Long-Term Spatio-Temporal Trends of Organotin Contaminations in the Marine Environment of Hong Kong

Hong Kong imposed a partial restriction on application of organotin-based antifouling paints in 1992. Since September 2008, the International Maritime Organization prohibited the use of such antifouling systems on all sea-going vessels globally. Therefore, it is anticipated a gradual reduction of organotin contamination in Hong Kong’s marine waters. Using the rock shell Reishia clavigera as a biomonitor, we evaluated the organotin contamination along Hong Kong’s coastal waters over the past two decades (1990–2015). In 2010 and 2015, adult R. clavigera were examined for imposex status and analysed for tissue concentrations of six organotins. We consistently found 100% imposex incidence in female R. clavigera across all sites. Tissue triphenyltin (TPT) concentrations were high in most samples. A probabilistic risk assessment showed that there were over 69% of chance that local R. clavigera would be at risk due to exposure to phenyltins. Comparing with those of previous surveys (2004–2010), both imposex levels and tissue concentrations of organotins did not decline, while the ecological risks due to exposure to organotins were increasing. We also observed high concentrations of monobutyltin and TPT in seawater and sediment from locations with intense shipping activities and from stormwater or sewage discharge. Overall, organotins are still prevalent in Hong Kong’s marine waters showing that the global convention alone may be inadequate in reducing organotin contamination in a busy international port like Hong Kong. Appropriate management actions should be taken to control the use and release of organotins in Hong Kong and South China.


Introduction
Organotin compounds (OTs), such as tri-butyltin (TBT) and tri-phenyltin (TPT), have been extensively used as active ingredients in antifouling paints on ship hulls, open-sea mariculture cages and other submerged marine infrastructures, and as pesticides since the 1960s [1]. These analysis, decision to publish, or preparation of the manuscript.
Competing Interests: The authors have declared that there are no competing interests between the employment, consultancy, patents, products in development, marketed products of the John Swire & Sons Limited, a company which kindly provided the James Henry Scott (Hong Kong) PhD Scholarship to one of our co-authors, Dr. Elvis G. B. Xu, and the objective and results of this research project. This does not alter the authors' adherence to PLOS ONE policies on sharing data and materials.

Imposex determination
Each animal was measured for its shell length and fresh weight using vernier calipers (SPI 31-415-3, USA) and an electronic balance (Libror EB-430HU, Japan), respectively, then was cracked using a bench vice. Their reproductive features were examined using a dissecting microscope (Olympus SZH10, Japan). Penis length was determined using the scale in the microscopic lens. Vas Deferens Sequence Index (VDSI) and Relative Penis Size Index (RPSI) were evaluated (S3 Table) by the same person to ensure consistency. Seawater, in/effluent and sediment samples Seawater and sediment were collected from five sampling stations which are proximal to shipping activities (Fig 1; see S4 Table for geographical coordinates of sites A to E) during January 2014. One litre of mixture of surface and sub-surface water (ca. 0.5 m below the water surface) and one kilogram of sediment were collected in duplicates using Van Dorn water sampler/glass bottle and Ekman grab, respectively. The seawater sampling was repeated in 2015, at 10 selected locations covering different potential sources of OTs. The influent and effluent of Stanley Sewage Treatment Works were also collected (S4 Table with geographical coordinates of all sampling sites). All samples were collected in two or three replicates, and were kept in ice and sent to the laboratory for analysis within 6 h of collection.

Chemical analysis
Rock shells from each site were pooled as three or four replicates, each having 8-15 individuals. We adopted established methods for analyses for OTs ( [21] for rock shell and sediment; [22] for water with slight modifications; see S5 Table). HPLC grade solvents were used (Tedia, USA). Chemical standards were purchased from Sigma-Aldrich (St. Louis, USA) and Chiron (Trondheim, Norway). Quantification of six OTs, including MBT, DBT, TBT, monophenyltin (MPT), diphenyltin (DPT) and TPT, was conducted using a gas chromatograph (GC; Agilent 6890, USA) equipped with a mass-selective detector (Agilent 5973). A DB-5MS fused silica capillary GC column (J&W Scientific Inc., USA) was used with film thickness of 0.25 mm i.d. × 30 m × 0.25 μm (more analytical parameters are available in S6 Table). Satisfactory recovery rates were obtained for the six OTs ranging from 64.6%-93.5% except DPT (S7 Table). The method was further validated by using a certified reference material [21]. A procedural blank was simultaneously analysed every batch of five samples. The detection limits were estimated at 0.2-1.5 μg kg -1 dry weight (dw) for tissue and sediment samples, and 0.4-2.6 ng L -1 for water samples (except 29.1 ng L -1 for MBT). All concentrations were reported without correction to the recovery rates.

Statistical analysis
Mean VDSI, RPSI, percentage of sterile female, tissue TBT, TPT and total OTs concentrations were temporally compared among surveys in 2004 [14], 2005/06 [20], 2010 and 2015 (present study) using paired-samples t test (with log 10 -transformed data) or Wilcoxon signed-rank test. Correlations among tissue OTs concentrations, imposex indices and condition index (= fresh tissue weight x 100 / (fresh tissue weight + dry shell mass) [24] were tested using Spearman's rank correlation analyses followed by sequential Bonferroni correction [25]. Relationships of the distance between the sampling site and its nearest harbour or marina (estimated using Google Map) and imposex indices, OTs concentrations and condition index were tested using Spearman's rank correlation analyses with Bonferroni correction.
All statistical tests were conducted using SPSS Statistics 19.0 (SPSS Inc., USA) and Microsoft EXCEL 2003 (Microsoft Corporation, USA).

Probabilistic ecological risk assessment
To evaluate the ecological risk of BTs and PTs on the populations of R. clavigera, a probabilistic risk assessment was conducted by calculating the risk quotient (RQ; see S8-S10 Tables):

RQ ¼
Measured tissue concentration ðMTCÞ Predicted no effect tissue concentration ðPNETCÞ If the calculated RQ 1, the R. clavigera population is at risk due to the exposure to BTs or PTs [14]. Distributions of RQs (for BTs and PTs respectively), based on the distributions of respective MTCs and PNETCs which were fitted with log-logistic or Pareto models, were computed using Monte Carlo simulation with 10,000 iterations for 10 times. The RQ distribution was truncated as only RQ 0 was relevant. The simulations were carried out using @Risk 5.7 (Palisade Corporation, USA).

Correlations among OT concentrations, imposex indices and condition index
Tissue concentration of total OTs was positively correlated with TPT concentration (in 2010: r s = 0.994, p < 0.001; S14 Table and in 2015: r s > 0.999, p < 0.001; S15 Table). There was, in general, good agreement among mean VDSI, median VDSI and RPSI although sometimes it was statistically insignificant after Bonferroni corrections. A significant positive relationship was also found between tissue TBT concentration and RPSI (r s = 0.560, p < 0.01) in 2010, and between tissue TPT concentration and RPSI (r s = 0.855, p < 0.01) in 2015 (S14 and S15 Tables).

Relationships between shipping activities and OT contaminations
In general, imposex status (VDSI, RPSI and percentage of sterile females) and tissue concentrations of OTs decreased with the distance to the nearest facility with high shipping activities as shown in 2010 and 2015, although some combinations showed statistically insignificant results after Bonferroni corrections (S16 Table).

OT concentrations in seawater and sediment
Total OTs concentrations ranged from 83.7 to 255.5 ng g -1 dw in sediment, and 20.5 to 41.9 ng L -1 in seawater, respectively measured in 2014. The two sampling stations in Aberdeen (A and B) showed the highest OTs concentrations in seawater and sediment (Fig 2B and 2C). TPT was the most abundant among all OT residues, accounting for 45-63% in seawater and 71-72% in sediment, respectively. TBT accounted for a higher percentage among OTs in seawater (25-37%) than that in sediment (13%).
In 2015, the range of total OTs concentrations was between 74.5 and 753.3 ng L -1 in seawater and sewage (Fig 2C; S4 Table). Among all, the influent of the sewage treatment plant showed the highest OTs concentration. MBT was the predominating compound (75-95% of total OTs) in these samples.

Ecological risk assessment
Cumulative distributions of six PNETC values were constructed for BTs (S8 Table) and PTs (S9 Table), respectively. In 2010, all 29 sites were seriously impacted by PTs, as shown by the overlap of MTC and PNETC distributions (Fig 3A). This situation continued in 2015, where all 10 sites had MTCs higher than the lowest PNETC. The results of the probabilistic risk assessment suggested that local R. clavigera collected in 2010 had 17.6% of chance to be at risk from exposure to PTs (i.e., RQ 1; Fig 3B), while the chance that R. clavigera being at risk has significantly increased to 69.4% in 2015 (Fig 3C).
Using total BTs instead of TBT, we reconstructed the distribution of tissue concentrations of total BTs and the computed risk quotients for 2004 (data extracted from Leung et al. [12]). The number of sites that was severely impacted by BTs reduced from five (2004) to two (2010), while all 10 sites had MTCs lower than the smallest PNETC in 2015 (Fig 4A). The chance for R. clavigera being impacted by BTs decreased from 1.6% to 0.7% during 2004 to 2010, but slightly increased to 3.4% in 2015 (Fig 4B-4D).

Spatial variation of OT contamination
Generally, the eastern waters of Hong Kong, except inner Sai Kung, exhibited lower OT contamination (i.e., low VDSI and RPSI) in R. clavigera, due to less intense shipping activities and the prevailing ocean current helped mix, dilute and wash away the pollutants. The western waters, however, were more polluted due to (1) intense shipping activities around the container terminals in Kwai Chung, Tsing Yi and Tuen Mun and (2) contaminated freshwater input  Table 1 Table 1 from the Pearl River Delta (PRD) originated from agricultural and industrial activities. Conventionally, the water, sediment and biota in PRD were heavily contaminated by OTs [26]. For example, TBT concentration in PRD's estuarine water could reach 38.5 ng L -1 which was 4-20 folds higher than the environmental standards in the United States and United Kingdom [27]. We also detected higher concentrations of OTs in seawater samples (over 200 ng L -1 ) from the western waters of Hong Kong (sites F, G, H and I; Fig 2C).
Specifically, areas of fish villages and typhoon shelters with intense shipping activities such as Sok Kwu Wan and Aberdeen, exhibited the highest levels of VDSI in R. clavigera. Sai Kung Pier, which recorded the highest RPSI and tissue concentrations of TPT and total OTs, is a hotspot for ferries and leisure boats especially during summer. The highest TBT levels in R. clavigera were observed in Kadoorie Beach and Butterfly Beach, where are close to the typhoon shelter in Tuen Mun. These results were in line with the lowest PDI and BDI found in these areas, indicating recent inputs of TPT and TBT, respectively. The findings were also supported by the highest TBT concentration and the second highest TPT concentration found in seawater in the same location (site H). Moreover, the positive association between OT contamination and proximity to shipping activities was consistent with previous studies [28], and this further supported that marine shipping activities are highly possibly a major source of OT contamination. The input of TPT might reflect the replacement of TPT from TBT as the active antifouling agent in Hong Kong and South China which deserves further investigations on the ingredients of the currently used antifouling paints.
The concentrations of OTs in sediment and in seawater measured in this territorial-wide study, i.e., in sediment (83.7-255.5 ng g -1 dw) and in seawater samples of 2014 (20.5-41.9 ng L -1 ) and 2015 (74.5-461.1 ng L -1 ), were much higher than those measured in a parallel study that sampled in four marine protected areas in Hong Kong (46.2-126.4 ng g -1 dw in sediment and 0.8-7.8 ng L -1 in seawater; Xu et al., unpublished results). Considering the results in 2015, the highest concentrations of butyltins (mainly MBT and DBT) in water were linked to stormwater drainage (site F, G), nullah discharge (site K and I) and sewage (influent and effluent of treatment plant). This could be due to the industrial and domestic releases of such compounds from PVC pipes, antifungal coatings on textiles and wood preservatives [29] especially during the wet summer season when there were increased amounts of rainfall and surface runoff. Similarly, we observed increased concentrations of nonylphenols in a marine reserve in Hong Kong during summer [30]. In contrast, phenyltins were likely associated with shipping activities as sites C, F and H showed the highest concentrations among seawater samples, as well as in the influent of the sewage treatment plant (S4 Table). In Hong Kong, over 80% of population uses seawater for toilet flushing and such saline wastewater enters sewage treatment plants [31]. Thus, it is not surprised to see the sewage influent being contaminated with organotins. Overall, OT contamination is still a widespread problem in Hong Kong which requires more in-depth investigations on the sources of these pollutants.

Temporal trends of OT contamination
We consistently showed 100% imposex in female R. clavigera and further noted the increases of imposex indices and the percentage of sterile females in 2010 and 2015, suggesting that OT contamination is persistent. Interestingly, although there was a reduction of TBT concentration in R. clavigera during 2010 to 2015, the ecological risks of the rock shell populations exposed to BTs increased over the period. This is attributed to the increases in concentrations of MBT and DBT, in which the former compound is the most abundant among all OTs found in the seawater samples. However, in certain less impacted sites such as Shek Mei Tao, Clear Water Bay and Chek Chau, there were elevated tissue concentrations of TBT in 2010 comparing to those in 2004 and the maximum amount of increase was 384.5 μg kg -1 dw, or 195 folds. This may be attributed to the proximity of these sites towards the fish culture zones in Po Toi O and Tap Mun. Although TBT was detected at low concentrations in the seawater samples in the present study, it could be due to the fact that TBT is easily degraded into DBT, and further into MBT in seawater as compared to their degradation rates in sediment and biota [30]. Moreover, marine organisms can readily take up the pollutants once they passively diffused from sediment into seawater [32] in addition to the terrestrial sources of butyltins. The lack of regulations to control the use and release of these compounds [33], or the lack of proper enforcement of existing regulations [34] worsens the situation. Not surprisingly, high levels of TBT were still detected in Asia where there is a lack of comprehensive regulations on its use and release [35,36], or there are antifouling fragments remained in shipyards [37] and illegal uses of TBT in places such as mariculture farms [38].
As TPT is applied as co-toxicant in antifouling systems with TBT [39], these two compounds often coexisted in marine organisms with high concentrations [40], and both have comparative potencies in triggering imposex in R. clavigera [5,41]. The present study was the first one quantifying concentrations of various PTs (i.e., MPT, DPT and TPT) in the tissues of R. clavigera and in coastal water and sediment from Hong Kong. In the past, TPT and its degradation products were seldom measured in any compartment (i.e., water, sediment and biota) of local waters. Nakayama et al. first reported higher TPT concentration (up to 400 ng g -1 ww) in the liver of the finless porpoise (Neophocaena phocaenoides) collected locally as compared to cetaceans from other Asian locations [42]. Recent studies also documented TPT as the dominating compounds among OTs in several gastropod species including R. clavigera collected in Shenzhen [43,44], and in several seafood species sampled in Hong Kong [23] suggesting that the coastal marine environment of South China was severely contaminated with TPT. The present study confirmed the high concentrations of TPT not only in marine organisms but also in seawater and sediment from Hong Kong waters, whereas the concentrations of TBT relatively low.

Comparison of OT contamination with other regions
We found fluctuations in concentrations of TBT in R. clavigera. This does not agree with the global decreasing trend [45] which could possibly due to the terrestrial input of TBT and the passive diffusion from sediment. While plentiful information of TBT is available, only limited monitoring studies had incorporated measurements of TPT. Shim et al. recorded TPT concentration in R. clavigera up to 2,460 ng Sn g -1 dw (ca. 7,256 μg TPT kg -1 dw) in Korea [40], which was comparable to the highest TPT concentration (i.e., 15,060 μg kg -1 dw) measured in the present study. In Japan, however, tissue TPT concentrations in R. clavigera were relatively low and decreasing [46]. Nonetheless, notable amount of TPT, with comparison of TBT, was detected in sediment from a Japanese fishing port [47]. In agreement with the present study, higher tissue concentrations of TPT, compared to TBT, were also detected in R. clavigera from Korea [40] and Taiwan [48]. In fishes collected from Bohai Bay, China, TPT concentrations (34.7 ng g -1 wet weight) were higher than those of TBT [49], but they were much lower than the concentrations measured in gastropods in the present study. This suggested that contamination of TPT could be a widespread problem in Asia. A more comprehensive study covering a wider range of areas is needed to understand the environmental fate and contemporary sources of OTs especially TPT in Asian coasts.

Implications for management action
Being an associate member of IMO which has yet to ratify the AFS Convention [50], Hong Kong demonstrated to other countries or areas of similar status, such as Thailand, Vietnam and Philippines, that only internationally-bound regulatory measures are not sufficient to control chemical contamination, such as OTs, in regional scale. Based on the experience from the US, Europe and Japan, local legislations in controlling the use of OTs as antifouling paints, pesticides, fungicides and other industrial uses should be more promising and effective in reducing OT contamination [51]. The Legislative Council of Hong Kong passed the Merchant Shipping (Control of Harmful Anti-Fouling Systems on Ships) Regulation (under Cap. 413) on 3 June 2015, and the government is now requesting the Central People's Government of China to notify the IMO the extension of the Convention to Hong Kong [52]. The present study, therefore, call for urgent implementation of the Convention and appropriate management actions to remediate the pollution and impacts of these compounds in the marine environment of this region.

Conclusions
Although OTs, as antifouling agents, were banned globally, they have been contaminating the coastal marine environment of Hong Kong over the past two decades and causing ecological risk to local rock shell populations. This study demonstrated that the implementation of the international regulation alone may not be effective enough in controlling widespread environmental problems, local and regional management actions are urgently needed to remediate the pollution of these compounds from various sources.
Supporting Information S1   Table. Spearman's rank correlation analyses between the distance to major shipping activities and imposex status (including Vas Deferens Sequence Index (VDSI) and Relative Penis Size Index (RPSI)), condition index and tissue concentrations of organotins. (DOCX) S17 Table. List of references of Supporting Information (S1-S16 Tables). (DOCX)