Differences between Trypanosoma brucei gambiense Groups 1 and 2 in Their Resistance to Killing by Trypanolytic Factor 1

Background The three sub-species of Trypanosoma brucei are important pathogens of sub-Saharan Africa. T. b. brucei is unable to infect humans due to sensitivity to trypanosome lytic factors (TLF) 1 and 2 found in human serum. T. b. rhodesiense and T. b. gambiense are able to resist lysis by TLF. There are two distinct sub-groups of T. b. gambiense that differ genetically and by human serum resistance phenotypes. Group 1 T. b. gambiense have an invariant phenotype whereas group 2 show variable resistance. Previous data indicated that group 1 T. b. gambiense are resistant to TLF-1 due in-part to reduced uptake of TLF-1 mediated by reduced expression of the TLF-1 receptor (the haptoglobin-hemoglobin receptor (HpHbR)) gene. Here we investigate if this is also true in group 2 parasites. Methodology Isogenic resistant and sensitive group 2 T. b. gambiense were derived and compared to other T. brucei parasites. Both resistant and sensitive lines express the HpHbR gene at similar levels and internalized fluorescently labeled TLF-1 similar fashion to T. b. brucei. Both resistant and sensitive group 2, as well as group 1 T. b. gambiense, internalize recombinant APOL1, but only sensitive group 2 parasites are lysed. Conclusions Our data indicate that, despite group 1 T. b. gambiense avoiding TLF-1, it is resistant to the main lytic component, APOL1. Similarly group 2 T. b. gambiense is innately resistant to APOL1, which could be based on the same mechanism. However, group 2 T. b. gambiense variably displays this phenotype and expression does not appear to correlate with a change in expression site or expression of HpHbR. Thus there are differences in the mechanism of human serum resistance between T. b. gambiense groups 1 and 2.


Introduction
Trypanosoma brucei is an important parasite of sub-Saharan Africa. The species is commonly divided into three distinct subspecies that differ in geographical distribution, host range and the disease they cause. T. b. brucei is the most widespread form of the parasite, infecting a wide range of mammals, but is not able to infect humans because it is susceptible to trypanosome lytic factors (TLF) found in normal human serum [1]. T. b. gambiense and T. b. rhodesiense are, in contrast, resistant to lysis by human serum and so are able to infect humans. The West African trypanosome T. b. gambiense is the most prevalent form of the human-infective parasite and is responsible for greater than 90% of reported African sleeping sickness cases [2]. An important facet of African trypanosome biology is that the adaptive immune response of vertebrates is rendered largely ineffective due to the parasite's ability to cyclically evade the immune response by changing the variable surface glycoprotein (VSG) antigens on the cell surface [3,4]. Several primates, including humans, therefore rely on innate immunity to prevent infection in the form of TLFs in serum that kill most species of trypanosome [5,6,7].
There are two distinct forms of TLF particle (TLF-1 and 2) in human serum. TLF-1 is a member of the high density lipoprotein (HDL) family of particles [8,9] while TLF-2 is a related high molecular weight serum protein binding complex [10]. Considerable effort and much debate has been focused on determining the cytotoxic component of these TLFs, predominantly focusing on TLF-1 due to the difficulty of obtaining active TLF-2 [10].
Investigation has centered on two primate-specific proteins found in TLF; haptoglobin-related protein (HPR) and apolipoprotein L1 (APOL1). While there has been some controversy concerning the roles of these two proteins, the current consensus is that both proteins are necessary for optimal lysis and that HPR and APOL1 have complementary roles [11,12]. HPR is the ligand that facilitates uptake of TLF-1 via a parasite haptoglobin-hemoglobin receptor (HpHbR) [13,14] and it has also been shown to have some trypanosome specific toxicity [15], possibly due to an uncleaved signal peptide that affects membrane fluidity [16]. After internalization, the TLF-1 particle is then trafficked to the lysosome where APOL1 is activated by a pH-mediated conformational change to form pores in the lysosome membrane. This leads to perturbation of the osmotic balance of the organelle and subsequent lysis [17,18]. As both HPR and APOL1 are found in TLF-2, it is likely that the lysis method of this particle involves both of these proteins, although the method of entry for TLF-2 does not directly involve the HpHbR receptor [13,19].
The two human infective trypanosome sub-species have evolved counter-measures to overcome the innate defense factors. Understanding the mechanism of how these parasites resist lysis may lead to the development of new treatment strategies. For example, the resistance mechanism for T. b. rhodesiense is dependent upon the expression of a serum resistance associated (SRA) gene. This has led to new diagnostic techniques [20] and therapeutic possibilities targeting SRA [21,22]. The SRA gene has not been found in any T. b. gambiense isolates however [23,24,25], so an alternate and still unknown mechanism of human serum resistance exists in this subspecies.
A complicating factor is that T. b. gambiense possesses two distinct ''groups'' that differ in genotype and phenotype. Isoenzyme and molecular data show that group 1 and group 2 T. b. gambiense populations are reliably distinguishable from each other, and that group 2 is more akin to T. b. brucei than group 1 T. b. gambiense [26,27,28,29,30,31,32,33,34,35]. In group 1 T. b. gambiense, human serum resistance is a stable trait [31,33] and it has been suggested that a reduced uptake of TLF-1 may be partially responsible for resistance. These parasites exhibit reduced expression of HbHpR, which may contribute to their ability to infect humans by avoiding TLF-1 [36]. In addition, mutant or transgenic T. b. brucei lines with reduced expression of the HpHbR receptor gene are resistant to TLF-1 particles [13,36], although these lines are still lysed by normal human serum and TLF-2. This raises the possibility that group 1 T. b. gambiense parasites are sensitive to APOL1 and that they avoid lysis by failing to take up APOL1 whether it is in TLF-1 or 2. However, uptake of TLF-2 in these parasites has not so far been examined. In contrast, group 2 T. b. gambiense exhibit a variable human serum resistance phenotype in a manner superficially similar to T. b. rhodesiense, raising the question: do groups 1 and 2 T. b. gambiense share a common mechanism of human serum resistance? To address this question we have compared the properties of the human serum resistance phenotypes of group 1 and group 2 T. b. gambiense by examining the effects and uptake of both TLF-1 and recombinant APOL1 in these parasites.

Methods
T. b. gambiense, T. b. brucei and T. b. rhodesiense cell lines The human serum resistant group 2 T. b. gambiense strain STIB386 was originally isolated in 1978 from an infected patient in the Ivory Coast. ELIANE is a group 1 T. b. gambiense strain isolated from a human infected while in Côte d'Ivoire in 1952 [37]. STIB247 is a human serum sensitive T. b. brucei clone, first isolated from a hartebeest in Serengeti in 1971 and the T. b. rhodesiense strain Baganzi was isolated from a human in South-Eastern Uganda in 1990. All lines were maintained in vitro in HMI9 medium [38] supplemented by 1.5 mM glucose, 1 mM methyl cellulose, 250 mM adenosine, 150 mM guanosine and 20% serum (fetal bovine serum (FBS) for T. b. brucei and sensitive group 2 T. b. gambiense or human serum for the other strains) . All lines were regularly assayed for human serum sensitivity/resistance.

Human serum resistance & APOL1 assays
Trypanosomes were diluted to 10 6 per ml in modified HMI9 and incubated in 25% human serum or FBS in a 1 ml volume in a standard 24 well plate. The numbers of cells in each well were counted with a hemocytometer for the zero time point. The cells were incubated with 5% CO 2 at 37uC and the number of viable motile trypanosomes in each well at 6 hours was quantified by microscopy using a hemocytometer and the percentage of viable cells calculated compared to time zero.
For the APOL1 lysis assays, a dilution series of recombinant APOL1 (5-50 mg/ml) was formulated that and made up to equal volumes with protein free buffer (0.2 M acetic acid and 0.05% Tween20). A control containing an equal volume of buffer was also prepared. Each protein volume and the control were aliquoted into different wells and the cells incubated with 5% CO 2 at 37uC in HMI9 medium containing 25% FBS. The number of viable motile trypanosomes in each well was recorded at 24 hours and compared to the control wells containing no APOL1 to determine percentage survival. In each assay, cells were incubated in 25% normal human serum as a positive control.

Generation of recombinant APOL1
An Invitrogen GatewayH-compatible entry vector containing the APOL1 open reading frame (ORF) (Genecopoeia Inc, USA) was used in conjunction with the Invitrogen GatewayH expression system. The APOL1 was cloned into pDest17 destination vector containing an N-terminal 66His-tag and transformed into BL21-AI competent E. coli. Protein expression was induced using 0.1% L-Arabinose for 16 hours at 37uC. Cells were lysed with guanidinium lysis buffer pH 7.8 (6 M Guanidine Hydrochloride,

Author Summary
The sub-species of Trypanosoma brucei are important pathogens of humans and animals in sub-Saharan Africa. T. b. brucei is not able to infect humans due to sensitivity to trypanosome lytic factors (TLF) containing the lytic protein apolipoprotein L1 (APOL1). T. b. gambiense is the most prevalent human infective sub-species, although there are two distinct sub-groups of T. b. gambiense that differ genetically and in terms of their human serum resistance phenotype stability. How they resist lysis by TLF is unknown. Previous data indicates that some group 1 T. b. gambiense parasites are able to avoid lysis due in part to reduced expression and activity of the gene encoding a TLF receptor (HpHbR). The authors showed that despite group 1 T. b gambiense displaying avoidance of some TLF particles, this sub-species group is resistant to the lytic protein of TLF, APOL1. This suggests that avoidance is not the complete story in this sub-species, as previously theorised. Group 2 T. b. gambiense is also innately resistant to APOL1 and there is no evidence of avoidance of TLF in resistant parasites. Group 2 T. b. gambiense variably displays the human serum resistance phenotype that does not appear to correlate with expression of HbHpR.
0.02 M Sodium Phosphate, 0.5 M NaCl) for 5 minutes and the cellular debris removed with a 0.2 mm filter (Sartorius). The cell lysate was bound to Ni-NTA beads (Invitrogen) for one hour at pH 7.8 and then washed twice with wash buffer (8 M Urea, 0.02 M Sodium Phosphate, 0.5 M NaCl) at pH 7.8, followed by two washes at pH 6 and two washes at pH 5.8. Finally, bound protein was eluted with wash buffer at pH 4. The eluate was dialyzed overnight against 0.2 M acetic acid and 0.05% Tween20 and concentrated using 10,000MW Vivaspin columns (Sartorius). Protein purity was estimated using a NanodropH spectrometer (Nanodrop) and SDS-PAGE. A Western blot using an antibody raised against an APOL1 peptide (Sigma-Aldrich) was used to check that the single band present in the purified preparation was APOL1. Activity of the recombinant protein was determined by its ability to lyse T. b. brucei. The specific activity of the protein was estimated using a previously defined method using the amount protein needed to lyse 50% of the cell within 2 hours [17]. This value was determined to be 0.143 units/mg, which is comparable to purified native APOL1, although much lower than an intact TLF-1 particle that has a specific activity of 83.3 units/mg [39].
Fluorescence microscopy of TLF-1 uptake TLF-1 was prepared as previously described [10] and the purity of preparation was verified by both western blot and SDS PAGE using an APOL1 specific antibody (Sigma-Aldrich). TLF-1 was labeled with AlexaFluorH488 (Molecular Probes, Invitrogen) using the manufacturer's instructions. In control experiments, bovine HDL of a density comparable to TLF-1 [10] was also AlexaFluor tagged. Trypanosomes were re-suspended in serum-free HMI9 medium at a concentration of 10 6 cells/ml and incubated in 10 mg/ml of LysotrackerH (Invitrogen) and 5 mg/ml of the purified AlexaFluor tagged human TLF-1 [10]. The cells were incubated at 37uC for 30 minutes, 1, 2 and 4 hours. At each time point, cells were washed once in serum-free HMI9 medium and fixed by immersion in chilled 2.5% gluteraldhyde (Sigma-Aldrich) in phosphate buffered saline for 5 minutes. The cells were resuspended in 50% glycerol, 0.1% DAPI, 2.5% DABCO in phosphate buffered saline (PBS) and spread onto lysine-coated slides which were then protected with cover-slips sealed using ethyl acetate.
Slides were imaged using the Deltavision Core system and SoftWorx package (Applied Precision) with standard filter sets (DAPI/FITC/Texas-Red and Light transmission). Approximately 30 serial sections through each trypanosome were taken for each filter. The images were composited and the brightness, contrast and color levels normalized between samples and exposures using the ImageJ software package (US National Institute of Health). Approximately thirty trypanosomes were imaged per time point to give an indication of the uptake in the population. The Pearson's correlation co-efficient between the TLF and LysotrackerH fields was calculated using the Pearson-Spearman Correlation (PSC) Plug-in for ImageJ (http://www.cpib.ac.uk/,afrench/coloc. html).

Fluorescence microscopy of APOL1 uptake
Trypanosomes were re-suspended in HMI9 medium containing 20% fetal bovine serum at a concentration of 10 6 cells/ml and incubated in 10 mg/ml of LysotrackerH (Invitrogen) and purified recombinant APOL1 (5-50 mg/ml). The cells were incubated at 37uC for 4 hours. After this period, cells were washed once in serum-free HMI9 medium and fixed by immersion in chilled 2.5% gluteraldhyde (Sigma-Aldrich) in phosphate buffered saline for 5 minutes. The cells were washed once more in chilled PBS and then re-suspended in PBS with an AlexaFluorH488 His-Tag antibody (Molecular Probes, Invitrogen). The cells were gently agitated for 1 hour and then washed twice with chilled PBS and spread onto slides as described above.

Real-time PCR
Total trypanosome RNA was extracted from approximately 5610 6 cells using a Qiagen RNAeasyH mini kit following the manufacturer's instructions. The RNA was subjected to three DNase I (Invitrogen) digests to remove all genomic contamination. OmniscriptH RT Kit (Qiagen) was used to generate cDNA from 1 mg of total RNA as per the manufacturer's instructions. Realtime-PCR was performed using cDNA from an equivalent of 50 ng of total RNA, 3 mM forward and reverse primers (HpHbR_RT_F & HpHbR_RT_R, Table 1) and 12.5 ml of SYBR green PCR master mix (Applied Biosystems) to a final volume of 25 ml. Real-time PCR conditions were: one cycle of 50uC for 2 mins, 95uC for 10 mins, followed by 40 cycles of 95uC for 15 s, 60uC for 1 min. The relative amounts of specific cDNA between samples were calculated using C T methodology [40] calculated with the Applied Biosystems SDS v1.4 software. The endogenous control gene was GPI8 using primers GPI8F and GPI8R primers (Table 1). Four biological replicates were performed for each parasite line. All primers were designed using the Primer3 software [41]. Mean levels of expression were compared using 1-way ANOVA and Tukey's post hoc test.

Sequencing and PCR
To test for the presence/absence of SRA and T. b. gambiense specific glycoprotein (TgSGP) [42], primer sets SRA-F with SRA-R and TgSGP-F with TgSGP-R [20,37,43] were used to amplify from prepared genomic DNA. The SRA gene was amplified using the SRA primer set under the following conditions with Taq polymerase for 30 cycles; 95uC for 50 seconds, 55uC for 50 seconds and 65uC for 60 seconds while the TgSGP primers were amplified with the conditions for 30 cycles; 95uC for

Human serum resistance phenotypes of different subspecies
To investigate human serum resistance/sensitivity in the two different groups of T. b. gambiense STIB386 (group 2 T. b. gambiense) and ELIANE (group 1 T. b. gambiense) were compared with each other and with representative isolates of each of the other subspecies; STIB247 (T. b. brucei) and Baganzi (T. b. rhodesiense). Strains were cultured and examined for the presence/absence of the subspecies specific genes SRA (T. b. rhodesiense specific [46]) and TgSGP (group 1 T. b. gambiense specific [37]) by PCR amplification in order to verify their sub-species classification (Figure 1). The SRA gene was found to be present in the T. b. rhodesiense Baganzi strain and the TgSGP gene was present in the group 1 T. b. gambiense ELIANE strain, confirming their sub-species status. The strains T. b. brucei STIB247 and group 2 T. b. gambiense STIB386 do not possess either gene (Figure 1). The human serum resistance phenotype for the T. b. brucei and T. b. gambiense strains were established by monitoring survival in the presence of human serum over 6 hours ( Figure 2). The group 1 T. b. gambiense strain ELIANE was always resistant to human serum with 100% survival and the T. b. brucei STIB247 was always sensitive with a mean survival of less than 25% (Figure 2). These results confirm the sub-species classification of each of the strains used in the subsequent experiments, (as previously described for STIB247, ELIANE and STIB386 [46]). The marked variation in human serum resistance phenotype previously described in T. b. rhodesiense is also seen in group 2 T. b. gambiense. By culturing the strain T. b. gambiense STIB386 in the presence of 10% human serum in an in vitro system we selected a population of human serum resistant parasites (STIB386R; Figure 2). Several clonal populations of the original STIB386 clone gave a range of human serum resistance phenotypes of which one sensitive line was chosen that was consistently sensitive to human serum in continuous culture without human serum (STIB386S; Figure 2). In order to demonstrate that the lines were indeed isogenic, a series of four microsatellite markers that are used routinely to genotype isolates, Tb5/4, Tb2/20, Tb2/10 and Tb3/1, were amplified using the conditions previously described [47], giving identical genotypes (unpublished). Thus isogenic lines of sensitive and resistant parasites were available for comparison.

Bloodstream expression site analysis
Previous studies indicate that the switching of resistance in T. b. rhodesiense strains correlated with a change in the variable antigen type (VAT) [23,24,48,49]. In order to determine if a change in VSG expression site (ES) has occurred, the ESAG6 and ESAG7 genes are routinely used as ES markers as they contain polymorphisms that are associated with each expression site [45,50,51,52]. However, while these markers are useful for T. b. brucei it is possible that the group 2 T. b. gambiense strain used in our study has a reduced repertoire of ESAG6 and ESAG7 variants compared to T. b. brucei, in a similar manner to group 1 T. b. gambiense, which has been shown to possess a smaller repertoire of ESAG6 and ESAG7 than other T. brucei sub-species [53]. The ESAGs of group 2 T. b. gambiense have not been studied in detail, although an estimate of the variation of ESAG6 and ESAG7 can be determined from the available genome sequence for this strain (Berriman, unpublished). The genome sequence for the other subspecies were also available for analysis: the T. b. brucei strains STIB247 (Berriman, unpublished), the genome reference TREU927 [54] and the group 1 T. b. gambiense strain DAL972 [55]. Sequence reads were aligned to the published variants of ESAG6 and ESAG7 [53,54,56,57,58]. Sequence reads from STIB247, TREU927 and STIB386 aligned to 21, 22 and 22 unique ESAG6 and 21, 22 and 21 unique ESAG7 sequences, respectively ( Figure S1). However, the group 1 T. b. gambiense strain DAL972 [55] sequencing analysis only showed aligned reads for eight unique ESAG6 sequences and nine unique ESAG7 sequences. This parallels previously published data revealing that group 1 T. b. gambiense have a less variable ESAG complement than T. b. brucei [53]. Alternatively, it may be the case that group 1 T. b. gambiense has an ESAG complement containing variants that differ from the T. brucei ESAG6 and ESAG7 sequences previously published. Whichever hypothesis is correct, both suggest that the group 2 T. b. gambiense strain STIB386 possess a T. b. bruceilike rather than a group 1-like ESAG repertoire. This is consistent with other studies suggesting that group 2 T. b. gambiense are more similar to T. b. brucei than group 1 T. b. gambiense [59,60] including the possession of metacyclic ESAGs found in T. b. brucei and T. b. rhodesiense, but not group 1 T. b. gambiense [57]. These results suggest that ESAG6 and ESAG7 variants are suitable ES markers.
In order to determine if a switch in ES is associated with the human serum resistance phenotype in STIB386, we examined the ES used in both sensitive and resistant lines using ESAG6 and ESAG7 as ES markers. RNA from both sensitive and resistant isogenic lines was extracted and used to generated cDNA. The hyper-variable regions of ESAG6 and ESAG7 genes were then amplified by PCR and sequenced. No sequence differences were detected between the hyper-variable regions of ESAG6 and ESAG7 in T. b. gambiense STIB386S and STIB386R lines. Multiple clones from the PCR reactions were sequenced confirming that the dominant ESAGs from the active ES were detected ( Figure S2). These results are consistent with the view that both isogenic lines are using the same dominant ES, suggesting that human serum resistance in group 2 T. b. gambiense is not associated with ES switching.

HpHbR expression
Recent evidence indicates that T. b. gambiense group 1 parasites do not take up TLF-1 due to reduced expression and function of HpHbR [36]. A reduction of expression of this gene in T. b. brucei by RNAi or gene knockout confers TLF-1 resistance [13,36]. Do group 2 T. b. gambiense parasites employ a similar mechanism? To address this question, the expression of the HpHbR gene was examined in the different lines by measuring the relative amounts of transcript for the gene using quantitative real-time PCR. The T. b. brucei line (STIB247), isogenic lines of T. b. gambiense group 2 (STIB386S and STIB386R) and T. b. rhodesiense (Baganzi) strains all expressed similar levels of transcript for the HpHbR gene relative to a GPI8 endogenous control ( Figure 3). Expression levels of the GPI8 control did not show much variation across the different isolates. The STIB386R strain showed a slight decrease in mean level of expression of the gene compared to STIB386S but this was not statistically significant using a one-way ANOVA (F 7 = 4.03, p = 0.091). However, the group 1 T. b. gambiense strain expressed a five-fold lower mean transcript level compared to the other parasite lines and comparing expression across all five strains identified a significant difference (F 18 = 36.51, p,0.01). Post-hoc Tukey's tests indicated the group 1 T. b. gambiense strain ELIANE to be different to all other lines whilst all comparisons between the other four lines were not significant.

TLF-1 uptake and localization
In order to determine if group 2 T. b. gambiense parasites have a functional HpHbR and so take up TLF-1 and traffic the complex to the lysosome in a similar manner to T. b. brucei, the uptake of labeled TLF-1 was examined in STIB386S and STIB386R lines in relation to the other parasite strains. TLF-1 was tagged with alexafluorH488 and its uptake analyzed by microscopy in conjunction with the commercial dye LysotrackerH, which labeled acidic vesicles including the lysosome. As previously reported, no labeled TLF-1 was detected in the group 1 T. b. gambiense strain parasites indicating a lack of TLF-1 uptake [36]. However the T. b. brucei strain STIB247, T. b. rhodesiense strain, Baganzi, and both the isogenic group 2 T. b. gambiense lines STIB386S and STIB386R lines all showed internalization of TLF-1 within one hour and colocalization of the TLF-1 with acidic vesicles (Figure 4). The mean correlation co-efficients (R) estimating the degree of correlation between the position of TLF-1 and LysotrackerH were: STIB247, R = 0.9 (n = 14); STIB386S, R = 0.68 (n = 18); STIB386R, R = 0.63 (n = 19); Baganzi, R = 0.72 (n = 28). This indicates that the HpHbR protein of group 2 is functional and that reducing uptake of TLF-1 is not a potential resistance mechanism for group 2 T. b. gambiense strain STIB386 in contrast to group 1 T. b. gambiense, which has been shown to not only have reduced expression of the gene but also a reduced function of the protein due to several amino acid substitutions within the open reading frame (ORF) [36]. Sequence analysis of the ORF of the HpHbR gene revealed several differences in sequence ( Figure S3). In order to investigate if there was a difference in the rate of uptake of TLF-1 between the different strains, TLF-1 uptake was examined over a four-hour time course and compared to the uptake of bovine HDL. The STIB247, Baganzi and STIB386S lines all took up TLF-1 rapidly whereas STIB386R took up TLF-1 at a slower rate that reached an asymptote at approximately 60% ( Figure S4). In contrast, type 1 T. b. gambiense showed no visible uptake of TLF-1 during the four-hour exposure.

Recombinant APOL1 activity, uptake and localization
The knowledge that group 1 T. b. gambiense parasites have evolved to avoid TLF-1 uptake raises the possibility that this sub-species group's resistance strategy is based on avoidance and it is inherently susceptible to APOL1 mediated lysis. However, this fact is difficult to reconcile with what is known about TLF-1 and TLF-2 uptake in these parasites. TLF-1 enters trypanosomes not only via the HpHbR but also via a low affinity receptor that is as yet unidentified; therefore a reduction in two TLF-1 receptors is required for protection from TLF-1. In addition, TLF-2 is internalized in a different manner to TLF-1 that does not involve HPR binding [14]. Unfortunately we were not able to work Figure 2. Human serum resistance phenotype. The survival of parasites exposed to 25% lytic human serum for 6 hours as a percentage of the starting cell count. Each value is the mean result from a separate experiment (for ELIANE, n = 4). The two isolate populations that were selected to create T. b. gambiense STIB386S (circled in blue) and T. b. gambiense STIB386R (circled in red) are indicated. doi:10.1371/journal.pntd.0001287.g002 Differences T.b.gambiense Groups 1&2 HS Resistance www.plosntds.org  directly with active TLF-2 but it is possible to assay the different strains for susceptibility to APOL1 using recombinant protein.
Recombinant APOL1 has been shown in numerous experiments to be toxic to T. b. brucei and is internalized via fluid phase endocytosis [21,61,62]. Therefore, uptake of this recombinant protein would be unavoidable. The capacity for recombinant APOL1 to lyse group 1 T. b. gambiense, in addition to both sensitive and resistant forms of group 2 T. b. gambiense and T. b. brucei, was examined for several concentrations of protein. Recombinant APOL1 is able to lyse T. b. brucei and sensitive group 2 T. b. gambiense fully after 24 hours. Lysis occurs at even low concentrations of protein, which approach physiological levels [63]. The resistant group 2 T. b. gambiense strain and the group 1 T. b. gambiense strain were unaffected by APOL1, even at high concentrations ( Figure 5 and Figure S5). This indicates that the difference between sensitive and resistant forms of group 2 T. b. gambiense is not due to differences in uptake but a difference in sensitivity to APOL1.
Although in other T. brucei lines recombinant APOL1 was internalized and trafficked to the acidic compartments [21,61,62], it is possible that the group 1 T. b. gambiense and resistant group 2 T. b. gambiense are able redirect APOL1 away from these organelles and avoid APOL1 activation. It is also possible that the recombinant protein is not being trafficked to the acidic compartments so would not reach the site of action. To investigate this, the internalization and location of recombinant APOL1 was visualized in several T. brucei lines using fluorescence microscopy ( Figure 6). All sub-species showed visible uptake of recombinant APOL1 within four hours. The position of highest fluorescence correlates with the lysotracker dye, including the group 1 T. b. gambiense strain ELIANE and the resistant group 2 T. b. gambiense strain STIB386R. The mean correlation coefficient (R) of the recombinant APOL1 and LysotrackerH field was estimated for each parasite line: STIB247, R = 0.87 (n = 16); STIB386S, R = 0.81 (n = 13); STIB386R, R = 0.89 (n = 15); Baganzi, R = 0.76 (n = 15); ELIANE, R = 0.85 (n = 18). This would indicate that both group 1 T. b gambiense and resistant group 2 T. b. gambiense possess the ability to resist the lytic effects of APOL1 in the lysosome. These results are in stark contrast to data suggesting TLF-1 shows reduced uptake in group 1 T. b. gambiense cells over short-term due to reduced expression and activity of HpHbR.
In summary, T. b. gambiense group 2 parasites appear to express the HpHbR gene and have a functional protein that facilitates the uptake of TLF-1, unlike that of T. b. gambiense group 1. These results indicate that important differences in the cell biology of human serum resistance exist between the T. b. gambiense group 2 parasites and T. b. gambiense group 1.

Discussion
The group 2 T. b. gambiense strain STIB386 exhibits a variable phenotype as previously described [31,64], but continuous exposure to a low concentration of normal human serum maintains the resistant phenotype indefinitely allowing comparisons between isogenic sensitive and resistant forms. Analysis of these parasites indicates that both sensitive and resistant parasites internalize TLF-1. As resistance to lysis by normal human serum and TLF-1 correlates with resistance to recombinant APOL1, it would appear that the resistance mechanism of group 2 T. b. gambiense is able to either neutralize or compensate for the effects of the protein. This allows it to overcome both TLF-1 and TLF-2 particles. This is in contrast to group 1 T. b. gambiense parasites that avoid uptake of the TLF-1 particle. However the innate resistance of both group 1 and group 2 T. b. gambiense to APOL1 might be based on the same mechanism. This resistance is distinct from the SRA gene mediated mechanism in T. b. rhodesiense, although the mechanism could still involve an inhibitory protein similar to SRA.
Previous attempts to characterize human serum resistance/ sensitivity in group 2 T. b. gambiense have relied on examination of the TxTat strain [64]. However, some TxTat lines have subsequently been shown to be T. b. rhodesiense [65]. Examination of the human serum resistance phenotype in TxTat suggested that a switch between resistant and sensitive forms was always accompanied by a switch in variable antigen type (VAT), indicating that the resistance mechanism is closely related to antigenic variation [49,64], consistent with these isolates being T. b. rhodesiense. Our sequencing data show that identical ESAG6 and ESAG7 gene copies are expressed in both strains of STIB386 suggesting that, irrespective of VSG expression, the same dominant ES is being used in both sensitive and resistant populations and that resistance may be unrelated to the use of a particular ES, as it is in T. b. rhodesiense. While it is intuitive to assume any variability in expression in trypanosomes is due to expression site switching, expression stochasticity is a common feature to many eukaryotes and can arise from several different mechanisms [66].
As previously reported, the group 1 T. b. gambiense strain ELIANE does not appear to internalize and concentrate TLF-1 to any discernable degree over the short term and would appear to employ an avoidance of uptake strategy to prevent lysis by TLF-1. This can be explained by both substantial down-regulation and/or loss of function in the HpHbR involved in internalizing TLF-1 Differences T.b.gambiense Groups 1&2 HS Resistance www.plosntds.org [36]. Alternately, TLF-1 may be differentially trafficked by ELIANE to RAB-11 compartments rather than the lysosome, rapidly removing the toxic particles from the cell and so would not be detected in this assay. There are two group 1 T. b. gambiense specific polymorphisms in the 39 UTR of HpHbR that could potentially contribute to mRNA instability and reduced expression of HpHbR to affect short-term uptake of TLF-1. There are also several polymorphisms in the ORF of HpHbR that appear to affect the function of HpHbR in group 1 T. b. gambiense [36]. This has been demonstrated by ectopically expressing the group 1 T. b. gambiense version of the gene into an artificially selected L427 HpHbR null mutant, resulting in a failure to restore the full TLF-1 sensitivity phenotype [36].
Interestingly, despite several strains exhibiting a loss of function and expression in the HpHbR receptor [36], group 1 T. b. gambiense appear to be fully resistant to the lytic effects of recombinant APOL1. Several hypotheses can be formulated as to why group 1 T. b. gambiense avoid TLF-1 despite its inherent resistance to APOL1; these include the hypothesis that the TLF-1 particle evolved before TLF-2 in primates. T. b. gambiense may initially have evolved primate infectivity by simply avoiding the TLF-1 particle by modifying receptors for HP/HPR. The sub-species was then unable to use a similar mechanism to avoid uptake of TLF-2 as this particle is not internalized in the same manner [14]. Instead, a second resistance mechanism to counteract APOL1 evolved. Another hypothesis is that the T. b. gambiense resistance mechanism is dose dependent and by avoiding TLF-1 uptake is able to avoid the majority of the APOL1 in the human bloodstream. This would suggest that group 1 T. b. gambiense may be susceptible to APOL1 lysis if internalized at high concentrations.
The results presented here demonstrate some remarkable features of the variation in human serum resistance mechanisms that have evolved in trypanosome strains. In Eastern Africa, trypanosomes have evolved the SRA protein to neutralize APOL1 and resist lysis by TLF. However, this is not the entire story and some human infective trypanosome strains in these T. b. rhodesiense foci do not possess the SRA mechanism [23,67]. The resistance mechanism for these isolates is unknown. While it has been suggested that avoidance of TLF-1 is a resistance strategy in the most prevalent human infective trypanosome, group 1 T. b. gambiense [36], this mechanism is not exhibited by the related and sympatric group 2 T. b. gambiense suggesting a resistance strategy that is different. It is now clear that human serum resistance in both groups of T. b. gambiense involves the ability to either neutralize or resist the effects of APOL1, although resistance is variably expressed in group 2. While it may be that groups 1 and 2 T. b. gambiense share a resistance mechanism, population studies suggest that the two groups are not closely related [26,27,28,29,30,31,32,33,35,59,68,69,70]. Taking all of the evidence into account, it would appear that human infectivity has likely evolved in the field on at least four occasions, suggesting there are multiple ways in which these parasites can avoid innate immunity. These facts, coupled with the relative ease with which TLF resistant parasites can be selected from T. b. brucei parasites [71] indicates that novel human infective parasites could evolve from the T. brucei population. Understanding the processes involved in overcoming these factors will aid in the development of novel intervention strategies.