Genome
Considerations for incorporating real-time PCR assays into routine marine biosecurity surveillance programmes: a case study targeting the Mediterranean fanworm (Sabella spallanzanii) and club tunicate (Styela clava)
Journal: Genome
Manuscript ID gen-2018-0021.R3
Manuscript Type: Article
Date Submitted by the 19-Jun-2018 Author:
Complete List of Authors: WOOD, Susanna; Cawthron Institute, Pochon, Xavier;Draft Cawthron Institute; University of Auckland Ming, Witold; Cawthron Institute von Ammon, Ulla; Cawthron Institute; University of Auckland Woods, Chris; National Institute of Water & Atmospheric Research Ltd Carter, Megan; National Institute of Water & Atmospheric Research Ltd Smith, Matt; National Institute of Water & Atmospheric Research Ltd Inglis , Graeme ; National Institute of Water & Atmospheric Research Ltd Zaiko, Anastasija ; Cawthron Institute; University of Auckland
Environmental DNA and RNA, Non-indigenous species, Occupancy Keyword: models, Real-time Polymerase Chain Reaction, Surveillance
Is the invited manuscript for consideration in a Special 7th International Barcode of Life Issue? :
Note: The following files were submitted by the author for peer review, but cannot be converted to PDF. You must view these files (e.g. movies) online.
Fig 1 - compressed.tif
https://mc06.manuscriptcentral.com/genome-pubs Page 1 of 38 Genome
1 Considerations for incorporating real-time PCR assays into routine marine biosecurity
2 surveillance programmes: a case study targeting the Mediterranean fanworm (Sabella
3 spallanzanii) and club tunicate (Styela clava)
4
5
6 Susanna A Wood1*, Xavier Pochon1,3, Witold Ming1, Ulla von Ammon1,2, Chris Woods4, Megan
7 Carter4, Matt Smith4, Graeme Inglis4, Anastasija Zaiko1,2
8 1 Coastal and Freshwater Group, Cawthron Institute, Nelson, New Zealand 9 2 School of Biological Sciences, University of Auckland, Auckland, New Zealand 10 3 Institute of Marine Science, University of Auckland, Auckland, New Zealand 11 4 National Institute of Water & Atmospheric Research Ltd, New Zealand 12 13 *corresponding authors: Coastal and Freshwater Group, Cawthron Institute, 98 Halifax Street East, 14 7010, Nelson, New Zealand. [email protected] 15 16
17
18
19
20
21
22
23
24
25
26
27
1 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 2 of 38
28
29
30
31
32 Abstract
33 Molecular techniques may provide effective tools to enhance marine biosecurity surveillance.
34 Prior to routine implementation, evidence-based consideration of their benefits and limitations is
35 needed. In this study, we assessed the efficiency and practicality of visual diver surveys and real-time PCR
36 assays (targeting DNA and RNA) for detecting two marine invasive species whose infestation levels varied
37 between species and location; Sabella spallanzanii and Styela clava. Filtered water samples (n=171)
38 were collected in parallel with dive surveys atDraft two locations as part of the New Zealand Marine High Risk Site
39 Surveillance programme: Nelson Harbour (27 sites) and Waitemata Harbour (30 sites). Diver surveys
40 resulted in a greater number of detections compared to real-time PCR: S. clava – 21 versus 5 sites
41 in Nelson, 6 versus 1 in Auckland; S. spallanzanii – 18 versus 10 in Auckland, no detections in
42 Nelson. Occupancy modelling derived detection probabilities for the real-time PCR for S. clava
43 were low (14%), compared to S. spallanzanii (66%). This could be related to abundances, or
44 species-specific differences in DNA shedding. Only one RNA sample was positive, suggesting
45 that most detections were from extracellular DNA or non-viable fragments. While molecular
46 methods cannot yet replace visual observations, this study shows they provide useful
47 complementary information.
48
49 Key words: Environmental DNA and RNA, Non-indigenous species, Occupancy models, Real-
50 time Polymerase Chain Reaction, Surveillance
2 https://mc06.manuscriptcentral.com/genome-pubs Page 3 of 38 Genome
51
Draft
3 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 4 of 38
52 Introduction
53 The arrival and establishment of marine non-indigenous species can have dramatic effects on the
54 structure and functioning of coastal ecosystems (Galil 2007; Wallentinus and Nyberg 2007;
55 Ehrenfeld 2010). Early detection and monitoring of marine non-indigenous species has become a
56 priority in many countries. Surveillance programmes commonly rely on techniques that use
57 morphological identification of organisms detected in samples or in situ, e.g., visual surveys by
58 divers (Cohen et al. 2001; Hewitt and Martin 2001; Inglis et al. 2006). Divers can detect moderate-
59 sized organisms even when present at relatively low densities, with limited water visibility (> 0.8
60 m Secchi depth) (Inglis et al. 2006). However, smaller individuals, especially when they are
61 amongst complex biofouling assemblages, may be overlooked. Utilization of divers to search for
62 marine non-indigenous species may alsoDraft not be possible at certain times/locations due to health
63 and safety concerns.
64
65 Molecular approaches are being advocated to circumvent some of the limitations of
66 morphological-based detection and monitoring methods in aquatic systems (e.g., Comtet et al.
67 2015). A suite of different techniques has been developed and applied in the last decade, with the
68 most commonly employed now being barcoding, real-time (or quantitative) polymerase chain
69 reaction (PCR) and metabarcoding using high-throughput sequencing (Ficetola et al. 2008;
70 Thomsen et al. 2012a; Wood et al. 2013). In aquatic systems, these techniques rely on collecting
71 and detecting either: (i) entire organisms (particularly in the case of micro-organism such as
72 bacteria, micro-algae, zooplankton), (ii) cells of organisms which may originate from various
73 sources, including scales, faeces, epidermal mucus, urine, saliva and gametes (Barnes et al. 2014),
74 or (iii) extracellular DNA which can be free-floating or particle-bound. These techniques have
4 https://mc06.manuscriptcentral.com/genome-pubs Page 5 of 38 Genome
75 been applied to determine aquatic biodiversity or the presence (and in some cases abundance) of
76 specific taxa in a wide range of habitats (e.g., Thomsen et al. 2012b; Dowles et al. 2016; Laroche
77 et al. 2016; Ulibarri et al. 2017; Keeley et al. 2018). Studies have shown that in some situations
78 molecular methods can be more efficient for detecting species than traditional approaches (Dejean
79 et al. 2012; Keskin 2014; Zaiko et al. 2016), making them a promising tool for the early detection
80 of newly introduced species and for monitoring the dispersal of established taxa (Blanchet 2012;
81 Piaggio et al. 2014; Rees et al. 2014; Comtet et al. 2015).
82
83 To date, most studies in aquatic environments investigating molecular techniques, DNA stability,
84 and appropriate sampling methods have targeted fish (Thomsen et al. 2012a; Takahara et al. 2013),
85 crustaceans (Forsströma and VasemägiDraft 2016), amphibians (Dejean et al. 2012), and
86 macroinvertebrates (Mächler et al. 2014). While sensitive and specific molecular assays have been
87 developed to detect sessile marine non-indigenous species (Gillium et al. 2014; Simpson et al.
88 2017; Wood et al. 2017), experimental and field studies to assess applicability of using these in
89 routine marine biosecurity programmes are lacking. Few studies have directly compared molecular
90 techniques with traditional visual survey methods (Ulibarri et al. 2017). Before molecular methods
91 can be routinely applied for surveillance, more information is required on achievable detection
92 rates and optimal sampling methods. Better understanding of what the assays are detecting (i.e.,
93 free-floating DNA, non-viable fragments of organisms, or living organisms (propagules)) is also
94 needed for providing robust advice to effectively guide management decisions.
95
96 In the present study, we focus on two marine non-indigenous species that are already established
97 in New Zealand coastal waters; Sabella spallanzanii and Styela clava. The Mediterranean fanworm
5 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 6 of 38
98 S. spallanzanii (Gmelin, 1791) (Polychaeta: Sabellidae) is a large, tube-dwelling polychaete worm
99 native to the Mediterranean Sea and Atlantic coast of Europe (Patti and Gambi 2001). Once
100 established, it can form dense populations (100’s to 1000’s per m2) covering a variety of marine
101 habitats (e.g., Holloway and Keough 2002). It was first detected in New Zealand in 2008 (Read et
102 al. 2011) and has been detected at multiple locations, including several at which it is now well
103 established (for specific distribution data please refer to https://www.marinebiosecurity.org.nz).
104 The club tunicate S. clava, Herdman, 1881 (Ascidiacea: Styelidae) is thought to be native to the
105 northwest Pacific (Japan, Korea, Northern China, and Siberia). It too can form dense populations
106 (100’s to 1000’s per m2) covering a variety of marine habitats once established (e.g., Clarke and
107 Therriault 2007; Davis and Davis 2010). It was first detected in New Zealand in 2005 (Davis and
108 Davis 2006), and is nowDraft relatively widespread and established
109 (https://www.marinebiosecurity.org.nz). Both these non-indigenous species present significant
110 ecological, economic and societal values risk to New Zealand (e.g., Soliman and Inglis 2018), and
111 are part of a suite of non-indigenous species subject to targeted surveillance in 11 major New
112 Zealand ports/marinas within the national Marine High Risk Site Surveillance (MHRSS)
113 programme (Inglis et al. 2006; Woods et al. 2017).
114
115 To compare the efficiency of species detection by visual diver surveys and real-time PCR assays, we targeted
116 two locations (Nelson and Auckland) with varying prevalence of the target species. In Nelson,
117 despite occasional detections of S. spallanzanii (usually on transient vessels or marina pontoons),
118 no established populations are known. In contrast, S. spallanzanii is well established in Auckland
119 (Woods et al. 2017), with populations reaching extremely high densities at some sites (e.g., 100-
120 600 individuals per m2, NIWA unpubl. data). Styela clava is present throughout Nelson Harbour,
6 https://mc06.manuscriptcentral.com/genome-pubs Page 7 of 38 Genome
121 particularly in the port and marina areas (Woods et al. 2017), but only at low densities (typically
122 ≤ 2 individuals per m2, but up to 50 per m2, NIWA unpubl. data). In Auckland S. clava populations
123 are present throughout the harbour, particularly in port and marina areas, but only at low densities
124 (typically ≤ 5 individuals per m2, but up to 50 per m2, NIWA unpubl. data).
125
126 Real-time PCR assays can provide semi-quantitative data; however, for larger organisms this is
127 rarely used because gene copy numbers may differ between organisms, cells and life stages. In this
128 study, we used real-time PCR assays with standard curves to generate data on gene copy numbers
129 to enable comparison with the abundance of individuals observed via visual diver surveys. We
130 also incorporated RNA analysis into this study to infer the source of any positive signals.
131 Environmental RNA is thought to degradeDraft within minutes to hours, and therefore is expected to
132 provide a better proxy for characterizing living organisms (Pochon et al. 2017). Thus, a signal
133 from only the DNA samples may indicate the detection was due to extracellular DNA or fragments
134 of the target organisms, whereas detection in both the DNA and RNA samples would suggest the
135 presence of living individuals, most likely larvae.
136
137 The objectives of the present study were: (i) to evaluate the applicability and sensitivity of two
138 surveillance approaches by comparing detection success in two locations with varying levels of
139 infestation; (ii) to determine the source of any molecular signal (living organisms, or non-viable
140 fragments or extracellular DNA), and (iii) to define considerations required when incorporating
141 real-time PCR assays into marine biosecurity surveillance programmes.
142
143 Methods
7 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 8 of 38
144 Sites and sampling
145 Sampling was undertaken at ports, marinas and surrounding coastline at two locations in New
146 Zealand, Nelson and Auckland (Fig. 1). Triplicate water samples were collected from a boat from
147 27 sites in Nelson on 12 and 13 February 2017, and 30 sites in Auckland on 19 and 20 April 2017.
148 Water sampling sites were within 3–5 m of the point of entry of the dive surveys (see below). The
149 water depth at each site was measured, and 15 L of seawater was pumped (5.0 GPM Washdown
150 Pump, Seaflow, China) through a pre-filter (20-µm netting) from three depths; 1 m above the
151 seafloor, 1 m below the water surface, and mid-way between these two depths (total volume = 45
152 L). Seawater (ca. 50 mL) from the site was added to the filtrate captured on the mesh and this was
153 filtered a second time (sterile Whatman GF/C, c. 1.6 µm pore size, 47 mm). The filter was halved
154 using sterile scissors, and half filters wereDraft placed into two separate cryotubes (2 mL), immediately
155 frozen in liquid nitrogen and then stored frozen (-80°C) for later DNA and RNA extraction (in
156 Nelson), or placed in sterile Eppendorf tubes containing LifeGuard Preservation Solution (1 mL;
157 QIAGEN, CA, U SA) and within 36 hours stored at -80°C for later DNA and RNA extraction (in
158 Auckland). Logistical constraints prevented us travelling with liquid nitrogen. Preliminary
159 experiments with samples collected from Nelson marina and stored in both LifeGuard Preservation
160 Solution and liquid nitrogen showed similar amplification of S. clava (data not shown). Prior to
161 sampling at each site, all sampling equipment was thoroughly washed using 2% bleach (sodium
162 hypochlorite) solution for at least 5 mins, and rinsed in seawater from the sampling site.
163
164 At each site, two occupationally-certified scientific divers using self-contained underwater
165 breathing apparatus (SCUBA) visually assessed up to 10 wooden pilings each or the equivalent
166 area of a continuous substratum (typically ca. 50 m of floating pontoon, steel wall, rock/concrete
8 https://mc06.manuscriptcentral.com/genome-pubs Page 9 of 38 Genome
167 breakwall or riprap) to a maximum water depth of 5 m (see Inglis et al. 2006). These dive surveys
168 were conducted as part of the bi-annual MHRSS surveying of each location (Woods et al. 2017).
169 When jetty/wharf piles were surveyed by the divers, the total number of piles with one or more
170 individuals of S. spallanzanii or S. clava observed as epibionts, not the number of individuals
171 observed per pile, are given. For pontoons or other continuous substratum, the average data from
172 the two divers (in individuals per m2) are given. These surveys were undertaken on 12 and 13
173 February 2017 (Nelson) and on 23 and 24 May 2017 (Auckland).
174
175 DNA and RNA extractions and cDNA synthesis
176 Each step of the molecular analysis (i.e., DNA extraction, real-time PCR setup, template addition
177 and real-time PCR analysis) was conductedDraft in a separate sterile laboratory dedicated to that step
178 with sequential work flow to ensure no cross-contamination. Each room was equipped with ultra-
179 violet sterilization which was switched on for a minimum of 15 min before and after each use. The
180 PCR set-up and template addition was undertaken in laminar flow cabinets with HEPA filtration.
181 Aerosol barrier tips (Axygen BioScience, CA, USA) were used throughout.
182
183 The filters from Nelson were placed into ZR BashingBead Lysis Tubes (2.0 mm; Zymo Research,
184 CA, USA) containing Lysis Buffer (1 mL) from the ZR-Duet™ DNA/RNA MiniPrep Kit (Zymo
185 Research). For the Auckland samples, the tubes containing LifeGuard were centrifuged (3,000 ×
186 g, 2 min) and the supernatant removed. The filter was transferred to the BashingBead Lysis Tube
187 using sterile forceps and any remaining pellet was mixed with Lysis Buffer and then transferred to
188 the same tube. The tubes were then placed on a bead beater (1600MiniG Spex SamplePrep, NJ,
189 USA) for 2 min at 1500 RPM. DNA and RNA were co-extracted from filters using the ZR-Duet™
9 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 10 of 38
190 DNA/RNA MiniPrep Kit (Zymo Research), according to the manufacturer’s protocol. The quality
191 and purity of isolated DNA and RNA were checked using a BioPhotometer (Eppendorf, Leipzig,
192 Germany; FileS1).
193
194 Extracted RNA of all corresponding DNA samples that tested positive using the S. spallanzanii or
195 S. clava real-time assays (see below) were then processed by eliminating trace DNA molecules
196 carried over in RNA extracts using two DNase treatments as in Langlet et al. (2013). The efficiency
197 of the DNase treatment was verified by running a real-time PCR using the reagents and conditions
198 described below for the S. spallanzanii or S. clava assay. The RNA samples were then reverse
199 transcribed into cDNA using the SuperScript III reverse transcriptase (Life Technologies, CA,
200 USA). To confirm that cDNA was produced,Draft the universal primers Uni18SF and Uni18SR (Zhan
201 et al. 2013) were used to amplify the eukaryotic V4 region of the nuclear small subunit ribosomal
202 DNA (18S rRNA) gene. These PCR amplifications were undertaken on an Eppendorf Mastercycler
203 (Eppendorf, Germany) in a total volume of 25 μL using 12.5 µL AmpliTaq Gold® 360 PCR Master
204 Mix (Life Technologies), 2.5 μL GC enhancer, 1 µL of each primer (10 µM, Integrated DNA
205 Technologies [IDT], CA, USA), 5 µL RNA/DNA free water (UltraPure, Life Technologies), 2 µL
206 Bovine Serum Albumin (BSA; 0.2 mg/mL, Sigma, USA) and 1 µL of template cDNA. Reaction
207 cycling conditions were: 95°C for 10 min, followed by 35 cycles of 95°C for 30 s, 50°C for 30 s,
208 72°C for 60 s, and a final extension of 72°C for 7 min. PCR products were visualized on a 1.5%
209 agarose gel, stained with RedSafe™ Nucleic Acid Staining Solution (iNtRON Biotechnology,
210 Gyeonggi-do, Korea).
211
212 Real-time PCR analysis of samples
10 https://mc06.manuscriptcentral.com/genome-pubs Page 11 of 38 Genome
213 All samples were screened in duplicate for inhibition using an internal inhibition control assay.
214 Each 10 µL reaction contained 6.25 µL KAPA Probe Fast QPCR Kit Master Mix (2×), 0.5 µL of
215 primers targeting the Internal Transcribed Spacer region 2 of the rRNA gene operon of
216 Oncorhynchus keta salmon sperm (10 µM, Sketa F2 and Sketa R3, IDT, USA, Haugland et al.
217 2005), 0.2 µL TaqMan probe (10 µM) labelled at the 5’ end with the fluorescent reporter dye
218 FAM-6-carboxyfluorecein and at the 3’ end with a non-fluorescent quencher with the Black Hole
219 Quencher-1 (IDT, USA), 0.8 µL DNA/RNA free water (Life Technologies, USA), 0.75 µL BSA
220 (Sigma), 1 µL extracted salmon sperm DNA (15 ng; Sigma, USA) and 1 µL of template DNA.
221 The cycling profile was: 95°C for 3 min, followed by 40 cycles at 95°C for 3 s and 60°C for 20 s.
222 All samples showed some inhibition and were diluted 1:10 with DNA/RNA free water (Life
223 Technologies, USA) and reanalysed forDraft inhibition as described above.
224
225 Each sample was then analysed in triplicate for S. spallanzanii and S. clava using the real-time
226 assays described in Wood et al. (2017) and Gillium et al. (2014), both targeting the mitochondrial
227 cytochrome c oxidase I (COI) gene. Each specific reaction consisted of 10 µL containing: 6.25 µL
228 KAPA Probe Fast QPCR Kit Master Mix (2×), 0.4 µL of the primers Sab3-QPCR-F and Sab3-
229 QPCR-R for S. spallanzanii (10 µM, IDT, Wood et al. 2017) and SC1F and SC1R for S. clava (10
230 µM, IDT, Gillum et al. 2014), 0.3 µL TaqMan probe synthesised with a FAM reporter dye at the
231 5´-end and a Black Hole Quencher-2 at the 3´-end, i.e. Sab3-QPCR-Probe for S. spallanzanii (10
232 µM, IDT, Wood et al. 2017) and SC1 for S. clava (10 µM, IDT, Gillum et al. 2014), 0.75 µL BSA
233 (Sigma), 0.8 µL DNA/RNA free water (Life Technologies, USA) and 1 µL of template DNA. The
234 cycling profile used for both assays was: 95°C for 3 min, followed by 40 cycles at 95°C for 3 s
235 and 60°C for 20 s. Five point standard curves ranging from 7.56 x 102 to 106 gene copies per µL
11 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 12 of 38
236 (S. spallanzanii) and 1.66 x 103 to 107 gene copies per µL for S. clava and no template controls
237 were analyzed in triplicate on each real-time PCR run. The standard curve was constructed using
238 a purified (AxyPrep PCR Clean-up Kit, Axygen Biosciences, USA) PCR product generated using
239 the S. spallanzanii and S. clava primers described above. The number of copies in the PCR product
240 used for the standard curves were determined using: (A × 6.022 × 1023) / (B × 1×109 × 650), with
241 A being the concentration of the PCR product, 6.022 × 1023 (Avogadro's number), B being the
242 length of the PCR product, 1×109 used to convert to ng, and 650 the average molecular weight per
243 base pair. The standard curves were linear (R2>0.98) and real-time PCR efficiency ranged between
244 0.80 to 0.96. When positive samples were obtained using the DNA samples, the corresponding
245 cDNA sample was analysed in triplicate using the corresponding real-time PCR assays described
246 above. Draft
247
248 Data analysis
249 Samples were considered positive when the target gene was present in at least two of the three
250 triplicates analysed. To compare detection probabilities for the two target genes and two detection
251 methods (visual surveys by divers and real-time PCR assays), a multi-method occupancy model
252 (Nichols et al. 2008) was used based on the mark-recapture-like approach (Mac-Kenzie et al. 2002)
253 implemented in PRESENCE v12.7 (Hines 2006). In the model, each sampling site (i.e., wharf,
254 breakwall, pontoon pier etc.; Figs. 1 and 2) represented a sample unit. For visual observations,
255 each diver’s survey at a site was treated as a sampling occasion. For the real-time PCR assay
256 method, each of the triplicate water samples was considered as an independent sampling occasion.
257
258 The following parameters were defined in the models for each target gene:
12 https://mc06.manuscriptcentral.com/genome-pubs Page 13 of 38 Genome
m 259 pt = probability of detection at occasion t by method m, given that sample unit is occupied and
260 the species is present at the immediate sampling site;
261 ψ = probability of sample unit being occupied by a species;
262 θt = probability of a species being present at a sampling site at occasion t, given that sample unit
263 is occupied (expected to approach 1 for sessile species);
264 휓 = naïve estimate of occupancy probability, calculated as proportion of sample units where the
265 species was detected over all units surveyed.
266
267 A small set of a priori single season multi-method models to describe the detection data were
268 defined. Since the analysis was primarily aimed at comparing species detection efficiency by two
269 methods, the pre-defined models do notDraft necessarily represent all environmental factors that
270 influenced the probability of occupancy of the two species. For S. clava, location was considered
271 as a single unit-specific covariate and its effect on large-scale probability of occupancy tested: ψ(H)
272 and ψ, respectively for location-dependent and location-independent occupancy. For S.
273 spallanzanii, no covariates were applied (no detection reported for Nelson). For both species, the
274 probability of detection was modelled as either constant or different between detection methods:
275 p(m) and p, respectively for method-dependent and method- independent detection. All fitted
276 models were ranked according to Akaike Information Criterion (AIC) values calculated in
277 PRESENCE.
278
279 Results
280 Styela clava
13 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 14 of 38
281 Styela clava was observed by divers at 21 of the 27 surveyed Nelson sites (Table 1, Fig. 1). The
282 highest densities reported were 2 individuals per m2 and when piles were surveyed, at two sites
283 one or more individuals were observed on 5 of the 10 piles surveyed (Table 1). In contrast, S. clava
284 was only detected in 5 of the 27 Nelson sites by real-time PCR. At four of these sites, it was only
285 detected in one of the triplicate samples collected at that location (Table 1). Calculated gene copy
286 numbers ranged from 3,840 to 14,400 per sample. At sites with sufficient replicates, where
287 standard errors could be calculated, variability in gene copy number per site was high (23-67%;
288 Table 1). There were three sites (sites 4, 14 and 16) where both diver surveys and the real-time
289 PCR assays detected S. clava, and two sites (sites 20, 22) where it was detected only by real-time
290 PCR (Table 1 and Fig. 1). These two sites were in the inner regions of the marina (Fig. 1). The
291 limited number of sites (n=3) where bothDraft methods detected S. clava prevented any statistical
292 analysis of the relationship between visual density and copy number to be undertaken; however,
293 there was some indication that higher gene copy numbers were associated with denser settlements
294 (e.g., site 4 and 14). The cDNA of all DNA samples that were positive returned no real-time PCR
295 signal.
296
297 In Auckland, S. clava was identified by divers at 6 of the 19 sites surveyed with a maximum density
298 of 1 per m2, and when piles were surveyed, at one site, one or more individuals were observed on
299 5 of the 10 piles surveyed (Table 1). The real-time PCR assay was only positive for one of these
300 19 sites (site 39; Table 1). There were an additional three sites where S. clava was detected by the
301 real-time PCR assay, but diver surveys were not undertaken due to operational/safety reasons
302 (Table 1, Fig. 1). Variability in gene copy number within a site was high (Table 1), and gene copy
303 numbers per sample were on average ca. 50-fold higher than those measured in Nelson (Table 1).
14 https://mc06.manuscriptcentral.com/genome-pubs Page 15 of 38 Genome
304 There was only one site (site 39) where both the visual survey and real-time PCR assay (only one
305 of the three replicates) detected S. clava, and this was the site with the highest density of S. clava
306 observed by divers (Table 1). The cDNA from one of the samples from site 29 was positive for S.
307 clava using the real-time PCR assay. All three replicates were positive with an average gene copy
308 number per sample of 71,200 ± 11,520 (standard error).
309
310 Sabella spallanzanii
311 No S. spallanzanii was observed during the dive survey or detected using the real-time PCR at any
312 of the Nelson sites. In Auckland, S. spallanzanii was identified by divers at 18 of the 19 sites
313 surveyed, with maximum densities of 30 per m2. When piles were surveyed, at one site, one or
314 more individuals were observed on 10 Draftof the 10 piles surveyed (Table 1, Fig. 2). The real-time
315 PCR assay detected S. spallanzanii at 20 of the 30 total sites surveyed in Auckland, including 10
316 of the 19 sites also surveyed by divers. In most instances S. spallanzanii was detected in 2 or 3
317 replicate samples taken at each site, with gene copy numbers ranging from 18,750 ± 7,400 to
318 1,205,740 ± 150,013 per sample (Table 1).
319
320 The limited number of samples with positive detection, and difficulties with using the visual data
321 in two different formats (i.e. per m2 and per no. of piles), prevented any robust statistical correlation
322 to be undertaken between visual survey density and real-time PCR copy numbers. However,
323 inspection of the data suggests no or very weak relationships. No S. spallanzanii was detected by
324 real-time PCR assay at three sites with the highest densities determined visually (sites 37, 38 and
325 42; Table 1), and conversely high copy numbers were detected at sites where visual surveys
326 reported only moderate densities (i.e., sites 35, 45; Table 1). There were no obvious spatial patterns
15 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 16 of 38
327 to explain the lack of detection via real-time PCR where divers observed this species. For example,
328 there were positive and negative detections using the real-time assay within the same marina
329 (distance between sites ca. 1,000 m), when all visual surveys detected the species (sites 53, 54).
330 No positive S. spallanzanii signals were detected in the cDNA samples.
331
332 Occupancy and detection probabilities
333 Naïve estimates of large-scale occupancy (휓 ) were considerably higher for visual surveys
334 compared to real-time PCR assay results for both species and both locations (Fig. 3). However, for
335 S. clava, occupancy estimates were consistently (ca. two-fold) higher for Nelson compared to
336 Auckland. Modelled estimates of small-scale occupancy (θ), as expected, were always 1, since
337 targeted sessile benthic species cannot vacateDraft the sampling site between sampling occasions. Both
338 large-scale occupancy (ψ) and detection probability (p) estimates yielded good precision in the
339 top-ranked models, as evidenced by the standard error values (never exceeded 30% of the estimate
340 value, Table 2). Consistent with naïve estimates, there was evidence that ψ for S. clava differed
341 between Auckland and Nelson. For both species, models accounting for varying detection
342 probabilities between methods had considerably higher Akaike weights compared to those not
343 including method as a factor, suggesting very strong evidence for method-specific detection. In all
344 cases probability of detection was higher for visual surveys. This was overall noticeably higher for
345 S. spallanzanii compared to S. clava, reaching 100% detection using visual surveys.
346
347 Discussion
348 The present study investigated the effectiveness of collecting and filtering water samples and using
349 real-time PCR assays to detect two non-indigenous marine benthic sessile species: the
16 https://mc06.manuscriptcentral.com/genome-pubs Page 17 of 38 Genome
350 Mediterranean fan worm S. spallanzanii, and the club tunicate S. clava. Our results concur with the
351 study of Mächler et al. (2014) who observed that for some species (they studied a gastropod, an
352 amphipod and two Trichoptera), traditional detection methods are more efficient than eDNA
353 methods. Likewise, Ulibarri et al. (2017) compared the American Fisheries Society Standard
354 Snorkelling technique with DNA sampling (15 mL water samples taken every 10 m over a 100-m
355 reach) combined with real-time PCR analysis, and also showed greater detection rates with the
356 traditional method. Here, we discuss the implications of our findings in the context of factors that
357 should be considered when including eDNA approaches into routine marine biosecurity
358 surveillance programmes.
359
360 In the present study, there were numerousDraft sites where divers observed the target species but these
361 were not detected by the real-time PCR assays. Considering detection rates reported from other
362 similar studies e.g., 57% for the mud crab (Forsström and Vasemägi 2016), 82% for dragonflies,
363 100% for toads (Thomsen et al. 2012b) and 59% for crayfish (Tréguier et al. 2014), the detection
364 probabilities derived for real-time PCR from the best-fitted occupancy model were relatively low
365 for S. clava (14%), but comparable for S. spallanzanii (66%). Direct comparisons with other
366 studies are challenging due to differences in target species morphology, density and life stage,
367 habitat, and sampling detection and analysis methods (Kelly et al 2014; Tréguier et al. 2014;
368 Deiner et al. 2015). The lower detection probabilities yielded by real-time PCR assays compared
369 to visual diver surveys in the current study suggest that molecular-based analyses should not be
370 used as the only detection method for the two considered species for marine surveillance purposes.
371 It is important to understand the reasons for lower detection using a molecular-based technique,
17 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 18 of 38
372 and to consider the potential for improvement with further development of sampling
373 methodologies and molecular-based technologies.
374
375 Reasons for the low detection rate using the real-time PCR assays and differences between the two
376 species are likely the result of a combination of factors. Although in this study the limited sample
377 numbers prevented any correlative statistical analysis between the methodologies, the data suggest
378 there would only be a weak relationship between copy number and numbers of individuals at a
379 site. However, at a population scale there may be relationships. For example, for higher density of
380 S. spallanzanii in Auckland, the detection probabilities were much higher compared to similar
381 occupancy levels of S. clava in Nelson. Other researchers have shown that the amount of target
382 DNA detected in a water body is influencedDraft by species density (Takahara et al. 2012; Pilliod et al.
383 2013).
384
385 There is limited knowledge on how the amount of cell/extracellular DNA released from different
386 organisms varies. Biological form and life-style may affect this process; for example, S.
387 spallanzanii has a two-layered crown of hundreds of fine tentacles which extend out of the tube
388 during feeding. It seems more likely that pieces may break off these and be distributed in ambient
389 environment, compared to the more rigid and streamlined form of S. clava. However, S. clava
390 expels water through its siphon which may include DNA. In a study of three marine fish Sassoubre
391 et al. (2016) demonstrated variable eDNA shedding and decay rates, and attributed this to different
392 sources i.e., scales versus mucus, as well as fish size and possibly physiologies, metabolic rates
393 and feeding activities. DNA degradation rates are also likely to be species-specific (Sassoubre et
394 al. 2016) and affected by environmental factors such as water temperature, pH and sunlight
18 https://mc06.manuscriptcentral.com/genome-pubs Page 19 of 38 Genome
395 (Dejean et al. 2011; Strickler et al. 2015; Andruszkiewicz et al. 2017). More information on the
396 source and longevity of DNA in the marine environment is urgently needed to assist in optimizing
397 sampling efforts and aid in the interpretation of results in routine marine biosecurity surveillance
398 programmes.
399
400 Research is required to determine the source of DNA from S. spallanzanii and S. clava. This may
401 also influence the most optimal sample collection methodology. In this study, we used a 20 µm
402 pre-filter followed by filtration on to a GF/C filter (1.6 µm pore size) to enable relatively large
403 volumes of water to be sampled. Using these methods any ‘free’ extracellular DNA would have
404 passed directly through the pre-filter, and therefore we were likely capturing extracellular DNA
405 attached to particles, or non-viable fragmentsDraft of the organisms. Previous studies have shown that
406 there is no ideal sampling method to capture all environmental DNA in a sample (Deiner et al.
407 2015; Sassoubre et al. 2016). Additionally, larger sampling volumes may not result in higher
408 detection rates; for example, Mächler et al. (2016) showed that the volume of water sampled had
409 no effect on detection for two of three studied species, whereas detection improved for the third
410 species as the volume of water filtered increased. Site topography and location may have also had
411 an effect on the real-time PCR detections. For example, sites 24-27 in Nelson are exposed to the
412 ocean with water column being constantly mixed by wind and tidal currents. Although much of
413 the water in the Nelson marina would be exchanged through tidal fluxes, either side of the tide
414 stand there would be periods where the water would be relatively stagnant, perhaps increasing
415 chances of detection. Timing molecular sampling close to the slack water period of the tidal cycle
416 may enhance detection.
417
19 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 20 of 38
418 At each site, there was high variability in detection via real-time PCR and a number of instances
419 where only one or two of the samples were positive. Copy numbers also varied markedly among
420 biological replicates. This, in contrast to other studies which have shown high consistency in
421 biological replicates (Forsströma and Vasemägi 2016), further highlights the likelihood that there
422 is species-specific variability in the quantities and distributions of DNA in marine systems. To
423 date, most environmental DNA studies in marine environments have focused on motile species,
424 e.g. fish and crustaceans (Forsström and Vasemägi 2016; Sassoubre et al. 2016), and little attention
425 has been given to sessile species which are often associated with marine bioinvasions. The high
426 variability in copy number among biological replicates also reinforces the need to carefully
427 consider adequate sampling effort, taking into account specific biological traits of target species
428 and environmental characteristics of theDraft sampling locations.
429
430 In Nelson, there were two sites where S. clava was detected by real-time PCR but divers did not
431 observe this species. These sites were in the inner regions of the marina with relatively limited
432 visibility (ca. 1–2 m), with potentially lower likelihood of species detection by divers. It is also
433 likely that there were specimens at the nearby locations (i.e., site 21, where S. clava was observed
434 by divers, was less than 100 m away) and that the samples contained DNA or fragments of
435 organisms from this close-by site. As noted above, the presence of DNA signal may have been
436 greater at these sites in the inner region of the marina due to reduced water exchange with the
437 coastal waters. DNA is not constrained by any barriers in aquatic systems; thus a limitation of
438 DNA sampling is that it does not necessarily provide any quantitative information, nor does it
439 provide accurate information on the exact location of the target taxa. Therefore, in a marine
440 biosecurity surveillance programme, a positive signal from a molecular technique would need to
20 https://mc06.manuscriptcentral.com/genome-pubs Page 21 of 38 Genome
441 be followed up with a more specific detection technique (e.g., visual method), or target sampling
442 of other matrices, e.g., biofouling, with molecular detection to confirm the presence of the target
443 taxa.
444
445 The lack of positive signals in all but one (site 29, S. clava) of the cDNA samples suggests that
446 most of the real-time PCR detections were from extracellular DNA (either free DNA, DNA bound
447 to organisms/inorganic particles), or non-viable fragments of organisms, rather than living
448 organisms. Environmental RNA is known to degrade within minutes to hours (Stoeck et al. 2007;
449 Orsi et al. 2013), and therefore is expected to provide a better proxy for the detection of living
450 organisms than DNA (Pochon et al. 2017). This detection possibly corresponds to the capture of
451 larvae as our sampling took place duringDraft the known spawning season for S. clava in New Zealand
452 (late summer to early autumn, Wong et al. 2011), although we might have expected more positive
453 detections given this scenario. Specialized collection, storage and workflow protocols (e.g.,
454 preservation buffers, dedicated instruments and sample preparation rooms), are required when
455 working with RNA, and the reverse transcription step adds considerable expense and time to the
456 sample processing. For routine marine surveillance programme a more affordable approach should
457 involve focusing primarily on DNA with additional samples (for both molecular and
458 morphological analysis) collected following positive detections.
459
460 Most marine organisms have planktonic gamete/larval stages, and although these will be widely
461 dispersed by physical processes, appointing sampling to this propagule dispersal period would
462 almost certainly enhance chances of collecting and detecting target organisms (especially those
463 not constantly present in the water column). However, planktonic life stages of many marine
21 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 22 of 38
464 invasive species are relatively limited (2–3 weeks), and it would be hard to co-ordinate large
465 national scale surveillance programmes to fit these time frames, especially given constraints such
466 as the need for relatively calm and stable weather for sampling. For many invasive species, the
467 reproductive season (in a new environment) is unknown and unlikely to match that of their native
468 locations due to environmental differences. In countries such as New Zealand, that span a
469 latitudinal gradient of twelve degrees (between -35° and -47° south) and where the temperature of
470 coastal waters varies from more than 24°C in the north to less than 7°C in the south, reproductive
471 periods can vary markedly. Timing national marine biosecurity surveillance programmes to align
472 with reproductive season for multiple species would be very challenging. Additionally, many
473 species arrive attached to vessels/structures as mature adults, and many do not reach reproductive
474 stages for many months; therefore, relyingDraft on sampling methods that capture only larval stages is
475 unlikely to be successful and advantageous from a biosecurity management perspective. As
476 suggested by Hayes et al. (2005), it is advisable to use multiple methods targeting different life-
477 stages and different environmental matrices where traces of species presence can be detected.
478
479 The results of the visual dive surveys are instantaneous, albeit in some cases specimens might need
480 confirmative identification by a skilled taxonomist or via confirmatory molecular testing, i.e. real-
481 time PCR or barcoding. In contrast, with the number of molecular samples collected per location
482 (81–90) and using the real-time PCR methods employed in this study, full molecular results took
483 2–3 weeks to generate post sampling. The addition of the dual DNA/RNA extraction, and when
484 required reverse transcription steps and testing of cDNA samples, added additional time that would
485 not be required in a routine surveillance programme. Even for the present study, where a relatively
486 limited number of sites were sampled, this constituted >2,300 real-time PCR reactions; at an
22 https://mc06.manuscriptcentral.com/genome-pubs Page 23 of 38 Genome
487 estimated 15 NZ dollars per sample (not including capital, machine running and labour expenses).
488 Digital droplet PCR (ddPCR) offers potential for rapidly expediting this process as no inhibition
489 assays are required, no standard curves are needed, and samples do not need to be run in triplicate
490 as essentially every result is an average of ca. 20,000 individual PCR reactions (Doi et al. 2015;
491 Te et al, 2015).
492
493 Technology is advancing rapidly and there is an increasing push for the development of portable
494 and cheap real-time PCR machines and assays that allow for on-site analysis (e.g., Tsai et al. 2012;
495 Chua et al. 2016). More regular sampling with the ability of managers to analyse samples rapidly
496 on site offers huge potential to enhance marine biosecurity surveillance in the future. One
497 limitation of real-time PCR assays is Draft that they only target one or a few species, and rely on
498 predictions of which species are most likely to arrive and become problematic at a given location.
499 While methods such as metabarcoding can overcome these limitations and allow characterisation
500 of entire biological communities (Taberlet et al. 2012; Díaz-Ferguson et al. 2014; Comtet et al.
501 2015; Pochon et al. 2017), the steps involved in sample preparation are relatively protracted and it
502 is unlikely they will be employed for on-site rapid detection in the near future.
503
504 The occupancy-modelling approach employed in the present study proved a useful method to
505 account for imperfect detection of the two target species. It provided a valuable analytical
506 framework for estimating the different sensitivities of the visual and real-time PCR methods for
507 the two study species, and we recommend its application to similar comparative studies.
508
23 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 24 of 38
509 This study has highlighted that there are uncertainties related to the use of molecular methods to
510 detect marine non-indigenous species, and limitations that may need further development to
511 overcome. Critical questions that need to be addressed include:
512 - Which sampling methods are most effective and is there one method that can capture the
513 DNA of multiple target species?
514 - Should sampling be undertaken at specific times of the day or year?
515 - How long does environmental DNA/RNA remain intact in marine environments and does
516 this vary between species?
517 Considerable research is required to answer these uncertainties, and it’s clear that molecular
518 methods cannot yet be used as a substitute for observations made by experienced field personnel
519 (Ulibarri et al. 2017). However, these Draft methods could be used to complement existing marine
520 surveillance techniques. This would enhance marine biosecurity surveillance through their ability
521 for sensitive and specific detection, and because they are able to identify cryptic species or life
522 stages (Smith et al. 2012). The incorporation of these methods into routine sampling programmes
523 would also expedite the development of more rapid and robust sampling techniques, and aid in the
524 development of novel molecular-focused sampling equipment.
525
526 Acknowledgements
527 We thank Janet Adamson (Cawthron) for technical assistance in the laboratory and Marc Jary
528 (Cawthron) for assistance preparing for field sampling. We thank the New Zealand Ministry for
529 Primary Industries for access to data generated as part of the Marine High Risk Site Surveillance
530 programme, and Abraham Growcott as MPI’s Operational Liaison. This work was supported by
24 https://mc06.manuscriptcentral.com/genome-pubs Page 25 of 38 Genome
531 the National Institute of Water and Atmospheric Research Ltd under Coasts and Oceans Research
532 Programme 6, Marine Biosecurity (SCI 2016-18).
533
534 References
535 Andruszkiewicz, E.A., Sassoubre, L.M., and Boehm, A.B. 2017. Persistence of marine fish
536 environmental DNA and the influence of sunlight. PLoS ONE 12: e0185043.
537 doi.org/10.1371/journal.pone.0185043.
538 Barnes, M.A., Turner, C.A., Jerde, C.L., Renshaw, M.A., Chadderton, W.L., and Lodge, D.M.
539 2014. Environmental conditions influence eDNA persistence in aquatic systems. Environ. Sci.
540 Technol. 48: 1819–1827. doi: 10.1021/es404734p.
541 Blanchet, S. 2012. The use of molecularDraft tools in invasion biology: an emphasis on freshwater
542 ecosystems. Fish Manage. Ecol. 19: 120–132. 10.1111/j.1365-2400.2011.00832.x.
543 Chua, K.H., Lee, P.C., and Chai, H.C. 2016. Development of insulated isothermal PCR for rapid
544 on-site malaria detection. Malaria J. 15: 134. doi: 10.1186/s12936-016-1183-z.
545 Clarke, C.L., and Therriault, T.W. 2007. Biological synopsis of the invasive tunicate Styela clava
546 (Herdmann 1881). Canadian Manuscript Report of Fisheries and Aquatic Sciences 2801.
547 Cohen, A.N., Berry, H.D., Mills, C.E., Milne, D., Britton-Simmons, K., et al. 2001. Washington
548 State Exotics Expedition 2000: A Rapid Survey of Exotic Species in the Shallow Waters of
549 Elliot Bay, Totten and EldInlets, and Willapa Bay. Available from:
550
551 Comtet, T., Sandionigi, A., Viard, F., and Casiraghi, M. 2015. DNA (meta)barcoding of biological
552 invasions: a powerful tool to elucidate invasion processes and help managing aliens. Biol.
553 Invasions 17: 905–922. doi.org/10.1007/s10530-015-0854-y.
554 Davis, M.H., Davis, M.E. 2006. Styela clava (Tunicata: Ascidiacea) a new edition to the fauna of
25 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 26 of 38
555 New Zealand. Porcupine Mar. Nat. Hist. Soc. Newsl. 20: 19−22
556 Davis, M.H., and Davis, M.E. 2010. The impact of the ascidian Styela clava Herdman on shellfish
557 farming in the Bassin de Thau, France. J. Appl. Ichthyol. 26: 12–18. doi: 10.1111/j.1439-
558 0426.2010.01496.x.
559 Deiner, K., Walser, J.-C., Mächler, E., and Altermatt, F., 2015. Choice of capture and extraction
560 methods affect detection of freshwater biodiversity from environmental DNA. Biol. Conserv.
561 183: 53–63. doi.org/10.1016/j.biocon.2014.11.018.
562 Dejean, T., Valentini, A., Duparc, A., Pellier-Cuit, S., Pompanon, F., Taberlet, P., and Miaud, C.
563 2011. Persistence of environmental DNA in freshwater ecosystems. PLoS ONE 6: e23398.
564 doi.org/10.1371/journal.pone.0023398.
565 Dejean, T., Valentini, A., Miquel, C., Taberlet,Draft P., Bellemain, E., and Miaud, C. 2012. Improved
566 detection of an alien invasive species through environmental DNA barcoding: the example of
567 the American bullfrog Lithobates catesbeianus. J. Appl. Ecol. 49: 953–959.
568 Díaz-Ferguson, E.E., and Moyer, G.R. 2014. History, applications, methodological issues and
569 perspectives for the use environmental DNA (eDNA) in marine and freshwater environments.
570 Revista de Biología Tropical. 62: 1273–1284. doi:10.1111/j.1365-2664.2012.02171.x.
571 Doi, H., Takahara, T., Minamoto, T., Matsuhashi, S., Uchii, K., and Yamanaka, H. 2015. Droplet
572 digital Polymerase Chain Reaction (PCR) outperforms real-time PCR in the detection of
573 environmental DNA from an invasive fish species. Environ. Sci. Technol. 49: 5601–5608.
574 doi:10.1021/acs.est.5b00253.
575 Dowle, E., Pochon, X., Banks, J., Shearer, K., and Wood, S.A. 2016. Targeted gene enrichment
576 and high throughput sequencing for environmental biomonitoring. Mol. Ecol. Res. 16: 1240-
577 1254. doi:10.1111/1755-0998.
26 https://mc06.manuscriptcentral.com/genome-pubs Page 27 of 38 Genome
578 Ehrenfeld, J.G. 2010. Ecosystem consequences of biological invasions. Annu. Rev. Ecol. Evol.
579 41: 59-80. doi: 10.1016/j.marpolbul.2016.05.054.
580 Ficetola, G.F., Miaud, C., Pompanon, F., Taberlet, P. 2008. Species detection using environmental
581 DNA from water samples. Biol. Let. 4: 423-425. doi:10.1098/rsbl.2008.0118.
582 Forsström, T., and Vasemägi, A. 2016. Can environmental DNA (eDNA) be used for detection
583 and monitoring of introduced crab species in the Baltic Sea? Mar. Pollut. Bull. 109: 350-355.
584 doi:10.1016/j.marpolbul.2016.05.054.
585 Galil, B.S. 2007 Loss or gain? Invasive aliens and biodiversity in the Mediterranean Sea. Mar.
586 Poll. Bull. 55: 314-322. doi.org/10.1016/j.marpolbul.2006.11.008.
587 Gillum, J.E., Jimenez, L., White, D.J., Goldstien, S.J., and Gemmell, N.J. 2014. Development and
588 application of a quantitative real-timeDraft PCR assay for the globally invasive tunicate Styela
589 clava. Manag. Biol. Invasion 5: 133-142. Doi:10.3391/mbi.2014.5.2.06.
590 Haugland, R.A., Siefring, S.C., Wymer, L.J., Brenner, K.P., and Dufour, A.P. 2005. Comparison
591 of Enterococcus measurements in freshwater at two recreational beaches by quantitative
592 polymerase chain reaction and membrane filter culture analysis. Water Res. 39: 559-568.
593 doi:10.1016/j.watres.2004.11.011.
594 Hayes, K.R., Cannon, R., Neil, K., and Inglis G. 2005. Sensitivity and cost considerations for the
595 detection and eradication of marine pests in ports. Mar. Pollut. Bull. 50: 823-834.
596 https://doi.org/10.1016/j.marpolbul.2005.02.032.
597 Hewitt, C.L., and Martin, R.B. 2001. Revised protocols for port surveys for introduced marine
598 species: survey design, sampling protocols and specimen handling. CRIMP Technical Report
599 Number 22, CSIRO Division of Marine Research, Hobart, Australia.
600 Hines, J. E. 2006. PRESENCE- Software to estimate patch occupancy and related parameters.
27 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 28 of 38
601 USGS-PWRC. http://www.mbr-pwrc.usgs.gov/software/presence.html.
602 Holloway, M.G., and Keough, M.J. 2002. An introduced polychaete affects recruitment and larval
603 abundance of sessile invertebrates. Ecol. Appl. 12: 1803–1823. doi:10.1890/1051-
604 0761(2002)012[1803:AIPARA]2.0.CO;2
605 Inglis, G., Hurren, H., Gust, N., Oldman, J., Fitridge, I., Floerl, O., and Hayden, B. 2006.
606 Surveillance design for early detection of unwanted exotic marine organisms in New Zealand.
607 Biosecurity New Zealand Technical Paper No. 2005-17.
608 Keeley, N., Wood, S.A., and Pochon, X. 2018. Development and validation of a multi-trophic
609 metabarcoding biotic index for benthic organic enrichment biomonitoring. Ecol. Ind. 85:
610 1044-1057. https://doi.org/10.1016/j.ecolind.2017.11.014.
611 Kelly, R.P., Port, J.a., Yamahara, K.M.,Draft and Crowder, L.B. 2014. Using environmental DNA to
612 census marine fishes in a large mesocosm. PLoS One 9: e86175.
613 doi.org/10.1371/journal.pone.0086175.
614 Keskin, E. 2014. Detection of invasive freshwater fish species using environmental DNA survey.
615 Biochem. Syst. Ecol. 56: 68–74. doi.org/10.1016/j.bse.2014.05.003.
616 Langlet, D., Geslin, E., Baal, C.,, Metzger E., Lejzerowicz, F., Riedel, B., Zuschin, M., Pawlowski,
617 J., Stachowitsch, M., and Jorissen, F. 2013. Foraminiferal survival after long-term in situ
618 experimentally induced anoxia. Biogeosci. 10: 7463-7480. doi.org/10.5194/bg-10-7463-2013.
619 Laroche, O., Wood, S.A., Tremblay, L.A., Ellis, J.I., Lejzerowicz, F., Pawlowki, J., Lear, G.,
620 Atalah, J., and Pochon, X. 2016. First evaluation of foraminiferal metabarcoding for
621 monitoring benthic health at an offshore oil drilling site. Mar. Environ. Res. 120: 225-235. doi:
622 10.1016/j.marenvres.2016.08.009.
623 Mächler, E., Deiner, K., Steinmann, P., and Altermatt, F. 2014. Utility of environmental DNA for
28 https://mc06.manuscriptcentral.com/genome-pubs Page 29 of 38 Genome
624 monitoring rare and indicator macroinvertebrate species. Freshw. Sci. 33: 1174–1183.
625 https://doi.org/10.1086/678128.
626 MacKenzie, D.I., Nichols, J.D., Lachman, G.D., Droege S., Royle J.A., and Langtimm C.A. 2002.
627 Estimating site occupancy rates when detection probabilities are less than one. Ecol. 83: 2248-
628 2255. doi: 10.1890/0012-9658(2002)083[2248:ESORWD]2.0.CO;2.
629 Nichols J.D., Bailey L.L., O’Connell A.F. Jr., Talancy N.W., Campbell Grant E.H., Gilbert A.T.,
630 Annand E. M., Husband T.P., and Hines J.E. 2008. Multi-scale occupancy estimation and
631 modelling using multiple detection methods. J. App. Ecol. 45: 1321–1329. doi:
632 10.1111/j.1365-2664.2008.01509.x.
633 Orsi, W., Biddle, J.F., and Edgcomb, V. 2013. Deep Sequencing of subseafloor eukaryotic rRNA
634 reveals active fungi across marineDraft subsurface provinces. PLoS ONE 8: e56335.
635 doi.org/10.1371/journal.pone.0056335.
636 Patti, F.P., and Gambi, M.C. 2001. Phylogeography of the invasive polychaete Sabella
637 spallanzanii (Sabellidae) based on the nucleotide sequence of internal transcribed spacer 2
638 (ITS2) of nuclear rDNA. Mar. Ecol. Prog. Ser. 215: 169-177.
639 doi.org/10.1017/S0025315417000261P.
640 Piaggio, A.J., Engeman, R.M., Hopken, M.W., Humphrey, J.S., Keacher, K.L., Bruce, W.E., and
641 Avery, M.L. 2014. Detecting an elusive invasive species: a diagnostic PCR to detect Burmese
642 python in Florida waters and an assessment of persistence of environmental DNA. Mol. Ecol.
643 Resour. 14: 374–380. doi: 10.1111/1755-0998.12180.
644 Pilliod, D.S., Goldberg, C.S., Arkle, R.S., Waits, L.P., and Richardson, J. 2013. Estimating
645 occupancy and abundance of stream amphibians using environmental DNA from filtered water
646 samples. Can. J. Fish. Aquat. Sci. 70: 1123–1130. doi.org/10.1139/cjfas-2013-0047.
29 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 30 of 38
647 Pilliod, D.S., Goldberg, C.S., Arkle, R.S., and Waits, L.P. 2014. Factors influencing detection of
648 eDNA from a stream-dwelling amphibian. Mol. Ecol. Resour. 14: 109–116. doi:10.1111/1755-
649 0998.12159.
650 Pochon, X., Zaiko, A., Fletcher, L.M., Laroche, O., and Wood, S.A. 2017. Wanted dead or alive?
651 Using metabarcoding of environmental DNA and RNA to distinguish living assemblages for
652 biosecurity applications. PLoS ONE 12: e0187636. doi: 10.1371/journal.pone.0187636.
653 Read, G.B., Inglis, G.J., Stratford, P., and Ahyong, S.T. 2011. Arrival of the alien fanworm Sabella
654 spallanzanii (Gmelin, 1791) (Polychaeta: Sabellidae) in two New Zealand harbours. Aquat
655 Inv. 6: 273-279. doi:10.3391/ai.2011.6.3.04.
656 Rees, H.C., Maddison, B.C., Middleditch, D.J., Patmore, J.R.M., and Gough, K.C. 2014. The
657 detection of aquatic animal species usingDraft environmental DNA - a review of eDNA as a survey
658 tool in ecology. J. Appl. Ecol. 51: 1450–1459. doi:10.1111/1365-2664.12306
659 Sassoubre, L.M., Yamahara, K.M., Gardner, D., Block, B.A., and Boehm, A/B. 2016.
660 Quantification of Environmental DNA (eDNA) shedding and decay rates for three marine fish.
661 Environ Sci Technol. 50:10456-10464. doi:10.1021/acs.est.6b03114
662 Simpson, T.J.S., Dias, P.J., Snow, M., Muñoz, J., and Berry, T. 2017. Real-time PCR detection of
663 Didemnum perlucidum (Monniot, 1983) and Didemnum vexillum (Kott, 2002) in an applied
664 routine marine biosecurity context. Mol. Ecol. Resour. 3: 443-453. doi: 10.1111/1755-0998.
665 Smith, K.F., Wood, S.A., Mountfort, D., and Cary, S.C. 2012. Development of a real-time PCR
666 assay for the detection of the invasive clam, Corbula amurensis, in environmental samples. J.
667 Exp. Mar. Biol. Ecol. 412: 52-57. doi.org/10.1016/j.jembe.2011.10.021.
668 Soliman, T., and Inglis, G.J. 2018. Forecasting the economic impacts of two biofouling invaders
669 on aquaculture production of green-lipped mussels Perna canaliculus in New Zealand.
30 https://mc06.manuscriptcentral.com/genome-pubs Page 31 of 38 Genome
670 Aquacul. Environ. Interact. 10: 1-12. doi.org/10.3354/aei00249
671 .Stoeck, T., Zuendorf, A., Breiner, H.W., and Behnke, A. 2007. A molecular approach to identify
672 active microbes in environmental eukaryote clone libraries. Microbial Ecol. 53: 328–339. doi:
673 10.1007/s00248-006-9166-1.
674 Strickler, K.M., Fremier, A.K., and Goldberg, C.S. 2015. Quantifying effects of UV-B,
675 temperature, and pH on eDNA degradation in aquatic microcosms. Biol. Conserv. 183: 85–92.
676 doi.org/10.1016/j.biocon.2014.11.038.
677 Taberlet, P., Coissac, E., Pompanon, F., Brochmann, C., and Willerslev, E. 2012. Towards next-
678 generation biodiversity assessment using DNA metabarcoding. Mol. Ecol. 21: 2045–2050.
679 doi:10.1111/j.1365-294X.2012.05470.x
680 Takahara, T., Minamoto, T., Yamanaka,Draft H., Doi, H., and Kawabata, Z., 2012. Estimation of fish
681 biomass using environmental DNA. PLoS One 7: e35868. doi:10.1111/j.1365-
682 294X.2012.05470.x.
683 Takahara, T., Minamoto, T., and Doi, H. 2013. Using environmental DNA to estimate the
684 distribution of an invasive fish species in ponds. PLoS One 8: e56584.
685 doi:10.1371/journal.pone.0056584
686 Te, S.H., Chen, E.Y., and Gin, K.Y. Comparison of Quantitative PCR and droplet digital PCR
687 multiplex assays for two genera of bloom-forming cyanobacteria, Cylindrospermopsis and
688 Microcystis. Appl. Environ. Microbiol. 81: 5203-5211. doi:10.1128/AEM.00931-15.
689 Thomsen, P.F., Kielgast, J., Iversen, L.L., Møller, P.R., Rasmussen, M., and Willerslev, E. 2012a.
690 Detection of a diverse marine fish fauna using environmental DNA from seawater samples.
691 PLoS One 7: e41732. doi.org/10.1371/journal.pone.0041732.
692 Thomsen, P., Kielgast, J., Iversen, L., Wiuf, C., Rasmussen, M., Gilbert, M., and Willerslev, E.
31 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 32 of 38
693 2012b. Monitoring endangered freshwater biodiversity using environmental DNA. Mol. Ecol.
694 21: 2565–2573. doi:10.1111/j.1365-294X.2011.05418.x.
695 Tréguier, A., Paillisson, J.-M., Dejean, T., Valentini, A., Schlaepfer, M.a., and Roussel, J.-M.,
696 2014. Environmental DNA surveillance for invertebrate species: advantages and technical
697 limitations to detect invasive crayfish Procambarus clarkii in freshwater ponds. J. Appl. Ecol.
698 51: 871–879. doi:10.1111/1365-2664.12262.
699 Tsai, Y.-L., Wang, H.-T. T., Chang, H.-F. G., Tsai, C.-F., Lin, C.-K., Teng, P.-H., Su, C., Jeng, C-
700 C., Lee, P.-Y. 2012. Development of TaqMan probe-based insulated isothermal PCR (iiPCR)
701 for sensitive and specific on-site pathogen detection. PLoS ONE, 7: e45278.
702 doi.org/10.1371/journal.pone.0045278.
703 Ulibarri, R.M., Bonar, S.A., Rees, C., DraftAmberg, J., Ladell B., and Jackson C. 2017. Comparing
704 efficiency of American Fisheries Society standard snorkeling techniques to environmental
705 DNA sampling techniques. N. Am. J. Fish. Manag. 37: 644-651.
706 doi.org/10.1080/02755947.2017.1306005.
707
708 Wallentinus, I., and Nyberg, C.D. 2007. Introduced marine organisms as habitat modifiers. Mar.
709 Pollut. Bull. 55: 323-332. doi.org/10.1016/j.marpolbul.2006.11.010.
710 Wong, N.A., McClary, D., and Sewell, M.A. 2011. The reproductive ecology of the invasive
711 ascidian, Styela clava, in Auckland Harbour, New Zealand. Mar. Biol. 158: 2775–2785.
712 doi.org/10.1007/s00227-011-1776-6.
713 Wood, S.A., Smith, K.F., Banks, J.C., Tremblay, L., Rhodes, L., Mountfort, D., Cary, C.S., and
714 Pochon, X. 2013. Molecular genetic tools for environmental monitoring of New Zealand’s
715 aquatic habitats, past, present and the future. NZJ Mar. Freshwat. Res. 47: 90-119.
32 https://mc06.manuscriptcentral.com/genome-pubs Page 33 of 38 Genome
716 doi.org/10.1080/00288330.2012.745885.
717 Wood, SA, Zaiko, A., Richter I., Inglis, G., Pochon, X. 2017. Development of a real-time
718 Polymerase Chain Reaction assay for the detection of the invasive Mediterranean fanworm,
719 Sabella spallanzanii, in environmental samples. Environ. Sci. Poll. Res. 24: 17373-17382.
720 doi:10.1007/s11356-017-9357-y.
721 Woods, C., Seaward, K., Inglis, G., Rodgers, L.P. (2017). Marine High Risk Site Surveillance
722 Programme: Annual report for all High Risk Sites 2016–17 (SOW18048). MPI Technical
723 Paper No: 2017/45MPI.
724 Zaiko, A., Schimanski, K., Pochon, X., Hopkins, G.A., Goldstien, S., Floerl, O., and Wood S.A.
725 2016. Metabarcoding improves detection of eukaryotes from early biofouling communities:
726 implications for pest monitoringDraft and pathway management. Biofoul. 32: 671-684.
727 doi:10.1080/08927014.2016.1186165.
728 Zhan, A., Hulak, M., Sylvester, F., Huang, X., Adebayo, A.A., Abbott, C., Adamowicz, S.J.,
729 Heath, D.D., Cristescu, M.E., and MacIsaac, H.J. 2013. High sensitivity of 454 pyrosequencing
730 for detection of rare species in aquatic communities. Meth. Ecol. Evol. 4: 558-565.
731 doi:10.1111/2041-210X.12037.
732
33 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 34 of 38
Table 1. Diver survey and real-time PCR results from DNA samples. No Sabella spallanzanii was detected via either method in Nelson. p = number of piles surveyed (data are a total from two dive surveyors and shows the number of piles with S. clava, not the number of individuals). Data in m2 are average values from two divers who each surveyed ca. 50 m of continuous substratum (e.g., pontoon). S = number of samples (from a total of 3) taken per site that the target was detected in. Real-time PCR data are an average of n=2 or 3 replicates per sample. Samples where less than two of the triplicates were positive were excluded. The volume of water sampled was 45 L.
Nelson Auckland Real-time PCR Real-time PCR Real-time PCR Site Visual Site Visual S. Visual S. S. clava S. clava (copies/sample ± S. spallanzanii no. S. clava no. clava spallanzanii (copies/sample ± 1SE) 1SE) (copies/sample ±1 SE) 1 1.5 /m2 0 28 NS 117,000 (1S) NS 451,010 (±124,430; 3S) 2 1 /m2 0 29 NS 213,600 (1S) NS 603,840 (± 94,770; 2S) 3 2 (10 p) 0 30 NS 0 NS 1,094,690 (± 230,850; 3S) 4 3 (10 p) 14,400 (1S) 31 NS 1,372,800 (± 897,600; 2S) NS 715,260 (±189,120; 3S) 5 1 (10 p) 0 32 0 0 20 /m2 31,320 (± 21,060; 2S) 6 0.5 /m2 0 33 NS 0 NS 278,060 (± 39,500; 3S) 7 2 /m2 0 34 NS 0 NS 282,600 (± 118,700; 2S) 8 5 (10 p) 0 35 1 (10 p) 0 4.5 (10 p) 1,205,740 (± 150,013; 3S) 9 1 (10 p) 0 36 NS 0 NS 39,210 (± 12,870; 2S) Draft 2 10 5 (10 p) 0 37 0 0 25 /m 0 11 2 (5 p and pontoon) 0 38 0 0 30 /m2 0 12 1 /m2 0 39 5 (10 p) 352,000 (±143,980; 1S) 12 (20 p) 374,350 (± 85,170; 3S) 13 0 0 40 0 0 20 /m2 35,360 (± 13,490; 2S) 14 3 /m2 14,400 (± 9,700; 1S) 41 0 0 20 /m2 144,920 (± 75,580; 2S) 15 1 (1 p and pontoon) 0 42 NS 0 NS 233,050 (± 102,420; 3S) 16 1 (1 p and pontoon) 5,400 (1S) 43 NS 0 NS 100,020 (± 26,630; 2S) 17 4 (10 p) 0 44 3 (10 p) 0 10 (20 p) 72,710 (± 12,970; 3S) 18 2 (10 p) 0 45 0 0 6 (20 p) 770,280 (± 78,789; 3S) 19 0 0 46 0 0 30 /m2 0 20 0 3,840 (± 960; 2S) 47 NS 0 NS 512,222 (± 69,680; 2S) 21 0.5 /m2 0 48 0 0 10 (10 p) 0 22 0 9,200 (± 2,120; 1S) 49 0 0 10 (10 p) 0 23 0 0 50 2 (5 p) 0 10 (10 p) 18,750 (± 7,400; 1S) 24 0 0 51 1 /m2 0 3 /m2 136,560 (± 62,590; 1S) 25 0.5 /m2 0 52 0 0 5 /m2 113,480 (± 46,060; 3S) 26 0.5 /m2 0 53 0.5 /m2 0 10 /m2 0 27 1.25 /m2 0 54 0 0 17.5 /m2 0 55 0 0 0 0 56 NS 0 NS 0 57 0 0 2 /m2 0
34 https://mc06.manuscriptcentral.com/genome-pubs Page 35 of 38 Genome
Table 2. Summary of model selection statistics for Steyla clava and Sabella spallanzanii detection data. Akaike Information Criteria (AIC), calculated conservatively using the number of sampling sites as the effective sample size (57 and 30 for S. clava and S. spallanzanii, respectively). AICw represents Akaike weight of the model, K the number of parameters in the model and -2Log(L) is twice the negative log- likelihood value. Detection probabilities may vary among methods (m) and large-scale probability of occupancy, for S. clava may vary between locations. For the top-ranked models the following parameter estimates are shown: large-scale probability of occupancy (ψ), small-scale probability of occupancy (θ), detection probability (p). Where applicable, standard errors (SE) are given.
Styela clava Sabella spallanzanii
Model AIC AICw K -2Log(L) Model AIC AICw K -2Log(L) ψ(H),θ,p(m) 190.9 0.95 5 180.9 ψ,θ,p(m) 90.92 1 4 82.92 ψ,θ,p(m) 197.2 0.04 4 189.2 ψ,θ,p 110.62 0 3 104.62 ψ(H),θ,p 213.8 0 4 205.8 ψ,θ,p 222.29 0 3 216.3
Averaged estimates from the top-ranked models Location Real-time PCR Visual survey Real-time PCR Visual survey Auckland ψ(SE) 0.45 (0.12) 0.45 (0.12)Draft 0.96 (0.04) 0.96 (0.04) θ(SE) 1 1 1 1 p(SE) 0.14 (0.04) 0.54 (0.09) 0.66 (0.06) 1
Nelson ψ(SE) 0.92 (0.11) 0.92 (0.11) θ(SE) 1 1 p(SE) 0.14 (0.04) 0.54 (0.09)
35 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 36 of 38
Figure captions
Fig. 1. Location of sampling sites in Nelson (a) and Auckland (b), and results of diver surveys and
real-time PCR assay for the detection of Styela clava. See Table 1 for density and gene copy
number information. Diver surveys were not undertaken at all Auckland sampling sites due to
operational/safety restrictions.
Fig. 2. Location of sampling sites in Auckland (New Zealand), and results of diver surveys and
real-time PCR assay for the detection of Sabella spallanzanii. See Table 1 for density and copy
number information. No S. spallanzannii was detected at any of the Nelson sites via real-time or diver survey. Diver surveys were not undertaken at all Auckland sites due to operational/safety restrictions. Draft
Fig. 3. Occupancy probability for (a) Steyla clava, and (b) Sabella spallanzanii in Nelson (dark grey) and Auckland (light grey): naïve estimates (휓 , mean + 1 standard error (SE) from real-time
PCR and visual observations data and model estimates (ψ, mean + 1 SE) from the best selected
model (see Table 2).
36 https://mc06.manuscriptcentral.com/genome-pubs
Page 37 of 38 Genome
Draft
317x230mm (300 x 300 DPI)
https://mc06.manuscriptcentral.com/genome-pubs Genome Page 38 of 38
Draft
209x258mm (300 x 300 DPI)
https://mc06.manuscriptcentral.com/genome-pubs