Genotyping Giardia duodenalis Isolates from Dogs: Lessons from a Multilocus Sequence Typing Study

Abstract

Giardiasis is a common infection of dogs, and the occurrence of both zoonotic and host-adapted assemblages of Giardia duodenalis is well documented in this host. In the current study, G. duodenalis isolates from dogs collected in Croatia from both private owners (n=44) and kennels (n=52) were analyzed at four genetic loci: the ITS1-5.8S-ITS2 (ITS), the glutamate dehydrogenase (gdh), the triosephosphate isomerase (tpi), and the beta-giardin (bg). Both generic and assemblage D specific primers were used for the amplification of the tpi gene. All data were stored in a dedicated database, and analyzed to evaluate (1) the rate of amplification of G. duodenalis DNA from dogs at the four loci; (2) the distribution of assemblages and the occurrence of mixed infections; (3) the genetic variability at the intra-assemblage level; and (4) the zoonotic potential. We found that only half of the isolates could be amplified at either the gdh or the bg gene, whereas the combined use of generic and D-specific tpi primers yielded the highest amplification rate (85%). Sequence analysis showed that assemblages C and D are largely predominant in both kennel and household dogs, thus suggesting a minor role of dogs in zoonotic transmission of giardiasis. However, in many kennel dogs, incongruent results were obtained by using different markers, a result that is more likely explained by mixed infections rather than by genetic recombination. Phylogenetic analysis based on single or multiple loci failed to reveal the presence of distinct subpopulations within assemblages C and D. Our study illustrates the problems associated with the characterization of G. duodenalis isolates from dogs, and it casts doubts on the interpretation of genotyping results based on the analysis of single markers. We concluded that the current typing scheme is not suited to distinguish between recombinants and mixed infections in field isolates.

Introduction

G iardia is a genus of intestinal flagellates that infects a wide range of vertebrate hosts. Among the six species currently recognized in the genus, only Giardia duodenalis (syn. Giardia intestinalis, Giardia lamblia) infects humans as well as domestic and wild mammals (Thompson and Monis 2004).

The molecular characterization of G. duodenalis isolates has revealed the existence of at least seven genetically distinct assemblages in this morphologically invariant species complex (Monis et al. 2009). Among these, assemblages A and B are found in humans and many other mammalian hosts. Although the host ranges are not strict, assemblages C and D infect dogs and wild carnivores, assemblage E is restricted to livestock, assemblage F is restricted to cats, and assemblage G is restricted to rats (Cacciò and Ryan 2008). Recently, another assemblage (H) has been described and proposed to be specifically associated with infection of marine mammals (Lasek-Nesselquist et al. 2010).

A still debated aspect of the complex epidemiology of giardiasis is the extent of zoonotic transmission (Cacciò and Ryan 2008), and a number of species have been considered as potential reservoirs, including livestock, pets, and aquatic animals (e.g., beavers and muskrats). However, and although the World Health Organization has considered G. duodenalis to have a zoonotic potential since over 20 years (WHO 1979), direct evidence for animal-to-human transmission is still scarce (Sprong et al. 2009, Cooper et al. 2010). Due to the limitations of cross-transmission studies, more reliance is placed on the molecular identification of zoonotic assemblages in animals: in particular, pets have been the subject of a number of investigations because of their close contacts with humans and the fact that infection is very common in these animals (recently reviewed by Ballweber et al. 2010). So far, most studies have been mainly based on analysis of the small subunit ribosomal locus (SSU-rRNA), which, however, permits identification only at the level of the assemblage, and, therefore, lacks the resolution necessary for informative comparisons.

In the current study, we present the first multilocus genetic characterization of G. duodenalis from a large number of dog isolates.

Materials and Methods

Source of isolates

A total of 96 dog fecal samples, collected in Croatia over a 5 year period (2005–2009) from both private owners (n=44) and kennels (n=52), were found to be positive for the presence of Giardia spp. cysts by using a commercial kit (Merifluor; Meridian Bioscience). The dog fecal samples from private owners were collected at the Faculty of Veterinary Medicine, University of Zagreb, Croatia. Fecal samples were submitted for routine coprological analysis due to clinical signs of gastrointestinal disturbances, which varied from increased frequency of defecation to watery diarrhea. The second group of samples were collected in four different shelter houses, and only from dogs kept in individual boxes to avoid sample mixture. Positive samples were concentrated by flotation on a sucrose gradient (specific gravity 1.06), and nucleic material in the cysts was visualized by staining with 4′-6-diamino-2-phenylindole.

Molecular methods

DNA was extracted from fecal samples according to a published procedure (da Silva et al. 1999). Briefly, an aliquot of each fecal sample (0.4 mL) was homogenized by using the FP120 Fast Prep Cell disruptor (Savant; Thermo Electro Corporation). The DNA released after the lysis step was purified by using the Fast DNA extraction kit (Qbiogene). Protocols for the amplification of a ∼315 bp fragment encompassing the ITS1-5.8S-ITS2 (ITS) region in the ribosomal unit, of a 511 bp fragment of the beta-giardin gene (bg), and of a 530 bp fragment of the glutamate dehydrogenase (gdh) gene were as previously described (Lalle et al. 2005, Cacciò et al. 2008, 2010). For the triose phosphate isomerase (tpi) gene, two protocols were used, one with broad specificity (Sulaiman et al. 2003) and the other that specifically targets G. duodenalis assemblages C and D (Lebbad et al. 2010). In all cases, the primary polymerase chain reaction (PCR) consisted of 25 μL of 2× PCR master mix (Promega), 10 pmol of each primer, and 1–3 μL of DNA in a total reaction volume of 50 μL. For the nested PCR, 2.5–5 μL of the first PCR were used as template. PCR products were separated by electrophoresis in 1.5% agarose gels stained with ethidium bromide. PCR products were purified by using the Qiaquick purification kit (Qiagen) and sequenced on both strands by using the ABI Prism BIGDYE Terminator Cycle Sequencing Kit (Life Technologies) according to the manufacturer's instructions. The sequencing reactions were analyzed by using the ABI 3100 automatic sequencer (Applied Biosystems), and sequences were assembled by using the software program SeqMan II (DNASTAR). Data were stored and analyzed in Bionumerics (Version 6.01; Applied Math).

Phylogenetic analysis

Sequences were aligned by using Clustal X (Thompson et al. 1997), distance-based analyses were conducted by using Kimura 2-parameters distance estimates, and trees were constructed by using the Neighbor-Joining algorithm, implemented in the MEGA program version 4.0 (Tamura et al. 2007). Bootstrap proportions were calculated by the analysis of 1,000 replicates for neighbor-joining trees.

Nucleotide sequence accession numbers

The nucleotide sequences determined in this work were deposited in GenBank under the accession numbers JN416511-JN416562 (beta-giardin), JN587349-JN587401 (gdh), JN587402-JN587492 (tpi), and JN603679-JN603734 (ITS).

Results

Molecular genotyping at the ITS region

Amplification of the ∼315 bp fragment encompassing the ITS1-5.8S-ITS2 region was obtained from 56 of the 96 (58%) G. duodenalis isolates (Table 1). Sequence analysis of the amplification products revealed assemblage C in 14 isolates, assemblage D in 40 isolates, and 1 case of a mixed C and D infection. There was no sequence variability among isolates belonging to either assemblage C or assemblage D (i.e., no intra-assemblage variability; Table 2).

Table 1.

ist of the Dog Isolates and Genotyping Data at the Four Investigated Markers

Isolate code	ITS	BG	GDH	TPI	TPI-D
Kennel dogs
ISSGdA179	neg	neg	neg	B	nd
ISSGdA345	neg	neg	C^ash	neg	D^ash
ISSGdA348	D	neg	neg	B^ash	neg
ISSGdA349	neg	neg	D^ash	B	D
ISSGdA353	C	neg	C	A^pz	neg
ISSGdA358	neg	C	neg	B^pz	D
ISSGdA362	neg	neg	C	B^pz	neg
ISSGdA364	neg	neg	neg	B^pz	neg
ISSGdA369	neg	neg	neg	B^ash	neg
ISSGdA371	D	neg	neg	B^pz	neg
ISSGdA372	nd	C^ash	C	C	C
ISSGdA373	D^ash	neg	D	C	D
ISSGdA374	neg	neg	neg	B^ash	neg
ISSGdA376	D	neg	neg	B^pz	neg
ISSGdA379	neg	C	C^ash	C	C
ISSGdA382	neg	D^ash	D^ash	neg	D
ISSGdA383	D	neg	D	B	D^ash
ISSGdA387	neg	neg	D^ash	C	D^ash
ISSGdA389	D	D	D	neg	D
ISSGdA396	neg	neg	neg	B^ash	neg
ISSGdA397	neg	D	C	neg	D^ash
ISSGdA401	D	D^ash	D	C^ash	D^ash
ISSGdA402	neg	neg	D	C	C
ISSGdA405	D	C^ash	neg	C^ash	D
ISSGdA406	D	D^ash	D	C	D^ash
ISSGdA407	D	D	neg	C^ash	D^ash
ISSGdA412	neg	neg	C	C^ash	D
ISSGdA414	neg	neg	C	C	D
ISSGdA416	D	neg	neg	C^ash	D^ash
ISSGdA418	neg	neg	A^pz	B	neg
ISSGdA419	C	neg	neg	B	neg
ISSGdA425	neg	neg	neg	A^ash	neg
ISSGdA427	neg	neg	neg	B	neg
ISSGdA429	neg	neg	neg	B	neg
ISSGdA431	neg	neg	neg	B	neg
ISSGdA433	neg	neg	neg	C	neg
ISSGdA449	neg	neg	D	B^pz	neg
ISSGdA452	neg	A^pz	neg	B^ash	neg
ISSGdA454	D	neg	neg	B^pz	D
ISSGdA456	D	D^ash	neg	B	D
ISSGdA460	D	neg	D	B	neg
ISSGdA464	neg	neg	neg	A^ash	neg
ISSGdA466	D	D	D	B^pz	D^ash
ISSGdA469	D	D^ash	D	B^pz	D^ash
ISSGdA470	D	neg	neg	C^ash	D^ash
ISSGdA471	neg	A^pz	neg	neg	neg
ISSGdA475	neg	A^pz	C	B^pz	neg
ISSGdA478	neg	neg	neg	B	neg
ISSGdA480	neg	A^ash	neg	neg	neg
ISSGdA482	D	A^pz	A^pz	B^pz	neg
ISSGdA486	D	D	neg	neg	D
ISSGdA489	C	C	C	C	C
Household dogs
ISSGdA717	neg	D	C	C^ash	neg
ISSGdA751	neg	neg	D	neg	D^ash
ISSGdA752	neg	neg	neg	neg	D^ash
ISSGdA754	D	D	neg	neg	neg
ISSGdA755	D	neg	neg	neg	D^ash
ISSGdA756	neg	B	neg	neg	neg
ISSGdA757	C	C	C^ash	C^ash	nd
ISSGdA758	C	C	C	C^ash	C^ash
ISSGdA760	C	C	C	C	C
ISSGdA761	C	B^pz	neg	neg	nd
ISSGdA764	neg	C	C^ash	C	neg
ISSGdA768	neg	C^ash	neg	neg	C^ash
ISSGdA769	C	C^ash	C^ash	C^ash	C^ash
ISSGdA770	neg	B^pz	C	C	C
ISSGdA771	D	neg	neg	neg	D
ISSGdA772	D	neg	neg	neg	D
ISSGdA773	neg	C	C	C^ash	nd
ISSGdA776	neg	B^ash	neg	neg	C
ISSGdA777	neg	C	C	C^ash	neg
ISSGdA778	C	neg	C	C	C
ISSGdA779	D	B^pz	neg	neg	neg
ISSGdA781	D	B	neg	C	neg
ISSGdA782	C	B	neg	neg	neg
ISSGdA785	D	D^ash	neg	neg	D^ash
ISSGdA856	D	neg	neg	neg	nd
ISSGdA871	D	D	neg	neg	neg
ISSGdA872	D	D	neg	C	nd
ISSGdA886	D	D^ash	D^ash	neg	D^ash
ISSGdA887	C	C	neg	C	C
ISSGdA888	D	C	C^ash	neg	D
ISSGdA889	D	neg	neg	neg	D
ISSGdA890	C	neg	neg	neg	neg
ISSGdA891	D	C^ash	C	C^ash	C^ash
ISSGdA892	D	D	D	neg	D
ISSGdA893	C	C^ash	C	C^ash	C^ash
ISSGdA894	D	neg	neg	neg	neg
ISSGdA895	D	D	D	neg	D
ISSGdA896	D	D^ash	neg	neg	neg
ISSGdA898	C	neg	neg	neg	neg
ISSGdA899	D	D	D	neg	D^ash
ISSGdA903	D	neg	D	neg	neg
ISSGdA905	C	C	C	C^ash	C^ash
ISSGdA907	D	D	D^ash	neg	D
ISSGdA908	C+D	C	neg	C	Neg

neg, negative; nd, not done; ^pzpotentially zoonotic: identical sequence has been found in humans before; ^ashallelic sequence heterogeneity

Table 2.

Occurrence of Sequence Variants at the Four Investigated Markers in the Different Giardia duodenalis Assemblages

Sequence variants	A	B	C	D	Total
BG	3	5	8	8	24
GDH	1	0	16	14	31
TPI	2	17	27	27	73
ITS	0	0	1	1	2

Molecular genotyping at the bg gene

Amplification of a 511 bp fragment of the bg gene was obtained from 52 of 96 (54%) G. duodenalis isolates (Table 1). Sequence analysis revealed assemblage A in 5 isolates, assemblage B in 7 isolates, assemblage C in 19 isolates, and assemblage D in 21 isolates. Among isolates typed as assemblage A, three different subtypes were identified, whereas five subtypes were found among isolates typed as assemblage B, eight subtypes among isolates typed as assemblage C, and eight subtypes among isolates typed as assemblage D (Table 2). A detailed description of the nucleotide substitutions for assemblages C and D is provided in Table 3.

Table 3.

Nucleotide Substitutions as Found in the Beta-Giardin, Glutamate Dehydrogenase and Triose Phosphate Isomerase Genes in Assemblages C and D

BG assemblage C (n=19)	Position	135	221	315	327	549
T		0	0	11	1	1
A		1	18	0	0	0
C		0	0	5	18	18
G		18	1	0	0	0
N		0	0	5	0	0
BG assemblage D (n=21)	Position	156	198	204	313
T		0	0	3	0
A		19	7	0	20
C		0	0	15	0
G		0	11	0	0
N		2	3	3	1
GDH assemblage C (n=24)	Position	558	624	662	690	691	705	810	835	846	939
T		0	1	0	0	0	1	1	0	18	5
A		2	0	21	1	21	0	0	20	0	0
C		0	23	0	0	0	23	23	0	1	0
G		21	0	1	22	1	0	0	0	0	18
N		1	0	2	1	2	0	0	2	5	1
GDH assemblage D (n=20)	Position	543	549	603	615	633	637
T		0	0	4	0	12	0
A		1	0	0	10	6	0
C		0	19	11	0	0	0
G		19	0	0	5	0	19
N		0	1	5	5	2	1
TPI assemblage C (n=33)	Position	77	87	90	114	138	150	215	273	288	330	382	383	393	408	451	496	522
T		0	0	1	16	0	4	0	5	0	16	0	19	0	0	0	0	4
A		0	1	0	0	1	0	32	0	0	0	30	0	21	0	0	1	0
C		30	0	32	7	0	0	0	22	31	5	0	9	5	0	32	0	29
G		0	32	0	0	32	27	1	0	0	0	2	0	0	30	0	32	0
N		3	0	0	10	0	6	0	6	2	12	1	5	7	3	1	0	0
TPI assemblage D (n=35)	Position	54	90	93	209	219	233	262	333	346	352	354	468	483
T		34	0	1	0	1	0	0	16	0	16	0	0	34
A		0	0	0	59	0	1	0	0	0	0	15	34	0
C		1	34	34	0	31	0	0	9	34	9	0	0	1
G		0	0	0	1	0	31	34	0	0	0	11	1	0
N		0	1	0	0	3	3	1	10	1	10	9	0	0

Underlined positions are those where allelic sequence heterozygosis (ASH) was found.

Molecular genotyping at the gdh gene

The amplification of a 530 bp fragment of the gdh gene was obtained from 46 of 96 (48%) G. duodenalis isolates (Table 1). Sequence analysis revealed assemblage A in 2 isolates, assemblage C in 24 isolates, and assemblage D in 20 isolates. The two isolates typed as assemblage A had a sequence identical to that of genotype AI, which is zoonotic. Among the isolates typed as assemblage C, 16 different sequences were identified, whereas 14 different sequences characterized the isolates belonging to assemblage D (Table 2). A detailed description of the nucleotide substitutions for assemblages C and D is provided in Table 3.

Molecular genotyping at the tpi gene

The amplification of a 530 bp fragment of the tpi gene was obtained from 62 of 96 (64.5%) G. duodenalis isolates (Table 1). Sequence analysis of the PCR products revealed assemblage A in 3 isolates, assemblage B in 27 isolates, and assemblage C in 32 isolates (Table 1). Amplification with the assemblage D specific primers was obtained from 50 of 90 isolates (55%; for 6 isolates, the amplification was not possible due to DNA exhaustion). Sequencing of these products revealed assemblage C in 15 isolates and assemblage D in 35 isolates. Of note, 20 isolates that were PCR negative using the generic assay were amplified with the D-specific primers, and classified as C (2 isolates) or D (18 isolates). Therefore, the combined use of generic and D-specific primers resulted in the amplification of 82 of 96 isolates (85.4%). A total of 7 isolates typed as assemblage B, and of 10 isolates typed as assemblage C with generic primers, were reclassified as assemblage D after PCR with assemblage-specific primers and sequencing of PCR products (Table 1).

Among the isolates typed as assemblage A, 2 subtypes were found, whereas 17 subtypes were found among the isolates typed as assemblage B, 27 subtypes among isolates typed as assemblage C, and 27 subtypes among isolates typed as assemblage D (Table 2). A detailed description of the nucleotide substitutions for assemblages C and D is provided in Table 3.

Combined genotyping results at four loci

The combined results showed that amplification at the four loci tested was obtained from 20 of the 96 isolates (20.8%). Amplification at three loci was obtained from 24 isolates (25%): failure in amplification reactions occurred at the ITS locus (10 isolates), at the gdh locus (9 isolates), and at the bg locus (5 isolates). Amplification at 2 loci was obtained from 32 isolates (33%): failure in amplification reactions occurred at the ITS/bg loci (10 isolates), at the ITS/gdh loci (4 isolates), at the bg/gdh loci (11 isolates), at the gdh/tpi loci (6 isolates), and at the bg/tpi loci (1 isolate). Finally, amplification at a single locus was observed in 20 isolates (20.8%): in 13 isolates only, the amplification at the tpi locus was obtained; in 4 isolates only, the ITS locus was amplified; and in 3 isolates only, the bg locus was amplified (Table 1). Interestingly, 14 of the 20 isolates that could be amplified at a single locus were classified as zoonotic (10 as B, 4 as A), mostly at the tpi locus (11 isolates, 9 as B and 2 as A). Among household dogs, the occurrence of zoonotic assemblages was only supported by results at the bg locus, and only assemblage B was identified in 7 of 44 isolates (15%). On the contrary, among kennel dogs, zoonotic assemblages were identified in 31 of 52 isolates (59.6%), mainly at the tpi locus (25 isolates typed as assemblage B and 3 as assemblage A), but also at the bg and gdh loci (5 and 2 isolates typed as assemblage A, respectively).

Single infections with assemblage C were found in 15 isolates (7 typed at 4 loci, 7 typed at 3 loci, and 1 typed at 2 loci), whereas single infections with assemblage D were found in 18 isolates (6 typed at 4 loci, 3 typed at 3 loci, and 9 typed at 2 loci). The detection of 2 assemblages was observed in 38 isolates (18 as C+D, 12 as B+D, 6 as B+C, A+B in 2, and A+C in 1), whereas 3 different assemblages were found in 4 isolates. The remaining 20 isolates could only be typed at a single locus.

Phylogenetic analysis of single and concatenated genes

The sequences of single genes from assemblages C and D were aligned with all homologous sequences available from previous studies, and phylogenetic analyses were performed by using neighbor joining and maximum likelihood methods. The nucleotide substitutions found in the different genes of all assemblages C and D isolates are listed in Table 4.

Table 4.

Nucleotide Substitutions as Found in the Beta-Giardin, Glutamate Dehydrogenase and Triose Phosphate Isomerase Genes from all Available Assemblages C and D Isolates

BG assemblage C (n=42)	Position	135	221	234	315	327	330	528	549
T		0	0	0	26	1	0	1	2
A		2	41	0	0	0	1	0	0
C		0	0	0	11	41	0	41	39
G		40	1	41	0	0	41	0	0
N		0	0	1	5	0	0	0	1
BG assemblage D (n=103)	Position	129	156	162	167	198	199	204	269	279	313	324	384	387	420	494	495	516	519	534	535
T		102	1	102	0	0	0	16	0	0	0	0	0	102	1	1	1	102	0	0	0
A		0	92	0	1	30	100	13	1	102	101	0	102	0	0	102	0	0	0	0	102
C		1	0	0	0	0	0	64	0	1	0	1	0	1	102	0	0	1	101	0	0
G		0	6	1	102	56	1	0	102	0	0	102	1	0	0	0	102	0	1	102	0
N		0	4	0	0	8	2	4	0	0	2	0	0	0	0	0	0	0	1	1	1
GDH assemblage C (n=38)	Position	558	585	586	624	662	690	691	705	711	810	835	846	849	939
T		0	1	0	2	0	0	0	2	1	1	0	25	0	9
A		2	0	2	0	34	2	34	0	0	0	29	0	3	0
C		0	37	0	36	0	0	0	36	37	37	0	7	0	0
G		35	0	35	0	2	35	1	0	0	0	5	0	35	28
N		1	0	1	0	2	1	3	0	0	0	4	6	0	1
GDH assemblage D (n=67)	Position	534	537	543	549	603	615	624	633	636	637
T		1	0	0	5	32	0	0	56	1	0
A		0	2	4	0	0	28	3	8	0	0
C		66	0	0	61	28	0	0	0	66	0
G		0	65	63	0	1	31	63	0	0	66
N		0	0	0	1	6	8	1	3	0	1
TPI assemblage C (n=63)	Position	77	87	90	105	114	138	150	215	273	288	324	330	382	383	393	408	451	452	492	496	522
T		3	0	2	0	38	0	4	0	7	1	0	34	0	29	0	0	0	0	61	0	8
A		0	1	0	0	0	1	0	62	0	0	1	0	59	0	40	2	0	62	2	1	0
C		53	0	61	62	11	0	0	0	47	59	0	10	0	22	11	0	62	0	0	0	52
G		0	62	0	0	0	61	49	1	0	0	62	0	2	0	0	57	0	1	0	62	0
N		7	0	0	1	14	1	10	0	9	3	0	19	2	12	12	4	1	0	0	0	3
TPI assemblage D (n=60)	Position	54	66	77	90	93	120	147	182	209	219	233	262	333	346	352	354	394	447	451	464	468	483
T		56	0	0	0	2	59	59	0	0	2	0	0	31	0	27	0	4	0	0	59	0	59
A		0	0	0	0	0	0	0	57	59	0	2	0	0	0	0	26	0	1	0	0	59	0
C		2	0	59	59	58	0	0	0	0	54	0	0	15	59	21	0	48	0	59	0	0	1
G		0	59	0	0	0	0	0	3	1	0	54	59	0	0	0	22	0	59	0	0	1	0
N		2	1	1	1	0	1	1	0	0	4	4	1	14	1	12	12	8	0	1	1	0	0

Underlined positions are those where ASH was found.

As shown in Figure 1, assemblage C and D were significantly separated from each other at the bg, gdh, and tpi genes, and also at the ITS1-5.8-ITS2 region (not shown). Based on bootstrap values, no significant substructuring within assemblage C and D was observed for the bg and gdh genes (Fig. 1, upper panels). A small, but significant, substructuring was observed in Assemblage D at the tpi gene (Fig. 1, right upper panel), albeit no correlation was observed with available epidemiologic information, such as the country of origin or the year of collection. Similar results were obtained when the analyses were performed on the aminoacid sequences (not shown). Next, the individual genes of single isolates were concatenated, and phylogenetic analyses were performed under two different assumptions: (1) a G. duodenalis field isolate is one unique isolate (i.e., a clonal strain), and the sequences at the different loci are derived from a single genotype; and (2) a G. duodenalis field isolate is a mixture of coexisting, nonrecombining populations, when more than one assemblage is detected. In other words, when the different genes of a field isolate are not from the same assemblage, then the isolate is considered to be a mixed infection. These two assumptions result in two different phylogenetic trees. When a field isolate corresponds to a unique genotype, then a complex phylogenetic tree is generated with eight significant Multi-Locus Genotypes (Fig. 1, left bottom panel). When a field isolate is considered a mixed infection, then the phylogenetic tree becomes relatively simple with only two significantly different MLGs and is very similar to the phylogenetic trees on the three individual genes (Fig. 1, right bottom panel).

FIG. 1.

Phylogenetic analysis based on single (A) and concatenated gene sequences (B), as obtained by using neighbor joining. Bootstrap values are indicated. Terminal branches were collapsed for clarity.

Discussion

Giardiasis is a very common infection of dogs, albeit prevalence varies considerably among different studies (recently reviewed by Feng and Xiao 2011). It is generally accepted that dogs from shelters or intensive breeding establishments show higher levels of infection compared with household dogs, and that young animals are significantly more affected than older ones (Tangtrongsup and Scorza 2010). For example, in a national survey conducted in the United States on >1 million pet dogs, an average rate of infection of 4% was reported, but this value rises to 13.1% for puppies and decreases to 1% for dogs older than 3 years of age (Little et al. 2009). Similarly, a European study revealed an overall prevalence of 24.8% (34.7% in shelter dogs), and again this value rises to 42.9% for very young dogs (Epe et al. 2010). Infection with G. duodenalis in dogs is often asymptomatic but has been associated with the occurrence of diarrhea and ill thrift in puppies (Thompson 2004).

From a public health perspective, it is necessary to distinguish G. duodenalis cysts that primarily infect dogs and other carnivores (assemblages C and D) from those that have zoonotic potential (assemblages A and B), and this can only be based on their molecular characterization. To date, >600 dog samples have been genotyped, mainly using PCR and sequencing of the SSU-rRNA gene (reviewed by Feng and Xiao 2011). Although the studies were quite heterogeneous in terms of sample origin and methods used, about 25% of dog isolates were classified as assemblage A, whereas the rest of the isolates were assemblage C or D, with the latter being more common overall (Sprong et al. 2009). Notably, assemblage B, which is commonly found in humans worldwide, has been detected in a limited number of dogs in Europe, Asia, and Australia (Ballweber et al. 2010, Feng and Xiao 2011), and, therefore, appears to be rare in this host.

To allow useful comparisons of G. duodenalis isolates, a genotyping scheme that discriminates within assemblages (i.e., that identifies subtypes) should be used, and this is not achievable when very conserved genes, such as the SSU-rRNA and the elongation factor 1-α, are investigated. Despite this, almost half of the studies performed to date were based on the analysis of a short (130 bp) fragment of the SSU-rRNA locus (Ballweber et al. 2010), due to the high sensitivity in the amplification of this multicopy target. We have previously shown that a multilocus typing scheme should be used to increase discrimination (Cacciò et al. 2008), and other research groups have further confirmed this (Lasek-Nesselquist et al. 2009, Lebbad et al. 2010). Recently, we developed a database to store genetic and epidemiologic information on Giardia spp. and demonstrated that genetic information from multiple loci is necessary to infer zoonotic potential (Sprong et al. 2009).

In the current study, a genetic characterization of a large number of dog isolates was attempted for the first time by PCR amplification of four polymorphic loci. In terms of the applicability of these assays, we found that PCR at the bg and gdh loci had the poorest performance, thus resulting in the amplification of only half of the isolates. Amplification at the ITS1-5.8S-ITS2 region was also obtained from about half (58%) of the isolates, but since this PCR assay was developed in our laboratory in 2009 (Cacciò et al. 2010), many samples were tested several years after DNA extraction. Indeed, the percentage of amplified DNA increases to 87% when only the DNA extracted during 2007–2009 are considered, thus indicating a possible deterioration of DNA over time. Finally, the combined use of generic and assemblage-specific primers for tpi amplification yielded good results (85.4%). Since all the primers used in this study (with the exception of the tpi assemblage D specific primers) have been designed to allow amplification of DNA from all G. duodenalis assemblages, the observed difference in amplification rates are likely due to mismatches in the binding regions of the primers, similar to what was observed for the tpi gene (Lebbad et al. 2010). It is difficult to make a general comparison between our findings and those previously published, mainly because only “representative” samples were sequenced in most studies (e.g., Palmer et al. 2008): therefore, the relative performance of different sets of primers in the amplification of G. duodenalis DNA from dogs remains to be determined.

Next, the occurrence of different G. duodenalis assemblages in the dog samples was analyzed: as shown in Table 1, assemblages C and D predominate in both kennel dogs and household dogs. It has been speculated that the circulation of zoonotic versus host-adapted assemblages is influenced by several factors. On the one hand, dogs kept in kennels will be more exposed to infection due to the ease of spread of G. duodenalis in those typically crowded environments, and assemblages C and D can outcompete other assemblages due to their better adaptation to the host. On the other hand, household dogs would have more frequent contact with zoonotic assemblages originating from humans or other animals in urban settings, and the low pressure of infection will make infection with zoonotic or host-adapted genotypes equally probable (Thompson 2004). However, the available data do not show a clearly different distribution of assemblages in these dog populations: for example, only 1 of 12 household dogs was infected with assemblage A in a Dutch study, the remaining were all typed as assemblages C and D (Overgaauw et al. 2009). In a study in Peru, Cooper et al. (2010) analyzed the transmission of G. duodenalis in 22 families and in dogs living in the same households: genotyping revealed only assemblages C and D in dogs, whereas humans were infected only with assemblages A and B. On the contrary, a study in Belgium (Claerebout et al. 2009) found that 33 of 41 household dogs were infected with assemblage A, whereas 31 of 33 kennel dogs were infected with assemblages C and D. Future studies are necessary to understand the factors that account for these data, and adapting a defined genotyping scheme will represent an advantage.

The availability of substantial DNA sequence information from multiple loci, both from the current study and from previously published data (see Sprong et al. 2009), allowed us to investigate the existence of subgroups within assemblages C and D by performing phylogenetic analysis. Trees were first inferred from single genes (Fig. 1), but no statistically supported sub-groups were found in either assemblage C or D, with the exception of the tpi gene tree that shows two clusters in assemblage D, which, however, do not correlate with either the origin of the isolates or the year of collection (data not shown). When the analysis was done on concatenated genes, it was necessary to treat the data under two different hypothesis: (1) recombination was considered to have occurred, and the resulting tree was complex with different subgroups identified (Fig. 1); (2) mixed infections were considered as the explanation for the incongruent assignment of isolates to single assemblages, and the concatenated tree is essentially identical to those generated by single genes (Fig. 1). In a study based on isoenzymes (Monis et al. 2003), three subgroups within assemblage C were observed in 10 dog isolates from Australia, whereas the analysis was not possible for assemblage D, as only a single isolate was available. The fact that subgroups cannot be identified by using DNA sequence may be attributed to the nature of the substitution pattern, as the predominance of mutations at third codon positions (Tables 3 and 4) may blur phylogenetic relationships due to homoplasy. When a protein-based tree was generated, no evident subgroups were observed, and this might be attributed to the small number of nonsynonymous substitutions in these rather conserved genes, and to the necessity to exclude the ITS locus from this analysis. Therefore, the recognition of any significant population structure in either assemblage C or D will require the investigation of more genetic loci from isolates collected in different areas of the world.

Finally, we evaluated our results with regard to the zoonotic potential of G. duodenalis. In household dogs, only assemblage B was detected (at the bg locus) at a relatively low frequency (16%), whereas assemblage A was not found. Importantly, the assignment to assemblage B was not confirmed when the other loci were tested, as only assemblage C or D were detected in these isolates. In kennel dogs, assemblage B was detected in about half of the samples and always at the tpi locus, whereas assemblage A was identified at a lower frequency (20%) and was detected at all loci except ITS. Of note, only zoonotic assemblages were found in 15 isolates (29%), albeit in most of those cases only amplification of a single locus was obtained.

The presence of mixed infections is supported by results obtained using tpi assemblage specific PCR and sequencing. Indeed, assemblage B was detected in 5 isolates using tpi generic primers, whereas the same isolates were typed as assemblage D using tpi assemblage specific primers; since the bg, gdh, and ITS loci supported classification of these isolates as assemblage D, it is very likely that the detection of assemblage B at the tpi locus was due to a mixed B+D infection. However, genetic exchanges can also account for this finding; there are evidences for recombination between isolates of the same assemblage (Cooper et al. 2007), but opinions differ about the occurrence of recombination between different G. duodenalis assemblages (Lasek-Nesselquist et al. 2009, Cooper et al. 2010). To what extent mixed infections and genetic exchanges are contributing to the complexity of genotyping results is currently unknown. In this context, the sensitivity and specificity of real-time PCR can be exploited to design assays able to detect multiple assemblages in a single amplification. For the genotyping of Giardia isolates from dogs, detection of assemblages A, B, C, and D will be necessary: an assay based on four different TaqMan probes can, therefore, be developed. Alternatively, assemblage-specific markers can be identified by the analysis of whole genome sequence from assemblages C and D representatives. We contend that until a more robust methodology is developed, caution should be used in interpreting genotyping data from dog isolates of Giardia, particularly when only single loci are investigated.

Footnotes

Acknowledgments

This study was supported by the European Commission (contract SANCO/2006/FOODSAFETY/032). The authors thank Daniele Tonanzi for his excellent technical support.

Disclosure Statement

No competing financial interests exist.

References

Ballweber

, Xiao

, Bowman

, Kahn

et al. Giardiasis in dogs and cats: update on epidemiology and public health significance. Trends Parasitol, 2010; 26:180–189.

Cacciò

, Beck

, Almeida

, Bajer

et al. Identification of Giardia species and Giardia duodenalis assemblages by sequence analysis of the 5.8S rDNA gene and internal transcribed spacers. Parasitology, 2010; 17:1–7.

Cacciò

, Beck

, Lalle

, Marinculic

et al. Multilocus genotyping of Giardia duodenalis reveals striking differences between assemblages A and B. Int J Parasitol, 2008; 38:1523–1531.

Cacciò

, Ryan

. Molecular epidemiology of giardiasis. Mol Biochem Parasitol, 2008; 160:75–80.

Claerebout

, Casaert

, Dalemans

, De Wilde

et al. Giardia and other intestinal parasites in different dog populations in Northern Belgium. Vet Parasitol, 2009; 161:41–46.

Cooper

, Adam

, Worobey

, Sterling

. Population genetics provides evidence for recombination in Giardia. Curr Biol, 2007; 17:1984–1988.

Cooper

, Sterling

, Gilman

, Cama

et al. Molecular analysis of household transmission of Giardia lamblia in a region of high endemicity in Peru. J Infect Dis, 2010; 202:1713–1721.

da Silva

, Bornay-Llinares

, Moura

, Slemenda

et al. Fast and reliable extraction of protozoan parasite DNA from fecal specimens. Mol Diagn, 1999; 4:57–64.

Epe

, Rehkter

, Schnieder

, Lorentzen

et al. Giardia in symptomatic dogs and cats in Europe-results of a European study. Vet Parasitol, 2010; 173:32–38.

10.

Feng

, Xiao

. Zoonotic potential and molecular epidemiology of Giardia species and giardiasis. Clin Microbiol Rev, 2011; 24:110–140.

11.

Lalle

, Pozio

, Capelli

, Bruschi

et al. Genetic heterogeneity at the beta-giardin locus among human and animal isolates of Giardia duodenalis and identification of potentially zoonotic subgenotypes. Int J Parasitol, 2005; 35:207–213.

12.

Lasek-Nesselquist

, Welch

, Sogin

. The identification of a new Giardia duodenalis assemblage in marine vertebrates and a preliminary analysis of G. duodenalis population biology in marine systems. Int J Parasitol, 2010; 40:1063–1074.

13.

Lasek-Nesselquist

, Welch

, Thompson

, Steuart

et al. Genetic exchange within and between assemblages of Giardia duodenalis. J Eukaryot Microbiol, 2009; 56:504–518.

14.

Lebbad

, Mattsson

, Christensson

, Ljungström

et al. From mouse to moose: multilocus genotyping of Giardia isolates from various animal species. Vet Parasitol, 2010; 168:231–239.

15.

Little

, Johnson

, Lewis

, Jaklitsch

et al. Prevalence of intestinal parasites in pet dogs in the United States. Vet Parasitol, 2009; 166:144–152.

16.

Monis

, Andrews

, Mayrhofer

, Ey

. Genetic diversity within the morphological species Giardia intestinalis and its relationship to host origin. Infect Genet Evol, 2003; 3:29–38.

17.

Monis

, Cacciò

, Thompson

RCA

. Variation in Giardia: towards a taxonomic revision of the genus. Trends Parasitol, 2009; 25:93–100.

18.

Overgaauw

, van Zutphen

, Hoek

, Yaya

et al. Zoonotic parasites in fecal samples and fur from dogs and cats in The Netherlands. Vet Parasitol, 2009; 163:115–122.

19.

Palmer

, Traub

, Robertson

, Devlin

et al. National study of the gastrointestinal parasites of dogs and cats in Australia. Vet Parasitol, 2008; 151:181–190.

20.

Sprong

, Cacciò

, van der Giessen

. Identification of zoonotic genotypes of Giatrdia duodenalis. PLOS Neg Trop Dis, 2009; 3:e558.

21.

Sulaiman

, Fayer

, Bern

, Gilman

et al. Triosephosphate isomerase gene characterization and potential zoonotic transmission of Giardia duodenalis. Emerg Infect Dis, 2003; 9:1444–1452.

22.

Tamura

, Dudley

, Nei

, Kumar

. MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0. Mol Biol Evol, 2007; 24:1596–1599.

23.

Tangtrongsup

, Scorza

. Update on the dagnosis and management of Giardia spp infections in dogs and cats. Top Companion Anim Med, 2010; 25:155–162.

24.

Thompson

, Gibson

, Plewniak

, Jeanmougin

et al. The CLUSTAL_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res, 1997; 25:4876–4882.

25.

Thompson

. The zoonotic significance and molecular epidemiology of Giardia and giardiasis. Vet Parasitol, 2004; 126:15–35.

26.

Thompson

, Monis

. Variation in Giardia: implications for taxonomy and epidemiology. Adv Parasitol, 2004; 58:69–137.

27.

[WHO] World Health Organization. Parasitic Zoonoses. Report of a WHO Expert Committee with the participation of FAO. Technical Report Series No. 637. Geneva: WHO, 1979.