β-Tubulin Genes from the Parasitic NematodeHaemonchus contortusModulate Drug Resistance inCaenorhabditis elegans

JMB—MS 358 Cust. Ref. No. CAM 427/94

[SGML] J. Mol. Biol. (1995) 246, 500–510

b-Tubulin Genes from the Parasitic Nematode Haemonchus contortus Modulate Drug Resistance in Caenorhabditis elegans Marcel S. G. Kwa, Jetty G. Veenstra, Marjon Van Dijk and Marleen H. Roos* Department of Parasitology and Tropical Veterinary Medicine, Institute of Infectious Diseases and Immunology, Faculty of Veterinary Medicine University of Utrecht P.O. Box, 80.165, 3508 TD Utrecht, The Netherlands

*Corresponding author

Resistance to antimitotic chemotherapeutics in pathogenic nematodes, fungi and mammalian cells is closely associated with structural changes in cytoskeletal b-tubulin. We investigated the possibility of using the well-characterised free-living nematode Caenorhabditis elegans as a model for studying the mechanism of resistance against benzimidazole (BZ) drugs in the parasitic nematode Haemonchus contortus. Functional analysis of a conserved b-tubulin isotype (tub-1) mutation near GTP-binding domain II, which is linked to BZ resistance, was carried out in C. elegans by heterologous expression of: (1) parasite BZ-sensitive alleles; (2) BZ-resistant alleles; and (3) in vitro mutagenised b-tubulin gene constructs. The injected heterologous gene constructs were not only stably maintained, but also expressed as shown by reverse transcriptase-polymerase chain reaction analysis. The degree of BZ drug susceptibility of the transformants was assayed and quantified by incubation with both benomyl and thiabendazol. All H. contortus tub-1 constructs, which encoded Phe at position 200, conferred susceptibility to thiabendazole in BZ-resistant C. elegans ben-1 mutants. In contrast, constructs carrying Tyr200 did not alter the BZ drug phenotype. From these experiments we conclude that: (1) C. elegans can be used as an expression host, since injected parasite genes were biologically active; and (2) the single Phe to Tyr mutation at position 200 in b-tubulin isotype 1 is the cause of BZ resistance in H. contortus. Keywords: b-tubulin; Caenorhabditis elegans; drug resistance; Haemonchus contortus; parasitic nematode

Introduction Infections caused by parasitic nematodes account for a significant portion of human and animal diseases (Bird, 1991; Jackson, 1993). Treatment of these infections relies mainly on a limited number of chemotherapeutics, since development of new drugs is slow and practical vaccines are not readily available. Large-scale use and improper dosing have led to parasite resistance against benzimidazole (BZ) and levamisole (LEV) drugs (Jackson, 1993; Lacey, Present address: M. H. Roos, Department of Molecular Biology, Institute for Animal Science and Health (ID-DLO), P.O. Box 65, 8200 AB Lelystad, The Netherlands. Abbreviations used: BZ, benzimidazole; BZ-S, BZ-R, BZ-susceptible, BZ-resistant; LEV, levamisole; RT-PCR, reverse transcriptase-polymerase chain reaction; TBZ, thiabendazole. 0022–2836/95/090500–11 $08.00/0

1988). The use of ivermectin is at present widespread. However, the first signs of decreased efficacy in treating nematode infections have been reported (Shoop, 1993). Understanding the mechanisms and dynamics of the development of resistance, on both molecular and population genetic levels, is needed in order to develop adequate control programs. Initial studies on the mechanism of BZ resistance in Haemonchus contortus, a parasite of sheep, have revealed that resistance is correlated with a selection for individuals that possess a specific b-tubulin allele (Roos, 1990; Lubega et al., 1994). In vitro selection studies with increasing drug concentrations showed that selection for resistant H. contortus individuals occurred within only a few generations, resulting in a population homogeneous for a particular b-tubulin isotype 1 allele (Kwa et al., 1993a; Roos et al., 1990). Sequence analysis of this b-tubulin locus, designated b-tubulin isotype 1, in both BZ-susceptible (BZ-S) 7 1995 Academic Press Limited

JMB—MS 358 Analysis of H. contortus b-Tubulin in C. elegans

and BZ-resistant (BZ-R) H. contortus populations showed three amino acid differences at positions 76, 200 and 368 (Geary et al., 1991; Kwa et al., 1993b). The Phe200 (BZ-S) to Tyr200 (BZ-R) change is of importance, since it is strictly conserved between different BZ-S and BZ-R H. contortus populations (Kwa et al., 1994). Moreover, the same mutation in codon 200 is one of the changes discriminating between BZ-S and BZ-R b-tubulins in Caenorhabditis elegans (Driscoll et al., 1989) as well as several species of fungi (Jung et al., 1992; Koenraadt et al., 1992). However, so far no functional analyses of the change at codon 200 have been performed. In order to prove definitely the functional importance of the amino acid 200 mutation in H. contortus, in vitro mutagenesis and expression studies have to be carried out. Unfortunately, these genetic studies can, at present, not be performed in the parasite itself, due to the lack of inbred strains and in vitro culturing techniques. An alternative choice is the well-characterised free-living nematode C. elegans. The advantages of C. elegans, as compared to other expression hosts such as Escherichia coli and yeast, are that: (1) it is a higher eukaryote that is closely related to H. contortus (Wood, 1988); (2) C. elegans performs trans-splicing using the same spliced leader (SL1) as H. contortus (Bektesh et al., 1988; Nilsen, 1993); (3) several b-tubulin mutant strains exist (Driscoll et al., 1989; Savage et al., 1989); and (4) gene expression and function in cell and tissue development and differentiation

501 can be studied through transformation by microinjection (Fire & Waterston, 1989; Mello et al., 1991). The derived amino acid sequences (amino acid residues 1 to 430) of three b-tubulin genes present in C. elegans show, apart from the last ten C-terminal amino acid residues, 90 to 98% identity with b-tubulins from parasites like H. contortus (Figure 1). Driscoll et al. (1989) have shown that BZ-R C. elegans strains, obtained by gamma irradiation or ethyl methane sulfonate induced mutagenesis, can result from deletion of the b-tubulin locus ben-1. Loss of ben-1 activity, which leads to BZ resistance, produces no other mutant phenotype, a result suggesting that it is redundant for all other b-tubulin functions. These BZ-R ben-1 mutant strains offered the opportunity to characterise the functional importance of mutations in the b-tubulin genes of H. contortus. Several authors have indicated the use of the free-living nematode C. elegans as a model for studying the biology of the cuticle and surface antigens of parasitic nematodes (Blaxter et al., 1992; Politz & Phillip, 1992). However, thus far only preliminary data have been reported employing transgenic C. elegans to analyse parasitic nematode genes (Grant, 1992). In the present study we show that C. elegans can indeed be used as an expression host for parasitic nematode genes and that this approach could be used successfully for elucidation of the mechanism of the development of BZ resistance in H. contortus.

Figure 1. Comparison of nematode b-tubulin derived amino acid sequences. H. contortus (H.c.) tub-1(RU) and tub-1(iSE) were deduced from genomic DNA sequences (Kwa et al., 1993b). The amino acid sequence of tub-1(8-9) and tub-2(12-16) originated from cDNA clones (Geary et al., 1991); ben-1, tub-1 and mec-7 were derived from 3 b-tubulin loci from C. elegans (C.e.) (Driscoll et al., 1989; Savage et al., 1989). Nucleotide binding domains (I to IV) are overlined.

JMB—MS 358 502

Results Sequence analysis In order to carry out transformational rescue studies with parasite genes, a complete H. contortus BZ-S b-tubulin isotype 1 gene was required. Identification of the proper BZ-S gene(s) was complicated by the fact that BZ-S H. contortus field populations are highly polymorphic, containing multiple BZ-S alleles as well as a small portion of BZ-R alleles (Kwa et al., 1993a,b; Roos et al., 1990). To decrease the probability of isolating BZ-R b-tubulin alleles, we constructed a plasmid DNA library from DNA from a partially (81.5%) inbred BZ-S H. contortus strain. We isolated a clone (pISE3) containing a 5.5 kb NsiI-SalI restriction fragment carrying a complete H. contortus b-tubulin isotype 1 gene. The nucleotide sequence (data not shown but submitted under numbers X80046, X80047, X80048 and X80049 to the EMBL data base) and the derived amino acid sequence of the gene positioned in the 5.5 kb NsiI-SalI fragment were determined. The gene was designated tub-1(iSE). The amino acid differences between tub-1(iSE) and several other nematode b-tubulins are shown in Figure 1. The nucleotide and deduced amino acid sequences of tub-1(iSE) and tub-1(RU) (Kwa et al., 1993b) were identical, with only one exception at codon 200. Here, Phe in tub-1(iSE) was found that was apparently replaced by Tyr in the previously sequenced tub-1(RU) gene, exactly as the case in resistant fungi (Koenraadt et al., 1992). We therefore conclude that also in H. contortus only one mutation at amino acid 200 could be the cause of benzimidazole resistance. This mutation was then investigated further. In addition, both genes probably represent two alleles of the same b-tubulin isotype 1 locus, although we have no formal proof for this contention (i.e. genetic linkage data). Transformational rescue of the ben-1 mutation Driscoll et al. (1989) demonstrated that gene ben-1 confers benomyl sensitivity in C. elegans, but they did not attempt transformational rescue studies of BZ-R ben-1 mutants. Therefore, before expressing any heterologous parasitic nematode genes, we needed to demonstrate such rescue. The ben-1 mutant strain TU1054, carrying the mutation ben-1(u462) that disrupts the ben-1 gene, was transformed simultaneously with plasmid pTU59, which contains the complete ben-1 gene, and with the selectable marker plasmid pRF4. The latter plasmid has a dominant rol-6 allele, which allows the selection of transformants, since these display the roller phenotype (Mello et al., 1991). This transformation resulted in one transgenic line. Presence of the ben-1 plasmid was verified by PCR using primers ben-1 89 and ben-1 92. In addition, RT-PCR analysis using primers SL1 and ben-1 92 confirmed that the injected ben-1 gene was transcribed and apparently

Analysis of H. contortus b-Tubulin in C. elegans

trans-spliced (data not shown). The drug phenotype of this transformant was subsequently tested by a two day incubation at 25°C on NGM agar plates containing 2.9 pg/ml benomyl. The results are shown in Table 1A. First, the host strain injected solely with the rol-6 marker gene showed no alteration in drug phenotype, indicating that the selectable marker did not have any detectable effect in the drug assays. Second, growth and movement were normal in wild-type C. elegans N2 (Bristol) grown on normal NGM agar plates. However, the addition of the benzimidazole drug benomyl in NGM plates resulted in paralysis of the wild-type strain. Third, the host strain (TU1054) without the injected ben-1 gene was, in contrast, not effected by the incubation on benomyl plates. Strain TU1054 injected with the intact ben-1 gene yielded animals that were paralysed in the presence of benomyl (Table 1A). Analysis of individual nematodes showed that all animals that were benomyl-S carried the injected ben-1 gene, whether exhibiting the roller phenotype or not. All benomyl-R individuals did not carry the ben-1 gene, whether showing the roller phenotype or not. From these data we concluded that rescue of the ben-1 mutation does indeed occur. We next tested whether a parasite b-tubulin could function in C. elegans. We used the same transformation procedure as described above but with plasmid pISE3. This plasmid contains the BZ-S H. contortus b-tubulin isotype 1 gene tub-1(iSE). Twelve TU1054 animals were injected with plasmid pISE3, and three independent stable clones (III 1, III 3, III 6) were obtained. The drug phenotype of these transformants was determined as described above. In contrast to the transgenic animals transformed with the homologous C. elegans ben-1 gene, we did not observe an alteration to benomyl sensitivity in the animals transformed with the heterologous tub1(iSE) gene (Table 1A). Since there are no reports on the activity of benomyl in parasitic nematodes, we also investigated the effect of a different benzimidazole drug, thiabendazole (TBZ), which has been used frequently against parasitic nematodes in vivo, as well as for in vitro selection for BZ resistance in H. contortus (Roos et al., 1990; Kwa et al., 1993a). The transgenic animals showed an increase in the proportion of F1 eggs that remained unhatched on the TBZ drug plates relative to the control plates. This difference, expressed as percentage of hatching, enabled us to develop a quantitative test for determining the degree of sensitivity to TBZ. Optimisation of drug concentrations and incubation conditions showed that after three days incubation at 25°C the most pronounced difference in hatching between wild-type and BZ-R TU1054 C. elegans strains was observed. After standardisation of the assay, we determined the ED50 value, that is the TBZ concentration at which 50% of the eggs fail to hatch. The values determined for the control strains wild-type N2 and TU1054, rol-6 transgenes, as well as the animals transformed with tub-1(iSE), are given

JMB—MS 358 Analysis of H. contortus b-Tubulin in C. elegans

503

Table 1 Transformation of C. elegans with b-tubulin genes from C. elegans and H. contortus Host strain

BZstatus

Introduced gene(s) or construct

A. Transformation with cloned BZ-S and BZ-R beta-tubulin genes Wild-type N2 S — ben-1-TU1054 R — ben-1-TU1054 R C.e. rol-6 ben-1-TU1054 R C.e. ben-1 + rol-6 ben-1-TU1054 R Clone III 1 H.c. tub-1(iSE) + rol-6 Clone III 3 H.c. tub-1(iSE) + rol-6 Clone III 6 H.c. tub-1(iSE) + rol-6 Wild-type N2 S H.c. tub-1(RU) + rol-6 ben-1-TU1054 R H.c. tub-1(RU) + rol-6

NGM

Benomyl (2.9 pg/ml)

Normal Normal Normal Normal

Paralysed Normal Normal Paralysed

15.7 2 4.4 69.7 2 12 57.7 2 7.8 24.8 2 8.7

Normal Normal Normal Normal Normal

Normal Normal Normal Paralysed Normal

26.5 2 9.0 29.7 2 7.0 20.6 2 3.2 19.0 2 5.1 64.0 2 14

Normal Normal Normal

56.6 2 8.2 58.5 2 12 71.8 2 6.1

Normal Normal Normal Normal Normal Normal

25.4 2 1.8 36.2 2 6.0 22.8 2 3.7 25.8 2 8.3 23.2 2 2.7 29.6 2 7.5

Normal Normal Normal

16.6 2 6.3 29.9 2 3.6 31.9 2 4.2

B. Transformation with in vitro mutagenized b-tubulin genes at amino acid residue 200 ben-1-TU1054 R Clone I 5 H.c. pISE200Y* + rol-6 Normal Clone II 2 H.c. pISE200Y* + rol-6 Normal Clone II 3 H.c. pISE200Y* + rol-6 Normal ben-1-TU1054 R Clone I 1 H.c. pRU200F* + rol-6 Normal Clone III 11 H.c. pRU200F* + rol-6 Normal Clone IV 11 H.c. pRU200F* + rol-6 Normal Clone VIII 5 H.c. pRU200F* + rol-6 Normal Clone IX 1 H.c. pRU200F* + rol-6 Normal Clone X 1 H.c. pRU200F* + rol-6 Normal C. Transformation with in vitro mutagenized b-tubulin genes at the carboxy terminus ben-1-TU1054 R Clone II 4 H.c./C.e. pISE + benCOOH + rol-6 Normal Clone IV 6 H.c./C.e. pISE + benCOOH + rol-6 Normal Clone VI 2‡ H.c./C.e. pISE + benCOOH + rol-6 Normal

ED50 (mg TBZ/ml) mean 2 SD†

† n = 5. ‡ Integrated.

in the last column of Table 1A. Wild-type C. elegans showed a mean ED50 value of 15.7 mg TBZ/ml. The ED50 value of 69.7 mg TBZ/ml of the BZ-R strain TU1054 was significantly higher (p = 0.001). Transformation with the rol-6 selectable marker alone did not alter the TBZ sensitivity (compared to TU1054, p = 0.161), analogous to the results obtained with the assays with benomyl. Introduction of the ben-1 gene resulted in an almost threefold, significant (p = 0.001), drop of the ED50 (24.8 mg TBZ/ml) when compared with the BZ-R host strain TU1054 (69.7 mg TBZ/ml). However, the mean value of the rescued mutant (ED50 , 24.8) was somewhat higher than wild-type strain (ED50 , 15.7). This was probably caused by the fact that, although transformed genes are stably maintained, always 15 to 30% of the offspring of rol-6 selected parents lose the transformed genes (Mello et al., 1991). Taken together, these data show convincingly that the egg-hatch assay can be used to quantitatively determine the degree of drug sensitivity or resistance in the C. elegans strains and transformants. We used this test to detect the TBZ sensitivity of clones III 1, III 3 and III 6 transformed with the parasite tub-1(iSE) b-tubulin gene. The mean ED50 values were, respectively, 26.5, 29.7 and 20.6 mg TBZ/ml and were similar to the value obtained with the ben-1 transformant. This strongly indicates that

the parasite tub-1(iSE) gene is also able to confer TBZ-sensitivity in C. elegans, and can apparently rescue the ben-1 mutation with regard to TBZ, but not to benomyl. To exclude the possibility that the incomplete rescue with regard to benomyl was caused by the ben-1 mutant strain that we used (TU1054 ben1(u462)), we repeated the tub-1(iSE) transformations with two different ben-1 mutant strains. In total 21 animals of both strain TU1055 ben-1(u463) and CB474 ben-1(e1880) were injected with pISE3, resulting in eight and four stable transformed lines, respectively. None of the tub-1(iSE) transformed ben-1 mutants TU1055, CB474 (not shown) and TU1054 differed significantly in either their benomyl nor their TBZ sensitivity. The absence of sensitivity to benomyl in the transformants thus appeared to be independent of the host strain used. Finally, we transformed the H. contortus BZ-R gene tub-1(RU) to both BZ-S wild-type N2 and BZ-R TU1054 C. elegans strains. Results from the drug tests (Table 1) show that the transformation with plasmid pRU1 carrying the parasite tub-1(RU) gene did not have any effect on the phenotype of the host strains: wild-type N2 remained sensitive to both benomyl and TBZ (ED50 of 19.0 mg TBZ/ml), whereas TU1054 animals remained resistant (ED50 of 64.0 mg TBZ/ml).

JMB—MS 358 504 Expression of the heterologous tub-1(iSE) gene The presence of the heterologous transcript from the injected plasmid pISE3 was investigated by RT-PCR analysis. Total RNA was reverse transcribed with the H. contortus tub-1 primer b4 (Figure 2A). Detection of the entire 1.5 kb transcript by amplification with primers b17 and b4 indicated that introns are correctly removed, however several aberrant shorter fragments were also present, probably due to aspecific amplification (data not shown). To decrease the chance of aspecific amplification, only the 5' half of the transcript was amplified with primers b17 and b14 (Figure 2A). Amplification products were characterised by Southern analysis using a H. contortus specific tub-1 cDNA probe. In Figure 2B a representative autoradiograph of the RT-PCR analysis is shown, including two positive and five negative controls. The two positive controls, genomic plasmid clone pRU1, and cDNA plasmid clone pRSA1

Analysis of H. contortus b-Tubulin in C. elegans

yielded a fragment of the expected size of 1.95 kb and 0.5 kb, respectively. As negative controls we included: C. elegans wild-type N2, TU1054, a TU1054 strain transformed with either pTU59 (ben-1) or pRF4 (rol-6), and a blank containing no DNA. These negative controls did not show any amplification product. PCR amplification using cDNA from the animals injected with the H. contortus construct pISE3 (clone III 1) yielded the expected 0.5 kb product, indicating that the transcript was processed correctly. Aside from the correct 0.5 kb product, a faint 2 kb band was also weakly detectable, probably caused by unavoidable contamination with genomic DNA present in the cDNA preparation. We then investigated whether trans-splicing of the heterologous tub-1 gene, which occurs in H. contortus (Kwa et al., 1993b), also takes place in C. elegans. The primer SL1 was used instead of the upstream b17 primer for PCR amplification (Figure 2A), employing the same cDNA preparations as described above. Figure 2C shows that the positive control plasmid

Figure 2. Reverse transcriptase (RT)-PCR characterisation of the 5' termini of the transcripts present in the C. elegans strains and transformants. Correctly processed transcripts yield a 0.5 kb amplification product. Shown are: A, the localisation of the primers used (sizes of the primers are not depicted in scale); B, the amplification products using primers b17 and b14; and C, the amplification products using the primer combination SL1 and b14. Both blots were hybridised with probe b17-14.

JMB—MS 358 Analysis of H. contortus b-Tubulin in C. elegans

505

Figure 3. Gene constructs used for functional analysis of H. contortus b-tubulin genes. tub-1 coding regions are represented by open and black boxes. The 3' region of ben-1, encoding the carboxy terminus (amino acid 393 to 444), is depicted by the shaded box. All constructs were made in a pBluescript II SK+ vector. Only the inserts are shown. Location of the substitution of codon 200 is indicated by the arrow. The scale bar represents 0.5 kb. Restriction enzyme sites: N, NsiI; St, SstI; H, HindIII; X, XhoI; EV, EcoRV; Sl, SalI.

pRSA1 yields a strong signal of 0.5 kb, as expected. PCR amplification of cDNA prepared from the tub-1(iSE) transformant showed a signal of the same size of approximately 0.5 kb, indicating that the trans-spliced leader SL1 was present just upstream of the start codon of the heterologous H. contortus tub-1 transcript. The genomic clone plasmid pRU1, as well as the negative controls, showed no amplification products. Functional analysis of the amino acid 200 mutation The transformation protocols described above, allowed us to identify the b-tubulin regions involved in either BZ sensitivity or resistance. We carried out in vitro mutagenesis of the H. contortus b-tubulin isotype 1 alleles, thereby changing codon 200 of tub-1(iSE) from Phe into Tyr. This yielded the construct pISE200Y* (Figure 3). In total 18 TU1054 animals were co-injected with pISE200Y* and rol-6 (pRF4). Three stable transformed clones were obtained (I 5, II 2 and II 3), of which drug sensitivity to both benomyl and TBZ was determined (Table 1B). In contrast to the transformation with the unchanged H. contortus BZ-S gene tub-1(iSE), we did not observe an alteration in drug phenotype of the host strain. The pISE200Y* transformed animals remained resistant to benomyl as well as to TBZ. ED50 values (56.6, 58.5 and 71.8 mg TBZ/ml) were all significantly higher than the values of the tub-1(iSE) transformed TU1054 animals (p-values ranging from