There is no author summary for this article yet. Authors can add summaries to their articles on ScienceOpen to make them more accessible to a non-specialist audience.
Abstract
Barriers to microbial migrations can lead adaptive radiations and increased endemism.
We propose that extreme unbalanced nutrient stoichiometry of essential nutrients can
be a barrier to microbial immigration over geological timescales. At the oasis in
the Cuatro Ciénegas Basin in Mexico, nutrient stoichiometric proportions are skewed
given the low phosphorus availability in the ecosystem. We show that this endangered
oasis can be a model for a
lost world. The ancient niche of extreme unbalanced nutrient stoichiometry favoured survival
of ancestral microorganisms. This extreme nutrient imbalance persisted due to environmental
stability and low extinction rates, generating a diverse and unique bacterial community.
Several endemic clades of
Bacillus invaded the Cuatro Cienegas region in two geological times, the late Precambrian
and the Jurassic. Other lineages of
Bacillus,
Clostridium and
Bacteroidetes migrated into the basin in isolated events. Cuatro Ciénegas Basin conservation is
vital to the understanding of early evolutionary and ecological processes.
eLife digest
Water is a rare sight in a barren land, but there are many more reasons that make
the Cuatro Cienegas Basin, an oasis in the North Mexican desert, a puzzling environment.
With little phosphorous and nutrients but plenty of sulphur and magnesium, the conditions
in the turquoise blue lagoons of the Basin mimic the ones found in the ancient seas
of the end of the Precambrian. In fact, Cuatro Cienegas is one of the rare sites where
we can still find live stromatolites, a bacterial form of life that once dominated
the oceans. Many bacteria of marine origin exist alongside these living fossils, prompting
scientists to wonder if the Basin could be a true lost world, a safe haven where ancient
microorganisms found refuge and have kept evolving until this day. But to confirm
whether this is the case would require scientists to hunt for clues within the genetic
information of local bacteria.
Souza, Moreno-Letelier et al. came across these hints after sampling for bacteria
in a small (about 1km
2) lagoon named Churince, and analysing the DNA collected. The results yielded an astonishing
amount of biodiversity: 5,167 species representing at least two-third of all known
major groups of bacteria were identified, nearly as much as what was found in over
2,000 kilometres in the Pearl River in China. This is unusual, as most other extreme
environments with little nutrients have low levels of diversity.
Closer investigation into the genomes of 2,500 species of
Bacillus bacteria revealed that the sample increased by nearly 21% the number of previously
known species in the group. Most of these bacteria were only found in the Basin. These
native or ‘endemic’ species have evolved from ancestors that came to the area in two
waves. The oldest colonization event happened 680 million years ago, as the first
animal forms just started to emerge. The most recent one took place while dinosaurs
roamed the Earth about 160 million years ago, when geological events opened again
the Basin to the ancient Pacific Ocean.
Previous experiments have shown that different species of bacteria in the Churince
have evolved to form a close-knit community which ferociously competes with microbes
from the outside world. Paired with the extreme conditions found in the lagoon, this
may have prevented other microorganisms from proliferating in the environment and
replacing the ancient lineages.
The days of this lost world may now be numbered. Drained by local farming, the wetlands
of the Basin have shrunk by 90% over the past five decades. The Churince lagoon, the
most diverse and fragile site where the samples were collected, is now completely
dry. Human activities also disrupt the delicate and unique balance of nutrients in
the oasis. But all may not be lost – yet. Local high school students have become involved
in the research effort to describe and protect these unique microbial communities,
and to change agricultural traditions in the area. Closing the canals that export
spring water out of the Basin could give the site a chance to recover, and the microbes
that are now seeking refuge in underground waters could re-emerge. Maybe there will
still be time to celebrate, rather than mourn, the unique life forms of the Cuatro
Cienegas Basin.
ABSTRACT Reductive genomic evolution, driven by genetic drift, is common in endosymbiotic bacteria. Genome reduction is less common in free-living organisms, but it has occurred in the numerically dominant open-ocean bacterioplankton Prochlorococcus and “Candidatus Pelagibacter,” and in these cases the reduction appears to be driven by natural selection rather than drift. Gene loss in free-living organisms may leave them dependent on cooccurring microbes for lost metabolic functions. We present the Black Queen Hypothesis (BQH), a novel theory of reductive evolution that explains how selection leads to such dependencies; its name refers to the queen of spades in the game Hearts, where the usual strategy is to avoid taking this card. Gene loss can provide a selective advantage by conserving an organism’s limiting resources, provided the gene’s function is dispensable. Many vital genetic functions are leaky, thereby unavoidably producing public goods that are available to the entire community. Such leaky functions are thus dispensable for individuals, provided they are not lost entirely from the community. The BQH predicts that the loss of a costly, leaky function is selectively favored at the individual level and will proceed until the production of public goods is just sufficient to support the equilibrium community; at that point, the benefit of any further loss would be offset by the cost. Evolution in accordance with the BQH thus generates “beneficiaries” of reduced genomic content that are dependent on leaky “helpers,” and it may explain the observed nonuniversality of prototrophy, stress resistance, and other cellular functions in the microbial world.
Introduction Understanding how species adapt to novel environments is an important task both for understanding the dynamics of living systems and for predicting biotic responses to anthropogenic changes in the natural environment [1]–[3]. However, most studies of evolutionary adaptation consider single species in isolation. Although this approach is useful for uncovering genetic mechanisms, virtually all species co-occur with many other species. Faced with a new abiotic environment, communities might respond by evolution of component species, but ecological changes in species' abundances and distributions can also occur. If ecological interactions such as competition affect evolutionary responses [4],[5], then results from single species studies might not accurately predict evolutionary dynamics in diverse assemblages. Although there has been growing interest in how evolution affects ecological dynamics [6]–[10], most studies have still considered single species or pairs of interacting species. In addition, the question of how ecological interactions affect evolutionary responses to novel abiotic environments has received even less attention [11],[12]. If ecological interactions among species are weak, then evolutionary changes should be the same as those predicted in single species studies. However, if species use overlapping resources or otherwise interact, the extent and type of evolutionary responses might differ from those predicted if the same set of species each adapted to the new abiotic conditions in isolation [13],[14]. Several mechanisms might influence evolutionary dynamics in mixtures of species. First, species in diverse communities might have their resource use restricted by competitors, lowering effective population sizes and therefore reducing the rate at which beneficial mutations arise and the species adapts to a novel environment [5],[15]. In this scenario, species in communities should adapt to the new environment as they would in isolation, but the rate of adaptation would be reduced. Second, if trait variation among species exceeds variation within species, a new abiotic environment might act on the relative abundance of different species (ecological sorting) rather than on genetic variation within species [4],[5]. In models of this mechanism, pre-adapted species increase in abundance at the expense of less well-adapted species and the average amount of evolution in surviving species is typically reduced compared to responses of the same species in monoculture (although in rare scenarios the amount of evolution can increase [4]). Third, there might be a trade-off between adaptation to biotic and abiotic components of the environment [16]. Such trade-offs might result from the production of costly adaptations involved in species interactions such as defences [17],[18] or from selective interference between adaptations to the biotic and abiotic environment [19]. In this case, species that evolved in communities should be less well adapted to the abiotic environment than if they adapted in isolation and vice versa. In the most extreme case, species might adapt to use resources generated by other species [20], in which case they will evolve entirely different resource use depending on whether other species are present. These mechanisms could change both the magnitude and direction of evolutionary change in communities compared to predictions from single species studies. However, evidence for an effect of diversity is currently scarce. Experiments have shown that diversity can inhibit evolution; for example, Brockhurst et al. [21] showed that niche occupation restricts adaptive radiation of a single bacterial strain. Similarly, Collins [16] found that diversity limits adaptation to elevated CO2 in algae and Perron et al. [22] showed that diversity limits the evolution of multi-drug resistance (although this effect was alleviated by horizontal transfer of resistance mutations). However, these studies considered genetic diversity within species rather than species diversity within communities. There is abundant evidence that coevolution drives fast evolution between species with strong ecological interactions [23],[24] and that pairwise coevolution can change the direction of evolution compared to adaptation in isolation [14]. Furthermore, studies of diffuse coevolution have shown that adaptation of a focal species to particular interacting species, such as insect herbivores, is influenced by interactions with other species, such as vertebrate herbivores [25]–[28]. For example, character displacement of limnetic and benthic species pairs of sticklebacks only occurred in lakes with low species diversity of other fish [29]. To the best of our knowledge, however, the evolution of interactions among multiple species in a community has not been investigated using an experimental evolution approach. Evolutionary dynamics in diverse systems will have important consequences for ecosystem functioning in altered environments. Ecosystem functions such as decomposition and productivity emerge from the degree to which species are adapted to their biotic and abiotic environments [30]–[32]. Following a change in the environment, ecosystem functioning might be disrupted either because the species abundances change or because component species fail to adapt to the new environmental optimum. Alternatively, coevolution among species might act to enhance ecosystem properties, for example if species evolve complementary resource use and thereby increase ecosystem productivity [33]. Understanding of these processes is needed to predict how ecosystem functioning will respond to environmental changes over evolutionary timescales. Here, we test whether species diversity influences environmental adaptation and ecosystem functioning using naturally co-occurring decomposer bacteria from temporary pools around the roots of beech trees (Fagus sylvatica), which have previously been used successfully for experimental ecology [34],[35]. We chose five species of bacteria differing in colony colour and shape so that each species could be isolated from species mixtures (Tables S1 and S2). Sequencing of 16S rDNA showed that isolates belong to five different families (Table S1). We refer to them as species since they represent genetically and phenotypically distinct clusters that co-occurred naturally. Monocultures of each species and polycultures containing all five species were allowed to adapt to laboratory conditions by regular serial transfer on beech-leaf extract (Figure 1). Laboratory conditions represent a new environment and differ from wild tree-holes in several ways: tree-holes receive a larger quantity and variety of resources, are spatially complex, and have an unpredictable input of water and leaves, whereas laboratory cultures experienced regular dilution with uniform medium in a shaken container. Growth assays were used to determine evolutionary responses. We predicted that species should adapt to laboratory conditions by evolving faster growth rates on the beech tea medium, but that the presence of other species might change the direction and extent of adaptation by one of the mechanisms outlined above. 10.1371/journal.pbio.1001330.g001 Figure 1 Experimental design for the evolution experiments. (i) Stocks of wild isolates were grown up, each comprising a single starting genotype of each species. (ii) Experiments were started with each species in monoculture or in polyculture (all five species mixed together). (iii) To stimulate active growth and promote adaptation to the laboratory conditions, each culture was diluted 20-fold in fresh medium twice weekly for 8 wk. Tubes were shaken to prevent the formation of biofilms and maintain spatial homogeneity. Numbers of generations ranged from 60.9 to 82.2 across cultures and effective population sizes ranged from 5.3×105 to 9.9×106 (Table S3). (iv) Final cultures were plated on agar. (v) Single colonies of each species were isolated for growth assays described in the main text. To measure species interactions and changes in resource use, our approach was to grow one species on beech tea, then to filter-sterilize the medium and to assay the growth of a second species on the “used” beech tea. If the second species used similar resources to the first (i.e., if their niches overlapped), the second species should grow less well on “used” beech tea than on “unused” beech tea because its resources would have been consumed. If the two species were specialized on different resources (i.e., occupied different niches), the second species should grow equally well on “used” and “unused” tea. Finally, if the second species used resources produced by the first (called facilitation or cross-feeding [36]), the second species should grow better on “used” tea than on “unused” tea. While this method does not provide direct information on competitive interactions in mixtures, it provides a tractable and reproducible measure of changes in resource use of each species during evolution. Because other types of interaction, such as direct inhibition by bacteriocides [37], might also affect growth rates, we also used nuclear magnetic resonance (NMR) spectroscopic profiling of “used” and “unused” tea to investigate changes in resource use directly. Finally, we tested whether adaptation to the presence of other species affected productivity (rate of production of CO2) by reassembling communities with different evolutionary histories using isolates that either evolved in monoculture or co-evolved in the same polyculture. If adaptation increased community productivity, we expected communities reassembled with isolates that evolved in polycultures to be more productive than those reassembled with isolates that had evolved in monoculture. Results Growth Rates on Beech Tea of Monoculture Isolates Although able to grow on beech tea in the lab at the start of the experiment, one species (E) dwindled to low cell densities during the evolution experiment (Table S3) and was excluded from growth assays and subsequent experiments because it failed to re-grow from frozen cultures. Species A to D were recoverable in all treatments and were used for subsequent experiments. Across species, final isolates that evolved in monoculture grew on average faster than ancestral isolates of the same species on unused beech tea (dark bars, first and second rows, Figure 2), consistent with the prediction that they adapted to laboratory conditions of serial dilution in beech tea medium by increasing growth rates on this medium. The effect was significant in species B, C and D, which grew between 47% and 120% faster after evolving in monoculture compared to their ancestral isolates. Growth rates of evolved monoculture isolates of species A were not significantly different from its ancestral isolate. Note that phenotypic plasticity and parental effects can be discounted as explanations for differences among treatments. In all our assays, frozen isolates were first grown in beech tea medium for 4 d ( = 4 to 6 generations, Table S3), and then an aliquot was taken from these cultures to start the assay cultures. Differences in phenotypes between treatments were therefore maintained after several generations of growth in identical environments and cannot be readily explained by phenotypically plastic responses. 10.1371/journal.pbio.1001330.g002 Figure 2 Maximum growth rates of isolates after evolution under each diversity treatment. Maximum rate of growth from low densities, VMAX , of each species grown on unused beech tea under assay conditions. Dark bars, growth rates of ancestral isolates. Mid grey bars, growth rates of monoculture isolates. Pale bars, growth rates of polyculture isolates. Standard error bars are shown. Tukey Honest Significant Difference test contrasts between treatments: *** p 0.95, which might indicate multiple peaks derived from the same compound. Contaminant peaks derived from methanol and acetonitrile were removed from the dataset. Variation in resource use and production across isolates was explored using principal components analysis of unscaled variances implemented with the prcomp() function in R [55]: we used unscaled rather than scaled variances to focus on compounds showing larger changes in their absolute concentrations. Productivity of Assembled Communities MicroResp kits were used to measure community respiration. Respired CO2 results in a change in colour of cresol red indicator dye suspended above each well of a 96-well plate. Ten replicates were used per treatment in a single plate and the experiment was repeated in triplicate. Each well contained 840 µl of 1/32× beech tea and 40 µl of each species from a stock culture of standard density. The plate was sealed and the change in optical density (OD) at 570 nm of the indicator gel measured after 6 h as recommended by the manufacturers [56]. The change in OD of blank wells (filled with 1 ml 1/32× beech tea) was used to account for the base level of CO2 in the vials. The rate of CO2 respiration per ml of culture medium was calculated using the formula provided in the MicroResp manual [56]. Statistical Analysis To calibrate OD600 in terms of cell density per ml of culture medium [57], we performed serial dilution and colony counts of stock cultures of isolates of each species from each treatment. We fitted a linear model with log (colony count)/ml as the response variable and species, treatment, and OD600 as explanatory variables, including interaction terms. The model simplified to retain species and OD600, but no interaction terms (i.e., different intercept for calibration line for each species, but same slopes, F 4,67 = 32.9, p<0.0001, r 2 = 0.64, Figure S8). The fitted lines were used to calibrate in units of log(number of cells) per ml. We used linear mixed effects models of repeated measures of cell density over time to compare growth of bacteria among treatments and species in the growth assays (Text S1). To report the direction and effect size of differences among treatments, we used the rate of change in density over the first 48 h as a simple measure of VMAX —that is, the maximum rate of growth from low densities (Figure S6). Analysis of variance (ANOVA) and Tukey's Honest Significant Difference tests were used to identify significant contrasts between particular treatments of interest. There was no evidence of different evolutionary trends in carrying capacity of isolates (i.e., using density at 96 h) as opposed to growth rate (Figure S1 versus Figure S7). To test for significant differences in NMR profiles between treatments, we used Monte Carlo simulation tests shuffling profiles randomly among species and treatments. The Euclidean distance between samples was recorded, and the mean distance between both evolved treatments in turn and ancestral isolates was used to measure the amount of evolution, and the mean distance between each species within a treatment was used to measure the amount of divergence in resource use among species. Observed values were compared to randomised values from 10,000 random permutations. Two-tailed tests were used. Supporting Information Figure S1 Maximum growth rates for each species and evolution treatment when grown in “used” and “unused” substrate. Boxplots of maximum growth rates, VMAX , in cell doublings per day across evolution treatments, species, and substrates. The dark line shows the median, the box limits show the inter-quartile range, and whiskers/points indicate extreme values. (TIF) Click here for additional data file. Figure S2 Amounts of compounds identified from distinct peaks in the NMR spectrum of unused beech tea. Bars show the size of the major peak for each distinct compound relative to the size of the standard, DSS; hence peak heights are dimensionless. The location of each peak on the spectrum is shown after each name (peak shift in parts per million). (TIF) Click here for additional data file. Figure S3 NMR peaks for each species and treatment. The difference in the size of NMR peaks between tea used by ancestral (dark grey), monoculture (mid grey), and polyculture (light grey) in turn and the size of peaks in unused beech tea. Positive values indicate production of a compound, and negative values indicate consumption of a compound. Peak sizes are expressed relative to the size of the standard, DSS, and hence are dimensionless. (TIF) Click here for additional data file. Figure S4 Contribution of each compound to variation between treatments. Loadings of the first four principal components of resource use and production of the four surviving species across ancestral, monoculture, and polyculture treatments. The input data were the difference between the size of the peak in medium used by the isolate and the size of the peak in the beech tea (i.e., the data in Figure S3). Bars indicate the correlation coefficient between variation in each compound and the relevant principal component. The percentage of total variation described by each principal component is shown above each plot; together they explain 90.1% of the total variation. (TIF) Click here for additional data file. Figure S5 Changes in substrate composition after use by a first species and then species B or D. The difference in the relative size of NMR peaks between tea used by a first species' ancestral (red), monoculture (green), and polyculture (blue) in turn and the relative size of peaks in unused beech tea; together with the change in the size of the peak after a second species grew on medium already used by the first species (then filter sterilised) for the same treatments (ancestral, pink; monoculture, light green; polyculture, light blue). The order of bars for each compound is first species ancestral, second species ancestral, first species monoculture, second species monoculture, first species polyculture, and second species polyculture. Positive values indicate production of a compound, and negative values indicate consumption of a compound relative to the starting medium. To improve clarity of the figure and focus on compounds of interest for cross-feeding, only compounds in which at least one isolate generated an increase in peak size of 0.5 are shown. Only species B and species D were used as the second species, chosen to represent two species showing different results in the growth assays. Evidence of evolved cross-feeding in polyculture is apparent when high blue peaks (generation of the compound by the first species) are associated with low purple peaks (use of the compound by the second species). For example, the species A polyculture isolate produces formate, which in turn is used up by both species B and D. (TIF) Click here for additional data file. Figure S6 Growth of replicates of each species in assays on unused beech tea across the three treatments. The y-axes are log(cell counts per ml), and x-axes are time since start in hours. Ancestral isolates of all four species grew linearly over the assay period on unused beech tea (ANOVA comparing a model with time as a factor versus a model with time as a continuous variable, likelihood ratio = 6.9, df = 13 and 21, p = 0.55). The monoculture isolates displayed significantly non-linear growth (ANOVA comparing models with time as a factor and as a continuous variable, L-ratio 39.3, p<0.0001). In species A, B, and C there was a reduction in growth rate between day 2 and 3 followed by recovery by day 4. In species D, there was a successive decline in growth rate. In each case, growth between day 0 and day 2 was faster than at any later period. Polyculture isolates grew linearly over the assay period (ANOVA comparing models with time as a factor and as a continuous variable, L-ratio 27.7, p<0.001). (TIF) Click here for additional data file. Figure S7 Boxplots of the density after 4 d (log10) across species and substrates. The dark line shows the median, the box limits show the inter-quartile range, and whiskers/points indicate extreme values. Key findings based on comparing Vmax remain the same when comparing amount of growth by day 4: species A grows well on unused tea in ancestral and monoculture treatments, but not when it has evolved in polyculture. Species B and C shift from having reduced growth on used tea in ancestral and monoculture isolates to having enhanced growth in polyculture treatments. Species D evolves to have stronger negative effects of used tea in monoculture than in ancestral isolates, but evolves even better growth on unused tea when it evolves in polycultures than in either ancestral or monoculture isolates. (TIF) Click here for additional data file. Figure S8 Scatter plot showing the linear relationship between OD600 and log colony counts. The model simplified to retain species and OD600, but no interaction terms (i.e., different intercept for calibration line for each species, but same slopes, F 4,67 = 32.9, p<0.0001, r 2 = 0.64). The fitted lines were used to calibrate in units of log(number of cells) per ml. (TIF) Click here for additional data file. Table S1 Molecular identification of bacterial isolates. (DOCX) Click here for additional data file. Table S2 Description and photographs of growth morphology of each species on agar plates. (DOCX) Click here for additional data file. Table S3 Densities, doubling rates, and effective population sizes of each species during the evolution experiments. (DOCX) Click here for additional data file. Table S4 Linear mixed effects model comparisons. (DOCX) Click here for additional data file. Text S1 Additional methods and references. (DOC) Click here for additional data file.
Phosphorus is a biolimiting nutrient that has an important role in regulating the burial of organic matter and the redox state of the ocean-atmosphere system. The ratio of phosphorus to iron in iron-oxide-rich sedimentary rocks can be used to track dissolved phosphate concentrations if the dissolved silica concentration of sea water is estimated. Here we present iron and phosphorus concentration ratios from distal hydrothermal sediments and iron formations through time to study the evolution of the marine phosphate reservoir. The data suggest that phosphate concentrations have been relatively constant over the Phanerozoic eon, the past 542 million years (Myr) of Earth's history. In contrast, phosphate concentrations seem to have been elevated in Precambrian oceans. Specifically, there is a peak in phosphorus-to-iron ratios in Neoproterozoic iron formations dating from ∼750 to ∼635 Myr ago, indicating unusually high dissolved phosphate concentrations in the aftermath of widespread, low-latitude 'snowball Earth' glaciations. An enhanced postglacial phosphate flux would have caused high rates of primary productivity and organic carbon burial and a transition to more oxidizing conditions in the ocean and atmosphere. The snowball Earth glaciations and Neoproterozoic oxidation are both suggested as triggers for the evolution and radiation of metazoans. We propose that these two factors are intimately linked; a glacially induced nutrient surplus could have led to an increase in atmospheric oxygen, paving the way for the rise of metazoan life.
Publication date
(Electronic, pub):
20
November
2018
Publication date Collection: 2018
Volume: 7
Electronic Location Identifier: e38278
Affiliations
[1
]deptDepartamento de Ecología Evolutiva, Instituto de Ecología Universidad Nacional Autónoma de México CoyoacánMexico
[2
]Jardín Botánico, Instituto de Biología Universidad Nacional Autónoma de México CoyoacánMexico
[3
]deptDepartment of Ecology, Evolution and Behavior University of Minnesota Saint PaulUnited States
[4
]deptLaboratorio Nacional de la Ciencias de la Sostenibilidad, Instituto de Ecología Universidad Nacional Autónoma de México CoyoacánMexico
[5
]deptLaboratorio de Biología Molecular y Ecología Microbiana, Departamento de Ingeniería
Genética Unidad Irapuato Centro de Investigación y Estudios Avanzados GuanajuatoMexico
Harvard Medical School United States
Max Planck Institute for Chemical Ecology Germany
Harvard Medical School United States
Author notes
[‡]
Departamento de Biología Celular, Facultad de Ciencias, Universidad Nacional Autónoma
de México, Mexico City, Mexico.
This article is distributed under the terms of the
Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author
and source are credited.
History
Date
received
: 11
May
2018
Date
accepted
: 12
October
2018
Funding
Funded by: FundRef http://dx.doi.org/10.13039/501100005201, WWF International;
Award ID: desunam7
Award Recipient
:
Valeria Souza
Funded by: FundRef http://dx.doi.org/10.13039/501100003141, Consejo Nacional de Ciencia y Tecnología;
Award ID: 238245
Award Recipient
:
Valeria Souza
Funded by: FundRef http://dx.doi.org/10.13039/501100003141, Consejo Nacional de Ciencia y Tecnología;
Award ID: 220536
Award Recipient
:
Gabriela Olmedo
Fundacion Carlos Slim along with WWF Mexico funded a 5 year study of the biodiversity
of Cuatro Cienegas in order to obtain the total inventarium of this extremely important
and endangered site. They did not dictate any rules on the research or hold any rights
on its publication.
Categories
Subject:
Research Article
Subject:
Ecology
Custom metadata
Author impact statement At Cuatro Ciénegas, the ancestral niche conditions of the ancient shores of Laurentia
persisted along with its microbial communities that have been coevolving for a very
long time.
scite shows how a scientific paper has been cited by providing the context of the citation, a classification describing whether it supports, mentions, or contrasts the cited claim, and a label indicating in which section the citation was made.
Smart Citations