Differential Expression of the p44 Gene Family in the Agent of Human Granulocytic Ehrlichiosis

Abstract

Using reverse transcription-PCR targeting of the p44 genes of the agent of human granulocytic ehrlichiosis (HGE) with primers flanking the hypervariable region, we show differential expression in a murine model of HGE infection and during tick transmission. The p44 genes were differentially expressed in salivary glands of infected nymphal ticks removed during transmission feeding but not in nonfeeding infected ticks. Similarly, the p44 genes were differentially expressed in infected C3H mice, in SCID mice, and in cultured HGE bacteria. Thus, differential p44 expression exists in vivo and in vitro and could provide a basis for antigenic variation.

In areas of endemic infection, human granulocytic ehrlichiosis (HGE) is the second most common tick-borne disease in the United States (1, 3, 6, 10, 13). Recent evidence suggests that the agent of HGE is able to interfere with the innate immune responses during early stages of infection as it invades neutrophilic granulocytes (2, 4, 17) but that specific antibody responses to the HGE agent may provide partial immunity in vivo (21). Antigenic variation of its outer surface proteins may be responsible for ineffective elimination of ehrlichial pathogens, as reported for the surface protein family MSP-2 of Anaplasma marginale, a genetically closely related organism (5, 7-9, 20). The p44 gene family of the agent of HGE shows a high degree of sequence homology to the MSP-2 family (11, 18, 23, 24). We hypothesized that differential expression of the p44 genes may be the basis antigenic variation leading to incomplete elimination by the host. Therefore, we investigated whether p44 genes are differentially expressed at different times during infection in vivo as well as during transmission by the tick vector.

Differential p44 expression during murine infection.

A group of five mice (C3H/HeN) were inoculated intraperitoneally with 100 μl of HGE-infected SCID mouse blood (NCH1 strain) (22). On day 10,500 μl of blood was recovered from each mouse. Infection with the HGE agent was confirmed by visualization of characteristic morulae in 2 to 5% of the peripheral neutrophils (200 neutrophils were counted per mouse). A total of 200 μl of blood from each mouse was used for total RNA extraction. RNA was isolated from white cells from whole blood and/or spleen obtained from infected mice, using an RNeasy isolation kit (Qiagen, Valencia, Calif.). For reverse transcription-PCR (RT-PCR) analysis, RNA was digested with RQ1 DNase (Promega, Madison, Wis.) and the reactions were performed with and without reverse transcriptase (Bioline, Springfield, N.J.) in a total volume of 20 μl, in the presence of RNase inhibitor. One mouse was studied in more detail. RT-PCR using primers HF and HB flanking the hypervariable region (Fig. 1A) showed that one or more p44 members were expressed (Fig. 1B). The primer sequences are HF (5′ CTA CTA GCT AAG GAG TTA GC 3′) and HB (5′ CAC AGA AGT AGA AGA AAC CG 3′), yielding a 537-bp product based on the published hge-44 sequence (11). Primers targeting the 16S RNA gene published previously were used as control (19). PCR products were ligated into pCR2.1 by using a TA cloning kit (Invitrogen, Carlsbad, Calif.). Competent Escherichia coli bacteria were transformed with the ligated vector and plated with 5 mM isopropyl-β-d-thiogalactopyranoside (PTG) for blue-white screening. Twenty inserts from randomly picked colonies were sequenced in both directions, resulting in at least 6 different sequences obtained from the 20 inserts.

FIG. 1.

(A) Schematic representation of the p44 gene family. The hypervariable region (HVR) is indicated with conserved sequences on either side (shaded boxes). Positions of primer sets are indicated by arrows. The 8F and 9B primers are specific for the hypervariable sequence of one p44 sequence, hge-44 (11); primers HF and HB anneal to the conserved portions of the p44 genes and thus will anneal to all p44 sequences; primers CF and CB will amplify a complete p44 sequence. (B) RT-PCR of RNA extracted from a mouse infected intraperitoneally with HGE bacteria, using the primer sets 16S (yielding a 227-bp product) and HF and HB (yielding a 537-bp product). Lanes: M, molecular mass marker; 1, total RNA prior to DNase treatment; 2, RNA after DNase treatment without reverse transcriptase; 3, RNA after DNase treatment with reverse transcriptase; 4, HGE DNA (positive control); 5, no template (negative control).

Differential p44 expression during tick transmission.

Twenty flat nymphs previously infected with NCH-1 were placed on each of a group of five C3H/HeN mice. On day 4 after placement, the fully engorged ticks (but not yet fed to completion) were removed and collected in five groups of 20 ticks from each mouse. Salivary glands of all 20 ticks per group were isolated, pooled in each group, and immersed in RLT buffer (Qiagen) for RNA isolation. RT-PCR was performed as described above. Similarly, RT-PCR was performed on RNA extracted from infected but nonfeeding (unengorged) nymphs. RT-PCR of salivary glands of unengorged nymphs showed a product with the 16S primers but not with the HF and HB primers, suggesting that the p44 gene family is not expressed or is expressed at a very low level (Fig. 2A). RT-PCR of salivary glands of engorged ticks yielded products for both the 16S and p44 primers, suggesting that during tick feeding, the p44 gene mRNAs can be detected (Fig. 2B). RT-PCR products of engorged tick salivary glands were cloned and sequenced, revealing that six different sequences were present within 20 inserts analyzed. Each sequence was present more than once. Subsequently, blood from mice that were infected by tick feeding was collected on days 3 and 10 after the ticks were removed. Sequence comparisons of the p44 RT-PCR products from ticks and the infected mice are shown in Fig. 3.

FIG. 2.

(A) RT-PCR of RNA extracted from salivary glands from nonfeeding ticks infected with the HGE agent. Lanes: M, molecular mass marker; 1, no template (negative control); 2, total ehrlichial DNA (positive control); 3, 4, and 8, tick RNA without reverse transcriptase; 5 to 7 the same tick RNA as lanes 3, 4, and 8 but with reverse transcriptase. (B) RT-PCR of RNA extracted from salivary glands of ticks infected with HGE, during active tick feeding. Lanes: 1, molecular mass marker; 2, total RNA prior to DNase treatment; 3, total RNA after DNase treatment but without reverse transcriptase; 4, RNA after DNase treatment with reverse transcriptase; 5, total ehrlichial DNA (positive control); 6, no template (negative control).

FIG. 3.

Protein sequence comparisons based on DNA sequences of cloned RT-PCR products. The p44 sequence of HGE-44 (AF037599) was used as consensus sequence. Six different sequences from a total of 20 sequenced inserts were identified from tick salivary glands (T1 to T21); the sequences from infected mice on days 3 and 10 are designated by C3H. The asterisk indicates that sequence T6 was identified as the same sequence as AF029323 (18). Arrows represent the primer positions in the corresponding DNA sequence. Dots denote sequence identity, and dashes denote gaps in the sequence to improve the overall alignment. Sequential GenBank accession numbers are AY112683 through AY112692.

HGE-44-MBP vaccination studies.

In a separate experiment, 10 C3H/HeN mice were vaccinated with HGE-44-MBP, a recombinant p44 protein fused to maltose-binding protein (MBP), and 10 control mice were vaccinated with MBP only (12). Mice vaccinated with HGE-44-MBP showed a strong antibody response (titer, ≥1:10,000) to p44 proteins of cultured HGE organisms by immunoblotting, while the (MBP-only) control group did not show any antibody reactivity to p44 proteins (21). Infected nymphal ticks were placed on both groups of mice (five ticks per mouse) and allowed to feed to repletion. On day 15 after tick placement, mouse blood was obtained for peripheral smear analysis and for DNA extraction and PCR. Of the 10 mice in the control group, 9 were infected, whereas of the 10 mice in the experimental group, 8 were infected. These results indicate that immunization with HGE-44-MBP did not provide protection against HGE infection. Blood was also collected on day 25 after tick placement for RNA extraction and RT-PCR. In all infected mice (nine in the control group and eight in the experimental group), RT-PCR using the HF and HB primers was positive in all cases, demonstrating expression of one or more p44 genes. RT-PCR products from one infected mouse in the vaccinated group were sequenced, and several different sequences were identified. The use of primers 8F (5′ TCA AGA CCA AGG GTA TTA GAG ATA G 3′) and 9B (5′ GCC ACT ATG GTT TTT TCT TCG GG 3′) annealing specifically to the hypervariable region of the hge-44 sequence (16) did not yield any product (data not shown), suggesting that mRNA identical to that for the protein used for vaccination was not present (the coding sequence of HGE-44-MBP is based on the published HGE-44 sequence) (11).

Differential p44 expression in culture and SCID mice.

RT-PCR was performed on total RNA from cultured ehrlichiae. Sequence analysis of the cloned RT-PCR products showed that 15 different p44 genes were expressed and that they remained expressed 20 days later (GenBank accession numbers AF512484 through AF512498). Further, differential expression of the p44 gene family was investigated in C3H/SCID mice to study the effect of the absence of an intact immune system in vivo with regard to the p44 expression. Two SCID mice (SC1 and SC2) were infected intraperitoneally, and blood was collected on days 10 and 30 for RNA extraction followed by RT-PCR and by cloning and sequencing of the RT-PCR products (for SC1, see GenBank numbers AF512671 through AF512678; for SC2, see GenBank numbers AF512679 through AF512683). As in the C3H/HeN mice, several different p44 genes were expressed at both time points (AF512671, AF512672, AF512675, AF512676, and AF512678 in SC1 and AF512679, AF512680, and AF512682 in SC2), and on day 30 three new different p44 sequences were present in SC1 (AF512673, AF512674, and AF512677), and two new sequences were present in SC2 (AF512681 and AF512683) that had not been identified on day 10 postinoculation, demonstrating differential expression over time of p44 genes in vivo in the absence of an intact immune system.

In this study we have investigated the in vivo differential expression of the p44 gene family of the agent of HGE in the tick as well as in the murine host. Several p44 transcripts are present in ticks during transmission feeding but not in nonfeeding ticks, suggesting that tick engorgement induces the expression of p44 genes. This is consistent with the finding that HGE bacteria require a period of about 36 h of attachment before infecting the mammalian host (14). Second, in the ensuing infection in the mice, different p44 genes were expressed at various time points and some of these were different from the ones that were expressed in the ticks during transmission. These data seem to support the hypothesis that continued differential expression of the p44 genes could serve as the basis for antigenic variation. Antigenic variation may provide a mechanism to escape host responses during infection of the vertebrate host, leading to delayed or incomplete clearance. Furthermore, differential p44 expression occurred in infected SCID mice as well as in in vitro culture, even though fewer different p44 transcripts were identified in SCID mice, suggesting an active recombination mechanism leading to a constant generation of new p44 variants independently of immune pressure.

Vaccination with one p44 protein (HGE-44-MBP) does not provide protection against subsequent exposure of ehrlichial organisms, despite a well-developed antibody response to both the conserved regions and the variable region of the HGE-44-MBP. However, vaccination of mice with HGE-44-MBP and subsequent challenge with HGE bacteria prevents the detection of mRNA for this specific p44 variant, suggesting that vaccination with a p44 protein can influence p44 gene expression. Thus, vaccination with a combination of several p44 proteins may suppress the expression of several p44 genes and hence may affect the overall p44 expression. Perhaps a combination of multiple peptides from within the HVR may be necessary to induce greater protection.