Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2012 May 1.
Published in final edited form as: Int J Parasitol. 2011 Feb 12;41(6):645–655. doi: 10.1016/j.ijpara.2011.01.005

Genetic analyses of atypical Toxoplasma gondii strains reveal a fourth clonal lineage in North America

Asis Khan a, JP Dubey b, Chunlei Su c, James W Ajioka d, Benjamin M Rosenthal *, L David Sibley a,*
PMCID: PMC3081397  NIHMSID: NIHMS275145  PMID: 21320505

Abstract

Toxoplasma gondii is a widespread parasite of animals that causes zoonotic infections in humans. Previous studies have revealed a strongly clonal population structure in North America and Europe, while strains from South America are genetically separate and more diverse. However, the composition within North America has been questioned by recent descriptions of genetically more variable strains from this region. Here, we examined an expanded set of isolates using sequenced-based phylogenetic and population analyses to re-evaluate the population structure of T. gondii in North America. Our findings reveal that isolates previously defined by atypical restriction fragment length polymorphism patterns fall into two discrete groups. In one case, these new isolates represent variants of an existing lineage, from which they differ only by minor mutational drift. However, in the second case, it is evident that these isolates define a completely new lineage that is common in North America. Support for this new lineage was based on phylogeny, principle components analysis, STRUCTURE analyses, and statistical analysis of gene flow between groups. This new group, referred to as haplogroup 12, contains divergent genotypes previously referred to as A and X, isolated from sea otters. Consistent with this, group 12 was found primarily in wild animals, as well as occasionally in humans. This new lineage also has a highly clonal population structure. Analysis of the inheritance of multilocus genotypes revealed that different strains within group 12 are the products of a single recombination event between type 2 and a unique parental lineage. Collectively, the archetypal type 2 has been associated with clonal expansion of a small number of lineages in the North, as a consequence of separate but infrequent genetic crosses with several different parental lines.

Keywords: Genotype, Phylogeny, Linkage disequilibrium, Network analysis, Clonality, Population

1. Introduction

Toxoplasma gondii is a widespread protozoan parasite in the phylum Apicomplexa, which contains more than 5000 species, only a few of which cause disease in humans (Dubey, 2010). Toxoplasma gondii commonly infects a wide range of warm-blooded animals and causes zoonotic infection in humans. Infections in humans are normally sub-clinical; however, they can lead to severe disease in immunocompromised and congenitally infected individuals (Dubey, 2010). Additionally, toxoplasmosis has recently been recognized as a cause of severe recurrent ocular disease, especially in some regions of southern Brazil (Silveira et al., 2001; Jones et al., 2006). Toxoplasma gondii propagates by both asexual replication, which occurs in a variety of hosts and sexually, which occurs only in feline intestinal epithelial cells (Dubey and Frenkel, 1972). Despite the presence of a sexual cycle, T. gondii maintains a highly clonal population structure. To date, the majority of isolates from North America and Europe belong to one of only three, closely related clonal lineages (referred to as types I, II, III or 1, 2 and 3 herein) (Howe and Sibley, 1995; Ajzenberg et al., 2002). Clonal propagation is likely favored by the ability of T. gondii to be transmitted between intermediate hosts via ingestion of tissue cysts, a trait that distinguishes it from related parasites (Su et al., 2003a). The major lineages in North America are thought to have resulted from a few natural genetic crosses between highly similar parental types, the progeny of which expanded to give rise to the clonal population structure during the past ~ 10,000 years (Su et al., 2003a; Boyle et al., 2006).

Although clonality in T. gondii is strongly evident in North America and Europe, much greater genetic diversity, likely reflecting more frequent recombination, is evident in South America (Khan et al., 2006, 2007; Lehmann et al., 2006; Pena et al., 2008). Initial studies supporting this conclusion were conducted using microsatellite (Lehmann et al., 2006) or restriction fragment length polymorphism (RFLP) markers (Pena et al., 2008). Although such markers provide a valuable tool for surveys, they do not capture all of the genetic diversity at a given locus. Consequently, the use of RFLP markers suggested that Southern strains might be recombinant versions of those in the North. Sequencing of selectively neutral introns revealed that Southern strains were not simply mixtures of those found in the North (Khan et al., 2007). Instead, these studies established that strong geographic separation exists between strains from North and South America with each region being typified by distinct, predominant strain types (Khan et al., 2007). Sequence-based analyses also identified 11 distinct haplotypes and population modeling suggested that they were derived from four ancestral founders through a small number of recombination events (Khan et al., 2007). Among the extant lineages are the three archetypal clonal lineages from North America and Europe, as well as several lineages that are predominant in South America (Khan et al., 2007). The biallelism previously noted in the North is also mirrored by distinct allelic types in the South, consistent with genetic drift occurring within each of these distinct populations following their separation several million years ago (Sibley and Ajioka, 2008).

One difference among isolates that have been studied from these two regions is that most strains sampled in North America and Europe have come from domestic animals or humans in urban or suburban environments, while those from South American include samples from more remote regions. Recent sampling of isolates from both wild and domestic animals in more geographically diverse regions of North America has revealed greater genetic diversity than encountered previously. For example, isolates that have been associated with disease in sea otters from western United States and Canada show additional genetic diversity, and these strains have been given the designations of X and A, based on RFLP variants at particular loci (Miller et al., 2004; Sundar et al., 2008). Isolates bearing the type X pattern were also found in terrestrial carnivores from coastal California, suggesting a link with infections in sea otters (Miller et al., 2008). Additionally, atypical RFLP patterns have been detected in strains that have been isolated from wild animals in Canada (Dubey et al., 2008b) and from domestic animals including sheep (Dubey et al., 2008c) and pigs (Dubey et al., 2008a) in the United States. Such findings are not entirely novel as even the original RFLP-based studies revealed examples of genotypes that differed from the archetypal clonal lineages (Sibley and Boothroyd, 1992; Howe and Sibley, 1995). However, establishing the relationship of these so-called atypical strains has been hampered by the fact that they were typed using RFLP analyses, which tend to under-represent the true genetic diversity. Hence, it is unclear whether these variants diversified via infrequent recombination with the archetypal lineages, by acquiring new mutations that subtly alter otherwise common genotypes, or through long-term coexistence as genetically separate populations with overlapping ranges.

To better understand the genetic diversity of North American strains, we sequenced several introns and coding sequences in an expanded set of isolates, including those previously designated as “atypical”. This expanded data set was analyzed using well-established criteria to evaluate the population structure. Our findings reveal that while some strains are merely minor variants of the established lineages, others comprise a fourth clonal lineage that circulates in North America.

2. Materials and methods

2.1. Toxoplasma gondii strains

Strains of T. gondii (Supplementary Table S1) were grown in monolayers of human foreskin fibroblast cells propagated in DMEM (Invitrogen, Carlsbad, CA, USA) supplemented with 10% FBS, 2 mM glutamine, 20 mM HEPES pH 7.5) and 10 µg/ml gentamicin, and harvested after natural egress (Su et al., 2003a; Khan et al., 2007). Parasites were harvested by passing through 3.0 micron polycarbonate filters to remove host cell debris and resuspended in PBS at a concentration of approximately 107 cell/ml. To prepare lysates for PCR, parasites were digested with 10 µg/ml proteinase K (Sigma, St. Louis, MO, USA) at 55°C for 2 h and heat inactivated at 95°C for 15 min (Su et al., 2002, 2003b).

2.1.1. Mouse infections

To determine how virulent these isolates were in a mouse model, groups of five 8-week-old female outbred mice (CD1) were inoculated i.p. with 10, 100 or 1000 tachyzoites grown in culture, as described previously (Taylor et al., 2006). Mice were observed for 30 days and surviving mice were bled to collect sera. Infections were confirmed by Western blotting against whole lysate of Me49 strain parasites separated by SDS PAGE and detected using horseradish peroxidase-conjugated goat anti-mouse immunoglobulin G (Amersham Pharmacia/GE Healthcare, Piscataway, NJ, USA), as described previously (Taylor et al., 2006). Cumulative mortality was defined as the number of animals that succumbed / total number of animals infected (those that died + those that were seropositive but survived) for all doses combined. Animals were cared for by the Division of Comparative Medicine, Washington University School of Medicine, USA, and all procedures were approved by the Animal Studies Committee at Washington University School of Medicine, USA.

2.2. PCR amplification and sequencing

Lysates were used as template DNA for PCR amplification of five introns (UPRT, MIC, BTUB, HP and EF) as described previously (Khan et al., 2007), combined with one surface antigen (SAG1) and two dense granule proteins (GRA6 and GRA7). Gene identification numbers are listed in Supplementary Table S2. Amplified PCR products were sequenced using BigDye cycle sequencing (Applied Biosystems, Foster City, CA, USA), performed by SeqWright DNA Technology Services (Houston, TX, USA), as described previously (Khan et al., 2007).

2.3. DNA polymorphism analysis

Clustal W/X (Higgins et al., 1996) was used to align the sequences using default settings. Aligned sequences were directly incorporated into Molecular Evolutionary Genetic Analysis (MEGA) Version 3.1 for neighbor-joining analyses (Kumar et al., 2001) using both distance and parsimony methods. One thousand bootstrap replicates were conducted and consensus trees were drawn with an arbitrary root according to the bootstrap 50% majority rule.

2.4. Linkage disequilibrium

Linkage disequilibrium (LD) was calculated from the concatenated intron sequences to estimate linkage across wide regions of the genome. D’, a scaled version of the disequilibrium coefficient D, was measured and LD plots were constructed using DnaSP 4.0 (Rozas et al., 2003). To demonstrate the significance of association among different informative loci, both two-tailed Fisher’s exact test and the Chi-squared test were performed. The average LD was detected using the ZnS statistic (Kelly, 1997), which compares LD over all pair-wise comparisons for S polymorphisms in N sequences.

To determine the extent of association between different informative alleles at different loci, the Index of Association (IA) (Smith et al., 1993) was calculated using Multilocus 1.3 (Agapow and Burt, 2001). IA values that were significantly different from zero were considered to reflect clonal populations. Analyses of fixation due to population substructure (iFST) were conducted using standard population methods (Weir, 1996). GDA software (http://hydrodictyon.eeb.uconn.edu/people/plewis/) was used to analyze NEXUS files containing the polymorphisms obtained from sequencing. Strains were grouped by region, haplotype, host-origin in order to calculate FST values between potential subpopulations.

2.5 Population structure analysis

The population structure of T. gondii was modeled using a Bayesian clustering algorithm implemented in the program STRUCTURE 2.2 (Falush et al., 2003a,b) to analyze intron sequence data as described previously (Khan et al., 2007). Five replicate simulations were conducted for each value of K (the number of founding groups) ranging from two to 10 using a burn-in of 104 and final run of 105 Markov chain Monte Carlo steps, under the admixture model with independent allele frequencies. The output patterns for K = 4–7 provided the best fit for the data and these were used for further analyses.

Principal components analysis (PCA) was performed using the program GenAlEX 6 (Peakall and Smouse, 2005) based on a pair-wise genetic distance matrix calculated as described previously (Huff, 1993). Distance was calculated as D = [1 − 2nxy/2n] where 2nxy, corresponds to the number of shared loci between two individuals and 2n the total number of loci.

3. Results

3.1. Phylogenetic analyses of T. gondii strains based on introns

To compare the genetic diversity of T. gondii, we analyzed a total of 66 strains that were isolated from a wide range of animal hosts and humans from North America, Europe and South America (Supplementary Table S1). Included among these were strains of T. gondii from North America that have been described as types X and A (Miller et al., 2004; Sundar et al., 2008), as well as more diverse isolates reported from domestic (Dubey et al., 2008a, c) and wild animals (Dubey et al., 2008b) and isolates that have shown additional polymorphism in previous studies (Howe and Sibley, 1995). For comparison, we included a set of reference strains spanning the 11 haplogroups previously characterized from North America-Europe and South America (Khan et al., 2007). Collectively, these isolates were chosen to reflect the maximum level of genetic diversity that has been detected to date in previous surveys; however, many of these new isolates have not been examined in detail or using sequence-based markers.

Genetic diversity was estimated by sequencing eight introns from five unlinked genes that encode house-keeping functions, collectively comprising 3,787 bp from each strain. The relationship among strains was evaluated using distance-based methods to reconstruct nearest-neighbor joining trees based on polymorphisms in the intron sequences. Phylogenetic reconstruction revealed 12 major nodes which were supported by bootstrap analysis and were defined as haplogroups (Fig. 1). These results are largely consistent with previous findings showing strong geographic separation between North and South and the presence of three major clonal groups in the North (referred to as “previous clonal” in Figs. 1–2, 5). New isolates from North America that were originally described as atypical based on RFLP patterns were found to comprise two groups (referred to as “new atypical” in Figs. 1–2, 5). First a set of strains found in pigs and found to have slightly divergent genotypes based on RFLP analysis (referred to as 3*, Supplementary Table S1), closely clustered with clonal type 3 (Fig. 1). Isolation of such strains from pigs is consistent with previous studies indicating that type 3 strains are common in domestic animals from North America (Howe and Sibley, 1995). Additionally, strain types A and X, which have been described from sea otters, were found to comprise a sister group of the clonal type 2 lineage (Fig. 1). These isolates were also highly similar to several isolates previously described from wild animals (white tailed deer, bears, naturally infected rodents (see Supplementary Table S1), and occasionally found in humans (ARI, RAY); these isolates were previously grouped as type 2-like strains (referred to as 2*) (Fig. 1, Supplementary Table S1). Both the monophyletic nature of isolates formerly referred to as types A, X and 2*, and their clear separation from the archetypal type 2, are supported by strong bootstrap values (Fig. 1). These findings indicate that although these strains are related to type 2, they are not members of this haplogroup. In keeping with the previously designated haplogroup nomenclature proposed for T. gondii (Khan et al., 2007), this new group is referred to here as type 12 (Fig. 1).

Fig. 1.

Fig. 1

Open in a new tab

Phylogenetic analysis based on intron sequences for 66 strains of Toxoplasma gondii. Strong geographic separation is evident between North American (blue lettering) and South American (red lettering) T. gondii strains. The major branches were classified into 12 haplogroups (numbers in boxes), of which groups 1, 2, 3, 11 and 12 are common to North America and Europe, whereas 4, 5, 6, 8, 9 and 10 are common to South America. An unrooted neighbor-joining tree was developed using MEGA with 50% majority rule and 1000 bootstrap replicates (indicated at each node as percentage values). Strains previously recognized as being highly clonal groups in North America (i.e. types 1, 2, 3) are referred to as “previous clonal”; while those comprising atypical strains based on previous restriction fragment length polymorphism typing, and which have been reanalyzed here with additional data, are referred to as “new atypical”. Strains previously listed as 2-like (2*), 3-like (3*) or A, X, are indicated and further defined in Supplementary Table S1.

Fig. 2.

Fig. 2

Open in a new tab

Phylogenetic analysis of Toxoplasma gondii strains using antigen-encoding genes. A) Combined phylogenetic tree for surface antigen 1 (SAG1) and dense granule proteins 6 and 7 (GRA6 and GRA7) for major North American and European groups. Previously defined groups A, X or 2* are part of group 12 and a sister taxa of type 2. Strains previously grouped as 3* are contained within group 3. An unrooted neighbor-joining tree was developed using MEGA with 50% majority rule and 1000 bootstrap replicates (indicated at each node as percentage values). Previous clonal and new atypical groupings are as defined in Fig. 1.

Fig. 5.

Fig. 5

Open in a new tab

Population analyses of Toxoplasma gondii. A) Analysis o f current population structure of T. gondii reveals 12 haplogroups (HG), which closely correspond to the haplogroups depicted by phylogenetic analysis (Fig. 1). STRUCTURE analysis was conducted using concatenated intron sequences analyzed with a Bayesian clustering algorithm using ancestral population sizes of K = 4 to 7. Group 12 is clearly separate from group 2 at all K values, while strains classified as 3* group closely with 3. B) Population structure between different haplogroups of T. gondii determined using principal coordinate analysis. Colored symbols indicate different haplogroups (blue, primarily from North America; red, primarily from South America). Percentage of the variance explained by axis one is ~54% whereas axes two and three each comprise ~14% of the variance. Previous clonal and new atypical groupings are as defined in Fig. 1.

3.2. Phylogenetic analyses of T. gondii strains based on antigen-encoding genes

Phylogenetic analyses of T. gondii based on selectively neutral loci such as introns is thought to provide an unbiased estimate of relatedness. However, other studies have emphasized the use of more variable loci encoding surface and secretory antigens (Miller et al., 2004; Pena et al., 2008; Sundar et al., 2008), which are often under selective pressure and thus capable of revealing greater genetic diversity. Phylogenetic analyses of a subset of 42 representative strains, based on sequences of the antigen-encoding genes SAG1, GRA6 and GRA7, revealed greater diversity in North America and Europe. However, the major groups were still preserved together with the overall branching patterns (Fig. 2). In particular, the 3* isolates from pigs were still closely clustered with type 3. Additionally, while greater divergence was detected with group 12, it remained a monophyletic sister group to type 2 (Fig. 2). The reconstructed ancestry based on a group of genes can often differ from gene trees built from single loci, reflecting the potential for different ancestries between distinct regions of the genome. Previous studies of T. gondii have stressed the preponderance of biallelic patterns, where most loci are comprised of only two distinct alleles (Boyle et al., 2006). Similarly, analysis of trees for individual loci revealed a common biallelic pattern between types 2 and 12 (Fig. 3). At some loci (UPRT, SAG1), types 2 and 12 shared the same allele. However, at most loci, some members of type 12 shared a new allele that was distinct, while others contained the same allele as type 2 (Fig. 3). Notably, the locus encoding GRA7 diverges from this pattern and appears to comprise several distinct alleles (Fig. 3). However, examination of the sequence alignment of this locus reveals that these differences are due to variable insertions/deletions within a 45 nucleotide block encoding 15 amino acids, some of which are absent or present in different strains. The occurrence of these patterns is not monophyletic, suggesting changes in this locus have arisen multiple times independently; hence, it is not a reliable index of common ancestry.

Fig. 3.

Fig. 3

Open in a new tab

Separate gene trees for each locus of the Toxoplasma gondii genome that was sequenced for the type 2 and related strains. Inheritance patterns reflect acquisition of the type 2 allele by all strains (UPRT, SAG1), or of either type 2 or γ allele at other loci. Unrooted neighbor-joining trees were generated using MEGA with 50% majority rule. Members of clonal group 2 are indicated in red.

3.3. Analysis of gene flow between groups

To further analyze whether new isolates constitute separate populations, we evaluated the variation in allele frequencies between groups using the FST statistic, which provides a measure of population subdivision between groups based on divergence of allele frequencies from random mating (Weir, 1996). FST varies from 0–1, with values > 0.25 indicating significant population subdivision. Consistent with the marked separation between the clonal lineages, high FST values were observed in pair-wise comparisons between groups 1, 2 and 3 (Table 1). Similar high values for FST were observed when comparing type 12 with I, 2 or 3, and when comparing all four groups together (Table 1). In contrast, when comparing previous clonal members of type 3, with new atypical isolates within this group (i.e. 3* isolates), FST remained relatively low (Table 1). Overall, these analyses provide strong support for type 12 being a lineage that is genetically separate from the other major lineages, while type 3* strains are merely minor variants of type 3 and do not denote a separate population. Similar genotypes were found in humans and animals, indicating an absence of strong separation between these groups (FST = 0.093). However, modest population structure was detected between domestic and wild animals (FST = 0.366), suggesting that these may constitute different populations.

Table 1.

Pairwise comparison of Fixation due to population substructure (Fst) and Index of Association (IA)

Haplogroups Fsta Iaa Polymorphic sitesb Two-tailed Fisher's exact testc Chi-squarec ZnS Number of Isolates
1 versus 2 0.90 21.98 25 276/0 276/0 0.9202d 12
1 versus 3 0.90 18.90 22 210/0 210/0 0.9333d 11
2 versus 3 0.91 17.01 19 171/0 171/0 0.9721d 13
1, 2, versus 3 0.90 10.85 33 241/73 392/151 0.4663 18
1 versus 12 0.77 24.07 23 229/172 231/172 0.8072d 21
2 versus 12 0.73 6.71 4 6/4 6/4 0.5501 23
3 versus 12 0.73 20.23 18 136/106 136/136 0.8313d 22
1, 2, 3, versus 12 0.79 11.33 32 308/117 346/173 0.3963 34
3 (3* versus 3) 0.03 −0.06 2 0/0 0/0 0.3750 10
Open in a new tab
a

1000 replicates used for calculation

b

Number of sites is 3787 bp

c

Number of significant comparisons without / with Bonferroni correction

d

P ≤ 0.05 (Kelly, J.K., 1997. A test of neutrality based on interlocus associations. Genetics 146, 1197–1206.)

3* refers to strains that are similar to type 3 but with additional plymorphism(s)

3.4. Evidence for clonality of type 12

Previous studies have stressed the clonal population structure of T. gondii in North America and Europe based on the presence of shared genotypes, absence of recombination, and high levels of LD (Sibley and Boothroyd, 1992; Howe and Sibley, 1995; Khan et al., 2007). Given the increased evidence of genetic diversity among animal isolates from North America, we re-evaluated the population structure based on diversity in intron sequences. Phylogenetic analysis of group 12 revealed that it contains highly related genotypes, some of which are identical or nearly identical (Fig. 1). To assess long-range LD, we concatenated the intron sequences and performed a standard test of D’, which measures the difference in association between alleles at separate loci from expected values based on random mating. As expected, the groups 1, 2 and 3 showed strong LD even across these otherwise unlinked loci (Fig. 4). This pattern was preserved when group 12 was included with the archetypal clonal lineages (Fig. 4 and data not shown). In contrast, strains of T. gondii from South America show much lower levels of LD, consistent with a higher rate of recombination among these strains (Fig. 4). We also estimated the IA, which provides a measure of independence of alleles at different loci among members of a population (Smith et al., 1993). Deviations of IA from 0 are reflective of low levels of recombination, supporting a clonal population structure. Pair-wise comparisons of groups 1, 2 and 3 gave high IA values, consistent with their highly clonal population structure, as reported previously (Sibley and Boothroyd, 1992; Howe and Sibley, 1995). Pairs of polymorphic loci were also strongly correlated when comparing haplogroup 12 with either haplogroups 1 or 3 (IA >20) as occurred when comparing haplogroups 1, 2 and 3 with each other (pair wise IA > 17) (Table 1). Similarly, using the ZnS statistic, which measures interlocus associations (Kelly, 1997), pairs of loci were strongly correlated when type 12 isolates were compared with those of types 1 or 3 (ZnS > 0.80, P < 0.5), approaching levels observed when types 1, 2 and 3 were compared with each other (ZnS > 0.90, P < 0.5). Analysis of IA or ZnS within group 3, containing both prior isolates and new isolates that cluster there, revealed low values, consistent with an absence of partitioning within this group (Table 1). Collectively, the combination of phylogenetic analysis, high FST values and elevated IA, and ZnS values, provides strong support for the conclusion that type 12 is a separate genetic lineage that has a clonal population structure.

Fig. 4.

Fig. 4

Open in a new tab

Linkage disequilibrium (LD) among different Toxoplasma gondiii population groups. Introns were concatenated for analyzing LD across the combined sequences. Strong LD was evident in groups 1, 2, 3, 12 where D', the scaled version of the disequilibrium coefficient, approached one among nucleotide pairs from unlinked loci, irrespective of their physical proximity.

3.5. Alternative assessments of population structure of T. gondii

To provide an alternative method for defining the current population structure of T. gondii, we implemented a Bayesian statistical method called STRUCTURE (Falush et al., 2003a) to analyze the genetic variation among strains based on intron sequences. STRUCTURE is a clustering method that estimates shared ancestry of subgroups within a larger population. Color blocks represent major ancestral contributions and mixed genotypes are revealed by the pattern of color combinations. We utilized an ancestral model with admixture to estimate the number of founder populations that would be required to construct the current population structure (K = 4–7 are shown in Fig. 5A). The most dominant feature evident in the output is the marked homogeneity of strains from North America (left side of figure, blue) compared with those from South America (right side, red). This reflects the strongly clonal population structure in the North, with greater genetic recombination being found in the South, as reported previously (Lehmann et al., 2006; Khan et al., 2007). In contrast, group 12 is dominated by a single color pattern in STRUCTURE, regardless of the K value, thus providing strong support for it being a separate population with a unique ancestry. Similarly, types 2 and 3 show strong within-group homogeneity, reflecting distinct ancestries with little mixture. The type 3* strains do not define a separate population by STRUCTURE analysis, but rather group closely with type 3. Other clonal groups such as type 1 appear to be a mixture of ancestral types, the exact origin of which varies depending on how many ancestral populations are modeled (different K values). Groups from South America typically show greater admixture, reflecting greater recombination in their ancestries. The ancestral population size K= 6 was considered the best estimate of the current population structure (Fig. 5A) as this most closely resembles the population structure derived from phylogenetic analysis (Fig. 1).

We also used a multivariate approach based on PCA to provide information about shared variance between major groups. PCA analysis combines variation from multiple traits (polymorphisms here) into a smaller number of “components” and then groups strains based on the extent to which they share these components by plotting their distribution along major axes. Although haplogroups 5, 6, 8, 9 and 10 were separated by phylogenetic analysis (Fig. 1), they clustered together in PCA, indicating their relatedness (Fig. 5B). This is also reflected in STRUCTURE analysis where haplogroup 5 was largely similar to 6, while groups 8, 9, 10 shared much of their ancestry in common. Haplogroup 3, which contains both prior clonal and new atypical strains, was tightly clustered in the PCA analysis, which was consistent with both phylogenetic and STRUCTURE analyses. Group 12 was related to group 2 along all three major axes (i.e. components) but formed a clearly distinct group (Fig. 5B).

3.6. Infrequent recombination and clonal expansion of T. gondii in North America

Previous studies have shown that the three archetypal clonal lineages share two distinct alleles at each locus, indicating they arose when the ancestor of type 2 (referred to here as II*) recombined separately with two other lineages (Boyle et al., 2006). Based on genetic similarity, it has previously been proposed that the alternative parental strain that gave rise to type 3 was similar to the strain P89 (haplogroup 9), designated as β (Fig. 6) (Boyle et al., 2006). Our data are consistent with this and furthermore suggest that atypical strains related to type 3 arose from this same event and have undergone slight additional diversification by mutational drift (Fig. 6). The alternative parental strain giving rise to type I has not been identified, but it has been designated as α (Fig. 6) (Boyle et al., 2006). Subsequent expansion of the progency from these crosses gave rise to the clonal linages, with minor variants arising by mutation within distinct strains. Our analysis suggests that a distinct genetic cross occured with between a unique parental strain, designated here as γ and the II* ancestral strain to give rise to group 12 (Fig. 6).

Fig. 6.

Fig. 6

Open in a new tab

Proposed model of recent expansion of clonal populations of Toxoplasma gondii in North America and Europe. Four major clonal groups including types 1, 2, 3 and 12 have expanded throughout North America. Three separate crosses between the ancestral II* and separate parental types designated as α, β and γ, gave rise to types 1, 3 and 12, respectively. Age estimates are based on previous studies (Su et al., 2003a; Khan et al., 2007).

One of the predictions of this scenario is that strains from group 12 should contain alleles that exactly match the type 2 lineage at some loci, while others should be unique, having been derived from the γ ancestral type. Comparison of separate gene trees for each of the loci analyzed fits this prediction (Fig. 3). For example, all of the type 12 strains share the same type 2 allele at the UPRT and SAG1 loci; at other loci, such as EF1 and HP2, they all differ from type 2 but share a common γ allele (Fig. 3). The separate subgroups all appear to be the progeny of a single cross and their respective genotypes differ by inheritance of type 2 versus γ alleles through reassortment and/or recombination (Fig. 3).

Strains of T. gondii differ dramatically in acute virulence in mice, despite their close genetic similarity. For example, type 1 strains are uniformly lethal at all doses, while type 2 strains are relatively less virulent (Sibley and Boothroyd, 1992). Given the somewhat close relationship between type 2 and type 12 strains, we examined their phenotypes following i.p. inoculation in mice. At low inoculation doses (i.e. 10 tachyzoites) no infections were noted with any of the type 12 strains, while infections with Me49 were routinely observed at this dose, as verified by serological status (data not shown). Whether this was due to lower viability in culture or an intrinsic limitation in vivo is uncertain. Consequently, we chose to compare the survival data for doses of 100 and 1,000 parasites only, in each case adjusting the cumulative mortality to remove animals that showed no sign of infection (i.e. those that remained seronegative). Surprisingly, type 12 strains showed a range of pathogenesis including those with low virulence such as Me49, as well as isolates with intermediate to high levels of cumulative mortality (Fig. 7). Despite these high levels of mortality, all type 12 isolates also gave rise to chronic infections characterized by animals that survived acute infection and remained seropositive (data not shown), thus distinguishing them from the highly virulent type 1 lineage (Sibley and Boothroyd, 1992).

Fig. 7.

Fig. 7

Open in a new tab

Acute virulence of Toxoplasma gondii type 12 strains as monitored in mice. Survival of CD1 outbred mice was monitored for 30 days following i.p. inoculation with different doses of tachyzoites indicated. The type 2 strain Me49 is shown for comparison. Cumulative mortality was defined as the number of animals that succumbed / total number of animals infected (those that died + those that were seropositive but survived) for both doses combined.

4. Discussion

We have undertaken a re-examination of the population structure of T. gondii in North America in order to evaluate the relationship of atypical strains initially identified through RFLP typing. Sequence-based phylogenetic reconstructions and a variety of population genetic analyses revealed that some of these atypical isolates constitute a new lineage while others are merely subtle variants of previously well-characterized ones. To provide a basis for future comparative analyses, we define criteria for grouping strains into major haplogroups based on shared genotypes, common ancestry and restricted gene flow. Finally, our findings reaffirm the clonal population pattern of strains in North America and reveal that the ancestral type 2 lineage has been the common parental stock for at least three independent crosses that led to clonal expansion.

Previous studies based on RFLP typing revealed that although most North American isolates belong to one of three canonical clonal lineages (1, 2 or 3), others possess unique alleles or uncommon combinations of alleles. Although extremely valuable for performing population surveys, RFLP analysis captures only a fraction of possible genetic variation. Hence, it is difficult to tell from such data how these atypical patterns relate to other major groupings in the population. Sequence-based analysis of a select group of these atypical strains revealed several underlying relationships with existing strain types. Some isolates, or groups of isolates, differ from others by having accumulated one or a few distinguishing mutations. Based on a variety of criteria, these strains do not constitute new lineages but instead are minor variants within existing groups. However, the lineages they belong to still fit the definition of “clonal” despite having some members that are not identical. These two conditions can coexist because “clonal” strains are not of necessity genetically identical but rather share a preponderance of alleles at otherwise unlinked loci. Because they are propagating asexually, recombination is limited and they show high levels of IA, LD and are typically represented by over abundance of similar genotypes among independent isolates. The new variants from domestic animals (3*) that group with type 3 are an example of this pattern.

By contrast, atypical isolates that were formerly designated as X, A and 2* were found to comprise an entirely new and unique lineage designated here as 12, which is abundant in North America. Criteria for type 12 being a separate lineage include phylogenetic support for a separate major branch, high levels of FST indicating population subdivision and a distinct grouping as revealed in STRUCTURE. Further analysis of group 12 reveals that it also has a clonal population structure based on over-representation of highly similar genotypes, strong LD and high levels of IA. This pattern is likely to be similar for other related isolates that share the atypical RFLP patterns of this group (Miller et al., 2004; Dubey et al., 2008a, b, c; Sundar et al., 2008), although it is also possible that additional diversity will be revealed by examining further isolates. To date, type 12 strains have not been reported from Europe, which otherwise share a preponderance of the three clonal types, with type 2 being especially common (Howe and Sibley, 1995; Ajzenberg et al., 2002). This suggests that type 12 may be endemic to North America, possibly the result of genetic recombination with a native North American isolate and ancestral type 2. Despite the presence of a fourth major group in North America, the overall population structure remains highly clonal.

Interestingly, group 12 includes a number of isolates from wild animals as well as several from humans (ARI, RAY), which were formerly typed as atypical isolates based on RFLP analysis (Howe and Sibley, 1995). This group also contains members of types X and A, which have previously been associated with severe encephalitis in sea otters in the western United States (Miller et al., 2004; Sundar et al., 2008). Interestingly, there was a strong association of type 12 with wild animals, while more domestic animals were infected with type 3, leading to population substructure revealed by FST. It has previously been noted that type X (referred to as 12 here) strains appear to be more common in wild felids, as opposed to domestic cats (Miller et al., 2008). Whether type 12 represents a sylvatic cycle that is genetically separate from a domestic cycle will require further isolates for more thorough analysis. Although it has been suggested that such strains are more pathogenic in sea otters (Miller et al., 2004), similar isolates are also found in asymptomatic wild animals (white tailed deer, turkey, black bear, rodents (Supplementary Table S1)). Moreover, type 2 strains of T. gondii, which are relatively avirulent in outbred laboratory mice, have also been associated with severe disease in otters (Cole et al., 2000). Direct testing of type 12 strains in mice revealed that some express intermediate or high levels of acute virulence. As type 2 strains are generally of low virulence in outbred mice, this suggests that the γ parental strain that contributed to the type 12 strains may have had high levels of virulence, or that unique combinations of genes are responsible for the observed phenotypes.

Although early studies stressed the high degree of similarity among T. gondii strains, the advent of methods to define different genotypes has led to a plethora of data on isolates and their unique genetic characteristics. Emerging sequencing technologies will accelerate discovery of polymorphisms among isolates of T. gondii, increasing the importance of systems to meaningfully classify and analyze such variation. Aside from defining population structure, there are also important biological reasons to group strains that are genetically related as they are likely to share major phenotypic traits. This has been aptly illustrated by the mapping of genes that confer important differences in virulence and gene expression between abundant lineages of T. gondii (Saeij et al., 2006, 2007; Taylor et al., 2006). Defining genes that underlie major phenotypic differences between archetypal strains has also been useful for establishing patterns of inheritance that may have shaped the spread of mouse virulence through larger populations (Khan et al., 2009). Application of similar genetic analyses to type 12 strains may reveal the molecular basis for their enhanced pathogenicity.

In an effort to support the rational grouping of strains into major genetic types, we have illustrated a series of genetic analyses that are capable of clearly defining new groups as distinct from previous ones and also of defining to what extent they may be clonal versus sexually propagating. The outcomes of such analyses best reflect organismal history when applied to selectively neutral loci. Introns, and perhaps other noncoding regions, meet these criteria, but other loci such as those encoding antigenic genes may be less suitable. For example, GRA7 is known to be highly antigenic (Tan et al., 2010) and under strong selection pressure (Khan et al., 2009). Phylogenetic analysis of this gene suggests that it has undergone frequent insertion/deletion or rearrangement and does not provide an accurate index of ancestry. Including such markers may obscure meaningful genetic relationships among strains, and yet such incongruent loci are readily identifiable by the methods outlined here.

Collectively, it is evident from our studies and previous work that the existing common lineages in North America were derived from a very few crosses, each of which involved the ancestral type 2 lineage as one of the parents. Limited genetic exchange appears to have occurred after this point during which the strains expanded clonally, although this does not discount the possibility that under some situations, mating may be more common in the wild. Our findings furthermore suggest that the offspring of clones derived from the ancestral type 2 background enjoy some unusual fitness that is associated with clonal propagation. Wider comparisons with more divergent strains from other localities will be necessary to define the molecular basis for the success of type 2-derived lineages.

Supplementary Material

01
02

Acknowledgements

We thank Jack Remington (Stanford University Sch. Med, USA), Marie Laure Dardé (University of Limoges, France), and Jack Frenkel (Santa Fe, New Mexico, USA) for T. gondii isolates, Michael Grigg (NIH, Bethesda MD, USA) for helpful discussions, and Julie Nawas for technical assistance. This work was supported by a grant from the National Institutes of Health, USA (AI059176).

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Agapow PM, Burt A. Indices of multilocus disequilibrium. Molecular Ecology Notes. 2001;1:101–102. [Google Scholar]
  2. Ajzenberg D, Cogné N, Paris L, Bessieres MH, Thulliez P, Fillisetti D, Pelloux H, Marty P, Dardé ML. Genotype of 86 Toxoplasma gondii isolates associated with human congenital toxoplasmosis and correlation with clinical findings. J. Infect. Dis. 2002;186:684–689. doi: 10.1086/342663. [DOI] [PubMed] [Google Scholar]
  3. Boyle JP, Rajasekar B, Saeij JPJ, Ajioka JW, Berriman M, Paulsen I, Sibley LD, White M, Boothroyd JC. Just one cross appears capable of dramatically altering the population biology of a eukaryotic pathogen like Toxoplasma gondii. Proc. Natl. Acad. Sci. (USA) 2006;103:10514–10519. doi: 10.1073/pnas.0510319103. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Cole RA, Lindsay DS, Howe DK, Roderick CL, Dubey JP, Thomas NJ, Baeten LA. Biological and molecular characterizations of Toxoplasma gondii strains obtained from southern sea otters (Enhydra lutris nereis) J. Parasitol. 2000;86:526–530. doi: 10.1645/0022-3395(2000)086[0526:BAMCOT]2.0.CO;2. [DOI] [PubMed] [Google Scholar]
  5. Dubey JP, Frenkel JF. Cyst-induced toxoplasmosis in cats. J. Protozool. 1972;19:155–177. doi: 10.1111/j.1550-7408.1972.tb03431.x. [DOI] [PubMed] [Google Scholar]
  6. Dubey JP, Hill DE, Sundar N, Velmurugan GV, Bandini LA, Kwok OC, Pierce V, Kelly KA, Dulin M, Thulliez P, Iwueke C, Su C. Endemic toxoplasmosis in pigs on a farm in Maryland: isolation and genetic characterization of Toxoplasma gondii. J. Parasitol. 2008a;94:36–41. doi: 10.1645/GE-1312.1. [DOI] [PubMed] [Google Scholar]
  7. Dubey JP, Quirk T, Pitt JA, Sundar N, Velmurugan GV, Kwok OC, Leclar D, Hill R, Su C. Isolation and genetic characterization of Toxoplasma gondii from raccoons (Procyon lotor), cats (Felis domesticus), striped skunk (Mephitis mephitis), black bear (Ursus americanus), and cougar (Puma concolor) from Canada. J. Parasitol. 2008b;94:42–45. doi: 10.1645/GE-1349.1. [DOI] [PubMed] [Google Scholar]
  8. Dubey JP, Sundar N, Hill D, Velmurugan GV, Bandini LA, Kwok OC, Majumdar D, Su C. High prevalence and abundant atypical genotypes of Toxoplasma gondii isolated from lambs destined for human consumption in the USA. Int. J. Parasitol. 2008c;38:999–1006. doi: 10.1016/j.ijpara.2007.11.012. [DOI] [PubMed] [Google Scholar]
  9. Dubey JP. Toxoplasmosis of animals and humans. Boca Raton: CRC Press; 2010. [Google Scholar]
  10. Falush D, Stephens M, Pritchard JK. Inference of population structure using multilocus genotype data: linked loci and correlated allele frequencies. Genetics. 2003a;164:1567–1587. doi: 10.1093/genetics/164.4.1567. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Falush D, Wirth T, Linz B, Pritchard JK, Stephens M, Kidd M, Blaser MJ, Graham DY, Vacher S, Perez-Perez GI, Yamaoka Y, Megraud F, Otto K, Reichard U, Katzowitsch E, Wang X, Achtman M, Suerbaum S. Traces of human migrations in Helicobacter pylori populations. Science. 2003b;299:1582–1585. doi: 10.1126/science.1080857. [DOI] [PubMed] [Google Scholar]
  12. Higgins DG, Thompson JD, Gibson TJ. Using CLUSTAL for multiple sequence alignments. Methods Enzymol. 1996;266:382–402. doi: 10.1016/s0076-6879(96)66024-8. [DOI] [PubMed] [Google Scholar]
  13. Howe DK, Sibley LD. Toxoplasma gondii comprises three clonal lineages: correlation of parasite genotype with human disease. J. Infect. Dis. 1995;172:1561–1566. doi: 10.1093/infdis/172.6.1561. [DOI] [PubMed] [Google Scholar]
  14. Huff DR, Peakall R, Smouse PE. RAPD variation within and among populations of outcrossing buffalograss (Buchloë dactyloides (Nutt.) Engelman) Theor. Appl. Genetics. 1993;96:827–834. doi: 10.1007/BF00211043. [DOI] [PubMed] [Google Scholar]
  15. Jones JL, Muccioli C, Belfort R, Jr, Holland GN, Roberts JM, Silveira C. Recently acquired Toxoplasma gondii infection. Brazil. Emerg. Infect. Dis. 2006;12:582–587. doi: 10.3201/eid1204.051081. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Kelly JK. A test of neutrality based on interlocus associations. Genetics. 1997;146:1197–1206. doi: 10.1093/genetics/146.3.1197. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Khan A, Jordan C, Muccioli C, Vallochi AL, Rizzo LV, Belfort R, Jr, Vitor RW, Silveira C, Sibley LD. Genetic divergence of Toxoplasma gondii strains associated with ocular toxoplasmosis Brazil. Emerg. Infect. Dis. 2006;12:942–949. doi: 10.3201/eid1206.060025. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Khan A, Fux B, Su C, Dubey JP, Darde ML, Ajioka JW, Rosenthal BM, Sibley LD. Recent transcontinental sweep of Toxoplasma gondii driven by a single monomorphic chromosome. Proc. Natl. Acad. Sci. (USA) 2007;104:14872–14877. doi: 10.1073/pnas.0702356104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Khan A, Taylor S, Ajioka JW, Rosenthal BM, Sibley LD. Selection at a single locus leads to widespread expansion of Toxoplasma gondii lineages that are virulence in mice. PLoS Genetics. 2009;5 doi: 10.1371/journal.pgen.1000404. e1000404. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Kumar S, Tamura K, Jakobsen IB, Nei M. MEGA2: molecular evolutionary genetics analysis software. Bioinformatics. 2001;12:1244–1245. doi: 10.1093/bioinformatics/17.12.1244. [DOI] [PubMed] [Google Scholar]
  21. Lehmann T, Marcet PL, Graham DH, Dahl ER, Dubey JP. Globalization and the population structure of Toxoplasma gondii. Proc. Natl. Acad. Sci. (USA) 2006;103:11423–11428. doi: 10.1073/pnas.0601438103. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Miller MA, Grigg ME, Kreuder C, James ER, Melli AC, Crosbie PR, Jessup DA, Boothroyd JC, Brownstein D, Conrad PA. An unusual genotype of Toxoplasma gondii is common in California sea otters (Enhydra lutris nereis) and is a cause of mortality. Intl. J. Parasitol. 2004;34:275–284. doi: 10.1016/j.ijpara.2003.12.008. [DOI] [PubMed] [Google Scholar]
  23. Miller MA, Miller WA, Conrad PA, James ER, Melli AC, Leutenegger CM, Dabritz HA, Packham AE, Paradies D, Harris M, Ames J, Jessup DA, Worcester K, Grigg ME. Type X Toxoplasma gondii in a wild mussel and terrestrial carnivores from coastal California: new linkages between terrestrial mammals, runoff and toxoplasmosis of sea otters. Int. J. Parasitol. 2008;38:1319–1328. doi: 10.1016/j.ijpara.2008.02.005. [DOI] [PubMed] [Google Scholar]
  24. Peakall R, Smouse PE. GenAIEx 6: Genetic analysis in Excel. Population genetic software for teaching and research. Molec. Ecol. Notes. 2005;6:288–295. doi: 10.1093/bioinformatics/bts460. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Pena HF, Gennari SM, Dubey JP, Su C. Population structure and mouse-virulence of Toxoplasma gondii in Brazil. Intl. J. Parasitol. 2008;38:561–569. doi: 10.1016/j.ijpara.2007.09.004. [DOI] [PubMed] [Google Scholar]
  26. Rozas J, Sanchez-DelBarrio JC, Messeguer X, Rozas R. DnaSP DNA polymorphism analyses by the coalescent and other methods. Bioinformatics. 2003;19:2496–2497. doi: 10.1093/bioinformatics/btg359. [DOI] [PubMed] [Google Scholar]
  27. Saeij JPJ, Boyle JP, Coller S, Taylor S, Sibley LD, Brooke-Powell ET, Ajioka JW, Boothroyd JC. Polymorphic secreted kinases are key virulence factors in toxoplasmosis. Science. 2006;314:1780–1783. doi: 10.1126/science.1133690. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Saeij JPJ, Coller S, Boyle JP, Jerome ME, White ME, Boothroyd JC. Toxoplasma co-opts host gene expression by injection of a polymorphic kinase homologue. Nature. 2007;445:324–327. doi: 10.1038/nature05395. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Sibley LD, Boothroyd JC. Virulent strains of Toxoplasma gondii comprise a single clonal lineage. Nature (Lond.) 1992;359:82–85. doi: 10.1038/359082a0. [DOI] [PubMed] [Google Scholar]
  30. Sibley LD, Ajioka JW. Population structure of Toxoplasma gondii: Clonal expansion driven by infrequent recombination and selective sweeps. Ann. Rev. Microbiol. 2008;62:329–351. doi: 10.1146/annurev.micro.62.081307.162925. [DOI] [PubMed] [Google Scholar]
  31. Silveira C, Belfort R, Jr, Muccioli C, Abreu MT, Martins MC, Victora C, Nussenblatt RB, Holland GN. A follow-up study of Toxoplasma gondii infection in southern Brazil. Am. J. Ophthalmol. 2001;131:351–354. doi: 10.1016/s0002-9394(00)00830-8. [DOI] [PubMed] [Google Scholar]
  32. Smith JM, Smith NH, O'Rouke M, Spratt BG. How clonal are bacteria? Proc. Nat. Acad. Sci. (USA) 1993;90:4384–4388. doi: 10.1073/pnas.90.10.4384. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Su C, Howe DK, Dubey JP, Ajioka JW, Sibley LD. Identification of quantitative trait loci controlling acute virulence in Toxoplasma gondii. Proc. Natl. Acad. Sci. (USA) 2002;99:10753–10758. doi: 10.1073/pnas.172117099. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Su C, Evans D, Cole RH, Kissinger JC, Ajioka JW, Sibley LD. Recent expansion of Toxoplasma through enhanced oral transmission. Science. 2003a;299:414–416. doi: 10.1126/science.1078035. [DOI] [PubMed] [Google Scholar]
  35. Su C, Hott C, Brownstein BH, Sibley LD. Typing single nucleotide polymorphisms in Toxoplasma gondii by allele-specific primer extension and microarray detection. Methods Mol. Biol. Parasite Genomics Protocols. 2003b;270:249–262. doi: 10.1385/1-59259-793-9:249. [DOI] [PubMed] [Google Scholar]
  36. Sundar N, Cole RA, Thomas NJ, Majumdar D, Dubey JP, Su C. Genetic diversity among sea otter isolates of Toxoplasma gondii. Vet . Parasitol. 2008;151:125–132. doi: 10.1016/j.vetpar.2007.11.012. [DOI] [PubMed] [Google Scholar]
  37. Tan TG, Mui E, Cong H, Witola WH, Montpetit A, Muench SP, Sidney J, Alexander J, Sette A, Grigg ME, Maewal A, McLeod R. Identification of T. gondii epitopes, adjuvants, and host genetic factors that influence protection of mice and humans. Vaccine. 2010;28:3977–3989. doi: 10.1016/j.vaccine.2010.03.028. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Taylor S, Barragan A, Su C, Fux B, Fentress SJ, Tang K, Beatty WL, Haijj EL, Jerome M, Behnke MS, White M, Wootton JC, Sibley LD. A secreted serine-threonine kinase determines virulence in the eukaryotic pathogen Toxoplasma gondii. Science. 2006;314:1776–1780. doi: 10.1126/science.1133643. [DOI] [PubMed] [Google Scholar]
  39. Weir BS. Genetic Data Analysis II Methods for Discrete Population Genetic Data. Sunderland, MA: Sinaur Assoc.; 1996. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

01
NIHMS275145-supplement-01.xls (126.5KB, xls)
02
NIHMS275145-supplement-02.xls (69KB, xls)

RESOURCES