Dual silencing of long and short Amblyomma americanum acidic chitinase forms weakens the tick cement cone stability

Ticks are important ectoparasites that transmit diverse animal and human disease agents such as protozoa, bacteria, spirochaetes and viruses (Jongejan and Uilenberg, 1994) and are considered second to mosquitoes in terms of the impact on medical and veterinary transmitted diseases (Sonenshine, 1993; Jongejan and Uilenberg, 2004). Amblyomma americanum (Linnaeus) is an important tick distributed in North America. This tick transmits multiple tick-borne disease (TBD) agents, Ehrlichia chaffensis (Anderson et al., 1993), E. ewingii (Wolf et al., 2000), Francisella tularensis (Taylor et al., 1991), an unknown causative agent of the disease called southern tick associated rash illness (STARI) (James et al., 2001), Rickettsia amblyommii (Apperson et al., 2008) and Theileria cervi (Laird et al., 1988). Amblyomma americanum has also been linked to transmission of the recently described human tick-borne Heartland virus (Savage et al., 2013). The distribution of A. americanum has expanded from southeastern United States (Mixson et. al., 2006) to northeastern regions as far as Maine (Keirans and Lacombe, 1998). The expanding geography and its role as a vector of numerous human TBD agents makes this tick important in public health policy. Currently methods of tick control and prevention of human and animal TBD infections is dependent on use of chemical acaricides. While acaricides are effective in the short term, they do not provide a permanent solution due to multiple limitations such as ticks developing resistance to acaricides, environmental contamination, cost of developing new acaricides, and the inconvenience of application procedures. These limitations have necessitated the search for alternative novel tick control methods that will provide a permanent solution (Graf et al., 2004; de la Fuente and Kocan, 2006; de la Fuente et al., 2007). Immunization of animals against tick infestation has been validated as a sustainable alternative tick control method (Opdebeeck et al., 1988; Willadsen, 2004; de la Fuente et al., 2010). The pre-requisite to this is a deeper understanding of tick feeding biology and physiology as a means to discover weak links in tick biology that can be targeted for tick vaccine development. In our laboratory we are studying molecular events of early stage tick feeding physiology that precedes key facets of tick parasitism, TBD agent transmission, blood meal uptake, and reproduction. Towards this goal subtractive hybridization analysis was used to identify 40 A. americanum (Aam) genes that were differentially up-regulated in ticks that were preparing to start feeding (Mulenga et al., 2007). One of these 40 genes is the focus of this study, and is a putative A. americanum tick acidic chitinase (AamAch).

Chitinases are enzymes that hydrolyse the β-1,4 glycosidic linkages of N-acetylglucosamines primarily found in chitin. Chitin is an insoluble structural polysaccharide that is important as a supporting element in arthropod exoskeleton (Neville et al., 1976), fungal and bacterial cell walls (Debono and Gordee, 1994; Gomaa, 2012), microfilaria sheath (Araujo et al., 1993) and the lining of the digestive tracts of many arthropods (Souza-Neto et al., 2003; Zimoch et al., 2005; Khajuria et al., 2010). Arthropod development and morphogenesis rely on remodeling chitin and in the process requires chitin synthases and chitinases to control this process (Merzendorfer and Zimoch, 2003). Insect chitinases are also involved in cuticle turnover, digestion and degradation of peritrophic membrane during molting (Kramer and Muthukrishnan, 1997; Merzendorfer and Zimoch, 2003; Rao et al., 2004). In fungi, chitinases were involved in autolysis, nutrition, morphogenesis and parasitism (Ghormade et al., 2000). Most bacterial chitinases are involved in degrading chitin to provide nitrogen and carbon (Patil et al., 2000). In plants, chitinases were involved in defence (Gooday, 1999) and morphogenesis (Grover, 2012), while in nematodes, chitinases were associated with development (Geng et al., 2002; Tachu et al., 2008). Mammals may not synthesize chitin, but the human genome encodes eight chitinases that are involved in T-cell-mediated inflammation and asthma (Kawada et al., 2007; Reese et al., 2007) and tissue remodeling and injury (Lee et al., 2011). Emerging data suggest that chitinases are important to tick physiology. In the tick Haemaphysalis longicornis, immunization of rabbits with recombinant virus expressing a chitinase was used as a bioacaricide, which caused reduced feeding efficiency (You et al., 2003; Assenga et al., 2006), and prevented molting (You and Fujisaki, 2009). The purpose of this study was to gain insight into the role(s) of a tick feeding stimuli responsive putative acidic chitinase (Ach) in A. americanum (Aam) tick feeding physiology. We demonstrate that A. americanum expresses the long and short AamAch and that both forms may be important to maintenance of stability of A. americanum tick cement cone.

A partial cDNA encoding putative acidic chitinase was identified among A. americanum feeding stimuli responsive genes (Mulenga et al., 2007). In this study rapid amplification of cDNA ends (RACE) was used to amplify the full-length cDNA (not shown). While analysing DNA sequences of cDNA clones, we identified long (L) 1959 base pair (bp) (GenBank accession no. KF819831) and short (S) 1718 bp (GenBank accession no. KF819830) forms, with AamAch-S characterized by a 210 bp deletion (Fig. 1A). AamAch-L and AamAch-S full-length cDNA respectively contain 1332 and 1104 bp open reading frame (ORF) that encodes 443 and 367 amino acid residue proteins (Fig. 1B). To verify if the 210 bp deletion was physiological and not a sequencing error, specific AamAch-L and AamAch-S polymerase chain reaction (PCR) primers were designed around the deleted region, as indicated in Fig. 1A. Forward primers were designed to specifically anneal Aam Ach-L or AamAch-S with the reverse primers annealing to both forms. Consistent with the observed 210 bp deletion (Fig. 1A), PCR primers amplified ~500 and 300 bp AamAch-L and AamAch-S cDNA fragments, respectively (Fig. 2), confirming that both forms were expressed in A. americanum.

When subjected to amino acid motif scanning on the ScanProsite Expasy server (de Castro et al., 2006) AamAch-L and AamAch-S did not retrieve any matches (not shown). However, when scanned against entries in GenBank, both forms showed identity to annotated chitinase and chitinase-like proteins in the glycoside hydrolase-18 (GH-18) family with amino acid identity levels ranging from 40 to 85% (not shown). When scanned on the SignalP server version 4.01 (Petersen et al., 2011), a 17 amino acid residue putative signal peptide was predicted in AamAch-L, but not AamAch-S (Fig. 1B). The GH-18 family is characterized by: (1) signal peptide, (2) Ser/Thr-rich linker regions that may be heavily glycosylated located in the N- or C-terminal region linking the signal peptide, catalytic domain and chitin binding domain (Arakane et al., 2003; Huang et al., 2012), (3) one to several GH-18 catalytic domains that have conserved DXDXE and/or SXGG motif (Henrissat, 1991; Watanabe et al., 1993; Tsuji et al., 2010), and (4) none to five chitin binding domains characterized by the CX₆₇-FX₆₇ -CX₂ -YX₇ -CX₉₁₉-CX₄₆-FX₄₇-CX₅₁₂-C motif (Fuhrman et al., 1995; Tellam, 1996). In arthropods, the typical GH-18 chitinase catalytic domain is characterized by four signature amino acid sequence motifs including active site ‘FDG(L/F)DLDWE(Y/F)P’ containing the conserved DXDXE motif (Thomas et al., 2000; Lu et al., 2002), ‘K(F/V)M(V/L/I)AVGGW’ containing the conserved SXGG motif, ‘M(S/T)YDL(R/H)G’ and ‘GAM(T/V)WA(I/L)D’ (Arakane and Muthukrishnan, 2010). Except for putative chitin binding domains, the four consensus amino acid motifs and putative glycosylation sites are present in both AamAch-L and AamAch-S (Fig. 1B). Of the four amino acid motifs, three (‘LDGVDMAWPFP’, ‘QAYDLRG’ and ‘GVYVRNND’) are conserved up to 55, 67 and 38%, respectively, while the fourth motif ‘KTLLSVGFW’ is 71% conserved in AamAch-L but absent in AamAch-S. As shown in Fig. 1B, both AamAch-L and AamAch-S have 2 and 1 putative N-linked (circled) and 24 and 6 O-linked glycosylation sites (marked as bold lowercase letters) located in the N-terminal region, which may indicate a S/T linker region connecting the signal peptide with the catalytic domain.

Fig. 3 summarizes the phylogeny relationship of AamAch-L and AamAch-S putative proteins with other tick GH18-like chitinases downloaded from GenBank and Uniprot (marked with [+]). The tree out-rooted from Homo sapiens GH-18 chitobiase (accession number: AAA35684.1) was constructed by the maximum likelihood method set to default parameters in the Molecular Evolutionary Genetics Analysis (MEGA) 5.2.2 online software (http://www.megasoftware.net) (Fig. 3). As shown in Fig. 3, sequences segregated into three clusters: A, B, and C supported by 94, 85 and 77% bootstrap values, respectively. Both AamAch-L and AamAch-S amino acid sequences segregated with Rhipicephalus sanguineus (ACX33152.1), Amblyomma maculatum (G3MSH7, G3MSG3), Ixodes scapularis (XP_002407798.1, XP_002407799.1) and R. pulchellus (L7MD51). Within cluster A, AamAch-L and AamAch-S show a respective 86 and 83% amino acid identity to R. sanguineus (ACX33152.1). However, when compared with remaining sequences, amino acid identity levels decreased to between 34 and 51% in cluster A and 24–45% in clusters B and C (not shown).

Qualitative reverse transcription PCR (qRT-PCR) using PCR primers in Fig. 1A revealed that both AamAch-L and AamAch-S mRNA were expressed through the first 5 days of feeding in all tested tick organs (not shown). Attempts to find optimal AamAch-S qRT-PCR primers for AamAch-S were not successful. Given the importance of the salivary gland and mid-gut to tick physiology, we determined quantitative AamAch-L mRNA expression levels in these organs (Fig. 4). For qRT-PCR analysis, we synthesized cDNA of dissected salivary glands (SG), mid-gut (MG) and carcass (CA) (tick remnant after removal of SG and MG) pooled from ten unfed and 24 h fed, eight 48–96 h fed, and five 120 h fed ticks in triplicate. As summarized in Fig. 4, AamAch-L displayed a dichotomous expression pattern. In MG, AamAch-L mRNA is significantly down-regulated in response to tick feeding activity dropping to near zero by the 120 h time point. In contrast, in the SG and CA (Fig. 4), AamAch-L mRNA expression levels did not significantly change through the 120 h feeding time point when compared with unfed ticks. Please note that exceptions to this were between unfed versus 24 h (P=0.0377) or 96 h (P=0.0477) in SG, and unfed versus 72 h (P=0.0277) in CA time points where transcript abundance was significantly higher in unfed ticks.

We are interested in characterizing extracellular proteins, and thus expression of recombinant (r) AamAch-S, predicted to function in intracellular protein, was not pursued. Fig. 5 summarizes expression and affinity purification of rAamAch-L in Pichia pastoris. Pilot expression showed a progressive daily increase in rAamAch-L expression levels, with the highest level observed at day 5 (Fig. 5A). Spent media of a 3 l large-scale expression was precipitated by ammonium sulphate saturation at 4°C and affinity purified using NiCl₂⁺ charged Hi-Trap column (GE Healthcare Biosciences, Piscataway, NJ, USA). Following purification, elution fractions with relatively pure rAamAch-L as judged by SDS-PAGE with silver staining were pooled and concentrated using JumboSep™ centrifugal device filters (Pall Corporation, Port Washington, NY, USA). Concentrated proteins were subjected to SDS-PAGE on 10% acrylamide gel followed by silver staining to verify purity as shown in Fig. 5B. The predicted molecular mass of rAamAch is approximately ~53 kDa: comprised of ~2.5 kDa N-terminus fusion, ~47 kDa of mature rAamAch protein and ~3.5 kDa of C-terminal tag containing the hexa-histidine sequence. However, when rAamAch was electrophoresed on SDS-PAGE and developed by silver staining, a band at about 70 kDa is observed (Fig. 5B). Sequence analysis shown in Fig. 1B revealed that AamAch-L had two and 24 putative N- and O-linked glycosylation sites, respectively. Thus there is a possibility that the observed band shift from ~53 to ~70 kDa was due to post-translational glycosylation of rAamAch-L. When rAamAch-L was treated with a deglycosylation mix that cleaves off both N- and O-linked glycans, there was a downshift in the molecular weight to about ~60 kDa (Fig. 5C), confirming that rAamAch-L was glycosylated.

Western blotting analysis of rAamAch-L using antibodies to 48 h (Chalaire et al., 2011) and replete fed (Mulenga et al., 2013) A. americanum tick saliva proteins was carried out to investigate the possibility of native AamAch-L being injected into the host during tick feeding. The expectation is that if native AamAch-L was immunogenic and injected into the host during tick feeding, rAamAch-L would specifically bind antibodies to 48 h and/or replete fed A. americanum tick saliva proteins. However, rAamAch-L non-specifically bound non-immune sera of rabbit (Fig. 6), bovine (Fig. 6) and chicken (not shown) as well as rabbit antibodies to 48 h and replete fed (Fig. 6) A. americanum tick saliva proteins. It is interesting to note that Rhipicephalus microplus saliva proteome contained a homolog to AamAch-L (Tirloni et al., 2014), while ongoing work in our laboratory has also found AamAch-L in A. americanum saliva proteome (T.K.K., L. Tirloni and A.M., unpublished observations), confirming that AamAch-L is a secreted tick saliva protein.

To determine if native AamAch-L was an active chitinase, rAamAch-L was subjected to chitinase function assay using a colorimetric substrate chitinase assay kit. There was no apparent chitinase activity against exochitinase activity substrates, 4-nitrophenyl N,N′-diacetyl-β-d-chitobioside and 4-nitrophenyl N-acetyl-β-d-glucosaminide or the endochitinase activity substrate, 4-nitrophenyl β-d-N,N′,N′′-triacetylchitotriose. The positive control chitinase from Trichoderma viride provided in the assay kit showed activity against all substrates (not shown). The substrate hydrolysis assay buffer in the kit was at pH 4.8. To investigate the possibility that the kit's reaction buffer pH was not optimal for rAamAch-L activity, we assayed at a broad pH range of 4–6 as described by Carmody (Carmody, 1963). Under these conditions no activity was observed for rAamAch-L, while the positive control T. viride chitinase showed activity to all substrates (not shown).

To investigate the significance of AamAch-L and AamAch-S in tick feeding physiology, we synthesized double stranded RNA (dsRNA) targeting the region that was conserved between both forms as summarized in Fig. 7. Please note that the RNA interference (RNAi) silencing experiment was done twice. Results summarized in Fig. 8 represent qRT-PCR validation of AamAch-S/L mRNA suppression in three randomly sampled ticks that were injected with AamAch-S/L-dsRNA from the second experiment. AamAch-S/L mRNA was apparently suppressed by 77.5–91.2% in the salivary gland (SG), 79.9–96.6% in the synganglion (SYN), 56.3–92.3% in the mid-gut (MG), and 0–91.2% in Malpighian tubules (MT). Silencing in the ovary (OV) and carcass (CA) were 32.2–54.3 and 0–73%, respectively. It is interesting to note that while in tick 1 AamAch-S/L mRNA was suppressed in the SG, SYN, MG and OV, no silencing occurred in the MT and CA.

Investigation of tick feeding parameters revealed that silencing of AamAch-S/L mRNA did not affect the ability of ticks to attach onto host skin and start feeding because by visual inspection at 24 h after being placed on rabbits, 100% green fluorescent protein (GFP)-dsRNA and AamAch-S/L-dsRNA-injected ticks were attached (not shown). The tick RNAi-mediated silencing feeding phenotypes were documented every 24 h with pictures (Fig. 9). Fig. 9A,B presents documentation of physical phenotypes of AamAch-S/L-dsRNA and GFP-dsRNA injected ticks feeding on the rabbit host in two independent experiments. In both experiments leakage of blood around the mouthparts was observed by day 13 post-attachment in AamAch-S/L-dsRNA injected ticks (indicated by red arrows in Fig. 9A,B). No leakage of blood was observed in GFP-dsRNA injected control ticks. In the first experiment, bleeding was so intense that by the day 15, veterinarians at the Comparative Medicine Program (CMP) facility advised to halt the experiment due to the uncontrolled seepage of blood. Therefore, all the remaining ticks on the host were removed. In the second experiment bleeding had continued, but it was not as intense to stop the experiment when compared with the first experiment. While manually detaching AamAch-S/L-dsRNA injected ticks, they appeared to be weakly attached onto the host skin, in that ticks would fall off the host with a gentle touch.

As an index to measure the amount of blood imbibed by ticks, engorgement mass (EM) of replete fed and spontaneously detached AamAch-S/L-dsRNA and GFP-dsRNA injected ticks were determined. EM of AamAch-S/L-dsRNA and GFP-dsRNA injected ticks ranged from 362.7 to 751.2 mg (N=7) and 145.4 to 777.9 mg (N=7), respectively. Although there is an apparent difference between mean EM of AamAch-S/L-dsRNA (441.1±92.27 mg) and GFP-dsRNA (504.2±53.66 mg) injected ticks (not shown), Student's unpaired t-test showed that differences were not statistically significant (P=0.5656). To determine the effect of RNAi-mediated silencing on fecundity, ticks were allowed to oviposit for 30 days and egg clutches were weighed. As an index of the tick's ability to convert its blood meal to eggs, the egg mass conversion ratio (EMCR) was calculated by dividing the egg weight (not shown) by EM. Calculated EMCR for GFP-dsRNA and AamAch-S/L-dsRNA injected ticks ranged from 594.9 to 449.1 mg (N=3) and 532.7 to 258.2 mg (N=3), respectively. While the EMCR of AamAch-S/L-dsRNA (417.0±82.12 mg) injected ticks is apparently smaller than GFP-dsRNA injected ticks (539.3±45.52 mg) (not shown), the observed difference is not statistically significant as revealed by Student's unpaired t-test (P=0.2626). Please note that, although seven spontaneously detached ticks were incubated to lay eggs from each treatment group, only three ticks each from control and treatment groups were able to oviposit, and the remaining four ticks did not.

A putative acidic chitinase-like encoding cDNA fragment was found among feeding stimuli responsive A. americanum genes (Mulenga et al., 2007). In this study we show that A. americanum constitutively and ubiquitously expresses the long and short putative acidic chitinase (Ach) forms. AamAch-L and AamAch-S were provisionally identified as members of the GH-18 family based on amino acid identity to previously annotated GH-18 chitinases, and conservation of amino acid motifs that characterize the GH-18 family (Arakane and Muthukrishnan, 2010). Based on data mining in this study, the tick GH18-like chitinases are a multi-member family. The observation that with exception of R. sanguineus GH-18 chitinase sequence (ACX33152.1) that showed >80 amino acid identity, AamAch-L and AamAch-S amino acid identities to other tick GH-18 chitinase sequences were ≤50% may suggest that tick GH-18 chitinases regulate multiple non-redundant pathways. At the time of this write-up genome sequence data for many tick species was not available, with exception of I. scapularis, and thus the comparative sequence and phylogeny analyses presented here may be incomplete.

Expression of AamAch-L and AamAch-S mRNA in multiple tick organs observed in this study signals the possibility that native AamAch-L and AamAch-S proteins could be important to tick physiology in general. Our quantitative RT-PCR analysis data are suggestive of the possibility that AamAch-L proteins could be associated with tick feeding physiology at the salivary gland level, and not at the mid-gut level. In tick research, genes that are up-regulated in response to feeding are thought to be associated with blood meal feeding (Mulenga et al., 2007; Aljamali et al., 2009; Mulenga and Khumthong, 2010; Konnai et al., 2011) and those that are down-regulated are believed to play other roles in tick biology, and are not associated with blood meal feeding (Umemiya et al., 2008; Aljamali et al., 2009). Given that AamAch-L mRNA abundance in the SG did not apparently change with feeding could signal the importance of this protein in unfed and during the parasitic stages of A. americanum. We were unable to quantitatively determine AamAch-S expression levels, and the relationship of AamAch-S to the tick feeding cycle remains unknown.

The observation in this study that dual silencing affected the stability of A. americanum tick cement was significant. Seepage of blood around mouthparts, and the fact that ticks easily detached with a light touch was suggestive of malformed or weakened tick cement cone, thus suggesting that AamAch-L and/or AamAch-S were important to tick feeding success. Within 5–30 min of attaching onto the host, a tick secretes an amorphous adhesive substance called cement to anchor itself onto host skin (Sonenshine, 1993). Secretion of tick cement is by far one of the most significant biological adaptations that make ticks successful as pests and vectors. Without tick cement, hard ticks will easily be groomed off the host before they complete feeding, acquiring and/or transmitting disease agents. Hard ticks are long-term blood feeders that remain securely attached onto host skin for long periods of time, 4–7 days for larva and nymphs, and 10–14 days for adult ticks with the help of an intact cement plug (Sonenshine, 1993). Without the cement cone, ticks can be easily groomed off the host. The compositions of tick cement or molecular cascades that lead to tick cement formation remain unknown. Thus the finding that AamAch-S/L could be associated with tick cement stability advances our knowledge in this regard. Finding ways to disrupt tick cement is an attractive target for development of novel tick control methods. It is important to note here that leakage of blood was not observed in all AamAch-S/L-dsRNA injected ticks. We are of the opinion that this could be explained by the possibility that ticks did not receive the same AamAch-S/L-dsRNA dosage. This is supported by our RNAi silencing validation data that show that AamAch-S/L mRNA was disrupted to different levels in different tick individuals.

Another notable observation in this study is that the observed effect of silencing is relatable to temporal and spatial transcription of AamAch-L. Bleeding around mouthparts in AamAch-S/L-dsRNA injected ticks was observed at day 13, late in the feeding cycle. Interestingly, transcription analysis show an initial decrease in AamAch-L transcript abundance during the first 3 days of feeding, before it increased, albeit not statistically significant in salivary gland and other tissues. Based on these observations it is logical to speculate that AamAch-L and potentially AamAch-S are functionally important late during the tick feeding process. It is also interesting to note that four of the seven ticks that were incubated to lay eggs did not oviposit for both GFP and AamAch-S/L-dsRNA injected ticks. We would like to note here that our injection site was right behind the fourth leg coxa. The tick ovary loops around the body near to our injection site, and thus there is potential that we caused traumatic injury to ovaries. This could have affected egg development.

We are interested in understanding the role(s) of tick saliva proteins in tick feeding physiology. Thus we wanted to functionally characterize rAamAch-L, which is predicted to be extracellular. Consistent with bioinformatics prediction of multiple putative N- and O-linked glycosylation sites, rAamAch-L was significantly glycosylated. This meant that rAamAch-L was potentially well folded with appropriate post-translational modifications, and thus functional analysis data could potentially reflect in vivo events. Data in this study suggest that native AamAch-L is potentially an inactive GH-18 chitinase. Within the GH-18 family functional domain amino acid motif ‘FDG(L/F)DLDWE(Y/F)P’, aspartic acid (D) and glutamic acid (E) (highlighted in bold) amino acid residues were thought to be important in activity of GH-18 chitinases (Arakane and Muthukrishnan, 2010). It is notable that in AamAch-L predicted functional domain (LDGVDMAWPFP), ‘D’ is replaced by ‘A’, while ‘E’ is replaced by ‘P’. Further investigations are needed to determine if these mutations are indeed important for AamAch-L function. The observation here is consistent with reported studies where GH-18 chitinase-like proteins in Tribolium castaneum, Autographa californica and Manduca sexta that have mutations similar to AamAch-L were not functional chitinases (Thomas et al., 2000; Lu et al., 2002; Arakane and Muthukrishnan, 2010). It is also important to note that H. longicornis tick chitinase (accession number: BAC06447.1) reported in the phylogeny analysis retains amino acid residues that are important for GH-18 chitinase function (You et al., 2003). It is interesting to note that all the sequences within cluster B and C contained aspartic acid (D) and glutamic acid (E) for the functional amino acid motif FDG(L/F)DLDWE(Y/F)P (not shown). However, sequences in cluster A did not retain the aspartic acid (D) and glutamic acid (E). This may imply that ticks encode both a functional and non-functional GH-18-like chitinases. It is also interesting to note that both AamAch-L and AamAch-S do not have putative chitin-binding domains, which are thought to bind chitin prior to its degradation. The absence of the domain does not affect the ability of the enzyme to hydrolyse the soluble substrate triacetylchitotriose, but abolishes its ability to hydrolyse insoluble chitin (Tjoelker et al., 2000). Whether or not the absence of these sequence features caused the lack of chitinase activity in rAamAch-L remains to be investigated.

Our western blotting analysis experiments to validate if native AamAch-L was immunogenic and injected into the host during tick feeding was inconclusive, in that we observed that rAamAch-L bound both pre-immune and tick saliva protein immune sera. Interestingly, in an ongoing study in our group, we have found a putative acidic chitinase in an I. scapularis tick that shows 61% amino acid identity to AamAch-L (T.K.K., L. Tirloni and A.M., unpublished observations) and have verified the presence of AamAch-L in unfed and fed A. americanum tick saliva proteomes (T.K.K., L. Tirloni and A.M., unpublished observations). Additionally in a recent study, Tirloni et al. (Tirloni et al., 2014) found a putative acidic chitinase that shows 79% amino acid identity to AamAch-L in saliva of R. microplus. Based on these observations we have concluded that AamAch-L is a tick saliva protein that may be part of the tick's system to eliminate host immunoglobulins during feeding. There is precedence in R. appendiculatus where host antibody elimination was documented (Wang and Nuttall, 1994).

In conclusion this research has made a contribution towards understanding the molecular basis of the tick attachment phase. This research has opened up opportunities to further investigate the role(s) of putative AamAch-L and AamAch-S in maintaining the stability of the tick cement cone. Important questions that remain to be resolved are if AamAch-L is part of the tick machinery to eliminate host antibodies, and if it is part of the cement cone or associated with upstream tick cement formation cascades. Given the apparent importance in maintenance of tick cement cone stability it will be interesting to investigate if this protein or its interacting molecular partners can be targeted for tick control. Based on data mining in this study, it is also apparent that the tick GH-18 family has multiple members. Thus to understand the role(s) of the GH-18 family in tick physiology, these genes must be investigated as a group.

Amblyomma americanum ticks were purchased from tick laboratories at Texas A&M University and Oklahoma State University (College Station, TX, USA and Stillwater, OK, USA, respectively). Routinely, ticks were fed on rabbits according to animal use protocols approved by Texas A&M University Institutional Animal Care and Use Committee. To feed, A. americanum ticks were placed onto the outer part of the ear of specific pathogen-free (SPF) New Zealand rabbits. Ticks were restricted onto the outer ears using orthopedic stockinette glued onto the rabbit ear with Kamar Adhesive (Kamar Products Inc., Zionsville, IN, USA). Six male ticks were pre-fed for 3 days prior to introducing 15 female ticks on each ear stocking (total of 30 female ticks per rabbit).

Two rounds of dissections were done to produce material for qualitative and quantitative RT-PCR analysis. In the first round of dissections to prepare material for qualitative RT-PCR, five ticks per time point were sampled at 24, 48, 72, 96 and 120 h. Tick organs including salivary glands (SG), mid-guts (MG), ovary (OV), synganglion (SYN) Malpighian tubules (MT) and carcass (CA, the remnants after removal of other organs) were isolated and placed in 1 ml of the RNA extraction solution, Trizol (Life Technologies, Carlsbad, CA, USA) and stored at −80°C until total RNA extraction. In the second round of dissections to prepare material for quantitative RT-PCR analysis, triplicate pools of 10, eight and five ticks were dissected at the unfed and 24, 48–96 and 120 h fed time points, respectively. Tick organs SG, MG and CA were pooled in 1 ml Trizol and stored at −80°C until total RNA extraction. Within the first hour of detachment, ticks were prepared for dissections as described by Mulenga et al. (Mulenga et al., 2013). Prior to dissecting, tick mouthparts were inspected to remove any remaining rabbit tissue and washed in RNase inhibitor diethylpyrocarbonate (DEPC)-treated water.

Total RNA was extracted using the Trizol reagent according to the manufacturer's instructions (Life Technologies), resuspended in DEPC-treated water, and quantified using a UV-VIS Spectrophotometer DU-640B (Beckman Coulter, Brea, CA, USA). Total RNA (1 μg) was used to synthesize cDNA using the Verso cDNA Synthesis Kit following the manufacturer's instructions (Thermo Scientific, Waltham, MA, USA). The synthesized cDNA was quantified as above and stored at −20°C until use.

The full-length cDNA sequence of putative AamAch cDNA was cloned using the RACE technique (ClonTech, Mountain View, CA, USA) according to the manufacturer's protocol using 120 fed tick cDNA. PCR fragments were routinely cloned into the pGEM-T TA cloning vector (Promega, Madison, WI, USA) and sequenced with T7 and SP6 promoter primers. Sequences were analysed using the MacVector program (MacVector Inc., Cary, NC, USA).

Preliminary sequence analysis identified AamAch-S with a 210 bp nucleotide deletion between position 817–1026 bp of AamAch-L. To determine if both AamAch-L and AamAch-S mRNAs were expressed, respective specific forward 5′-GAGGGGAGTCTGGAACGGAG-3′ and 5′-CGCCCAAGCCCACGCCCTCGTC-3′ and the universal reverse 5′-GAAGTCGTCCAGGTCGTTGTTG-3′ PCR primers were designed as illustrated in Fig. 1A. The AamAch-S specific forward primer overlapped the deletion site, while the AamAch-L forward primer was designed inside the deletion site.

BlastX and BlastP homology scanning of AamAch-L and AamAch-S at the National Center for Biotechnology Information and UniProt were used to identify protein sequences in other organisms that showed identity to AamAch. Scanning on SignalP version 4.1 (Petersen et al., 2011) and NetNGlyc 1.0 and NetOGlyc 4.0 (Steentoft et al., 2013) servers detected presence of a signal peptide and putative N- and O-linked glycosylation sites, while scanning in GenBank and visual inspection of motifs described in Arakane and Muthukrishnan (Arakane and Muthukrishnan, 2010) detected annotated amino acid motifs. Calculated molecular weights were determined using the MacVector (MacVector Inc.) protein toolbox.

Tick GH-18-like chitinases in public databases GenBank (N=5) and UniProt (N=5) sequence were downloaded and used to construct a guide phylogeny tree using MEGA 5.2.2 online software (http://www.megasoftware.net). Based on preliminary blast results, AamAch-L and AamAch-S were categorized as a GH-18 chitinase-like protein, therefore Homo sapiens (accession number: AAA35684.1) GH-18 chitobiase was used as an out-group in the phylogeny tree. Sequences were aligned using ClustalW of ~350 amino acid conserved cysteine amino acid residue region C_70–417 and the phylogeny tree constructed using the maximum likelihood method using the Whelan and Goldman (WAG) model option (Whelan and Goldman, 2001). To estimate bootstrap values, replications were set to 1000.

Qualitative and quantitative RT-PCR analyses determined temporal and spatial mRNA profiles. In qualitative RT-PCR, AamAch-L and AamAch-S specific PCR primers (Fig. 1A) were used with cDNA of 24, 48, 72, 96 and 120 h fed tick dissected SG, MG, OV, SYN and CA. For both putative AamAch-L and AamAch-S, PCR cycling conditions were an initial denaturing step at 94°C followed by 30 amplification cycles of 94°C for 30 s, 53°C for 30 s, 72°C for 1 min, and a final extension step of 72°C for 5 min. PCR products were electrophoresed on 2% agarose gels containing 1 μg ml⁻¹.

In quantitative RT-PCR analysis triplicate pools of unfed, 24, 48, 72, 96 and 120 h tick SG, MG and CA were analysed using an Applied Biosystems 7300 Real Time PCR System (Life Technologies) using specific AamAch-L forward (5′-CTGGAACGGAGTTACCGATGTGC-3′) and reverse (5′-TGCGGACCGATATCAATGGAT-3′) qRT-PCR primers. In preliminary experiments efforts to find optimal qRT-PCR primers for AamAch-S failed, and thus this analysis was abandoned. Reaction volumes in triplicate contained 10-fold diluted cDNAs that were originally synthesized from 1 μg total RNA, 350 nmol l⁻¹ of forward and reverse AamAch-L primers, and 2X SYBR Green Master Mix (Life Technologies). For internal control, forward (5′-GGCGCCGAGGTGAAGAA-3′) and reverse (5′-CCTTGCCGTCCACCTTGAT-3′) primers targeting a 55 amplicon of A. americanum 40S ribosomal protein S4 (RPS4; accession number: GAGD01011247.1) that is stably expressed in I. scapularis during feeding (Koči et al., 2013) was used.

Relative quantification (RQ) of AamAch-L transcript was determined using the comparative C_T (2^−ΔΔC_T) method (Livak and Schmittgen, 2001) on a Microsoft Excel spreadsheet (Microsoft Corporation, Redmond, WA, USA). Based on the C_T values of the three biological replicates for each dissected tissues, the ΔC_T values were determined by the formula C_T (AamAch) – C_T (RPS4) per time point. Subsequently, the ΔΔC_T values were determined using the formula ΔC_T (replicate) – C_T (average of replicates per tissue) per time point for each individual replicate. The fold change was determined for each replicate in each tissue by using the formula 2^−ΔΔC_T and plotted on an Excel spreadsheet to generate a chart. Statistical analyses were performed using non-parametric Student's t-tests with Prism software (GraphPad Software, San Diego, CA, USA).

Recombinant AamAch-L protein was expressed using the P. pastoris and pPICZα plasmid expression system (Life Technologies). The expression plasmid was constructed by subcloning the mature AamAch-L coding domain into pPICZαA KpnI and NotI sites, using forward (5′-GGTACCATGCCCCAGCAAGACGGGGATG-3′) and reverse (5′-GCGGCCGCAGCGGGTGATGCGGTAATCTCG-3′) primers with added restriction enzyme sites in bold. The pPICZαA-AamAch-L expression plasmid was linearized with PmeI and used to transform P. pastoris X-33 strain (Life Technologies) by electroporation as described previously (Mulenga et al., 2013). Likewise, induction, validation and affinity purification of recombinant protein expression were done as described previously (Mulenga et al., 2013). Affinity purified putative rAamAch-L was dialysed against 1×phosphate-buffered saline (PBS) at pH 7.4 for downstream assays. Routinely, affinity purified rAamAch-L was resolved on a 10% SDS-PAGE gel and silver stained to verify purity and background contamination. Samples with least background were selected and concentrated by either ammonium sulphate precipitation or by centrifugation using MicroSep Centrifugal Concentration Devices (Pall Corporation).

To determine if rAamAch-L was N-glycosylated and/or O-glycosylated, 2.5 μmol l⁻¹ affinity purified rAamAch-L was treated with protein deglycosylation enzyme mix according to the manufacturer's instructions (New England Biolabs, Ipswich, MA, USA). Deglycosylation was verified by western blotting analysis using the antibody to c-terminus hexa histidine-tag (Life Technologies) and the positive signal detected using horseradish peroxidase (HRP) chromogenic substrate (Thermo Scientific).

To determine if native AamAch-L is injected into the host during feeding, affinity purified rAamAch-L was subjected to routine western blotting analyses using antibodies to 48 h and replete fed A. americanum tick saliva proteins. Antibodies used here were produced by repeated tick infestation of rabbits every 48 h (Chalaire et al., 2011) and repeatedly allowing ticks to feed to repletion (Mulenga et al., 2013). Our preliminary analysis showed that rAamAch-L non-specifically reacted with both pre-immune (PI) serum and immune sera to tick saliva proteins. To validate if rAamAch-L non-specifically binds immunoglobulins, deglycosylated and glycosylated rAamAch-L was further subjected to western blot analyses using PI sera of chicken and bovine. In all western blotting analyses, 1:2000 antibody dilutions were used.

Chitinase function assays were done by substrate hydrolysis using a commercial kit from Sigma (St Louis, MO, USA; catalogue no. CS0980). Substrates used were nitrophenyl N,N′-diacetyl-β-d-chitobioside and 4-nitrophenyl N-acetyl-β-d-glucosaminide for exochitinase activity and 4-nitrophenyl β-d-N,N′,N′′-triacetylchitotriose substrate for endochitinase activity. Prior to beginning the assay, standard and substrate solutions were equilibrated to 37°C by incubating for 10 min in a 37°C water bath. The pre-warmed substrates 50 μl (1 mg ml⁻¹) were mixed with rAamAch-L (1.91 and 0.477 μmol l⁻¹), and positive (0.8 and 0.2 μmol l⁻¹) or negative (1.91 and 0.477 μmol l⁻¹) controls into a 96-well plate. Validated T. viride active chitinase was used as a positive control. The standard reaction was placed into separate wells on the plate and the whole plate was mixed by shaking in the plate reader and incubated for 30 min at 37°C. The reaction was stopped by addition of 100 μl of stop solution (0.4 mol l⁻¹ sodium carbonate) and the end-point was measured at A₄₀₅ using the Infinite M200 Pro plate reader (Tecan, Männedorf, Switzerland) pre-warmed to 37°C. The experiment was also repeated using different mixtures of 0.1 mol l⁻¹ citric acid (A) and 0.2 mol l⁻¹ dibasic phosphate (B) buffer at pH values of 4 (61.45 ml A plus 38.55 ml B), 5 (48.5 ml A plus 51.5 ml B), 5.4 (44.25 ml A plus 55.75 ml B) and 6 (36.85 ml A plus 63.15 ml B) as described by Carmody (Carmody, 1963).

RNAi-mediated silencing was performed as previously published (Mulenga et. al., 2008). dsRNA targeting both AamAch-L and AamAch-S was synthesized in vitro using the Megascript RNAi kit (Life Technologies) according to the manufacturer's protocol. A 466 bp cDNA fragment from position 155–620 of AamAch-L and 136–602 AamAch-S cDNA sequences (see Fig. 7) was used for AamAch-S/L-dsRNA synthesis. Templates for dsRNA synthesis were amplified using cloned AamAch-L plasmid DNA using specific primers with added T7 promoter sequence in bold (forward: 5′-TAATACGACTCACTATAGGGCCACGTCAAGCACCACCACC-3′ and reverse: 5′-TAATACGACTCACTATAGGGGCGTAGTTTTCCTTATCGCG-3′). Enhanced green fluorescent protein coding cDNA (EGFP; accession number: JQ064510.1) was amplified as described previously (Mulenga et al., 2008). Two test groups of 15 female A. americanum ticks were injected with GFP-dsRNA or AamAch-S/L-dsRNA. Ticks were injected with 0.5–1 μl (~3 μg μl⁻¹) of dsRNA on the ventral side of the lower right coxa using a 33-gauge/0.5 inches/45 deg angle bevelled needle using a model 701 Hamilton syringe (Hamilton Company, Reno, NV, USA). Injected ticks were kept for 24 h at room temperature in 85% humidity to recover before being placed on SPF New Zealand rabbits to feed.

Ticks injected with AamAch-S/L-dsRNA (N=15) and GFP-dsRNA (N=15) were fed on rabbits as described above. Effect of RNAi silencing on tick feeding success was assessed by investigating tick attachment and mortality rates, time to feed to repletion, EM as an index for amount of blood taken in by tick, and EMCR as a measure of utilizing blood meal to produce eggs. Attachment rates were determined by daily counts of unattached ticks, and then subtracting from the total number of ticks that were placed on the animal less those that were found dead. Mortality rates were determined by calculating dead ticks as a fraction of the total number of ticks. EM was the weight of spontaneously detached ticks. To determine EMCR, ticks were allowed to lay eggs for 3–4 weeks at room temperature in 85% humidity. EMCR was determined by dividing egg mass by EM. Tick phenotypes during feeding were documented daily using the Canon EOS Rebel XS camera attached to a Canon Ultrasonic EF 100 mm 1:2:8 USM Macro Lens (Canon USA Inc., Melville, NY, USA). This was repeated one more time.

To validate if injection of dsRNA caused disruption of AamAch-S/L mRNA, three ticks each of GFP dsRNA and AamAch dsRNA injected females were sampled at 48 h post-attachment by manual detachment. Ticks were processed individually for dissection of tick organs as described above. Dissected organs, SG, MG, SYN, MT, OV and remnants labelled as CA, preserved in 200–400 μl RNAlater (Life Technologies), were processed for mRNA extraction using the Dynabead mRNA Direct Kit (Life Technologies). Concentration of mRNA was determined using the NanoQuant Plate in the Infinite M200 Pro plate reader (Tecan) and template cDNA synthesized from ~200 ng of mRNA using the qScript cDNA SuperMix (Quanta Biosciences, Gaithersburg, MD, USA).

Validating of RNAi silencing by qRT-PCR was done as described above. Approximately 50 ng or 20-fold of cDNA, and 900 nmol l⁻¹ of AamAch-L forward 5′-GGGAACTGGCTGGTGCATT-3′ and reverse 5′-GAGCGGCACGACGACTGT-3′ primers were added to 2×SYBR Green Master Mix (Life Technologies) in triplicate. To determine the apparent level of AamAch mRNA suppression, the formula S = 100 – (RQ^T/RQ^C × 100) was used, where S is mRNA suppression, and RQ^T and RQ^C are RQ of tissues in AamAch-S/L-dsRNA and GFP-dsRNA injected ticks, respectively. For each tick organ, mRNA suppression was determined as the mean ± s.d. of two S values. The two S values were generated by independently comparing treatment RQ to two separate control tick RQs.

To investigate if observations were significantly different between treatments, in the substrate hydrolysis, effects of RNAi silencing on tick feeding and validation of RNAi silencing analyses were subjected to non-parametric Student's t-test set to 95% confidence interval. One-way ANOVA and Tukey's honestly significant difference test were used to determine if differences in transcript abundance between unfed ticks and ticks fed for 24, 48, 72, 96 and 120 h were statistically significant. The statistical software package Prism version 6 (GraphPad Software Inc.) was used.