Screening of nanobody targeting alkaline phosphatase PhoX in Microcystis and detection of the PhoX in situ by fluorescence immunoassays

Cyanobacterial alkaline phosphatases (APases) play a key role in organophosphate utilization in freshwater. Tracking the distribution of APases can provide insights into the physiological response of phytoplankton to phosphorus nutrition. Extracellular APase PhoX, one of three prokaryotic APase families, is important for organophosphate utilization. Because the existing methods only give information on bulk APases activity, the specific contribution of the PhoX is little evaluated. To develop an immunoassay for PhoX detection, the phoX gene of Microcystis aeruginosa PCC7806 was expressed in Escherichia coli BL21 and the purified PhoX was used as a coated antigen for screening anti-PhoX nanobodies from an alpaca antibody library. After three rounds of panning, a nanobody 3H with the highest affinity was selected. Further tests showed 3H could specifically bind to Microcystis PhoX and be used for PhoX detection by immunoblotting analysis. Then we constructed two different fluorescent-labeled 3H, EGFP-3H (with an enhanced green fluorescent protein, EGFP) and FITC-3H (with fluorescein 5-isothiocyanate, FITC), as specific fluorescent dyes for PhoX. The fluorescence staining tests with laboratory strains and water bloom samples showed that both fluorescent-labeled nanobodies could visualize the distribution and evaluate relative expression levels of PhoX in Microcystis. The nanobody 3H could be a useful regent to develop fluorescence immunoassays for in situ analyses of Microcystis PhoX.


Introduction
Harmful Microcystis blooms are occurring more frequently and intensively as a result of nutrient loading and global warming (Harke et al. 2016;Sinha et al. 2017;Huo et al. 2021). Phosphorus (P) is a critical nutrient for phytoplankton growth and proliferation. Therefore, P restriction is commonly regarded as an efficient strategy for controlling harmful phytoplankton blooms (Schindler et al. 2008). Dissolved P mainly comprises inorganic P (DIP) and organic P (DOP). DIP (primarily orthophosphate [Pi])) is the most important P source for phytoplankton, owing to its bioavailability and rapid uptake (Falkowski and Raven 2007). However, Pi concentration is too low to support phytoplankton growth in some freshwater ecosystems (Cotner and Wetzel 1992;Correll 1998;Ivančić et al. 2021). Despite all this, many cyanobacteria can continue to grow under Pi-limited stress because the DOP acts could as a "P reserve" necessary for cyanobacterial community succession (Ji et al. 2017;Wang et al. 2019). Increasing the synthesis of alkaline phosphatases (APases), a class of catalysts for organophosphate mineralization, may serve as a compensatory strategy for P deficiency in cyanobacteria (Willis et al. 2019).
Prokaryotic APases are mainly grouped into three distinct families (PhoA, PhoD and PhoX) (Luo et al. 2009). PhoX is widely distributed among cyanobacteria and its expression is regulated by DIP concentration (Sebastian and Ammerman 2009;Lin et al. 2018). Therefore, it widely used as an indicator of P starvation (Zhang et al. 2021a). For measuring the activity of APases in phytoplankton, colorimetric and fluorometric detection with p-nitrophenyl phosphate (pNPP) and 4-methylumbelliferyl phosphate (4-MUP) substrates, respectively, have been widely employed (Davis and Mahaffey 2017; Yao Zu and Sujuan Hong contributed equally to this work. Krasaesueb et al. 2021). Additionally, the extracellular phosphatase in phytoplankton was detected at the single-cell level using an enzyme-labelled fluorescence approach (Bar-Yosef et al. 2010). However, those methods only give total activities of APases without identifying the specific APases responsible for these activities. Therefore, exploring a novel method for detecting alkaline phosphatase in situ is of great significance.
Immunoassays based on high specificity of antigen-antibody binding are powerful tools for the immunodetection of small molecule pollutants , biomacromolecules (Zhang et al. 2021b) and organisms (Chin Chwan Chuong et al. 2022). In recent years, phage display technology is a popular method to obtain sensitive and specific antibodies (Ledsgaard et al. 2022). The antibodies are recovered by multiple rounds of panning, then unique antibody sequences are identified and cloned into suitable vectors conveniently for antibody expression . The types of genetically engineered antibodies from the phage display library contain single-domain antibodies (sdAb) (Qiu et al. 2018), single-chain variable fragments (ScFv) (Xu et al. 2019) and peptides (Cheng et al. 2022). Single-domain antibodies (commonly referred to as nanobodies) derived from the camelid family and the sharks are small (~ 15 kDa) protein binders that retain the structural and functional properties of naturally occurring heavy chain-only antibodies (Hamers-Casterman et al. 1993;Greenberg et al. 1995). Nanobodies are appealing in the field of diagnostics and molecular imaging due to their favorable qualities such as facile production, excellent thermal stability and solubility, and easy tailoring (Bao et al. 2021;Muyldermans 2021). Nanobodies against native surface epitopes of cyanobacteria (Microcystis aeruginosa), green algae (Chlamydomonas reinhardtii) and dinoflagellates (Alexandrium minutum) were selected and conveniently converted into application-friendly reagents for intact cell imaging and flow cytometry detection (Jiang et al. 2013(Jiang et al. , 2014Mazzega et al. 2019;Folorunsho et al. 2021).
In this study we isolated nanobodies against Microcystis PhoX by panning an alpaca naïve single domain antibody library. The most positive nanobody 3H was expressed in E. coil BL21 and used for PhoX detection by immunoblotting analysis. Subsequently, we constructed two fluorescentlabeled 3H proteins (EGFP-3H and FITC-3H nanobodies) and used them to in situ investigate the PhoXs distribution in different Microcystis by fluorescence immunoassays.

Reagents and strains
Alpaca naïve single domain antibody library whose capacity is 2 × 10 10 cfu mL −1 and Escherichia coli TG1 (E. coli TG1) were purchased from Chengdu NB Biolab Co., Ltd.  (Waterbury 2006) at 28 ± 2℃ with a continuous illumination of 20 μmol photons m −2 s −1 . To analyze growth in Pi-starved medium, K 2 HPO 4 was omitted from BG-11 medium and replaced by equimolar amounts of KCl. Natural surface blooms with more than 90% of Microcystis cells were collected in the summer of 2021 from Lake Taihu and in the spring and early summer of 2022 from Lake Xuanwuhu (Nanjing city), China.

Expression and purification of PhoX
PhoX of PCC7806 was generated as described by Hong et al. (2021). The purified protein was checked by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and quantified by the Bradford assay (Bradford 1976).

Enrichment and selection of anti-PhoX phage particles
The enrichment and selection of anti-PhoX phage particles were conducted as recorded by Xu et al. (2018) with some modifications. Briefly, 100 μL phage particles mixed with 1.9 mL MPBS solution [containing 5% skim milk in phosphate-buffered saline (PBS)] were added into a six-well plate which was coated with the purified PhoX [the first, second and third round were 100, 50 and 25 μg mL -1 in carbonatebuffered saline (CBS), respectively]. The plate was shaken at 300 rpm for 2 h and 37℃ before being washed ten times with PBST solution (containing 0.1% Tween 20 in PBS). The binding particles were eluted by 1 mL trypsin (200 μg mL -1 ) and used to infect E. coil TG1 for amplification. Each round of input library particles was quantified to about 10 9 cfu mL -1 and enriched effects were determined by calculating the recovery rate.
The E. coil TG1 infected with the third-round enriched particles was spread on 2 × TY-AG solid medium (16 g tryptone, 10 g yeast extract and 5 g NaCl in 1 L double distilled water with 1.5% agar powder, 100 μg mL −1 ampicillin and 1% glucose) and cultured at 37℃ overnight. The next day, individual colonies were randomly picked into 96-well plates with 100 μL well −1 of 2 × TY-AG medium (containing 100 μg mL −1 ampicillin and 1% glucose in 2 × TY medium) and cultured for 2 h at 250 rpm and 37℃. KM13K07 helper phages were added 20 μL well −1 (approximately 10 12 cfu mL −1 ) and rescued for 2 h at 250 rpm and 37℃. The plates were centrifuged at 3 300 × g for 30 min at room temperature (RT). Then the pellets were resuspended with 200 μL well −1 of 2 × TY-AK medium (containing 100 μg mL −1 ampicillin and 50 μg mL −1 kanamycin in 2 × TY medium) followed by incubating overnight at 250 rpm and 30℃. The next day, after being centrifuged at 3 300 × g for 30 min, the supernatants were employed for monoclonal phage ELISA.

Colony PCR and DNA sequencing
The positive phage particles were identified by colony PCR and DNA sequencing. PCR reaction conditions were 98℃ for 5 min, 35 cycles of 98℃ for 15 s, 56℃ for 15 s and 72℃ for 15 s, and final extension at 72℃ for 2 min. The specific primers were pComb3xss-F and Gback-R and listed in Table 1. Subsequently, the PCR products were checked and sequenced. The antibody sequence alignment was performed by the DNAMAN software (version 9.0).
Preparation of nanobody 3H and EGFP-3H fusion protein (with enhanced green fluorescent protein).
The most positive nanobody 3H gene was amplified by polymerase chain reaction (primers: 3H-F and 3H-R) and cloned between Nco I and Not I sites of pET-29a( +) vector using a ClonExpress®Ultra One Step Cloning Kit. To construct the pET-29-EGFP vector, the EGFP gene was amplified (primers: EGFP-F and EGFP-R) and cloned between Nco I and Not I sites of pET-29a( +) vector. Subsequently, the 3H nanobody gene was amplified (primers: 3H-1-F and 3H-1-R) and cloned into Xho I of the pET-29-EGFP vector. The recombinant vectors were transformed into E. coli BL21. The EGFP and the nanobodies (3H and EGFP-3H) were expressed by adding 0.2 mM and 0.4 mM isopropylβ-D-thiogalactoside (IPTG) to induce expression for 12 h at 28℃ and 16 h at 16℃, respectively. All proteins were purified from the supernatant of whole cell lysate. Finally, the proteins were assessed by SDS-PAGE and quantified by the Bradford assay (Bradford 1976). All primer sequences are listed in Table 1.

Phosphate stress of PCC7806 and immunoblotting analysis
PCC7806 strain at the exponential growth phase was starved in the P-depleted BG-11 medium under the conditions described above for 0, 6 and 24 h, respectively. The cells were harvested by centrifugation at 8 000 × g for 5 min and ground with liquid nitrogen. The cell debris was resuspended with PBS solution before being centrifuged at 8 000 × g for 20 min. The supernatants and the purified PhoX were separated on SDS-PAGE gels, then the separated proteins were transferred to an immobilon-P transfer membrane (ABM, catalog number B500) for immunodetection. The nanobody 3H (5 μg mL −1 ) was used as the detection antibody.

Whole-cell ELISA
Microcystis strains (PCC7806, FACHB-929 and FACHB-1757) and the control strains (PCC6803 and green algae Chlorella sp. QZ and Scenedesmus sp. JK-37) were cultivated both in the P-depleted and P-repleted BG-11 medium for 6 h. The cells were collected by centrifugation in 1.5 mL

Synthesis of FITC-3H (with fluorescein 5-isothiocyanate)
The antibody 3H was dialyzed against 25 mmol L −1 Na 2 CO 3 / NaHCO 3 solution (pH 9.8) overnight at RT. Then, the protein solution was dialyzed against 100 mL FITC solution (Na 2 CO 3 /NaHCO 3 solution with 0.1 mg mL −1 FITC) for 24 h at RT in the dark. Finally, the yielding product was intensively dialyzed against PBS solution until the absorbance OD 480 of the mixture was almost zero. The synthesized FITC-3H was stored at -20℃ before further experiments.

Immunofluorescence staining
Colonial Microcystis sp. TH-1 and unicellular PCC7806 cultivated in the P-repleted and P-depleted BG-11 medium for 6 h were collected by centrifugation and washed with PBST three times. The strains were fixed using 4% paraformaldehyde for 15 min and blocked with 5% skim milk for 1 h. After being washed with PBST three times, the cells were incubated with 10 μg mL −1 EGFP-3H nanobody for 2 h in the dark. For the group of blank control, 10 μg mL −1 EGFP replaced EGFP-3H nanobody. After washing as noted above, the cells were analyzed by fluorescence microscopy using a Nikon ECLIPSE Ci-L and images were captured by NIS elements software (Nikon). Quantification of the fluorescence signals was conducted by the ImageJ software (version 1.52v). Natural samples were collected by centrifugation and washed with PBST three times. The samples were fixed using 4% paraformaldehyde for 15 min and blocked with 5% skim milk for 1 h. After being washed with PBST three times, the cells were incubated in turn with the nanobody 3H (10 μg mL −1 ) at RT for 2 h, anti-S-tag polyclonal antibody (1:5 000 dilutions in PBS) and goat anti-rabbit IgG antibody [Alexa Fluor® 488] (1:100 dilutions in PBS) at RT for 1 h. For direct immunofluorescent assays, the cells blocked with skim milk were incubated with 10 μg mL −1 EGFP-3H and FITC-3H nanobody for 2 h in the dark, respectively. After washing as noted above, the cells were analyzed by fluorescence microscopy using a Nikon ECLIPSE Ci-L and images were captured by NIS elements software (Nikon). Quantification of the fluorescence was conducted by the ImageJ software (version 1.52v).

Screening and identification of anti-PhoX phage particles
Microcystis PhoX was successfully expressed in E. coli BL21 and recovered from the supernatant of whole cell lysate, which displayed an approximately 74 kDa protein bank on an SDS-PAGE gel as shown in Fig. 1A. The final concentration of the recovered PhoX was 248 μg mL −1 , which met the need for phage screening.
To enrich anti-PhoX phage particles, three rounds of panning were conducted and the desired effect was achieved. Comparing the third round to the first round, the recovery rate was increased about 9 688-fold (Table 2). About 384 clones (4 × 96-well plates) from the third enriched library were employed for monoclonal phage ELISA and four positive particles (P/N > 2.8) were captured successfully (Fig. 1B). The most positive (3H) OD 450 value was 1.946 (control was 0.234). Four nanobodies displayed a real and similar size of nanobody gene fragments by colony PCR on a gel (data not shown). Then the nanobodies were identified by DNA sequencing and their amino acid sequences were displayed in Fig. 1C.

Generation and validation of the nanobody
In order to obtain nanobody, the most positive nanobody 3H was soluble expressed in E. coli BL21 and it was abundant in the supernatant of whole cell lysate ( Fig. 2A). The nanobody 3H was successfully recovered from the supernatant and its molecular weight plus that of the tags (S-tag and His tag) were approximately 15 kDa on an SDS-PAGE gel ( Fig. 2A). The final concentration of the nanobody in 250 mM imidazole solution was 1544 μg mL −1 , i.e., 9.26 mg L −1 in the 500 mL original culture. Subsequently, the PhoX-targeting capability of nanobody 3H was verified by both Western blotting and ELISA analysis. As shown in Fig. 2B, a clear band of about 72 kDa was present in the nitrocellulose membrane after incubating with nanobody 3H, confirming that nanobody 3H could be used for PhoX detection. In addition, a comparable result was given in ELISA. The absorbance OD 450 increase corresponded with a dose-dependent increase in the PhoX concentration (Fig. 2C).

Immunological test of PhoX in M. aeruginosa PCC7806
To clarify the response of Microcystis PhoX to environmental P levels, PhoX of PCC7806 cultivated both in the P-repleted and P-depleted conditions was evaluated by Western blotting analysis with nanobody 3H. PhoX was hardly expressed under P-replete condition, while it was increased significantly when PCC7806 was cultivated in a P-depleted medium for 6 h and lightly reduced after 24 h treatment (Fig. 3). The results indicated that nanobody 3H exhibited positive immunoreactivity with the PhoX and could

Specificity of the nanobody to different Microcystis strains
To evaluate the specificity of nanobody 3H, three Microcystis strains were tested by whole-cell ELISA, PCC6803 and two green algae as controls. All strains yielded very low fluorescence (almost same as background value) when they were cultivated in a P-repleted medium (Fig. 4). While the strains were starved in a P-depleted medium for 6 h, nanobody 3H recognized Microcystis strains (OD 450 1.6 ~ 2.1) and the control strains gave weaker signals (OD 450 about 0.5) almost the same as the background (Fig. 4). The results indicated that nanobody 3H specifically recognized Microcystis PhoX and gave no cross-reactivity with any of the other strains.

Generation of EGFP-3H and its application in Microcystis PhoX detection
To apply nanobody 3H conveniently for in situ observation of PhoX, we constructed an EGFP-3H fusion expression vector (Fig. 5A). The EGFP-3H nanobody and EGFP were successfully expressed and purified from the supernatant of whole cell lysate, which were approximately 42 kDa and 27 kDa protein bands on SDS-PAGE gels, respectively ( Fig. 5B and C).
To examine the performance of the EGFP-3H in immunofluorescence analysis, colonial and unicellular Microcystis were incubated with EGFP-3H, respectively, after cultivation with or without phosphate. EGFP replaced EGFP-3H as a control. For colonial Microcystis, an intense green fluorescence signal was illustrated in the extracellular polymeric substances (EPS) after strain was grown in a P-depleted BG-11 medium. However, neither the phosphate-grown strain nor the control exhibited any green fluorescence (Fig. 6A). In addition, unicellular Microcystis with all treatments also displayed no fluorescence (data not shown). Quantification of the fluorescence intensities showed a 9.37-fold increase after the colonial Microcystis was P-starved (Fig. 6B).

Immunofluorescence detection of PhoX in Microcystis blooms
PhoXs of Microcystis blooms were immunolocalized by immunofluorescent assays indirectly utilizing the nanobody (1:5 000 v/v dilutions in PBST) were used as the second and third antibodies in the ELISA, respectively. The OD 450 values were the means ± SDs from triplicate measurement 3H and directly using EGFP-3H and FITC-3H nanobodies. As shown in Fig. 7A, all nanobodies displayed intense green fluorescence signals in the EPS or on the cell surfaces of Microcystis. Notably, different fluorescence intensities appeared on Microcystis colonies stained with both EGFP-3H and FITC-3H, suggesting that the expression levels of PhoX could be diversified among Microcystis colonies originated from the same environment ( Fig. 7A and B). The results indicated that nanobody 3H might be utilized for in situ detection of Microcystis PhoX in natural blooms.

Comparison of the immunoassays with the chemical methods for APases detection
Various chemical methods have been developed for APase detection (Štrojsová et al. 2003;Ma et al. 2016). However, they merely report the total activity of bulk APases without identifying the nature or source of the APases involved. Typically, cyanobacteria have more than two APase-encoding genes (Lin et al. 2018). For example, two and four putative APase genes were found in M. aeruginosa PCC7806 and Anabaena sp. FACHB 709, respectively, and they may simultaneously involve in P metabolism and regulation under P-depleted stress (Liu and Wu 2012;Hong et al. 2021). Moreover, APases-producing bacteria are abundant in shallow eutrophic freshwater and some of them may adhere to the exopolysaccharide mucilage of Microcystis colonies (Dai et al. 2014;Smith et al. 2021). If this is the case, they might contribute a considerable proportion to total ATPase of Microcystis blooms since we were unable to absolutely exclude the bacteria from the viscous colonies (Cao et al. 2010). Immunoassays based on specific antibodies are sensitive and specific methods. In this study, PhoX-targeting nanobody 3H was selected and then conveniently converted to two fluorescent-labeled 3H proteins (EGFP-3H and FITC-3H). Immunoassays employing the fluorescent nanobodies eliminate the need for secondary antibodies and minimize time-consuming washing procedures. With the specificity of nanobody 3H to Microcystis PhoX, the fluorescent nanobodies could visualize the distribution of Microcystis PhoX free from the interference of adhered bacteria (Figs. 6A and 7A). Although both fluorescent nanobodies perform satisfactorily in immunofluorescence detection of Microcystis PhoX from natural blooms, EGFP-3H was superior to FITC-3H. EGFP-3H was produced as an EGFP fusion protein by recombinant DNA technology in E. coil, thereby eliminating the need for a laborious fluorescent-labeling process. On the other hand, FITC is susceptible to cyanobacterial autofluorescence and easily quenched by light. EGFP with a larger molecular weight (about 27 kDa) may be prone to nonspecific binding, thereby, in-situ detection by immunoassay based on the EGFP-3H could give a relatively high fluorescence background value for tracking Microcystis PhoX in complex environmental samples. This limits the practical application of EGFP-3H in Microcystis PhoX detection. To reduce the fluorescent background as much as possible, the working concentration of EGFP-3H needs to be optimized carefully. Furthermore, alternative fluorescent dyes including lanthanide chelates (Hagan and Zuchner 2011) and nanomaterials (Elmizadeh et al. 2019;Luo et al., 2021) with different absorbance/emission characteristics could be substituted for EGFP to satisfy additional applications. Generally, genetically engineered antibody selected from naïve and immune libraries tend to possess lower affinity than their parental counterparts. To improved antibody sensitivity, the isolation method must be optimized or the antibody genes must be affinitymatured (Sheedy et al. 2007). Furthermore, some signal amplification strategies in immunoassays have been developed, for example multivalent antibodies (Li et al. 2020). In our study, the relative expression level of Microcystis PhoX in single colony could be determined by quantitative fluorescence imaging (Figs. 6B and 7B), however, the immunological methods failed to achieve PhoX quantitative detection. An enzyme-labeled fluorescence alcohol standard has been introduced for quantitatively evaluating the single-cell phosphatase activity of microplankton (Diaz-de-Quijano et al. 2020). The nanobody 3H could be a useful regent to develop fluorescence immunoassays for quantitatively detecting Microcystis PhoX. We believe that immunoassay should open up new ways for alkaline phosphatase detection.

Subcellular localization of Microcystis PhoX
Subcellular localization of APases would provide insights into the physiological response of cyanobacteria to P deficiency and improve our understanding of the P cycle. Exceeding 50% of PhoX in marine bacteria were distributed in extracellular matrix (Luo et al. 2009). Extracellular APases are either dissolved enzymes liberated into the extracellular microenvironment or cell surface-localized enzymes (Chróst 1991). Microcystis PhoX carries a Tat-way signal peptide that might facilitate PhoX release into the extracellular environment, corresponding with the properties of bacterial PhoX (Wu et al. 2007;Hong et al. 2021). However, no studies reliably describe the location of PhoX in Microcystis.
In this study two fluorescent-labeled 3H proteins (EGFP-3H and FITC-3H) were used for the microscopic observation of Microcystis PhoX from both laboratory cultures and natural samples. Intense green fluorescence was observed in the EPS or on the cell surfaces of colonial Microcystis. However, no fluorescence was detectable in unicellular Microcystis, which was consistent with the results described by Ding et al. (2019) and Wu et al. (2009). In the culture of M. aeruginosa, the APases released to the medium was the major contributor to total APases when available Pi was scarce TH-1 cultivated with and without phosphate. Images were acquired after incubating cells with the nanobody 3H (10 μg mL −1 ). Green signals were from EGFP-3H and the red corresponding to the autofluorescence of chlorophyll a. Scale bars represent 50 μm. (B) Mean PhoX fluorescence intensity from Microcystis was determined by quantitative fluorescence imaging. The fluorescence was normalized to Microcystis cultivated without phosphate for 6 h and stained by EGFP. Data are represented as means ± SDs from triplicate measurement. For statistical analysis, the student's t-test was performed, **P < 0.01 a Fig. 7 (A) Immunofluorescence images of PhoX in Microcystis from natural blooms. Images were acquired after incubating cells with the nanobody 3H (10 μg mL −1 ), EGFP-3H (10 μg mL −1 ) and FITC-3H (10 μg mL −1 ), respectively. Green signals were from the goat antirabbit IgG [Alexa Fluor 488], EGFP-3H and FITC-3H, respectively, and the red corresponding to the autofluorescence of chlorophyll a. Scale bars represent 50 μm. Mean PhoX fluorescence intensities from colonial Microcystis were determined by quantitative fluorescence imaging. The fluorescence was normalized to colonial Microcystis ① (B) and Microcystis ④ (C), respectively. Data represented as means ± SDs from triplicate measurement. Different letters on the histograms indicate statistically significant differences at P < 0.05 by one-way ANOVA followed by Tukey's test (Wan et al. 2019). Therefore, unicellular Microcystis could contain too little cell-bound PhoX to be detected by immunofluorescent assay.
In natural blooms several hundreds of Microcystis cells aggregate and form colonies surrounded by EPS-containing mucilage. However, strains would lose colonial characteristics in long-term laboratory culture (Xiao et al. 2017(Xiao et al. , 2018. Colonial strains possess higher affinity and lower consumption for phosphate than unicellular strains with lower Pi, which may attribute to the difference of the capsules in nutrient sequestration and processing (Shen and Song 2007). Colonial Microcystis form a mucilaginous coat to enlarge the size of cells for storing more extracellular enzymes, which was helpful to capture nutrient compounds from the surrounding water (Flemming and Wingender 2010). This may also be one of the reasons why Microcystis colonies have a competitive advantage and become a dominant population in water.

Application of immunofluorescence staining for Microcystis blooms
Microcystis collected from natural blooms were stained by the fluorescent nanobodies (EGFP-3H and FITC-3H) and intense green fluorescence was observed in the EPS or on the cell surfaces of colonial Microcystis (Fig. 7A). The APases activity for colonial Microcystis isolated from the lake were detectable with ELF microscopy with low-P treatment (Horst et al. 2014). Contrary to our results for Microcystis spp., Wan et al. (2019) found that no Microcystis spp. cells were stained with the ELF-labelling test. A possible reason is that APases generation response to P availability might be strain-specific. Microcystis display a different affinity for phosphate and some strains produced little APases even under low P conditions (Hernández et al. 1999;Shen and Song 2007). Besides, the total phosphorus was different among strains and positively correlated with the growth rate ). Our study indicated that the expression level of PhoX varies among colonial Microcystis from the same environment and may be correlated with the physiological state (Fig. 7). The loose colony which should be at its older stage, was not labeled (white arrow) (Fig. 7A). Microcystis strains may have different phosphate availability requirements and display various phosphate deficiency responses. This physiological diversity could facilitate Microcystis adaptation to different environments.

Conclusions
We selected a PhoX-targeting nanobody 3H from an alpaca antibody library. The expressing 3H from E. coil was specifically bound to Microcystis PhoX. Two fluorescent-labeled 3H proteins (EGFP-3H and FITC-3H nanobodies) displayed reasonable results when used to detect Microcystis PhoX in situ. The nanobody 3H could be used to developed a fluorescence immunoassay for in situ analysis of Microcystis PhoX in natural blooms.