Frequency of Secondary Symbiont Infection in an Invasive Psyllid Relates to Parasitism Pressure on a Geographic Scale in California

Abstract

Two endosymbionts, an obligate primary symbiont and a facultative secondary symbiont, are harbored within the invasive red gum (eucalyptus) lerp psyllid, Glycaspis brimblecombei, in California. An extensive survey of diversity and frequency of G. brimblecombei's secondary symbiont in multiple populations throughout the state of California was conducted using PCR detection, restriction enzymes, cloning, and sequencing. A total of 380 G. brimblecombei individuals in 19 populations were screened for secondary symbionts. Based on molecular screening results, only one type of secondary symbiont was present in G. brimblecombei populations in California. Overall, 40% of the 380 psyllids screened were infected with the secondary symbiont. Interestingly, secondary symbiont infection frequencies in G. brimblecombei populations varied dramatically from 0 to 75% and were significantly related to parasitism pressure by Psyllaphaegus bliteus, a solitary endoparasitoid of the psyllid.

Insect bacterial endosymbionts are harbored in a wide diversity of insects (2, 7, 14, 32). Despite the prevalence of endosymbionts throughout many insect orders, little is known about the role and maintenance of endosymbionts within insect hosts. In general, an endosymbiont-host association can be obligate for survival and/or reproduction of the host (termed a primary symbiont here) or facultative and therefore not necessary for host survival and/or reproduction (termed a secondary symbiont [SS] here) (7, 26). The role of primary symbionts to their host is thought to be nutritional enrichment (4, 5, 14, 15, 19, 20, 27). Alternatively, the roles of facultative SSs are not well understood (2, 3). Nevertheless, SSs have recently been found to exert a variety of effects on the phenotype of Acyrthosiphon pisum Harris (Hemiptera: Aphidae) (8, 17, 18, 25, 29, 35).

Little is known about the roles SSs play within their insect hosts other than A. pisum. Although SSs are known to occur in 33 psyllid species (26, 30, 34), no information has yet been published on the frequency of SSs within and among psyllid populations or their effect on psyllid phenotypes. It is important to understand insect-endosymbiont interactions because endosymbionts may play major evolutionary roles in eukaryotes by influencing interspecific interactions, distributions, and fitness during ecological time scales in populations.

In the state of California, the invasive red gum lerp psyllid Glycaspis brimblecombei Moore (Hemiptera: Psylloidea), its preferred and prevalent host plant Eucalyptus camaldulensis Dehnhardt (Myrtales: Myrtaceae) (6), and its parasitoid Psyllaphaegus bliteus Riek (Hymenoptera: Encyrtidae) provide an ideal system for the study of endosymbiont-host interactions on a geographic scale. Eucalyptus spp. were introduced from Australia to California as seeds around 1850 (13); E. camaldulensis is the most abundant host plant species of G. brimblecombei and has been planted throughout the state (13). In 1998, G. brimblecombei invaded California from Australia (24) and quickly became established throughout the range of its host plant. The psyllid has caused serious defoliation of red gum eucalyptus, leading to significant tree death in some populations (9, 23, 24). A statewide release effort of G. brimblecombei's parasitoid, P. bliteus, was implemented in 2001 to control this psyllid pest. Based on 3 years of pre- and postrelease data, successful establishment of P. bliteus was found in some G. brimblecombei populations, but not in others, despite repeated introductions (9). Consequently, variation in parasitism pressure by P. bliteus naturally exists in G. brimblecombei populations in California. In addition, G. brimblecombei is known to possess at least one type of secondary endosymbiont in California (26, 30, 34). In light of what has been found in the A. pisum-endosymbiont system, where a SS of A. pisum induces resistance to a parasitoid wasp (21, 22), it is important to investigate the distributions and frequencies of potential SSs in G. brimblecombei, especially if SSs influence the efficacy of biological control against this psyllid pest.

The objective of this study was to characterize the variation of secondary endosymbiont(s) within and among G. brimblecombei populations in California. In addition, important abiotic and biotic selection pressures that vary in G. brimblecombei populations are related to observed SS infection frequencies in populations.

MATERIALS AND METHODS

Insect and environmental data collection.

Psyllid individuals were collected from 19 localities in California (Table 1 and Fig. 1) from 19 to 29 June 2006. All psyllids were collected from the host tree E. camaldulensis, except for two populations in which psyllids were collected from Eucalyptus rudis (Riverside and San Juan Bautista). At each locality, over 400 infested leaves were collected from multiple branches of at least three trees. The only exceptions to this were at Atascadero and Weldon where psyllids were present at low densities, and we were able to collect only 40 infested leaves from two trees at each location. For SS infection frequency analysis, to prevent sampling of siblings, only one live third- to fifth-instar psyllid was collected per infested eucalyptus leaf. Glycaspis brimblecombei nymphs are relatively sessile and are covered by a carbohydrate and wax dome-shaped structure (a lerp) produced by the nymph. From each location, 20 psyllids were randomly chosen to determine the frequency and variation of SS within and among California populations. Only unparasitized psyllids, based on morphological observation and dissection under a stereomicroscope, were used in SS infection frequency analysis. Psyllids were preserved immediately after collection in 95% ethanol and subsequently stored at −20°C. From two populations, Riverside and Tustin, several parasitoids were also preserved to check for the presence of SSs in P. bliteus.

TABLE 1.

Locations in California where G. brimblecombei was collected in June 2006

Site no.a	Locationa	Latitude (N)	Longitude (W)	Level of infestationb	SS infection frequency	Total parasitism (%)c
1	Sylmar	34°19′	118°28′	High	0.40	7.5
2	Ojai	34°22′	119°18′	Moderate	0.55	15
3	Atascadero	35°3′	120°5′	Light	0.70	−
4	San Juan Bautista	36°5′	121°33′	Very high	0.45	8.5
5	Gilroy	37°	121°33′	Very high	0.70	48.5
6	Ardenwood	37°487′	122°03′	Very high	0.45	22
7	Redwood City	37°29′	122°12′	High	0.20	7
8	Sonoma	38°15′	122°27′	Moderate	0.30	0
9	Shasta	40°36′	122°29′	Moderate-high	0.25	27
10	Sacramento Wildlife Refuge	39°26′	122°11′	Very high	0.10	6
11	Sacramento	38°27′	121°22′	Very high	0.05	16
12	Fresno	36°43′	119°55′	Light-moderate	0.10	0
13	Bakersfield	35°23′	119°06′	High	0.0	13.5
14	Riverside	33°58′	117°2′	Moderate	0.40	20
15	Tustin	33°45′	117°46′	Moderate	0.75	72.5
16	Escondido	33°7′	117°6′	Moderate	0.65	40.5
17	Quivira Basin	32°46′	117°14′	High	0.55	27.5
18	Scripps Ranch	32°53′	117°9′	Moderate	0.65	46.5
19	Weldon	35°39′	118°20′	Light	0.35	−

^aLocations and site numbers shown in Fig. 1.

^bThe level of infestation was qualitatively measured (see Materials and Methods for details) at the time of collection.

^cPercent parasitism of G. brimblecombei by P. bliteus was measured from 200 lerps (a − sign indicates that the sample size was too small to measure total parasitism).

FIG. 1.

Locations in California where G. brimblecombei was collected in June 2006. The locations are indicated by pie graphs. The dark portions within pie graphs represent the SS infection frequency (n = 20 psyllids per location). Locations where psyllids were not present during sampling in 2006 are indicated (x). The numbers to the right of the pie graphs are the site numbers shown in Table 1. The map is modified, with permission, from a University of California, Agriculture and Natural Resources, map.

Percent parasitism was estimated at each location based on random samples selected from the 400 collected leaves. For each location, 200 occupied lerps (or unoccupied lerps if a parasitoid exit hole was present) were examined for the presence of parasitoids (9). Parasitism was not measured in the leaves from the Atascadero and Weldon sites due to the small sample sizes.

The level of infestation was qualitatively observed at each location and is based on two factors; the number of lerps that occur per leaf and the distribution of infested leaves present throughout the canopy of the tree. Qualitative classifications are as follows: light, 1 to 10 lerps per leaf and patchy distribution of infested leaves; moderate, 10 to 20 lerps per leaf and the majority of leaves (3/4) in the canopy covered with lerps; high, 20 to 30 lerps per leaf and nearly all leaves covered by lerps; and very high, >30 lerps per leaf, severe defoliation, and all leaves covered with lerps. Environmental data, including yearly average air temperature (minimum, maximum, and average), humidity (minimum, maximum, and average), dew point (minimum, maximum, and average), and total precipitation, were obtained for each locality using a state government weather database (http://www.cimis.water.ca.gov/cimis/data.jsp). Also, latitude and longitude coordinates were obtained from each location using a global positioning system unit (Garmin, Ltd., Olathe, KS). Environmental data were related to infection frequency data using stepwise regression with SPSS statistical software (31). Ratings of the level of infestation were converted to dummy variables for use in regression. All response variables were tested for normality, homogeneity of variance, and multicollinearity. Proportional data were arc-sine square root transformed prior to analysis.

Detection of a secondary endosymbiont(s).

Detection and characterization of a secondary symbiont(s) involved several techniques including PCR, gel electrophoresis, digestion of PCR products with restriction enzymes, and subsequent cloning and sequencing of the restriction fragments. Total DNA was extracted from individual psyllids and wasps using the Chelex-based method detailed by Vavre et al. (37). Initial screening of symbionts began with eubacterium-specific primers (standard 10F and the universal reverse primer 480R) (28) (Table 2) using Taq DNA polymerase (NEB, Ipswich, MA). These primers were selected because they do not amplify G. brimblecombei's primary endosymbiont, “Candidatus Carsonella ruddii,” which is an obligate endosymbiont that occurs in all psyllid individuals (2). Primers 10F and 480R amplify a fragment that spans the intergenic spacer region (ISR) between 16S and 23S rRNA subunits. In principle, bands of different lengths, i.e., resulting from variation in the ISR, will separate when subjected to gel electrophoresis, indicating the presence of more than one SS or a single SS with multiple ribosomal operon regions. PCR was performed in 25-μl reaction mixtures containing 2 μl of DNA template (concentration not determined), 0.2 mM each deoxynucleoside triphosphate, 0.2 μM each primer, 1× PCR buffer (NEB), and 1 U Taq polymerase (NEB). Amplifications were performed in a Mastercycler 5331 (Eppendorf, Hamburg, Germany) programmed as described by Sandström et al. (28). Amplified DNA was visualized after electrophoresis on a 1% agarose gel stained with ethidium bromide (run at ∼4.9 V/cm for at least 1 h). In all cases (of infected psyllids), this yielded a single visible band ∼2.5 kb in length. To guard against false-negative results, primers that are specific to the 28S D2 rRNA of insects were used to confirm successful DNA extraction (Wolbachia FIBR Project [http://troi.cc.rochester.edu/∼wolb/FIBR/]). A known positive control (PCR and automated sequence-confirmed SS extracted DNA) and negative control (double distilled water) were also included in all assays.

TABLE 2.

Diagnostic primers used in this study for screening eubacteria (excluding the primary symbiont “Candidatus Carsonella”) and the secondary symbiont of G. brimblecombei

Primer	Primer sequence (5′-3′)	Target symbionta	Target gene	Product size (kb)	Reference
10F	AGTTTGATCATGGCTCAGATTG	Universal eubacteria	16S-23S gene	2.502	28
480R	CACGGTACTGGTTCACTATCGGTC				28
Gb1F	GGTTGAACAAAAGGGCGTTA	SS Gb	Partial ClpB protease gene	0.752	This study
Gb1R	CCATGTGTAGCGGTGAAATG		Partial 16S gene		This study
Gb2F	ACAGAAGAAGCACCGGCTAA	SS Gb	Partial 16S gene	1.105	This study
Gb3R	TTGTCTGATGAAATAGCGCG		Partial ISR		This study
gyrBF	TAARTTYGAYGAYAACTCYTAYAAAGT	Universal eubacteria	Gyrase subunit B gene	0.932	11
gyrBR	CMCCYTCCACCAARGTAMAGTTC				11

^aSS Gb, secondary symbiont of G. brimblecombei.

To verify the identities of secondary endosymbionts, PCR products of the entire 16S rRNA gene of a single SS-infected psyllid from each population were directly sequenced using two overlapping primer sets (Gb1F-Gb1R and Gb2F-Gb3R) (see Table 2). Again, PCR was performed in 25-μl reaction mixtures with the same composition of reagents detailed above. Thermocycler conditions with primer sets Gb1F-Gb1R and Gb2F-Gb3R were as follows: an initial denaturing step of 94°C for 5 min; followed by 40 cycles of 94°C for 1 min, 58°C for 1 min, and 72°C for 2 min; and a final extension step of 72°C for 5 min. Amplified DNA was cleaned using the Wizard PCR Preps DNA purification system (Promega, Madison, WI) and directly sequenced in both directions at the University of California Riverside Genomics Institute Core Instrumentation Facility using an Applied Biosystems 3730 DNA analyzer with a Big-Dye V3.1 kit (Applied Biosystems, Foster City, CA).

Since more than one ribosomal operon (and therefore more than one PCR template) may be expected in the SS of G. brimblecombei (in addition to other eubacteria of similar chromosome sizes) (10; N. Moran, personal communication), additional assays were conducted to more effectively characterize the secondary endosymbiont. Since some SSs cannot be cultured, this extra characterization during the screening method is warranted to ensure that only one SS is present (versus falsely claiming that two SSs are present due to two different rRNA PCR products). A potential problem with Taq DNA polymerase is its propensity to preferentially amplify one template at the expense of others, e.g., templates which have a lower titer or are longer than the “amplified” template (in the present study this may be further confounded, since Taq DNA polymerase only amplifies template effectively up to 3 kb in length). However, so-called proofreading DNA polymerases have a much higher affinity for such “additional” templates. Therefore, Phusion DNA polymerase (Finnzymes Oy, Espoo, Finland) was used to amplify these potential templates in a single infected and uninfected psyllid individual (as detected by our initial screening; see above) from each population. Reagents for PCR using Phusion polymerase with 10F and 480R primers include 20-μl reaction mixtures containing 0.4 μl of DNA template (concentration not determined), 0.2 mM each deoxynucleoside triphosphate, 0.25 μM each primer, 1× Phusion GC buffer (Finnzymes), and 0.4 U Phusion DNA polymerase (Finnzymes). Reaction conditions for Phusion polymerase using 10F and 480R primers are as follows: an initial denaturing step of 98°C for 30 seconds; followed by 35 cycles of 98°C for 10 seconds, 50.1°C for 30 seconds, and 72°C for 1.5 min; and a final extension step of 72°C for 10 min. After this assay, the same results were attained when both Phusion polymerase and Taq polymerase were used, suggesting that larger or lower-titer templates were not present.

In a further assay, the Phusion-amplified 10F and 480R PCR product of a SS-infected psyllid individual was run on a 0.7% agarose gel at ∼1.7 V/cm. After 14 h, two bands of slightly different sizes could be distinguished; they were between 2.5 and 2.6 kb in length. This PCR product was then cloned using the GeneJET PCR cloning kit (Fermentas International Inc., Burlington, Ontario, Canada). Over 30 clones were screened via gel electrophoresis, and only two band lengths were consistently found. To obtain complete sequences for these two products, clones representing both band lengths were first cut using the restriction endonuclease Eco88I (Fermentas). This yielded four restriction fragments between 400 bp and 1,000 bp, which were again cloned and then sequenced. In addition, a single-copy gene, DNA gyrase subunit B (gyrB), was amplified from this psyllid individual using eubacterium-specific gyrB primers (Table 2) (11). Reagents for PCR using Phusion polymerase with gyrB primers are the same as described above for Phusion polymerase with 10F and 480R primers. Reaction conditions for Phusion polymerase using gyrB primers are as follows: an initial denaturing step of 98°C for 30 seconds; followed by 35 cycles of 98°C for 10 seconds, 53°C for 30 seconds, and 72°C for 42 seconds; and a final extension step of 72°C for 10 min. The gyrB PCR product was cloned as described above, and 10 clones were sequenced to ensure that only one secondary endosymbiont species was present.

Nucleotide sequence accession numbers.

Nucleotide sequences were deposited in GenBank and given accession numbers as follows: DNA gyrase subunit B (gyrB) clones, EU036653 to EU036662; rRNA operon product clones detected using primers 10F and 480R, EU30463 to EU30464; and the PCR product from direct sequencing of two overlapping primer sets (Gb1F-Gb1R and Gb2F-Gb3R) of the 16S rRNA gene, EU043378.

RESULTS

Variation, frequency, and distribution of SSs in G. brimblecombei populations.

Only one SS was found in 19 populations of G. brimblecombei. Sequences of the entire 16S rRNA gene were identical for infected individuals in all populations. Based on electrophoresis, cloning, restriction enzymes, and sequencing (see Materials and Methods) two 16S-23S ribosomal RNA gene operons are present within “the secondary endosymbiont of G. brimblecombei” (34). These two rRNA operons primarily differ in their ISRs, particularly the tRNA genes they encode. One rRNA operon possesses an ISR (ISR1) encoding a tRNA^Glu alone, and the second rRNA operon possesses an ISR (ISR2) encoding both tRNA^Ile and tRNA^Ala. The former rRNA intergenic spacer region (ISR1) has 100% sequence identity with the sequence given GenBank accession number AF263561 (“the secondary endosymbiont of Glycaspis brimblecombei”) (34), and the latter rRNA intergenic spacer region (ISR2) has 97% sequence identity with the sequence given GenBank accession number AY264663.1 [Arsenophonus endosymbiont of Acanthaleyrodes styraci Takahashi (Hemiptera: Aleyrodidae) (33)]. All sequences derived from overlapping primer sets Gb1F-Gb1R and Gb2F-Gb3R were identical, since these primers were specific for amplifying only one type of rRNA operon (i.e., the short rRNA operon, ISR1). The single-copy gene gyrB was identical for all 10 clones derived from an infected psyllid individual.

Interestingly, SS infection frequency varies dramatically in psyllid locations. SS infection frequencies range from 0 to 75% with an overall average of 40% infection in 19 locations (Fig. 1). In total, out of 380 psyllids, 152 tested positive for the SS. By using primers 10F and 480R and electrophoresis, no SS was found in P. bliteus.

SS infection frequencies in populations in relation to environmental gradients.

Based on step-wise regression with 14 environmental variables, SS infection frequencies are significantly and positively related only to total parasitism (F = 21.740, df = 16, P < 0.0005, R = 0.769, adjusted R² = 0.565, standard error of estimate = 0.1760, β = 0.952 ± 0.204) (Fig. 2). Other environmental variables, such as the level of infestation, latitude, longitude, total precipitation, three humidity variables, three dew point variables, and three air temperature variables, were not significantly related to infection frequency. That is, in populations with high parasitism pressure, there is a high probability of sampling psyllid nymphs (i.e., individuals that either escaped or resisted parasitism, since unparasitized individuals were screened) that harbor the SS. However, in populations with low or no parasitism, there is a high probability of sampling unparasitized psyllid nymphs that do not harbor the SS (Fig. 2).

FIG. 2.

Relationship between SS infection frequency and total parasitism from 17 G. brimblecombei locations in June 2006. For infection frequency, there were 20 psyllids per location; for parasitism, there were 200 occupied and unoccupied lerps (see text for details).

DISCUSSION

There was a positive relationship on a geographic scale between SS infection frequency and parasitism pressure in field populations of G. brimblecombei. Whether the secondary endosymbiont of G. brimblecombei confers resistance or induces susceptibility to parasitism in G. brimblecombei cannot be determined from our data. This question will be the subject of continuing investigations. Nevertheless, resistance of an infected psyllid to a parasitoid may be an important parameter to test in manipulated trials, because it has been found in A. pisum that SSs confer resistance to parasitoids and fungal entomopathogens (16, 21, 22, 29). The proposed mechanism inducing resistance in A. pisum to the parasitoid Aphidius ervi may be associated with the bacteriophage designated APSE-2 which is intimately associated with the SS “Candidatus Hamiltonella defensa” (19, 22). Interestingly, based on our preliminary data, the secondary endosymbiont of G. brimblecombei is associated with a similar phage (GenBank accession number EU039462), which has 91% sequence identity to A. pisum bacteriophage APSE-2 DNA polymerase (P45) based on BLASTN 2.2.16 (1) using APSE-2 DNA polymerase primers 30.6f and 31.9r (12).

In regard to other environmental variables in this study, average annual temperature variables did not significantly relate to SS infection frequency in G. brimblecombei populations in California. Likewise, infection frequencies of S. symbiotica in the pea aphid were not correlated with temperature in populations in California (18) or Japan (36). Serratia symbiotica is known to confer heat tolerance and resistance to a parasitoid (22) in A. pisum; therefore, multiple abiotic and biotic selection pressures on a geographic scale may determine infection frequencies for the pea aphid.

Corroboration of laboratory and field studies have shown that the aphid SS “Candidatus Regiella insecticola” (PAUS, U-type) is significantly related to host plant specialization. Initially, a geographic cline was found in the frequency of “Ca. R. insecticola” infection in A. pisum and was significantly correlated with host plant species, temperature, and amount of precipitation in populations across Japan (36). In laboratory assays, Tsuchida et al. (35) found that “Ca. R. insecticola”-infected A. pisum displayed higher fecundity than uninfected aphids on an unsuitable host plant (white clover). Furthermore, they found that there was a higher occurrence of “Ca. R. insecticola”-infected aphids on white clover than on vetch at three locations in central Japan, where white clover and vetch occurred sympatrically.

In the present study, psyllids were collected from their most favorable and prevalent host plant in California, E. camaldulensis. Glycaspis brimblecombei is known to colonize a variety of red gum eucalyptus species (6), but the most ubiquitous red gum species that has been planted in its introduced range throughout the state of California is E. camaldulensis. Consequently, even if G. brimblecombei's SS influences host plant specialization, it may not be a major influence on G. brimblecombei's infection frequencies in California due to a lack of host tree species diversity.

In summary, G. brimblecombei is infected with one type of SS in California. In addition, this SS possesses two types of rRNA operons. Also, infection frequencies of this SS vary dramatically in populations. This variation in infection frequency of the SS of G. brimblecombei is positively related to parasitism pressure on a geographic level in field populations.