. gingivalis mutant has an insertion sequence element IS 1126 [141] at the promoter region of kgp. The resulting MSM-3 mutant [138] did not express Lys-gingipain activity and did not utilize hemin and hemoglobin for growth. In contrast, it expressed increased Arg-gingipain activity. A P. gingivalis rgpA rgpB kgp mutant was unable to grow in the presence of serum as a sole iron source, while an rgpA kgp hagA mutant was significantly diminished in hemagglutination and hemoglobin binding activities [142]. These results strongly confirm the involvement of gingipains in P. gingivalis pigmentation and indicate that the hemagglutinin/adhesion domains of gingipains are important for P. gingivalis cells to agglutinate red blood cells and bind hemoglobin, allowing for heme acquisition.

6.2.2 HbR receptor

The gingipains appear to be mostly involved in the acquisition of hemoglobin, as compared to other iron/heme sources successfully utilized by P. gingivalis. Studies by Nakayama et al. [143] have revealed that the hemagglutinin domain of HRgpA contains a 19-kDa protein (HbR; Fig. 4), which is responsible for the majority of the gingipains properties concerning hemin/hemoglobin utilization. The HbR protein is encoded by an internal region (termed HGP15 or HA2 domain) of HRgpA, and is also found in Kgp and HagA. The HA2 domain was shown to bind hemoglobin (Kd ∼ 10−9 M) and hemin (Kd ∼ 10−8 M) very efficiently [143]. Overexpression of this domain in E. coli permitted this organism to bind hemoglobin in a pH dependent manner. Mutation of the HA2 region resulted in the production of non-pigmented P. gingivalis cells and decreased ability of these cells to bind hemoglobin [143]. Recent studies have also shown that deletion of the portion of the kgp gene, which encodes HA2 domain, resulted in generation of pigment-less P. gingivalis mutants [144].

While some proteins involved in heme capture bind directly to the iron center, the P. gingivalis HbR recognizes heme by a mechanism that is porphyrin-mediated, since in vitro porphyrin binding to this domain is iron independent [130]. Porphyrins that differed from PPIX in only the vinyl group of the tetrapyrrole ring (e.g. deuteroporphyrin IX) showed comparable effects in competing with hemoglobin for the HbR and stimulated growth recovery of bacteria, while PPIX derivatives with modifications at both propionic acid side chains did not compete with hemoglobin for the HbR and did not support bacterial growth. The affinity of gallium deuteroporphyrin IX (GaDPIX; Fig. 1 (c)) for the HbR was comparable to other porphyrins, but surprisingly, this metalloporphyrin supported the growth of P. gingivalis. In contrast, Stojiljkovic et al. [145] showed that porphyrin capture and transport systems may act as an entrance for non-iron metalloporphyrins such gallium protoporphyrin IX (GaPPIX; Fig. 1 (c)), a compound exhibiting a bacteriostatic activity. GaPPIX inhibits the growth of microorganisms, particularly those expressing active heme transport systems, due to replacement of iron with gallium in cytochromes. It is likely that in P. gingivalis gallium may be removed from deuteroporphyrin by a ferrochelatase, thus inactivating the bacteriostatic effect of GaDPIX [130].

HbR has also been demonstrated to influence iron acquisition from lactoferrin. The concentration of lactoferrin is increased in the gingival crevicular fluid of patients with periodontal diseases [146, 147] and one would suspect that it may serve as an iron source for periodontal pathogens. However, to date, contradictory results concerning lactoferrin utilization as an iron source by P. gingivalis have been published. P. gingivalis cells have an ability to adsorb lactoferrin [148]. Several reports [72, 83, 149, 150] have shown that this bacterium may utilize lactoferrin as an iron source for growth. However, other investigators demonstrated that lactoferrin was not able to support the growth of P. gingivalis [151]. Interestingly, growth of P. gingivalis cells in a rich medium containing hemoglobin, but not hemin, was inhibited by lactoferrin [151]. Moreover, lactoferrin inhibited P. gingivalis aggregation [152], which is considered to be a virulence factor and a putative protective mechanism against the cellular defenses of the host in the gingival crevice. Usually, lactoferrin chelates iron and may compete with microorganisms in iron acquisition [153]. This process is mediated by the N-terminal fragment of lactoferrin, lactoferricin. Studies performed by Shi et al. [151] may at least in part explain these contradictory results. Lactoferrin has been shown to bind to and remove the HbR polypeptide from the P. gingivalis cell surface, which resulted in P. gingivalis growth retardation. An analysis of truncated HbR mutants suggests that the cationic lactoferricin region of lactoferrin may interact with the N-terminal anion-rich fragment of the HbR in an electrostatic action, loosen the non-covalent bond between HbR and other components of the complex and release the protein from the cell surface [151].

7 Heme- and hemoglobin-binding proteins other than gingipains identified in P. gingivalis

7.1 Binding of hemoglobin and hemin to P. gingivalis cells

Porphyromonas gingivalis appears to lack siderophore activities [20] and must therefore use alternate mechanisms to sequester and transport exogenous iron and heme (Fig. 3). In addition to gingipains, which can bind hemin and hemoglobin, a variety of other cell envelope heme/hemoglobin-binding proteins have been described in this anaerobic bacterium (Table 1) [e.g. 138,154-163]. Although hemoglobin is immediately bound by haptoglobin in the host environment, the hemolytic and proteolytic capabilities of P. gingivalis may contribute to the degradation of this host hemoglobin-sequestering protein and liberate hemoglobin for binding and heme uptake into the cell. Amano et al. [164] found that binding of human hemoglobin to P. gingivalis 381 strain occurred rapidly, reversibly, and specifically, with an apparent Kd of 10−6 M. Hemoglobin binding was inhibited by unlabeled human hemoglobin but not by hemin and protoporphyrin IX. The binding was only partially inhibited by human serum albumin, transferrin, lactoferrin, catalase, and cytochrome c, suggesting that hemoglobin binding may not only occur through the heme moiety.

Tompkins et al. [165] have shown that heme-starved P. gingivalis cells express high- and low-affinity heme-binding activities (Kd ∼ 10−10 and ∼ 10−7 M, respectively). Bramanti and Holt [154, 155, 166] have demonstrated that growth of P. gingivalis W50 in the presence of normal to excess hemin (i.e., hemin-replete conditions) resulted in the repression of several surface proteins. In contrast, under heme-limited conditions (i.e., hemin-deplete conditions) the induction of at least 10 outer membrane proteins was observed [158, 159]. Two of these proteins, at approximately 26 kDa (OMP26) and 80 kDa, were recognized by SDS–PAGE analysis as major hemin-regulated P. gingivalis W50 outer membrane proteins [154, 158, 159]. OMP26 has been suggested to bind hemin and to move across the outer membrane depending upon the heme concentration in the medium. Two other heme-binding proteins have been identified in P. gingivalis outer membrane when cells were grown under conditions of hemin limitation [156, 167]. Smalley et al. [167] identified a 32-kDa hemin-binding protein, and Kim et al. [156] found that the expression of a 30-kDa hemin-binding protein was hemin and iron regulated. Both proteins were identified by staining of SDS–PAGE gels with the chromogenic substrate tetramethylbenzidine, a method that utilizes the intrinsic peroxidase activity possessed by heme. But in the case of these proteins neither their genes nor their mutants have been examined. Thus, the role of these proteins in heme transport has not been defined.

7.2 IhtB protein

Porphyromonas gingivalis ihtB (iron heme transport) protein has formerly been reported as an antigenic 30-kDa protein (Pga30) [168]. It has also been proposed that IhtB may represent the 30- and 32-kDa heme-binding proteins reported by Kim et al. [156] and Smalley et al. [167], respectively. This hemin-binding lipoprotein is thought to be anchored to the outer membrane and is potentially involved in iron assimilation. In contrast, Roper et al. [19] postulated that the IhtB may be an intracellular cobalt chelatase. Analysis of the deduced amino acid sequence of the IhtB protein showed significant similarity to the Salmonella typhimurium protein CbiK, a cobalt chelatase that is structurally related to the ATP-independent family of ferrochelatases [169]. It is also likely that IhtB may function to remove iron from heme prior to uptake by P. gingivalis. The ihtB gene is located in an operon that also encodes an iron transport system containing IhtA (a TonB-linked receptor homologous to iron siderophore and vitamin B12 receptors), IhtC (a periplasmic binding protein), IhtD (a permease), and IhtE (an ATP-binding protein) [168]. The ihtABCDE operon shows sequence similarity to ABC systems identified in other bacteria. The proteins encoded by this operon may act as a two component receptor system that removes iron from transferrin prior to transport into the periplasm, similar to proteins of pathogenic Neisseria species [170, 171].

7.3 RagB and RagA proteins

Another P. gingivalis protein, RagB, which may be involved in heme binding, has been characterized [160, 172]. It has been identified as the 55-kDa immunodominant antigen of W50 and W83 strains [160, 172], and recently has been found in ATCC33277 strain [173]. The protein has been implicated in the severe destructive disease process of pathogenic strains of P. gingivalis [160, 174]. RagB has been classified as a lipoprotein because it exhibits a signal peptide of the lipoprotein type [175]. Based on the similarity to proteins found in other bacteria [160, 176] and its recognition by sera from patients infected with P. gingivalis [160, 172], the protein is thought to be anchored to the exterior side of the outer membrane. RagB is encoded in an operon together with RagA protein. The N- and C-terminal regions of RagA showed homology to TonB-linked receptors involved in iron uptake in other Gram-negative bacteria. The co-localization and co-transcription of ragA and ragB strongly suggest that they would be functionally linked and their products may form a complex on the surface of P. gingivalis cells [160]. The genetic arrangement at the ragAB locus is similar to those for the lactoferrin and transferrin binding systems in Neisseria and Haemophilus species, respectively [160, 177179]. In these organisms, a TonB-linked outer membrane receptor is also co-transcribed with an outer membrane lipoprotein, and the resulting complex is involved in the active transport of iron from lactoferrin and transferrin. However, no significant homology between the RagB and analogous transferrin and lactoferrin binding lipoproteins, Tbp2 and Lbp2, has been found. Database searches with RagA revealed high homology to SusC, a TonB-linked receptor involved in maltose uptake in Bacteroides thetaiotamicron [180]. This suggests that RagAB complex may participate in either the uptake or recognition of a specific carbohydrate or glycoprotein. Since P. gingivalis is an asaccharolytic bacterium, it seems that carbohydrate uptake may be linked to an anabolic process. This may be also explained by a hypothesis that the ragAB may have arisen by a horizontal gene transfer from a foreign source, such as other bacteria of the periodontal flora [172].

7.4 HBP35 (35-kDa protein)

The gene for a 35-kDa protein (HBP35), which appears to confer colonizing activities to P. gingivalis, has been identified and characterized in the 381 strain [181183]. The HBP35 protein was capable of binding hemin and with lower efficiency selected metalloporphyrins, suggesting that the binding of porphyrins might be affected by the metal ion differentials [183]. The HBP35 did not bind hemoglobin and little binding of PPIX was observed. The protein contains a typical heme regulatory motif (YCPGGK; consensus KCPVDH) that functions as heme direct binding site [184]. The Cys-Pro residues are commonly found in heme-binding proteins to mediate the regulation of their activity by heme [184, 185]. Insertional inactivation of the hbp35 gene produced a beige mutant, which showed decreased virulence. It exhibited little hemagglutination, decreased Arg- and Lys-specific proteolytic activities, lower self-aggregation and the inability to co-aggregate with human gingival cells [186]. These results suggest that hemin binding capability of this protein may be important for the expression of P. gingivalis virulence.

Interestingly, in the P. gingivalis W83 this protein has been recognized as a putative thioredoxin (AAQ65800), a protein which can be reversibly oxidized and reduced [187]. The HBP35 contains the motif called CxxC (WCGYC; consensus WCGPC), which is conserved among thioredoxin molecules isolated from both bacteria and mammals [188]. Typically, thioredoxins or thioredoxin-like proteins from different organisms show very low homology, and their targets or substrates share no apparent structural similarity. The reduced form of E. coli thioredoxin is required for defense against H2O2, possibly by scavenging free radicals generated in the cell [187]. In addition, thioredoxin proteins have been suggested to remove peroxides directly, as well as regulate protein function by binding to target proteins without affecting their redox state. It is likely that HBP35 may play a protective role in oxidative stress in P. gingivalis, similar to other proteins such B. fragilis thioredoxin peroxidase [189].

7.5 Tla and Tlr proteins

Several other putative outer membrane proteins, which may contribute to hemin acquisition in P. gingivalis, have been described. Aduse-Opoku et al. [176] initially reported on the identification of the tla (TonB-linked adhesion) gene from P. gingivalis W50 strain. The C-terminal region of the Tla protein exhibited 98% identity to the arginine-specific protease Rgp, and the N-terminus showed regional similarity to TonB-linked receptors, which are frequently involved in translocation of hemin, iron siderophores, or vitamin B12 in other bacteria. An isogenic tla mutant was demonstrated to grow with high levels of hemin and was unable to grow in media containing low concentrations of hemin (<2.5 mg/liter). Hemin-depleted cells of this mutant failed to respond to hemin in an agar diffusion plate assay, suggesting a role for Tla in hemin acquisition and utilization [176]. In addition, the tla mutant produced significantly less arginine-specific protease activity than the wild type strain.

Slakeski et al. [163] screened a genomic library of P. gingivalis W50 strain and found that the previously reported nucleotide sequence of the tla gene consisted of two separate sequences. The C-terminus of the protein, designated Tlr (TonB-linked receptor), possessed no similarity to the Rgp protease. In addition, the full sequence of the tlr gene is present in the P. gingivalis W83 genome, whereas the sequence of tla is absent. The tlr gene is located in the operon (htrABCD) that encode a putative iron compound ABC system with sequence similarity to heme transport systems of other bacteria [163]. However, the involvement of Tla and Tlr in iron/heme utilization in P. gingivalis remains still not clear. The investigators were unable to find tla transcript [176], and in the W83 strain the tlr gene sequence shows an authentic frame shift. These data suggest that the genes could undergo conversions resulting from an insertional sequence transposition. Interestingly, copies of IS 1126 have been found in flanking regions of tla and kgp genes [176]. This rearrangement strategy might cause a novel combination of available protein capacities by the fusion of different functional domains.

7.6 Hemin regulated protein HemR

Karunakarun et al. [190] have described an additional gene, hemR, which has been proposed to encode an outer membrane protein in P. gingivalis 53977 strain. The N-terminus of the hemR gene exhibited homology to genes involved in iron acquisition from several Gram-negative microorganisms. However, the C-terminus of the hemR gene showed similarity with a prtT gene, which encodes an arginine-specific protease, distinct in nucleotide sequence from the major arginine-specific proteases possessed by P. gingivalis, HRgpA and RgpB. Since an isogenic hemR mutant could not be created and characterized, and the hemR sequence is absent in the W83 or A7436 strains [20, 138], a role for the HemR in hemin accumulation in P. gingivalis cannot be established. It is also likely that hemR, similar to tla, may have resulted from a rearrangement caused by transposition of an insertional element.

7.7 Heme/hemoglobin-binding receptor HmuR

In addition to the above-described heme/hemoglobin-binding proteins, an outer membrane receptor HmuR (hemin utilization receptor) involved in the acquisition of both hemin and heme from hemoglobin in P. gingivalis A7436 strain has been identified [138, 162]. The HmuR sequence shows extensive homology to TonB-linked receptors involved in iron/heme uptake in other bacteria [162]. The amino terminal region of the hmuR gene exhibits 100% identity with the previously reported hemR gene [190]; however, the C-terminal regions of hemR and hmuR are strikingly dissimilar. Despite the fact that previous studies have determined that hemR is present in strains 53977, W50, and 381 [190], it is absent in P. gingivalis A7436 and W83 strains [162]. This suggests that in the latter strains, heme transport can occur independently on HemR.

The role of the hmuR gene in heme accumulation has been extensively studied through mutant construction [129, 139, 162, 191] and biochemical analysis [129, 191]. The hmuR mutant was defective in growth in the presence of either hemin or hemoglobin as iron sources, and exhibited a decreased ability to bind hemin and hemoglobin [162]. E. coli cells expressing the HmuR protein have the ability to bind hemin and hemoglobin [162]. Purified recombinant HmuR has been shown to bind to hemin and hemoglobin, as well as with lower capability to hemin-hemopexin, haptoglobin-hemoglobin and hemin-albumin complexes, and did not bind to human serum albumin or transferrin [129]. E. coli cells expressing membrane-associated recombinant HmuR were capable of binding metalloporphyrins such as hemin (Kd= 2.4 × 10−5 M), zinc protoporphyrin IX, copper protoporphyrin IX with higher affinity than protoporphyrin IX alone. All these results indicate that HmuR may bind heme through histidine residues, similar to other hemoproteins [77]. Amino acid comparisons of the conserved motifs of several different heme/hemoglobin receptors and the P. gingivalis HmuR protein revealed that HmuR contains highly conserved amino acid residues including invariant histidine residues (His95, His434), glutamic acid residues (Glu427, Glu448, Glu458), the YRAP (consensus FRAP), and NPDL (consensus NPNL) amino acid motifs, which may be involved in heme/hemoglobin binding. The histidine residues have been previously found to be critical for heme binding to other bacterial heme/hemoglobin receptors [192]. To define specific amino acid residues/motifs for heme utilization, a series of hmuR mutants was constructed by site-directed mutagenesis and characterized by functional analysis [191]. The experimental data suggest that His95 and His434, as well as the conserved motifs NPDL and YRAP seem to be involved in the utilization of heme from serum hemoproteins.

To further analyze structural bases of HmuR function, a three-dimensional model of HmuR was constructed using a homology modeling method and the conserved residues of the HmuR protein were analyzed (Fig. 5) [191]. Comparison of the HmuR model with the crystal structures of E. coli ferric enterobactin receptor FepA (PDB ID: 1FEP) [193] and cobalamin receptor BtuB (1NQE) [194] indicates their structural similarity (Fig. 5). Each of these TonB-dependent receptors exhibit barrel-like structures composed of two polypeptide domains. The first domain is a β-barrel unit built up of β-strands connected by long loops on the extra-cellular side and by short turns on the periplasmic side. The second domain is a globular N-terminal domain which folds into the barrel pore. The β-barrel structures are relatively conserved among HmuR, FepA, BtuB, and FhuA (Fig. 5), indicating their similar role for ligand transport across the outer membrane. In contrast, the extra-cellular loops of these receptors exhibit high variety, indicating that specificity for ligand binding lies within these regions in TonB-dependent receptors. The presence of conserved histidines and the NPDL motif on the extra-cellular loops of HmuR indicated their probable function for hemoprotein binding, while the presence of the conserved YRAP motif on the β-barrel strand may indicate its role for heme transport. The homology model of HmuR provides structural information and confirms the results from mutational analysis [191].

Figure 5

Comparison of HmuR homology model with the crystal structures of TonB-dependent receptors FepA (a), BtuB (b) and FhuA (c). The molecular modeling of HmuR was based on the crystal structures of two homologous E. coli proteins, the ferric enterobactin receptor FepA (PDB ID: 1FEP) and the cobalamin receptor BtuB (1NQE), using three-dimensional structure-modeling program MODELLER [191]. The structural comparison of HmuR with E. coli FepA (1FEP), BtuB (1NQE), and ferric hydroxamate receptor FhuA (1QJQ) was performed using the DeepView/Swiss-Pdb viewer software. All these proteins show similar barrel-like structures, with conserved β-strand barrel structure and variable extra-cellular loop regions. The figure shows the superimposition of HmuR (blue) with FepA (yellow), BtuB (red), and FhuA (green), respectively.

7.7.1 Interaction between HmuR and gingipains

Several studies suggest that soluble Kgp may function as a heme scavenger or hemophore-like protein, and may be capable of interacting with an outer membrane receptor [82, 88, 130, 143]. Interestingly, ELISA studies detected the interaction between recombinant HmuR and purified Kgp and HRgpA [129]. Recombinant HmuR appeared to more readily bind to Kgp, as higher concentrations of HRgpA were needed to observe the interaction. Since the reactivity between HmuR and RgpB was not detected, one may suggest that the interaction between HmuR and Kgp or HRgpA actually occurs at the hemagglutinin domains of the gingipains [129].

The precise role of the Kgp–HmuR complex in heme acquisition was further elucidated through the construction and analysis of a double hmuR kgp mutant [139]. The mutant exhibited reduced ability to grow with either hemin or hemoglobin as iron sources [139]. The removal of the proteins, either singly or in combination, diminished the ability of the organism to bind hemin and hemoglobin [139, 162]. It has been reported that P. gingivalis kgp mutants display a non-pigmented colony phenotype, and that this phenotype correlates with a decreased ability to bind hemoglobin [137]. In agreement with these studies, the kgp and the double kgp hmuR mutants constructed in P. gingivalis A7436 strain were observed to be non-pigmented when grown on blood agar plates [139]. In contrast, the hmuR mutant was darkly pigmented, most likely due to the absence of HmuR and significantly decreased ability to transport the heme moiety across the outer membrane, which caused an excessive accumulation of hemoglobin and hemin by Kgp and HRgpA on the bacterial cell surface [162].

Physiological and biochemical studies demonstrated that Kgp may play a more crucial role in the utilization of hemin than previously suspected. Based on hmuR and kgp mutants analysis [139, 162], the finding that hemin/hemoglobin receptor HmuR interacts with gingipains Kgp and HRgpA [129], and characterization of the HbR [130, 143], a hypothesis that gingipains may function as hemophore-like proteins that capture and shuttle heme to HmuR has been proposed. It is likely that the gingipain first binds hemoglobin and then degrades the globin molecule with the ultimate release of hemin that is transported into the cell. This is consistent with the report showing the presence of a hemoglobin protease (Hbp) in a human pathogenic strain of E. coli that has been proposed to function in a similar manner [195]. Numerous investigators have reported that Kgp binds hemoglobin (Kd= 1.45 × 10−8–2.1 × 10−9 M) [88, 131, 143] and hemin [88, 129], and the HbR receptor has the ability to bind hemoglobin and hemin (Kd= 2.1 × 10−9–4.9 × 10−8 M and Kd= 1.6 × 10−8 M, respectively) [88, 130]. The tight binding of hemoglobin and hemin and the ability of Kgp to interact with HmuR may indicate that Kgp serves to release and bind hemin and deliver it to HmuR, similar to a hemophore HasA secreted by S. marcescens [196, 197]. Several Gram-negative bacteria produce extracellular heme/hemoglobin/hemopexin binding proteins (hemophores), which are secreted as apo-proteins. They function to capture and shuttle heme to a specific outer membrane receptor. HasA hemophores from Serratia marcescens [196200] and Pseudomonas aeruginosa [201] extract heme from hemoglobin, while HuxA from H. influenzae removes heme from heme-hemopexin complex [202, 203]. The outer membrane receptor HasR from S. marcescens alone can bind and transport heme directly; however, the efficiency of heme uptake is increased when HasR cooperates with HasA [197]. It is also likely that Kgp–HmuR complex in P. gingivalis may represent two-component system similar to those present in other Gram-negative bacteria, which are composed of a TonB-linked outer membrane receptor and an accessory lipoprotein, such as the Neisseria transferrin, lactoferrin [178], hemoglobin [204], and haptoglobin-hemoglobin [205] receptors.

8 Other putative P. gingivalis proteins that may be involved in iron and heme utilization

RT-PCR analysis of total P. gingivalis RNA revealed that the hmuR gene is co-transcribed with a hmuY gene located immediately upstream [162]. Gene expression under the control of one promoter for hmuY and hmuR was supported by the absence of −10 and −35 promoter sequences upstream from the putative transcriptional start site of the hmuR gene, and then confirmed by the sequencing of the P. gingivalis W83 genome. The hmuY gene encodes a 142-amino acid protein. Internal regions of HmuY exhibited 76% and 89% identities to peptides derived from a 24-kDa protein purified from P. gingivalis 381 strain (Table 1), which has been demonstrated to bind hemin and has been shown to be regulated by hemin and iron [156]. No significant homology of HmuY was found to other proteins whose sequences are deposited in protein databases. Sequence analysis of the predicted 15.5-kDa HmuY protein revealed an ATP/GTP-binding site motif A (P-loop), suggesting that it may bind phosphate and function as an ATP-binding protein and/or DNA-binding regulatory protein. The transcription of the hmuY hmuR operon has been demonstrated to be repressed by hemin [162], and preliminary data suggest that HmuY protein may be involved in heme utilization in P. gingivalis A7436 in cooperation with HmuR (T. Olczak, A. Sroka, and C.A. Genco, unpublished data).

In addition to the previously described proteins that may be engaged in iron utilization in P. gingivalis, it seems that this bacterium also uses other systems such as two ferrous iron transport proteins (FeoB-1 and FeoB-2; Table 1) [20]; however, the precise role of these proteins in iron uptake remains to be experimentally elucidated. Although P. gingivalis W83 possesses multiple iron/heme-binding putative proteins, it is likely that the bacterium may use only a few of them, as has been reported for other bacteria [206].

9 Regulation of expression of genes involved in iron/heme utilization in P. gingivalis

Expression of bacterial iron uptake systems is often regulated in response to the level of iron in the microenvironment [207]. Failure to acquire iron limits the capability of the bacterial pathogen to colonize and proliferate. Iron overload, on the other hand, can be harmful due to the toxic properties of this nutrient. For a bacterium to adhere to the host tissue surfaces, replicate, and colonize host cells, it is critical that virulence genes are expressed during certain periods of the infection process. In many microorganisms, the regulation of iron uptake genes and virulence factors is under the control of the ferric uptake regulator (Fur) protein (Fig. 6) [208, 209]. Fur contains the N-terminal domain, which functions in DNA binding and the C-terminal region responsible for metal binding [210, 211]. The latter is also important for dimerization of the protein. E. coli Fur possesses a regulatory iron-binding site and at least one zinc-binding site in each monomer [212]. Zinc is invariably associated with the DNA-binding domain and is required for maintaining the polypeptide fold, which recognizes a specific DNA sequence. Fur acts as a transcriptional repressor of iron-regulated promoters by virtue of its Fe2+-dependent DNA-binding activity [213]. Under iron-rich conditions Fur binds the divalent iron ion, acquires a configuration capable of binding target DNA sequences designated as Fur or iron boxes (E. coli consensus Fur box sequence: 5′-GATAATGATAATCATTATC-3′), and inhibits transcription from virtually all genes and operons repressed by the metal (Fig. 6). Recently, it has been suggested that two Fur dimers may bind at each Fur box on opposite faces of the double DNA helix [214]. In contrast, when iron is scarce, the equilibrium is displaced to release Fe2+, the RNA polymerase accesses cognate promoters, and the genes are expressed (Fig. 6). Fur also positively regulates genes of iron-containing proteins by repressing synthesis of an anti-sense RNA [215217].

Figure 6

Schematic representation of Fur-mediated gene repression. Iron responsive genes are under control of the transcriptional repressor Fur, which binds DNA under iron replete conditions and negatively regulates transcription of iron transport genes. Fur is a dimeric metalloprotein in the presence or absence of Fe2+ that changes conformation in response to divalent transition metal cations. RNA pol., RNA polymerase.

Homologs for the fur gene have been described for Gram-negative and Gram-positive bacteria. Most of these homologs are capable of complementing an E. coli fur mutant, confirming that the molecular mechanisms, which control transcriptional regulation by iron, are shared by many microorganisms. A P. gingivalis fur homolog has also been identified (W. Simpson and C.A. Genco, unpublished data). The P. gingivalis fur gene (501 bp) in A7436 strain encodes a 19.2-kDa protein, which reveals homology to Fur proteins from a variety of Gram-negative bacteria. The P. gingivalis Fur protein can also partially complement the functional activity of E. coli Fur (W. Simpson and C.A. Genco, unpublished data). Typically, Fur proteins are rich in histidine residues which have been demonstrated to be involved in the binding of ferrous iron [218]. In contrast, the Fur protein of P. gingivalis contains only one histidine residue at amino acid position 24. This suggests that this histidine may be critical to the iron/metal-binding capability of the protein, or alternatively, other amino acids may be involved in the binding of metal. Several studies have indicated the sequence CXXCG found within the carboxy terminus of Fur proteins that may function as a metal binding motif [211]. Analysis of the primary structure of the P. gingivalis Fur protein revealed a similar motif (CTECA) located within its C-terminal region.

Located 90 bp upstream of the start codon of P. gingivalis fur, a region (5′-GATAAAAACCAATCGGGTT-3′), which exhibited 42% homology to the consensus Fur box of E. coli, was identified suggesting an autoregulatory function of Fur protein. In P. gingivalis A7436, a 19-bp putative Fur box was also identified (5′-GATAATTATGAAAAAAATC-3′) upstream of the hmuY start codon. This fragment exhibits 68% identity to the E. coli consensus Fur binding sequence. The presence of the putative sequence suggests that Fur protein may regulate the expression of both hmuY and HmuR genes. The P. gingivalis Fur protein is currently being investigated also through mutational analysis and biochemical characterization of an overexpressed protein. Results from these collective studies will greatly contribute to identification of genes of the Fur regulon and understanding of iron-responsive gene regulation in P. gingivalis.

A novel putative iron-dependent repressor has also been found in P. gingivalis W83 strain (AAQ66163), but the detailed function of this regulatory protein has not been examined. This protein shows homology to FeoA, a protein involved in Fe2+ transport in other bacteria [219]. It is also similar to iron dependent repressors that include the diphteria toxin repressor DtxR from Corynebacterium species [220], and transcriptional repressor SirR from Staphylococcus species [221].

Iron has been well documented to regulate the expression of virulence factors of pathogenic bacteria, with increased expression occurring under low-iron conditions [207]. In P. gingivalis, hemin binding and transport are induced also by hemin [43, 158, 159], suggesting that the regulation of genes involved in iron/heme transport may be regulated by the environmental level of this compound. The existence of several different levels of regulation by hemin in this organism is consistent with reports describing the production of both hemin-repressible and hemin-inducible proteins [65, 68, 135, 154, 157159, 222, 223]. Smalley et al. [157] suggested that binding of hemin may occur through both high- and low-affinity binding sites. P. gingivalis hemin-repressible proteins may represent low-affinity binding receptors and function as components of heme uptake that is induced only in the presence of low hemin levels, and the hemin-inducible proteins may be involved in high-affinity binding of hemin. However, P. gingivalis heme/hemoglobin-binding proteins neither have been classified into these two groups of receptors nor have been experimentally examined in this regard.

Recently, it has been reported that regulation of expression of gingipains and their subsequent involvement in heme acquisition may be controlled by mechanisms, which are independent of iron or hemin levels. Preliminary analysis of P. gingivalis grown in the presence of different iron sources indicated that the kgp and rgpA gene expression was not tightly regulated by iron, but more likely regulated by growth phase (X. Liu and C.A. Genco, unpublished data). Non- kgp mutations leading to less pigmented strains have also been identified. Chen et al. [137] found that transposon Tn 4351 was inserted into a putative glycosyl (rhamnosyl) transferase-encoding gene, an enzyme involved in the synthesis of O-antigen side chains of the lipopolysaccharide component [224]. The gene in P. gingivalis W83 is located upstream from two putative genes, porR and porS. Investigation of a novel non-pigmented P. gingivalis mutant revealed that porR was responsible for cell pigmentation [225]. porR shares similarity with genes, whose products are involved in biosynthesis of sugar portions of cell surface polysaccharides and aminoglycosides. In addition, PorR shows a homology to DegT, a pleiotropic regulatory protein from Bacillus stearothermophilus, which enhances the production of extracellular enzymes [226], as well as to aminotransferases from other bacteria, which are involved in the biosynthesis of O-antigen component of lipopolysaccharide [227]. A porR mutant showed no hemagglutination and altered distribution of Rgp, Kgp and hemagglutinins in P. gingivalis cell fractions. Interestingly, an increased level of Rgp and Kgp proteases in culture supernatants of the mutant cells has been found. These data suggest that porR, which is involved in biosynthesis of cell surface polysaccharide, may produce components for anchorage of HRgpA, Kgp, hemagglutinins and heme/hemoglobin receptors to the outer membrane.

Abaibou et al. [228, 229] have shown that a recA locus may also be involved in expression and regulation of gingipain activity in P. gingivalis. A recA mutant was non-pigmented, lacked hemolytic activity and produced significantly less proteolytic activity. It has been found that a gene vimA (virulence modulating) located downstream of the recA is responsible for P. gingivalis pigmentation [229, 230]. The mutant showed increased autoaggregation and significantly reduced proteolytic, hemolytic and hemagglutinating activities. Rgp and Kgp were found in these cells mostly in the soluble forms, suggesting that the enzymes are secreted as proenzymes.

Hasegawa et al. [231] identified another gene, a sensor histidine kinase (gppX), which may be involved in the regulation of gingipain activities and black pigmentation. The gppX mutant formed non-pigmented colonies, and the mutant cells exhibited significantly reduced gingipain activities. Moreover, gingipain activities were detected mostly in the culture supernatant. This indicates that GppX could be involved in maturation and proper localization of the gingipains to the outer membrane. However, the precise mechanism of this regulation is still not known.

The above-described genes may form a class of regulatory proteins that influence the activity of gingipains through changes in their attachment to the outer membrane of P. gingivalis. This might be also a novel post-transcriptional or translational regulatory mechanism for virulence factor expression in P. gingivalis.

10 Role of iron and heme capture in the virulence of P. gingivalis

In P. gingivalis both iron and heme appear to regulate the expression of several putative virulence factors, including gingipains [232]. Kesavalu et al. [233] reported that P. gingivalis strains, including W50, grown under iron-deplete conditions for multiple passages showed significantly decreased lesion size in mice as compared to cells grown under iron-normal or iron-replete conditions. On the other hand, higher hemin level resulted in elevated cell-associated iron, which increased the capability of the pathogen to survive at times of iron deprivation. It has been speculated that under hemin-excessive conditions putative hemin-inducible components might be operative in P. gingivalis and utilized for both hemin binding and transportation of this molecule from the surface of the cell membrane to the cytosolic fraction [43, 157, 234]. P. gingivalis growth under conditions of heme excess has been reported to enhance the virulence of the bacterium in a murine model of infection [235]. In contrast, others have suggested that iron or heme limitation resulted in increased virulence and increased expression of outer membrane proteins [68, 166]. Under hemin limitation, P. gingivalis showed elevated formation of extracellular vesicles and increased levels of hemolytic and protease activities, as well as increased hemagglutinating activity [67, 86, 157, 167, 222, 232, 235]. These discrepancies may be explained by several reasons. Isolates of P. gingivalis may have various abilities to induce infections in an animal model. It has been shown that pathogenic strains of P. gingivalis could multiply under iron-limited conditions, while non-pathogenic strains could not [67]. Bacteria possessing a low requirement for iron are likely to have a higher potential for initiating periodontal infections [67]. It has been reported that under iron-limiting conditions, the pathogenic strains had a much lower requirement for human iron-loaded transferrin or hemin than the non-pathogenic strains [67]. In addition, differences in binding capacity for iron(II) and iron(III) protoporphyrin IX between strains may exist, especially when different experimental conditions are used [236]. Smalley et al. [61] found that the hemin-binding capacity of P. gingivalis was stronger under reducing conditions. Also protoporphyrin availability in vivo may modulate membrane protein expression and in turn affect host immune responses against P. gingivalis [237]. It has been reported that PPIX limitation induced the expression of new proteins of 42, 34, 30, 29, and 18 kDa and suppressed the production of proteins of 47, 27, 17, and 15 kDa; however, PPIX availability did not affect proteolytic enzyme production or virulence in a mouse abscess model [237]. All these data suggest that hemin and iron availability regulates P. gingivalis virulence and influences growth, survival, gingipain levels and iron storage.

Porphyromonas gingivalis cells carrying a surface layer of μ-oxo dimer of heme are less susceptible to peroxidation by hydrogen peroxide [236]. Both, monomeric Fe(III)PPIX and μ-oxo dimer of heme can catalytically degrade H2O2. It has been demonstrated that in medium supplemented with 50 μM μ-oxo dimer of heme, P. gingivalis growth was observed at H2O2 concentrations up to 5 mM. Cells carrying a surface heme layer, as a result of growth in the presence of μ-oxo dimer of heme, survived hydrogen peroxide up to concentrations of 0.6 mM. Other investigators [238] showed that the growth of P. gingivalis was only moderately affected when hydrogen peroxide was added at concentrations up to 30 mM in a complex culture medium, whereas when defined basal medium was used, hydrogen peroxide at a concentration 3 mM completely inhibited the growth. It has been suggested that bacteria grown under poor nutritional conditions are much more sensitive to the deleterious effect of an oxidizing agent. These results clearly show that μ-oxo dimer of heme in both soluble form and as a cell surface layer of P. gingivalis cells inactivates hydrogen peroxide and may aid in the survival of the organism during neutrophil attack. The fact that μ-oxo dimer of heme can protect P. gingivalis against hydrogen peroxide and may function as a protective barrier against assault by reactive oxidants generated by neutrophils [60, 236] makes it an important virulence factor for this organism. It is noteworthy that P. gingivalis cells grown in vitro under conditions of heme excess accumulate more iron PPIX than those grown under heme limitation, the former being more resistant to phagocytic killing [60, 239]. Thus, cells carrying the surface layer of μ-oxo dimer of heme growing in vivo in the presence of blood may demonstrate higher ability to survive challenge with reactive oxidants.

The major component of the black pigment, μ-oxo dimer of heme [240], is generated by P. gingivalis through the reaction of hemoglobin-derived iron(II) or iron (III) protoporphyrin IX monomers and oxygen from both oxy- and deoxyhemoglobin and deposited on the cell surface [241]. It has been proposed that the liberation of the heme molecule from these species and subsequent formation of μ-oxo dimer of heme may serve as a mechanism through which hemoglobin-derived heme captures oxygen and its toxic derivatives to generate an anaerobic microenvironment [60, 241, 242]. This would serve as an oxidative buffer and permit P. gingivalis to maintain a local anaerobic environment.

11 Concluding remarks

Recent years have brought considerable advances in understanding of the mechanism of iron and heme acquisition in P. gingivalis. A great deal of research effort is now concentrated on two major aspects of iron/heme utilization: transport mechanisms and the regulation of gene expression in response to iron and heme. The mechanism of iron and heme capture employed by P. gingivalis may represent a novel mechanism for iron/heme utilization by bacterial pathogens in general. An iron/heme utilization model in P. gingivalis has recently been suggested [77]. Initially, P. gingivalis cells adhere to red blood cells using hemagglutinins. Proteases next digest surface proteins of red blood cells, which results in the release of hemoglobin. The released hemoglobin is captured by receptors on the cell surface and subjected to digestion by proteases such as gingipains to yield heme, which is transported into the cell through heme/hemoglobin receptors, such as HmuR. Surplus heme is stored on the cell surface as μ-oxo dimers of heme, resulting in black pigmentation.

To date, genetic and biochemical data demonstrated that HmuR and Kgp may function as major proteins involved in hemin/hemoglobin utilization in P. gingivalis. It is likely that Kgp is mostly engaged in hemoglobin and/or hemin binding and probably delivering heme to HmuR, and HmuR is responsible for heme transport into the cell. The ability of P. gingivalis to bind hemoglobin and/or hemin is also attributed to proteins other than HmuR and Kgp (Table 1). The probability that multiple proteins comprise the heme/hemoglobin-binding machinery of P. gingivalis correlates well with the involvement of several hemoglobin-binding proteins reported for other Gram-negative microorganisms [205, 243, 244], and we cannot rule out the engagement in iron/heme utilization of additional P. gingivalis proteins that have recently been identified [e.g. 20, 160,161,163,183]. It is also noteworthy that the mechanism of iron and heme utilization may differ in various P. gingivalis strains. Based on numerous studies and on the P. gingivalis W83 genome sequence one may suggest that some proteins may be absent or strain-to-strain differences in protein sequences or structures may exist.

The complete genome sequence of P. gingivalis W83 strain will significantly enhance our knowledge of the metabolism of this pathogen. It will also facilitate an increased understanding of the virulence of P. gingivalis and will enable the development of improved diagnostics and therapeutics. Several groups are investigating the potential of different P. gingivalis cell surface components (e.g. fimbriae, gingipains, hemagglutinins, capsular polysaccharides) to elicit protective antibody responses in animal models. To date, vaccines based on the gingipains are at early stages of development [245, 246]. Based on experimental data and sequence homology of putative proteins discovered in the P. gingivalis W83 genome, designing of antimicrobials, which abolish or interfere with iron/heme systems or use outer membrane receptors as drug delivery systems, might be one of the most important future perspectives.


This work was supported by Grant No. 3 P05A 113 24 from the State Committee for Scientific Research (KBN), Poland (T. Olczak) and Public Health Service Grant No. DE 09161 from the National Institute of Dental and Craniofacial Research, USA (C.A. Genco). We thank Dr. M. Olczak, A. Sroka, and Dr. W. Watorek for helpful scientific discussions and critical reading of the manuscript.


  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. [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. [127].
  128. [128].
  129. [129].
  130. [130].
  131. [131].
  132. [132].
  133. [133].
  134. [134].
  135. [135].
  136. [136].
  137. [137].
  138. [138].
  139. [139].
  140. [140].
  141. [141].
  142. [142].
  143. [143].
  144. [144].
  145. [145].
  146. [146].
  147. [147].
  148. [148].
  149. [149].
  150. [150].
  151. [151].
  152. [152].
  153. [153].
  154. [154].
  155. [155].
  156. [156].
  157. [157].
  158. [158].
  159. [159].
  160. [160].
  161. [161].
  162. [162].
  163. [163].
  164. [164].
  165. [165].
  166. [166].
  167. [167].
  168. [168].
  169. [169].
  170. [170].
  171. [171].
  172. [172].
  173. [173].
  174. [174].
  175. [175].
  176. [176].
  177. [177].
  178. [178].
  179. [179].
  180. [180].
  181. [181].
  182. [182].
  183. [183].
  184. [184].
  185. [185].
  186. [186].
  187. [187].
  188. [188].
  189. [189].
  190. [190].
  191. [191].
  192. [192].
  193. [193].
  194. [194].
  195. [195].
  196. [196].
  197. [197].
  198. [198].
  199. [199].
  200. [200].
  201. [201].
  202. [202].
  203. [203].
  204. [204].
  205. [205].
  206. [206].
  207. [207].
  208. [208].
  209. [209].
  210. [210].
  211. [211].
  212. [212].
  213. [213].
  214. [214].
  215. [215].
  216. [216].
  217. [217].
  218. [218].
  219. [219].
  220. [220].
  221. [221].
  222. [222].
  223. [223].
  224. [224].
  225. [225].
  226. [226].
  227. [227].
  228. [228].
  229. [229].
  230. [230].
  231. [231].
  232. [232].
  233. [233].
  234. [234].
  235. [235].
  236. [236].
  237. [237].
  238. [238].
  239. [239].
  240. [240].
  241. [241].
  242. [242].
  243. [243].
  244. [244].
  245. [245].
  246. [246].
  247. [247].
  248. [248].
  249. [249].
  250. [250].
  251. [251].
  252. [252].
View Abstract