Bryophyte Communities along a Tropical Urban River Respond to Heavy Metal and Arsenic Pollution

Aquatic and rheophilous bryophytes can indicate water pollution as they bioaccumulate toxic water elements. We evaluated (1) bioaccumulation of eight heavy metals and arsenic by Marchantia polymorpha L., and (2) changes in bryophyte community structure, as responses to urban pollution in southern Ecuador. To this end, we registered presence/absence and coverage of submerged bryophytes in 120 quadrats across three zones of the Zamora river inside Loja city, and a control zone in a nearby forest. We found that the concentrations of five (Al, Cd, Cu, Fe, and Zn) of the eight chemical elements and arsenic were highest in urban M. polymorpha. Moreover, bryophyte species richness decreased in urban zones. Bryophyte community structure also differed between control and city zones, but no differences were found among city zones. The control zone was composed by a more distinct set of bryophyte species, e.g., an indicator species analysis showed that 16 species had high and significant indicator values for control zone, but only 11 species were indicators of at least one of the three urban zones. We concluded that bryophytes, in general, and M. polymorpha, in particular, can be suitable biomonitors of water quality in tropical urban rivers.


Introduction
Communities occurring along urban rivers are increasingly near sources of pollution such as municipal wastewater, domestic garbage, and agricultural and industrial discharges [1][2][3].Most of these effluents contain toxic substances, like heavy metals and metalloids, which are a considerable threat to the environment [4,5], and human health [6][7][8][9].As such, looking for a sensitive and effective indicator of water pollution is an important task for scientists and local authorities that can inform decision making and city planning [10].
Bryophytes are well-known bioaccumulators of toxic elements due to its eco-physiology (i.e., rapid absorption and slow desorption of pollutants [10][11][12]), and morphology (bryophytes lack of epidermis that allows them to accumulate toxins present in water [13,14]).Recent studies have shown that some bryophyte species living in contaminated rivers, like the thallose liverwort Marchantia polymorpha L., can be enriched by heavy metals like Cu, Zn, and Cd [12,13].Moreover, bryophyte community structure can also respond to natural and anthropogenic variability throughout a river's profile.For instance, it has been shown that bryophyte communities change with turbidity, water temperature, and pH in the Italian Alps and Apennines [15,16].Water pollution [10] and level of urbanization [15,[17][18][19][20] also have an impact in other European bryophyte community structures.
Despite their great importance as bioindicators and bioaccumulators, most bryophyte studies have been carried out in temperate regions of the world [10].In Ecuador, most bryophyte studies have been taxonomic in scope [21][22][23][24] and few studies has focused on monitoring water contamination using aquatic macroinvertebrate communities [25].Here, we present the first analysis of the effects of water pollution on bryophyte bioaccumulation of eight heavy metals and one metalloid.We further test if bryophyte community structure, beyond effects on individual species, responds to river pollution.To this end, we measured bryophyte species richness and community composition in the Zamora river, passing through Loja city, a major city in southern Ecuador.We asked the following questions: (1) Is M. polymorpha bioaccumulating toxic elements inside Loja city?(2) How do bryophyte species richness and community structure respond to water pollution?(3) Are there bryophyte species strictly associated either to forest or city areas?Previous research has shown a higher concentration of pollutants in centric zones of cities [10,14,20,[26][27][28][29].

Study Area
We performed this study in the streams and river banks of the Zamora river, passing through Loja city; which has been eroded by wastewater contamination, garbage, and extraction of stony material [30].The study area corresponds to urban zones and a forest fragment of Loja city.The approximate area of the city is 10,790 km 2 .The annual average temperature of Loja is 15 • C, and the annual precipitation is about 900 mm [30].The study consisted of three study zones (south, center, and north) within city limits and along the river, and a control zone (forest) outside the city.We sampled 3 sites within each zone, for a total of 12 sites (Figure 1).
The control zone (F) includes the upper parts of the river basin, with banks dominated by forest remnants [25].The south zone (S; 1035.000m 2 of area or recreational parks; 68,919 inhabitants), is characterized by recent urban development and lack of adequate supply of sewage.According to the water quality index Canadian Council Ministers (CCME-WQI) and water quality index (WQI-C), the water is considered to be regularly contaminated or poor in quality.Arsenic, mercury, and lead in the rivers reach values up to 0.326 ppm, 0.022 ppm, and 0.389 ppm, respectively [31].The center zone (C; 635,000 m 2 of area; 54,576 inhabitants) is characterized by a high degree of urbanization and high number of effluents of sewage.According to the CCME-WQI and the WQI-C, the water is considered to be highly contaminated or poor in quality.In this area, levels of arsenic, mercury, and lead can reach values up to 0.043 ppm, 0.175 ppm, and 0.664 ppm, respectively [31].The north zone (N; 1060.000m 2 of area; 26,527 inhabitants), finally, is an urban area with a high storage of microbiological load due to sewage [30]; nevertheless, the zone still has some recreational parks.According to the CCME-WQI and WQI-C, the water is considered to be highly contaminated or poor in quality.The levels of arsenic, mercury, and lead in this zone can reach values up to 0.022 ppm, 0.043 ppm, and 0.688 ppm, respectively [31].

Elemental Bioccumulation in Marchantia polymorpha L.
Elemental bioaccumulation was studied in the subcosmopolitic thallose liverwort Marchantia polymorpha L. We chose this species because it has proved elsewhere to be an effective bioaccumulator of heavy metals [12,13,32].To sample M. polymorpha, we took four samples (0.5-1 g) of the species in the same sites (two separate sites by zone), where bryophytes frequency and cover were sampled at each of the four study zones.They were rinsed with Milli-Q water, then stored in paper bags for oven drying at 60 °C for 3-4 days.Fieldwork was conducted between April and December 2015.
The dried samples were mechanically milled using a stainless-steel grinder for digestion.Subsequently, 0.2 g of dry sample was weighed, then a preparation of aqua regia (HCl and HNO3)

Elemental Bioccumulation in Marchantia polymorpha L.
Elemental bioaccumulation was studied in the subcosmopolitic thallose liverwort Marchantia polymorpha L. We chose this species because it has proved elsewhere to be an effective bioaccumulator of heavy metals [12,13,32].To sample M. polymorpha, we took four samples (0.5-1 g) of the species in the same sites (two separate sites by zone), where bryophytes frequency and cover were sampled at each of the four study zones.They were rinsed with Milli-Q water, then stored in paper bags for oven drying at 60 • C for 3-4 days.Fieldwork was conducted between April and December 2015.
The dried samples were mechanically milled using a stainless-steel grinder for digestion.Subsequently, 0.2 g of dry sample was weighed, then a preparation of aqua regia (HCl and HNO 3 ) was performed in a 3:1 ratio (v/v) [33].Each sample was digested twice with 30 ml of aqua regia subjected to a heating plate with a temperature of 250 • C.After the samples were allowed to cool, the solutions were filtered through filter paper in a 100 ml distillation flask calibrated with distilled water.The concentrations of eight heavy metals-aluminum (Al), cadmium (Cd), copper (Cu), iron (Fe), magnesium (Mn), mercury (Hg), lead (Pb), and zinc (Zn), and one metalloid, arsenic (As)-in digested solutions were analyzed using a flame atomic absorption spectrophotometer Perkin Elmer AAnalyst 400 (Shelton, CT, USA).The respective wavelength (nm), precision (as the relative standard deviation, %), and limit of detection (µg g −1 ) of the elements were as follows: Pb 283. 31

Bryophyte Community Structure
In each of the 12 zones, ten 20 × 30 cm quadrats were selected along the banks of the Zamora river [34].In each quadrat, we registered presence/absence and coverage of rheophilous (permanently submerged) and aquatic (periodically submerged) bryophytes.The samples were identified using general [35,36] and specific keys [37] and were deposited as vouchers in the herbarium HUTPL at Universidad Técnica Particular de Loja.

Data Analysis
To determine changes in heavy metal accumulation of M. polymorpha across the different zones, we used one-way analysis of variance (ANOVA) followed by a Tukey post-hoc test.Shapiro Wilk and Bartlett's test confirmed the model met normality and homogeneity of variance assumptions.However, when data were non-normal, we used Kruskal-Wallis one-way ANOVA on ranks followed by Dunn's pairwise comparison.Analyses were done in R 3.2.2[38] using the "dunn.test"package [39].
We calculated species richness as the total number of different bryophytes species occurring in a quadrat.We also calculated sampling completeness with the Chao 2 species richness estimator using the R package 'vegan' [40].The effect of environmental variables (zone, site, slope, and plant cover) on species richness was modeled by fitting generalized linear mixed models (GLMMs) [41] as implemented in the R package 'nlme' [42].The minimal adequate model was selected based on Akaike's information criterion (AIC).Data were analyzed from a multi-level approach, considering site as a random factor and introducing the environmental variables as fixed factors (zone, plant cover, and slope).To identify significant richness differences between zone, a post hoc Tukey multiple comparison tests as implemented in the R package 'lsmeans' [43].The richness models of bryophytes were fitted with Poisson errors.
To detect for changes in bryophyte community structure among zones we used a permutational multivariate analysis of variance (PERMANOVA) on the species cover matrix [44].Experimental design included two factors: Zone (four levels, fixed factor), site (three levels, random factor nested within zone), plant cover (fixed factor), and slope (fixed factor), with 10 replicate quadrats for each locality (120 quadrats).We used the Bray-Curtis distance measure and 999 Monte Carlo permutations.We visualized community level composition with a non-metric multidimensional scaling (NMDS) on the species cover matrix.NMDS and PERMANOVA were conducted in vegan.
To determine which bryophyte species was associated with specific zone, we used the indicator species analysis [45] as implemented in the R package 'labdsv' [46].This analysis calculates an indicator value for each species based on the mean cover of each species per zone, which results from multiplying the relative abundance for each species by the frequency for each species in each zone.The indicator value ranges from 0 to 1, or 0 to 100.The significance was tested using a Monte Carlo permutation with 1000 replicates.

Elemental Bioaccumulation by Marchantia Polymorpha
The concentrations of heavy metals (Cd, Cu, Mn, Pb, Zn, Al, and Fe) and the metalloid arsenic in M. polymorpha significantly differed in the four zones (Figure 2).A greater concentration of Al, Cu, Fe, Mn, and Zn in M. polymorpha was measured in the three urbanized zones when compared with the control (Table 1 and Figure 2).No traces of Hg were detected across the sites, thus we removed Hg from further analysis.

Species Richness
A total of 44 bryophyte species were recorded, including 24 mosses, 18 liverworts, and 2 anthocerotes.The control zone showed the highest richness with 34 species, followed by the south zone (29 spp.), the central zone (24 spp.), and the north zone (21 spp.) (Figure 3).A similar pattern was observed with the Chao 2 richness estimator, confirming a high number of species estimated in the control zone (41 estimated species), followed by the south, center, and north zones (with 33, 26, and 22 estimated species, respectively).

Species Richness
A total of 44 bryophyte species were recorded, including 24 mosses, 18 liverworts, and 2 anthocerotes.The control zone showed the highest richness with 34 species, followed by the south zone (29 spp.), the central zone (24 spp.), and the north zone (21 spp.) (Figure 3).A similar pattern was observed with the Chao 2 richness estimator, confirming a high number of species estimated in the control zone (41 estimated species), followed by the south, center, and north zones (with 33
The NMDS plot (Stress = 0.20) showed that species composition of the Control zone (Forests) was significantly different from that of the other three zones (Figure 4).Results of the mixed models showed that the most relevant predictor for the bryophytes richness was the control zone (Table 2).

Bryophyte Community Structure
PERMANOVA showed that bryophyte communities varied significantly both at the zone (R 2 = 0.22; p = 0.001) and at the site (R 2 = 0.16; p = 0.001) but not plant cover (R 2 = 0.01; p = 0.096) (Table 3).Slope, while significant (p = 0.01), explained only 1% of the variance.The NMDS plot (Stress = 0.20) showed that species composition of the Control zone (Forests) was significantly different from that of the other three zones (Figure 4).

Indicator Species Analysis
We determined 16 indicator species of the control zone, 3 indicator species in the south, and 4 indicator species both in the north and central zones (Table 3).Fissidens serratus Müll.Hal and Monoclea gottschei Lindb., were the best indicator of Control zone.Noteroclada confluens (Hook.f. and Taylor) Spruce showed high indicator value for South zone.Campylopus pauper (Hampe) Mitt.and Marchantia polymorpha L. showed high indicator value for Center zone and Fissidens elegans Brid., for the North zone (Table A1).

Discussion
We found higher concentrations of four heavy metals (Al, Cu, Fe, and Zn) in M. polymorpha in city zones, which indicated its ability to bioaccumulate toxic elements.Thus, this species can be used as an effective biomonitor of heavy metal pollution [12].Our study adds M. polymorpha to the list of suitable species for this task.For example, several studies have highlighted the ability of Fontinalis antipyretica to absorb heavy metals (i.e., Al, Cu, Cr, Cd, and Zn) and As in rivers with urban, industrial, and agricultural residual discharges [14,27,47].Other bryophyte species used for this purpose in temperate regions include: Platyhypnidium riparioides, Fissidens polyphyllus, Brachythecium rivulare, Hygrohypnum ochraceum, Fontinalis antipyretica, F. squamosa, F. dalecarlica, F. duriaei, Thamnobryum alopecurum, and Taxiphyllum barbieri [10,27,28,[48][49][50].Our study, however, comes with a caveat; M. polymorpha is sometimes exposed to air, thus it can bioaccumulate air pollution in addition to water pollution.Future research should evaluate both the average time M. polymorpha remains submerged in a year and M. polymorpha's abilities to accumulate air contamination.

Indicator Species Analysis
We determined 16 indicator species of the control zone, 3 indicator species in the south, and 4 indicator species both in the north and central zones (Table 3).Fissidens serratus Müll.Hal and Monoclea gottschei Lindb., were the best indicator of Control zone.Noteroclada confluens (Hook.f. and Taylor) Spruce showed high indicator value for South zone.Campylopus pauper (Hampe) Mitt.and Marchantia polymorpha L. showed high indicator value for Center zone and Fissidens elegans Brid., for the North zone (Table A1).

Discussion
We found higher concentrations of four heavy metals (Al, Cu, Fe, and Zn) in M. polymorpha in city zones, which indicated its ability to bioaccumulate toxic elements.Thus, this species can be used as an effective biomonitor of heavy metal pollution [12].Our study adds M. polymorpha to the list of suitable species for this task.For example, several studies have highlighted the ability of Fontinalis antipyretica to absorb heavy metals (i.e., Al, Cu, Cr, Cd, and Zn) and As in rivers with urban, industrial, and agricultural residual discharges [14,27,47].Other bryophyte species used for this purpose in temperate regions include: Platyhypnidium riparioides, Fissidens polyphyllus, Brachythecium rivulare, Hygrohypnum ochraceum, Fontinalis antipyretica, F. squamosa, F. dalecarlica, F. duriaei, Thamnobryum alopecurum, and Taxiphyllum barbieri [10,27,28,[48][49][50].Our study, however, comes with a caveat; M. polymorpha is sometimes exposed to air, thus it can bioaccumulate air pollution in addition to water pollution.Future research should evaluate both the average time M. polymorpha remains submerged in a year and M. polymorpha's abilities to accumulate air contamination.
As expected, bryophyte species richness was highest in the Control (forest) zone.Ceschin et al. [51] have showed that greater bryophyte richness occurs in fast-flowing rivers, clean waters with high oxygenation, and good habitat characteristics for the growth of bryophytes (e.g., temperature, pH, turbidity).However, low bryophyte richness in urban zones of the Loja river might be explained in part by the absence of natural vegetation in the banks of the river, which results in less available habitat for the growth of bryophytes, and in part to the presence in the riverbed of metal walls and wooden covers [52].For example, Downes et al. [53] found that urban parts of rivers affect negatively bryophyte establishment and distribution.
Bryophyte community structure also changed significantly between control and urban zones, but no differences as one enters to the city zones.These differences were attributed to a greater frequency and coverage of sensitive species restricted to zones of natural forests where less urbanization and water pollution allow them to grow more.Thus, Scarlett and O'Hare [17] and Ceschin et al. [51] pointed that most sensitive bryophyte prefer clear, oxygenated waters with low concentrations of ammonium and phosphates.In our case, bryophytes communities in the control zone were better represented by species such as Symphyogyna brongniartii, Rhodobryum huillense, Plagiochila laetevirens, Monoclea gottschei, Dumortiera hirsuta, Neesioscyphus argillaceus, and Fissidens serratus (Appendix A), which have been previously identified as sensitive to changes in water quality [54][55][56].
Nonetheless, tolerant species were able to grow in waters of the center and north zones of the river, which present high levels of contamination, lack of oxygen, and greater amount of organic waste.For instance, Platyhypnidium aquaticum, M. polymorpha, Rhynchostegium scariosum, Thuidium delicatulum, and Riccia crassifrons are typical of disturbed habitats [51,[55][56][57], which agrees with our results (Appendix A).In addition, some studies point that Platyhypnidium aquaticum and Brachythecium rivulare are characteristic species of open environment with poor water conditions caused by urban spills [11,51,58].
The indicator species analysis showed that Clasmatocolea vermicularis, Fissidens serratus, Lophocolea bidentata, Monoclea gottschei, Noteroclada confluens, Symphyogyna brongniartii, Symphyogyna brasiliensis, Plagiochila laetevirens, and Rhodobryum huillense were the best indicators of the control zone, related to uncontaminated water.Accordingly, Lophocolea bidentata, Symphyogyna brongniartii, and Monoclea gottschei occur more abundantly in shady habitats with vegetation of undisturbed forests, due to the fact that they are sensitive to environmental changes [59,60].Conversely, Fissidens elegans, Riccia crassifrons, and Marchantia polymorpha were the best indicators of more urbanized zones such as in southern, central, and northern areas with greater impact of anthropogenic disturbances.Some authors point out that these species are considered tolerant to pollution, and therefore can be indicative of intense urban activity related with wastewater [19].
Our results revealed significant changes in bryophyte community structure, reduced species richness, and increases in heavy metal concentration in M. polymorpha in the Zamora river as you move inside Loja City.These results are likely a consequence of a greater degree of urbanization and wastewater deposition in the river and its tributaries.We conclude that bryophyte species richness decreases in urban zones; similarly, species composition changes significantly in the control zone with respect to urban zones of the river.These changes likely reflect higher concentration of heavy metals (Al, Cd, Cu, Fe, and Zn) in urban zones, which likely reflect the impact of human settlements and wastewater related water pollution.The complementary use of the diversity and bioaccumulation of heavy metals in bryophytes can provide key information of water pollution of Zamora River of Ecuador, which, in the long term, will allow for the establishment of monitoring zones for adequate management and strategies of mitigation of water pollution.Although our study demonstrates the efficacy of bryophyte in passive biomonitoring, we suggest that future work should evaluate bryophyte as part of active biomonitoring where transplantation of bryophytes allows for the evaluation and monitoring of the magnitude of water pollution due to the known exposure period [10,61].

Figure 1 .
Figure 1.Study area in Loja Province in southern Ecuador showing the location of four study zones.Control zone (F), south zone (S), center zone (C), and north zone (N).

Figure 1 .
Figure 1.Study area in Loja Province in southern Ecuador showing the location of four study zones.Control zone (F), south zone (S), center zone (C), and north zone (N).

Figure 2 .
Figure 2. Boxplot of heavy metals and metalloid accumulation in the Marchantia polymorpha collected from different zones in the Zamora River.Al, Cd, Fe, Mn, and Zn: One-way ANOVA followed by a Tukey test.As, Cu, and Pb: Kruskal-Wallis ANOVA followed by a Dunn's test.Different colors for each zone are considered significant (p < 0.05).

Figure 2 .
Figure 2. Boxplot of heavy metals and metalloid accumulation in the Marchantia polymorpha collected from different zones in the Zamora River.Al, Cd, Fe, Mn, and Zn: One-way ANOVA followed by a Tukey test.As, Cu, and Pb: Kruskal-Wallis ANOVA followed by a Dunn's test.Different colors for each zone are considered significant (p < 0.05).

Figure 3 .
Figure 3. Boxplot of bryophyte species richness at different zones along the Zamora river.Different colors correspond to post hoc Tukey groups (p < 0.05) after a GLMM.Control zone (F); south zone (S); center zone (C); and north zone (N).

Figure 3 .
Figure 3. Boxplot of bryophyte species richness at different zones along the Zamora river.Different colors correspond to post hoc Tukey groups (p < 0.05) after a GLMM.Control zone (F); south zone (S); center zone (C); and north zone (N).

Table 1 .
Mean concentration (µg g −1 ) and standard error (SE) of heavy metals and arsenic in Marchantia polymorpha by different zones.

Table 1 .
Mean concentration (µg g −1 ) and standard error (SE) of heavy metals and arsenic in Marchantia polymorpha by different zones.

Table 3 .
Results of PERMANOVA analysis of species composition and environmental variables (zone, site, slope and plant cover).p < 0.05 are considered significant.Df = degrees of freedom; SS = sum of squares; MS = Mean squares; F = statistical; R 2 = coefficient of variation.

Table 2 .
Results of the generalized linear mixed models (GLMM) on bryophytes richness and environmental variables (zone, site, slope, and plant cover).The random variable site was non-significant.

Table 3 .
Results of PERMANOVA analysis of species composition and environmental variables (zone, site, slope and plant cover).p < 0.05 are considered significant.Df = degrees of freedom; SS = sum of squares; MS = Mean squares; F = statistical; R 2 = coefficient of variation.

Table A1 .
Indicator values (IV) for bryophyte species.Species with indicator value >25 are considered as the best indicators.Species with both high indicator value and a significant p value (p < 0.05) are shown in bold.