Avian Malaria and Related Parasites from Resident and Migratory Birds in the Brazilian Atlantic Forest, with Description of a New Haemoproteus Species

Determining the prevalence and local transmission dynamics of parasitic organisms are necessary to understand the ability of parasites to persist in host populations and disperse across regions, yet local transmission dynamics, diversity, and distribution of haemosporidian parasites remain poorly understood. We studied the prevalence, diversity, and distributions of avian haemosporidian parasites of the genera Plasmodium, Haemoproteus, and Leucocytozoon among resident and migratory birds in Serra do Mar, Brazil. Using 399 blood samples from 66 Atlantic Forest bird species, we determined the prevalence and molecular diversity of these pathogens across avian host species and described a new species of Haemoproteus. Our molecular and morphological study also revealed that migratory species were infected more than residents. However, vector infective stages (gametocytes) of Leucocytozoon spp., the most prevalent parasites found in the most abundant migrating host species in Serra do Mar (Elaenia albiceps), were not seen in blood films of local birds suggesting that this long-distance Austral migrant can disperse Leucocytozoon parasite lineages from Patagonia to the Atlantic Forest, but lineage sharing among resident species and local transmission cannot occur in this part of Brazil. Our study demonstrates that migratory species may harbor a higher diversity and prevalence of parasites than resident species, but transportation of some parasites by migratory hosts may not always affect local transmission.


Introduction
Brazil is one of the world's richest countries in terms of bird species diversity, with 1919 known species [1], of which 234 species are endangered [2]. In Brazil, the Atlantic Forest biome has the second largest number of species (934) and second highest level

Microscopic examination
We collected 195 thin blood smears in 2019 from 135 bird samples. Haemosporidian parasites were only identified by microscopy in 11 birds: two were infected with Plasmodium nucleophilum (identified by the nucleophilic pattern of blood stages and the presence of all the other main characteristics of this species (see [15]), three with Plasmodium sp., four with Haemoproteus (Parahaemoproteus) sp., one with mixed infection (Plasmodium sp. + Haemoproteus sp.), and one with Haemoproteus (Parahaemoproteus) nucleocentralis n. sp. which is described below.
Among the Plasmodium-infected individuals for which we analysed blood smears, three were from resident host species (Tachyphonus coronatus, Turdus rufiventris, and Zonotrichia capensis) and three from one migratory species (Elaenia albiceps). We found Plasmodium gametocytes in the blood smears of one Curucutu resident bird (Zonotrichia capensis) and two migratory birds (Elaenia albiceps). In contrast, among the Haemoproteus-infected individuals for which we analysed blood smears, one was a resident species (Tangara desmaresti) and four were migratory species (Elaenia albiceps). The Haemoproteus parasite found in the E. albiceps is a new lineage (see discussion below) and likely a new species, but, due to the poor quality of staining, it was not possible to describe this species. Even with poor staining, it was possible to notice some distinctive morphological features of this parasite, such as the presence of a readily visible vacuole in the macrogametocytes (Figure 2a

Microscopic Examination
We collected 195 thin blood smears in 2019 from 135 bird samples. Haemosporidian parasites were only identified by microscopy in 11 birds: two were infected with Plasmodium nucleophilum (identified by the nucleophilic pattern of blood stages and the presence of all the other main characteristics of this species (see [15]), three with Plasmodium sp., four with Haemoproteus (Parahaemoproteus) sp., one with mixed infection (Plasmodium sp. + Haemoproteus sp.), and one with Haemoproteus (Parahaemoproteus) nucleocentralis n. sp. which is described below.
Among the Plasmodium-infected individuals for which we analysed blood smears, three were from resident host species (Tachyphonus coronatus, Turdus rufiventris, and Zonotrichia capensis) and three from one migratory species (Elaenia albiceps). We found Plasmodium gametocytes in the blood smears of one Curucutu resident bird (Zonotrichia capensis) and two migratory birds (Elaenia albiceps). In contrast, among the Haemoproteus-infected individuals for which we analysed blood smears, one was a resident species (Tangara desmaresti) and four were migratory species (Elaenia albiceps). The Haemoproteus parasite found in the E. albiceps is a new lineage (see discussion below) and likely a new species, but, due to the poor quality of staining, it was not possible to describe this species. Even with poor staining, it was possible to notice some distinctive morphological features of this parasite, such as the presence of a readily visible vacuole in the macrogametocytes (Figure 2a this parasite, such as the presence of a readily visible vacuole in the macrogametocytes (Figure 2a, b), sub-terminal position of nuclei in macrogametocytes (Figure 2a, b), close adherence of advanced gametocytes, both to the nuclei and envelope of erythrocytes (Figure 2a-d), and slight displacement of nuclei in infected erythrocytes (Figure 2a-d). For further parasite description, additional material is needed.  Vector: Probably Culicoides biting midges; species is unknown (see Discussion). Etymology: The species name refers to a distinctive character of this species, which is the predominantly central position of nuclei in fully-grown macrogametocytes (see Description below).

Description
Young gametocytes (Figure 3a,b) were rare in the type material; they develop in mature erythrocytes. Earliest gametocytes are broadly oval in form; they adhere to the nuclei of erythrocytes and were seen in a subpolar position in infected erythrocytes ( Figure 3a); pigment granules were readily visible. Advanced growing gametocytes closely adhere both to the erythrocyte nuclei and envelope; they extend longitudinally along the nuclei and slightly displace them laterally (Figure 3b). Gametocyte outline is even. Pigment granules are prominent; some of them reach size of pigment granules present in mature gametocytes (compare Figure 3b  (l-p) Microgametocytes. Note that nuclei were located in central or close to central position in all mature macrogametocytes (e-k), but they were occasionally seen in sub-terminal position in the growing parasites (d). Long arrows: Gametocyte nuclei; arrowheads: Pigment granules; triangle arrowhead: Nuclei of infected erythrocytes. Giemsa-stained thin blood films. Scale bar: 10 μm.
Macrogametocytes (Figure 3c-k) develop in mature erythrocytes. The cytoplasm is homogeneous or slightly granular in appearance. Volutin granules and vacuoles were not seen. Outline is predominantly even (Figure 3e-g), occasionally slightly wavy at gametocyte ends (Figure 3d, e). Gametocytes grow along nuclei of infected erythrocytes, enclose nuclei with their ends, but do not encircle them completely (Figure 3c-k). Advanced and fully-grown macrogametocytes were closely appressed both to the envelop and nuclei of host cells; they displace the nuclei laterally (Table 1) and slightly deform them, resulting in acceptance of a slightly roundish shape in comparison to nuclei in non-infected erythrocytes (Figure 3c, e, k). Fully-grown gametocytes fill erythrocytes up to their poles (Figure 3e-k). Gametocytes nucleus is relatively small (Table 1), variable in form, predominantly assumes a central or close to the central position (Figure 3d-g), a characteristic feature of this species development; occasionally the nuclei were observed in slightly sub- Macrogametocytes (Figure 3c-k) develop in mature erythrocytes. The cytoplasm is homogeneous or slightly granular in appearance. Volutin granules and vacuoles were not seen. Outline is predominantly even (Figure 3e-g), occasionally slightly wavy at gametocyte ends (Figure 3d,e). Gametocytes grow along nuclei of infected erythrocytes, enclose nuclei with their ends, but do not encircle them completely (Figure 3c-k). Advanced and fully-grown macrogametocytes were closely appressed both to the envelop and nuclei of host cells; they displace the nuclei laterally (Table 1) and slightly deform them, resulting in acceptance of a slightly roundish shape in comparison to nuclei in non-infected erythrocytes (Figure 3c,e,k). Fully-grown gametocytes fill erythrocytes up to their poles (Figure 3e-k). Gametocytes nucleus is relatively small (Table 1), variable in form, predominantly as-sumes a central or close to the central position (Figure 3d-g), a characteristic feature of this species development; occasionally the nuclei were observed in slightly sub-central position (Figure 3d,i,k), nucleolus was not seen. Pigment granules can be scattered throughout the cytoplasm (Figure 3g), but also often seen in groups (Figure 3c-f). Pigment granules are variable in form and shape, they usually are oval or slightly elongate; predominate medium-size granules (0.5-1.0 µm), but a few of large-size (1.0-1.5 µm) granules were often seen. Fully-grown gametocytes displace nuclei of host-cells laterally, but do not influence shape of the cells in comparison to uninfected erythrocytes (Figure 3e-h, Table 1).
a Number of measurements (n) was 21, except indicated otherwise. All measurements are given in micrometres, except for pigment granules. Minimum and maximum values are provided, followed in parentheses by the arithmetic mean and standard deviation. b Microgametocyte nuclei were hardly defined and difficult to measure. c Nucleus displacement ratio (NDR) according to [16].
Microgametocytes (Figure 3l-p, Table 1). General configuration and other characters are as in macrogametocytes with the usual haemosporidian sexual dimorphic characters, which are the large diffuse nuclei and relatively pale staining of the cytoplasm.

Taxonomic Remarks
Haemoproteus nucleocentralis n. sp. is the first haemosporidian parasite reported in Tangara desmaresti. The most similar partial cytb sequences (GenBank accessions MN459077 and KJ466075) are of 99% similarity (or 2 bp); they were reported in the Orange-bellied Euphonia Euphonia xanthogaster (Passeriformes, Thraupidae) in the Chilean Andes. However, there is no morphological characterization of these parasites and no publication associated with these DNA sequences. In the MalAvi database, the closest lineage to the new species is hOCHLEU01, with 98% similarity (or 5 bp); it was reported in several species of Thraupidae in South American Andean birds.
Haemoproteus coatneyi was often reported in birds belonging to the Tangara genus and other Thraupidae birds [17], thus, should be distinguished from the new species. A characteristic feature of H. nucleocentralis n. sp. is the predominantly central position of nuclei in fully-grown gametocytes. This character is relatively rare in haemoproteids parasitizing passerine birds and has not been reported in haemoproteids parasitizing New World passerines as of yet, so is worthy of attention during H. nucleocentralis identification. Based on this character, gametocytes of H. nucleocentralis can be readily distinguished from Haemoproteus coatneyi [18], the common parasite of New World passerines. In H. coatneyi, nuclei are strictly sub-terminal in macrogametocytes [12]. The same is seen in, Haemoproteus paruli [18] and Haemoproteus thraupi [18], which have gametocytes morphologically indistinguishable from H. coatneyi [12]. Another distinctive feature of H. nucleocentralis n. sp. is the deformed infected host-cell nuclei, which assume a roundish form (compare Figure 3a,b with Figure 3e,k), but this feature remains insufficiently investigated in the other above-mentioned Haemoproteus parasites.
Haemoproteus nucleocentralis n. sp. can be readily distinguished from Haemoproteus erythrogravidus, a common parasite of New World passerines, due to the absence of protrusion in the envelope of the infected erythrocyte (so-called the gravid morphology of infected host cells) [19]. Additionally, nuclei located strictly in sub-terminal position in macrogametocytes of H. erythrogravidus, and this is not the case in the new species.
Gametocytes of H. nucleocentralis n. sp. share some similar features with Haemoproteus witti, a common parasite of hummingbirds (Apodiformes), in the New World. In both species, nuclei are predominantly of central position in macrogametocytes [20]. Interestingly, several lineages of H. witti were reported in passerines, but gametocytes were not, indicating possible incomplete (abortive) development of this hummingbird parasite in passerines. In H. witti gametocytes, the average number of pigment granules is close to 25, which is significantly less than (about 10) in H. nucleocentralis (Table 1).
Haemoproteus nucleocentralis n. sp. can be readily distinguished from Haemoproteus vireonis, a common parasite of South American passerines [12,21]. In the latter parasites, the growing gametocytes often assume the dumbbell-shape and nuclei locate strictly sub-terminally in macrogametocytes. Both these features are not characteristics of H. nucleocentralis n. sp.
Phylogenetic inference showed that cytb sequences of H. nucleocentralis n. sp. clustered with Haemoproteus paruli (see 2.3. Molecular and Phylogenetic Analysis). These two lineages are only of 96% similarity (21 bp difference) in partial cytb sequences. Gametocytes of the latter parasite are indistinguishable morphologically from those of H. coatneyi [12]. Cytb sequences of H. coatneyi and H. paruli are available, and are located in different branches in the phylogenetic tree. However, it should be noted that morphological data were not provided when linking these sequences with corresponding morphospecies [22], and the accuracy of the molecular characterization of both H. coatneyi and H. paruli needs further support.

Molecular and Phylogenetic Analysis
All samples were analysed by nested PCR and subsequent sequencing to detect the presence of Plasmodium, Haemoproteus, and Leucocytozoon. Among these, 68 were positive: 48.5% (33) for Plasmodium, 23.5% (16) for Haemoproteus, and 26.5% (18) for Leucocytozoon, in addition to a mixed infection with Plasmodium and Haemoproteus (1.5%).
The lineage pBASCUL01 was found in Basileuterus culicivorus and hTANDES01 was reported here in Tangara desmaresti. This last one represents the first description of haemosporidian parasite in this species of passerines.
Based on the Plasmodium lineage phylogenetic analysis (Figure 4), it is likely that most of the lineages found in Núcleo Curucutu belong to the subgenera Haemamoeba and Novyella. According to the MalAvi Database, the pDENPET03 (P. nucleophilum) lineage has already been described in 68 hosts belonging to 59 genera, 23 families, and eight orders in North and South America. In this study, the lineage was found in eight individuals of five species, belonging to five families, all of which are passerines. Although not statistically significant as mentioned above, it is also important to note a trend of higher occurrence of Plasmodium sp. in migratory species than in resident birds (18.6% versus 15.4%) (Supplementary Materials, Raw Data Spreadsheet S2).
All the 20 positive samples for Leucocytozoon were detected by PCR in Elaenia albiceps. Of these, 18 unique DNA sequences were obtained, since two contained overlapping Leucocytozoon sequences. Two of these 18 included mixed infections: one with Plasmodium sp. and another with P. nucleophilum. Five different lineages were obtained (lDIUDIU11, lELAALB02, lELAALB05, lZOLPYR01, and lTROAED02) with previous records in Argentina, Chile, Colombia, and Peru, also in Elaenia albiceps or Elaenia frantzii ( Figure 5, Table 2 and Supplementary Table S1).
The  Figure 6). The hELAALB01 lineage shows 99% identity with the lineages hPHSIB2 (H. homopalloris), hCOLL2 (H. pallidus), and hSYAT03 (H. pallidulus), and exist within a larger clade within the phylogenetic tree ( Figure 6). All these species are known to have pale staining gametocytes. Moreover, the new hTANDES01 lineage described here as a new species (H. nucleocentralis) appeared in the clade of haemoproteids belonging to subgenus Parahaemoproteus, indicating that species of Culicoides (Ceratopogonidae) are involved in its transmission.  All the 20 positive samples for Leucocytozoon were detected by PCR in Elaenia albiceps. Of these, 18 unique DNA sequences were obtained, since two contained overlapping Leucocytozoon sequences. Two of these 18 included mixed infections: one with Plasmodium sp. and another with P. nucleophilum. Five different lineages were obtained (lDIUDIU11, The lineages reported during the present study were identified by the symbols "triangle" and "square", for migratory and resident species, respectively. Lineages given in bold refer to new lineages. ogens 2020, 9, x FOR PEER REVIEW 14 of 22 lELAALB02, lELAALB05, lZOLPYR01, and lTROAED02) with previous records in Argentina, Chile, Colombia, and Peru, also in Elaenia albiceps or Elaenia frantzii ( Figure 5, Table 2  and Supplementary Table S1).  (Figure 6). The hELAALB01 lineage shows 99% identity with the lineages hPHSIB2 (H. homopalloris), hCOLL2 (H. pallidus), and hSYAT03 (H. pallidulus), and exist within a larger clade within the phylogenetic tree ( Figure 6). All these species are known to have pale staining gametocytes. Moreover, the new hTANDES01 lineage described here as a new species (H. nucleocentralis) appeared in the clade of haemoproteids belonging to subgenus Parahaemoproteus, indicating that species of Culicoides (Ceratopogonidae) are involved in its transmission.

Discussion
We analysed blood samples from birds from Núcleo Curucutu belonging to seven orders, 21 families, and 66 species, the majority belonging to the family Tyrannidae (53.4%). A great diversity of haemosporidian parasites was found, including Plasmodium and Haemo-proteus lineages widely described from passerines in the Neotropics and Europe (MalAvi database) and the first molecular detection of Leucocytozoon lineages in São Paulo State. The haemosporidian lineages pDENPET03, hELAALB01, and lELAALB05 were most frequently found, with the first two considered as generalist lineages, as they were previously reported in bird species belonging to different families and, sometimes, even orders [24,25], however, lELAALB05 has been described only in Argentina [26]. In fact, many lineages from this study had been previously detected only in Argentina (pCURCUR01, lDIUDIU11, lELAALB02, hELAALB05, and hELAALB01). Concerning, Haemoproteus sp. hELAALB01, it is important to mention that this lineage from Fecchio et al. [27] (GenBank #MK695429 and #MK695430, hELAALB01) is different (96% of identity) from [28] (GenBank #MK981643, #MK264397, #JX029900, hELALB01). The hELALB01 lineage is currently named hMY-ISWA01 according to MalAvi database. Not all infections detected by PCR were confirmed by microscopy. We detected one Turdus rufiventris with pTUMIG03 lineage, which has been associated with P. unalis [29]. However, although positive for Plasmodium sp., characteristics of this species were not observed during microscopic examination of blood smears. Additionally, a smear that was positive for Haemoproteus sp. was PCR positive only for Plasmodium (pSPMAG06 #HM031936, [30], currently identified as Plasmodium lutzi [31]. It is widely known that in samples with co-infections, probably like the latter, general PCR protocols tend to favor the amplification of the parasite with the higher parasitemia or the amplification of the lineage that best matches primer sequences, and mixed infections frequently are overlooked in PCR-based studies [32,33]. Here, the presence of gametocytes of P. nucleophilum pDENPET03 was documented in Elaenia albiceps (Tyrannidae) and Zonotrichia capensis (Emberizidae). These findings imply that this parasite has the capacity to complete its life cycle and produce infective gametocytes in the mentioned host species. Elaenia albiceps is an endemic bird in the Neotropical region, migrating north between February and March, transiting the Atlantic coast to the Amazon, spending the winter in northeast and northern of Brazil [5,34]. The individuals sampled here were found in the Serra do Mar mountainous forest region in March of each year, during the migration of this species between its breeding areas in southern South America and its winter areas in northeastern Brazil [35,36]. During migration, the birds stay for a few days in the Curucutu region, where they feed on fruits high in the mountainous forest (F. Schunck pers. obs.).
Leucocytozoon lineages from E. albiceps showed total similarity with lineages previously described in Argentina, indicating a possible flow of these parasites between this country and Brazil. We do not have evidence of this species being competent hosts (presence of gametocytes in blood smears), but we know that E. albiceps from Argentina first fly to the Atlantic coast of Brazil, then to the Cerrado region of central Brazil and no birds overwinter to the west of the Andes mountain range (i.e., Peru, Ecuador and Colombia) [34]. We also found lTROAED02 in E. albiceps, a lineage initially identified in Peru in Troglodytes aedon [37], but quite generalist in Colombia, where it can infect 25 different species [38].
Fecchio et al. [39] linked the scarcity of Leucocytozoon infections with warmer temperatures in the tropical lowlands rather than a lack of transmission opportunities, as the vectors for this genus (black flies in the family Simuliidae) are abundant and diverse in lowland regions. However, the warm temperature conditions in the tropical lowlands hardly are limited factors for this infection transmission because Leucocytozoon parasites are prevalent in similarly warm tropical rainforests in Africa and Asia [40]. The dissimilarity of vector species could explain the lack of transmission in the study area. We hypothesize that Leucocytozoon spp. infect E. albiceps in Patagonia (where is cold and there is greater abundance of black flies) and the Leucocytozoon transmission ends when the birds enter Brazil. In fact, Leucocytozoon sp. presents an inverse latitudinal gradient in the probability of infection and phylogenetic diversity in New World birds, with higher prevalence and lineage diversity toward the poles [26]. The greater probability of a bird becoming infected with Leucocytozoon sp. in regions with colder summers and towards the poles, such as the Patagonia region of Argentina, which has a higher prevalence of these parasites than Brazil, indicates that it is possible that E. albiceps may be transporting these pathogens during their migratory trips. However, our failure to detect these parasites in blood smears may indicate that Leucocytozoon sp. did not evolve to complete its life cycle and produce gametocytes in this host species due to possible abortive development in non-adapted hosts [41]. In this case, only tissue stages develop, and their merozoites or remnants of tissue stages (syncytia) appear in circulation providing templates for PCR amplification, but parasite cannot inhabit red blood cells and thus are difficult to detect by microscopic examination of blood films [24,42]. The birds of genus Elaenia would then be a dead end host for the parasite, as it would be unable to infect vector species. It is possible, that parasites may persist in tissue stages during migration and gametocytes are absent, but a relapse may occur at breeding sites and gametocytes would re-appear. In fact, Leucocytozoon sp. has not yet been reported in blood smears by microscopic examination in Brazilian birds, potentially missing infections in migratory species where gametocytes were absent. Although, possible not the case for Leucocytozoon parasites, our work suggests that E. albiceps does indeed have the potential to disperse haemosporidians over long distances, but that such dispersion may be taxonomically restricted to certain genera and lineages, probably generalist lineages such as pDENPET03.

Sampling
This study was performed in the Núcleo Curucutu, Parque Estadual Serra do Mar

Microscopic Examination
Thin blood smears were fixed with 100% methanol on the same day of collection and stained with a 10% Giemsa solution, within 30 days after collections, for 1 h [12]. Blood smears were then examined microscopically for 20-25 min by viewing 100 fields at low magnification (400×) and 100 fields at high magnification (1000×) [12], using a Leica ® DM3000LED light microscope. Morphological identification of parasite species was per-

Microscopic Examination
Thin blood smears were fixed with 100% methanol on the same day of collection and stained with a 10% Giemsa solution, within 30 days after collections, for 1 h [12]. Blood smears were then examined microscopically for 20-25 min by viewing 100 fields at low magnification (400×) and 100 fields at high magnification (1000×) [12], using a Leica ® DM3000LED light microscope. Morphological identification of parasite species was performed according to Valkiūnas [12] and Valkiūnas and Iezhova [43].

Molecular Detection and Genotyping of Haemosporidian Infections
DNA from blood samples was extracted with the Wizard ® SV 96 Genomic DNA Purification System (Promega, Madison, WI, USA) with modifications. Briefly, FTA cards with 10 µL of blood were incubated with Whole Blood Lysis Buffer (400 µL) for 15 min in a shaker at 90 • C. The initial lysis was completed with Proteinase K and incubated overnight in a shaker at 37 • C. The lysates were transferred to columns and washed according to the manufacturer's instructions. DNA was eluted in 50 µL of Nuclease-FreeWater and stored at −20 • C.
Polymerase chain reactions (PCR) were conducted using a nested protocol targeting the mitochondrial cytochrome b (cytb) gene of Plasmodium, Haemoproteus, and Leucocytozoon species [44]. The first reaction used the primers HaemNFI/HaemNR3 and 50 ng of genomic DNA. In the nested reaction, performed with a second pair of primers (HaemF/HaemR2 for Plasmodium and Haemoproteus or HaemFL/HaemR3L for Leucocytozoon), 1 µL of the product from the first reaction was used as a template. In each PCR, positive controls were carried out in parallel, containing Plasmodium, Haemoproteus, and Leucocytozoon DNA, and ultrapure water served as a negative control. PCR products were sequenced by BigDye ® Terminator v3.1 Cycle Sequencing Kit in ABI PRISM ® 3500 Genetic Analyzer (Applied Biosystems, Carlsbad, CA, USA), using nested PCR primers. The cytb sequences (~480 bp) were obtained and aligned with sequences from the MalAvi database (http://130.235.244.92/Malavi/), in order to verify parasite lineage identity and identify new lineages. The sequences possessing at least one different nucleotide were considered unique lineages and were named according to the MalAvi nomenclature [45] and deposited in GenBank and MalAvi database.
A comparison between the prevalence of haemosporidian infections (detected by PCR) between resident birds and migratory birds was analysed using a chi-square test with Yates' correction for smaller samples as warranted. Findings were considered statistically significant if p < 0.05.

Phylogenetic Analysis
The phylogenetic relationship among reported parasites was inferred using partial cytb gene sequences. GenBank accessions of the used sequences are given in the phylogenetic trees. The phylogenetic reconstruction was performed separately for Plasmodium, Haemoproteus, and Leucocytozoon parasites using the Bayesian inference method implemented in MrBayes v3.2.0 [46]. Bayesian inference was executed with two Markov Chain Monte Carlo searches of 3 million generations, with each sampling 1 of 300 trees. After a burn-in of 25%, the remaining 15,002 trees were used to calculate the 50% majority-rule consensus tree. The phylogeny was visualized using FigTree version 1.4.0 [47].