Introduction
Invasive species continue to
cause biodiversity loss (Kearney et al. 2018), loss of ecosystem
services (Fei et al. 2019) and reduced profitability within agriculture
(Pimentel et al. 2001). While Australia’s large land mass may be
uniquely protected by its natural isolation, increasing movement of
humans and associated traded goods pose an increasing threat to
biosecurity (Hulme 2016). Exotic pests were conservatively estimated to
cost A$13.6 billion in 2011/12 in Australia, including both economic
losses and management costs (Hoffmann and Broadhurst 2016). Arthropod
pests alone are estimated to cost Australia’s agricultural industries
A$8 billion annually (Bureau of Rural Sciences 2007; Canyon et al.
2011).
During a biological invasion, the area occupied (and associated impacts)
accelerate through time (Fleming et al. 2017), with early-implemented
prevention measures such as quarantine and surveillance frequently more
cost-effective than eradication, containment and asset-based protection
(Epanchin-Niell and Hastings 2010; Moore et al. 2010). The cost of
eradication is highest for late detections when the area invaded is
wider, and the likelihood of eradication success is lower (Rejmánek and
Pitcairn 2002; Timmins and Braithwaite 2003). Conversely, early
detection of exotic species, while populations are small and isolated,
improves the success and cost of management, but requires high
sensitivity, coverage, and robustness of surveillance methods (Dodd et
al. 2015). The biology and ecology of the target species imposes limits
on ease of detection and diagnosis. For example, highly polyphagous
pests can impede both surveillance and diagnostics when hosts are
unknown, cryptic exotic species are difficult to distinguish from
pre-existing native species causing delays in controlling the incursion
(Hauser 2011), and the phenology of a pest may create a narrow window of
detection or diagnosis (International Plant Protection Convention 2014).
Molecular methods that sequence DNA extracted from species tissue
samples have become vital to biosecurity efforts (Armstrong and Ball
2005; Floyd et al. 2010) as they can overcome many of the limitations
surrounding traditional diagnosis. Although routinely part of
surveillance activities, collection of tissue samples can
nonetheless be problematic due to small species sizes, rapid lifecycles,
and elusive behaviours (Rajan 2006; Augustin et al. 2012). When
specimens are difficult to collect, an alternative approach is to sample
for biological excreta, including scat, hair, or feeding residues (Höss
et al. 1992; Waits and Paetkau 2005; Rodgers and Janečka 2013; Valentin
et al. 2018). More recently, advances in molecular technology permit the
identification of species using only trace amounts of DNA found in the
environment (Valentin et al. 2018).
Environmental DNA (eDNA) sampling utilises a species’ residual DNA
remaining in the environment. Environmental DNA has already shown itself
to be a powerful tool for wildlife managers working with cryptic or
elusive species (Sigsgaard et al. 2015; Smart et al. 2015). It has been
widely applied to aquatic and semi-aquatic ecosystems (Ficetola et al.
2008; Thomsen et al. 2012; Piaggio et al. 2014; Goldberg et al. 2015)
due to the relative ease of isolating eDNA from water samples (Rodgers
and Mock 2015). Environmental DNA monitoring can be more sensitive than
traditional monitoring methods (Jerde et al. 2011; Dejean et al. 2012)
and is capable of detecting species at low densities (Pilliod et al.
2013; Smart et al. 2015; Dougherty et al. 2016). Applying eDNA sampling
to terrestrial environments can be challenging, although novel
applications that exploit the unique ecology or lifecycle of target
species are accumulating, such as isolating eDNA from predator wounds
(Williams et al. 2003), browsed twigs (Nichols et al. 2015), soil
(Andersen et al. 2012), leeches (Schnell et al. 2012), carrion-flies
(Calvignac-Spencer et al. 2013; Schubert et al. 2015), crop surfaces
(Valentin et al. 2018), and even the surface of flowers (Thomsen and
Sigsgaard 2019). Here we explore the possibility of applying an eDNA
approach to early detection and diagnosis of a globally significant
agricultural pest, the polyphagous vegetable leafminer, Liriomyza
sativae Blanchard.
The Agromyzidae are a well-studied group of small flies whose larvae
feed internally on plants, often as leaf and stem miners. Nearly all
agromyzid species are very host-specific, although some are highly
polyphagous and have become globally significant agricultural pests
(Spencer 1973, 1990). This includes L. sativae, L. bryoniae(Kaltenbach), L. huidobrensis (Blanchard), L. trifolii(Burgess) and Chromatomyia horticola (Goureau), all of which mine
plant leaves as larvae. While ecologically distinct, agromyzids are
morphologically similar and can often only be distinguished via
genitalia of adult flies (International Plant Protection Convention
2016). Moreover, when high confidence is necessary, the damage created
by agromyzids is functionally indistinguishable between species. Of the
exotic polyphagous leafminer, only L. sativae has been recorded
in Australia (International Plant Protection Convention 2017), and it
remains under quarantine. The cryptic nature of Liriomyza pests,
and wide host ranges have made them historically difficult to detect and
contain, allowing them to spread unchecked after an incursion (Powell
1981). Live specimens necessary for morphological identification are
difficult to obtain. Adult flies are less than 2 mm in length,
relatively inconspicuous in the environment and are typically only
active at certain times of the day (Zehnder and Trumble 1984). Larvae
will generally exit a plant leaf after one week, and mines are
frequently detected without living populations (Johnson et al. 1980).
However, leafmining damage remains until the plant leaf dies. Traces of
DNA within the mines will accumulate during defecation, moulting and if
parasitised by hymenopteran wasps, making agromyzid leafminers an ideal
candidate to test an eDNA approach.
Here, with an aim to enhance the
sensitivity and robustness of leafminer diagnosis and biosecurity
objectives, we present a novel eDNA-based diagnostic method for L.
sativae based on residual DNA inside empty leaf mines. Specifically, we
answer how the sensitivity of the diagnostic is affected by: (i)
different field preservation methods; (ii) DNA degradation due to field
exposure time; and (iii) different plant hosts.