Elsevier

Bioresource Technology  生物资源技术

Volume 238, August 2017, Pages 389-398
2017 年 8 月，第 238 卷，第 389-398 页
Bioresource Technology

Full-scale photobioreactor for biotreatment of olive washing water: Structure and diversity of the microalgae-bacteria consortium
用于橄榄清洗水生物处理的全规模光生物反应器：微藻-细菌共生体的结构与多样性

https://doi.org/10.1016/j.biortech.2017.04.048Get rights and content  获取权限和内容
Full text access  全文访问

Highlights  亮点

  • •
    A full-scale photobioreactor (PBR) was applied for the removal of OWW pollutants.
    一个全规模光生物反应器（PBR）被用于去除 OWW 污染物。
  • •
    The microalgae-bacteria consortium in the PBR was studied by Illumina-sequencing.
    通过 Illumina 测序研究了光生物反应器中的微藻-细菌联合体。
  • •
    The selection of the DNA extraction method was crucial for community identification.
    DNA 提取方法的选择对于群落鉴定至关重要。
  • •
    OWW bioremediation relied on the Chlorophyta/Cyanobacteria/Proteobacteria interplay.
    OWW 的生物修复依赖于绿藻门/蓝细菌门/变形菌门的相互作用。

Abstract  摘要

The performance of a full-scale photobioreactor (PBR) for the treatment of olive washing water (OWW) was evaluated under different HRTs (5–2 days). The system was able to treat up to 3926 L OWW day−1, and consisted of an activated-carbon pretreatment column and a tubular PBR unit (80 tubes, 98.17 L volume, 2-m height, 0.25 m diameter). PBR was an effective and environmentally friendly method for the removal of phenols, COD, BOD5, turbidity and color from OWW (average efficiencies 94.84 ± 0.55%, 85.86 ± 1.24%, 99.12 ± 0.17%, 95.86 ± 0.98% and 87.24 ± 0.91%, respectively). The diversity of total bacteria and microalgae in the PBR was analyzed using Illumina-sequencing, evaluating the efficiency of two DNA extraction methods. A stable microalgae-bacteria consortium was developed throughout the whole experimentation period, regardless of changes in HRT, temperature or solar radiation. MDS analyses revealed that the interplay between green algae (Sphaeropleales), cyanobacteria (Hapalosiphon) and Proteobacteria (Rhodopseudomonas, Azotobacter) played important roles in OWW bioremediation.
对全规模光生物反应器（PBR）在处理橄榄清洗水（OWW）中的性能进行了评估，研究在不同水力停留时间（HRT，5–2 天）下的表现。该系统每天最多可处理 3926 升 OWW，由活性炭预处理柱和管状 PBR 单元（80 根管，体积 98.17 升，高 2 米，直径 0.25 米）组成。PBR 是一种高效且环保的方法，可去除 OWW 中的酚类、化学需氧量（COD）、生化需氧量（BOD）、浊度和颜色，平均去除效率分别为 94.84 ± 0.55%、85.86 ± 1.24%、99.12 ± 0.17%、95.86 ± 0.98% 和 87.24 ± 0.91%。通过 Illumina 测序技术对 PBR 中的细菌和微藻多样性进行了分析，并评估了两种 DNA 提取方法的效率。在整个实验期间，无论 HRT、温度或太阳辐射的变化如何，均成功构建了一个稳定的微藻-细菌共生体系。多维尺度分析（MDS）显示，绿藻（球藻目）、蓝藻（Hapalosiphon）和变形菌（红假单胞菌、固氮菌）之间的相互作用在 OWW 的生物修复中发挥了重要作用。

Keywords  关键词

Full-scale photobioreactor (PBR)
Olive washing water (OWW)
Microalgae-bacteria consortium
Illumina-sequencing
Non-metric multidimensional scaling (MDS)

全规模光生物反应器 (PBR) 橄榄清洗水 (OWW) 微藻-细菌联合体 Illumina 测序 非度量多维尺度分析 (MDS)

1. Introduction  1. 引言

Photobioreactors (PBRs) are culture systems which incorporate a natural or artificial light source to allow the growth of phototrophic microorganisms (Chang et al., 2017). The idea to use microalgae-bacteria consortia for wastewater treatment is not new and the first studies were reported in the 1950s decade (Oswald and Gotaas, 1957). This technology is an attractive choice for bioremediation due to its many advantages. Microalgae and cyanobacteria generate the oxygen necessary for aerobic heterotrophic bacteria to oxidize organic matter, while at the same time they capture CO2 released from bacterial respiration. This alternative to support the growth of aerobic bacteria is cheaper than mechanical aeration, because the oxygen is supplied through photosynthesis, and additionally it contributes to CO2 emission mitigation (Guieysse et al., 2002). In recent years, several studies have demonstrated that microalgae-bacteria associations are able to successfully remove from wastewaters a variety of pollutants, such as phenol, black oil, acetonitrile, naphthalene or benzopyrene (Essam et al., 2014, Mahdavi et al., 2015, Ramanan et al., 2016, Ryu et al., 2015).
光生物反应器（PBRs）是利用自然或人工光源来促进光养微生物生长的培养系统（Chang 等，2017）。利用微藻-细菌联合体进行废水处理的想法并不新鲜，最早的研究可追溯到 20 世纪 50 年代（Oswald 和 Gotaas，1957）。由于其诸多优势，这项技术是生物修复的一个有吸引力的选择。微藻和蓝藻产生了需氧异养细菌氧化有机物所需的氧气，同时捕获细菌呼吸释放的 CO₂。与机械曝气相比，这种支持需氧细菌生长的替代方法成本更低，因为氧气通过光合作用提供，此外还可有助于减缓 CO₂排放（Guieysse 等，2002）。近年来，多项研究表明，微藻-细菌联合体能够成功去除废水中的多种污染物，例如酚类、重油、乙腈、萘或苯并芘（Essam 等，2014；Mahdavi 等，2015；Ramanan 等，2016；Ryu 等，2015）。
Olive washing water (OWW) is generated as a by-product of the olive oil industry and is a complex effluent characterized by an intense dark-green color and high turbidity, and containing organohalogenated contaminants, long-chain fatty acids, and recalcitrant phenol compounds such as lignin and humic acids, well known for their phytotoxic effects (Maza-Márquez et al., 2014, Maza-Márquez et al., 2016a). Since the olive oil industry has the major food sector production in Mediterranean countries, significant amounts of OWW are generated (approximately 1 m3 per processed ton, Ochando-Pulido and Martinez-Ferez, 2015). Moreover, the olive oil industry is mainly composed of small and dispersed factories, which require simple and cost-efficient technologies to treat their effluents. In this respect, the combination of microalgae with bacteria is an effective, economical and straightforward solution for the olive oil sector. In a recently published study, Maza-Márquez et al. (2016a) demonstrated that a microalgae-bacteria consortium was able to grow fast in laboratory-scale PBRs operated with real OWW, achieving excellent removal rates of color, turbidity, phenols, chemical oxygen demand (COD) and biological oxygen demand at 5 days (BOD5), and generating effluents complying with the specific guidelines and EU legislation for its discharge to the environment (Maza-Márquez et al., 2016a). However, the widespread exploitation of this technology has some limitations, such as the high cost associated to the artificial illumination of the PBR, or the difficulties to implement the technology in small local industries which operate under varying conditions of solar radiation, temperature or pollutants’ concentration.
橄榄清洗水（OWW）是橄榄油工业产生的副产品，这是一种复杂的废水，具有深绿色的颜色、高浑浊度，并含有有机卤化物污染物、长链脂肪酸以及难降解的酚类化合物，例如木质素和腐殖酸，这些物质因其植物毒性效应而广为人知（Maza-Márquez 等，2014；Maza-Márquez 等，2016a）。由于橄榄油工业是地中海国家主要的食品行业之一，产生了大量的 OWW（每加工一吨橄榄约产生 1 立方米，Ochando-Pulido 和 Martinez-Ferez，2015）。此外，橄榄油行业主要由小型且分散的工厂组成，这些工厂需要简单且经济高效的技术来处理其废水。在这方面，微藻与细菌的结合对橄榄油行业来说是一种有效、经济且简单的解决方案。在最近发表的一项研究中，Maza-Márquez 等（2016a）证明，在实验室规模的光生物反应器（PBR）中，微藻-细菌联合体能够在真实的 OWW 中快速生长，在 5 天内实现对颜色、浑浊度、酚类、化学需氧量（COD）和 5 日生化需氧量（BOD）的卓越去除率，并产生符合环境排放相关法规和欧盟法律要求的废水（Maza-Márquez 等，2016a）。然而，该技术的广泛应用仍面临一些限制，例如与 PBR 人工照明相关的高成本，或在小型地方性工厂中实施技术的难度，这些工厂通常在太阳辐射、温度或污染物浓度等条件下存在较大变化。
In this study, a full-scale PBR was designed to meet the demands of a local olive oil industry in Southern Spain. The system was setup outdoors exposed to sunlight, avoiding electricity cost, and was long-term operated under changing conditions of solar radiation, temperature and OWW composition. The ability of the PBR to remove OWW pollutants was assessed during 7 months testing different hydraulic retention times (HRTs). Illumina-sequencing was used to examine the microbial communities’ structure and population dynamics in the PBR and identify the major key players of the microalgae-bacteria consortium. Multivariate analysis (non-metric multidimensional scaling, MDS, and BIO-ENV) were performed to assess changes in the microbial community structure, and to test the dynamic relationships between the relative abundances of bacteria and microalgae populations and the changes of the environmental/operational variables influencing the PBR.
在本研究中，设计了一套全规模的光生物反应器（PBR），以满足西班牙南部当地橄榄油行业的需求。该系统被安装在户外，利用阳光照射，避免了电力成本，并在太阳辐射、温度和橄榄油废水（OWW）成分变化的条件下进行了长期运行。通过为期 7 个月的试验，在不同的水力停留时间（HRTs）下评估了 PBR 去除 OWW 污染物的能力。使用 Illumina 测序技术分析了 PBR 中微生物群落的结构和种群动态，并确定了微藻-细菌共生体中的主要关键成员。通过多变量分析（非度量多维缩放分析 MDS 和 BIO-ENV 分析），评估了微生物群落结构的变化，并测试了细菌和微藻种群的相对丰度与影响 PBR 的环境/操作变量变化之间的动态关系。

2. Materials and methods  2. 材料与方法

2.1. Full-scale PBR experimental plant
2.1. 全规模 PBR 实验装置

A full-scale PBR experimental plant was installed and operated under real conditions at the olive oil factory Nuestra Señora de los Desamparados, located in Puente Genil, Córdoba (Southern Spain). A schematic diagram of the plant is displayed in Fig. 1 and a picture is shown in Fig. S1. It included an activated carbon pretreatment column and a tubular PBR unit. Their specific characteristics are described in detail in the Supplementary material (Table S1, Fig. S2).
在西班牙南部科尔多瓦省普恩特亨尼尔的橄榄油工厂 Nuestra Señora de los Desamparados，安装并运行了一座全规模 PBR 实验工厂以模拟真实条件。工厂的示意图如图 1 所示，实物照片见图 S1。该工厂包括一个活性炭预处理柱和一个管式 PBR 装置。它们的具体特性在补充材料中进行了详细描述（表 S1，图 S2）。
  1. Download: Download high-res image (165KB)
    下载：下载高分辨率图片（165KB）
  2. Download: Download full-size image
    下载：下载原始尺寸图片

Fig. 1. Scheme of the full-scale photobioreactor (PBR) used in the study.
图 1. 本研究中使用的全尺寸光生物反应器（PBR）示意图。

Briefly, the activated carbon column (FCC-125-CR) was built in a compact configuration (1670 mm height, 406 mm diameter, 0.129 m2 total surface, Figs. 1, S2B) and the residence time of the influent was 20 min. The influent flow rate and pressure were controlled by a rotameter and a manometer. The column was also provided with an automatic cleaning system, which used clean water stored in two reservoirs of 45 L volume each. An additional security filter (308 cm2 total filtration surface) at the end of the column was installed to protect the system’s pumps (Fig. 1).
简要来说，活性炭柱（FCC-125-CR）被设计为紧凑结构（高度 1670 mm，直径 406 mm，总表面积 0.129 m²（图 1，S2B）），进水的停留时间为 20 分钟。进水流量和压力通过转子流量计和压力计进行控制。该柱还配备了一个自动清洗系统，该系统使用储存在两个 45 升水箱中的清水。此外，在柱体末端安装了一个附加的安全过滤器（总过滤表面积 308 cm²），以保护系统的水泵（图 1）。
The tubular PBR unit (Figs. 1, S1, S3) was able to treat 1570–3926 L day−1 of OWW, depending on the selected hydraulic retention time (HRT) (Table S2A). The configuration of the PBR was designed and optimized in previous work (Maza-Márquez et al., 2016a) to reduce the risk of microalgae inhibition by insufficient exposure to light and provide high photosynthetic rates. It was composed of 8 modules, each of which comprised ten polyvinyl chloride (PVC) tubes (2 m height, 0.25 m diameter, 98.17 L volume). The modules were connected using flexible pipes (Fig. S3B). In each module, the water entered upflow, circulating at the same speed through the different tubes. In addition to the 8 modules, the PBR unit included a feeding tank, a recirculation tank, and 2 pumps (feeding and recirculation pumps) (Fig. S2A). The pumps were monitored through a control box, which also included a security alarm and an emergency stop control.
管式 PBR 装置（图 1、S1、S3）能够根据所选择的水力停留时间（HRT）（表 S2A）处理 1570–3926 L/天的油橄榄废水（OWW）。PBR 的配置在之前的研究中已设计并优化（Maza-Márquez 等，2016a），以降低因光照不足导致微藻抑制的风险，同时提供高光合作用效率。该装置由 8 个模块组成，每个模块包含 10 根聚氯乙烯（PVC）管（高度 2 m，直径 0.25 m，体积 98.17 L）。模块之间通过柔性管连接（图 S3B）。在每个模块中，水以向上流动的方式进入，并以相同的速度在不同管道中循环。除了 8 个模块外，PBR 装置还包括一个供水箱、一个回流箱和两台泵（供水泵和回流泵）（图 S2A）。泵通过一个控制箱进行监控，控制箱还配备了安全警报及紧急停止控制功能。

2.2. Experimental set-up  **2.2 实验装置**

The system worked in two steps: 1-pre-treatment with activated carbon, and 2- PBR biological treatment. OWW was stored in a pool (ca. 45 m3) (Fig. S4), located higher than the pre-treatment feeding tank (T1, 5000 L), in order to avoid the energetic cost of pumping the influent. To avoid the damage of pumps, the OWW was filtered through a 1-mm brush screen before being pumped through the activated carbon column (Fig. 1). Pump 1 speed was adjusted according to the selected HRT (Table S2A). After the pre-treatment step, the water went through a security filter column (Fig. 1) to eliminate the particles with a size >130 μm before entering the PBR feeding tank (T2, 1000 L). The water was then routed to the PBR unit by pump 2. Once the water passed through all the PBR modules, it was sent to the recirculation tank (T3, 1000 L), where one part was recycled to the PBR feeding tank (pump 3110 L min−1).
该系统分两步运行：1. 使用活性炭进行预处理；2. PBR（填充床反应器）生物处理。OWW（橄榄废水）储存在一个约 45 立方米的池中（图 S4），该池位于预处理供料箱（T1，5000 升）上方，以避免进水泵送的能源成本。为了防止泵损坏，OWW 在通过活性炭柱泵送之前，先通过 1 毫米的刷式滤网进行过滤（图 1）。根据选定的 HRT（表 S2A），调整泵 1 的速度。经过预处理步骤后，水通过一个安全过滤柱（图 1），以去除粒径大于 130 微米的颗粒，然后进入 PBR 供料箱（T2，1000 升）。随后，水通过泵 2 输送到 PBR 单元。一旦水通过所有 PBR 模块后，就被送至循环罐（T3，1000 升），其中一部分被回收并通过泵输送回 PBR 供料箱（泵流量为 3110 升/分钟）。
The PBR unit was initially filled with diluted OWW (60% in water), in order to propitiate biofilm formation based on microalgae. The PBR was operated in recirculation for a period of 7 days, until a biofilm was visible on the surface of the plastic PBR tubes. From that moment, the system was continuously fed with real OWW and operated in continuous flow. Four experiments were set up operating the PBR at different HRTs (5, 4, 3 and 2 days), in order to analyze the influence of this parameter on the efficiency of removal of OWW pollutants. The operational conditions in each experimental phase are summarized in Table S2A.
最初，PBR 装置用稀释的 OWW（水中占 60%）填充，以促进基于微藻的生物膜形成。PBR 在循环模式下运行了 7 天，直到塑料 PBR 管表面可见生物膜。从那时起，系统开始连续供给真实的 OWW，并以连续流模式运行。设置了四个实验，在不同的水力停留时间（HRT）（5 天、4 天、3 天和 2 天）下运行 PBR，以分析该参数对 OWW 污染物去除效率的影响。每个实验阶段的运行条件汇总于表 S2A 中。

2.3. Wastewater characteristics
2.3. 废水特性

The full-scale PBR was fed the OWW generated by the olive oil factory nearby, which showed the average characteristics summarized in Table S2B.
全规模的 PBR 使用了附近橄榄油工厂产生的 OWW，其平均特性总结在表 S2B 中。

2.4. Sample collection and physico-chemical analysis
2.4. 样品采集和理化分析

Samples (1 L) from raw OWW and PBR effluent were weekly collected in sterile bottles, carried at 4 °C from the experimental plant to the laboratory and immediately processed. The parameters monitored included: pH, temperature, dissolved oxygen concentration (DO), BOD5, COD, color, turbidity, total phenols, total N and total P. Analyses were conducted in accordance to the Standard Methods for Examinations of Water and Wastewater (APHA, 2005). The concentration of total volatile suspended solids (VSS) was measured to monitor the microbial growth in the PBR, according to the Standard Method 2540. Total phenols were measured spectrophotometrically (760 nm), according to the Folin-Ciocalteu procedure (Maza-Márquez et al., 2016a). Temperature and DO were checked in situ using a multiple P4 Oxical-SL Universal Meter (WTW, Germany), while pH was measured with a Crison micro pH 2002 instrument (Crison Instruments, Spain). Solar radiation data (kWh/m2) were retrieved from the public database of Agencia Andaluza de la Energía (Junta de Andalucía, Spain, https://www.agenciaandaluzadelaenergia.es/Radiacion).
每周从原始 OWW 和 PBR 出水中取样（1 升），使用无菌瓶收集样品，在 4°C 条件下从实验装置运送至实验室并立即处理。监测的参数包括：pH 值、温度、溶解氧浓度（DO）、生化需氧量（BOD）、化学需氧量（COD）、颜色、浊度、总酚、总氮和总磷。分析按照《水和废水检验标准方法》（APHA, 2005）进行。按标准方法 2540 测定挥发性悬浮固体总量（VSS）浓度，以监测 PBR 中的微生物生长。总酚含量根据 Folin-Ciocalteu 方法（Maza-Márquez 等，2016a）采用分光光度法（760 nm）测定。温度和溶解氧使用多功能 P4 Oxical-SL Universal Meter（WTW 公司，德国）进行现场检测，pH 值使用 Crison micro pH 2002 仪器（Crison Instruments 公司，西班牙）测量。太阳辐射数据（kWh/m²）从安达卢西亚能源署（Junta de Andalucía, Spain, https://www.agenciaandaluzadelaenergia.es/Radiacion）的公共数据库中获取。

2.5. DNA extraction  2.5 DNA 提取

In order to investigate the changes of the diversity and population dynamics of total Bacteria and microalgae in the PBR, biological samples were collected after 35 days of operation in each experimental phase, when the PBR was working under stable conditions. A total of six repeated biological samples were retrieved from the system at each sampling time, which were processed for DNA extraction and subsequent massive parallel sequencing using Illumina.
为了研究光生物反应器 (PBR) 中细菌和微藻多样性及种群动态的变化，在每个实验阶段运行稳定 35 天后采集生物样本。在每次采样时，从系统中总共取了 6 份重复的生物样本，这些样本用于 DNA 提取，并随后通过 Illumina 进行大规模平行测序。
The extraction of total DNA represents a crucial step in molecular biology for microbial community analyses (Tanase et al., 2015). Since PBR biofilms are commonly composed of both eukaryotic and prokaryotic microorganisms with cell walls of varying chemical composition and structure, two different DNA extraction kits were used and compared: the FastDNA-SPIN kit (herein algae DNA kit), and the FastDNA-SPIN kit for Soil (herein soil DNA kit) (MP-Bio, Santa Ana, CA, USA). According to the manufacturer, the FastDNA-SPIN kit is recommended for the isolation of genomic DNA from plant and animal tissue, bacteria, yeast, algae and fungi, while the FastDNA-SPIN kit for Soil was designed to efficiently isolate bacterial, fungal, plant and animal genomic DNA from soil and other environmental samples (http://www.mpbio.com). The FastPrep24 instrument (MP-Bio, Santa Ana, CA, USA) was used for the extraction and purification of total genomic DNA following previous protocols (Maza-Márquez et al., 2016a), using 3 repeated biological samples of each operational period for each kit. The quality and concentration of the extracted DNA samples were measured spectrophotometrically (260 nm), using NanoDrop ND-1000 (Thermo Scientific, Waltham, MA USA). The DNA extracted from the different repeated samples was pooled together in equimolar proportions for Illumina sequencing.
DNA 的提取是微生物群落分子生物学分析中的关键步骤（Tanase 等，2015）。由于 PBR 生物膜通常由具有不同化学成分和结构的细胞壁的真核和原核微生物组成，研究中使用并比较了两种不同的 DNA 提取试剂盒：FastDNA-SPIN 试剂盒（以下简称“藻类 DNA 试剂盒”）和 FastDNA-SPIN Soil 试剂盒（以下简称“土壤 DNA 试剂盒”）（MP-Bio，美国加利福尼亚州圣安娜）。根据制造商的说明，FastDNA-SPIN 试剂盒推荐用于从植物和动物组织、细菌、酵母、藻类和真菌中提取基因组 DNA，而 FastDNA-SPIN Soil 试剂盒则专为从土壤和其他环境样本中高效提取细菌、真菌、植物和动物的基因组 DNA 而设计（http://www.mpbio.com）。使用 FastPrep24 设备（MP-Bio，美国加利福尼亚州圣安娜）按照之前的研究方法（Maza-Márquez 等，2016a）提取和纯化总基因组 DNA，每个操作阶段各使用 3 份重复的生物样本。通过 NanoDrop ND-1000 分光光度计（Thermo Scientific，美国马萨诸塞州沃尔瑟姆）在 260 nm 波长下测量提取的 DNA 样本的质量和浓度。从不同重复样本中提取的 DNA 以等摩尔比例混合后，用于 Illumina 测序。

2.6. Analysis of the diversity of total bacteria and microalgae by Illumina-sequencing
### 2.6 通过 Illumina 测序分析总细菌和微藻的多样性

Genomic DNA samples (430–480 ng/μL) were sent to RTL Genomics (Lubbock, Texas, USA, http://www.researchandtesting.com) for massive parallel sequencing. A tag-encoded paired-end procedure was followed, using the Illumina MiSeq Reagent Kit v3 and the Illumina MiSeq apparatus (Illumina, Hayward, CA, USA). The molecular markers selected for the survey of microbial diversity in the PBR samples were the16S rRNA gene for total Bacteria and the 18S rRNA gene for Chlorophyta. Chlorophyta (green algae) were selected as target group because microscopic analyses and identification of isolates by traditional Sanger sequencing showed that they were the dominant microalgae in a previous study (Maza-Márquez et al., 2016a). Partial 16S-rRNA gene amplicons (ca. 500 bp) were generated by amplification of the hypervariable V1-V3 regions of the 16S rRNA gene of Bacteria, using primers 28F (5′-GAGTTTGATCNTGGCTCAG-3′) and 519R (5′-GTNTTACNGCGGCKGCTG-3′) (Fan et al., 2012). The PCR conditions were: 94 °C for 3 min, followed by 32 cycles of 94 °C for 30 s; 60 °C for 40 s and 72 °C for 1 min; and a final elongation step at 72 °C for 5 min. For the amplification of partial 18S rRNA-genes (473 bp) of Chlorophyta, primers Chloro-F (5′-TGG CCT ATC TTG TTG GTC TGT-3′) and Chloro R (5′-GAA TCA ACC TGA CAA GGC AAC-3′) were selected (Ryu et al., 2015). The cycling conditions were as follows: 94 °C for 3 min, followed by 35 cycles of 94 °C for 15 s, 59 °C for 45 s and 72 °C for 1 min, and a final extension at 72 °C for 8 min.
将基因组 DNA 样品（430–480 ng/μL）送至 **RTL Genomics**（美国德克萨斯州拉伯克市，http://www.researchandtesting.com）进行大规模平行测序。实验采用了标签编码的双端测序方法，使用 Illumina MiSeq 试剂盒 v3 和 Illumina MiSeq 测序仪（Illumina 公司，加利福尼亚州海沃德）。用于分析 PBR 样品中微生物多样性的分子标记是针对总细菌的 16S rRNA 基因和针对绿藻（Chlorophyta）的 18S rRNA 基因。选择绿藻作为目标群体，是因为在之前的研究中（Maza-Márquez 等人，2016a），通过显微镜分析和传统 Sanger 测序鉴定发现绿藻是主要的微藻。 通过扩增细菌 16S rRNA 基因的高变区 V1-V3 区域，生成了部分 16S rRNA 基因扩增子（约 500 bp）。所使用的引物是 28F（5′-GAGTTTGATCNTGGCTCAG-3′）和 519R（5′-GTNTTACNGCGGCKGCTG-3′）（Fan 等人，2012）。PCR 反应条件如下：94°C 预变性 3 分钟，随后进行 32 个循环，每个循环包括 94°C 变性 30 秒、60°C 退火 40 秒和 72°C 延伸 1 分钟，最后在 72°C 延伸 5 分钟。 为了扩增绿藻 18S rRNA 基因的部分片段（473 bp），选择了引物 Chloro-F（5′-TGG CCT ATC TTG TTG GTC TGT-3′）和 Chloro-R（5′-GAA TCA ACC TGA CAA GGC AAC-3′）（Ryu 等人，2015）。PCR 循环条件如下：94°C 预变性 3 分钟，随后进行 35 个循环，每个循环包括 94°C 变性 15 秒、59°C 退火 45 秒和 72°C 延伸 1 分钟，最后在 72°C 延伸 8 分钟。

2.7. Data processing  ### 2.7 数据处理

Raw data obtained through Illumina-sequencing were demultiplexed and quality-filtered using the mothur v1.34.4 software (Schloss et al., 2009), using a modified version of the protocol described by Rodríguez-Sánchez et al. (2016). Sequences passed several quality screening controls: reads with more than one ambiguous base, homopolymers with a size >8 or significant alignment errors, and lower quality reads were removed. Chimeras were removed using UCHIME v4.2.40 implemented in mothur v1.34.4 (Edgar et al., 2011). The number of reads after this step for all samples was in the range of 12,165 to 29,054. For a proper ecological analysis, samples were normalized by random selection of a total number of 12,165 reads for each of the samples using the subsampling function implemented in Mothur. For each sample, a distance Phylip matrix was generated using the rarefied 12,165 high-quality reads, and sequences were clustered to define phylotypes (PHYs) at 97% similarity level (Maza-Márquez et al., 2016b). Representative sequences from all PHYs were taxonomically assigned using the Classifier tool from the SILVA database.
通过 Illumina 测序获得的原始数据使用 mothur v1.34.4 软件（Schloss 等，2009）进行了样本拆分和质量过滤，采用了 Rodríguez-Sánchez 等（2016）描述的协议的修改版本。序列通过了多个质量筛选控制：含有超过一个不确定碱基的读取、同聚物长度大于 8 或具有显著比对错误的读取，以及质量较低的读取均被移除。嵌合体使用 UCHIME v4.2.40（Edgar 等，2011）工具在 mothur v1.34.4 中删除。此步骤后，各样本的读取数范围为 12,165 至 29,054。为进行适当的生态学分析，使用 Mothur 中的随机抽样功能对样本进行标准化处理，每个样本随机抽取总计 12,165 条读取。对于每个样本，使用稀释后的 12,165 条高质量读取生成了一个 Phylip 距离矩阵，并将序列聚类以 97%的相似性水平定义为类群（PHYs）（Maza-Márquez 等，2016b）。所有 PHYs 的代表序列通过 SILVA 数据库的 Classifier 工具进行了分类赋值。
The calculation of the Alpha- (Chao-1, H′, Shannon index, Simpson’s diversity) and Beta- (Bray-Curtis) diversity indices and the construction of rarefaction curves for each sample were performed using the PAST software (version 3.14) (Hammer et al., 2001) and the Primer software (PRIMER-E v. 6.0, Plymouth, UK). Good's coverage indices were also calculated as previously described (Maza-Márquez et al., 2016b). Heatmaps representing the relative abundances of total Bacteria and Chlorophyta in the samples at several taxonomic levels were generated using the gplots and RColor Brewer packages for R v.3.2.0 (http://www.r-project.org)
使用 PAST 软件（版本 3.14）（Hammer 等，2001）和 Primer 软件（PRIMER-E v6.0，Plymouth，英国）计算了 Alpha 多样性指数（Chao-1、H′、Shannon 指数、Simpson 多样性）和 Beta 多样性指数（Bray-Curtis），并为每个样本构建了稀释曲线。同时根据 Maza-Márquez 等（2016b）的描述计算了 Good's 覆盖指数。样本中总细菌和绿藻（Chlorophyta）在多个分类水平上的相对丰度热图使用 R v.3.2.0（http://www.r-project.org）的 gplots 和 RColor Brewer 包生成。

2.8. Statistical analysis
2.8. 统计分析

IBM SPSS Statistics v. 19 (SPSS Inc., IBM, USA) was used to test for the normality of data using the Shapiro-Wilk’s test. As most of the data sets fitted non-normal distributions, the Kruskal-Wallis test was chosen for comparisons among groups of samples, in search of significant differences. A 95% level of significance (p < 0.05) was selected. The Primer software (PRIMER-E v. 6.0, Plymouth, UK) was used to analyze the sets of biotic data derived from the Illumina-sequencing platform. The biotic data sets were standardized by total and sample-resemblance matrices were generated using the Bray Curtis coefficient of similarity. Based on the similarity matrix, ANOSIM analysis and non-metric multidimensional scaling (MDS) were used to calculate significant differences between groups of samples and construct 2D-ordination plots, following methods described previously (Maza-Márquez et al., 2016b). A stress level of the MDS plots <0.2 indicates a good fit of the 2D-representation, (Clarke and Warwick, 2001). The contribution of single PHYs to the dis(similarity) between groups of samples was evaluated using SIMPER (similarity percentages analysis) (Clarke and Warwick, 2001).
使用 IBM SPSS Statistics v.19 (SPSS Inc., IBM, USA) 通过 Shapiro-Wilk 正态性检验对数据的正态性进行测试。由于大多数数据集呈现非正态分布，因此选择 Kruskal-Wallis 检验比较样本组之间的差异，显著性水平设定为 95%（p < 0.05）。使用 Primer 软件 (PRIMER-E v.6.0, Plymouth, UK) 分析来自 Illumina 测序平台的生物数据集。这些生物数据集经过标准化，并利用 Bray-Curtis 相似性系数生成样本相似矩阵。基于相似性矩阵，采用 ANOSIM 分析和非度量多维尺度分析 (MDS) 计算样本组间的显著性差异，并构建二维排序图，方法参照此前研究（Maza-Márquez 等, 2016b）。MDS 图的应力水平 <0.2 表明二维表示拟合良好（Clarke 和 Warwick, 2001）。使用 SIMPER（相似性百分比分析）评估单个 PHYs 对样本组之间相似性或差异性的贡献（Clarke 和 Warwick, 2001）。
All the environmental/operational variables (except pH) were transformed (log(x + 1)) and normalized, and were represented by vectors overlaid over the MDS plots to show the potential monotonic correlations among them. The vectors were generated by a multiple partial correlation algorithm, which takes into account the correlation of all the other variables, based on the Spearman's rank coefficients. Pearson’s product moment correlation coefficients (r) showing the strength of the linear associations between the biotic and environmental/operational variables were also calculated. Finally, a BIO-ENV analysis was performed (Maza-Márquez et al., 2016b), in search of the environmental/operational variables mostly contributing to explain the changes of the relative abundance of PHYs occurring throughout the sample set. The statistical significance of vectors and BEST values was tested by a global permutation test (using the default 499 permutations).
所有环境/操作变量（除了 pH 值）经过转换（log(x + 1)）并标准化后，以向量形式叠加在 MDS 图上，显示它们之间可能存在的单调相关性。这些向量是通过多重偏相关算法生成的，该算法基于 Spearman 等级相关系数，考虑了所有其他变量之间的相关性。此外，还计算了 Pearson 积矩相关系数（r），以显示生物变量与环境/操作变量之间线性关联的强度。最后，进行了 BIO-ENV 分析（Maza-Márquez 等，2016b），以寻找对解释样本集中 PHYs 相对丰度变化贡献最大的环境/操作变量。向量和 BEST 值的统计显著性通过全局置换检验进行测试（使用默认的 499 次置换）。

3. Results and discussion
3. 结果与讨论

3.1. Biofilm formation and stabilization in the PBR
3.1. 生物膜的形成与在 PBR 中的稳定化

The experimental plant was operated for a period of 7 months (January to July of 2015). Prior to evaluate the influence of HRT, a start-up period was maintained for 40 days, during which the accumulation of microbial biomass in the PBR was monitored. Strong and stable biofilms were developed after three days (Fig. S5A). Since OWW was stored in pools at the same facility than the PBR, indigenous microorganisms were already adapted to analogous environmental stresses and selected for their resistance to OWW pollutants. In this sense, the fast formation of stable biofilms by indigenous bacteria and microalgae isolated from OWW was observed previously in a lab-scale PBR (Maza-Márquez et al., 2016a, Maza-Márquez et al., 2016b). Since the full-scale PBR was receiving natural sunlight, the development of an intensely green-colored biofilm was faster than previously observed under laboratory conditions using artificial light (Maza-Márquez et al., 2016a). Fig. S6 shows the growth of the microbial consortium biomass, expressed as VSS, and the food-to-microorganisms ratio (F/M ratio, kg of BOD5 kg of VSS−1 day−1). After the first 7 days, the concentration of VSS in the PBR reached 8.4 g L (Fig. S6), and widespread biofilm matrices were observable in the surface of the tubes (Fig. S5A). In contrast to previous experiments done in a laboratory-scale PBR (Maza-Márquez et al., 2016a), no lag phase was observed in the full-scale plant; on the contrary, an exponential period of biofilm growth was detected. As expected, the VSS concentration further increased from 8.4 to 10.2 g/L when the PBR started being fed with real OWW, and was kept stable for the rest of the start-up period (Fig. S6). From day 15th to 180th, VSS ranged between 8.7 and 10.2 g/L, and the F/M ratio was kept between 0.08 and 0.12. The extensive adherence and robustness of the biofilms were maintained during the whole experimentation period (Fig. S5B). These results were in agreement with previous studies (Maza-Márquez et al., 2016a).
实验装置运行了 7 个月（2015 年 1 月至 7 月）。在评估水力停留时间（HRT）影响之前，进行了为期 40 天的启动阶段，在此期间监测了 PBR 中微生物生物量的积累。强大且稳定的生物膜在 3 天后形成（图 S5A）。由于油橄榄废水（OWW）与 PBR 位于同一设施内的池中储存，本地微生物已适应类似的环境压力，并被选择为对 OWW 污染物具有抗性的菌种。在这一背景下，此前在实验室规模的 PBR 中观察到了由 OWW 中分离出的本地细菌和微藻快速形成稳定生物膜（Maza-Márquez 等，2016a，2016b）。由于全规模 PBR 接收到自然阳光，生物膜的绿色着色发展速度比使用人工光的实验室条件下更快（Maza-Márquez 等，2016a）。图 S6 显示了微生物群落生物量的增长情况（以 VSS 表示）以及食物与微生物比（F/M 比率，以 BOD 5 、VSS −1 、天数 −1 为单位）。在前 7 天内，PBR 中的 VSS 浓度达到 8.4 g/L（图 S6），并且在管道表面可以观察到广泛分布的生物膜结构（图 S5A）。与此前在实验室规模 PBR 中进行的实验相比（Maza-Márquez 等，2016a），全规模装置中未观察到滞后期；相反，检测到生物膜的指数增长阶段。如预期，当 PBR 开始以真实 OWW 为进料时，VSS 浓度从 8.4 g/L 进一步增加到 10.2 g/L，并在启动阶段的其余时间保持稳定（图 S6）。从第 15 天到第 180 天，VSS 维持在 8.7 至 10.2 g/L 之间，而 F/M 比率保持在 0.08 至 0.12 之间。生物膜的强附着性和稳健性在整个实验期间均得以保持（图 S5B）。这些结果与此前的研究一致（Maza-Márquez 等，2016a）。

3.2. Efficiency of PBR for the removal of pollutants from OWW
3.2. PBR 对 OWW 污染物去除的效率

The PBR was operated for 140 days, divided into four operational phases of 35 days each using different HRTs (5, 4, 3 and 2 days). Changes in COD, BOD5, total phenol, color and turbidity monitored in the PBR influent and effluent are displayed in Fig. 2, and the notched box-plots in Fig. S7 compare the removal efficiencies of all the aforementioned parameters under operation at different HRTs. No significant differences were found among the four experimental phases (Kruskal-Wallis test, p > 0.05). The average removal rates achieved were: 85.86 ± 1.24% for COD, 95.86 ± 0.98% for turbidity, 99.12 ± 0.17% for BOD5, 94.84 ± 0.55% for total phenols, and 87.24 ± 0.91% for color (Table S2B). Total N and P in the effluent (Fig. 2) were <3 and <2 mg L−1, respectively, regardless of the operational stage, suggesting the uptake of these nutrients by the microbial biomass. Throughout the whole full-scale experiment, the dissolved oxygen concentration was >8 mg L−1, indicating an intense photosynthetic activity associated to microalgae and cyanobacteria in the biomass. The pH of the effluents remained close to neutrality (Table S2B), making their discharge into the environment safe.
PBR 运行了 140 天，分为四个运行阶段，每阶段 35 天，采用不同的 HRT（5 天、4 天、3 天和 2 天）。PBR 进水和出水中 COD、BOD、总酚、颜色和浊度的变化如图 2 所示，图 S7 中的凹槽箱线图比较了不同 HRT 条件下上述参数的去除效率。四个实验阶段之间没有显著差异（Kruskal-Wallis 检验，p > 0.05）。平均去除率分别为：COD 为 85.86 ± 1.24%，浊度为 95.86 ± 0.98%，BOD 为 99.12 ± 0.17%，总酚为 94.84 ± 0.55%，颜色为 87.24 ± 0.91%（表 S2B）。无论运行阶段如何，出水中的总氮和总磷分别低于 3 mg L 和 2 mg L（图 2），表明这些营养物质被微生物生物量吸收。在整个全规模实验过程中，溶解氧浓度保持在>8 mg L，表明生物量中的微藻和蓝藻与光合作用相关的活性较强。出水的 pH 值接近中性（表 S2B），确保了其排放到环境中的安全性。
  1. Download: Download high-res image (690KB)
    下载：下载高分辨率图片（690KB）
  2. Download: Download full-size image
    下载：下载原始尺寸图片

Fig. 2. Time course of color, turbidity, BOD5, COD, total phenols, total N and total P in the influent olive washing water (OWW) and photobioreactor (PBR). Vertical dashed lines separate the phases operated under different HRTs (5d, 4d, 3d and 2d).
图 2. 进水橄榄清洗水（OWW）和光生物反应器（PBR）中颜色、浊度、BOD 5 、COD、总酚、总氮和总磷的时间变化。竖虚线分隔了在不同水力停留时间（HRTs：5 天、4 天、3 天和 2 天）下运行的阶段。

Our results confirmed that the biofilm formed in the PBR showed a high capacity to remove pollutants from OWW. Additionally, the system reached stability in a short period of time. Ten days after the system started being fed with real OWW, an F/M ratio of ca. 0.1 was maintained, which is a typical value for treating industrial wastes (Ferrer-Polonio et al., 2015). Drastic changes of VSS concentration and F/M ratio were not further detected. Under these conditions, steady and high removal rates were achieved for all the pollutants measured, until the end of the study (Fig. 2, Table S2B). The observed high removal rates of phenolic compounds by the microalgae-bacteria consortium were consistent with previously published studies and can be explained by synergetic relationships (Essam et al., 2014, Mahdavi et al., 2015, Ryu et al., 2015). The ability of microalgae to remove N and P nutrients is also well documented in earlier literature (Vasconcelos-Fernandes et al., 2015, Delrue et al., 2016).
我们的研究结果证实，PBR 中形成的生物膜对去除 OWW 中的污染物具有很高的能力。此外，该系统在短时间内达到了稳定状态。在系统开始以实际 OWW 作为进料后的第 10 天，F/M 比维持在约 0.1，这是处理工业废物的典型值（Ferrer-Polonio 等，2015）。此后未检测到 VSS 浓度和 F/M 比的剧烈变化。在这些条件下，所有测量的污染物均实现了稳定且高效的去除率，直至研究结束（图 2，表 S2B）。微藻-细菌共生体系对酚类化合物的高去除率与先前研究结果一致，这可通过协同关系进行解释（Essam 等，2014；Mahdavi 等，2015；Ryu 等，2015）。关于微藻去除 N 和 P 营养物质的能力，早期文献中也有详细记载（Vasconcelos-Fernandes 等，2015；Delrue 等，2016）。
The PBR was able to treat approximately 4000 L d−1 at 2d HRT, covering the needs of the factory in which was installed and reducing its ecological footprint. Moreover, the PBR worked with minimum supervision, making it a low-cost system. The system was flexible enough to be adapted to the particular needs of other small local companies.
PBR 能够在 2 天水力停留时间（HRT）下处理约 4000 升的 d −1 ，满足其安装工厂的需求并减少生态足迹。此外，PBR 仅需最少的监管即可运行，使其成为一种低成本系统。该系统具有足够的灵活性，可以适应其他小型本地公司的特定需求。

3.3. Microbial community structure analyzed by Illumina-sequencing
3.3. 通过 Illumina 测序分析微生物群落结构

To better understand the biology of the PBR biofilms, the performance of two different strategies used for DNA extraction (algae DNA and soil DNA extraction kits) were compared. Subsequently, Illumina-sequencing was done, based on the amplification of total Bacteria 16S rRNA and Chlorophyta 18S rRNA partial genes from community DNA extracted by both methods.
为了更好地了解 PBR 生物膜的生物学特性，对用于 DNA 提取的两种不同策略（藻类 DNA 提取试剂盒和土壤 DNA 提取试剂盒）的性能进行了比较。随后，通过这两种方法提取的群落 DNA，对细菌 16S rRNA 和绿藻门 18S rRNA 部分基因进行了扩增，并进行了 Illumina 测序。
The normalized number of high quality reads for both total Bacteria 16S rRNA and Chlorophyta 18S rRNA amplicons was 97,320 (12,165 sequences per sample). The rarefaction curves (Fig. S8) and the values of Good’s coverage (>99%, Table S3) indicated that the sequencing depth achieved sufficed to properly describe the diversity of the target communities represented in the environmental DNA samples.
**以下为简体中文翻译：** 总细菌 16S rRNA 和绿藻门（Chlorophyta）18S rRNA 扩增子的标准化高质量测序数为 97,320 条（每个样本 12,165 条序列）。稀释曲线（图 S8）和 Good’s 覆盖度值（>99%，表 S3）表明，测序深度足以充分描述环境 DNA 样本中目标群落的多样性。
The calculated species richness, Chao-1 and Shannon-Wiener H′ indices, and Simpson’s index were summarized in Table S3. A maximum of 240 and 332 PHYs (algae and soil DNA kits, respectively) were detected for the total Bacterial community at 97% identity of the 16S rRNA gene sequences (3% cut-off level). Overall, the number of PHYs was quite stable in the samples under all the operational conditions tested. According to the average values of the Chao-1 and H′ indices, the diversity and species richness in the samples was medium, regardless of the method of DNA extraction used (Chao-1: 437.4 ± 175.4 and 512.6 ± 138.4; H′: 1.77 ± 0.42 and 2.5 ± 0.11, for the algae and soil DNA kits, respectively). In contrast, the Simpson’s index described a medium or low functional organization of the community, depending on extraction method selected (0.49 ± 0.12 and 0.79 ± 0.02 for the algae and soil DNA kits, respectively). The Chlorophyta community was composed of a low number of species (9–10 PHYs), regardless of the extraction method used. The average values of Chao-1 (9.5 ± 1 and 8 ± 0, for the algae and soil DNA kits, respectively) and H′ (0.35 ± 0.09 and 0.56 ± 0.06 for the algae and soil DNA kits, respectively) also described low richness and diversity of this microbial group, and the Simpson’s index (0.15 ± 0.04 and 0.27 ± 0.03 for the algae and soil DNA kits, respectively) evidenced a high functional organization of the community (Table S3).
计算的物种丰富度、Chao-1 指数、Shannon-Wiener H′指数和 Simpson 指数的结果总结在表 S3 中。在 97%的 16S rRNA 基因序列相似性水平（3%的差异阈值）下，总细菌群落中检测到的最大物种操作分类单元（PHYs）数分别为 240（藻类 DNA 提取试剂盒）和 332（土壤 DNA 提取试剂盒）。总体而言，在所有测试的操作条件下，样本中的 PHYs 数量相对稳定。根据 Chao-1 和 H′指数的平均值，无论使用何种 DNA 提取方法，样本中的多样性和物种丰富度均属中等水平（藻类 DNA 提取试剂盒：Chao-1 为 437.4 ± 175.4，H′为 1.77 ± 0.42；土壤 DNA 提取试剂盒：Chao-1 为 512.6 ± 138.4，H′为 2.5 ± 0.11）。相比之下，Simpson 指数则描述了群落的功能组织化程度在中等或较低水平，具体取决于所选的提取方法（藻类 DNA 提取试剂盒为 0.49 ± 0.12，土壤 DNA 提取试剂盒为 0.79 ± 0.02）。 绿藻门群落的物种数较少（9–10 个 PHYs），无论使用哪种提取方法。Chao-1（藻类 DNA 提取试剂盒为 9.5 ± 1，土壤 DNA 提取试剂盒为 8 ± 0）和 H′指数（藻类 DNA 提取试剂盒为 0.35 ± 0.09，土壤 DNA 提取试剂盒为 0.56 ± 0）也表明这一微生物群体的物种丰富度和多样性较低。而 Simpson 指数（藻类 DNA 提取试剂盒为 0.15 ± 0.04，土壤 DNA 提取试剂盒为 0.27 ± 0）则显示该群落具有较高的功能组织化程度（表 S3）。
The Bray-Curtis based dendrograms in Fig. S9 display the similarity between the PBR biofilm samples analyzed, based on the relative abundance of populations of total Bacterial and Chlorophyta detected by Illumina-sequencing. Remarkable differences of the total Bacteria community structure were detected depending on the DNA extraction kit chosen. The global similarity among samples was >75% using the algae DNA kit, while it was >94% with the soil DNA kit. Moreover, significant differences of the similarity of the Bacteria community were detected between the operational periods with the algae DNA kit, but not with the soil DNA kit. In contrast, the global similarity of the Chlorophyta community among samples was >94%, regardless of the DNA kit used, indicating a high stability of the microalgae community throughout all the experimental phases.
Bray-Curtis 基于分析的树状图（图 S9）显示了对 PBR 生物膜样本的相似性分析结果，该分析基于 Illumina 测序检测到的总细菌和绿藻门（Chlorophyta）群落的相对丰度。研究发现，根据所使用的 DNA 提取试剂盒的不同，总细菌群落结构存在显著差异。使用藻类 DNA 提取试剂盒时，样本之间的总体相似性超过 75%；而使用土壤 DNA 提取试剂盒时，相似性高达 94%以上。此外，使用藻类 DNA 提取试剂盒时，不同运行周期之间的细菌群落相似性存在显著差异；但使用土壤 DNA 提取试剂盒时，这种差异并不显著。相比之下，无论使用哪种 DNA 提取试剂盒，绿藻门群落样本之间的总体相似性均超过 94%，表明在整个实验阶段中，微藻群落具有较高的稳定性。
According to ANOSIM, significant differences were found among the groups of samples belonging to each DNA extraction method (R = 1 and R = 0.844, for total Bacteria and Chlorophita, respectively). These results showed the remarkable influence of the DNA kits on the evaluation of the microbial community structure. However, in spite of their different resolution, both extraction methods reflected that the structure of the communities was quite stable (similarity > 75%) throughout the full experimental period. These results are in agreement with the stability of the performance of pollutants’ removal observed throughout time and using different HRTs (Table S2B).
根据 ANOSIM 分析，不同 DNA 提取方法的样本组之间存在显著差异（总细菌和绿藻门的 R 值分别为 1 和 0.844）。这些结果表明，DNA 提取试剂盒对微生物群落结构的评估具有显著影响。然而，尽管两种提取方法的分辨率不同，结果均表明，在整个实验期间，群落结构的稳定性较高（相似性 > 75%）。这些结果与实验中观察到的污染物去除性能的稳定性一致（使用不同的 HRT，见表 S2B）。

3.4. Diversity of total bacteria
3.4. 总细菌多样性

Fig. 3, Fig. 4 display the heatmaps of the relative abundance of bacterial PHYs for both DNA kits in the four experimental phases, at the Phylum, Order and Genus levels. Clear differences were found between the two methods of DNA extraction, regarding the relative abundances of several of the taxonomic groups detected.
图 3 和图 4 展示了在四个实验阶段中，两种 DNA 试剂盒在菌门、目和属水平上细菌 PHYs 相对丰度的热图。对于检测到的多个分类群的相对丰度，两种 DNA 提取方法之间存在明显差异。
  1. Download: Download high-res image (409KB)
    下载：下载高分辨率图片（409KB）
  2. Download: Download full-size image
    下载：下载原始尺寸图片

Fig. 3. Heatmaps showing the relative abundances of bacterial phylotypes detected by Illumina-sequencing in the photobioreactor (PBR) operated under different HRTs (5d, 4d, 3d, 2d) using the algae DNA kit. A. Classification at Phylum level. B. Classification at Order level. C. Classification at Genus level.
图 3. 热图显示了通过 Illumina 测序检测到的光生物反应器（PBR）在不同水力停留时间（HRTs，5 天、4 天、3 天、2 天）下使用藻类 DNA 试剂盒得到的细菌类群的相对丰度分布情况。A. 门水平分类。B. 目水平分类。C. 属水平分类。

  1. Download: Download high-res image (339KB)
    下载：下载高分辨率图片（339KB）
  2. Download: Download full-size image
    下载：下载原始尺寸图片

Fig. 4. Heatmaps showing the relative abundances of bacterial phylotypes detected by Illumina-sequencing in the photobioreactor (PBR) operated under different HRTs (5d, 4d, 3d, 2d) using the soil DNA kit. A. Classification at Phylum level. B. Classification at Order level. C. Classification at Genus level.
图 4. 热图显示了使用土壤 DNA 试剂盒通过 Illumina 测序检测到的细菌分类群在不同 HRT（5 天、4 天、3 天、2 天）条件下光生物反应器（PBR）中的相对丰度。A. 门水平分类。B. 目水平分类。C. 属水平分类。

When the algae DNA kit was used, the bacterial populations detected were affiliated to 4 different Phyla (Fig. 3A). Cyanobacteria were dominant (80.09% average relative abundance), followed by Proteobacteria (17.05%), Bacteroidetes (1.55%) and Actinobacteria (1.29%). In contrast, Proteobacteria was the dominant Phylum when the soil DNA kit was used (98.18% relative abundance), followed by Firmicutes (1.10%) and Bacteroidetes (0.71%) (Fig. 4A).
使用藻类 DNA 试剂盒时，检测到的细菌群落隶属于 4 个不同的门（图 3A）。蓝细菌占主导地位（平均相对丰度为 80.09%），其次是变形菌门（17.05%）、拟杆菌门（1.55%）和放线菌门（1.29%）。相比之下，当使用土壤 DNA 试剂盒时，变形菌门是主要的优势门（相对丰度为 98.18%），其次是厚壁菌门（1.10%）和拟杆菌门（0.71%）（图 4A）。
At Order level, differences in taxonomic resolution were also detected based on the DNA kit selected. Overall, 9 Orders were detected by the algae DNA kit, and 7 with the soil DNA kit. Sphingomonadales and Rhizobiales were the only orders detected by both methods. The dominant Order detected with the algae DNA kit was Nostocales (76.50% relative abundance) (Fig. 3B), while Rhizobiales and Pseudomonadales were the dominant orders detected with the soil DNA kit (51.66% and 39.07% relative abundance, respectively) (Fig. 4B). At the Genus level, the highest relative abundances were observed in all samples for PHYs affiliated to the cyanobacteria Hapalosiphon (69.58%) when the algae DNA kit was used (Fig. 3C). The sequencing of samples processed using the soil DNA kit reflected a community dominated by Proteobacteria related to the genera Azotobacter (32.39%) and Rhodopseudomonas (30.57%), followed by Nitrobacter (7.49%), Sphingobium (3.34%) and Rhizobium (2.90%) (Fig. 4C).
在目（Order）水平，不同 DNA 试剂盒的选择也导致了分类学分辨率的差异。总体而言，使用藻类 DNA 试剂盒检测到 9 个目，而使用土壤 DNA 试剂盒检测到 7 个目。Sphingomonadales（鞘脂单胞菌目）和 Rhizobiales（根瘤菌目）是两种方法均检测到的目。使用藻类 DNA 试剂盒检测到的优势目是念珠藻目（Nostocales），其相对丰度为 76.50%（图 3B）；而使用土壤 DNA 试剂盒检测到的优势目是根瘤菌目（Rhizobiales）和假单胞菌目（Pseudomonadales），其相对丰度分别为 51.66%和 39.07%（图 4B）。 在属（Genus）水平，使用藻类 DNA 试剂盒时，所有样本中相对丰度最高的是隶属于蓝藻门（Cyanobacteria）的 Hapalosiphon 属（69.58%）（图 3C）。而使用土壤 DNA 试剂盒处理的样本测序结果反映了一个以变形菌门（Proteobacteria）为主的群落，优势属为 Azotobacter（固氮菌属，32.39%）和 Rhodopseudomonas（红假单胞菌属，30.57%），其次是 Nitrobacter（硝化杆菌属，7.49%）、Sphingobium（鞘脂单胞菌属，3.34%）和 Rhizobium（根瘤菌属，2.90%）（图 4C）。
In order to evaluate which PHYs contributed most to the dis(similarity) between the PBR biofilm samples, SIMPER analyses were carried out. The global similarity between the samples according to the results derived from the algae DNA kit was 85.64%. PHY 1 (Hapalosiphon) contributed 81.21% to explain the similarity between the samples, while the contribution of other PHYs was minor (<2%) (Table S4A). The global similarity between samples taking into consideration only the data derived from the soil DNA kit was 96.00%, and it was mostly explained by PHY 32 (Azotobacter, 37.23% contribution) and PHY 21 (Rhodopseudomonas, 33.99%) (Table S4B). SIMPER was also used to evaluate which PHYs were the major contributors to the dissimilarities of the biofilm bacterial community detected comparing the two DNA kits used (Table S5). Three PHYs explained >76% of these differences (PHY 1, Hapalosiphon; PHY 32, Azotobacter; PHY 21, Rhodopseudomonas).
为了评估哪些 PHY 对 PBR 生物膜样本之间的相似性或差异性贡献最大，进行了 SIMPER 分析。根据藻类 DNA 试剂盒的结果，样本之间的总体相似性为 85.64%。PHY 1（Hapalosiphon）对样本相似性的解释贡献为 81.21%，而其他 PHY 的贡献较小（<2%）（表 S4A）。仅根据土壤 DNA 试剂盒数据计算的样本总体相似性为 96.00%，主要由 PHY 32（Azotobacter，贡献 37.23%）和 PHY 21（Rhodopseudomonas，贡献 33.99%）解释（表 S4B）。SIMPER 还用于评估在比较两种 DNA 试剂盒时，哪些 PHY 对生物膜细菌群落差异的主要贡献最大（表 S5）。有三个 PHY 解释了>76%的差异（PHY 1，Hapalosiphon；PHY 32，Azotobacter；PHY 21，Rhodopseudomonas）。
When the experimental phases operated at different HRTs were compared pairwise, SIMPER corroborated the stability of the bacterial community structure in the PBR, showing average dissimilarity <25% (Table S6). According to the data derived from the algae DNA kit, the PHYs mostly explaining these dissimilarities were PHY 1 (Hapalosiphon, 7.99–50.65% contribution), PHY 15 (Lysobacter, 6.86–10.70%), PHY 19 (Rhizobium, 4.14–9.56%) and PHY 21 (Rhodopseudomonas, 5.02–19.42%) (Table S6A). In the case of data generated from the soil DNA kit, the major contributors were PHY 21 (Rhodopseudomonas, 21.20–58.09), PHY 32 (Azotobacter, 5.52–24.63%) and PHY 36 (Nitrobacter, 9.65–21.94%) (Table S6B).
当对比在不同 HRT 下运行的实验阶段时，SIMPER 分析证实了 PBR 中细菌群落结构的稳定性，显示平均差异性小于 25%（表 S6）。根据藻类 DNA 试剂盒获得的数据，主要导致这些差异的 PHY 为 PHY 1（Hapalosiphon，贡献率为 7.99–50.65%）、PHY 15（Lysobacter，贡献率为 6.86–10.70%）、PHY 19（Rhizobium，贡献率为 4.14–9.56%）和 PHY 21（Rhodopseudomonas，贡献率为 5.02–19.42%）（表 S6A）。而根据土壤 DNA 试剂盒生成的数据，主要贡献者为 PHY 21（Rhodopseudomonas，贡献率为 21.20–58.09%）、PHY 32（Azotobacter，贡献率为 5.52–24.63%）和 PHY 36（Nitrobacter，贡献率为 9.65–21.94%）（表 S6B）。
The results obtained showed that the selection of the DNA extraction kit for the analysis of community structure and diversity in the PBR biofilms was a critical step, since each method used rendered a very different picture (94.45% average dissimilarity between methods, Table S5). Both DNA kits used are based on bead-beating disruption of the microbial cell envelopes; however, they differ in the nature of the lysing matrix used, which is remarkably harder in the case of the algae DNA kit, recommended by the manufacturer for organisms with dense cell walls. According to the results derived from this method, the microbial community in the PBR was dominated by Cyanobacteria of the Nostocales, which are characterized by a branched morphology and mucilaginous cell walls, and are often resistant to standard methods applied for the extraction of bacterial DNA from environmental samples (Srivastava et al., 2007). When the rRNA amplicons derived from DNA extracted using the soil DNA kit were sequenced, no Cyanobacteria were detected, most probably due to the softer lysing matrix failing to break their cell walls.
研究结果表明，在分析 PBR 生物膜的群落结构和多样性时，选择 DNA 提取试剂盒是一个关键步骤，因为不同的方法会呈现出完全不同的结果（方法之间的平均差异性为 94.45%，见表 S5）。两种使用的 DNA 试剂盒都基于珠磨破碎微生物细胞包膜的原理，但它们在裂解基质的性质上有所不同。藻类 DNA 试剂盒使用的裂解基质明显更硬，制造商推荐用于具有致密细胞壁的生物体。根据这种方法得到的结果，PBR 中的微生物群落以具有分枝形态和粘液状细胞壁的念珠藻目蓝细菌为主，这类蓝细菌通常对从环境样本中提取细菌 DNA 的标准方法具有抗性（Srivastava 等，2007）。而当使用土壤 DNA 试剂盒提取的 DNA 进行 rRNA 扩增子测序时，没有检测到蓝细菌，这很可能是由于较软的裂解基质未能破坏其细胞壁所致。
Among the Proteobacteria, populations of the Alphaproteobacteria and particularly the Rhizobiales were prevalent, and PHYs related to the genera Daeguia, Rhizobium, Rhodopseudomonas, Shinella and Tardiphaga were detected by both DNA kits. Strikingly, PHYs affiliated to some genera displayed a high relative abundance among the Proteobacteria according to the results of the soil DNA kit, but were not detected (Azotobacter, Nitrobacter) or detected at much lower relative abundances (Rhodopseudomonas) when using the algae DNA kit (Fig. 3, Fig. 4). These results further highlight the need to use more than one DNA extraction protocol when the diversity of complex microbial biofilm communities is analyzed by molecular methods.
在变形菌门中，α变形菌类群尤为显著，尤其是根瘤菌目（Rhizobiales）的种群占据了主导地位。通过两种 DNA 提取试剂盒均检测到了与 **Daeguia**、**Rhizobium**（根瘤菌）、**Rhodopseudomonas**（红假单胞菌）、**Shinella** 和 **Tardiphaga** 属相关的 PHYs（细菌群落物种）。值得注意的是，根据土壤 DNA 试剂盒的结果，某些属的 PHYs 在变形菌门中表现出较高的相对丰度，但在使用藻类 DNA 试剂盒时却要么未被检测到（如固氮菌**Azotobacter**、硝化杆菌**Nitrobacter**），要么检测到的相对丰度明显较低（如红假单胞菌**Rhodopseudomonas**）（图 3，图 4）。这些结果进一步表明，当通过分子方法分析复杂微生物生物膜群落的多样性时，有必要采用不止一种 DNA 提取方案。
Proteobacteria and Cyanobacteria were the prevailing Phyla detected in the PBR, which was consistent with the results reported in previous studies (Maza-Márquez et al., 2016a). Ryu et al. (2015) assessed a microalgae-bacteria consortium for the degradation of hazardous compounds in synthetic water using pyrosequencing, concluding that Proteobacteria was the dominant Phyla. Besides, recent studies demonstrated that Proteobacteria and Bacteroidetes are more prone than other bacterial Phyla to associate with green algae in symbiotic mutualistic relationships (Ramanan et al., 2016).
在 PBR（光生物反应器）中检测到的主要门类是变形菌门和蓝藻门，这与之前的研究结果一致（Maza-Márquez 等，2016a）。Ryu 等人（2015）使用焦磷酸测序技术评估了微藻-细菌联合体在合成废水中降解有害化合物的性能，发现变形菌门是主要的门类。此外，最近的研究表明，与其他细菌门相比，变形菌门和拟杆菌门更容易与绿藻形成共生互利关系（Ramanan 等，2016）。
PHYs related to Hapalosiphon, Rhodopseudomonas and Azotobacter were the dominant populations in the PBR biofilm community and consequently, the possible key players for the removal of the toxic compounds of OWW. Interestingly, these three microorganisms are known for fixing N2 under aerobic conditions, and additionally, Rhodopseudomonas may fix nitrogen also under anaerobic conditions (Emerich and Wall, 1985, Kargi and Ozmihci, 2004, Srivastava et al., 2007). Rhodopseudomonas spp. had been associated with the degradation of aromatic compounds in wastewater treatment (Adessi et al., 2016, Vincenzini et al., 1981). On the other hand, it is well document in literature that Azotobacter spp. have a potential for biological treatment of olive-oil wastewater and other complex industrial effluents (Cerrone et al., 2010, Gapes et al., 1999, Juárez et al., 2005). Minor dominant taxa detected in this study such as Rhizobium, Brevundimonas and Simplicispira have been reported to contribute to biodegradation of pollutants in association with microalgae (Ryu et al., 2015). Besides, Rhizobium is also able to fix nitrogen and is often described to improve the metabolism of plants and microalgae by different mechanisms, often not completely elucidated (Kim et al., 2014).
在 PBR 生物膜群落中，与**Hapalosiphon**、**Rhodopseudomonas**（红假单胞菌）和**Azotobacter**（固氮菌）相关的 PHYs 是主要种群，因此可能是去除橄榄油废水（OWW）中有毒化合物的关键微生物。值得注意的是，这三种微生物在有氧条件下均能够固定氮（N2），此外，红假单胞菌在厌氧条件下也可以进行氮固定（Emerich 和 Wall，1985；Kargi 和 Ozmihci，2004；Srivastava 等，2007）。已有研究表明，红假单胞菌属（Rhodopseudomonas spp.）与废水处理中芳香族化合物的降解有关（Adessi 等，2016；Vincenzini 等，1981）。另一方面，文献中广泛报道固氮菌属（Azotobacter spp.）在橄榄油废水和其他复杂工业废水的生物处理方面具有潜力（Cerrone 等，2010；Gapes 等，1999；Juárez 等，2005）。本研究中检测到的次要优势类群（如根瘤菌**Rhizobium**、短芽孢杆菌**Brevundimonas**和**Simplicispira**属）已被报道能够与微藻协同降解污染物（Ryu 等，2015）。此外，根瘤菌（Rhizobium）也能够固定氮，且经常被描述为通过不同的机制（许多尚未完全阐明）改善植物和微藻的新陈代谢（Kim 等，2014）。
The nature of the positive interactions among cyanobacteria and heterotrophic bacteria in either nature or engineered systems are still poorly known and go beyond the metabolic exchange of O2 and CO2. Recent work based on transcriptomic analyses gave insight into the possible mechanisms of the mutualistic association among Synechococcus and Shewanella, evidencing the exchange of micronutrients and growth factors such as Fe and amino acids, as well as interactions between both organisms in the response to oxidative stress (Beliaev et al., 2014).
在自然或工程系统中，蓝藻与异养细菌之间正向相互作用的本质仍然知之甚少，并且超出了 O 2 和 CO 2 的代谢交换。基于转录组分析的最新研究揭示了聚球藻（Synechococcus）和希瓦氏菌（Shewanella）之间可能存在的互利关联机制，显示了微量营养素和生长因子的交换，例如铁和氨基酸，以及两种生物在应对氧化应激中的相互作用（Beliaev 等，2014）。

3.5. Diversity of microalgae
3.5. 微藻的多样性

PHYs of the Chlorophyta detected in the samples of the PBR biofilms were classified into 4 different Classes: Chlorodendrophyceae, Chlorophyceae, Trebouxiophyceae and Ulvophyceae. Ulvophyceae were not detected when using the algae DNA kit (Fig. 5). Chlorophyceae was the dominant Class, for both methods used (92.24 and 99.72% average relative abundances, for the algal and soil DNA kits, respectively).
在 PBR 生物膜样本中检测到的绿藻门（Chlorophyta）的 PHYs 被分类为四个不同的类别：绿藻纲（Chlorodendrophyceae）、网翼藻纲（Chlorophyceae）、栅藻纲（Trebouxiophyceae）和石莼纲（Ulvophyceae）。使用藻类 DNA 提取试剂盒时未检测到石莼纲（图 5）。绿藻纲是主要的类别，两种方法的平均相对丰度分别为 92.24%（藻类 DNA 提取试剂盒）和 99.72%（土壤 DNA 提取试剂盒）。
  1. Download: Download high-res image (394KB)
    下载：下载高分辨率图片（394KB）
  2. Download: Download full-size image
    下载：下载原始尺寸图片

Fig. 5. Heatmaps showing the relative abundances of phylotypes of Chlorophyta detected by Illumina-sequencing in the photobioreactor (PBR) operated under different HRTs (5d, 4d, 3d, 2d). A. Classification at Class level using the algae DNA kit. B. Classification at Order level using the algae DNA kit. C. Classification at Class level using the soil DNA kit. D. Classification at Order level using the soil DNA kit.
图 5. 热图显示了在不同 HRT（5 天、4 天、3 天、2 天）条件下光生物反应器（PBR）中通过 Illumina 测序检测到的绿藻门分类群的相对丰度。A. 使用藻类 DNA 试剂盒进行分类至纲水平。B. 使用藻类 DNA 试剂盒进行分类至目水平。C. 使用土壤 DNA 试剂盒进行分类至纲水平。D. 使用土壤 DNA 试剂盒进行分类至目水平。

At the Order level, significant differences were found depending of the method of DNA extraction used (Fig. 5). PHYs classified in 8 different Orders were detected with the algae DNA kit: Sphaeropleales (91.65% average relative abundance), Chlorellales (3.71%), Chaetophorales (0.36%), Chaetopeltidales (0.15%), Chlamydomonadales (0.07%), Prasiolales (0.02%), Tetrasporales (0.006%), and Chlorodendrales (0.005%). Additionally, 4.02% of the PHYs were unclassified at the Order level. In contrast, Chlamydomonadales was the dominant order when the soil DNA kit was used (84.21% average relative abundance). Moreover, the Prasiolales and Chlorellales were not detected, while in contrast Ulotrichales were only identified in the PBR samples with this DNA kit. As in the case of the analysis of the Bacteria community, the selection of the DNA extraction method drastically influenced the output of the massive parallel sequencing.
在目级别上，根据所使用的 DNA 提取方法发现了显著差异（图 5）。使用藻类 DNA 试剂盒检测到的 PHYs 分属于 8 个不同的目：球藻目（平均相对丰度 91.65%）、小球藻目（3.71%）、毛藻目（0.36%）、毛盘藻目（0.15%）、衣藻目（0.07%）、绿膜藻目（0.02%）、四分藻目（0.006%）和绿树藻目（0.005%）。此外，有 4.02%的 PHYs 在目级别未分类。相比之下，使用土壤 DNA 试剂盒时，衣藻目是主要目（平均相对丰度 84.21%）。此外，未检测到绿膜藻目和小球藻目，而相反，乌藻目仅在使用该 DNA 试剂盒的 PBR 样本中被检测到。与细菌群落分析的情况类似，DNA 提取方法的选择对高通量测序的结果有显著影响。
SIMPER analysis calculated the average dissimilarity of the Chlorophyta community in the PBR samples and the contribution of each particular PHY (Table S7). In all cases, the community was highly stable (similarity >95%). For the samples processed with the algae DNA kit, the changes of the relative abundance of PHY 8 (Sphaeropleales) mostly explained the similarities among samples (94.54% contribution). When the DNA soil kit was used instead, the explanation provided by PHY 8 was lower (11.45%), since the major contributor was PHY 3 (Chlamydomonadales, 85.84%), which was only detected with this extraction method. As in the case of Bacteria, the Chlorophyta communities were highly dissimilar (87.88%, Table S7C) depending on the DNA kit chosen. SIMPER also demonstrated a high stability of the community, at different HRT and for both DNA kits (<5% dissimilarity).
SIMPER 分析计算了 PBR 样品中绿藻门群落的平均差异性以及每个特定 PHY 的贡献（表 S7）。在所有情况下，群落的稳定性都很高（相似性 >95%）。对于使用藻类 DNA 试剂盒处理的样品，PHY 8（球藻目）的相对丰度变化主要解释了样品间的相似性（贡献率为 94.54%）。而当使用土壤 DNA 试剂盒时，PHY 8 的解释力较低（11.45%），因为主要贡献者是 PHY 3（衣藻目，85.84%），该目仅通过这种提取方法检测到。与细菌的情况类似，绿藻门群落的差异性很大（87.88%，表 S7C），这取决于所选择的 DNA 试剂盒。SIMPER 还表明，在不同的 HRT 和两种 DNA 试剂盒条件下，群落的稳定性都很高（差异性 <5%）。
The class Chlorophyceae and specifically the Orders Sphaeropleales and Chlamydomonadales were the dominant green microalgae identified in the PBR. The Chlorophyceae is a morphologically and ecologically diverse class of green algae and are found in a wide range of habitats (Leliaert et al., 2012). The literature on the biodegradation abilities of microalgae is also very abundant, since they are often used in urban, agricultural and industrial wastewater treatments (Di Caprio et al., 2015, Min et al., 2011, Pérez et al., 2015). However, although there is an extended bibliographic record on the abilities of microalgae to treat pollutants, most of the studies are based on two well- documented genera, Scenedesmus and Chlorella (Maza-Márquez et al., 2016a). To date, only a few microalgae have been described, but available data suggest that their biodiversity is immensely underestimated (Sharma and Rai, 2011). This fact explains why the PHYs detected in this study could be only classified up to the Order level. Many questions remain also unanswered regarding the complex relationships among bacteria and microalgae, and further investigations are needed in this regard.
在 PBR 中鉴定出的主要绿色微藻为绿藻纲（Chlorophyceae），特别是球藻目（Sphaeropleales）和衣藻目（Chlamydomonadales）。绿藻纲是一个形态和生态多样性极高的绿藻类群，分布于各种不同的生境中（Leliaert 等，2012）。关于微藻生物降解能力的文献也非常丰富，因为它们常被用于城市、农业和工业废水处理（Di Caprio 等，2015；Min 等，2011；Pérez 等，2015）。然而，尽管关于微藻处理污染物能力的文献记录广泛，但大多数研究集中在两个研究充分的属——栅藻属（Scenedesmus）和小球藻属（Chlorella）（Maza-Márquez 等，2016a）。迄今为止，仅有少数微藻被描述，但现有数据表明，它们的生物多样性被极大低估了（Sharma 和 Rai，2011）。这一事实也解释了为什么本研究中检测到的 PHYs 只能分类到目这一层级。此外，关于细菌与微藻之间复杂关系的许多问题仍未解答，需进一步研究。

3.6. Linking microbial community structure and population dynamics in the PBR with operational parameters and environmental variables (MDS and BIO-ENV)
3.6. 将 PBR 中的微生物群落结构和种群动态与操作参数及环境变量（MDS 和 BIO-ENV）联系起来

The plots in Fig. 6 display the MDS ordinations of the samples retrieved from the PBR biofilms, according to the data on the relative abundance of Bacteria and Chlorophyta PHYs identified by Illumina-sequencing. Since temperature and solar radiation data displayed a high positive correlation (>0.95), their effects cannot be statistically differentiated (Clarke and Ainsworth, 1993) and the influence of this two variables on the ordination was represented as a single vector in the plots.
图 6 中的图显示了根据 Illumina 测序鉴定的细菌和绿藻门（Chlorophyta PHYs）相对丰度数据，从 PBR 生物膜中获取的样本的多维尺度分析（MDS）排序。由于温度和太阳辐射数据表现出高度正相关（>0.95），其影响在统计上无法区分（Clarke 和 Ainsworth, 1993），因此这两个变量对排序的影响在图中以单一向量表示。
  1. Download: Download high-res image (572KB)
    下载：下载高分辨率图片（572KB）
  2. Download: Download full-size image
    下载：下载原始尺寸图片

Fig. 6. Non-metric multidimensional scaling (MDS) plots, illustrating the ordinations of the samples retrieved from the PBR operated under different HRTs (5d, 4d, 3d, 2d), according to the relative similarity of total bacterial and Chlorophyta communities analyzed by Illumina-sequencing. A. MDS plots for total Bacteria derived from the amplification of the partial 16S rRNA gene. B. MDS plots for Chlorophyta phylum derived from the amplification of the partial 18S rRNA gene. In the left column, the MDS plots correspond to the samples processed with the soil DNA kit, while in the right column, MDS plots represent the samples processed with the algae DNA kit. The black vectors in the plots represent the direction throughout the ordination of the relative abundances of the bacterial and Chlorophyta phylotypes. The red vectors represent the strength and directional influence throughout the ordination of abiotic and operational variables: hydraulic retention time (HRT), phenols removal, color removal, food/microbial ratio (F/M), temperature/solar radiation (T/SR), COD, BOD5 and pH. The variables which best explained the distributions of the biotic data according to BIO-ENV analysis are marked which an asterisk (*). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
图 6. 非度量多维尺度分析（MDS）图，显示了在不同水力停留时间（HRTs，5 天、4 天、3 天、2 天）下运行的 PBR 中采集样本的排序情况，这些样本基于 Illumina 测序分析的细菌总群落和绿藻门群落的相对相似性。 A. 基于部分 16S rRNA 基因扩增的细菌总群落的 MDS 图。 B. 基于部分 18S rRNA 基因扩增的绿藻门群落的 MDS 图。 左侧列的 MDS 图对应使用土壤 DNA 提取试剂盒处理的样本，而右侧列的 MDS 图对应使用藻类 DNA 提取试剂盒处理的样本。图中的黑色向量表示细菌和绿藻门类群相对丰度在排序中的方向。红色向量表示在排序中非生物和操作变量的强度和方向性影响，包括水力停留时间（HRT）、酚去除率、颜色去除率、食物/微生物比值（F/M）、温度/太阳辐射（T/SR）、化学需氧量（COD）、生化需氧量（BOD 5 ）和 pH 值。根据 BIO-ENV 分析，最能解释生物数据分布的变量用星号（*）标注。（关于此图例中颜色的参考说明，请参阅本文的网络版。）

When the samples were ordinated based on the relative abundance of bacterial PHYs, BIO-ENV analysis detected that the F/M ratio and COD provided the best explanation (37.2%) for data derived using the algae DNA kit, while the F/M ratio, COD and pH were the best explaining variables (65.7%) when the soil DNA kit was used. The ordination of samples based on the relative abundance of PHYs of Chlorophyta was 100% explained by the variables F/M ratio and COD when the algae DNA kit was selected. Using the soil DNA kit, a high level of explanation was also provided (94.3%) by the variables COD and color removal.
当基于细菌 PHYs 的相对丰度对样本进行排序时，BIO-ENV 分析发现，使用藻类 DNA 试剂盒时，F/M 比和 COD 可以最好地解释数据（37.2%）；而使用土壤 DNA 试剂盒时，F/M 比、COD 和 pH 是最佳解释变量（65.7%）。当选择藻类 DNA 试剂盒时，基于绿藻门 PHYs 相对丰度对样本进行排序，可以通过 F/M 比和 COD 变量 100%解释。使用土壤 DNA 试剂盒时，COD 和脱色率变量也提供了较高的解释度（94.3%）。
Overall, BIO-ENV revealed that F/M ratio and COD were the variables most influential for the ordination of the PBR biofilm samples, regardless of the microbial community targeted (Bacteria or Chlorophyta) or the DNA extraction method used. In addition, increasing temperature/solar radiation and shorter HRTs, together with high F/M ratios, were always positively correlated with the removal efficiency of color and phenols (Fig. 6). The availability of nutrients, pH or microbial cell density has been reported as an important factor that controls the activity of biodegrading consortia (Burgess and Pletschke, 2008).
总体而言，BIO-ENV 分析表明，无论目标微生物群（细菌或绿藻）或 DNA 提取方法如何，F/M 比和 COD 是对 PBR 生物膜样本排序影响最大的变量。此外，温度/太阳辐射的升高和更短的水力停留时间（HRT），与高 F/M 比一起，总是与颜色和酚类去除效率的提高呈正相关（图 6）。营养物质的可用性、pH 值或微生物细胞密度已被报道为控制生物降解菌群活性的一个重要因素（Burgess 和 Pletschke，2008）。
A few studies found in the literature have described the HRT as the most important operating parameter in wastewater treatment using microalgae (Delrue et al., 2016, Matamoros et al., 2015, Muñoz and Guieysse, 2006), and both temperature and solar radiation are regarded as crucial environmental factors for the performance of outdoor PBRs (Chang et al., 2017, Muñoz and Guieysse, 2006). Interestingly, in this study the changes of the HRT and varying conditions of solar radiation and temperature (Table S2A) did not drastically influence the biology (Fig. 4, Fig. 5) or performance of the PBR (Fig. 2). These results were also supported by BIO-ENV and BEST analyses (Fig. 6). Algae, cyanobacteria and bacteria coexist in a wide range of extreme habitats and fight unfavorable environmental conditions by an array of mutualistic mechanisms (reviewed by Ramanan et al., 2016). As stated in Section 2.2, stable biofilms composed of indigenous species adapted to OWW were developed in the PBR, being able to offer resistance to environmental stress while providing optimal removal rates of pollutants. In this sense, the efficiency of microalgae-based treatments is known to decrease at lower temperatures due to slowing down of biological activities; however, this effect can be overcome by using cold-adapted photosynthetic strains (Muñoz and Guieysse, 2006). Finally, the longer HRT applied in the PBR in the first experiment, conducted in winter, may compensate for the lower solar radiation intensities (Muñoz and Guieysse, 2006).
一些文献研究将水力停留时间（HRT）描述为利用微藻进行污水处理时最重要的运行参数（Delrue 等，2016；Matamoros 等，2015；Muñoz 和 Guieysse，2006），而温度和太阳辐射则被认为是室外光生物反应器（PBR）性能的关键环境因素（Chang 等，2017；Muñoz 和 Guieysse，2006）。有趣的是，本研究中 HRT 的变化以及太阳辐射和温度的不同条件（表 S2A）并未显著影响生物特性（图 4、图 5）或 PBR 的性能（图 2）。这些结果也得到了 BIO-ENV 和 BEST 分析的支持（图 6）。藻类、蓝藻和细菌能够在各种极端环境中共存，并通过一系列互利机制应对不利的环境条件（Ramanan 等，2016 综述）。正如第 2.2 节所述，在 PBR 中形成了由适应油污废水（OWW）的本地物种组成的稳定生物膜，能够抵抗环境压力，同时提供最佳的污染物去除率。在这一方面，众所周知，基于微藻的处理效率在较低温度下会因生物活动减缓而降低；然而，通过使用适应低温的光合菌株可以克服这一影响（Muñoz 和 Guieysse，2006）。最后，在冬季进行的首次实验中，PBR 中应用的较长 HRT 可能弥补了较低太阳辐射强度的影响（Muñoz 和 Guieysse，2006）。
The Pearson product-moment correlations between the pairs of vectors displayed in Fig. 6 are detailed in Table S8. Strong correlations were found between the relative abundance of the dominant PHYs of bacteria and microalgae detected in the PBR biofilms (Fig. 3, Fig. 4, Fig. 5) and the environmental/operational parameters (Table S8). In that context, it was observed that the relative abundances of bacterial PHYs classified as Rhodopseudomonas and Azotobacter and microalgae of the Sphaeropleales displayed positive correlations with the removal of phenols and color (r > 0.60). The relative abundances of Hapalosiphon were also strongly correlated with color removal (r = 0.95). These results confirmed that the taxa displaying the highest relative abundances in the PBR appear to play crucial functions on the removal of pollutants from OWW. Microorganisms phylogenetically related to those identified in the present study were described previously as the key players in wastewater treatments where phenols and other toxic compounds were present (Adessi et al., 2016, Cerrone et al., 2010, Mahdavi et al., 2015).
图 6 中显示的向量对之间的 Pearson 积矩相关性详见表 S8。在 PBR 生物膜中检测到的主要细菌和微藻门类（PHY）的相对丰度（图 3、图 4、图 5）与环境/操作参数之间存在较强的相关性（表 S8）。在此背景下，观察到分类为红假单胞菌（Rhodopseudomonas）和固氮菌（Azotobacter）的细菌门类以及球藻目（Sphaeropleales）的微藻的相对丰度与酚类和色度的去除呈正相关（r > 0.60）。此外，Hapalosiphon 的相对丰度与色度去除也呈现出很强的相关性（r = 0.95）。这些结果证实，在 PBR 中显示最高相对丰度的分类群似乎在去除 OWW 污染物方面发挥了重要作用。本研究中鉴定出的微生物的系统发育相关种类此前已被描述为在处理含有酚类和其他有毒化合物的废水过程中起关键作用的微生物（Adessi 等，2016；Cerrone 等，2010；Mahdavi 等，2015）。
To the best knowledge of the authors, little previous research is available regarding the relationships between the structure of the microalgae-bacteria consortia in PBRs and the environmental/operational variables. The results of this study demonstrated that few populations became dominant in the PBR and correlated strongly with the removal of pollutants, thus bringing valuable information about their potential role in the process.
据作者所知，关于光生物反应器（PBR）中微藻-细菌共生体结构与环境/操作变量之间关系的研究较少。本研究结果表明，只有少数种群在 PBR 中占据主导地位，并与污染物的去除密切相关，从而为其在这一过程中的潜在作用提供了宝贵的信息。

4. Conclusions  4. 结论

The results presented here show the robustness of the PBR technology to treat OWW and its potential to be implemented in olive oil industries as an efficient and reliable method to treat their complex effluents. Illumina-sequencing provided novel data on the PBR microbial communities, improving the current knowledge of their biodiversity, and their population dynamics were related to changes of the environmental/operational variables using MDS. The selection of the DNA extraction method was a critical step for the survey of microbial diversity, in particular for the detection of Cyanobacteria with tough cell walls, which predominated in the PBR biofilms.
本文研究结果表明，PBR（光生物反应器）技术在处理橄榄油废水（OWW）方面具有较高的稳健性，并且有潜力作为一种高效且可靠的方法应用于橄榄油行业，用于处理其复杂的废水。Illumina 测序技术提供了关于 PBR 微生物群落的新数据，丰富了当前对其生物多样性的认识。同时，通过多维尺度分析（MDS），将其种群动态与环境/操作变量的变化联系了起来。在研究微生物多样性时，DNA 提取方法的选择是一个关键步骤，特别是在检测具有坚韧细胞壁的蓝藻（Cyanobacteria）方面，这类蓝藻在 PBR 的生物膜中占主导地位。

Acknowledgements  致谢

This research was supported by European Algatec project (FP7) SME/2008/1/232331. Paula Maza-Márquez would like to specially thank her beloved teacher and friend Mavi, for guiding and helping her to develop this work. “Thanks for trusting me. Always in our memories.”
本研究得到了欧洲 Algatec 项目（FP7）SME/2008/1/232331 的支持。Paula Maza-Márquez 特别感谢她敬爱的老师和朋友 Mavi，感谢她在本研究的指导和帮助。“感谢您对我的信任。永远铭记于心。”

Appendix A. Supplementary data
附录 A. 补充数据

Download: Download Acrobat PDF file (2MB)

Supplementary data 1. Supplementary Figures and Tables.

References

Cited by (40)

  • Microalgae-based wastewater treatment – Microalgae-bacteria consortia, multi-omics approaches and algal stress response

    2022, Science of the Total Environment

  • A comprehensive review on cultivation and harvesting of microalgae for biodiesel production: Environmental pollution control and future directions

    2020, Bioresource Technology
    Citation Excerpt :

    Meanwhile, there still exist some disadvantages such as poor light utilization, requirement of large areas of land, etc., which hinders the development of microalgae cultivation. PBRs are the most efficient and productive closed systems to improve the algal cultivation yields (Chen et al., 2011; Maza-Márquez et al., 2017). PBRs offer better culturing conditions control and growth parameters such as pH, temperature, CO2 and O2.

  • Microalgae-based wastewater treatment for nutrients recovery: A review

    2019, Bioresource Technology

  • A critical review on designs and applications of microalgae-based photobioreactors for pollutants treatment

    2019, Science of the Total Environment

  • A comprehensive review on the use of algal-bacterial systems for wastewater treatment with emphasis on nutrient and micropollutant removal

    2022, Bioengineered

  • Effects of the use of ornamental plants and different substrates in the removal of wastewater pollutants through microcosms of constructed wetlands

    2018, Sustainability Switzerland
View all citing articles on Scopus
View Abstract
© 2017 Elsevier Ltd. All rights reserved.