Marine Biology

, Volume 158, Issue 2, pp 461–470

Microsporidia infections in the amphipod, Echinogammarus marinus (Leach): suggestions of varying causal mechanisms to intersexuality

Authors

  • Gongda Yang
    • Institute of Marine Sciences, School of Biological SciencesUniversity of Portsmouth
    • Environmental Research Institute, North Highland CollegeUHI Millennium Institute
  • Stephen Short
    • Institute of Marine Sciences, School of Biological SciencesUniversity of Portsmouth
  • Peter Kille
    • School of BiosciencesUniversity of Wales Cardiff
    • Institute of Marine Sciences, School of Biological SciencesUniversity of Portsmouth
Original Paper

DOI: 10.1007/s00227-010-1573-7

Cite this article as:
Yang, G., Short, S., Kille, P. et al. Mar Biol (2011) 158: 461. doi:10.1007/s00227-010-1573-7

Abstract

A notable body of research has established a clear link between intersexuality and the feminising influence of microsporidia infection in amphipods. In this study, we investigated the relationship between microsporidia infection and intersexuality in the amphipod Echinogammarus marinus (Leach) from Portsmouth, southern England. The analysis revealed a male-biased population (~2:1) harbouring both Dictyocoela berillonum and Dictyocoela duebenum microsporidia, with ~38 and 6% of animals displaying ‘high’ levels of infection, respectively. We also reveal the presence of several intersex phenotypes: intersex females (1%) that possess genital papillae. Two male intersex phenotypes—internal intersex (8.2%), possessing an oviduct structure on their testes, and external intersex males (4.4%), which possess external brood plates. We found a statistically significant relationship between D. berillonum infection and the external intersex male phenotype; however, the male-biased population and low levels of female infection suggest that the correlation may not be the result of incomplete feminisation. In addition, we found that the internal and external intersex characteristics are rarely seen on the same specimen, suggesting that the male intersex phenotypes are caused by distinct mechanisms. In combination, these findings are suggestive of a more complex relationship between amphipod intersexuality and their microsporidia than had previously been recognised.

Introduction

Intersexuality, an atypical sex phenotype including both male and female secondary sex characteristics on the same individual, has been widely reported in various crustaceans, such as lobsters, crabs and amphipods (Bulnheim 1965; Ginsburger-Vogel 1991; Sangalang and Jones 1997; Zou and Fingerman 2000). The specific mechanism causing intersex phenotypes in crustaceans still remains unknown; however, several factors have been linked to intersexuality, including genetic factors, environmental conditions (photoperiod, temperature and salinity), as well as cytoplasmic parasites (Bulnheim 1978; Ginsburger-Vogel 1991; Dunn et al. 1996; Rodgers-Gray et al. 2004). Feminising parasites have been discovered in some crustacean groups, for example Wolbachia bacteria in isopods and microsporidia in amphipods (Bouchon et al. 1998; Rigaud and Juchault 1998; Terry et al. 2004). Strong evidence establishing a relationship between the occurrence of intersex and the incidence of feminising parasite infections indicates that intersexuality in crustaceans is a possible result of incomplete feminisation by cytoplasmic parasites (Rigaud and Juchault 1998; Kelly et al. 2002, 2004).

Microsporidia are a group of single-cell eukaryotes, which are obligate intracellular parasites infecting a variety of animal groups, ranging from protists to humans (Sprague and Vavra 1977; Canning 1990). Due to their highly specialised parasitic lifestyle, the survival of a microsporidium is critically dependent on the transmission success amongst its hosts (Dunn and Smith 2001). Generally, two transmission strategies have been revealed in microsporidia, namely vertical and horizontal transmission. Horizontal transmission occurs between hosts of the same or different species and generations, and vertically transmitted (principally transovarial) microsporidia are passed from parental generation to their offspring (Dunn and Smith 2001). In most occasions, parasites applying horizontal transmission have higher copy numbers and stronger virulence than vertically transmitted ones (Kellen et al. 1965; Ebert and Herre 1996). In contrast, vertically transmitted parasites have subtle or neutral influence as their survival largely relies on their hosts’ fitness and reproductive success (Lipsitch et al. 1996; Dunn and Smith 2001; Rigaud and Haine 2005). Most vertically transmitted parasites, with the exception of some viruses, for example the sigma virus in Drosophila melanogaster (Bangham et al. 2007), follow a maternal line. In order to increase the chance of being passed to the next generation by female hosts, some vertically transmitted microsporidia are capable of distorting the host’s sex ratio in favour of females, since male hosts are a biological “dead-end” (Kellen et al. 1965; Dunn et al. 1996). Strategies employed by microsporidia to distort their host’s sex ratio have been reported as male-killing in mosquitoes (Andreadis and Hall 1979; Bandi et al. 2001) and feminising effects in amphipods (Dunn et al. 1993; Terry et al. 1998; Rodgers-Gray et al. 2004).

It has been revealed that microsporidia infection is wide spread in a variety of amphipod species, such as Gammarus duebeni, Corophium volutator and Orchestia mediterranea (Ginsburger-Vogel 1991; Dunn et al. 1993; Mautner et al. 2007). Terry et al. (2004) reported 11 species of microsporidia in 16 amphipod species (100% of investigated amphipod populations), with 5 out of 8 vertically transmitted microsporidia showing asymmetrical distribution in male and female hosts, which indicated they had potential sex-distorting effects. Coexistence of different microsporidia species in the same amphipod population has been reported by several studies, but dual infection by two microsporidia species in a single amphipod host is reported to be very rare (Ironside et al. 2003; Terry et al. 2004; Haine et al. 2004). According to a theoretical model proposed by Ironside et al. (2003), stable co-occurrence of two vertically transmitted parasites in a single host population is not expected to exist, since the less competitive parasite will be completely replaced by the dominant species. Some microsporidia species, such as Nosema granulosis and Dictyocoela duebenum, have been demonstrated to exhibit feminising effects on their amphipod hosts (Terry et al. 1999; Ironside et al. 2003; Kelly et al. 2004; Dunn et al. 2006). Kelly et al. (2002) reported that female Gammarus duebeni infected by Nosema granulosis produced female-biased broods, whilst uninfected females had male-biased offspring. Female-biased broods in microsporidia-infected female amphipods were also reported by other studies (Ironside et al. 2003; Mautner et al. 2007). The fact that intersex individuals were only found in the offspring produced by microsporidia-infected females suggested intersexuality was induced by incomplete feminising effects of the microsporidia (Kelly et al. 2004). Ginsburger-Vogel (1991) reported that intersexuality could be induced by grafting tissues from intersex amphipods onto animals presenting normal phenotypes, indicating that the factor causing intersexuality can be readily transmitted between hosts. Various tissues, such as muscle and gonads of Orchestia gammarellus, could be used to induce intersexuality, whilst only testes can be utilised to induce intersexuality in O. aestuarensis or O. mediterranea.

Echinogammarus marinus is an intertidal amphipod that is widely distributed in north-west Europe. Ford et al. (2003) reported two intersex phenotypes, intersex male and intersex female, in E. marinus. The reproductive physiology of intersex males and intersex females follows their normal counterparts, with intersex males having spermatozoa and intersex female producing oocytes. In addition, depending on whether the female sex characteristics could be observed externally (brood plates) or only internally (oviduct structure) in male intersex phenotypes, two further categories of male intersexuality have been subsequently described, namely external and internal intersex males (Yang et al. 2008). Previous studies have observed the occurrence of intersexuality in E. marinus to be strongly related to microsporidia infection using histological techniques; however, the parasite species has not been identified (Ford et al. 2006, 2007).

Preliminary studies into an E. marinus population in Portsmouth Harbour (southern England, UK) revealed different population parameters in relation to sex ratios and prevalence of intersexuality to previous published data. These included male-biased sex ratios and a higher ratio of male to female intersexuality not previously observed in E. marinus populations (Ford et al. 2006, 2007).

This study aimed at assessing the prevalence of intersex phenotypes and their relationship with microsporidia infection. And on the basis of published literature, we made the following four hypotheses/assumptions:
  1. 1.

    Microsporidia were the cause of the observed intersexuality through incomplete feminisation and thus intersex specimens would be more likely infected than normal (non-intersex) specimens.

     
  2. 2.

    If internal intersex males were a prerequisite to having external intersex characteristics and occurred under the same causal mechanism, then all intersex specimens would have similar/equal proportions of microsporidian infection.

     
  3. 3.

    Any microsporidia would be more prevalent in females than males if they were to be considered feminisers and the possible cause of intersexuality.

     
  4. 4.

    A population with feminising microsporidia would most likely be female-biased presuming some males would have converted into fully functional females.

     

Materials and methods

Sampling

A total of 476 E. marinus were collected underneath seaweeds from the intertidal zone nearby Tipner Lane, Portsmouth Harbour (50°49′ 37″N, 1°5′ 43″W), during summer 2009. Animals were sexed after being anaesthetised in carbonated seawater and were initially classified into four sexual phenotypes, namely normal male (NM, n = 299), normal female (NF, n = 151), external intersex male (EIM, n = 21) and intersex female (IF, n = 5), according to the presence of their secondary sex characteristics (Ford et al. 2003). Males were further dissected, and depending on the presence of an oviduct on the testes, internal intersex males (IIM, n = 39) were further categorised (Yang et al. 2008). Intersex females have well developed brood plates, ovaries and one or two genital papillae. Internal intersex males have a normal male appearance but with an oviduct structure on their testes, whilst external intersex males can be distinguished from their rudimental brood plates. A subsample of amphipods (N = 48) were used to assess the prevalence of microsporidia infection using a PCR-based parasite screening technique. In order to assess the transmission route, identification of microsporidia infection was also carried out on the embryos of female amphipods. Due to the low number of intersex female specimens found in Portsmouth Harbour, only four sex phenotypes (NM, NF, IIM and EIM) were statistically investigated for microsporidia infection in this study.

DNA extraction and parasite screening

The genomic DNA was separately purified from muscle and gonadal tissues of each of the amphipod specimens, by employing the DNeasy Blood & Tissue Kit (Qiagen) and following the manufacturer’s protocol. The genomic DNA was quantified by spectrophotometry (Nanodrop, Labtech), and 20 ng of nucleic acid was applied as template to conduct the PCR. Primers amplifying an 890-bp fragment of microsporidia SSU rDNA were adopted from previous studies (18sf, Baker et al. 1995; 981r MacNeil et al. 2003). Thirty-two cycles were carried out on the 20-μl PCR reaction, using a 62°C annealing temperature. A 4-μl aliquot of the PCR was loaded on a 1.0% agarose gel for electrophoresis, and the infection status was determined by the presence of the corresponding band. To investigate the distribution of microsporidia in their amphipod hosts, the parasite screening was carried out on both gonadal and muscle tissues. In order to confirm the quality of purified DNA in samples showing negative microsporidia infection, PCR was conducted using primers (LCO1490-HCO2198, Folmer et al. 1994) to amplify the host’s cytochrome c oxidase I (COI) gene.

Microsporidia species identification

The amplified SSU ribosomal DNA (rDNA) fragment was purified by using the QIAquick PCR purification kit (Qiagen). Samples were sent to the DNA Sequencing Core, Cardiff University, and sequenced by the Sanger method. The SSU rDNA sequence was blasted against GenBank (NCBI) in order to identify the species of microsporidia parasite. Preliminary sequencing results revealed two SSU sequences. Two species-specific reverse primers (BMR and DMR, Table 1) were designed on the most variable site between the two sequences. Both of the reverse primers were applied with forward primer V1f (Weiss et al. 1994). The 48 samples that were originally screened for microsporidia infection were then further examined using the species-specific primers. A total of 20 ng of host genomic DNA was used as template, and 40 cycles were carried out using a 58°C annealing temperature. The infection status was determined by analysing the PCR products using agarose gel electrophoresis as described earlier (2.2).
Table 1

The primers used for amplifying and sequencing microsporidia SSU rDNA, as well as for CO1 gene of amphipod host

Primer

Sequence

Reference

V1f

5′-CACCAGGTTGATTCTGCCTGAC-3′

Weiss et al. (1994)

1342AC

5′-ACGGGCGGTGTGTACAAGGTACAG-3′

Adapted from McClymont et al. (2005)

18sf

5′-GTTGATTCTGCCTGACGT-3′

Baker et al. (1995)

981r

5′-TGGTAAGCTGTCCCGCGTTGAGTC-3′

MacNeil et al. (2003)

LCO1490

5′- GGTCAACAAATCATAAAGATATTGG-3′

Folmer et al. (1994)

HCO2198

5′-TAAACTTCAGGGTGACCAAAAAATCA-3′

Folmer et al. (1994)

DMR

5′-GATTTCTCTTCCGCAATACCAAT-3′

a

BMR

5′-GATTTCTCTTCCGCAATACAGA-3′

a

ITSF

5′-AAAGGAAATTGACGGAGGAACACC-3′

a

ITS2R

5′-TTCAAGGAGTAYYCGARCATC-3′

a

aPrimers designed by this study

For both microsporidia species, the whole small subunit (SSU) and internal transcribed spacer (ITS) region rDNA were amplified within two fragments (V1f-1342AC, annealing temperature 62°C; ITSF-ITS2R, annealing temperature 58°C, Table 1). The fragments were then purified and sequenced as described previously and were then subsequently clustered to retrieve the entire SSU-ITS sequence. All the primers utilised in this study are listed in the following table (Table 1).

Results

Sex phenotype ratio of E. marinus

A total of 476 Echinogammarus marinus were collected from Portsmouth Harbour, by two samplings carried out in June and August 2009. Of these amphipods, three previously described intersex phenotypes were observed, namely intersex female, internal intersex male and external intersex male (Ford et al. 2003; Yang et al. 2008). The number of each sex phenotype is shown in Table 2. There was no significant difference (χ2 = 4.75, df = 4, P = 0.31) observed in the proportions of the five sex phenotypes between the two sampling periods (June and August 2009).
Table 2

The number and proportion of five sex phenotypes of E. marinus specimens collected from Portsmouth Harbour (England) in summer 2009

 

Normal males

Intersex males

Normal females

Intersex females

Total

EIM

IIM

June

117 (59.7%)

10 (5.1%)

13 (6.6%)

54 (27.5%)

2 (1%)

196

August

143 (51.1%)

11 (3.9%)

26 (9.3%)

97 (34.6%)

3 (1.1%)

280

Total

260 (54.6%)

21 (4.4%)

39 (8.2%)

151 (31.7%)

5 (1%)

476

EIM external intersex male, IIM internal intersex male

Male amphipods outnumbered females in both sampling occasions (Table 2). Intersex females were found to be very rare in this site, making up only 1% of 476 amphipods collected. The proportion of internal intersex male (8.2%) was higher than external intersex male (4.4%). Amongst the 12 external intersex males examined in this study, 11 displayed no signs of an oviduct developing on the testes, and only one sample had both internal (oviduct) and external (brood plates) intersex characters.

Microsporidia infection in normal and intersex E. marinus

A subsample of 48 amphipods, with 12 specimens for each of the 4 phenotypes (normal male, normal female, internal intersex and external intersex male), were examined for microsporidia infection by a PCR-based screening method, using a non-species-specific primer set (18sf–981r). Due to the low number found within the field, intersex females were not statistically investigated in this study. According to the band intensity, the infection status of amphipods was classified into three groups, namely high infection intensity, low infection intensity or uninfected (Fig. 1). Both high and low infection were categorised as “infected” but given the considerable difference in band intensity between high and low levels of infection (without intermediates; Fig. 1), we also compared specimens of each sexual phenotype presenting only high levels of infection.
https://static-content.springer.com/image/art%3A10.1007%2Fs00227-010-1573-7/MediaObjects/227_2010_1573_Fig1_HTML.jpg
Fig. 1

Typical examples of band intensities using the PCR-based microsporidia parasite screening method. Lane 1, 2 and 6 were classified as uninfected; 5 and 7 as low infection; 3 and 4 as high infection

The majority of external intersex males (83.7%) were found to be infected, and the proportion of microsporidia infection in internal intersex males and normal males was 50 and 33.3%, respectively (Table 3). The initial parasite screening on muscle and gonadal tissue gave out the identical results for the three male phenotypes. In normal females, a greater proportion of microsporidia infection rate was observed in gonads (50%) than in muscle tissue (25%), but the additional infections in gonads were contributed by individuals with low infection level (Table 3). All the samples showing negative result for microsporidia infection gave out substantial bands for the CO1 gene amplification, suggesting adequate DNA extraction procedures.
Table 3

The number of microsporidian-infected amphipod in 4 sexual phenotypes—normal male (NM), internal intersex male (IIM), external intersex male (EIM) and normal female (NF)—of E. marinus collected from Portsmouth Harbour (England)

Tissue

Sex phenotype

Microsporidia infection

Infection rate (high only)

High

Low

Uninfected

Gonads***

NM

2

2

8

33.3% (16.6%)

IIM

6

0

6

50% (50%)

EIM

10

0

2

83.7% (83.7%)

NF

1

5

6

50% (8.3%)

Muscle**

NM

2

2

8

33.3% (16.6%)

IIM

6

0

6

50% (50%)

EIM

10

0

2

83.7% (83.7%)

NF

2

1

9

25% (16.6%)

** P < 0.01; *** P < 0.001

A significant difference in microsporidia infection was observed amongst the four sexual phenotypes, in both gonadal (Fisher’s exact test, P < 0.001) and muscle tissue (Fisher’s exact test, P < 0.01). To assess where significant differences lie between phenotypes, pairwise Fisher’s exact tests were carried out on the four groups in both tissues. Bonferroni correction was employed, and a significant difference was determined when the P value was below 0.0083. Significant differences were found between external intersex males and normal males (P = 0.005, in both tissues), as well as between external intersex males and normal females (P < 0.001, gonads; P = 0.003, muscle) in both tissues. No significant difference was observed on the proportion of microsporidia infection between any other combinations. The statistical significance of these relationships holds irrespective of whether the categories high and low infection are grouped together as “infected” or we considered highly infected specimens only.

Microsporidia species identification

Six PCR products from the initial parasite screening were sequenced by the Sanger method, and two types of SSU sequence were revealed, sharing 92.9% identity out of 890-bp length sequence. Each SSU rDNA gene was obtained by PCR amplification (primers V1f-1342AC and ITSF-ITS2R) followed by sequencing of the purified fragments. The two overlapped fragments were then clustered, and the sequences were blasted against Genbank (NCBI). One sequence was found to have 98.7% similarity to Dictyocoela berillonum, and the other shared 98.6% identity to Dictyocoela duebenum. A phylogenetic tree was generated by aligning the whole SSU rDNA sequences of the two microsporidium parasites to other microsporidia species from the genus Dictyocoela on GenBank. The two microsporidia species revealed in this study were found to be closely related to D. berillonum and D. duebenum, respectively (NCBI, MEGA 4, Fig. 2). D. berillonum was found in all of the four sexual phenotypes, whilst D. duebenum was only observed in normal females. The proportion of individuals infected by D. berillonum and D. duebenum was 72.9 and 8.3%, respectively, or 37.5 and 6.25% when considering high infection only.
https://static-content.springer.com/image/art%3A10.1007%2Fs00227-010-1573-7/MediaObjects/227_2010_1573_Fig2_HTML.gif
Fig. 2

A phylogenetic tree generated on the SSU rDNA sequence of microsporidia from genus Dictyocoela, and two SSU rDNA sequenced by this study (D. duebenum Portsmouth and D. berillonum Portsmouth). Scale bar 1% sequence divergence

Tissue- and species-specific infection

On the basis of the rDNA sequences, two species-specific primers (BMR, DMR) were designed, which were successfully used to discriminate the two parasites species by PCR. Both primers showed a high level of fidelity at 58°C, and no evidence of cross-binding was observed even after 45 cycles of PCR amplification (Fig. 3). The 48 amphipods initially examined for microsporidia infection were re-screened using species-specific PCR. Of these 48 amphipods, only one microsporidia species, D. berillonum, was found in three male phenotypes, whilst both D. berillonum and D. duebenum were revealed in the normal females (Table 4).
https://static-content.springer.com/image/art%3A10.1007%2Fs00227-010-1573-7/MediaObjects/227_2010_1573_Fig3_HTML.jpg
Fig. 3

PCR products of two samples from a 45-cycle species-specific PCR. The template for sample 1 was genomic DNA from the muscle tissue of a D. duebenum-infected normal female. An intense band was observed from the PCR using DMR (a primer specific to D. duebenum), and no band for BMR (a primer specific to D. berillonum). On the other hand, a bright band was observed for sample 2 (gnomic DNA of muscle tissue from a D. berillonum-infected normal male) when the primer BMR was applied, but no band for DMR. No evidence of cross-binding of the two species-specific primers was observed even after 45-cycle amplification, suggesting a high level of fidelity for both primers

Table 4

The number of amphipods infected by D. berillonum or D. duebenum in 4 sexual phenotypes—normal male (NM), internal intersex male (IIM), external intersex male (EIM) and normal female (NF)—of E. marinus collected from Portsmouth Harbour (England)

Microsporidia species

Tissue

Sex phenotype

High

Low

Uninfected

Infection rate (high only)

D. berillonum

Gonads***

NM

2

6

4

66.7% (16.6%)

IIM

6

2

4

66.7% (50%)

EIM

10

2

0

100% (83.7%)

NF

0

7

5

58.3% (0%)

Muscle***

NM

2

3

7

41.7% (16.6%)

IIM

6

0

6

50% (50%)

EIM

10

0

2

83.7% (83.7%)

NF

0

2

10

16.7% (0%)

D. duebenum

Gonads

NM

0

0

12

0 (0)

IIM

0

0

12

0 (0)

EIM

0

0

12

0 (0)

NF

2

1

9

25% (16.6%)

Muscle

NM

0

0

12

0 (0)

IIM

0

0

12

0 (0)

EIM

0

0

12

0 (0)

NF

3

1

8

33.3% (25%)

** P < 0.01; *** P < 0.001

Both tissues showed significant difference in the D. berillonum infection amongst the four sexual phenotypes (Fisher’s exact test, P < 0.001). Pairwise comparison was made amongst the four sexual phenotypes with a Bonferroni correction (P value = 0.0083). External intersex males had significantly higher proportion of D. berillonum-infected individuals than normal males (Fisher’s exact test, P = 0.003 for both tissues). No significant difference was observed between any other pairwise comparisons (Fisher’s exact test, P > 0.0083).

D. duebenum was only found in normal females, and the different proportion of D. duebenum-infected individuals between normal males and normal females was not significant (Fisher’s exact test, P > 0.0083). Bi-infection by D. berillonum and D. duebenum was observed in gonads of two normal females, whilst the muscle tissue of both females was found to be infected by D. duebenum only. In addition to the statistically studied 48 specimens, three intersex females were screened for parasite infection. All of the three intersex females were found to be infected by D. berillonum, and no D. duebenum was revealed.

In order to assess the transmission route of the two microsporidia species, PCR employing species-specific primers was conducted on the genomic DNA of eggs from one D. berillonum-infected intersex female, one D. duebenum-infected and one uninfected normal female. Eggs were found to have the same parasite infection status as their mothers, with D. berillonum infection in the eggs from the intersex female and D. duebenum in those from the infected normal female. No microsporidium infection was detected on eggs from the uninfected normal female.

Discussion

The aim of this study was to determine the relationship between phenotypic differences in intersexuality and microsporidia infection in an Echinogammarus marinus population in Portsmouth Harbour, southern England.

Over the sampling period, normal and intersex male phenotypes made up 67.2% of the population of sexually dimorphic adults, which is unusual compared to previous reported populations of E. marinus (Ford et al. 2006; Martins et al. 2009). Internal (8.2%) and external (4.4%) intersex males represented approximately 12–13% of the population, which also could be considered noteworthy when all previous reports of intersexuality in E. marinus found intersex females to be the most common phenotype (Ford et al. 2006). Normal (30.7%) and intersex females (1.1%) in the population from Portsmouth Harbour only represented just under a third of the population.

During this study, a PCR-based screening method was used to reveal the presence of microsporidia, and two species-specific primer sets were developed to identify two parasite species from the genus Dictyocoela. Species-specific PCR developed by this study combined the parasite screening and species differentiation into a single reaction and clearly distinguished the two parasite species without requiring a restriction enzyme digestion of the PCR products. Two microsporidia species, D. berillonum and D. duebenum, were revealed coexisting in the E. marinus population in Portsmouth Harbour. D. berillonum was found to correlate with male intersexuality in this amphipod population, and D. duebenum was observed in females only; however, the prevalence and sample size was very low.

Following the previously stated hypotheses that: (1) intersex specimens would more likely be infected than normal specimens; (2) internal and external male intersexes would have similar/equal proportions of infection; (3) females would more likely be infected than males, and (4) the population would be predominantly female biased do not entirely concur with the results observed in this study.

External intersex male specimens were significantly more likely to be infected by D. berillonum than normal males and females, supporting our first hypothesis. Grouping high and low infection together as infected indicates 100% infection in external intersex males, 66.7% infection in normal and internal intersex males, plus 58.3% infection in normal females. However, the majority of infection in normal individuals (males and females) and internal intersex males presented very low band intensities after 40 PCR cycles as represented in Fig. 1. Therefore, considering we observed no evidence of intermediate infection and given the vast differences in band intensity classified as either high or low, we cannot rule out that low infection category represents the presence of background spores in the environment picked up in the PCR. Given this possibility, we decided to also classify low infection alongside those considered uninfected. This revealed no infection in females and a gradient of infection between normal (16.6%) < internal intersex males (50%) < external intersex males (83.3%) both in gonadal and muscle tissue.

These data present some interesting findings with regard to the original hypotheses. Intersex specimens were significantly more likely to be infected than normal specimens, indicating microsporidia as the possible cause of reproductive dysfunction. However, the population is not female biased and neither is there a female bias in infection status as might be expected from having vertically transmitted feminising microsporidia in the population. If internal and external intersexes shared a similar causal mechanism, we might expect them to have an equally high likelihood of being infected by microsporidia. External intersex males were significantly more likely to be infected than normal males, and external intersex males were more likely to be infected than internal intersexes; however, this second result was not significant.

Internal and external intersex males differing in their infection status suggest that one (internal) may not be a prerequisite to the other (external) and that they may occur independently via different mechanisms. Internal and external intersex males have been reported in other E. marinus populations, and the majority of external intersex males were found to have an oviduct developed internally (Yang et al. 2008). In contrast, this study revealed that 11 out of the 12 examined external intersex males did not have the internal intersex characteristics (an oviduct), demonstrating that intersex males possessing both internal and external female characters are comparatively rare in this population. This suggests that internal and external intersex males might be two distinct phenotypes, and the development of external brood plates and internal oviducts in male amphipods is possibly caused by different mechanisms.

Juvenile amphipods have been reported to have undifferentiated gonads with the anlagen of both oviduct and vas deferens (Lockwood 1968). Oviducts in normal male Orchestia gammarella were revealed to degenerate quickly after sex differentiation, but persisted for 2 or 3 moult cycles in young intersex males (Ginsburger-Vogel 1972a, b). Oviducts on the internal intersex male E. marinus from this study is possibly due to delayed degeneration, but specimens with very well developed oviducts indicated that the later development was also possible or that the degeneration of oviducts has been arrested. Brood plates observed in external intersex males are possibly induced by ovarian-hormone-like effects. The development of brood plates (oostegites) has been reported to be controlled by the ovaries in amphipods and isopods (Charniaux-Cotton and Payen 1988; Suzuki and Yamasaki 1997), although to date ovarian hormones have not been identified in Crustacea (Suzuki 1999). It is possible that the oviducts observed in internal intersex males of this E. marinus population were caused by delayed degeneration, whilst the rudimental brood plates in external intersex males were possibly induced by feminising effects of microsporidia parasites or de-masculinisation due to inadvertent disruption of reproductive physiology.

Intersexuality in amphipods has been suggested as a consequence of incomplete feminisation by feminising parasites (Kelly et al. 2002). Contrary to the idea that D. berillonum observed in this study is indeed a feminiser is the lack of female-biased sex ratios and female bias in infection status mentioned earlier. It is possible that D. berillonum exerted a certain level of feminising effects in this E. marinus population but exists as a strong feminiser in another amphipod species. However, to date, despite D. berillonum being reported in a variety of amphipods, it has never been shown to be a potential feminiser. This raises possibilities that intersexuality may occur via a different mechanism or that infection comes subsequent to reproductive dysfunction. Ford et al. (2004) proposed that increased microsporidia infection may occur through immunosuppressed hosts. This theory has recently been supported by Jacobson et al. (2010) who observed increased infection with a closely related microsporidia to D. berillonum in the amphipod Monoporeia affinus during exposure to an industrial surfactant (perfluoroctane sulphonate). Therefore, alternative explanations exist for the link between intersexuality and infection in these populations. Firstly, intersex specimens may have a weakened immune response directly as a result of their internal physiological disruption irrespective of the initial cause of intersexuality (e.g. genetic, pollution induced or as a result of environmental sex determination). Therefore, intersex specimens could have a greater risk of infection than normal specimens. Secondly, intersexuality may be a bi-product of internal disruption caused by horizontally transmitted parasites damaging the internal physiology of male specimens (e.g. their androgenic gland) in a similar fashion to the de-masculination caused by Rhizocephalan parasites in decapods (Høeg 1995).

The two microsporidia parasites revealed in this study have been reported in various amphipods, and the two species share a list of host species, for example Echinogammarus berelloni, Gammarus duebeni and Gammarus tigrinus (Terry et al. 2004). D. duebenum has been revealed as a feminising parasite in some amphipod hosts, such as Gammarus duebeni and Gammarus tigrinus (Terry et al. 2004). In this study, D. duebenum was only found in normal females consistent with this parasite being a vertical transmitter and potential sex-distorter. However, the prevalence and sample sizes were too low to confirm this with any certainty.

Both an adult female infected with D. duebenum and D. berillonum were found to show matching infection in their broods suggesting vertical transmission. According to a theoretical model, multiple purely vertically transmitted parasites are not able to stably coexist in a single host population, since a less efficient species will be excluded by others (Ironside et al. 2003). In fact, the co-occurrence of different vertically transmitted microsporidia species in the same amphipod population has been reported by several studies (Hogg et al. 2002; Ironside et al. 2003; Haine et al. 2004). The empirical evidence of multiple vertical transmission parasites coexisting in a single host population suggests that some assumptions of the theoretical model have been violated, for example the ratio of infected and uninfected hosts might be prevented from reaching equilibrium by seasonal variations (Ironside et al. 2003). Some microsporidia parasites are able to apply both vertical and horizontal transmissions (Dunn and Smith 2001; Smith 2009); however, feminising microsporidia have been reported to be solely vertically transmitted in amphipod hosts (Dunn et al. 2001). It is clear that D. berillonum is more prevalent than D. duebenum and despite our evidence to support vertical transmission in D. berillonum, it is likely, based on our findings and the literature (Terry et al. 2004), that D. berillonum is predominantly horizontally transmitted in E. marinus.

Microsporidia from the genus Dictyocoela have been reported to infect only gonadal tissue of their amphipod hosts (Terry et al. 2004). In contrast, this study revealed the two microsporidia species, D. berillonum and D. duebenum, infect both muscle and gonadal tissues of E. marinus. Generally, D. berillonum infection was more often observed in the hosts’ gonadal tissue than in muscle. However, when considering specimens displaying only high levels of infection, there was an exact match between muscle and gonadal tissues. Interestingly, two normal females showed evidence of bi-infection (via PCR performed using species-specific primers) by D. berillonum and D. duebenum in their ovaries, whilst only D. duebenum was found in their muscle tissue. Although coexistence of multiple microsporidia species in a single amphipod population have been revealed by several studies, the coinfection of different microsporidia parasites in the same host individual was reported to be very rare (Haine et al. 2004; Terry et al. 2004) and would present an interesting model for studying parasites with competing vertical and horizontal strategies.

To summarise, we have developed a PCR-based method for determining two closely related microsporidian species that have been shown to frequency share amphipod host species. The E. marinus population observed in Portsmouth Harbour is interesting in that it harbours both D. berillonum and D. duebenum in the population and on several occasions, the same host. In addition, this E. marinus population is male biased with intersex male phenotypes occurring at a higher frequency than female intersexes. Despite a relationship, between infection and the intersex male phenotypes by the microsporidia D. berillonum, the male bias population and low levels of female infection suggest that the observed intersexuality may be caused via an alternative mechanism to incomplete feminisation. Explanations for the observed results would be firstly, D. berillonum may be a poor feminiser in this amphipod species; secondly, intersexes are more likely to become infected than normal specimens due to suppressed immune systems; thirdly, that intersexuality results from non-specific internal disruption of tissues caused by infection. The majority of the external male intersexes were infected compared to only half of the internal intersex males. This coupled with the fact that the majority of external intersex males lack oviducts on their testes (which are found in internal intersex males) suggests these two phenotypes could result from different causal mechanisms. This amphipod model and population with their varying phenotypes of intersexuality offer unique opportunities for investigating the role of parasites and the environment on sex determination and differentiation.

Acknowledgments

The greatly appreciate the comments of two anonymous referees. This work was conducted at University of Portsmouth and Cardiff University and funded through a studentship awarded to GY from the UHI Millennium Institute ARC Programme (an initiative jointly supported by Highlands and Islands Enterprise, the Scottish Funding Council and the European Regional Development Fund).

Copyright information

© Springer-Verlag 2010