Directional Plk1 Inhibition-Driven Cell Cycle Interruption Using Amphiphilic Thin-Coated Peptide-Lanthanide Upconversion Nanomaterials as in vivo Tumor Suppressors

Please note that technical editing may introduce minor changes to the text and/or graphics, which may alter content. The journal’s standard Terms & Conditions and the Ethical guidelines still apply. In no event shall the Royal Society of Chemistry be held responsible for any errors or omissions in this Accepted Manuscript or any consequences arising from the use of any information it contains. Accepted Manuscript Journal of Materials Chemistry B


Introduction
2][3][4][5] However, the actual role of Plk1 in carcinogenesis is still yet to be fully elucidated.Some pre-clinical evidence suggests that overexpression of Plk1 is linked to an increased survival rate of cells, 6 and this is supported by the fact that Plk1 is normally highly expressed in actively proliferating tissue. 7trategic inactivation of Plk1 results in cell cycle arrest and subsequently apoptosis, thus Plk1 has emerged as a favorable therapeutic target for cancer treatment. 8ithin Plk1, there are two distinct domains-an aminoterminal kinase domain and a unique carboxyl terminal polo box domain (PBD).Several organic small molecules which can inhibit the Plk1 kinase domain and the PBD have been identi-ed.Molecules such as BI2536, CBB2001 and MLN0905 target the kinase domain of Plk1 causing inhibition as well as pronounced dose-limiting toxicity (DLT), [9][10][11] however unspecic activities and unwanted side-effects have really limited the further development of Plk1 kinase domain targeting molecules.On the other hand, peptides and their analogues that can bind to the PBD present a novel and promising avenue for Plk1 targeting and inhibition. 12,13Peptides can be employed as tailormade inhibitors and potential drugs but there are issues in working with this class of biological molecules.Peptides oen display physical and chemical instabilities, 14 poor cellular uptake, and localization issues. 15Real-time assessments, including MTT, in vitro imaging and tumor inhibition, are also crucial elements in cancer prognosis and treatment.Conventionally, for imaging purposes complex methods of analysis such as immunostaining of Plk1 using antibodies or fusion uorescent proteins have been employed. 16,17Recently, E. K. Ryu et al. demonstrated that a Plk1-specic peptide (PLHSpT) conjugated with the a v b 3 integrin receptor inhibitor peptide (RGD) and a radioligand can achieve the much sought-aer theranostic effect. 14However, this is a rather indirect method of analysis as it only tests cancer cells which involves multiple steps to interrupt the cell cycle, inhibit the tumor size, and diagnose the cancer tissues with PET.Most recently, we have reported a porphyrin conjugated to a Plk1-specic peptide (e.g.PLHSpT) as a multi-modal bio-probe, with the molecular system serving as the vector. 18Its overall tumor growth inhibition efficiency through the singlet oxygen generation has, indeed, still room for further improvement.
In the present study we have prepared lanthanide-doped upconversion nanoparticles (UCNPs) which can act as delivery vehicles for Plk1-specic peptides.The nano-particle-peptide conjugates were shown to inhibit Plk1 activity, while cell-cycle arrest and visualization can also be achieved simultaneously.Owing to the excellent photo-physical properties of lanthanide ions, the dopant ions (ytterbium/erbium) can be excited and emit in the near-infrared (NIR) region.0][21][22][23][24][25][26][27][28] By conjugating the PBD binding peptide onto UCNPs, the cellular uptake problem of the individual peptides can also be addressed because the resulting organic-inorganic nanomaterial is highly biocompatible, 29,30 and the peptides are protected from enzymatic degradation. 31Although, in the case of cyclin D, the inhibitory effects of our previously established UCNPs (NaGdF 4 :Yb 3+ , Er 3+ ) conjugated with cyclin D-specic peptides were not as satisfactory as expected, 32 we have rationalized that this could be ascribed to unwanted cross-binding among various cyclins and their cyclin-dependent kinases (CDKs), 33 in addition, the in vitro emission signal was also quenched aer bio-conjugation when compared to the unmod-ied NaGdF 4 :Yb 3+ , Er 3+ .With these factors in mind, we have now developed a new type of novel peptide-functionalized UCNP which can achieve termination and visualization of cancer cells concurrently through binding and inhibition of Plk1.
Herein, we report the synthesis of eight UCNPs, using two solid systems of different sizes-60 nm NaGdF 4 :Yb 3+ , Er 3+ (NaGdF 4 ) and 10 nm BaGdF 5 :Yb 3+ , Er 3+ (BaGdF 5 ).These two types of UCNPs are square, doped with Yb 3+ and Er 3+ at molar ratio 18 : 2 and coated with polyethyleneimine (PEI), PLHSpT (P 1 ), 13,14 PLHSD (P 2 ) and GGPLHSpT (P 3 ) respectively (Fig. 1).All three peptides have already shown strong binding with PBD in Plk1. 13P 3 , which contained an additional Gly-Gly motif, was used primarily to check whether any signicant impact on the binding affinity between the peptide (i.e.PLHSpT) and Plk1, as well as the anti-tumor activity was observed when the linking distance between the UNCPs and the peptides was increased in length and exibility.Comprehensive in situ (photophysical properties, protein pull down assay), in vitro (emission spectra, imaging, co-staining with GFP-Plk1, ICP-MS, cellular toxicity) and in vivo studies (imaging and solid tumor inhibition) in both cancer and normal cell lines have been conducted and the data are presented.Experimental results revealed that our 60 nm UNCPs have a better performance than the 10 nm UCNPs in terms of uptake efficiency and imaging quantity.In particular for NaGdF 4 -P 1 coated with the P 1 peptide; control experiments have illustrated that NaGdF 4 -P 1 is selective in its mode of action and it was demonstrated to have no anti-growth effect in normal cells.

Results and discussion
Through conjugation to various inorganic molecules such as nanoparticles, peptides can be effectively shielded from enzymatic degradation.In addition peptide coated UNCPs can be excited at 980 nm enabling their real-time imaging.In this study, we intended to exploit both these aforementioned properties coupled with Plk1 targeting to design and prepare nanosystems capable of simultaneous Plk1 inactivation and real time imaging, something that is not currently easily achievable.Two sets of UNCPs of different sizes have been conjugated with three kinds of Plk1 specic peptides and photophysical properties, cellular uptake efficiency, in vitro imaging and selectivity for Plk1 of these systems have been evaluated.The UCNPs have been characterized and veried by a series of experiments including spectroscopic measurements, microscopic analyses and competitive assays through protein binding and western blotting.In addition to investigating the selection of Plk1 specic peptides and different UCNPs, we also wished to probe the effects of the size and shape of UCNPs themselves.Thus we used 60 nm and 10 nm UCNPs as both of them have a similar crystal lattice and they are square in shape.Through this comprehensive study, we aimed to satisfy our hypothesis that Plk1 specic coated nanomaterials can act as promising systems in the future development of multimodal anti-cancer upconversion nano-drugs.2.1 Synthesis and characterization of nanoparticles conjugated with Plk1 (P n ) specic peptides The surface groups of our NaGdF 4 and BaGdF 5 were examined by FTIR.The corresponding IR transmission spectra of aminefunctionalized NaGdF 4 and BaGdF 5 are shown in Fig. S1.† The broad absorption band centered at 3459 cm À1 is due to the O-H/N-H stretching vibration.The IR absorption peaks at 1637 and 1436 cm À1 are attributable to the N-H bending of the amino group (NH 2 ) and the C-N bond stretching vibration.Therefore, the IR transmission spectrum fully supports the capping of poly(ethyleneimine) (PEI) on the surfaces of NaGdF 4 and BaGdF 5 .The amine-functionalized NaGdF 4 and BaGdF 5 were then modied into amide-functionalized NaGdF 4 -ene and BaGdF 5 -ene.The conversion of amine to amide was veried by the IR spectrum in Fig. S1(b).† The overall spectrum is similar to Fig. S1(a).† In fact, the absorption peaks at 1648 cm À1 newly appear under the stretching vibration of the C]O bond. 34ig. S1(c) † shows the IR spectrum recorded aer addition of the target specic peptide P n on the surfaces of NaGdF 4 and BaGdF 5 .The small absorption peaks around 3022 to 3417 cm À1 represent the N-H bond stretching while the broad peak at 3307 cm À1 stands for carboxylic acid.Hence, we had conrmed the successful coating of target specic peptide P n onto our NaGdF 4 and BaGdF 5 .
The shape and morphology of NaGdF 4 or BaGdF 5 and peptide coated NaGdF 4 -P n or BaGdF 5 -P n , on the other hand, were investigated by TEM.In Fig. 2a and b, the morphology of NaGdF 4 and BaGdF 5 was square in shape which corresponds to the cubic phase of NaGdF 4 and BaGdF 5 .This observation is similar to the previously reported synthesis of the cubic phase NaGdF 4 (ref.30) and BaGdF 5 (ref.22) which was veried by the SAED pattern of NaGdF 4 and BaGdF 5 as shown in Fig. S2.† From the high resolution TEM (HRTEM), a thin layer is observed (as indicated by red arrows) and the size distribution in TEM images reveals also that the overall sizes of Plk1 specic peptide coated NaGdF 4 -P n and BaGdF 5 -P n increase by 1-2 nm compared with the bare NaGdF 4 and BaGdF 5 , thereby justifying the conjugation of NaGdF 4 and BaGdF 5 with peptide P n .As demonstrated by the dynamic light scattering analysis, the size of the nanomaterials is also increased by 1-3 nm aer the peptide conjugation (Fig. S5 †).Besides, the surface charges of NaGdF 4 and BaGdF 5 should be different before and aer the peptide conjugation.In Fig. 3, there is clearly a drop of zeta potential which again qualies the successful peptide conjugation onto our nanomaterials.
The phase composition of NaGdF 4 or BaGdF 5 and peptide coated NaGdF 4 -P n or BaGdF 5 -P n was revealed by powder X-ray diffraction (XRD).In Fig. 2e and f and BaGdF 5 (JSPDF#24-0098) respectively.The diffraction peaks agree well with the standard pattern, with the matching results indicating that our NaGdF 4 and BaGdF 5 are present in the pure cubic phase with no other phase impurities.Aer the coating of Plk1 specic peptides on the surface of NaGdF 4 -P n or BaGdF 5 -P n , the XRD patterns show no changes in the 2-theta angles and this implies that the coating of peptides did not induce changes in the crystal structures of both NaGdF 4 and BaGdF 5 .Also, this observation was consistent with the non-crystalline nature of peptides.

Photophysical properties of upconversion nanoparticles
In Yb 3+ and Er 3+ co-doped nanoparticles, under 980 nm laser diode excitation, Yb 3+ ions possessing a large absorption crosssection, much longer excited state lifetime and the energy ( 2 F 5/2 / 2 F 7/2 ) well resonant with higher-energy states of other lanthanide ions can transfer energy continuously and efficiently to Er 3+ ions, leading to bright upconversion luminescence in the visible region.Fig. 4 shows the emission spectra of the peptide coated NaGdF 4 and BaGdF 5 (NPs), as well as the spectra of all the coated NPs indicating the same structures.The three emission peaks at 521 nm, 541 nm, and 650 nm arise from 2 H 11/2 / 4 I 15/2 , 4 S 3/2 / 4 I 15/2 , and 4 F 9/2 / 4 I 15/2 transitions respectively, which are ascribed to the two-photon upconversion process under 980 nm excitation with 340 mW and commonly present in nanomaterials doped with Yb 3+ and Er 3+ ions.The nearby peaks originated from transitions between other Stark levels of the corresponding excited energy levels and the ground state.In order to understand the upconversion mechanism well, we measured the decay time curves for the peptide coated UCNPs.The insets of Fig. 4a and b display the luminescence decay curves of the 4 S 3/2 / 4 I 15/2 transition of Er 3+ ions in NaGdF 4 -P n and BaGdF 5 -P n .Each of the down edges of the decay curves can be tted well into a single exponential function I(t) ¼ I 0 exp(Àt/s), where I 0 is the initial emission intensity at t ¼ 0, and s is the decay time.The lifetime of the 541 nm emission peak is $30 ms.Moreover, all the decay curves consist of rising edges, indicating that the population of 4 S 3/2 originates from the energy transfer of the sensitizer Yb 3+ ions.The upper level 4 F 7/2 was populated by two-step upconversion energy transfer of Yb 3+ ions and then nonradiatively relaxed to 4 S 3/2 , which resulted in the 541 nm emission.Thus the population of the 4 S 3/2 level is closely related to the phonon state density and the emission would be sensitive to temperature.From Fig. 4a, the nanoparticles of NaGdF 4 coated with different peptides show remarkably different emission intensities, thereby implying the obvious inuence of the surface coating on the upconversion properties of the nanoparticles.In short, the peptide coated nanoparticle emitting long lifetime bright visible light under near infrared excitation could be an ideal material to achieve termination and visualization of cancer cell simultaneously.

In vitro cell based assay of UCNPs
The application of peptides as potential anti-cancer agents is not a new concept; but, there are very limited methods for probing directly the uptake of such peptides in vitro.In our system, the lanthanide content in peptide-functionalized UCNPs offers an in-built and direct monitoring opportunity for ICP-MS evaluation of peptide uptake efficiency in vitro (Fig. 5).From the analysis of the ICP-MS results (Fig. 5), it can be seen that the uptake of two unfunctionalized nanoparticles is similar in both the cancer (HeLa) and the normal (liver derived QSG-7701) cells lines.A variation in the uptake rate occurs when the UNCPs are coated with peptides, especially peptides P 1 and P 3 (see arrows in Fig. 5).In fact, in general, there are three main differences between peptide coated and unfunctionalized UCNPs.First, coating with peptides facilitates cell permeability in both cancer (HeLa) and normal (liver derived QSG-7701) cells.Secondly, the 60 nm NaGdF 4 -P n shows slightly higher cellular uptake efficiency in comparison with 10 nm BaGdF 5 -P n in both cancer and normal cell lines.The underlying reason for this could simply be that in the 60 nm systems a larger surface area is available, and thus more peptides are coated on the nanoparticle.Lastly, peptides P 1 and P 3 helped to improve the selectivity of the uptake, most possibly due to the fact that hydrophilic peptides enhanced the distribution of nanoparticles.In addition, cellular Plk1 may also increase nanoparticle accumulation through specic peptide-protein binding (i.e.greater difference between uptake in cancer versus normal cells).From this later point it also appears that the specic sequence of the peptides, but not their actual length, is important in the binding interactions with Plk1.Given that our nanomaterials demonstrated interesting photo-physical properties within the biological window, high selectivity for Plk1 (by western blotting), and that they are more readily taken up by cancer cells, the potential of NaGdF 4 -P n or BaGdF 5 -P n to act as in vitro imaging bio-probes for Plk1 appeared very promising.To investigate this further, in vitro experiments were conducted in cervical carcinoma HeLa cells with two-photon microscopes (Fig. S3 †).NaGdF 4 -P n or BaGdF 5 -P n have, as expected, detectable emissions within the biological window (500 nm to 700 nm) via the near-infrared excitation at 980 nm.Such upconversion inside the cells can also help us to conrm the cell permeability of our UCNPs.The near-infrared (excitation) to near-infrared (emission) in vitro imaging of NaGdF 4 -P n or BaGdF 5 -P n can be correlated with the uptake results in Fig. S3.† Under the NIR excitation by the multi-photon confocal microscope, impressive green ( 4 S 3/2 / 4 I 15/2 ) and red (Fig. 6, 2 F 9/2 / 4 I 15/2 ) f-f emissions can be observed in HeLa cells due to the presence of NaGdF 4 -P n or BaGdF 5 -P n aer 24 hours incubation.The in vitro emission spectra were recorded with the lambda scan system on a confocal microscope (resolution ¼ 3 nm), the erbium emission bands 4 S 3/2 / 4 I 15/2 and 2 F 9/2 / 4 I 15/2 are found which is the same as our measurements with these UCNP-P n in aqueous solution.Obviously, the in vitro emission of 60 nm NaGdF 4 -P n is far more intense than the 10 nm BaGdF 5 -P n in vitro.This observation can be found from the in vitro imaging and its emission spectra.NaGdF 4 -P 1 and NaGdF 4 -P 3 show impressive red emission inside the cell.
The selectivity of the prepared UCNPs toward Plk1 binding was conrmed by co-staining experiments with GFP-Plk1 specic uorescent protein.(GFP-Plk1 protein is not available with the excitation at 980 nm).The co-staining experiments were carried out with the transient expression of the GFP-Plk1 protein in the cancer cells through plasmid transfection, followed by incubation with the UCNPs for another 24 hours.In Fig. 6a, the red emission from bare NaGdF 4 (l ex ¼ 980 nm) is not overlapped with the green emission from the GFP-Plk1 protein (l ex ¼ 488 nm).However, in Fig. 6b and d, the red emission from NaGdF 4 -P 1 and NaGdF 4 -P 3 overlapped with the green emission from GFP-Plk1 well (red circle and white arrow in Fig. 6b and d).
From in vitro microscopy, the up-conversion spectra showing the uptake behaviours of those nanomaterials are similar to what we found in ICP-MS uptake assays.The up-conversion red emission from NaGdF 4 -P 1 and NaGdF 4 -P 3 can be overlayed with the green GFP emission, further supporting the nano- material selectivity towards Plk1.In order to further verify the binding selectivity of NaGdF 4 -P n or BaGdF 5 -P n towards Plk1, the nanoparticle-peptide bound Plk1 proteins in the total cell lysate were examined using anti-Plk1 antibodies and western blotting by pull down assay.Among the various NaGdF 4 -P n or BaGdF 5 -P n tested, the NaGdF 4 -P 1 shows the highest binding affinity towards Plk1 (about 10-fold higher) as revealed by the graph of relative Plk1 intensity (Fig. 6m and n).Therefore, we propose that by conjugating the Plk1 specic peptide P 1 onto the surface, the cellular uptake of our UCNPs should be much higher rather than conjugating other peptides (P 2 and P 3 ).Hence, NaGdF 4 -P 1 is the best candidate for in vivo imaging nanomaterials compared with the other Plk1 specic peptide analogues.

Inhibition of cell viability and cancer cell cycle interruption
Plk1 expression in different kinds of tumor cells is higher than that in normal healthy cells.Therefore, before further studies and investigation of applications, the inhibitory effect of NaGdF 4 -P n or BaGdF 5 -P n on cancer cells had to be investigated.Firstly, the cell viability was examined using an MTT assay in various cell lines (details are given in the Experimental section).Cell inhibition rates were calculated via viability related absorbance.Fig. 7 shows that NaGdF 4 -P n or BaGdF 5 -P n exerted a higher inhibitory effect on cell viability, in comparison with the results using NaGdF 4 -PEI and BaGdF 5 -PEI in tested cell lines.Among the different Plk1 specic peptides conjugated to the NaGdF 4 -P n or BaGdF 5 -P n systems, NaGdF 4 -P n , showed a better inhibitory effect on cancer (HeLa and HK-1) cells in contrast to BaGdF 5 -P n as highlighted in Fig. 7.The results from the MTT assays carried out in normal cell lines, namely MRC-5 and QSG-7701, demonstrated the selectivity of NaGdF 4 -P n or BaGdF 5 -P n towards cancer cell lines (which overexpress Plk1).In fact, there is a comparable amount of asprepared nanoparticles taken up by HeLa cells and QSG-7701 cells but the inhibitory effect on HeLa is much higher than that on QSG-7701, thereby proving that our NaGdF 4 -P n or BaGdF 5 -P n can selectively kill cancer cells but remained relatively nontoxic to the normal cells.The IC 50 values of NaGdF 4 -P n or BaGdF 5 -P n are provided in ESI Table S1.† Furthermore, cell cycle interruption was evaluated by ow cytometry.All of the Plk1 specic peptide coated NaGdF 4 -P n or BaGdF 5 -P n (0-100 mg mL À1 ) and control (NaGdF 4 -PEI or BaGdF 5 -PEI) had been tested for the inhibitory activity and the data are shown in Fig. 7e.Normally, the cell phase distribution of usual cell division is $67%, $10% and $23% in G1, G2 and S phases respectively.However, when treated with various NaGdF 4 -P n or BaGdF 5 -P n , only the NaGdF 4 -P 1 showed a signicant G2 phase arrest at a dosage concentration of 100 mg mL À1 (indicated with statistical signicance) while other NaGdF 4 -P n or BaGdF 5 -P n showed milder cell cycle interruption.

In vivo bio-distribution and mouse xenogra tumor inhibition
In order to study the metabolism of NaGdF 4 -P n or BaGdF 5 -P n in an animal model, gadolinium concentration in different mouse tissues was analyzed with ICP-MS.Fig. 8a-c show that NaGdF 4 -Fig.8 (a-c PEI and NaGdF 4 -P 1 accumulate in mice spleen and liver 6 hours aer caudal vein injection, while BaGdF 5 -PEI shows low accumulation in these organs.The reason for this is that BaGdF 5 -PEI was quickly metabolized and excluded.Since liver and spleen are the primary organs responsible for detoxication, our nanomaterials were captured and precipitated in these organs.Interestingly, compared with NaGdF 4 -PEI, NaGdF 4 -P 1 not only accumulates in the spleen and liver, but it is also excreted in high concentration in the urine which suggests a longer retention time in the body (Fig. 8a and c).In addition, mouse xeno-gra tumor inhibition was also investigated using our nanomaterials which were subcutaneously injected around the tumor.Fig. 8d clearly shows that treatment with NaGdF 4 -P n (especially NaGdF 4 -P 1 ) results in tumor shrinkage while BaGdF 5 -P n shows only a slight effect on the tumor.Therefore, the tumor inhibition performance of the NaGdF 4 -P n systems was better than that of the BaGdF 4 -P n systems which is completely in line with the results from the in vitro cell line based inhibition studies.

Synthesis of NaGdF 4 and BaGdF 5 upconversion nanoparticles
Water-dispersed NaGdF 4 was prepared by the one-step hydrothermal method.Typically, 1 mmol of NaCl and 1 mmol of Gd(NO 3 ) 3 (0.5 M), Yb(NO 3 ) 3 (0.5 M) and Er(NO 3 ) 3 (0.1 M) with the molar ratio of 78 : 20 : 2 were dissolved in 20 mL of ethylene glycol (EG) and stirred for 30 min.Aer capping with polyethyleneimine (PEI, branched 25 kDa), approximately 10 mL of EG containing 5.5 mmol of NH 4 F was then added to the above solution and stirred for another 30 min.The obtained mixture was then transferred into a 50 mL stainless Teon-lined autoclave, which was sealed and kept at 190 C for 24 h.Aer that, the autoclave was cooled to room temperature naturally and the reaction mixture was separated through centrifugation.Ethanol and DI-water were used to wash the precipitate several times and dried at 70 C for 12 h to obtain the as-prepared samples.As a result, NaGdF 4 with PEI as the surface modier was prepared by the hydrothermal method.The synthesis of BaGdF 5 is similar to the above procedure except for replacing NaCl with BaCl 2 .The zeta-potential for the NaGdF 4 or BaGdF 5 colloidal solution was measured to be around +23.6 mV and +31.5 mV respectively, indicating the successful conjugation of positively charged PEI on the surface of NPs.

Synthesis of Plk1 specic peptides
The three peptides were obtained by stepwise elongation of the peptide chain by the method outlined below.For peptide P 1 , 500 mg of the rink amide resin (0.45 mmol g À1 loading) was suspended in a 20% solution of piperidine in DMF, stirred for 20 min at room temperature and washed with DMF prior to subsequent steps.The rst amino acid to be coupled, Fmoc-(Bnp)-Thr-OH (420 mg, 4 equiv.),was dissolved in DMF and coupled to the resin in the presence of PyBOP (645 mg, 4 equiv.)and NMM (135 mL, 4 equiv.)with the use of microwave chemistry as described previously. 15,32Other Fmoc amino acid derivatives, Fmoc-Ser( t Bu)-OH (420 mg, 4 equiv.),Fmoc-His(Trt)-OH (438 mg, 4 equiv.),Fmoc-Leu-OH (350 mg, 4 equiv.),and Fmoc-Pro-OH (320 mg, 4 equiv.),were connected to the resin using an analogous synthetic strategy.Following the nal removal of the Na-Fmoc group, the peptide-resin was washed with DMF.The free peptide was obtained by cleavage through the use of 9 mL of TFA in the presence of 750 mL of water and 750 mL of TIPS according to the standard procedure.Aer cleavage, the peptide was puried by preparative reverse phase HPLC.The main peptide-containing fractions were collected and lyophilized.The purity of all nal products was conrmed by analytical reverse phase HPLC.P 2 and P 3 were obtained and puried by the same methodology as the preparation of P 1 .

Characterization
Powder X-ray diffraction (XRD) patterns of the as-prepared NaGdF 4 or BaGdF 5 and NaGdF 4 -P n or BaGdF 5 -P n were recorded using a Rigaku smart lab 9 kW (Rigaku Japan) with Cu-Ka radiation (l ¼ 1.5406 Å).The step scan covered the angular range from 25 to 60 in steps of 0.02 .The morphology of the as-prepared samples was characterized by using a Tecnai G2 20 S-TWIN Transmission Electron Microscope operated at 200 kV.A fourier transform infrared (FT-IR) spectrometer (Nicolet Avator 630) was used to record the infrared spectra of the samples by using a KBr pellet technique.The powder materials were pressed into a tungsten mesh grid and installed in situ FT-IR transmittance cell, and the samples were degassed in a vacuum system with a residual pressure of less than 3 Â 10 À4 Torr at room temperature.UC spectra were recorded using an FLS920P Edinburgh analytical instrument apparatus equipped with a 980 nm diode laser as an UC pump source.

Study of cellular uptake of NaGdF 4 -P n or BaGdF 5 -P n by ICP-MS
To measure the intracellular concentration of the NaGdF 4 -P n or BaGdF 5 -P n . 1 Â 10 5 cells were plated in each well and incubated with the complex for at 100 mg mL À1 , which was the same as in vitro imaging.Aer co-incubation, the cell culture medium containing NaGdF 4 -P n or BaGdF 5 -P n was removed and the exposed cells were further washed with 1 mL PBS 3 times to remove NaGdF 4 -P n or BaGdF 5 -P n adhering to the outer cell membrane and the bottom of the confocal dishes.Then the cells were trypsinized and dispersed into 1 mL of culture medium.The exposed cells were collected by centrifugation at 4000 rpm and the cell pellet was digested in 100 mL of concentrated HNO 3 (69%) at 70 C for 3 hours.The cellular uptake of NaGdF 4 -P n or BaGdF 5 -P n was determined using an Agilent 7500 series of inductively coupled plasma mass spectroscopy (ICP-MS).All ICP experiments were performed in triplicate and the values obtained were averaged.The concentration of Gd per cell was calculated by determining the concentration of Gd in the cell lysate by ICP-MS and then dividing it by the number of cells which counted using a haematocytometer.

In vitro bio-imaging
To investigate the suitability of the obtained and Plk1 specic peptides coated UCNPs as bio-probes, imaging of HeLa cells (1 Â 10 5 ) was conducted.Aer the incubation with NaGdF 4 -P n or BaGdF 5 -P n for 24 hours, the cells were washed by PBS 3 times before imaging.The in vitro imaging was performed on a confocal laser scanning microscope, Leica TCS SP5, equipped with a Ti:sapphire laser (Libra II, coherent).The samples were excited by a 980 nm wavelength laser.

Plk1 GFP co-staining in vitro bio-imaging
To validate the binding between NaGdF 4 -P n or BaGdF 5 -P n and Plk1, GFP co-staining experiment was carried out.Plk1 full length cDNA was PCR amplied from pOTB7-Plk1 and inserted into multiple cloning sites (MCS) of pEGFP-C3 (CLONETECH, #6082-1) using restriction digestion sites of XhoI/EcoRI for GFP-Plk1 expression.The positive recombinant was selected through kanamycin proved by sequencing.Plasmids were puried using a StarPrep Plasmid Miniprep Kit (Genstar).For transfection, a 3.5 cm dish of HeLa cells (with 70-80% conuence) was transfected with 8 mg DNA using lipofectamine2000 according to the manufacturer instructions.Then, 500 mL of the lipofectamine 2000 and plasmid were added to the confocal dish with 500 mL cell medium.Aer incubation for 8 hours, 100 mg mL À1 of NaGdF 4 -P n or BaGdF 5 -P n was added and incubated for 24 hours.The imaging procedures and parameters were the same as given in the previous in vitro bio-imaging section.

MTT inhibition rate assay
1 Â 10 5 of HK-1, HeLa, QSG-7701 and MRC-5 were treated with NaGdF 4 -PEI or BaGdF 5 -PEI and NaGdF 4 -P n or BaGdF 5 -P n for 24 hours.Then the cells were treated with nanoparticles at 4 concentrations (100 mg mL À1 , 50 mg mL À1 , 20 mg mL À1 and 5 mg mL À1 ).The cell monolayers were rinsed with phosphate buffer saline (PBS) once and incubated with 0.5 mg mL À1 MTT solution.The cellular inhibitory potency of the nanoparticles was examined by the formation of formazan aer addition of MTT (3-(4,5-dimethylthiazol-2-yl)-2 and 5-diphenyltetrazolium bromide) for 4 hours to allow formazan production during cell metabolism.Aer that, formazan was fully dissolved in DMSO through oscillating.Finally, the absorbance of solution was measured with a Biotek Power wave xsMicroplate Reader at a wavelength of 570 nm and 690 nm.

Antibodies and western blotting
Anti-b-actin (#66009-1-Ig) antibodies were purchased from Proteintech Group, Inc. Anti-CDK1 (#9116) and anti-Plk1 (#4535) antibodies were from Cell Signaling Technology.HeLa cells in a 12 well-plate were treated with nanoparticles (1, 10, 50, 100 mg mL À1 each) for 24 hours, and then rinsed in PBS twice aer the culture medium was discarded.Direct lysis in 50 mM Tris-HCl, pH 6.8, 1% SDS, 10% glycerol and thorough denaturation in a boiling water bath were followed.The supernatant was collected as a clear lysate aer centrifugation steps.Equal amounts of total proteins were loaded onto the 10% nitrocellular membranes, which was followed by western blotting.All the results nally were analyzed using a Gel-Pro analyzer and GraphPad Prism 5 soware.

Pull down assay
Five 10 cm dishes of cells were lysed in RIPA buffer (50 mM Tris pH ¼ 7.4, 150 mM NaCl, 1% NP-40, phenylmethanesulfonyl uoride (PMSF) 1 mM, cocktail protease inhibitors) for 15 minutes on ice, and then centrifuged at 13 000 rpm for 20 minutes.The clear supernatant was collected and the total 5 mL cellular lysate was divided into eight samples that were subjected to incubation with the NaGdF 4 or BaGdF 5 nanoparticle series (100 mg each) at 4 C for 2 hours.Nanoparticle precipitated proteins were collected by short centrifugation and washed with RIPA buffer three times to remove unbound proteins.Finally, the pull-down proteins underwent SDS-PAGE separation followed by antibody detection.

Animals and ethics
BALB/c nude mice were provided by Guangdong Medical Laboratory Animal Center, with the licensing number of SCXK-2008-0002 approved by Guangdong Science and Technology Department.Mice were kept in the isolated bio-safety facility for specic pathogen-free animals.All animal feeding and experiments were performed according to the Procedures for the Care of Laboratory Animals approved by the Shenzhen Institute of The Hong Kong Polytechnic University (with Use Permit of Experiment Animal license SCXK-2010-0062).

Mouse xenogra tumor inhibition and ICP-MS biodistribution analysis
For tumor inhibition assays, human HeLa cells were trypsinized, harvested and resuspended in DMEM. 1 Â 10 6 cells in 100 mL DMEM were injected s.c.into the anks of the mice (6 weeks old).Aer 14 days when the tumors were about 150 mm 3 , the animals were randomized into groups for experiments.NaGdF 4 and BaGdF 5 series of nanoparticles of Plk1 (500 mg for a tumor) were injected around the tumors and the mice were kept for another 7 days.Finally, the tumors were extracted and imaged.For ICP-MS bio-distribution analysis, 200 mg nanoparticles were i.p. injected into the mice, and the tissues as well as urine aer 24 hour post-injection were collected.Tissues were digested with nitric acid and subjected to ICP-MS for measuring the concentrations of Gd.

Conclusions
Plk1 has emerged as a promising target for the development of new anti-cancer agents.The polo box domain (PBD) of Plk1 provides an excellent target for the development of molecules to inhibit Plk1 activity and as such we have rationally designed multifunctional lanthanide-doped upconversion nanomaterials to exploit this.Specically our aim was to design novel nanomaterials that could deliver Plk1 specic peptides into cancer cells causing Plk1 inhibition while simultaneously allowing direct NIR imaging and monitoring.NaGdF 4 :Yb 3+ , Er 3+ (NaGdF 4 ) and BaGdF 5 :Yb 3+ , Er 3+ (BaGdF 5 ) nanoparticles of two different sizes (60 nm and 10 nm, respectively) were prepared and then conjugated with Plk1 specic peptides (-P 1 ¼ PLHSpT, -P 2 ¼ PLHSD, and -P 3 ¼ GGPLHSpT) to generate novel nanomaterials.Systematic and comparative studies of the nanomaterials prepared were carried out, and structural morphology, phase composition, upconversion photophysical properties, cellular uptake efficiency, cytotoxicity, tumor selectivity, and Plk1 binding affinity and inhibitory potency were also investigated in detail.The experimental results obtained showed that several of the nanomaterials prepared could achieve our original dual-function aim of anti-cancer activity and NIR imaging.Of the novel bio-conjugated nanomaterials investigated in our study, NaGdF 4 -P 1 , demonstrated considerable promise as a system for further development.NaGdF 4 -P 1 emerged as the best candidate for Plk1 binding and imaging and it could effectively exert cell cycle G2/M arrest and thus tumor inhibition both in vitro and in vivo.It is also worth noting that NaGdF 4 -P 1 is selective in its mode of action and that no anti-growth effects on the normal cells lines used in this study were observed.In summary, this study represents one of few examples where systems have been developed that are capable of both Plk1 inhibition and simultaneously direct NIR imaging and monitoring of cancer cells.

Fig. 2
Fig. 2 Surface characterization by (a and b) TEM analysis; (c and d) size distribution and (e and f) XRD pattern analysis.

Fig. 3 Fig. 4
Fig. 3 Peptide conjugation leads to a drop of zeta potential since peptides are relatively more negative compared with the PEI on NaGdF 4 or BaGdF 5 .

Fig. 5
Fig. 5 HeLa and QSG-7701 cellular uptake profile of NaGdF 4 -P n and BaGdF 5 -P n by ICP-MS.The y-axis is the cellular uptake concentration at the ppt level per cell.

Fig. 6
Fig. 6 In vitro images of (a-d) expressed GFP-Plk1; (e-h) upconversion emission of NaGdF 4 -PEI, -P 1 , -P 2 , -P 3 respectively and (i-l) their overlay images in HeLa cells.(m and n) The pull down assay compared the binding affinity among NaGdF 4 -P n or BaGdF 5 -P n towards Plk1 and their relative Plk1 intensity respectively.

)
Fig. 8 (a-c) In vivo uptake profiles of NaGdF 4 -PEI, NaGdF 4 -P 1 and BaGdF 5 -PEI analyzed by ICP-MS respectively.(d) Tumor inhibition effect of NaGdF 4 -P n and BaGdF 5 -P n via subcutaneous injection around the tumor.(e) Xenograft tumor weight after the treatment of NaGdF 4 -P n and BaGdF 5 -P n .

P 1 :
PLHSpT MALDI-TOF HRMS ([M + 1] + , m/z): calcd for C 24 H 42 N 8 O 10 P, 633.27; found for [M + H] + , 633.26; P 2 : PLHSD MALDI-TOF HRMS ([M + 1] + , m/z): calcd for C 23 H 37 N 8 O 8 , 568.37; found for [M + H] + , 568.45; P 3 : GGPLHSpT MALDI-TOF HRMS ([M + 1] + , m/z): calcd for C 28 H 47 N 10 O 12 P, 746.31; found for [M + H] + 747.330.3.4Bio-conjugation of Plk1 specic peptides onto NaGdF 4 or BaGdF 5 50 mg PEI NaGdF 4 nanocrystals were added to 744 mg DCC, 200 mg acrylic acid and 266 NHS in DMF to react at room temperature overnight.The excess DCC, acrylic acid and NHS were removed by washing with DMF and methanol.The precipitate was then vacuum dried.30.8 mg DCC, 15.9 mg NHS and 40 mg 3-(tritylthio)propanoic acid were added to 50 mg Plk1 peptide in DMF and stirred at room temperature overnight.Excess DCC, NHS and 3-(tritylthio)propanoic acid were removed by washing with DMF and methanol.The resin of the modied Plk1 peptide was removed by reacting with TFA, TIS and de-ionized water in the ratio of 90 : 5 : 5 for 4 hours at room temperature.Finally, the modied NaGdF 4 nanocrystal and Plk1 peptide reacted with each other in PBS buffer at pH 7.4 at room temperature overnight.The product was washed with water and methanol three times and vacuum dried overnight.The bioconjugation of Plk1 specic peptides (P n ) onto BaGdF 5 is similar to that onto NaGdF 4 except for replacing NaGdF 4 with BaGdF 5 .The characterization of the NaGdF 4 -P n or BaGdF 5 -P n was performed with FTIR transmittance and zeta potential analysis.

3. 11
Flow cytometry 10 6 cells were harvested by trypsine digestion, washed twice in PBS, and then xed in 70% ethanol for 2 hours at 4 C. Aer removal of the ethanol, cells were resuspended in PBS and then subjected to PI staining (1% triton-100, 50 mg mL À1 RNase A, 20 mg mL À1 propidium iodide) for 40 minutes at 37 C. Cell phase distributions were analyzed in BD FACS Calibur and the results were processed by Flowjo 7.6.1,shown as the offset histogram and column.